Journal of Experimental Botany

Battey, Nicholas H.

doi: 10.1093/jxb/erg075pmid: 12493845

Abstract To a remarkable extent current science, and the attitudes of current scientists, correspond to an idea conceived by Francis Bacon 400 years ago. Bacon dreamed of using science to subjugate nature for mankind. This is illustrated here by a science-eye's view of bark, a feature of plants particularly apparent in January. But much of the beauty and value of plants is missed by Baconian analysis. The touching story of Daphne's metamorphosis into a laurel, her skin becoming bark, shows one such dimension, highlighted in the sculpture by Bernini. This content is only available as a PDF. © 2003

Miyao, Mitsue

doi: 10.1093/jxb/erg026pmid: 12493846

Abstract The majority of terrestrial plants, including many important crops such as rice, wheat, soybean, and potato, are classified as C3 plants that assimilate atmospheric CO2 directly through the C3 photosynthetic pathway. C4 plants, such as maize and sugarcane, evolved from C3 plants, acquiring the C4 photosynthetic pathway in addition to the C3 pathway to achieve high photosynthetic performance and high water‐ and nitrogen‐use efficiencies. Consequently, the transfer of C4 traits to C3 plants is one strategy being adopted for improving the photosynthetic performance of C3 plants. The recent application of recombinant DNA technology has made considerable progress in the molecular engineering of photosynthetic genes in the past ten years. It has deepened understanding of the evolutionary scenario of the C4 photosynthetic genes. The strategy, based on the evolutionary scenario, has enabled enzymes involved in the C4 pathway to be expressed at high levels and in desired locations in the leaves of C3 plants. Although overproduction of a single C4 enzyme can alter the carbon metabolism of C3 plants, it does not show any positive effects on photosynthesis. Transgenic C3 plants overproducing multiple enzymes are now being produced for improving the photosynthetic performance of C3 plants. Key words: C4 photosynthesis, gene evolution, phosphoenolpyruvate carboxylase, transgenic plants. Received 27 May 2002; Accepted 5 September 2002 Introduction Terrestrial plants are classified into three major photosynthetic types, namely, C3, C4 and Crassulacean acid metabolism (CAM) plants, according to the mechanism of their photosynthetic carbon assimilation. About 90% of terrestrial plant species, which include major crops such as rice (Oryza sativa), wheat (Triticum aestivum), soybean (Glycine max), and potato (Solanum tuberosum), are classified as C3 plants, and they assimilate CO2 directly through the C3 photosynthetic pathway, also called the Calvin cycle or the photosynthetic carbon reduction (PCR) cycle. C4 and CAM plants possess a unique photosynthetic pathway, in addition to the C3 pathway, which allows them to adapt to specific environments. While C3 plants grow well in temperate climates, CAM plants such as stonecrops and cactus adapt to extreme arid conditions, but their photosynthetic capacity is very low (Black, 1973). By contrast, C4 plants such as maize (Zea mays) and sugarcane (Saccharum officinarum) adapt to high light, arid and warm environments and achieve higher photosynthetic capacity and higher water‐ and nitrogen‐use efficiencies compared with C3 plants (Black, 1973). Both C4 and CAM plants evolved from ancestral C3 species in response to changes in environmental conditions that caused a decrease in CO2 availability. C4 plants evolved in response to the low atmospheric CO2 concentrations, while the CAM plants evolved either in response to the selection of increased water‐use efficiency or for increased carbon gain (Ehleringer and Monson, 1993). In leaves of C3 plants, all of the photosynthetic reactions from the capture of solar light energy to assimilation of carbon into carbohydrates (triosephosphates) proceed in the chloroplasts of the mesophyll cells (Fig. 1A). The primary CO2 fixation step in the C3 pathway is catalysed by ribulose‐1,5‐bisphosphate carboxylase/oxygenase (Rubisco). However, Rubisco also reacts with O2 at its catalytic site (oxygenase reaction), leading to photorespiration. Photorespiration plays a role in protecting photosynthesis from photoinhibition (Osmond and Grace, 1995), but it wastes fixed carbon as released CO2 and decreases the efficiency of photosynthetic CO2 assimilation in C3 plants (Leegood et al., 1995). Under current atmospheric conditions (0.036% CO2, 21% O2), up to 50% of the fixed carbon is lost by photorespiration. C4 plants have evolved the C4 photosynthetic pathway, a mechanism to concentrate CO2 at the site of the reaction of Rubisco, and thereby overcame photorespiration. This CO2‐concentrating mechanism, together with modification of leaf anatomy, enabled C4 plants to achieve high photosynthetic efficiency (Hatch, 1987). Leaves of C4 plants have two types of photosynthetic cells, the mesophyll and bundle sheath cells that contain chloroplasts of different functions. While all the photosynthetic enzymes are confined in the mesophyll cells in C3 plants, they are localized in the mesophyll and/or bundle sheath cells in C4 plants. The enzymes involved in the C3 pathway are located in the chloroplasts of the bundle sheath cells while those involved in the C4 pathway (C4 photosynthetic enzymes) in the mesophyll and/or bundle sheath cells (Fig. 1B). The C4 pathway consists of three key steps: (i) the initial fixation of CO2 in the cytosol of the mesophyll cells by phosphoenolpyruvate carboxylase (PEPC) to form a C4 acid, oxaloacetate (OAA), (ii) decarboxylation of a C4 acid in the bundle sheath cells to release CO2, and (iii) regeneration of the primary CO2 acceptor phosphoenolpyruvate (PEP) (Hatch, 1987; Fig. 1B). As a whole, one molecule of CO2 is pumped up from the cytosol of the mesophyll cells into the vicinity of Rubisco in the chloroplast of the bundle sheath cells, consuming two molecules of ATP. The decarboxylation reaction is catalysed by one or more of the three enzymes, namely, NADP‐malic enzyme (NADP‐ME), NAD‐malic enzyme (NAD‐ME), and phosphoenolpyruvate carboxykinase (PEP‐CK), and C4 plants are classified into three subtypes depending on the major decarboxylation enzyme. The C4 acid exported from the mesophyll to bundle sheath cells are also different. Before being exported, OAA is reduced to malate by NADP‐malate dehydrogenase (NADP‐MDH) or transaminated to aspartate by aspartate aminotransferase (AspAT) in the NADP‐ME type and the NAD‐ME and PEP‐CK types, respectively. Regeneration of PEP is catalysed by pyruvate, orthophosphate dikinase (PPDK) located in the mesophyll cell chloroplasts in all subtypes, although PEP‐CK in the bundle sheath cell cytosol also participates in this process in the PEP‐CK type. Maize and sugarcane use NADP‐ME for the decarboxylation and these are classified as the NADP‐ME type. The rate‐limiting steps of the C4 pathway can be assessed from the control coefficients (Cj) determined for the individual enzymes. Cj is defined as the ratio between the fractional change in metabolite flux and the fractional change in enzyme activity, and a Cj of one indicates that the flux is fully controlled at this step while the value of zero indicates no control over the flux (ap Rees and Hill, 1994). The Cj values for C4 enzymes can be determined from the plot of the CO2 assimilation rate against enzyme activity, which has been altered by mutation and/or transgenic techniques. The results indicate that the C4 pathway is controlled by reactions of Rubisco, PEPC and PPDK (reviewed in Matsuoka et al., 2001). Under saturating illumination and ambient CO2, the Cj values for Rubisco, PEPC and PPDK were 0.5–0.6, 0.35, and 0.2–0.4, respectively. By contrast, two other enzymes examined so far appear to exert little or no control: the Cj values for NADP‐MDH of Flaveria bidentis (NADP‐ME type) and for the decarboxylation enzyme of Amaranthus edulis (NAD‐ME type) were almost zero. The activities of the rate‐limiting enzymes are strictly regulated in the leaves of C4 plants. The activity of Rubisco is controlled by the Rubisco activase as it is in C3 plants (Salvucci et al., 1987). Those of PEPC and PPDK are regulated through reversible protein phosphorylation by their specific regulatory proteins, being up‐regulated in the light (Burnell and Hatch, 1985; Vidal and Chollet, 1997). Since the discovery of the C4 pathway, it has been postulated that the transfer of C4 traits to C3 plants should improve the photosynthetic performance of C3 plants. Initially, conventional hybridization between C3 and C4 plants was carried out. This approach was available only in several plant genera and most C3‐C4 hybrids were infertile (Brown and Bouton, 1993). Another approach that has been adopted in the last ten years involves the use of recombinant DNA technology. With this technology, understanding of the evolution of C4 photosynthetic genes has been expanded and it is now possible to express C4 enzymes at high levels and in desired locations in the leaves of C3 plants. The evolution of C4 genes together with techniques with which to overproduce C4 enzymes in the leaves of C3 plants is summarized here. The regulation and physiological impacts of overproduced C4 enzymes in transgenic rice plants are also presented. The physiological impacts of the overproduction in potato, tobacco (Nicotiana tabacum) and Arabidopsis thaliana as well as rice have previously been reviewed in detail by Häusler et al. (2002). Evolution of C4 photosynthetic genes C4 photosynthetic genes had previously been considered to be specific for C4 plants, since the activities of the corresponding enzymes are low in C3 plants (Hatch, 1987) and their kinetic properties are usually different from those of C4 enzymes (e.g. for PEPC, see Svensson et al., 1997; Dong et al., 1998). However, recent comparative studies have revealed that C3 plants have at least two different types of genes, one encoding enzymes of ‘housekeeping’ function and the other very similar to the C4 genes of C4 plants, though expression of the latter is very low or even undetectable in C3 plants. Based on this finding, it is postulated that the C4 genes evolved from a set of pre‐existing counterpart genes in ancestral C3 plants, with modifications in the expression level in the leaves and kinetic properties of enzymes (Ku et al., 1996). Hereafter, the C4 genes in C4 plants and their homologues in C3 plants are designated C4‐specific and C4‐like genes, respectively. In addition to C4‐specific or C4‐like genes, both C3 and C4 plants have other homologous genes for the housekeeping function. These are designated as C3‐specific genes. The number of homologous genes and the evolutionary origins of C4‐specific genes are different among C4 enzymes and plant species (Monson, 1999). By contrast, modifications of C4‐like genes required for functioning in the C4 pathway probably share common features in all the C4‐specific genes examined so far. In the following, the evolutionary origin of the maize C4‐specific PPDK gene is considered as an example. Evolution of the maize C4‐specific PPDK gene Maize has three different isoforms of PPDK, namely, the chloroplastic isoform involved in the C4 pathway and two cytosolic isoforms (Sheen, 1991). The chloroplastic and one cytosolic isoforms are encoded by a single gene that has a dual promoter system (Glackin and Grula, 1990; Sheen, 1991; Fig. 2). This gene (designated Pdk1 hereafter) has two transcription initiation sites and transcription from these sites is regulated by different promoters located at their respective 5′‐flanking regions. Transcription at the first initiation site produces large transcripts for the chloroplastic isoform, while that from the second site produces small transcripts for the cytosolic isoform. The large transcripts are expressed highly specifically in the mesophyll cells of green leaves at a high level and expression is induced by light, while the small transcripts are expressed at a high level in roots but at a low level in the mesophyll cells (Sheen, 1991). By definition, the genes encoding the chloroplastic and cytosolic isoforms in the maize Pdk1 are C4‐specific and C3‐specific genes, respectively: they have previously been designated C4ppdkZm1 and cyppdkZm1, respectively, by Sheen (1991). The second cytosolic isoform is encoded by a gene with a single promoter (previously designated cyppdkZm2 by Sheen (1991) but Pdk2 hereafter), which shows high homology to the C3‐specific gene in Pdk1. This gene is expressed at a very low level in the mesophyll cells (Sheen, 1991) and thus C3‐specific. The cytosolic isoform of PPDK accumulates significantly in kernels (Aoyagi and Bassham, 1984; Aoyagi and Chua, 1988), though it is uncertain which of the two genes is expressed in this organ. Rice also has three different isoforms of PPDK, and two different genes have been identified. One has a dual promoter system and encodes the chloroplastic and cytosolic isoforms (Imaizumi et al., 1997) and the other, with a single promoter, encodes the cytosolic isoform (Moons et al., 1998). The former gene is highly homologous to the maize Pdk1 (Imaizumi et al., 1997). It has 21 exons and the positions of introns are essentially the same as those in the maize gene, except that the exon 1 and 3 of the maize gene are split into two exons in the rice gene (Fig. 2). The deduced amino acid sequences are 88% homologous in the mature protein portion and 56% homologous in the transit peptide portion. In addition, this gene is expressed in rice plants essentially in the same way as the maize Pdk1 does in maize, except for the expression level of the large transcripts in green leaves: the large transcripts are expressed specifically in photosynthetic organs but at low levels, while the smaller ones in reproductive organs at high levels and roots at a low level (Imaizumi et al., 1997). Thus, this gene is a counterpart of the maize Pdk1. The other gene encoding the cytosolic isoform has been identified as a cDNA clone from rice roots (osppdka; Moons et al., 1998), but its genomic clone has not yet been isolated. Since the 3′ region of the cDNA is highly homologous to the exons in the 3′ region of the rice Pdk1 (Moons et al., 1998), it seems likely that this gene is a counterpart of the maize Pdk2. From the comparison between the maize and rice genes, the postulated evolutionary origin of PPDK genes is depicted in Fig. 3. From a single ancestral gene, two genes encoding a cytosolic isoform were derived. One was an ancestral Pdk2 gene, which would evolve to become the Pdk2 genes of C3 and C4 plants. The other was an ancestral Pdk1 gene with a single promoter. This gene subsequently evolved to become the Pdk1 gene with a dual promoter system in an ancestral C3 plant, acquiring a sequence for the transit peptide, and finally evolved to become the Pdk1 gene of a C4 plant by acquiring a mechanism(s) for high‐level expression. The evolutionary scenario for the C4‐specific PPDK gene might differ from this in the C4 species of Flaveria. Until now, only Pdk1 with a dual promoter system, but not Pdk2, has been identified in both C3 and C4 species of Flaveria (Rosche et al., 1994; Rosche and Westhoff, 1995). If Pdk2 were to be missing, the evolution in this species probably proceeded without gene duplication. Modifications required for the evolution of the maize C4‐specific PPDK gene have been investigated by comparing expression patterns of C4‐like and C4‐specific genes in rice and maize plants. When a reporter gene (GUS gene) was expressed in rice plants under the control of the promoter of the C4‐specific gene in the maize Pdk1 (–1032 to +71, relative to the transcription initiation site), GUS was expressed highly specifically in the mesophyll cells of green leaves at a high level and in a light‐responsive manner (Matsuoka et al., 1993). Its expression level was even higher than that under the control of the cauliflower mosaic virus 35S promoter. On the other hand, when the GUS gene was expressed in maize plants under the control of the promoter of the C4‐like gene in the rice Pdk1 (–1419 to +512, a region upstream from the initiation codon), GUS was expressed in both the mesophyll and bundle sheath cells at low levels, although its expression was light responsive (Nomura et al., 2000a). These results clearly show that a cis‐acting element(s) for light‐responsive expression is present in the rice promoter, but that those for cell‐specific and high‐level expression are missing and had to be acquired during the course of evolution from a C4‐like to a C4‐specific gene (Fig. 3). Some of these cis‐acting elements in the promoter of the maize C4‐specific gene have been identified (Sheen, 1991; Matsuoka and Numazawa, 1991; Imaizumi et al., 1997; Nomura et al., 2000a). Another important implication of the results is that trans‐acting elements (e.g. transcription regulators) required for the expression of the C4‐specific gene are present in the leaves of the C3 plant, rice. The expression pattern of the promoter of the C3‐specific gene in Pdk1, on the other hand, does not differ much between the maize and rice genes. The C3‐specific promoters from the maize and rice Pdk1 both directed expression of the GUS gene in non‐photosynthetic organs such as grains and roots in transgenic rice (Nomura et al., 2000b). Thus, modifications of Pdk1 required for the evolution from a C4‐like to a C4‐specific gene are relatively simple; namely, gain of the cis‐acting elements for cell‐specific and high‐level expression in the promoter region. As described later, however, cis‐acting elements for high‐level expression are not restricted to the promoter region. Evolution of other C4‐specific genes The same evolutionary scenario can be applied to the C4‐specific PEPC gene. The promoter of the maize C4‐specific gene (–1212 to +78) directed high‐level, mesophyll‐cell‐specific and light‐inducible expression of the GUS gene in transgenic rice (Matsuoka et al., 1994). The C4‐specific PPDK and PEPC are both located in the mesophyll cells of C4 plants. Quite recently, it has been found that C4‐specific genes for the enzymes located in the bundle sheath cells might evolve in similar ways. The promoter of the C4‐specific PEP‐CK gene of a turf grass Zoysia japonica (PEP‐CK type) directed specific expression of the GUS gene in the bundle sheath cells and vascular bundles in transgenic rice (M Nomura, M Matsuoka, unpublished results; Fig. 4). Similar results have been obtained with the C4‐specific gene for the mitochondrial AspAT of Panicum miliaceum (NAD‐ME type; M Nomura, M Matsuoka, unpublished results). It seems quite likely that, irrespective of the enzyme location and the subtype of the C4 pathway, gene modifications for the evolution of C4‐specific genes followed similar mechanisms. How to overproduce C4 enzymes in the mesophyll cells of C3 plants In leaves of C3 plants, photosynthesis and subsequent carbon and nitrogen metabolism proceed mainly in the mesophyll cells. To alter carbon metabolism in leaves of C3 plants, C4 enzymes have to be overproduced in these cells. From the evolutionary origins of the C4‐specific genes and the expression patterns of their promoters in the leaves of C3 plants, it was anticipated that the introduction of the intact C4‐specific genes into C3 plants would lead to high‐level and cell‐specific expression of C4 enzymes. In fact, this strategy was effective in overproducing C4 enzymes specific to the mesophyll cells of C4 plants. By contrast, to overproduce enzymes specific to the bundle sheath cells, conventional techniques with gene constructs containing a strong promoter fused to a cDNA were effective (Table 1). Enzymes located in the mesophyll cells of C4 plants The first trial of this kind was conducted with the intact maize PEPC gene expressed in transgenic rice plants (Ku et al., 1999). The maize gene of 8.8 kb that contained all exons and introns and its own promoter and terminator sequences was introduced into rice plants. As expected, the activity of PEPC in the leaf protein extract was greatly increased up to 110‐fold that of non‐transformants or 3‐fold the maize activity. The level of the PEPC protein accounted for 12% of total leaf soluble protein at most. The introduction of the intact maize gene was also effective in overproducing PPDK in rice leaves (Fukayama et al., 2001). The introduction of the intact maize Pdk1 of 7.3 kb increased the PPDK activity in rice leaves up to 40‐fold that of non‐transformants or about half of the maize activity. In a homozygous transgenic line, the PPDK protein accounted for 35% of total leaf soluble protein or 16% of total leaf nitrogen, much above the levels of foreign protein in transgenic plants reported previously. The C4‐specific gene of the maize Pdk1 was exclusively expressed in the leaves of these transgenic rice plants while the C3‐specific gene was expressed in grains, an indication that the maize Pdk1 is expressed in rice plants with an organ specificity similar to that in maize plants. To examine whether or not the promoter sequence of the C4‐specific gene is sufficient for the high‐level expression, the full‐length cDNA encoding the maize C4‐specific PPDK was expressed under the control of the promoter of the C4‐specific gene of the maize Pdk1 or the rice Cab promoter (Fukayama et al., 2001). Cab encodes the light‐harvesting chlorophyll‐binding protein and is expressed at high levels in photosynthetically active organs (Sakamoto et al., 1991). The introduction of these gene constructs, however, increased the activity of PPDK in rice leaves only up to several fold that of non‐transformants (Fukayama et al., 2001). Thus, the transcriptional activity of the C4‐specific promoter cannot be the prime reason for high‐level expression. The 5′‐ and 3′‐noncoding regions by themselves cannot lead to high‐level expression either, since the constructs containing the full‐length cDNA with these regions were not effective. Therefore, it is quite possible that, in addition to the promoter region, the presence of introns or the terminator sequence, or a combination of both, is required for high‐level expression. In transgenic rice plants containing the C4‐specific PEPC or PPDK gene, the levels of transcripts and protein and the activity of C4 enzyme in the leaves all correlated well with the copy number of the introduced gene (Ku et al., 1999; Fukayama et al., 2001). It was also found that the levels of transcripts in the leaves per copy of the maize C4‐specific gene were comparable in both maize and transgenic rice plants (Ku et al., 1999; Fukayama et al., 2001). As described above, the promoters of C4‐specific PPDK and PEPC genes have the cis‐elements for organ‐ and cell‐specific expression. It is likely that the maize C4‐specific genes behave in a qualitatively and also quantitatively similar way in both maize and transgenic rice plants. Overproduction by the introduction of the intact C4‐specific gene, however, seems to have some limitation in that transgenes from phylogenetically closely related plants have to be used to achieve high‐level expression of the C4 enzyme in C3 plants. The intact maize C4‐specific PEPC gene was not expressed at high levels in tobacco leaves, because of incorrect transcription initiation (Hudspeth et al., 1992). Not only incorrect initiation and termination of transcription, but also incorrect splicing could occur when genes from monocots are introduced into dicots (Goodall and Filipowicz, 1991). Thus, phylogenetic distance may hamper the expression of genes from C4 plants in the leaves of C3 plants. Conventional techniques for overproduction of transgenes, namely, the introduction of a chimeric gene containing cDNA for a C4 enzyme, fused to a strong promoter alone or together with enhancer sequences, can also increase the activity of C4 enzymes in the leaves of C3 plants, although the increase does not exceed several fold that of non‐transformants (for a review see Matsuoka et al., 2001). Enzymes located in the bundle sheath cells of C4 plants Since the intact C4‐specific genes for these enzymes would be expressed specifically in the bundle sheath cells of C3 plants, they cannot be used for overproduction in photosynthetically active mesophyll cells. More conventional techniques were applied and have proven successful. The expression of the maize C4‐specific NADP‐ME cDNA under the control of the rice Cab promoter increased the activity of NADP‐ME in rice leaves to 30‐ or 70‐fold that of non‐transformants (Takeuchi et al., 2000; Tsuchida et al., 2001). The level of the NADP‐ME protein was also increased up to several per cent of total leaf soluble protein. Such high‐level expression was unique to the cDNA for the C4‐specific NADP‐ME, and the expression of the cDNA for the rice C3‐specific isoform under the control of the same promoter increased the activity only some fold (Tsuchida et al., 2001). This observation suggests that expression of the rice C3‐specific NADP‐ME is suppressed at co‐ and/or post‐transcriptional levels by some regulation mechanisms intrinsic to rice, while that of the foreign C4‐specific isoform can escape from such suppression. The Zoysia C4‐specific PEP‐CK was also overproduced by introduction of a cDNA construct (M Miyao et al., unpublished results). Overproduction of C4 enzymes in a different intracellular compartment The C4 enzymes described above were all overproduced in the same intracellular compartment in C3 plants as in C4 plants, namely, PEPC and PEP‐CK in the cytosol and PPDK and NADP‐ME in the chloroplasts. The intracellular location of foreign enzyme can be altered by use of targeting signals. To overproduce PEP‐CK in the chloroplasts of the mesophyll cells of rice leaves, the cDNA of the C4‐specific PEP‐CK from Urochloa panicoides was fused to a sequence of the transit peptide for targeting to chloroplasts, and expressed under the control of the maize C4‐specific PEPC or PPDK promoter (Suzuki et al., 2000). The PEP‐CK activity of transgenic rice leaves reached about half of that in the Urochloa leaves. Similarly, bacterial enzymes were overproduced in the chloroplasts of transgenic C3 plants (Häusler et al., 2001; Panstruga et al., 1997). Factors affecting the expression levels of transgenes In general, expression of transgenes is hampered by many mechanisms including the positional effects (Gelvin, 1998), silencing (Gallie, 1998; Chandler and Vaucheret, 2001) and rearrangement (Hiei et al., 1994) of transgenes. During the course of the study of overproducing C4 enzymes, it was found that the rearrangement occurs frequently during the gene transfer mediated by Agrobacterium tumefaciens. A significant fraction of transgenic rice plants introduced with the intact maize C4‐specific gene showed activities of C4 enzymes comparable to or even lower that that of non‐transformants (Ku et al., 1999; Fukayama et al., 2001). DNA gel‐blot analysis of these low‐expressing lines showed that transgenes in all lines tested sustained partial deletion and/or chimeric linking (Fukayama et al., 2001). Such rearrangement is not peculiar to long transgenes with complex exon‐intron structures, and it did occur in five out of nine transgenic rice plants introduced with a cDNA construct of 4.4 kb (Miyao et al., 2001). It is possible that cis‐acting elements and/or the transit sequence are selectively deleted from an introduced gene, altering the level and/or location of a C4 enzyme in transgenic C3 plants. As described above, overproduction of C4 enzymes in C3 plants can be achieved by introducing appropriate gene constructs. It is also necessary to screen a number of transgenic plants to obtain a desired expression level of a C4 enzyme and to confirm the enzyme location in the leaves of C3 plants. Regulation and physiological impacts of C4 enzymes overproduced in C3 plants PEPC The activity of PEPC in higher plants is regulated by two different mechanisms; one the reversible protein phosphorylation of a conserved serine residue near the N‐terminus, and the other through various metabolite effectors such as glucose‐6‐phosphate (Glc6P), malate, aspartate, and glutamate (Vidal and Chollet, 1997). Upon phosphorylation, PEPC becomes more sensitive to the activator Glc6P and less sensitive to the feedback inhibitor malate, being more active in vivo (Vidal and Chollet, 1997). The phosphorylation itself is also inhibited by malate through the conformational change of PEPC (Bakrim et al., 1998). In leaves of C4 plants, PEPC is phosphorylated in the light and dephosphorylated in darkness and its activity is modulated in response to changes in light intensity (Vidal and Chollet, 1997). The maize PEPC expressed in transgenic rice leaves also underwent activity regulation via phosphorylation, but in an opposite manner (Fukayama et al., 2002). It remained dephosphorylated and less active during the daytime and became phosphorylated and more active in the night in transgenic rice leaves. Since the activity of the endogenous rice PEPC was also down‐regulated during the daytime, it is likely that both the maize and rice PEPC undergo activity regulation by the same mechanisms in rice leaves. Bacterial PEPC lacks the phosphorylation site (Vidal and Chollet, 1997) and can escape from down‐regulation via phosphorylation. Another issue affecting potential activity of foreign PEPC in the leaves C3 plants is the cytosolic concentrations of potential inhibitors and activators. The concentrations of inhibitors of higher plant PEPC are high in the cytosol of the mesophyll cells of C3 plants, about 1 mM for malate and around 40 mM for aspartate and glutamate (Heineke et al., 1991). Bacterial PEPC is also inhibited by aspartate, and in addition, it requires acetyl‐CoA as an activator (Chen et al., 2002). Taken together, the actual in vivo activity of foreign PEPC, either from higher plants or bacteria, in the leaves of transgenic C3 plants is lower than maximum extractable activities, especially when measured in the presence of activators. Until now, several transgenic C3 plants overproducing PEPC have been produced and analysed precisely (Matsuoka et al., 2001). All the transformants analysed so far show a higher level of malate (Hudspeth, et al., 1992; Kogami et al., 1994; Häusler et al., 1999) or OAA (Fukayama et al., 2002) in the leaves compared to that of non‐transformants, an indication that foreign PEPC is partly active in these transformants. Physiological impacts of the elevated PEPC activity reported to date have varied with no clear consensus. Among these, stimulation of respiration in the light and destabilization of stomatal opening have been observed in different plant species overproducing PEPC of different origin, namely, transgenic potato overproducing PEPC from Corynebacterium glutamicum (Gehlen et al., 1996) and transgenic rice overproducing the maize C4‐specific PEPC (Fukayama et al., 2002). The stimulated respiration is consistent with an anaplerotic function of the C3‐specific PEPC, which replenishes the tricarboxylic acid cycle with organic acids to meet the demand of carbon skeletons for amino acid synthesis (Champigny and Foyer, 1992). The effects on stomatal movement were observed only under non‐steady‐state conditions. Transient closure of stomata at the onset of gas‐exchange measurements and after a step increase in light intensity were reported in the transgenic potato and rice, respectively. Accelerated stomatal opening, by contrast, was observed in the transgenic potato, but not in the transgenic rice. PEPC has long been implicated in the synthesis of malate as an osmotically active solute in stomata (Assmann, 1993). At present, it remains obscure if foreign PEPC is expressed in guard cells of these transformants. In view of the critical function of PEPC in photosynthesis of C4 and CAM plants, some researchers had expected that overproduction of PEPC alone would improve the photosynthetic performance of C3 plants. A research group claims that the photosynthetic rate under saturating light can be greatly increased by overproduction of the maize PEPC in transgenic rice plants (Jiao et al., 2001). Their results, however, require reconsideration, since the correlation between the photosynthetic rate and the level of the PEPC protein in transgenic rice leaves has not yet been confirmed. Other groups all reported negative effects on photosynthesis, and the photosynthetic rate was slightly lowered by the overproduction. It has previously been reported that the O2 inhibition of photosynthesis was mitigated by overproduction of the maize PEPC, and suggested that the maize PEPC may participate in photosynthetic CO2 fixation in transgenic rice leaves (Ku et al., 1999). Later experiments, however, indicated that the initial CO2 fixation product, determined by 14CO2 labelling experiments with transgenic rice plants showing a 50‐fold elevation in extractable PEPC activity, was exclusively the C3 compound 3‐phosphoglycerate (Fukayama et al., 2002). The apparent reduction of O2 inhibition can be explained by more marked suppression of photosynthesis by overproduction of PEPC at 2% O2 than at 21% O2 (Matsuoka et al., 2000). In a recent paper (Agarie et al., 2002), it has been proposed that the lower photosynthetic rate resulted from the reduced capacity of regeneration of inorganic phosphate. However, a significant reduction of the photosynthetic rate was observed at very low intercellular CO2 concentrations and a low O2 concentration where the Rubisco activity but not the phosphate regeneration limits photosynthesis (Fukayama et al., 2002). It is more likely that the suppression of photosynthesis results from the enhanced respiration by elevated PEPC. Transgenic Arabidopsis overproducing PEPC from a cyanobacterium Synechococcus vulcanus has recently been reported (Chen and Izui, 2002). The Synechococcus PEPC has some advantages for raising the in vivo activity of PEPC in the leaves of C3 plants. It does not undergo activity regulation via phosphorylation or, in contrast to other bacterial PEPCs, is not inactivated by malate, and moreover, it does not require acetyl‐CoA for activation (Chen et al., 2002). Expression of this PEPC in Arabidopsis led to stunting and bleaching of leaf colour (Chen and Izui, 2002). Similar phenomena were observed in transgenic potato overproducing the potato enzyme that had been modified for a higher affinity for PEP and lowered sensitivity toward malate (Rademacher et al., 2002). It has been demonstrated that carbon flow in these transgenic potato plants was redirected from soluble sugars and starch to organic acids and amino acids (Rademacher et al., 2002). PPDK In leaves of C4 plants, the activity of PPDK is rapidly modulated in response to changes in light intensity by reversible protein phosphorylation, which is mediated by a bifunctional regulatory protein (Burnell and Hatch, 1985). Unlike phosphorylation of PEPC, PPDK is dephosphorylated in the light and dephosphorylated in darkness, and upon phosphorylation it is completely inactivated. Such a strict activity regulation is a prerequisite for proper operation of the C4 pathway since the synthesis of PEP by PPDK consumes two molecules of ATP. The activity of the maize C4‐specific PPDK expressed in rice leaves was also light–dark regulated as it is in maize, being up‐regulated in the light (Fukayama et al., 2001). There are four reports on transgenic plants, which overproduce PPDK derived from higher plants; transgenic Arabidopsis (Ishimaru et al., 1997), potato (Ishimaru et al., 1998), rice (Fukayama et al., 2001) overproducing the maize C4‐specific PPDK, and transgenic tobacco overproducing PPDK from a CAM plant Mesembryanthemum crystallinum (Sheriff et al., 1998). Physiological impacts were minimal and no changes in the photosynthetic characteristics were observed in these transformants, even in the transgenic rice with a 40‐fold increase in activity (Fukayama et al., 2001). In general, the reaction of PPDK is freely reversible, depending on concentrations of substrates, activators, and inactivators (Burnell and Hatch, 1985). This could be the reason why the overexpression of PPDK does not result in significant effects on carbon metabolism in the leaves. NADP‐ME In contrast to PEPC and PPDK, there is no specific regulatory protein for higher plant chloroplastic NADP‐ME. The activity of the C4‐specific NADP‐ME is increased under illumination through increases in pH and concentration of Mg2+ in the chloroplast stroma (Edwards and Andreo, 1992). Two sets of transgenic rice plants overproducing the maize C4‐specific isoform (Takeuchi et al., 2000; Tsuchida et al., 2001) and the rice C3‐specific isoform of NADP‐ME (Tsuchida et al., 2001) have been reported. The transformants overproducing the rice enzyme with some fold increase in activity did not show any detectable differences in their growth, while those overproducing the maize enzyme at the same activity level showed serious stunting and bleaching of leaf colour, due to enhanced photoinhibition of photosynthesis under natural light conditions. It is proposed that the action of the maize NADP‐ME in the chloroplasts increases the NADPH/NADP ratio and suppresses photorespiration, rendering photosynthesis more susceptible to photoinhibition (Takeuchi et al., 2000; Tsuchida et al., 2001). The C4‐specific NADP‐ME has a higher Vm value, lower Km values for substrates, and higher optimum pH, as compared with the C3‐specific isoform (Casati et al., 1997). Such features are suitable for strict regulation of the enzyme activity in the bundle sheath cell chloroplasts of C4 plants, but they allow the enzyme to continue operating in the leaves of C3 plants even when serious damage occurs. Future applications of overproduction of C4 enzymes A major objective of overproduction of C4 enzymes in C3 plants is to improve the photosynthetic performance. As described above and reviewed recently (Häusler et al., 2002), none of the positive effects on photosynthesis have been observed in transgenic C3 plants overproducing a single C4 enzyme. Transgenic C3 plants overproducing multiple enzymes are being produced and analysed in some research groups (Häusler et al., 2002). Although the introduction of the ‘C4‐like’ pathway into the mesophyll cells of C3 plants is one strategy being adopted (Mann, 1999; Surridge, 2002), whether or not this pathway can operate with desirable effects on C3 photosynthesis is a matter of controversy (Edwards, 1999; Leegood, 2002; Häusler et al. 2002). Considering the C4 pathway operating in a single cell found in some aquatic organisms (for a review see Leegood, 2002), it might be possible that the C4‐like pathway could support C3 photosynthesis under some stress conditions such as drought, in which the CO2 availability is limited. Apart from photosynthesis, overproduction of a single C4 enzyme seems to have some positive effects on physiology of C3 plants. It has been reported that overproduction of the chloroplastic, but not cytosolic, PPDK increased the number of seeds per seed capsule and the weight of each seed capsule in transgenic tobacco (Sheriff et al., 1998), and that overproduction of the maize C4‐specific PEPC improved resistance to aluminium of root elongation in transgenic rice (Miyao et al., 2001). Of course, it is of prime importance to elucidate mechanisms for these effects and to confirm whether or not similar phenomena can be generally observed in different plant species. Taking account of a variety of housekeeping functions of the C3‐specific enzymes, it is not unlikely that overproduction of C4 enzymes could improve various features of C3 plants. Acknowledgements The author is grateful to Professor Makoto Matsuoka, Nagoya University, Japan, and Professor RE Häusler, University of Köln, Germany, for providing unpublished information, and to Ms Hiroko Tsuchida for her assistance in preparing the manuscript. Open in new tabDownload slide Fig. 1. Simplified illustrations of the C3 photosynthetic pathway (A) and the C4 photosynthetic pathway of the NADP‐ME type C4 plants (B). In the C4 pathway, one molecule of CO2 is pumped up from the cytosol of the mesophyll cell into the chloroplast of the bundle sheath cell where Rubisco is present. This process consumes two molecules of ATP (one consumed by PPDK and the other required for the conversion of AMP produced by PPDK to ADP), but it does not consume or produce any other substances. TP, triosephosphate. Open in new tabDownload slide Fig. 2. Comparison of the rice and maize Pdk1 genes that encode the chloroplastic and cytosolic isoforms of PPDK. Both genes have a dual promoter system and transcription starts at two different sites indicated by bent arrows, giving rise to transcripts different in size. The larger transcript encodes the chloroplastic isoform and the smaller one encodes the cytosolic isoform. The coding regions common to the two transcripts are represented by filled boxes, and the 5′‐ and 3′‐non‐coding regions by open boxes. Hatched boxes in maize exon 1 and rice exon 1′ represent regions that encode the transit peptide, and those in maize exon 2 and rice exon 2 represent the coding regions unique to the small transcript. ATG and TGA indicate the initiation and termination codons, respectively. The gene structures reported previously (Imaizumi et al., 1997) are modified. Open in new tabDownload slide Fig. 3. A schematic representation of evolution of the maize Pdk1 and Pdk2 genes that encode PPDK. Pdk1 genes of the existing C3 and C4 plants have a dual promoter system and encode the chloroplastic and cytosolic isoforms, while Pdk2 genes have a single promoter and encode the cytosolic isoform. Chloroplastic and cytosolic genes represent genes that encode the chloroplastic and cytosolic isoforms, respectively. Open in new tabDownload slide Fig. 4. The promoter of the C4‐specific gene for the enzyme specific to the bundle sheath cells directs the expression of a reporter gene in the bundle sheath cells and vascular bundles in rice plants. The 5′ region (–1447 to +227, an upstream region from the initiation codon) of the C4‐specific PEP‐CK gene from Zoysia japonica (PEP‐CK type) was fused to the 5′ side of the GUS gene and introduced into rice plants by Agrobacterium‐mediated gene transfer. Histochemical localization of GUS activity is shown. Cross‐sections of leaf blade (a), leaf sheath (b) and stem (c). Scale bars 0.1 mm. M Nomura, M Matsuoka, unpublished results. Table 1. Increase in activities of C4enzymes in transgenic rice leaves C4 enzyme (location in C4 plants) Introduced construct Highest enzyme activitya (increase in fold) References Over rice activity Over maize activity PEPC (MC) Intact maize gene 110 3–4 Ku et al., 1999 PPDK (MC) Rice Cab prom::maize FL C4 cDNA 5 <0.1 Fukayama et al., 2001 Maize Pdk1 C4 prom::maize FL C4 cDNA 5 <0.1 Fukayama et al., 2001 Intact maize gene 40 0.5 Fukayama et al., 2001 NADP‐ME (BSC) Rice Cab prom::rice FL C3 cDNA <5 <0.1 Tsuchida et al., 2001 Rice Cab prom::maize FL C4 cDNA 30 0.6 Tsuchida et al., 2001 70 – Takeuchi et al., 2000 PEP‐CK (BSC) Rice Cab prom::Zoysia FL C4 cDNA – 0.1b Miyao M. et al.,  unpublished results Maize C4 PEPC prom::Urochloa C4 cDNAc – 0.5d Suzuki et al., 2000 Maize Pdk1 C4 prom::Urochloa C4 cDNAc – 0.5d Suzuki et al., 2000 C4 enzyme (location in C4 plants) Introduced construct Highest enzyme activitya (increase in fold) References Over rice activity Over maize activity PEPC (MC) Intact maize gene 110 3–4 Ku et al., 1999 PPDK (MC) Rice Cab prom::maize FL C4 cDNA 5 <0.1 Fukayama et al., 2001 Maize Pdk1 C4 prom::maize FL C4 cDNA 5 <0.1 Fukayama et al., 2001 Intact maize gene 40 0.5 Fukayama et al., 2001 NADP‐ME (BSC) Rice Cab prom::rice FL C3 cDNA <5 <0.1 Tsuchida et al., 2001 Rice Cab prom::maize FL C4 cDNA 30 0.6 Tsuchida et al., 2001 70 – Takeuchi et al., 2000 PEP‐CK (BSC) Rice Cab prom::Zoysia FL C4 cDNA – 0.1b Miyao M. et al.,  unpublished results Maize C4 PEPC prom::Urochloa C4 cDNAc – 0.5d Suzuki et al., 2000 Maize Pdk1 C4 prom::Urochloa C4 cDNAc – 0.5d Suzuki et al., 2000 a Highest enzyme activities among the primary transgenic plants are listed. b Highest level of the enzyme protein relative to the level in Zoysia leaves is presented. c The Urochloa C4‐specific PEP‐CK cDNA was fused to a sequence of the transit peptide for targeting to chloroplasts. d Highest activities of the secondary transgenic plants relative to the activity of Urochloa leaves are presented. MC, mesophyll cells; BSC, bundle sheath cells; prom, promoter; FL, full‐length. Open in new tab Table 1. Increase in activities of C4enzymes in transgenic rice leaves C4 enzyme (location in C4 plants) Introduced construct Highest enzyme activitya (increase in fold) References Over rice activity Over maize activity PEPC (MC) Intact maize gene 110 3–4 Ku et al., 1999 PPDK (MC) Rice Cab prom::maize FL C4 cDNA 5 <0.1 Fukayama et al., 2001 Maize Pdk1 C4 prom::maize FL C4 cDNA 5 <0.1 Fukayama et al., 2001 Intact maize gene 40 0.5 Fukayama et al., 2001 NADP‐ME (BSC) Rice Cab prom::rice FL C3 cDNA <5 <0.1 Tsuchida et al., 2001 Rice Cab prom::maize FL C4 cDNA 30 0.6 Tsuchida et al., 2001 70 – Takeuchi et al., 2000 PEP‐CK (BSC) Rice Cab prom::Zoysia FL C4 cDNA – 0.1b Miyao M. et al.,  unpublished results Maize C4 PEPC prom::Urochloa C4 cDNAc – 0.5d Suzuki et al., 2000 Maize Pdk1 C4 prom::Urochloa C4 cDNAc – 0.5d Suzuki et al., 2000 C4 enzyme (location in C4 plants) Introduced construct Highest enzyme activitya (increase in fold) References Over rice activity Over maize activity PEPC (MC) Intact maize gene 110 3–4 Ku et al., 1999 PPDK (MC) Rice Cab prom::maize FL C4 cDNA 5 <0.1 Fukayama et al., 2001 Maize Pdk1 C4 prom::maize FL C4 cDNA 5 <0.1 Fukayama et al., 2001 Intact maize gene 40 0.5 Fukayama et al., 2001 NADP‐ME (BSC) Rice Cab prom::rice FL C3 cDNA <5 <0.1 Tsuchida et al., 2001 Rice Cab prom::maize FL C4 cDNA 30 0.6 Tsuchida et al., 2001 70 – Takeuchi et al., 2000 PEP‐CK (BSC) Rice Cab prom::Zoysia FL C4 cDNA – 0.1b Miyao M. et al.,  unpublished results Maize C4 PEPC prom::Urochloa C4 cDNAc – 0.5d Suzuki et al., 2000 Maize Pdk1 C4 prom::Urochloa C4 cDNAc – 0.5d Suzuki et al., 2000 a Highest enzyme activities among the primary transgenic plants are listed. b Highest level of the enzyme protein relative to the level in Zoysia leaves is presented. c The Urochloa C4‐specific PEP‐CK cDNA was fused to a sequence of the transit peptide for targeting to chloroplasts. d Highest activities of the secondary transgenic plants relative to the activity of Urochloa leaves are presented. MC, mesophyll cells; BSC, bundle sheath cells; prom, promoter; FL, full‐length. Open in new tab References AgarieS, Miura A, Sumikura R, Tsukamoto S, Nose A, Arima S, Matsuoka M, Miyao‐Tokutomi M. 2002 . Overexpression of C4 PEPC caused O2‐insensitive photosynthesis in transgenic rice plants. Plant Science 162, 257 –265. AoyagiK, Bassham JA. 1984 . Pyruvate orthophosphate dikinase mRNA organ specificity in wheat and maize. Plant Physiology 76, 278 –280. AoyagiK, Chua N‐H. 1988 . Cell‐specific expression of pyruvate, Pi dikinase. In situ mRNA hybridization and immunolocalization labelling of protein in wheat seed. Plant Physiology 86, 364 –368. ap ReesT, Hill SA. 1994 . Metabolic control analysis of plant metabolism. Plant , Cell and Environment 17, 587 –599. AssmannSM. 1993 . Signal transduction in guard cells. Annu al Rev iew of Cell Biol ogy 9, 345 –375. BakrimN, Nhiri M, Pierre J‐N, Vidal J. 1998 . Metabolite control of Sorghum C4 phosphoenolpyruvate carboxylase catalytic activity and phosphorylation state. Photosynthesis Research 58, 153 –162. BlackJr CC. 1973 . Photosynthetic carbon fixation in relation to net CO2 uptake. Annual Review of Plant Physiology 24, 253 –286. BrownRH, Bouton JH. 1993 . Physiology and genetics of interspecific hybrids between photosynthetic types. Annual Review of Plant Physiology and Plant Molecular Biology 44, 435 –456. BurnellJN, Hatch MD. 1985 . Light‐dark modulation of leaf pyruvate, Pi dikinase. Trends in Biochemical Sciences 10, 288 –291. CasatiP, Spampinato CP, Andreo CS. 1997 . Characteristics and physiological function of NADP‐malic enzyme from wheat. Plant and Cell Physiology 38, 928 –934. ChampignyM‐L, Foyer C. 1992 . Nitrate activation of cytosolic protein kinases diverts photosynthetic carbon from sucrose to amino acid biosynthesis. Basis for a new concept. Plant Physiology 100, 7 –12. ChandlerVL, Vaucheret H. 2001 . Gene activation and gene silencing. Plant Physiology 125, 145 –148. Chen L‐M, Izui K. 2002 . Phosphoenolpyruvate carboxylase (PEPC) from a thermophilic cyanobacterium, Synechococcus vulcanus: unusual allosteric properties and its gene expression in Arabidopsis thaliana. In: Proceedings of 12th International Congress on Photosynthesis, Canberra, Australia: CSIRO Publishing, S17 –021. ChenL, Omiya T, Hata S, Izui K. 2002 . Molecular characterization of a phosphoenolpyruvate carboxylase from a thermophilic cyanobacterium, Synechococcus vulcanus with unusual allosteric properties. Plant and Cell Physiology 43, 159 –169. DongL‐Y, Masuda T, Kawamura T, Hata S, Izui K. 1998 . Cloning, expression, and characterization of a root‐form phosphoenolpyruvate carboxylase from Zea mays: comparison with the C4‐form enzyme. Plant and Cell Physiology 39, 865 –873. EdwardsG. 1999 . Tuning up crop photosynthesis. Nature Biotechnology 17, 22 –23. EdwardsGE, Andreo CS. 1992 . NADP‐malic enzyme from plants. Photochemistry 31, 1845 –1857. EhleringerJR, Monson RK. 1993 . Evolutionary and ecological aspects of photosynthetic pathway variation. Annual Review of Ecology and Systematics 24, 411 –439. FukayamaH, Hatch MD, Tamai T, Tsuchida H, Sudoh S, Furbank RT, Miyao M. 2002 . Activity regulation and physiological impacts of the maize C4‐specific phospho enolpyruvate carboxylase overproduced in transgenic rice plants. Photosynthesis Research (in press). FukayamaH, Tsuchida H, Agarie S,et al. 2001 . Significant accumulation of C4‐specific pyruvate, orthophosphate dikinase in a C3 plant, rice. Plant Physiology 127, 1136 –1146. GallieDR. 1998 . Controlling gene expression in transgenics. Current Opinion in Plant Biology 1, 166 –172. GehlenJ, Panstruga R, Smets H,et al. 1996 . Effects of altered phosphoenolpyruvate carboxylase activities on transgenic C3 plant Solanum tuberosum. Plant Molecular Biology 32, 831 –848. GelvinSB. 1998 . The introduction and expression of transgenes in plants. Current Opinion in Biotechnology 9, 227 –232. GlackinCA, Grula JW. 1990 . Organ‐specific transcripts of different size and abundance derive from the same pyruvate, orthophosphate dikinase gene in maize. Proceedings of the Nati o nal Academy of Sciences, USA 87, 3004 –3008. GoodallGJ, Filipowicz W. 1991 . Different effects of intron nucleotide composition and secondary structure on pre‐mRNA splicing in monocot and dicot plants. EMBO Journal 10, 2625 –2644. HatchMD. 1987 . C4 photosynthesis: a unique blend of modified biochemistry, anatomy and ultrastructure. Biochimica et Biophysica Acta 895, 81 –106. HäuslerRE, Hirsch HJ, Kreuzaler F, Peterhänsel C. 2002 . Overexpression of C4‐cycle enzymes in transgenic C3 plants: a biotechnological approach to improve C3‐photosynthesis. Journal of Experimental Botany 53, 591 –607. HäuslerRE, Kleines M, Uhrig H, Hirsch H‐J, Smets H. 1999 . Overexpression of phosphoenolpyruvate carboxylase from Corynebacterium glutamicum lowers the CO2 compensation point (Γ*) and enhances dark and light respiration in transgenic potato. Journal of Experimental Botany 50, 1231 –1242. HäuslerRE, Rademacher T, Li J, Lipka V, Fischer KL, Schubert S, Kreuzaler F, Hirsch HJ. 2001 . Single and double overexpression of C4‐cycle genes had differential effects on the pattern of endogenous enzymes, attenuation of photorespiration and on contents of UV protectants in transgenic potato and tobacco plants. J ournal of Experimental Botany 52, 1785 –1803. HeinekeD, Riens B, Grosse H, Hoferichter P, Peter U, Flügge UI, Heldt HW. 1991 . Redox transfer across the inner chloroplast envelope membrane. Plant Physiology 95, 1131 –1137. HieiY, Ohta S, Komari T, Kumashiro T. 1994 . Efficient transformation of rice (Oryza sativa L.) mediated by Agrobacterium and sequence analysis of the boundaries of the T‐DNA. The Plant Journal 6, 271 –282. HudspethRL, Grula JW, Dai Z, Edwards GE, Ku MSB. 1992 . Expression of maize phosphoenolpyruvate carboxylase in transgenic tobacco. Plant Physiology 98, 458 –464. ImaizumiN, Ku MSB, Ishihara K, Samejima M, Kaneko S, Matsuoka M. 1997 . Characterization of the gene for pyruvate, orthophosphate dikinase from rice, a C3 plant, and a comparison of structure and expression between C3 and C4 genes for this protein. Plant Molecular Biology 34, 701 –716. IshimaruK, Ichikawa H, Matsuoka M, Ohsugi R. 1997 . Analysis of a C4 maize pyruvate, orthophosphate dikinase expressed in C3 transgenic Arabidopsis plants. Plant Science 129, 57 –64. IshimaruK, Ohkawa Y, Ishige T, Tobias DJ, Ohsugi R. 1998 . Elevated pyruvate, orthophosphate dikinase (PPDK) activity alters carbon metabolism in C3 transgenic potatoes with a C4 maize PPDK gene. Physiologia Plantarum 103, 340 –346. JiaoD, Li X, Huang X, Chi W, Kuang T, Ku MSB. 2001 . The characteristics of CO2 assimilation of photosynthesis and chlorophyll fluorescence in transgenic PEPC rice. Chinese Science Bulletin 46, 1080 –1084. KogamiH, Shono M, Koike T, Yanagisawa S, Izui K, Sentoku N, Tanifuji S, Uchimiya H, Toki S. 1994 . Molecular and physiological evaluation of transgenic tobacco plants expressing a maize phosphoenolpyruvate carboxylase gene under the control of the cauliflower mosaic virus 35S promoter. Transgenic Research 3, 287 –296. KuMSB, Agarie S, Nomura M, Fukayama H, Tsuchida H, Ono K, Hirose S, Toki S, Miyao M, Matsuoka M. 1999 . High‐level expression of maize phosphoenolpyruvate carboxylase in transgenic rice plants. Nature Biotechnology 17, 76 –80. KuMSB, Kano‐Murakami Y, Matsuoka M. 1996 . Evolution and expression of C4 photosynthesis genes. Plant Physiology 111, 949 –957. LeegoodRC, Lea PJ, Adcock MD, Häusler RE. 1995 . The regulation and control of photorespiration. Journal of Experimental Botany 46, 1397 –1414. LeegoodRC. 2002 . C4 photosynthesis: principles of CO2 concentration and prospects for its introduction into C3 plants. Journal of Experimental Botany 53, 581 –590. MannCC. 1999 . Genetic engineers aim to soup up crop photosynthesis. Science 283, 314 –316. MatsuokaM, Fukayama H, Tsuchida H, Nomura M, Agarie S, Ku MSB, Miyao M. 2000 . How to express some C4 photosynthesis genes at high levels in rice. In: Sheehy JE, Mitchell PL, Hardy B, eds. Redesigning rice photosynthesis to increase yield. Amsterdam: Elsevier Science Publishers, 167 –175. MatsuokaM, Furbank RT, Fukayama H, Miyao M. 2001 . Molecular engineering of C4 photosynthesis. Annual Review of Plant Physiology and Plant Molecular Biology 52, 297 –314. MatsuokaM, Kyozuka J, Shimamoto K, Kano‐Murakami Y. 1994 . The promoters of two carboxylases in a C4 plant (maize) direct cell‐specific, light‐regulated expression in a C3 plant (rice). The Plant Journal 6, 311 –319. MatsuokaM, Numazawa T. 1991 . CIS‐acting elements in the pyruvate, orthophosphate dikinase gene from maize. Molecular and General Genetics 228, 143 –152. MatsuokaM, Tada Y, Fujimura T, Kano‐Murakami Y. 1993 . Tissue‐specific light‐regulated expression directed by the promoter of a C4 gene, maize pyruvate, orthophosphate dikinase, in a C3 plant, rice. Proceedings of the Nati o nal Academy of Sciences, USA 90, 9586 –9590. MiyaoM, Fukayama H, Tamai T, Matsuoka M. 2001 . High level expression of C4 photosynthesis enzymes in transgenic rice. In: Proceedings of 12th International Congress on Photosynthesis, Canberra, Australia: CSIRO Publishing, S39 –001. MonsonRK. 1999 . The origins C4 genes and evolutionary pattern in the C4 metabolic phenotype. In: Sage RF, Monson RK, eds. C4 plant biology. San Diego: Academic Press, 377 –410. MoonsA, Valcke R, van Montagu M. 1998 . Low‐oxygen stress and water deficit induce cytosolic pyruvate orthophosphate dikinase (PPDK) expression in roots of rice, a C3 plant. The Plant Journal 15, 89 –98. NomuraM, Sentoku N, Nishimura A,et al. 2000a. The evolution of C4 plants: acquisition of cis‐regulatory sequences in the promoter of C4‐type pyruvate, orthophosphate dikinase gene. The Plant Journal 22, 211 –221. NomuraM, Sentoku N, Tajima S, Matsuoka M. 2000b. Expression patterns of cytoplasmic pyruvate, orthophosphate dikinase of rice (C3) and maize (C4) in a C3 plant, rice. Australian Journal of Plant Physiology 27, 343 –347. OsmondCB, Grace SC. 1995 . Perspectives on photoinhibition and photorespiration in the field: quintessential inefficiencies of the light and dark reactions of photosynthesis? Journal of Experimental Botany 46, 1351 –1362. PanstrugaR, Hippe‐Sanwald S, Lee Y‐K, Lataster M, Lipka V, Fischer R, Liao YC, Häusler RE, Kreuzaler F, Hirsch H‐J. 1997 . Expression and chloroplast‐targeting of active phosphoenolpyruvate synthetase from Escherichia coli in Solanum tuberosum. Plant Science 127, 191 –205. RademacherT, Häusler RE, Hirsch H‐J, Zhang L, Lipka V, Weier D, Kreuzaler F, Peterhänsel C. 2002 . An engineered phosphoenolpyruvate carboxylase redirects carbon and nitrogen flow in transgenic potato plants. The Plant Journal 32, 25 –39. RoscheE, Streubel M, Westhoff P. 1994 . Primary structure of the photosynthetic pyruvate orthophosphate dikinase of the C3 plant Flaveria pringlei and expression analysis of pyruvate orthophosphate dikinase sequences in C3, C3–C4 and C4 Flaveria species. Plant Molecular Biology 26, 763 –769. RoscheE, Westhoff P. 1995 . Genomic structure and expression of the pyruvate, orthophosphate dikinase gene of the dicotyledonous C4 plant Flaveria trinervia (Asteraceae). Plant Molecular Biology 29, 663 –678. SakamotoM, Sanada Y, Tagiri A, Murakami T, Ohashi Y, Matsuoka M. 1991 . Structure and characterization of a gene for light‐harvesting Chl a/b binding protein from rice. Plant and Cell Physiology 32, 385 –393. SalvucciME, Werneke JM, Ogren WL, Portis Jr AR. 1987 . Purification and species distribution of Rubisco activase. Plant Physiology 84, 930 –936. SheenJ. 1991 . Molecular mechanisms underlying the differential expression of maize pyruvate, orthophosphate dikinase genes. The Plant Cell 3, 225 –245. SheriffA, Meyer H, Riedel E, Schmitt JM, Lapke C. 1998 . The influence of plant pyruvate, orthophosphate dikinase on a C3 plant with respect to the intracellular location of the enzyme. Plant Science 136, 43 –57. SurridgeC. 2002 . The rice squad. Nature 416, 576 –578. SuzukiS, Murai N, Burnell JN, Arai M. 2000 . Changes in photosynthetic carbon flow in transgenic rice plants that express C4‐type phosphoenolpyruvate carboxykinase from Urochloa panicoides. Plant Physiology 124, 163 –172. SvenssonP, Bläsing O, Westhoff P. 1997 . Evolution of the enzymatic characteristics of C4 phosphoenolpyruvate carboxylase. A comparison of the orthologous PPCA phosphoenolpyruvate carboxylases of Flaveria trinervia (C4) and Flaveria pringlei (C3). European Journal of Biochemistry 246, 452 –460. TakeuchiY, Akagi H, Kamasawa N, Osumi M, Honda H. 2000 . Aberrant chloroplasts in transgenic rice plants expressing a high level of maize NADP‐dependent malic enzyme. Planta 211, 265 –274. TsuchidaH, Tamai T, Fukayama H,et al. 2001 . High level expression of C4‐specific NADP‐malic enzyme in leaves and impairment of photoautotrophic growth of a C3 plant, rice. Plant and Cell Physiology 42, 138 –145. VidalJ, Chollet R. 1997 . Regulatory phosphorylation of C4 PEP carboxylase. Trends in Plant Science 2, 230 –237. © 2003

Garwe, Dahlia; Thomson, Jennifer A.; Mundree, Sagadevan G.

doi: 10.1093/jxb/erg013pmid: 12493847

Abstract The strategy of ‘complementation by functional sufficiency’ was used to isolate a cDNA designated XVSAP1 from a cDNA library constructed from dehydrated Xerophyta viscosa Baker leaves. Analysis of the cDNA sequence indicated a highly hydrophobic protein with six transmembrane regions. Southern blot analysis revealed that there are at least two copies of XVSAP1 in X. viscosa. The deduced amino acid sequence showed 49% identity to WCOR413, a low‐temperature‐regulated protein from wheat. The protein also showed between 25% to 56% identity to WCOR413‐like proteins from Arabidopsis thaliana. Expression of XVSAP1 in Escherichia coli (srl::Tn10) conferred osmotic stress tolerance when the cells were grown in 1 M sorbitol. Analysis of gene expression using semi‐quantitative RT‐PCR indicated that XVSAP1 is induced by dehydration, salt stress (100 mM), both low (4 °C) and high temperature (42 °C) and high light treatment (1500 µmol m–2 s–1). These results suggest that XVSAP1 may have a significant role to play in the response of X. viscosa to abiotic stresses. Key words: Cold stress, desiccation stress, heat stress, resurrection plant, salinity stress. Received 27 May 2002; Accepted 14 August 2002 Introduction Adverse environmental factors such as extremes of temperature and osmotic stress resulting from conditions of high salinity and drought, affect the growth, productivity and distribution of plants (Boyer, 1982). As sessile organisms, plants have evolved a wide variety of mechanisms that enable them to grow and reproduce under hostile environmental conditions. The response to abiotic stresses is mediated through physiological, morphological and metabolic modifications occurring in all plant organs. Expression of genes in response to different environmental stimuli results from a complex signal transduction cascade that commences with perception of the stimulus followed by processing including amplification and integration of the signal. The final step is a response reaction in the form of de novo gene expression (Ingram and Bartels, 1996). Although the general response to abiotic stress is similar in all plants, there is a group of plants known as ‘resurrection plants’ that have developed mechanisms that enable them to withstand severe water deficit. These plants are unique in their ability to tolerate drying of their vegetative tissues. Resurrection plants can lose over 90% of their water content, survive in their dried state for prolonged periods and then resume active life when water becomes available again (Bartels et al., 1990; Gaff, 1971; Sherwin and Farrant, 1996). The molecular basis of desiccation tolerance has been studied in a few species representing different groups: the moss Tortula ruralis, the monocotyledonous Sporobolus stapfianus (Blomstedt et al., 1998; Oliver, 1996; Neale et al., 2000) and the dicotyledonous Craterostigma plantagineum (Bartels et al., 1990). It is thought that two basic mechanisms exist which allow desiccation‐tolerant plants to survive water deprivation. The first involves the protection of cellular integrity through inducible and constitutive mechanisms, while the second involves the repair of desiccation or rehydration induced damage. However, both mechanisms are probably employed for desiccation tolerance with different plants utilizing one strategy more than the other (Oliver and Bewley, 1997). Theories on the mechanisms by which resurrection plants tolerate dehydration have mostly been derived from observations of the cellular processes that occur during drying of the plant (Ingram and Bartels, 1996). It has been demonstrated that several genes are differentially expressed in response to dehydration (Blomstedt et al., 1998; Itturiaga et al., 1992; Schneider et al., 1993). Both genetic and biochemical studies have established that the phytohormone ABA is crucial in the response to desiccation, salt, and cold (Bray, 1997). Characterization of the ABA inducible genes Em from wheat and rab16A from rice by expression studies and analysis of protein binding in vitro, showed that a cis‐regulatory ABA‐responsive element (ABRE) is important for transcription (Marcotte et al., 1989; Mundy et al., 1990). The element consists of 8–10 base pairs with the core sequence ACGT. Exogenous application of the plant hormone leads to the expression of most of the dehydration‐induced proteins (Bartels et al., 1990). ABA also plays an important role in physiological processes such as the closing of guard cells under drought stress and the regulation of several events during late seed development (Zeevaart and Creelman, 1988; McCarty, 1995). Blomstedt et al. (1998) found that a number of genes activated in the early stages of dehydration in resurrection plants are similar to those expressed in the desiccating seed of most plants. The synthesis of globular and extremely hydrophilic proteins known as late embryogenesis abundant (LEA) proteins is one of the well‐documented responses to dehydration, salinity and cold stress as is the accumulation of osmolytes. It has been suggested that LEA proteins have a role to play in the maintenance of protein or membrane structure, the sequestration of ions and the binding of water. Osmolytes are thought to function by raising the osmotic potential of the cell and also by stabilizing proteins and membranes when water deficit occurs (Le Rudulier et al., 1984; McNeil et al., 1999). Low temperature is one of the major environmental factors limiting plant growth. Freezing temperature induces injuries, particularly to the cellular membrane systems, that result largely from the severe dehydration that occurs upon ice formation within the cells. (Gilmour et al., 1988; Thomashow, 2001). Low temperature also affects the normal functioning of integral membrane proteins such as transporters and receptor proteins whose activity is dependent on the fluidity of the membrane (Hazel, 1995). The products of cold‐regulated (cor) genes such as the Arabidopsis COR6.6 and COR78, may protect and help plants to adapt to cold stress (Thomashow, 1999). Studies on cold‐regulated gene expression in Arabidopsis resulted in the discovery of a DNA regulatory element, the C‐repeat (CRT) dehydration responsive element (DRE) which has a conserved core sequence of CCGAC. Transcriptional activators that bind the CRT/DRE designated CBF1, CBF2 and CBF3 or DREB1A, DREB1C and DREB1A, respectively, were subsequently identified. DRE confers responsiveness to low temperature and dehydration (Liu et al., 1998; Stockinger et al., 1997). Danyluk et al. (1994) identified a low‐temperature responsive dehydrin‐like gene, wcor410, belonging to a family of homologous members, wcor410, wcor410b and wcor410c. The results from their work suggested that the protein was involved in the cryoprotection of the plasma membrane against freezing or dehydration stress. It was shown that water stress, polyethylene glycol, ABA and, to a lesser extent, salt and wounding also resulted in the up‐regulation of members of the wcor410 family. Similarly the wcs120 transcript, which codes for a protein homologous to dehydrins, was found to accumulate in response to cold stress and its promoter was found to be stress‐inducible (Oullet et al., 1998). The processes associated with tissue recovery on rehydration in resurrection plants have been less extensively studied with only a few rehydration‐associated proteins identified (Oliver et al., 1998). Most of the information available is on the fully desiccation‐tolerant moss T. ruralis. It has been postulated that the moss relies more on the activation of pre‐existing repair mechanisms for desiccation tolerance than it does on either pre‐established or activated protection systems (Oliver, 1991). In desiccation‐tolerant angiosperms recovery is more complex. Studies by Tuba et al. (1993) showed that in poikilochlorophyllous plants such as Xerophyta, the chloroplasts were extensively altered after a period of desiccation. The rebuilding of chloroplasts and the photosynthetic apparatus occurs on rehydration. However, Craterostigma wilmsii, a modified desiccation‐tolerant resurrection plant, retains its chlorophyll on drying. Protective mechanisms during dehydration rather than repair on rehydration appear to dominate in modified desiccation tolerant resurrection plants (Sherwin and Farrant, 1996). X. viscosa is a monocotyledonous resurrection plant that is capable of tolerating severe abiotic stress conditions (Sherwin and Farrant, 1996; Mundree et al., 2000). Mundree et al. (2000) used an approach called ‘complementation by functional sufficiency’ to isolate genes from X. viscosa that conferred functional sufficiency to osmotically–stressed E. coli (srl::Tn10). XVSAP1 was isolated from a cDNA library constructed from dehydrated X. viscosa leaves using this strategy. The protein shows 49% identity to a cold‐tolerance protein, WCOR413, from Triticum aestivum (Danyluk and Sarhan, 1996). It also bears close identity (56%) to other uncharacterized proteins from Arabidopsis which themselves have 64% identity to WCOR413. In this report the molecular characterization of XVSAP1 is described. Materials and methods Plant material, growth conditions and XVSAP1 cDNA isolation X. viscosa plants were collected from the Buffelskloof Nature Reserve (Mpumalanga Province, South Africa). The plants were potted and grown under greenhouse conditions as described by Sherwin and Farrant (1996). Experimental plants were watered to ensure full hydration prior to the stress experiments. Relative water content (RWC) determination was as described by Sherwin and Farrant (1996). Construction and screening of the cDNA library was as previously described (Mundree et al., 2000). A cDNA insert in pBluescriptSK+ (Stratagene, La Jolla, CA) named XVSAP1 was used for the experiments described in this paper. XVSAP1 expression in osmotically stressed Escherichia coli XVSAP1 was cloned into the pProEXHT Prokaryotic Expression Vector System (Life Technologies, Inc, USA). E. coli (srl::Tn10) cells were transformed with the pProEXHT‐XVSAP1 construct and grown in M9 minimal medium supplemented with 1 mM MgSO4.7H2O, 0.2% glycerol, 0.1% vitamin B, and 100 µg ml–1 ampicillin. Cell cultures were induced in duplicate by adding 0.2 mM isopropyl thiogalactopyranoside (IPTG) after the OD600 of the cells had reached approximately 0.5. The cells were allowed to grow for a further 2 h before an osmotic stress was imposed by adding 4 M sorbitol to a final concentration of 1 M. The growth of the cells was monitored by taking absorbance readings at 600 nm over a 48 h period. The experiment was repeated three times. XVSAP1 sequence analysis The nucleotide sequence of XVSAP1 was determined using the ALFexpress automated DNA sequencer (Pharmacia Biotech AB, Uppsala, Sweden) as described by Mundree et al. (2000). The BLAST program of the National Centre for Biotechnology Information (Altschul et al., 1990) was used to search databases for sequence similarities. Nucleotide and amino acid sequence comparisons were done using CLUSTAL W (1.5) on the BCM server. The ProfileScan tool on the ISREC bioinformatics server was used to scan XVSAP1 for conserved motifs. DNAMAN (Version 4.3, Lynnon BioSoft) was used to construct the homology tree. Southern blot analysis Genomic DNA from X. viscosa was isolated using the plant DNA preparation procedure described by Dellaporta et al. (1983) except that in all cases approximately 1 g of leaf tissue was used and DNA was precipitated with isopropanol. Isolated DNA was quantitated spectrophotometrically. 15 µg DNA aliquots were restricted with EcoRI, XhoI, BglII, HindIII and EcoRV restriction endonucleases, electrophoresed on 1% agarose gels and blotted onto nylon membranes (Hybond ‐N, Amersham) by capillary transfer (Sambrook et al., 1989). DNA was fixed using a cross‐linker (Stratalinker 1800, Stratagene). The complete XVSAP1 cDNA was labelled with digoxigenin (DIG) using the random primed method according to the manufacturer’s instructions (Roche Diagnostics GmbH, Germany). Blots were hybridized with the labelled XVSAP1 probe for 16 h at 42 °C. The blots were subsequently washed with 2× SSC (sodium citrate buffer), 0.1% SDS at room temperature and stringently with 0.5× SSC, 0.1% SDS at 68 °C. The chemiluminescent alkaline substrate disodium 3‐(4‐methoxyspiro(1,2‐dioxetane‐3,2′‐(5′‐chloro)tricyclo[3,3.1.13.7]decan}‐4‐yl) phenyl phos phate [CSPD (Roche Diagnostics GmbH, Germany)], was used for detection according to the manufacturer’s instructions. Stress induction X. viscosa plants were dehydrated by withholding water for a period of 2 weeks. Leaves were detached from the plants at 90%, 78%, 63%, 51%, 44%, and 4% RWC. Leaf samples were also collected at 4%, 32%, 42%, 85%, and 92% RWC on re‐hydrating. For the heat treatment, fully hydrated plants were kept in a phytotron at 42 °C (humidity 50–70%, 16/8 h light/dark cycle). The plants were watered regularly to maintain them at full hydration. To determine the effect of cold stress, plants were kept at 4 °C and leaf samples taken every 6 h for 60 h. To test the response of X. viscosa to high salinity, the plants were irrigated with 100 mM NaCl daily for 7 d. The high light treatment was carried out by exposing the plants to light at 1500 µmol m–2 s–1 for 4 d in a phytotron (25 °C, humidity 50–70%). Plants were irrigated with water daily to keep them fully hydrated. To determine if abscisic acid (ABA) had an effect on the expression of XVSAP1, X. viscosa leaves were sprayed with the phytohormone at a concentration of 100 µM in water once every 24 h. In all cases, leaf samples were taken from the experimental plants just before commencing treatments (time 0). Samples were collected every 24 h thereafter except for the cold treatment, where samples were collected every 6 h. In the case of ABA, samples were taken every 3 h after treatment of the leaves. All leaf samples collected were frozen in liquid nitrogen and stored at –70 °C until further use. RNA isolation Total RNA was isolated using the Trizol reagent (Gibco‐BRL). X. viscosa leaves (200 mg) were ground in liquid nitrogen and homogenized in 0.75 ml of the reagent. Following incubation for 5 min at room temperature, 0.2 ml chloroform was added followed by a further incubation at room temperature for 10 min. Samples were centrifuged at 12 000 g for 10 min at 4 °C and the RNA was precipitated using isopropanol. RNA was quantitated spectrophotometrically, separated on a 1.2% agarose formaldehyde gel and stained with ethidium bromide to verify quantitation. Semi‐quantitative RT‐PCR All RNA samples were treated with DNase I (Roche Diagnostics GmbH, Germany) according to the manufacturer’s instructions to eliminate DNA contamination. In each case, 2 µg RNA was used for the reverse transcription reaction. The internal control RNA was prepared by deleting a 473 bp (NdeI restriction) fragment from XVSAP1 in pBluescriptSK+ and performing in vitro transcription. Two picograms of the truncated clone were used in all the RT reactions except for the ABA RT‐PCR where 0.5 pg of the internal standard was used. The reverse transcription reactions were performed using the Omniscript reverse transcriptase kit according to the manufacturer’s directions (Qiagen GmbH, Germany). RNase inhibitor was obtained from Roche Diagnostics GmbH, Germany. The cDNA (5 µl) from the RT step was used in 50 µl PCR reactions undiluted. The primer pair (forward primer, 5′‐GCACGAGGCA GATTTGAA TTG‐3′; reverse primer, 5′‐ATATGGACACGCAT GACCCA‐3′) produced an 829 bp product from XVSAP1 and a 342 bp product from the truncated clone. Reactions were conducted using a Gene Amp 9700 (Perkin Elmer Applied Biosystems, CA, USA) thermocycler under the following conditions: 95 °C for 2 min followed by 23 cycles of 95 °C for 30 s, 61 °C for 40 s and 72 °C for 45 s and a final extension step for 6 min. The linear portion of the reaction was determined to be between 18 and 25 cycles and 23 cycles were used for all the experiments. After PCR, the samples were resolved by electrophoresis on a 0.8% agarose gel and stained with ethidium bromide. Gel pictures were obtained using the Gel documentation system GDS 2000 (UVP Ltd, Cambridge, UK). Results XVSAP1 expression in osmotically stressed E. coli E. coli (srl::Tn10) (Csonka and Clark, 1979) cannot grow on minimal media containing high concentrations of sorbitol. To confirm the osmo‐protection function of XVSAP1, the cDNA was cloned into a prokaryotic protein expression vector to yield pPROEXHT‐XVSAP1. E. coli (srl::Tn10) cells transformed with this plasmid exhibited significantly better growth in the presence of 1 M sorbitol over a period of 48 h, compared to E. coli (srl::Tn10) transformed with the vector only, after induction with IPTG (Fig. 1). Although the imposition of osmotic stress by the addition of sorbitol caused an initial decrease in the growth rate of both cultures, 2 h after the stress was imposed there was a steady increase in the growth rate of the experimental cultures. Sequence analysis of XVSAP1 The nucleotide sequence of XVSAP1 is 942 bp with an open reading frame of 867 bp (Fig. 2A). The deduced amino acid sequence encodes a basic protein of 264 amino acids with a molecular weight of 29.6 kDa and a predicted pI of 9.12. A Prosite motif search revealed that the protein has two prokaryotic membrane lipoprotein lipid attachment sites between amino acid residues 149–159 and 239–249. One possible N‐myristoylation site was also found and this is indicated on the XVSAP1 sequence at position 42–47. A hydropathic plot [based on the method of Kyte and Doolittle, 1982 (window of 19 amino acid residues)] (Fig. 2B) predicted a protein rich in hydrophobic residues with an average hydrophobicity index of 0.81. The sequence consists of at least six transmembrane helices (Fig. 2A) suggesting that XVSAP1 is likely to be an integral membrane protein. A computer search of protein sequence databanks revealed that XVSAP1 showed 49% identity to WCOR413, a cold‐responsive protein isolated from wheat and between 25–56% identity to cold associated proteins identified in A. thaliana (Fig. 3). The protein also has 53% identity to a cold associated protein from rice. Results from the BLAST program indicate that the region extending from the lysine residue at position 36 to the phenylalanine residue at 119 (Fig. 2A) bears 12% identity with K+ potassium transporter family that is conserved across phyla (Quintero and Blatt, 1997). A homology tree based on the amino acid sequences of XVSAP1 and its homologues (Fig. 3B) indicates that the first three ATCAPs from Arabidopsis are the most closely related with over 70% identity. XVSAP1 is closest to these three homologues, The rice cold associated protein (RCAP) and WCOR413 are closely related with nearly 70% identity. Southern analysis Southern blot analysis of the X. viscosa genomic DNA probed with XVSAP1 cDNA was carried out to determine the gene copy number. Of the restriction enzymes used, only BglII has a predicted restriction site within XVSAP1. At least seven hybridization bands were detected with this enzyme (Fig. 4). A double‐digestion with EcoRI and XhoI and restriction with HindIII and EcoRV, each resulted in at least four hybridization fragments of varying intensities. These results indicate that there are multiple copies of XVSAP1 in X. viscosa. Semi‐quantitative RT‐PCR Semi‐quantitative RT‐PCR was used to compare the relative transcript levels after various stress treatments. Competitor RNA prepared as described in the Materials and methods was used as a control for variations in the RT and PCR reactions. XVSAP1 was induced by dehydration with the transcript appearing only at 51% and 44% RWC (Fig. 5A). There was no evidence of the transcript at 4% RWC. XVSAP1 was not detected during rehydration of X. viscosa (Fig. 5B). The dehydration–hydration curve (Fig. 5C) for the above treatment revealed that the plant took 12 d to dehydrate to 4% RWC and then completely rehydrated within 4 d after watering. Heat (Fig. 6A), salt (Fig. 6B) and high light (Fig. 6C) resulted in significant induction of XVSAP1. The transcripts took 3 d to appear with heat shock and had declined by day 8 of the treatment. Salt shock resulted in the induction of XVSAP1 expression within 24 h. During the treatment, the transcripts were evident for 7 d, but began to decline on the sixth day of the treatment. The transcripts appeared within 48 h with high light treatment, whereas with the cold treatment (Fig. 6D), the transcripts were evident within the first 24 h. Levels of XVSAP1 transcription during cold treatment remained fairly steady for the duration of the experiment. Discussion The cDNA designated XVSAP1 was isolated from dehydrated X. viscosa leaves using complementation by functional sufficiency as described by Mundree et al. (2000). The E. coli (srl::Tn10) mutant strain used in this study lacks a specific sorbitol transport system and is unable to catabolize this osmoticum. The cells are therefore unable to grow in minimal media in which sorbitol is the sole carbon and energy source (Csonka and Clark, 1979). XVSAP1 cloned in a prokaryotic expression vector was able to rescue E. coli (srl::Tn10) cells growing in media containing a high concentration of sorbitol confirming this study’s hypothesis that the gene is associated with osmotic stress tolerance. Comparable experiments with a construct expressing the related WCOR413 from wheat showed that the protein has similar effects to those of XVSAP1 (data not shown). The predicted XVSAP1 is a highly hydrophobic protein that is probably anchored in the plasma membrane. XVSAP1 showed significant identity to proteins identified in wheat, rice and Arabidopsis. Only the wheat protein, WCOR413, has been partially characterized in this group. The WCOR series of stress‐inducible proteins from wheat resembles soluble hydrophilic dehydrins in contrast to XVSAP1, which is predicted to be an integral membrane protein. Data supplied with the protein sequence in the GenBank (Accession number T06810) indicate that the WCOR413 is a cold‐regulated protein. The proteins from rice and Arabidopsis are classed as cold‐associated proteins on the basis of their similarity to WCOR413 at the amino acid level. As is evident from the results of the homology analysis, XVSAP1 is more closely related to the first three ATCAPs from Arabidopsis than to WCOR413. The level of identity between XVSAP1 and its homologues suggests that the protein may also be a cold‐regulated protein. An analysis of the genomic organization of XVSAP1 by Southern blotting confirmed that the gene is indeed present in the X. viscosa genome. As with the Arabidopsis homologues, there is more than one copy of the gene in the nuclear genome suggesting that XVSAP1 belongs to a small gene family. An examination of the sequence of XVSAP1 revealed very few clues to the possible functions of the protein in conferring stress tolerance. It is possible that XVSAP1 may be involved in the transport of substances or ions across the plasma membrane as a region stretching from amino acid residue 36 to 119 bears 12% identity with a K+ potassium transporter family. The family is conserved across phyla, having plant (AtKT), yeast (HAK) and bacterial (KUP) sequences as members (Quintero and Blatt, 1997). However, no data are currently available to support this possibility. It is conceivable that XVSAP1 may be modified for full function as is indicated by the presence of two prokaryotic membrane lipoprotein lipid attachment sites. Modification of proteins by the covalent attachment of lipids appears to be widespread in living systems (Hayashi and Wu, 1990). The presence of these sites supports the concept that XVSAP1 associates closely with the cell membrane. It is interesting to note that the Arabidopsis homologues also have prokaryotic membrane lipoprotein lipid attachment sites. XVSAP1 is possibly further processed for full function by the attachment of lipids as is implied by the existence of one putative N‐myristoylation site. The process of N‐myristoylation is a cotranslational modification that involves the covalent reaction of myristate and the amino‐terminal glycine residue of a growing polypeptide. Proteins so modified have diverse functions and the myristate appears to be critical for mediating protein–protein and/or protein–membrane interactions (Johnson et al., 1994; Ishitani et al., 2000). However, there is currently no evidence that myristoylation is required for the function of XVSAP1 or indeed that it occurs at all. It has been shown that many genes induced by drought are also induced by salt and cold stresses (Zhu et al., 1997). All these factors result in osmotic stress or water deficit in plant cells. As XVSAP1 was isolated on the basis of its ability to confer tolerance to osmotically stressed E. coli (srl::Tn10), it was necessary to establish if XVSAP1 was involved in the response to these abiotic stresses. An examination of the expression of XVSAP1 during dehydration and rehydration of X.viscosa showed that only dehydration and not rehydration induces the expression of the gene. Interestingly, XVSAP1 expression was strongly induced at 51% and 44% RWC and not at any other stage. This indicates that XVSAP1 is not required in the initial stages of dehydration, but is only expressed when dehydration becomes severe and the plant has dried down to approximately 50% RWC. No expression of the gene was evident at 4% RWC. Since XVSAP1 is likely to be an integral membrane protein, one of the roles it could play is the stabilization of membranes during the drying process. As the plant dries further, its metabolic processes decline and eventually stop. This could explain why the expression of XVSAP1 is not observed at 4% RWC. As no expression of XVSAP1 was observed during rehydration, this indicates that XVSAP1 has no role to play during this process. However, the absence of the transcript does not imply absence of the protein product. It has been observed that most of the components required for recovery from desiccation are already present in the plant during dehydration in X. humilis (Dace et al., 1998). In X. viscosa, XVSAP1 could be one of those components involved in the repair of membrane damage that results from severe water deficit. Studies of expression at the protein level would clarify whether XVSAP1 is involved in just dehydration or whether the protein is also part of the rehydration process. Heat and high light stresses both strongly induced the accumulation of XVSAP1 mRNA. In both cases, the transcript only started to accumulate at least 48 h after imposition of the stress. The results obtained suggest that XVSAP1 is not involved in the initial stages of the response to heat or high light intensity. It is expected that X. viscosa would have mechanisms in place to deal with such stresses in the short term as the extremophile grows in environments where it is regularly exposed to high temperature and high light intensity. However, when the duration of these stresses increases, other mechanisms that would have a protective effect come into play. It is known that heat stress affects most cellular processes and causes denaturation of proteins, cellular enzymes, and damage to membranes. The damage is due to the temperature change itself as well as heat‐induced oxidative stress (Karim et al., 1999; Munro and Pelham, 1985). Survival after heat stress requires an ability to tolerate or repair oxidative damage as well other kinds of heat‐induced damage. It is expected that XVSAP1 would have a role to play in the protection of membranes against heat damage. In a similar manner, high light intensity can result in the formation of reactive oxygen species (ROS). If the free radicals are not quenched, damage in the form of photo‐bleaching and lipid peroxidation occurs (Smirnoff, 1993). X. viscosa appears to withstand damage caused by light by a combination of protective and avoidance mechanisms. The poikolochlorophyllous resurrection plant loses its chlorophyll and dismantles its photosynthetic apparatus, while the levels of anthocyanins and antioxidant enzymes increase, affording the plant a degree of protection (Sherwin and Farrant, 1998). As with heat stress, it is proposed that XVSAP1 is involved in the protection of membranes possibly by maintaining structural integrity. High exogenous salt concentrations cause an imbalance of cellular ions resulting in ion toxicity, osmotic stress and production of ROS (Hasegawa et al., 2000). Various molecules including proteins that protect membrane integrity, control ion homeostasis and play a role in ROS scavenging have been reported to attenuate salt stress effects (Hasegawa et al., 2000; Ingram and Bartels, 1996; Ishitani et al.,1997). Studies in both wheat and barley showed that the induction of genes by salt occurs within 2 h and that many transcripts decrease in abundance within 24 h (Robinson et al., 1990). However, in the case of XVSAP1, the transcript appeared within 24 h after salt shock and persisted for 7 d. The response XVSAP1 to salt is again delayed compared to other salt‐responsive genes. In addition, it also lasts for a longer period, supporting the earlier theory that the gene is expressed on persistence of a particular abiotic stress. XVSAP1 has a relatively high identity to cold‐responsive WCOR413 and the uncharacterized Arabidopsis homologues (Fig. 3A). It was therefore reasonable to consider that the gene could be induced by cold. This proved to be the case. XVSAP1 was detected within 24 h after the commencement of the treatment. The results correlate well with those obtained with other COR genes. Cold‐induced mRNAs generally begin to accumulate within a few hours at low temperature and remain at high levels until removal of the stress. The CBF genes are induced within 15 min of plants being exposed to low temperatures followed, at about 2 h, by the induction of cold‐regulated genes that contain the CRT/DRE regulatory element (Gilmour et al., 1998; Thomashow, 1998). It is expected that in the natural habitats of X. viscosa, temperatures at night could go well below zero on occasion. XVSAP1 would therefore form part of the mechanism that assists X. viscosa to cope with the stress, particularly since chilling injury is mainly a consequence of destablization of cell membranes. It has been established that many genes that respond to drought and/or cold stress are also induced by exogenous applications of ABA (Bray, 1997; Chandler and Robertson, 1994). However, in the case of X. viscosa, ABA treatment in planta failed to induce XVSAP1. Moreover, placing of excised leaves in a 100 µM solution of ABA did not have an effect on the expression of XVSAP1 despite the fact that less competitor RNA was used in the RT‐PCR reaction (results not shown). Shinozaki and Yamaguchi‐Shinozaki (1997) suggested that there are at least four independent signal pathways that function in the activation of stress‐inducible genes. Two of these are ABA‐dependent (pathways I and II) and two are ABA‐independent (pathways III and IV). The fact that XVSAP1 was not induced by exogenous applications of ABA suggests that XVSAP1 responds to environmental stresses through an ABA‐independent pathway. It is also possible that the response to ABA is transient and was not detectable under the experimental conditions used. The data presented here show that XVSAP1 is a stress‐associated gene in X. viscosa. The fact that the gene is induced by heat, high salt, cold, and dehydration is not surprising since the gene was isolated on the basis of its response to osmotic stress. It is known that the above abiotic stresses all result in water deficit in the cell. It is predicted that the protein product would play a protective role possibly in stabilizing cell membranes during dehydrative stresses. Acknowledgements We thank Professor JM Farrant for collecting and maintaining the X. viscosa plants and Mrs D James for sequencing of the cDNA. We also thank Professor F Sarhan for the vector construct expressing WCOR413.This work was partially supported by the Tobacco Research Board, Zimbabwe. SGM acknowledges the financial support received from the National Research Foundation (RSA) and the University of Cape Town Research Committee. Open in new tabDownload slide Fig. 1. Growth analysis of E. coli (srl::Tn10) cells transformed with the prokaryotic protein expression vector pPROEXHT (open circles) and cells transformed with pPROEXHT‐XVSAP1 (closed circles) in minimal media. The expression of XVSAP1 was induced with IPTG at time zero. The cells were allowed to grow for a further 2 h before sorbitol was added to a final concentration of 1 M. Samples were taken at intervals and the absorbance at 600 nm determined. Error bars represent standard deviation based on the average of triplicate samples. Open in new tabDownload slide Fig. 2. (A) The nucleotide and deduced amino acid sequences of XVSAP1. The start and stop codons are indicated by shading. The N‐myristoylation site is indicated by broken lines and the prokaryotic membrane lipoprotein lipid attachment sites are underlined by bold lines. The regions of the six transmembrane helices are indicated in bold letters. (B) A hydrophobicity profile of XVSAP1 protein as determined by the method of Kyte and Doolittle (1982) using a window of 19 amino acid residues. The six putative transmembrane domains are indicated by Roman numerals. Open in new tabDownload slide Fig. 3. (A) Comparison of the deduced amino acid sequence of XVSAP1 with related proteins. The proteins are WCOR413 (GenBank Accession No. T06810), a cold‐regulated protein from wheat, cold associated proteins (ATCAP1–4: GenBank Accession Nos T08404, AAD41971, CAB16776, and T02423, respectively) from Arabidopsis and an Oryza sativa cold‐acclimation WCOR413‐like protein (RCAP: GenBank Accession No. AAG13395). Only Arabidopsis proteins showing the highest homology are shown. Percentages following the sequences indicate the per cent identity to XVSAP1. Identical amino acid residues are indicated by asterisks. Similar amino acids are shown by points. (B) Homology tree for XVSAP1 and related proteins. Open in new tabDownload slide Fig. 4. Southern blot analysis of genomic DNA from X. viscosa.15 µg of DNA was cut with EcoRI/XhoI (lane 1), BglII (lane 2), HindIII (lane 3), and EcoRV (lane 4), electrophoresed on a 1% agarose gel, transferred to a nylon membrane and probed with DIG‐labelled XVSAP1 cDNA. Open in new tabDownload slide Fig. 5. Expression of the XVSAP1 transcript during dehydration and rehydration of X. viscosa was determined using semi‐quantitative RT‐PCR. PCR products were visualized on agarose gels stained with ethidium bromide. The native XVSAP1 product was 830 bp and the competitor 345 bp. M refers to the marker lane. Percentages refer to the relative water content (RWC). (A) X. viscosa plants were dehydrated from 90% RWC to 4% RWC by withholding water for a period of 2 weeks. (B) The same X. viscosa plants were rehydrated over 5 d by watering. (C) The RWC variation during the dehydration–hydration of X. viscosa. Open in new tabDownload slide Fig. 6. Induction of XVSAP1 by heat, NaCl, high light intensity, and low temperature treatments in X. viscosa was compared using semi‐quantitative RT‐PCR. PCR products were visualized on agarose gels stained with ethidium bromide. The native XVSAP1 product was 830 bp and the competitor 345 bp. M refers to the marker lane. (A) Plants were kept in a 42 °C incubator for 7 d to induce heat stress. (B) Salt shock was induced by irrigating potted plants with 100 mM NaCl. (C) Plants were exposed to high light at 1500 µmol m–2 s–1 for 4 d in a phytotron (25 °C, humidity 50–70%) for the high light treatment (D) Cold treatment was achieved by keeping plants at 4 °C for 60 h. References AltschulSF, Gish W, Miller W, Myers EW, Lipman DJ. 1990 . Basic local alignment search tool. Journal of Molecular Biology 215, 403 –410. BartelsD, Schneider K, Terstappen G, Piatkowski D, Salami F. 1990 . Molecular cloning of abscisic acid modulated genes which are induced during desiccation of the resurrection plant Craterostigma plantagineum. Planta 181, 27 –34. BlomstedtCK, Gianello RD, Hamill JD, Neale AD, Gaff DF. 1998 . Drought‐ stimulated genes correlated with desiccation tolerance of the resurrection grass Sporobolus stapfianus. Plant Growth Regulation 24, 210 –228. BoyerJS. 1982 . Plant productivity and environment. Science 218, 443 –448. BrayEA. 1997 . Plant responses to water deficit. Molecular Plant Science 22, 48 –54. ChandlerPM, Robertson M. 1994 . Gene expression regulated by abscisic acid and its relation to stress tolerance. Annual Review of Plant Physiology and Plant Molecular Biology 45, 113 –141. CsonkaLN, Clark AJ. 1979 . Deletions generated by the transposon Tn10 in the srl recA region of Escherichia coli K‐12 chromosome. Genetics 93, 321 –343. DaceH, Sherwin H, Illing N, Farrant JM. 1998 . Use of metabolic inhibitors to elucidate mechanisms of recovery from desiccation stress in the resurrection plant Xerophyta humilis. Plant Growth Regulation 24, 171 –177. DanylukJ, Houde M, Sarhan F. 1994 . Differential expression of a gene encoding an acidic dehydrin in chilling‐sensitive and freezing‐tolerant Gramineae species. FEBS Letters 344, 20 –24. DanylukJ, Sarhan F. 1996 . Identification and characterization of a low temperature‐regulated gene encoding a hydrophobic protein from wheat. Montreal, Canada: Universite de Quebec a Montreal, Sciences Biologiques. DellaportaSJ, Wood J, Hicks JB. 1983 . A plant DNA mini preparation: version II. Plant Molecular Biology Reporter 1, 19 –21. GaffDF. 1971 . Desiccation‐tolerant plants in southern Africa. Science 174, 1033 –1034. GilmourSJ, Hajela RK, Thomashow MA. 1988 . Cold acclimation in Arabidopsis. Plant Physiology 87, 745 –750. GilmourSJ, Zarka DG, Stockinger EJ, Salazar MP, Houghton JM, Thomashow MF. 1998 . Low temperature regulation of the Arabidopsis CBF family of AP2 transcriptional activators as an early step in cold‐induced COR gene expression. The Plant Journal 16, 433 –443. HasegawaPM, Bressan RA, Zhu J‐K, Bohnert HJ. 2000 . Plant cellular and molecular responses to high salinity. Annual Review of Plant Physiology and Plant Molecular Biology 51, 463 –499. HayashiS, Wu HC. 1990 . Lipoproteins in bacteria. Journal of Bioenergetics and Biomembranes 22, 3451 –3471. HazelJR. 1995 . Thermal adaptation in biological membranes: is homeoviscous adaptation the explanation? Annual Review of Physiology 75, 19 –42. IngramJ, Bartels D. 1996 . The molecular basis of dehydration tolerance in plants. Annual Review of Plant Physiology and Plant Molecular Biology 47, 377 –403. IshitaniM, Liu J, Halfter U, Kim C‐S, Shi W, Zhu J‐K. 2000 . SOS3 function in plant salt tolerance requires N‐myristoylation and calcium binding. The Plant Cell 12, 1667 –1677. IshitaniM, Xiong L, Stevenson B, Zhu JK. 1997 . Genetic analysis of osmotic and cold stress signal transduction in Arabidopsis: interactions and convergence of abscisic acid‐dependent and abscisic acid‐independent pathways. The Plant Cell 9, 1 –16. ItturiagaG, Schneider K, Salamini F, Bartels D. 1992 . Expression of desiccation‐related proteins from the resurrection plant Craterostigma plantagenium in transgenic tobacco. Plant Molecular Biology 20, 555 –558. JohnsonDR, Bhatnagarm RS, Knoll LJ, Gordon JI. 1994 . Genetic and biochemical studies of protein N‐myristoylation. Annual Review of Biochemistry 63, 869 –914. KarimMA, Fracheboud Y, Stamp P. 1999 . Photosynthetic activity of developing leaves of Zea mays is less affected by heat stress than that of developing leaves. Physiologia Plantarum 105, 685 –693. KyteJ, Doolittle RF. 1982 . A simple method for displaying the hydropathic character of a protein. Journal of Molecular Biology 157, 105 –132. Le RudulierD, Strom AR, Dandekar AM, Smith LT, Valentine RC. 1984 . Molecular biology of osmoregulation. Science 224, 1064 –1068. LiuQ, Kasuga M, Sakuma Y, Abe H, Miura S, Yamaguchi‐Shinozaki K, Shinozaki K. 1998 . Two transcription factors, DREB1 and DREB2, with an EREBP/AP2 DNA binding domain separate two cellular signal transduction pathways in drought‐ and low‐temperature responsive gene expression, respectively, in Arabidopsis. The Plant Cell 10, 1391 –1406. MarcotteJr WD, Russel SH, Quatrano RS. 1989 . Abscisic acid‐responsive sequences from the Em gene of wheat. The Plant Cell 1, 969 –976. McCartyDR. 1995 . Genetic control and integration of maturation and germination pathways in seed development. Annual Review of Plant Physiology and Plant Molecular Biology 46, 71 –93. McNeilSD, Nuccio ML, Hanson AD. 1999 . Betaines and related osmoprotectants. targets for metabolic engineering of stress resistance. Plant Physiology 120, 945 –959. MundreeSG, Whittaker A, Thomson JA, Farrant JM. 2000 . An aldose reductase homolog from the resurrection plant Xerophyta viscosa Baker. Planta 211, 693 –700. MundyJ, Yamaguchi‐Shinozaki K, Chua N‐H. 1990 . Nuclear proteins bind conserved elements in the abscisic acid‐responsive promoter of a rice RAB gene. Proceedings of the National Academy of Sciences, USA 87, 1406 –1410. MunroS, Pelham H. 1985 . Molecular genetics: what turns on heat shock genes. Nature 317, 477 –478. NealeAD, Blomstedt T, Bronson TN, Guthridge LK, Evans J, Gaff, Hamill JD. 2000 . The isolation of genes from the resurrection grass Sporobolus stapfianus which are induced during severe drought stress. Plant, Cell and Environment 23, 265 –277. OliverMJ. 1991 . Influences of protoplasmic water loss on the control of protein synthesis in the desiccation‐tolerant moss Tortula ruralis: ramifications for a repair‐based mechanism of desiccation‐tolerance. Plant Physiology 97, 1501 –1511. OliverMJ. 1996 . Desiccation tolerance in vegetative plant cells. Physiologia Plantarum 97, 779 –787. OliverMJ, Bewley JD. 1997 . Desiccation tolerance of plant tissues: a mechanistic overview. Horticultural Reviews 18, 171 –214. OliverMJ, Wood AJ, O’Mahony P. 1998 . ‘To dryness and beyond’—preparation for the dried state and rehydration in vegetative desiccation‐tolerant plants. Plant Growth Regulation 24, 193 –201. OulletF, Vasquez‐Tello A, Sarhan F. 1998 . The wheat wcs120 promoter is cold‐inducible in both monocotyledonous and dicotyledonous species. FEBS Letters 423, 324 –328. QuinteroFJ, Blatt MR. 1997 . A new family of K+ transporters from Arabidopsis that are conserved across phyla. FEBS Letters 415, 206 –211. RobinsonNl, Tanaka CK, Hurkaman WJ. 1990 . Time‐dependent changes in polypeptide and translatable mRNA levels caused by NaCl in barley roots. Physiologia Plantarum 78, 128 –134. SambrookJ, Fritsch EF, Maniatis T. 1989 . Molecular cloning: a laboratory manual, 2nd edn. New York: Cold Spring Harbor Laboratory Press. SchneiderK, Wells B, Schmelzer E, Salamini F, Bartels D. 1993 . Desiccation leads to the rapid accumulation of both cytosolic and chloroplastic proteins in the resurrection plant Craterostigma plantagineum Hochst. Planta 189, 120 –131. SherwinH, Farrant JM. 1996 . Differences in rehydration of three desiccation‐tolerant angiosperm species. Annals of Botany 78, 703 –710. SherwinHW, Farrant JM. 1998 . Protection mechanisms against excess light in the resurrection plants Craterostigma wilmsii and Xerophyta viscosa. Plant Growth Regulation 24, 203 –210. ShinozakiK, Yamaguchi‐Shinozaki K. 1997 . Gene expression and signal transduction in water‐stress response. Plant Physiology 115, 327 –334. SmirnoffN. 1993 . Role of active oxygen in the response of plants to water deficit and desiccation. New Phytologist 125, 27 –58. StockingerEJ, Gilmour SJ, Thomashow MF. 1997 . Arabidopsis thaliana CBF1 encodes an AP2 domain‐containing transcriptional activator that binds to the C‐repeat/DRE, a cis‐acting DNA regulatory element that stimulates transcription in response to low temperature and water deficit. Proceedings of the National Academy of Sciences, USA 94, 1035 –1040. ThomashowMF. 1998 . Role of cold‐responsive genes in plant freezing tolerance. Plant Physiology 118, 1 –7. ThomashowMF. 1999 . Plant cold acclimation, freezing tolerance genes and regulatory mechanisms. Annual Review of Plant Physiology and Plant Molecular Biology 50, 571 –599. ThomashowMF. 2001 . So what’s new in the field of plant cold acclimation? Lots! Plant Physiology 125, 89 –93. TubaZ, Lichtenthaler HK, Marioti I, Csintalan Z. 1993 . Resynthesis of thylakoids and functional chloroplasts in the desiccated leaves of the poikilochlorophyllous plant Xerophyta scabrida upon rehydration. Journal of Plant Physiology 142, 742 –748. ZeevaartJAD, Creelman RA. 1988 . Metabolism and physiology of abscisic acid. Annual Review of Plant Physiology and Plant Molecular Biology 39, 439 –473. ZhuJ‐K, Hasegawa PM, Bressan RA. 1997 . Molecular aspects of osmotic stress in plants. Critical Review of Plant Science 16, 253 –277.

Moyano, Elisabet; Jouhikainen, Katja; Tammela, Päivi; Palazón, Javier; Cusidó, Rosa M.; Piñol, M. Teresa; Teeri, Teemu H.; Oksman‐Caldentey, Kirsi‐Marja

doi: 10.1093/jxb/erg014pmid: 12493848

Abstract In order to increase the production of the pharmaceuticals hyoscyamine and scopolamine in hairy root cultures, a binary vector system was developed to introduce the T‐DNA of the Ri plasmid together with the tobacco pmt gene under the control of CaMV 35S promoter, into the genome of Datura metel and Hyoscyamus muticus. This gene codes for putres cine:SAM N‐methyltransferase (PMT; EC. 2.1.1.53), which catalyses the first committed step in the tropane alkaloid pathway. Hairy root cultures overexpressing the pmt gene aged faster and accumulated higher amounts of tropane alkaloids than control hairy roots. Both hyoscyamine and scopolamine production were improved in hairy root cultures of D. metel, whereas in H. muticus only hyoscyamine contents were increased by pmt gene overexpression. These roots have a high capacity to synthesize hyoscyamine, but their ability to convert it into scopolamine is very limited. The results indicate that the same biosynthetic pathway in two related plant species can be differently regulated, and overexpression of a given gene does not necessarily lead to a similar accumulation pattern of secondary metabolites. Key words: Datura metel, hairy root cultures, hyoscyamine, Hyoscyamus muticus, 35S‐pmt gene, scopolamine, tropane alkaloids. Received 5 March 2002; Accepted 14 August 2002 Introduction Hairy roots result from the transfer and integration of the genes located on the root‐inducing plasmid Ri of Agrobacterium rhizogenes into the plant genome and their expression therein (White and Nester, 1980). These types of roots are characterized by fast growth, frequent branching, plagiotropism, and the ability to synthesize the same compounds as the roots of the intact plant (David et al., 1984). Roots of Solanaceae plants are the main site of tropane alkaloid biosynthesis, and hence hairy root cultures are also capable of accumulating high levels of these metabolites (Oksman‐Caldentey and Arroo, 2000). On the contrary, since tropane alkaloid production is very low in unorganized in vitro cultures of Solanaceae (Hashimoto and Yamada, 1983; Yamada and Endo, 1984; Oksman‐Caldentey and Strauss, 1986), it seems that the tropane alkaloid pathway requires root organization in order to be developed completely. This has earlier been demonstrated by Palazón et al. (1995) when Datura stramonium L. root cultures treated with auxin lost their organization and were not able to synthesize tropane alkaloids. Similarly, hairy root cultures of several Solanaceae plants displaying callus‐like morphology achieved very low tropane alkaloid contents (Moyano et al., 1999; Jouhikainen et al., 1999). Until very recently the lack of understanding of the regulation of secondary metabolite pathways has limited the general use of metabolic engineering in medicinal plants. Hyoscyamine is usually the main alkaloid in transgenic root cultures of many Solanaceae plants including Hyoscyamus, while scopolamine is only produced in small amounts, except Datura metel L. and Duboisia sp. which accumulate this alkaloid at high levels (Muranaka et al., 1993; Celma et al., 2001). However, recently it has been shown that an increase in the expression of the hyoscyamine‐6β‐hydroxylase (h6h) gene coding for the last enzyme involved in the tropane alkaloid biosynthesis, can considerably enhance the production of scopolamine in hairy root cultures of Hyoscyamus muticus L. (Jouhikainen et al., 1999). Putrescine‐N‐methyltransferase (PMT) is the enzyme catalysing the first committed step in the tropane alkaloid pathway converting putrescine into N‐methylputrescine. Although PMT occurs at higher activity levels in D. stramonium hairy root cultures than either arginine decarboxylase or ornithine decarboxylase, and in this sense does not appear to be rate‐limiting, this enzyme can limit the flux of putrescine into tropane alkaloids since its expression is very sensitive to the environment where the roots are growing (Piñol et al., 1999). Previously Robins et al. (1991) reported that hairy root cultures of Datura supplemented with auxin lost PMT activity before showing any morphological alteration. Moreover, Biondi et al. (2002) have documented a considerable increase of PMT activity as well as an increase in polyamine contents and tropane alkaloids after methyljasmonate treatment in transformed roots of H. muticus. However, it has recently been shown that overexpression of PMT in Atropa belladonna L. does not affect tropane alkaloid levels either in transgenic plants or in hairy roots (Sato et al., 2001). In the present work, the gene from Nicotiana tabacum L. coding for PMT has been inserted into hairy roots of D. metel and H. muticus under the control of the constitutive CaMV 35S promoter, in order to influence tropane alkaloid production. Moreover, the morphology and alkaloid production capacity of these engineered hairy roots have been compared with the respective ones of control hairy roots. As far as it is known, this is the first time that the overexpression of the tobacco pmt gene has been demonstrated to improve tropane alkaloid production in hairy root cultures in a plant species‐dependent manner. Materials and methods Bacterial strains and cultures The Escherichia coli strain DH5α and Agrobacterium tumefaciens strain C58C1(pRiA4) were used. The latter is a cured derivative of the nopaline strain C58 (Van Larebeke et al., 1974) containing the pRi from A. rhizogenes A4. The E. coli strain was grown at 37 °C in LB medium (Miller, 1972) and the Agrobacterium strain at 28 °C in YEB medium (Maniatis et al., 1982). Construction of the plasmid pBMI The pTVPMT plasmid carrying the cDNA which codes for putrescine‐N‐methyltransferase from tobacco was a gift from Professor T Hashimoto of the Nara Institute of Science and Technology (Kyoto, Japan). In order to construct the binary vector known as pBMI, the NcoI‐BamHI fragment of 1400 bp, corresponding to the cDNA for PMT was isolated from pTV‐PMT and then subcloned in pET‐3d (Novogen), after separating the NcoI‐BamHI fragment from this plasmid. The plasmid pET‐3d was used because the strong expression vector for plants, pRoc 2275‐C does not have a NcoI site. The resulting plasmid pET‐3d‐PMT was introduced into competent cells of E. coli DH5α strain following the method described by Hanahan (1983). Transformants were screened for ampicillin resistance and subsequently characterized by restriction endonuclease analysis. The 1439 bp XbaI‐BamHI fragment from the plasmidic cDNA of the ampicillin‐resistant E. coli cells was subcloned into the plasmid pRoc‐2275‐C, after separating the same fragment from this plasmid. In the subcloning steps carried out for constructing the pBMI plasmid, ligations were cohesive ended. The T‐DNA region of the resulting binary vector (flanked by the 25 bp right and left border sequences), in addition to the cDNA for tobacco PMT subcloned between the 35S promoter and the terminal sequence of the gene for nopaline synthesis, also carried the gene for neomycine phosphotransferase (NPTII) under the control of the nopaline synthesis promoter (Nos‐Pro) and the terminal sequence of the enzyme itself (Nos‐Ter). This plasmid, known as pBMI was kept in E. coli DH5α and then transferred to the A. tumefaciens C58C1(pRiA4) strain as described by Mozo and Hooykaas (1991). The recombinant clones of A. rhizogenes were selected following Kado and Liu (1981). Establishment and culture of transformed roots of Datura metel Sterile leaf sections of D.metel L. strain metel were inoculated with A. rhizogenes strain A4 or A. tumefaciens strain C58C1(pRiA4, pBMI) carrying the pmt gene (Cusidó et al., 1999). Roots which appeared 4–6 weeks after the inoculation, were cultured separately on solid Gamborg B5 medium (Gamborg et al., 1986) supplemented with carbenicillin (500 mg l–1) to eliminate the bacteria and, in the case of transformed roots carrying the pmt gene, with kanamycin 50 mg l–1. Rapidly growing clones with no bacterial contamination were used to establish the cultures of hairy root clones. After several subcultures, transformed roots were transferred to half‐strength B5 solid medium (B5/2). Each hairy root clone was also cultured in liquid medium (40 ml of B5/2 hormone‐free medium in Erlenmeyer flasks), maintained on an orbital shaker at 100 rpm at 25 °C in the dark, and subcultured every 4 weeks. Establishment and culture of transformed roots of Hyoscyamus muticus Young plants of H. muticus L. strain Cairo were used for transformation with Agrobacterium as for D. metel (Vanhala et al., 1995). Each induced hairy root was cultivated separately in liquid culture (5–10 ml), first in the presence of cefotaxime (500 mg l–1) to remove excess bacteria and then in modified Gamborg’s B50 medium (20 ml) in 100 ml conical flasks (Oksman‐Caldentey et al., 1991). The hairy root clones were routinely subcultured every 3 weeks as described above. Extraction and determination of alkaloids The extraction and determination of scopolamine and hyoscyamine in the transformed roots of D. metel were carried out as reported earlier (Piñol et al., 1996). In H. muticus, quantitative tropane alkaloid determinations were performed on methanol extracts of hairy root clones and on samples of the medium using enzyme immunoassay (Vanhala et al., 1998) and radioimmunoassay (Oksman‐Caldentey et al., 1987) for scopolamine and hyoscyamine, respectively. Polymerase chain reaction analysis The presence of the transferred Agrobacterium rol and pmt genes was tested by PCR as described previously Sevón et al. (1995) and Moyano et al. (1999). Genomic DNA was extracted from the putative engineered hairy root clones and wild‐type hairy root clones according to Edwards et al. (1991). The oligonucleotide primers used for amplification of the pmt gene were 5′‐GCCATT CCCATGAACGGCC‐3′ (position 108–127 nt) and 5′‐CCTCCG CCGATGATCAAAACC‐3′ (position 569–549 nt), according to the sequence data of the pmt gene from Nicotiana tabacum L. (Hibi et al., 1994). The complete PCR mixture contained 200 ng total DNA, 12.5 pmol µl–1 of each oligonucleotide primer, 200 µM dNTPs, 1.5 U Taq Polymerase (Pharmacia Biotech.) and buffer supplied by the enzyme manufacturer (1/10 V) in a total volume of 25 µl. PCR was carried out in the following conditions: 1 cycle of 95 °C, 5 min; 30 cycles of 94 °C, 1 min; 62 °C, 1 min; 72 °C 1.5 min; 1 cycle of 72 °C, 5 min. Products (10 µl) were analysed on 1.5% agarose/TBE gel where the expected size was 450 bp. Putrescine N‐methyltransferase (PMT) activity assay PMT activity in transgenic root cultures of H. muticus was measured principally with the method previously employed by Feth et al. (1985) and Hibi et al. (1992). All extraction and purification procedures were performed at +4 °C. First, root material (2 g FW) was washed with tap water and crushed in 2.5 ml of potassium‐phosphate buffer A (100 mM, pH 7.5) containing 0.25 M sucrose, 5 mM EDTA, 0.5% (w/v) sodium ascorbate, 3 mM dithiothreitol (DTT), and protease inhibitor mix (Complete™ Mini, Boehringer Mannheim). This suspension was homogenized with a pinch of sea sand and mixed with 10% (w/v) insoluble polyvinylpolypyrrolidone (PVPP). After 10 min centrifugation at 13 000 rpm, the supernatant was diluted with potassium‐phosphate buffer B (50 mM, pH 7.5, containing 1 mM EDTA and 1 mM DTT), and added to a PD‐10 column (Pharmacia) previously equilibrated with the same buffer. Proteins were eluted with 3.5 ml of buffer B and collected into four fractions (2 mM DTT and protease inhibitor mix added). Protein concentration of each fraction was determined according to Bradford (1976). The protein‐richest fraction was then selected for PMT activity testing performed as in Feth et al. (1985). Briefly, the enzyme reaction was performed in buffer B with putrescine and SAM concentrations of 3.6 mM and 0.6 mM, respectively. After 30 min incubation at 30 °C, the reaction was stopped by 1 min heating at 100 °C. Samples were stored at –20 °C. N‐methylputrescine concentrations of the samples were determined with HPLC analysis after dansylation. The HPLC system consisted of an Agilent Hypersil BDS‐C18 column (5 µm, 4.6×150 mm), Waters 600 Pump and Waters 2487 Dual λ Absorbance Detector (detection wavelength 217 nm). The isocratic mobile phase consisted of an aqueous solution of 2% H3PO4 adjusted to pH 5.2 by triethylamine and acetonitrile (40:60, flow rate 1 ml min–1). Northern blot analysis and quantification of pmt gene product Northern blot analysis was carried out with seven randomly selected transformed clones of D. metel and with all the transformed root clones of H. muticus. Total RNA was isolated from the hairy roots after 2 weeks of cultivation using the RNeasy kit (Qiagen). Ten micrograms total RNA were loaded per lane of a denaturating formaldehyde gel. Northern blots were made on Hybond‐N+ nylon transfer membranes (Amersham) and hybridized with a 32P‐labelled probe specific for pmt (full cDNA). Hybridization was performed at 42 °C in the presence of 50% formamide. The blots were washed at 42 °C with 2× saline sodium citrate buffer (SSC), 0.1% SDS (30 min) and with 1× SSC, 0.1% SDS (15 min). The radioactivity on the filter was imaged using Hyperfilm™‐MP (Amersham). The band of 1.4 kb that appears in the Northern blot corresponded to the mRNA of pmt gene. The intensity of this band was quantified using Phoretix software as described by Palazón et al. (1998). Results Growth and root morphology Transformed root cultures of D. metel and H. muticus overexpressing the pmt gene appeared 4–6 weeks after the Agrobacterium inoculation in 80–90% of the leaf explants. Several phenotypic characteristics, e.g. branching, plagiotropism, colour, and callus formation of the hairy root clones, were followed during their growth in order to compare the morphology and the alkaloid production of the clones. Transgenic root cultures of Datura were preselected by culturing them in a medium supplemented with kanamycin. Later, the presence of the pmt gene in the root genome was confirmed in 100% of root clones tested by PCR. In Hyoscyamus, 15 hairy root clones out of 20 (75%) were found to be positive for the pmt transgene in the PCR studies. The phenotype and morphology of the Datura and Hyoscyamus root clones carrying the pmt gene differed from control clones (carrying Agrobacteriumrol genes but not the 35S‐pmt transgene). At the start of growth on solid medium, purple pigment formation, which later turned brown, most probably indicating the presence of phenolic compunds such as anthocyanins, was observed in the roots of Hyoscyamus (Fig. 1). There was also slight purple pigmentation in the control roots induced by strain C58C1(pRiA4); thus this peculiar colour formation in the hairy roots might be due to the pRiA4 plasmid in this plant species. This phenomenon has never been found in Hyoscyamus hairy root clones transformed with other plasmids. In the case of Datura root cultures, a very high percentage (80%) of the pmt transformed root cultures established, aged faster and turned brown when cultured on solid medium. These root clones were discarded and only root cultures showing a normal growth capacity were considered for further experiments. In Hyoscyamus the majority of the clones were brownish, but otherwise their appearance and growth was normal. Regarding the biomass production of Datura root cultures overexpressing the tobacco pmt gene, these results did not show any clear effect of the transgene on the root growth ability after 4 week’s cultivation (Fig. 2), since the fresh and dry weight values achieved by the transgenic roots were very similar to the average of those obtained by the three control root lines studied. However, considerable variation in growth capacity between the individual clones was observed, for example, clone PMT22 achieved two‐fold higher fresh weight at the end of the growth period than clone PMT15. As previously reported (Sevón et al., 1998; Moyano et al., 1999), this fact is very frequent in transformed root cultures, since each clone arises from an independent transformation event. Depending on the presence of certain T‐DNA genes, mainly the aux1 gene, in the genome of the root clone obtained, growth can be very different. Growth capacity, measured both as fresh and dry weight, of Hyoscyamus roots was very much higher when compared to Datura clones (Fig. 2). This fact revealed that transformed roots of Hyoscyamus constitute a very effective system to produce biomass of hairy roots in vitro. On the other hand, all Hyoscyamus root clones carrying the 35S‐pmt gene reached significantly (P ≥0.01) lower fresh and dry weight than control roots (three control clones) after 4 weeks of culture, showing a negative effect of the transgene on the growth capacity. Tropane alkaloid production The pmt transformed root clones of Datura displayed an alkaloid spectrum similar to that of the transformed control roots. As also mentioned in a previous report (Cusidó et al., 1999), scopolamine was the main alkaloid obtained in all the clones (Fig. 3). Clone PMT10 was the most productive among all the Datura root clones established. Great differences in alkaloid contents among the clones were observed, but the majority of PMT clones reached higher scopolamine and hyoscyamine levels than the control ones after 4 weeks of growth, with only few clones (e.g. PMT22 and PMT31) presenting lower alkaloid contents than control ones at the end of the culture period (Fig. 3). Transformed roots of Hyoscyamus accumulated much higher hyoscyamine amounts than the ones of Datura (approximately 10‐fold), whereas scopolamine contents were very low (lower than 0.2 mg l–1). This hyoscyamine‐rich genotype is typical for hairy root cultures of H. muticus (Vanhala et al., 1995; Sevón et al., 1998) indicating very low activity of the h6h gene in this plant species (Jouhikainen et al., 1999). The contents of hyoscyamine in the pmt transformed hairy roots were on average 2–3‐fold higher (Fig. 3), but the scopolamine contents were similar to the controls. These results, in accordance with this study’s predictions, showed that the overexpression of the tobacco pmt gene under the control of the 35S promoter enhanced alkaloid production of D.metel‐transformed roots, and this improvement affected both hyoscyamine and scopolamine contents. Furthermore, the results indicated the capacity of Datura roots partially to convert the additional amount of hyoscyamine into scopolamine and revealed that the H6H activity in these roots is not low compared to the ones of Hyoscyamus. The pmt gene expression In order to establish a possible relationship between transgene expression and the capacity of root cultures to synthesize tropane alkaloids, the levels of the tobacco pmt gene product in the transformed roots were studied. For this reason Datura and Hyoscyamus root clones with differing capacities to biosynthesize tropane alkaloids were selected. RNA gel blot results showed (Fig. 4) a clear band of 1.4 kb corresponding to cDNA of pmt in the selected transgenic root clones, whereas in the control ones no band of 1.4 kb was visible. The same quantity of total RNA was charged in each lane of the gel and the intensity of the pmt band was analysed photodensitometrically by Phoretix software. These experiments were carried out in triplicate although only one is shown. Figure 5 represents the results of densitometrical analysis, showing in both types of roots, a relationship between pmt gene expression and the capacity of the culture to biosynthesize alkaloids. In all cases, roots with a high intensity of the pmt band reached a high alkaloid content (Fig. 5). For example, the two best producing clones of Hyoscyamus KC2 and KC17 also showed the strongest pmt expression in the Northern blot experiments. In addition, the PMT enzymatic activity and the alkaloid content in transgenic Hyoscyamus hairy roots were measured simultaneously (Fig. 6). The PMT activity was the highest after 7 d of cultivation when the alkaloid contents were still low, and gradually decreased with the time. However, the ratio of the PMT activity of individual clones remained the same through the whole cultivation period. The highest PMT activity was observed in clone KC17, which had the strongest pmt expression according to northern blot as well. This clone also presented the second highest hyoscyamine content among all the clones studied. However, in some transgenic clones this kind of positive correlation seems to be absent suggesting that PMT activity is not the sole factor contributing to elevated alkaloid contents. Discussion A binary vector was constructed to introduce the tobacco pmt gene and the T‐DNA of pRiA4 of Agrobacterium rhizogenes, into D. metel and H. muticus explants. Results of agroinfection revealed an excellent capacity of the construction to transfer the T‐DNA of pRiA4 into the plant genome and to develop the hairy root syndrome. The majority of the established transformed root cultures, also overexpressed the pmt gene under the control of the 35S promoter and showed several altered metabolic traits, such as rapid ageing, reduction of growth in Hyoscyamus cultures and modifications in secondary metabolite pathways. According to Kholodenko et al. (1998) an undesirable effect of metabolic engineering is the promotion of metabolic flux alterations that can induce cell death. This fact produces metabolic changes not directly related to transgene presence. For this reason, brownish root clones of Datura were rejected and the remaining ones were maintained for further studies. As previously mentioned, in Hyoscyamus this kind of selection was not performed and transgenic roots carrying the pmt gene grew less rapidly than the control roots. Rapid ageing of transformed roots has previously been connected with high alkaloid production (Jouhikainen et al., 1999), but the altered phenotype obtained in this study might be due to the formation of other compounds. On the other hand, due to the role of polyamines in plant development (Martin‐Tanguy, 1997) premature senescence of root cultures in both plant species and slower growth capacity in Hyoscyamus can be caused by pmt overexpression in polyamine metabolism, because the substrate of PMT enzyme (putrescine) is shared with polyamine metabolism (Sato et al., 2001). On the other hand, alkaloid content was closely related to the presence of the 35S‐pmt transgene, showing that ectopic expression of tobacco pmt increased the biosynthetic flux towards the tropane alkaloids. Consequently, tropane alkaloid contents of transformed roots in both D. metel and H. muticus were enhanced up to 5‐fold. In H. muticus, the enhancement concerned only hyoscyamine, which may be due in this species to a limiting amount of H6H activity that would convert hyoscyamine to scopolamine (Jouhikainen et al., 1999). In H. muticus transgenic hairy roots, both PMT enzymatic activity and alkaloid content were measured. Although all lines transgenic for 35S‐pmt contained elevated levels of hyoscyamine, there was not always correlation between the amount of extractable PMT and the amount of alkaloids. The biosynthesis of tropane alkaloids in Hyoscyamus takes place in specific cells of the pericycle (Kanegae et al., 1994). In A. belladonna, it has been shown that regulation of the plant’s endogenous pmt is also pericycle specific (Suzuki et al., 1999). It is evident that the transcriptional control by the 35S promoter of the transgenic pmt gene in the hairy root lines analysed here is not cell type specific. The fact that the altered expression pattern of pmt rather than the amount of PMT enzyme increases tropane alkaloid production in H. muticus hairy roots, shows that tropane alkaloid biosynthesis is very complex and might be slightly different in various plant species. Furthermore, it may also indicate that the transgene allows bypassing of the endogenous control of metabolic flux to the alkaloids that would take place at the level of the first committed enzymatic step in their biosynthesis. Acknowledgements This work has been partially supported by a grant from the Boehringer Ingelheim Industry and the Spanish CICYT (PB‐1243). The work of KJ was supported by a grant from the Pharmacy Graduate School Program of the University of Helsinki. The authors are very grateful to Professor T Hashimoto of the Nara Institute of Science and Technology (Kyoto, Japan) for supplying the pTVPMT plasmid carrying the cDNA which codes for putrescine‐N‐methyltransferase from tobacco. Open in new tabDownload slide Fig. 1. A typical purple pigment formation in a Hyoscyamusmuticus hairy root clone induced by C58C1 (pRiA4, pBMI) 7 d after subculturing. Open in new tabDownload slide Fig. 2. Fresh and dry weight (g l–1) of transformed root culture clones at the end of the culture period (28 d). Control values represent the mean of three individual control lines. Measurements are based on 3–6 replicates. Standard deviations were 8–10% of the values represented. Open in new tabDownload slide Fig. 3. Tropane alkaloid production (mg l–1) of transformed root culture clones at the end of the culture period (28 d). Control values represent the mean of three individual control lines. Measurements are based on 3–6 replicates. Standard deviations were 3–10% of the values represented. Open in new tabDownload slide Fig. 4. Northern blot results corresponding to Datura clones carrying 35S‐pmt gene. Each line was charged with the same quantity of total RNA. Open in new tabDownload slide Fig. 5. Comparative study of alkaloid production (mg g–1 DW) and densitometrical analysis of Northern blot results corresponding to selected root clones of Datura and Hyoscyamus. Experiments were carried out three times. Open in new tabDownload slide Fig. 6. Comparison of the PMT activity and hyoscyamine content in transgenic (filled diamonds) and control (filled triangles) hairy root clones of Hyoscyamus muticus. References BlondiS, Scaramagli S, Oksman‐Caldentey K‐M, Poli F. 2002 . Secondary metabolism in root and callus cultures of Hyoscyamus multicus L.: the relationship between morphological organization and response to methyl jasmonate. Plant Science 163, 563 –569. BradfordMM. 1976 . A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding. Analytical Biochemistry 72, 248 –254. CelmaRC, Palazón J, Cusidó RM, Piñol MT, Keil M. 2001 . Decreased scopolamine yield in field‐grown Duboisia plants regenerated from hairy roots. Planta Medica 67, 249 –253. CusidóRM, Palazón J, Piñol MT, Bonfill M, Morales C. 1999 . Datura metel: in vitro production of tropane alkaloids. Planta Medica 65, 144 –148. DavidC, Chilton MD, Tempé J. 1984 . Conservation of T‐DNA in plants regenerated from hairy root cultures. Bio/Technology 2, 73 –76. EdwardsE, Johnstone C, Thompson A. 1991 . A simple and rapid method for the preparation of plant genomic DNA for PCR analysis. Nucleic Acids Research 19, 1349 . FethF, Arfmann H‐A, Wray V, Wagner KG. 1985 . Determination of putrescine N‐methyltransferase by high performance liquid chromatography. Phytochemistry 24, 921 –923. GamborgOL, Miller RA, Ojima K. 1986 . Nutrient requirements of suspension cultures of soybean root cells. Experimental Cell Research 50, 151 –158. HanahanD. 1983 . Studies on transformation of Escherichia coli with plasmids. Journal of Molecular Biology 166, 557 –562. HashimotoT, Yamada Y. 1983 . Scopolamine production in suspension cultures and rediffentiated roots of Hyoscyamus niger. Planta Medica 47, 121 –141. HibiN, Fujita T, Hatano M, Hashimoto T, Yamada Y. 1992 . Putrescine N‐methyltransferase in cultured roots of Hyoscyamus albus. Plant Physiology 100, 826 –835. HibiN, Higashiguchi S, Hashimoto T, Yamada Y. 1994 . Gene expression in tobacco low‐nicotine mutants. The Plant Cell 6, 723 –735. JouhikainenK, Lindgren L, Jokelainen T, Hiltunen R, Teeri TH, Oksman‐Caldentey K‐M. 1999 . Enhancement of scopolamine production in Hyoscyamus muticus L. hairy root cultures by genetic engineering. Planta 208, 545 –551. KadoCI, Liu ST. 1981 . Rapid procedure for detection and isolation of large and small plasmids. Journal of Bacteriology 145, 1365 –1373. KanegaeT, Kajiya H, Amano Y, Hashimoto T, Yamada Y. 1994 . Species‐dependent expression of the hyoscyamine 6β‐hydroxylase gene in the pericycle. Plant Physiology 105, 483 –490. KholodenkoBN, Cascante M, Hoek JB, Westerhoff HV, Schwaber J. 1998 . Desired metabolite concentrations and fluxes. Biotechnology and Bioengineering 59, 239 –247. ManiatisT, Friotsh EF, Sambrook J. 1982 . Molecular cloning, a laboratory manual. New York: Cold Spring Harbor Laboratory. Martin‐TanguyJ. 1997 . Conjugated polyamines and reproductive development: biochemical, molecular and physiological approaches. Physiologia Plantarum 100, 675 –688. MillerJH. 1972 . Experiments in molecular genetics. New York: Cold Spring Harbor Laboratory. MoyanoE, Fornalé S, Palazón J, Cusidó RM, Bonfill M, Morales C, Piñol MT. 1999 . Effect of Agrobacterium rhizogenes T‐DNA on alkaloid production in Solanaceae plants. Phytochemistry 52, 1287 –1292. MozoT, Hooykaas PJJ. 1991 . Electroporation of megaplasmids into Agrobacterium. Plant Molecular Biology 16, 917 –918. MuranakaT, Ohkawa H, Yamada Y. 1993 . Continuous production of scopolamine by a culture of Duboisia leichhardtii hairy root clone in a bioreactor system. Applied Microbiology and Biotechnology 40, 219 –223. Oksman‐CaldenteyK‐M, Arroo R. 2000 . Regulation of tropane alkaloid metabolism in plants and plant cell cultures. In: Verpoorte R, Alfermann AW, eds. Metabolic engineering of plant secondary metabolism. Dordrecht: Kluwer Academic Publishers, 253 –281. Oksman‐CaldenteyK‐M, Kivelä O, Hiltunen R. 1991 . Spontaneous shoot organogenesis and plant regeneration from hairy root cultures of Hyoscyamus muticus. Plant Science 78, 129 –136. Oksman‐CaldenteyK‐M, Strauss A. 1986 . Somaclonal variation of scopolamine content in protoplast–derived cell culture clones of Hyoscyamus muticus. Planta Medica 52, 6 –12. Oksman‐CaldenteyK‐M, Vuorela H, Strauss A, Hiltunen R. 1987 . Variation in the tropane alkaloid content of Hyoscyamus muticus plants and cell culture clones. Planta Medica 53, 349 –354. PalazónJ, Altabella T, Cusidó RM, Ribó M, Piñol MT. 1995 . Growth and tropane alkaloid production in Agrobacterium transformed roots and derived callus of Datura. Biologia Plantarum 37, 161 –168. PalazónJ, Cusidó RM, Gonzalo J, Bonfill M, Morales C, Piñol MT. 1998 . Relation between the amount of rolC gene product and indole alkaloid accumulation in Catharanthus roseus transformed root cultures. Journal of Plant Physiology 153, 712 –718. PiñolMT, Palazón J, Cusidó RM, Ribó M. 1999 . Influence of calcium ion‐concentration in the medium on tropane alkaloid accumulation in Datura stramonium hairy roots. Plant Science 141, 41 –49. PiñolMT, Palazón J, Cusidó RM, Serrano M. 1996 . Effects of Ri T‐DNA from Agrobacterium rhizogenes on growth and hyoscyamine production in Datura stramonium root cultures. Botanica Acta 109, 133 –138. RobinsRJ, Bent FG, Rhodes MJC. 1991 . Studies on the biosynthesis of tropane alkaloids by Datura stramonium L. transformed root cultures. 3. The relationship between morphological integrity and alkaloid biosynthesis. Planta 185, 390 –395. SatoF, Hashimoto T, Hachiya A, Tamura K, Choi K‐B, Morishige T, Fujimoto H, Yamada Y. 2001 . Metabolic engineering of plant alkaloid biosynthesis. Proceedings of the National Academy of Sciences, USA 98, 367 –372. SevónN, Oksman‐Caldentey K‐M, Hiltunen R. 1995 . Efficient plant regeneration from hairy root‐derived protoplasts of Hyoscyamus muticus. Plant Cell Reports 14, 738 –742. SevónN, Hiltunen R, Oksman‐Caldentey K‐M. 1998 . Somaclonal variation in transformed root cultures and protoplast‐derived hairy root cultures of Hyoscyamus muticus. Planta Medica 64, 37 –41. SuzukiK, Yamada Y, Hashimoto T. 1999 . Expression of Atropa belladonna putrescine N‐methyltransferase gene in root pericycle. Plant Cell Physiology 40, 289 –297. Van LarebekeN, Engler G, Holsters M, Van den Esacker S, Schilperoort RA, Schell J. 1974 . Large plasmid in Agrobacterium tumefaciens essential for crown gall‐inducing ability. Nature 252, 169 –170. VanhalaL, Eeva M, Lapinjoki S, Hiltunen R, Oksman‐Caldentey K‐M. 1998 . Effect of growth regulators on transformed root cultures of Hyoscyamus muticus. Journal of Plant Physiology 153, 475 –481. VanhalaL, Hiltunen R, Oksman‐Caldentey K‐M. 1995 . Virulence of different Agrobacterium strains on hairy root formation of Hyoscyamus muticus. Plant Cell Reports 14, 236 –240. WhiteF, Nester EW. 1980 . Hairy root: plasmid encodes virulence in Agrobacterium rhizogenes. Journal of Bacteriology 141, 1134 –1141. YamadaY, Endo T. 1984 . Tropane alkaloid production in cultured cells of Duboisia leichhardtii. Plant Cell Reports 3, 186 –188.

Laval, Valérie;Masclaux, Frédéric;Serin, Arnaud;Carrière, Marguerite;Roldan, Caroline;Devic, Martine;Pont‐Lezica, Rafael F.;Galaud, Jean‐Philippe

doi: 10.1093/jxb/erg018pmid: 12493849

Abstract The membrane receptor protein from pea, peabp80, has been shown to function by in vitro binding studies, and in vivo in yeast mutant, as a vacuolar sorting receptor (VSR). Families of proteins with homology to peabp80 have been identified in many plants including Arabidopsis. The family of membrane receptors, atbp80a–f (Arabidopsis thaliana binding protein 80 kDa) is highly homologous to peabp80 and may also function as vacuolar sorting receptors. Interactions with vacuolar sorting determinants have been shown only in vitro for atbp80b. In this paper, atbp80b was over‐ and under‐expressed in Arabidopsis. Transgenic plants that over‐expressed atbp80b showed no visible phenotype. However, antisense transformants were defective in germination. In non‐germinating antisense transformants the embryo appeared to be normal, but, using several methods, it was not possible to rescue the non‐germinating seeds, indicating that the mechanisms were probably independent of a seed‐coat‐imposed inhibition. To make a correlation between the lack of germination and gene expression, transcription analysis of all atbp80 genes was performed in the non‐germinating antisense seeds indicating that all the normally transcribed genes were not detected. Then, a gene expression study of atbp80s genes was carried‐out following seed imbibition and in various organs during wild‐type plant development showing that all the genes from the family were transcribed and differentially expressed. Key words: Antisense, Arabidopsis thaliana, embryogenesis, vacuolar sorting receptor, germination. Received 14 March 2002; Accepted 26 August 2002 Introduction A family of type I membrane receptors identified as putative vacuolar sorting receptors (VSRs) by Paris et al. (1997), or as ELPs for EGF‐ Like receptor Proteins by Ahmed et al. (1997), is found in plants. The Arabidopsis genome sequencing project has identified, on the basis of sequence homology, seven putative genes named atbp80a (previously named AtELP2b), atbp80a′ (AtELP2a), atbp80b (AtELP1), atbp80c (AtELP4), atbp80d (AtELP5), atbp80e (AtELP6), and atbp80f (AtELP3). The nomenclature is still confusing and that proposed by Hadlington and Denecke (2000) was chosen. EST sequencing programmes have only identified transcripts from atbp80b, atbp80a and atbp80f. Several laboratories independently identified the first member of this family from pea clathrin‐coated vesicles and different preparations from Arabidopsis (Paris et al, 1997; Ahmed et al., 1997; Neuhaus and Rogers, 1998; Laval et al., 1999). Bp80 genes encode integral membrane proteins with a large lumenal or extracellular domain containing three cys‐rich regions similar to epidermal growth factor (EGF)‐like repeats, and a short cytoplasmic domain. At the subcellular level, atbp80 has been found in clathrin‐coated vesicles in pea cotyledons and Arabidopsis cell cultures (Ahmed et al., 1997; Hinz et al., 1999; Kirsch et al., 1994). They have also been found in the trans Golgi network, and in a prevacuolar compartment in stigmas from Nicotiana alata (Ahmed et al., 1997; Hillmer et al., 2001; Miller et al., 1999; Paris et al., 1997). Finally, atbp80 was also localized on unidentified endomembrane fractions as well as a plasma membrane rich‐fraction of Arabidopsis cell cultures (Laval et al., 1999). Several publications point to a role for atbp80 as VSRs in plants, equivalent to animal mannose 6‐phosphate receptors (Kornfeld, 1992), or the VPS10p receptor from yeast (Marcusson et al., 1994). Peabp80 (also named BP80 or VSR‐PS‐1) binds as a monomer to polypeptide ligands containing a central NPIR motif at pH 6.0 to 6.5 and this binding was abolished at acidic pH leading to ligand release into the prevacuolar compartment (Kirsch et al., 1994; Cao et al., 2000). In vitro binding to this motif has been also demonstrated for atbp80b (Ahmed et al., 2000). Humair et al. (2001) have recently demonstrated by using an in vivo system, that peabp80 was able to function as a VSR in yeast mutant. Unlike other eukaryotes, many plants cells contain two distinct vacuolar compartments. One functions as a digestive organelle, similar to the lysosome in animal cells. It contains hydrolytic enzymes and accumulates secondary metabolites (Wink, 1993). The second type functions as a protein storage compartment, especially in seeds but also in vegetative tissues such as roots and tubers (Muntz, 1998; Staswick, 1994; Neuhaus and Rogers, 1998). It is now clear that these two types of vacuoles are separate organelles (Hoh et al., 1995; Paris et al., 1996). The bp80 proteins have been found in vesicles containing proteins targeted to the lytic vacuole, but not in the dense vesicles containing storage proteins (Hinz et al., 1999; Jiang and Rogers, 1999; Miller et al., 1999). However, there is one report that suggests that these proteins may function as sorting receptors for pumpkin storage proteins (Shimada et al., 1997). With the exception of peabp80, none of the other proteins have been shown to function as VSRs in vivo. In contrast to the vacuolar sorting receptors in animal cells or in yeast, atbp80s, the putative plant vacuolar sorting receptors in Arabidopsis, are members of a large gene family. Are these proteins expressed differentially during plant development, and do all members of this family share the same function? In this paper, a gain or loss of function strategy was used on whole plants to get new data on the biological role of this protein family inplanta. Transgenic Arabidopsis plants were constructed over‐ or under‐expressing atbp80 genes, using sense or antisense atbp80b constructs. Antisense transformants produced apparently normal seeds, but in some lines fewer than 10% of the seeds germinated. A gene expression analysis of all the atbp80 was performed in non‐germinating antisense seeds indicating that none of the atbp80 is detected. At present, only three atbp80 genes (atbp80b, a and f) are known to be transcribed referring to the EST sequencing programs. Atbp80 gene expression was analysed during the first days following imbibition in various organs of wild‐type plants. All the genes were expressed and showed different patterns of expression. Materials and methods Plant material Arabidopsis thaliana, ecotype Columbia (Col‐0), was cultured in a grown chamber with 16/8 h light/dark photoperiod, given by fluorescent tubes 36W (12 W m–2) at 20 °C temperature. Seeds were sterilized by 5 min incubation in 2.6% sodium hypochlorite and washed three times with 70% ethanol. Plants were grown in pots filled with TKS2 peat Floratorf supplemented with 1‰ (w/w) nitrate. Sense and antisense constructs For the antisense construct, the coding region of atbp80b (accession No. U 86700; Laval et al., 1999), from the ATG to the stop codon was introduced in antisense orientation in the expression vector pBin19pJR1, which contains the 35S CaMV promoter (Smith et al., 1988). Ten pmoles of primer 5′‐ATATGTCGACATGAAGCTTG GGCTTTTCAC‐3′ and 5′‐TAATGGTACCCCACTTATTTCTG TTTGTGGC‐3′, allowing the introduction of SalI and KpnI restriction sites, were used for PCR amplification. The amplified DNA was cloned in pGEM‐T vector (Promega) before sequencing (Sanger et al., 1977). The vector was digested and the fragment of interest was cloned into the predigested KpnI‐SalI pBin19pJR1 vector. A similar procedure was followed for the sense construct, using primers 5′‐ATATGTCGACGCACAGTTGAAGTGAACT TGC‐3′ and 5′‐ATTAGGTACCATGAAGCTTGGGCTTTTCAC‐3′ to introduce the coding region of atbp80b from the ATG to poly A tail in the forward orientation under the 35S‐promoter control in pBin19pJR1. Plant transformation The constructs were mobilized in Agrobacterium tumefaciens strain C58Ci (pMP90) (Koncz and Schell, 1986) by the heat shock method derived from (Holsters et al., 1978). Agrobacterium‐mediated transformation of Arabidopsis was performed by incubating standard floral tip. Sterilized seeds obtained from self‐fertilized primary transformant (T1) were germinated on MS medium supplemented with 50 µg ml–1 kanamycin, 10% sucrose and 7% agar. After 2 weeks, kanamycin T1‐resistant seedlings were individually transplanted into pot filled with TKS2, grown to maturity self‐fertilized and T2 seeds were harvested. Germination assays Sterilized T2 antisense (AST2) seeds were sown on MS culture medium, complemented or not with 10–4 M GA (Sigma), or with a mix of amino acids (Sigma). Seeds were incubated 2 d at 4 °C prior to the transfer to culture chamber at 20 °C with 16/8 h light/dark photoperiod. Germination was scored after 10 d of incubation. To determine if seed integuments inhibit germination, wild‐type and T2 antisense integuments were mechanically removed. Embryos were placed on MS culture medium recovered by 50 µl of MS liquid and observed every day for 3 weeks. Analysis of embryo cell viability To estimate embryo viability, integuments from wild‐type and T2 antisense seeds were removed and embryos were incubated in solution containing 1% FDA (fluorescein diacetate) for 30 min, washed in distilled water and observed under fluorescent microscopy at 470–490 nm excitation. Histological analysis Seeds or plantlets were fixed with 3% glutaraldehyde in sodium cacodylate 0.1 M pH 7.2 for 2 h, post‐fixed with osmium tetroxyde (1%) for 1 h, then dehydrated by serial incubation in ethanol for 15 min increasing the alcohol concentration and embedded in spurr resins for 1 week (Spurr, 1969). Resin polymerization was carried out for 48 h at 60 °C. Semi‐thin sections were cut with an ultracut E microtome and stained with toluidine blue 0.5% in water before microscopic observation. DNA analysis PCR assays on genomic DNA were performed to verify the presence of the transgene in the different plants or seeds. The presence of the NptII gene was confirmed by PCR by using primers 5′‐GA GGCTATTCGGCTATGACTG‐3′ and 5′‐ATCGGGAGCGGCGA TACCGTA‐3′. Genomic DNA was prepared according to Dellaporta et al. (1983). Fifty ng of genomic DNA were used for PCR amplification with 50 pmol of the corresponding primers. PCR amplification was carried for 40 cycles of 1 min denaturation at 95 °C, 1 min annealing at 50 °C and 1 min extension at 72 °C. The amplification products were loaded on 2% agarose gel, separated by electrophoresis, and stained with ethidium bromide. RT‐PCR analysis Total RNA was extracted from various organs at different stages. First strand cDNA synthesis was performed using oligo‐dT and specific oligonucleotides used for the PCR reactions were complementary to two separated exons, allowing the products from genomic DNA and cDNA to be differentiated. The primers used for atbp80b (at3g52850) were 5′‐TTGGTGACGGTTAC ACTCAC‐3′ and 5′‐TCCATATGGTGACCACTTGTGTTGG‐3′; primers 5′‐AAGTGAGATCAGCATGGGC‐3′ and 5′‐AGGCGA GAAAACATCGTCGTT‐3′ for atbp80a’ (at2g14740); primers 5′‐AAGTGAAATCAGCGTGGGC‐3′ and 5′‐CAGTTCTCCGTAG CTACATCG‐3′ for atbp80a (at2g14720); primers 5′‐ACGGCTTA ACTTTCTCTGCTTGC‐3′ and 5′‐GATCTCAAACATTTCTTT GTGTATG‐3′ for atbp80f (at4g20110); primers 5′‐GAGTGA AGCGAGTGTAGCCTCC‐3′ and 5′‐CGGTGGGGTGGAGTTC ATGCACAGGGAGG‐3′ for atbp80c (at2g30290); primers 5′‐GGT GATCCTGATGCTGATGTAGAGAATG‐3′ and 5′‐GAGGGCA ACGACATCCTGATGTCTCTGAG‐3′ for atbp80d (at2g34940); primers 5′‐GTGCAGAGAGAGTTGTTGAATCTCTAG‐3′ and 5′‐GAGATCACAGTCATTGGATCGGTTGTG‐3′ for atbp80e (at1g30900). Primers 5′‐GTCCAGTGTCTGTGATATTGCACC‐3′ and 5′‐GCTTACGAATCCGAGGGTGCC‐3′ for β‐tubulin (M21415) were used as a control. PCR amplification was carried for 35 cycles of 1 min denaturation at 95 °C, 1 min annealing at 50 °C and 1 min extension at 72 °C. Immunodetection Antibodies were raised, as previously described by Laval et al. (1999), against selected peptides derived from atbp80b, peptide 63 (AEQESQIGKGSRGDC) and peptide 64 (NNRQYRGKLEC) which are conserved in most of the members of the family. A mixture of those two antibodies was used to detect atbp80 in wild‐type, antisense T2 lines, and sense T2 lines. To extract proteins, tissues from 4‐week‐old plants were ground in liquid nitrogen and extracted with 62.5 mM Tris‐HCl, pH 6.8, and 2% SDS at 65 °C for 30 min. Extracts were centrifuged for 30 min at 10 000 g and the protein concentration was determined (Smith et al., 1985). Gel electrophoresis and antibody binding competition assays were carried‐out as previously reported (Laval et al., 1999). Results Analysis of atbp80b T2 sense and antisense transformants Transgenic plants with constructs designed to over‐ or under‐express atbp80b, were obtained. The nucleotide sequence of atbp80b is 61–84% identical to the other members of the family in Arabidopsis. The complete coding region of atbp80b (1872 bp) was used for the antisense construct to inhibit the expression of multiple members of the gene family (Fig. 1A). For the sense transformants, the coding region of atbp80b including the 3′‐UTR (which is different for each gene in the family) was used (Fig. 1B). Around 500 seeds for each transformant were tested for germination. Wild type seeds show between 80% and 100% germination as did five sense transformants (ST2 lines C, D, E, F, and G) tested (Fig. 1C). All of the ST2 transformants were similar to wild type in appearance. Most of the eleven AST2 seed lines showed significantly reduced germination (Fig. 1C, AST2, lines M, D, O, G, N, F, L, K, I, E, J), ranging from 4% for line J to 100% for line M. It is assumed that the defect of germination is not the consequence of a T‐DNA disruption of a functional gene because most of the antisense transformants have the same phenotype. Line to line variability in germination is probably due to different levels of expression, as expected for antisense transformants. Transformants AST2E, J, I and K showed the strongest inhibition of germination. In these four lines, less than 25% of the seeds germinated and developed as in the wild type. PCR on antisense lines using primers to the NPTII and atbp80b constructs confirmed that all contained a complete construct (data not shown). Also, southern blots (data not shown) revealed that AST2E, I and K developing plants contain at least two insertions and AST2J, four insertions of the T‐DNA. Analysis of AST2 non‐germinating seeds. Four antisense lines showing the lowest level of germination: line K (25% germination), lines I and E (10% germination), and J (4% germination) were analysed further. In these lines, germinating seeds were distinguished from non‐germinating, sowing the T2 on MS medium for 10 d. The germinated ones were isolated and T3 seeds were collected from these individuals. Eighty to 100% of these T3 seeds germinated. Possibly, in those seeds that germinated, the antisense construct has been heritably silenced at meiosis by epigenetic mechanisms (Kooter et al., 1999). To establish a correlation between the absence of germination and atbp80 gene expression level, a transcriptional analysis by RT‐PCR was carried‐out on non‐germinating AST2K and wild type dry seeds, imbibed wild‐type seeds, and on wild‐type plants at the cotyledon stage (Fig. 2). Results show that only the atbp80b, atbp80a′ and atbp80a genes were transcribed in mature dry seed and following imbibition. In AST2K non‐germinating seeds, none of these genes were expressed, indicating that the antisense strategy was efficient to inhibit the expression of several members of the atbp80 family. From this analysis, it was showed that atbp80d and atbp80e mRNAs were detected only when the plants reached the cotyledon stage. At present, these genes were not found in EST databases and it can be concluded that these genes are not pseudogenes. However, for atbp80c, no mRNA was detected in the conditions tested. Failure to germinate may be caused by various factors, seed viability, dormancy, or seed‐coat‐imposed inhibition (Dubreucq et al., 1996). To check if the sterilization procedure can influence the germination rate of non‐germinating antisense seeds, these were sown directly in soil, but the germination rate was unmodified. To evaluate seed viability, embryos were isolated from seeds that failed to germinate after 8 d, as well as from non‐transformed seeds, and incubated with 1% FDA. Wild‐type embryos boiled for 15 min did not fluoresce, but untreated wild‐type, AST2E, and AST2I embryos fluoresced strongly (Fig. 3) indicating that the embryo cells are living. Histological analysis of wild‐type, AST2E, and AST2I embryos from non‐germinating seeds confirmed that the shape of the embryos, and the cellular organization was similar between types (data not shown). To break putative dormancy, chilling or gibberellin treatment (10–4 M) was carried out; however, neither treatment improved germination. To distinguish between seed‐coat‐imposed inhibition and embryo‐blocked germination, naked embryos from wild‐type, AST2E, I, and J seeds were sown in MS medium. Under such conditions, less than 10% of the AST2 transformants germinated compared to 80% of wild‐type embryos. Also, both control and antisense seeds show a normal uptake of water in the first 24 h of imbibition confirming that the seed coat was not impermeable. Atbp80 has been implicated in the transport of vacuolar peptidases (Ahmed et al., 2000; Paris et al., 1997), and antisense transformants defective in the VSR might be unable to mobilize reserve proteins for germination. The medium was supplemented with a mixture of the 20 amino acids to support protein synthesis during germination, but no change in the germination pattern was observed. AST2 seedlings show an arrest in development. As stated before, about 1% of AST2 lines E, and J germinated, but an abnormal development of the seedling was observed. Seedlings showed a long root, a thick but short hypocotyl, and no photosynthetic tissues. Cotyledons never developed and became green. They were frequently enclosed in the integument (Fig. 4A). Histological sections (Fig. 4B) of 2‐month‐old AST2E transformants were compared to 5‐d‐old wild‐type plants. At the root level, vascular tissue contained bigger cells than the wild type (Fig. 4B1, 2), and the cortex and epidermal cells were collapsed and without structure. Hypocotyl sections (Fig. 4B3, 4) showed a disorganized structure in the AST2 seedlings. The epidermal layer was detached from the cortex cells. Vascular tissues were not clearly organized and many cells looked collapsed (cf. Fig. 4B5 and 4B6). Atbp80 protein content in the various different transgenic lines was detected using antibodies raised against two different synthetic peptides derived from atbp80b (Laval et al., 1999). Extracts from 1‐month‐old plants transformed with the sense construct showed a strong band of the expected size of 80 kDa (Fig. 5). However, the antibodies can potentially recognize any of seven atbp80 gene products, since the peptide sequence used for raising the antibodies is highly conserved in all atbp80 proteins. Some weaker upper and lower bands were detected. These could correspond to modified forms of atbp80 glycosylated product, degradation products and/or cross‐reacting unrelated proteins. As in the wild‐type plants, atbp80 protein was also detectable serologically in 5–10% T2 plants of antisense lines AST E, I, and J that germinate. These data suggest that the level of atbp80s is sufficient for germination and agreed with a probable silencing of antisense constructs. Atbp80 gene expression in various organs during plant development At present, only atbp80b, atbp80a, and atbp80f are found in EST databases. Are all atbp80 transcribed, or some of them are pseudogenes? An expression study of the all atbp80 genes was performed on roots, young and old leaves, floral stalk, flowers, immature siliques (days 4–9 after pollination) or pre‐mature siliques (days 10–17 after pollination). The beta tubulin gene was used as a control (Fig. 6). Results indicate that all the atbp80 are transcribed either in all the organs analysed (atbp80a′, atbp80a and atbp80e) or in a particular organ such as atbp80c, which is only expressed in flowers. Except for atbp80a′ and atbp80a genes, which are expressed in all the conditions and in all the organs analysed (Figs 2, 6), the other genes present a specific expression profile. Atbp80b gene was transcribed in all the conditions except in pre‐mature siliques. Atbp80f mRNAs were detected in roots, floral stalk, and young organs (leaves and siliques), but not in old leaves or in old‐siliques or in the flowers. A relationship can be established between atbp80f gene expression and the age of the tissue. Atbp80d mRNAs were detected in every tissue except in pre‐mature siliques. Finally, the atbp80e gene was expressed in each tissue except in dry seeds and after imbibition as reported in Fig. 2. Discussion Arabidopsis was transformed with sense and antisense atbp80b. All the sense lines germinated, grew, and reproduced normally. The sense transformants indicated an increase in atbp80 proteins as revealed by atbp80 antibodies, suggesting that over‐expression of the gene and accumulation of atbp80b protein does not change the phenotype in the growing conditions described here. Eleven independent antisense transformants showed variable degrees of germination from 4–100%. From each line seeds could be divided into three categories; (1) those that germinated, grew, and reproduced normally; (2) non‐germinating seeds; and (3) seeds that germinated but then failed to develop further. In antisense plants derived from seeds that germinated normally (group 1), atbp80 proteins were detectable serologically. Thus, it can be assumed that the non‐germinating phenotype (group 2) may be linked to the lack of atbp80s in those antisense transformants as indicated by the atbp80 gene expression analysis. Among the seven members of atbp80 genes, only the atbp80b, atbp80a′, and atbp80a transcripts are present in dry seed and following imbibition. Even if atbp80 proteins are well conserved at the amino acid level, the expression data reported here clearly suggest that atbp80 proteins may participate in specific developmental steps. In non‐germinating antisense seeds, none of these genes were detected. This indicates that the complete coding region of atbp80b, used for the antisense construct, was able to inhibit the expression of other members of the family. The embryo appeared to be normal in group 2 antisense transformants, and it was not possible to rescue the non‐germinating seeds, indicating that the defect in germination was independent of a seed‐coat‐imposed inhibition. One per cent of the AST2 germinating‐seeds from lines E and J did not develop and the seedlings remained at the cotyledon stage (group 3). The level of gene expression or atbp80 protein content in those plants cannot be evaluated. Histological studies revealed alterations in the structure of cell walls, mainly in the root and hypocotyl cortex. In the hypocotyl, the epidermis detached from cortical cells suggesting weak cell adhesion in these lines. The mechanism by which the absence or low levels of atbp80s may alter cell wall structure is unclear, but two hypotheses can be stated. First, if atbp80 is missing, some vacuolar hydrolases may take the default pathway to the extracellular space and degrade cell wall components. Second, some members of the atbp80 family may be located either in the plasma membrane or associated vesicles (Laval et al., 1999) and therefore could participate in the deposition of cell wall material or trafficking between the ER and the plasma membrane. Recent results published by Brandizzi et al. (2002) show the importance of the length of the hydrophobic domain on subcellular location of peabp80. By using the same predictive method (TMHMM version 2.0 program), all the transmembrane domains of atbp80 proteins are 23 amino acids long and with reference to Brandizzi et al. (2002), these proteins may potentially be located on the plasma membrane. It can reasonably be proposed that suppression of atbp80 expression is responsible for the germination defect and abnormal development. These results are the first demonstration of phenotypes with a reduced level of atbp80inplanta. Bp80 proteins are believed to function as VSRs. They target to the lytic vacuole soluble proteins containing the sorting signal (NPIR) within the NH2‐terminal propeptide (Ahmed et al., 2000; Cao et al., 2000). The cargo proteins transported by the receptors can be proteases such as aleurain in barley and Arabidopsis. In aleurone cells of barley, the degradation of the storage proteins in the protein storage vacuoles depends on the synthesis of cysteine proteases. It was not possible to verify the presence of the protease AtALEU (a protein highly homologous to barley aleurain) in either wild‐type or antisense seeds, using an antibody raised against the barley protein (kindly supplied by J Rogers, Washington State University). Humair et al. (2001) showed that the peabp80 protein functions as a VSR in vivo using the vps10p yeast mutant. Peabp80 was able to target proteins sharing the motif NPIR and not the typical C‐terminal vacuolar determinant signal. The non‐germinating antisense phenotype described in this paper seems to confirm that function a priori, since the absence of specific proteases in the seed will produce an embryo unable to mobilize reserve proteins. A low level of atbp80 may be responsible for a failure to target cysteine proteases to the lytic vacuole. However, the cotyledons developmental arrest observed in the group 3 antisense transformants suggests that germination itself is independent of mobilization of cotyledon reserves. It has been suggested that all the members of the atbp80 family may share the same function. This is based on the presence of a much‐conserved region on the C‐terminal cytoplasmic tail, conserved not only within the Arabidopsis proteins, but also in orthologous proteins from monocots and dicots (Cao et al., 2000; Hadlington and Denecke, 2000). However, if the conserved cytoplasmic sequence is certainly involved in the recruitment of coat proteins for clathrin coated vesicle formation and, such as in yeast, for the recycling of the receptor, only the transmembrane domain seems to be necessary for aleurain targeting to the vacuole (Jiang and Rogers, 1998). It will be important to determine if other bp80s are also capable of functioning in the vps10p yeast mutant targeting to the vacuole NPIR‐containing or other vacuolar proteins. This will support the putative VSR function. Atbp80 proteins are members of a multigenic family and these results on expression of atbp80s during plant development in Arabidopsis indicate that all the genes are expressed differently and this implies different functions during development. Atbp80s may have a role in early germination, in the deposition of cell wall material or trafficking between the ER and the plasma membrane and in mobilization of storage proteins. Acknowledgements The University Paul Sabatier and the CNRS supported this work. VL and FM are fellows from the Ministère de l’Education Nationale, de la Recherche et de la Technologie, France. We are grateful to A Jauneau (IFR40, Toulouse, France) for the microscopic analysis and to Dr JJ Milner for valuable advice and for reading this manuscript. View largeDownload slide Fig. 1. (A) Scheme of the atbp80b antisense construction. The open reading frame of atbp80b (U86700) was cloned in antisense orientation under CaMV 35S promoter (P35S) between the left border (LB) and the right border (RB) of the T‐DNA. NptII gene, under the control of the NOS promoter (PNOS), confers kanamycin resistance. T: NOS terminator. (B) Scheme of the atbp80b sense construction. The open reading frame of atbp80b (from ATG to the 3′UTR) was cloned in sense orientation under CaMV 35S promoter as previously described. (C) Germination assays. Sterilized seeds of WT (open column), atbp80b antisense lines (AST2, grey columns), and atbp80b sense transformants (ST2, filled columns) were sown in kanamycin‐containing MS medium, and germination was scored after 10 d. Each letter indicates an independent transformant. View largeDownload slide Fig. 1. (A) Scheme of the atbp80b antisense construction. The open reading frame of atbp80b (U86700) was cloned in antisense orientation under CaMV 35S promoter (P35S) between the left border (LB) and the right border (RB) of the T‐DNA. NptII gene, under the control of the NOS promoter (PNOS), confers kanamycin resistance. T: NOS terminator. (B) Scheme of the atbp80b sense construction. The open reading frame of atbp80b (from ATG to the 3′UTR) was cloned in sense orientation under CaMV 35S promoter as previously described. (C) Germination assays. Sterilized seeds of WT (open column), atbp80b antisense lines (AST2, grey columns), and atbp80b sense transformants (ST2, filled columns) were sown in kanamycin‐containing MS medium, and germination was scored after 10 d. Each letter indicates an independent transformant. View largeDownload slide Fig. 2. Atbp80s RT‐PCR gene expression. RNA was extracted from non‐germinating AST2K antisense seeds and compared to RNA from wild‐type dry seed, 24 h‐imbibed seeds, or from seedlings at cotyledon stage. First strand cDNA synthesis was carried‐out with oligo‐dT and specific primers were used to amplify atbp80b, atbp80a′, atbp80a, atbp80f, atbp80c, atbp80d, and atbp80e cDNA. Beta tubulin was used as control. View largeDownload slide Fig. 2. Atbp80s RT‐PCR gene expression. RNA was extracted from non‐germinating AST2K antisense seeds and compared to RNA from wild‐type dry seed, 24 h‐imbibed seeds, or from seedlings at cotyledon stage. First strand cDNA synthesis was carried‐out with oligo‐dT and specific primers were used to amplify atbp80b, atbp80a′, atbp80a, atbp80f, atbp80c, atbp80d, and atbp80e cDNA. Beta tubulin was used as control. View largeDownload slide Fig. 3. Analysis of embryo cell viability. Seed integuments were removed and embryos were immediately incubated in a 1% FDA solution before observation under fluorescence microscopy at 470–490 nm excitation. (A) WT seeds heated at 100 °C for 15 min.; (B) WT seeds; (C) AST2E non‐germinating seeds; (D) AST2I non‐germinating seeds. View largeDownload slide Fig. 3. Analysis of embryo cell viability. Seed integuments were removed and embryos were immediately incubated in a 1% FDA solution before observation under fluorescence microscopy at 470–490 nm excitation. (A) WT seeds heated at 100 °C for 15 min.; (B) WT seeds; (C) AST2E non‐germinating seeds; (D) AST2I non‐germinating seeds. View largeDownload slide Fig. 4. AST2E and J transformant plants showing an inhibition of development. (A) Two‐months‐old AST2E (1) and AST2J (2) plantlets. (B) Histological comparison of AST2E to WT. Transversal sections were made as described in Materials and methods, from root (1, 2), hypocotyl (3, 4) and cotyledon (5, 6). View largeDownload slide Fig. 4. AST2E and J transformant plants showing an inhibition of development. (A) Two‐months‐old AST2E (1) and AST2J (2) plantlets. (B) Histological comparison of AST2E to WT. Transversal sections were made as described in Materials and methods, from root (1, 2), hypocotyl (3, 4) and cotyledon (5, 6). View largeDownload slide Fig. 5. Immunodetection of atbp80 proteins extracted from wild type (WT), ST2 lines E and F plants and AST2 lines E, I, and J. Thirty µg of total proteins were separated on 11% SDS‐PAGE gel, blotted onto nitrocellulose and incubated with specific purified antibodies raised against synthetic peptides derived from atbp80b sequence. View largeDownload slide Fig. 5. Immunodetection of atbp80 proteins extracted from wild type (WT), ST2 lines E and F plants and AST2 lines E, I, and J. Thirty µg of total proteins were separated on 11% SDS‐PAGE gel, blotted onto nitrocellulose and incubated with specific purified antibodies raised against synthetic peptides derived from atbp80b sequence. View largeDownload slide Fig. 6. Atbp80s RT‐PCR gene expression analysis in various organs during plant development. RNA was extracted from wild‐type roots, young or old leaves, floral stalk, flowers, immature (days 4–9 after pollination) and pre‐mature siliques (days 10–17 after pollination). First strand cDNA synthesis was carried out with oligo‐dT and specific primers were used to amplify atbp80b, atbp80a′, atbp80a, atbp80f, atbp80c, atbp80d, and atbp80e cDNA. Beta tubulin was used as control. View largeDownload slide Fig. 6. Atbp80s RT‐PCR gene expression analysis in various organs during plant development. RNA was extracted from wild‐type roots, young or old leaves, floral stalk, flowers, immature (days 4–9 after pollination) and pre‐mature siliques (days 10–17 after pollination). First strand cDNA synthesis was carried out with oligo‐dT and specific primers were used to amplify atbp80b, atbp80a′, atbp80a, atbp80f, atbp80c, atbp80d, and atbp80e cDNA. Beta tubulin was used as control. References AhmedSU, Bar‐Peled M, Raikhel NV. 1997. Cloning and subcellular location of an Arabidopsis receptor‐like protein that shares common features with protein‐sorting receptors of eukaryotic cells. Plant Physiology  114, 325–336. Google Scholar AhmedSU, Rojo E, Kovaleva V, Venkataraman S, Dombrowski JE, Matsuoka K, Raikhel NV. 2000. The plant vacuolar sorting receptor AtELP is involved in transport of NH2‐terminal propeptide‐containing vacuolar proteins in Arabidopsis thaliana. Journal of Cell Biology   149, 1335–1344. Google Scholar BrandizziF, Frangne N, Marc‐Martin S, Hawes C, Neuhaus J‐M, Paris N. 2002. The destination for single‐pass membrane proteins is influenced markedly by the length of the hydrophobic domain. The Plant Cell  14, 1077–1092. Google Scholar CaoX, Rogers SW, Butler J, Beevers L, Rogers JC. 2000. Structural requirements for ligand binding by a probable plant vacuolar sorting receptor. The Plant Cell  12, 493–506. Google Scholar DellaportaSL, Wood J, Hicks JB. 1983. A plant DNA minipreparation: version II. Plant Molecular Biology  Reporter   1, 19–21. Google Scholar DubreucqB, Grappin P, Caboche M. 1996. A new method for the identification and isolation of genes essential for Arabidopsis thaliana seed germination. Molecular and General Genetics  252, 42–50. Google Scholar HadlingtonJL, Denecke J. 2000. Sorting of soluble proteins in the secretory pathway of plants. Current Opinion in Plant Biology  3, 461–468. Google Scholar HillmerS, Movafeghi A, Robinson DG, Hinz G. 2001. Vacuolar storage proteins are sorted in the cis‐cisternae of the pea cotyledon Golgi apparatus. Journal of Cell Biology  152, 41–50. Google Scholar HinzG, Hillmer S, Baumer M, Hohl I. 1999. Vacuolar storage proteins and the putative vacuolar sorting receptor BP‐80 exit the Golgi apparatus of developing pea cotyledons in different transport vesicles. The Plant Cell  11, 1509–1524. Google Scholar HohB, Hinz G, Jeong BK, Robinson DG. 1995. Protein storage vacuoles form de novo during pea cotyledon development. Journal of Cell Science  108, 299–310. Google Scholar HolstersM, de Waele D, Depicker A, Messens E, van Montagu M, Schell J. 1978. Transfection and transformation of Agrobacterium tumefaciens. Molecular and General Genetics  163, 181–187. Google Scholar HumairD, Felipe DH, Neuhaus JM, Paris N. 2001. Demonstration in yeast of the function of BP‐80, a putative plant vacuolar sorting receptor. The Plant Cell  13, 781–792. Google Scholar JiangL, Rogers JC. 1998. Integral membrane protein sorting to vacuoles in plant cells: evidence for two pathways. Journal of Cell Biology  143, 1183–1199. Google Scholar JiangL, Rogers JC. 1999. Functional analysis of a Golgi‐localized Kex2p‐like protease in tobacco suspension culture cells. The Plant Journal  18, 23–32. Google Scholar KirschT, Paris N, Butler JM, Beevers L, Rogers JC. 1994. Purification and initial characterization of a potential plant vacuolar targeting receptor. Proceedings of the National Academy of Sciences, USA  91, 3403–3407. Google Scholar KonczC, Schell J. 1986. The promoter of TL‐DNA gene5 controls the tissue‐specific expression of chimaeric genes carried by a novel type of Agrobacterium binary vector. Molecular and General Genetics  204, 383–396. Google Scholar KooterJM, Matzke MA, Meyer P. 1999. Listening to the silent genes: transgene silencing, gene regulation, and pathogen control. Trends in Plant Science  4, 340–347. Google Scholar KornfeldS. 1992. Structure and function of the mannose 6‐phosphate/insulin like growth factor II receptors. Annual Review of Biochemistry  61, 307–330. Google Scholar LavalV, Chabannes M, Carriere M, Canut H, Barre A, Rouge P, Pont‐Lezica R, Galaud JP. 1999. A family of Arabidopsis plasma membrane receptors presenting animal beta‐integrin domains. Biochimica et Biophysica Acta  1435, 61–70. Google Scholar MarcussonEG, Horazdovsky BF, Cereghino JL, Gharakhanian E, Emr SD. 1994. The sorting receptor for yeast vacuolar carboxypeptidase Y is encoded by the VPS10 gene. Cell  77, 579–586. Google Scholar MillerEA, Lee MC, Anderson MA. 1999. Identification and characterization of a prevacuolar compartment in stigmas of Nicotiana alata. The Plant Cell  11, 1499–1508. Google Scholar MuntzK. 1998. Deposition of storage proteins. Plant Molecular Biology  38, 77–99. Google Scholar NeuhausJM, Rogers JC. 1998. Sorting of proteins to vacuoles in plant cells. Plant Molecular Biology  38, 127–144. Google Scholar ParisN, Rogers SW, Jiang L, Kirsch T, Beevers L, Phillips TE, Rogers JC. 1997. Molecular cloning and further characterization of a probable plant vacuolar sorting receptor. Plant Physiology  115, 29–39. Google Scholar ParisN, Stanley CM, Jones RL, Rogers JC. 1996. Plant cells contain two functionally distinct vacuolar compartments. Cell  85, 563–572. Google Scholar SangerF, Nicklen S, Coulson AR. 1977. DNA sequencing with chain‐terminating inhibitors. Proceedings of the National Academy of Sciences, USA  74, 5463–5467. Google Scholar ShimadaT, Kuroyanagi M, Nishimura M, Hara‐Nishimura I. 1997. A pumpkin 72 kDa membrane protein of precursor‐accumulating vesicles has characteristics of a vacuolar sorting receptor. Plant and Cell Physiology  38, 1414–1420. Google Scholar SmithPK, Krohn RI, Hermanson GT, Mallia AK, Gartner FH, Provenzano MD, Fujimoto EK, Goeke NM, Olson BJ, Klenk DC. 1985. Measurement of protein using bicinchoninic acid. Analytical Biochemistry  150, 76–85. Google Scholar SmithXJS, Watson CF, Ray J, Bird CR, Morris XC, Schuch W, Grierson D. 1988. Antisense RNA inhibition of polygalacturonase gene expression in transgenic tomatoes. Nature  25, 724–726. Google Scholar SpurrAR. 1969. A low viscosity epoxy resin embedding medium for electron microscopy. Journal of Ultrastructural Research   26, 31–43. Google Scholar StaswickPE. 1994. Storage proteins of vegetative plant tissues. Annual Review of Plant Physiology and Plant Molecular Biology  45, 303–322. Google Scholar WinkM. 1993. The plant vacuole: a multifunctional compartment. Journal of Experimental Botany  44, 231–246. Google Scholar

Lunn, John E.; Gillespie, Vanessa J.; Furbank, Robert T.

doi: 10.1093/jxb/erg023pmid: 12493850

Abstract Sucrose‐phosphate synthase (SPS) from the cyanobacterium Synechocystis sp. PCC 6803 lacks all of the Ser residues known to be involved in the regulation of higher plant SPS by protein phosphorylation. The Synechocystis SPS is also not allosterically regulated by glucose 6‐phosphate or orthophosphate. To investigate the effects of expressing a potentially unregulated SPS in plants, the Synechocystis sps gene was introduced into tobacco, rice and tomato under the control of constitutive promoters. The Synechocystis SPS protein was expressed at a high level in the plants, which should have been sufficient to increase overall SPS activity 2–8‐fold in the leaves. However, SPS activities and carbon partitioning in leaves from transgenic and wild‐type plants were not significantly different. The maximal light‐saturated rates of photosynthesis in leaves from tomato plants expressing the Synechocystis SPS were the same as those from wild‐type plants. Tomato plants expressing the maize SPS showed 2–3‐fold increases in SPS activity, increased partitioning of photoassimilate to sucrose and up to 58% higher maximal rates of photosynthesis. To investigate the apparent inactivity of the Synechocystis SPS the enzyme was purified from transgenic tobacco and rice plants. Surprisingly, the purified enzyme was found to have full catalytic activity. It is proposed that some other protein in plant cells binds to the Synechocystis SPS resulting in inhibition of the enzyme. Key words: Carbon partitioning, Lycopersicon esculentum Mill., Oryza sativa L., Nicotiana tabacum L., photosynthesis, rice, sucrose‐phosphate synthase, Synechocystis sp. PCC 6803, tobacco, tomato. Received 1 July 2002; Accepted 27 August 2002 Introduction Sucrose‐phosphate synthase (SPS, EC 2.4.1.14) catalyses the penultimate reaction in the pathway of sucrose biosynthesis, in which sucrose‐6F‐phosphate (Suc6P) is synthesized from UDPglucose (UDPGlc) and fructose 6‐phosphate (Fru6P). The Suc6P is then irreversibly hydrolysed by sucrose‐phosphatase (SPP, EC 3.1.3.24) to give sucrose (Lunn and ap Rees, 1990a). Studies of SPS‐antisense or co‐sense suppressed transgenic potato (Solanum tuberosum L.), rice (Oryza sativa L.) and Arabidopsis thaliana (L.) Heynh. plants indicate that SPS makes a major contribution to the control of flux through the pathway of sucrose biosynthesis in both photosynthetic and non‐photosynthetic tissues (Geigenberger et al., 1999; Geigenberger and Stitt, 2000; Ono et al., 1999a; Strand et al., 2000). SPS is a highly regulated enzyme in plants. In addition to allosteric regulation by glucose 6‐phosphate (Glc6P) and orthophosphate (Pi), the spinach (Spinacia oleracea L.) leaf SPS is activated in the light and deactivated in the dark by dephosphorylation and rephosphorylation, respectively, of Ser158 (Siegl et al., 1990; Huber and Huber, 1996). In maize (Zea mays L.) leaves, Ser162 has been identified as the phosphorylation site involved in light–dark regulation of the enzyme (Huber and Huber, 1996). There appear to be differences between species in the details of light–dark regulation of SPS by covalent modification (Lunn et al., 1997; Lunn and Furbank, 1999), but equivalent Ser residues are present in almost all known SPS sequences from higher plants (Lunn et al., 1999; Langenkämper et al., 2002). This suggests that there is at least the potential for phosphorylation and regulation of SPS at this site in most plants. There are at least two other known regulatory phosphorylation sites in spinach SPS; Ser424 is involved in osmotic regulation and Ser 229 in binding of 14‐3‐3 proteins (Toroser and Huber, 1997; Toroser et al., 1998). Expression of the maize leaf SPS in transgenic tomato (Lycopersicon esculentum Mill.) plants, under the control of the tobacco RbcS promoter, increased SPS activity in the leaves and shifted photoassimilate partitioning away from starch towards sucrose (Worrell et al., 1991). The maize SPS expressed in the transgenic tomato plants was not light–dark regulated, presumably because the tomato SPS‐protein kinases were unable to phosphorylate Ser162 in the maize enzyme (Worrell et al., 1991). Some lines of these, and other tomato plants expressing the maize leaf SPS under the control of the cauliflower mosaic virus (CaMV) 35S promoter, showed increased light and/or CO2 saturated rates of photosynthesis, higher relative growth rates and higher fruit yields (Galtier et al., 1993, 1995; Laporte et al., 1997, 2001; Micallef et al., 1995; Murchie et al., 1999; Nguyen‐Quoc et al., 1999). Over‐expression of SPS in A. thaliana also altered carbon partitioning (Signora et al., 1998) and has been reported to improve photosynthetic performance at low temperature, and to increase freezing tolerance (Strand et al., 2001). Transgenic cotton (Gossypium hirsutum L.) plants with increased SPS activities in fibre cells had improved fibre quality and higher fibre yields than untransformed plants (Haigler et al., 2000), and hybrid poplar (Populus tremula L.×P. tremuloides Michx.) trees with increased SPS activity showed increased photosynthesis and higher growth rates (Mouillon and Hurry, 2001). Attempts to over‐express SPS in other species have been less successful. When the spinach leaf SPS was expressed in tobacco (Nicotiana tabacum L.) plants the enzyme was found to be inactivated by protein phosphorylation (Frommer and Sonnewald, 1995; Toroser et al., 1999), as was the maize SPS expressed in transgenic rice plants (Ono et al., 1999b). Site‐directed mutagenesis of the spinach SPS, substituting Ala for Ser158 (S158A), or of the maize SPS (S162A), has been one approach adopted to overcome SPS inactivation by protein phosphorylation (Toroser et al., 1999; Takahashi et al., 2000). However, expression of the maize S162A SPS in transgenic rice did not significantly alter carbon partitioning (Takahashi et al., 2000). Transformation of canola (Brassica napus L.) with the spinach S158A‐SPS gene gave some lines with higher SPS activity in the siliques, but not in the leaves (King, 1997). Several of the transgenic canola lines showed co‐suppression of the endogenous SPS activity (King, 1997). Similarly, transgenic sugarcane (Saccharum officinarum L.) and rice plants containing the spinach S158A‐SPS gene showed no significant increases in leaf SPS activity (CPL Grof, CSIRO Plant Industry, Brisbane, JE Lunn, RT Furbank, unpublished results). The problems of SPS deactivation by protein phosphorylation and co‐suppression of the endogenous SPS by introduced plant SPS genes have influenced the search for alternative sources of genes for achieving overexpression of SPS in plants. The SPS from the cyanobacterium Synechocystis sp. PCC 6803 differs from the plant enzyme in being insensitive to Glc6P and only weakly inhibited by Pi (Lunn et al., 1999). However, an even more interesting feature of the Synechocystis SPS is that it lacks all of the known phosphorylation sites found in SPS from higher plants (Lunn et al., 1999). The Synechocystis sps gene also shows low overall identity with plant SPS gene coding regions; for example, 47% identity with the maize and spinach SPS genes (Lunn et al., 1999). These observations suggested that the cyanobacterial SPS would not be subject to allosteric regulation or inactivation by protein phosphorylation if it were expressed in transgenic plants. The low nucleotide sequence identity with plant SPS genes also suggested that expression of the Synechocystissps gene would be less likely to cause co‐suppression of the endogenous plant SPS. To investigate the effects of expressing this potentially unregulated SPS in plants, the Synechocystissps gene has been introduced into tobacco and tomato under the control of the CaMV 35S promoter and into rice under the control of the maize Ubi1 promoter. For each species, several independent lines of plants expressing the cyanobacterial SPS were obtained and the analysis of these plants is presented in this paper. Materials and methods Materials Tomato (Lycopersicon esculentum cv. UC82B) seeds were obtained from Lefroy Valley (Tyabb, Vic., Australia) and rice (Oryza sativa subsp. japonica cv. Taipei 309) seeds were obtained from Dr Narayana Upadhyaya (CSIRO Plant Industry, Canberra). The maize SPS cDNA clone and anti‐maize SPS antiserum were obtained from Dr Christine Foyer (IACR‐Rothamsted, Harpenden, UK). Plant growth conditions Tobacco (Nicotiana tabacum cv. Wisconsin 38) and tomato plants were grown in 25 cm pots of compost containing 5 g of Osmocote slow‐release fertilizer per pot. After flowering, tomato plants were also fertilized three times per week with liquid fertilizer (Phostrogen tomato fertilizer). Rice plants were grown in 15 cm plastic pots containing a mixture of soil, perlite, sand, and peat moss (50:25:15:10 by vol) and Osmocote, submerged in water. All plants were grown in a naturally illuminated glasshouse with 28 °C day and 20 °C night temperatures. Gene construction and plant transformation Standard cloning procedures were carried out as in Sambrook et al. (1989). For expression of the Synechocystis SPS or maize SPS in tobacco and tomato under the control of the CaMV 35S promoter, the Synechocystis sps (Lunn et al., 1999) or maize SPS (Worrell et al., 1991) coding regions were inserted into the SmaI site of pDH51 (Pietrzak et al., 1986). The resulting 35S‐Synsps‐tm35S or 35S‐ZmSPS‐tm35S gene constructs were excised and inserted into the EcoRI site of the binary vector pPLEX502, containing the nptII selectable marker gene under the control of the clover stunt virus Sc1 promoter (Schünmann et al., 2002). For expression of the Synechocystis SPS in rice under the control of the maize Ubi1 promoter, the sps coding region was inserted between the KpnI and EcoRI sites of pWUbi1.tm1 (Wang et al., 1998). The resulting Ubi1‐Synsps‐tm1 gene construct was excised and inserted into the HindIII site of the binary vector pWBVec8, containing the hpt selectable marker gene under the control of the CaMV 35S promoter (Wang et al., 1998). Binary vectors pPLEX502/35S‐sps‐tm35S, pPLEX502/35S‐ZmSPS‐tm35S and pWBVec8/Ubi1‐sps‐tm1 were introduced into Agrobacteriumtumefaciens strain AGL1 by triparental mating. The binary vectors were re‐isolated from the A. tumefaciens, and the gene coding regions were sequenced to check that no mutations had been introduced during construction of the vectors. Agrobacterium‐mediated transformation of tobacco leaf segments was carried out as described by Horsch et al. (1985), of tomato seedling cotyledons as described by Fillatti et al. (1987) and of rice callus as described by Wang et al. (1998). Transformed tobacco and tomato cells were selected on kanamycin‐containing media, and transformed rice cells were selected on hygromycin‐containing media. Neomycin‐phosphotransferase activity was measured as described in McDonnell et al. (1987). Southern blot analysis was carried out as described in Aoki et al. (2002), using PCR products comprising bases 511 to 1291 of the Synechocystissps coding region or bases 1500 to 1995 of the maize SPS coding region as probes. Expression of the Synechocystis SPS in E. coli The Synechocystis SPS was expressed in E. coli as a fusion protein, linked at the C‐terminus by a self‐cleaving intein domain to a chitin‐binding protein (Lunn, 2002). The fusion protein was purified from E. coli cell extracts by binding to chitin beads, and intein‐mediated cleavage of the fusion protein was induced by incubating with 50 mM dithiothreitol (DTT) for 16 h at 4 °C to release the free Synechocystis SPS protein. SDS‐polyacrylamide gel electrophoresis (SDS‐PAGE) and immunoblot analysis Proteins were separated by SDS‐PAGE on 7% or 9% (w/v) polyacrylamide gels as described in Laemmli (1970) and either stained with Coomassie Brilliant Blue R‐250 or transferred to a nitrocellulose membrane by electroblotting. Membranes were probed with either anti‐Synechocystis SPS antibody (1:10 000 dilution in blocking buffer) or anti‐maize SPS antiserum (1:500 dilution) as described in Lunn et al. (1999). Anti‐phosphoSer, phosphoThr and phosphoTyr antibodies were obtained from Sigma‐Aldrich and used at a dilution of 1:500 in 25 mM Tris‐HCl, pH 7.5, 150 mM NaCl, 0.2% (v/v) Tween 20, and 0.5% (w/v) bovine serum albumin. Extraction and assay of SPS Leaves or leaf discs (0.2–0.5 g) were frozen in liquid N2, and extracted by grinding in a mortar with 1–2 vols of ice‐cold extraction buffer (50 mM Tricine‐KOH, pH 8.0, 10 mM MgCl2, 1 mM EDTA, 5 mM DTT, 1 mM phenymethylsulphonylfluoride (PMSF), 1 mM benzamide, 1 mM benzamidine, 5 mM ϵ‐aminocaproic acid, 10 µM leupeptin, 10 µM antipain, and 2% (w/v) polyvinylpolypyrrolidone (PVPP)) with about 0.1 g of quartz. The crude extract was centrifuged at 11 600 g for 1 min. A 500 µl aliquot of the supernatant was desalted by passage through a column (bed volume 3 ml) of Sephadex G‐25 M (Pharmacia, Uppsala, Sweden), equilibrated with extraction buffer minus PVPP. All procedures were carried out at 4 °C. SPS was assayed in tobacco and tomato leaf extracts by measuring the UDPGlc, ADPGlc or GDPGlc‐dependent synthesis of [14C]Suc6P from [14C]Fru6P as described by Lunn et al. (1997). SPS was assayed in rice leaf extracts by measuring the Fru6P‐dependent production of UDP or GDP from UDPGlc or GDPGlc as described by Lunn and Hatch (1997). The assay buffer was 50 mM Tricine‐KOH, pH 8.0 and 10 mM MgCl2. Chlorophyll was determined in methanolic extracts as described by Porra et al. (1989). Protein was determined by the dye‐binding assay of Bradford (1976) with bovine γ‐globulin as standard. Photosynthesis and carbon partitioning measurements Rates of photosynthetic O2 evolution were measured in a leaf disc oxygen electrode (Hansatech Ltd, Kings Lynn, UK), under 5% CO2, as described in Delieu and Walker (1981). Illumination was provided by a 150 W quartz halogen projector lamp attenuated by neutral density filters. The third and fourth fully expanded leaves were chosen for photosynthesis measurements, 2–3 d after flowering had commenced. Two 3.4 cm2 leaf discs were taken from each side of the mid‐vein of the two leaves and rates of O2 evolution measured at light intensities of 172 and 2000 µmol quanta m–2 s–1 (providing data in quadruplicate). A single leaf disc was used for the full light response curve of photosynthesis. A representative individual transgenic or untransformed plant, from which sucrose‐phosphate synthase activity had been assayed, was chosen for these measurements. Sugars and starch were extracted and assayed as described in Lunn and Hatch (1997). Immunoaffinity chromatography Rabbit polyclonal antibodies, raised against the His6‐Synechocystis SPS (Lunn et al., 1999), were purified from the antiserum by affinity chromatography on Protein G‐Sepharose (Pharmacia) and then coupled to cyanogen bromide‐activated Sepharose (Pharmacia) according to the manufacturer’s instructions. Leaf tissue (2 g) from 35S‐Synsps tobacco (1‐17) was extracted in 3 ml of ice‐cold PBS (10.1 mM Na2HPO4, 1.8 mM KH2PO4, 137 mM NaCl, and 3 mM KCl, pH 7.4), containing 1 mM EDTA, 1 mM DTT, 2% (w/v) PVPP, and protease inhibitors as described above. All subsequent procedures were carried out at 4 °C. The extract was centrifuged at 35 000 g for 5 min and the supernatant was desalted on a Sephadex G‐25M column equilibrated with the PBS extraction buffer minus PVPP. The desalted extract was applied to the antibody‐Sepharose column (12 ml) that had been equilibrated with PBS. The column was washed with PBS at a flow rate of 1.4 ml min–1 until the A280 of the effluent returned to the baseline, then washed successively with 30 ml each of 0.1 M glycine (pH 2.5), PBS, 2.5 M KSCN and PBS. Proteins in samples of the original extract (20 µl) and fractions collected from the column (200 µl) were precipitated with acetone (final concentration 80%) and analysed by SDS‐PAGE and immunoblotting as described above. Purification of Synechocystis SPS from transgenic tobacco and rice leaves Approximately 90 g of deribbed leaf material from 35S‐Synsps tobacco or Ubi1‐Synsps rice plants were blended for 2 min in a Waring blender in 350 ml of ice‐cold extraction buffer (50 mM Hepes‐NaOH, pH 7.5, 10 mM MgCl2, 1 mM EDTA, 5 mM DTT, and 2% (w/v) PVPP) containing protease inhibitors as above. After filtering through four layers of cheesecloth the crude extract was centrifuged at 35 000 g for 10 min. The supernatant was decanted and fractionated with polyethylene glycol 8000 (PEG). Protein in the 10–20% PEG fraction was dissolved in 20 ml of buffer A (20 mM Tris‐HCl, pH 8.0, 1 mM EDTA, and 2 mM DTT) containing protease inhibitors. All procedures to this point were carried out at 4 °C and subsequent procedures were carried out at room temperature. The 10–20% PEG fraction from tobacco was applied to a 5 ml Econo Q anion‐exchange column (BioRad, Hercules, CA, USA), equilibrated with buffer A, at a flow rate of 2 ml min–1. Proteins were eluted with a linear salt gradient (0–1 M NaCl over 50 ml), collecting 2 ml fractions. The bulk of the Synechocystis SPS protein eluted between 280–340 mM NaCl. The corresponding fractions were pooled and desalted by passage through a 20 ml column of Sephadex G25M equilibrated with buffer A. The desalted EconoQ fractions from tobacco or the 10–20% PEG fraction from rice were applied to a 1 ml MonoQ anion‐exchange column (Pharmacia), equilibrated with buffer A, at a flow rate of 1 ml min–1. Proteins were eluted with a linear salt gradient (0–1 M NaCl over 30 ml), collecting 0.5 ml fractions after 6 min. The Synechocystis SPS protein eluted between 350–380 mM NaCl. The two peak fractions (1 ml) were combined and concentrated to a volume of about 200 µl using a Millipore Ultrafree‐MC centrifugal concentrator (10 000 nominal MW cutoff; Millipore, Bedford, MA, USA). The concentrated protein was applied to a 24 ml Superdex 200 (Pharmacia) gel filtration column equilibrated with buffer A containing 150 mM NaCl, at a flow rate of 0.5 ml min–1. Fractions of 0.2 ml were collected after 20 min. The Synechocystis SPS peak eluted at 27.6 min (14.3 ml). Results Expression of the Synechocystis SPS in tobacco The Synechocystissps gene was introduced into tobacco under the control of the constitutive CaMV 35S promoter via Agrobacterium‐mediated transformation. Twenty plants, regenerated on kanamycin‐containing media, showed neomycin phosphotransferase activity indicating that they contained the nptII selectable marker gene. The plants were screened for expression of the Synechocystis SPS with an antibody raised against the purified His6‐SynSPS (Lunn et al., 1999). Ten of the transgenic plants contained an immunoreactive 82 kDa protein, whereas untransformed tobacco plants did not. Figure 1A shows representative results from one transgenic tobacco plant expressing the Synechocystis SPS (lane 2) and one untransformed tobacco plant (lane 1). Southern blot hybridization confirmed that these ten plants contained the 35S‐Synsps gene construct, and originated from independent transformation events (data not shown). Of these ten T0 plants, individuals 1‐18 and 2‐22 were estimated to contain three and at least six copies of the transgene, respectively, but all the other plants contained single transgene loci (data not shown). SPS activity was measured in leaf extracts from these ten plants with UDPGlc or GDPGlc as the substrate. Both the endogenous tobacco SPS and the Synechocystis SPS would contribute to the UDPGlc‐dependent SPS activity, whereas only the Synechocystis SPS would show GDPGlc‐dependent activity (Curatti et al., 1998; Lunn et al., 1999). In this initial screening the UDPGlc‐dependent SPS activity in the transgenic tobacco plants ranged from 102–151% of that in wild‐type plants (Fig. 2). Low GDPGlc‐dependent SPS activity was detected in some of the transgenic plants, but even in plant 1‐17 where this was highest, 0.089 µmol min–1 mg–1 Chl, it was less than 12% of the UDPGlc‐dependent activity. There was no correlation between GDPGlc‐dependent activity and the apparent increase in UDPGlc‐dependent activity over wild‐type plants. For example, plant 2‐5 consistently showed higher UDPGlc‐dependent SPS activity than wild‐type plants on a chlorophyll basis, but no detectable GDPGlc‐dependent activity. These results indicated that, even in the highest Synechocystis SPS‐expressing plants, very little of the total SPS activity was attributable to the Synechocystis SPS. The immunoblot signal from 20 µg of leaf protein from tobacco plant 1‐17 was comparable to that from about 50 ng of the purified His6‐SynSPS. From the specific activities of the His6‐SynSPS (Table 1), it was estimated that the Synechocystis SPS should contribute about 43 nmol min–1 mg–1 protein UDPGlc‐dependent SPS activity and 27 nmol min–1 mg–1 protein GDPGlc‐dependent SPS activity to the total leaf activity in this plant. This translates to a total expected SPS activity of about 1.7 µmol min–1 mg–1 Chl with UDPGlc as substrate and 0.7 µmol min–1 mg–1 Chl with GDPGlc as substrate. Clearly, there is a large discrepancy between the SPS activities detected in plant 1‐17 and those predicted from these estimates. A recovery experiment was carried out in which purified His6‐SynSPS was added to a tobacco leaf sample before extraction. SPS activity was assayed in both the supplemented extract and in a control leaf extract without added enzyme. The UDPGlc‐dependent SPS activity in the His6‐SynSPS‐supplemented extract was much higher than in the control extract (Table 2), with the difference in activity corresponding to 109% of that expected from the amount of enzyme added. There was no detectable ADPGlc‐ or GDPGlc‐dependent SPS activity in the control extract, whereas the supplemented extract showed high activity with both substrates, corresponding to 98% and 138% of the expected activity of the added His6‐SynSPS (Table 1). In this experiment a leaf extract from 35S‐Synsps tobacco line 1‐17 showed no detectable ADPGlc‐ or GDPGlc‐dependent SPS activity. A homozygous line was established from the primary transformant 35S‐Synsps 1‐17, for further analysis. T1 progeny from 35S‐Synsps 1‐17 were screened for expression of the Synechocystis SPS by immunoblotting. Five out of 13 T1 plants showed high level expression of the Synechocystis SPS protein, comparable to the parent plant (data not shown). T2 plants were grown from seed of these five T1 plants and screened by immunoblotting. Two lines showed segregation of Synechocystis SPS‐expressing and non‐expressing progeny, indicating that their T1 parents were hemizygous. For each of the other three lines, 16 out of 16 plants tested all showed expression of the Synechocystis SPS, indicating that the parental T1 plants were homozygous for the 35S‐Synsps gene. One of these lines (35S‐Synsps 1‐17‐2) was selected at random for further analysis. This tobacco line contains a single transgenic locus and expresses the Synechocystis SPS protein at a high level, comparable to the parental plant. The UDPGlc‐dependent SPS activity in leaves of these plants was 0.71±0.16 µmol min–1 mg–1 Chl compared to 0.57±0.13 µmol min–1 mg–1 Chl in leaves from wild‐type plants (means ±SD, n=3). Activity in the roots was 11.8±5.5 nmol min–1 mg–1 protein and 6.5±2.1 nmol min–1 mg–1 protein, respectively. Carbon partitioning in the leaves of these plants was determined by measuring the soluble sugar and starch content of leaf samples harvested before sunrise and in the late afternoon. No significant differences were observed in the soluble sugar and starch contents of wild‐type and Synechocystis SPS‐expressing plants (data not shown). The total amounts of carbohydrate accumulated in the leaves of the wild type and transgenic plants during the day were 41.3 and 42.9 µmol (hexose) mg–1 Chl, and the ratios of starch to sucrose accumulated were 5.5 and 6.2, respectively (n=6). Expression of the Synechocystis SPS in rice The Synechocystissps gene was introduced into rice under the control of the constitutive maize Ubi1 promoter via Agrobacterium‐mediated transformation. Fifty‐four rice plants were regenerated on hygromycin‐containing media from 31 separate calli. Of these, 28 plants showed expression of the Synechocystis SPS protein by immunoblotting with the anti‐Synechocystis SPS antibody. Figure 1A shows representative results from one transgenic rice plant expressing the Synechocystis SPS (lane 4) and one untransformed rice plant (lane 3). Southern blot hybridization with probes for the hpt selectable marker gene and the Ubi1‐Synsps gene confirmed that all of the plants were transformed. Of the 19 pairs of plants derived from single calli, 17 pairs were judged to be sibling clones from the Southern blot hybridization. Seven of the lines were estimated to contain between 8 and 12 copies of the Ubi1‐Synsps gene, and none of these lines showed detectable expression of the Synechocystis SPS protein. Those lines that did express the Synechocystis SPS protein contained between one and four copies of the transgene. Overall, there were 16 independent lines expressing detectable Synechocystis SPS protein. None of these showed significantly different UDPGlc‐dependent SPS activity in the leaves compared to untransformed plants (Fig. 3). No ADPGlc‐ or GDPGlc‐dependent SPS activity was detected in any of the plants. Expression of the Synechocystis SPS in tomato These results from tobacco and rice, and the previously reported difficulties in over‐expressing SPS in these two species (Frommer and Sonnewald, 1995; Ono et al., 1999b) are in marked contrast with the successful over‐expression of SPS in transgenic tomato plants (Worrell et al., 1991; Galtier et al., 1995). This led to the suggestion that, for some reason, tomato is more amenable for achieving over‐expression of SPS than some other species. Therefore, it was decided to investigate whether the Synechocystis SPS could be expressed and show activity in tomato. As a positive control, the maize SPS was also expressed in tomato. The 35S‐Synsps and 35S‐ZmSPS gene constructs were introduced into tomato (cv. UC82B) via Agrobacterium‐mediated transformation. Twenty‐eight plants were regenerated on kanamycin‐containing medium from transformation with the 35S‐Synsps gene construct, of which six showed expression of the Synechocystis SPS protein by immunoblotting. Figure 1A shows representative results from one transgenic tomato plant expressing the Synechocystis SPS (lane 6) and one untransformed tomato plant (lane 5). The strength of the immunoblot signal was comparable to that in 35S‐Synsps tobacco line 1‐17. Similarly, 27 plants were recovered from transformation with the 35S‐ZmSPS gene construct, of which six showed expression of the maize SPS protein by immunoblotting. Figure 1B shows representative results from one transgenic tomato plant expressing the maize SPS (lane 8) and one untransformed tomato plant (lane 7). SPS assays on leaves of the primary transformants showed that two of the plants expressing the maize SPS protein, numbers 50 and 71, had activities of 1.9 and 1.7 µmol min–1 m–1g Chl, respectively. These activities were about 4‐fold higher than that in leaves of wild‐type plants, 0.4 µmol min–1 mg–1 Chl. On a protein basis, the activities in plants 50 and 71 were 40 nmol min–1 mg–1 protein, which is about 2‐fold higher than in the wild‐type plants (22 nmol min–1 mg–1 protein). One of the transgenic plants expressing the Synechocystis SPS protein at a high level, plant 5, showed higher SPS activity on a Chl basis, (1.0 µmol min–1 mg–1 Chl), than the wild‐type plant, but slightly lower activity than wild type on a protein basis (17 nmol min–1 mg–1 protein). The primary transformants were recovered from tissue culture over several months, and it was difficult to obtain good replicate samples from plants at different developmental stages for comparison of SPS activities. Therefore, further analysis was carried out on T1 progeny from plants 5, 50 and 71, all grown under the same conditions. Plant 35S‐Synsps 5 had a single transgene locus, judged by Southern blot hybridization, and the 35S‐ZmSPS 50 and 71 plants had five and three transgene loci, respectively. Twelve T1 progeny from each of these three parents were screened for expression of the Synechocystis or maize SPS proteins by immunoblotting. SPS activities were measured in all of the immuno‐positive plants and several of the immuno‐negative plants. None of the T1 progeny from 35S‐Synsps plant 5 (5‐1 to 5‐11) showed higher leaf SPS activity than the wild‐type plants (Fig. 4). Also, there were no significant differences in SPS activity between the T1 plants that were expressing the Synechocystis SPS protein (marked by asterisks) and those that were not (Fig. 4). By contrast, six of the T1 plants that expressed the maize SPS protein showed 2–3‐fold higher SPS activities than wild‐type plants (Fig. 4). SPS activities in three of the 35S‐ZmSPS T1 plants that did not express detectable amounts of the maize SPS protein were similar to or lower than those in wild‐type plants. One plant, 71‐1, that was immuno‐positive for the maize SPS showed slightly lower SPS activity than the wild type. However, the activities measured in the three samples from this plant varied considerably (1.35, 0.28 and 0.22 µmol min–1 mg–1 Chl). The fact that two of these samples had only about 25% of the activity in samples from wild‐type plants might indicate that there was suppression of the endogenous tomato SPS in some sectors of the leaves of this plant. Screening of T2 progeny from the 35S‐Synsps 5‐8 T1 plant, as described above, indicated that this plant was homozygous for the single transgene locus. However, the T2 progeny from 35S‐ZmSPS T1 plants 50‐2, 50‐8, 50‐9, 71‐3, 71‐5, and 71‐6 all showed segregation of maize SPS expression. Photosynthetic activity and carbon partitioning were measured in wild type, 35S‐Synsps 5‐8 (T2), and 35S‐ZmSPS 50‐9 (T2) and 71‐6 (T2) plants, which showed UDPGlc‐dependent SPS activities of 0.68, 0.60, 2.02, and 2.48 µmol min–1 mg–1 Chl, respectively. The rates of photosynthetic O2 evolution at a saturating concentration of CO2 (5% v/v) were the same in wild‐type tomato plants and 35S‐Synsps plants expressing the Synechocystis SPS protein (Fig. 5). However, 35S‐ZmSPS plants expressing the maize SPS showed higher rates of photosynthesis, particularly at saturating irradiance (Fig. 5). For example, the 35S‐ZmSPS line 50‐9 had 9% and 58% higher rates of photosynthesis than the wild‐type plant at irradiances of 172 µE m–2 s–1 and 2000 µE m–2 s–1, respectively (Fig. 5). To investigate photoassimilate partitioning, leaves were sampled from the plants before sunrise and in the late afternoon to measure their soluble sugar and starch contents. The starch contents of the tomato leaves varied enormously, reflected in the large error bars (Fig. 6). Similarly high variability had been found in the leaf starch content of the T1 tomato plants. However, it was clear that leaves from the two independent lines expressing the maize SPS (50‐9 and 71‐6) had much less starch than either the wild‐type plants or the transgenic plants expressing the Synechocystis SPS (5‐8) (Fig. 6). The differences between plants in the absolute amounts of sucrose accumulated in the leaves were relatively small. Expression of the Synechocystis SPS in E. coli Even though the Synechocystis SPS protein was apparently expressed at high levels in some lines of all three species, tobacco, rice and tomato, there was little evidence of its activity either in extracts from the plants or from any changes in carbon partitioning. The estimates of expected SPS activity in the 35S‐Synsps tobacco were made on the assumption that the enzyme expressed in the plants would have the same specific activity as the His6‐SynSPS enzyme expressed in E. coli (Lunn et al., 1999). If the presence of the N‐terminal His6‐tag were to increase the specific activity of the enzyme, this would lead to overestimation of the expected SPS activity in the transgenic tobacco plants. To test this possibility, the kinetic properties of the Synechocystis SPS without a His6‐tag were determined. The Synechocystis SPS was expressed in E. coli and purified as described in the Materials and methods. The resulting 82 kDa protein differs from the native Synechocystis SPS, and that expressed in the transgenic tobacco plants, in having a single, extra glycine residue at the C‐terminus arising from the cloning strategy used to make the gene expression construct. The Synechocystis SPS expressed in E. coli has a broad pH optimum centred around 8.5, the same as that of the His6‐SynSPS, and is not activated by Glc6P. It shows activity with UDPGlc, ADPGlc and GDPGlc, with similar Km values to the His6‐tagged enzyme (Table 1). The specific activity with UDPGlc is also the same as that of the His6‐SynSPS, although the specific activities with ADPGlc and GDPGlc as substrates are slightly lower (Table 1). From immunoblots, comparing the amount of Synechocystis SPS protein in plant leaf extracts with known amounts of the enzyme expressed in E. coli, the expected activity of the Synechocystis SPS expressed in 35S‐Synsps tobacco line 1‐17 was estimated to be about 0.17 µmol min–1 mg–1 protein, equivalent to 4 µmol min–1 mg–1 Chl. Similarly, the expected Synechocystis SPS activity in Ubi1‐Synsps rice line 43, which had the strongest immunoblot signal, was estimated to be about 0.06 µmol min–1 mg–1 protein, or 2 µmol min–1 mg–1 Chl. These estimates indicated that leaf SPS activity should have been about 8‐fold higher than wild type in the highest expressing tobacco lines and over 2‐fold higher in the transgenic rice, if the Synechocystis SPS enzyme expressed in the plants were fully active. Analysis of the Synechocystis SPS expressed in plants The failure to detect any enzyme activity reproducibly in vitro from the Synechocystis SPS expressed in plants, and the lack of evidence that the enzyme was active in vivo, led to the investigation of the enzyme’s apparent inactivity. Although the Synechocystis SPS does not have any of the known phosphorylation sites of the higher plant enzyme, the possibility that it does have cryptic phosphorylation sites cannot be excluded. Antibodies that recognize phosphorylated‐Ser, Thr or Tyr residues cross‐reacted with several proteins in leaf extracts from wild‐type tobacco and rice plants. However, they did not specifically recognize any 82 kDa protein from the 35S‐Synsps tobacco or Ubi1‐Synsps rice plants (data not shown). Although no evidence for phosphorylation of the Synechocystis SPS was found, it was decided to purify the protein from transgenic tobacco plants in order to investigate whether there was some other post‐translational modification to the protein that could explain its apparent inactivity. An anti‐Synechocystis SPS antibody‐Sepharose column was prepared for immunoaffinity purification of the Synechocystis SPS. A desalted leaf extract from 35S‐Synsps tobacco (1‐17) was applied to the column, which was then washed successively with PBS, 0.1 M glycine (pH 2.5), PBS, 2.5 M KSCN, and PBS. Samples of each fraction were analysed by SDS‐PAGE and immunoblotting with the anti‐Synechocystis SPS antibody. The Synechocystis SPS was clearly detectable in the crude extract applied to the column, but not in the flow‐through fraction (Fig. 7B), indicating that the enzyme had bound to the column. The 0.1 M glycine and subsequent PBS wash fractions did not contain any detectable protein, but the 2.5 M KSCN and final PBS wash fractions did show a single protein band by SDS‐PAGE, with a molecular mass of 73 kDa (Fig. 7A). Evidently, this protein was specifically and tightly bound to the anti‐Synechocystis SPS antibody column, but it was smaller than the Synechocystis SPS (82 kDa) and was not recognized by the anti‐Synechocystis SPS antibody in immunoblots (Fig. 8B). The antibody did not show a reaction with any 73 kDa protein in crude tobacco leaf extracts either (Fig. 1). The 73 kDa protein was unlikely to be derived from antibodies leached from the column as it differed in size from a typical IgG antibody (150–160 kDa) or its component light (24 kDa) and heavy (49 kDa) chains, and it was not recognized by the alkaline‐phosphatase conjugated, goat anti‐rabbit‐IgG secondary antibody used for detection on the immunoblot. As immunoaffinity chromatography had failed to purify the Synechocystis SPS, an alternative procedure was developed using PEG precipitation and anion‐exchange chromatography on an EconoQ and then a MonoQ column. Initially, fractions containing the Synechocystis SPS protein were identified by immunoblotting. The crude extract and PEG fractions showed UDPGlc‐dependent SPS activity, attributable to the endogenous tobacco SPS, and in immunoblots contained a single, 82 kDa protein band that was recognized by the anti‐Synechocystis SPS antibody (Fig. 8A). Fractions from the anion‐exchange columns that contained the Synechocystis SPS protein showed ADPGlc and GDPGlc‐dependent SPS activities, as well as UDPGlc‐dependent activity. The EconoQ fractions contained a 54 kDa immunoreactive protein in addition to the 82 kDa protein. In several experiments the 54 kDa protein was the only immunoreactive protein detected in MonoQ fractions with SPS activity, and these did not contain any of the full length, 82 kDa Synechocystis SPS (Fig. 8A). This showed that the 54 kDa truncated form of the Synechocystis SPS must be catalytically active. The truncated enzyme was purified further by gel filtration on a Superdex 200 column. In this experiment, the fraction from the MonoQ column that was applied to the Superdex 200 column contained only two immunoreactive proteins, the 54 kDa protein and a minor 30 kDa protein (Fig. 8B). Activity eluted from the gel filtration column in a single peak with a retention time corresponding to a molecular mass of 105 kDa. The peak fractions contained two major protein bands by SDS‐PAGE, with molecular masses of 54 and 41 kDa (Fig. 8C). The kinetic properties of this highly purified Synechocystis SPS (Superdex 200 fractions 17‐19) were investigated. The enzyme showed a broad pH optimum around 8.0 to 8.5, was active in the presence of 10 mM EDTA and in the absence of added Mg2+, and was not activated by 17.5 mM Glc6P. It showed activity with UDPGlc, ADPGlc and GDPGlc as substrates, with similar Km values to the enzyme expressed in E. coli (Table 1). The Vmax activities with all three substrates were higher than for the enzyme expressed in E. coli, but the relative activities with each substrate were comparable, with UDPGlc>ADPGlc> GDPGlc (Table 1). It should be noted that the 54 kDa form of the Synechocystis SPS enzyme probably represented only about half of the protein in these preparations (Fig. 8C), therefore the Vmax values of the pure truncated enzyme would be about 2‐fold higher than the values shown in Table 1. The 54 and 30 kDa immunoreactive proteins (Fig. 8A, B) were presumably cleavage products arising from proteolysis of the 82 kDa Synechocystis SPS during the purification procedure. To investigate this process further, two leaf extracts were prepared from 35S‐Synsps 1‐17 tobacco, one in the usual way including serine and cysteine‐protease inhibitors in the extraction and desalting buffers (see Materials and methods) and the other without protease inhibitors. The two extracts were incubated at 25 °C for 1 h with samples taken at intervals for immunoblotting and SPS assays. In both extracts only an 82 kDa protein was recognized by the anti‐Synechocystis SPS antibody, even after incubation at 25 °C for 1 h without protease inhibitors (data not shown). The initial activities in the extracts with and without protease inhibitors were 0.39 and 0.37 µmol min–1 mg–1 Chl, respectively. In the presence of protease inhibitors, 92% of the initial SPS activity remained after 1 h, whereas in the absence of protease inhibitors it had fallen to 49% of the initial activity after 1 h. The Synechocystis SPS was also partially purified from Ubi1‐SynSPS rice plants by PEG fractionation and anion‐exchange chromatography on a MonoQ column. Fraction number 14 from the MonoQ column contained the most Synechocystis SPS protein, as judged by immunoblotting (Fig. 9A), and showed SPS activities of 393, 218 and 144 nmol min–1 mg–1 protein with UDPGlc, ADPGlc and GDPGlc as substrates, respectively (Fig. 9B). The peak of UDPGlc‐dependent activity was in fraction 13, which showed much lower relative activities with ADPGlc and GDPGlc (Fig. 9C). This activity is largely attributable to the endogenous rice SPS that was not fully resolved from the Synechocystis enzyme. Fractions 13 and 14 showed two minor immunoreactive protein bands of about 60 and 62 kDa in addition to the main 82 kDa protein band (Fig. 9). However, the presence of these bands did not correlate with ADPGlc or GDPGlc‐dependent SPS activity, as these activities were higher in fraction 15 than 13, but the former contained only the 82 kDa immunoreactive protein and not the smaller proteins (Fig. 9). Discussion The Synechocystissps gene promised to have several advantages over higher plant SPS genes for achieving over‐expression of SPS in transgenic plants, which previous work on tomato had shown could have a beneficial effect on plant productivity. Tobacco and rice were initially chosen for expression of the Synechocystissps gene as representative plants that accumulate mostly starch (tobacco) or sucrose (rice) in the leaves, and are readily transformable via Agrobacterium. The Synechocystissps gene was expressed at high enough levels in the plants, judged by immunoblotting, to achieve 2–8‐fold increases in SPS activity. Despite the high levels of expression of the Synechocystis SPS, little evidence was found that the enzyme was active. A few of the plants did appear to have higher UDPGlc‐dependent activity (Fig. 2), but no ADPGlc or GDPGlc‐dependent SPS activity could reproducibly be found in these plants. Such activities would have been definitive evidence that the cyanobacterial enzyme was active, as the endogenous plant SPS is specific for UDPGlc. Recovery experiments with the purified His6‐tagged enzyme expressed in E. coli showed that the extraction and assay procedures were suitable for detecting these activities (Table 2). Evidence that the Synechocystis SPS was also inactive in vivo was consistent with the lack of detectable enzyme activity in vitro. The Synechocystis SPS was also expressed in tomato (Fig. 1), but no reproducible evidence could be found of its activity in tomato either (Fig. 4). Maximal rates of photosynthesis were also unaffected by expression of the Synechocystis SPS (Fig. 5). The leaf starch contents of the tomato plants were extremely variable, which made it difficult to judge whether there was any significant difference in carbon partitioning between wild type and Synechocystis SPS plants (Fig. 6). By contrast, tomato plants expressing the maize SPS had 2–3‐fold higher SPS activities, higher rates of photosynthesis and lower leaf starch contents (Figs 4–6), in agreement with previous reports (Worrell et al., 1991; Galtier et al., 1993, 1995; Laporte et al., 1997; Micallef et al., 1995; Murchie et al., 1999; Nguyen‐Quoc et al., 1999). None of the plants showed large differences in the amounts of sucrose in their leaves (Fig. 5). Even if more sucrose were synthesized, the high acid invertase activity present in tomato leaves (Gao et al., 1998) would probably limit the amount of sucrose accumulated (Winter and Huber, 2000). However, from the rates of photosynthesis and decreased starch accumulation in the 35S‐ZmSPS plants, it could be inferred that a higher proportion of photoassimilate was being exported from the leaves, reflecting greater partitioning into sucrose. The lack of detectable Synechocystis SPS activity even in tomato, where we and others have been able to increase SPS activity by expressing the maize SPS, suggested that there is a general problem with expression of the cyanobacterial enzyme in plants. No evidence was found that the enzyme was phosphorylated on any Ser, Thr or Tyr residues, which was one possible explanation for the lack of activity in plants. Therefore, it was decided to purify the enzyme from the transgenic tobacco plants for further investigation. The initial attempt at purification by immunoaffinity chromatography yielded a 73 kDa protein, which although evidently tightly bound to the antibody‐Sepharose column, was not recognized by the Synechocystis SPS antibody (Fig. 7). After this puzzling result, more conventional procedures were used to purify the enzyme. Surprisingly, the Synechocystis SPS purified from leaves of transgenic tobacco plants showed full SPS activity with UDPGlc, ADPGlc and GDPGlc as substrates (Table 1) with a broad pH optimum around 8.0–8.5, and was not activated by Glc6P. These distinctive properties are characteristic of the Synechocystis SPS (Lunn et al., 1999). Similar results were also obtained from rice (Fig. 9). These results conclusively showed that the Synechocystis SPS protein expressed in the transgenic plants was inherently active. The lack of activity in crude extracts must, therefore, be due to inhibition of the enzyme by some factor in the plant cells, which almost certainly also inhibited the enzyme in vivo. No Synechocystis SPS activity was detectable even in desalted extracts, so the inhibitor was unlikely to be a low molecular weight compound. The most likely alternative seemed to be another protein, and the unexpected result from the immunoaffinity chromatography experiment suggested a possible candidate. If the 73 kDa protein purified in this experiment were bound to the Synechocystis SPS, this could explain why it was retained on the column, but not recognized by the anti‐Synechocystis SPS antibodies. If plant cells do contain a protein that binds and inhibits the Synechocystis SPS, an inhibition of the enzyme might have been expected in the recovery experiment (Table 2). However, the rate of association between the putative inhibitory protein and the Synechocystis SPS could be too slow to be seen in the time scale of the recovery experiment. Another interesting observation from the purification of the Synechocystis SPS from tobacco was the appearance of truncated forms of the enzyme. The appearance of the 54 kDa and 30 kDa immunoreactive proteins coincided with the disappearance of the 82 kDa form of the enzyme, suggesting that the Synechocystis SPS had been proteolytically cleaved (Fig. 8A, B). The 54 kDa‐protein was the only immunoreactive protein in some fractions, but these contained high SPS activity showing that the truncated enzyme was catalytically active. The Synechocystis SPS, like the higher plant enzyme, contains two domains; a larger, N‐terminal glucosyltransferase domain and a smaller, C‐terminal domain that shows similarity to SPP (Lunn et al., 2000). The glucosyltransferase domain of the Synechocystis SPS contains about 464 amino acid residues while the SPP‐like domain contains about 256 residues, with predicted molecular masses of 52 and 29 kDa, respectively. These values are very similar to the 54 and 30 kDa proteins derived from the Synechocystis SPS, suggesting that the latter had been cleaved close to the junction between the two domains. This implies that the 54 kDa protein fragment contains the intact glucosyltransferase domain, consistent with it retaining SPS activity. The truncated form of the enzyme was observed in several purification experiments from tobacco, and seemed to be linked to the appearance of SPS activity. An experiment to test this correlation, by incubating a crude leaf extract at 25 °C without protease inhibitors to favour proteolysis of the Synechocystis SPS, was inconclusive, because the protein was not cleaved. However, when the enzyme was partially purified from rice leaves there was no correlation between the presence of truncated Synechocystis SPS protein and enzyme activity (Fig. 9). This indicates that it was purification of the enzyme not proteolytic cleavage that allowed activity to be revealed, and that the latter was coincidental and not essential for seeing activity. From these results, it is proposed that the Synechocystis SPS expressed in plants is inherently active, but is inhibited in vivo by binding of an endogenous plant protein, possibly the 73 kDa protein purified on the antibody‐Sepharose column. Removal of the putative inhibitor protein during purification of the Synechocystis SPS reveals its activity, and perhaps allows proteolytic cleavage of the enzyme to give the truncated 54 kDa form that retains catalytic activity. This form of the enzyme appears to be dimeric, whereas the intact enzyme is monomeric. As mentioned in the Introduction, Ser229 in the spinach leaf SPS is the phosphorylation site involved in binding of 14‐3‐3 proteins (∼30 kDa) to the enzyme (Toroser et al., 1998). When A. thaliana suspension cells were starved of sugar, the amount of 14‐3‐3 protein bound to SPS decreased and the enzyme was proteolytically cleaved, but still retained catalytic activity (Cotelle et al., 2000). The truncated SPS had a molecular mass of about 90–95 kDa, which is very similar to the predicted size of the N‐terminal, glucosyltransferase domain of a typical higher plant SPS. This suggests that the enzyme had been cleaved close to the junction between the glucosyltransferase domain and the C‐terminal, SPP‐like domain (Lunn et al., 2000). There are several reports of preparations of SPS purified from plants containing proteolytic fragments of about 90 and 30 kDa (Lunn and ap Rees, 1990b; Bruneau et al., 1991). The observation that the Synechocystis SPS, like the higher plant enzyme, is also prone to specific cleavage in this junction region suggests that proteolytic processing of SPS to remove the SPP‐like domain could be a widespread mechanism for regulating SPS activity. The significance of this for control of sucrose synthesis is unclear and warrants further investigation. Acknowledgements We thank Dr Narayana Upadhyaya, Dr Petra Schünmann and Ms Judy Gaudron (CSIRO Plant Industry, Canberra) for their help with the rice transformation, and Dr Christine Foyer (IACR‐Rothamsted, Harpenden, UK) for the generous gifts of the maize SPS cDNA clone and anti‐maize SPS antiserum . Open in new tabDownload slide Fig. 1. Expression of the Synechocystis SPS and maize SPS in transgenic plants. The Synechocystis SPS (A) and maize SPS (B) proteins were detected in leaf extracts (20 µg protein) from transgenic plants by immunoblotting. (1) Wild‐type (WT) tobacco cv. Wisconsin 38; (2) 35S‐Synsps tobacco 1‐17 (T2); (3) WT rice cv. Taipei 309; (4) Ubi1‐Synsps rice 43 (T0); (5) WT tomato cv. UC82B; (6) 35S‐Synsps tomato 5 (T1); (7) WT tomato; (8) 35S‐ZmSPS tomato 50 (T1). Open in new tabDownload slide Fig. 2. SPS activity in leaves of wild‐type and 35S‐Synsps tobacco cv. Wisconsin 38 plants. UDPGlc‐ and GDPGlc‐dependent SPS activity was measured in young, fully expanded leaves, harvested at 10.00 h in full sunlight. Data are single leaf measurements. Open in new tabDownload slide Fig. 3. SPS activity in leaves of wild‐type and Ubi1‐Synsps rice cv. Taipei 309 plants. UDPGlc‐dependent SPS activity was measured in young, fully expanded leaves, harvested at 10.00 h in full sunlight. Data are means ±SD (n=6, except for plants 1, 19, 39, 41, 52 where n=3). Open in new tabDownload slide Fig. 4. SPS activity in leaves of wild‐type and T1 tomato plants. UDPGlc‐dependent SPS activity was measured in young, fully expanded leaves, harvested at 10.00 h in full sunlight. WT, wild type; 5‐1 to 5‐11, T1 progeny from 35S‐Synsps plant 5; 50‐1 to 50‐9 and 71‐1 to 71‐8, T1 progeny from 35S‐ZmSPS plants 50 and 71. Data are mean ±SD (n=3). Asterisks indicate plants expressing the Synechocystis SPS or maize SPS proteins. Open in new tabDownload slide Fig. 5. Photosynthesis in wild‐type and transgenic tomato plants. Rates of photosynthetic O2 evolution were measured using a leaf disc oxygen electrode under 5% CO2, and are expressed on a per mg Chl (A) and per unit leaf area (B) basis. Rates were measured at light intensities of 172 and 2000 µmol quanta m–2 s–1 (values are means ±SD, n=4). Single leaf discs were used for the full light response curve of photosynthesis. Open in new tabDownload slide Fig. 6. Carbon partitioning in wild‐type and transgenic tomato plants. Leaf samples were harvested from wild type (WT), 35S‐Synsps line 5‐8 (T2) and 35S‐ZmSPS lines 50‐9 (T2) and 71‐6 (T2) tomato plants at 05.45 h and at 17.00 h. The glucose, fructose, sucrose, and starch contents of the leaf samples are expressed as µmol hexose equivalents on a chlorophyll basis. Data are means ±SD (n=3). Open in new tabDownload slide Fig. 7. Immunoaffinity chromatography of 35S‐Synsps tobacco leaf extract. (A) SDS‐polyacrylamide gel (9% w/v) stained with Coomassie Blue R‐250. (B) Immunoblot of duplicate gel probed with anti‐Synechocystis SPS antibody. Lane 1, crude extract; lane 2, PBS wash 1; lane 3, 0.1 M glycine, pH 2.5, fraction; lane 4, PBS wash 2; lane 5, 2.5 M KSCN fraction; lane 6, PBS wash 3. Open in new tabDownload slide Fig. 8. Purification of the Synechocystis SPS from tobacco. (A) Immunoblot of fractions from the purification of the Synechocystis SPS from 35S‐Synsps tobacco (1‐17) leaves, probed with anti‐Synechocystis SPS antibody. Lane 1, crude extract (10 µg); lane 2, PEG precipitate (5 µg); lane 3, EconoQ fractions 15‐20 (2 µg); lane 4, MonoQ fractions 9‐11 (2 µg). (B) Immunoblot of MonoQ fractions 10‐11 (4 µg) applied to the Superdex 200 column. (C) SDS‐polyacrylamide gel (9% w/v) of samples (150 µl) from fractions 15‐20 eluted from the Superdex 200 column, stained with Coomassie Blue R‐250. UDPGlc‐dependent SPS activities in each fraction are shown at the top. Open in new tabDownload slide Fig. 9. Purification of the Synechocystis SPS from rice. (A) Immunoblot of samples (20 µl) from the purification of the Synechocystis SPS from Ubi1‐Synsps rice leaves, probed with anti‐Synechocystis SPS antibody. Cr, crude extract; 11‐15, MonoQ fractions 11‐15. (B) SPS activities in MonoQ fractions 11‐16, assayed with 20 mM UDPGlc, ADPGlc or GDPGlc and 5 mM Fru6P. Table 1. Kinetic properties of the Synechocystis SPS expressed in E. coli and tobacco The Synechocystis SPS was expressed in E. coli with a C‐terminal intein‐chitin binding protein tag. The fusion protein was bound to chitin beads and the SynSPS released by intein‐mediated cleavage induced by DTT. The Synechocystis SPS was purified from 35S‐Synsps tobacco (1‐17) plants by PEG precipitation, anion‐exchange chromatography and gel filtration as described in Materials and methods. Enzyme Vmax (µmol min–1 mg–1 protein) Km (mM) UDPGlca ADPGlca GDPGlca UDPGlca ADPGlca GDPGlca Fru6Pb His6‐SynSPSc 17.0 15.0 9.3 2.9 2.5 1.8 0.2 SynSPSE. coli 16.9 13.4 6.8 2.2 2.7 1.5 0.2 SynSPStobacco 23.8 19.1 12.0 1.8 2.0 2.4 0.3 Enzyme Vmax (µmol min–1 mg–1 protein) Km (mM) UDPGlca ADPGlca GDPGlca UDPGlca ADPGlca GDPGlca Fru6Pb His6‐SynSPSc 17.0 15.0 9.3 2.9 2.5 1.8 0.2 SynSPSE. coli 16.9 13.4 6.8 2.2 2.7 1.5 0.2 SynSPStobacco 23.8 19.1 12.0 1.8 2.0 2.4 0.3 a Assayed with 5 mM Fru6P. b Assayed with 20 mM UDPGlc. c From Lunn et al. (1999). Open in new tab Table 1. Kinetic properties of the Synechocystis SPS expressed in E. coli and tobacco The Synechocystis SPS was expressed in E. coli with a C‐terminal intein‐chitin binding protein tag. The fusion protein was bound to chitin beads and the SynSPS released by intein‐mediated cleavage induced by DTT. The Synechocystis SPS was purified from 35S‐Synsps tobacco (1‐17) plants by PEG precipitation, anion‐exchange chromatography and gel filtration as described in Materials and methods. Enzyme Vmax (µmol min–1 mg–1 protein) Km (mM) UDPGlca ADPGlca GDPGlca UDPGlca ADPGlca GDPGlca Fru6Pb His6‐SynSPSc 17.0 15.0 9.3 2.9 2.5 1.8 0.2 SynSPSE. coli 16.9 13.4 6.8 2.2 2.7 1.5 0.2 SynSPStobacco 23.8 19.1 12.0 1.8 2.0 2.4 0.3 Enzyme Vmax (µmol min–1 mg–1 protein) Km (mM) UDPGlca ADPGlca GDPGlca UDPGlca ADPGlca GDPGlca Fru6Pb His6‐SynSPSc 17.0 15.0 9.3 2.9 2.5 1.8 0.2 SynSPSE. coli 16.9 13.4 6.8 2.2 2.7 1.5 0.2 SynSPStobacco 23.8 19.1 12.0 1.8 2.0 2.4 0.3 a Assayed with 5 mM Fru6P. b Assayed with 20 mM UDPGlc. c From Lunn et al. (1999). Open in new tab Table 2. Activity of the Synechocystis SPS in tobacco leaf extracts SPS activity was measured in desalted extracts from duplicate leaf samples (90–110 µg Chl) from wild‐type tobacco plants, to one of which was added 50 ng of purified His6‐SynSPS (0.87 µmol min–1) before extraction. Activity was also measured in a leaf sample from a 35S‐Synsps tobacco plant (line 1‐17). SPS assays contained 10 mM UDPGlc, ADPGlc or GDPGlc and 5 mM Fru6P. Plant extract SPS activity (µmol min–1 mg–1 Chl) UDPGlc ADPGlc GDPGlc Wild type 0.51 <0.01 <0.01 Wild type + His6‐SynSPS 7.39 6.01 5.58 35S‐Synsps line 1‐17 0.71 <0.01 <0.01 Plant extract SPS activity (µmol min–1 mg–1 Chl) UDPGlc ADPGlc GDPGlc Wild type 0.51 <0.01 <0.01 Wild type + His6‐SynSPS 7.39 6.01 5.58 35S‐Synsps line 1‐17 0.71 <0.01 <0.01 Open in new tab Table 2. Activity of the Synechocystis SPS in tobacco leaf extracts SPS activity was measured in desalted extracts from duplicate leaf samples (90–110 µg Chl) from wild‐type tobacco plants, to one of which was added 50 ng of purified His6‐SynSPS (0.87 µmol min–1) before extraction. Activity was also measured in a leaf sample from a 35S‐Synsps tobacco plant (line 1‐17). SPS assays contained 10 mM UDPGlc, ADPGlc or GDPGlc and 5 mM Fru6P. Plant extract SPS activity (µmol min–1 mg–1 Chl) UDPGlc ADPGlc GDPGlc Wild type 0.51 <0.01 <0.01 Wild type + His6‐SynSPS 7.39 6.01 5.58 35S‐Synsps line 1‐17 0.71 <0.01 <0.01 Plant extract SPS activity (µmol min–1 mg–1 Chl) UDPGlc ADPGlc GDPGlc Wild type 0.51 <0.01 <0.01 Wild type + His6‐SynSPS 7.39 6.01 5.58 35S‐Synsps line 1‐17 0.71 <0.01 <0.01 Open in new tab References AokiN, Whitfeld P, Hoeren F, Scofield G, Newell K, Patrick J, Offler C, Clarke B, Rahman S, Furbank RT. 2002 . Three sucrose transporter genes are expressed in the developing grain of hexaploid wheat. Plant Molecular Biology 50, 453 –462. BradfordMM. 1976 . A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding. Analytical Biochemistry 72, 248 –254. BruneauJM, Worrell AC, Cambou B, Lando D, Voelker TA. 1991 . Sucrose‐phosphate synthase, a key enzyme for sucrose biosynthesis in plants. Plant Physiology 96, 473 –478. CotelleV, Meek SEM, Provan F, Milne FC, Morrice N, MacKintosh C. 2000 . 14‐3‐3s regulate global cleavage of their diverse binding partners in sugar‐starved Arabidopsis cells. EMBO Journal 19, 2869 –2876. CurattiL, Folco E, Desplats P, Abratti G, Limones V, Herrera‐Estrella L, Salerno G. 1998 . Sucrose‐phosphate synthase from Synechocystis sp. PCC 6803: identification of the spsA gene and characterization of the enzyme expressed in Escherichia coli. Journal of Bacteriology 180, 6776 –6779. DelieuT, Walker DA. 1981 . Polarographic measurement of photosynthetic oxygen evolution by leaf discs. New Phytologist 89, 165 –175. FillattiJJ, Kiser K, Rose R, Comai L. 1987 . Efficient transfer of a glyphosate tolerance gene into tomato using a binary Agrobacterium tumefaciens vector. Bio/technology 5, 26 –30. FrommerWB, Sonnewald U. 1995 . Molecular analysis of carbon partitioning in solanaceous species. Journal of Experimental Botany 46, 587 –607. GaltierN, Foyer CH, Huber J, Voelker TA, Huber SC. 1993 . Effects of elevated sucrose‐phosphate synthase activity on photosynthesis, assimilate partitioning, and growth in tomato (Lycopersicon esculentum var. UC82B). Plant Physiology 101, 535 –543. GaltierN, Foyer CH, Murchie E, Alred R, Quick P, Voelker TA, Thépenier C, Lascève, Betsche T. 1995 . Effects of light and atmospheric carbon dioxide enrichment on photosynthesis and carbon partitioning in the leaves of tomato (Lycopersicon esculentum L.) plants over‐expressing sucrose‐phosphate synthase. Journal of Experimental Botany 46, 1335 –1344. GaoZ, Sagi M, Lips SH. 1998 . Carbohydrate metabolism in leaves and assimilate partitioning in fruits of tomato (Lycopersicon esculentum L.) as affected by salinity. Plant Science 135, 149 –159. GeigenbergerP, Reimholz R, Deiting U, Sonnewald U, Stitt M. 1999 . Decreased expression of sucrose‐phosphate synthase strongly inhibits the water stress‐induced synthesis of sucrose in growing potato tubers. The Plant Journal 19, 119 –129. GeigenbergerP, Stitt M. 2000 . Diurnal changes in sucrose, nucleotides, starch synthesis and AGPS transcript in growing potato tubers that are suppressed by decreased expression of sucrose phosphate synthase. The Plant Journal 23, 795 –806. HaiglerCH, Hequet EF, Krieg DR,et al. 2000 . Transgenic cotton with improved fiber micronaire, strength, and length and increased fiber weight. Proceedings of the Beltwide Cotton Conference 1, 483 –483. HorschRB, Fry JE, Hoffmann NL, Eichholtz D, Rogers SG, Fraley RT. 1985 . A simple and general method for transferring genes into plants. Science 227, 1229 –1231. HuberSC, Huber JL. 1996 . Role and regulation of sucrose‐phosphate synthase in higher plants. Annual Review of Plant Physiology and Plant Molecular Biology 47, 431 –444. KingSP. 1997 . Carbon and nitrogen partitioning in canola (Brassica napus L.). PhD thesis, Australian National University, Australia, 114 –147. LaemmliUK. 1970 . Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680 –685. LangenkämperG, Fung RWM, Newcombe RD, Atkinson RG, Gardner RC, MacRae EA. 2002 . Sucrose‐phosphate synthase genes in plants belong to three different families. Journal of Molecular Evolution 54, 322 –332. LaporteMM, Galagan JA, Prasch AL, Vanderveer PJ, Hanson DT, Shewmaker CK, Sharkey TD. 2001 . Promoter strength and tissue specificity effects on growth of tomato plants transformed with maize sucrose‐phosphate synthase. Planta 212, 817 –822. LaporteMM, Galagan JA, Shapiro JA, Boersig MR, Shewmaker CK, Sharkey TD. 1997 . Sucrose‐phosphate synthase activity and yield analysis of tomato plants transformed with maize sucrose‐phosphate synthase. Planta 203, 253 –259. LunnJE. 2002 . Evolution of sucrose synthesis. Plant Physiology 128, 1490 –1500. LunnJE, ap Rees T. 1990a. Apparent equilibrium constant and mass‐action ratio for sucrose‐phosphate synthase from seeds of Pisum sativum. The Biochemical Journal 267, 739 –743. LunnJE, ap Rees T. 1990b. Purification and properties of sucrose‐phosphate synthase from seeds of Pisum sativum. Phytochemistry 29, 1057 –1063. LunnJE, Ashton A, Hatch MD, Heldt HW. 2000 . Purification, molecular cloning and sequence analysis of sucrose‐6‐phosphate phosphohydrolase from plants. Proceedings of the National Academy of Sciences, USA 97, 12914 –12919. LunnJE, Furbank RT. 1999 . Sucrose biosynthesis in C4 plants. New Phytologist 143, 221 –237. LunnJE, Furbank RT, Hatch MD. 1997 . Adenosine 5′‐triphosphate‐mediated activation of sucrose‐phosphate synthase in bundle sheath cells of C4 plants. Planta 202, 249 –256. LunnJE, Hatch MD. 1997 . Primary partitioning and storage of photosynthate in sucrose and starch in leaves of C4 plants. Planta 197, 385 –391. LunnJE, Price GD, Furbank RT. 1999 . Cloning and expression of a prokaryotic sucrose‐phosphate synthase gene from the cyanobacterium Synechocystis sp. PCC 6803 . Plant Molecular Biology 40, 297 –305. McDonnellRE, Clark RD, Smith WA, Hinchee MA. 1987 . A simplified method for the detection of neomycin phosphotransferase II activity in transformed plant tissues. Plant Molecular Biology Reporter 5, 380 –386. MicallefBJ, Haskins KA, Vanderveer PJ, Roh KS, Shewmaker CK, Sharkey TD. 1995 . Altered photosynthesis, flowering, and fruiting in transgenic tomato plants that have an increased capacity for sucrose synthesis. Planta 196, 327 –334. MouillonJM, Hurry V. 2001 . Effects of elevated sucrose‐phosphate synthase activity on photosynthesis, carbohydrate partitioning and growth rate in hybrid aspen (Populus tremula×P. tremuloides). PS2001 Proceedings of the 12th International Congress on Photosynthesis, S28–007. MurchieEH, Sarrobert C, Contard P, Betsche T, Foyer CH, Galtier N. 1999 . Overexpression of sucrose‐phosphate synthase in tomato plants grown with CO2 enrichment leads to decreased foliar carbohydrate accumulation relative to untransformed controls. Plant Physiology and Biochemistry 3, 251 –260. Nguyen‐QuocB, N‘Tchobo H, Foyer CH, Yelle S. 1999 . Overexpression of sucrose phosphate synthase increases sucrose unloading in transformed tomato fruit. Journal of Experimental Botany 50, 785 –791. OnoK, Ishimaru K, Aoki N, Ohsugi R. 1999a. Transgenic rice with low sucrose‐phosphate synthase activities retain more soluble protein and chlorophyll during flag leaf senescence. Plant Physiology and Biochemistry 37, 949 –953. OnoK, Ishimaru K, Aoki N, Takahashi S, Ozawa K, Ohkawa Y, Ohsugi R. 1999b. Characterization of a maize sucrose‐phosphate synthase protein and its effect on carbon partitioning in transgenic rice plants. Plant Production Science 2, 172 –177. PietrzakM, Shillito RD, Hohn T, Potrykus I. 1986 . Expression in plants of two bacterial antibiotic resistance genes after protoplast transformation with a new plant expression vector. Nucleic Acids Research 14, 5857 –5868. PorraRJ, Thompson WA, Kriedemann PE. 1989 . Determination of accurate extinction coefficients and simultaneous equations for assaying chlorophylls a and b extracted with four different solvents: verification of the concentration of chlorophyll standards by atomic absorption spectroscopy. Biochimica et Biophysica Acta 975, 384 –394. SambrookJ, Fritsch EF, Maniatis T. 1989 . Molecular cloning: a laboratory manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. SchünmannPHD, Surin B, Defeyter R, Llewellyn DJ, Waterhouse PM. 2002 . A suite of novel promoters and terminators for plant biotechnology. Functional Plant Biology (in press). SieglG, MacKintosh C, Stitt M. 1990 . Sucrose‐phosphate synthase is dephosphorylated by protein phosphatase 2A in spinach leaves. FEBS Letters 270, 198 –202. SignoraL, Galtier N, Skøt L, Lucas H, Foyer CH. 1998 . Over‐expression of sucrose phosphate synthase in Arabidopsis thaliana results in increased foliar sucrose/starch ratios and favours decreased foliar carbohydrate accumulation in plants after prolonged growth with CO2 enrichment. Journal of Experimental Botany 49, 669 –680. StrandÅ, Foyer CH, Gardeström P, Hurry V. 2001 . Increased expression of sucrose‐phosphate synthase in transgenic Arabidopsis thaliana results in improved photosynthetic performance and increased freezing tolerance at low temperatures. Proceedings of the 12th International Congress on Photosynthesis, S35–003. (http://www.rsbs.anu.edu.au/rsbsweb/conferences/ps2001) StrandÅ, Zrenner R, Trevanion S, Stitt M, Gustafsson P, Gardeström P. 2000 . Decreased expression of two key enzymes in the sucrose biosynthesis pathway, cytosolic fructose‐1,6‐bisphosphatase and sucrose‐phosphate synthase, has remarkably different consequences for photosynthetic carbon metabolism in transgenic Arabidopsis thaliana. The Plant Journal 23, 759 –770. TakahashiS, Ono K, Ugaki M, Ishimaru K, Aoki N, Ohsugi R. 2000 . Ser162‐dependent inactivation of overproduced sucrose‐phosphate synthase protein of maize leaf in transgenic rice plants. Plant and Cell Physiology 41, 977 –981. ToroserD, Athwal GS, Huber SC. 1998 . Site‐specific regulatory interaction between spinach leaf sucrose‐phosphate synthase and 14‐3‐3 proteins. FEBS Letters 435, 110 –114. ToroserD, Huber SC. 1997 . Protein phosphorylation as a mechanism for osmotic‐stress activation of sucrose‐phosphate synthase in spinach leaves. Plant Physiology 114, 947 –955. ToroserD, McMichael Jr R, Krause KP, Kurreck J, Sonnewald U, Stitt M, Huber SC. 1999 . Site‐directed mutagenesis of serine 158 demonstrates its role in spinach leaf sucrose‐phosphate synthase modulation. The Plant Journal 17, 407 –413. WangMB, Li ZY, Matthews PR, Upadhyaya NM, Waterhouse PM. 1998 . Improved vectors for Agrobacterium tumefaciens‐mediated transformation of monocot plants. Acta Horticulture 461, 401 –407. WinterH, Huber SC. 2000 . Regulation of sucrose metabolism in higher plants: localization and regulation of activity of key enzymes. Critical Reviews in Biochemistry and Molecular biology 35, 253 –289. WorrellAC, Bruneau JM, Summerfelt K, Boersig M, Voelker TA. 1991 . Expression of a maize sucrose‐phosphate synthase in tomato alters leaf carbohydrate partitioning. The Plant Cell 3, 1121 –1130.

Robbins, Mark Paske; Paolocci, Francesco; Hughes, John‐Wayne; Turchetti, Valentina; Allison, Gordon; Arcioni, Sergio; Morris, Phillip; Damiani, Francesco

doi: 10.1093/jxb/erg022pmid: 12493851

Abstract Anthocyanins and condensed tannins are major flavonoid end‐products in higher plants. While the transactivation of anthocyanins by basic helix‐loop‐helix (bHLH) transcription factors is well documented, very little is known about the transregulation of the pathway to condensed tannins. The present study analyses the effect of over‐expressing an Sn transgene in Lotus corniculatus, a model legume, with the aim of studying the regulation of anthocyanin and tannin end‐products. Contrary to expectation, effects on anthocyanin accumulation were subtle and restricted to the leaf midrib, leaf base and petiole tissues. However, the accumulation of condensed tannin polymers was dramatically enhanced in the leaf blade and this increase was accompanied by a 50‐fold increase in the number of tannin‐containing cells in this tissue. A detailed analysis of selected lines indicated that this transactivational phenotype correlated with high steady‐state transcript levels of the introduced transgene and the introduction of a single copy of the CaMV35S‐Sn construct into these clonal genotypes. While the levels of condensed tannins in leaves were increased by up to 1% of the dry weight, other major secondary end‐products (flavonols, lignins and inducible phytoalexins) were unaltered in transactivated lines. These results give an initial insight into the developmental and higher‐order regulation of polyphenolic metabolism in Lotus and other higher plant species. Key words: Anthocyanins, condensed tannins, Lotus, metabolic engineering, transactivation. Received 26 June 2002; Accepted 19 August 2002 Introduction Secondary products perform a wide range of functions essential to the survival of higher plants (Burbulis and Winkel‐Shirley, 1999). Of these secondary product pathways, those leading to flavonoid and anthocyanin end‐products have been the most extensively studied due to their importance in flower pigmentation and the excellent range of genetic resources in petunia, Antirrhinum, maize, and Arabidopsis (Holton and Cornish, 1995). In maize, two major classes of transcription factors have been described. The R/B family, which shows sequence homology to the basic helix‐loop‐helix (bHLH) DNA binding proteins found in animal MYC protooncogenes (Ludwig et al., 1989) and the C1/P family, which encodes proteins with similarity to the DNA binding domain of the mammalian MYB oncogene products (Paz‐Ares et al., 1987). By contrast, there is very little information on the molecular mechanisms that control the expression of the closely related biosynthetic pathway to condensed tannins (syn. proanthocyanidins) (Robbins and Morris, 1999). This lack of knowledge regarding condensed tannins is unfortunate as these are one of the world’s major biopolymers and they perform important functions in plant–insect interactions (Muir et al., 1999), in plant–herbivore interactions (Furstenburg and van Hoven, 1994), as well as having a range of biotechnological applications (Robbins et al., 1999). A major interest of scientists studying forage legumes is to investigate whether this pathway can be introduced into the foliar tissues of white clover and lucerne with the aim of producing bloat‐safe field crops with enhanced nitrogen efficiency when grazed by ruminant livestock (Aerts et al., 1999). Previous work has shown that the tissue‐specific accumulation of anthocyanin pigments in maize and Antirrhinum is controlled by a range of anthocyanin regulatory genes that are responsible for the co‐ordinate induction of structural genes in this metabolic sequence (Holton and Cornish, 1995). Also of note is the observation that the ectopic expression of bHLH transcription factors has been reported to enhance anthocyanin levels in non‐tannin‐containing species such as tomato, tobacco and petunia (Mooney et al., 1995; Bradley et al., 1998). In maize cell cultures, the expression of bHLH and MYB anthocyanin transactivators in combination has been reported to result in the accumulation of anthocyanin end‐products contained within anthocyanoplasts (Grotewold et al., 1998). In Arabidopsis the expression of R, a maize anthocyanin MYC, increased levels of anthocyanin while the expression of R and C1 resulted in anthocyanin accumulation in novel tissue locations such as root, petal and stamen (Lloyd et al., 1992). Recently, two genes have been identified in Arabidopsis, TT8 (encoding a bHLH domain transcription factor) and TT2 (encoding a R2R3 MYB domain protein) and these have been identified as key determinants for the proanthocyanidin accumulation in developing seeds (Nesi et al., 2000, 2001). In this paper, the effect of the over‐expression of Sn, a maize anthocyanin bHLH gene, has been analysed in Lotus corniculatus which is a metabolic model for condensed tannin biosynthesis. This model legume accumulates a wide range of flavonoid end‐products as outlined in Fig. 1. Most notably, this species biosynthesizes condensed tannins and clonal transformable genotypes have been produced which permit the analysis of transgenic interventions in a uniform genetic background (Robbins et al., 1999). Previous work has shown that this pathway is readily manipulable using structural genes of the pathway (Carron et al., 1994; Colliver et al., 1997) and that the over‐expression of pathway genes can enhance condensed tannin content and also modify tannin polymer structure (Bavage et al., 1997). In this paper, it is demonstrated that Sn, which encodes a heterologous bHLH involved in the control of anthocyanin biosynthesis, can transactivate both anthocyanin and condensed tannin pathways in Lotus. Increases in end‐products occur in a tissue‐specific manner and strongly suggest that there are shared regulatory mechanisms, which control the biosynthesis of anthocyanin and condensed tannin end‐products. Materials and methods Production and growth of experimental material Hairy root transformation was performed in three well characterized genotypes (S33, S41, S50) belonging to the L. corniculatus cultivar Leo (Carron et al., 1994). The plasmid 121.Sn (Damiani et al., 1998), derived from pBI121.1 where the gus gene was replaced with the full length cDNA of Sn (Tonelli et al., 1991), was introduced through triparental mating into the wild‐type A. rhizogenes strain 1855. Axenic plants were inoculated in the stem by puncturing with a needle previously dipped in the bacterial culture. Independently derived hairy roots were detached and cultured in the dark for 2 weeks on a hormone‐free medium consisting of Gamborg’s B5 salts (Sigma) supplemented with sucrose (2%), agar (1.5%), kanamycin (50 mg l–1) and carbenicillin (1 g l–1) (Gamborg et al., 1968). Shoot formation from each hairy root started when plates were transferred to light (50 µM m–2 s–1 light intensity). Following 2 months of in vitro culture in the light, plants were moved to soil in a mist chamber to prevent plant desiccation and then grown on for further analysis. Typically, experimental plant material comprised mature Lotus plants grown in triplicate at 18/15 °C day/night with a day length of 14 h and a light level of 750 µmol m–2 s–1. Screening for condensed tannin phenotype Detached leaves from mature plants were stained for condensed tannins using 4‐dimethylaminocinnamaldehyde (DMACA) according to the method of Li et al. (1996). This reagent produces a blue coloration on reaction with condensed tannins under acidified conditions and from the intensity of staining each plant was scored on a 0–6 scale. DNA analysis and measurements of transcript levels For Southern blotting DNA was extracted according to Cluster et al. (1996), restricted, electrophoresed and blotted to Amersham Hybond N+ filter according to standard procedures. Filters were then probed with the 32P labelled XbaI fragment of the Sn cDNA. For RNA analysis, total RNA was isolated from leaves of transgenic lines 50/11, 50/10, 50/9, and 50/6 and from control plant S50 according to Chang et al. (1993). One µg of PolyA+ RNA, isolated from total RNA using the ‘Dynabeads mRNA purification Kit’ (Dynal), was hybridized with Sn and Rubisco SSU probes as described by Damiani et al. (1999). The DFR probe was a 610 bp PCR fragment amplified from the plasmid pGMcDFR (GenBank Accession No. AY117027) containing a full length DFR cDNA of L. corniculatus, with the primers DFRFW1/DFRGSP2 (DFRFW1: 5′‐ CTAACATGAAGAAGGTGAAG‐3′; DFRGSP2: 5′‐TGGCATTG TCGGCATTAGAAAGG‐3′). RT‐PCR amplification from DNA‐free total RNA was carried out, run on agarose gel, blotted and hybridized basically as described by Damiani et al. (1999). As control for the presence of DNA contaminating the RNA preparations, RNA samples were processed both in the presence and/or in the absence of MMLV reverse transcriptase (Gibco BRL) and submitted to PCR amplification. The primer pairs used both for RT‐PCR analysis and probe preparation for Southern of the RT‐PCR blots were: Sn1/Sn2 (Sn1: 5′‐TCT GGCTGTGCAACGCGCACC‐3′ and Sn2: 5′‐ CTTCTCTCGTCG CTTTCGCTC‐3′) to amplify a 810 bp Sn fragmentand EF1F/EF1B (EF1F: 5′‐ATTGTGGTCATTGGCCACGT‐3′; EF1B: 5′‐ CCAA TCTTGTACACATCCTG‐3′) to amplify a 710 bp fragment of EF‐1α gene. The primer pair used for RT‐PCR analysis of DFR mRNA levels was DFRFW1/DFRGSP2, which can amplify all known L. corniculatus DFR cDNA sequences. Cellular location of anthocyanin‐containing and tannin‐containing cells in transgenics and cell counts Cell counts were performed on trifoliate leaves selected at random. Anthocyanin‐containing cells were red in colour and visible with the naked eye. Cells containing condensed tannins were visible after staining with 4‐dimethylaminocinnamaldehyde (DMACA). Shoot tissues were harvested and decolourized in ethanol overnight. After discarding the ethanol, samples were stained using 0.3% (w/v) DMACA in 6 M HCl for 1 h. After four changes of distilled water, samples were then analysed and blue colour indicated the presence of condensed tannin polymers. For quantification, mature leaves were mounted on a slide and viewed using an Olympus BH light microscope (10× magnification). Quantification of phenolic and flavonoid end‐products Condensed tannin levels were determined in freeze‐dried samples as described by Terrill et al. (1992). Flavonols were quantified essentially as described by Robbins et al. (1998) by summing peaks after performing diode array high performance liquid chromatography using a Waters 996 machine. Thioglycolic acid (TGA) lignin was determined using the method of Whitmore (1978). Inducible phytoalexins were analysed after elicitation with 10mM glutathione as described by Robbins et al. (1995). Results Initial screening of Lotus corniculatus for condensed tannin phenotype Standard transformation procedures were employed to introduce a CaMV35S‐Sn gene construct into Lotus corniculatus. Clonal genotypes were used in order to assess effects in a uniform genetic background. Regenerated plants were analysed and scored for condensed tannin content as shown in Fig. 2. In genotype S50 lines were noted with reduced condensed tannin content and these appear to be similar to lines described by Damiani et al. (1999) where the introduction of Sn apparently reduced end‐product accumulation in leaves. More notable in this study were a number of co‐transformed lines in the S50 background with markedly higher levels of condensed tannin than found in control lines (Fig. 2a). Similar effects were noted in the S33 background with a number of lines showing enhanced tannin content relative to controls (Fig. 2b). Interestingly when a high tannin genotype (S41) was transformed, no lines were noted with levels of tannins higher than controls (Fig. 2c). A limited number of lines were subjected to RT‐PCR analysis in order to determine whether there was any obvious correlation between transgene expression and derived chemical phenotype. In the S33 background high levels of steady‐state transcript were noted for lines 33/19 and 33/1. Curiously one other enhanced line, 33/6, showed no obvious expression of transgene, however, evidence suggests that this was due to silencing of transgene between initial screening for condensed tannin phenotype and sampling for molecular analysis by RT‐PCR (data not shown). In the S50 transgenics Sn expression was clearly detected in lines with high tannin scores, 50/23, 50/11, 50/10; but not in lines with control or reduced levels, 50/14, 50/13, 50/9, 50/8, 50/6, and 50/5. Line 50/1 was noted to silence between initial analysis and RT‐PCR analysis (data not shown). Only two S41 lines were analysed for transgene expression, 41/6 and 41/17 and neither showed detectable transgene expression. However, analysis of CaMV‐gus lines in this genotype confirmed expression of the CaMV promotor in the S41 genetic background. Detailed molecular analysis of enhanced and suppressed lines in S50 background In view of phenotypes noted in the S50 genotype, a subset of enhanced and suppressed lines was selected for a more focused analysis. In particular, two enhanced lines (50/10 and 50/11) were selected and compared with two suppressed lines (50/6 and 50/9) together with a control line. Southern blot analysis (Fig. 3a) showed that lines contained between one and six transgene copies and this is typical for co‐transformation experiments in Lotus (Carron et al., 1994). One aspect of interest is that, while suppressed lines contained multiple copies of CaMV‐Sn, the two enhanced lines had single transgene copies within their genomic DNA complement. In order to clarify the molecular basis of the phenotype, Northern analysis was performed (data not shown) and Sn transcript was clearly detectable in PolyA+ RNA extracted from leaf tissues of 50/10 and 50/11. Probing of the same blots with a probe encoding dihydroflavonol reductase (DFR), a gene common to anthocyanin and condensed tannin pathways, showed enhanced steady‐state levels of this gene which would be consistent with up‐regulation conferred by the introduced transcriptional activator. RT‐PCR analysis was also carried out on control and selected lines for Sn, DFR and a housekeeping gene (EF‐1α), for normalization, and confirmed the results from Northern analysis, i.e. expression of the Sn transgene in 50/10 and 50/11 combined with enhanced levels of DFR mRNA in these two lines when compared with control and suppressed lines (Fig. 3b). Analysis of anthocyanin and condensed tannin cell types in S50 background Careful observation of lines grown under control conditions confirmed a subtle anthocyanin phenotype in Sn‐expressing lines in the S50 genetic background, 50/10 and 50/11. Similar general phenotypes were also noted in the S33 background. When grown in tissue culture under high light conditions, control Lotus plants did not accumulate visible quantities of anthocyanin pigment (Fig. 4b). However, lines which had been identified with enhanced levels of condensed tannin exhibited a characteristic anthocyanin pigmentation which was particularly marked in juvenile leaf and stem tissues (Fig. 4a and insert). Close examination of mature control plants showed some anthocyanin‐containing cells in leaf petiole tissues adjacent to leaf base tissues (Fig. 4d). When line 50/10 was analysed, intense pigmentation was noted in leaf bases, at the end of the petiole and along the midrib of the leaf; no alteration in anthocyanin accumulation was noted in other parts of the plant (Fig. 4c). By contrast, line 50/9 which had been identified with reduced tannin content showed a marked reduction in anthocyanin cells in the petiole (Fig. 4f) relative to a control line (Fig. 4e). Anthocyanins were induced in a tissue‐specific manner and cross‐sections of petiole (Fig. 4i), leaf (Fig. 4g) and leaf base (Fig. 4h) showed that anthocyanin‐containing cells were restricted to the subepidermal cell layer. In order to examine the effect of Sn upon the distribution of condensed tannin cells in leaves, one Sn‐suppressed line (50/9) was compared with an Sn‐enhanced line (50/10) and also an S50 control. The presence of cells containing condensed tannins was determined using DMACA, a reagent which specifically stains cell types that accumulate condensed tannin polymers (Li et al., 1996). Under normal conditions Lotus leaves can contain tannin cells in three positional locations: adjacent to vascular tissues, distributed through the palisade mesophyll and in a matrix formation in the spongy mesophyll (Robbins and Morris, 1999). These observations were confirmed in this study. Leaves from a control plant contained condensed tannin cells in leaf mesophyll tissues (Fig. 5b, e) and analysis of cell distributions at the base of the trifoliate leaf indicated that tannin‐containing cells were restricted to the vascular mesophyll (Fig. 5h, k). By contrast, control leaves contained very few tannin‐containing cells at the leaf tip (Fig. 5n, q) while petioles had tannin‐containing cells adjacent to central vascular tissue (Fig. 5t). In Sn‐suppressed leaves numbers of tannin‐containing cells in leaves were reduced (Fig. 5c, f, i, l). No tannin‐containing cells were noted at the leaf tip and some reduction in the numbers of tannin‐containing cells was evident in the petiole (Fig. 5u). By contrast, Sn‐enhanced lines had higher numbers of tannin cells than suppressed or control lines and had a modified developmental expression profile for these cell types. Increases in numbers of tannin cells were clear after initial staining of trifoliate leaves (Fig. 5a, d) and more detailed analysis indicated that expression of Sn had induced the production of tannin‐containing cells both in palisade and spongy mesophyll cell layers (Fig. 5g, j). Sn‐enhanced lines also contained tannins at the leaf tip (Fig. 5m), in contrast to control lines and this was accompanied by the appearance of tannin‐containing cells, which were predominantly located in the spongy mesophyll (Fig. 5p). Analysis of the tissue‐specific modulation of anthocyanin and tannin‐containing cells in Sn transgenics In view of changes in anthocyanin and tannin cell numbers, cell counts were performed in a range of tissue locations within the trifoliate leaf and these data are presented in Fig. 6. In the enhanced line selected for study, 50/10, increases in anthocyanin cell counts were recorded in petiole and leaf base and anthocyanin cells were noted in layers directly above leaf midrib and this confirms initial observations (Fig. 4d). However, no anthocyanin containing cells were found in the leaf blade, i.e. in areas outside the leaf midrib. By contrast, tannin‐containing cells were increased in number relative to controls in the petiole, leaf base and in tannin‐containing cells directly adjacent to vascular tissues. More notably, however, tannin‐containing cells were present in the lamina of the leaf and a dramatic increase in the numbers of these cells was noted in 50/10. By contrast, lower numbers of tannin‐containing cells were found in 50/9, the Sn‐suppressed line included in this analysis. Effects of Sn transgene upon other flavonoid and phenolic end‐products in Lotus leaves This study has shown that the introduction of Sn modifies the accumulation of condensed tannins and anthocyanin end‐products. In order to analyse the effects on other secondary products, both phenolic and flavonoid levels were determined in leaves from three control lines, a phenotype negative line (50/13) together with 50/10 and 50/9. These data are displayed in Fig. 7. In control lines, the major flavonoid end‐products were condensed tannins (0.98 mg g–1 DW) and flavonols, primarily kaempferol glycosides (82±3 mg g–1 DW). No isoflavonoid end‐products were detectable in leaf tissues, but after elicitation with glutathione measurable levels of vestitol could be determined (515±178 µg g–1 FW) together with trace amounts of sativan (typically 20±16 µg g–1 FW). Free and wall‐bound phenolics were present at levels below detectability, but levels of thioglycolic acid (TGA) lignin were measured at 5.5±0.5 mg g–1 DW in control lines. Line 50/13 had similar levels of flavonols, lignins and inducible isoflavans to control lines. 50/10 contained 10.3±1.4 mg g–1 DW condensed tannin which corresponds to an increase of nearly 1% of the dry weight of this particular end‐product. In other experiments under environmental conditions where plants accumulate higher levels of tannins, leaves from 50/10 accumulated over 2% DW more condensed tannin than corresponding controls. In this up‐regulated line, levels of lignin and inducible phytoalexins were unaltered (data not shown). Mean flavonol levels in 50/10 were reduced relative to controls, 72±9 mg g–1 DW as compared with 82±3 mg g–1 DW. However, this difference was not significant and there is no other evidence that increases in condensed tannins are matched by a corresponding decrease in flavonols in other Sn up‐regulated lines (data not shown). Discussion Previous studies have implicated bHLH transcription factors as anthocyanin pathway regulators in a number of higher plant systems. The maize R gene family (R, B, Lc, and Sn) has been extensively studied and ectopic expression of Lc results in enhanced pigmentation in tomato and petunia (Goldsbrough et al., 1996; Bradley et al., 1998), while expression of R increases anthocyanin production in tobacco and Arabidopsis (Lloyd et al., 1992). By contrast, expression of Lc resulted in no visible phenotypic alteration in vegetative or floral pigmentation in pelargonium or lisianthus (Bradley et al., 1999). Expression of B‐Peru in white clover has been reported to induce a novel pattern of anthocyanin accumulation in leaf tissues (de Majnik et al., 2000). Sn hairy roots produced from a range of dicotyledonous plant species showed patterns of red pigmentation dependent upon species, genotype and transformation event (Damiani et al., 1998). The phenotypes resulting from the expression of Sn in Lotus as outlined in this paper are in general agreement with this class of effects upon anthocyanin pathways, however, in this system, alterations in pigmentation were restricted to petiole, leaf base and, in some cases, the leaf midrib. One interpretation of the data in this paper is that Lotus contains an orthologue to maize R transcription factors and that the ectopic expression of Sn can functionally complement the expression of this Lotus orthologue. An attempt to clone an endogenous Lotus bHLH gene was made, but so far the only one cloned was found to be homologous to PG1, a ubiquitous bHLH protein not functionally related to anthocyanin regulation (Kawagoe and Murai, 1996). However, preliminary quantitative analyses have shown that the cloned bHLH gene is down‐regulated in suppressed lines (F Paolocci, unpublished results). In the absence of cloned Sn orthologues, and appropriate mutants for complementation studies, no direct conclusions regarding the effects of Sn upon anthocyanin biosynthesis in Lotus can be drawn. Evidence has also been provided that both the introduction of Sn and expression of Sn can modulate condensed tannin biosynthesis. In two of the genotypes under study, lines were noted with enhanced levels of condensed tannin. Analysis of two lines, 50/10 and 50/11, indicated that these were single copy transformation events and that leaves from these two lines contained detectable levels of the Sn transgene. Similarly, the analysis of two lines, 50/6 and 50/9, that are suppressed for condensed tannin accumulation, showed that there had been complex transformational events, which are associated with transgene suppression. Levels of DFR mRNA in leaf tissues were enhanced in these two lines and in view of the limited anthocyanin phenotype in leaves, it can be deduced that a major proportion of this Sn‐mediated induction of DFR transcript is related to the condensed tannin pathway. Increases in tannin content were accompanied by increases in tannin cell number and these effects were particularly dramatic in the leaf lamina, a tissue that does not biosynthesize anthocyanin end‐products. Increases in cell numbers occurred in a cell lineage, which normally contains tannin cells, i.e. the vascular mesophyll. However, tannin cells were also found in cell layers which had no tannin cells in control lines, i.e. in the leaf spongy and palisade mesophyll. It is noted, however, that glasshouse‐grown Lotus plants of the high tannin S41 genotype contain tannin cells in all three mesophyll layers, i.e. vascular, palisade and spongy (data not shown). Therefore, in S50 and S33, cells in spongy and palisade mesophyll may be competent for the biosynthesis of condensed tannins, but have no expression of an Sn orthologue or functionally equivalent genes in these cell layers. Finally, taking the anthocyanin and tannin data together, there is evidence that, in these experiments, Sn has modulated the anthocyanin and condensed tannin pathways. Anthocyanin cell numbers are enhanced in the subepidermal cell layer while alterations in tannin‐containing cells occur in leaf mesophyll layers. This co‐ordinate induction of independent biosynthetic pathways by a bHLH class gene is surprising, but, nevertheless, gives an interesting insight into the higher order regulation of condensed tannin biosynthesis. Until recently, most anthocyanin bHLH genes have been assumed to have fairly well described biochemical functions. However, it was noted that the TT8 locus in Arabidopsis encodes a bHLH domain protein, which has been reported to regulate the biosynthesis of both anthocyanins and proanthocyanidins (syn. condensed tannins) in siliques (Nesi et al., 2000). This dual metabolic function of TT8 is similar to the effects reported in this paper where Sn can up‐regulate tannin and anthocyanins but does not appear to alter lignin, isoflavonoid or flavonoid pathways. In conclusion, the use of Sn and related anthocyanin bHLH transcription factors may give rise to useful approaches for the modification of levels of condensed tannins in crop species. Additionally, these transgenics may be a valuable resource for cloning the terminal steps of the condensed tannin pathway and also for cloning genes involved in the design of cells that biosynthesize anthocyanin and condensed tannin end‐products. Acknowledgements IGER is grant‐funded by BBSRC, JWH was funded by a BBSRC‐RASP studentship ref 4648. We would like to thank other members of the laboratory for helpful assistance; Rolando Barahona, Gordon Allison, Teri Davies, and others. Thanks also to Dr Helen Ougham and Dr Joe Gallagher for constructive comments on this manuscript. IRMGPF is funded by Consiglio Nazionale della Richerche. Open in new tabDownload slide Fig. 1. Schematic illustration of the major flavonoid and phenolic end‐products found in Lotus. Isoflavan phytoalexins are induced as a result of pathogen attack while other end‐products, for example, lignins, anthocyanins and condensed tannins accumulate in specific cell types within the plant. Open in new tabDownload slide Fig. 2. Screening of CaMV‐Sn transgenics for alterations in condensed tannin content. (a) S50 genotype; (b) S33 genotype; (c) S41 genotype. Plants were stained for numbers of condensed tannin‐containing cells using dimethylaminocinnamaldehyde and scored as described in experimental procedures. Control=CaMV‐gus transformant, Recip=untransformed recipient genotype. Values are from triplicate determinations and bars indicate standard error of mean values. Open in new tabDownload slide Fig. 3. Molecular analysis of selected CaMV‐Sn lines with enhanced and suppressed levels of condensed tannins. (a) 10 µg genomic DNA restricted with HindIII and probed with Sn fragment amplified from Sn cDNA. (b) Reverse transcriptase PCR analysis of selected CaMV‐Sn lines. Amplification and hybridization using homologous sequences performed as described in experimental procedures. Open in new tabDownload slide Fig. 4. Localization of anthocyanin‐containing cells in transgenic Lotus plants. (a) line 50/10 plantlet from tissue culture; insert, juvenile tissues from the same plant. (b) Control S50 plantlet grown in tissue culture. (c) Schematic diagram of a trifoliate Lotus corniculatus leaf; 1, junction of leaf to petiole via leaf base; 2, leaf tip; 3, petiole; 4, base of leaflet; wide bar, 1 cm. (d) Leaf–petiole interface, line 50/10. (e) Leaf–petiole interface, control line. (f) Leaf–petiole interface, line 50/9. (g) Section through junction of leaf base to petiole, line 50/10. (h) Transverse section through leaf base, line 50/10. (i) Transverse section through petiole, line 50/10. Open in new tabDownload slide Fig. 5. DMACA localization of cells containing condensed tannins in selected Lotus lines harbouring the CaMV‐Sn construct. A comparison of an enhanced line (50/10), a control and a suppressed line (50/9). (a) Trifoliate leaf from line 50/10. (b) Control line. (c) Line 50/9. (d) Condensed tannin cells as viewed from the adaxial side, line 50/10. (e) Control line. (f) Line 50/9. (g, j) Transverse section of leaf (position 4), 50/10. (h, k) Control. (i ,l) 50/9. (m) Leaf tip (position 2), adaxial view, 50/10. (n) Control. (o) 50/9; (p) Transverse section of leaf tip, 50/10. (q) Control. (r) 50/9. (s) Transverse section of petiole (position 3), 50/10. (t) Control. (u) 50/9. Open in new tabDownload slide Fig. 6. Cell counts in control and selected CaMV‐Sn Lotus lines. (a) Numbers of anthocyanin‐containing cells in leaf and petiole. (b) Numbers of condensed tannin‐containing cells in leaf and petiole. Open in new tabDownload slide Fig. 7. Levels of major flavonoid and phenolic end‐products in S50 control and S50‐Sn transformed lines. (a) Condensed tannins; (b) flavonols; (c) lignin; (d) isoflavan phytoalexin (values after elicitation). Light bars=vestitol, dark bars=sativan. Columns represent mean values and the control column corresponds to the mean of triplicate control plants. Values displayed are from triplicate determinations and bars indicate standard error of mean values. References AertsRJ, Barry TN, McNabb WC. 1999 . Polyphenols and agriculture; beneficial effects of proanthocyanidins in forages. Agriculture, Ecosystems and Environment 75, 1 –12. BavageAD, Davies IG, Robbins MP, Morris P. 1997 . Expression of an Antirhinnum dihydroflavonol reductase gene results in changes in condensed tannin structure and accumulation in root cultures of Lotus corniculatus (bird’s foot trefoil). Plant Molecular Biology 35, 443 –458. BradleyJM, Davies KM, Deroles SC, Bloor SJ, Lewis DH. 1998 . The maize Lc regulatory gene up‐regulates the flavonoid biosynthetic pathway of petunia. The Plant Journal 13, 381 –392. BradleyJM, Deroles SC, Boase MR, Bloor S, Swinny E, Davies KM. 1999 . Variation in the ability of the maize Lc regulatory gene to up‐regulate flavonoid biosynthesis in heterologous systems. Plant Science 140, 31 –39. BurbulisIE, Winkel‐Shirley B. 1999 . Interactions among enzymes of the Arabidopsis flavonoid biosynthetic pathway. Proceedings of the National Academy of Sciences, USA 96, 12929 –12934. CarronTR, Robbins MP, Morris P. 1994 . Genetic modification of condensed tannin biosynthesis in Lotus corniculatus. 1. Heterologous antisense dihydroflavonol reductase down‐regulates tannin accumulation in ‘hairy root’ cultures. Theoretical and Applied Genetics 87, 1006 –1015. ChangS, Puryear J, Cairney J. 1993 . A simple and efficient method for isolating RNA from pine trees. Plant Molecular Biology Reports 11, 113 –116. ClusterPD, Calderini O, Pupilli F, Crea F, Damiani F, Arcioni S. 1996 . The fate of ribosomal genes in three interspecific somatic hybrids of Medicago sativa: three different outcomes including the rapid amplification of new spacer‐length variants. Theoretical and Applied Genetics 93, 801 –808. ColliverSP, Morris P, Robbins MP. 1997 . Differential modification of flavonoid and isoflavonoid biosynthesis with an antisense chalcone synthase construct in transgenic Lotus corniculatus. Plant Molecular Biology 35, 509 –522. DamianiF, Paolocci F, Consonni G, Crea F, Tonelli C, Arcioni S. 1998 . A maize anthocyanin transactivator induces pigmentation in hairy roots of dicotyledonous species. Plant Cell Reports 17, 339 –344. DamianiF, Paolocci F, Cluster PD, Arcioni S, Tanner GJ, Joseph RJ, Li YG, de Majnik J, Larkin PJ. 1999 . The maize transcription factor Sn alters proanthocyanidin synthesis in transgenic Lotus corniculatus plants. Australian Journal of Plant Physiology 26, 159 –169. de MajnikJ, Weinman JJ, Djordjevic MA, Rolfe BG, Tanner RJ, Joseph RG, Larkin PJ. 2000 . Anthocyanin regulatory gene expression in transgenic white clover can result in an altered pattern of pigmentation. Australian Journal of Plant Physiology 27, 659 –667. FurstenburgD, van Hoven W. 1994 . Condensed tannin as anti‐defoliate agent against browsing by giraffe (Giraffa camelopardalis) in the Kruger National Park. Comparative Biochemistry and Physiology107A, 425 –431. GamborgOL, Miller RA, Ojna K. 1968 . Nutrient requirements of suspension cultures of soybean root cells. Experimental Cell Research 50, 151 –158. GoldsbroughAP, Tong Y and Yoder JI. 1996 . Lc as a non‐destructive visual reporter and transposition excision marker gene for tomato. The Plant Journal 9, 927 –933. GrotewoldE, Chamberlin P, Snook M, Siame B, Butler L, Swenson J, Maddock S, StClair G, Bowen B. 1998 . Engineering secondary metabolism in maize cells by ectopic expression of transcription factors. The Plant Cell 10, 721 –740. HoltonTA, Cornish C. 1995 . Genetics and biochemistry of anthocyanin biosynthesis. The Plant Cell 7, 1071 –1083. KawagoeY, Murai N. 1996 . A novel basic region/helix‐loop‐helix protein binds to a G‐box motif CACGTG of the bean seed storage protein β‐phaseolin gene. Plant Science 16, 47 –57. LiYG, Tanner GJ, Larkin PJ. 1996 . The DMACA‐HCl protocol and the threshold proanthocyanidin content for bloat safety in forage legumes. Journal of the Science of Food and Agriculture 70, 89 –101. LloydAM, Walbot V, Davis RW. 1992 . Arabidopsis and Nicotiana anthocyanin production activated by maize regulators R and C1. Science 258, 1773 –1775. LudwigSR, Habera LF, Dellaporta SL, Wessler SR. 1989 . Lc, a member of the maize R gene family responsible for tissue specific anthocyanin production, encodes a protein similar to transcription activators and contains the myc‐homology region. Proceedings of the National Academy of the Sciences, USA 86, 7092 –7096. MooneyM, Desnos T, Harrison K, Jones J, Carpenter R, Coen E. 1995 . Altered regulation of tomato and tobacco pigmentation genes caused by the delila gene of Antirrhinum. The Plant Journal 7, 333 –339. MuirAD, Gruber MY, Hinks CF, Lees GJ, Onyilagha J, Soroka J, Erlandson M. 1999 . Effect of condensed tannins in the diets of major crop insects. In: Gross GG, Hemingway RW, Yoshida T, eds. Plant polyphenols. 2. Chemistry, biology, pharmacology, ecology. New York: Kluwer, 867 –881. NesiN, Debeaujon I, Jond C, Pelletier G, Caboche M, Lepiniec L. 2000 . The TT8 gene encodes a basis helix‐loop‐helix domain protein required for expression of DFR and BAN genes in Arabidopsis siliques. The Plant Cell 12, 1863 –1878. NesiN, Jond C, Debeaujon I, Caboche M, Lepiniec L. 2001 . The Arabidopsis TT2 gene encodes an R2R3 MYB domain protein that acts as a key determinant for proanthocyanidin accumulation in developing seed. The Plant Cell 13, 2099 –2114. Paz‐AresJ, Ghosal D, Wienand U, Peterson PA, Saedler H. 1987 . The regulatory C1 locus of Zea mays encodes a protein with homology to myb proto‐oncogenes products and with structural similarities to transcription activators. EMBO Journal 6, 3553 –3558. RobbinsMP, Bavage AD, Strudwicke C, Morris P. 1998 . Genetic manipulation of condensed tannins in higher plants. 2. Analysis of birdsfoot trefoil plants harboring antisense dihydroflavonol reductase constructs. Plant Physiology 116, 1133 –1144. RobbinsMP, Bavage AD, Morris P. 1999 . Designer tannins: using genetic engineering to modify levels and structures of condensed tannins in Lotus corniculatus. In: Gross GG, Hemingway RW, Yoshida T, eds. Plant polyphenols. 2. Chemistry, biology, pharmacology, ecology. New York; Kluwer, 301 –314. RobbinsMP, Morris P. 1999 . Metabolic engineering of condensed tannins and other phennolic pathways in forage and fodder crops. In: Verpoorte R, Alfermann AW, eds. Metabolic engineering of plant secondary metabolism. The Netherlands: Kluwer, 165 –177. RobbinsMP, Thomas B, Morris P. 1995 . Phenylpropanoid defence responses in transgenic Lotus corniculatus 2. Modelling plant defence responses in transgenic root cultures using thiol and carbohydrate elicitors. Journal of Experimental Botany 46, 513 –524. TerrillTH, Douglas GB, Foote AG, Purchas RW, Wilson GF, Barry TN. 1992 . Determination of extractable and bound condensed tannin concentrations in forage plants, protein concentrate meals and cereal grains. Journal of the Science of Food and Agriculture 58, 321 –329. TonelliC, Consonni G, Dolfini SF, Dellaporta SL, Viotti A, Gavazzi G. 1991 . Genetic and molecular analysis of Sn, a light‐inducible tissue‐specific regulatory gene in maize. Molecular and General Genetics 225, 401 –410. WhitmoreFW. 1978 . Lignin–carbohydrate complex formed in isolated cell walls of callus. Phytochemistry 17, 421 –425. © 2003

De Gara, Laura; de Pinto, Maria C.; Moliterni, Vita M. C.; D’Egidio, Maria G.

doi: 10.1093/jxb/erg021pmid: 12493852

Abstract Metabolic changes during the development and maturation of Triticum durum Desf. (L.) kernels were studied, with particular emphasis on changes in the redox state of ascorbate and glutathione, as well as in the activities of the enzymes responsible for the recycling of their oxidized forms (ascorbic free radical reductase, EC 1.6.5.4; dehydroascorbate reductase, EC 1.8.5.1; glutathione reductase, EC 1.6.4.2) and for detoxification or utilization of hydrogen peroxide (ascorbate peroxidase, EC 1.11.1.11; catalase, EC 1.11.1.6). In parallel with this analysis, the production and storage of reserve compounds was studied, in particular, soluble carbohydrates (mono‐ di‐saccharides and fructans) and the transition from sulphydryl groups to disulphide bridges into proteins. The results indicate that both the activities of the ascorbate and glutathione redox enzymes and that of catalase are high before the start of drying maturation, after which they decrease. Moreover, analysis of the redox state of ascorbate and glutathione pairs and the sulphydryl to disulphide transition into proteins suggests that these three parameters are tightly related during kernel maturation, thus confirming the involvement of the two redox pairs in protein maturation as well as in protection against reactive oxygen species. The physiological implications of the changes in cellular redox state and in soluble carbohydrates for the acquisition of desiccation tolerance and reaching the resting phase in orthodox seeds are also discussed. Key words: Ascorbate, ascorbate peroxidase, catalase, glutathione, kernel maturation, Triticum. Received 9 April 2002; Accepted 27 August 2002 Introduction Seed development and maturation is a highly orchestrated multi‐step process during which embryos are formed and supplied with the carbohydrates, proteins and lipids needed for the subsequent germination. During seed development the increase in size is due to cell division and expansion, followed by a progressive accumulation of storage compounds. In many plants, the last step of seed maturation is dehydration, a process allowing seeds to delay germination until there are suitable environmental conditions (Steeves, 1983). Seeds that reach maturity in a highly dehydrated state, orthodox seeds, survive for long periods and can be stored for years under cold dry conditions without losing viability. By contrast, the so‐called recalcitrant seeds do not undergo drying maturation, are very sensitive to desiccation, and lose viability in a very short time (Roberts, 1973). The desiccation tolerance acquired by orthodox seeds may be analogous in some aspects to the defence mechanisms carried out in vegetative tissues by tolerant species subjected to drying. In fact, the increase in abscisic acid (Mansfield and McAinsh, 1995; Rock and Quatrano, 1995) and the accumulation of specific polypeptides (Skriver and Mundy, 1990; Bradford and Chandler, 1992) are examples of processes that are common to both orthodox seed development and vegetative tissue responses to water depletion. Among orthodox seeds, cereal kernels have been intensively investigated because of their economic importance. In particular, data are available on their capability of accumulating storage compounds and acquiring desiccation tolerance (Black et al., 1999; Gebbing and Schnyder, 1999). The aim of this paper was to analyse some parameters characterizing the development of Triticum durum Desf. (L.) kernels, including the storage of soluble and insoluble carbohydrates and the accumulation and maturation of proteins. Particular attention has been paid to ascorbate (ASC) and glutathione (GSH) metabolisms. These two redox pairs are involved in plant developmental processes as well as in the protection against reactive oxygen species (ROS), the production of which is well known to increase strongly in vegetative tissues subjected to water stress (Sgherri and Navarri Izzo, 1995; Noctor and Foyer, 1998; Asada, 1999; De Gara and Tommasi, 1999). ASC and GSH are responsive to ROS scavenging both directly and by means of enzymatic reactions. Among the latter, much attention has been given to the so‐called ascorbate–glutathione cycle, a set of reactions in which ASC and GSH undergoing oxidation are continuously regenerated. The first enzyme of the cycle is ascorbate peroxidase (APX), that removes hydrogen peroxide by oxidizing ASC to the ascorbic free radical (AFR). The AFR formed is reduced to ASC by an NAD(P)H‐dependent AFR‐reductase (AFRR), or it disproportionates to ASC and dehydroascorbate (DHA). DHA, the final product of ASC oxidation, is then reduced to ASC by DHA‐reductase (DHAR), an enzyme using GSH as reductant. Finally, GSH is regenerated from the glutathione disulphide (GSSG) produced in the latter reaction by GSSG‐reductase (GR) by using the reducing power of NADPH. Changes in the ASC–GSH cycle enzymes occur in response to several kinds of stresses. Moreover, genetically modified plants over‐expressing one of these enzymes are more resistant to abiotic stresses, whereas plants under‐expressing one of them become more sensitive (Noctor and Foyer, 1998; Smirnoff and Wheeler, 2000). Changes in the ASC–GSH cycle also occur in developmental processes such as seed germination (Tommasi et al., 2001) or leaf senescence (Borraccino et al., 1994). In addition, a co‐ordinated increase in their activities characterize the exponential growing phase in cultured tobacco cells, whereas all the enzymes have much lower activities in cells that have reached the stationary phase (de Pinto et al., 2000). On these bases, changes in the ASC and GSH metabolisms could also be required for kernel development, during which, phases of intense metabolic activities are followed by programmed dehydration and achievement of a resting phase. Materials and methods Plant growth Plants of Triticum durum Desf. (L.) (cv. Simeto) were grown in experimental fields in 1998–1999 in Rome on 10 m2 plots with a sowing density of up to 450 seed m–2. The plants were arranged in a randomized block design and collected on two segments of a 25 cm length on the same row. Irrigation, fertilization and plant protection were performed to ensure optimal plant growth. Ear flowering started on 4 May and plants were hand collected from 13 d after anthesis (DAA) (physiological stage of early milky phase) to 45 DAA (complete kernel development). The ears were separated from the stems and stored at –80 °C for analysis. Mid‐ear kernels (grains) were used for the measurements of metabolite contents and enzyme activities. Dry weight, water status and chlorophyll contents Kernels (20–50) were dried at 100 °C until constant dry weight was achieved (24–36 h). Water content was expressed on a percentage fresh weight basis. Chlorophyll contents (chlorophyll a plus chlorophyll b) were assayed according to Zhang and Kirkham (1996) using 0.5–1 g of kernels for each experiment. Carbohydrate and protein contents The kernels previously dehydrated by lyophilization were ground in a laboratory mill. The carbohydrate‐soluble analyses were carried out by a preliminary extraction with 96° ethanol for 1 h at 80 °C, followed by a water extraction for 2 h at 105 °C. The ethanol‐soluble fraction mainly contained low molecular weight carbohydrates (e.g. mono – and di‐saccharides), while, high molecular weight carbohydrates (e.g. fructans and oligosaccharides) were mainly present in the water‐soluble fraction. Glucose and fructose contents were determined in the two fractions using enzymatic (glucose oxidase/peroxidase kit; Sigma Diagnostic, St Louis, USA) or chemical (resorcine‐HCl) procedures, respectively. Since fructans are exclusively constituted by fructose and glucose, their quantification allowed the measurement of the total content of fructan (Virgona and Barlow, 1991). Starch determination was performed using the enzymatic assay procedure Total Starch (amyloglucosidase/α‐amylase method) (American Association of Cereal Chemists, 2000) proposed by Megazyme (Warriewood, Australia) Protein content was determined by the Dumas method (American Association of Cereal Chemists, 2000) with the automatic Leco FP‐428 system (Leco Corporation, St Joseph, MI USA). Carbohydrate and protein contents were calculated on a per‐kernel basis. Analysis of ascorbate and glutathione pools The kernels (0.5–1 g) were homogenized with eight volumes of cold 5% meta‐phosphoric acid at 4 °C. The homogenate was centrifuged at 20 000 g for 15 min at 4 °C, and the supernatant was collected for analysis of ascorbate and glutathione contents and redox state as described in de Pinto et al. (1999). Ascorbate and glutathione contents were calculated on a per‐kernel basis. Enzyme assays Kernels were ground in liquid N2 with a mortar and pestle. Four volumes of a medium which consisted of 50 mM Tris‐HCl (pH 7.8), 0.05% (w/v) cysteine, and 0.1% (w/v) BSA, were added just as the last trace of liquid N2 disappeared. The thawed mixture was then ground and centrifuged at 20 000 g for 15 min. The supernatant was used for both spectrophotometric and native‐PAGE analysis. DHA reductase (glutathione: dehydroascorbate oxidoreductase, EC 1.8.5.1), AFR reductase (NADH: ascorbate free radical oxidoreductase, EC 1.6.5.4) and ascorbate peroxidase (l‐ascorbate: hydrogen peroxide oxidoreductase, EC 1.11.1.11) were assayed according to Di Cagno et al. (2001). Catalase (hydrogen peroxide: hydrogen peroxide oxidoreductase, EC 1.11.1.6) activity assay was performed according to Beaumont et al. (1990) with minor modification, by following the H2O2 dismutation at 240 nm in a reaction mixture, which consisted of 0.1 M phosphate buffer, pH 7.0, 50–100 µg protein and 18 mM H2O2 (extinction coefficient 23.5 mM–1 cm–1). Glutathione reductase (NADPH: glutathione disulphide oxidoreductase, EC 1.6.4.2) was measured as indicated in Osswald et al. (1992). Enzyme activities were calculated on a per‐kernel basis. Native polyacrylamide gel electrophoresis (PAGE) for APX and DHAR was performed as previously reported (De Gara et al., 1997). Gels were photographed by using the gel documentation system GEL DOC 2000 by Biorad. Analysis of proteic sulphydric groups The identification of proteic –SH groups was performed by their labelling with mBBr according to Kobrehel et al. (1992). Kernels (0.5–2 g) were ground in liquid N2 with a mortar and pestle. Four volumes of 2 mM mBBr (dissolved in acetonitrile) in 100 mM Tris‐HCl, pH 7.5, buffer were added just as the last trace of liquid N2 disappeared. The thawed mixture was then ground for 1 min, transferred to a microfuge tube and centrifuged at 30 000 g for 15 min at 4 °C. Ten µl of 10% SDS and 10 µl of 100 mM 2‐mercaptoethanol were added to 80 µl of the mBBr‐labelled extracts to stop the reaction and derivatize excess mBBr. Samples were then applied to gels for electrophoretic analysis. SDS‐PAGE analysis of mBBr‐labelled extracts was performed at pH 8.5 in 12.5% gels of 1.5 mm thickness for approximately 14 h at a constant current of 8 mA. Electrophoresis was stopped when the solvent front, marked with bromophenol blue tracking die, migrated to approximately 1 cm from the bottom of the gel. After electrophoresis, gels were placed in 12% (w/v) trichloroacetic acid for 30 min to fix the proteins and were then transferred to a solution of 40% ethanol (v/v), 10% acetic acid (v/v) for 4–10 h to remove excess of mBBr. The fluorescence of proteins bound to mBBr was visualized by placing gels on a light box fitted with a UV light source (365 nm). Gels were photographed by using the gel documentation system GEL DOC 2000 by Biorad. Results Maturation parameters Kernel development was studied from 13 day after anthesis (DAA) (early milky phase) to complete maturation (45 DAA). As expected, during the first 28 d of development, an increase in fresh weight and size of the kernels was observed. After this stage the water loss due to the dehydration process exceeded the dry weight increase due to reserve storage and, therefore, kernel dimensions decreased in terms of both fresh weight and size (Table 1). Chlorophyll content was relatively stable until 17 DAA. At 21 DAA it decreased by about 20%, after which it progressively declined and, in the mature kernels, chlorophylls were not detectable (Table 1). During the first period of maturation, kernels were particularly rich in low molecular weight carbohydrates including fructans (Table 2). As the development proceeded, fructan content per kernel decreased, whereas the mono‐ and di‐ saccharide contents first decreased from 17 to 28 DAA, then returned to a value comparable to the 13 DAA value in the mature kernels (Table 2). As expected, the starch content increased during maturation (Table 2). Under these growth conditions, this storage polysaccharide reached values of 35 mg per kernel that is about 60% of mature kernel dry weight. During kernel development, protein contents also increased (Table 2). In order to analyse changes in the sulphydryl status of proteins during kernel maturation, the proteic –SH groups were labelled with monobromobimane (mBBr), which reacts stoichiometrically with cysteine residues (Crawford et al., 1989). The analysis showed a progressive increase in the number of proteic –SH groups during the first 21 DAA; this increase was particularly pronounced for proteins with a molecular mass ranging from 43–94 kDa (Fig. 1). At 28 DAA the proteic –SH groups available for interacting with mBBr started to decrease, particularly in the proteins with high molecular weight. Such a decrease is very probably due to the transition of sulphydryl groups to disulphide bridges. The changes in proteic redox state had the same behaviour when they were analysed on per kernel basis (Fig. 1) and when the same amount of proteins was loaded to the gel lanes (data not shown). Ascorbate and glutathione redox state The increase in weight and in the storage of reserve proteins occurring during kernel maturation renders the measurement of metabolites and enzymes on a per weight or per mg protein basis only slightly indicative of the physiological changes occurring during kernel maturation. For this reason, both ascorbate and glutathione contents and enzyme activities were calculated on a per kernel basis. Remarkable variations in the ascorbate and glutathione pools were also observed during kernel development and maturation. The total ascorbate content (ASC+DHA) gradually increased from 13 to 21 d, remained constant for the following 7 d then decreased, reaching very low values in the mature kernels (Fig. 2A). In addition, the redox balance of the ascorbate pool significantly changed during the maturation period. At the beginning of kernel development the reduced form of vitamin C was predominant. In the following period of maturation, the ASC/DHA ratio decreased and DHA was the only form of vitamin C present in the mature kernels (Fig. 2A). The glutathione pool also transiently increased during kernel maturation, reaching its maximum at 28 DAA (Fig. 2B). The redox balance of the glutathione pool was shifted more towards the oxidized form (GSSG) than that of ascorbate, since the GSH/GSSG ratio was around 1–1.5 during the whole maturation period, with the exception of the kernels collected at 28 DAA, the GSH/GSSG ratio of which decreased to about 0.4. Enzymes of H2O2 removal and of the ascorbate and glutathione cycle The redox enzymes of the ascorbate–glutathione cycle were also tested in order to follow their behaviour during kernel maturation. When determined on a per kernel basis, no significant differences were detected for APX during the first 28 DAA, but this H2O2 scavenging enzyme decreased until it was no longer detectable in the mature kernels (Fig. 3). Since three APX isoenzymes were active in wheat seedlings (Paciolla et al., 1996; De Gara et al., 1997), the presence of APX isoenzymes in wheat kernels and their pattern during maturation were analysed by native PAGE. Three APX bands, which had the same migration rate as seedling or germinating embryo isoenzymes (data not shown), were evident from 13–28 DAA of kernel development, but disappeared during the dehydration period (Fig. 4). The activity of AFRR increased by about 50% from 13 to 17 DAA, remained constant for the following 10 d, after which it decreased to values lower than that of day 13 (Fig. 3). DHAR activity also transiently increased during kernel development, but unlike AFRR, this ASC recycling enzyme activity increased 5‐fold over the period from 13 to 28 DAA and remained higher in mature kernels than at 13 d (Fig. 3). Several proteins with DHA‐reducing capability seemed to be involved in kernel maturation. At 13 DAA three proteins with DHA reducing capability were visible by native PAGE (Fig. 5). During maturation two new bands with lower migration rate appeared whereas the band with the highest migration rate disappeared (Fig. 5). The activity of GR was constant from 13–21 DAA, after which it started to decrease, reaching values about 50% lower in the mature kernels than those reported during the first 21 d of kernel development (Fig. 3). To obtain information on the global H2O2 removal capability of kernel cells during maturation, the activity of catalase (CAT) was also measured. As shown in Fig. 6, CAT almost doubled from 13 to 21–28 d, after which it rapidly decreased. Unlike APX, CAT was still present in the completely dehydrated kernels, even if with low activity. Discussion Data reported in this paper show that mono‐ and di‐saccharides, fructans and starch change in different ways during kernel maturation. The behaviour of fructans mirrors that of ASC, indeed, the content of both these metabolites increases during the first 21 DAA. Since fructans are polymers of fructose, and sucrose is the fructosyl donor in their biosynthesis (Pollock and Cairns, 1991), the glucose release occurring in the tissues that actively synthesize fructans could be used as a precursor of ASC. However, at present, it is not known whether kernels are able autonomously to synthesize ASC or whether they import this metabolite from the mother tissues. It has been reported that the rise in mono‐ and di‐saccharides detected from 28–45 DAA is mainly due to an increase in sucrose (Abou‐Guendia and D’Appolonia, 1972) and, according to the data present in the literature, this could be related to the acquisition of desiccation tolerance, a feature of particular relevance for orthodox seeds that reach maturity after a dehydration process. Indeed, the increase in sucrose, as well as in raffinose and dehydrins is concomitant and proportional to the induction of desiccation tolerance during seed maturation (Brenac et al., 1997; Black et al., 1999; Buitink et al., 2000). The synthesis and degradation of fructans mark different stages of kernel maturation. Fructans, as well as being a form of carbohydrate reserve stored in vegetative organs by 15% of flowering plants, wheat included (Vijn and Smeekens, 1999), are also involved in drought tolerance (Wiemken et al., 1995). Tobacco plants transformed with a bacterial gene for fructan synthesis have a higher resistance to drought stress than non‐transformed plants (Pilon‐Smits et al., 1995). Under water stress, plants hydrolyse fructans as a strategy for decreasing osmotic potential into cells (Virgona and Barlow, 1991). The decrease in fructans observed in the kernel from 21 DAA could be triggered by the dehydration process. On the other hand, the changes in the soluble carbohydrate content occurring as a consequence of the decrease in fructans (that are soluble oligosaccharides) and the concomitant increase in the synthesis of starch (insoluble polysaccharide) could contribute to up‐regulate phloem unloading and be part of the cross‐talk between the whole plant and the developing seeds. Indeed, it has been proposed that a turgor‐homeostat model might regulate phloem unloading and sugar uptake in embryos of several species (Weber et al., 1997). As far as the ROS–scavenger processes are concerned, the activities of APX and CAT suggest that the H2O2‐detoxification requirement is almost constant during the first 28 d of kernel maturation; whereas, it decreases during dehydration, since the activities of the two enzymes responsible for H2O2‐removal fall remarkably. The differences observed in their behaviours could be due to their different localization in cells and, as a consequence, to their involvement in metabolic pathways taking place at different times during kernel maturation. CAT is mainly present in the microbodies (Scandalios, 1994), organelles that, being involved in the fatty acid metabolism, are particularly active in lipid synthesizing cells. This could explain an increase in CAT activity during the first period of kernel development, when the synthesis of the storage lipids also increases (Morrison, 1988). On the other hand, APX, being more widely shared out in cell organelles (De Gara and Tommasi, 1999, and references reported here), could intervene in the removal of H2O2 produced by metabolic pathways with different timing during kernel maturation. Moreover, APX is always very active in dividing cells and tissues undergoing differentiation (De Gara et al., 1996; de Pinto et al., 2000), whereas it decreases in senescent tissues (Borraccino et al., 1994). These reports perfectly agree with a high and quite stable APX activity from the beginning of kernel formation until the start of the dehydration phase, the period during which the cells undergo division and differentiation/storing processes. The loss in the H2O2‐scavenging capability in the physiological context of orthodox seed dehydration is not a critical event for seed viability, since the absence or very low activities of ROS‐scavenging enzymes are typical features of all mature and viable orthodox seeds (Cakmak et al., 1993; De Gara et al., 1997; Tommasi et al., 1999, 2001). A decrease in ROS generation also occurs during seed dehydration (Vertucci and Farrant, 1995). Indeed, the oxidative metabolism, that is the main source of ROS in non‐green tissues, decreases together with the dehydration process, until reaching very low values in dehydrated seeds (Vertucci, 1989; Leprince et al., 2000). Photosynthesis, another process responsible for ROS generation, is active during the first stages of kernel maturation, but at 28 DAA kernels contain a much lower amount of chlorophylls compared with the previous period and no chlorophylls are detectable after 45 DAA, thus confirming that a progressive decrease in this other physiological source of ROS also occurs during dehydration. However, the direct measurement of the changes in ROS production and ROS‐scavenging enzymes in the same developing seed will be necessary in order to verify the existence of a causal relationship between a decrease in ROS production and APX and CAT decline in ripening seeds. The presence in dry kernels of the enzymes of the ascorbate and glutathione recycle (i.e. AFRR, DHAR and GR) is of great importance for the first stage of germination. Since dry kernels are provided with DHA, these three recycling enzymes allow the ascorbate–glutathione cycle to be active as soon as they are reactivated by seed imbibition. ASC supplied by means of the reduction of its oxidized forms is necessary for the activation of the ASC‐dependent metabolic pathway at the beginning of germination, since a lag of several hours is required for the start of ASC biosynthesis (De Gara et al., 1997; Tommasi et al., 2001). The ASC–GSH recycling enzymes seem to have a different relevance during kernel maturation. AFRR activities resemble those of ASC utilization and, indeed, AFR production. Besides APX, several other enzymes, utilizing ASC as a physiological electron donor, show transient increase during kernel maturation. Among these are dioxygenases involved in the synthesis of gibberellin and abscissic acid (Prescott and John, 1996). The synthesis of these two phytohormones transiently increases during seed ripening (Frydman et al., 1974; Rock and Quatrano, 1995); and, at least in the case of abscisic acid, this increase is due to the seed biosynthetic activity, not to hormone translocation from the parental tissues (Hole et al., 1989). Another phytohormone synthesized by an ASC‐depending enzyme is ethylene. It has been reported that, in wheat endosperm, production of ethylene reaches its highest level in the middle phase of kernel development (between 16–25 DAA), after which it decreases to undetectable values during the dehydration period (30–40 DAA) (Young and Gallie, 1999), when ASC is no longer available. Among the enzymes whose activities increase during kernel development, DHAR has the highest increase in percentage terms. It is worth noting that protein disulphide isomerase has DHA reducing activity, because it utilizes DHA as an electron acceptor for maintaining two cysteine residues at its catalytic site in the oxidized form (Wells and Xu, 1994). In seed development, protein disulphide isomerase plays a key role in the maturation of storage proteins. The analysis of mBBR‐labelled proteic –SH indicates that the transition from sulphydryl groups to disulphide bridges within proteins starts at 28 DAA, when the DHA reducing capability has the highest value. Thus, it is possible that some of the proteins, present during kernel maturation, that are detectable for their DHA‐reducing activity by native PAGE, may be protein disulphide isomerases. The GSH/GSSG pair is also involved in protein folding by sulphydryl groups–disulphide bridges transition (Kunert and Foyer, 1993). It has been reported that a low GSH/GSSG ratio facilitates the folding and assembly of newly synthesized secretory proteins in the endoplasmic reticulum (Hwang et al., 1992). The GSH/GSSG ratio found in wheat kernels is similar to that detected in the endoplasmic reticulum (around 2) or even lower. It is also known that GSSG stimulates the rate of assembly of polymeric proteins by oxidizing specific cysteine residues (Jung et al., 1997); moreover, it is responsible for the protection of specific thiol groups from the irreversible formation of intramolecular disulphide bonds (Kranner and Grill, 1996). In this reaction that occurs in seed dehydration and in vegetative tissues undergoing water stress conditions, a GSSG molecule reacts with free thiol groups of proteins producing a protein–glutathione complex and a molecule of GSH. Interestingly, the increase in the GSH/GSSG ratio that occurs after 28 d of kernel maturation (from a value of 0.4 at 28 d to a value of 1.6 at 45 d), in spite of GSSG‐reductase activity decreasing during this period, can be explained by means of an involvement of GSSG reduction in the SH–S‐S transition that mainly occurs 28 d after anthesis, as well as in protein‐S‐SG formation. The data reported here also show that the redox state of the ascorbate and glutathione pools of kernels shifts towards the oxidized forms during maturation, unlike that which occurs in vegetative tissues or recalcitrant seeds, where the reduced forms represent 90–95% of the total pools (Hendry et al., 1992; Tommasi et al., 1999). Recently, it has been reported that the induction of an increase in DHA content in cultured cells is responsible for the decrease in their mitotic activity (de Pinto et al., 1999; Potters et al., 2000). The availability of ASC and GSH and changes in their redox state also affect gene expression (Noctor and Foyer, 1998; Catani et al., 2001); thus, during kernel maturation, the levels and redox balance of the two redox pairs could also be involved in the activation/inactivation of specific metabolic pathways. On this basis, the peculiar redox state occurring during orthodox seed development and, in particular, the fact that at the end of their maturation processes only DHA is present in the seeds, raises the question as to whether such a shift toward the oxidized state is only a consequence of ASC utilization or whether it is a strategy carried out by orthodox seeds that contributes to reaching the resting stage. Acknowledgements We thank Vincenzo Cellamare for technical assistance. This work was partially supported by University of Bari and by Consiglio Nazionale delle Ricerche. Open in new tabDownload slide Fig. 1. Sulphydryl group–disulphide bridge transition in proteins of developing wheat kernels. SDS‐PAGE of proteins extracted from kernels at different DAA. The presence of the –SH groups was labelled with mBBr. One‐fifth of the total protein amount of a kernel was loaded in each lane. Open in new tabDownload slide Fig. 2. Ascorbate (A) and glutathione (B) pools in wheat kernels at different stages of maturation. The reported values are the means of five experiments ±standard error (SE). Open in new tabDownload slide Fig. 3. Ascorbate–glutathione redox enzymes in wheat kernels at different stages of maturation. The reported values are the means of five experiments ± standard error (SE). 1 unit=1 nmol ASC oxidized min–1 kernel–1 (APX); 1 nmol NADH oxidized min–1 kernel–1 (AFRR), 1 nmol DHA reduced min–1 kernel–1 (DHAR); 1 nmol NADPH oxidized min–1 kernel–1 (GR). Open in new tabDownload slide Fig. 4. Native PAGE of ascorbate peroxidase isoenzymes in wheat kernels at different stages of maturation. Half of the protein content of a kernel was loaded in each lane. Open in new tabDownload slide Fig. 5. Native PAGE of dehydroascorbate reducing proteins in wheat kernels at different stages of maturation. A quarter of the protein content of a kernel was loaded in each lane. Open in new tabDownload slide Fig. 6. Catalase activity in wheat kernels at different stages of maturation. The reported values are the means of five experiments ±standard error (SE). 1 unit=1 nmol H2O2 dismutated min–1 kernel–1. Table 1. Growth parameters, hydric state and chlorophyll content in wheat kernels collected from ears at different stage of maturation Days after anthesis Fresh weighta (mg kernel–1) Dry weighta (mg kernel–1) Water content (% fresh weight) Chlorophylla (µg g–1 FW) Kernel sizesb (mm kernel–1) 13 33±0.4 7.5±0.3 77 112±4 5.8±0.1 17 52±0.8 14.6±0.6 72 103±6 7.6±0.3 21 64±0.5 28.2±0.2 56 84±5 8.0±0.4 28 90±1.2 47.1±0.4 52 36±2 8.6±0.3 45 64±0.7 58.2±0.6 9 n.d.c 8.1±0.2 Days after anthesis Fresh weighta (mg kernel–1) Dry weighta (mg kernel–1) Water content (% fresh weight) Chlorophylla (µg g–1 FW) Kernel sizesb (mm kernel–1) 13 33±0.4 7.5±0.3 77 112±4 5.8±0.1 17 52±0.8 14.6±0.6 72 103±6 7.6±0.3 21 64±0.5 28.2±0.2 56 84±5 8.0±0.4 28 90±1.2 47.1±0.4 52 36±2 8.6±0.3 45 64±0.7 58.2±0.6 9 n.d.c 8.1±0.2 a The reported results are the mean values obtained from five different experiments ±standard error (SE). b The reported results are the mean values obtained from the measurement of 50 kernels ±standard error (SE). c n.d.=not detectable. Open in new tab Table 1. Growth parameters, hydric state and chlorophyll content in wheat kernels collected from ears at different stage of maturation Days after anthesis Fresh weighta (mg kernel–1) Dry weighta (mg kernel–1) Water content (% fresh weight) Chlorophylla (µg g–1 FW) Kernel sizesb (mm kernel–1) 13 33±0.4 7.5±0.3 77 112±4 5.8±0.1 17 52±0.8 14.6±0.6 72 103±6 7.6±0.3 21 64±0.5 28.2±0.2 56 84±5 8.0±0.4 28 90±1.2 47.1±0.4 52 36±2 8.6±0.3 45 64±0.7 58.2±0.6 9 n.d.c 8.1±0.2 Days after anthesis Fresh weighta (mg kernel–1) Dry weighta (mg kernel–1) Water content (% fresh weight) Chlorophylla (µg g–1 FW) Kernel sizesb (mm kernel–1) 13 33±0.4 7.5±0.3 77 112±4 5.8±0.1 17 52±0.8 14.6±0.6 72 103±6 7.6±0.3 21 64±0.5 28.2±0.2 56 84±5 8.0±0.4 28 90±1.2 47.1±0.4 52 36±2 8.6±0.3 45 64±0.7 58.2±0.6 9 n.d.c 8.1±0.2 a The reported results are the mean values obtained from five different experiments ±standard error (SE). b The reported results are the mean values obtained from the measurement of 50 kernels ±standard error (SE). c n.d.=not detectable. Open in new tab Table 2. Changes in carbohydrates and proteins contents during wheat kernel maturation The values are the mean of five experiments ±SE. Days after Mono‐ and di‐saccharides Fructans Starch Proteins anthesis (mg kernel–1) (mg kernel–1) (mg kernel–1) (mg kernel–1) 13 0.92±0.02 1.66±0.01 1.3±0.30 1.1±0.01 17 0.95±0.01 1.87±0.01 5.8±0.30 2.2±0.01 21 0.75±0.03 1.82±0.01 12.9±0.20 7.1±0.01 28 0.58±0.02 1.02±0.01 24.4±0.40 7.1±0.03 45 0.92±0.06 0.80±0.04 34.4±0.32 9.5±0.03 Days after Mono‐ and di‐saccharides Fructans Starch Proteins anthesis (mg kernel–1) (mg kernel–1) (mg kernel–1) (mg kernel–1) 13 0.92±0.02 1.66±0.01 1.3±0.30 1.1±0.01 17 0.95±0.01 1.87±0.01 5.8±0.30 2.2±0.01 21 0.75±0.03 1.82±0.01 12.9±0.20 7.1±0.01 28 0.58±0.02 1.02±0.01 24.4±0.40 7.1±0.03 45 0.92±0.06 0.80±0.04 34.4±0.32 9.5±0.03 Open in new tab Table 2. Changes in carbohydrates and proteins contents during wheat kernel maturation The values are the mean of five experiments ±SE. Days after Mono‐ and di‐saccharides Fructans Starch Proteins anthesis (mg kernel–1) (mg kernel–1) (mg kernel–1) (mg kernel–1) 13 0.92±0.02 1.66±0.01 1.3±0.30 1.1±0.01 17 0.95±0.01 1.87±0.01 5.8±0.30 2.2±0.01 21 0.75±0.03 1.82±0.01 12.9±0.20 7.1±0.01 28 0.58±0.02 1.02±0.01 24.4±0.40 7.1±0.03 45 0.92±0.06 0.80±0.04 34.4±0.32 9.5±0.03 Days after Mono‐ and di‐saccharides Fructans Starch Proteins anthesis (mg kernel–1) (mg kernel–1) (mg kernel–1) (mg kernel–1) 13 0.92±0.02 1.66±0.01 1.3±0.30 1.1±0.01 17 0.95±0.01 1.87±0.01 5.8±0.30 2.2±0.01 21 0.75±0.03 1.82±0.01 12.9±0.20 7.1±0.01 28 0.58±0.02 1.02±0.01 24.4±0.40 7.1±0.03 45 0.92±0.06 0.80±0.04 34.4±0.32 9.5±0.03 Open in new tab References Abou‐GuendiaM, and D’Appolonia BL. 1972 . Changes in carbohydrate components during wheat maturation. I. Changes in free sugars. Cereal Chemistry 49, 664 –676. American Association of Cereal Chemists. 2000 . Approved methods of the AACC. St Paul, Minesota, USA: American Association of Cereal Chemists. AsadaK. 1999 . The water–water cycle in chloroplasts: scavenging of active oxygen and dissipation of excess photons. Annual Review of Plant Physiology and Plant Molecular Biology 50, 601 –639. BeaumontF, Jouve HM, Gagnon J, Gaillard J, Pelmont J. 1990 . Purification and properties of a catalase from potato tubers (Solanum tuberosum). Plant Science 72, 19 –26. BlackM, Corbineau F, Gee H, Gee H, Côme D. 1999 . Water content, raffinose, and dehydrins in the induction of desiccation tolerance in immature wheat embryos. Plant Physiology 120, 463 –471. BradfordKJ, Chandler PM. 1992 . Expression of ‘dehydrine like’ proteins in embryos and seedlings of Zizania palustris and Oryza sativa during dehydration. Plant Physiology 99, 488 –494. BrenacP, Horbowicz M, Downer SM, Dickerman AM, Smith ME, Obendorf RL. 1997 . Raffinose accumulation related to desiccation tolerance during maize (Zea mays L.) seed development and maturation. Journal of Plant Physiology 150, 481 –488. BorraccinoG, Mastropasqua L, De Leonardis S, Dipierro S. 1994 . The role of ascorbic acid system in delaying the senescence of oat (Avena sativa L.) leaf segments. Journal of Plant Physiology 144, 161 –166. BuitinkJ, Hemminga MA, Hoekstra F. 2000 . Is there a role for oligosaccharides in seed longevity? An assessment of intracellular glass stability. Plant Physiology 122, 1217 –1224. CakmakI, Strbac D, Marschner H. 1993 . Activities of hydrogen peroxide–scavenging enzymes in germinating wheat seeds. Journal of Experimental Botany 44, 127 –132. CataniMV, Rossi A, Sabatini S, Levriero M, Melino G, Avigliano L. 2001 . Induction of gene expression via activator protein‐1 in the ascorbate protection against UV‐induced damage. Biochemical Journal 356, 77 –85. CrawfordNA, Droux M, Kosower NS, Buchanan BB. 1989 . Evidence for function of the ferredoxin/thioredoxin system in the reductive activation of target enzymes of isolated intact chloroplasts. Archives of Biochemistry and Biophysics 271, 223 –239. De GaraL, de Pinto MC, Arrigoni O. 1997 . Ascorbate synthesis and ascorbate peroxidase activity during the early stage of wheat germination. Physiologia Plantarum 100, 894 –900. De GaraL, de Pinto MC, Paciolla C, Cappetti V, Arrigoni O. 1996 . Is ascorbate peroxidase only a scavenger of hydrogen peroxide? In: Obinger C, Burner U, Ederman R, Penel C, Greppen H, eds. Plant peroxidases: biochemistry and physiology. Geneve: University of Geneve, 157 –162. De GaraL, Tommasi F. 1999 . Ascorbate redox enzymes: a network of reaction involved in plant development. Recent Research Development in Phytochemistry 3, 1 –15. de PintoMC, Dennis F, De Gara L. 1999 . The redox state of ascorbate‐dehydroascorbate pair as a specific sensor of cell division in tobacco TBY‐2 cells. Protoplasma 209, 90 –97. de PintoMC, Tommasi F, De Gara L. 2000 . Enzymes of the ascorbate biosynthesis and ascorbate–glutathione cycle in cultured cells of tobacco bright yellow 2. Plant Physiology and Biochemistry 38, 541 –550. Di CagnoR, Guidi L, De Gara L, Soldatini G. 2001 . Combined cadmium and ozone treatments affect photosynthesis and ascorbate‐dependent defences in sunflower. New Phytologist 151, 627 –636. FrydmanVM, Gaskin P., MacMillan J. 1974 . Qualitative and quantitative analysis of gibberellins through seed maturation in Pisum sativum cv. Progress no. 9. Planta 118, 123 –132. GebbingT, Schnyder H. 1999 . Pre‐anthesis reserve utilization for protein and carbohydrate synthesis in grains of wheat. Plant Physiology 121, 871 –878. HendryGAF, Finch‐Savage WE, Thorpe PC, Atherton NM, Buckland SM, Nilsson KA, Seel WE. 1992 . Free radical processes and loss of seed viability during desiccation in the recalcitrant species Quercus robur L. New Phytologist 122, 273 –279. HoleDJ, Smith JD, Cobb BG. 1989 . Regulation of embryo dormancy by manipulation of abscissic acid in kernels and associated cob tissue of Zea mays L. cultured in vitro. Plant Physiology 91, 101 –105. HwangC, Sinskey AJ, Lodish HF. 1992 . Oxidized redox state of glutathione in endoplasmic reticulum. Science 257, 1496 –1502. JungR, Nam YW, Saalbach I, Müntz K, Nielse NC. 1997 . Role of sulphydryl redox state and disulphide bonds in processing and assembly of 11S seed globulins. The Plant Cell 9, 2037 –2050. KobrehelK, Wong JH, Balogh A, Kiss F, Yee BC, Buchanan BB. 1992 . Specific reduction of wheat storage proteins by thioredoxin h. Plant Physiology 99, 919 –924. KrannerI, Grill D. 1996 . Significance of thiol‐disulphide exchange in resting stages of plant development. Botanica Acta 109, 8 –14. KunertKJ, Foyer C. 1993 . Thiol disulphide exchange in plants. In: De Kok LJ, ed. Sulphur nutrition and assimilation in higher plants. The hague, The Netherlands: Academic Publishing bv, 139 –150. LeprinceO, Hoeksra FA, Harren FJM. 2000 . Unravelling the responses of metabolism to dehydration point to a role for cytoplasmic viscosity in desiccation tolerance. In: Black M, Bradford KJ, Vasquez‐Ramos J, eds. Seed biology: advances and application. Oxon, New York: CABI Publishing, 57 –66. MansfieldTA, McAinsh MR. 1995 . Hormones as regulator of water balance. In: Davies PJ, ed. Plant hormones–physiology, biochemistry, molecular biology. The Netherlands: Kluwer Academic Publishers, 671 –697. MorrisonWR. 1988 . Lipids. In: Pomeranz Y, ed. Wheat chemistry and technology. St Paul, Minesota, USA: American Association of Cereal Chemists, 373 –439. NoctorG, Foyer CH. 1998 . Ascorbate and glutathione: keeping active oxygen under control. Annual Review of Plant Physiology and Plant Molecular Biology 49, 729 –737. OsswaldWF, Kraus R, Hippeli S, Benz B, Volpert R, Elstner EF. 1992 . Comparison of the enzymatic activity of dehydroascorbic acid reductase, glutathione reductase, catalase, peroxidase, and superoxide dismutase of healthy and damaged spruce needles (Picea abies (L.) Karst.). Journal of Plant Physiology 139, 742 –748. PaciollaC, De Gara L, De Tullio M, Arrigoni O. 1996 . Distribution of cytosolic ascorbate peroxidase in angiosperms. Giornale Botanico Italiano 130, 729 –737. Pilon‐SmitsEAH, Ebskamp MJM, Paul MJ, Jeuken MJW, Weisbeerk PJ, Smeekens SCM. 1995 . Improved performance of transgenic fructan‐accumulating tobacco under drought stress. Plant Physiology 107, 125 –130. PollockCJ, Cairns AJ. 1991 . Fructan metabolism in grasses and cereals. Annual Review of Plant Physiology and Plant Molecular Biology 42, 77 –101. PottersG, Horemans N, Caubergs RJ, Asard H. 2000 . Ascorbate and dehydroascorbate influence cell cycle progression in a tobacco cell suspension. Plant Physiology 124, 17 –20. PrescottAG, John P. 1996 . Dioxygenases: molecular structure and role in plant metabolism. Annual Review of Plant Physiology and Plant Molecular Biology 47, 245 –271. RobertsEH. 1973 . Predicting the storage life of seeds. Seed Science and Technology 1, 499 –514. RockCD, Quatrano RS. 1995 . The role of hormones during seed development. In: Davies JP, ed. Plant hormones–physiology, biochemistry, molecular biology. The Netherlands: Kluwer Academic Publishers, 671 –697. ScandaliosJG. 1994 . Regulation and properties of plant catalases. In: Foyer CH, Mullineaux PM, eds. Causes of photo‐oxidative stress and amelioration of defense system in plant. Boca Raton: CRC Press, 275 –315. SgherriCLM, Navarri‐Izzo F. 1995 . Sunflower seedlings subjected to increasing water deficit stress: oxidative stress and defence mechanisms. Physiologia Plantarum 93, 25 –30. SkriverK, Mundy J. 1990 . Gene expression in response to abscisic acid and osmotic stress. The Plant Cell 2, 503 –512. SmirnoffN, Wheeler GL. 2000 . Ascorbic acid in plants: biosynthesis and function. Critical Reviews in Plant Sciences 19, 267 –290. SteevesTA. 1983 . The evolution and biological significance of seeds. Canadian Journal of Botany 61, 3550 –3560. TommasiF, Paciolla C, Arrigoni O. 1999 . The ascorbate system in recalcitrant and orthodox seeds. Physiologia Plantarum 105, 193 –198. TommasiF, Paciolla C, de Pinto MC, De Gara L. 2001 . A comparative study of glutathione and ascorbate metabolism during germination of Pinus pinea L. seeds. Journal of Experimental Botany 52, 1647 –1654. VertucciCW. 1989 . The effect of water contents on physiological activities of seeds. Physiologia Plantarum 77, 172 –176. VertucciCW, Farrant JM. 1995 . Acquisition and loss of desication tolerance. In: Kigel J. Galili G, eds. Seed development and germination. New York: Marcel Dekker, 237 –271. VijnI, Smeekens S. 1999 . Fructan: more than a reserve carbohydrate? Plant Physiology 120, 351 –359. VirgonaJM, Barlow EWR. 1991 . Drought stress induces changes in the non‐structural carbohydrate composition of wheat stem. Australian Journal of Plant Physiology 18, 239 –247. WeberH, Borisjuk L, Wobus U. 1997 . Sugar import and metabolism during seed development. Trends in Plant Science 2, 169 –174. WellsWW, Xu DP. 1994 . Dehydroascorbate reduction. Journal of Bioenergetics and Biomembranes 26, 369 –377. WiemkenA, Sprengen N, Bollen T. 1995 . Fructan: an extension of sucrose by sucrose. In: Pontis HG, Salerno GL, Echeverria EJ, eds. Sucrose metabolism, biochemistry, physiology and molecular biology. Rockville, MD: American Society of Plant Physiologists, 178 –189. YoungTE, Gallie DR. 1999 . Analysis of programmed cell death in wheat endosperm reveals differences in endosperm development between cereals. Plant Molecular Biology 39, 915 –926. ZhangF, Kirkham MB. 1996 . Antioxidant responses to drought in sunflower and sorghum seedlings. New Phytologist 132, 361 –373. © 2003

Marillia, Elizabeth‐France; Micallef, Barry J.; Micallef, Malgre; Weninger, Alan; Pedersen, Kalie K.; Zou, Jitao; Taylor, David C.

doi: 10.1093/jxb/erg020pmid: 12493853

Abstract Pyruvate dehydrogenase kinase (PDHK), a negative regulator of the mitochondrial pyruvate dehydrogenase complex (mtPDC), plays a pivotal role in controlling mtPDC activity, and hence, the TCA cycle and cell respiration. Previously, the cloning of a PDHK cDNA from Arabidopsis thaliana and the effects of constitutively down‐regulating its expression on plant growth and development has been reported. The first detailed analyses of the biochemical and physiological effects of partial silencing of the mtPDHK in A. thaliana using antisense constructs driven by both constitutive and seed‐specific promoters are reported here. The studies revealed an increased level of respiration in leaves of the constitutive antisense PDHK transgenics; an increase in respiration was also found in developing seeds of the seed‐specific antisense transgenics. Both constitutive and seed‐specific partial silencing of the mtPDHK resulted in increased seed oil content and seed weight at maturity. Feeding 3‐14C pyruvate to bolted stems containing siliques (constitutive transgenics), or to isolated siliques or immature seeds (seed‐specific transgenics) confirmed a higher rate of incorporation of radiolabel into all seed lipid species, particularly triacylglycerols. Neither constitutive nor seed‐specific partial silencing of PDHK negatively affected overall silique and seed development. Instead, oil and seed yield, and overall plant productivity were improved. These findings suggest that a partial reduction of the repression of the mtPDC by antisense PDHK expression can alter carbon flux and, in particular, the contribution of carbon moieties from pyruvate to fatty acid biosynthesis and storage lipid accumulation in developing seeds, implicating a role for mtPDC in fatty acid biosynthesis in seeds. Key words: Arabidopsis thaliana, early flowering, mitochondrial pyruvate dehydrogenase complex, plant development, pyruvate dehydrogenase kinase (EC 2.7.1.99), respiration, seed oil content, seed weight. Received 22 April 2002; Accepted 30 August 2002 Introduction The pyruvate dehydrogenase complex (PDC) is a large multienzyme system catalysing the oxidative decarboxylation of pyruvate with the concomitant release of acetyl‐CoA and reduction of NAD+ through the following reaction sequence: pyruvate+NAD++CoA‐SH→acetyl‐CoA+NADH+CO2. The complex contains three primary components: pyruvate dehydrogenase (PDH; E1, EC 1.2.4.1); dihydrolipoamide transacetylase (E2, EC 2.3.1.12); and dihydrolipoamide dehydrogenase (E3, EC 1.8.1.4). Plants are unique in having PDCs in two isoforms, one located in the mitochondrial matrix as in other eukaryotic cells, and the other located in the plastid stroma (Lernmark and Gardeström, 1994). Although both plastidic and mitochondrial PDC (mtPDC) isoforms are very sensitive to product feedback regulation (by NADH and acetyl‐CoA), only the mtPDC is regulated through inactivation/reactivation by phosphorylation/dephosphorylation (Randall et al., 1981; Budde et al., 1988). Current knowledge about the molecular structure of plant mtPDCs is largely based on studies of its mammalian counterparts. As in mammalian systems, plant mtPDC contains the regulatory proteins, PDH kinase (PDHK, EC 2.7.1.99) and the PDH phosphatase (PDHP, EC 3.1.3.43) (Budde and Randall, 1990). PDHK is the negative regulator of the PDC. PDH has an α2β2 structure and it is the E1α component which is phosphorylated by PDHK, resulting in the inactivation of PDH. PDHP reactivates PDH through dephosphorylation. The in situ steady‐state level of phosphorylation of the E1 subunit can be regulated by pyruvate, ATP, and the particular substrates being oxidized by the mitochondria (Budde et al., 1988). Light also exerts an indirect effect on the inhibition of mtPDC activity, likely associated with photosynthesis and/or photorespiration (Budde and Randall, 1990). Maximum PDC activity appears to vary developmentally, with the highest catalytic activity observed during seed germination and early seedling development (Hill et al., 1992; Grof et al., 1995). MtPDC acts as a link between glycolytic carbon metabolism and the tricarboxylic acid (TCA) cycle. Because of the irreversible nature of the reaction, PDC is a particularly important site for metabolic regulation. The importance of the TCA cycle in plant development, especially for the initiation of the reproductive phase of plant growth has been demonstrated (Landschütze et al., 1995). In that study, antisense repression of the mitochondrial citrate synthase gene resulted in the delayed flowering phenotype and a specific disintegration of the ovary tissues of flowers. The reaction catalysed by plastidial PDC provides acetyl‐CoA for fatty acid biosynthesis both in vegetative and reproductive tissues, including seeds, but a role for mtPDC in fatty acid biosynthesis in seeds has not been conclusively demonstrated. Based on the sequence information from the molecular characterization of mammalian mtPDCs, the cloning of a number of plant cDNAs encoding mtPDC components has been reported (Luethy et al., 1994, 1995; Grof et al., 1995; Guan et al., 1995; Thelen et al., 1998a, b, 1999; Zou and Taylor, 1998; Zou et al., 1999). Rather than typical Ser/Thr kinase domains for phosphorylation, PDHKs from plants contain five conserved domains more characteristic of prokaryotic two‐component histidine kinases (Thelen et al., 1998a, 2000; Zou et al., 1999). Prokaryotic histidine kinases typically autophosphorylate a His residue, followed by phosphotransfer to Asp or Glu residues of their response regulator protein, thereby inducing signal transduction (Stock et al., 1989). Recently, through the use of histidine modifying agents, it has been shown that in the A. thaliana PDCK, His residues are critical for both the autophosphorylation and the phosphotransfer mechanisms associated with PDCK, and thus confirmed the A. thaliana PDHK to be a His kinase (Mooney et al., 2000). Recombinant A. thaliana PDCK was shown to inactivate kinase‐depleted corn mtPDC, by phosphorylating Ser residues. When using site‐directed mutagenesis on a conserved His‐121 residue (changed to alanine or glutamine), both the autophosphorylation and transphosphorylation functions of the PDCK were reduced by about 50%, implicating the His‐121 residue as at least one residue critical for PDCK function. By covalent alkylation of both His residues 121 and 168, serine phosphorylation and kinase function were reduced to 0–12%, thereby implicating both His 121 and His 168 (the only two invariant His residues amongst the cloned PDHKs) as essential for complete PDCK function (Thelen et al., 2000). In a previous study, using a transgenic approach to characterize mtPDHK, it was demonstrated that the activity of the A. thaliana PDC could be manipulated by modulating the expression level of its regulator, the PDCK gene. The effects of partial repression of PDHK expression on plant growth and development were described for transgenic plants expressing an antisense construct encoding an A. thaliana PDHK (AtPDHK) cDNA using a CaMV 35S promoter for constitutive expression in the plant. Constitutive antisense expression of the AtPDHK cDNA in transgenic plants resulted in reduced PDHK expression and elevated PDC activity based on measurements of in vitro enzyme activity. These transgenic plants also had an altered growth phenotype with reduced accumulation of vegetative mass, early flower initiation, and a shortened generation time (Zou et al., 1999). Immunoblot analyses of control and transgenic antisense PDHK mitochondrial protein preparations were probed with a monoclonal antibody to the E1α subunit of maize mitochondrial PDH. The maize antibody had shown a strong cross‐reactivity with the Arabidopsis 43 kDa PDH E1α subunit (Thelen et al., 1998b). However, as reported previously, a comparison of non‐transformed wild‐type (n‐WT), pBI121 plasmid‐only transgenic control and antisense PDHK transgenic line 9‐5 mitochondrial proteins showed that there was no evidence that the amount of the E1α subunit was increased in the antisense PDHK transgenic line. Similar results were obtained in western analyses of the other antisense PDHK transgenic lines. Therefore the increase in mtPDC activity was not associated with an increase in the amount of PDH protein, but in the level of activated enzyme (Zou et al., 1999). This previous work did not examine whether the metabolic flux through the reaction catalysed by PDC was actually increased in the transgenics. Studies of the A. thaliana antisense PDHK transgenic plants that contain the AtPDHK cDNA under the control of either a constitutive or seed‐specific promoter are reported here. Particular emphasis is placed on the physiological and biochemical changes and alterations of carbon flux resulting from altered mtPDHK activity and it is shown that these changes can lead to altered phenotypes exhibiting increased harvest index, seed storage lipids and seed weight, thereby improving plant productivity. Further more, the data also support the involvement of mtPDC in fatty acid biosynthesis in developing seeds. Materials and methods Plant material Arabidopsis thaliana plants, ecotype Columbia, both wild type and transgenics, were grown at the same time on Terra‐lite Redi‐earth (WR Grace & Co., Canada Ltd., Ajax, ON) in a growth chamber set at 22 °C with a diurnal photoperiod of 16 h light (200 µE m–2 s–1) and 8 h dark, unless otherwise stated. Under these growth conditions, one generation of wild‐type Columbia plants took 75–77 d (Zou et al., 1999). Molecular cloning of the AtPDCK gene and production of transgenic plants General molecular techniques such as plasmid DNA isolation, restriction digestion, modification and ligation of DNA, polymerase chain reaction, agarose gel electrophoresis, Northern blots, transformation and culture of E. coli were carried out according to standards protocols (Sambrook et al., 1989). An Arabidopsis mtPDHK clone (YA5) previously isolated and characterized, was first cloned in an antisense orientation behind the double 35S promoter and used to transform Arabidopsis plants with Agrobacterium as described (Zou et al., 1999). For seed‐specific down‐regulation, the same fragment (975 bp) of the PDHK YA5 clone was inserted in an antisense orientation behind the napin promoter of the pDH 1 vector (a customized vector derived from pUC19) at the XbaI and BamHI sites. The orientation of the PDCK fragment in the pDH 1 vector was verified by restriction digestion and DNA sequencing. A HindIII/EcoRI digestion excised the expression cassette which was then ligated to the corresponding sites of the transforming vector pRD400 (Datla et al., 1992; courtesy of Dr R Datla, PBI/NRC). Agrobacterium transformation (strain GV3101 pMP10) and plant vacuum infiltration were the same as described previously (Zou et al., 1999), with the plasmid‐only control being pRD400. Selection of transgenic plants, propagation and the convention for designating the transgenic generation (T1, T2, T3, etc) are as described by Katavic et al. (1994). For this study, the analyses were performed on progeny of the T3 generation. The analysis of the transgenic plants, pyruvate dehydrogenase and citrate synthase enzyme assays were conducted as described by Zou et al. (1999). Seed oil content and lipid species analyses were performed as cited previously (Katavic et al., 1995; Zou et al., 1997). Substrate incorporation studies Feeding experiments were conducted on either developing siliques, as in the case of the constitutive antisense transgenic line 9‐5‐A, and its corresponding plasmid‐only control (pBI121), or on both developing siliques and individual immature seeds from the seed‐specific antisense transgenic line 5‐6‐E, and its corresponding plasmid‐only control (pRD400). All experiments were carried out twice. For silique feedings, bolted stems were harvested 2 weeks after the first flower and fed at room temperature with 1 µCi 3‐14C sodium pyruvate (American Radiolabelled Chemicals) in 100 µl water, followed by a 500 µl ‘chase’ of distilled water, and incubated for 48 h in a growth chamber as described above. For each line, siliques were harvested and pooled within the same stage of development, numbered from one to eight as described previously ((Zou et al., 1996) for preparation of a total lipid extract (TLE). The different families of lipids in the TLE were resolved by thin layer chromatography (TLC). The silica gel areas corresponding to triacylglycerols (TAGs), diacylglycerols (DAGs), free fatty acids (FFAs), and polar lipids (PLs; primarily phosphatidylcholines (PCs) and phosphatidic acids (PAs)) were scraped and placed in scintillation vials with 7 ml of Omni‐Solv (DuPont) to assess the incorporation of radioactivity using a scintillation counter (LSC 1219 Rackbeta, LKB Wallac) as described previously (Taylor et al., 1992). In the case of immature seed feedings, two weeks after the first flower about 10 developing seeds were excised from three to four siliques of the primary stem, and pooled within the same developmental stage. Each pool of seeds was incubated in 2 µCi 3‐14C sodium pyruvate or 5 µCi U‐14C sucrose (American Radiolabeled Chemicals) in 500 µl HEPES pH 7.4, for 20 h in the light, at 25 °C, with shaking at 100 rpm, and washed three times with distilled water before preparing the TLE as described by Taylor et al. (1992). A portion of the TLE was transmethylated using 3 N methanolic HCl and the resulting fatty acid methyl esters (FAMEs) were redissolved in acetonitrile and separated using reverse‐phase radio‐HPLC analysis as described previously (Taylor et al., 1992). Respiration measurements Transgenic lines 3‐1‐J, 9‐5‐A, 10‐4‐C (constitutive expression) and the pBI121 plasmid‐only control, and transgenic lines 1‐1‐B, 5‐6‐E, 10‐2‐K (seed specific expression) and pRD400 plasmid‐only control were grown under environmental conditions favouring the full development of the leaves, i.e. short days (12 h light), in order to facilitate gas exchange measurements. Individual leaves or 25–30 fully immature siliques containing embryos at mid‐development (developmental stages 5 to 6 as described previously (Zou et al., 1996) were placed in a cuvette linked to a Li‐Cor (Lincoln, NE 68504) model LI‐6400 portable gas exchange system (open gas exchange system) for respiration and photosynthetic measurements. It was necessary to humidify the air going into the leaf cuvette for the Arabidopsis measurements to maintain a vapour pressure deficit (VPD) between the leaf and air of 1–1.2 kPa (at the temperature used for the measurements, this is equivalent to around 60% RH). The humidity was maintained by using a system involving two water baths where the incoming air from the growth chamber was first passed through a flask containing distilled water at 25 °C to add water to the air, and the air was then passed through a coil in a second water bath set at a lower temperature to give the appropriate water vapour pressure in the air. (The Arabidopsis leaves do not transpire enough to depend only on that water vapour source in the chamber.) Gas exchange measurements were performed after the stabilization of the CO2 exchange rate, i.e. 10–15 min after the leaf or siliques were placed in the cuvette. Respiration rates were determined in two ways. First, during the daylight cycle, the CO2 exchange rate was measured with the plant material in the cuvette exposed to normal light, and then again with the cuvette covered to prevent any exposure to light. Second, during the dark period, the plant material was previously in the dark for at least 2 h prior to measurements. In the latter case, the whole plant was already acclimated to dark conditions, whereas in the first case, only the measured leaf was placed in the dark. All gas exchange measurements were performed on eight sets of leaves and on four to six sets of siliques for each line studied. To calculate CO2 exchange rates, the leaf area in the cuvette was determined by tracing the leaf, and for inflorescences the portion of inflorescence measured was cut, dried at 60 °C, and the weight determined. Dry weights of leaves, inflorescences and roots were determined by drying the tissue at 60 °C for 3 d and then weighing the dried tissue. Results Studies of the tandem 35S:antisense AtPDHK transgenic lines In the present study, three of the constitutively repressed PDHK lines (3‐1‐J, 9‐5‐A and 10‐4‐C) were selected, as well as the non‐transformed wild type (n‐WT) and the plasmid only control (pBI121), and the respiration rates of leaves and siliques were measured using a Li‐Cor gas measuring system. As shown in Fig. 1A, there was a significant increase (up to 1.6‐fold) in dark respiration of the leaves of the constitutively‐repressed antisense PDHK transgenic lines compared to the non‐transformed wild type (n‐WT) and to the plasmid‐only (pBI121) transgenic control lines. There was also some indication that leaf photosynthetic rates were lower in these PDHK transgenic lines, but the differences were not statistically significant. Figure 1B shows that constitutively‐repressed antisense PDHK plants exhibited a significantly reduced leaf dry weight (40–60% of that of the controls). Root and inflorescence dry weights were reduced in the PDHK transgenics compared to both controls, most notably in transgenic line 9‐5‐A . It is important to note that the leaf measurements were performed on plants grown with 16 h light and 8 h dark. Thus the rates of photosynthesis can automatically be doubled to normalize the rates of photosynthesis versus respiration over time. In addition, the rates of respiration were made on plants in the first half of the dark period, and thus it is unlikely that the respiration rates measured occurred over the whole night period, since starch in the leaf can be depleted relatively quickly at night. Certainly the size of the leaves was greatly reduced in the constitutive antisense transgenics, indicating that the net difference between the average absolute values of photosynthesis and respiration in the two controls versus the three transgenics was slightly lower in the transgenics, as the data indicate. The increased dark respiration in leaves of the antisense PDHK transgenics was directly correlated with increased relative activities of mitochondrial PDC and of citrate synthase in the TCA cycle (Fig. 2). Since mtPDC represents the primary entry point of carbohydrates into the TCA cycle, it also plays an important role in modulating this respiratory process. In contrast to leaves, the difference in the rate of dark respiration in developing siliques (seeds) of the constitutively‐repressed antisense PDHK transgenic lines was not significantly different compared to the non‐transformed control or pBI121 plasmid‐only transgenic control lines (Fig. 3). When the mature seeds from these constitutively repressed PDHK lines were examined, two consistent trends were found (1) a significant increase in the accumulation of seed storage lipids on a per‐100‐seed basis (Fig. 4) and (2) an increase in the average 1000‐seed weight (Fig. 5). In order to study the link between respiratory flux, lipid deposition and sink size in more detail, and to facilitate the interpretation of data, antisense line 9‐5‐A was chosen because it had a single insert and was homozygous for the antisense trait. Transgenic line 9‐5‐A transformed with the constitutively‐expressed antisense PDHK construct exhibited increases of 13% and 20% in seed oil content and seed weight, respectively, in the T3 generation, compared with n‐WT and pBI121 control transformants. Furthermore, the average number of siliques per 15 cm segment of bolted stem (30±3) and the average number of seeds per silique (50±7) were not significantly different from the n‐WT and pBI121 control transgenics. Though it flowered earlier (Fig. 6A) and reached maturity sooner (68–70 d versus 75–77 d for pBI121 controls) (Zou et al., 1999), the inflorescence of the antisense PDHK line 9‐5‐A was as robust at maturity as the pBI121 and non‐transformed control plants (Fig. 6B). This indicated that seed yield was not adversely affected in the antisense PDHK transformants. In fact, the harvest index is increased in the constitutively‐repressed PDHK transgenics despite the fact that leaf and root dry weights are significantly reduced (Fig. 1A). The reduction in leaf dry weight is consistent with the fewer and smaller rosette leaves, as reported previously (Zou et al., 1999). To study carbon flux into seed storage lipids, bolted stems from antisense PDHK line 9‐5‐A and pBI121 control transgenics weighing 85–90 mg FW and containing 24–27 developing siliques were fed with 3‐14C pyruvate. The results clearly show that there was a general increase in radiolabel found in all lipid classes, but in particular, 3‐ and 4‐fold increases in the accumulation of radiolabelled polar lipids and TAGs, respectively (Fig. 7). Studies of the napin:antisense AtPDHK transgenic lines Once again, from a group of transgenics, three napin:antisense PDHK lines (1‐1‐B, 5‐6‐E and 10‐2‐K) were chosen together with the corresponding non‐transformed wild‐type control (n‐WT) and plasmid only transgenic control line pRD400, for more detailed studies. Physiologically, the antisense PDHK transgenic lines, where the construct was driven by a seed‐specific promoter (napin), were indistinguishable from the pRD400 and non‐transformed controls with respect to plant development: rosette leaf size and number, flowering, and generation time were unaffected (data not shown). While leaf respiration was not significantly different than the controls, respiration in the developing siliques of the napin:antisense PDHK transgenic lines was consistently higher (Fig. 8). In particular, lines 5‐6‐E and 10‐2‐K were statistically different at a 95% confidence level. Based on the absolute values for n‐WT, the three antisense PDHK lines 1‐1B, 5‐6‐E and 10‐2‐K showed rates of respiration of 0.010, 0.011 and 0.013 µmol CO2 s–1 g–1 DW greater than the n‐WT control. The connection to seed lipid deposition was able to be studied more directly by using transgenic lines wherein the repressed PDHK expression was seed‐specific. Once again, a significant increase was observed in the accumulation of seed storage lipids on a per 100‐seed basis (cf. Fig. 4) and an increase in the average 1000‐seed weight (cf. Fig. 5). Single insert homozygous line 5‐6‐E was chosen to be studied in more detail, having exhibited 50% and 25% increments in seed oil content and seed weight, respectively, in the T3 generation, compared with n‐WT controls and the pRD400 control transformants. Line 5‐6‐E developed identically to wild type and the pRD400 controls, with no visible vegetative or reproductive phenotype. Like the controls, line 5‐6‐E matured at about 77 d after planting; the number of siliques per plant and number of seeds per silique were also indistinguishable from those of n‐WT and pRD400 plants. Green siliques were harvested at specific developmental stages (as described by Zou et al., 1996) and fed with 3‐14C pyruvate. The results indicated that as the feedings were conducted on the less mature (stages 3 and 4) to the mid‐developing (stages 5 and 6) siliques, there was a progressive increase in the 14C label found in total lipid extracts of the transgenic line compared to the pRD400 control. At stages 3 and 4, the increase was the strongest in the PL and DAG fractions; by stages 5 and 6, the relative increase of radiolabel was most obvious in the TAG fraction of Nap‐5‐6 (Fig. 9). These gradual changes in the size of the labelled pools of each lipid class generally followed a progressive movement of label through the known Kennedy pathway intermediates and into TAGs, as the developing seeds matured. To measure the flux of carbon from pyruvate directly into developing embryos, without a contribution or effect of the maternal silique tissues, embryos were harvested from the various stages of developing siliques and the embryos incubated directly, with shaking in the presence of 3‐14C pyruvate for 20 h. Previous time‐course studies on the incorporation of labelled pyruvate and sucrose into components of immature seeds showed that over a period of 18 h, the incorporation rate was linear (Focks and Benning, 1998). Once again, the antisense napin:PDHK transgenic line exhibited a dramatic increase in label incorporation into seed lipids (radio‐specific activity of the lipid pool) at the mid‐development stages (5 and 6) compared to the corresponding lipid pool found in the pRD400 control embryos (Fig. 0). By stage 7, there continued to be a very strong increase in the size of the endogenous fatty acid pools. Line 5‐6‐E at stage 7 exhibited a 45% increase compared to stage 6, while the pRD400 line showed a smaller, but significant corresponding increase from stage 6 to stage 7 (30%). Thus, even though the activity in line 5‐6‐E is still about 2.5–3‐fold higher than in pRD400 seeds, the specific activity on a µmol basis is strongly reduced in both lines. A similar trend of increased movement of carbon into seed storage lipids was observed when the transgenic line 5‐6‐E was fed with radiolabelled sucrose, except that the accumulation pattern was typically maximized one to two stages behind that observed during direct pyruvate feeding (data not shown). This is not unexpected given that sucrose must first be broken down by glycolysis before providing pyruvate moieties for fatty acid biosynthesis and other biosynthetic processes. While the flux of acyl moieties into seed lipids was enhanced in the napin:antisense PDHK lines, the acyl composition of TAGs and other seed lipids in mature seeds was unchanged. For example, the proportions of saturates, mono‐unsaturates, polyunsaturates and very long chain fatty acids (C20 or >) were, repectively, 13.3±0.4, 34.2±0.7, 52.5±0.9, and 23.1±0.8 in the pRD400 controls versus 13.5±0.4, 35.8±1.0, 50.7±1.0, and 23.2±0.4 in the napin:antisense PDHK line 5‐6‐E. Discussion Biochemically, respiration can be said to involve glycolysis, the oxidative pentose phosphate pathway, the TCA cycle, and the mitochondrial electron transport system. In addition to generating energy and reducing equivalents, the respiration process also provides an array of intermediates (carbon skeletons) as the building blocks for many essential biosynthetic processes. Mitochondrial PDC acts as a link between glycolytic carbon metabolism and the TCA cycle. Because of the irreversible nature of the reaction, PDC is a particularly important site for metabolic regulation. Acetyl‐CoA, the product of PDC, is the starting point for the synthesis of fatty acids, isoprenoids and a number of secondary products. The present study is the first to demonstrate that partial silencing of PDHK can increase the metabolic flux through the reaction catalysed by mtPDC, based on measured increases in respiratory CO2 evolution and fatty acid production, in the anti‐sense PDHK transgenics. In a previous study (Zou et al., 1999), it was noted that flowering time could be hastened by reducing PDHK expression, and it was hypothesized that this was due to enhanced availability of mitochondrial acetyl‐CoA via increased pyruvate oxidation, and correlated with increased TCA cycle activity. These data supported earlier findings (Landschütze et al., 1995) suggesting that flowering time could be delayed by reduced TCA cycle activity and reduced respiration. This is in general agreement with the observation that an increase in the number of mitochondria and respiratory activity in the shoot apex upon induction of flowering appears to be causally linked to flower formation (Havelange, 1980). However, respiration rates had not been directly measured in the transgenic lines, and such measurements would be necessary to test the hypothesis. Collectively, the results in this and a previous study (Zou et al., 1999) where partial silencing of PDHK was achieved by an antisense approach, indicated that the antisense transgenics exhibited higher PDC activity, higher respiration rates, higher TCA enzyme activities and, in the case of the constitutively expressed antisense transgenics (e.g. 35S35S:antisense line 9‐5‐A), earlier flowering and shorter maturity times. Therefore, these data support the earlier conclusion by Landschütze et al. (1995), that the TCA cycle is of major importance during the transition from the vegetative to the generative phase and that the regulation of PDC activity directly affects mitochondrial respiration. As far as is known, the current results achieved by down‐regulating the expression of a regulator of mitochondrial respiration, rather than focusing on a specific catabolic enzyme in the respiratory pathway, are the first to support this hypothesis. The increase in respiratory CO2 evolution measured for the intact siliques of the seed‐specific anti‐sense PDHK trangenics can be explained by the theoretical respiratory CO2 evolution associated with increased seed oil production in these transformants. Using an average increase of 300 µg total fatty acids per 100 seeds as a basis for the calculation (cf. Fig. 4), and assuming that each triacylglycerol has an average chemical formula of C57H110O6 (MW=890; i.e. 3 C18 fatty acids attached to a glycerol backbone), it can be calculated that there is an increase of 336 nmol of triacylglycerol per 100 seeds or 1000 nmol of C18 fatty acids per 100 seeds. Since there is 1 pyruvate required per 2 carbons in the fatty acid skeleton, there are 9 mol of pyruvate required per C18 fatty acid, or approximately 9 µmol (1000 nmol×9) of additional pyruvate for the extra 300 µg of fatty acids. Since there is one CO2 released per pyruvate oxidized by the PDC, this is equivalent to an extra 9 µmol CO2 evolved per 300 µg fatty acids. In addition, the extra respiratory CO2 evolution expected in response to the NADPH and ATP requirements for this additional fatty acid synthesis can be calculated. For each C18 fatty acid synthesized, there are 16 NADPH required and approximately 19 ATP (1 ATP per conversion of acetyl‐CoA to malonyl Co‐A (8ATP per C18 fatty acid), 1 ATP per acetyl‐CoA shunted out of the mitochondria (9 ATP per C18), and 2 ATP per ester linkage between fatty acids and glycerol. To generate 16 NADPH equivalents and 19 ATP, considering glycolysis and then either the TCA cycle or the oxidative pentose phosphate pathway as sources of reducing equivalents and energy, would require approximately another 9 mol of CO2 released per mol C18 fatty acid formed (or another 9 µmol CO2 evolved per 300 µg fatty acids). Therefore, a total of 18 µmol CO2 would be evolved per 300 µg fatty acids formed. Since an average A. thaliana seed weighs 20 µg, 100 seeds is equivalent to 2 mg. Also, assuming that oil deposition in a seed occurs over a 4 week period (2.4 million seconds), the average rate of CO2 evolution would be 18 µmol CO2 per 2 mg per 2.4 million seconds or 0.004 µmol CO2 g–1 dry weight s–1. From Fig. 8, it can be calculated that the average increase in silique respiration rates for the seed‐specific PDHK transgenics compared to the nt‐WT control is 0.011 µmol CO2 s–1 g–1 dry weight. This calculation demonstrates that the increased respiratory rate can more than account for that required to support the increased oil accumulation in the developing seeds. Thus, greater respiration is occurring overall in the antisense PDHK transgenic seeds both for oil production and general respiration. Using isolated intact plastids, it has been shown that acetate is the preferred substrate for fatty acid synthesis, and there is evidence that a multienzyme system including acetyl‐CoA synthetase and acetyl‐CoA carboxylase, exists in plastids, which channels acetate into lipids (Roughan and Ohlrogge, 1996). It is almost certain that at least some of the acetyl‐CoA in plastids is formed by plastidic pyruvate dehydrogenase, using pyruvate imported from the cytosol or produced locally by plastidial glycolysis. A further possibility, especially in non‐photosynthetic tissues (e.g. roots and developing embryos), is that acetyl‐CoA, generated in the mitochondria, is an alternative means to provide acetate moieties for fatty acid synthesis (Ohlrogge and Browse, 1995). Mitochondrially‐generated acetyl‐CoA could be hydrolysed to yield free acetate, which could move into the plastid for conversion to acetyl‐CoA via plastidial acetyl‐CoA synthetase, an enzyme with 5‐15‐fold higher activity than the in vivo rate of fatty acid synthesis (Roughan and Ohlrogge, 1994). Liedvogel (1986) demonstrated the presence of a short‐chain acyl‐CoA hydrolase localized specifically in the matrix space of spinach mitochondria. It is inferred, therefore, that acetyl‐CoA in the mitochondria is hydrolysed by this enzyme to yield free acetate. Free acetate may move to the plastid for conversion to acetyl‐CoA by a plastidic acetyl‐CoA synthase, thereby contributing, in part, to the acetyl‐CoA pool required for fatty acid biosynthesis. Alternatively, the mitochondrial acetyl‐CoA could be converted to acetylcarnitine and transported directly into the plastid. Hence, in theory, the mitochondrial pyruvate dehydrogenase complex has an important role to play in fatty acid biosynthesis. One of our major interests in pursuing the characterization of the AtPDHK gene and the results of its partial silencing, has been to determine whether or not mtPDH plays a role in converting pyruvate to acetate moieties which may contribute to fatty acid biosynthesis in the primary lipid sink of oilseeds (developing embryos). Results in the present study support the hypothesis that mtPDH plays an important role in fatty acid biosynthesis in the developing embryos of oilseeds. As shown in the case of the constitutively‐silenced PDCK transgenics, silique respiration was not significantly increased, yet these lines still displayed a significant increase in lipid accumulation in the seeds. It is believed that this is because these lines have a much reduced leaf area and overall rate of photosynthesis; thus carbon supply to the silique is probably reduced in these plants. Therefore, the increased lipid level cannot be attributed to the increased level of NADH or ATP, resulting from an increased TCA activity and respiration rate (a secondary effect of the elevated PDC activity). Rather, the radiolabelling data support the hypothesis that the observed increase in seed lipid biosynthesis results from an increased supply of acetyl‐CoA from the mitochondria due to enhanced pyruvate decarboxylation. Given that there is some evidence that plants can sense and alter gene expression in response to the size of metabolic pools (e.g. organic acids; see Landschütze et al., 1995), it is possible that some of the effects observed in the present study might be secondary consequences of perturbed metabolite (e.g. organic acid) contents. Further studies are required to examine this possibility. In any case, these data provide evidence that alterations in PDC activity in the mitochondria can act as a signal to affect lipid biosynthesis in seeds by a mechanism not directly related to increased respiratory activity. To summarize, the collective data from the present studies conducted with transgenic lines where PDHK was partially silenced in both a constitutive and a seed‐specific manner support the hypothesis that enhanced mitochondrial PDC activity can result in an increased availability of acetyl moities from mitochondrial pyruvate turnover which are utilized in the synthesis of fatty acids and storage lipids in the developing seed. The proof of this hypothesis has been hindered until now. This question has been addressed directly using a transgenic approach. The antisense PDHK transgenics exhibit an increase in fatty acid biosynthesis and accumulation as well as a significant increase in the average seed weight of the progeny. When looking at the constitutive antisense PDHK transgenics, it is indeed interesting that, despite an altered vegetative phenotype (smaller leaf size and number) and a significant increase in leaf (but not silique) respiration, there was no penalty with respect to seed yield and oil deposition. On the contrary, both traits showed increases in the antisense PDHK transgenics. The same can be said with respect to the seed‐specific antisense PDHK transgenics, wherein fatty acid biosynthesis, lipid deposition and sink size (seed weight) were all increased over that observed in the seed of plasmid‐only transgenic controls (silique respiration was increased, although respiration in the vegetative tissues was not affected). These findings contrast with those observed in plants where sink tissues accumulate starch (e.g. potato tubers). Starch biosynthesis takes place exclusively in plastids that are the sole location for starch synthases and starch branching enzymes (Preiss, 1997). While current understanding of how sucrose to starch conversion is regulated in potato tubers is limited, some transgenic studies have shed some light on this process. Starch deposition has been altered by genetic engineering. By expression of a mutant E. coliglgC16 gene encoding an ADP glucose pyrophosphorylase in potato tubers, an increase in starch accumulation was observed (Stark et al., 1992). More recently, studies of potato tubers in which the expression of the plastidic adenylate transporter was down‐regulated, revealed that tubers contained drastically reduced starch content, starch grains of only about 50% normal size, and an increased level of soluble sugars, all of which was accompanied by a 2‐fold increase in the respiration rate observed in tuber slices and a large reduction in tuber volumetric weight (Tjaden et al., 1998; Geigenberger et al., 2001). In another study, it was shown that combined over‐expression of a yeast invertase and a glucokinase from Zhymomonas mobilis, in potato tubers resulted in a large increase in glycolytic metabolites, 2–3‐fold increases in the activities of key enzymes of respiratory pathways and 3–5‐fold increases in carbon dioxide production. This increased respiration resulted in a dramatic decrease in starch content. The double transgenic lines showed starch levels that were as little as 35% of that found in wild‐type tubers (Trethewey et al., 1998). The fundamental difference in productivity observed between the A. thaliana antisense PDHK transgenics (this study) and transgenically engineered potatoes may lie in the nature of the storage product accumulated in developing seed embryos and tubers, the respective sink organs, and also in the fact that the result of partial silencing of PDHK is an increase in acetate moieties, the building blocks for fatty acid synthesis. In the case of organs storing starch, there is a requirement for sucrose moieties for conversion to starch in the plastids. Any process which causes an unexpected turnover in fixed carbon to CO2 (e.g. higher respiration in tuber tissue) will detrimentally affect the proportion of sucrose destined for starch biosynthesis, and the size or volumetric weight of the sink (tuber). In contrast, A. thaliana, like other members of the Brassicaceae, accumulates oil (triacylglycerol) as the primary storage component in its seeds. It has been shown that in rapeseed embryos, there is an import of sucrose for the biosynthesis of starch which occurs very early in embryo development (at the ‘milky’ stage); however, this starch is quickly broken down to provide fixed carbon as acetate, for fatty acid and storage lipid bioassembly as the embryo reaches mid‐maturity (Kang and Rawsthorne, 1994). Indeed, mature Brassica seeds contain little or no detectable starch. Yet this initial import and metabolism of carbohydrate is essential for the eventual switchover to fatty acid biosynthesis and storage lipid deposition: Focks and Benning (1998) have shown that an A. thaliana mutant defective in the seed‐specific regulation of carbohydrate metabolism develop wrinkled seeds with drastically reduced oil content. In another study, overexpressing ADP glucose pyrophosphorylase in canola prolonged the period of starch accumulation at the expense of oil biosynthesis (Boddupalli et al., 1995). These studies, in which the mtPDHK is partially silenced, strongly support the hypothesis that if mitochondrial PDC activity is increased, the resulting enhanced conversion of pyruvate to acetate results in a net increment in the size of the mitochondrially‐generated acetate pool available for lipid biosynthesis. Thus the antisense PDHK transgenics exhibit both an increase in lipid deposition and sink size (seed weight). Besides the increased productivity observed as a result of PDHK silencing, there are other potentially important applications of this concept for crop improvement. Given that the generation time in Arabidopsis control plants is about 76 d under the growth conditions used in these studies, a 7 d earlier flowering and earlier maturing phenotype in the antisense PDHK plants represents a shortening of generation time by about 10%. Similar modification of flowering time to extend the geographical range of cultivation is an important goal for Brassica crops (Lagercrantz et al., 1996). In related Brassicaceae (e.g. canola), this would provide the advantage of an earlier harvest and permit more northerly cultivation (Murphy and Scarth, 1994). Late season frost damage in temperate climates (e.g. on the Canadian Prairies) could be avoided with cultivars that mature earlier, and this could also significantly alleviate problems associated with late‐season clearing of chlorophyll from the maturing seeds (which can lead to ‘green oil’ during processing and necessitate expensive bleaching steps). Yet another application of silencing the negative regulation of mitochondrial PDC may be to increase respiration in sink tissues (e.g. potato tubers transformed with an antisense PDHK construct under control of the strong tissue‐specific promoter, patatin B33; Trethewey et al., 1998), thereby reducing starch deposition while providing additional acetyl‐CoA moieties for the biosynthesis and deposition of plant‐produced bioplymers, including bioplastics (e.g. based on potatoes transformed with bacterial gene(s) allowing acetate to be re‐channelled into polyhydroxyalkanoates; (Padgette et al., 1997). There is still an important question that remains to be answered: during the partial silencing of the negative regulator of mtPDC, can an increased synthesis of fatty acids occur directly within the mitochondria and, if so, can these fatty acids be exported to the ER for triacylglycerol bioassembly? An answer to this question is important in the light of reports by other researchers (Wada et al., 1997; Gueguen et al., 2000), regarding the presence of fatty acid synthase machinery within the mitochondrial matrix, and the capacity for this organelle to make 16:0‐ACP and 18:0‐ACP, for example (Gueguen et al., 2000). However, given the logistics of extracting significant numbers of highly purified mitochondria to address this question, these experiments must await the production of antisense PDHK transgenic lines in a higher Brassica species, like B.napus canola, a project currently underway in this laboratory. Acknowledgements The authors gratefully acknowledge M Giblin, D Barton, D Reed, and Dr V Katavic for technical assistance, and D Schwab, B Panchuk and Dr L Pelcher of the PBI DNA Technology Group for primer synthesis and DNA sequencing. We thank Dr WA Keller for his critical review of this manuscript. This is National Research Council of Canada Publication No. 43919. Open in new tabDownload slide Fig. 1. (A) Gas exchange measurements. A comparison of leaf photosynthesis and dark respiration rates as measured in the constitutive anti‐sense transgenic lines 3‐1‐J, 9‐5‐A and 10‐4‐C compared with the non‐transformed wild‐type (n‐WT) control and the pBI121 (plasmid only) control transgenic line. Dark respiration and photosynthetic rates for leaves are depicted as negative and positive carbon assimilation, respectively. (B) Dry matter measurements. Organs were harvested and dried as described in Materials and methods. Values are the average of 6–8 determinations per control or transgenic line, ±SE. Open in new tabDownload slide Fig. 2. A comparison of relative leaf dark respiration rates with leaf mitochondrial PDC and citrate synthase (Cit Syn) activities measured in constitutive anti‐sense transgenic lines 3‐1‐J, 9‐5‐A and 10‐4‐C compared with the non‐transformed wild‐type (n‐WT) plants and to the pBI121 (plasmid only) control transgenic line. The rates for each parameter are depicted relative to the respective values measured in the leaves of the pBI121 transgenic control line, each set at 100%. Dark respiration rates are the average of 8 determinations per transgenic line, ±SE. PDC and citrate synthase activity data are the means of replicate experiments with three independent enriched mitochondrial preparations from each line. PDC and citrate synthase activities are presented ±SD, relative to the respective activities observed in the pBI121 control mitochondria (PDC: 890±57 nmol min–1 mg–1 protein; Citrate synthase: 1.6±0.1 µmol min–1 mg–1 protein), each set at 100%. Open in new tabDownload slide Fig. 3. A comparison of relative rates of dark respiration measured in siliques of the constitutive anti‐sense transgenic lines 3‐1‐J, 9‐5‐A and 10‐4‐C compared with the non‐transformed wild type (n‐WT) and to the pBI121 (plasmid only) control transgenic line. The relative silique dark respiration is depicted as a ratio (the rate of the transgenic/rate of the n‐WT)) with the n‐WT having a value of 1.00. The actual base value for the WT is 0.084 (µmol CO2 s–1 g–1 DW; (SE=0.007). Values are the means of two experiments, each comprising nine measurements ±SE, for each transgenic line. Open in new tabDownload slide Fig. 4. A comparison of lipid deposition (expressed as µg total fatty acids per 100 seeds) measured by by transmethylation of the TAG fraction from mature T3 seeds. On the left, the three constitutive antisense PDHK lines, 3‐1‐J, 9‐5‐A and 10‐4‐C are compared with the non‐transformed wild type (n‐WT) and to the plasmid only control (pBI121). On the right, the three seed‐specific antisense PDHK lines, 1‐1‐B, 5‐6‐E and 3‐1‐K are compared with the non‐transformed wild type (n‐WT) and to the corresponding plasmid only control (pRD400). Values represent the average of 3–4 measurements, ±SE. Arrows identify the single copy antisense transgenics (35S AS9‐5 and Nap AS5‐6) chosen for more detailed study as described in the Results. Open in new tabDownload slide Fig. 5. A comparison of the average 1000‐seed weights of mature T3 seeds. On the left, the three constitutive antisense PDHK lines 3‐1‐J, 9‐5‐A and 10‐4‐C, are compared with the non‐transformed wild type (n‐WT) and to the plasmid only control (pBI121). On the right, the three seed‐specific antisense PDHK lines 1‐1‐B, 5‐6‐E and 3‐1‐K, are compared with the non‐transformed wild type (n‐WT) and to the corresponding plasmid only control (pRD400). Values represent the average of 3–4 measurements (±SE), with the respective plasmid‐only control weights set at 100%. Open in new tabDownload slide Fig. 6. (A) Constitutive A. thaliana transgenic antisense PDHK line 9‐5 (left) and the plasmid‐only control pBI121 (right), at 35 d after planting (dap). The antisense transgenic is fully bolted with developing flowers and is well into the generative phase, while the control has not yet initiated flowering. (B) A comparison of the fully mature A. thaliana non‐transformed control (left) at 77 dap, the constitutive antisense PDHK transgenic line 9‐5 (middle) at 68 dap, and the pBI121 control transgenic line (right) at 77 dap. Open in new tabDownload slide Fig. 7. Results of feeding experiments in which bolted stems from constitutive antisense line 9‐5‐A and pBI121 control transgenics weighing 85–90 mg FW and containing 24–27 developing siliques were fed with 3‐14C pyruvate. A total lipid extract was prepared, lipid classes were resolved by TLC, and the regions scraped and counted as described previously (Taylor et al., 1992). Triplicate experiments were performed. While the trends were identical, the absolute total radiolabel incorporation values from one experiment to the next were not (a range from 6–60 nCi g–1 fresh weight of stem material was observed). Thus, data from one such experiment are depicted here with activity expressed as nCi g–1 fresh weight of stem material. Open in new tabDownload slide Fig. 8. A comparison of relative rates of dark respiration measured in siliques of the seed‐specific antisense transgenic lines 1‐1‐B, 5‐6‐E and 10‐2‐K in comparison to the non‐transformed wild type (n‐WT) and to the pRD400 (plasmid only) control transgenic line. The relative silique dark respiration is depicted as a ratio (the rate of the transgenic/rate of the n‐WT) with the n‐WT having a value of 1.00. The actual base value for the n‐WT is 0.069 µmol CO2 s–1 g–1 DW; SE=0.002. Values are the means of two experiments, each comprising nine measurements ±SE, for each transgenic line. Open in new tabDownload slide Fig. 9. Results of feeding experiments in which siliques from seed‐specific antisense line 5‐6‐E and pRD400 control transgenics were fed with 3‐14C pyruvate. Siliques were staged and selected as described by Zou et al. (1996). After 20 h of feeding time, the siliques were rinsed with distilled water and a total lipid extract was prepared. Lipid classes were resolved by TLC, and the regions scraped and counted as described previously (Taylor et al., 1992). Triplicate experiments were performed. While the trends were identical, the total absolute radiolabel incorporation values from one experiment to the next were not (a range from 100–400 nCi g–1 fresh weight of stem material was observed). Thus data from one such experiment are depicted here with activity expressed as nCi g–1 fresh weight of stem material. Open in new tabDownload slide Fig. 10. Results of feeding experiments in which staged developing seeds from seed‐specific antisense line 5‐6‐E and pRD400 control transgenics were harvested and fed with 3‐14C pyruvate. Siliques were staged and selected as described by Zou et al. (1996), the seeds harvested and pooled from each silique stage, and kept on ice before beginning the feedings. After a 20 hour feeding time, seeds were washed with distilled water, a total lipid extract was prepared containing an internal standard (15:0 FFA), lipid classes were resolved by TLC, and the regions scraped and counted as described previously (Taylor et al., 1992). At the same time, a portion of the TLE was retained and run on a vented GC to determine the total lipid pool size as described previously (Taylor et al., 1992; Katavic et al., 1995). Replicate experiments were performed. While the trends were identical, the total absolute radiolabel incorporation values from each experiment were not (ranges from 0.2–6 nCi µmol–1 total fatty acids were observed). Thus data from one experiment are depicted here with radio‐specific activity expressed as nCi µmol–1 total fatty acids) at each developing seed stage. References BoddupalliSS, Stark DM, Barry GF, Kishore GM. 1995 . Effect of overexpressing ADPGlc pyrophosphorylase on the oil biosynthesis in canola. In: Ohlrogge JB, Jaworski JG, eds. Proceedings of the biochemistry and molecular biology of plant fatty acids and glycerolipids symposium, June1–4, 1995 . S. Lake Tahoe, CA. National Plant Lipid Co‐operative, Abstract P‐102 . BuddeRJA, Fang TK, Randall DD. 1988 . Regulation of the phosphorylation of mitochondrial pyruvate dehydrogenase complex in situ. Plant Physiology 88, 1031 –1036 . BuddeRJA, Randall DD. 1990 . Pea leaf mitochondrial pyruvate dehydrogenase complex is inactivated in vivo in a light‐dependent manner. Proceedings of the National Academy of Sciences, USA 8, 673 –676. DatlaRSS, Hammerlindl JK, Panchuk B, Pelcher LE, Keller WA. 1992 . Modified binary plant transformation vectors with the wild‐type gene encoding NPTII. Gene 21, 383 –384. FocksN, Benning C. 1998 . wrinkled1: A novel, low‐seed‐oil mutant of Arabidopsis with a deficiency in the seed‐specific regulation of carbohydrate metabolism. Plant Physiology 118, 91 –101. GeigenbergerP, Stamme C, Tjaden J, Scultz A, Quick PW, Betsche T, Kersting HJ, Neuhaus HE. 2001 . Tuber physiology and properties of starch from tubers of transgenic potato plants with altered plastidic adenylate transporter activity. Plant Physiology 12, 1667 –1678. GrofCPL, Winning BM, Scaysbrook TP, Hill SA, Leaver CJ. 1995 . Mitochondrial pyruvate dehydrogenase: molecular cloning of the E1α subunit and expression analysis. Plant Physiology 108, 1623 –1629. GuanY, Rawsthorne S, Scofield G, Shaw P, Doonan J. 1995 . Cloning and characterization of a dihydrolipoamide acetyltransferase (E2) subunit of the pyruvate dehydrogenase complex from Arabidopsis thaliana. Journal of Biological Chemistry 270, 5412 –5417. GueguenV, Macherel D, Jaquinod, Douce R, Bourguignon J. 2000 . Fatty acid and lipoic acid biosynthesis in higher plant mitochondria. Journal of Biological Chemistry 275, 5016 –5025. HavelangeA. 1980 . The quantitative ultrastructure of the meristematic cells of Xanthium strumarium during the transition to flowering. American Journal of Botany 67, 1171 –1178. HillSA, Grof CPL, Bryce JH, Leaver CJ. 1992 . Regulation of mitochondrial function and biogenesis in cucumber (Cucumis sativus L.) cotyledons during early seedling growth. Plant Physiology 99, 60 –66. KangF, Rawsthorne S. 1994 . Starch and fatty acid synthesis in plastids from developing embryos of oilseed rape (Brassica napus L.). The Plant Journal 6, 795 –805. KatavicV, Haughn GW, Reed D, Martin M, Kunst L. 1994 . In planta transformation of Arabidopsis thaliana. Molecular and General Genetics 245, 363 –370. KatavicV, Reed DW, Taylor DC, Giblin EM, Barton DL, Zou J‐T, MacKenzie SL, Covello PS, Kunst L. 1995 . Alteration of fatty acid composition by an EMS‐induced mutation in Arabidopsis thaliana affecting diacylglycerol acyltransferase activity. Plant Physiology 108, 399 –409. LagercrantzU, Putterill J, Coupland G, Lydiate D. 1996 . Comparative mapping in Arabidopsis and Brassica, fine scale genome collinearity and congruence of genes controlling flowering. The Plant Journal 9, 13 –20. LandschützeV, Willmitzer L, Müller‐Röber B. 1995 . Inhibition of flower formation by antisense repression of mitochondrial citrate synthase in transgenic potato plants leads to a specific disintegration of the ovary tissues of flowers. EMBO Journal 14, 660 –666. LernmarkU, Gardeström P. 1994 . Distribution of pyruvate dehydrogenase complex activities between chloroplasts and mitochondria from leaves of different species. Plant Physiology 106, 1633 –1638. LiedvogelB. 1986 . Acetyl coenzyme A and isopentenyl diphosphate as lipid precursors in plant cells–biosynthesis and compartmentation. Journal of Plant Physiology 124, 211 –222. LuethyMH, Miernyk JA, Randall DD. 1994 . The nuclotide and deduced amino acid sequences of a cDNA encoding the E1 beta‐subunit of the Arabidopsis thaliana mitochondrial pyruvate dehydrogenase complex. Biochimica and Biophysica Acta 1187, 95 –98. LuethyMH, Miernyk JA, Randall DD. 1995 . The mitochondrial pyruvate dehydrogenase complex: nucleotide and deduced amino‐acid sequences of a cDNA encoding the Arabidopsis thaliana E1α‐subunit. Gene 164, 251 –254. MooneyBP, David NR, Thelen JJ, Miernyk JA, Randall DD. 2000 . Histidine modifying agents abolish pyruvate dehydrogenase kinase activity. Biochemical and Biophysical Research Communications 267, 500 –503. MurphyL, Scarth R. 1994 . Vernalization response in spring oilseed rape (Brassica napus L.) cultivars. The Canadian Journal of Plant Science 74, 275 –277. OhlroggeJ, Browse J. 1995 . Lipid biosynthesis. The Plant Cell 7, 957 –970. PadgetteSR, Gruys KJ, Mitsky TA, Tran M, Taylor NB, Slater SC, Kishore GM. 1997 . Strategies for production of polyhydroxyalkanoate polymers in plants. Plant Physiology 114, Supplement 3 (abstract 10003). PreissJ. 1997 . Manipulation of starch synthesis. In: Quick WP, Foyer CH, eds. A molecular approach to primary metabolism in plants. London: Taylor and Francis Publishers, 81 –103. RandallDD, Williams M, Rapp BJ. 1981 . Phosphorylation–dephosphorylation of pyruvate dehydrogenase complex from pea leaf mitochondria. Archives of Biochemistry and Biophysics 20, 437 –444. RoughanPG, Ohlrogge JB. 1994 . On the assay of acetyl‐CoA synthestase in chloroplasts and leaf extracts. Analytical Biochemistry 21, 77 –82. RoughanPG, Ohlrogge JB. 1996 . Evidence that isolated chloroplasts contain an integrated lipid‐synthesizing assembly that channels acetate into long‐chain fatty acids. Plant Physiology 110, 1239 –1247. SambrookJ, Fritsch EF, Maniatis T. 1989 . Molecular cloning, a laboratory manual. 2nd edn. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press. StarkDM, Timmerman KP, Barry GF, Preiss J, Kishore GM. 1992 . Regulation in the amount of starch in plant tissues by ADP‐glucose pyrophosphorylase. Science 258, 287 –292. StockJB, Ninfa AJ, Stock AM. 1989 . Protein phosphorylation and regulation of adaptive responses in bacteria. Microbiology Review 53, 450 –490. TaylorDC, Barton DL, Rioux KP, MacKenzie SL, Reed DW, Underhill EW, Pomeroy MK, Weber N. 1992 . Biosynthesis of acyl lipids containing very‐long chain fatty acids in microspore‐derived embryos of Brassica napus L. cv. Reston. Plant Physiology 99, 1609 –1618 ThelenJJ, Miernyk JA, Randall DD. 1998b. Nucleotide and deduced amino acid sequences of the pyruvate dehydrogenase kinase from Arabidopsis thaliana. Plant Physiology 118, 1533 . ThelenJJ, Miernyk JA, Randall DD. 1999 . Molecular cloning and expression analysis of the mitochondrial pyruvate dehydrogenase from maize. Plant Physiology 11, 635 –644. ThelenJJ, Miernyk JA, Randall DD. 2000 . Pyruvate dehydrogenase kinase from Arabidopsis thaliana: a protein histidine kinase that phosphorylated serine residues. Biochemical Journal 349, 195 –201. ThelenJJ, Muszynski MG, Miernyk JA, Randall DD. 1998a. Molecular analysis of two pyruvate dehydrogenase kinases from maize. Journal of Biological Chemistry 27, 26618 –26623. TretheweyRN, Geigenberger P, Riedel K, Mohammad‐Reza H, Sonnewald U, Stitt M, Riesmeier JW, Willmitzer L. 1998 . Combined expression of glucokinase and invertase in potato tubers leads to a dramatic reduction in starch accumulation and a stimulation of glycolysis. The Plant Journal 15, 109 –118. TjadenJ, Möhlmann T, Kampfenkel K, Henrichs G, Neuhaus HE. 1998 . Altered plastidic ATP/ADP‐transporter activity influences potato (Solanum tuberosum L.) tuber morphology, yield and composition of tuber starch. The Plant Journal 16, 531 –540. WadaH, Shintani, Ohlrogge J. 1997 . Why do mitochondria synthesize fatty acids? Evidence for involvement in lipoic acid production. Proceedings of the National Academy of Sciences, USA 94, 1591 –1596. ZouJ‐T, Brokx SJ, Taylor DC. 1996 . Cloning of a cDNA encoding the 21.2 kDa oleosin isoform from Arabidopsis thaliana and down‐regulation of its expression in a mutant defective in diacylglycerol acyltransferase activity. Plant Molecular Biology 31, 429 –433. ZouJ‐T, Katavic V, Giblin EM, Barton DL, MacKenzie SL, Keller WA, Hu X, Taylor DC. 1997 . Modification of seed oil content and acyl composition in Brassicaceae by expression of a yeast sn‐2 acyltransferase gene. The Plant Cell 9, 909 –923. ZouJ‐T, Qi Q, Katavic V, Marillia E‐F, Taylor DC. 1999 . Effects of antisense repression of an Arabidopsis thaliana pyruvate dehydrogenase kinase cDNA on plant development. Plant Molecular Biology 41, 837 –849. ZouJ‐T, Taylor DC. 1998 . Plant pyruvate dehydrogenase kinase gene. International Patent Application No. PCT/CA98/00096; Abstract published: 09/02/98; Priority date: 10/02/97.

Spolaore, Silvia; Trainotti, Livio; Pavanello, Anna; Casadoro, Giorgio

doi: 10.1093/jxb/erg025pmid: 12493854

Abstract Two endo‐β‐1,4‐glucanase (EGase; EC 3.2.1.4.) genes, highly expressed during ripening of the non‐climacteric strawberries (Fragaria×ananassa Duch. cv. Chandler), were isolated. Serial promoter deletions of both genes (i.e. FaEG1 and FaEG3) fused to GUS were transiently assayed in strawberry fruits by using a technique recently developed in this laboratory. Although differences were observed with the short fragments, GUS activity became comparable with the largest fragments of both promoters. The apparently similar strength of the two largest promoter fragments was in contrast with previous results of Northern analyses which demonstrated different transcripts amounts for the two genes. The inclusion of the 3′ flanking region of both genes in the transient assays showed that, in the case of FaEG3, the 3′ region had a down‐regulating effect on the expression of GUS, and this might account for the lower amount of FaEG3 mRNA usually observed in ripe fruits compared to that of FaEG1. Downstream instability elements might be involved in such down‐regulation. Key words: Endo‐β‐1,4‐glucanase genes, Fragaria× ananassa Duch., strawberry, transient promoter analysis. Received 11 June 2002; Accepted 3 September 2002 Introduction Higher plants endo‐β‐1,4‐glucanases (EGases; EC 3.2.1.4.), also known as ‘cellulases’, have long been studied in events where the separation of cells is required. Although such events are found in a wide range of physiological processes, particular attention has been given to the abscission of organs and the softening of fleshy fruits due to their economic relevance. As regards the involvement of endo‐β‐1,4‐glucanases in the softening of fruits, a good deal of work at the molecular level has involved, among others, the fruits of tomato (reviewed in Rose and Bennett, 1999), avocado (Fisher and Bennett, 1991; Abeles et al., 1992), pepper (Ferrarese et al., 1995; Harpster et al., 1997), and strawberry (Harpster et al., 1998; Manning, 1998; Llop‐Tous et al., 1999; Trainotti et al., 1999). Tomato has been, and still is, a model plant for a number of reasons. Among them is the fact that its berry is a climacteric fruit. This characteristic makes tomato interesting for studies of ripening since the appearance of the climacteric peak is usually considered to mark the onset of ripening. Accordingly, in their review of the cell wall modifications that occur in growing and ripening tomato fruits, Rose and Bennett (1999) made a distinction between genes encoding cell wall modifying enzymes that are expressed during fruit development and are up‐regulated by auxin, and genes encoding cell wall modifying enzymes that are expressed during ripening and are up‐regulated by ethylene. The absence of both respiratory and ethylene climacteric peaks to mark a possible boundary between growth and ripening makes the non‐climacteric fruits more intriguing. However, the pattern of endo‐β‐1,4‐glucanase gene expression in the non‐climacteric strawberry fruits seems partly to contradict the idea that in these fruits the passage from growth to ripening is smooth and gradual. In strawberry, two endo‐β‐1,4‐glucanase genes are significantly expressed in fruits (Llop‐Tous et al., 1999; Trainotti et al., 1999) and one of them (FaEG3) is also expressed in growing vegetative tissues (Trainotti et al., 1999). In fruits, FaEG3 is switched on at a very young stage and shows an increasing expression up to a maximum in red ripe fruits. In other words, its gradual expression in fruits seems to parallel the gradual evolution of fruits from growth to ripening. The second endo‐β‐1,4‐glucanase gene (FaEG1) is fruit‐specific and starts to be expressed at the stage of white fruits, that is when the fruits stop their growth and enter into the ripening phase. The well‐defined start of expression of FaEG1 suggests that some yet unknown signal might mark the beginning of ripening, as the climacteric peak does in the climacteric fruits. The different spatial and temporal expression pattern of the two strawberry endo‐β‐1,4‐glucanase genes makes them valuable. In particular, the fruit ripening specificity and the high rate of expression of FaEG1 make its promoter a good candidate for biotechnological uses that require a fruit and ripening specific expression of a given gene in strawberry. As regards FaEG3, the possibility that its early expression in fruits might lead to the production of oligosaccharides acting as signals for the subsequent start of ripening (Dunville and Fry, 2000) might offer a tool to modulate the timing of fruit ripening. In parallel with the good deal of work done with climacteric fruits, ripening related genes have mostly been studied in this type of fruits. In the present work, data are presented about the characterization of two genes encoding endo‐β‐1,4‐glucanases in the non‐climacteric strawberry. In particular, chimeric gene fusions of 5′‐ and 3′‐ flanking regions of both genes with a β‐glucuronidase (GUS) gene were prepared. The different constructs were used in transient expression assays in strawberry fruits, also treated with the auxin analogue NAA. Materials and methods Plant material and auxin treatment Strawberry fruits (Fragaria×ananassa Duch. cv. Chandler) were obtained from farms either near Verona or at Pergine (Trento). The different developmental stages of fruits were determined according to Huber (1984). Fruits that had to be used in the experiments with auxin were first dipped in an antifungal solution containing promycidon (0.06 g l–1). The synthetic auxin 1‐naphthalene acetic acid (NAA, 2 mM), and Silwet L‐77 (200 µl l–1) as surfactant, were sprayed on a pool of fruits every 12 h over a period of 48 h. Control fruits were treated in the same way but NAA was omitted. When not used immediately, fruits were quartered, frozen in liquid nitrogen and stored at –80 °C. Nucleic acids extraction When extracting DNA by using the Nucleon PhytoPure system (Amersham Pharmacia Biotech, England), RNA was obtained as a by‐product of the DNA extraction by means of an overnight precipitation in 2 M LiCl at 4 °C. Isolation of genomic clones, DNA sequencing and determination of the transcription start site The genomic clones were isolated from a strawberry (Fragaria×ananassa Duch. cv. Chandler) genomic library constructed by cloning Mbo I partially digested DNA into the Xho I (partially filled‐in) site of the Lambda FIX II vector (Stratagene, USA). The library was a gift from Dr Juan Munoz Blanco (University of Cordoba, Spain). The FaEG1 and FaEG3 cDNAs (Trainotti et al., 1999) were used as probes to screen the library following standard procedures (Sambrook et al., 1989). DNA from the purified lambda clones was extracted using a Qiagen kit (Qiagen, Germany), digested and, after electrophoresis and blotting, probed once more with either FaEG1 or FaEG3. The hybridizing bands were subcloned in the pGEM (Promega, USA) plasmid vector. DNA sequencing was performed at the CRIBI sequencing facility of the University of Padua using a PCR‐based dideoxynucleotide terminator protocol and an automated sequencer (Applied Biosystems 377). Sequences were determined on both strands using, as templates, plasmids containing inserts of different lengths prepared by progressive unidirectional deletions using the ‘Erase‐a‐Base’ system (Promega, USA) and, when necessary, chemically synthesized oligonucleotides. Sequence manipulations, analyses and alignments were performed using the ‘Lasergene’ software package (DNASTAR, USA). The transcription start site of both genes was determined using the ‘AMV Reverse Transcriptase Primer Extension System’ (Promega). Each reaction was carried out with 200 fmol µl–1 of 32P‐end‐labelled primer and 20 µg of total RNA from red fruits. A control without RNA was also included. Preparation of glucuronidase gene constructs Plasmids used for transformation experiments contained the GUS reporter gene interrupted by a plant intron described by Vancanneyt et al. (1990). This gene can be driven by the CaMV 35S promoter (35S) in plasmid p35SGUS‐INT or it can be used without promoter, as in plasmid pPR97 (Szabados et al., 1995), to carry out negative controls. Plasmid pPR97 was also used as the cloning vector for both the FaEG1 and FaEG3 promoter deletions. These promoter deletions were prepared by PCR using a high fidelity polymerase (Pfu, Stratagene). Five promoter fragments were prepared for each gene. In the case of FaEG1, the fragments were amplified from either lambda or plasmid DNA using oligo SS11 (5′‐TTTTTTTTCTCTCTCG TTTTTGCTGG‐3′, annealing from base 3213 to base 3188 of the antisense strand) at the 3′ end and either sequence‐specific or universal oligos (annealing to vector sequences) at the 5′. The resulting fragments contained 46 bp of the 5′ untranslated region (UTR) of the FaEG1 mRNA and different portions of the 5′ untranscribed region. These fragments were cloned into the vector pPR97, a few bases upstream of the starting ATG of the GUS gene. The resulting constructs were named to reflect both the length and the origin of the inserted promoter fragment. So pEG1 refers to the promoter fragments of the FaEG1 gene. pEG1‐30 contains 2980 bp, pEG1‐20 2113 bp, pEG1‐10 953 bp, pEG1‐4.6 462 bp, and pEG1‐2 216 bp of the 5′ untranscribed region of FaEG1. Constructs pEG1‐30‐3′ and 35S‐3′EG1 were obtained by replacing the NOS polyadenylation sequence present in constructs pEG1‐30 and 35SGUS‐INT with the 1281 bp of FaEG1 ranging from 6825 to 8105, and corresponding to the 3′ UTR plus 1028 bp of 3′ untranscribed region. Also in this case the fragments were prepared by means of PCR with appropriate restriction sites, useful for the subsequent cloning steps, added at the 5′ end of the oligos used for the amplification. To produce the five promoter fragments of gene FaEG3, the 3′ oligo was SS21 (5′‐ ACTAAAACACTGGTCTATACTA‐3′, annealing from base 3379 to base 3358 of the antisense strand). The resulting fragments contained 134 bp of the 5′ UTR of the FaEG3 mRNA and different portions of the 5′ untranscribed region. Also these fragments were cloned into the vector pPR97 and the resulting constructs were named pEG3‐n (i.e. promoter of gene FaEG3). In particular, pEG3‐30, pEG3‐20, pEG3‐14, pEG3‐8, and pEG3‐4 contained 3025, 2025, 1420, 814, and 490 bp of the FaEG3 untranscribed region, respectively. Constructs pEG3‐30‐3′ and 35S‐3′EG3 were obtained by replacing the NOS polyadenylation sequence present in constructs pEG3‐30 and 35SGUS‐INT with the 1458 bp of FaEG3 ranging from 6733 to 8190, and corresponding to the 3′ UTR plus 957 bp of 3′ untranscribed region. Also in this case the fragments were prepared by means of PCR with appropriate restriction sites, useful for the subsequent cloning steps, added at the 5′ end of the oligos used for the amplification. Analysis of transient gene expression A new protocol, developed in the laboratory (Spolaore et al., 2001), was used to test the strength of the different promoter deletions of the two strawberry endo‐β‐1,4‐glucanase genes. In brief, Agrobacterium cells were transformed with plasmids containing the constructs to be transiently assayed. Afterwards, a suspension of transformed Agrobacterium cells was evenly injected throughout the entire fruits by using a sterile 1 ml hypodermic syringe. After 2 d incubation at 22 °C with a 16 h light photoperiod, the injected fruits were frozen and ground in a mortar. The extracted proteins were then used to assay both GUS and luciferase activities. In the case of the auxin treatments, fruits were injected after 24 h from the first hormone treatment; thereafter, the Agrobacterium infiltrated fruits were treated with auxin every 12 h throughout the incubation period. Proteins for reporter gene assay were extracted and assayed as previously described (Spolaore et al., 2001). In order to compare values obtained in different infiltration experiments, the GUS activities measured were normalized to the corresponding luciferase activities, so the plotted values are expressed as nmol 4‐MU released min–1 pg–1 of luciferase. Results and discussion The two endo‐β‐1,4‐glucanase encoding cDNAs (FaEG1 and FaEG3) previously characterized in the laboratory (Trainotti et al., 1999) were used as probes to screen a strawberry genomic library. Two DNA fragments, each containing one of the two EGase genes, were the result of this screening. In the case of FaEG1, 8627 bp were sequenced, while 8215 were the bases sequenced for FaEG3. The sequences of the two genes are not shown here and are available in public databases with the following accession numbers: AJ414708 (FaEG3) and AJ414709 (FaEG1). The transcription start site of FaEG1 was determined by the 5′ primer extension technique and it was found to correspond to the T in position –46 from the initial ATG. One TATA box was found at position –26 from the transcription start site (–72 from the initial ATG). The transcription start site of FaEG3 corresponds to the C in position –134 from the initial ATG. No standard TATA box could be evidenced, though a putative one (TATATA) was found at –27 bp (relative to the transcription start site, –161 bp from the initial ATG). In order to evaluate the regulatory capacity of the two promoters, serial promoter fragments were fused to the GUS reporter gene. Pink strawberry fruits were used for these experiments since ripening continued after the transformation with Agrobacterium, and the fruits became red by the time of tissue sampling for the reporter assays. To this purpose, it had previously been shown that the transition from pink to red fruits corresponds to the maximum rate of expression for both genes (Llop‐Tous et al., 1999; Trainotti et al., 1999). Two different reporter genes were used, one of them (GUS‐INT, Vancanneyt et al., 1990) for the promoter analysis and the other one (LUC‐INT, Hanson et al, 1999) to check the transformation procedure. The different constructs and the results of the transient transformation assays are shown in Fig. 1. All the GUS values were normalized to the luciferase activity measured in the same protein extract. In the case of the FaEG1 promoter, whose fragments were named pEG1‐n (Fig. 1, upper panel), GUS activity just above the background started to be detected with the smallest fragment (pEG1‐2). Thereafter, in parallel with the increase in fragment size, the promoter strength increased steadily until about 1.0 kb (pEG1‐10). The next promoter fragment (pEG1‐20) caused a dramatic increase in GUS activity which appeared even higher than that obtained with the 35S promoter. Finally, a further increase in GUS expression was obtained with the largest promoter fragment (pEG1‐30). Though 462 bp were sufficient to drive a significant expression of GUS, the above analysis demonstrated that a particularly high efficiency was obtained when the promoter fragment changed from about 1.0 kb to about 2.0 kb. Accordingly, the latter region might contain elements for a positive regulation of the FaEG1 expression as observed for other genes (Montgomery et al., 1993; Nicholass et al., 1995; Atkinson et al., 1998). The lower panel of Fig. 1 shows both the fragments of the FaEG3 promoter (named pEG3‐n) and the results of the transient expression assays. GUS activity above the background was detected with the fragment of about 0.5 kb (pEG3‐4). Also in the case of the FaEG3 promoter a steady GUS increase parallelled the increase in fragment size. However, in this case a jump in GUS activity was observed when the promoter size changed from about 2.0 kb to about 3.0 kb, and this activity was higher than that obtained with the 35S promoter. This finding suggests that elements involved in a strong positive regulation should be located in the distal 1.0 kb of the analysed FaEG3 promoter, a situation different from that of the FaEG1 promoter where the maximum regulatory capacity was found by adding the two most distal fragments. With both promoters the strength of the largest 5′ fragments (about 3.0 kb) was greater than that of the constitutive 35S promoter. This result makes both EGase promoters useful for biotechnological applications aimed at the improvement of the strawberry fruit quality. However, the promoter of FaEG1 might be especially interesting since the expression of this gene is both fruit‐ and ripening‐specific, while FaEG3 is also expressed in young fruits and in growing vegetative tissues. A comparable regulatory capacity of the two largest promoter fragments seems to contrast with the amount of their related transcripts as determined by Northern analysis. In fact, previous research had shown that, in spite of an earlier start, expression of FaEG3 led to an amount of mRNA in red fruits which was apparently lower than that of the transcript related to FaEG1 (Llop‐Tous et al., 1999; Trainotti et al., 1999). A possible explanation for the discrepancy between Northern and promoter analyses could be obtained by including sequences at the 3′ of the genes in the analysis of both promoters. The 3′ flanking region of each EGase gene was also added to a 35S‐GUS‐INT construct in order to enucleate their own regulatory activity. While the 3′ sequence of FaEG1 had a minor effect on the expression of the reporter gene, the 3′ sequence of FaEG3 caused a marked decrease in the expression of GUS, and such a decrease was also shown with the 35S promoter (Fig. 1, upper and lower panels). Accordingly, the down‐ regulating effect of the FaEG3 3′ flanking region might explain the observed lower amount of the FaEG3 mRNA compared to that of FaEG1. It is known that genes can contain sequences able to affect the stability of mRNAs (Abler and Green, 1996). In particular, it has been shown that the 3′ UTRs of SAUR transcripts contain a sequence [i.e. the DST (downstream) element] able to function as an instability element. This element consists of three separated subdomains, with the second having the sequence ATAGAT highly conserved in all the described SAUR DST elements (Abler and Green, 1996). A sequence analysis of the 3′ UTRs of the two strawberry EGase genes provided some clues as to the down‐regulating effect shown by the FaEG3 3′ flanking region. The complete second subdomain ATAGAT and a partial third subdomain GTA were found in the FaEG3 3′ UTR, but not in the same region of the FaEG1 gene. The presence of possible instability elements suggests that the FaEG3 mRNA might be less stable than the FaEG1 encoded transcript, and this could explain the apparent down‐regulating effect of the 3′ flanking region of FaEG3. Of course, the authors are aware that other as yet unknown sequences might contribute to the different regulation of the two strawberry EGase genes. While apparently insensitive to ethylene (Abeles et al., 1992), in the non‐climacteric strawberry the ripening process can be delayed by auxin (Given et al., 1988), a hormone that has been shown to have a down‐regulating effect on the expression of a number of ripening‐related genes (Reddy and Poovaiah, 1990; Manning, 1994; Medina‐Escobar et al., 1997; Harpster et al., 1998; Trainotti et al., 1999). To ascertain whether the two isolated promoters contained some information involved in the hormonal control of both FaEG1 and FaEG3 gene expression, two different fragments for each promoter (pEG1‐10, pEG1‐30 and pEG3‐14, pEG3‐30, respectively) were used to drive expression of GUS in auxin‐treated strawberry fruits. The effect of auxin treatments was analysed in white fruits, a stage where the expression of both EGase genes is far from having reached a maximum, thus allowing possible changes in reporter gene activity to be easily detected following the hormonal treatment. In all the analysed situations the auxin analogue NAA induced a decrease in the amount of reporter activity (Fig. 2). Interestingly, the auxin down‐regulating effect is greater in percentage terms with the promoter of FaEG1, which is the gene more related to the ripening phase. The pattern of spatial expression of the two EGase genes was shown to overlap in fruits by using the tissue printing technique (Trainotti et al., 2000), while differences were found with regard to the pattern of temporal expression (Trainotti et al., 1999). White fruits were, therefore, used in experiments aimed at understanding whether regulatory elements involved in this stage‐specific expression might be present in the largest promoter fragments (i.e. pEG1‐30 and pEG3‐30, respectively) of the two genes. Contrary to the comparable strength shown by these two promoters in the transition from pink to red fruits (see Fig. 1, upper and lower panels), in white fruits pEG3‐30 showed a much greater strength than pEG1‐30 (Fig. 3). The much higher amount of GUS activity observed with the FaEG3 promoter is in accordance with the higher amount of the FaEG3 mRNA observed in white fruits. This result suggests that sequences controlling the temporal expression of the two EGase genes should be present in both pEG1‐30 and pEG3‐30 promoter fragments. Another possibility is that trans acting factors regulating the expression of FaEG1 might be available only at the ripening stage. Transient expression analyses showed a different behaviour of the two EGase promoters, and suggested that important regulatory elements might be located in the 1.0 kb regions spanning from 1.0–2.0 kb of the FaEG1 promoter and from 2.0–3.0 kb of the FaEG3 promoter, respectively. An informatic analysis of the two promoters was therefore performed (Hehl and Wingender, 2001). In the case of FaEG3, a number of putative domains involved in the binding of different transcription factors were found to be scattered throughout the promoter. However, in the most active region (i.e. from 2.0–3.0 kb) one DOF1 (DNA‐binding with One Finger 1) domain was found (Fig. 1, lower panel), and it is known that Dof proteins can be involved in a number of regulatory circuits in plants. For instance, it has been shown that the pumpkin Dof protein AOBP binds to the promoter of an abscorbate oxidase gene (Kisu et al., 1998; Shimofurutani et al., 1998), while the maize Dof1 and Dof2 proteins control the expression of genes involved in carbon metabolism (Yanagisawa and Sheen, 1998; Yanagisawa, 2000). Accordingly, by binding to a yet unknown Dof protein, the DOF1 binding domain of the most distal 1.0 kb of the FaEG3 promoter might be responsible for the observed high activity. The analysis of the FaEG1 promoter revealed that the most active promoter region (i.e. from 1.0–2.0 kb) contains important domains (Fig. 1, upper panel). Among them it is worth mentioning a domain for the binding of the AGL3 protein (Huang et al., 1995), and a G‐box binding site (Sablowski et al., 1994). It is also interesting to note the presence of a DOF2 binding domain at the 5′ end of the pEG1‐10 promoter fragment. Such a terminal localization might have made its possible binding to strawberry Dof proteins ineffective, while a binding and consequent effects on the transcription might have been effective in the case of the bigger pEG1‐20 promoter fragment. Acknowledgements We wish to thank Dr P Perini (Consorzio Verde Europa, Verona) and Dr P Faletti (Cooperativa S. Orsola, Pergine, Trento) for their gift of the plant material used in this research. Dr Juan Munoz Blanco (University of Cordoba, Spain) is thanked for the gift of the strawberry genomic library. Dr Pascal Ratet and Dr Peter Mergaert (ISV, CNRS Gif‐sur‐Yvette, France) are thanked for the gift of plasmids p35SGUSINT and pPR97. Dr N Gutterson (DNA Plant Technology, Oakland, California, USA) is thanked for the gift of plasmid pDNAP1136 containing the LUCint gene, while Dr Michael Schultze (Department of Biology, University of York, UK) is thanked for the gift of the vector pISV2678. This work has been financially supported by the European Union FAIR Program (contract FAIR‐CT97‐3005). Open in new tabDownload slide Fig. 1. GUS activity measured in protein extracts obtained from strawberries transformed at the pink stage with constructs harbouring different promoter deletions of FaEG1 (top) and FaEG3 (bottom). The different promoter fragments used in each construct are depicted on the left; the indicated bp length is meant from the starting ATG. Also putative cis elements are marked (G‐box; AGL3: box recognized by Agamous‐like proteins; Dof1 and Dof2: box recognized by Dof transcription factors). In the various experiments the GUS activity (expressed as nmol MU min–1 µg–1 protein) was normalized to the luciferase activity (expressed as pmol of luciferase in 1 µg of protein) measured in the same protein extract. The activities measured with the strawberry promoter fragments were expressed as a percentage of the activity obtained with the CaMV 35S promoter, set arbitrarily to 100%. EA: endogenous activity (i.e. GUS activity measured in fruits transformed only with a luciferase gene). All values are the average of four independent experiments. Bars represent standard errors. Open in new tabDownload slide Fig. 2. GUS activity measured in protein extracts obtained from strawberries transformed at the white stage with constructs pEG1‐10 and pEG1‐30, respectively, for FaEG1 and pEG3‐14 and pEG3‐30, respectively, for FaEG3. 36 h before transformation and throughout the subsequent incubation period, fruits have been treated with the auxin analogue naphthalene acetic acid (NAA). The GUS activity (expressed as nmolMU min–1 µg–1 protein) was normalized to luciferase activity (expressed as pmol of luciferase in 1 µg of protein) measured in the same protein extracts. All GUS activities were expressed as a percentage of the activity obtained with the CaMV 35S promoter, set arbitrarily to 100%. All values are the average of four independent experiments. Bars represent standard errors. Open in new tabDownload slide Fig. 3. GUS activity measured in proteins extracted from strawberries transformed at the white stage with different constructs (i.e. pEG1‐30 and pEG3‐30, respectively). The GUS activity (expressed as nmolMU min–1 µg–1 protein) was normalized to luciferase activity (expressed as pmol of luciferase in 1 µg of protein) measured in the same protein extract. Activities were expressed as a percentage of the activity obtained with the CaMV 35S promoter, set arbitrarily to 100%. EA: endogenous activity (i.e. GUS activity measured in fruits transformed only with a luciferase gene). All values are the average of four independent experiments. Bars represent standard errors. References AbelesFB, Morgan PW, Saltveit Jr ME. 1992 . Ethylene in plant biology, 2nd edn. San Diego: Academic Press. AblerML, Green PJ. 1996 . Control of mRNA stability in higher plants. Plant Molecular Biology 32, 63 –78. AtkinsonRG, Bolitho KM, Wright MA, Iturriagagoitia‐Bueno T, Reid SJ, Ross GS. 1998 . Apple ACC‐oxidase and polygalacturonase: ripening‐specific gene expression and pro moter analysis in transgenic tomato. Plant Molecular Biology 38, 449 –460. DunvilleJC, Fry SC. 2000 . Uronic acid‐containing oligo saccharins: their biosynthesis, degradation and signalling roles in non‐diseased plant tissues. Plant Physiology and Biochemistry 38, 125 –140. FerrareseL, Trainotti L, Moretto P, Polverino de Laureto P, Rascio N, Casadoro G. 1995 . Differential ethylene‐inducible expression of cellulase in pepper plants. Plant Molecular Biology 29, 735 –747. FisherRL, Bennett AB. 1991 . Role of cell wall hydrolases in fruit ripening. Annual Review of Plant Physiology and Plant Molecular Biology 42, 675 –703. GivenNK, Venis MA, Grierson D. 1988 . Hormonal regulation of ripening in the strawberry, a non‐climateric fruit. Planta 174, 402 –406. HansonB, Engler D, Moy Y, Newman B, Ralston E, Gutterson N. 1999 . A simple method to enrich an Agro bacterium‐transformed population for plants containing only T‐DNA sequences. The Plant Journal 19, 727 –734. HarpsterMH, Brummell DA, Dunsmuir P. 1998 . Expression analysis of a ripening‐specific, auxin‐repressed endo‐β‐1,4‐glucanase gene in strawberry. Plant Physiology 118, 1307 –1316. HarpsterMH, Lee KY, Dunsmuir P. 1997 . Isolation and characterization of a gene encoding endo‐β‐1,4‐glucanase from pepper (Capsicum annuum L.). Plant Molecular Biology 33, 47 –59. HehlR, Wingender E. 2001 . Database‐assisted promoter analysis. Trends in Plant Science 6, 251 –255. HuangH, Tudor M, Weiss CA, Hu Y, Ma H. 1995 . The Arabidopsis MADS‐box gene AGL3 is widely expressed and encodes a sequence‐specific DNA‐binding protein. Plant Molecular Biology 28, 549 –567. HuberDJ. 1984 . Strawberry fruit softening: the potential roles of polyuronides and hemicelluloses. Journal of Food Science 49, 1310 –1315. KisuY, Ono T, Shimofurutani N, Suzuki N, Esaka M. 1998 . Characterization and expression of a new class of zinc finger protein that binds to silencer region of ascorbate oxidase gene. Plant Cell Physiology 39, 1054 –1064. Llop‐TousI, Domínguez‐Puigjaner E, Palomer X, Vendrell M. 1999 . Characterization of two divergent endo‐β‐1,4‐glucanase cDNA clones highly expressed in the nonclimateric strawberry fruit. Plant Physiology 119, 1415 –1421. ManningK. 1994 . Changes in gene expression during strawberry fruit ripening and their regulation by auxin. Planta 194, 62 –68. ManningK. 1998 . Isolation of a set of ripening‐related genes from strawberry: their identification and possible relationship to fruit quality traits. Planta 205, 622 –631. Medina‐EscobarN, Cárdenas J, Moyano E, Caballero JL, Muñoz‐Blanco J. 1997 . Cloning, molecular characterization and expression pattern of a strawberry ripening‐specific cDNA with sequence homology to pectate lyase from higher plants. Plant Molecular Biology 34, 867 –877. MontgomeryJ, Pollard V, Deikmann J, Fischer RL. 1993 . Positive and negative regulatory regions control the spatial distribution of polygalacturonase transcription in tomato fruit pericarp. The Plant Cell 5, 1049 –1062. NicholassFJ, Smith CJS, Schuch W, Bird CR, Grierson D. 1995 . High levels of ripening‐specific reporter gene expression directed by tomato fruit polygalacturonase gene‐flanking regions. Plant Molecular Biology 28, 423 –435. ReddyASN, Poovaiah BV. 1990 . Molecular cloning and sequencing of a cDNA for an auxin‐repressed mRNA: correlation between fruit growth and repression of the auxin‐regulated gene. Plant Molecular Biology 14, 127 –136. RoseJKC, Bennett AB. 1999 . Cooperative disassembly of the cellulose‐xyloglucan network of plant cell walls: parallels between cell expansion and fruit ripening. Trends in Plant Science 4, 176 –183. SablowskiRWM, Moyano E, Culianez‐Macia FA, Schuch W, Martin C, Bevan MA. 1994 . A flower‐specific Myb protein activates transcription of phenylpropanoid biosynthetic genes. EMBO Journal 13, 128 –137. SambrookJ, Fritsch EF, Maniatis T. 1989 . Molecular cloning: a laboratory manual, 2nd edn. New York: Cold Spring Harbor Laboratory Press. ShimofurutaniN, Kisu Y, Suzuki M, Esaka M. 1998 . Functional analyses of the Dof domain, a zinc finger DNA‐binding domain, in a pumpkin DNA‐binding protein AOBP. FEBS Letters 430, 251 –256. SpolaoreS, Trainotti L, Casadoro G. 2001 . A simple protocol for transient gene expression in ripe fleshy fruit mediated by Agrobacterium. Journal of Experimental Botany 52, 845 –850. SzabadosL, Charrier B, Kondorosi A, de Bruijn FJ, Ratet P. 1995 . New plant promoter and enhancer testing vectors. Molecular Breeding 1, 419 –423. TrainottiL, Spolaore S, Pavanello A, Baldan B, Casadoro G. 1999 . A novel E‐type endo‐β‐1,4‐glucanase with a putative cellulose‐binding domain is highly expressed in ripening strawberry fruits. Plant Molecular Biology 40, 323 –332. TrainottiL, Spolaore S, Pavanello A, Casadoro G. 2000 . Molecular aspects of the strawberry fruit softening. In: de Vries GE, Metzlaff K, eds. Phytosfere ’99—highlights in European plant biotechnology research and technology transfer. 133 –139. VancanneytG, Schmidt R, O’Connor‐Sanchez A, Willmitzer L, Rocha‐Sosa M. 1990 . Construction of an intron‐containing marker gene: splicing of the intron in transgenic plants and its use in monitoring early events in Agrobacterium‐mediated plant transformation. Molecular and General Genetics 220, 245 –250. YanagisawaS. 2000 . Dof1 and Dof2 transcription factors are associated with expression of multiple genes involved in carbon metabolism in maize. The Plant Journal 21, 281 –288. YanagisawaS, Sheen J. 1998 . Involvement of maize Dof zinc finger proteins in tissue‐specific and light‐regulated gene expression. The Plant Cell 10, 75 –99. © 2003