Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Infection of Arabidopsis thaliana leaves with Albugo candida (white blister rust) causes a reprogramming of host metabolism

Infection of Arabidopsis thaliana leaves with Albugo candida (white blister rust) causes a... Introduction Biotrophic fungi are agriculturally important pathogens often causing severe losses in yield. These pathogens are not thought to produce toxins and it is likely that they lower the yield of their hosts by two major processes. Firstly, they often lower the rate of photosynthesis and hence the production of carbon by the host and secondly they compete very effectively for host carbon; they are an effective sink. This acquisition of host carbon by the fungus results in a dynamic interaction between the physiology of the host and that of the parasite and to a reprogramming of host metabolism. Many rust and powdery mildew fungi cause a progressive decline in the rate of photosynthesis, an increase in invertase activity and an accumulation of hexose sugars (and sometimes sucrose) in infected leaves (for reviews see Farrar and Lewis, 1987 ; Scholes, 1992 ; Whipps and Lewis, 1981 ). Recently a model linking these metabolic symptoms has been proposed ( Scholes , 1994 ; Scholes, 1992 ; Tang , 1996 ). In this model, an increase in invertase activity leads to an accumulation of soluble carbohydrates, and/or to a change in the flux through these metabolic pathways. This in turn initiates a signal transduction pathway(s) leading to a repression of photosynthetic gene expression and hence to a decline in the rate of photosynthesis in infected leaves. Although it is clear that sugars control the expression of many plant genes ( Koch, 1996 ) much less is known about the precise mechanisms by which plants sense sugars. Currently there is evidence for the existence of a hexokinase‐dependent sugar sensing mechanism ( Jang , 1997 ; Sheen, 1994 ), a hexose‐transport associated sensor and a sucrose‐specific sensing pathway (reviewed by Smeekens and Rook, 1997 ). Although the content of hexose sugars and invertase activity are known to increase in leaves infected by many different biotrophic pathogens, a number of questions concerning the sugar sensing hypothesis remain unanswered. For example, is the increase in invertase activity of host or fungal origin? Is it confined to regions of the leaf invaded by fungal mycelium or is the activity induced systemically throughout the infected leaf? When and where does invertase activity increase in relation to changes in the amount or fluxes of hexose sugars? Is there a repression of photosynthetic gene expression in regions of the leaf where hexose sugars accumulate and over a time scale which is consistent with the hypothesis? What is the evidence that sugar sensing pathways are an important component of a compatible plant–pathogen interaction? A number of defence‐related genes are inducible by soluble sugars ( Herbers and Sonnewald, 1998 ; Herbers , 1996 ; Johnson and Ryan, 1990 ) suggesting that sugars play a role not only in the repression of photosynthetic genes but also in the induction of defence responses. Herbers . (1996) showed that, in transgenic tobacco plants over‐expressing yeast invertase in the apoplast, defence‐related gene transcripts, callose content, peroxidase activities and salicylic acid were all elevated. The accumulation of pathogenesis‐related proteins and the repression of photosynthetic gene transcripts were inversely correlated, and required the same threshold level of hexoses for both induction and repression. This raises the question of whether defence–related proteins are induced to any extent, at any stage, during a compatible host–pathogen interaction and, if not, whether the pathogen actively represses the induction of defence proteins. This study aims to examine the effect of Albugo candida (white blister rust) on the regulation of photosynthesis, carbohydrate metabolism and gene expression in discrete regions of A. thaliana leaves during a compatible interaction. Specifically we: (i) examine the spatial changes in the rate of photosynthesis in an infected leaf as the disease develops by quantitative imaging of chlorophyll fluorescence, (ii) quantify changes in the amounts of soluble and storage carbohydrates in different regions of the infected leaf as the disease progresses, (iii) quantify changes in the activity of wall‐bound and soluble invertases and in mRNA levels of the host wall‐bound invertase, and (iv) examine changes in the expression of genes encoding photosynthetic and defence proteins. The results are discussed in relation to the hypothesis and questions outlined above. Results The effect of A. candida on photosynthesis in different regions of an A. thaliana leaf Images of chlorophyll fluorescence were captured at a range of irradiances from healthy and infected plants 6, 8, 11 and 14 DAI (days after inoculation). Carbon dioxide assimilation was measured simultaneously using an infra‐red gas analyser (IRGA). To determine whether Ø II could be used to assess CO 2 assimilation, the product of Ø II and irradiance determined from an entire leaf was plotted against the assimilation rate simultaneously measured by gas exchange ( Fig. 1 ). This plot was linear for both healthy and infected leaves for all irradiances examined (ranging from 0 to 1045 µmol/m 2 /s). Therefore the rate of photosynthesis could be accurately determined in different regions of an infected leaf using chlorophyll fluorescence imaging. 1 The relationship between the rate of CO 2 assimilation determined by infra‐red gas analysis and Ø II × irradiance determined by analysis of chlorophyll fluorescence images. The two measurements were linearly related for healthy (○) and A. candida ‐infected (▪) leaves. The solid line represents a linear regression ( r 2 = 0.926, gradient = 20.4). Figure 2A shows the development of visible symptoms for representative leaves used for imaging. At least two leaves were imaged for each time‐point. Leaves were inoculated towards the tip of the adaxial surface with four 10 µL droplets of zoospore suspension. The region of the leaf inoculated with zoospores was termed region A, the uninoculated region of the same leaf, region B and similar regions of a control leaf, C and D. Fungal mycelium was confined to region A of the infected leaf. Six DAI, visible symptoms were not apparent, nor were symptoms apparent on images obtained by chlorophyll fluorescence analysis (data not shown). Spores first became visible on the lower surface of the leaf 8 DAI, and developed to cover approximately 30% of the leaf by day 11. No symptoms were apparent on the upper surface of the leaf 8 DAI. By 11 days a slight chlorosis was visible in the infected region which became much more pronounced by day 14. 2 Images of Ø II and nonphotochemical quenching (ΔFm/Fm′) for healthy and A. candida ‐infected leaves of A. thaliana . (A) Images of Ø II and ΔFm/Fm′ were calculated from measurements of chlorophyll fluorescence made after photosynthesis had reached steady‐state (1260 s) at an actinic irradiance of 380 µmol/m 2 /s and ambient CO 2 (350 µmol/mol). Upper two rows: photographs of the upper and lower surfaces of control and infected leaves (8, 11 and 14 DAI) used for imaging. Middle row: images of Ø II (and calculated assimilation rate derived from Fig. 1 ). The scales show the relationship between colour and the relevant chlorophyll fluorescence measurement. Bottom row; images of nonphotochemical energy dissipation (ΔFm/Fm′). (B) Images of Ø II and ΔFm/Fm′ are shown at selected time points throughout the induction of photosynthesis for control and A. candida ‐infected leaves 11 DAI. Times are indicated in seconds after the actinic light was first switched on. Induction was performed at low irradiance (93 µmol/m 2 /s) and ambient CO 2 (350 µmol/mol). The first two rows show images of Ø II and ΔFm/Fm′ for a control leaf and the bottom two rows for an infected leaf. The scales show the relationship between colour and the relevant chlorophyll fluorescence measurement. Representative images of Ø II and nonphotochemical quenching (NPQ), taken when photosynthesis had reached a steady‐state value at an actinic irradiance of 380 µmol/m 2 /s are shown in Fig. 2A . A slight reduction in Ø II was first apparent 8 DAI and was limited to the inoculated area (region A) where spores were visible on the lower leaf surface. This reduction became more widespread by day 11 and by day 14 photosynthesis was severely inhibited. In all cases there was a close correlation between the area where Ø II was reduced and the visible symptoms on the leaf; there was little impact on uninvaded regions of the leaf except at day 14. NPQ was elevated in regions where Ø II was reduced, indicating that light energy which was not used for photosynthetic electron transport was being lost by nonphotochemical processes. Images of Fv/Fm were similar in both control and infected leaves (data not shown) indicating that A. candida had not caused photodamage in any region of the leaf. Average values of Ø II and NPQ were extracted from different regions of images of control and infected leaves at irradiances from 0 to 1450 µmol/m 2 /s and are shown in Fig. 3 with corresponding measurements of whole‐leaf gas exchange. Results are shown for leaves 8, 11 and 14 DAI; there was no difference between control and infected leaves before this time. The rate of photosynthesis (measured by gas exchange or imaging) remained constant in control leaves throughout the duration of the experiment. In contrast, gas exchange measurements showed that there was a progressive decline in CO 2 assimilation by infected leaves, such that by day 14 the rate was only 30% of the control ( Fig. 3A–C ; upper panel). Images of Ø II showed that the reduction in photosynthesis was largely limited to the region invaded by the fungal mycelium and became more pronounced as the disease developed ( Fig. 3B,C ; middle panel). Only on day 14 was photosynthesis also reduced in the uninvaded region of the infected leaf. The decline in Ø II in region A was apparent at all irradiances except the lowest (93 µmol/m 2 /s). 3 The effect of A. candida on the rate of CO 2 assimilation, Ø II × irradiance and NPQ (ΔFm/Fm′) in localized regions of leaves of A. thaliana (A) 8; (B) 11 and (C) 14 DAI. Control leaf (○); whole infected leaf (▪); infected region A (▴) and uninfected region B (◆) of an infected leaf. Nonphotochemical quenching did not differ between control and infected leaves 8 DAI ( Fig. 3A ; lower panel). At low irradiances, NPQ was low (0.5) and it increased to a maximum of 2.5 as the irradiance increased ( Fig. 3A ; lower panel). A similar pattern was observed in control leaves 11 and 14 DAI. In infected region A, 11 and 14 DAI, NPQ was slightly reduced at low irradiances compared to uninfected region B or control leaves ( Fig. 3B,C ; lower panel). However, as the irradiance increased, NPQ increased much more rapidly in infected region A. The maximal value of NPQ observed at high irradiance 11 DAI was the same as that observed in control leaves but was significantly higher (3.2 compared with 2.5) by day 14. The NPQ of uninfected regions was also increased 14 DAI compared with control leaves but the increase was smaller than that observed in the infected region. To examine the effect of A. candida on photosynthetic induction, leaves were illuminated at an irradiance of 93 µmol/m 2 /s for 1260 s. Images of Ø II and NPQ were captured during this induction process and representative images of selected time‐points are shown in Fig. 2B . In contrast to measurements taken at steady state, during photosynthetic induction photosynthetic electron transport through photosystem II (and hence Ø II ) is not strictly proportional to the rate of CO 2 assimilation (data not shown) most probably due to the many different processes being activated during this period (generation of intermediates in the Calvin cycle, Rubisco activation, stomatal opening and generation of a transthylakoid pH gradient). In control leaves, Ø II increased rapidly and reached a steady‐state value of 0.8 after 500 s. This corresponded to a steady‐state rate of CO 2 assimilation of approximately 3.5 µmol/m 2 /s. In infected leaves, the steady‐state value of Ø II was not significantly different from that of control leaves but the pattern observed during induction was more complex. The infected leaf showed three distinct regions. In the infected region, where spores were visible on the underside of the leaf, Ø II was induced more slowly than other parts of the leaf. Surrounding this was a zone 3–4 mm wide, where Ø II was induced very rapidly when compared with both the control leaf and other regions of the infected leaf. The uninfected part of the leaf behaved in the same way as the control leaf. The induction of NPQ was also complex. In the control leaf, NPQ exhibited an initial transient increase but then declined to a uniform low value. This transient increase in NPQ was greater in infected region A but the final value of NPQ attained was lower. The induction of NPQ in uninfected region B was similar to that of the control leaves, although there was a slight difference in the time taken to reach a steady‐state value. Chlorophyll content of different regions of A. candida ‐infected leaves The chlorophyll content of all regions of infected and uninfected leaves was similar until 12 days after inoculation. Thereafter both chlorophyll a and b were lost from fungally invaded regions of the leaf (A) such that by day 14 the total chlorophyll content was 40% ( P < 0.001) of that in equivalent region C of a control leaf (310.1 ± 13.3 and 125.1 ± 10.3 mg/m 2 , respectively). The effect of A. candida on soluble carbohydrates and starch in localized regions of A. thaliana leaves Figure 4 shows the amount of sucrose, glucose and fructose in different regions of control and A. candida ‐infected leaves throughout the infection cycle. There was a significant increase in the amount of all soluble carbohydrates in infected region A when compared with the equivalent control region C from 6 DAI. The accumulation of carbohydrates was most pronounced on day 14. There was no increase in the soluble carbohydrate content of the uninfected region (B) of the infected leaf until day 14. 4 The effect of A. candida on the amount of soluble carbohydrates in localized regions of leaves of A. thaliana throughout the course of infection. Infected leaf; regions A (▴) and B (◆). Control leaf; regions C (○) and D (□). The results are the means ± the standard error of six replicate measurements. *, ** and *** indicate that data are significantly different, P < 0.05, 0.01 and 0.001, respectively. To examine the distribution and amount of starch, leaves were either cleared and stained for starch with I–KI solution, or analysed for starch content enzymically. Figure 5A shows control and infected leaves which have been stained for starch 8, 11 and 14 DAI and Fig. 5B shows a scan of the intensity of the black stain along a 2.5‐cm transect of the leaf starting from the middle of a pustule (or equivalent region) and extending longitudinally down the leaf. Eight DAI starch staining of leaves showed that the distribution of starch within the infected leaf was relatively uniform ( Fig. 5 B, i). Although the starch content appeared slightly higher in the infected than in the control leaf ( Fig. 5 B, i) enzymatic analysis revealed little difference between the content of either tissue ( Fig. 5C ). By day 11 it was apparent that starch had been lost from the infected region of leaves. In addition, the uninfected region of the infected leaf appeared to contain more starch than control leaves and in some leaves it was clear that there was a darker staining ring approximately 1–2 mm in width immediately around the pustule ( Fig. 5A,B ii). These results were confirmed by starch assay in different regions of the leaves ( Fig. 5C ). By day 14 the amount of starch in infected regions of the leaf was lower than comparable regions of control leaves and in uninfected regions of infected leaves it was higher than comparable regions of control leaves ( Fig. 5B , iii and C.). 5 The effect of A. candida on the starch content of localized regions of leaves of A. thaliana throughout the course of infection. (A) Leaves of A. thaliana were cleared and stained for starch. Left column control leaves; right column A. candida ‐infected leaves (i) 8 (ii) 11 and (iii) 14 DAI. (B) The intensity of the stain was measured along a 2.5‐cm transect positioned as shown in (A). Control leaf (‐ ‐ ‐); infected leaf (—). (C) Starch was extracted from different regions of control and infected leaves and quantified by enzyme‐linked assay. Infected leaf; regions A (▴) and B (◆). Control leaf; regions C (○) and D (□). The results are the means ± the standard error of five replicate measurements. The effect of A. candida on invertase activity, isoform pattern and gene expression in localized regions of A. thaliana leaves Figure 6B,C shows the changes in activity and isoform pattern of soluble invertases in infected and control A. thaliana leaves. There was an increase in the activity of soluble invertase in region A from 9 DAI but no change in region B when compared to the control leaf. Up until 6 DAI the soluble invertase isoform pattern was the same in infected and control regions; there were five major isoforms in each region ( Fig. 6C ). From day 10 onwards a new isoform with a pI of 5.1 appeared in the infected region of the leaf. The activity of the new isoform increased gradually as the fungus developed and the activity of the other five isoforms decreased as the leaves aged suggesting that it was of fungal origin ( Fig. 6A–D .). 6 The effect of A. candida on invertase activity in localized regions of leaves of A. thaliana . Samples for analysis of (A) wall‐bound and (B) soluble invertases were harvested from infected leaves, regions A (▴) and B (◆) and control leaves, regions C (○) and D (□) throughout the timecourse of infection. The results are the means ± the standard error of six replicate measurements. * and *** indicate that data are significantly different, P < 0.05 and 0.001, respectively. (C) Soluble invertase isoforms in different regions of A. thaliana leaves infected with A. candida . Soluble invertase isoforms were separated by isoelectric focusing over a pH range of 4.0–6.5. Invertase activity was visualized by staining reducing sugars with 2, 3, 5 tetrazolium chloride. Samples were prepared from infected region A, uninfected region B and equivalent control regions C and D, 0, 6, 10, 14 and 19 DAI. Samples were loaded on an equal leaf area basis. Table 1 shows the expression of the ATβFRUCT 1 apoplastic invertase gene in different regions of an A. thaliana leaf following infection by A. candida . An increase in gene expression was first detected in region A 8 DAI and by day 13 was approximately 40‐fold greater than in the control leaf. There was no change in gene expression in the uninfected region of the infected leaf. The activity of the apoplastic invertase mirrored the change in gene expression; activity was increased in region A from 6 DAI and by day 14 was fourfold greater than in the control leaf ( Fig. 6A ). Again, there was no increase in apoplastic invertase activity in region B of the infected leaf. 1 Expression of the ATβFRUCT 1 apoplastic invertase gene in leaves of A. thaliana infected with A. candida . The amount of expression in regions A and B of infected leaves and C and D of control leaves was analysed throughout the infection cycle using semiquantitative RT‐PCR. Results are expressed as pg/µg total RNA. Data are shown as the mean ± SE of three measurements. Infected leaf (pg/µg total RNA) Control leaf (pg/µg total RNA) Days after inoculation Region A Region B Region C Region D 4 4.5 ± 0.39 10.0 ± 3.7 < 2.0 15.9 ± 2.0 6 6.5 ± 1.7 5.6 ± 0.9 < 2.0 < 2.0 8 56.5 ± 3.6 17.6 ± 4.4 6.4 ± 0.48 14.6 ± 0.8 10 68.7 ± 8.8 2.2 ± 2.0 11.9 ± 2.0 11.9 ± 0.3 13 436.1 ± 105.7 15.0 ± 1.5 < 2.0 < 2.0 The effect of A. candida on the expression of genes encoding photosynthetic and defence proteins in A. thaliana leaves The expression of photosynthetic and defence proteins was examined by Northern blot analysis of total RNA from control and A. candida ‐infected leaves. Whereas RNA prepared from control leaves is exclusively of plant origin, infected leaves contain both plant and fungal RNA. The internal transcribed spacer (ITS) probes from A. thaliana and A. candida were used to determine the relative amounts of plant and fungal RNA in the different samples. The plant and fungal ITS probes were completely specific ( Fig. 7A ). The amount of plant RNA in all samples was similar. Four DAI, fungal RNA was present in the infected leaf but this had increased substantially as the fungus developed. 7 Northern hybridizations showing the expression of genes encoding (A) defence proteins and (B) photosynthetic proteins in leaves of A. thaliana infected with A. candida . Species‐specific ITS probes were employed to determine the proportion of RNA of the plant and fungus present in infected leaves. The use of the A. thaliana ITS probe showed equal loading of A. thaliana RNA. The expression of the gene encoding acidic chitinase was not detectable in either control or infected leaves (data not shown). The expression of genes encoding basic chitinase, basic glucanase and PAL was very low and signals were only observed in both control and infected leaves after blots had been exposed for 5 days. Clearly the expression of these defence genes was not induced by infection with A. candida ( Fig. 7A ). The expression of both photosynthetic genes cab and rbcS exhibited an age‐dependent decline in control leaves from day 8 onwards ( Fig. 7B ). However, in infected leaves both cab and rbcS gene expression were repressed earlier (from 6 DAI) and to a greater extent than in control leaves ( Fig. 7B ). Discussion Albugo candida lowers the rate of host photosynthesis in discrete regions of the infected leaf Biotrophic pathogens often lower the rate of whole leaf photosynthesis of their hosts ( Farrar and Lewis, 1987 ) and consistent with these studies, A. candida lowered the rate of whole leaf gas exchange. This was first apparent 8 days after inoculation when the leaf was illuminated with high irradiances. At this stage there was no loss of chlorophyll from the leaf and transpiration was not affected by the disease (data not shown). This suggests that the decline in photosynthesis may have been due to a decline in the content and/or activity of Rubisco, or to end‐product inhibition of photosynthesis. However, infected leaves are heterogeneous; they comprise cells directly associated with fungal mycelium and cells remote from the point of infection. Thus, in order to understand the complex changes in the regulation of metabolism in infected leaves it is essential to take account of this heterogeneity. In this study, quantitative imaging of chlorophyll fluorescence was used to examine spatial changes in the rate of photosynthesis in infected leaves as the pathogen developed. Chlorophyll fluorescence imaging is of most use when quantitative measurements of photosynthetic parameters can be obtained. Ø II is a measure of the proportion of absorbed light flowing through photosystem II ( Genty , 1989 ) hence Ø II × irradiance is a linear measure of photosynthetic electron transport. In this study, Ø II × irradiance was linearly related to CO 2 assimilation (determined by gas exchange) during steady state photosynthesis in both healthy and infected leaves at different stages of infection; therefore images of Ø II × irradiance could be used as a quantitative measure of CO 2 assimilation in different regions of the infected leaf. A decline in the rate of photosynthesis as determined by chlorophyll fluorescence imaging was first apparent 8 DAI and was closely correlated with the region where fungal mycelium was present and where spores were being produced on the underside of the leaf. As the area invaded by fungal mycelium increased, the decline in the rate of photosynthesis became more marked. Only at 14 DAI was there any impact of the presence of the fungus on uninfected regions of the leaf. Therefore the decline in the rate of photosynthesis was restricted closely, both temporally and spatially, to invaded regions of the leaf. This contrasts with the pattern observed in rust‐infected oat leaves ( Scholes and Rolfe, 1996 ) where the rate of photosynthesis declined in uninfected regions of the infected leaf at an earlier stage of infection and was not tightly coupled with the extent of fungal mycelium in the leaf. As photosynthesis decreased in the A. candida ‐infected region there was a corresponding increase in NPQ when the leaf was illuminated with irradiances greater than 300 µmol/m 2 /s. This indicated that light energy which was not being used for photosynthetic electron transport was being lost by nonphotochemical processes. NPQ arises from a number of processes that are induced upon exposure of leaves or isolated chloroplasts to light. The most important of these is high‐energy‐state quenching (qE), which depends upon the presence of the trans ‐thylakoid proton gradient ( Ruban and Horton, 1995 ). Thus one interpretation of the observed changes in NPQ is that there was a greater reduction in the activity of the Calvin cycle compared to other components of the photosynthetic apparatus. This would lead to a lower demand for ATP and NADPH and result in an increased transthylakoid pH gradient which in turn would result in an increase in NPQ ( Ruban and Horton, 1995 ). Such an effect would be most apparent at intermediate and high irradiances where the amount and activity of Rubisco and/or the amount and fluxes of carbohydrates often limit the rate of photosynthesis ( Stitt , 1991 ). Certainly the amount of mRNA encoding the small subunit of Rubisco was lower in infected as compared with control leaves at this stage of the life cycle and there was some accumulation of soluble carbohydrates. At low irradiances the maximum rate of photosynthesis is limited by the ability of the leaf to capture light, thus ATP and NADPH production are limiting rather than the capacity of the Calvin cycle. Consistent with this hypothesis, an increase in NPQ was not apparent in the A. candida ‐infected regions of the leaf when it was illuminated at lower irradiances; in contrast, NPQ was lower than in control leaves or in the uninfected region of the infected leaf. This can be seen clearly in the images of NPQ taken during photosynthetic induction at 93 µmol/m 2 /s. A low value of NPQ was also observed in all regions of a rust‐infected oat leaf at the flecking stage, only in the infected regions of the leaf at the sporulation stage ( Scholes and Rolfe, 1996 ) and in virus‐infected tobacco leaves ( Balachandran , 1994 ). Low values of NPQ may have been due to an increased demand for ATP which would result in a lower trans ‐thylakoid pH gradient. An increased demand for ATP is consistent with much evidence that suggests that biotrophic fungi stimulate host metabolic activity ( Roberts and Walters, 1988 ; Scholes and Farrar, 1986 ). Interpretation of fluorescence quenching parameters during photosynthetic induction is more difficult than in the steady state as many different processes within the leaf are being activated. However, the induction of Ø II and NPQ in control leaves was relatively simple. Ø II increased upon illumination reaching a maximum value after approximately 3 min. This correlated closely with an increase in transpiration as stomata opened. NPQ increased transiently and then declined as the Calvin cycle was activated using ATP and hence reducing the transthylakoid pH gradient. In contrast, a much more complicated pattern was observed in infected leaves with considerable heterogeneity within fungally invaded regions. Three distinct regions were visible in images of both Ø II and NPQ in infected leaves. Most notable was a 2–3 mm wide ring of cells surrounding the fungal pustule where Ø II was induced more rapidly than was observed in a control leaf. The alterations in the rate of induction of Ø II were unlikely to have resulted from altered rates of stomatal opening as the effect was first seen after just 2 s. Weis . (1998) demonstrated that zones of rapid photosynthetic induction are correlated with metabolically active regions of leaves where carbon export is low and respiration high. Thus, the distinct fluorescence characteristics may reflect high metabolic activity in this ring of cells at the leading edge of the colony. This is consistent with the starch print obtained for this leaf which revealed that there was greater amount of starch in these cells when compared to other regions of the infected leaf and to the control leaf. Thus images of NPQ and Ø II during photosynthetic induction reveal complex metabolic heterogeneity within the infected leaf. How does A. candida alter the amounts of storage and soluble carbohydrates in an infected leaf? In this study, the starch content of A. candida ‐infected regions of A. thaliana leaves declined, whilst in regions surrounding the infected area it increased as infection progressed. The loss of starch from the infected region was similar to that observed in leaves of Senecio squalidis infected with Albugo tragopogonis where starch disappeared from infection sites even before the disease symptoms were apparent ( Whipps and Cooke, 1978 ). It contrasts however, with the pattern of starch accumulation in radish cotyledons infected with Albugo candida ( Saettler and Pound, 1966 ) and barley leaves infected with brown rust Puccinia hordei ( Scholes and Farrar, 1987 ) where starch accumulation was localized to chloroplasts in cells within the fungal pustule. An increase or decline in the starch content of healthy leaves has been attributed to both a direct effect of Pi on the activity of chloroplast enzymes, e.g. ADPG pyrophosphorylase, and to the influence of cytosolic Pi on the availability of triose phosphates within the chloroplast. In addition, the amount of starch in a leaf will depend upon the sink or metabolic status of the leaf. In A. thaliana leaves infected with A. candida , the mechanism underlying the increase in starch in uninfected regions is unknown but is most likely to reflect a change in fluxes of sugars within the leaf. It is possible that Pi may have been limiting in these regions but the concentration of soluble sugars and the rate of photosynthesis were similar to control leaves. In infected regions of the leaf where starch was depleted, the rate of photosynthesis was low. It is likely that breakdown of starch contributed to the elevated amount of glucose in these regions, possibly due to an increase in demand for sugars by the fungus during sporulation. An increase in the amount of soluble carbohydrates (hexoses and sucrose) was apparent in infected regions of A. thaliana leaves from 6 days after inoculation. In the uninfected region of the infected leaf the soluble carbohydrate content was similar to control leaves except towards the end of the infection cycle (14 DAI) when an increase in all sugars was observed. An accumulation of glucose, fructose and sucrose has also been reported for whole infected leaves of Senecio squalidus infected with either Albugo tragopogonis ( Long and Cooke, 1974 ) or Puccinia lagenophorae ( Whipps and Cooke, 1978 ), vine leaves infected with Uncinula necator ( Brem , 1986 ), barley ( Scholes , 1994 ) and wheat ( Wright , 1995 ; Zulu , 1991 ) leaves infected with powdery mildew. Can an increase in host or fungal invertase activity account for the elevated amounts of hexoses in infected regions of the leaf? In A. thaliana leaves infected with A. candida there was an increase in both apoplastic and soluble invertase activity in infected areas of the leaf and the increase in soluble activity was partly contributed to by a new isoform. It is therefore likely that a large proportion of the increase in hexose content of invaded regions resulted from the increased invertase activity, since the timing of the increase correlated with the appearance of glucose and fructose, and the increase in activity was localized to invaded regions. Invertase activity did not increase in uninfected regions of the infected leaf and hexose sugars did not accumulate until the end of the infection cycle. There was also a good correlation between an increase in hexose concentrations and invertase activity in whole leaves of wheat infected with rust and powdery mildew ( Heisterüber , 1994 ; Wright , 1995 ; Zulu , 1991 ), oat and barley infected with rust ( Mitchell , 1978 ; Scholes , 1994 ) and vine infected with powdery mildew ( Brem , 1986 ). In addition, in infected leaves where hexoses did not accumulate there was also no stimulation in invertase activity, for example in rusted poplar leaves ( Roberts and Mitchell, 1979 ) and in vine leaves infected with downy mildew ( Brem , 1986 ). It is clear from this study that the increase in apoplastic invertase activity was of host origin, as mRNA levels of the ATβFRUCT1 gene increased dramatically in infected regions of the leaf. The induction of host apoplastic invertase activity and/or gene expression has been observed in two other host–pathogen interactions; carrot tap roots infected with the bacterial pathogen Erwinia carotovora ( Sturm and Chrispeels, 1990 ) and tomato roots infected with Fusarium oxysporum ( Benhamou , 1991 ). In the roots of some plants apoplastic invertase activity is thought to facilitate phloem unloading by maintaining a steep sucrose gradient between source (photosynthetic cells) and sink (heterotrophic cells) regions of the plant ( Eschrich, 1980 ) however, its role and cellular location in mature actively exporting leaves is less clear. Current evidence suggests that apoplastic invertase activity may be located in the leaf vasculature ( Kingston‐Smith and Pollock, 1996 ; Ramloch‐Lorenz , 1993 ; Zhang , 1996 ). Several studies have shown that apoplastic invertase gene expression declines suddenly as the leaves mature (although some activity remains) coinciding with the transition of the leaf from a sink to a source organ ( Godt and Roitsch, 1997 ; Sturm , 1995 ). In diseased leaves this situation is reversed and will have a number of consequences for both the plant and pathogen. Firstly, an increase in apoplastic invertase activity will reduce the amount of sucrose exported from the leaf, as has been shown in tobacco plants which over‐express yeast invertase in the apoplast ( Von Schaewen , 1990 ) but, in addition, it may facilitate phloem unloading of sucrose into cells adjacent to fungal mycelium thus converting regions of a source leaf into a sink for carbon. Such a strategy would aid efficient nutrient acquisition by the fungus and may lead to the down‐regulation of photosynthetic metabolism observed in these regions. At present the identity of the signal or the signal transduction pathways involved in the induction of invertase activity are unknown. A potential signal is the phyto‐hormone cytokinin which is produced by a number of fungi ( Pegg, 1981 ). Ehness and Roitsch (1997) showed that the application of cytokinin to suspension cultured cells of Chenopodium rubrum resulted in the direct stimulation of CIN 1 apoplastic invertase gene expression. Alternatively, elicitors such as cell wall fragments (oligogalacturonides, chitosans and pectic substances) have been demonstrated to stimulate the production of a range of wound‐induced and pathogenesis‐related proteins including invertase ( Ehness and Roitsch, 1997 ). Further studies are required to elucidate the signal transduction pathways involved. In addition to the increase in apoplastic invertase activity in A. candida ‐infected regions of A. thaliana leaves , there was an increase in the activity of soluble invertases which was again confined to the region of the leaf where fungal mycelium was present. To try to distinguish between host and fungal enzymes the profile of soluble invertase isoforms was examined in the different regions of the leaf by isoelectric focusing (IEF). Healthy A. thaliana leaves and the uninfected region of the infected leaf contained at least five major soluble isoforms which have characteristics typical of vacuolar invertases (e.g. acidic pIs). In the infected region, the host isoforms disappeared more rapidly than in control leaves and a new isoform with a higher pI appeared. Thus the majority of the increase in soluble activity measured in the infected region was due to this new isoform. If the new isoform was of host origin it had either resulted from the post‐translational modification of an existing soluble invertase isoform or it represented the expression of a hitherto unidentified host invertase gene. However, circumstantial evidence suggests that it is probably of fungal origin. Firstly, the kinetics of appearance of the isoform was very similar to the increase in biomass of the fungal mycelium and secondly, the isoform could not be induced in uninfected leaves by perturbations such as sugar feeding, wounding or accelerated senescence (N.B., unpublished data). Assuming that the new isoform was of fungal origin then its role in the metabolism of the infected leaf would critically depend upon its subcellular location. If located within the fungal hyphae then its ability to alter the carbohydrate metabolism of the host would be limited as it would not be able to hydrolyse sucrose in the extra‐haustorial matrix prior to uptake. However, if it was bound either to the hyphal membrane or excreted it would further contribute to the invertase activity already present within the host. What is the evidence that sugar sensing pathways may trigger a repression of photosynthetic gene expression and an induction of defence gene expression? Carbohydrate responsive genes have been classified as initiating ‘feast or famine’ responses ( Koch, 1996 ). Genes encoding photosynthetic proteins and those involved in resource mobilization are induced by carbohydrate depletion (the famine response) and repressed by an accumulation of sugars or an increase in fluxes of sugars through metabolic pathways. Conversely, a specific set of genes is up‐regulated in response to an accumulation of sugars (the feast response) most notably those involved in carbon storage and utilization and in plant defence. Signal transduction pathways for the repression of photosynthetic genes and for the induction of defence genes are largely unknown, but current evidence suggests the existence of a hexokinase‐dependent sugar sensing mechanism ( Jang , 1997 ; Sheen, 1994 ), a hexose‐transport associated sensor and a sucrose‐specific sensing pathway (reviewed by Smeekens and Rook, 1997 ). Evidence for hexose signalling associated with transport across a membrane into the cell comes from the use of glucose analogs such as 6‐deoxyglucose and 3‐O methyl glucose. These analogues can be transported across the membrane but are not metabolized. When added to cell suspension cultures or to intact leaves they initiate a sugar signal transduction pathway controlling the expression of some ‘feast’ and pathogen‐related genes ( Ehness , 1997 ; Roitsch , 1995 ). In a number of systems it has been shown that entry of hexoses into intermediary metabolism through the action of hexokinase initiates a sugar signal transduction pathway leading to the repression of typical ‘famine’ genes such as cab and rbcS (see Smeekens, 1998 ). Further evidence that hexoses may act as an internal sensor comes from the work of Jang . (1997) , who have over‐expressed and antisensed hexokinase genes in A. thaliana . Plants with reduced hexokinase activity are less sensitive to glucose whilst those with increased hexokinase activity are hypersensitive to glucose. To determine if the decline in the rate of photosynthesis and the accumulation of sugars observed in A. candida ‐infected regions of A. thaliana leaves were correlated with alterations in photosynthetic gene expression, cab and rbcS mRNA were measured by Northern blot analysis. In control leaves there was an age‐dependent decline in the abundance of both cab and rbcS gene expression which was evident from approximately 8–10 DAI. In infected leaves the expression of these genes was strongly reduced as early as 6 days after inoculation. The increase in the activity of acid invertases (apoplastic and soluble), the accumulation of hexoses and the down‐regulation of both cab and rbcS gene expression are consistent with the sugar sensing model. Although the content of hexoses in the infected regions did not reach the threshold level of hexoses (≈ 4.5 mmol/m 2 ) required to repress cab gene expression in transgenic tobacco expressing vacuolar or apoplastic yeast‐derived invertase ( Herbers , 1996 ), it was not far below. Moreover, the content of hexoses (≈ 3 mmol/m 2 ) in cold girdled spinach leaves did not reach this threshold even though the expression of photosynthetic genes was repressed ( Krapp and Stitt, 1995 ). In A. candida ‐infected A. thaliana leaves there was no induction of the defence proteins measured, despite the accumulation of carbohydrates in infected leaves. This suggests that either these proteins are not induced by sugars—at least via the same signal transduction pathway which may be involved in the repression of photosynthetic gene expression—or that defence gene expression was actively suppressed during the compatible interaction. Basse . (1992) reported that defence responses in tomato cells could be induced by elicitors (glycopeptides prepared from yeast invertase cleaved with alpha‐chymotrypsin) but inhibited by suppressors (oligosaccharides prepared from the elicitors by treatment with endo‐β‐N‐acetyl glucosaminidase H). It is also important to note that in plants undergoing a resistant interaction, the induction of invertase activity and defence proteins occurs very rapidly after the initial challenge (e.g. Benhamou , 1991 ) whereas in compatible interactions invertase activity, accumulation of sugars and the repression of photosynthetic gene expression occurs several days after infection. In conclusion, although the timing of events is consistent with a role for sugar signalling in this compatible host–pathogen interaction, more direct evidence is required. We are currently examining the effect of this pathogen on photosynthesis, carbohydrate metabolism and gene expression in A. thaliana plants with altered hexokinase activity ( Jang , 1997 ) and in plants which lack apoplastic invertase activity. Experimental Procedures Growth and inoculation of plants Seeds of Arabidopsis thaliana (ecotype OY0) were sown in Erin multipurpose compost and kept at 4 °C for 4 days to trigger germination. Seeds were then transferred to a growth room at a day temperature of 22 ± 1 °C and a night temperature of 15 ± 1 °C with a 9.5‐h photoperiod. Irradiance (200 µmol/m 2 /s) was provided by fluorescent lamps (OSRAM L 58 W/77 Fluora, Germany). Seedlings were transplanted into pots (6 × 6 cm 2 ) 2 weeks after sowing and grown in the same growth room. When the plants were 5 weeks old, eight of the youngest fully expanded leaves were inoculated towards the tip of the adaxial surface with four 10 µL droplets of zoospore suspension (10 6 zoospores/mL) as described in Tang . (1996) . Following inoculation, plants were kept in a propagating chamber for 24 h to maintain high humidity. Control plants were inoculated with water and treated in an identical manner. The region of the leaf inoculated with zoospores was termed region A, the uninoculated region of the same leaf, region B and similar regions of a control leaf, C and D. Fungal mycelium was confined to region A of the infected leaf. Determination of chlorophyll content The chlorophyll and carotenoid content of regions A–D were determined 4, 6, 8, 10 12 and 14 DAI following the procedure of Scholes . (1994) . The concentrations of chlorophyll a , chlorophyll b , and carotenoids were calculated from the equations of Lichtenthaler and Wellburn (1983) . Simultaneous measurement of CO 2 assimilation and chlorophyll fluorescence imaging Chlorophyll fluorescence imaging was performed as described in Rolfe and Scholes (1995) . Infected and control plants were removed from the growth room just prior to use, 6, 8, 11 and 14 DAI. Leaves (still attached to the plant) were placed in the chamber of an IRGA (LCA‐4 with PLCA4 leaf chamber; Analytical Development Company, Hoddesdon, Herts, UK) set to record CO 2 assimilation every 30 s. The CO 2 concentration was maintained at 340 µmol/mol with 60% relative humidity. The leaf was maintained in darkness for 5 min. During this time an image was captured representing zero fluorescence. The leaf was then exposed to a 1.5‐s pulse of saturating illumination and an image of chlorophyll fluorescence captured. Thirty seconds later the actinic light (93 µmol/m 2 /s) was switched on. The leaf was exposed to a saturating flash 2.5, 40 and then every 60 s after the commencement of actinic illumination. Two duplicate images were captured just before and at the peak of the saturating flash. During this time the leaf went through photosynthetic induction. The experiment lasted 1260 s in total. For the measurement of photosynthesis at different irradiances, the protocol described above was performed. Thereafter, the actinic irradiance was increased step‐wise up to a maximum of 1450 µmol/m 2 /s. The leaf was illuminated for 10 min at each irradiance and images of chlorophyll fluorescence were captured and stored every minute for the last 4 minutes. Analysis of soluble carbohydrates and starch Leaf discs (0.8 cm 2 ) were harvested from regions A, B, C and D of five control or infected plants 6 h into the photoperiod, 4, 6, 8, 10, 12 and 14 DAI. Samples were immediately frozen in liquid N 2 . Soluble carbohydrates were extracted in buffered 80% ethanol (50 m m Hepes‐NaOH 5 m m MgCl 2 pH 7.5) at 70 °C. Sucrose, glucose and fructose were measured by enzyme‐linked assay as described in Scholes . (1994) . Leaf discs remaining after the extraction of soluble carbohydrates were washed in distilled water and then ground in liquid N 2 in a pestle and mortar. One millilitre of 0.5 m Mes (2‐(N‐Morpholino) ethanesulphonic acid) buffer (pH 4.5) was then added together with 14 units of amyloglucosidase and 0.4 units of α‐amylase (Sigma UK). The samples were agitated overnight at room temperature. An aliquot (50 µL) was then assayed for glucose as described in Scholes . (1994) . After control and A. candida ‐infected leaves had been imaged (6, 8, 11 and 14 DAI) leaves were detached from the plant and placed in 80% ethanol for 7 days to remove chlorophyll. Leaves were then immersed in an iodine‐potassium iodide‐lactic acid mixture as described by Lindner . (1959) . Areas of the leaf containing starch turned black. In order to obtain semiquantitative data on the amount of starch in the different regions, leaves were imaged by placing them on a lightbox and capturing an image using a monochrome CCD camera. The intensity of staining along a 2.5‐cm transect was measured using the Optimas 4.0 image analysis package. This provided a quantitative measure of stain intensity and hence a semiquantitative measure of starch content. Measurement of invertase activity and soluble invertase isoform pattern Leaf discs (≈ 7.07 cm 2 ) were harvested from regions A, B, C and D from six control and six infected plants, 3 h into the photoperiod 0, 3, 6, 9, 14 and 18 DAI. Soluble and wall‐bound invertase activities were extracted and assayed, and isoelectric focusing (IEF) of proteins (to obtain the isoform pattern) performed, as described in Tang . (1996) . In order to visualize the invertase isoforms, the gel was stained for invertase activity as described in Faye (1981) . Quantification of ATβFRUCT1 gene expression by RT‐PCR A. thaliana contains two genes encoding apoplastic invertase, ATβFRUCT1 which is expressed in all tissues and ATβFRUCT2 which is expressed solely in reproductive tissues (N.B., unpublished data). The low abundance of the ATβFRUCT1 gene precluded detection by Northern blot analysis, hence semiquantitative reverse transcriptase linked polymerase chain reaction (RT‐PCR) was used to measure expression. To compensate for different reaction efficiencies during RT‐PCR, an internal RNA standard derived from a fragment of the ATβFRUCT1 gene was added to each sample. To generate this internal RNA standard, a 782‐bp portion of the ATβFRUCT1 gene was amplified by PCR using gene‐specific primers (primer 1: 5′‐CCCTTTCCACCGAAACTCTCCACT‐3′; primer 2: 5′‐GTGAGAGACTTTGGAGAAAGCGGAT‐3′), cloned and a small internal fragment (105 bp) removed by digestion with restriction enzymes SnaBI and Bpu1102I. The 5′ overhangs generated by the restriction enzyme digest were filled in using the large fragment of the DNA polymerase I and the DNA self‐ligated. This modified portion of the ATβFRUCT1 gene was designated pδβFRUCT1. The plasmid was then linearized by digestion with SalI and RNA (sense strand) generated by in vitro transcription with T7 RNA polymerase according to the manufacturer’s instructions (Promega). Template DNA was removed by digestion with RQ1 DNase and the RNA purified by extraction with phenol:chloroform followed by ethanol precipitation. The purified RNA, 15 µg produced from 5 µg template (DNA), was resuspended in 200 µL of RNAse‐free dH 2 O, quantified by absorption spectroscopy, and stored at −80 °C until use. RNA produced by in vitro transcription (677 bp) could be distinguished on the basis of size from both genomic DNA (782 bp) and authentic RT‐PCR product of the ATβFRUCT1 mRNA (349 bp). For RT‐PCR, RNA was extracted from infected and uninfected leaves of A. thaliana using the method of Loening (1969) . Five µg of sample RNA was mixed with 100 pg of standard RNA and brought to a total volume of 8 µL with RNAse‐free dH 2 O. The sample was incubated at 37 °C for 30 min with 1 unit of RQ1 RNAse‐free DNAse (Promega) and then incubated at 75 °C for 5 min to inactivate the nuclease. cDNA was then generated using random hexamer primers and commercially available first strand cDNA synthesis kit (Promega). Between 0.5 and 5 µg of cDNA were used in the subsequent amplification reactions. The PCR consisted of cDNA, 0.05 µ m primer 1, 0.05 µ m primer 2 (primers as above) 1.5 m m MgCl 2 and 150 µ m each of dATP, dCTP, dGTP and dTTP and supplied PCR buffer in a total volume of 100 µL. The reaction was brought to 90 °C before the addition of 2.5 units of Taq DNA polymerase. Cycling conditions were 94 °C for 30 s, 65 °C for 1 min and 72 °C for 1 min repeated 30 times. Reactions were held at 72 °C for 5 min at the end of the cycling reactions. These conditions had been empirically determined to result in linear amplification of the cDNA. Thirty µL of each PCR reaction mixture was electrophoresed through 1% (w/v) agarose gels, 40 m m Tris, 20 m m acetic acid, 1 m m EDTA buffer and then stained for 15 min in 0.5 µg/mL ethidium bromide. The gels were destained for 30 min in dH 2 O and then fluorescent DNA bands visualized by UV trans ‐illumination. An image of fluorescence intensity of each band was detected using a cooled CCD camera (C5985, Hamamatsu, Japan) using a Tiffin Orange 15 (Tiffin, USA) and BG18 (Schott, Mainz, Germany) filters. Optimas 4.0 imaging processing software (Bioscan Inc, Edmonds, USA) was used to analyse the intensity of each band and calculate the fluorescence intensity of the ATβFRUCT1 RT‐PCR product relative to that of the internal RNA standard. PCR amplification of genes encoding internal transcribed spacer (ITS) regions Amplification of the internal transcribed spacer (ITS) regions of A. thaliana and A. candida was carried out using the primers ITS1 and ITS4 ( White , 1990 ). Primer pairs were used in combination with DNA prepared from either control or infected leaves in PCR reactions according to the protocol of Wright . (1999) . After amplification, samples were electrophoresced through 1% (w/v) agarose gels and visualized using ethidium bromide. DNA fragments of the correct size (≈ 600 bp from A. thaliana and 700 bp from A. candida ) were isolated from the gel using a Wizard PCR Preps DNA purification kit (Promega UK). Fragments were cloned and sequenced as described in Wright . (1999) . The sequences were confirmed as ITS regions by homology searches using the NCBI BLAST programme ( Altschul , 1997 ). Northern blot analysis of photosynthetic and defence proteins Four leaf discs (approximately 100 mg fresh weight) were harvested from control and infected leaves 3 h into the photoperiod, 6, 8, 12 and 14 DAI for analysis of the expression of cab and rbcS , and 4 and 8 DAI for the analysis of the expression of defence proteins. Total RNA was isolated following the procedures of Loening (1969) . Northern blots were prepared using standard procedures ( Sambrook , 1989 ). Five µg of total RNA from each sample was separated by denaturing electrophoresis through 1% (w/v) agarose gels containing formaldehyde. The RNA was transferred to a nylon membrane (Zetaprobe, Biorad Ltd, UK) by capillary transfer overnight in 10 × SSC buffer (1.5 m NaCl, 0.15 m sodium citrate, pH 7.0) and then immobilized by UV cross‐linking. Radioactive DNA probes were prepared by random‐prime labelling of isolated DNA fragments (Megaprime, Amersham International Plc, UK). The probes were: a 600 and 700 bp Eco RI fragment from cDNA clones encoding the ITS region of A. thaliana and A. candida, respectively; an 1800‐bp Eco RI fragment isolated from the cDNA clone, CAB180, encoding a chlorophyll a/b binding protein ( Leutwiler , 1986 ); an 1800‐bp Hin dIII fragment isolated from a genomic DNA clone, pATS17, encoding the small subunit of Rubisco ( Krebbers , 1988 ); a 1398‐bp Bam HI/ Sac I fragment isolated from a genomic DNA clone, pSKPAL, encoding phenyl ammonia lyase (PAL) ( Dong , 1991 ); 1000 bp Bam HI/ Kpn I fragments isolated from cDNA clones encoding an acidic chitinase ( Genbank I.D. T46824) and a basic chitinase ( Genbank I.D. T21751); and a 1200‐bp Bam HI/ Kpn I fragment isolated from a cDNA clone encoding a basic glucanase (BG2) ( Genbank I.D. N37986). The membranes were hybridized with the radioactive probes for 18 h in 0.5 m Na‐phosphate buffer pH 7.2, 7% (w/v) SDS at 65 °C and then washed 3 times at 65 °C for 30 min in 80 m m Na‐phosphate buffer, pH 7.2, 5% (w/v) SDS and three times in 80 m m Na‐phosphate buffer, pH 7.2, 1% (w/v) SDS. Bound radioactivity was detected by autoradiography for 30 min (ITS), 24 h (photosynthetic proteins) or 5 days (defence proteins) using preflashed Biomax MS‐1 Film (Eastman Kodak Company UK) at −80 °C using one intensifying screen ( Sambrook , 1989 ). Statistical analysis The mean, standard deviation and standard error were calculated for each data set where appropriate. Standard error bars were plotted except where smaller than the symbol size. anova was used to identify the difference between infected region A, uninfected region B, and equivalent control regions C and D. Where appropriate, the student t ‐test was used to identify the difference between regions A and C, and regions B and D. The statistical analysis was carried out using the package Microsoft Excel 5.0. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Molecular Plant Pathology Wiley

Infection of Arabidopsis thaliana leaves with Albugo candida (white blister rust) causes a reprogramming of host metabolism

Loading next page...
 
/lp/wiley/infection-of-arabidopsis-thaliana-leaves-with-albugo-candida-white-HZBpV4wn20

References (50)

Publisher
Wiley
Copyright
Copyright © 2000 Wiley Subscription Services, Inc., A Wiley Company
ISSN
1464-6722
eISSN
1364-3703
DOI
10.1046/j.1364-3703.2000.00013.x
pmid
20572957
Publisher site
See Article on Publisher Site

Abstract

Introduction Biotrophic fungi are agriculturally important pathogens often causing severe losses in yield. These pathogens are not thought to produce toxins and it is likely that they lower the yield of their hosts by two major processes. Firstly, they often lower the rate of photosynthesis and hence the production of carbon by the host and secondly they compete very effectively for host carbon; they are an effective sink. This acquisition of host carbon by the fungus results in a dynamic interaction between the physiology of the host and that of the parasite and to a reprogramming of host metabolism. Many rust and powdery mildew fungi cause a progressive decline in the rate of photosynthesis, an increase in invertase activity and an accumulation of hexose sugars (and sometimes sucrose) in infected leaves (for reviews see Farrar and Lewis, 1987 ; Scholes, 1992 ; Whipps and Lewis, 1981 ). Recently a model linking these metabolic symptoms has been proposed ( Scholes , 1994 ; Scholes, 1992 ; Tang , 1996 ). In this model, an increase in invertase activity leads to an accumulation of soluble carbohydrates, and/or to a change in the flux through these metabolic pathways. This in turn initiates a signal transduction pathway(s) leading to a repression of photosynthetic gene expression and hence to a decline in the rate of photosynthesis in infected leaves. Although it is clear that sugars control the expression of many plant genes ( Koch, 1996 ) much less is known about the precise mechanisms by which plants sense sugars. Currently there is evidence for the existence of a hexokinase‐dependent sugar sensing mechanism ( Jang , 1997 ; Sheen, 1994 ), a hexose‐transport associated sensor and a sucrose‐specific sensing pathway (reviewed by Smeekens and Rook, 1997 ). Although the content of hexose sugars and invertase activity are known to increase in leaves infected by many different biotrophic pathogens, a number of questions concerning the sugar sensing hypothesis remain unanswered. For example, is the increase in invertase activity of host or fungal origin? Is it confined to regions of the leaf invaded by fungal mycelium or is the activity induced systemically throughout the infected leaf? When and where does invertase activity increase in relation to changes in the amount or fluxes of hexose sugars? Is there a repression of photosynthetic gene expression in regions of the leaf where hexose sugars accumulate and over a time scale which is consistent with the hypothesis? What is the evidence that sugar sensing pathways are an important component of a compatible plant–pathogen interaction? A number of defence‐related genes are inducible by soluble sugars ( Herbers and Sonnewald, 1998 ; Herbers , 1996 ; Johnson and Ryan, 1990 ) suggesting that sugars play a role not only in the repression of photosynthetic genes but also in the induction of defence responses. Herbers . (1996) showed that, in transgenic tobacco plants over‐expressing yeast invertase in the apoplast, defence‐related gene transcripts, callose content, peroxidase activities and salicylic acid were all elevated. The accumulation of pathogenesis‐related proteins and the repression of photosynthetic gene transcripts were inversely correlated, and required the same threshold level of hexoses for both induction and repression. This raises the question of whether defence–related proteins are induced to any extent, at any stage, during a compatible host–pathogen interaction and, if not, whether the pathogen actively represses the induction of defence proteins. This study aims to examine the effect of Albugo candida (white blister rust) on the regulation of photosynthesis, carbohydrate metabolism and gene expression in discrete regions of A. thaliana leaves during a compatible interaction. Specifically we: (i) examine the spatial changes in the rate of photosynthesis in an infected leaf as the disease develops by quantitative imaging of chlorophyll fluorescence, (ii) quantify changes in the amounts of soluble and storage carbohydrates in different regions of the infected leaf as the disease progresses, (iii) quantify changes in the activity of wall‐bound and soluble invertases and in mRNA levels of the host wall‐bound invertase, and (iv) examine changes in the expression of genes encoding photosynthetic and defence proteins. The results are discussed in relation to the hypothesis and questions outlined above. Results The effect of A. candida on photosynthesis in different regions of an A. thaliana leaf Images of chlorophyll fluorescence were captured at a range of irradiances from healthy and infected plants 6, 8, 11 and 14 DAI (days after inoculation). Carbon dioxide assimilation was measured simultaneously using an infra‐red gas analyser (IRGA). To determine whether Ø II could be used to assess CO 2 assimilation, the product of Ø II and irradiance determined from an entire leaf was plotted against the assimilation rate simultaneously measured by gas exchange ( Fig. 1 ). This plot was linear for both healthy and infected leaves for all irradiances examined (ranging from 0 to 1045 µmol/m 2 /s). Therefore the rate of photosynthesis could be accurately determined in different regions of an infected leaf using chlorophyll fluorescence imaging. 1 The relationship between the rate of CO 2 assimilation determined by infra‐red gas analysis and Ø II × irradiance determined by analysis of chlorophyll fluorescence images. The two measurements were linearly related for healthy (○) and A. candida ‐infected (▪) leaves. The solid line represents a linear regression ( r 2 = 0.926, gradient = 20.4). Figure 2A shows the development of visible symptoms for representative leaves used for imaging. At least two leaves were imaged for each time‐point. Leaves were inoculated towards the tip of the adaxial surface with four 10 µL droplets of zoospore suspension. The region of the leaf inoculated with zoospores was termed region A, the uninoculated region of the same leaf, region B and similar regions of a control leaf, C and D. Fungal mycelium was confined to region A of the infected leaf. Six DAI, visible symptoms were not apparent, nor were symptoms apparent on images obtained by chlorophyll fluorescence analysis (data not shown). Spores first became visible on the lower surface of the leaf 8 DAI, and developed to cover approximately 30% of the leaf by day 11. No symptoms were apparent on the upper surface of the leaf 8 DAI. By 11 days a slight chlorosis was visible in the infected region which became much more pronounced by day 14. 2 Images of Ø II and nonphotochemical quenching (ΔFm/Fm′) for healthy and A. candida ‐infected leaves of A. thaliana . (A) Images of Ø II and ΔFm/Fm′ were calculated from measurements of chlorophyll fluorescence made after photosynthesis had reached steady‐state (1260 s) at an actinic irradiance of 380 µmol/m 2 /s and ambient CO 2 (350 µmol/mol). Upper two rows: photographs of the upper and lower surfaces of control and infected leaves (8, 11 and 14 DAI) used for imaging. Middle row: images of Ø II (and calculated assimilation rate derived from Fig. 1 ). The scales show the relationship between colour and the relevant chlorophyll fluorescence measurement. Bottom row; images of nonphotochemical energy dissipation (ΔFm/Fm′). (B) Images of Ø II and ΔFm/Fm′ are shown at selected time points throughout the induction of photosynthesis for control and A. candida ‐infected leaves 11 DAI. Times are indicated in seconds after the actinic light was first switched on. Induction was performed at low irradiance (93 µmol/m 2 /s) and ambient CO 2 (350 µmol/mol). The first two rows show images of Ø II and ΔFm/Fm′ for a control leaf and the bottom two rows for an infected leaf. The scales show the relationship between colour and the relevant chlorophyll fluorescence measurement. Representative images of Ø II and nonphotochemical quenching (NPQ), taken when photosynthesis had reached a steady‐state value at an actinic irradiance of 380 µmol/m 2 /s are shown in Fig. 2A . A slight reduction in Ø II was first apparent 8 DAI and was limited to the inoculated area (region A) where spores were visible on the lower leaf surface. This reduction became more widespread by day 11 and by day 14 photosynthesis was severely inhibited. In all cases there was a close correlation between the area where Ø II was reduced and the visible symptoms on the leaf; there was little impact on uninvaded regions of the leaf except at day 14. NPQ was elevated in regions where Ø II was reduced, indicating that light energy which was not used for photosynthetic electron transport was being lost by nonphotochemical processes. Images of Fv/Fm were similar in both control and infected leaves (data not shown) indicating that A. candida had not caused photodamage in any region of the leaf. Average values of Ø II and NPQ were extracted from different regions of images of control and infected leaves at irradiances from 0 to 1450 µmol/m 2 /s and are shown in Fig. 3 with corresponding measurements of whole‐leaf gas exchange. Results are shown for leaves 8, 11 and 14 DAI; there was no difference between control and infected leaves before this time. The rate of photosynthesis (measured by gas exchange or imaging) remained constant in control leaves throughout the duration of the experiment. In contrast, gas exchange measurements showed that there was a progressive decline in CO 2 assimilation by infected leaves, such that by day 14 the rate was only 30% of the control ( Fig. 3A–C ; upper panel). Images of Ø II showed that the reduction in photosynthesis was largely limited to the region invaded by the fungal mycelium and became more pronounced as the disease developed ( Fig. 3B,C ; middle panel). Only on day 14 was photosynthesis also reduced in the uninvaded region of the infected leaf. The decline in Ø II in region A was apparent at all irradiances except the lowest (93 µmol/m 2 /s). 3 The effect of A. candida on the rate of CO 2 assimilation, Ø II × irradiance and NPQ (ΔFm/Fm′) in localized regions of leaves of A. thaliana (A) 8; (B) 11 and (C) 14 DAI. Control leaf (○); whole infected leaf (▪); infected region A (▴) and uninfected region B (◆) of an infected leaf. Nonphotochemical quenching did not differ between control and infected leaves 8 DAI ( Fig. 3A ; lower panel). At low irradiances, NPQ was low (0.5) and it increased to a maximum of 2.5 as the irradiance increased ( Fig. 3A ; lower panel). A similar pattern was observed in control leaves 11 and 14 DAI. In infected region A, 11 and 14 DAI, NPQ was slightly reduced at low irradiances compared to uninfected region B or control leaves ( Fig. 3B,C ; lower panel). However, as the irradiance increased, NPQ increased much more rapidly in infected region A. The maximal value of NPQ observed at high irradiance 11 DAI was the same as that observed in control leaves but was significantly higher (3.2 compared with 2.5) by day 14. The NPQ of uninfected regions was also increased 14 DAI compared with control leaves but the increase was smaller than that observed in the infected region. To examine the effect of A. candida on photosynthetic induction, leaves were illuminated at an irradiance of 93 µmol/m 2 /s for 1260 s. Images of Ø II and NPQ were captured during this induction process and representative images of selected time‐points are shown in Fig. 2B . In contrast to measurements taken at steady state, during photosynthetic induction photosynthetic electron transport through photosystem II (and hence Ø II ) is not strictly proportional to the rate of CO 2 assimilation (data not shown) most probably due to the many different processes being activated during this period (generation of intermediates in the Calvin cycle, Rubisco activation, stomatal opening and generation of a transthylakoid pH gradient). In control leaves, Ø II increased rapidly and reached a steady‐state value of 0.8 after 500 s. This corresponded to a steady‐state rate of CO 2 assimilation of approximately 3.5 µmol/m 2 /s. In infected leaves, the steady‐state value of Ø II was not significantly different from that of control leaves but the pattern observed during induction was more complex. The infected leaf showed three distinct regions. In the infected region, where spores were visible on the underside of the leaf, Ø II was induced more slowly than other parts of the leaf. Surrounding this was a zone 3–4 mm wide, where Ø II was induced very rapidly when compared with both the control leaf and other regions of the infected leaf. The uninfected part of the leaf behaved in the same way as the control leaf. The induction of NPQ was also complex. In the control leaf, NPQ exhibited an initial transient increase but then declined to a uniform low value. This transient increase in NPQ was greater in infected region A but the final value of NPQ attained was lower. The induction of NPQ in uninfected region B was similar to that of the control leaves, although there was a slight difference in the time taken to reach a steady‐state value. Chlorophyll content of different regions of A. candida ‐infected leaves The chlorophyll content of all regions of infected and uninfected leaves was similar until 12 days after inoculation. Thereafter both chlorophyll a and b were lost from fungally invaded regions of the leaf (A) such that by day 14 the total chlorophyll content was 40% ( P < 0.001) of that in equivalent region C of a control leaf (310.1 ± 13.3 and 125.1 ± 10.3 mg/m 2 , respectively). The effect of A. candida on soluble carbohydrates and starch in localized regions of A. thaliana leaves Figure 4 shows the amount of sucrose, glucose and fructose in different regions of control and A. candida ‐infected leaves throughout the infection cycle. There was a significant increase in the amount of all soluble carbohydrates in infected region A when compared with the equivalent control region C from 6 DAI. The accumulation of carbohydrates was most pronounced on day 14. There was no increase in the soluble carbohydrate content of the uninfected region (B) of the infected leaf until day 14. 4 The effect of A. candida on the amount of soluble carbohydrates in localized regions of leaves of A. thaliana throughout the course of infection. Infected leaf; regions A (▴) and B (◆). Control leaf; regions C (○) and D (□). The results are the means ± the standard error of six replicate measurements. *, ** and *** indicate that data are significantly different, P < 0.05, 0.01 and 0.001, respectively. To examine the distribution and amount of starch, leaves were either cleared and stained for starch with I–KI solution, or analysed for starch content enzymically. Figure 5A shows control and infected leaves which have been stained for starch 8, 11 and 14 DAI and Fig. 5B shows a scan of the intensity of the black stain along a 2.5‐cm transect of the leaf starting from the middle of a pustule (or equivalent region) and extending longitudinally down the leaf. Eight DAI starch staining of leaves showed that the distribution of starch within the infected leaf was relatively uniform ( Fig. 5 B, i). Although the starch content appeared slightly higher in the infected than in the control leaf ( Fig. 5 B, i) enzymatic analysis revealed little difference between the content of either tissue ( Fig. 5C ). By day 11 it was apparent that starch had been lost from the infected region of leaves. In addition, the uninfected region of the infected leaf appeared to contain more starch than control leaves and in some leaves it was clear that there was a darker staining ring approximately 1–2 mm in width immediately around the pustule ( Fig. 5A,B ii). These results were confirmed by starch assay in different regions of the leaves ( Fig. 5C ). By day 14 the amount of starch in infected regions of the leaf was lower than comparable regions of control leaves and in uninfected regions of infected leaves it was higher than comparable regions of control leaves ( Fig. 5B , iii and C.). 5 The effect of A. candida on the starch content of localized regions of leaves of A. thaliana throughout the course of infection. (A) Leaves of A. thaliana were cleared and stained for starch. Left column control leaves; right column A. candida ‐infected leaves (i) 8 (ii) 11 and (iii) 14 DAI. (B) The intensity of the stain was measured along a 2.5‐cm transect positioned as shown in (A). Control leaf (‐ ‐ ‐); infected leaf (—). (C) Starch was extracted from different regions of control and infected leaves and quantified by enzyme‐linked assay. Infected leaf; regions A (▴) and B (◆). Control leaf; regions C (○) and D (□). The results are the means ± the standard error of five replicate measurements. The effect of A. candida on invertase activity, isoform pattern and gene expression in localized regions of A. thaliana leaves Figure 6B,C shows the changes in activity and isoform pattern of soluble invertases in infected and control A. thaliana leaves. There was an increase in the activity of soluble invertase in region A from 9 DAI but no change in region B when compared to the control leaf. Up until 6 DAI the soluble invertase isoform pattern was the same in infected and control regions; there were five major isoforms in each region ( Fig. 6C ). From day 10 onwards a new isoform with a pI of 5.1 appeared in the infected region of the leaf. The activity of the new isoform increased gradually as the fungus developed and the activity of the other five isoforms decreased as the leaves aged suggesting that it was of fungal origin ( Fig. 6A–D .). 6 The effect of A. candida on invertase activity in localized regions of leaves of A. thaliana . Samples for analysis of (A) wall‐bound and (B) soluble invertases were harvested from infected leaves, regions A (▴) and B (◆) and control leaves, regions C (○) and D (□) throughout the timecourse of infection. The results are the means ± the standard error of six replicate measurements. * and *** indicate that data are significantly different, P < 0.05 and 0.001, respectively. (C) Soluble invertase isoforms in different regions of A. thaliana leaves infected with A. candida . Soluble invertase isoforms were separated by isoelectric focusing over a pH range of 4.0–6.5. Invertase activity was visualized by staining reducing sugars with 2, 3, 5 tetrazolium chloride. Samples were prepared from infected region A, uninfected region B and equivalent control regions C and D, 0, 6, 10, 14 and 19 DAI. Samples were loaded on an equal leaf area basis. Table 1 shows the expression of the ATβFRUCT 1 apoplastic invertase gene in different regions of an A. thaliana leaf following infection by A. candida . An increase in gene expression was first detected in region A 8 DAI and by day 13 was approximately 40‐fold greater than in the control leaf. There was no change in gene expression in the uninfected region of the infected leaf. The activity of the apoplastic invertase mirrored the change in gene expression; activity was increased in region A from 6 DAI and by day 14 was fourfold greater than in the control leaf ( Fig. 6A ). Again, there was no increase in apoplastic invertase activity in region B of the infected leaf. 1 Expression of the ATβFRUCT 1 apoplastic invertase gene in leaves of A. thaliana infected with A. candida . The amount of expression in regions A and B of infected leaves and C and D of control leaves was analysed throughout the infection cycle using semiquantitative RT‐PCR. Results are expressed as pg/µg total RNA. Data are shown as the mean ± SE of three measurements. Infected leaf (pg/µg total RNA) Control leaf (pg/µg total RNA) Days after inoculation Region A Region B Region C Region D 4 4.5 ± 0.39 10.0 ± 3.7 < 2.0 15.9 ± 2.0 6 6.5 ± 1.7 5.6 ± 0.9 < 2.0 < 2.0 8 56.5 ± 3.6 17.6 ± 4.4 6.4 ± 0.48 14.6 ± 0.8 10 68.7 ± 8.8 2.2 ± 2.0 11.9 ± 2.0 11.9 ± 0.3 13 436.1 ± 105.7 15.0 ± 1.5 < 2.0 < 2.0 The effect of A. candida on the expression of genes encoding photosynthetic and defence proteins in A. thaliana leaves The expression of photosynthetic and defence proteins was examined by Northern blot analysis of total RNA from control and A. candida ‐infected leaves. Whereas RNA prepared from control leaves is exclusively of plant origin, infected leaves contain both plant and fungal RNA. The internal transcribed spacer (ITS) probes from A. thaliana and A. candida were used to determine the relative amounts of plant and fungal RNA in the different samples. The plant and fungal ITS probes were completely specific ( Fig. 7A ). The amount of plant RNA in all samples was similar. Four DAI, fungal RNA was present in the infected leaf but this had increased substantially as the fungus developed. 7 Northern hybridizations showing the expression of genes encoding (A) defence proteins and (B) photosynthetic proteins in leaves of A. thaliana infected with A. candida . Species‐specific ITS probes were employed to determine the proportion of RNA of the plant and fungus present in infected leaves. The use of the A. thaliana ITS probe showed equal loading of A. thaliana RNA. The expression of the gene encoding acidic chitinase was not detectable in either control or infected leaves (data not shown). The expression of genes encoding basic chitinase, basic glucanase and PAL was very low and signals were only observed in both control and infected leaves after blots had been exposed for 5 days. Clearly the expression of these defence genes was not induced by infection with A. candida ( Fig. 7A ). The expression of both photosynthetic genes cab and rbcS exhibited an age‐dependent decline in control leaves from day 8 onwards ( Fig. 7B ). However, in infected leaves both cab and rbcS gene expression were repressed earlier (from 6 DAI) and to a greater extent than in control leaves ( Fig. 7B ). Discussion Albugo candida lowers the rate of host photosynthesis in discrete regions of the infected leaf Biotrophic pathogens often lower the rate of whole leaf photosynthesis of their hosts ( Farrar and Lewis, 1987 ) and consistent with these studies, A. candida lowered the rate of whole leaf gas exchange. This was first apparent 8 days after inoculation when the leaf was illuminated with high irradiances. At this stage there was no loss of chlorophyll from the leaf and transpiration was not affected by the disease (data not shown). This suggests that the decline in photosynthesis may have been due to a decline in the content and/or activity of Rubisco, or to end‐product inhibition of photosynthesis. However, infected leaves are heterogeneous; they comprise cells directly associated with fungal mycelium and cells remote from the point of infection. Thus, in order to understand the complex changes in the regulation of metabolism in infected leaves it is essential to take account of this heterogeneity. In this study, quantitative imaging of chlorophyll fluorescence was used to examine spatial changes in the rate of photosynthesis in infected leaves as the pathogen developed. Chlorophyll fluorescence imaging is of most use when quantitative measurements of photosynthetic parameters can be obtained. Ø II is a measure of the proportion of absorbed light flowing through photosystem II ( Genty , 1989 ) hence Ø II × irradiance is a linear measure of photosynthetic electron transport. In this study, Ø II × irradiance was linearly related to CO 2 assimilation (determined by gas exchange) during steady state photosynthesis in both healthy and infected leaves at different stages of infection; therefore images of Ø II × irradiance could be used as a quantitative measure of CO 2 assimilation in different regions of the infected leaf. A decline in the rate of photosynthesis as determined by chlorophyll fluorescence imaging was first apparent 8 DAI and was closely correlated with the region where fungal mycelium was present and where spores were being produced on the underside of the leaf. As the area invaded by fungal mycelium increased, the decline in the rate of photosynthesis became more marked. Only at 14 DAI was there any impact of the presence of the fungus on uninfected regions of the leaf. Therefore the decline in the rate of photosynthesis was restricted closely, both temporally and spatially, to invaded regions of the leaf. This contrasts with the pattern observed in rust‐infected oat leaves ( Scholes and Rolfe, 1996 ) where the rate of photosynthesis declined in uninfected regions of the infected leaf at an earlier stage of infection and was not tightly coupled with the extent of fungal mycelium in the leaf. As photosynthesis decreased in the A. candida ‐infected region there was a corresponding increase in NPQ when the leaf was illuminated with irradiances greater than 300 µmol/m 2 /s. This indicated that light energy which was not being used for photosynthetic electron transport was being lost by nonphotochemical processes. NPQ arises from a number of processes that are induced upon exposure of leaves or isolated chloroplasts to light. The most important of these is high‐energy‐state quenching (qE), which depends upon the presence of the trans ‐thylakoid proton gradient ( Ruban and Horton, 1995 ). Thus one interpretation of the observed changes in NPQ is that there was a greater reduction in the activity of the Calvin cycle compared to other components of the photosynthetic apparatus. This would lead to a lower demand for ATP and NADPH and result in an increased transthylakoid pH gradient which in turn would result in an increase in NPQ ( Ruban and Horton, 1995 ). Such an effect would be most apparent at intermediate and high irradiances where the amount and activity of Rubisco and/or the amount and fluxes of carbohydrates often limit the rate of photosynthesis ( Stitt , 1991 ). Certainly the amount of mRNA encoding the small subunit of Rubisco was lower in infected as compared with control leaves at this stage of the life cycle and there was some accumulation of soluble carbohydrates. At low irradiances the maximum rate of photosynthesis is limited by the ability of the leaf to capture light, thus ATP and NADPH production are limiting rather than the capacity of the Calvin cycle. Consistent with this hypothesis, an increase in NPQ was not apparent in the A. candida ‐infected regions of the leaf when it was illuminated at lower irradiances; in contrast, NPQ was lower than in control leaves or in the uninfected region of the infected leaf. This can be seen clearly in the images of NPQ taken during photosynthetic induction at 93 µmol/m 2 /s. A low value of NPQ was also observed in all regions of a rust‐infected oat leaf at the flecking stage, only in the infected regions of the leaf at the sporulation stage ( Scholes and Rolfe, 1996 ) and in virus‐infected tobacco leaves ( Balachandran , 1994 ). Low values of NPQ may have been due to an increased demand for ATP which would result in a lower trans ‐thylakoid pH gradient. An increased demand for ATP is consistent with much evidence that suggests that biotrophic fungi stimulate host metabolic activity ( Roberts and Walters, 1988 ; Scholes and Farrar, 1986 ). Interpretation of fluorescence quenching parameters during photosynthetic induction is more difficult than in the steady state as many different processes within the leaf are being activated. However, the induction of Ø II and NPQ in control leaves was relatively simple. Ø II increased upon illumination reaching a maximum value after approximately 3 min. This correlated closely with an increase in transpiration as stomata opened. NPQ increased transiently and then declined as the Calvin cycle was activated using ATP and hence reducing the transthylakoid pH gradient. In contrast, a much more complicated pattern was observed in infected leaves with considerable heterogeneity within fungally invaded regions. Three distinct regions were visible in images of both Ø II and NPQ in infected leaves. Most notable was a 2–3 mm wide ring of cells surrounding the fungal pustule where Ø II was induced more rapidly than was observed in a control leaf. The alterations in the rate of induction of Ø II were unlikely to have resulted from altered rates of stomatal opening as the effect was first seen after just 2 s. Weis . (1998) demonstrated that zones of rapid photosynthetic induction are correlated with metabolically active regions of leaves where carbon export is low and respiration high. Thus, the distinct fluorescence characteristics may reflect high metabolic activity in this ring of cells at the leading edge of the colony. This is consistent with the starch print obtained for this leaf which revealed that there was greater amount of starch in these cells when compared to other regions of the infected leaf and to the control leaf. Thus images of NPQ and Ø II during photosynthetic induction reveal complex metabolic heterogeneity within the infected leaf. How does A. candida alter the amounts of storage and soluble carbohydrates in an infected leaf? In this study, the starch content of A. candida ‐infected regions of A. thaliana leaves declined, whilst in regions surrounding the infected area it increased as infection progressed. The loss of starch from the infected region was similar to that observed in leaves of Senecio squalidis infected with Albugo tragopogonis where starch disappeared from infection sites even before the disease symptoms were apparent ( Whipps and Cooke, 1978 ). It contrasts however, with the pattern of starch accumulation in radish cotyledons infected with Albugo candida ( Saettler and Pound, 1966 ) and barley leaves infected with brown rust Puccinia hordei ( Scholes and Farrar, 1987 ) where starch accumulation was localized to chloroplasts in cells within the fungal pustule. An increase or decline in the starch content of healthy leaves has been attributed to both a direct effect of Pi on the activity of chloroplast enzymes, e.g. ADPG pyrophosphorylase, and to the influence of cytosolic Pi on the availability of triose phosphates within the chloroplast. In addition, the amount of starch in a leaf will depend upon the sink or metabolic status of the leaf. In A. thaliana leaves infected with A. candida , the mechanism underlying the increase in starch in uninfected regions is unknown but is most likely to reflect a change in fluxes of sugars within the leaf. It is possible that Pi may have been limiting in these regions but the concentration of soluble sugars and the rate of photosynthesis were similar to control leaves. In infected regions of the leaf where starch was depleted, the rate of photosynthesis was low. It is likely that breakdown of starch contributed to the elevated amount of glucose in these regions, possibly due to an increase in demand for sugars by the fungus during sporulation. An increase in the amount of soluble carbohydrates (hexoses and sucrose) was apparent in infected regions of A. thaliana leaves from 6 days after inoculation. In the uninfected region of the infected leaf the soluble carbohydrate content was similar to control leaves except towards the end of the infection cycle (14 DAI) when an increase in all sugars was observed. An accumulation of glucose, fructose and sucrose has also been reported for whole infected leaves of Senecio squalidus infected with either Albugo tragopogonis ( Long and Cooke, 1974 ) or Puccinia lagenophorae ( Whipps and Cooke, 1978 ), vine leaves infected with Uncinula necator ( Brem , 1986 ), barley ( Scholes , 1994 ) and wheat ( Wright , 1995 ; Zulu , 1991 ) leaves infected with powdery mildew. Can an increase in host or fungal invertase activity account for the elevated amounts of hexoses in infected regions of the leaf? In A. thaliana leaves infected with A. candida there was an increase in both apoplastic and soluble invertase activity in infected areas of the leaf and the increase in soluble activity was partly contributed to by a new isoform. It is therefore likely that a large proportion of the increase in hexose content of invaded regions resulted from the increased invertase activity, since the timing of the increase correlated with the appearance of glucose and fructose, and the increase in activity was localized to invaded regions. Invertase activity did not increase in uninfected regions of the infected leaf and hexose sugars did not accumulate until the end of the infection cycle. There was also a good correlation between an increase in hexose concentrations and invertase activity in whole leaves of wheat infected with rust and powdery mildew ( Heisterüber , 1994 ; Wright , 1995 ; Zulu , 1991 ), oat and barley infected with rust ( Mitchell , 1978 ; Scholes , 1994 ) and vine infected with powdery mildew ( Brem , 1986 ). In addition, in infected leaves where hexoses did not accumulate there was also no stimulation in invertase activity, for example in rusted poplar leaves ( Roberts and Mitchell, 1979 ) and in vine leaves infected with downy mildew ( Brem , 1986 ). It is clear from this study that the increase in apoplastic invertase activity was of host origin, as mRNA levels of the ATβFRUCT1 gene increased dramatically in infected regions of the leaf. The induction of host apoplastic invertase activity and/or gene expression has been observed in two other host–pathogen interactions; carrot tap roots infected with the bacterial pathogen Erwinia carotovora ( Sturm and Chrispeels, 1990 ) and tomato roots infected with Fusarium oxysporum ( Benhamou , 1991 ). In the roots of some plants apoplastic invertase activity is thought to facilitate phloem unloading by maintaining a steep sucrose gradient between source (photosynthetic cells) and sink (heterotrophic cells) regions of the plant ( Eschrich, 1980 ) however, its role and cellular location in mature actively exporting leaves is less clear. Current evidence suggests that apoplastic invertase activity may be located in the leaf vasculature ( Kingston‐Smith and Pollock, 1996 ; Ramloch‐Lorenz , 1993 ; Zhang , 1996 ). Several studies have shown that apoplastic invertase gene expression declines suddenly as the leaves mature (although some activity remains) coinciding with the transition of the leaf from a sink to a source organ ( Godt and Roitsch, 1997 ; Sturm , 1995 ). In diseased leaves this situation is reversed and will have a number of consequences for both the plant and pathogen. Firstly, an increase in apoplastic invertase activity will reduce the amount of sucrose exported from the leaf, as has been shown in tobacco plants which over‐express yeast invertase in the apoplast ( Von Schaewen , 1990 ) but, in addition, it may facilitate phloem unloading of sucrose into cells adjacent to fungal mycelium thus converting regions of a source leaf into a sink for carbon. Such a strategy would aid efficient nutrient acquisition by the fungus and may lead to the down‐regulation of photosynthetic metabolism observed in these regions. At present the identity of the signal or the signal transduction pathways involved in the induction of invertase activity are unknown. A potential signal is the phyto‐hormone cytokinin which is produced by a number of fungi ( Pegg, 1981 ). Ehness and Roitsch (1997) showed that the application of cytokinin to suspension cultured cells of Chenopodium rubrum resulted in the direct stimulation of CIN 1 apoplastic invertase gene expression. Alternatively, elicitors such as cell wall fragments (oligogalacturonides, chitosans and pectic substances) have been demonstrated to stimulate the production of a range of wound‐induced and pathogenesis‐related proteins including invertase ( Ehness and Roitsch, 1997 ). Further studies are required to elucidate the signal transduction pathways involved. In addition to the increase in apoplastic invertase activity in A. candida ‐infected regions of A. thaliana leaves , there was an increase in the activity of soluble invertases which was again confined to the region of the leaf where fungal mycelium was present. To try to distinguish between host and fungal enzymes the profile of soluble invertase isoforms was examined in the different regions of the leaf by isoelectric focusing (IEF). Healthy A. thaliana leaves and the uninfected region of the infected leaf contained at least five major soluble isoforms which have characteristics typical of vacuolar invertases (e.g. acidic pIs). In the infected region, the host isoforms disappeared more rapidly than in control leaves and a new isoform with a higher pI appeared. Thus the majority of the increase in soluble activity measured in the infected region was due to this new isoform. If the new isoform was of host origin it had either resulted from the post‐translational modification of an existing soluble invertase isoform or it represented the expression of a hitherto unidentified host invertase gene. However, circumstantial evidence suggests that it is probably of fungal origin. Firstly, the kinetics of appearance of the isoform was very similar to the increase in biomass of the fungal mycelium and secondly, the isoform could not be induced in uninfected leaves by perturbations such as sugar feeding, wounding or accelerated senescence (N.B., unpublished data). Assuming that the new isoform was of fungal origin then its role in the metabolism of the infected leaf would critically depend upon its subcellular location. If located within the fungal hyphae then its ability to alter the carbohydrate metabolism of the host would be limited as it would not be able to hydrolyse sucrose in the extra‐haustorial matrix prior to uptake. However, if it was bound either to the hyphal membrane or excreted it would further contribute to the invertase activity already present within the host. What is the evidence that sugar sensing pathways may trigger a repression of photosynthetic gene expression and an induction of defence gene expression? Carbohydrate responsive genes have been classified as initiating ‘feast or famine’ responses ( Koch, 1996 ). Genes encoding photosynthetic proteins and those involved in resource mobilization are induced by carbohydrate depletion (the famine response) and repressed by an accumulation of sugars or an increase in fluxes of sugars through metabolic pathways. Conversely, a specific set of genes is up‐regulated in response to an accumulation of sugars (the feast response) most notably those involved in carbon storage and utilization and in plant defence. Signal transduction pathways for the repression of photosynthetic genes and for the induction of defence genes are largely unknown, but current evidence suggests the existence of a hexokinase‐dependent sugar sensing mechanism ( Jang , 1997 ; Sheen, 1994 ), a hexose‐transport associated sensor and a sucrose‐specific sensing pathway (reviewed by Smeekens and Rook, 1997 ). Evidence for hexose signalling associated with transport across a membrane into the cell comes from the use of glucose analogs such as 6‐deoxyglucose and 3‐O methyl glucose. These analogues can be transported across the membrane but are not metabolized. When added to cell suspension cultures or to intact leaves they initiate a sugar signal transduction pathway controlling the expression of some ‘feast’ and pathogen‐related genes ( Ehness , 1997 ; Roitsch , 1995 ). In a number of systems it has been shown that entry of hexoses into intermediary metabolism through the action of hexokinase initiates a sugar signal transduction pathway leading to the repression of typical ‘famine’ genes such as cab and rbcS (see Smeekens, 1998 ). Further evidence that hexoses may act as an internal sensor comes from the work of Jang . (1997) , who have over‐expressed and antisensed hexokinase genes in A. thaliana . Plants with reduced hexokinase activity are less sensitive to glucose whilst those with increased hexokinase activity are hypersensitive to glucose. To determine if the decline in the rate of photosynthesis and the accumulation of sugars observed in A. candida ‐infected regions of A. thaliana leaves were correlated with alterations in photosynthetic gene expression, cab and rbcS mRNA were measured by Northern blot analysis. In control leaves there was an age‐dependent decline in the abundance of both cab and rbcS gene expression which was evident from approximately 8–10 DAI. In infected leaves the expression of these genes was strongly reduced as early as 6 days after inoculation. The increase in the activity of acid invertases (apoplastic and soluble), the accumulation of hexoses and the down‐regulation of both cab and rbcS gene expression are consistent with the sugar sensing model. Although the content of hexoses in the infected regions did not reach the threshold level of hexoses (≈ 4.5 mmol/m 2 ) required to repress cab gene expression in transgenic tobacco expressing vacuolar or apoplastic yeast‐derived invertase ( Herbers , 1996 ), it was not far below. Moreover, the content of hexoses (≈ 3 mmol/m 2 ) in cold girdled spinach leaves did not reach this threshold even though the expression of photosynthetic genes was repressed ( Krapp and Stitt, 1995 ). In A. candida ‐infected A. thaliana leaves there was no induction of the defence proteins measured, despite the accumulation of carbohydrates in infected leaves. This suggests that either these proteins are not induced by sugars—at least via the same signal transduction pathway which may be involved in the repression of photosynthetic gene expression—or that defence gene expression was actively suppressed during the compatible interaction. Basse . (1992) reported that defence responses in tomato cells could be induced by elicitors (glycopeptides prepared from yeast invertase cleaved with alpha‐chymotrypsin) but inhibited by suppressors (oligosaccharides prepared from the elicitors by treatment with endo‐β‐N‐acetyl glucosaminidase H). It is also important to note that in plants undergoing a resistant interaction, the induction of invertase activity and defence proteins occurs very rapidly after the initial challenge (e.g. Benhamou , 1991 ) whereas in compatible interactions invertase activity, accumulation of sugars and the repression of photosynthetic gene expression occurs several days after infection. In conclusion, although the timing of events is consistent with a role for sugar signalling in this compatible host–pathogen interaction, more direct evidence is required. We are currently examining the effect of this pathogen on photosynthesis, carbohydrate metabolism and gene expression in A. thaliana plants with altered hexokinase activity ( Jang , 1997 ) and in plants which lack apoplastic invertase activity. Experimental Procedures Growth and inoculation of plants Seeds of Arabidopsis thaliana (ecotype OY0) were sown in Erin multipurpose compost and kept at 4 °C for 4 days to trigger germination. Seeds were then transferred to a growth room at a day temperature of 22 ± 1 °C and a night temperature of 15 ± 1 °C with a 9.5‐h photoperiod. Irradiance (200 µmol/m 2 /s) was provided by fluorescent lamps (OSRAM L 58 W/77 Fluora, Germany). Seedlings were transplanted into pots (6 × 6 cm 2 ) 2 weeks after sowing and grown in the same growth room. When the plants were 5 weeks old, eight of the youngest fully expanded leaves were inoculated towards the tip of the adaxial surface with four 10 µL droplets of zoospore suspension (10 6 zoospores/mL) as described in Tang . (1996) . Following inoculation, plants were kept in a propagating chamber for 24 h to maintain high humidity. Control plants were inoculated with water and treated in an identical manner. The region of the leaf inoculated with zoospores was termed region A, the uninoculated region of the same leaf, region B and similar regions of a control leaf, C and D. Fungal mycelium was confined to region A of the infected leaf. Determination of chlorophyll content The chlorophyll and carotenoid content of regions A–D were determined 4, 6, 8, 10 12 and 14 DAI following the procedure of Scholes . (1994) . The concentrations of chlorophyll a , chlorophyll b , and carotenoids were calculated from the equations of Lichtenthaler and Wellburn (1983) . Simultaneous measurement of CO 2 assimilation and chlorophyll fluorescence imaging Chlorophyll fluorescence imaging was performed as described in Rolfe and Scholes (1995) . Infected and control plants were removed from the growth room just prior to use, 6, 8, 11 and 14 DAI. Leaves (still attached to the plant) were placed in the chamber of an IRGA (LCA‐4 with PLCA4 leaf chamber; Analytical Development Company, Hoddesdon, Herts, UK) set to record CO 2 assimilation every 30 s. The CO 2 concentration was maintained at 340 µmol/mol with 60% relative humidity. The leaf was maintained in darkness for 5 min. During this time an image was captured representing zero fluorescence. The leaf was then exposed to a 1.5‐s pulse of saturating illumination and an image of chlorophyll fluorescence captured. Thirty seconds later the actinic light (93 µmol/m 2 /s) was switched on. The leaf was exposed to a saturating flash 2.5, 40 and then every 60 s after the commencement of actinic illumination. Two duplicate images were captured just before and at the peak of the saturating flash. During this time the leaf went through photosynthetic induction. The experiment lasted 1260 s in total. For the measurement of photosynthesis at different irradiances, the protocol described above was performed. Thereafter, the actinic irradiance was increased step‐wise up to a maximum of 1450 µmol/m 2 /s. The leaf was illuminated for 10 min at each irradiance and images of chlorophyll fluorescence were captured and stored every minute for the last 4 minutes. Analysis of soluble carbohydrates and starch Leaf discs (0.8 cm 2 ) were harvested from regions A, B, C and D of five control or infected plants 6 h into the photoperiod, 4, 6, 8, 10, 12 and 14 DAI. Samples were immediately frozen in liquid N 2 . Soluble carbohydrates were extracted in buffered 80% ethanol (50 m m Hepes‐NaOH 5 m m MgCl 2 pH 7.5) at 70 °C. Sucrose, glucose and fructose were measured by enzyme‐linked assay as described in Scholes . (1994) . Leaf discs remaining after the extraction of soluble carbohydrates were washed in distilled water and then ground in liquid N 2 in a pestle and mortar. One millilitre of 0.5 m Mes (2‐(N‐Morpholino) ethanesulphonic acid) buffer (pH 4.5) was then added together with 14 units of amyloglucosidase and 0.4 units of α‐amylase (Sigma UK). The samples were agitated overnight at room temperature. An aliquot (50 µL) was then assayed for glucose as described in Scholes . (1994) . After control and A. candida ‐infected leaves had been imaged (6, 8, 11 and 14 DAI) leaves were detached from the plant and placed in 80% ethanol for 7 days to remove chlorophyll. Leaves were then immersed in an iodine‐potassium iodide‐lactic acid mixture as described by Lindner . (1959) . Areas of the leaf containing starch turned black. In order to obtain semiquantitative data on the amount of starch in the different regions, leaves were imaged by placing them on a lightbox and capturing an image using a monochrome CCD camera. The intensity of staining along a 2.5‐cm transect was measured using the Optimas 4.0 image analysis package. This provided a quantitative measure of stain intensity and hence a semiquantitative measure of starch content. Measurement of invertase activity and soluble invertase isoform pattern Leaf discs (≈ 7.07 cm 2 ) were harvested from regions A, B, C and D from six control and six infected plants, 3 h into the photoperiod 0, 3, 6, 9, 14 and 18 DAI. Soluble and wall‐bound invertase activities were extracted and assayed, and isoelectric focusing (IEF) of proteins (to obtain the isoform pattern) performed, as described in Tang . (1996) . In order to visualize the invertase isoforms, the gel was stained for invertase activity as described in Faye (1981) . Quantification of ATβFRUCT1 gene expression by RT‐PCR A. thaliana contains two genes encoding apoplastic invertase, ATβFRUCT1 which is expressed in all tissues and ATβFRUCT2 which is expressed solely in reproductive tissues (N.B., unpublished data). The low abundance of the ATβFRUCT1 gene precluded detection by Northern blot analysis, hence semiquantitative reverse transcriptase linked polymerase chain reaction (RT‐PCR) was used to measure expression. To compensate for different reaction efficiencies during RT‐PCR, an internal RNA standard derived from a fragment of the ATβFRUCT1 gene was added to each sample. To generate this internal RNA standard, a 782‐bp portion of the ATβFRUCT1 gene was amplified by PCR using gene‐specific primers (primer 1: 5′‐CCCTTTCCACCGAAACTCTCCACT‐3′; primer 2: 5′‐GTGAGAGACTTTGGAGAAAGCGGAT‐3′), cloned and a small internal fragment (105 bp) removed by digestion with restriction enzymes SnaBI and Bpu1102I. The 5′ overhangs generated by the restriction enzyme digest were filled in using the large fragment of the DNA polymerase I and the DNA self‐ligated. This modified portion of the ATβFRUCT1 gene was designated pδβFRUCT1. The plasmid was then linearized by digestion with SalI and RNA (sense strand) generated by in vitro transcription with T7 RNA polymerase according to the manufacturer’s instructions (Promega). Template DNA was removed by digestion with RQ1 DNase and the RNA purified by extraction with phenol:chloroform followed by ethanol precipitation. The purified RNA, 15 µg produced from 5 µg template (DNA), was resuspended in 200 µL of RNAse‐free dH 2 O, quantified by absorption spectroscopy, and stored at −80 °C until use. RNA produced by in vitro transcription (677 bp) could be distinguished on the basis of size from both genomic DNA (782 bp) and authentic RT‐PCR product of the ATβFRUCT1 mRNA (349 bp). For RT‐PCR, RNA was extracted from infected and uninfected leaves of A. thaliana using the method of Loening (1969) . Five µg of sample RNA was mixed with 100 pg of standard RNA and brought to a total volume of 8 µL with RNAse‐free dH 2 O. The sample was incubated at 37 °C for 30 min with 1 unit of RQ1 RNAse‐free DNAse (Promega) and then incubated at 75 °C for 5 min to inactivate the nuclease. cDNA was then generated using random hexamer primers and commercially available first strand cDNA synthesis kit (Promega). Between 0.5 and 5 µg of cDNA were used in the subsequent amplification reactions. The PCR consisted of cDNA, 0.05 µ m primer 1, 0.05 µ m primer 2 (primers as above) 1.5 m m MgCl 2 and 150 µ m each of dATP, dCTP, dGTP and dTTP and supplied PCR buffer in a total volume of 100 µL. The reaction was brought to 90 °C before the addition of 2.5 units of Taq DNA polymerase. Cycling conditions were 94 °C for 30 s, 65 °C for 1 min and 72 °C for 1 min repeated 30 times. Reactions were held at 72 °C for 5 min at the end of the cycling reactions. These conditions had been empirically determined to result in linear amplification of the cDNA. Thirty µL of each PCR reaction mixture was electrophoresed through 1% (w/v) agarose gels, 40 m m Tris, 20 m m acetic acid, 1 m m EDTA buffer and then stained for 15 min in 0.5 µg/mL ethidium bromide. The gels were destained for 30 min in dH 2 O and then fluorescent DNA bands visualized by UV trans ‐illumination. An image of fluorescence intensity of each band was detected using a cooled CCD camera (C5985, Hamamatsu, Japan) using a Tiffin Orange 15 (Tiffin, USA) and BG18 (Schott, Mainz, Germany) filters. Optimas 4.0 imaging processing software (Bioscan Inc, Edmonds, USA) was used to analyse the intensity of each band and calculate the fluorescence intensity of the ATβFRUCT1 RT‐PCR product relative to that of the internal RNA standard. PCR amplification of genes encoding internal transcribed spacer (ITS) regions Amplification of the internal transcribed spacer (ITS) regions of A. thaliana and A. candida was carried out using the primers ITS1 and ITS4 ( White , 1990 ). Primer pairs were used in combination with DNA prepared from either control or infected leaves in PCR reactions according to the protocol of Wright . (1999) . After amplification, samples were electrophoresced through 1% (w/v) agarose gels and visualized using ethidium bromide. DNA fragments of the correct size (≈ 600 bp from A. thaliana and 700 bp from A. candida ) were isolated from the gel using a Wizard PCR Preps DNA purification kit (Promega UK). Fragments were cloned and sequenced as described in Wright . (1999) . The sequences were confirmed as ITS regions by homology searches using the NCBI BLAST programme ( Altschul , 1997 ). Northern blot analysis of photosynthetic and defence proteins Four leaf discs (approximately 100 mg fresh weight) were harvested from control and infected leaves 3 h into the photoperiod, 6, 8, 12 and 14 DAI for analysis of the expression of cab and rbcS , and 4 and 8 DAI for the analysis of the expression of defence proteins. Total RNA was isolated following the procedures of Loening (1969) . Northern blots were prepared using standard procedures ( Sambrook , 1989 ). Five µg of total RNA from each sample was separated by denaturing electrophoresis through 1% (w/v) agarose gels containing formaldehyde. The RNA was transferred to a nylon membrane (Zetaprobe, Biorad Ltd, UK) by capillary transfer overnight in 10 × SSC buffer (1.5 m NaCl, 0.15 m sodium citrate, pH 7.0) and then immobilized by UV cross‐linking. Radioactive DNA probes were prepared by random‐prime labelling of isolated DNA fragments (Megaprime, Amersham International Plc, UK). The probes were: a 600 and 700 bp Eco RI fragment from cDNA clones encoding the ITS region of A. thaliana and A. candida, respectively; an 1800‐bp Eco RI fragment isolated from the cDNA clone, CAB180, encoding a chlorophyll a/b binding protein ( Leutwiler , 1986 ); an 1800‐bp Hin dIII fragment isolated from a genomic DNA clone, pATS17, encoding the small subunit of Rubisco ( Krebbers , 1988 ); a 1398‐bp Bam HI/ Sac I fragment isolated from a genomic DNA clone, pSKPAL, encoding phenyl ammonia lyase (PAL) ( Dong , 1991 ); 1000 bp Bam HI/ Kpn I fragments isolated from cDNA clones encoding an acidic chitinase ( Genbank I.D. T46824) and a basic chitinase ( Genbank I.D. T21751); and a 1200‐bp Bam HI/ Kpn I fragment isolated from a cDNA clone encoding a basic glucanase (BG2) ( Genbank I.D. N37986). The membranes were hybridized with the radioactive probes for 18 h in 0.5 m Na‐phosphate buffer pH 7.2, 7% (w/v) SDS at 65 °C and then washed 3 times at 65 °C for 30 min in 80 m m Na‐phosphate buffer, pH 7.2, 5% (w/v) SDS and three times in 80 m m Na‐phosphate buffer, pH 7.2, 1% (w/v) SDS. Bound radioactivity was detected by autoradiography for 30 min (ITS), 24 h (photosynthetic proteins) or 5 days (defence proteins) using preflashed Biomax MS‐1 Film (Eastman Kodak Company UK) at −80 °C using one intensifying screen ( Sambrook , 1989 ). Statistical analysis The mean, standard deviation and standard error were calculated for each data set where appropriate. Standard error bars were plotted except where smaller than the symbol size. anova was used to identify the difference between infected region A, uninfected region B, and equivalent control regions C and D. Where appropriate, the student t ‐test was used to identify the difference between regions A and C, and regions B and D. The statistical analysis was carried out using the package Microsoft Excel 5.0.

Journal

Molecular Plant PathologyWiley

Published: Mar 1, 2000

There are no references for this article.