TY - JOUR
AU - Haro, Rosario
AB - Abstract This study aims to increase our understanding of the functions of CHX transporters in plant cells using the model plant Physcomitrella patens, in which four CHX genes have been identified, PpCHX1–PpCHX4. Two of these genes, PpCHX1 and PpCHX2, are expressed at approximately the same level as the PpACT5 gene, but the other two genes show an extremely low expression. PpCHX1 and PpCHX2 restored growth of Escherichia coli mutants on low K+-containing media, suggesting that they mediated K+ uptake that may be energized by symport with H+. In contrast, these genes suppressed the defect associated with the kha1 mutation in Saccharomyces cerevisiae, which suggests that they might mediate K+/H+ antiport. PpCHX1–green fluorescent protein (GFP) fusion protein transiently expressed in P. patens protoplasts co-localized with a Golgi marker. In similar experiments, the PpCHX2–GFP protein appeared to localize to tonoplast and plasma membrane. We constructed the ΔPpchx1 and ΔPpchx2 single mutant lines, and the ΔPpchx2 ΔPphak1 double mutant. Single mutant plants grew normally under all the conditions tested and exhibited normal K+ and Rb+ influxes; the ΔPpchx2 mutation did not increase the defect of ΔPphak1 plants. In long-term experiments, ΔPpchx2 plants showed slightly higher Rb+ retention than wild-type plants, which suggests that PpCHX2 mediates the transfer of Rb+ either from the vacuole to the cytosol or from the cytosol to the external medium in parallel with other transporters. The distinction between these two possibilities is technically difficult. We suggest that K+ transporters of several families are involved in the pH homeostasis of organelles by mediating either K+/H+ antiport or K+–H+ symport. Introduction Potassium is the most abundant cation in all types of living cells. In primitive organisms, the earliest and primary functions of K+ were probably electrical charge balance and osmotic adjustments. The fulfillment of these functions led to a K+-rich cellular environment, in which further physiological evolution resulted in a large number of physiological functions that became K+ dependent. Because of this dependence, K+ uptake is a vital process for all living cells; its failure results in a physiological collapse. In contrast to K+, in normal growth conditions including seawater, Na+ is maintained at low concentrations in the cytosol. Because living cells are not impermeable to Na+, Na+ efflux is another basic function in living cells. The same situation applies to H+, which is also maintained at a very low concentration in the cytosol, where the pH is around 7.0. In contrast, the pH in the lumen of organelles may differ significantly from neutrality. For all these functions, living cells exchange K+ for H+, K+ for Na+, and Na+ for H+, across both the plasma membrane and organellar membranes. Although all these functions are apparently simple, they must be performed in a wide variety of environmental conditions. This is especially important in organisms that can thrive in different environments with different ion concentrations, in which the K+, Na+ and H+ concentrations might be hundreds of times lower or higher than in cytosolic water. Probably as an adaptation to this variability, a large number of highly similar genes that may encode variants of K+ and Na+ transporters have evolved in different types of organisms, in which many of these putative transporters have unknown functions. This occurs in plants, in which a large research effort is currently focused on identifying the function of many of these putative transporters both in the plasma membrane (Chanroj et al. 2012, Gómez-Porras et al. 2012) and in the membranes of intracellular compartments (Orlowski and Gristein 2007, Casey et al. 2010). One of the largest families of putative K+ transporters in flowering plants is the CHX family (Sze et al. 2004, Chanroj et al. 2012). These transporters show sequence similarity to bacterial cation/H+ exchangers (Brett et al. 2005, Chanroj et al. 2012), and it has been shown that they are involved in K+ and H+ movements across endomembranes (Chanroj et al. 2011) and in K+ uptake (Zhao et al. 2008). Interestingly, the former function may overlap with the function of NHX and KEA transporters (Pardo et al. 2006, Chanroj et al. 2012) and the latter with that of HAK transporters (Grabov 2007). Moreover, because recently it has been found that HAK transporters also mediate the movement of K+ and H+ across endomembranes (Haro et al. 2013), an interesting question is whether transporters of the HAK and CHX families fulfill similar functions in endomembranes. Interestingly, it seems that plants with a higher number of CHX genes have a lower number of HAK genes, e.g. 28 and 13, respectively, in Arabidopsis (Ahn et al. 2004, Chanroj et al. 2012), and vice versa, e.g. 17 and 27, respectively, in rice (Yang et al. 2009, Chanroj et al. 2012). To study the cellular functions of plant K+ or Na+ transporters, we chose the moss plant model Physcomitrella patens (Garciadeblás et al. 2007, Fraile-Escanciano et al. 2010). This plant is anatomically much simpler than flowering plants, and large amounts of protonemal tissue can be grown easily. Although protonemal cells do not have counterparts in flowering plants, their growth depends on the same basic functions as the cells of flowering plants, e.g. bioenergetics and transport, cell wall synthesis, vesicle traffic, etc. In addition, many genetic tools, mainly gene disruption, are more efficacious in this plant than in flowering plants (Cove 2005, Cove et al. 2006, Quatrano et al. 2007). Regarding the CHX transporters, only four CHX genes have been identified in the genome of P. patens (Chanroj et al. 2012). This low number of genes in comparison with flowering plants offers an additional advantage for the study of the basic functions of CHX transporters. Results The CHX transporters of P. patens A phylogenetic study of the CHX transporters, including those of P. patens, has been recently published (Chanroj et al. 2012). Taking this report as the starting point of our study, Fig. 1 shows a phylogenetic tree restricted to the Arabidopsis and P. patens transporters, and the exon–intron structures of the P. patens CHX genes. Both types of analyses suggest that the four CHX genes of P. patens evolved via an early duplication of a single ancestor gene followed by two independent more recent duplications. These duplications gave rise to two distant pairs of genes: PpCHX1-2 and PpCHX3-4. Interestingly, the PpCHX1 and PpCHX2 transporters are closer to the Arabidopsis CHX16-20 than to the P. patens CHX3-4 transporters. Fig. 1 View largeDownload slide Sequence analysis of the four CHX transporters and exon–intron analysis of the corresponding genes of P. patens. (A) Phylogenetic tree of the P. patens (Pp) and Arabidopsis (At) CHX transporters, including the NhaS4 transporter of Synechococcus elongatus PCC 6301 (see Chanroj et al. 2012 for more details; alignment generated by the Clustal X program) (A). Exon–intron structure (inverted triangles denote introns) of the four P. patens CHX genes (B). Fig. 1 View largeDownload slide Sequence analysis of the four CHX transporters and exon–intron analysis of the corresponding genes of P. patens. (A) Phylogenetic tree of the P. patens (Pp) and Arabidopsis (At) CHX transporters, including the NhaS4 transporter of Synechococcus elongatus PCC 6301 (see Chanroj et al. 2012 for more details; alignment generated by the Clustal X program) (A). Exon–intron structure (inverted triangles denote introns) of the four P. patens CHX genes (B). To determine whether the four CHX genes were differentially expressed, we performed a search in the expressed sequence tag (EST) database of the NCBI website (http://www.ncbi.nlm.nih.gov/, accessed in June, 2012) using the sequences of the four CHX genes. The search retrieved 27 ESTs for CHX1 and 54 for CHX2, but none for either CHX3 or CHX4. Therefore, we investigated the existence of transcripts of these two genes by reverse transcription–PCR (RT–PCR), finding that in normal conditions of pH (from 5 to 7), or K+ and Na+ concentrations (from 0.1 to 10 mM) the expression of the CHX3 or CHX4 genes was not detectable when the other two genes showed very clear bands. In view of these results, we concentrated our study on the CHX1 and CHX2 genes. A real-time PCR study of the transcript expression of PpCHX1 and PpCHX2 revealed that both genes were expressed almost to the level of the ACT5 gene of P. patens. In standard growth conditions, PpCHX1 showed a 4-fold higher expression than PpCHX2. The expression of PpCHX1 was constant in most growth conditions, except for a 3-fold reduction at pH 4.0. In contrast, the expression of PpCHX2 was lower in standard conditions and slightly increased when the plants were stressed (Table 1). These results suggest that both PpCHX1 and PpCHX2 are housekeeping genes whose expression is required in all growing conditions. Although the transcript levels showed variability, the detected variations seem too low to have any biological relevance. Table 1 Real-time PCR determination of transcript expression of CHX1 and CHX2 genes of P. patens at different pH values in the presence and absence of Na+ or K+ Conditions  PpCHX1  PpCHX2  pH 5.8  0.84  0.19  pH 5.8 –K+  1.00  0.59  pH 5.8 + 100 mM Na+  0.94  0.73  pH 4.0  0.34  0.62  pH 9.0  0.98  0.42  Conditions  PpCHX1  PpCHX2  pH 5.8  0.84  0.19  pH 5.8 –K+  1.00  0.59  pH 5.8 + 100 mM Na+  0.94  0.73  pH 4.0  0.34  0.62  pH 9.0  0.98  0.42  Plants were grown for 15 d in the recorded conditions. Reported values are the ratio between the transporter transcript abundance and actin transcript abundance. Data represent results of two independent experiments with two replicates each. View Large Alternative PpCHX2 proteins The amino acid alignment of some conceptual translations of the PpCHX2 gene (e.g. PHYPADRAFT_191426, NCBI) revealed that the translated polypeptide lacked 29 residues in the C-terminal tail with reference to other conceptual translations. The missing residues, from 656 to 694, seemed to correspond to a putative intron that might be incorrectly eliminated by some conceptual translations. Consistent with this notion, we found that the corresponding mRNA fragment was present in all ESTs in the database. Furthermore, by RT–PCR, we could not detect the presence of CHX2 mRNAs in which this putative intron had been processed. However, the alternative splicing of this intron could not be absolutely excluded because some CHX transporters showed a gap coinciding with this 29 residue fragment that interrupts a region where CHX transporters retain high sequence homology (Supplementary Fig. S1). Therefore, an alternative splicing in PpCHX2 at the point described above was kept in mind throughout this study. Cloning of the PpCHX1 and PpCHX2 cDNAs and localization of the PpCHX1–GFP and PpCHX2–GFP proteins Next, we cloned the PpCHX1 and PpCHX2 cDNAs from plants growing in standard conditions. The two open reading frames (ORFs) of the cDNA sequences corresponded exactly with the two CDS sequences Pp1s123_30V6.1 and Pp1s162_22V6.1 in http://www.phytozome.net/ (accessed on September 27, 2012). The translated amino acid sequences, which showed 78% identity, presented a hydrophobic N-terminal half of 424 and 429 residues and a hydrophilic C-terminal tail of 420 and 398 residues (Supplementary Fig. S1). Furthermore, for the reasons discussed above, we constructed the shorter PpCHX2.1 clone, deleting the 87 bases of the putative intron in the PpCHX2 gene (Supplementary Fig. S1). To localize the proteins, we constructed expression vectors with the three CHXs tagged at the C-terminal end with green fluorescent protein (GFP) under the expression control of the promoter of the PpACT5 gene, which shows an expression level similar to those of PpCHX1 and PpCHX2 (Table 1). These constructs were then expressed in P. patens protoplasts. In these experiments, PpCHX1–GFP showed a punctuate pattern that was compatible with the Golgi network (Fig. 2A). Although the details of the images varied depending on the observed protoplast, the pattern shown in Fig. 2A was observed in all protoplasts showing the GFP signal. The Golgi network localization was confirmed because the GFP signal co-localized with the GmMan1–yellow fluorescent protein (YFP) fluorescent marker of the Golgi complex (Fig. 2B, C). Fig. 2 View largeDownload slide Localization of PpCHX1–GFP and PpCHX2–GFP fusion proteins in protoplasts of P. patens. Images of the green fluorescence of PpCHX1–GFP (A), PpCHX2–GFP (D) and PpCHX2.1–GFP (G). Images of the yellow fluorescence of the GmMan1–YFP marker of the Golgi complex (B) and the γ-TIP–YFP marker of the tonoplast (E, H). Merged images of the GFP, YFP marker fluorescence and chloroplast fluorescence are in red (C, F, I). Images show the maximum projection of 11 consecutive sections (A–C), 25 consecutive sections (D–F) and nine consecutive sections (G–I). Frequency of the green fluorescence patterns: (A) pattern appeared in all protoplasts that showed fluorescence with small differences; (D) pattern, most fluorescent protoplast showed peripheral and internal labelings, some protoplasts showed only internal labeling and very few only peripheral labeling; (G) pattern appeared in all protoplasts that showed fluorescence, with small differences. Fig. 2 View largeDownload slide Localization of PpCHX1–GFP and PpCHX2–GFP fusion proteins in protoplasts of P. patens. Images of the green fluorescence of PpCHX1–GFP (A), PpCHX2–GFP (D) and PpCHX2.1–GFP (G). Images of the yellow fluorescence of the GmMan1–YFP marker of the Golgi complex (B) and the γ-TIP–YFP marker of the tonoplast (E, H). Merged images of the GFP, YFP marker fluorescence and chloroplast fluorescence are in red (C, F, I). Images show the maximum projection of 11 consecutive sections (A–C), 25 consecutive sections (D–F) and nine consecutive sections (G–I). Frequency of the green fluorescence patterns: (A) pattern appeared in all protoplasts that showed fluorescence with small differences; (D) pattern, most fluorescent protoplast showed peripheral and internal labelings, some protoplasts showed only internal labeling and very few only peripheral labeling; (G) pattern appeared in all protoplasts that showed fluorescence, with small differences. The GFP signal of PpCHX2–GFP showed a dual localization, to the cell periphery, enclosing the chloroplasts, and to internal round structures (Fig. 2D, G). This double localization was found in most protoplasts; however, some showed only the internal round structures while very few showed only the peripheral signal. Interestingly, the PpCHX2.1–GFP protein localized only to the internal round structures in all protoplasts, which suggests that the missing residues with reference to PpCHX2 have localization determinants. However, currently it cannot be predicted whether this finding reflects a physiological process, e.g. alternative splicing, or if it is a fortuitous experimental response. In any case, the missing residues in PpCHX2.1 did not affect the functional expression of the protein in either Escherichia coli or yeast mutants (see below). The peripheral signal was characteristic of the plasma membrane, while the internal GFP signal appeared to correspond to the vacuolar membrane. To confirm the latter localization, we co-expressed either PpCHX2–GFP or PpCHX2.1–GFP and the γ-TIP–YFP marker of the tonoplast. These co-expression experiments showed that the GFP internal round structures and YFP signals co-localized, demonstrating the tonoplast localization of PpCHX2 (Fig. 2D–F) and PpCHX2.1 (Fig. 2G–I). Unfortunately, we could not co-localize PpCHX2 with a positive control of the plasma membrane. Therefore, the temporal or circumstantial localization of PpCHX2 to the plasma membrane is currently only a possibility, which does not have physiological support (see below). Growth rescue of Escherichia coli mutants The TKW4205 mutant of E. coli, defective in the K+ transport systems Kdp, TrkA and Kup (Schleyer and Bakker 1993), requires a high K+ concentration to grow at pH 5.5; this defect can be highly reduced by heterologous K+ transporters in an expression level-dependent manner (Senn et al. 2001). To test whether the PpCHX proteins transported K+, we cloned the PpCHX1, PpCHX2 and PpCHX2.1 cDNAs into plasmid pBAD24 (Guzman et al. 1995) under the control of the arabinose-responsive PBAD promoter. PpCHX1 partially rescued the growth of TKW4205 at 10 µM arabinose (not shown in Fig. 3), but when increasing the arabinose concentration to 100 µM the growth of TKW4205 was excellent at 5 mM (Fig. 3). As shown with other K+ transporters (Senn et al. 2001), K+ became toxic to E. coli when the expression level of PpCHX1 was increased by increasing the arabinose concentration up to 13 mM (Fig. 3 shows the growth inhibition at 13 mM arabinose, and 5 mM and higher concentrations of K+). However, when K+ was decreased to 2 mM, the PpCHX1 clone grew fairly well at 13 mM arabinose but not at 100 µM arabinose. Fig. 3 View largeDownload slide Growth at low K+, pH 5.5 of E. coli mutant strain TKW4205 transformed with empty vector pBAD24 or PpCHX1, PpCHX2 or PpCHX2.1 constructs. A colony of fresh transformed cells was picked and grown for approximately 3 h in LB medium supplemented with 50 mM KCl. Cultures were brought to a uniform cell density and 3-fold serial dilutions were placed on media at 100 µM or 13 mM arabinose and varying concentrations of KCl of 2, 5, 10, 20 or 50 mM. Photos were taken after 4 d of growth at 37°C. Fig. 3 View largeDownload slide Growth at low K+, pH 5.5 of E. coli mutant strain TKW4205 transformed with empty vector pBAD24 or PpCHX1, PpCHX2 or PpCHX2.1 constructs. A colony of fresh transformed cells was picked and grown for approximately 3 h in LB medium supplemented with 50 mM KCl. Cultures were brought to a uniform cell density and 3-fold serial dilutions were placed on media at 100 µM or 13 mM arabinose and varying concentrations of KCl of 2, 5, 10, 20 or 50 mM. Photos were taken after 4 d of growth at 37°C. PpCHX2 and PpCHX2.1 partially rescued the growth of TKW4205 at low K+ in the presence of 100 µM arabinose (Fig. 3). When increasing the arabinose concentration to 13 mM, the growth of the TKW4205 transformants was excellent at 5 mM, slightly positive at 2 mM K+ (Fig. 3), and was inhibited at 50 mM K+. In some kinetic studies it is necessary to measure zero-trans influxes, which cannot be performed with K+. In these studies, Rb+ might substitute for K+ if it is transported. Therefore, for future experiments, it was convenient to know whether PpCHX1 and PpCHX2 transported Rb+. This possibility can be tested in growth experiments because Rb+ can substitute for cellular K+ in a large number of cases without any toxic effects. Therefore, we repeated the growth experiments with TKW4205 described above using Rb+ instead of K+. The results showed that both PpCHX1 and PpCHX2 rescued the growth of TKW4205 at Rb+ concentrations that were only slightly higher than those found for K+, 10 mM Rb+ instead of 5 mM K+, at 100 µM arabinose for PpCHX1 and 13 mM arabinose for PpCHX2 (Fig. 4). These results demonstrated that both PpCHX1 and PpCHX2 are K+ transporters that show a notable capacity to transport Rb+. Fig. 4 View largeDownload slide Growth at low Rb+, pH 5.5 of E. coli mutant strain TKW4205 transformed with empty vector pBAD24 or PpCHX1, PpCHX2 or PpCHX2.1 constructs at low Rb+, pH 5.5. A colony of fresh transformed cells was picked and grown for approximately 3 h in LB medium supplemented with 50 mM KCl. Cultures were brought to a uniform cell density and 2-fold serial dilutions were placed on media at 100 µM or 13 mM arabinose and varying concentrations of RbCl of 10, 20 or 50 mM. Photos were taken after 4 d of growth at 37°C. Fig. 4 View largeDownload slide Growth at low Rb+, pH 5.5 of E. coli mutant strain TKW4205 transformed with empty vector pBAD24 or PpCHX1, PpCHX2 or PpCHX2.1 constructs at low Rb+, pH 5.5. A colony of fresh transformed cells was picked and grown for approximately 3 h in LB medium supplemented with 50 mM KCl. Cultures were brought to a uniform cell density and 2-fold serial dilutions were placed on media at 100 µM or 13 mM arabinose and varying concentrations of RbCl of 10, 20 or 50 mM. Photos were taken after 4 d of growth at 37°C. PpCHX1 and PpCHX2 complement the kha1 mutation in yeast Yeast mutants have been extensively used to characterize plant transporters functionally (Dreyer et al. 1999). The trk1 trk2 (to test K+ and Na+ influxes) and ena1-4 nha1 (to test Na+ and K+ effluxes) mutants have been widely used, but these mutants can be complemented only with plasma membrane transporters. Although only PpCHX2p might localize to the plasma membrane, we tested the function of the PpCHX1, PpCHX2 and PpCHX2.1 cDNAs in the above-mentioned yeast mutants, but none of them improved their defective growths at restrictive conditions, low K+, or high Na+ or K+. Negative results in this type of experiment may be explained because either the protein is not targeted to the yeast plasma membrane or the transporter is not active in yeast cells. The latter occurs frequently with HAK transporters (Rubio et al. 2000, Garciadeblás et al. 2007). Yeast mutants in the KHA1 gene have also been used for the functional expression of endomembrane transporters because KHA1 localizes to Golgi-like structures. In the absence of additional mutations, kha1 mutants do not show a clearly defective phenotype. In contrast, the kha1 ena1-4 nha1 (strain LMB01) or kha1 ena1-4 nha1 trk1 trk2 tok1 (strain LMM04) mutants show defective growth in selected conditions (Maresova and Sychrova 2005). This defective growth is suppressed by some plant CHX transporters (Maresova and Sychrova 2006, Zhao et al. 2008, Chanroj et al. 2011). To test whether PpCHX1, PpCHX2 and PpCHX2.1 suppressed the defect of the kha1 mutation, we transformed the kha1 ena1-4 nha1 mutant (LMB01) with the PpCHX1, PpCHX2 and PpCHX2.1 cDNAs. At pH 7.5, the three transporters improved the growth at low K+, 50 and 100 µM, and the effect was still observable at 1 mM K+ (Fig. 5A). It is worth highlighting that the defect that was suppressed by the PpCHX clones was exclusively due to the kha1 mutation, because the ena1-4 nha1 KHA1 strain (B3.1) grew at 50 µM K+, pH 7.5; in this strain, the PpCHX clones had no effect (Fig. 5B). At pH 6.5, the only defect of the kha1 ena1-4 nha1 mutant was a barely detectable impairment of growth at 50 µM K+, which was not significantly modified by the PpCHX clones. In contrast to these results with the ena1-4 nha1 kha1 mutant, the PpCHX1, PpCHX2 and PpCHX2.1 clones had no effect in the kha1 ena1-4 nha1 trk1 trk2 tok1 mutant strain (LMM04). Fig. 5 View largeDownload slide Functional expression of PpCHX1, PpCHX2 and PpCHX2.1 in yeast mutants LMB01 and B3.1. (A) LMB01 or (B) B3.1 cells transformed with empty pYPGE15 vector or PpCHX1, PpCHX2 or PpCHX2.1 constructs were grown overnight in growth medium (SD + adenine and tryptophan) supplemented with 50 mM KCl. On the following day, cultures were brought to a uniform cell density of 0.3, and 3-fold serial dilutions were placed on solid AP medium, pH 7.5, supplemented with KCl, 50 µM, 100 µM or 1 mM. Photos were taken after 4 d of growth at 28°C. Fig. 5 View largeDownload slide Functional expression of PpCHX1, PpCHX2 and PpCHX2.1 in yeast mutants LMB01 and B3.1. (A) LMB01 or (B) B3.1 cells transformed with empty pYPGE15 vector or PpCHX1, PpCHX2 or PpCHX2.1 constructs were grown overnight in growth medium (SD + adenine and tryptophan) supplemented with 50 mM KCl. On the following day, cultures were brought to a uniform cell density of 0.3, and 3-fold serial dilutions were placed on solid AP medium, pH 7.5, supplemented with KCl, 50 µM, 100 µM or 1 mM. Photos were taken after 4 d of growth at 28°C. Functional analyses of ΔPpchx1 and ΔPpchx2 plants To elucidate the in planta roles of PpCHX1 and PpCHX2, we disrupted the PpCHX1 and PpCHX2 genes individually, using the disruption fragments shown in Supplementary Fig. S2. We isolated four ΔPpchx1 and seven ΔPpchx2 lines, in which the hygromycin and zeocin resistance cassette substituted for the coding region of PpCHX1 or PpCHX2, respectively. These mutant lines were named ΔPpchx1-(1–4) and ΔPpchx2-(1–7). In all these lines, the absence of the corresponding transcripts was verified by RT–PCR. During the process of obtaining and multiplying all these lines, the growth and morphological characteristics of the mutant plants were absolutely normal. We found that in Ppchx1 or Ppchx2 plants the disruption of one of the two genes did not significantly affect the expression level of the other. In the case of the PpCHX3 and PpCHX4 genes, their transcripts were approximately 500 and 2,000 times less abundant, respectively, than PpCHX1 transcripts in wild-type plants (real-time PCR as described in the Materials and Methods, pH 5.8). Disruption of either PpCHX1 or PpCHX2 did not change these low expression levels of PpCHX3 and PpCHX4. With the aim of identifying any growth defect of the ΔPpchx1 or ΔPpchx2 plants, we applied a battery of growth tests including: high K+ or Na+ concentrations, K+ starvation, high and low pH values, high Ca2+ concentrations and the combination of some of them. In all cases, we found no differences between wild-type and mutant plants. Next, we studied K+ and Rb+ uptake, because a mild defect in uptake might not affect the growth of the mutant plants, and again we found no differences between wild-type and mutant plants. In its plasma membrane location, PpCHX2 might mediate either K+ influx or K+ efflux. In the former case, the absence of PpCHX2 should result in a defective K+ uptake. However, PpHAK1 mediates ‘active’ K+ influx (Garciadeblás et al. 2007), and the presence of this transporter might conceal any defect produced by the disruption of PpCHX2. Therefore, we constructed the ΔPphak1 ΔPpchx2 double mutant, which was tested in parallel with the ΔPpchx2 single mutant. First, we tested K+ or Rb+ uptake in three different preparations of plants: normal, K+ starved and Rb+ loaded, performing the tests at neutral, low and high pH values. In these tests, ΔPpchx2 plants did not show any appreciable defect. Then we used conditions of high Ca2+, in which hak1 plants showed more clearly its defective Rb+ uptake, but again the addition of the ΔPpchx2 mutation did not reveal any detectable effect due to the lack of the PpCHX2 transporter (Supplementary Fig. S3). The described tests were aimed at detecting defective K+ or Rb+ influxes, but the PpCHX2 transporter might mediate K+ efflux. Therefore, we tested the K+ content in wild-type and ΔPpchx2 plants, finding no significant differences. Next, to investigate other functions, we tested alterations in K+/Rb+ exchanges. For this purpose, we substituted 10 mM Rb+ for the K+ content of the culture medium and followed the external K+ concentration and the internal Rb+/K+ ratio for several days. Because at 10 mM Rb+ the re-uptake of the K+ lost by the plants is inhibited, a defective K+ efflux should result in a slower increase of external K+, but again we did not find differences that reveal a defective K+ efflux. In contrast, when we recorded the Rb+/K+ ratio, we found that the ratio was slightly higher in ΔPpchx2 plants. In general, the variability was lower in experiments carried out with the same batch of plants—e.g. a time course experiment (Fig. 6)—than in experiments with independent batches of plants (Fig. 6, inset). Although the differences were small, the statistical analyses of independent 5 d exchange experiments revealed that the Rb+/K+ ratio in ΔPpchx2 plants was significantly higher than in wild-type pants (2.24 ± 0.46 vs. 1.75 ± 0.21; n = 7; t-test, P = 0.024). In these experiments, the K+ loss to the external medium was determined with high precision and no differences were found between ΔPpchx2 and wild-type plants; on the other hand, we found that Rb+ uptake was not affected in ΔPpchx2 plants. Thus, the defect underlying the higher Rb+/K+ ratio in ΔPpchx2 plants seems to be a higher retention of Rb+ in the steady state—influx vs. efflux—that determines the Rb+ content. This may result from a slower transfer of Rb+ either from the vacuole to the cytosol, resulting in a higher vacuolar Rb+ content, or from the cytosol to the external medium, resulting in a higher cytosolic Rb+ content. Unfortunately, it is not currently possible to distinguish between these two possibilities by measuring Rb+ contents. Fig. 6 View largeDownload slide Rb+/K+ exchange in wild-type and ΔPpchx2 plants. Plants were grown at 10 mM K+ for 1 week then transferred to K+-free medium with 10 mM Rb+. Internal contents of Rb+ and K+ were measured and the Rb+/K+ ratio was calculated. Time course of the Rb+/K+ content in a typical experiment. Open circles, wild-type plants; filled circles, ΔPpchx2 plants. Inset, means of Rb+/K+ ratios in seven independent experiments measured on the fifth day for wild-type plants (filled bars) and ΔPpchx2 plants (open bars). Standard errors are shown and means are significantly different (P = 0.024) according to t-test. Fig. 6 View largeDownload slide Rb+/K+ exchange in wild-type and ΔPpchx2 plants. Plants were grown at 10 mM K+ for 1 week then transferred to K+-free medium with 10 mM Rb+. Internal contents of Rb+ and K+ were measured and the Rb+/K+ ratio was calculated. Time course of the Rb+/K+ content in a typical experiment. Open circles, wild-type plants; filled circles, ΔPpchx2 plants. Inset, means of Rb+/K+ ratios in seven independent experiments measured on the fifth day for wild-type plants (filled bars) and ΔPpchx2 plants (open bars). Standard errors are shown and means are significantly different (P = 0.024) according to t-test. Discussion Functional studies of CHX transporters have started recently, so far limited to Arabidopsis (Chanroj et al. 2012); the emerging model suggests that these transporters play important roles in the cellular homeostasis of K+. Although two of these transporters, AtCHX13 (Zhao et al. 2008) and AtCHX21 (Hall et al. 2006), are apparently functional in the plasma membrane, most CHX transporters appears to reside in pre-vacuolar, endoplasmic reticulum and Golgi membranes, or in unspecific endomembranes (Padmanaban et al. 2007, Chanroj et al. 2011, Lu et al. 2011). The localization of PpCHX1 to the Golgi complex and PpCHX2 to the tonoplast and possibly to the plasma membrane is consistent with the Arabidosis model. Some chx mutants of Arabidopsis show defects in the opening of stomata or the guidance of the pollen tube (Padmanaban et al. 2007, Lu et al. 2011, Evans et al. 2012). These defects cannot be studied in protonemal cultures of P. patens, in which these structures do not exist. However, the basic cellular functions mediated by the CHX transporters that are primarily altered in these mutants (e.g. membrane trafficking events that affect protein and cargo sorting; Chanroj et al. 2012) are probably identical or very similar in Arabidopsis and P. patens. The mechanism of CHX transporters The functional expression of PpCHX1 and PpCHX2 in E. coli mutants (Fig. 4) was similar to previous findings with AtCHX17, AtCHX20 (Chanroj et al. 2011) and AtCHX23 (Chanroj et al. 2011, Lu et al. 2011), which demonstrates that all these proteins mediate K+ transport. In E. coli, PpCHX1 and PpCHX2 seem to have identical functions but with different specific activities, as suggested by the different expression levels of the arabinose-responsive PBAD promoter that are required in each case. For rapid growth on 5 mM K+, the PpCHX1 clone required 100 µM arabinose while the PpCHX2 clone required 13 mM arabinose, which represents >10-fold different transcript expression levels (Guzman et al. 1995). Although 13 mM arabinose was toxic for the PpCHX1 clone at K+ concentrations of ≥5 mM, growth at 2 mM K+ occurred at 13 mM but not at 100 µM arabinose. These results and the slow growth of the PpCHX2 and PpCHX2.1 clones at 2 mM K+ can be explained by the rate of K+ uptake, which depends on the number of transporters, K+ concentration and K+ influx kinetics, i.e. the Km and specific activity of the system. If the rate of K+ influx is too low or too high, the cells do not grow. These observations are important regarding the mechanism of transport discussed below because they strongly suggest that the growth at low K+ is limited by the kinetics and not by the thermodynamics of the system. To understand the physiological function of CHX transporters, the study of their functional mechanism is of crucial importance; for this issue, the mechanism underlying the CHX-mediated K+ uptake in E. coli constitutes the most basic information. In our experiments (Fig. 3), this uptake may be mediated by either a K+ uniporter—‘passive’ transport—or a K+–H+ symporter—‘active’ transport—but not by an electroneutral K+/H+ antiporter, which would mediate K+ efflux. The simplest way to distinguish between the uniport and symport mechanisms is to calculate whether the value of the membrane potential can explain the internal/external ratio of K+ concentrations in growing cells of E. coli (Rodríguez-Navarro et al. 1986). To calculate this ratio, the external K+ concentration is 2 mM, because the PpCHX1 clone was found to grow at this concentration. In a medium without Na+, the internal concentration of K+ in growing cells of E. coli is 211 mM (Schultz and Solomon 1961). This concentration might be reduced at K+-limited growth rates, but not very much considering that, even in the presence of NH+4 and Na+, the K+ content of chemostat cultures of Enterobacter aerogenes is not greatly reduced (Tempest and Strange 1966). According to these data, an internal K+ concentration <150 mM is absolutely unlikely, which corresponds to a minimal internal/external ratio of 75. To obtain this ratio by ‘passive’ uptake would require a membrane potential of −112 mV, which is unlikely to be generated by E. coli at pH 5.5. The membrane potential in E. coli has been extensively studied (Padan et al. 1976, Zilberstein et al. 1979, Felle et al. 1980, Bakker and Mangerich 1981, Kashket 1982, Setty et al. 1983). According to these studies, a likely value of the membrane potential of E. coli at pH 5.5 is between −90 and −95 mV. The most negative value is given in the study of Felle et al. (1980) by microelectrode recording, reporting a measured value of −80 mV that is corrected to −100 mV after calculating the effect of the current leakage through the membrane seal. To interpret these calculations, it is worth noting that a positive growth is an unequivocal proof of fulfilling an energetic requirement, while a negative growth might be the consequence of an influx that is too slow to support a detectable growth rate. We discussed above that the growth of the PpCHX1 and PpCHX2 transformants of TKW4205 at low K+ is probably limited by the kinetics and not by the thermodynamics of the system. The discussion above can be applied to the Arabidopsis transporter AtCHX20 based on similar growth tests of E. coli transformants (Chanroj et al. 2011). In summary, the thermochemical calculations do not prove, but strongly suggest, that some plant CHX transporters mediate K+–H+ symport in E. coli. This mechanism is of crucial importance to develop a comprehensive functional model of these transporters that explains the results obtained with the yeast mutants. In contrast, it is worth noting that if the mechanism of the CHX transporters was K+ uniport they would be equivalent to K+ channels (without voltage gating), through which K+ moves but not H+. Using a different approach, the same issue has been previously discussed by Chanroj et al. (2011) for the Arabidopsis transporters AtCHX17 and AtCHX20. Both PpCHX1 and PpCHX2 suppressed the defect of the kha1 ena1-4 nha1 mutant of Saccharomyces cerevisiae at pH 7.5 (Fig. 5), which again coincides with previously reported results for several Arabidopsis CHX transporters (Maresova and Sychrova 2006, Padmanaban et al. 2007, Chanroj et al. 2011). Assuming that the KHA1 gene of S. cerevisiae encodes a K+/H+ antiporter, a tentative interpretation of the results obtained with the heterologous expression would be that CHX transporters mediate K+–H+ symport in E. coli and K+/H+ antiport in S. cerevisiae. However, an alternative explanation would be that K+–H+ symporters and K+/H+ antiporters fulfill similar functions in endosomal compartments and that these two mechanisms show a certain degree of functional redundancy and reciprocal substitution. This means that a K+–H+ symporter could substitute for KHA1, even if KHA1 mediates K+/H+ antiport. This reasoning would apply not only to CHX transporters but also to PpHAK2, which also suppresses the kha1 mutation of S. cerevisiae (Haro et al. 2013). The basis of this notion is that these two mechanisms participate in the pH control of organelles. NHX, HAK and CHX transporters are probably present in the membrane of most organelles (see below). Therefore, the possibility that the functional mechanisms associated with these transporters might be either K+–H+ symport or K+/H+ antiport raises the question of how these two mechanisms can participate in the pH control of the organellar lumen. In the most likely model, the organelle pH is established by the steady state that results from H+ pumping into the organelle and returning to the cytosol, in parallel with K+ and Cl− conductances (Demaurex 2002, Paroutis et al. 2004, Casey et al. 2010, Ohgaki et al. 2011). H+ pumping and a parallel influx of anions would decrease the organelle pH to very low values, but the effective control of the pH requires the return of H+ to the cytosol. H+ pumping is mediated by the electrogenic V-ATPase, while the return of H+ can be mediated by multiple systems: H+ passive leaks and fluxes coupled to Cl− or K+ fluxes. In plant cells, a specific type of pyrophosphatase might cooperate with the H+ pump V-ATPase (Segami et al. 2010). The question raised above refers to K+ coupling and can be addressed with a simple model including the pump, the coupled H+ efflux and a K+ channel, if necessary (Fig. 7). Assuming that the membrane potential drives all the other movements, a simple calculation shows that both a K+/H+ antiporter and a K+–H+ symporter are similarly effective to return H+ from the organelle lumen to the cytosol, assuming the existence of a K+ channel for K+ recirculation. If the equilibrium is reached, which is the limit of the gradient that can generate the system, the organelle pH could be higher than the cytosolic pH (ΔpH would be 1 for a ΔΨ of −60 mV). It is worth noting that the three couplings, antiport plus channel, symport plus channel and antiport plus symport, would be similarly effective to return H+ to the cytosol (Fig. 7), but they would be associated with different K+ contents. Fig. 7 View largeDownload slide Alternative models of action for PpCHXs in endomembrane compartments. (A) Coupling of: a K+–H+ symporter with a K+ channel, (B) a K+/H+ antiporter with a K+ channel and (C) a K+–H+ symporter with a K+/H+ antiporter. The equations for the systems in equilibrium are used to calculate the ΔpH that the system can attain. ΔΨ denotes the membrane potential, negative in the cytosolic side; for calculations, 2.3 RT/F = 60 mV. Fig. 7 View largeDownload slide Alternative models of action for PpCHXs in endomembrane compartments. (A) Coupling of: a K+–H+ symporter with a K+ channel, (B) a K+/H+ antiporter with a K+ channel and (C) a K+–H+ symporter with a K+/H+ antiporter. The equations for the systems in equilibrium are used to calculate the ΔpH that the system can attain. ΔΨ denotes the membrane potential, negative in the cytosolic side; for calculations, 2.3 RT/F = 60 mV. Functions of PpCHX1 and PpCHX2 in planta To analyze the function of PpCHX1 and PpCHX2 in plant cells, we disrupted these genes and also constructed the ΔPpchx2 ΔPphak1 double mutant. PpCHX1 localized to the Golgi complex, and ΔPpchx1 plants showed no growth defects when growth was tested in a large variety of conditions of pH, and K+, Na+ and Ca2+ concentrations. This suggests that the function of PpCHX1 may be replaced by other transporters. As already discussed, the mechanism of the redundant transporters might be either K+–H+ symport or K+/H+ antiport. PpHAK3 also localizes to the Golgi (Haro et al. 2013), which makes it a candidate substitute for PpCHX1, but there are also other candidates. In Arabidopsis, NHX5 and NHX6 are associated with the Golgi and the trans-Golgi network (Bassil et al. 2011), and two NHX transporters in P. patens show high sequence homology with AtNHX5 and AtNHX6 (Chanroj et al. 2012). Therefore, it might be possible that CHX, HAK and NHX transporters are working in parallel in the Golgi membrane, returning H+ from the lumen of the Golgi apparatus to the cytosol. In the case of ΔPpchx2 plants, no defects were found that connect CHX2 to K+ (or Rb+) uptake, but we found that these plants showed a higher cellular retention of Rb+. When ΔPpchx2 plants were exposed to Rb+, they showed an increased Rb+/K+ ratio but the same K+ loss. Considering that Rb+ influx was not increased by the mutation, the increased Rb+ content of ΔPpchx2 plants must be the consequence of either a slower Rb+ efflux through the plasma membrane or a slower vacuole to cytosol Rb+ transfer, which results in a higher vacuolar Rb+ retention. Because we did not detect differences in K+ contents and the differences in Rb+ contents between ΔPpchx2 and wild-type plants were small, the most likely hypothesis is that PpCHX2 mediates K+ and Rb+ movements in parallel with other transporters that exhibit a higher K+/Rb+ discrimination either in the plasma membrane or in the tonoplast. These transporters might fully substitute PpCHX2 for K+ transport, but only partially for Rb+ transport. PpCHX2–GFP localized to the tonoplast, but it might also localize to the plasma membrane. The functional difference between these two localizations is mechanistic because Rb+ efflux (K+ efflux in physiological conditions) across the plasma membrane must be Rb+/H+ antiport while vacuole to cytosol Rb+ transfer (K+ transfer in physiological conditions) must be either Rb+ uniport or Rb+–H+ symport, assuming that, in physiological conditions, the vacuole has low pH, high K+ content and a weak membrane potential that is positive with reference to the cytosol. Considering these observations and the previous discussion about the mechanism of CHX transporters, the most likely possibility is that PpCHX2 mediates K+ (or Rb+)–H+ symport in the tonoplast. Unfortunately, taking into account the large volume of the vacuole, the technical tools to distinguish from a slightly higher Rb+ content in the cytosol or in the vacuole between mutant and wild-type plants are practically non-existent. Therefore, at the current level of knowledge, a precise conclusion cannot be reached and it is doubtful that it can be reached from biochemical experiments. Most probably, the identification of the functions of both PpCHX1 and PpCHX2 will come from a genetic approach, by constructing double or triple mutants that show clear defects. However, the problem of this approach is that the number of non-CHX transporters that could substitute for the CHX transporters is high. In the plasma membrane, K+ efflux is mediated by the PpENA1 ATPase (Fraile-Escanciano et al. 2009), and in the Golgi complex, we have already discussed the presence of PpHAK3 and NHX transporters. Regarding the vacuole, P. patens have five NHX transporters showing high sequence homology to the NHX1–NHX4 transporters of Arabidopsis, which play vital K+/H+ exchange roles in vacuoles (Barragán et al. 2012, Chanroj et al. 2012). In conclusion, despite the phylogenetic distance between Arabidopsis and P. patens, the functional basis of CHX transporters in both species seems to be very similar. Taken together, the present study and other studies on CHX, HAK and NHX transporters in P. patens and Arabidopsis provide compelling evidence about the existence of a complex series of K+ transporters in plant organelles. Although these systems might not be essentially redundant considering their functional conditions, they might replace each other in mutant plants. Although some double CHX mutants in Arabidopsis are clearly defective (Lu et al. 2011, Evans et al. 2012), the construction of double mutants involving two different transporter families seems the only way to unravel the complexity of the individual functions of endomembrane K+ transporters in plant cells. Materials and Methods Plant growth conditions The moss P. patens was routinely grown in BCDAT medium (Ashton et al. 1979), supplemented with 7 g l−1 agar when required, as described elsewhere (Garciadeblás et al. 2007). Physiological tests were performed in liquid medium with continuous white light (Haro et al. 2010). Plants were normally grown in glass bottles with air bubbling in a phytochamber at 25°C. The bottles were inoculated with moss suspensions that were fragmented using a Politron PT2100 homogenizer (Kinematica AG). KFM is a K+- and Na+-free medium (Garciadeblás et al. 2007) that was used for K+ starvation and supplemented with K+ or Na+ for growing plants at controlled cation concentrations. Control plants were grown in KFM with 4 mM KCl. K+-starved plants were obtained by transferring moss plants to KFM medium supplemented with no KCl and grown there for 7–10 d. KFM pH was adjusted to 5.8 as a standard condition. In the case of the different physiological tests performed, KFM was adjusted to either pH 9.0 with 20 mM TAPS [N-Tris(hydroxymethyl)methyl-3-aminopropanesulfonic acid] or to pH 4.5 with 5 mM tartaric acid. The inocula with which we started all experiments were almost exclusively protonemata. Yeast and E. coli strains, plasmids, media and growth conditions Functional expression of the PpCHX1 and PpCHX2 cDNAs was performed in derivatives of the yeast strains W303-1A (MATa ade2-1 can1-100 his3-11,15 leu2-3,112 trp1-1 ura3-1 mall0), B3.1 (ena1Δ::HIS3::ena4Δ,nha1Δ::LEU2) (Bañuelos et al. 1998), WΔ6 (trk1Δ::LEU2, trk2Δ::HIS3) (Haro and Rodríguez-Navarro 2003), LMB 01 (Mata ade2-1 can1-100 his3-11,15 leu2-3,112 trp1-1 ura3-1 mall0 ena1Δ::HIS3::ena4Δ nha1Δ::LEU2 kha1Δ::kanMX:tok1Δ) and LMM04 (Mata ade2-1 can1-100 his3-11,15 leu2-3,112 trp1-1 ura3-1 mall0 ena1Δ::HIS3::ena4Δ nha1Δ::LEU2 trk1Δ::LEU2 trk2Δ::HIS3 kha1Δ::kanMX:tok1Δ) (Maresova and Sychrova 2005) which were used for functional tests. For this purpose the full-length cDNAs were cloned into vector pYPGE15 (Brunelli and Pall 1993) as previously described (Garciadeblás et al. 2007). Yeast transformants were routinely grown in SD medium (Sherman 1991) supplemented with 50 mM K+. For growth experiments, the yeast strains were inoculated on arginine phosphate (AP) medium (Rodríguez-Navarro and Ramos 1984) supplemented with the indicated K+ concentrations and with 10 mM MES-Ca2+ for pH 6.0, 10 mM TAPS-Ca2+ for pH 7.5 and 10 mM tartaric acid-Ca2+ for pH 4.5. AP medium was solidified with 1.5% Difco™ Agar Noble (Becton, Dickinson and Company, http://www.bd.com/). This solid medium contained 16 µM K+ and 770 µM Na+. For functional expression in E. coli, the cDNAs were cloned into plasmid pBAD24, a bacterial multicopy plasmid of inducible expression containing the pBAD promoter, its regulatory gene araC, which is dependent on arabinose concentration, and an ampicillin resistance cassette (Guzman et al. 1995). These constructs were transformed into bacterial strain TKW4205 (thi rha lacZ nagA recA Sr::Tn10 ΔkdpABC5 trkA405 kup1), which is deficient in the three K+ uptake systems, Kdp, TrkA and Kup (Schleyer and Bakker 1993). The expression level was modulated varying the arabinose concentration of the growth medium as previously described (Senn et al. 2001). Recombinant DNA techniques The E. coli strain DH5α was routinely used for plasmid DNA propagation. The manipulation of nucleic acids was performed by standard protocols or, where appropriate, according to the manufacturer’s instructions. Total P. patens RNA was prepared using the RNeasy Plant Kit and DNeasy Plant Kit (Qiagen USA, http://www.quiagen.com). The PCR amplification of cDNAs was performed with a double-stranded cDNA Synthesis System Kit (GE Healthcare, http://www.gehealthcare.com). PCR was performed in a Perkin-Elmer thermocycler, using the Expand-High-Fidelity PCR System (Roche Molecular Biochemicals, http://www.roche-applied-science.com). The resulting PCR fragments were first cloned into the PCR2.1-Topo vector using the TOPO TA Cloning Kit (Invitrogen, http://www.invitrogen.com/site/us/en/home.html). DNA sequencing was performed in an automated ABI PRISM 377 DNA sequencer (Perkin-Elmer Applied Biosystems). Genomic DNA from Physcomitrella was obtained using the cetyltrrimethylammonium bromide (CTAB) method (Aono et al., http://moss.nibb.ac.jp/). Physcomitrella patens protoplasts were transformed following the polyethylene glycol method (Hohe et al. 2004). For expression in yeast cells, the full-length PpCHX1 and PpCHX2 cDNAs were amplified from P. patens mRNA using the specific forward and reverse primers that amplified fragments that included the ATG and STOP triplets triplet codons (Supplementary Table S1). The resulting PCR fragments were first cloned into the PCR2.1-Topo vector. For expression in yeast, the PCR2.1-Topo constructs were digested with BamHI and KpnI, and the fragments containing the cDNAs were then ligated into the yeast expression vector pYPGE15. For expression in E. coli, the cDNAs were amplified by PCR using primers that included the appropriate restriction enzyme sequences to direct the cloning into the corresponding sites of the pBAD24 vector polylinker. The forward primer used to amplify the 5′ end of any the cDNAs contained an optimized environment of nucleotide sequences, including the NheI site of pBAD24 situating the ATG triplet at 7 bp from the Shine–Dalgarno box present in the plasmid. The reverse primer used to amplify the 3′ end of the cDNAs included the stop codon with the KpnI site sequence (Supplementary Table S1). In all cases, the fidelity of PCR amplifications was confirmed by DNA sequencing. The in vitro construction of the PpCHX2.1 cDNA was carried out by a two-step PCR method using the PpCHX2 cDNA as a template, as detailed in Supplementary Fig. S4. The fragment removed extended from nucleotide positions 1,988 to 2,075 of the cDNA. The primers used to produce removal of the desired fragment are described in Supplementary Table S1. The resulting PCR fragments were first cloned into the PCR2.1-Topo vector using the TOPO TA Cloning Kit (Invitrogen). For expression in yeast and E. coli, the verified PpCHX2.1 cDNA (cloned in PCR2.1-Topo) was treated as described above for PpCHX1 and PpCHX2. Localization of PpCHX1–GFP and PpCHX2–GFP Localization of the PpCHX1–GFP and PpCHX2–GFP fusion proteins in P. patens was performed by transient expression in protoplasts. The PpCHX1-GFP and PpCHX2-GFP genes were cloned under the control of the PpACT5 promoter containing a large intron in the 5′ untranslated region (5′ UTR). The expression plasmid, pRHACT5, was constructed from the PCR2.1-Topo vector. For this purpose, the 5′ UTR-PpAct5 fragment was amplified by PCR using the primers published by Weiss et al. (2006): number 13 as the forward primer and number 14 as the reverse primer; the latter was modified to contain a BamHI cloning site at the 3′ end of the primer. Then, a BamHI–HindIII fragment from the pMF6 vector containing the GFP:nos-3′ terminator (Rubio-Somoza et al. 2006) was cloned downstream of the PpACT5 promoter. Finally, the PpCHX1 and PpCHX2 cDNAs were cloned in-frame into the BamHI site at the 5′ end of the GFP gene. Experiments of co-localizations were performed with organelle markers used in others plants (Nelson et al. 2007, http://www.bio.utk.edu/cellbiol/markers/). The tonoplast marker was achieved by fusing the coding region of YFP to the C-terminus of γ-TIP, an aquaporin of the vacuolar membrane (Saito et al. 2002). The Golgi localization was based on the fusion of YFP with the cytoplasmic tail and transmembrane domain (first 49 amino acids) of GmMan1, soybean α-1,2-mannosidase I (Saint-Jore-Dupas et al. 2006). The resulting constructs were used for transient expression in P. patens protoplasts. After transformation, the protoplasts were kept in the dark for 24 h in BCDAT medium supplemented with 6% mannitol and 5% glucose, followed by cultivation in the same medium for 3–4 d under normal growth conditions. Z-series of protoplasts were obtained on a Leica TCS SP8 confocal microscope (LeicaMicrosystems). Images were processed using the LAS AF Lite 3.1.0 (Leica Microsystems). The GFP signal and Chl autofluorescence were collected simultaneously under the laser excitation lines of 488 and 633 nm, respectively. To rule out any cross-talk between GFP and YFP, the YFP signal was collected temporally separated (sequential scan) under the laser excitation line of 514 nm. However, we had internal controls. Because the CHX–GFP fusion proteins and YFP markers were expressed from different plasmids, we always had more protoplasts expressing only the GFP or YFP protein than protoplasts expressing both proteins. Protoplasts expressing the proteins separately were used to check the absence of cross-talk in our experimental conditions. Real-time PCR assays To analyze the expression of the PpCHX genes, the plants were grown for 1 week in the conditions shown in Table 1. cDNA was synthesized from RNA extracted from plants grown under these conditions as described above. Real-time PCRs were performed using an ABI Prism 7000 Sequence Detection System (Perkin-Elmer Applied Biosystems) and SYBR Green PCR Master Mix (Perkin-Elmer Applied Biosystems). PCR amplifications were carried out according to standard procedures as described in Garciadeblás et al. (2003). The results were expressed as the transcript level ratios between the studied PpCHX and actin genes in the same preparation. PCR primers were designed to amplify the following fragments (Supplementary Table S1, numbered according to their CDS sequences): PpACT5 (from nucleotide 1,281 to 1,422, sequence 109052 in the JGI P. patens genome), PpCHX1 (from nucleotide 1,903 to 2,051), PpCHX2 (from nucleotide 2,102 to 2,260), PpCHX3 (from nucleotide 1,501 to 1,654) and PpCHX4 (from nucleotide 1,510 to 1,621). Each pair of primers was used in PCR experiments with all the purified cDNAs to check that they amplified only the fragment for which they were designed. Generation of ΔPpchx1 and ΔPpchx2 knockout lines The pTN186 vector (T. Nishiyama, http://moss.nibb.ac.jp/) was used to construct the PpCHX1:Hyg knockout fragment (Supplementary Fig. S6). In this fragment, the hygromycin resistance cassette, which contained the aph4 gene under the control of the promoter and terminator of the 35S gene, was flanked by two fragments of the non-coding 5′ and 3′ regions of the gene PpCHX1. These fragments were amplified by PCR using primers that included the restriction enzyme sequences to direct the cloning into the corresponding sites of the pTN186 vector polylinker. The 5′ fragment extended from nucleotide positions −975 to 80 and was inserted between the KpnI and HindIII sites. The 3′ fragment extended from position 3,034 to 4,103 and was inserted between the restriction sites XbaI and SacI of the pTN186 vector polylinker. The ΔPpchx1 knockout lines were generated by transforming P. patens protoplasts with 25 µg of the linear fragment obtained by digesting the pTN186:Ppchx1 plasmid with the KpnI and SacI restriction enzymes. Stable antibiotic-resistant clones were selected after two rounds of incubation in BCDAT medium supplemented with 30 µg ml−1 of hygromycin. The p35S-Zeo vector (T. Nishiyama, http://moss.nibb.ac.jp/) was used to construct the PpCHX2:Zeo knockout fragment (Supplementary Fig. S6). In this fragment, the zeocin resistance cassette, which contained the ble gene under the control of the promoter and terminator of the 35S gene, was flanked by two fragments of the non-coding 5′ and 3′ regions of the PpCHX2 gene. These fragments were amplified by PCR using primers that included the restriction enzyme sequences to direct the cloning into the corresponding sites of the vector polylinker. The 5′ fragment extended from nucleotide positions −1,066 to 314 of the ATG and was inserted between the XhoI and EcoRI sites. The 3′ fragment extended from positions 2,081 to 3,114 and was inserted between the restriction sites SacII and NotI. The ΔPpchx2 knockout lines were generated by transforming P. patens protoplasts with 25 µg of the linear fragment obtained by digesting the p35S-Zeo:Ppchx1 plasmid with the XhoI and NotI restriction enzymes. Stable antibiotic-resistant clones were selected after two rounds of incubation in BCDAT medium supplemented with 50 µg ml−1 of zeocin. The screening of putative disrupted clones was performed by PCR on genomic DNA purified from transformant plants. Four independent PCRs were performed to confirm that both the 5′ and 3′ insertion sites were sequentially correct. For both sites, one primer corresponded to a chromosomal fragment outside the knockout construction and was used in both PCRs; of the other two primers, one was specific for the marker and the other for the wild-type gene (Supplementary Fig. S6). Clones in which the sequences of the knockout insertions were correct and the fragments of the wild-type gene could not be amplified were selected. Four ΔPpchx1-(1–4) and seven ΔPpchx2-(1–7) lines were isolated and studied; however, repeated experiments were performed only with ΔPpchx1-1 and ΔPpchx2-1 plants. Cation contents and cation uptake experiments To determine the K+ content, samples of P. patens were transferred to a filter, washed with 10 mM MgCl2, weighed, and extracted with 0.1 M HCl (Garciadeblás et al. 2007). The K+ concentrations in the supernatant were determined by atomic emission spectrophotometry. For Rb+/K+ exchange experiments, protonema cells growing in BCDAT (12 mM K+) were transferred to K+-free (20 µM K+) KFM medium containing 10 mM Rb+ and daily samples were taken from day 1 to day 7. Plant samples were washed, extracted with 0.1 M HCl and analyzed to determine the K+ and Rb+ concentrations. In all these experiments, each data point was the mean of three independent measurements, in which the deviations were extremely low. The reported statistical analysis was performed with data from experiments that were absolutely independent, including the starting plant inocula. Cation uptake tests at micromolar K+ or Rb+ concentrations were carried out by following the decrease of the cations in the external medium. The experiments were started by the addition of the selected cation; at intervals, samples of medium were taken and the K+ and Rb+ concentrations in the cell-free medium were determined by atomic emission spectrophotometry. Protein alignments and phylogenetic tree constructions Protein alignments and phylogenetic trees were obtained by using the Clustal X program (Thompson, 1997). Funding This work was supported by the Spanish Ministerio de Ciencia e Innovación; the ERDF European program [grant No. AGL2007-61075, and a fellowship to S.A.M.]; the DGUI-UPM Research Group Program. Acknowledgments We would like to thank Blanca Garciadeblás and Rocío Álvarez-Aragón for their help in some experiments, Pablo González-Melendi for his help with the confocal microscope, Mitsuyasu Hasebe and Tomoaki Nishiyama for their gift of plasmids pTN186 and p35S-Zeo, Hana Sychrova for her gift of strains LMB01 and LMM04, and the Arabidopsis Biological Resource Center for providing the fluorescent markers. Disclosures The authors have no conflicts of interest to declare. Abbreviations Abbreviations ACT actin EST expressed sequence tag GFP green fluorescent protein RT–PCR reverse transcription–PCR TAPS N-Tris(hydroxymethyl)methyl-3-aminopropanesulfonic acid UTR untranslated region YFP yellow fluorescent protein References Ahn SJ,  Shin R,  Scachtman DP.  Expression of KT/KUP genes in Arabidopsis and the role of root hairs K+ uptake,  Plant Physiol. ,  2004, vol.  134 (pg.  1135- 1145) Google Scholar CrossRef Search ADS PubMed  Ashton NW,  Cove DJ,  Featherstone DR.  The isolation and physiological analysis of mutants of the moss Physcomitrella patens,  Planta ,  1979, vol.  144 (pg.  437- 442) Google Scholar CrossRef Search ADS PubMed  Bakker EP,  Mangerich WE.  Interconversion of components of the bacterial proton motive force by electrogenic potassium transport,  J. Bacteriol. ,  1981, vol.  147 (pg.  820- 826) Google Scholar PubMed  Bañuelos MA,  Synchrová H,  Bleykasten-Grosshans C,  Souciet J-L,  Potier S.  The Nha1 antiporter of Saccharomyces cerevisiae mediates sodium and potassium efflux,  Microbiology ,  1998, vol.  144 (pg.  2749- 2758) Google Scholar CrossRef Search ADS PubMed  Barragán V,  Leidi EO,  Andrés Z,  Rubio L,  Luca AD,  Fernández JA, et al.  Ion exchangers NHX1 and NHX2 mediate active potassium uptake into vacuoles to regulate cell turgor and stomatal function in Arabidopsis,  Plant Cell ,  2012, vol.  24 (pg.  1127- 1142) Google Scholar CrossRef Search ADS PubMed  Bassil E,  Ohto M,  Esumi T,  Tajima H,  Zhu Z,  Cagnac O, et al.  The Arabidopsis intracellular Na+/H+ antiporters NHX5 and NHX6 are endosome associated and necessary for plant growth and development,  Plant Cell ,  2011, vol.  23 (pg.  224- 239) Google Scholar CrossRef Search ADS PubMed  Brett CL,  Donowitz M,  Rao R.  Evolutionary origins of eukaryotic sodium/proton exchangers,  Amer. J. Physiol. Cell Physiol. ,  2005, vol.  288 (pg.  C223- C239) Google Scholar CrossRef Search ADS   Brunelli JP,  Pall ML.  A series of yeast/Escherichia coli λ expression vectors designed for directional cloning of cDNAs and cre/lox-mediated plasmid excision,  Yeast ,  1993, vol.  9 (pg.  1309- 1318) Google Scholar CrossRef Search ADS PubMed  Casey JR,  Grinstein S,  Orlowski J.  Sensors and regulators of intracellular pH,  Nat. Rev. Mol. Cell Biol. ,  2010, vol.  11 (pg.  50- 61) Google Scholar CrossRef Search ADS PubMed  Cove D.  The moss Physcomitrella patens,  Annu. Rev. Genet. ,  2005, vol.  39 (pg.  339- 358) Google Scholar CrossRef Search ADS PubMed  Cove D,  Bezanilla M,  Harries P,  Quatrano R.  Mosses as model systems for the study of metabolism and development,  Annu. Rev. Plant Biol. ,  2006, vol.  57 (pg.  497- 520) Google Scholar CrossRef Search ADS PubMed  Chanroj S,  Lu Y,  Padmanaban S,  Nanatani K,  Uozumi N,  Rao R, et al.  Plant-specific cation/H+ exchanger 17 and its homologs are endomembrane K+ transporters with roles in protein sorting,  J. Biol. Chem. ,  2011, vol.  286 (pg.  33931- 33941) Google Scholar CrossRef Search ADS PubMed  Chanroj S,  Wang G,  Venema K,  Zhang MW,  Delwiche CF,  Zse H.  Conserved and diversified gene families of monovalent cation/H+ antiporters from algae to flowering plants,  Front. Plant Sci. ,  2012, vol.  3 pg.  25  Google Scholar CrossRef Search ADS PubMed  Demaurex N.  pH homeostasis of cellular organelles,  News Physiol. Sci. ,  2002, vol.  17 (pg.  1- 5) Google Scholar PubMed  Dreyer I,  Horeau C,  Lemaillet G,  Zimmermann S,  Bush DR,  Rodríguez-Navarro A, et al.  Identification and characterization of plant transporters using heterologous expression systems,  J. Exp. Bot. ,  1999, vol.  50 (pg.  1073- 1078) Evans AR,  Hall D,  Pritchard J,  Newbury HJ.  The roles of the cation transporters CHX21 and CHX23 in the development of Arabidopsis thaliana,  J. Exp. Bot. ,  2012, vol.  63 (pg.  59- 67) Google Scholar CrossRef Search ADS PubMed  Felle H,  Porter JS,  Slayman CL,  Kaback HR.  Quantitative measurements of membrane potential in Escherichia coli,  Biochemistry ,  1980, vol.  19 (pg.  3585- 3590) Google Scholar CrossRef Search ADS PubMed  Fraile-Escanciano A,  Garciadeblás B,  Rodríguez-Navarro A,  Benito B.  Role of ENA ATPase in Na+ efflux at high pH in bryophytes,  Plant Mol. Biol. ,  2009, vol.  71 (pg.  599- 608) Google Scholar CrossRef Search ADS PubMed  Fraile-Escanciano A,  Kamisugi Y,  Cuming AC,  Rodríguez-Navarro A,  Benito B.  The SOS1 transporter of Physcomitrella patens mediates sodium efflux in planta,  New Phytol. ,  2010, vol.  188 (pg.  750- 761) Google Scholar CrossRef Search ADS PubMed  Garciadeblás B,  Barrero-Gil J,  Benito B,  Rodríguez-Navarro A.  Potassium transport systems in the moss Physcomitrella patens: pphak1 plants reveal the complexity of potassium uptake,  Plant J. ,  2007, vol.  52 (pg.  1080- 1093) Google Scholar CrossRef Search ADS PubMed  Garciadeblás B,  Senn ME,  Bañuelos MA,  Rodríguez-Navarro A.  Sodium transport and HKT transporters: the rice model,  Plant J. ,  2003, vol.  34 (pg.  788- 801) Google Scholar CrossRef Search ADS PubMed  Gómez-Porras JL,  Riaño-Pachón DM,  Benito B,  Haro R,  Sklodowski K,  Rodríguez-Navarro A, et al.  Phylogenetic analysis of K+ transporters in bryophytes, lycophytes, and flowering plants indicates a specialization of vascular plants,  Front. Plant Sci. ,  2012, vol.  3 pg.  167  Google Scholar CrossRef Search ADS PubMed  Grabov A.  Plant KT/KUP/HAK potassium transporters: single family–multiple functions,  Ann. Bot. ,  2007, vol.  99 (pg.  1035- 1041) Google Scholar CrossRef Search ADS PubMed  Guzman L-M,  Belin D,  Carson MJ,  Beckwith J.  Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter,  J. Bacteriol. ,  1995, vol.  177 (pg.  4121- 4130) Google Scholar CrossRef Search ADS PubMed  Hall D,  Evans AR,  Newbury HJ,  Pritchard J.  Functional analysis of CHX21: a putative sodium transporter in Arabidopsis,  J. Exp. Bot. ,  2006, vol.  57 (pg.  1201- 1210) Google Scholar CrossRef Search ADS PubMed  Haro R,  Bañuelos MA,  Rodríguez-Navarro A.  High-affinity sodium uptake in land plants,  Plant Cell Physiol. ,  2010, vol.  51 (pg.  68- 79) Google Scholar CrossRef Search ADS PubMed  Haro R,  Fraile-Escanciano A,  González-Melendi P,  Rodríguez-Navarro A.  The potassium transporters HAK2 and HAK3 localize to endomembranes in Physcomitrella patens. HAK2 is required in some stress conditions,  Plant Cell Physiol. ,  2013, vol.  54 (pg.  1441- 1454) Google Scholar CrossRef Search ADS PubMed  Haro R,  Rodríguez-Navarro A.  Functional analysis of the M2D helix of the TRK1 potassium transporter of Saccharomyces cerevisiae,  Biochim. Biophys. Acta ,  2003, vol.  1613 (pg.  1- 6) Google Scholar CrossRef Search ADS PubMed  Hohe A,  Egener T,  Lucht JM,  Holtof H,  Reinhard C,  Schween G, et al.  An improved and highly standarised transformation procedure allows efficient production of single and multiple targeted gene-knockouts in a moss, Physcomitrella patens,  Curr. Genet. ,  2004, vol.  44 (pg.  339- 347) Google Scholar CrossRef Search ADS PubMed  Kashket ER.  Stiochiometry of the H+-ATPase of growing and resting aerobic Escherichia coli,  Biochemistry ,  1982, vol.  21 (pg.  5534- 5538) Google Scholar CrossRef Search ADS PubMed  Lu Y,  Chanroj S,  Zulkifli L,  Johson MA,  Uozumi N,  Cheung A, et al.  Pollen tubes lacking a pair of K+ transporters fail to target ovules in Arabidopsis,  Plant Cell ,  2011, vol.  23 (pg.  81- 93) Google Scholar CrossRef Search ADS PubMed  Maresova L,  Sychrova H.  Physiological characterization of Saccharomyces cerevisiae kha1 deletion mutants,  Mol. Microbiol. ,  2005, vol.  55 (pg.  588- 600) Google Scholar CrossRef Search ADS PubMed  Maresova L,  Sychrova H.  Arabidopsis thaliana CHX17 gene complements the kha1 deletion phenotypes in Saccharomyces cerevisisae,  Yeast ,  2006, vol.  23 (pg.  1167- 1171) Google Scholar CrossRef Search ADS PubMed  Nelson BK,  Cai X,  Nebenführ A.  A multicolored set of in vivo organelle markers for co-localization studies in Arabidopsis and other plants,  Plant J. ,  2007, vol.  51 (pg.  1126- 1136) Google Scholar CrossRef Search ADS PubMed  Ohgaki R,  Ijzendoorn SCV,  Matsushita M,  Hoekstra D,  Kanazawa H.  Organellar Na+/H+ exchangers: novel players in organelle pH regulation and their emerging functions,  Biochemistry ,  2011, vol.  50 (pg.  443- 450) Google Scholar CrossRef Search ADS PubMed  Orlowski J,  Gristein S.  Emerging roles of alkaline cation/proton exchangers in organellar homeostasis,  Curr. Opin. Cell Biol. ,  2007, vol.  19 (pg.  483- 492) Google Scholar CrossRef Search ADS PubMed  Padan E,  Zilberstein D,  Rottenberg H.  The proton electrochemical gradient in Escherichia coli cells,  Eur. J. Biochem. ,  1976, vol.  63 (pg.  533- 541) Google Scholar CrossRef Search ADS PubMed  Padmanaban S,  Chanroj S,  Kwak JM,  Li X,  Ward JM,  Sze H.  Participation of endomembrane cation/H+ exchanger AtCHX20 in osmoregulation of guard cells,  Plant Physiol. ,  2007, vol.  144 (pg.  82- 93) Google Scholar CrossRef Search ADS PubMed  Pardo JM,  Cubero B,  Leidi EO,  Quintero FJ.  Alkali cation exchangers: roles in cellular homeostasis and stress tolerance,  J. Exp. Bot. ,  2006, vol.  57 (pg.  1181- 1199) Google Scholar CrossRef Search ADS PubMed  Paroutis P,  Touret N,  Grinstein S.  The pH of the secretory pathway: measurement, determinants, and regulation,  Physiology ,  2004, vol.  19 (pg.  207- 215) Google Scholar CrossRef Search ADS PubMed  Quatrano RS,  McDaniel SF,  Khandelwal A,  Perround P,  Cove DJ.  Physcomitrella patens: mosses enter the genomic age,  Curr. Opin. Plant Biol. ,  2007, vol.  10 (pg.  182- 189) Google Scholar CrossRef Search ADS PubMed  Rodríguez-Navarro A,  Blatt MR,  Slayman CL.  A potassium–proton symport in Neurospora crassa,  J. Gen. Physiol. ,  1986, vol.  87 (pg.  649- 674) Google Scholar CrossRef Search ADS PubMed  Rodríguez-Navarro A,  Ramos J.  Dual system for potassium transport in Saccharomyces cerevisiae,  J. Bacteriol. ,  1984, vol.  159 (pg.  940- 945) Google Scholar PubMed  Rubio-Somoza I,  Marinez M,  Diaz I,  Carbonero P.  HvMCD1, a R1MYB transcription factor from barley with antagonistic regulatory functions during development and germination,  Plant J. ,  2006, vol.  45 (pg.  17- 30) Google Scholar CrossRef Search ADS PubMed  Rubio F,  Santa-María GE,  Rodríguez-Navarro A.  Cloning of Arabidopsis and barley cDNAs encoding HAK potasium transporters in root and shoot cells,  Physiol. Plant. ,  2000, vol.  109 (pg.  34- 43) Google Scholar CrossRef Search ADS   Saint-Jore-Dupas C,  Nebenführ A,  Boulaflous A,  Follet-Gueye ML,  Plasson C,  Hawes C, et al.  Plant N-glycan processing enzymes employ different targeting for their spatial arrangement along the secretory pathway,  Plant Cell ,  2006, vol.  18 (pg.  3182- 3200) Google Scholar CrossRef Search ADS PubMed  Saito C,  Ueda T,  Abe H,  Wada Y,  Kuroiwa T,  Hisada A, et al.  A complex and mobile structure forms a distinct subregion within the continous vacuolar membrane in young cotyledons of Arabidopsis,  Plant J. ,  2002, vol.  29 (pg.  245- 255) Google Scholar CrossRef Search ADS PubMed  Schleyer M,  Bakker EP.  Nucleotide sequence and 3′-end deletion studies indicate that the K+-uptake protein Kup from Escherichia coli is composed of a hydrophobic core linked to a large and partially essential hydrophilic C terminus,  J. Bacteriol. ,  1993, vol.  175 (pg.  6925- 6931) Google Scholar CrossRef Search ADS PubMed  Schultz SG,  Solomon AK.  Cation transport in Escherichia coli: I. Intracellular Na+ and K+ concentrations and net cation movement,  J. Gen. Physiol. ,  1961, vol.  45 (pg.  355- 369) Google Scholar CrossRef Search ADS PubMed  Segami S,  Nakanishi Y,  Sato MH,  Maeshima M.  Quantification, organ-specific accumulation and intracellular localization of type II H+-pyrophosphatase in Arabidopsis thalina,  Plant Cell Physiol. ,  2010, vol.  51 (pg.  1350- 1360) Google Scholar CrossRef Search ADS PubMed  Senn ME,  Rubio F,  Bañuelos MA,  Rodríguez-Navarro A.  Comparative functional features of plant potassium HvHAK1 and HvHAK2 transporters,  J. Biol. Chem. ,  2001, vol.  276 (pg.  44563- 44569) Google Scholar CrossRef Search ADS PubMed  Setty OH,  Hendler RW,  Shrager RI.  Simultaneous measurements of proton motive force, delta pH, membrane potential, and H+/O ratios in intact Escherichia coli,  Biophys. J. ,  1983, vol.  43 (pg.  371- 381) Google Scholar CrossRef Search ADS PubMed  Sherman F.  Getting started with yeast,  Methods Enzymol. ,  1991, vol.  194 (pg.  3- 21) Google Scholar PubMed  Sze H,  Padmanaban S,  Cellier F,  Honys D,  Cheng N,  Bock KW, et al.  Expression patterns of a novel AtCHX gene family highlight potential roles in osmotic adjustment and K+ homeostasis in pollen development,  Plant Physiol. ,  2004, vol.  136 (pg.  2532- 2547) Google Scholar CrossRef Search ADS PubMed  Tempest DW,  Strange RE.  Variation in content and distribution of magnesium, and its influence on survival, in Aerobacter aerogenes grown in a chemostat,  J. Gen. Microbiol. ,  1966, vol.  44 (pg.  273- 279) Google Scholar CrossRef Search ADS PubMed  Thompson JD.  The Clustal_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools,  Nucleic Acids Res. ,  1997, vol.  25 (pg.  4876- 4882) Google Scholar CrossRef Search ADS PubMed  Weiss A,  Rodríguez-Franco M,  Timm B,  Hermann M,  Link S,  Jost W, et al.  Use of Physcomitrella patens actin 5′ regions for high transgene expression: importance of 5′ introns,  Appl. Microbiol. Biotechnol. ,  2006, vol.  70 (pg.  337- 345) Google Scholar CrossRef Search ADS PubMed  Yang Z,  Gao Q,  Sun C,  Li W,  Gu S,  Xu C.  Molecular evolution and functional divergence of HAK potassium transporter gene family in rice (Oryza sativa L.),  J. Genet. Genomics ,  2009, vol.  36 (pg.  161- 172) Google Scholar CrossRef Search ADS PubMed  Zhao J,  Cheng N-H,  Motes CM,  Blancaflor EB,  Moore M,  Gonzalez N, et al.  AtCHX13 is a plasma membrane K+ transporter,  Plant Physiol. ,  2008, vol.  148 (pg.  796- 807) Google Scholar CrossRef Search ADS PubMed  Zilberstein D,  Schuldiner S,  Padan E.  Proton electrochemical gradient in Escherichia coli cells its relation to active transport of lactose,  Biochemistry ,  1979, vol.  18 (pg.  669- 673) Google Scholar CrossRef Search ADS PubMed  © The Author 2013. Published by Oxford University Press on behalf of Japanese Society of Plant Physiologists. All rights reserved. For permissions, please email: journals.permissions@oup.com
TI - Knockouts of Physcomitrella patens CHX1 and CHX2 Transporters Reveal High Complexity of Potassium Homeostasis
JF - Plant and Cell Physiology
DO - 10.1093/pcp/pct096
DA - 2013-07-27
UR - https://www.deepdyve.com/lp/oxford-university-press/knockouts-of-physcomitrella-patens-chx1-and-chx2-transporters-reveal-UJGnA1G37b
SP - 1455
EP - 1468
VL - 54
IS - 9
DP - DeepDyve
ER -