A Cation-regulated and Proton Gradient-dependent Cation Transporter from Chlamydomonas reinhardtii Has a Role in Calcium and Sodium Homeostasis*

The CrCAX1 gene encoding a Ca2+/H+ and Na+/H+ exchanger was cloned and characterized from the unicellular green alga Chlamydomonas reinhardtii to begin to understand the mechanisms of cation homeostasis in this model organism. CrCAX1 was more closely related to fungal cation exchanger (CAX) genes than those from higher plants but has structural characteristics similar to plant Ca2+/H+ exchangers including a long N-terminal tail. When CrCAX1-GFP was expressed in Saccharomyces cerevisiae, it localized at the vacuole. CrCAX1 could suppress the Ca2+-hypersensitive phenotype of a yeast mutant and mediated proton gradient-dependent Ca2+/H+ exchange activity in vacuolar membrane vesicles. Ca2+ transport activity was increased following N-terminal truncation of CrCAX1, suggesting the existence of an N-terminal auto-regulatory mechanism. CrCAX1 could also provide tolerance to Na+ stress when expressed in yeast or Arabidopsis thaliana because of Na+/H+ exchange activity. This Na+/H+ exchange activity was not regulated by the N terminus of the CrCAX1 protein. A subtle tolerance by CrCAX1 in yeast to Co2+ stress was also observed. CrCAX1 was transcriptionally regulated in Chlamydomonas cells grown in elevated Ca2+ or Na+. This study has thus uncovered a novel eukaryotic proton-coupled transporter, CrCAX1, that can transport both monovalent and divalent cations and that appears to play a role in cellular cation homeostasis by the transport of Ca2+ and Na+ into the vacuole.

example, tightly controlled levels of Ca 2ϩ play a critical role in cell signaling, but high concentrations of Ca 2ϩ are very toxic to the cell (1). Likewise a careful balance between K ϩ and Na ϩ ions is required to prevent osmotic stress and the toxic effects of Na ϩ (2,3). In unicellular organisms particularly, failure of Ca 2ϩ or Na ϩ homeostasis will lead to death of that organism; therefore the tight control of these ions is paramount. This is certainly the case for many of these organisms that exist in potentially harsh ionic conditions. For example, many of the unicellular green algae of the Chlamydomonas genus can tolerate and adapt to multiple ion stresses. The acidophile Chlamydomonas acidophila and the halotolerant Chlamydomonas sp. W80 can tolerate very severe ion stresses (4,5). The soil and freshwater living Chlamydomonas reinhardtii (herein referred to as Chlamydomonas), although not as tolerant to these stresses as the extremophile species, have adaptive mechanisms to many ion stress conditions (6), either by direct removal of an ion through efflux or by mediating Ca 2ϩ -signaling processes.
Some of the mechanisms of Na ϩ and Ca 2ϩ homeostasis and transport are well understood in model eukaryotes such as yeast (Saccharomyces cerevisiae) and the plant Arabidopsis thaliana. In both species, ⌬pH-dependent cation/H ϩ exchangers are important in mediating the removal of excess Na ϩ and Ca 2ϩ from the cytosol across the plasma membrane or the vacuolar membrane (7). Na ϩ , K ϩ , or Ca 2ϩ transporting cation/H ϩ exchangers derive from at least four phylogenetic superfamilies. One of these is the CaCA (Ca 2ϩ /cation antiporter) superfamily, which includes mammalian Na ϩ /Ca 2ϩ exchanger genes and cation/H ϩ exchanger (CAX) 2 genes (8,9). CAX genes have been identified in bacteria, fungi, protozoa, plants, and lower vertebrates but are absent from higher animals (9). The majority of the CAX genes characterized to date encode H ϩ -coupled exchangers that transport Ca 2ϩ . Many are highly specific for Ca 2ϩ such as the yeast and Arabidopsis vacuolar Ca 2ϩ /H ϩ exchangers ScVCX1 and AtCAX1 (10,11), whereas others have a broader specificity and can transport a range of divalent cations (12). An interesting exception is the yeast CAX family transporter ScVNX1, which can transport Na ϩ and K ϩ but appears not to transport Ca 2ϩ (13).
A disparate feature of these transporters is the mechanism of regulation. Ca 2ϩ transport by ScVCX1 is negatively regulated by the protein phosphatase calcineurin (10), whereas a post-

* This work was supported by Biotechnology and Biological Sciences
Research Council Grants BB/B502152/1 (to J. K. P.) and BB/C502030/1 (to C. M. B.) and Leverhulme Trust Grant F/00 120/BG (to J. K. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental Fig. S1 and Tables S1-S3. The nucleotide sequence(s) reported in this paper has been submitted to the Gen-Bank TM /EBI Data Bank with accession number(s) FM253128. 1 (14). Further analysis in yeast and plants confirmed that the N-terminal tail regulates AtCAX1 Ca 2ϩ transport activity through an autoinhibitory mechanism (15)(16)(17). In addition, many of the plant CAX transporters are transcriptionally regulated by alterations in ion levels (18 -20). Chlamydomonas is a model organism for studying photosynthesis and other aspects of cell biology and should be an excellent model to study adaptive responses to stress and the evolutionary relationships of cation transporters among eukaryotes. The recently sequenced Chlamydomonas genome has indicated the wide range of transporters present including members of the CaCA superfamily (21,22), the majority of which remain uncharacterized. Here we describe the first functional analysis of a cation/H ϩ exchanger from Chlamydomonas and demonstrate that it functions as a H ϩ -coupled Ca 2ϩ and Na ϩ transporter that provides a role in cation homeostasis.

EXPERIMENTAL PROCEDURES
DNA Manipulations-CrCAX1 and N-terminally truncated CrCAX1 (sCrCAX1) cDNA was amplified by PCR from cDNA template derived from isolated Chlamydomonas RNA using primers CrCAX1F, CrCAX1R, and sCrCAX1F (see supplemental Table S1 for primer sequences). PCR products were cloned into pGEM-T vector for propagation and sequencing. CrCAX1 and sCrCAX1 cDNAs were subcloned into the XbaI and SacI sites of the yeast expression vector piHGpd and into a p35S-CAMBIA2300 plant kanamycin-resistant expression vector that had been modified to contain the CaMV 35S promoter sequence and nos terminator sequence. To identify the position of translation initiation, CrCAX1 cDNA constructs were generated that encode an N-terminal fragment. A CrCAX1 cDNA that encodes the first predicted 180 amino acids was amplified using primers CrCAX1TnTF and CrCAX1TnTR. A mutant variant was generated using the forward primer CrCAX1M1TnTF where the predicted Met 1 was mutated to Ala so that translation would be predicted to initiate at Met 91 . The PCR products were subcloned into the XbaI and BamHI sites of pSP64 (Promega) for in vitro transcription-translation analysis. For comparison with AtCAX1, an AtCAX1 cDNA that encodes the first 67 amino acids was amplified using primers CAX1TnTF and CAX1aa67R. This PCR product was fused to a previously amplified rsGFP sequence containing a 5Ј BglII site and 3Ј SacI site (23). The AtCAX1-67-GFP cDNA was subcloned into the PstI and SacI sites of pSP64. A mutant construct in which Met 37 was substituted with Ala was generated using AtCAX1-67-GFP as template using a QuikChange sitedirected mutagenesis kit (Stratagene) using primers CAX1M37AF and CAX1M37AR. To generate C-terminal green fluorescent protein (GFP) fusions to CrCAX1 and sCrCAX1, a HindIII site was introduced into the 3Ј end of the cDNA using the primer CrCAX1HR in combination with CrCAX1F or sCrCAX1F. A GFP cDNA was amplified from a synthetic GFP (S65G, S72A) plasmid (24) and was cloned into the HindIII and SacI sites of CrCAX1-and sCrCAX1-pGEM-T. The CrCAX1-GFP and sCrCAX1-GFP constructs were then subcloned into piHGpd and p35S-CAMBIA2300.
Cell and Plant Growth-C. reinhardtii wild type strain 137Cϩ (CC-125 or CCAP 11/32C) was obtained from the UK Culture Collection of Algae and Protozoa (CCAP). The cells were grown in Tris-acetate-phosphate (TAP) liquid medium with agitation and on plates at 25°C in 100 mol of photons/m 2 ϫ s Ϫ1 light provided by fluorescent tubes. Cell growth was determined by measuring at A 750 nm . The yeast (S. cerevisiae) strains K667 (cnb1::LEU2 pmc1::TRP1 vcx1⌬) (10), AXT3 (nhx1::TRP1 ena1-4::HIS3 nha1::LEU2) (25), and nhx1⌬::kanMX4 (Euroscarf) were used for heterologous expression. Transformed yeast strains were grown in synthetic defined medium minus appropriate amino acids for selective growth for the expression plasmid and the mutations. For metal tolerance assays, serial dilutions of K667 yeast were grown at 30°C on solid yeast extract-peptone-dextrose medium containing a range of metal salts including CaCl 2 , NaCl, CoCl 2 , and CdCl 2 for 4 days. For determination of AXT3 yeast growth rate in liquid NaCl solutions, yeast strains of the same starting cell density were inoculated in arginine-phosphate-dextrose medium containing 1 mM KCl and a range of NaCl concentrations and grown at 30°C, shaking for 48 h in 24-well flat bottomed plates, and cell growth was determined by measuring at A 600 nm . A. thaliana accession Col-0 was grown on soil or on solid 0.5ϫ strength Murashige and Skoog medium at 22°C in a 16 h/8 h light/dark cycle in 100 mol of photons/m 2 ϫ s Ϫ1 light provided by fluorescent tubes. Total chlorophyll (chlorophyll aϩb) measurements were determined as described (26).
DNA Transformations-Yeast strains were transformed with piHGpd plasmid constructs using the lithium acetate/polyethylene glycol method as described previously (27). Arabidopsis was transformed with p35S-CAMBIA2300 plasmid constructs using the standard Agrobacterium tumefaciens-mediated floral dipping method. Following screening for kanamycin-resistant seeds, at least four homozygous lines were selected for each independent transformation with CrCAX1 and sCrCAX1.
RNA Extraction and RT-PCR-RNA was isolated from Arabidopsis using a RNA isolation kit (Qiagen). RNA was isolated from yeast by acid phenol extraction. RNA was isolated from Chlamydomonas cells grown in liquid TAP medium supplemented with various metal salts. The cells were collected by centrifugation, resuspended in water, and lysed in a lysis solution containing 50 mM Tris/HCl, pH 7.5, 150 mM NaCl, 15 mM EDTA, 2% SDS, and 40 pg/ml proteinase K for 20 min at room temperature. RNA was purified from the lysed cells by four extractions with equal volumes of phenol/chloroform/isoamyl alcohol followed by two extractions of the aqueous phase with an equal volume of chloroform. RNA was precipitated in 2 volumes of 100% isopropanol, and the pellet was collected by centrifugation. First strand cDNA was produced from 1 g of DNase-treated total RNA using Superscript II reverse transcriptase (Invitrogen) and an oligo(dT) primer. To determine CrCAX1 expression in transgenic Arabidopsis lines or yeast cells, RT-PCR was performed using sCrCAX1F and CrCAX1TnTR primers and actin or tubulin primers as a constitutive control. To determine relative CrCAX1 expression in metal-treated Chlamydomonas cells, RT-PCR was performed using CrCAX1TnTF and CrCAX1TnTR primers and 18 S rRNA primers as a constitutive control. PCR products were amplified with a 60°C annealing temperature for 25 cycles and detected with ethidium bromide.
Vacuolar Membrane Vesicle Isolation and Transport Measurements-Vacuolar-enriched membrane vesicles were isolated from K667 or nhx1⌬ yeast cells expressing sCrCAX1 and CrCAX1, as described previously (27). Ca 2ϩ /H ϩ exchange activity was determined by measuring pH gradient-dependent 45 Ca 2ϩ uptake into K667-derived vesicles, as described previously (28). Na ϩ /H ϩ exchange activity was determined in nhx1⌬-derived vesicles by measuring Na ϩ -dependent acridine orange fluorescence quench, essentially as described in Ref. 29 using a Jasco FP750 fluorescence spectrometer.

RESULTS
Identification of a CAX Gene from C. reinhardtii-Analysis of sequence from Chlamydomonas expressed sequence tag cDNA clones and the completed genome sequence has identified the presence of open reading frames homologous to Arabidopsis and S. cerevisiae CAX genes. One of these annotated open reading frames designated CrCAX1 (JGI I.D. 157233) was analyzed further by comparison with known CAX genes. Following DNA and amino acid sequence alignment, the accuracy of the sequence prediction was ascertained. PCR primers were designed to amplify the predicted CrCAX1 cDNA from C. reinhardtii 137Cϩ (wild type) RNA isolated from cells grown in liquid TAP medium. CrCAX1 (accession number FM253128) has a 1344-bp open reading frame and encodes a protein of 447 amino acids. Sequence comparisons of CrCAX1 found highest sequence identity with fungal and protozoan CAX genes (40% sequence identity with PbCAX1 from Plasmodium berghei, 35% identity with ScVCX1, and 30% identity with AtCAX1; supplemental Table S2). However, sequence alignment and transmembrane span prediction comparison of CrCAX1 with CAX genes from fungi and higher plants indicated that CrCAX1 encodes a protein with similar topology to higher plant CAX transporters including a long hydrophilic tail predicted to be cytosolic and 11 predicted transmembrane spans (supplemental Fig. S1). The CrCAX1 hydrophilic N-terminal tail sequence is longer (94 amino acids) than that of ScVCX1 (32 amino acids) and AtCAX1 (67 amino acids).
Detailed phylogenetic analysis was recently performed using 138 CAX gene sequences that were previously cloned or identified from genome sequences of bacteria, fungi, plants, protozoa, and lower vertebrates (9). This analysis indicated that CAX genes could be separated into three main phylogenetic groups (Type I, Type II, and Type III). Arabidopsis and S. cerevisiae CAX genes formed part of the Type I group, which was further divided into subtypes (Type I-A to Type I-H). Phylogenetic analysis performed with the CrCAX1 sequence and using the classification of Shigaki et al. (9) indicates that CrCAX1 falls within clade Type I-C, which includes CAX genes from protozoa, including Plasmodium falciparum and Cryptosporidium hominis, the red algae Thalassiosira pseudonana (Fig. 1), and the diatom Phaeodactylum tricornutum (data not shown).
In Vitro Translation Analysis of CrCAX1-To confirm that translation of CrCAX1 did initiate from the predicted first AUG start codon (Met 1 ) rather than the downstream AUG (Met 91 ), an in vitro cell-free transcription and translation experiment was performed. For comparison translation analysis was performed using the previously characterized AtCAX1.
To improve translation efficiency, partial length constructs were used that encode the first 180 amino acids of CrCAX1 and the first 67 amino acids of AtCAX1. A single 35 S-labeled protein band of the expected size was identified for CrCAX1, confirm-  (9) ing that translation occurred solely from Met 1 ( Fig. 2A), unlike for AtCAX1 where some translation (16% of total AtCAX1 protein) initiated from the second AUG (Met 37 ) (Fig. 2B). Analysis of an AtCAX1-M37A mutant confirmed that the lower protein band was due to initiation from Met 37 . When CrCAX1 Met 1 was mutated to Ala, translation was able to initiate from Met 91 , and a smaller protein product was observed ( Fig. 2A).
Functional Analysis of CrCAX1 in S. cerevisiae-To assess the transport function of CrCAX1, a yeast heterologous expression approach was used. This has been extremely successful previously to ascertain the function of higher plant cation/H ϩ exchangers (11,14,31). The K667 yeast mutant lacks vacuolar Ca 2ϩ /H ϩ exchange and Ca 2ϩ -ATPase activity and is therefore unable to sufficiently sequester Ca 2ϩ into the vacuole in response to Ca 2ϩ stress and cannot grow on high Ca 2ϩ containing media (10). Many plant CAX transporters including AtCAX1 possess an N-terminal regulatory domain that inhibits Ca 2ϩ transport activity when expressed in a heterologous system (14); therefore for comparison an N-terminal truncation mutant of CrCAX1 (termed sCrCAX1) was generated in which translation was initiated from the first downstream AUG, which encodes Met 91 immediately prior to the first predicted transmembrane span (supplemental Fig. S1). C-terminally tagged CrCAX1-GFP and sCrCAX1-GFP constructs were also generated and expressed in yeast. Both CrCAX1-GFP and sCrCAX1-GFP localized equivalently to the vacuole (Fig. 3). As predicted transmembrane proteins, CrCAX1 and sCrCAX1 are likely to be located at the vacuolar membrane, although imaging by epifluorescence microscopy was unable to unequivocally confirm this.
Comparison of CrCAX1 and sCrCAX1 in yeast grown on yeast extract-peptone-dextrose medium supplemented with a high concentration (200 mM) of CaCl 2 found that sCrCAX1 could efficiently suppress the Ca 2ϩ hypersensitivity of K667 yeast in a manner equivalent to N-terminally truncated AtCAX1 (sAtCAX1) (Fig. 4A). Full-length CrCAX1 could only weakly suppress the Ca 2ϩ -sensitive phenotype. Transformed K667 yeast growth was evaluated on other metal conditions. CrCAX1 or sCrCAX1 could not suppress sensitivity to Mn 2ϩ or Zn 2ϩ but could provide some Cd 2ϩ tolerance to yeast (data not shown). Growth of yeast on 0.75 M NaCl was significantly enhanced following expression of either CrCAX1 or sCrCAX1, but there was no difference between full-length and truncated CrCAX1 in the ability to provide Na ϩ tolerance (Fig. 4A). Likewise, slight tolerance to 2.5 mM CoCl 2 was observed by expression of either CrCAX1 or sCrCAX1. sAtCAX1 was unable to provide tolerance to either Na ϩ or Co 2ϩ stress, indicating that tolerance was not linked to altered Ca 2ϩ homeostasis. The Ca 2ϩ and Na ϩ tolerance by CrCAX1 and sCrCAX1 was clearly due to their expression in the yeast (Fig. 4B). To confirm the Na ϩ tolerance phenotype of CrCAX1, plasmids were transformed into AXT3 yeast, which lacks plasma membrane and endomembrane Na ϩ transporters and is hypersensitive to Na ϩ (25). Transformed AXT3 yeast strains were grown in liquid arginine-phosphate-dextrose medium supplemented with various concentrations of NaCl, and cell growth was measured. Both sCrCAX1 and CrCAX1 could provide AXT3 yeast with significant tolerance to high concentrations of NaCl, whereas AXT3 expressing empty vector or sAtCAX1 was unable to grow on this salt medium (Fig. 4C).
Ca 2ϩ /H ϩ Exchange Activity of CrCAX1-To confirm that growth of yeast on Ca 2ϩ -containing medium was due to enhanced vacuolar Ca 2ϩ /H ϩ exchange activity, ⌬pH-dependent 45 Ca 2ϩ uptake in the presence of the Ca 2ϩ -ATPase inhibitor vanadate was measured in tonoplast-enriched membrane vesicles isolated from K667 yeast expressing sCrCAX1. Ca 2ϩ /H ϩ exchange activity was determined for sCrCAX1, and this 45 Ca 2ϩ uptake was significantly inhibited in the presence of 5 M of protonophore FCCP (Fig. 5A). In contrast, Ca 2ϩ /H ϩ exchange activity mediated by full-length CrCAX1 was significantly reduced (by 78%) but was detectable over basal levels (Fig. 5B). A competition 45 Ca 2ϩ uptake assay was performed to assess the substrate specificity of sCrCAX1. A 10-fold excess concentration of nonradioactive Ca 2ϩ and Cd 2ϩ could signifi-  cantly inhibit 10 M 45 Ca 2ϩ uptake (Fig. 5C). A 10-fold excess of Na ϩ (as 0.1 mM NaCl), in addition to 0.1 mM NaN 3 present in the uptake buffer, making the total Na ϩ concentration 0.2 mM, could inhibit ϳ60% of 45 Ca 2ϩ uptake, whereas Ca 2ϩ and Cd 2ϩ could inhibit ϳ85-95% of uptake. A 10-fold excess of Co 2ϩ could only marginally inhibit sCrCAX1 45 Ca 2ϩ uptake.
Expression of CrCAX1 in Arabidopsis-CrCAX1 and sCrCAX1 were expressed highly under the control of the constitutive CaMV 35S promoter in wild type Arabidopsis Col-0 ( Fig. 6B) to assess the impact that this cation transporter had to higher plant stress tolerance. CrCAX1-or sCrCAX1-expressing plants had no obvious alteration in morphology in nonstressed conditions. No phenotypic changes were observed between empty vector control and CAX-expressing lines under most ion stress conditions tested, except when grown under salt stress conditions. Multiple CrCAX1 and sCrCAX1 lines were able to provide significant tolerance when seedlings were germinated and grown on half-strength Murashige and Skoog medium containing 100 mM NaCl, as observed by enhanced growth and increased chlorophyll content compared with the vector control lines (Fig. 6, A and C). No significant difference was observed between the CrCAX1 and sCrCAX1 lines.
Na ϩ /H ϩ Exchange Activity of CrCAX1-The salt tolerance phenotype of yeast and Arabidopsis seedlings expressing CrCAX1, coupled with the ability of Na ϩ to inhibit Ca 2ϩ /H ϩ exchange activity, suggested that Na ϩ may also be a substrate for transport. Na ϩ /H ϩ exchange activity was assessed in vacuolar-enriched membrane vesicles isolated from nhx1⌬ yeast lacking the endogenous endomembrane Na ϩ /H ϩ exchanger NHX1 and expressing sCrCAX1 and CrCAX1. Na ϩ -dependent H ϩ flux was determined by the fluorescence quenching of acridine orange. After H ϩ pumping into the vesicles was initiated by the addition of Mg 2ϩ -ATP, fluorescence quenching was observed until a steady-state pH gradient was obtained. The addition of 5 M of the protonophore FCCP or 0.1% Triton X-100 completely abolished the pH gradient (data not shown). In the vector control vesicles from nhx1⌬ yeast, basal levels of quench recovery were observed following Na ϩ addition, presumably because of the remaining endogenous Na ϩ /H ϩ exchange activity (Fig. 7). Significant quench recovery following the addition of 50 mM NaCl was observed in yeast express- A, saturated liquid cultures of K667 (cnb1 pmc1 vcx1) yeast expressing CrCAX1 in piHGpd, N-terminally truncated sCrCAX1 in piHGpd, sAtCAX1 in piHGpd and empty vector alone (piHGpd) were serially diluted to the cell densities as indicated and then spotted onto selection medium lacking histidine (Ϫhis) and yeast extract-peptone-dextrose medium containing 200 mM CaCl 2 , 0.75 M NaCl, and 2.5 mM CoCl 2 . Yeast growth at 30°C is shown after 4 days. A representative experiment is shown. B, RT-PCR analysis of CrCAX1 and sCrCAX1 expression in K667 yeast compared with yeast expressing sAtCAX1 and empty vector control. C, AXT3 (ena1 nha1 nhx1) yeast expressing CrCAX1 in piHGpd, N-terminally truncated sCrCAX1 in piHGpd, sAtCAX1 in piHGpd, and empty vector alone (piHGpd) were diluted to a cell density of A 600 nm 0.5 and inoculated into arginine-phosphate-dextrose medium containing 1 mM KCl and various concentrations of NaCl as indicated. Yeast cell density was determined by A 600 nm measurements following growth shaking at 30°C for 48 h. The data represent the means of six to eight replicates, and the bars indicate S.E. ing CrCAX1 and sCrCAX1 (Fig. 7). No significant K ϩ /H ϩ or Li ϩ /H ϩ exchange activity was observed. In the K667 yeast background lacking the Ca 2ϩ /H ϩ exchanger ScVCX1 but with the endogenous Na ϩ /H ϩ exchanger ScNHX1 still present, significant background Na ϩ /H ϩ exchange activity was measured, although the expression of CrCAX1 did slightly enhance Na ϩ /H ϩ exchange activity (data not shown).
Transcriptional Regulation of CrCAX1 by Metal Stress Conditions-Chlamydomonas cells were grown in TAP medium supplemented with various metal salts, and CrCAX1 mRNA transcript expression was determined following 16 h of treatment by RT-PCR. CrCAX1 expression was not enhanced by any metal treatment tested (data not shown). Following treatment with a range of CaCl 2 concentrations, there was no change in CrCAX1 expression until treatment with 50 mM CaCl 2 and higher when a significant reduction of transcript was observed (Fig. 8A). Excess concentrations of NaCl above 100 mM similarly caused a significant reduction in CrCAX1 expression, but excess concentrations of other metal salts such as KCl did not. Despite a reduction in CrCAX1 transcript in 100 mM NaCl, growth of Chlamydomonas in this medium was not significantly impaired, although cell growth and chlorophyll content was reduced when cells were grown in 200 mM NaCl and in the 50 mM and 100 mM CaCl 2 conditions (Fig. 8B).

DISCUSSION
Carefully regulated Ca 2ϩ partitioning is an essential requirement in all cells to prevent toxic accumulation of Ca 2ϩ in the cytosol but also to allow Ca 2ϩ to fulfill a role as a cellular signal.   A complement of Ca 2ϩ -binding proteins, Ca 2ϩ influx channels, and Ca 2ϩ efflux transporters comprising of Ca 2ϩ -ATPases and ion-coupled Ca 2ϩ exchangers, allow cytosolic Ca 2ϩ levels to be dynamically shaped (1). In addition, various Ca 2ϩ sensing and effector proteins including Ca 2ϩ -dependent kinases and phosphatases allow the generated Ca 2ϩ signals to be decoded and mediate downstream responses (32). Combinations of these components are likely to be present in all eukaryotic cells where Ca 2ϩ -mediated signaling occurs, and indeed genomic sequence and proteomic analyses have begun to identify genes and proteins likely to be involved in Ca 2ϩ homeostasis and signaling in Chlamydomonas (21,33). These include putative Ca 2ϩ -ATPases, Ca 2ϩ exchangers, and Ca 2ϩ channels, some of which are similar to animal-type transporters not found in higher plants (21,33). To date, all of these transporters are uncharacterized, one exception being a light-activated, Ca 2ϩ -permeable cation channel COP4 that plays a role in photoreceptor mediated Ca 2ϩ signaling (34). In this report we have described the first characterization of a Chlamydomonas Ca 2ϩ efflux transporter and described the first analysis of a Type I-C CAX transporter (9).
An unexpected and intriguing feature of the CrCAX1 transporter is that in addition to having Ca 2ϩ transport activity, this exchanger could also transport Na ϩ (Fig. 7), and this Na ϩ /H ϩ exchange activity allowed CrCAX1 to provide significant tolerance to NaCl when expressed in yeast or plants (Figs. 4 and 6). The physiological significance of putative Na ϩ transport activity by CrCAX1 in Chlamydomonas is likely to be due to the requirement of robust mechanisms of salt stress tolerance in a unicellular organism, particularly one such as Chlamydomonas that lives in aquatic environments with potentially harsh ionic conditions. By analogy with other unicellular organisms such as yeast, it is likely that Chlamydomonas will have multiple pathways for Na ϩ extrusion from the cytosol. For example, genome sequence analysis indicates that Chlamydomonas also has NHX-type Na ϩ /H ϩ exchangers (data not shown), although their function in Na ϩ transport and tolerance has yet to be demonstrated. Further work will need to elucidate the relative roles of both NHX-and CAX-mediated Na ϩ transport in Chlamydomonas.
Although some CAX transporters such as ScVCX1 are specific for Ca 2ϩ (10), others such as AtCAX2 have a broader substrate specificity and can transport a variety of divalent cations like Ca 2ϩ , Mn 2ϩ , Cd 2ϩ , and possibly Zn 2ϩ (12). CrCAX1 is the first eukaryotic CAX transporter that has been shown to directly efflux both divalent and monovalent cations. There is some precedence for Na ϩ transport by CAX-type exchangers like ScVNX1 (13), indicating that Na ϩ /H ϩ exchange activity is not restricted to NHX-type exchangers (7). Similarly, a plant CAX transporter from soybean, GmCAX1, was previously shown to provide slight tolerance to Na ϩ when expressed in Arabidopsis, but no direct Na ϩ transport activity was demonstrated (20). In addition, some bacterial CAX-and NHX-type exchangers may transport both Na ϩ and Ca 2ϩ , as with chaA from Escherichia coli (35,36), Aa-caxA from Alkalimonas amylolytica N10 (37), and ApNhaP from Aphanothece halophytica (38).
We might expect that the ability to bind and efflux Ca 2ϩ and Na ϩ is determined by the amino acid sequence. Previous anal-ysis of plant CAX transporters has identified two highly conserved sequence repeats named c-1 and c-2 (supplemental Fig.  S1) that have been suggested to function as cation selectivity filters (39). These regions are similar and show sequence conservation with the ␣-1 and ␣-2 regions required for cation binding in mammalian Na ϩ /Ca 2ϩ exchangers (8,39). Specific residues within the c-1 and c-2 regions have been shown to be required for Ca 2ϩ transport (39,40), and these residues are conserved in the CrCAX1 c-1 and c-2 regions (Fig. 9). It is unknown whether these domains will also determine Na ϩ binding. A comparison of the CrCAX1 c-1 and c-2 region with other CAX transporters, none of which have been shown to transport Na ϩ , highlights two residues in CrCAX1 that differ from the consensus sequence for these regions: Met 170 in c-1, which is Ile in the consensus sequence, and Cys 350 in c-2, which is Ala in the consensus sequence (Fig. 9). Future analysis will be able to determine whether these differences determine functional characteristics to CrCAX1. The fact that Na ϩ transport is not regulated by the putative N-terminal auto-regulatory domain may suggest differences in the mechanisms of Na ϩ and Ca 2ϩ transport. Thus, other domains of CrCAX1 may determine Na ϩ selectivity and transport. A cytosolic region called the acidic domain between transmembrane spans 6 and 7 has been suggested to be involved in Ca 2ϩ binding (35), possibly analogous to the role of the cytosolic Ca 2ϩ -binding ␤ repeat domains of Na ϩ /Ca 2ϩ exchanger proteins (41). The CrCAX1 acidic domain is shorter and possesses fewer negatively charged residues than in AtCAX1 or ScVCX1 (supplemental Fig. S1). . a denotes aliphatic residues (Leu, Ile, and Val), and h denotes hydroxylic (Ser and Thr) residues. The alignments were performed using ClustalW2. Shading was performed by Boxshade. Black shading denotes identical residues, and gray shading denotes similar residues. Residues shown in AtCAX1 and OsCAX1a to be essential for Ca 2ϩ transport (39,40) are identified above the sequence by a ϩ. Asterisks highlights residues in CrCAX1 (Met 170 and Cys 350 ) that differ from the consensus sequence. The numbers describe the sequence positions of the amino acids. The accession numbers of CAX gene sequences are given in supplemental Table S3.
Whether this region plays a role in cation selectivity also remains to be determined.
The identification and analysis of CrCAX1 allows further functional and regulatory comparison between the CAX transporters, particularly with those that have been studied in detail: S. cerevisiae VCX1 and Arabidopsis AtCAX1. Despite having greater sequence similarity to ScVCX1 than AtCAX1 ( Fig. 1 and supplemental Table S2), CrCAX1 has a broader substrate specificity than ScVCX1, showing both Ca 2ϩ /H ϩ and Na ϩ /H ϩ exchange activity, and possibly the ability to transport Co 2ϩ and Cd 2ϩ (Figs. 5 and 7). A particularly intriguing feature of CrCAX1 is the structural similarities with the higher plant CAX transporters with regard to the extended hydrophilic N-terminal tail (supplemental Fig. S1). Comparison of full-length and an N-terminally truncated CrCAX1 mutant indicates that the autoinhibitory regulation mechanism is shared between Chlamydomonas and Arabidopsis, suggesting that this mechanism might be conserved in CAX transporters in all algae and higher plant lineages. Analysis of CAX transporters from mung bean and rice also show N-terminal regulation (31,42), thus supporting this view. However, the N-terminal domain of CrCAX1 only regulated Ca 2ϩ /H ϩ exchange activity but not Na ϩ /H ϩ exchange, suggesting that the autoinhibitory mechanism is more complex than just blocking cation transport.
Like many plant CAX transporters, CrCAX1 was transcriptionally regulated by excess ions (Ca 2ϩ and Na ϩ ) in the growth medium (Fig. 8A). AtCAX1 mRNA transcript level is significantly increased when seedlings are grown in high Ca 2ϩ medium (18), presumably to allow sequestration of Ca 2ϩ into the plant vacuole and provide Ca 2ϩ tolerance. However, CrCAX1 transcript decreased as cells were grown in excess Ca 2ϩ . This is analogous to the observed inhibition of ScVCX1 by calcineurin in response to elevated cytosolic Ca 2ϩ (10). Although CrCAX1 can clearly play a role in providing tolerance to excess concentrations of Ca 2ϩ and Na ϩ , as shown by the yeast expression experiments, CrCAX1 levels may be moderated in vivo to prevent excessive accumulation of ions into the cell. This may be important for unicellular organisms that have a finite capacity for intracellular ion storage. This may also suggest that CrCAX1 plays an important role in the generation of Ca 2ϩ signals rather than solely for providing Ca 2ϩ tolerance.
These studies also suggest that there may be divergent mechanisms of regulation between CrCAX1 and AtCAX1 at the level of translation. The in vitro transcription and translation experiment confirmed that translation of CrCAX1 did initiate from the predicted AUG start codon at position 1 with no translation initiation from the downstream AUG codon (Fig. 2). As a comparison, we performed the same experiment with an AtCAX1 construct, which has two AUG codons in close proximity at the 5Ј end, with a second AUG 109 bp from the first, encoding a Met residue at position 37 (supplemental Fig. S1). Although the majority of translation initiated from the first AUG (M1), a small proportion (16%) of translation initiated from the second AUG (M37). Such alternate translation initiation from a single mRNA transcript has previously been observed. Alternate initiation of Arabidopsis DNA ligase 1 alters the intracellular targeting of the AtLIG1 protein (23). The experiment in this report suggests that alternate translation initiation could also be a means to regulate activity of some proteins, such as AtCAX1. We have previously demonstrated that removal of the first 36 residues of AtCAX1 up to Met 37 significantly activates Ca 2ϩ transport activity (14); thus this apparent alteration in translation of AtCAX1 could signify an additional level of regulation. Further experiments will be required to determine whether AtCAX1 is alternatively translated in planta.
Expression of CrCAX1-GFP in yeast showed localization at the vacuole, consistent with the ability of CrCAX1 to suppress the Ca 2ϩ -hypersensitive phenotype of a yeast mutant lacking two vacuolar Ca 2ϩ transport pathways. Unfortunately we were unsuccessful in attempts to observe expression of CrCAX1-GFP in Chlamydomonas despite using a codon-optimized GFP tag. However, localization of CrCAX1-GFP to the tonoplast in yeast and Arabidopsis (data not shown) indicates that CrCAX1 is similarly localized to a vacuolar membrane in vivo. In addition, CrCAX1 is predicted to be vacuolar localized by computer-based prediction (22). Antibodies raised against plant tonoplast proteins (vacuolar H ϩ -pyrophosphatase) detect proteins at the tonoplast in Chlamydomonas (43), suggesting some conservation of protein vacuolar targeting between Chlamydomonas and higher plants. Chlamydomonas are not as vacuolated as typical plant or yeast cells. They possess two contractile vacuoles that are involved in osmoregulation and numerous small acidic vacuoles that appear similar to acidocalcisomes (43,44). Acidocalcisomes are acidic organelles that contain a matrix of pyrophosphate and polyphosphates bound to Ca 2ϩ (45). They have been identified and characterized predominantly in protozoan parasites and are proposed to play a role in cation storage and adaptation to environmental stress. The Chlamydomonas acidocalcisome membrane, like that of the plant vacuolar membrane, contains both a H ϩ -pyrophosphatase and a V-type H ϩ -ATPase that drive H ϩ transport into the organelle and establish a pH gradient (43). This gradient would therefore be sufficient to energize cation/H ϩ exchange by CrCAX1 of Ca 2ϩ and Na ϩ into the acidocalcisome. It is also interesting to note that the genomes of many of the protozoan species that store Ca 2ϩ in acidocalcisomes including P. falciparum possess single CAX genes that cluster in the Type I-C clade alongside CrCAX1 (Fig. 1). However, we cannot rule out the localization of CrCAX1 at another membrane such as the contractile vacuole.
We have provided evidence that Chlamydomonas posses a CrCAX1 protein that can transport both Ca 2ϩ and Na ϩ by a cation/H ϩ exchange mechanism and have suggested a role for this transporter in providing tolerance to cation stress based on heterologous expression experiments in yeast and Arabidopsis. The identification of CrCAX1 as a Na ϩ transporter may help understand osmotic stress regulation by this unicellular organism. In addition, the identification of a Chlamydomonas Ca 2ϩ transporter gene provides the potential to advance our understanding of Ca 2ϩ signaling in this model eukaryote. Chlamydomonas is already established as an excellent model for Ca 2ϩ signaling studies. Ca 2ϩ signaling is required in phototaxis and aspects of flagellar function including excision require specific Ca 2ϩ signals (34, 46 -48). It was recently demonstrated that the efficient delivery of Ca 2ϩ indicator dyes into Chlamydomonas cells provides the ability to monitor cytosolic Ca 2ϩ oscillations, such as in response to deflagellation (48,49). Further charac-terization of CrCAX1 and homologous Ca 2ϩ transporters coupled with the ability to monitor Ca 2ϩ changes should further our knowledge in the generation and specificity of these Ca 2ϩ signals in this species.