Research article

Comparative analysis of CAX2-like cation transporters indicates functional and regulatory diversity

Clare Edmond, Toshiro Shigaki, Sophie Ewert, Matthew D. Nelson, James M. Connorton, Vesela Chalova, Zeenat Noordally, Jon K. Pittman


Internal compartmentalization of metals is an important metal tolerance mechanism in many organisms. In plants and fungi, sequestration into the vacuole is a major detoxification mechanism for metals. Cation transport into the vacuole can be mediated by CAX (cation exchanger) transporters. The Arabidopsis thaliana AtCAX2 transporter was shown previously to transport Ca2+, Cd2+ and Mn2+. To assess the conservation of the functional and regulatory characteristics of CAX2-like transporters in higher plants, we have characterized AtCAX2 orthologues from Arabidopsis (AtCAX5), tomato (LeCAX2) and barley (HvCAX2). Substrate specificity and regulatory activity were assessed using a yeast heterologous-expression assay. Each CAX could transport Ca2+ and Mn2+ into the yeast vacuole, but they each had different cation transport kinetics. Most notably, there was variation in the regulation of the transporters. As found with AtCAX2 previously, only expression of an N-terminally truncated form of AtCAX5 in yeast was able to mediate Ca2+ and Mn2+ transport, indicating that activity may be controlled by an autoregulatory region at the N-terminus. In contrast, either full-length or truncated LeCAX2 could efficiently transport Ca2+, although Mn2+ transport was controlled by the N-terminus. HvCAX2 did not appear to possess an N-terminal regulatory domain. Expression of AtCAX2 was not significantly modulated by metal stress; however, AtCAX5 and HvCAX2 were transcriptionally up-regulated by high Mn2+ treatment, and by Ca2+ and Na+ stress respectively. It is therefore apparent that, despite the high sequence identity between plant CAX2 orthologues, there is significant diversity in their functional characteristics, particularly with regard to regulatory mechanisms.

  • Arabidopsis
  • barley
  • calcium and manganese transport
  • cation/proton antiporter
  • regulation
  • tomato


All organisms have an essential requirement for a wide range of metals, such as various trace metals including Zn2+, Fe2+ and Mn2+, which function as enzyme co-factors and protein components, and metals such as Ca2+, that have cell signalling roles. These essential metals are extremely toxic at elevated concentrations, and thus many organisms have sophisticated mechanisms which they use to tolerate metal stress. The removal of metals out of the cytoplasm is an important tolerance mechanism. In plants and fungi, the sequestration of metal ions or metal complexes into the vacuolar compartment is a critical response to metal stress. For example, when yeast (Saccharomyces cerevisiae) is deleted for the vacuolar Cd2+–conjugate transporter YCF1, it is extremely sensitive to Cd2+ stress [1]. Likewise, the accumulation of excess Ca2+ or Mn2+ into plant vacuoles can provide tolerance to these stresses [2,3]. The molecular and biochemical characterization of such vacuolar sequestration pathways is, however, limited to just a few transporters from a small number of species.

Many ion sequestration pathways utilize the electrochemical proton gradient generated by proton pumps [4]. Higher plants, such as Arabidopsis thaliana, possess a variety of gene families which together encode many putative proton-coupled ion transporters, some of which are predicted to be vacuolar localized and high-capacity cation transporters [4,5]. One such gene family is the CAX (cation exchanger) family, which has 6 members in Arabidopsis and a similar number in rice [6,7]. AtCAX1 was identified as a high-capacity low-affinity Ca2+/H+ antiporter from a yeast Ca2+-hypersensitivity suppression screen [8] and has subsequently been shown to be involved in Ca2+ signalling events and to provide tolerance to excess Ca2+ [3,911]. AtCAX2 has a lower affinity for Ca2+ transport than AtCAX1 [8] and can transport a range of cations into the vacuole, including Mn2+ and Cd2+ [1215]. Knockout analysis suggests that AtCAX2 does not play a major physiological role in Ca2+ homoeostasis, but is important for vacuolar Mn2+ accumulation [14]. All plant CAX transporters characterized to date appear to be able to transport Ca2+, but further analysis is needed to determine whether a broad metal-substrate range is a common characteristic of CAX proteins.

Other notable features to emerge from the characterization of Arabidopsis CAX transporters are their modes of regulation. AtCAX1 is expressed at high levels in leaf and flower tissues [16] and is induced by elevated Ca2+ [3]. In contrast, AtCAX2 is detected at fairly low levels in all tissues, but is not greatly induced by any metal [12]. AtCAX1 is also regulated post-translationally [17]. A domain on the N-terminal tail of AtCAX1 regulates Ca2+ transport activity through an autoinhibitory process [18]. If expressed in yeast or tobacco, full-length AtCAX1 is inactive, whereas N-terminal truncated or mutated AtCAX1 shows deregulated activity [10,19]. Interaction with activator proteins or phosphorylation appears to activate AtCAX1 [19,20]. AtCAX2 is likely to be regulated in a similar manner, as it is unable to suppress Ca2+ and Mn2+ sensitivity of yeast unless it is N-terminally truncated [14,21]. There is some evidence that CAX transporters from other plants are regulated by a similar mechanism [7,22]; however, it is still unclear whether this mode of regulation is ubiquitous among plant CAX-type transporters.

Previous phylogenetic analysis has shown that higher-plant CAX genes can be clustered into two groups, type IA and type IB [6]. The majority of the plant CAX genes characterized to date are from the type IA group, which includes AtCAX1, AtCAX3, VCAX1 and OsCAX1a [16,23,24]. The type IB group includes AtCAX2 and OsCAX3 [7]. The hypothesis of the present study is that these two groups of transporters have different functions and characteristics, but further evidence is needed to clarify this. It is also unclear how much functional variation exists within each group and whether the characteristics of the Arabidopsis genes can be used as a paradigm for orthologous CAX genes from other species within each group. This is important if we wish to utilize effectively the wealth of knowledge gained from the model plant Arabidopsis and translate this into economically important crop plant species. The identification of CAX2-like transporters in crops, such as rice or barley, that can mediate the vacuolar sequestration of essential metal nutrients will be attractive candidates for future plant improvement studies to generate plants with improved nutritional content. We have therefore focused further analysis on the characterization of AtCAX2-related genes. In the present study, we have characterized plant CAX2-like cDNAs, including an Arabidopsis homologue of AtCAX2, termed AtCAX5, and identified two orthologous cDNA clones from a solanaceous dicot plant, tomato (LeCAX2), and the monocot cereal barley (HvCAX2). Analysis of the substrate specificity and regulatory characteristics of these related transporters suggests that, although some characteristics are conserved amongst CAX2-like genes, there is still significant variation in the function of members of this CAX type IB phylogenetic clade.


Plant and yeast material

A. thaliana (accession Col-0) was grown at 22 °C in 18 h light/6 h dark on solid 0.5×MS (Murashige and Skoog) medium or on soil. Tomato (Lycopersicon esculentum cv. Microtom) and barley (Hordeum vulgare cv. Maris Otter) were grown on soil under glasshouse conditions. Barley seedlings were also grown on 0.5×MS plates. For metal-stress treatments, Arabidopsis plants grown on 0.5×MS plates for 2 weeks or barley seedlings grown on 0.5×MS medium for 2 weeks were transferred on to 0.5×MS medium containing the metal salt and incubated for 12 h until the tissue was harvested. For AtCAX6::GUS (β-glucuronidase) analysis, transgenic Arabidopsis generation and the histochemical GUS assay were performed as described previously [14]. The S. cerevisiae strain K667 (cnb1::LEU2 pmc1::TRP1 vcx1Δ) [25] was used for heterologous expression using the yeast expression vector p2HGpd. Yeast transformation and growth analysis were carried out as described previously [13]. Protein was isolated from yeast, and Western blotting was performed as described previously [22]. YFP (yellow fluorescent protein)- and CFP (cyan fluorescent protein)-tagged CAX proteins expressed in yeast were visualized by confocal and epifluorescence microscopy using Leica SP5 and DMR microscopes (Leica Microsystems) and an argon laser with standard YFP and CFP excitation and emission wavelengths (for confocal microscopy) or YFP and CFP filter cubes (Chroma Technology) for epifluorescence microscopy. The fluorescent yeast vacuole marker stain carboxy-DCFDA [5-(and 6-)carboxy-2′,7′-dichlorofluorescein diacetate] [26] was detected using a L4 FITC filter cube.

Plasmid DNA constructs

The full-length and truncated AtCAX2 cDNA constructs were generated as described previously [13,14]. The AtCAX5 cDNA was identified as an EST (expressed sequence tag) clone (Entrez Nucleotide EST database accession number BG459283) obtained from Professor C. Benning (Department of Biochemistry and Molecular Biology, Michigan State University, MI, U.S.A.). The primers sCAX5FOR and CAX5REV (see Supplementary Table S1 at for primer sequences) were used to amplify the sAtCAX5 cDNA (where sAtCAX5 encodes N-terminally truncated AtCAX5). Full-length AtCAX5 was amplified using CAX5FOR and CAX5REV. sAtCAX5 and AtCAX5 were cloned into pGEM-T Easy (Promega) for DNA sequencing. Intron number 5 was removed from sAtCAX5 and AtCAX5 by PCR mutagenesis [27] using CAX5Mut primers. sAtCAX5 and AtCAX5 were subcloned into the XbaI and NotI sites of p2HGpd. To generate an N-terminal YFP–AtCAX2 fusion construct, EYFP (enhanced YFP) cDNA (Clontech), amplified using YFPFOR and YFP/CFPREV, was cloned into the BamHI and NcoI sites of AtCAX2 (amplified using CAX2tagFOR with CAX2REV [14]). To generate an N-terminal CFP–AtCAX5 fusion construct, ECFP (enhanced CFP) cDNA (Clontech), amplified using CFPFOR and YFP/CFPREV, was cloned into the XbaI and NcoI sites of AtCAX5 (amplified using CAX5tagFOR and CAX5REV). Both constructs were subcloned into p2HGpd. To generate AtCAX6–GUS, the 0.8 kb promoter fragment of AtCAX6 was amplified using the primers CAX6PROMF and CAX6PROMR and transcriptionally fused to GUS by subcloning into the modified pBI121 plasmid [14]. Barley EST sequences (including Entrez Nucleotide EST database accession numbers AV834864 and AV910358) were identified with high sequence identity with the 5′- and 3′-ends of AtCAX2. The primers HvCAX2FOR and HvCAX2REV amplified HvCAX2 cDNA from barley root and leaf RNA by RT-PCR (reverse transcription-PCR). HvCAX2 was cloned into pGEM-T Easy (Promega) for DNA sequencing, and then subcloned into p2HGpd. HvCAX2 cDNA was also obtained from a barley embryo cDNA library (from Dr C. Bray, Faculty of Life Sciences, University of Manchester, Manchester, U.K.) using a PCR-based cDNA-screening method [28]. 5′-RACE (rapid amplification of cDNA ends) was performed using a RACE kit (Roche) and the gene-specific primers HvCAX2SP1 and HvCAX2SP2. A barley cDNA clone designated ‘CAX’ (Entrez Nucleotide EST database accession number AB218888) identified by a BLAST search with HvCAX2 search was identical at the amino acid level with HvCAX2 cDNA. A C-terminal tagged sAtCAX2–c-Myc fusion protein was generated previously [13]. HvCAX2–c-Myc was generated using the same method by cloning the c-Myc sequence into HvCAX2 amplified using HvCAX2FOR and HvCAX2tagREV. Various tomato EST sequences with high identity with 5′- and 3′-ends of AtCAX2, and an unnamed full-length cDNA sequence (Entrez Nucleotide EST database accession number BT014476) with high sequence identity with AtCAX2, were identified by BLAST. The primers LeCAX2FOR and LeCAX2REV amplified LeCAX2 cDNA by RT-PCR from tomato leaf RNA. LeCAX2 was cloned into pGEM-T Easy (Promega) for DNA sequencing, and then subcloned into p2HGpd. N-terminally truncated sLeCAX2 was amplified using sLeCAX2FOR and LeCAX2REV.

RNA extraction and cDNA amplification

RNA was isolated from Arabidopsis, barley and tomato tissues using an RNA extraction kit (Qiagen) and was DNase-treated prior to RT-PCR using SuperScriptII reverse transcriptase (Invitrogen) and an oligo(dT)primer. Expand (Roche) DNA polymerase (for Arabidopsis and tomato) or Phusion (Finnzymes) DNA polymerase with GC buffer (for barley) were used with the following PCR conditions: 94 °C for 2 min, followed by 40 cycles of 94 °C for 30 s, 60 °C for 1 min and 72 °C for 2 min, for the amplification of all full-length cDNA. Arabidopsis and barley quantitative CAX transcript expression was performed by qRT-PCR (quantitative real-time PCR) using a qPCR SYBR Green kit (Eurogentec) and an ABI 7000 detection system (Applied Biosystems). As an internal standard, actin (ACT2) primers for Arabidopsis (AtACT2F and AtACT2R) or barley (HvACTF and HvACTR) were used. AtCAX2 primers (AtCAX2F and AtCAX2R) were designed using the 3′-UTR (untranslated region) sequence, and the AtCAX5 (AtCAX5F and AtCAX5R) and AtCAX6 (AtCAX6F and AtCAX6R) primers were designed using the 5′-UTR sequence. HvCAX2 primers (HvCAX2F and HvCAX2R) were designed against the ORF (open reading frame).

Preparation of membrane vesicles and transport analysis

Vacuolar-enriched membrane vesicles were prepared from yeast as described previously [18]. Time-dependent 10 μM 45CaCl2 uptake measurements into membrane vesicles were performed as described previously [9]. The 200 μM MnCl2-dependent Acridine Orange fluorescence recovery assay was performed as described previously [29], except using vacuolar-enriched membrane vesicles. Metal competition experiments were performed as described previously [13], except using vacuolar-enriched membrane vesicles.


Expression of an Arabidopsis CAX2-like gene in response to Mn2+ stress

It has been found previously that vacuolar Mn2+/H+ antiport activity in the Arabidopsis cax2 knockout mutant is significantly reduced compared with wild-type, but is not completely absent, suggesting the presence of additional vacuolar transporters that are responsible for Mn2+/H+ transport [14]. Arabidopsis has two uncharacterized genes which cluster with AtCAX2 (At3g13320) in the type IB group, named AtCAX5 (At1g55730) and AtCAX6 (At1g55720) [6]. These three genes are highly similar to each other (AtCAX5 is 87% identical with AtCAX2 and 88% identical with AtCAX6 at the predicted amino acid level), suggesting that they may have similar functions. The expression level of AtCAX5 and AtCAX6 in comparison with AtCAX2 was compared in Arabidopsis seedlings. Using qRT-PCR and UTR-sequence primers specific to each CAX gene, we could show that AtCAX5 and AtCAX6 were both expressed, but at lower expression levels compared with AtCAX2 (Figure 1a). Expression of AtCAX6 was barely detectable, although AtCAX6 promoter–GUS reporter analysis and RT-PCR confirmed that AtCAX6 was expressed at very low levels, predominantly in the leaf petiole (see Supplementary Figures S1a and S1b at

Figure 1 Expression of AtCAX5 in Arabidopsis and in response to metal stress

(a) Expression levels of AtCAX2, AtCAX5 and AtCAX6 in 2-week-old Arabidopsis seedlings grown on 0.5×MS growth medium. (b) Expression of AtCAX2 and AtCAX5 in various tissues of Arabidopsis. (c) Expression of AtCAX2 and AtCAX5 in Arabidopsis seedlings in response to different metal stress conditions. RNA was isolated from whole-plant tissue 12 h after treatment with 0.5×MS growth medium alone (basal), 0.5×MS growth medium supplemented with various metal salts (Ca, 100 mM CaCl2; Cd, 10 μM CdCl2; Co, 1 mM CoCl2; K, 80 mM KCl; Mn, 2 mM MnCl2; Na, 100 mM NaCl; and Zn, 5 mM ZnCl2), or 0.5×MS growth medium prepared with Ca2+ omitted (no Ca). In all experiments, expression of each CAX gene was determined by qRT-PCR relative to actin expression and is shown as a percentage of actin. All results are means±S.E.M., with each experiment repeated three or four times.

As AtCAX5 had the highest expression level of the two AtCAX2-like genes, this gene was characterized further. We wished to examine the tissue-specific expression pattern of AtCAX5 and its expression in response to various metal stresses. Like AtCAX2, AtCAX5 was expressed in all tissues, particularly in the stem and root, but less in the leaf (Figure 1b). This expression profile was equivalent to that determined by publicly available microarray analyses. It is important to note, however, that most Arabidopsis microarrays cannot differentiate between AtCAX5 and AtCAX6.

Some of the Arabidopsis CAX transporters are transcriptionally up-regulated in response to particular metal stress conditions [3]. Although AtCAX2 is able to transport Ca2+, Cd2+ and Mn2+, its expression is not induced by these metals [12] (Figure 1c). Expression of AtCAX5 was induced slightly by Cd2+ and Ca2+ compared with water control treatment, and significantly following Mn2+ treatment (Figure 1c). In contrast, some metal treatments, notably excess Zn2+, caused a reduction in AtCAX5 expression. Ca2+ depletion yielded a small increase in AtCAX5 expression, whereas AtCAX2 was slightly down-regulated by this treatment (Figure 1c).

Ca2+ and Mn2+ stress tolerance of yeast by AtCAX5

An EST cDNA clone for AtCAX5 was obtained (accession number BG459283) that appeared to be full-length by restriction-digest analysis; however, this AtCAX5 cDNA was not completely spliced, as a 102 bp intron (intron 5) was still present, which would cause a truncation of the encoded protein at transmembrane domain 5. The intron was removed by mutagenesis, and the 1326 bp cDNA and the predicted amino acid sequence were identical with the predicted sequence annotation of At1g55730. AtCAX5 was also amplified by RT-PCR from RNA isolated from various Arabidopsis tissues. A single transcript was identified in stem and leaf tissue, whereas two bands of approx. 1.3 and 1.4 kb in size were amplified from fruit RNA (results not shown). The 1.4 kb band was AtCAX5 containing intron 5, indicating that both the fully processed and the incompletely spliced form of AtCAX5 are expressed in certain tissues. Whether a truncated protein is expressed from this transcript and has a function is unclear. The AtCAX5 amino acid sequence has significant similarity to AtCAX2 and an equivalent predicted topology, with 11 predicted transmembrane-spanning domains, a longer loop region between transmembrane spans 6 and 7, which has multiple acidic residues (the acidic motif), and a long hydrophilic N-terminal tail (see Supplementary Figure S2 at The three-amino-acid (Cys-Ala-Phe) Mn2+-specificity determinant of AtCAX2, positioned within transmembrane-spanning domain 4 [13], is also conserved in AtCAX5, suggesting that it may also transport Mn2+.

The cation/H+ antiport activity of AtCAX2 is largely undetectable when expressed in yeast unless the N-terminus is truncated [14]. To examine whether AtCAX5 was similarly regulated, the Ca2+- and Mn2+-hypersensitive yeast mutant strain K667 was transformed with full-length AtCAX5 and sAtCAX5, in which translation was initiated at Met43. AtCAX5 was unable to confer growth on high Ca2+- and Mn2+-containing medium, whereas sAtCAX5 was able to suppress both the Ca2+- and Mn2+-hypersensitivity of K667 (Figure 2a). In comparison with sAtCAX2, the ability of sAtCAX5 to suppress these metal phenotypes was not as efficient. Growth of sAtCAX5-expressing K667 yeast was weaker on 250 mM CaCl2 and on 10 mM MnCl2 than the sAtCAX2-expressing strain, whereas growth of the sAtCAX5 strain on 15 mM MnCl2 was extremely weak (see Supplementary Figure S3a at This reduced tolerance efficiency was not the result of an altered expression level or membrane localization. N-terminal YFP–AtCAX2 and CFP–AtCAX5 fusion proteins were both localized to the vacuole when expressed in yeast, as shown by the equivalent localization pattern with the vacuolar marker stain [26] carboxy-DCFDA (Supplementary Figure S3b), specifically at the vacuolar membrane (Figure 2b).

Figure 2 Suppression of Ca2+ and Mn2+ hypersensitivity of yeast by N-terminally truncated AtCAX5

(a) Suppression of Ca2+ and Mn2+ sensitivity of the pmc1 cnb vcx1 yeast mutant (K667) by full-length AtCAX2 and AtCAX5 or N-terminally truncated sAtCAX2 and sAtCAX5 transporters. Saturated liquid cultures of K667 expressing various plasmids were diluted to the cell density (OD600) indicated and then spotted on to selection medium lacking histidine (−His) and YPD [1% (w/v) yeast extract/2% (w/v) peptone/2% (w/v) glucose] medium containing 175 mM CaCl2 (175 mM Ca) or 5 mM MnCl2 (5 mM Mn). Yeast growth at 30 °C is shown after 3 days, and a representative experiment is shown in each case. (b) Confocal fluorescence imaging of yeast cells expressing YFP–AtCAX2 and CFP–AtCAX5, demonstrating localization at the yeast vacuolar membrane.

AtCAX5 confers cation/H+ antiport activity

To confirm that the AtCAX5-mediated Ca2+ and Mn2+ tolerance of K667 was the result of cation transport activity, vacuolar-enriched membrane vesicles were isolated from sAtCAX5 expressing K667 yeast and the accumulation of 10 μM 45Ca2+ was measured. The sAtCAX5 Ca2+ transport was H+-dependent, as 45Ca2+ accumulation was significantly inhibited by the protonophore FCCP (carbonyl cyanide p-trifluoromethoxyphenylhydrazone) (see Supplementary Figure S4 at sAtCAX5 Ca2+ transport activity was reduced compared with sAtCAX2 (Figure 3a), although the Km for Ca2+ was equivalent for the two truncated transporters (Table 1). sAtCAX5 had a lower Vmax compared with sAtCAX2. No significant Ca2+/H+ antiport activity was detectable from yeast expressing full-length AtCAX5, AtCAX2 or vector only (Figure 3a).

Figure 3 Ca2+ and Mn2+ proton-dependent transport activity of AtCAX5

(a) Single 10-min time point measurements of ΔpH-dependent uptake of 10 μM 45Ca2+ into vacuolar membrane vesicles prepared from K667 yeast expressing sAtCAX2, sAtCAX5, AtCAX2, AtCAX5 and empty vector. Results are shown following subtraction of the protonophore FCCP background values. All experiments were performed following the establishment of a pH gradient in the presence of the Ca2+–ATPase inhibitor vanadate. Results are means±S.E.M. (n=3). (b) Time course of Mn2+/H+ antiport activity determined by Mn2+-dependent H+ transport. Acridine Orange fluorescence recovery was determined in vacuolar membrane vesicles prepared from K667 yeast expressing sAtCAX2 (■), sAtCAX5 (▲) and empty vector (○), following the establishment of a pH gradient and the addition of 200 μM MnCl2 at 0 min time point. Fluorescence recovery is shown as a percentage of fluorescence following the addition of Mn2+. Results are means±S.E.M. (n=3).

View this table:
Table 1 Kinetic characteristics of CAX-dependent Ca2+/H+ antiport activity

Initial rates of ΔpH-dependent 45Ca2+ uptake into vacuolar membrane vesicles isolated from K667 yeast expressing various CAX genes were calculated over a range of Ca2+ concentrations from 1 to 100 μM in the presence of protonophore (5 μM FCCP). All Michaelis–Menten kinetic values are means±S.E.M. of 2–3 independent experiments.

The Mn2+/H+ antiport activity of sAtCAX5 was determined by an Acridine Orange fluorescence quench assay, which monitored the Mn2+-dependent dissipation of the proton gradient in vesicles expressing sAtCAX5. The addition of Mn2+ dissipated the proton gradient significantly more rapidly in sAtCAX5-expressing membrane vesicles compared with empty vector control vesicles, although at a reduced level when compared with sAtCAX2 (Figure 3b). Addition of FCCP or Triton X-100 detergent caused a complete and rapid fluorescence recovery as a result of the loss of the proton gradient, whereas the addition of water alone did not give fluorescence recovery (results not shown).

The substrate specificity of sAtCAX5 was further analysed by competition of 45Ca2+ uptake with non-radioactive metals. 45Ca2+/H antiport activity was determined in the presence or absence of excess concentrations of non-radioactive metals (Ca2+, Mn2+, Cd2+, Zn2+, Ni2+ and Co2+) (Figure 4). Ca2+ and Cd2+ significantly inhibited 45Ca2+ transport by sAtCAX5 to levels equivalent to that seen with sAtCAX2; however, the degree of inhibition with Mn2+ was slightly less with sAtCAX5 than was observed with sAtCAX2. Furthermore, no inhibition of 45Ca2+ transport by Zn2+ was observed with sAtCAX5, whereas 45Ca2+ transport by sAtCAX2 was significantly inhibited by Zn2+ (Figure 4). Ni2+ and Co2+ did not inhibit Ca2+ transport by sAtCAX2 or sAtCAX5.

Figure 4 Substrate specificity of sAtCAX5 compared with sAtCAX2 as determined by competition analysis

ΔpH-dependent uptake of 10 μM 45Ca2+ into vesicles isolated from sAtCAX2-expressing K667 yeast (a) and from sAtCAX5-expressing K667 yeast (b). Ca2+/H+ antiport activity, estimated as the difference between uptake with and without the protonophore (5 μM gramicidin), was measured in the absence [control (Cont)] or the presence of 10× or 100×of non-radioactive CaCl2 (Ca), MnCl2 (Mn), CdCl2 (Cd), ZnCl2 (Zn), NiCl2 (Ni) or CoCl2 (Co) after 10 min. Ca2+ uptake values are shown following subtraction of the gramicidin background values and expressed as a percentage of the control in the absence of any excess non-radiolabelled metals. Results are means±S.E.M. (n=3) with each experiment performed with two to six replications.

Identification of barley and tomato CAX2-like genes

Sequence information from genomes and EST sequences from a variety of plant species indicates that CAX2-like (type IB) genes are present in most higher plants [6]. To gain insight into the functions of orthologous CAX2 genes in other plant species, cDNAs were amplified by RT-PCR from barley and tomato by using available cDNA (EST) sequence information. cDNA sequences with high sequence identity to AtCAX2 were identified by BLAST searches, and primers were designed to amplify predicted full-length cDNA sequences from barley and tomato tissues. Sequence comparison (Supplementary Figure S2) and phylogenetic analysis confirmed that these sequences were highly similar to AtCAX2 and were named HvCAX2 and LeCAX2. LeCAX2 appeared to be a full-length cDNA which encoded a protein of significant similarity to AtCAX2 and AtCAX5, including a long N-terminal tail (Supplementary Figure S2). In contrast, HvCAX2 encoded a protein with significant similarity to AtCAX2 and AtCAX5, but with a much shorter N-terminus. Further screening of a barley embryo cDNA library and the use of 5′-RACE obtained equivalent HvCAX2 cDNA clones, confirming that HvCAX2 was not missing any 5′ cDNA sequence and thus did possess a short N-terminus. Both HvCAX2 and LeCAX2 possess the Cys-Ala-Phe Mn2+ determinant motif [13], shared with AtCAX2 and AtCAX5, and an acidic motif region between transmembrane spans 6 and 7 (Supplementary Figure S2). These sequence characteristics are also shared with the rice orthologue OsCAX3. Two additional domains, named c-1 and c-2 (Supplementary Figure S2), which are predicted to function as substrate-selectivity filters [30], are highly conserved among these CAX2-like genes.

LeCAX2 was expressed at moderately high levels in the leaf and fruit of tomato plants (Supplementary Figure S1c). Similarly, expression of HvCAX2, as determined by qRT-PCR, showed high HvCAX2 expression in first leaves and roots of seedlings grown on artificial medium, and high expression in leaves, main shoot, immature spike and seeds, in mature soil-grown plants (Supplementary Figure S1d). This expression profile corresponds with that identified from microarray analysis (see details for Contig9559_at at In addition, microarray data indicate that HvCAX2 is expressed highly in the stamen and anther, and in the grain it is preferentially expressed in the embryo rather than the endosperm. To assess whether HvCAX2 expression is altered in response to metal stress, barley seedlings were treated with high concentrations of various metal salts and HvCAX2 expression was analysed in shoot tissue. No significant change was observed in response to Mn2+, Cd2+ or other heavy metals (Figure 5). In contrast, high Ca2+ (100 mM) and salt (100 mM NaCl) treatment caused a significant increase in HvCAX2 expression.

Figure 5 HvCAX2 expression in response to metal stress

Expression of HvCAX2 in barley seedlings in response to different metal stress conditions. RNA was isolated from whole plant tissue 12 h after treatment with 0.5×MS growth medium alone (basal) or supplemented with various metal salts (Ca, 100 mM CaCl2; Cd, 10 μM CdCl2; Co, 1 mM CoCl2; Cu, 0.1 mM CuCl2; Fe, 5 mM FeCl2; Mn, 2 mM MnCl2; Na, 100 mM NaCl; and Zn, 5 mM ZnCl2). Expression was determined by qRT-PCR relative to actin and is shown as a percentage of actin. Results are means±S.E.M. (n=3).

Cation transport activity of HvCAX2 and LeCAX2

HvCAX2 and LeCAX2 were expressed in the K667 mutant yeast strain to assess whether metal tolerance could be provided. An N-terminal truncated version of LeCAX2 (sLeCAX2) lacking the first 32 amino acids and with a M33E substitution for translation initiation was also expressed. K667 yeast expressing HvCAX2 were able to grow on high-Ca2+-containing medium (200 mM CaCl2); however, the growth efficiency was reduced compared with sAtCAX2 yeast (Figure 6a). Both full-length LeCAX2 and N-terminal truncated sLeCAX2 could efficiently suppress the Ca2+ hypersensitivity of K667 when grown on 200 mM CaCl2, equivalent to that provided by sAtCAX2. On high-Mn2+ medium (3.5 mM MnCl2), HvCAX2 could provide only relatively weak growth compared with sAtCAX2, although this was stronger than that mediated by full-length AtCAX2. This relatively inefficient Ca2+ and Mn2+ tolerance activity of HvCAX2 was not the result of reduced protein level in the yeast relative to sAtCAX2 (Figure 6b). LeCAX2 was unable to provide any Mn2+ tolerance to K667 yeast, even at low Mn2+ concentrations (1.5 mM MnCl2), but this inability was overcome when the N-terminus was removed, as sLeCAX2-expressing yeast could grow strongly on high Mn2+ (Figure 6a).

Figure 6 Suppression of Ca2+ and Mn2+ hypersensitivity of yeast by HvCAX2 and LeCAX2

(a) Suppression of Ca2+ and Mn2+ sensitivity of the pmc1 cnb vcx1 yeast mutant (K667) by full-length HvCAX2 and LeCAX2 or N-terminally truncated sLeCAX2 transporters. Suppression was compared with AtCAX2 and sAtCAX2. Saturated liquid cultures of K667 containing various plasmids were diluted to the cell densities (OD600) indicated and then spotted on to selection medium lacking histidine (−His) and YPD [1% (w/v) yeast extract/2% (w/v) peptone/2% (w/v) glucose] medium containing CaCl2 (Ca) or MnCl2 (Mn) concentrations as indicated. Yeast growth at 30 °C is shown after 3 days. A representative experiment is shown in each case. (b) Western blotting of AtCAX2–c-Myc, sAtCAX2–c-Myc and HvCAX2–c-Myc fusion proteins isolated from yeast with an anti-(c-Myc) antibody.

To confirm that the cation-tolerance phenotypes observed in yeast by HvCAX2 and LeCAX2 were the result of direct transport activity, proton-dependent Ca2+ and Mn2+ transport activity was measured in vacuolar membrane vesicles isolated from the yeast strains. In comparison with sAtCAX2, HvCAX2 Ca2+/H+ antiport activity was reduced, whereas LeCAX2 and sLeCAX2 activity was equivalent to that of sAtCAX2 (Figure 7a). Furthermore, Ca2+ transport kinetics for LeCAX2, sLeCAX2 and sAtCAX2 were all similar, but HvCAX2 had a slightly higher Km for Ca2+ (Table 1). Mn2+/H+ antiport activity by HvCAX2 was reduced compared with that of sAtCAX2 (Figure 7b). No Mn2+ transport activity could be measured for LeCAX2, whereas the activity of sLeCAX2 was equivalent to that of sAtCAX2.

Figure 7 Ca2+ and Mn2+ proton-dependent transport activity of HvCAX2 and LeCAX2

(a) Single 10 min time point measurements of Ca2+/H+ antiport activity determined as the ΔpH-dependent uptake of 10 μM 45Ca2+ into vacuolar membrane vesicles prepared from K667 yeast expressing sAtCAX2, HvCAX2, LeCAX2, sLeCAX2 and empty vector. Results are shown following subtraction of the protonophore (5 μM FCCP) background values. All experiments were performed following the establishment of a pH gradient in the presence of the Ca2+–ATPase inhibitor vanadate. Results are means±S.E.M. (n=3) (b) Mn2+/H+ antiport activity determined as Mn2+-dependent H+ transport. Acridine Orange fluorescence recovery after 15 min was measured in vacuolar membrane vesicles prepared from K667 yeast expressing sAtCAX2, HvCAX2, LeCAX2, sLeCAX2 and empty vector, following the establishment of a pH gradient and the addition of 200 μM MnCl2 at 0 min time point. Fluorescence recovery is shown as a percentage of fluorescence following the addition of Mn2+. Results are means±S.E.M. (n=3).


In the present study, we have analysed the AtCAX2 gene homologues in Arabidopsis, barley and tomato. Like AtCAX2, all three transporters were able to suppress the Ca2+ and Mn2+ tolerance phenotypes of a metal-hypersensitive yeast mutant as a result of Ca2+ and Mn2+/H+ antiport activity. Furthermore, they all provide metal tolerance by vacuolar sequestration of the cations. Our results therefore suggest that the type IB CAX proteins, including those from monocots and dicots, transport both Ca2+ and Mn2+, and this is likely to be a conserved trait throughout this group. Despite this overall similarity in function, we have identified variation in the transport kinetics of these transporters, and more notably, significant variation in the regulatory characteristics of these transporters.

Full-length AtCAX2 is not able to confer growth on high Mn2+ or Ca2+ concentrations when expressed in yeast compared with N-terminal truncated versions (Figure 3) [14,21], suggesting that the N-terminus regulates transport activity rather than cation selectivity. In the present study, we have demonstrated that AtCAX5 and LeCAX2 can similarly provide Mn2+ tolerance in yeast only when truncated at their N-termini (Figures 2 and 6). The N-terminus of AtCAX5 also appears to regulate Ca2+ transport, although this is not the case for LeCAX2 (see below). This inhibited transport activity of AtCAX2 and AtCAX5 was equivalent to the regulatory mechanism found previously for AtCAX1, whereby the N-terminal domain regulates Ca2+ transport activity as a result of a process of autoinhibition [10,18,19]. Autoregulatory domains have also been observed on plant and mammalian Ca2+–ATPases and H+–ATPases [31] and are a very effective means to regulate transport activity rapidly. This is important for Ca2+ transporters that play a role in Ca2+ signalling, as rapid activation and deactivation of the Ca2+ flux is central to the generation of Ca2+ oscillations [32]. The relevance of post-translational regulation of Mn2+ transport is unclear. Mn2+ is essential for many physiological plant functions, including water oxidation during photosynthesis, but it is unclear whether Mn2+ has a signalling role like that of Ca2+ [33]. Analysis of wild-type and deregulated AtCAX1 in tobacco [10] and Arabidopsis [19] has confirmed that the AtCAX1 regulatory mechanism is functional in planta. Further studies are needed to similarly confirm the regulatory mechanisms of type IB CAXs in plant cells.

The present study also suggests that this regulatory mechanism is not fully conserved for all plant CAX transporters. LeCAX2 has significant sequence similarity to AtCAX2 and AtCAX5, including a long N-terminal tail (Figure 8 and Supplementary Figure S2). It is therefore intriguing that LeCAX2 can efficiently transport Ca2+ at the equivalent level of activity as sLeCAX2 (Figures 6 and 7), although Mn2+ transport by LeCAX2 does appear to be constrained by the presence of the N-terminus. On the basis of AtCAX1 studies, the model of activation of transport activity is one where the N-terminally inhibited CAX transporter is activated following an interaction with an activator protein [20]. These activators appear to be plant specific, since none of the full-length Arabidopsis CAX transporters studied to date show high activity when expressed in yeast, unless they are mutated or an activator protein is co-expressed. It is possible that LeCAX2 can mediate Ca2+ transport in yeast as a result of activation by a yeast protein. Alternatively, there may be differences in the mechanism by which transport activity is regulated for this protein. Secondary-structure prediction of the N-terminus of selected CAX proteins supports this. Four type IA CAX proteins (AtCAX1, AtCAX3, VCAX1 and OsCAX1a), which show increased Ca2+ transport activity when N-terminally truncated [19,22,30], have similar N-terminal secondary structures with an equivalent coil-helix-coil structure predicted at the extreme N-terminus (Figure 8). AtCAX2 and AtCAX5 also have predicted coil-helix-coil regions close to the N-terminus, but the predicted secondary structure of LeCAX2 N-terminus is clearly different, indicating that structural characteristics could explain the apparent regulatory differences. Future mutagenesis analysis of this protein region should be able to discern whether this structure does explain the regulatory characteristics of LeCAX2.

Figure 8 Comparison of predicted secondary structure of the N-termini of CAX proteins

Secondary-structure prediction of N-terminal domains of plant type IA CAX proteins from Arabidopsis (AtCAX1 and AtCAX3), mung bean (VCAX1) and rice (OsCAX1a), and of plant type IB CAX proteins from Arabidopsis (AtCAX2 and AtCAX5), barley (HvCAX2), maize (ZCAX2), rice (OsCAX2 and OsCAX3) and tomato (LeCAX2). All the type IA proteins shown have been shown previously to be regulated by an N-terminal domain. The S. cerevisiae Ca2+/H+ antiporter ScVCX1 (a type IF CAX protein), which does not possess an N-terminal regulatory domain, is shown for comparison. Amino acid sequences up to the first predicted transmembrane span are shown. Secondary protein structure was predicted using PSIPRED [40]. A single line denotes a coil, a grey cylinder denotes a helix and a grey arrow denotes a strand.

HvCAX2 has a much shorter N-terminal tail compared with LeCAX2 and the Arabidopsis proteins, and is more analogous to the type IB gene OsCAX3 from rice cloned previously [7]. PCR analysis suggested that HvCAX2 is not a partial-length cDNA, but does indeed encode a protein with a short N-terminus. Full-length HvCAX2 could efficiently transport both Ca2+ and Mn2+, although activity was reduced compared with sAtCAX2 (Figure 7). Likewise, full-length OsCAX3 can provide tolerance to moderate concentrations of Ca2+ and Mn2+ in yeast [7]. It will be interesting to see if truncations of HvCAX2 and OsCAX3 results in increased transport activity. Alternatively, these transporters may be regulated by other means. The yeast vacuolar Ca2+/H+ antiporter VCX1 has a short N-terminus with predicted secondary protein structure similar to HvCAX2 and OsCAX3 (Figure 8). VCX1 is negatively regulated by the Ca2+-dependent protein phosphatase calcineurin, apparently at the post-transcription level [25]. Plants do not possess an orthologue of calcineurin, but these transporters could conceivably be regulated in a similar manner.

The significant HvCAX2 transcript induction by high concentration of Ca2+ and Na+ (Figure 5) may indicate a role of this transporter in the response to these metal stresses. HvCAX2 may provide Ca2+ tolerance by Ca2+ sequestration. Direct salt tolerance can be mediated predominantly by Na+ transporters encoded by Na+/H+ exchanger genes [4], whereas CAX genes may play a more indirect role. Adaptation and tolerance to salt stress in plants is controlled in part by Ca2+-signalling pathways. High-salt treatment leads to an increase in cytosolic Ca2+ levels that are then returned to low resting levels via the efflux of Ca2+ from the cytosol, e.g. by transport into the vacuole. Such a role has been proposed previously for the vacuolar Ca2+/H+ antiporter AtCAX3, which is induced by Na+ and is important for plant salt tolerance [11,34]. The lack of transcript induction by Mn2+ does not rule out HvCAX2 having a physiological function in Mn2+ transport. AtCAX2 is similarly not up-regulated by Mn2+, yet is important for vacuolar Mn2+ sequestration [14]. HvCAX2 was found to be expressed highly in young seedling tissue and also in seed tissue (Supplementary Figure S1d). The presence of HvCAX2 in barley seeds, particularly in the embryo, suggests that this transporter may be a determinant for Ca2+ and/or Mn2+ content in this cereal grain. Future analysis of HvCAX2 barley transgenic lines will allow this to be examined further.

The present study has provided the first characterization of AtCAX5. It has been shown previously by knockout analysis that AtCAX2 provides Mn2+ accumulation into the plant vacuole, yet vacuolar Mn2+ transport was not completely abolished in cax2 [14]. AtCAX5 is therefore a good candidate for providing this residual activity. sAtCAX5 had lower Ca2+ and Mn2+ transport activity than sAtCAX2 (Figure 3) with a reduced Ca2+ transport capacity (Table 1). Although the competition experiment suggests that sAtCAX2 may be able to transport Zn2+, there was no inhibition of sAtCAX5-mediated 45Ca2+ transport by either concentration of non-radioactive Zn2+ (Figure 4). Thus, despite the significant sequence similarity between AtCAX2 and AtCAX5, there appear to be differences in function. Another clear distinction between AtCAX2 and AtCAX5 was the transcriptional up-regulation of AtCAX5 in response to Mn2+ and Cd2+ (Figure 1). This is interesting, as very few plant genes have been identified previously with a significant Mn2+-induction profile, particularly Mn2+ transporters [33]. In addition, there was a slight increase in AtCAX5 transcript following Ca2+ depletion, and a reduction in response to excess Zn2+. The relevance of this is unclear, but future genetic analysis of AtCAX5 using mutant plant lines may uncover the physiological function of this cation transporter.

These genes may provide novel tools for future biotechnological manipulation of plants. Ca2+ is an essential nutrient, and manipulation of plant Ca2+ levels can have dietary advantages [35]. For example, manipulation of the sAtCAX1 Ca2+ transporter can be utilized to improve the Ca2+ nutrition of carrots by enhancing Ca2+ content [36]. Although sAtCAX1 was amenable for the manipulation of carrot, overexpression of sAtCAX1 in tomato induced various deleterious phenotypes as a result of high vacuolar Ca2+ accumulation [37]. Manipulation of endogenous transporters, such as LeCAX1, may be a means to overcome unwanted phenotypes in some crop species. Enhancement of other essential minerals can also improve crop plant traits. Although Mn2+ deficiency is not a problem in human diets, the manipulation of plant Mn2+ homoeostasis has applications for improved growth under non-optimal Mn2+ conditions [33]. Sequestration of toxic concentrations of Mn2+ can provide tolerance to plants grown on high-Mn2+-containing soils, such as acidic soils [2,12,29]. In addition, the ability of plants to accumulate Mn2+ efficiently can alleviate poor growth on Mn2+-deficient soils [38]. A higher Mn2+ content in seeds can improve plant germination and seedling establishment on Mn2+-poor soils, as has been demonstrated with barley seeds previously [39]. Seed-expressed HvCAX2 may therefore be a good candidate gene for improving seed Mn2+ content.

In summary, in the present study we have provided biochemical analysis of novel cation transporters from three plant species: Arabidopsis and two economically important crop plants. The present study demonstrated the ability of these proteins to transport Ca2+ and Mn2+ and has indicated differential transcriptional and post-translational regulation of transport activity for each transporter. Future characterization of these genes in planta will clarify these functional and regulatory characteristics. This comparative analysis shows that, although the study of Arabidopsis CAX genes can provide functional information, the variation in related genes from different species is significant.


This work was supported by a David Phillips Fellowship from the Biotechnology and Biological Sciences Research Council [grant number BB/B502152/1] (to J. K. P.). T. S. was supported by a National Science Foundation grant (grant number 0209777).


We thank Dr Kendal Hirschi (Department of Pediatrics, Baylor College of Medicine, Houston, TX, U.S.A.) in whose laboratory the initial cloning of AtCAX5 was performed, and for his comments on the manuscript. We thank Dr Wanda Waterworth and Dr Cliff Bray (Faculty of Life Sciences, University of Manchester, Manchester, U.K.) for providing barley seed and a barley cDNA library, and Thurston Heaton at the Firs Experimental Grounds (University of Manchester, Manchester, U.K.) for barley and tomato plant maintenance.

Abbreviations: carboxy-DCFDA, 5-(and 6-)carboxy-2′,7′-dichlorofluorescein diacetate; CAX, cation exchanger; CFP, cyan fluorescent protein; EST, expressed sequence tag; FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone; GUS, β-glucuronidase; MS, Murashige and Skoog; qRT-PCR, quantitative real-time PCR; RACE, rapid amplification of cDNA ends; RT-PCR, reverse transcription-PCR; UTR, untranslated region; YFP, yellow fluorescent protein


View Abstract