The Rice Cation/H+ Exchanger Family Involved in Cd Tolerance and Transport

Cadmium (Cd), a heavy metal toxic to humans, easily accumulates in rice grains. Rice with unacceptable Cd content has become a serious food safety problem in many rice production regions due to contaminations by industrialization and inappropriate waste management. The development of rice varieties with low grain Cd content is seen as an economic and long-term solution of this problem. The cation/H+ exchanger (CAX) family has been shown to play important roles in Cd uptake, transport and accumulation in plants. Here, we report the characterization of the rice CAX family. The six rice CAX genes all have homologous genes in Arabidopsis thaliana. Phylogenetic analysis identified two subfamilies with three rice and three Arabidopsis thaliana genes in both of them. All rice CAX genes have trans-member structures. OsCAX1a and OsCAX1c were localized in the vacuolar while OsCAX4 were localized in the plasma membrane in rice cell. The consequences of qRT-PCR analysis showed that all the six genes strongly expressed in the leaves under the different Cd treatments. Their expression in roots increased in a Cd dose-dependent manner. GUS staining assay showed that all the six rice CAX genes strongly expressed in roots, whereas OsCAX1c and OsCAX4 also strongly expressed in rice leaves. The yeast (Saccharomyces cerevisiae) cells expressing OsCAX1a, OsCAX1c and OsCAX4 grew better than those expressing the vector control on SD-Gal medium containing CdCl2. OsCAX1a and OsCAX1c enhanced while OsCAX4 reduced Cd accumulation in yeast. No auto-inhibition was found for all the rice CAX genes. Therefore, OsCAX1a, OsCAX1c and OsCAX4 are likely to involve in Cd uptake and translocation in rice, which need to be further validated.


Introduction
Cadmium (Cd) is one of the most toxic heavy metal elements in the environment and has an inhibitory effect on the photosynthesis rate, enzyme activity and ion absorption of plant [1][2][3][4]. Cd stress causes the overproduction of reactive oxygen species (ROS), resulting in oxidative stress and negatively affects the defense system of plants [5][6][7], which, in turn, affects the growth and development, and ultimately reduces the yield, nutritional quality and taste of rice [8]. Moreover, Cd can be easily absorbed by rice and accumulated in grains, which are the staple food for more half of the world population [9,10]. Rice with unacceptable Cd content has become a serious food safety problem in many rice this study have important theoretical significance and application value for rice genetics and breeding.

Bioinformatics Analyses of CAX Family Genes in Oryza sativa and Arabidopsis thaliana
Both Arabidopsis and rice have six CAX genes. All identified CAX members were named based on their phylogenetic relationships ( Figure 1). The two subfamilies (Type IA and Type IB) formed by the two main branches of the phylogenetic tree each have three genes from Arabidopsis and three genes from rice ( Figure 1). Chromosome mapping showed that the six rice CAX genes are distributed on chromosomes 1, 2 (two genes), 3, 4 and 5 ( Figure 1 and Table 1). The six CAX genes of Arabidopsis are distributed on chromosomes 1 (two genes), 2, 3 (two genes) and 5 ( Figure 1 and Table 1). The CDS regions of the rice CAX genes range in length from 1089 to 1362 bp and encoded proteins with lengths of 363-454 amino acid residues, molecular weights of 39.07- 49.11 KDa and pI values of 4.57-7.00 ( Table 1). The length of CDS, encoded proteins with lengths, molecular weights and pI values of the Arabidopsis CAX genes range in length from 1326 to 1428 bp, 441-475 amino acid residues, 48.10-51. 84 KDa and pI values of 4.45-6.51, respectively (Table 1). AtCAX1, with a relatively long length, appears to be distinct from the other five members of the Subfamily IA proteins (Table 1). OsCAX4, with a relatively short length, appears to be divergent from the other five members of the Subfamily IB proteins (Table 1). Gene structures were different in each of the two main subfamilies, as illustrated in Figure 1. Subfamily IA members have 8-11 exons and 7-10 introns. Subfamily IB members have 10-12 exons and 9-11 introns. The four members of the IB subfamily (AtCAX2, AtCAX5, AtCAX6 and OsCAX4) all have 12 exons and 11 introns. In particular, OsCAX4 was classified into Type IB phylogenetically and has no UTR ( Figure 1). The rice CAX proteins have 8-11 putative transmembrane domains (TMDs) ( Table 1 and Figure  S1). Except for AtCAX3, which contains 11 TDMs, all other Arabidopsis CAX proteins have 10 putative TDMs (Table 1 and Figure S1). yeast strains BY4741. In addition, the functions of the CAX family genes in ion transport in multiple species were summarized to assist in studying the roles of the rice CAX genes in other metals and ions. The results of this study have important theoretical significance and application value for rice genetics and breeding.

Bioinformatics Analyses of CAX Family Genes in Oryza sativa and Arabidopsis thaliana
Both Arabidopsis and rice have six CAX genes. All identified CAX members were named based on their phylogenetic relationships ( Figure 1). The two subfamilies (Type IA and Type IB) formed by the two main branches of the phylogenetic tree each have three genes from Arabidopsis and three genes from rice ( Figure 1). Chromosome mapping showed that the six rice CAX genes are distributed on chromosomes 1, 2 (two genes), 3, 4 and 5 ( Figure 1 and Table 1). The six CAX genes of Arabidopsis are distributed on chromosomes 1 (two genes), 2, 3 (two genes) and 5 ( Figure 1 and Table 1). The CDS regions of the rice CAX genes range in length from 1089 to 1362 bp and encoded proteins with lengths of 363-454 amino acid residues, molecular weights of 39.07- 49.11 KDa and pI values of 4.57-7.00 ( Table 1). The length of CDS, encoded proteins with lengths, molecular weights and pI values of the Arabidopsis CAX genes range in length from 1326 to 1428 bp, 441-475 amino acid residues, 48.10-51.84 KDa and pI values of 4.45-6.51, respectively (Table 1). AtCAX1, with a relatively long length, appears to be distinct from the other five members of the Subfamily IA proteins (Table 1). OsCAX4, with a relatively short length, appears to be divergent from the other five members of the Subfamily IB proteins (Table 1). Gene structures were different in each of the two main subfamilies, as illustrated in Figure 1. Subfamily IA members have 8-11 exons and 7-10 introns. Subfamily IB members have 10-12 exons and 9-11 introns. The four members of the IB subfamily (AtCAX2, AtCAX5, AtCAX6 and OsCAX4) all have 12 exons and 11 introns. In particular, OsCAX4 was classified into Type IB phylogenetically and has no UTR ( Figure 1). The rice CAX proteins have 8-11 putative transmembrane domains (TMDs) ( Table 1 and Figure S1). Except for AtCAX3, which contains 11 TDMs, all other Arabidopsis CAX proteins have 10 putative TDMs (Table 1 and Figure S1).  The phylogenetic tree was constructed based on sequence alignment of CAX homologs from Arabidopsis (At) and rice (Os), using the neighbor-joining method with bootstrapping analysis implemented in MEGA 7.0. The CAX proteins are clustered into two groups (Type IA and Type IB). Gene structures were drawn, using Gene Structure Display Server 2.0 with genomic sequences and CDS sequences. Introns and exons are represented by black lines and yellow boxes, respectively. Blue boxes represent UTRs. In BLASTP analysis, the identity between the CAX genes is high with a minimum value of 54%, indicating that they are relatively conservative ( Figure S2a). The Subfamily IB members have a higher degree of similarities, reaching 71% ( Figure S2b), in comparison to the Subfamily IA members, which have their highest identity as 64% ( Figure S2c). Sequence identity in aligned regions ranges from 31 to 88% outside of selfhits (Table 2), with the highest percentage of identity being between AtCAX5 and AtCAX6. The lowest identity is between OsCAX1c and OsCAX4 (Table 2). These results indicated that the CAX gene family has high homology in Oryza sativa and Arabidopsis thaliana.  OsCAX1a  100  OsCAX1b  69  100  OsCAX1c  59  58  100  OsCAX2  50  52  43  100  OsCAX3  49  51  40  74  100  OsCAX4  36  38  31  55  55  100  AtCAX1  64  62  57  48  49  38  100  AtCAX2  46  49  44  72  72  54  46  100  AtCAX3  67  66  56  50  48  39  79  48  100  AtCAX4  56  59  49  46  47  34  57  47  59  100  AtCAX5  48  49  44  72  73  52  48  87  49  47  100  AtCAX6  47  45  38  66  65  53  46  82  47 45 88

Response of Rice CAX Family Genes to Cd Stress
The CAX genes of Arabidopsis and a few other species to Cd stress have been widely reported [58,65]. We analyzed the expression profiles of all the six rice CAX family genes, using qRT-PCR, under different CdCl 2 treatments. The transcription levels of all the six genes were significantly upregulated in roots under the 30 and 100 µM CdCl 2 treatments, and the expressions were increased with increasing Cd concentrations ( Figure 2A). The highest expression was observed for OsCAX1a in roots treated with 100 µM CdCl 2 ( Figure 2A). Under 10 µM CdCl 2 treatment, the expression of OsCAX1b, OsCAX1c and OsCAX4 in roots was similar to the untreated control ( Figure 2A). The transcription levels in leaves were also significantly induced by all Cd treatments ( Figure 2B). OsCAX1b and OsCAX2 showed relatively high expression in leaves under 100 µM CdCl 2 , with OsCAX1b being the highest among the six genes, while the other four CAX genes were highly expressed in leaves when treated at 30 µM CdCl 2 ( Figure 2B). These results suggested that the rice CAX genes could respond to Cd stress.
in roots was similar to the untreated control ( Figure 2A). The transcription levels in leaves were also significantly induced by all Cd treatments ( Figure 2B). OsCAX1b and OsCAX2 showed relatively high expression in leaves under 100 μM CdCl2, with OsCAX1b being the highest among the six genes, while the other four CAX genes were highly expressed in leaves when treated at 30 μM CdCl2 ( Figure 2B). These results suggested that the rice CAX genes could respond to Cd stress.

Functional Analysis of Rice CAX Genes in Yeast
Since the phenomena of auto-inhibition at the N-terminus have been reported for CAX genes in multiple species [58,[66][67][68][69], the full-length rice CAX genes and their 15 Nterminal truncated versions were included in our functional test by using heterologous yeast assay. To create the truncated version without affecting the transmembrane structure, we removed several amino acids from the N-terminal ( Figure S4). The truncated versions used were three ΔOsCAX1a (OsCAX1a-12AA, OsCAX1a-28AA and OsCAX1a-40AA), three ΔOsCAX1b (OsCAX1b-11AA, OsCAX1b-22AA and OsCAX1b-33AA), three ΔOsCAX1c (OsCAX1c-14AA, OsCAX1c-24AA and OsCAX1c-37AA), two ΔOsCAX2 (Os-CAX2-16AA and OsCAX2-40AA), three ΔOsCAX3 (OsCAX3-10AA, OsCAX3-19AA and OsCAX3-40AA) and one ΔOsCAX4 (OsCAX4-29AA). Then they were expressed in wildtype yeast strain BY4741(MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0). On a SD-Gal medium without Cd, there were no differences in growth between the control carrying an empty vector (vector control) and the CAX-carrying yeast strains ( Figure 3A). For the truncated versions of genes, the yeast cells expressing OsCAX1a, OsCAX1c, OsCAX4 had better growth than the vector control on SD-Gal medium containing 160 μmol/L CdCl2, while the other three genes showed no significant difference with the vector control on the SD-Gal medium containing 40 or 160 μmol/L CdCl2 ( Figure 3A). Consequently, we did not find N-terminal auto-inhibition of the rice CAX genes.
Further experiments using liquid media with different Cd concentrations were conducted to confirm their Cd transport activity. Without Cd, all CAX genes expressing yeast cells showed similar growth to the control. However, the growth of BY4741 was significantly promoted by OsCAX1a (∆OsCAX1a), OsCAX1c (∆OsCAX1c) and OsCAX4 (∆OsCAX4) when 10, 20 or 40 µmol/L CdCl 2 was supplemented ( Figure 3B-D and Figure  S3). In addition, N-terminal auto-inhibition was still not observed in experiments using liquid medium either ( Figure S3). Under the 24 h Cd exposure, OsCAX1a and OsCAX1c enhanced Cd accumulation, while OsCAX4 reduced Cd accumulation ( Figure 3E). These results indicated that OsCAX1a, OsCAX1c and OsCAX4 had Cd transport activity in yeasts and might be involved in Cd uptake, transport and accumulation in rice. Further experiments using liquid media with different Cd concentrations were conducted to confirm their Cd transport activity. Without Cd, all CAX genes expressing yeast cells showed similar growth to the control. However, the growth of BY4741 was significantly promoted by OsCAX1a (ΔOsCAX1a), OsCAX1c (ΔOsCAX1c) and OsCAX4 (ΔOsCAX4) when 10, 20 or 40 μmol/L CdCl2 was supplemented ( Figure 3B-D and Figure  S3). In addition, N-terminal auto-inhibition was still not observed in experiments using liquid medium either ( Figure S3). Under the 24 h Cd exposure, OsCAX1a and OsCAX1c enhanced Cd accumulation, while OsCAX4 reduced Cd accumulation ( Figure 3E). These results indicated that OsCAX1a, OsCAX1c and OsCAX4 had Cd transport activity in yeasts and might be involved in Cd uptake, transport and accumulation in rice.

Expression Pattern and Subcellular Localization of the Rice CAX Genes
To investigate the expression patterns of rice CAX genes, the expression in tissues of seedlings, booting, flowering and grain filling stages grown under normal field conditions were measured by qRT-PCR. We found that OsCAX1c was specifically expressed in leaves, with very limited expression elsewhere, while other genes were detected in all tissues of different growth stages ( Figure 4A). It was obvious that OsCAX1c, OsCAX3 and OsCAX4 were strongly expressed in leaves ( Figure 4A). Both OsCAX1a and OsCAX2 were The absorbance at 600 nm (OD600) of cell cultures was measured every 5 h. (E) Cd accumulation in the wild-type BY4741 (gray) and BY4741 expressing OsCAX1a (green), OsCAX1C (red) and OsCAX4 (orange) treated with 5 µM CdCl 2 for 24 h. Statistical comparison was performed by one-side t-test (* p < 0.05 and ** p < 0.01). Error bar indicates standard deviation, and the data are presented as means ± SD (n = 3).

Expression Pattern and Subcellular Localization of the Rice CAX Genes
To investigate the expression patterns of rice CAX genes, the expression in tissues of seedlings, booting, flowering and grain filling stages grown under normal field conditions were measured by qRT-PCR. We found that OsCAX1c was specifically expressed in leaves, with very limited expression elsewhere, while other genes were detected in all tissues of different growth stages ( Figure 4A). It was obvious that OsCAX1c, OsCAX3 and OsCAX4 were strongly expressed in leaves ( Figure 4A). Both OsCAX1a and OsCAX2 were expressed at a high level in flowering spikelet but weakly in leaves. The expression of OsCAX1b in roots and leaves was slightly higher than in other tissues ( Figure 4A). expressed at a high level in flowering spikelet but weakly in leaves. The expression of OsCAX1b in roots and leaves was slightly higher than in other tissues ( Figure 4A). To further investigate their tissue-specific expression, we developed stable transgenic rice lines expressing the GUS reporter gene driven by the OsCAX1a, OsCAX1b, OsCAX1c, OsCAX2, OsCAX3 or OsCAX4 (−2000 bp from the start codon) promoter. Transgenic seedlings were cultured in deionized water for 14 days, and then various tissues were collected and stained for GUS detection ( Figure 4B). All the six genes were observed in roots including the root tip (0.2 cm from the root tip), root hair (2 to 3 cm from the root tip) and root lateral branching zones. OsCAX3 and OsCAX4 expressed abundantly throughout the primary and lateral roots. OsCAX3 strongly expressed throughout the stele in primary roots, whereas other genes weakly expressed or hardly expressed at the base of the stele ( Figure 4B). All genes, except OsCAX1b, expressed in the primary root tip. In the rootshoot junction, OsCAX1a, OsCAX1b and OsCAX3 strongly expressed, with OsCAX3 showing slightly higher expression than OsCAX1a and OsCAX1b. OsCAX1a and OsCAX2 hardly expressed in leaves; OsCAX1c, OsCAX3 and OsCAX4 were more strongly expressed in leaves than OsCAX1b ( Figure 4B). These results were consistent with the (qRT-qPCR) results shown in Figure 4A.
To determine the subcellular localization of OsCAX1a, OsCAX1c and OsCAX4 proteins in plant cells, we constructed GFP fusion protein and transformed it into rice-sheath protoplasts. OsCAX1a, OsCAX1c and OsCAX4 fused with GFP was observed at the periphery protoplasts ( Figure 5). Moreover, OsCAX1a and OsCAX1c were co-localized with the tonoplast membrane marker AtTPK, suggesting that both OsCAX1a and OsCAX1c were localized in the vacuolar membrane ( Figure 5). The merge images of OsCAX4-GFP and FM TM 4-64 indicated the plasma membrane subcellular localization of OsCAX4 (Figure 5). Vacuolar localization of OsCAX1a has been reported previously by Kamiya et al. (2005) [63]. The membrane localizations of OsCAX1a, OsCAX1c and OsCAX4 were consistent with their cation/H + exchanger functions, which suggested that the rice CAX genes might play important roles in Cd regulation. To further investigate their tissue-specific expression, we developed stable transgenic rice lines expressing the GUS reporter gene driven by the OsCAX1a, OsCAX1b, OsCAX1c, OsCAX2, OsCAX3 or OsCAX4 (−2000 bp from the start codon) promoter. Transgenic seedlings were cultured in deionized water for 14 days, and then various tissues were collected and stained for GUS detection ( Figure 4B). All the six genes were observed in roots including the root tip (0.2 cm from the root tip), root hair (2 to 3 cm from the root tip) and root lateral branching zones. OsCAX3 and OsCAX4 expressed abundantly throughout the primary and lateral roots. OsCAX3 strongly expressed throughout the stele in primary roots, whereas other genes weakly expressed or hardly expressed at the base of the stele ( Figure 4B). All genes, except OsCAX1b, expressed in the primary root tip. In the root-shoot junction, OsCAX1a, OsCAX1b and OsCAX3 strongly expressed, with OsCAX3 showing slightly higher expression than OsCAX1a and OsCAX1b. OsCAX1a and OsCAX2 hardly expressed in leaves; OsCAX1c, OsCAX3 and OsCAX4 were more strongly expressed in leaves than OsCAX1b ( Figure 4B). These results were consistent with the (qRT-qPCR) results shown in Figure 4A.
To determine the subcellular localization of OsCAX1a, OsCAX1c and OsCAX4 proteins in plant cells, we constructed GFP fusion protein and transformed it into rice-sheath protoplasts. OsCAX1a, OsCAX1c and OsCAX4 fused with GFP was observed at the periphery protoplasts ( Figure 5). Moreover, OsCAX1a and OsCAX1c were co-localized with the tonoplast membrane marker AtTPK, suggesting that both OsCAX1a and OsCAX1c were localized in the vacuolar membrane ( Figure 5). The merge images of OsCAX4-GFP and FM TM 4-64 indicated the plasma membrane subcellular localization of OsCAX4 ( Figure 5). Vacuolar localization of OsCAX1a has been reported previously by Kamiya et al. (2005) [63]. The membrane localizations of OsCAX1a, OsCAX1c and OsCAX4 were consistent with their cation/H + exchanger functions, which suggested that the rice CAX genes might play important roles in Cd regulation.

The CAX Family Genes in Rice Have High Identity to Their Homologues in Arabidopsis Thaliana
It is well-known that the homeostasis of intracellular Ca 2+ plays an important role in signal transduction of stresses and is necessary for normal cell growth [44]. Hence, control of Ca 2+ concentration is critical to cellular function. The Ca 2+ /cation antiporter (CaCA) superfamily members, including MHX (Mg 2+ /H + exchanger), CCX (cation/Ca 2+ exchanger) and CAX, are the most important Ca 2+ transporters. The CaCA exchangers family structure is highly conserved in related plants [70]. Here, we explored the structural characteristics of CAX transporters through phylogenetic tree analysis, transmembrane structure and homology identification of CAX family proteins in rice and Arabidopsis. It was found that the two subfamilies formed by the two main branches of the phylogenetic tree each have three Arabidopsis genes and three rice genes (Figure 1), and the length of the CDS region and the encoded proteins, the molecular weight and the range of pI values were all relatively close in rice and Arabidopsis ( Table 1). The gene structures of the two main subfamilies are slightly different. The members of Subfamily IA have 8-11 exons and 7-10 introns. The members of Subfamily IB have 10-12 exons and 9-11 introns, which were one more than those of the Subfamily IA members (Figure 1). CAX family proteins have 8-11 putative transmembrane domains ( Figure S1). It is worth noting that the identity between

The CAX Family Genes in Rice Have High Identity to Their Homologues in Arabidopsis Thaliana
It is well-known that the homeostasis of intracellular Ca 2+ plays an important role in signal transduction of stresses and is necessary for normal cell growth [44]. Hence, control of Ca 2+ concentration is critical to cellular function. The Ca 2+ /cation antiporter (CaCA) superfamily members, including MHX (Mg 2+ /H + exchanger), CCX (cation/Ca 2+ exchanger) and CAX, are the most important Ca 2+ transporters. The CaCA exchangers family structure is highly conserved in related plants [70]. Here, we explored the structural characteristics of CAX transporters through phylogenetic tree analysis, transmembrane structure and homology identification of CAX family proteins in rice and Arabidopsis. It was found that the two subfamilies formed by the two main branches of the phylogenetic tree each have three Arabidopsis genes and three rice genes (Figure 1), and the length of the CDS region and the encoded proteins, the molecular weight and the range of pI values were all relatively close in rice and Arabidopsis ( Table 1). The gene structures of the two main subfamilies are slightly different. The members of Subfamily IA have 8-11 exons and 7-10 introns. The members of Subfamily IB have 10-12 exons and 9-11 introns, which were one more than those of the Subfamily IA members (Figure 1). CAX family proteins have 8-11 putative transmembrane domains ( Figure S1). It is worth noting that the identity between CAX genes is high, the sequence identity in the aligned region ranges from 31 to 88% outside of selfhits (Table 2) and the homology among all CAX genes reached 54% (Mate Figure S2). Taken together, the CAX family genes in rice have high identity to their homologues in Arabidopsis thaliana, indicating that functional similarities are expectable, which is advantageous in leveraging the relatively more abundant information accumulated in Arabidopsis to speed up studies in rice.

N-Terminal Auto-Inhibition Was Not Found for the Rice CAX Genes Tested Using Cd Treatment
N-terminal auto-inhibition has been widely reported for CAX genes in different species [58,[67][68][69], and it is most widely studied in Arabidopsis [66,71,72]. For instance, we found that deleting 36 amino acids at the N-terminus of AtCAX1 ORF improved Ca and Mn tolerance compared to the full-length AtCAX1 ORF by heterologous transformation of yeast [71,72]. The modified AtCAX2 (sCAX2A) with a domain of AtCAX2 and without the N-terminal autoinhibitory domain elevated Ca accumulation in the fruits of tomato. Mn transport in yeast was controlled by the N-terminus of tomato LeCAX2 [58]. The absence of N-terminal regulatory region (NRR) of the cotton (Gossypium hirsutum) GhCAX3 altered cold tolerance compared to full-length of GhCAX3 [68]. It has also been reported in rice that the N-terminus of CAX family genes (Type IA) regulates Ca tolerance and transport in yeast. OsCAX1a∆27 and OsCAX1c∆47 strongly enhanced the Ca tolerance of yeast, and OsCAX1b was more sensitive to 100 mM CaCl 2 than OsCAX1b∆36 was [63]. We transformed the yeast strain BY4741 with all members of the rice CAX genes and their corresponding truncated versions with the removal of the 10-40 amino acids from the N-terminal. The Cd transport activity in yeast did not show any differences, indicating that the N-terminal auto-inhibition was not shown for Cd in rice. However, it is best not to interpret this as an indication that the N-terminal auto-inhibition is not present in the rice CAX genes, since we only tested for Cd. Kamiya et al. (2005) [63] reported that OsCAX2 and the N-terminal truncation mutants of OsCAX2 (OsCAX2∆26) expressed in yeast did not show Ca tolerance. The modified Arabidopsis CAX2 (sCAX2A) also showed almost no changes in Mn, Cu and Fe accumulation in the fruits, compared to the overexpression of AtCAX2 in tomato [73]. More detailed studies are needed to (dis)validate this observation.

Rice CAX Genes, Particularly OsCAX1a, OsCAX1c and OsCAX4, Might Be Important in Controlling Cd Uptake and Translocation in Rice
Existing studies have shown that the CAX genes in plants are involved in ion tolerance, transport and homeostasis, as shown in Figure 6 and Table S3. However, the response of rice CAX genes to Cd has not been studied. Our results showed that the expression of all the six rice CAX genes were upregulated by high-concentration CdCl 2 treatments. This was particularly true in roots, for which gene expression increased with increasing Cd concentration ( Figure 2). By transforming the BY4741 yeast strain, we found that OsCAX1a, OsCAX1c and OsCAX4 had better Cd tolerance than the control on a SD-Gal medium containing 160 µmol/L CdCl 2 and more active in liquid media containing Cd ( Figure 3A-D). OsCAX1a and OsCAX1c increased, while OsCAX4 decreased, the accumulation of Cd in yeast ( Figure 3E). Thus, OsCAX1a, OsCAX1c and OsCAX4 involve in Cd tolerance and transport in yeast, implying that they might involve in Cd tolerance and accumulation in rice as well. As discussed above, the rice CAX family genes have high identity to their homologous genes in Arabidopsis, indicating that their functions may be conserved. Three Arabidopsis CAX genes, namely AtCAX1 [48,49], AtCAX2 and AtCAX4 [50][51][52], were verified by transgenic methods to participate in Cd tolerance, transport and accumulation ( Figure 6 and Table S3). OsCAX1a has very high homology with AtCAX1 and AtCAX4, reaching 64% and 56%, respectively ( Figure 1 and Table 2). OsCAX1c has high homology with AtCAX1 (57%) and AtCAX4 (49%) (Figure 1 and Table 2). OsCAX1a strongly expressed in spikelet and nodes, while OsCAX1c very strongly expressed in leaves ( Figure 4A). Both of OsCAX1a and OsCAX1c are most likely to locate on the vacuolar membrane. OsCAX4 on plasma membrane has the highest identity (54%) to AtCAX2 (Figure 1 and Table 2) and is strongly expressed in roots and grains. Taken together, we speculate that OsCAX1a, Os-CAX1c and OsCAX4 may play different roles in the key processes, including the absorption by the roots, the xylem loading, the distribution and transportation via the nodes, and the redistribution in the leaf via phloem. These processes collectively determine the plant tolerance and accumulation of Cd. leaves ( Figure 4A). Both of OsCAX1a and OsCAX1c are most likely to locate on the vacuolar membrane. OsCAX4 on plasma membrane has the highest identity (54%) to AtCAX2 (Figure 1 and Table 2) and is strongly expressed in roots and grains. Taken together, we speculate that OsCAX1a, OsCAX1c and OsCAX4 may play different roles in the key processes, including the absorption by the roots, the xylem loading, the distribution and transportation via the nodes, and the redistribution in the leaf via phloem. These processes collectively determine the plant tolerance and accumulation of Cd. Although OsCAX1b, OsCAX2 and OsCAX3 did not show Cd transport activity in our yeast assay (Figure 3), their roles in Cd regulation could not be completely ruled out. Firstly, their expression levels in roots and leaves were all induced by Cd treatments (Figure 2). Secondly, OsCAX1b and OsCAX3 were previously reported to involve in tolerance to Ca, and OsCAX3 also confers Mn tolerance in yeast [63]. Thirdly, CAX proteins function as Ca 2+ /H + and/or heavy metal/H + exchangers [56,73]. The importance of Ca 2+ acts as a secondary messenger in the signal transduction of various biotic and abiotic stimuli in plants has been well documented [74][75][76]. Changes in apoplastic pH or Ca 2+ concentrations can be expected to strongly affect Cd binding capacity, since Cd 2+ and Ca 2+ have similar ionic radii [77]. Fourthly, CAX genes interact with other genes to form regulatory network controlling tolerance to metals. Yeast two-hybrid analysis showed that AtCAX1 directly interacts with AtSOS2, which is a serine/threonine kinase whose function is essential for salt tolerance in Arabidopsis [78]. Co-expression of AtSOS2 specifically activated AtCAX1 and integrated Ca transport in yeast and salt tolerance in Arabidopsis [79]. Therefore, it is likely that CAX genes interact with key genes for Cd absorption and accumulation, or are regulated by transcription factors or phosphorylated by receptor kinases to affect Cd transport. Fifthly, since CAX genes have high degree of genetic identity, it is likely that some of the genes may have functional redundancy (Table 2). Thus, any test/assay using a single gene, as we performed, may fail to identify the function of a gene. Indeed, it has been reported that the AtCAX1 and AtCAX3 function as dimers, and co-expressing both AtCAX1 and Although OsCAX1b, OsCAX2 and OsCAX3 did not show Cd transport activity in our yeast assay (Figure 3), their roles in Cd regulation could not be completely ruled out. Firstly, their expression levels in roots and leaves were all induced by Cd treatments (Figure 2). Secondly, OsCAX1b and OsCAX3 were previously reported to involve in tolerance to Ca, and OsCAX3 also confers Mn tolerance in yeast [63]. Thirdly, CAX proteins function as Ca 2+ /H + and/or heavy metal/H + exchangers [56,73]. The importance of Ca 2+ acts as a secondary messenger in the signal transduction of various biotic and abiotic stimuli in plants has been well documented [74][75][76]. Changes in apoplastic pH or Ca 2+ concentrations can be expected to strongly affect Cd binding capacity, since Cd 2+ and Ca 2+ have similar ionic radii [77]. Fourthly, CAX genes interact with other genes to form regulatory network controlling tolerance to metals. Yeast two-hybrid analysis showed that AtCAX1 directly interacts with AtSOS2, which is a serine/threonine kinase whose function is essential for salt tolerance in Arabidopsis [78]. Co-expression of AtSOS2 specifically activated AtCAX1 and integrated Ca transport in yeast and salt tolerance in Arabidopsis [79]. Therefore, it is likely that CAX genes interact with key genes for Cd absorption and accumulation, or are regulated by transcription factors or phosphorylated by receptor kinases to affect Cd transport. Fifthly, since CAX genes have high degree of genetic identity, it is likely that some of the genes may have functional redundancy (Table 2). Thus, any test/assay using a single gene, as we performed, may fail to identify the function of a gene. Indeed, it has been reported that the AtCAX1 and AtCAX3 function as dimers, and co-expressing both AtCAX1 and AtCAX3 mediated Ca, lithium (Li) and salt tolerance in yeast [61,80], and affected element distribution and abundant Arabidopsis seeds [81].
In conclusion, we investigated all rice CAX genes and demonstrated that OsCAX1a, OsCAX1c and OsCAX4 were associated with Cd tolerance and Cd transport in yeast. They are also highly likely to involve in Cd tolerance, absorption and accumulation in rice; however, further validation through transgenic methods in rice is needed.

Plant Materials and Growth Conditions
Rice (Oryza sativa L. cv. Nipponbare) seeds were surface-sterilized with 10% (v/v) hydrogen peroxide solution for 30 min, thoroughly rinsed, washed six times with deionized water and germinated for 48 h under dark conditions and a temperature of 30 • C [82]. The dew-white and uniformly growing seeds were sown in a 96-well polymerase chain reaction (PCR) plate (8 × 12) with perforated wells at the bottom to facilitate the roots to fully contact with the nutrient solution [83]. The seedlings that grew consistently were moved to a rectangle 3.5 L box containing 2 L of normal nutrient solution, with a 3 × 8 foam board being used as bed, and grown for a week. Then the seedlings were grown under treatment conditions of 0, 10, 30 or 100 µM CdCl 2 . The nutrient solution was changed every two days and the night before sampling. Culture was conducted in a greenhouse, under natural light, at day/night temperatures of 30 • C/22 • C and 60% relative humidity. After a 2-week treatment, root and leave samples were rapidly taken, frozen in liquid nitrogen and stored in a refrigerator, at −80 • C, for RNA extraction and real-time quantitative reverse-transcription PCR (qRT-PCR) analysis.

RNA Extraction and Real-Time PCR
Total RNA was extracted by using the TRIzol reagent (Vazyme Biotech Co. Ltd., Nanjing, China). DNaseI-treated total RNAs were subjected to reverse transcription (RT) with the HiScript II Q Select RT SuperMix for qPCR (+gDNA wiper) kit (Vazyme Biotech Co. Ltd., Nanjing, China). Transcript levels of selected genes were measured by qRT-PCR, using the CFX96 ® Real-Time PCR System using the 2 × T5 Fast qPCR Mix (SYBRGreenI) kit (Vazyme Biotech Co. Ltd., Nanjing, China). Rice actin gene (Rac1) was used for normalization. The ∆∆Ct method was used to calculate the relative transcript abundance [78]. The primers for qRT-PCR are given in Table S1.

Functional Analysis of Rice CAX Genes in Yeast
The heterologous yeast assay was utilized to identify the Cd transport ability of rice CAX proteins by transforming the BY4741 yeast strain (MATa his3∆1 leu2∆0 met15∆0 ura3∆0). All CAX genes were amplified from full-length cDNA clones, using the PCR primers lying downstream from the start codon and just upstream of the stop codon (Table S2). The protocol for removing some amino acids from the N-terminus of the CAX transporters was described in detail by Shigaki et al. (2010) [72]. Using BamH I and EcoR I, we ligated each rice CAX gene into the pYES2 vector (Invitrogen), with correct direction, using a ClonExpress ® Ultra One Step Cloning Kit (Vazyme Biotech Co. Ltd., Nanjing, China), resulting in the pYES2-OsCAXs construct. The vectors truncated were verified by Sanger sequencing. Empty vector pYES2 and the CAX gene vectors were then transformed into BY4741 yeast cells, respectively, according to the manufacturer's protocol (Yeast Transformation Kit; Beijing Coolaber Technology Co. Ltd., Beijing, China). Transformants were selected on synthetic dextrose medium without uracil (SD-Ura) and verified by PCR with yeast plasmid extraction (Yeast Plasmid Extraction Kit; Beijing Solarbio Technology Co. Ltd., Beijing, China). Positive clones were cultured on an SD-Ura liquid medium until the early logarithmic phase. For plate growth tests, yeast transformants were diluted to an OD600 of 1.0, 0.1, 0.01 and 0.001, step by step, with sterile water; 6 µL of the cell suspension was spotted on SD-Ura plates containing 0, 40 or 160 µmol/L CdCl 2 , respectively. The plates were incubated at 30 • C for 2 days before the growth phenotypes were evaluated.
To quantify the growth of BY4741 yeast strain transformed with various plasmids in liquid SD-Ura media containing CdCl 2 overnight yeast cells were prepared and the optical density (OD) at 600 nm was adjusted to 0.8 with sterile distilled water. Then 20 µL of cell suspensions was added to 20 mL liquid SD-Ura media containing 0, 10, 20 or 40 µmol/L CdCl 2 . The OD values at 600 nm were determined at the indicated time.
For the Cd concentration determination, the yeast transformants were cultured overnight with a liquid SD-Ura medium with 2% galactose, at 30 • C and 200 rpm, until the OD600 reached 0.8. Then the yeast transformants were treated with 5 µM CdCl 2 for 24 h. The yeast strain carrying empty vector was used as control. The cultures of each set were harvested in pre-weighed microfuge tubes by centrifugation and washed with sterile water for three times. After aspirating the supernatant, pelleted cells were dried in the oven, overnight, at 60 • C. The samples were wet-digested by using 5 mL concentrated HNO 3 +HClO 4 (4 + 1), at 120 • C, in a heating block for 30 min. After cooling, the digested solution was diluted to 15 mL with deionized water. The Cd concentrations were determined by inductively coupled plasma-mass spectrometry (ICP-MS). All the assays were performed at least three times.

Tissue Expression Assay
For histochemical analysis of GUS activity, the upstream 2.0 kb genomic fragment of each gene was cloned into pCAMBIA1300 to generate OsCAXs promoter : GUS vector. The vectors were then transformed into Nipponbare to produce transgenetic plants. The GUS staining Solution (1 mg/mL) from (Beijing Coolaber Technology Co., Ltd., Beijing, China) was used to determine the activity of GUS. The samples were incubated at 37 • C for 24 h, and the solution containing (X-gluc dry powder + X-gluc solution): GUS staining buffer = 1:50. After staining, green tissue materials (such as leaves) were treated 2 or 3 times by 75% ethanol to remove chlorophyll and decolorize until the negative control material was white. GUS activity was detected by a stereoscopic fluorescence microscopy (TL5000; Leica Microsystems).

Subcellular-Localization Assay
Subcellular localization was investigated by transiently expressing GFP-OsCAXs fusion into rice protoplasts. The ORF of OsCAX1a, OsCAX1c and OsCAX4 were amplified by PCR from rice ('Nipponbare') root cDNA, using primers (Table S2). Protoplasts generated from the young stems of 3-week-old rice seedlings grown under light were transformed with transient expression plasmids, according to Bart et al. (2006) [84]. GFP fluorescence signals were detected with a TCS SP5 confocal laser scanning microscope (Leica Microsystems) at 500-535 nm, after excitation at 488 nm, while FM TM 4-64 (plasma membrane specific localization dye) was excited at 543 nm and scanned at 600-630 nm. Double staining using AtTPK (red signal) as the tonoplast membrane marker was used for further confirmation of the subcellular localization. For co-localization experiments, sequential scanning was done for both of the channels and then merged together to show overlapping signals. All the images were further processed by using Leica LAS AF Lite software.