Isolation and Characterization of Maize PMP3 Genes Involved in Salt Stress Tolerance

Plasma membrane protein 3 (PMP3), a class of small hydrophobic polypeptides with high sequence similarity, is responsible for salt, drought, cold, and abscisic acid. These small hydrophobic ploypeptides play important roles in maintenance of ion homeostasis. In this study, eight ZmPMP3 genes were cloned from maize and responsive to salt, drought, cold and abscisic acid. The eight ZmPMP3s were membrane proteins and their sequences in trans-membrane regions were highly conserved. Phylogenetic analysis showed that they were categorized into three groups. All members of group II were responsive to ABA. Functional complementation showed that with the exception of ZmPMP3-6, all were capable of maintaining membrane potential, which in turn allows for regulation of intracellular ion homeostasis. This process was independent of the presence of Ca2+. Lastly, over-expression of ZmPMP3-1 enhanced growth of transgenic Arabidopsis under salt condition. Through expression analysis of deduced downstream genes in transgenic plants, expression levels of three ion transporter genes and four important antioxidant genes in ROS scavenging system were increased significantly in transgenic plants during salt stress. This tolerance was likely achieved through diminishing oxidative stress due to the possibility of ZmPMP3-1's involvement in regulation of ion homeostasis, and suggests that the modulation of these conserved small hydrophobic polypeptides could be an effective way to improve salt tolerance in plants.


Introduction
Plant growth and development are affected by various abiotic stresses, such as high salinity, drought, low temperature and heavy metals. In particular, salinity is a global environmental challenge, affecting crop production over 800 million hectares, or a quarter to one third of all agricultural land on the earth [1]. Recently, physiological and genetic mechanisms of salt tolerance have been intensively investigated, and it is believed that high concentration of salts often causes ion imbalance and hyperosmotic stress to plants [2,3].
Ion homeostasis is fundamental to physiological processes of living cells. The living cells often maintain high concentration of K + and low concentration of Na + in the cytosol, which is important for activities of many cytosolic enzymes. However, under salt stress, Na + accumulates extremely in cells and consequently disrupts ion homeostasis. Thus, the maintenance of Na + and K + homeostasis is crucial under salt stress for plants to survive.
There are two kinds of mechanisms for plants to maintain ion homeostasis under salt stress. Firstly, at the organismal level, the ability of regulating Na + uptake and transporting Na + from roots to the shoots is critical in all plants. In saline soil, plant roots are inclined to minimize Na + accumulation in plants. For instance, sodium influx of halophyte roots is much lower than non-halophyte roots because the width of the Casparion band is two or three times larger in halophytes than in non-halophytes, which effectively prevent the excessive Na + from entering into the apoplastic space [4]. When the Na + ion enters the apoplastic space of roots, the excessive Na + is restricted to old tissues to prevent Na + from accumulating in reproductive and delicate organs, which can lead to irreversible damage. Secondly, at the cellular level, the maintenance of appropriate Na + accumulation in cells is due to diffusion and active transport. Recently, some factors responsible for ion transport, such as nonselective cation channels (NSCCs), ion transporters and membrane-potential modulators, have been characterized. In plants, the NSCCs, which catalyze ion influx, can be divided into three groups according to their physical stimuli. These are respectively cyclic-nucleotide-gated NSCCs (CNGSs), amino-acidgated NSCCs (AAG-NSCCs) and reactive-oxygen-species-activated NSCCs (ROS-NSCCs) [5]. Among these NSCCs, CNGSs are perhaps the best studied. Known examples include AtCNGS3, which localizes in root epidermal and cortical cells and contributes to Na + uptake at the initial stage of salt stress [6,7]. Other examples include AAS-NSCCs and ROS-NSCCs, which support the role of Ca 2+ transporter [8]. Previous studies also revealed that there are several ion transporters that play important roles in retrieving intracellular ion homeostasis under saline conditions. As a plasma membrane Na + /H + -antiporter, SOS1 is an important tolerance determinant involved in the exclusion of sodium ions from cells [9,10]. The transcription level of SOS1 is up-regulated by salt stress but not by drought and cold stress [11]. In the presence of calcium, SOS3 activates the substrate phosphorylation activity of SOS2 [12], and then the SOS3/SOS2 complex in turn activates SOS1 probably via phosphorylation, which catalyzes sodium efflux from plant cells. Additionally, several other transporters are also involved in sodium ion transport, such as AtNHX1 and AtNHX2 [13].
Plasma membrane protein 3 (PMP3), a class of small molecular weight hydrophobic proteins in higher plants, responds to various stresses, such as low temperature, salt and dehydration. [14][15][16]. All these proteins are highly conserved at both sequential and structural levels and contain the common conservative domain of UPF0057. The PMP3 protein (Pmp3p) was also identified in yeast (Saccharomyces cerevisiae), which functions in the maintenance of plasma membrane potential. Deletion of PMP3 increased Na + and K + sensitivity of yeast cells and resulted in excessive concentration of Na + and K + [16]. Complementation analysis in yeast revealed that plant homologues, such as RCI2 in Arabidopsis thaliana, Lti6a/b in Oryza sativa, AcPMP3-1 in sheep grass (Aneurolepidium chinense, a monocotyledonous halophyte), PutPMP3-1 and PutPMP3-2 in alkali grass (Puccinellia tenuiflora), are capable of restoring salt sensitivity of yeast mutant lacking PMP3/SNA1 [16][17][18][19][20][21][22]. It was also demonstrated that AcPMP3 regulated cellular Na + and K + accumulation in the Dnha1Dpmr2Dpmp3 yeast mutant under salt stress [22]. In Arabidopsis, eight PMP3 homologues (AtRCI2A-H) have been identified [23]. Over-expression of AtRCI2A in Arabidopsis restricted Na + uptake into cells and resultantly enhanced salt resistance. On the contrary, the disruption of AtRCI2A led to over accumulation of Na + and increased Na + sensitivity of Arabidopsis mutants [24,25]. These reports indicate that PMP3 homologues participate in maintaining intracellular ion homeostasis. However, due to the small size of these polypeptides, they are likely not ion transporters [22]. While it is generally agreed that Pmp3p are involved in cation uptake indirectly, the precise mechanism is currently unknown.
In maize (Zea mays L.), important genes that participate in regulation of ion homeostasis, such as ZmCIPK16, ZmCBL4 and ZmNHXs, have been identified [26][27][28]. ZmNHX genes are likely to play a role in ion homeostasis regulation, as they are induced by salt stress in roots or stems [28]. Additionally, increased activities of Na + /H + -antiporter and H-pump enhance salt tolerance of maize seedlings [29]. Although it is known that PMP3 genes likely play an important role in ion homeostasis regulation, there was no systemic collection and characterization of PMP3 genes in maize. In this study, we systematically investigated functions of eight ZmPMP3 genes. The result showed that these genes could complement the function of the yeast Pmp3p deletion mutant under salt stress. Over-expression of ZmPMP3-1 enhanced salt tolerance of transgenic Arabidopsis plants. The up-regulations of ion transporters genes in transgenic plants, such as NHX1, AVP1 and AHA2, are probably suggestive of the regulation of ion homeostasis by ZmPMP3-1.

Results
Isolation and sequence analysis of ZmPMP3 family genes To identify salt-or drought-stresses responsive genes, a cDNA library was constructed from salt-, drought-, and alkali-treated maize inbred line YQ7-96. An up-regulated EST (EC858820.1) homologous to PMP3 was selected. After rapid amplification of cDNA 39 end (39RACE), a full-length sequence of ZmPMP3-1 (EU364580) encoding a putative hydrophobic protein of 58 amino acids was obtained. There are seven other members of this gene family in maize according to the submitted sequences in GenBank, namely from ZmPMP3-2 to ZmPMP3-8 (Table 1). Open reading frames (ORFs) of the eight genes ranged from 165 bp to 228 bp and encoded peptides of 54 to 75 amino acids. Comparison among genomic and coding sequences revealed that the ZmPMP3 genes, except ZmPMP3-1, harbored two exons separated by an intron of variable size. The ZmPMP3s encoded by these genes exhibited identities with each other from 42.17% to 95.38%. Particularly, the UPF0057 domain (LIVF-x-STAC-LIVF (3)-P-PF-LIVA-GAV-IV-x (4)-GKN) was conserved among these putative proteins of ZmPMP3 genes ( Figure 1A). The results of the TMHMM 2.0 analysis showed that these putative proteins contained two transmembrane domains composed of alpha-helices, being linked by a putative turn structure. In addition, the orientations of N-and C-termini of the members were different ( Figure S1) [30]. Comparison of the amino acid sequences of these eight genes revealed that the membrane spanning regions were highly conserved and the difference in protein lengths was caused by variable lengths of the C-terminal ( Figure 1A).
Homologues of the ZmPMP3s were also present in many other plants species, such as Arabidopsis, Oryza sativa L., Aneurolepidium chinense, and Puccinellia tenuiflora. The maize ZmPMP3s showed sequential similarity with rice Oslti6b (AY607690) from 36% to 87.93%.
With the completion of the B73 maize genome sequencing, the ZmPMP3 genes were located on chromosomes by blasting in MaizeGDB database (http://www.maizegdb.org/). The eight genes were located on eight different chromosomes but chromosome 4 and 5 did not contain any ZmPMP3 gene (Table 1).

Subcellular localization of ZmPMP3s
To investigate the localization of ZmPMP3s, the full-length coding regions without terminating codons of the ZmPMP3 genes were cloned into the pBI221-GFP plasmid [31]. The ZmPMP3s-GFP fusion proteins were transiently expressed in onion epidermal cells and the subcellular localizations of these fusion proteins were visualized by confocal microscopy. As showed in Figure 2, GFP protein was found in the nucleus, cytoplasm, and cell membrane (Figure 2A, S2C). All of eight ZmPMP3-GFP fusion proteins showed GFP signals in the peripheral layer of cells, a location of which may correspond to either cell wall or plasma membrane. To further examine the location of ZmPMP3s, the epidermal cells expressing ZmPMP3-GFP fusion proteins were plasmolyzed by 30% sucrose solution. In these cells, the fluorescence was still detected in regions corresponding to the distorted plasma membrane ( Figure 2B, C, S2A, and S2B), suggesting that the eight proteins were membrane proteins.

Responses to various abiotic stresses
The eight ZmPMP3 genes displayed different expression patterns under abiotic stresses ( Figure 3). Under salt treatment, ZmPMP3-1 and ZmPMP3-5 were up-regulated in shoot and root; ZmPMP3-6 was only up-regulated in root and ZmPMP3-8 was only up-regulated in shoot; ZmPMP3-4 and ZmPMP3-7 transcripts decreased slightly at 0.5 h or 1 h and then increased significantly at 9 h.
The ZmPMP3-7 transcript in root accumulated to maximal at 24 h followed by a reduction at 48 h. The transcript levels of ZmPMP3-4 showed a minor peak at 1 h, and the accumulation increased significantly at 24 h. The ZmPMP3-5 in root was repressed at 5 h followed by an increase with its peak at 24 h.
There were only four genes responsive to cold stress. The ZmPMP3-1 and ZmPMP3-2 transcripts in shoot and root were accumulated gradually under cold stress, while ZmPMP3-4 and ZmPMP3-6 were induced rapidly at 0.5 h or 1 h and then fell.
RT-PCR revealed that ZmPMP3-1, ZmPMP3-5 and ZmPMP3-6 were early responsive to these stresses. Under salt and drought conditions, the stronger inductions of ZmPMP3-5 and ZmPMP3-6 were observed only in root. While the expression of ZmPMP3-3 was unaffected by those stresses.
ZmPMP3-2, ZmPMP3-4, ZmPMP3-5 and ZmPMP3-6 were responsive to exogenous ABA treatment but with different patterns. ZmPMP3-2 and ZmPMP3-6 were up-regulated under ABA treatment. The transcript levels of ZmPMP3-2 and ZmPMP3-6 increased to the maximum level at 5 h and 1 h, respectively, under ABA treatment. ZmPMP3-5 was down-regulated. The transcript level of ZmPMP3-4 was dependent on time-point under ABA treatment since it decreased after 1 h and began to increase to the maximum level after 2 h. These results suggested that the four genes were regulated by ABA signal.  2), Aneurolepidium chinense AcPMP3-1/2 (BAD34658.1, BAD34659.1), Puccinellia tenuiflora PutPMP3-1/2 (BAG54793.1, BAG54794.1), Hordeum vulgare Blt101 (CAA80984.1). The amino acid identity in dark grey is larger than 75%, and the amino acid identity in light grey is larger than 50%. Straight arrows marked predicted transmembrane domains. The curved arrow indicated a putative loop structure. B) Phylogenetic analysis based on putative amino acid sequences of PMP3-related proteins in rice, Arabidopsis, Hordeum vulgare, Aneurolepidium chinense, and Puccinellia tenuiflora. The analysis was performed by using the MEGA 4.0 program with neighbor joining method and with 1000 replicates. Numbers on the figure 1B were bootstrap values. doi:10.1371/journal.pone.0031101.g001

Expression patterns in different tissues
To investigate spatial expression patterns of the ZmPMP3 genes in maize, different tissues including leaves, roots, ears, tassels and stalks were selected and detached from CN165 at V12 stage due to yield reduction caused by crop sensitivity to abiotic stresses [32]. Silks, as an important female tissue, are sensitive to abiotic stresses. Thus, silks at VT stage were also included in this experiment. The results showed that the transcript levels of the eight ZmPMP3 genes were all at their peaks in the young organs and tissues. The transcript levels of ZmPMP3-1, ZmPMP3-2, ZmPMP3-3, ZmPMP3-5, ZmPMP3-7 and ZmPMP3-8 were detected at the highest levels in young stalks, and were detected at very low levels in leaves or roots ( Figure 4). However, the ZmPMP3-7 transcripts were detected at very low level in ears and silks. In addition, the . Expression patterns of the eight ZmPMP3 genes in CN165 under various stresses. One-week old shoots and roots of maize seedlings, treated by 200 mM NaCl, 4uC, 30% PEG-6000, and 100 mM ABA, were harvested respectively. The qRT-PCR was performed to determine the relative expression levels of the eight genes under various treatment. The maize GAPDH gene was selected as internal control. Data from qRT-PCR experiments were analyzed according to the 2 2DDCT method. Data represents the average of three independent experiments 6SE. when the relative expression levels go up to 2 fold or down to 0.5 fold, the difference of the expression is regarded as significant (P value,0.05). doi:10.1371/journal.pone.0031101.g003 transcripts of ZmPMP3-4 and ZmPMP3-6 were detected at the highest levels in tassels and silks, respectively. This result indicated that the activities of the different ZmPMP3 isoforms were confined to specific organs or tissues. Therefore, it was most likely that they functioned in different organs under normal or stress conditions.

Functional complementation of ZmPMP3s in yeast mutant
The PMP3 gene of S. cerevisiae encodes a homologue of the maize ZmPMP3s. Deletion of PMP3 in yeast causes the hyperpolarization of cell membrane and consequently disrupted ion homeostasis. The corresponding mutant strain is sensitive to the toxic Na + and hygromycin B [16]. Thus, function analysis of ZmPMP3s could be conducted by investigating the capacity of the ZmPMP3 genes to complement the deletion of PMP3 in yeast. The ZmPMP3 genes with ADH1 promoter were strongly expressed in the YR93-31 (YR93-1 Dpmp3) cells (the strains ZmPMP3-1 to ZmPMP3-8 genes over-expressed were named as YR93-31M1 to YR93-31M8, respectively). YR93-1 (MATa ade2 his3-D200 leu2-3,112 lys2-D201 ura3-52 gal2 Dpmr2-2::HIS3 Dnha1::LEU2) was used as a control.
Both 10-fold dilution dot experiment and dose-response curves demonstrated that the expression of the ZmPMP3 genes could suppress the sensitivity of mutant strains to the Na + and hygromycin B ( Figure 5). As shown in Figure 5A and B, the salt sensitivity of YR93-31M1-5 and YR93-31M7-8 were restored. But the relative growth rate of YR93-31M6 was 18%, which was similar to YR93-31's. The salt sensitivity of YR93-31M7 was partly restored. Under 5 mg/L hygromycin B ( Figure 5C) conditions, the sensitivity of the yeast YR93-31Ms to hygromycin B was reversed to the levels which were the same as that of the passive control YR93-1 under hygromycin B condition. YR93-31M6 also showed similar hygromycin B tolerance to YR93-31 strain. These results suggested that the ZmPMP3s abolished the hyperpolarization of cell membrane resulting from PMP3 deletion and consequently maintained intracellular ion homeostasis.
Salt tolerance of yeast was reported to be affected by Ca 2+ [9]. Therefore, we investigated the salt tolerance of YR93-31M1-8 under high and low calcium conditions ( Figure 5D). On low calcium (YPDA+5 mM EGTA+25 mM NaCl), the salt sensitivity of all mutant strains YR93-31, YR93-1 and YR93-31M1-8 were exacerbated. However, the expression of the ZmPMP3 genes still stimulated the growth of mutant strain YR93-31. On high calcium (YPDA+10 mM CaCl 2 +25 mM NaCl), the salt sensitivity of YR93-31 strain was reversed, the result being consistent with previous reports [16,22]. Thus, under either high or low Ca 2+ conditions, the expression of the ZmPMP3 genes decreased the sensitivity of mutant trains YR93-31 to salt stress. These results suggested that although over-expression of the ZmPMP3 genes in yeast enhanced the growth of yeast deletion mutant under salt stress, this process was not dependent on the presence of Ca 2+ .
Salt tolerance of ZmPMP3-1 overexpressed transgenic Arabidopsis ZmPMP3-1 was selected to investigate whether the ZmPMP3 genes could improve salt tolerance of plants because it was responsive to various abiotic stresses and encoded protein showed higher identity to AtRCI2 genes. The full-length ORF of ZmPMP3-  Figure 6E).
Furthermore, the fresh weights of two week old seedlings sowed on MS media containing 150 mM and 175 mM NaCl were monitored ( Figure 7A, B). The fresh weights of L6, and L11 lines were significantly higher than the fresh weight of the WT. To investigate the phenotype of roots, the transgenic and the WT seedlings were transferred to the MS and 200 mM NaCl plates ( Figure 8). Under normal condition, there was no difference observed between the transgenic and the WT seedlings. However, under salt stress condition, the numbers of lateral roots of the transgenic lines were significantly larger than the WT. These results clearly revealed that overexpression of ZmPMP3-1 improved salt tolerance of transgenic Arabidopsis plants and eliminated growth retardation caused by salt stress.
Plants at silique stage were irrigated by 300 mM NaCl solution to investigate salt tolerance of mature plants. After one week, only ,18% of the wild type plants survived, while 22%-42% of the transgenic plants survived. Under normal condition, the transgenic and the WT plants all grew well ( Figure 9A, C). As it shown in Figure 9B, most of the transgenic lines continued to grow well at Photosynthetic rate and CMS of ZmPMP3-1 overexpressed transgenic Arabidopsis During the period of increased salt stress, plants often exhibit stress symptom, including increased leaf senescence and cellular damage due to the photo-oxidative stress [33][34][35][36][37]. The chlorophyll florescence and CMS (cell membrane stability) as important physiological probes were determined to investigate effects of salt stress on plants. To determine the chlorophyll florescence, the 4week old Arabidopsis plants were treated for one week by 300 mM NaCl and then Fv and Fm values were measured to calculate the maximum florescence ratios (Fv/Fm), which represents the photosynthetic rate of plants. The results showed that the Fv/ Fm values of the transgenic plants were significantly higher than the WT, indicating that the chlorophyll of the WT was damaged more easily by salt stress ( Figure 10A). The difference of the CMS between the transgenic and the WT plants was observed until the fifth day, and the CMS of the transgenic lines was significantly higher than the WT ( Figure 10B). Taken above, these results indicated that the transgenic lines could be less damaged than the WT under high salinity environment.
Expression patterns of stress-and ion transporter-related genes in ZmPMP3-1 over-expressed transgenic plants To understand the molecular mechanism underlying the enhanced salt tolerance by 35S-ZmPMP3-1, we tested transcript levels of a set of selected abiotic stress-responsive genes and antioxidants genes in the transgenic and WT plants. Because transgenic yeasts reestablish ion homeostasis under high salt conditions, several genes modulating ion homeostasis were also selected ( Table 2, S1). Under salt conditions, several stressresponsive genes including the stress-response gene DREB2A and its target gene RD29B were up-regulated more highly in the transgenic plants than in the WT. Expression of an ABA-response gene RAB18 was also elevated ( Table 2). The ion transporterrelated genes, including Na + /H + -antiporter gene NHX1, and H-ATPase genes AVP1, and AHA2, were up-regulated in the transgenic plants than in the WT. The up-regulation of these genes in the transgenic plants suggested that over-expressing ZmPMP3-1 affected the expression of the ion transport-related genes. In addition, several antioxidant genes, such as P5CS1-2, CAT1, CCS, and CSD1-2, exhibited up-regulation in the transgenic plants (Table 2). Under normal conditions, the transcript levels of all genes were similar between the transgenic and WT plants.
The expression of ZmPMP3 homologues, such as Blt101 in barely, wpi6 in wheat, Oslti6a and Oslti6b in rice, AcPMP3-1 and AcPMP3-2 in sheep grass, accumulated under salt, drought, cold, and ABA conditions [15,21,39], indicating their roles in abiotic stresses tolerance. In this study, ZmPMP3s were mainly responsive to salt and drought stresses. Only four genes were responsive to cold stress. Under salt and drought stresses, the inductions or repressions of the ZmPMP3 genes were dependent on stress duration. The rapid and transient rise of the transcripts of ZmPMP3-1, ZmPMP3-5 and ZmPMP3-6 suggested that these ZmPMP3s mediate an early protection against sudden abiotic stresses before an overall and permanent adaption. The stronger inductions of ZmPMP3-5 and ZmPMP3-6 were observed only in root, indicating their organ specific expression in response to abiotic stresses at seedling stage. The differences of expression patterns among the eight genes under abiotic stresses implied that although the genes are conserved structurally, their regulation in response to abiotic stresses is perhaps altered due to specific adaptations to environmental conditions.
Under abiotic stresses, plants may produce hormones such as ABA which in turn causes stomatal closure and induces expression of stress-related genes. Some genes are induced by extracellular ABA treatment, whereas others are not [40]. In this study, four of the eight members of these ZmPMP3 genes, including ZmPMP3-2, ZmPMP3-4, ZmPMP3-5 and ZmPMP3-6, were regulated by ABA, and the others were not. This result showed that these genes were differently responsive to ABA, suggesting that they were involved in different signaling pathways. The ZmPMP3-2, ZmPMP3-4, ZmPMP3-5 and ZmPMP3-6 may be involved in ABA-dependent signal pathway, and the others were not.
Additionally, the transcript levels of ZmPMP3s showed different degrees in various organs. The transcripts of ZmPMP3-1, ZmPMP3-2, ZmPMP3-3, ZmPMP3-5, ZmPMP3-7 and ZmPMP3-8 were accumulated mainly in stalks, and moderately in tassels. ZmPMP3-4 and ZmPMP3-6 were accumulated mainly in tassels and silks, respectively. These results indicated that in spite of their high sequence similarity and conserved subcellular localization, the eight ZmPMP3s did not have the exactly same function.

ZmPMP3s have a common function of maintaining ion homeostasis and this process is not dependent on the presence of Ca 2+
The Pmp3p in yeast, a homologue of ZmPMP3s, contributes to the regulation of intercellular ion homeostasis through controlling the plasma membrane potential, which prevents excessive build-up of the Na + influx [16]. When PMP3 was absent, YR93-31 showed membrane hyperpolarization, and which in turn lead to high sensitivity to toxic ions, such as Na + and hygromycin B [16]. Hygromycin B is a lipophilic toxic cation, which is taken up according to the membrane potential [41]. Recent research found that PMP3-type proteins, including AtRCI2A, Oslti6a/b, AcPMP3-1, and PutPMP3-1/2, have the capability of complementing the function of PMP3 lacking mutant yeast strain and restore salt sensitivity of this lose-function strain [16,22]. It is suggested that these PMP3-type proteins were most likely to be involved in regulation of membrane potential and consequently regulated the ion homeostasis under ionic stress condition. The complementation analysis in this study revealed that ZmPMP3s, except ZmPMP3-6, could restore salt and hygromycin B sensitivity of YR93-31 strain. It is suggested that ZmPMP3s have a common function of regulating membrane potential, consequently maintaining intracellular ion homeostasis.
The calcium ion (Ca 2+ ) is firmly established as a ubiquitous intracellular second messenger in plants, which acts not only the regulator of intercellular osmotic potential, but also as an important signal factor in stress signal transduction pathway in plants, such as Ca 2+ -CaM signal pathway [42], Ca 2+ -CDPK signal pathway [43], and SOS system [44]. All of these pathways are linked to the downstream cell response to abiotic stresses. It is reported that the yeast has a similar stress-related Ca 2+ -signaling pathway to plants [9]. Our research showed that the salt sensitivity phenotype of YR93-31 was relieved in the presence of Ca 2+ . It is possible that Ca 2+ activates a Ca 2+ -and calmodulin-dependent protein phosphatase (CaN), which promotes the expression of stress-related genes, and then confers high salt tolerance to yeasts [45]. Alternatively, there is a class of non-selected ion channel like NSC1 (non-selected ion channel 1) that is sensitive to Ca 2+ , which could inhibit Na + uptake in the presence of millimolar Ca 2+ [46]. However, the function of ZmPMP3s seems to be independent on the presence of Ca 2+ . Under either high calcium or low calcium conditions, YR93-31Ms expressing ZmPMP3s still showed lower salt sensitivity than YR93-31 strain. Therefore, it suggested that the process of ZmPMP3s regulating membrane potential was independent on the presence of Ca 2+ .
Physiological and molecular mechanisms of ZmPMP3-1 which enhanced salt tolerance Salt stress has three different effects: it reduces water potential, causes ion imbalance and toxicity, and then results in a clear stunting of plants [47]. In this study, ZmPMP3-1 over-expression in Arabidopsis relieved growth inhibition caused by salt stress, as evident from green cotyledon rate, biomass and root morphology of the transgenic plans at seedling stage.
Physiological indices, including Fv/Fm and CMS [36,37], are typical physiological parameters for evaluating abiotic stresses tolerance and resistance in crop plants. The maximum efficiency of PSII photochemistry, measured as Fv/Fm, is a direct reflection of the PSII activity, and environmental stresses are associated with decreased Fv/Fm ratio. In this study, the transgenic plants under salt condition exhibited lower decline of photosynthetic rate (Fv/ Fm). Thus ZmPMP3-1 plants might have more robust photosynthetic capabilities than the WT. CMS determinations were consistent with Fv/Fm results and suggested that the transgenic plants had enhanced salt-stress tolerance.
Ion cytotoxicity is caused by replacement of K + by Na + in biochemical reactions and conformal changes and loss of function of proteins as Na + interfere with noncovalent interaction between their amino acids. Metabolic imbalance caused by ionic toxicity may also lead to oxidative stress [44]. Plants possess a number of antioxidants that protect against the potentially cytotoxic species of activated oxygen, such as superoxide dismutase (SOD), catalase (CAT), guaicol peroxidase (POD), and so on. The plants with higher activation of antioxidants have been reported to show higher salt tolerance [48][49][50]. In this study, the transcripts of P5CS1-2, CAT1, CCS, and CSD1-2 were more highly accumulated in the transgenic plants than in the WT. The CDS, CAT and CCS genes encode superoxide dismutase (SOD), catalase (CAT), and copper chaperone for SOD1, respectively, which are important antioxidants in ROS scavenging system. P5CS1-2 genes encode delta-1-pyrroline-5-carboxylate synthase, which plays a key role in biosynthesis of proline. Transgenic plants overexpressing these genes always show higher tolerance of salt stress than nontransformed control [51,52]. Thus the enhanced salt tolerance of transgenic plants might be attributed to the up-regulation of these antioxidants genes.
Theoretically, three mechanisms can be explained for ion homeostasis in plant symplast. They are respectively, (i) preventing the Na + entry from cell membrane; (ii) compartmenting the excessive Na + into vacuoles; and (iii) exporting Na + back to external medium [53]. In Arabidopsis thaliana, the function of NHX1 and AVP1 genes in maintenance ion homeostasis were well established [54][55][56][57]. AtNHX1 has the activity of Na + /H +antiporter, it might compartmentalize the excessive Na + into vacuole [56]. AtAVP1 encodes P-ATPase, which participates indirectly in the maintenance of ion homeostasis by generating the H + gradient [57]. Brini et al. found that co-expression of wheat NHX1 and AVP1 could improve salt tolerance of transgenic Arabidopsis by compartmentalizing excessive Na + into vacuole [58]. In this study, the accumulation of AtNHX1and AtAVP1 transcripts was significantly higher in the transgenic lines than in WT under salt conditions. The yeast functional complementation showed that ZmPMP3s had the function of controlling membrane potential. In fact, previous research on guard cells by voltage-clamp demonstrated that the main transporters in membrane were responsive to membrane voltage [59]. Thus, it is speculated that ZmPMP3-1 affects the expression of some ion transporter genes, including NHX1 and AVP1, by regulating membrane potential, and consequently affect intracellular ion homeostasis. In further research, more efforts should be given to the deciphering the signals and molecular mechanism of ZmPMP3 in regulating ion homeostasis.
The present study identified eight ZmPMP3 genes in maize. Possibly, these genes share a common function of cell membrane depolarization, and consequently maintain intercellular ion homeostasis. The functions of these genes were similar and related, but not exactly the same. These eight genes were responsive to abiotic stresses differently, and accumulated mainly in different organs. The investigation of molecular mechanism in ZmPMP3-1 over-expressed transgenic plants could aid in the elucidation of the regulatory mechanism of PMP3 in response to abiotic stresses and maintenance of ion homeostasis in plants.

Materials and Methods
Plant material, grown conditions, and stress treatment A maize inbred line ''CN165'' was used in this study, which is the female parent of commercial maize hybrid ''Zhongyu 4'' and have strong stress tolerance under poor soil condition during the vegetative and reproductive stages.
For salt and drought stress, the maize seedlings at V3 stage in green house (28uC, humidity 40-50%, ,150 mmol/m 22 s 21 under a 12-h light/12-h darkness photocycle) were transferred into 200 mM NaCl and 30% PEG-6000 solution and then harvested at given time points. For cold stress, the maize seedlings at V3 stage were transferred into a 4uC chamber and harvested at given time points. For ABA treatment, maize seedlings at V3 stage were sprayed with 100 mM ABA solution. Shoots and roots which were sampled respectively from the seedlings, were frozen immediately with liquid nitrogen, and stored in 280uC for RNA isolation and real-time quantitative PCR (qRT-PCR) analysis.
To investigate the spatial expression patterns of the ZmPMP3s in maize, different tissues were selected and detached from CN165 at V12 stage, including leaf, root, ear, tassel and stalk. In addition, the silk at VT stage was included in this analysis. The maize plants were well irrigated (volumetric water content 22%-40%, TDR300 Soil Moisture Meter produced by Spectrum Technologies. Inc) during sampling and five individual plants were used in tissue sampling. Tissues and organs were frozen immediately with liquid nitrogen, and stored in 280uC for RNA isolation and qRT-PCR analysis.
Arabidopsis thaliana ecotype Columbia (Col) was used as WT and phenotypic assays of 35S-ZmPMP3-1 plants in all experiments. Isolation and bioinformatic analysis of sequences of ZmPMP3 genes Based on an EST homologous to the plasma membrane protein 3 (Pmp3p) which was obtained from a previously constructed cDNA library [60], the specific full-length primers were designed as follows: F, 59-AGCGAAAGGAGAGAAGGAATC-39 and R, 59-GATGGGGTGGGTACGGTAG-39. The full-length cDNA fragment was obtained in CN165 by rapid amplification of cDNA 39 end. The conditions for amplification were 95uC for 5 min, then 30 cycles at 95uC for 30 s, 58uC for 30 s, and 72uC for 1 min, followed by 72uC for 10 min. The cDNA full-length fragment was purified, sequenced, and named as ZmPMP3-1.
To explore other ZmPMP3-related genes in maize, blastx and blastn were conducted by comparing the sequences of ZmPMP3-1 and homologous genes such as AtRCI2a, AtRCI2b, Oslti6a, Oslti6b, AcPMP3-1, AcPMP3-2, Blt101, PutPMP3-1, and PutPMP3-2, with the entries in NCBI GenBank (http://www.ncbi.nlm.nih.gov). Seven putative protein sequences with E-value#10 26 and their corresponding full-length cDNA sequences were obtained. The specific primers were designed based on the seven corresponding cDNA sequences (Table S2). Full-length cDNA fragments were amplified from salt, cold and drought-induce first-strand cDNA by RT-PCR (Dataset S1). Amplification and sequencing of the fulllength cDNA fragments were repeated three times. The calculation of the putative protein sequences was carried out by NCBI ORF finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html) (Dataset S2) and the conserved domains were found by blast in the PROSITE database (http://expasy.org/prosite/) [61]. The prediction of the transmembrane helices of these putative protein sequences was executed by the TMHMM 2.0 tool on line (http:// www.cbs.dtu.dk/services/TMHMM/). The ProtScale tool (http://web.expasy.org/protscale/) was used to analyze hydrophobic domains ( Figure S1) [62]. Phylogenetic tree based on the resultant alignments was constructed by using the neighbor-joining method with the MEGA 4 program [63,64]. The genomic sequences of ZmPMP3 genes were PCR-amplified from genomic DNA of the inbred CN165 using specific primers (Table S2 & Dataset S3).

Subcellular localization of ZmPMP3s-GFP fusion proteins
The ORF (open reading frame) sequences without terminating codons of the ZmPMP3 genes were cloned into XbaI/ScaII sites of pBI221-GFP vector [31], to generate ZmPMP3-GFP fusion genes driving by CaMV 35S promoter. The specific primers containing XbaI and ScaII were shown in Table S3. The resulting constructs were transferred into onion epidermal cells by particle bombardment with gene gun (Biorad Helios TM ). The pBI221-GFP vector was used as control. After incubation on MS plate for 16-24 hours, the onion epidermal cells were observed with a laser scanning confocal microscope (Leica TCS-NT). For imaging the fluorescence of GFP, the excitation line was 488 nm for GFP, and GFP signal was collected under 465-490 nm bandwidths.
The YR93-1, YR93-31 and YR93-31M1-8 strains were cultured in YPDA liquid medium to an OD600-0.3, and then 20 ml of culture was transferred into 2 ml fresh YPDA liquid medium containing various concentrations of NaCl and 5 mg/L hygromycin B. Yeast was grown for 15 h at 30uC, and then absorbance (OD600) was measured. The absorbance (OD600) of culture under normal condition (without additional NaCl) was used as control. The values of relative growth rate were calculated by use of the formula: the relative growth rate = absorbance (OD600) of culture under NaCl or hygromycin B conditions/ absorbance (OD600) of culture under control.
The 10-fold dilutions of YR31-1, YR93-31 and YR93-31M strains were spotted onto solid YPDA plates and YPDA supplemented with 25 mM NaCl. The strains were cultured overnight to an OD600-1.0, and then diluted to OD600-0.3. The diluted yeast strains were dotted onto solid YPD plates containing 2% agar. The plates were cultured for 2-4 days at 30uC, and then photographed.
The relative growth rates of YR93-1, YR93-31 and YR93-31M1-8 yeast strains on low calcium and high calcium. The histogram showed the values of relative growth rate of low calcium and high calcium under salt conditions. The absorbance (OD600) of culture in YPDA with 25 mM NaCl was used as control. The values were calculated by use of the formula: relative growth rate = absorbance (OD600) of culture under low calcium or high calcium conditions/absorbance (OD600) of culture under control. The values represented the averages of three replications.

Generation of transgenic plants
The ORF of ZmPMP3-1 was cloned into NcoI/BstII sites of pCAMBIA3301 vector replacing the GUS gene, and ZmPMP3-1 was controlled by the CaMV35S promoter. The primers with NcoI and BstII sites were designed as follows: F-Nco, 59-CATCGG-CACCATGGCGGAGGGGACT-39, R-Bst, 59-ATGGGTCAC-CGTCGGCAAGGATGAC-39. The reconstructed vector was transferred into Agrobacterium GV3101 and then transferred into wild type Arabidopsis (Columbia 0 ecotype) plants by Agrobacterium tumaefaciens mediated transformation. Positive transgenic plants over-expressing ZmPMP3-1 were firstly screened on MS plates containing basta. T4 homozygous plants were used for further analysis.

Phenotyping of stress tolerance and physiological parameters
Seeds were vernalized for 3 days in 4uC on MS plates, 150 mM and 175 mM NaCl plates, then cultured for 3, 5 and 14 days in green house (22uC, humidity 40-50%, 120-150 mmol/m 22 s 21 under 16-h light and 8-h darkness). Green cotyledons rate and fresh weights of these plants were evaluated during different periods. Seedlings grown 7 days on MS plate were transferred to vertical MS and 200 mM NaCl plates for 10 days, and then the numbers of lateral roots morphology were monitored. To characterize salt tolerance of mature plants, 7 day old seedlings grown on MS plate were transferred to the soil (nutrition earth: vermiculite = 1:1) and cultured in green house. Plants at silique stage were irrigated with 300 mM NaCl solution for 7 days. Survival rates of these plants were evaluated and pictures were taken. Each value of green cotyledons rate represented the average of 40 seeds with at least three replicates. Each fresh weight was calculated based on the average weight of 20 individual plants with three replicates and the numbers of lateral roots were results from the averages of 8-10 plants with three replicates.
Plant cell membrane stability (CMS) was determined with a conductivity meter (DDS-1, YSI) and formula CMS (%) = (12initial electrical conductivity/electrical conductivity after boiling)6100. Four-old-week Arabidopsis plants irrigated with 300 mM NaCl for each line were selected to determine CMS and the values were calculated based on the averages of six individual plants with two replicates [65]. Chlorophyll florescence was measured with OS-30p Chlorophyll Fluorometer (OPTI-SCIENCE, USA) according to the protocol, and maximum efficiencies of PSII photochemistry, Fv/ Fm = (Fm2F0)/Fm, were used to assess changes in primary photochemical reactions of photosynthetic potential. Four-weekold Arabidopsis plants irrigated with 300 mM NaCl were chosen to collect Fv/Fm data. Fv/Fm value of each line represents average of eight individual plants with three replicates.

Real time RT-PCR analysis
Total RNAs were isolated from maize and Arabidopsis plants by TRIZOL reagent (Invitrogen, Carlsbad, CA), and then reversetranscribed by M-MLV reverse transcriptase according to the manufacture's protocol (Invitrogen, CA). Real-time PCR was performed in the presence of Power SYBR green PCR Master Mix (TaKaRa, Japan) and the amplification was performed in ABI 7300 sequence detection system. ZmGAPDH and AtActin were used as internal control to normalize all data. The primers, used for evaluating the transcript levels of ZmPMP3s, CSD1, CSD2, CSD3, CCS, APX1, APX2, P5CS1, P5CS2, CAT1, RD29B, DREB2A, KIN, PAB18, AVP1, AVP2, SOS1, HKT1, NHX1, ATK1, HKT1, CNGC12, AHA1, AHA2, AVA-P4, in qRT-PCR experiment, were shown in Table S5. qRT-PCR analysis was performed at least three times using sets of cDNA samples from independent plants. Data from qRT-PCR experiments were analyzed according to the 2 2DDCT method [66]. Specific primers were designed using PRIMER 5.0.     Dataset S1 Fasta file of all of the ZmPMP3 CDS' nucleic acid sequences. Accession numbers of sequences in GenBank were described in Table 1