Characterization of the Sesbania rostrata Phytochelatin Synthase Gene: Alternative Splicing and Function of Four Isoforms

Phytochelatins (PCs) play an important role in detoxification of heavy metals in plants. PCs are synthesized from glutathione by phytochelatin synthase (PCS), a dipeptidyltransferase. Sesbania rostrata is a tropical legume plant that can tolerate high concentrations of Cd and Zn. In this study, the S. rostrata PCS gene (SrPCS) and cDNAs were isolated and characterized. Southern blot and sequence analysis revealed that a single copy of the SrPCS gene occurs in the S. rostrata genome, and produces four different SrPCS mRNAs and proteins, SrPCS1–SrPCS4, by alternative splicing of the SrPCS pre-mRNA. The SrPCS1 and SrPCS3 proteins conferred Cd tolerance when expressed in yeast cells, whereas the SrPCS2 and SrPCS4 proteins, which lack the catalytic triad and the N-terminal domains, did not. These results suggested that SrPCS1 and SrPCS3 have potential applications in genetic engineering of plants for enhancing heavy metal tolerance and phytoremediation of contaminated soils.


Introduction
Phytochelatins (PCs) have been found in all plant species investigated, a few fungal species and some marine diatoms [1]. PCs can act as high affinity metal chelators and facilitate the vacuolar sequestration of heavy metals, especially Cd 2+ ions, and hence play in the plant kingdom an important role in detoxification of some heavy metals [2,3]. PCs have the general structure (γ-Glu-Cys) n -Gly, where n has been reported to be as high as 11, but generally ranges from 2 to 5 [4]; they are synthesized posttranslationally in the presence of heavy metal ions from glutathione (GSH) by PC synthase (PCS), a γ-glutamyl-cysteine transpeptidase [5]. So far, four different models for the catalytic mechanism of PCS have been proposed [5][6][7][8][9][10], although a consistent model that clearly demonstrates how PCS is activated and how PCs are synthesized has not been established. Alignment of PCS protein sequences reveals a high degree of similarity in the N-terminal domain, whereas the C-terminal domain which turns out to be extremely variable, is missing in the prokaryotic PCS proteins [11,12]. The N-terminal core domain of PCS, which can be referred to the catalytic domain sufficient to confer a PCS activity, has been confirmed by limited proteolysis and mutant analyses [6,8,11,13]. The variable C-terminal region was shown to ensure higher PCS activity, enhance protein stability and respond to a broader spectrum of heavy metals [11]. The stereostructure of the PCS from the cyanobacterium Nostoc sp. PCC7120 (NsPCS) in its native and γ-Glu-Cys-acylated condition was reported [14]. This study represented the first crystal structure of a PCS and established that prokaryotic NsPCS and the eukaryotic Arabidopsis thaliana PC synthase1 (AtPCS1) belong to the papain superfamily of cysteine proteases, with a conserved catalytic machinery that had been deduced from kinetic, protein chemical, and site mutagenic studies of the prototypical AtPCS1 [8,10,15].
PCS genes have been cloned and characterized in Arabidopsis thaliana (AtPCS1, [3,16]), Schizosaccharomyces pompe (SpPCS, [17]), Triticum aestivum (TaPCS1, [6]), Lotus japonicus (LjPCS, [18][19][20], the model nematode Caenorhabditis elegans (CePCS1, [21,22]) and the prokaryote Nostoc sp. PCC7120 (NsPCS, [8,12]). It has been proposed that PCS is a constitutive enzyme in general and its gene expression is not induced in response to heavy metal exposure [23]. This proposal was supported by the analysis of AtPCS1 expression, which showed that the level of AtPCS1 mRNA was unchanged on exposure to heavy metals such as Cd, Zn, or Cu, or to oxidative stress, salt stress, jasmonic acid or salicylic acid [6,10,24]. During the early stages of plant development, however, the steady-state level of AtPCS1 mRNA increased by 2-fold in 5-day-old Cd-treated Arabidopsis compared to non-treated seedlings [25]. Clemens et al. [17] found that the level of TaPCS1 in wheat roots treated with 100 μM Cd 2+ was 5-to 10-fold higher than those of untreated roots. These studies indicated that in specific organs and developmental stages of plants, in which a strong defense against exposure to toxic heavy metals is required, the expression of the PCS genes may be regulated in response to heavy metals.
Sesbania rostrata, a legume native to tropical West Africa, is an annual species that bears stem nodules as well as root nodules associated with Azorhizobium caulinodans, a specialized fast-growing strain of Rhizobia that can fix nitrogen in its free-living form [26]. Yang et al. [27] found that S. rostrata can tolerate some heavy metals, especially Zn and Cd. To understand the mechanism of heavy metal tolerance in S. rostrata, in this study the corresponding PCS gene and its cDNAs were isolated and characterized.
An unrooted phylogenetic tree ( Figure 2) was constructed including the majority of the predicted PCS sequences available in the databases. The tree revealed separate clusters, paralleling taxonomic distance, for the PCSs of cyanobacteria, nematodes, yeast, ferns, and the families of vascular plants Poaceae, Allaceae-Typhaceae, Solanaceae, Brassicaceae, and Leguminosae. As shown in Figure 2, SrPCS proteins (in red dashed ellipse of Figure 2) are highly homologous to leguminous soybean GmPCS [7] and Lotus LjPCS1 [18]. The structure and function of PCS genes have only been studied in detail in the model plant Arabidopsis. Two functional PCS genes, AtPCS1 and AtPCS2, which are localized in different chromosomes (At5g44070 and Atlg03980), were found in Arabidopsis [6,10,17,24]. In order to study the structure of S. rostrata PCS gene (SrPCS), we cloned the genomic sequence of SrPCS from the genomic DNA of S. rostrata by genome walking [28] with primers based on the conserved sequence of SrPCS1-4 cDNAs. Unexpectedly, only a 5,112-bp genomic DNA fragment of SrPCS (GenBank accession no. GQ204307) was obtained.

Genomic Organization and Pre-mRNA Splicing of SrPCS Gene
Alignment of the genomic sequence of SrPCS with those of four cDNA sequences of SrPCS1-4 revealed that all four mRNAs (SrPCS1-SrPCS4) were produced from the one SrPCS gene through alternative splicing (Figure 3).
By comparing with the sequence of the canonically spliced form SrPCS3 mRNA, eight exons and seven introns were found in SrPCS gene. As shown in Figure 3, SrPCS1 and SrPCS2 are generated through alternative AG acceptor splice sites. And the SrPCS4 is produced by atypical splice mechanism employing direct repeats at splice sites (CTCC, Figure 3), which is similar to the atypical mRNA processing of p53 transcripts in the human brain [29].
The presence of a single copy of the SrPCS gene in S. rostrata genome was further confirmed by Southern blot analysis (Figure 4). Only a single hybridizing band was detected in DNA samples digested with either EcoRI, BamHI or XbaI, which do not cut within the probe region ( Figure 4, lane 1-3). When the genomic DNA was digested with PstI that cut once within the probe region, two hybridizing bands were observed (Figure 4, lane 4). These results indicate that the S. rostrata genome contains only a single copy of the SrPCS gene.

Expression Analysis of SrPCS Gene
Alternative splicing of pre-mRNA is often regulated by specific factors in response to developmental or environmental cues [30][31][32]. To find out whether the alternative splicing of SrPCS pre-mRNA is regulated by heavy metals, we investigated SrPCS expression in plants in the presence and absence of Cd 2+ . As in wheat [17] and Brassica juncea [33] plants, SrPCS1−SrPCS4 mRNA expression was very low and could not be detected by RNA blot analysis of total RNA from different organs of S. rostrata. Therefore a more sensitive real time-PCR method was used to estimate the abundance of SrPCS transcripts. Unfortunately, because of very high sequence similarity of the four SrPCS mRNA, only specific primers for SrPCS3 can be designed for real time-PCR analysis.   Unlike in Arabidopsis and Brassica juncea, where higher amounts of Arabidopsis AtPCS1 transcript [6] and B. juncea BjPCS1 protein [33] were detected in roots than in leaves, SrPCS3 is more highly expressed in the leaves than roots and stems under normal growth conditions without Cd 2+ treatment ( Figure 5). SrPCS3 expression was significantly reduced in roots after exposure to different concentration of Cd 2+ for 24 hours. No significant change of SrPCS3 expression was observed in stems, whereas SrPCS3 expression was also significantly down-regulated in leaves by 0.3 mM Cd 2+ treatment. These observations are contrary to TaPCS1 from wheat [17] and LjPCSs from the model legume Lotus japonicus [18]. Clemens et al. [17] found that TaPCS1 showed a 5-to 10-fold higher expression in wheat roots treated with 0.1 mM Cd 2+ for 6 h. Ramos et al. [18] investigated the expression of three PCS genes in roots of L. japonicus, and found that the mRNA expression of LjPCS1-8R increased only slightly (2.1-fold) after 6 h of 0.1 mM Cd 2+ treatment, whereas the mRNA levels of LjPC2-7R and LjPCS3-7N increased 2.3-to 2.9-fold and 2.5-to 3.5-fold between 6 and 96 h, respectively. The relative expression of SrPCS3 to an internal control glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in S. rostrata plants, treated with different concentration of Cd 2+ for 24 hours, was analyzed by real-time RT-PCR. Mean values plus standard errors are given. * indicates significance at the 5% level.

Functionality of Four SrPCS Isoforms SrPCS1−SrPCS4
To test the functionality of four SrPCS isoforms, the full-length cDNAs of SrPCS1−SrPCS4 were expressed in S. cerevisiae under the control of the inducible pGAL1 promoter. Cd tolerance assay showed that S. cerevisiae cells expressing SrPCS1 or SrPCS3, but not SrPCS2 and SrPCS4, exhibited strong tolerance to Cd 2+ (600 μM), which completely inhibited the growth of cells transformed with the empty pYES2 plasmid or cells expressing SrPCS2 or SrPCS4 (Figure 6). Growth curve experiments (Figure 7, a-d) also demonstrated that expression of SrPCS1 and SrPCS3 conferred enhanced Cd tolerance to yeast cells (Figure 7, a and c), whereas the expression of SrPCS2 and SrPCS4 did not make yeast cells more Cd tolerant (Figure 7, b and d).  Dose-response analyses showed that SrPCS3-expressing yeast cells tolerated higher Cd 2+ concentrations than the SrPCS1 expressing cells and control cells (Figure 7, e), which may result from the lack of the C-domain that has been shown to ensure higher PCS activity [11]. Alternative splicing, unlike gene promoter activity that primarily regulates the amount of transcripts, changes the structure of transcripts and their encoding proteins [34]. Therefore alternative splicing is one of the most important mechanisms to generate a large number of mRNAs and protein isoforms from a single or few genes [35]. Several studies have shown that about 20% of plant genes have one or more alternative transcript [36][37][38][39]. And based on the alignments of ESTs from other legume species against the Medicago truncatula genome sequence, hundreds of legume genes were recently predicted to be alternatively spliced [39].
In this study, we found that a single copy of the SrPCS gene occurs in S. rostrata genome, and that alternative splicing of the SrPCS pre-mRNA produced four different SrPCS mRNAs and proteins, SrPCS1−SrPCS4 (Figure 1). The SrPCS1 and SrPCS3 proteins, whose N-terminal domains contain a catalytic triad of Cys-56, His-162 and Asp-180 (Figure 1), conferred Cd tolerance when expressed in yeast cells ( Figure 6 and 7), whereas the SrPCS2 and SrPCS4 proteins, which contain only a Cys-56 of the catalytic triad and lack the N-terminal domains (Figure 1), did not render yeast cells more Cd tolerant (Figure 6 and 7). This observation is similar to the findings of Ramos et al. [18], who showed that a PCS gene (LjPCS2) in model legume Lotus japonicus can be alternatively spliced to encode two protein isoforms, one (LjPCS2-7N) conferring enhanced Cd tolerance to yeast cells while the other (LjPCS2-7R) did not. SrPCS1 and SrPCS3 characterized in this study have potential applications in genetic engineering of plants for enhancing heavy metal tolerance and phytoremediation of contaminated soils [40]. Although we detected the mRNA expression of the four SrPCS isoforms in the current study, the corresponding proteins of these mRNAs and their physiological functions remain to be characterized. In particular, the presence of SrPCS2 and SrPCS4 proteins, which lack the catalytic triad and the N-terminal domains and do not confer enhanced Cd tolerance to yeast cells, needs to be determined in S. rostrata plants, since they are small variants of the full-length protein (about one-third of the SrPCS3) and may not be stable if translated in plant cells.

Plant Materials
Seed testae of Sesbania rostrata were scalpelled and surface-sterilized in 70% alcohol for 1 minute, followed by three rinses with sterile distilled water, and then soaked in 10% sodium hypochlorite for 15 minutes with gentle agitation. After washing three times with sterile distilled water, the seeds were germinated on 1/2 MS medium [41] for 6 days. These seedlings were then transferred to 1/8 liquid MS without sugar to grow in this medium for 15 days, changing medium every three days, and then transferred to 1/4 liquid MS containing different concentration of cadmium for 1-2 days as needed at 24 ± 2 o C under fluorescent light with a 16-h photoperiod.

Cloning of PCS cDNAs and Gene from S. rostrata
Total RNA was extracted from seedlings treated with Cd for 2 days as described above using the Trizol reagent (Invirtogen). The cDNA was synthesized from 2 µg of total RNA that was treated by DNaseI before reverse transcription (RT) with the M-MLV (Progema). RT-PCR amplification of the conserved partial sequence of SrPCS cDNAs was performed with a pair of primers (forward primer ZF168, 5'-GCGTTAAGAGACAATGGCGATG-3'; reverse primer ZF169, 5'-GGGAGATGCTAAG-AAAAGGGTACA-3') based on the sequence of soybean PCS [7]. The amplification parameters were as follows: 5 min at 95 o C, 30 cycles of 40 s at 95 o C, 40 s at 54 o C and 2 min at 72 o C, 10 min at 72 o C in a 20-µl reaction mixtures. The 5'-and 3'-ends of SrPCS cDNAs were isolated with a modified RACE procedure based on template-switching effect and inverse PCR [42]. Genomic DNA from S. rostrata was extracted by a modified CTAB method [43]. SrPCS genomic clone was obtained using the Universal GenomeWalker Kit (Clontech) according to the manufacturer's protocol. PCR products were cloned into pGEM-T easy (Progema) and sequenced. Sequence identity was calculated by the Ident and Sim program in the Sequence Manipulation Suite (http://www.bioinformatics.org/sms/).

Real Time RT-PCR of SrPCS3
Real time RT-PCR was used for the quantification of the SrPCS3 expression using the Premix Ex Taq TM (Perfect Real Time, TakaRa) with a pair of SrPCS3 specific primers (forward primer ZF332, 5'-GCTTTATACTCTGAGCTG-3'; reverse primer ZF335, 5'-TTAGATGGCAG-TGATGT-3'). As an internal control, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA was simultaneously amplified to normalize SrPCS3 expression with a pair of GAPDH specific primers (forward primer ZF371, 5'-GCTGGTGCTGATTATGTT-3'; reverse primer, ZF372 5'-GCTCTGGCTTGTATTCCT-3'). PCR cycling conditions were: 15 s at 95 o C; 40 cycles of 5 s at 96 o C, 20 s at 58 o C and 30 s at 72 o C in a 25-µl reaction mix containing 1x SYBR Green I. Each sample was performed in triplicates at least three times. Amplification products were measured and analyzed with an ABI Prism 7900HT sequence detection system and SDS software (Applied Biosystems).

Southern Blot Analysis of SrPCS
For Southern blot analysis, 25 µg genomic DNA was digested with restriction enzymes and separated on 0.8% agarose gel and transferred to Hybond-N membrane (Amersham) according to Sambrook and Russell [44]. The conserved SrPCS cDNA fragment (nt 47-919 of GenBank GQ204309) was labeled with [α-32 p] dCTP using Random Primer Kit (TakaRa) and used to detect DNA on the membrane as previously described [45].

Expression of Recombinant SrPCS1-4 in Yeast and Cd Tolerance Assay
Four ORFs of SrPCS1-4 cDNAs were amplified by PCR with a pair of SrPCS-specific primers carrying a BamHI and a KpnI restriction enzyme site (forward primer ZF268, 5'-GGTACCATGGCG-