To identify and isolate J. curcas genes involved in salt tolerance, yeast cells were used for functional screening. Yeast cells have been shown to exhibit salt stress, which is reflected as retardation in cell growth. Wild type yeast strains have been reported to show salt sensitivity from 500 mM depending on growth conditions. In our assays we scored for survival of yeast-transformants that could survive at and beyond 750 mM NaCl. This condition selected for the screen at a stringent concentration of 750 mM NaCl, subjected the yeast transformants to ionic as well as osmotic stress, and imposed stringent conditions for salt and drought stresses. Further, to enhance sensitivity of the screen, three salt hypersensitive mutants shs-2, shs-6 and shs-8 were generated by random mutagenesis that exhibited growth retardation beyond 500 mM NaCl (Figure 1). The salt hypersensitive mutant shs-2 was used to demonstrate the ability of recovered J. curcas genes to complement the salt sensitive defect in this mutant.

The underlying basis for the screen is the ability to use a catabolic regulated GAL1 promoter 15 to conditionally express a J. curcas gene in the yeast cells, cloned into a yeast shuttle vector 16; while scoring the relative survival of yeast cells exposed to salinity or other abiotic stress to the same yeast transformants, grown under repressed state, subject to similar stress or under control conditions. Scoring the relative growth patterns allows the identification and characterization of the tolerant transformants by plasmid-rescue (Figure 2).

To understand the early salt-stress phenotypes and characterize the genes involved in early responses to salt stress on J. curcas, 3-4 week old seedlings were treated with various levels of salt stress (see additional file 1: salinity treatment to seedlings). The phenotypic effects of salt stress resulted in observable wilting and drooping in J. curcas seedlings. The effect of salt stress was evident at a concentration of 150 mM NaCl, and the effects were pronounced at higher salt concentrations, or when the seedlings were treated for prolonged time (see additional file 1). Further, from this time-course analysis, for salinity stress phenotype in J. curcas it was identified that the exposure of roots to 150 mM NaCl for 2 h, at a relative humidity of 40-50% was sufficient to observe early effects of salt stress, which was marked by low level-wilting of young leaf tissue prior to stem drooping (see additional file 1: salinity treatment to seedlings).

Salt and stress pathways in plants are perceived by signalling networks and regulated by transcription machinery 17. To enrich cDNA libraries for transcripts that represent a greater proportion of early acting factor to later responding regulators, we sampled root tissue at the early-time points of salt stress. J. curcas root tissue samples treated at 150 mM, for 2 h time point were used in construction of cDNA libraries using the SMART cDNA synthesis Kit 18 (see Methods).

While plant genomes display large size variations, the number of expressed complement of genes involved in cellular processes are more conserved. Based on the information available for model plant Arabidopsis and rice, ~3000-5000 genes out of a transcriptome of ~27000 genes are expressed in root tissue 1920. Recent analysis indicates a modest genome size for the J. curcas, among Euphorbiaceae members, with an estimated genome size of ~450 Mb that provide an approximation of transcriptome complexity 21. To adequately represent root transcriptome, titers representing ~48000 c.f.u were recovered in each of the un-amplified pools that largely comprised of full-length cDNA sequences in 45-50% of the genes. This library was used for subsequent functional analysis in yeast.

To increase the probability of representation of each individual clone within these yeast expression libraries and to efficiently transform yeast, we amplified the copy numbers of clones contained in these libraries in E. coli (see Methods for details). A total of 20,000 yeast transformants (for the plasmid borne URA3 marker) representing the J. curcas root cDNA library was screened for salinity tolerance using a replica printing method (Figure 3). The replica printing process was devised to identify and isolate yeast transformants expressing heterologous gene sequences derived from J. curcas root cDNA libraries, with a low-rate of false negatives. Screening criteria included scoring relative survival advantage, for each transformant when subject to stress and gene expression status (Figures 2 and 3). This method increased efficiency of gene recovery and reduced the screening errors by providing additional, more exhaustive controls that allowed the identification and elimination of both false-positive as well as false-negative phenotypes. The screening conditions select yeast transformants that are capable of survival under stress (acquired due to expression of J. curcas heterologous cDNA) and are able to grow; while eliminating events due to lethality, non-expression and other experimental artefacts.

Using the replica-printing based functional screen, 345 candidate yeast transformants expressing cDNA (derived from J. curcas roots) were isolated from a pool of 20,000 yeast transformants. A representative data set showing results of phenotypic selection of salt tolerance is presented in (Figure 4). The transformed yeast cells displayed an enhanced ability to tolerate unfavourable conditions imposed by salinity stress at a stringent selection of 750 mM NaCl concentration.

To determine and characterize J. curcas cDNA gene sequences that conferred enhanced salinity tolerance at 750 mM NaCl to yeast transformants plasmid rescue was performed as described in Bates 22. Analysis of the DNA sequence information, recovered from the yeast transformants showed that 32 sequences from 345 clones encoded full-length genes with assignable ORF's derived from J. curcas; while the remaining were truncated cDNA sequences. The ability to recover genes from J. curcas involved in these processes demonstrates the efficacy of our strategy.

To demonstrate the robustness of the screening method, we have re-transformed the rescued plasmids into wild-type yeast (data not shown) and salt hypersensitive mutant shs-2 (Figure 5). We evaluated these transformants for growth at high-salt stress of 750 mM or higher concentration. Several transformants of shs-2 mutant allowed growth at 750 mM NaCl. The data demonstrated the ability of the plasmid encoded genes derived from J. curcas permitted survival of wild-type yeast transformants (data not shown), and complemented several salt sensitive phenotypes of the shs-2 mutants. The data provided functional support for the consistent ability of the recovered genes to confer salt tolerance, and support cell growth at conditions otherwise detrimental to cell survival.

Further, sequence analysis and computational annotation reveals presence of novel genes from J. curcas involved in abiotic stress tolerance (Table 1). To annotate the sequences, computational searches were performed for similarity to sequences deposited in plant sequence databases (see Methods), that suggest the presence of plant specific regions within these sequences (Table 1). In this analysis 31 out of 32 (~96%) sequences with assignable hit to the sequence databases showed matches greater than the accepted threshold as elaborated by Altshul et al. 23 to other known plant genes; while for a single sequence, significant matches to existing sequences in the sequence database was not apparent (Table 1). This sequence could represent novel determinant of salinity and drought stress tolerance, whose role needs to be further established. Of the 31 sequences, 7 sequences showed match to encoded ribosomal proteins, while 1 sequence was annotated as a eukaryotic translation initiation factor SUI1. Twenty three sequences have been reported to have a strong role in conferring abiotic stress tolerance in other plant species (Table 1).

To understand the possible role of the genes and regulatory pathway for abiotic stress tolerance in J. curcas, we assigned functional classes for the gene sequences recovered in the present study. Comparison of functional annotation for several of the sequences obtained from plasmid rescue indicated a possible role in tolerance to abiotic stress. Some of the sequence orthologs of these genes have been reported to be involved in conferring stress tolerance in other plant species. We find among the list of known functions sequence classes mechanistically involved in abiotic stress tolerance due to wounding, response to cellular redox or oxidative damage, and modulation of cell physiology in response to stresses. For example, previous investigations indicate the involvement of metallothioneins 24, thioredoxins, cellular redox machinery 2526 and factors involved in the regulation of protein translation 27 to be associated with stress resistance in rice and Arabidopsis. Among gene sequences listed in Table 1, the most notable examples with implied roles in stress tolerance are discussed below: we isolated a sequence ortholog corresponding to allene oxide cyclase. Sequences of allene oxide cyclase, isolated from mangrove have been reported to confer salinity stress tolerance in multiple organisms 28. Previous research data also indicate that accumulation of late embryogenesis abundant protein (LEA-5) transcripts in response to stress conditions such as cold, drought, UV light, salinity and wounding 29. In this report, we isolated J. curcas LEA-5 like sequences, suggesting a conservation in stress response pathway. Additionally, we identified genes encoding major allergen Pru ar1-like proteins, defensin and protease inhibitor/seed storage/lipid transfer protein family reported to get up-regulated upon pathogen attack and/or environmental stimuli 30. Plants have been known to have specific mechanism for detoxification of reactive oxygen species (ROS) induced in stress response which includes activation expression of antioxidant enzymes such as cytosolic ascorbate peroxidase-1 (Apx1). Experiments involving over-expression of Apx-1 in transgenic tobacco, conferred enhanced tolerance to salt and water stress, by effectively detoxifying accumulated H₂O₂ in the chloroplasts 31. Mitochondrial ATP synthase 6 KD subunit (MtATP6) is reported to be involved in NaHCO₃ induced stress, whose over-expression in tobacco led to greater salt stress tolerance 17. Al-induced protein and copper chaperone have been implicated to provide protection to cellular function during cytotoxic reactions caused by metal ions 32. We identified J. curcas sequence ortholog of a profilin. Profilins have previously been implicated in root elongation and root hair system, a response that is directly linked to response of plants to adapt and source water during drought resistance 33. The high frequency of occurrence of functionally conserved genetic components indicates a possible conservation in some of the genetic pathways and/or the genes regulating salinity and drought stress among diverse plants 2. The fact that some of these factors are conserved suggests that certain aspects of mechanism of salt and water deficit stress sensing and response may be evolutionarily conserved among diverse taxa.

To understand and correlate the gene expression pattern of selected genes to functional roles in stress determination, we performed gene expression analysis in leaf and root tissues of J. curcas. In the present study, we selected five genes i.e., late embryogenesis abundant protein 5 (LEA-5), cytosolic ascorbate peroxidase (Apx-1), metallothionein, profilin and annexin, based on the diverse functional classes and varied mechanisms implicated in stress avoidance/tolerance pathways in model plants. While our functional screen has been designed to identify functional genes to abiotic stress tolerance, rather than rely on changes in gene expression; several reports have correlated changes in gene expression to possible gene functions mediating stress tolerance 8. To correlate gene function and gene-expression status for each of the genes, semi-quantitative RT-PCR analysis was conducted. We assayed for changes in transcript abundance for each of the genes that have been normalized to actin, in both leaf and root tissue, that were either control treated or salt stressed, before sampling at various time-intervals (see Methods). Changes in expression for each transcript normalized to actin have been plotted in Figure 6 and Figure 7.

Analysis of the gene-expression data suggests dynamic changes in the transcript abundance of these genes, with changes in transcript level being apparent from 2 h time-point, indicating an early regulation of the genes in response to salt stress. Gene expression of the normalized transcripts suggests both up-regulation as well as repression of the transcripts. We noted increased expression of profilin in root tissue between 2 h to 8 h time-points of exposure to 150 mM salt stress; whereas profilin was repressed in leaves (Figure 7 vs. Figure 6) consistent with previously reported function during salt stress 33. In contrast, we observed an initial repression in the expression of annexin that was subsequently induced in the root-tissue, while the expression was down-regulated upon salt-stress in the leaves (Figure 6 vs. Figure 7). Transcript abundance in the levels of Apx-1 and metallothioneins were down-regulated in roots, while a more dynamic expression pattern was observed in the leaf tissues, implicating a role in stress adaptive responses 31. Among the transcripts we profiled in this study, only LEA-5 was up-regulated in leaves at the 8 h time interval, after initial down-regulation at the 2 h time-point; while in roots LEA-5 expression was unaffected by salt stress. Such an induction of LEA-5 in leaf-tissue demonstrated either a longer range signal or the terminal effects of long exposure to ions by the roots. Further experiments are required to de-lineate these effects. The gene-expression data, indirectly suggests varied modes of gene-regulation between the leaf and root tissue (Compare Figure 6 to Figure 7). The data indicates a complex framework for gene regulation during adaptation to salt stress in different tissues. The data in conjunction with gene function may help unravel certain aspects of plant abiotic stress response.

Seeds of J. curcas (S003) were collected from Reliance Life Sciences' agricultural farm at Kakinada, South of Andhra Pradesh, India. Seed coats were removed from the seeds, surface sterilized with 70% ethanol followed by 5% hypochloride solution prior to being placed onto MS basal salts 34 distributed into 150-mm culture tubes. MS basal salts supplemented with 3% sucrose and solidified with 0.8% agar (Hi-Media, India) were used to grow seeds. The cultures were maintained at 23-25°C at 50-60% RH, under long day conditions as described previously 35.

Young 3-4 week-old in vitro germinated J. curcas seedlings were removed from the media and randomly separated into groups of 15-20 plantlets. Roots (both main and lateral roots) of these plants were either placed in sterile water or into 150 mM NaCl solution for 2 h. The treated Jatropha root tissue from plants that were untreated or treated were dissected and frozen in liquid nitrogen, prior to being stored at -80°C.

J. curcas root samples control treated or challenged with 150 mM NaCl were used to generate pools of cDNA. Total RNA for cDNA synthesis was extracted from J. curcas root samples using Qiagen RNA Miniprep Kit (Qiagen, Germany). Briefly, the root tissue samples were homogenized to a fine powder in liquid nitrogen and total RNA was extracted as described in the Plant mini RNA prep kit (Qiagen, Germany). Total RNA was estimated spectrophotometrically using Nanodrop.

To prepare poly (A+) mRNA pools suitable for cDNA synthesis, 10 μg total RNA was treated with RNAase free DNAaseI (Sigma-Aldrich, St Louis, USA) for 15-20 min at 37°C, and mRNA fraction was enriched using oligo-d(T) beads (Oligotex, Qiagen, Germany). First strand cDNA pools were synthesized from normalized amounts of RNA either derived from untreated root tissue or from tissue challenged with salt stress, using PowerScript reverse transcriptase (Takara, CloneTech) as described in the Super SMART cDNA synthesis Kit 18. Double stranded DNA was generated through PCR amplification using conditions described in Table 218. The cDNA pools were size separated using NucleoSpin columns (BD CloneTech, USA). The yield and quality of the amplicons were monitored on agarose gel analysis. A profile of amplification patterns obtained after 22 and 25 cycles of PCR amplifications as described in the Super SMART cDNA synthesis Kit 1836 is shown in Figure 8. A schematic diagram elaborating the construction of Jatropha root cDNA libraries is outlined (see additional file 2:-process flow).

To identify and define the condition that lead to salt stress to wild-type yeast BY4741 (See additional file 3: genotype details) (yeast strain obtained from EUROSCARF, Germany), we measured relative growth of yeast (YPD media 2% Peptone, 1% Yeast extract, 2% Dextrose, solidified with 1.5% Agar, Hi-Media, Mumbai, India) under a range of salinity stress, from 0.0 mM NaCl to 2.0 M NaCl. Salt-hyper-sensitive mutants were isolated through UV induced random mutagenesis and selected on salt containing plates by replica-plating.

To express the cDNA libraries in yeast, the cDNA library amplicons were cloned into a yeast expression vector, pYES 2.1 TOPO TA (Invitrogen, Carlsbad, USA). The pYES 2.1 TOPO TA is a E. coli-Yeast shuttle vector 16, propagated in E. coli using bacterial selection marker for Amp^r; while the transformants in the yeast BY4741 strain background are selected for the URA3 marker. Cloning of gene of interest downstream of the GAL1 promoter allows regulated gene expression of the library in yeast (Invitrogen, Carlsbad, USA). The J. curcas root cDNA libraries in pYES 2.1 TOPO TA were transformed into E. coli TOP 10F' (Invitrogen, Carlsbad, USA). Subsequently, the cDNA library were recovered from the plates and plasmid DNA representing the pooled libraries, derived from J. curcas root cDNA was extracted using Plasmid Midi preparation Kit (Qiagen, Germany).

Plasmid transformation of yeast (Saccharomyces cerevisiae) was accomplished using PEG-lithium acetate based transformation protocols 37, while the plasmid selection in yeast was based on the URA3 marker borne on the yeast expression plasmid pYES2.1 TOPO TA (Invitrogen, Carlsbad, USA). Amplified plasmids containing cloned inserts derived from J. curcas, regulated by the galactose-inducible GAL1 promoter, were transformed into wild-type yeast strain BY4741; following heat-shock at 15 min at 42°C, the yeast cells were revived in YPD media and plated on synthetic minimal medium plates lacking uracil (see additional file 4: stock composition), and placed at 23-25°C for 48-96 h.

The yeast (BY4741) containing J. curcas root cDNA library clones, in GAL1 regulated yeast expression systemwere screened for salinity and drought resistance. Twenty thousand yeast transformants were picked and inoculated into sterile 96 well U-bottom microtitre plates (Nunc, USA) containing synthetic selection media (see additional file 5: composition of synthetic selection media). Subsequently, the individual yeast transformants were replica printed on quadruplet selection plates containing either 0 mM NaCl or 750 mM NaCl using a 96-pin replicator (Nunc, USA)(Figure 3). Under these screening conditions each individual yeast transformant (arrayed from 96-well plates) was tested for its ability to grow under four conditions described earlier. The J. curcas cDNA expression in yeast could be regulated by the yeast GAL1 system 38.

Isolated yeast transformants were grown on synthetic selection media containing 2% dextrose without salt (see additional file 5: composition of synthetic selection media) for 36-72 h. The transformants were lysed with 10 U/μl lyticase (Sigma-Aldrich, St Louis, USA) and 2-4% SDS. Nucleic acid fraction recovered from yeast were purified with two sequential rounds of phenol:chloroform:isoamyl alcohol (25:24:1) extraction, followed by ethanol precipitation of nucleic acids. To analyze the inserts cloned in the pYES2.1 TOPO TA based yeast expression plasmid, nucleic acid preparations recovered from individual yeast transformants showing resistance to salinity stress were backtransformed into E. coli via electroporation (GenePulser II, BioRad, USA) 39.

E. coli backtransformants for yeast expression plasmids were analyzed for inserts by colony PCR analysis with the GAL1 and V5/6XHIS primers as described in the pYES2.1 TOPO TA kit (Invitrogen, Carlsbad, USA). The conditions for colony PCR analysis were provided in Table 2. E. coli re-transformants (representing the yeast plasmid expression vector containing Jatropha cDNA) were grown and plasmid DNA extracted as described in Qiagen plasmid miniprep kit (Qiagen, Germany).

Sequencing of Jatropha cDNA conferring salinity tolerance was performed as described with BigDye Sequencing Kit (ABI, USA), and analyzed on Genetic Analyzer (ABI 3100, USA). Sequencing reaction was with vector specific GAL1 and V5/6XHIS primers

Prior to annotation, the sequences were subjected to quality check and vector masking using NCBI's UniVec http://www.ncbi.nlm.nih.gov/blast/. To understand and assign functional classes to the sequence information determined in this screen, we performed computational searches against sequence databases at NCBI http://www.ncbi.nlm.nih.gov/blast/ and TIGR plant transcript assemblies http://tigrblast.tigr.org/euk-blast/plantta_blast.cgi using the BLAST algorithm 2340.

Plasmid transformation of yeast hyper-salt sensitive mutant (shs-2) with five selected genes viz., LEA 5, mitochondrial ATP synthase 6 KD subunit, cytosolic ascorbate peroxidase, metallothionein and profilin was done as mentioned earlier. Single colony of each transformants and mutant strain was dissolved in 100 μl sterile water. YPD media containing range of NaCl concentrations (0 mM, 250 mM, 500 mM, 750 mM and 1 M) were prepared in Petri dishes. Each Petri plate was divided into six sectors. 10 μl of each yeast transformants and the mutant strain were patched on the salt series plates uniformly and incubated overnight at 30°C incubator. Survival of transformants was scored against hyper-salt sensitive mutant strain.

The semi-quantitative RT-PCR expression of selected genes was performed with specific oligonucleotide primers (Table 3) on first strand cDNA synthesised from RNA isolated from root and leaf tissues samples of 21-day-old J. curcas seedlings. The total RNA was extracted from the 150 mM salt treated leaves and roots at different durations of 0 h, 2 h and 8 h as described in the Plant mini RNA prep kit (Qiagen, Germany). The total RNA was estimated spectrophotometrically at 230, 260 and 280 nm (Nanodrop). Eight hundred ng of total RNA was taken to synthesize the first strand cDNA with oligodT using Superscript reverse transcriptase (Invitrogen) as per manufacturer's protocol. For PCR, 1 μL of cDNA was used as DNA template in a reaction volume of 50 μL using PCR master mix (Lucigene) with cycling conditions of 95°C for 5 min, 95°C for 1 min, 60°C for 1 min, 72°C for 1 min. The amplification reaction was carried out for 32 cycles for all genes with a final extension at 72°C for 10 min. J. curcas actin gene was used as an internal control for analysis of gene expression. RT-PCR analyses were carried out with three independent total RNA samples.

Quantitation of RT-PCR product was determined by densitometer. Two μl of RT-PCR products derived from target gene and actin gene were resolved on 2%agarose gels stained with ethidiumbromide. Densitometeric scan analysis was carried out using Kodak MI Imaging software program as per the supplier's instructions. Percent gene expression was determined by normalizing values against internal control actin. Values were represented as per cent of gene expression with respect to corresponding controls and plotted using Microsoft Excel.

Sequence data from this article can be found in the GenBank data libraries under accession numbers FJ489601-FJ489611, FJ619041-FJ619057 and FJ623457-FJ623460.