A Seed Preferential Heat Shock Transcription Factor from Wheat Provides Abiotic Stress Tolerance and Yield Enhancement in Transgenic Arabidopsis under Heat Stress Environment

Reduction in crop yield and quality due to various abiotic stresses is a worldwide phenomenon. In the present investigation, a heat shock factor (HSF) gene expressing preferentially in developing seed tissues of wheat grown under high temperatures was cloned. This newly identified heat shock factor possesses the characteristic domains of class A type plant HSFs and shows high similarity to rice OsHsfA2d, hence named as TaHsfA2d. The transcription factor activity of TaHsfA2d was confirmed through transactivation assay in yeast. Transgenic Arabidopsis plants overexpressing TaHsfA2d not only possess higher tolerance towards high temperature but also showed considerable tolerance to salinity and drought stresses, they also showed higher yield and biomass accumulation under constant heat stress conditions. Analysis of putative target genes of AtHSFA2 through quantitative RT-PCR showed higher and constitutive expression of several abiotic stress responsive genes in transgenic Arabidopsis plants over-expressing TaHsfA2d. Under stress conditions, TaHsfA2d can also functionally complement the T-DNA insertion mutants of AtHsfA2, although partially. These observations suggest that TaHsfA2d may be useful in molecular breeding of crop plants, especially wheat, to improve yield under abiotic stress conditions.


Introduction
Plants are exposed to several environmental stresses during their life cycle that hampers their growth and development, resulting in considerable yield losses. Some of the common abiotic stresses that plants encounter are drought, flooding, salinity stress and extreme temperatures. As a consequence of a combination of different abiotic stresses, there is a dramatic yield loss in many crop plants. High temperature stress is one of the most severe stresses adversely affecting crop production worldwide [1]. However, morphological, physiological, biochemical and molecular changes occur in some plant species, that enable them to survive under high temperature conditions. A common response to high temperature stress is the synthesis of Heat Shock Proteins (HSPs). HSP encoding genes are found to be under the control of Heat Shock Factors (HSF), transcription factors that regulate the expression of HSP genes. The number of HSFs differs in organisms, for example, while vertebrates have four, Drosophila, C. elegans and yeast have only one HSF [2][3]. In plants, the numbers of HSFs are characteristically high, while Arabidopsis has 21, rice 25, Brachypodium 24, soybean 52 and Poplar 27 members [4][5][6]. Plant HSFs are represented by conserved structural features such as presence of highly conserved N-terminal, helix turn helix type DNA binding domain (DBD), followed by an oligomerization domain composed of two hydrophobic heptad repeats (HR-A and HR-B) connected to DBD by a linker of 15-80 residues and presence of nuclear localization (NLS) and nuclear export signals [4]. In contrast to Nterminal architecture, the C-terminal trans-activation domain (CTD), possessing AHA motif, is less conserved in plants. Based on sequence homology and domain architecture, plant HSFs have been divided into three conserved classes. Type A HSFs are characterized by exclusive presence of AHA transactivator motif in their C-terminal trans-activation domain (CTD), whereas class B and class C HSFs do not possess a trans-activator domain. Class B HSFs have a longer HR-A/B region due to insertion of 21 and 7 residues, respectively, and are more similar to non-plant HSFs than class A and C [4,7]. When exposed to heat stress, these HSFs form trimers and bind with high affinity to DNA at the heat stress elements (HSEs) in the promoter sequences [8,9]. The functional HSE sequence is inverted repeats of 5'-nGAAnnTTCn-3' motif in promoters of the Hsp genes [10]. In plants, HSF family has been described in detail in two model organisms, Arabidopsis and rice [4,5,[11][12].
There is functional diversification found among different HSF members. HSFA1a is described as the master regulator in tomato for acquired thermotolerance [13]. It forms a complex with other HSFs, A2 and B1, and thus regulates thermal response. Unlike tomato, Arabidopsis has four HsfA1 and, recently, Liu et al. [14] developed a quadruple mutant to remove functional redundancy of all HsfA1 genes in Arabidopsis. It was found that more than 65% heat shock up-regulated genes were HsfA1 dependent and, apart from severely reduced thermo tolerance, the quadruple mutant was also defective in overall growth and development, thus confirming the role of HsfA1 as the master regulator. In another recent report, triple and quadruple mutants of 4 HsfA1-type proteins were generated and it was found that HS responsiveness of DREB2A disappeared, indicating again Arabidopsis HsfA1 as the main positive regulator in HS-responsive gene expression [15]. In a genome-wide expression analysis of Hsf gene family in rice, OsHsfA1 showed a stable expression across 28 different tissues, representing major plant developmental stages and abiotic stress conditions [5]. Unlike HsfA1, HsfA2 is expressed only in heatstress induced plants and functions as a regulator in other environment stresses as well [16][17]. HsfA2 is strongly induced upon heat stress in tomato, Arabidopsis and rice [18][19]5] and is found to play a role during heat stress or in recovery after HS [20]. It has also been reported that under oxidative stress, 26 proteosome function and/or heat shock protein (HSP90) activity inhibits, which is a trigger for the HsfA2 expression [21]. HsfA2 is also considered as an important linker between heat and oxidative stress responses [22]. AtHsfA2 knockout mutants display reduction in basal and acquired thermotolerance as well as oxidative stress while, its overexpression leads to increased tolerance under both these stressed conditions. Schramm et al. [23] functionally validated the transcription activation potential of HsfA2 on GUS reporter construct containing target promoters and their deletion analysis to verify HsfA2 binding site. This heat inducible HsfA2, when overexpressed in Arabidopsis, shows more tolerance to anoxia and submergence stress as well [24]. Transgenic Arabidopsis overexpressing OsHsfA2e showed thermotolerance in cotyledons, rosette leaves, inflorescence stems and seeds [25]. Charng et al. [26] reported direct evidence of the biological function of HsfA2 in sustaining the transcript level of HSP genes during prolonged heat stress and recovery. Further, Cohen-Peer et al. [27] reported physical interaction of AtHsfA2 and AtSUMO1, which represses its transcriptional activity. Recently, the role of OsHsfA2c and OsHsfB4b has been implicated in OsClpB-cyt expression [28]. Apart from heat stress, class A HSFs are also induced in Arabidopsis and rice by other abiotic stresses, such as cold, oxidative, salt and drought [5,[29][30]. In rice, spl17 mutant defective in OsHsfA4d, leads to disease mimic spotted leaf phenotype under high temperature [31]. The only known HSF gene from wheat, TaHsfA4a is involved in cadmium tolerance [32]. In Arabidopsis, over-expression of HsfA4a lead to decreased expression of cytosolic H 2 O 2 scavenging ascorbate peroxidase and it was hypothesized that HSFs may act as H 2 O 2 sensors in the plants [33,34]. In plants, class A HSFs have been characterized in more detail than class B and class C HSFs. Tomato HsfB1 is a transcription co-activator, functioning along with HsfA1 and hypothesized as an HS induced factor, essential for maintenance and restoration of housekeeping gene transcription during stress conditions [35]. In Arabidopsis, HsfB1 is heat inducible among all five class B HSF, however, its over-expression had no effect on HSP induction and thermotolerance [19,36]. In rice also OsHsfB1, B2a and B2b were found responsive to heat and oxidative stresses [5,30]. Kumar et al. [37] suggested that class B HSFs act as negative regulator of biotic stress response, since down-regulation of class B Hsf resulted in increased expression of plant defense genes.
Wheat is one of the most important food crops in the world, and its complete genome has not been sequenced. Moreover, compared to rice and Arabidopsis, there have been few studies investigating the stress response associated genes, especially heat stress related genes in wheat, partly because of its hexaploid nature, larger genome size and relatively longer life cycle. In the present investigation, the cloning of a seed preferential heat shock factor, identified through subtractive hybridization [38] and its characterization in transgenic Arabidopsis plants was undertaken.

Plant material, growth conditions and stress treatments
Bread wheat (Triticum aestivum L.) cv. CPAN1676 and Arabidopsis ecotype Col0 were used throughout the present study. For seedling based studies, wheat plants were grown in plastic trays in a growth chamber (Conviron, Canada) and Arabidopsis plants were grown on half-strength MS medium in Petri plates at 20uC and 16 hour photoperiod in daily cycle. For other stages of plant development, plants were grown in potted soil and Soilrite for wheat and Arabidopsis, respectively. For stress treatments, plants were transferred to growth chambers set at specified temperatures, and heat stress provided for different time periods. After heat stress treatment, the control and stressed tissues were sampled, flash frozen in liquid nitrogen and stored at -80uC until RNA isolation. Remaining seedlings were transferred back to the growth chamber for recovery. For developing grains, potted wheat plants grown in departmental garden were transferred for two hours to the growth chambers maintained at 37uC and 42uC at 3,7, 14, 21 and 28 days after anthesis. Sampling of developing grains was done from spikes of control and stressed plants and developing seeds were stored at -80uC until RNA isolation. For germination assay under salt stress, Arabidopsis seeds were germinated on K MS for control condition and K MS supplemented with different concentrations of NaCl in Petri plates. For seedling response under salinity stress, salt stress was provided by immersing two-week-old seedlings in K MS medium supplemented with 150 mM and 300 mM NaCl for stress conditions. Similarly, for drought stress experiment, the seeds were germinated on K MS for control condition and K MS supplemented with 300 mM of mannitol for water stress condition.

Cloning of wheat HSF cDNA and genomic fragments
Total RNA from developing wheat seeds was isolated as described by Singh et al. [39]. For RACE-PCR, previously identified wheat EST [38] showing homology with rice HSF was used as a template for designing the primers. Both 5and 3RACE-PCR was performed using SMART TM RACE amplification kit (Clontech, Palo Alto, USA) as per the manufacturer's instructions. The resultant PCR fragments were then cloned using pGEM-T Easy vector (Promega, USA) and sequenced. For cloning of genomic fragment, primers were designed from cloned full-length cDNA and PCR was conducted by using wheat genomic DNA as template by using Long-PCR Enzyme mix (Fermentas, Lithuania). The PCR products were purified using the PCR purification kit (Qiagen, Germany) and cloned into the pGEM-T Easy vector and then sequenced.

Sequence analysis and phylogenetic relationship
The nucleotide and protein sequences were analyzed using DNA analysis software DNASTAR. Nucleotide and the deduced amino acid sequences searched for their homology with the existing rice sequences in the NCBI database using the BlastN and BlastP programs [40]. Phylogenetic tree was constructed using neighbourhood joining method with boot strap values in CLUSTALX program by using full length protein sequences of different rice HSFs as reported earlier [5].

Expression analysis by Northern and Semi-quantitative RT-PCR
Twenty micrograms of total RNA from control and high temperature treated samples were resolved in 1.2% agarose gel containing formaldehyde and transferred to Hybond N membrane (Amersham Pharmacia Biotech, UK). PCR-amplified cDNA fragment of EST [38] representing wheat Hsf was purified from agarose gel. Probes were labeled with a 32 P-dATP using Megaprime DNA labelling system (Amersham Pharmacia Biotech, UK) and purified through Probquant G-25 column (Amersham Pharmacia Biotech, UK). For expression analysis of wheat HSF in different developing seed tissues, a two-step RT-PCR was employed. An aliquot of 2 mg of RNA from each sample was used to synthesize the first strand cDNA using the SuperScriptIII first strand cDNA synthesis kit (Life Technologies, USA), in a 20-mL reaction volume containing 1x reaction buffer, 1 mL of oligo dT (20) primer (50 mM), 2.5 mM each of dCTP, dTTP, dGTP, dATP, and 200 units of reverse transcriptase SuperscriptIII. After incubation at 50uC for 1 h, RNA was removed by incubating with RNase H at 37uC for 20 min. After RNase H treatment, 80 mL MQ water was added to each tube. 1 mL of this diluted cDNA template was added into each PCR reaction with TaHsf gene-specific primers (Table S1). PCR was conducted with the following program using Taq DNA polymerase (Roche, Germany): initial denaturation at 94uC for 5 min, followed with 94uC for 30 s, 60uC for 30 s, 72uC for 1 min with 25 cycles. A wheat actin was used as internal control (Table S1). The PCR products were checked on 1.2% agarose gel in 1x TAE buffer with Ethidium Bromide.

Transactivation analysis in yeast
For the measurement of the transactivation activity, the complete TaHsf coding sequence fragment amplified by PCR from pGEM-T Easy:TaHsf plasmid was cloned between the NdeI and SmaI sites of vector pGBKT7 (Clontech, USA). The recombinant plasmids were transformed into the yeast strain AH109 harboring the LacZ and HIS3 reporter genes. The HIS3 reporter gene activity was confirmed by a viability test on a medium lacking histidine. The LacZ activity was observed in a medium containing 200 mg/L X-gal.

Construction of binary vectors and plant transformation
For over-expression studies, a 1.39 kb fragment (cDNA) containing the TaHsf open reading frame along with 5' and 3' UTR, was PCR amplified from pGEM-T Easy:TaHsf plasmid with Topo-TaHsf-F and TaHsf-R primers (Table S1). The amplicon was inserted into pENTR/D-TOPO vector (Invitrogen) for construction of entry clone and sequenced. This cassette was then mobilized into the destination vector pMDC32 by LR-clonase reaction to generate pMDC32:TaHsfA2d vector. Arabidopsis thaliana Col0, wild type and T-DNA mutant plants for AtHsfA2/ AT2G26150 (SALK_008978; The Salk Institute Genomic Analysis Laboratory, http://signal.salk.edu/tabout.html) were transformed by floral dip method [41]. For analysis of different transgenic lines under various abiotic stresses, T3 homozygous seeds were used and the results presented represents at least three independent experiments. The data presented represents at least 10 replicates and values are represented with average means of these experiments with standard error bars. P values were calculated by student t-test using functions in MS-Excel and represented in the bar diagrams. Expression analysis by RT-qPCR for HSF target genes Total RNA from different Arabidopsis plants was isolated using the RNeasy plant mini kit (Qiagen, Germany) according to the manufacturer's instructions, including on-column DNaseI treatment to remove genomic DNA contamination. Twomg of the total RNA was used as template to synthesize cDNA employing the High Capacity cDNA Archive kit (Applied Biosystems, USA) and mixed with 200 nM of each primer and SYBR Green PCR Master Mix (Applied Biosystems) for real-time PCR analysis, using the ABI Prism 7000 Sequence Detection System and Software (PE Applied Biosystems) according to the manufacturer's protocol. The specificity of the reactions was verified by melting curve analysis. At least two independent RNA isolations were used for cDNA synthesis, and each cDNA sample was subjected to real-time PCR analysis in triplicate. The primer sequences for AtHsfA2 target genes were taken from an earlier report related to characterization of AtHsfA2 target genes [16].

Chlorophyll Fluorescence Measurements
PSII activity was measured according to Krause and Weis [42]. Measurements of modulated chlorophyll fluorescence emission from the upper surface of the leaf were made using a pulse amplitude modulation fluorometer (Junior-PAM chlorophyll fluorometer, H.Walz, Germany). Leaves were dark-adapted for 20 min before measuring the induction of fluorescence. Measurements of various PSII functions such as maximum photosynthetic efficiency (Fv/Fm), effective photosynthetic efficiency (YII) and Electrone Transport Rate (ETR) were recorded in rosette leaves at specified time points in at least ten plants per line viz. WT, mutant and transgenics.

Measurements of total chlorophylls
Chlorophylls were estimated by the non-maceration method according to Hiscox and Israelstam [43]. Leaf samples (0.05 g) from control and salt stress treated plants in five different replicates were incubated in 5 mL of DMSO at 65uC for 4 hr in dark. Absorbance was recorded at 645, 663nm in a (Beckman DU TM 640B, Beckman Instruments Inc., USA) spectrophotometer, and chlorophyll contents were calculated according to the following formula where, V = volume of DMSO in mL, X = path length, 1 cm,and W = fresh weight in grams.

Identification and cloning of seed preferential wheat Hsf
Previously, we had generated a large number of wheat ESTs through subtractive hybridization of RNA extracted from control and heat stressed tissues of three different growth stages of wheat, viz. seedling, flower and developing grains [38]. One such EST showing high homology with rice Hsf, was identified from developing seed tissue. We checked the tissue specificity of this transcript by RNA gel blot analysis. Fig 1A shows that this wheat Hsf is expressed only in developing seed tissues and absent from shoot and root tissues. Although, a significant amount is also present in the control tissue, its transcript was further enhanced by heat stress at 37uC and 42uC. We then checked the inducibility of this gene at different stages of wheat grain development under control and heat stress conditions. This wheat Hsf is induced by heat stress in the early stages of seed development, and gradually become constitutive as the seed matures (Fig. 1B).
To complete the full length cDNA, 5' and 3' RACE-PCR (Fig.  S1) was performed and a full length cDNA of 1396 bp was cloned harboring 1026 bp ORF along with 5' UTR of 120 bp and 3' UTR of 205 bp, along with poly-A tail. The complete nucleotide and deduced amino acid sequence was submitted to NCBI with accession number KF061193. Subsequently, a genomic clone of 2682 bp was also obtained by performing PCR using genomic DNA as template (Fig. S1). The cDNA and genomic clone alignment revealed the presence of a single intron of 1256 bp (Document S1). The deduced amino acid sequence of 341 residues was analyzed by BLAST in NCBI and TIGR databases and this transcript was confirmed as a putative HSF. To classify this wheat HSF, a phylogenetic tree was constructed by using all 25 rice HSFs. This new wheat HSF segregated as a separate clade in the resultant phylogenetic tree along with other rice class A2 type of HSFs, OsHsfA2d being the closest (Fig. 2). This was further confirmed by BLASTP analysis in NCBI and TIGR databases. TaHsfA2d shows a similarity of 69% with rice (Oryza sativa japonica) OsHsfA2d (Os03G06630) and a similarity of 65% with Arabidopsis HsfA2. With this analysis this wheat HSF was termed as TaHsfA2d.

TaHsfA2d possesses transactivation activity
The amino acid alignment of TaHsfA2d with the corresponding rice homolog revealed the characteristic features of class A type HSFs, such as the presence of an N-terminal DNA binding domain, middle part containing HR-A/B, followed by potential NLS and NES and a C-terminal part containing AHA type transactivation motif (Fig. 3A). In order to test the transactivation potential of TaHsfA2d, we used the yeast assay system. The complete ORF of TaHsfA2d was cloned in vector pGBKT7 with GAL4 DBD fusion and introduced into the yeast reporter strain AH109. The HIS3 reporter gene activity was confirmed by a viability test on histidine lacking medium, while yeast harboring pGBKT7::TaHsf2d grows on histidine lacking medium, the empty vector control showed no growth (Fig. 3B). Thus, it was hypothesized that TaHsfA2d has a functional AHA motif and acts as trans-activator similarly to other class A Hsfs.

Overexpression of TaHsfA2d increases basal thermotolerance of transgenic Arabidopsis plants
To determine the agronomical usefulness of TaHsfA2d in molecular breeding, the WT and mutant hsfa2 plants of Arabidopsis were transformed with TaHsfA2d. The expression of native and transgene was confirmed by RT-PCR in T 2 transgenic and WT seedlings ( Fig. 4A and Figure S2). We selected and tested mutant plants for expression of AtHsfA2 after a heat shock of 30 minutes at 42uC (Fig. 4A). This was done to select the knockout (KO) hsfa2 line (nhsfa2) and it was then complemented with TaHsfA2d by transformation. Subsequently, the effect of a sudden heat shock of 42uC in two-week-old plants was studied by measuring the maximum photochemical efficiency (Fv/Fm) of PSII at various time points. Fig. 4B shows that sudden heat shock caused a gradual decrease in PSII activity in terms of maximum photochemical efficiency (Fv/Fm) in all the plants, however, TaHsfA2d overexpressing plants represented by two independent lines viz. OE6 and OE7 maintained a higher Fv/Fm at all the time points and showed a faster and better recovery upon returning to control temperature. In contrast, the WT and mutant plants showed drastic reduction in PSII activity right from initial phase of stress treatment, while WT plants could recover well after 2 days, mutant plants showed no measurable PSII activity.
In a separate experiment, two-week-old seedlings grown on K strength MS medium were transferred in small plastic tray pots in Soilrite. After establishment in soil the trays were shifted to a growth chamber maintained at moderate high temperature of 30uC for two weeks. In this moderate heat stress environment also, the overexpressing plants showed better biomass accumulation and faster growth as measured by fresh weight and number of rosette leaves (Fig. 5A). This is also evident from the steady state PSII activity measured as effective photosynthetic efficiency (YII) and ETR. Transgenic Arabidopsis plants overexpressing wheat TaHSFA2d showed higher PSII activity under constant heat stress conditions, showing less reduction in both YII and ETR (Fig. 5B). The mutant plants either almost ceased growth and development upon transfer to non-lethal heat stress or died, while overexpression plants had almost double the biomass than the WT plants. To further check the thermotolerance of different Arabidopsis plants during growth and development throughout their life cycle, seeds were germinated in three inch plastic pots with 15 seeds (after germination) in each pot, and kept continuously at 30uC temperature in a growth chamber. In this assay also, the overexpression transgenic plants showed faster germination and seedling establishment. Although all the plants matured in around 8 weeks, the overexpression transgenic plants accumulated more biomass, had longer and higher number of productive siliques and showed considerably higher seed yield (Fig. 6). In comparison, the KO mutant hsfa2 plants showed very poor growth and produced almost half the seeds as compared to WT plants and the yield was much less when compared to overexpression transgenics. This analysis showed that for most of the parameters the overexpression plants and complimented mutant plants differed significantly from WT and mutant plants, whereas WT and mutant plants behaves in a similar manner under various stresses.

Overexpression of TaHsfA2d provides salt and drought tolerance to transgenic Arabidopsis
TaHsfA2d transgenic Arabidopsis plants were tested for salt stress tolerance in both germinating and vegetative stages at different concentrations of NaCl. The germination potential of different Arabidopsis seeds viz., WT, KO, OE and KO-Complemented was checked by germinating them on K strength MS medium supplemented with 150 and 300 mM NaCl. OE plants showed better and faster germination of 80% as compared to 60% and 35% for WT and KO seeds respectively. The difference in per cent germination was even greater under higher salt stress condition, where OE seeds maintained the similar germination potential while only 12% of the KO plants could germinate (Fig.  7A). The salt tolerance in vegetative tissues was checked by using two-weeks-old seedlings grown on K strength MS media. Twoweeks-old plants of Arabidopsis were then dipped in K strength MS medium supplemented with NaCl. After four days observations were taken in terms of greening of the tissue and total chlorophyll content ( Fig. 7B and 7C). The WT and KO seedlings showed a faster depletion of photosynthetic pigments than the OE plants. After 4 days exposure to salt stress, OE seedlings still maintained almost half of the original total chlorophyll content, while WT and mutant plants were almost completely devoid of chlorophyll (Fig.  7C). In a different set of experiment, we also assessed the response of different Arabidopsis plants towards drought stress simulated by 300 mM of mannitol. As shown in Figure 8, transgenic Arabidopsis plants of both the overexpression lines performed much better than both WT and mutant plants. There was a faster and robust germination in overexpressing plants and subsequently more biomass accumulation and growth was observed as measured by plant height, root-shoot length and rosette diameter (Fig. 8A and  8B).

Constitutive expression of AtHsfA2 target genes in transgenic Arabidopsis overexpressing TaHsfA2d
To understand the molecular basis of TaHsfA2d mediated stress tolerance in transgenic Arabidopsis, we checked the expression of AtHsfA2 target genes in WT, KO, OE and complemented KO plants by quantitative real time PCR. These genes were carefully selected on the basis of previous reports related to target genes of AtHsfA2 in Arabidopsis [16,24]. For this analysis, heat stress was given at 42uC for 20 minutes only, as Arabidopsis class A HSFs may have functional redundancy in activating these genes. The expression of 21 target genes was examined in three different tissues, viz. seedling, flower and siliques under normal growth conditions and on exposure to heat stress conditions in these tissues (Fig. 9). These genes broadly represent different abiotic stress related genes, such as those coding for heat shock proteins (HSPs), reactive oxygen scavengers and secondary metabolites. The highest fold-change was observed for a chloroplast localized small heat shock protein (sHSP26.5); siliques showed the maximum expression, followed by seedling and flower (Fig. 9). Almost similar expression pattern was observed in case of other organelle localized sHSPs, such as ER localized sHSP22 and mitochondria localized sHSP23.5. Many of the target genes showed higher expression in siliques tissues only, such as Sti1 which is exclusively present in siliques tissue. This can be explained by the fact that TaHsfA2d was primarily identified and cloned from developing wheat seed tissues. It is noteworthy that not all the target genes have higher expression in OE plants as compared to the WT plants such as HSP70B and GolS2, suggesting that these genes may not be the primary target of TaHsfA2d in wheat. Overall, mutant plants showed little or no expression under heat stress conditions in all the three tissues in terms of relative fold change; it was also reflected by their higher Ct values when compared to WT and OE plants in the assay (data not shown).

TaHsfA2d can partially complement Arabidopsis KO hsfa2 plants
It is clearly evident from the above analysis, that Arabidopsis KO plants for hsfa2 are defective in heat and salt stress response. Since TaHsfA2d showed 65% homology with AtHsfA2 at the amino acid level, we tried to complement the KO plants with wheat TaHsfA2d. The complemented plants performed better than KO plants in all the experiments conducted in the present study and their response was almost at par with the WT plants ( Fig. 4-8). However, we expected these plants to perform better than WT plants as the expression of the transgene, TaHsfA2d is governed by a strong, constitutive promoter CaMV35S. The partial recovery displayed by these complemented KO plants can be explained on the basis of relatively lower expressions of the target genes in transgenics as shown in Fig. 9. Almost all the genes showed a lower expression in terms of relative fold change than the corresponding WT and OE tissue. One exception to this is the expression of Sti gene in silique tissue. This discrepancy in the expression values of WT and complemented KO plants can be explained by the possible different target gene specificities of TaHsfA2d, as the wheat TaHsfA2d expresses preferentially in seed and its transcript

Discussion
Heat shock factors have been functionally analyzed in many plants, such as Arabidopsis, rice, tomato and sunflower [4, 20, 25, 31, and 44]. However in wheat, only one heat shock factor, TaHsfA4d, has been functionally characterized so far [32]. In this study, we cloned and characterized a class A Hsf from wheat, TaHsfA2d, from developing wheat seeds. Through northern analysis we found that this Hsf is seed specific. Although, previous expression analysis based on EST showed that TaHsfA2d is heat inducible in both shoot and roots, nevertheless, the expression of TaHsfA2d was much higher in flower and developing seed under heat stress conditions [38]. Interestingly, in our previous study we have shown that among various class A HSFs in rice, OsHsfA2d also showed induced expression in developing rice seeds [5]. Apart from heat stress, we found that TaHsfA2d is also inducible by salt stress and by ABA [38]. Further, we demonstrate in this study that this TaHsfA2d functions as the transcriptional activator in yeast (Fig. 3). Unlike class B and class C Hsfs, class A Hsfs are known to be trans-activators due to the presence of AHA motif. We found that TaHsfA2d has a functional AHA motif and acts as transactivator similarly to other class A Hsfs.
There are 21 Hsfs in Arabidopsis, 25 in rice and 24 in Brachipodium, thus we expect 70-75 Hsfs in wheat owing to its hexaploid nature and presence of three homeologous genomes. While both Arabidopsis and tomato have a single HsfA2, rice has as many as 5 different HsfA2 (a-e) [4,12,30,45]. In both Arabidopsis and tomato, HsfA2 is highly inducible by abiotic stresses and constitutive up-regulation of this single Hsf in transgenic plants  improved the tolerance to high temperature, while its down regulation showed the negative effect [13,22,26]. Especially in Arabidopsis, overexpression of AtHsfA2 increased tolerance to several types of environmental stresses, such as high temperature [19,22,23,26], oxidative stress [16] and anoxia stress [24]. In rice also, different HsfA2 were found to be heat inducible and in addition, OsHsfA2d expressed in developing seeds [5]. Overexpression of OsHsfA2e increases the heat and salinity stress tolerance in Arabidopsis [25]. In the present investigation, Arabidopsis plants overexpressing TaHsfA2d showed remarkable tolerance to moderate high temperature provided continuously (Fig. 5 and 6). This could be due to the constitutive expression of several abiotic stresses related genes in the transgenic plants (Fig. 9). It has been shown earlier that overexpression of AtHsfA1a and A1b improved basal thermotolerance in transgenic Arabidopsis plants [46]. In the present study, a positive effect was also found on seed yield and silique length as well as silique number under moderate heat stress conditions (Fig. 6). It is known that many heat shock proteins are accumulated during seed maturation [47]. We also found higher expression of some HSPs in the siliques of transgenic Arabidopsis plants overexpressing TaHsfA2d (Fig. 9) such as HSP 26.5P, HSP 101, HSP70-T2, HSP 23.6-M and HSP17. Kotak et al. [48] found that accumulation of HSPs in developing seeds is regulated by AtHsfA9 in Arabidopsis. Whether TaHsfA2d also governs accumulation of HSPs in wheat developing grains remains to be elucidated.
Overexpression of TaHsfA2d in Arabidopsis also increased plant tolerance against salinity and drought stress, as reflected by better germination, higher chlorophyll retention and biomass accumulation ( Fig. 7 and 8). This can again be explained by the constitutive expression of several salt/osmotic stresses related genes in transgenic plants (Fig. 8). Sun et al. [49] reported that overexpression of molecular chaperon HSP17.7CII increased salt and osmotic stress tolerance. Overexpression of GolS1 increased water stress tolerance in Arabidopsis through accumulation of higher amounts of raffinose [50]. In rice, it has been shown that myo-inositol 1-P-synthase (MIPS) gene has important protective functions under salinity stress [51,52]. We have recently found that in heat stress also, the levels of myo-inositol increase significantly and which is also correlated well with TaMIPS expression in wheat [53]. It is therefore likely that salinity stress tolerance showed by Arabidopsis plants overexpressing TaHsfA2d is due to higher expression of these genes (Fig. 9).
Target genes for HsfA2 were identified through microarray analysis in transgenic plants overexpressing Arabidopsis HsfA2 [16] and rice HsfA2e [25]. It has been found that Arabidopsis HsfA2 governs expression of 46 genes, which were constitutively expressed in transgenic plants. Similarly, 37 genes were constitutively up-regulated in transgenic Arabidopsis plants overexpressing rice OsHsfA2e. There is a fair similarity in the diversity of genes governed by these two Hsf A2. On the basis of these reports we checked the expression of 21 target genes of HsfA2. These genes mainly represent heat shock proteins, reactive oxygen scavengers and other stress related proteins. We found that 16 out of 21 genes examined, have constitutively higher expression in seedling, flower and siliques of transgenic Arabidopsis plants overexpressing TaHsfA2d. A chloroplastic sHSP gene showed remarkably high expression in all the tissues tested (Fig. 9). Recently, we have shown that over-expression of a wheat chloroplastic sHSP26 improves photosynthetic efficiencies under heat stress conditions in transgenic Arabidopsis plants and also take part in seed development [54]. It is worth mentioning here that in our previous report [38], we reported many seed specific genes induced by heat stress, are also constitutively expressed in transgenic Arabidopsis plants overexpressing TaHsfA2d, such as sHSPs, HSP70, HSP101 and Sti. It is therefore likely that heat, drought and salt stress tolerance of the transgenic Arabidopsis plants overexpressing TaHsfA2d is provided by constitutive expression of these stress related genes.
Arabidopsis plants overexpressing TaHsfA2d, which preferentially expresses in developing wheat grains, showed improved heat and salt stress tolerance at various growth stages, and higher yields under moderate heat stress conditions. These findings suggest that TaHsfA2d gene may be useful in molecular breeding of wheat for enhanced thermotolerance especially during terminal heat stress. Figure S1 Cloning of wheat Hsf cDNA and Genomic clones. 5 and 3 RACE-PCR was done by using RNA from developing seeds from the heat stressed plants. Both 5and 3RACE-PCR was performed using SMART TM RACE amplification kit (Clontech, Palo Alto, USA) as per the manufacturer's instructions. The resultant PCR fragments were then cloned using pGEM-T Easy vector (Promega, USA) and sequenced. For cloning of genomic fragment, primers were designed from cloned full-length cDNA and PCR was conducted by using wheat genomic DNA as template by using Long-PCR Enzyme mix (Fermentas, Lithuania). The PCR products were purified using the PCR purification kit (Qiagen, Germany) and cloned into the pGEM-T Easy vector and then sequenced.