Transcriptional regulation of heat shock proteins and ascorbate peroxidase by CtHsfA2b from African bermudagrass conferring heat tolerance in Arabidopsis

Heat stress transcription factor A2s (HsfA2s) are key regulators in plant response to high temperature. Our objectives were to isolate an HsfA2 gene (CtHsfA2b) from a warm-season grass species, African bermudagrass (Cynodon transvaalensis Burtt-Davy), and to determine the physiological functions and transcriptional regulation of HsfA2 for improving heat tolerance. Gene expression analysis revealed that CtHsfA2b was heat-inducible and exhibited rapid response to increasing temperature. Ectopic expression of CtHsfA2b improved heat tolerance in Arabidopsis and restored heat-sensitive defects of Arabidopsis hsfa2 mutant, which was demonstrated by higher survival rate and photosynthetic parameters, and lower electrolyte leakage in transgenic plants compared to the WT or hsfa2 mutant. CtHsfA2b transgenic plants showed elevated transcriptional regulation of several downstream genes, including those encoding ascorbate peroxidase (AtApx2) and heat shock proteins [AtHsp18.1-CI, AtHsp22.0-ER, AtHsp25.3-P and AtHsp26.5-P(r), AtHsp70b and AtHsp101-3]. CtHsfA2b was found to bind to the heat shock element (HSE) on the promoter of AtApx2 and enhanced transcriptional activity of AtApx2. These results suggested that CtHsfA2b could play positive roles in heat protection by up-regulating antioxidant defense and chaperoning mechanisms. CtHsfA2b has the potential to be used as a candidate gene to genetically modify cool-season species for improving heat tolerance.


Materials and Methods
Growth conditions and stress treatment. Plants of African bermudagrass were collected from field plots at Nanjing Agricultural University and were propagated from stolons with at least one node per segment and placed into a hydroponic system with half-strength Hoagland's nutrient solution 19 on September 2, 2013. The nutrient solution was changed weekly to maintain adequate nutrient supply and aerated using air pumps to provide sufficient oxygen to the plants. Clonal plants were maintained in a growth chamber at 30/25 °C (day/ night), 14 h photoperiod, 65% humidity and 650 μ mol m −2 s −1 photosynthetically active radiation for two weeks to allow for establishment.
To induce expression of heat-responsive genes, plants were subjected to heat stress at 42 °C for 6 h followed by recovery at 30 °C on September 15, 2013. Leaves and roots of African bermudagrass were collected at 0, 0.5, 1, 3, 6, 8, 12 and 24 h and flash frozen in liquid nitrogen for RNA extraction.
Isolation of CtHsfA2b coding sequence and gene expression analysis. Due to the rapid response of HsfA2 genes to heat stress 16 , African bermudagrass leaves exposed to 42 °C for 1 h were used for gene isolation. Total RNA was extracted using Tripure Isolation Reagent Kit (Roche Diagnostic, Basel, Switzerland) and cDNA was synthesized by a Transcriptor First Strand cDNA Synthesis Kit (Roche Diagnostic, Basel, Switzerland). One pair of primers (CtHsfA2b-F/R, see Supplementary Table S1) was used to isolate the 3′ region of CtHsfA2b cDNA sequence according to the expressed sequence tag (EST) database of bermudagrass 20 . The remaining sequence was obtained by using a SMARTer ® RACE 5′ /3′ Kit (Clontech Laboratories, Mountain View, CA, USA) with specific primers CtHsfA2b-GSP and -NP. The open reading frame (ORF) of CtHsfA2b was amplified using CtHsfA2b-S and CtHsfA2b-A primers. All positive clones confirmed by colony PCR were sequenced by Sangon Co. Ltd (Shanghai, China). Primers were designed by using Premier Primer 5.
For qRT-PCR analysis, total RNA was extracted using RNApure fast isolation Kit (YuanPingHao, Tianjin, China) and first strand cDNA was synthesized by the PrimeScript RT reagent Kit with gDNA Eraser (Perfect Real Time) (Takara, Otsu, Japan). qRT-PCR was performed with SYBR Green I Master reaction system (Roche Diagnostic, Rotkreuz, Switzerland) on Roche LightCycler480 II (Roche Diagnostic, Rotkreuz, Switzerland). Each sample was performed with three biological replicates, each with three technical replicates. Relative expression level of genes was determined by 2 −ΔΔCT method using CtEF1α 21 and AtActin2 as respective reference genes of African bermudagrass and Arabidopsis. Primers for qRT-PCR are listed in Supplementary Table S1.
Plant expression vector construction and transformation. The ORF of CtHsfA2b with Sal I and Not I sites was first linked to a pENTR1A Dual Selection Vector and then transformed into a pEarleyGate 103 22 destination plasmid using LR Clonase II enzyme mix (Invitrogen, Carlsbad, CA, USA). The destination plasmid was transformed into 8 Arabidopsis plants (T0) mediated by Agrobacterium tumefaciens using the floral dip method 23 .
For screening the first generation of transgenic lines (T1), more than 600 seeds collected from transformed Arabidopsis were sown on MS medium with 20 μ g ml −1 glufosinate ammonium in October, 2013. Seventeen positive transgenic lines out of 20 surviving T1 seedlings were confirmed by PCR analysis using primers CtHsfA2b-S and pEarleyGate103-R. Hundreds of T2 seeds collected in January 2014 from T1 plants were also screened for green fluorescent protein (GFP) observation to remove the non-transgenic plants that derived from allele segregation. T2 resistant-seedlings that expressed the CtHsfA2b gene were used for further analysis.  Confirmation of CtHsfA2b functions in improving heat tolerance in Arabidopsis. Arabidopsis ecotype Colombia was used as the wild type (WT). The HsfA2 (At2g26150) T-DNA insertion mutant SALK_008978 (hsfa2) 24 was obtained from the Arabidopsis Biological Resource Center (ABRC; Ohio State University). The T-DNA insertion and homozygous line were confirmed by PCR amplification with primer LP, RP and LBp1.3 (see Supplementary Table S1) and designed according to the SIGnAL website (http://signal.salk.edu/tdnaprimers.2.html).
To compare heat tolerance among WT, hsfa2 and transgenic lines, Arabidopsis seeds were grown on Murashige and Skoog (MS) Medium under a 12-h photoperiod at 25 °C for 6 d. One set of the seedlings was used for protein localization using GFP observation and the examination of survival rates under heat treatment. Another set of plants was transferred to plastic pots (7 cm diameter and 7 cm depth) filled with a mixture of 25% vermiculite and 75% peat moss and maintained in a growth chamber at 25/20 °C (day/night), 14 h photoperiod, 65% humidity and 150 μ mol m-2 s-1 photosynthetically active radiation for further examination of heat responses in mature plants and for gene expression analysis.
Physiological measurements of Arabidopsis. 30-day-old wild type and 35S:CtHsfA2b transgenic Arabidopsis were treated with heat stress of 45/40 °C (day/night) for 72 h, and physiological measurements were performed after 7 d of recovery. Leaf net photosynthetic rate (Pn) was measured using a Licor 6400 photosynthetic system (LI-6400, Li-Cor, Lincoln, NE, USA). Leaf photochemical efficiency was expressed as the variable fluorescence (Fv) to the maximal fluorescence (Fm) ratio (Fv/Fm) with a fluorescence induction monitor (OPTI-Sciences, Hudson, NY, USA) on dark-adapted leaves. Chlorophyll content was measured using a spectrophotometer (GE Healthcare Life Sciences, Cambridge, UK) after leaves were soaked in dimethylsulfoxide for 48 h according to previous methods 25 . Cellular membrane stability was estimated based on leaf electrolyte leakage (EL) 26 . The initial (Ci) and maximum (Cmax) levels of EL were measured using a conductance meter (Thermo Scientific, Beverly, MA, USA). Relative EL was calculated as EL = (Ci/Cmax) × 100.
Subcellular localization of CtHsfA2b. The pENTR1A Dual Selection Vector containing CtHsfA2b was LR-recombined with a small binary vector p2GWF7.0 for fusion-expression of CtHsfA2b and enhanced green fluorescent protein (eGFP) in Arabidopsis protoplasts. Healthy and mature leaves of 4-week-old Arabidopsis were used for protoplast isolation. Arabidopsis protoplast isolation and transfection were performed according to a modified method 27 . The transfected protoplasts were mixed with DAPI solution (Yeasen Biotech Co., Ltd, Shanghai, China) to a final concentration of 10 μ g/ml for a five-minute incubation at room temperature and then observed under a Confocal Laser Scanning Microscope (Carl Zeiss, Jena, Germany). In order to verify the permanent subcellular localization of CtHsfA2b, root tips of 3-day-old transgenic Arabidopsis seedlings grown on agar medium were cut and placed on slides for GFP observation using a fluorescence microscope BX53 (Olympus, Tokyo, Japan).

Yeast one-hybrid (Y1H) analysis of CtHsfA2b binding to HSE.
In order to determine DNA binding-ability of CtHsfA2b to HSEs on the promoter of AtApx2 (ProAtApx2), an Y1H assay was performed according to manufacturer's instructions using a One-Hybrid Library Construction & Screening Kit (Clontech Laboratories, Mountain View, CA, USA) with modification of pGADT7-Rec2 replaced by modified pGADT7 vector that had LR-recombination construction. For bait construction, three tandem HSE (TTTCtgGAAG) elements of the AtApx2 promoter were linked to a pHis2.1 vector with EcoR I and Sac I sites and confirmed by sequencing with primers of pHis2.1-F/R. For the construction of the prey, the coding sequence of CtHsfA2b was LR-recombined into the pGADT7 vector. The bait pHis2.1-HSE and prey pGADT7-CtHsfA2b then were co-transformed into yeast strain Y187. Interaction between bait and prey was determined by cell growth on SD medium lacking Trp, Leu, and His (SD-WLH). Positive yeast clones verified by PCR were diluted by distilled water and potted on SD-WLH medium containing 50 mM 3-amino-1,2,4-triazole (3-AT, a competitive inhibitor of the HIS3 protein).

Luciferase (LUC) assay of the binding of CtHsfA2b and ProAtApx2 in vivo.
To further confirm the binding of CtHsfA2b and ProAtApx2 in vivo, the coding sequence of CtHsfA2b was transferred from a pENTR1A vector to a pMDC32 vector via a LR Clonase reaction. ProAtApx2, obtained with primers ProAtApx2-LUC-F/R, was introduced into a pCAMBIA 1381Z-LUC 28 vector with BamH I and Pst I sites. Then, pMDC32-CtHsfA2b and pCAMBIA 1381Z-LUC-ProAtApx2 were transformed into Arabidopsis protoplasts with same transfection method described above. Beetle Luciferin Potassium Salt (Promega, Madison, Wisconsin, USA) was dissolved in sterile water to a final concentration of 7.8 mM and stored at − 80 °C in separate aliquots. The stock solution was diluted 10 times before use. 10 μ l of diluted beetle luciferin was mixed with 100 μ l transfected protoplasts in each well of a 12-well strip. After incubation at 37 °C in the dark for 15 min, the luminescence was captured in a CCD Video Camera Imaging System (Vilber Lourmat, Marne la Vallée, France) and quantified by an ELIASA, Infinite 200 PRO (Tecan, Mannedorf, Switzerland).
Detection of reactive oxygen species. Detection of superoxide free radicals was performed as described previously 29 with minor modifications. Fifteen-day-old seedlings exposed to 25 °C or 37 °C for 1 h were harvested and then vacuum-infiltrated with 0.1 mg ml −1 nitroblue tetrazolium (NBT; Sigma,USA) in 25 mM HEPES buffer (pH 7.6). Samples were subsequently incubated at room temperature for 2 h in darkness. Then, chlorophyll was removed with 95% alcohol and photographs were taken with using a stereomicroscope (Olympus, Tokyo, Japan).
The accumulation of Hydrogen peroxide (H 2 O 2 ) in Arabidopsis leaves was detected by 3,3′ -diaminobenzidine (DAB; Sigma, USA) staining as described previously 29 with minor modifications. Samples of the same batch of plants used for superoxide detection were vacuum-infiltrated with 0.1 mg ml −1 DAB in 50 mM Tris-acetate buffer (pH 5.0). Samples in DAB solutions were incubated at room temperature in the darkness for 24 h. Thereafter, chlorophyll was removed, and photos were taken as described above.

Results
Heat-inducible CtHsfA2b gene in African bermudagrass. According to the bermudagrass EST database, we obtained a full-length sequence of HsfA2 gene that consisted of 1386 bp nucleotides with a 1095 bp ORF encoding a 365-residue polypeptide via the RACE PCR method from African bermudagrass. The gene was classified into A2b subgroup of HSF by cluster analysis (Fig. 1a) and named African bermudagrass HsfA2b (CtHsfA2b, GenBank accession number: KU156630). Homologous alignment of the deduced amino acid sequences (Fig. 1b) revealed that CtHsfA2b had high sequence identity to HsfA2b of Setaria italica, Oryza sativa and Brachypodium distachyon, and it had a DNA binding domain (DBD), an oligomerization domain (OD) with a hydrophobic region (HR-A and B), a nuclear localization signal (NLS), a nuclear export signal (NES) and an activator motif (AHA motif).
qRT-PCR analysis demonstrated that CtHsfA2b exhibited rapid response to heat stress (Fig. 1c), as the expression level increased within 30 min and reached the highest level by 1 h of 42 °C heat treatment in leaves and roots of African bermudagrass. After reaching the highest expression level at 1 h, the gene expression attenuated in both leaves and roots and returned to the initial levels following recovery at 30 °C normal growth temperature.

GFP-fused CtHsfA2b protein localized in nucleus.
In order to determine the cellular localization of CtHsfA2b, we first constructed a p2GWF7.0-CtHsfA2b-eGFP plasmid and transformed the plasmid into Arabidopsis mesophyll protoplasts. The completely-overlapped fluorescence signals of DAPI and GFP demonstrated that CtHsfA2b was localized to the nucleus (Fig. 2a). After obtaining CtHsfA2b transgenic Arabidopsis with the plant expression vector pEarleyGate 103 that also has a GFP encoding region, we further confirmed that CtHsfA2b protein was localized in the nucleus by scanning the roots of transgenic plants (Fig. 2b). Improved Arabidopsis heat tolerance through ectopic expression of CtHsfA2b. To characterize the function of CtHsfA2b involved in heat tolerance, we transformed this gene into Arabidopsis (termed as 35S:CtHsfA2b) for ectopic expression analysis. 7-day-old seedlings of transgenic lines (35S:CtHsfA2b-a, b and c) had higher survival rate (76%, 64% and 50% respectively) than the WT (28%) after 7 d recovery from 4 h heat stress at 45 °C (Fig. 3a). qRT-PCR analysis confirmed that CtHsfA2b was expressed only in the transgenic lines, but not in the WT (Fig. 3b). GFP-fused CtHsfA2b protein was also expressed in 35S:CtHsfA2b transgenic lines (Fig. 3c). GFP exhibit bright green fluorescence in roots, but was not as obvious in leaves due to the interference by chlorophyll fluorescence.
Ectopic expression of CtHsfA2b also improved tolerance to heat stress and recuperative ability of adult Arabidopsis after relief of heat stress. Leaves of both 35S:CtHsfA2b transgenic lines and WT wilted after 3 d of 45/40 °C (day/night) heat stress, but the transgenic lines generated new leaves after returning to normal growth temperature for a week, whereas, the WT almost completely withered up and died with a significantly lower survival rate (Fig. 4a and see Supplementary Fig. S1). Net photosynthetic rate, photochemical efficiency and chlorophyll content of 35S:CtHsfA2b transgenic lines were significantly higher than that of WT after heat treatment, which demonstrated more-stable photosynthetic system in the transgenic plants (Fig. 4b). In respect to cell membrane stability, 35S:CtHsfA2b had less relative electrolyte leakage compared to the WT after heat stress. The results revealed that AtApx2, AtHsp18.1-CI, AtHsp22.0-ER, AtHsp25.3-P, AtHsp26.5-P(r), AtHsp70b and AtHsp101-3 were expressed in transgenic lines but not expressed or expressed at lower levels in the WT before heat stress (Fig. 5). After treatment at 37 °C, these genes were all up-regulated both in the WT and transgenic plants, but there was a greater extent of up-regulation of AtApx2, AtHsp18.1-CI, AtHsp22.0-ER, AtHsp25.3-P and AtHsp26.5-P(r) in transgenic lines compared to the WT. 35S:CtHsfA2b-a, one of the transgenic lines, had equal or lower expression levels of AtHsp70b and AtHsp101-3 than the WT, while 35S:CtHsfA2b-b had greater up-regulation of these two genes. All the genes exhibited typical heat response characteristics in that the genes were induced by high temperature rapidly increased to the highest levels and then decreased to reduced levels within 2 h of heat treatment.

Ectopic expression of CtHsfA2b rescued heat-sensitive defects of Arabidopsis hsfa2 mutant.
The positive roles of CtHsfA2b in improving plant tolerance to heat stress were further confirmed by ectopically expressing CtHsfA2b in the Arabidopsis hsfa2 mutant (termed as 35S:CtHsfA2b/hsfa2). The hsfa2 mutant plants were severely injured with a lower survival rate (23%) than the WT (44%) after heat stress (Fig. 6a). However the 35S:CtHsfA2b/hsfa2 transgenic lines rescued hsfa2 heat-sensitive defects and had a significantly higher survival rate (53-86%) than the WT. Results of qRT-PCR and GFP observation showed that the CtHsfA2b gene and its GFP-fused protein were expressed in 35S:CtHsfA2b/hsfa2, but not in the hsfa2 mutant and WT (Fig. 6b,c).
Adult seedlings grown in the soil suffered heat stress and were also examined, which confirmed that the heat treatment at 45/40 °C (day/night) for 60 h was lethal to the hsfa2 mutant plants, but had less damage to WT and 35S:CtHsfA2b/hsfa2, since hsfa2 mutant plants did not recover even after relief of heat stress (Fig. 6d and see Supplementary Fig. S1). The survival rate of hsfa2 mutant after 10 d recovery was significantly lower than WT and the transgenic lines. Ectopic expression of CtHsfA2b rescued the heat-sensitive phenotype of hsfa2 mutant to the WT level at the mature seedling stage.
The expression of downstream genes was examined at 1 h of heat stress (37 °C). Under the control temperature condition (25 °C), these genes were all expressed in 35S:CtHsfA2b/hsfa2, but most were not in the WT and hsfa2 mutant plants, except AtHsp101-3 (Fig. 7). All these genes were up-regulated in the WT, hsfa2 and 35S:CtHsfA2b/ CtHsfA2b binds the HSE on the promoter of AtApx2. We detected the ROS-scavenging ability that includes APX contribution under heat stress. The accumulation of superoxide radical and hydrogen peroxide in WT plants was higher than that in transgenic plants after 1 h of heat stress (Fig. 8a). Transcription levels of AtApx2 were also up-regulated by ectopic expression of CtHsfA2b in Arabidopsis, as previously describe in Figs 5 and 7. Therefore, we examined CtHsfA2b potential binding to ProAtApx2 in vitro by using an Y1H assay with the typical active HSE motif on ProAtApx2 (Fig. 8b) as bait. The results showed that CtHsfA2b could bind to the HSE on ProAtApx2, and then triggered the activation of the HIS reporter-gene expression on SD medium with WLH deficiency and even with 50 mM 3-AT added (Fig. 8c).
To further confirm the function of CtHsfA2b in regulation of ProAtApx2 in vivo, we performed a luciferase (LUC) assay in Arabidopsis protoplasts using the whole length sequence of ProAtApx2 as the promoter of LUC reporter gene. The different negative controls (lane 1, 2 and 3) had no fluorescence (Fig. 8d). When 1381ZLUC-ProAtApx2 was transformed into protoplast with luciferin added as a substrate, obvious fluorescence was captured (Lane 4) because ProAtApx2 itself can initiate transcription of the LUC reporter gene. Furthermore, there exhibited brighter fluorescence when 1381ZLUC-ProAtApx2 and PMDC32-CtHsfA2b were co-transformed into protoplasts (lane 5) than 1381ZLUC-ProAtApx2 transformed alone (lane 4), which demonstrated that CtHsfA2b binding enhanced transcriptional activity of ProAtApx2. Fluorescence intensity measured by an ELIASA also showed a similar significant difference between lane 4 and 5.

Discussion
In our study, CtHsfA2b was rapidly up-regulated within 1 h of 42 °C heat treatment and transcripts of CtHsfA2b attenuated quickly when plants were exposed to the normal temperature and remained at low levels during the recovery phase. These results confirmed that CtHsfA2b was heat inducible and could possibly trigger heat responses of downstream target genes quickly in the early stage of heat stress. Rice A2-type HSF genes (OsHsfA2c, d and e) also had quick heat response patterns similar to CtHsfA2b, while OsHsfA2b exhibited a different gradual increase of expression level under heat stress of 42 °C 16 . This difference implies that, due to the redundancy of the same type of HSFs, gene response patterns to HS may vary with the specific family member and plant species as well as the level of temperature stress 30 . Feedback regulations are common in the complicated heat response pathways. After the heat shock peak, transcripts of CtHsfA2b attenuated quickly and remained at low level in the recovery stage in African bermudagrass. This may be due to the inactivation state of CtHsfA1. In previous studies of Arabidopsis and tomato, HsfA1 was activated by release from Hsp90/Hsp70 chaperone complexes as a result of triggering the transcription of HsfA2 in the initial stage of heat response that forms active homotrimers (HsfA1, HsfA2 and HsfB1) 31,32 . In the attenuation/recovery phase, transcription of HsfA2s is decreased by complex formation of HsfA1 with Hsp70 and the negative regulator heat shock factor binding protein1 (HSBP1) resulting in disintegration of active homotrimers to inactive monomers 10,33 . However, the transcript levels of the potential target genes of CtHsfA2b in transgenic Arabidopsis also attenuated rapidly within 2 h during heat treatment (Fig. 5), which indicated that the CtHsfA2b protein could become inactive as well with continuous heat stress. It can be deduced that the inactivation may due to interaction with Hsp17-CII, which was shown to act as a repressor for the transcription potential of HsfA2 34 , to form large cytoplasmic aggregates (heat stress granule, HSG). Furthermore, in the recovery stage, release of HsfA2 from HSG requires Hsp17-CI and probably the participation of Hsp70 and 101 10,34 . When heat stress re-occurs, the circulation is reinitiated, including changes between active HSFs triggering transcription and feedback regulation to inactive state 4,35 . CtHsfA2b had a typical characteristic of heat shock genes that are inducible by heat stress and become inactivated under normal temperatures to play a central role in the feedback regulation of HS.
HsfA2 is an important and irreplaceable regulator of a subset of heat-stress responsive proteins that maintain cellular stability and balance during heat stress and recovery stage 36,37 . The up-regulation of HsfA2 may markedly enhance transcriptional induction of a subset of HS genes, causing long-term heat protection effects 34,38 . In our study, the difference in phenotypes between transgenic lines and the WT plants was demonstrated after the relief of heat stress, even though the transcription of CtHsfA2b decreased to a low level in all plants in the recovery stage. In previous studies, overexpression of OsHsfA2e and AtHsfA2 in Arabidopsis up-regulated 37 and 46 genes respectively 16,36 , and among these genes Apx2 and HSPs also had elevated transcription levels. The induction of CtHsfA2b by heat stress could lead to subsequent induction of downstream genes with significant biological functions for heat adaptation, such as antioxidant enzyme AtApx2 for ROS scavenging and HSPs for protein protecting as discussed in detail below.
Heat stress induces oxidative damages due to excessive production of ROS 35 . Ascorbate peroxidase is one of the most important antioxidant enzymes in plants that detoxifies hydrogen peroxide using ascorbate as a reductant 39 , and Apx2 has been considered the strongest HsfA2-dependent HS-responsive gene in Arabidopsis 38 . Apx2 is one of the two cytosol isoforms of ascorbate peroxidase and another one is Apx1, which is considered a central factor of the reactive oxygen network, but Apx1 is heat labile and its function might be compensated by Apx2 under HS 38 . Transgenic Arabidopsis expressing CtHsfA2b exhibited higher expression levels of AtApx2 and also had lower levels of hydrogen peroxide accumulation than the WT plants indicating improved antioxidant capacity during heat stress. In a previous study, Apx2 mRNA levels also showed the strongest dependence on enhancement of all APX genes in HSF3-transgenic Arabidopsis plants 39 . We confirmed that CtHsfA2b was bound to the HSE on the ProAtApx2 and could regulate its transcription level during heat response. Additionally, transcription activity may be enhanced by other HSEs in the AtApx2 promoter or other transcription factors that interact with CtHsfA2b. CtHsfA2b may be an important HSF during elevated temperatures that induces antioxidant mechanisms to protect the plants.
Our results suggested that CtHsfA2b could be involved in enhancing heat tolerance of plants through regulating multiple HSPs [AtHsp18.1-CI, AtHsp22.0-ER, AtHsp25.3-P and AtHsp26.5-P(r), AtHsp70b and AtHsp101-3], which have previously been shown to play important roles in heat tolerance. These HSPs were previously demonstrated as being targeted to the cytosol, chloroplasts and endoplasmic reticulum 2,38 , indicating the protection of most cellular compartments. Most of the up-regulated members by CtHsfA2b are small HSPs (sHSPs), whose monomers are only around 16-30 kDa but most need to form large oligomers (of 8 or more monomers) in the native state to function 40 , which may be the reason they were induced to very high transcription levels by heat stress. As an sHSP in chloroplast, Hsp25.3P (former name HSP21) was assumed to play important roles in protecting chloroplast stroma and facilitating protein refolding of photosystem II in tomato and Arabidopsis 41,42 . Hsp70 is a ubiquitous and important chaperone for normal plant growth and heat stress tolerance 40 , and Hsp101 was identified as a key component involved in resolubilizing protein aggregates for both basal and acquired thermotolerance in plants 43,44 . Additionally Hsp18.1 from pea (Pisum sativum) can stably bind heat-denatured proteins and maintain them in a folding-competent state for the further refolding by Hsp70/Hsp100 complexes 45 . Therefore, besides their respective important roles, the different classes of HSPs regulated by CtHsfA2b may cooperate in the cellular protection from heat stress.
In summary, the CtHsfA2b gene from heat-tolerant warm-season African bermudagrass was heat-inducible, and its ectopic expression in Arabidopsis up-regulated downstream genes involved in protein chaperone functions and antioxidant protection resulting in improved heat tolerance. The specificity and importance of CtHsfA2b lie in its rapid induction by heat stress, its strong accumulation and systemic transcriptional regulation on target genes. These results suggest that CtHsfA2b may be an important regulator of heat stress responses and the activation of heat tolerance mechanisms. CtHsfA2b could be potentially used as a candidate gene for genetic transformation or molecular breeding to improve heat tolerance in cool-season grasses and develop improved cultivars.