Introduction

Abiotic stress usually leads to the dysfunction of proteins. Thus, maintaining their appropriate conformation and preventing the aggregation of non-native proteins are particularly important for cell survival in plants growing under stress conditions. Small heat shock proteins (sHSPs) act as molecular chaperones, facilitating the synthesis and folding of proteins in many normal cellular processes. Thus, they may play a crucial role in protecting plants against stress by re-establishing normal protein conformation (Wang et al. 2003, 2004). Based on their approximate molecular masses, HSPs are divided into five families: HSP100s (105 ± 5 kDa), HSP90s (approximately 90 kDa), HSP70s (73 ± 5 kDa), HSP60s (60 ± 5 kDa), and sHSPs (≤40 kDa) (Trent 1996). The sHSPs are perhaps the most widespread members, with monomers ranging in size from 12 to 40 kDa. Bioinformatics analysis has revealed that an average sHSP has 161 amino acids, and is characterized by a C-terminal “a-crystallin” domain (ACD) that consists of 90 amino acids (Basha et al. 2012). Plant sHSPs are divided into six classes. Classes I, II, and III are localized in the cytosol or nucleus while CIV through CVI occur in the plastids, endoplasmic reticulum, and mitochondria (Mahmood et al. 2010).

The sHSPs are synthesized ubiquitously in prokaryotic and eukaryotic cells in response to heat and other abiotic stresses. In tolerant plants, sHSPs can act as molecular chaperones and bind target proteins that have become damaged (Horwitz 1992). Expression patterns for some sHSPs can vary according to tissue type or stage of development (Waters et al. 1996). This includes most sHSPs from rice (Sarkar et al. 2009). Zou et al. (2009) have reported that rice sHSPs display different levels of expression in response to NaCl or mannitol treatment. Furthermore, transgenic rice plants over-expressing OsHSP17.0 and OsHSP23.7 show greater tolerance to drought and salt stresses when compared with the WT (Zou et al. 2012).

The phytohormone abscisic acid (ABA) plays a crucial role in regulating plant responses to abiotic stress and in controlling seed germination, plant growth, and stomatal behavior (Verslues and Zhu 2007). Expression of some sHSP genes from rice differs in response to ABA treatment. For example, transcripts of OsHSP18.03 are more abundant in the presence of ABA while levels of those for OsHSP24.1 are decreased (Zou et al. 2009; Ye et al. 2012).

To investigate the expression of several abiotic stress-related genes, we used leaves and panicles sampled from ‘Pei’ai 64S’ rice at various stages of development and under different stress conditions. In particular, OsMSR4 (Oryza sativa Multi-Stress-Responsive 4) encodes a Class III sHSP and is highly induced under cold, drought, and heat treatments. We also examined the role that OsMSR4 might have in regulating the response of transgenic Arabidopsis to exogenous ABA and drought stress.

Materials and methods

Plant materials, growing conditions, and stress treatments

Seeds of rice (Oryza sativa L. ssp. indica) cultivar Pei’ai 64S were treated with 0.1 % HgCl2 (w/v) for 10 min and washed three times with sterile water. The sterilized seeds were kept in distilled water for 72 h at 25 °C in the dark, then germinated at 37 °C under darkness for 48–72 h. The germinants were transferred to plastic pots (20 cm tall, 10 cm in diameter) containing soil and placed in a growth chamber at 28 °C/22 °C (day/night) under long-day conditions (16-h photoperiod). For microarray analysis, stress treatments were applied at the five-leaf, booting, and heading stages. Drought stress was imposed by spilling excess water out of the pots and then withholding further irrigation. The leaves of rice at the booting stage and panicles at the booting and heading stages were harvested 16 h after their leaves began to curl. For the cold-stress experiments, rice plants at the five-leaf, booting, and heading were placed in a PCG15.5 Percival growth chamber (USA) for 12 h where conditions included 4 °C and darkness. For heat stress treatments, rice plants at the booting and heading stages were placed in the same type of chamber for 2 h with conditions of 45 °C and darkness. The unstressed control plants remained in the standard growing environment described above.

In a separate trial, seeds of Arabidopsis thaliana (Columbia 0) were surface-sterilized with 10 % bleach (v/v) for 25 min and washed five times with distilled water. The sterilized seeds were sown on a ½-strength Murashige and Skoog (MS) medium supplemented with 1 % (w/v) sucrose and 0.8 % (w/v) agar (pH 5.8). When seedlings achieved the four-leaf stage, they were transplanted into soil and placed in a growth room at 26 °C/22 °C (day/night), under a 16-h photoperiod.

Three previously recognized stress-related genes RD29A, P5CS1, and NCED3 were monitored in our qRT-PCR analysis. Three-week-old Arabidopsis seedlings were immersed in ½-strength MS solutions containing standard MS components (unstressed control), 10 % polyethylene glycol (PEG) for drought treatment, or 100 µmol L−1 ABA. Samples were collected after 2 h of exposure.

Microarray analysis

Total RNAs were isolated from rice tissues using TRIzol reagent (Invitrogen). RNA samples were processed according to instructions from the technical manual for Affymetrix GeneChip expression analysis (Affymetrix, Inc.). CRNA was synthesized and labeled using an Affymetrix GeneChip IVT Labeling Kit, and was hybridized to the probe sets (Affymetrix Rice Genome Array; http://www:Affymerix.com//rice.affx). After the GeneChip array was washed and stained, it was scanned using an Affymetrix GeneChip Scanner 30007G. The microarray data were first analyzed via GeneChip Operating Software (GCOS1.2). Further analyses were performed with the R/Bioconductor (http://www.bioconductor.org/). In particular, Robust Multi-Array Average and probe sequence information, known as gcRNA, was used for correcting and normalizing the microarray background (Irrizary et al. 2003; Cope et al. 2004). Each experiment was repeated three times.

Quantitative real-time PCR (qRT-PCR) analysis

The DNase-treated RNAs were employed for first-strand cDNA synthesis, using Superscript II reverse transcriptase (Invitrogen) according to the manufacturer’s instructions. Our qRT-PCR analysis of OsMSR4, NCED3, RD29A, and P5CS1 expression was performed with a Platinum® SYBR® Green qPCR SuperMix-UDG (Invitrogen). The gene for 18S rRNA was used as the endogenous control. All primer pairs are listed in Table 1 of electronic supplementary material. Each qRT-PCR was conducted according to the manufacturer’s instructions in an ABI 7900HT (Applied Biosystems, Foster City, CA, USA) under the following conditions: 95 °C for 10 min, then 40 cycles of 95 °C for 15 s and 58 °C for 40 s, with a final 20 s at 72 °C. The data were analyzed via the comparative Ct method. Amplification experiments were conducted in triplicate, and dissociation curve analysis (60–95 °C) was done to verify the fidelity of the amplification. PlantCARE software (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) was used to identify cis-elements related to stress responses in the putative promoter region of OsMSR4.

Vector construction and Arabidopsis transformation

To construct the plant transformation vector, we utilized reverse-transcription PCR (RT-PCR) to amplify a cDNA fragment containing the entire open reading frame (ORF) of OsMSR4 from ‘Pei’ai 64S’ rice. Forward primer 5′-CCATGGAAATGGCGGACCAGCTCTCC-3′ and reverse primer 5′-GGATCCTTACAGGATCACGCACTTCTGGC-3′ were used (the underlined bases indicate restriction sites Nco I and BamH I, respectively). After the PCR fragment was obtained, it was sub-cloned into the pMD18-T (Takara) vector and sequenced. Following sequence verification, the pMD18-T plasmid harboring the cloned cDNA was digested by Nco I and BamH I and ligated into vector pJIT163, which carries the double 35S promoter. The resulting plasmid containing the cloned cDNA was digested with Kpn I and Xho I, and the purified DNA fragment was ligated into the pCAMBIA1300 plasmid, downstream of the 35S promoter between the Kpn I and Sal I sites.

The construct was introduced into Agrobacterium tumefaciens strain GV3101. Agrobacterium-mediated transformation of Arabidopsis plants was performed via the floral dip method (Clough and Bent 1998), with some modifications. Transgenic Arabidopsis seedlings were verified by plating seeds on plant nutrient agar plates supplemented with 25 mg L−1 hygromycin. Transgenic lines displaying a segregation ratio of 3:1 (resistant:sensitive) were selected to produce T3 seeds. Those that exhibited 100 % resistance to hygromycin were used for further experiments.

Stress tolerance assays for Arabidopsis

For the germination assays, seeds from WT Arabidopsis and two transgenic lines were sown in triplicate on Petri dishes containing 1/2 MS media and different concentrations of either mannitol (0, 200, 250, 300, or 325 mM) or ABA (0–0.6 µM). The seeds were stratified by incubating them at 4 °C for 2 days to synchronize germination before the dishes were transferred to a growth room. To test drought tolerance, germination was recorded after 6 days based on radicle emergence. For determining ABA sensitivity, we recorded the presence of seedlings with leaves as a percentage of total seed sown at 10 days after transfer.

Root growth was monitored in response to drought treatment. Briefly, Arabidopsis seeds were germinated on 1/2 MS agar media in Petri dishes containing different concentrations (0, 225, 275, or 300 mM) of mannitol. Dishes were placed on shelves in a vertical orientation to facilitate comparisons of their root development.

To induce drought stress, 3-week-old plants were transferred to a growth chamber under normal illumination. Standard irrigation was withheld for 12 days before watering resumed for 7 days.

Proline measurements

Free proline concentrations in leaf extracts from drought-stressed transgenic and WT Arabidopsis were determined as described by Bates et al. (1973). For each sample, the measurement was repeated three times.

Statistical analyses

All data were examined by ANOVA, using the SAS statistics program. Differences in values were considered statistically significant at P = 0.05.

Results

Expression analysis of OsMSR4

To identify the genes related to abiotic stresses, we used Affymetrix rice genome arrays for ‘Pei’ai 64S’ rice. Our data from this microarray analysis showed that imposition of drought increased the expression of OsMSR4 by 17.4- and 20.0-fold in leaves and panicles, respectively, at the booting stage. Heat stress (45 °C) caused expression in the leaves to rise by 45.6- and 18.0-fold at the booting and heading stages, respectively. Results from qRT-PCR analysis revealed patterns of expression similar to those found in our microarray analysis (Supplemental Fig. S1). This suggested that OsMSR4 is a multiple stress-responsive gene in rice.

OsMSR4 encodes a Class III sHSP

To study the possible roles of OsMSR4, we used RT-PCR to clone a corresponding cDNA containing its ORF from ‘Pei’ai 64S’ rice. Sequence analysis showed that this cDNA was 829 bp long. Comparison of the cDNA and its corresponding genomic DNA sequence from GenBank showed that OsMSR4 contains one intron and is located on Chromosome 2. Several putative cis-elements related to stress responses were identified in the putative promoter region of OsMSR4, approximately 1 kb upstream of the putative TATA box. They included MBS, involved in drought-inducibility, and an RY-element that functions in seed-specific regulation (Chen et al. 2012).

The OsMSR4 cDNA sequence encodes a protein of 172 amino acid residues with a calculated molecular mass of 18.6 kDa and a pI of 8.448. No putative functional domain was found in the protein sequence except for the ACD predicted by InterProScan (http://www.ebi.ac.uk/InterProScan/). Alignment of the deduced amino acid sequence from OsMSR4 with other representative CIII sHSPs showed that identities ranged from 75 to 92 % (Fig. 1). The nuclear localization signal (NLS) sequence was highly conserved in the proteins from all of those plant species. Consistent with this alignment, cluster analysis indicated that the protein encoded by OsMSR4 is more closely related to sHSPs from CIII than from any other class (Fig. 2). Subcellular localization of the encoded protein was predicted to be within the nucleus, based on an online investigation with a web server (http://chemdata.shu.edu.cn/subcell/).

Fig. 1
figure 1

Alignment of deduced amino acid sequences for OsMSR4 with other Class III sHSPs from Solanum peruvianum (AF399821.1), Arabidopsis thaliana (AAD25777.1), and Glycine max (XP_003528707 and XP_003547857). Numbers on right side indicate positions of amino acids. Protein sequences were aligned online via Clustalw2 (http://www.ebi.ac.uk/Tools/msa/clustalw2/). Consensus keys: ‘*’, single, fully conserved residue; ‘:’, conservation of strong groups; ‘.’, conservation of weak groups; and ‘–’, no consensus. ‘NLS’, nuclear localization signal

Full size image
Fig. 2
figure 2

Phylogenetic tree of deduced amino acid sequences from OsMSR4 and other plant sHSPs. Tree was constructed using MEGA software (Version 5.1) and Neighbor-Joining method, with pairwise deletion and Poisson correction model. Accession numbers: Lp16.1-III (Solanum peruvianum, AF399821.1), Gm18.2-III (Glycine max, XP_003528707), Gm17.9-III (G. max, XP_003547857), At17.4-III (Arabidopsis thaliana, AAD25777.1), Os18.0-1 (Oryza sativa, AAC78393), Os17.9-1 (O. sativa, EU846987), Os16.9-1 (O. sativa, GU120337), Hv17.0-1 (Hordeum vulgare subsp. vulgare, CAA69172), Zm17.2-I (Zea mays, NP_001105442), At17.8-1 (A. thaliana, NP_172220), Zm18.0-II (Z. mays, CAA38012), Os18.0-II (O. sativa, ABA29610), Ta17.3-II (Triticum aestivum, CAA41218), Ps17.7-II (Pisum sativum, AAA33670), Zm-17.5-II (Z. mays, ACG43108), and At17.6-II (A. thaliana, CAA45039.1)

Full size image

Drought tolerance is enhanced in transgenic Arabidopsis

To elucidate the biological functions of OsMSR4, we generated Arabidopsis plants expressing that gene. Two independent transgenic lines, L-3 and L-4, were chosen for further study because of their higher levels of expression (Supplemental Fig. S2). Under normal growing conditions, the transgenic plants were indistinguishable in phenotype from the WT. However, in the presence of 300 mM mannitol, the WT germination rate was 58 % versus 85 % and 87 % for the two transgenic lines (Fig. 3a). When exposed to 275 mM mannitol, the primary roots from WT seedlings were only 2.46 cm long compared with 3.0 and 3.2 cm for the transgenics (Fig. 3b). Whereas the modified plants had a 38 % survival rate after water was withheld for 12 days, only 15 % of the WT plants survived the drought treatment (Fig. 3c).

Fig. 3
figure 3

Effect of expression by OsMSR4 on drought tolerance in Arabidopsis plants. a Germination rates for WT and transgenic lines on media with and without mannitol. b Lengths of primary roots from individual plants exposed to mannitol. c Survival rates after drought treatment. d Proline concentrations in transgenic lines and WT. e Phenotypes of transgenic lines and WT grown in soil for 3 weeks after germination under normal conditions. f Seedlings produced from transgenic lines and WT after watering resumed for 7 days. Asterisks indicate differences statistically significant from WT (P = 0.05) under same conditions

Full size image

To investigate the physiological basis for this improvement in drought tolerance by transgenic Arabidopsis, we measured proline levels after irrigation was withheld for 8 days. Proline concentrations were 0.112 and 0.124 mmol g−1 FW in transgenic lines versus 0.068 mmol g−1 FW in the WT control (Fig. 3d).

ABA sensitivity is decreased in transgenic Arabidopsis plants

To determine whether OsMSR4 is involved in regulating the response of transgenic Arabidopsis to ABA, we compared sensitivity between WT and transgenic plants during the germination and seedling stages. Seeds from the transgenic lines germinated 2 days earlier than those of the WT when placed on treatment media supplemented with ABA. Moreover, the average percentage of seedlings with leaves was 6.3 % versus 29.2 % and 31.3 % for the transgenics when plants were exposed to media containing 0.6 µM ABA (Fig. 4a).

Fig. 4
figure 4

Sensitivity of WT and transgenic Arabidopsis to ABA. a Amount of seedlings with leaves as percentage of total number of seeds sown on 1/2 MS media supplemented with different concentrations of ABA. For each concentration, asterisks indicate differences statistically significant from WT (P = 0.05). Comparisons between WT and transgenic Arabidopsis seedlings on control MS medium (b) and treatment medium with 0.6 µM ABA (c) at 10 days after stratification

Full size image

Expression of stress-related genes is altered in transgenic Arabidopsis

We used qRT-PCR to investigate whether changes in sensitivity to ABA and drought tolerance in transgenic plants are accompanied by altered expression of three stress-related genes. Under both normal and stress conditions, transcript levels for RD29A, P5CS1, and NCED3 were significantly higher in the transgenic Arabidopsis lines than in the WT (Fig. S3).

Discussion

Heat shock proteins are the most ubiquitous and evolutionarily conserved molecular chaperones across all species in response to excessive temperatures and other abiotic stresses (Lindquist and Craig 1988). In plants, the sHSPs are the most prevalent type of HSPs; their abundance and diversity enable plants to survive in adverse environments (Waters et al. 1996). The CIII–related genes in rice, tomato, and Arabidopsis contain short introns in the region that encodes the β4 strand of the ACD. An important characteristic of CIII sHSPs is the highly conserved motif NGKRKR between β5 and β6 of the ACD (Siddique et al. 2003). We isolated and cloned a full-length cDNA of OsMSR4 from rice. Sequence analysis indicated that the encoded protein (Accession AK119621.1) contains an NIL sequence and an ACD in the C-terminal region. It shares significant homology with CIII sHSPs from other plants. Based on alignment and phylogenetic comparisons with sHSP sequences from other species, we confirmed that OsMSR4 is a member of the plant CIII sHSP gene family, a finding consistent with those from earlier experiments (Sarkar et al. 2009).

The HSPs are produced in cells upon exposure to elevated temperatures and other stresses. Thus, they may play an important role in the plant response to a wide range of stresses (Mahmood et al. 2010). Guan et al. (2004) have reported that plant sHSPs display different expression levels under heat shock treatment and other abiotic stresses. Furthermore, expression of OsMSR4 is up-regulated in response to salt, drought, or heat (Sarkar et al. 2009). We found that this gene was strongly induced in the leaves and panicles of ‘Pei’ai 64S’ rice when plants were exposed to cold, heat, or drought stresses at different developmental stages. The greatest accumulation of transcripts was detected in leaves at the booting and heading stages in response to high temperature. This was evidence of a correlation between OsMSR4 expression and heat shock treatment, similar to activity by other sHSPs. Expression was also significantly increased in leaves and panicles at the booting and heading stages under drought and chilling conditions. This suggested that OsMSR4 functions in those tissues where it plays an important role in regulating plant responses to such stresses. However, differences in expression patterns observed here versus the findings reported by Sarkar et al. (2009) most likely arose because the plant materials and methods were not the same between these analyses.

We also noted several matches to stress-related cis-acting elements, including ABRE and MBS, in the 1-kb putative promoter region of OsMSR4. These regulatory elements are pivotal in regulating the transcription of stress-inducible genes involved in ABA signaling and abiotic stress responses (Yamaguchi-Shinozaki and Shinozaki 2005). Therefore, our results suggested that OsMSR4 participates in modulating plant responses to abiotic stress and exogenous ABA.

Extensive studies have demonstrated that constitutive overexpression of sHSPs in plants is associated with enhanced tolerance to abiotic stress (Jiang et al. 2009; Perez et al. 2009; Xue et al. 2010). For example, transgenic Arabidopsis plants expressing RCHSP17.8 are more tolerant of heat, salt, drought, and osmotic stress (Jiang et al. 2009). Although we saw no apparent phenotypic differences between our transgenic and WT Arabidopsis plants under normal growing conditions, the transformed lines showed significantly higher rates for seed germination and plant survival, longer primary roots, and greater drought tolerance when irrigation was withheld. We considered these results to be evidence that expression of OsMSR4 in that species enhances tolerance to water stress at the seedling stage. Furthermore, the transgenic plants accumulated more free proline under drought when compared with the WT.

To gain a better understanding of the molecular basis for this enhanced tolerance by plants transformed with OsMSR4, we also examined the expression patterns of the well-characterized, stress-related RD29A and P5CS1 and found that transcripts levels of both genes were significantly higher in the OsMSR4 transgenics than in the WT under drought conditions (Fig. S3). Expression of RD29A is known to be induced when plants are dehydrated or exposed to low temperature, high salt, or exogenous ABA (Yamaguchi-Shinozaki and Shinozaki 1994). The Arabidopsis gene P5CS1 appears to catalyze the rate-limiting step in proline biosynthesis and is required for proline accumulations under osmotic stress, activity that is critical to the development of stress tolerance (Kishor et al. 1995). In response to abiotic stress, osmotic adjustments are a crucial process by which plants adapt to such challenges (Chaves et al. 2003). Under drought stress, reactive oxygen species (ROS) are generated and act as messengers, triggering the expression of genes in the signal transduction pathways during the stress-response process (Swindell et al. 2007). Therefore, we propose that excessive OsMSR4 may bind and hold proteins of transcription factors that are responsible for the induction of acolytes needed for osmotic balance under drought stress. Alternatively, OsMSR4 expression might reduce the extent of damage of stress-related transcription factors that results from ROS accumulation. Afterward, RD29A, P5CS1, and possibly other stress-related genes are induced in the transgenic plants.

ABA plays diverse roles in plant development and the adaption to environmental stresses such as drought, high salinity, and low temperature (Lee and Hwang 2009). Under abiotic stress conditions, this hormone is rapidly accumulated and then functions as a secondary messenger in abiotic stress signaling. Exogenous ABA treatment can induce numerous genes that respond to dehydration stress (Li et al. 2013). For example, NCED3 is thought to be involved in ABA biosynthesis, acting as an important regulator of ABA levels in Arabidopsis during periods of water stress, such that transcription is induced (Huo et al. 2013). The involvement of P5CS1 and RD29A in drought tolerance is regulated by ABA-mediated pathways or other pathways (Yoshiba et al. 1999; Shinozaki et al. 2003). We found that Arabidopsis plants expressing OsMSR4 had significantly decreased sensitivity to exogenous ABA. Moreover, expression by RD29A, P5CS1, and NCED3 was stronger in the transgenics than in the WT plants under either drought stress or ABA treatment. Therefore, the information obtained from our study suggests that OsMSR4 functions in response to drought conditions and exposure to ABA by altering the expression of stress-related genes through either ABA-dependent or -independent signal transduction pathways. Future research will focus on transgenic rice plants with OsMSR4 overexpression and/or RNAi-mediated inhibition while we continue to examine the functioning of that gene.