Introduction

Heat shock transcription factors (Hsfs) are a group of proteins conserved in all eukaryotic organisms which mediate the activation of genes responsive to heat stress as the terminal components of the signal transduction chain. Hsfs recognize the heat shock elements (HSEs) presented in promoters of heat shock inducible genes of all eukaryotes (von Koskull-Doring et al. 2007). Compared with other eukaryotes such as yeast, Drosophila and vertebrates, which contain one to three Hsfs, plants possess more than 20 Hsf members (Baniwal et al. 2004). Hsfs have a modular structure with an N-terminal DNA binding domain (DBD) that specifically binds to HSE, a heptad hydrophobic repeat (HR A/B) required for oligomerization, a nuclear localization signal (NLS) enriched in basic amino acids and in most cases, a C-terminal transcriptional activation domain (CTAD) (Nover et al. 1996, 2001). Based on details of their oligomerization domains, plant Hsps are grouped into class A, B and C (Baniwal et al. 2004; Nover et al. 2001; von Koskull-Doring et al. 2007).

Plant Hsfs function in multiple processes of stress response. Several members of Hsfs have been reported to play central roles in the heat stress response (HSR). Among those widely studied, LpHsfA1a is a master regulator for induced thermotolerance in tomato and AtHsfA2 is the dominant Hsf in acquired thermotolerance in Arabidopsis. Both genes can enhance thermotolerance when over-expressed and abolished the thermotolerance when knocked-out or interfered (Charng et al. 2007; Mishra et al. 2002; Nishizawa et al. 2006; Ogawa et al. 2007). Besides, plant Hsfs are also involved in other environmental stress tolerance. For example, over-expression of OsHsfA2e and AtHsfA3 elevated tolerance to salt stress in transgenic Arabidopsis (Yokotani et al. 2008; Yoshida et al. 2008); over-expression of AtHsfA2 improved heat and osmotic stress tolerance of plants simultaneously (Ogawa et al. 2007). In addition, HsfA9 was reported to be essential for embryogenesis and seed maturation in sunflower and Arabidopsis, which increased our understanding of the diversity of Hsf function (Kotak et al. 2007; Prieto-Dapena et al. 2008).

Growth inhibition is generally observed when plants are stressed. This is a mechanism for plants to prolong energy supply to cope with indisposed environment (Osorio et al. 1998). Several Hsfs including AtHsfA3, AtHsfA2 and OsHsfA2e, among many other types of stress-inducible transcription factors, also confer dwarf phenotype (Ogawa et al. 2007; Yokotani et al. 2008; Yoshida et al. 2008). Final size of plant organ depends on cell number and cell size, which are controlled by two successive processes referred to as cell proliferation and cell expansion, respectively (Anastasiou and Lenhard 2007; Mizukami 2001; Tsukaya 2006). Many genes affecting organ growth by cell division and/or cell expansion are well studied (Breuer et al. 2007; Deprost et al. 2007; Fleury et al. 2007; Jing et al. 2009; West et al. 2004), but the underlying mechanism of Hsf-induced growth retardation is still far from understanding.

A group of so-called resurrection plants is capable of withstanding severe water loss in vegetative tissues (Bartels and Sunkar 2005; Gaff 1971; Oliver et al. 2000). It has been observed that the growth is arrested and metabolic reactions are largely ceased in leaves of resurrection plants soon after desiccation, and resumed rapidly upon rehydration (Bartels and Sunkar 2005; Oliver et al. 2000). Therefore, resurrection plants can be used to identify the possible link between the stress tolerance and the rapid regulation of growth, which will deepen our understanding on the molecular basis of stress tolerance. Previously, Hsf fragments had been isolated from the resurrection plant Craterostigma plantagineum and Boea hygrometrica (Bockel 2001; Wang et al. 2009). In this study, we describe the cloning of the full length cDNA of the Hsf from B. hygrometrica and functional characterization based on phenotypic analysis and putative target genes using transgenic approach. The aim of our research is to estimate the possibility that BhHsf1 may function in integration of the processes of growth retardation and stress tolerance.

Materials and methods

Plant materials and growth conditions

Boea hygrometrica plants were collected from their natural habitat in Beijing and maintained in a greenhouse with regular irrigation as described previously (Jiang et al. 2007). Tobacco (Nicotiana tabacum cv. SR-1 and Nicotiana benthamiana) and Arabidopsis thaliana (Col-0) were used for transformation. Seeds were sterilized in 1% sodium hypochlorite for 15 min, and were then germinated on MS medium. 7-day-old Arabidopsis and 14-day-old tobacco seedlings were transferred to soil, and grown in a chamber at 22 ± 1°C in a photoperiod of 16 h light and 8 h darkness.

Molecular cloning and sequence analysis

A cDNA fragment of BhHsf1 was identified by differential screening of dehydration-inducible genes from a cDNA library prepared from desiccated leaves of B. hygrometrica using macroarray hybridization method as described previously (Wang et al. 2009). The full length coding region was subsequently obtained using the system for Rapid Amplification of cDNA Ends Kit (Invitrogen, CA). Reverse transcription was performed using gene specific primer 5′-TATCCGACGGCTGCTT-3′. Primers 5′-GGGATCCTTTGGAGGGATTTGAT-3′ and 5′-CTGGATCCTCCTTTTGATGTGTTTC-3′ were used in the first and second rounds of PCR amplification with anchor primers 5′-GGCCACGCGTCGACTAGTACG14-3′ and 5′-GGCCACGCGTCGACTAGTAC-3′, respectively. The conditions were 94°C denaturation for 2 min, 35 cycles of 94°C 30 s; 55°C 30 s; 72°C 1 min, and 72°C elongation for 10 min for both the first and second PCR amplifications.

The 5′-flanking region of BhHsf1 was obtained by TAIL-PCR (Liu et al. 1995) using a random primer 5′-(AGCT)GTCGA(GC)(AT)GA(AGCT)A(AT)GAA-3′ and two pairs of nested specific primers 5′-CTGCAGCTACCACAGAAACCAT-3′, 5′-CAGCCTCATCAAGAAACACCAAG-3′ and 5′-CTGTTACACGCCCACATTTTC-3′, 5′-GTGTTATTCTTACGTATGATGTGCG-3′.

cDNA and protein sequences were analyzed using BLAST program. Amino acid comparison and multiple alignment was performed using ClustalW program. Phylogenetic tree was constructed by Neighbor Joining method using MEGA 4.0. Cis-elements were predicted by PLACE software.

Generation of transgenic plants

BhHsf1 coding sequence was fused between 35S promoter and terminator and the cassette was cloned into pBin19 binary vector (Frisch et al. 1995) to create the 35S:BhHsf1 plasmid. Based on this vector, GFP coding region was in-frame fused downstream of BhHsf1 to generate 35S:BhHsf1-GFP construct. Besides, the 786 bp promoter of BhHsf1 was amplified from genomic DNA, and ligated upstreame of GUS coding region in pBISN1 vector to generate ProBhHsf1:GUS fusion. The resulting plasmids were introduced into tobacco and Arabidopsis by Agrobacterium-mediated transformation (Clough and Bent 1998; Kapila et al. 1997). At least 30 independent lines for each construct were selected on MS agar plates containing 50 and 100 mg l−1 kanamycin (for Arabidopsis and tobacco, respectively) and homozygous T3 generations of transgenic plants were used for further analysis.

Phenotypic analyses

To determine the size of the organs, stem height, leaf length and width were measured with a ruler, fully expanded cotyledons and fifth leaves were photographed for area measurement by ImageJ software (http://rsb.info.nih.gov/ij). To examine the growth kinetics of leaves, the fifth leaves were used to measure blade length and width in 4-day intervals after emergence. In order to study cell number and size, fully expanded fifth leaves (5-week-old plants grown in soil) were cleared to observe palisade cells between the midvein and the leaf margin. The number of palisade cells per unit area of a region was determined, and the area of the leaf blade was multiplied by this value to calculate the total number of palisade cells in the subepidermal layer. Eight regions per leaf and five leaves were analyzed (modified from Horiguchi et al. 2005; Jing et al. 2009). Areas of organs and cells were measured with ImageJ software. Flowers and fully expanded siliques from the same positions on WT and transgenic plants were compared.

Thermotolerance assays

Seedlings were grown on MS agar in Petri dishes. The Petri dishes were sealed with transparent plastic film and then immerged into water bath at different temperatures for heat stress tolerance assays. For tobacco, 2-week-old plants were exposed to 48°C for 2.5 h for the basal thermotolerance assays, and to 40°C for 3 h followed by 50°C for 2 h for the acquired thermotolerance assays (modified from Prieto-Dapena et al. 2008). For Arabidopsis, 2-week-old plants were exposed to 45°C for 1 h, for the basal thermotolerance assay and to 37°C for 1 h, 22°C for 3 h, and 49°C for 1 h for the acquired thermotolerance assays (modified from Ogawa et al. 2007). These plants were grown at 22°C after HS treatment for 3 days (for basal thermotolerance assays) and 6 days (for acquired thermotolerance assays) before taking photographs and counting the number of surviving plants.

GUS histochemical analysis

Plants from six independent transgenic Arabidopsis lines containing ProBhHsf1:GUS were used for GUS staining using X-Gluc as a substrate. After staining, the tissues were soaked in several changes of 95% ethanol and examined under a microscope (Bertrand et al. 2003). For heat-shock experiments, the plants were placed at 37°C for 8 h and were allowed to recover at growth condition for 3 h before histochemical staining. Arabidopsis plants that harbor AtCYCB1;1:GUS (Colon-Carmona et al. 1999) were crossed with 35S:BhHsf1 homozygous plants. Seven-day-old seedlings of crossed progenies were used for histochemical GUS staining.

GFP visualization

Samples were incubated with 0.1 mg/ml DAPI and illuminated with an argon laser (488 nm wavelength for detection of GFP fluorescence and 405 nm for DAPI, respectively) using a confocal laser scanning microscope (Leica TCS SP5, Germany).

RNA blot and RT-PCR

Total RNA was prepared by a guanidine thiocyanate method (Hu et al. 2000). For RNA blot analysis, total RNA (5 μg per lane) was denatured for 1 h at 55°C, and, separated on a 1% (w/v) denaturing agarose gel, transferred onto HybondTM-N+ nylon membrane (Amersham Phamacia Biotech, UK), fixed by incubating for 2 h at 80°C. Probe labeled with fluorescein and fluorescent signal detection were done using Gene Images Random Prime Labelling Module and Gene Images CDP-Star Detection Module (Amersham Phamacia Biotech, UK) according to manufacture’s instruction.

Reverse-transcribed PCR (RT-PCR) was performed to monitor the expression of BhHsf1 in transgenic tobacco and Arabidopsis. ACTIN2 and 18S rRNA were used as the internal controls for Arabidopsis and tobacco, respectively. Quantitative real time PCR analyses were performed using RNA isolated with Trizol reagent (Invitrogen, US) as previously described (Wang et al. 2009). The primers used are listed in Table S1.

Microarray analysis

Total RNA was isolated with Trizol reagent (Invitrogen) from the aerial part of the 3-week-old seedlings of 35S:BhHsf1 and WT plants grown in parallel and then analyzed using an A-MEXP-1038 Agilent Arabidopsis 3 Oligo Microarray 4x44K 015059 G2519F (Agilent Technologies Inc. USA). For each biological replicate, materials from 15 plants were pooled for RNA purification. Two independent experiments were carried out using two independent transgenic lines (OE-15 and OE-21). Cy5- and Cy3-labeled cDNA probes were prepared from total RNA using the Low RNA Input Fluorescent Linear Amplification Kit (Agilent Technologies Inc. USA). Microarray experiments including labeling, hybridizations and data analysis were performed according to the manufacturer’s manual. Agilent G2566AA Feature Extraction Software was used to locate and delineate every spot in the array and to integrate each spot’s intensity, filtering, and normalization by the Lowess method. All changes in gene expression were statistically significant at P < 0.01. Microarray data were submitted and can be accessed from the MIAME checklist format (http://www.mged.org/Workgroups/MIAME/miame.html) under the ArrayExpress accession of E-MTAB-106. Pathways analysis was performed with GeneSpring10 software (http://www.silicongenetics.com) using KEGG and Reactome database.

Flow cytometric analysis

The fully expanded fifth leaves (corresponding to 5-week-old plants) of WT and transgenic Arabidopsis OE plants were chopped with a razor in parallel, suspended in cold nuclear isolation buffer (Galbraith et al. 1983) and filtered through a 20 μm filter. Then 4′,6-diamidino-2-phenylindole (DAPI; 0.5 mg ml−1) staining buffer was added to the filtered solution. Nuclear ploidy was analyzed with a Cell Lab Quanta™ SC flow cytometer (Beckman coulter, US).

Results

Cloning and sequence analysis of BhHsf1

A cDNA fragment encoding C-terminal of an Hsf was identified previously as a dehydration-inducible gene from a cDNA library from dehydrated leaves of B. hygrometrica using a macroarray screen approach (Wang et al. 2009). Subsequently, the corresponding full-length cDNA sequence containing 14 bp of 5′ UTR, 287 bp of 3′ UTR and 1,152 bp of open reading frame (ORF) was cloned using 5′RACE and designated as BhHsf1 (Accession No. FJ947047). The deduced amino acid sequence includes an Hsf specific DNA-binding domain (DBD), an oligomerization domain with the hydrophobic heptad repeat (HR-A/B), and an adjacent putative nuclear localization signal (NLS) enriched in arginine and lysine (Fig. 1a). The plant Hsfs unique stretch (DSDR) between the third and fourth β-strands within the DBD (Nover et al. 1996) also exists in BhHsf1 (Fig. 1b). Besides, several features for class A Hsfs were found in BhHsf1, including the 21-amino acid insertion linking the cores of HR-A and HR-B, and the putative C-terminal transcriptional activation domain (CTAD motif), which is characterized by aromatic, large hydrophobic, and acidic amino acid residues. However, this protein is lack of any nuclear export signal (NES), which is required for keeping the balance of nuclear export and import for some Hsfs. BLASTp search in GenBank database revealed that BhHsf1 was most closely related to HaHsfA9, a seed-specific Hsf from sunflower, with 45% identity over the whole sequence. Phylogenetic relationship of 20 Hsfs from A. thaliana, tomato, sunflower and B. hygrometrica based on the sequence of the conserved DBD domain and HR-A/B region has indicated that BhHsf1 belongs to the subgroup containing HaHsfA9 and AtHsfA9 (Fig. 1c).

Fig. 1
figure 1

BhHsf1 sequence characterization and phylogenetic analysis. a Full-length cDNA sequence and characteristics of the predicted BhHsf1 protein. BhHsf1 cDNA contains a short 5′-untranslated region (UTR), a coding region and a poly (A) tail. The arrow indicates the position of the conserved intron that is present in corresponding genomic sequence. The deduced translation of the BhHsf1 protein is showed below the coding region sequence (CDS). We also indicate putative domains identified by sequence comparison with other Hsfs, the DNA binding domain (DBD); the oligomerization domain, with overlapping heptad repeats (HR-A/B) indicated by filled triangle and hollow triangle; a putative nuclear localization sequence (NLS) (underlined); the 21-amino acid insertion between HR-A and HR-B (boldface); and finally the putative AHA motifs (underlined). b Alignment of the conserved DNA binding domain (DBD) sequences of BhHsf1, HaHsfA9 and AtHsfA9. Amino acids highlighted in black are conserved in all three proteins; those highlighted in grey are conserved in two of them. The four β-sheets and three α-helixs are indicated. Dots above indicated the residues with conserved and important functions. c Phylogenetic relationship of Hsfs based on amino acid sequence comparison of the DBDs and HR-A/B regions. Amino acid comparison and multiple alignments were performed using the ClustalW program. Phylogenetic tree was generated by Neighbor Joining method using MEGA 4.0. Plant species are abbreviated as follows: At, Arabidopsis thaliana; Lp, Lycopersicon esculentum (cotton); Ha, Helianthus annuus (sunflower); Bh, Boea hygrometrica. Sequence data were taken from the GenBank database for the following accession numbers: BhHsf1, FJ947047; HaHsfA9, AAM43804; AtHsfA1a, CAB10555; AtHsfA1b, CAA74397; AtHsfA2, AAC31222; LpHsfA2, CAA47807; AtHsfA3, CAB82937; LpHsfA3, AAF74563; AtHsfA4a, CAA16745; AtHsfA4c, BAB09213; AtHsfA5, CAB10177; AtHsfA6a, BAB11313; AtHsfA6b, BAB01258; AtHsfA7a, CAB41311; AtHsfA7b, CAB86436; AtHsfA8, AFF16564; AtHsfA9, BAA97129; AtHsfB1, CAB16764; AtHsfB2a, CAB63802; AtHsfC1, BAB02003

Full size image

BhHsf1 expression is regulated by stress and development

A RNA blot was performed to investigate the expression of BhHsf1 gene in response to dehydration. The transcripts of BhHsf1 was absent in untreated fresh leaves, but appeared in leaves dehydrated for 2 h, and accumulated to high levels in leaves dehydrated for 8–72 h (Fig. 2), and disappeared after rehydration (data not shown).

Fig. 2
figure 2

RNA blot analysis of BhHsf1 in dehydrated B. hygrometrica leaves. The numbers above each lane indicate the hours after leaves were dehydrated for 0, 2, 8, 24 and 72 h. Each lane was loaded with 5 μg mRNA

Full size image

To further characterize the regulation of BhHsf1 gene expression, a 786 bp 5′ flanking sequence of BhHsf1 was amplified from B. hygrometrica genomic DNA (Fig. S1), and cloned in-frame upstream to a β-glucuronidase (GUS) reporter gene. The ProBhHsf1:GUS cassette was stably transformed into Arabidopsis. Histochemical staining of transgenic plants revealed GUS activity at almost all developmental stages (Fig. 3). In young seedlings, GUS staining was detected throughout the plant with strong signals in the actively proliferating tissues of the shoot and root, such as leaf primordial, root tip and the lateral root differentiation zones, as well as cotyledons and vascular tissues (Fig. 3a–d). In adult plants, GUS signals were detected very faint (Fig. 3e). Upon heat, GUS signals were increased in leaves (Fig. 3f). Wound can also induce stronger GUS signal in leaves (Fig. 3g). In old plants, GUS signals were found in anther and pistil but not in mature siliques (Fig. 3h, i). However, the intensity of GUS signal is not significantly altered after drought stress in Arabidopsis (data not shown). One possible reason for this is that the promoter GUS construct may miss regulatory sequences such as the coding sequences (De Almeida et al. 1989) and the promoter-proximal introns (Rose et al. 2008), which play important roles in the regulation of gene expression in some cases.

Fig. 3
figure 3

Histochemical localization of GUS expression driven by BhHsf1 promoter in transgenic Arabidopsis. a 9-day-old seedling. Bar 5 mm. b Shoot apex showing the emerging first pair of leaf primordia. Bar 20 μm. c Meristematic region in root of 7-day-old seedling. Bar 20 μm. d Lateral root formation of early to late stages (6–12 days after germination, from left to right). Bar 20 μm. Three-week-old seedling untreated (e) and heat stressed (f). Bar 5 mm. g Rosette leaves treated with wounding. Bar 10 mm. h Flower. Bar 5 mm. i Siliques at different developmental stages. Bar 5 mm. GUS staining was observed for at least four different transgenic lines for each experiment

Full size image

BhHsf1 is localized in the nucleus

The subcellular localization of BhHsf1 was examined by expressing the BhHsf1-GFP fusion protein in tobacco epidermal cells. The BhHsf1-GFP signals were detected mainly in the nuclei, while the control GFP was detected in both nuclei and cytosol (Fig. 4). These results indicate that BhHsf1 protein is targeted to the nucleus, which is consistent with the absence of NES in BhHsf1.

Fig. 4
figure 4

Nuclear localization of BhHsf1-GFP fusion protein. Tobacco epidermal cells transiently expressing GFP alone (upper) and BhHsf1-GFP (bottom) proteins were stained with DAPI and observed with a confocal microscope. Note that BhHsf1 is localized within nucleus, whereas that of control GFP is evenly dispersed in the whole epidermal cells. DAPI is stained pseudocolored blue. Bars 10 μm

Full size image

Over-expression of BhHsf1 in Arabidopsis and tobacco results in reduced size of aerial organs

Several Hsfs have been transformed to Arabidopsis for functional assessment. In order to gain an insight on the basic function of BhHsf1 in comparison with other plant Hsfs, transgenic Arabidopsis ectopically expressing BhHsf1 under the 35S promoter were constructed. Among a collection of 36 T1 transgenic lines, nearly all showed dwarf phenotype to various extents: 8 lines were severe, 19 lines moderate and 7 lines weak, as represented by OE-7, OE-15 and OE-21, respectively (Fig. 5b, c), which is in consistence with the expression levels of BhHsf1 transcripts (Fig. 5a). Plants of these lines were selfed, and kanamycin-resistant T3 seeds were collected for further analyzed.

Fig. 5
figure 5

Phenotypes of 35S:BhHsf1 transgenic Arabidopsis lines. a Expression of BhHsf1 in WT and three OE lines by RT-PCR, ACTIN2 was used as an internal control. b Seedlings of 9-day-old. Bar 5 mm. c Seedlings of 4-week-old. Bar 5 mm. d Roots of 9-day-old WT and OE lines. Bar 10 mm. e Rosette leaves from WT and OE lines arranged in order (left to right) from the oldest to the youngest. Bar 10 mm. WT, OE-7, OE-15 and OE-21 are shown from top to bottom. Dimensions (f, h) and areas (g, i) of cotyledons (f, g) and the fifth leaves (h, i) in WT and BhHsf1 OE lines. Cotyledons and leaves from eight plants of each genotype were measured: average values ± SD are shown; Students’ t-test, *P < 0.05, **P < 0.01. Flowers (j) and siliques (l) of 7-week-old WT and 35S:BhHsf1 plants. Bars 5 mm. k A normal flower of WT (left) and a sterile flower of OE-7 line (right) with reduced stamens that is defect in self-pollination. Bar 2 mm. m Normal ovules in silique of WT (left) and aborted ovules in silique of OE-7 (right), respectively. Bar 2 mm. n Seven-week-old plants of WT and OE lines. Bar 2 cm. WT, OE-7, OE-15 and OE-21 are shown from left to right unless indicated

Full size image

When grown in parallel, the OE lines exhibited obvious smaller aerial organs compared with wild type (WT) (Fig. 5). For example, cotyledons appeared at approximately the same time after germination in both WT and OE lines, but the final sizes were smaller and the shape was rounder in OE lines (Fig. 5b). The areas of fully expanded cotyledons in 9-day-old seedlings of OE-7, OE-15 and OE-21 were about 40, 55 and 70% of that in the WT, respectively (Fig. 5f). Further measurement revealed that cotyledon length in OE lines was significantly smaller than that in WT, while the cotyledon width was not (Fig. 5g). Consistently, the area of the fully expanded fifth leaves of 5-week-old plants of the OE lines were less than 45% of that in WT of the same age (Fig. 5h); and the reduction in leaf length was more significant than that in the width (Fig. 5i). The sizes of other aerial organs, including stems (Fig. 5n), flowers (Fig. 5j) and siliques (Fig. 5l), were also decreased to some extent, resulting in dwarfish phenotype. No difference between OE plants and WT were seen on the processes of development. In addition, it was observed that the stamens failed to elongate over the pistils in a portion of flowers of OE lines, which resulted in sterile to 46% in OE-7, 25% in OE-15, and 13% in OE-21, respectively (Fig. 5k, m). Overall, compared with WT, BhHsf1 OE plants exhibited pleiotropic phenotypes. However, the lengths of the primary roots of OE lines were not significantly different from that of the WT (Fig. 5d), which is consistent with our quantification of the primary root length (data not shown). Thus, no defects were found in the length of the primary root in the OE plants.

To confirm the effects of BhHsf1 on plant growth, transgenic tobacco plants over-expressing BhHsf1 were also generated. Three homologous lines, 2-10, 4-9 and 8-2 were obtained (Fig. 6a). In agreement with the Arabidopsis OE lines, these transgenic tobacco lines also displayed small-sized leaves (Fig. 6b and c), flowers (Fig. 6d) and stems (Fig. 6e). The length of stamens were also reduced (Fig. 6d), which resulted in the failure of self-pollination.

Fig. 6
figure 6

Phenotype of 35S:BhHsf1 transgenic tobacco plants. a Relative expression of BhHsf1 in WT and three 35S:BhHsf1 lines by RT-PCR. 18S rRNA was used as an internal control. b WT and 35S:BhHsf1 plants of 18-day-old. Bar 5 mm. c The tenth leaves of 3-month-old plants. Bar 2 cm. d Flowers of 3-month-old WT and 35S:BhHsf1 plants. Reduced stamens of various extents are observed in all OE lines. Bar 1 cm. e WT and 35S:BhHsf1 plants of 3-month-old. Bar 10 cm. WT, 8-2, 2-10 and 4-9 are shown from left to right

Full size image

Reduced organ size in BhHsf1 OE transgenic plants is attributed mainly to decreased cell proliferation

The most noticeable phenotype of BhHsf1 OE transgenic plants is the size reduction in aerial organs. Since leaf has been used as a model to investigate the genetic and environmental control of organ size and shape in the plants (Tsukaya 2006), we performed a growth kinetic analysis of the fifth leaf to determine the nature of the defect in leaf growth of the BhHsf1 OE lines (Fig. 7a, b). The data revealed that the size difference in both leaf blade length and width between OE lines and WT occurred at comparatively early growing stages, but the time that leaves to stop growing in the longitudinal and transversal direction are similar between OE lines and WT (Fig. 7a, b), suggesting that the OE plants may have altered the cell division and thus the leaf expansion rates during the proliferation phase of leaf development but did not affect the duration of proliferation or leaf expansion. To assess the contribution of cell division to the small-sized leaves of OE lines, the palisade cell area and cell number of the fully expanded fifth leaves in WT and OE lines was compared. It was observed that the average area of palisade cells in OE lines was approximately the same as that in WT (Fig. 7c–g). However, the total palisade cell numbers in OE lines were only one-third of that in WT (Fig. 7h). These findings suggest that the over-expression of BhHsf1 mainly impedes the process of cell proliferation, rather than cell expansion, during leaf growth. And the effect is correlated with the dosage of BhHsf1 transcripts in each OE line (Fig. 7). A B-type cyclin CYCB1;1 promoter fused GUS reporter gene, which has been used as a tool to monitor mitotic activity in actively proliferating tissues (Himanen et al. 2002), was introduced into OE-15 and WT by crossing to evaluate the meristem size and activity. The range and intense of the GUS staining in young developing leaves had been obviously reduced in OE plants (Fig. 8), suggesting the cell mitotic division decreased in the site of vigorous division and growth in OE plants.

Fig. 7
figure 7

Growth rate and anatomical analysis of the fifth leaf of WT and BhHsf1 OE Arabidopsis. a, b Kinetics of leaf growth. The length and width of the fifth leaves of WT and OE lines were examined. For each genotype, more than 12 leaves were continuously observed in 4-day intervals after emergence. Average values ± SE are shown. c–f Palisade cells of the fully expanded fifth leaves in WT and OE lines. The fifth leaves from 5-week-old plants were cleared and visualized under a microscope. Bar 50 μm. g Area of palisade cells of leaves in WT and OE lines. Five blades of the fifth leaf from each genotype were cleared, and at least 50 palisade cells were measured in each leaf under a microscope. Average values ± SD are shown. h Mesophyll cell number in WT and OE lines. The area of the leaf blade was multiplied by the number of palisade cells per unit area to calculate the total number of palisade cells in the subepidermal layer. Average values ± SD are shown

Full size image
Fig. 8
figure 8

CYCB1;1:GUS expression in young developing leaves of WT (a) and 35S:BhHsf1 Arabidopsis lines (b), respectively. Bars 20 μm

Full size image

35S:BhHsf1 Arabidopsis and tobacco exhibit increased thermotolerance

The previously described experiments demonstrated that Hsfs from various plants were required for basal thermotolerance and/or acquired thermotolerance (Ogawa et al. 2007; Prieto-Dapena et al. 2008; Yokotani et al. 2008). To determine the effect of BhHsf1 on plant heat stress responses, we compared the acquired and basal thermotolerance of transgenic Arabidopsis with those of WT: 100% seedlings of OE lines but none of WT display the acquired thermotolerance (Fig. 9a); meanwhile 50% seedlings of OE line but none of WT survived when exposed to high temperature without a conditioning pretreatment (Fig. 9b, c). Similarly, we found that over-expression of BhHsfA1 conferred transgenic tobacco increased both acquired (Fig. 9d) and basal (Fig. 9e) thermotolerance. The observation implicated that BhHsf1, like other Hsfs, is involved in tolerance to heat stress.

Fig. 9
figure 9

Thermotolerance of WT and 35S:BhHsf1 transgenic Arabidopsis and tobacco plants. Comparison of acquired (a) and basal (b) thermotolerance among WT and transgenic Arabidopsis OE lines. c Percentages of the surviving transgenic Arabidopsis plants after HS stress at 45°C for 1 h. For each experiment, more than 60 plants were used; values are means ± SD from three independent measurements. Acquired (d) and basal (e) thermotolerance of WT and 35S:BhHsf1 transgenic tobacco lines

Full size image

Transcriptome analysis of aerial organs in BhHsf1 OE Arabidopsis

To investigate the possible molecular mechanisms of BhHsf1 function in plant growth and stress responses, a genome-wide expression analysis on the aerial organs of young plants (3-week-old) was carried out to identify the genes with altered expression levels in the OE lines. Approximately 27,718 expressed genes were detected. The statistical analysis identified a total of 687 differentially expressed genes between 35S:BhHsf1 and WT at P < 0.01, among which 240 (34.9%) were down-regulated and 447 (65.1%) were up-regulated with fold changes higher than 1.5.

Table 1 listed 59 genes that were up-regulated in OE lines for at least sevenfold of that in WT. These genes are involved in diverse aspects of stress response, metabolism, protein fate, transcription regulation and signal transduction, implying that BhHsf1 is involved in multiple processes of expression regulation. Many of them that have been reported to be also induced by other Hsfs include AtAPX2, AtGolS1, MKP6.25, and Hsps (Nishizawa et al. 2006; Ogawa et al. 2007). There were 11 Hsp genes that were up-regulated in OE lines to up to 189-folds, including genes encoding two Hsp70s and seven sHsp with molecular weight around 16–30 kDa. sHsps and Hsp70 can function as molecular champions preventing aggregation of denatured proteins and refold stress-denatured proteins (Gaitanaris et al. 1990; Goloubinoff et al. 1999; Sheffield et al. 1990). AtAPX2 encodes cytosolic ascorbate peroxidase that scavenges reactive oxygen species (Mittler et al. 2004). AtGolS1 encodes galactinol synthase, the rate-limiting enzyme of raffinose oligosaccharide synthesis, plays an important role in compatible solute accumulation in response to stress (Taji et al. 2002). AtMKP6.25 encodes a member of late embryogenesis abundant (LEA) proteins, which accumulate in response to drought, freezing and salt stress as protective molecules (Shao et al. 2005; Tunnacliffe and Wise 2007). AtAPX2 and AtGolS1 were both induced by stresses such as HS, dehydration, and salt (Chang et al. 2004; Panikulangara et al. 2004). Beside, some other stress-inducible genes were also found to be up-regulated by BhHsf1. For example, genes encoding C2 domain-containing protein and glycine-rich protein have been reported to be induced by dehydration (Collett et al. 2004; Neale et al. 2000; Wang et al. 2009). These results suggested that BhHsf1 may regulate a large group of stress related genes, which is in agreement with our observation that the BhHsf1 OE lines conferred tolerance to heat stress. Five genes with different expression level were further analyzed by real time PCR and the results confirmed the microarray data (Fig. 10).

Table 1 Up-regulated genes in BhHsf1 OE plants compared with WT
Full size table
Fig. 10
figure 10

Quantitative real time PCR analysis of AtGOLS1, AtMKP6.25, AtAGL68, AtHsp22.0-ER and AtAPX2 in WT, OE-15 and OE-21 seedlings

Full size image

We also focused on the genes that were down-regulated in the BhHsf1 over-expression lines (Table S2). Pathway analysis revealed that mitotic cell cycle and DNA replication are the pathways containing the largest number of down-regulated genes. For example, seven and five down-regulated genes were identified from OE-15 falling into pathways of mitotic cell cycle and DNA replication, respectively (Table 2). Although the induction levels of these genes were low, down-regulation of both pathways was statistically significant. Among the genes listed in Table 2, six were present in G1/S Transition and S phase and one was in G2/M Transition. The six genes in G1/S Transition and S phase also involved in DNA replication. Similar sets of genes were also down-regulated in OE-21 with fewer folds, which was consistent with the fact that dwarf phenotype and the expression level of BhHsf1 in OE-15 is more significant than that in OE-21. The repression of CDC45 and MCM10 by BhHsf1 were further confirmed by real time PCR, which were down-regulated to 1.35 ± 0.11 and 1.44 ± 0.06 fold in OE15 when compared to that in WT, respectively. These results indicate that mitotic cell cycle might be blocked in BhHsf1 OE plants.

Table 2 Down-regulated cell cycle related genes in BhHsf1 OE plants compared with WT
Full size table

BhHsf1 over-expression affects endoreduplication

During Arabidopsis leaf development, a certain degree of endoreduplication is generously observed. Molecular evidences have indicated that cell cycle was disturbed in BhHsf1 OE lines due to deficiency in DNA replication, so we examined whether nuclear endoreduplication was affected in cells of 35S:BhHsf1 transgenic Arabidopsis. Flow cytometric analyses with nuclei of fully expanded fifth leaves revealed differences in the polyploidy among cells of WT and OE lines. It was observed that the ploidy of WT and OE plant cells ranged from 2C to 16C. OE and WT possessed similar frequency of cells at 2C and 4C, however, OE lines showed declined frequency of cell numbers at the higher ploidy levels 8C and 16C (Fig. 11). These observations suggested that over-expression of BhHsf1 in Arabidopsis impaired nuclear endoreduplication.

Fig. 11
figure 11

Percentage of cells with different nuclear ploidy. Nuclear DNA Content was analyzed using flow cytometry in fully expanded fifth leaves of 32-day-old WT and BhHsf1 OE seedlings. Average values ± SDs are shown (n = 3)

Full size image

Discussion

BhHsf1 improves stress resistance

Hsfs play a critical role in response to heat stress in all organisms, especially in sessile plants (von Koskull-Doring et al. 2007). BhHsf1 comprise motifs such as DBD, HR-A/B, NLS and CTAD, which are the modular structures in Hsfs. The existence of 21 amino acids between HR-A/B and the putative AHA motifs enable us to group BhHsf1 into class A. BhHsf1 possesses highest sequence similarity with HaHsfA9 and AtHsfA9, which expressed during seed development. Recent studies have suggested that DT in seeds may be adopted by resurrection plants to develop vegetative DT (Moore et al. 2009; Oliver et al. 2000). The relevant anagenesis of BhHsf1 with seed-specific HsfA9 and AtHsfA9 fits this hypothesis.

Previous studies have demonstrated a link between Hsfs and different environment stresses (Li et al. 2005; Nishizawa et al. 2006; Ogawa et al. 2007). In this study, evidence from Northern blot and GUS staining supports the link between BhHsf1 and abiotic stress. Although no significant drought-triggered changes of BhHsf1 promoter activity was detected in Arabidopsis, which might be due to the potential flaw of the promoter-GUS expression system. The fact that BhHsf1 transcript was induced by drought treatment in B. hygrometrica and the proBhHsf1:GUS activity was up-regulated by heat shock and wound suggests a role for this putative transcription factor in regulation of plant responses to these abiotic stresses. Besides, like its homologue HaHsfA9, BhHsf1 is able to confer heat tolerance when over-expressed. Gene expression comparison between Hsf over-expression lines and WT had revealed many common genes that were up-regulated by BhHsf1 and other plant Hsfs that had been investigated. For example, AtAPX2, AtGolS1, MKP6.25 and almost the same set of Hsps including Hsp22.0-ER, Hsp18.1-CI, Hsp17.6B-CI, Hsp17.6C-CI, Hsp25.3-P, Hsp26.5-P, Hsp70 and Hsp70T-2 were up-regulated by AtHsfA2, AtHsfA3, OsHsfA2e and BhHsf1 (Nishizawa et al. 2006; Ogawa et al. 2007; Yokotani et al. 2008; Yoshida et al. 2008). HSEs are located in 5′ upstream region of all these genes, suggesting that they could be direct target genes of these Hsfs. This indicates the evolutionally conserved mechanism of plant stress adaptation via Hsf-mediating activation of genes that function in ROS scavenging, osmo-adjust and macromolecular protection.

BhHsf1 leads to retarded organ growth

The growth of plant aerial organs is always arrested during stress conditions (Bartels and Sunkar 2005). It is considered that plants may preserve carbohydrates for sustained metabolism, prolonged energy supply, and for better recovery after stress relief by inhibition of shoot growth (Osorio et al. 1998).

Over-expression of many stress-induced transcription factors resulted in enhanced stress tolerance and also growth retardation to some extent. For example, over-expression of AtDREB2A (Liu et al. 1998), AtAREB1 (Furihata et al. 2006), OsHsfA2e (Yokotani et al. 2008) exhibited slightly retarded growth; over-expression of AtDREB1A (Kasuga et al. 1999), AtHsfA3 (Yoshida et al. 2008) and AtHsfA2 (Ogawa et al. 2007) led to severe dwarf phenotype. The function of these factors in stress resistance has been investigated intensively; however, the molecular and cellular mechanism underlying the growth inhibition by these factors was less studied.

When ectopically expressed BhHsf1 in Arabidopsis, cell numbers decreased significantly in the leaves. In addition, other organs of Arabidopsis and tobacco, such as stems, siliques and flowers, also showed smaller sizes, suggesting that BhHsf1 over-expression may generally decrease cell proliferation in aerial organs. GUS expression driven by the BhHsf1 promoter was detected in the actively proliferating tissues such as root tips, root hair differentiation zones, and leaf primordial, but was less in mature rosette leaves. The association of GUS expression with proliferating regions suggested a role of BhHsf1 in regulating cell proliferation. Considering its role as a transcription factor, BhHsf1 likely regulate cell cycle related genes to restrict cell proliferation. Genome-wide genes expression profile uncovered that the over-expression of BhHsf1 repressed the genes involved in the pathways of the meiotic cell cycle and DNA replication, thus providing a strong evidence to support the role of BhHsf1 in regulation of cell proliferation via transcriptional regulation of target gene expression. We found the genes in these pathways were only slightly down-regulated, which could be explained that this type of genes are mainly limited in apical meristems and their meristem expression may be largely diluted in the sample of aerial part of the plants used for the microarray experiment. This is testified by the fact that AtCYCB1;1:GUS expression was decreased significantly in the actively proliferating zone of 35S:BhHsf1 lines, but AtCYCB1;1 transcripts were not detected differently expressed in OE lines and WT in microarray analysis.

Previously, it was suggested that endoreduplication was coupled with cell size determination (Jing et al. 2009; Tsukaya 2008); however, controversial results were also reported (De Veylder et al. 2001; Jasinski et al. 2002; Schnittger et al. 2003). For example, over-expression of the cyclin-dependent kinase (CDK) inhibitor genes NtKIS1a and AtKRPs in Arabidopsis leaves resulted in inhibited endoreduplication but the cells were relatively larger (De Veylder et al. 2001; Jasinski et al. 2002). In this study, we found that the DNA replication was deficient and the cell division and endoreduplication were affected in BhHsf1 OE Arabidopsis lines; however, the cell size was not different from that in WT. Thus our data suggested a partial uncoupling of cell size and nuclear DNA content.

The decline of cell proliferation is often observed when plants are exposed to drought, high salinity and freezing (Xiong and Zhu 2002). For example, mitotic and DNA-synthetic activities had declined when the moss Polytrichum formosum was subjected to dehydration (Mansour and Hallet 2001); soybean developed smaller leaves with less cell numbers under drought condition (Randall and Sinclair 1988). In plants, many components of the cell cycle machinery have been demonstrated to be conserved throughout the eukaryotes; however, it is not clear whether or how these proteins are regulated by stress response. Cyclin-dependent kinase (CDK) is a key factor in control of cell division (Cockcroft et al. 2000; West et al. 2004) and has been reported to be repressed during water stress (Setter and Flannigan 2001). CDK inhibitor ICK1 can block cells at G2/M but allow S-phase. The over-expression of ICK1 in Arabidopsis resulted in dwarf plants and enhanced drought tolerance (Kang et al. 2002; Wang et al. 1998, 2000).

The fact that several Hsfs from different plant species display dwarf phenotype suggests that the regulation of cell cycle may be another conserved mechanism of Hsfs to perform fundamental roles in the control of plant growth. According to our data on the stress-inducible expression and dual roles on enhancement of stress tolerance and growth retardation, it is possible that BhHsf1 may conduct signal transduction in both processes of stress response and growth inhibition, serving as positive regulator of genes related to stress tolerance and negative regulator of genes related to DNA replication and mitotic cell cycle synergistically. Therefore, it is likely that the rapid dehydration induction of BhHsf1 may attribute to the fast acquisition of desiccation tolerance and quick cease of growth in B. hygrometrica upon water loss. However, whether this is the case in resurrection plants is still a misery due to the lack of efficient genetic transformation methods.