Introduction

Coping with plant environment stress is the foundation of sustainable agriculture. Of all abiotic stresses, water deficit is regarded as the most adverse environmental factors limiting plant growth and crop productivity (Shinozaki et al. 2003; Bartels and Sunkar 2005; Gosal et al. 2009). Plants may actively respond and adapt to this stress through a series of morphological, physiological, cellular, and molecular processes (Reddy et al. 2004). These responses include stomatal closure, repression of cell growth and photosynthesis, activation of respiration, as well as accumulation of osmolytes and proteins specifically involved in stress tolerance (Shinozaki and Yamaguchi-Shinozaki 2007; Nakashima et al. 2009; Saibo et al. 2009), among which the molecular mechanism had been the major subject of many studies in the past decade (Bartels and Sunkar 2005; Xiao et al. 2007; Nakashima et al. 2009). Various genes are involved in water stress response in plants. The products of these genes function both in the stress response and in the establishment of plant stress tolerance (Shinozaki and Yamaguchi-Shinozaki 2007). The elucidation of molecular basis of plant responses to water stress might provide new strategies to improve the drought tolerance of agriculturally important plants (Xiong et al. 2002; Zheng et al. 2004; Rahaie et al. 2010). Up to date, a number of genes with diverse functions that responded to drought at the transcriptional level had been described in many annual plants (Shinozaki et al. 2003; Zheng et al. 2004; Wang et al. 2007), however, fewer cases had been obtained from trees (Liu et al. 2009). The existence of a variety of drought-inducible genes suggested that the responses of plants to water stress were extensively sophisticated.

Pigeon pea (Cajanus cajan), a perennial tree of the family Fabaceae, is an important grain legume of the semi-arid tropics and sub-tropics and is a hardy, widely adapted and drought-tolerant crop, which allow its cultivation in a wide range of environments and cropping systems (Varshney et al. 2011). Although compared with other grain legumes, pigeon pea ranks only the sixth in area and production; it is used in more diverse ways than others (Wu et al. 2009; Domoguen et al. 2010). In addition to its main use as protein-rich food, its immature green seeds and pods are also consumed fresh as a green vegetable (Ritchie et al. 2000). The crushed dry seeds are fed to animals while the green leaves form a quality fodder. The plant of pigeon pea also replenishes soil nutrients and controls soil erosion (ICRISAT 1998). With so many benefits at low cost, pigeon pea had become an ideal crop for sustainable agriculture systems in Karst areas of southwest China, where frequently suffers from the formidable water deficiency (Li et al. 2001; Gong et al. 2005).

An arbuscular mycorrhizae (AM) is a type of mycorrhiza in which the fungus penetrates the cortical cells of the roots of a vascular plant. The colonized plant is better nourished and better adapted to its environment. Recent evidence suggested that colonization of root systems by AM fungi (AMF) obviously afford host plants greater resistance to environmental stresses like drought stress (Augé 2001, 2004; Wu and Xia 2006). In our previous investigation, AMF colonization might substantially elevate the tolerance to drought of pigeon pea via cumulative effects (Qiao et al. 2011). However, in recent years, little has been known regarding the molecular mechanism of AMF colonization enhanced plant water deficit tolerance. Porcel et al. (2006) showed that AM plants responded to drought stress by down-regulating the expression of the plasma membrane genes and anticipating its down-regulation as compared to non-AM plants.

To isolate the drought-responded genes, as well as to unravel the molecular mechanism that AMF colonization further enhanced drought tolerance of pigeon pea plants, in the present study, suppression subtractive hybridization (SSH) and semi-quantitative reverse transcription (RT)-PCR strategies were used to identify differential expression genes implicating in drought stress from young pigeon pea seedlings inoculated by AMF in exposure to water deficit conditions.

Materials and methods

Plant material and stress treatment

Seeds of pigeon pea (cv. Guimu 1) that obtained from Guangxi Academy of Agricultural Sciences were surface sterilized in 0.1% (w/v) mercuric chloride for 5 min and washed five times (2 min each) in distilled water, then pre-cultured at 30°C for 24 h in incubator. Subsequently, the seeds were germinated in plastic pots (15 cm × 12 cm) containing either sterilized loam (ca. 1,000 g) collected from Karst area alone (non-AM) or sterilized loam inoculated with 50 g Glomus mosseea (GM, one of the common AMF) inoculum, which was purchased from the Plant Nutrition and Resource Research Institute, Beijing Academy of Agriculture and Forestry (China). All seedlings were grown under 70–80% (w/w) of field water capacity (FWC) in the greenhouse with a maximum photosynthetic active radiation (PAR) of 1,200 μmol m−2 s−1 and temperature of 25 ± 2°C before experiment treatment. After 90 days of growth, the colonization status was detected using magnified intersection method (Mcgonigle et al. 1990). Then, both AM-colonized (AMD) and non-AM (NAMD) seedlings were randomly chosen to carry out drought stress, whose water content of soil was kept around 50% FWC. The unstressed plants, both mycorrhizal (AMC) and non-mycorrhizal (NAMC), were watered daily to maintain 70–80% FWC based on the loss of soil weight.

Isolation of total RNA and mRNA

Young leaves were harvested separately at the 5th, 10th, 15th, 20th, 25th and 30th day after stress treatment, and kept at −80°C after being frozen in liquid nitrogen. Total RNA was extracted using PlantRNA reagent (Tiangen Biotech, China) according to the manufacturer’s protocol. The quantity and quality of isolated total RNA was examined by spectrophotometry and gel electrophoresis, respectively. Equal amount of total RNA from the six time points were mixed for each four treatment (AMC, AMD, NAMC and NAMD), and the mRNA was purified from the mixed RNA using Oligotex™ mRNA Midi Kit (Qiagen, USA) based on the manufacturer’s instructions.

Suppressive subtraction hybridization

SSH was performed using the PCR-Select cDNA Subtraction Kit (Clontech, USA) followed the manufacture’s instructions. The first-strand cDNA synthesis was carried out using 2 μg of the mixed mRNA which mentioned above. The forward library’ tester was made using mRNA from the AMD, and the driver was prepared using that from the AMC. The final PCR product was designated as subtracted cDNA-forward, which was enriched in genes upregulated in the stressed sample. Similarly, the reverse subtraction library was constructed except that the mRNA used for the tester and driver was reversed, and the obtained genes were downregulated on stress. Actin gene (Qiao and Wen 2010) was amplified for 18, 23, 28, and 33 cycles to test the subtraction efficiency of the library before cloning.

Construction of the subtracted cDNA library

Products of the final PCR from both forward and reverse subtraction were purified using QIAquick PCR Purification Kit (Qiagen, Germany). The purified products were then ligated into a pMD-18T vector (Takara Biotechnology, China) to transform Escherichia coli DH-5α cells. Recombinant white colonies were selected and the length of inserted segments in the library was determined by PCR using nested primers. PCR products were analyzed by electrophoresis on 1.0% (w/v) agarose gel to confirm the amplification quality and quantity.

Dot-blotting analysis

In order to screen positive clones, dot blotting was performed. The final PCR products were denatured by incubated at 95°C for 5 min and then cooled on ice immediately. Two microlitres of denatured PCR products were pipetted into the Hybond-N+ nylon membrane in a 96-well format and fixed by baking at 80°C for 2 h. Two forward-subtracted membranes were hybridized with two subtracted cDNA probes (forward- and reverse-subtracted), so were the two reverse-subtracted membranes. The probes were labeled with 32P α-dATP using Strip-EZ DNA Kit (Ambion, USA). Then hybridization and washing were carried out according to the manufacturer’s instructions. Result analysis and classification of differentially screened clones were performed according to the protocol recommended.

Sequencing and sequence analysis

Differentially expressed clones were sequenced by Shanghai Oebiotech in China. BLASTX of these sequences against protein database was performed at the NCBI (http://www.ncbi.nlm.nih.gov/BLAST). The functional categorization of sequence was carried out based on the AMIGO (http://amigo.geneontology.org/cgi-bin/amigo/go.cgi) and MIP functional catalog (http://mips.gsf.de/proj/funcatDB/search_main_frame.html).

Semi-quantitative RT-PCR

Semi-quantitative RT-PCR was employed to characterize 35 drought-induced ESTs, which probably implicated to abiotic stress tolerances based on the sequence analysis. In order to identify the contributions and genes affected by the presence of AMF to increased drought tolerance in pigeon pea, the NAMD and NAMC were used as reference. The actin gene was used as an internal standard. The mixed total RNA was reverse transcribed by M-MLV RTase. Then 0.5 μL of first strand cDNA were adopted as templates for PCR in 10 μL reaction volume. The parameters used for semi-quantitative RT-PCR analysis were listed in Table 1. Each semi-quantitative PCR was repeated at least twice. The PCR products were separated on 1% (w/v) agarose gel. Images of the electrophoresis gels were captured using BioRad Quantity One software, and the intensity of bands was quantified using Band Leader software (Magnitec) and normalized against actin gene band intensity. Expression profiles of these ESTs were also analyzed by the hierarchical average linkage clustering and k mean clustering in the Genesis software (Sturn et al. 2002).

Table 1 Parameters used for semi-quantitative RT-PCR analysis
Full size table

Results

Qualitative analysis of RNA

Using the RNA extraction kit described above, total RNA was isolated from leaves of both stressed- and unstressed AM-colonized seedling. Electrophoresis of the extracted RNA on agarose gel showed distinct 28 s rRNA band which was more abundant than the 18 s rRNA band, indicating little or no RNA degradation. Typical A260/A280 absorbance ratios of the RNA ranged from 1.88 to 1.93. Yields were in the range of 0.60–1.19 mg/g fresh weight. The A260/A280 ratios of purified mRNA were greater than 2.0. A clear smear greater than 0.5 kb was present on the 1% (w/v) agarose gel, justifying that the high quality mRNA was isolated from total RNA.

Evaluation of subtraction efficiency

Forward and reverse subtractions were conducted between stressed and unstressed seedlings infected by GM. The actin gene of pigeon pea was used to identify the subtraction efficiency. PCR amplification of the expected band showed that it appeared after 23 cycles when the unsubtracted tester cDNA was used as a template, conversely, this band did not appear until 28 cycles when the subtracted cDNA was taken as a template (Fig. 1), which indicated that cDNA homologous to both tester and driver had been considerably eliminated.

Fig. 1
figure 1

Evaluation of the subtraction efficiency of the subtracted cDNAs. PCRs with pigeon pea actin gene-specific primers were performed using subtracted and unsubtracted cDNAs from 30 days water deficit stress-induced pigeon pea leaves. Lane M, DNA size markers. Lane 1, 5, 9 and 13: 18 cycles. Lane 2, 6, 10 and 14: 23 cycles. Lane 3, 7, 11 and 15: 28 cycles. Lane 4, 8, 12 and 16: 33 cycles

Full size image

Differential screening of SSH library

The second PCR products of SSH were cloned into pMD-18T vectors after purification and preserved in E. coli DH-5α. Screening by blue-white spots demonstrated that approximately 98% of transformants contained inserts. A total of 768 clones were randomly picked from two SSH library. As the evidence of PCR screening, the average insert size of the SSH clones ranged from 250 to 1,000 bp, however, most were around 500 bp. Consequently, all the clones were further screened by dot-blotting analysis, among which 142 clones over expressed and 49 repressed were chosen to carry out the following analysis.

Sequence analysis of selected cDNA clones

A total of 191 clones were selected to have their inserts sequenced, and 182 differentially expressed cDNA section clones, including 133 upregulated and 49 downregulated, were obtained (Table 2). The sequences of these EST clones were submitted to NCBI as dbEST IDs 74605977–74606158 and GenBank accession nos. JK086989–JK087170. These inserts were searched against GenBank using BLASTX program. In order to best identify the functions of these ESTs, a moderate e value cutoff (1e−8) was set up to assign the ESTs as “significant homology” or “no hit” based on the BLAST results. A total of 142 ESTs consisting of 102 upregulated and 40 downregulated exhibited high homology to previously identified or putative proteins in Arabidopsis, rice, maize, Glycine max, etc. However, 31 upregulated and 9 downregulated ESTs showed no homology in the database, which probably was short fragments of cDNA derived from the 3′ UTR or was more likely novel ESTs in AM-colonized pigeon pea.

Table 2 Partial significantly changed transcripts in the suppression subtractive hybridization (SSH) library from pigeon pea (Cajanus cajan) inoculated by arbuscular mycorrhizae (AM)
Full size table

According to the MIP standards, the forward library and reverse library cDNAs with significant protein homology were sorted into 16 (Fig. 2a) and 12 (Fig. 2b) functional categories, respectively. Functional categories showed that the genes represented by those ESTs involved in a broad spectrum of biological pathways. For the upregulated genes, the metabolism category, energy and biogenesis of cellular components were the higher proportion of the known genes. For the downregulated genes, the largest set was assigned to the metabolism category. Cellular transport, transport facilities and transport routes and protein with binding function or cofactor requirement represented the second largest group in upexpressed and downregulated genes, respectively. The unknown proteins still represented a large group in both forward library and reverse library.

Fig. 2
figure 2

Functional classification of upregulated (a) and downregulated (b) clones identified from subtractive library based on MIPS functional categories. The percentage of gene transcripts in each group is listed

Full size image

Semi-quantitative RT-PCR

Semi-quantitative RT-PCR was carried out using specific primers based on the 35 selected ESTs which probably related to the abiotic stress tolerances according to the sequence analysis. Hierarchical cluster and k mean cluster analysis of the expression profiles demonstrated differences both AM-colonized and non-colonized seedlings under drought and control treatment. The results indicated that there were at least six patterns (Fig. 3). Six ESTs represented type I which had higher expression in AMC and NAMC than other treatments. However, type IV (6 ESTs) displayed higher expression in AMD and NAMD. Type V (4 ESTs) and type VI (6 ESTs) had similar expression patterns, which generally showed their highest expression in AMC, but the lowest expression in NAMD and NAMC, respectively. The expression level of four other ESTs (type III) peaked in AMD and bottomed in NAMC. Nine ESTs (type II) showed the lowest expression levels in NAMD and highest levels in AMD or NAMC.

Fig. 3
figure 3

Hierarchical cluster and k mean cluster analysis of selected ESTs derived from AM-colonized and non-AM pigeon pea under drought and control treatment. I, II, III, IV, V, and VI represent different types of expression patterns. The rows represent individual genes. The up- and downregulated proteins are indicated in red and green respectively. Color brightness represents the degree of the difference, as shown in the bar at the top (color figure online)

Full size image

Discussion

Abiotic stresses present a major challenge in the quest for sustainable food production as these may bring about 70% reductions in plant yield. Among the abiotic stresses, drought was labeled as the first adverse factors (Gosal et al. 2009). Pigeon pea was proved to be an ideal crop for Karst areas of southwest China. Previously, AM-colonized pigeon pea physiologically demonstrated a further tolerance enhancement to drought stress (Qiao et al. 2011). The illumination of molecular foundation of its responses to water deficit might facilitate the improvement of the drought tolerance. Recently, innovative biotechnological approaches had enhanced our understanding of the processes underlying plant responses to drought stress at molecular level (Gosal et al. 2009). In current investigation, SSH strategy was used to identify drought-responded genes preferentially expressed in AM-colonized pigeon pea seedlings. The 142 upregulated ESTs and the 49 downregulated ESTs were obtained after dot-blotting analysis.

In current investigation, we also detected semi-quantitative RT-PCR data agreed with the SSH data for 29 out of 35 data points, which indicated that 83% of clones of SSH were authentic positive. Seventeen differentially expressed genes, which upregulated based on the SSH analysis were more abundant in AMD in comparison with those of the AMC, such as cationic peroxidase 2 (VC-CF1), lipid binding protein (VC-DA8), ribosomal protein L14 (VC-AE4). Conversely, 12 differentially expressed genes were downregulated, e.g., peroxisomal glycolate oxidase (VD-CA4), myo-inositol 1-phosphate synthase (VD-DA11), cyclin-dependent kinase inhibitor (VD-BG4). The rests (6 differentially expressed genes) demonstrated contrary tendency compared with SSH data. The semi-quantitative RT-PCR data proved that SSH technology was also reliable to identify the differentially expressed genes involved in drought stress of pigeon pea inoculated by AMF.

Water stress tolerance might be ascribed to a complex cascade of molecular events including genes activation (Ramanjulu and Bartels 2002), and/or expression inhibition (Shinozaki et al. 2003). In the current case, most of the genes, irrespectively up or downregulated, demonstrated significant homology to the known genes from other plants (Table 2). Functional categories showed that the known genes represented by those cDNAs had a wide range of roles in different cell activities, e.g., metabolism, biogenesis of cellular components and energy. Although most of differentially expressed genes belonged to different categories, they presumably functioned not only in protecting cells from water deficit by the production of important metabolic proteins, but also in the regulation of genes for signal transduction in the water-stress response based on the sequence analysis. The former included proteins that probably function in stress tolerance, e.g., glutathione S-transferase (VC-CA6), ubiquitin-activating enzymes (VC-DC10) and s-adenosylmethionine synthetases (VD-BF5). Ubiquitin-activating enzymes played vital role in the degradation of many cellular proteins. Together with transcription control and protein phosphorylation, it regulated many basic cellular processes including cell cycle and antigen presentation (Glickman and Ciechanover 2002). The expression of s-adenosylmethionine synthetases during drought and re-watering after serious drought indicated that it was involved in the response to drought and re-watering and might be one of the key genes for drought tolerance and water use efficiency (Lin et al. 2008). The latter contained proteins involved in further regulation of signal transduction and gene expression that probably function in stress response, e.g., myo-inositol 1-phosphate synthase (VD-DA11), calmodulin-binding proteins (VC-AB10), and phosphatidylserine synthase (VC-AE12). Myo-inositol 1-phosphate synthase was a precursor of inositol phospholipids and played a crucial role in signal transduction, actin remodeling, phytic acid biosynthesis, membrane, and cell wall biosynthesis (Downes et al. 2005; Suzuki et al. 2007). Ca2+ and IP3 were the most probable candidates as second messengers in water-stress responses in plant cells, thus calmodulin-binding proteins and phosphatidylserine synthase were regarded to have important roles in various signal transduction cascades in plant as well as in yeasts and animals (Song and Yang 2006).

Currently, a total of 35 differentially expressed genes of pigeon pea were homologous to the known genes that putatively involved in the abiotic stress tolerances. The relative differences in mRNA levels between AM-colonized and non-colonized pigeon pea subjected to drought stress and stress free was determined by semi-quantitative RT-PCR. We found that the expression level of these selected genes were different in all treatment (Fig. 3). For instance, the protein phosphatase-2c (VC-AA9) and receptor-like protein kinase (VC-CB1) reached the highest expression in AMD and the lowest in NAMD. Biochemical and molecular genetic studies had identified that protein phosphatase-2c enzymes (Rodriguez 1998) and plant receptor-like protein kinases (Joshi et al. 2010) played a fundamental role in sensing external environmental signals to regulate gene expression. Therefore, the present evidences indicated the enhanced tolerance of AMF pigeon pea to drought stress might at least partially be ascribed to the changes in expression level of stress-related genes triggered by the colonization of AMF.

It is worthy to note that two transcription factors among water stress-modulated genes were identified, zinc finger proteins (VD-BD2) and AP2 domain-containing protein (VC-DF3). Zinc finger proteins played a key regulatory role in ABA signaling under drought stress (Golldack et al. 2011). In rice and tomato transgenic modification of zinc finger proteins transcription factor modified the tolerance of plants to water deficit and to salt stress (Amir Hossain et al. 2009; Hsieh et al. 2010). AP2/EREBP genes formed a large multigene family, and they played a variety of roles throughout the plant life cycle (Riechmann and Meyerowitz 1998). The expression of some members in AP2 domain-containing protein family was water stress inducible (Gorantla et al. 2005; Pandey et al. 2005). There had been many instances where overexpression of some AP2 domain-containing genes enhanced stress tolerance, including water stress tolerance (Fu et al. 2007; Dai et al. 2009).

Interestingly, we found that several ESTs had been reported to be induced by other forms of abiotic stress, not only by water stress. Myo-inositol 1-phosphate synthase (VD-DA11) and ribulose-1, 5-bisphosphate carboxylase small subunit (VD-BC12) had been identified in the salt-stressed halophyte smooth cordgrass (Baisakh et al. 2008) and alfalfa (Jin et al. 2010). Light-harvesting chlorophyll-binding proteins (VC-AB9) had been isolated from Lepidium latifolium inhibited by cold stress (Aslam et al. 2010). It was suggested that the different environmental stresses might result in similar non-specific responses at the cellular and molecular level, in addition to stress-specific responses, e.g., the elevation in the endogenous level of free proline under salinity condition (Gomathi et al. 2010), and in the malondialdehyde (MDA) accumulation by heat stress (Ding et al. 2010), or by oxidative stress (Guo et al. 2007). Similarly, our previous investigation demonstrated that the proline and MDA content of the drought-stressed seedlings were higher than that of the stress free (Qiao et al. 2011). It maybe due to the fact that multiple stressors trigger similar downstream signal transduction chains (Jin et al. 2010).

The current investigation focused on the differentially expressed genes that were isolated from young pigeon pea seedlings inoculated by AMF in exposure to drought stress, which may probably inflected that AMF colonization further enhanced drought tolerance. The following work will aim to further investigate whether the expressions of these differentially expressed genes are strictly triggered by AMF or not. Additionally, a large number of response genes regulated by water deficit stress in this study encode unknown proteins, and the ongoing work will also extend to verify how they involved in the drought tolerance, which are essential to fully elucidate the molecular mechanism of the enhanced tolerance to drought stress of pigeon pea caused by AMF, as well as to further improve its stress tolerance via genetic manipulation.

Author contribution

Xiaopeng Wen designed the research and instructed the experiment. Guang Qiao and Lifei Yu performed the experiment and analyzed the data. Xiangbiao Ji gave materials support. Guang Qiao and Xiaopeng Wen prepared the manuscript.