Introduction

Plants have an ability to combat stresses of diverse nature by modulating defense responses physiologically and morphologically through induced and preformed defensive strategies. Plants also show such defense response to combat pathogen infection. The mechanism involves complex interplay of signaling cascades such as nitrous oxide, salicylic acid, ethylene and jasmonic acid. In most cases, the host arrests the growth of the pathogen at the site of penetration; such defense mechanism is termed as pattern triggered immunity (PTI) which includes morphological barriers such as defense structures produced before infection and antimicrobial compounds such as phytoanticipins and phytoalexins produced after infection. However, in few cases, the pathogen bypasses host immunity in the absence of host-resistant proteins or R proteins by secretion of effecter molecules, which lead to effector-triggered susceptibility (ETS) (Jones and Dangl 2006). On the other hand, effector proteins produced by the pathogen are recognized by host proteins and inhibit pathogen growth, which is known as effector-triggered immunity (ETI) (Jørgensen 1994). However, the defense mechanism of ETI or PTI is due to altered protein synthesis and its time bound degradation. Hence, a change in protein level is believed to be an indicator of host–pathogen interaction. Transcriptome analysis at time intervals of infection helps in the identification and isolation of the genes expressed in the host.

Chickpea (Cicer arietinum L.) is the third most important pulse crop of the world after peas and common bean and is grown under arid and semi-arid environmental conditions. India is a major producer, accounting for approximately 70% of world’s chickpea production with an average productivity of 889 kg/ha (Directorate of Economics and Statistics DAC 2016). The susceptibility of the crop to many biotic and abiotic factors is the main reason for its low yield (Dubey and Suresh 2006). Among biotic factors, wilt caused by Fusarium oxysporum f. sp ciceris (Padwick) Matuo and K. Sato (Foc) is considered as one of the main causes for low productivity of chickpea. In India, this disease is prevalent in all the chickpea-growing states and causes an estimated annual loss of 10% (Singh and Dahiya 1973). However, the losses due to Fusarium wilt depends on the stage of the crop infected. Early wilting causes 77–94% losses, while late wilting causes 24–65% losses (Haware and Nene 1980). All growth stages of the crop are susceptible to the disease, but its incidence is more severe at flowering and pod formation stages when temperature is relatively high (> 24 °C), particularly under drought conditions (Govil and Rana 1994). Global distribution of the disease is correlated with the presence of designated races of Foc (Jimenez-Gasco and Jimenez-Diaz 2003), and eight races of the pathogen (race 0, 1A, 1B/C, 2, 3, 4, 5 and 6) have been reported worldwide. Out of these, 0 and 1 B/C cause yellowing and rest of the races cause the wilting syndrome (Haware and Nene 1982; Jimenez-Diaz et al. 1993). Recently, Dubey et al. (2012) identified eight races of Foc from India with a new set of chickpea differential cultivars.

Studies on host response to pathogen at the molecular level are poorly understood in chickpea. The suppression subtractive hybridization (SSH) allows rapid identification and isolation of differentially expressed genes in a single experiment. The SSH is a powerful technique widely used in comparing two populations of mRNA and to isolate clones of differentially expressed genes (Diatchenko et al. 1996). It has been utilized widely to isolate and characterize differentially expressed genes during defense response against various pathogens in different crops, viz, coffee against Hemileia vastatrix (Fernandez et al. 2004), chickpea against Ascochyta rabiei (Coram and Pang 2005; Jaiswal et al. 2011), wheat against Puccinia striiformis (Yu et al. 2010) and cotton against Verticillium dahliae (Xu et al. 2011). Hence, the present study has been designed to comprehend chickpea–Fusarium interaction at the transcriptome level by the method of SSH. This study was aimed to identify and validate differentially regulated genes predicted to be involved in defense response in chickpea against Foc race 4 during compatible and incompatible interaction at early stage of the infection. Since differences in race are expected to yield specific results; it is therefore necessary to study against different races.

Materials and methods

Plant cultivars, fungal strain and inoculation

Chickpea (Cicer arietinum L.) cultivars, JG 62 (susceptible) and WR 315 (resistant), were selected for the present studies. Liquid inoculation method developed by Gurjar et al. (2012) was followed with little modification. Chickpea seeds were wrapped in wet blotter paper and stored at room temperature (24–26 °C) till germination. The germinated seedlings were transferred to trays containing half-strength Hoagland’s nutrient solution. These seedlings were kept under glasshouse conditions with 16 and 8 h photoperiod of day and night, respectively, at 25 °C and relative humidity of 40% at the National Phytotron Facility (NPF), Indian Agricultural Research Institute, New Delhi, India. Twelve days old seedlings were inoculated by root dipping with freshly prepared spore suspension of a highly virulent isolate, Foc53 (Race 4) of F. oxysporum f. sp. ciceris at 1 × 106 spores/ml for 10 min. The inoculated seedlings were shifted to Hoagland’s solution by adding 0.01% dextrose. The seedlings grown on a tray without pathogen inoculation served as the control. Chickpea root was collected from pathogen-inoculated and -uninoculated samples at 48, 72, 96 and 120 h post-inoculation (hpi).

RNA extraction and cDNA library construction

A total of 200 µg RNA was isolated by the RNA purification mini kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instructions. Equal amounts of total RNA from the pathogen-inoculated and -uninoculated samples harvested at different intervals were pooled separately. Two microgram (μg) of poly(A)+ mRNA was purified by mRNA isolation midi kit columns (Qiagen, Valencia, CA, USA) according to the manufacturer’s protocol.

An SSH library was constructed using a PCR select cDNA subtraction kit (Clontech., Palo Alto, CA, USA) following the manufacturer’s protocol. The subtracted cDNAs were then immediately cloned directly into pGEM-T Easy Vector (Promega, Madison, WI, USA) and transferred into DH5α E. coli electrocompetent cells. Detection of recombinant cells was carried out based on blue/white colony selection. Positive transformants were arrayed on 96-microtiter deep well plates and resultant cDNA clones were stored at − 80 °C.

cDNA inserts and differential screening

The presence of cDNA inserts was checked by PCR amplification in a 96-well Bio-Rad thermal cycler (Bio-Rad, CA, USA) using adapter primers (Table 1). PCR products (1.0 μl) of recombinant clones were arrayed onto Hybond-N+ nylon membranes (Millipore). The pathogen-inoculated and -uninoculated cDNA was labeled with ά-P32 radioisotope and hybridized with recombinant clones. The results from the hybridizations were recorded for each clone, and clones showing an intense hybridization signal with probes were selected for sequencing.

Table 1 Primer sequences of stress related genes used in semi-quantitative and quantitative PCR to analyze the expression patterns in chickpea against Fusarium oxysporum f. sp. ciceris
Full size table

Sequencing and sequence analysis

The positive clones were sequenced in both forward and reverse direction with T7 and SP6 primer sequence (Sigma, USA) and assembled into contigs using CAP3 sequence assembly tool (Huang and Madan 1999). These sequences thus generated were used for homology search on National Center for Biotechnology Information (NCBI) database using Basic Local Alignment Search Tools (BLASTx and BLASTn) (Altschul et al. 1990). E value higher than 10−5 were designated as nonsignificant homology with databases. The functional classification of annotated expressed sequence tags (ESTs) was done according to the classification described by Bevan et al. (1998).

Validation of representative defense-related genes

Total RNA was extracted from pathogen-inoculated and -uninoculated samples and the genomic DNA was removed from total RNA using RNase free DNase I (Fermentas, USA). The primers were designed with Primer3 plus software according to homology cDNA sequences and synthesized by Sigma, USA (Table 1). Synthesis of cDNA was done from 500 ng of RNA using Thermo Scientific Verso™ cDNA Synthesis Kit (ABgene, UK) according to the manufacturer’s protocol. Thirty nanogram of final cDNA concentrations was used for semi-quantitative PCR analysis and 18S ribosomal gene was used as the reference gene. The PCR conditions were 94 °C for 5 m and 20–25 cycles at 94 °C for 30 s, 58–62 °C for 30 s and 72 °C for 45 s and final extension at 72 °C for 10 m. The expression level based on the intensity of ethidium bromide staining was determined by visual observations (Fernandez et al. 2004).

Quantitative real-time PCR (qRT-PCR) experiments were performed on Bio-Rad iCycler (Bio-Rad, CA, USA) with SyBr green. Reaction mix (20 μl) containing 1 μl of cDNA (30 ng), 1 μl of each primer (10 μM) and 10 μl of SyBr Green PCR master mix (2×) and 7 μl of water. The following PCR conditions were performed: 95 °C for 10 min, followed by 40 cycles of 95 °C for 20 s, 50–55 °C for 30 s and 68 °C for 30 s. The PCR obtained products were subjected for melting curve analysis to evaluate primer specificity. Negative control without cDNA template was run for each analysis. Cicer arietinum 18S ribosomal gene was used as the housekeeping gene to normalize the expression of target genes and as a calibrator. The relative fold change was calculated from three replicates using the \(2^{{ - \Delta \Delta C_{\text{T}} }}\) method (Livak and Schmittgen 2001).

Results

Isolation of differentially expressed genes

Two cDNA libraries were constructed using resistant and susceptible cultivar against Foc to capture a wide spectrum of differentially expressed genes. These libraries contained 1143 (~ 90%) and 989 (~ 92%) cDNA inserts from compatible and incompatible interaction, respectively. The results also showed that the insertion size also ranged between 200 and 1000 bp and was mainly from 350 to 500 bp (Fig. 1). In Southern blot technique, clones whose intensity of hybridization was relatively different between two probes were selected for sequencing. A total of 202 ESTs were sequenced from 2132 recombinant clones, among which 102 were from compatible and 100 from incompatible interaction. After processing these sequences, 15.8% of the sequences were found redundant and 3.9% genes of fungal origin proteins were eliminated. As a result, a total of 162 unique sequences with 18 contigs and 144 singletons were obtained, of which 10 contigs and 68 singletons from compatible and 8 contigs and 70 singletons incompatible interaction were identified.

Fig. 1
figure 1

Agarose gel (1.25%) profile showing the amplified cDNA inserts of chickpea cultivar WR-315. Lanes 1–48 inserts of subtractive library; M—1 kb DNA ladder at both sides

Full size image

Characterization of expressed sequenced tags (ESTs)

Sequence analysis of the clones showed that more than 90% were of good quality and matched with the ESTs of different plant species of the NCBI database involved in a variety of cellular processes. Representative 162 ESTs were grouped according to their putative physiological functions, viz., defense-related genes, signal transduction-related genes, energy and metabolism-related genes, transcription and translation-related genes, cellular transport-related genes and hypothetical proteins. A total of 11.1% from compatible and incompatible library did not show homology with the known sequences of the gene bank database. The percent distribution of various classes of genes is depicted in Fig. 2. The largest category consisted of genes, which encode for proteins related to hypothetical proteins (22.8%), followed by energy and metabolism (20.3%)-related genes, defense and cell rescue-related genes (17.9%) and signal transduction-related genes (16%). The list of genes encoding for various functional categories for incompatible interaction are given in Table 2 and for compatible interaction in Table 3. Conserved domain search proved that the ESTs contains coiled coil, CCCH, leucine-rich repeats, U box, glycosyl hydrolase family 18, DnaJ/HSP40 family and Bet v 1 domain-containing proteins. These family proteins are reported to be involved in defense-related activity. Among both the libraries, some of the putative genes were found common.

Fig. 2
figure 2

Functional classification and relative distribution of identified genes in compatible and incompatible interactions

Full size image
Table 2 Representative resistance and stress related genes obtained by SSH of the incompatible interaction between chickpea and F. oxysporum f. sp. ciceris race 4
Full size table
Table 3 Representative resistance and stress-related genes obtained by SSH of the compatible interaction between chickpea and F. oxysporum f. sp. ciceris race 4
Full size table

Validation of differentially expressed genes

Twelve ESTs encoding defense-related proteins were further analyzed for the temporal expression by semi-quantitative and quantitative PCR analysis. In the present study, expression of putative genes encoding for superoxide dismutase (SOD), pathogenesis-related protein 4 (PR4), pathogenesis-related protein 10 (PR10), leucine rich repeat protein kinase, proline-rich cell wall protein, squalene monooxygenase, cysteine proteinase and cinnamate 4 hydroxylase expression was much higher in incompatible compared to compatible interaction (Fig. 3). These candidate genes were up-regulated at 48, 72 and 96 h after infection in the incompatible interaction. The expression of a putative superoxide dismutase expression was induced as early as 48 hpi in both the interactions. A putative PR10 protein expression was induced at 72 hpi and the highest expression was observed at 120 hpi in the incompatible interaction. The transcript level of a putative protein kinase expression in the present study was up-regulated at 48 and 72 hpi in the compatible and the incompatible interaction, respectively, and then dropped abruptly. The transcript level of a putative PR 4 protein expression was observed to be uniform in the incompatible interaction at all the hpi, and in the compatible interaction its expression was much higher at 48 and 96 hpi. In the resistant cultivar, proline-rich cell wall protein, cysteine proteinase and cinnamate 4 hydroxylase expression was maximum at 72 hpi, followed by 96 hpi and sharply declined at 120 hpi, whereas, squalene monooxygenase peak was observed at 96 hpi followed by 48 hpi. On the other hand, the expression of 18S ribosomal RNA was uniform throughout the experiment (Fig. 4).

Fig. 3
figure 3

Agarose gel (1.25%) showing RT-PCR analysis of putative genes of defense from susceptible (JG-62) and resistant (WR-315). UI uninoculated, hai hours after inoculation

Full size image
Fig. 4
figure 4

Quantification of relative expression level of defense-related genes by qRT-PCR. UI uninoculated, hai hours after inoculation

Full size image

Discussion

Chickpea root transcript was studied to understand molecular mechanism involved in resistance and susceptibility upon pathogen infection. For the first time, an attempt was made to develop the cDNA library in chickpea against Foc race four, a predominant race of North India. The diversity of partial sequences identified in the present studies would provide valuable insights into the biology of chickpea crop against Fusarium wilt. Based on previous work of Chatterjee et al. (2014) on transcriptome analysis in chickpea against Foc race 1, samples were collected at 48, 72, 96 and 120 h time points for capturing early response genes Gupta et al. (2009) also reported that Foc pathogen colonizes xylem vessels 96 h after infection, while significant transcriptomic alterations were observed at 48 h after infection. In the present study, homology search showed that the maximum hits were with Medicago truncatula which is the closest neighbor of legume. However, functional annotation is constantly being updated in this legume crop; hence, it reflected in the scores of chickpea. Though a complete draft genome sequence of chickpea has been reported recently, the functional annotations of genes need to be carried out to understand the host–pathogen interaction in the crop. A total of 162 unique sequences with 18 contigs and 144 singletons were taken into consideration for functional clustering. Out of these, 17.1 and 18.7% in compatible and incompatible interactions, respectively, encode for proteins related to defense processes, including regulation of oxidative burst, antimicrobial compounds and protein degradation. These genes include PR 10, PR 4, cysteine proteinase, superoxide dismutase (SOD), proline-rich cell wall protein, HSP 70, cytochrome 450, metallothionein, chalcone synthase, mitogen-activated protein kinases (MAPK), 4-coumarate-CoA ligase, and chitinases. These genes were also reported in the previous studies as differentially expressed in several host–pathogen interactions (Fernandez et al. 2004; Coram and Pang 2005; Jaiswal et al. 2011; Yu et al. 2010; Xu et al. 2011). Candidate genes identified in the present study such as superoxide dismutase, ubiquitin and 26S proteasome, cysteine proteases, metallothionein and BTB/POZ domain-containing protein at the early stage of the infection suggests the presence of reactive oxygen species during host–pathogen interaction. Some of the genes of this category have been shown to be involved in cross talk in various defense pathways (He et al. 1998). During stress conditions, peroxidation of polyunsaturated fatty acids occurs leading to cross talk of various genes and results in generation of reactive oxygen species (ROS). Putative BTB/POZ domain-containing gene was also shown to participate in defense mechanism in cotton against Verticillium dahliae (Xu et al. 2011) and BTB/POZ domain in NPR1 protein plays key role in regulation of systemic acquired resistance (Rochon et al. 2006). Genes encoding for ubiquitin and 26S proteasome play a key role in hormone signaling, oxidative burst and gene induction and programmed cell death (Jrujillo and Shirasu 2010). Putative cysteine proteases identified during early stage of infection in the present study were reported to play an active role in programmed cell death in Arabidopsis (Clarke et al. 2000). An oxidative burst takes place by production of ROS which was proven to cause cellular damage to both host plant and pathogen during various abiotic and biotic stresses (Barna et al. 2003). ROS precede the hypersensitive reaction, which is associated with the defense process during host–pathogen interaction. In plants, ROS was reported to damage carbohydrates, lipids, proteins and nucleic acids (Blokhina et al. 2003). The putative SOD obtained in the present study is also reported to be involved in the production of H2O2 that functions as signal molecule for programmed cell death and protecting neighboring cells from ROS by oxidative cross-linking of cell wall (Borden and Higgins 2002). Further, metallothionein proteins are also involved in the protection of neighboring cells from oxidative damage by scavenging ROS (Kumari et al. 1998). Several heat shock protein gene induction was also correlated with oxidative stress (Scarpeci et al. 2008). This clearly indicates the crucial role of candidate genes in protecting the plants against oxidative burst. The GTP-binding protein and ADP ribosylation factor identified in the present study were involved in the signaling process (Hou et al. 2007). Putative LRR receptor protein kinase proteins are involved in the signaling process by activating downstream MAP kinases. These signaling events converge into an MAPK cascade through phosphorylation and dephosphorylation, which confers resistance to both fungal and bacterial pathogens (Asai et al. 2002). The pathogenesis-related protein 10 (PR10) identified in the study is involved in the defense against diverse groups of pathogens in chickpea (Saikia et al. 2005) and also reported to be induced in roots of rice by biotic and abiotic stress through jasmonic signaling pathway (Hashimoto et al. 2004), which was up-regulated upon infection of rice blast fungus. A strong induction of PR10 in response to bacteria, fungi, wounding, jasmonic acid and ABA treatment in Lithospermum erythrorhizon was observed (Hwang et al. 2003). During the study, several classes of genes were involved in the production of antimicrobial compounds. Pathogenesis-related 4 protein is reported to have chitinase activity (Legrand et al. 1987) and is involved in defense against biotic stress and abiotic stress in rice (Wang et al. 2011). Flavonoids obtained in this study are secondary metabolites produced during host–pathogen interaction and are involved in defense mainly through the production of phytoalexins. A chalcone synthase is a key enzyme involved in isoflovonoid production during biotic stress (Gurjar et al. 2012). Cinnamate 4 hydroxylase and 4-coumarate-CoA ligase play an important role in the a key reaction of the phenylpropanoid pathway, which leads to the production of several secondary metabolites. Similar findings were reported by Lu et al. (2006) and Soria-Guerra et al. (2010), where putative 4-coumarate-CoA ligase was also involved in the biosynthesis of jasmonic acid, which plays a vital role in plant defense. Aoki et al. (2000) reported that chytochrome 450 plays an important role in biosynthesis of (iso)flavonoids such as medicarpin, glyceollins, genistein and daidzein, which interns play a central role in plant defense mechanisms and also helps in the transportation of toxic materials formed during pathogen infection into the vacuole. Hence, a similar function might be proposed for chickpea responses to Foc infection. 3.9% of genes originated from the fungus, indicating the presence of fungal growth and proliferation in host plant. Similarly, Gurjar et al. (2012) obtained 18% fungal originated genes during chickpea and Foc race 1 interaction in resistance cultivar.

The quantitative and semi-quantitative analysis of defense-related genes such as PR 10, PR 4, LRR protein kinase, cinnamate 4 hydroxylase, proline-rich cell wall, cysteine proteinase, SOD and squalene monooxygenase gene expression was significantly higher in incompatible interaction. The majority of the genes showed peak expression at 48 and 72 hpi after inoculation. The findings of present studies are consistent with Coram and Pang (2005), who reported peak expression of these genes in chickpea against A. rabiei at 24 hpi, which returned to normal level of expression at 96 hpi. They also reported up-regulation of PR 10 protein in the incompatible interaction as compared to the compatible interaction. Lo et al. (1999) obtained peak expression of PR 10 at 36 h and at 48 h after inoculation with Cochliobolus sublineolum in sorghum. PR 10 and PR 3 genes were overexpressed in a partially field-resistant cotton cultivar compared to highly susceptible cultivar after inoculation with F. oxysporum f. sp. vasinfectum (Zambounis et al. 2012). Compared with their expression levels in uninfected and compatible-type fungal-infected roots, all of the genes were highly expressed in incompatible-type interaction (Wang et al. 2014).

Conclusion

The present investigation has provided insights into the pathogen-responsive genes in chickpea. This study reported the ESTs involved in defense mechanisms elicited in response to Foc race 4. The ESTs which are submitted to the GenBank database may be great genomic resources for academics working on chickpea. The ESTs will also serve as potential resources for future genetic improvement for resistance to Fusarium wilt in chickpea cultivars. Functional characterization of unknown ESTs can be further characterized by 5′ rapid amplification of cDNA ends (RACE) to know their role in the defense mechanism. Temporal and quantitative change in the expression of proteins involved in defense probably determines the net outcome of interaction. Thus, identification of root-specific promoters and driving the defense-related genes in the root tissue would be a promising strategy for Fusarium wilt management.