Introduction

In natural ecosystems, all plants appear to be symbiotic with fungal endophytes (Petrini 1986; Rodriguez et al. 2009). Such symbiotic associations can have profound impacts on plants through increasing fitness by conferring abiotic and biotic stress tolerance (Brundrett 2006). Epichloë (Neotyphdium) endophytes, found in many cool-season grasses (Schardl et al. 2004), have been shown to confer enhanced fitness by protecting the host plant against certain nematodes, fungal pathogens, insect herbivores and mammalian herbivores (Siegel et al. 1985; Latch 1993; Schardl and Phillips 1997), as well as enhanced tolerance to drought and high temperatures (Arechavaleta et al. 1989; Bacon 1993). In addition, there are some evidences that colony growth of plant-pathogenic fungi is inhibited by Epichloë endophytes (White and Cole 1985; Christensen et al. 1991; Christensen and Latch 1991; Holzmann-Wirth et al. 2000), and that disease tolerance can be imparted by some but not all Epichloë species (Gwinn and Gavin 1992, Welty et al. 1991). Enhanced tolerance obtained from these mutualistic symbioses have been hypothesized to be the result of a diverse range of fungal secondary metabolites including ergot alkaloids, the L-aminopyrrolizidine (loline) alkaloids, the pyrrolopyrazine alkaloid peramine and so on that are induced by endophyte infection (Bush et al. 1997; Siegel and Bush 1997).

Drunken horse grass (Achnatherum inebrians) is a perennial bunchgrass distributed on alpine and subalpine grasslands in China (Shi 1997), which can occur in natural symbiosis with Epichloë gansuensis (Li et al. 2004b) and has been detected mainly in seeds, leaf sheaths and peduncles (Bruehl et al. 1994). Due to the increased tolerance to stress conditions provided by endophyte infection (Li et al. 2007a, b; Zhang et al. 2010), the distribution of A. inebrians infected by E. gansuensis (E+) is more extensive than non-infected A. inebrians (E−), especially in harsh conditions such as arid or semi-arid grasslands (Li et al. 2004a). Interestingly, the high concentration of the ergot alkaloids ergonovine and lysergic acid amide were detected in E+ A. inebrians which could contribute to stress resistance similar to the other mutualistic symbiosis (Miles et al. 1996).

In contrast to our knowledge about some physiological responses of E+ A. inebrians to environmental stress, the information available on the molecular basis of stress tolerance is not strong. With the development of next generation high-throughput sequencing technology, which is also referred to as Solexa deep-sequencing, a number of researchers have uncovered molecular-based mechanisms involved in the environmental stress response in plants and animals (Xiao et al. 2010; Luan et al. 2011). Millions of short RNAs and differentially expressed genes can be identified in a sample using this technology (Bennett et al. 2005). Taking advantage of this technology, we present here the genome-wide analysis of the differential expression of genes between E+ and E− A. inebrians when germinated under low temperature. The difference of the germination rate at 10 °C between E+ and E− A. inebrians was investigated and the differential expression genes were also identified using Solexa deep-sequencing.

Materials and methods

Plant materials

Achnatherum inebrians plants with mature seeds were harvested in Sangke grassland, Gansu province, China in October 2007. After harvest, E+ and E− A. inebrians seeds were isolated according to the method described by Li et al. (2004b). The two specimens then were kept under constant 4 °C storage at the Official Forage Seed Testing Center of Ministry of Agriculture, Lanzhou, China. No specific permissions were required for the locations/activities in this study. The field studies did not involve endangered or protected species.

Seed germination at different temperatures

Both E+ and E− A. inebrians seeds which were washed with sterilized deionized distilled water twice were incubated at 5, 10, 15, 20 and 25 °C in Petri dishes (15-cm diameter) containing wet sterilized filter paper with 12 h light and 12 h dark cycle. After imbibition, germination rate of seeds was monitored every 48 h until 25 days. The seeds showing radical extension of 1 mm were scored and removed from the Petri dishes. After 25 days, when no additional seeds had germinated for 5 days, the germination experiment was terminated. Each treatment was repeated five times independently with 100 seeds in each replicate.

Sample preparation and RNA extraction

Both E+ and E− A. inebrians seeds were washed with sterilized deionized distilled water twice and sown on sterilized filter paper and grown at 10 °C in constant temperature incubators for 2 days. Total RNAs were isolated from about 100 seeds of either E+ or E− drunken horse grass using the TRIzol reagent (Invitrogen) and incubated with DNase I (Takara) to remove residual genomic DNA according to manufacturer’s instruction. The concentration and quantification of all RNA were determined spectrophotometrically. Three independent replicates were performed to each sample.

Tag library construction and Solexa sequencing

For Solexa deep sequencing, the 3′-tag digital gene expression libraries were prepared as described in the Illumina DGE protocol. In brief, about 6 μg of total RNA was used to fractionate mRNA with oligo(dT) magenic beads. First- and second-strand cDNA was synthesized and bead-bound cDNA was subsequently digested with NlaIII. And then, the 3′-cDNA fragments attached to the oligo(dT) beads were ligated to the Illumina GEX NlaIII Adapter 1 containing a recognition site for the MmeI which cuts 17–18 bp downstream from the NlaIII site (CATG site) to produce tags with adapter 1. Subsequently, the restriction enzyme MmeI was applied to produce 21–22 bp tags starting with the NlaIII recognition sequence. After removing 3′ fragments with magnetic beads precipitation, an Illumina GEX adapter 2 was ligated at the site of MmeI cleavage. The resulting adapter ligated cDNA tags were enriched by 15 cycles of PCR performed with two primers that anneal to the ends of the two adapters. The resulting 85 bp fragments were purified using a 6 % Novex TBE PAGE gel and subjected to the Illumina/Solexa sequencing system (Illumina Inc., San Diego, CA, USA). Image recognition and base calling were performed using the Illumina Pipeline.

Data processing and statistical analysis

To obtain high quality data in the two libraries, raw sequence reads were filtered by the Illumina pipeline, removing 3′ adaptor tags, low quality tags and one copy tags. After filtering, we classified the clean tags according their copy number in the library and showed their percentage as total clean tags. Subsequently, due to the limited listing of A. inebrians mRNAs and ESTs in GenBank, all tags were mapped to rice reference sequences GCA_000005425.2 by SOAP (Li et al. 2008) as rice is relatively close to A. inebrians in genetic relationship. Only one mismatch was allowed and tags mapped to more than one transcript were excluded from our analysis. The ArrayExpress (http://www.ebi.ac.uk/arrayexpress/) accession number is E-MTAB-4199.

Cloning of selected transcripts 3′ end and semi-quantitative RT-PCR analysis

3′ end fragment sequence information of selected transcripts was obtained via 3′ RACE according to the GeneRcerTM Kit (Invitrogen). Briefly, reverse transcription was performed with about 1 μg of the total RNA of A. inebrians seeds using the GeneRacerTM oligo-dT primer [5′-GCTGTCAACGATACGCTACGTAACGGCATG ACAGTG (T)18-3′], to create RACE-ready first-strand cDNA with known priming sites at the 3′ ends. To obtain the 3′ ends, the PCR reaction was performed using the first strand of cDNA with forward gene-specific primer as shown in Table S1 and the GeneRacerTM 3′ reverse primer (5′-GCTGTCAACGATACGCTACGTAACG-3′), and then nested PCR was carried out with the same forward primers and 3′ nested primer (5′-CGCTACGTAACGGCATGACAGTG-3′). Individual PCR products were cloned into PMD18-T vector (TaKaRa) and sequenced with the universal T7 primer on an ABI 3730 DNA Sequence Analyzer. For RT-PCR analysis, primers were designed based on 3′ end sequence of selected genes using primer 5.0 version software, and a housekeeping gene, ACTIN, used as the internal control. The PCR primers are shown in Table S1.

Results

Seed germination at different temperatures

The germination rate of Achnatherum inebrians seeds was influenced by temperature (Fig. 1). With decreasing temperature, the germination rate of A. inebrians seeds decreased gradually. Interestingly, the decreased germination rate of E+ A. inebrians seeds was significantly slow compared with E− A. inebrians seeds at 10 °C (Fig. 1). However, a similar germination rate was observed in E+ and E− A. inebrians seeds at other temperature (15, 20, 25 and 30 °C) conditions. Thus, endophyte can improve the germination rate of A. inebrians seeds within a certain range of low temperature.

Fig. 1
figure 1

Germination of E+ (filled shape) and E− (unfilled shape) A. inebrians seeds in relation to incubation temperature. Each point is the average of n = 5 (100 seeds each) ± SE

Full size image

Analysis of differential gene expression libraries

To obtain a global view of differences in the transcriptional profile of E+ and E− A. inebrians seeds germinating under low temperature, Solexa deep-sequencing was employed to construct two gene expression profiles from E+ and E− A. inebrians seeds germinated for 48 h at 10 °C. Major characteristics of the two libraries were summarized (Table 1). In total, 3,649,469 (E+) and 3823829 (E−) tags were sequenced and filtered by the Illumina pipeline removing 3′ adaptor tags, low quality tags and one copy tags, and a total of 3,480,742 (E+) and 3,643,237 (E−) clean tags were obtained. The number of corresponding distinct tags were 112,754 (E+) and 113,028 (E−). All clean tags of the two libraries were subsequently mapped to the rice genome by the tolerances of one mismatch in each alignment. With this criteria set only 12.32 % (E+) and 12.40 % (E−) total clean tags were aligned to the reference genome, 10.38 % (E+) and 10.14 % (E−) of the distinct clean tags mapped unambiguously to the rice genes, and more than 80 % of the distinct clean tags in both libraries did not map to the reference database (Table 1).

Table 1 Summary of Solexa deep-sequencing profiles and their mapping to the reference genes
Full size table

Since the copy number of a tag reflects the quantitative level of gene expression, the distribution of tag expression was used to evaluate the normality of the differential gene expression data (’t Hoen et al. 2008). As shown in Fig. 2, similar patterns of the distribution of total tags and distinct tags over different tag abundance categories were found for the two libraries, suggesting there was no bias in the construction of the libraries from the two specimens. In addition, high-expression tags with copy numbers larger than 100 were overwhelmingly dominant (Fig. 2a, c) whereas low-expression tags with copy numbers less than 5 occupied the majority of distinct tag distributions (Fig. 2b, d). These results suggest that while the majority of mRNA is expressed at low levels, a small proportion of mRNA is highly expressed.

Fig. 2
figure 2

Distribution of total and distinct clean tags from imbibed E+ and E− A. inebrians seeds. a and b, distribution of E+ total and distinct clean tags respectively; c and d, distribution of E− total and distinct clean tags respectively

Full size image

To identify genes showing significant change in expression level, we mapped all clean tag sequences of these two gene expression libraries to the rice genome using the method described by Audic and Claverie (1997). Our results indicated that a total 254 genes showing differential expression with a log2 ratio >1 or <−1 (P < 0.001) were found between these two specimens (Fig. 3). Of those genes, 152 genes were considered significant with log2 ratio >2 or <−2, with 112 genes up-regulated and 40 genes down-regulated in the E+ A. inebrians.

Fig. 3
figure 3

Comparison of gene expression levels between imbibed E+ and E− A. inebrians seeds. TPM (transcript per million clean tags): the number of mapped tags per transcript per million clean tags (Li and Dewey 2011). With an estimated false-discovery rate (FDR) of <0.001 and |log2Ratio| ≥ 1 (ratio: E+/E−), the red represents up-regulated genes, the green represents down-regulated genes and the blue represents genes that are not differentially expressed. (Color figure online)

Full size image

Validation of representative genes by semi-quantitative RT-PCR

In order to experimentally validate differentially expressed genes derived from Solexa sequencing, the expression level of 6 modulated genes was analyzed further by semi-quantitative RT-PCR with gene-specific primers (Table S1). Due to the limited information available of A. inebrians genes, 3′ RACE was performed to obtain 3′ end sequence of corresponding gene (Table S2). The gene-specific primers were designed based on 3′ end sequence using Primer 5 software. The results normalized against ACTIN expression levels, which are among the more stably-expressed genes, show that expression of seven genes behaved similarly between RT–PCR and Solexa sequencing results (Fig. 4). Only one gene did not show consistent expression in the two results. Overall, the RT-PCR results mainly supported the Solexa sequencing results and confirmed differentially expressed genes in the E+ A. inebrians.

Fig. 4
figure 4

RT-PCR analysis of gene expression level for six selected genes. All data were normalized to the ACTIN expression level. The RT-PCR products were separated on agarose gels and were visualized with ethidium bromide under UV light

Full size image

Functional categorization of differentially expressed genes

To characterize the functional consequences of differentially expressed genes in E+ A. inebrians, all 152 differentially expressed genes were functionally annotated based on careful analysis of the Gene Ontology database and the scientific literature (Ashburner et al. 2000). All of the genes were manually assigned to eight categories of observed differences (Fig. 5). It was found that approximately 9 % of these genes are involved in the biosynthesis of alkaloids, 6 % are involved in biosynthesis of fatty acids, 10 % belong to stress response genes, about 24 % are involved in protein turnover, 11 % belong to transcription and signal transduction and further 13 % genes are proteins of unknown function.

Fig. 5
figure 5

Pie chart of functional categories of differential expressed genes from E+ and E− A. inebrians seeds in the germination process at low temperature

Full size image

Biosynthesis of alkaloids

One of the important differences in the transcriptional profiling between E+ and E− specimens was genes involved in the biosynthesis of tryptophan which is the precursor of ergot alkaloids (Robbers et al. 1972). Previous research has shown that exogenous tryptophan feeding or over-expression of genes involved in the biosynthesis of tryptophan can result in a significant increase of indole alkaloid accumulation (Robbers et al. 1972; Whitmer et al. 1998, 2002). In our present studies, almost all of the genes in this category were up-regulated dramatically an exception being a UDP-glucosyl transferase in the E+ specimen. Most of these genes regulate different key steps in the tryptophan biosynthesis pathway (Table S3), including glycolysis (two genes identical to 2, 3-bisphosphoglycerate-independent phosphoglycerate mutase and UDP-glucoronosyl/UDP-glucosyl transferase), shikimate synthesis (phospho-2-dehydro-3-deoxyheptonate aldolase, 3-phosphoshikimate 1-carboxyvinyltransferase and indole-3-glycerol phosphate synthase) and tryptophan conversion (transketolase and aminotransferase). In summary, up-regulation of tryptophan synthesis related genes may result in increased ergot alkaloids biosynthesis.

Fatty acid metabolism

The regulation of biosynthesis of fatty acids is important for the acclimation of higher plants to low temperature. In the present work we identified a series of genes involved in fatty acid metabolism (Table S3). Among these genes, two genes identical to acyl-desaturase which are involved in unsaturated fatty acids synthesis (Keil et al. 2010) were dramatically up-regulated. Furthermore, some genes related to fatty acid synthesis were also up-regulated in the E+ specimen, including one gene similar to acetyl-CoA carboxylase which catalyzes the formation of malonyl-CoA. In addition, an acyl-coenzyme A oxidase related gene was significantly down-regulated in the E+ specimen. Goto et al. (2011) have demonstrated that down-regulation of acyl-coenzyme A oxidase gene could enhance the fatty acid oxidation which results in the decrease of the levels of circulating and accumulating lipids in mammals. Different expression of these genes in E+ seeds of A. inebrians could change the composition and accumulation of fatty acids consequently improving its fitness to low temperatures.

Stress response

In our investigation, a set of modulated genes related to stress response were detected (Table S3). Among them, several genes associated with ROS scavenging function were upregulated in the E+ specimen, including homologs of peroxiredoxin, peroxidase precursor and copper/zinc superoxide dismutase. It is well known that stress conditions including cold, drought and salt can induce the accumulation of ROS such as super oxide, hydrogen peroxide, and hydroxyl radicals (Mittler 2002; Deng et al. 2014; Garg and Chandel 2015). Therefore, those ROS, which are known as signals, induced up-regulation of ROS scavenger related genes (Prasad et al. 1994). Moreover, two genes similar to heat shock proteins (HSPs) were up-regulated in E+ samples. HSPs, often called the stress proteins, are now recognized as important for stress resistance (Grover 2002). In addition, several stress respond proteins also dramatically increased, including a universal stress protein which is accumulated in plants infected with fungi under various environmental stress condition (Kvint et al. 2003). However, two genes similar to DnaJ domain containing protein were down-regulated. Blumberg and Silver (1991) have proved that the increased expression of SCJ1, a homologue of the bacterial heat-shock gene DnaJ, results in mis-sorting of a nuclear-targeted test protein.

Protein turnover

Genes related to protein turnover were also identified in our experiment. Most of them were prevalently up-regulated in E+ specimen (Table S3), including the genes identical to protein synthesis (e.g. ribosomal proteins) and protein degradation genes (e.g. protease and ubiquitin-conjugating enzymes). On the one hand, the levels of several proteins associated with protein translation were increased including ribosomal proteins (e.g. subunit of 26, 40 and 60 s ribosome and eukaryotic translation initiation factor) and protein translation initiation factor. Regulation of the translational machinery is considered as an integral component in the cellular stress response (Wood and Oliver 1999; Dhanaraj et al. 2007). On the other hand, there are several modulated genes related protein degradation were detected in our analysis, including ubiquitin-conjugating enzymes and some proteases (e.g. FtsH protease, Serine Carboxypeptidase and T1 family peptidase). Protein degradation during stress progress is a highly conserved and regulated phenomenon in all the organisms reported so far (Vierstra and Callis 1999).

Discussion

Numerous studies have shown that endophytes provide their grass partners with a series of fitness improvements (Siegel et al. 1985; Latch 1993; Schardl and Phillips 1997; Pedro et al. 2006). Germination is a complex process that could be affected by endophytes. It has been found that endophyte infection can increase the germination ability under environmental stress in Festuca arundinacea (Soleimani et al. 2010), Lolium perenne (Clay 1987) and Bromus setifolius (Novas et al. 2003). However, little is known regarding the molecular mechanisms involved in germination capability under stress conditions in endophyte-infected plants. In the present study, we analyzed the germination superiority of A. inebrians seeds infected with Epichloë endophyte at low temperature stress using Solexa deep-sequencing to identify candidate genes showing differential expression compared to endophyte-free A. inebrians. Germination experiments between two populations showed that E+ A. inebrians seeds have significantly higher germination ability at low temperatures. Especially at 10 °C, germination rate of E+ A. inebrians seeds is almost up to 50 %, whereas little is germinated in E− A. inebrians. Previous work has shown that Epichloë endophyte infection was of benefit to the germination of A. inebrians under high CdCl2 concentrations stress (Zhang et al. 2010).

Analysis of Solexa deep-sequencing identified candidate genes that may contribute to low temperature tolerance due to endophyte infection. In general, approximately eight thousands genes were investigated (Table 1), 152 of which (1.9 %) were differentially expressed (up or down-regulated). Moreover, semi-quantitative RT-PCR analysis of selected genes confirmed the reliability of the Solexa deep-sequencing data (Fig. 4). Further functional annotation indicated that a number of genes could be assigned to categories concordant with low temperature response. Induced changes of alkaloid biosynthesis genes were detected in the host plant, including induction of genes related to biosynthesis of tryptophan (Table S3) which is the precursor of ergot alkaloids (Robbers et al. 1972). In previous studies, the over-expression of tryptophan decarboxylase in Trichoderma harzianum resulted in enhancing alkaloids accumulation in a transgenic cell line (Canel et al. 1998; Whitmer et al. 1998, 2002). Our earlier work has demonstrated that an endophytic fungus and ergot alkaloids existed in A. inebrians (Li et al. 2007b). The presence of such fungi and associated high levels of ergot alkaloids (Miles et al. 1996), have become central in mechanisms to explain beneficial effects of endophytes on host plants (Schardl et al. 2004, 2007). Therefore, the increasing ergot alkaloid accumulation result from over production of tryptophan should contribute to the low temperature resistance in seed germination progress.

Several changes which are commonly associated with responses to various stress conditions were found in our study. Firstly, differential expression of genes related to fatty acid biosynthesis was observed (Table S3). Interestingly, some genes involved in unsaturated fatty acids synthesis were up-regulated in E+ A. inebrians. In higher plants, the most thoroughly accepted mechanism as a response to low temperature is increasing the level of unsaturated fatty acids of membrane lipids. For example, in the progress of maize germination at low temperature, Santis et al. (1999) found that the content of unsaturated fatty acid was dramatically increased with the decrease of the temperature. In addition, Prasad (1996) also suggested a possible involvement of the content of the unsaturated fatty acids in chilling tolerance of maize seedlings. Therefore, differential expression of those fatty acids related genes maybe result in a changed plasma membrane lipid composition of E+ A. inebrians seeds which could contribute to the enhancement of low temperature resistance. Moreover, most of the universal stress response related genes were up-regulated in E+ A. inebrians seeds (Table S3), including some genes homologous to antioxidase and heat shock proteins which are involved in scavenging ROS and maintaining proteins in their functional conformations respectively under various stress conditions. In addition, the expression of many genes related to protein turnover were modulated in E+ specimen, which may have resulted in accelerated protein translation and degradation (Table S3). It has been reported that the levels of protein synthesis is increased by stress conditions, in which many defensive proteins are produced resulting in improved stress resistance of plants (Ndimba et al. 2005). A number of previous studies have shown that protein degradation of numerous short-lived and regulatory proteins that control these cellular functions are involved in the regulation of stress response (Bochtler et al. 1999; Liu 2004; Reinstein 2004).

In summary, based on sequence similarities to genes in the rice genome database, a number of genes modulated in their response to low temperature in E+ A. inebrians were identified by Solexa deep-sequencing. The Solexa deep-sequencing technology is a powerful tool for high-throughput genome research, which has been also adapted for differentially expressed genes (Velculescu and Kinzler 2007; Licatalosi and Darnell 2010; Metzker 2010; Wang et al. 2016). Based on the knowledge of these differentially expressed genes, we have speculated their possible functional role in the germination process and stress tolerance of A. inebrians-Epichloë symbioses under low temperature. However, due to the limited genome information of A. inebrians, more than 60 % of candidate genes that may have novel functional roles in endophyte-grass associations could not be analyzed. Once more EST and genome projects of A. inebrians are completed, some of these differentially expressed genes could be further identified and functionally characterized in future. Briefly, this study provides the first global catalogue of modulated genes of A. inebrians infected with endophytes in germination under low temperature, together with their functional annotations and will help to elucidate the molecular basis of the stress resistance enhancement that endophytes provide their host plants.