Introduction

Tamarix species, which include several shrubs and small trees, are widely distributed in drought-stricken areas and saline or saline-alkali soil in Central Asia and China. They are highly tolerant of salt, drought and high temperature. Tamarix hispida, commonly called the Kashgar tree, grows to approximately 1 m in height and can thrive in soda or saline soil. Like many in the genus, its tolerance of saline conditions indicates the presence of well-developed salt-resistance mechanisms and may be a potential source of information about the genetic determinants for saline tolerance. Approximately 6.4% of the world’s terrestrial surface area is covered by saline-alkali soils where plants must endure concomitant saline and alkali stress, making it difficult to grow (Zhang et al. 2006). The cloning of saline-alkali tolerance genes and determination of their expression patterns under saline-alkali stress, may offer some attractive candidate genes and valuable information for improving saline-alkali tolerance of other species through genetic engineering.

Expressed sequence tags (ESTs) analysis is an effective method to discover novel genes and investigate gene expression in different organs and tissues. This technology is a commonly-used approach to identify genes involved in specific biological functions, especially in organisms where genomic data are not available (Gueguen et al. 2003). The EST approach has been successfully used to identify stress tolerance genes and to determine the expression patterns induced through abiotic stress in several plants, including Selaginella lepidophylla (Iturriaga et al. 2006), Leymus chinensis (Jin et al. 2006), Triticum aestivum (Houde et al. 2006), Populus nigra (Nanjo et al. 2004), Tortula ruralis (Oliver et al. 2004), Avicennia marina (Mehta et al. 2005), Mesembryanthemum crystallinum (Kore-eda et al. 2004), Suaeda salsa (Zhang et al. 2001) and Thellungiella halophila (Wang et al. 2004).

In this study, we constructed three cDNA libraries from T. hispida leaf tissue after exposure to NaHCO3 for 0, 24, and 52 h, and obtained 3,152, 3,307, and 2,988 valid EST sequences from the three libraries, respectively. These ESTs were assembled into 986 contigs and 2,959 singlets. A total of 1,752, 1,558 and 1,675 unigenes were obtained from the three libraries, respectively. Among them, 7,365 ESTs (77.96%) showed significant similarities (E-value < 10−4) to gene sequences in the non-redundant (Nr) GenBank database, and the other 2,082 (22.04%) had little, or no, similarity. The gene expressions were compared among the three libraries, and the gene expression patterns following NaHCO3 stress were obtained, resulting in the identification of some pathways where stress tolerance may be involved. The results showed that saline-alkali stress promoted pronounced differences in the gene expression profiles. The genetic information derived from T. hispida, a remarkably salt-tolerant plant, can contribute to the characterization of genes responding to saline-alkali stress, and is expected to aid our understanding of the potential genetic determinants of saline-alkali resistance in woody plants in general, and of Tamarix species in particular.

Materials and methods

Material treatments and cDNA library construction

The T. hispida seedlings were grown in a mixture of turfy peat and sand (2:1 v/v) in a greenhouse with 75% relative humidity and an average temperature of 24°C. The well-watered seedlings were exposed to a solution of 0.4 M NaHCO3 for 0 (control), 24, and 52 h. Leaves were harvested and immediately stored at −70°C until RNA isolation. Total RNA was extracted by the CTAB methods and total RNA was digested with DNase to remove DNA contamination. The cDNA libraries were constructed with a CreatorTM SMARTTM cDNA Library Construct Kit (Clothtech, PT3357-1) according to the manufacturer’s instructions. Total RNA was reverse transcribed, and double-stranded cDNA was synthesized. After digestion with Sfi I, the cDNA was size-fractionated, over 400 bp in length, using a drip column containing Sepharose CL-2B gel filtration medium, and ligated into the pDNR-LIB vector. Ligated plasmids were transformed by electroporation into Escherichia coli DH5α using an electroporator 2510.

DNA sequencing and analysis

The cDNA library clones were plated onto LB-chloromphenicol plates. Individual colonies were transferred to 96-well microtiter plates, containing 1 ml LB media supplemented with 34 μg chloromphenicol, and incubated overnight at 220 rpm and 37°C. Plasmid DNA was prepared using an alkaline lysis preparation protocol (Sambrook et al. 1998). The sequencing reaction used the DYEnamic ET Dye Terminator Kit (GE Healthcare) according to the manufacturer’s instructions. DNA Sequencing was performed in a MegaBACE1000 DNA capillary sequence. Single-pass sequencing was performed from the 3′-end of the cDNA using a M13 primer.

DNA sequence data processing and analysis

Raw single-pass sequence data demonstrating poor quality vector sequences or sequences less than 100 bp were removed. The remaining sequences were analyzed. Contigs were built using the CAP3 assembly program (Huang and Madan 1999). The tentative unique genes (TUGs) were compared against the Nr GenBank database using BlastX and BlastN (Altschul et al. 1997) with a cutoff E-value of 10−4. All the ESTs were assigned to functional groups according to Bohnert et al. (2001) with modification based on top BLASTX hit alignments. To investigate the genetic relationship between T. hispida and other plants, all the tentative unique genes from T. hispida were compared against the genomes of Arabidopsis thaliana, Avena sativa, Beta vulgaris, Glycine max, Hordeum vulgare, Lotus japonicus, Solanum lycopersicum, Tritucum aestivum, Zea mays, Oryza sativa and Manihot esculenta using Blastn (http://www.ncbi.nlm.nih.gov/BLAST/Genome/PlantBlast.shtml?10), with a cutoff E-value of 10−4.

The comparison of gene expression from 0, 24, and 52 h libraries was analyzed according to Stekel et al. (2000). The statistic, denoted R j for gene j, is given by the expression

$$ R_{j} = {\sum\limits_{i = 1}^m {x_{{i,j}} } }\log {\left( {\frac{{x_{{i,j}} }} {{N_{i} f_{j} }}} \right)} $$

where m is the number of cDNA libraries, x i,j is the number of transcript copies of gene j in the ith library and N i is the total number of cDNA clones sequenced in the ith library. f j is the frequency of gene transcript copies of gene j in all of the libraries, given by the formula

$$ f_{j} = \frac{{{\sum\nolimits_{i = 1}^m {x_{{i,j}} } }}} {{{\sum\nolimits_{i = 1}^m {N_{i} } }}} $$

In a library in which there are no observed copies of the gene, that is, x i,j  = 0, its contribution to R j is zero.

Real-time quantitative RT-PCR

To investigate the differences in gene regulation between T. hispida exposed to NaCl or NaHCO3, nine transcripts, differentially regulated by NaHCO3 stress, were selected for expression analysis using real time PCR. Well watered T. hispida seedlings were treated with a solution containing 0.4 M NaCl for 0, 24, and 52 h. Total RNA from harvested leaves was isolated and digested with DNaseI to remove DNA contamination. About 5μg of total RNA from each pool were reverse transcribed in the presence of Oligo (dT) and nine random primers in a volume of 20 μl. The synthesized cDNA was diluted with 180 μl of water and used as the template in real-time RT-PCR. PCR primers were designed to amplify target cDNA fragments (Table 1). The 18S rRNA (http://www.ncbi.nlm.nih.gov/entrez/viewer.fcgi?db=nucleotide&val=126032140) gene was used as an internal reference to normalize the amount of total RNA present in each reaction. The reactions were conducted in triplicate on a MJ Research OpticonTM2 machine (MJ Research). PCR was performed in a total volume of 12 μl containing 6 μl of QuantiTect SYBR-green PCR Master Mix (Qiagen), 0.5 μM of forward and reverse primers, and 2 μl cDNA template (equivalent to 0.05 μg of total RNA). The amplification was completed with the following cycling parameters: 94°C for 30 s followed by 45 cycles at 94°C for 12 s, 58°C for 30 s, 72°C for 45 s and 1 s at 78.5°C for plate reading. A melting curve was generated for each sample at the end of each run to assess the purity of the amplified products. After reactions, all samples were also electrophoresed in agarose gels to verify implication of the target fragments. The expression levels of the clones were calculated from the threshold cycle according to 2−ΔΔCt (Kenneth and Thomas 2001).

Table 1 Nucleotide sequences of primers used in RT-PCR to analyse gene expression under NaCl
Full size table

Results and discussions

Construction of cDNA libraries and single-pass sequencing of cDNAs

Three cDNA libraries were constructed using data from the leaf tissue of T. hispida exposed to NaHCO3 for 0, 24, and 52 h, and were designated 0 h, 24 h and 52 h libraries, respectively. PCR amplification revealed that the average length of the insert in the three cDNA libraries was 1.1, 1.2 and 1.2 kb, respectively. A total of 9,447 valid EST sequences were obtained from the three libraries. They were submitted to the GenBank databases with accession numbers EH048198 to EH057644. The 9,447 ESTs were assembled into 986 contigs and 2,959 singlets. The numbers of unigenes from the 0, 24, and 52 h libraries were 1,752, 1,558 and 1,675, respectively (Table 2). And there are 526 genes with R > 1 and 16 genes with R > 3 in the 3,945 unigenes (Fig. 1). The majority of the clusters (R > 3) are annotated, and their functions mainly involve in cell rescue and defense and photosynthesis, suggesting that NaHCO3 stress strongly affected this two pathways.

Table 2 General characteristics of Tamarix hispida ESTs
Full size table
Fig. 1
figure 1

The number of genes for a given value of the test statistic R is plotted as a function of R

Full size image

EST sequences analysis

In order to compare the frequency of ESTs present within each of the three cDNA libraries, the total unigene sets were analyzed for expression overlap. Only a limited number (244 unigenes) were found to overlap among the three libraries (Fig. 2). The overlap of transcripts between the 24 and 52 h libraries (437 transcripts) was slightly larger than between other library combinations (387 between 0 and 24 h; 412 between 0 and 52 h). In contrast, large numbers of library-specific unigenes and singletons: 1,173 (67.9%), 964 (62.4%) and 1,060 (63.7%) were specifically expressed in the 0, 24, and 52 h libraries, respectively. These results suggest that the saline-alkali exposure strongly affected the gene expression in stressed leaves; a likely result of the diversity of transcripts after stress stimulates encoding of stress-adaptive determinants.

Fig. 2
figure 2

Distribution of EST contigs among three cDNA libraries from unstressed (for 0 h) and salt-stressed (for 24 and 52 h) leaves. Numbers in two or more rings indicate the overlap of EST-derived contigs present within each individual library

Full size image

Genetic similarities between T. hispida and some plants

The BLASTX search indicated that the known ESTs are mainly matching proteins from Arabidopsis thaliana and Oryza sativa, etc. The top six matched plants are shown in Table 3. To investigate the genetic similarities of a nucleic acid sequence between T. hispida and other species, all our ESTs were searched against the plant genomes having data in the National Center for Biotechnology Information (NCBI) database using the basic local alignment search tool (BLAST) program, with a cut off E-value of 10−4. The analysis revealed that 2,999 (31.64%) ESTs matched the genome sequence of Oryza sativa, Arabidopsis thaliana and Manihot esculenta. Among them, 15.36% of T. hispida ESTs were similar to the genome sequences of Oryza sativa, a larger proportion than between T. hispida and the other two species.

Table 3 Genetic similarities analysis
Full size table

Characterization of gene expression following saline-alkali stress

The EST sequences were compared to the NCBI Nr database using the BLASTX algorithms (Altschul et al. 1997). A total of 7,365 ESTs (77.96%) showed significant similarities (E-value < 10−4) to gene sequences in the Nr database; the remaining 2,082 (22.04%) ESTs showed little or no similarity. The database-matched ESTs in each library were classified into 12 categories (Table 4).

Table 4 Number (abundance) of ESTs in 0 h and salt-stressed (24 h and 52 h) leaf cDNA libraries of Tamarix hispid a
Full size table

The abundance of cell growth, division and metabolism genes increased obviously following saline-alkali stress for 52 h (Table 4), implying that these two pathways may be enhanced to mediate stress. However the abundance of energy and protein destination genes obviously decreased following stress (Table 4), suggesting that the saline-alkali stress weakened the energy metabolism and protein destination. The abundance of defense, photosynthesis and transcription ESTs was highly up-regulated after exposure to NaHCO3 for 24 h. The temporarily-enhanced activation of these pathways may strengthen the ability of T. hispida to tolerate NaHCO3. Additionally, the proportion of transport facilitation ESTs in the 52 h library was 8.23%, which was higher than in the 0 h library (6.85%). T. hispida is a salt-secreting halophyte. Its leaf tissues can accumulate and sequester high concentrations of salt. The increase of transport facilitation ESTs in the saline-stress library may indicate that the ion and water transport pathways are enhanced following NaHCO3 exposure, possibly as a mechanism to compensate for high salt concentrations.

Highly redundant ESTs

The EST analysis is a rapid and powerful tool for discovery of gene expression. If a large number of ESTs are obtained randomly from an unbiased cDNA library, the number of ESTs fitting a particular gene should reflect the abundance of the corresponding transcript in the tissue from which the library was derived (Beisson et al. 2003). The top 10 most redundant ESTs out of 9,447 ESTs (84 copies or more) are listed in Table 5. These abundant ESTs accounted for the highest proportion in the 24 h library (20.08%), followed by the 52 h (15.93%) and 0 h libraries (12.32%), indicating that most of the abundant transcripts displayed a temporary increase in response to salinity stress. These transiently-increased genes may play roles in adaptation to stress conditions.

Table 5 The 10 largest EST clusters and their representation among the three T. hispida libraries
Full size table

Stress tolerance involves various pathways

The abundances of transcripts in each library were compared. Genes expressed differentially after exposure to NaHCO3, with some showing an increase in frequency, while several others decreased in frequency after the 0, 24, or 52 h exposure period (Table 6). The genes that were up-regulated were involved in a variety of functional areas, including defense, protein synthesis and destination, photosynthesis, transport, and some with an unknown function. That so many genes responded (both up-regulated and down-regulated) after NaHCO3 exposure provides evidence that the saline-alkali stress tolerance mechanism in T. hispida is a complex network involving multiple physiological and metabolic pathways.

Table 6 The redundant transcripts whose expression increased or decreased after exposed to NaHCO3 treatment
Full size table

Analysis of the unknown ESTs was a useful starting point from which to isolate new genes governing salt tolerance (Mehta et al. 2005). In this study, we identified some unknown function and novel genes that were up-regulated under saline-alkali stress. The obtained sequences and expression patterns of these genes may (1) provide a foundation for understanding their roles in saline-alkali tolerance in T. hispida, and (2) provide a good candidate pool from which to choose novel stress tolerance genes.

The up-regulated transcripts included Jasmonate-induced protein (EH056009), auxin-repressed protein-like protein ARP1 (EH056131) and Gibberellin-regulated protein (EH057297), suggesting that hormone signaling transduction may mediate stress tolerance in T. hispida. Jasmonic acid (JA) and its derivatives, commonly called jasmonates, are hormonal regulators involved in plant responses to abiotic and biotic stresses (Creelman and Mullet 1997; Wasternack and Parthier 1997). An endogenous rise in jasmonates is functionally relevant because of jasmonate-induced gene expression. There is a strict correlation of the expression of JIP23 (jasmonate-induced protein, a 23-kD protein) and enhanced endogenous JA levels (Kramell et al. 2000). In this study, the transcript representing Jasmonate-induced protein was not found in the 0 h library, but it was present in the 24 and 52 h libraries (1.03% and 1.84%, respectively) (Table 6). The significant increase in endogenous JA levels after induction of salt stress suggests an important role in stress mediation.

Glycine-rich RNA-binding proteins (GRPs) have been implicated in the post-transcriptional regulation of gene expression in plants under various stress conditions. Previous studies suggested that glycine-rich RNA-binding protein (GRP) may play a role in plant responses under changing environmental conditions, as their mRNA levels increased following exposure to cold, wounding, water stress, abscisic acid, viral infection, UV radiation and heavy metals (Sachetto-Martins et al. 2000; Kwak et al. 2005). We also identified a GRP (EH054828) genes as a saline-responsive gene, since its expression was up-regulated more than three-fold following salt stress. The up-regulation of GRP indicates that it may exhibit RNA chaperone activity during the saline-alkali adaptation process.

Late embryogenesis abundant (LEA) proteins accumulate during the late stages of embryogenesis. It has been proposed that they play a role in desiccation tolerance based on their accumulation and physicochemical properties. Dehydrins (DHN), a class of LEA proteins, were also involved in desiccation tolerance of plants (Momma et al. 2003). In Ginkgo biloba, five different DHN genes were found to be expressed in stressed plants but not in well-watered ones (Deng et al. 2006). Similarly, we observed that a DHN gene, while not present in the 0 h library, increased progressively in the 24 and 52 h libraries (0.09% and 0.13%, respectively). Another LEA gene, (LEA-18), was also highly increased following stress for 24 and 52 h (Table 6). LEA family genes, therefore, may play important roles in saline-alkali tolerance in T. hispida.

Salt stress induces the generation of reactive oxygen species (ROS) in plants, and therefore it is important for plants to have effective ROS-scavenging mechanisms. There is a high proportion of ROS-scavenging ESTs included in the three cDNA libraries (3.08–4.60%), and some ROS-scavenging ESTs were up-regulated following stress. The metallothionein (MT) (EH052305) and thioredoxin (EH051022) genes were up-regulated following saline-alkali stress for 24 h (Table 6). The MTs are capable of scavenging ROS generated by abiotic stress (Mir et al. 2004; Wong et al. 2004). Thioredoxins are involved in oxidative damage avoidance as components of the plant antioxidant network and may also act to regulate scavenging mechanisms (Vieira and Rey 2006). The up-regulation of MTs and thioredoxins suggests they play important roles in scavenging ROS generated by T. hispida under stress conditions.

Salt-tolerance also involves changes in the levels and composition of fatty acids of the major glycerolipids in roots and leaves of a wide range of plants. The function of nonspecific lipid transfer proteins (LTP) in vivo remains unknown, but they bind and catalyze the transfer of lipids in vitro, and they are responsive to abiotic environmental changes such as drought, cold, heat and salt and also infections by bacterial and fungal pathogens (Jung et al. 2003; Jang et al. 2004; Wu et al. 2004; Liu and Lin 2003; Lindorff-Larsen and Winther 2001). We investigated the expression of two nonspecific LTP genes (EH053076, EH051895) after exposure to NaHCO3 and NaCl for different periods of time. The resulting data showed that they were all up-regulated following NaHCO3 or NaCl salt stress for 24 h (Fig. 3), suggesting an active lipid metabolism or/and salinity defense during salt stress.

Fig. 3
figure 3

Comparison of gene expression patterns in response to NaCl and NaHCO3 The X-axes lists treatment time points and Y-axes shows gene expression level compared with control. (▲) show gene expression in libraries in response to NaHCO3 (calculated by formula: gene expression levels = copy numbers in these libraries/unstressed library) and (◆) show gene expression in response to NaCl. (a) lipid transfer protein (EH051895); (b) ethylene responsive element binding protein (EH056776); (c) drought-induced protein SDi-6 (EH051954); (d) heat shock protein DnaJ (EH055899); (e) glyceraldehye-3-phosphate dehydrogenase (EH054170); (f) ultraviolet-B-repressible protein (EH056631); (g) metallothionein (EH056653); (h) lipid transfer protein (EH053076); (i) Salt tolerant protein (EH057224)

Full size image

In addition to short-term response and regulatory mechanisms, a functional system for reestablishing homeostasis is key to long-term tolerance. Both ionic and osmotic homeostasis must be restored. Many genes involved in ion homeostasis were found in T. hispida, such as vacuolar H+-ATPases (EH055782), chloroplast ATP synthase (EH056157), plasma membrane H+-ATPase (EH049661) and Na/H antiporter Nhx1 (EH051170), showing that T. hispida has efficient homeostatic pathways. One of the functions of vacuolar H+-ATPases in plants is to help create a proton electrochemical gradient across the tonoplast. The activity of a vacuolar H+-ATPase in salt cress increased by a factor of 1.5–2.5 after treatment with 200 and 400 mM NaCl for 2 weeks, respectively (Vera-Estrella et al. 2005). Similarly, we also found the expression level of a vacuolar H+-ATPase (EH055782) was increased two-fold in response to NaHCO3 stress for 52 h. This compound may function to establish an electrochemical H+ gradient across the tonoplast that energizes sodium sequestration into the central vacuole during NaHCO3 stress.

Transcriptional regulation of ribosomal proteins may also be involved in saline-alkali stress. Ribosomal proteins are closely involved with replication, transcription, DNA repair, RNA processing, regulation of development and malignant transformation (Blasi et al. 2002). Transcriptional regulation of ribosomal protein expression is an important mechanism in controlling ribosome assembly and function (Posas et al. 2000). Also, ribosomal proteins have been shown to play roles in stress tolerance and salt-adaptation in plants (Herruer et al. 1998; Kawasaki et al. 2001; Wu et al. 2005). In this study, many stress-responsive ribosomal proteins were identified. Interestingly, some ribosomal proteins were up-regulated while others were down-regulated, suggesting that ribosomal protein genes expression may be enhanced the saline-alkali tolerance of T. hispida.

Heat shock proteins (HSP) are a group of proteins whose expression is increased when the cells are exposed to elevated temperatures. Production of high levels of heat shock proteins can also be triggered by exposure to different kinds of environmental stress conditions, such as infection, cold stress and salt stress (Hwang et al. 2005). In this study, we identified three HSP genes that respond to NaHCO3 stress. The transcripts of a 70 KD heat shock cognate protein (EH054733), heat shock factor DnaJ (EH055899) and heat shock protein (EH052669) were up-regulated following NaHCO3 stress (Table 6). These proteins may affect stress tolerance through folding and assisting in the establishment of proper protein conformation or prevention of unwanted protein aggregation.

Photosynthesis regulation may be important in stress tolerance

Photosynthesis is strongly affected by salt stress. Water stress, reduction of chloroplast stromal volume and generation of active oxygen species are also thought to play important roles in inhibiting photosynthesis, and some photosynthesis genes have been reported as being stress responsive (Tsugane et al. 1999). We found that many photosynthesis genes were differentially regulated following NaHCO3 stress. These photosynthesis genes mainly belong to photosystem II, and demonstrated different expression patterns, suggesting that they were not co-regulated. Light-harvesting chlorophyll a/b-binding proteins are major components of antenna complexes that collect and deliver light energy to the photosynthetic reaction center in chloroplasts. The fact that several types of these binding proteins were either up- or down-regulated suggests saline-alkali stress strongly affected the ability of plants to collect and deliver light energy. Stress-enhanced proteins (Sep) are considered members of the chlorophyll a/b-binding gene family. Previous studies found that Sep genes was highly up-regulated by light-stress and UV-A illumination and down-regulated by mechanical wounding (Heddad and Adamska 2000). In this study, we found one Sep gene (EH055770) in T. hispida was down-regulated by saline-alkali stress (Table 6). Photosynthesis regulation, therefore, may be an important stress-regulation pathway for T. hispida. Additionally, Rubisco catalyzes the CO2 fixation of photosynthesis in plants. The abundance of small Rubisco subunits were significantly changed during saline-alkali stress (Table 6), implying that NaHCO3 stress strongly affected CO2 fixation in T. hispida.

Comparing gene expression under NaCl and NaHCO3 stress

In many global environments, plants face simultaneous saline and alkali stress, and the response to these two factors may be different. Research in the area of gene expression may provide important information regarding the differential responses of plants. To investigate this, nine genes responsive to NaHCO3 stress were selected for expression analysis under NaCl stress using real-time RT-PCR. Results indicate that while some gene expression patterns were similar, others were quite different (Fig. 3). Post-exposure gene expression trends were very similar at 24 h for both NaCl- and NaHCO3-exposed test plants, but largely different at 52 h, indicating that short-term stress response to either saline or saline-alkali circumstances is similar, but prolonged stress triggers alternative mechanisms.

In summary, we have constructed three cDNA libraries from T. hispida after exposure to NaHCO3 for 0, 24, and 52 h, and obtained 9,447 high quality EST sequences. Gene expression comparisons were conducted among the three cDNA libraries. The genes responsive to saline-alkali stress were identified and encompassed diverse functional areas, such as hormone signaling transduction, transcription regulation, ROS scavenging, reestablishment of homeostasis, photosynthesis regulation and transcriptional regulation of ribosomal protein, suggesting that the response to saline-alkali stress in T. hispida is a complex one, involving multiple physiological and metabolic pathways. The EST analysis can also be used to establish this Tamarix species as a model system for the molecular genetic studies of woody plant saline-alkali tolerance. These results will enrich our knowledge of the stress tolerance of woody plants on a molecular level and provide new insight into saline-alkali tolerance in plants.