A Bowman-Birk type protease inhibitor is involved in the tolerance to salt stress in wheat

Authors


G. Xia. Fax: 86-531-88565610; e-mail: xiagm@sdu.edu.cn
S. Zhao. e-mail: syzhao@sdu.edu.cn

ABSTRACT

A salt-responsive gene WRSI5 was characterized from salt-tolerant cultivar Shanrong No. 3 (SR3), an introgression line via asymmetric somatic hybrid between Triticum aestivum L. cv. Jinan177 (JN177) and Thinopyrum ponticum Podp. The peptide encoded by WRSI5 contains a Bowman-Birk domain sharing a high level of sequence identity to monocotyledonous protease inhibitors. When expressed in vitro, the WRSI5 gene product exhibited trypsin, but not chymotrypsin inhibition. The expression level of WRSI5 was increased in SR3 roots exposed to salt, drought or oxidative stress. In situ hybridization showed that it is induced in the endodermal cells of the mature region of the SR3 root tip, with no signal detectable in the corresponding region of the salt-susceptible cultivar JN177. SR3 has a higher selectivity for K+ over Na+, and therefore limits the transport of Na+ from the root to the shoot. When overexpressed in Arabidopsis thaliana, WRSI5 improves the ability of seedlings to grow on a medium containing 150 mm NaCl. We suggest that WRSI5 plays an important role in regulating the plant growth rate or long-distance Na+ transport in SR3 plants exposed to salt stress.

Abbreviations
BBI

Bowman-Birk type protease inhibitor

DDRT

differential display reverse transcription

JA

jasmonic acid

INTRODUCTION

Abiotic stress, especially high salinity, has a major influence on crop growth and development, and can be responsible for substantial losses in economic yield. Over 6% of arable land is salt-affected, and this proportion is increasing in response to land clearance and the use of irrigation (Munns 2005). Bread wheat is among the more salt tolerant of crops, but its yield is severely reduced where the level of soil salinity exceeds the equivalent of 100 mm NaCl (Munns, James & Lauchli 2006). Within the Triticeae, certain genes have been reported as being strongly induced by high salinity, and thus have been suggested as candidates for tolerance. Some of these appear to limit the uptake and transport of salt, some have an osmotic or protective function, and some have been related to salt-stress signal transduction and the regulation of gene expression (Munns 2005).

Protease inhibitors respond to a number of cellular physiological processes by regulating protease activity. The expression level and activity of protease targets have a major influence on plant growth and development (Ryan 1990; Qi, Song & Chi 2005; Fear, Komarnytsky & Raskin 2007). Currently, 59 distinct protease inhibitor families have been recognized (Christeller 2005), among which is the cysteine-rich serine protease inhibitor known as the Bowman-Birk (BBI) family (Prakash et al. 1996). Dicotyledenous BBIs are double-headed and 8 kDa in size, whereas monocotylenous ones can be divided into two classes, the single-headed ∼8 kDa type and the double-headed ∼16 kDa type (Mello, Tanaka & Silva-Filho 2003; Qi et al. 2005). BBI-type protease inhibitors are common in both developing seeds and wounded plant tissue, and their in vivo roles include protease regulation during germination, storage of sulphur amino acids during dormancy, radio-protection, programmed cell death, and defence against herbivores, pathogens and insects (Solomon et al. 1999; Chilosi et al. 2000; Dittmann et al. 2001; Xu et al. 2003; Deshimaru et al. 2004; Otlewski et al. 2005). The expression of protease inhibitors in response to drought or salinity stress has been studied by various groups. Cystatin (a cysteine protease inhibitor) in chestnut and barley, and rice OCPI1 (Oryza sativa chymotrypsin inhibitor-like 1) can both be induced by drought, salinity, exogenous abscisic acid and cold stress (Pernas, Sánchez-Monge & Salcedo 2000; Gaddour et al. 2001; Huang, Xiao & Xiong 2007). BBI, in addition to its well-studied role in defence against pathogenesis, may also be involved in the abiotic stress response. Thus, Shitan et al. (2007) showed that in yeast, BBI was responsible for tolerance to excess cadmium, while the wheat BBI-type protease inhibitors wali3, wali5 and wali6 were all induced by wounding or by the imposition of aluminium ion or toxic metal stress (Richards, Snowden & Cardner 1994; Snowden, Richards & Cardner 1995).

We have shown how asymmetric somatic hybridization can be used to introduce genes from the salt- and drought-tolerant grass Thinopyrum ponticum into bread wheat (Xia et al. 2003; Chen et al. 2004). One particular introgression line, selected on the basis of its improved salt and drought tolerance trait, has recently been registered as the cultivar Shanrong No. 3 (SR3) in Shandong Province, China (Lu-Nong-Shen-Zi No. [2004] 030). This genotype yields significantly more than both its bread wheat parent Jinan177 (JN177) and the standard salt tolerance check cultivar in saline alkaline soils in Shandong (Chen et al. 2004; Shan et al. 2006). In the present study, we show that when SR3 and JN177 gene expression profiles under salt stress were compared, one of the main differentially expressed genes was a BBI-type protease inhibitor. We also describe the cloning of this gene, and its contribution to the enhanced salt tolerance of SR3.

MATERIALS AND METHODS

Plant and growth conditions

The wheat cultivar SR3 was bred by our group, and the seeds of bread wheat parent JN177 were supplied by Crop Research Institute, Shandong Academy of Agricultural Sciences, China.

Seeds of SR3 and JN177 were germinated for 3 d at room temperature, and then grown hydroponically in half-strength Hoagland's solution at 25 °C under 900 µmol m−2 s−1 illumination, with a 16 h light/8 h dark cycle. Seedlings at the three-leaf stage were exposed to 200 mm NaCl/0.8 mm CaCl2, added in daily increments of 50 mm NaCl. Shoot and root tissues were harvested from eight plants of each line after treatment, snap frozen in liquid nitrogen, and kept at −80 °C.

Differential expression of cDNAs and cloning of differentially expressed products

Total RNA was extracted from salt-treated and control JN177 and SR3 roots, harvested 6, 12, 24, 48 and 72 h after the salt treatment was imposed, using the TRIzol (Invitrogen, Carlsbad, CA, USA) reagent, and then was equally mixed to form stressed root RNA pools. DDRT-polymerase chain reaction (PCR) were performed using the fluoroDD kit (Genomyx; Beckman Instruments, Fullerton, CA, USA), following the manufacturer's instructions. The 31 nt anchoring reverse transcription primer was AP5 and the 26 nt random PCR primer ARP10 (Table 1). DNase I treatment was applied to remove any contaminating DNA from the preparation of the total RNA. The DDRT-PCR products were separated by 6% denaturing polyacrylamide gel electrophoresis (PAGE), critical fragments were excised from the gel, and eluted by soaking in 30 µL TE (10 mm Tris-HCl, pH 8.0; 1 mm EDTANa2, pH 8.0) buffer at 37 °C for 1 h. For re-amplification, 2 µL of eluate was used as the template for a new PCR, employing as primers PR1 and PR2 (Table 1). The amplification products were ligated into the vector pGEM-T easy (Promega, Madison, WI, USA), which was used to transform competent Escherichia coli DH10B cells by heat-shock treatment (Sambrook & Russell 2001a). Insert lengths were determined by agarose gel electrophoresis of EcoRI digests of positive clones, and representative inserts were sequenced. The recombinant plasmids were designated pGEMT-WRSI5.

Table 1.  Sequences of the polymerase chain reaction (PCR) primers used in this study
NameSequence(5′-3′)Length (nt)Usage
AP5ACGACTCACTATAGGGC(dT)12CA31Anchoring primer at 3′ end for mRNA differential display
ARP10ACAATTTCACACAGCAGATCTCAGAC26PCR random primer at 5′ end for mRNA differential display
RP1GTAATACGACTCACTATAGGGC22Upstream primer for re-amplification
RP2AGCGGATAACAATTTCACACAGGA24Downstream primer for re-amplification
GSP3-10AGCACGACCCACAGAGAAGCATAGG255′RACE gene-specific primer
SMART IIAAGCAGTGGTAACAACGCAGAGTACGCGGG (T)25N-1N (N=A,C,G,orT;N-1=A,G,or C)30Oligonucleotide primer for first-strand cDNA synthesis of 5′RACE
5′CDSCTAATACGACTCACTATAGGGCAAGCAGTGGTA27Primer at 3′ end for first-strand cDNA synthesis of 5′RACE
SMART IACAACGCAGAGT45Long general primer for 5′RACE

Full-length cDNA cloning with 5′-RACE

Following the manufacturer's (Clontech, Palo Alto, CA, USA) instructions of the SMART RACE cDNA Amplification Kit, the amplification primers for first-strand cDNA synthesis were 5′ CDS and SMARTII (Table 1). The gene-specific primer GSP3-10 (Table 1) was designed on the basis of the sequence of the differentially expressed fragment. The 5′ end linker primer SMARTI was provided with the kit (Table 1). A touchdown PCR method was used, in which the reactions were held at 94 °C for 30 s and at 72 °C for 2 min over five cycles; then at 94 °C/30 s, 70 °C/15 s and 72 °C/120 s over the next five cycles; and finally at 94 °C/30 s, 65.5 °C/15 s and 72 °C/120 s for 30 cycles. The PCR products were sequenced.

Reverse transcription-PCR analysis following different stresses

Plants were grown as described earlier, but were exposed to various stress agents (200 mm NaCl, 15% PEG6000, 100 µm H202 and 200 µm AlCl3 for 0, 1, 6 and 24 h). RNA was extracted using the TRIzol reagent. The reverse transcription (RT)-PCR procedure was performed according to the manufacturer's recommendation (RT-PCR Kit; Takara, Dalian, China). The PCR involved an incubation at 65 °C for 5 min, then at 42 °C for 50 min, and at 70 °C for 10 min; followed by 30 cycles of 94 °C/30 s, 54 °C/30 s and 72 °C/60 s. The wheat actin gene served as an RT-PCR control. Amplicons were separated by agarose gel electrophoresis, and visualized by EtBr staining.

Northern blot analysis

An estimated 30 µg of SR3 total root RNA (sampled 0, 6, 12, 24, 48, 72 and 96 h after the imposition of 200 mm NaCl) was separated by 1.5% formaldehyde-denaturing agarose gel electrophoresis, and was transferred to a Nylon membrane with a negative charge. Using a random primer labelling method (Promega), discrepant cDNA fragment labelling with α-32P-dCTP was used as a probe. Membrane/probe hybridization was performed using standard procedures (Sambrook & Russell 2001b), and hybridization signals were visualized by autoradiography.

In situ hybridization

After 0, 6, 12, 24 or 48 h treatment in 200 mm NaCl, the roots, including the tip part of 0.3–0.5 cm length from root apex, the part with a length about 0.5 cm upper tip part, and the upper part of the first lateral root genesis, were fixed in 4.0% v/v polyformaldehyde/PBS (140 mm NaCl, 2.7 mm KCl, 10 mm Na2HPO4 and 1.8 mm KHPO4, pH 7.4) buffer, dehydrated through an ethanol series, and embedded in paraffin (melting point 57–59 °C) following the method of Cox & Goldberg (1988). Longitudinal and transverse sections of thickness 8 µm were cut and mounted on poly-lysine-coated slides. Digoxigenin (DIG)-labelled sense and antisense RNA probes were synthesized in vitro using, respectively, SP6 or T7 RNA polymerase (Roche Diagnostics GmbH, Mannheim, Germany). Hybridizations were performed overnight at 45 °C in 4 × SSC (20 × SSC: 3.0 m NaCl and 0.3 m sodium citrate, pH 7.0), 40% deionized formamide, 1 × Denhardt's solution, 10% w/v dextran sulphate, 0.5 gL−1 denatured and sheared salmon sperm DNA, and 0.15 gL−1 yeast tRNA. The slides were washed twice in 2 × SSC (37 °C for 5–10 min), then in 2 × SSC and 1 × SSC for 15 min each (37 °C), treated with buffer containing 20 µm L−1 RNaseA in NTE (500 mm NaCl, 10 mm Tris-HCl, pH 8.0, and 1 mm EDTANa2, pH 8.0) for 30 min at 37 °C, and rinsed twice for 10 min in the same buffer. The slides were then washed three times with 2 × SSC containing 50% v/v formamide (15 min at 52 °C), and twice with 0.1 × SSC (30 min at 37 °C). Hybridization and signal detection were achieved via an alkaline phosphatase-linked immunoassay (DIG Nucleic Acid Detection Kit; Roche) in accordance with the manufacturer's instructions.

Inhibition assay of the in vitro expressed WRSI5 protein

Recombinant bacterial cells containing pET-WRSI5 were grown in LB medium containing 50 mg L−1 kanamycin at 37 °C until the OD600 reached 0.6. IPTG was added to a final concentration of 0.5 mm, and the cells held at 37 °C for 5–6 h. The bacterial cells were harvested and resuspended in 40 mL start buffer (20 mm potassium phosphate buffer pH 7.4, 0.5 mm NaCl). Lysozyme (5 mg) was added to lyse the cells (30 min at room temperature), and the preparation was then sonicated in an ice-water slurry (20 200 W sonication bursts each of 10 s duration, with an interburst interval of 10 s), and centrifuged at 12 k r.p.m. for 15 min at 4 °C. A Hi-Trap column (Amersham Biosciences, Uppsala, Sweden) was equilibrated with 0.1 mm NiSO4 in start buffer in order to purify the fusion proteins, according to the method described by Chu et al. (2003). Bradford's (1976) method was used to determine the protein content using bovine serum albumin as a standard. The eluted fractions were separated by 15% SDS-PAGE and visualized by Coomassie Brilliant Blue R-250 staining, following the standard methods. Inhibitory activity against 2 µg trypsin or chymotrypsin (Sigma-Aldrich, St. Louis, MO, USA) was determined by incubating with purified heterologous protein at 25 °C for 10 min in a 3 mL reaction volume. Residual trypsin activity was measured with TAME (Sigma-Aldrich) at 247 nm according to Hummel's (1959) method, and residual chymotrypsin activity with ATEE (Sigma-Aldrich) at 237 nm, according to Schwert & Takenaka's (1955) method. The activity of commercial soybean BBI (Sigma-Aldrich) was assayed under the same conditions to act as a positive control.

Measurement of Na+/K+ content

Plants at the three-leaf stage were transferred to half-strength Hoagland's solution with 0, 150, 200 or 250 mm NaCl for 6 d, and then separated into roots and shoots to measure the tissue ratios of Na+ to K+. The material was dried at 105 °C for 15 min and at 65 °C for 2 d, and was ground to a powder. The powder was dissolved in HN03 at 80 °C for 15 min, and the Na+ and K+ concentrations measured by atomic emission spectrometry (Calanber type 180–80). Each experiment consisted of at least eight individual plants, and was repeated three times.

Arabidopsis thaliana transformation and the identification of transgenically expressed salt tolerance

Arabidopsis thaliana ecotype Columbia seed was germinated on filter paper at room temperature for 3 d, before growing in soil until flowering. Sense or antisense WRSI5, driven by the CaMV 35S promoter, was introduced as a transgene by the floral dip method (Clough & Bent 1998). Transgenic progeny were selected by germination on MS0 medium containing 50 µg mL−1 kanamycin. The salt tolerance of T2 transformants was assessed by their ability to germinate successfully on MS0 medium containing 150 mm NaCl. Both wild type and transgenics carrying an empty vector acted as controls.

RESULTS

The SR3 WRSI5 gene

From a template of cDNA from NaCl-stressed SR3, the AP5/ARP10 primers generated an amplicon of ∼300 bp, named w3-10. Based on the w3-10 sequence, the full-length coding region (WRSI5) was obtained by 5′ RACE. This sequence (GenBank accession AY549888) contains a coding region for a 90-residue cysteine-rich polypeptide. The deduced sequence includes a 23-residue signal peptide at the N terminus, 10 of the conserved cysteine residues diagnostic for BBIs, and a BBI domain between residues 32 and 87. An amino-acid alignment showed that the WRSI5 product is ∼87% identical to that of the wheat aluminum stress-related gene wali5 (AAA50850) and ∼86% identical to a protease inhibitor-related protein of barley (CAA88619). A CLUSTAL W (http://www.ebi.ac.uk/clustalw/) alignment with 22 related BBI type protease inhibitors from wheat, barley, maize and rice (Fig. 1a) shows that the cysteine residues are highly conserved, while the other residues are highly variable. A neighbor-joining unrooted tree phylogenetic analysis identified three major groups – the first group captures BBIs of higher molecular weight (181–254 residues), the second consists of wheat IBB1 and IBB2, and barley IBB_HORVU, while the third may be divided into two subgroups, one including SR3 WRSI5, wali3 and wali5, a barley putative protease inhibitor, and the other a rice wound-induced protease inhibitor, along with maize WIP1 and IWP1 (Fig. 1b).

Figure 1.

Sequence analysis of WRSI5. (a) Multiple alignment of WRSI5 with wheat and other species homologs. The signal peptide domains are underlined. The conserved cysteine residues are indicated by shading. (b) Phylogenetic analysis of the coding region of WRSI5. The phylogenetic tree was built from the coding sequences of WRSI5 (AAS49905) and other BBI protease inhibitors from wheat (AAA50850, AAA50848, P09863, P09864 and P81713), rice (BAA99379, BAA99380, BAB18291, BAB21173, BAB21176, BAB21177, Q0JR25, CAB88208, CAB88209, CAB88391, CAB88392 and CAB88393), barley (CAA88619 and P12940) and maize (AAL01255 and P31862). The classification of BBIs follows that of Mello et al. (2003), who defined six distinct groups of inhibitors in monocotyledonous plants. BBIs responding to stress (including WRSI5) do not belong to any of these groups.

Inhibitory activity of heterologous WRSI5

WRSI5 was overexpressed in E. coli to obtain a sufficient heterologous protein for assaying its biological activity. It exhibited in vitro a higher level of trypsin-inhibiting activity than soybean BBI (Fig. 2), but did not inhibit chymotrypsin.

Figure 2.

Trypsin (left) and chymotrypsin (right) in vitro inhibition assay for the WRSI5 fusion protein. Commercial soybean BBI (SBBI) was assayed in parallel as a positive control. Crude total proteins were extracted from the BL21 (DE3) cells containing the recombinant plasmid pET-WRSI5. The trypsin and chymotrypsin activity was determined by measuring the increase of absorbance at 247 and 237 nm with TAME (Sigma-Aldrich, St. Louis, MO, USA) and ATEE (Sigma-Aldrich).

WRSI5 expression in salt-stressed SR3

Although WRSI5 mRNA transcript was detectable by Northern blotting in the roots of non-stressed SR3 plants, its expression level increased rapidly in response to the imposition of salt stress. The maximum level of expression was maintained for 1.5 d, following 12–48 h of stress. The level was less as the duration of stress was increased, so that by 96 h after stress imposition, the WRSI5 mRNA level had returned to its pre-stress level (Fig. 3). The WRSI5 transcript level responded in a similar fashion to the stress imposed by PEG, H2O2 or AlCl3. A strong induction was observed in the roots after 1 h of H2O2 stress, 6 h of salt stress, and 24 h of PEG or AlCl3 stress. No expression was detectable in SR3 shoots after either PEG, H2O2 or AlCl3 stress (Fig. 4).

Figure 3.

Northern blot analysis of WRSI5 expression in the roots of SR3 after various times following the imposition of salt stress (200 mm NaCl). About 30 µg total RNA from the roots of the salt-tolerant wheat cultivar SR3 under 200 mm NaCl stress for 0, 6, 12, 24, 48, 72 and 96 h in one step was separated through 1.5% formaldehyde-denatured agarose gel electrophoresis, and discrepant cDNA fragment labelling with α-32P-dCTP was used as a probe. The two bands situated at the lower part of the figure are 28s rRNA and 18s rRNA.

Figure 4.

A time course of WRSI5 expression in the root and shoot of SR3 in response to various abiotic stresses. NaCl (200 mm), PEG (15% PEG6000), H202 (100 µm) and AlCl3 (200 µm). Time courses of treatments are indicated up the lanes. Plants are transferred to the 1/2 Hoagland solution with 200 mm NaCl, 15% PEG6000, 100 µm H202 and 200 µm AlCl3 for 0, 1, 6 and 24 h, with eight single plants at least. The WRSI5 transcript level was rapidly increased after all the stress treatments in roots.

Morphological changes and mRNA accumulation in salt-stressed root tips

The treatment with 200 mm NaCl stunted root growth (data not shown), and the root microstructure was also affected. After 24 h of salt stress, the epidermal cells of root tip were peeled away (Fig. 5). mRNA in situ hybridization to the roots of SR3 and JN177 showed that WRSI5 mRNA was detectable by the antisense probe other than the sense one in special endodermal cells in the mature region of SR3 root tips, but not in those of JN177 (Fig. 6).

Figure 5.

Microstructure of the SR3 root tip. Longitudinal section (a) non-stressed plant; (b) plant exposed to 200 mm NaCl for 24 h. The arrows in (b) indicate that the epidermis cells have started to desquamate peel away. V, vascular tissue (stele); C, cortex; E, epidermis.

Figure 6.

In situ localization of WRSI5 expression on longitudinal and transverse root sections of SR3 and JN177. Sections (a) to (h), (k) to (n) probed with antisense WRSI5; (i) and (j) with sense WRSI5. Positive hybridization signals are brown or purple. Sections (a), (b) and (k), non-stressed SR3; (c) to (f), (i) to (j) and (l), SR3 treated with 200 mm NaCl for 24 h; (m), non-stressed JN177; (g), (h) and (n), JN177 treated with 200 mm NaCl for 24 h. (a) to (j), root tip; (k) to (n), mature root. Arrows in (e) and (f) indicate the restained micrographs with safranin and fast green of section (c) and magnification micrograph of the boxed part of section (c), respectively. RM, root meristem; V, vascular tissue (stele); C, cortex; LR, lateral root genesis.

K+/Na+ content and distribution in SR3 and JN177 under salt stress

Under various levels of salt stress, SR3 maintained a better growth than JN177 (Fig. 7). The Na+content of both SR3 and JN177 shoots increased as the NaCl concentration was increased, but was significantly lower in those of SR3 than those in JN177 (20% lower under 200 mm NaCl). The Na+ content of SR3 roots was significantly higher than in those of JN177, under a given level of salinity stress (Fig. 8). SR3 showed an improved level of selectivity for K+ against Na+ in transporting ions from the root to the shoot, with the result that it retained more K+ in the shoot, and accumulated more K+ and less Na+ in its leaves. The leaf K+/Na+ ratio of SR3 was about 20% higher than that of JN177. Whereas the K+ content in the seedling decreased in the root and shoot of both SR3 and JN177 as the level of salinity rose, the decrease in K+ content was less in the shoot than in the root.

Figure 7.

Seedling growth of SR3 and JN177 stressed by NaCl treatment for 6 d. Plants are transferred to the 1/2 Hoagland solution with 0, 150, 200 and 250 mm NaCl, with eight single plants at least. Under different levels of salt stress, SR3 could maintain a better growing state than salt-susceptible parent JN177.

Figure 8.

Na+ and K+ content of SR3 and JN177 after 6 d of salt stress. (a), K+ content of the SR3 and JN177 shoot; (b), K+ content of the SR3 and JN177 root; (c), Na+ content of the SR3 and JN177 shoot; (d), Na+ content of the SR3 and JN177 root. Data are mean ± SE. *Indicates significant differences (P < 0.05).

Salt tolerance of transgenic A. thaliana

The WRSI5 gene was overexpressed in salt-stressed A. thaliana by driving its expression with a strong constitutive promoter. When exposed to 150 mm NaCl, both the shoots and roots of the transgenic plants fared better than those of the non-transgenic control, although a variable amount of leaf curling was observed in some of the transgenic plants (Fig. 9). An RT-PCR analysis of a sample of salt-tolerant transgenic plants indicated that the expression level of WRSI5 was correlated with plant growth under salt stress (Fig. 10). In contrast, when WRSI5 was inserted in an antisense direction, no improvement in salt tolerance over the wild type could be observed (Fig. 9).

Figure 9.

Growth of transgenic Arabidopsis thaliana on a medium containing 150 mm NaCl. (a), non-transgenic wild type; (b), plants expressing antisense WRSI5; (c), plants overexpressing sense WRSI5.

Figure 10.

RT-PCR analysis of transgenic Arabidopsis thaliana expressing WRSI5. The level of expression is variable between the lines.

DISCUSSION

Even though the amino acid composition of the BBIs has diverged during evolution, their cysteine residues have remained highly conserved. The WRSI5 gene encodes a BBI domain at positions 32–87, and carries 10 of the conserved cysteine residues. It lacks the four cysteine residues at positions 4, 5, 10 and 11, which are carried by most of the other similar length monocotyledonous BBIs. Certain wound-induced BBIs, which lack both the C4–C5 and C10–C11 disulphide bridges, sited in the inhibitory loops, nevertheless maintain their protease inhibitory activity (Rohrmeier & Lehle 1993). Similarly, the WRSI5 product has trypsin-inhibiting activity, at least in vitro, and like the majority of monocotyledonous BBIs, inhibits either trypsin or chymotrypsin, but not both. It is a more effective trypsin inhibitor than soybean BBI, in relation to the structure of 3:13 disulphide bond but not the 4:5 one. This is consistent with another report, in which the same structure could only suppress the trypsin activity (Ragg et al. 2006).

Protein degradation and the recycling pathway are also commonly induced when plants are subjected to abiotic stress (Ingram & Bartels 1996). Salt stress has been shown to result in the accumulation of protease inhibitors and the expression of other wound-related genes in tomato (Dombrowski 2003). Similarly, chestnut cystatin is not only involved in the defence response against fungal infection, but also in the response to abiotic stresses such as cold, heat and salinity (Pernas et al. 2000). The expression of barley cysteine proteases is also affected by drought, anaerobiosis, darkness and cold shock (Gaddour et al. 2001). In rice, the chymotrypsin inhibitor gene OCPI1 was strongly induced by dehydration as well as by the exogenous supply (Huang et al. 2007). Three wheat WIP1-like genes (wali3, wali5 and wali6) are induced not only by the presence of Al3+, but also by other ions, and by wounding (Richards et al. 1994; Snowden et al. 1995). WRSI5 is induced by salt stress, drought, Al3+ and H2O2 stress (Fig. 5), although with different response times and degrees. Sorting by the peak time of mRNA transcription, it should be H2O2 (1 h) > salt (6 h) > AlCl3 (24 h) > PEG (24 h) in roots. H2O2 is known to directly regulate the expression of numerous genes, including four proteinase inhibitors, and acts as a second messenger for the activation of defence genes (Orozco-Cárdenas, Narváez-Vásquez & Ryan 2001). It is well known that stresses by salt, drought and Al3+ ion are relative with oxidative stress (Mittler 2002). Therefore, these stresses firstly resulted in the second oxidative stress and then produced H2O2, which triggers the expression of some stress-related genes, including WRSI5, based on the sorting result earlier. It also showed that the crosstalk of signals produced by salt stress and drought stress or Al3+ ion stress may exist at some junction in different signal transduction pathways. Besides, other possible mechanism may be involved in the WRSI5-induced responses to salt stress. For example the transcript level of lipoxygenase, an enzyme that is involved in the conversion of linolenic acid to JA, is increased in SR3 subjected to salt stress (L. Yan, personal communication). It is generally accepted that increases in proteolytic activity are induced by wounding and by the exogenous supply of JA, and that most JA-responsive proteins accumulate when plants are subjected to salt stress (Moons et al. 1997). Thus, it appears that the accumulation of protease inhibitors in response to the imposition of salt stress may be a JA-dependent process. JA may play an important role in the crosstalk of abiotic stress and biotic stress by controlling the proteolysis.

The ability to restrict the absorption and distribution of Na+ ions in the plant is an important determinant of salt tolerance. In wheat, salt tolerance is associated with a decreased rate of Na+transport to the shoot, and with a high level of selectivity for K+ over Na+, achieved by controlling the flow of Na+ ions into the xylem (Davenport et al. 2005; Munns 2005). Certain salt-tolerant wheat lines derived from a hybrid between wheat and Lophopyrum elongatum apparently succeed in restricting the shoot concentration of Na+ by preventing its entry into the xylem from the root cortex (Gorham, Wyn & Bristol 1990; Santa-María & Epstein 2001). Endodermal cells in the root tip with a Casparian strip are responsible for this trait (Tester & Leigh 2001). In our work, SR3 grew better under salt stress than the parent wheat line JN177, and the leaves of salt-stressed SR3 accumulated less Na+ and had a higher K+/Na+ than those of JN177, whereas the roots of SR3 contained more Na+ than those of JN177. Significantly, WRSI5 expression was induced in the endodermal cells of the mature region of the SR3 root tip, whereas no expression was observed in the corresponding region of JN177. This preferential expression of WRSI5 implies that it plays a role in regulating Na+ transport. Furthermore, to a different extent, the overexpression of WRSI5 improved the salt tolerance of A. thaliana.

In conclusion, it is likely that WRSI5 effects on Na+ concentration indirectly. It increased the leaf growth rate, and so led to a lower concentration of Na+ in the leaves of SR3. On the other hand, WRSI5 may function on some proteases to regulate the sodium uptake from roots and transport to shoots directly.

ACKNOWLEDGMENTS

This work was supported by the funds of the Major Program of the Natural Science Foundation of China (No. 30530480), the National Basic Research 973 Program of China (2006CB100100), the Science Foundation of Shandong Province (No. Q2006D02) and the National Key Technology R&D Program (No. 2007BAD59B00).

Ancillary