Introduction

The Bet v 1 family, also known as the pathogenesis related 10 (PR10)-like protein family, contains proteins with low sequence similarity, but similar three-dimensional (3D) structures. The family is divided into 11 subfamilies, including the major latex protein (MLP), PR10, cytokinin-specific binding protein (CSBP) and plant polyketide cyclase-like subfamilies (Radauer et al. 2008). Bet v 1 family proteins are usually acidic, approximately 15–18 kDa, with similar theoretical isoelectric points (pI = 4.75–6.65) and structures (Osmark et al. 1998). Homologies between subfamilies of the Bet v 1 family are usually low; for example, homology between the MLP and PR10 subfamilies is less than 25%, but members of both families have similar Y-shaped cavity structure (Osmark et al. 1998; Radauer et al. 2008). The function of Bet v 1 proteins, which includes binding ligands, such as cytokinins, brassinolide and secondary metabolite, is closely related to their structure (Fujimoto et al. 1998; Gonneau et al. 2001; Marković-Housley et al. 2003; Koistinen et al. 2005; Radauer et al. 2008). In Arabidopsis, regulatory components of abscisic acid (ABA) receptor 1 (RCAR1) that share amino acid and structural similarity with Bet v 1 from brich pollen, can bind abscisic acid (Ma et al. 2009), also the same as the PYR/PYL family of START protein (Park et al. 2009). Bet v 1 family proteins may also be involved in phosphorylation or dephosphorylation reactions (Jwa et al. 2001), for example, RCAR1, act as ABA receptor, could antagonize phosphatases type 2C (PP2C) action in plants (Ma et al. 2009). Bet v 1 family proteins are also reported to have ribonuclease activity (Moiseyev et al. 1994). Bet v 1 family proteins are often involved in response to biotic and abiotic stress. PR10, MLP and CSBP subfamily proteins are induced by pathogen invasion (Somssich et al. 1986; Ruperti et al. 2002; Liu and Ekramoddoullah 2006; Pasternak et al. 2006), and by abiotic stress, such as mechanical stimulation, salt stress and drought (Hashimoto et al. 2004; Kimbrough et al. 2004). Bet v 1 family proteins also take part in plant developmental processes (Barratt and Clark 1991; Crowell et al. 1992; Nessler and Burnett 1992; Breda et al. 1996; Ruperti et al. 2002). Thus, Bet v 1 family proteins may have multiple functions in physiological processes including disease and stress resistance, and development.

Verticillium wilt is a devastating disease of cotton. A large number of disease resistance genes are induced by Verticillium dahliae invasion. We used suppression subtractive hybridization to isolate 15 differentially expressed sequences, named Vdrg1–Vdrg15, from Gossypium hirsutum cv. Zhongmian 12 induced by V. dahliae toxin. Homology analysis showed that Vdrg6 is similar to genes in the MLP subfamily of the Bet v 1 family (Wang et al. 2004). Here, we cloned the full-length Vdrg6 cDNA from cv. Zhongmian 12. We investigated expression of this gene, called Gh-MLP, after V. dahliae toxin treatment and under salt stress conditions. Transgenic Arabidopsis were used to verify the function of Gh-MLP under V. dahliae invasion and salt stress. Preliminary analysis of Gh-MLP suggested it might be involved in flavonoid content changes.

Materials and methods

Toxin purification

Verticillium dahliae strains Vd991 (original host G. hirsutum L., high-toxicity, defoliation, from the Institute of Plant Protection, Jiangsu Academy of Agricultural Sciences, China) were inoculated into Czapek’s medium containing 30.0 g sucrose, 3.0 g NaNO3, 0.5 g MgSO4·7H2O, 0.5 g KCl, 0.01 g FeSO4·7H2O, 1.0 g K2HPO4, and in 1 liter of distilled water. After 21 days at 25°C, the culture was centrifuged at 1500g for 15 min, and the supernatant leach-vacuumed with a leaching bottle containing two layers of filter paper. The filtered liquid was sterilized using a 0.22 μm Millipore filter (Millipore, Bedford, MA, USA), for non-concentrated toxin, and subsequently vacuum freeze-dried for concentrated toxin. Freeze-dried samples were dissolved in deionized water (1.5 g/5 ml) and toxin solution was filtered again with 0.22 μm Millipore filter. Toxin content was calculated by total protein in solution by the Bradford method (Bradford 1976).

Plant material and experimental treatments

Cotton seeds (G. hirsutum cv. Zhongmian 12; from the Cotton Research Institute of Chinese Academy of Agricultural Sciences, China) were dipped in 20% NaClO for 20 min in an aseptic manipulation cabinet and washed five times with sterilized water. The seed hull was removed and the seed was deposited into triangular flasks containing MS medium (Murashige and Skoog 1962) under 16 h light/8 h dark at 28°C for 7 days. Seedlings were treated with 2 ml of 75 mM NaCl, or non-concentrated Vd991 toxin for 1/6, 1/2, 1, 3/2, 2, 4, 6, 8, 12, 24, 36, 48, 60 or 72 h. Treated tissue material was quick-frozen in liquid nitrogen and stored at −80°C.

RNA, cDNA synthesis and DNA extraction

Total RNA was extracted with a plant total RNA Kit (H.Q.&.Q. Total RNA Kit II) from U-gene (Jixi, Anhui, China). The cDNA was synthesized by RevertAid™ First Strand cDNA Synthesis Kit from MBI (Fermentas, Glen Burnie, MD, USA). The DNAsecure Plant Kit (TianGen Biotech, Ltd., Beijing, China) was employed for genomic DNA extraction. The purity and concentration of RNA and DNA were determined by agarose gel electrophoresis and spectrophotometry.

Full-length cDNA cloning, intron analysis and promoter cloning

The 5′ Rapid Amplification of cDNA Ends (RACE) System, Version 2.0 (Invitrogen, Carlsbad, CA, USA) was used for full-length cDNA cloning. After toxin treating cv. Zhongmian 12 for 24 h, total root RNA was extracted and 5 μg reverse transcribed (RT) using MLP-GSP1 primers (Table 1). The RT product was treated with RNase mix, purified and linkers added. cDNA was amplified using a two-step PCR method. MLP-GSP2 and AAP were used as primers (Table 1) for the first round 5′RACE PCR reaction. MLP-GSP3 and AUAP primers were used for the second round 5′RACE nested PCR reaction. The PCR of 30 cycles at 94°C for 2 min, 94°C for 30 s, 55°C for 30 s, 72°C for 2 min and extension at 72°C for 5 min. The 250-bp PCR product was cloned into the pGEMT-Easy vector (Promega, Madison, WI, USA), and confirmed by sequencing. The full-length Gh-MLP cDNA sequence was amplified by primers MLP-QF and MLP-QR (Table 1) designed according to sequencing results. Intron was detected with cotton genomic DNA sample by the same PCR reaction. The PCR product was cloned into the pGEMT-Eeasy vector (Promega) and sequenced. Genome walking was used to clone the promoter of Gh-MLP gene (Takara Biochemicals, Tokyo, Japan). Primers for the three rounds PCR reaction were MLP-SP1, MLP-SP2 and MLP-SP3 (Table 1), respectively. Amplified sequences were retrieved for sequencing.

Table 1 Primers for cloning and detection of Gh-MLP
Full size table

Bioinformatics analysis

Open reading frames (ORFs) were identified using ORF finder (NCBI), and the protein sequence deduced from the DNA sequence. Homology comparison was carried out by BLASTN and BLASTP (http://www.ncbi.nlm.nih.gov/Blast.cgi). Preliminary properties of the encoded protein were predicted by ProtParam (Appel et al. 1994). Signal peptides and membrane-spanning structures were predicted by SignaIP (Nielsen et al. 1997) and TMHMM (Sonnhammer et al. 1998). Subcellular localization was predicted by Psort (Horton and Nakai 1997). Clustal_X 1.83 software was used for multiple sequence alignments (Thompson et al. 1997). SWISS-MODEL was used for tertiary structure prediction (Guex and Peitsch 1997).

Semi-quantitative RT-PCR

The primers MLP-F and MLP-R (Table 1) were designed according to the full-length cDNA sequence. Semi-quantitative RT–PCR was used to detect expression of Gh-MLP in cv. Zhongmian 12 after NaCl or non-concentrated Vd991 treatment, using Act-F and Act-R primers (Table 1) for the internal control β-actin. PCR reactions were 20 μl, for 30 cycles at 94°C for 2 min, 94°C for 30 s, 55°C for 30 s, and 72°C for 2 min, followed by final extension at 72°C for 5 min, and 10 μl of the PCR product was load onto 1.2% agarose gel for result checking, scanning and analysis were carried out by Gel Doc™ XR imaging system (BioRad, Hercules, CA, USA) after ethidium bromide staining.

Vector construction and Arabidopsis transformation

The full-length Gh-MLP cDNA sequence was amplified with MLP-BF and MLP-SR (Table 1). Template cDNA was retrotranscribed from 24 h toxin-treated cotton root, and the plasmid 35S::Gh-MLP was made by ligating the Gh-MLP cDNA into the binary vector pBI121 (Calbiochem) at BamHI and SacI sites. The plasmid was transformed into Agrobacterium tumefaciens strain LBA4404 and used to transform Arabidopsis thaliana (L.) ecotype Columbia (Col-0) by floral dipping (Clough and Bent 1998). Transgenic lines were screened (Murashige and Skoog 1962) on 1/2 MS medium containing 50 mg/l kanamycin. The molecular screening method can be referred to semi-quantitative RT-PCR. We screened out false positive and multicopy insertion lines with kanamycin and PCR. A T3 generation transgenic Arabidopsis line was used for this study.

Evaluation of salt tolerance

Salt stress test was carried as described (Srivastava et al. 2004). Sterilized filter paper was placed onto Ø90 mm plates, and 5 ml H2O and 75 mM NaCl were added. Sterilized Arabidopsis seeds were evenly planted on paper and the plates were put in a chamber at 22°C with 16 h light/8 h dark. Seed germinations were observed after 7 days.

Disease resistance evaluation

Sterilized Arabidopsis seeds were planted on MS medium with 10 μg/ml Vd991 toxin. Resistance was measured after 2 weeks. Arabidopsis inoculation with Vd991 and Vd114 V. dahliae strains (original host G. hirsutum L., low-toxicity, non-defoliation, from the Plant Protection Institute of Chinese Academy of Agricultural Sciences, China) was as described (Veronese et al. 2003). V. dahliae spore suspension solutions were spread onto potato dextrose agar (PDA) plates and cultured at 25°C for 4–5 days. Sterilized water was added to collect spores using two layers of sterile cheesecloth. The number of spores was calculated with a blood-counting chamber, and spores diluted to 5 × 106 cfu/ml in sterilized water. Arabidopsis seeds were vernalized and planted in steam-sterilized nutrition soil in a chamber at 22°C, with 16 h light/8 h dark, and 60-70% humidity. After 4 weeks, seedlings were gently uprooted and rinsed in sterile water, and roots were dipped for 1 min in 20 ml conidial suspensions. The spore suspension was renewed for every 32 seedlings treated. The control plants were treated with sterile water. Treated seedlings were moved into new nutrition pots containing sterilized nutrient soil and were kept at high humidity for 2 days. Phenotype observations and analyses, including plant inflorescence height, rosette diameter and leaf chlorosis levels, were taken at 4 weeks.

Determination of total flavonoid content

A NaNO2–Al(NO3)3 method was used to determine the total flavonoid content of the transgenic material (Zhang et al. 2008). A standard curve was made by drying rutin to a set weight at 105°C, and dissolving in an appropriate amount of 60% ethanol in a 50°C water bath. Standard solutions were brought to 50 ml with 60% ethanol after cooling to room temperature and 0, 1.0, 2.0, 3.0, 4.0, and 5.0 ml standard rutin solutions were aliquoted into six 10 ml tubes, and brought to 5 ml with 60% ethanol. NaNO2 solution (5%, 0.3 ml) was added, the tubes were shaken, and 0.3 ml 10% Al(NO3)3 solution was added. After shaking, the solutions were incubated for 3 min. Finally, 4 ml 1 mol/l NaOH and 0.4 ml double distilled water were added and the solution mixed and incubated for 10 min. Absorption at 510 nm was determined with 60% ethanol as blank. The standard curve used the absorption of solution as the ordinate and rutin concentration as the abscissa, and regression analysis obtained the standard curve. The total flavonoid content of the transgenic Arabidopsis plants were determined by collecting 28-day-old Arabidopsis plant leaves, drying them to a set weight, and grinding to a powder. One gram of Arabidopsis plant leaf powder was put into a Soxhlet extractor, 70% ethanol was added, and after refluxing at 80°C for 3 h, brought to 50 ml with 60% ethanol. Quantification of 2 ml of extract was as described above and the content of flavonoid calculated by regression analysis.

Results

Gh-MLP cloning and bioinformatics analysis

The full-length Gh-MLP cDNA sequence (GeneBank NO. DQ123838) comprised 757 bp, with an ORF of 474 bp that encodes 157 amino acids, a 39-bp 5′-untranslated region (UTR), and a 244-bp 3′ UTR containing a predicted AATAAA polyadenylation signal. Investigation of the genomic DNA demonstrated that the gene contains a 109-bp intron (Fig. 1). Multiple alignment analysis showed that the predicted protein sequence has homology to MLP subfamily proteins, including 66% identity with MLP28 from A. thaliana (Kimbrough et al. 2004). Its homology to other subfamily proteins from the Bet v 1 family is low, at less than 20% (Table 2). However, Gly50, Gly53, Gly56, Gly93, Gly114 and Tyr125, which are involved in maintaining structural stability, are conserved (Fig. 2). SWISS-MODEL predictions showed that the structural properties of Gh-MLP were similar to other Bet v 1 family proteins, which contain seven antiparallel β-sheets and three α-helices (Fig. 3). The basic properties of Gh-MLP, as predicted by Protparam, are 17.6 kDa molecular weight, a pI of 4.86 and a GRAVY value of −0.363, which indicated a hydrophilic protein that might be a monomer in vivo. SignaIP and TMHMM analysis found no signal peptide or transmembrane domain. Genome walking was used to cloned 1537 bp of the promoter region of the Gh-MLP gene. Sequence analysis showed that several potential cis-acting elements, such as an elicitor responsive element, a PR-10a binding factor 2 binding (PB) element, a salicylic acid-inducible element, a salt responsive element, cold responsive elements and a wound-responsive element (WUN-motif) (Fig. 1).

Fig. 1
figure 1

Nucleotide sequence of the Gh-MLP gene and flanking regions. Nucleotides are numbered on the left, with the position of the first nucleotide of the open reading frame designated as +1. The first nucleotide of the cDNA at −39 is in bold. The potential cis-element and polyadenylation signal are shaded in gray. The intron is boxed. Amino acids are numbered on the right. The nucleotide sequence data reported here will be GenBank Nucleotide Sequence Database accession number DQ123838

Full size image
Table 2 Homology analysis of cotton Gh-MLP with eleven Bet v 1 subfamily proteins
Full size table
Fig. 2
figure 2

Multiple sequence alignment between Gh-MLP and other Bet v 1 subfamily proteins. Shaded amino acids are conserved sites that maintain protein stability. The glycine-rich motif is underlined. The Uniport accession number and origin of Bet v 1 subfamily proteins are Arath_MLP-28_1, Q9SSK9, Arabidopsis thaliana; Cucsa_Csf-2, Q9SXL8, Cucumis sativus; Betve_Betv1f, P43179, Betula verrucosa; Lillo_PR-10.1, Q9ZPP9, Lilium longiflorum; Picgl_PR-10.2, Q9SNX6, Picea glauca; Orysa_PR-10a, Q9LKJ9, Oryza sativa; Luplu_CSBP, Q9FYU3, Lupinus luteus; Papso_NCS1, Q4QTJ2, Papaver somniferum; Bacce_BCE_3426, Q734H9, Bacillus cereus; Nicta_c17, Q53HY7, Nicotiana tabacum; Nitmu_Nmul_A1581, Q2Y8N9, Nitrosospira multiformis; Phypa_PR10_1, Q9AXI3, Physcomitrella patens

Full size image
Fig. 3
figure 3

Three-dimensional model of the Gh-MLP protein as constructed by SWISS-MODEL. The Y-shaped hydrophobic cavity is formed by a seven-stranded β-sheet wrapped around a long C-terminal helix, and is closed at one end by two short helices, and is highly similar to other Bet v 1 protein structures

Full size image

Induction of Gh-MLP in the cotton root by NaCl and Vd-toxin

To test for induction of Gh-MLP by salt stress or pathogens, 7-day-old cv. Zhongmian 12 seedlings were treated with NaCl or non-concentrated Vd991 toxin, and semi-quantitative RT-PCR was used to analyze Gh-MLP expression in root tissues. Gh-MLP expression was induced by salt stress, with an increase after 10 min of 75 mM NaCl treatment. Initial induction was relatively weak and upregulation was less than two-fold after 6 h, but increased more markedly between 8 and 24 h. During this period, the induction expression level was greater than twofold, and was followed by a gradual weakening expression (Fig. 4a). This result suggested that Gh-MLP plays a role in the salt stress response. In addition, Gh-MLP was rapidly induced in root tissues by Vd991 toxin, with the expression of Gh-MLP was upregulated more than twofold compared to control after 1.5 h, reaching a peak value after 12 h of treatment, with continue high expression from 12 to 72 h (Fig. 4b). Thus, Gh-MLP expression was rapidly induced and maintained at high levels by Vd991 toxin, suggesting that Gh-MLP might be involved in resistance to cotton diseases.

Fig. 4
figure 4

Semi-quantitative RT-PCR analysis of Gh-MLP expression induced by 75 mM NaCl and non-concentrated Vd991 toxin. Treatment time is 1/6, 1/2, 1, 3/2, 2, 4, 6, 8, 12, 24, 36, 48, 60 and 72 h. a Expression of Gh-MLP in cotton root tissues after 75 mM NaCl, with 0 as the control. b Gh-MLP expression after cotton root tissues were treated with by Vd991 unconcentrated toxin, 0 is the control

Full size image

Overexpression of Gh-MLP improves salt tolerance of A. thaliana

Bet v 1 family proteins improve the tolerance of plants to many stresses, including salt stress and drought (Dubos and Plomion 2001; Srivastava et al. 2004). We examined the germination of transgenic Arabidopsis seeds under salt stress to determine the role of Gh-MLP. Seeds were cultured for 7 days in H2O with NaCl at 75, 100 or 150 mM. Both wild type Col-0 and Gh-MLP transgenic plants germinated normally in H2O, but under NaCl treatment, Col-0 did not germinate. However, four transgenic Arabidopsis lines sprouted normally (Fig. 5). The 4-2-5 line also sprouted in 100 mM NaCl, although its growth was inhibited after germination, and no lines sprouted in 150 mM NaCl (data not shown). Taken together, overexpression of Gh-MLP appeared to improve the salt tolerance of Arabidopsis, so we hypothesized that the gene might be involved in cotton salt tolerance.

Fig. 5
figure 5

Phenotype determination of Gh-MLP transgenic Arabidopsis lines under salt stress. Arabidopsis seeds were germinated on plates containing 5 ml H2O or 75 mM NaCl, at 22°C, 16 h light/8 h dark, with observation of seed germination after 7 days. Col-0 was wild type Columbia, 1-1-1, 2-2-1, 3-1-4 and 4-2-5 are four homozygous transgenic Arabidopsis lines

Full size image

Gh-MLP transgenic A. thaliana cannot significantly improve its resistance to V. dahliae

We investigated the tolerance of the transgenic Arabidopsis plants to Vd991 toxin to verify that Gh-MLP was involved in V. dahliae resistance. Gh-MLP transgenic plants exhibited Verticillium wilt symptoms as control in MS medium containing toxin (Fig. 6a). The transgenic T3 Arabidopsis lines 1-1-1 and 4-2-5, which had stable phenotypes, were chosen for inoculation with the high-toxicity, defoliation Vd991 strain, and the low-toxicity, non-defoliation Vd114 strain, using the root-dipping method. Resistance to disease was measured by plant inflorescence height, rosette diameter, and extent of leaf chlorosis. Overexpression of Gh-MLP did not significantly improve Arabidopsis resistance to V. dahliae, as infection with Vd991 and Vd114 caused transgenic plants to display symptoms of Verticillium wilt (Fig. 6b). The inflorescence height of wild type Col-0 inoculated with Vd991 was reduced by 51% over mock-treated plants and rosette diameter was reduced by 50%, while the inflorescence height of the two transgenic Arabidopsis lines were also reduced by 50% and rosette diameters were reduced by 47 and 48%. Both wild type and transgenic plants reached level 4 leaf chlorosis, and results with Vd114 were similar. Compared to the control, the inflorescence height of Col-0 inoculated with Vd114 was reduced by 34%, and rosette diameter was reduced by 14%, while the inflorescence heights of the two transgenic Arabidopsis lines were reduced by 27 and 29%, and rosette diameter was reduced by 8%, with both plants at level 3 for leaf chlorosis (Table 3). Transgenic Arabidopsis plants did not show improved resistance to V. dahliae, nor did the Gh-MLP gene alleviate the symptoms of Verticillium, only partially reducing the growth inhibition of V. dahliae in transgenic Arabidopsis.

Fig. 6
figure 6

Determination of resistance of Gh-MLP transgenic Arabidopsis to V. dahliae toxin and V. dahliae. a Symptoms of transgenic Arabidopsis in MS medium containing Vd991 toxin at 10 μg/ml for 2 weeks. b Determination of resistance of transgenic Arabidopsis inoculated with Vd991 and Vd114, by 28-day-old seedling root dipping. Similar results were seen in other trangenic lines

Full size image
Table 3 Statistical analysis of symptoms from Gh-MLP transgenic Arabidopsis plants resistant to Vd991 and Vd114
Full size table

Overexpression of Gh-MLP changed the total flavonoid content in transgenic A. thaliana

To determine the effect of overexpression of Gh-MLP on the total flavonoid content, we extracted total flavonoid from leaves from nine transgenic Arabidopsis plants using an ethanol distillation method, and determined total flavonoid using NaNO2–Al(NO3)3 colorimetry. The average flavonoid content was 19.79 mg/g with a high value of 27.50 mg. In the positive control, the total flavonoid content of the empty vector pBI121 transgenic Arabidopsis line was 11.25 mg, while the negative control Col-0 content was 9.38 mg (Fig. 7). Thus, the average total flavonoid content of leaves from transgenic Arabidopsis lines was double that of the control, suggesting that overexpression of Gh-MLP was involved in changing content of total flavonoid, a secondary metabolite in plants.

Fig. 7
figure 7

Determination of total flavonoid content from Gh-MLP transgenic Arabidopsis lines. A NaNO2–Al (NO3)3 method was employed with rutin as a control. CK(−) is wild type Col-0, CK(+) is the empty vector pBI121 transgenic Arabidopsis line, 1-1-1, 2-2-1, 3-1-4, 4-2-5, 6-1-2, 8-2-1, 9-1-5, 10-3-1 and 11-3-4 are nine homozygous transgenic Arabidopsis lines

Full size image

Discussion

In this study, the full-length cDNA sequence and promoter region of Gh-MLP were cloned by RACE and genomic walking. BLAST analysis found the protein encoded by Gh-MLP had a similarity of 66% to MLP28 from A. thaliana, and 57% to Csf2 from cucumber (Suyama et al. 1999). These proteins are all belong to the MLP subfamily of the Bet v 1 family. An MLP gene has also been discovered in peach, strawberry, melon, and soybean (Radauer et al. 2008). Although the Gh-MLP protein has low similarity (less than 20%) to Bet v 1 family proteins (Table 1), sites important for maintaining protein structure stability, such as Gly50, Gly53, Gly56, Gly93, Gly114 and Tyr125, are highly conserved (Fig. 2), and the protein has the same polar amino acid and nonpolar amino acid distribution, and distribution of hydrophobic clusters as Bet v 1 family proteins, suggesting that the Gh-MLP protein has a similar 3D structure to the Bet v 1 family proteins (Fig. 3) (Osmark et al. 1998; Radauer et al. 2008). In addition, Gh-MLP, like other MLP subfamily proteins, had no glycine-rich loop (GxGGxGT) compared to Bet v 1 family proteins, but did have a GxxxxxG sequence, where the third amino acid is usually Trp (Osmark et al. 1998; Lytle et al. 2009). In Gh-MLP, the glycine-rich loop was replaces with GEWGKVG (Fig. 1). Gh-MLP has no signal peptide, so it is predicted to be an intracellular cytoplasmic protein, unlike other PR proteins, which are usually extracellular proteins (Ziadi et al. 2001).

At present, studies of MLP gene expression characteristics have mainly focused on responses to abiotic stresses and participation in developmental regulation. For example, expression of wild strawberry MLP gene was highly induced during fruit ripening (Nam et al. 1999), and Csf-2 of cucumber and Sn-1 of bell pepper were induced by injury to immature tissues (Pozueta-Romero et al. 1995; Suyama et al. 1999). Expression of MLP28 (At1g70830), MLP43 (At1g70890) and MLP (At4g23670) from A. thaliana are rapidly upregulated by gravity or mechanical stimulus (Kimbrough et al. 2004). In our study, the Gh-MLP gene from cotton also showed involvement in abiotic stresses. With 75 mM NaCl stress, the Gh-MLP was induced to low levels in a short time in root tissues, and transcripts were highly accumulated after 8 h. The salt-responsive element contained in the Gh-MLP promoter region also suggested induction by salt stress (Fig. 1), demonstrating that Gh-MLP in cotton might be involved in the salt stress response (Fig. 4a). This was verified with transgenic Arabidopsis plants, which germinated and grew normally after 75 mM NaCl treatment, which inhibited the wild type (Fig. 5). This result agreed with the findings from overexpression of pea PR10.1 in Brassica napus (Srivastava et al. 2004). Our results showed that Gh-MLP, like other Bet v 1 family proteins, appear to be involved in resistance to abiotic stresses, and improved the resistance of transgenic Arabidopsis to salt.

Reports on the expression of MLP subfamily genes in response to pathogen invasion are rare. Our study demonstrated that cotton Gh-MLP belongs to the MLP subfamily, and was rapidly induced by pathogen exposure. Gh-MLP had similar expression characteristics to Bet v 1 family genes, indicating that Gh-MLP might be involved in cotton disease defenses. In root tissues treated with Vd991 toxin, Gh-MLP expression was rapidly induced within 10 min, and the expression level of Gh-MLP increased more than twofold over the control after 1.5 h of toxin treatment, and remained at this level up to 72 h. Compared to the expression of PR10 subfamily genes, transient induction of the Gh-MLP was more rapidly, reaching a peak value 12 h after treatment (Fig. 4b). The expression of the edible alfalfa MsPR10-1 gene, which belongs to the PR10 subfamily, was not induced after invasion by Pseudomonas syringae pv. pis until after 3 h, and transcripts did not reach a peak until 24 h (Breda et al. 1996). Thus, the response of Gh-MLP to pathogen invasion suggests a more rapid response than that seen for PR10 subfamily genes. Moreover, an elicitor responsive element and PB element in promoter region, similar to the CHTC2 gene in Solanum tuberosum, and the LlPR10.1 gene in Lupinus luteus, and Mal d 1 gene in Malus pumila (Pühringer et al. 2000; Desveaux et al. 2005), possibly indicate that Gh-MLP responds to pathogen invasion (Fig. 1). However, whether Bet v 1 family proteins confer disease resistance is controversial. On the one hand, some Bet v 1 family proteins are directly involved in the defensive response to pathogens; for example, the Ocatin protein in Oxalis tuberose and CaPR-10 in hot pepper (Capsicum annuum), which have anti-microbial activity (Flores et al. 2002; Park et al. 2004). On the other hand, some Bet v 1 family proteins have no pathogen resistance activity (Pühringer et al. 2000). We found that Gh-MLP transgenic Arabidopsis lines did not show improved resistance to the V. dahliae toxin (Fig. 6a). After inoculation with V. dahliae, both transgenic Arabidopsis lines and the wild type displayed Verticillium wilt symptoms (Fig. 6b), including etiolated leaves and growth delay (dwarf plant and decreased diameter of lotus throne). Statistical analysis demonstrated that overexpression of Gh-MLP only slightly relieved the growth inhibition of Arabidopsis caused by V. dahliae (Table 3). This was similar to the effects seen for overexpression of STH-2 in potato and pea PR10.1 in Brassica napus, in which the resistance of the transgenic lines to pathogens did not obviously change (Constabel et al. 1993; Wang et al. 1999). Thus, similar to other Bet v 1 family proteins, overexpression of Gh-MLP did not significantly improve the resistance of the transgenic Arabidopsis lines to V. dahliae.

Based on these findings, Gh-MLP is more likely to be involved in the salt stress response than in defense against pathogenic invasion. However, we cannot exclude the possibility that Gh-MLP also has a disease resistance function. We hypothesize that Gh-MLP might be involved in resistance to other pathogens rather than V. dahliae, or play only a minor role in disease resistance, because Gh-MLP expression was much more strongly induced by V. dahliae toxin treatment than by salt stress (Fig. 4).

The role of Bet v 1 family protein in resistance to biotic and abiotic stresses is thought to be that, as a type of receptor, they take part in binding or transport molecules like plant hormones and flavonoid (Radauer et al. 2008). In this study, overexpression of Gh-MLP caused the total flavonoid content in transgenic Arabidopsis plants to increase by two-fold (Fig. 7). This showed that Gh-MLP activity was involved in the variation of total flavonoid content, although the mechanism is not clear. One possibility is that Gh-MLP acts as a receptor, binding plant hormone and activating the transduction signal, and then affecting flavonoid biosynthesis, leading to a change in the ability to respond to biotic and abiotic stresses. New evidence showed that RCAR1 and PYL5, which contain a Bet v 1 domain, bind ABA, and activate ABA signaling through direct inhibition of PP2C, which activates the plant stress response (Ma et al. 2009; Santiago et al. 2009). Another possibility for flavonoid content variation is that Gh-MLP might participate in binding or transport of flavonoid. Both Bet v 1 in birch pollen allergen and Fra a 1 from strawberry can bind flavonoid (Mogensen et al. 2002; Hjernø et al. 2006). Hyp-1 protein in Hypericum perforatum binds emodin and converts it to hypericin (Bais et al. 2003). Bet v 1 family proteins take part in binding or transport of ligands through the Y-type hydrophobic cavity structure (Osmark et al. 1998; Koistinen et al. 2005). Thus, we propose that the increase in total flavonoid content in transgenic Arabidopsis might be due to a large number of Gh-MLP proteins binding to flavonoid. Flavonoid can act as structural components of plants or phytoalexins, to protect against pathogen invasion (Arima et al. 2002). Similar to salt stress conditions, genes involved in flavonoid synthesis can be induced and ultimately cause a significant increase in total flavonoid content (Walia et al. 2005; Hichem et al. 2009). Therefore, overexpression of Gh-MLP causes an increase in total flavonoid content that may improve both salt stress and potential disease resistance in transgenic Arabidopsis. Taken together, these results suggested that flavonoid biosynthesis in transgenic Arabidopsis was directly or indirectly affected through Gh-MLP binding or transport of ligands, thereby changing its resistance to salt stress by varying the flavonoid levels. The lack of significant improvement in resistance to V. dahliae might be because the species of flavonoid compounds affected by Gh-MLP have no anti-fungal activity.

In conclusion, the MLP subfamily-like gene Gh-MLP cloned from upland cotton had low sequence similarity to other subfamilies of Bet v 1, but similar predicted 3D structure. Like PR10, the promoter region of this gene had potential cis-acting elements for response to biotic and abiotic stress, such as a salt responsive element, an elicitor responsive element, and a PB element. Gh-MLP gene expression was rapidly induced by NaCl and V. dahliae toxin, suggesting a role in salt stress and pathogen invasion responses. Overexpression of Gh-MLP improved the resistance of the transgenic Arabidopsis to salt, but did not significantly improve resistance to V. dahliae. Determination of total flavonoid content in transgenic Arabidopsis showed that overexpression of Gh-MLP upregulated the total flavonoid content by 100%. These results suggest that Gh-MLP plays a role in response to biotic and abiotic stresses, possible as a receptor that binds or transports ligand, and is involved in changing flavonoid change.