Molecular evidence for the involvement of a polygalacturonase-inhibiting protein, GhPGIP1, in enhanced resistance to Verticillium and Fusarium wilts in cotton

Polygalacturonase-inhibiting protein (PGIP), belonging to a group of plant defence proteins, specifically inhibits endopolygalacturonases secreted by pathogens. Herein, we showed that purified GhPGIP1 is a functional inhibitor of Verticillium dahliae and Fusarium oxysporum f. sp. vasinfectum, the two fungal pathogens causing cotton wilt. Transcription of GhPGIP1 was increased in cotton upon infection, wounding, and treatment with defence hormone and H2O2. Resistance by GhPGIP1 was examined by its virus-induced gene silencing in cotton and overexpression in Arabidopsis. GhPGIP1-silenced cotton was highly susceptible to the infections. GhPGIP1 overexpression in transgenic Arabidopsis conferred resistance to the infection, accompanied by enhanced expression of pathogenesis-related proteins (PRs), isochorismate synthase 1 (ICS1), enhanced disease susceptibility 1 (EDS1), and phytoalexin-deficient 4 (PAD4) genes. Transmission electron microscopy revealed cell wall alteration and cell disintegration in plants inoculated with polygalacturonase (PGs), implying its role in damaging the cell wall. Docking studies showed that GhPGIP1 interacted strongly with C-terminal of V. dahliae PG1 (VdPG1) beyond the active site but weakly interacted with C-terminal of F. oxysporum f. sp. vasinfectum (FovPG1). These findings will contribute towards the understanding of the roles of PGIPs and in screening potential combat proteins with novel recognition specificities against evolving pathogenic factors for countering pathogen invasion.

Scientific RepoRts | 7:39840 | DOI: 10.1038/srep39840 of 10 imperfect modules characterized by an extracytoplasmic type LRR consensus sequence (domain C), and a C-terminal domain (domain D; Fig. 1). Phylogenetic analysis of GhPGIP1 was performed by constructing a phylogenetic tree that consisted of several main branches (Fig. 2), in which cotton PGIP was observed to have a high degree (80%) of amino acid identity with PGIPs from Arabidopsis 15   GhPGIP1 expression for the first 5 d, the expression reaching the highest level on 3 d (Fig. 3A). However, inoculation with F. oxysporum f. sp. vasinfectum caused a transient increase for the first 0.5 h, which was followed by a continued increase until the seventh day (Fig. 3B). The transcription level in cotton seedlings treated with MeJA increased markedly at 3 h (Fig. 3C) whereas upon SA treatment, GhPGIP1 levels reached the highest at 6 h ( Fig. 3D) and gradually decreased thereafter. Upon H 2 O 2 treatment, the expression increased until 6 h, reached the maximum (Fig. 3E), and then decreased. On wounding, the transcription increased after 0.5 h and the increase continued until 24 h (Fig. 3F).
In vitro inhibition profile of GhPGIP1. The inhibition of two different PGs from V. dahliae (VdPG1) and F. oxysporum f. sp. vasinfectum (FovPG1) by GhPGIP1 was assessed in vitro. The purified GhPGIP1 protein was obtained by cleavage of His-tag from His-GhPGIP1 fusion protein using protease digestion with biotinylated thrombin (Fig. 4A,B, Supplementary Fig. 2). The ability to inhibit VdPG1 and FovPG1 was carried out by agarose diffusion assay 36 . The results revealed that GhPGIP1 inhibited both VdPG1 and FovPG1 (Fig. 4C,D), with an IC 50 of 44.37 and 50.13 μ g/mL, respectively (Table 1). GhPGIP1 displayed a differential activity profile against the two fungal PGs (Fig. 5A); the effects of various parameters on VdPG1 and FovPG1 inhibition are presented below. The VC (vector control) as well as the lowest GhPGIP1 concentration tested did not inhibit either of the enzymes. A gradual increase in inhibition was observed with the increasing inhibitor concentration (Fig. 5A). A significant increase in inhibition was observed at 0.81, 1.62, and 3.24 nM GhPGIP1 concentrations, with inhibition being 5.50%, 20.57%, and 30.69% for VdPG1 and 2.03%, 16.10%, and 27.39% for FovPG1, respectively. However, further increase in GhPGIP1 concentration (up to 6.48 and 12.96 nM) resulted in marginal increase in inhibition. GhPGIP1 at 3.24 nM concentration was used for further studies. pH optima of VdPG1 and FovPG1 inhibition and thermal stability of GhPGIP1. Inhibition of VdPG1 by GhPGIP1, assayed over a pH range of 3.5-5.0, revealed that the optimum pH for inhibition was between 4.2 and 4.8, with an inhibition of 26.88 and 31.45%, respectively (Fig. 5B). FovPG1 inhibition by the same concentration of GhPGIP1 over a pH range of 7.0-10.0 showed that the pH optimum for its inhibition was between 8.0 and 9.0, with 23.79% and 29.61% inhibition, respectively (Fig. 5C). GhPGIP1 (3.24 nM) was equally active from 20 °C to 60 °C (Fig. 5D). The activity dropped significantly to 21.2% at 70 °C and further to 10.5% and 2.2% at 80 °C and 90 °C, respectively. No inhibition of PGs was observed beyond this temperature.

GhPGIP1 is capable of restricting fungal infection in leaves of transgenic Arabidopsis plants by acting as a functional inhibitor.
To test whether the accumulation of GhPGIP1 in transgenic Arabidopsis plants resulted in an increased resistance to tissue degradation caused by the two pathogens, the leaves of line 2 and line 6 ( Supplementary Fig. 3) were drop-inoculated with a suspension of V. dahliae and F. oxysporum f. sp. vasinfectum conidia as described previously 37 . The transgenic plants exhibited enhanced resistance to these pathogens (Fig. 6A,B). Notably, leaf etiolation symptoms in the transgenic plants decreased at the inoculation site, particularly in line 6. Trypan blue staining was performed for the infected leaves to observe the fungal growth and disease lesions. The results revealed that V. dahliae infection increased the number of dead cells, with larger and expanding cell death areas observed beyond the inoculation site in the wild-type (WT) plants at 72 hours post-infection (hpi; Fig. 6C). However, in transgenic Arabidopsis plants, dead cells were restricted to the inoculation site that showed no noticeable spread (Fig. 6D). Similarly, the germination and growth of F. oxysporum f. sp. vasinfectum was significantly attenuated in the transgenic plants compared to that in the WT plants (Fig. 6E,F). Leaves were also treated with 3,3-diaminobenzidine (DAB), which forms a precipitate in the presence of H 2 O 2 38,39 .
After inoculation with V. dahliae, less H 2 O 2 accumulated in the transgenic leaves compared to the WT leaves (Fig. 6G,H). Transgenic lines inoculated with F. oxysporum f. sp. vasinfectum showed no H 2 O 2 accumulation at 72 hpi whereas accumulation was detected in infected WT plants and was accompanied with deregulated resistance against the pathogen and cell death (Fig. 6I,J). GhPGIP1 is required for expression of genes in the disease-resistance pathway. To understand the role of GhPGIP1 in the defence response of plants, the expression levels of Arabidopsis defence genes in WT , and wound induction, the samples after inoculation with double distilled water were used as control; however, for Vd991 and Fusarium f. sp. vasinfectum, samples treated with czapeks medium were used as control. Data were collected from three independent biological replicates. The data are means ± standard errors (n = 3). Asterisks indicate significant differences compared to the control [least significance differences (LSD), *P < 0.05, ** P < 0.01 ]. and overexpression lines inoculated with V. dahliae were detected under normal growth conditions. These genes included the pathogenesis related protein genes PR1 and PR5, ICS1, EDS1, and PAD4. In the over expressing lines, expression of some of these genes was significantly up-regulated (Fig. 7). In transgenic line 6, PR1, PR5, ICS1, and  Silencing of GhPGIP1 by Agrobacterium-mediated VIGS. We employed VIGS to examine the cellular function of GhPGIP1. After one month, the plants treated with pTRV1 and pTRV2-GhPGIP1 or GhCLA1 showed characteristic phenotype that was uniformly distributed across all true leaves (Fig. 8A,B), but the plants infiltrated with pTRV2 empty vector did not display this phenotype and grew normally.
To verify whether GhPGIP1 has been silenced completely, total cellular RNA isolated from the plants was transcribed and the expression of GhCLA1 and GhPGIP1 was detected by semi-quantitative RT-PCR. The expression of GhPGIP1 in the RNAi cotton lines (GhP1, GhP2, GhP3, and GhP4), especially in GhP1 and GhP3, was significantly lower than that in the WT plants (Ctrl; Fig. 8C). The gene UBQ7 was used as an internal control.
Enhanced susceptibility of GhPGIP1-silenced cotton plants to VdPG1 and FovPG1. To demonstrate that GhPGIP1 is essential for resistance to V. dahliae and F. oxysporum f. sp. vasinfectum in cotton, four weeks after inducing GhPGIP1-silencing, WT plants were administered VdPG1 and FovPG1, respectively. The results showed that the leaves of WT cotton plants displayed a strong resistance to VdPG1 and FovPG1 (Fig. 9A,G), whereas the GhPGIP1-silenced leaves appeared to be highly susceptible, 3 d after the administration (Fig. 9B,H). Five days after treatment with the two PGs, the GhPGIP1-silenced leaves still exhibited severe disease symptoms, including larger water soaked necrotic areas, compared to the WT leaves (Fig. 9C,I). Fungal growth was analysed by trypan blue staining in the inoculated leaves. Similar to previous reports 40 , our results revealed Inspection of the cell wall by transmission electron microscopy revealed that it was thinner in the cotton leaves treated with the two PGs than in the control leaves ( Fig. 10A,B,C). After treatment with either VdPG1 or FovPG1, the cell wall in PGIP-silenced cotton plants appeared thinner than in the WT (Fig. 10D,E,F,G,H,I).
Homology modeling. Based on the crystal structure of PvPGIP2 (1OGQ), which shares high sequence similarity with GhPGIP1, crystal structure of the latter was built. VdPG1 and FovPG1 were modelled based on the endopolygalacturonase from Aspergillus tubingensis (4C2L) and Aspergillus aculeatus (1IA5), respectively. In each case, the chosen models supported the criteria mentioned previously 13 . The final models having ideal geometry were found by MolProbity. These models were subsequently used in the docking experiments. Docking studies of GhPGIP1-VdPG1 and GhPGIP1-FovPG1 complexes. To predict the conformation of GhPGIP1 and its putative interactions with VdPG1 and FovPG1, two different protein-protein docking programmes, PatchDock (Schneidman-Duhovny) and FireDock were used. Firstly, top 20 solutions of the complexes were generated by PatchDock, according to the scoring function used. Subsequently, FireDock was used for the fast interaction refinement of the top 10 complexes that were rescored according to the criteria described previously 41 . Then the prediction candidate models were re-ranked due to re-refinement and optimization (Supplementary Table 1). The best-docked conformations obtained in the exact expected orientations were selected and are depicted in Fig. 11A,B.
The docked protein complexes suggested that both VdPG1 (Fig. 11A) and FovPG1 (Fig. 11B) interacted with GhPGIP1 at the B1-sheet. However, the C-terminus of VdPG1 bound the B1 sheet with stronger affinity, as most parts of the C-terminus were embedded in the B1-sheet of GhPGIP1, whereas the C-terminus of FovPG1 was only partially embedded. However, the substrate-binding sites exposed in the two enzymes appeared less different (circled in black). From the docking score, the binding of GhPGIP1 with VdPG1 was predicted to be Figure 7. Expression of disease resistance-related genes and plant defensin genes. Total RNAs were extracted from the soil grown WT and transgenic Arabidopsis that inoculation with V. dahliae 5-weeks later. The relative transcription levels were detected with Arabidopsis EF-1-α as an endogenous reference for normalization. The expression level of WT was set to 1. The experiment occurred with three independent biological replicates, and error bars indicate standard error (n = 3). Significance of difference compared with WT was analysed statistically (Student's t-test, *P < 0.05, **P < 0.01).
Scientific RepoRts | 7:39840 | DOI: 10.1038/srep39840 stronger than with FovPG1. The docked model, therefore, was in accordance with the experimental observation that GhPGIP1 had the ability to inhibit the enzymatic activity of VdPG1 and FovPG1.

Discussion
In the present study, we characterized GhPGIP1 from cotton, an LRR protein classified into different families in plants. The conserved LRR domains of GhPGIP1 exhibited high similarity with the LRR regions of AtPGIP1 and AtPGIP2 (Fig. 1), with 80% similarity among them (Fig. 2). The LRR domains are conserved in all PGIPs, indicating that protein-protein interaction is involved in systemic resistance and in the recognition of non-self-molecules in plants 11 . GhPGIP1, like PGIPs of other plants, belongs to a small cluster of proteins that are individually composed of several conserved LRR domains and exhibited distinct and diverse functions 11 . These results are also supported by the notion that PGIP contains at least 10-tandemly-repeated extracellular LRR motifs 9 .
Previous reports have shown that some PGIPs are induced by biotic stress in soybean 42 , Arabidopsis 15 , and strawberry 43 , and might be essential for general resistance to abiotic stress 40 . In the present study, we examined the expression patterns of GhPGIP1 in cotton after treatment with various abiotic and biotic stresses. Our results revealed that GhPGIP1 was significantly induced by V. dahliae and F. oxysporum f. sp. vasinfectum infection on 3 and 7 d, respectively (Fig. 3A,B), providing protection against the diseases. The observed expression of GhPGIP1 at different times, post MeJA treatment (Fig. 3C), was different from that reported for CaPGIP in C. annuum 18 .
In Arabidopsis 44 and bean 20 , PGIPs accumulated in response to SA treatment, providing disease resistance. In our study, SA treatment induced the up-regulation of GhPGIP1, and its expression peaked at 6 h (Fig. 3D), similar to that of CaPGIP in C. annuum 18 . Hydrogen peroxide plays a central role in launching the defence response during stress in plants 45 ; our study showed increased transcription of GhPGIP1 with a peak at 6 h (Fig. 3E). Additionally, wounding can induce defence proteins to provide protection from stresses 15 . In our study, wounding induced up-regulation of GhPGIP1 (Fig. 3F). These findings suggest that plants induce sophisticated defence mechanisms to regulate their responses when interacting with the surrounding environment.
The optimum pH for the inhibition of Aspergillus niger PG by PGIP-1 and PGIP-2 from guava fruit was determined to be 4.2, whereas that for PGIP-3 was 4.4, when assayed over a pH range of 4.4-4.5 46 . However, PGIP from chilli retained > 50% activity at pH 3.0 and 8.0 47 . Previous studies on thermal tolerance of PGIPs have shown them to be thermostable. PGIPs isolated from orange retained considerable activity at 60 °C 48 , and that from tomato retained partial inhibitory activity at 100 °C 49 . In this study, GhPGIP1 inhibited VdPG1 effectively with an inhibition rate of 26.88-31.45% over a pH range of 4.2-4.8, and inhibited FovPG1 over a pH range of 8.0-9.0, with an inhibition rate of 23.79-29.61% (Fig. 5B,C). GhPGIP1 remained active from 20 °C to 60 °C; the activity dropped from 70 °C to 90 °C (Fig. 5D), which was consistent with the results of a previous study 50 . This diversity in PGIP activity at different pH and temperature could be crucial in countering the multitude of PGs encountered from various pathogens. Such differential PGIP activities have been proposed to generate a steady-state concentration of biologically active oligogalacturonides that participate in producing a better host defence response 51 .
To determine the effect of gain-of-function of GhPGIP1 in plants, we generated GhPGIP1 transgenic Arabidopsis lines and examined their response to V. dahliae and F. oxysporum f. sp. vasinfectum; these pathogens can cause yellowing, wilting, defoliation, and eventual death in cotton plants 52 . Arabidopsis was chosen for the ease of its transformation. The qRT-PCR analysis revealed high expression levels of GhPGIP1 in the different transgenic lines (Supplementary Fig. 3). Moreover, the results revealed that the accumulation of GhPGIP1 in transgenic Arabidopsis plants could endow them with enhanced inhibition capabilities against PGs, reducing the disease symptoms in transgenic plants challenged with the two fungal pathogens (Fig. 6). This provides indirect evidence that PGs are important pathogenic factors in the infection of Arabidopsis by the two fungal pathogens. Previous studies have reported the effectiveness of PGIP in reducing the disease symptoms in other transgenic plants infected with B. cinerea and Cercospora nicotianae 18,41 . Here, we provide evidence that GhPGIP1 gene can  of V. dahliae has been demonstrated to possess dual functions: that of a toxin and an elicitor, at high and low concentrations, respectively 53,54 . Numerous studies have shown that SA plays a vital role in defence signalling pathways and its levels increase with the activation of PR gene 55 . Plants are believed to reserve SA as glucoside SA, which releases free SA under the action of SA GTase, thereby, inducing the PR gene expression. ICS1 is an essential enzyme in pathogen-induced biosynthesis of SA and is necessary for understanding the SA-mediated defence responses 56,57 . EDS1-regulated, SA-antagonized, and SA-promoted processes are indispensable for resistance to pathogens. EDS1 regulates SA accumulation by elevating its own expression and that of other genes 39 . For instance, EDS1 is indispensable for the accumulation of pathogen-induced PAD4. In the EDS1/PAD4-dependent R gene-mediated response, both these genes positively regulate the accumulation of SA. Furthermore, in the induction of plant hypersensitive response (HR), EDS1 is especially necessary 58,59 . Our experiments demonstrated that all the SA-responsive genes were up-regulated during V. dahliae infection (Fig. 7), suggesting that GhPGIP1 contributes to the SA-mediated defence responses and plays a positive role in the regulation of gene expression.
We used VIGS approach to investigate the effect of loss of function of GhPGIP1 in cotton infected with highly virulent strains of V. dahliae and F. oxysporum f. sp. vasinfectum. The control plants showed relatively high GhPGIP1 expression, whereas various treatments caused a decrease in its transcript levels in PGIP-silenced cotton (Fig. 8). Moreover, PGIP-silenced cotton exhibited enhanced susceptibility to the purified PGs (Fig. 9B,C,H,I) and the control cotton showed enhanced resistance to the two PGs (Fig. 9A,G). Therefore, we concluded that basal disease resistance of cotton was possibly activated on exposure with the two pathogens, and GhPGIP1 might be participating in inhibition at different stages of the infection. To further investigate the cell wall alterations of the vascular elements, the two PGs purified from the two vascular fungi were inoculated in the control and PGIP-silenced cotton. The results showed thinner cell walls after inoculation with both the PGs, compared to the control; however, the PGIP-silenced cotton revealed thinner cell walls than the control cotton (Fig. 10). The marked decrease in thickness of the cell walls in PGIP-silenced cotton suggested that PGs mainly degrade cell walls, and can be effectively inhibited by PGIP; this adversely affects the growth and development of the pathogen protects the cell wall integrity. These observations provide indirect evidence for the role of GhPGIP1 in protecting the plant cell walls against fungal pathogens.
The PG-PGIP interaction, considered a model protein-protein recognition in plant-pathogen interactions 60 , is paradigmatic for studying the key recognition events underlying plant immunity 60,61 . The concave face of PGIPs participates in establishing the specificity for binding to the invading PG, whereas the convex face imparts the required flexibility 13 . To further examine the interaction between GhPGIP1 and the two PGs, docking studies were performed. The results indicated remarkable differences between the docking complexes, reflected in the interaction modes of enzyme-inhibitor complexes in Fig. 11. GhPGIP1 was observed to contact VdPG1 only at the C-terminal of B1-sheet, in the VdPG1-GhPGIP1 complex; the other parts of VdPG1 were still accessible to the substrate, which had the potential to form ternary complexes with GhPGIP1 (Fig. 11A). In the FovPG1-GhPGIP1 complex, most part of the active site was almost covered by the inhibitor (Fig. 11B). These studies demonstrated the structural flexibility and specificity of GhPGIP1 binding interactions with the two PGs.
In conclusion, we have demonstrated that transgenic Arabidopsis plants overexpressing GhPGIP1 were more resistant to V. dahliae and F. oxysporum f. sp. vasinfectum, and conversely, cotton with partially silenced GhPGIP1 expression was more sensitive to VdPG1 and FovPG1 infection. Moreover, we investigated the interaction between GhPGIP1 and the two PGs, which enhanced our understanding of this interaction at the structure level. However, structural analyses would be necessary for drawing structure-function correlation. Furthermore, the multiplicity of GhPGIP1 functions would be crucial for screening host proteins with improved antifungal potential against the evolving PGs.

Methods
Plant growth and preparation of fungal polygalacturonases. The state cotton 2006001 (original strain no. GK44), was provided by the Cotton Research Institute, Chinese Academy of Agricultural Sciences and its seedlings were used for inoculation with the two fungi. RNA was extracted from the inoculated seedlings for gene expression analysis and for the preparation of cDNA library. The virulent strains of Verticillium dahliae Vd991 and Fusarium oxysporum f. sp. vasinfectum were cultured on potato dextrose agar on cellophane sheets for 7 d. New hyphae were collected and used for the extraction of RNA with Trizol; the extracted RNA was subsequently reverse transcribed.
RNA extraction and cloning of GhPGIP1. Total RNA was extracted from the cotton plants according to the manufacturer's instructions (Promega, WI, USA). A PolyATract mRNA Isolation System was used to obtain polyadenylated mRNA following the specifications of the supplier (Promega). Thereafter, a cDNA library was constructed as described previously 41,62 , based on a 21-amino acid sequence (5′ -FDXSYFHNKCLCGAPLPSCK-3′ ) conserved at the C-terminus of all the previously characterized PGIPs 9,15 .

Sequence manipulation and phylogenetic analyses.
A homology search against the NCBI database was conducted to verify whether the obtained sequences encoded GhPGIP1. Multiple protein sequence alignment was performed using Clustal Omega (http://www.ebi.ac.uk/Tools/msa/clustalo/) and the software SMART (http://smart.embl-heidelberg.de/) was used for the prediction of domain structure. A phylogenetic tree based on the protein sequences was constructed using MEGA 5.1.

RNA isolation and qRT-PCR analysis of plants treated with various stresses.
The cotton seedlings were grown at day/night temperatures of 25 °C/22 °C in a growth chamber under a light intensity of 150 μ mol·m −2 ·s −1 provided during a 16/8-h photoperiod. Total RNA was extracted (Biomed) from the complete stool of cotton plants that were exposed to one of the following treatments for induction of stress response: infliction with wounds, administration of 5 mM salicylic acid 18 65 . The total RNA was reverse transcribed (TIANGEN BIOTECH CO., LTD, Beijing, China). The qRT-PCR analysis for GhPGIP1 was performed using specific primers, qGhPGIP-F: 5′ -TCTGGTACAATCCCTGCCTC-3′ and qGhPGIP-R: 5′ -CAGATCCAGCCTTGCCAAAC-3′ ; the endogenous control used was UBQ7 gene (DQ116441) from cotton, which was detected using the sense primer UBQ-F: 5′ -GAAGGCATTCCACCTGACCAAC-3′ and antisense primer UBQ-R: 5′ -CTTGACCTTCTTCTTCTTGTGCTTG-3′ . The transcription levels of the target gene (GhPGIP1) relative to the reference gene (UBQ7) were analysed by the comparative CT (2 −ΔΔCT ) method, where Δ Δ CT = (C T target-C T reference) Sample X − (C T target-C T reference) Sample 1 . Sample 1 was the calibrator sample without any treatment and Sample X represents the samples from the plants exposed to the different stress treatments. All the experiments were repeated three times from three independent plants. The data were analysed statistically using SPSS (IBM Corp., Armonk, NY). The data analyses were performed by one-way analysis of variance (ANOVA). The asterisk indicates a significant difference at P < 0.5, based on the least significant difference (LSD) method.

Expression and purification of recombinant GhPGIP1 protein. GhPGIP1 was amplified
using the primers 5′ -CTACATATGgACCACTgCAACgCTCAAgACAAg -3′ and 5′ -TAAGGATCCT TATTACTTgCAgACgTCgAgCggAg-3′ . The amplified fragment was subsequently digested with NdeI and BamHI, and cloned into PET-32a vector with a six-His tag. The construct was transformed into Escherichia coli BL21DE3 following the protocol of the manufacturer for competent BL21 DE3 cells (CW BIO, Beijing, China). The single transformed colonies obtained were cultured at 37 °C until the optical density reached 0.6 to 0.8, and the cells were then induced with 1.0 mM IPTG. The culture was further incubated for 6 h at 28 °C and was harvested, thereafter, by centrifugation at 10,000 × g for 20 min at 28 °C. The expression of soluble GhPGIP1 was verified by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and the recombinant protein was purified using a 6× His-Tagged Protein Purification Kit (CW BIO) 41 . The 6× His fusion tags, trx• tag and S• tag, were removed using a Thrombin Cleavage Capture Kit (Novagen, Madison, WI).
Inhibitory activity of GhPGIP1 and agarose diffusion experiment. The GhPGIP1 activity was determined by reducing the end groups with a modified DNS reagent 66 . The plate assay was performed using a reaction mixture containing 0.8% agarose and 0.5% polygalacturonic acid 18 . VdPG1 and FovPG1 were added with or without GhPGIP1 on the agarose plates, the mixture was incubated at 30 °C for 12 h, and then 0.05% ruthenium red was used to stain the agarose plates, followed by thorough washing with sterile water 66 . The PG activity was expressed as agarose diffusion units, with a ring of 0.5 cm radius defined as one agarose diffusion unit. One unit activity of PGIP was defined as the amount of PGIP required to reduce the PG activity by 50% 18 . The GhPGIP1 mixture containing 0.0011 reducing units was used to assess the inhibitory activity of GhPGIP1.
PGIP activity measurements. The purified VdPG1 (15 ng) and FovPG1 (25 ng) were incubated separately with 0.1 mg/mL of polygalacturonic acid substrate (Sigma) at 30 °C in 50 mM sodium acetate buffer (pH 4.8 and pH 8.9, respectively), in reaction volumes of 200 μ L 50 . The PG activity was determined by analysis of the reducing end-group according to Anthon and Barrett 67 . The activity of the PGs, pre-incubated with GhPGIP1, for 20 min at 30 °C, was used to assay the inhibitory activity of PGIP. The inhibitory activity was expressed as percentage reduction in the number of reducing ends (in μ kat/mg of protein) liberated by PGs in the presence of PGIP 50 . The PGs pre-incubated with VC served as the control.
In separate experiments, the temperature and pH stability of GhPGIP1 were studied by pre-incubating the GhPGIP1 solutions separately for 1 h at temperatures ranging from 20 to 100 °C, and for 16 h at pH ranging from 2.0 to 11.0 at 4 °C, respectively. The pre-incubated GhPGIP1 solutions were reconstituted in the appropriate assay buffer and their inhibition potential was assayed at 30 °C 50 . All the experiments were performed twice, each time in triplicates. The data of a representative experiment was subjected to Tukey's Honestly Significant Difference (HSD) test followed by analysis of variance at P < 0.05.

Disease assays and lesion-size determination in transgenic Arabidopsis plants. GhPGIP1was
amplified using the forward (5′ -gCCAAgCTTATgAAgATATATCCAgCTTTCCT-3′ ) and reverse (5′ -ATATAAACTAgTCTTgCAgACgTCgAgCggAg-3′ ) primers, which incorporated HindIII and SpeI cleavage sites, respectively. The amplified gene was subcloned into Super-pCAMBIA1300 ( Supplementary Fig. 4). V. dahliae (Vd991) and F. oxysporum f. sp. vasinfectum were grown on potato dextrose agar at room temperature (23 °C) for 3-4 d. The spore suspensions were prepared as described elsewhere 21 . The wild type and transgenic Arabidopsis plants were grown for 5 weeks in growth chambers at 24 °C under conditions of 70% relative humidity and 16 h light/8 h dark cycle. The plants were then challenged with a conidial suspension of V. dahliae and F. oxysporum f. sp. vasinfectum by syringe inoculation as reported earlier 41 . The disease lesion areas were measured after 3-5 d. All the disease assays were repeated at least four times.
Histochemical staining and cell death quantification. The leaves from transgenic (line 2 and line 6) and wild-type Arabidopsis plants inoculated with V. dahliae and F. oxysporum f. sp. vasinfectum were collected for histochemical staining. Lactophenol trypan blue was used to detect the dead or dying cells as described elsewhere 39,68 . To identify the dead cells, the inoculation spots (1 mm 2 ) were marked on the leaves, as reported previously 39 . The accumulation of H 2 O 2 was detected following the protocol of Torres 69 . Briefly, the leaves were infiltrated with 3-3′ -diaminobenzidine (Sigma) solution and placed in a lucifugal beaker for 8-10 h, and then dipped in an ethanol/lactic acid/glycerol solution (3:1:1) for 6-8 h, followed by decolouration with chloral hydrate 39 (Sigma). The stained leaves were then observed by an imaging system and the photographs were taken with a Nikon Eclipse Ti microscope (Japan) using a 4 × 0.25 numerical aperture objective. RNA extraction, reverse transcription, and qRT-PCR. RNA was isolated from the complete stools of wild-type and transgenic Arabidopsis plants inoculated with V. dahliae, using Trizol reagent (TIANGEN BIOTECH CO., LTD, Beijing, China) as per the manufacturer's instructions. Total RNA was reverse transcribed by FastQuant RT Kit (with gDNase; TIANGEN BIOTECH CO., LTD, Beijing, China). qPCR was performed with a SYBR ® Premix Ex Taq (Takara, Japan) according to the manufacturer's protocol using gene specific primers; Arabidopsis EF-1-α was used as an endogenous reference. The qPCR assays were performed for at least three biological replicates and a minimum of three technical replicates for each sample. The primers used are listed in Supplementary Table 2.
Scientific RepoRts | 7:39840 | DOI: 10.1038/srep39840 To facilitate the infiltration, small holes were punched on the underside of the cotyledon using a needle 21 . The assays were performed with at least six plants for each construct and the experiments were repeated at least thrice.
The expression of GhCLA1 and GhPGIP1 in the VIGS silenced-lines was detected in the detached cotton leaves using semi-quantitative RT-PCR. The leaves were collected five weeks after Agrobacterium infiltration and subjected to RNA extraction using an RNA Extraction Kit (Biomed, Beijing, China), and the total RNA was reverse transcribed (TIANGEN BIOTECH CO., LTD, Beijing, China). The GhUBQ7 gene was used as the internal control. The primers used in the RT-PCR analysis are listed in Supplementary Table 2.

VdPG1 and FovPG1 infiltration of cotton plants.
Infiltrations were performed with a needle-less 1-mL or 2-mL syringe placed against the upper side of the leaves 72 ; the lower side of the leaves were inoculated with the respective PG dissolutions and treated with a bacterial lysis buffer, followed by binding, and were subsequently treated with an elution buffer 18 . The solutions of VdPG1 and FovPG1 in sodium acetate, pH 4.8 and pH 8.9, respectively, were applied in varying concentrations. Before infiltration, the enzyme activities were determined using 0.25% (w/v) polygalacturonic acid, as described previously 72 . In each experiment, the PG and control solutions were inoculated in two sections per leaf, two leaves per plant, and more than three plants for each treatment. The experiments were performed at least thrice with different batches of the enzymes. Trypan blue staining was performed as mentioned above.
Homology modelling of GhPGIP1, VdPG1, and FovPG1. The X-ray crystallographic structures of PGIP2 (PDB ID: 1OGQ), enzymes from Aspergillus tubingensis (PDB ID: 4C2L), and Aspergillus aculeatus (PDB ID: 1IA5) were used as models for GhPGIP1 ( Supplementary Fig. 5A), and polygalacturonases from V. dahliae (Supplementary Fig. 5B) and F. oxysporum f. sp. vasinfectum ( Supplementary Fig. 5C), respectively. The structures were modelled using SWISS-MODEL 41 . The post-refinement of the structure models was carried out using the KoBaMIN server (Rodrigues), and further geometric accuracy of the models was evaluated using MolProbity 4.02b 73 . The final structures were visualised in Pymol.
Docking studies. We used PatchDock (Schneidman-Duhovny) beta 1.3 version molecular docking algorithm based on shape complementarity principles for docking GhPGIP1 to the two PGs, individually. For the protein-protein interactions, the receptor molecule was GhPGIP1 and the ligand molecules were VdPG1 and FovPG1. The output of the analysis contained 20 models with the highest-scoring ones being the most probable 13 . To select the most suitable model from among the top-20 docked complexes, the output was resubmitted to FiberDock for further optimization and re-refinement.