GbSOBIR1 confers Verticillium wilt resistance by phosphorylating the transcriptional factor GbbHLH171 in Gossypium barbadense

Identification of the GbSOBIR1 gene and its expression profile

We previously characterized the eukaryotic P450 gene SSN (SILENCING-INDUCED STEM NECROSIS); a lesion mimic phenotype was observed in knock-down mutant ssn lines (Sun et al., 2014). A set of genes were identified that are differentially expressed in roots between the knock-down mutant (Ri15) and WT seedlings. Among them, one receptor-like kinase gene was further studied, and we identified the homolog of this gene in the cotton cultivar G. barbadense cv 7124. This is found to encode an RLK that contains five LRR domains (Figure 1a). Phylogenetic analysis of this cotton RLK and several RLKs from other plants showed that it was evolutionarily close to the Arabidopsis SOBIR1 (Figure 1b), which plays important roles in plant defence responses. Amino acid sequence alignment of GbSOBIR1 and previously reported SlSOBIR1, AtSOBIR1, NbSOBIR1 shows that each pair was highly homologous (more than 60% identical) (Figure S1). Based on the phylogenetic results, this gene was designated GbSOBIR1.

The expression of GbSOBIR1 in various tissues of the cotton plant was examined by RT-PCR. As shown in Figure 1c, the gene was expressed in all tissues investigated. We also checked V. dahliae-induced expression of GbSOBIR1 by qRT-PCR analysis, and it was found to be slightly induced during early stages of V. dahliae infection (Figure 1d). To further investigate the involvement of GbSOBIR1 in disease responses, gene expression was analysed after exogenous treatment with the defence-related phytohormones SA and JA. As shown in Figure S2, expression of GbSOBIR1 was enhanced slightly by exogenous application of SA but was down-regulated following JA treatment. These results suggest that GbSOBIR1 may be involved in defence against V. dahliae infection in the cotton plant.

Reduced GbSOBIR1 expression results in increased susceptibility of cotton to V. dahliae infection

To elucidate the role of GbSOBIR1 during cotton defence to V. dahliae, tobacco rattle virus (TRV)-based virus-induced gene silencing (VIGS) was employed to knock down the transcript of GbSOBIR1. Cotton seedlings were grown for 10 days, after which they were treated with either an empty recombinant TRV vector (TRV:00) or a TRV vector targeting GbSOBIR1 (TRV: GbSOBIR1). Two weeks after treatment, the reduced expression of the GbSOBIR1 gene was confirmed through RT-PCR analysis (Figure 2a). The plants were then inoculated with the V. dahliae strain V991. Control plants displayed typical disease symptoms such as wilted leaves and darkened vascular bundles. However, the GbSOBIR1-silenced plants were more severely affected than the control plants (Figure 2a,b); 24% and 60% of the leaves from TRV:GbSOBIR1 and TRV:00 plants, respectively, showed no necrosis or chlorosis, whereas 60% and 32% of the leaves from TRV:GbSOBIR1 and TRV:00 plants, respectively, were totally necrotic or shed (Figure 2c). The phenotypes correlated with the degree of Verticillium colonization, as determined by a fungal biomass analysis (Figure 2d). These results show that knock-down of the GbSOBIR1 gene attenuates the resistance of cotton plants to V. dahliae infection.

Overexpression of GbSOBIR1 in transgenic Arabidopsis plants enhances V. dahliae resistance

An overexpression strategy was also employed to assess the function of the GbSOBIR1 gene. Unfortunately, we were unable to obtain transgenic plants with RNAi or overexpression of GbSOBIR1 due to abnormal development or lethality during the plant tissue culture process necessary for cotton transformation. Instead, we generated Arabidopsis lines that heterologously expressed GbSOBIR1 and more than 20 independent transgenic lines were obtained. Four lines with high expression levels of GbSOBIR1 (Figure 3a) were chosen for further analysis. Wild-type and transgenic plants were inoculated with V. dahliae strain V991, and disease development was monitored up to 21 days after inoculation. A more resistant phenotype could be observed in the transgenic plants with less stunting, wilting, anthocyanin accumulation, chlorosis, early senescence and necrosis (Figure 3b,c); this was confirmed by fungal biomass analysis (Figure 3d). These results support the view that GbSOBIR1 contributes positively to plant resistance to V. dahliae.

GbSOBIR1 interacts with GbbHLH171 in vitro and in vivo

To understand the regulatory network by which GbSOBIR1 confers resistance to V. dahliae, the kinase region of GbSOBIR1 was cloned into the pGBKT7 vector to construct a bait plasmid encoding a fusion protein with the DNA-binding domain of GAL4. The fusion plasmid was used to screen a V. dahliae-inoculated cotton root library using the yeast-two-hybrid (Y2H) system, and one bHLH transcript factor was identified as an interacting protein of GbSOBIR1. The cDNA encoding this protein is 2697 bp in length with an open reading frame of 1941 bp, encoding a protein that contains 646 amino acid residues. A bHLH and R2R3-MYB transcription factor N-terminal domain and a helix-loop-helix domain can be found in the predicted protein. Based on a previous phylogenetic analysis (Yan et al., 2015), we named this protein GbbHLH171 (He et al., 2018). The ORF of GbbHLH171 was cloned and mated to GbSOBIR1 in yeast to confirm the interaction (Figure 4a). To verify the interaction in vivo, GbSOBIR1 fused to the Myc epitope tag and GbbHLH171 fused to the HA epitope tag were generated and transiently co-expressed in N. benthamiana leaves to perform co-immunoprecipitation experiments. HA-tagged GbbHLH171 was co-immunopurified with GbSOBIR1 when Myc-tagged GbSOBIR1 was used to capture the GbbHLH171 complex (Figure 4b).

How GbbHLH171, a transcription factor, interacts with membrane-located GbSOBIR1 needs to be explored. Another bHLH transcriptional factor, AaMYC2, has been found to form complexes with several proteins in the plasma membrane (Shen et al., 2016). To test whether GbbHLH171 interacts similarly with GbSOBIR1, bimolecular fluorescence complementation (BiFC) assays were carried out to assess the interaction between GbSOBIR1 and GbbHLH171 in plant cells. When GbSOBIR1-nYFP and GbbHLH171-cYFP were co-expressed in N. benthamiana, YFP fluorescence was detected at the plasma membrane, while no fluorescence was detected when GbSOBIR1-nYFP and cYFP or nYFP and GbbHLH171-cYFP were co-expressed (Figure 4c). These results indicate that GbSOBIR1 interacts with GbbHLH171 both in vitro and in vivo, and the subcellular location of this interaction is at the plasma membrane.

GbbHLH171 plays a positive role in cotton resistance to V. dahliae

To determine whether GbbHLH171 is involved in resistance against V. dahliae, we first examined its expression pattern. As shown in Figure S3A, GbbHLH171 is predominantly expressed in cotton roots and is induced within 1 h after inoculation with V. dahliae (Figure S3B). VIGS was employed to knock down GbbHLH171 expression. TRV constructs targeting GbbHLH171 were generated, and cotton plants were infiltrated with both TRV:GbbHLH171 and the negative control TRV:00. Two weeks after TRV infiltration, plants were inoculated with the V. dahliae strain V991 and subsequently monitored for the development of disease symptoms. Knock-down of GbbHLH171 resulted in more severe disease symptoms compared with the TRV:00 plants (Figure 5a,b). We also obtained GbbHLH171 overexpression cotton lines. When challenged with V991, plants overexpressing GbbHLH171 displayed a more resistant phenotype with less leaves showing chlorosis or necrosis (Figure 5c,d). These results indicate that GbbHLH171 is required for resistance to V. dahliae in cotton.

Phosphorylation of GbbHLH171 by GbSOBIR1 is required for its functional role

To test whether GbbHLH171 is a substrate for phosphorylation by GbSOBIR1, proteins of GbSOBIR1 and GbbHLH171 were obtained through high-yield wheat germ cell-free and prokaryotic protein expression systems, respectively. An in vitro phosphorylation assay was carried out and visualized by SDS-PAGE containing Phos-tag and MnCl₂. As shown in Figure 6a, GbbHLH171 was phosphorylated by GbSOBIR1. GbSERK1 protein was used as a negative control. Furthermore, the immunopurified proteins from the co-immunoprecipitation experiments were also subject to SDS-PAGE containing Phos-tag and MnCl₂ and then visualized by Western blot using anti-HA antibodies. An additional phosphorylated protein band is observed (Figure 6b). Collectively, these results demonstrate that GbbHLH171 can be phosphorylated by GbSOBIR1 both in vitro and in vivo.

The phosphorylation of AtMYC2 was previously reported to be important for its function to regulate gene transcription (Zhai et al., 2013). As another bHLH transcription factor, it was investigated whether the GbSOBIR1-mediated phosphorylation similarly affects the transcriptional activity of GbbHLH171 and its role in resistance to V. dahliae. As previously reported, the Arabidopsis JAZ2 promoter contains G-boxes that are sufficient for activation by MYC bHLH transcription factors (Figueroa and Browse, 2012). Because there are G-boxes in the promoter region of GbJAZ2 as well (Figure S4), we chose GbJAZ2 as a target to assess the roles of GbbHLH171 and GbSOBIR1 using a LUC assay in vivo. The promoter sequence of GbJAZ2 was isolated from cotton genomic DNA and inserted into the vector pGreenII 0800-LUC with a Luc reporter gene. Agrobacterium harbouring the plasmids (Figure 6c) was simultaneously injected into N. benthamiana leaves, and Luc fluorescence intensity was examined 60 h later. As shown in Figure 6c, the GbJAZ2 promoter on its own drove Luc expression weakly, while the fluorescence signal increased when it was cotransformed with GbbHLH171, implying that GbbHLH171 can enhance the transcription levels of GbJAZ2. When agrobacterium cells containing 35S:GbbHLH171, 35S:GbSOBIR1 and GbJAZ2Pro:Luc were simultaneously injected, the fluorescence intensity of Luc was even higher than that in the cells cotransformed with GbJAZ2: Luc and 35S:GbbHLH171. These results demonstrate that GbSOBIR1 can enhance the transcriptional activity of GbbHLH171 in vivo.

To test whether GbSOBIR1-mediated phosphorylation of GbbHLH171 influences resistance to V. dahliae in cotton plants, we performed VIGS experiments on GbbHLH171-overexpression lines. When the cotyledons were fully expanded, TRV:00- and TRV:GbSOBIR1-containing agrobacterium cells were injected into cotyledons of GbbHLH171-overexpressing plants followed by V. dahliae inoculation 2 weeks later. As shown in Figure 6d, the enhanced resistance caused by GbbHLH171 overexpression was attenuated by targeting GbSOBIR1. More wilt and chlorotic leaves could be observed from the TRV:GbSOBIR1 plants, and the statistic results were consistent with this (Figure 6e).

GbbHLH171 phosphorylation at Ser413 by GbSOBIR1 is essential for its transcriptional activity

To identify the GbSOBIR1-mediated phosphorylation site of GbbHLH171, the co-immunoprecipitated proteins from tobacco leaves transiently expressing GbSOBIR1 and GbbHLH171 were subjected to mass spectrometric analysis, which revealed several sites with high probability (Figure 7a,b). To determine the physiological function of these potential phosphorylation sites and whether they are GbSOBIR1-dependent, phosphorylation defective GbbHLH171 mutants carrying serine to alanine mutations (i.e. M1:bHLH171^S403A, M2:bHLH171^S410A, M3:bHLH171^S412A, M4:bHLH171^S413A) were examined in Luc reporter assays. As shown in Figure 7c, co-expressing bHLH171^S413A, GbSOBIR1 and GbJAZ2Pro:Luc resulted in lower fluorescence intensity than co-expressing the original GbbHLH171, GbSOBIR1 and GbJAZ2Pro: Luc. The remaining GbbHLH171 mutants had similar results to the wild-type protein, revealing that the S413A mutation, but not the other three mutations, affected the transcriptional activity of GbbHLH171. These results support the view that Ser413 is an in vivo phosphorylation site of GbbHLH171 by GbSOBIR1.

Plant materials, VIGS experiments and disease assays

Cotton (Gossypium barbadense) ‘7124’ and Gossypium hirsutum ‘YZ1’ plants were used in this study. For the VIGS experiments and disease assays, plants were grown in a growth room with a 16-h-day/8-h-night cycle at 25 °C.

The TRV vectors and Agrobacterium tumefaciens for VIGS were prepared according to Fradin et al. (2009). Inserts to generate TRV: GbSOBIR1, TRV:GbbHLH171 were amplified from the cDNA of G. barbadense cv 7124. Primer pairs to generate TRV vectors are shown in supplemental Table S1. PCR fragments were cloned into the TRV:00 plasmid using BamHI and KpnI. The constructs were transformed to A. tumefaciens GV3101 by electroporation. The vectors were agroinfiltrated into the cotyledons of 10-d-old cv 7124 plants using a needleless syringe as described previously (Gao et al., 2013a). Two weeks after infiltration, RNA was extracted from cotton roots to measure the expression of the target genes.

The Verticillium dahliae strain V991 was cultured on potato dextrose agar at 25 °C for 4 days and then further incubated on fresh potato dextrose agar medium for another 7 days. The conidia of V. dahliae were collected and resuspended in distilled water. The V. dahliae infection assays were performed by root dipping in spore suspension (2 × 10⁵ spores/mL) as described previously (Xu et al., 2011). Leaves of Inoculated plants were classified according to the disease symptoms, Lv0: healthy leaves, Lv1: leaves showing some necrosis or chlorosis, Lv2: leaves showing severe necrosis or chlorosis, Lv3: dead/shed leaves. The in planta V. dahliae biomass quantification in Arabidopsis and cotton was carried out according to (Fradin et al., 2011) and (Song et al., 2018), respectively.

Gene cloning, vector construction and plant transformation

The full-length sequence of GbbHLH171 was identified in the Gossypium barbadense database (Yuan et al., 2015) and was obtained through PCR using V. dahliae-infected cotton root cDNA as the template. The full-length coding sequence of GbbHLH171 was inserted into the binary plant vector pK2GW7.0 (Ghent University, http://www.plantgenetics.rug.ac.be/gateway/) to construct the vector 35S:GbbHLH171 for overexpression. The overexpression vector was introduced into G. hirsutum ‘YZ1’ plants by A. tumefaciens (strain EHA105)-mediated transformation as described previously (Jin et al., 2006). The overexpression vector was transferred into A. tumefaciens strain GV3101 to transform the Arabidopsis (Arabidopsis thaliana) ecotype Columbia-0 using the floral dip method (Clough and Bent, 1998).

Nucleic acid extraction and expression analysis

Total RNA was isolated as described previously (Tan et al., 2013). For RT-PCR and qRT-PCR analyses, RNA was reverse-transcribed to cDNA using SuperScript III reverse transcriptase (Invitrogen). qRT-PCR was performed using the ABI Prism 7000 system (Applied Biosystems, CA). The values are given relative to the housekeeping genes UB7 in cotton. The primers that were used in the RT-PCR and qRT-PCR are listed in Supplemental Table S1.

Yeast-two-hybrid assays

The V. dahliae-inoculated cotton root library for Y2H screening was constructed with the Matchmaker Gold Yeast-Two-Hybrid System (Clontech, Cat. no. 630489). The kinase domain of the cotton GbSOBIR1 gene was fused with the GAL4 DNA-binding domain in pGBKT7 to ensure that there was no autoactivation and toxicity due to the X-a-Gal assay in yeast; the GbSOBIR1 fusion protein was used as bait to identify interacting proteins. To detect protein–protein interactions between GbSOBIR1 and the identified proteins, the full-length GbbHLH171 genes were cloned into pGADT7. pGBKT7 fused with GbSOBIR1 was used to transform the Y2H yeast strain, and pGADT7 fused with GbbHLH171 was transferred into the Y187 yeast strain using the Transformation System (Clontech, Cat. no. 630489). Interactions between these proteins after mating were determined by growth on SD medium with the -Trp/-Leu/-His/-Ade/-X-α-Gal assay as described by the manual (Clontech, Cat. no. 630489).

Co-immunoprecipitation

The Agrobacterium tumefaciens GV3101 strains containing pGWB417-35s-GbSOBIR1-Myc, pGWB415-35s-GbbHLH171-HA or an empty pGWB417-35s-Myc vector were grown in fresh LB medium supplemented with appropriate antibiotics and grown to the optical density of OD600 = 1. Cultures were pelleted and resuspended in buffer containing 10 mm MES (pH 5.6) and 150 μm acetosyringone. Agrobacteria carrying different constructs were mixed 1:1 (OD600 = 0.8) and infiltrated into leaves of 4-week-old N. benthamiana plants. Samples were harvested 60 h later, and 5 g of tissue was ground in liquid nitrogen and resuspended in 10 mL volume of grinding buffer (50 mm Tris-HCl 7.5, 150 mm NaCl, 1 mm EDTA, 1% NP-40, one tablet of protease inhibitor mixture; Roche). The extracted proteins were subjected to co-immunoprecipitation using a Pierce magnetic c-Myc-tag IP/Co-IP kit (Thermo Scientific, Cat. no. 88844). The eluted proteins were subsequently analysed by Western blot using anti-Myc or anti-HA antibodies.

BiFC analysis

For bimolecular fluorescence complementation (BiFC) analysis experiments, the full-length coding region of GbSOBIR1 was cloned into pS1301nYFP, and the GbbHLH171 sequence was cloned into pS1301cYFP. The fusion constructs were transformed into A. tumefaciens strain GV3101. Then, the positive-transformed cells were cultured at 28 °C in LB medium with kanamycin (50 μm) and rifampicin (50 μm). The cell cultures were then infiltrated into tobacco leaves as described above. The interactions were detected under fluorescence light using a Leica TCS SP2 confocal spectral microsystems laser-scanning microscope (Leica, Heidelberg, Germany).

In vitro phosphorylation assays

Cell-free protein expression and in vitro kinase assays were performed as previously reported (Min et al., 2015). The cell-free-expressed kinase GbSOBIR1 and the E. coli-induced and -expressed substrate GbbHLH171-GST were mixed with kinase assay buffer (40 mm HEPES, pH 7.5, 130 mm KCl, 10 mm MgCl₂, 0.1 mm ATP, 5 mm dithiothreitol, 5 mm b-glycerophosphate and 0.2 mm sodium orthovanadate). The reactions were incubated at 37 °C for 15 min and then stopped by boiling for 5 min. The phosphorylation assay products were separated by electrophoresis using Phos-tag SDS-PAGE (8% acrylamide gels, 0.375 m Tris, 0.1% SDS, 0.1 mm MnCl₂ and 0.05 mm Phos-tag AAL-107), and the gels were stained with Coomassie Brilliant Blue, destained and visualized.

Dual-luciferase reporter assays

The transient dual-luciferase reporter assays were performed as described previously (Hellens et al., 2005). The promoter sequence of GbJAZ2 was amplified by PCR using cv7124 genomic DNA and was cloned into pGreenII 0800-LUC at the PstI and BamHI sites. The coding regions of GbbHLH171 and GbSOBIR1 were amplified by PCR and cloned into pGreenII 62-SK at the PstI and BamHI sites to generate 35::GbbHLH171 and 35::GbSOBIR1. For the mutated version of GbbHLH171, different primers were used for site-directed mutagenesis by overlap extension PCR. The pGreenII 62-SK clone harbouring the wild-type GbbHLH171 insert was used as a template for overlap extension PCR. The PCR product was then cloned into pGreenII 62-SK as well. Leaves of 4-week-old N. benthamiana plants were infiltrated with A. tumefaciens carrying different vectors. Sixty hour later, these leaves were collected and were ground in liquid nitrogen. Firefly luciferase and Renilla spp. luciferase activities were quantified using the dual-luciferase assay reagents (Promega, Madison, WI) with a multimode plate reader (PerkinElmer, Waltham, MA).

Identification of phosphorylation site

Purified protein from the Co-IP experiment was digested according to the FASP procedure described previously (Wisniewski et al., 2009). LC-MS/MS experiments were performed on a quadrupole-Orbitrap-liner ion trap hybrid mass spectrometer system (Orbitrap Fusion Lumos) that was coupled to Ultimate 3000 RSLCnano system (Thermo Fisher Scientific, Germany). The peptide mixture (2 μg) was loaded onto an Acclaim PepMap 100 analytical column (75 μm × 15 cm, C18, 3 μm, Thermo Scientific) in buffer A (0.1% formic acid) and separated with a linear gradient of buffer B (80% acetonitrile and 0.1% formic acid) at a flow rate of 300 nl/min over 60 min. Data-dependent acquisition with full scans in the 350–1500 m/z range was carried out using an Orbitrap mass analyser at a mass resolution of 60 000 at 200 m/z. Most intense precursor ions were selected using the top speed data-dependent mode with a maximum cycle time of 3 s. Peptides with charge 2–7 were selected, and dynamic exclusion time was 40 s. Precursor ions were fragmented using higher-energy collision dissociation (HCD), and MS/MS ions were detected using Orbitrap at a mass resolution of 30 000 at 200 m/z. For data analysis, MS/MS spectra were searched using MASCOT engine (Matrix Science, London, UK) embedded into Proteome Discoverer 2.1 (Thermo Electron, San Jose, CA).