lncRNA7 and lncRNA2 modulate cell wall defense genes to regulate cotton resistance to Verticillium wilt

Abstract

In plants, long noncoding RNAs (lncRNAs) regulate disease resistance against fungi and other pathogens. However, the specific mechanism behind this regulation remains unclear. In this study, we identified disease resistance-related lncRNAs as well as their regulating genes and assessed their functions by infection of cotton (Gossypium) chromosome segment substitution lines with Verticillium dahliae. Our results demonstrated that lncRNA7 and its regulating gene Pectin methylesterase inhibitor 13 (GbPMEI13) positively regulated disease resistance via the silencing approach, while ectopic overexpression of GbPMEI13 in Arabidopsis (Arabidopsis thaliana) promoted growth and enhanced resistance to V. dahliae. In contrast, lncRNA2 and its regulating gene Polygalacturonase 12 (GbPG12) negatively regulated resistance to V. dahliae. We further found that fungal disease-related agents, including the pectin-derived oligogalacturonide (OG), could downregulate the expression of lncRNA2 and GbPG12, leading to pectin accumulation. Conversely, OG upregulated the expression of lncRNA7, which encodes a plant peptide phytosulfokine (PSK-α), which was confirmed by lncRNA7 overexpression and Ultra Performance Liquid Chromatography Tandem Mass Spectrometry (UPLC-MS) experiments. We showed that PSK-α promoted 3-Indoleacetic acid (IAA) accumulation and activated GbPMEI13 expression through Auxin Response Factor 5. Since it is an inhibitor of pectin methylesterase (PME), GbPMEI13 promotes pectin methylation and therefore increases the resistance to V. dahliae. Consistently, we also demonstrated that GbPMEI13 inhibits the mycelial growth and spore germination of V. dahliae in vitro. In this study, we demonstrated that lncRNA7, lncRNA2, and their regulating genes modulate cell wall defense against V. dahliae via auxin-mediated signaling, providing a strategy for cotton breeding.

Cotton long noncoding RNAs and their regulating genes modulate cell wall defense against Verticillium dahliae via auxin-mediated signaling.

Introduction

Cotton (Gossypium) is the world’s most important natural textile and an important oilseed crop, responsible for an estimated $500 billion of annual economic output worldwide (Chen et al., 2007). Verticillium wilt is a serious soil-borne vascular disease caused by Verticillium dahliae and afflicts >400 plant species (Zhang et al., 2016a, 2016c; Li et al., 2017). Verticillium wilt has become a primary factor in the decline of cotton production. The associated reduction in fiber quality and yield losses can reach 30% (Shi, 1995; Cai et al., 2009; Xu et al., 2011; Sun et al., 2013; Yu, 2013; Li et al., 2014). Therefore, a method of effectively controlling Verticillium wilt is urgently needed, and the only economically viable control is through host plant resistance. However, the mechanism of resistance to Verticillium wilt in cotton remains unclear.

Several strategies, including distant relative hybridization breeding with chromosome segment substitution lines (CSSLs) and molecular genetics, have been used to develop wilt-resistant cotton varieties (Li et al., 2019). Previous research assessing the functional genomics and mechanism of cotton resistance to Verticillium wilt has primarily focused on protein-coding genes (Li et al., 2017; Hu et al., 2018a, 2018b; Zhou et al., 2019; Xiao et al., 2021), although some studies have assessed the resistance-related roles of small RNAs (Zhu et al., 2013; He et al., 2014; Zhang et al., 2015, 2016b) and circRNAs (Xiang et al., 2018). In addition to these RNA types, some long noncoding RNAs (lncRNAs), which are defined as transcripts greater than 200 nt in length and lack long or conserved open-reading frames (ORFs), could also play a disease-resistance role in plants (Wang et al., 2014, 2021; Zhang et al., 2014). lncRNAs have been associated with disease resistance mechanisms in several plant species. For example, tomato (Lycopersicon esculentum) lncRNA16397 enhanced resistance to Phytophthora infestans by inducing glutaredoxin gene expression (Cui et al., 2017), and there is a Brassica napus lncRNA (TCONS-00000966) that has 90% overlap with a plant defensin gene, suggesting its possible role in responding to biotic stresses (Joshi et al., 2016).

Moreover, the co-induction of several lncRNAs with neighboring genes and putative defense response roles has been reported: a co-expression analysis of 559 lincRNAs responsive to Pectobacterium carotovorum infection in potato identified 17 lincRNAs highly associated with defense-related genes (Kwenda et al., 2016), while 20 Fusarium oxysporum responsive lncRNAs were identified in Arabidopsis (Arabidopsis thaliana), five of which were related to disease development (Zhu et al., 2014). The study found that the virus-induced gene silencing (VIGS)-based knockdown of two lncRNAs in cotton seedlings, GhlncNAT-ANX2 and GhlncNAT-RLP7, enhanced resistance toward V. dahliae and Botrytis cinerea, possibly by increasing the expression of the sense cognate disease-resistance genes lipoxygenase 1 (LOX1) and LOX2 (Zhang et al., 2018a, 2018b). A recent study found that GhlncLOX3-silenced seedlings displayed a reduced resistance to V. dahliae, with downregulated expression of GhLOX3 and decreased content of JA (Wang et al., 2021). However, little is currently known about differentionally regulated lncRNAs closely associated with Verticilium wilt resistance.

The first obstacle encountered by pathogens attempting to colonize plant tissues is the plant cell wall, which is comprised of complex structures subject to dynamic remodeling when responding to developmental and environmental factors and plays an important role in responding to diseases (Bacete et al., 2018; Molina et al., 2021). Molina et al. (2021) analyzed how plant cell walls relate to immunity by assessing the susceptibility of Arabidopsis cell wall mutants (cwms) to various pathogens. This included identifying how the wall composition determines disease-resistance phenotypes and how wall-mediated defensive pathways contribute to cwm resistance (Molina et al., 2021). Bacete et al. (2020) described the functions of Arabidopsis Response Regulator 6 (ARR6) relating to plant disease resistance and cell wall composition. The cell wall composition of plants impaired in ARR6 function (arr6), and pectin-enriched cell wall fractions extracted from arr6 walls triggered more intense immune responses than those activated by the walls of wild-type plants, suggesting that arr6 pectin function is enriched in wall-related damage-associated molecular patterns (DAMPs), which triggers immune responses (Bacete et al., 2020). Gamir et al. (2021) analyzed the dynamics of local and systemic responses to OG perception in tomato roots or shoots, exploring both their impact on the plant and their relevance in pathogen resistance. Metabolomics and gene expression analysis in plants treated with purified OGs revealed that these treatments differentially activated the synthesis of defense-related hormones and secondary metabolites including flavonoids, alkaloids, and lignans (Gamir et al., 2021). Ohara et al. (2021) generated OsPG-FOX rice (Oryza sativa L.) lines with little pectin due to overexpression of the gene encoding a pectin degrading enzyme (OsPG) and found that under weak-light conditions, OsPG overexpression decreased both pectin content and cell adhesion, resulting in abnormally large intercellular gaps and facilitating the invasion of rice blast fungus.

In this study, we identified lncRNAs and their regulating genes associated with cell wall defense using CSSLs (the resistant CSSL-1 and the susceptible CSSL-4), both of which were inoculated with V. dahliae via RNA-seq, which reduced genetic background interference and helped with identification.

We identified multiple lncRNAs and their regulating genes based on RNA-seq and weighted gene co-expression network analysis (WGCNA). Functional characterization via VIGS demonstrated that lncRNA2 and its regulating gene GbPG12 negatively regulated CSSL-1 resistance to Verticillium wilt, while lncRNA7 and its regulating gene GbPMEI13 positively regulated V. dahliae infection.

Both of these genes have functional annotations relating to plant cell wall physiology, and subsequent experiments demonstrated that lncRNA2 and lncRNA7 accumulation were altered via treatment with the pathogen-induced pectin breakdown molecular signal oligogalacturonide (OG). OG could downregulate the expression of lncRNA2, leading to the decreased expression of GbPG12, which reduces pectin degradation. In contrast, OG upregulates the expression of lncRNA7, which could code the plant peptide hormone phytosulfokine (PSK-α). PSK-α caused IAA accumulation and activated GbPMEI13 expression with Auxin Response Factor 5 (ARF5), while GbPMEI13 inhibits GbPME enzyme activity, increasing the degree of methylation of pectin and increasing CSSL-1 resistance to V. dahliae. We also demonstrated that GbPMEI13 could inhibit the mycelial growth and spore germination of V. dahliae in vitro and that the ectopic overexpression of GbPMEI13 increases Arabidopsis resistance to V. dahliae.

This study identified multiple wilt-disease-related lncRNAs and their regulating genes and how they can inform both traditional and molecular genetics-based efforts to develop wilt-resistant cotton varieties. This study also provides insights into the molecular relationship between fungal infections, plant cell wall catabolism, and phytohormone signaling in plants.

Results

Identification and characterization of lncRNAs in cotton roots

The CSSLs CSSL-1 and CSSL-4 were inoculated with a defoliating V. dahliae isolate Vd991 by dipping the roots. The disease index of CSSL-1 and CSSL-4 was 7.14% and 66.25% at the floral bud forming stage, which demonstrated that the CSSL-1 line is resistant to Verticillium wilt, while CSSL-4 is susceptible to Verticillium wilt. This confirms that these materials are suitable to identify disease-resistance-related biomolecules (Supplemental Figure S1). The V. dahliae-inoculated roots of CSSL-1 (Root_Resistant_Inoculated [RRI]) and CSSL-4 (Root_Susceptible_Inoculated [RSI]), and the control roots of CSSL-1(Root_Resistant_Mock-treated [RRM]) and CSSL-4 (RSM [Root_Susceptible_Mock-treated]) were used for RNA sequencing (RNA-seq). Analyzing the transcriptome data identified a total of 3,049 lincRNAs, 158 lncNATs, and 163 sense lncRNAs in CSSL-1, as well as 3,041 lincRNA, 153 lncNAT, and 162 sense lncRNAs in CSSL-4 (Figure 1a). The GC content of the lncRNAs from both lines was lower than for the coding genes (Figure 1b). The average lengths of the coding genes were similar to that of the lncNATs and lincRNAs, but longer than those of the sense lncRNAs, in both CSSL-1 and CSSL-4 (Figure 1c).

Figure 1.

Identification and characterization of lncRNAs and their target mRNAs. a, Distribution of lncRNAs in CSSL-1 and CSSL-4. b, GC content of different transcripts in cotton. The horizontal line within the box plots and box limits are the median value, upper and lower quartiles of GC value, respectively. The “dots” at the end of the boxplot represent outliers, which have values larger than the upper inner fence of the box plot. c, Density plot showing the length distribution of different transcripts. d, Venn diagram for screening lncRNAs related to Verticillium wilt resistance. e, Venn diagram for screening mRNAs related to Verticillium wilt resistance. f, Sequence information of lncRNAs and structure information of GbPG12 and GbPMEI13.

Differentially expressed lncRNAs and the associated genes

We next identified the differentially expressed lncRNAs and mRNAs between the RRI, RRM, RSI, and RSM samples (Table 1) and verified the reliability of the differential expression trends from the RNA-seq data based on the Reverse transcription quantitative PCR (RT-qPCR) testing of 33 lncRNAs and mRNAs randomly selected by transcriptome analysis of RRI and RRM transcriptomes (Supplemental Figure S2). The expression trends of these selected transcripts were consistent with the results of RNA-seq, indicating that the experimental data have high reliability.

Table 1.

Differentially expressed lncRNAs and mRNAs

Group	Upregulated lncRNAs	Downregulated lncRNAs	Upregulated mRNAs	Downregulated mRNAs
RRI-versus-RRM	60	93	3,293	1,621
RSI-versus-RSM	58	197	3,256	3,685
RRM-versus-RSM	95	216	4,201	4,392
RRI-versus-RSI	79	86	2,683	1,419

Candidate regulating genes of lncRNAs

lncRNAs regulate gene expression in several ways (Wang and Chang, 2011). We used the Tbtools software (https://github.com/CJ-Chen/TBtools) to predict interactions between lncRNAs and their neighboring genes within 10,000 bp upstream or downstream, ultimately identifying 1,592 lncRNAs and 2,757 neighboring coding genes (Supplemental Table S1). We identified 3,425 lncRNAs with corresponding positive and negative co-expressed candidate target or regulating coding genes by performing a WGCNA on lncRNAs trans-regulated genes (Supplemental Table S2). A total of 163 lncRNAs matched the sense strand of the coding genes (Supplemental Table S3 and Supplemental Figure S3, a–d), and 20 lncRNAs could be premiRNA and generate 22 kinds of microRNAs (miRNA) (Supplemental Table S4).

Comparison of pathogen-induced expression profiles of lncRNAs in CSSL-1 and CSSL-4

A Venn diagram demonstrated that 59 lncRNAs were differentially induced by the pathogen in the resistant CSSL-1 line that was not induced in the susceptible CSSL-4 line (Figure 1d). Target gene predictions for these 59 candidate wilt-resistance-related lncRNAs identified 109 putative regulating genes (Supplemental Table S5). Analysis of the differential expression profiles of these targets or regulating genes in the four experimental groups (RSM-versus-RRM, RSM-versus-RSI, RRM-versus-RRI, and RSI-versus-RRI) demonstrated that seven differentially regulated mRNAs were induced by the pathogen in CSSL-1 but were not induced in CSSL-4. Therefore, these seven mRNAs are candidates for resistance to Verticillium wilt (Figure 1e). We combined the candidate disease-related lncRNAs with these seven candidates and found seven paired lncRNA and predicted regulating gene pairs (Supplemental Table S6). Of these, we were particularly interested in lncRNA2 and lncRNA7, which were predicted as regulators and could alter resulting expression levels, including a predicted cotton homolog of a polygalacturonase enzyme (henceforth named GbPG12) and the homolog of a pectin methylesterase (PME) inhibitor protein (henceforth GbPMEI13) (Supplemental Figure S4). Previous reports have assessed polygalacturonase (PG) and PME inhibitor (PMEI) proteins in cotton pectin metabolism (Nicholass et al., 1995; Liu et al., 2018).

Cloning and expression profiling of lncRNAs and their regulating genes

We cloned and analyzed lncRNA2, PG12, lncRNA7, and PMEI13 (Figure 1f). lncRNA2 is 1,126 bp long and contains a 159 bp ORF in the sense strand, as well as two ORFs of 297 and 315 bp in the antisense strand (MK836304). The PG12 is 1,587 bp in length and encodes a 529 amino acid (aa) protein with a C-terminal putative enzymatic domain typical of polygalacturonases (PLN02793 domain) (MK836302). lncRNA7 is 331 bp long and contains a 222-bp ORF, putatively encoding a 74-aa polypeptide that shares 40% identity with the PSK precursor (pPSK) 3 of A. thaliana (MK836305). PMEI13 is 606 bp long and putatively encodes a 202-aa protein that has a C-terminal PMEI-like_3 domain. PMEIs belong to a large multi-gene encoded protein family, PF04043 (MK836303). In plants, PMEIs are targeted to the extracellular matrix, typically inhibit plant PMEs to control endogenous PME activity, and play an important physiological role in the regulation of pectin methyl esterification and fungal resistance (Lionetti et al., 2007, 2014, 2017).

lncRNAs and their regulating genes are primarily expressed in cotton roots and respond to V. dahliae infection or OG

RT-qPCR analysis of the total RNA extracted from various cotton organs with gene-specific primers demonstrated that lncRNA2, PG12, lncRNA7, and PMEI13 are expressed in the organs we examined, and were particularly pronounced in the roots of CSSL-1 (Figure 2, a–d). This is consistent with the roots being the primary organ defending against Verticillium wilt.

Figure 2.

RT-qPCR results. a–d, lncRNA2, PG12, lncRNA7, and PMEI13 tissue expression analysis. e–h, lncRNA2, PG12, lncRNA7, and PMEI13 expression levels after inoculation with V. dahliae. i–j, lncRNA2, PG12, lncRNA7, and PMEI13 expression levels following OGs treatment. Error bars indicate SD with three biological replicates in all the subplots. * and ** indicate statistical significances at the 0.05 and 0.01 probability levels, respectively, using Student’s t test.

Our RNA-seq results from the aforementioned V. dahliae inoculation experiment indicated that both lncRNA2 and PG12 were significantly downregulated upon infection; this was in direct contrast with the significantly increased levels we observed for lncRNA7 and PMEI13. We conducted subsequent follow-up V. dahliae infection experiments and monitored the expression levels of lncRNAs and their regulating genes with RT-qPCR and found that lncRNA2 and PG12 levels were significantly lower in infected plants compared with mock-inoculated controls, at time periods of up to 5  days post inoculation (DPI; Figure 2, e and f). This RT-qPCR analysis demonstrated significant increases in lncRNA7 and PMEI13 levels upon infection with V. dahliae, both of which exhibited bi-modal accumulation peaks at 0.5–3 hours post inoculation (HPI) and 3 DPI (Figure 2, g and h), which is consistent with the RNA-seq results.

Previous reports have found that the PG and PMEI proteins are involved in cotton pectin metabolism (Choi et al., 2018; Liu et al., 2018). Therefore, we profiled lncRNAs and the expression of their regulating genes in cotton seedlings, which were spray-treated with OG—a pectin-breakdown product. We found that the OG treatment of CSSL-1 seedlings significantly suppressed the expression of lncRNA2 and PG12 but that lncRNA7 and PMEI13 levels continuously increased from 6 to 24 HPI (Figure 2, i and j). Therefore, we can conclude that lncRNA2, PG12, lncRNA7, and PMEI13 were all regulated by OG and could be involved in Verticillium wilt resistance through the DAMPs pathway.

Silencing of lncRNA2 and PG12 enhances cotton resistance to V. dahliae

To assess the potential functional impacts of lncRNAs and their regulating mRNA transcripts during a V. dahliae infection, we manipulated the cellular levels of both RNAs in V. dahliae-infected cotton seedlings. We started with the VIGS-based silencing of lncRNA2 and PG12 expression, and used an empty pTRV2 vector for a negative control and a pTRV:CLA silencing for a positive control in 2-week-old seedlings of the CSSL-1 cotton line. These VIGS-knockdown seedlings were subsequently inoculated with V. dahliae Vd991. As shown in Figure 3a, the newly emerged true leaves of plants infiltrated with Agrobacterium GV3101 carrying GbCLA exhibited an albino phenotype, indicating that the VIGS system worked under our experimental conditions. No substantial phenotypic differences were observed prior to the V. dahliae Vd991 infection between the empty vector and lncRNA2-KD plants, while the PG12-KD plants grew more slowly (Figure 3, b–d). On 20 DPI, the leaves of the empty vector control plants began yellowing, with some defoliation. These plants were more susceptible to Verticillium wilt than the lncRNA2-KD and PG12-KD plants (Figure 3, e–g), which displayed reduced yellowing and better retained their leaves.

Figure 3.

VIGS results of lncRNA2 and PG12. Error bars indicate SD with three biological replicates in all the subplots. ** indicates statistical significance at the 0.01 probability level using Student’s t test. a–g, Disease symptoms. h–k, Percentage of diseased plants and disease index of the control, lncRNA2-KD, and PG-KD plants. l–n, Symptoms of Verticillium wilt in stems of the control, lncRNA2-KD, and PG-KD plants. o, Relative biomass of V. dahliae in stems of the control, lncRNA2-KD, and PG-KD plants, which was determined by qPCR. p–s, Analysis of lncRNA2 and PG12 expression levels in the lncRNA2-KD and PG-KD plants inoculated with V. dahliae Vd991. t, Determination of PG enzyme activity in lncRNA2-knockdown and control plants.

Disease severity was determined using a standard visual assessment method (Xu et al., 2011; Hu et al., 2018). The lncRNA2-KD and PG12-KD plants had significantly lower disease rates and disease index values than the control plants (Figure 3, h–k). In addition, lncRNA2-KD and PG12-KD plants showed less browning of the vascular bundles than the control plants (Figure 3, l–n), and RT-qPCR analysis demonstrated that the relative amounts of V. dahiae cells in the lncRNA2-KD and PG12-KD plants were significantly lower than in the empty vector line (Figure 3o). Collectively, these results demonstrate that lncRNA2 and its regulating mRNA transcript PG12 negatively regulate cotton resistance to Verticillium wilt.

We examined the expression levels of lncRNA2 and its PG gene and observed that there were significant reductions in lncRNA2 and PG12 levels in the lncRNA2-KD plants compared with the control plants (Supplemental Figure S5, a and b). However, in the PG12-KD plants, PG12 expression levels significantly decreased, while the expression level of lncRNA2 had no significant difference compared with the control plants (Supplemental Figure S5, c and d). In inoculated lncRNA2-KD and PG12-KD plants with V. dahliae Vd991, similar RT-qPCR results were obtained (Figure 3, p–s). These results demonstrate that lncRNA2 could have a positive regulatory effect on the PG12 gene. We further measured PG enzyme activity in the lncRNA2-KD plants and control plants. Our results demonstrated that when lncRNA2 was knocked down, the PG enzyme activity of plants also decreased (Figure 3t).

Knockdown of lncRNA7 and PMEI13 compromises the resistance of cotton to V. dahliae

We conducted similar VIGS-based knockdown experiments with lncRNA7 and PMEI13. We confirmed the expected albino phenotype of the positive control pTRV:CLA plants, and observed no obvious phenotypic differences between the empty vector control, lncRNA7-KD, and PMEI13-KD plants before inoculation. (Figure 4, b–d). After 20 DPI with V. dahliae Vd991, the control plants exhibited some yellowing and defoliation; however, the knockdown of lncRNA7 and PMEI13 resulted in increased yellowing and withering (Figure 4, e–g).

Figure 4.

The VIGS results of lncRNA7 and PMEI13. Error bars indicate SD with three biological replicates in all the subplots. ** indicates statistical significance at the 0.01 probability level using Student’s t test. a–g, Disease symptoms. h–k, Percentage of diseased plants and disease index of the control, lncRNA7-KD, and PMEI13-KD plants. l–n, Symptoms of Verticillium wilt in stems of the control, lncRNA7-KD, and PMEI13-KD plants. o, Relative biomass of V. dahliae in stems of the control, lncRNA7-KD, and PMEI13-KD plants, which was determined by qPCR. p–s, Analysis of lncRNA7 and PMEI13 expression levels in the lncRNA7-KD and PMEI13-KD plants inoculated with V. dahliae Vd991.

The lncRNA7-KD and PMEI13-KD plants had higher disease rates and disease index values than the control plants (Figure 4, h–k). In addition, the vascular bundles of the lncRNA7-KD and PMEI13-KD plants browned more than the control plants (Figure 4, l–n) and an RT-qPCR-based assay indicated that the V. dahiae content in lncRNA7-KD and PMEI13-KD plants was significantly higher than in the control plants (Figure 4o). These results indicate that lncRNA7 and PMEI13 positively regulate cotton resistance to Verticillium wilt.

We found that expression levels of lncRNA7 significantly decreased in the lncRNA7-KD line and the expression levels of PMEI13 also decreased (Supplemental Figure S6, a and b). However, in the TRV: PMEI13 line, the expression levels of PMEI13 significantly decreased, while the expression levels of lncRNA7 have no significant difference compared to the control line (Supplemental Figure S6, c and d). Inoculated lncRNA7-KD and PMEI13-KD plants with V. dahliae Vd991, similar RT-qPCR results were obtained (Figure 4, p–s). These results demonstrate that lncRNA7 could have a positive regulatory effect on the PMEI13 gene. The next step is to assess how lncRNA7 regulates the expression of PMEI13 and how to increase plant resistance to Verticillium wilt.

lncRNA7 encodes PSK-α

Our bioinformatics analysis predicted that lncRNA7 encodes a 74 aa pPSK peptide; this peptide sequence has all of the prerequisite aas to enable its subsequent hydrolysis and modification to generate PSK-α (Yang et al., 2001). To confirm whether lncRNA7 can encode PSK-α, we transiently overexpressed lncRNA7 in CSSL-1 seedlings and used RT-qPCR and UPLC–MS to detect the transcription level of lncRNA7 and the content of PSK-α, respectively. Our results demonstrated that the transcription level of lncRNA7 increased in the OE:lncRNA7 lines and the content of PSK-α increased by 2.19 times compared with the control group. This indicates that lncRNA7 can encode PSK-α (Figure 5, a–d).

Figure 5.

lncRNA7 encodes PSK PSK-α; PSK-α increases the resistance of cotton to V. dahliae. Error bars indicate SD with three biological replicates in all the subplots. * and ** indicate statistical significances at the 0.05 and 0.01 probability levels, respectively, using Student’s t test. a, Relative expression level of lncRNA7 in transient overexpression plants. b, MS detection results of PSK-α standard. c, MS detection results of PSK-α extracted from CSSL-1. d, Relative content of PSK-α in transient overexpression plants. e–h, Disease symptoms of PSK-α treated plants. i and j, Percentage of diseased plants and disease index of the control and PSK-treated plants. k and l, Symptoms of Verticillium wilt in stems of the control and PSK-treated plants.

Exogenous application of PSK increases the resistance of cotton to V. dahliae

We treated the CSSL-1 seedlings with PSK-α and found that the exogenous application of PSK-α increased cotton resistance to V. dahliae and did not cause any obvious significant changes in the phenotypes of these plants. Before initiating the infection, we observed no significant differences between the synthetic PSK-α treated cotton and the untreated control plants before inoculation (Figure 5, e and f). Following inoculation with V. dahliae Vd991, the PSK-α treated cotton showed milder symptoms of Verticillium wilt with less yellowing of the leaves compared with control plants (Figure 5, g and h). PSK-α-treated cotton plants had lower mortality rate and disease index values (Figure 5, i and j), while the vascular bundles of PSK-α-treated plants had less browning compared with the control plants (Figure 5, k and l).

PSK-α treatment decreases lncRNA7 accumulation but increases the accumulation of IAA and auxin-signaling-related transcription factors ARF5 and PMEI13

Mature PSK-α is perceived by the transmembrane receptor protein PSKR1, which triggers an increase in cytosolic Ca²⁺ levels and activates auxin-mediated pathways. Both of these processes enhance tomato plant resistance to B. cinerea (Zhang et al., 2018a). We found that the PSK-α treatment of CSSL-1 seedlings significantly induced the expression of PMEI13 and the classic auxin signaling transcription factor ARF5 but that lncRNA7 levels decreased within 9 h of PSK-α treatment, before returning to normal after 12 h (Figure 6a). The results of the high-performance liquid chromatography-mass spectrometry (HPLC-MS/MS) analysis demonstrated that the IAA level of CSSL-1 seedlings significantly increased 6 h after PSK-α treatment (Figure 6b). These results demonstrate PSK-α’s ability to reduce lncRNA7 levels and confirm that PSK-α can cause accumulation of IAA. This indicates that PSK-α could upregulate PMEI13 expression by an auxin-mediated mechanism.

Figure 6.

PSK upregulated PMEI13 expression through the IAA signaling pathway via ARF5. Error bars indicate SD with three biological replicates in all the subplots. * and ** indicate statistical significances at the 0.05 and 0.01 probability levels, respectively, using Student’s t test. a, Relative expression levels of lncRNA7, PMEI13, and ARF5 in CSSL-1 after treatment with PSK. b, The IAA content in the control and PSK-treated plants. c, The relative expression level of lncRNA7, PMEI13, and ARF5 in CSSL-1 after NAA treatment. d, cis-acting element analysis of the 2,000 bp upstream of PMEI13. e, Yeast one-hybrid results. Yeast transfected with pABAi-PMEI13P plasmid grew well on SD/-Ura plates and growth was inhibited on SD/-Ura with AbA (500 ng/mL) plates. Yeast transfected with pABAi-PMEI13P and pGADT7-ARF5 plasmids grew well on SD/-Ura and SD/-Ura with AbA (500 ng/ml) plates. f, A single colony of transformants was randomly picked and diluted in SD/-Ura and SD/-Ura with AbA (500 ng/mL) plates after gradient dilution. The growth of pABAi-PMEI13P transformants on SD/-Ura with AbA (500 ng/mL) plates was significantly inhibited and the colony growth was much less than that on the SD/-Ura plates, indicating that PMEI13P did not self-activate under these experimental conditions. Since the co-transformation of pABAi-PMEI13P and pGADT7-ARF5 can activate the ABAi reporter gene, they can grow normally on SD/-Ura with AbA (500 ng/mL) plates. g, Transient transcriptional activity assay. Four-week-old N. benthamiana plants were transiently transformed with different combinations of constructs, as indicated in the picture. Relative LUC activity was observed and calculated by indiGO software. h, 10 ng/mL IAA treatment enhanced the resistance of CSSL-1 to V. dahliae. After trypan blue staining, the number of dead cells in the treatment group was significantly lower than that of the control group.

Treatment with the synthetic auxin NAA does not affect lncRNA7 expression but upregulates expression of both PMEI13 and ARF5

We explored the potential involvement of auxin signaling in the regulation of lncRNA7 and PMEI13 by treating CSSL-1 seedlings with the synthetic auxin 1-Naphthaleneacetic acid (NAA). The results indicated that NAA treatment had little effect on lncRNA7 accumulation at most time points, though small but significant increases were detected at 24 h. This indicates that the NAA pathway was located downstream of the PSK. In contrast, NAA treatment significantly induced the accumulation of both PMEI13 and ARF at most time points over a 24-h treatment period (Figure 6c). This indicates that NAA could upregulate the expression of PMEI13 by ARF. Analysis using the jaspar website (http://jaspar.genereg.net/) demonstrated that a 2-kb upstream fragment of the transcription start site of the GbPMEI13 locus contains two putative ARF5 cis-acting elements and seven ARF8 cis-acting elements (Figure 6d). Yeast one-hybrid assay results demonstrated that ARF5 can bind to the PMEI13 promoter to initiate transcription (Figure 6, e and f). Further transient transcriptional activity assays demonstrated similar results (Figure 6g). Furthermore, we measured the change in resistance of CSSL-1 to V. dahliae when applying a certain amount of IAA. When treated with 10 ng/mL IAA, the number of dead cells in CSSL-1 in the experimental group was lower than in the control group when inoculated with V. dahliae (Figure 6h).

PMEI13 is mainly expressed in the vascular system and roots, and PMEI13 is localized in the cell wall and interacts with PME to regulate the degree of methylation of pectin

The expression pattern of GbPMEI13 was further confirmed by β-glucuronidase (GUS) expression analysis in transgenic Arabidopsis plants. The 2-kb upstream fragment of GbPMEI13 was cloned and expressed in the GUS cDNA under the control of this fragment in the plant expression vector PKGWFS7.0. In transgenic Arabidopsis, expression of the GbPMEI13 pro::GUS fusion was detected in the roots, stem, and leaves (Figure 7a). Furthermore, as shown in the enlarged image, GbPMEI13 was primarily concentrated in the vascular system and roots (Figure 7, b–d).

Figure 7.

PMEI13 is mainly expressed in the cell wall of plant roots, which can inhibit the enzymatic activity of PME and change the degree of methyl esterification of pectin. a–d, The GUS results of the PMEI13 promoter. e, Subcellular localization demonstrated that PMEI13 is located in the cell membrane and cell wall. f, After the wall separation, subcellular localization indicated that PMEI13 is located in the cell wall. g, Docking analysis of PMEI13 and PME31 demonstrated that PMEI13 and PME31 could form a 1:1 stoichiometric complex. h, PME enzyme activity of PMEI13-knockdown cotton plants, control plants, WT, and PMEI13-OE Arabidopsis plants. i, The degree of methylation of pectin of PMEI13-knockdown cotton plants, control plants, WT, and PMEI13-OE Arabidopsis plants. Error bars in (h) and (i) indicate the SD with at least three biological replicates (n ≥ 3). * and ** indicate statistical significances at the 0.05 and 0.01 probability levels, respectively, using Student’s t test.

A subcellular localization experiment using a GbPMEI13– green fluorescent protein (GFP) fusion protein construct in Nicotiana benthamiana demonstrated that PMEI13–GFP is found at the cell periphery near the cell wall (Figure 7e), and plasmolysis experiments with confocal microscopy analysis indicated that GbPMEI13-GFP is primarily concentrated in the cell wall (Figure 7f).

Docking analysis of PMEI13 and PME31 demonstrated that PMEI13 and PME31 could form a 1:1 stoichiometric complex (Figure 7g). PME enzyme activity was measured in PMEI13-DOWN cotton plants and PMEI13-OE Arabidopsis plants. The results demonstrated that PMEI13 expression was negatively corresponded with PME enzyme activity, indicating that PMEI13 can inhibit PME activity (Figure 7h). Our results demonstrated that the degree of pectin methylation was positively corresponded with the expression level of PMEI13 (Figure 7i).

Previous studies have demonstrated that methyl-esterified pectin is less susceptible to fungal PGs than pectin with lower degree of methoxylation (DM) (Limberg et al., 2000; Wietholter et al., 2003; Chiara Volpi et al., 2011). Liu et al. (2018) also reported that GhPMEI3 could interact with PMEs and that the ectopic expression of GhPMEI3 increased pectin methyl esterification and limited fungal disease in Arabidopsis. Our results confirmed that PMEI13 can prevent the demethylation of high-DM pectin by interacting with PME, indicating that it plays a role in the resistance to Verticillium wilt in cotton.

Purification and characterization of recombinant GbPMEI13, and PMEI13 could inhibit the mycelial growth and spore germination of V. dahliae

The GbPMEI13 was cloned into the pET32a plasmid, producing the recombinant PMEI13 protein in the Escherichia coli BL21 DE3 strain. The molecular mass of the purified protein was close to the value inferred from its gene sequence, and the identity of the purified PMEI13 protein was further confirmed by Western blot analysis: its His6-tag was recognized by a His-tag antibody (Figure 8a).

Figure 8.

Heterologously expressed GbPMEI13 inhibits the mycelial growth and spore germination of V. dahliae in vitro. a, GbPMEI13 heterologous expression; M: protein maker; line 1, IPTG no induced; line2: IPTG induced; lines 3–4: purified GbPMEI13; lines 5–6: Western blot of GbPMEI13. b, The effect of GhPMEI 13 on anti- V. dahliae. c, Verticillium dahliae mycelia grew well without GbPMEI13. d, GhPMEI 13 inhibited the mycelial growth of V. dahliae at 120 ug/mL. e, GhPMEI 13 inhibited spore germination of V. dahliae at 240 ug/mL.

Purified PMEI13 was used to characterize the capacity of anti-V. dahliae. As shown in Figure 8b, the growth of V. dahliae gradually slowed as the concentration of PMEI13 increased. A microscopic photograph demonstrated that V. dahliae without protease treatment grew well after 72 h of incubation at 25°C (Figure 8c). Additionally, V. dahliae mycelium growth was substantially inhibited by PMEI13 treatment at concentration levels of 120 ug/mL (Figure 8d), and the V. dahliae spore germination was completely inhibited by PMEI13 treatment at concentration levels of 240 ug/mL (Figure 8e).

Ectopic overexpression of GbPMEI13 in Arabidopsis promotes plant growth and enhances resistance to V. dahliae

We obtained and analyzed three overexpression GbPMEI13 transgenic lines, which exhibited different phenotypes from the wild-type and GbPMEI13 knockout mutant plants and demonstrated that GbPMEI13 overexpression promotes growth in Arabidopsis (Figure 9a). We tested the resistance to V. dahliae of various transgenic and knockout lines and observed that all three GbPMEI13-OE lines had less yellowing on their leaves than the wild-type and Atpmei13 plants 15 DPI with V. dahliae. The average disease index value of the three GbPMEI13-OE lines was 23.96, while the DI values of the wild-type and Atpmei13 group were 35.94 and 51.39, respectively (Figure 9, a–c). These results demonstrate that the ectopic overexpression of GbPMEI13 can significantly enhance Arabidopsis resistance to Verticillium wilt.

Figure 9.

Ectopic overexpression of GbPMEI13 in Arabidopsis enhances resistance to V. dahliae. a, Disease symptoms. b, Percentage of different disease levels in plants. c, Disease index of GbPMEI13 over-expression Arabidopsis lines, WT, and At PMEI13 lines. Error bars indicate the SD with at least three biological replicates (n ≥ 3). * indicate statistical significance at the 0.05 probability level using Student’s t test.

Discussion

Verticillium wilt is a soil-borne disease that can decrease the production of cotton and other crops. Previous studies assessing the mechanism of cotton resistance to Verticillium wilt have primarily focused on functional coding genes, miRNAs, small RNA, and the circRNAs (Tim and Tmo, 1999; Rohrig et al., 2002; Liu et al., 2017; Choi et al., 2018; Xiang et al., 2018; Xiao et al., 2021). However, the mechanism behind how lncRNAs and their target genes regulate cotton resistance to Verticillium wilt remains unclear.

Resistant island cotton and a susceptible upland cotton cultivar were used to study lncRNAs associated with resistance to Verticillium wilt, and the number of differentially induced lncRNAs in Gossypium barbadense and Gossypium hirsutum were ∼1,200 and 1,900, respectively (Zhang et al., 2018b). This study used two cotton introgression lines with nearly identical genetic backgrounds but significant differences in disease resistance (Li et al., 2019), which helped to reduce interference from genetic background noise and facilitate the accurate identification of resistance-related genes and lncRNAs.

We sequenced the roots of cotton plants inoculated with V. dahliae or water, which was used as a control, and we obtained the mRNA and lncRNAs differentially expressed between the two materials and treatments. There were 153 and 165 differentially expressed disease-responsive lncRNAs in the resistant and susceptible lines, respectively. Additionally, we found that the number of differentially expressed genes or lncRNAs in the resistant and susceptible lines was greater in the control than in V. dahliae-inoculated plants (Table 1). This indicates that fungal induction reduced the number of differentially expressed genes between the two lines.

The most-studied functions of lncRNAs and their related cis-regulation relate to their involvement with their target genes. For example, lncRNA Evx1 was found to regulate its neighboring gene EVX1 (Luo et al., 2016) while GhlncNAT was found to regulate the overlapping receptor-like kinase gene GhANX2 (Zhang et al., 2018b). The latest research found that lnc-Ga13g0352 might act as a dual-functional regulator that either activates or inhibits the transcription of target genes Ga12g02719 and Ga05g03705 (Zheng et al., 2021). We used bioinformatics analysis and RT-qPCR verification to demonstrate that lncRNA2 and lncRNA7, has transcriptional regulatory relationships with PG12 and PMEI13, respectively (Supplemental Figures S5, a–d and S6, a–d; Figures 3, p–s and 4, p–s).

The expression levels of lncRNA2, PG12, lncRNA7, and PMEI13 are regulated by the infection of V. dahliae (Figure 2, e–h). The functions of lncRNA2, PG12, lncRNA7, and PMEI13 were characterized by VIGS. Silencing lncRNA2 and PG12 enhances the resistance of cotton to V. dahliae, while the silencing of lncRNA7 and PMEI13 compromises the resistance of cotton to V. dahliae. Ectopic overexpression of GbPMEI13 in Arabidopsis promotes growth and enhances resistance to V. dahliae. Previous reports have suggested that PG and PMEI proteins are involved in cotton pectin metabolism and defend against pathogenic fungi (Odat et al., 2014; Choi et al., 2018; Liu et al., 2018; Chiniquy et al., 2019; Lu et al., 2019; Del Corpo et al., 2020). DAMPs were proposed to describe plant elicitors like OGs, where pectin homogalacturon are broken down by pectinases. OGs have several signaling steps to elicit defenses and provide protection against pathogens (Bacete et al., 2018; Gamir et al., 2021). Therefore, we speculate that lncRNA2, PG12, lncRNA7, and PMEI13 could be related to the DAMPs pathway, which was confirmed by the RT-qPCR results of OG treatment (Figure 2, i and j).

Over the past decade, tens of thousands of lncRNAs have been annotated in animal and plant genomes. Some of these reports have revealed the existence of stable and functional small peptides translated from lncRNAs (Johnson and Douglas, 2007; Nelson et al., 2016; Choi et al., 2019; Zhang et al., 2021). In soybean (Glycine max), two small peptides (12- and 24-aa-long) translated from the ENOD40 transcript were found to interact with sucrose synthase, which is required for plant symbiosis (Tian et al., 2004). Nelson et al. (2016) discovered a putative muscle-specific lncRNA that encodes a peptide of 34 aas DWORF, DWORF localizes to the SR membrane, where it enhances SERCA activity by displacing the SERCA inhibitors. Matsumoto et al. (2017) reported that LINC00961 generates SPAR polypeptides that act via the lysosome to suppress aa-mediated mTORC1 activity, thereby modulating skeletal muscle regenerative responses following injuries. Our work demonstrates that GblncRNA7 encodes the small peptides of 75 aas, which is similar to the pPSK 3 of A. thaliana. Additionally, lncRNA7 overexpression and UPLC–MS experimental results indicate that lncRNA7 can encode PSK-α (Figure 5, a–d).

PSK is an important signaling molecule composed of disulfated pentapeptide and has recently been implicated in plant defenses to pathogenic infections. Zhang et al. (2018a) demonstrated that PSK can act as a DAMP and was primarily perceived by PSKR1, which increases cytosolic [Ca²⁺] and activated auxin-mediated pathways to enhance the immunity of tomato plants to B. cinerea. Our RT-qPCR results demonstrated that lncRNA7, which encodes the cotton pentapeptide PSK, was induced by V. dahliae (Figure 2g). Further research found that OG treatment can significantly increase the expression level of PSK (Figure 2j), and that PSK can regulate PMEI13 expression in the IAA pathway using ARF5 (Figure 6, a–g).

ARFs are the key factor of the IAA pathway and can activate or repress transcription by binding with auxin-response elements (Dorokhov et al., 2006). AtRF5 is involved in early embryogenesis and vascular tissue formation (Vorwerk et al., 2004). Two poplar (Populus trichocarpa) genes are homologous to AtARF5: both are involved in vascular tissue development and could play similar roles in the development of the secondary xylem of poplar (Johnson and Douglas, 2007). AtRF8 affects hypocotyl elongation and root growth (Tian et al., 2004). Our experimental results confirmed that GbARF5 can bind to the promoter region of GbPMEI13 and positively regulate the expression of GbPMEI13.

GbPMEI13 is a member of the PMEI family of proteins. Liu et al. (2018) reported that GhPMEI3, another member of the PMEI family, is involved with cotton resistance to Verticillium wilt at the protein level. GhPMEI3 is a 240-aa protein with no signal peptides; the theoretical M_r of GhPMEI3 is 25.96 kD with a pI of 5.07. Sequence alignment found that the nucleic acid and protein similarity between GbPMEI13 and GhPMEI3 was 37.53%, and 21.95%, respectively. In addition, PMEI13 has a 28 aa signal peptide, with a theoretical M_r of 21.48 kD and a pI of 6.81. In addition to differences in their sequences, the pI of PMEI13 is close to neutral. One of the important factors affecting the stability of PMEI is ion interactions: subtle pH changes in the microenvironment significantly impact the stability of partially buried ionic interactions (Bonavita et al., 2016). Most plant PMEs were neutral to alkaline, therefore, PMEI13 could better inhibit PME activity.

The subcellular localization experiment on GbPMEI13 revealed that its protein was mainly localized on the cell wall. GUS reporter-aided analysis of the activity of the GbPMEI13 promoter demonstrated that the gene is primarily expressed in the vascular bundle. Verticillium wilt is a vascular disease, and GbPMEI13 could prevent the degradation of the vascular bundle cell walls. This was verified by the increased browning of the cotton stem vascular tissues once GbPMEI13 was silenced in cotton (Figure 4n).

The plant cell wall is the primary barrier responsible for guarding against pathogens. Pectin is the main component of the cell wall and plays an important role in plant resistance to pathogen infection. PMEs can demethylate methyl esterified pectin and are involved in cell wall modification and plant defense response (Johnson and Douglas, 2007; Odat et al., 2014). Plant PMEIs, including pepper CaPMEI1 and cotton GhPMEI3, not only inhibit PMEs by forming a 1:1 ratio complex but also possess antifungal activity and play an important role in disease resistance (An et al., 2008; Liu et al., 2018). Our experiments confirmed that PMEI13 overexpression can inhibit PME enzyme activity and increase pectin methylation while silencing PMEI13 expression has the opposite effect. We also demonstrated that GbPMEI13 could inhibit the mycelial growth and spore germination of V. dahliae in vitro (Figure 8, b–e).

Several studies have reported that highly methylated pectin can effectively prevent the degradation of PG enzymes, helping plants resist pathogenic fungal infections (Lionetti et al., 2007; Liu et al., 2018). Our results demonstrated that silencing the expression of the PMEI13 gene decreased the methylation of pectin in the cotton cell wall, weakening the resistance of cotton to Verticillium wilt. This is consistent with the results of previous studies.

The plant cell wall is composed of cellulose, hemicellulose, and pectin, and is the first point of contact between plants and their microbial pathogens (Bacete et al., 2018). When pathogens breach the cell walls, host peptides and oligosaccharide fragments are released as DAMPs. One of the most well-characterized DAMPs is pectin-derived OGs, which can elicit defenses and provide protection against pathogens (Bonavita et al., 2016; Gao et al., 2016; Gamir et al., 2021). DAMPs are the typical response of plant cell wall reinforcements and can be beneficial to plants. In this study, we report how cotton changed the methyl esterification of pectin through the DAMPs pathway and reconstructed the cell wall structure to resist Verticillium wilt following V. dahliae infection. While infesting cotton, V. dahliae secretes the pectin lyase, VdPG to degrade the cell walls of cotton. Pectin in the cotton cell wall was degraded to produce OG, OG can downregulate lncRNA2 expression, which can positively regulate GbPG12 expression, while low-level PG expression attenuates the degradation of pectin by cotton plants. OGs can upregulate the expression of GblncRNA7, which encodes PSK. PSK upregulates GbPMEI13 expression by ARF5 via the IAA pathway. Highly expressed PMEI inhibited PME enzyme activity, increasing pectin methylation in the cell wall. High-methyl esterified pectin enhanced resistance to VdPG degradation; furthermore, GbPMEI13 can inhibit the mycelial growth and spore germination of V. dahliae, increasing cotton resistance to Verticillium wilt (Figure 10).

Figure 10.

Model of lncRNAs and their regulating genes involved in cotton resistance to V. dahliae. VdPG to degrade the cell wall of cotton. The pectin in the cotton cell wall was degraded to produce OG. OGs can downregulate the expression of GbPG12 by downregulating the expression of GblncRNA2, which can positively regulate the expression of GbPG12. Low levels of PG expression attenuate the degradation of pectin by cotton plants. OGs can upregulate the expression of GblncRNA7, which encodes the pPSKs. Mature PSK (PSK-α) can be perceived by its receptor (PSKR), and upregulates GbPMEI13 expression by ARF5 via the IAA pathway. High-methyl esterified pectin enhanced resistance to VdPG degradation; furthermore, GbPMEI13 could inhibit the mycelial growth and spore germination of V. dahliae, increasing cotton resistance to Verticillium wilt.

In summary, this study demonstrates that lncRNAs in cotton can regulate the genes involved in plant resistance to V. dahliae by reinforcing cell walls. lncRNA2, lncRNA7, and their regulating genes are all involved in the resistance of cotton plants to multiple fungal pathogens. Our study demonstrates that lncRNAs are important regulators of the immune response system and provide a strategy for breeding cotton varieties with durable broad-spectrum resistance.

Materials and methods

Plant materials and fungal infection assays

The CSSLs were developed by backcross and repeated selfing using “CCRI36” (G. hirsutum) as the recipient parent and “Hai1” (G. barbadense) as the donor parent (Li et al., 2019). We used two of these CSSLs in this study: CSSL-1 (MBI8255), which was highly resistant to Verticillium wilt, and CSSL-4 (MBI8232) which was highly susceptible to Verticillium wilt. Both were provided by Professor Youlu Yuan of the Cotton Institute of Chinese Academy of Agricultural Science.

Seeds of these two lines were sown in vermiculite-filled pots, watered with Hoagland’s solution, and grown in a greenhouse at 25°C under a 16-h light and 8-h dark photoperiod for 4 weeks. A culture of the defoliating strain of Verticillium dahliae, Vd991, was initiated by plating stock culture stored at −80°C on a Potato Dextrose Agar medium at 25°C. After 3–4 d, the actively growing hyphal fronts were transferred to Czapek’s medium and grown for 3 d at 25°C. The culture was then filtered through two layers of Kimwipes. The conidial concentration was determined using a hemocytometer and the filtrate was adjusted to 1 × 10⁶ conidia per milliliter by adding water, which was used as the inoculum.

When two fully expanded leaves appeared, the whole plants were uprooted from the vermiculite and the roots were dipped in the inoculum for 5 min. The inoculated plants were then returned to the field disease nursery. The roots of the plants postinoculation were harvested at the floral bud forming stage. The roots of the plants treated with distilled water and planted in nonV. dahliae fields were collected at predetermined time points and were used as controls. All samples were immediately stored at −80°C until further use.

Pathogenicity assays of CSSL-1 and CSSL-4 were conducted in the field at the florescence stage; the number of plants with Verticillium wilt symptoms and the total numbers of plants assayed were counted as previously reported (Xu et al., 2011; Hu et al., 2018a, 2018b).

Construction of the lncRNA-seq library and RNA-seq

For whole-genome RNA-seq, the V. dahliae-inoculated roots of CSSL-1 (RRI) and CSSL-4 (RSI), and the control roots of CSSL-1 (RRM) and CSSL-4 (RSM) were used. Three biological replicates were performed for each treatment.

Total RNA was isolated from each cotton seedling sample using RNAprep Pure Plant Kits (TIANGEN Biotech, Beijing, China). Ribosomal RNA was removed using the Ribo-Zero Gold Kit (Epicentre, Madison, WI, USA). Sequencing libraries were constructed using the NEB Next Ultra Directional RNA Library Prep Kit for Illumina (NEB, Ipswich, MA, USA), according to the manufacturer’s instructions. The libraries were sequenced on the Illumina HiSeq 4000 platform and 150 bp paired-end reads were generated.

Identification of lncRNAs

The raw data were processed by filtering adaptors, removing the low-quality reads, and trimming the reads whose number of N bases accounted for more than 5% of the total (quality score, Q ≥ 30). The reference G. barbadense genome and the annotation files were downloaded from the CottonGen database (http://www.cottongen.org; Yuan et al., 2015). The genome index was built with Bowtie2 version 2.2.3 and clean reads were mapped to the G. barbadense genome using TopHat version 2.0.12. A transcript was deemed to be noncoding if its protein-coding potential was ˂0, which meant that the transcript cannot encode a protein.

Expression analysis

The expression levels of lncRNA and the mRNA transcripts were estimated using Cufflinks and all mapped reads by Cufflinks. All RNA-seq datasets were respectively aligned to the cotton genome using TopHat version 2.0. The transcriptome from each dataset was then independently assembled using Cufflinks version 2.0. The differential gene expression analyses were conducted using DESeq version 1.16, based on a negative binomial distribution. Genes with Q ≤ 0.05 and |log2_ratio| ≥ 1 were considered to be differentially expressed genes.

lncRNAs regulate gene prediction

The lncRNAs cis-regulated target genes were predicted by Tbtools (Chen et al., 2018). The lncRNAs trans-regulated target genes were predicted using a Pearson correlation coefficient analysis between the expression of lncRNA and its protein-coding genes, which was calculated using a cor. test() in R (Kwenda et al., 2016). The antisense of lncRNAs was analyzed by the RNAplex software (Tafer et al., 2008). lncRNA targets of miRNAs were predicted by comparing sequence matches between the identified lncRNAs and the known miRNAs using psRNATarget (http://plantgrn.noble.org/psRNATarget/), while less than three mismatches and G/U pairs were allowed within the lincRNA and miRNA pairing regions.

Identification of unique VW-resistant lncRNAs and mRNAs of CSSL-1

A Venn diagram was used to screen for the specific lncRNAs that had a significant difference in expression levels between the resistant line CSSL-1 infected samples and mock samples while simultaneously having no significant difference in expression levels between the susceptible line CSSL-4 infected samples and mock samples.

The target genes of these CSSL-1 specific lncRNAs were analyzed with differential expression genes (DEGs) in the four comparisons: RSM-versus-RRM, RSM-versus-RSI, RRM-versus-RRI, and RSI-versus-RRI. The mRNAs which had significant differences in expression levels in the comparisons RSM-versus-RRM, RRM-versus-RRI, and RSI-versus-RRI, but no significant differences in RSM-versus-RSI were filtered out.

VIGS of lncRNA2, PG12, lncRNA7, and PMEI 13 in cotton inoculated with Vd991

VIGS vectors were constructed and Agrobacterium tumefaciens-mediated gene silencing in cotton was performed as previously described (Xie et al., 2018). Spore suspensions were prepared at 1 × 10⁶ conidia/mL in a solution containing 0.001% (v/v) Tween 20; both the control and VIGS-silenced plants were root-inoculated. The assays were performed with at least 20 plants for each constructed vector, and the experiments were repeated at least three times.

Overexpression of lncRNA7 and extraction of PSK-α

We used the transient transformation system first used by Li to assess lncRNA7 overexpression (Li et al., 2018). The PSK-α extraction method is as follows: 50 g CSSL-1 was added to liquid nitrogen to freeze-grind it, after which the ground CSSL-1 powder was placed into a beaker and 50 mL double distilled H2O (ddH2O) was added. The beaker was placed into an ultrasonic breaker at a frequency of 5 s and 5-s intervals at 60 W for 2 h. The crushed mixture was left at 4°C for 3 d, after which it was centrifuged at 10,000 rpm for 1 h at 4°C. The supernatant was freeze-dried to a powder and 2 mL of ddH₂O was added to dissolve the lyophilized powder, after which it was centrifuged at 10,000 rpm for 10 min. The supernatant was filtered with a 0.22 um filter membrane and sent for testing.

Determination of PSK-α

The PSK-α standard was synthesized by Nanjing Peptide Biotechnology Co., Ltd. (Nanjing, China). The PSK-α extracted from CSSL-1 was identified by UPLC–MS according to a previously published method (Dong et al., 2018). PSK-α were quantified using a Dionex Ultimate 3000 Rapid Resolution liquid chromatography system (Thermo Fisher Scientific, Waltham, MA, USA) and a compact accurate-mass quadrupole time-of-flight mass spectrometer (Bruker, Billerica, MA, USA) equipped with an electrospray ionization interface. Chromatographic separation was performed using a Dionex Acclaim RSLC 120 C18 (100 mm long, 2.1 mm internal diameter, 2.2 mm particle size) column (Thermo Fisher Scientific). Data were analyzed using the Bruker Data Analysis version 4.2 software.

GUS staining

The promoter sequence of GbPMEI13 was cloned into the pKGEFS version 7.0 vector and translated into the A. tumefaciens strain GV3101. Transformation of Arabidopsis plants was performed using the floral dip method. For selection, the seeds were planted in aseptic conditions on Murashige and Skoog basal salt mixture (MS) agar containing 25 mg L⁻¹ kanamycin. The selected positive seedlings were histochemically stained for GUS activity based on the method used by Niwa et al. (2013).

Subcellular localization of GbPMEI13

The GbPMEI13, without the termination codon, was cloned and inserted into the pK7FWG2.0 vector and translated into the A. tumefaciens strain GV3101. Nicotiana benthamiana plants were grown under a photoperiod of 16-h light and 8-h dark at 22°C. The A. tumefaciens strain GV3101, which contained an appropriate construct, was grown at 28°C in L-broth to the stationary phase. Bacterial cells were harvested by centrifugation and resuspended in 10 mM MgCl₂ and 10 mg/mL acetosyringone and left at room temperature for 3 h. The bacterial suspensions were infiltrated into the abaxial air spaces of the leaves of 2- to 4-week-old plants using syringes. The plants were maintained after infiltration in the dark at 22°C for 48 h. GFP fluorescence was detected via confocal laser scanning microscopy (Zeiss LSM710; Carl Zeiss, Oberkochen, Germany). GFP fusion fluorescence was excited with a 488-nm laser, 2%, collection bandwidth: 493–598 nm, gain: 752.0.

PG and PME enzyme activity assay and identifying the degree of pectin methylation

The polygalacturonase test kit (BC2660; Solarbio life sciences, Beijing, China) was used to measure PG activity, according to the manufacturer’s instructions. The Plant Pectin content detection kit (ml076399; Shanghai Enzyme Biotechnology Co., Ltd., Shanghai, China) was used to measure PME activity, according to the manufacturer’s instructions. The degree of pectin methylation was quantified according to Lionetti’s method, as previously reported (Lionetti et al., 2017).

Purification of the recombinant GhPMEI3 and determination of in vitro antifungal activity

The GhPMEI13 ORFs were amplified and cloned into the PET-32a vector to produce the fusion protein harboring a His6-tag. The recombinant GhPMEI3 was purified using a His6-tagged protein purification kit (CW BIO, Boston, MA, USA) and blotted using an anti-His6-tag monoclonal antibody.

The antifungal activity of GhPMEI13 was determined according to methods previously reported (An et al., 2008; Liu et al., 2018). Spore germination and hyphal growth were examined under a laser scanning confocal microscope (Nikon ECLIPSE Ti2).

Statistical analyses

The number of plants with Verticillium wilt symptoms and the total numbers of plants assayed were counted as in Hu et al. (2018a, 2018b). The results in RT-qPCR were calculated by the 2^−ΔΔT method. Standard deviation (SD) was calculated by the function STDEV in Excel, and statistical significances were determined by paired t test in Excel.

Data availability

The following information was supplied regarding data availability: Data is available at NCBI BioProject. ID: PRJNA514818 (http://www.ncbi.nlm.nih.gov/bioproject/514818)

Accession numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers: MK836304, for lncRNA2; MK836302, for GbPG12; MK836305, for lncRNA7; MK836303, for GbPMEI13.

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure S1. Disease symptoms of CSSL-1 and CSSL-4 at the floral bud forming stage.

Supplemental Figure S2. RT-qPCR results verify correctness of transcriptome data.

Supplemental Figure S3. Some antisense-lncRNAs and their target genes.

Supplemental Figure S4. Heatmap of lncRNA2 and lncRNA7 and their target genes.

Supplemental Figure S5. Analysis of lncRNA2 and PG12 expression levels in the lncRNA2-KD and PG-KD plants.

Supplemental Figure S6. Analysis of lncRNA7 and PMEI13 expression levels in the lncRNA7-KD and PMEI13-KD plants.

Supplemental Table S1. lncRNAs and their neighboring coding genes.

Supplemental Table S2. lncRNAs and their corresponding co-expressed coding genes.

Supplemental Table S3. lncRNAs and their antisense coding genes.

Supplemental Table S4. lncRNAs and their corresponding premiRNAs.

Supplemental Table S5. Fifty-nine wilt-resistance-related lncRNAs and their 109 target coding genes.

Supplemental Table S6. Seven pairs of wilt-resistance-related lncRNAs and their corresponding coding genes.

Supplemental Table S7. The primer sequences used in this study.

 

Y.C., Y.S. and Y.Y. conceived the study. L.Z. and Y.S. participated in experiment materials preparation. L.Z., J.C., J.L.L., P.W., and Z.C. participated in the VIGS experiment. L.Z., C.C., and Q.S. analyzed the data. L.Z., H.L., Z.Z., and J.C. performed RT-qPCR. J.L.L. overexpressed and purified PSK-α and PMEI13, Y.Z. identified PSK-α by UPLC–MS. Q.S. wrote the original draft. J.G.L guided research, analyzed data, and re-wrote the manuscript. X.Z. and Y.S. guided research. All authors read and approved the final manuscript.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plphys/pages/General-Instructions) is Yingfan Cai (yingfancai@outlook.com).