HyPRP1 performs a role in negatively regulating cotton resistance to V. dahliae via the thickening of cell walls and ROS accumulation

Background Developing tolerant cultivars by incorporating resistant genes is regarded as a potential strategy for controlling Verticillium wilt that causes severe losses in the yield and fiber quality of cotton. Results Here, we identified the gene GbHyPRP1 in Gossypium barbadense, which encodes a protein containing both proline-rich repetitive and Pollen Ole e I domains. GbHyPRP1 is located in the cell wall. The transcription of this gene mainly occurs in cotton roots and stems, and is drastically down-regulated upon infection with Verticillium dahliae. Silencing HyPRP1 dramatically enhanced cotton resistance to V. dahliae. Over-expression of HyPRP1 significantly compromised the resistance of transgenic Arabidopsis plants to V. dahliae. The GbHyPRP1 promoter region contained several putative phytohormone-responsive elements, of which SA was associated with gene down-regulation. We compared the mRNA expression patterns of HyPRP1-silenced plants and the control at the global level by RNA-Seq. A total of 1735 unique genes exhibited significant differential expression. Of these, 79 DEGs involved in cell wall biogenesis and 43 DEGs associated with the production of ROS were identified. Further, we observed a dramatic thickening of interfascicular fibers and vessel walls and an increase in lignin in the HyPRP1-silenced cotton plants compared with the control after inoculation with V. dahliae. Additionally, silencing of HyPRP1 markedly enhanced ROS accumulation in the root tips of cotton inoculated with V. dahliae. Conclusions Taken together, our results suggest that HyPRP1 performs a role in the negative regulation of cotton resistance to V. dahliae via the thickening of cell walls and ROS accumulation. Electronic supplementary material The online version of this article (10.1186/s12870-018-1565-1) contains supplementary material, which is available to authorized users.


Background
The plant cell wall contains a large set of structural proteins that are involved in defense response. Cell wall proteins are unusually rich in one or two amino acids and contain highly repetitive sequence domains. Currently, much is known about their sequence information, but there is little direct evidence of the functions of these proteins [1]. Hybrid proline-rich proteins (HyPRPs) form a subgroup of putative plant cell wall glycoproteins enriched in proline. HyPRPs are composed of three different domains: a hydrophobic signal peptide, a repetitive proline-rich domain in the N-terminus, and a hydrophobic C-terminal domain, not specifically rich in proline or glycine but containing cysteine. Based on the abundance and specific distribution of cysteine in this C-terminal domain, the HyPRPs are subdivided into A and B groups. In group A, four or six cysteine residues are present in a specific pattern (-CXXC-C-C-C-C-), while group B contains eight cysteine residues (termed the eight-cysteine motif, 8CM) in a specific order (-C-C-CC-CXC-C-C-) in the C-ter minal domain [2]. Multiple studies have indicated that group B HyPRPs play various functional roles in specific developmental stages and in response to biotic and abiotic stresses. For example, CaHyPRP1 performs distinct dual roles as a negative regulator of basal defense and in the positive regulation of cell death in Capsicum annuum against Xanthomonas campestris [3]. GmHyPRP is involved in triggering the soybean resistance response to Phakopsora pachyrhizi [4]. The HyPRP gene EARLI1 (Early Arabidopsis Aluminum Induced 1) is induced in Arabidopsis by low temperature and salt stress [5]. GhHyPRP4 has been reported to take part in the cold stress response of Gossypium hirsutum [6]. Overexpression of CcHyPRP from Cajanus cajan increased resistance to multiple abiotic stresses in yeast and Arabidopsis [7]. However, the roles of group A HyPRPs in plant development and defense against pathogen attacks remain unclear, in contrast to the relatively better characterized proteins of group B. As of now, only one group A protein, PvPRP1, has been researched extensively. PvPRP1 can be down-reglated by Colletotrichum lindemuthianum and up-regulated by wounding in the hypocotyls of French bean (Phaseolus vulgaris) [8]. The down-regulated expression of PvPRP1 in response to fungal infection is due to mRNA destabilization through the binding of PRP-BP (PvPRP1 mRNA binding protein) to a 27-nucleotide U-rich domain in the 3′ untranslated region of PvPRP1 mRNA [9].
Verticillium dahliae Kleb is a soil-borne pathogenic fungus capable of causing vascular wilt disease in cotton (Gossypium spp.). In most cotton-growing areas, Verticillium wilt has become the most important disease of cotton [10]. Unfortunately, currently available fungicides are not effective in protecting cotton from this vascular disease infection. Therefore, developing tolerant cotton cultivars by incorporating genes from resistant germplasm is now regarded as the most effective strategy for controlling this disease. Genetic dissection of Verticillium wilt resistance at the molecular level, as mediated by the relevant genes, will enhance our ability to utilize the existing germplasm to reduce cotton yield losses [11].
In recent years, high-throughput technology has been used to systematically monitor expression profiles and screen a wide spectrum of differentially expressed genes/ proteins in cotton inoculated with V. dahliae with the goal of ultimately using genetic engineering to breed resistant cultivars. A total of 188 differentially expressed proteins were identified in the roots of Gossypium barbadense upon infection with V. dahliae based on comparative proteomics analysis [12]. Moreover, 3442 unigenes related to defense responses against V. dahliae were identified in G. barbadense cv. 7124 using RNA-Seq [13]. In addition, a total of 3027 Veticillium-resistance unigenes were identified from a full-length cDNA library of G. barbadense cv Pima90-53 [14]. These compre hensive gene and protein expression data provide helpful molecular information. However, deeper insights into understanding the defense mechanisms of cotton in response to V. dahliae are needed.
Previously, we obtained a substantial number of transcript sequences from G. barbadense related to defense responses against V. dahliae [14]. Of these, a cDNA clone that encodes a group A HyPRP protein, designated as GbHyPRP1, whose expression was significantly downregulated in cotton after V. dahliae inoculation. In the present study, GbHyPRP1 and its homologous genes were cloned from other G. hirsutum cultivars. The transcriptional expression of HyPRP1 was investigated in different tissues and in response to treatment with different hormones and V. dahliae. A potential role of HyPRP1 in negatively regulating plant resistance to V. dahliae was examined by overexpression in Arabidopsis and by virus-induced gene silencing (VIGS) in cotton. We applied transcriptomic analysis to systematacially explore the molecular mechanisms underlying the HyPRP1-mediated cotton defensive response to V. dahliae. The important function of HyPRP1 involved in cotton resistance to V. dahliae via the thickening of the cell wall and ROS accumulation was proved.

Results
GbHyPRP1 is a cell wall protein down-regulated by V. dahliae challenge Previously, we isolated a full-length cDNA clone from a full-length cDNA library of G. barbadense Pima90-53 challenged with V. dahliae [14]. The cDNA clone has a 5′ untranslated region (5′ UTR) of 41 bp, 3′ UTR of 186 bp with a polyA tail and an open reading frame (ORF) of 945 bp that potentially encodes a 314-amino acid protein. The protein consists of a 26-residue signal peptide (SP) sequence with an initial ATG codon, a 165-residue proline-rich domain (PRD) (containing a basic histidine-rich domain) at the N-terminus and a 123-residue hydrophobic C-terminus Pollen Ole e I domain (cysteine-containing domain) (Fig. 1a). Based on the sequence features and a homology search, the protein belongs to group A HyPRPs, designated as GbHy PRP1 (GenBank accession number KP162172). Gb HyPRP1 has a predicted molecular weight of~33.59 kDa with a theoretical pI of 9.97. The alignment results showed that GbHyPRP1 from G. barbadense shared a high sequence identity with the other GhHyPRP1 proteins from six cultivars of G. hirsutum (Additional file 1: Figure S1).
A GFP gene fused to the C-terminal end of GbHyPRP1 under the control of the constitutive CaMV 35S promoter was successfully transformed into tobacco epidermal cells and transiently expressed. The control GFPs appeared to be distributed throughout the whole cell, which indicated that GFPs were localized to plasma membrane and cytoplasm. By contrast, the GFP signal indicated that the fusion HyPRP1-GFP proteins were localized to the cell periphery (Fig. 1b). In addition, GbHyPRP1 belongs to group A HyPRPs (Fig. 1a), which form a subgroup of plant cell wall glycoproteins enriched in proline. Thus, we inferred that HyPRP1 localized to the cell wall.
As shown in Fig. 1c, GbHyPRP1 was expressed in all tested parts of the cotton plants, with a significantly higher expression level in the stem compared with the leaf. In addition, the expression profiles of GbHyPRP1 in response to the highly aggressive defoliating V. dahliae strain were examined in infected G. barbadense roots and stems. The qPCR (real-time quantitative PCR) analysis showed that the expression level of GbHyPRP1 significantly decreased in either stems or roots after inoculation with V. dahliae (Fig. 1d). These data indicate that GbHyPRP1 is involved in the cotton-V. dahliae interaction.
V. dahliae-responsive expression of HyPRP1 exhibits the same trend in G. hirsutum as in G. barbadense We further determined the V. dahliae-responsive expression of HyPRP1 in G. hirsutum using a resistant cv. ND601(DI = 22.63 ± 2.28) and a susceptible cv. CCRI8 (DI = 57.59 ± 2.76) [15] inoculated with V. dahliae ( Fig. 2)a. The results showed that the expression of HyPRP1 dramatically decreased in both resistant and susceptible cotton stems and roots after inoculation with V. dahliae compared to the control (Fig. 2b). Moreover, the expression of HyPRP1 in the resistant cv. ND601 was significantly lower compared to the susceptible cv. CCRI8 (Fig. 2b). These results ulteriorly indicated that HyPRP1 is a negative regulator involved in the cotton-V. dahliae interaction. Fig. 1 Characterization of GbHyPRP1. a Schematic structure of GbHyPRP1. b Subcellular localization of GFP alone or GbHyPRP1-GFP fusion in tobacco leaves transiently transformed by Agrobacterium infiltration. The green fluorescence were monitored using a confocal laser scanning microscope. Bars = 10 μm. c Tissue-specific expression of GbHyPRP1 by qPCR. Two-week-old G. barbadense Pima90-53 plants were used for sampling. The values were normalized to gene PP2A1. The significant differences in expression level of GbHyPRP1 in different tissues were evaluated by the non-parametric Kruskal-Wallis test followed by Dunn's multiple comparisons test. The bar represents mean ± SE from three biological replicates (*P < 0.05). d Transcriptional analysis of GbHyPRP1 were measured in response to V. dahliae (Vd) compared to the control by qPCR. The roots and stems of two-week-old Vd-exposed Pima90-53 seedlings and control (water-treated) were collected. The values were normalized to gene PP2A1. The bar represents mean ± SE from three biological replicates. Sidak's multiple comparisons test demonstrated that there were significant differences (**P < 0.01, ***P < 0.001) between Vd and control at 0, 6, 12, 24, 36 and 48 h post-inoculation (hpi) Silencing of HyPRP1 enhances cotton resistance to Verticillium wilt We further tested whether HyPRP1 was required for cotton resistance to Verticillium wilt using VIGS, which is frequently employed as a reverse genetics technique. At approximately two weeks post-infiltration, marker gene CLA1-VIGS plants started to display the albino phenotype in the true leaves (Additional file 2: Figure S2A). At the same time, the expression of HyPRP1 in susceptible CCRI8 was not detected using semi-RT-PCR, indicating that HyPRP1 had been silenced (Additional file 2: Figure  S2B). HyPRP1-silenced plants were used for Verticillium inoculation. At 15 days post inoculation (dpi), less chlorosis and fewer wilting leaves were observed in VIGS plants compared to the control (Fig. 3a). The DI of VIGS plants (32.67 ± 1.96) was significantly lower than that of the control (81.00 ± 2.13), i.e., susceptible CCRI8 became tolerent (Fig. 3b). VIGS assays indicated that silencing of HyPRP1 significantly enhances cotton resistance to Verticillium wilt.
Over-expression of GbHyPRP1 compromises Arabidopsis resistance to Verticillium wilt To further determine the involvement of GbHyPRP1 in Verticillium wilt resistance, plant over-expression vectors containing the GbHyPRP1 ORF driven by a constitutive Fig. 2 Time-related changes in the expression of HyPRP1 in response to V. dahliae (Vd) in G. hirsutum.a ND601 showed higher tolerence to V. dahliae (left panel) with lower disease index (right panel) compared to CCRI8 at 14 dpi. Two-week-old seedlings were dip-infected with the V. dahliae spores. The disease indices were presented means ± SE from three biological replications with at least 15 plants per replication. b, The expression of HyPRP1 changed to a lower degree both in stems and roots of tolerent ND601 and susceptible CCRI8 after infection using V. dahliae comparing to the control at the same hpi. Meanwhile, the relative expression of HyPRP1 displayed more lower both in stems and roots of tolerent ND601 comparing to the susceptible CCRI8 after infection using V. dahliae at the same hpi. The relative gene expression was calculated using the comparative 2 -ΔΔCt method with PP2A1 as endogenous control gene. Values are shown as the mean ± SE of three biological replicates. Sidak's multiple comparisons test demonstrated that there were significant differences (**P < 0.01, ***P < 0.001) between Vd and control 35S promoter were constructed and transformed (Agrobacterium-mediated) into Arabidopsis thaliana. Three lines, L1, L2 and L3, were used for further analyses. After 15 dpi with V. dahliae, all the transgenic lines exhibited more wilting and etiolation compared to WT (wild type) (Fig. 4a). The average DI of transgenic plants, i.e., 69.5 (classifed as susceptable), was significantly higher than that of the control (29.8; classified as tolerant) (Fig. 4b).
These results suggest that over-expression of GbHyPRP1 compromised Arabidopsis resistance to Verticillium wilt.
GbHyPRP1 contains hormone elements in the promoter and is down-regulated by SA The upstream region of GbHyPRP1, named pGbHyPRP1, was 1431 bp in length based on sequencing (Additional file 3: Figure S3). pGbHyPRP1 fused with the reporter gene GUS introduced into Arabidopsis plants, and GUS staining confirmed that the promoter of GbHyPRP1 functioned well with respect to the expression of the reporter gene (Fig. 5a). Promoter search using PLANTCARE indicated that several potentially inducible cis-regulatory elements corresponding to hormone, defense and pathogen elicitor responses were found in pGbHyPRP1 (Fig. 5b). Thus, we investigated the expression profile of GbHyPRP1 following treatments with plant hormones, including SA (salicylic acid), abscisic acid (ABA), jasmonic acid (JA) and ethylene (ET). The results indicated that the expression of GbHyPRP1 was strongly down-regulated by SA at 12, 24, 36 and 48 hps (hours post-spraying), whereas it was significantly up-regulated by ABA, JA and ET at various time points (Fig. 5c). To our surprise, the HyPRP1 transcript levels in the SA-treated plants were quite similar to those in the V. dahliae-inoculated cotton seedlings (Figs. 1d and 2b). These results imply that HyPRP1 may play roles in cotton resistance to V. dahliae through the SA-mediated signaling pathway.

Transcriptome analysis of VIGS cotton indicated HyPRP1 influences genes related to cell wall remodeling and ROS balance
To better understand the molecular mechanisms of HyPRP1-mediated cotton defense response to V. dahliae, we empoyed a combined approach of VIGS and RNA-Seq to compare the mRNA expression patterns of HyPRP1-silenced and control plants at the global level. Transcriptome libraries yielded 56,364,572 to 63,092,290 raw reads. After cleaning and quality checks, 55,741,348 to 62,384,322 high quality reads were obtained, and 79.66-80.87% of these reads were uniquely mapped to G. hirsutum L. acc. TM-1 [16], representing more than forty thousand unigenes(RPKM ≥1)( Table 1). Under V. dahliae stress, a total of 1735 unique genes exhibited significant differential expression based on VIGS compared to the control, including 816 up-regulated and 919 down-regulated genes (P-value ≤ 0.05) (Additional file 4: Data S1).
Of these, we identified 79 differentially expressed genes (DEGs) that might be involved in cell wall  Table 2). Expansins and xyloglucan-modifying enzymes are usually up-regulated by various stresses and then become involved in cell wall remodeling [17]. Our RNA-Seq results indicated that the expression of six expansin genes (Gh_A10G1374, Gh_A05G2385, Gh_A05G3493, Gh_A1 3G0050, Gh_D04G1924 and Gh_D05G2650) and seven xyloglucan endotransglucosylase/hydrolase (XEH) genes (Gh_A02G1426, Gh_A11G1910, Gh_D03G0294, Gh_D1 1G2065, Gh_A11G0455, Gh_D05G1444 and Gh_D02G1 371) were significantly up-regulated in HyPRP1 VIGS cotton seedlings inoculated with V. dahliae compared to the control. Additionally, reactive oxygen species (ROS) have been shown to be associated with cotton  Properties of pGbHyPRP1. a Histochemical analysis of GUS activity in Arabidopsis expressing the pGbHyPRP1-GUS chimeric gene. GUS activity was present in whole 7-day-old seedlings. b Location of cis-regulatory elements involved in hormone-and elicitor-responsive elements found in pGbHyPRP1. The conserved fungus-responsive elements ELI-box3 (− 216) were present in the promoter region of WRKY in the Populus, while the EIRE (elicitor responsive element) (−879) are the binding site of WRKY and required for elicitor responsiveness in the promoter of PR (pathogenesis-related protein) in the parsley [64,65]. TC-rich repeat elements (− 502) are putatively involved in plant defense and stress response. Furthermore, ABRE (ABA-responsive element) (− 292), TCA-element and ERE (ethylene responsive element) (− 1128) have been reported in the upstream region of many genes that show regulated expression in response to ABA, SA and Et, respectively [26][27][28]. c Expression profiles of GbHyPRP1 in leaves of 15-day-old Pima90-53 seedlings subjected to SA, ABA, JA and ET. Control indicates water treatment. Three plants at each time point were sampled and analysed. Values are presented as the means ± SE in three independent experiments. Asterisks represent significant differences with respect to the control as determined by Tukey's multiple comparisons test (*P < 0.05, ** P < 0.01; *** P < 0.001; ns = not significant) resistance to fungal pathogens [18]. Here, we found 43 DEGs associated with the production and scavenging of H 2 O 2 and O 2 − , which allows regulation of dynamic changes in ROS levels (Table 3). Thus, our results suggest that HyPRP1 as a cell wall structural protein might have a potential role in cotton resistance to V. dahliae infection through remodeling of the cotton cell wall and ROS production.

Silencing of HyPRP1 causes a drastic increase in cell wall thickness and the lignin content
To further determine the effect of silenced HyPRP1 on the cell wall, the cotton stems were histologically examined. 10-day-old seedlings with two fully expanded cotyledons were used for VIGS. After two weeks later HyPRP1-silenced plants were inoculated with V. dahliae. At 14 dpi, cross sections of the basal part of stems revealed that the thickness of interfascicular fiber walls increased obviously in the HyPRP1-silenced plants compared with the control (Fig. 6a and a′). Moreover, we examined vessel walls in ultra-thin sections using transmission electron microscopy, which revealed a dramatic reduction in thickness ( Fig. 6b and b′). These results demonstrated that silencing of HyPRP1 does not affect the development of the cell wall under a pathogen-free condition, but repression of HyPRP1 expression significantly enhances cell wall thickening in response to V. dahliae. Further, lignin contents were estimated in cell wall residues. The data showed that HyPRP1-silenced plants had higher lignin content compared with the control at 0 dpi, 7 dpi and 14 dpi (Fig. 7a-c). Accordingly, the autofluorescence of lignified cell walls in the xylem and vascular bundles of HyPRP1-silenced plants was more intense and covered a greater area compared with the control (Fig. 7d-i).
Silencing of HyPRP1 enhanced ROS accumulation in root tips of cotton infected with V. dahliae We evaluated whether HyPRP1 silencing would result in cotton generating more ROS in response to V. dahliae infection. DAB and NBT staining showed that HyPRP1 VIGS cotton had significantly higher levels of O 2 − and H 2 O 2 compared to the control under V. dahliae infection ( Fig. 8a and b). In particular, VIGS plants not only showed darker staining approximately one millimeter from the tip, but staining was also detected in the upper part of the root ( Fig. 8a and b). DCFH-DA staining further confirmed enhanced ROS levels in HyPRP1 VIGS cotton roots compared to the control (Fig. 8c). These results suggest that HyPRP1 is involved in ROS production in cotton infected with V. dahliae.

Discussion
HyPRP1 is a negative regulator in cotton resistance to Verticillium wilt The roles of HyPRPs in response to multiple abiotic and biotic factors such as cold, salinity, drought and pathogens have been inferred primarily from their expression profiles [3,8,[19][20][21][22]. Remarkably, all of these HyPRP genes have been reported to be up-regulated by abiotic factors and down-regulated upon infection with pathogens, e.g., HyPRP1 in C. annuum and N. benthamiana following P. capsici infection [3] and PvPRP1 in cell cultures of P. vulgaris treated with an elicitor [8]. Similarly, in our study, transcriptional suppression of HyPRP1 in roots of both resistant and susceptible cotton was also detected after V. dahliae infection (Figs. 1d and 2b). This indicates that HyPRP1 is involved in the interaction between cotton and Verticillium. Further, based on loss-and gain-of function studies comprising VIGS and over-expression in cotton and Arabidopsis plants, respectively, we showed that HyPRP1 was negatively correlated with resistance to V. dahliae (Figs. 3 and 4). Thus, it is reasonable to speculate that HyPRP1 functions in the process of cotton resistance to Verticillium wilt as an important negative regulator.  HyPRP1 is potentially down-regulated by SA in cotton resistance against V. dahliae Upon pathogen attack, plants can rapidly initiate an immune response that is regulated by specific phytohormones, which vary greatly in composition, quantity and timing [23,24]. Through analysis of differential gene expression and transcription profiling of cotton inoculated with V. dahliae, SA-, ABA-, JA-and ET-mediated signaling pathways have been proven to contribute to V. dahliae resistance [14,25]. ABRE (ABA-responsive element)  (− 292), TCA-element and ERE (ethylene responsive element) (− 1128) have been reported in the upstream region of many genes that showed regulated expression in response to ABA, SA and ET, respectively [26][27][28]. In this paper, these potential cis-regulatory elements were also found in pGbHyPRP1 (Fig. 5b). Furthermore, the expression profiles of GbHyPRP1 following treatment with these phytohormones were examined. The expression of GbHyPRP1 was significantly down-regulated by SA but up-regulated by ABA, JA and ET (Fig. 5c). Activation of complicated and concerted phytohormone signaling networks is an important regulatory mechanism of immunity employed by plants. In many cases, these hormones interact antagonistically or synergistically with each other [29]. Generally, pathogens that require a living host (biotrophs) are implicated in SA-mediated defense responses, whereas pathogens that kill the host and feed on the contents (necrotrophs) are associated with JA/ET-mediated defenses [23,30]. Interestingly, however, V. dahliae is a hemibiotrophic phytopathogenic fungus. Thus, we infer that HyPRP1 taking part in cotton resistance to V. dahliae is probably regulated by a complex phytohormone signaling network. The HyPRP1 transcript levels showed quite similar down-regulation in SA-treated and V. dahliae-inoculated cotton seedlings (Figs. 1d, 2b and 5c). Additionally, V. dahliae infection significantly increased SA levels in G. thurberi [31], G. hirsutum and G. barbadense seedlings (our unpublished data). Thus, we speculate that HyPRP1 may be mainly and negatively regulated by SA signaling in cotton resistance against V. dahliae.

HyPRP1 participates in complex interactions within the cell wall polymer network in cotton infected with V. dahliae
Cell wall proteins are essential constituents of plant cell walls and are involved in modification of the cell wall structure [32]. Arabinogalactan protein 31 (AGP31), a remarkable plant cell-wall protein, comprises an SP, a short AGP domain, an His-stretch, a PRD and a PAC (PRP-AGP containing Cys) domain [33][34][35]. Arabidopsis AGP31 is able to bind methylesterified polygalacturonic acid, possibly through its His-stretch, and to interact with itself in vitro through its PAC domain [34]. Similarly, cotton HyPRP1 also contains an SP, a basic histidine-rich domain embedded within the PRD and a cysteine-containing domain embedded within the Pollen Ole e I domain ( Fig. 1a and Additional file 1: Figure S1). The cell wall, a physical barrier that pathogens need to breach to colonize the host plant, is typically reinforced with the phenolic polymer lignin [36,37]. Lignin is believed to play a critical role in the resistance of cotton to V. dahliae [13]. By comparative transcriptome analysis of HyPRP1-silenced cotton plants and the control inoculated with V. dahliae, 79 DEGs potentially involved in cell wall biogenesis were identified ( Table 2). Of these genes, Gh_D13G1445 was described as a Mannose-1-phosphate guanylyltransferase catalyzing the production of GDP-mannose. Arabidopsis cyt1 (encodes mannose-1-phosphate guanylyltransferase) mutant cause changes in the cell wall composition, such as dramatic decrease in cellulose content [38]. UDP-glycosyltransferases (UGTs) can influence the resistance of plants to infection by pathogenic microorganisms through regulating the glycosylation of phenylpropanoid and phenylpropanoid-de rived compounds, which are essential for the synthesis of lignin [39]. The xyloglucan endotransglucosylase/hydrolases (XTHs), specifically hydrolyzing xyloglucan as a substrate, are considered to be involved in the construction and restructuring of xyloglucan cross-link in plant cell wall [40]. Calmodulin, a highly conserved Ca 2+ -binding protein, acts as an intermediary connecting Ca 2+ signals involved in plant defence reactions [41]. Most of DEGs mentioned above have higher transcript levels in HyPRP1-silenced plants, suggesting that HyPRP1 is a negative regulator of cell wall-related genes. Further, we observed a dramatic thickening of interfascicular fiber walls and vessel walls (Fig. 6) and an increase in lignin (Fig. 7) in the HyPRP1-silenced cotton plants compared with the control after inoculation with V. dahliae. On the other hand, some DEGs were down-regulated, which seems to be a negative impact on cell-wall thickening and lignin accumulation. For example, phenylalanine ammonia-lyase (PAL) is the first committed enzyme in the phenylpropanoid biosynthesis, which engenders a variety of precursors of important secondary metabolites, mainly including flavonol glycosides and lignin. In our study, PAL2 (Gh_D06G0758) was shown to be significantly downregulated. We inferred that PAL transcription was feedback regulated by particular biosynthetic intermediates [42], which means that the expression of PAL may not be always positive correlation with lignin accumulation. Although the roles of many cell wall proteins have been studied broadly, the knowledge on the interaction between components is lacking. Thus, we speculated that HyPRP1 participates in complex interactions within the cell wall polymer network in cotton infected with V. dahliae.

HyPRP1 contributes to Veticillum defense by enhancing ROS accumulation
ROS have been studied extensively for their roles in interactions between plants and foliar pathogens. Little is known about ROS synthesis and function in defense reactions of the root, but ROS are consistently observed to accumulate in a plant after the perception of pathogens [18]. In the Verticillium-cotton interaction, the generation of H 2 O 2 was observed in the roots of cotton infected with V. dahliae [43]. Moreover, transgenic tomato plants expressing the Ve resistance gene accumulated  [44]. Likewise, transgenic cotton plants expressing a fungal endochitinase gene were more resistant and accumulated ROS faster than the control following pathogen inoculation [45].
These results indicate that cotton plants infected with V. dahliae are accompanied by increased ROS accumulation. In our study, silencing of HyPRP1 markedly enhanced ROS accumulation in the root tips (Fig. 8). Therefore, we reasonably inferred that cotton negatively modulates HyPRP1 transcription to generate ROS used for defense against Veticillum. Further studies are needed to elucidate the roles of ROS in cotton resistance to Verticillium. However, we suggest that ROS perform three possible functions: (i) ROS are primary immune signaling molecules [46]; (ii) ROS mediate cell wall modifications [47]; and (iii) ROS are important modulators that play a role in defense-related protein post-trans lational modifications [48].

Conclusions
Based on our research and existing developments on how plants resist Verticillium wilt, we propose a model of HyPRP1-mediated cotton defense against V. dahliae. Upon V. dahliae attack, recognition by cotton plants results in the activation of immune responses, including the production of a specific combination of the signals Fig. 8 Comparison of ROS accumulation in root tips of HyPRP1-silenced cotton seedlings inoculated with V. dahliae. a Representative microphotographs of NBT staining. The roots were inoculated with NBT for 40 min and were then subjected to microscopic observation to determine O 2 − production. Scale bar = 1 mm. b Representative microphotographs of DAB staining. The roots were placed in DAB for 10 h and were then subjected to microscopic observation to determine H 2 O 2 production. Scale bar = 1 mm. c Representative confocal images of DCFH-DA staining. The roots were placed in DCFH-DA for 30 min and were then examined to determine ROS production. Scale bar = 100 μm. At each time point of each treatment, nine roots from three separate plants were examined. Data represent the means ± SE from three biological replicates; asterisks indicate statistically significant differences according to Tukey's multiple comparisons test (*P < 0.05, **P < 0.01, ***P < 0.001). 10-day-old seedlings with two fully expanded cotyledons were used for VIGS. After two weeks later HyPRP1-silenced plants were inoculated with V. dahliae such as SA and JA, which have been proved to be involved in the plant-V. dahliae interaction [14,25]. The expression of HyPRP1 is most likely and mainly down -regulated by SA signaling. A significant reduction in HyPRP1 may affect the cell wall polymer network, including an increase in the thickness of the cell wall and the content of lignin required to prevent V. dahliae infection. Alternatively, down-regulation of HyPRP1 obviously enhances the accumulation of ROS, which could mediate the establishment of a cotton defensive response to V. dahliae.

Plant materials and growth conditions
The cotton seeds G. barbadense cv. Pima90-53, G. hirsutum cv. CCRI8 and ND601 were preserved at the North China Key Laboratory for Crop Germplasm Resources of Education Ministry, Hebei Agricultural University, Baoding, China. For transcriptional analysis of HyPRP1 in different tissues and response to V. dahliae by qPCR, two-week-old cotton seedlings were cultivated on Murashige and Skoog (MS) medium and were inoculated with V. dahliae as described by Zhang et al. [14]. For VIGS and the hormone treatment experiment, seeds were sterilized in 20% (V/V) commercial bleach (the final concentration of sodium hypochlorite was approximately 1%) for 20 min followed by washing four times with distilled water. Seeds were soaked in distilled water for 2 days and then germinated on wet towels for another 2 days at 25°C. Germinant seeds were transferred to pots containing commercially sterilized soil (a mixture of soil, peat, and composted pine bark) and covered with a plastic dome in a growth room at 25°C under a 14-h light/10-h dark cycle. The A. thaliana accession Columbia was grown in commercially sterilized soil at 22°C, 70% relative humidity, and~150 μE m − 2 s − 1 under a 9-h photoperiod.

V. dahliae cultivation
A highly aggressive defoliating fungus, V. dahliae strain Linxi2-1, was isolated from a symptomatic upland cotton plants growing in agricultural fields near Linxi, Hebei Province, China, and preserved in North China Key Laboratory for Crop Germplasm Resources of Education Ministry, Hebei Agricultural University. V. dahliae was cultivated on potato dextrose agar (PDA) plates for 10 d and then inoculated into Czapek's broth on a shaker at 150 rpm for 1 week at 25°C in the dark. Spores were harvested by filtration through folded Fisherbrand™ lens paper and were resuspended in sterile distilled water to a specific density.
Cloning of HyPRP1 and the GbHyPRP1 promoter region GbHyPRP1 was identified from a full-length cDNA library of G. barbadense Pima90-53 [14]. The full-length cDNA of HyPRP1 from other upland cotton cultivars was obtained by homology-based cloning. The genomic walking method was performed to amplify its 5′ flanking (promoter) region using the Genome Walking Kit (Takara, Dalian, China) according to the manufacturer's instructions. Nested sequence-specific primers designed on the basis of the known GbHyPRP1 gene sequence and shorter arbitrary degenerate primers were chemically synthesized or provided by the kit. The HyPRP1 protein sequences were identified using NCBI Web BLAST (http://www.ncbi.nlm.nih.gov/) and were a ligned using the Clustal W program (http://www.clustal.org/). The promoter sequence was analyzed using the software programs PlantCARE and PLACE to define putative cis-elements or binding sites for transcription factors [49,50].

qPCR and semi-RT-qPCR analysis
Total RNA was extracted using TRIzol® reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Each treatment, imposed on three pooled root, stem or leaf samples, was repeated at least three times in all experiments. RNA was quantified using a NanoDrop™ 1000 Spectrophotomete (Thermo Fisher Scientific). Subsequently, first-strand cDNA was synthesized from 1 μg of total RNA using the PrimeScript™ RT Reagent Kit with gDNA Eraser (TaKaRa, Dalian, China). qPCR was performed using a CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) and SYBR® Green reagent (TaKaRa, Dalian, China) as the reporter dye. Data were collected using CFX Manager™ software (Bio-Rad, Hercules, CA, USA). Target gene relative expression was normalized using PP2A1 (catalytic subunit of protein phosphatase 2A) [51]. For semi-RT-qPCR, the reactions were run with a denaturation step of 95°C for 5 min, followed by 25 cycles of 94°C for 1 min, 55°C for 30 s, 72°C for 1 min, with a final extension at 72°C for 10 min. PCR products were electrophoresed on 1% agarose gels.

Subcellular localization analysis of transiently expressed fusion proteins
For subcellular localization studies, the GbHyPRP1 ORF was amplified using PCR. The resulting product was inserted into the vector pDONR™207. Subsequently, the fragment was recombined into the destination vector pK7WGF2 [52] using L/R-Clonase. The construct was verified by sequencing and was transferred to the Agrobacterium tumefaciens GV3101 strain using the freeze/ thaw method [53]. Transient transformation of tobacco leaf epidermal cells was performed as described in [54].
Localization of fluorescent proteins was monitored 3 days after infiltration using a confocal laser scanning microscope (FluoView FV1000; Olympus). A pCAMBIA derivative (pCamE) carrying a cauliflower mosaic virus 35S-driven GFP was used as the control [55].

Hormone treatments
Cotton seedlings at the 2-cotyledon stage were sprayed with 100 μM SA, ABA, JA or ET and were covered with plastic bags to maintain 100% humidity. Cotyledon tissues were collected from hormone-treated plants at 6, 12, 24, 36 and 48 hps, immediately frozen in liquid nitrogen and then stored at − 80°C until RNA extraction. The control seedlings were sprayed with distilled water.

Generation of transgenic Arabidopsis lines
The coding sequence of GbHyPRP1 was amplified from G. barbadense Pima90-53 cDNA and cloned into a pBI121 vector. A genomic GbHyPRP1 upstream fragment was amplified and cloned into the pBI121 vector containing beta-glucuronidase (GUS) gene coding sequences, where the 35S promoter region was excised by digestion using PstI and BamHI restriction enzymes and was replaced with the sequence of the GbHyPRP1 upstream fragment. The recombinant plasmid was transformed into A. thaliana Columbia wild type (WT) plants through Agrobacterium tumefaciens strain G V3101-mediated plant transformation using the floral dip method [56]. Primary transformants were selected for survival on ½ MS medium with 50 μg ml − 1 kanamycin. Homozygous plants were isolated using gene-specific primers. Semi-RT-qPCR was then performed on cDNA from overexpression lines to confirm the status of transcription using the full-length GbHyPRP1 primers. Expression was normalized to the expression of actin. T3 transgenic A. thaliana plants carrying the GbHyPRP1 promoter were used in histochemical assays for GUS staining as described by [57].

VIGS assays in cotton
pTRV1 and pTRV2 from the Yule Liu research group at Tsinghua University (China) were used for the VIGS assays [58]. HyPRP1 fragments (364 bp) were amplified (Additional file 5: Table S1) and inserted into the pTRV2 vector to generate the derivative pTRV-HyPRP1, which were transformed into A. tumefaciens strain GV3101. An Agrobacterium-mediated VIGS assay in cotton was performed as previously described [59]. Cotton seedings with two fully expanded cotyledons were utilized. At this stage, the true leaves had not yet emerged. The cloroplastos alterados 1 gene (CLA1) was used as a marker to monitor the silencing efficiency. The seedlings injected with Agrobacterium cultures harboring the pTRV1 and pTRV2 (empty vector) were used as control. VIGS assays were repeated at least three times using more than 30 plants from each treatment per repeat. Twenty-four-day-old cotton seedlings were subjected to V. dahliae inoculation by root dipping in a spore suspension (10 7 spores ml − 1 ) for 2 min and were then returned to their original pots. Four-week-old Arabidopsis plants were infected with V. dahliae by soil drenching using a 10-ml conidial suspension (10 6 spores ml − 1 ) per pot (80 ml). Control plants were inoculated with distilled water in the same way. Cotton and Arabidopsis symptoms and disease index (DI) were scored at 15 dpi. The DI was calculated based on five disease grades as described previously [60].

Primers
All primers used in this paper are listed in Additional file 5: Table S1.

RNA-Seq library construction and analysis of transcriptome sequencing data
Two independent sets from both control and HyPRP1silenced plants were sampled to generate two biological replicas. For each sample, the first true leaves from at least six plants were collected at 12 hpi and pooled to minimize plant-to-plant variation. Then, total RNA was extracted using TRIzol® reagent following the manufacturer's instructions, after which genomic DNA was removed using DNase I (Invitrogen, Carlsbad, CA, USA). RNA purity was checked using a NanoPhotometer® spectrophotometer (IMPLEN, CA, USA), and concentration was measured in a Qubit® 2.0 Flurometer using the Qubit® RNA Assay Kit (Life Technologies, CA, USA). RNA integrity was assessed using the RNA Nano 6000 Assay Kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). Transcriptome sequencing was performed using an Illumina HiSeqTM 2000 sequencer (Illumina, San Diego, CA, USA).
Sequence tag preprocessing was performed according to a previously described protocol [61]. Paired-end clean reads were aligned to the reference genome of G. hirsutum L. acc. TM-1 (https://www.cottongen.org/) using Bowtie v2.0.6 and TopHat v2.0.9. Differential expression analysis of HyPRP1-silenced and unsilenced (mock) groups (two biological replicates per group) was performed using the DESeq R package (1.10.1). The resulting P-values were adjusted using Benjamini and Hochberg's approach for controlling the false discovery rate. Genes with an adjusted P < 0.05 according to DESeq were assigned as differentially expressed.

Examination of cell walls
Stems were cut into pieces (1 mm 3 ) and fixed in 2.5% (wt/vol) glutaraldehyde in 0.1 M phosphate buffer, postfixed in 1% osmium tetroxide, and then embedded in Spurr's resin. Semi-thin sections (1 μm), prepared using a Leica UC6 (Leica, Illinois, USA), were hot-stained with a 1% toluidine blue water solution. Images were captured using a digital camera (Digital sight DS-L1, Nikon, Japan) attached to a Olympus BX51 microscope (Tokyo, Japan). Ultra-thin sections (70 nm thick), prepared using a Leica Ultracut R ultramicrotome (Leica, Illinois, USA), were stained with uranyl acetate and lead citrate and examined under a JEM-1230 electron microscope (JEOL Ltd., Tokyo, Japan). Approximately 60 cells were measured per sample to determine the thickness of the cell walls.

Lignin extraction and quantification
Stem samples were collected from both HyPRP1-silen ced and control plants at 0, 7 and 14 dpi. Three samples were collected per plant at different time intervals; the samples were crushed immediately in liquid nitrogen and freeze-dried. The extraction and quantification of lignin in cell walls were performed according to the method of Schenk et al., 2014 [62].

Identifiction of lignin deposition in cell walls using histochemistry and autofluorescence
Standard transverse free-hand sections were collected from the stems of cotton seedlings (5 cm from the stem base). Lignin autofluorescence was imaged using a BX51 microscope with both UV excitation and brightfield modes.

Biological analysis of ROS accumulation
Since ROS include various forms of reduced and chemically reactive molecules such as hydrogen peroxide (H 2 O 2 ) and the superoxide oxygen anion (O 2 − ), we used NBT and DAB dyes, as well as the ROS-reactive fluorescent probe DCFH-DA, to detect the accumulation of ROS. Yellow, water-soluble NBT can be reduced by O 2 − to blue, water-insoluble formazan. DAB polymerizes at sites of peroxidase activity into a reddish brown polymer in the presence of H 2 O 2 . Fluorescence-free DCFH-DA can cross the cell membrane and, after hydrolyzation, be oxidized by ROS to form highly fluorescent DCF [18]. ROS (H 2 O 2 ) accumulation was visualized by incubating intact cotton roots in 5 μM DCFH-DA (dichlorofluorescein diacetate; Beyotime, Nanjing, China) for 20 min at 37°C in the dark or in 10 mM K-citrate buffer (pH 6.0) containing 2.5 mM DAB (3,3′-diaminobenzidine) for 12 h, as described by [63]. ROS (O 2 − ) production was also visualized by incubating intact roots in 10 mM K-citrate buffer (pH 6.0) containing 0.5 mM NBT (nitrotetrazolium blue chloride) for 30 min at 37°C. An Olympus FV1000 laser confocal microscope with an excitation wavelength of 488 nm was used for the analysis of DCF fluorescence. For DAB and NBT staining, the roots were boiled in 95% ethanol for 10 min and then rinsed twice with 50% ethanol, after which images were captured under a light microscope (BX51; Olympus, Tokyo, Japan) using a digital camera (DP71; Olympus, Tokyo, Japan) and Image-Pro® Plus 6.0 software (Media Cybernetics, Rockville, MD, USA). Quantization of fluorescence and coloration intensity was performed using ImageJ software version 1.48 (National Institute of Mental Health, Bethesda, MD, USA).

Statistical analysis
All experiments were repeated at least three times for each determination. Statistical analysis of data was conducted with the software of GraphPad Prism® 6 (Graph Pad, San Diego, CA, USA). The P-value less than 0.05 was considered to be statistically significant.