Transcriptome analysis reveals the mechanism by which the biocontrol fungus Chaetomium globosum CEF-082 controls Verticillium wilt in cotton

Background: Verticillium wilt of cotton is a serious soil-borne disease that causes a substantial reduction in cotton yield. A previous study showed that the endophytic fungus Chaetomium globosum CEF-082 could control Verticillium wilt of cotton, but the molecular mechanism by which CEF-082 controls Verticillium wilt is still unknown. Results: To study the mechanism by which CEF-082 controls Verticillium wilt, the transcriptome of cotton seedlings pretreated with CEF-082 was sequenced. The results revealed 5638 DEGs 24 h post-inoculation with CEF-082, and 2921 and 2153 DEGs 12 and 48 h post-inoculation with Verticillium dahliae , respectively. At twenty-four hours post-inoculation with CEF-082, KEGG enrichment analysis indicated that the DEGs were mainly enriched in plant-pathogen interaction, MAPK signalling pathway-plant, avonoid biosynthesis, and phenylpropanoid biosynthesis. There were 1209 DEGs specically induced after inoculation with CEF-082 and V. dahliae . GO enrichment indicated that these DEGs were mainly enriched in the terms reactive oxygen species metabolic process, hydrogen peroxide metabolic process, defence response, superoxide dismutase activity, and antioxidant activity. Here, many genes, such as ERF, CNGC, FLS2, MYB, GST and CML, were identied that regulate crucial points in defence-related pathways and that may contribute to V. dahliae resistance in cotton. These results provide a basis for the understanding of the molecular mechanism by which biocontrol fungi control Verticillium wilt. Conclusions: In this study, we found that CEF-082 could regulate multiple metabolic pathways in cotton. After treatment with Verticillium dahliae , the defence response of cotton plants pre-inoculated with CEF-082 was strengthened.

. K-165 induced the resistance of Arabidopsis thaliana to Verticillium wilt [9]. Bacillus subtilis DZSY21 reduced the disease severity of southern corn leaf blight and upregulated the expression levelof PDF1.2 in DZSY21-treated plants [10]. The preinoculation of cauli ower with Verticillium Vt305 reduced symptom development and the colonization of plant tissues by V. longisporum [11]. However, the mechanism of the biological control of plant diseases remains unclear.
It has been reported that cotton plants infected with V. dahliae induce a series of immune reactions. In recent years, transcriptomic studies of the defence responses of plants infected with V. dahliae have becoming increasing common, and several signal transduction pathways and key genes have been identi ed, including plant hormone signal transduction, plant-pathogen interaction, and phenylpropanoidrelated and ubiquitin-mediated signals in cotton; additionally, these studies have investigated the key regulatory gene families, such as receptor-like protein kinases (RLKs), WRKY transcription factors and cytochrome P450s (CYPs) [3]. PAL, 4CL, CAD, CCoAOMT, and COMT in the phenylalanine metabolism pathway have been shown to be upregulated in sea-island cotton [2]; 401 transcription factors, mainly in the MYB, bHLH, AP2-EREBP, NAC, and WRKY families, have been shown to be up-or downregulated by V. dahliae in A. thaliana [12]; and CNGC, RBOH, FLS2, JAZ, MYC2, NPR1 and TGA have been shown to be induced by V. dahliae in sun ower [13]. However, there are few studies on the transcriptome level in plants induced by biocontrol fungi or induced by biocontrol fungi and V. dahliae at the same time.
In previous studies, we found that the endophytic fungus CEF-082 isolated from upland cotton plants could control Verticillium wilt in cotton. However, the molecular mechanism of biocontrol is unknown.
Therefore, the purpose of this study is to reveal the molecular mechanism by which CEF-082 controls Verticillium wilt in cotton via RNA sequence analysis to provide a basis for the understanding of the biological control of plant diseases.

Fungal strain culture
Endophytic Chaetomium globosum CEF-082 of cotton was cultured on potato dextrose agar (PDA) plates for 20 d. Spores were washed with sterile water and diluted to a 1×10 5 CFU/mL spore suspension. V. dahliae VD1070-2 was cultured on PDA for 7 d, inoculated into liquid Czapek-Dox medium [14], and cultured in the dark at 25°C and 150 rpm for 7 d. The mycelia were ltered out and diluted to a 1×10 7 CFU/mL spore suspension.

Cotton inoculation treatment
Jimian 11, a highly Verticillium wilt-susceptible upland cotton variety, was provided by Professor Heqin Zhu from State Key Laboratory of Cotton Biology, Institute of Cotton Research of Chinese Academy of Agricultural Sciences. It is a hybrid [(Jihan 4× Ke 4104) F_2 × 74Yu102]. The seeds were sterilized with 70% alcohol for 1 minute and with 1.05% sodium hypochlorite for 10 minutes and then washed with sterile water 5 times. The cotton seeds were planted in vermiculite and transferred to a plastic pot (25 cm×15 cm) containing 2000 mL of culture solution after emergence. The culture solution was prepared according to the methods of Zhang et al. [15], with some modi cations. In this study, 2 mM NaCl was used instead of 2.5 mM KCl, while the other 9 mineral nutrients were the same. Twenty plants were cultured per pot. The cotton plants were inoculated with the CEF-082 spore suspension by soaking the cotton roots in the spore suspension for 40 minutes prior to the rst true leaf attening, and water, instead of the CEF-082 spore suspension, was used as the control group. Zero, 6 and 24 h later, leaf samples were taken, and 24 h was considered 0 h before inoculated with V. dahliae. After that, the same method was used to inoculate V. dahliae VD1070-2 into the treatment group and the control group. Then, leaf samples were collected at 12 h, 1 d, 2 d, 3 d, 5 d and 7 d. Determination of the hydrogen peroxide (H 2 O 2 ) content H 2 O 2 content was estimated according to Anket Sharma et al. [16] with minor modi cations.
Approximately 0.1 g of cotton leaves was weighed and added to 1 mL of acetone for ice bath homogenization. Then, the samples were centrifuged at 8000 g and 4°C for 10 minutes, and the supernatant was collected. Then, 25 μL of 20% titanium chloride in concentrated HCl and 200 μL of ammonia solution (17 M) were added. The precipitate was then washed 3 times with acetone. The washed precipitates were dissolved in 1.5 mL of H 2 SO 4 (2 N), and the absorbance was read at 415 nm.
Control effect of biocontrol fungus CEF-082 on Verticillium wilt of cotton The above hydroponic seedlings were investigated at 14 d post-inoculation (dpi) with VD1070-2. The method of investigation was consistent with that of Zhu et al. [17].

RNA-seq
A polysaccharide polyphenol RNA extraction kit (TianGen) was used to extract RNA from cotton leaves.
Electrophoresis was performed, and Drop one was used to detect the concentration and quality of RNA. Transcriptome sequencing was performed for the 24 h (0 h (T0h, C0h)), 12 h (T12h, C12h) and 48 h (T48h, C48h) samples. Three replicates were performed, and there were 18 samples. The construction of the DNA library and sequencing were performed by BGI company. The raw reads obtained from sequencing were ltered to obtain clean reads, which were spliced and compared to the reference genome.
Screening and analysis of differentially expressed genes (DEGs) To improve the accuracy of the identi cation of DEGs, we de ned DEGs with a fold change > = 2 and Qvalue < 0.001 as signi cant. Kyoto Encylopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analyses of these genes were carried out.

Principal component analysis (PCA)
The princomp function in R software was used for PCA analysis, and the ggplot2 package in R software was used to draw gures.
Quantitative reverse-transcription-PCR (qRT-PCR) analysis Some genes were selected for RT-PCR to calculate whether the trend of gene expression was consistent with the transcriptome sequencing results. RNA was extracted from sample leaves and re-transcribed into cDNA. Primers were obtained from the upland cotton gene uorescence quantitative speci c primer database (https://biodb.swu.edu.cn/qprimerdb/) (Table S1), and gene expression was calculated by 2 -ΔCt .

Control effect of CEF-082 on Verticillium wilt of cotton and the H 2 O 2 content
The disease index was 18.61 in the control group (water+ V. dahliae) and 7.62 in the treatment group (CEF-082+ V. dahliae) 14 d after V. dahliae inoculation (Fig. 1A). The results showed that CEF-082 could control Verticillium wilt of cotton, and the control effect was 59.1% (Fig. 1C).
The H 2 O 2 content in the treatment group was higher than that in the control group throughout the majority of the duration of the experiment and lower than that in the control group at 5 dpi with V. dahliae. The H 2 O 2 content in the treatment group was highest at 2 dpi (12.80 μmol/g), while the H 2 O 2 content in the control group was highest at 1 dpi (10.38 μmol/g). The changes in the two groups were similar and were stable 5 d later (Fig. 1B).

PCA
The minimum correlation between the three replicates was 95.5% (Fig. S1). According to the results of the PCA, at 0 h and 12 h, the gene expression difference between the control group and the treatment group was relatively small, but there was a big difference at 48 h (Fig. 2). These results revealed a gene expression difference between the CEF-082-treated group and the non-CEF-082-treated group.
qRT-PCR RNA was extracted from cotton transcriptome samples and reverse transcribed into cDNA. Twelve DEGs were selected. The gene expression levels in the control and treatment groups were compared by qRT-PCR. The results showed that nine of the 12 genes were upregulated, which was consistent with the results of their upregulated expression in the transcriptome, while three genes were downregulated, which was inconsistent with the expression of the transcriptome genes, namely, Gh_D12G2793, Gh_D08G2484 and Gh_D05G3615 (Fig. 3). In addition, the level of upregulation of 5 genes in the qRT-PCR data was lower than that in the RNA-seq data. The qRT-PCR data were consistent with the transcriptome data up to 75%.

Functional annotation and enrichment analysis of the DEGs
The average clean reads of the 18 samples was 62.08. The lowest Q20 value of the clean reads was 97.93, and the lowest Q30 value was 90.06 (Table S2). A total of 47183 new transcripts were found, of which 7288 belonged to new protein-coding genes (Table S3).
There were 3480 upregulated and 2158 downregulated DEGs at 0 h, 1716 upregulated and 1205 downregulated DEGs at 12 h, and 1524 upregulated and 629 downregulated DEGs at 48 h (Fig. 4). The highest number of DEGs were identi ed after inoculation with CEF-082 for 24 h. After inoculation with V. dahliae, the number of DEGs gradually decreased.
The GO enrichment analysis revealed that the 5638 genes were mainly enriched in 86 terms, including the intrinsic component of membrane, integral component of membrane, membrane part, membrane, catalytic activity, response to biotic stimulus, cell wall, oxidoreductase activity, defence response, response to stimulus, response to stress, and response to fungus (Q-value <0.001), and the rst 15 terms are listed in Table 2. Of the 16 genes in the response to fungus term, 15 were upregulated and 1 was downregulated. The GO classi cation showed that there were 18, 14 and 12 terms in biological process, cellular component and molecular function, respectively, and the KEGG classi cation indicated that the DEGs mainly belonged to the metabolism pathway (2856 DEGs).

DEGs speci cally induced by CEF-082
A total of 1209 speci c DEGs were identi ed at 12 h and 48 h after removing all of the shared DEGs in the three stages in the control group, and the cluster thermogram showed the expression patterns of these genes at different stages (Fig. 7). KEGG classi cation showed that these DEGs mainly belonged to metabolism (672 DEGs) and were signi cantly enriched in 5 KEGG pathways, including avonoid biosynthesis, indole alkaloid biosynthesis, MAPK signalling pathway-plant, plant-pathogen interaction, and phenylpropanoid biosynthesis (Table 4). GO classi cation showed that there were 14, 12 and 9 terms in the biological process, cellular component and molecular function, respectively. GO enrichment indicated that these DEGs were enriched in reactive oxygen species metabolic process (14 DEGs), hydrogen peroxide metabolic process (12 DEGs), hydrogen peroxide catabolic process (12 DEGs), defence response (31 DEGs), superoxide dismutase activity (5 DEGs), antioxidant activity (19 DEGs), oxidoreductase activity, acting on superoxide radicals as acceptor (5 DEGs), cofactor binding (75 DEGs) and DNA binding (121 DEGs) (Fig. S2).
At 12 h and 48 h, 96 shared DEGs were obtained by eliminating the shared DEGs in the CK at different stages (Fig. 8). KEGG analysis of the 96 DEGs indicated that they were mainly enriched in glutathione metabolism and avonoid biosynthesis (Table 5). GO analysis showed that the DEGs were enriched in superoxide dismutase activity, oxidoreductase activity, acting on superoxide radicals as acceptors, and antioxidant activity terms. Of the 96 DEGs, there were 9 transcription factors (TFs) and 20 plant resistance genes (PRGs) ( Table S4).

Protein interaction network induced by CEF-082
A protein-protein interaction network (Fig. 10) was obtained by using the 96 DEGs shared at 12 and 48 h and genes interacting with them in cotton. Six hub genes were obtained: Gh_A05G1020, Gh_D09G0858, BGI_novel_G004376, Gh_A08G0125, Gh_D07G1197, and Gh_A05G3508. Among them, Gh_D07G1197 was enriched in the avonoid biosynthesis pathway.

Discussion
The number of DEGs identi ed at 12 h and 48 h was lower than that identi ed at 0 h. It has been suggested that the number of DEGs decreased because both plants were infected with V. dahliae and began to respond defensively. For CEF-082 treatment and CEF-082+ V. dahliae treatment, DEGs were mainly enriched in 5 signalling pathways, plant-pathogen interaction, MAPK signalling pathway-plant, avonoid biosynthesis, phenylpropanoid biosynthesis, and glutathione metabolism. The pathways of plant-pathogen interaction and avonoid biosynthesis were also induced in sun ower infected with V. dahliae [13], and the results were also consistent with those of Tan [18], who reported that most DEGs in tomato were associated with phenylpropanoid metabolism and plant-pathogen interaction pathways. However, the glutathione metabolism pathway has rarely been reported in the transcriptome of cotton plants treated with V. dahliae.
DEGs related to ET, SA, JA, brassinosteroid (BR) and cytokinin were upregulated or downregulated upon V. dahliae infection in cotton [3]. In this study, we also found that DEGs in ABA, auxin and gibberellin were signi cantly induced not only after treatment with CEF-082 but also after inoculation with V. dahliae. The 8 plant hormones were also induced after infection with V. dahliae in sun ower [13]. The responses of the A. thaliana auxin receptors TIR1, AFB1 and AFB3 and auxin transporter AXR4 were impaired upon infection with V. dahliae [19]. Therefore, both CEF-082 and V. dahliae can induce changes in hormones.
Previously, it was shown that after plants were infected with pathogens, the FLS2 pattern recognition receptors recognized pathogens, and the hypersensitive response (HR) was activated through ROS, JA, WRKYs and the NO signalling pathways [20][21] and mediated by CNGC, RBOH, CaM/CML and FLS2 [22][23][24]. These results are consistent with the results from this study. In this study, 24 h after treatment with CEF-082, the DEGs of FLS2, Rboh, CDPK, CNGCs and GST in the plants were also upregulated or downregulated to varying degrees (Fig. 5). In addition, most of the genes coding peroxidase (POD), superoxide dismutase (SOD), and catalase (CAT) were also upregulated. These genes were related to the accumulation of ROS. Forty-eight hours after treatment with V. dahliae, the genes encoding CNGC, CaM/CML and FLS2 were upregulated. However, in this study, the NO signalling pathway was not induced.
Phenylpropane synthesis is related to cotton defence mechanisms [25], while avonoids are known were to buffer substantial stress-induced alterations in ROS homeostasis and to modulate the ROS-signalling cascade [26]. Plant CNGC subunits and CaM constitute a molecular switch that either opens or closes calcium channels [27]. Previous reports have shown that calcium-dependent CDPK4 and CDPK5 regulate ROS production by phosphorylating NADPH oxidase in potatoes [28]. ROS are important not only for signalling mechanisms for defence [29] but also for regulating programmed cell death via the establishment of the HR [30]. MAPK family members can improve resistance to Verticillium wilt in cotton [31]. In this study, 24 h after CEF-082 inoculation, certain signal transduction pathways may have been involved in the plant response to CEF-082 (Fig. 11). After inoculation with CEF-082, FLS2 recognized CEF-082, MAPK signal transduction was induced, and calcium channels were opened. Then, H 2 O 2 was produced, leading to ROS burst. Plant hormones were also induced, including ET, SA, JA, ABA, BR, auxin, gibberellin and cytokinin. The signalling pathways of avonoids and phenylpropane synthesis were also involved in this process. In addition, lignin synthesis was also induced after treatment with CEF-082 (Fig.  12). C4H and C3H were not induced in T0h-vs-C0h, T12h-vs-C12h, or T48h-vs-C48h but were induced in C12h-vs-C0h, which was similar to the results of Xu et al. [32], who indicated that C4H-1 and C4H-3 were upregulated after treatment with V. dahliae. Three days after inoculation with V. dahliae, lignin was detected, and the pith diameter of CEF-082 + V. dahliae-treated plants was slightly larger than that of water + V. dahliae-treated plants (Fig. S3). The defence response at T12h and T48h was similar to that at T0h, and only some key points induced were different in the pathways shown in Fig. 11 and Fig. 12. Thus, it is speculated that CEF-082 can control cotton Verticillium wilt because inoculation with CEF-082 can prime signalling pathways to defend against V. dahliae upon its infection.
When pathogens infect plants, they induce a series of defence responses. GST participates in plant defences and can remove ROS [34]. Plant GSTs can be subdivided into eight categories, phi, zeta, tau, theta, lambda, dehydroascorbate reductase (DHAR), elongation factor 1 gamma (EF1G) and tetrachlorohydroquinone dehalogenase (TCHQD) [35]. GSTF8 was used as a marker in early stress and defence responses [36], and salicylic acid, methyl jasmonate, ABA and H 2 O 2 can induce GST expression [37][38][39]. LrGSTU5 was obviously upregulated after treatment with Fusarium oxysporum [40], and the GST genes were also upregulated in G. barbadense treated with V. dahliae [41]. In this study, the GST genes were also signi cantly induced 24 h after treatment with CEF-082 (Fig. 5), and GST genes were upregulated in cotton treated with Water + V. dahliae. These results are consistent with those of Han et al.
and . Certain GST genes were also signi cantly induced in the treatment group but were not signi cantly induced in the control group after treatment with V. dahliae. The GST gene Gh_A09G1509 enhanced resistance against Verticillium wilt in tobacco [42]. Hence, we suggest that CEF-082 can induce speci c GST genes to protect cotton from V. dahliae.
V. dahliae can induce a defence response after it infects cotton [3]. In this study, susceptible cotton varieties were inoculated with the biocontrol fungus CEF-082 and V. dahliae, which also induced a series of defence responses. Compared with plants inoculated with water +V. dahliae, the plants inoculated with CEF-082 + V. dahliae had signi cantly upregulated or downregulated expression levels of resistancerelated genes. Therefore, it is speculated that the defence response was strengthened after inoculation with the biocontrol fungus CEF-082. In addition, we obtained 1209 speci c DEGs, which could not be induced in plants inoculated with water +V. dahliae, and GO enrichment showed that these genes were involved in the metabolic process of ROS. The disease resistance of cotton was enhanced after CEF-082 treatment, and thus, we inferred that these speci c DEGs might be genes related to plant disease resistance.

Conclusion
CEF-082 can induce defensive responses in cotton, and pretreated with CEF-082 with appropriate concentration 10 5 CFU/mL can improve the resistance of cotton (Jimian 11) to Verticillium wilt.
Transcriptome analysis revealed that genes in cotton leaves involved in ROS burst, Ca 2+ , lignin biosynthesis, avonoids and phenylpropane synthesis were signi cantly upregulated or downregulated. In this study, the transcriptome was used to study the expression of genes in cotton when CEF-82 and V. dahliae coexisted, which provided a basis for understanding the mechanism by which biocontrol fungi prevent Verticillium wilt in cotton.

Declarations
This work was supported by National Key Research and Development Program of China (2017YFD0201900) and National Natural Science Foundation of China (31901938).We thank all the foundation of economic support. The funding organizations provided the nancial support to the research projects, and involved in the design of the study.
Authors' contributions YZ, LZ, HZ and CT conceived the study. YZ and NY performed the experiments. YZ analysed the results and wrote the manuscript, with feedback from all authors. All authors have read and approved the manuscript.

Availability of data and materials
Most data supporting the results and conclusions are included in the article and additional les.

Ethics declarations
Ethics approval and consent to participate     Pathways with a Q-value < 0.05 are shown.