Genome-wide identification of cyclophilin genes in Gossypium hirsutum and functional characterization of a CYP with antifungal activity against Verticillium dahliae

Background Cyclophilins (CYPs), belonging to the peptidyl prolyl cis/trans isomerase (PPIase) superfamily, play important roles during plant responses to biotic and abiotic stresses. Results Here, a total of 79 CYPs were identified in the genome of Gossypium hirsutum. Of which, 65 GhCYPs only contained one cyclophilin type PPIase domain, others 14 GhCYPs contain additional domains. A number of cis-acting elements related to phytohormone signaling were predicated in the upstream of GhCYPs ORF. The expression analysis revealed that GhCYPs were induced in response to cold, hot, salt, PEG and Verticillium dahliae infection. In addition, the functional importance of GhCYP-3 in Verticillium wilt resistance was also presented in this study. GhCYP-3 showed both cytoplasmic and nuclear localization. Overexpression of GhCYP-3 in Arabidopsis significantly improved Verticillium wilt resistance of the plants. Recombinant GhCYP-3 displayed PPIase activity and evident inhibitory effects on V. dahliae in vitro. Moreover, the extracts from GhCYP-3 transgenic Arabidopsis displayed significantly inhibit activity to conidia germinating and hyphal growth of V. dahliae. Conclusions Our study identified the family members of cotton CYP genes using bioinformatics tools. Differential expression patterns of GhCYPs under various abiotic stress and V. dahliae infection conditions provide a comprehensive understanding of the biological functions of candidate genes. Moreover, GhCYP-3 involved in the resistance of cotton to V. dahliae infection presumably through antifungal activity. Electronic supplementary material The online version of this article (10.1186/s12870-019-1848-1) contains supplementary material, which is available to authorized users.

ROS (reactive oxygen species) regulation [7]. With the availability of whole genome sequencing, the identification and characterization of plant CYPs are carried out mainly in Arabidopsis thaliana (29 AtCYPs) [8], Oryza sativa (27 OsCYPs) [9] and Glycine max (62 GmCYPs) [10]. The majority of studies reveal the involvement of plant CYPs mostly in different types of abiotic stress. For example, Arabidopsis CyPs showed evidence of response to wounding [11]. Rice OsCYP19-4 showed over 10-fold upregulation in response to cold. Overexpressing of OsCYP19-4 could enhance rice plants cold-resistance with significantly increased tiller and spike numbers, and consequently enhanced grain weight [12]. Transgenic plants overexpressing OsCYP21-4 exhibited increased tolerance to salinity and hydrogen peroxide treatment [13]. Ectopic expression of pigeon pea (Cajanus cajan L.) CYP, CcCYP, in Arabidopsis exhibited high-level tolerance against drought, salinity and extreme temperatures [14]. Against biotic stress, especially against pathogen infection, only several plant CYPs have been studied in plant-pathogen system. Pepper cyclophilin (CACYP1) gene expression increased in response to Xanthomonas campestris pv. vesicatoria and Colletotrichum gloeosporioides [15]. Fungal infection with Fusariumsolani f. speumartii increased the level of Solanum tuberosusm CyP gene StCyP mRNA in tubers [16]. The expression of V. vinifera VviCyP was highly induced by Plasmopara viticola [17]. In cotton, a cyclophilin-like gene GhCyp1 was cloned from G. hirsutum. Overexpression of GhCyp1 in transgenic tobacco plants conferred higher tolerance to salt stress and Pseudomonas syringae pv. tabaci infection compared with control plants [18].
In 2015, the genome of G. hirsutum L. acc. Texas Marker-1 (TM-1) was sequenced, more than 70, 000 protein-coding genes were predicted (NAU version 1.1) [19,20]. Recently, an improved de novo-assembled genome for G. hirsutum L. acc. TM-1 were generated (NAU version 2.1) [21]. The genome-sequencing project facilitates the survey of all CYP genes in cotton. In the present study, the CYP gene family members in G. hirsutum and their expression patterns under various abiotic stresses and on V. dahliae infection were systematically investigated. Furthermore, the function of GhCYP-3 was analyzed to reveal its role in cotton resistance to V. dahliae infection. Our study will enlighten the novel insights into the function of CYP genes in plant against multivariate stress responses in the future and provide more candidate genes for resistance breeding in cotton.

Results
Up to 79 CYPs were identified in the genome of G. hirsutum TM-1 A local BLASTP search was performed with the Arabidopsis CYP proteins as query, which resulted in 79 CYP candidates from G. hirsutum NAU version 1.1, 74 CYPs from G. hirsutum JGI version 1.0 and 78 CYPs from G. hirsutum NAU version 2.1 (Table 1). These candidates were submitted to Pfam to confirm the existence of cyclophilin type PPIase domain (CLD, PF00160) and named GhCYP-1 to GhCYP-79. The characteristics of the individual CYP, including CDS length, protein length, molecular weight, and isoelectric point (pI) were presented in Table 1. The protein length varied from 69 amino acid (aa) residues (GhCYP-70) to 801 aa (GhCYP-55). The molecular weight ranged from 7.5 kDa (GhCYP-70) to 90.5 kDa (GhCYP-55), and the pI values ranged from 4.6 to 12.0. Most of the GhCYPs were expected to be in the cytoplasm. Also, some GhCYPs exhibited chloroplast, mitochondrial, nuclear and extracellular localization. Of the 79 GhCYPs, 65 only contained one CLD domain, but the remaining 14 GhCYPs contain additional domains, including tetratricopeptide-like repeats (TPR, PF00515, PF07719, PF13181, PF13414), Zinc finger (zf-CCHC, PF00098), RNA recognition motif (RRM, PF00076) and WD40 (PF00400) (Fig. 1).

Cis-elements potentially related to hormonal signal for GhCYPs
Here we surveyed the presence of cis-elements potentially related to the hormonal signal, in the -2 kb 5′ flanking region upstream to the start codon of these GhCYPs. In total six types of hormones related cis-elements in the promoters were predicted (Fig. 2). Of these GhCYPs, 66 GhCYPs had ethylene (ET) responsive element (ERE), 35 GhCYPs contained salicylic acid (SA) responsive element (TCA-element), 47 GhCYPs harbored abscisic acid (ABA) responsive element (ABRE), 38 GhCYPs possessed gibberellin (GA) responsive element (P-box; TATC-box), 47 GhCYPs contained methyl jasmonate (MeJA) responsive element (CGTCA-motif ), and 26 possessed auxin responsive element (TGA-box; AuxRR-core). In total, 217 cis-elements related to ET, 150 cis-elements related to MeJA, 132 cis-elements related to ABA, 57 cis-elements related to GA, 43 cis-elements related to SA and 34 cis-elements related to auxin were identified in all GhCYPs (Fig. 2). The enrichment of hormone-responsive cis-elements in the upstream of these GhCYPs suggests that they are likely to be involved in plant responses to various hormone signal pathways.

Expression patterns of GhCYPs under various abiotic stresses
Expression profiles of GhCYPs were examined in roots of cotton plants under four different abiotic stress conditions using high-throughput RNA-seq data (Additional file 2: Table S2). The transcripts with low Fragments Per Kilobase of exon per Million fragments mapped (FPKM) were probably false assembly. Therefore, in this study GhCYPs with FPKM >10 and present in at least two   Fig. 3. In the cold treatment, 14 GhCYPs were commonly down-regulated significantly apart from GhCYP-49 at 1 hct (hours after cold treated) and 3 hct and GhCYP-52 at 6 hct, which showed up-regulated significantly. At 1 hht (hours after hot treated), 22 of 26    and GhCYP-75 were down-regulated. Only the expression of GhCYP-38 showed be down-regulated at 3 hpt (hours after PEG treated) and 6 hpt, and then be upregulated at 12 hpt. These expression patterns suggest that CYPs undertake multiple functions to help the cotton counter various complex environmental challenges.
Expression profiles of GhCYPs under the stress of V. dahliae To gain a better understanding of the roles of CYP family genes in the cotton resistance against V. dahliae, the expression profiles were obtained by RNA-Seq. A total of 30 GhCYP genes showed differential expression at least one time point (hpi, hours post inoculation), including 19 upregulated genes and 11 downregulated genes (Fig. 4). Notably, levels of GhCYP-10, GhCYP-22, GhCYP-48 and GhCYP-59 were up-regulated at all five treatment time points. Additionally, the expression of GhCYP-11, GhCYP-27, GhCYP-37, GhCYP-45 and GhCYP-74 was down-regulated in most time points. These results revealed that GhCYPs were associated with the interaction between cotton and V. dahliae.

GhCYP-3 contains conserved amino acid residues and has PPIase activity
GhCYP-3, cloned from G. hirsutum cv. JM20, contains a single cyclophilin domain, 173 amino acid residues with a calculated molecular mass of 18.2 kDa and a pI of 8.34. Alignment with previously characterized Arabidopsis and human CYP (AtCYP19-1 and hCypA) revealed that GhCYP-3 contains seven conserved amino acid residues that critically affect PPIase activity (Fig. 5a). GhCYP-3 was found to be expressed in root, stem and leaf of unchallenged cotton plants with V. dalihae (Fig. 5b). Translation fusion of GhCYP-3 with GFP was constructed under the control of 35S promoter, and then transiently expressed in onion epidermal cells. Fluorescent imaging of the GhCYP-3-GFP bombarded onion epidermal cells showed both cytoplasm and nuclear localization (Fig. 5c). The GhCYP-3 ORF was cloned into pET-32a at SacI and BglII sites, thus the recombinant plasmid was constructed (Fig. 6a). SDS-PAGE and western blot analysis revealed that the fusion TrxA-6×His-S-tag-6×His (THS, 20.4 kDa; empty vector as control) and TrxA-6×His-Stag-GhCYP3 (THS-CYP, 34.6 kDa) proteins were highly expressed in the E. coli BL21(DE3) at 37°C with 1 mM IPTG for 4 h (Fig. 6b). The recombinant THS-CYP protein was purificated using the nickel-affinity. Further, we measured the PPIase activity of the recombinant protein by a coupled assay using synthetic peptide synthetic peptide succinyl-Ala-Ala-Pro-Phe-p-nitroanilide. The isomerization of the peptide substrate was observed in the presence of recombinant GhCYP-3, showing OD values higher than that of the blank, a negative control (Fig. 6c). These results indicated that GhCYP-3 has PPIase activity in vitro.
GhCYP-3 expression is upregulated in response to V. dahliae invasion To analyze the expression pattern of GhCYP-3 in three cotton cultivars with different degrees of resistance to V. dahliae, the roots of two-week-old cotton seedlings were sampled at 0, 4, 8, 12, 24 and 48 hpi. The qRT-PCR results revealed that the transcriptional levels of GhCYP-3, compared with 0 hpi, were significantly upregulated in two-resistant cultivars Pima90-53 and JM-20 at 4 hpi. However, at the same inoculation time, GhCYP-3 in susceptible Han-208 was strongly downregulated and reached the highest expression level until 24 hpi (Fig 7). These results suggested that GhCYP-3 is upregulated and earlier involved in the cotton interaction with V. dahliae in resistant than in susceptible cultivars.

Overexpression of GhCYP-3 in Arabidopsis improves plant Verticillium wilt resistance
To further evaluate whether the GhCYP-3 functions in plant resistance to Verticillium wilt, we designed primers to amplify the ORF of GhCYP-3 into the plant expression vector pBI121 (Fig. 8a), which will make GhCYP-3 overexpression in transgenic Arabidopsis plants. By selection on kanamycin-containing medium (Fig. 8b), ten potential T 1 transgenic lines were generated. After PCRverification of gene insertion (Fig. 8c), two lines of the T 3 generation displaying GhCYP-3 overexpression (Fig. 8d) were selected for analysis of PPIase activity and Verticillium wilt resistance.
In response to V. dahliae infection, two independent overexpressing lines displayed less wilting and smaller degree of leaf etiolation in comparison with WT plants at 15 dpi (Fig. 9a). Disease evaluation further indicated that the disease indices of transgenic lines were significantly lower than that of the WT (Fig. 9b). Furthermore, we examined the colonization in stems of infected WT and transgenic plants by isolation and cultivation of V. dahliae on PDA (potato dextrose agar) for 8 days. As a result, less fungal colonies came out from transgenic plants comparing to WT (Fig. 9c and d). These fungi recovery assay indicated that GhCYP-3 has a specialized effect of growth inhibition on V. dahliae.

GhCYP-3 exhibited obvious inhibitory effects on V. dahliae
We further assessed the inhibitory effects of GhCYP-3 on V. dahliae. As showed in Fig. 10a, distinct inhibition zones formed around the disc containing recombinant GhCYP-3. Additionally, V. dahliae spores were inoculated with plant extracts from WT and transgenic lines and then be spread on PDA plates. After 48 h, extracts from all the plants significantly reduced the number of colonies compared to the H 2 O (as control). However, the number of fungal colonies on plates contain transgenes extracts was significantly less than the WT (Fig. 10b and c). These results indicated that GhCYP-3 can efficiently inhibit conidia germinating and hyphae growth of V. dahliae in vitro.

Discussion
CYP genes family has been systematically analyzed in several plants, such as Arabidopsis [8], rice [22] and soybean [10]. Recently, 78 CYP genes were also identified in G. hirsutum L. acc. TM-1 from Illumina paired-end genomic sequencing (NAU version 1.1) by Chen et al. [23]. Now, NAU version 2.1, highly accurate reference grade genome assemblies and annotations for G. hirsutum, was generated [21]. A fairly large number of gaps and erroneous assemblies were successfully filled and corrected in new NAU version 2.1 [21]. Therefore, in the present study we identified a total of 79 CYP genes by integrating NAU version 1.1, JGI database and NAU version 2.1 ( Table 1). Compared to the previous report from Chen et al., the supplementary three CYPs (GhCYP-49, GhCYP- Fig. 4 Expression profiles of CYPs from upland cotton inoculated with V. dahliae. Differential expression analysis was performed using the DESeq R package (1.10.1). The fold change (FC) is the ratio of treatment FPKM to control FPKM. Expression data are shown as log 2 (FC) . The resulting Pvalues were adjusted using the 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 and marked with a red arrow (up-regulated) or green arrow (down-regulated). hpi, hours post inoculation 60 and GhCYP-70) were identified in this study. Besides, two CYPs (Gh_Sca140771G01 and Gh_A12G1281), undiscoverable in JGI database and NAU version 2.1, were removed. It thus made the candidate CYP genes in G. hirsutum more reliable.
All 79 CYPs in G. hirsutum have conservative PPIase domain (CLD). Of which, 14 GhCYPs are multi-domain proteins (Fig. 1). In addition to the CLD, GhCYP-14 and GhCYP-51 contain RRM and zf-CCHC. The RRM domain was found in proteins involved in RNA processing where it mediates binding to various RNAs to execute both housekeeping functions and regulatory mechanisms [24]. Proteins that contain zinc fingers typically interact with DNA and RNA, and serve primarily to alter the binding specificity of a particular protein. In addition, GhCYP-14 and GhCYP-51 were predicted to be localized in the nuclear (Table 1). Therefore, these two CYPs possibly mediate ribosomal association while the CLD catalyzes peptidyl prolyl cis-trans isomerization of nascent polypeptides. GhCYP-19 and GhCYP-56 contain WD40 domain that generally serve as a rigid scaffold for protein interactions [25]. Other 10 multi-domain GhCYPs contain TPR domain, which mainly act as interactive scaffolds in the formation of protein complexes and regulators of RNA metabolism involved in the immune response [26,27].
Regulation of gene expression via specific cis-regulatory elements in the promoter regions has evolved as a major adaptive mechanism to respond to environmental stress in plants [28]. Phytohormones are critical to the regulation of plant development and defense [29]. Thus, the analysis of the putative cis-regulatory elements relating to hormone helps to advance our understanding of GhCYPs involving stress tolerance in cotton. Six hormones (ET, MeJA, ABA, GA, SA and Auxin) responsive regulatory elements were detected in the potential promoter regions of GhCYPs (Fig. 2), indicating that GhCYPs involve different hormone-mediated signaling pathways.
In 2010, we isolated 203 ESTs from a cDNA library using suppression subtractive hybridization (SSH) with a resistant upland cotton cultivar Jimian20 induced with V. dahliae. Of which, an EST encoded a partial polypeptide with homology to CYP [30]. This is the first report that CYP is involved in the interaction between cotton and phytopathogen. In 2011, GhCyp1 was cloned from G. hirsutum cv. Zhongmian 35. Overexpression of GhCyp1 in transgenic tobacco plants conferred higher tolerance to salt stress and P. syringae pv. tabaci infection compared with control plants [18]. Here the RNAseq expression analysis of the cyclophilin gene families for G. hirsutum revealed that many of the genes potentially play important roles in various stress response (Figs. 3 and 4). So far as we know, almost all of the reported CYP genes were involved in plant stress reactions by their up-regulation expression, such as OsCYP19-4 and OsCYP21-4 from rice against cold [12] and salt [13], CcCYP from pigeon pea against drought, salinity and extreme temperatures [14], CyPs from Arabidopsis against wounding [11], and several cyclophilin genes in response to phytopathogen infection [15][16][17]. Therefore, significantly higher expression of CYP genes in cotton, such as   GhCYP-67 to V. dahliae suggests likely functional importance under these stress conditions. However, further functional studies are required to unravel the precise role of these candidates during cotton response to biotic and abiotic stress conditions. GhCYP-3 contained the above-mentioned EST sequence related to V. dahliae-infection and showed 98% similarity in amino acid sequences with GhCyp1. Thus, GhCYP-3 potentially plays an important role in regulating cotton immune response. The results of expression (Fig. 7) and overexpression analysis (Fig. 9) further supposed that GhCYP-3 was involved in the cotton defence to V. dahliae. Plant CYPs could locate in multiple cell organelles, such as ER [31], chloroplast [9], Golgi [13], cytoplasm and nucleus [32,33]. Interestingly, the GhCYP3-GFP fusion protein was localized in the cell cytoplasm as well as the cell nucleus in onion epidermal cells (Fig. 5c). Nuclear localization was proposed to play a possible role in the regulation of gene expression [33]. GmCYP1 was as a "helper" that activates the enzymatic activity of a Phytophthora sojae RXLR effector Avr3b in a PPIase activity-dependent manner [34]. Furthermore, GmCYP1 was demonstrated to interact with the isoflavonoid regulators GmMYB176 and 14-3-3 protein, suggesting that it participates in isoflavonoids metabolism and plays role in defense [33]. AtCYP57 was proved to be involved in the A. thaliana response to P. syringae infection by influencing callose accumulation and PAD4 (peptidyl arginine deiminase type 4) expression, which interacts with EDS1 (enhanced disease susceptibility 1) to provide basal immune response in plants. Nucleus location makes authors inferred that AtCYP57 could directly regulate the translation of defence genes [32]. Therefore, we inferred that GhCYP-3 play resistance function, may like AtCYP57 and GmCYP1, in the cell nucleus by directly interacting with some transcription factor to regulate the translation of defence genes, which lead to the production of antimicrobial metabolites.
Alternatively, GhCYP-3 was also located in the cytoplasm, indicating that it needs to play some extra roles, presumably including direct antifungal activity. Antifungal activities of CYPs have been reported from some plants including ginseng [35], Chinese cabbage [36], chickpea [37] and black-eyed pea [38]. In our study,  (Fig. 10a). In a state of nature, the concentration of CYP in plant is hardly so high in the plate. Otherwise, there was no definite evidence to suggest that CYP is involved in plant resistance to pathogens by direct antifungal activity in vivo at present. However, we do not rule it out, because the extracts from GhCYP-3 transgenic Arabidopsis displayed significantly inhibit activity to conidia germinating and hyphae growth of V. dahliae (Fig. 10b). Many characterized antifungal proteins active on the fungal cell wall, plasma membrane and or intracellular targets [39]. For example, Buforin 2 from the stomach tissue of Bufo bufo gargarizans can translocate through the plasma membrane and exhibits antifungal activity upon interaction with fungal DNA and RNA [40]. During the infection process, V. dahliae forms penetration peg and specialized fungus-host interface to secret secretory effector proteins [41]. This is probably a chance for antifungal proteins including GhCYP-3 produced by host plant encountering and entering into the cytoplasm of V. dahliae. Additionally, we identified 10 putative CYPs (VdCYPs) in V. dahliae strain VdLs.17 genome with very high sequence similarity with GhCYP-3 (Additional file 3: Table S3). These mean that GhCYP-3 may replace VdCYPs and carry out the same biological function in the cytoplasm of V. dahliae. Thus, we inferred that the antifungal activity of GhCYP-3 has been shown to be due to the effect on the normal function of VdCYPs, which is essential for the development of V. dahliae. Nevertheless, further research is needed to confirm this bold deduction.

Conclusions
This is the first systematic analysis of CYP family genes in cotton aiming to help clarifying the gene sequence characteristics and expression patterns. The putative cisregulatory elements predication and expression divergence Error bars indicate the SE (n>30) of three biological replicates. Asterisks indicate statistically significant differences as determined by Dunnett's multiple comparisons test (**P < 0.01). c, 10-dpi stem sections were plated on PDA medium. Colonies were formed after incubating 7 days at 25°C . Vd, V. dahliae. d, Percentage of colony formation. Error bars indicate the SE (n = 18) of three biological replicates. Asterisks indicate statistically significant differences as determined by Dunnett's multiple comparisons test (**P < 0.01) suggested that the GhCYP genes are involved in multiple phytohormone regulation pathways and responses to various abiotic stress and V. dahliae infection. These results will provide potential clues for the selection of candidate genes for further in-depth study on the functional characterization. Furthermore, GhCYP-3 showed both cytoplasmic and nuclear localization. Heterologous overexpression of GhCYP-3 in Arabidopsis significantly improved Verticillium wilt resistance of the plants. Recombinant GhCYP-3 and the extracts from GhCYP-3 transgenic Arabidopsis displayed significantly inhibit activity to V. dahliae. These results indicated that GhCYP-3 was associated with the resistance of cotton to V. dahliae infection presumably through antifungal activity, and it will offer an important candidate gene for Verticillium wilt tolerance in cotton molecular breeding.

Methods
Identification and characterization of CYP family genes in G. hirsutum Genome assemblies of G. hirsutum TM-1 from Nanjing Agricultural University (NAU version 1.1 and version 2.1) and JGI (version 1.0) were retrieved from the CottonFGD website (https://cottonfgd.org/). Arabidopsis CYPs, accessed from TAIR website (https://www.Arabidopsis. org/) were used as query to identify putative CYPs in G. hirsutum genomes by local BLAST using BioEdit software (Ibis Biosciences, Carlsbad, CA, USA). The identified GhCYPs were further verified through the Pfam database (http://pfam.xfam.org). The molecular weight (MW) and isoelectric point (pI) of each protein were calculated using ExPASy program (http://www.expasy.org/). The signal peptide was predicted with the program of SignalP 4.1. The amino acid sequences were aligned with DNAMAN software (Vers. 7; Lynnon Corporation, Quebec, Canada), using default parameters. CELLO v2.5 (http://cello.life.nctu.edu. tw/) was used for the subcellular localization prediction of GhCYPs. The putative cis-acting elements in the promoter regions were predicted using NAU version 2.1 database with the Plant CARE (http://bioinformatics.psb.ugent.be/ webtools/plantcare/html/).

Genome-wide expression analysis of GhCYPs
A genome-wide expression analysis of the cotton CYP genes in various abiotic stresses and V. dahliae infection was performed using high-through RNA sequence data, which was downloaded from NCBI databases (SRP044705) The inhibition results were observed at 5 days after incubation. Clearing inhibition zones were formed around the disc containing recombinant GhCYP-3 (1 & 2). The disc containing water was as control (3 & 4). b and c, Extracts from transgenic plants significantly reducing the number of V. dahliae colonies. V. dahliae spores were inoculated with plant extracts and then be spread on PDA plates. After 48 h, the number of fungal colonies on plates contained transgenes extracts was drastically less than the H 2 O and WT. Data represent mean ± SE of three biological replicates. Asterisks indicate statistically significant differences as determined by Dunnett's multiple comparisons test (*P < 0.05, **P < 0.01, ***P < 0.001) and extracted from our RNA-seq data [42]. Genes with FPKM ≥ 10 were used for further expression analysis.

Plant materials and V. dahliae strains
The cotton seeds G. hirsutum cv. Ji Mian 20 (JM20), Han208, CCRI8 and G. barbadense cv. Pima90-53 were preserved at the North China Key Laboratory for Crop Germplasm Resources of Education Ministry, Hebei Agricultural University, Baoding, China. Cotton seedlings were grown in commercial sterilized soil at 28°C /25°C (day/night) temperatures with a 16-h-light/8-hdark regime. A. thaliana was grown in pots containing vermiculite soil with temperature at 23°C day and 20°C night, under a 16/8 h photoperiod. V. dahliae strain Linxi2-1 was isolated from a symptomatic upland cotton plants growing in agricultural fields in Linxi county, Hebei Province, China [43]. These highly aggressive defoliating V. dahliae strains were maintained on PDA. The conidial suspension was prepared according to previous description [44] and adjusted to 10 7 spores per milliliter and 10 6 spores per milliliter with distilled water was used to the inoculation of cotton and Arabidopsis, respectively. The plant was infected with V. dahliae using soil drench method [44]. 10 mL of the conidial suspension was directly injected with a needle without piercing into the bottom of each pot. Seedlings received sterile water in the same manner were used as control.

Gene cloning and subcellular localization
Total RNA was extracted from leaf tissues of JM20 with an RNA plant plus reagent (TIANGEN Biotech, China). First-strand cDNA was synthesized from an aliquot of 1 μg of total RNA with a PrimeScript™ RT Reagent Kit and gDNA Eraser (TaKaRa, China). GhCYP-3 was amplified with primers CYP-F1 and CYP-R1 (Additional file 1: Table  S1), designed based on the sequences of Gh_A01G1361 (G. hirsutum L. acc. TM-1) [20]. A GhCYP-3-GFP fusion construct under the control of the 35S promoter was generated by cloning the ORF into the SalI and BamHI sites of the binary vector pCamE. The vector expressing GFP alone served as control. Protein subcellular localization in onion (Allium cepa) epidermal cells was determined according to the protocol of Yang (2015) [44].

qRT-PCR and semi-quantitative PCR
Total RNA and cDNA were prepared by the method described above. The qRT-PCR was performed on a CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) using the qPCR kit for SYBR Green (TaKaRa, China). qPCR conditions consisted of one cycle of 3 min at 95°C, followed by 40 cycles of 15 s at 95°C, 15 s at 58°C, and 20 s at 72°C. The cotton ubiquitin 14 (UBQ14) gene served as an internal standard [45]. Fold-changes in expression were calculated via the 2 −ΔCt method. Semi-quantitative RT-PCR was done with an Applied Biosystems® 2720 Thermal Cycler with Arabidopsis AtACT2 (AT3G18780) as the internal standard [46]. The resulting products were resolved on a 1.5% agarose gel. All primers are listed in Additional file 1: Table S1 Three biological and three technical replicates were analyzed for all quantitative experiments.

Generation and evaluation resistance of transgenic Arabidopsis
GhCYP-3 was cloned into pBI121 vector at the XbaI and SacI sites by PCR with primers CYP-X and CYP-S (Additional file 1: Table S1). The chimeric construct was introduced into Agrobacterium tumefaciens strain GV3101 for Arabidopsis transformation using the floral dip method [47]. Putative transformants were selected on MS (Murashige and Skoog) medium containing 50 mg L -1 kanamycin, and then be further verified by PCR for gene insertion and semi-quantitative RT-PCR for gene expression. Independent T 1 transgenic lines were used to produce the T 3 generations, which were randomly chosen as representative lines and subjected to analysis. Disease severity of Arabidopsis plants caused by V. dahliae was assessed according to symptoms manifested on the leaves. The disease index (DI) was calculated as previously described [48]. Fungal recovery assay was also used to evaluate the resistance of the plant through assessment of Verticillium colonization recovered from stem sections according to the method of Fradin (2009) [49]. In each treatment, 18 individual plants were used and all the experiments were repeated thrice.

Purification and PPIase activity assay of recombinant protein
The GhCYP-3 ORF was cloned into expression vector pET-32a (+) (Novagen, Darmstadt, Germany) with forward primer CYP-Bg-F and reverse primer CYP-Sa-R (Additional file 1: Table S1), which will introduce the BglII and SacI site into the 5′ and 3′ end of the ORF, respectively. To induce the expression of GhCYP-3 protein with a His-tag in E. coli BL21(DE3) (TransGen Biotech, Beijing, China), a final concentration of 1.0 mmol·L −1 isopropyl-β-D-thiogalactopyranoside (IPTG) was added into the culture when the OD 600 value reached 0.4-0.6, and the culture was allowed to continue growing for 4 -6 h before harvesting. The proteins were separated on a SDS-PAGE gel and detected by western blot using anti His-Tag mouse monoclonal antibody (1:5000; CW Biotech, Beijing, China). The recombinant protein was purified using a 6×His-Tagged Protein Purification Kit (CW Biotech, Beijing, China). The PPIase activity of the recombinant protein was assayed in vitro using the tetrapeptide substrate Suc-AAPF-pNA (N-succinyl-Ala-Leu-Pro-Phe-p-nitroanilide; Sigma-Aldrich, Ontario, Canada) in a Shimadzu UV-2450 spectrophotometer (Shimadzu, Kyoto, Japan) as described by Yoon (2016) [12]. Three biological and two technical replicates were analyzed for all measurement.

Assay of antifungal activity
The antifungal activity of the recombinant protein was tested against V. dahliae using filter paper disc diffusion method. V. dahliae spores were uniformly spread on the PDA medium plates and then cultured for 60 h at 28°C. Sterilized blank paper discs of 6 mm diameters, impregnated with tested protein, were placed on the surface of PDA medium previously spread with V. dahliae. The plates were inoculated at 25°C for 10 d. The antifungal activity of extracts from Arabidopsis plants transformed with GhCYP-3 was performed as described previously [50] with the following modifications. Conidial suspension adjusted to a density of 10 5 conidia ml -1 , and germinated on PDA overnight at 25°C prior to assay. Total homogenates (5 g) from ten Arabidopsis plants were prepared by directly grinding plant leaves into a fine powder in liquid nitrogen with no buffer added. Subsequently, extracts were collected by centrifugation at 10000 g for 10 min at 25°C. Conidial suspensions (25 μl) were mixed with 225 μl of extract, and incubated for 1 h at 25°C. The mixture (50 μl) were spread onto PDA plates and incubated at 25°C for 48 h and fungal colonies enumerated. All experiments were repeated three times independently.