In this study, S. sclerotiorum isolate 1980 (Godoy et al., 1990) was used as the wild-type strain. Strains were routinely cultured at 20°C on potato dextrose agar (PDA; Difco Laboratories, Detroit, MI, USA). Gene deletion transformants were cultured on PDA with hygromycin B at 100 μg/ml (Calbiochem, San Diego, CA, USA). Gene complemented strains were cultured on PDA with hygromycin B at 100 μg/ml and G-418 at 30 μg/ml (Sigma, St. Louis, MO, USA).

SsXyl1 gene was obtained through the BLAST searches for homologous sequences in S. sclerotiorum genome using reported xylanases gene in B. cinere (Brito et al., 2006) and some other plant pathogenic fungi. The signal peptide sequence and conserved domain were predicted using the SignalP 4.1 Server¹ and PFAM², respectively. The sequence alignments were carried out using DNAMAN software (Lynnon BioSoft, Vaudreuil, QC, Canada) and displayed with GeneDoc software (Nicholas and Nicholas, 1997). Conserved amino acids were shown with a shaded background. The phylogenetic tree was constructed using maximum likelihood analysis in MEGA (Tamura et al., 2011).

To assay the xylanase activity of SsXyl1, the cDNA fragment encoding the amino acid T²² to S²²² of SsXyl1 (without signal peptide) was artificially synthesized (Shengong, Shanghai, China) and inserted into pPICZαA. The resulting vector was transformed into Pichia pastoris X33 strain via electroporation. The strain was cultured in BMMY medium (1% yeast extract, 2% peptone, 100 mM potassium phosphate (pH 6.0), 1.34% yeast nitrogen base with ammonium sulfate, 0.00004% biotin, and 0.5% methanol) at 28°C under shaking (220 rpm) for 48 h. The culture filtrate was collected and the xylanase activity was measured via the 3, 5-dinitrosalicylic acid (DNS) method (Miller, 1959; Karaoglan et al., 2014). To measure the xylanse activity, 900 μl 1.0% beechwood xylan (Sigma, St. Louis, MO, USA) in 50 mM sodium citrate buffer (pH 5.0) were incubated 5 min at 50°C. One hundred microliters of samples (culture filtrate) were added in the xylan solution followed by incubation 5 min at 50°C. Next, 9 ml DNS solution was added to the reaction mixture. The mixture was then boiled for 5 min, and the absorbance was measured at 540 nm. Xylose (0–6 μmol) was used to create a standard curve. One unit of xylanase activity was defined as the amount of enzyme catalyzing the formation of 1.0 μmol of xylose per minute at pH 5.0 at 50°C. The culture filtrate of the strain transformed with pPICZαA as the control.

A reverse-transcriptase polymerase chain reaction (RT-PCR) was applied to determine the relative expression levels of SsXyl1 during the different development stages and the infection processes of the S. sclerotiorum. The wild-type strains were inoculated on the PDA medium and Arabidopsis thaliana as described by Yu et al. (2015) followed by mycelia harvest. To compare the expression level of SsXly1 on different carbon sources, the wild-type strains were cultured on different minimum medium in which the glucose was replaced with xylan or tomato leaf extract. The total RNA in each sample was extracted with Trizol (Huashun, China). The first-strand of cDNA synthesis was performed with RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, Waltham, MA, USA). The real-time RT-PCR was applied on a Bio-Rad CFX96^TM Realtime System (Hercules, CA, USA). The SsXyl1 cDNA abundance was normalized using the β-tubulin gene (tub1, SS1G_04652) as an internal control. The primer pairs qRT-Xyl1fp (TTCTCGGCAATTTCTACATCC)/qRT-Xyl1rp (CCCAATCTACAGTGAAGGAGC) and qRT-Tubfp (GTGAGGCTGAGGGCTGTGA)/qRT-Tubrp (CCTTTGGCGATGGGACG) were designed with Primer Premier 6.0 (Premier, Canada) and used to amplify the cDNA of SsXyl1 and tub1. The amplification mixtures were composed of 10 μl of SYBR Green Realtime PCR Master Mix (Toyobo, Japan), with 4 pM primer, 1 μl of cDNA, and ddH₂O water to a final volume of 20 μl. Amplification steps were as follows: 95°C for 2 min (1 cycle) followed by 95°C for 20 s, 58°C for 15 s and 72°C for 20 s (40 cycles). Each sample was analyzed over three biological replicates and each real-time RT-PCR analysis was repeated three times.

The RT-PCR was used to analyze SsXyl1 expressions in the gene deletion and complemented strains. The transforms were cultured on PDA supplemented with different antibiotics for 4 days, and the mycelia were harvested. The total RNA extraction and cDNA synthesis were performed as mentioned before. The primer pairs RT-Xyl1fp (CTTCACTGTAGATTGGGACACC)/RT-Xyl1rp (ACTAACCGTTTGCGTAGCAG) were amplified a 460-bp fragment of SsXyl1 cDNA. The tub expressions amplified with qRT-Tubfp/qRT-Tubrp served as the control.

The gene deletion vector was constructed based on vector pSKH (Hamid et al., 2013). A primer pair Xyl1DF (CGCAAGCTTGGGGAGAGAGATATGCAAATGT) and Xyl1DR (CCGCTCGAGAGAGGCTTTTAGTTTCACAATG) were designed to amplify a 1000-bp of 3′ untranslated region (UTR) of SsXyl1 gene. The fragment was digested with HindIII and XhoI and then inserted into pSKH to produce the pSKHXYL1. The primer pairs Xyl1UF (GCCGAGCTCATGGAAATAAACACGCCTAGC) and Xyl1UR (GCCGTCGACCCACGCAGTGCTTTATATACT) were amplified a 1000-bp of 5′ UTR of SsXyl1 gene. The fragment was then digested with SacI and SalI and inserted into pSKHXYL1 to generate the pSKHXYL2. The construct was used as a PCR template to generate split-marker fragments. The 5′UTR::N-HY was amplified using primers Xyl1U (CATGGAAATAAACACGCCTAGC) and HY (AAATTGCCGTCAACCAAGCTC). The 3′UTR::C-YG was generated using YG (TTTCAGCTTCGATGTAGGAGG) and Xyl1D (GAGAGGCTTTTAGTTTCACAATG). The overlap between the two hygromycin resistance gene (hph) fragments was 741 bp. The two fragments were used to transform into S. sclerotiorum wild-type protoplasts as described by Rollins (2003).

To construct the gene complement vector, a neomycin-resistance gene was inserted into pCAMBIA3300 at the XbaI site to produce pCAMBIA3300NEO. A pair of primers 07749FP (CCGGAATTCTTCTATGCACGCAGACTTATTG) and 07749RP (CCGCTCGAGGTCAAATACAACCTGCCACCT) were designed to amplify the genome DNA from the wild-type strain and yield 2.8-kb fragments that contained the SsXyl1 coding sequence and 1 kb of 5′ and 1 kb of 3′ untranslated region. The fragment was digested with EcoRI and XhoI and then inserted into pCAMBIA3300NEO to produce the pCXYL1. The vector was then linearized with XhoI and transformed into the KO52 protoplasts with PEG methods (Rollins, 2003).

The pathogenicity assays were performed on leaves of Brassica napus Zhongshuang 9 and A. thaliana Columbia-0. B. napus were grown in a greenhouse at 25°C to 35°C for about 10 weeks. A. thaliana were cultivated in a growth chamber at 25°C with a 12 h light/12 h dark cycle for 5 weeks. The leaves of B. napus and A. thaliana were inoculated with mycelia-colonized agars (φ = 6 mm) obtained from actively growing colony margins of the wild-type, gene deletion, and complemented strains. The inoculated plants were maintained at 90% relative humidity. Photographs were taken at 72 h post inoculation (hpi) for rapeseed leaves and at 120 hpi for A. thaliana. The experiment was repeated three times and five plants were inoculated with each strain for each repeat.

The SsXyl1 gene contains an ORF of 669 bp and encodes a 222-amino-acid polypeptide. The N-terminal of SsXyl1 was predicted to contain a typical peptide with SignalP 4.1 Server (Petersen et al., 2011). The predicted cleavage site was between amino acid position 21 and 22. This generates a putative mature protein with a calculated molecular mass of 21.69 kDa and an isoelectric point (pI) of 4.82. Residues 46–220 in the protein were predicted by PFAM (Finn et al., 2015) to be a Glyco_hydro_11 glycosyl hydrolase motif. Sequence comparison showed that SsXyl1 exhibited high sequence similarity with endo-β-1, 4-xylanase of B. cinerea XYN11A (53% identities, E-value: 5e-50) (Brito et al., 2006), Trichoderma reesei XYN2 (45% identities, E-value: 2e-55) (Enkerli et al., 1999) and Fusarium graminearum XYL2 (41% identities, E-value: 3e-46) (Dong et al., 2012). The sequence alignment and phylogenetic tree are shown in Figure 1.

To confirm the xylanase activity of SsXyl1, the SsXyl1 cDNA (without signal peptide) was linked into pZICZαA. The resulting vector was transformed into the yeast P. pastoris X33 strain. The strains were cultured with shaking, and the xylanase activity of culture filtrate was determined. The results showed that the enzyme activity for the strain carrying SsXyl1 was 39.17 U/ml, and the wild-type strain was 0.47 U/ml. These results indicate that SsXyl1 encodes a xylanase in S. sclerotiorum.

The expression levels of SsXyl1 during the different development stages and the infection process were determined with real-time RT-PCR. Figure 2A shows that the SsXyl1 expression was slightly increased during the development and carpogenic germination stages of sclerotia. When inoculated on the A. thaliana, the SsXly1 expression was dramatically increased at 3 hpi and then decreased gradually. The expression was still higher than at 0 dpi (Figure 2B). The results suggest that SsXyl1 expression is strongly induced during infection.

SsXyl1 expression was also determined in the mycelia of S. sclerotiorum grown on different carbon sources. Figure 3 shows that the expression level of SsXyl1 for the mycelia grown in xylan was significantly higher than that grown in glucose. The level was maximal when xylan was the only carbon source. The expression pattern indicates that SsXyl1 expression was induced by xylan. The SsXly1 expression is also induced when the tomato leaf extracts were the carbon source.

To functional analysis the SsXly1 in S. sclerotiorum, SsXyl1 gene deletion vector was constructed based on the vector pSKH as described in the Section “Materials and Methods.” The construct was used as the PCR template to generate split-marker fragments, which transformed the protoplasts of the wild-type strain (Supplementary Figure S1A). Several transformants were obtained and confirmed by amplifying of the hygromycin resistance gene. Some strains were chosen randomly and the SsXyl1 expressions were determined with RT-PCR. The results showed that KO52 strains lacked the SsXyl1 transcript (Supplementary Figure S1B). The complementation strain KOC5 was generated via the transformation of the SsXyl1 complemented vector into the KO52 protoplasts. RT-PCR revealed that the SsXyl1 transcript was present in KOC5 (Supplementary Figure S1B).

Figure 4A shows the colonial morphology of the wild-type, SsXyl1 deletion (KO52), and complemented (KOC5) strains on PDA plates. The gene deletion strain showed abnormal morphology versus the wild-type strain. The strain produced fewer sclerotia than the wild-type strain. Figures 4B,C indicates that the number and weight of sclerotia for KO52 were approximately half of those of the wild-type strain. Most importantly, sclerotia produced by the KO52 cannot produce apothecia under standard conditions of carpogenic germination. The KO52 exhibited a dense hyphal branch when observed by microscopy (Figure 5A). It has a slower growth rate versus the wild-type strain (Figure 5B).

The wild-type, SsXyl1 deletion, and complemented strains were all inoculated on the detached leaves of oilseed rape and A. thaliana plants. Figure 6 shows that the gene deletion mutant KO52 lost the virulence to the hosts, while the re-induction of SsXyl1 in the deletion mutant almost completely rescued the phenotype. The results suggest that SsXly1 plays an important role in the virulence of S. sclerotiorum.