Verticillium dahliae wild type XS11 was isolated from a smoke tree, Cotinus coggygria in Fragrant Hills Park, Beijing (Wang et al., 2013). The spores of the wild type and its derivative mutants and its complementation strains were stored in 15% (v/v) glycerin at -80°C. To acquire conidia, all strains were activated and cultured on potato dextrose agar medium (PDA, containing 200 g of potato, 20 g of glucose, and 15 g of agar per liter) at 25°C and then collected after 7 days for generation of fresh hyphae, germination tests, and etc. For all stress assay, strains were cultured on solid complete medium (CM, 50 ml of 20× nitrate salts, 1 ml of 1000× trace elements, 10 g of glucose, 2 g of peptone, 1 g of yeast extract, 1 g of casamino acids, and 1 ml of vitamin solution per liter). To test sensitivity to osmotic stress, all strains were grown for 24 days on CM containing 0.8 M NaCl and 1.2 M sorbitol. For cell wall stress assay, all strains were grown on CM with 20 μg/ml Calcofluor White (CFW) (Sigma-Aldrich) and 50 μg/ml Congo Red (CR) (Sigma-Aldrich) for 3 and 7 days, respectively. For oxidative stress, agar diffusion tests were performed to measure the sensitivity of strains to H₂O₂, the same spore suspension (10⁵ spores/ml) of each strain were spread on PDA plates, and filter paper discs containing H₂O₂ (6, 12, and 18 mM) were placed in the center of each plate. The inhibition zone was determined after 3 days post inoculation (dpi). For fungicides assay, four different fungicides, such as 5 μg/ml difenoconazole (Sigma-Aldrich), 2 μg/ml chlorothalonil (Sigma-Aldrich), 10 μg/ml fludioxonil (Sigma-Aldrich), and 5 μg/ml iprodione (Sigma-Aldrich) were used. Three independent experiments of three replicates each were performed. To observe microsclerotia formation, conidia were sprayed onto the cellulose membrane (Ø = 80 mm; pore size = 0.22 μm) overlaid on solid basal medium (10 g of glucose, 0.2 g of sodium nitrate, 0.52 g of KCl, 0.52 g of MgSO₄.7H₂O, 1.52 g of KH₂PO₄, 3 μmol thiamine HCl, 0.1 μmol biotin, and 15 g of agar per liter). The microsclerotia formation were observed and photographed after incubation for every 48 h intervals. At 7 dpi, the observations were conducted every 7 days. All experiments were repeated at least three times.

Information regarding VdPbs2 was obtained from JGI¹. Homologs of VdPbs2 were identified using BLASTP searches of home databases of other fungal species (Broad Institute and Joint Genome Institute). Multiple sequence alignments were conducted using ClustalX 2.0 (Larkin et al., 2007). The phylogenetic tree was constructed using Mega6.0 (Tamura et al., 2013) with the Neighbor Joining algorithm under default settings and 1000 bootstrap replications.

To delete VdPbs2 in the genome of V. dahliae, we used the split-marker method. First, the 1476 bp upstream (5′) and 1494 bp downstream (3′) flanking sequences of VdPbs2 were amplified with primer pairs LY105/LY106 and LY107/LY108, respectively (Supplementary Table S1). The geneticin-resistance cassette was amplified with the Geneticinfor/Geneticinrev primers for deletion, which include approximately 20 bp that overlaps with the 5′ and 3′ flanking sequences, respectively. The two deletion cassettes resulting from fusion PCR with primer pairs LY105/Geneticinrev and Geneticinfor/LY108 (Supplementary Table S1) were used for protoplast transformation after sequencing. To obtain ΔVdPbs2 complementation strains, the 3804 bp segment and the VdPbs2-GFP fusion construct were constructed containing the native promoter and coding region of VdPbs2. The 3804 bp segment for native complementation, amplified with primer pair LY109/LY166 (Supplementary Table S1) is used to restore the defects of ΔVdPbs2 mutant. The VdPbs2-GFP fusion plasmid was constructed as follows. Firstly, a 3.76 kb genomic fragment was amplified with the primer pair LY105/LY167 (Supplementary Table S1), including the native promoter and the full VdPbs2 open reading frame region. Then, it was inserted into the pKD5-GFP digested with SmaI. Confirmations were performed using PCR with the primer pairs LY137/LY165-RB, restriction digestion and sequencing. Finally, the native complementary segments of VdPbs2 and VdPbs2-GFP fusion constructs were transformed along with a hygromycin-resistance cassette into ΔVdPbs2 protoplasts using the PEG method (Wang et al., 2013). All transformants were verified using external screening primer pair LY137/LY138 and internal screening primers pair LY145/LY146 (Supplementary Table S1). The ΔVdPbs2/Pbs2GFP strain was preliminarily screened for GFP fluorescence and then verified using the external screening primer pair LY137/LY138 and the internal screening primer pair LY145/LY146 (Supplementary Table S1). Finally, southern blotting was performed to confirm the deletion of VdPbs2 with the DIG High Prime DNA Labeling and Detection Starter Kit I in accordance with the manufacturers’ protocol (Roche, Germany). The genomic DNA of wild type and the deletion of VdPbs2 strain was digested with KpnI and hybridized with a probe amplified from the V. dahliae strain XS11 genomic DNA with LY170up/LY170down (Supplementary Table S1) and labeled with the DIG primer.

Fresh mycelium of ΔVdPbs2 mutants and wild type were cultured in CM at 25°C for 5 days and collected with single-layer miracloth. Mycelia were subjected to RNA extraction using TRIzol reagent (Invitrogen) and purified with the RNA Mini Kit (Ambion). RNA integrity was confirmed by agarose gel electrophoresis. Reverse-transcription PCR was performed with Oligo-DT and SuperScript III reverse transcriptase (Invitrogen). Quantitative real-time PCR (qRT-PCR) was performed with SuperReal Premix Plus (TIANGEN, China) on an ABI 7500 real-time PCR system (Applied Biosystems, USA). The β-tubulin of V. dahliae is used as an internal reference. Relative expression levels were calculated using the ΔΔCT method (Livak and Schmittgen, 2001). All primers used in this study are listed in Supplementary Table S1.

To test the ability of penetration of ΔVdPbs2 mutant, spores were dropped onto onion epidermis at the concentration of 10⁴ conidia/ml. At 32 hpi, the penetration was observed after staining with aniline blue under light microscopy (DM2500, Leica). To determine the pathogenicity of the ΔVdPbs2 mutant, spores were filtered from liquid CM after 10 days of cultivation and then diluted to 10⁶/ml with distilled water. One-year-old smoke tree seedlings were selected for inoculation and soaked in the conidia suspension for 10 min. The seedlings were then replanted in autoclaved soil and observed at regular intervals. To determine whether specific strains could invade the seedlings, the seedling stems were clipped into tiny fragments for isolation 14 days after inoculation (Xiong et al., 2015). Tobacco seedlings were also used for virulence tests using the same methods. The height of tobacco seedlings were measured at 30 dpi.

To analyze the response to stress, mycelium of the wild type and VdPbs2 deletion mutant were inoculated in the CM with 0.8 M NaCl for 4 days, then myceliua were collected for observation. Pictures were taken using the microscope (Leica DM 2500). To analyze of subcellular localization of VdPbs2, conidia and hyphae were collected from liquid CM. Then the fluorescence of mycelium and conidia treated with 0.8 M NaCl for 2 h were observed. The pictures were acquired using a Leica SP5 confocal laser-scanning microscope. A diode laser, Argon/2 (458, 477, 488, 496, 514 nm) was used, and the fluorescence filters were EX 488; EM 510/40. The quantification of image fluorescence was performed using the Adobe Photoshop software.

The melanized area fraction was measured using ImageJ² under the default settings (all the threshold of image was 42.589) (Papadopulos et al., 2007). Data were expressed as mean values ± standard error of the mean. Statistical analyses were performed by using Student’s t-test. A p-value < 0.05 was considered as statistically significant.

To investigate whether the other component of HOG signaling pathway affects the physiology and morphology of V. dahliae, we identified the homolog of S. cerevisiae Pbs2 in the V. dahliae genome database. A gene encoding a MAPK kinase (VDAG_02783) was designated as VdPbs2. The protein contains two kinase motifs (residues 258–280 and 325–563) and a tyrosine kinase domain, Pkinase_Tyr (residues 322–559, marked with dashed lines in Supplementary Figure S1). Subsequent phylogenetic analysis and amino-acid sequence alignments revealed that VdPbs2 has high sequence similarity with Pbs2 homologs in other fungi, particularly those in V. alfalfae and N. crassa. Moreover, RNA-Seq revealed that expression levels of VdPbs2 increase during microsclerotial development at 60 h, 72 h, 96 h, and 14 days in XS11 strain (Xiong et al., 2014).

Two deletion mutants (ΔVdPbs2-22 and ΔVdPbs2-32) were verified by PCR and Southern blots (Supplementary Figure S2). The complemented strain ΔVdPbs2/Pbs2GFP was confirmed to harbor the full-length VdPbs2 gene (Supplementary Figure S2) and restore phenotypes of the ΔVdPbs2 mutant (Supplementary Figure S3). The results showed that the deletion mutants and complementation strain (ΔVdPbs2/Pbs2 and ΔVdPbs2/Pbs2GFP) were successfully generated.

To investigate the role of VdPbs2 in microsclerotia formation, we first paid our attention to the connection between VdPbs2 function and axenic growth on plate media. Similar to the ΔVdHog1 mutant, ΔVdPbs2 mutants exhibited no significant difference in growth rate but delayed to form microsclerotia on PDA compared with the wild type (Figure 1A). Few melanized microsclerotia can form in the ΔVdPbs2 mutant; by contrast, abundant melanized microsclerotia were produced in the wild type and the ΔVdPbs2/Pbs2 strain (Figure 1A). To determine the influence of VdPbs2 on microsclerotia in detail, we observed the microsclerotia formation on BM. The wild type and the ΔVdPbs2/Pbs2 strain started to accumulate a small amount of melanized microsclerotia at 3 dpi, however, a small number of melanized microsclerotia were observed in the ΔVdPbs2 mutant at 7 dpi (Figure 1B). Furthermore, at 24 dpi, the ΔVdPbs2 and ΔVdHog1 strains still had significant defects in microsclerotia formation, and the melanized area fraction of each strain revealed the deficiency in the melanin accumulation in ΔVdPbs2 and ΔVdHog1 mutants when compared with wild type and the ΔVdPbs2/Pbs2 strain (Figures 1C,D). Strikingly, the melanized microsclerotia were significantly less in the ΔVdHog1 mutant than that of in the ΔVdPbs2 mutant (Figure 1), indicating that VdHog1 may play a more prominent role in the formation of melanized microsclerotia.

Consistent with reduced melanin accumulation in the ΔVdPbs2 mutant, genes associated with melanin synthesis were expressed at significantly lower levels in ΔVdPbs2 mutant (Figure 2A). Notably, of five melanin-related genes, four genes (VDAG_00190, VDAG_03665, VDAG_03393, and VDAG_00183) were more than 50-fold down-regulated in ΔVdPbs2 mutant compared with the wild type and the ΔVdPbs2/Pbs2 complementation strain (Figure 2A). The result was consistent with expression profiles of these genes in ΔVdHog1 mutant (Wang et al., 2016). Furthermore, we tested the expression analysis and subcellular localization of VdPbs2 fused with GFP under the control native promoter of VdPbs2. The results demonstrated that VdPbs2 was significantly upregulated during microsclerotia formation and green fluorescence remained a higher level at the early stage of microsclerotia formation (Figures 2B,C). Taken together, these observations indicate VdPbs2 is required for melanized microsclerotia formation via the Hog1-mediated pathway.

To investigate the function of VdPbs2 in the response to hyperosmotic stress, strains were grown on CM supplemented with 0.8 M NaCl and 1.2 M sorbitol, respectively. When grown on minimal media containing 0.8 M NaCl and 1.2 M sorbitol, respectively, ΔVdPbs2 mutant, compared to the wild type and the ΔVdPbs2/Pbs2 strain was dramatically reduced for growth, which was similar to ΔVdHog1 mutant (Figures 3A,B). Besides, clear hyphal lysis occurred in both ΔVdPbs2 and ΔVdHog1 mutants indicated by hyphae deformities visible on the above media (Figure 3C). As shown in Figures 3D,E, cytoplasmic distribution of VdPbs2 was clearly observed after treated with 0.8 M NaCl. Collectively, the results suggested that VdPbs2-VdHog1 module contributes to the response to osmotic stress in V. dahliae.

To determine whether deletion of VdPbs2 affects the response to cell wall stress in V. dahliae, we tested cell viability of the ΔVdPbs2 mutant under cell wall stressors such as CFW and CR. Conidia (10⁵ conidia/ml and 10⁶ conidia/ml) of ΔVdPbs2, ΔVdHog1, wild type, and ΔVdPbs2/Pbs2 strain were spotted on CM media containing CFW (20 μg/ml) and CR (50 μg/ml), respectively. Enhanced growth on media with CFW (20 μg/ml) and CR (50 μg/ml), respectively, was observed for ΔVdPbs2 and ΔVdHog1 mutants. By contrast, reduced growth was observed for the wild type and ΔVdPbs2/Pbs2 strain (Figures 4A,B), suggesting VdPbs2 and VdHog1 are involved in the response to cell wall stress.

We next sought more evidence for a functional connection between VdPbs2 and cell wall assembly. To determine the expression profiles of genes encoding chitin synthase, we used qPCR to analyze RNA extracted from wild type and ΔVdPbs2 mutant strains grown in liquid shake CM for 5 days. Loss of VdPbs2 function induced the expression of chitin synthase genes (VDAG_08591 and VDAG_03141) compared to wild type (Figure 4C). Thus, genes for chitin synthase are misregulated in ΔVdPbs2 mutant when compared with wild type, accounting for enhanced resistance to cell wall stressors. Summarily, these results demonstrate that VdPbs2 may negatively regulate cell wall synthesis.

To evaluate the responses of the ΔVdPbs2 mutant to oxidative stress, the inhibition zone was measured on the media containing H₂O₂. As shown in Figures 5A,B, the ΔVdPbs2 mutant exhibited the larger inhibition zones than the wild type and the ΔVdPbs2/Pbs2 strain at a different concentration of H₂O₂ suggesting that VdPbs2 is required for H₂O₂ detoxification. In addition, consistent with our previous observations, loss of VdHog1 did not abolish oxidative sensitivity in V. dahliae (Figure 5A). Furthermore, based on sequence homology, we identified genes encoding H₂O₂ detoxification in V. dahliae. Transcript analysis revealed that three genes (VDAG_08724, VDAG_03661, and VDAG_06340) were consistently down-regulated in the ΔVdPbs2 mutant compared to that of the wild type after treated with 1 mM H₂O₂ for 30 min (Figure 5C). Thus, VdPbs2 is essential for the oxidative stress response, but not VdHog1.

VdHog1 deletion mutant is highly resistant to the fungicide fludioxonil (Wang et al., 2016). To determine if deletion of VdPbs2 affects the response to fungicides, we tested the sensitivity of the ΔVdPbs2 mutant to various fungicides. Similar to the response of the ΔVdHog1 mutant to fungicides, the ΔVdPbs2 mutant exhibited enhanced resistance to fludioxonil and iprodione and increased sensitivity to chlorothalonil and difenoconazole, respectively, when compared with the wild type and the ΔVdPbs2/Pbs2 strain (Figure 6), suggesting that VdPbs2 is involved in accumulation of osmoprotectant molecules of fungal cell in the response to fungicidal compounds.

We next sought to address whether VdPbs2 plays a role in virulence in plants. We used seedlings of smoke tree and tobacco to carry out the virulence experiments. On both hosts, the ΔVdPbs2 mutant exhibited striking reduced virulence (Figures 7A,B) and only less 20% mortality of plants at 45 dpi (Figure 7D). By contrast, at 45 dpi, up to 80% mortality of which inoculated with the wild type and the ΔVdPbs2/Pbs2 strain showed clear wilt symptoms, including chlorosis (Figures 7A,B) and obviously reduced plant height (Figures 7B,C). Due to the limitations, we just further observed the penetration of the strain on onion epidermis. The wild type could infect epidermal cells and expand into the epidermal tissues, whereas the ΔVdPbs2 mutant hardly infects epidermal cells even though the mutants produced long germ tubes (Supplementary Figure S4). Together, these results indicated that VdPbs2 may be involved in the penetration process during plant infection.