Colletotrichum gloeosporioides from Hevea was isolated and kept previously (BioSample: SAMN17266943). Knockout mutants ΔCgnoxA, ΔCgnoxB, and ΔCgnoxR were constructed in our previous study (Guo and An, 2018). The strains were kept on potato agar (PDA) or cultured in a liquid medium. For the microscope analysis, conidia were cultured on Yeast Casein Sucrose (YCS) medium (1 g l^-1 yeast extract, 1 g l^-1 acid-hydrolyzed casein, 2% w/v sucrose, pH 6.9).

To generate a double mutant of CgnoxA and CgnoxB, the ΔCgnoxA strain was used as the recipient strain in which CgnoxB was knocked out using a split-marker strategy as described in our established protocol. Briefly, the flanking sequences of CgnoxB were amplified and fused with the split fragments of the neomycin phosphotransferase gene (NPTII) which confers resistance to Geneticin (G418) (Thermo Fisher, Waltham, MA, USA). Then the two recombinant fragments were co-transformed into protoplasts of the ΔCgnoxA strain for the gene knockout.

To generate the complementation strain, the vector pMD-PgTt which contains the terminator of trpC from A. nidulans and the hygromycin phosphotransferase gene (HPT) was used. The nucleotide sequences of Cgnox genes together with their native promoters were amplified and ligated into the vector pMD-PgTt, respectively. Then, the plasmids were linearized before the protoplast transformation. Positive complementation strains were named as Res-ΔCgnoxA, Res-ΔCgnoxB, and Res-ΔCgnoxR.

To label the actin structure, the Lifeact-EGFP-expressing strain was used as the recipient strain, and CgnoxA, CgnoxB, and CgnoxR were knocked out respectively as described previously (Guo and An, 2018). In addition, the double-mutant strains of CgnoxA and CgnoxB were generated as described above. Protoplast preparation and transformation were performed as described in our established protocol (Wang et al., 2018). The primers that were used are listed in Supplementary Table S1 .

For the colony growth assay, disks of mycelium with a diameter of 0.5 cm were inoculated onto the PDA medium (2 g l^-1 NaNO₃, 0.5 g l^-1 KCl, 1 g l^-1 KH₂PO₄, 0.5 g l^-1 MgSO₄·7H₂O, 0.01 g l^-1 FeSO₄·7H₂O, pH 6.9), and colony morphology and diameter were recorded. Each strain contained three replicates, and all of the experiments were performed twice.

The pathogenicity assay was carried out as described in our previous report (Gao et al., 2022). Briefly, conidia were collected, washed two times with ddH₂O, and resuspended in a solution of 0.5% Sabouraud Maltose Broth (Difco, Franklin Lakes, NJ, USA) to a final concentration of 2 × 10⁵ conidia ml⁻¹. The detached leaves from rubber tree variety 73-3-97 were used for inoculation. The leaves were divided into two groups, with one group of leaves being pre-wounded with a sterile needle and the other group without being wounded. Then, droplets (5 μl) of the conidial suspensions were inoculated onto the leaves. The leaves were kept in a moist chamber at 28°C under natural illumination for 4 days, and the disease symptoms were recorded. Each treatment contained three replicates of 10 leaves, and the entire experiment was repeated three times.

For the calculation of the appressorium formation ratio, conidia resuspended with ddH₂O at a concentration of 5 × 10⁵ conidia ml^-1 were incubated on hydrophobic plastic plates. After incubation for 12 and 24 h, the germination behavior was observed using Leica DM2000 microscopy. For penetration assays, the conidium droplets (3 × 10⁵ conidia ml^-1) were inoculated on the onion epidermis that was plated on water agar plates. After incubation for 16 h, the infection structures were observed using Leica DM2000 microscopy. Each treatment contained three replications, and the entire experiment was conducted twice.

For the RNA extraction from vegetative mycelia, the conidial suspension was inoculated into the liquid complete medium and cultured at 120 rpm, 28°C, for 2 days. Then, the mycelium was collected, disrupted in liquid nitrogen by grinding in a mortar with a pestle, and used for RNA extraction. For the RNA extraction from appressoria, conidia suspension in ddH₂O at a concentration of 1 × 10⁶ conidia ml^-1 was incubated on polyester; after incubation for 24 h, appressoria were collected by an RNase-free scraper and used for RNA extraction. For RNA extraction from infectious mycelia, the conidial suspension was inoculated on rubber leaves and incubated for 3 days, then the lesion area was cut from the leaves, disrupted in liquid nitrogen, and used for RNA extraction. The RNA was extracted using the RNAprep Pure Plant Plus Kit (TIANGEN Biotech, Beijing, China). For cDNA synthesis, 1 μg of total RNA from different samples was used for reverse transcription with FastKing gDNA Dispelling RT SuperMix (TIANGEN Biotech, Beijing, China) according to the manufacturer’s recommendations. To analyze the transcription levels of CgnoxA, CgnoxB, and CgnoxR during different stages, a quantitative real-time PCR was performed with QuantStudio 6 (Thermo Fisher, Waltham, MA, USA) in a 20-μl reaction volume using ChamQ SYBR Color qPCR Master Mix (Vazyme, Nanjing, China). The expression levels of chitin synthase genes in the mycelium of mutant strains were analyzed as demonstrated above. The β2-tubulin-coding gene was used as an endogenous control for normalization, and relative expression levels were estimated using the 2^-ΔΔCt method (Livak and Schmittgen, 2001). The primers are listed in Supplementary Table S1 .

For DAB staining, conidium droplets (3 × 10⁵ conidia ml^-1) were inoculated on the onion epidermis that was plated on water agar plates. After incubation for 12 h, the infection structures were stained with 1 mg/ml DAB (pH3.8, 30 μl) for 12 h under darkness, then the accumulation of ROS was observed with Leica DM2000 microscopy and the average optical density (AOD) of dark-brown polymers were quantified using ImageJ software. For AOD quantification, all the images were changed to eight-bit type at first. Then we selected “Area” and “Integrity density” in “Set Measurements.” After clicking “Calibrate” in “Analyze,” we selected “Uncalibrated OD” in “Function,” and then the gray value 255 equals OD value 0, and gray value 0 becomes OD value 2.71. Each strain contained three replications, and at least 60 appressoria were measured for each replicate.

For visualization of O₂•- production, the conidial suspension was inoculated on a YCS medium-coated glass slide and incubated in a moist chamber at 28°C of 5 h before being stained with 0.05% (w/v) NBT aqueous solution for 10 min. Then the O₂•- production of hyphae was observed by Leica DM2000 microscopy. In order to analyze the distribution of formazan precipitate in the top of germinated hyphae of different strains, the AOD in two different zones in the hyphal tip was calculated as follows: apex and subapex 7 μm, 0–7 μm from the tip; and shank 8 μm, 7–15 μm from the tip. For each strain, more than 10 hyphae were measured in each experiment, and the entire experiment was conducted twice.

To investigate the actin filament structure in germinated hypha, conidial suspensions of WT, ΔCgnoxA, ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB strains expressing Lifeact-EGFP (5 × 10⁵ conidia ml^-1) were incubated on YCS-coated glass slides for 5 h before observation under the confocal. To investigate the actin filament structure in appressoria, conidial suspensions of each strain (2 × 10⁵ conidia ml^-1) were incubated on the hydrophobic borosilicate glass coverslips (Thermo Fisher, Waltham, MA, USA) for 24 h before observation. For diphenyleneiodonium (DPI) treatment, the DPI solution was added into conidial suspension to the concentration of 40 μmol l^-1. For the microscope analysis, the tip of germinated hypha and appressorium were captured through the Leica TCS SP8 laser scanning confocal microscope, with excitation of 488-nm argon laser and emission wavelength range of 505–525 nm. The projection of z-stack images was performed with ImageJ (http://rsbweb.nih.gov/ij/, version 1.47g). To compare the distribution of actin filaments in the top of germinated hyphae of different strains, the fluorescence intensity in three different zones in the hyphal tip was calculated as follows: apex 2 μm, 0–2 μm from the tip; subapex 5 μm, 2–7 μm from the tip; and shank 8 μm, 7–15 μm from the tip. For each strain, more than 10 hyphae were measured in each experiment, and the entire experiment was conducted twice.

For staining with Calcofluor white (CFW), hyphae that incubated on YCS-coated glass slides as mentioned above were stained with a 10-μg ml^-1 CFW aqueous solution (Sigma-Aldrich, Merck, USA) for 10 min in the dark. The fluorescence was imaged via the Leica TCS SP8 laser scanning confocal microscope, with excitation of 405-nm UV laser and emission wavelength range of 430–460 nm. The projection of z-stack images was performed with ImageJ (http://rsbweb.nih.gov/ij/, version 1.47g). Quantification of the fluorescent intensity was performed by measuring the mean gray value using ImageJ software. For each strain, more than 10 hyphae were measured in each experiment, and the entire experiment was conducted twice.

Conidial suspensions were inoculated into 100 ml liquid complete medium to the initial concentration of 10⁵ conidia ml^-1. After incubation at 120 rpm, 28°C, for 16 h, mycelium was collected by miracloth, washed with ddH₂O, and drained with filter paper. Then, 0.2 g mycelium was incubated in 10 ml Glucanex solution at 100 rpm, 28°C for 3 h. Then protoplasts were collected by filter with miracloth, and the concentration was measured with a hemocytometer under a microscope. Each strain contained three replications.

Statistical significance analyses were performed in PASW Statistics (IBM, USA). Data with a single variable were analyzed by one-way ANOVA, and mean separations were performed by Duncan’s multiple-range test. Differences at P < 0.05 were considered significant.

To investigate whether CgNOXs are involved in vegetative growth, the mutant strains were cultured on PDA medium and the colony growth was recorded. After culture for 5 days, all the mutants showed a similar colony morphology to WT. The colony diameters of ΔCgnoxA and ΔCgnoxB were nearly the same as those of WT, while those of ΔCgnoxR and ΔCgnoxAnoxB were slightly decreased compared with WT ( Figure 1 ). These results suggested that the CgNOX complex is required for vegetative growth.

The roles of CgNOXA, CgNOXB, and CgNOXR in the pathogenicity of C. gloeosporioides were investigated via inoculation assay on detached leaves with or without wounds. The results showed that when inoculated on the leaf wounds, all the mutants successfully infect the leaves and developed typical anthracnose lesions. However, the lesions caused by ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB were smaller than those of WT and ΔCgnoxA ( Figures 2A–C ). Moreover, when inoculated on the intact leaves, the disease incidence of ΔCgnoxR was significantly decreased compared with that of WT and ΔCgnoxA, whereas ΔCgnoxB and ΔCgnoxAnoxB could not even infect the leaves at all ( Figures 2D–F ). These results revealed that CgNOXB and CgNOXR play vital roles in the pathogenicity and especially the penetration ability of C. gloeosporioides to host plants.

The expression patterns of CgnoxA, CgnoxB, and CgnoxR during vegetative growth in vitro, appressorium formation, and colonization in plant leaves were investigated via a qRT-PCR assay. The results ( Figure 3 ) revealed that in appressoria, the transcription level of CgnoxB was about 2.5-fold higher than that during in vitro growth and in vivo colonization, whereas the transcription levels of CgnoxA and CgnoxR were about 0.5-fold lower than in the other stages. Meanwhile, the expressions of the three genes during in vivo colonization were all down-regulated compared with that during in vitro growth. These results enlightened the role of CgNOXB in the appressorium formation of C. gloeosporioides.

To explore the roles of NOX in appressorium formation, the germination rates and appressorium formation of the mutants were investigated. The mutants were cultured on hydrophobic plastic plates to induce appressorium formation. After incubation for 12 h, over 80% conidia of WT formed typical appressoria, whereas ΔCgnoxA, ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB showed decreases in appressorium formation, with about 73.6%, 51.7%, 47.6%, and 52.8%, respectively. After incubation for 24 h, approximately 82.3% and 75.0% conidia from WT and ΔCgnoxA formed appressoria; in comparison, the rates of ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB were only 52.8%, 65.5%, and 57.7%, respectively ( Figures 4A, B ). The results revealed that CgNOXs are required for appressorium development.

The formation of penetration peg was further investigated by incubation of the mutants on the onion epidermis. The result ( Figures 4C, D ) showed that, after incubation for 16 h, most of the appressoria of WT and ΔCgnoxA formed typical invasive hyphae (also named primary hyphae) and successfully penetrated the plant tissue. By contrast, appressoria of ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB failed to penetrate and invade onion cells. The results suggested that CgNOXB and CgNOXR are required for the penetration ability of C. gloeosporioides to plant.

The appressoria of the mutants were stained with DAB to investigate the ROS generation. After incubation for 24 h on onion epidermis, WT and ΔCgnoxA showed a little accumulation of dark-brown polymers around appressoria. ( Figures 5A, B ), whereas for the ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB mutants, the appressoria accumulated an amount of dark-brown polymers, with higher AOD levels than those of WT and ΔCgnoxA ( Figures 5A, B ). In addition, these three mutants did not form invasive hyphae as mentioned above. To further assess O₂•-, the direct product of NOXs, the hyphae of the mutants were stained with NBT. Microscopic observation showed that the WT and ΔCgnoxA accumulated blue formazan precipitate intensively in the apex of hyphal tips ( Figures 5C, D ), suggesting the polarity distribution of O₂•-, whereas for ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB, the O₂•- generation was uniformly distributed in the hyphal tips.

In our previous work, a widely used Lifeact-EGFP gene fusion was introduced into C. gloeosporioides to observe the organization of F-actin by live-cell imaging (Liu et al., 2021). To understand the roles of CgNOXs in F-actin organization, we then generated the ΔCgnoxA, ΔCgnoxB, and ΔCgnoxR and ΔCgnoxAnoxB mutants that express Lifeact-EGFP. Then the F-actin structure in mutants was investigated.

In the conidia of WT and the mutants, F-actin showed a typical filamentous network and patches, suggesting that the knockout of NOX genes did not influence the F-actin network in conidia ( Figure S1 ). In hyphal tips of WT and ΔCgnoxA, F-actin showed a polarized distribution with patches and cable structures ( Figure 6A ), whereas in ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB mutants, this kind of organized distribution was diminished ( Figure 6A ). To quantitate this polarity distribution of F-actin, the hyphal tips were divided into three regions of apex, subapex, and shank, and the relative fluorescence intensity was calculated. The results were in accordance with that of the microscope observation, revealing that WT and ΔCgnoxA employed a higher intensity at the apex and subapex regions than in the shank region, while ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB employed uniform intensity all through the hyphal tips ( Figure 6A ). To further verify whether this kind of F-action polarity was mediated by NOXs, the WT strain was treated with DPI, the inhibitor of NOXs. The microscopic analysis showed that DPI led to a decrease in fluorescence intensity in all three zones of hyphal tips; moreover, the polarity of F-actin was also diminished by DPI.

Then the F-actin structures in the appressoria were investigated. The results showed that in the appressoria of WT, F-actin was reorganized to a ring structure around the appressorium pore. This ring-shaped F-actin network was also observed in ΔCgnoxA ( Figure 7A ), whereas in ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB, F-actin showed a diffused distribution and did not form a ring structure in appressoria. Furthermore, DPI treatment also interfered with the organization of the F-actin network, resulting in a fuzzy ring structure with very low fluorescence intensity ( Figure 7A ). During incubation on onion epidermis for 15 h, WT and ΔCgnoxA formed typical appressoria and invasive hyphae; furthermore, F-actin in appressoria was diffused with low intensity, suggesting that there was a reorganization of the F-actin structure in appressoria after the formation of invasive hyphae ( Figure 7B ). In comparison, appressoria of ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB showed diffused F-actin structures of actin patches in conidia and appressoria; meanwhile, there were no invasive hyphae formed. In addition, to investigate whether the transcription of Lifeact-GFP was interfered in the mutants, a qRT-PCR analysis was conducted. The result showed that the relative expression levels of Lifeact-GFP in these mutants were all below twofold, suggesting that the knockout of NOX-coding genes did not influence the transcription of Lifeact-GFP ( Figure S2 ). These results demonstrate the important roles of CgNOXB and CgNOXR in F-actin organization.

To investigate the roles of CgNOXs in fungal cell wall synthesis, the hyphae of the strains were observed with Calcofluor white (CFW) staining. In WT, CFW fluorescence was intensively distributed at the apex regions of hyphal tips, indicating the polarized deposition of the cell wall material, whereas in ΔCgnoxA, the fluorescence is mainly located in subapex regions. Moreover, this kind of polarity was diminished in the hyphal tips of ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB, in which the fluorescence was uniformly distributed all through the hyphal tips ( Figure 8A ).

The protoplast release assay was conducted to assess the sensitivity of the cell wall to enzymatic degradation. The results showed that ΔCgnoxA released a similar number of protoplasts as WT after the treatment with lyases. However, ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB released fewer protoplasts than WT ( Figure 8B ). The results indicated the change in the cell wall composition of ΔCgnoxB, ΔCgnoxR, and ΔCgnoxAnoxB. Chitin synthases participate in the biosynthesis of chitin and are involved in the cell wall integrity of filamentous fungi (Yang and Zhang, 2019; Wang et al., 2021). Therefore, the expression changes of chitin synthases in the mutants were investigated. There were seven chitin synthases (CgCHS) identified in C. gloeosporioides (Wang et al., 2021). The qRT-PCR assay ( Figure 8C ) showed that in ΔCgnoxA, the relative expression levels of CgCHS2, CgCHS3, CgCHS4, CgCHS6, and CgCHS7 increased over twofold; meanwhile, all seven genes were upregulated in ΔCgnoxB. However, in ΔCgnoxR, all the CgCHS genes were dramatically down-regulated. These data suggest that CgNOXA, CgNOXB, and CgNOXR are all involved in the cell wall integrity in C. gloeosporioides.