The E2 ubiquitin-conjugating enzyme CfRad6 regulates the autophagy and pathogenicity of Colletotrichum fructicola on Camellia oleifera

Abstract

Anthracnose is a common disease found in Camellia oleifera producing areas across China, whose primary pathogen is Colletotrichum fructicola. We previously revealed that autophagy is essential for the pathogenicity of C. fructicola. However, the function of ubiquitin–proteasome system (UPS), which is a parallel protein degradation pathway to autophagy, remains elusive. Here, we report that CfRad6, an E2 conjugating enzyme in UPS, interacts with three putative E3 ubiquitin ligases, namely CfRad18, CfUbr1, and CfBre1. Importantly, we presented evidence showing that CfRad6 negatively regulates autophagy, revealing the first link between UPS and autophagy in pathogenic fungi. Targeted gene deletion showed that CfRad6 is involved in growth and conidiation. We further found that the ΔCfrad6 mutant is defective in appressoria formation and responses to environmental stresses. These combined effects, along with the abnormal autophagy level, lead to the pathogenicity defects of the ΔCfrad6 mutant. Taken together, our study indicates the pleiotropic functions of CfRad6 in the development and pathogenicity of C. fructicola.

Background

Camellia oleifera is a woody plant cultivated for its edible oil plant in the southern provinces of China (Wu et al. 2020). Anthracnose is one of the most important diseases of C. oleifera (Li et al. 2016). Our previous research has shown that several species of pathogenic fungi cause anthracnose on C. oleifera, with Colletotrichum fructicola being the most prevalent and pathogenic (Li et al. 2016; Jiang et al. 2018).

The autophagy and ubiquitin–proteasome system (UPS) are the main cellular quality control systems (Pohl and Dikic 2019). Recently, we revealed that autophagy is essential for the pathogenicity of C. fructicola (Guo et al. 2022; Zhang et al. 2022). Autophagy is a common and conserved process in which organelles and proteins are degraded and recycled in vacuoles (lysosomes) (Mizushima and Komatsu 2011; Ohsumi 2014). The UPS is a parallel protein degradation pathway to autophagy in eukaryotic cells (Ciechanover 2005). Furthermore, increasing evidence suggests that the ubiquitination of proteins is not limited to the function of degradation but also participates in many processes, such as intracellular transcription, DNA repair and cell cycle regulation (Mukhopadhyay and Riezman 2007; Ulrich and Walden 2010). However, the roles of UPS in C. fructicola remain elusive, and the relationship between UPS and autophagy in pathogenic fungi is also unknown.

The UPS pathway is composed of several components, including ubiquitin, E1 ubiquitin-activating enzyme, E2 ubiquitin-conjugating enzyme, E3 ubiquitin ligase, 26S proteasome and deubiquitinating enzyme. It is a major non-lysosome (vacuoles) pathway for degrading proteins (Jia et al. 2023). Typically, the ubiquitination process of a target protein involves a three-step sequential transmission reaction of the three ubiquitin enzymes of E1, E2, and E3 (Jentsch et al. 1987). E2 plays a critical role in this process by controlling the conversion of ubiquitin chains from initiation to extension, and determining the mission of the modified target protein by influencing the formation of ubiquitin chain topology (Imura et al. 2015; Stewart et al. 2016).

Previous research revealed that Rad6, an important ubiquitin-conjugating enzyme E2, plays an important role in DNA damage repair by ubiquitinating various target protein (Shukla et al. 2022). In yeast, Rad6 does not function alone but interacts with the other three E3 ubiquitin ligase proteins Bre1, Rad18, and Ubr1 to regulate mono-ubiquitination of histone proteins, DNA replication repair and ubiquitination substrate proteins in cells (Shi et al. 2016). Similarly, Rad6 in rice cells interacts with OsSgt1 to regulate protein degradation and DNA repair (Yamamoto et al. 2004). Moreover, when the soybean GmUBC2, the yeast Rad6 homologous gene, was expressed in Arabidopsis, it improved the tolerance of Arabidopsis cells to osmotic stress (Zhou et al. 2010; Zhang et al. 2018). The studies demonstrate that Rad6 plays a significant role in yeast and plant. Additionally, Rad6-mediated ubiquitination has been found to be crucial in phytopathogens, such as Magnaporthe oryzae, where it is required for development and pathogenicity (Shi et al. 2016). In Fusarium graminearum, Rad6 mediates the H2B ub1 level and deoxynivalenol (DON) production (Ma et al. 2021). The Fusarium verticllioides Ubl1, a homologous of Rad6-mediated E3 Ubr1, is also involved in its pathogenicity on hosts (Ridenour et al. 2014). However, the roles of Rad6 in autophagy and the relationship between UPS with autophagy are unknown in phytopathogens.

In the present study, we analyzed the Rad6-related ubiquitination pathway in C. fructicola and characterized its pleiotropic functions. We also provided evidence indicating the links between autophagy and the ubiquitination pathway.

Results

The phylogenetic analysis and the domain prediction of CfRad6

We identified CfRad6 in the database of C. fructicola by using the known ScRad6 amino acid sequence from Saccharomyces cerevisiae, and a phylogenetic tree of Rad6 proteins was constructed using Neighbor-Joining Algorithm. The result revealed that CfRad6 has the highest homology with C. gloeosporioides and is the most distant from the unicellular yeast (Fig. 1a). The CfRad6 protein is composed of 151 amino acids and contains a ubiquitin-coupled catalysis (UBCc) domain that spans amino acids 7 to 150 (Fig. 1b).

Fig. 1

Phylogenetic analysis of Rad6 protein in fungi and domain prediction of CfRad6. a Based on protein sequences of diverse fungi, the phylogenetic tree was constructed by Mega 11.0 program with the Neighbor-joining method. The scale is 0.02. The Latin names of the corresponding species and their NCBI accession numbers are as follows: C. fructicola (XP_007273365.1), C. gloeosporioides (KAH9242036.1), F. graminearum (EYB31908.1), M. oryzae (XP_003714754.1), A. oryzae (XP_001826971.1), A. nidulans (Q96UP5.1), S. cerevisiae (AJS07636.1). b Domain prediction of CfRad6 by SMART website. The oval (7–150 amino acids) represents the conserved UBCc domain

Targeted CfRAD6 gene deletion and complementation

In order to study the biological function of CfRad6 in C. fructicola, we utilized homologous recombination to replace the CfRAD6 coding region with the hygromycin resistance cassette (HPH) (Additional file 1: Figure S1a), and the resulting transformants were confirmed by PCR, RT-PCR (Reverse Transcription PCR) and RT-qPCR (Additional file 1: Figure S1b–d). In addition, the wild-type CfRAD6 gene was re-introduced into the knockout mutant, which restored all defects. We thus obtained the mutant ΔCfrad6 and the complemented strain ΔCfrad6/CfRAD6. We first examined the role of CfRad6 in total protein ubiquitination level but found that the ubiquitination level within ΔCfrad6 was similar to that of WT and ΔCfrad6/CfRAD6 (Additional file 1: Figure S2).

CfRad6 interacts with CfRad18, CfBre1, and CfUbr1 in C. fructicola

Rad6 is known to interact with the other three E3 ubiquitin ligase proteins Rad18, Bre1, and Ubr1 (Enserink and Kolodner 2012). In order to elucidate whether there is also an interaction between ubiquitin-conjugating enzyme E2 Rad6 and the three E3 ubiquitin ligases in C. fructicola, yeast two-hybrid (Y2H) assays were performed. The results revealed that yeast cells transformed with CfRad6/CfRad18, CfRad6/CfUbr1 and CfRad6/CfBre1 all showed growth on the SD-Leu-Trp-His-Ade medium, indicating that CfRad6 does indeed interact with CfRad18, CfUbr1, and CfBre1 (Fig. 2). This finding suggests that CfRad6 shares the same conserved interaction with E3 ubiquitin ligases in C. fructicola.

Fig. 2

CfRad6 interacts with CfRad18, CfBre1, and CfUbr1. The pGADT7 and pGBKT7 vectors, fused with the related genes, were co-transformed into the yeast strain AH109. Ten-fold serial dilutions of transformants were spotted in equal volume (5 μL) on synthetic two-deficient media (SD-Leu-Trp) and synthetic quadruple dropout selection media (SD-Leu-Trp-Ade-His). Leu: leucine; Trp: tryptophan: Ade, adenine; His: histidine

CfRad6 is involved in response to rapamycin

To investigate whether there is a link between CfRad6-related UPS and autophagy, we tested the role of CfRad6 in the response to rapamycin, an autophagy inducer that inactivates Tor (the target of rapamycin) pathway (Marroquin-Guzman and Wilson 2015; Wang and Wang 2015). We cultured the wide-type (WT), ΔCfrad6 and ΔCfrad6/CfRAD6 strains on PDA medium supplemented with 25 nM rapamycin. The results showed that the inhibition rate of ΔCfrad6 was significantly lower than that of WT and ΔCfrad6/CfRAD6 (Fig. 3). The result indicated that CfRad6 is involved in the response to rapamycin of C. fructicola.

Fig. 3

CfRad6 is involved in the response to rapamycin. a The WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were incubated on PDA plates with 25 nM rapamycin at 28℃ for 3 days. b Statistics analysis of growth inhibition rates of the strains to rapamycin stress; Error bars represent the standard deviation (SD) of three replicates, and asterisks indicate the difference is significant (P < 0.01)

CfRad6 negatively regulates the autophagy process

To further investigate the role of CfRad6 in autophagy, the GFP-tagged CfATG8, which was used as a marker for autophagy in our previous studies (Guo et al. 2022; Zhang et al. 2022), was transformed into the WT and ΔCfrad6 mutant strains. When grown in CM medium, the ΔCfrad6 mutant exhibited more autophagosomes than WT. After MM-N induction for 4 h, most of the GFP fluorescence was transferred into the vacuoles, both in WT and ΔCfrad6, but WT still showed some cytoplasmic GFP fluorescence and significantly more autophagosomes than the ΔCfrad6 mutant (Fig. 4a,b). We further evaluated autophagic flux by performing GFP-CfAtg8 proteolysis assays through western blotting. In CM condition, WT expressing GFP-CfATG8 exhibited lower amounts of free GFP than full-length GFP-CfAtg8, but the ΔCfrad6 mutant expressing GFP-CfATG8 exhibited comparable free GFP to GFP-CfAtg8 (Fig. 4c), suggesting the higher autophagy level in the ΔCfrad6 mutant under rich nutrients. When induced in MM-N for 4 h, WT showed comparable free GFP to GFP-CfAtg8 while ΔCfrad6 showed higher free GFP than GFP-CfAtg8 (Fig. 4c). Additionally, the autophagic flux was also quantitatively calculated by measuring the GFP relative to GFP + GFP-CfAtg8 ratio. The ratios of ΔCfrad6 were both higher than that in WT under CM and MM-N conditions (Fig. 4c). These results suggest that CfRad6 plays a negative role in autophagy.

Fig. 4

CfRad6 negatively regulates autophagy. a Micrographs of GFP-CfAtg8 labeled autophagosomes in the WT and ΔCfrad6 mutant strains. b Statistical analysis of autophagosome numbers in the WT and ΔCfrad6 mutant strains. On average, more than 50 hyphal tips (about 50 μm in length) were observed. The assays were repeated three times, and the results were similar. c Immunoblot analysis of GFP-CfAtg8 proteolysis in the WT and ΔCfrad6 mutant. The upper lanes point to the full GFP-Atg8 (46 kDa), and the lower lanes point to the free GFP (26 kDa). The autophagy level was estimated by calculating the amount of free GFP relative to the total amount of full GFP-CfAtg8 plus free GFP. d Micrographs of GFP-CfAtg8 labeled autophagosomes in WT with or without MG132 treatment. e Statistical analysis of autophagosome numbers in WT with or without MG132 treatment. f Immunoblot analysis of GFP-CfAtg8 proteolysis in the WT with or without MG132 treatment

We next tested whether CfRad6 plays a direct role in autophagy by interacting with CfAtg8. However, Y2H assays revealed that yeast cells transformed with CfRad6/CfAtg8 did not grow on the SD-Leu-Trp-His and SD-Leu-Trp-His-Ade medium (Additional file 1: Figure S3a), indicating that there was no direct interaction between CfRad6 and CfAtg8. BiFC assays also confirmed that CfRad6 did not interact directly with CfAtg8 in autophagy induction conditions (Additional file 1: Figure S3b).

The studies conducted in Homo sapiens showed that proteasome malfunction activates the autophagy pathway (Zheng et al. 2011; Ji and Kwon 2017; Su and Wang 2020). We thus use the 26S proteasome inhibitor MG132 (z-Leu-Leu-Leu-CHO) to verify its roles in the autophagy of C. fructicola. Although the treatments with MG132 showed similar autophagosome numbers to those in CM conditions (Fig. 4d,e), the autophagy level of the treatments with MG132 was higher than in CM conditions (Fig. 4f).

CfRad6 is important for growth and conidiation

To investigate whether CfRad6 participates in regulating the vegetative growth of C. fructicola, the WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were cultured on PDA and MM plates for 3 days and measured their colony diameters. The results showed that the colonies of ΔCfrad6 on PDA and MM medium were significantly smaller than those of WT and complemented strains (Fig. 5a,b). The results showed that CfRad6 is involved in regulating the vegetative growth of C. fructicola.

Fig. 5

CfRad6 is important for growth and conidiation. a The growth of WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains in PDA and MM medium for 3 days. b The diameters of strains were measured and statistically analyzed. Asterisks mean the difference is significant (P < 0.01). c Statistics analysis of conidia

Conidia are crucial for infecting host cells of C. fructicola. To investigate whether CfRad6 affects the conidiation of C. fructicola, we quantified the conidia of WT, ΔCfrad6, and ΔCfrad6/CfRAD6 that were cultured on a 28°C and 180 rpm shaker for 2 days. The results showed that ΔCfrad6 produced significantly fewer conidia than WT and ΔCfrad6/CfRAD6 (Fig. 5c). These results indicate that CfRad6 is important for the conidiation of C. fructicola.

CfRad6 is required for the full pathogenicity of C. fructicola

We are most concerned about the contribution of CfRad6 to the pathogenicity of C. fructicola, so we inoculated the WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains on the leaves of wounded and healthy C. oleifera leaves. The ΔCfrad6 mutant only formed very small disease spots on the wounded leaves but did not develop lesions on the healthy leaves, in contrast to the typical and large lesions produced by the WT and ΔCfrad6/CfRAD6 strains (Fig. 6a–d). To further investigate the specificity of CfRad6 in pathogenicity, we also inoculated the above strains on apples and found that the lesions caused by the ΔCfrad6 mutant were significantly smaller than those formed by the WT and ΔCfrad6/CfRAD6 strains (Fig. 6e,f). These results suggest that CfRad6 is required for the full pathogenicity of C. fructicola on its hosts.

Fig. 6

CfRad6 is required for full pathogenicity. a The WT, ΔCfrad6, ΔCfrad6/CfRAD6 strains were inoculated on wounded C. oleifera leaves. b The WT, ΔCfrad6, ΔCfrad6/CfRAD6 strains were inoculated on healthy C. oleifera leaves. c The disease spot areas of strains on wounded C. oleifera leaves were measured by ImageJ and statistically analyzed. d The disease spot areas of strains on healthy C. oleifera leaves were measured by ImageJ and statistically analyzed. e The strains were inoculated on wounded apples and photographed. f Statistical analysis of disease spots on apples. CK: compared control. Asterisks indicate the difference is significant (P < 0.01)

CfRad6 mediates the appressoria formation and lipid and glycogen metabolism of C. fructicola

Appressoria play a crucial role in the infection of hosts by Colletotrichum spp. (Fukada and Kubo 2015; Fukada et al. 2019; Ryder et al. 2022). To investigate the underlying mechanism of the pathogenicity defects of ΔCfrad6, we examined the appressoria formation of WT, ΔCfrad6, and ΔCfrad6/CfRAD6 after they were cultured on hydrophobic glass slides for 24 h. The result revealed that ΔCfrad6 showed a significant decrease in germination rates for ΔCfrad6 compared to WT and ΔCfrad6/CfRAD6 (Fig. 7a,b). In addition, we also observed that ΔCfrad6 hardly formed appressoria, unlike the more than 70% appressoria formation rate for WT and ΔCfrad6/CfRAD6 (Fig. 7a,b). This suggests that CfRad6 mediates the appressoria formation of C. fructicola.

Fig. 7

CfRad6 mediates the appressoria formation and lipid transfer. a The conidia of WT, ΔCfrad6, ΔCfrad6/CfRAD6 strains were cultured on hydrophobic glasses for 48 h, and appressoria formation were observed. b Statistical analysis of appressoria formation rates. Asterisks indicate the difference is significant (P < 0.01). Bar = 40 μm. c Lipids in conidia and appressoria of the WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were stained with Nile red at different time points. Bar = 20 µm

The lipids and glycogen in conidia are critical sources of nutrients for appressorium (Wang et al. 2007; Li et al. 2017). We used Nile red to stain lipids and found that lipids began transferring from conidia to appressoria at 8 hpi and almost disappeared at 48 hpi in WT, but lipids were still clearly observed in conidia and germ tubes of ΔCfrad6 (Fig. 7c). Similar results were also found in the glycogen staining with KI/I₂ (Additional file 1: Figure S4)_. This indicates that CfRad6 mediates the lipid and glycogen metabolism of C. fructicola.

CfRad6 is involved in response to environmental stresses

The endoplasmic reticulum (ER) is the site of intracellular protein modification, and studies have revealed that pathogens face host-derived ER stress during infection (Tang et al. 2020; Yin et al. 2020). To analyze whether CfRad6 is involved in responding to ER stress, the WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were cultured on PDA medium with 5.0 mM DTT. The results showed that the inhibition rate of ΔCfrad6 was significantly higher than that of WT and ΔCfrad6/CfRAD6 (Fig. 8a,b). This indicates that CfRad6 mediates the response of C. fructicola to ER stress.

Fig. 8

CfRad6 is involved in responses to environmental stress. a The WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were inoculated in PDA, and PDA supplemented with 5.0 mM DTT for 3 days. b Statistical analysis of inhibition rates of the strains to endoplasmic reticulum stress against untreated control. c The WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were inoculated in PDA and PDA plus with 400 μg/mL CR and 0.1% SDS for 3 days. d Statistical analysis of inhibition rates of the strains to cell wall integrity stress against untreated control. e The WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were inoculated in PDA and PDA plus with 0.7 M NaCl and 0.6 M KCl for 3 days. f Statistical analysis of inhibition rates of the strains to osmotic stress against untreated control. Error bars represent the SD of three replicates, and asterisks indicate the difference is significant (P < 0.01)

The cell wall is the first barrier between the fungi and the environment, acting as a defense against external stress (Levin 2011; Malavazi et al. 2014). To assess whether CfRad6 plays a role in the cell wall integrity stress response, the WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were cultured on PDA medium with 400 μg/mL CR and 0.1% SDS. The results showed that the ΔCfrad6 mutant had a significantly lower inhibition rate on CR than the WT and ΔCfrad6/CfRAD6 strains and a significantly higher inhibition rate on SDS (Fig. 8c,d). Additionally, we observed that the ΔCfrad6 mutant are more sensitive to osmotic stress (Fig. 8e,f). These findings indicate that CfRad6 plays a role in responding to cell wall integrity stress and osmotic stress.

Discussion

In the present study, we identified an E2 protein Rad6 in C. fructicola, and found that CfRad6 interacts with three E3 proteins to regulate the growth, conidiation, appressoria formation, stress responses, autophagy, and pathogenicity of C. fructicola.

Similar to the homologs of Rad6 proteins in S. cerevisiae and M. oryzae, CfRad6 contains a ubiquitin-coupled catalysis (UBCc) domain (Shi et al. 2016). By combining the conserved UBCc domain with phylogenetic analysis, we concluded that there is a sequence similarity among Rad6 proteins. Additionally, we found that CfRad6 directly interacts with CfRad18, CfUbr1, and CfBre1, which is consistent with the facts that Rad6 functions in conjunction with three E3 proteins in yeast and the rice blast fungus (Enserink and Kolodner 2012; Shi et al. 2016). Therefore, we proposed that there is also functional similarity among Rad6 proteins.

As expected, the deletion of CfRAD6 caused a decrease in the growth rate, which is consistent with the positive roles on growth of its homologs in Caenorhabditis elegans and S. cerevisiae (Enserink and Kolodner 2012; Skibinski and Boyd 2012). Additionally, the ΔCfrad6 mutant also showed a similar reduction in conidial production to the RAD6-deleted strain in M. oryzae (Shi et al. 2016). Together, these results suggest that Rad6 proteins not only share similarity in amino acid sequence but also have conserved functions in growth and conidiation.

Our findings suggest that CfRad6 is involved in response to autophagy inducer rapamycin, indicating its involvement in autophagy, which is further corroborated by the increased autophagosome abundance in the ΔCfrad6 mutant. Moreover, the immunoblot analysis also revealed a higher autophagy level in the ΔCfrad6 mutant. Thus, we hypothesized that the abnormal UPS could activate the autophagy pathway, which is in line with studies showing that proteasome inhibition or proteasome functional insufficiency can activate compensatory autophagy in Homo sapiens (Cha-Molstad et al. 2015; Wang and Wang 2015; Bao et al. 2017). However, CfRad6 did not directly interact with CfAtg8. Further studies are warranted to illustrate how the ubiquitin system links autophagy. A recent study in M. oryzae has shown that Ubp3-mediated de-ubiquitination regulates ribophagy but not autophagy (Cai et al. 2022). Nonetheless, we have demonstrated the crosstalk between UPS and autophagy in phytopathogenic fungi for the first time. Recently, we revealed the essential roles of autophagy in the pathogenicity of C. fructicola (Guo et al. 2022; Zhang et al. 2022). The disordered autophagy flux might foretell its defects in pathogenicity.

We have found that ΔCfrad6 is non-pathogenic on the healthy leaves of C. oleifera and shows decreased pathogenicity on the wounded leaves of C. oleifera or apples. We reason that the pathogenicity defects of ΔCfrad6 are directly caused by the decreased appressoria formation rate. Furthermore, the irregular response to environmental stresses of ΔCfrad6, especially the ER stress, which is likewise derived from the host under pathogenic fungi infection (Tang et al. 2020; Yin et al. 2020), may also be a reason for the pathogenicity defect of ΔCfrad6. Intriguingly, the non-pathogenicity caused by ΔCfrad6 on healthy leaves showed a slight difference with the study in M. oryzae, which revealed that the MoRAD6 gene deletion mutant still formed disease symptoms (Shi et al. 2016). This slight contradiction also occurred in our previous studies on CfGcn5. We found that the CfGCN5 gene deletion mutant showed non-pathogenicity on healthy C. oleifera leaves (Zhang et al. 2021, 2022), whereas the deletion strain in M. oryzae still showed lesions on the host, although the lesions were decreased compared to WT (Zhang et al. 2017). Therefore, we presume that it may be because the epidermis of C. oleifera leaves is thicker than that of rice leaves, resulting in ΔCfrad6 failing to infect C. oleifera leaves, but further studies are highly warranted.

Conclusions

Our study revealed that CfRad6 shows pleiotropic functions in C. fructicola, playing an important role in growth, conidiation, appressoria formation, stress response, autophagy and pathogenicity of C. fructicola. Therefore, CfRad6 can be considered as a potential target for fungicide development to control anthracnose.

Methods

Strains and culture conditions

The strain that was used as wild-type (WT) for transformation in this study was CFLH16. The WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were cultured on PDA or MM plates at 28℃. The liquid PDB was used to prepare mycelia for DNA extraction. The liquid CM was used to prepare mycelia for protein extraction.

Phylogenetic tree construction and domain prediction

The E2 CfRad6 protein in C. fructicola was identified using the known ScRad6 amino acid sequence in yeast as a bait, and its homologous sequences in C. gloeosporioides, F. graminearum, M. oryzae, A. oryzae, A. nidulans, and S. cerevisiae were obtained from NCBI database (https://www.ncbi.nlm.nih.gov/) and constructed the phylogenetic tree by Mega 11.0 program with the Neighbor-joining method. The domain prediction of CfRad6 was performed by SMART website (http://smart.embl-heidelberg.de/).

Gene deletion and complementation

The CfRAD6 gene deletion was performed using the standard one-step strategy (Zhang et al. 2019, 2021; Li et al. 2021). First, we overlapped the two ~ 1.0 kb sequences flanking CfRAD6 to the two sides of the HPH gene (~ 1.0 kb). Then, we amplified and transformed the ~ 3.4 kb fragments containing the flanking sequences and hygromycin resistance cassette into protoplasts of WT. Finally, the complement fragments were amplified by PCR with primers (Additional file 2: Table S1) and inserted into pYF11 (bleomycin resistance) to complement the mutant strain.

RT-qPCR analysis

Total RNA was extracted from WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains cultured on liquid CM for 36 h, and then the total RNA was quality controlled and transcribed to cDNA. The RT-qPCR with primers (Additional file 2: Table S1) was run on an ABI QuantStudio 3, and the relative quantification of transcription was normalized to the stable expression ACTIN gene. The experiment was repeated three times, and each experiment used independent biological replicates.

Growth, conidiation and appressoria formation assays

The strains of WT, ΔCfrad6, and ΔCfrad6/CfRAD6 were cultured on PDA and MM plates for 3 days, then the colony diameters were measured and statistically analyzed. All experiments were repeated three times, each with three replicates. For conidia production assays, mycelia were grown on liquid PDB in an incubation shaker for 48 h, then were filtered with three layers of lens paper, and conidia were collected and quantified. For appressoria formation assays, droplets of conidial suspension were dripped on hydrophobic glasses and incubated at 28°C for 24 h, followed by observation under the microscope (ZEISS, Axio Observer. A1).

Nile red and glycogen staining

For glycogen staining, the conidia of WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were incubated on hydrophobic glasses for different time points and stained with a solution consisting of 120 mg/mL of KI and 20 mg/mL of I2 in distilled water for 30 min. Yellowish brown deposits can be observed in the bright field immediately. For lipid bodies staining, samples above were stained with Nile red (N8440 Solarbio) to detect neutral lipids for 5 min, then orange-red was visible under a fluorescent microscope. The Nile red was used at 10 µg/mL (stock 1 mg/mL in ethanol) in PBS buffer. All samples were examined and photographed with the fluorescent microscope (ZEISS, Axio Observer. A1).

Stresses response analysis

The WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were cultured on PDA, and PDA supplemented with 25 nΜ rapamycin, endoplasmic reticulum stress (5.0 mM DTT), cell wall integrity stress (0.1% SDS, 400 μg/mL CR) and osmotic stress (0.7 M NaCl, 0.6 M KCl) for 3 days, then the colony diameters were measured, and the inhibition rates were statistically analyzed.

Pathogenicity assays

The WT, ΔCfrad6, and ΔCfrad6/CfRAD6 strains were inoculated onto the edge of wounded and healthy tea-oil leaves, moisturizing culture in an incubator for 2 days, then its lesions were observed, photographed and analyzed. For the pathogenicity on wounded apples, the apples were first punched into several 4-mm-diameter holes, then the same strains were inoculated onto it. The apples were cultured for 4–5 days, and then the lesions were observed and photographed.

Yeast two-hybrid and BiFC assays

The yeast two-hybrid assay was performed according to the instructions of Alkali-Cation ™ Yeast Transformation Kit. The coding sequences of each candidate gene were amplified with primers listed in Table S1 from cDNA library of WT, inserted into the bait vector pGBKT7 digested by BamHI and prey vector pGADT7 digested by BamHI, respectively, and then ligated to the linearized plasmid vector using the one-step ligase of Novozymes, and the plasmid was extracted in E. coli. After the sequencing, pairs of plasmids were co-transformed into the yeast strain AH109 and then cultured on SD-Leu-Trp medium. After growing colonies, it was transferred on SD-Leu-Trp-Ade-His medium. The positive control is ADRECT + BD ( +), and the negative control is ADRECT + BD (-).

For the BiFC assays, the pHZ68-CfAtg8 and pHZ65-CfRad6 vectors were constructed and used to co-transform the WT strains. The WT strains co-expressing pHZ68-CfAtg8 and pHZ65-CfRad6 were cultured in liquid CM medium for 36 h and washed with sterile water, followed by incubation in MM-N for 4 h and 8 h. The fluorescent signals of the hypha were observed under the fluorescent microscope.

Autophagy induction assays

The WT and ΔCfrad6 strains that expressed the GFP-CfATG8 gene were cultured on liquid CM medium for 36 h, and mycelium pellets of that were washed with sterile water, followed by incubated in liquid MM-N or liquid CM plus with 50 µM MG132 for 4 h. The autophagosomes and GFP signals of the hypha were observed under the fluorescent microscope and analyzed.

Western blot analysis

The proteins were exacted from the above-treated mycelium pellets according to our previous method (Zhang et al. 2022), then proteins were analyzed by western blotting with the primary antibodies anti-GFP (rabbit, 1:5,000, Abways, AB0045) or anti-ubiquitin (rabbit, 1:1,000, Zenbio, R26024), and the secondary antibodies HRP-labeled goat anti-rabbit lgG (H1L) (1:10,000, Abways, AB0101). Finally, an Omni-ECL Femto Light Chemiluminescence Kit (EpiZyme, SQ201) was used to detect the proteins, and the gray analysis was carried out with ImageJ software.

Statistical analysis

All statistical data were performed using a one-way ANOVA (Analysis of Variance) with Duncan’s new multiple range test.