The DNA and protein sequences from B. cinerea B05.10 were downloaded from the Ensembl Fungi Botrytis database (Amselem et al., 2011; Staats and van Kan, 2012; van Kan et al., 2017).¹ Other fungal sequences were from the database of the National Centre for Biotechnology (NCBI).² The amino acid sequences of the putative orthologues genes were analyzed using the BlastP algorithm (Altschul et al., 1990) at the NCBI. The functional protein domains of BAC protein were predicted using Interpro and SMART.^3,4 Additionally, the phosphorylation site of BAC was predicted using Netphos-3.1.⁵ The analysis of WCC binding sites was conducted according to the previous reports (Froehlich et al., 2003; Wang et al., 2015; Baek et al., 2019).

Supplementary Table S1 lists the wild-type and mutant strains of B. cinerea used in the present study. The wild type of B. cinerea strain B05.10 (Büttner et al., 1994), the BAC single mutant strain bac^S1407P, and the complemented strain bac^P1407S (Chen et al., 2020) were used in the present experiments. The bac^S1407A and bac^S1407D strains were generated in this work by introducing homologous recombination fragments into the B05.10 strain according to the previous method (Schumacher et al., 2012).

All strains of B. cinerea were ordinarily cultivated on solid potato dextrose agar (PDA) or complete medium (CM), the PDA medium for strain subculture, and the CM medium were used to avoid unknown nutrients (Pontecorvo et al., 1953). The strains were grown at 23°C using Percival incubators, which were equipped with cool white light fluorescent tubes (light intensity up to 100 μM/m²/s; wavelength 400–720 nm) at 12-h intervals of light and dark (12:12-h LD).

The growth rate of the mycelial colony was determined by measuring the increase of the colony diameter per day for each strain growing on the solid medium in a Petri dish. To determine the yield of conidia, the colonies cultured for 14 days were washed twice with 5 ml of water to harvest the conidia suspension, which was subsequently filtered by four layers of gauze to remove the mycelial debris. The conidia suspended in the final filtrate (10 ml) were counted under a hemocytometer.

The PP2C and AC domain-coding sequences of BAC and BAC^S1407P were isolated from the wild-type and mutant strains and inserted into pET28a+ (Supplementary Figure S4) to produce the BAC-His₆ PP2C-AC and BAC^S1407P-His₆ PP2C-AC vectors, respectively. These vectors were transformed into Escherichia coli BL21 cells by a heat-shock method. Target proteins were produced by the addition of 1.5 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) at 30°C. The BAC PP2C-AC domain and BAC^S1407P PP2C-AC domain were individually linked to six histamines (His6) and expressed as fusion proteins; eventually, the BAC-His₆ PP2C-AC domain and BAC^S1407P-His₆ PP2C-AC domain fusion proteins were produced and then purified using the Ni-NTA 6FF Sefinose (TM) Resin Kit (Shenggong, Shanghai, China).

For the gel retardation assay, the purified protein was run in SDS-PAGE or mixed with Phos-tag (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) to examine the phosphorylation levels of different proteins following the reported method (Gou et al., 2015). In this assay, the phosphate group on the protein binds the Phos-tag with the manganese ion in the gel. Eventually, the relative mass of the protein becomes larger, and therefore, the mobility in the gel becomes slower. The variance of the position of the protein is representative of the phosphorylation levels between different proteins. The fusion proteins were specified by binding with anti-His₆ and secondary antibody goat anti-mouse IgG HRP (AB-M-M100, GOOD HERE, Hangzhou, China) according to the manufacturer’s instructions. This binding was detected by the Western Blot test reagent dye solution (34,580, Thermo Scientific™, Shanghai, China), where the second antibody was bound to HRP, and the substrate of ECL produced chemiluminescence after being catalyzed by HRP.

For the total protein phosphorylation level test, the mycelia growth on cellophane-covered CM medium for 3 days under light or dark conditions, and the protein extraction buffer (Yang et al., 2013) was used to extract mycelia protein. The anti-phosphoserine antibody (ab9332, Abcam, United Kingdom) was used to analyze the total phosphorylation level.

The adenylyl cyclase activity assay was performed as described previously with some modifications (Korkhov and Qi, 2019). The reaction mixtures contained 50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 0.1% digitonin, 5 mM MnCl₂, 10 nM to 100 μM total ATP (Shenggong, Shanghai, China), and 0.01 mg/ml purified BAC. The reactions were started by adding ATP to the protein. The reactions were incubated for 10–30 min at 23°C and were terminated by treatment at 85°C for 10 s. The cAMP content was measured using a highly sensitive ELISA kit (Biosamite, Shanghai, China) according to the manufacturer’s instructions.

Mutants of B. cinerea were constructed using protoplasts by the method reported previously (Schumacher et al., 2012). The mutant construction strategy is shown in Figure 1A and Supplementary Figure S4. The bac^S1407A and bac^S1407D homologous recombination fragments were constructed by PCR-amplifying their coding regions from chromosomal DNAs using 5′- and 3′-corresponding primers and recombining them with a NAT (nourseothricin resistance cassette) of the pNAN-OGG. These products were transferred to the protoplasts of B. cinerea to generate the mutants mentioned above. The pNAN-OGG was used as a template to amplify the NAT fragments. The recombinant sequences were detected by PCR amplification with specific primer sets, as shown in Figure 1A; Supplementary Figure S4. Supplementary Table S2 shows the sequences of these primers.

The pathogenicity assay was performed as described previously (Chen et al., 2020). The healthy mature fruits of grape (Vitis vinifera), apple (Malus pumila Mill), tomato fruit (Solanum lycopersicum), 4-week-old tomato leaves, and 2-week-old A. thaliana leaves were used in the inoculation assay. Prior to inoculation, the fruits were rinsed in sterile water, submerged in 70% alcohol for the 30 s, and soaked in sterile water for 2 min. The 10 μl droplets of conidial suspensions (1 × 10⁶/ml GB) were dropped onto the inoculation site of the fruit and leaves surface. The inoculated fruits and leaves were placed in a box to keep 90% relative humidity. The lesion area of the inoculation site has been measured via Image J software.

The banding pattern denotes the alternated, ring-shaped grayish and white regions formed on radially growing mycelial colonies in the presence and absence of light illumination, respectively (Canessa et al., 2013). In order to observe the banding pattern, all strains were incubated for 7 days under LD conditions on the solid CM medium supplemented with 0.02% SDS, which reduces the mycelial growth to approximately 50% of radial growth on the SDS-free medium. This SDS treatment was conducted to make the banding pattern more distinct.

The banding patterns of different strains were further analyzed using a race tube method. The race tube was a transparent cylindrical tube (inner diameter, 9 mm) with air-permeable plugs at both ends. A solid medium was laid inside the tube and placed horizontally, and mycelia clumps of the test strains were separately placed at one end of the tube and incubated at 23°C to examine the change in the banding pattern under different conditions.

To observe the effect of the cAMP signal pathway on the banding pattern, an inhibitor of the cAMP decomposing enzyme PDE, 3-isobutyl-1-methylxanthine (IBMX) (Agarwal et al., 2010) was dissolved in the CM medium with Dimethyl sulfoxide (DMSO) to final concentration 10 μM in solid CM medium and used for the culture of the strains at 23°C under LD condition for a week.

For isolating RNAs from the strains (B05.10, Δbcwcl1, Δbcwcl1-com, and bac^S1407P), they were grown on cellophane membrane-covered PDA plates under the alternation of 12 h-light and 12 h-dark conditions. Samples were harvested in a temperature-controlled (23°C) darkroom equipped with low-intensity red-safety lights. The collected mycelial samples were immediately frozen in liquid nitrogen and subsequently kept at −80°C until further usage.

To test protein–protein interactions in yeast cells, the CDS of BcpkaR, Bcwcl1, and Bcfrq1 were cloned to pGBKT7 and pGADT7 vectors and then transferred to yeast strain AH109, which has a tryptophan (Trp) and leucine (Leu) deficiency for screening the transformants. Namely, the cells transformed with the constructed pGADT7 and pGBDT7 plasmids were able to grow on the medium lacking His, Ade, Trp, and Leu. In this experiment, the colonies from SD/−Trp/−Leu (-LW) plates were transferred to SD/-Trp//-Leu/-His//-Ade/3-amino-1,2,4-triazole (3-AT) (-LWH/) plates to confirm protein–protein interactions under 28°C for 1 week.

To test the binding activity of transcription factors on the promoters of target genes, the CDS of Bcwcl1 was cloned to pB42AD, and the promoters of Bcfrq1, Bcltf1, Bcltf2, and Bcltf3 were cloned to pLacZ vector and then transferred to yeast strain EGY48. The yeast was cultured at 30°C until it grew monoclonal at SD/−Trp/-Ura plates. Moreover, the colonies from SD/−Trp/-Ura plates were transferred to SD/−Trp/-Ura plates containing X-gal for culture at 30°C for 2–3 days to confirm the binding activities of transcription factors to promoters, and the combined yeast colonies that can be transcriptionally activated will turn blue.

For isolating RNAs from the strains, the cellophane membrane-covered PDA plates were used to culture strains. The collected mycelial samples were immediately frozen in liquid nitrogen and subsequently kept at −80°C until further usage. Frozen mycelia were ground to powder in liquid nitrogen, and total RNA was isolated using TRIzol reagent (Invitrogen) as described by Canessa et al. (2013). Total RNA quantity and quality were verified using NanoDrop (Thermo Scientific). All experiments were replicated three times. The expression of different genes in different strains at different times under LD conditions was compared with the expression of 3 h Bcfrq1 in wild type (B05.10) according to the method described previously (Hevia et al., 2015).

Statistical data were expressed as means ± standard errors (SE) from three repetitions. Bar charts represent mean values with standard deviations, using Tukey’s honestly significant (HSD) test to examine if differences between groups of samples were significant at a p-value of <0.01.

Our previous study demonstrated that mutation at the S1407 residue in BAC of B. cinerea caused obvious defects in growth, development, and pathogenicity (Chen et al., 2020). Here, we used NetPhos-3.1 to predict that the BAC S1407 site was located in the type 2C serine/threonine phosphatases (PP2C) domain of BAC (Figure 2A). The present prediction analysis showed that S1407 was a phosphorylation site of BAC (Supplementary Figure S1). Simultaneously, a comparison of the protein sequences of AC among different fungi by sequence alignment showed that the S1407 was a highly conserved residue (Figure 2B).

To verify the possibility of S1407 as a phosphorylation site, we expressed the BAC-His₆ PP2C-AC domains (subsequently referred to as BAC) and BAC^S1407P-His₆ PP2C-AC domains (subsequently referred to as BAC^S1407P) using a prokaryotic expression system. Western blot analysis after the protein phosphorylation level assay showed that the position of the BAC band was higher than that of BAC^S1407P in SDS-PAGE with Phos-tag; however, the bands were parallel in SDS-PAGE results (Figure 2C). These results indicate that mutations at the S1407 site in the PP2C domain of BAC affected the overall phosphorylation levels of BAC. By adding ATP substrates with different concentration gradients, the change curves of cAMP synthesized by BAC and BAC^S1407P were plotted (Figure 2D). The results showed that mutation at the S1407 site of BAC affected the ability of the BAC to convert ATP to cAMP. Taken together, the aforementioned results demonstrate that mutation of the S1407 site in the PP2C domain altered the phosphorylation level of BAC and affected its ability to convert ATP into cAMP.

Replacing the serine residue with alanine can cause this site to dephosphorylate continuously, while mutation to aspartic acid leads to continuous phosphorylation (Kamada et al., 2010). BAC S1407 phosphomimetic strain bac^S1407D, and BAC S1407 phosphodeficient strain bac^S1407A, were obtained by homologous recombination (Figures 1A,B). The intracellular cAMP contents of bac^S1407P and bac^S1407D were lower than those of the wild type, bac^S1407A, and bac^P1407S (Figure 1C). This is consistent with the results observed with the phosphorylation levels of BAC. In addition, the growth rates of the wild-type (B05.10) and mutant strains (bac^S1407P, bac^S1407D, bac^P1407S) were observed in CM medium supplemented with or without exogenous cAMP. The growth rate of bac^S1407P and bac^S1407D was slightly restored by exogenous cAMP (Supplementary Figure S2). After 7 days of culture, the colony morphology of bac^S1407P and bac^S1407D was recovered to some extent (Supplementary Figure S2). These results indicate that the S1407 site of BAC is a phosphorylation site, and the phosphorylation level of this site can affect the BAC activity of cAMP synthesis and mycelial growth.

PP2C domain is a domain that can remove the phosphates from substrate protein. The mutation of the S1407 site of the PP2C domain of BAC affects the phosphorylation of BAC and the ability of BAC to synthesize cAMP. It may also affect the phosphorylation level of substrate protein of the PP2C domain. We detected the total protein phosphorylation level of strains B05.10, bac^S1407P, bac^S1407A, and bac^S1407D (Figure 3). The results showed that the total protein phosphorylation level of bac^S1407P and bac^S1407D was significantly higher than that of wild type, and the phosphorylation levels of some protein band increased significantly in bac^S1407P and bac^S1407D strains under light and dark conditions (Figure 3), which indicated that the PP2C domain of BAC might have the ability to remove the substrate protein phosphorylation level.

Due to mutation at the S1407 residue, the phosphorylation level and function of BAC are greatly affected. The growth rate of bac^S1407D was significantly lower than that of the wild type and even lower than that of bac^S1407P. In addition, bac^S1407D mutant continuously produced conidia under LD conditions (Figures 4A–D). The pathogenicity and response to light of bac^S1407D were weaker than those of bac^S1407P, while the pathogenicity, conidiation, and sclerotia production and light responses of the bac^S1407A mutant were similar to those of the wild type (Figures 4E,F). The above results prove that the phosphorylation level of BAC can regulate the production of conidia and sclerotia. The light signaling pathway is also involved in regulating the development of conidia and sclerotia. We compared the conidia-producing phenotype of the Δbcwcl1 mutant with the BAC S1407 site mutant strains and found that the yield of conidia increased in Δbcwcl1 but decreased in bac^S1407P and bac^S1407D (Supplementary Figure S2C). The cAMP signaling pathway of B. cinerea may be involved in the regulation of photomorphogenesis, which is crucial for the formation of conidia and sclerotia of B. cinerea.

We further explored the effects of varied site mutations of BAC S1407 on growth rate under different light conditions. The colony appearances of different strains grown on CM medium under Blue light (BL), Red light (RL), DD, and LL conditions are shown (Figure 5A). The growth rates of bac^S1407P and bac^S1407D were significantly lower than those of the wild type under BL, RL, LL, and DD (Figure 5B). Interestingly, the growth rates of bac^S1407P and bac^S1407D under BL and LL conditions were higher than those of RL and DD (Figure 5B), suggesting that cAMP participates in mediating the response to light in B. cinerea.

The conidia of B. cinerea are important propagules for infection and transmission, and their formation is regulated by light (Williamson et al., 2007). A study on the interaction between B. cinerea and A. thaliana showed that the pathogenicity of B. cinerea was significantly higher at dusk than at dawn, and this phenomenon is regulated by the circadian clock of B. cinerea (Hevia et al., 2015). In the bac^S1407P strain, the yield of conidia was significantly affected, and the ability to produce sclerotia was lost. The conidia of wild type (B05.10) and BAC with S1407 site mutations (bac^S1407P, bac^S1407A, bac^P1407S) were inoculated on A. thaliana leaves (Figures 6A–C) and tomato fruits (Figures 6D–F). The pathogenicity of wild-type, bac^S1407A, and bac^P1407S conidia was significantly different at dawn and dusk (Figures 6A,D). However, the differences in pathogenicity between bac^S1407P and bac^S1407D at dawn and dusk were significantly greater than those of the wild type (Figures 6B,C,E,F). These results indicate that the BAC-mediated cAMP signaling pathway is involved in the regulation of differences in the pathogenicity of B. cinerea between dawn and dusk.

The light signaling pathway is one of the most common environmental factors that regulate the circadian rhythm of fungi (Hevia et al., 2016). Under alternating light and dark conditions, fungi form a weak ring which is due to different growth states and is recognized as the growth rhythm. This growth rhythm is more apparent if SDS is added to the medium because the mycelium growth rate is reduced (Canessa et al., 2013). We compared the differences in circadian rhythms between the wild type (B05.10) and mutant (bac^S1407P, bac^S1407D, bac^S1407A, and Δbcwcl1) under LD conditions (Figure 7A). The photoreceptor BcWCL1 is sensitive to light and is an important component of the circadian clock. Therefore, Δbcwcl1 showed no apparent growth rhythm on the CM supplemented with SDS. The bac^S1407A showed no significant changes compared with the wild type, and the growth of bac^S1407D was seriously affected. Hence, it was difficult to observe the rhythm phenotype. However, the growth rhythm of bac^S1407P was more significant than that of wild-type B05.10 (Figure 7A). This is consistent with the results of circadian rhythm observed in different PDA and CM media without SDS (Supplementary Figure S3). These results indicate that the BAC-mediated cAMP signaling pathway is involved in the operation of the circadian clock.

The 45 h-expression curve of the circadian clock core components Bcwcl1 and Bcfrq1 in wild type and bac^S1407P under conditions with light–dark alternation was verified at this point (Figures 7B–E). Transcript levels of Bcfrq1 and Bcwcl1 showed rhythmic changes in the wild type (Figures 6B,C). Compared with that in bac^S1407P, the expression of Bcfrq1 and Bcwcl1 was significantly higher, and the expression of Bcwcl1 increased significantly in bac^S1407P (Figure 7D). The rhythm of Bcfrq1 in bac^S1407P was faster than the wild type, and the expression peak was reduced (Figure 7B). The expression of Bcfrq1 in Δbcwcl1 was significantly decreased without rhythm, which was consistent with its rhythmic phenotype (Figure 7A). This result is consistent with the phenotype of the mutant strains observed at the S1407 site of BAC, with elevated BAC phosphorylation levels contributing to the more significant difference in pathogenicity between dawn and dusk (Figure 6). These results further demonstrate that BAC is involved in the regulation of rhythmic growth and pathogenicity.

Phosphodiesterases (PDEs), which can decompose cAMP, are inhibited by isobutylmethylxanthine (IBMX) for the disruption of the cAMP signaling pathway (Agarwal et al., 2010). The inhibitor IBMX was added to the CM medium in the race tube to verify the role of the BAC-mediated cAMP-PKA signaling pathway in the circadian clock. In the race tube containing CM medium supplemented with IBMX, the banding pattern of bac^S1407P disappeared under LD conditions (Figure 8A). Owing to the lack of operation of the circadian clock, the surface of the mycelial colony of all strains became flattered in comparison to that observed with the IBMX-free CM medium under LD conditions (Supplementary Figure S3B). These results verify the regulatory effect of cAMP on the circadian clock.

Studies in N. crassa have shown that PKA can phosphorylate FRQ1 and WC1 (Huang et al., 2007), and nucleotide sequence comparison has shown that the phosphorylation sites of WCL1 are largely conserved in comparison to those in WC1 (Huang et al., 2007; Figure 8B). The cAMP-PKA signaling pathway regulatory subunit BcPKAR can be activated by cAMP to transfer the signal (Liñeiro et al., 2016). The present yeast two-hybrid (Y2H) assay system analysis indicated that BcPKAR interacted with BcWCL1 and BcFRQ1 (Figure 8C). These results suggest that B. cinerea may have a similar regulatory mechanism as N. crassa.

The phenotypes of BAC S1407 mutations indicate that BAC is crucial to the photomorphogenesis of B. cinerea and that Bcltf1, Bcltf2, and Bcltf3 are the key transcription regulators downstream of WCC that regulate conidia production under light and sclerotia under dark conditions in B. cinerea (Schumacher et al., 2014; Cohrs et al., 2016; Brandhoff et al., 2017). The changes in the expression of Bcltf1, Bcltf2, and Bcltf3 in wild type and bac^S1407P observed under LD conditions showed that these three genes were significantly different from those in the wild type (Figure 9). In the wild type, compared with the expression curves of core components of the circadian clock, namely, Bcfrq1 and Bcwcl1 (Figures 9B,C), the expression curves of Bcltf1 and Bcltf2 showed that they were more remarkably controlled by the circadian clock (Figures 9A,D). However, Bcltf3 was less sensitive to the clock in the wild type, which was more relevant to light regulation (Figure 9H). The expression oscillation amplitude of Bcltf1 increased (Figure 9B), and the expression oscillation of Bcltf3 was significantly reduced (Figure 9I) in bac^S1407P, indicating that the BAC-mediated cAMP signaling pathway has a regulatory effect on it. Unlike the predicted results, the oscillation of Bcltf2 expression was remarkably reduced, and the rhythm disappeared in bac^S1407P compared to that in B05.10 (Figure 9E). This also explains why the production of conidia sharply declined due to the remarkable reduction in Bcltf2 expression, which was influenced by the cAMP signaling pathway in bac^S1407P.

In bac^S1407P, the expression curve of Bcwcl1 is significantly higher than that of wild type (Figure 7D), while Bcfrq1 in Δbcwcl1 is almost not expressed (Figure 7E), and the expression of LTF in bac^S1407P and Δbcwcl1 is significantly changed compared with that of wild type (Figures 9C,F,J), which indicates that BcWCL1, the core component of the circadian clock, can regulate LTF, but its regulatory mechanism has not been reported.

The zinc finger domain of WCC is highly conserved, and the amino acid sequence for DNA binding has been identified in this domain (Wang et al., 2015). Based on the WCC consensus-binding site reported by Baek et al. (2019), the reported WCC downstream gene (Bcfrq1, Bcltf1, Bcltf2, and Bcltf3) promoter regions were analyzed. Bcfrq1 had three putative binding sites on the promoter, whereas Bcltf1 and Bcltf2 possessed one putative site (Figure 10A). In contrast, Bcltf3 did not carry a WCC consensus-binding site (Figure 10A). These results indicate that Bcltf3 is indirectly regulated by WCC. The yeast one-hybrid assay was used to verify the transcriptional regulation of Bcfrq1, Bcltf1, Bcltf2, and Bcltf3 in BcWCL1 (Figure 10B). Consistent with the promoter binding site prediction of Bcfrq1, Bcltf1, Bcltf2, and Bcltf3 in BcWCL1, the yeast one-hybrid assay proved that BcWCL1 could regulate the promoters of Bcfrq1, Bcltf1, and Bcltf2, but not that of Bcltf3 (Figure 10B). In conclusion, the BAC mediated cAMP signaling pathway can regulate the expression of Bcwcl1 and LTF to change the survival strategy of B. cinerea in the natural environment.