Snf1 Kinase Differentially Regulates Botrytis cinerea Pathogenicity according to the Plant Host

The Snf1 kinase of the glucose signaling pathway controls the response to nutritional and environmental stresses. In phytopathogenic fungi, Snf1 acts as a global activator of plant cell wall degrading enzymes that are major virulence factors for plant colonization. To characterize its role in the virulence of the necrotrophic fungus Botrytis cinerea, two independent deletion mutants of the Bcsnf1 gene were obtained and analyzed. Virulence of the Δsnf1 mutants was reduced by 59% on a host with acidic pH (apple fruit) and up to 89% on hosts with neutral pH (cucumber cotyledon and French bean leaf). In vitro, Δsnf1 mutants grew slower than the wild type strain at both pH 5 and 7, with a reduction of 20–80% in simple sugars, polysaccharides, and lipidic carbon sources, and these defects were amplified at pH 7. A two-fold reduction in secretion of xylanase activities was observed consequently to the Bcsnf1 gene deletion. Moreover, Δsnf1 mutants were altered in their ability to control ambient pH. Finally, Δsnf1 mutants were impaired in asexual sporulation and did not produce macroconidia. These results confirm the importance of BcSnf1 in pathogenicity, nutrition, and conidiation, and suggest a role in pH regulation for this global regulator in filamentous fungi.


Introduction
The infection process of necrotrophic fungi lies in the synergy of several molecular mechanisms, such as secretion of degrading enzymes, production of toxins, oxidative burst, or modulation of environmental pH. These mechanisms perturb the host immunity and induce host cell death, from which the fungus will retrieve nutrients for its growth.
In plant-pathogenic fungi, the secretion of a wide spectrum of plant cell wall degrading enzymes (CWDEs) is the first and most studied mechanism involved in the penetration and colonization of the plant tissue. Cellulose, hemicelluloses, and pectin are the main polysaccharide components of the plant cell wall [1]. Cellulose is a β-1,4-linked D-glucose polymer, forming microfibrils in the primary and secondary cell wall, and is degraded by fungal cellulases. Xylan, the major hemicellulose, is built up by β-1,4-linked xyloses, arabinoses, and glucuronic acids, and is degraded by fungal xylanases. Pectin is a complex polysaccharide, mostly composed of D-galacturonic acids and mainly found in the middle lamella and primary cell wall, as well as being degraded by fungal pectinases.
Studies on the function of CWDEs in plant pathogenic fungi have been difficult, due to gene redundancy. CWDE encoding genes are repressed by glucose but derepressed by the function of the sucrose non-fermenting protein kinase 1 gene (Snf1). Snf1 is highly thought to be a differentiated organ, dedicated to the massive secretion of CWDE [40]. Indeed, in infection cushion formation, CWDE secretion and nutrient assimilation are supposed to together play major roles in B. cinerea pathogenicity, as their simultaneous impairment is a phenotypic signature of non-pathogenic mutants [41,42]. B. cinerea also has a remarkable capacity for modulating the ambient pH during its infection process, by secreting either organic acids (e.g., oxalic acid) or ammonia, enabling this fungus to colonize plants with acidic or neutral tissues [43]. To confirm the role of Snf1 in B. cinerea pathogenicity and nutrition, two independent ∆Bcsnf1 mutants were obtained by gene replacement, and the in vitro and in planta growth of these mutants was examined in neutral or acidic pH conditions.

Fungal Strains and Growth Conditions
B. cinerea strain B05.10, mutants and complemented strains were kept on potato dextrose agar (PDA) (Oxoid, CM0139) at 21 • C in the dark. For radial growth and pathogenicity tests, three-day-old plates with mycelia were used; mycelium plugs were taken from the margin of actively growing fungal colonies.

Construction of Deletion Cassettes by PCR Fusion
The orthologue of the yeast Snf1 gene was identified in B. cinerea by bidirectional blast hit (BBH) and named Bcsnf1, corresponding to the Bcin14g03370 reference gene (strain B05.10). For targeted gene replacement, the split-marker technique was used [44]. With this approach, two constructs are required per transformation, each containing a flanking region of the target gene and a truncated selectable marker cassette (Figure 1). Homologous recombination between the overlapping regions of the selectable marker gene and between the flank regions and their genome counterparts results in a targeted gene deletion and replacement with an intact marker gene. Genomic DNA from B. cinerea B05.10 strain was extracted with a DNeasy Plant Mini Kit (Qiagen, Germany) and used as a template. In the first PCR round, the 5 and 3 flanking regions were amplified with the primer pairs Snf1-5 -For and Snf1-5 -Rev, and the Snf1-3 -For and Snf1-3 -Rev (Table S1). The hygromycin resistance gene was amplified with the primers Hyg-For and Hyg-Rev, using a hygromycin cassette as a template [45]. Double-joint PCR [46] was performed with iProofTM High-Fidelity DNA Polymerase (BIO-RAD, Hercules, CA, USA), following the manufacturer's instructions, and the DNA fragments were purified from agarose gel using GFX PCR and a Gel Band Purification Kit (GE Healthcare, Chicago, IL, USA). Equimolar amounts of the purified fragments were mixed and fused in a second round of linear PCR. The joint fragments were amplified with nested primers in a third round of PCR. The primer pair Snf1-5 -For and Hyg-Nest-Rev was used to amplify the 5 flanking region and the 5 fragment of the hygromycin gene, and the Hyg-Nest-For and Snf1-3 -Rev primers to amplify the 3 fragment of the hygromycin gene and 3 flanking region (Table S1). The DNA fragments were purified from agarose gel using GFX PCR and a Gel Band Purification Kit.  0  3  2  0  3  2  2  3  1  0  3  3  2  3  3  3  3   Pathogenicity   2  3  3  0  2  1  3  2  1  3  3  3  2  3  3  2  3  2  3 In vitro radial growth on different carbon sources  Two-week-old sporulating mycelium of B05.10 wild type strain was scraped from plates and filtered through a 100 µm cell strainer (BD Falcon, Swedesboro, NJ, USA). The filtered spores were transferred in a 250 mL Erlenmeyer flask containing 100 mLNY medium (2 g L −1 yeast extract and 20 g L −1 malt extract). The fungus was grown at 23 • C for 24 h at 120 rpm. Lysing enzymes (β-1,3-glucanases mixture) from T. harzianum (Sigma, St. Louis, MO, USA) were prepared by dissolving 0.2 g in 10 mL of KCl/NaP buffer [47] and preheated at 37 • C, then the enzyme solution was filtered through a 20-µm filter (Sartorius, Germany) and diluted to 20 mL with KCl/NaP buffer. Mycelium from an Erlenmeyer flask was harvested on a Nitex bolting cloth and washed with KCl/NaP buffer. The mycelium was transferred and digested into a 100 mL Erlenmeyer flask, where the lysing enzyme was added. After 3 h of incubation at 23 • C and 70 rpm, the protoplasts were filtered through a 40-µm cell strainer (BD Falcon, Swedesboro, NJ, USA) and washed with 2 mL TMS buffer (1 M sorbitol and 10 mM MOPS, pH 6.3). The protoplasts were centrifuged at 4 • C for 5 min at 3500 rpm. The supernatant was carefully discarded, the pellet was diluted in 10 mL TMS buffer, and the suspension was centrifuged again at 4 • C for 5 min at 3500 rpm. The pellet was diluted in 300 µL TMS buffer, and the protoplast concentration was determined using a Thoma cell counting chamber. For cell transformation, 2 × 10 7 protoplasts were suspended in 100 µL TMSC buffer (1 M sorbitol, 10 mM MOPS, and 40 mM CaCl 2 , pH 6.3), mixed with 2 µg of each replacement cassette and diluted in TE CaCl 2 2× buffer (20 mM Tris-HCl, 2 mM EDTA, 80 mM CaCl 2 ·2H 2 O, pH 7.5). The transformation mixture was incubated for 20 min on ice. Then, 160 µL of PEG solution (1.2 g PEG6000 dissolved in 800 µL MS buffer (0.6 M sorbitol and 10 mM MOPS, pH 6.3)) was added to the mixture, and after incubation at RT for 15 min, 1 mL pre-cooled TMSC buffer was added. The sample was centrifuged for 5 min at 5000 rpm and the pellet was resuspended in 400 µL TMSC buffer. Then, 4 × 3 mL MMV Top medium (2 g L −1 NaNO 3 , 1 g L −1 K 2 HPO 4 , 0.5 g L −1 KCl, 0.5 g L −1 MgSO 4 ·7H 2 O, 0.01 g L −1 FeSO 4 ·7H 2 O, 20% (w/v) saccharose, 2% (w/v) glucose, and 0.4% (w/v) agar) containing 100 µg ml −1 hygromycin was preheated at 42 • C. Next, 100 µL of the sample was transferred into one tube of MMV Top medium and was poured onto a Petri dish containing MMV medium (MMV Top medium with 1.5% (w/v) agar). The plates were kept at 21 • C in the dark for several days. When transformants appeared, they were transferred to MM medium containing 100 µg ml −1 hygromycin.

Molecular Validation of Gene Deletion by PCR and Southern Blot
Transformant colonies were first checked by PCR, to confirm the insertion of the hygromycin resistance gene using the Hyg-For and Hyg-Rev primers (Table S1). The positive colonies were validated for the correct insertion of the hygromycin resistance gene at the 5 and 3 Snf1 flanking regions with the primer pairs Snf1-M-5 -For, Snf1-M-5 -Rev; and Snf1-M-3 -For, Snf1-M-3 -Rev (Table S1). As non-transformed nuclei can be maintained in hygromycin resistant transformants, we checked the presence/absence of the wild type snf1 ORF sequence by PCR, with the Snf1-WT-For and Snf1-WT-Rev primers (Table S1). PCR conditions are reported in Table S1, according to the information sheet of Taq polymerase (MP Biomedicals, Santa Ana, CA, USA).
Selected homokaryotic mutants determined by PCR were then verified by Southern Blot analysis, using a PCR DIG Probe Synthesis Kit, DIG Easy Hyb, and DIG Luminescent Detection Kit (Roche, Basel, Switzerland). Genomic DNA from the wild type and mutant strains were digested with SpeI or with SnaBI in two different experiments. The SpeIdigested genomic DNAs were hybridized with an 818 bp hygromycin specific probe (prepared using the Hyg-Probe-For and Hyg-Probe-Rev primers) and the SnaBI-digested samples were hybridized with an 1167 bp 5 flanking region-specific probe (prepared using the Snf1-5 -For and Snf1-5 -Rev primers). The DIG signal was detected with a ChemicDoc XRS camera (Bio-Rad).

Functional Complementation
Genomic DNA from B. cinerea wild type was extracted with a DNeasy Mini Plant Kit and used as a template to amplify the Bcsnf1 gene together with 1 Kb of promoter and 1 Kb of terminator region, using the Snf1-comp-For and Snf1-Comp-Rev primers (Table S1). The nourseothricin resistance gene was amplified from pONT vector with the Nourseo For and Nourseo Rev primers (Table S1). PCRs were performed using iProofTM High-Fidelity DNA Polymerase (BIO-RAD, Hercules, CA, USA), following the manufacturer's indications. Fragments of the expected sizes (Table S1) were purified from agarose gel using GFX PCR and a Gel Band Purification Kit. ∆snf1.1 mutant was grown on PDA covered with cellophane at 21 • C for 2 days. Mycelium was harvested and ground, and then grown in 100 mL NY medium at 21 • C for 24 h at 110 rpm. Mycelium was filtered and digested to obtain protoplasts, as described above. The Bcsnf1 gene and nourseothricin resistance cassettes were transferred into the ∆snf1.1 genome by co-transformation. Transformants were selected on MMII medium containing 100 µg ml −1 nourseothricin, and insertion of the Bcsnf1 gene was verified by PCR with the Snf1-WT-For and Snf1-WT-Rev, the Snf1-5 -For and Snf1-WT-Rev, and the Snf1-WT-For and Snf1-3 -Rev primers (Table S1).

Pathogenicity Assays on Plants
Pathogenicity tests were performed on detached cotyledons of 6-day-old cucumber (cv. Petit vert de Paris) plants and detached leaves of 7-day-old French bean (cv. Saxa) by placing a 3-mm diameter plug in a drop of water on the surface of plants. Inoculated leaves and cotyledons were placed on moist filter paper in plastic boxes and incubated into a climatic chamber with 16/8 h light/dark cycle at 21 • C and 70-75% relative humidity. Apple fruits (cv. Golden Delicious) were superficially wounded with a scalpel, and mycelium plugs of 7-mm of diameter were placed above the wounds. At 4 dpi, leaves and fruits were photographed, and the lesion areas were measured using the ImageJ program [48]. Mycelium on the French bean leaf surface was stained with lactic blue cotton solution and photos of the lesions at 2 dpi were taken with a SteREO Discovery.V20 microscope (Zeiss, Jena, Germany).

Xylanase Enzymatic Assay
Each strain was first grown on the surface of cellophane sheets overlaying potato dextrose agar (PDA) (Oxoid, CM0139). After three days, the cellophane membranes with the mycelium were transferred to 10 mL liquid MM (medium described above without agar), supplemented with 1% (w/v) of xylan from beechwood (Carlroth). The pH was buffered at 5 or 7. The supernatant and the mycelium of each culture were collected after 4 dpi. The mycelium was lyophilized and weighed. Enzymatic reactions were performed in a mixture containing 0.625% (w/v) of xylan in McIlvaine buffer at pH 5 or pH 7 and 100 µL of the culture supernatant in a final volume of 400 µL. The enzymatic reaction mixtures were incubated at 37 • C for 0, 15, 30, 45, 60, 75, and 90 min, and enzymatic reactions were stopped at 95 • C. Total xylanase activity was assayed by measuring the release of xylose reducing sugars from the xylan substrate with the 4-hydroxybenzoic acid hydrazide (PAHBAH) method [49]. Two hundred µL aliquots of each enzymatic reaction mixture were added to 1800 µL of 0.5% (w/v) PAHBAH (in 500 mM NaOH). The dosage mixture was incubated at 95 • C for 10 min and its absorbance was measured at 410 nm. Xylanase activity was expressed as µg of xylose/min/mg mycelium dry weight using xylose as a standard. Three biological replicates of the wild type and ∆snf1.1 mutant were prepared.

Monitoring Ambient pH Changes in Liquid Culture
As the ∆snf1 mutants do not sporulate, the mycelia of the wild type and the mutant strains were first grown on sporulation malt medium (5 g L −1 glucose, 20 g L −1 malt extract, 1 g L −1 tryptone, 1 g L −1 casamino acids, 1 g L −1 yeast extract, 0.2 g L −1 ribonucleic acid sodium salt, and 15 g L −1 agar-agar) at 21 • C in dark. After four days of incubation, the mycelium was cut into small pieces and transferred into 30 mL of liquid sporulation malt medium. It was then incubated at 21 • C and 110 rpm for 44 h. The mycelium was collected by centrifugation at 3500 rpm for 5 min and rinsed with 35 mLsterile H 2 O two times. The washed mycelium was then grown in 30 mL of autoclaved Gamborg medium [50] adjusted to pH 6 (not buffered) and including six-days-old cucumber cotyledons to mimic the plant environment. Cultures were incubated at 21 • C and 110 rpm and their pHs were measured at 1, 2, and 3 dpi.

Quantitative Real-Time RT-PCR
qRT-PCR was performed in order to compare the Bcsnf1 gene expression level of the wild type and the complemented strains in vitro. In vitro, four-day-old mycelium was grown and harvested on MM containing 1% PGA or xylan at pH 5 or pH 7. Samples were frozen immediately in liquid nitrogen and stored at −80 • C until utilization. Total RNA was extracted following the protocol of Reid et al. (2006) [51]. First-strand cDNAs were synthesized with an ImProm-II™ Reverse Transcription System (Promega, Madison, WI, USA), following the manual's instructions, then the samples were treated with RQ1 (RNA Qualified) RNase-Free DNase (Promega, Madison, WI, USA). qRT-PCR was performed with SYBR ® Green master mix (BIO-RAD, Hercules, CA, USA) on a Rotor-Gene Q real-time PCR cycler (Qiagen, Hilden, Germany) with the Bc-Tub-For, Bc-Tub-Rev; and Snf1-ORF-For, Snf1-ORF-Rev primer pairs (Table S1). qPCR conditions were as follows: 40 cycles of 95 • C for 20 s, and 57 • C for 20 s and 72 • C for 30 s. Relative expression of the Bcsnf1 gene compared to the tubulin reference gene was determined using the 2 −∆∆Ct method [52].

Statistical Analysis
To investigate the significant difference of in vitro and in planta growth of the wild type, mutant, and complemented strains, one-way analyses of variance (ANOVA) was performed. Tukey-Kramer multiple comparisons were accomplished at a 99% significance level. To investigate the significant difference of xylanase activity from the wild type and mutant strains, a Student test was performed.

Targeted Gene Deletion of Bcsnf1 Gene in Botrytis Cinerea
To study the role of B. cinerea Snf1 protein kinase, the ORF of the encoding gene Bcsnf1 (locus Bcin14g03370) was replaced with a deletion cassette containing the hygromycin resistance gene, by using the PEG/CaCl 2 chemical transformation of protoplasts. A splitmarker approach [44] was used to fuse and integrate two truncated deletion cassettes at the Bcsnf1 locus by homologous recombination. Hygromycin-resistant transformants were first checked by PCR for the presence of the hygromycin resistance gene and the absence of the Bcsnf1 gene (data not shown); four transformants were analyzed by Southern blot, and deletion of the Bcsnf1 gene was confirmed in two of them ( Figure 1). SnaBI-digested genomic DNAs hybridized with a 5 flanking region-specific probe showed the expected band of 5.57 Kb for the wild type, while the mutants showed a band of 11.3 Kb, indicating insertion of the hygromycin cassette and corresponding replacement of the Bcsnf1 gene ( Figure 1B). SpeI-digested genomic DNAs hybridized with a hygromycin specific probe showed only one band at the expected size of 6.35 Kb in the mutant strains, confirming that no other ectopic integration of the cassette had occurred ( Figure 1C). The two independent ∆snf1 homokaryotic deletion mutants (∆snf1.1 and ∆snf1.4) with no additional ectopic insertion of the hygromycin cassette were, thus, selected for further phenotypic characterization. To complement the ∆snf1 mutation, the entire Bcsnf1 gene containing 1 Kb of the promoter and 1 Kb of the terminator regions flanked with the nourseothricin resistance gene was reintroduced into the deletion strain ∆snf1.1. Transformants able to grow on medium containing nourseothricin were selected, and amplification of the Bcsnf1 gene was confirmed by PCR (data not shown).

Bcsnf1 Gene Deletion Abolishes Asexual Sporulation and Production of Macroconidia
In the B. cinerea life cycle, asexual reproduction is supported by the differentiation of macro-conidiophores that produce macroconidia representing the main inoculum source of the fungus. Indeed, macroconidia disseminate and germinate at the surface of plants, where they initiate the disease. The B. cinerea ∆snf1 mutants and wild type strains were grown on synthetic solid minimal media (MM) supplemented with 1% (w/v) carbon source (glucose, sucrose, carboxymethyl cellulose (CMC), galacturonic acid, polygalacturonic acid, xylose, or xylan). pH was adjusted and buffered at 5 or 7. At 7 dpi, in all the media tested, the wild type asexually sporulated (Figure 2A,B), but the ∆snf1 mutants did not show any macroconidiophores nor macroconidia. However, abundant microconidia and their reproductive structures (micro-conidiophores) were seen instead ( Figure 2C,D). Indeed, B. cinerea can develop micro-conidiophores, producing microconidia that are implicated in the sexual reproduction of the fungus. Microconidia (male parent) are not able to germinate and initiate in vitro growth or disease on plants and are consequently not infectious. Functional complementation of Bcsnf1 fully restored the asexual reproduction and ability to produce macroconidia ( Figure 2E,F).

Bcsnf1 Gene Deletion Does Not Affect in Planta Penetration by the Fungus, but Alters Virulence According to the Host
With asexual sporulation being abolished in the ∆snf1 mutants, penetration on French bean leaves was checked by using mycelium plugs as inoculum instead of macroconidia. Compound appressoria, called infection cushions, are dedicated to the penetration of B. cinerea mycelium through the plant cell wall barriers ( Figure 3A [40]). No defect in penetration was observed for the ∆snf1 mutants, as many infection cushions were visible at the margin of the mycelium plugs ( Figure 3B).
Next, the in planta colonization was checked on three different organs of plants displaying neutral or acidic pH: one-week-old cucumber cotyledons (pH 6.7), primary French bean leaves (pH 6.3), and golden apple fruits (pH 3.9). Surface lesions were measured daily up to 4 dpi for the wild type, two independent ∆snf1 mutants, and the complemented ∆snf1.1-C strain. In comparison with the wild type strain, a strong reduction of fungal colonization was observed for both ∆snf1 mutants on cucumber cotyledons, with a lesion area severely reduced by 89% and only a small infected tissue visible around the mycelium plugs ( Figure 4A). Virulence of the complemented strain was partially restored by 44%. On French bean leaves, the lesion area produced by the mutant strains was 85% smaller than that of the wild type at 4 dpi, and the virulence was partially restored (by 40%) in the complemented strain ( Figure 4B). On apple fruits, the mutant strains showed only a 59% reduction of the lesion size ( Figure 4C), and surprisingly no restoration of the virulence was observed in apple for the complemented strain.

Bcsnf1 Gene Deletion Alters Xylanase Secretion and Carbon Nutrition
The total xylanase activity secreted at 4 dpi by the wild type and the ∆snf1.1 mutant strain was determined in vitro at pH 5 or pH 7 with the PAHBAH method. At pH 5, the xylanase activity of the ∆snf1.1 mutant significantly decreased, by two-times compared to the wild type strain ( Figure 5; p < 0.05). At pH 7, the xylanase activity secreted by the wild type strain was three-times lower in comparison with at pH 5. For the ∆snf1.1 mutant, the xylanase activity was very low at pH 7, with a decrease by two times compared to the wild type, although this difference was not statistically significant ( Figure 5).   In vitro radial growth of the wild type, two independent Δsnf1 mutants, Δsnf1.1-C complemented strain were followed for four days. Strains were inocul synthetic media containing 1% (w/v) simple sugars (glucose, xylose, sucrose, and polysaccharides (PGA, xylan, cellulose) or non-fermentable carbons (tween 80, o triolein, and NaAc). The media were buffered at pH 5 or pH 7, and the experimen performed in three biological replicates. On simple sugars and polysaccharides, the mutants showed a mild reduction of growth diameter, by 14-34%, compare wild type, and no significant restoration was determined for the complemente ( Figure 6A). However, at pH 7 the growth of the mutants was moderately reduced 50%, compared to the wild type, and at least 45% of restoration was observed growth of the complemented strain ( Figure 6B). On non-fermentable carbon sou pH 5, the mutants also showed a mild reduction of growth diameter, by 13-37% pared to the wild type, and no significant restoration was determined for the mented strain ( Figure 7A). However, at pH 7 the growth of the mutants was mod reduced, by 35-59%, compared to the wild type, and between 49-80% of partial tion was observed for the growth of the complemented strain ( Figure 7B). In vitro radial growth of the wild type, two independent ∆snf1 mutants, and the ∆snf1.1-C complemented strain were followed for four days. Strains were inoculated on synthetic media containing 1% (w/v) simple sugars (glucose, xylose, sucrose, and GA), or polysaccharides (PGA, xylan, cellulose) or non-fermentable carbons (tween 80, olive oil, triolein, and NaAc). The media were buffered at pH 5 or pH 7, and the experiments were performed in three biological replicates. On simple sugars and polysaccharides, at pH 5, the mutants showed a mild reduction of growth diameter, by 14-34%, compared to the wild type, and no significant restoration was determined for the complemented strain ( Figure 6A). However, at pH 7 the growth of the mutants was moderately reduced, by 18-50%, compared to the wild type, and at least 45% of restoration was observed for the growth of the complemented strain ( Figure 6B). On non-fermentable carbon sources, at pH 5, the mutants also showed a mild reduction of growth diameter, by 13-37%, compared to the wild type, and no significant restoration was determined for the complemented strain ( Figure 7A). However, at pH 7 the growth of the mutants was moderately reduced, by 35-59%, compared to the wild type, and between 49-80% of partial restoration was observed for the growth of the complemented strain ( Figure 7B).

Bcsnf1 Deletion Alters the Ability of the Fungus to Modulate the Alkaline pH
B. cinerea is known to display a dual behavior in modulating the ambient pH. It, first, acidifies its environment and, second, increases pH to neutral values [43]. To measure the pH changes during B. cinerea liquid growth, mycelia of the wild type and the two mutant strains were grown in a modified Gamborg liquid medium adjusted at pH 6 and containing cucumber cotyledons as the only source of carbohydrates. At 1 dpi, the ∆snf1 mutants were able to initially acidify the medium similarly to the wild type strain (Table 2). However, at 2 and 3 dpi the mutants appeared to abnormally increase the pH of the medium to higher values than the wild type. Indeed, an increase of one pH unit was observed in the medium at 2 dpi.  . Mycelium colony diameters were measured at 4 dpi (excluding the mycelium plug diameter). The mean was calculated from nine measures from three independent biological experiments for each strain and condition. Bars indicate the standard error and letters indicate the significant difference (p < 0.01) between the strains. NaAc = sodium acetate. Bcsnf1 gene expression was checked in vitro at pH 5 and pH 7 for the wild type and the complemented strain using RT-qPCR analysis. Bcsnf1 gene was similarly expressed in both strains at pH 5 or pH 7 on xylan or PGA (Figure 8). Thus, transcriptional regulation of Bcsnf1 gene seems to be not affected by pH. The expression of the Bcsnf1 gene was restored in the complemented strain (Figure 8).

Discussion
To characterize the role of Snf1 kinase in the biology of Botrytis cinerea, two independent deletion mutants of the encoding gene were obtained. Analysis of their phenotypes was similar, and functional complementation was performed in one of them.

Role of Snf1 in Conidiation of B. cinerea
The absence of Bcsnf1 gene abolished the asexual reproduction of B. cinerea: the ∆snf1 mutants did not produce macro-conidiophore with macroconidia, representing the main inoculum of the disease. Instead, the ∆snf1 mutants produced micro-conidiophores with microconidia that are implicated in the sexual reproduction of the fungus but not in host infection. The complementation of one of the mutants by the reintroduction of a Bcsnf1 copy restored the capacity to produce macroconidia. In other filamentous fungi, 11 out of 19 ∆snf1 mutants generated were also moderately to severely decreased in their capacity of sporulation (Table 1). Moreover, several Snf1 mutants display an abnormal morphology of conidia [10,11,17]. In P. digitatum, conidiation (asexual sporulation) of the ∆Snf1 mutant represented only 10% of the wild type, and 90% of the observed conidiophores did not branch at their tips. Conidia were produced directly at the tips of hyphae [13]. To explore the potential function of Snf1 in P. digitatum asexual reproduction, expression levels of the regulators brlA and fadA were analyzed by qPCR. brlA is a signaling gene that positively regulates conidiation and conidiophore morphogenesis. The transcription level of brlA was significantly lower in the ∆snf1 mutant, suggesting that the regulatory role of Snf1 on conidiation may be correlated with the regulation of brlA expression (BrlA regulator would be likely positioned downstream of the Snf1 regulator). FadA is an α-subunit of a heterotrimeric G protein that mediates growth signaling and negatively regulates conidiation. The expression level of fadA in the ∆Snf1 mutant was higher than that found in the wild type, indicating that the expression of the FadA-signaling pathway is negatively regulated by Snf1, and that Snf1 is required to activate the conidiation-signaling pathway and inactivate the growth-signaling pathways in P. digitatum [13]. In the entomopathogenic fungus B. bassiana, RNAseq transcription levels of the regulators of conidiation were reduced in the ∆Snf1 mutant (abaA; flbC; fluG; [53]). Thus, Snf1 probably affects sporulation by regulating the expression of key regulators in the sporulation signaling pathway, and we can imagine that the Snf1 kinase also regulates conidiation in B. cinerea. It should be noted that this role in the regulation of the conidiogenesis of B. cinerea seems not to be pHdependent, because the mutant failed to sporulate at either pH 5 or pH 7. As macroconidia of B. cinerea are considered as the main inoculum of the grey mold disease, targeting Snf1 would interrupt the disease cycle of the fungus.

Role of Snf1 in Pathogenicity of B. cinerea
Pathogenicity tests demonstrated that Snf1 is an important virulence factor in many filamentous fungi. Indeed, a significant reduction of pathogenicity was observed for a dozen plant-pathogenic fungi (Table 1). In M. oryzae and C. fructicola, the decrease in pathogenicity observed for the ∆snf1 mutant was attributed to a defect in developing the appressorium, a structure dedicated to the plant penetration [10,16,25]. During plant infection, mycelium plugs of the ∆snf1 mutants of B. cinerea were still able to produce compound appressoria (Figure 3), called infection cushions [40]. As these structures are dedicated to the penetration of the pathogen into the host tissue, it was not surprising to observe that the ∆snf1 mutants were still able to penetrate the plant tissues and initiate infection. However, colonization of the mutant was strongly reduced, from 59% to 89%, according to the plant tissues tested. In particular, the lesion produced by the mutants slowed down in comparison with the wild type strain and stopped after 2 dpi (data not shown) when the infection cushions were differentiated ( Figure 3). These results suggest that Snf1 does not regulate the penetration of the fungus, but more likely the colonization of the mycelium in plant tissues. Indeed in V. dahliae, microscopic observation of the infection behavior of a green fluorescent protein (GFP)-labeled ∆Snf1 mutant showed that it was defective in the initial colonization of roots, xylem vessels, and cotyledons [12]. Fluorescence microscopy studies also revealed that the ∆Snf1 mutant of F. virguliforme failed to successfully colonize the vascular vessels and adjacent tissues of infected soybean roots [15]. As already proposed in many other fungi, defects in host colonization observed for the B. cinerea ∆snf1 mutants are the consequence of an inability to progress through, and feed on, the plant tissues, likely due to a deregulation of CWDEs and sugar transporters. Indeed, we confirmed in vitro that B. cinerea ∆snf1 mutant secretes a two-fold reduced xylanase activity compared to the wild type strain, although this significant difference was statistically confirmed at pH 5, but not at pH 7. We can suppose that other plant CWDEs might also be affected in the ∆snf1 mutant.

Role of Snf1 in B. cinerea Growth on Different Carbon Sources
In vitro, radial growth experiments were performed to confirm defaults of ∆snf1 mutants for feeding on polysaccharides and simple sugars. Growth on carbon sources such as polysaccharides (cellulose, PGA, xylan) or simple sugars (glucose, sucrose, galacturonic acid, xylose) was partially impaired in the mutants, as observed in many other fungi (Table 1). In M. oryzae, it was shown that Snf1 also controls the β-oxidation, and the ∆snf1 mutant was unable to metabolize fatty acids and sodium acetate [25]. We tested the role of B. cinerea Bcsnf1 gene on lipid metabolism, by cultivating the strains on different lipidic non-fermentable carbon sources. Compared to the wild type, the B. cinerea ∆snf1 mutants exhibited a reduced growth rate on all fatty acid sources tested; while on sodium acetate, a significant reduction was observed only at pH 7. Considering all the carbon sources tested, the growth of wild type and ∆snf1 mutants was always lower at pH 7 than at pH 5. Moreover, in comparison with the wild type, the growth of the ∆snf1 mutants was always much more affected at pH 7 than at pH 5. This in vitro observation is in agreement with the observation in planta, because the ∆snf1 mutants were much more impaired in the colonization of tissues with neutral pH (about 6.3-6.7), such as bean leaf and cucumber cotyledon, than of tissues with acidic pH (about 3.9), such as apple fruit. These results suggest that the role of Snf1 in B. cinerea is more important at neutral pH than at acidic pH. The surprising observation that complementation of the ∆snf1 mutant by reintroduction of a Bcsnf1 copy partially restored the defects observed at neutral pH, but not at acidic pH, deserves further investigation. It could be suggested that the defects observed at neutral pH for the ∆snf1 mutants were severe enough to see a partial restoration with the complemented strain, but the milder defects observed at acidic pH did not allow this.

A Suggested Role of Snf1 on Alkaline pH Modulation
Our results indicate, not only the role of Bcsnf1 gene in the uptake of simple sugars and lipid metabolism, but also suggest its importance in the adaptive response to ambient pH variations in a filamentous fungus. This behavior is similar to that observed with the budding yeast Saccharomyces cerevisiae, where about 75% of the genes induced by high pH were also induced when glucose was depleted. Therefore, the function of Snf1 protein kinase appears to be crucial not only in adaptation to glucose scarcity but also for neutral/alkaline pH tolerance [24]. Uptake of many nutrients is perturbed by alkalinization of the environment that represents a stress condition for S. cerevisiae. This organism responds to this stress with a profound remodeling of gene expression involving several signaling pathways, including the Snf1 pathway. Yeast cells lacking Snf1 are markedly sensitive to neutral/alkaline pH, and Snf1 is known to be activated by alkaline stress: exposure to high pH results in increased Snf1 phosphorylation [24,54]. The role of Snf1 in the glucose metabolism in yeast thus appears to be important for its function in high pH tolerance. Moreover, Snf1 kinase inhibits Nrg1, a transcriptional repressor downstream Rim101 (ortholog of PacC in filamentous fungi) in the signaling pathway of adaptation to alkaline pH [54]. Thus, the Snf1 pathway and Rim101/PacC pathway seem to converge at the Nrg1 regulator in yeast.
The ability to adapt to and thrive in a broad range of environmental pH conditions is a hallmark of fungal biology. This is particularly important in the case of pathogenic fungi, which can modify the pH of infected tissues as an attack strategy. During its interaction with the host plant, B. cinerea is known to modulate its ambient pH by secreting organic acids (e.g., oxalic acid) or ammonia [43]. The ambient pH acts as a regulatory element, assisting B. cinerea in tuning its virulence machinery to the composition of its host tissue by differentially regulating the synthesis of CWDEs. pH measurements of in vitro liquid cultures of B. cinerea showed that ∆snf1 mutants over-alkalinize the medium at 2 and 3 dpi compared to the wild type strain. Thus, B. cinerea Bcsnf1 gene might have a role in the control and repression of ambient pH alkalinization. This result seems contradictory with that obtained with the entomopathogenic fungus B. bassiana showing extracellular over-acidification by the ∆snf1 mutant at 3 and 4 dpi [18]. The ∆snf1 mutant of B. bassiana had enhanced production of lactic, pyruvic, and citric acid, but oxalic acid production was partially repressed. Transcriptional analysis showed that a set of genes involved in organic acids biosynthesis and secretion was changed in this mutant, indicating that Bbsnf1 gene in B. bassiana controls extracellular acidification by the production of different organic acids [18].
RT-qPCR analysis in vitro at pH 5 and 7 showed that B. cinerea Bcsnf1 gene expression is not modulated by pH, since it is similarly expressed at both pH values. Therefore, the higher impact of Snf1 in fungal growth at neutral pH does not seem to depend on pH regulation of gene expression, but possibly on an increased Snf1 phosphorylation and activation at neutral/alkaline pH, as observed in yeast [24,54]. In a previous study, we observed that inactivation of the alkaline pH-signaling pathway PacC in B. cinerea resulted in a defect in virulence, depending on the pH of the host tissues [55]. The deletion of the pH regulator BcpacC resulted in virulence defects in hosts characterized by tissues with neutral pH, but not in hosts with acidic pH tissues. This result is quite similar to what we observed in this study for the deletion of Bcsnf1 gene, and it would be interesting to confirm the interactions between Snf1 and PacC signaling pathways in B. cinerea or other filamentous fungi, as was described in yeast [54]. In another study, we developed a RNA-seq approach comparing the transcriptomes from the ∆pacC mutant and the wild type strain of B. cinerea at acidic or neutral pH conditions (unpublished data, N. Poussereau personal communication). We observed that Bcsnf1 transcription was not regulated by the PacC transcription factor nor by pH, as we also concluded from our RT-qPCR analysis in this study.

Other Roles of the SNF1 Complex in Yeast and Filamentous Fungi
More recently, Snf1 was also proposed as a key regulator of filamentous fungi for other very diverse functions, such as cell wall integrity, stress tolerance to osmotic, oxidative or heat shocks, and biosynthesis of secondary metabolites in P. microspora [26], C. fructicola [16], A. alternata [17], P. anserina [23], and C. militaris [20]. These additional pleiotropic effects reveal that Snf1 kinase is an important global regulator of fungal biology and that it can be considered an attractive antifungal target. Additionally, in yeast the ATG autophagy pathway may collaborate with the SNF1 pathway to enhance survival under adverse environmental conditions, and inhibition of SNF1 would likely induce fungal degeneration over time [20]. Moreover, a contribution of Snf1 to yeast cell tolerance to freezing was also demonstrated [56].
Snf1 kinase is the α-subunit of a larger SNF1 protein complex, also including a βsubunit encoded by sip1 or sip2 or gal83 genes, and a γ-subunit encoded by the snf4 gene. In the yeast S. cerevisiae, the three subunits are equally important for SNF1 complex function [57]. In filamentous fungi, very few studies have analyzed all the components of the SNF1 complex. If the α-subunit FgSNF1 is mainly required for SNF1 complex functions in F. graminearum, the β-subunit FgGAL83 and the γ-subunit FgSNF4 have only adjunctive roles in sporulation and vegetative growth; however, they have major role in virulence [58]. In M. oryzae, the null mutants ∆Mosip2 and ∆Mosnf4 showed multiple disorders as ∆Mosnf1, suggesting that complex integrity is essential for SNF1 kinase function in this fungus [25]. One may wonder if the same situation might be present in B. cinerea, and it would, therefore, be interesting to study the βand γ-subunits in the SNF1 complex of this fungus.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/microorganisms10020444/s1, Table S1: Sequences of the primer pairs used in this study.