pH controls both transcription and post-translational processing of the protease BcACP1 in the phytopathogenic fungus Botrytis cinerea

The first barrier encountered by a phytopathogenic fungus during infection is the host cuticule and the primary plant cell wall. Successful infection depends not only on the pathogen's ability to penetrate and colonize the host tissues, but also on its ability to degrade the plant cell wall polymers (polysaccharides and several classes of structural proteins: Carpita & Gibeaut, 1993) into nutrients required for its development. Enzymic degradation of the parietal polymers is therefore considered an important aspect of plant fungal infections. Phytopathogenic fungi secrete a wide range of cell-wall-degrading enzymes that act at different stages of the infection. Among these enzymes are pectinases, cellulases and xylanases that are involved in polysaccharide degradation (St Leger et al., 1997; Cotton et al., 2003; Olivieri et al., 2004; Brito et al., 2006). Several classes of proteases have also been described (Billon-Grand et al., 2002; Bindschedler et al., 2003; Ten Have et al., 2004), whose proposed role would be to facilitate host tissue penetration and colonization by degrading structural plant cell wall proteins, and by releasing amino acids, which represent the pathogen's main source of nitrogen and sulfur during infection (Rauscher et al., 1995).

Fungi can secrete several proteases, each characterized by an optimal pH for activity, and many produced in correlation with the environmental pH. In Sclerotinia sclerotiorum for example, secreted proteases with an acidic optimum pH can be detected when the environment has been acidified via oxalic acid production by this phytopathogenic fungus (Marciano et al., 1983; Magro et al., 1984; Poussereau et al., 2001; Cotton et al., 2003). Similarly, in the insect pathogen Metarhizium anisopliae, a secreted subtilisin with an alkaline optimum pH can be detected when the ambient pH has become alkaline following production of ammonia by the fungus (St Leger et al., 1999). In order to ensure the secretion of these proteases in their optimal environment, their production needs to be controlled in response to the environmental pH. Although the effect of ambient pH on fungal growth, physiology and differentiation is well established, the molecular responses to environmental pH are only now being elucidated. In Aspergillus nidulans, alkaline pH is perceived and transmitted via the pal signalling pathway, which activates the zinc finger transcription factor PacC (Penalva & Arst, 2002). PacC is a transcriptional activator of alkaline-induced genes and a transcriptional repressor of acid-induced genes (Tilburn et al., 1995). PacC and its regulatory pathway are conserved in fungi, and it has been demonstrated that this regulator participates in the virulence of some pathogenic fungi (Rollins, 2003; Caracuel et al., 2003; You & Chung, 2007; Miyara et al., 2008).

The ascomycete Botrytis cinerea is a pathogen of more than 200 plant species (Elad et al., 2004). It can infect different organs including cotyledons, leaves, stems, roots, flowers and fruits at different stages of their development. B. cinerea, like the closely related necrotrophic fungus S. sclerotiorum, uses an infection strategy based on the acidification of the ambient environment via the production of organic acids (oxalic or citric acids) (Verhoeff et al., 1988) and on the secretion of numerous cell-wall-degrading enzymes, including proteases (Manteau et al., 2003). It has been reported that B. cinerea secretes several acidic proteases during pathogenesis. Movahedi & Heale (1990) purified and characterized one aspartic protease from infected carrots, while Ten Have et al. (2004) described a family of five aspartic proteases which they detected during the infection of several plant species.

In this paper, we report on an unusual acidic protease of B. cinerea, BcACP1, which belongs to the G1 family of proteases, a newly discovered class of endopeptidases found exclusively in fungi, and more specifically in ascomycete species (Sims et al., 2004; Pillai et al., 2007). Members of this family of proteases are characterized by an optimal activity at acidic pH and by their insensitivity to pepstatin A (inhibitor of the aspartic proteases). The environmental cues that control the enzyme production are still unknown. We show herein that BcACP1 is regulated by the environmental pH at both the transcriptional and post-translational level.

Biological material and culture conditions.
Botrytis cinerea strain BO5.10 was maintained at 21 °C on solid sporulation medium containing 1.5 % malt extract, 0.5 % glucose, 0.1 % tryptone, 0.1 % casein hydrolysate, 0.1 % yeast extract and 0.02 % tRNA, with 8 h black light illumination per day. B. cinerea transformant strains were maintained on malt agar medium (2 % malt extract, 1.5 % agar) supplemented with 100 µg hygromycin B ml^–1 (InvivoGen) and 200 µM cefotaxim (Sigma). For Bcacp1 expression, B. cinerea was grown on a rich solid medium containing 1.5 % agar, 2 % malt extract, 2 % glucose, and 100 mM NH₄Cl on the surface of cellophane sheets. After 3 days of incubation at 22 °C, the mycelium was transferred onto the surface of 2 ml Gamborg medium (Gamborg et al., 1968) supplemented with 2 % glucose (minimal medium) but lacking nitrogen and sulfur, and buffered at different pHs as indicated in the text. At 8–16 h after the transfer (as indicated in the text), the mycelia were harvested by filtration, frozen in liquid nitrogen and kept at –70 °C for protein or RNA extraction as described previously (Rolland et al., 2003). The pH values after fungal growth were measured using a pH electrode. Escherichia coli strain Sure S (Stratagene) was used for cloning and plasmid propagation and was grown in LB broth, supplemented with ampicillin (50 µg ml^–1) (Sigma) when required (Sambrook et al., 1989).

Phytopathogenicity tests.
Phytopathogenicity assays were performed on apple fruits. B. cinerea strains were grown for 15 days on sporulation medium. Harvested conidia (2x10⁶ conidia ml^–1) were germinated in liquid sporulation medium with agitation (150 r.p.m.) for 12 h at 22 °C. The surface of apple fruits was disinfected with 70 % ethanol prior to the infection. Slices of apple (approx. 1 cm thick) were inoculated with 10 µl of the germinated conidia suspension and kept in a Petri dish under high-humidity conditions at 22 °C.

Preparation of protein extracts from infected fruits.
Slices of apple collected at different times after inoculation were suspended in sterile deionized water, sonicated for 5 min at 4 °C (Vibra Cells, Sonics & Materials Inc.) and centrifuged at 4 °C for 15 min at 8000 g. The proteins present in the supernatant were precipitated overnight with (NH₄)₂SO₄ (85 % saturation) and centrifuged at 10 000 g for 20 min. The precipitate was solubilized in 3 ml water, dialysed for 6 h against 1 l distilled water twice at 4 °C and stored at –20 °C before use.

BcACP1 immunodetection.
Detection of BcACP1 by Western blotting was performed using rabbit anti-SsACP1 polyclonal antibodies at 1/1000 dilution as described previously (Rolland et al., 2003). Proteins from concentrated filtrates were separated by 12 % SDS-PAGE and transferred to a nitrocellulose membrane. The membrane was probed for 16 h with anti-SsACP1 antibodies, washed and incubated for 45 min in the presence of rabbit anti-IgG secondary antibodies as described in the manufacturer's instructions (ECL Western detection kit, Pierce Super Signal Substrate; Perbio). Chemiluminescent reaction was used to reveal the immunoreactive proteins.

RNA isolation and Q-PCR.
The mycelium of interest (100 mg) was frozen in liquid nitrogen, ground, and total RNA was extracted by phenol/chloroform separation and lithium chloride precipitation (Verwoerd et al., 1989). For Q-PCR experiments, 20 µg total RNA of each sample were treated with DNase I (Ambion) to remove genomic DNA. Absence of genomic DNA was controlled by PCR using primers (actA-F, 5'-CCGTGCTCCAGAAGCTTTGT-3', and actA-R, 5'-GTGGATACCACCGCTCTCAAG-3') specific for the actin-encoding gene BcactA (accession number AJ000335) and a Taq DNA polymerase (MP Biomedicals). Quality of total RNA was verified using an Agilent 2100 Bioanalyser, Agilent RNA 6000 Nano reagents and Agilent RNA Chips from Agilent. Total cDNA was produced by treating 5 µg DNase I-treated total RNA with Thermoscript RT (Invitrogen) as described by the manufacturer. Q-PCR experiments were performed on an ABI PRISM 7900HT (Applied Biosystems) using primers specific for the Bcacp1 gene (acp1-F, 5'-TGACGGAGACCTTTGCGAGTA-3', and acp1-R, 5'-TGCCCTCCTCGAAATCTTCA-3'), using the Power SYBR Green PCR Master Mix (Applied Biosystems) and according to the manufacturer's instructions. Following examination of the primer efficiencies, the amplification reaction was carried out as follows: 95 °C 10 min, 95 °C 15 s and 60 °C 1 min (50 cycles), 95 °C 15 s, 60 °C 15 s and 95 °C 15 s. Relative quantification was based on the method using BcactA as reference; the accuracy of the method could be expected from the high and similar amplification efficiencies of Bcacp1 and BcactA. Three independent biological replicates were analysed.

Northern analysis.
Samples (15 µg) of total RNAs were separated on 1.5 % formaldehyde agarose gel, transferred to Nytran membrane (Schleicher and Schuell) and UV cross-linked. RNA gel blot hybridizations were performed using standard protocols (Sambrook et al., 1989). The RNA hybridization probe was ³²P-labelled using a random-primed DNA labelling kit (Promega) and purified by exclusion chromatography (Spin column, Amersham Pharmacia Biotech). The membrane was hybridized overnight, in the presence of the α-³²P-labelled probe, at 42 °C in hybridization buffer [50 % (w/v) formamide, 5x SSC, 0.2 % (w/v) SDS, 1x Denhardt's solution and 100 µg denatured salmon sperm DNA ml^–1] and washed sequentially with 2x SSC, 0.1 % (w/v) SDS and 0.2x SSC, 0.1 % (w/v) SDS. The signals were detected using a phosphoimager (Cyclone, Packard).

Construction of the reporter gene vectors.
The β-glucuronidase (GUS)-encoding gene was fused to a 0.4 kb fragment of the promoter of the Bcacp1 gene of B. cinerea. Genomic DNA from B. cinerea and two specific primers, ACP4BC (5'-ACACCGCGGTGATCTGTATGGGTCTATACGGGT-3') and ACP7BC (5'-GCGCCATGGTTTGATTGATTGATTGAA-3') were used in a PCR to amplify a 400 bp fragment upstream of the start codon of the Bcacp1 gene. The PCR product was cloned upstream of the gus gene. The resulting plasmid was digested with XbaI and EcoRI to excise the gus cassette and subclone it into the Agrobacterium-mediated transformation vector pBHt2 (Mullins et al., 2001) to yield pBc400. Plasmid p18, corresponding to an Agrobacterium-mediated transformation vector containing the gus reporter gene under the control of 1.1 kb of the promoter of the Ssacp1 gene of S. sclerotiorum, has been previously described (Rolland et al., 2003). In this study, we refer to it as pSs1100. Another vector containing a promoter region of 400 bp upstream of the start codon of the Ssacp1 gene was also constructed. The reporter cassette was amplified by PCR from the vector pSs1100 using the primers ACPD2 (5'-CTGTCTAGATGAAGTCCCGTTCTGT-3') and TERNOS (5'-CTTCTAGACTAGTAACATAGATGACA-3'). This fragment was digested with XbaI and cloned into plasmid pBHt2 (Mullins et al., 2001) to yield pSs400.

Fungal transformation.
Transformation of B. cinerea BO5.10 was performed as described by Rolland et al. (2003). Agrobacterium tumefaciens LBA1126 carrying pBc400 or pSs400 was used to transform conidia of B. cinerea. The transformants arose after 15 days of incubation at 21 °C on solid malt agar medium containing hygromycin B (70 µg ml^–1) and 200 µM cefotaxim (Sigma).

GUS extraction and GUS assays.
GUS activity was measured essentially as described by Rolland et al. (2003). Mycelia were harvested, frozen in liquid nitrogen and ground. After adding 1 ml GUS extraction buffer (50 mM sodium phosphate pH 7.0, 0.1 % Triton X-100, 0.1 % N-lauroyl sarcosine, 10 mM EDTA, 10 mM β-mercaptoethanol) to 100 mg mycelium, the cells were disrupted and centrifuged at 13 000 g at 4 °C for 10 min. Samples (100 µl) of supernate were incubated at 37 °C in 1 ml GUS extraction buffer containing 1 mM p-nitrophenyl β-D-glucuronide (Sigma-Aldrich). At various time points, 100 µl samples of the reaction were mixed with 900 µl stop buffer (0.2 M Na₂CO₃, pH 9.5). Absorbance was measured at 405 nm, and the protein concentration was determined according to Bradford (1976).

BcACP1 maturation assay.
B. cinerea was grown on solid medium containing 2 % malt extract on the surface of cellophane sheets. After 3 days incubation at 21 °C, the cellophane sheets were transferred onto the surface of 2 ml Gamborg medium supplemented with 2 % glucose and buffered to pH 7. Sixteen hours after transfer, the medium was recovered, put on ice and its pH was adjusted as indicated in the text by addition of HCl. The mixture was then incubated at different temperatures, in the presence or absence of protease inhibitors, and an aliquot of 1.4 ml was recovered at different times. The reaction was stopped by precipitating the proteins (Rolland et al., 2003) and analysed by Western blotting.

Isolation and characterization of the Bcacp1 gene
Two degenerate oligonucleotides corresponding to two conserved regions of the G1 proteases (WYEWY and NAEWI) were previously used successfully to clone the acp1 gene of S. sclerotiorum. This gene encodes the protease SsACP1, whose characteristics are an acidic optimal pH for activity (pH 2.0) and an insensitivity to pepstatin A (inhibitor of aspartic proteases) (Poussereau et al., 2001). Since B. cinerea and S. sclerotiorum are closely related necrotrophic fungi, we used the same strategy to amplify a 200 bp fragment from the genomic DNA of B. cinerea. A SON (single oligonucleotide nested) PCR approach (Antal et al., 2004) was then used to clone the entire coding sequence flanked by 1 kb and 200 bp of the promoter and the terminator sequences, respectively (GenBank accession no. DQ151453). The gene was named Bcacp1 while the acp1 gene of S. sclerotiorum was renamed Ssacp1 to avoid confusion. Analysis of the recently released B. cinerea (strain BO5.10) genome sequence on the website of the Broad Institute confirmed the sequence of Bcacp1 and revealed that it is the only G1 protease-encoding gene in this fungus.

Bcacp1 encodes a protein of 257 amino acids sharing 89 % identity with SsACP1, 55 % identity with PrtA of Aspergillus niger (Takahashi et al., 1991) and 45 % identity with SGP of Scytalidium lignicolum (Maita et al., 1984). Despite high conservation between the sequences of SsACP1 and BcACP1, two notable differences exist. BcACP1 contains four extra amino acids in its putative pro-peptide sequence and two cysteines in its carboxyl-terminal region (Fig. 1A). By comparison with the other G1 proteases, BcACP1 is likely to be secreted as a zymogen and to contain an N-terminal pre-pro-sequence of 57 amino acids. The unprocessed BcACP1 is predicted to contain 239 amino acids and to have a molecular mass of 24.75 kDa. The mature BcACP1 is predicted to contain 200 amino acids and to have a molecular mass of 20.57 kDa. As previously described (Rolland et al., 2003), the secreted BcACP1 can be detected in B. cinerea culture filtrate as a single protein of 28 kDa by using antibodies raised against SsACP1 of S. sclerotiorum. The apparent molecular mass of BcACP1 is therefore 7–8 kDa higher than that predicted from its sequence. This difference could be due to post-translational modifications.

(44K): Fig. 1. BcACP1 and SsACP1 share high similarity at both the protein and the promoter level. (A) Alignment of BcACP1 and SsACP1. The predicted signal peptide sequence and the pro-sequence are indicated. The two major differences between the BcACP1 and SsACP1 sequences are highlighted in bold. (B) Schematic of the Bcacp1 and Ssacp1 promoters. The positions of the AREA, CREA, CYS3, PacC and GAL4 consensus binding sites are indicated.

Secretion of BcACP1 during plant infection
We have shown that SsACP1 is secreted by S. sclerotiorum during pathogenesis (Poussereau et al., 2001). To test whether this was also true for B. cinerea, apple fruits were infected with conidia and the production of BcACP1 was followed by Western blotting of the proteins collected from the infected tissues (Fig. 2). Primary infection occurred within 16 h post-inoculation (p.i.), a time at which necrotic areas appeared around the inoculation point. The infection then progressed to near-complete maceration of the fruit disc in 96 h. BcACP1 was detected at 96 h p.i., but a faint band could already be observed at 72 h p.i. These results indicate that BcACP1 is indeed produced when B. cinerea attacks plants. In parallel, the pH of the infection sites was monitored by using a flat contact electrode placed directly onto the necrotic area. A decrease in pH from 4.2 to 3.8 was recorded, suggesting the secretion of acids by B. cinerea. When bean leaves were used as hosts, the pH also dropped (from 6 to 4.5) during infection over 3 days, and again secretion of BcACP1 could be observed (data not shown).

(43K): Fig. 2. BcACP1 is produced during pathogenesis. (A) Pathogenicity assay. Apple fruit discs were inoculated with germinated conidia of B. cinerea. The infection was followed for 96 h post-inoculation (hpi) and the pH at the centre of the necrotic area was monitored. (B) Detection of BcACP1. Proteins collected from the infected tissues were analysed by Western blotting using anti-SsACP1antibodies.

pH regulation of Bcacp1 transcription
We investigated in vitro the role of ambient pH in the control of Bcacp1 expression. B. cinerea was grown for 3 days on solid rich medium under conditions where the presence of sulfur and nitrogen repressed Bcacp1 expression (Rolland & Bruel, 2008). The mycelium was then transferred to minimal medium buffered to different pHs. Eight hours after transfer, the mycelium was harvested and total RNA was extracted for quantitative PCR analysis. The actin-encoding gene BcactA was used as reference, and we verified that its expression did not change with pH (C_t value of 17.96 at pH 3 and 18.3 at pH 7). In contrast, expression of Bcacp1 varied significantly with pH: expression could be observed at pH 3, 4 and 5 while barely any expression occurred at pH 6 or 7 (Fig. 3A). Similarly to what was found for Ssacp1 in S. sclerotiorum (Poussereau et al., 2001), these results indicate that Bcacp1 transcription is pH dependent, requires acidic conditions to proceed, and could be maximal at pH 4.

(20K): Fig. 3. Bcacp1 transcription is controlled by the environmental pH independently of the canonical PacC consensus binding site. (A) Q-PCR analysis of Bcacp1 regulation. Following growth under nitrogen and sulfur conditions that repressed Bcacp1 expression, B. cinerea was transferred (8 h) to minimal medium buffered to pH 3 to pH 7. Total RNA was extracted and submitted to quantitative PCR analysis. Bcacp1 expression is reported relative to that of the actin-encoding reference gene (BcactA). (B) Functional activity of the Bcacp1 promoter. B. cinerea transformants carrying the GUS reporter gene under the control of a 0.4 kb promoter region of Bcacp1 were grown on rich solid medium and then transferred (16 h) to minimal medium buffered to pH 4.5 or 7.5. The mycelium was harvested and GUS specific activity was measured. (C) Functional activity of the Ssacp1 promoter. The same approach was adopted with the Ssacp1 promoter, using fragments of 0.4 kb or 1.1 kb. Means±SD are plotted (n=3).

It is well established in fungi that the regulation of gene expression in response to changes in ambient pH is mediated by a conserved signal transduction pathway acting through the transcription factor PacC and its 5'-GCCARG-3' consensus DNA-binding sequence (Penalva & Arst, 2002). PacC is present in B. cinerea (accession number AY647218), but, curiously, analysis of 1.5 kb of the Bcacp1 promoter failed to reveal any putative PacC binding site in this DNA sequence. To further investigate the regulation of Bcacp1 by pH, we constructed a reporter system by fusing 0.4 kb of the Bcacp1 promoter to the β-glucuronidase-encoding gene. This pBcacp1-GUS cassette was introduced into B. cinerea by Agrobacterium-mediated transformation (Rolland et al., 2003). Several transformants were obtained and Southern-blot analysis revealed the presence of only one copy of the transgene in each transformant's genome (data not shown). Five transformants were grown, transferred for 16 h to buffered conditions (pH 4.5 or pH 7.5), and GUS activity was measured. In all strains, a higher activity was detected when the fungus was transferred to acidic rather than neutral conditions (Fig. 3B). This result correlates with the expression profile of the endogenous Bcacp1 gene (Fig. 3A), and it indicates that pH regulation of the Bcacp1 promoter activity can occur in the absence of a consensus PacC binding site and in the presence of only 0.4 kb of the promoter sequence.

The Bcacp1 and Ssacp1 promoters share 75 % identity, and the two genes might hence share some regulators (Fig. 1B). However, one major difference is the presence of one PacC consensus binding site in position –834 bp of the Ssacp1 promoter, and the role of this site in pH regulation was examined. We first compared the capacity of 0.4 kb of the Ssacp1 promoter to support pH regulation. B. cinerea transformants carrying pSsacp1(0.4 kb)-gus constructs were generated and GUS activity was measured as described above. As shown in Fig. 3(C), the results obtained were identical to those obtained with pBcacp1-gus. Transformants carrying the gus gene under the control of 1.1 kb of pSsacp1 were next generated and studied. Placed in either an acidic or neutral environment, these transformants always exhibited a threefold higher GUS production than those carrying the short version of the promoter (Fig. 3C); this suggests the existence of sequences enhancing Ssacp1 expression within the –1.1 to –0.4 kb region of the promoter. Disregarding this difference, pH regulation still occurred with this construct and was not different from that observed with the shorter version of the promoter. pH regulation of Ssacp1 hence appears to be independent of the presence of a consensus PacC binding site, similar to what was observed with Bcacp1.

pH-dependent maturation of BcACP1
Besides being controlled by ambient pH, Bcacp1 expression is also controlled by nitrogen and sulfur catabolic repression (Rolland & Bruel, 2008). Indeed, Bcacp1 expression is activated in the absence of readily metabolizable sulfur or nitrogen sources such as sulfate or ammonia. Conversely, the presence of such sources of both nitrogen and sulfur abolishes expression. In accordance with these known types of regulation, culture of B. cinerea in malt extract medium supplemented with MgSO₄ and NH₄Cl led to no expression of Bcacp1 while culture in non-supplemented medium led to accumulation of Bcacp1 mRNA (Fig. 4, top). When the mycelia that were collected from these two different cultures were transferred to minimal medium lacking nitrogen and sulfur sources, and buffered to pH 4, secreted BcACP1 could be immunodetected in the culture broth in both cases (Fig. 4, bottom). Under acidic conditions, and in the absence of nitrogen and sulfur, this result confirmed at the protein level the Bcacp1 gene expression results shown in Fig. 3(A). When the mycelium grown in the presence of sulfur and nitrogen was transferred to minimal medium buffered to pH 7, BcACP1 could not be detected among the secreted proteins (Fig. 4, bottom left), and this again confirmed the results presented in Fig. 3(A). Lastly, and interestingly, a protein of apparent molecular mass 35 kDa was immunodetected among the secreted proteins when the mycelium grown in the absence of nitrogen and sulfur was transferred to minimal medium buffered to pH 7 (Fig. 4, bottom right). In the absence of Bcacp1 expression at this pH (Fig. 3A), this result could be explained by the translation of the mRNA that had accumulated during pre-culture and the artificial trapping of a pro-protein form of BcACP1 at this neutral pH. To test this hypothesis, the pH 7 culture broth was collected and acidified in vitro to pH 4 with HCl, incubated at 4 °C for 1–50 min, and analysed by Western blotting. As shown in Fig. 5(A), the 35 kDa band progressively disappeared over time while a band corresponding to a 28 kDa protein appeared; 50 % of the shift occurred in 30–40 min. We next investigated whether the processing was sensitive to temperature. The same experiment was performed at 10 °C and 20 °C, and 50 % of the shift occurred in 15 and 5 min, respectively, indicating that the processing was indeed temperature dependent (data not shown). The pH dependence of the observed processing was then investigated. The pH 7 culture broth was acidified to different pH values and the processing was monitored after 1 h of incubation at 4 °C (Fig. 5B). Whereas the 28 kDa band did not appear at a pH value of 5.8 or higher, it did appear at a pH value of 5.6 or lower. At pH 5.6, only about 15 % of the protein was processed under the conditions used.

(28K): Fig. 4. BcACP1 is produced as a pro-enzyme. B. cinerea was grown on solid malt extract (MA) medium supplemented or not with MgSO4 (1 mM) and NH4Cl (50 mM). Using ethidium bromide staining of rRNA to approximate the amount of total RNA, the Bcacp1 mRNA level was monitored at the end of the preculture by Northern blotting (top panel). The mycelium was then transferred to minimal medium lacking nitrogen and sulfur, and buffered to pH 4 or pH 7. After 16 h of culture under these conditions, the secreted proteins were analysed by Western blotting using anti-ACP1 antibodies (bottom panel).

(40K): Fig. 5. The maturation of BcACP1 is pH dependent. (A) Kinetics of BcACP1 pH-dependent maturation. The pH 7 culture medium containing the secreted 35 kDa form of BcACP1 was acidified in vitro with HCl to pH 4 and incubated at 4 °C for 1 to 50 min. Precipitation and Western-blot analysis of the proteins using anti-ACP1 antibodies were performed on six aliquots over time. (B) pH dependence of the maturation process. The same culture medium was acidified to different pH values (as indicated) and incubated for 1 h at 4 °C. The proteins were precipitated and analysed by Western blotting.

Transformation of the 35 kDa protein into its 28 kDa counterpart most likely resulted from proteolysis. Whether this proteolysis relied on proteases present in the culture broth or on BcACP1 itself was finally explored. The in vitro processing experiment was performed in the presence of pepstatin A (aspartic protease inhibitor), leupeptin (serine and cysteine protease inhibitor), E64 (cysteine protease inhibitor), NEM (N-ethylmaleimide, cysteine protease inhibitor) or EDTA (metalloprotease inhibitor), a series of inhibitors that block most if not all the main families of proteases, except the G1 family of proteases. None of these inhibitors prevented the appearance of the 28 kDa band (data not shown). Dilution of the sample before acidification was also carried out, and no effect on the maturation kinetics was observed (data not shown). Taken together, these results suggest first that the 35 kDa protein detected by the anti-SsACP1 antibodies corresponds to a pro-protein form of BcACP1. Second, the results suggest that this form can mature to BcACP1 (28 kDa) via a temperature- and pH-dependent maturation process which probably relies on autocatalytic proteolysis. We have cloned the Bcacp1 gene encoding a member of the G1 family of proteases in B. cinerea. This gene is orthologous to the S. sclerotiorum Ssacp1 gene previously studied (Poussereau et al., 2001). The G1 proteases are endopeptidases found exclusively in ascomycete species (Sims et al., 2004). Comparison of Bcacp1 and Ssacp1 revealed that the coding regions of these two genes exhibit high DNA similarity. The BcACP1 and SsACP1 proteins are 89 % identical, allowing the polyclonal anti-SsACP1 antibodies to recognize the BcACP1 protein. Remarkably, high conservation is also observed in the promoter regions of the two genes, suggesting possible binding of conserved regulators. Comparison of the now available genomes of B. cinerea and S. sclerotiorum (Broad Institute) revealed that the two genes encode a unique G1 protease in each fungus. Using apple fruit or bean leaves as hosts, we showed that BcACP1 is produced during infection concomitantly with acidification of plant tissues. This is reminiscent of what has been observed with SsACP1 in S. sclerotiorum (Poussereau et al., 2001) and suggests that these two closely related fungi behave similarly during infection with respect to the production of this G1 protease.

We investigated the role of environmental pH in the regulation of Bcacp1. Q-PCR analysis of B. cinerea transferred to different pH conditions showed expression of Bcacp1 at pH 3, 4 and 5 but little or no expression at pH 6 and pH 7. Bcacp1 is therefore an acid-expressed gene and its expression profile is similar to that of Ssacp1 (Poussereau et al., 2001). Based on this result, BcACP1 production would be expected during plant infection as soon as the fungus met an acidic pH environment. Surprisingly, BcACP1 is not detected at the early time points of apple fruit infection while pH is permissive to its production. This could be explained by a shortage of fungal biomass and the production of subdetectable amounts of protease. Alternatively, other regulations besides pH could also occur in planta that could prevent Bcacp1 expression in the early phases of infection.

In fungi, the zinc finger transcription factor PacC plays an important role in the cell's response to ambient pH. Through binding to a 5'-GCCARG-3' consensus DNA sequence, PacC is considered an activator of alkaline-expressed genes (Penalva & Arst, 2002) and it was shown to repress one acid-expressed gene (Espeso & Arst, 2000). Other acid-expressed genes, however, lack the consensus PacC binding site upstream of their coding sequence (Sarkar et al., 1996). The absence of such a site in the Bcacp1 promoter and its presence at position –834 in the Ssacp1 promoter prompted us to perform a functional analysis of these two promoters. Transcriptional fusions using the GUS reporter gene were introduced in B. cinerea, and GUS activity measurements showed that pH regulation of Ssacp1 and Bcacp1 is independent of the presence of a consensus PacC binding site. We cannot rule out that the results on Ssacp1 regulation could differ when investigated in S. sclerotiorum, but our results would suggest that Bcacp1 and Ssacp1 add to previously described acid-expressed genes whose pH regulation does not operate through a consensus PacC binding site. Our results could indicate that PacC regulation of Bcacp1 and Scacp1 proceeds through another DNA motif. One such motif exists in Candida albicans (5'-CCAAGAA-3'; Ramon & Fonzi, 2003), but it is absent in the Bcacp1 and Ssacp1 promoters. Alternatively, the results could indicate that Bcacp1 and Ssacp1 are not regulated via PacC, as shown for several genes in C. albicans (Davis et al., 2000). However, a change in Ssacp1 transcription in a strain of S. sclerotiorum carrying an activating allele of the PacC orthologue (Kim et al., 2007) suggests that PacC plays a role in Ssacp1 regulation, but the molecular details of this regulation remain unknown and an indirect intervention of PacC cannot be ruled out.

In the course of this work, we found experimental conditions that allowed trapping and visualization of the pro-protein form of BcACP1. When grown at acidic pH in the absence of simple sulfur and nitrogen sources, B. cinerea produces high levels of transcripts, and subsequent transfer of the fungus to minimal medium buffered to pH 7.0 results in the secretion of a 35 kDa form of BcACP1. Transfer of this protein to acidic pH results in its conversion into the 28 kDa form that is detected in the fungus secretome during plant infection or growth in vitro under acidic conditions. We showed that this conversion is pH dependent and most likely autocatalytic. Indeed, inhibition of most if not all known protease families did not block the 35 kDa to 28 kDa processing of BcACP1, and dilution had no effect on the maturation kinetics. This result is consistent with previous studies performed with other members of the G1 family of proteases (Kataoka et al., 2005; Huang et al., 2007), but we cannot exclude the existence of an unknown family of proteases involved in the processing of BcACP1. We tried to use the same experimental approach to reveal the pro-protein form of SsACP1 of S. sclerotiorum but without success. This could be due to the difference in the stability of Ssacp1 mRNA or SsACP1 protein compared to their Bcacp1 counterparts.

In conclusion, the present study highlights the importance of pH in the control of the production of a new member of the G1 family of proteases, BcACP1. Ambient pH regulates both transcription and post-translational processing of this secreted enzyme, and such a tuning system probably participates in the adaptation of B. cinerea to a dynamic pH environment.

This work was supported by grants from the CNRS (UMR 5240) and the Université Claude Bernard, Lyon 1, France. S. R. was supported by a grant from the French MENRT. The B. cinerea strain was kindly provided by Professor P. Tudzynski (Institut für Botanik, Westfläische Wilhelms-Universität, Münster, Germany).

Edited by: J.-R. Xu

Footnotes

†Present address: Dartmouth Medical School, Genetics Department, Remsen 7400 Building, Hanover, NH 03755, USA.