Expression of Five Endopolygalacturonase Genes and Demonstration that MfPG1 Overexpression Diminishes Virulence in the Brown Rot Pathogen Monilinia fructicola

Monilinia fructicola is a devastating pathogen on stone fruits, causing blossom blight and fruit rot. Little is known about pathogenic mechanisms in M. fructicola and related Monilinia species. In this study, five endopolygalacturonase (endo-PG) genes were cloned and functionally characterized in M. fructicola. Quantitative reverse-transcriptase PCR (qRT-PCR) revealed that the five MfPG genes are differentially expressed during pathogenesis and in culture under various pH regimes and carbon and nitrogen sources. MfPG1 encodes the major endo-PG and is expressed to significantly higher levels compared to the other four MfPGs in culture and in planta. MfPG1 function during pathogenesis was evaluated by examining the disease phenotypes and gene expression patterns in M. fructicola MfPG1-overexpressing strains and in strains carrying the β-glucuronidase (GUS) reporter gene fused with MfPG1 (MfPG1-GUS). The MFPG1-GUS reporter was expressed in situ in conidia and hyphae following inoculation of flower petals, and qRT-PCR analysis confirmed MfPG1 expression during pathogenesis. MfPG1-overexpressing strains produced smaller lesions and higher levels of reactive oxygen species (ROS) on the petals of peach and rose flowers than the wild-type strain, suggesting that MfPG1 affecting fungal virulence might be in part resulted from the increase of ROS in the Prunus–M. fructicola interactions.


Introduction
Brown rot blossom blight and fruit rot caused by the fungal pathogen Monilinia fructicola (G. Wint.) Honey and related Monilinia species is a severe disease problem of stone fruits worldwide. The disease causes extensive tissue maceration in affected blossoms and fruits. Biochemical analyses suggest that cell wall-degrading enzymes (CWDEs) produced by the pathogen play a critical role during M. fructicola pathogenesis [1]. M. fructicola is able to produce and secrete pectin lyases and polygalacturonases in axenic cultures. The addition of pectic substances increases the production of CWDEs substantially, implicating an important role of CWDEs in nutrient acquisition. The exact function of CWDEs in M. fructicola pathogenesis is remains largely uncertain, although our previous studies have shown that the principal cutinase, MFCUT1, is a virulence factor [2].
Fungi produce a wide array of CWDEs and a number of these are critical for virulence in phytopathogenic fungi [3]. Endopolygalacturonases (endo-PG; EC 3.2.1.15), which cleave internal O-glycosidic bonds of pectate polymers in plant cell walls, are among the most important CWDEs in the plant-microbe interactions. Multiple endo-PG genes have been isolated and characterized in Botrytis cinerea and Sclerotinia sclerotiorum which, like M. fructicola, are members of the family Sclerotiniaceae. B. cinerea contains six differentially regulated endo-PG genes, including BcPG1 and BcPG2 that are expressed constitutively in culture, BcPG4 and BcPG6 that are strongly induced by galacturonic acid, and BcPG3 that is preferentially expressed at low pH [4]. Inactivation of the BcPG1 gene in B. cinerea resulted in a strain that causes significantly smaller lesions on tomato leaves relative to the wild-type, supporting a virulence function of this endo-PG in gray mold disease [5]. In contrast, genetic analysis of T4BcPG1, a BcPG1 homologue in the T4 strain of B. cinerea, revealed that this endo-PG can act as an elicitor to induce defense responses in grapevine [6]. S. sclerotiorum has four endo-PG coding genes [7]. Among them, only SSPG1d is highly expressed during infection [7]. These findings point to the complex role of fungal endo-PGs during pathogenesis and suggest that the expression of these CWDEs must be tightly coordinated for optimal colonization of the host by the pathogen.
Previously we have established a DNA transformation system for M. fructicola [8] and demonstrated that formation of appressorium and expression of MfCUT1 (a cutinase gene) are required to penetrate into plant tissue [2,9]. To provide a better understanding of M. fructicola pathogenicity, we focused on genes that are involved in plant cell wall degradation and investigated their function on brown rot disease development. Since the genome sequence of M. fructicola is not yet available, we cloned five M. fructicola endoPG genes-MfPG1, MfPG2, MfPG3, MfPG5 and MfPG6 by PCR and determined their expression in axenic culture and in planta during pathogenesis. De Cal and colleagues have shown that M. fructicola could acidify the infected tissue during colonization on peach and nectarine fruit and that four of the MfPGs were regulated by pH in axenic culture [10]. However, the involvement of the five MfPGs in fungal growth and pathogenesis remains to be determined. The pathological role for MfPG1 in lesion development was assessed by examining the host reaction after inoculation with wildtype and MfPG1-overexpressing strains of M. fructicola. Spatial and quantitative expression of MfPG1 in inoculated host tissue was determined by in situ detection of β-glucuronidase (GUS) fused with MFPG1 and by qRT-PCR. The results indicate that overexpression of MfPG1 actually increases reactive oxygen species (ROS) accumulation and reduces lesion development in M. fructicola host interactions.

Cloning and characterization of MfPG genes
Five endopolygalacturonase (endo-PG) genes, designated MfPG1, MfPG2, MfPG3, MfPG5 and MfPG6, were cloned from M. fructicola (S1 Table). Analysis and comparison of the assembled sequences from cDNA and genomic DNA revealed that MfPG1 is an intronless gene, while the other four genes contain one to four introns with sizes from 49 to 62 bp. The MFPG1, MFPG2, MFPG5 and MFPG6 proteins have predicted molecular weights ranging from 35 to 37 kDa. In contrast, the MFPG3 protein has a predicted molecular weight of 50.1 kDa. MFPG1, MFPG2 and MFPG3 are likely basic PG proteins with a predicted pI of 9. MFPG5 and MFPG6 have a predicted pI of 5.
The conceptually translated MFPGs belong to the glycosyl hydrolase family 28 which, after protein domain annotation, includes polygalacturonase and rhamnogalacturonase A [11]. All MFPGs contain sites for substrate hydrolysis (NT/SD, DDC and GGHGLS) and for substrate binding (RI/VK) (Fig 1). The consensus sequence (CSGGHGLSI/VGS) required for polygalacturonase activity was found at the C-terminus of all MFPGs. A secretory signal sequence of 16-21 residues was found at the N-terminus of all MFPGs.
Phylogenetic relationships of the M. fructicola MFPGs to other PG protein homologs from fungi, oomycetes and plants revealed that all PG proteins found in the family of Sclerotiniaceae can be grouped into six monophyletic clades (S1 Fig). Interestingly, PG3 and PG6 are less related to PG1, PG2, and PG5 than to the PG proteins from Penicillium griseoroseum (PgGII) and Aspergillus niger (PGAII) Analysis of the 687-bp or 1-kb sequences upstream of the putative ATG translational start codon of each of the MfPG1, MfPG2, MfPG3 and MfPG5 genes identified several putative binding sites for diverse transcriptional regulators (S2 Fig). All MfPG promoter regions contain at least one carbon catabolite repressor (CreA) binding site and the nitrogen-inducible activator (AreA/NIT2) sites. A pH-responsive transcription factor pacC binding site was found 876 and 90 nucleotides upstream from the ATG codon of MfPG3 and MfPG5, respectively. No pacC binding sites was found in the promoter of MfPG1 and MfPG2. The promoter region of all MfPGs except MfPG3 contain multiple TATA boxes (TBP; -76 and -26), a binding site for the activator of RNA polymerases. The MfPG6 upstream region was not available for this analysis since only 60-bp sequence was obtained. The expression of polygalacturonase gene family members has been known to be regulated by pH in many fungal pathogens [12]. Northern blot hybridization revealed that the MfPG1 gene was preferentially expressed under acidic conditions. However, the expression of MfPG2, MfPG3, MfPG5 and MfPG6 could not be detected by Northern blotting. The expression level of each MfPG gene was further analyzed by qRT-PCR (S3 Fig). MfPG1 expression was dominant at pH 3.5 to 4.0. MfPG2 expressed preferentially at pH values 3.5 and 6.0. The expression of MfPG3 and MfPG6 was favored when the fungus was grown at neutral pH, while the expression of MfPG5 was slightly higher when the fungus was grown at pH 4.0 than at the other pH's tested. Comparison of the expression levels among five MfPGs confirmed further that MfPG1 was the predominant homolog expressed in vitro (S2 Table). At the end of the growth period, the pH in all media did not change significantly.
Expression of the polygalacturonase genes has been known to be regulated by carbon and nitrogen sources [13]. Because CreA binding sites were found in the 5' untranslated region of MfPGs, the expression of MfPGs was examined in M. fructicola grown in medium amended with various carbon sources. Expression of the MfPG1 gene was higher when the fungus was grown in pectin-containing medium than in PDB or medium containing galacturonic acid or glucose as the sole carbon source (Fig 2A). The MfPG3, MfPG5 and MfPG6 genes were expressed to high levels when the fungus was grown in medium amended with galacturonic acid. Expression of the MfPG6 also was enhanced by pectin. The MfPG2 gene was expressed at low levels regardless of carbon sources. There was no significant difference in the accumulation of the MfPG2 gene transcript when the fungus was grown in medium containing various carbon sources. Glucose apparently repressed the expression of MfPG5.
Because AreA/NIT2 binding sites were found in the 5'-untranslated region of MfPGs, expression of MfPGs was also examined in M. fructicola grown in medium amended with different nitrogen sources (Fig 2B). Only expression of the MfPG1 gene was greatly enhanced when the fungus was grown in medium containing ammonium nitrate or yeast extract as the sole nitrogen source.
Our previous studies have shown that H 2 O 2 and caffeic acid impact gene expression and virulence of M. fructicola [14]. The expression of MfPG1 did not change after M. fructicola was shifted to a H 2 O 2 -containing medium. However, qRT-PCR analysis revealed that expression of MfPG1 was slightly inhibited when H 2 O 2 was added directly to M. fructicola cultures ( Fig 2C). As assessed by Northern blotting, MfPG1 transcript levels increased when M. fructicola was shifted to medium containing caffeic acid for 3 h and slightly decreased after 24 h ( Fig 2D). The pH changes in culture filtrates among the treatments were within 0.2 units.
MfPGs are differentially expressed during M. fructicola pathogenesis All MfPG gene transcripts could be detected to various levels in M. fructicola inoculated onto peach petals at 6 h post inoculation (hpi) (Fig 3A). Accumulation of the MfPG1, MfPG2 and MfPG3 gene transcripts increased significantly during symptom development. Although the MfPG5 and MfPG6 transcript levels did not change dramatically, their expression was up-regulated in M. fructicola during pathogenesis in peach petals. On rose petals inoculated with M. fructicola, MfPG1 transcript levels were abundant and increased over time during colonization ( Fig 3B). The expression level of all five MfPG genes was compared, revealing that MfPG1 was expressed at the highest level at each time point on peach and rose petals (S3 Table). Overall, MfPG1 gene was expressed equally well both in culture and in planta (S2 and S3 Tables). Both MfPG2 and MfPG3 were highly expressed in planta compared to their expression in axenic cultures. The expression of MfPG5 and MfPG6 was lower in planta than in axenic cultures.

Expression of the MfPG1-GUS reporter during pathogenesis
An MFPG1-GUS fusion protein was constructed to examine the expression pattern of MfPG1 during M. fructicola pathogenesis. Transformation of pNC-MFPG-GUS plasmid into fungal protoplasts recovered five transformants (PG20, PG50, PG55, PG213, and PG249) showing blue pigments on pectin-containing medium after staining with X-gluc, indicative of GUS activity. The colonies of the wild-type strain were not stained blue under similar culture conditions. GUS activity was detected in these transformants inoculated on peach petals at 5 hpi (Fig 4). Expression of MFPG1-GUS was also detected in transformants on apple and peach fruits at 4 dpi. Microscopic examination revealed MFPG1-GUS expression in conidia, germ tubes and appressoria during penetration on peach petals. GUS activity was not detectable in wild-type inoculated on peach petals or apple fruit. The MfPG1-overexpressing strains (4-1, 5-1, and 8-2) produced significantly smaller brown rot lesions than those induced by the wild-type strain on rose and peach petals (Table 1). qRT-PCR analysis revealed that MfPG1 mRNA levels were significantly higher in both rose and peach petals inoculated with the 4-1, 5-1 and 8-2 strains than in host tissues infected with the wild-type strain M1 (Fig 5A). Expression of the MfPG2, MfPG3, MfPG5 and MfPG6 genes was up-regulated at 24 hpi in rose petals inoculated with the MfPG1-overexpressing strain (Fig 5B and S4 Table). The accumulation of transcripts for MfPG2 and MfPG5 was significantly higher in the overexpression strains (4-1 and 5-1) than wild-type inoculated on rose and peach petals. The pH of petal tissue changed during the course of pathogenesis, decreasing from pH 5.5 at 0 hpi to 4.5 at 48 hpi. Assays for enzyme activity revealed that the MfPG1-overexpressing strains displayed higher polygalacturonase activity than the wild-type strain (S6 Fig). No notable differences in growth, germination, and appressorium formation were observed between the wild-type strain and the overexpression strains in axenic culture and in planta.
Reactive oxygen species (ROS) are associated with cell death and often with the development of hypersensitive resistance. ROS accumulation was measured in rose petals after inoculation with the MfPG1 overexpressors and the wild-type. MfPG1-overexpressing strains induced ROS accumulation at levels higher than the wild-type around the infection site at 6 hpi ( Fig 6A and S7 Fig). The relative expression level of MfPG1 in the MfPG1-overespressing strain was analyzed by qRT-PCR. The results revealed that the MfPG1-overexpressing strain expressed higher level of MfPG1 than those of the wild-type strain at all examined time points (Fig 6B). Rose petals inoculated with the wild-type strain induced chlorotic lesions at the site of infection 10 hpi. Rose petals challenged with the MfPG1-overexpressing strain developed lesions at rates much slower than the wild-type strain (Fig 6C).   functional analysis suggested that MfPG1 impacts fungal virulence, as the MfPG1-overexpressing strains induced smaller brown rot lesions than the wild-type strain on rose and peach petals. The five PG genes were found to be differentially expressed in culture and in inoculated host tissue. These different patterns of expression may reflect a degree of selectivity in substrate utilization among the various homologs during saprophytic and parasitic growth. All five MfPG genes from M. fructicola share sequence similarities of endo-PG genes from B. cinerea and S. sclerotiorum, both members of the Sclerotiniaceae as is M. fructicola. However, the endo-PGs found in B. cinerea and S. sclerotiorum share greater similarity with each other than with their counterparts in M. fructicola. This suggests that B. cinerea and S. sclerotiorum are evolutionarily closer than M. fructicola, and analysis of rDNA sequences corroborates this genetic correspondence among these species [15].
The production of plant CWDEs, often regulated by transcription factors, such as CreA, AreA/NIT2 and pacC, is important for fungal growth and pathogenicity in vitro and in planta [12,13,16,17,18,19,20]. Analysis of the upstream regions of MfPG1, MfPG2, MfPG3 and MfPG5 revealed a number of potential regulatory motifs that may contribute to the differential expression of these genes in response to experimental treatments. The promoter regions of MfPG1, MfPG2 and MfPG3 contain one or two CRE binding sites, but the promoter region of MfPG5 has four CRE binding sites. qRT-PCR analysis revealed that expression of MfPG5, but not MfPG1, MfPG2 and MfPG3, was down-regulated in medium containing glucose as the sole carbon source, providing strong evidence for carbon catabolite repression. When grown on pectin amended medium, MfPG1, MfPG2 and MfPG5 are preferentially expressed under acidic conditions, while MfPG3 and MfPG6 are expressed to high levels at neutral pH. The pacC binding site found in the promoter region of MfPG3 might contribute to the up-regulation of this gene at neutral to alkaline condition. Although the promoter region of MfPG5 also contains a putative pacC binding site, immediately adjacent to a TBP binding site, the expression of MfPG5 was repressed at alkaline environment. pacC is a transcription factor that up-regulates alkaline-expressed genes and downregulates acid-expressed genes [16]. The expression patterns of all MfPG genes in response to pH are consistent with the reported findings [10]. The pH regulation on the expression of MfPG genes might in part contribute to their expression during pathogenesis. For example, MfPG1, MfPG2 and MfPG5 were acid-inducible genes and upexpressed at the late stage of infection, in which the pH of infected petal tissue became acidic during pathogenesis. MfPG1, MfPG3, MfPG5 and MfPG6 were up-regulated in the presence of galacturonic acid and MfPG5 transcripts were accumulated at higher level in MfPG1-overexpressing strain than the wild-type strain during infection. This indicates that MfPG5 expression could be up-regulated by galcturonate released by MFPG1 during fungal infection. Galacturonic acid is a monosaccharide that is released from pectin after PG hydrolysis and has been shown to up-regulate the inducible gene family members in S. sclerotiorum [21]. Because the expression of MfPG1, MfPG3, MfPG5 and MfPG6 is up-regulated by galacturonic acid and/or pectin in axenic culture, it is tempting to speculate that these MFPGs are functioning in cell wall degradation and nutrient acquisition during M. fructicola colonization within plant hosts.
Nitrogen has long been known to influence the production of pectolytic enzymes in fungi [13]. For example, ammonium enhanced polygalacturonase production by Penicillium expansum [22]. The transcription factors AreA in Aspergillus nidulans and NIT2 in Neurospora 2,7-dichlorofluorescin diacetate (DCFH-DA) at 4, 6, 10 and 24 hpi. (B) Expression of MfPG1 in M. fructicola strains was detected by qRT-PCR. The expression level of MfPG1 in the overexpressing strain was compared to that of wild-type using the comparative C T method. (C) The wild-type (M1) and overexpressing (4-1) strains were paired inoculated on rose petals and lesion developments at 6, 10 and 24 hpi were presented. doi:10.1371/journal.pone.0132012.g006 crassa are important regulatory proteins for nitrogen metabolism [13]. Four putative AreA/ NIT2 binding sites are found in the MfPG1 promoter region and one or two AreA/NIT2 binding sites are found in the promoter regions of MfPG2, MfPG3 and MfPG5. However, only the expression of MfPG1 is up-regulated when ammonium nitrate or yeast extract is used as the sole nitrogen source, suggesting that multiple AreA/NIT2 binding sites contribute to the unique regulation of MfPG1 in response to different nitrogen sources both in axenic culture and in planta.
All five MfPG genes are up-regulated to different degrees in M. fructicola during host infection, suggesting a role during M. fructicola pathogenesis. The high level of MfPG1 expression relative to other MfGP genes in axenic culture is similar to that observed during infection and colonization of host tissue. An interesting finding of this study is that MfPG1 seems to play a negative role in fungal virulence. MfPG1 is constitutively expressed in hyphae, conidia, appressoria and germ tubes during fungal colonization, as evidenced by MFPG1-GUS fusion protein expression. Although M. fructicola can be genetically transformed, generation of disruptants with homogeneous nuclei has proven to be difficult [2,8]. Similarly, we were unable to recover any homogeneous MfPG1 mutants in this study after numerous attempts using a split-marker gene replacement approach [23]. To evaluate the role of MfPG1 in pathogenicity, fungal strains overexpressing MfPG1 were created and analyzed. The MfPG1-overexpressing strains accumulated higher levels of the MfPG1 transcript than wild-type when they were inoculated onto rose and peach petals. The MfPG1-overexpressing strains induced smaller lesions than those induced by the wild-type, supporting that MfPG1 suppresses virulence in these interactions. The smaller lesions observed in rose and peach petals inoculated with the MfPG1-overexpressing strains relative to the wild-type could be in part attributable to higher levels of ROS accumulation induced in the host tissue. The results suggest that the product of MfPG1 could function as an elicitor, inducing plant defense, which partially limits lesion development. As demonstrated in the T4 strain of B. cinerea, a BcPG1 homologue indeed elicits defense responses in grapevine [6]. In other phytopathogenic fungi, endo-PGs can function as virulence factors [5,6,24,25]. MfPG1 could be required for generating the initial pectic fragments (oligogalacturonides) that in turn induce the expression of other MfPG genes as well. A recent study [26] also found that expression of S. sclerotiorum SsSPG1 is significantly higher in plant tissues infected by a superoxide dismutase mutant than the wild-type strain. The superoxide dismutase mutant is less virulent and induces higher ROS accumulation than the wild-type strain on pea leaves.
Our previous work has demonstrated that the expression of MfCUT1, a virulence factor that encodes the principle cutinase is regulated by redox status in M. fructicola [2,14,27]. The expression of MfCUT1 is induced by H 2 O 2 and repressed by the fruit phenolic caffeic acid, a pattern that is opposite to what we observed with MfPG1. In contrast, expression of MfPG1 is down-regulated by H 2 O 2 and up-regulated by caffeic acid, suggesting a relation between the redox regulation and pathogenicity roles of MfPG1 and MfCUT1 during pathogenesis. Because overexpression of MfPG1 in strains results in smaller lesions induced by those strains, the display of MfPG1 must be correctly orchestrated during infection for optimal virulence.
In spite of their worldwide impact as devastating pathogens of orchard crops, there have been relatively few studies of pathogenic mechanisms in M. fructicola and related Monilinia species. However, recent advances in experimental methods and the genomic tools coming online are enabling mechanistic studies that will lead to a deeper understanding of pathogenesis by Monilinia species.

Strains and cultivation
The wild-type MUK-1 (M1) strain of M. fructicola used in this study was single-spore isolated from diseased peach fruit and has been previously characterized [28]. Fungal strains were cultured at 23°C. Plasmids were propagated in Escherichia coli XL1-Blue competent cells (Stratagene, La Jolla, CA).

Cloning and sequencing of MfPGs and their flanking regions
Oligonucleotide primers used in this study are given in S5 Table. A 1.0-kb MfPG1 DNA fragment was amplified from genomic DNA of M. fructicola with the primers MFPG-F2 and MFPG-R1 that are complementary to highly conserved regions of the endo-PG-coding genes from B. cinerea (BcPG1; Accession No. U68715) and S. sclerotiorum (SsPG1d; Accession No. L29041). A 0.7-kb MfPG2 DNA fragment was amplified with the primers MFPG2-F2 and MFPG2-R1.

MFPG1-GUS fusion and MfPG1 overexpression constructs and fungal transformation
The MFPG1-GUS fusion construct was made in the binary vector pNC1381Xa carrying an nptII cassette and gusA ORF [2]. A 2.4-kb fragment encompassing the entire MfPG1 ORF and its 5'-untranslated region was amplified from genomic DNA with the primers MFPG-EcoRI-901 and MFPG-SpeI-3026. After digestion with EcoRI and SpeI, the amplicon was cloned into pNC1381Xa to produce pNC-MFPG-GUS. For MfPG1 overexpression, a DNA fragment containing the entire MfPG1 ORF and its 5'-(1 kb) and 3'-untranslated (0.3 kb) regions was amplified from genomic DNA with the primers MFPG1-SmaI and MFPG1-SacI using the KOD HiFi polymerase kit. The amplicon was digested with SmaI and SacI and then cloned into pUCATPH [29] to generate the plasmid pUCATPH-MFPG1. The resultant plasmids were transformed into protoplasts prepared from the wild-type strain of M. fructicola as previously described [2]. For transformation, 10 μg DNA was mixed with 10 μl spermidine (50 mM) and 200 μl protoplast (5 x 10 7 /ml). After 30 min incubation in ice, 1.25 ml PEG solution (40% PEG, 50 mM CaCl 2 , 50 mM Tris-HCl, pH 8.0) was added into the mixture and incubated at room temperature for an additional 25 min. The protoplasts were spread on SH medium [30] supplemented with G418 sulfate or hygromycin.

Expression of MfPGs
To examine the influence of pH, nitrogen source and exogenous redox compounds on the expression of MfPGs, freshly prepared conidial suspensions (2×10 5 /ml) were transferred to a 9-cm petri dish with 20-mL starter medium containing 0.1% yeast extract, 0.1% glucose, and modified Czapek-Dox salts [14]. After incubating for 3 days, fungal mycelium was washed twice and then immersed in 20 mL modified Czapek-Dox salts supplemented with 1% citrus pectin (second medium). The pH of second medium was adjusted to appropriate values with a 0.1 M sodium citrate/phosphate buffer (pH 3.5, 4.0, 5.0, 6.0 and 7.0) or 0.1 M phosphate buffer (pH 8.0). To determine the effect of nitrogen sources, the second medium (pH 6.0) was supplemented with 0.1% yeast extract or 0.1% ammonium nitrate. To determine the effect of caffeic acid, the second medium was supplemented with 1% pectin as previously described [14]. Mycelia were harvested at 3 or 24 h post-incubation (hpi) for RNA isolation.
To assess the effect of different carbon sources on the expression of MfPGs, fungal strain was grown in 20 ml PDB medium (pH 6.0) or medium containing modified Czapek-Dox salts (pH 6.0), 0.1% yeast extract and a carbon source (1% glucose, 1% pectin or 1% galacturonic acid) for 48 h. To verify the effect of H 2 O 2 on MfPG1 expression, fungal spores were inoculated into a 20 ml medium containing modified Czapek-Dox salts (pH 6.0), 0.1% yeast extract and 1% pectin for 48 h cultivation and then treated with H 2 O 2 . Each treatment had three replicates within an experiment and at least two independent experiments were performed.

Extraction and manipulation of nucleic acids
Fungal DNA and RNA extraction was conducted as described (Lee et al., 2010). RNA was extracted from inoculated plant tissues using a RNA extraction kit (GeneMark, Taiwan) or from mycelia as described previously [14]. Standard procedures were used for endonuclease digestion of DNA, electrophoresis, and Southern-and Northern-blot hybridizations. DNA probes used for hybridization were labeled with digoxigenin (DIG)-11-dUTP (Roche Applied Science) by PCR with gene-specific primers. Probe labeling, hybridization (at 55°C), post-hybridization washing and immunological detection of the probe using a CSPD chemiluminescent substrate for alkaline phosphatase were conducted according to the manufacturer's recommendations (Roche Applied Science).

Real-Time quantitative reverse transcriptase PCR
Purified RNA was treated with DNase and used to synthesize cDNA with the MMLV high performance reverse transcriptase (Epicentre, USA) and an oligo-dT primer (5'-T25VN-3'). Primers used for quantitative PCR (qPCR) were designed using the Roche on-line design tool (http://www.roche-applied-science.com/sis/rtpcr/upl/index.jsp?id=UP030000) (S5 Table). The qPCR was conducted with a Rotor-Gene Q PCR machine (QIAGEN) using HOT FIREPol Eva-Green qPCR Mix Plus kit (Solis BioDyne). The cycling profile for amplification was the following: 95°C for 15 min, followed by 45 cycles of 95°C for 20 s, 58°C for 15 s and 72°C for 15 min. Amplification of specific transcripts was confirmed by melting curve and agarose gel electrophoresis staining with ethidium bromide. Amplification of the M. fructicola β-tubulin gene was used as an internal control. The ΔCt values (MfPG Ct-β-tubulin Ct) from three replicates in each experiment were used to calculate the relative expression levels using a comparative C T method [31].

Plant inoculations
Assays for fungal pathogenicity were performed on detached flower petals of peach (cv. "Tian-Tao") and rose (Rosa "Flaming Peace") as well as apple fruits (Malus domestica cv. "Fuji"). The hybrid tea rose cv. Flaming Peace was identified as an alternative host to complement pathogenicity assays of M. fructicola on stone fruits, which are seasonally restricted [14]. Apple fruits were wounded prior to inoculation. Inoculation was performed by applying droplets (5 μl) of conidial suspensions (10 5 spores/ml) onto host surface. Lesion size was measured as previously described [2] and analyzed statistically with the Excel paired t-test.

Detection of ROS
Conidia suspension (1x10 5 /ml) prepared from the wild-type strain and MfPG1-overexpressing strains were inoculated onto rose petals (30 μl per drop). Reactive oxygen species (ROS) was detected by 2,7-dichlorofluorescin diacetate (DCFH-DA) applying on the inoculation site at a final concentration of 60 μM. The treated petals were incubated at 37°C for 30 min and examine for fluorescence. The fluorescence was measured with a multimode reader (Infinite M200, Tecan Group) with excitation wavelength at 485 nm and emission wavelength at 535 nm.

Enzyme activity assays
Fungal strains were cultured in modified Czapek-Dox salts supplemented with 0.1% yeast extract and 1% polygalacturonate for six days and culture filtrate was collected for enzymatic activity detection using a cup plate assay. The culture filtrate was added into a 3-mm-diameter well of agar plate containing polygalacturonate (0.5%), agar (2%) and sodium acetate (0.1 M, pH 5.3) and then incubated for 8 h at 37°C. Area of clear zone, indicative of polygalacturonase activity, was measured after the plate was flooded with 1% cetyl trimethyl ammonium bromide [32].
The upstream regions of MfPG genes were analyzed by the TESS program using the standard filters. Only domain sequences in the MFPG promoters resulting from this analysis having La/ parameter values of 2.0 and La parameter values higher than 12 were considered to be potential transcription factor binding sites. Multiple sequence alignments were performed using ClustalX 1.81 software (http://www.clustal.org/). Phylogenetic tree was constructed by the neighbor-joining methods with the PAUP Ã 4.0β10 software (Sinauer Associates, Sunderland, MA, USA). The confidence values of the branches were determined by bootstrap analysis with 1000 replicates.