Gene 479 (2011) 29–36

Contents lists available at ScienceDirect

Gene

journal homepage: www.elsevier.com/locate/gene

Methods paper Evolutionary analysis of glycosyl hydrolase family 28 (GH28) suggests lineage-specific expansions in necrotrophic fungal pathogens

Daniel D. Sprockett, Helen Piontkivska, Christopher B. Blackwood ⁎

Department of Biological Sciences, Cunningham Hall, Kent State University, Kent, OH 44240, USA article info abstract

Article history: Glycosyl hydrolase family 28 (GH28) is a set of structurally related enzymes that hydrolyze glycosidic bonds Accepted 13 February 2011 in pectin, and are important extracellular enzymes for both pathogenic and saprotrophic fungi. Yet, very little Available online 25 February 2011 is understood about the evolutionary forces driving the diversification of GH28s in fungal genomes. We reconstructed the evolutionary history of family GH28 in fungi by examining the distribution of GH28 copy Keywords: number across the phylogeny of fungi, and by reconstructing the phylogeny of GH28 genes. We also examined Gene family the relationship between lineage-specific GH28 expansions and fungal ecological strategy, testing the Genome analysis Birth-and-death model hypothesis that GH28 evolution in fungi is driven by ecological strategy (pathogenic vs. non-pathogenic) and Pectinase pathogenic niche (necrotrophic vs. biotrophic). Our results showed that GH28 phylogeny of Ascomycota and Fungi Basidiomycota sequences was structured by specific biochemical function, with endo-polygalacturonases and Plant pathogens endo-rhamnogalacturonases forming distinct, apparently ancient clades, while exo-polygalacturonases are more widely distributed. In contrast, Mucoromycotina and Stramenopile sequences formed taxonomically- distinct clades. Large, lineage-specific variation in GH28 copy number indicates that the evolution of this gene family is consistent with the birth-and-death model of gene family evolution, where diversity of GH28 loci within genomes was generated through multiple rounds of gene duplication followed by functional diversification and loss of some gene family members. Although GH28 copy number was correlated with genome size, our findings suggest that ecological strategy also plays an important role in determining the GH28 repertoire of fungi. Both necrotrophic and biotrophic fungi have larger genomes than non-pathogens, yet only necrotrophs possess more GH28 enzymes than non-pathogens. Hence, lineage-specific GH28 expansion is the result of both variation in genome size across fungal species and diversifying selection within the necrotrophic plant pathogen ecological niche. GH28 evolution among necrotrophs has likely been driven by a co-evolutionary arms race with plants, whereas the need to avoid plant immune responses has resulted in purifying selection within biotrophic fungi. © 2011 Elsevier B.V. All rights reserved.

1. Introduction incorporate deleterious mutations are often purged from the genome via purifying selection, while copies that acquire novel or enhanced Gene families are sets of genes that have arisen through the functionality can become fixed in the population through positive or accumulation of duplicate copies of a single ancestral gene. Following diversifying selection (Hughes, 2002; Lynch and Conery, 2003). This duplication, the paralogous genes represent a genetic redundancy pattern of duplication and expansion of advantageous gene copies that may result in relaxation of selection pressure on one or both of paired with the expulsion of dysfunctional gene copies has been the gene copies, allowing mutations to accumulate in one or both described as evolution via the birth-and-death process (Nei et al., copies (Ohno, 1970; Prince and Pickett, 2002). Gene copies that 1997; Nei and Rooney, 2005). Glycosyl hydrolase family 28 (GH28) has interesting functional diversity and is variable in copy number among related organisms,

Abbreviations: GH, glycosyl hydrolase family; GalA, galacturonic acid; PG, making this gene family a likely candidate for birth-and-death polygalacturonase; RG, rhamnogalacturonase; XG, xylogalacturonase; PGIPs, evolution. GH28 enzymes are involved in degradation of pectin, a polygalacturonase-inhibiting proteins; MRCA, most recent common ancestor; SSU rRNA, major structural constituent of the plant cell wall. Pectin is a long small subunit ribosomal RNA; ML, maximum likelihood; JTT, Jones–Taylor–Thorton; LG, Le polysaccharide chain that is composed of α-linked galacturonic acid and Gascuel; aLRT, approximate likelihood ratio test; Sh-like, Shimodaira and Hasegawa; (GalA) monomers, with some regions of GalA alternating with ANOVA, analysis of variance; Mb, mega-bases. ⁎ Corresponding author. Tel.: +1 330 672 3895; fax: +1 330 672 3713. rhamnose or branched xylose side-chains (Willats et al., 2006). E-mail address: [email protected] (C.B. Blackwood). Various enzymes in family GH28 degrade these bonds by catalyzing a

0378-1119/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.gene.2011.02.009 30 D.D. Sprockett et al. / Gene 479 (2011) 29–36 series of functionally distinct reactions. Endo-glycosidases catalyze 2. Materials and methods hydrolysis of internal glycosidic bonds at random locations within the polysaccharide, while exo-glycosidases catalyze hydrolysis of termi- 2.1. Fungal SSU rRNA-based species tree and ecological characteristics nal bonds that attach individual sugars to the ends of the polysaccharides. GH28 enzymes are also categorized into polygalac- We first reconstructed a species tree using small subunit ribosomal turonases (PG), which hydrolyze GalA-GalA linkages (E.C.'s 3.2.1.15 RNA (SSU rRNA) sequences from 69 fully sequenced fungal genomes [endo-PG] and 3.2.1.67 [exo-PG]), rhamnogalacturonases (RG), which (Supplementary File 1), including 57 Ascomycetes, 8 Basidiomycetes, 3 hydrolyze GalA-rhamnose bonds (E.C. 3.2.1.-), and xylogalacturo- Mucoromycotinas, and 1 Chytridiomycota, as well as an additional 2 nases (XG), which hydrolyze GalA-xylose bonds (E.C. 3.2.1.-) Oomycetes (Stramenopiles) and 1 plant. Relevant ribosomal sequences (Markovič and Janeček, 2001). were extracted from the SILVA database (Pruesse et al., 2007)and Genes encoding GH28 enzymes have been identified in genomes GenBank (Benson et al., 2005) and aligned using the MUSCLE sequence of a wide range of plants, bacteria, and fungi. Fungal GH28 pectinases alignment tool using default parameters available through the Geneious are degradative enzymes used by both saprotrophs growing on Pro 4.7.1 software package (Drummond et al., 2009). A maximum senesced leaf tissue (Kjøller and Struwe, 2002) and plant pathogens likelihood-based species tree was reconstructed using the program seeking to gain access to plant intracellular nutrients (Herron et al., PhyML (Guindon and Gascuel, 2003), utilizing the Tamura–Nei 2000; Reignault et al., 2008). The contribution of pectinases to substitution model (TN93). The reliability of internal branches was pathogenic infectivity remains unclear. Targeted gene disruption evaluated using 1000 bootstrap replications (Felsenstein, 1985a). studies have demonstrated that various GH28 functional types do Genome size and ecological strategy of each fungal species were play a role in virulence (Garcia-Maceira et al., 2001; Kars et al., 2005; identified from published studies, as summarized in Supplementary Shieh et al., 1997; ten Have et al., 1998), while several other studies File 1. Fungal species were categorized as plant pathogen or have not found GH28 expression to be critically important during non-pathogen, and plant pathogens were further categorized into cell wall invasion (Scott-Craig et al., 1990; Gao et al., 1996). This necrotrophic and biotrophic subgroups. disagreement may be due to lack of consideration of the differing strategies adopted by phytopathogenic fungi (Glazebrook, 2005). Necrotrophic pathogens acquire carbon and energy by extensively 2.2. GH28 genomic distribution and ancestral copy number degrading plant tissue, often resulting in the host's death. In contrast, biotrophic fungi obtain nutrients from living plant tissues Translated GH28 amino acid sequences were collected from 69 and depend on continued survival of their host (Oliver and Ipcho, fungal and 2 stramenopile genome sequences in Supplementary File 2004). Therefore, while necrotrophs indiscriminately destroy cell 1. BLASTP (Altschul et al., 1990) was used to identify putative GH28 wall constituents, biotrophs colonize intact plant cells while sequences based on their sequence similarity to six diverse, yet well avoiding immune responses. Extracellular enzymes like GH28s can described GH28 members from the filamentous fungus Aspergillus elicit a wide array of plant immunological responses, such as the niger [GenBank: XM_001389525.1, XM_001395147.1, DQ374426.1, production of extracellular plant proteins that inhibit fungal PGs, DQ374422.1, DQ374431.1, DQ374425.1] (Bussink et al., 1992; de Vries known as polygalacturonase-inhibiting proteins or PGIPs and Visser, 2001; Martens-Uzunova et al., 2006; Martens-Uzunova (De Lorenzo et al., 2001; Federici et al., 2001). GH28 gene products and Schaap, 2009). These six A. niger sequences were chosen as GH28 also stimulate the PGIP-mediated hypersensitive immune response, queries because they are representative of previously described or localized cell death, in host plants. Both PGs and PGIPs are highly classifications of A. niger pectinases (endo-PG, exo-PG, endo-RG, polymorphic, resulting in a high level of PGIP specificity for exo-RG, endo-XG, exo-XG) (Martens-Uzunova and Schaap, 2009). particular PG isozymes (Casasoli et al., 2009; Cook et al., 1999; Additionally, A. niger has been used in numerous structural and De Lorenzo et al., 2001; Di Matteo et al., 2006; Raiola et al., 2008). biochemical glycosyl hydrolase studies (Bussink et al., 1992; de Vries This molecular co-evolutionary arms race could be a source of strong et al., 2002; de Vries, 2003; Kusters-van Someren et al., 1991; van diversifying selection on the GH28 repertoires of fungal necrotrophs. Pouderoyen et al., 2003), and its genome has been shown to possess Conversely, plant immune system activity likely acts to reduce GH28 up to 20 putative GH28 members (Martens-Uzunova and Schaap, diversity in biotrophic fungi via purifying selection (Oliver and 2009). Hits with an expectation E value below the cut-off 0.001 were Ipcho, 2004). Furthermore, saprotrophic fungi that degrade dead retained for further analyses. Biochemical activity of gene products plant material would not experience the strong diversifying was inferred from existing functional annotations in GenBank, where selection associated with a co-evolutionary arms race, and should available, as commonly performed in evolutionary analyses of gene therefore also possess a low level of GH28 diversity. families (e.g., Karlsson and Stenlid, 2008; Parrent et al., 2009). Because of the importance of pectinase in liberating carbon Putative GH28 amino acid sequences were aligned using MUSCLE and energy and the wide distribution of PGIPs, we hypothesize that (Edgar, 2004). MUSCLE was chosen because it is computationally the distribution of GH28 genes in fungi may be closely linked to expedient yet attains alignment accuracy equal to or better than a the evolution of ecological strategy and pathogenic niche. In this standard CLUSTAL alignment (Edgar, 2004). This alignment was then study, we examine the pattern of GH28 gene family evolution by manually curated for accuracy. Thirty three sequences (representing investigating its occurrence and distribution in fungal genomes 21 genomes) were considered probable pseudogenes because they and by comprehensively reconstructing the long-term evolutionary were shorter than 250 residues in length, missing the GH28 active-site history of the GH28 family in fungi. Using GH28 genes from motif predicted in PROSITE (Hulo et al., 2008), or missing the GH28 completely sequenced fungal genomes, we infer the ancestral gene domain as defined in the Conserved Domain Architecture Retrieval copy number in most recent common ancestors (MRCA) of major Tool (CDART; Geer et al., 2002); these sequences were excluded from taxonomic groups. Additionally, using a reconstructed phylogeny of further analyses. The final alignment encompassed 293 GH28 fungal GH28 sequences, we infer the ancestral enzymatic mode(s) of members from 40 species. The results of this selection process were action and subsequent functional diversification within this gene generally consistent with other studies examining the pectin- family. We then test the hypothesis that the occurrence of GH28s, or degrading networks of A. niger (Martens-Uzunova and Schaap, particular functional categories of GH28, is correlated with ecological 2009), the white rot basidiomycete Phanerochaete chrysosporium strategy (pathogenic vs. non-pathogenic) and pathogenic niche (Wymelenberg et al., 2005), and the zygomycete Rhizopus oryzae (necrotrophic vs. biotrophic) against the null hypothesis that GH28 (Mertens et al., 2008), implying that all relevant GH28 homologs were copy number is simply the result of variation in overall genome size. identified. D.D. Sprockett et al. / Gene 479 (2011) 29–36 31

Using the species tree described above, a parsimony based hypothesis that the two traits evolved independently can be rejected if ancestral character state reconstruction was performed using the the correlated model fits the data significantly better than the Mesquite software package (Maddison and Maddison, 2009) to infer independent model (Pagel, 1994). Distribution of GH28 presence/ the most likely number of GH28 gene copies in the fungal MRCA. absence was examined separately for correlation with several categories Phytophthora spp. GH28s were excluded from this analysis because of pathogenicity (plant pathogen, biotrophic or necrotrophic plant Phytophthora GH28s were likely horizontally transferred from fungi pathogen, and animal pathogen). (Andersson, 2006, and see analysis in Section 3.3.) and its inclusion would bias the estimation of GH28 gene copy number in the MRCA of 3. Results and discussion fungi. 3.1. GH28 distribution and ancestral copy number 2.3. GH28 phylogenetic reconstruction The survey of GH28 members revealed the presence of at least one The alignment of 293 GH28 members from 40 species described GH28 homolog in 40 of the 69 completed fungal genome sequences above was used to reconstruct the evolutionary history of this gene examined, with GH28 copy number per genome ranging from 0 to 20 family. Additionally, seven peach (Prunus persica) GH28s, retrieved (Supplementary File 1). Such large variation is suggestive of a from the CAZy database, were used as an outgroup for this dataset. dynamic evolutionary history, and is consistent with evolution via P. persica was selected as an outgroup because it had been shown to the birth-and-death process, resulting in lineage-specific gene family harbour a diverse set of GH28s, possessing both exo- and endo-acting expansions and contractions (Nei et al., 1997; Nei and Rooney, 2005). PGs (Hadfield and Bennett, 1998; Pressey and Avants, 1973). We used a maximum parsimony approach with our fungal species Maximum likelihood (ML) phylogenetic trees were reconstructed tree to reconstruct ancestral gene copy number at key points in the with the program PhyML (Guindon and Gascuel, 2003)available evolutionary history of fungi. Our species tree was essentially identical through Geneious Pro 4.7.1 (Drummond et al., 2009) using the Jones– to recently published fungal phylogenies (James et al., 2006; Hibbett et Taylor–Thorton (JTT)+gamma (Fig. 2 and Supplementary File 2) and al., 2007). Our analysis indicates that the MRCA of fungi possessed 2 to 4 the Le and Gascuel (LG)+gamma (Supplementary File 3) amino acid GH28 members (Fig. 1). Following the diversification of fungi from substitution models, taking into account rate heterogeneity among sites plants and animals ~1.5 billion years ago (Wang et al., 1999), there was (Jones et al., 1992; Le and Gascuel, 2008). The support for internal differential expansion and contraction of this gene family among fungal branches was calculated using an approximate likelihood ratio test lineages. Overall, examining our estimates of MRCA GH28 abundance (aLRT) based on Shimodaira and Hasegawa (Sh-like) test reveals at least 14 independent instances of gene birth and at least 28 as implemented in PhyML 3.0 (Guindon and Gascuel, 2003). Both independent instances of gene death. However, these estimates are trees revealed similar clustering patterns, with minor differences likely underestimates, as indicated by our finding 33 pseudogenes, likely among poorly supported nodes. Thus, we focus our discussion on the non-functional but still recognizable as being originally a GH28 gene. For JTT+gamma-based tree (Fig. 2 and Supplementary File 2). Tree example, our results suggest that the MRCA of Saccharomycotina likely topologies were visualized using Geneious Pro 4.7.1 (Drummond possessed 2 to 4 GH28 copies (Fig. 1). However, following their split et al., 2009) and FigTree 1.2.1 (Rambaut, 2007). ~773 million years ago from other Ascomycota lineages (Blair, 2009), Saccharomycotina experienced a large-scale gene family “death” due to 2.4. Statistical analysis of GH28 evolution, genome size, and ecological a reduction in GH28 copy number, a trend also seen in other strategy Saccharomycotina gene families (Cliften et al., 2006). Conversely, organisms in the order Eurotiales, class Eurotiomycetes To test the hypothesis that GH28 copy number primarily depends (Geiser et al., 2006) have experienced a large increase in both the on genome size, we calculated Pearson's correlation coefficient number of GH28 gene family members and their functional diversity between GH28 copy number and genome size in basepairs. We also since their divergence from Sordariomycetes ~673 million years ago tested this hypothesis using Felsenstein's independent contrasts, a (Fig. 1)(Blair, 2009). The diverse genus of filamentous fungi Aspergillus method that accounts for evolutionary relatedness (Felsenstein, has undergone an especially large GH28 expansion. A. niger possesses 20 1985b). The latter test was performed in the PDAP package in putative GH28 homologs, making this Eurotiales clade an ideal Mesquite (Midford et al., 2008) with continuous character values phylogenetic location to search for novel GH28 enzymes that might be log10 transformed. useful in industrial applications. An analysis of variance (ANOVA) was used to test the hypothesis that GH28 copy number and genome size differed among necrotrophs, 3.2. Evolution of GH28 functional diversity biotrophs, and non-pathogens. This was followed by post-hoc pairwise comparisons using two-tailed t-tests. Pagel's correlation method (Pagel, The ML phylogenetic tree of GH28 sequences (Fig. 2 and Supple- 1994) was also used to evaluate the correlation between ecological mentary Files 2 and 3) revealed 8 well-supported, distinct clades, with P. strategy and GH28 occurrence (or occurrence of GH28 functional persica GH28s forming a well-supported monophyletic outgroup categories) while accounting for phylogeny. As implemented in (strongly supported by 99% bootstrap value, Fig. 2). However, relation- Mesquite (Maddison and Maddison, 2009), Pagel's correlation test ships among some of these clades remain poorly resolved, as indicated accounts for branch length while estimating the rate of change in two by both the low bootstrap support values and disagreement between binary, discrete characters, and tests the hypothesis that traits are the JTT+gamma (Fig. 2) and LG+gamma (Supplementary File 3) ML undergoing correlated evolution (Pagel, 1994). Using the species tree tree topologies. However, within clade relationships are both consistent topology derived from the SSU rRNA phylogenetic tree, the log- and well supported. likelihoods for two statistical models are calculated (Maddison and Notably, annotated sequences in Clade A, the most basal clade, have Maddison, 2009). The first model assumes the two binary characters exclusively endo-RG activity (Fig. 2). The fact that both Basidiomycete evolved in an independent manner, while the second model assumes and Ascomycete sequences occur in Clade A (Supplementary File 2) they are evolving as correlated traits (Pagel, 1994). Next, a null suggests that the MRCA of the Dikarya (Fig. 1) possessed at least one distribution is created by randomly assigning character states to the endo-RG that has since diversified into the present day form. terminal tips of the tree via a Monte-Carlo simulation. Finally, a Clade B contains exo-PGs and a single endo-PG (Fig. 2), although log-likelihood test is used to evaluate whether the independent or based on this analysis we suggest that this endo-PG sequence could be correlated models of character evolution best fit the given data. The null a mis-annotation. The single endo-PG is a gene from Botrytis cinerea, 32 D.D. Sprockett et al. / Gene 479 (2011) 29–36

Plants - Prunus persica Stramenopiles 15 Phytophthora ramorum Chytridiomycetes 23 Phytophthora sojae Mucoromycotina 0 Batrachochytrium dendrobatidis 7 Phycomyces blakesleeanus 2 Mucor circinelloides 18 Rhizopus oryzae Basiodiomycetes 1 Ustilago maydis 0 Puccinia graminis f. sp. tritici Dikarya 0 Sporobolomyces roseus 1.5 BYA 1 Cryptococcus neoformans var. grubii 2 Phanerochaete chrysosporium 2 Schizophyllum commune 2 Coprinopsis cinerea 4 Laccaria bicolor Taphrinomycotina 0 Schizosaccharomyces japonicus 0 Schizosaccharomyces octosporus 0 Schizosaccharomyces pombe Saccharomycotina 0 Yarrowia lipolytica 773 MYA 0 Candida lusitaniae 0 Candida guilliermondii 0 Debaryomyces hansenii 0 Pichia stipitis 0 Lodderomyces elongisporus 0 Candida parapsilosis 0 Candida albicans 0 Candida tropicalis 0 Saccharomyces kluyveri 0 Kluyveromyces thermotolerans 0 Eremothecium gossypii 0 Kluyveromyces lactis 0 Zygosaccharomyces rouxii Ascomycota 0 Candida glabrata 0 Saccharomyces castellii 1 Saccharomyces kudriavzevii 3 Saccharomyces bayanus 1 Saccharomyces cerevisiae Leotiomycetes 15 Botrytis cinerea 13 Sclerotinia sclerotiorum 5 Mycosphaerella fijiensis Dothideomycetes 2 Mycosphaerella graminicola 4 Stagnospora nodorum 4 Cochliobolus heterostrophus 5 Pyrenophora tritici-repentis 1 Magnaporthe grisea Sodariomcyetes 1 Chaetomium globosum 2 Neurospora crassa GH28 Gene Copies 0 Podospora anserina 0 11 Verticillium albo-atrum 1 11 Verticillium dahliae 2-3 4 Trichoderma reesei 9 Nectria haematococca 4-5 6 Fusarium graminearum 6-9 13 Fusarium oxysporum f. sp. lycopersici 10-13 7 Fusarium verticillioides 14+ 0 Microsporum gypseum 0 Microsporum canis 0 Trichophyton equinum Eurotiomycetes 0 Paracoccidioides brasiliensis 0 673 MYA 0 Blastomyces dermatitidis 0 Histoplasma capsulatum 0 Uncinocarpus reesii 0 Coccidioides immitis Non-Phytopathogen 0 Coccidioides posadasii str. Silveira Biotroph 9 Aspergillus nidulans Necrotroph Eurotiales 19 Aspergillus flavus 17 Aspergillus oryzae 7 Aspergillus terreus 20 Aspergillus niger 3 Aspergillus clavatus 11 Aspergillus fumigatus 12 Neosartorya fischeri

Fig. 1. Maximum parsimony reconstruction of ancestral glycosyl hydrolase family 28 (GH28) copy number. The phylogenetic tree was constructed from SSU rRNA sequences using the maximum likelihood method utilizing the Tamura–Nei substitution model (TN93). Branches supported by N70% bootstrap replication frequencies are thickened. Ancestral state reconstruction was inferred using maximum parsimony. The color of branches indicates the inferred range of GH28 copy number in the most recent common ancestor. The shape located to the left of the species name indicates the ecological strategy of each extant fungal species, while GH28 copy number is indicated in white within each shape. D.D. Sprockett et al. / Gene 479 (2011) 29–36 33 while the orthologous protein in the closely related species Sclerotinia These conclusions are based on the assumption that the function of sclerotiorum is annotated as an exo-PG (Govrin and Levine, 2000). unannotated sequences is in agreement with the patterns observed Eleven other GH28 genes from B. cinerea are paired with orthologous for annotated sequences in Fig. 2. Because the functionally important genes in S. sclerotiorum in our GH28 phylogeny, and the pair in Clade B regions of a gene are the ones most likely to be preserved by natural is the only example of homolog disagreement in biochemical activity selection, and also directly determine phylogenetic tree topology, we (Supplementary File 2). The two closely related Sordariomycetes, can infer the function of a gene based on phylogenic distance (Watson Verticillium albo-atrum VaMs.102 and Verticillium dahliae VdLs.17 also et al., 2005; Lee et al., 2007). Although a direct functional analysis co-occur in 10 out of 11 instances. The pattern exhibited by these would be needed to determine the unique biochemistry of each species pairs suggests that even though this dynamically evolving putative GH28 homolog, our assumption is also bolstered by the well- gene family has experienced variable rates of gene family birth-and- supported clustering of endo-RG and exo-PG sequences into individ- death between lineages, recently diverged species still retain many ual clades. orthologous gene copies derived from a speciation event. Biochemical studies have indicated that there is a high level of mutability in enzyme active sites and that endo- and exo-acting 3.3. Distribution of GH28 clades among phyla isozymes of glycosyl hydrolases could have arisen independently multiple times throughout the evolutionary history of fungi by Within the Dikarya, phylogenetic grouping of GH28 sequences convergent evolution (Dias et al., 2004; Gherardini et al., 2007; corresponds primarily with enzyme function, and not species Proctor et al., 2005; Rouvinen et al., 1990; Todd et al., 2002). However, phylogeny (Fig. 2). However, both Stramenopiles and Mucoromyco- our results do not support this scenario. The exo-PG mode of tina form distinct, well-supported clades. Despite possessing many enzymatic activity is present in several clades, including B, C, D, and morphological similarities to fungi, Stramenopiles are a distantly E, but is not present in Clade F (Fig. 2). If we assume that the endo-PG related taxonomic group that includes water molds (Oomycetes; in Clade B is a mis-annotation as described above, then endo-PGs only Beakes and Sekimoto, 2009). Unannotated GH28s from the Strame- occur in Clade F. This pattern suggests that repeated evolutionary nopiles Phytophthora ramorum and Phytophthora sojae reliably form a transitions between endo- and exo-acting forms of fungal of GH28s sister-group with Clade F (100% bootstrap support, Fig. 2), and Clade F have not occurred. Based on the widespread species distribution includes endo-PGs from a wide range of distantly related Dikarya represented in Clade F (Supplementary File 2) and its well-supported (Supplementary File 2). This topology supports an analysis by phylogenetic proximity to the Mucoromycotina, a basal group of Gotesson et al. (2002) showing that GH28s from the Stramenopile fungi, it is likely that at least one of the 2–4 GH28s in the MRCA of Phytophthora cinnamomi are more closely related to fungal GH28s fungi (Fig. 1) was an endo-PG. In addition, Clade C possesses both exo- than with those from plants. Furthermore, Phytophthora species have RGs and exo-PGs, while not including endo-RGs. This pattern suggests acquired multiple gene families laterally from distantly related plant- that exo-RGs have evolved from the exo-PG form of GH28 through degrading fungi (Andersson, 2006; Soanes et al., 2007; Belbahri et al., active-site conversion, rather than through the acquisition of exo- 2007; Richards and Talbot, 2007), and similar processes likely resulted activity by the endo-RG isoenzyme. in the observed distribution of the pectinase gene family.

Fig. 2. Maximum likelihood-based (JTT+gamma) phylogenetic tree of glycosyl hydrolase family 28 (GH28) gene family. This phylogeny summarizes relationships among 293 GH28 gene sequences present in 40 fully sequenced fungal genomes. Eight well-supported clades are shown here, with bootstrap values N70% indicated at each node. Prunus persica was used as an outgroup. Also indicated is the total number of GH28s in each clade, as well as the number of representatives of enzymatic functional categories (i.e., modes of action) in each clade. PG=polygalacturonase; RG=rhamnogalacturonase, XG=xylogalacturonase. Further details are available in Supplementary File 2. The asterisk (*) indicates a possible mis-annotation of functional category, as explained in the text. 34 D.D. Sprockett et al. / Gene 479 (2011) 29–36

Mucoromycotina is a basal group of fungi “incertae sedis” (Hibbett non-pathogenic fungi (Pb0.0001), and no detectable difference et al., 2007), and GH28s from this group form a well-supported clade between necrotrophic and biotrophic pathogens (P=0.91). This is of unannotated GH28 homologs (Supplementary File 2). This consistent with studies showing extensive gene family expansion in phylogenetic pattern supports the conclusion of Mertens et al. pathogens as compared with non-pathogenic relatives (Haas et al., (2008) that GH28s from R. oryzae were phylogenetically distinct 2009; Karlsson and Stenlid, 2008; Powell et al., 2008; Soanes et al., from GH28s of non-Mucoromycotina fungi. The parsimony analysis of 2008). Ecological strategy also significantly influenced GH28 copy the ancestral GH28 copy number and the observed gene copy number number (Pb0.0001; Fig. 4b). However, in this case necrotrophs had in this clade indicates a large increase in R. oryzae GH28 repertoire (23 significantly higher GH28 copy numbers than both biotrophs copies) following the Mucoromycotina/Dikarya split. In fact, recent (P=0.0007) and non-pathogens (Pb0.0001), while biotrophs were genomic analyses suggest that R. oryzae underwent a whole genome statistically indistinguishable from non-pathogens (P=0.75). These duplication, along with more recent duplications of virulence factor results suggest that the increase in GH28 copy number may have been genes such as secreted aspartic protease and subtilase protein (Ma positively selected for in necrotrophs, but not in non-plant pathogens et al., 2009). However, it is currently unclear whether there was a or biotrophic pathogens. GH28 expansion in the lineage leading to R. oryzae, or if there was an Because genome size only explains a small portion of GH28 earlier expansion in Mucoromycotina, followed by gene loss in Mucor variation, we tested the hypothesis that the occurrence of GH28 is circinelloides, which possess only 2 GH28 copies. Greater taxon correlated with ecological strategy (pathogenic vs. non-pathogenic) sampling in this basal fungal group will help to clarify this issue. and pathogenic niche (necrotrophic vs. biotrophic). Pagel's test for correlated binary characters showed that GH28 presence is not 3.4. GH28 evolution: Genome size and ecological strategy correlated with either a plant (P=0.9) or animal (P=0.9) pathogenic niche. However, possession of at least one GH28 copy was highly Expansion of gene families involved in host cell wall degradation is correlated with a necrotrophic niche (Pb0.0001), whereas it was not one way that fungi adapt to a pathogenic niche (Haas et al., 2009; correlated with biotrophic niche (P=0.74). This supports the findings Machida et al., 2005; Powell et al., 2008; Soanes et al., 2008). However, of other studies that show GH28s are an important virulence factor in genomic-scale evolutionary forces can also influence gene family necrotrophs such as B. cinerea (Kars et al., 2005) and Aspergillus flavus expansion and contraction, in which case we should observe that (Shieh et al., 1997), and suggests that GH28s are important enzymes organisms with larger genomes possess correspondingly higher during necrotrophic pathogenicity. numbers of GH28 copies. Indeed, both linear regression (Pb0.001, The presence of GH28s in Clade F (endo-PG) and Clade A (endo-RG) r2 =0.15) and Felsenstein's independent contrasts (P b0.0001, was also significantly correlated with a necrotrophic ecological strategy r2 =0.19) showed a significant correlation between genome size and (Pb0.001), suggesting that, in addition to the presence of GH28s, the GH28 copy number. However, the relatively low amount of variance in diversity of GH28s present in a genome is important for a necrotrophic GH28 copy number explained by genome size suggests that other niche. In fact, we found that GH28 copy number is significantly factors, such as ecological strategy, may contribute significantly. For positively correlated with GH28 functional diversity (r2 =0.61, example, Fig. 3 shows that biotroph genomes sampled in this study Pb0.001), indicating that the GH28 repertoire of necrotrophic patho- range in size from 11.74 to 88.64 Mb, yet possess relatively few GH28s. gens may be experiencing diversifying selection. Conversely, necrotrophs have a comparatively narrow range of genome sizes between 32.83 and 61.36 Mb, yet possess a far wider range of 4. Conclusions GH28 copy numbers. Differences among ecological strategies were further investigated In this study, we comprehensively surveyed GH28 gene family using a one-way ANOVA. Ecological strategy had a significant effect on distribution and diversity within the fungal kingdom, and characterized genome size (Pb0.0001, Fig. 4a), with phytopathogens (biotrophic its evolution as being consistent with the birth-and-death model. The and necrotrophic fungi) having significantly larger genomes than initial appearance of GH28s predates the evolution of fungi, which

Fig. 3. Scatterplot showing the relationship between genome size (Mb) and GH28 copy number. Rectangles show increased ranges of values for genome size of biotrophs (red) and GH28 copy number of necrotrophs (green). D.D. Sprockett et al. / Gene 479 (2011) 29–36 35

diversifying selection in all pathogens, not just necrotrophs, because sucrose can be captured from host phloem and represents a major energy and carbon resource during both biotrophic and necrotrophic growth. Hence, it is necessary to integrate an evolutionary perspective with organismal ecology and enzyme biochemistry to obtain insights into the evolution of gene families and genomes. Supplementary materials related to this article can be found online at doi:10.1016/j.gene.2011.02.009.

Acknowledgments

The authors thank Bess Heidenreich for assistance in compilation of data. Financial support for the work was provided by Kent State University.

References

Altschul, S.F., Gish, W., Miller, W., Myers, E.W., Lipman, D.J., 1990. Basic local alignment search tool. J. Mol. Biol. 215, 403–410. Andersson, J.O., 2006. Convergent evolution: gene sharing by eukaryotic plant pathogens. Curr. Biol. 16, 804–806. Beakes, G.W., Sekimoto, S., 2009. The evolutionary phylogeny of oomycetes—insights gained from studies of holocarpic parasites of algae and invertegrates. In: Lamour, K., Kamoun, S. (Eds.), Oomycete Genetics and Genomics: Diversity, Interactions, and Research Tools. John Wiley & Sons, New York, pp. 1–24. Belbahri, L., Calmin, G., Mauch, F., Andersson, J.O., 2007. Evolution of the cutinase gene family: evidence for lateral gene transfer of a candidate Phytophthora virulence factor. Gene 408, 1–8. Benson, D.A., Karsch-Mizrachi, I., Lipman, D.J., Ostell, J., Wheeler, D.L., 2005. GenBank. Nucleic Acids Res. 33, D34–D38 (Database Issue). Blair, J.E., 2009. Fungi. In: Hedges, S.B., Kumar, S. (Eds.), The Timetree of Life. Oxford University Press, Oxford, pp. 215–219. Bussink, H., Buxton, F., Fraaye, B., Graaff, L., Visser, J., 1992. The polygalacturonases of Aspergillus niger are encoded by a family of diverged genes. Eur. J. Biochem. 208, 83–90. Casasoli, M., Federici, L., Spinelli, F., Di Matteo, A., Vella, N., Scaloni, F., Fernandez-Recio, J., Cervone, F., De Lorenzo, G., 2009. Integration of evolutionary and desolvation energy analysis identifies functional sites in a plant immunity protein. Proc. Nat. Acad. Sci. 106, 7666–7671. Cliften, P.F., Fulton, R.S., Wilson, R.K., Johnston, M., 2006. After the duplication: gene loss Fig. 4. Genome size (A) and GH28 copy number (B) in fungi with differing ecological and adaptation in Saccharomyces genomes. Genetics 172, 863–872. fi strategies. Bars with different letters indicate signi cant differences according to t-test Cook, B.J., Clay, R.P., Bergmann, C.W., Albersheim, P., Darvill, A.G., 1999. Fungal b (P 0.05). Non-pathogen n=48, biotrophic n=7, necrotrophic n=17. polygalacturonases exhibit different substrate degradation patterns and differ in their susceptibilities to polygalacturonase-inhibiting proteins. Mol. Plant Microbe Interact. 12, 703–711. places the origin of this gene family at more than 1.5 billion years ago. De Lorenzo, G., D'Ovidio, R., Cervone, F., 2001. The role of polygalacturonase-inhibiting Ancient gene families such as GH28 continue to radiate and expand into proteins (PGIPs) in defense against pathogenic fungi. Annu. Rev. Phytopathol. 39, new molecular niches. Our analysis indicates that the earliest fungi 313–335. already possessed much of the functional diversity seen today in GH28, de Vries, R.P., Visser, J., 2001. Aspergillus enzymes involved in degradation of plant cell wall polysaccharides. Microbiol. Mol. Biol. Rev. 65, 497–522. including exo-PG, endo-PG, and endo-RG acting enzymes. More recent de Vries, R.P., Jansen, J., Aguilar, G., Pařenicová, L., Joosten, V., Wülfert, F., Benen, J.A.E., radiations have expanded or reduced GH28s in a lineage-specific Visser, J., 2002. Expression profiling of pectinolytic genes from Aspergillus niger. – manner—a finding that will inform future searches for industrially FEBS Lett. 530, 41 47. de Vries, R.P., 2003. Regulation of Aspergillus genes encoding plant cell wall useful pectin-degrading enzymes. Fungi in Eurotiales are likely to polysaccharide-degrading enzymes; relevance for industrial production. Appl. contain an expanded GH28 network, so investigation of organisms in Microbiol. Biotechnol. 61, 10–20. that clade may be an efficient strategy for discovery of novel GH28s. Dias, F., Vincent, F., Pell, G., Prates, J.A.M., Centeno, M.S.J., Tailford, L.E., Ferreira, L., Fontes, C.M.G.A., Davies, G.J., Gilbert, H.J., 2004. Insights into the molecular determinants of Furthermore, since plant immune systems and PGIP's have likely substrate specificity in glycoside hydrolase family 5 revealed by the crystal structure contributed to the diversifying selection pressure on necrotrophic and kinetics of Cellvibrio mixtus mannosidase 5A. J. Biol. Chem. 279, 25517–25526. fungal GH28s, leading to expansion of GH28 repertoires, screening the DiMatteo,A.,Bonivento,D.,Tsernoglou,D.,Federici,L.,Cervone,F.,2006. Polygalacturonase-inhibiting protein (PGIP) in plant defence: a structural view. genomes of necrotrophs may also yield novel pectin-degrading Phytochemistry 67, 528–533. enzymes. Drummond, A.J., Ashton, B., Cheung, M., Heled, J., Kearse, M., Moir, R., Stones-Havas, S., Even though GH28 variability is partially explained by fluctuations Thierer, T., Wilson, A., 2009. Geneious Pro 4.7.1. [http://www.geneious.com]. Edgar, R.C., 2004. MUSCLE: multiple sequence alignment with high accuracy and high in overall genome size, ecological strategy is a key factor when throughput. Nucleic Acids Res. 32, 1792–1797. accounting for the GH28 repertoire of contemporary fungi. GH28 copy Federici, L., Caprari, C., Mattei, B., Savino, C., Di Matteo, A., De Lorenzo, G., Cervone, F., number has expanded in necrotrophic fungi compared to biotrophs Tsernoglou, D., 2001. Structural requirements of endopolygalacturonase for the and non-pathogens, and the failure to recognize these fundamental interaction with PGIP (polygalacturonase-inhibiting protein). Proc. Nat. Acad. Sci. 98, 13425–13430. differences in ecological strategy obscures evolutionary insights into Felsenstein, J., 1985a. Confidence limits on phylogenies: an approach using the factors driving the molecular evolution of this gene family and fungi in bootstrap. Evolution 39, 783–791. – general. However, the importance of any particular aspect of the Felsenstein, J., 1985b. Phylogenies and the comparative method. Am. Nat. 125, 1 15. fi Gao, S., Choi, G.H., Shain, L., Nuss, D.L., 1996. Cloning and targeted disruption of enpg-1, ecological niche depends on the speci c functions encoded by a gene encoding the major in vitro extracellular endopolygalacturonase of the chestnut family. A recent analysis found evidence of expansion of glycosyl blight fungus, Cryphonectria parasitica. Appl. Environ. Microbiol. 62, 1984–1990. hydrolase family 4 in both pathogenic and endophytic fungal lineages Garcia-Maceira, F., Di Pietro, A., Huertas-Gonzalez, M.M.D., Ruiz-Roldan, M.C., Roncero, M.I.G., 2001. Molecular characterization of an endopolygalacturonase from (Parrent et al., 2009). GH4 encodes invertase enzymes that hydrolyze Fusarium oxysporum expressed during early stages of infection. Appl. Environ. sucrose. In contrast with pectinase, invertase is likely to be subject to Microbiol. 67, 2191–2196. 36 D.D. Sprockett et al. / Gene 479 (2011) 29–36

Geer, L.Y., Domrachev, M., Lipman, D.J., Bryant, S.H., 2002. CDART: protein homology by Midford, P.E., Garland, T., Maddison, W.P., 2008. PDAP:PDTREE: A translation of the domain architecture. Genome Res. 12, 1619–1623. PDTREE application of Garland et al'.s Phenotypic Diversity Analysis Program 1.14. Geiser, D.M., Gueidan, C., Miadlikowska, J., Lutzoni, F., Kauff, F., Hofstetter, V., Fraker, E., [http://mesquiteproject.org]. Schoch, C.L., Tibell, L., Untereiner, W.A., et al., 2006. Eurotiomycetes: Eurotiomy- Nei, M., Gu, X., Sitnikova, T., 1997. Evolution by the birth-and-death process in multigene cetidae and Chaetothyriomycetidae. Mycologia 98, 1053–1064. families of the vertebrate immune system. Proc. Nat. Acad. Sci. 94, 7799–7806. Gherardini, P.F., Wass, M.N., Helmer-Citterich, M., Sternberg, M.J.E., 2007. Convergent Nei, M., Rooney, A.P., 2005. Concerted and birth-and-death evolution of multigene evolution of enzyme active sites is not a rare phenomenon. J. Mol. Biol. 372, 817–845. families. Annu. Rev. Genet. 39, 121–152. Glazebrook, J., 2005. Contrasting mechanisms of defense against biotrophic and Ohno, S., 1970. Evolution by Gene Duplication. Springer-Verlag, New York. necrotrophic pathogens. Annu. Rev. Phytopathol. 43, 205–227. Oliver, R.P., Ipcho, S.V.S., 2004. Arabidopsis pathology breathes new life into the Gotesson, A., Marshall, J.S., Jones, D.A., Hardham, A.R., 2002. Characterization and necrotrophs-vs.-biotrophs classification of fungal pathogens. Mol. Plant Pathol. 5, evolutionary analysis of a large polygalacturonase gene family in the oomycete 347–352. plant pathogen Phytophthora cinnamomi. Mol. Plant Microbe Interact. 15, 907–921. Pagel, M., 1994. Detecting correlated evolution on phylogenies: a general method for Govrin, E.M., Levine, A., 2000. The hypersensitive response facilitates plant infection by the comparative analysis of discrete characters. Proc. R. Soc. B Biol. Sci. 255, 37–45. the necrotrophic pathogen Botrytis cinerea. Curr. Biol. 10, 751–757. Parrent, J.L., James, T.Y., Vasaitis, R., Taylor, A.F.S., 2009. Friend or foe? Evolutionary Guindon, S., Gascuel, O., 2003. A simple, fast, and accurate algorithm to estimate large history of glycoside hydrolase family 32 genes encoding for sucrolytic activity in phylogenies by maximum likelihood. Sys. Biol. 52, 696–704. fungi and its implications for plant-fungal symbioses. BMC Evol. Biol. 9, 148–164. Haas, B.J., Kamoun, S., Zody, M.C., Jiang, R.H.Y., Handsaker, R.E., Cano, L.M., Grabherr, M., Powell, A.J., Conant, G.C., Brown, D.E., Carbone, I., Dean, R.A., 2008. Altered patterns of gene Kodira, C.D., Raffaele, S., Torto-Alalibo, T., et al., 2009. Genome sequence and analysis of duplication and differential gene gain and loss in fungal pathogens. BMC Genomics 9, the Irish potato famine pathogen Phytophthora infestans. Nature 461, 393–398. 147–162. Hadfield, K.A., Bennett, A.B., 1998. Polygalacturonases: many genes in search of a Pressey, R., Avants, J.K., 1973. Separation and characterization of endopolygalacturonase function. Plant Physiol. 117, 337–343. and exopolygalacturonase from peaches. Plant Physiol. 52, 252–256. Herron, S.R., Benen, J.A.E., Scavetta, R.D., Visser, J., Jurnak, F., 2000. Structure and function of Prince, V.E., Pickett, F.B., 2002. Splitting pairs: the diverging fates of duplicated genes. pectic enzymes: virulence factors of plant pathogens. Proc. Nat. Acad. Sci. 97, Nat. Rev. Genet. 3, 827–837. 8762–8769. Proctor, M.R., Taylor, E.J., Nurizzo, D., Turkenburg, J.P., Lloyd, R.M., Vardakou, M., Davies, Hibbett, D.S., Binder, M., Bischoff, J.F., Blackwell, M., Cannon, P.F., Eriksson, O.E., Huhndorf, G.J., Gilbert, H.J., 2005. Tailored catalysts for plant cell-wall degradation: redesigning S., James, T., Kirk, P.M., Lücking, R., et al., 2007. A higher-level phylogenetic the exo/endo preference of Cellvibrio japonicus arabinanase 43A. Proc. Nat. Acad. Sci. classification of the fungi. Mycol. Res. 111, 509–547. 102, 2697–2702. Hughes, A.L., 2002. Adaptive evolution after gene duplication. Trends Genet. 18, 433–434. Pruesse, E., Quast, C., Knittel, K., Fuchs, B.M., Ludwig, W., Peplies, J., Glockner, F.O., 2007. Hulo, N., Bairoch, A., Bulliard, V., Cerutti, L., Cuche, B., De Castro, E., Lachaize, C., Langendijk- SILVA: a comprehensive online resource for quality checked and aligned ribosomal Genevaux, P.S., Sigrist, C.J.A., 2008. The 20 years of PROSITE. Nucleic Acids Res. 36, RNA sequence data compatible with ARB. Nucleic Acids Res. 35, 7188–7196. D245–D249 (Database Issue). Raiola, A., Sella, L., Castiglioni, C., Balmas, V., Favaron, F., 2008. A single amino acid James, T.Y., Kauff, F., Schoch, C.L., Matheny, P.B., Hofstetter, V., Cox, C.J., Celio, G., substitution in highly similar endo-PGs from Fusarium verticillioides and related Gueidan, C., Fraker, E., Miadlikowska, J., et al., 2006. Reconstructing the early Fusarium species affects PGIP inhibition. Fung. Genet. Biol. 45, 776–789. evolution of fungi using a six-gene phylogeny. Nature 443, 818–822. Rambaut, A., 2007. FigTree 1.2.1. [http://tree.bio.ed.ac.uk/]. Jones, D.T., Taylor, W.R., Thornton, J.M., 1992. The rapid generation of mutation data Reignault, P., Valette-Collet, O., Boccara, M., 2008. The importance of fungal pectinolytic matrices from protein sequences. Comput. Appl. Biosci. 8, 275–282. enzymes in plant invasion, host adaptability and symptom type. Eur. J. Plant Pathol. Karlsson, M., Stenlid, J., 2008. Comparative evolutionary histories of the fungal chitinase 120, 1–11. gene family reveal non-random size expansions and contractions due to adaptive Richards, T.A., Talbot, N.J., 2007. Plant parasitic oomycetes such as Phytophthora species natural selection. Evol. Bioinform. 4, 47–60. contain genes derived from three eukaryotic lineages. Plant Signal. Behav. 2, 112–114. Kars, I., Krooshof, G., Wagemakers, L., Joosten, R., Benen, J.A.E., van Kan, J.A.L., 2005. Rouvinen, J., Bergfors, T., Teeri, T., Knowles, J.K., Jones, T.A., 1990. Three-dimensional Necrotizing activity of five Botrytis cinerea endopolygalacturonases produced in structure of cellobiohydrolase II from Trichoderma reesei. Science 249, 380–386. Pichia pastoris. Plant J. 43, 213–225. Scott-Craig, J.S., Panaccione, D.G., Cervone, F., Walton, J.D., 1990. Endopolygalacturonase is Kjøller, A.H., Struwe, S., 2002. Fungal communities, succession, enzymes, and decompo- not required for pathogenicity of Cochliobolus carbonum on maize. Plant Cell 2, sition. In: Burns, R.G., Dick, R.P. (Eds.), Enzymes in the Environment: Activity, Ecology, 1191–1200. and Applications. Marcel Dekker, New York, pp. 267–284. Shieh, M.T., Brown, R.L., Whitehead, M.P., Cary, J.W., Cotty, P.J., Cleveland, T.E., Dean, R.A., Kusters-van Someren, M.A., Harmsen, J.A.M., Kester, H.C.M., Visser, J., 1991. Structure of the 1997. Molecular genetic evidence for the involvement of a specific polygalacturonase, Aspergillus niger pelA gene and its expression in Aspergillus niger and Aspergillus P2c, in the invasion and spread of Aspergillus flavus in cotton bolls. Appl. Environ. nidulans. Curr. Genet. 20, 293–299. Microbiol. 63, 3548–3552. Le, S.Q., Gascuel, O., 2008. An improved general amino acid replacement matrix. Mol. Soanes, D.M., Richards, T.A., Talbot, N.J., 2007. Insights from sequencing fungal and Biol. Evol. 25, 1307–1320. oomycete genomes: what can we learn about plant disease and the evolution of Lee, D., Redfern, O., Orengo, C., 2007. Predicting protein function from sequence and pathogenicity? Plant Cell 19, 3318–3326. structure. Nat. Rev. Molec. Cell Biol. 8, 995–1005. Soanes, D.M., Alam, I., Cornell, M., Wong, H.M., Hedeler, C., Paton, N.W., Rattray, M., Lynch, M., Conery, J.S., 2003. The origins of genome complexity. Science 302, Hubbard, S.J., Oliver, S.G., Talbot, N.J., 2008. Comparative genome analysis of 1401–1404. filamentous fungi reveals gene family expansions associated with fungal Ma, L.-J., Ibrahim, A.S., Skory, C., Grabherr, M.G., Burger, G., Butler, M., Elias, M., Idnurm, pathogenesis. PLoS ONE 3, e2300. A., Lang, B.F., Sone, T., et al., 2009. Genomic analysis of the basal lineage fungus ten Have, A., Mulder, W., Visser, J., van Kan, J., 1998. The endopolygalacturonase gene Rhizopus oryzae reveals a whole-genome duplication. PLoS Genet. 5, e1000549. Bcpg1 is required for full virulence of Botrytis cinerea. Mol. Plant Microbe Interact. Machida, M., Asai, K., Sano, M., Tanaka, T., Kumagai, T., Terai, G., Kusumoto, K.I., Arima, 11, 1009–1016. T., Akita, O., Kashiwagi, Y., et al., 2005. Genome sequencing and analysis of Todd, A.E., Orengo, C.A., Thornton, J.M., 2002. Plasticity of enzyme active sites. Trends Aspergillus oryzae. Nature 438, 1157–1161. Biochem. Sci. 27, 419–426. Maddison, W.P., Maddison, D.R., 2009. Mesquite: a modular system for evolutionary van Pouderoyen, G., Snijder, H.J., Benen, J.A.E., Dijkstra, B.W., 2003. Structural insights into analysis 2.71. [http://mesquiteproject.org]. the processivity of endopolygalacturonase I from Aspergillus niger.FEBSLett.554, Markovič, O., Janeček, Š., 2001. Pectin degrading glycoside hydrolases of family 28: 462–466. sequence-structural features, specificities and evolution. Protein Eng. 14, 615–631. Wang, D.Y., Kumar, S., Hedges, S.B., 1999. Divergence time estimates for the early Martens-Uzunova, E.S., Zandleven, J.S., Benen, J.A.E., Awad, H., Kools, H.J., Beldman, G., history of animal phyla and the origin of plants, animals and fungi. Proc. R. Soc. B Voragen, A.G.J., Van Den Berg, J.A., Schaap, P.J., 2006. A new group of exo-acting family Biol. Sci. 266, 163–171. 28 glycoside hydrolases of Aspergillus niger that are involved in pectin degradation. Watson, J.D., Laskowski, R.A., Thornton, J.M., 2005. Predicting protein function from Biochem. J. 400, 43–52. sequence and structural data. Curr. Opin. Struct. Biol. 15, 275–284. Martens-Uzunova, E.S., Schaap, P.J., 2009. Assessment of the pectin degrading enzyme Willats, W.G.T., Knox, J.P., Mikkelsen, J.D., 2006. Pectin: new insights into an old network of Aspergillus niger by functional genomics. Fung. Genet. Biol. 46, polymer are starting to gel. Trends Food Sci. Technol. 17, 97–104. S170–S179. Wymelenberg, A.V., Sabat, G., Martinez, D., Rajangam, A.S., Teeri, T.T., Gaskell, J., Kersten, Mertens, J.A., Burdick, R.C., Rooney, A.P., 2008. Identification, biochemical characterization, P.J., Cullen, D., 2005. The Phanerochaete chrysosporium secretome: database predic- and evolution of the Rhizopus oryzae 99-880 polygalacturonase gene family. Fung. tions and initial mass spectrometry peptide identifications in cellulose-grown Genet. Biol. 45, 1616–1624. medium. J. Biotechnol. 118, 17–34.