Saprophytic and pathogenic fungi in the Ceratocystidaceae differ in their ability to metabolize plant-derived sucrose

Abstract

Background: Proteins in the Glycoside Hydrolase family 32 (GH32) are carbohydrate-active enzymes known as invertases that hydrolyse the glycosidic bonds of complex saccharides. Fungi rely on these enzymes to gain access to and utilize plant-derived sucrose. In fungi, GH32 invertase genes are found in higher copy numbers in the genomes of pathogens when compared to closely related saprophytes, suggesting an association between invertases and ecological strategy. The aim of this study was to investigate the distribution and evolution of GH32 invertases in the Ceratocystidaceae using a comparative genomics approach. This fungal family provides an interesting model to study the evolution of these genes, because it includes economically important pathogenic species such as Ceratocystis fimbriata, C. manginecans and C. albifundus, as well as saprophytic species such as Huntiella moniliformis, H. omanensis and H. savannae.

Results: The publicly available Ceratocystidaceae genome sequences, as well as the H. savannae genome sequenced here, allowed for the identification of novel GH32-like sequences. The de novo assembly of the H. savannae draft genome consisted of 28.54 megabases that coded for 7 687 putative genes of which one represented a GH32 family member. The number of GH32 gene family members appeared to be related to the ecological adaptations of these fungi. The pathogenic Ceratocystis species all contained two GH32 family genes (a putative cell wall and a putative vacuolar invertase), while the saprophytic Huntiella species had only one of these genes (a putative cell wall invertase). Further analysis showed that the evolution of the GH32 gene family in the Ceratocystidaceae involved transposable element-based retro-transposition and translocation. As an example, the activity of a Fot5-like element likely facilitated the assembly of the genomic regions harbouring the GH32 family genes in Ceratocystis.

Conclusions: This study provides insight into the evolutionary history of the GH32 gene family in Ceratocystidaceae. Our findings suggest that transposable elements shaped the evolution of the GH32 gene family, which in turn determines the sucrolytic activities and related ecological strategies of the Ceratocystidaceae species that harbour them. The study also provides insights into the role of carbohydrate-active enzymes in plant-fungal interactions and adds to our understanding of the evolution of these enzymes and their role in the life style of these fungi.

Fig. 1

Glycoside hydrolase 32 (GH32) gene family expansions and contractions mapped onto the Sordariomycetes chronogram. Significant (P< 0.05) expansions (indicated with red lines) and contractions (indicated with green lines) were inferred using CAFE v3.1 (Computational Analysis of gene Family Evolution) [45]. The probable ancestral gene family size for each node is indicated within white circles, while the family sizes in extant species are indicated at the tips of terminal branches. Colour-coding designates the Sordariomycetes taxa to either order or family level. The chronogram was inferred in this study (see Additional file 2: Figure S1). The sequences from Dothideomycetes were used for outgroup purposes

Fig. 2

Alignment of the conserved motifs of the glycoside hydrolase 32 (GH32) enzymes. These include conserved regions (labelled A-G) and various amino acids (shown with black stars). The N-terminal β-propeller module (indicated in the blue block) and the C-terminal β-sandwich module (indicated in the red block) are also highlighted. The translated sequences of one group of the Ceratocystis GH32 gene possess a trans-membrane domain (shown with dotted lines) characteristic of vacuolar invertases [5], while the translated sequences of the other Ceratocystis GH32 gene and the Huntiella GH32 gene possess an eukaryotic secretion signal (shown with dotted lines) needed for secretion [60]

Fig. 3

3D structure of the C. manginecans invertase (CmINV-CW). Roman numerals (I–V) show the five blades of the β-propeller module, while the C-terminal β-sandwich module is indicated in dark red. These structures were inferred with the Swiss-Model Web server (http://www.expasy.org/swissmod/SWISS-MODEL.html) by making use of a fructosyltransferase from Aspergillus japonicus (PDB id: 3lfi.1) as template

Fig. 4

The predicted genes flanking the Glycoside hydrolase 32 (GH32) gene family members in Huntiella and Ceratocystis. Genes present on the scaffolds harbouring the putative invertases were predicted using AUGUSTUS [32] and annotated using Blast2GO [44]. Note that the genes are not drawn to scale. The Huntiella GH32 family gene is flanked by putative G1/S-specific cyclin Pcl5 (Colletotrichum orbiculare, ENH86823), RNAse P Rpr2/Rpp21 subunit domain-containing protein (Gaeumannomyces gramini, EPQ63823), Malate synthase-like protein (Acremonium chrysogenum, XP003651419), serine/threonine-protein kinase (Metarhizium acridum, EFY93082.1), nitrogen response regulator (Colletotrichum gloeosporioides, ELA29612.1), DEAD/DEAH box helicase (Colletotrichum sublineola, KDN64774), 2-dehydropantoate 2-reductase (Colletotrichum gloeosporioides, EQB48758), and structural maintenance of chromosomes 5 (Villosiclava virens, KDB17190) genes. The two Ceratocystis GH32 family genes were flanked by putative Phosphatidylinositol-specific phospholipase (Metarhizium anisopliae, KFG82763), putative WD domain-containing protein (Togninia minima, EOO00810.1), reverse transcriptases (Sclerotinia sclerotiorum, XP_001588999 and Blumeria graminis, CCU77161), transcription elongation factor 5 (Scedosporium apiospermum, KEZ42236), adenylate kinase (Magnaporthe oryzae, XP003716198), and Fot5 transposase (Colletotrichum gloeosporioides, ELA33194.1) genes

Fig. 5

Maximum likelihood phylogeny of the Sordariomycetes Glycoside Hydrolase 32 (GH32) gene family. Representative sequences of the 8 groups that span the fungal GH32 gene phylogeny [2] were included in this study. GenBank accession numbers or sequence identifiers from genome projects for each of these sequences are provided in parentheses. Percentage bootstrap support (based on a 1000 repeats) is indicated at the internodes. The exon-intron structure of the genes is diagrammatically indicated next to each taxon where gaps within solid lines indicate intron positions. Colour-coding designates the groups previously identified [2]. The sequences from Arabidobsis thaliana were used for outgroup purposes

Fig. 6

The inferred evolutionary history of the Ceratocystidaceae Glycoside hydrolase 32 (GH32) gene family and the orthology relationships among these genes. a The evolutionary tree shows nine homologous genes from six species (A). The Huntiella cell wall invertase genes are depicted as HsINV-CW, HmINV-CW and HoINV-CW, while the Ceratocystis vacuolar invertase genes are depicted as CaINV-V, CmINV-V and CfINV-V and those encoding the Ceratocystis cell wall invertases as CaINV-CW, CmINV-CW and CfINV-CW. As indicated by CAFE, the genome of the Ceratocystidaceae ancestor likely encoded two invertase (INV) genes, one of which (depicted by the grey line) was subsequently lost from both the Ceratocystis and Huntiella lineages (depicted by grey broken line) before the radiation of species. However, the remaining invertase gene (depicted in orange) was duplicated in the Ceratocystis ancestor resulting in the two invertase genes encoded by the genomes of the extant species. This duplication was also reconstructed using NOTUNG 2.6 which detects duplications based on gene tree to species tree reconciliation [89] (results not shown). All of the invertase genes in the extant Ceratocystis and Huntiella species thus evolved from the same ancestral gene in the Ceratocystidaceae ancestor (depicted by the orange line). The respective Ceratocystis genes each evolved through vertical decent after their emergence (i.e., gene duplication) in the last common ancestor. b Following the standard nomenclature for duplicated genes (reviewed by Koonin [36]), the Huntiella cell wall invertase genes share an orthologous relationship (i.e., orthologs are related via speciation and are derived via vertical decent from the common ancestor). The same is also true for the respective cell wall and vacuolar invertase genes of Ceratocystis, where each represent a set of orthologs. Because the duplication that gave rise to the Certocystis genes occurred before radiation of this genus, the Ceratocystis cell wall and vacuolar invertase genes represent outparalogs (i.e., homologs that derive from a gene duplication event that precedes lineage radiation/speciation) [36]. However, all of the Ceratocysistis invertase genes represent co-orthologs of the gene in Huntiella. This is because the lineage-specific duplication in Ceratocystis gave rise to a set of genes that are collectively orthologous to those of Huntiella [36]

Fig. 7

Maximum likelihood phylogeny of the Fot5 DDD catalytic domain. This analysis was done using the WAG substitution model [49] and gamma correction to account for among site rate variation. The Ceratocystis Fot5 sequences are included in the grey area and indicated according to species (green dots = C. albifundus, blue dots = C. fimbriata, red dots = C. manginecans). The branch labelled with an asterisk received 81 % bootstrap support based on the analysis of 1000 pseudoreplicates (see Additional file 4: Figure S2 for full information regarding bootstrap support for the tree, as well as the sequence identifiers of putative Ceratocystis Fot5 homologs and Additional file 3: Table S2 for their genomic coordinates). GenBank accession numbers or for previously identified Fot5 homologs are: Fot2 [Genbank:JN624854, F. oxysporum), Fot5 [Genbank:CAE55867, F. oxysporum], Fot1 [Genbank:X64799, F. oxysporum], Fot4 [Genbank:AF076632, F. oxysporum], Fot9 [JGI:2517, F. graminearum], Fotyl [Genbank:CAG33729.1 Yarrowia lipolytica], Molly [Genbank:CAD32687, Parastagonospora nodorum], Ophio [Genbank:ABG26269, Ophiostoma novo-ulmi], PABRA [Genbank:ACY56713, Paracoccidioides brasiliensis], Pixie [Genbank:CAD32689, Parastagonospora nodorum], Pot2 [Genbank:CAA83918, Magnaporthe grisea], Pot3 [Genbank:AAC49418, M. grisea], SCSCL [Genbank:XP001592252, Sclerotinia sclerotiorum], Taf1 [Genbank:AAX83011, Aspergillus fumigatus], Tan1 [Genbank:U58946, Aspergillus awamori] USMA [Genbank:UM03882, Ustilago maydis), Flipper [Genbank:AAB63315, Botryotinia fuckeliana] and Cirt1 [Genbank:XP710204, Candida albicans]