Comparative genomics to explore phylogenetic relationship, cryptic sexual potential and host specificity of Rhynchosporium species on grasses

Abstract

Background: The Rhynchosporium species complex consists of hemibiotrophic fungal pathogens specialized to different sweet grass species including the cereal crops barley and rye. A sexual stage has not been described, but several lines of evidence suggest the occurrence of sexual reproduction. Therefore, a comparative genomics approach was carried out to disclose the evolutionary relationship of the species and to identify genes demonstrating the potential for a sexual cycle. Furthermore, due to the evolutionary very young age of the five species currently known, this genus appears to be well-suited to address the question at the molecular level of how pathogenic fungi adapt to their hosts.

Results: The genomes of the different Rhynchosporium species were sequenced, assembled and annotated using ab initio gene predictors trained on several fungal genomes as well as on Rhynchosporium expressed sequence tags. Structures of the rDNA regions and genome-wide single nucleotide polymorphisms provided a hypothesis for intra-genus evolution. Homology screening detected core meiotic genes along with most genes crucial for sexual recombination in ascomycete fungi. In addition, a large number of cell wall-degrading enzymes that is characteristic for hemibiotrophic and necrotrophic fungi infecting monocotyledonous hosts were found. Furthermore, the Rhynchosporium genomes carry a repertoire of genes coding for polyketide synthases and non-ribosomal peptide synthetases. Several of these genes are missing from the genome of the closest sequenced relative, the poplar pathogen Marssonina brunnea, and are possibly involved in adaptation to the grass hosts. Most importantly, six species-specific genes coding for protein effectors were identified in R. commune. Their deletion yielded mutants that grew more vigorously in planta than the wild type.

Conclusion: Both cryptic sexuality and secondary metabolites may have contributed to host adaptation. Most importantly, however, the growth-retarding activity of the species-specific effectors suggests that host adaptation of R. commune aims at extending the biotrophic stage at the expense of the necrotrophic stage of pathogenesis. Like other apoplastic fungi Rhynchosporium colonizes the intercellular matrix of host leaves relatively slowly without causing symptoms, reminiscent of the development of endophytic fungi. Rhynchosporium may therefore become an object for studying the mutualism-parasitism transition.

Keywords: CAZymes; Effectors; Host specificity; Leotiomycetes; Non-ribosomal peptide synthetases; Phylogenetic evolution; Polyketide synthases; Rhynchosporium; Sex-related genes; Whole genome sequencing.

Fig. 1

Rhynchosporium rDNA structures. The sequences of the rDNA regions of all Rhynchosporium species were obtained by Sanger sequencing. 18S, 5.8S and 28S genes are separated by ITS1 and ITS2, respectively (grey lines). The introns (black lines) in the 18S and 28S genes of R. lolii and R. orthosporum are identical except for 1 additional nucleotide in the R. lolii 28S intron. The 28S intron of R. agropyri is highly similar to those of the CCG species (87% identity) except for 5′- and 3′-terminal extensions of 99 and 87 bp, respectively

Fig. 2

Rhynchosporium evolutionary relationships. a Rhynchosporium subtree of the Leotiomycetes phylogeny (cf. Additional file 1: Figure S1). The nucleotide sequences of 18S rDNA, 28S rDNA, ITS region, elongation factor EF1-a and RNA polymerase II subunits RPB1 and RPB2 were concatenated. b SNP-based phylogeny of the BCG species including the three R. commune isolates UK7, AU2 and 13-13. The evolutionary history was inferred using the Minimum Evolution method (optimal tree with sum of branch length = 1.022). The tree is drawn to scale. All nucleotide positions containing gaps and missing data were eliminated, leaving a final dataset of 5,904,161 positions. Scale: number of substitutions per site and SNPs per position, respectively. Bootstrap numbers are given above branching points, divergence times in italics below branching points. BCG, beaked conidia group; CCG, cylindrical conidia group

Fig. 3

Integration of the genus Rhynchosporium into the fungal systematics. The concatenated amino acid sequences of elongation factor EF1-a and of the polymerase II subunits RPB1 and RPB2 from 21 taxa were used to construct the phylogenetic tree. Numerals on the nodes represent the percentages from 500 bootstraps. Scale: number of substitutions per nucleotide

Fig. 4

Rhynchosporium species and their hosts from the Poaceae family. The Rhynchosporium systematics is contrasted with the section of the grass systematics [37] containing all known host species. Numbers in brackets indicate the total number in this order. *Two R. commune isolates were described to be able to cross the Poodae-Triticodae border, being pathogenic to Lolium multiflorum and Hordeum vulgare [16]. BCG: beaked conidia group, CCG: cylindrical conidia group, n/a: not a host species

Fig. 5

Rhynchosporium MAT gene loci. MAT1-1, R. commune UK7 (13,701 bp); MAT1-2, R. agropyri (12,801 bp). Dotted lines mark the idiomorphic regions. The genes flanking the MAT loci code for a cytoskeletal protein (SLA2) and a DNA lyase (APN2), respectively

Fig. 6

Cell wall-degrading enzymes of R. commune. a Genes coding for secreted CAZyme. AA, auxiliary activities, CBM only, proteins with carbohydrate-binding modules but lacking known enzyme activities, CE, carbohydrate esterases, GH, glycoside hydrolases, PL, polysaccharides lyases. b Genes coding for enzymes that target the different cell wall components

Fig. 7

PKS phylogenetic tree. The amino acid sequence of the KS domains from 114 fungal and bacterial PKS were used to construct the phylogenetic tree. Numerals on the nodes represent the percentages from 500 bootstraps. Numerals <50 were omitted. Scale: number of substitutions per nucleotide. Colored backgrounds indicate enzyme groups: blue, reducing PKS clades I-IV; orange, non-reducing PKS clades NRI-III and NR bI + II [81]; grey, bacterial PKS; green, fatty acid synthases (FAS)

Fig. 8

NRPS phylogenetic tree 1. The amino acid sequences of the A domain from 153 mostly mono-modular fungal and bacterial NRPS were used to construct the phylogenetic tree. Numerals at the nodes represent the percentages from 500 bootstraps. Numeral <50 were omitted. Scale: number of substitutions per nucleotide. Colored backgrounds indicate enzyme groups: Sid, siderophore synthetases; NPS11/NPS12, NPS11/12-like NRPS and ETP toxin synthetases; Cyclo, cyclosporine synthetases; MBC, major bacterial clade; AAR, α-amino-adipate reductases; NPS10, NRPS 10-like NRPS; PKS-NPS, hybrid enzymes. Orange-framed boxes mark members of outgroups

Fig. 9

NRPS phylogenetic tree 2. The amino acid sequences of the A domain from 151 mostly oligo-modular specifically fungal NRPS were used to construct the phylogenetic tree. Numerals at the nodes represent the percentages from 500 bootstraps. Numerals <50 were omitted. Colored backgrounds indicate enzyme groups: blue, NPS8-like; pink, NPS6-like, orange, ergot alkaloid synthetases, green, peptaibol synthetase TEX1; grey, AM-toxin synthetase. Red and blue dots mark the modules of HC-toxin and peramine synthetase, respectively. Coloured boxes indicate modules of the 4 Rhynchosporium NRPS

Fig. 10

PKS4 gene cluster of R. commune. The cluster harbors genes coding for two secondary metabolism key enzymes, PKS4 and DMATS3, along with several decorating enzymes, a putative transporter and a regulatory gene. DH, dehydrogenase; DO, dioxygenase; HL, hydrolase; MT, methyl transferase; P450, cytochrome P-450 enzymes; TF, transcription factor; TP, transporter; UF, unknown function. SMURF and MDM, see text

Fig. 11

Phylogeny of NIPs. The amino acid sequences of 39 mature NIPs were aligned using the MUSCLE algorithm. For clarity reasons and due to their high similarity only one NIP2 protein from the 3 R. commune isolates was taken into consideration. The tree is drawn to scale. Numerals on the nodes represent the percentages from 500 bootstraps. Asterisks mark the CCG species. Scale: number of substitutions per nucleotide

Fig. 12

Sequence comparison of NIP2 and NIP2-like proteins (NLP). For clarity reasons the amino acid sequences of the mature proteins (-SP) from R. commune UK7 were aligned with the sequences only available in other isolates/species. Arrows indicate the domains identified by MEME. The CRS domain in position 61-63 (marked by the black box on the arrow) differentiates the two protein groups

Fig. 13

Flow chart for the identification of candidate effector genes in the R. commune genome (for details see text). RcSP7 and RcSP8 did not match the ≥ 2% Cys criterion. RcSP4 (mis-annotated), RcSP7 (very low expression) and RcSP8 (no expression) were not submitted to functional analysis

Fig. 14

Expression of RcSP genes during pathogenesis. Relative RNA abundance was measured by qRT-PCR during growth of fungal isolate UK7 on barley cv. ‚Ingrid’. Dotted line indicates the development of fungal biomass

Fig. 15

Growth acceleration of deletion mutants. Relative biomass of fungal deletion mutants and wild-type isolate UK7 was determined by qPCR at 14 dpi on barley cv. ‘Ingrid’. Results from independent mutants are combined (cf. Additional file 9: Figure S3). Bars represent the 95% confidence intervals. n-values are given at the base of the bars

Fig. 16

Growth of ΔRcSP9 mutants. Relative biomasses of three independent deletion mutants and wild-type isolate UK7 were determined by qPCR during pathogenesis on barley cv. ‘Ingrid’. Bars represent 95% confidence intervals. n_ΔRcSP9 = 9, n_UK7 = 4

Fig. 17

Disease phenotype of RcSP9 deletion mutants. Primary leaves of barley cv. ‘Ingrid’ were inoculated with spores of wild-type isolate UK7 or of the mutants (#1036, #1067, #1231) and photos were taken at indicated times post inoculation. C, mock inoculation