- Research article
- Open Access
The genome sequence of the biocontrol fungus Metarhizium anisopliae and comparative genomics of Metarhizium species
BMC Genomics volume 15, Article number: 660 (2014)
Metarhizium anisopliae is an important fungal biocontrol agent of insect pests of agricultural crops. Genomics can aid the successful commercialization of biopesticides by identification of key genes differentiating closely related species, selection of virulent microbial isolates which are amenable to industrial scale production and formulation and through the reduction of phenotypic variability. The genome of Metarhizium isolate ARSEF23 was recently published as a model for M. anisopliae, however phylogenetic analysis has since re-classified this isolate as M. robertsii. We present a new annotated genome sequence of M. anisopliae (isolate Ma69) and whole genome comparison to M. robertsii (ARSEF23) and M. acridum (CQMa 102).
Whole genome analysis of M. anisopliae indicates significant macrosynteny with M. robertsii but with some large genomic inversions. In comparison to M. acridum, the genome of M. anisopliae shares lower sequence homology. While alignments overall are co-linear, the genome of M. acridum is not contiguous enough to conclusively observe macrosynteny. Mating type gene analysis revealed both MAT1-1 and MAT1-2 genes present in M. anisopliae suggesting putative homothallism, despite having no known teleomorph, in contrast with the putatively heterothallic M. acridum isolate CQMa 102 (MAT1-2) and M. robertsii isolate ARSEF23 (altered MAT1-1). Repetitive DNA and RIP analysis revealed M. acridum to have twice the repetitive content of the other two species and M. anisopliae to be five times more RIP affected than M. robertsii. We also present an initial bioinformatic survey of candidate pathogenicity genes in M. anisopliae.
The annotated genome of M. anisopliae is an important resource for the identification of virulence genes specific to M. anisopliae and development of species- and strain- specific assays. New insight into the possibility of homothallism and RIP affectedness has important implications for the development of M. anisopliae as a biopesticide as it may indicate the potential for greater inherent diversity in this species than the other species. This could present opportunities to select isolates with unique combinations of pathogenicity factors, or it may point to instability in the species, a negative attribute in a biopesticide.
Metarhizium anisopliae is a globally distributed, entomopathogenic fungus that infects many important crop pests including aphids, scarabaeoid beetle larvae and western flower thrips [1–4] (Figure 1). The species was one of the first to be investigated for its use as a biological control agent and advances in the understanding of its biology and ecology have led to improved biocontrol applications . M. anisopliae is regarded as asexual as no teleomorph has been observed . In cases such as these, phylogenetic species boundaries are often used to taxonomically characterize anamorphic fungi  and M. anisopliae has been well characterized in this regard [5, 6, 8, 9]. Analysis of mating type loci (idiomorphs) however, can enhance our understanding of the genetic mechanisms behind sexual or asexual lifestyles and the potential pathways of genetic exchange.
The Metarhizium species complex is diverse, including generalist species with a broad host range and specialist species with narrow host ranges. Furthermore, individual isolates can also exhibit a range of cultural variability. In our laboratory, variability of cultures is minimized through the use of single spore isolates and mother cultures from long term storage. Despite best practice, significant variability arises in the colour and amount of sporulation from replicates of the same single spore isolates (BRIP 53293) cultured on SDAY plates of identical composition and grown under identical temperature regimes (Figure 2). Two main morphologies have been observed: 1) highly sporulating olive green cultures and 2) low sporulating tri-colour cultures, that is orange, pale green and olive green. In addition to these morphologies, some cultures also exhibit more abundant fluffy mycelial growth while others tend to sector, a sign of aging . Culture degeneration has been shown to affect the stability of enzyme production (e.g. cuticle degrading enzymes) and secondary metabolite production (e.g. destruxins) and results in extensive downstream gene regulation [10, 11].
While possible mechanisms responsible for sectoring may include changes due to physiological and environmental adaptations or differences in transposable elements [11–13], our observations of cultural phenotype variation under laboratory conditions are not well understood. Phenotypic variability which affects sporulation in the laboratory could also affect the multiplication of blastospores in the host haemocoel, isolate pathogenicity, as well as commercial conidial production, which has important implications for maintenance, preservation and the selection of isolates for commercialization [11, 14]. The existence of unique divergent gene sets and expansion of gene families within M. anisopliae may provide an important genetic resource for the continued development of this important entomopathogen as a biopesticide.
The recent publication of reference genomes and comparative genomics of the generalist M. anisopliae and the locust-specific pathogen M. acridum has further enhanced our understanding of the biology of these entomopathogenic fungi and the molecular basis of host-specificity [5, 13]. However the isolate sequenced as the reference strain for M. anisopliae (ARSEF23) was subsequently re-classified to M. robertsii, a phylogenetically distinct species within the Metarhizium PARB clade (M. pingshaense, M. anisopliae, M. robertsii and M. brunneum) .
In light of the well-supported divergence of ARSEF23 from the main M. anisopliae clade [5, 9] and re-classification to M. robertsii, the primary aim of this study was to assemble an accurate, annotated, draft reference genome for M. anisopliae using our isolate Ma69. We used a comparative genomics approach, not to repeat the excellent work of the previous publication , but to complement it, add to the body of knowledge and establish new genomic resources for the Metarhizium genus. In this study, we compare the genomes of M. anisopliae, M. robertsii and M. acridum to identify the key suite of genes which differentiate M. anisopliae from the other two species. We used a multi-faceted bioinformatics approach to identify genes that were divergent between the three Metarhizium species and to assign putative functions to them where possible. We identify a suite of effector-like genes that are predicted to be specific to M. anisopliae, significant differences in the repetitive DNA complements, repeat-induced point mutations and mating type gene composition between the three species and discuss the implications of these findings.
Phylogenetic validation, genome sequencing and assembly of M. anisopliaeisolate Ma69
The rDNA internal transcribed spacer (ITS) region of M. anisopliae isolate Ma69 was sequenced and analyzed by BLASTN to confirm its identity . One hundred hits to M. anisopliae ITS sequences with an e-value of 0.00 were obtained, confirming isolate Ma69 as M. anisopliae (Additional file 1. Top ten hits only).
The genome of isolate Ma69 was then shotgun-sequenced by generating two Illumina libraries, a 100 bp paired-end library and a 3 kb mate-paired library, comprising 5.61 Gb and 9.29 Gb of raw data respectively. This corresponded to approximately 380X coverage of the final genome assembly. A total of 142.9 million reads (96.02%) were retained for assembly after quality-control filters were applied for the removal of adapter sequence and regions of low base-call quality or low sequence complexity. Initial de novo assembly using only paired-end reads produced an assembly containing 1,567 scaffolds with an N50 of 138 and an N50 length of 85,355 bp. Contigs were scaffolded with the 3 kb mate-pair library, resulting in 577 scaffolds with an N50 of 11 and an N50 length of 1,243,138 bp. The genome assembly statistics and general features were tabled along with those of M. robertsii and M. acridum (Table 1). The whole genome shotgun project of the organism M. anisopliae, isolate BRIP 53293 EFD69 SSC31 (Ma69), was deposited at DDBJ/EMBL-Bank/GenBank under the accession APNB00000000.
Whole-genome synteny comparisons between Metarhizium species
Synteny dot-plots were generated for pair-wise comparisons between the genomes of Ma69, ARSEF23 and CQMa102. Significant macrosynteny and high levels of sequence homology (≥95% sequence identity) were observed between the genomes of M. anisopliae and M. robertsii (Figure 3). Macrosynteny was also observed between M. anisopliae and M. acridum as well as between M. robertsii and M. acridum, albeit with lower levels of sequence homology (at ~90-95% sequence identity).
Gene prediction and predicted protein orthology
A total of 11,415 protein-encoding genes were predicted within the M. anisopliae Ma69 genome assembly, compared with 10,582 and 9,849 from M. robertsii and M. acridum, respectively . Orthology relationships between the translated proteins of the three Metarhizium species were predicted. While the number of predicted proteins in M. anisopliae was greater than those of M. robertsii and M. acridum (an increase of 833 and 1,566 respectively), most of these appeared to belong to gene families that are expanded in M. anisopliae (Additional file 2). A total of 127 proteins from M. anisopliae were predicted to have no orthologs in either M. robertsii or M. acridum (Additional file 3). These proteins were ‘unique-by-orthology’ to M. anisopliae and are referred to in this study as ‘divergent’. Additionally, groups of orthologs in which the number of proteins belonging to one species was greater than the other two were classified as ‘expanded’. There were 297 expanded groups containing 603 proteins in M. anisopliae, 257 groups containing 562 proteins in M. robertsii and 250 groups containing 540 genes in M. acridum (Additional file 4).
Functional annotation of species-specific (divergent) genes, expanded gene families and effector-like proteins in M. anisopliae, M. robertsii and M. acridum
Orthology relationships between the predicted proteins of the three Metarhizium species were used as a basis for predicting ‘divergent’ and ‘expanded’ gene families. Expanded gene families in all 3 species were abundant in proteins with similar Pfam annotations. These were generally related to membrane transport, sugar metabolism, protein kinases, cytochrome p450s, and transcription factors (Additional file 5). Putative functional annotations were assigned to divergent genes by comparison to multiple databases and algorithms, including: BLASTp versus NCBI Protein and Swissprot, gene ontologies (GOs), Interpro, Pfam, SignalP, WolfPsort and BioPerl::SeqStats. Functional annotations were then ‘manually curated’ based on the sum collection of supporting evidence for each gene, with a view, where possible, to intelligibly describe its putative role in pathogenicity. For the purposes of summarizing this analysis, the divergent genes were then sorted into generalized categories based on their putative role and/or function. Of the 127 genes found to be divergent in M. anisopliae, 56 (44.1%) contained motifs known to be associated with pathogenicity in other species, 8 (6.3%) had homology to genes in PHIbase , 70 (55.1%) were low-molecular weight proteins of ≤30 KDa and 10 (7.87%) had predicted signal peptides (Additional file 6). The divergent genes of M. anisopliae were found, compared to its complete gene content, to be relatively more abundant in genes encoding membrane-anchored proteins, transposable elements, or having unknown functions and other functions described in more detail in sub-sections below. Genes encoding DNA/RNA binding factors and degradative enzymes were more abundant in the divergent genes of M. robertsii relative to its complete gene content (Additional file 7).
DNA/RNA binding – transcriptional regulation
Thirteen divergent genes of M. anisopliae were broadly characterized as having DNA/RNA binding functions, that is, potentially involved in the regulation of gene expression and intracellular signaling. No signal peptides were identified on these genes. Four of these genes were putative helicases, 2 were putative PIWI-like argonaut/dicer proteins, 4 had endonuclease/exonuclease/phosphatase (EEP) domains, 2 were putative fungal transcriptional factors and one was a putative ribonuclease H homolog.
M. anisopliae had six divergent genes that were grouped in three pairs of paralogs that encoded enzymes with a degradative function. The products of gene paralogs Ma69_03389/Ma69_04038 were putative peptidases (GO: 0008238). Gene paralogs Ma69_06536/Ma69_11051 also encoded putative peptidases, each with a signal peptide and with homology to a Phytophthora sojae GIP1-like effector in PHIbase (PHI:653). Gene paralogs Ma69_00354/Ma69_02206 encoded putative subtilisins. Ma69_00354 in particular was assigned GO terms indicating: serine type endopeptidase (GO: 0004252); active evasion of host immune response via regulation of host complement system (GO: 0042874); alkaline serine protease alp1 (GO: 0005576) and pathogenesis (GO: 0009405).
Twelve genes, divergent in M. anisopliae were classified as encoding membrane-anchored proteins. These genes included a putative arrestin-like G-protein coupled receptor, several membrane-associated proteins and a major facilitator superfamily (MFS) membrane transporter protein. Two genes were homologous to CaNAG4, a membrane transport protein (PHI: 511) and CaMDR1, a multidrug transporter protein (PHI: 26) both found in Candida albicans.
Transposable elements, other genes and unknown genes
Fourteen genes divergent in M. anisopliae were identified as transposons and were therefore not considered of interest to this study. Of the remaining divergent genes in M. anisopliae, 44 had functions including protein kinases, histidine kinases and heat-shock proteins. Four divergent genes had homology to fungal pathogenicity genes in PHIbase. Two of these were homologous to TOXF from Cochliobolus carbonum and two had homology to CTB3 from Cercospora nicotianae (PHI: 157 and PHI: 1051 respectively). Due to a lack of homology to the datasets screened, 38 divergent genes, lacked functional annotations and as such their putative biological roles are unknown. Two of these unknown genes had predicted signal peptides.
Divergent candidate secreted effector proteins in M. anisopliaeMa69
Candidate secreted effector proteins are defined here as small protein molecules (≤300 amino acids in length) which are putatively capable of being secreted by the pathogen via the eukaryotic secretory pathway. Secretion to the apoplast relies on the presence of an N-terminal signal peptide and some of these molecules are transferred into plant cells, facilitated by an amino acid motif, downstream of the signal peptide. Of the 127 divergent genes in M. anisopliae, 10 were found to encode putative signal peptides. Nine of these were ≤ 300 amino acids in length, four had pathogenicity motifs and these proteins were deemed candidate secreted effector proteins. These candidate secreted effector proteins were assigned to various functional categories including: degradative enzymes (3), membrane anchored proteins (2), other (2) and unknown genes (2).
Repetitive DNA content and repeat-induced point mutation (RIP)
Repetitive DNA analysis was performed on the scaffold sequences of M. anisopliae and the published reference genomes of M. robertsii and M. acridum. Repetitive sequences were predicted de novo and RIPCAL was used to determine genome-wide dinucleotide frequencies and to quantify RIP-like polymorphisms (SNP mutation biased towards CpA → TpA or its reverse complement TpG → TpA) within alignments of each repeat family (using the deRIP consensus as the model sequence for comparison) [16, 17]. To facilitate comparison of repeat types between Metarhizium spp., their respective genomic matches to characterized repeat sequences in RepBase  are also shown in Figure 4. Overall, the repetitive contents of the genome of M. anisopliae Ma69 and M. robertsii ARSEF23 were similar and relatively low in comparison to other Pezizomycotina. In contrast, the M. acridum genome assembly contained approximately twice as much repetitive DNA. In total, 2.4% of the genome assembly was predicted to be repetitive for M. anisopliae, 2.12% for M. robertsii, and 4.42% in M. acridum. In the three Metarhizium genomes, the most abundant repeat type was simple/low-complexity repeats, followed by retroelements and DNA transposons. The increased repetitive content of M. acridum was largely due to increased amounts of simple/low-complexity sequences. Interestingly, M. acridum was also relatively depleted in transposons and though maintaining a comparable level of LTR retroelements was relatively deficient in LINEs and DNA transposons.
Repeat families were scanned for repeat-induced point mutation (RIP)-like dinucleotide changes using two methods. The first involved genome-wide comparison of dinucleotide frequencies between de novo-identified repetitive sequences and non-repetitive sequences (Additional file 8). The RIP index TpA/ApT, which measures the frequency of the common RIP-product TpA versus a non-RIP like control ApT, was 1.02 within repeats of M. anisopliae and M. robertsii and 1.31 in M. acridum. Dinucleotide frequency analysis also showed an overall increase in TpA dinucleotides in M. acridum relative to M. anisopliae and M. robertsii and a corresponding depletion of the RIP-target dinucleotides CpA and TpG (Additional file 8).
The second RIP-quantitation method used the RIPCAL alignment-based method  versus a ‘deRIPped’ consensus of each family as a reference for comparison . RIP mutation statistics for all repeat families in all three species are tabled in Additional file 9. RIP levels were on the whole relatively low in all three species, however all three species had a rid (cytosine-5 methyltransferase) homolog . RIPCAL analysis showed that M. anisopliae and M. robertsii had similar total levels of RIP mutations and both species exhibited elevated levels of mutation of CpA dinucleotides (as well as CpT in some cases) typical of RIP in the Pezizomycotina in some repeat families. It should be noted that RIP mutation requires a minimum repeat length and we would not expect it to occur in repeats less than 400 bp in length [20, 21]. However assemblies containing short scaffolds such as those used in this study means these Metarhizium assemblies are likely to contain a high number of short repeat families which are incomplete versions of full length repeats. A total of 40 repeat families of M. anisopliae, ranging in average repeat family length of 47 to 375 bp, had CpA↔TpA (RIP-like) dominance scores of 1 or greater (Additional file 9 - summary). In M. robertsii, 12 repeat families had a similar RIP-like dominance score, with lengths ranging from 54 to 173 bp. In contrast, M. acridum had 56 repeat families with a RIP-like dominance score ≥1, lengths ranging from 45 to 1225 bp. Interestingly, the total number of RIP-like mutations in M. robertsii (7624) (relative to ‘deRIPped’ repeat consensus sequences), was approximately five times fewer than those of M. anisopliae and M. acridum (34960 and 39934 respectively).
Mating (MAT) type gene analysis
Analysis of orthologous relationships between the genes of the three Metarhizium species in this study identified incongruence in the number and presence of MAT genes (Figure 5). Ma69 had three genes putatively identified as MAT 1-1-1 (MA69_8894), MAT1-1-2 (MA69_8895) and MAT1-1-3 (MA69_8896) with no orthologous pairs in M. acridum, and orthologs of MAT 1-1-1(MAA_03718) and MAT1-1-3 (MAA_03719) in M. robertsii. Ma69 also had one gene (MA69_3509) putatively identified as MAT1-2 which had an ortholog (MAC_07229) in M. acridum, but not in M. robertsii.
The four putative MAT genes from Ma69 were subjected to BLASTp analysis and were found to have near identical sequence homology to ascomycetous mating type (MAT) genes (Additional file 10). The top ten sequence hits for each MAT gene were aligned to confirm the identity of the genes and identify conserved domains (Additional file 11). The M. anisopliae MAT1-1 alpha box domain [Pfam: PF04769] was confirmed by alignment to known MAT1-1 sequences and carries a conserved intron (Figure 6) [22, 23]. The M. anisopliae MAT1-2 HMG-box domain [Pfam: PF00505] was also confirmed by alignment with other MAT1-2 sequences and also carries a conserved intron (Figure 7) [22, 24]. The presence of both MAT idiomorphs in the genome of the M. anisopliae isolate sequenced indicates it to be putatively homothallic.
Prediction of secreted effector protein candidates via species-specificity and secretion predictions
To our knowledge, pathogenicity effector-like motifs have not yet been examined in Metarhizium species. Some experimentally-validated pathogenicity effector proteins exhibit short conserved amino-acid motifs (Additional file 12), that in some cases have been demonstrated to facilitate host-cell import . We combined protein matches to these known ‘effector motifs’, with predictions of signal-peptides (which infer extracellular secretion of the protein), to arrive at a conservative set of predicted secreted effector protein candidates.
In total, 1,620 proteins in M. anisopliae Ma69 contained sequences matching to a motifs associated with pathogenicity in other fungal species (Additional file 13). The most abundant motifs found were [YWF]XC, [LI]XAR and RXLR. These motifs are short (generally 3–4 amino acids), thus a substantial number of false-positive matches to proteins would be expected, in particular in long, intracellular proteins unrelated to pathogenicity. However these motifs were only considered and found in high numbers among groups of genes that had been previously bioinformatically-filtered for protein properties relevant to pathogenicity. Potential pathogenicity genes were identified using orthology as a basis for predicting species-specific (‘divergent’) genes in all 3 species, which are strong candidates as determinants of their respective differences in host-range phenotype. Seven amino acid motifs known to be associated with pathogenicity proteins were found to match several divergent genes of M. anisopliae, M. robertsii and M. acridum (Additional file 14).
Furthermore, 1,295 proteins were predicted to be secreted in M. anisopliae Ma69 (Additional file 15). We identified 242 of these proteins that we term ‘candidate effectors’, which both contained a pathogenicity effector motif and were predicted to be secreted (Additional file 16). Among these 242 candidate effectors, 6 motifs were found, in descending order of abundance: [YFW]XC (166 proteins), [LI]XAR (54), RXLR (18), CHXC (2), KECXD (1) and YXSL[RK] (1).
M. anisopliae is a fungus with significant commercial and industrial applications as a biopesticide. The host and environmental range of M. anisopliae is regarded as entomologically cosmopolitan, compared to that of M. acridum which is host-specific. Recent molecular analysis of the M. anisopliae complex has proposed its division into nine taxa, or distinct species, which individually may have narrower host specificities [5, 8]. In a multigene phylogeny, M. anisopliae clustered with three other species M. pingshaense, M. robertsii and M. brunneum forming a distinct clade (PARB) . Whether or not these represent discrete biological species or merely formae specialis is not known, as M. anisopliae has no known teleomorph  and the basis of its assumed asexuality has not been determined. Furthermore, advancing the currently poor understanding of the underlying mechanisms of genetic exchange could have important implications for understanding the persistence of traits of biopesticides in commercial applications.
Given that the isolate ARSEF23 sequenced as M. anisopliae has subsequently been re-classified as M. robertsii, and is sufficiently divergent from M. anisopliae[5, 9], an alternative genome reference is required for this important biocontrol fungus to differentiate it from and complement the existing Metarhizium genomic resources. Thus, we sequenced the genome of M. anisopliae isolate Ma69 using Illumina short-read sequence data, producing an annotated, de novo draft genome assembly of 577 scaffolds with an N50 of 11 and an N50 length of 1.24 Mb. The whole genome assembly and its predicted gene content was compared to the published genomes of M. robertsii (ARSEF23) and M. acridum (CQMa102) (Table 1), with particular focus on the species-specific genes of M. anisopliae Ma69, mating type and bioinformatic prediction of its candidate secreted effector proteins.
Whole genome synteny, homology and divergent genes in M. anisopliae
The comparative genomics of M. robertsii and M. acridum has been comprehensively examined  and we have refrained from duplicating previous efforts, except to include and compare M. anisopliae in that context. The previously published genome analysis of M. robertsii showed syntenic conservation with gene clusters of M. acridum but did not present any whole-genome scale synteny comparisons. This was likely due to the poorer contiguity (as indicated by N50, Table 1) of the M. acridum genome assembly, which would prevent whole-genome scale synteny from being accurately observed. In this study, whole genome comparisons including the new assembly of M. anisopliae, in which the largest scaffolds are of a comparable level of contiguity with those of M. robertsii, indicate a generally macrosyntenic conservation pattern (Figure 3, Additional file 17 and Additional file 18). A low level of intra-chromosomal rearrangement is also observed, however this is similar to levels of degraded macrosynteny between other Pezizomycotina species of the same genus, such as previously observed between the Aspergilli . We found that both M. anisopliae and M. robertsii had nucleotide sequence identities ≥ 90% to M. acridum, but sequence identity to each other was ≥ 95%, confirming that M. anisopliae and M. robertsii are more closely related to each other than to M. acridum.
A total of 11,415 proteins were predicted in the M. anisopliae genome assembly, 833 and 1,566 more than M. robertsii and M. acridum respectively. We examined the orthology of genes within the three Metarhizium species and found 127 genes in Ma69 that had no orthologs in M. robertsii and M. acridum, which we refer to in this study as ‘divergent’ genes. We also defined a set of ‘expanded’ genes in which the number of inparalogs of a single species was greater than corresponding outparalogs in the other two species, of which for M. anisopliae there were 603 expanded genes within 297 ortholog groups.
Repetitive DNA, TE classes and RIP analysis
Repetitive sequences are associated with transposable elements (TEs) which play a central role in the evolutionary restructuring of fungal genomes due to their ability to move within the host genome, causing a range of mutations . Excluding deleterious insertions, the mutational activity of TEs may promote genetic diversity and speed up adaptative evolution in the host . Class I (retroelements) use a ‘copy and paste’ mechanism to transpose via the reverse-transcription of an RNA intermediate and include long terminal repeats (LTRs), non LTRs, and long and short interspersed nuclear elements (LINEs and SINEs, respectively [27, 29]. Class II TEs (DNA transposons) ‘cut and paste’ directly through a DNA form, using the enzyme transposase [27, 29]. Earlier, we detailed our observations of phenotypic variation within single spored cultures of M. anisopliae (Figure 2). Due to their ability to directly excise from DNA and re-insert elsewhere in the genome, DNA transposons can generate a wide range of DNA sequence variation which may result in phenotypic changes and may account for the variety of cultural phenotypes observed. Alternatively, the abundance of Class II TEs in M. anisopliae may also affect cultural morphology, as evidence by LINE mediated mutations causing changes to conidial pattern formation in Magnaporthe grisea.
At the gene level, insertion of TEs either in or adjacent to genes may cause partial or total gene inactivation, resulting in new phenotypes [27, 29]. On a genome-wide scale, TEs may be associated with large scale chromosomal modifications such as deletions, inversions and translocations . To prevent or minimize potential deleterious effects of TEs, some fungi possess a gene silencing mechanism known as repeat induced point (RIP) mutation which targets duplicated DNA sequences > 400 bp long with > 80% shared identity . Comparisons of the repeat content of the three Metarhizium spp., using reciprocal blast clustering, revealed surprisingly few common repetitive sequences. Further analysis of RIP-like polymorphism in all three Metarhizium spp. confirmed a common dinucleotide bias for CpA which conforms with expectations for species of the Pezizomycotina. As found in a number of species of Pezizomycotina , all three Metarhizium species examined here possessed a rid homolog which is essential for RIP and although the function of these genes is untested, we assume that RIP is active in all three species. In silico analysis of repetitive DNA and RIP across the three Metarhizium spp. indicated that repetitive content in M. anisopliae and M. robertsii was low at around 2% of genomic DNA and doubled in M. acridum at around 4%. Dinucleotide frequency analysis supports active RIP in all 3 species, with elevated frequencies of the RIP-product TpA and depleted levels of the primary RIP targets CpA and TpG within their respective repetitive DNA complements. M. acridum also exhibited a significant increase in TpA and decrease in CpA and TpG relative to the other two species, perhaps indicative of its higher transposon content. However a more complex picture emerges after alignment-based RIPCAL analysis, which supports similar levels of RIP mutations between M. anisopliae and M. acridum, with a predicted 5-fold decrease in RIP-like polymorphism in M. robertsii. We speculate that while M. anisopliae and M. robertsii appear to contain similar levels of repetitive DNA, RIP activity may be greater in M. anisopliae than in M. robertsii. Higher numbers of RIP-like polymorphism in M. anisopliae and M. acridum may suggest a slightly greater adaptive potential compared to M. robertsii, as pathogenicity-related genes have previously been demonstrated to be affected by RIP mutations leaking outwards from flanking repetitive DNA [33, 34].
MAT gene orthology
Sexual function in filamentous ascomycetes is determined by mating type loci (MAT) which has been extensively described [35–37]. Typically, the single locus determining mating behaviour between mating partners contains different genes which are not allelic and are therefore known as idiomorphs [35, 36]. In filamentous ascomycetes (Pezizomycotina), mat genes encode DNA binding motifs (high-mobility group (HMG) boxes and α domains)  and are responsible for the control of both mating and incompatibility, cell-cell recognition and recognition between nuclei . The lack of an observed sexual lifestyle in Metarhizium species may be the result of a loss of gene function, the lack of an opposite mating type, or merely the inability to induce a teleomorph under laboratory conditions. Functional MAT genes have been discovered in other ascomycetes previously assumed to be asexual (formerly known as Deuteromycetes) but the potential for a sexual cycle in these remains enigmatic [37–40] and indeed, its manifestation may occur on a continuum of sexuality in fungi, ranging from common to rare .
We identified both MAT1-1 and MAT1-2 idiomorphs in M. anisopliae, indicating that it is putatively homothallic and possibly capable of sexual reproduction. The Ma69 MAT1-1 idiomorph encodes three proteins: MAT1-1-1, an α domain protein; MAT1-1-2, an amphipathic α-helical protein; and MAT1-1-3, an HMG box protein . At this point, the functionality of these genes in M. anisopliae is not known, however sequence analysis has yielded a high level of conservation of alpha box and HMG domains with other ascomycetes. This warrants further investigation, including gene expression and/or transformation of closely related teleomorph species.
The M. acridum isolate CQMa102 possessed the MAT1-2 gene but lacked the MAT1-1 idiomorph. The MAT1-2 gene encoded an HMG-domain protein that was highly conserved with M. anisopliae and other ascomycetes. Until an opposite mating type can be found and functionality is confirmed, heterothallism in M. acridum is deemed putative at this juncture. In contrast to both M. anisopliae and M. acridum, the M. robertsii isolate ARSEF23 possessed an incomplete MAT1-1 idiomorph and lacked an ortholog to MAT1-2. Absence of the MAT1-2 idiomorph would indicate that the isolate sequenced was MAT1-1 heterothallic, however again, in the absence of an opposite mating type and without confirmation of functionality, heterothallism is deemed putative. The relatively similar extents of RIP-like polymorphism in M. anisopliae and M. acridum, given the putative homothallism of M. anisopliae, is perplexing. RIP occurs only during meiosis, and as such, evidence of RIP may indicate the occurrence of meiosis in an ancestral or cryptic sexual stage, or alternatively, the existence of a similar process in vegetative cells . The results of RIP analysis for M. robertsii, which indicate approximately 5 times less RIP than in the other Metarhizium spp., lends support to impaired or less frequent sexual activity in this species.
The missing MAT1-1-2 gene in M. robertsii, a phenomenon also observed in Pyrenopeziza brassicae and Cochliobolus heterostrophus, two heterothallic species with sexual stages  is intriguing. A functional MAT1-1-1 gene is critical to mating identity, sexual development  and vegetative incompatibility  as the alpha box domain, which is subsequently processed into mature pheromone molecules is located here . However the requirement for a functional MAT1-1-2 is less clear. In the aforementioned example of P. brassicae, the absence of MAT1-1-2 does not impede out crossing, indeed the teleomorph is found naturally in oilseed rape . In another example, Neurospora crassa, the homologs of MAT1-1-2 and MAT1-1-3 (reported as matA-2 and matA-3, respectively) were reported to be non-essential for mating or ascospore production [43, 46], however their expression increased the efficiency of sexual development . It was also reported that homologs of MAT1-1-2 and MAT1-1-3 in Podospora anserina (reported as α-helical genes and HMG-1 genes, respectively) were not required for mating but were required for sexual development and biparental progeny, as mutations in the α-helical gene lead to barren fruiting bodies [37, 47]. The evidence suggests that while MAT1-1-1 is crucial to mating, the MAT1-1-2 gene may have a supporting role in some ascomycetes in the successful development of sexual bodies, post mating, while in other species, its absence does not impede the development of fit progeny. The effect of the absence of MAT1-1-2 in M. robertsii is unknown at this point, however it is intriguing that its closely related sibling species, M. anisopliae possesses the full complement of MAT1-1 genes. In future, the discovery of isolates of M. robertsii and M. acridum with complementary MAT idiomorphs followed by evolutionary analysis of these idiomorphs may help to further our understanding of the role of MAT1-1-2 in the Metarhizium species complex, and the potential effect on genetic exchange and perhaps, the observable phenotypic variation between cultures.
Candidate secreted effector proteins
In this analysis, we identified a suite of genes which have the characteristics of effector proteins. Effector proteins are defined as molecules produced by a pathogen, which can alter host cell structure or function, thereby facilitating infection and/or initiating defense mechanisms . Putative effector proteins are characterized in this study as being ≤ 300 amino acids in length and have a predicted signal peptide, which would facilitate secretion into the pathogen’s extracellular space. In addition to these characteristics, effector predictions may also be supported by matches to proteins with previously identified pathogenicity motifs.
Proportionately, the genome of M. anisopliae had similar numbers of putative secreted effector-like proteins to M. robertsii, and more secreted effector-like proteins than M. acridum suggesting a similar capability or role of secreted proteins between M. anisopliae and M. robertsii. Ten predicted secreted effectors in M. anisopliae were divergent from M. robertsii and M. acridum. In particular, 3 of these divergent secreted effectors were determined to be putative degradative enzymes, which is of significant biological interest to Metarhizium biopesticide research and warrants further investigation to identify their function and potential effect on host range.
We found six motifs known to be associated with pathogenicity in other species, to be present in 242 candidate secreted effector-like proteins of M. anisopliae. Of these, 166 matched [YFW]xC, 54 matched [LI]XAR and 18 matched RXLR and represent the first analysis of effector-assosciated motifs in an entomopathogenic fungus. The best characterized of these motifs is the RXLR motif found in the oomycete Phytophthora infestans, although conserved RXLR motif effectors have yet to be observed in fungi. In pathogenicity effectors of P. infestans, the RXLR motif is found adjacent to the signal peptide, on the N-terminal end of the mature cleaved protein, and has been shown to facilitate translocation of the effector protein across the host membrane into the host cell. The [LI]XAR motif was identified in effector proteins secreted by the rice blast pathogen Magnaporthe oryzae, however the function of this motif is unknown . As such, the function of these motifs in Metarhizium remains speculative.
While the high number of [YFW]XC motifs among the two predicted sets of genes, divergent and effector candidates, may simply be an artifact of random background matches to this short, three residue motif, there is some evidence suggesting that there may be an important, albeit, still unknown role for [YFW]XC motifs in secreted proteins. The [YFW]XC motif was first reported in the powdery mildew fungus Blumeria graminis. It was predominantly expressed in the haustoria and was also over-represented among its predicted secretome . Since then, the [YFW]XC motif has also been discovered among effector candidates from the haustoria-producing rust fungi Puccinia graminis f.sp. tritici, P. striiformis f.sp. tritici and Melampsora larici-populina (as cited by Pedersen). M. anisopliae also produces haustoria in vitro under nutrient deprivation  however the link between the [YFW]XC motif and production of haustoria is still yet to be resolved. The function of the [YFW]XC motif still remains unclear, however it may have arisen from an extracellular RNase ancestor  and may be involved in establishing disulphide bridges with a C-terminal cysteine residue (Thordal-Christensen pers. comm.) thereby assisting protein folding and enhancing extracellular stability. Pedersen et al.,  hypothesize that a secreted fungal ribonuclease appears to be the common origin of many of their candidates for secreted effector proteins (CSEPs). They suggested that some of these CSEPs could still be involved in interactions with host RNAs and modulate host immunity via this route. They also go on to suggest that extracellular ribonucleases are very stable molecules, resistant to proteolytic degradation, thereby providing a rigid scaffold ideal for evolving an effector arsenal, in which exposed loop regions subjected to positive selection allow diversification and evasion of host recognition. The structural conservation among effector candidates from diverse plant pathogenic species supports the hypothesis of an ancient common ancestor. The entomopathogenic M. anisopliae is likely to have arisen from a plant-pathogenic predecessor  and the discovery of these motifs in Metarhizium is consistent with the high level of conservation and putative functional requirement of these for pathogenicity.
Genomics, phenotypic variability and biopesticides
The key to successful commercialization of biopesticides is the identification and selection of virulent microbial isolates which are amenable to industrial scale production and formulation . Information from comparative genomics studies can be used to identify genes which may contribute to isolate virulence and fitness as well as other characteristics which affect cultural variability and stability of potential biopesticides.
The discovery of both MAT idiomorphs in M. anisopliae raises more questions about its perceived asexuality, the potential pathways of genetic exchange in this species, impact on virulence, and implications for industrial applications. Understanding the molecular basis of mating genes not only gives insight into fundamental processes such as evolution of homothallism, heterothallism and asexuality, but also facilitates research on ascomycetous species of industrial interest and subsequent applied aspects . Proof of function or not in these genes, will enhance the understanding of genetic exchange pathways in this species.
Overall, the repetitive contents of the genome of M. anisopliae Ma69 and M. robertsii ARSEF23 were similar and relatively low in comparison to other Pezizomycotina. In contrast, the M. acridum genome assembly contained approximately twice as much repetitive DNA. RIP levels were on the whole relatively low in all three species, however all three species had a rid (cytosine-5 methyltransferase) homolog and we observed SNP polymorphism consistent with active RIP in all three species. M. anisopliae and M. robertsii had similar total levels of RIP –like mutations and both species exhibited elevated levels of mutation of CpA dinucleotides (as well as CpT in some cases) typical of RIP in the Pezizomycotina in some repeat families. Interestingly, the total number of RIP-like mutations in M. robertsii was approximately five times fewer than those of M. anisopliae and M. acridum, lending support to impaired or less frequent sexual activity in this species. RIP occurs only during meiosis, and as such, evidence of RIP in these species may indicate the occurrence of meiosis in an ancestral or cryptic sexual stage, or alternatively, the existence of a similar process in vegetative cells.
The availability of an annotated whole genome sequence for M. anisopliae adds value to the already published genomes of M. robertsii and M. acridum; however in and of itself, represents a significant resource for future research into this agriculturally important fungal biopesticide. The nomenclature of the species M. robertsii needs to be widely adopted immediately to prevent further confusion with M. anisopliae, therefore the publication of the genome reference of M. anisopliae will serve as a valuable reference to differentiate it from M. robertsii.
Origin and culture of fungal strain Ma69
An Australian isolate of M. anisopliae known to be pathogenic to aphids was selected for whole genome sequencing. Originally, M. anisopliae isolate BRIP 53293 was isolated from soil in the Kingaroy region, Queensland and was obtained from the Queensland Department of Employment, Economic Development and Innovation (DEEDI). A single spore culture of BRIP 53293 (BRIP 53293 EFD 69 SSC31) was prepared on Sabouraud Dextrose Agar (SDA) before being transferred to Sabouraud Dextrose Broth (SDB), where cultures were shaken (150 rpm) at 25°C for 5 days prior to DNA extraction. For brevity, isolate BRIP53293 EFD69 SSC31 will be referred to hereafter as Ma69.
Genome sequencing and assembly
Genomic DNA was extracted from the single-spored fungal culture using NucleoSpin Plant II (Machery Nagel) DNA extraction kit as per the manufacturer’s instruction. The ITS region of ribosomal DNA (rDNA) was sequenced  and subjected to BLAST  analysis to confirm the identity of the isolates. Sufficient genomic DNA was then prepared for paired-end and mate-pair sequencing at the Australian Genome Research Facility (AGRF), Brisbane according to Illumina protocols. Raw read sequences were trimmed via cutadapt  for: Illumina adaptor sequences; PCR primer and barcode contaminant sequences; homopolymers and runs of unknown bases > 5 bp; base quality > Q30 and; a minimum read length after trimming ≥ 50 bp. Reads with a discarded pair (i.e. with a length of < 50 bp after trimming) were removed from paired-end datasets, but were retained as singleton reads.
De novo assembly of Ma69
Paired-end reads were assembled de novo with Velvet (version 1.1.2) . The k-mer length parameter was tested between 20 and 70 bp in 2 bp increments and optimized for minimal N50 and maximal N50 length. Velvet paired-end scaffolds were then scaffolded with reads from the 3 kb mate-paired library via SSPACE (version 1.1) .
Comparative genomics: whole assembly alignment
Whole genome alignments between the genome assemblies of Ma69, ARSEF23 and CQMa102 were performed with MUMMER 3.23 , using the nucmer algorithm with the –mum parameter. Dot-plots were generated with mummerplot with the –filter, -colour and –fat parameters.
Putative protein-coding regions of the Ma69 assembly were predicted with GeneMark-ES . Predicted gene models were filtered for a minimum protein product length of 50 amino acids. Predicted genes encoding products < 50 amino acids in length were discarded.
Protein family classification and orthology
Predicted proteins of Ma69 were compared for orthology to the ARSEF23 and CQMa102 published protein datasets, obtained from the NCBI NR protein database. Orthologous relationships between genes of these three species were inferred with ProteinOrtho4 using the –pair and –selfblast parameters . A list of divergent genes specific to each species was generated according to non-orthology. In order to assign putative functions to the species-specific gene datasets, the datasets were screened and compared to a number of databases and were then manually-curated based on the sum collection of evidence for each gene. The genes were then sorted into generalized categories based on their putative role/function. The whole proteome datasets for each species were screened via EMBOSS for sixteen motifs known to be involved in fungal pathogenicity (Additional file 9). Homology to PHIbase (v 3.2) protein sequences were assessed by BLASTp (e-value threshold 1e-5). The best BLAST hit was assigned on the basis of highest bit score and the disruption phenotype was recorded. Gene Ontology (GO) terms were assigned to the species-specific gene translated sequences via Blast2GO (default settings with BLASTp, Annex augmentation). Interproscan was also enabled within Blast2GO assigning additional GO term, functional domain and structure annotations. Secretion, sub-cellular localization, molecular weight and iso-electric point were calculated for each species-specific dataset via SignalP 3.0 , WolfPsort  and EMBOSS IEP . Amino acid compositions were calculated using the SeqStats BioPerl module .
Repetitive DNA analysis
De novo prediction of repetitive sequences of Ma69 and the published reference genomes of ARSEF23 (M. robertsii) and CQMa102 (M. acridum) were generated using RepeatScout  while overlapping and redundant consensus repeats were combined via CAP3 . These de novo repeats were mapped to the genome assembly via RepeatMasker . For the purpose of comparison of characterised repeat types, RepeatMasker was also run versus each Metarhizium spp. assembly using RepBase sequences corresponding to the “fungi” taxon. Multiple alignments of repeat families were generated by RIPCAL  using ClustalW [71, 72]. deRIP  was used to predict the pre RIP consensus of each aligned repeat family. RIPCAL  was run on each repeat family alignment using the deRIP consensus as a model sequence for comparison. Repeats for all species were analyzed using TEClass  to predict the most probable class for each repeat family. Repeats common between all species were identified using proteinortho4 using the ‘--p blastn’ parameter .
Mating (MAT) type gene analysis
MAT genes were identified by homology to characterized mating-type sequences using BLASTP . The amino-acid sequences top ten BLASTp hits for each putative Ma69 MAT protein were obtained from NCBI and multiple-alignment on these was performed in CLC bio Genomics Workbench v4.0 (CLC bio, Denmark).
Search for known effector-associated motifs
Proteins were assessed for the presence of domains known to be associated with effectors outlined in Additional file 9 using EMBOSSpreg v 6.5.7 (http://emboss.sourceforge.net/apps/release/6.3/emboss/apps/preg.html).
Shan LT, Feng MG: Evaluation of the biocontrol potential of various Metarhizium isolates against green peach aphic Myzus persicae (Homoptera: Aphididae). Pest Manag Sci. 2010, 66: 669-675.
Roberts DW, St Leger RJ: Metarhizium spp., Cosmopolitan Insect-Pathogenic Fungi: Mycological Aspects. Advances in Applied Microbiology. 2004, Academic Press, 1-70. doi:10.1016/S0065-2164(04)54001-7
Vestergaard S, Gillespie AT, Butt TM, Schreiter G, Eilenberg J: Pathogenicity of the hyphomycete fungi Verticillium-lecanii and Metarhizium-anisopliae to the western flower thrips, Frankliniella-occidentalis. Biocontrol Sci Tech. 1995, 5: 185-192.
Zimmermann G: Review on safety of the entomopathogenic fungus Metarhizium anisopliae. Biocontrol Sci Tech. 2007, 17: 879-920.
Kepler RM, Sung G-H, Ban S, Nakagiri A, Chen M-J, Huang B, Li Z, Spatafora JW: New teleomorph combinations in the entomopathogenic genus Metacordyceps. Mycologia. 2012, 104 (1): 182-197.
Driver F, Milner RJ, Trueman JWH: A taxonomic revision of Metarhizium based on phylogenetic analysis of rDNA sequence data. Mycol Res. 2000, 104 (2): 134-150.
Taylor JW, Jacobson DJ, Fisher MC: The evolution of asexual fungi: reproduction, speciation and classification. Annu Rev Phytopathol. 1999, 37: 197-246.
Bischoff JF, Rehner SA, Humber RA: A multilocus phylogeny of the Metarhizium anisopliae lineage. Mycologia. 2009, 101 (4): 512-530.
Inglis GD, Duke GM, Kabaluk JT, Geottel MA: Genetic diversity of Metarhizium anisopliae var. anisopliae in southwestern British Columbia. J Invertebr Pathol. 2008, 98: 101-113.
Wang CS, Butt TM, St. Leger RJ: Colony sectorization of Metarhizium anisopliae is a sign of ageing. Microbiology - SGM. 2005, 151: 3223-3236.
Ryan MJ, Bridge PD, Smith D, Jeffries P: Phenotypic degeneration occurs during sector formation in Metarhizium anisopliae. J Appl Microbiol. 2002, 93: 163-168.
Wang S, O’Brien TR, Pava-Ripoll M, St. Leger RJ: Local adaptation of an introduced transgenic insect fungal pathogen due to new beneficial mutations. Proc Natl Acad Sci U S A. 2011, 108 (51): 20449-20454.
Gao Q, Jin K, Ying S-H, Zhang Y, Xiao G, Shang Y, Duan Z, Hu X, Xie X-Q, Zhou G, Peng G, Luo Z, Huang W, Wang B, Fang W, Wang S, Zhong Y, Ma L-J, St Leger RJ, Zhao G-P, Pei Y, Feng M-G, Xia Y, Wang C: Genome sequencing and comparative transcriptomics of the model entomopathogenic fungi Metarhizium anisopliae and M. acridum. PLoS Genet. 2011, 7 (1): e1001264-
Ryan MJ, Smith D, Bridge PD, Jeffries P: The relationship between fungal preservation method and secondary metabolite production in Metarhizium anisopliae and Fusarium oxysporum. World J Microbiol Biotechnol. 2003, 19: 839-844.
Winnenburg R, Urban M, Beacham A, Baldwin TK, Holland S, Lindeberg M, Hansen H, Rawlings C, Hammond-Kosack K, Kohler J: PHI-base update: additions to the pathogen-host interaction database. Nucleic Acids Res. 2008, 36 (SI): D572-D576.
Hane JK, Oliver RP: RIPCAL: a tool for alignment-based analysis of repeat-induced point mutations in fungal genomic sequences. BMC Bioinformatics. 2008, 9: 478-
Hane JK, Oliver RP: In silico reversal of repeat-induced point mutation (RIP) identifies the origins of repeat families and uncovers obscured duplicated genes. BMC Genomics. 2010, 11: 655-
Jurka J, Kapitonov VV, Pavlicek A, Klonowski P, Kohany O, Walichiewicz J: Repbase Update, a database of eukaryotic repetitive elements. Cytogenet Genome Res. 2005, 110: 462-467.
Freitag M, Williams RL, Kothe GO, Selker EU: A cytosine methyltransferase homologue is essential for repeat-induced point mutation in Neurospora crassa. Proc Natl Acad Sci U S A. 2002, 99 (13): 8802-8807.
Watters MK, Randall TA, Margolin BS, Selker EU, Stadler DR: Action of repeat-induced point mutation on both strands of a duplex and on tandem duplications of various sizes in Neurospora. Genetics. 1999, 153 (2): 705-714.
Cambareri E, Singer M, Selker E: Recurrence of repeat-induced point mutation (RIP) in Neurospora crassa. Genetics. 1991, 127: 699-710.
Woo PCY, Chong KTK, Tse H, Cai JJ, Lau CCY, Zhou AC, Lau SKP, Yuen K-Y: Genomic and experimental evidence for a potential sexual cycle in the pathogenic thermal dimorphic fungus Penicillium marneffei. FEBS Lett. 2006, 580: 3409-3416.
Paoletti M, Rydholm C, Schwier EU, Anderson MJ, Szakacs G, Lutzoni F, Debeaupuis J-P, Latge J-P, Denning DW, Dyer PS: Evidence for sexuality in the opportunistic fungal pathogen Aspergillus fumigatus. Curr Biol. 2005, 15: 1242-1248.
Poggeler S: Genomic evidence for mating abilities in the asexual pathogen Aspergillus fumigatus. Curr Genet. 2002, 42: 153-160.
Kale SD, Tyler BM: Entry of oomycete and fungal effectors into plant and animal host cells. Cell Microbiol. 2011, 13 (12): 1839-1848.
Hane JK, Rouxel T, Howlett BJ, Kema GHJ, Goodwin SB, Oliver RP: A novel mode of chromosomal evolution peculiar to filamentous Ascomycete fungi. Genome Biol. 2011, 12: R45-
Daboussi M-J, Capy P: Transposable elements in filamentous fungi. Annu Rev Microbiol. 2003, 57: 275-299.
Santana MF, Silva JCF, Batista AD, Ribeiro LE, da Silva GF, de Araujo EF, de Queiroz MV: Abundance, distribution and potential impact of transposable elements in the genome of Mycosphaerella fijiensis. BMC Genomics. 2012, 13: 720-
Favaro LCL, de Araujo WL, de Azevedo JL, Paccola-Meirelles LD: The biology and potential for genetic research of transposable elements in filamentous fungi. Genet Mol Biol. 2005, 28 (4): 804-813.
Nishimura M, Hayashi N, Jwa N-S, Lau GW, Hamer JE, Hasebe A: Insertion of the LINE retrotransposon MGL causes a conidiophore pattern mutation in Magnaporth grisea. Mol Plant Microbe Interact. 2000, 8: 892-894.
Novikova OS, Fet V, Vlinov AG: Homology-dependent inactivation of LTR retrotransposons in Aspergillus fumigatus and A. nidulans genome. Mol Biol. 2007, 41: 886-893.
Clutterbuck AJ: Genomic evidence of repeat-induced point mutation (RIP) in filamentous ascomycetes. Fungal Genet Biol. 2011, 48: 306-326.
Fudal I, Ross S, Brun H, Besnard AL, Ermel M, Kuhn ML, Balesdent MH, Rouxel T: Repeat-induced point mutation (RIP) as an alternative mechanism of evolution toward virulence in Leptosphaeria maculans. Mol Plant Microbe Interact. 2009, 22 (8): 932-941.
Van de Wouw AP, Cozijnsen AJ, Hane JK, Brunner PC, McDonald BA, Oliver RP, Howlett BJ: Evolution of linked avirulence effectors in Leptosphaeria maculans is affected by genomic environment and exposure to resistance genes in host plants. PLoS Pathog. 2010, 6 (11): e1001180-
Coppin E, Debuchy R, Arnaise S, Picard M: Mating types and sexual development in filamentous ascomycetes. Microbiol Mol Biol Rev. 1997, 61 (4): 411-428.
Kronstad JW, Staben C: Mating type in filamentous fungi. Annu Rev Genet. 1997, 31 (1): 245-276.
Shiu PKT, Glass NL: Cell and nuclear recognition mechanisms mediated by mating type in filamentous ascomycetes. Curr Opin Microbiol. 2000, 3: 183-188.
Arie T, Kaneko I, Yoshida T, Noguchi M, Nomura Y, Yamaguchi I: Mating-type genes from asexual phytopathogenic Ascomycetes Fusarium oxysporum and Alternaria alternata. Mol Plant Microbe Interact. 2000, 13 (12): 1330-1339.
Yun SH, Arie T, Kaneko I, Yoder OC, Turgeon BG: Molecular organisation of mating type loci in heterothallic, homothallic, and asexual Gibberella/Fusarium species. Fungal Genet Biol. 2000, 31: 7-20.
Sharon A, Yamaguchi K, Christiansen S, Horwitz BA, Yoder OC, Turgeon BG: An asexual fungus has the potential for sexual development. Mol Gen Genet. 1996, 251: 60-68.
Dyer PS, Paoletti M: Reproduction in Aspergillus fumigatus: sexuality in a supposedly asexual species?. Med Mycol. 2005, 43: S7-S17.
Glass N, Grotelueschen J, Metzenberg R: Neurospora crassa A mating-type region. Proc Natl Acad Sci U S A. 1990, 87: 4912-4916.
Ferreira AV, An Z, Metzenberg RL, Glass NL: Characterization of mat A-2, mat A-3 and deltamatA mating-type mutants of Neurospora crassa. Genetics. 1998, 148 (3): 1069-1079.
Jones SK, Bennett RJ: Fungal mating pheromones: Choreographing the dating game. Fungal Genet Biol. 2011, 48: 668-676.
Lacey ME, Rawlinson CJ, McCartney HA: First record of the natural occurrence in England of the teleomorph of Pyrenopeziza brassicae on oilseed rape. Trans Br Mycol Soc. 1987, 89 (1): 135-140.
Bobrowicz P, Pawlak R, Correa A, Bell-Pedersen D, Ebbole DJ: The Neurospora crassa pheromone precursor genes are regulated by the mating type locus and the circadian clock. Mol Microbiol. 2002, 45 (3): 795-804.
Arnaise S, Debuchy R, Picard M: What is a bona fide mating-type gene? Internuclear complementation of mat mutants in Posospora anserina. Mol Gen Genet. 1997, 256 (2): 169-178.
Kamoun S: A catalogue of the effector secretome of plant pathogenic oomycetes. Annu Rev Phytopathol. 2006, 44: 41-60.
Whisson SC, Boevink PC, Moleleki L, Avrova AO, Morales JG, Gilroy EM, Armstrong MR, Grouffaud S, van West P, Chapman S, Hein I, Toth IK, Pritchard L, Birch PRJ: A translocation signal for delivery of oomycete effector proteins into host plant cells. Nature. 2007, 450: 115-119.
Yoshida K, Saitoh H, Fujisawa S, Kanzaki H, Matsumura H, Yoshida K, Tosa Y, Chuma I, Takano Y, Win J, Kamoun S, Terauchi R: Association genetics reveals three novel avirulence genes from the rice blast fungal pathogen Magnaporthe oryzae. Plant Cell. 2009, 21: 1573-1591.
Godfrey D, Bohlenius H, Pedersen C, Zhang Z, Emmersen J, Thordal-Christensen H: Powdery mildew fungal effector candidates share N-terminal Y/F/WxC motif. BMC Genomics. 2010, 11 (1): 317-
St. Leger RJ, Staples RC, Roberts DW: Cloning and regulatory analysis of starvation-stress gene, ssgA, encoding a hydrophobin-like protein from the entomopathogenic fungus, Metarhizium anisopliae. Gene. 1992, 120: 119-124.
Pedersen C, Ver Loren van Themaat E, McGuffin LJ, Abbott JC, Burgis TA, Barton G, Bindschedler LV, Lu X, Maekawa T, Wessling R, Cramer R, Thordal-Christensen H, Panstruga R, Spanu PD: Structure and evolution of barley powdery mildew effector candidates. BMC Genomics. 2012, 13: 694-
Ash GJ: The science, art and business of successful bioherbicides. Biol Control. 2010, 52: 230-240.
Poggeler S: Mating-type genes for classical strain improvements of ascomycetes. Appl Microbiol Biotechnol. 2001, 56: 589-601.
White TJ, Bruns T, Lee S, Taylor J: Amplification and Direct Sequencing of Fungal Ribosomal RNA Genes for Phylogenetics. PCR Protocols: A Guide to Methods and Applications. Edited by: Innis MA, Gelfand DH, Sninsky JJ, White TJ. 1990, San Diego: Academic Press, 315-322.
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ: Basic local alignment search tool. J Mol Biol. 1990, 215: 403-410.
Martin M: Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. 2011, 17 (1): 10-12. http://journal.embnet.org/index.php/embnetjournal/article/view/200,
Zerbino DR, Birney E: Velvet: Algorithms for de novo short read assembly using de Bruijn graphs. Genome Res. 2010, 18 (5): 821-829.
Boetzer M, Henkel CV, Jansen HJ, Butler D, Pirovano W: Scaffolding pre-assembled contigs using SSPACE. Bioinformatics. 2010, 27 (4): 578-579.
Kurtz S, Phillippy A, Delcher AL, Smoot M, Shumway M, Antonescu C, Salzberg SL: Versatile and open software for comparing large genomes. Genome Biol. 2004, 5 (2): R12-
Ter-Hovhannisyan V, Lomsadze A, Chernoff YO, Borodovsky M: Gene prediction in novel fungal genomes using an ab initio algorithm with unsupervised training. Genome Res. 2008, 18: 1979-1990.
Lechner M, Findeib S, Steiner L, Marz M, Stadler PF, Prohaska SJ: Proteinortho: detection of (Co-)orthologs in large-scale analysis. BMC Bioinformatics. 2011, 12: 124-
Petersen TN, Brunak S, von Heijne G, Nielsen H: SignalP 4.0: discriminating signal peptides from transmembrane regions. Nat Methods. 2011, 8: 785-786.
Horton P, Park K-J, Obayashi T, Fujita N, Harada H, Adams-Collier CJ, Nakai K: WoLF PSORT: protein localization predictor. Nucleic Acids Res. 2007, 35 (S): W585-W587.
Rice P, Longden I, Bleasby A: EMBOSS: the European molecular biology open software suite. Trends Genet. 2000, 16 (6): 276-277.
Stajich JE, Block D, Boulez K, Brenner SE, Chervitz SA, Dagdigian C, Fuellen G, Gilbert JG, Korf I, Lapp H, Lehvaslaiho H, Matsalla C, Mungall CJ, Osborne BI, Pocock MR, Schattner P, Senger M, Stein LD, Stupka E, Wildinson MD, Birney E: The Bioperl toolkit: Perl modules for the life sciences. Genome Res. 2002, 12 (10): 1611-1618.
Price AL, Jones NC, Pevzner PA: De novo identification of repeat families in large genomes. Bioinformatics. 2005, 21 (Suppl 1): i351-i358.
Huang W, Madan A: CAP3: a DNA sequence assembly program. Genome Res. 1999, 9 (9): 868-877.
RepeatMasker Open-3.0. http://www.repeatmasker.org,
Chenna R, Sugarwarra H, Koike T, Lopez R, Gibson TJ, Thompson JD: Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res. 2003, 31 (13): 3497-3500.
Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, Thompson JD, Gibson TJ, Higgins DG: Clustal W and Clustal X version 2.0. Bioinformatics. 2007, 23 (21): 2947-2948.
Abrusan G, Grundmann N, DeMester L, Makalowski W: TEclass- a tool for automated classification of unknown eukaryotic transposable elements. Bioinformatics (Oxford, England). 2009, 25 (10): 1329-1330.
JP thanks Mr Darryn Webster for advice and assistance with spreadsheet manipulation. This work was funded by EH Graham Centre for Agricultural Innovation, an alliance between Charles Sturt University and New South Wales Department of Primary Industries. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
The authors declare that they have no competing interests.
JP, BW, BS and GA conceived of the study and participated in its design and coordination. JP carried out the laboratory work. JH and AW performed the bioinformatics analysis. All authors helped to draft and approved the final manuscript.
Electronic supplementary material
Additional file 4: Orthologs and expanded genes: Groups of orthologs in which the number of proteins belonging to one species was greater than the other two were classified as ‘expanded’. There were 297 expanded groups containing 603 proteins in M. anisopliae, 257 groups containing 562 proteins in M. robertsii and 250 groups containing 540 genes in M. acridum.(XLS 3 MB)
Additional file 6: Metarhizium species-specific gene summary: Putative function annotations were assigned to divergent genes by comparison to multiple databases and algorithms, including: BLASTp versus NCBI Protein and Swissprot, gene ontologies (GOs), Interpro, Pfam, SignalP, WolfPsort and BioPerl::SeqStats. Functional annotations were then ‘manually curated’ based on the sum collection of supporting evidence for each gene, with a view to intelligibly describe its putative role in pathogenicity. For the purposes of summarizing this analysis, the divergent genes were then sorted into generalized categories based on their putative role and/or function. (XLS 90 KB)
Additional file 9: RIP mutation statistics for all repeat families in all three species: Repeat families were scanned for repeat-induced point mutation (RIP)-like dinucleotide changes using two methods. The second RIP-quantitation method used the RIPCAL alignment-based method versus a ‘deRIPped’ consensus of each family as a reference for comparison. (XLSX 369 KB)
Authors’ original submitted files for images
Below are the links to the authors’ original submitted files for images.
About this article
Cite this article
Pattemore, J.A., Hane, J.K., Williams, A.H. et al. The genome sequence of the biocontrol fungus Metarhizium anisopliae and comparative genomics of Metarhizium species. BMC Genomics 15, 660 (2014). https://doi.org/10.1186/1471-2164-15-660
- Metarhizium anisopliae