Secondary metabolite gene clusters in the entomopathogen fungus Metarhizium anisopliae: genome identification and patterns of expression in a cuticle infection model

Abstract

Background: The described species from the Metarhizium genus are cosmopolitan fungi that infect arthropod hosts. Interestingly, while some species infect a wide range of hosts (host-generalists), other species infect only a few arthropods (host-specialists). This singular evolutionary trait permits unique comparisons to determine how pathogens and virulence determinants emerge. Among the several virulence determinants that have been described, secondary metabolites (SMs) are suggested to play essential roles during fungal infection. Despite progress in the study of pathogen-host relationships, the majority of genes related to SM production in Metarhizium spp. are uncharacterized, and little is known about their genomic organization, expression and regulation. To better understand how infection conditions may affect SM production in Metarhizium anisopliae, we have performed a deep survey and description of SM biosynthetic gene clusters (BGCs) in M. anisopliae, analyzed RNA-seq data from fungi grown on cattle-tick cuticles, evaluated the differential expression of BGCs, and assessed conservation among the Metarhizium genus. Furthermore, our analysis extended to the construction of a phylogeny for the following three BGCs: a tropolone/citrinin-related compound (MaPKS1), a pseurotin-related compound (MaNRPS-PKS2), and a putative helvolic acid (MaTERP1).

Results: Among 73 BGCs identified in M. anisopliae, 20 % were up-regulated during initial tick cuticle infection and presumably possess virulence-related roles. These up-regulated BGCs include known clusters, such as destruxin, NG39x and ferricrocin, together with putative helvolic acid and, pseurotin and tropolone/citrinin-related compound clusters as well as uncharacterized clusters. Furthermore, several previously characterized and putative BGCs were silent or down-regulated in initial infection conditions, indicating minor participation over the course of infection. Interestingly, several up-regulated BGCs were not conserved in host-specialist species from the Metarhizium genus, indicating differences in the metabolic strategies employed by generalist and specialist species to overcome and kill their host. These differences in metabolic potential may have been partially shaped by horizontal gene transfer (HGT) events, as our phylogenetic analysis provided evidence that the putative helvolic acid cluster in Metarhizium spp. originated from an HGT event.

Conclusions: Several unknown BGCs are described, and aspects of their organization, regulation and origin are discussed, providing further support for the impact of SM on the Metarhizium genus lifestyle and infection process.

Keywords: Biological control; Cattle tick; Infection process; Metarhizium spp; Secondary metabolite biosynthetic gene clusters; Transcriptome analysis.

Fig. 1

Pseurotin-related compound BGC (MaNRPS-PKS2). a Phylogenetic analysis was performed using Maximum-likelihood and Bayesian methods, based on the pseurotin-related backbone gene and orthologous sequences exhibited by several fungi. The orthologous sequences were classified according to fungal lifestyle trait, represented by different colors. The Bayesian tree is displayed, and branch support values (bootstrap proportions and Bayesian posterior probability) are associated with nodes. The Bayesian inference ran for 9,997,000 generations. Species in bold in (a) were used for the cluster conservation analysis presented in (b). b Some genes from M. anisopliae MaNRPS-PKS2 BGC resembled the characterized pseurotin BGC from A. fumigatus (34–81 % identity) and putative BGCs from A. nomius (49–85 % identity), S. apiospermum (63–84 % identity) and P. solitum (59–81 % identity). The M. anisopliae Zn(II) 2-Cys(6) transcription factor resembles the embedded transcription factor found in A. fumigatus (34 % identity), and the putative transcription factor from A. nomius (49 % identity). Interestingly, S. apiospermum and P. solitum do not have orthologs for this transcription factor. Orthologous genes were assigned the same color; white boxes represent genes that were not predicted to be part of M. anisopliae cluster, and blue boxes represent the conserved Zn(II)2-Cys(6) transcription factor

Fig. 2

Conservation of supercluster regions in several species. a Comparison of the fumagillin/pseurotin supercluster region among M. anisopliae, A. fumigatus and T. ophioglossoides. The backbone gene from fumagillin (green) is absent in M. anisopliae and T. ophioglossoides, but some accessory genes are present and intertwined with the well-conserved pseurotin BGC (red). These accessory genes appear to participate in metabolite biosynthesis; therefore, the final product of this cluster was speculated to be a pseurotin-related compound. Upstream of the pseurotin-related BGC, the fumitremorgin cluster (yellow) is present in A. fumigatus, but absent in M. anisopliae, and there is a putative tropolone/citrinin-related BGC at this location in T. ophioglossoides. The tropolone/citrinin-related BGC has orthologous sequences (MaPKS1) in M. anisopliae, although they are located in a different genomic region. The MaPKS14 (light-blue) BGC is located downstream the pseurotin-related BGC only in M. anisopliae. b Comparison of a putative supercluster region in M. anisopliae, A. fumigatus, A. niger, and P. ipomoeae. Three BGCs (helvolic acid, MaPKS18, and MaNRPS-PKS6) were assigned to this M. anisopliae region. The helvolic acid (pink) and MaPKS18 (purple) clusters appear to be co-regulated. Additionally, both are conserved in A. fumigatus and P. ipomoeae, although the BGCs are distantly located in chromosome 4 in A. fumigatus. * This locus was inverted to fit in the figure

Fig. 3

Putative helvolic acid (MaTERP1) conservation and synteny. The MaTERP1 cluster from M. anisopliae resembled the characterized helvolic acid cluster from A. fumigatus (41–65 % identity), and putative BGCs from N. fischeri (41–64 % identity) and P. ipomoeae (80–90 % identity). Notably, the BGC found in P. ipomoeae exhibits a strong synteny with clusters from the Metarhizium genus (e.g., M. anisopliae and M. guizhouense). The locus tags for the four backbone genes are given

Fig. 4

Species and helvolic acid BGC (MaTERP1) phylogeny. a Supermatrix tree of nine genes (MANI_010495/ MANI_010512/ MANI_010527/ MANI_010530/ MANI_010531/ MANI_010532/ MANI_010536/ MANI_010537/ MANI_010594) involved in helvolic acid biosynthesis. This supermatrix tree resembles the generated supertree (Additional file 13). The orthologous sequences were classified according to fungal lifestyle trait, represented by different colors. The Bayesian tree is displayed, and branch support values (bootstrap proportions and Bayesian posterior probability) are associated with nodes. The Bayesian inference ran for 1,000,000 generations. The cluster tree was compared with the species tree presented in (b). Note that the helvolic acid BGC is present in few Eurotiales and Hypocreales species. b The phylogeny of tef1, a barcode gene, showing established species relationships. Branch support values (Bayesian posterior probability) are associated with nodes. The Bayesian inference ran for 43,000 generations

Fig. 5

Tropolone/citrinin-related compound BGC (MaPKS1). a Phylogenetic analysis was performed using Maximum-likelihood and Bayesian methods, based on the tropolone/citrinin-related backbone gene and orthologous sequences in several fungi. Additionally, two PKS outgroup sequences were added: cichorine (Aspergillus nidulans FGSC A4) and mycophenolic acid (Penicillium brevicompactum). The orthologous sequences were classified according to fungal lifestyle trait, represented by different colors. The Bayesian tree is displayed, and branch support values (bootstrap proportions and Bayesian posterior probability) are associated with nodes. The Bayesian inference ran for 120,000 generations. Species in bold in (a) also have their domain organization shown with abbreviations (KS: Keto-synthase; AT: Acyltransferase; ACP: Acyl carrier protein; MT: Methyltransferase O- or C-; TD: Thioester reductase), and were used for the cluster conservation analysis presented in b. These clusters have characterized or partially characterized biosynthetic routes. b Some genes from M. anisopliae MaPKS1 BGC resembled the characterized stipitatic acid (tropolone) BGC from T. stipitatus and the citrinin BGC from M. purpureus. These conserved genes are involved in the first steps of the biosynthesis of their compound, as described in c. Note that the mrl1 gene of the citrinin biosynthetic pathway is absent in M. anisopliae. Additionally, the gene MANI_112402 resembles the ctnA citrinin regulator from M. purpureus (59 % identity), and putative transcription factors from C. posadasii, T. stipitatus, M. purpureus and M. pilosus as demonstrated in (b). Orthologous genes were assigned the same color; white boxes represent genes that are not predicted to be part of M. anisopliae cluster; and blue boxes represent the conserved transcription factor