Prevalence and diversity of TAL effector-like proteins in fungal endosymbiotic Mycetohabitans spp

Abstract

Endofungal Mycetohabitans (formerly Burkholderia) spp. rely on a type III secretion system to deliver mostly unidentified effector proteins when colonizing their host fungus, Rhizopus microsporus. The one known secreted effector family from Mycetohabitans consists of homologues of transcription activator-like (TAL) effectors, which are used by plant pathogenic Xanthomonas and Ralstonia spp. to activate host genes that promote disease. These 'Burkholderia TAL-like (Btl)' proteins bind corresponding specific DNA sequences in a predictable manner, but their genomic target(s) and impact on transcription in the fungus are unknown. Recent phenotyping of Btl mutants of two Mycetohabitans strains revealed that the single Btl in one Mycetohabitans endofungorum strain enhances fungal membrane stress tolerance, while others in a Mycetohabitans rhizoxinica strain promote bacterial colonization of the fungus. The phenotypic diversity underscores the need to assess the sequence diversity and, given that sequence diversity translates to DNA targeting specificity, the functional diversity of Btl proteins. Using a dual approach to maximize capture of Btl protein sequences for our analysis, we sequenced and assembled nine Mycetohabitans spp. genomes using long-read PacBio technology and also mined available short-read Illumina fungal-bacterial metagenomes. We show that btl genes are present across diverse Mycetohabitans strains from Mucoromycota fungal hosts yet vary in sequences and predicted DNA binding specificity. Phylogenetic analysis revealed distinct clades of Btl proteins and suggested that Mycetohabitans might contain more species than previously recognized. Within our data set, Btl proteins were more conserved across M. rhizoxinica strains than across M. endofungorum, but there was also evidence of greater overall strain diversity within the latter clade. Overall, the results suggest that Btl proteins contribute to bacterial-fungal symbioses in myriad ways.

Keywords: Mycetohabitans; Rhizopus; effectors; endofungal bacteria; long-read sequencing; meta-assembled genomes.

Fig. 1.. Genomic relationships, species and origins of the sequenced strains. (a) Output from the reference sequence alignment-based phylogeny builder (REALPHY) for the nine strains sequenced in this study and the M. rhizoxinica type strain B1 (denoted by *). B5 is the type strain for M. endofungorum. The geographical and substrate origin of the host isolate for each strain is listed when known. Burkholderia pseudomallei strain K96243 was used as an outgroup. Scale bar indicates substitutions per site. (b) Pangenome of the ten strains with indicated average nucleotide identity (ANI) heatmap. The blue, green and purple layers represent M. rhizoxinica, Mycetohabitans sp. B46 and M. endofungorum genomes respectively, where gene clusters are arranged based on synteny with B1. The outermost dark red layer highlights single-copy core gene (SCG) clusters found in all ten genomes. In the heatmap, blue dotted lines indicate clusters of strains with 94.6 % or greater ANI to one another.

Fig. 2.. Btl proteins do not exclusively cluster by species. Unrooted trees depicting (a) DisTAL and (b) FuncTAL phylogenetic analyses based on the amino acid sequence with RVDs removed and RVDs only, respectively, for the putative Btl proteins encoded in ten Mycetohabitans genomes. More information on these proteins can be found in Table S1. Clades present (with the same members) in both trees are highlighted by background shapes of the same colour. Scale bars correspond to DisTAL and FuncTAL scores.

Fig. 3.. Maximum likelihood phylogenetic analysis of 46 intact Btl protein sequences. Bootstrap values (as a percentage of 1000) are shown at the nodes for those with >50 % support. As shown in the key at right, coloured circles at branch ends indicate the source of the data for each sequence, and coloured boxes detail Mycetohabitans species, number of repeats, and presence or absence of key protein motifs, including the nuclear localization signal (NLS), shown using the single letter amino acid code. An asterisk indicates a missing amino acid where one would be expected in a typical conserved repeat, i.e. ‘N*’. A diagram of a Btl protein with coloured, corresponding motif locations is in the top left.

Fig. 4.. Alignments of Btl protein repeat variable diresidue (RVD) sequences of the first, central and final repeats by clade from Fig. 3. RVDs matching the consensus are in black font. RVDs that vary from the consensus are in bold pink font, or bold orange to indicate a second variant at that position. An asterisk (*) represents a missing amino acid at the respective position of the RVD. Coloured circles indicate species, following the key at bottom right.

Fig. 5.. Genomic relationships, species and btl gene content of Mycetohabitans strains examined in this study. (a) Output from the reference sequence alignment-based phylogeny builder (REALPHY) for the nine strains we sequenced, the metagenome bacterial assemblies and publicly available Mycetohabitans spp. genomes. Taxa are labelled by bacterial strain name (B#), or fungal accession number where a bacterial strain has not been named. Circles to the right of each taxon represent the btl gene content by clade (I–VI from left to right): white represents no gene, black represents one gene from that clade, or two in the case of strain B14, and grey represents an incompletely assembled gene that closely match that clade. Plus signs (+) indicate additional incompletely assembled fragments that were not assigned to a clade. Burkholderia pseudomallei strain K96243 was used as an outgroup. Scale bar indicates substitutions per site. (b) Average nucleotide identity analysis of the ten strains. Dotted lines indicate clusters of strains with 95 % or greater ANI to one another, indicative of species. The source of a strain’s sequence is indicated by coloured dots along the left axis. B1 is the type strain for M. rhizoxinica and B5 is the type strain for M. endofungorum.

Fig. 6.. Relationship of Btl proteins to TALEs, RipTALs and other TAL-like proteins. (a) Maximum likelihood tree based on amino acid sequences of the indicated proteins. Nodes with bootstrap values >50 % (out of 1000 replicates) are labelled. Scale bar indicates substitutions per site. Where available, identifiers from UniProt or GenBank are given in grey font next to the protein names. Branches are coloured and clades are labelled based on the genus from which the sequence was taken, except for the marine metagenome sequences, which are in orange. (b) DisTAL analysis of the same group of protein sequences presented as an unrooted tree. Line colouring matches that in (a). Proteins too divergent from Xanthomonas TAL effectors to be parsed correctly by DisTAL are absent from the tree.