Clonal reproduction of Moniliophthora roreri and the emergence of unique lineages with distinct genomes during range expansion

Abstract

The basidiomycete Moniliophthora roreri causes frosty pod rot of cacao (Theobroma cacao) in the western hemisphere. Moniliophthora roreri is considered asexual and haploid throughout its hemibiotrophic life cycle. To understand the processes driving genome modification, using long-read sequencing technology, we sequenced and assembled 5 high-quality M. roreri genomes out of a collection of 99 isolates collected throughout the pathogen's range. We obtained chromosome-scale assemblies composed of 11 scaffolds. We used short-read technology to sequence the genomes of 22 similarly chosen isolates. Alignments among the 5 reference assemblies revealed inversions, translocations, and duplications between and within scaffolds. Isolates at the front of the pathogens' expanding range tend to share lineage-specific structural variants, as confirmed by short-read sequencing. We identified, for the first time, 3 new mating type A locus alleles (5 in total) and 1 new potential mating type B locus allele (3 in total). Currently, only 2 mating type combinations, A1B1 and A2B2, are known to exist outside of Colombia. A systematic survey of the M. roreri transcriptome across 2 isolates identified an expanded candidate effector pool and provided evidence that effector candidate genes unique to the Moniliophthoras are preferentially expressed during the biotrophic phase of disease. Notably, M. roreri isolates in Costa Rica carry a chromosome segment duplication that has doubled the associated gene complement and includes secreted proteins and candidate effectors. Clonal reproduction of the haploid M. roreri genome has allowed lineages with unique genome structures and compositions to dominate as it expands its range, displaying a significant founder effect.

Keywords: Theobroma cacao; cacao frosty pod; genome evolution; mating type loci; pathogenomics.

Fig. 1.

Colinearity between M. roreri isolates syntenic groups. Colinear regions between the 5 isolates have been depicted across the 5 different M. roreri isolates evidencing distant translocation events (blue) and inversions (red). For each of the genotypes, the 3 tracks reported indicate a) the local repeat density in terms of percentage of bases associated to a repetitive element, b) the local gene density in terms of bases ascribed to gene loci, and c) the effector gene density in terms of number of genes in a 100 Kb window.

Fig. 2.

Phylogenetic trees relating the M. roreri isolates. Phylogenetic trees describing the relationships present between the M. roreri isolates (A and A) and M. perniciosa (B). Relationships identified with both methods have been highlighted with the same color. a) Tree built on the binary distance between isolates according to the presence of shared breakpoints of the SVs identified between genomes. b) Phylogenetic tree based on amino acid sequence conservation among the single-copy orthologous proteins.

Fig. 3.

Function of genes affected by SNPs and SVs. a) Distribution of the gene functions associated with the genes affected by SVs in long-read assemblies when comparing MrC26 with the other 4 isolates. Most of the genes are associated with transposable elements, no homolog in RefSeq database, or no biological function, often related to fungal pathogenicity (Pusztahelyi et al. 2015; Macheleidt et al. 2016). Expression and modulation in the RNA-Seq samples have been also reported, showing that many of the genes related to the secondary metabolism are active, with DE meaning significantly differentially expressed in at least 1 condition, EX meaning expressed in at least 1 condition even if not significantly modulated, and NO meaning not expressed. b) Secondary metabolite clusters found in MrC26 with antiSMASH. Expression and modulation in the RNA-Seq samples have been also reported as long as the presence of SVs in the locus.

Fig. 4.

SVs identified between syntenic groups 2 and group 10 in the different M. roreri isolates. Syntenic group 2 and syntenic group 10 show the largest SV identified among the M. roreri isolates assembled with long reads. a) The dotplots evidence the presence of a translocation event specific of MrB3 and a duplication event specific of MrC26 among the 5 long-read assemblies. b) and c) Raw read alignments confirm the translocation event in MrB3 and evidence that is shared only with the closely related isolate MrP5 using both MrC26 (b) and MrB3 (c) genome assemblies as reference. d) Detailed view of the translocation breakpoint and the effect on the gene content of MrC26 and MrB3.

Fig. 5.

Secreted protein in M. roreri. a) Secretome identification pipeline used for M. roreri. b) Number of secreted proteins, effectors, and effector clusters identified in the 22 M. roreri isolates, in the M. perniciosa Mp-MCCS1 isolate, and in the species used as outgroup (F. oxysporum forma specialis Lycopersici 4286, M. fiardii PR-910, R. solani AG-1 IA, and T. nigripes CBS291.85), c) Effector clusters shared across the different species; the effectors specific of M. roreri are evidenced in blue, the effector clusters shared between the Moniliophthorae in violet, and the effector clusters shared between M. roreri and at least 1 other Marasmiaceae species with or without M. perniciosa and not F. oxysporum or R. solani in red.

Fig. 6.

Phylogeny of mating type loci. a) Location and structure of mating type loci. Moniliophthora roreri mating type locus A genes (HD1 and HD2) are located in syntenic group 2, and locus B genes (STE3_Mr4 and Mr_Ph4) are located in syntenic group 5. b) Phylogenetic trees describing the relationships present between the mating type genes identified in the 22 M. roreri isolates. The HD1 and HD2 (A locus) and STE3_Mr4 and Mr_Ph4 (B locus) proteins previously described in Díaz-Valderrama and Aime (2016a) for A1B1 and A2B2 mating types are evidenced in red. A1, A2, and alleles for A1 locus were assigned using HD1 and HD2 gene sequence homology, and X3, X4, and X4 IDs were assigned to the newly identified alleles. B locus alleles B1 and B2 were assigned based on homology of STE3_Mr4 and Mr_Ph4 genes alleles, and the newly identified locus was denominated BX3.

Fig. 7.

Moniliophthora roreri mating type loci variability across 22 isolates. Comparison of mating type loci structure across different isolates in relation to the phylogeny calculated on the conservation of single-copy ortholog proteins. By aligning the raw Illumina reads of each isolate on the different A locus structures (a) and B locus structures (b) identified for the long-read assemblies, it is possible to identify the closest allele structure. The location of both genes of each mating type locus is highlighted in gray in the plot. B locus shows a lower amount of variation among the isolates and the presence of 3 different alleles, and A locus counts 5 alleles that do closely follow the relationships described by the single-copy orthologs.

Fig. 8.

Moniliophthora roreri mating type loci allele geographical distribution. Geographical distribution of the different mating type alleles. Boxes sharing the same color carry the same mating type alleles. Color coding has been used to distinguish the mating type A locus alleles (A1, A2, and the newly identified AX3, AX4, and AX5); icon shape was used to distinguish the B locus alleles (B1, B2, and the newly identified BX3).