Fusion of two divergent fungal individuals led to the recent emergence of a unique widespread pathogen species

Abstract

In a genome alignment of five individuals of the ascomycete fungus Zymoseptoria pseudotritici, a close relative of the wheat pathogen Z. tritici (synonym Mycosphaerella graminicola), we observed peculiar diversity patterns. Long regions up to 100 kb without variation alternate with similarly long regions of high variability. The variable segments in the genome alignment are organized into two main haplotype groups that have diverged ∼3% from each other. The genome patterns in Z. pseudotritici are consistent with a hybrid speciation event resulting from a cross between two divergent haploid individuals. The resulting hybrids formed the new species without backcrossing to the parents. We observe no variation in 54% of the genome in the five individuals and estimate a complete loss of variation for at least 30% of the genome in the entire species. A strong population bottleneck following the hybridization event caused this loss of variation. Variable segments in the Z. pseudotritici genome exhibit the two haplotypes contributed by the parental individuals. From our previously estimated recombination map of Z. tritici and the size distribution of variable chromosome blocks untouched by recombination we estimate that the hybridization occurred ∼380 sexual generations ago. We show that the amount of lost variation is explained by genetic drift during the bottleneck and by natural selection, as evidenced by the correlation of presence/absence of variation with gene density and recombination rate. The successful spread of this unique reproductively isolated pathogen highlights the strong potential of hybridization in the emergence of pathogen species with sexual reproduction.

Fig. 1.

(A) Frequencies of 1-kb windows with 0, 1, 2, 3…. segregating sites in five isolates of Z. pseudotritici. The plot illustrates that there is not a gradual decline in the number of polymorphic sites, but rather windows with 0 segregating sites is a distinct category. (B) Genome-wide distribution of variable (red) and nonvariable (green) segments across 13 aligned chromosomes of five isolates of Z. pseudotritici. Windows where the frequencies of gaps exceed 50% (gray) were excluded from the analysis.

Fig. 2.

(A) Close-up of 400 aligned 1-kb windows on chromosome 1 (Fig. 1B) in five Z. pseudotritici isolates illustrating the alternating patterns of variable (red) and nonvariable (green) segments. The vertical axis gives the distance in each window between the two haplotypes defined as H1 and H2 (see below). This distance between H1 and H2 reflects, according to our model, the distance between the parents of the hybrid species. (B) The pattern of association of sequences in the variable segments of Z. pseudotritici can be illustrated as topologies. With five individuals, there are 15 possible topologies. In Z. pseudotritici the five sequences always segregate into two haplotypes either where one single sequence differs from the remaining four or where two identical sequences differ from the remaining three. The two “ends” of the topology trees represent the haplotypes H1 and H2. For each aligned window in Z. pseudotritici the specific topology was assigned (A).

Fig. 3.

Distribution of windows assigned to each of the 15 possible topologies. In total, 13,579 variable windows could be assigned to one of the topologies. For the topologies 1–5 one sequence differ from the other four and for the topologies 6–15 two sequences different from the other three (Fig. 2B).

Fig. 4.

(A) The genomic pattern in Z. pseudotritici is consistent with a speciation history where a common ancestor of the two hybrid parents diverged recently after the emergence of the Z. tritici lineages. Two haploid individuals from these diverged lineages formed viable hybrid offsprings through one sexual cross ∼500 y ago. (B) Initial fusion between the two individuals gave rise to a diploid zygote. Following meiotic recombination the genomes of the haploid individuals in the F₁ hybrid swarm are mosaics of the parental chromosomes. Some segments of the genomes may be inherited and fixed from only one of the two parents. These segments will remain nonvariable in the subsequent generations of the hybrid swarm, and they exist alongside variable segments contributed by both parental species. Through time, further crossover and random genetic drift will lead to the genomic mosaic observed today. In the extant population of Z. pseudotritici, variable windows show only two haplotypes, suggesting that the hybrid swarm did not backcross to the parental species, but instead propagated as a saltatory biological species. (C) Parental haplotypes (H1 and H2) are more closely related to each other (blue bars) than to Z. tritici (red bars) as shown by the two histograms of divergence estimates for all aligned windows. In the subplot the divergence between the two Z. pseudotritici haplotypes H1 and H2 are plotted against the net divergence between Z. tritici and the two haplotypes for windows on chromosome 1. The majority of the 7,076 compared segments are found below the line of identity, showing the differentiation between the two Z. pseudotritici haplotypes H1 and H2 to be consistently lower than the divergence to Z. tritici.

Fig. 5.

(A) Frequency of variable (red) and nonvariable (green) windows differs across the 13 core chromosomes. (B) Average lengths of the variable segments (red dots) and and nonvariable (green dots) are shown for the 13 core chromosomes. Variable and nonvariable segments are within the same size range, whereas the average lengths of the topology blocks (blue) are considerably smaller. (C) Average length of segments is correlated with the local recombination rate measured as ρ. Longer segments are located in regions of lower recombination rates and shorter segments in regions of high recombination rate.