Clonality, recombination, and hybridization in the plumbing-inhabiting human pathogen Fusarium keratoplasticum inferred from multilocus sequence typing

Abstract

Background: Recent work has shown that Fusarium species and genotypes most commonly associated with human infections, particularly of the cornea (mycotic keratitis), are the same as those most commonly isolated from plumbing systems. The species most dominant in plumbing biofilms is Fusarium keratoplasticum, a cosmopolitan fungus known almost exclusively from animal infections and biofilms. To better understand its diversity and population dynamics, we developed and utilized a nine-locus sequence-based typing system to make inferences about clonality, recombination, population structure, species boundaries and hybridization.

Results: High levels of genetic diversity and evidence for recombination and clonality were detected among 75 clinical and 156 environmental isolates of Fusarium keratoplasticum. The multilocus sequence typing system (MLST) resolved 111 unique nine-locus sequence types (STs). The single locus bifactorial determinants of mating compatibility (mating types MAT1-1 and MAT1-2), were found in a ratio of 70:30. All but one of the 49 isolates of the most common ST (FSSC 2d-2) came from human infections, mostly of the cornea, and from biofilms associated with contact lenses and plumbing surfaces. Significant levels of phylogenetic incongruence were found among loci. Putative clonal relationships among genotypes were estimated, showing a mixture of large clonal complexes and unrelated singletons. Discordance between the nuclear ribosomal rRNA and other gene genealogies is consistent with introgression of ribosomal RNA alleles of phylogenetic species FSSC 9 into F. keratoplasticum. No significant population subdivision based on clinical versus non-clinical sources was found.

Conclusions: Incongruent phylogenetic trees and the presence of both mating types within otherwise identical STs were observed, providing evidence for sexuality in F. keratoplasticum. Cryptic speciation suggested in a published three-locus MLST system was not supported with the addition of new loci, but evidence of introgression of ribosomal RNA genes from another strongly supported phylogenetic species (FSSC 9), also known from plumbing systems and human infections, was detected in two isolates. Overall, F. keratoplasticum is a diverse and geographically unstructured species with a mixed clonal and recombinant life history.

Figure 1

Plot of mean number of genotypes of F. keratoplasticum as a function of number of characters (SNPs and indels) sampled.

Figure 2

Population snapshots of F. keratoplasticum inferred using Phyloviz beta. Each node represents a unique ST (the prefix ″2-″ has been omitted for clarity); areas of nodes are proportional to the number of isolates. STs with 8/9 loci in common (SLVs) are connected with solid black lines, STs with 7/9 loci in common (DLVs) are connected with dashed lines, STs with 6/9 loci in common (TLVs) are connected with dotted lines. STs different at more than three loci not connected to a group. 9 SLV groups are shown. In A) light gray indicates proportion of STs that are MAT1-1; darker gray shading indicates proportion of STs that is MAT1-2. In B) light gray indicates isolates from clinical sources; darker gray indicates isolates from all other sources. Plus signs denote several STs containing isolates known to produce cyclosporin in vitro. Asterisks denote the ST of the ISO standard ATCC 36031 strain. White stars denote STs containing isolates successfully crossed in mating experiments.

Figure 3

Neighbor net of 111 unique nine-locus haplotypes inferred using SPLITSTREE4. A PHI test for recombination (7) implemented in SPLITSTREE4 suggests a statistically high likelihood of recombination in the full dataset (p < 0.0001).

Figure 4

Nine individual locus unrooted maximum parsimony trees inferred using parsimony ratchet (72, 73) implemented in PAUP* 4.0. Isolates S-2406 and S-2510 are strongly supported members of F. keratoplasticum (FSSC 2) in eight of the trees. In the rDNA tree, however, these two isolates are strongly supported members of FSSC 9.