As humans, we often think of genetic mutations as being bad. Over the past several decades we have seen health warnings issued on a variety of environmental exposures, from cigarettes to tanning beds, and with good reason because they cause mutations. For multicellular organisms like humans, these mutations are strongly associated with cancer. But in bacteria, this is not true. In fact, the rate at which mutations occur sometimes increases to help bacteria cope with stressful environments.

Unlike bacteria, humans are eukaryotes – the name given to organisms whose cells contain different compartments separated by membranes, such as the nucleus of the cell. For years, we have assumed that eukaryotic microbes, like fungi and parasites, act more like humans than like bacteria because work in budding yeast (another eukaryote) has suggested this to be the case. However, recent work in disease-causing fungi has shown that, much like bacteria, elevated mutation rates may help them to respond to stress. This could also enable fungi to become resistant to drugs used to treat fungal infections.

Cryptococcus deuterogattii is a fungus that causes human diseases including meningoencephalitis and a lung infection called pulmonary cryptococcosis. An ongoing outbreak of the fungus began in the Pacific Northwest of Canada in the late 1990s and emerged in the United States in 2006/2007. Among isolates closely related to those fungi causing the outbreak, three were found that appear to have a specific mutation in their DNA mismatch repair pathway, meaning that they may also experience a higher mutation rate. These strains are also less able to cause disease than others.

Billmyre et al. now demonstrate experimentally that all three isolates have a specific DNA mismatch repair defect, and show that these fungi experience elevated mutation rates, resulting in what is known as a hypermutator state. Furthermore, whole genome sequencing and phylogenetic analysis showed that these hypermutator strains are derived from the outbreak-causing fungi, and that their reduced ability to cause disease is likely a result of accumulating mutations and the loss of the ability to grow at the higher temperatures found in the human body.

Fungal infections are difficult to treat, in part because there are a limited number of available drugs. Elevated mutation rates will likely increase how often and how rapidly fungi develop resistance to these drugs. Understanding how commonly fungi exhibit a hypermutator state that could impact the development of drug resistance will therefore be important for treating patients with fungal infections, which account for millions of infections and hundreds of thousands of deaths annually worldwide.

Mutation is the raw material of evolution. As a result, all organisms must strike a balance between allowing enough random mutations for selection to act upon, and the fact that most of these mutations are likely to be deleterious and must be purged from the population. This is particularly true for pathogenic microbes that take part in Red Queen conflicts with their hosts and require a continuous supply of mutations to remain competitive. In this case, increased pressure may exist to maintain an elevated mutation rate to increase the rate of adaptation. Increases in mutation rate may also serve to accelerate adaptation when microbes are introduced into new environments or encounter novel stresses, such as antimicrobial therapy.

Adaptive variation in mutation rate is common in bacteria. One example is the Long Term Evolution Experiment (LTEE), where defects in DNA repair emerged and the resulting hypermutator phenotype swept the population in six out of twelve E. coli lines (Tenaillon et al., 2016). Mutator alleles likely emerge frequently but are typically purged from the population because individual mutations are more likely to be deleterious than adaptive; as a result hypermutators generally produce less fit offspring than non-hypermutators. However, occasionally a sufficiently beneficial mutation is potentiated by the presence of the hypermutator and the hypermutator allele is able to hitchhike to higher frequency within the population. This is more likely to occur in populations at local evolutionary minima, where many different large-effect adaptive mutations are possible. For example, hypermutator phenotypes are common in isolates of Pseudomonas aeruginosa growing within the lungs of patients with Cystic Fibrosis, where environmental conditions are constantly changing (Oliver et al., 2000). This changing environment causes an ongoing need to adapt, and increases the likelihood of hypermutators emerging by increasing the target size for adaptive variation. However, once beneficial mutations have fixed, and organisms have adapted to their new environment, antimutator suppressor alleles can emerge to reduce the mutation rate or defective mutator alleles can even be replaced by functional alleles through horizontal gene transfer (Wielgoss et al., 2013; Denamur et al., 2000). This balance is likely critical to evolution in microbes as they face diverse stresses and environmental changes, and as a result over 1% of natural bacterial isolates display hypermutator phenotypes (LeClerc et al., 1996). Even more extreme cases also exist, like in Mycoplasma pneumoniae and M. genitalium, where the mismatch repair machinery has been completely lost (Himmelreich et al., 1997).

In fact, many mutator phenotypes are the result of defects in mismatch repair, commonly referred to as MMR. In bacteria, MMR requires the MutHLS system, where MutS binds to mismatches and recruits MutL, which subsequently activates MutH to excise the mispair (Su and Modrich, 1986; Au et al., 1992; Grilley et al., 1989; Cox et al., 1972). In eukaryotes, both the MutS and MutL families have expanded, yielding MutS homolog (MSH) (Reenan and Kolodner, 1992) and MutL homolog (MLH) families (Strand et al., 1993). Heterodimers of various MSH family members participate in mismatch repair, but Msh2 in particular is a core component of the majority of mismatch repair pathways in yeast (Harfe and Jinks-Robertson, 2000). As a result, defects in MMR, and MSH2 mutations in particular, result in elevated rates of simple mismatches, but also dramatically elevated rates of repeat tract instability (Strand et al., 1993), which are associated with hereditary colon cancer in humans. MMR also plays a role in rejecting heteroduplexes of divergent sequences during both mitosis (Datta et al., 1997) and meiosis (Alani et al., 1994), meaning that loss of the complex can lower species boundaries and allow increased recombination between divergent chromosomes (Alani et al., 1994; Hunter et al., 1996; Rayssiguier et al., 1989).

In contrast to bacteria, substantially less is known about hypermutators in natural populations of eukaryotes. Evolutionary theory predicts that the presence of sex abrogates selection for hypermutator strains by eliminating genetic linkage between the mutator allele and the associated beneficial mutations it hitchhikes upon to high frequency (Tenaillon et al., 2000). Studies of allele incompatibility within the model fungus Saccharomyces cerevisiae have supported this traditional paradigm. Incompatible alleles of PMS1 and MLH1 exist within the population, such that a cross of two strains with wildtype mutation frequency can give rise to progeny with elevated mutation rates (Heck et al., 2006; Bui et al., 2015). However, the homozygous incompatible arrangement was not found among any of the original 65 strains examined, and was only identified within one clinical isolate of 1010 global diverse yeast isolates (Bui et al., 2017) and four clinical strains out of 93 sequenced in a separate study (Skelly et al., 2017). This frequency is far below that of the individual alleles in the population, and in each case observed suppressors had arisen to restore a wildtype mutation rate despite the incompatibility. In contrast, eukaryotic mutators can adapt and thrive in clonally expanding populations such as in human tumor environments (Bielas et al., 2006). Recent studies have begun to challenge this dogma in eukaryotic microorganisms as well, with over 50% of clinical C. glabrata isolates harboring loss of function mutations in MSH2, resulting in loss of mismatch repair function (Healey et al., 2016). Two groups also recently independently identified hypermutators in the fungal pathogen Cryptococcus neoformans. In the first case, long branch lengths were observed in genome sequencing of serial isolates from a recurrent infections that are correlated with nonsense mutations in MSH2, MSH5, and RAD5 Rhodes et al., 2017.In the second case, two of eleven clinical isolates exhibited elevated mutation rate that is linked to either an MSH2 nonsense mutation or multiple MSH2 missense mutations (Boyce et al., 2017). In previous work we also identified a clade of candidate hypermutators within the Pacific Northwest Cryptococcus deuterogattii outbreak that contain a coding single base deletion within the critical mismatch repair component MSH2 (Billmyre et al., 2014).

Cryptococcus deuterogattii, is a basidiomycete human fungal pathogen previously known as Cryptococcus gattii VGII and recently elevated to species level (Hagen et al., 2015), although this decision is disputed by a portion of the community (Kwon-Chung et al., 2017; Hagen et al., 2017). Unlike the other species of the Cryptococcus pathogenic species complex that infect immunocompromised hosts, C. deuterogattii is characterized by its ability to cause infection in otherwise healthy hosts (Springer et al., 2012) and by loss of the RNAi pathway (Feretzaki et al., 2016; D'Souza et al., 2011). C. deuterogattii is responsible for an ongoing outbreak in the Pacific Northwest region of the United States and Canada, which began in the late 1990's (Kidd et al., 2004). This outbreak was originally analyzed using Multi-locus sequence typing (MLST) and shown to be comprised of three clonal expansions, denoted VGIIa, VGIIb, and VGIIc (Fraser et al., 2005; Byrnes et al., 2010). These three subtypes differ in terms of virulence and total number of cases, with VGIIa responsible for the majority of infections. More recently, whole genome sequencing studies of these subpopulations showed that the clonal subtypes appeared to have different proximal geographic origins, and most interestingly, that VGIIa has three very closely related isolates, here referred to as VGIIa-like, that were considerably less virulent and differed in isolation location or date from the outbreak (Billmyre et al., 2014). We identified a candidate potential large effect mutation shared by these three diminished virulence isolates that resulted in a predicted non-functional Msh2. Here we show that these VGIIa-like isolates exhibit a bona fide hypermutator phenotype. Furthermore, we show that homopolymer runs are particularly unstable, and that these runs are common within the coding regions of genes in the Cryptococcus genome. We predict that this results in dramatic phenotypic diversity from inactivation and possibly also activation or reactivation of genes. Finally, we show that the mutant msh2 allele is not directly responsible for the decrease in virulence of the VGIIa-like isolates, nor does it appear to have directly played a role in the evolution of virulence in the VGIIa clonal outbreak. Rather, it appears to represent a parallel route of adaptation to a new environment in a successful lineage isolated from two continents over two decades and from both the environment and patient samples. This work suggests that hypermutator phenotypes are not limited to prokaryotes, but may represent a frequent avenue to evolutionary change and phenotypic diversity in eukaryotic microbes as well.

We previously identified a sublineage of strains highly related to the VGIIa outbreak in the Pacific Northwest denoted VGIIa-like, in part because they differed either in isolation location or time from the typical outbreak isolates (Billmyre et al., 2014). The VGIIa-like sublineage is composed of three isolates: NIH444, a clinical isolate from Seattle in 1975, well before the start of the outbreak; ICB107, a clinical isolate from Brazil in 1981; and CBS7750, an environmental isolate from California in 1990. All three isolates are characterized by diminished virulence relative to the VGIIa clonal outbreak strains (Fraser et al., 2005; Byrnes et al., 2010; Hagen et al., 2013). The analysis here includes additional genomes for a South American outgroup to both the VGIIa-like and VGIIa groups (Engelthaler et al., 2014)(Figure 1A) as well as alignment of these genomes to the improved C. deuterogattii genome assembly (Farrer et al., 2015). This allowed determination of the ancestral state of the multiple alleles differentiating the VGIIa and VGIIa-like groups. We also generated a second independent genome sequence for our original NIH444 isolate (NIH444(1) and NIH444(2)), as well as an NIH444 isolate acquired from a different lab (NIH444(v)) (Springer et al., 2010). We identified a total of 19 mutations that are shared by all of the VGIIa-like isolates sequenced: 13 SNPs and 6 INDELs (Figure 1—source data 1). Five SNPs are noncoding while 7 of the 8 remaining SNPs were nonsynonymous mutations. In addition, we confirmed that one of the INDELs in this branch is a single base deletion in a coding exon of MSH2 (26). This deletion is unique to the VGIIa-like group.

Figure 1

Download asset Open asset

Figure 1—source data 1

Mutations shared by VGIIa-like strains relative to VGIIa.

Mutations identified on the branch separating VGIIa-like from VGIIa strains with predicted effects and functional domain information about genes affected by the changes.

https://doi.org/10.7554/eLife.28802.004

Download elife-28802-fig1-data1-v1.docx

We confirmed both the presence of this mutation and showed that it results in a hypermutator phenotype. Sanger sequencing verified that this single base deletion is not an Illumina sequencing artifact (Figure 1B). This single base deletion occurs within an early exon of MSH2 and results in a nonsense mutation, with both an out of frame message and an early stop codon present in the predicted transcript (Figure 1C).

MSH2 defect results in hypermutator phenotype

Because Msh2 is a critical component of the mismatch repair complex, we next tested whether the VGIIa-like strains displayed an increased mutation rate. We utilized a standard fluctuation assay approach to determine the rate of 5-FOA resistance, as previously applied in Cryptococcus (Magditch et al., 2012). The resistance rate was increased in the hypermutator lineage between 4.4-fold at the minimum (CBS7750) and 6.6-fold at the maximum (ICB107) compared to the type strain VGIIa R265 (Figure 2A). The majority of 5-FOA resistance in Cryptococcus neoformans is acquired through mutation of URA5 (Kwon-Chung et al., 1992); the URA5 locus was PCR amplified and sequenced in independently derived 5-FOA resistant isolates to determine the genetic basis of resistance. Mutations were characterized as either substitutions, or indels of either single or multiple bases (Figure 2B). The majority of resistant isolates were the result of substitutions, and the mutation profile was similar between the hypermutator VGIIa-like strains and the non-hypermutator VGIIa strain EJB17 (EJB17 vs NIH444: p=1.00, EJB17 vs CBS7750 p=0.60, EJB17 vs ICB107 p=1.00, Fisher's Exact Test).

Figure 2

Download asset Open asset

MSH2 mutants accumulate indels in homopolymer runs

Next, the whole genome data was analyzed to ascertain whether the similarity in mutation spectrum observed in the URA5 locus was recapitulated at multiple loci or if the URA5 locus had unique properties. To do this, we focused on two relatively long branches within the C. deuterogattii phylogeny: the branch separating the outgroup VGIIb strain R272 from the VGIIa clonal cluster (57,373 variants) and the private ICB107 variants, which represented the longest available hypermutator branch (719 variants). The frequency of individual SNPs was examined first. While a high rate of transitions as compared to transversions was maintained in the hypermutator branch, there was a slight reduction in the frequency of A->G and T->C mutations relative to C->T and G->A mutations (Figure 3A). However, much more striking was a dramatic increase in indels within homopolymer runs (Figure 3B). While shifts in homopolymer runs in the R272 branch account for only 0.84% of the variants, in ICB107 they account for 45.3%, exceeding even the proportion of SNPs.

Figure 3

Download asset Open asset

By looking further at the context of the homopolymer run mutations within the ICB107 branch, it is apparent that longer base runs appear to be less stable than shorter runs, as approximately 24% of the mutations occur within the context of a nine base homopolymer run (Figure 3C). This peak is likely a function of both the decreased stability of longer runs and the low quantity of longer homopolymer runs in the genome. Many of these longer runs occur within intergenic regions in the genome, but a substantial portion of the coding genes in C. deuterogattii contain at least one longer coding homopolymer run (Figure 3C). The longest homopolymer run in the URA5 locus contains only four bases, which is predicted to be both relatively stable and among the coding genes with relatively shorter homopolymer runs (Figure 3C). In contrast, the FRR1 gene that encodes the FKBP12 homolog responsible for the antifungal action of both FK506 and rapamycin has a run of seven cytosines (7C) within its coding region. This places it both within the range of commonly mutated homopolymers in the ICB107 branch, and in the upper half of homopolymer containing genes in C. deuterogattii, suggesting that FRR1 is an appropriate locus to test the effect of mismatch repair mutants on homopolymers containing genes (Figure 3C).

Genes containing homopolymer runs are highly unstable in msh2 mutants

The FRR1 locus was utilized to further assess whether VGIIa-like strains were hypermutators at more than one locus and to test the hypothesis that homopolymer runs are particularly unstable in these isolates. The protein product of FRR1, FKBP12, binds to either FK506 or rapamycin to form a protein-drug complex that inhibits calcineurin or TOR, respectively. By selecting with both drugs at the same time at the non-permissive temperature of 37°C, loss of function mutations in FRR1 are selected as the only single step mutation conferring resistance to both drugs. Here we utilized a semi-quantitative assay to identify gross differences in mutation rate whereby independent colonies were grown in liquid culture and then swabbed onto quadrants of a selective plate. On FK506/rapamycin media, the wildtype strains produced a small number of resistant colonies (Figure 4A). In contrast, the NIH444 hypermutator strain produced prolific FK506/rapamycin resistant colonies. Surprisingly, the CBS7750 hypermutator was completely resistant to both drugs. Upon examining the FRR1 locus in the whole genome sequence data we discovered that there was already a single base deletion within the 7C homopolymer run in this strain. This may represent an unselected drug resistance phenotype potentiated by the msh2 defect. We attempted to use this assay for the ICB107 strain as well but were unsuccessful, likely as a result of its inability to grow at the higher growth temperature necessary for this assay that is based upon calcineurin-dependent growth at 37°C (Figure 7B).

Figure 4

Download asset Open asset

Next a standard fluctuation assay approach was employed to quantify the mutation rate difference between the VGIIa and VGIIa-like hypermutators. There was a greater than 120-fold increase in mutation rate to FK506/rapamycin resistance in NIH444 compared to R265, in contrast with the maximum of 6.6-fold increase observed for 5-FOA resistance (Figure 4C). As before, we selected independent resistant colonies, PCR amplified, and sequenced the FRR1 gene to determine the mechanism of resistance. All resistance was explained by mutations within FRR1 as expected, but while R265 still primarily acquired resistance through substitutions, NIH444 now almost exclusively underwent single base additions or single base deletions, all within the 7C homopolymer run (R265 vs NIH444, p=0.0003, Fisher's Exact Test). These indels resulted in nonsense mutations and resistance to both FK506 and rapamycin.

Hypermutation phenotype is linked to the msh2 defect

Once we established a bona fide hypermutator phenotype in the VGIIa-like strains, we sought to verify that this phenotype was linked to the msh2 del131 allele. We crossed the NIH444 strain to a very closely related R265a congenic strain recently generated by serial backcrossing (Zhu et al., 2013). C. deuterogattii matings are less fertile than C. neoformans, but we successfully dissected 14 viable spores from this cross. We typed these strains for the MSH2 allele by sequencing and observed 1:1 segregation of the MSH2 and the msh2 del131 alleles. The spores were then tested for rate of FK506/rapamycin resistance using the same semi-quantitative swabbing assay described above (Figure 5A). While five of the msh2 del131 strains displayed the predicted hypermutator phenotype, surprisingly two of the spores (#4 and 5) appeared wildtype despite inheritance of a defective copy of MSH2.

Figure 5 with 1 supplement see all

Download asset Open asset

Figure 5—source data 1

Aneuploidy in NIH444 x R265a cross.

Predicted aneuploid scaffolds for 14 spore progeny from an NIH444 by R265a cross based on read depths from whole genome sequencing. 8 of 14 demonstrate aneuploidy for at least one scaffold.

https://doi.org/10.7554/eLife.28802.009

Download elife-28802-fig5-data1-v1.docx

Mating in Cryptococcus has previously been demonstrated to produce phenotypic plasticity through the generation of aneuploid progeny at a high rate (Ni et al., 2013). These aneuploids often display defects in growth at high temperature; thus, we tested whether the high growth temperature of 37°C employed in our FK506/rapamycin resistance assay may have masked the ability of these meiotic progeny to develop drug resistance. Congruent with this hypothesis, both spore products #4 and #5 demonstrated growth defects at 37°C and 39°C (Figure 5C). Whole genome sequencing confirmed that both segregants carried an extra copy of one chromosome, suggesting that a 1N + 1 aneuploidy was causing temperature sensitivity (Figure 5B). In fact, all of the progeny from this mating had an unusually high rate of aneuploidy with 8 of the 14 progeny aneuploid for at least one scaffold/chromosome (Figure 5—source data 1). By passaging all of the spores on YPD at 37°C we were able to restore the ability to grow at high temperatures in all of the progeny after four to nine passages after which all no longer demonstrated a high temperature growth defect (Figure 5C). As a result the passaged derivatives of isolates #4 and #5 now demonstrated a hypermutator phenotype as expected (Figure 5D) and all 14 progeny produced the results predicted through linkage of the msh2 defect to the hypermutator phenotype (Figure 5—figure supplement 1). Further, whole genome typing of the 14 meiotic progeny demonstrated that sufficient recombination had occurred to establish linkage. All 14 progeny demonstrated unique SNP profiles and only one SNP demonstrated inheritance congruent with the hypermutator phenotype: a SNP on scaffold 3 that was linked to the msh2 del131 indel (~175 kb) (Figure 5E).

As a final verification of phenotypic linkage, we constructed two independent deletions of the MSH2 gene in the R265 VGIIa background via biolistic transformation and homologous recombination replacing the MSH2 ORF with a neomycin resistance cassette. A complete deletion of MSH2 resulted in the same hypermutator phenotype as the msh2 del131 allele, providing further evidence that the hypermutator phenotype was linked to the loss of function in MSH2 (Figure 5F). A fluctuation assay was used to determine the mutation rate for resistance to FK506 and rapamycin and demonstrated that the null mutants did not have an elevated mutation rate in comparison to the NIH444 strain (Figure 5G). This suggests that suppressors have not arisen on the NIH444 background to moderate the effects of the msh2 del131 allele.

Hypermutation enables phenotypic diversity including reversion of mutations

During growth of the msh2Δ::NEO strains a spontaneous ade2 mutant was fortuitously identified based on its classic red pigmentation (Figure 6A). We verified that this was the result of a defect in adenine biosynthesis as this strain grows on YNB supplemented with adenine, but not YNB media (Figure 6B). As expected, sequencing of the ADE2 locus revealed that the ade^- phenotype was linked to a single base mutation that results in a His->Arg amino acid change (Figure 6C).

Figure 6 with 1 supplement see all

Download asset Open asset

The red pigment produced by ade2 mutants is a toxic intermediate in adenine synthesis that accumulates in the vacuoles of mutant cells. As a result, these mutants have a growth defect, and suppressor mutations can readily be isolated that eliminate production of this toxic intermediate and now produce white colonies (Figure 6C). We isolated two red and two white derivative colonies and sequenced the ADE2 locus. While the red colonies retained the causative substitution, one of the white colonies retained it and the other had undergone reversion at this site back to the functional nucleotide (Figure 6C). We confirmed that the revertant isolate was no longer an adenine auxotroph and that the second white isolate was still auxotrophic, despite the lack of pigmentation (Figure 6D). This is likely the result of a second mutation upstream in the adenine biosynthetic pathway, resulting in further inactivation and loss of the ability to produce the toxic red intermediate. We tested a number of independent white reverted colonies for the ability to grow on media lacking adenine and observed that direct reversion to functional adenine biosynthesis was more common than additional inactivating mutations, with 37/42 (88%) white colonies demonstrating adenine prototrophy (Figure 6E). Taken together, these results suggest that hypermutation allows both frequent inactivation of pathways, and also the means to either adapt to the consequences or to simply directly revert the original mutation. In the context of antifungal drug action or activity, this phenotypic switching could be highly important.

Hypermutation does not directly affect virulence

We previously demonstrated that the VGIIa-like mutants exhibit diminished virulence in a nasal instillation model of virulence (Fraser et al., 2005; Byrnes et al., 2010). Here, we tested whether loss of Msh2 function was directly responsible for the decrease in virulence but observed no difference in virulence between R265 and two independent de novo deletions of msh2 (Figure 7A). This suggested that the defects in murine virulence in the VGIIa-like isolates could be the result of multiple independent deleterious mutations unique to each lineage. We next tested the ability to grow at higher temperature, a critical virulence factor, in the VGIIa-like isolates. All three isolates showed defects in high temperature growth, varying from minor defects in NIH444, the isolate with the shortest branch to R265, to moderate defects in CBS7750, the isolate with an intermediate length branch, and finally severe defects in ICB107, the isolate with the longest branch affected by the hypermutator (Figure 7B). This suggests that instead of an immediate change in virulence, defects in mismatch repair may instead result in loss of virulence over time as mutations accumulate in critical pathways. Thus, rather than the virulence difference being explained by a single change or set of changes shared by all three VGIIa-like strains, instead they may all be avirulent for their own unique reasons.

Figure 7

Download asset Open asset

Hypermutator lineage is derived and hypermutation did not cause the VGIIa outbreak

Historically, C. deuterogattii has been thought of as a tropical and subtropical pathogen. As a result, the origin of the VGIIa outbreak in the Pacific Northwest, a non-tropical environment, was surprising. In addition, (1) the South American origin of the ICB107 strain (within the proposed cradle of the C. deuterogattii species [Hagen et al., 2013]), (2) the isolation of NIH444 in Seattle near the outbreak origin, and (3) the diminished virulence of the VGIIa-like subclade (Fraser et al., 2005; Byrnes et al., 2010), all combined to suggest that the VGIIa-like group might have been the immediate precursors to the clonal VGIIa outbreak cluster (Billmyre et al., 2014; Engelthaler et al., 2014). Identification of the hypermutator phenotype further suggested that the defect in Msh2, carried by the older VGIIa-like group, may have played a role in adaptation to the climate of the Pacific Northwest and may have even potentiated the increase in virulence. To test this, a maximum parsimony phylogeny of the VGIIa-like and VGIIa strains was constructed, including additional sequences for a related South American VGII clade not included in our previous analysis (Engelthaler et al., 2014) and using the VGIIb R272 strain as an outgroup (Figure 8). As described above, the msh2 mutation can allow restoration of function mutations at the exact mutation site that are indistinguishable from the original sequence (Figure 6). Likewise, mating could also reintroduce a functional MSH2 allele. For these reasons, we tested for the possible presence and impact of the past msh2 del131 mutator allele throughout the phylogeny by examining the imputed mutation spectrum.

Figure 8

Download asset Open asset

As discussed above (Figure 3), branches with defects in mismatch repair show an increased frequency of shifts in homopolymer runs. Increases in homopolymer run shifts were observed only on the most proximal branch ancestral to the VGIIa-like mutator lineage. This suggests that the mutator phenotype is congruent with the presence of the allele throughout the VGIIa-like phylogeny, and that the VGIIa group did not experience a transient period of msh2 mediated hypermutation, followed by repair or mating-mediated replacement. Instead, the VGIIa-like lineage may represent a unique pathway to adaptation distinct from that followed by the VGIIa group that resulted in diminished rather than enhanced virulence.

In this study we have identified and characterized a successful lineage of eukaryotic hypermutators. Elevated mutations rates are a common adaptive mechanism in bacteria, but are typically thought of as transient states that allow beneficial mutations in the short term but are selected against in the long term (Wielgoss et al., 2013). Bacteria solve this problem through horizontal gene transfer of genes from the mismatch repair pathway, allowing the initial beneficial mutation to be separated from the deleterious mutator allele (Denamur et al., 2000). In contrast, few cases have been identified in natural isolates of eukaryotes, suggesting that variation in mutation rate may play a less substantial role than in bacteria. Recent studies have identified exceptions to this rule. In S. cerevisiae, incompatible alleles of PMS1 and MLH1 result in elevated mutation rates if present in combination (Heck et al., 2006; Bui et al., 2015). However, in the rare cases where the incompatible arrangement is found in nature, additional suppressors have arisen to restore wildtype mutation rate, suggesting that hypermutator phenotypes may not be tolerated over long periods of time (Bui et al., 2017; Skelly et al., 2017). In addition, a recent study of Candida glabrata has demonstrated that a substantial proportion of clinical isolates (>50%) carry nonsynonymous mutations in the MSH2 gene, some causing elevated mutation rates similar to the null phenotype and others with intermediate changes in mutation rate (Healey et al., 2016). The authors correlate the mismatch repair defects with multi-drug resistance. A second study with a different sample cohort confirmed the presence of MSH2 mutations, but concluded that drug resistance was better correlated with drug exposure than with mismatch repair defects (Dellière et al., 2016).

In contrast with the studies in S. cerevisiae and C. glabrata, here we identified a group of viable hypermutator strains resulting from a nonsense mutation in MSH2. These strains were isolated over a period of fifteen years from two different continents and include both clinical and environmental strains. Increases in mutation rate were similar, although slightly lower in magnitude, to those observed previously in the model yeast S. cerevisiae, where msh2 mutations increase resistance rate to Canavanine by 40 fold compared to 6.8 fold for 5-FOA here, or reversion of a 7xT run in a HOM3 mutant by 662 fold compared to 120 fold for 7xC in FRR1 here (Marsischky et al., 1996). The mutation spectrum was also similar, as homopolymer run indels were responsible almost exclusively for mutations within run-containing genes in msh2 mutants, but not in wildtype strains as previously reported in S. cerevisiae (Tran et al., 1997). The msh2 del131 allele carried by these strains appears to result in a complete loss of function, rather than simply reducing the efficiency of mismatch repair. This is distinct from the mutators identified in C. glabrata, and suggests that the VGIIa-like lineage is a successful and relatively long-lived hypermutator lineage, capable of both disseminating over a large area and persisting in the environment. Cryptococcus deuterogattii may also represent an intermediate between mutator-rich C. glabrata and mutator-poor S. cerevisiae. C. deuterogatttii is not an obligate pathogen and can cycle between an environmental lifestyle and an infectious lifestyle as an ‘accidental pathogen’ (Springer et al., 2012). Elevated mutation rates may be selected for by Red Queen interactions with a host, meaning that pathogens like C. glabrata that grow only in association with their hosts would experience higher mutation rates than facultative pathogens like C. deuterogattii, or effectively nonpathogenic yeasts like S. cerevisiae. In fact, a putative recurrent infection of Cryptococcus neoformans was recently demonstrated to contain nonsense mutations in MSH2, MSH5, and RAD5, predicted to result in a mutator phenotype (Rhodes et al., 2017).

In addition, a second recent study also identified two other hypermutator strains from clinical isolates of C. neoformans as well (Boyce et al., 2017). Boyce et al. went on to examine changes in mutation rate and spectrum with similar results to those presented here, although with a much larger induction of mutation rate in URA5 than reported here (around 200 fold vs 6.8 fold). This could reflect differences between DNA repair in C. neoformans and C. deuterogattii. They also observed changes in stability of homopolymer runs, similar to those described here but in a second homopolymer-containing locus. Interestingly, this study also observed loss of ability to grow at high temperatures in one of three replicates of msh2 mutants in an experimental evolution assay, like that observed here in the VGIIa-like hypermutator clade. Notable as well is that hypermutators in C. neoformans have thus far been observed only in clinical isolates, even in studies that included matched environmental sample populations like Boyce et al. This provides further evidence that microbial lifestyle may be affecting the frequency of mutator strains, with pathogens being more likely to evolve elevated mutation rate, perhaps in response to drug or host challenges.

An alternate hypothesis is that ploidy may play a role in the success of mutator strains. C. deuterogattii, C. neoformans, and C. glabrata exist primarily as haploids in nature, while S. cerevisiae is predominantly a diploid. Diploid strains are buffered from the effects of loss of function mutations like those observed in homopolymer runs in this study, which may reduce the supply of large effect beneficial mutations for mutator alleles to hitchhike upon in diploids. However, past work in S. cerevisiae suggests that diploids adapt more quickly in msh2 mutant backgrounds, rather than less, suggesting that ploidy may play the opposite role, at least in S. cerevisiae (Thompson et al., 2006).

Evolutionary theory predicts that the admixture provided by sex in a population nullifies the ability of mutator alleles to hitchhike to high frequency (Tenaillon et al., 2000). Unlike obligately sexual animals, or asexual bacteria, fungi can reproduce both sexually and asexually, with the frequency of sex varying substantially between different species. We previously described C. deuterogattii as a species characterized by long periods of mitotic clonal expansion and only intermittent sexual crosses (Billmyre et al., 2014). A contibuting factor is likely the highly biased mating type distribution in C. deuterogattii, as the vast majority of C. deuterogattii isolates are MATα, with only a handful of MATa isolates described globally (Byrnes et al., 2010). In Cryptococcus deneoformans, which shares a similar biased mating type distribution, this has resulted in the development of a unisexual α-α sexual cycle, dispensing with the obligate need for a MATa partner (Lin et al., 2005). This unisexual cycle can result in both de novo variation, but also recombination and admixture at similar levels to that observed in typical bisexual crosses (Ni et al., 2013; Lin et al., 2005; Sun et al., 2014). However, no laboratory unisexual cycle has been observed in C. deuterogattii to this point. Consequently, as strains are introduced to new locales, they may need to survive and adapt via mitotic growth without sexual crosses for long periods of time. This could elevate linkage between mutator alleles and the beneficial mutations they elicit, but also eliminate the ability to separate those beneficial mutations from additional deleterious alleles.

We also showed that hypermutators can allow both inactivating mutations in genes, and reversion of those mutations to the wildtype. Mismatch repair-defective mutator strains are characterized by particularly high rates of slippage within homopolymer runs, and these occur at run lengths that are common within fungal genomes. This may indicate that eukaryotes may harbor contingency loci featuring homopolymer runs, like those observed in bacteria (Moxon et al., 2006). Fungal contingency loci could be critical for responses to antifungal agents, both in the environment and the clinical setting. In addition, variation could be important in host-pathogen interactions as well. For example, in S. cerevisiae, tandem repeats are enriched within cell surface genes, which could enable alteration of antigenic diversity during pathogenic interactions (Verstrepen et al., 2005). While these loci are unstable even in wildtype populations, defects in the MMR pathway could enhance this instability and result in increased diversity. In C. neoformans, a mutator phenotype has been previously described in a lab-passaged strain notable for frequent mutations in the RAM pathway, which enables a dimorphic transition between pseudohyphal and yeast phase growth, although the molecular basis of the elevated mutation rate is unknown (Magditch et al., 2012). This transition is highly important for survival in the face of environmental threats such as amoeba, but compromises pathways responsible for high temperature growth, suggesting that oscillation between RAM+ and RAM- states may allow populations to survive as conditions change.

In addition, we provided evidence that the VGIIa-like lineage was not a subvirulent progenitor of the Pacific Northwest outbreak, in contrast with previous hypotheses from two groups (Billmyre et al., 2014; Engelthaler et al., 2014). Rather, the VGIIa-like lineage is derived and may represent either an alternative pathway of adaptation or potentially an accelerated stage of the evolutionary trajectory of the outbreak. The ability to grow at high temperatures is a critical component of virulence for fungal pathogens. However, many pathogenic fungi are thought to be ‘accidental’ pathogens where virulence factors are selected for by other environmental factors. In this model, thermotolerance could be selected by high ambient temperatures in a non-host environment (Casadevall et al., 2003). Moving from a tropical/sub-tropical environment in South America to the more temperate climate of the Pacific Northwest may relieve this high temperature selection. Consequently, the loss of ability to grow under thermal stress observed in the VGIIa-like strains may be adaptive for the majority of their growth conditions in the environment, but also render them relatively avirulent in mammals. This could suggest that as the VGIIa outbreak strains continue to adapt to the Pacific Northwest, they will gradually lose virulence potential, like the VGIIa-like lineage, making the outbreak self-limiting over time. Alternatively, the defects could simply reflect a decrease in viability caused by a long period of growth with an elevated mutation rate, resulting in mutational meltdown. The most severe temperature defect was observed in ICB107, a strain isolated from a patient in Brazil and also the strain with the largest number of mutations separating it from the VGIIa group, which would support the second hypothesis. These strains may have been able to persist in the Pacific Northwest in spite of the high mutation load because of a bottleneck effect during their introduction.

Our original hypothesis was that the common variants that differentiated the VGIIa group from the progenitor of the VGIIa-like mutator group would explain the decline in virulence observed in the VGIIa-like strains (Billmyre et al., 2014; Byrnes et al., 2010). However, the msh2 mutation, the only predicted large effect mutation, has no obvious immediate effect on virulence in a murine inhalation model. This is interesting, at least in part, because msh2 mutants of C. neoformans were previously reported to exhibit a modest growth advantage in mouse lungs in pooled signature tagged mutant experiments (Liu et al., 2008). This suggests that the mutator strains are better able to adapt to growth in the mouse lung, but not in a way that correlates directly with virulence in a dissemination and survival model of virulence. Boyce et al. obtained similar results in C. neoformans to those reported here, with msh2Δ and mlh1Δ mutants appearing neutral in a murine infection model, and pms1Δ mutants even relatively less virulent (Boyce et al., 2017). In addition, our results suggest that loss of virulence in the VGIIa-like strains may be a consequence of independent private mutations, i.e. each VGIIa-like strain is avirulent for its own reason. In previous work, CBS7750 and ICB107 have the most substantial virulence defects, while NIH444 has a more modest defect (Byrnes et al., 2010). This is congruent with the fact that NIH444 has the shortest divergence from the VGIIa clonal group, while CBS7750 and ICB107 have much longer branches, and is also congruent with the high temperature growth defects we observed.

Finally, we demonstrated that mutation spectrum analysis could be utilized within a phylogeny to infer the presence of a hypermutator allele through an increase in the proportion of indels within homopolymer runs. In bacteria, the hypermutator state is often transient and wildtype mismatch repair is often restored via horizontal gene transfer (Denamur et al., 2000). This paradigm could also operate in fungi, with sexual recombination replacing direct DNA transfer. However, these events are difficult to detect if the hypermutator allele is purged from the population by selection after it is separated from its linked beneficial allele. The widespread availability of whole genome sequencing could now allow changes in mutation spectrum to be used to detect episodic hypermutation throughout entire populations of microbes. We suspect that these episodes are common throughout the evolution of eukaryotic microbes and may be even more common among pathogenic microbes, reflecting their natural history as well as the result of Red Queen host-pathogen conflicts.

1. 1. Billmyre RB
   2. Clancey SA
   3. Heitman J

   (2017) Cryptococcus gattii VGII Genome sequencing

   Publicly available at the NCBI BioProject database (accession no: PRJNA387047).

   https://www.ncbi.nlm.nih.gov/bioproject/PRJNA387047/

1. 1. Billmyre RB
   2. Croll D
   3. Li W
   4. Mieczkowski P
   5. Carter DA
   6. Cuomo CA
   7. Kronstad JW
   8. Heitman J

   (2014) Cryptococcus gattii VGIII

   Publicly available at the NCBI BioProject database (accession no. PRJNA254979).

   https://www.ncbi.nlm.nih.gov/bioproject/PRJNA254979
2. 1. Engelthaler DM
   2. Hicks ND
   3. Gillece JD
   4. Roe CC
   5. Schupp JM
   6. Driebe EM
   7. Gilgado F
   8. Carriconde F
   9. Trilles L
   10. Firactive C
   11. Ngamskulrungroj P
   12. Castaneda E
   13. Lazera MDS
   14. Melhem MSC
   15. Perez- Bercoff A
   16. Huttley G
   17. Sorrell TC
   18. Voelz K
   19. May RC
   20. Fisher MC
   21. Thompson GR
   22. Lockhart SR
   23. Keim P
   24. Meyer W

   (2014) Cryptococcus gattii VGIII

   Publicly available at the NCBI BioProject database (accession no. PRJNA244927).

   https://www.ncbi.nlm.nih.gov/bioproject/?term=PRJNA244927

1. 1. Alani E
   2. Reenan RA
   3. Kolodner RD

   (1994)

   Interaction between mismatch repair and genetic recombination in Saccharomyces cerevisiae

   Genetics 137:19–39.

   • PubMed
   • Google Scholar
2. 1. Au KG
   2. Welsh K
   3. Modrich P

   (1992)

   Initiation of methyl-directed mismatch repair

   The Journal of Biological Chemistry 267:12142–12148.

   • PubMed
   • Google Scholar
3. 1. Bielas JH
   2. Loeb KR
   3. Rubin BP
   4. True LD
   5. Loeb LA

   (2006) Human cancers express a mutator phenotype

   PNAS 103:18238–18242.

   https://doi.org/10.1073/pnas.0607057103

   • PubMed
   • Google Scholar
4. 1. Billmyre RB
   2. Croll D
   3. Li W
   4. Mieczkowski P
   5. Carter DA
   6. Cuomo CA
   7. Kronstad JW
   8. Heitman J

   (2014) Highly recombinant VGII Cryptococcus gattii population develops clonal outbreak clusters through both sexual macroevolution and asexual microevolution

   mBio 5:e01494-14.

   https://doi.org/10.1128/mBio.01494-14

   • PubMed
   • Google Scholar
5. 1. Boyce KJ
   2. Wang Y
   3. Verma S
   4. Shakya VPS
   5. Xue C
   6. Idnurm A

   (2017) Mismatch Repair of DNA Replication Errors Contributes to Microevolution in the Pathogenic Fungus Cryptococcus neoformans

   mBio 8:e00595-17.

   https://doi.org/10.1128/mBio.00595-17

   • PubMed
   • Google Scholar
6. 1. Bui DT
   2. Dine E
   3. Anderson JB
   4. Aquadro CF
   5. Alani EE

   (2015) A genetic incompatibility accelerates adaptation in yeast

   PLOS Genetics 11:e1005407.

   https://doi.org/10.1371/journal.pgen.1005407

   • PubMed
   • Google Scholar
7. 1. Bui DT
   2. Friedrich A
   3. Al-Sweel N
   4. Liti G
   5. Schacherer J
   6. Aquadro CF
   7. Alani E

   (2017) Mismatch repair incompatibilities in diverse yeast populations

   Genetics 205:1459–1471.

   https://doi.org/10.1534/genetics.116.199513

   • PubMed
   • Google Scholar
8. 1. Byrnes EJ
   2. Li W
   3. Lewit Y
   4. Ma H
   5. Voelz K
   6. Ren P
   7. Carter DA
   8. Chaturvedi V
   9. Bildfell RJ
   10. May RC
   11. Heitman J

   (2010) Emergence and pathogenicity of highly virulent Cryptococcus gattii genotypes in the northwest United States

   PLoS Pathogens 6:e1000850.

   https://doi.org/10.1371/journal.ppat.1000850

   • PubMed
   • Google Scholar
9. 1. Casadevall A
   2. Steenbergen JN
   3. Nosanchuk JD

   (2003) 'Ready made' virulence and 'dual use' virulence factors in pathogenic environmental fungi--the Cryptococcus neoformans paradigm

   Current Opinion in Microbiology 6:332–337.

   https://doi.org/10.1016/S1369-5274(03)00082-1

   • PubMed
   • Google Scholar
10. 1. Cingolani P
    2. Platts A
    3. Wang leL
    4. Coon M
    5. Nguyen T
    6. Wang L
    7. Land SJ
    8. Lu X
    9. Ruden DM

    (2012) A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3

    Fly 6:80–92.

    https://doi.org/10.4161/fly.19695

    • PubMed
    • Google Scholar
11. 1. Cox EC
    2. Degnen GE
    3. Scheppe ML

    (1972)

    Mutator gene studies in Escherichia coli: the mutS gene

    Genetics 72:551 LP–567.

    • PubMed
    • Google Scholar
12. 1. Danecek P
    2. Auton A
    3. Abecasis G
    4. Albers CA
    5. Banks E
    6. DePristo MA
    7. Handsaker RE
    8. Lunter G
    9. Marth GT
    10. Sherry ST
    11. McVean G
    12. Durbin R
    13. 1000 Genomes Project Analysis Group

    (2011) The variant call format and VCFtools

    Bioinformatics 27:2156–2158.

    https://doi.org/10.1093/bioinformatics/btr330

    • PubMed
    • Google Scholar
13. 1. Datta A
    2. Hendrix M
    3. Lipsitch M
    4. Jinks-Robertson S

    (1997) Dual roles for DNA sequence identity and the mismatch repair system in the regulation of mitotic crossing-over in yeast

    PNAS 94:9757–9762.

    https://doi.org/10.1073/pnas.94.18.9757

    • PubMed
    • Google Scholar
14. 1. Davidson RC
    2. Cruz MC
    3. Sia RA
    4. Allen B
    5. Alspaugh JA
    6. Heitman J

    (2000) Gene disruption by biolistic transformation in serotype D strains of Cryptococcus neoformans

    Fungal Genetics and Biology 29:38–48.

    https://doi.org/10.1006/fgbi.1999.1180

    • PubMed
    • Google Scholar
15. 1. Dellière S
    2. Healey K
    3. Gits-Muselli M
    4. Carrara B
    5. Barbaro A
    6. Guigue N
    7. Lecefel C
    8. Touratier S
    9. Desnos-Ollivier M
    10. Perlin DS
    11. Bretagne S
    12. Alanio A

    (2016) Fluconazole and echinocandin resistance of Candida glabrata correlates better with antifungal drug exposure rather than with msh2 mutator genotype in a french cohort of patients harboring low rates of resistance

    Frontiers in Microbiology 7:2038.

    https://doi.org/10.3389/fmicb.2016.02038

    • PubMed
    • Google Scholar
16. 1. Denamur E
    2. Lecointre G
    3. Darlu P
    4. Tenaillon O
    5. Acquaviva C
    6. Sayada C
    7. Sunjevaric I
    8. Rothstein R
    9. Elion J
    10. Taddei F
    11. Radman M
    12. Matic I

    (2000) Evolutionary implications of the frequent horizontal transfer of mismatch repair genes

    Cell 103:711–721.

    https://doi.org/10.1016/S0092-8674(00)00175-6

    • PubMed
    • Google Scholar
17. 1. D'Souza CA
    2. Kronstad JW
    3. Taylor G
    4. Warren R
    5. Yuen M
    6. Hu G
    7. Jung WH
    8. Sham A
    9. Kidd SE
    10. Tangen K
    11. Lee N
    12. Zeilmaker T
    13. Sawkins J
    14. McVicker G
    15. Shah S
    16. Gnerre S
    17. Griggs A
    18. Zeng Q
    19. Bartlett K
    20. Li W
    21. Wang X
    22. Heitman J
    23. Stajich JE
    24. Fraser JA
    25. Meyer W
    26. Carter D
    27. Schein J
    28. Krzywinski M
    29. Kwon-Chung KJ
    30. Varma A
    31. Wang J
    32. Brunham R
    33. Fyfe M
    34. Ouellette BF
    35. Siddiqui A
    36. Marra M
    37. Jones S
    38. Holt R
    39. Birren BW
    40. Galagan JE
    41. Cuomo CA

    (2011) Genome variation in Cryptococcus gattii, an emerging pathogen of immunocompetent hosts

    mBio 2:e00342–10.

    https://doi.org/10.1128/mBio.00342-10

    • PubMed
    • Google Scholar
18. 1. Engelthaler DM
    2. Hicks ND
    3. Gillece JD
    4. Roe CC
    5. Schupp JM
    6. Driebe EM
    7. Gilgado F
    8. Carriconde F
    9. Trilles L
    10. Firacative C
    11. Ngamskulrungroj P
    12. Castañeda E
    13. Lazera MS
    14. Melhem MS
    15. Pérez-Bercoff A
    16. Huttley G
    17. Sorrell TC
    18. Voelz K
    19. May RC
    20. Fisher MC
    21. Thompson GR
    22. Lockhart SR
    23. Keim P
    24. Meyer W

    (2014) Cryptococcus gattii in North American Pacific Northwest: whole-population genome analysis provides insights into species evolution and dispersal

    mBio 5:e01464-14.

    https://doi.org/10.1128/mBio.01464-14

    • PubMed
    • Google Scholar
19. 1. Farrer RA
    2. Desjardins CA
    3. Sakthikumar S
    4. Gujja S
    5. Saif S
    6. Zeng Q
    7. Chen Y
    8. Voelz K
    9. Heitman J
    10. May RC
    11. Fisher MC
    12. Cuomo CA

    (2015) Genome evolution and innovation across the four major lineages of Cryptococcus gattii

    mBio 6:e00868-15.

    https://doi.org/10.1128/mBio.00868-15

    • PubMed
    • Google Scholar
20. 1. Feretzaki M
    2. Billmyre RB
    3. Clancey SA
    4. Wang X
    5. Heitman J

    (2016) Gene network polymorphism illuminates loss and retention of novel rnai silencing components in the Cryptococcus pathogenic species complex

    PLOS Genetics 12:e1005868.

    https://doi.org/10.1371/journal.pgen.1005868

    • PubMed
    • Google Scholar
21. 1. Fraser JA
    2. Giles SS
    3. Wenink EC
    4. Geunes-Boyer SG
    5. Wright JR
    6. Diezmann S
    7. Allen A
    8. Stajich JE
    9. Dietrich FS
    10. Perfect JR
    11. Heitman J

    (2005) Same-sex mating and the origin of the Vancouver Island Cryptococcus gattii outbreak

    Nature 437:1360–1364.

    https://doi.org/10.1038/nature04220

    • PubMed
    • Google Scholar
22. 1. Grilley M
    2. Welsh KM
    3. Su SS
    4. Modrich P

    (1989)

    Isolation and characterization of the Escherichia coli mutL gene product

    The Journal of Biological Chemistry 264:1000–1004.

    • PubMed
    • Google Scholar
23. 1. Hagen F
    2. Ceresini PC
    3. Polacheck I
    4. Ma H
    5. van Nieuwerburgh F
    6. Gabaldón T
    7. Kagan S
    8. Pursall ER
    9. Hoogveld HL
    10. van Iersel LJ
    11. Klau GW
    12. Kelk SM
    13. Stougie L
    14. Bartlett KH
    15. Voelz K
    16. Pryszcz LP
    17. Castañeda E
    18. Lazera M
    19. Meyer W
    20. Deforce D
    21. Meis JF
    22. May RC
    23. Klaassen CH
    24. Boekhout T

    (2013) Ancient dispersal of the human fungal pathogen Cryptococcus gattii from the Amazon rainforest

    PLoS One 8:e71148.

    https://doi.org/10.1371/journal.pone.0071148

    • PubMed
    • Google Scholar
24. 1. Hagen F
    2. Khayhan K
    3. Theelen B
    4. Kolecka A
    5. Polacheck I
    6. Sionov E
    7. Falk R
    8. Parnmen S
    9. Lumbsch HT
    10. Boekhout T

    (2015) Recognition of seven species in the Cryptococcus gattii/Cryptococcus neoformans species complex

    Fungal Genetics and Biology 78:16–48.

    https://doi.org/10.1016/j.fgb.2015.02.009

    • PubMed
    • Google Scholar
25. 1. Hagen F
    2. Lumbsch HT
    3. Arsic Arsenijevic V
    4. Badali H
    5. Bertout S
    6. Billmyre RB
    7. Bragulat MR
    8. Cabañes FJ
    9. Carbia M
    10. Chakrabarti A
    11. Chaturvedi S
    12. Chaturvedi V
    13. Chen M
    14. Chowdhary A
    15. Colom MF
    16. Cornely OA
    17. Crous PW
    18. Cuétara MS
    19. Diaz MR
    20. Espinel-Ingroff A
    21. Fakhim H
    22. Falk R
    23. Fang W
    24. Herkert PF
    25. Ferrer Rodríguez C
    26. Fraser JA
    27. Gené J
    28. Guarro J
    29. Idnurm A
    30. Illnait-Zaragozi MT
    31. Khan Z
    32. Khayhan K
    33. Kolecka A
    34. Kurtzman CP
    35. Lagrou K
    36. Liao W
    37. Linares C
    38. Meis JF
    39. Nielsen K
    40. Nyazika TK
    41. Pan W
    42. Pekmezovic M
    43. Polacheck I
    44. Posteraro B
    45. de Queiroz Telles F
    46. Romeo O
    47. Sánchez M
    48. Sampaio A
    49. Sanguinetti M
    50. Sriburee P
    51. Sugita T
    52. Taj-Aldeen SJ
    53. Takashima M
    54. Taylor JW
    55. Theelen B
    56. Tomazin R
    57. Verweij PE
    58. Wahyuningsih R
    59. Wang P
    60. Boekhout T

    (2017) Importance of resolving fungal nomenclature: the case of multiple pathogenic species in the Cryptococcus genus

    mSphere 2:e00238-17.

    https://doi.org/10.1128/mSphere.00238-17

    • PubMed
    • Google Scholar
26. 1. Hall BM
    2. Ma CX
    3. Liang P
    4. Singh KK

    (2009) Fluctuation analysis CalculatOR: a web tool for the determination of mutation rate using Luria–Delbrück fluctuation analysis

    Bioinformatics 25:1564–1565.

    https://doi.org/10.1093/bioinformatics/btp253

    • PubMed
    • Google Scholar
27. 1. Harfe BD
    2. Jinks-Robertson S

    (2000) DNA mismatch repair and genetic instability

    Annual Review of Genetics 34:359–399.

    https://doi.org/10.1146/annurev.genet.34.1.359

    • PubMed
    • Google Scholar
28. 1. Healey KR
    2. Zhao Y
    3. Perez WB
    4. Lockhart SR
    5. Sobel JD
    6. Farmakiotis D
    7. Kontoyiannis DP
    8. Sanglard D
    9. Taj-Aldeen SJ
    10. Alexander BD
    11. Jimenez-Ortigosa C
    12. Shor E
    13. Perlin DS

    (2016) Prevalent mutator genotype identified in fungal pathogen Candida glabrata promotes multi-drug resistance

    Nature Communications 7:11128.

    https://doi.org/10.1038/ncomms11128

    • PubMed
    • Google Scholar
29. 1. Heck JA
    2. Argueso JL
    3. Gemici Z
    4. Reeves RG
    5. Bernard A
    6. Aquadro CF
    7. Alani E

    (2006) Negative epistasis between natural variants of the Saccharomyces cerevisiae MLH1 and PMS1 genes results in a defect in mismatch repair

    PNAS 103:3256–3261.

    https://doi.org/10.1073/pnas.0510998103

    • PubMed
    • Google Scholar
30. 1. Himmelreich R
    2. Plagens H
    3. Hilbert H
    4. Reiner B
    5. Herrmann R

    (1997) Comparative analysis of the genomes of the bacteria Mycoplasma pneumoniae and Mycoplasma genitalium

    Nucleic Acids Research 25:701–712.

    https://doi.org/10.1093/nar/25.4.701

    • PubMed
    • Google Scholar
31. 1. Hsueh YP
    2. Idnurm A
    3. Heitman J

    (2006) Recombination hotspots flank the Cryptococcus mating-type locus: implications for the evolution of a fungal sex chromosome

    PLoS Genetics 2:e184.

    https://doi.org/10.1371/journal.pgen.0020184

    • PubMed
    • Google Scholar
32. 1. Hunter N
    2. Chambers SR
    3. Louis EJ
    4. Borts RH

    (1996)

    The mismatch repair system contributes to meiotic sterility in an interspecific yeast hybrid.

    The EMBO Journal 15:1726–1759.

    • PubMed
    • Google Scholar
33. 1. Kidd SE
    2. Hagen F
    3. Tscharke RL
    4. Huynh M
    5. Bartlett KH
    6. Fyfe M
    7. Macdougall L
    8. Boekhout T
    9. Kwon-Chung KJ
    10. Meyer W

    (2004) A rare genotype of Cryptococcus gattii caused the cryptococcosis outbreak on Vancouver Island (British Columbia, Canada)

    PNAS 101:17258–17263.

    https://doi.org/10.1073/pnas.0402981101

    • PubMed
    • Google Scholar
34. 1. Kwon-Chung KJ
    2. Bennett JE
    3. Wickes BL
    4. Meyer W
    5. Cuomo CA
    6. Wollenburg KR
    7. Bicanic TA
    8. Castañeda E
    9. Chang YC
    10. Chen J
    11. Cogliati M
    12. Dromer F
    13. Ellis D
    14. Filler SG
    15. Fisher MC
    16. Harrison TS
    17. Holland SM
    18. Kohno S
    19. Kronstad JW
    20. Lazera M
    21. Levitz SM
    22. Lionakis MS
    23. May RC
    24. Ngamskulrongroj P
    25. Pappas PG
    26. Perfect JR
    27. Rickerts V
    28. Sorrell TC
    29. Walsh TJ
    30. Williamson PR
    31. Xu J
    32. Zelazny AM
    33. Casadevall A

    (2017) The case for adopting the "species complex" nomenclature for the etiologic agents of cryptococcosis

    mSphere 2:e00357-16.

    https://doi.org/10.1128/mSphere.00357-16

    • PubMed
    • Google Scholar
35. 1. Kwon-Chung KJ
    2. Varma A
    3. Edman JC
    4. Bennett JE

    (1992) Selection of ura5 and ura3 mutants from the two varieties of Cryptococcus neoformans on 5-fluoroorotic acid medium

    Medical Mycology 30:61–69.

    https://doi.org/10.1080/02681219280000081

    • PubMed
    • Google Scholar
36. 1. LeClerc JE
    2. Li B
    3. Payne WL
    4. Cebula TA

    (1996) High mutation frequencies among Escherichia coli and Salmonella pathogens

    Science 274:1208–1211.

    https://doi.org/10.1126/science.274.5290.1208

    • PubMed
    • Google Scholar
37. 1. Li H
    2. Durbin R

    (2009) Fast and accurate short read alignment with burrows-wheeler transform

    Bioinformatics 25:1754–1760.

    https://doi.org/10.1093/bioinformatics/btp324

    • PubMed
    • Google Scholar
38. 1. Li H
    2. Handsaker B
    3. Wysoker A
    4. Fennell T
    5. Ruan J
    6. Homer N
    7. Marth G
    8. Abecasis G
    9. Durbin R
    10. 1000 Genome Project Data Processing Subgroup

    (2009) The Sequence alignment/map format and samtools

    Bioinformatics 25:2078–2079.

    https://doi.org/10.1093/bioinformatics/btp352

    • PubMed
    • Google Scholar
39. 1. Lin X
    2. Hull CM
    3. Heitman J

    (2005) Sexual reproduction between partners of the same mating type in Cryptococcus neoformans

    Nature 434:1017–1021.

    https://doi.org/10.1038/nature03448

    • PubMed
    • Google Scholar
40. 1. Liu OW
    2. Chun CD
    3. Chow ED
    4. Chen C
    5. Madhani HD
    6. Noble SM

    (2008) Systematic genetic analysis of virulence in the human fungal pathogen Cryptococcus neoformans

    Cell 135:174–188.

    https://doi.org/10.1016/j.cell.2008.07.046

    • PubMed
    • Google Scholar
41. 1. Magditch DA
    2. Liu TB
    3. Xue C
    4. Idnurm A

    (2012) DNA mutations mediate microevolution between host-adapted forms of the pathogenic fungus Cryptococcus neoformans

    PLoS Pathogens 8:e1002936.

    https://doi.org/10.1371/journal.ppat.1002936

    • PubMed
    • Google Scholar
42. 1. Marsischky GT
    2. Filosi N
    3. Kane MF
    4. Kolodner R

    (1996) Redundancy of Saccharomyces cerevisiae MSH3 and MSH6 in MSH2-dependent mismatch repair

    Genes & Development 10:407–420.

    https://doi.org/10.1101/gad.10.4.407

    • PubMed
    • Google Scholar
43. 1. McKenna A
    2. Hanna M
    3. Banks E
    4. Sivachenko A
    5. Cibulskis K
    6. Kernytsky A
    7. Garimella K
    8. Altshuler D
    9. Gabriel S
    10. Daly M
    11. DePristo MA

    (2010) The genome analysis toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data

    Genome Research 20:1297–1303.

    https://doi.org/10.1101/gr.107524.110

    • PubMed
    • Google Scholar
44. 1. Moxon R
    2. Bayliss C
    3. Hood D

    (2006) Bacterial contingency loci: the role of simple sequence DNA repeats in bacterial adaptation

    Annual Review of Genetics 40:307–333.

    https://doi.org/10.1146/annurev.genet.40.110405.090442

    • PubMed
    • Google Scholar
45. 1. Ni M
    2. Feretzaki M
    3. Li W
    4. Floyd-Averette A
    5. Mieczkowski P
    6. Dietrich FS
    7. Heitman J

    (2013) Unisexual and heterosexual meiotic reproduction generate aneuploidy and phenotypic diversity de novo in the yeast Cryptococcus neoformans

    PLoS Biology 11:e1001653.

    https://doi.org/10.1371/journal.pbio.1001653

    • PubMed
    • Google Scholar
46. 1. Oliver A
    2. Cantón R
    3. Campo P
    4. Baquero F
    5. Blázquez J

    (2000) High frequency of hypermutable Pseudomonas aeruginosa in cystic fibrosis lung infection

    Science 288:1251–1253.

    https://doi.org/10.1126/science.288.5469.1251

    • PubMed
    • Google Scholar
47. 1. Pitkin JW
    2. Panaccione DG
    3. Walton JD

    (1996) A putative cyclic peptide efflux pump encoded by the TOXA gene of the plant-pathogenic fungus Cochliobolus carbonum

    Microbiology 142:1557–1565.

    https://doi.org/10.1099/13500872-142-6-1557

    • PubMed
    • Google Scholar
48. 1. Rayssiguier C
    2. Thaler DS
    3. Radman M

    (1989) The barrier to recombination between Escherichia coli and Salmonella typhimurium is disrupted in mismatch-repair mutants

    Nature 342:396–401.

    https://doi.org/10.1038/342396a0

    • PubMed
    • Google Scholar
49. 1. Reenan RA
    2. Kolodner RD

    (1992)

    Isolation and characterization of two Saccharomyces cerevisiae genes encoding homologs of the bacterial HexA and MutS mismatch repair proteins

    Genetics 132:963–973.

    • PubMed
    • Google Scholar
50. 1. Rhodes J
    2. Beale MA
    3. Vanhove M
    4. Jarvis JN
    5. Kannambath S
    6. Simpson JA
    7. Ryan A
    8. Meintjes G
    9. Harrison TS
    10. Fisher MC
    11. Bicanic T

    (2017)

    A population genomics approach to assessing the genetic basis of within-host microevolution underlying recurrent cryptococcal meningitis infection

    G3: Genes|Genomes|Genetics 7:1165–1176.

    • Google Scholar
51. 1. Skelly DA
    2. Magwene PM
    3. Meeks B
    4. Murphy HA

    (2017) Known mutator alleles do not markedly increase mutation rate in clinical Saccharomyces cerevisiae strains

    Proceedings of the Royal Society B: Biological Sciences 284:20162672.

    https://doi.org/10.1098/rspb.2016.2672

    • Google Scholar
52. 1. Springer DJ
    2. Phadke S
    3. Billmyre B
    4. Heitman J

    (2012) Cryptococcus gattii, no longer an accidental pathogen?

    Current Fungal Infection Reports 6:245–256.

    https://doi.org/10.1007/s12281-012-0111-0

    • PubMed
    • Google Scholar
53. 1. Springer DJ
    2. Ren P
    3. Raina R
    4. Dong Y
    5. Behr MJ
    6. McEwen BF
    7. Bowser SS
    8. Samsonoff WA
    9. Chaturvedi S
    10. Chaturvedi V

    (2010) Extracellular fibrils of pathogenic yeast Cryptococcus gattii are important for ecological niche, murine virulence and human neutrophil interactions

    PLoS One 5:e10978.

    https://doi.org/10.1371/journal.pone.0010978

    • PubMed
    • Google Scholar
54. 1. Strand M
    2. Prolla TA
    3. Liskay RM
    4. Petes TD

    (1993) Destabilization of tracts of simple repetitive DNA in yeast by mutations affecting DNA mismatch repair

    Nature 365:274–276.

    https://doi.org/10.1038/365274a0

    • PubMed
    • Google Scholar
55. 1. Su SS
    2. Modrich P

    (1986) Escherichia coli mutS-encoded protein binds to mismatched DNA base pairs

    PNAS 83:5057–5061.

    https://doi.org/10.1073/pnas.83.14.5057

    • PubMed
    • Google Scholar
56. 1. Sun S
    2. Billmyre RB
    3. Mieczkowski PA
    4. Heitman J

    (2014) Unisexual reproduction drives meiotic recombination and phenotypic and karyotypic plasticity in Cryptococcus neoformans

    PLoS Genetics 10:e1004849.

    https://doi.org/10.1371/journal.pgen.1004849

    • PubMed
    • Google Scholar
57. 1. Tamura K
    2. Stecher G
    3. Peterson D
    4. Filipski A
    5. Kumar S

    (2013) MEGA6: molecular evolutionary genetics analysis version 6.0

    Molecular Biology and Evolution 30:2725–2729.

    https://doi.org/10.1093/molbev/mst197

    • PubMed
    • Google Scholar
58. 1. Tenaillon O
    2. Barrick JE
    3. Ribeck N
    4. Deatherage DE
    5. Blanchard JL
    6. Dasgupta A
    7. Wu GC
    8. Wielgoss S
    9. Cruveiller S
    10. Médigue C
    11. Schneider D
    12. Lenski RE

    (2016) Tempo and mode of genome evolution in a 50,000-generation experiment

    Nature 536:165–170.

    https://doi.org/10.1038/nature18959

    • PubMed
    • Google Scholar
59. 1. Tenaillon O
    2. Le Nagard H
    3. Godelle B
    4. Taddei F

    (2000) Mutators and sex in bacteria: conflict between adaptive strategies

    PNAS 97:10465–10470.

    https://doi.org/10.1073/pnas.180063397

    • PubMed
    • Google Scholar
60. 1. Thompson DA
    2. Desai MM
    3. Murray AW

    (2006) Ploidy controls the success of mutators and nature of mutations during budding yeast evolution

    Current Biology 16:1581–1590.

    https://doi.org/10.1016/j.cub.2006.06.070

    • PubMed
    • Google Scholar
61. 1. Tran HT
    2. Keen JD
    3. Kricker M
    4. Resnick MA
    5. Gordenin DA

    (1997) Hypermutability of homonucleotide runs in mismatch repair and DNA polymerase proofreading yeast mutants

    Molecular and Cellular Biology 17:2859–2865.

    https://doi.org/10.1128/MCB.17.5.2859

    • PubMed
    • Google Scholar
62. 1. Verstrepen KJ
    2. Jansen A
    3. Lewitter F
    4. Fink GR

    (2005) Intragenic tandem repeats generate functional variability

    Nature Genetics 37:986–990.

    https://doi.org/10.1038/ng1618

    • PubMed
    • Google Scholar
63. 1. Wielgoss S
    2. Barrick JE
    3. Tenaillon O
    4. Wiser MJ
    5. Dittmar WJ
    6. Cruveiller S
    7. Chane-Woon-Ming B
    8. Médigue C
    9. Lenski RE
    10. Schneider D

    (2013) Mutation rate dynamics in a bacterial population reflect tension between adaptation and genetic load

    PNAS 110:222–227.

    https://doi.org/10.1073/pnas.1219574110

    • PubMed
    • Google Scholar
64. 1. Zhu P
    2. Zhai B
    3. Lin X
    4. Idnurm A

    (2013) Congenic strains for genetic analysis of virulence traits in Cryptococcus gattii

    Infection and Immunity 81:2616–2625.

    https://doi.org/10.1128/IAI.00018-13

    • PubMed
    • Google Scholar

Author details

1. R Blake Billmyre

   Department of Molecular Genetics and Microbiology, Duke University School of Medicine, Duke University Medical Center, Durham, United States

   Present address

   Stowers Institute for Medical Research, Kansas City, United States

   Contribution

   Conceptualization, Data curation, Software, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing—original draft, Writing—review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-4866-3711

2. Shelly Applen Clancey

   Department of Molecular Genetics and Microbiology, Duke University School of Medicine, Duke University Medical Center, Durham, United States

   Contribution

   Formal analysis, Validation, Investigation, Methodology, Writing—review and editing

   Competing interests

   No competing interests declared

3. Joseph Heitman

   Department of Molecular Genetics and Microbiology, Duke University School of Medicine, Duke University Medical Center, Durham, United States

   Contribution

   Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Methodology, Writing—original draft, Project administration, Writing—review and editing

   For correspondence

   heitm001@duke.edu

   Competing interests

   No competing interests declared

• 2,209

  views

• 332

  downloads

• 76

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.