The study of quantitative traits provides a window on the interactions between multiple unlinked genetic loci. The interaction between hosts and pathogenic microbes, such as fungi, involves aspects of quantitative genetics for both partners in this dynamic equilibrium. One important pathogenic fungus is Cryptococcus neoformans, a basidiomycete yeast that can infect the human brain and whose mating system has two mating type alleles, a and α. The α mating-type allele has previously been linked to increased virulence potential. Here congenic C. neoformans strains were generated in the two well-characterized genetic backgrounds B3501α and NIH433a to examine the potential influence of genes outside of the mating-type locus on the virulence potential of mating type. The congenic nature of these new strain pairs was established by karyotyping, amplified fragment length polymorphism genotyping, and whole-genome molecular allele mapping (congenicity mapping). Virulence studies revealed that virulence was equivalent between the B3501 a and α congenic strains but the α strain was more virulent than its a counterpart in the NIH433 genetic background. These results demonstrate that genomic regions outside the mating type locus contribute to differences in virulence between a and α cells. The congenic strains described here provide a foundation upon which to elucidate at genetic and molecular levels how mating-type and other unlinked loci interact to enable microbial pathogenesis.
HUMAN pathogenic fungi are increasing in prevalence as the population of immunocompromised individuals escalates due to human immunodeficiency virus (HIV)/AIDS and to immunosuppression associated with cancer and its therapy. Like their human hosts, pathogenic fungi are eukaryotic cells and therefore current antifungal treatments are limited and often either stimulate the emergence of drug-resistant isolates or are quite toxic to humans. Thus, an understanding of the mechanisms by which fungal pathogens have adapted to survive and cause disease in their hosts is of paramount importance. For pathogenic fungi, which like most fungi can reproduce both sexually and asexually, recent studies have begun to forge a link between mating and virulence.
The role of mating in pathogenicity differs among human pathogenic fungi. For example, mating is thought to be an integral part of the Pneumocystis infection cycle on the basis of morphological analysis of infected lung tissue (reviewed in Cushion 2004; Thomas and Limper 2004). In other pathogenic fungi, such as Coccidioides immitis or Aspergillus fumigatus, a sexual cycle has not yet been described in the laboratory but population genetics studies provide evidence of actively recombining populations (Burt et al. 1996; Koufopanou et al. 1997; Varga and Toth 2003; Paoletti et al. 2005). The propagation of some organisms, such as Candida albicans and Cryptococcus neoformans, is largely clonal although there is evidence of recombination and both organisms have retained mating-type loci, mating machinery, and either a complete sexual cycle or at least a parasexual one (Kwon-Chung 1975; Pujol et al. 1993; Gräser et al. 1996; Franzot and Casadevall 1997; Hull and Johnson 1999; Xu et al. 1999; Hull et al. 2000; Magee and Magee 2000; Lengeler et al. 2002; Litvintseva et al. 2003). Mating of C. neoformans has thus far been observed only in the laboratory, although isolation of intervarietal hybrid strains reflects mating events that can occur in nature (Lengeler et al. 2001; Hull and Heitman 2002). There is also substantial evidence that several components involved in Cryptococcus mating are associated with virulence, including the transcription factor Ste12 and the PAK kinase Ste20 (Yue et al. 1999; Chang et al. 2000; Wang et al. 2002; Davidson et al. 2003). Here we address the role of mating type in the virulence of C. neoformans.
C. neoformans occurs in two varieties—grubii (serotype A) and neoformans (serotype D)—and diverged from the sibling species C. gattii ∼40 million years ago (Xu et al. 2000). The grubii and neoformans varieties have different disease epidemiologies with var. grubii causing the vast majority of cryptococcosis worldwide and >99% of infections in AIDS patients (Casadevall and Perfect 1998). However, in Europe var. neoformans can account for up to 20% of cryptococcosis cases, many in the context of an AD hybrid background.
Cryptococcus has two mating types—a and α. Yet, the vast majority of human cryptococcosis is caused by strains of the α mating type. Mating-type alleles in Cryptococcus are determined by a MAT locus that is >100 kb and contains >20 genes (Lengeler et al. 2002; Fraser et al. 2004). Analysis of markers within and flanking the MAT locus has shown that recombination is suppressed in the MAT locus (Lengeler et al. 2002). Due to the large size and complexity of the MAT locus, simple gene exchange experiments may not be sufficient to elucidate the role of this large genomic region in virulence. Here, a and α congenic strains were generated by a series of 10 backcrosses, yielding strains that are identical except at the MAT locus. In a similar previous study, the var. neoformans α mating-type strain JEC21α was found to be more virulent than the congenic a mating-type strain JEC20a, suggesting that α strains are more virulent in mice than a strains (Kwon-Chung et al. 1992). In contrast, the var. grubii congenic strains KN99a and KN99α showed no difference in murine virulence (Nielsen et al. 2003). However, following co-infection with the var. grubii congenic strains, the α strain more efficiently colonized the central nervous system than the a mating-type strain and provided evidence for a contribution of the MAT locus at least during co-infection (Nielsen et al. 2005).
These observations lead to a number of questions. Are there innate differences in the pathogenicity of varieties grubii and neoformans? Does the genetic background of the strain affect the role of mating type in virulence? Are 10 backcrosses sufficient to produce congenic strains? Can minor differences between the genomes of the congenic strains account for their virulence differences? To address these questions we generated two additional congenic strain pairs in var. neoformans (serotype D) and examined their levels of congenicity and their virulence. Our findings reveal that the virulence of the congenic strains differs from that of the parental strains and that genetic background can determine whether a virulence difference is observed between a and α cells in a murine model of cryptococcosis. The congenic strains described here can be used in future studies to identify other unlinked loci that interact with the mating-type locus to quantitatively affect virulence of Cryptococcus.
MATERIALS AND METHODS
Strains and media:
Parental var. neoformans strains used in this study were NIH433a, B3501α, JEC20a, and JEC21α (Kwon-Chung et al. 1992). B3501Aα denotes the freezer isolate of strain B3501α that was sequenced at Stanford, and it is identical to strain B3501α with the exception of one minor chromosome length polymorphism noted by pulsed-field gel electrophoresis (PFGE) (Loftus et al. 2005). Strains were grown on yeast extract-peptone-dextrose (YPD) medium. Matings were on V8 medium [5% (v/v) V8 juice, 3 mm KH2PO4, 4% (w/v) bactoagar] (Kwon-Chung et al. 1982).
Congenic strain construction:
To generate the congenic strain pair KN3501a and KN3501α, the parental strain JEC20a was crossed with strain B3501α and single basidiospores were isolated (Figure 1B). One of the F1 a progeny from this mating was backcrossed to B3501α and single basidiospores were isolated. The process of isolating a single-basidiospore cultures and backcrossing to B3501α was repeated an additional eight times. After the last backcross sibling a and α progeny were selected and designated KN3501a and KN3501α. The process was repeated to generate the congenic KN433a and KN433α strain pair from the parental strains NIH433a and JEC21α (Figure 1C). The α progeny strains were backcrossed to NIH433a and sibling strains were isolated from the tenth backcross.
PFGE was as described previously (Marra et al. 2004). Cells were grown in YPD at 30° with shaking at 250 rpm to an optical density at 600 nm of 0.5. Spheroplasts were prepared according to Lengeler et al. (2000), with modifications based on Wickes et al. (1994). Plugs containing ∼1 μg DNA were electrophoresced in a 13 × 14-cm 1% PFGE-grade agarose gel (Bio-Rad, Hercules, CA) in 0.5× Tris-borate-EDTA (TBE). Chromosomes were separated in 0.5× TBE using contour-clamped homogeneous electric field (CHEF) in a Bio-Rad DRII apparatus with the following settings: initial A time, 75 sec; final A time, 150 sec; start ratio, 1.0; run time, 40 hr; mode, 10; initial B time, 200 sec; final B time, 400 sec; start ratio, 1.0; run time, 56 hr; mode, 11. The voltage was set to 4 V/cm and the buffer temperature was set to 12°. The gel was stained in 0.5 μg/ml ethidium bromide for 0.5–1.0 hr, visualized, and photographed using ultraviolet light on a Gel-Doc imager (Bio-Rad).
Amplified fragment length polymorphisms (AFLPs) were carried out as previously described (Boekhout et al. 2001). Briefly, the restriction and ligation reactions were performed simultaneously on 10 ng DNA using MseI, EcoRI, and T4 DNA ligase, combined with EcoRI and MseI adaptors from PE Biosystems AFLP microbial fingerprinting kit. The first PCR was performed using EcoRI and MseI core sequence preselective primers. The second PCR used a more selective EcoRI primer labeled at the 5′ end with 6-carboxyfluorescein (FAM) combined with a more selective MseI primer. The AFLP products were electrophoresced on a 5% polyacrylamide gel on an ABI310 sequencer (PE Biosystems).
Genomic DNA was prepared using the Camgen yeast genomic DNA extraction kit (Whatman Bioscience) with the addition of 425- to 600-μm glass beads and vortexing for 1 min as a preparatory step. DNA concentration was adjusted to 10 ng/μl and diluted 10-fold for PCRs in 10 mm Tris-Cl, pH 8.0. A subset of microsatellites and restriction fragment length polymorphisms (RFLPs) identified as polymorphic in the B3501/B3502 linkage map (Marra et al. 2004) were selected to span each linkage group.
Microsatellite polymorphisms were characterized as described by Marra et al. (2004). Briefly, 100–300 bp of sequence flanking the microsatellite locus was amplified by PCR. The PCR amplicons were evaluated for size polymorphisms by electrophoresis in 6.7% polyacrylamide in 1× TBE in a 30 × 38-cm Bio-Rad SequiGen GT vertical gel apparatus. After electrophoresis the gels were silver stained and then exposed for 1–2 sec with white light onto X-ray duplication film (Kane X-Ray) and developed according to standard procedures.
RFLPs were characterized as described by Marra et al. (2004). Briefly, ∼800 bp of sequence flanking the RFLP was amplified by PCR. The PCR amplicons were then digested with the appropriate restriction enzyme and electrophoretically separated in 1% agarose gels in 1× TBE. Gels were visualized with UV light and photographed in a MultiImage light cabinet (Alpha Innotech).
Virulence studies were performed using the murine tail-vein injection model. Four- to 6-week-old female DBA mice (10 per strain) were injected directly in the lateral tail vein with 5 × 106 or 1 × 106 cells. The concentration of cells in the inoculum was confirmed by plating serial dilutions and enumerating colony-forming units (CFUs). Mice were monitored twice daily and those that showed signs of severe morbidity (weight loss, abnormal gait, extension of the cerebral portion of the cranium) were killed by CO2 inhalation. The animal protocol was approved by the Duke University Animal Use Committee. Survival data from the mouse experiments were analyzed by the Kruskal-Wallis test, and for animals that survived to the termination of the experiment the last day was considered the date of death for these analyses. Statistical results did not change if we assumed that the surviving animals at the termination of the experiment survived an additional 1000 days.
Congenic strain development:
Additional congenic strains were developed to determine whether genetic background affects the virulence of the a and α mating types in C. neoformans. The original var. neoformans congenic strains JEC20a and JEC21α were generated by crossing the environmental strain NIH433a, which has relatively low virulence in mice, with the clinical strain NIH12α (high virulence) to generate B3502a (low virulence) and B3501α (high virulence). Mating-type α progeny from a cross of B3501α to B3502a were backcrossed nine times to generate JEC21α and JEC20a, which is analogous to B3502a (Figure 1A). JEC21α was originally found to be more virulent than JEC20a but both strains had relatively low virulence (Kwon-Chung et al. 1992). The low virulence observed in the JEC20/21 congenic strains has at least two possible explanations. First, the low virulence could be due to the B3502a genetic background. Alternatively, the low virulence could be due to the multiple passages involved in generating the congenic strains. Previous studies have shown that continuous in vitro culture can reduce the virulence of C. neoformans strains (Franzot et al. 1998).
To differentiate between these two possibilities, and to determine the role of genetic background in mating-type virulence potential, we isolated congenic strains in two related strain backgrounds, B3501α and NIH433a (Figure 1A). The B3501α strain background has high virulence whereas the environmental isolate NIH433a has low virulence. Congenic strains were not isolated in the clinical NIH12α genetic background due to abundant monokaryotic fruiting by this strain. Because recombination is suppressed in the MAT locus (Lengeler et al. 2002), the sequenced JEC20a MAT locus should be identical to the NIH433a MAT locus. Similarly, the sequenced JEC21 MATα allele is identical to the MATα allele of B3501Aα (Loftus et al. 2005). Therefore, JEC20a and JEC21α were used as the parental strains to generate the congenic strains. Figure 1B illustrates how the JEC20a MAT locus was backcrossed into the B3501α genetic background to generate the sibling strains KN3501a and KN3501α. The KN3501 congenic strains, like their B3501α parental strain, are derived from a 50% clinical and 50% environmental genetic background. The JEC21α MAT locus was backcrossed into the NIH433a genetic background to generate the sibling strains KN433a and KN433α (Figure 1C), which are in an environmental genetic background.
Congenicity of strain pairs:
The KN3501 and KN433 congenic strain pairs were compared to each other and to the parent strains to identify any differences between the strains. First, the karyotype of the strains was analyzed using PFGE. Figure 2 shows that the KN3501 parental strains JEC20a and B3501α have different karyotypes. The most striking differences are the sizes of chromosomes 2, 5, 6, 8, 11, and 13. The sibling congenic strains KN3501a and KN3501α have identical karyotypes and their karyotype is most similar to B3501α. However, chromosomes 13 and 14 in the KN3501 congenic strains comigrate with those present in JEC20a.
The KN433 parental strains NIH433a and JEC21α also have different karyotypes. Chromosome 4 is larger in NIH433a than in JEC21α but chromosomes 5, 6, and 9 are all smaller than those of JEC21α. The sibling congenic strains KN433a and KN433α have identical karyotypes and show all the characteristics of the NIH433a karyotype. Thus, the KN3501 and KN433 congenic strain pairs are identical at the chromosome level, although the KN3501 strains differ from either parental strain.
AFLPs were used to genotype each strain to analyze the congenicity of the genomes. Comparison of the parental strains JEC20a with B3501α (Figure 3A) and of NIH433a with JEC21α (Figure 3B) reveals polymorphic fragments between the parental strains (Figure 3, A and B, arrows). The KN3501 congenic strain pair has AFLP genotypes that differ at only one location (Figure 3A, *) and their genotypes most resemble B3501α. The KN433 congenic strains have identical AFLP genotypes. Interestingly, one band present in NIH433a is absent in the congenic strains (Figure 3B, star), indicating that the KN433 strains might not be exactly identical to NIH433a. These data suggest that at a genomic level there is at least one difference between KN3501a and KN3501α but no difference could be detected between KN433a and KN433α.
One advantage of constructing congenic strains in the NIH433a and B3501α genetic backgrounds is the availability of the B3501/B3502 linkage map (Marra et al. 2004). This linkage map is based on microsatellites, RFLPs, and insertions/deletions (INDELs) between the B3501α and B3502/JEC20a genomes. To identify any detectable differences between the KN3501a and KN3501α congenic strains, microsatellite and RFLP markers that span the linkage groups and chromosomes were chosen (Figure 4). Chromosome 10 could not be analyzed because it is monomorphic (and presumably congenic) between B3501α and B3502a. All markers found to be polymorphic between B3502a and B3501α on the linkage map were also monomorphic between JEC20a and B3502a, indicating that these strains are isogenic. After 10 backcrosses into the B3501α genetic background, both KN3501a and KN3501α retained three regions (six markers in KN3501a but five markers in KN3501α) with JEC20a identity (Figure 4) that constitute ∼626 kb or ∼3% of the genome. Thus, KN3501a and -α are 97% congenic with B3501α.
To determine whether the remaining 3% of the genome could be rendered congenic we analyzed strains from an additional cross of KN3501a to B3501α. We sought to identify an a progeny strain that was congenic to B3501α for the two largest regions of dissimilarity on chromosomes 6 and 11. One a strain was found to contain the B3501 allele at both regions but this strain was sterile when crossed to B3501α. These regions of the genome do not appear to contain genes that might be implicated in self/nonself recognition or other mating processes so it is unclear what role, if any, these polymorphisms might play in fertility.
Nonetheless, comparison of the sibling strains KN3501a and KN3501α showed that the two strains differ from each other only at the MAT locus and in an ∼15-kb region on chromosome 9 and thus are 99.93% identical. These data indicate that KN3501a and KN3501α are congenic with each other, even though they differ by ∼3% from the B3501α parental background. Because the genomes of both B3501α and JEC20/21 are known, we can infer the complete genome sequence of the KN3501 congenic strains.
Since B3501α and B3502a are siblings from a cross of NIH12α to NIH433a, the polymorphisms observed between B3501α and B3502a are due to the presence of the NIH12α or NIH433a allele at each locus. Therefore, B3502a received 50% of its genome from NIH12α and the other 50% from NIH433a. Because B3502a and JEC20/21 are isogenic (Figure 1 and Marra et al. 2004), when we compared NIH433a and JEC21α we anticipated that 50% of the markers would be monomorphic (Figure 5, black bars) and 50% would be polymorphic (Figure 5, blue bars vs. red bars), in accord with our experimental observations. After 10 backcrosses into the NIH433a genetic background, both KN433a and KN433α were identical to each other and to NIH433a at all markers examined except those in the MAT locus.
In summary, the KN3501 and KN433 congenic strains were mapped and both strain pairs were estimated to be >99.9% identical. Furthermore, no differences were observed between the KN433 congenic strains and their parental strain NIH433a but the KN3501 congenic strains differed from the B3501α parental strain.
Congenic strain virulence:
We next compared the genomic profile of the strains with their virulence in a murine tail-vein model of cryptococcosis. Figure 6A shows that the congenic KN3501a and KN3501α strains produced very similar survival curves with no significant difference (P = 1.00). Similar results were obtained when the experiment was repeated with an inoculum of 5 × 106 cells (data not shown). Interestingly, the congenic strains were significantly less virulent than the genetic background strain B3501α (P ≤ 0.009), supporting the conclusion that their reduced virulence was attributable to either multiple passages or the 626-kb region persisting from JEC20a. The parental strains JEC20a and B3501α differed slightly in their survival curves with the α strain appearing to cause lethal infections of all animals faster than the a strain. However, this apparent difference in virulence between JEC20a and B3501α did not meet statistical significance (P = 0.427). Thus, the congenic KN3501a and KN3501α strains showed no difference in virulence in two independent experiments.
The survival curves in Figure 6B demonstrate that KN433α is significantly more virulent than the congenic strains NIH433a or KN433a (P = 0.001 and 0.009, respectively). These data support the conclusion that the NIH433 genetic background somehow influences the observed differences in virulence between a and α cells. No difference in growth at high temperature, auxotrophy, capsule size, or melanin production was observed in vitro that could account for the difference in virulence between KN433α and KN433a or NIH433a (data not shown). There was no statistically significant difference in virulence between the parental strains NIH433a and JEC21α (P = 0.473). We also note that the KN433 congenic strains were as or more virulent than both parental strains, suggesting that passage has contributed to enhanced virulence of the NIH433 environmental background.
The model yeast Saccharomyces cerevisiae and the human pathogenic fungus C. albicans have relatively small MAT loci consisting of just a few genes, which facilitates the generation of isogenic strains by allele exchange or single-gene disruptions. In contrast, the Cryptococcus MAT locus spans >100 kb and contains >20 genes, which makes the generation of isogenic strains considerably more challenging. To overcome this, we generated C. neoformans var. neoformans congenic strains in different, but related, genetic backgrounds by a backcrossing approach in which the MAT locus alleles were introgressed onto different genetic backgrounds. This approach yields strains in which the MAT locus a or α allele is present on otherwise identical genetic backgrounds.
The completely sequenced genomes of the C. neoformans strains B3501Aα and JEC21α (and by inference its congenic partner JEC20a) represent unique and invaluable tools for conducting genetic and molecular studies (Loftus et al. 2005). These genomic sequences were used to generate a highly saturated C. neoformans var. neoformans linkage map for B3501α and B3502/JEC20a (Marra et al. 2004). The congenic strains developed in this article are all closely related to the sequenced and mapped genomes. Genetic manipulation of the JEC21α genome was previously possible using the JEC20/21 congenic strain pair. The KN3501 congenic strain pair now enables similar approaches for the other sequenced strain, B3501Aα. Production of the congenic strains in sequenced and mapped genetic backgrounds has allowed us to scrutinize the congenicity of the strains and also to determine the precise differences between the congenic strains and their relationship to the parental strains.
Given that the meiotic progeny of a genetic cross inherit, on average, half of their genome from each parent, 10 backcrosses should result in congenic strains with 99.9% identity. Our data confirm this for NIH433 and the congenic strains KN433a and KN433α. However, the KN3501 congenic strains contain remnants from the opposite parental genome that persisted after 10 backcrosses and account for ∼3% of the genome. It is unclear why these portions of the JEC20a genome have persisted in the KN3501 congenic pair, but when additional backcrossing was conducted to obtain isolates in which this genomic region corresponded to the B3501α allele, the isolates were sterile. Despite this small difference from the B3501α parental strain, KN3501a and KN3501α are 99.9% identical. These data clearly demonstrate that the KN433 and KN3501 strain pairs are highly similar except at the MAT locus and therefore can be used to examine the role of mating type in virulence without significant confounding genetic variables.
In earlier studies, the α strain was more virulent than the a strain in the var. neoformans congenic strains JEC20a and JEC21α but both mating types displayed equivalent virulence in the var. grubii congenic strains KN99a and KN99α (Kwon-Chung et al. 1992; Nielsen et al. 2003). A number of hypotheses have been proposed to explain these differences (Nielsen et al. 2003; McClelland et al. 2004). First, the JEC20a and JEC21α strains might not be identical and differences in the genetic background of the strains could account for their virulence differences. This possibility seems unlikely since α progeny from a cross of JEC20a and JEC21α are more virulent than a progeny, providing evidence that the observed virulence difference is linked to mating type in these strains (Kwon-Chung et al. 1992).
Alternatively, there could be innate differences between variety grubii and neoformans strains, which have diverged for ∼20 million years. Mating type might play a role in virulence of var. neoformans strains but not of var. grubii strains. In single infections, the var. grubii strain KN99α was equivalent to KN99a in virulence. However, KN99α cells more readily penetrate the blood-brain barrier than KN99a cells during co-infection (Nielsen et al. 2005). Thus, differences in virulence potential between a and α strains in var. grubii are apparent during co-infection.
Finally, the genetic background of the congenic strains could affect the impact of the mating-type allele on virulence. This phenomenon could apply to either var. neoformans or var. grubii. In this model, different congenic strains could have different mating-type virulence characteristics. Our data support this hypothesis. Similar to the var. neoformans JEC20/21 results, the KN433 congenic strains showed that the α strain is more virulent than the a strain. Likewise, the KN3501 congenic strains, as well as an analogous pair of independently derived congenic strains in the B3501Aα genetic background (K. J. Kwon-Chung, personal communication), resemble the var. grubii KN99 congenic strains where no difference was observed in the virulence of a and α strains.
A correlation was observed between the overall virulence of the genetic background and differences in virulence associated with mating type. For example, both the JEC20/21 and KN433 genetic backgrounds had an overall low virulence, and in both cases the difference in virulence between congenic a and α strains was significant (Kwon-Chung et al. 1992 and this article). By contrast, the KN3501 and KN99 genetic backgrounds had high overall virulence, with no significant difference in virulence between the congenic a and α strains (Nielsen et al. 2003 and this article). Thus, the contribution of the α allele of the MAT locus may be more apparent in strains with lower virulence. In this case, var. grubii α congenic strains generated in genetic backgrounds with decreased virulence potential may exhibit enhanced virulence compared to their congenic a partner strains.
In contrast, no link was observed between the virulence of the congenic strains compared to the parental strains and mating-type virulence potential. The KN433 and KN99 congenic strains are as virulent or more virulent than their parental strains NIH433a and H99α but the KN433 pair shows a difference in mating-type virulence potential whereas the KN99 pair does not. Likewise, the KN3501 and JEC20/21 congenic strains are less virulent than their parental strains but the KN3501 pair shows no difference in virulence between mating types whereas the α strain is more virulent in the JEC20/21 pair.
Taken together, these data support the conclusion that genetic background plays a significant role in determining the potential effect of mating type on virulence in C. neoformans. Interestingly, the α strain was more virulent in all cases where a difference in virulence between mating types was observed. This observation suggests that there is an overall virulence advantage to being mating-type α that is more pronounced in certain genetic contexts. These studies provide a foundation from which to identify the genetic determinants that influence the virulence impact of the MAT locus α allele, and it will be of considerable interest to elucidate how the α allele collaborates with other unlinked genetic determinants to enhance the virulence composite, which on the basis of these findings represents a quantitative trait.
Mating type also contributes to virulence of other human pathogenic fungi. Both mating types of Histoplasma capsulatum (+ and −) are found in environmental soil samples in an equal ratio yet the vast majority of clinical isolates possess the − mating type (Kwon-Chung 1973; Kwon-Chung et al. 1974). Similar to some of the C. neoformans congenic strains, no difference in virulence was observed between H. capsulatum + and − mating types in murine infection experiments, but this issue has not yet been examined rigorously with congenic strains (Kwon-Chung 1981). In the diploid pathogenic fungus C. albicans, most clinical isolates are a/α and thus heterozygous at the mating-type locus (Lockhart et al. 2002). Recent studies have revealed that homozygous a/a and α/α strains are less virulent than a/α heterozygous strains and this difference could explain how the a/α mating type is conserved in the natural population (Lockhart et al. 2005). These data clearly implicate the mating-type locus in the virulence of divergent human pathogenic fungi. The next challenge will be to unravel how the MAT locus collaborates with other unlinked loci and to define the molecular differences that alter the impact of the a and α alleles on this process. With this understanding, we can better predict the genetic virulence composite of strains, which will aid in the identification of antifungal and vaccine targets for human pathogenic fungi.
We thank Johnny Huang for technical assistance and John Perfect and James Fraser for comments on the manuscript. This work was supported in part by National Institute of Allergy and Infectious Diseases R01 grants AI50113 and AI25783 to Joseph Heitman and Tom Mitchell, respectively. Gary Cox was a Burroughs Wellcome new investigator, and Joseph Heitman was a Burroughs Wellcome scholar in molecular pathogenic mycology and an investigator of the Howard Hughes Medical Institute.
↵1 Present address: Plant Pathology and Ecology, Connecticut Agricultural Experiment Station, New Haven, CT 06511.
Communicating editor: A. Mitchell
- Received May 3, 2005.
- Accepted June 1, 2005.
- Copyright © 2005 by the Genetics Society of America