The α-Specific Cell Identity Factor Sxi1α Is Not Required for Virulence of Cryptococcus neoformans

Abstract

Cryptococcus neoformans is a human fungal pathogen that has two mating types (a and α). Experiments have shown that in some backgrounds α strains are more virulent than a strains. Our studies reveal that the only known α-specific factor, SXI1α, is not necessary for virulence.

Cryptococcus neoformans is a fungus found in the environment associated with bird guano, soil, and certain tree species that has the capacity to infect humans and cause disease (1). Disease occurs primarily in immunocompromised individuals and leads to cryptococcal meningoencephalitis, which is uniformly fatal without treatment. C. neoformans has two mating types, a and α, which fuse with one another and undergo sexual differentiation that results in the production of spores (6). Several virulence traits have been identified in C. neoformans, including being of the α mating type.

Strikingly, over 95% of all clinical and environmental isolates of C. neoformans are α (7), and analysis of a congenic strain pair (a versus α) revealed increased virulence by α strains in a mouse model (8). The role that mating type plays in C. neoformans virulence has sparked interest in understanding the differences between a and α cells.

In fungi, the region of the genome that distinguishes mating types is the mating type (MAT) locus (4). The C. neoformans MAT locus contains over 20 genes, spans more than 100 kb, and is composed almost entirely of divergent alleles of a common gene set (9). Hence, there are alleles of each gene in the MAT locus that reside in both MATa and MATα, with one exception: the SXI1α gene is unique to MATα and is necessary for specifying a/α cell identity and controlling sexual development (5).

Given that the Sxi1α regulator of mating type is the only α-specific gene identified thus far, it represented an ideal candidate for an α-specific factor contributing to increased virulence of α cells. We addressed the role that Sxi1α might play in virulence by comparing sxi1αΔ strains to their wild-type counterparts in a mouse model of infection.

Although SXI1α had previously been deleted from the background most commonly used for genetic analyses (C. neoformans var. neoformans serotype D) (5), virulence affects can be assessed more readily in a more virulent background (C. neoformans var. grubii serotype A). We therefore deleted the SXI1α gene from the serotype A strain H99 via biolistic transformation of a PCR-generated deletion construct (URA5 selectable marker) into a 5-fluoroorotic acid-resistant H99 derivative (F99) as described by Davidson et al. (2). Primer pairs used to create the deletion construct were CHO467 (CACCGTTATCCAAGGTCCTCA) and CHO468 (CAAACAATAAGCAGCTGATCAAATAGAGCGAAGTTGCTCGACC), CHO473 (TTGCGCAGATGCATGGTGTATATGTACTTGCCTCCAGGAGGTGG) and CHO474 (GACCGTCGATGGAGCTAGAGG), and CHO469 (GGTCGAGCAACTTCGCTCTATTTGATCAGCTGCTTATTGTTTG) andCHO472 (GGTCGAGCAACTTCGCTCTATTTGATCAGCTGCTTATTGTTTG). Using a recently developed serotype A strain (10), we screened transformants for a mating defect based on the known sxi1αΔ phenotype in serotype D. Crosses between 100 Ura⁺ transformants and the congenic serotype A strain KN99-5a were carried out on V8 (pH 5) agar as described previously (10), and this screen resulted in the isolation of 12 transformants with defects in sexual development. These strains were then subjected to Southern analysis, and nine were shown to have restriction patterns consistent with deletions of SXI1α. Two representative strains are shown in Fig. 1A. An additional probe to the SXI1α open reading frame confirmed the absence of the wild-type SXI1α gene in the deletion strains (data not shown). A wild-type strain (H99), two independent sxi1αΔ strains (CHY773 [no. 1] and CHY774 [no. 2]), and an ectopic integrant (CHY775), in which the marker integrated randomly into a different location in the genome, were chosen for further evaluation.

FIG. 1.

Sxi1α controls sexual development in C. neoformans var. grubii. (A) Schematic diagram of anticipated fragments on a Southern blot of wild-type and sxi1α deletion genomic DNA restriction digested with BamHI and EcoRI. The probe (created by PCR with primers CHO467 and CHO468) is indicated as a bar in the promoter region of the SXI1α gene. The Southern digest shows predicted fragments for the wild type (2.2 kb) and two deletion strains (6.2 kb). (B) Mating tests with wild-type and sxi1α deletion strains. Left panel, wild-type a strain crossed with a wild-type α strain. Center and right panels, wild-type a strain crossed with two independent sxi1α deletion strains.

Mating assays were carried out in more detail after the initial screening of transformants. Figure 1B reveals that while sxi1αΔ strains undergo limited filamentation, they do not form basidia or spores, even after extended incubation periods on V8 agar. The serotype A sxi1αΔ strains were also tested in a series of assays designed to reveal any defects in several virulence traits. In each case, an attribute known to be important for virulence was tested. The ability to grow at 37°C, prototrophic growth, melanin production, capsule production, and urease production were evaluated; however, none of these testable virulence traits was influenced by the deletion of SXI1α in serotype A or D (data not shown).

Although none of the virulence traits tested in vitro was affected by the deletion of SXI1α, it is possible that pathways controlled by Sxi1α contribute to virulence in the animal host. To test this hypothesis, we used a mouse model of infection to evaluate deletion strains for their ability to cause cryptococcal meningitis. Experiments were carried out with the serotype A wild-type, sxi1αΔ, and Ura⁺ ectopic integrant strains. A/Jcr mice were infected via inhalation with a suspension containing 5 × 10⁵ cells in 0.05 ml of phosphate-buffered saline. Animals were monitored over the course of 26 days for signs of infection, and moribund animals were sacrificed between 20 and 26 days postinfection (Fig. 2A). There were no significant differences observed between the wild-type and sxi1αΔ strains, indicating that Sxi1α plays no obvious role in virulence under these conditions.

FIG. 2.

sxi1αΔ strains are as virulent as wild-type strains in vivo. (A) Wild-type, sxi1αΔ, and control strains in the serotype A background were inoculated into A/Jcr mice intranasally. Mice were monitored for the development of symptoms consistent with cryptococcal meningitis, and moribund animals were sacrificed. The median survival time of mice infected with the wild-type strain was 21.5 days compared to 22.3 and 23.0 days for the independent sxi1αΔ test strains and 24.0 days for the ectopic control strain. P values from a Mann-Whitney test of statistical significance comparing each test strain with the wild-type control were >0.05, confirming that there were no significant differences among the strains tested. (B) Wild-type and sxi1αΔ strains in the serotype D background were inoculated into DBA mice by tail vein injection. Mice were monitored for the development of symptoms consistent with cryptococcal meningitis, and moribund animals were sacrificed. The median survival time of mice infected with the wild-type strain was 39.5 days compared to 32.4 and 28.0 days for the independent sxi1αΔ test strains. P values from a Mann-Whitney test of statistical significance comparing each test strain with the wild-type control were >0.05, indicating that there were no significant differences among the strains tested.

To address the possibility that differences in serotype could affect the role of SXI1α in virulence, we tested the serotype D wild-type reference strain JEC21 (8) and two independent sxi1αΔ strains (CHY610 [no. 1] and CHY611 [no. 2]) constructed in a JEC21 ura5 strain (JEC43) in virulence experiments with a murine model of infection. DBA mice were infected via lateral tail vein injections with 10⁷ cells in 0.05 ml of phosphate-buffered saline. Animals were monitored over the course of 56 days for signs of infection, and moribund animals were sacrificed (Fig. 2B). In the case of the serotype D strains, the profile of disease progression was more variable. Some animals developed morbidity as early as day 16 after infection, whereas others survived to the end of the experiment on day 56. Mice surviving to day 56 were apparently healthy with no signs of disease. The wild-type strain took the longest to affect the mice, and even in this case, nearly one-third of animals did not succumb to disease. Both of the sxi1αΔ strains began to affect the mice before the wild-type strain, suggesting that these strains could be moderately hypervirulent. A Mann-Whitney test of the data indicates that the apparent differences are not statistically significant for either of the test strains (P = 0.1656 for sxi1αΔ no. 1; P = 0.0935 for sxi1αΔ no. 2), but we cannot exclude the possibility that sxi1αΔ mutants may be somewhat more virulent in the serotype D background. We do conclude, however, that SXI1α is not required for virulence in either serotype A or serotype D strains of C. neoformans.

In light of these results, a fundamental question remains. What properties of α cells allow them to be more prevalent in nature and clinical isolates and to exhibit increased virulence in some backgrounds? Perhaps it is sexual development that leads to the predominance of α strains. Although there are no apparent differences in spore viability between a and α strains in the laboratory, there may be differences in nature that contribute to α prevalence. However, population genetics studies of C. neoformans suggest that growth is primarily clonal in nature (3), diminishing the role that sexual development might play in the wild. Alternatively, it has been proposed that the production of haploid spores during haploid α fruiting could explain the prevalence of α strains (11). Such spore-generating α strains would have advantages over a strains with respect to survival in harsh environmental conditions and distribution in the environment, increasing opportunities for human exposure. Haploid fruiting is unaffected in serotype D sxi1α deletion strains, suggesting that other α-specific factors contribute to the control of fruiting (5). In this case, it seems likely that other genes in the MAT locus, although not unique to α cells, have mating type-specific functions that allow differential α behavior.

With respect to the documented increase in virulence of α cells compared to a cells in mice, this phenomenon appears to be specific to serotype D and independent of sxi1α. This observed dichotomy between the serotypes offers an opportunity to investigate the emerging differences between serotypes A and D in comparative studies to identify serotype D, α-specific properties that lead to virulence distinctions in this background. Such factors may also play roles in virulence in other backgrounds, but their contributions may vary. Determining these disparities will likely reveal factors important to the virulence process in all serotypes and help to explain the differences in virulence profiles and disease progression between serotypes. In summary, what is clear for both C. neoformans serotypes A and D is that Sxi1α is not an essential factor in the virulence process, and it remains to be discovered what role SXI1α and mating type play in the life cycle and distribution of this unusual human fungal pathogen.