Emerging evidence supports the view that the genome of C. neoformans exhibits plasticity beyond the haploid euploidy commonly observed in most clinical and environmental isolates [18, 25–28, 33]. As demonstrated here, genomic variation such as aneuploidy appears to have a major impact on the virulence of C. neoformans in a mammalian host. However, detailed investigations are needed to link specific changes to pathogenesis and the clinical outcome of cryptococcosis. In our study, we made the key observation that the black (monosomic) and white (disomic) variants of CBS7779 differed in their virulence in mice. This difference may be attributed to melanin formation because mutants defective in melanin are known to have reduced virulence and impaired dissemination to the brain [34–38]. However, the numbers of yeast cells in lung and brain tissue were similar for the black and white strains. These observations argue against melanin as the sole reason for the difference in virulence. Rather, it appears that melanin formation, although particularly sensitive to variation in chromosome copy number in CBS7779, may be just one of a number of traits affected by chromosomal changes. That is, aneuploidy in general may impair virulence through a number of mechanisms. For example, in other fungi, such as S. cerevisiae, aneuploidy results in a decreased growth rate and an imbalance in gene expression [39, 40]. For C. neoformans, we found that white strains grew more slowly in culture but their proliferation in the host was not impaired. Thus our in vitro growth conditions did not reflect proliferation in vivo. Our histopathological observations also suggested other contributors to the observed difference in virulence. For example, it appeared that infections with black strains resulted in more airway remodeling and this damage may account for the greater virulence. It is also possible that the enhanced expression of genes on chr 13 that we observed in the white strain created an imbalance in gene expression that directly or indirectly influenced the elaboration of known or unknown virulence factors. For example, it is likely that secreted factors contribute to pathology in the lung, and our observation that white, disomic strains have increased sensitivity to brefeldin A suggests an underlying perturbation of the secretory pathway.
The discovery of disomy in strain CBS7779 raises the question of whether this strain is particularly prone to genome instability, perhaps due to background defects in chromosome replication and segregation. The strain was initially selected for CGH analysis because Boekhout and van Belkum  reported that it had an unusually small genome (15 Mb versus 18-27 Mb for other strains). However, CGH did not reveal differences in content relative to the reference genome of strain H99 . In the current study, the ease in detecting additional chromosome changes in our analysis of second-generation variants supports the possibility of general genome instability in CBS7779. Conversely, chromosome and ploidy changes may be more common than previously appreciated in species of Cryptococcus, as indicated by reports of extensive genome variation including karyotype variability and changes in clinical and environmental strains of C. neoformans and C. gattii in culture and in mammalian hosts [17, 41–46]. The study of Fries et al.  is particularly informative because they found differences in karyotypes (rearrangements and different numbers of chromosomes) in sequential isolates of C. neoformans from AIDS patients. In addition, they passaged strains through mice and documented similar changes in karyotypes in cells recovered for three of the six isolates. The karyotype variations likely reflect translocations and segmental changes, and it is possible that stress conditions in mammalian hosts (e.g., oxidative, nitrosative and temperature stresses) promote chromosome variation. In our mouse passage experiments, the percentage of black colonies recovered from tissue was higher than in the starting inoculum, suggesting that they arose during infection. Although we are cautious about drawing conclusions about the frequency of copy number variation based solely on melanin as a surrogate indicator, our subsequent qPCR and CGH analyses revealed multiple chromosome changes in variants after passage in mice. More detailed studies with tagged chromosomes and fluctuation analysis will be needed to more closely examine the frequency of copy number change in animals and in response to stress.
More recently, Sionov et al.  analyzed fluconazole heteroresistance in the commonly studied C. neoformans strain H99 and found copy number changes for several chromosomes. In this case, selection for high-level resistance was correlated with disomy for chrs 1 and 4, a reduced growth rate and lower virulence. They also found a difference in virulence between a fluconazole resistant derivative H99 R64 and the parental H99 strain. Specifically, ~20% of the mice infected with the disomic strain died by the end of the experiment compared with ~50% inoculated with the wild-type strain. The appearance of disomic chromosomes in fluconazole-resistant isolates also occurs in the related species C. gattii[47, 48]. In our study, we demonstrated that disomy could be detected in strain H99 by tagging chr 13 with a neomycin resistance marker and selecting on high levels of neomycin. However, some isolates showed the expected amplification of chr 13 and others had changes in different chromosomes. Together, these studies suggest that copy number variation is a common property of C. neoformans.
Genomic variation in C. neoformans extends beyond aneuploidy to include the occurrence of diploid strains arising from a-α and same-sex mating, endoreplication or clonal mating, as well as the formation of polyploid cells in infected animals [19, 20, 25, 49]. For example, Lengeler et al.  examined 10 hybrid strains thought to arise from the fusion of serotype A and D mating partners, and they found that six were diploid and four were aneuploid with DNA content between 1N and 2N. We subsequently examined three of these AD hybrids by CGH and found that some chromosomes were preferentially retained from one parental genome or the other . For example, all three strains retained chr 1 from the serotype A parent and lost the serotype D homologue. A survey of an additional 16 AD hybrid strains revealed that chr 1 from the A serotype genomes was preferentially retained in 11 of the isolates . Recently, Lin et al.  surveyed 489 clinical and environmental strains of A serotype and found that 7.8% were diploid based on fluorescence flow cytometry. Thus diploid strains are relatively common, although in contrast, only six of the isolates appeared to be aneuploid. Characterization of laboratory-constructed diploids demonstrated that elevated ploidy has a minor negative influence on virulence in mice, and it is clear that diploids contribute to the disease spectrum because of the occurrence of these strains and mixed infections in AIDS patients [28, 49]. Interestingly, Desnos-Ollivier et al.,  showed that mixed infections with strains of different mating type, serotype, genotype and ploidy occur at a high frequency (18.4%) in patients; many of the isolates (8/23) in this study were diploid. Multilocus sequence typing indicated that some of the isolates may have arisen through diploidization (endoreplication). In our analysis, one strain from an AIDS patient (HC-6) appeared to be diploid with a copy number difference at chr 6. An isolate from a second patient (HC-4) also contained an elevated copy number for a segment of chr 9. Additional complexity comes from the discovery of tetraploid or octoploid cells that arise as a proportion of cryptococcal cells in the lungs of infected mice [25, 26]. It is possible that disomy may result from the reduction in chromosome number as cells transition from the diploid or polyploid states to the haploid complement of chromosomes. Overall, the emerging view is that aneuploidy and changes in ploidy are common features of C. neoformans.
The occurrence of aneuploidy in C. neoformans is reminiscent of the situation in other fungal pathogens, and especially in the pathogenic yeast Candida albicans. The diploid genome of this fungus exhibits a high level of plasticity, as strains with chromosomal rearrangements, loss of heterozygosity and aneuploidy are frequently observed. In addition, mating between diploid strains yields tetraploid strains that do not undergo meiosis but instead display random chromosome loss. As in C. neoformans, clinical isolates of C. albicans show considerable genomic diversity with the detection of translocations, chromosomal truncations and the formation of extra chromosomes. Elegant studies have shown that whole-chromosome or segmental aneuploidy can readily be detected by CGH and that this variation is associated with a growth advantage on L-sorbose or D-arabinose and with resistance to fluconazole [7, 11–15, 50–52]. Considerable genomic variation also occurs during growth in mammalian hosts and it has been found that aneuploidy influences virulence [6, 8, 9, 11, 53, 54]. Changes during DNA transformation also are observed in C. albicans and our observations on the appearance of disomic strains during neomycin selection indicates that similar variation may occur in C. neoformans[55, 56]. Chromosome variation is also a feature of the haploid yeast pathogen Candida glabrata. Specifically, Poláková et al.  examined the chromosome complements of 40 clinical isolates and found considerable variation including frequent examples of translocations, segmental duplications and the appearance of novel chromosomes. The sequences in regions of variation may influence interactions with the host and disease outcomes because they encode proteases, phospholipases and potential antifungal drug transporters. In fact, an isolate carrying a novel minichromosome showed increased tolerance to fluconazole.