OUP user menu

Cryptococcus neoformans population includes hybrid strains homozygous at mating-type locus

Massimo Cogliati, Maria Carmela Esposto, Anna Maria Tortorano, Maria Anna Viviani
DOI: http://dx.doi.org/10.1111/j.1567-1364.2006.00085.x 608-613 First published online: 1 June 2006


Recent attempts to characterise the hybrid strains of Cryptococcus neoformans have led to the identification of a cryptic population of hybrid strains (‘H strains’) with double DNA content but only a single mating-type allele. To verify a set of hypotheses concerning their origin, we investigated 14 previously isolated H strains and ten F1-progeny strains arising from H99 and JEC20 mating. The double DNA content was tested by flow cytometry; the presence of only one mating type was tested by amplifying 12 mating-type-specific genes and one gene unlinked with the mating-type locus (URA5). Analysis of the F1 progeny identified two H strains, and electrophoretic karyotyping confirmed the occurrence of genetic recombination. The simultaneous presence of the homozygous and heterozygous loci, and the fact that all of the F1-progeny strains presented a recombinant karyotype, suggest that the H strains originated from the post-meiotic random fusion of two of the four recombinant nuclei. Further studies are required to elucidate the role of the homozygous mating-type loci in the virulence of C. neoformans.

  • Cryptococcus neoformans
  • hybrid strains
  • mating type


Cryptococcus neoformans is a heterothallic basidiomycete that is capable of causing life-threatening meningitis mainly in immunocompromised hosts. The original distinction of var. neoformans and var. gattii (Kwon-Chung, 1975, 1976) has been recently revised, and the two varieties have now been elevated to species level (Boekhout, 2001; Kwon-Chung, 2002). However, the proposal to group all serotype D strains as var. neoformans and all serotype A strains as var. grubii (Franzot, 1999) is still disputed, and conflicts with several studies reporting the isolation of numerous hybrid AD strains (Viviani, 1997; Barò, 1999; Tanaka, 1999; Cogliati, 2000; Diaz, 2000; Steenbergen & Casadevall, 2000; Nishikawa, 2003). Molecular analysis has confirmed that the hybrid strains are heterozygous at the mating-type locus (MATa/MATα), one inherited from the parental A strain and the other from the D strain (Cogliati, 2001; Lengeler, 2001), and some epidemiological studies carried out using molecular tools have confirmed that the hybrid AD strains are relatively frequent and play an important role in the epidemiology of C. neoformans (Brandt, 1996; Meyer, 1999; Boekhout, 2001; FIMUA Cryptococcosis Network, 2002; Yan & Xu, 2002; Meyer, 2003). Finally, it has been shown that these strains are the result of recent hybridisation between serotype A and D populations (Xu, 2002). Taken together, these data show that, although the serotype A and D populations show significant genetic divergence, they are able to mate in the geographical areas in which both populations coexist.

Recent attempts to characterise the AD hybrid strains of C. neoformans have led to the identification of a further cryptic population of hybrid strains (here called ‘H strains’) with a double DNA content but only a single mating-type allele (Aa or Da or Aα or Dα) instead of a pair of a and α alleles, one inherited from the serotype A parent and one from the serotype D parent. In a previous paper, we described the isolation of 12 H strains in which all of the four allelic patterns were represented (Cogliati, 2001), and other authors (Yan & Xu, 2002) subsequently reported the isolation of three H strains (Aα, Dα and Da) during an epidemiological study carried out in the USA. Similarly, molecular typing of the strains collected in the European Confederation of Medical Mycology (ECMM) cryptococcosis survey revealed the presence of six H strains in the European population of C. neoformans (Cogliati, 2005).

It is possible to formulate a set of hypotheses concerning the origin of the H strains: (i) they may be diploid strains with homozygous genome originating from an incomplete mitotic event; (ii) they may be AD hybrid strains that have lost or modified one mating-type allele as a result of a mutation event; (iii) they may be AD hybrid strains originating from an incomplete meiotic event; or (iv) they may be AD hybrid strains originating from a post-meiotic event.

To investigate these hypotheses, we further studied some of the previously isolated H strains to confirm the presence of only one mating-type allele, and also studied the F1 progeny originating from the mating of the H99 (serotype A) and JEC20 (serotype D) reference strains.

Materials and methods


The 14 Cryptococcus neoformans H strains tested included eight that have been previously reported (IUM 84-3040, IUM 89-0384, IUM 95-1584, IUM 94-1912, IUM 96-4769, RV 28949, RV 29192, RV 52755) (Cogliati, 2001), and six identified in this study. One strain for each mating type was included as a reference haploid strain: JEC20 (Da), JEC21 (Dα), IUM 96-2828 (Aa), and H99 (Aα).


Genomic DNA was extracted and PCR-fingerprinted using the (GACA)4 primer as previously described (Viviani, 1997). The strains were grouped into four different genotypes on the basis of the combination of four major PCR-fingerprinting bands: the VN1 genotype (540 and 420 bp) was associated with serotype D haploid strains; the VN6 genotype (800, 540 and 475 bp) with serotype A haploid strains; and the VN3 (800, 540 and 420 bp) and VN4 genotypes (800, 540, 475 and 420 bp) were associated with AD hybrid strains.

Flow cytometry

The DNA content of all of the strains was determined by flow cytometry as described elsewhere (Tanaka, 1996) using a Becton Dickinson cytofluorimeter. The JEC20 and JEC21 strains were used as reference haploid strains.

PCR of mating-type-specific genes

The mating type was determined by amplifying a set of genes embedded in the mating-type locus: MFa, MFα, STE3a, STE3α, STE11a, STE11α, STE12a, and STE12α.

In addition, the STE20a and STE20α alleles underwent PCR as previously described (Lengeler, 2001). The primers and annealing temperatures are shown in Table 1. The PCR were performed using the following 50 μL PCR mix: 10X PCR buffer (500 mM KCl, 100 mM Tris-HCl, pH 8.3, Applied Biosystems, Monza, Italy), 3 mM MgCl2 (Applied Biosystems), 400 μM of each deoxynucleotide (Roche, Milan, Italy), 20 pmol of each primer, 2.5 U AmpliTaq polymerase (Applied Biosystems), and 400 ng of the DNA sample. The amplifications were run in a PCR Sprint thermal cycler (Thermo Hybaid, UK) using a 5-min denaturing cycle at 94°C, 30 cycles at 94°C for 30 s, 30 s at the annealing temperature (Table 1), and 1 min at 72°C, and a final 5-min extension cycle at 72°C. The amplicons were separated in 1.4% agarose gel at 60 V for 2 h.

View this table:
Table 1

Oligonucleotides used to amplify the URA5 gene and 12 mating-type-specific genes of Cryptococcus neoformans

Primer nameNucleotide sequenceTa (°C)Reference
STE20αA F5′-AGGACATCTATAGCAGAT-3′55Lengeler (2001)
STE20αA R5′-CCAAAAGCTGATGCTGTGGA-3′55Lengeler (2001)
STE20aA F5′-TCCACTGGCAACCCTGCGAG-3′64Lengeler (2001)
STE20αD F5′-AAATCGGCTACGGCACGTC-3′58Lengeler (2001)
STE20αD R5′-GATTTATCTCAGCAGCCACG-3′58Lengeler (2001)
STE20aD F5′-GATCTGTCTCAGCAGCCAC-3′64Lengeler (2001)
STE20aD R5′-AATATCAGCTGCGCAGGTGA-3′64Lengeler (2001)
  • Ta=annealing temperature.

URA5 gene restriction

The URA5 gene underwent PCR and restriction in order to test the presence of only one allele in a gene unlinked with the mating-type locus. The mix and thermal cycling were the same as those described above. The amplified band (720 bp) was purified using a Microcon 50 centrifugal filter (Millipore, Bedford, MA), and digested by the MspI restriction enzyme (Appligene, Illkirch, France). The restriction products were run in 2% agarose at 60 V for 2 h. The strains with a band at 440 bp were considered to be serotype A alleles and those with a band at 370 bp as serotype D alleles; the heterozygous strains presented both bands. All of the digestion products showed a non-discriminating 150 bp band and a series of bands of less than 100 bp that were not considered in the analysis.

Mating assay

The H99 and JEC20 reference strains were crossed in V8 medium, and incubated in the dark at 25°C for 4 weeks (Alspaugh, 1997). The culture was observed periodically to identify the presence of filaments and basidiospores and, when sporulation occurred, the basidiospores were dissected by micromanipulation and grown on Sabouraud dextrose agar.

Electrophoretic karyotyping

Ten strains of the F1 progeny from H99 and JEC20 were analysed by means of electrophoresis karyotyping to test their chromosomal recombination. The chromosomes were extracted as previously described (Wickes, 1994), and separated on 1% chromosome-grade agarose (Bio-Rad, Hercules, CA) using a CHEF Mapper (Biorad) pulse-field electrophoretic apparatus and the following run conditions: initial and final A-time 75 s and 150 s, linear ramp run time 30 h; initial and final B-time 150 s and 300 s, linear ramp run time 54 h. The voltage was set at 125 V, and the buffer temperature was 12°C. The karyotype profiles were analysed using Diversity One software (PDI, Huntington Station, NY).


Flow cytometry showed that all of the H strains contained twice as much DNA as the reference haploid strains, thus confirming their diploid or aneuploid nature (Table 2).

View this table:
Table 2

Molecular characteristics of Cryptococcus neoformans strains included in the present study

StrainGenotypePloidySTE20MFSTE3STE11STE12URA5 RFLP
IUM 95-1584VN42nααααA
IUM 97-4898VN42nααααAD
IUM 98-4437VN42nααααA
IUM 98-5004VN42nααααA
IUM 01-4728VN42nααααAD
RV 29192VN42nααααAD
RV 52755VN32nααααAD
IUM 89-0384VN42nααααD
IUM 96-4769VN32nααααAD
IUM 97-4875VN42nααααD
IUM 97-4897VN42nααααA
RV 28949VN42nααααAD
IUM 84-3040VN42nDaaaaaAD
IUM 94-1912VN32nAaaaaaD
IUM 96-2828VN6nAaaaaaA
  • * By PCR-fingerprinting with (GACA)4;

  • By flow cytometry.

  • S1-10=F1 progeny from H99 and JEC20 mating.

Table 2 shows the results of the amplifications used to identify a set of mating-type-specific genes. All eight previously reported strains were confirmed to be H strains, as were the further six strains described in this study. Seven isolates were identified as Aα, five as Dα, one as Da, and one as Aa.

Genotyping by PCR fingerprinting (Table 2) showed that all of the H strains had the VN3 or VN4 pattern, which identify the hybrid strains originating from the mating of serotype A and serotype D strains. These results were confirmed for seven isolates for which URA5 gene restriction analysis (Table 2) produced two bands: one characteristic of the serotype A and one of the serotype D allele (Fig. 1). Three isolates presented the serotype A allele and two the serotype D allele, the same found for the mating-type locus; by contrast, the IUM 94-1912 and IUM 97-4897 strains, respectively presented the URA5 D and URA5 A alleles, although the mating-type-locus allele was of the opposite serotype.


URA5 gene restriction profiles of nine Cryptococcus neoformans H strains. The reference haploid strains H99 (serotype A) and JEC21 (serotype D) are shown in lanes 1 and 2. The allelic pattern, and molecular size of the two discriminating bands, are shown below, and on the left side of the figure, respectively.

The ten basiodiospores (S1-S10) of the F1 progeny originating from the mating of H99 and JEC20 were isolated and analysed as described above (Table 2): three strains were haploid, mating type Da and presented the VN1 genotype characteristic of serotype D; five were diploid hybrid strains with mating type Aα/Da; and two were identified as H strains presenting the mating type Aα.

Analysis of the electrophoretic karyotype of all of the F1 basidiospores (including the two H strains) showed the presence of chromosomal bands randomly inherited from both parental strains. A similarity matrix confirmed that all of the F1 progeny strains shared chromosomal bands with both parental strains (Fig. 2, Table 3), thus demonstrating that the F1 isolates are recombinants of the H99 and JEC20 strains.


Electrophoretic karyotype of H strain S2 in comparison with that of the parental strains H99 and JEC20. SC = the Saccharomyces cerevisiae reference karyotype.

View this table:
Table 3

Similarity between F1 progeny and parental strains

StrainMating typeSimilarity vs. H99 (%)Similarity vs. JEC20 (%)


We here discuss the results on the basis of the hypotheses mentioned in the introduction.

The first hypothesis can be rejected easily because the H strains isolated from the F1 progeny originated from basidiospores: i.e. a meiotic, not a mitotic event. Furthermore, a mitotic event can be expected to lead to complete homozygosis of the whole genome, whereas the H strains in our study have a hybrid genotype (VN3 or VN4), and some are heterozygous at the URA5 locus.

The second hypothesis proposed the loss or a substantial modification of the mating-type locus, which was not detected in the H strains. This conflicts with the high frequency of isolation of H strains: we detected two H strains from ten F1 basidiospores, and a similar rate has been reported by other authors (Tanaka, 2003) who analysed the F1 progeny obtained by mating serotype A and D strains, whereas a lower isolation rate would be expected if the H strains originate from such a mutation event. Furthermore, the presence of only one mating type was confirmed for 12 different genes embedded in the mating-type locus, and so a deletion or modification should have involved all or most of the mating-type locus.

The third hypothesis speculates that H strains may originate from premeiotic karyogamy and subsequent interruption of the meiosis. However, the H strains isolated from the F1 progeny originated from basidiospores (the final product of meiosis) and, in this instance, the originating hybrids can be expected to be strains with a completely heterozygous genome, whereas our results show that they had only one mating-type allele inherited from one of the parents, and that a number of them also contained a single allele at the URA5 gene.

The most probable hypothesis is therefore that H strains originate from the postmeiotic random fusion of two of the four recombinant nuclei (Fig. 3). Three combinations of AD-hybrids are generated throughout this pathway, namely a/α, a/a and α/α, with an expected frequency of 67, 16.5 and 16.5%, respectively. The results of the present study showed that two out of seven (28%) hybrid F1 progeny were α/α H strains and five (70%) were heterozygous (a/α) at mating-type locus. Although these frequencies are very close to those expected and the small sample size is not statistically sufficient, the relatively high percentage of H strains seems to be consistent with a recombinant rather than a mutation event model. This is also suggested by the simultaneous presence of homozygous and heterozygous loci, and supported by the finding that all of the F1 progeny (including the two H strains) showed a recombinant karyotype. In addition, such a mechanism of origin may involve the formation of differently sized basiodiospores (with diploid basiodiospores being larger than haploid basiodiospores), as previously reported (Kwon-Chung, 1978). The formation of nucleus-deficient basiodiospores also cannot be excluded, and would explain the poor viability of F1 hybrid strains reported by some authors (Lengeler, 2001). As our hypothesis is supported by a set of indirect proofs, further studies are ongoing to detect recombinant nuclei during meiosis and to follow their migration towards basidiospores to exclude other possible mechanisms.

Figure 3

Hypothesis of the origin of AD hybrid strains: (1) the two nuclei from serotype A and serotype D parental strains fuse inside the basidium; (2) four recombinant nuclei are produced by meiosis: two containing the MATα locus, and two the MATa locus; (3) the four recombinant nuclei remain in the basidium and duplicate by mitosis; and (4) the four recombinant nuclei migrate randomly in the basidiospores. Two nuclei (MATa and MATα, MATa and MATa, or MATα and MATα) may fuse to give rise to a diploid hybrid strain.

The existence of H strains in the C. neoformans population is an important step toward elucidating the role of mating type in virulence because the interactions of a single mating-type locus with hybrid backgrounds has not yet been investigated. Other authors have studied diploid serotype D strains that were homozygous at the mating-type locus (Hull, 2002): these strains were constructed by inducing the fusion of cells with the same mating type by means of assisted mating, and were used to evaluate the effect of ploidy on the mating process, but their virulence was not determined.

In conclusion, H strains are a subpopulation of C. neoformans that originate from the hybridisation of serotype A and D populations, and can be isolated from clinical samples. Further studies are required to elucidate the role of homozygous mating-type alleles in the virulence of C. neoformans.


  • Editor: Stuart Levitz


View Abstract