Fungal genome and mating system transitions facilitated by chromosomal translocations involving intercentromeric recombination

Abstract

Species within the human pathogenic Cryptococcus species complex are major threats to public health, causing approximately 1 million annual infections globally. Cryptococcus amylolentus is the most closely known related species of the pathogenic Cryptococcus species complex, and it is non-pathogenic. Additionally, while pathogenic Cryptococcus species have bipolar mating systems with a single large mating type (MAT) locus that represents a derived state in Basidiomycetes, C. amylolentus has a tetrapolar mating system with 2 MAT loci (P/R and HD) located on different chromosomes. Thus, studying C. amylolentus will shed light on the transition from tetrapolar to bipolar mating systems in the pathogenic Cryptococcus species, as well as its possible link with the origin and evolution of pathogenesis. In this study, we sequenced, assembled, and annotated the genomes of 2 C. amylolentus isolates, CBS6039 and CBS6273, which are sexual and interfertile. Genome comparison between the 2 C. amylolentus isolates identified the boundaries and the complete gene contents of the P/R and HD MAT loci. Bioinformatic and chromatin immunoprecipitation sequencing (ChIP-seq) analyses revealed that, similar to those of the pathogenic Cryptococcus species, C. amylolentus has regional centromeres (CENs) that are enriched with species-specific transposable and repetitive DNA elements. Additionally, we found that while neither the P/R nor the HD locus is physically closely linked to its centromere in C. amylolentus, and the regions between the MAT loci and their respective centromeres show overall synteny between the 2 genomes, both MAT loci exhibit genetic linkage to their respective centromere during meiosis, suggesting the presence of recombinational suppressors and/or epistatic gene interactions in the MAT-CEN intervening regions. Furthermore, genomic comparisons between C. amylolentus and related pathogenic Cryptococcus species provide evidence that multiple chromosomal rearrangements mediated by intercentromeric recombination have occurred during descent of the 2 lineages from their common ancestor. Taken together, our findings support a model in which the evolution of the bipolar mating system was initiated by an ectopic recombination event mediated by similar repetitive centromeric DNA elements shared between chromosomes. This translocation brought the P/R and HD loci onto the same chromosome, and further chromosomal rearrangements then resulted in the 2 MAT loci becoming physically linked and eventually fusing to form the single contiguous MAT locus that is now extant in the pathogenic Cryptococcus species.

Fig 1. Genome comparison between the 2 Cryptococcus amylolentus isolates CBS6039 and CBS6273.

Shown here are results of dot plot visualization of alignments between the 2 C. amylolentus genomes with Nucmer maximum gap size set at 10. Data used to generate the figure can be found at NCBI BioProject with accession no. PRJNA200571 and EBI with study accession no. PRJEB19939.

Fig 2. Synteny map of the MAT loci in Cryptococcus amylolentus and closely related species.

Shown at the top and the bottom are results of synteny analyses between sequences from strains CBS6039 and CBS6273 for the HD and P/R loci, respectively. Red color highlights the genes that define the HD locus (SXI1 and SXI2) and P/R locus (mating pheromones and STE3); blue and gray colors highlight the genes that are present or absent from the mating type (MAT) locus in the human pathogenic Cryptococcus species complex, respectively. Shown in the middle are results of dot plot analyses between strains CBS6039 and CBS6273 for the chromosomal regions encompassing the genes that define the HD (SXI1 and SXI2) and P/R (mating pheromone and STE3) loci, respectively. Compared to the flanking regions showing complete synteny between strains CBS6039 and CBS6273, the regions highlighted in green (CBS6039 chromosome 11:155396–177087) and red (CBS6039 chromosome 10:624796–720709) exhibit significantly elevated levels of sequence divergence and chromosomal rearrangements, and define the HD and P/R loci, respectively, in C. amylolentus.

Fig 3. Identification and characterization of centromeres on each of the 14 chromosomes in the CBS6039 genome.

(A) Live cell direct fluorescence microscopy images of centromere binding protein Cse4 (CENP-A) at 3 different stages of the mitotic cycle. (B) Plots of read depths when mCherry-CENP-A chromatin immunoprecipitation sequencing (ChIP-seq) data were mapped against the CBS6039 genome assembly are presented. All of the centromeric regions identified in the CBS6039 genome (except for chromosome 1; see Results for more details) showed significantly higher read depth when compared to flanking non-centromeric regions (see S3 Fig for plots of whole chromosomes). Red plots (chipD) are based on signals obtained from ChIP-seq analysis, while blue plots (contD) indicate the negative control. (C) The diagram depicts the structures of the 6 unique centromere-specific Long Terminal Repeat (LTR) retrotransposons, Tcen (transposons in centromeres) 1–6, identified in the Cryptococcus amylolentus centromeric regions. While Tcen1 contains only LTRs (shown in grey), all of the other 5 Tcen elements consist of various genes/domains found in retrotransposons (RH, RNaseH; RT, Reverse Transcriptase; INT, Integrase). On the far right are the corresponding centromeres in the CBS6039 genome within which the full-length Tcen elements have been identified. (D) Schematic illustrating the distributions of the 6 Tcen elements, as well as their remnants, on the identified centromere regions in the CBS6039 genome. These intervals were defined as the longest ORF-free regions on the respective chromosomes and contain mostly retrotransposons or their remnants, and show enrichment of CENP-A binding based on ChIP-seq analyses. (E) RNA sequencing (RNA-seq) analysis reveals that the identified CBS6039 centromere regions also had reduced levels of transcriptional activity when compared to flanking non-centromeric regions. The blue bars indicate RNA-seq read depth. Please see S3 Table for coordinates of the centromeres in C. amylolentus.

Fig 4. Genome comparison between Cryptococcus amylolentus strain CBS6039 and Cryptococcus neoformans strain H99.

Shown here are distributions of BLAST hits in the CBS6039 genome, using protein sequences of the ORFs from each of the 14 chromosomes in the H99 genome as query. The x-axis shows the numerical order of the ORFs on each H99 chromosome; the y-axis illustrates the 14 chromosomes in the CBS6039 genome. The red vertical bars indicate locations of the centromeres in the H99 genome, the green dots indicate the presence of BLAST hits that are centromere-flanking in the CBS6039 genome, and the numbers beside the green dots indicate the CBS6039 chromosomes from which the centromere-flanking hits are located (see Results and S4 Fig for further details).

Fig 5. Model for the transition from tetrapolar to bipolar mating system organization.

(I) In the ancestor, the P/R and HD loci were located on different chromosomes, which had regional centromeres that shared common transposable/repetitive elements. (II) and (III) Ectopic recombination occurred between the 2 chromosomes within the centromeric regions, possibly mediated by the common transposable/repetitive elements, bringing the 2 mating type (MAT) loci onto the same chromosome. (IV) Subsequent chromosomal rearrangements (e.g., inversions and transpositions) bring the 2 MAT loci next to each other. (V) Eventually the P/R and HD loci fuse to form a single contiguous MAT locus that is present in the derived bipolar mating system. (VI) The resulting chromosome with the contiguous MAT locus could undergo additional intercentromeric recombination events. The numbers in the parentheses next to the “CEN” indicate the C. amylolentus chromosome on which those centromeric flanking regions are located.

Fig 6. Distribution and frequency of crossovers during sexual reproduction in C. amylolentus.

(A) The SNP distribution along chromosome 1 in meiotic progeny from a cross between strains CBS6039 and CBS6273 suggests that 1 meiotic event occurs per basidium during sexual reproduction in C. amylolentus. Blue color indicates SNPs that correspond to the genomic sequence of strain CBS6039, and red color indicates SNPs that correspond to the genomic sequence of strain CBS6273. Meiotic progeny from 2 individual basidia (#1 and #2), as well as 2 random basidiospores, were analyzed. For basidium #2, 2 additional basidiospores, #4 and #6, that are genetically identical to basidiospores #3 and #5, respectively, were also included. On the right are the number of estimated crossovers that occurred along chromosome 1 during meiosis in each progeny. (B) The frequencies of crossovers along each chromosome. The data are summarized from S5 Fig. (C) Percentage of progeny that had different numbers of crossovers along chromosomes 10 and 11. Data used to generate the figures can be found at NCBI BioProject with accession no. PRJNA200571 and at EBI with study accession no. PRJEB19939.

Fig 7. Meiotic recombination frequencies observed on chromosomes 10 and 11.

For both chromosomes 10 (A) and 11 (B), the top panel shows the recombination frequencies (cM/kb) at different locations along the chromosome, calculated based on the physical locations of the genetic markers on the chromosome (middle panel; see S5 Table for detailed information on the locations of the markers) and the genetic distances between markers that were estimated from the genetic linkage map (bottom panel). The green and red blocks indicate the locations of the mating type (MAT) loci (the P/R locus on chromosome 10 and the HD locus on chromosome 11) and the centromeres, respectively.