Mechanisms of chromosome number evolution in yeast

Abstract

The whole-genome duplication (WGD) that occurred during yeast evolution changed the basal number of chromosomes from 8 to 16. However, the number of chromosomes in post-WGD species now ranges between 10 and 16, and the number in non-WGD species (Zygosaccharomyces, Kluyveromyces, Lachancea, and Ashbya) ranges between 6 and 8. To study the mechanism by which chromosome number changes, we traced the ancestry of centromeres and telomeres in each species. We observe only two mechanisms by which the number of chromosomes has decreased, as indicated by the loss of a centromere. The most frequent mechanism, seen 8 times, is telomere-to-telomere fusion between two chromosomes with the concomitant death of one centromere. The other mechanism, seen once, involves the breakage of a chromosome at its centromere, followed by the fusion of the two arms to the telomeres of two other chromosomes. The only mechanism by which chromosome number has increased in these species is WGD. Translocations and inversions have cycled telomere locations, internalizing some previously telomeric genes and creating novel telomeric locations. Comparison of centromere structures shows that the length of the CDEII region is variable between species but uniform within species. We trace the complete rearrangement history of the Lachancea kluyveri genome since its common ancestor with Saccharomyces and propose that its exceptionally low level of rearrangement is a consequence of the loss of the non-homologous end joining (NHEJ) DNA repair pathway in this species.

Figure 1. Phylogeny of the Saccharomycetaceae species used in this study.

Parentheses show the numbers of chromosomes in extant species, and the inferred numbers at nodes in the tree. Negative numbers in red show chromosome number reductions. The black dot indicates the position of the WGD and the Ancestral genome sequence. Node ‘B’ is an older node that is the common ancestor of all non-WGD and post-WGD species. Lowercase letters represent specific rearrangements that differentiate L. kluyveri from the WGD Ancestor (black dot) as shown in Figure 2. Species whose names are underlined are those for which end-to-end complete chromosome sequences are available. The phylogeny used is that of Hedtke et al .

Figure 2. Cartoon showing the rearrangements indicated by lowercase letters in Figure 1 .

Monocolored chromosomes belong to the WGD Ancestor. Chromosomes in gray boxes are extant L. kluyveri chromosomes. Events encircled by a color correspond to events on branches of the same color in Figure 1. Black crossed lines between chromosomes represent points of interchromosomal translocations, and square brackets along chromosomes (events c, f and h) represent inversions. Arrows point to the products resulting from each rearrangement. The rearrangement for event o (marked with two asterisks) is not shown as it involves a reciprocal translocation located one gene from the edge of the Ancestral inference, which essentially swaps the telomeres of Anc3 and Anc8 at the ends of Lklu3 and Lklu4.

Figure 3. Progression of rearrangements and chromosome fusions leading to the loss of a centromere in Z. rouxii.

Two non-reciprocal telomeric translocations and a telomere-to-telomere fusion gave rise to the extant chromosome structures in Z. rouxii. Chromosomes in green boxes are those that underwent rearrangements, while those in gray boxes are finished translocation products (i.e., extant regions in Z. rouxii). The edges of the breakpoints are labelled with both the Ancestral and current Z. rouxii gene names. In the bottom step, the loss of a centromere occured contemporaneously with the two chromosomes fusing at their telomeres. All three rearrangements led to the internalisation of previously telomeric genes. The panels on the right show details of the gene orders and internalized telomeric genes at the junctions.

Figure 4. Loss of a centromere in A. gossypii by the breakage of a chromosome at its centromere.

The green chromosome at the top represents chromosome 5 at Node ‘B’ of the tree (Figure 1), which is identical to chromosome 5 of the WGD Ancestor (see Figure 2). After A. gossypii diverged from K. lactis, this chromosome broke in the intergenic region containing its centromere. To avoid losing large numbers of genes during cell division, both arms of the split chromosome fused their broken edges to the telomeres of Ancestral chromosomes 6 and 8, which gave rise to the organisation on the extant A. gossypii chromosomes 1 and 3. The timing of loss of the centromere is unclear: it may have happened as a part of the rearrangement, or the centromere may have been carried on one of the chromosome arms and lost after fusion to the telomere of another centromere-containing chromosome. The mechanism of the fission event is also ambiguous: it may have occurred by the chromosome actually breaking into two, or by two separate translocations to other chromosome ends that separated the centromere from its neighboring genes.

Figure 5. CDE conservation in the Saccharomycetaceae.

(A) Sequence logo showing base frequencies at each position in all annotated CDEI and CDEIII regions from 10 species. (B) Rate of chromosome loss per mitotic cell division caused by mutagenesis of individual residues in CDEI and CDEIII sequences (gray letters) of S. cerevisiae CEN6 (redrawn from [61]). Sites conserved in the logo tend to have the largest effects on chromosome loss when mutated. (C) Variation of CDEII lengths in species with identifiable point centromeres. The number of points is fewer than the number of chromosomes in each species because some chomosomes have identical CDEII length.