PerspectiveMolecular Biology

Putting the breaks on meiosis

See allHide authors and affiliations

Science  20 Nov 2015:
Vol. 350, Issue 6263, pp. 913
DOI: 10.1126/science.aad5404

Sexual reproduction involves the production of haploid gametes from diploid cells through a series of genome divisions, called meiosis. Accurate meiotic chromosome segregation requires homologous recombination, initiated by programmed DNA double-strand breaks (DSBs). DSBs are focused at sites, called hotspots, where recombination preferentially occurs (1). Because DSB repair by recombination involves copying information from an unbroken chromosome, one prevailing view is that, over time, DSB hotspots should be replaced by “cold” sequence variants that reduce DSBs at the same site (2). Consistent with this paradigm, hotspot patterns in mice and primates display high divergence between species and even individuals (3). This view is now challenged by two research articles in this issue, by Lam and Keeney (page 932) and Singhal et al. (page 928), that characterize recombination hotspot patterns in species of budding yeast (4) and birds (5), respectively. Both papers document remarkable hotspot pattern stability over evolutionary time, suggesting that the picture in mammals may be the exception rather than the rule.

Different mechanisms, different outcomes.

In mammals, meiotic recombination hotspots contain sequences where PRDM9 binds and forms chromatin marks (stars) that promote double-strand breaks. Both PRDM9 and its target sequences evolve rapidly. In yeast, birds, and other eukaryotes, breaks form at preexisting genomic elements with recombination-independent functions that ensure evolutionary stability—in this example, a nucleosome-depleted region (NDR) with a gene promoter that already contains break-promoting chromatin marks.


Lam and Keeney compared genome-wide meiotic DSB maps in Saccharomyces cerevisiae strains with roughly the same sequence divergence—a measure of separation in evolutionary time—as humans and chimpanzees, and in four Saccharomyces species, including a pair (S. cerevisiae and S. kudriavzevii) with roughly the same sequence divergence as mammals and birds. Hotspot patterns among S. cerevisiae strains showed almost complete overlap, and even the distantly related S. cerevisiae and S. kudriavzevii showed a greater than 80% overlap in hotspot locations. Even more remarkably, Lam and Keeney found that different species often displayed similar fine-structure break distributions and DSB frequencies at individual hotspots. Thus, in contrast to the rapid divergence seen in mammals, hotspot patterns in budding yeast show a high degree of conservation.

Singhal, Leffler, and colleagues derived meiotic recombination maps from single-nucleotide polymorphism distributions in small populations of two Australian birds, the zebra finch and long-tailed finch, again with sequence divergence similar to that between humans and chimpanzees. As with yeast, a large fraction of hotspots in the two finches (>70%) colocalized. Using elevated GC content, a hallmark of GC-biased gene conversion (6), as a signal for hotspots, the authors examined the ensemble of hotspots shared between the two finches in three other species: the Australian double-barred finch, the medium ground finch from the Galapagos, and the Old World collared flycatcher. Despite their broad geographical and evolutionary separation, all five species showed elevated GC content at these loci, consistent with the retention of a substantial subset of hotspots. Thus, in birds as in yeast, but unlike in mammals, meiotic recombination patterns appear to be highly conserved over broad swaths of evolutionary time.

The cause of this difference may be found in the different ways that yeast and mammals designate DSB hotspots (1). Yeast hotspots do not share specific sequences, but instead are located in the nucleosome-depleted regions (NDRs) already present in mitotic cells. NDRs contain gene promoters and are flanked by nucleosomes that contain histone H3 methylated at lysine 4 (H3K4me). H3K4me is implicated primarily in promoter function but is also used to recruit DSB-forming proteins during meiosis. In contrast, in most mammals, hotspots show no particular correlation with preexisting chromatin elements. Instead, they contain binding sites for positive-regulatory domain zinc finger protein 9 (PRDM9), a meiotic protein with a highly variable zinc finger array that recognizes specific sequences and that catalyzes the formation of H3K4me in its vicinity. Thus, yeast hotspots are likely maintained because their constituent elements perform other important functions. PRDM9-designated hotspots in mammals are under no such constraints, and thus can undergo rapid evolution in which hotspot evaporation is balanced by the rapid evolution of new binding specificity in PRDM9 (7). The finding that bird hotspots are also conserved suggests that birds, like yeast, designate hotspots using genomic elements that are under functional selection.

The existence of at least two different modes of hotspot designation, with accompanying differences in evolutionary stability, raises a question: If birds and yeast do it one way, and most mammals do it another way, what about other eukaryotes? In this regard, it is worth noting that hotspots also appear to be sequence-independent, and are associated with functional genomic elements, in plants (8); in dogs, which naturally lack a functional PRDM9 (9); and even in Prdm9−/− mutant mice (10). These observations, together with the exciting findings of Lam and Keeney and of Singhal et al., raise the intriguing possibility that the yeast and bird paradigm for hotspot designation may be the primordial one, and that the mammalian mechanism of sequence-based designation may be a relative latecomer to the game.


Navigate This Article