Studies of highly diverged species have revealed two mechanisms by which meiotic recombination is directed to the genome—through PRDM9 binding or by targeting promoter-like features—that lead to dramatically different evolutionary dynamics of hotspots. Here, we identify PRDM9 orthologs from genome and transcriptome data in 225 species. We find the complete PRDM9 ortholog across distantly related vertebrates but, despite this broad conservation, infer a minimum of six partial and three complete losses. Strikingly, taxa carrying the complete ortholog of PRDM9 are precisely those with rapid evolution of its predicted binding affinity, suggesting that all domains are necessary for directing recombination. Indeed, as we show, swordtail fish carrying only a partial but conserved ortholog share recombination properties with PRDM9 knock-outs. DOI: http://dx.doi.org/10.7554/eLife.24133.001, eLife digest The genetic information of Eukaryotic organisms (animals, plants and fungi) is encoded on strands of DNA called chromosomes. In animals that sexually reproduce, most cells carry two copies of each chromosome, with one inherited from each of their parents. Sex cells such as sperm and egg cells are the exception, and contain only a paternal or maternal set respectively. These chromosomes are not exact copies of the parental chromosomes, but are instead combinations of both of them, generated by a process called meiotic recombination. Meiotic recombination begins by breaking the chromosomes, and the repair of those breaks shuffles DNA segments between the two chromosomes. This shuffling is known as a “recombination event”. In humans, apes and mice, the location of recombination events depends on where a protein called PRDM9 binds to the DNA. Over the course of evolution, this binding location has changed relatively rapidly so that even closely related species such as humans and chimpanzees localize recombination events to different DNA regions. In contrast, closely related species that do not produce PRDM9 tend to direct recombination events to similar DNA regions. It remains unclear when PRDM9 first evolved its role in recombination, or why different methods of directing recombination have developed. To begin answering these questions, Baker, Schumer et al. investigated whether 225 species of vertebrates (backboned animals) have a gene that encodes PRDM9. This analysis revealed that even distantly related animals have genes that produce equivalents of the complete PRDM9 protein. However, several species have independently lost the ability to produce PRDM9. In certain other species, particular regions of the gene have been removed or shortened. Notably, only species that carry genes that contain regions called the KRAB and SSXRD domains show relatively rapid evolution of where PRDM9 binds in the DNA. To investigate this phenomenon further, Baker, Schumer et al. constructed a map of recombination events in swordtail fish, which carry a version of the gene that lacks the KRAB and SSXRD domains. The PRDM9 protein produced by this gene does not direct where recombination events occur. Overall, it appears that the KRAB and SSXRD domains are necessary for PRDM9 to direct meiotic recombination. Furthermore, Baker, Schumer et al. predict that those species that have complete versions of PRDM9 use this protein to localize recombination events. Knowing which species use PRDM9 in this way is the first step towards understanding why recombination mechanisms change in evolution, and with what consequences. DOI: http://dx.doi.org/10.7554/eLife.24133.002