Selection coefficients at the mammalian Y chromosome typically do not deviate strongly from neutrality. Here we show that strong balancing selection, maintaining intermediate frequencies of DNA sequence variants, acts on the Y chromosome in two populations of African buffalo (Syncerus caffer). Significant correlations exist between sequence variant frequencies and annual rainfall in the years before conception, with five- to eightfold frequency changes over short time periods. Annual rainfall variation drives the balancing of sequence variant frequencies, probably by affecting parental condition. We conclude that sequence variants confer improved male reproductive success after either dry or wet years, making the population composition and dynamics very sensitive to climate change. The mammalian Y chromosome, interacting with ecological processes, may affect male reproductive success much more strongly than previously thought.

Abstract Background The Y-chromosomal diversity in the African buffalo (Syncerus caffer) population of Kruger National Park (KNP) is characterized by rainfall-driven haplotype frequency shifts between year cohorts. Stable Y-chromosomal polymorphism is difficult to reconcile with haplotype frequency variations without assuming frequency-dependent selection or specific interactions in the population dynamics of X- and Y-chromosomal genes, since otherwise the fittest haplotype would inevitably sweep to fixation. Stable Y-chromosomal polymorphism due one of these factors only seems possible when there are Y-chromosomal distorters of an equal sex ratio, which act by negatively affecting X-gametes, or Y-chromosomal suppressors of a female-biased sex ratio. These sex-ratio (SR) genes modify (suppress) gamete transmission in their own favour at a fitness cost, allowing for stable polymorphism. Results Here we show temporal correlations between Y-chromosomal haplotype frequencies and foetal sex ratios in the KNP buffalo population, suggesting SR genes. Frequencies varied by a factor of five; too high to be alternatively explained by Y-chromosomal effects on pregnancy loss. Sex ratios were male-biased during wet and female-biased during dry periods (male proportion: 0.47-0.53), seasonally and annually. Both wet and dry periods were associated with a specific haplotype indicating a SR distorter and SR suppressor, respectively. Conclusions The distinctive properties suggested for explaining Y-chromosomal polymorphism in African buffalo may not be restricted to this species alone. SR genes may play a broader and largely overlooked role in mammalian sex-ratio variation.