Searching for sex-reversals to explain population demography and the evolution of sex chromosomes





Sex determination can be purely genetic (as in mammals and birds), purely environmental (as in many reptiles), or genetic but reversible by environmental factors during a sensitive period in life, as in many fish and amphibians (Wallace et al. 1999; Baroiller et al. 2009a; Stelkens & Wedekind 2010). Such environmental sex reversal (ESR) can be induced, for example, by temperature changes or by exposure to hormone-active substances. ESR has long been recognized as a means to produce more profitable single-sex cultures in fish farms (Cnaani & Levavi-Sivan 2009), but we know very little about its prevalence in the wild. Obviously, induced feminization or masculinization may immediately distort population sex ratios, and distorted sex ratios are indeed reported from some amphibian and fish populations (Olsen et al. 2006; Alho et al. 2008; Brykov et al. 2008). However, sex ratios can also be skewed by, for example, segregation distorters or sex-specific mortality. Demonstrating ESR in the wild therefore requires the identification of sex-linked genetic markers (in the absence of heteromorphic sex chromosomes) followed by comparison of genotypes and phenotypes, or experimental crosses with individuals who seem sex reversed, followed by sexing of offspring after rearing under non-ESR conditions and at low mortality. In this issue, Alho et al. (2010) investigate the role of ESR in the common frog (Rana temporaria) and a population that has a distorted adult sex ratio. They developed new sex-linked microsatellite markers and tested wild-caught male and female adults for potential mismatches between phenotype and genotype. They found a significant proportion of phenotypic males with a female genotype. This suggests environmental masculinization, here with a prevalence of 9%. The authors then tested whether XX males naturally reproduce with XX females. They collected egg clutches and found that some had indeed a primary sex ratio of 100% daughters. Other clutches seemed to result from multi-male fertilizations of which at least one male had the female genotype. These results suggest that sex-reversed individuals affect the sex ratio in the following generation. But how relevant is ESR if its prevalence is rather low, and what are the implications of successful reproduction of sex-reversed individuals in the wild?

ESR in the wild could have several major consequences, even at low prevalence. First, ESR may profoundly impact the evolution of chromosomes, because what happens to a genome is strongly influenced by the phenotypic sex. The phenotypic sexes are under different kinds of natural and sexual selection (Moore & Moore 2006), they differ in their mutation rates (Ellegren 2007; Cotton & Wedekind 2010), and there are sex-specific patterns and rates of recombination (Burt et al. 1991; Isberg et al. 2006; Berset-Brändli et al. 2008). The latter could even be demonstrated in sex-reversed amphibians and fish (reviewed in Perrin 2009). Recombination between X and Y (or W and Z, respectively) is to a large degree suppressed in males (Charlesworth & Charlesworth 2000) but seems often possible in sex-reversed individuals. Structural differences may prevent recombination between heteromorphic sex chromosomes in some species. However, heteromorphic sex chromosomes are not very frequent in lower vertebrates (Schartl 2004). In fact, 96% of all amphibians seem to have homomorphic sex chromosomes (Perrin 2009). If they allow for recombination between sex chromosomes in sex-reversed individuals, recombination followed by natural and sexual selection would prevent the evolutionary decay of Y and W chromosomes. Sex reversal would then be a ‘fountain of youth’ for these chromosomes (Perrin 2009).

Second, ESR often leads to varying population sex ratios, and skewed sex ratios can favour turnovers of sex chromosomes (Mank & Avise 2009), i.e. replacement of old sex chromosomes with autosomes or a shift between a XY and a ZW system. Such transitions require that the Y (or the W, respectively) is not considerably degenerated, which is often the case in species with ESR (Cnaani & Levavi-Sivan 2009) and potentially explained by Perrin’s (2009) hypothesis. Indeed, transitions between sex determining mechanisms seem to have recently happened in some fish and amphibians (Miura 2007; Baroiller et al. 2009b; Schultheis et al. 2009).

Third, even if it is obvious that ESR can immediately influence the population sex ratio, the long-term effects of ESR on demographics and population growth are not always obvious and need to be modelled. Different kinds of long-term effects can result from correlated shifts in genotypic frequencies. For example, the frequency of XY males can decline rapidly already at some modest pressure of masculinization, up to the point where the Y chromosome disappears entirely from the population (Hurley et al. 2004; Cotton & Wedekind 2009). Such populations are then vulnerable to extinction even if a large census size suggests otherwise, because a suddenly altered environment may no longer provide the stimuli required for ESR to occur, e.g. in the event of a clean up of effluents. Moderate environmental feminization can increase population growth rate, but such populations are still threatened. If the ESR ceases, sex-reversed XY or YY individuals can produce strongly male-biased sex ratios in the following generations. This effect of feminization has even been suggested as a potential control of introduced exotic species (Gutierrez & Teem 2006; Cotton & Wedekind 2007).

Alho and colleagues used their findings to predict the development of the sex ratio of their study population over a period of 30 years. Building on existing and rather general models, they incorporated several aspects that are specific to their study population, e.g. male and female age at maturation. They found that a masculinization of 9% in a population like theirs would lead, on the long term and depending on the model settings, to either a modest female bias or an excess of phenotypic males. However, there is currently a distinct excess of females in the population. This discrepancy between prediction and observation remains unexplained but is not necessarily surprising. In general, there are a number of possible pitfalls that complicate the search for ESR or the demonstration of a link between ESR and population demography in the wild (Chowen & Nagler 2004; Williamson & May 2005; Alho et al. 2010; Stelkens & Wedekind 2010). For example, temporal variation in the intensity or direction of ESR can produce short-term and long-term effects that are hard to predict. Also, the applicability of genetic markers may well be population specific, and the possibility of null alleles or of recombination between X and Y needs to be appropriately dealt with. Inheritance studies to confirm the diagnostic value of sex-linked markers can easily be confounded by sex-specific mortalities before offspring can be sexed, and even if offspring mortality can be kept low in such experiments, the rearing conditions should take the possibility of genetic variation for reaction norms into account (Conover & Van Voorhees 1990; Magerhans et al. 2009). Such genetic variation could induce strong evolutionary dynamics and may even allow populations to adapt to environmental changes that would otherwise increase the rate of ESR (Conover & Van Voorhees 1990).

  • image(1)

[  Female common frog (Rana temporaria). Photo taken by Juha Merilä. ]

ESR is likely to play a key role in shaping the genetics, ecology, and evolution of sex determination in many species. A better understanding of ESR in the wild may therefore help us to answer some important evolutionary questions (Wedekind & Stelkens 2010). How have different sex determination mechanisms evolved, when did that happen, why does this often lead to a loss of recombination and a genetic decay of the Y or W in the heterogametic sex, and why are there still relatively undamaged Y and W chromosomes in so many species? We also need to know what kind of anthropogenetic effects induce ESR in the wild and how populations respond. What is the relevance of, for example, changing temperatures? The current prevalence of ESR may well differ from the one before anthropogenetic effects became important, and we clearly need good estimates of both frequencies. We need studies like Alho et al.’s and further build on them to fully understand the evolutionary relevance of ESR and to better protect vulnerable populations.

C.W. studies the selective forces that currently act on fish in the wild, including sexual selection, host-pathogen interactions, and the direct and indirect effects of human activities (e.g. non-random harvesting, supplementary breeding, species invasions, or climate change). His primary empirical systems are salmonid fishes. He also studies the evolution of cooperation in conflict situations.