Introduction
Ectothermic vertebrates are highly vulnerable to climate change and chemical pollution, because several aspects of individual development, including sex differentiation depend on environmental conditions in numerous species of reptiles, amphibians and fish (Bókony, Kövér, Nemesházi, Liker, & Székely, 2017; Bókony et al., 2018; Eggert, 2004; Holleley, Sarre, O’Meally, & Georges, 2016; Orton & Routledge, 2011; Ospina-Álvarez & Piferrer, 2008; Tamschick et al., 2016). In species with genetic sex determination, thermal and chemical disturbances during embryonic or larval development can cause sex reversal, meaning that genetically female individuals become phenotypic males, or vice versa (Eggert, 2004; Holleley et al., 2016; Ospina-Álvarez & Piferrer, 2008). Laboratory experiments show that sex-reversed individuals of some species may have reduced reproductive success (Harris et al., 2011; Senior, Nat Lim, & Nakagawa, 2012), and theoretical studies suggest that sex reversals may lead to serious consequences for natural populations, including changes in genetic variability, distorted sex ratios, and even extinction (Bókony et al., 2017; Quinn, Sarre, Ezaz, Marshall Graves, & Georges, 2011; Wedekind, 2017). Therefore, it is imperative to gain information on the prevalence and fitness of sex-reversed individuals in natural populations, to be able to assess and forecast the effects of anthropogenic environmental changes.
For studying sex reversal, one needs to identify not only the phenotypic sex but also the genetic sex of each individual. The latter can be especially difficult in non-model organisms, due to lack of information on sex-linked DNA sequences. Because of their highly conserved sex chromosome system, universal sex-linked DNA markers have long been available for birds and mammals (Fridolfsson & Ellegren, 1999; Griffiths & Tiwari, 1993; Shaw, Wilson, & White, 2003), making molecular sexing a routine in these taxa. However, in the majority of ectothermic vertebrates, sex chromosome turnover (i.e. the swapping of the chromosome used for genetic sex determination) is common and the sex chromosomes of many species are homomorphic (Devlin & Nagahama, 2002; Holleley et al., 2016; Jeffries et al., 2018; Miura, 2017). Consequently, there is often little homologous sex-linked variation between and sometimes even within species, making molecular sexing challenging (Ezaz, Stiglec, Veyrunes, & Marshall Graves, 2006; Perrin, 2009; Stöck et al., 2013). Furthermore, type of sex-chromosome system (i.e. male or female heterogamety) can differ between closely related species or even between different populations of the same species, especially in amphibians (Holleley et al., 2015; Rodrigues et al., 2017; Sarre, Ezaz, & Georges, 2011).
For the above reasons, genetic sexing methods need to be developed and validated species by species in amphibians. Recombination between the sex chromosomes (Ezaz et al., 2006; Perrin, 2009; Stöck et al., 2013) is expected to be reduced in the vicinity of the ‘master sex-determination gene’ (Bachtrog, 2006; Bachtrog et al., 2014; van Doorn & Kirkpatrick, 2007), providing a preferential target for sex marker development. Unfortunately, the master sex-determination gene remains elusive in all but a few amphibian species (Eggert, 2004; Miura, 2017; Nakamura, 2013; Yoshimoto et al., 2010), and the size of the non-recombining region around it can be small. Thus, in order to find markers which make reliable identification of the sex chromosomes possible in the species of interest, researchers must test high numbers of loci across the genome (Lambert, Skelly, & Ezaz, 2016; Olmstead, Lindberg-Livingston, & Degitz, 2010; Stöck et al., 2011). Owing to these challenges, reliable sex-linked markers only exist for a handful of amphibian species so far (Alho, Matsuba, & Merilä, 2010; Berset-Brändli, Jaquiéry, Dubey, & Perrin, 2006; Brelsford, Lavanchy, Sermier, Rausch, & Perrin, 2017; Eggert, 2004; Lambert et al., 2016; Ma, Rodrigues, Sermier, Brelsford, & Perrin, 2016; Olmstead et al., 2010; Rodrigues et al., 2017; Stöck et al., 2011).
Due to this general lack of sex markers, we know troublingly little about sex reversals in nature: how widespread they are, which environmental factors they are associated with, and how they affect individual fitness and population viability. To our knowledge, the frequency of sex reversal in the wild has been published for only two amphibian species so far: 9% of genetic females were phenotypically male in a Finnish common frog (Rana temporaria ) population, while 8.5% female-to-male and 3% male-to-female sex reversal was found in green frogs (Rana clamitans ) in the USA (Alho et al., 2010; Lambert, Tran, Kilian, Ezaz, & Skelly, 2019).
In this study, we investigated sex reversals in the agile frog(Rana dalmatina) . This species is widespread in Europe, but its population sizes show a decreasing tendency (Kaya et al., 2009). It inhabits light deciduous woodlands, but also occurs near or in urbanized areas. Similarly to most Rana species, its diploid karyotype consists of 26 chromosomes (Spasić-Bošković, Tanić, Blagojević, & Vujošević, 1997); its sex chromosomes were identified only recently, showing a male-heterogametic (XX/XY) sex-determination system (Jeffries et al., 2018). Because no molecular sexing method has been published for agile frogs yet, first we searched for sex-linked markers using an existing Restriction Site Associated sequencing (RADseq) dataset (Jeffries et al., 2018) and validated them to provide a reliable genetic sexing method for this species. Subsequently, we studied the occurrence of sex reversals in wild agile frog populations in North-Central Hungary, and tested if sex reversals are more common in populations associated with anthropogenic land use. Finally, we examined if sex reversal was associated with fitness costs by comparing fitness-related traits between sex-reversed and normal individuals.