The devastating infectious disease chytridiomycosis has caused declines of amphibians across the globe, yet some populations are persisting and even recovering. One understudied effect of wildlife disease is changes in reproductive effort. Here we aimed to understand if disease has plastic effects on reproduction and if reproductive effort could evolve with disease endemism. We compared the effects of experimental pathogen exposure (trait plasticity) and population-level disease history (evolution in trait baseline) on reproductive effort using gametogenesis as a proxy in the declining and endangered frog Litoria verreauxii alpina. We found that unexposed males from disease-endemic populations had higher reproductive effort, which is consistent with an evolutionary response to chytridiomycosis. We also found evidence of trait plasticity, where males and females were affected differently by infection: pathogen exposed males had higher reproductive effort (larger testes), whereas females had reduced reproductive effort (smaller and fewer developed eggs) regardless of the population of origin. Infectious disease can cause plastic changes in reproductive effort at an individual level, and population-level disease exposure can result in changes to baseline reproductive effort; therefore, individual- and population-level effects of disease should be considered when designing management and conservation programs for threatened and declining species. --

Experimental design

We conducted a common garden experiment in which individuals from four populations (one disease-naïve: Grey Mare; and three disease-endemic: Eucumbene, Kiandra and Ogilvies) were collected as eggs from the wild, brought into captivity, and raised in a bio-secure, pathogen-free environment (Bataille et al. 2015; Grogan et al. 2018a,b,c). At approximately eight months after metamorphosis, we experimentally exposed the frogs to 750,000 B. dendrobatidis zoospores in 25mL of dilute salt solution for 18hrs, or mock exposed (exposed n=215, unexposed-controls n=39, see Table S1). Sample size between exposed and unexposed-control groups are not even because these animals were part of a larger clinical infection experiment (Bataille et al. 2015; Grogan et al. 2018a,b,c). We tested the animals for B. dendrobatidis infection every 14 days using quantitative PCR until the end of the experiment (first signs of morbidity or day 86 after inoculation) (see Text S1 husbandry details).

Sample fixation and processing

On the day that animals first showed clinical signs of infection (from day 19 onwards after inoculation) or at the end of the experiment if they remained healthy (day 86 after inoculation), we swabbed the animals for B. dendrobatidis infection, weighed them to the nearest 0.01g, humanely euthanized them with buffered Tricaine Methanesulfonate (MS-222), and fixed them in neutral buffered formalin. We used mass as a proxy for animal size (Text S1), and measured mass on the day of experimental pathogen exposure and on the day of euthanasia for all individuals.

We dissected the formalin fixed animals and removed the left gonad. We conducted visual gross examination blind to experimental pathogen exposure and population of origin. All males dissected had well developed testes, which indicated sexual maturity. We assigned females to one of three developmental categories based on gross examination of the ovaries: 1) small, underdeveloped ovaries, 2) mid-sized almost developed ovaries with some developed eggs, and 3) ovaries with developed eggs.

We processed a subset of those dissected individuals (16 unexposed females, 16 unexposed males, 21 exposed females and 23 exposed males, see Table S1) for histological examination following standard methods (Woods and Ellis 1994). We analysed the histological samples blinded to experimental exposure and population of origin.  

Gonad and gamete characteristics

We assessed six histosections per male individual to determine testis area, number of seminiferous tubules, density of seminiferous tubules, and germinal epithelium depth. We measured the largest tubule size and identified spermatogenesis cell clusters to stage (spermatogonia, spermatocytes, spermatids, spermatozoa) following Brannelly et al. (2016c), (see Text S1). For most of the characteristics measured, larger depth, area, or higher count indicates higher reproductive output. However, a lower density of seminiferous tubules and a lower proportion of spermatogonia indicates higher reproductive output: lower tubule density per unit of testis area allows for more spermatogenesis to occur due to increased intra-tubule space (i.e., more space inside the testis devoted to spermatogenesis) and is therefore associated with higher reproductive effort when spermatogenesis is active (Marion 1982). Because spermatogonia are an early stage of spermatogenesis, a lower proportion of spermatogonia cell clusters (i.e., more cells in later stages of development) are indicative of higher spermatogenesis effort (Delgado et al. 1992; Rastogi et al. 2005).

We assessed six histosections per female individual to determine ovary size, number of eggs, size of the largest eggs, oviduct width, density of eggs within the ovary, and oogenesis stages (Stage I – pre-vitellogenic; Stage II-V – vitellogenic; Stage VI – post-vitellogenic) (Text S1). Similar to the interpretation of the male gonad/gamete characteristic analyses, for females, a lower density of eggs per ovary correlates with larger average egg size, which indicates higher reproductive effort because larger egg cells are associated with later oogenesis. A smaller proportion of pre-vitellogenic eggs indicates an overall higher proportion of combined vitellogenic and post-vitellogenic eggs (i.e., eggs that have yolk development), which we interpreted as higher reproductive effort. In both males and females, we analysed gonad size relative to body size, rather than unscaled gonad size.