High chytrid prevalence and infection intensities in tadpoles of Mixophyes fleayi

Introduction

The global invasion of the pathogenic amphibian chytrid fungus (Batrachochytrium dendrobatidis, Bd) has strongly affected frog species around the world (Berger et al. 1998; Scheele et al. 2019). However, where some species have been decimated or even driven extinct, others appear to incur fewer costs owing to the fungus. Moreover, different life stages present heterogeneity in vulnerability to mortality; where post-metamorphic stages are frequently cited as being more vulnerable, particularly around metamorphosis itself, larval stages (hereafter interchangeably used with tadpoles) experience fewer detrimental effects (Garner et al. 2009; Ortiz-Santaliestra et al. 2013; Sauer et al. 2020).

Although Bd infections only rarely directly lead to mortality in tadpoles (Blaustein et al. 2005; Arellano et al. 2017), infections are still likely to affect them negatively by limiting feeding ability (Cashins 2009; Venesky et al. 2009; Hagman and Alford 2015; Harjoe et al. 2022), possibly reducing growth and size at metamorphosis (Garner et al. 2009), which may affect individuals post-metamorphosis (Altwegg and Reyer 2003; Garner et al. 2009; Kilpatrick et al. 2010; Sauer et al. 2020). Given the immune system reorganisation that takes place at metamorphosis (Rollins-Smith 1998), highly infected tadpoles may experience considerable costs around this stage, and larval Bd infections may further reduce fitness of these individuals when they metamorphose (Garner et al. 2009; Briggs et al. 2010; Fernández-Loras et al. 2017; Sauer et al. 2020). Because the fungus infects only keratinised body parts, which in tadpoles only occur in mouthparts, dekeratinisation of oral structures (hereafter interchangeably used with mouthpart loss) caused by Bd may be the mechanism that negatively affects larval feeding ability (Marantelli et al. 2004; Cashins 2009; Venesky et al. 2009; Harjoe et al. 2022). For example, torrent-adapted tadpoles in tropical streams were reported to have decreased body condition with increased oral dekeratinisation, with body condition improving on the regrowth of oral structures (Cashins 2009). On metamorphosis, the distribution of keratin shifts from the mouthparts to the skin, which rapidly becomes infected with Bd on development (Marantelli et al. 2004; McMahon and Rohr 2015). Dekeratinisation of the oral structures, in particular when it is observed in the jaw sheaths, has frequently been noted as an indicator of Bd infection in tadpoles (Fellers et al. 2001; Rachowicz 2002; Marantelli et al. 2004; Blaustein et al. 2005; Knapp and Morgan 2006; Drake et al. 2007; Cashins 2009). However, the extent to which dekeratinisation is a good indicator appears to be species-specific (Navarro-Lozano et al. 2018).

In addition to the costs accrued during the larval phase, tadpoles may act as a reservoir for Bd in ecosystems when they occur in permanent water bodies, particularly when tadpoles are long-lived (Rachowicz and Vredenburg 2004; Briggs et al. 2010; Catenazzi et al. 2013; Valencia-Aguilar et al. 2016; Courtois et al. 2017; Sapsford et al. 2018; das Neves-da-Silva et al. 2021). Reservoirs, i.e. hosts that permanently maintain the pathogen and transmit it to other hosts (Haydon et al. 2002), can greatly influence disease impact by increasing the exposure to susceptible individuals (Gog et al. 2002; De Castro and Bolker 2004). When terrestrial habitats seasonally warm up and dry out, becoming unfavourable for Bd growth (Piotrowski et al. 2004), the prevalence and intensity (measured as pathogen load) of Bd infections frequently decreases in adult frogs (Kriger and Hero 2007). However, aquatic environments may maintain lower temperatures and the requisite moisture requirements for Bd growth, facilitating the maintenance of Bd in ecosystems (Sapsford et al. 2018). Tadpoles may therefore remain infected with (and exposed to) Bd under conditions that are favourable to the fungus for long periods of time, and have a higher infection prevalence than do post-metamorphic individuals.

Previous research has found associations between Bd infections, tadpole body size, and temperature (Cashins 2009; Catenazzi et al. 2013; Valencia-Aguilar et al. 2016; Sapsford et al. 2018). Larger tadpoles are reportedly more likely to be infected with Bd, likely owing to the cumulative exposure older tadpoles have accrued to Bd zoospores in the water over time (Valencia-Aguilar et al. 2016; Sapsford et al. 2018). Infections with Bd tend to be more common, and infection intensities higher, with lower water temperatures, likely owing to the thermal preferences of Bd (Sapsford et al. 2018). Although both temperature and body size appear to be influential variables, ambiguity remains about the relative contributions of each on both infection prevalence and intensity in tadpoles (Sapsford et al. 2018; das Neves-da-Silva et al. 2021).

In this study, we assessed the seasonal Bd infection prevalence and infection intensity of tadpoles over 2 years in two independent streams. We focused on tadpoles of Mixophyes fleayi (Fleay’s barred frog), an endangered rainforest species for which the Bd prevalence, intensity, and impacts on mortality have been intensively assessed in both adult and recently metamorphosed cohorts (Hollanders et al. 2023a, 2023b). We built a modified version of a recently developed Bayesian hierarchical model to estimate the effect of temperature, body size, and their interaction on Bd infections. We also assessed to what extent mouthpart loss can be predicted by Bd infection and the uncertainty associated with using mouthpart loss as a proxy for Bd infection. Our results provide valuable insight into Bd dynamics of a long-lived tadpole in subtropical stream environments.

Materials and methods

Field surveys

Mixophyes fleayi (Fleay’s barred frog) is a large (male snout–urostyle lengths [SUL] ± 60 mm, female SUL ± 80 mm), terrestrial, stream-breeding frog endemic to the Gondwanan rainforests of northern New South Wales and South East Queensland, Australia. The tadpoles of M. fleayi are bottom-dwellers that graze the stream substrates, and tadpoles are likely to use their oral disc to adhere to the bottom of fast-flowing stream sections (Anstis 2002). We conducted M. fleayi tadpole surveys at 6-weekly intervals from September 2019 to January 2021 at Brindle Creek, Border Ranges National Park, and Tuntable Creek, Nightcap National Park, New South Wales, Australia. For details on sites, see Hollanders et al. (2023b). At Brindle Creek, we sampled from two adjacent pools separated by a riffle section where adult frogs were known to breed. At Tuntable Creek, we sampled from four pools separated by riffle sections. We used a dipnet (Jonah’s Aquarium The Perfect Dipnet, 1/8” mesh) to sample each pool for 1 min.

Tadpoles were held in a plastic tub and up to 40 individuals were processed (photographed and sampled for Bd) per site per occasion. For 1 year at Brindle Creek, we took a top-down photograph of the full tub of tadpoles (prior to subsampling 40 for processing), so as to measure the length of each tadpole (from the tip of the nose to the base of the tail) by using the Ruler Tool in Adobe Photoshop 2023 to gauge the distribution of size classes (Fig. 1). We handled tadpoles following Cashins (2009), photographed their mouthparts (Nikon D750 and Sigma 105 mm macro at minimum focus), and sampled for Bd by using swabs (five horizontal strokes and five vertical strokes across the mouthparts). For 620 individuals sampled over the 2 years, we photographed mouthparts to identify mouthpart loss associated with Bd and to derive a body size proxy, for which we used the distance between the two mouth corners measured in pixels using the Ruler Tool in Adobe Photoshop 2020. We chose to measure body size indirectly by using this method to reduce handling time of the animals and to get a more precise measurement than when measuring body length in-hand by using calipers. We verified the method by comparing body-length measurements by using calipers with mouthpart widths measured using Photoshop of 136 individuals and found strong agreement between the two measurements (Supplementary material Fig. S1). We identified mouthpart loss by assessing for dekeratinisation of the jaw sheaths from photographs, and scored the amount of dekeratinisation of both the bottom (posterior) and top (anterior) jaw sheath separately on a scale of 1–5, where 1 was barely visible mouthpart loss and 5 was complete dekeratinisation (i.e. no black mouthparts visible) (Fig. 2).

Fig. 1.

Size distributions of Mixophyes fleayi tadpoles per survey at Brindle Creek over 1 year.

Fig. 2.

Various stages of oral dekeratinisation in tadpoles of Mixophyes fleayi. The oral dekeratinisation scores of top/bottom (anterior/posterior) jaw sheaths were (a) 0/0, (b) 5/3, (c) 5/3, and (d) 5/4.

We placed dataloggers (HOBO MX2201) along the edge of the stream to record temperature every 2 h for the duration of the study. We relied on terrestrial dataloggers for our analysis, because dataloggers in the creeks were swept away by floods.

Results

Bd infection prevalence

Average Bd prevalence was estimated as 0.58 [95% HPDI: 0.432, 0.713], with no significant site effect (log odds change 0 [95% HPDI: −1.009, 0.828], RJMCMC inclusion 0.551) (Table 1). Individual probability of being infected with Bd was strongly negatively associated with temperature (log odds change −3.271 [95% HPDI: −4.616, −1.943], RJMCMC inclusion 1) and strongly positively with body size (log odds change 3.966 [95% HPDI: 3.129, 4.891], RJMCMC inclusion 1) (Table 1, Fig. 4). We found weak evidence for interaction effects, with low RJMCMC inclusion probabilities and effect estimates overlapping 0 (Table 1). The Bayesian P-value of the prevalence component of the model was 0.293, indicating reasonable fit (Fig. S2, S3a). Model fit improved significantly with the addition of the random survey effect (comparison not shown), which was large (standard deviation [s.d.] of the effect 1.257 [95% HPDI: 0.788, 1.853]), suggesting a considerable amount of unexplained variation among surveys (Fig. 3a). The survey effects were somewhat positively correlated with those of infection intensity (correlation of 0.235 [95% HPDI: −0.307, 0.727]), suggesting that when infections were common the associated infection intensities were also high. The probability of detecting Bd by using qPCR was high (probability of detecting one log₁₀ gene copy in a qPCR replicate estimated as 0.619 [95% HPDI: 0.567, 0.677], yielding a 0.999 [95% HPDI: 0.999, 1] probability of detecting the average load after two qPCR replicates), reflected in that the model estimated that 460 individuals were infected compared with the 471 infections that we observed. The false-positive probability of detecting Bd in the qPCR protocol was estimated to be low (0.016 [95% HPDI: 0.003, 0.029]), but note that Bd detection errors in the swabbing process were not modelled because only single swabs were collected.

Fig. 3.

Survey-specific estimates (medians and 95% HPDIs) of (a) Bd infection prevalence, (b) average Bd infection intensity (log₁₀ gene copies per swab), (c) oral dekeratinisation prevalence, and (d) average oral dekeratinisation score of tadpoles with dekeratinisation. Estimates were averaged over individual-level predictions for each survey.

Discussion

We investigated the occurrence of Bd and associated infection intensities in endangered Mixophyes fleayi tadpoles over 2 years, and found that both average infection prevalence (0.58 [95% HPDI: 0.43, 0.71]) and infection intensity (3.88 [3.68, 4.07] log₁₀ gene copies per swab) were high. Compared with adults from the same sites reported in previous studies, which had an average reported prevalence of 0.15 [95% HPDI: 0.07, 0.27] and intensity of 3.036 [95% HPDI: 2.836, 3.234] (Hollanders et al. 2023b), the odds of tadpoles being infected with Bd were approximately eight times greater and average log₁₀ infection intensities were 1.28 times greater. We found evidence of seasonality for Bd infections in tadpoles, where lower temperatures were associated with higher Bd prevalence (but not Bd infection intensity). Additionally, we corroborated evidence for risk of infection increasing cumulatively with age, because we found a strong association between body size and infection prevalence. It is noteworthy that recently metamorphosed individuals of M. fleayi at Brindle Creek have a lower average infection prevalence (estimated at 0.42 [95% HPDI: 0.13, 0.7], after correcting for imperfect detection in the swabbing process) and lower infection intensities (estimated at 2.8 [2.45, 3.08] log₁₀ gene copies per swab) than do the tadpoles (Hollanders et al. 2023a). This discrepancy of infection patterns between larval and post-metamorphic individuals sharing a spatio-temporal space warrants further investigation and suggests considerable differences in immunocompetence, pathogen resistance, exposure history, or other mechanisms of persistence (Brannelly et al. 2021).

Individual Bd infections of tadpoles were associated with lower temperatures and larger body sizes, consistent with results from previous studies (Fig. 4) (Cashins 2009; Catenazzi et al. 2013; Valencia-Aguilar et al. 2016; Sapsford et al. 2018). Low temperatures also correlated with increased infection prevalence with adult amphibians at the same site, consistent with the regional pattern previously reported (Kriger and Hero 2007; Hollanders et al. 2023b). The average 6-weekly temperatures at the sites ranged from 8.6 to 19.8°C, which is within the optimal range for Bd growth; nevertheless, growth is reportedly faster at ~20°C than it is at ~10°C, so Bd physiology may explain the observed pattern of infection prevalence at our sites (Piotrowski et al. 2004; Stevenson et al. 2013). However, thermal optima vary among Bd strains and the performance of the local strain is unknown (Voyles et al. 2017). Alternatively, immune function of tadpoles may be improved with higher temperatures (Cohen et al. 2017). Although tadpole immune response to Bd is poorly understood, previous studies have demonstrated that low-protein diets increase the probability of gaining Bd infections (Venesky et al. 2012) and tadpoles kept at lower temperatures have higher infection intensities (Altman and Raffel 2019), suggesting that resistance to Bd is influenced by external factors. Further, some studies have reported regrowth of oral structures after initial dekeratinisation as a result of Bd infection, suggesting possible tolerance mechanisms after infection (Cashins 2009; Hagman and Alford 2015). However, if tadpoles mounted successful immune responses to Bd infection, lower probability of being infected would be expected with a larger body size, and/or lower infection intensities would be associated with these remaining infections. Contrary to this, we found evidence for infection status being strongly predicted by body size (a proxy for age), in line with previous studies (Valencia-Aguilar et al. 2016; Sapsford et al. 2018). This result is consistent with tadpoles experiencing a cumulative exposure risk and limited immune response to suppress Bd infections in the field.

Fig. 4.

Prediction curves (medians and 95% HPDIs) with the coefficients of (a) temperature and (b) body size on the probability of Mixophyes fleayi tadpoles being infected with Bd from logistic regression, with the other (scaled) predictors held at their respective means. Points are observed individuals with corresponding (a) temperature and (b) body size at observation, categorised by Bd infection status (infected, top; uninfected, bottom). Body sizes were transformed from mouthpart widths by using an equation (Fig. S2).

Unlike infection status, we did not find strong predictors influencing infection intensity. RJMCMC excluded all predictors except body size, which was found to have a weakly positive effect on infection intensity (Fig. S2); however, this may simply indicate that larger tadpoles have larger oral structures to swab. The lack of evidence for a temperature effect was noteworthy, given the strength of the effect of temperature on infection status. This may indicate that Bd growth is largely independent of the range of temperatures that occur in these streams once Bd becomes established on a tadpole. Tadpoles did have considerably higher average infection intensities than did adults (median of 3.88 vs 3.0 log₁₀ gene copies), which may in fact underestimate the difference, given that tadpoles have only their oral structures swabbed (10 strokes) compared with the entire ventral surface of post-metamorphic animals (45 strokes, Hollanders et al. 2023b). That tadpoles in populations where adults display limited susceptibility to Bd after initial epidemics have higher infection intensities than do adult frogs has also been found in Rana sierrae/muscosa in the United States (Briggs et al. 2010). One explanation is that tadpoles exist in an environment that is more conducive to Bd growth because of the moisture and lower temperatures (Piotrowski et al. 2004). Alternatively, an increased immune response may be present in adult individuals compared with tadpoles. With relatively fewer deleterious consequences of Bd infections for tadpoles, lower selective pressure would be expected than in post-metamorphic animals experiencing mortality owing to chytridiomycosis. Our results indicated that subtropical stream environments are suitable for high Bd prevalence in tadpoles throughout the year.

Tadpoles of Mixophyes fleayi may be an important Bd reservoir because of the high infection prevalence and intensity and year-round presence in the streams (Rachowicz and Vredenburg 2004; Cashins 2009; Briggs et al. 2010; Catenazzi et al. 2013; Courtois et al. 2017; Sapsford et al. 2018). However, without an explicit causal or mechanistic model that accounts for transmission (i.e. shedding rates, exposure rates of vulnerable hosts, etc.), we cannot make claims about their status as a reservoir (Wilber et al. 2022). On the basis of the body-size distributions from December 2019 to January 2021, pulses of reproduction occurred in the summer months (Fig. 1), consistent with previous studies (Stratford et al. 2010). Importantly, large tadpoles were present throughout the study period; given that larger tadpoles were more often infected, these individuals are likely to contribute to the maintenance of Bd in the stream environments. Experimental studies have shown that tadpoles readily infect other tadpoles when they share an aquatic environment (Rachowicz and Briggs 2007; Hagman and Alford 2015; Courtois et al. 2017). However, the extent to which adults become infected when entering these aquatic environments is less well understood. In laboratory studies, all post-metamorphic animals housed in an enclosure with infected tadpoles became infected within 3 weeks (Rachowicz and Vredenburg 2004). However, the transmission of tadpoles to terrestrial adults is currently poorly understood. For stream-breeding frogs where breeding congregations occur in and around streams, understanding the risk of exposure from the stream environment is important for understanding infection dynamics. This remains an open area of research across many Bd–amphibian systems.

To our knowledge, our study is the first to assess the effects of both Bd infection status and intensity on the probability of having oral dekeratinisation. Our results confirmed that oral dekeratinisation is a good indicator of Bd infection in Mixophyes fleayi, with a 0.15 [95% HPDI: 0.03, 0.3] probability of an infected tadpole not displaying oral dekeratinisation (Symonds et al. 2007; Navarro-Lozano et al. 2018). We found that infected tadpoles had, on average, 113 times higher odds to have mouthpart loss than did uninfected tadpoles (Fig. 5a). Additionally, we found that among those tadpoles infected with Bd, individuals with a higher infection intensity were significantly more likely to have mouthpart loss (Fig. 5b). Only 13 tadpoles that had oral dekeratinisation returned negative swabs, but some of these were likely to be false-negatives. Our study did not correct for Bd detection errors in the swabbing process, which is predicted to be roughly 0.67 for the average tadpole infection intensity estimated in this study (Hollanders et al. 2023a). Although other causes of oral dekeratinisation exist (Knapp and Morgan 2006), given the prevalence of Bd at our sites and the clear association with Bd, these individuals are likely to have lost their mouthparts as a result of infection. The positive effect of body size on the degree of oral dekeratinisation for uninfected tadpoles may also suggest that prior infections caused the dekeratinisation, and that tadpoles had successfully cleared infection or that our detection protocol failed to detect Bd in these tadpoles. Importantly, the probability of finding two tadpoles with oral dekeratinisation that are not infected with Bd is just 0.0025; consequently, the presence of mouthpart loss in a small sample of tadpoles should allow investigators to conclude the presence of Bd. Unfortunately, the relationship between oral dekeratinisation and Bd infection appears highly species-specific; although the relationship is present for many species, some species show mouthpart loss in the absence of Bd (Navarro-Lozano et al. 2018).

Fig. 5.

(a) Posterior distributions (with medians and 95% HPDIs) of the probability of having mouthpart loss, as determined by the presence of dekeratinisation of the jaw sheath, for Mixophyes fleayi tadpoles uninfected (left) and infected (right) with Bd. Jittered points are observed individuals, categorised by presence of mouthpart loss (present, top; absent, bottom) and Bd infection status. (b) Prediction curve (median and 95% HPDI) with the coefficient of Bd infection intensity on the probability of having mouthpart loss from logistic regression. Points are observed individuals with corresponding Bd infection intensities estimated by the model (summarised as the median of the posterior distributions) after accounting for measurement error in the sampling (swabbing) and diagnostic (qPCR) processes. (c) Posterior distribution (with median and 95% HPDI) of the average mouthpart loss intensity score for infected (right) tadpoles (there were not enough data to estimate the score of uninfected tadpoles). Points are observed mouthpart loss intensity scores.

In conclusion, we have shown that Bd infection prevalence and infection intensity in tadpoles are considerably higher than in adults and juveniles at the same sites. Our focal species Mixophyes fleayi declined significantly during the epidemic but has since recovered, having stable populations despite individual mortality occurring with high Bd intensities (Hollanders et al. 2023b). Nevertheless, the year-round presence of tadpoles that are more often infected and infected with higher pathogen loads in the stream environments may constitute a Bd reservoir, both in terms of indirect transmission through the water and directly into the adult population through metamorphosis. However, to establish the role of tadpoles as pathogen reservoirs, future research will need to quantify transmission to more vulnerable hosts. The results of this work suggest fruitful future research into mechanisms of immunocompetence across amphibian life stages.

Supplementary material

Supplementary material is available online.