Empirical evidence for effects of invasive American Bullfrogs on occurrence of native amphibians and emerging pathogens

eDNA-based monitoring of Batrachochytrium dendrobatidis and Batrachochytrium salamandrivorans with ddPCR in Luxembourg ponds: taking signals below the Limit of Detection (LOD) into account

BACKGROUND

Batrachochytrium dendrobatidis (Bd) and Batrachochytrium salamandrivorans (Bsal) are two pathogenic fungi that are a significant threat to amphibian communities worldwide. European populations are strongly impacted and the monitoring of the presence and spread of these pathogens is crucial for efficient decision-making in conservation management.

RESULTS

Here we proposed an environmental DNA (eDNA) monitoring of these two pathogenic agents through droplet digital PCR (ddPCR) based on water samples from 24 ponds in Luxembourg. In addition, amphibians were swabbed in eight of the targeted ponds in order to compare the two approaches at site-level detection. This study allowed the development of a new method taking below-Limit of Detection (LOD) results into account thanks to the statistical comparison of the frequencies of false positives in no template controls (NTC) and below-LOD results in technical replicates. In the eDNA-based approach, the use of this method led to an increase in Bd and Bsal detection of 28 and 50% respectively. In swabbing, this resulted in 8% more positive results for Bd. In some samples, the use of technical replicates allowed to recover above-LOD signals and increase Bd detection by 35 and 33% respectively for eDNA and swabbing, and Bsal detection by 25% for eDNA.

CONCLUSIONS

These results confirmed the usefulness of technical replicates to overcome high levels of stochasticity in very low concentration samples even for a highly sensitive technique such as ddPCR. In addition, it showed that below-LOD signals could be consistently recovered and the corresponding amplification events assigned either to positive or negative detection via the method developed here. This methodology might be particularly worth pursuing in pathogenic agents' detection as false negatives could have important adverse consequences. In total, 15 ponds were found positive for Bd and four for Bsal. This study reports the first record of Bsal in Luxembourg.

Widespread amphibian Perkinsea infections associated with Ranidae hosts, cooler months and Ranavirus co-infection

Amphibians suffer from large-scale population declines globally, and emerging infectious diseases contribute heavily to these declines. Amphibian Perkinsea (Pr) is a worldwide anuran pathogen associated with mass mortality events, yet little is known about its epidemiological patterns, especially in comparison to the body of literature on amphibian chytridiomycosis and ranavirosis. Here, we establish Pr infection patterns in natural anuran populations and identify important covariates including climate, host attributes and co-infection with Ranavirus (Rv). We used quantitative (q)PCR to determine the presence and intensity of Pr and Rv across 1234 individuals sampled throughout central Florida in 2017-2019. We then implemented random forest ensemble learning models to predict infection with both pathogens based on physiological and environmental characteristics. Perkinsea infected 32% of all sampled anurans, and Pr prevalence was significantly elevated in Ranidae frogs, cooler months, metamorphosed individuals and frogs co-infected with Rv, while Pr intensity was significantly higher in ranid frogs and individuals collected dead. Ranavirus prevalence was 17% overall and was significantly higher in Ranidae frogs, metamorphosed individuals, locations with higher average temperatures, and individuals co-infected with Pr. Perkinsea prevalence was significantly higher than Rv prevalence across months, regions, life stages and species. Among locations, Pr prevalence was negatively associated with crayfish prevalence and positively associated with relative abundance of microhylids, but Rv prevalence did not associate with any tested co-variates. Co-infections were significantly more common than single infections for both pathogens, and we propose that Pr infections may propel Rv infections because seasonal Rv infection peaks followed Pr infection peaks and random forest models found Pr intensity was a leading factor explaining Rv infections. Our study elucidates epidemiological patterns of Pr in Florida and suggests that Pr may be under-recognized as a cause of anuran declines, especially in the context of pathogen co-infection.