Introduction and Purpose

Amphibians are the most threatened vertebrates of the Anthropocene, with ~18 % of the 7530 described species listed as endangered or critically endangered (AmphibiaWeb 2016). A major cause of global amphibian declines is chytridiomycosis, a disease caused by the fungal pathogens Batrachochytrium dendrobatidis (Bd) (Daszak et al. 2003; Wake and Vredenburg 2008) and Batrachochytrium salamandrivorans (Bsal) (Martel et al. 2014). Chytridiomycosis caused by Bd is implicated in the extinction, or serious population declines, of more than 200 amphibian species worldwide (Skerratt et al. 2007; Crawford et al. 2010). Bd has been detected at 48 % of localities sampled from around the world, representing every continent where amphibians live (James et al. 2015). Fungal-borne diseases affecting both wild and domestic organisms are emerging at rates not seen with bacterial- or viral-borne disease (Fisher et al. 2012). In addition, generalist fungal pathogens, such as Bd and Bsal, can persist in host communities over a long term due to varying levels of host susceptibility. These also have the ability to reproduce very quickly during epizootics, potentially driving some host species to extinction (Fisher et al. 2012).

Batrachochytrium salamandrivorans is emerging as a highly virulent pathogen of salamanders (Martel et al. 2014) and has recently spread to, and devastated, wild populations of Salamandra salamandra in the Netherlands and Belgium (Martel et al. 2013; Spitzen-van der Sluijs et al. 2013). The pathogen has been progressing through the Netherlands and neighboring Belgium and Germany. Researchers report Bsal at 14 separate areas in these countries where it infects at least three species of salamanders (Spitzen-van der Sluijs et al. 2016). In Europe, three-fourths of salamander species belong to Salamandridae (AmphibiaWeb 2016), many of which have been shown to be susceptible to Bsal infection (Martel et al. 2014). Many of these will disperse seasonally, reproduce aquatically, and are associated with seeps, pools, and riparian areas, traits which could allow them to easily spread the pathogen. These factors make the Bsal threat a very serious conservation concern in Europe and beyond. Importation of 201 salamander species have recently been restricted by the U.S. federal government in response to this threat (http://www.fws.gov/policy/library/2016/2016-00452.pdf); imported salamanders which are hosts for Bsal represent sources of harm to endemic U.S. salamanders under the Lacey Act (1990; 16 U.S.C. sections 3371–3378 and 18 U.S.C. sections 42–43). Similarly, Switzerland has banned importation of all salamanders to prevent introduction of Bsal (Schmidt 2016).

Survey and study of the Bsal and Bd pathogens are a conservation priority (Sleeman 2013; Grant et al. 2015). Informed entities must facilitate a rapid response to Bsal outbreaks should they occur (Gray et al. 2015). We missed this opportunity with the earliest Bd epizootics. Given that Bsal may be easily missed without sufficient breadth of taxonomic and geographic coverages, collaboration between researchers and documentation of every detail is paramount. The Bsal strain responsible for the epizootic in Europe shows a thermal preference lower than Bd under laboratory conditions, with optimal growth occurring between 10 and 15°C; the pathogen is sensitive to temperatures above 20°C (Martel et al. 2013); infection can be cleared after prolonged host exposure to 25°C in laboratory trials (Blooi et al. 2015). Because of known affinity for low temperatures, we can hypothesize that Bsal thrives in colder climates. Thus, we consider mountainous regions rich in salamanders to be priority areas for surveillance. Surveys should also include sites where enigmatic salamander declines have been reported, such as North America (Rovito et al. 2009; Caruso and Lips 2013). Given these considerations, we sampled salamander populations from several locations in three mountain ranges representing North America (Appalachian Mountains), South America (Peruvian Andes), and Europe (Swiss Alps) for both Bd and Bsal from June 2013 to January 2016.

Methods

Field Sampling

We swabbed live wild salamanders in three mountain ranges (Fig. 1): the Alps in Switzerland; the Appalachians in Western North Carolina, USA; and the Cordillera Oriental of the Eastern slope of the Andes in Peru. We sampled salamander communities opportunistically; every available individual found within a discreet timeframe was included. We captured salamanders using powder-free gloves and placed them in individual plastic bags. After a preliminary visual examination for signs of chytridiomycosis, such as skin ulcerations, abnormal shedding, poor body condition, and lethargy (Martel et al. 2013; Sabino-Pinto et al. 2015), every salamander was swabbed with a dry swab (Medical Wire and Equipment Co. Ltd.). Swabs were stroked across the skin on the ventral side of abdomen, legs, and tail a total of 30 times. Individuals were then weighed, measured (snout-vent and tail length), and noted for location of capture. All information regarding this survey is available at http://www.amphibiandisease.org (ark:/21547/AaW2 for Bsal data, and ark:/21547/AaZ2 for Bd data).

Figure 1
figure 1

Shows the three locations of the survey and prevalence of infection for each.

Full size image

Quantification of Bd and Bsal Infection

Swabs were analyzed for the presence of Bd and Bsal using a real-time polymerase chain reaction duplex protocol (Blooi et al. 2016) with a detection limit of two copies of ITS gene DNA. We used gBlock® (Integrated DNA Technologies, Coralville, IA) synthetic double-stranded DNA sequences for the Bd and Bsal ITS genes as standards, and included four serial dilutions (each in triplicate) in each plate.

Statistical Analyses

We calculated the Bd and Bsal prevalence of infection for all locations by dividing the number of infected salamanders by the total number surveyed. Confidence intervals were calculated using Bayesian inference and Jeffrey’s non-informative priors in the R package binom. Tests comparing prevalence between sites were conducted using a chi-square complemented with a Fisher’s exact test in the R package stats.

Results

We did not detect Bsal in any of our samples. However, we found Bd in each mountain range (Table 1; prevalence = 3.5 %, CI 1.2–8.2 %, n = 113 in the Appalachian Mountains; prevalence = 1.5 %, CI 0.5–3.6 %, n = 265 in the Lepontine Alps; prevalence = 13.7 %, CI 8.7–20.4 %, n = 131 in the Peruvian Andes).

Table 1 Detection of Infections by Location and Salamander Species.
Full size table

The arboreal Andean salamander, Bolitoglossa cf. caldwellae, accounted for 18 of the total 26 global Bd positives. Within this species, prevalence of Bd varied significantly across sites (\( \chi ^{2} \) = 14.26, DF = 2, Fisher’s P < 0.01) and was largest at the highest elevation site compared to the other two sites combined (\( \chi ^{2} \) = 12.55, DF = 1, Fisher’s P < 0.01, odds ratio = 6.05). Salamanders at the higher elevation site show over 6 times the odds of being infected compared to the other two sites combined. The survey site in the Eastern Andean slope (900–1000 m) is located within the Tambopata–Manu wet spot (Killeen et al. 2007) and shows a 35 % (19.6–48.3 %) prevalence for Bd. We found Bd in salamander samples from one site in Switzerland. It is a wetland surrounded by forest at 290 m, just west of the city of Golino. Those positives are from Triturus carnifex 25 % (0.3–65.3 %) and Lissotriton vulgaris 30 % (7.5–58 %). We found Bd at one Appalachian site where four positive samples represent three species of stream-dwelling plethodontids: Desmognathus monticola 9.1 % (0.03–30.5 %), Desmognathus ocoee 22.2 % (2.8–50.3 %), and Eurycea wilderae 10 % (0.04–33.1 %). The two sites are located in Western North Carolina, part of the ancient Appalachian mountain range and home to at least 60 species of salamander.

Discussion

We tested 509 salamander swabs from the Appalachian Mountains, the Andes and the Alps, for both Bd and Bsal presence. We did not detect Bsal in any salamander swabs collected. This survey is the first to screen the Peruvian salamander B. cf. caldwellae and complements Martel et al. (2014), where 645 Appalachian salamanders were screened for Bsal from 2009 to 2011, and 1817 Swiss salamanders were screened from 2008 to 2013. These findings support the hypothesis that Bsal is an emerging novel pathogen, which seems to have spread from its endemic region, presumably Asia (Martel et al. 2014), to Northwestern Europe, but not to other regions. However, if Bsal is not extensively distributed, or if present at a low prevalence, it is very possible that we have missed it. The opportunistic nature of the data collection also inevitably left some species underrepresented.

Wild populations of salamanders in Europe have been found with a Bsal prevalence below 10 % and as high as 100 % (Spitzen-van der Sluijs et al. 2016). True prevalence could vary across populations and the susceptibility of individual salamander species may vary significantly from one to another. Sample sizes below 18 would likely fail to include Bsal-positive individuals, given a true prevalence of 10 % or less. This is evident in the upper limits of our 95 % confidence intervals (n = 18; UCI 10 %, n = 10; UCI 17.1 %, n = 5; UCI 30.6 %), as sample sizes decrease, the probability of missing Bsal increases, even when assuming a high true prevalence (Digiacomo and Koepsell 1986). Given that some species at some locations are lacking in sample size for our data, it is not appropriate to assume those populations are without Bsal infection. An ideal Bsal survey would encompass more individuals and populations of underrepresented salamander taxa. Surveys should also include species known to be susceptible in order to maximize detection probability. The North American salamander, Siren intermedia, may harbor the pathogen without succumbing to chytridiomycosis for over 1 week (Martel et al. 2014). Moreover, the same infection experiments show that Bsal causes mortality in Notophthalmus viridescens, a widespread species of wild North American newt, after several weeks of infection. Our survey does not include these North American species known to have the ability to harbor Bsal, though it provides good evidence of Bsal absence in the Swiss S. salamandra at Camorino (n = 223, CI 0–0.9 %) and Triturus carnifex (n = 18, CI 0–10 %) at Mt. Generoso, two members of Salamandridae known to carry the pathogen in the wild (Martel et al. 2013; Spitzen-van der Sluijs et al. 2016).

We report that B. cf. caldwellae at higher elevations in the Eastern Andes of Peru may have higher Bd prevalence than those from the lower elevation sites. The Bd pathogen has been found infecting amphibians across a wide variety of elevations and ecosystems throughout that area (Catenazzi et al. 2011), but may not occur at high prevalence at the lower elevations due to consistently warmer temperatures experienced there (Catenazzi et al. 2014). Although our data are consistent with this hypothesis, many other factors, such as pathogen strain, host densities, and site effects could influence the dynamics of Bd prevalence. There is some uncertainty about the risk Bd poses to Neotropical salamanders. Salamander communities in the upper cloud forest of Southern Mexico seemed to have collapsed sometime between the mid-1970s and early 1980s (Rovito et al. 2009). Overall, an enigmatic and seemingly regional reduction in the abundance of most of the species surveyed was reported. Changes in precipitation and temperature regimes seem a likely explanation (Rovito et al. 2009; Brinkman et al. 2016). No direct evidence was found linking Bd to these declines; however, past Bd epizootics could have left the salamander communities in their present state, similar to what happened to Neotropical anuran communities around the same timeframe (Rovito et al. 2009). Ranavirus has recently been reported from the Peruvian Andes (Warne et al. 2016) and could pose a risk to Neotropical amphibians. No declines in B. cf. caldwellae have been observed in our study area despite the presence of Bd.

This survey complements other Bd surveys in Switzerland (Garner et al. 2005; Lotters et al. 2012; Tobler et al. 2012; Woodhams et al. 2014). We found Bd in only one surveyed Alpine wetland, a lowland site (250 m). This runs counter to the increase in Bd prevalence with altitude we report from Peru. We failed to detect Bd at any other site in Switzerland, including the Camorino site, where we tested 223 S. salamandra individuals. The relatively low prevalence of Bd in Swiss salamanders detected by this survey complements previous findings by Woodhams et al. (2014), of which two species shared across surveys were found to have 24.8 % (Ichthyosaura alpetris) and 27.3 % (Lissotriton. vulgaris) prevalence of Bd infection.

We found Bd in very low prevalence and at only one site in the Appalachians of Western North Carolina. Our findings are similar to previous surveys in the Appalachian region that found very low Bd prevalence (Williams and Groves 2014; Moffitt et al. 2015). Underrepresented species in our data may be undergoing undetected epizootics; these hosts may be rare due to chytridiomycosis. Epizootics of other pathogens, such as Ranavirus, which has been detected in Appalachian salamanders (Hamed et al. 2013; Blackburn et al. 2015; Sutton et al. 2015), could be associated with past enigmatic declines in the area. Appalachian salamanders should continue to be monitored for population trends and disease in search of the reasons for these declines.

Conclusion

Emerging wildlife diseases can be a serious threat to biodiversity, as seen with Bd and Bsal. The pandemic Bd strains are currently found in over 90 % of the Bd-positive sites tested globally (James et al. 2015); there is fear that the Bsal pathogen could become similarly successful without preventative action. We need more eyes on the lookout for Bsal. There is especially the risk of Bsal spillover from captive salamanders into wild communities (Martel et al. 2014; Grant et al. 2015; Yap et al. 2015; Richgels et al. 2016) to consider. Susceptibility trials have previously demonstrated that popular traded newt species infected with Bsal exhibit none of the clinical signs of chytridiomycosis (Martel et al. 2014). It is likely that without proper diagnostic tests, infected salamanders, such as these, would go undiagnosed until after arriving at their final destination. The international wildlife trade is also commonly plagued by misidentification of animals and their origins (Pavlin et al. 2009; Gerson 2012), severely limiting the effectiveness of trade restrictions. Diagnostic testing for Bsal could be an effective alternative to salamander movement restrictions in some cases. Salamanders could be tested and certified as Bsal free. The swabbing part of the process requires minimal skill and can be performed by the public. Diagnostic labs already at work on wildlife pathogens could be employed to run the analyses of those samples. This will increase chances of early detection of Bsal in the established salamander trade, something that is crucial to the conservation of salamander diversity and their services in North America (Grant et al. 2015) and beyond. In surveying sites in Appalachia, the Peruvian Andes, and the Swiss Alps, we failed to detect Bsal. We contribute evidence that Bsal has yet to spread beyond Asia and Northwestern Europe.