Terrestrial Dispersal and Potential Environmental Transmission of the Amphibian Chytrid Fungus (Batrachochytrium dendrobatidis)

Dispersal and exposure to amphibian chytrid fungus (Batrachochytrium dendrobatidis, Bd) is not confined to the aquatic habitat, but little is known about pathways that facilitate exposure to wild terrestrial amphibians that do not typically enter bodies of water. We explored the possible spread of Bd from an aquatic reservoir to terrestrial substrates by the emergence of recently metamorphosed infected amphibians and potential deposition of Bd-positive residue on riparian vegetation in Cusuco National Park, Honduras (CNP). Amphibians and their respective leaf perches were both sampled for Bd presence and the pathogen was detected on 76.1% (35/46) of leaves where a Bd-positive frog had rested. Although the viability of Bd detected on these leaves cannot be discerned from our quantitative PCR results, the cool air temperature, closed canopy, and high humidity of this cloud forest environment in CNP is expected to encourage pathogen persistence. High prevalence of infection (88.5%) detected in the recently metamorphosed amphibians and frequent shedding of Bd-positive residue on foliage demonstrates a pathway of Bd dispersal between aquatic and terrestrial habitats. This pathway provides the opportunity for environmental transmission of Bd among and between amphibian species without direct physical contact or exposure to an aquatic habitat.


Introduction
Infection by the pathogenic amphibian chytrid fungus Batrachochytrium dendrobatidis (Bd) poses a major threat to global amphibian biodiversity [1,2]. Response to infection varies considerably between species; a minority of those tested generally carry Bd in the absence of morbidity and may serve as aclinical reservoir hosts, such as the American bullfrog, Lithobates catesbeianus [3], and the African clawed frog, Xenopus laevis [4], whereas others are highly susceptible to chytridiomycosis and have suffered dramatic decline following introduction in wild populations [5,6]. Variation in virulence has been observed, and exposure to certain isolates of the highly pathogenic BdGPL clade causes mortality in amphibians more quickly than others [7,8]. Bd demonstrates low host species specificity and as of 2012, infection had already been reported in 516 species from 52 countries [9], and evidence suggests this pathogen is native in some parts of its range but emerging and spreading in others [10,11]. Identified 15 years ago [12], the geographic origin and subsequent pathways of global and local Bd dispersal remain largely speculative, although recent studies show Bd is now commonly spread via the international and domestic trade in live amphibians [13][14][15][16]. However, mechanisms of dispersal outside the amphibian host and in the absence of anthropogenic assistance are more obscure.
Direct and indirect modes of Bd dispersal and transmission within wild amphibian populations have been postulated, but few have been demonstrated. Direct contact between animals engaged in amplexus or territorial confrontation is thought to be a common mode of transmission [17]. Contact with contaminated water is another avenue, and Bd's motile uniflagellated zoospores can disperse through a water body either by swimming short distances or by being carried in water currents [18]. Waterfowl might carry Bd between separate water bodies, either on their feathers or feet [19][20][21]. However, the high prevalence of Bd detected in terrestrial and arboreal amphibian species that infrequently contact each other and typically do not directly engage with other species or enter permanent water bodies [22][23][24][25], suggests the presence of additional avenues of Bd dispersal and environmental transmission. For example Burrowes et al. [26] detected a high prevalence of infection (44.1%) in Eleutherodactylus coqui, a directdeveloping terrestrial anuran inhabiting leaf litter in the cloud forest in Puerto Rico and McCracken et al. [27] found 33% of canopy-dwelling amphibians infected in a lowland Ecuadorian rainforest. Bd has also been detected on 62% of terrestrial soil-dwelling caecilians sampled in Cameroon [28,29]. Collectively, the detection of Bd on amphibians that inhabit the forest canopy, terrestrial leaf littler, and soil suggests a common terrestrial existence where its dispersal and transmission are not constrained by the absence of permanent water.
The spread of Bd through Central and South America is associated with dramatic amphibian declines and extirpations [5,6,30,31] and interestingly, affected sites include remote wilderness areas and national parks where anthropogenic-assisted Bd spread is expected to be minimal [31][32][33]. Although a wave of Bd appears to have swept southeast through Central America during the 1980's [10,32], relatively little is known of its present distribution and ecological impact in Honduras. Infected amphibians have been reported from two locations, Pico Bonito National Park [34] and Cusuco National Park (CNP) [24], but the country boasts a mosaic of additional montane cloud forests likely to be similarly affected, but not yet surveyed. It has been estimated that nearly 50% of 111 amphibian species in Honduras have suffered declines in recent years from a combination of factors, including chytridiomycosis, and seven endemic anuran species were believed extinct [35], although one (Craugastor milesi) was recently rediscovered [36]. Bd has been detected in Honduran terrestrial anurans that undergo direct metamorphosis in leaf litter, including Craugastor aurilegulus and C. rostralis [24,34], and the source of pathogen exposure to these species remains enigmatic. Similarly, Bd-positive terrestrial frogs have been detected in Costa Rica (Oophaga pumilio and Craugastor fitzingeri), prompting the authors to suggest that Bd can survive on the moist forest floor where transmission might occur [32].
Since Bd occurs in the superficial skin of infected metamorphosed amphibians, there appears to be potential for infectious zoospores and sporangia within shedding skin to contaminate environmental substrates. Newly post-metamorphic anurans, in particular, often exhibit both elevated Bd prevalence and zoospore loads [24,[37][38][39], so their emergence from water might represent a considerable pathway of Bd dispersal into the terrestrial zone. To explore this potential avenue of terrestrial Bd spread we investigated whether terrestrial vegetation becomes contaminated with Bd following contact with recently metamorphosed amphibians under natural field conditions.

Ethics
Amphibian sampling in CNP adhered to established protocols [40] and were authorized by the Instituto Nacional de Conservacion y Desarollo Forestal Areas Protegidas y Vida Silvestre (ICF) of Honduras as part of the long-term Biodiversity Monitoring Programme performed by Operation Wallacea. Permission to export samples was granted by Honduran permit #'s 44735 and 19987.

Study Site
This investigation was performed from 9 July to 6 August 2013 in Cusuco National Park (CNP), a montane rainforest located in the Sierra de Omoa of northwestern Honduras. The altitude of CNP ranges from 550 m to 2200 m and fieldwork was performed between 1300 m and 1600 m at three different river sites ( [24] and its presence in the region for approximately two decades or greater [41]. Recently metamorphosed individuals of four tree frog species susceptible to Bd were targeted for sampling (Duellmanohyla soralia, Plectrohyla dasypus, Plectrohyla exquisita, and Ptychohyla hypomykter). Of these species, P. dasypus, previously demonstrated the highest prevalence of infection both at the species level (78%) and among recently metamorphosed individuals (94%) [24]. Most sampling was performed at night when animals were more active and likely to be encountered on riparian vegetation, although some opportunistic sampling occurred in the day. Most frogs were encountered within 5 m of the water's edge, but some were found up to 50 m from the river. Sampling was restricted to frogs resting on leaves, and not those perched on stalks or branches.

Leaf and Amphibian Sampling
Recently metamorphosed amphibians were removed from leaves by inverting them above a new plastic bag, into which the amphibian either jumped or was guided by a gentle tap on the underside of the leaf. Care was taken not to exert pressure between the frog and leaf, to prevent increased potential Bd shedding. Vegetation was sampled first, and then the corresponding amphibian was sampled. Nitrile gloves were worn and changed between every swab collected to reduce the risk of sample cross contamination. Leaves and frogs were each sampled with sterile fine-tipped rayon swabs (Medical Wire & Equipment Co., #MW113). For leaves, each swab was drawn across the leaf surface 20 times, where the amphibian was perched and in most instances, had left a small film of moisture visible on the leaf's surface, approximately 0.5 cm in diameter, marking the amphibians' location (Fig 1). For amphibians, the hands, feet, and pelvic patch were swabbed five times each following protocols established by Hyatt et al. [40]. Swab buds were snapped off into 2 mL microcentrifuge tubes filled with 1 mL 70% ethanol as a preservative. After sampling was completed, each amphibian was replaced to its original position in the vegetation.

Temperature
Immediately upon encountering an amphibian perched on vegetation, the amphibian's dorsal body surface, the vegetation surface, and the air temperature were measured to characterize the environmental conditions Bd would be exposed to, if present. Temperatures were measured using a Raytek ST81 Non-contact Infrared Thermometer (RAYST81, emissivity set to 0.95), from a distance of 0.5 m or less. Accuracy of the thermometer is ± 1% of reading or ± 1°C, whichever is greater. This technique obtains amphibian body temperature readings within 0.5°C of cloacal temperatures [42]. Air temperature was measured with the attachable RTD temperature probe.

Water Sampling
Water samples from rivers at the three sites were collected and filtered for Bd detection. These samples were processed for Bd testing following established protocols [43], with the exception that a peristaltic pump was used to increase the efficiency of sampling efforts by maximizing the volume of water filtered. We used sterile silicone rubber peristaltic pump tubing and replaced a new length for the collection of each sample. Water was pumped through Sterivex filter capsules (0.22 micron pore size) until the flow rate greatly diminished. Then the volume filtered was measured and recorded. The content of each filter capsule was rinsed with 50 mL phosphate buffered saline solution and then pumped dry. A bead of clay sealant was used to plug the outlet spout of the capsule before being preserved by adding 0.9 mL of Qiagen ATL tissue lysis buffer with a sterile 1 mL syringe. Luer-Lok screw caps sealed the inlet spout of the capsules and a bead of quick-drying epoxy was applied behind each clay plug in the outlet spout to provide the seal with reinforcement during transit. Fresh pairs of Nitrile gloves were worn each time a water sample was collected. All water sampling was performed during daytime hours. Water temperature was measured at the time of sampling using the attachable RTD temperature probe of the Raytek ST81 Non-contact Infrared Thermometer.

Real-Time PCR Analysis
Samples were processed via a sensitive quantitative PCR assay (qPCR) specific to Bd following the established protocol [44] and with the addition of BSA to the qPCR master mix as per Garland et al. (2010) [45]. Samples were extracted with 100 μl Prepman Ultra (Applied Biosystems, California, USA), with a final 20 μl of supernatant removed for downstream use. An aliquot of this supernatant was diluted 1:10 in DNase-free water for qPCR. The qPCR protocol used Sensi-Mix II Low Rox (Bioline, Massachusetts, USA) as the qPCR master mix [46]. For each sample, 5 μl of 1:10 dilution of swab DNA was added to each well for a final total qPCR volume of 20 μl. Samples and controls were run in triplicate with three positive, standard control samples (100, 10, and 1 zoospore/well, made from JAM81 pure culture; see Boyle et al. 2004 for standard control construction) and one non-template control (DNase free, molecular-grade water). When the qPCR assay failed to detect Bd in all three wells, the sample was deemed negative for Bd. Samples that produced a positive signal for Bd in either two or three wells on the first run were considered positive for Bd. When only one of three replicates detected Bd, the sample was rerun (in triplicate again) in a subsequent plate. For rerun samples that had at least a cumulative total of two of six replicates positive for Bd (from at least two separate plates), the sample was deemed positive for Bd. All zoospore loads described in this report have not been converted and reflect the actual zoospore loads present in 5 μl DNA (1:10 dilution), placed into 20 μl reaction volumes.

Data Analysis
We applied Chi-square test on a 2x2 contingency table to determine whether row and column marginal frequencies were equal. The values in the matrix included: number of Bd-negative frogs associated with Bd-negative leaves (5)

Amphibian Swab Bd Results
Bd was detected on 46 of 52 (88.5%) amphibians and from all four species (Table 1). The average Bd zoospore equivalent load detected on Bd-positive amphibians was 103.94 and ranged from 0.06-1,574.62.

Leaf Swab Bd Results
Bd was detected on 36 of 52 (69.2%) leaves, 97.2% of which had a Bd-positive recently metamorphosed amphibian on them (35/36) (statistical significance of the association, df = 1, chi-squared = 6.23, p-value = 0.013) ( Table 1). Only one Bd-positive leaf had an amphibian that tested Bd-negative. The average Bd zoospore equivalent load detected on Bd-positive leaves was 40.48 and ranged from 0.12-1,040.45.

River Water Filter Bd Results
The presence of Bd was detected in all three river water samples ( Table 2). The average Bd zoospore equivalent load per liter of river water was 0.23 and ranged from 0.03-0.57. Daytime water temperature averaged 17.0°C and ranged from 16.3-17.5°C.

Discussion
We frequently detected Bd on leaf surfaces after removal of recently metamorphosed Bd-positive frogs, indicating their emergence does contribute towards the spread of Bd from aquatic into terrestrial locations. Average zoospore loads detected on leaf surfaces were comparable to those from corresponding amphibian skin swabs, and sometimes greater. The presence of Bd on riparian vegetation allows exposure to occur in the absence of direct physical contact with Bd-positive animals or contaminated water. Accordingly, this pathway of Bd dispersal and terrestrial exposure provides one possible explanation for the source of infection previously detected in amphibians that do not demonstrate a strong association with water. This pathway of Bd spread may occasionally facilitate transmission between aquatic and terrestrial species and from juvenile to adult frogs, if foliage maintains infectious Bd loads. On 11 July 2013, both a recently metamorphosed and adult Plectrohyla dasypus were observed perched together on the same plant at the same time, approximately 5 cm apart (Fig 2). The skin swab sample collected from this juvenile frog (HN13BD145) exhibited a considerable zoospore load (1,085.68), as did the leaf swab (43.30), demonstrating a high risk of exposure to the nearby adult which tested Bd-negative at the time of sampling. Following metamorphosis, this species leaves the aquatic habitat and moves into arboreal vegetation, reducing the likelihood of subsequent Bd exposure from contaminated river water. The high prevalence of infection in P. dasypus juveniles detected in this and previous surveys [24] suggests that their seasonal emergence en masse may release a substantial quantity of Bd into the riparian zone shared with amphibians that approach the water's edge, but do not typically enter it. Although we identified a potential mechanism of pathogen exposure to terrestrial amphibians, the role of contaminated vegetation in Bd transmission remains in question. Detection of Bd via qPCR indicates DNA presence, but does not reveal condition at the time of sampling. A lack of experimental work to evaluate the persistence and detectability of Bd DNA following cell death makes it difficult to discern whether we likely detected viable Bd or instead DNA fragments from expired cells that continued to react with Bd qPCR primers. This interpretation limitation is not exclusive to environmental Bd swab samples, but likewise applies to amphibian skin swabs; a positive qPCR result does not independently demonstrate the viability of Bd on that animal. Still, environmental conditions observed at all sampling localities in CNP were similar to those in the laboratory where Bd survived outside a host [19] and may aid persistence of Bd on leaf surfaces. Temperatures recorded in the field were all within or near the range for optimal in vitro growth of Bd (17-25°C) and well below its thermal maximum of 28°C [47], although optimal temperature regimes may vary between Bd isolates [48] and none from Honduras have yet been characterized. Desiccation poses the other well-defined abitoic limitation to Bd survival [19,49] but the presence of both high relative humidity and a dense forest canopy preventing direct sun exposure is correlated with higher Bd prevalence and infection loads [50]. These conditions are typical of CNP, a montane cloud forest, and expected to prolong drying. Lastly, laboratory experiments have shown that when maintained under suitable temperature and moisture levels (and without bacteria), Bd can survive in the absence of a host for at least two months in water or moist sand [19]. Thus, additional laboratory work is needed to test the survival times of cultured Bd on leaf surfaces to identify the potential duration of this form of environmental persistence and evaluate the average Bd loads needed to cause successful transmission under naturalistic conditions. Previous efforts to illustrate environmental Bd transmission have mainly focused on exposure to permanent water bodies inhabited by Bd-infected amphibians [18,19]. Laboratory trials demonstrated transmission of Bd between experimentally-infected and uninfected tadpoles of Rana muscosa and also from tadpoles to post-metamorphic animals, when occupying a shared water source [18]. Successful transmission required a 2-3 week duration of exposure, likely impeded by dilution of the pathogen in a naturalistic environment, similar to the low densities of Bd detected in the water samples collected at our survey sites in CNP (Table 2). To encourage transmission after short-term exposure, laboratory experiments have often employed highly concentrated inoculates of approximately 100 million Bd zoospores delivered in less than 100 mL of water [51][52][53] whereas the highest concentration detected in a natural body of water has been 3 million zoospores L -1 and less than 100 zoospores L -1 is common [54]. In this context, the concentrated Bd loads we detected on leaf surfaces in CNP relative to the adjacent Bd-positive river water suggests that contact with affected foliage might pose a greater threat of exposure and transmission to terrestrial amphibians than would a splash of water from these rivers.
We detected the presence of Bd on vegetation in the understory, but periods of heavy rain are expected to also flush Bd into the soil and leaf litter below. Surveys in CNP have identified the presence of live aquatic crustaceans (copepods and ostracods) inhabiting terrestrial water films on forest floor leaf litter [55,56], suggesting moisture persistence in this limnoterrestrial habitat. The persistence of these water films in humid rainforest environments would help protect Bd from desiccation in a seemingly terrestrial habitat, and also allow exposure to amphibians that occupy leaf litter and burrow into the ground. Accordingly, this mode of Bd dispersal and indirect exposure may explain the origins of infection documented in species of soil-dwelling salamanders [22,23,25] and caecilians [28,29].
Numerous biologic and abiotic factors are expected to influence the frequency of Bd dispersal from aquatic into terrestrial habitats and potential consequences. The prevalence and intensity of Bd detected in amphibian populations often demonstrates fluctuations due to seasonal changes in environmental conditions and these factors will affect the amount of zoospores available to be shed into the terrestrial environment [49,54,57]. Rowley et al. [58] investigated the presence of Bd in terrestrial retreat sites of two aquatic stream frog species (Litoria lesueuri and L. nannotis), and did not detect Bd in 122 environmental swab samples. As suggested by the authors, the observed Bd absence may have been influenced by the low prevalence and infection loads concurrently detected in the adult amphibians sampled at these locations. Our results show that in a locality where both Bd prevalence and infection loads are high, it is common for Bd to be shed into terrestrial locations, including amphibian retreat sites.
The presence of Bd in terrestrial habitats should be considered when identifying potential threats to amphibian species of concern. Although it has been suggested that Bd poses the greatest risk of infection to amphibians breeding in permanent streams [59], we caution against this generalization and encourage additional surveillance in terrestrial and arboreal amphibian habitats where animals continue to test positive for Bd, despite pathways of exposure being more obscure. The frequency of Bd exposure from terrestrial substrates is unknown but may be considerable where optimal environmental conditions are present, especially if it can survive as a saprobe as previously suggested [12]. An improved understanding of Bd dispersal and persistence in the natural environment is essential to better explain and predict the continued spread of this pathogen in regions where the anthropogenic-assisted exposure to Bd-positive amphibians or substrates is unlikely.