Exposure to Batrachochytrium dendrobatidis affects chemical defences in two anuran amphibians, Rana dalmatina and Bufo bufo

Background Batrachochytrium dendrobatidis (Bd) is the causative agent of chytridiomycosis, one of the major causes of worldwide amphibian biodiversity loss. Many amphibians exhibit skin-based chemical defences, which may play an important role against invading pathogens, but whether the synthesis of these chemical compounds is enhanced or suppressed in the presence of pathogens is largely unknown. Here we investigated direct and indirect effects of larval exposure to the globally distributed and highly virulent Bd-GPL strain on skin secreted chemical defences and life history traits during early ontogeny of agile frogs (Rana dalmatina) and common toads (Bufo bufo). Results Exposure to Bd during the larval stage did not result in enhanced synthesis of the antimicrobial peptide Brevinin-1 Da in R. dalmatina tadpoles or in increased production of bufadienolides in B. bufo tadpoles. However, exposure to Bd during the larval stage had a carry-over effect reaching beyond metamorphosis: both R. dalmatina and B. bufo froglets contained smaller quantities of defensive chemicals than their Bd-naïve conspecifics in the control treatment. Prevalence of Bd and infection intensities were very low in both larvae and metamorphs of R. dalmatina, while in B. bufo we observed high Bd prevalence and infection intensities, especially in metamorphs. At the same time, we did not find a significant effect of Bd-exposure on body mass or development rate in larvae or metamorphs in either species. Conclusions The lack of detrimental effect of Bd-exposure on life history traits, even parallel with high infection intensities in the case of B. bufo individuals, is surprising and suggests high tolerance of local populations of these two species against Bd. However, the lowered quantity of defensive chemicals may compromise antimicrobial and antipredatory defences of froglets, which may ultimately contribute to population declines also in the absence of conspicuous mass-mortality events.

Chytridiomycosis, a disease affecting amphibians, is caused by the chytrid fungi Batrachochytrium dendrobatidis (Bd) and Batrachochytrium salamandrivorans (Bsal) [10]. Because Bsal has only been discovered recently [11], we know comparatively little about it [12], so that here we concentrate on chytridiomycosis caused by Bd. Chytridiomycosis has already led to the decline or extinction of several hundred species [13] and continues to cause mass mortality events on five continents due to repeated introductions arising from human activities [14]. Bd infects keratinous epidermal layers of the skin [15] and impairs its osmoregulatory function. This effect can cause shifts in electrolyte balance leading to cardiac asystolic death in juveniles and adults [16]. Tadpoles exhibit keratinous elements only in their mouthparts, so that they are less susceptible to Bd infection than later life-stages [17,18], and can act as reservoirs in natural habitats [19][20][21].
In this study, our aim was to experimentally investigate whether exposure to Bd resulted in increased production of skin-borne chemical defences as expressions of phenotypic plasticity, or if it caused lowered synthesis of defensive chemicals due to costs of infection or because of immune suppression. Therefore we exposed agile frog (Rana dalmatina; Fitzinger 1838) and common toad (Bufo bufo; Linnaeus, 1758) larvae to a highly virulent Bd isolate and monitored consequences for chemical defences and life history traits in well-developed tadpoles. Potential effects of larval infection reaching beyond metamorphosis on the abovementioned parameters were also tested. Adult R. dalmatina individuals produce at least one skin secreted AMP, Brevinin-1 Da [39], and B. bufo secrete several bufadienolide compounds in the skin [35,36,40]. Therefore, these species are suitable for the investigation of interactions between Bd infection and chemical defences.

Rana dalmatina
Control individuals all remained uninfected. The low Bd dose treatment resulted in very low infection prevalence and intensity in both tadpoles and froglets, while the high Bd dose treatment resulted in higher prevalence and intensity values in tadpoles and especially so in froglets ( Table 1; Fig. 1A).
Body mass of tadpoles was positively affected by development stage, but treatment had no significant effect on it, either alone, or in interaction with development stage. In case of froglets, treatment had a marginally non-significant effect on body mass, where individuals exposed to the high Bd dose treatment tended to be heavier than others. Length of larval development had no effect on body mass, either alone or in interaction with treatment (For details see Table 2). Treatment had no significant  (Table 2).

Bufo bufo
None of the control individuals were infected, but we obtained relatively high infection prevalence and intensities in toadlets exposed to the low Bd dose treatment and in both life stages exposed to the high Bd dose treatment (Table 1; Fig. 1B and C).  We detected 22 different bufadienolide compounds in B. bufo extracts, three of which we identified with the help of the standards as arenobufagin, telocinobufagin and bufotalin (Table 3). The presence of individual compounds showed varied age-dependent patterns: some bufadienolides were present at both life stages in all (e.g., compound 7) or nearly all individuals (e.g., compound 13), while others occurred in a high proportion of individuals only after metamorphosis (e.g., compounds 10-12; Table 3). After metamorphosis, toadlets produced more bufadienolide compounds and also experienced a two to three fold increase in TBQ compared to tadpoles (Fig. 3).
The number of bufadienolide compounds (NBC) did not differ among treatments in either life stage (GLM; tadpoles: Body mass was affected by development stage in case of tadpoles: individuals that were less developed also had a lower body mass. Treatment and its interaction with development stage had no significant effect on body mass ( Table 2). Within the high Bd dose treatment, where Bd prevalence was sufficiently high to be analyzed, tadpole body mass was again positively related to development stage (F 1,15 = 29.6, P < 0.001), while infection intensity had no effect on mass (GLM: F 1,14 = 0.003, P = 0.96). Fourteen days after completion of metamorphosis, body mass of toadlets was in a negative relationship with the length of larval development ( Table 2). Treatment and its interaction with the length of larval development had no significant effect on toadlet body mass ( Table 2). Within the low Bd dose treatment group, neither infection intensity, nor the length of larval development had an effect on body mass of toadlets (GLM: infection intensity: F 1,10 = 2.01, P = 0.19; length of larval development:

Discussion
In the present study experimental exposure to Bd did not result in significantly increased production of the antimicrobial peptide Brevinin-1 Da in R. dalmatina tadpoles, nor did it influence bufadienolide toxin synthesis in B. bufo tadpoles. However, Bd-exposure during the larval stage negatively affected chemical defences in metamorphosed individuals of both species. These results suggest that neither larvae nor freshly metamorphosed individuals respond to Bd-exposure with enhanced synthesis of antimicrobial chemicals, and that infection during the larval stage may rather carries costs that manifest in decreased quantities of chemical defences in metamorphs. Our results further indicate that larvae of the agile frog (R. dalmatina) were resistant to infection with a highly virulent Bd isolate, as indicated by low prevalence and infection intensities. At the same time, tadpoles of the common toad (B. bufo) were not resistant, but tolerant, as suggested by high Bd prevalence and high infection intensities, but no malign effects on life history traits: exposure to Bd did not influence body mass or development rate in larvae or metamorphs in either species. The knowledge available regarding the occurrence of AMP synthesis in larval anurans is limited and controversial. Skin-associated granular glands and their ducts are mostly immature before metamorphosis in most species [41,42], suggesting no, or limited AMP production (but see [43]). However, at the same time, gland products may also be secreted by a merocrine process, where the secretum reaches the skin surface via exocytosis directly or through the epidermal interstitium [41,44,45]. Furthermore, the adaptive immune system is suppressed during metamorphosis to prevent immune responses against newly emerging tissue types [46], suggesting a greater reliance on innate immune defences against invading pathogens during this susceptible period. Schadich et al. [47] found no evidence of AMP synthesis in tadpoles of Litoria ewingii despite intense AMP production after metamorphosis, while Wabnitz et al. [48] demonstrated efficient AMP synthesis in the skin of Litoria splendida at both larval and adult life stages. While Woodhams et al. [49] did not find AMPs in an early larval stage (development stage 25 according to Gosner,[50]) in two closely related species, Rana arvalis and R. temporaria, we detected that R. dalmatina tadpoles produce de novo the same AMP as do adults [39] at least during late larval development (development stage 37 according to Gosner). Whether this discrepancy among studies indicates species-specific differences in AMP synthesis, or if mean ± SE) in tadpoles (A) and toadlets (B), and total bufadienolide quantity (TBQ; mean ± SE) in tadpoles (C) and toadlets (D) after exposure to zero (control), low or high Bd zoospore concentrations. Note that the scale is not continuous in case of NBC. Letters in lower case indicate homogeneous subsets according to Tukey HSD post-hoc tests AMP production starts in Ranids some time later during larval development remains to be determined.
Our expectation that AMP synthesis is boosted upon exposure to Bd in R. dalmatina was not met. Phenotypic plasticity in chemical defences is not well understood [51], but some studies suggest that environmental stressors such as competition [52] and pathogen presence [53,54] can induce an increase in AMP synthesis in metamorphosed anuran amphibians. Furthermore, adult European water frogs (Pelophylax lessonae and P. esculentus) are capable of elevating AMP synthesis in response to Bd, if the skin microbiota is suppressed [55]. Although during the tadpole stage we also observed a slight increase in the relative amount of Brevinin-1 Da upon Bd exposure, this was not significant, and we detected a sharp reduction in AMP quantity in both Bd-exposed groups in froglets. Wild caught, infected adults of the frog Litoria serrata exhibited similarly reduced quantities of AMPs compared to uninfected conspecifics [56], but whether this was a cause or consequence of infection remained unknown. Our results suggest that reduced AMP expression can be a consequence of exposure to Bd. A reduced synthesis of AMPs may result from immunosuppression by Bd, as demonstrated in case of the adaptive immune system [57][58][59]. Lowered AMP production may also be a direct cost of infection, or may be indirectly caused by elevated corticosterone levels resulting from Bdinfection [60,61] because elevated corticosterone levels can cause reduced AMP synthesis [62,63]. Whether the reduced quantity of Brevinin-1 Da is still large enough to be effective against Bd in metamorphs, or other agents of the immune system can take over the role of Brevinin-1 Da in preventing severe infection requires further investigation.
Interestingly, AMP synthesis was reduced in metamorphs arising from the low Bd dose treatment without detectable amounts of Bd on all but one froglet. There are, however, precedents for significant effects of Bd exposure on various traits in the absence of confirmed infection. For example, Garner et al. [64] experienced significant mortality in B. bufo tadpoles and freshly metamorphosed individuals due to Bd exposure without detectable infection loads. Also, Bd exposure reduced growth in European treefrogs (Hyla arborea) without detectable levels of infection [65]. Individuals are probably able to prevent initial infections or naturally clear Bd infections acquired but may suffer the costs of the mounted immune response [66]. Alternatively, Bd presence in the surrounding aquatic environment may be sufficient to induce pathology or responses (e.g., via waterborne chemicals) even if infection does not occur, as shown in case of crayfishes (Procambarus spp. and Orconectes virilis) which are alternative hosts of Bd [67].
Exposure to Bd had no detectable effect either on the number of bufadienolide compounds or on total bufadienolide quantity in case of B. bufo tadpoles. Similarly, NBC of toadlets was also not affected by Bd presence. However, TBQ in the high Bd dose treatment was significantly lower compared to the low Bd dose treatment in toadlets. These results clearly indicate that Bd-prevalence and higher infection loads did not induce enhanced toxin synthesis. Besides their role in the chemical defence system, bufadienolides also contribute to the osmotic homeostasis of toads [68,69]. The decreased TBQ in the high Bd dose may have resulted from a compensatory response to the altered electrolyte balance (reduced sodium and potassium concentrations) due to Bd infection [16], but this speculation needs experimental confirmation. Alternatively, Bd infection can lead to structural damage in the skin and its glands [10], which may have contributed to the lowered toxin production in the high Bd dose treatment. Whatever the cause is, heavily infected toadlets unable to produce the increase in toxin content after metamorphosis, may suffer from detrimental consequences because skin secreted bufadienolides can act as repellents against vertebrate predators [70,71], they play a role in immune defence [37,38,72] and are important for osmotic homeostasis [68,69]. Lowered TBQ in the high Bd dose treatment, thus, indicates a possible indirect negative effect of Bd infection, similarly to what we found in regard to the chemical defence of R. dalmatina.
Exposure to Bd did not affect the measured life history traits in either species at the tadpole stage. In larval anurans only the mouthparts are keratinized structures [17], thus, mortalities due to chytridiomycosis are rare and susceptibility varies among species [10,18,73,74]. However, sublethal negative effects of Bd infection can also occur due to mouthpart damage [75,76], lethargy and poor swimming performance, resulting in lowered body mass and growth [18,74,76,77]. In the present study only a very few tadpoles of R. dalmatina became infected and in these individuals infection intensity was very low in both Bd treatments. In case of B. bufo tadpoles, prevalence of infection and infection intensities were high, especially in the high Bd dose treatment, so that the lack of fitness consequences was somewhat surprising. Two out of the three identified bufadienolide compounds (arenobufagin and telocinobufagin) were previously documented to moderately inhibit the growth of Bd [38]. These and some of the other unidentified bufadienolide compounds may have contributed to the high Bd-tolerance of B. bufo individuals, which lack AMPs. These results suggest that the studied populations exhibit low susceptibility to Bd infection during larval development: tadpoles of R. dalmatina appear to be highly resistant, while B. bufo larvae may be highly tolerant.
During metamorphic climax, Bd starts to colonise newly keratinized skin surfaces and spreads out on the entire animal [17,78]. In the present study, both the prevalence of Bd (GPL, IA042) and infection intensities were high in case of B. bufo, which is consistent with previous studies that used the same isolate and resulted in significant negative effects on body mass and mortality of B. bufo originated from the United Kingdom [64,79]. Mortalities due to chytridiomycosis in the closely related B. spinosus (formerly a subspecies of B. bufo) were also observed in Spain [64,80]. Contrary to this, the length of larval development was not affected by Bd exposure in our experiment, and we did not observe significant negative effects on body mass of toadlets. Only in the high Bd dose treatment, infection intensity was negatively related to body mass: lighter individuals had higher infection intensities, than heavier ones. Whether this pattern was a cause or consequence of infection remains unclear. All in all, these results suggest that B. bufo individuals of the studied population in Central Europe may be more tolerant to Bd than those in Western European populations. Alternatively to this spatial hypothesis, the pattern may also arise due to temporal differences; toads may have adapted to the presence of Bd since the earlier studies [64,79,80]. Furthermore, populations may respond differently to Bd infection from year to year due to phenotypic plasticity or epigenetic changes. These processes could also have contributed to the observed differences between the present and former studies in toad susceptibility. Regarding R. dalmatina, we detected low prevalence of infection in the low Bd dose treatment and moderate prevalence with low infection intensities in the high Bd dose treatment, and no adverse effects of Bd-exposure on life history traits of froglets. These results are in line with those of previous field studies suggesting that R. dalmatina is resistant to chytridiomycosis [21,81,82]. The less keratinized skin as well as the presence of Brevinin-1 Da may make the skin of R. dalmatina less suitable for Bd growth, hence the lower probability and intensity of infection as compared to B. bufo, which speculations need further investigations in the future.

Conclusions
Our results provide evidence that exposure to Bd can have negative effects on two different types of chemical defences of phylogenetically distant species in a later life-stage, even in the absence of obvious immediate effects on life-history traits, or, indeed, actual Bd infection. Because the investigated chemical defences are widespread among amphibians, the results of this study are likely to be applicable to many species with similar chemical defences, and this hidden effect of Bd presence in aquatic habitats should be considered in the future. Whether weakened chemical defences lead to lowered fitness in affected individuals remains an open question that will need further investigation.

Experimental procedures
In March 2016, we collected 40 eggs from each of nine freshly laid egg clutches of R. dalmatina from a pond in the Pilis-Visegrádi-Hills, Hungary (47.767058 N, 18.981325 E). We transported them to the Experimental Station Júliannamajor of the Plant Protection Institute, Centre for Agricultural Research. The Közép-Duna-Völgyi KTVF issued the permission to conduct the study (PE/KTF:3596-6-8/2016) and the Ethical Commission of the ATK NÖVI approved the investigation in accordance with Good Scientific Practice guidelines and national legislation. We placed eggs from each clutch separately into plastic boxes (24 × 16 × 13 cm) holding 1 L of reconstituted soft water (RSW; [83]) at a constant temperature of 19 °C and a 12:12 h light:dark cycle. Eggs were disinfected by bathing them for 3 days in 10 mg/L chloramphenicol in order to prevent accidental Bd infection [84] because Bd is reportedly present in the study area [21], even if with low prevalence [82]. Although chloramphenicol is an antibiotic agent, it is also effective against Bd [84,85] and can be safely used for the disinfection of amphibian eggs [86].
Five days after hatching, when larvae were at development stage 25 (Gosner), we started the experiment with 12 healthy-looking tadpoles from each family. We maintained tadpoles individually in plastic boxes (15 × 12 × 12 cm) filled with 1 L of RSW, and fed tadpoles with slightly boiled and smashed spinach complemented with Spirulina powder (1 m/m%) ad libitum. We changed water twice a week using different dip nets for each treatment to prevent contamination across treatments. Temperature was 19.4 ± 0.7 °C (mean ± SD) during the experiment. The light:dark cycle was adjusted weekly to outdoor conditions, starting with 12:12 h light:dark in late March which we gradually changed to 14:10 h by the end of April. We exposed tadpoles during the entire larval development to sterile culture broth (control), or to a low or high zoospore concentration in liquid Bd culture (low Bd dose and high Bd dose treatments hereafter; for details, see below). We assigned tadpoles to the treatments using stratified randomization, and arranged rearing boxes into randomized spatial blocks, each containing one replicate from each treatment. We exposed individuals of nine families to the three treatments in four replicates, which resulted in a total of 108 experimental units. Hatchlings that were not used in the experiment were released at the site of origin.
Thirty days after start of the experiment, we randomly selected half of the tadpoles, sampled their skin secretions (for details see below) then gently blotted them dry and weighed them to the nearest mg using an OHAUS-PA213 analytical balance. Thereafter, we euthanized and preserved tadpoles in 70% ethanol and stored samples at 4 °C until further analysis.
We monitored development of remaining tadpoles daily. When an individual reached development stage 42 (emergence of forelimbs; according to Gosner) we poured the water off, placed back the tadpole into the same box covered with a transparent and perforated lid, added 100 ml RSW and lifted one side of the container by ca. 2 cm to provide metamorphs with both a body of water and a dry surface. Once metamorphs reached stage 46 (complete tail resorption; according to Gosner) we placed individuals into new, covered boxes of the same size as before, equipped with wet paper towels and a piece of cardboard egg-holder as a shelter. We fed the froglets with small crickets (Acheta domestica, instar stage 1-2) ad libitum. Dates of metamorphosis and of completion of tail resorption were registered daily. Fourteen days after completion of tail resorption we weighed animals and humanely euthanized them using the "cooling then freezing" method [87], preserved individuals in 70% ethanol and kept samples at 4 °C until further processing.
We conducted the same experiment using individuals of B. bufo. However, we observed high mortality independently from treatments (15 out of 36 tadpoles died both in the control and in the low Bd-dose treatment, and 10 out of 36 tadpoles died in the high Bd-dose treatment), presumably due to a bacterial bloom caused by the simultaneous presence of Spirulina and culture broth (an effect we did not observe in case of R. dalmatina tadpoles in this experiment, or when tadpoles were fed solely with spinach in previous experiments). Consequently, 17 days after start of the first experiment, we re-started this part of the study and conducted the same experiment as described above with 108 randomly chosen B. bufo tadpoles from the same location with the following additional differences: (1) Eggs originating from 14 different clutches were kept in mixed groups in outdoor mesocosms containing 130 L of aged tap water, 40 g beech leaves and 0.5 L pond water with no prior chloramphenicol treatment. (2) Once brought into the laboratory after hatching, we fed tadpoles with smashed and slightly boiled spinach only (no Spirulina). (3) Light exposure adjustment was different because of the delayed start (4) Because of their smaller size, we fed toadlets with springtails (Folsomia sp.) after metamorphosis.

Maintenance of Bd culture and experimental exposure
We experimentally infected tadpoles with the global pandemic lineage (GPL) of Bd. This isolate originated from a dead Alytes obstetricans (IA042) collected in 2004 from a mass mortality event in Spanish Pyrenees. Cultures were maintained in mTGhL broth (8 g tryptone, 2 g gelatinehydrolysate and 4 g lactose in 1000 ml distilled water) in 25 cm 2 cell culture flasks at 4 °C and passed every three months into sterile mTGhL. One week before use, we inoculated 100 ml mTGhL broth with 1-2 ml of these cultures in 175 cm 2 cell culture flasks and incubated them for seven days at 22 °C. We assessed the concentration of intact zoospores using a Bürker chamber at × 400 magnification. During inoculation of tadpoles' rearing boxes, the mean initial concentrations were ~ 1.8 × 10 6 (used to infect R. dalmatina) and ~ 2 × 10 6 (used to infect B. bufo) zoospores (zsp)/ml in the flasks. These cultures were used for the high Bd dose treatment and we prepared a 100-fold dilution with sterile mTGhL broth for the low Bd dose treatment. After each water change, we inoculated 1 ml of these cultures into the tadpoles' rearing boxes holding 1 L RSW, resulting in 18-20 zsp/ml in the low Bd dose treatment and 1800-2000 zsp/ml in the high Bd dose treatment. Similar zoospore concentrations have been used widely and successfully in studies involving experimental infection [18,75,88]. We inoculated controls with the same quantity of sterile mTGhL broth. Contaminated water and equipment were disinfected overnight with VirkonS before disposal [89].

Assessment of infection intensity using qPCR
We assessed infection intensity from dissected mouthparts in case of preserved tadpoles and from toe clips in case of metamorphs. We homogenized tissue samples, extracted DNA using PrepMan Ultra Sample Preparation Reagent (Thermo Fisher Scientific, Waltham, Massachusetts, USA) according to previous recommendations [90], and stored extracted DNA at − 20 °C until further analyses. We assessed infection intensity using real-time quantitative real-time polymerase chain reaction (qPCR) following a standard amplification methodology targeting the ITS-1/5.8S rDNA region [90] on a BioRad CFX96 Touch Real-Time PCR System. To avoid PCR inhibition by ingredients of PrepMan, samples were diluted ten-fold with double-distilled water. We ran samples in duplicate. In case the result was equivocal, we repeated reactions in duplicate. If it again returned an equivocal result, we considered that sample to be Bd positive [91]. Genomic equivalent (GE) values were estimated from standard curves based on four dilutions of a standard (100, 10, 1 and 0.1 zoospore genomic equivalents; provided by J. Bosch; Museo Nacional de Ciencias Naturales, Madrid, Spain).

Skin secretion sampling in R. dalmatina and analysis of Brevinin-1 Da
We analysed skin secretions in two ontogenetic stages: 30 days after start of the experiment, when larvae were in development stage 36.9 ± 1.4 (mean ± SD; according to Gosner) and 14 days after completion of metamorphosis. Following the procedure described in [92] we collected skin secretions non-invasively by bathing tadpoles or froglets individually in 5 ml polypropylene tubes containing 5 ml collection buffer with 0.1 mM norepinephrine (NE) bitartrate for 15 min. After removing animals, we acidified the NE solution by adding 50 µL of 99% trifluoracetic acid (TFA), to reach a final concentration of 1 V/V%. Samples were stored at − 20 °C until further analyses. As indicated by preliminary assessments of Brevinin-1 Da concentrations, we had to pool groups of three samples within treatments to obtain detectable quantities of the targeted AMP. As a pre-step of purification, we activated each reverse-phase Sep-Pak cartridges (200 mg, LiChrolut RP-18, Merck-Millipore) with 2 ml acetonitrile and subsequently rinsed with 2 ml solvent A (HPLC-grade water with 0.12 V/V% TFA; according to [93]). Next, we loaded the skin extracts onto the cartridges, saved the solution, rinsed cartridges with 2 ml solvent A, and loaded the saved solution onto the cartridges again. Finally, we rinsed cartridges with 4 ml solvent A. We eluted the purified skin secretion with 2 ml solvent B (70:30 acetonitrile:HPLC grade water, acidified with TFA to a final concentration of 0.1 V/V%) into 2 ml polypropylene tubes and dried samples using a vacuum centrifuge (Savant, Integrated Speed Vac System, ISS 100).
We accomplished peptide identification and quantification using nano-UHPLC-MS/MS liquid chromatography-mass spectrometry, with a Maxis II ETD QqTOF (Bruker Daltonics, Bremen, Germany) coupled to an Ultimate 3000 nanoRSLC system (Dionex, Sunnyvale, CA, USA) under the control of Hystar v.3.2 (Bruker Daltonics, Bremen, Germany). We dissolved samples in 2% acetonitrile and 0.1% formic acid in water, out of which 5 µl were injected onto an Acclaim PepMap100 C-18 trap column (100 µm × 20 mm, Thermo Scientific, Sunnyvale, CA, USA). We performed sample desalting and preconcentration with 0.1% TFA for 8 min with a flow rate of 5 µl/min. Peptides were separated on an ACQUITY UPLC M-Class Peptide BEH C18 column (130 Å, 1.7 µm, 75 µm × 250 mm, Waters, Milford, MA, USA) at 48 °C, using a flow rate of 300 nl/min. We used the following HPLC solvents: solvent A containing 0.1% formic acid in water and solvent B containing 0.1% formic acid in acetonitrile, with the gradient: 4% B from 0 to 11 min, followed by a 120 min gradient to 50% B, then elevated the concentration of solvent B to 90% in 1 min and kept it there for 10 min. After each sample a blank was run to avoid carry-over. Sample ionization was achieved in the positive electrospray ionization mode via a CaptiveSpray nanoBooster ion source with capillary voltage set to 1300 V, at 0.2 Bar nanoBooster pressure. The drying gas was heated to 150 °C with 3 L/min flow rate. For external mass calibration, we used the low concentration tuning mix from Agilent technologies via direct infusion. Internal mass calibration was performed via lock mass for each run using sodium formate with the following ion transfer parameters: prepulse storage 10 µs, collision transfer 10 µs, quadrupole ion energy 5 eV, Funnel 1 RF 400 Vpp, Multipole RF 400 Vpp. The collision RF was set to 1200 Vpp with 120 µs ion transfer time. We identified the chromatographic peaks as Brevinin-1 Da by comparing retention time and mass spectrum of the 453.3551 4+ fragment to a commercially purchased molecular standard (Biocenter Kft., Szeged, Hungary). We analysed chromatograms using the software Compass Data Analysis (version 4.0, Bruker daltonics Inc., Billerica, USA) to obtain chromatogram area values as quantity estimates for statistical analysis.

Skin toxin sampling in B. bufo and analysis of bufadienolides
We collected bufadienolide samples from tadpoles preserved at development stage 36.3 ± 0.9 (mean ± SD; according to Gosner) and from 14 day old toadlets (that finished metamorphosis 14 days earlier). We homogenized whole bodies with a homogenizer (VWR VDI 12) equipped with a dispersing tool (IKA S12N-7S), dried samples under vacuum at 45 °C using a rotary evaporator (Büchi Rotavapor R-134, Flawil, Switzerland), weighed dry mass (dry body mass henceforth) and re-dissolved samples in 1 ml absolute HPLC-grade methanol, aided by brief exposure to ultrasound in a bath sonicator (Tesla UC005AJ1). As the last step of sample preparation, we filtered samples through FilterBio nylon syringe filters (pore size = 0.22 μm) and stored them at − 20 °C until further analyses.
We analysed bufadienolide compounds by means of high-performance liquid chromatography coupled with diode-array detector and electrospray ionization mass spectrometry (HPLC-DAD-ESI-MS). We identified the chromatographic peaks as bufadienolides based on the UV spectrum [34] and by comparing their retention time and mass spectrum to those of the following commercially available standards: bufalin, bufotalin, resibufogenin, gamabufotalin, areno-and telocinobufagin (Biopurify Phytochemicals, Chengdu, China), cinobufagin (Chembest, Shanghai, China), cinobufotalin (Quality Phytochemicals, New Jersey, USA) and digitoxigenin (Santa Cruz Biotechnology, Dallas, TX, USA) or to compounds established by analysing a large sample obtained from an adult male common toad by gently massaging the parotoid glands.
We performed HPLC-MS measurements on a Shimadzu LC-MS 2020 instrument (Shimadzu, Kyoto, Japan) that consisted of a binary gradient solvent pump, a vacuum degasser, a thermostated autosampler, a column oven, a diode array detector and a single-quadrupole mass analyser with electrospray ionization (ESI-MS). Chromatographic separations were carried out at 35 °C on a Kinetex C18 2.6 µm column (100 mm × 3 mm i.d., Phenomenex) in series with a C18 guard column (4 mm × 3 mm i.d.) using 10 µL injections. Eluent A was 5% aqueous acetonitrile with 0.05% formic acid and eluent B was acetonitrile with 0.05% formic acid. The flow rate was 0.8 mL/min and the gradient was as follows: 0-2 min: 10.

Statistical analyses
We analysed the data on the two species and life stages separately. We excluded two individuals due to extremely slow development (development stage was lower than the mean by more than 3 SD), and accidentally lost three individuals. This resulted in the following sample sizes used in the analyses: R. dalmatina; control: 35 We averaged GE values obtained from qPCR runs for each sample and subsequently rank-transformed means, because GE values that fall outside the standard interpolation curve are not estimated reliably.
In R. dalmatina we compared Brevinin-1 Da quantity among treatments using general linear models (GLM) entering log-transformed values of the chromatogram areas as the dependent variable. We used log-transformation to enhance normality of model residuals and homogeneity of variances. In case of 14 days old froglets, we only had 5-5 replicates in Bd-exposed treatments because of mortality. For each triplet, we calculated mean values of body mass, development stage and infection intensity. The initial model included treatment as a fixed factor, and body mass, development stage and infection intensity as covariates.
To calculate the number of bufadienolide compounds (NBC) present in each B. bufo individual, we assumed a compound to be present if the signal-to-noise ratio (S/N) of its peak was at least three. We estimated the quantity of each compound from the area values of chromatogram peaks based on the calibration curve of the bufotalin standard, and summed up these values to obtain an estimate of total bufadienolide quantity (TBQ) for each individual (for a similar approach see [34,94]). We analysed effects of treatment on both toxin variables separated by life stages using GLMs. In case of TBQ, we entered as the dependent variable logtransformed values in case of tadpoles and square-root transformed values in case of metamorphs to enhance normality of model residuals and homogeneity of variances. Initial models included treatment as a fixed factor and dry mass, development stage and infection intensity as covariates. In case of a significant treatment effect, we used Tukey HSD post-hoc tests to reveal significant differences between treatment groups. When a covariate also had a significant effect, we ran post-hoc tests on residuals extracted from the regression of the dependent variable on the covariate.
We analysed body mass data using GLMs. Initial models included treatment as a fixed factor and an estimate of the speed of development (Gosner stage in case of tadpoles and length of larval development in case of metamorphs) as a covariate, as well as their interaction. To enhance normality of model residuals and homogeneity of variances, we entered log-transformed values of body mass in case of R. dalmatina tadpoles and B. bufo metamorphs. In case of the high Bd dose treatment in B. bufo tadpoles, and both Bd treatments in metamorphs, where Bd prevalence was high, we also investigated the effect of infection intensity on body mass in the Bd-exposed treatment groups with GLMs, including infection intensity and development stage as covariates. We analysed variation in development stage using GLMs with treatment as a fixed factor, except for R. dalmatina metamorphs where we used Kruskal-Wallis tests due to the non-normal distribution of model residuals and inhomogeneity of variances. In case of B. bufo tadpoles in the high Bd dose treatment we analysed the relationship between infection intensity and development stage using Spearman rank correlation and Pearson correlations in both Bd treatments of metamorphs.
We verified normal distribution of model residuals using Shapiro-Wilk tests and by inspecting diagnostic plots, and homogeneity of variances using Levene's tests. We applied a backward stepwise model simplification procedure [95] to avoid potential problems due to the inclusion of non-significant terms [96]. We obtained statistics for removed variables by re-entering them one by one to the final model. All tests were two-tailed. Statistics were calculated using SPSS Statistics 20.0 for Windows.