Direct and Latent Effects of Pathogen Exposure Across Native and Invasive Amphibian Life Stages

Emerging infectious diseases are one of the multiple factors contributing to the current “biodiversity crisis”. As part of the worldwide biodiversity crisis, amphibian populations are declining globally. Chytridiomycosis, an emerging infectious disease, caused by the fungal pathogen Batrachochytrium dendrobatidis (Bd), is a major cause of amphibian population declines. This fungus primarily affects keratinized structures in larval, juvenile, and adult amphibians as well as heart function. However, we know little about how Bd can impact embryos as well as potential latent effects of Bd exposure over ontogeny. Using two different Bd strains and multiple exposure times, we examined the effects of Bd exposure in Pacific chorus frog (Pseudacris regilla), Western toad (Anaxyrus boreas) and American bullfrog (Lithobates catesbeianus) life stages. Using a factorial experimental design, embryos of these three species were exposed to Bd at early and late embryonic stages, with some individuals re-exposed after hatching. Embryonic Bd exposure resulted in differential survival as a function of host species, Bd strain and timing of exposure. P. regilla experienced embryonic mortality when exposed during later developmental stages to one Bd strain. There were no differences across the treatments in embryonic mortality of A. boreas and embryonic mortality of L. catesbeianus occurred in all Bd exposure treatments. We detected latent effects in A. boreas and L. catesbeianus larvae, as mortality increased when individuals had been exposed to any of the Bd strains during the embryonic stage. We also detected direct effects on larval mortality in all three anuran species as a function of Bd strain, and when individuals were double exposed (late in the embryonic stage and again as larvae). Our results suggest that exposure to Bd can directly affect embryo survival and has direct and latent effects on larvae survival of both native and invasive species. However, these impacts were highly context dependent, with timing of exposure and Bd strain influencing the severity of the effects.


INTRODUCTION
In many organisms, exposure to stressors during embryonic or prenatal stages can result in both direct and latent effects on subsequent developmental stages. For example, in birds, conditions experienced at early life are determinants of fitness in adults (1); in snapping turtles (Chelydra serpentina), hatchling from eggs incubated on a wet substrate have an improved locomotor performance in comparison to hatchlings from drier substrates (2) and in fishes, embryos of pink salmon (Oncorhynchus gorbuscha) exposed to oil have a reduction in juvenile growth and survival (3). In amphibians, these effects can be on individual growth rates, behavior, locomotion or immunology (4)(5)(6)(7). Thus, exposure to predator cues in the pinewoods tree frog Hyla femoralis slowed larval growth and development, resulting in metamorphs with relatively smaller body sizes (8). In amphibians, repeated exposure at early life stages to other environmental stressors, such as contaminants, predator cues, and pathogens, can produce latent effects in juvenile and adult amphibians (4,9,10). Particularly in amphibian embryos, environmental cues can cause significant changes in hatching traits (11)(12)(13)(14). Time of hatching can change in response to risks and opportunities. For example, embryos can hatch early to escape predators and pathogens, and this life history shift can have effects that persist through later life stages (15)(16)(17). Additionally, embryos infected with water molds can suffer differential mortality rates relative to the timing of exposure to the pathogen (18). As such, the timing of pathogen exposure might play a critical role on host susceptibility to infection (19).
Host ontogeny is a key factor for examining or predicting disease dynamics. There is precedence in different systems than minimum changes in the history of exposure to pathogens during a particular developmental stage can drastically change host life history. Changes in individual susceptibility to pathogens occur throughout ontogeny in many organisms, including plants (20), insects (21), birds (22), reptiles (23), mammals (24) and amphibians (25,26). The key, however, to understanding temporal association between pathogens and susceptibility is to empirically discern latent and direct effects within and across life history stages. We posit that amphibians can be model systems for testing these questions as they are a taxon of conservation concern, have complex life histories, and are susceptible to multiple emerging infectious diseases.
We explored the direct and latent effects of Bd exposure on both the embryonic and larval stages using three amphibian species with differential susceptibility to native and invasive Bd strains (33,34,46,47). We posit that amphibian embryos will be susceptible to the chytrid fungus as Bd can produce enzymes that can destroy tissue (48)(49)(50). Further, the release of fungal toxin (30,51) could impact embryos by delaying growth, triggering key transitions resulting in ontogenetic shifts or latent effects on life history trajectories. Finally, we hypothesized that Bd can impact embryos by depletion of oxygen. The lack of dissolved oxygen slows down the development in Bufo bufo (52) and hypoxia can kill early life stages (13). Direct or latent effects may also vary with Bd strain and with host species, therefore evaluating different strains is critical to disentangle intrinsic aspects of the pathogen as virulence and how it changes among hosts. We also examined the influence of Bd exposure on larval survival predicting that repeated exposure to Bd across the embryonic/larval transition would result in decreased survival.

MATERIALS AND METHODS
We studied three anuran species found in the US Pacific Northwest (PNW). These are the Pacific chorus frog (Pseudacris regilla) a common species throughout its PNW range found in a variety of habitats from sea level to montane regions, the Western toad (Anaxyrus boreas) whose populations have experienced declines across much of its historic range and American bullfrogs (Lithobates catesbeianus) an introduced species in the PNW (53)(54)(55). Twenty clutches of P. regilla were collected from Little Three creeks on 19

Pre-hatch Exposure Regime
Bd exposure treatments were administered in either the early embryonic developmental stages or closer to hatching. Early exposure (early) corresponded to the late gastrula stages, or Gosner Developmental Stage 12 (56) while closer to hatching exposure (late) corresponded to embryos capable of muscular response, or Gosner Developmental Stage 18 (56). Bd strains (i.e., the isolate of the fungus used for the inoculation) included an endemic Bd strain (JEL 630, hereafter "West, " isolated from L. catesbeianus in Oregon), and a novel Bd strain to Oregon freshwater habitats (JEL 627, hereafter "East, " isolated from L. catesbeianus in Maine USA). These strains were identified as part of the North American clade (Bd-GPL-1) (57). We obtained cryogenically preserved culture plates from J. Longcore to prepare 1% sterile tryptone-agar media plates with 0.5 ml of stock Bd broth coming from each particular strain of the fungus (31). Bd cultures were allowed to grow for 5-8 days at 20 • C before used in the experiment (31). Using a hemocytometer, we quantified the zoospores from a pooled inoculation broth (8-12 plates per Bd strain). Five ml inoculations of the zoospore broth (30 K zoospores/ml) were then administered to experimental units (18 cm H × 10 cm OD high-density polyethylene graduated beakers) containing 800 ml of dechlorinated water. A similar dose was previously tested in larvae of P. regilla (47), A. boreas (33,58,59), and L. catesbeianus (32,60). Controls were inoculated with a sham inoculum created by rinsing the same number of sterile agar plates with 5 ml of dechlorinated water.
Using a factorial experimental design, each group of eggs was assigned to a time of exposure treatment (early, late) and a Bd strain treatment (West, East, Control) (Figure 1, Pre-Hatching). Sixty experimental units (581 total eggs) were assigned for P. regilla (10 replicates per Early and Late treatment groups), 51 experimental units (506 total eggs) for A. boreas (8 replicates per Early exposure treatment, 9 replicates per Late exposure treatment), and 30 experimental units (600 total eggs) for L. catesbeianus (5 replicates for all groups, except for the East and West/Late exposure groups, which had 6 and 4 replicates, respectively) ( Table 1). The length of the pre-hatching phase varied by species, lasting 19 days for P. regilla and A. boreas, and 22 days for L. catesbeianus. Embryos that died were preserved individually in 2.0 ml Eppendorf tubes with 95 % ethanol. No water changes were performed during the pre-hatching phase as movement associated with water changes can induce hatching, thus influencing our results. Upon hatching at Gosner stage 21, water changes were conducted weekly. We quantified the time of hatching by direct observation, and hatchling events and survival were recorded twice per day.
To analyze differences in hatching time, we compared proportions between treatments of eggs being exposed to Bd and control treatments (no exposure to the pathogen) using quasibinomial generalized linear models (GLM) performed independently per species. All analyses were run in R (61). To evaluate differences among strains and controls we calculated pairwise comparison using a Tukey HSD test.

Post-hatch Exposure Regime
Upon hatching, survivors were pooled within pre-hatch treatment groups (Early or Late; East, West, Control) to standardize sample sizes for the post-hatch phase of the experiment. In this phase, larvae were either re-exposed to the same pre-hatch Bd strain or held as controls to estimate latent effects (Figure 1, Post-Hatch). For P. regilla, we had a total of 328 surviving hatchlings distributed across the larval exposure treatments for a total of 82 experimental units, resulting in 33 control replicates, 21 East strain replicates, and 28 West strain replicates. Less than 10% of eggs hatched from the East/Late pre-hatch exposure treatment group; as such, there was no continuation of this treatment in the post-hatch phase. For A. boreas, we ran 42 control replicates, 28 East strain replicates, and 26 West strain that contained a total of 384 surviving hatchlings for a total of 96 experimental units. For L. catesbeianus, we ran 23 control replicates, 17 East strain replicates, and 16 West strain using a total of 228 surviving hatchlings with a total of 56 experimental units. Due to complete mortality in the East/Early and the West/Early pre-hatch phase, these treatments were not continued in the post-hatch phase ( Table 2).
Larvae were held individually and those that were re-exposed to Bd were re-inoculated once a week (every 7 d) for the duration of the experiment. Individuals were held in rectangular plastic containers (31 × 18 × 8 cm) filled with 2,000 ml dechlorinated water. Water changes occurred concurrently with re-inoculation using 5 ml of 50 K zoospores/ml. Animals that died during the experiment were preserved in 95% ethanol. At the end of the experiment, animals remaining alive were humanely euthanized in accordance with institutional animal care protocol in MS-222 (Tricaine methanesulfonate) and then preserved in 95% ethanol. The experimental trials for each species lasted until individuals reached Gosner stage 30-31 (distinctive rear limb bud) or death. Lithobates catesbeianus 5 (100) 5 (100) 5 (100) 6 (120) 5 (100) 4 (80) TABLE 2 | Number of replicated groups exposed in the different treatments per species as larvae.

Pre-hatching treatments
Bd strain × Time

Control East West
Early Late Early Late Early Late

Lithobates catesbeianus
Post-hatch Control 6 (24) 6 (24) 0 (0) 7 (28) 0 (0) 4 (16) Total duration for the experiment was 65 days for P. regilla, 59 days for A. boreas and 19 days for L. catesbeianus. We monitored survival twice per day and quantified developmental differences through time by staging all larvae (Gosner stage) every week during water changes. At the end of the post-hatch phase, we sampled a subset of all Bd-exposed animals of each species and also randomly sampled 5 control animals of each species to confirm no contamination happened. To assess infection load at the termination of the experiment, we dissected larvae mouthparts for P. regilla individuals, and we swabbed mouthparts using fine tipped sterile rayon swabs (Medical Wire and Equipment MW&E 113) for A. boreas and L. catesbeianus. Both protocols, swabbing and cutting mouthparts, are recommended as adequate protocols for assessing infection loads. Excised mouthparts and swabbing are similar in the likelihood of detecting Bd infection regardless of developmental stage and larval size (62)(63)(64).
Each sample was analyzed using quantitative polymerase chain reaction (qPCR) following the methods of (65). A small modification of the amount of Prepman Ultra (Applied Biosystems R , Life Technologies) was used to extract the DNA; we used 60 µL instead of 40 µL (66). Our extractions were diluted 1:10 and each sample was analyzed in triplicate to quantify the average number of genome equivalents per animal (7,500 realtime PCR Applied Biosystems instrument). To analyze infection loads, we log transformed the qPCR results as log (genome equivalents per individual + 1) to normalize data.
Effect of exposure on survivorship was analyzed independently by species using odds ratios calculated with a generalized linear mixed model, family: binomial (logit). The values of the ratios represent the likelihood or the risk of mortality due to exposure to the pathogen in comparison to the controls. Therefore, odds ratios higher than 1 represent an increased risk after exposure, odds ratios equal to one represent no difference in the risk, and odds ratios lower than 1 represent a lower risk of the exposed group. All analyses were run in R (version 4.0.3).

Pre-hatching Phase
Pseudacris regilla embryos exposed to both the East and West Bd strains in the Early exposure groups had a lower proportion of hatchlings relative to controls (East strain: t = −4.40, p < 0.001; West strain: t = 1.99, p = 0.04). A post hoc Tukey test showed that this proportion was different in embryos exposed to the East strain in contrast to the control (z = −4.45, p < 0.001) and the West strain (z = 6.14, p < 0.001), with approximately a 50% hatching rate (Figure 2A). Reduced hatching was also found in the Late/East treatment group (t = −11.03, p < 0.001) relative to the Late/West treatment (t = −1.29, p = 0.19). In fact, <10% of embryos hatched after being exposed late in development to the non-native East strain ( Figure 2B). In A. boreas, the proportion of embryos that hatched was similar across both strains in comparison to controls across the Early (East strain: t = -0.49, p = 0.62; West strain: t = 0.62, p = 0.53) ( Figure 2C) and Late exposure treatments (East strain: t = 1.31, p = 0.19; West strain: t = 0.73, p = 0.46) (Figure 2D).
The proportion of L. catesbeianus embryos that hatched was low when embryos were exposed early in development, with lower survival in the West Bd strain treatment relative to controls (West strain: t = 3.58, p < 0.001) (Figure 2E). There were no survivors in the East strain exposure treatment. The estimate of Bd strain as a factor in our model was high (5,329), potentially due to the 100% mortality, making the t and p-value not significant (t = 0.003, p = 0.99). The proportion of embryos that hatched in the Late exposure treatment was lower across both Bd strains in comparison to the controls (East strain: t = 2.89, p < 0.01; West strain: t = 2.13, p = 0.03) (Figure 2F). A post hoc Tukey test showed that this proportion was different in embryos exposed to the East strain in contrast to the control (z = 2.89, p < 0.01), but it was not different for embryos exposed to the West strain (z = 2.13, p = 0.08).

Post-hatching Phase
Our generalized linear mixed model quantified as odds ratios (OR) the effects of exposure to a particular strain on larvae Frontiers in Veterinary Science | www.frontiersin.org mortality in comparison to the controls given their history of exposure as embryos. As such, results are reported as an increase or decrease in odds of mortality.
Direct Effects on Larvae -Only Exposed to Bd as Larvae (Control-Bd) We found evidence for direct effects of Bd exposure on larval mortality for the three species. In P. regilla, post-hatching exposure to the East strain increased the odds of mortality (OR Early/Control−East 8.88, p = 0.01, CI. 1.43-54.85, Figure 3, Left panel). For A. boreas, we found that individuals exposed during the post-hatch phase to the West strain had lower odds of mortality relative to controls (OR Early/Control−West 0.12, p = 0.03, CI: 0.018-0.84, Figure 4, Left panel). In contrast, larvae coming from the Late control group and exposed posthatch to East or West had higher odds of mortality than controls (OR Late/Control−East 14.38, p = 0.03, CI: 1.19-173.65; OR Late/Control−West 19.56, p = 0.03, CI: 1.32-288, Figure 4, Right panel). Larvae of L. catesbeianus increased their odds of mortality when exposed to either East or West strain (OR East 9.9, p = 0.04, CI: 1.06-92; OR West 539, p < 0.001, CI: 29.64-9,801) ( Figure 5).

Latent Effects on Larvae -Only Exposed to Bd as Embryos (Bd-Control)
We did not find evidence for latent effects in P. regilla. In A. boreas odds of mortality changed according to the time of exposure and Bd strain. Odds of mortality for larvae decreased when embryos were exposed early in development to the West strain of Bd (OR Early/West−control 0.14, p = 0.02, CI: 0.026-0.73, Figure 4, Left panel). On the contrary, individuals exposed Late as embryos to the East strain had higher odds of mortality than controls (OR Late/East−control 10.62, p = 0.04, CI: 1.07-105, Figure 4, Right panel). In L. catesbeianus, we found higher odds of mortality than controls for both Bd strains (OR Late/East−control 31, p = 0.001, CI: 3.5-272, OR Late/West−control 23.21, p = 0.006, CI: 2.42-222.14, Figure 5).

Double Exposed Treatments -Exposed to Bd as Both Embryos and Larvae (Bd-Bd)
We found evidence that exposure to Bd in both the embryonic and larval stages affect the larval odds of mortality in all three species. We found P. regilla that were re-exposed to the West FIGURE 3 | Odds ratio (OR) for P. regilla tadpoles according to their exposure regimes. Label of x-axis includes a fraction that indicates in the numerator the exposure regime as embryo and in the denominator exposure regime post-hatching (no exposure = NE, East or West for Bd chytrid strain). A dashed line indicates value 1. OR > 1 higher risk after exposure, OR = 1 no risk difference, OR < 1 lower risk after exposure. A star (*) indicates treatments with significant odds ratios. Effects are represented as follows: Direct effects -only exposed to Bd as larvae-are represented by circles, latent effects -only exposed as embryos-are represented as four squares, and 2-fold effects -exposed as embryos and larvae-are represented with filled squares; controls are represented by triangles.

Infection Loads
We confirmed Bd infection using real-time PCR analyses of tadpole mouthparts for P. regilla and swabs for A. boreas and L. catesbeianus. All tadpoles from control treatments tested negative for Bd. Individuals from the three species only exposed as larvae (direct effects) were positive for the East strain. From individuals only exposed as embryos (latent effects), we found that A. boreas tested positive for Bd when exposed late as embryos with the East strain. In L. catesbeianus, individuals tested positive when exposed late to either of the strains ( Table 3). In the group of re-exposed individuals (2-fold effects), P. regilla was reported Bd positive for both East and West strain when exposed early as embryos. In A. boreas, individuals were positive for Bd when exposed early to the West strain and late when exposed to the East strain ( Table 3).

DISCUSSION
Life stage, time of exposure, and Bd strain influenced susceptibility to Bd in the embryo-larvae life history transition for three anuran species: P. regilla, A. boreas, and L. catesbeianus. We detected direct effects of Bd on embryonic and larval mortality, latent effects across the embryo/larval transition, and additive effects when double-exposed to Bd across both life stages. Exposure of embryos to Bd resulted in direct impacts on hatchling survivorship. We found direct, negative impacts of Bd strain and time of exposure on embryonic survival and proportion of hatching success for P. regilla and L. catesbeianus. Embryos of P. regilla were drastically affected by the nonnative East Bd strain, resulting in more than 90% mortality when exposed later in embryonic development. Interestingly, embryos of invasive L. catesbeianus died when exposed to either Bd strain (East or West). When exposed early in embryonic development to the East strain, the number of viable hatchlings was zero and we detected a mortality of 90% in hatchlings after early exposure to the West Bd strain. When exposed later in development (East or West strains), only 50% of embryos hatched. The influence of time of exposure may be explained by changes in the thickness of the jelly layer surrounding the embryo through development. This jelly layer becomes thinner over development, thus late-stage embryos may be more affected by exposure to a pathogen (15,18). Amphibian species-specific egg deposition forms (films, strings, clusters), FIGURE 4 | Odds ratio (OR) for A. boreas tadpoles according to their exposure regimes. Label of x-axis includes a fraction that indicates in the numerator the exposure regime as embryo and in the denominator exposure regime post-hatching (no exposure = NE, East or West for Bd chytrid strain). A dashed line indicates value 1. OR > 1 higher risk after exposure, OR = 1 no risk difference, OR < 1 lower risk after exposure. A star (*) indicates treatments with significant odds ratios. Effects are represented as follow: Direct effects -only exposed to Bd as larvae-are represented by circles, latent effects -only exposed as embryos-are represented as four squares, and 2-fold effects -exposed as embryos and larvae-are represented with filled squares; controls are represented by triangles.
as well as morphology of egg structure, provide differential protection from pathogens (67).
Post-hatching exposure resulted in both direct and latent impacts on larval survivorship. Direct effects on larvae are reported mainly as an increase in the odds of mortality for all three anuran species. P. regilla was negatively affected by exposure to the non-native East Bd strain, while A. boreas and L. catesbeianus were affected by both strains (East and West). Odds of mortality in A. boreas were higher when exposed to the West strain (19.56) than when exposed to the East strain (14.38). On the contrary, the odds of mortality in L. catesbeianus were higher when exposed to the East strain (9.9) than when exposed to the West strain (5.39). This result was not wholly unexpected as larvae mortality has been reported in experimental studies exposing these same species to Bd. A. boreas has been identified as particularly susceptible to Bd (30, 46) while P. regilla and L. catesbeianus larvae have relatively high survivorship (30,33,46,68). In this study, we found a direct effect of Bd on larval survivorship for all three species. The increase in the odds of mortality in P. regilla and L. catesbeianus larvae can be explained by the origin and characteristics of the East strain. Isolated from L. catesbeianus in Maine (USA), this strain has been identified as hypervirulent (57,69,70) and categorized as part of the North American clade in the Global Pandemic Lineage (GPL) (57). As such, we anticipated an increase in larval mortality due to a lack of evolutionary relationship with this strain. However, L. catesbeianus larvae were also susceptible to the East strain even though it was isolated from their conspecifics within their native range.
We detected an increase in the odds of mortality in P. regilla when exposed early in embryonic development to Bd while the odds ratio for larval mortality were lower for later exposure. We could not evaluate potential latent effects after late exposure of P. regilla embryos to the East strain. Effects of Bd differed in embryos according to the timing of exposure. However, these effects of timing of exposure and duration of stressors such as pathogens and its relation to critical window and disease is not broadly tested (71,72). We found an increase in the odds of larval mortality of both A. boreas and L. catesbeianus as a function of Bd strain and timing of embryonic exposure. In A. boreas, embryos exposed early to the West strain showed a decrease in the odds of mortality. Conversely, when A. boreas were exposed to the East strain late in embryonic development, larvae were almost 10 times more likely to die than control individuals. There was a similar increase in the odds of larval mortality in L. catesbeianus when exposed as embryos to any of the Bd strains. The high mortality rates in L. catesbeianus when exposed early to Bd prevented us from understanding potential latent effects for this invasive species. Our results support the hypothesis that timing of pathogen exposure is a major factor that influences host survivorship. Exposures of amphibian embryos to stressors such as pathogens at early stages of development can trigger effects over ontogeny (73). This exposure can later modify other characteristics, such as growth rates (74), mortality rates (75), mass (5), and development of immune response (76).
We also found effects of double Bd exposure (exposed in both the embryonic and larval stages) in all three anuran species. All species showed an increase in the odds of larval mortality when the first Bd exposure occurred at a later embryonic developmental stage (Gosner stage 18). In P. regilla, odds of mortality increased after double exposure to the West strain. Exposure to either strain (East or West) increased the FIGURE 5 | Odds ratio (OR) for L. catesbeianus tadpoles according to their exposure regimes. Label of x-axis includes a fraction that indicates in the numerator the exposure regime as embryo and in the denominator exposure regime post-hatching. A star (*) indicates treatments with significant odds ratios. Effects are represented as follow: Direct effects -only exposed to Bd as larvae-are represented by circles, latent effects -only exposed as embryosare represented as four squares, and 2-fold effects -exposed as embryos and larvae-are represented with filled squares; controls are represented by triangles.
odds of mortality in A. boreas and L. catesbeianus. Double exposure effects have been reported in experiments examining the larval/metamorph transition (77), thus our experiment provides additional information concerning other life history transitions providing a comprehensive view of the jeopardy through life.
The differential response of A. boreas to Early/Late and East/West Bd treatments may be explained by the presence of a potential critical window of vulnerability for this species and by the virulence of the Bd strain. Late exposure of A. boreas embryos to the East strain increased the odds of larval mortality of this species. Fernandez-Beneitez et al. (18) found that embryos of toads (Bufo calamita and Pelobates cultripes) exposed to Saprolegnia spp. at an early developmental stage (Gosner stage 12) suffered no increase in mortality, while embryos challenged at later stages of embryonic development (Gosner stages 15 and 19) were sensitive to the pathogen dying 72 h after exposure. Understanding which species experience latent effects will help target management efforts by identifying how exposure in particular life history stages can change host response. If a species is identified as being particularly susceptible to exposure to Bd as embryos, actions such as ex situ protection could be useful for its conservation.
Our findings complement the information on susceptibility of P. regilla to Bd as larvae of this species had previously been reported to be tolerant to certain Bd strains (30,46). Interestingly, we found that this tolerance can change with an individual's previous exposure regime to non-native strains. Our experimental evaluation revealed that Bd strains from an invasive species can have harmful consequences on native and even invasive conspecific hosts. Our findings for A. boreas support previous work showing species as being susceptible to both the East and West strains of Bd (30,46,47). In L. catesbeianus, larvae and adults have been reported as able to withstand infection loads of the chytrid in different regions (78) and this species is suggested as an asymptomatic carrier or reservoir of Bd (79,80). Our results indicate that larvae can also be susceptible to Bd but this response will be mediated by previous exposure in an early life stage. Individuals that were re-exposed to Bd were about 50 times more likely to die than individuals kept as controls. This contrasts with previous experimental studies reporting this species as a carrier of Bd (30,32,60). Generally, those studies directly exposed individuals in the larval stage ] without considering previous exposure regimes. In our study, L. catesbeianus were vulnerable to Bd exposure in response to direct exposure and across life history transitions. We found species-specific embryonic mortality after exposure to Bd. Many pathogens impact anuran embryos, including ranavirus (76), oomycetes (81)(82)(83)(84), filamentous ascomycetes (85) and microsporidia (86). The mechanisms of protection offered by the vitelline membrane against ranavirus are unknown. However, bacteriostatic activity of the egg membrane or the capsule (87) or just its role as structural barrier when exposed to contaminants have been proposed (88,89). In the case of Oomycetes, zoospores are chemotactic and move toward suitable substrates where they can germinate and grow. Similarly, ascomycetes are able to grow through the jelly. Few studies have suggested how Bd affects anuran embryos. Bd enzymatic action is one mechanism that could explain this result, as it can cause damage in skin tissue of hosts after exposure (30,49,90,91). A complex mix of proteolytic and hydrolyzing enzymes (esterases) that degrade amphibian tissue have been described from different Bd isolates (90)(91)(92). In addition, many hatching anurans release enzymes to assist with degradation of the egg capsule at the moment of hatching (93,94); this could potentially facilitate the enzymatic action of Bd to degrade tissues. Recently, dose-dependent mortality and proliferation in zebrafish (Danio rerio) tissue was reported with toxins secreted after the establishment of Bd sporangia (95).
The present study offers useful information about the complexity of host response to a pathogen, particularly with multiple exposures across life stages. Our study provides information about direct effects of Bd on anuran embryos, with significant impacts on mortality and the proportion of hatching success. Our results also quantified latent effects of Bd exposure over ontogeny (96). Despite being a relatively brief period, exposure to Bd in the egg led to increased mortality after hatching and species-specific differences were due to the timing of embryonic exposure and re-exposure in the larval stage. As we increase our understanding of how Bd impacts amphibians through direct and even latent effects, we are recognizing that the effects Bd may have on population dynamics and conservation of amphibians have been underestimated in wild populations.
Additional research exploring the mechanisms protecting the embryos is needed to better understand the susceptibility of this developmental stage to disease. Characteristics such as jelly thickness and composition, or size of the capsule, can be involved in resistance to chytrid. As eggs received material from their parents during oviposition, evaluating the role of parents in the immune response of their offspring can help us to understand more about embryonic immunity. As continued survey efforts have located Bd in populations of amphibians around the world, there is growing evidence that the risk of Bd cannot be simplified to species susceptibility. Instead, Bd risk includes strain, host life-stage, and specific exposure scenarios. Further studies are also required to better understand how variation in other environmental and biological parameters can affect the outcome of repeated Bd exposure in anuran species. Our results add information to the growing body of evidence concerning differential susceptibility to pathogens among amphibian species and across life stages.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by Institutional Animal Care and Use committee Oregon State University.

AUTHOR CONTRIBUTIONS
JU, AB, and TG conceptualized the manuscript. JU and EB performed the field work, experiments and data analysis. JU, EB, AB, and TG have made a significant, direct and intellectual contribution to the work. All authors contributed to the article and approved the submitted version.