Our results demonstrate that Ranavirus is extremely transmissible among juvenile wood frogs in terrestrial environments by both direct and indirect routes of transmission. In the direct contact experiment, nearly every frog (99.2% overall) cohoused with an infective acquired an infection, even in as little as little as an hour (95.8% for 1hr). Such high infection rates with short periods of exposure suggests that even briefly sharing cover in the forest (e.g., under leaf litter, in a mammal burrow) could result in transmission. Similarly, almost every frog (96.8% overall) became infected after being exposed to contaminated soil and leaf litter, even if the infective frog had been absent from the environment for two full days (100% for 48hr). These results provide evidence that Ranavirus can persist for multiple days and still retain the capacity to infect later-arriving frogs that are moving and sitting in the leaf litter. We discuss each hypothesis below.
Transmission efficiency and intensity during the direct contact experiment.
Our findings support our first hypothesis that transmission intensity increased (and therefore viral load increased) with the amount of time frogs were in direct contact (P = 1.9*10− 9) (Table 1). In addition to exposure time, cohort was the other significant covariate explaining viral load (Table 1). This cohort effect was likely due to high moisture content in the second cohort and demonstrates the importance of moisture in transmission. As for transmission efficiency, the predominantly positive results (99.2%) indicate that Ranavirus is highly transmissible among frogs via direct terrestrial exposure to an infective across all exposure times (Fig. 1). The only negative result in the direct contact experiment occurred in a 1-hour exposure, the shortest exposure period, which fits our hypothesis, but does not conclusively show that longer exposure periods increase transmission efficiency.
Direct contact among tadpoles is a well-established transmission route (Brunner and Yarber 2018). Here we demonstrate that terrestrial Ranavirus transmission among juvenile frogs can readily occur, which aligns with another recent study (Le Sage et al. 2022). The Le Sage et al. (2022) study used large terrariums to demonstrate that susceptible juvenile frogs behaviorally avoided infective frogs and they reported a lower positivity rate than we found in our direct contact experiment. Sample sizes differed among the studies: 7 of 10 susceptible frogs were infected (Le Sage et al. 2022), compared to 122 of 123 susceptible frogs were infected here (excluding NAs). Therefore, we conclude that the transmission risk from direct contact exposures is very high for juvenile frogs.
This direct contact experiment offers insights into how transmission may occur in nature. The damp soil and leaf litter containers force direct contact by limiting the space available for movement, making skin to skin contact likely even though the duration of contact was not quantified. Yet, this experimental design cannot rule out the possibility that susceptible frogs are exposed to Ranavirus from the soil, leaf litter, or air, in addition to skin contact with the infective frog. Direct contact could occur immediately after metamorphosis when juvenile frogs leave the water and move on to the moist soil surrounding wetlands, until rain permits larger movements. Wood frog juveniles may also encounter one another in leaf litter as they spend about 2 weeks traveling an average of 100m and up to 2.5km into forests, until they select a suitable overwintering habitat (Patrick et al. 2006, 2008). To reduce predation risk, wood frogs will remain in a set location for long periods of time without moving, until they make a large singular movement on a rainy night (Rittenhouse and Semlitsch 2007; Rittenhouse et al. 2009). This movement behavior may increase their risk of directly encountering each other during migration, because all frogs are moving during the same time (i.e., rainy nights). Settling juveniles preferentially select microhabitats that are desirable, regardless of how many frogs are already present, leading to very high densities of frogs in a given place (Patrick et al. 2008). Therefore, both juveniles and adults may contact other frogs in terrestrial environments which could lead to transmission events (Le Sage et al. 2022). The enclosed conditions of the experimental containers provide a reasonable model for frogs sheltering together in damp places on the forest floor.
Transmission efficiency and intensity in the indirect contact experiment.
Because Ranavirus was extremely transmissible (96.8% overall) across all wait times, we did not find strong support for our second hypothesis that transmission efficiency would decrease as the wait times between exposures increased. Every susceptible frog tested positive after entering the indirect contact experiments 48 hours after the infectives were removed, indicating that Ranavirus can persist in the environment and infect susceptible frogs. We expected that waiting longer would allow the virus in the environment to degrade, but the only 4 negative results were in the short rather than long wait time.
Similarly, we hypothesized that shorter time between exposures would correspond to higher viral loads and therefore higher transmission intensity, but this pattern was also not evident. Instead, longer waits corresponded to higher viral loads in susceptible frogs (P = 0.012) (Table 1). Qualitatively, viral load increased as wait time increased up until a peak at 24 hours, after which the viral load declined at 48-hour waits (Fig. 2). These results may be explained by variation in soil moisture content, viral loads of the infective partners, or individual responses to infection. Ambystoma Tigrinum Virus (ATV), a type of Ranavirus, can persist in moist wetland sediment and infected 86% of salamanders exposed to the soil (Brunner et al. 2007). Active Ranavirus can persist in sediment for several days under warm conditions to over a month in cool temperatures below 20 degrees C (Brunner and Yarber 2018). Wait times used here may not have been long enough for the virus to degrade or lose virulence.
Indirect contact could result in Ranavirus transmission among frogs in nature in many ways. First, sequential egg clutches or tadpoles that are slower at reaching metamorphosis will be exposed to the soil, mud, and leaf litter that earlier metamorphs have passed through and shed viral particles along. Second, given that recent wood frog metamorphs travel the furthest distances of any life stage, an infected metamorph could carry Ranavirus to new areas, inducing outbreaks (Berven and Grudzien 1990; Patrick et al. 2006). Third, moisture requirements for wood frogs lead to specific microhabitat preferences that may put them in proximity to each other, such as congregating within moist rocky ravines in oak-hickory forest (Rittenhouse and Semlitsch 2007). However, since wood frogs do not move for multiple days at a time (Rittenhouse and Semlitsch 2007), any viral particles left behind will have time to degrade before another susceptible frog passes through, which may reduce the likelihood of transmission. Fourth, indirect transmission might also occur by rainwater carrying viral particles from infected frogs located at the top of the drainage to those below. Interestingly, one study on the landscape ecology of Ranavirus demonstrated that higher elevation catchments (water systems higher up on the hill that receive little drainage from nearby surface waters) were more likely to have Ranavirus outbreaks than lower elevation drainage sites, and that proximity to a Ranavirus outbreak did not increase the likelihood of an outbreak, which are counterintuitive findings (Gahl and Calhoun 2008). Another study using manmade vernal pools to test which factors predict Ranavirus epidemics found that proximity to another outbreak site did increase Ranavirus prevalence, as expected (Youker-Smith et al. 2018). In addition, old tadpoles and high host densities had more Ranavirus (Youker-Smith et al. 2018), which means that tadpoles nearing metamorphosis may be more likely to carry Ranavirus as they metamorphose out of wetlands. The spatial ecology factors that increase the likelihood of Ranavirus disease outbreaks are still not well understood.
Like Ranavirus, the chytrid fungus Batrachochytridium dendrobatidis is a significant conservation threat to amphibians, and it is also primarily an aquatic pathogen (Kolby et al. 2015). Recent research found that B. dendrobatidis is shed from recent metamorphs onto vegetation perches, which can be used later by susceptible terrestrial adult frogs, and therefore this provides a potential indirect exposure route for infections from metamorphs to the adults (Kolby et al. 2015). Terrestrial indirect transmission of these aquatic pathogens may present more of a conservation risk than previously thought.
Comparing transmission efficiency and intensity for mode of transmission.
We hypothesized that direct contact would result in higher transmission efficiency and intensity than indirect contact. Surprisingly, both direct and indirect modes of transmission had high transmission efficiencies. Comparing modes of transmission, transmission intensity as measured by viral load was highest in the 12-hour, 24-hour, and 48-hour direct contact exposure periods, whereas the shorter direct contact exposure periods (1, 6, and 12hr) were comparable with indirect contact viral loads (Fig. 2). One caveat is that the semi-natural environment within the containers in the laboratory could result in slower breakdown of the virus than a natural forest floor due to the lack of wind, rain, UV exposure, temperature fluctuations, and other natural processes. Regardless, the similarly high transmission intensities between the direct and indirect contact experiments suggest that terrestrial transmission of Ranavirus may be of greater conservation concern than previously recognized.
Correlations between infective host viral load and transmission intensity.
Lastly, we did not find strong support for our fourth hypothesis that infectives with higher viral loads would cause proportionately higher viral loads in their susceptible partners (higher transmission intensity) independent of wait times or exposure periods. The indirect contact experiment had a borderline significant association between viral load and the paired infective animal viral load (P = 0.14), and so we included this variable in our linear regression since including it improved the model (Table 1). For the direct contact experiment, the paired infective animal viral load was not a significant covariate (Table 1).