Thermal acclimation has little effect on tadpole resistance to Batrachochytrium dendrobatidis
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Vol. 133: 207–216, 2019 DISEASES OF AQUATIC ORGANISMS
Published online March 28
https://doi.org/10.3354/dao03347 Dis Aquat Org
Thermal acclimation has little effect on tadpole
resistance to Batrachochytrium dendrobatidis
Karie A. Altman1, 2,*, Thomas R. Raffel1
1
Department of Biological Sciences, Oakland University, Rochester, MI 48309, USA
2
Present address: Department of Biological Sciences, University of Pittsburgh, Pittsburgh, PA 15260, USA
ABSTRACT: Given that climate change is predicted to alter patterns of temperature variability, it
is important to understand how shifting temperatures might influence species interactions, includ-
ing parasitism. Predicting thermal effects on species interactions is complicated, however, be-
cause the temperature-dependence of the interaction depends on the thermal responses of both
interacting organisms, which can also be influenced by thermal acclimation, a process by which
organisms adjust their physiologies in response to a temperature change. We tested for thermal
acclimation effects on Lithobates clamitans tadpole susceptibility to the fungus Batrachochytrium
dendrobatidis (Bd) by acclimating tadpoles to 1 of 3 temperatures, moving them to 1 of 5 perform-
ance temperatures at which we exposed them to Bd, and measuring Bd loads on tadpoles post-
exposure. We predicted that (1) tadpole Bd load would peak at a lower temperature than the tem-
perature for peak Bd growth in culture, and (2) tadpoles acclimated to intermediate temperatures
would have overall lower Bd loads across performance temperatures than cold- or warm-acclimated
tadpoles, similar to a previously published pattern describing tadpole resistance to trematode
metacercariae. Consistent with our first prediction, Bd load on tadpoles decreased with increasing
performance temperature. However, we found only weak support for our second prediction, as
acclimation temperature had little effect on tadpole Bd load. Our results contribute to a growing
body of work investigating thermal responses of hosts and parasites, which will aid in developing
methods to predict the temperature-dependence of disease.
KEY WORDS: Chytridiomycosis · Temperature · Green frog · Beneficial acclimation · Dormancy
Resale or republication not permitted without written consent of the publisher
1. INTRODUCTION due to nonlinear and thermal acclimation responses
to temperature (Rohr et al. 2013). Thermal acclima-
Temperature is one of the most important abiotic tion, which here refers to adaptive and plastic pheno-
factors influencing organism performance (any phys- typic adjustments made in response to temperature
iological trait of interest: e.g. metabolism, sprint changes, implies that organism performance at a
speed; Angilletta 2009, Huey & Kingsolver 2011). given temperature depends on which temperatures
Given the changing climate, a primary goal of wild- the organism experienced in the recent past
life managers is to predict organism performance (Angilletta 2009). Species interactions such as para-
under new temperature scenarios. However, several sitism compound this complexity, because both para-
factors complicate this, including organism responses sites and hosts might have nonlinear and thermal
to temperature variability. Temperature shifts on var- acclimation responses to temperature that could
ious timescales (e.g. day to day or month to month) drive disease outcomes in variable-temperature
make predicting organism performance difficult environments.
*Corresponding author: karie.altman@pitt.edu © Inter-Research 2019 · www.int-res.com208 Dis Aquat Org 133: 207–216, 2019 Thermal acclimation is an important adaptation for of acclimation effects for another component of tad- many organisms that live in variable-temperature pole resistance to trematode infection — their ability environments. However, whether and how organ- to prevent initial parasite encystment — that was con- isms acclimate can vary widely among populations sistent with the beneficial acclimation hypothesis but and species (Stillman 2003, Seebacher et al. 2012), not the dormancy hypothesis. These results indicate and several hypotheses have been proposed to ex- that host thermal responses can have distinct effects plain this variation. For the host species of interest in on different aspects of host resistance to infection. this study (green frogs Lithobates [Rana] clamitans), Building on the work of Altman et al. (2016), the goal 2 hypotheses were found to be potentially important of the present study was to determine whether and for driving the temperature-dependence of tadpole how thermal acclimation influences L. clamitans tad- resistance to encysted trematode parasites (Altman pole resistance to the amphibian chytrid fungus Batra- et al. 2016). (1) According to the ‘beneficial acclimation chochytrium dendrobatidis (Bd). Bd has been used in hypothesis’, an organism acclimated to a particular multiple studies of host thermal acclimation, the re- temperature will outperform unacclimated organisms sults of which have generally supported the beneficial at that temperature (Leroi et al. 1994). (2) According acclimation hypothesis (Raffel et al. 2013, 2015). How- to the ‘dormancy hypothesis’, an organism acclimated ever, these past studies have only used 2 acclimation to a particular temperature might adaptively decrease temperatures, making it impossible to detect potential its performance across temperatures, for example, to nonlinear effects of thermal acclimation. We therefore conserve energy during hibernation or aestivation used 3 acclimation temperatures in the current study (Geiser 2004, Storey & Storey 1990). to address whether host thermal acclimation can have These hypotheses are not mutually exclusive, and nonlinear effects on Bd susceptibility. We predicted the 2 mechanisms combined could lead to nonlinear that L. clamitans thermal acclimation responses might acclimation effects (Altman et al. 2016). For example, have nonlinear effects on tadpole resistance to Bd in- a species that undergoes a dormancy response when fection, similar to tadpole clearance of R. ondatrae acclimated to low temperatures might also experience metacercariae (i.e. intermediate-acclimated tadpoles a beneficial acclimation response when acclimated to would have the lowest average Bd loads overall rela- high temperatures. In this scenario, we would expect tive to cold- and warm-acclimated tadpoles). Alterna- both warm- and cold-acclimated organisms to have tively, thermal acclimation effects on Bd infection relatively poor performance at cold performance tem- might more closely resemble acclimation effects on peratures. This could lead to a nonlinear acclimation tadpole resistance to initial R. ondatrae encystment response in which organisms acclimated to intermedi- (i.e. beneficial acclimation). Importantly, peak Bd ate temperatures have the highest average perform- growth tends to occur at lower temperatures on am- ance relative to cold- or warm-acclimated organisms phibians (in vivo) than in culture (in vitro: optimum when all performance temperatures are considered temperature ~17−25°C; Piotrowski et al. 2004), pre- (note that warm-acclimated organisms undergoing sumably due to increased amphibian resistance at beneficial acclimation in this example could still have higher temperatures (Andre et al. 2008, Raffel et al. the highest peak performance at warm performance 2013, Sonn et al. 2017, but see Cohen et al. 2017). We temperatures relative to cold- and intermediate-accli- therefore also predicted that Bd load on L. clamitans mated organisms). This is exactly the pattern Altman tadpoles would peak at a temperature lower than that et al. (2016) found for the ability of L. clamitans tad- of optimum Bd growth in vitro. To test these predic- poles to clear encysted trematode metacercariae. tions, we acclimated L. clamitans tadpoles to 1 of 3 Tadpoles acclimated to intermediate temperatures temperatures, exposed each tadpole to Bd at 1 of 5 per- cleared a larger proportion of Ribeiroia ondatrae formance temperatures, and then quantified Bd infec- metacercariae than cold- or warm-acclimated tad- tion on each tadpole weekly for 4 wk post-exposure. poles at low and intermediate performance temp- eratures (tadpoles cleared similar proportions of metacercariae at high performance temperatures, re- 2. MATERIALS AND METHODS gardless of acclimation temperature; Altman et al. 2016). Such nonlinear acclimation effects are seldom 2.1. Frog collection and maintenance reported, perhaps in part because most experimental studies of thermal acclimation use only 2 acclimation Green frog tadpoles Lithobates clamitans were col- temperatures (e.g. Raffel et al. 2013, 2015). Interest- lected from ponds in Evart and Rochester, Michigan, ingly, Altman et al. (2016) observed a different pattern USA in August 2015 and October 2016, respectively.
Altman & Raffel: Tadpole thermal acclimation and parasite resistance 209 Both populations were maintained in the lab under 2.2. Temperature treatments similar conditions until the experiment began in November 2016. Similar proportions of tadpoles from We refer to the temperature at which an organism each collection location were assigned to each exper- was maintained before the temperature switch as its imental treatment, and tadpole population was in- ‘acclimation temperature’ and the temperature fol- cluded as a covariate in all analyses. Note that the lowing a temperature switch as its ‘performance tem- tadpole population term described both where and perature’. In a natural setting, both Bd and tadpoles when the tadpoles were collected (e.g. all individuals would undergo temperature fluctuations simultane- from the first population were collected at the same ously. However, the goal of this experiment was to time, and all individuals from the second population detect acclimation effects on tadpole resistance to were collected at another time). Bd; therefore, tadpoles were the only organisms to Before the experiment, tadpoles were housed in undergo a temperature shift during the experiment. groups within 14 l plastic boxes filled with aerated In this experimental design, any effects of acclima- artificial spring water (ASW; Cohen et al. 1980) that tion temperature can be attributed to the tadpole was changed weekly. During this time, tadpoles were acclimation treatments and not to Bd acclimation fed ground fish flakes and were maintained at room responses (Raffel et al. 2013, 2015). Onset HOBO temperature (~22°C) on a 12 h light:12 h dark cycle. temperature loggers (Bourne) recorded the tempera- To ensure that tadpoles were Bd negative at the tures within the incubators every 30 min throughout beginning of the experiment, tadpoles were main- the experiment and were used to verify that the incu- tained at 30°C for 7 d before the experiment to clear bators maintained our target treatment temperatures any existing Bd infection (Chatfield & Richards- throughout the experiment (see Fig. S1 in Supple- Zawacki 2011). A subset of the tadpoles (72 of 124 ment 1 at www.int-res.com/articles/suppl/d133p207_ total individuals) was also swabbed before the exper- supp.pdf). iment, and all were Bd negative. During the experiment, tadpoles were maintained individually in 0.5 l glass canning jars filled with 2.3. Bd maintenance 300 ml of ASW. Tadpoles were raised on a diet of ground fish flakes before the experiment; however, Twenty-five days before exposing frogs to Bd, cry- fish flakes quickly fouled the water at high treatment opreserved stock of Bd strain JEL 423 was removed temperatures. Therefore, tadpoles were fed thawed from −80°C storage and thawed according to Boyle et frozen spinach ad libitum during the experiment, and al. (2003). The thawed aliquot was divided between 2 soiled water was replaced weekly with ASW of the glass bottles, each containing 200 ml of 1% tryptone appropriate treatment temperature. When perform- broth, which were then incubated at 23°C for 18 d. ing water changes, nets were disinfected by soaking After this time, replicate 30 ml Bd cultures were pre- in 2% potassium permanganate for at least 5 min pared by transferring 3 ml of active Bd culture to after each use to prevent Bd transmission between 27 ml of sterile 1% tryptone broth in 50 ml centrifuge tadpoles (nets were rinsed in tap water after disin- tubes. The new Bd cultures were then divided among fection and before using them again). Incubators performance temperature treatments and placed into constructed from Styrofoam containers, heat tape, experimental incubators of the appropriate tempera- and adjustable thermostats were used to house tad- tures. Bd cultures were allowed to acclimate to their poles to ensure proper replication of temperature respective performance temperatures for 7 d before treatments (Raffel et al. 2013, Greenspan et al. 2016). being used to inoculate tadpoles. Tadpoles were checked daily for mortality through- out the experiment, and jars were rotated daily within the incubators to account for any within-incu- 2.4. Experimental design bator variation. The mass of each tadpole was meas- ured twice during the experiment: the day tadpoles L. clamitans tadpoles were randomly assigned to were exposed to Bd (or sham inoculum) and 28 d 1 of 3 acclimation temperatures (10, 16, or 22°C), after exposure. The tadpole mass predictor included where they were maintained for 14 d, which prior in statistical models was the average of these 2 val- results suggest is an adequate amount of time for ues. Results were unchanged when either initial or the tadpole immune system to acclimate to a new final mass was included in the models rather than temperature (e.g. all tadpoles performed similarly average mass. regardless of acclimation temperature by 14 d post-
210 Dis Aquat Org 133: 207–216, 2019
exposure to a parasite; Altman et al. 2016). After Table 1. Regression statistics from linear mixed-effects mod-
the acclimation period, tadpoles were moved to 1 of els (function lmer in R package lme4) describing the effects of
acclimation temperature (acc. temp.) and performance tem-
5 performance temperatures (13, 16, 19, 22, or perature (perf. temp.) on the average number of Batra-
25°C), where they were exposed to Bd zoospores chochytrium dendrobatidis (Bd) zoospores on frogs at each of
that had been acclimated to the performance tem- the 4 swab time points: 7, 14, 21, and 28 d post-exposure. Re-
perature (Fig. S2 in Supplement 1). This tempera- sults are from the simplest model at each time point (p < 0.1
for inclusion in the model). Block coefficients are reported as
ture range is consistent with pond temperatures
NA (not applicable) because block was a categorical variable
in which L. clamitans tadpoles live (Altman et with more than 2 treatment levels and therefore generated
al. 2016) and supports Bd growth both in vitro multiple-coefficients. Bold indicates significance (p < 0.05)
(Piotrowski et al. 2004, Voyles et al. 2017) and in
vivo (e.g. Raffel et al. 2015). Only 3 acclimation Swab time point Coefficient df F p
temperatures were used as opposed to 5 so that the and predictor
experiment could be conducted with a smaller sam-
7 d post-exposure
ple size, and acclimation temperatures were chosen
Mass −1.10 × 10−1 1, 64.9 2.12 0.150
to represent the lower, middle, and upper range of Tadpole population −2.41 × 10−2 1, 65.0 0.03 0.869
performance temperatures tested here. Note that a Block NA 2, 25.6 0.16 0.851
10°C performance treatment was intended for this Acc. temp. −1.50 × 10−4 1, 24.6 < 0.01 0.999
Acc. temp.2 4.56 × 10−3 1, 25.0 2.28 0.143
experiment as well; however, it was eliminated be-
Perf. temp. −9.51 × 10−3 1, 25.4 13.39 0.001
cause the Bd zoospore concentration in 10°C cul- Perf. temp.2 9.28 × 10−3 1, 24.9 8.01 0.009
tures was not high enough on the day of exposure Acc. temp. × perf. temp. −1.34 × 10−3 1, 50.2 0.31 0.581
to prepare inocula. Each acclimation by perform- Acc. temp.2 × perf. temp. −1.36 × 10−3 1, 52.0 3.81 0.056
ance temperature cross included 3 control tadpoles 14 d post-exposure
and at least 5 Bd-exposed tadpoles (Table S1). Tad- Mass 2.30 × 10−1 1, 63.9 8.30 0.005
Tadpole population −6.50 × 10−2 1, 67.2 0.19 0.667
poles were maintained for 6 h in their new per- Block NA 2, 25.6 1.36 0.276
formance incubators before being exposed to Bd Acc. temp. 4.04 × 10−4 1, 33.2 < 0.01 0.972
zoospores to ensure their water was at the new Perf. temp. −4.12 × 10−2 1, 25.1 11.32 0.002
performance temperature prior to adding the zoo- Perf. temp.2 1.21 × 10−2 1, 25.2 12.73 0.001
spore inoculum. 21 d post-exposure
Mass 9.74 × 10−2 1, 65.5 1.50 0.225
To accommodate time constraints, the experiment
Tadpole population 2.10 × 10−2 1, 62.1 0.02 0.886
was conducted in 3 temporal blocks. Temporal Block NA 2, 30.2 0.67 0.520
blocks had the same experimental timeline but Acc. temp. −1.34 × 10−2 1, 46.1 2.38 0.129
began on 3 consecutive days (e.g. Day 0 for Block Perf. temp. −1.66 × 10−2 1, 29.7 0.69 0.414
A occurred the day before Day 0 for Block B). Acc. temp. × perf. temp. 6.30 × 10−3 1, 45.3 8.80 0.005
Total sample sizes for each treatment (Table S1 in 28 d post-exposure
Mass 2.87 × 10−1 1, 64.3 7.20 0.009
Supplement 1) were spread out over the 3 blocks. Tadpole population 1.97 × 10−1 1, 67.9 0.95 0.334
Blocks contained similar numbers of tadpoles from Block NA 2, 26.4 0.23 0.799
each temperature treatment and tadpole collection Acc. temp. −1.97 × 10−2 1, 29.4 1.87 0.182
population. Tadpoles in each block were of similar Perf. temp. −2.77 × 10−2 1, 25.9 2.79 0.107
size (average mass ± SE: Block A, 0.82 ± 0.11 g;
Block B, 0.78 ± 0.11 g; and Block C, 1.20 ± 0.17 g;
p = 0.096 for the effect of block on average tad- 2.5. Bd exposure
pole mass using simple linear regression). Note
that there was a marginally nonsignificant effect of On the day of exposure, one 30 ml Bd culture from
block on tadpole developmental stage (p = 0.055), each performance temperature was passed through a
which reflects that tadpoles in Block C tended to 20 µm nylon filter (Spectrum Laboratories) to isolate
be more developed (mean Gosner stage ± SE: zoospores. The zoospore concentration of each ino-
31.17 ± 0.99) than those in Blocks A (28.28 ± 0.81) culum was determined by counting zoospores using
and B (28.63 ± 0.91). However, block was included a hemocytometer, and an appropriate volume of ster-
as a covariate in all analyses and was not a signifi- ile broth was added to dilute each culture to a con-
cant predictor of Bd load in any analysis (Table 1, centration of 106 zoospores ml−1. Three ml of inocu-
Tables S2−S5 in Supplement 1); therefore, this lum of the appropriate temperature was added to each
marginal difference was unlikely to dramatically tadpole’s water so that each tadpole was exposed to
influence our results. 3 × 106 zoospores. Three control tadpoles at eachAltman & Raffel: Tadpole thermal acclimation and parasite resistance 211
temperature were also exposed to 3 ml of sterile plement 2 at www.int-res.com/articles/suppl/d133
broth at the appropriate temperature. An extra water p207_supp.pdf).
change was performed 3 d after tadpoles were ex-
posed to Bd to prevent water fouling due to the pres-
ence of extra nutrients from the added broth. 2.7. Statistical analyses
Most of the tadpoles that died during the experi-
2.6. Bd swabbing and quantitative PCR ment died within 3 d of Bd or sham exposure (before
water had been changed following exposure). These
Each tadpole was swabbed 10 times across the animals were excluded from analyses, because swabs
mouthparts at 7, 14, 21, and 28 d post-exposure using were likely to pick up transient Bd DNA from zoo-
sterile fine-tip cotton swabs (Advantage Bundling spores that failed to establish in the tadpoles (Table S1
SP). Vinyl gloves were worn while swabbing, and be- details the number of tadpoles per treatment in-
tween each animal, gloves were disinfected and cluded in analyses). The only exception to this was 1
rinsed by successive immersion in 10% bleach, 1% tadpole in the 22°C acclimation × 19°C performance
AmQuel Plus (a dechlorinator; Kordon), and deion- temperature treatment, which died between 7 and
ized water (Raffel et al. 2013). Swabs were then 14 d post-exposure. This tadpole was therefore in-
frozen at −20°C until DNA extractions were per- cluded in the 7 d post-exposure analysis but was
formed. Bd DNA was extracted from swabs using excluded from subsequent analyses.
40 µl PrepMan Ultra extraction buffer (Applied Bio- R statistical software v.3.0.5 (R Core Team 2018)
systems), according to Kriger et al. (2006). was used to conduct all analyses. Infection data are
Quantitative PCR (qPCR) was run according to the usually non-normally distributed and often fit a neg-
protocol of Kriger et al. (2006), using a Bio-Rad CFX ative binomial distribution, and qPCR data are typi-
Connect system (Bio-Rad Laboratories). Samples were cally log-normally distributed. Past studies have used
run in singlicate to control costs (Kriger et al. 2006). zero-inflated negative binomial generalized linear
TaqMan® Exogenous Internal Positive Control Re- models (e.g. function glmmadmb in package glm-
agents (Applied Biosystems) were added to every mADMB or the recently updated function glmmTMB
reaction well to assess reaction inhibition (Kriger et in package glmmTMB) to analyze data similar to
al. 2006). This internal positive control (IPC) system those from this experiment (e.g. Raffel et al. 2013,
includes a standard concentration of artificial DNA 2015, McMahon et al. 2014); however, these models
with its own primers and a distinct fluorescent probe, would not converge when fit to our data. Regular
and the strength of the IPC reaction was used to negative binomial generalized linear models (func-
assess inhibition. All samples were initially run at tion glm.nb in package MASS) resulted in unreason-
1:10 dilution, and samples that were inhibited were ably low p-values for many terms in the model, indi-
re-run at 1:100 dilution. A sample was considered cating possible overfitting. We ultimately decided to
inhibited if its IPC cycle threshold (CT) value was analyze log-transformed zoospore numbers using
more than 4 cycles greater than the average CT value linear mixed-effects models assuming log-normally
of the negative control wells on the plate on which distributed data (function lmer in package lme4,
the sample was run. Furthermore, all sample curves Bates et al. 2015), which allowed us to include incu-
were visually inspected, and any samples with ab- bator as a random effect in all analyses. Performance
normally shaped curves were re-run at the original incubator was the random effect for all models except
dilution. those testing for main or quadratic effects of acclima-
Bd standards based on the internal transcribed tion temperature, where acclimation incubator was
spacer region of the Bd genome (ranging from 2.1 × instead specified as the random effect (Altman et al.
101 to 2.1 × 104 gene copies µl−1; Pisces Molecular) 2016).
were included on every qPCR plate. Standards were For the linear mixed-effects models, the response
not found to vary substantially among plates; there- variable in each analysis was Bd load, quantified as
fore, all plates from the experiment were combined the log number of zoospores detected on each swab.
to form a composite standard curve for use in esti- A 1 was added to zoospore numbers before log-trans-
mating the starting gene copy number of each formation to allow for the transformation of zeroes.
sample. Finally, the starting quantity units were con- Both acclimation and performance temperature data
verted from gene copies to zoospores using a conver- were centered prior to analyses by subtracting the
sion factor of 63.5 gene copies zoospore−1 (see Sup- mean temperature value from each data point to en-212 Dis Aquat Org 133: 207–216, 2019 sure the interpretability of both linear and quadratic effect. Furthermore, a binomial generalized linear effects in each model. Data from each of the 4 swab mixed-effects model (function glmer in package time points were analyzed separately. At each time lme4) was run to assess the temperature effects on point, both linear and quadratic effects of acclimation whether a tadpole tested positive for Bd at any point and performance temperature on Bd load were tested during the experiment (Table S4). This analysis used for. Interactions between acclimation and perform- χ2 rather than F-tests, and tadpoles with Bd loads
Altman & Raffel: Tadpole thermal acclimation and parasite resistance 213
1.8 a b
1.6
1.4
1.2
1.0
0.8
0.6
0.4
Log(Zoospores+1)
0.2
0.0
1.8 c d
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
10 13 16 19 22 25 28 10 13 16 19 22 25 28
Performance temperature (°C)
Fig. 1. Mean ± SE number of Batrachochytrium dendrobatidis (Bd) zoospores (log transformed) detected on tadpoles accli-
mated to 10°C (open circles), 16°C (gray circles), and 22°C (black circles) at each performance temperature (tadpoles were ex-
posed to Bd at the same performance temperatures, but points are offset for clarity). Each panel shows data from 1 of the 4
swabbing time points: (a) 7 d, (b) 14 d, (c) 21 d, and (d) 28 d post-exposure to Bd
significant interaction between acclimation and per- the average number of zoospores per animal detected
formance temperatures (F1, 45.3 = 8.80, p = 0.005). At on tadpoles ranged from 0.5 to 2835 zoospores, with
14 and 28 d post-exposure, larger tadpoles had an average of 307 zoospores (average log(zoospores)
higher Bd loads (14 d, F1, 63.9 = 8.30, p = 0.005; 28 d, = 2.49), indicating that our qPCR methods were able
F1, 64.3 = 7.20, p = 0.009). There was no interaction to detect Bd infection and that many of the tadpoles
between mass and acclimation or performance simply did not become infected.
temperature.
Results from the other analyses conducted were
generally consistent with those presented in the main 4. DISCUSSION
text (Tables S2−S6). In these analyses, performance
temperature was consistently a stronger predictor of Bd load on Lithobates clamitans tadpoles was
Bd load than acclimation temperature (Tables S2−S4), consistently highest at the lowest performance tem-
and tadpole mass was positively correlated with Bd perature (13°C) relative to the other performance
load when mass significantly contributed to the mod- temperatures tested and declined as performance
els (Tables S2 & S3). Results for Bd load averaged temperatures increased (Fig. 1). This supports our
through each time point (7−14, 7−21, and 7−28 d post- prediction that Bd growth on tadpoles would peak at a
exposure) are shown in Fig. S3 in Supplement 1. lower temperature than that at which it peaks in cul-
The results presented in Fig. 1 and Fig. S3 describe ture (17−25°C; Piotrowski et al. 2004). Note, however,
the average log-transformed number of zoospores on that we were unable to detect a true peak in Bd load
all Bd-exposed tadpoles, regardless of whether they in this study, because our highest observed Bd loads
tested positive for Bd. Because many tadpoles never occurred at the lowest temperature we tested. Our re-
became infected, the average Bd load is quite low. sults are consistent with those of other studies that
Among tadpoles that tested positive for Bd, however, have demonstrated a negative relationship between214 Dis Aquat Org 133: 207–216, 2019 temperature and Bd growth on amphibian hosts, acclimation studies should cover a broader range of which likely arises from a combination of enhanced temperatures, including those near organisms’ upper host immunity and reduced Bd growth (independent and lower thermal ranges, so that these effects can of host immunity) at high temperatures (Piotrowski et be detected if they exist. Note, however, that it is al. 2004, Rollins-Smith et al. 2011). For example, Row- unlikely that tadpole thermal acclimation would in- ley & Alford (2013) found that the probability of natu- fluence tadpole susceptibility to Bd at temperatures ral Bd infection in 3 amphibian host species declined above those tested in this study, because Bd load was as they spent more time at temperatures above 25°C. already very low on tadpoles at 25°C (Fig. 1), and Bd Furthermore, optimal Bd growth on several species of growth in vitro is typically low at temperatures above amphibians, or highest Bd-induced host mortality, oc- 25°C (Piotrowski et al. 2004, Voyles et al. 2017). curred at a lower temperature than that at which Bd At 7 d post-exposure in the present study, there growth peaks in vitro, likely because host resistance is was a marginally nonsignificant interaction between enhanced at higher temperatures (Andre et al. 2008, performance temperature and the quadratic effect of Raffel et al. 2013, Cohen et al. 2017, Sonn et al. 2017). acclimation temperature (Table 1). When exposed to However, a recent study showed that a cold-adapted Bd at 13°C, a trend showed that tadpoles acclimated amphibian species was instead more susceptible to to 16°C had lower Bd loads than tadpoles acclimated Bd infection at high temperatures, due to a thermal to either 10 or 22°C; however, tadpoles had similar mismatch between host and Bd thermal performance (low) Bd loads at higher performance temperatures curves (Cohen et al. 2017). These studies highlight (Fig. 1). This might indicate a beneficial acclimation the need to consider the thermal responses of both response by tadpoles acclimated to 22°C and a dor- host and parasite when predicting the temperature- mancy response by 13°C acclimated tadpoles. How- dependence of infection. ever, at 14 d post-exposure, tadpoles had similar Bd Interestingly, we saw very little evidence for an loads at 13°C regardless of acclimation temperature; effect of thermal acclimation of L. clamitans tadpoles therefore, any possible advantage of being accli- on their susceptibility to Bd. Performance tempera- mated to intermediate temperatures might only be ture was a much stronger predictor of Bd load on tad- temporary. Altman et al. (2016) found stronger evi- poles than acclimation temperature, no matter how dence for a nonlinear pattern in the ability of green the data were analyzed (Table 1, Tables S2−S5), and frog tadpoles to clear Ribeiroia ondatrae metacercar- even so, tadpoles only had appreciable Bd loads at iae, in which 7 d after parasite exposure, tadpoles the 13°C performance temperature (Fig. 1). In addi- acclimated to intermediate temperatures had cleared tion, there was no effect of performance temperature a larger proportion of metacercariae than either cold- on Bd load at 21 and 28 d post-exposure (Table 1). or warm-acclimated tadpoles (Altman et al. 2016). The effects of Bd are generally less adverse in tad- This pattern was especially evident at lower perform- poles than in adult amphibians (Berger et al. 1998), ance temperatures (13−23°C); tadpoles cleared simi- and tadpole species vary in their susceptibility to Bd lar proportions of metacercariae at the 2 highest per- (Blaustein et al. 2005). Therefore, low risk of severe formance temperatures (25 and 28°C), regardless of disease or low overall tadpole susceptibility to Bd acclimation temperature (Altman et al. 2016). Al- could have contributed to the lack of thermal accli- though the present study revealed only a weak pat- mation effects observed in our study. However, it is tern suggesting that acclimation to intermediate tem- important to note that our study was limited by the peratures might enhance tadpole resistance to initial fact that the lowest performance temperature that we Bd exposure at low temperatures, its similarity to the were able to test was 13°C. Based on our finding that results of Altman et al. (2016) is intriguing. Future tadpoles had high Bd loads at 13°C and the fact that studies that directly address potential parallels in Bd is capable of sustaining growth at temperatures as host thermal acclimation responses to multiple para- low as 2°C (Voyles et al. 2017), it is possible and even sites would therefore be valuable. likely that tadpoles would also exhibit high Bd loads Acclimation temperature had no effect on tadpole at temperatures lower than those tested in this exper- Bd load after 7 d post-exposure in the current study iment (e.g. 7−10°C). Furthermore, thermal acclima- (Fig. 1c,d). We did detect a significant interaction be- tion effects are sometimes evident only at very high tween acclimation and performance temperatures or low temperatures (e.g. Kaufmann & Bennett 1989). when the data from 21 d post-exposure were analyzed Therefore, tadpole acclimation temperature might in isolation, with tadpoles acclimated to 10°C exhibit- influence host Bd susceptibility at low temperatures ing higher Bd loads at 13°C (performance tempera- that were untested in this study. Future thermal ture) than tadpoles acclimated to 16 or 22°C (Fig. 1c).
Altman & Raffel: Tadpole thermal acclimation and parasite resistance 215
Estimates of the amount of time amphibians require In conclusion, our study demonstrates the impor-
for immune system acclimation range from days tance of accounting for thermal effects on both para-
(Maniero & Carey 1997, Altman et al. 2016, Greenspan site infectivity and host resistance when predicting
et al. 2017) to weeks (Bly & Clem 1991), and different the temperature-dependence of disease, as other
aspects of acclimation might require different amounts authors have previously emphasized (Raffel et al.
of time. It is therefore possible that tadpoles were not 2013, Sonn et al. 2017). Unlike previous studies that
fully acclimated to their performance temperatures by demonstrated beneficial acclimation effects in am-
21 d post-exposure. However, given that Bd levels of- phibians exposed to Bd following a temperature shift
ten fluctuate through time (e.g. Briggs et al. 2010, Ger- (Raffel et al. 2013, 2015), we found little evidence for
vasi et al. 2013, Catenazzi et al. 2017, Daversa et al. an effect of thermal acclimation on L. clamitans tad-
2018) and that no effect of acclimation temperature on pole susceptibility to Bd. However, there was an
Bd load was found at any other time point after 7 d interesting similarity between the weak effect of L.
post-exposure, it seems possible that the apparent ac- clamitans thermal acclimation on tadpole resistance
climation effect at 21 d post-exposure was caused by to Bd infection and previously reported effects on R.
random fluctuations in Bd load. ondatrae infection (Altman et al. 2016), which might
At both 14 and 28 d post-exposure in our study, warrant further study. It is unknown whether other
larger tadpoles had higher Bd loads (Table 1). We ectotherms might show similar thermal responses
chose to end our study at 28 d post-exposure because when challenged with various pathogens, but if gen-
we were focused on thermal acclimation effects, which eralizable, these results could help lead to better pre-
are typically short-term. In a natural setting, larger, dictions of the temperature-dependence of wildlife
older tadpoles (which have presumably been continu- and vector-borne diseases.
ously exposed to Bd throughout their larval period)
had increased probabilities of Bd infection (Sapsford
Acknowledgements. This work was conducted with Oakland
et al. 2018). However, it is unknown whether a similar
University Institutional Animal Care and Use Committee
pattern might be evident in a single-exposure labora- (IACUC) approval (IACUC no. 15101). We thank the anony-
tory study. Future studies tracking long-term (> 28 d) mous reviewers whose input greatly improved the manu-
Bd infection in tadpoles could provide insight into script. Thank you to the Zielinski family for allowing us to
whether the mass effects on tadpole Bd load that we collect tadpoles on their property and to Joyce Longcore for
providing Bd cultures. Finally, thanks to S. Brady, A. Foster,
found were due to random fluctuations in Bd load or if K. Julius, J. McBride, R. McWhinnie, M. Ostrowski, K. Rose,
they reflect a persistent pattern. H. Russell, J. Sckrabulis, R. Stepanian, J. Tituskin, and S.
Many outstanding questions remain regarding the Zielinski for assisting with animal maintenance, data collec-
effects of thermal acclimation on parasitism. Although tion, and qPCR. This work was supported by an NSF
CAREER award to T.R.R. (IOS-1651888). K.A.A. was sup-
we know that the amphibian immune response is
ported by a King-Chávez-Parks Future Faculty Fellowship.
temperature dependent (Rollins-Smith et al. 2011),
relatively little is known about how thermal acclima-
tion influences amphibian immunity (but see Raffel et LITERATURE CITED
al. 2006, Greenspan et al. 2017). Furthermore, thermal
Altman KA, Paull SH, Johnson PTJ, Golembieski MN,
acclimation effects on parasite infectivity remain rela- Stephens JP, LaFonte BE, Raffel TR (2016) Host and par-
tively unexplored. There is evidence that trematode asite thermal acclimation responses depend on the stage
cercaria infectivity is influenced by acclimation tem- of infection. J Anim Ecol 85:1014−1024
perature (Altman et al. 2016), but it is currently un- Andre SE, Parker J, Briggs CJ (2008) Effect of temperature
on host response to Batrachochytrium dendrobatidis
known whether thermal acclimation might influence
infection in the mountain yellow-legged frog (Rana mus-
the infectivity of Bd. Based on its small size, Bd is ex- cosa). J Wildl Dis 44:716−720
pected to acclimate faster than its relatively larger Angilletta MJ Jr (2009) Thermal adaptation: a theoretical
hosts (Rohr & Raffel 2010), and it is possible that Bd and empirical synthesis. Oxford University Press, New
York, NY
acclimation occurs so quickly following a temperature
Bates D, Mächler M, Bolker B, Walker S (2015) Fitting linear
shift that any effect of thermal acclimation on infectiv- mixed-effects models using lme4. J Stat Soft 67:1–48
ity would be negligible. Further study is needed to in- Berger L, Speare R, Daszak P, Green DE and others (1998)
vestigate the relative rates at which hosts and para- Chytridiomycosis causes amphibian mortality associated
sites acclimate to new temperatures to determine if with population declines in the rain forests of Australia
and Central America. Proc Natl Acad Sci USA 95:
parasites have an advantage over hosts in variable- 9031−9036
temperature environments, as previously postulated Blaustein AR, Romansic JM, Scheessele EA, Han BA, Pessier
(Rohr & Raffel 2010). AP, Longcore JE (2005) Interspecific variation in suscep-216 Dis Aquat Org 133: 207–216, 2019 tibility of frog tadpoles to the pathogenic fungus Batra- the beneficial acclimation assumption. Proc Natl Acad chochytrium dendrobatidis. Conserv Biol 19:1460−1468 Sci USA 91:1917−1921 Bly JE, Clem LW (1991) Temperature-mediated processes in Maniero GD, Carey C (1997) Changes in selected aspects of teleost immunity: in vitro immunosuppression induced immune function in the leopard frog, Rana pipiens, asso- by in vivo low temperature in channel catfish. Vet ciated with exposure to cold. J Comp Physiol B 167: Immunol Immunopathol 28:365−377 256−263 Boyle DG, Hyatt AD, Daszàk P, Berger L and others (2003) McMahon TA, Sears BF, Venesky MD, Bessler SM and oth- Cryo-archiving of Batrachochytrium dendrobatidis and ers (2014) Amphibians acquire resistance to live and other chytridiomycetes. Dis Aquat Org 56:59–64 dead fungus overcoming fungal immunosuppression. Briggs CJ, Knapp RA, Vredenburg VT (2010) Enzootic and Nature 511:224−227 epizootic dynamics of the chytrid fungal pathogen of Piotrowski JS, Annis SL, Longcore JE (2004) Physiology of amphibians. Proc Natl Acad Sci USA 107:9695−9700 Batrachochytrium dendrobatidis, a chytrid pathogen of Catenazzi A, Swei A, Finkle J, Foreyt E, Wyman L, Vreden- amphibians. Mycologia 96:9−15 burg VT (2017) Epizootic to enzootic transition of a fun- R Core Team (2018) R: a language and environment for sta- gal disease in tropical Andean frogs: Are surviving spe- tistical computing, version 3.0.5. R Foundation for Statis- cies still susceptible? PLOS ONE 12:e0186478 tical Computing, Vienna Chatfield MWH, Richards-Zawacki CL (2011) Elevated tem- Raffel T, Rohr J, Kiesecker J, Hudson P (2006) Negative perature as a treatment for Batrachochytrium dendroba- effects of changing temperature on amphibian immunity tidis infection in captive frogs. Dis Aquat Org 94:235−238 under field conditions. Funct Ecol 20:819−828 Cohen LM, Neimark H, Eveland LK (1980) Schistosoma Raffel TR, Romansic JM, Halstead NT, McMahon TA, mansoni: response of cercariae to a thermal gradient. Venesky MD, Rohr JR (2013) Disease and thermal accli- J Parasitol 66:362−364 mation in a more variable and unpredictable climate. Nat Cohen JM, Venesky MD, Sauer EL, Civitello DJ, McMahon Clim Chang 3:146−151 TA, Roznik EA, Rohr JR (2017) The thermal mismatch Raffel TR, Halstead NT, McMahon TA, Davis AK, Rohr JR hypothesis explains host susceptibility to an emerging (2015) Temperature variability and moisture synergisti- infectious disease. Ecol Lett 20:184−193 cally interact to exacerbate an epizootic disease. Proc R Daversa DR, Monsalve-Carcaño C, Carrascal LM, Bosch J Soc B 282:20142039 (2018) Seasonal migrations, body temperature fluctua- Rohr JR, Raffel TR (2010) Linking global climate and tem- tions, and infection dynamics in adult amphibians. PeerJ perature variability to widespread amphibian declines 6:e4698 putatively caused by disease. Proc Natl Acad Sci USA Geiser F (2004) Metabolic rate and body temperature reduc- 107:8269−8274 tion during hibernation and daily torpor. Annu Rev Phys- Rohr JR, Raffel TR, Blaustein AR, Johnson PTJ, Paull SH, iol 66:239−274 Young S (2013) Using physiology to understand climate- Gervasi S, Gondhalekar C, Olson DH, Blaustein AR (2013) driven changes in disease and their implications for con- Host identity matters in the amphibian−Batrachochy- servation. Conserv Physiol 1:cot022 trium dendrobatidis system: fine-scale patterns of varia- Rollins-Smith LA, Ramsey JP, Pask JD, Reinert LK, Wood- tion in responses to a multi-host pathogen. PLOS ONE 8: hams DC (2011) Amphibian immune defenses against e54490 chytridiomycosis: impacts of changing environments. Gosner KL (1960) A simplified table for staging anuran Integr Comp Biol 51:552−562 embryos and larvae with notes on identification. Her- Rowley JJ, Alford RA (2013) Hot bodies protect amphibians petologica 16:183−190 against chytrid infection in nature. Sci Rep 3:1515 Greenspan SE, Morris W, Warburton R, Edwards L and oth- Sapsford SJ, Alford RA, Schwarzkopf L (2018) Disentangling ers (2016) Low cost fluctuating temperature chamber for causes of seasonal infection prevalence patterns: tropical experimental ecology. Methods Ecol Evol 7:1567−1574 tadpoles and chytridiomycosis as a model system. Dis Greenspan SE, Bower DS, Webb RJ, Berger L, Rudd D, Aquat Org 130:83−93 Schwarzkopf L, Alford RA (2017) White blood cell pro- Seebacher F, Holmes S, Roosen NJ, Nouvian M, Wilson RS, files in amphibians help to explain disease susceptibility Ward AJ (2012) Capacity for thermal acclimation differs following temperature shifts. Dev Comp Immunol 77: between populations and phylogenetic lineages within a 280−286 species. Funct Ecol 26:1418−1428 Huey RB, Kingsolver JG (2011) Variation in universal tem- Sonn JM, Berman S, Richards-Zawacki CL (2017) The influ- perature dependence of biological rates. Proc Natl Acad ence of temperature on chytridiomycosis in vivo. Eco- Sci USA 108:10377−10378 Health 14:762−770 Kaufmann JS, Bennett AF (1989) The effect of temperature Stillman JH (2003) Acclimation capacity underlies suscepti- and thermal acclimation on locomotor performance in bility to climate change. Science 301:65 Xantusia vigilis, the desert night lizard. Physiol Zool 62: Storey KB, Storey JM (1990) Metabolic rate depression and 1047−1058 biochemical adaptation in anaerobiosis, hibernation and Kriger KM, Hero JM, Ashton KJ (2006) Cost efficiency in the estivation. Q Rev Biol 65:145−174 detection of chytridiomycosis using PCR assay. Dis Aquat Voyles J, Johnson LR, Rohr J, Kelly R and others (2017) Org 71:149−154 Diversity in growth patterns among strains of the lethal Leroi AM, Bennett AF, Lenski RE (1994) Temperature accli- fungal pathogen Batrachochytrium dendrobatidis across mation and competitive fitness: an experimental test of extended thermal optima. Oecologia 184:363−373 Editorial responsibility: Louise Rollins-Smith, Submitted: April 4, 2018; Accepted: January 9, 2019 Nashville, Tennessee, USA Proofs received from author(s): March 22, 2019
You can also read
