Electronic Supplementary Materials: Correlates of virulence in a frog-killing fungal pathogen: Evidence from a California amphibian decline
Contents
1. Supplementary Methods
2. Supplementary Figures and Tables
3. Index of Raw Data Files
4. Literature Cited
1. Supplementary Methods
1.1 Host-pathogen dynamics in the field
We evaluated R. cascadae population size at Section Line Lake, in the Klamath Mountains, and at Carter Meadow, in the southern Cascade Mountains, between 2008 and 2013. Data on Bd prevalence was collected between 2008 and 2011 and data on Bd infection intensity was collected between 2009 and 2011. Rana cascadae has experienced dramatic declines in the southern Cascade Mountains in the last few decades, and Carter Meadow is home to one of the remnant populations in this mountain range (Fellers et al 2008, Pope et al 2014). In contrast, R. cascadae remains widespread in the Klamath Mountains (Piovia-Scott et al 2011). We began collecting Bd samples from Carter Meadow in 2008, so this paper contains the earliest Bd data from that site that we are aware of; five Bd samples collected from Section Line Lake in 2006 predate the results presented here, all were Bd-negative (K. Pope, unpublished data).
We measured frog abundance using visual encounter surveys (VESs: Crump and Scott 1994). Each site was surveyed multiple times each summer (with the exception of Section Line Lake in 2012, when only one survey was conducted); we used the highest counts of adults and juveniles from a single survey as our abundance estimates for each year of the study. The count data presented in the results were collected under conditions conducive to observing frogs (mostly clear skies, warm air temperatures, and available patches of sunlight for basking). Animals between 27 and 45 mm in snout-urostyle length were considered juveniles; larger animals were considered adults, and smaller animals were considered young-of-year. These size-based distinctions between life stages are supported by detailed surveys of R. cascadae populations throughout northern California (Pope et al., unpublished data). We do not show counts for young-of-year frogs, as metamorphosis was not always complete when surveys were conducted.
During the study period frogs were marked with passive integrated transponder (PIT) tags and visible implant elastomer (VIE) as part of an ongoing population study. PIT tags uniquely identify individual animals, while VIE marks did not uniquely identify individuals and were routinely used on animals too small for PIT tags. These marks allowed us to account for the non-independence of multiple samples collected from the same individuals in our analyses of Bd prevalence and load. For the prevalence analysis, we removed observations of recaptured animals that were marked only with VIE, as we were not able to ascertain whether these animals had been swabbed previously. This gave us 317 observations from 239 unique individuals to analyze. The dataset for Bd load (which featured only Bd-positive animals) included very few observations from recaptured individuals (13 out of 150), so we removed these observations prior to analysis, allowing us to use a simpler model structure that omitted the random effect for individual animal. Analyses were conducted in R (R Development Core Team 2012).
1.2 Bd quantification using real-time PCR
We evaluated the amount of Bd on each swab in terms of zoospore equivalents (ZE) using real-time, quantitative PCR (Boyle et al 2004, Hyatt et al 2007, Retallick et al 2006). To remove DNA from swabs, each swab was placed in a vial containing 0.5ml of 1% Tris EDTA (TE) and shaken overnight. Swabs were then removed from the buffer and discarded, and the samples were centrifuged at 13,000 rpm for 10 min. The supernatant was discarded and 40ml of PrepMan Ultra (Applied Biosystems, Foster City, CA) was added to the pellets. The samples were vortexed, heated at 100°C for 10min, then centrifuged at 13,000 rpm for 2 minutes. The upper 20ml of each sample was removed and diluted in 180ml molecular grade water. Diluted DNA samples were analyzed on a Step-One Plus Real-time PCR machine (Applied Biosystems); we used 2.5 µl of dilute DNA with a total reaction volume of 12.5 µl. Samples were run singly (Kriger et al 2006) and Bd standards were run on each plate. The DNA quantity found by qPCR was multiplied by 160 to account for dilutions that occurred during processing, producing an estimate of the number of Bd zoospore equivalents (ZE) in each sample.
1.3 Laboratory exposure experiment
Amphibian husbandry Pre-metamorphosis
Rana cascadae eggs were collected from five locations May - July 2011 and transported to the University of California, Davis for hatching and rearing. Eggs and larvae from the same collection sites were housed communally in a standardized environment free of Bd. Field-collected eggs were placed in aerated glass aquaria filled with spring water and kept in a growth chamber on a 12 hr/12 hr light/dark cycle until hatching. After hatching, larvae from each population were transferred to separate 300 gallon plastic cattle tanks (Rotonics Manufacturing, Inc.) in a greenhouse at the UC Davis Vegetable Crops Facility and reared through metamorphosis. We used a single tank for each source population. While this design may confound tank effects with population effects, we felt it was more important to minimize the amount of space used and thus avoid introducing additional environmental variation. Tanks were prepared for larval frogs using the following procedure: approximately three weeks prior to introducing larval frogs, we filled each tank with 200 gallons of reverse osmosis water to which we added a handful of organic straw, enough water conditioner (RO Right, Kent Marine) to bring the hardness to approximately 8.5 dGH (100 ppm), 5-15 g of ground rabbit food (as a source of macronutrients), and an inoculum of Daphnia magna (Carolina Biological Supply Company). Tanks were aerated to maintain dissolved oxygen concentrations and stabilize pH. To supplement existing food sources (such as periphyton), larval frogs were fed fish flakes and boiled lettuce. When larvae began to form legs, wooden floats were added to each tank to provide a platform for animals in the process of metamorphosis. Metamorphosis began August - September 2011 -- because the collection site in the southern Cascades receives less winter snowfall than those in the Klamath Mountains, hatching and metamorphosis occurred approximately one month earlier for frogs from this collection site.
Amphibian husbandry post-metamorphosis
After metamorphosis individual frogs were placed in six-quart plastic shoe boxes (Sterilite Corportation) and housed at the UC Davis vivarium facility in Tupper Hall in a temperature-controlled room held at 200C with a 12 hr-12 hr light-dark cycle. Each box contained 150 mL of distilled de-ionized water to which we added water conditioner (RO Right, Kent Marine) at a rate of 1 tsp in each 5 gallons of water. A small plastic shelter was placed in each container. Containers were rinsed and water was changed weekly. Frogs were fed wingless fruit flies dusted with vitamins or calcium (Rep-Cal Research Labs), and occasionally small crickets, three times a week. Moribund frogs were euthanized using Ethyl 3-aminobenzoate methanesulfonic acid (MS-222).
Bd exposure
Bd was isolated from the skin of two juvenile R. cascadae, one collected at Carter Meadow on July 8, 2011 and the other collected at Section Line Lake on July 18, 2011 (see below for details on Bd isolation). Because Bd virulence can be attenuated after extensive passage in culture (Langhammer et al 2013) neither isolate was transferred more than five times prior to being used in frog exposures. To generate Bd zoospores for inoculation, we used both TGhL broth and agar plate cultures. Zoospores were harvested from broth using a SpectraMesh nylon 20 µm filter (Spectrum Laboratories, Inc., Rancho Dominguez, CA) and the filtrate collected in a sterile vial. Zoospores were harvested from plates by flooding each plate with 5 mL TGhL broth for 5 minutes, then aspirating the liquid. The filtrate and the aspirate were pooled and the zoospore density of the resulting inoculum was quantified using a hemocytometer. This inoculum was then diluted with TGhL to 5.5 x 104 zoospores per mL. Two mLs of the diluted inoculum were used in exposure treatments, resulting in a dose of 1.1 x 105 zoospores per exposed animal.
Bd treatments were applied October – November 2011, approximately two months after metamorphosis. Thus, we applied Bd treatments to frogs from the southern Cascades approximately one month before frogs from the Klamath Mountains (Table S3). Our experiment fully crossed Bd isolates with frog collection sites. A total of 171 frogs (19 - 40 frogs from each of the five collection sites; Table S3) were randomly assigned to treatments groups such that each treatment was applied to the same number of animals from each site; when the sets of frogs from a collection site could not be divided into equal groups, the remainder were assigned treatments at random. The three treatments were: 1) exposure to Bd isolated from Section Line Lake (hereafter “Section Line Bd”), 2) exposure to Bd isolated from Carter Meadow (hereafter “Carter Bd”), 3) exposure to sterile broth containing no Bd zoospores (hereafter “control”). Treatments were administered by placing each frog into a 100 X 20 mm petri dish with 18 mL of distilled, deionized water with electrolytes and 2mL of inoculum containing 1.1 x 105 zoospores of the appropriate Bd isolate for 15 hours (control animals were treated in a similar manner, but the inoculum contained no zoospores). Frogs remained halfway submerged while they were kept in the dish. The snout-urostyle length of each frog was measured two weeks after the experiment was initiated.
Following Bd exposure, frogs were monitored daily. Animals that did not appear to be alert or responsive were tested for a response to gentle, tactile stimulus and a righting reflex. Frogs that displayed a lethargic response to these stimuli were considered moribund and euthanized using MS-222. Using the same methods described for the field study, we evaluated the amount of Bd on each live frog in the two Bd exposure treatments 1, 2, 3, 4, 6, 10, and 14 weeks after treatment. At each sampling date, we also evaluated Bd on 10 - 42 control animals to confirm that they were not infected. In addition, 66 of the experimental animals were tested for Bd prior to treatment application to confirm that there were no animals infected before the experiment began. None of the control or pre-treatment samples tested positive for Bd.
Statisical analyses
Analyses were conducted in R (R Development Core Team 2012). Survival analyses were conducted using the ‘survival’ package (Therneau and Lumley 2011), and linear mixed models were analyzed using the ‘lme4’ package (Bates et al 2014). Where applicable, we used post-hoc Tukey test to evaluate differences between treatment groups using functions in the ‘multcomp’ package (Hothorn et al 2008).
Frog size may influence Bd infection intensity and mortality rate, and frogs from different collection sites differed in length (ANOVA of snout-urostyle length, Frog collection site effect: F4,161 = 42.41; P<0.0001). Because there was a strong association between length and collection site, we did not include length as a covariate in models that included frog collection site.
1.4 Bd isolation and culture
Frogs were euthanized using a 0.2% MS-222, adjusted to a pH of 7. Areas of the skin showing signs of infection were removed and cut into 1mm by 1mm pieces and a sterile needle was used to pull the skin through two different tryptone-gelatin hydrolysate-lactose (TGhL) agar plates. The first swipe (in plate one) was meant to reduce any bacteria present. After the second swipe, the skin was left in the agar of the second plate to allow any zoosporangium present to grow. Plates were incubated at 23 ⁰C and examined periodically for the presence of zoospores or zoosporangium. Once either was visible the resulting colony was transferred to new TGhL agar. The resulting isolate was then transferred to TGhL broth, stored at 4 ⁰C and transferred periodically to maintain healthy growth. All three northern California isolates described in this paper (from Section Line Lake, Carter Meadow, and Finley Lake) have been contributed to the chytrid culture collection maintained in the lab of Joyce Longcore at the University of Maine.
For in vitro growth assays, zoospores were collected by filtering Bd cultures through Whatman 4 filters (Kent, UK) that were UV sterilized. Zoospores were counted using a hemocytometer then diluted to 100 zoospores /µL with fresh TGhL broth; 100 µL of this zoospore solution was used to inoculate each replicate culture.
1.5 Bd genome sequencing and genomic analyses
Sequence alignment and SNP-calling were performed as in Rosenblum et al. (2013). Briefly, we used Seqyclean v. 1.8.10 (Zhbannikov et al 2013) to clean reads; remove PCR duplicates, contaminants, and adaptors; and trim sequences by quality scores. Reads were aligned to the reference genome of JEL423 (Broad Institute v. 17-Jan-2007) using Bowtie 2 v. 2.1.0 (Langmead et al 2009), and we used best practice protocol for variant calling in GATK v. 1.4 (McKenna et al 2010). We marked duplicate reads with Picard, and we realigned reads containing indels with GATK walkers RealignerTargetCreator and IndelRealigner. We made final variant calls, and filtered false positives, using GATK UnifiedGenotyper and VariantFiltration walkers with the same filter parameter value as in Rosenblum et al. (2013). Supercontigs (i.e, chromosomal segments) were defined as in Rosenblum et al. (2013).