The putatively positive association between host genetic diversity and the ability to defend against pathogens has long attracted the attention of evolutionary biologists. Chytridiomycosis, a disease caused by the chytrid fungus Batrachochytrium dendrobatidis (Bd), has emerged in recent decades as a cause of dramatic declines and extinctions across the amphibian clade. Bd susceptibility can vary widely across populations of the same species, but the relationship between standing genetic diversity and susceptibility has remained notably underexplored so far. Here, we focus on a putatively Bd-naive system of two mainland and two island populations of the common toad (Bufo bufo) at the edge of the species’ range, and use controlled infection experiments and dd-RAD sequencing of >10,000 SNPs across 95 individuals to characterise the role of host population identity, genetic variation and individual body mass in mediating host response to the pathogen. We found strong genetic differentiation between populations and marked variation in their susceptibility to Bd. This variation was not, however, governed by isolation-mediated genetic erosion, and individual heterozygosity was even found to be negatively correlated with survival. Individual survival during infection experiments was strongly positively related to body mass, which itself was unrelated to population of origin or heterozygosity. Our findings underscore the general importance of context-dependency when assessing the role of host genetic variation for the ability of defence against pathogens.

Experimental infections

Experimental procedures took place in a temperature controlled room at 18°C with a 12:12 h light:dark cycle. Ninety days after the start of metamorphosis, toadlets were weighed to the nearest 0.001 g, and transferred to individual housing in 0.7 L polypropylene boxes (Really Useful Products LTD, UK) lined with a moist paper towel and containing a cover object. Following a two-week period of acclimatisation, 202 toadlets were randomly allocated into one of two experimental treatments: (I) exposure to an active Bd culture, or (II) a control group exposed to culture media alone. Treatments broadly followed Garner et al. (2009, 2011). Each toadlet was placed into an individual Petri dish filled with 30 ml of aged tap water and 450 μl of culture media for four hours. In order to ensure a level of exposure sufficient to cause mortality, nine such treatments were carried out across 21 days. All individuals in the exposed group received the same dosage of the Bd isolate UK CORN'12 3.1, part of the hypervirulent global panzootic lineage BdGPL (Farrer et al., 2011). Dosages were calculated by counting live spores using a haemocytometer and varied between 40,500 and 382,500 zoospores depending on session. The experiment proceeded for 50 days after the first exposure (i.e., a 21-day period of inoculations followed by a further 29 days of daily monitoring of mortality). If a toadlet reached a humane end point (i.e., was unresponsive or was unable to support its own weight), it was euthanised by licenced personnel in accordance with the Animal (Scientific Procedures) Act 1986 using the non-schedule 1 method of immersion in buffered tricaine methanesulfonate (MS222) followed by fixation in 70% ethanol. All animals surviving to the end of the experiment were euthanised and fixed as described above.

SNP genotyping

DNA for genotyping was extracted from hind leg muscle of toadlets using a Qiagen DNEasy extraction kit following the manufacturer's protocol (Qiagen, UK). DNA concentration was assessed by fluorometry using a Qubit 3.0 (Thermo Fisher Scientific, MA, USA), and standardised to 20 ng/uL. The set of 101 exposed individuals (excluding the controls) was reduced to 95 by randomly removing six individuals from the best-represented population (MAK). The resulting panel of samples constituted 15 individuals from CRO, 31 from MAK, 27 from MAT, and 22 from SKB. Preparation of dd-RAD libraries and next generation DNA sequencing were performed by Floragenex (OR, USA) following the protocol of Truong et al. (2012). In brief, DNA was double digested with a combination of rare and frequent cutting endonucleases (PstI and MseI, respectively), followed by ligation with adaptors with individual indices. 1x100bp single end sequencing was performed on the resulting PCR-generated library using an Illumina HiSeq 4000 (CA, USA). 

Once obtained from Floragenex, a de novo catalogue of SNP loci was constructed using STACKS 2.0 (Catchen et al., 2013), given that the B. bufo genome (Streicher et al. 2021) was not yet available at the time of the analyses. The raw sequences were filtered and demultiplexed using the process_radtags pipeline. Reads with an uncalled base were discarded, as were reads containing a 15 bp window in which the average quality dropped below a phred score of 10 (i.e., a 90% probability of being correct). Barcodes and RAD-tags containing one mismatch to an expected sequence were retained. To identify an optimal set of parameter values in STACKS, we followed procedures described in Paris et al. (2017) and Rochette & Catchen (2017). Based on these considerations, a value of 4 was chosen for both the M and n parameters, which control the number of mismatches permitted between two alleles of an individual heterozygote locus, and two alleles in a locus across a population, respectively.