Marker development
We studied putatively sex-linked sequences that were identified by
RADseq from a sample of 40 agile frogs from a single clutch in
Switzerland (Jeffries et al., 2018). The 92 bp long RAD tags were mapped
to a genome assembly of the common frog (unpublished data, D.L.
Jeffries), a species closely related to the agile frog (Pyron & Wiens,
2011), using Magic-BLAST 1.3 (Boratyn, Thierry-Mieg, Thierry-Mieg,
Busby, & Madden, 2019). We concentrated on those tags that hit to the
agile frog’s sex chromosomes (Jeffries et al., 2018) uniquely or had an
e-value of at least 5 orders of magnitude lower than the next best hit.
First, we aimed to check if the putative sex-linked loci carried
sex-linked single nucleotide polymorphisms (SNPs) in our study
populations as well. Using the NCBI Primer-BLAST tool (Ye et al., 2012),
we designed primers for a total of 14 loci based on common frog genome
with aim for sequencing agile frog DNA around the sex-linked RAD tags so
we could sequence DNA fragments of about 220-1100 bp length. PCRs were
performed with these primers on DNA samples of morphologically sexed
adult agile frogs from Hungary in the laboratory of the Conservation
Genetics Group, Department of Ecology, University of Veterinary Medicine
Budapest (for PCR primers used for sequencing and detailed conditions
see Table S2; PCR programs are described in Table S3). Clear PCR
products in the expected length range were cut and purified from 2%
agarose gel, using NucleoSpin Gel and PCR Clean-up kit (Macherey-Nagel),
and ran on a 3130xl Genetic Analyzer (Thermo Fisher Scientific) at
BIOMI, Gödöllő, Hungary. Sequencer output files were analysed by the
STADEN package (Bonfield, Smith, & Staden, 1995; downloaded from
https://sourceforge.net/projects/staden/files) and sequences were
checked manually. In total, primers designed for 11 loci produced strong
PCR products close to the expected fragment sizes which were also
suitable for cutting from agarose gel, and 7 of these yielded
unambiguous DNA sequences of the target loci (Table S2). Three out of
these 7 loci contained sex-linked SNPs based on sequences from 10 agile
frogs (5 males and 5 females) from Hungary (Table S2), and we
denominated these Rds1, Rds2 and Rds3, according to their order on the
common frog’s chromosome 4 (that is corresponding to the agile frog’s
sex chromosome). Segregation of SNPs at all three loci matched
expectations for an XX/XY sex-determination system as found in Jeffries
et al. (2018). Because there is no sex chromosome sequence assembly
available for the agile frog, we estimated the distances between the
sex-linked SNPs based on the corresponding chromosome 4 of the common
frog (unpublished data of D. L. Jeffries; reported as chromosome 5 in
Jeffries et al. (2018).
We designed sexing primers for these three putatively sex-linked loci,
so that for each locus two fragments could be amplified in a single PCR:
one fragment amplified from both chromosomes X and Y, and a shorter
fragment amplified only from Y (i.e. if a Y-specific SNP was present).
The shorter amplicon is part of the X/Y-universal fragment (see Figure
S1). Using this method, successful amplification of the X/Y-universal
product means that the target locus is amplifiable in the investigated
DNA sample (i.e. positive control). If the Y-specific fragment is
amplified as well, that proves the presence of the Y-specific SNP (male
genotype). We designed primers specific for the Y-SNPs so that the SNP
was present at their 3’ end. To increase allelic specificity, a
mismatching base was artificially introduced at the
3rd position closest to the 3’ end of these primers
(replacing the original base in the sequence; following Liu et al.
(2012). On Rds2 two sex-linked SNPs were situated 11 nucleotides apart,
therefore the Y-specific primer binding to both of these SNPs did not
require the introduction of any artificial mismatch. PCR conditions for
each pool of sexing primers were optimized based on individuals with
known DNA sequence at the concerned locus. Specificity of the primers
was tested by agarose gel electrophoresis of PCR products from
individuals with known DNA sequence. Sexing PCRs were carried out in a
final volume of 16 µl containing 1.6 µl DreamTaq green buffer (10x,
Thermo Fisher Scientific), 0.65 µl dNTP (2 mM, Thermo Fisher
Scientific), primers of varying amount (Table 2), 0.065 µl DreamTaq DNA
polymerase (5U/µl, Thermo Fisher Scientific) and 20-100 ng genomic DNA.
PCRs were carried out on a Bioer Life ECO gene amplification instrument
(TC-96/G/H(b)C). Optimized sexing PCR profiles are described in Table
S3.
Because PCR optimization by the above method was insufficient for Rds3,
we developed an HRM-based (high-resolution melting) method for
sex-linked SNP-identification at this locus in the laboratory of the
Ruminant Genome Biology Research Group, NARIC Agricultural Biotechnology
Institute, Gödöllő, Hungary. Total HRM reaction volume was 15 µl,
containing 3 µl 5x HOT FirePol EvaGreen qPCR Mix Plus (ROX, Solis
BioDyne), 1 µl forward and 1 µl reverse primer (10 µM each; Table 2) and
80-100 ng genomic DNA. Reactions were performed in a Roche Light Cycler
96 Instrument (as described in Table S3) and the results were analysed
with the Light Cycler 96 v. 1.1.0.1320 software (Roche Diagnostics
International LTD). Detailed guidance for HRM-based sexing is available
in Figure S2. HRM allows us to differentiate not only between
individuals carrying and not carrying Y-SNP but it provides information
on further differences between individual genotypes as well (i.e.
presence of additional SNPs can be detected: Figure S2). Genotyping with
this method was validated by comparing the assumed genotype based on HRM
to DNA sequence data of 42 individuals. While PCR-based sexing allowed
us to detect the presence or absence of a Y-SNP (Figure S1), the HRM
method gave information on the presence of both the Y-SNP and the X-SNP
(Figure S2).