Methods

We trapped razorback musk turtles (Sternotherus carinatus) from the Leaf River in Hattiesburg, Mississippi using 3-ft-diameter hoop nets baited with sardines in soybean oil (Beach Cliff). We collected a 5-mm snippet of interdigital webbing from the hindfoot of each turtle to store as a genetic sample. The sample was preserved in 100% ethanol until total genomic DNA was extracted using a DNeasy Tissue Kit (Qiagen Inc). Genomic DNA from one razorback musk turtle was sent to the Savannah River Ecology Lab Molecular Ecology Lab (SREL MEL), where they prepared an Illumina pair-end shotgun library for sequencing (Lance et al. 2013) and subsequently used PAL_FINDER_v0.02.03 (Castoe et al. 2012) to identify di-, tri-, tetra-, penta-, and hexanucleotide microsatellite sequences (i.e., nucleotide repeat motifs) among the reads. Primer3 (Rozen and Skaletsky 1999) was used to develop primers for 6766 potential microsatellite loci. We screened a total of 60 penta- and tetranucleotide loci, focusing on loci whose sequences only appeared 1–2 times within the sequence reads, to reduce the likelihood of targeting repetitive loci.

We initially screened potential microsatellite loci for amplification success and variability across 5 individuals of Sternotherus carinatus. For those loci that amplified well, we subsequently genotyped 30 Sternotherus carinatus from the Leaf River in Hattiesburg, Mississippi (31°20′16″N, 89°16′48.2″W; WGS84). We also tested all loci for cross-amplification across 5 other kinosternids (Kinosternon baurii [n = 5], K. subrubrum [n = 5], S. depressus [n = 5], S. peltifer [n = 5], and S. odoratus [n = 4]). Amplifications were conducted in a total volume of 12.5 µl using 7.76 µl of dH₂O, 1.25 µl 10× standard Taq (Mg-free) buffer (New England Biolabs), 0.75 µl 2 mM dNTPs, 0.75 µl 25 mM MgCl₂, 0.25 units Taq polymerase (New England Biolabs), 0.4 µl of 10 mM M13 tailed forward (Boutin-Ganache et al. 2001) and reverse primers, 0.09 µl of 1 µM labeled M-13 primer (LICOR Co.), and 20–50 ng of DNA template. PCR cycling conditions were as follows: initial denaturation at 94°C for 2 min, 35 cycles of 30 sec at 94°C, 30 sec at a locus-specific annealing temperature of 54°C–56°C (see Table 1), and 1 min at 72°C with a final elongation of 10 min at 72°C. Microsatellites alleles were visualized on a polyacrylamide gel using a LICOR 4300 DNA Analyzer. Alleles were sized using GeneProfiler ver. 4.05 (LICOR Co). Some loci required further optimization in terms of PCR conditions with modifications made to the annealing temperature and the amount of Taq polymerase (Table 1). For the additional kinosternid species in this study, we only tested for amplification under standard PCR conditions with an annealing temperature of 56°C. We calculated summary statistics for each locus using GenAlEx 6.5 (Peakall and Smouse 2012), and we tested for Hardy-Weinberg equilibrium within loci and linkage disequilibrium between loci using the probability tests of GENEPOP for the web (Raymond and Rousset 1995; Rousset 2008). We adjusted our alpha values using a sequential Bonferroni correction to account for multiple comparisons (Rice 1989).

Table 1 Characteristics of 40 loci isolated for the razorback musk turtle (Sternotherus carinatus). Forward primers are reported without the M13 tail. Columns include repeat motif, number of individuals out of 30 (n) for which loci amplified, size range of alleles (Size bp), the number of alleles (N_A), the observed and expected heterozygosities (H_O and H_E), and the annealing temperature (T°C) for razorback musk turtles. Those loci marked with an asterisk (*) were optimized with increased Taq (0.1 units) and forward/reverse primers (0.5 µl). Dashes (—) represent loci that did not amplify in the razorback musk turtle but amplified in other kinosternid taxa at an annealing temperature of 56°C (Table 2).

View Fullscreen

Table 1 Continued.

View Fullscreen

Results

Of the 60 loci we tested, 40 amplified in one or more species (Table 1). Of these, we optimized reaction conditions for 25 microsatellite loci that amplified in Sternotherus carinatus samples (n = 30). Numbers of alleles for these loci ranged from 2 to 15 (mean = 7.6, SE = 3.4) for S. carinatus, with observed heterozygosity values of 0.138 to 0.963 (mean = 0.657, SE = 0.202) and expected heterozygosity values of 0.128 to 0.910 (mean = 0.687, SE = 0.207). Only one locus deviated significantly from Hardy-Weinberg equilibrium (Scar24), and no loci exhibited linkage disequilibrium. For the additional kinosternid species in this study, the total number of successfully amplified polymorphic loci was 26 for Sternotherus odoratus, 16 for S. peltifer, 13 for S. depressus, 14 for Kinosternon bauriii, and 6 for K. subrubrum (Table 2).

Table 2 All loci that cross-amplified in at least one kinosternid species. Columns represent species, parentheses indicate number of each species tested, numbers in columns represent number of alleles amplified at a locus. Dashes (—) = no amplification at this locus for the species.

View Fullscreen

Discussion

Mud and musk turtles of the family Kinosternidae are among some of the least-studied turtles in the world (Lovich and Ennen 2013), and this study is the first to optimize microsatellite loci designed specifically for a kinosternid turtle. The loci were designed by using DNA from the razorback musk turtle, but they cross-amplified well in 5 other species of mud and musk turtles. Microsatellite loci remain an affordable and effective method to study population genetics (Selkoe and Toonen 2006). They can be used at relatively little cost, which can be reduced further by multiplexing loci with alleles in different size ranges (e.g., Scar08 and Scar19; Scar24 and Scar36). Furthermore, in contrast to next-generation approaches like RAD-Seq, microsatellite loci are amenable to adding additional individuals to the data set as opportunity and funding allows.

Although populations of most kinosternids are considered stable across the southeastern United States (with the notable exception of the critically endangered Sternotherus depressus), there is still much we do not know about the evolutionary history of these species, as made evident by the recent description of Sternotherus intermedius (Scott et al. 2018). The loci described herein should prove useful for studies evaluating the contemporary and historic levels of connectivity (i.e., gene flow) in kinosternid species as well as for evaluating levels of genetic diversity within and among populations. Because lower genetic diversity has been correlated with lower fitness in turtles (Ennen et al. 2010), it is important to understand gene flow among isolated or fragmented populations that may be susceptible to genetic drift or inbreeding depression, particularly in threatened species (Ennen et al. 2010; Shaffer et al. 2015). A better understanding of the population genetics of these species will allow for more informed management decisions and will benefit the long-term survival of a species. We are currently using these loci to evaluate range-wide population structure in razorback musk turtles and to assess patterns of co-ocurrence and potential hybridization in S. carinatus and S. peltifer (G.J.B., unpubl. data, 2020).