Habitat degradation can have significant effects on native species inhabiting natural ecosystems. Within oak barrens and oak-pine barrens ecosystems, there is a complex interspecies interaction between the federally endangered Karner blue butterfly (Lycaeides melissa samuelis) and its obligate host plant, wild lupine (Lupinus perennis L.). Recruitment of wild lupine is critical for maintaining butterfly populations; however, this recruitment can be impeded by habitat fragmentation. Reduced recruitment can result in low genetic diversity in isolated populations, limiting its adaptive potential to respond to environmental change. This study was aimed at understanding the genetic diversity and population structure of wild lupine populations throughout central and west Michigan. We identified significant population structure across most of the populations sampled, with only two sites not significantly different from each other. No sites within our study area displayed statistically significant levels of inbreeding.  There are also at least two genetic clusters of wild lupine present within our study region, although there is significant overlap among these groups, indicating that genetic differentiation among clusters may be limited.

Tissue samples from lupine were collected between June and July 2018 from each of the seven locations, by taking 2-3 leaves from 22-30 plants across each location (Table 1). Leaves were placed in coin envelopes and dried in silica until DNA extraction. Genomic DNA was extracted using Qiagen DNeasy plant mini extraction kits following the manufacturer's protocols (QIAGEN, Hilden, Germany). The DNA was amplified at nine microsatellite loci. These microsatellite primers included primers GAC2, AT9, GAT5, GAT7 (Shi et al., 2004), primers developed for L. nanus (Douglas ex Benth.) (LUNA3, LUNA9, LUNA12, LUNA17) (Morris, 2009), and a primer developed for L. polyphyllus (Lindl.) (LUP4) (Li et al., 2013; GenBank Accession KC814580). PCR reactions consisted of 1X PCR buffer, 2.0 mM MgCl₂, 300 mM dNTPs, 0.08 mg/mL bovine serum albumin (BSA), 0.4 mM forward primer, 0.4 mM reverse primer, and 0.25 units Taq polymerase (ThermoFisher, Waltham, MA) in a 10 uL reaction volume. The loci were amplified using the following thermal cycler conditions: 94º C for 1 min, 35 cycles of 94º C for 1 min, 55º C for 1 min, 72º C for 1 min, and a final elongation step of 72º C for 5 min. Successful amplification was visually checked on a 2% agarose gel. We then analyzed the PCR product via fragment analysis on an ABI Genetic Analyzer 3130xl (Applied Biosystems) and scored individual genotypes in GeneMapper v5 (Applied Biosystem). For quality assurance, all individuals were scored by two different individuals and then a third individual checked a subset of the genotypes to verify peak calls. We analyzed the genotypes in MICRO-CHECKER v 2.2.3 (Van Oosterhout et al., 2004) to identify any potential error in scoring due to null alleles, large allele dropout, or microsatellite stutter. GAT7 showed potential for containing null alleles across multiple populations and was removed from the analysis. There was also some evidence of homozygous excess for some loci in the FR-MI-B population. Prior to analysis, all loci were checked for Hardy-Weinberg Equilibrium and linkage disequilibrium for each population. If a locus was out of HWE equilibrium at more than 60% of the populations, it was discarded. However, no loci met these criteria, and the remaining eight loci were maintained for further analysis.

Files can be opened in Excel or any text editing program (such as TextEdit). The data are formatted for GENALEX, which is a macro for population genetics analysis in Excel.