Geranium carolinianum L. (Geraniaceae) is a weedy winter annual herb native to North America and naturalized in East Asia, South America, and the Caribbean (Aedo, 2000). In China, it is naturalized along roadsides and in agricultural fields, and is considered an invasive species with minor environmental impacts (Liu et al., 2006). The origin and colonization history of these introduced populations is unknown. Although it is not closely related to the horticultural geranium (genus Pelargonium L'Hér. ex Aiton), genetic resources developed for G. carolinianum are potentially transferable to other Geranium L. species, which includes annuals and perennials that are both native and introduced in North America (Aedo, 2000). Here, we present 12 microsatellite loci developed using targeted enrichment and next-generation sequencing.

METHODS AND RESULTS

DNA was extracted from leaf tissue using a QIAGEN Plant Mini Kit (QIAGEN, Valencia, California, USA) and enriched for microsatellites using three probe mixes [Mix 2 = (AG)₁₂, (TG)₁₂, (AAC)₆, (AAG)₈, (AAT)₁₂, (ACT)₁₂, (ATC)₈; Mix 3 = (AAAC)₆, (AAAG)₆, (AATC)₆, (AATG)₆, (ACAG)₆, (ACCT)₆, (ACTC)₆, (ACTG)₆; Mix 4 = (AAAT)₈, (AACT)₈, (AAGT)₈, (ACAT)₈, (AGAT)₈] as in Glenn and Schable (2005). DNA from one individual was digested separately with RsaI (New England Biolabs, Ipswich, Massachusetts, USA) and AluI (New England Biolabs). Digests were pooled, a single adenosine was added with exo-Klenow fragment (New England Biolabs), and fragments were ligated to CAG-SimpleXT-13 adapters (upper oligo = GTTTCAGTCGGGCGTCATCA CCAGCACCGGAACAT; lower oligo = /5Phos/TGTTCCGGTGCTGG). Linker-ligated DNA was pooled, denatured, and hybridized to biotinylated microsatellite oligonucleotide mixes, which were subsequently captured on magnetic streptavidin beads (Dynabeads M-280; Invitrogen, Carlsbad, California, USA), and remaining DNA was washed away. The enriched DNA was eluted and amplified in PCR using Phos_Pig_CAG (/5Phos/GTTTCAGTCGGGCGTCATCA). PCR reactions were run in a 25 µL total volume with 2 µL eluted DNA, 0.2 µL Phusion Taq polymerase (5 units/µL; New England Biolabs), 1× Phusion reaction buffer (New England Biolabs), 0.025 mg/mL bovine serum albumin (BSA; New England Biolabs), 2.0 mM MgCl₂, 0.15 mM each dNTP (Sigma-Aldrich, St. Louis, Missouri, USA), and 0.5 µM Phos_Pig_CAG. Cycling conditions were 2 min at 95°C; 25 cycles of 20 s at 95°C, 20 s at 60°C, and 1.5 min at 72°C; and a final 30-min extension at 72°C. Multiple libraries with additional different SimpleX sequences were then pooled and ligated into a 454 RL MID-tagged library, pooled with additional 454 RL MID-tagged libraries, and sequenced on half a run of the Roche 454 FLX with Titanium chemistry using the manufacturer's protocols.

Following demultiplexing, 1115 sequences with microsatellites were identified from 5586 sequence reads using MSATCOMMANDER version 0.8.1 (Faircloth, 2008). All microsatellite-containing sequences are provided in  Appendix S1 (APPS1300006_AppendixS1.txt). MSATCOMMANDER also designs primers for microsatellite-containing sequences using Primer3 (Rozen and Skaletsky, 2000). Sequences were discarded if Primer3 could not design acceptable primers. All primers (excluding those tested as described below, a total of 433 pairs) are reported in  Appendix S2 (APPS1300006_AppendixS2.xlsx). To avoid selecting primers within highly repetitive regions of the genome, sequences were assembled into contigs by automatic assembly using default parameters in Sequencher version 4.7 (Gene Codes Corporation, Ann Arbor, Michigan, USA). Microsatellite-containing sequences that were at high frequency (i.e., members of contigs with >10 sequences) were not considered for screening.

Of the remaining sequences, we selected 37 primer pairs for testing. The majority of primers tested were based on unique sequences in the data set and amplified tri- or tetranucleotide repeats. Forward primers were tagged on the 5′ end with the universal “CAG” sequence (CAGTCGGGCGTCATCA) to allow fluorescent labeling of PCR product using a 3-primer protocol as detailed below. Additionally, a 5′GTTT “pigtail” was added to some reverse primers to ensure consistency in amplicon size (Brownstein et al., 1996). These 37 primer pairs were screened using DNA template from 12 samples across G. carolinianum's native range. All primer pairs were initially tested using a “touchdown 58” PCR protocol. Cycling conditions were 3 min at 95°C; nine touchdown cycles of 30 s at 95°C, 30 s at 65-58°C (starting at 65°C and decreasing one degree each cycle), and 45 s at 72°C; 30 cycles of 30 s at 95°C, 30 s at 50°C, and 45 s at 72°C; and a final extension of 20 min at 72°C. Primer pairs that did not amplify consistently at touchdown 58 were tested with a “touchdown 50” protocol, which was identical to touchdown 58 except the annealing temperatures for the touchdown cycles decreased from 58°C to 50°C, followed by 30 cycles at an annealing temperature of 50°C. All PCR reactions were carried out in a 12.5 µL total volume with 10-200 ng DNA template, 0.1875 units JumpStart Taq DNA polymerase (Sigma-Aldrich), 1× JumpStart reaction buffer without magnesium (Sigma-Aldrich), 1× (0.1 mg/mL) BSA, 2 mM MgCl₂, 0.15 mM each dNTP, 0.025 µM forward (CAG-tagged) primer, 0.25 µM reverse (untagged) primer, and 0.25 µM universal CAG primer labeled with a 5′ FAM, HEX, or NED fluorescent dye. Primer pairs with consistent amplification at either touchdown 58 or touchdown 50 were screened for polymorphism by fragment analysis. PCR products from the initial 12 DNA samples were sized on an ABI 3730xl capillary sequencer (Applied Biosystems, Carlsbad, California, USA) using a ROX-labeled internal size standard (GGF500R; Georgia Genomics Facility, Athens, Georgia, USA). Chromatograms were analyzed using GeneMapper version 3.7 (Applied Biosystems).

Of the 37 primer pairs tested, we identified eight polymorphic and four monomorphic loci with high-quality amplification (Table 1). Primers that performed poorly in the screening are listed in Appendix  S3 (APPS1300006_AppendixS3.docx). Monomorphic loci were tested further against either 48 or 96 individual DNA samples from across the native and invasive ranges, but no polymorphism was found. Polymorphic loci were used to assess genetic variation in 103 individual DNA samples from three native populations in the United States and three invasive populations in eastern China (Table 2). For GC7 and GC35, only a subset of DNA samples were available for one of the populations from Georgia, USA, and the population from Henan, China, so sample sizes from other populations were increased to reach a total sample size of roughly n = 100 per locus. Voucher specimens from the two Georgia and three China populations have been deposited at the University of Georgia Herbarium (GA). We were unable to obtain a voucher for the North Carolina population.

TABLE 1.

Characteristics of microsatellite primers developed in Geranium carolinianum.

We calculated the number of alleles, observed heterozygosity (H_o), and expected heterozygosity (H_e) for each polymorphic locus and population in GenAlEx version 6.2 (Peakall and Smouse, 2006). Results are given in Table 2. Across all samples, number of alleles per polymorphic locus ranged from two to seven, and five loci were monomorphic in at least one population. Mean population H_e(averaged across loci) ranged from 0.100 to 0.290. Within populations, H_o for individual loci ranged from zero to 0.857, and H_e ranged from 0.046 to 0.559. We did not test for Hardy—Weinberg equilibrium because this is a colonizing, mixed-mating species.

CONCLUSIONS

We report 12 microsatellite loci in G. carolinianum, eight of which are shown to be variable in native and invasive populations. Forward primer sequences are labeled with a 5′ universal sequence to allow fluorescent tagging of amplicons during PCR, and reverse primer sequences (with the exception of GC1-R, GC6-R, GC7-R, and GC10-R) are labeled with a 5′ GTTT pigtail sequence for consistent genotyping. Although not tested, addition of a 5′ pigtail to these primers should facilitate accurate genotyping. These microsatellite loci will be useful for studies of genetic variation, gene flow, and mating system in G. carolinianum and may be transferable to related species.

TABLE 2.

Results of primer screening of eight polymorphic microsatellite loci in three native (United States) and three invasive (China) populations of Geranium carolinianum.