Development, characterization, and cross-amplification of polymorphic microsatellite markers for North American Trachymyrmex and Mycetomoellerius ants

Objective The objective of this study is to develop and identify polymorphic microsatellite markers for fungus-gardening (attine) ants in the genus Trachymyrmex sensu lato. These ants are important ecosystem engineers and have been a model group for understanding complex symbiotic systems, but very little is understood about the intraspecific genetic patterns across most North American attine species. These microsatellite markers will help to better study intraspecific population genetic structure, gene flow, mating habits, and phylogeographic patterns in these species and potentially other congeners. Results Using next-generation sequencing techniques, we identified 17 and 12 polymorphic microsatellite markers from T. septentrionalis and Mycetomoellerius (formerly Trachymyrmex) turrifex, respectively, and assessed the genetic diversity of each marker. We also analyzed the cross-amplification success of the T. septentrionalis markers in two other closely related Trachymyrmex species, and identified 10 and 12 polymorphic markers for T. arizonensis and T. pomonae, respectively.


Introduction
Fungus-gardening ants (tribe Attini) have been subject to decades of research elucidating their major evolutionary transitions and radiations [1][2][3][4][5]. Despite these revolutionary advances, most attine research has occurred at higher taxonomic levels (species or higher). Consequently, we know very little about the microevolutionary processes that may underlie macroevolutionary (phylogenetic) patterns. This lack of understanding is partially technical since most genetic markers used to construct phylogenies typically exhibit little intraspecific variation, for both the ants and fungi [2,[6][7][8]. Mitochondrial DNA (mtDNA) markers, while useful for large scale geographic studies [9][10][11], may be generally problematic because some attines may have nuclear-incorporated mtDNA-like sequences or whose genomes contain pseudogenes [7,12]. Moreover, mtDNA is maternally inherited, which may or may not be indicative of intraspecific patterns of gene flow since male and female ants are known to exhibit different patterns of gene flow [13][14][15]. Finer-level intraspecific examinations have typically required specific diploid markers such as microsatellites or single-nucleotide polymorphisms (SNPs) [16][17][18][19]. Both approaches represent co-dominant expression and genome-wide distributions, but microsatellites are short, non-coding, highly variable repetitive motifs [20], whereas SNPs are shorter and found throughout coding and non-coding genomic regions [21].
SNPs are popular in the genomics age because of the large quantities of potentially informative sites quickly generated in a single sequencing run, but they require considerable time and effort to identify and analyze, and are typically restricted to single studies. Conversely, microsatellite markers, once developed, can be used repeatedly in multiple studies, which may reduce their overall expense on a per-project basis [22]. Moreover, recent comparative studies in insects, fungi, and birds reported no major population genetic structural differences between these two approaches [23][24][25][26], or that more SNPs were required to produce the same information as a single microsatellite marker [27].
Traditional microsatellite development techniques (i.e., constructing libraries, cloning) have yielded polymorphic microsatellite markers for both higher and lower attines [28][29][30][31][32]. However, because these markers were developed for tropical attines, they may or may not be useful for North American species (J.N. Seal, unpublished observations), which tend to belong to phylogenetically distinct lineages [3,33]. Next-generation sequencing (NGS) techniques have revolutionized microsatellite marker development in terms of decreased time and cost, and increased quantity of potential markers [34][35][36], although only a few studies have utilized these techniques for attines. Extensive marker sets exist for Acromyrmex lundii [37], Atta laevigata [38], and two species in the lower attine genus Mycocepurus [18,39].
Here, our aim was to use NGS techniques to develop microsatellite markers for two additional attine species, Trachymyrmex septentrionalis and Mycetomoellerius turrifex. Both are common and ecologically important fungus-gardening ants throughout southeastern North America [10,40,41]. We also explored the potential for cross-amplification of the T. septentrionalis markers in two closely related species: T. arizonensis and T. pomonae. These three species belong to Trachymyrmex sensu stricto, unlike M. turrifex which belongs to a new genus, Mycetomoellerius [3,33,42]. These markers will provide a new tool to study patterns of attine intraspecific genetic variation and phylogeographic structure.

Methods
Primer sets were developed from a single T. septentrionalis forager collected in north Florida ( Illumina paired-end shotgun library following standard Illumina TruSeq DNA Library Kit protocols and using a multiplex identifier adaptor index. Sequencing was conducted on an Illumina HiSeq 3000 platform (San Diego, CA, USA) producing 100-bp paired-end reads. PAL_ FINDER v.0.02.03 [43] was used to analyze five million resulting reads, and those with di-, tri-, tetra-, penta-, and hexanucleotide microsatellites were extracted. Positive reads were batched to a local installation of Primer3 v.2.0.0 [44,45] for primer design. Markers for which the primer sequence only occurred three or four times in the five million reads were selected to avoid duplication. From this selection of 9293 (T. septentrionalis) and 6943 (M. turrifex) markers, we chose a subset to test (46 for T. septentrionalis; 20 for M. turrifex) following an M13-tail PCR protocol [46].
We extracted genomic DNA from individual worker ants (from restricted portions of their respective ranges; Additional file 1) preserved in 95% ethanol at − 20 °C using a QIAamp DNA Micro Kit (QIAGEN). In particular, ants used in the development of T. septentrionalis markers were all collected in Florida (i.e., not Texas) ( Table 1) We scored alleles in Geneious v.10.2.3 [47]. For markers that resulted in high quality PCR product and exhibited polymorphism, we measured genetic variability for each species. We estimated the number of alleles per locus (K), observed and expected heterozygosity (H o and H e ), and the probability of identity (PI; the probability of two independent samples having the same genotype) using GenAlEx v.6.5 [48,49]. Sample sizes (numbers of individuals tested) for each marker are provided (Tables 1  and 2, Additional files 2 and 3). We assessed deviations from Hardy-Weinberg equilibrium (HWE) and tested for linkage disequilibrium (LD) across all pairs of markers using GENEPOP v.4.2 [50] employing Bonferroni corrections for multiple comparisons. The presence of null alleles was assessed using Microchecker [51].

Results
For Trachymyrmex septentrionalis markers tested on T. septentrionalis ants, 37 of the original 46 positively amplified with 17 being reliably amplified and scored. The other 20 either showed weak amplification, weak polymorphism (2-4 alleles per marker), monomorphism, or multi-peak profiles. The number of alleles across all samples of the 17 reliable markers ranged from 9 to 33 (average = 16.2). H o varied from 0.11 to 0.89 (average = 0.62), while H e ranged from 0.65 to 0.94 (average = 0.84). After Bonferroni correction, 9 markers showed significant deviations from HWE. There were two cases of LD detected across the 136 paired markers (between Ts4 and Ts5, and Ts3 and Ts43; Table 1; Additional file 4). Most of the loci tested significantly for the presence of null alleles, which suggests that T. septentrionalis may have even higher allelic diversity than what we captured in our sampling (Table 1).
Of the 20 original Mycetomoellerius turrifex markers, 17 positively amplified with 12 reliably amplified and scored. The number of alleles across all samples ranged  Table 2). We tested cross-species amplification of the 46 T. septentrionalis markers developed here in both T. arizonensis and T. pomonae. For T. arizonensis (n = 8), 24 markers were positively amplified with 10 reliably scored. Of those 10, all revealed at least 2 alleles, 8 revealed 3 or more alleles, and 9 exhibited some heterozygosity. There were 9 markers in HWE and one violated HWE (Ts38). There was no LD (Additional files 2, 4). For T. pomonae (n = 6), 23 markers were successfully amplified and 12 were reliably scored. Of those 12, all revealed at least 2 alleles, 8 revealed 3 or more alleles, and 10 exhibited some heterozygosity. There were 10 in HWE, one violated HWE (also Ts38), and one could not be computed (Ts7) because there was only one copy of the second allele. There was no LD (Additional files 3, 4). Only one of the loci in T. pomonae and only two in T. arizonensis appeared to have null alleles; this likely results from examining ants collected across a few hectares of pine oak forest (Southwest Research Station) (Additional files 2 and 3).

Discussion
These newly developed microsatellite markers provide a promising new tool for population genetic, mating, dispersal, and phylogeographic studies of Trachymyrmex septentrionalis and Mycetomoellerius turrifex populations across their ranges. Positive cross-species amplification also indicates that some of the T. septentrionalis markers appear to be conserved in at least two other congeneric species (T. arizonensis and T. pomonae). This indicates that future analyses for additional individuals across a wider geographic region may be informative for understanding similar questions for T. arizonensis and T. pomonae and perhaps other closely related Trachymyrmex. It also seems likely that the markers developed here for M. turrifex should cross amplify in closely related   Mycetomoellerius species, such as the highly studied M. zeteki [42,52,53], among others. While all markers across all four species showed polymorphism/allelic variation, several markers (particularly within T. septentrionalis) were hypervariable (over 20 alleles per locus; Table 1). These hypervariable markers, collectively with other polymorphic markers, may be informative for investigating intraspecific dynamics at an especially fine-scale (e.g., mating habits such as multiple-mating or multiple-queen hypotheses by comparing relatedness of individuals within colonies). Despite the extensive polymorphism illustrated by these markers, evidence of null alleles suggests that more alleles may be discovered. The frequency of null alleles may be more pronounced in populations that already demonstrate significant genetic structure [54].
The new markers developed here may also be used in conjunction with mtDNA to better understand dispersal, signatures of population expansion or contraction, and phylogeographic history. For example, male and female ants of several species may mediate gene flow differentially through sex-specific dispersal patterns [13,14,[55][56][57], but this has yet been tested in Trachymyrmex or Mycetomoellerius ants. By employing marker systems with different modes of inheritance, we may be able to discern how ecological dispersal patterns influence ant genetic structure across populations.
In a phylogeographic context, the application of both maternally and biparentally inherited markers may also help to uncover and track the progression of previously unrecognized or poorly understood historical processes such as species origin, divergence, colonization, and genetic drift. For example, the extensive evidence for null alleles in T. septentrionalis is likely driven by phylogeographic structure in this species across the relatively restricted range examined in this study (northern Florida) [10]. We explored the 17 T. septentrionalis markers across a small subset (total n = 12) of T. septentrionalis ants from Texas as compared to a random subset of T. septentrionalis ants from Florida. The allelic diversity appears to be much higher in Florida than in Texas (see Additional file 5), which may suggest some fundamental genomic differences in populations east and west of the Mississippi River. These preliminary results align with previous mtDNA results that suggest (1) a significant divergence between populations on either side of the Mississippi River and (2) limited genetic diversity west of the Mississippi River [10,11]. Future analyses comparing the allelic diversity of populations east and west of the Mississippi River would be insightful for better understanding these patterns.

Limitations
• The restricted sample size and geographic distribution for T. arizonensis and T. pomonae may underestimate potential polymorphic markers in these species. • Markers not deviating from HWE in our initial screenings could potentially deviate in other populations, which may reduce the number of usable markers in certain genetic studies; on the other hand, markers deviating from HWE in the studied populations might not deviate in other populations, which would increase the number of markers. • The presence of null alleles could be a result of potential phylogeographic structure in the populations studied (especially T. septentrionalis), as a result polymorphism may be underestimated. • Sample sizes per marker for the Texas T. septentrionalis subset were smaller compared to Florida, which may limit our interpretations; however, reduced microsatellite diversity would be consistent with lower mtDNA diversity in western populations (Texas).