Characterization of 16 novel microsatellite loci for Ephippiger diurnus (Orthoptera: Tettigoniidae) using pyrosequencing technology and cross-species amplification

. A novel panel of 16 microsatellite markers, obtained by pyrosequencing of enriched genomic libraries, is reported for the ﬂ ightless European bushcricket Ephippiger diurnus (Dufour) (Orthoptera: Tettigoniidae). Five multiplex and one simplex PCR protocols were optimized, and the polymorphism at the 16 loci was assessed in two natural populations from southern France. The mean allele number and (expected mean heterozygosity) were 8.94 (0.71) and 6.57 (0.70), respectively, in each population. Several loci were at Hardy-Weinberg disequilibrium (HWD), possibly due to the incidence of null alleles. The occurrence of null alleles has been previously reported for this species, and it is a common feature of microsatellite loci in Orthoptera. Cross-ampli ﬁ cation tests demonstrated the transferability of some of these loci to other ephippigerine species. The microsatellite loci reported here substantially increase the number of available loci for this species and will afford an accurate picture of E. diurnus phylogeography, the genetic structure of its populations, and an improved understanding of the evolution of male song and other sexually-selected traits in this highly variable species.

free of null alleles and conforming to Hardy-Weinberg expectations) has proven challenging in E. diurnus. Indeed, a set of 16 loci had been reported for this species before our study (Hockham et al., 1999;Hamill et al., 2006). According to the authors themselves, these loci displayed strong heterozygote defi cit, and the incidence of null alleles was considerable for some of them. In a preliminary trial we tested 13 of these available 16 loci on samples from highly divergent populations previously characterized for mitochondrial DNA COI variation (Party et al., 2015). Most loci failed to amplify and/or presented complex allelic patterns impeding their scoring. This situation signifi cantly reduced the number of available markers to only fi ve, which is a minimum value for population genetic analyses. We therefore applied highthroughput (pyrosequencing) technology to a partial genomic library enriched in microsatellite motifs in order to increase the number of loci and fi lter out those of low quality according to criteria detailed below.

Sample collection and DNA extraction
Fifty-one specimens of E. diurnus were collected from nine localities in southern France between 2011 and 2014 ( Fig. 1). Hind femora were dissected and preserved in 95% ethanol for DNA Characterization of 16 novel microsatellite loci for Ephippiger diurnus (Orthoptera: Tettigoniidae) using pyrosequencing technology and cross-species amplifi cation INTRODUCTION The European bushcricket Ephippiger diurnus (Ortho ptera: Tettigoniidae) has attracted considerable attention among behavioral and evolutionary biologists because of its diverse calling songs (Duijm, 1990;Ritchie, 1996), large spermatophore (Barbosa et al., 2016), and a strong population genetic structure (Spooner & Ritchie, 2006). E. diurnus are fl ightless, do not migrate, and have specifi c habitat preferences, and previous studies showed that they are distributed in geographically isolated, genetically differentiated populations throughout their range in southern France and northeastern Spain (Party et al., 2015). These geographically separate populations generally exhibit distinctive male songs that are characterized by a specifi c number of syllables per call (Ritchie, 1991(Ritchie, , 1996, and some attempts have been made to relate the song trait to phylogeography by evaluating mitochondrial DNA (COI) divergence (Party et al., 2015). The various populations can be crossed in the laboratory (Ritchie, 2000), but the full potential of such hybridization is unknown. To determine the phylogeography of E. diurnus with greater precision and to explore the evolution of song diversity, genetic markers that afford reliable, fi ne-level resolution of population differences are needed.

NOTE
Because of the high level of divergence among mtDNA COI clades (see Fig. 4 in Party et al., 2015) we tested PCR amplifi cation of the 100 loci in four specimens of E. diurnus collected in Mireval, Sode, Port de Lers, and Col de Mantet (Fig. 1), and belonging to the two main COI clades ( Fig. 1 in Party et al., 2015). Sequences and primers for the 100 loci are given in Table S1. Thus, we could retain only those loci amplifying unambiguously in all clades (Table S1). All amplifi cations were achieved with an ABI GeneAmp PCR System 2700 thermal cycler. PCR reactions were carried out in a 10-μl solution containing the following: 2 mM MgCl 2 , 2 mM DNTPs, 1 × PCR Buffer, 0.5 unit of GoTaq G2 polymerase (Promega, Charbonnieres, France), 2 μM of each forward and reverse primer and ~10 ng of template DNA. PCR cycling conditions were as follows: 95°C 3 min, followed by 30 cycles at 94°C 30 s, 60 s at 60°C, 72°C for 45 s, and a fi nal extension step of 10 min at 72°C. PCR products were resolved in a 2% agarose gel. Twenty-one primers not amplifying in all four specimens (i.e. partial PCR amplifi cation) were not considered. Forty-two primer pairs that showed clear, reproducible and unique fragments in the four specimens were retained for further analysis. Among them, eight loci showed an incidence of smear or amplifi ed nonspecifi c bands, thereby justifying a +3°C increase of the annealing temperature (Table S1). Lastly, 37 loci did not amplify at 60°C and were tested at 52°C using the same PCR conditions as above but were not tested in the following steps described below (Table S1).
Fragment analysis of 42 loci followed the cost-effective M13 fl uorescent protocol described by Schuelke (2000) with modifications described below. Each forward primer was tagged at its 5' end with one 18-19 bp tail described in Culley et al. (2013) and one fl uorescent label depending on the expected amplifi cation size to allow posterior PCR multiplexing (Table S1). The combination of tails and fl uorescent labels were as follows: M13 modA-NED, M13 modB-PET, T7 term-VIC and M13 (-21)-FAM (Applied Biosystems, Warrington, UK, see Table 1). Simplex PCR tests were performed on four to eight specimens to confi rm PCR amplifi cation with tailed primers. The 6.25 μl PCR reaction contained: 3.25 μl Multiplex PCR Master mix (Qiagen), 1 μl of a primer mix per locus containing: 2 μM of each reverse and labeled tail primer and 0.5 μM of the forward tailed primer (ratio: 1 : 1 : 1/4, see Culley et al., 2013), 1 μL H 2 O and 1 μl of DNA (~10 ng/μl). PCR cycling conditions followed a denaturing step of 15 min at 95°C, then 30 cycles at 94°C for 30 s, 60°C or 63°C (Table S1) for 45 s and 72°C for 45 s; and then 8 cycles of 94°C for 30 s, 53°C for 45 s and 72°C for 45 s and a fi nal elongation step of 10 min at 72°C. PCR products were visualized on a 2% agarose gel. Fragment analysis was conducted on a 3730 xl DNA Analyzer (Applied Biosystems) using the GeneScan 500 LIZ as internal size standard (Applied Biosystems) and 1 to 2 μl of PCR product (1 : 20 dilution). After visual screening of electropherogram profi les in GeneMapper version 5.0 (Applied Biosystems), 25 primers were selected for their scoring in additional specimens. Finally, 16 primer pairs showing polymorphism and unambiguous profi les were retained. For loci showing noisy electropherograms, the annealing temperature was increased by 3°C (Table S1).

Multiplex PCR amplifi cation
We used Multiplex Manager version 1.0 (Holleley & Geerts, 2009) to determine the best combination of loci in a multiplexed PCR amplifi cation protocol. Five multiplex PCR amplifi cation reactions and one simplex PCR were defi ned for the fi nal set of 16 loci, and these were amplifi ed in two populations of E. diurnus (Vias and Peyriac de Mer; Table 2) belonging to each of the two main mtDNA COI clades described in Party et al. (2015, see extraction. Whole genomic DNA was extracted using the DNA Easy Blood and Tissue kit (Qiagen, Hilden, Germany) following the manufacturer's instructions. DNA quality and molecular weight were assessed in a 1% agarose gel and with a Nanodrop 2000 spectrophotometer (Thermo Scientifi c, Villebon sur Yvette, France).

Microsatellite isolation
Five μg of DNA were obtained by pooling individual DNA extracts from eight insects sampled in eight of nine localities (Fig.  1). The DNA pool was sent to Genoscreen, Lille, France (www. genoscreen.fr) for microsatellite isolation through 454 GS-FLX Titanium pyrosequencing of enriched DNA libraries following the approach described by Malausa et al. (2011). Briefl y, enriched libraries were constructed using eight microsatellite probes (TG, TC, AAC, AAG, AGG, ACG, ACAT, ACTC), and the resulting library was sequenced on a GsFLX PTP. The resulting 72,447 reads were analyzed using the program QDD (Meglecz et al., 2010) and sorted according to the following criteria: number of microsatellite repeats ≥ 5, microsatellite motif not interrupted by any other bases or sequences, fragment size ≥ 80 bp. A fasta fi le with 5,027 reads containing a microsatellite repeat and a list of optimized primer pairs for 503 reads (size range: 90-319 bp) was provided by Genoscreen. Within these reads, 323 primer pairs with expected fragment sizes ≥ 120 bp were chosen. Special attention was paid to homologous sequences shared among distinct reads: short, repeated sequences in the vicinity of microsatellites are frequently shared among distinct loci and impede consistent single locus PCR amplifi cations if primers overlap them (Meglécz et al., 2007). To avoid this problem, sequences homologous among different reads identifi ed after an "all-against-all" BLASTn analysis (http://blast.ncbi.nlm.nih.gov/Blast.cgi) were masked before primer design. In the same line, all sequences were checked and masked for the presence of known annotated repeated elements in the fl anking regions by the RepeatMasker software (http://www.repeatmasker.org/) and the related domestic silkworm database. Finally, 100 loci were chosen to proceed with the fi rst PCR screening using unlabeled primers (Eurofi ns Genomics, Ebersberg, Germany). For some of these loci new primers were designed when the expected amplifi cation size was not adequate for the posterior multiplexing PCR procedure and or when primers provided by Genoscreen were located in zones of high homology according to our BLASTn analysis (Table S1).   Table  S1).

Polymorphism analysis
The allele number, heterozygosity and Hardy-Weinberg equilibrium (HWE) for each loci were computed in GENEPOP version 4.3 (Rousset, 2008). P values were adjusted for multiple tests of signifi cance using the sequential Bonferroni correction at the 5% nominal level (Rice, 1989). Incidence of null alleles was assessed on Micro-checker version 2.2.3 (Van Oosterhout et al., 2004), and when signifi cant their frequency was obtained using the Brookfi eld 1 method. GenBank numbers (KU512644-KU512669) were attributed only to primary sequences (Table S1).

Cross-amplifi cation
The fi nal set of 16 loci was cross-amplifi ed in two other species of Ephippiger and in one species of Uromenus using the simplex PCR amplifi cation protocol described above (Table 3).

RESULTS AND DISCUSION
Microsatellite loci isolated from E. diurnus showed moderate to high levels of allelic diversity and were polymorphic in both populations analyzed, except for locus Ediur86 in Vias. The number of alleles ranged from four to 20 in Vias (N = 23), with a mean of 8.94 alleles per locus, and from three to 13 in Peyriac de Mer (N = 21), with a mean of 6.57 alleles per locus. The expected heterozygosity ranged from 0 to 0.95 in Vias, with a mean of 0.71, and from 0.41 to 0.86 in Peyriac de Mer, with a mean of 0.70. Heterozygote defi ciency and signifi cant departure from HWE were detected for several loci in both populations after Bonferroni correction ( Table 2). The analysis of the distribution of homozygote size classes on Micro-checker suggested the incidence of null alleles that might contribute to the observed heterozygote defi ciency and HWD in both populations (Table 2). No scoring errors due to stuttering or large allele drop-out were detected. The frequency of null alleles ranged from 0.09 to 0.40 in Vias and from 0.12 to 0.31 in Peyriac de Mer. A high proportion and preva-lence of null alleles at microsatellite loci are common in Orthoptera (Zhang et al., 2003;Chapuis et al., 2005;Chapuis & Estoup, 2007), and Ephippiger diurnus seems not to be an exception. Previous reports on microsatellite characterization for this species also showed a considerable prevalence of null alleles (Hockham et al., 1999;Hamill et al., 2006). The distribution of E. diurnus in small, genetically differentiated populations, an outcome of low dispersal and specifi c habitat preferences, probably contributes to the observed heterozygote defi ciency.
The microsatellite markers we report for E. diurnus will be valuable for fi ne level phylogeographic analysis and studies at larger geographical scale, and for studying the diversity of the male calling song and female preferences. Cross-amplifi cation tests showed the transferability of this set of microsatellite markers to other Ephippiger species as well as to another ephippigerine, Uromenus rugosicollis (Table 3).