Electric eels galore: microsatellite markers for population studies

Abstract Fourteen novel microsatellite loci are described and characterized in two species of electric eels, Electrophorus variiand E. voltaifrom floodplains and rivers of the Amazon rainforest. These loci are polymorphic, highly informative, and have the capacity to detect reliable levels of genetic diversity. Likewise, the high combined probability of paternity exclusion value and low combined probability of genetic identity value obtained demonstrate that the new set of loci displays suitability for paternity studies on electric eels. In addition, the cross-amplification of electric eel species implies that it may also be useful in the study of the closely related E. electricus, and to other Neotropical electric fishes (Gymnotiformes) species as tested herein.


INTRODUCTION
Electric eels (Electrophorus Gill, 1864) share with other species of Neotropical electric fishes (Gymnotiformes) a specialized electrogenic-electrosensory system used to navigate, and communication (Crampton, 2019). In addition to low-voltage electric organ discharges (EODs), electric eels generate high-voltage EODs for stunning prey and defense, as reported in the field by Humboldt in the 18th Century, and elegantly demonstrated in the laboratory by Catania (2019). For centuries, electric eels captivate minds, inspire scientific innovation, like the electric battery, which has been used as a model for understanding bioelectrogenesis (Finger, Piccolino, 2011;Gallant et al., 2014). Despite the broad public and scientific community interest, only recently species diversity on Electrophorus began to be explored in extent (de Santana et al., 2019). As a result, three electric eel species occurring in very distinct ecological environments were recognized: E. electricus (Linnaeus, 1766)  The new finds offer an opportunity to study the genetics of populations of those distinct ecological and unique animals by characterizing their genetic variation, within and between populations, and the forces that affect their frequencies, such as migration, mutation, selection, and genetic drift. An excellent way to study the genetic composition of natural fish populations is by using molecular markers, which are powerful tools for quantifying genetic variation in individuals and populations, contributing to the management and conservation of species (Allendorf et al., 2010). According to Zane et al. (2002), the microsatellites (SSR -Simple Sequence Repeats), for instance, are considered useful for population studies because they are highly polymorphic markers. The population genetic analysis of species in the wild is of paramount importance for elucidating the factors and conditions that allow populations and species to be maintained and in the development of a strategy for its effective management (Moysés et al., 2005).
Published population genetic studies in Neotropical electric fishes are inexistent, and only a few attempts to develop microsatellite primers for Gymnotiformes were made (e.g. Moysés et al., 2005).
This study aims to develop candidate microsatellite loci to accurately access genetic diversity and help in future studies of population genetics of electric eels. Thus, this paper reports the development and characterization of novel microsatellite loci for E. varii and evaluates it in E. voltai to cross-amplification. Additionally, the primers were tested for cross-amplification in four species across Gymnotiformes.

MATERIAL AND METHODS
A partial enriched genomic library was constructed, and microsatellites were isolated and characterized following the protocol of Billotte et al. (1999). Tissue samples from E. varii and E. voltai were donated by the Instituto Nacional de Pesquisas da Amazônia (INPA), with invoice number: 009/96. Total genomic DNA was extracted from muscle tissue from a sample of E. varii (INPA 41112), according to Almeida (Almeida et al., 2010). Genomic DNA (5 µg) was digested, and the blunt-ended fragments were ligated to the adaptors (Edwards et al., 1996). Fragments were selected, amplified, and cloned into pGem-T Easy (Promega; www.promega.com) vectors using 5μL of the amplification product, 50 ng of vector, and 1 U of T4 DNA ligase in reaction buffer at 4°C (overnight). Cloning products were used to transform Escherichia coli (DH5 -α lineage) cells. The recombinant clones were selected and sequenced on an ABI 3500 XL automated sequencer. Sequences were analyzed, and primers were designed according to Hall (1999) and Rozen, Skaletsky (2000), respectively. The selected forward primers were marked with the M13 at the 5' end (Schuelke, 2000). To test the potential presence of hairpin structures and problems with the primer-dimer, we follow the protocol of Vallone, Butler (2004). PCR amplifications were carried out on a panel consisting of 13 individuals of E. varii (INPA 41112 -41122 and INPA 41124 -41125) from three localities along the Curiaú River; and 14 individuals of E. voltai (LIA 4802 -4806; INPA 41123; INPA 050453) -five from two localities of the Xingu River, one specimen collected in the Curiaú River and eight collected in the Iriri River. All specimens of electric eels were collected in the Amazon basin, Brazil. Cross-amplification tests were performed using four other Gymnotiformes species whose voucher specimens are deposited in the Apolinário-Silva et al. (2018). Amplifications were made with an initial denaturation step at 94ºC for 4 min, followed by 35 cycles at 94ºC for 40 s, 48ºC, 54ºC, or 60ºC (Tabs. 1-2) for 1 min, 72ºC for 1 min, and a final extension at 72ºC for 30 min. The PCR products were submitted to electrophoresis on an automated sequencer. GeneScan 600 Liz (Applied Biosystems) was used as the molecular weight standard. Individuals were genotyped with GeneMarker 1.85 (SoftGenetics, State College, PA), followed by manually editing. Tests for Hardy-Weinberg Equilibrium (HWE) and the presence of linkage disequilibrium among the pairs of loci were calculated using GENEPOP 4.0.10; P values were subsequently adjusted applying the sequential Bonferroni correction (Rice, 1989). GenAlEx v.6.41 was used to estimate the observed (H o ) and expected (H e ) heterozygosities and the average number of alleles per locus. The paternity exclusion probability (Q) (Weir, 1996) and genetic identity probabilities (I) (Paetkau et al., 1995) were estimated using Identity 1.0. Estimates of the polymorphic information content (PIC) and potential null alleles were obtained through Cervus v.3.0 and Micro-Checker v.2.2.3, respectively. Default settings were used for all tests.

RESULTS
A set of 13 polymorphic and highly informative microsatellite loci for genetic studies of populations of Electrophorus were developed: a total of 45 out of 96 clones sequenced contained microsatellite regions, with 25 being suitable for primer design and PCR reactions. After testing different amplification conditions, 14 loci (almost all dinucleotide repeats) were successfully amplified. From those, one was monomorphic, and 13 were polymorphic for two electric eel species.
In E. varii, a total of 85 different alleles were detected, varied from 2 (Elec24) to 15 (Elec39), with an average of 6.4 alleles per locus. The observed and expected heterozygosity ranged from 0.000 (Elec24) to 1.000 (Elec14) and from 0.334 (Elec49) to 0.902 (Elec39), respectively. After sequential Bonferroni correction for multiple comparisons (α = 0.05, k = 91), no evidence of linkage disequilibrium between any pair of loci examined was observed. In the HWE tests, two loci, Elec24 and Elec241, presented significant deviation after correction for multiple tests (sequential Bonferroni correction α = 0.05 and k = 14). These loci were also the only ones showing possible null alleles, inferred from excess homozygous genotypes, explaining the observed deviation from HWE. It was observed that the same loci that had a significant deviation in the HWE, plus loci Elec22 and Elec31, also had significant values of the endogamic coefficient (F IS ; Tab. 1). The mean PIC for the 13 polymorphic loci was 0.572 following a scale proposed by Botstein et al. (1980), 10 loci (Elec12, Elec14, Elec 21, Elec31, Elec39, Elec43, Elec53, Elec241, Elec246 and Elec247) were highly informative and three loci (Elec22, Elec24 and Elec49) were moderately informative. The probabilities of identity and paternity exclusion were equal to 2.665 -12 and 0.999, in that order (Tab. 1).
All 14 microsatellite primers developed for E. varii were successfully cross-amplified in E. voltai. Thirteen are polymorphic loci and produced a total of 74 different alleles, with allele number ranging from 2 (Elec31 and Elec451) to 12 (Elec49), with an average of 5.2 alleles per locus (Tab. 2). The observed and expected heterozygosity varied from 0.071 (Elec12, Ele24 and Elec451) to 1.000 (Elec14) and from 0.069 (Elec451) to 0.908 (Elec49), correspondingly. After Bonferroni sequential correction for multiple comparisons (α = 0.05, k = 91), no evidence of linkage disequilibrium between any pair of loci examined was detected.

DISCUSSION
Deviations of the HWE and significant F IS values for some loci, mainly in E. voltai, are likely to be caused by the mixture of individuals originating from different populations. Freeland (2005) suggested that the inclusion of elements of multiple genetic units in a single panel could cause the Wahlund effect, i.e., excess homozygosity and significant estimations of F IS . Similar results were observed by Apolinário-Silva et al. (2018), which   et al., 2019) as well as in other Gymnotiformes species not tested herein. Heterologous primers can be successfully used in different species of fishes, and the quality of amplification depends on the degree of genetic conservation of positions bordering microsatellite regions (Abdul-Muneer, 2014). Consequently, the low amplification rate primers in the four species of Gymnotiformes can be explained by the lack of conservation of microsatellite sites. Equally, the successful amplification described in E. voltai can be attributed to the elevate conservation of the microsatellite flanking regions, which according to Barbará et al. (2007), is expected among closely related species. Accordingly, the lowest cross-amplification found for Gymnotus, currently hypothesized as the putative sister taxon to Electrophorus (Alda et al., 2018), was unexpected (see discussion on Electrophorus interrelationships in de Santana et al., 2019), indicating that Electrophorus current hypothesis of interrelationships deserves further attention.