Development and characterization of microsatellite loci of Microglanis cottoides ( Siluriformes : Pseudopimelodidae ) and cross-species amplification

Thirteen microsatellite loci were isolated and characterized in Microglanis cottoides. Of these, two were monomorphic and 11 were polymorphic. These polymorphic loci tested on 24 individuals from a wild population produced a total of 108 different alleles, with levels of variability high, ranging from 2 to 20, with an average of 8.3 alleles per locus. The observed and expected heterozygosity ranged from 0.125 to 0.958 and from 0.119 to 0.931, respectively. A high combined probability of paternity exclusion value and a low probability combined genetic identity value obtained show that the set of loci described herein displays good suitability for paternity studies and differentiation of M. cottoides. Additionally, all thirteen microsatellite primers developed for M. cottoides were tested in four other Pseudopimelodidae species and successful cross-species amplification was achieved for the majority of loci.

With 21 valid species described to date (Ruiz & Shibatta, 2011), the genus Microglanis is also characterized by its wide mouth (gape width same as head width), short maxillary barbel (occasionally reaching pectoral-fin origin) and small eye without free orbital margin, among other particular characters (Shibatta, 2003a(Shibatta, , 2003b)).These small catfishes are widely distributed across South America, occurring from Guyana to the Rio de La Plata basin in Argentina (Shibatta & Benine, 2005;Sarmento-Soares et al., 2006).
Although our knowledge of the number of Microglanis species has increased in the last decade (Shibatta, 2003a(Shibatta, , 2003b;;Bertaco & Cardoso, 2005;Shibatta & Benine, 2005;Sarmento-Soares et al., 2006;Alcaraz et al., 2008;Ottoni et al., 2010;Shibatta & Ruiz, 2010;Shibatta & Ruiz, 2011), the biology of Microglanis still remains poorly known (Shibatta, 2003a).Recently, it has been shown that some species can occur in small populations patchily distributed along the river basins where they are found (Esguícero & Arcifa, 2010).According to these authors, such conditions which may be associated with the progressive environmental degradation of freshwater habitats could represent threats to populations of some Microglanis species.Despite this fact, up to this moment population genetic studies involving Microglanis species are still lacking.
Among the known species, Microglanis cottoides (Boulenger 1891) is especially interesting for phylogeographic and genetic population studies, owing to its geographical distribution in the Brazilian Atlantic Forest, being present in the Uruguay River basin and along the eastern coast of Brazil, with records of their occurrence in drainages across the southern and southeastern regions (Malabarba & Mahler, 1998;Shibatta, 2003bShibatta, , 2007;;Mori & Shibatta, 2006).These regions can be considered of great biogeographical significance, not only for the high degree of endemism of its fish fauna (Bizerril, 1994(Bizerril, , 1995)), but also for being a very populated area with high environmental degradation due to human activities.
Studies on geographic distribution of genealogical lineages have been widely used to describe historical events, such as habitat fragmentation or expansion of the range of species and populations.These studies also serve to understand migration events, vicariance and extinction of gene lineages, and other processes that affect the population structure or generate speciation in a spatial and temporal context (Avise, 2001(Avise, , 2009;;Hardy et al., 2002).
A good way to study the genetic structure of natural fish populations is through the use of molecular markers such as microsatellite (SSR-Simple Sequence Repeats) (O'Connell & Wright, 1997).In fact, SSR are considered a class of molecular markers highly polymorphic and useful for population studies (Zane et al., 2002).Therefore, the current work describes 13 SSR markers developed for Microglanis cottoides and tests their applicability for four other Pseudopimelodidae species.

Material and Methods
Microsatellites were isolated and characterized following the protocol of Billotte et al. (1999), with some minor modifications.Total genomic DNA was extracted from muscle tissue preserved in 95% ethanol.The DNA was isolated according to Almeida et al. (2001).Genomic DNA (5 μg) was digested with RsaI and the blunt-ended fragments were ligated to the adaptors Rsa 21 and Rsa 25 (Edwards et al., 1996).Fragments were selected with (AGA) 5 , (CT) 8 and (GT) 8 probes, and then cloned into the pGem®-T (Promega) vector.Such recombinant plasmids were used to transform the E. coli DH5α lineage.The recombinant clones were selected and sequenced using the BigDye Terminator™ kit (v.3.1 -Applied Biosystems).Sequencing was performed on an ABI 3500 XL automated sequencer.Sequences were analyzed using BioEdit v.7.0 software (Hall, 1999) and primers were designed using Primer 3 software (Rozen & Skaletsky, 2000).The selected forward primers were labelled with the M13 sequence (5'-TGTAAAACGACGGCCAGT-3') at the 5' end (Schuelke, 2000).The AutoDimer software (Vallone & Butler, 2004) was used to test the potential presence of hairpin structures and primer-dimer problems.The individual genotyping was performed on an ABI 3500 XL automated sequencer.

Results
Of 96 clones sequenced, 25 contained microsatellite repeats but only 18 were suitable for primer design and PCR reactions.After testing the different amplification conditions, 13 loci were successfully amplified.From these, two were monomorphic and 11 were polymorphic (Table 1).
The eleven polymorphic loci produced a total of 108 different alleles, with high levels of variability, ranging from 2 to 20, with an average of 8.3 alleles per locus.The observed and expected heterozygosity ranged from 0.125 to 0.958 and from 0.119 to 0.931, respectively.After Bonferroni correction for multiple comparisons, no evidence of linkage disequilibrium between any pair of loci examined was observed.Only the locus Mcot09 showed significant deviation from Hardy-Weinberg equilibrium (P < 0.05).This locus showed the presence of null alleles, inferred from excess homozygous genotypes which could explain the observed deviation from HWE.Additionally, the locus Mcot08 revealed an excess of homozygotes and the possible presence of null alleles, but did not show any significant deviation from the HWE.It was also observed that the loci Mcot08 and Mcot12 were the only ones showing significant values of endogamic coefficient (F IS ) (Table 1).

Discussion
A high combined probability of paternity exclusion value (0.999) and a low combined probability of genetic identity value (2.16 x 10 -6 ) were obtained, showing that the set of loci described herein, exhibits good applicability for studies of parentage and population differentiation of M. cottoides.
The average PIC was 0.528, ranging from 0.115 to 0.927.According to Botstein et al. (1980), PIC values above 0.5 indicate highly informative markers; from 0.25 to 0.50, moderately informative; and below 0.25, slightly informative.Of the 13 loci obtained, eight proved to be highly informative (Mcot02,Mcot03,Mcot04,Mcot07,Mcot08,Mcot09,Mcot10 and Mcott12) with PIC values equal to or higher than 0.589; two were moderately informative ( Mcot05 and Mcot11) presenting PIC values equal to or higher than 0.337.Only one locus (Mcot13) was slightly informative, exhibiting a PIC value equal to 0.115.
All 13 microsatellite primers developed for M. cottoides were successfully cross-amplified in two from four Pseudopimelodidae species (Table 2).Only two loci (Mcot04 and Mcot09) failed to amplify in P. pulcher and M. garavelloi, respectively.From the set of loci tested, Microglanis garavelloi and M. parahybae exhibited 10 polymorphic loci, while M. cibelae and P. pulcher showed, respectively, 11 and 12 polymorphic loci each.
Table 1.Description of 13 microsatellite loci (11 polymorphic and two monomorphic) isolated from the Neotropical fish Microglanis cottoides.Flanking primers, T a = optimal annealing temperatures, k = number of alleles, allele size range (bp), H o = observed heterozigosity, H e = expected heterozygosity estimated from 30 individuals, Q = paternity exclusion probability, I = probability of genetic identity, F IS = endogamy coefficient, PIC = polymorphic information content, GenBank accession numbers.* Significant deviations from Hardy-Weinberg equilibrium (P < 0.05).** Significant value for the endogamy coefficient (F IS ).The successful cross-species amplification of microsatellite loci described herein can be attributed to the high conservatism of the flanking microsatellite regions, which is expected among close related species, as reported elsewhere (Barbará et al., 2007).Thus, the set of primers presented in the current study appear as promising tools for future population genetic studies involving these five species of Neotropical fish.

Table 2 .
Cross-amplification of 13 loci in four species of Pseudopimelodidae.Amplifications and polymorphisms were tested in five individuals of each species.-indicates no amplification.k = the number of alleles per locus.