Evaluation of polymorphism in ten microsatellite loci in Uruguayan sheep flocks

The allele frequencies of 10 microsatellite loci previously described for sheep as BM1314, BM6526, OarFCB128, OarHH64, OarCP20, OarHH47, OarFCB48, OarHH35, OarHH72 and BM2508 were estimated for the Uruguayan flocks. A representative sample of 101 individuals composed by the two predominant breeds (76% Corriedale and 24% Australian Merino) was used. The sample did not show a significant tendency towards substructuring, in spite of presenting some significantly different allele frequencies between races. The Corriedale sample presents three loci in which the presence of null alleles is possible. The markers were highly variable, showing between 7 and 15 alleles each. The Polymorphism Information Content Index ranged from 0.63 to 0.87 and the Exclusion Probability from 0.39 to 0.75 for a cumulative Exclusion Probability of 99.98%. These results suggest the effectiveness of this set of loci for testing genetic relatedness. This is the first report of microsatellite variation in Corriedale.


Introduction
For every program of breeding and selection, it is desirable to rely on a precise assessment of parentage to assist, for example, in the elimination of undesirable recessive alleles, in early selection for future breeding, or to ensure accurate pedigrees and registration.Microsatellites, segments of the nuclear genome composed of tandem repeat of short-sequence motifs, have become excellent candidate markers for this kind of studies (Queller et al., 1993), since they are numerous, highly variable and easy to score.
A large number of highly polymorphic microsatellites have been characterized and mapped in domestic animals, including sheep, cattle and other ruminants (de Gortari et al., 1997(de Gortari et al., , 1998;;Hayes et al., 1996;Jenkings et al., 1997), facilitating the use of these markers in parentage testing.For this kind of test to be acceptable, it must provide a high degree of certainty in the assignment of parentage, which can be evaluated with the Exclusion Probability (Weir, 1996).This requires a previous knowledge of allele frequencies in the population of interest and certainty about the assumptions of the model, like independence of the markers utilized and Hardy-Weinberg equilibrium.
The goals of this study were to analyze the polymorphism of a set of microsatellites previously described as polymorphic in sheep and to evaluate their usefulness for relatedness testing in Uruguayan sheep flocks.

Material and Methods
We analyzed ten microsatellite loci in a sample of the Uruguayan flocks considering only pure-breed individuals.The analysis were carried out separately for two subsamples: one of 77 Corriedale individuals from a fair, and another of 24 Australian Merinos from an experimental field (all of them are expected to be unrelated ewes), as well as on the combined dataset.Corriedale, the predominant race in the country, has not been previously characterized using microsatellite markers.
Total DNA extractions were made with SDS/proteinase K/NaCl/alcohol precipitation (modified from Miller et al., 1988) from skin preserved in 95% ethyl-alcohol.The microsatellite loci OarHH35, OarHH64, OarHH47 (Henry et al., 1993), OarVH72 (Pierson et al., 1993), OarFCB128 (Buchanan and Crawford 1993), OarFCB48 (Buchanan et al., 1994), OarCP20, BM1314, BM6526, (Crawford et al., 1995), and BM2508 (Mulsant et al., 1998) were amplified by PCR (Polymerase Chain Reaction) for all animals; excluding OarHH35 and OarHH64 all of them map on different chromosomes.Amplification was carried out in a total volume of 12.5 µL containing the following constituents: 6.25 µL of DNA (0.4 µg/mL) used as a template, 1X Taq Polymerase Buffer, 240 µM of each dNTP, 240 nM of each primer, 0.5 units of Taq Polymerase and variable concentration of MgCl 2 depending on the locus (from 1.5 to 3.0 mM).PCR amplifications were performed in a Rapidcycler TM (Idaho Technology) by an initial denaturation of 3 min at 94 °C, followed by 35 cycles of 45 s of denaturation at 94 °C, 45 s of annealing at 50 °C and 45 s of extension at 72 °C, and a final extension of 7 min at 72 °C.
The amplified products were electrophoresed in 7%, 0.80 mm thick denaturing polyacrylamide gels (1000-1200 V, 1-2 h), and the DNA bands were visualized by silver staining (Sanguinetti et al., 1994).Genotypes were scored by at least two persons independently and the absolute sizes of the alleles were determined in relation to a 10bp DNA size standard (GIBCO BRL).
On the basis of allele and genotype frequencies, Hardy-Weinberg equilibrium and linkage disequilibrium were tested using the software GenePop, version 3.1b (Raymond and Rousset, 1995).A Bonferroni correction was used to adjust significance levels in all tests to an overall α = 0.05.Observed heterozygosity (H obs ), Polymorphism Information Content (PIC, Botstein et al., 1980) and Exclusion Probability assuming that the genotype of the lamb and one of its parents is known (Q, according to Weir, 1996) were calculated.
The software Structure (Pritchard et al., 2000) was used in order to assess levels of subdivision between two races (10,000 Burn-in period, 10,000 MCMC repetitions, and default parameters).

Results
Allele length in base pairs, allele frequencies, H obs , PIC, Q and the number of individuals analyzed per locus are given in Table I.All loci were highly polymorphic, showing between 7 and 15 alleles, PICs between 0.63 and 0.87 and H obs in the range of 0.62 to 0.84 for the whole sample.The two loci that map on the same chromosome (OarHH35 and OarHH64) did not show a significant linkage (p > 0.373, SE = 0.052).
The loci OarHH35, OarHH64 and OarFCB128 showed significant departures from Hardy-Weinberg equilibrium in the whole population and in the Corriedale sample (Table II).With the exception of OarFCB128, OarVH72 and BM6526, all loci presented significant differences in allele frequencies between Merino and Corriedale samples.Some alleles were much more frequent in one race than in the other (e.g., allele 123 in OarHH35 and 146 in OarHH47, which are 14 and 8 times more frequent, respectively in Merino than in Corriedale, or allele 75 in 38 Tomasco et al.OarCP20 which is seven times more frequent in Corriedale than in Merino).Other alleles were found only in one sample, and sometimes at high frequency (e.g., alleles 134 and 142 of OarHH47 in Corriedale, and allele 85 of OarCP20 in Merino).In several cases, these frequencies differed from those reported in the literature (see Henry et al., 1993, for OarHH47, OarHH35, OarHH64;Buchanan and Crawford 1993, for OarFCB128;Ede et al., 1995, for OarCP20;and Buchanan et al., 1994, for OarFCB48).There were also differences in one base pair between absolute allele sizes reported for OarFCB48 and OarFCB128 and our observations but it was easy to resolve the correspondences on the basis of allele frequencies.Direct comparisons would be needed to assess the reason for this discrepancy.Analyses of subdivision using the software Structure (Pritchard et al., 2000) did not find a significant tendency towards substructuring in the sample (data not shown), in spite of the presence of two breeds.

Discussion
The loci used in the analysis are effectively independent, since the only two that map on the same chromosome (OarHH35 and OarHH64) did not show significant linkage disequilibrium.The departures from Hardy-Weinberg equilibrium shown by OarHH35, OarHH64 and OarFCB128 in the Corriedale sample were probably the cause of the departure in the whole population.This could be due to the presence of null alleles in these loci in Corriedale or be the consequence of several years of intensive selection on these breeds (Usha et al., 1995).To avoid this problem, these loci can be excluded from parentage analysis and the cumulative Q would vary from 99.99% to 99.77% and from 99.98% to 99.75% for the whole population and the Corriedale stock, respectively.For Merinos, the cumulative Q with ten loci was 99.93%.
In spite of the presence of several differences in allele frequencies between races, these were insufficient to result in substantial separation between these breeds.These two races are known to be related, since Corriedale was originated from crosses between pure Merino and Lincoln sheep (Majala, 1997), so this result is not entirely surprising.Coupled with the fact that most of Uruguayan sheep are not completely pure, these observations suggest that allele 40 Tomasco et al. frequency estimations derived from the whole sample may be sufficient for future applications in relatedness tests.Also, such tests may exclude the loci suspected to have null alleles without a marked reduction in the probability of exclusion.
The fact that estimated frequencies differed from those reported in the literature in most loci analyzed, stresses the importance of independent estimates for different regions and races.
In sum, these results indicate that the selected system of markers is highly effective for relatedness studies, with a Probability of Exclusion greater than 99.9% considering all loci.The significance of this report, the first of this kind in the country, is that it offers interesting perspectives for the incorporation of molecular genetic techniques to animal breeding in Uruguay.In addition, our results represent an original contribution of allele frequencies for regional Corriedale and Australian Merino races.

Table I
-Loci, absolute allele sizes in base pairs, allele frequencies, observed heterozygosity (H 0bs ), polymorphism information content (PIC), exclusion probability (Q) and number of individuals analyzed per locus and sample.The whole population combines all the Merinos and Corriedales; p-values refer to the comparison of allele frequency distributions between Merino and Corriedale samples: *and NS means significant and non-significant difference, respectively.Markers are followed by ♦ if their absolute allele sizes are not concordant with those reported in the literature.

Table II -
Observed p-values and standard errors (SE) of Hardy-Weinberg exact tests for heterozygote deficit.