RAPD analysis of Nectomys squamipes ( Rodentia , Sigmodontinae ) populations

Random amplified of polymorphic DNA (RAPD) analysis was used to assess genetic distance and the genetic structure of populations of Nectomys squamipes, a semiaquatic rodent species distributed along watercourses. DNA samples of five populations were analyzed using three primers, producing 45 scorable bands, 31 of which were polymorphic. There was a significant differentiation among populations [FST = 0.17; ΦST = 0.14 (P < 0.004)] but gene flow (Nm = 1.25) was sufficient to overcome genetic drift effects. No fixed specific markers were found for any population. The Mantels test and UPGMA cluster analysis showed a lack of relationship between genetic and geographic distances. The apparent homogeneity indicated by RAPD markers coincided with morphometric data, despite the wide geographic range of N. squamipes. Alternative hypotheses for explaining our results include recurrent processes of local extinction and recolonization or a recent and sudden increase in the geographic distribution of this species. 1Divisão de Genética, Coordenação de Pesquisa, Instituto Nacional de Câncer, Rio de Janeiro, RJ, Brasil. 2Departamento de Ecologia, Laboratório de Vertebrados, Universidade Federal do Rio de Janeiro, Caixa Postal 68020, 21491-590 Rio de Janeiro, RJ, Brasil. Send correspondence to F.C.A. Fax: +55-21-290-3008. E-mail: falmeida@biologia.ufrj.br


INTRODUCTION
Water rats of the genus Nectomys (Rodentia, Sigmodontinae) comprise 11 species that can be distinguished from one another by their morphometric and karyotypic attributes (Bonvicino, 1994).This genus is widely distributed in the South American continent, with all species inhabiting watercourse banks (Ernest and Mares, 1986).Nectomys squamipes is a medium-sized rodent, with a body weight ranging from 160 to 400 g, distributed along the São Francisco and Paraná River basins and small independent basins in eastern Brazil (Musser and Carleton, 1993).Karyotypic analysis indicated that N. squamipes has a basic diploid number (2n) of 56 chromosomes with an autosomic fundamental number (FN) of 56 (Maia et al., 1984;Yonenaga-Yassuda et al., 1987;Bonvicino et al., 1996).
Nectomys squamipes is a natural host of Schistosoma mansoni, with high prevalence in endemic areas, and is an important experimental model for schistosomiasis (Rey, 1993;D'Andrea et al., 1996).Reproduction of N. squamipes occurs throughout the year, peaking during the rainy season, with a gestation period of approximately 30 days, and a most frequent litter size of five (ranging from two to seven; D' Andrea et al., 1996).This species is nocturnal, with an omnivorous diet that includes leaves, fruits, fungi, invertebrates and small vertebrates (D'Andrea et al., 1996).Field studies reported that the home range of N. squamipes is small, with a maximum recapture distance of 90.4 + 55.3 m for males and 42.8 + 28.9 m for females.A preliminary report on the geographic variation of N. squamipes showed limited morphometric differentiation among populations (Bonvicino, 1994).
Studies on genetic variability within and between populations can provide insights on speciation and a better understanding of the evolutionary trends.This is important for assessing prospective evolutionary potentials as well as for risk assessments and conservation strategies.During the course of a survey on the geographic variation of N. squamipes, liver samples were obtained from specimens of five different localities.We used random amplified of polymorphic DNA (RAPD; Williams et al., 1990;Welsh and McClelland, 1990) for analyzing the genetic structure of these populations and the genetic variability within and among them.We chose RAPD because it is expeditious and inexpensive and because it does not require a previous knowledge of the genome of species under study (Lynch and Milligan, 1994).Earlier comparisons showed that RAPD and isoenzyme results are very similar, with the RAPD technique revealing even higher amounts of variation (Baruffi et al., 1995;Sinclair et al., 1996).

DNA extraction
DNA from ethanol-preserved liver samples was extracted as described by Smith et al. (1987) while extraction from samples stored in liquid nitrogen and at -70°C followed the procedure of Sambrook et al. (1989).DNA was stored in TE buffer (10 mM Tris, 1 mM EDTA, pH 8.0) at 4°C.DNA samples were run in 0.8% agarose gels stained with ethidium bromide to assure that degradation had not occurred.DNA concentrations were estimated by comparisons with a standard sample in an agarose gel.

RAPD-PCR
PCR reactions were carried out with 200 ng of template DNA in 67 mM Tris, pH 8.8, 6.7 mM MgSO 4 , 16.6 mM (NH 4 ) 2 SO 4 , 10 mM β-mercaptoethanol, 1 mM of each dNTP, 250 ng of primer and 1 unit of Taq DNA polymerase in a final volume of 50 µl.Amplifications were carried out in a DNA Thermal Cycler (Perkin-Elmer 480) under the following conditions: one initial cycle of 3' at 95°C, 5' at 35°C, and 5' at 72°C followed by 43 cycles of 1' at 95°C, 1' at 35°C and 2' at 72°C, with a final extension period of 15' at 72°C.Tubes containing all reaction components except for template DNA were included as controls.Amplification products were observed in silver-stained 5% polyacrylamide gels following vertical electrophoresis.
Eight primers of different size were tested and those resulting in visible, reproducible and easily scorable bands were selected.Mendelian patterns of inheritance were tested in a family comprising both progenitors and two male offspring from the Sumidouro population.Patterns of inheritance were deduced by comparing fragment profiles and by careful verification that any band present in the offspring was at least shared with one progenitor (Stott et al., 1997).

Data analysis
Bands were scored on dried gels and analyzed twice by two different observers.One randomly chosen specimen was used as control for band migration and staining quality in all gels.Ambiguous and extremely light bands were disconsidered.For every individual, bands were scored as present (1) or absent (0) in a matrix for estimating genetic dissimilarity between individuals of each locality.The dissimilarity index (D) was estimated by D = NAB/(NA + NB), where NA and NB are the numbers of fragments of individuals A and B, respectively, and NAB is the number of fragments that are not shared between these individuals (Gilbert et al., 1990).Intrapopulational dissimilarity was estimated as the average D value of all pairwise comparisons within each population.Estimates of the mean standard error were calculated as in Gilbert et al. (1990).
Genotypic analysis based on allele frequencies (Lynch and Milligan, 1994) was carried out for assessing genetic variability among populations and the degree of subdivi- sion of genetic diversity among populations (F ST ; Wright, 1951) was calculated.Gene flow was estimated by Wright's (1951) formula, Nm = 0.25(1/F ST -1), where N is population size, and m is the immigration rate among populations, assuming equilibrium between migration and drift in the "island model".Subdivision of genetic variation was also assessed by analysis of molecular variance (AMOVA; Excoffier et al., 1992) using D values from pairwise comparisons.Geographically close populations (from Glicério, Sumidouro and Fazenda União) were grouped and compared with those from other localities in order to assess intergroup variance.AMOVA analyses were performed using the WINAMOVA 1.55 program (Excoffier et al., 1992).
The genetic distance between any two populations was estimated by averaging D values between pairwise comparisons from individuals of these populations (Gilbert et al., 1990).Genetic distances were used in UPGMA cluster analysis with the MEGA program (Kumar et al., 1993).The relationship between genetic and geographic linear distance was analyzed by a Mantel's test with the NTSYS-PC program (Rohlf, 1993).

RESULTS
PCR amplifications resulted in visible products with four of the eight primers tested.When analyzing amplified products from two separate, albeit identical, reactions on contiguous lanes, these four primers produced reproducible results (Figure 2).Moreover, these primers amplified fragments whose pattern of inheritance was clearly Mendelian, although one of these primers (ROJS2) was discarded because amplified fragments produced faint bands.
One hundred and ten different bands were scored in an analysis of 55 individuals with the three selected primers, but only 45 different bands were used for analyses.Assuming that these 45 bands represented different loci, 31 of them were found to be polymorphic.For each primer, the number of scored, useful and polymorphic bands was determined (Table I).
D estimates of pairwise comparisons within localities were used for constructing UPGMA clusters to analyze inter-individual relationships within populations, among trapping transects.None of the localities where trapping occurred in more than one transect showed transectspecific clusters, suggesting that each sample locality could be considered a single population sharing the same genetic pool.Estimates of the intrapopulational dissimilarity index were very similar among populations (Table II).
There were no population-specific alleles.The estimated F ST value was 0.17, indicating that approximately 17% of genetic variability was due to differences between populations.The Nm value, equivalent to the number of immigrants exchanged per generation, was 1.25.AMOVA indicated a significant level of differentiation, with an estimated Φ ST value among populations of 0.14 (P < 0.0004).This indicated that 86% of total variation occurred within  populations while only 14% was due to population subdivision.When we grouped the three geographically close populations (Sumidouro, Glicério and Fazenda União) and compared them to the two distant ones (Fazenda Canoas and Pedreira), the variance subdivision estimate among groups (Φ CT ) was not significant (P = 0.57).This indicated that the three geographically closest populations showed the same genetic differentiation as between the more distant ones.Distance estimates were made between populations (or interpopulation dissimilarity index values; Table III).The relationship between populations is represented in the UPGMA topology (Figure 3).This confirmed Mantel's test estimates in showing a lack of significant correlation between genetic and linear geographic distance (r = -0.228;t = -0.588;P = 0.278), which indicates that there was no isolation by distance in N. squamipes.

DISCUSSION
The RAPD procedures proved to be a useful tool for assessing genetic variability because band profiles with the three selected primers were reproducible and their patterns of inheritance proved to be Mendelian for a dominant marker.Despite the high proportion of excluded bands a substantial number of polymorphic markers was detected.
Few studies of terrestrial mammals using the RAPD technique have been reported and different similarity and dissimilarity indexes have been used for RAPD analysis.The lack of standard procedures for RAPD analysis makes comparisons with other studies very difficult.Baruffi et al. (1995) using the same D index herein used, found D values ranging from 0.158 to 0.387 in six wild population samples of fruit flies (Ceratitis capitata), each consisting of five individuals.Conversely, our within-population D values were very similar to one another, ranging between 0.118 and 0.128, despite differences in sample size.Sinclair et al. (1996) found a smaller F ST value (0.028) between flying-fox (Pteropus scapulatus) populations than the ones we found (F ST = 0.17, Φ ST = 0.14), while studies with bird species found similar values (F ST = 0.17, Haig et al., 1996; Φ ST = 0.20, Nusser et al., 1996).This is interesting because birds have a higher dispersion ability, with potentially higher gene flow between populations.
In the N. squamipes populations the F ST value corresponds to an Nm of 1.25, which is above the minimum number of migrants per generation needed to avoid differentiation by genetic drift (Nm > 1, Slatkin, 1987).Thus, although differentiation estimates were significant, the level of differentiation was not very high.This was coincident with the lack of population-specific alleles and with previous data showing absence of major morphometric differentiation in N. squamipes across a very extensive geographic distribution (Bonvicino, 1994).
The lack of substantial differentiation contrasts with the dispersion pattern in N. squamipes, which occurs along watercourses with a discontinuous distribution along hydrographic basins (Ernest and Mares, 1986).Based on the dispersion data, it was expected that genetic drift would result in a higher differentiation between populations than that found by us.Alternative hypotheses which might explain our results include recurrent processes of local extinction and recolonization (Maruyama and Kimura, 1980;Slatkin, 1987) or recent, sudden increases in the geographic distribution of this species (Slatkin, 1993).

Figure 2 -
Figure 2 -Silver-stained polyacrylamide gel with RAPD products for the control family, obtained using the primers: ψ1, ψ2 and universal.S 1 and S 2 = Offspring, F = father, M = mother, L = molecular size standard 100-bp DNA ladder.

Table II -
Intrapopulational genetic dissimilarity indexes (D), standard error of the mean (SEM), and expected heterozygosity (with standard error, SE) for each Nectomys squamipes population.

Table I -
Primers used and the number of scored, useful and polymorphic bands obtained in the RAPD analysis.

Table III -
Dissimilarity index between pairs of populations (D, in the lower-left matrix) and standard error of the mean (SEM, in the upper-right matrix).For abbreviations, see legend to Figure1.