AFLP diversity between the Novosibirsk and Tomsk chromosome races of the common shrew ( Sorex araneus )

Genetic diversity between of the Novosibirsk and Tomsk chromosome races of the common shrew (Sorex araneus) was analyzed using 39 polymorphic AFLP (amplifi ed fragments length polymorphism) markers. Exact and F-statistics tests for population differentiation demonstrated signifi cant interracial difference in allele frequencies and signifi cant subdivision between the races. The value of the genetic distance between the chromosome races observed in this study corresponds to that found between subspecies of mammals studied so far.


INTRODUCTION
The common shrew Sorex araneus Linnaeus, 1758 inhabits a huge territory including the whole north part of Europe and Asia up to Baikal Lake (Corbet, 1978) and displays a remarkable chromosomal polymorphism determined by local fi xation of various Robertsonian and whole-arm reciprocal translocations (Searle, Wójcik, 1998).The species area is subdivided into a mosaic of parapatric chromosome races that has been classifi ed into four main karyotypic groups: the West European (WEKG), the East European (EEKG), the North European (NEKG) and the Siberian (SKG) (Searle, 1984;Wójcik, 1993;Searle, Wójcik, 1998;Polyakov et al., 2001).Whereas no clear morphological divergence has been found between races within a karyotypic group (Hausser, 1984;Searle, Thorpe, 1987;Hausser et al., 1991;Meyer, Searle, 1994), signifi cant morphological differences have been detected between the shrews belonging to different karyotypic groups: WEKG and EEKG in Poland (Chętnicki et al., 1996) and EEKG and SKG in Russia (Pavlinov, 2004).

Radjabli
It was known before the use of karyological data that shrews of Siberian highlands differed morphologically from other Siberian and Uralian shrew samples (Dolgov, 1968(Dolgov, , 1985)).Yudin (1989), combining these earlier observations and his own analysis with pioneering descriptions of Siberian chromosome races by Aniskin and Volobouev (1981), fi rst assumed that the morphological distinction between populations of shrews could be connected with their karyotypic divergence.Emphasizing the pronounced morphological and karyotypic individuality of shrews from Altai Yudin (1989) considered them as a subspecies Sorex araneus rypheus in distinction to the nominal subspecies Sorex araneus araneus.
Subsequent investigations confi rmed that namely the shrews of the Tomsk chromosome race were quite distinct morphologically from the shrews of neighboring Novosibirsk race (Polyakov et al., 2002) as well as from some other EEKG races (Okulova et al., 2007).
In this study we estimated genetic diversity between samples from populations of the Novosibirsk and the Tomsk chromosome races of the common shrew using polymorphic AFLP (amplifi ed fragments length polymorphism) markers.This method can generate large numbers of molecular markers without any previous knowledge of the genetic constitution of the genotypes under investigation (Vos et al., 1995).Because of its universal applicability, reproducibility and a high power of discrimination this approach has been used successfully in numerous phylogenetic and ecological studies (see review of Bonin at al., 2007).This is the fi rst time this method is used to analyze populations of the common shrew.

AFLP procedure
We followed Vos et al. (1995) in the AFLP procedure.It was based on the selective PCR of restriction fragments from a total digest of genomic DNA by EcoRI and MseI endonucleases.Two combinations of three randomly chosen primers complementary to EcoRI and MseI adapters with three extra-nucleotide each (E-ACT/M-CCG and E-ACT/M-CGT) were used for the analytic amplifi cation.Amplifi ed fragments were resolved on 6% denaturing polyacrylamide gels and silver stained (Creste et al., 2001).Gels were then dried and scanned.Electrophoretic bands were scored with Gel-Pro Analyzer system version 3.0.00.00

Data scoring and statistical analysis
Only precisely and evenly expressed electrophoretic bands were chosen for analysis.Length of visible fragments of DNA on gels ranged between 60 and 800 nucleotides, however the area of fragments longer than 500 nucleotides was not clear enough.For this reason they were excluded from the analysis.We regarded each band as a locus with two alternative alleles: present (1) or absent (0).The identifi cation of 39 polymorphic bands (=loci) led to the construction of a 47 individuals x 39 loci data matrix, which was analyzed for diversity within and between races.The genetic data was analyzed using the Tools for Population Genetic analysis (TFPGA) program ver-

Comp. Cytogenet., 2009 3(2)
Comparative Cytogenetics sion 1.3 (Miller, 1997) for dominant markers in diploid organism, assuming Hardy-Weinberg equilibrium.Both band-based and allele frequency-based approaches were applied to extract the statistical information from AFLP data (Bonin at al., 2007).Exact test for population differentiation (Raymond, Rousset, 1995) was applied to determine signifi cant differences in the bands frequencies between races.Genetic distance D NL (Nei, Li, 1979) was calculated from the bands frequencies using the phylogenetic package PHYLIP 3.6 (RESTDIST and NEIGHBOR subroutines) (Felsenstein, 2005).Unbiased genetic distance D N (Nei, 1978) and gene diversity (Nei, 1973) were calculated on the base of allele frequencies.Frequency of the null allele was calculated as a square root of the frequency of null homozygotes (Weir, 1990).
To evaluate the substructure of the studied samples Weir and Cockerham's (1984) methods of calculating Wright's F-statistics were applied to the data.Two-level hierarchy (within and between races) was expressed using the terminology of Weir (1990) where θ correspond to unbiased Wright's F ST .For hierarchical data set two θ values were calculated: θ P measured differentiation within races and θ R estimated the differentiation between the races.Jackknifi ng and bootstrapping over loci (using 1000 replications) were generated to obtain variance estimates and confi dence intervals.

RESULTS
A summary of the genetic diversity estimated from the AFLP data is given in Table 1.Amount of polymorphic loci and the levels of gene diversity within two races were about similar with slight priority within the Tomsk race sample.Exact test however demonstrated the high signifi cance of the difference in bands frequencies between the races: χ 2 =189.6 (P< 0.001).Two-level hierarchy F-statistics also revealed the signifi cant differences among the intra and interracial substructures.Bootstrapping (1000 replications) confi dence intervals of within and between races jackknifi ng variances θ estimates (θ P =0.2237 ± 0.0366 and θ R =0.0315 ± 0.0441) did not overlap at the 95% confi dence level.
Two estimators of genetic distance between races gave the cognate data: D N = 0.0713 and D NL = 0.0983.

DISCUSSION
Summarizing the series of observations Nei (1975) designed the scale of relatedness between the genetic distance and taxonomic categories.By this scale the genetic distance between local populations of a species ranges from 0 to 0.02 and from 0.02 to 0.2 between subspecies.Genetic differentiation between the Novosibirsk and Tomsk races, estimated by D N and D NL was higher than that between local populations.Nearly the same values of genetic distance were published recently for four sibling species of the blind mole rat Spalax ehrenbergi: S. galili, S. golani, S. carmeli, and S. judaei (Polyakov et al., 2007).The latter result was also obtained with the use of AFLP method and is thus completely compatible with our data.Like the shrew chromosome races Novosibirsk and Tomsk these four species are distinguished by chromosome complements and form parapatric areas (Nevo et al., 2001).
Together with substantial genetic distance our analysis revealed statistically signifi cant difference between the studied races in contents of their DNA polymorphism (Exact test and F-statistics).Thus, the whole set of existing knowledge about the Tomsk chromosome race (morphological, karyotypic and molecular specifi city) indicate its taxonomical distinctiveness.This can be considered as evidence in favor of a subspecies status of the Tomsk chromosome race and its name Sorex araneus rypheus suggested by Yudin (1989).The shrews of other chromosome races of the Siberian Karyotypic Group may also belong to this subspecies.Analysis of Inter-SINE-PCR markers in a sample of shrews inhabiting at the right bank of the Yenisei river and apparently belonging to other than Tomsk chromosome race demonstrated a close clustering with the Tomsk race and was quite distinct from the races of the East European Karyotypic group (Bannikova et al., 2003).However, further studies and additional materials from Siberian populations are necessary to estimate genetic diversity within and between karyotypic groups and to test this suggestion.

Table 1 .
Nei's (1973)istics of the sample sizes, number and % (in parentheses) of polymorphic loci andNei's (1973)index of gene diversity for the Novosibirsk and the Tomsk chromosome races.