Genetica 119: 201–218, 2003.
© 2003 Kluwer Academic Publishers. Printed in the Netherlands. 201
Genetic variation of microsatellite loci in the major histocompatibility
complex (MHC) region in the southeast Asian house mouse
(Mus musculus castaneus)
Shiao-Wei Huang & Hon-Tsen Yu
∗Department of Zoology, National Taiwan University, Taipei 106, Taiwan, ROC; ∗Author for correspondence (Phone:+886-2-23630231, ext. 2127; Fax: +886-2-23636837; E-mail: ayu@ccms.ntu.edu.tw)
Received 27 February 2001 Accepted 15 April 2003
Key words: balancing selection, major histocompatibility complex, microsatellites, M. m. castaneus, Mus musculus, polymorphism, wild mouse
Abstract
Major histocompatibility complex (MHC) genes are the most polymorphic loci known for vertebrates. Here we employed five microsatellite loci closely linked to the MHC region in an attempt to study the amount of genetic variation in 19 populations of the southeast Asian house mouse (Mus musculus castaneus) in Taiwan. The overall polymorphism at the five loci was high ( ¯He = 0.713), and the level of polymorphism varied from locus to locus.
Furthermore, in order to investigate if selection is operating on MHC genes in natural mouse populations, we compared the extent and pattern of genetic variation for the MHC-linked microsatellite loci (the MHC loci) with those for the microsatellite loci located outside the MHC region (the non-MHC loci). The number of alleles and the logarithm of variance in repeat number were significantly higher for the MHC loci than for the non-MHC loci, presumably reflecting linkage to a locus under balancing selection. Although three statistical tests used do not provide support for selection, their lack of support may be due to low statistical power of the tests, to weakness of selection, or to a profound effect of genetic drift reducing the signature of balancing selection. Our results also suggested that the populations in the central and the southwestern regions of Taiwan might be one part of a metapopulation structure.
Introduction
The major histocompatibility complex (MHC) is a multigene family encoding cell-surface glycoproteins that mediate both humoral and cell-mediated immune responses. The MHC region of the mouse spans a region containing 2000–4000 kb of DNA on chromo-some 17 and comprises genes grouped as class I, class II, and class III (Hood, Steinmetz & Malissen, 1983). Some of the class II as well as the class I genes are the most polymorphic genes known to date. The char-acteristics of some of these highly polymorphic MHC genes – long persistence of allelic lineages, prevalence of non-synonymous over synonymous substitutions in the peptide-binding region (PBR), and departures from the level of homozygosity expected under neutrality – indicate that some sort of selection is acting to
main-tain their extreme diversity (Nei & Hughes, 1991; Satta et al., 1994; Paterson, 1998).
The mechanism maintaining polymorphism at the MHC loci has been widely debated. Due to their critical role in immune response, it is generally assumed that the selective pressures affecting MHC diversity arise from infectious disease, but whether that selection takes the form of heterozygote advantage (over-dominance selection) or negative frequency-dependent selection, remains controversial (Takahata & Nei, 1990; Takahata, Satta & Klein, 1992; Potts & Wakeland, 1993). On the other hand, in addition to parasite-mediated selection, reproduc-tive selection through MHC-based mating preferences and/or selective abortion, has also been contributed to the evolution of MHC diversity in mice (Potts & Wakeland, 1993).
However, most surveys of natural population poly-morphism in the extensive MHC region so far have been conducted on humans (Klitz, Thomson & Baur, 1986), with a few limited studies on other mammals (e.g., bighorn sheep, Boyce et al., 1997; domestic sheep, Paterson, 1998), due to lack of a simple and inexpensive MHC genotyping system for any an-imal species. The recent description of over 7000 microsatellites densely distributed across the house mouse genome (Dietrich et al., 1996) now makes it possible to perform large-scale microsatellite-based MHC typing for this species. Although microsatellite loci are generally thought to be selectively neutral, a few theoretical studies have already shown that the statistics of microsatellite allele distribution would be influenced by an adjacent locus under selection, suggesting that neutral microsatellite loci can be in-dicators of selective processes at closely linked loci (Slatkin, 1995; Schlötterer & Wiehe, 1999). Sev-eral studies have used microsatellites to infer selection indirectly, for example, Schlötterer, Vogl and Tautz (1997), Paterson (1998), Huttley et al. (1999) and Kohn, Pelz and Wayne (2000).
Furthermore, comparing patterns of genetic poly-morphism among populations at different types of loci is an increasingly popular approach to assess the role of selection in determining allelic variation (Lewontin & Krakauer, 1973; Karl & Avise, 1992; Spitze, 1993; Pogson et al., 1995; Lynch et al., 1999). As gene flow and genetic drift affect all loci equally whereas selection is more likely to be locus-specific (Lewontin & Krakauer, 1973), discordance between potentially selected and neutral loci may be taken as evidence of selection (Spitze, 1993; Lynch et al., 1999). This can be achieved by comparing the pat-terns of genetic differentiation at the microsatellite loci within the MHC (MHC-microsatellites) with those outside the MHC region (non-MHC microsatellites) among populations. If the selective factors acting on MHC have been of significantly greater magnitude than non-selective factors, the patterns of variation for MHC and non-MHC microsatellite loci may differ greatly. On the contrary, the hypothesis of neutral-ity (no selection or weak selection) shaping MHC genetic structure would in turn be supported by the absence of statistical differences between the patterns of population structure at MHC- and at non-MHC microsatellites.
musculus castaneus) in Taiwan. We describe the
pop-ulation genetics of these loci by using standard genetic parameters and estimate the relationship among the populations. Further, as the pattern and level of poly-morphism across the genome are useful for identifying genes or genomic regions subject to natural selection (Satta, Li & Takahata, 1998), we compare the ex-tent and pattern of genetic variation for MHC- and non-MHC microsatellite loci, in order to discover if selection is, or has been, acting on the MHC region. Materials and methods
Mouse samples
Since June 1995, we have collected a large number of house mice from 19 populations inhabiting centralized rice granaries in townships in lowland Taiwan (Chou et al., 1998). Mice trapped from the same township were considered as one population. We grouped all populations into five geographical regions on the is-land of Taiwan (Figure 1, Table 2 and Appendix A). In addition, mice from Jinmen were treated as a separate population on an offshore isle.
Microsatellite loci
Five microsatellite loci (D17Mit28, D17Mit21, D17Mit33, D17Mit233, and D17Mit124) closely
linked to the MHC genes were employed in this study. These microsatellites lie between 18.00 and 22.50 cM from the centromere of Chromosome 17 (Figure 2). All are simple (CA)n repeat loci and are described in detail on the website of the Whitehead Institute for Biomedical Research at http://www.genome.wi.mit.edu. Of the five loci, three (D17Mit28, D17Mit21, and D17Mit33) are located within the conventional H-2 region (bounded by the
K and L genes) where most of the highly polymorphic
antigen-presenting MHC loci are found, whereas the other two (D17Mit233 and D17Mit124) are located close to the boundary of the MHC region and near the M region. General information about these MHC microsatellite loci is summarized in Table 1.
In an earlier attempt to investigate population dif-ferentiation in these mice, we employed six unlinked autosomal microsatellite loci (D6Mit138, D10Mit20,
203
Figure 1. Map indicating the location of the 19 mouse populations grouped into six geographic regions in our survey. Northwestern region
(NW) includes: 1, Tayuan; 2, Hsinwu; 3, Guanshi; 4, Hsinpu. Central region (C) includes: 5, Miaoli; 6, Taichung; 7, Tsaotun; 8, Shijou; 9, Linnei. Southwestern region (SW) includes: 10, Tainan; 11, Kaoping. Southeastern region (SE) includes: 12, Chrshang; 13, Guanshan; 14, Huanlien. Northeastern region (NE) includes: 15, Wuchiai; 16, Tungshan; 17, Ilan City; 18, Toucheng. The relative position of an offshore island, Jinmen (19), is shown in the inset panel. Light gray areas, 1000–2000 m; dark gray area, above 2000 m in elevation.
The data from that study are compared with those reported here to reveal any difference in the levels of polymorphism between loci located within and outside the MHC region. These six microsatellite loci are assumed to be neutral. The repeat motifs of the six microsatellite loci are as follows (Yu & Peng, 2002): 150 [(ATT)n], 34 [(GAAG)n], D10Mit20 [(TAGA)n], D15Mit16 [(TAGA)n], 105 [(ATTTT)n], and D6Mit138 [(GA)n(GAAA)m].
Genotyping
Genomic DNA extraction from mice (from liver, spleen, etc.) followed the standard phenol–chloroform
procedures (Ausubel et al., 1995). PCR was carried out in a thermal cycler (AG9600) in 10µl reactions overlaid with mineral oil containing: 200 ng of geno-mic DNA; 1× PCR buffer; 0.5 mM MgCl2; 0.25 mM
dNTP; 0.05 U Taq polymerase (Promega); 0.3µM of each primer, one of which was end-labeled with [γ-P33]-ATP (2000 Ci/mmol). Standard cycling
condi-tions were 94◦C for 3 min followed by 21–25 cycles of 94◦C for 1 min, 59◦C (for D17Mit33, D17Mit21 and
D17Mit28) or 61◦C (for D17Mit233 and D17Mit124) for 1 min, and 72◦C for 1 min. A final 10 min exten-sion step was added to complete the thermal profile. PCR products were then separated on 8 or 10% de-naturing sequencing gels and visualized by exposure
Figure 2. Linkage map of the five microsatellite loci and selected MHC genes on the 17th chromosome of the house mouse. Traditional H-2
regions include K, Aβ, Aα, Eβ, Eα, D, and L genes. Scales are linkage distances (cM) from centromere. The MHC loci are based on a linkage map in Meagher and Potts (1997).
Table 1. Number of alleles, variance in repeat number corrected for length difference (Vlength-corrected), and other information for the five microsatellite loci in 19 mouse populations in Taiwan
Locus Size range (bp) No. of alleles Vlength-corrected Locus locationa Classification
D17Mit33 158–212 15 1.658 C4/Slp hybrid 3 gene, non-transcriptional region Class III
D17Mit21 106–204 27 0.855 H2-A-beta-2 segment, intron 3 Class II
D17Mit28 86–130 19 1.060 H2-K (bm1) gene, promoter region Class I
D17Mit124 147–169 18 0.114 H2-M region Class I
D17Mit233 102–136 17 0.557 H2-M region Class I
aFrom http: www.ncbi.nlm.nih.gov.
to Kodak X-OMAT Blue film for 1–2 days. DNA se-quences of plasmids (pUC 18 or pUC 19) were run along with the samples as markers to determine the allele sizes. Mice of two inbred strains (B6 and CBA/J) and their F1 hybrid (purchased from the Lab
An-imal Center, National Taiwan University) were used as controls to recognize the bands (each representing an allele) for homozygotes (inbred strains) and for het-erozygotes (F1hybrids). This procedure helps remove
ambiguity in scoring alleles.
Data analysis
Both observed heterozygosity (Ho) and unbiased
ex-pected heterozygosity (He; Nei, 1978) were
calcu-lated to estimate the genetic variability in the 19 mouse populations. Hardy–Weinberg expectation for
each locus was tested by the Markov-chain method in GENEPOP 3.1 (Raymond & Rousset, 1995; http://wbiomed.curtin.edu.au/genepop), which imple-ments Fisher’s exact tests for multiple alleles (Guo & Thompson, 1992).
Variance in the number of repeats was calculated as another measure of variability. To account for length dependence of microsatellite mutation rates at each locus, variance in repeat number (Vi) was divided by the maximal number of repeat (Xi) as follows:
Vlength-corrected = 1 k k i=1 Vini Xi(ni − 1)
where k is the number of population, and ni is the number of chromosomes analyzed in the ith population.
205 Two genetic distance measures were calculated:
Nei’s (1972) standard distance, Ds, assuming an
in-finite allele model (IAM) (Nei, 1972), and Goldstein et al.’s (1995) distance (ðµ)2, assuming the stepwise mutation model (SMM). Calculations were imple-mented in MICROSAT 1.4 (Minch et al., 1995).
Population differentiation at different geographic scales was examined by determining FSTvalues (Weir
& Cockerham, 1984) between all possible pairs of 18 populations, or among populations within geographic regions (northwestern, central, southwestern, south-eastern, and northeastern) in Taiwan. The calculations were implemented in GENEPOP 3.1 (Raymond & Rousset, 1995). Significant departures from zero of the
FSTvalues were tested using permutations (see Dallas
et al., 1995) and were implemented in FSTAT (Goudet, 1995).
Isolation by distance was tested by the Mantel’s test (Mantel, 1967), which examines the overall rela-tionship between population differentiation (FST/(1−
FST)) and the logarithm of the geographic distance
separating populations (Rousset, 1997). The Mantel’s test was implemented in GENEPOP 3.1 (Raymond & Rousset, 1995).
Balancing selection was tested using Watterson’s (1978) homozygosity test. The expected proportion of homozygotes under Hardy–Weinberg equilibrium,
Fobs = (pi2), in a sample of size 2n genes
con-taining k alleles, is used as a measure of the allele frequency distribution and is compared to the homozy-gosity (Fexp) expected in a sample drawn from a
pop-ulation in mutation-drift equilibrium under neutrality (Ewens, 1972). The distribution of the homozygos-ity statistic under the finite-alleles neutral model, for combination of 2n and k up to 500 and 40, has been obtained from computer simulations and tabulated extensively at http://allele5.biol.berkeley.edu/homozy-gosity/homozygosity.html. Significantly low p-values reject the null hypothesis of neutrality and suggest the presence of selection.
Results
Genetic variability
Genetic variability for the five MHC microsatellite loci in 19 mouse populations was generally high. For all populations combined, the numbers of alleles for each locus ranged from 15 to 27, and the variance in the repeat number corrected for length differences
at each locus (Vlength-corrected) ranged from 0.855 to
1.658 (Table 1).
For individual populations, mean allele numbers per locus ranged from 3.6 to 9.8 (Appendix A). The observed (Ho) and the expected heterozygosity (He)
for individual populations ranged from 0.219 to 0.768 ( ¯Ho= 0.601) and from 0.330 to 0.835 ( ¯He= 0.713),
respectively (Table 2).
The level of polymorphism at each locus seems to be related to its position (Tables 1 and 2). The highest level of polymorphism was observed for D17Mit21 located in the class II region, with 27 alleles and an expected heterozygosity (He) equal to 0.801. The
low-est level of polymorphism was observed for D17Mit33 located in the class III region, with 15 alleles and an expected heterozygosity (He) equal to 0.659. The
three loci (D17Mit28, D17Mit124 and D17Mit233) in the class I region showed an intermediate level of poly-morphism, with 17–19 alleles and averaged expected heterozygosities (He) 0.684–0.717.
The levels of polymorphism varied among in-dividual populations as measured by the expected heterozygosity (Table 2). Guanshi was clearly the least polymorphic, with He = 0.330. On the other
hand, four populations (Jinmen, Tainan, Taichung, and Kaoping) showed high expected heterozygosity, ranging from 0.800 to 0.835.
Private alleles were found in over half of the popu-lations (10/19) for the five loci. The number of private alleles was highest in Jinmen (n = 5) (Appendix A). While most of the private alleles had frequencies less than 0.1, the D17Mit233 allele∗162 in Miaoli had a frequency as high as 0.3 (Appendix A).
For the 19 populations at these five loci, 28 out of the 95 cases involved a single most common allele with frequency exceeding 0.5 (= ‘predominant’ allele) (Appendix A). For individual populations, Guanshi had the largest number of loci with a ‘predominant’ allele, that is, each of the five loci had its own pre-dominant allele. For each individual locus, while there were 8, 6, 7, and 6 populations containing a ‘pre-dominant’ allele at D17Mit33, D17Mit28, D17Mit124, and D17Mit233, respectively, there was only one (Guanshi) at D17Mit21. Otherwise, most populations contained more than four alleles, with roughly equal frequencies, at each locus.
The proportions of microsatellite genotypes ob-served in each population sample were compared with Hardy–Weinberg expectations (HWE) using Fisher’s exact test (Guo & Thompson, 1992). Significant de-partures (P < 0.05) were found in 28 out of the 95
Table 2. Observed heterozygosity (Ho), expected heterozygosity (He), and sample size for the five MHC microsatellite loci in 19 Taiwanese house mouse populations Region Population Observed heterozygosity (Ho) Expected heterozygosity (He) Sample size
D17Mit D17Mit D17Mit D17Mit D17Mit Mean D17Mit D17Mit D17Mit D17Mit D17Mit Mean D17Mit D17Mit D17Mit D17Mit D17Mit
33 21 28 124 233 33 21 28 124 233 33 21 28 124 233 Northwestern Tayuan 0.389 0.579 0.684 0.421 0.684 0.551 0.492 0.824 0.670 0.551 0.799 0.667 18 19 19 19 19 Hsinwu 0.533 0.667 0.400 0.933 0.933 0.693 0.632 0.798 0.492 0.772 0.825 0.704 15 15 15 15 15 Guanshi 0.143 0.143 0.095 0.571 0.143 0.219 0.138 0.307 0.304 0.558 0.343 0.330 21 21 21 21 21 Hsinpu 0.600 0.800 0.200 0.300 0.800 0.520 0.668 0.790 0.442 0.337 0.737 0.595 10 10 10 10 10 Central Miaoli 0.200 0.400 0.600 0.400 0.800 0.440 0.200 0.844 0.711 0.600 0.867 0.644 5 5 5 5 5 Taichung 0.875 0.500 0.875 0.625 0.750 0.725 0.817 0.892 0.850 0.750 0.792 0.820 8 8 8 8 8 Tsaotun 0.556 0.778 0.778 0.667 0.500 0.656 0.532 0.730 0.787 0.687 0.522 0.652 18 18 18 18 18 Shijou 0.569 0.778 0.585 0.769 0.569 0.654 0.689 0.878 0.814 0.828 0.660 0.774 65 63 65 65 65 Linnei 0.556 0.657 0.580 0.720 0.590 0.620 0.676 0.878 0.687 0.792 0.607 0.728 99 99 100 100 100 Southwestern Tainan 0.667 0.741 0.556 0.593 0.778 0.667 0.783 0.894 0.827 0.773 0.804 0.816 27 27 27 27 27 Kaoping 1.000 0.778 0.667 0.333 0.889 0.733 0.808 0.895 0.752 0.686 0.856 0.800 8 9 9 9 9 Southeastern Chrshang 0.611 0.333 0.444 0.556 0.722 0.533 0.798 0.767 0.687 0.703 0.757 0.743 18 18 18 18 18 Guanshan 0.462 0.333 0.538 0.385 0.462 0.436 0.563 0.815 0.726 0.748 0.532 0.677 13 12 13 13 13 Hualien 0.842 0.868 0.632 0.216 0.921 0.696 0.809 0.838 0.745 0.202 0.760 0.671 38 38 38 37 38 Northeastern Wuchiai 0.500 0.545 0.417 0.250 0.700 0.482 0.529 0.887 0.833 0.377 0.760 0.677 12 11 12 12 10 Tungshan 0.375 0.267 0.533 0.750 0.714 0.528 0.548 0.352 0.637 0.790 0.878 0.641 16 15 15 12 14 Ilan City 0.595 0.515 0.343 0.700 0.719 0.580 0.720 0.803 0.781 0.806 0.835 0.789 37 33 35 30 32 Toucheng 0.667 0.786 0.235 0.333 0.462 0.497 0.651 0.743 0.745 0.595 0.406 0.625 18 14 17 9 13 Offshore Jinmen 0.767 0.895 0.737 0.579 0.895 0.768 0.838 0.809 0.878 0.727 0.925 0.835 19 19 19 19 19 Mean (over population) 0.580 0.641 0.530 0.597 0.659 0.601 0.659 0.801 0.717 0.684 0.692 0.713 – – – – –
207
Table 3. Measures of genetic distance for all pairwise combinations of house mouse populations in Taiwana
Region Popu- Northwestern Central Southwestern Southeastern Northeastern Offshore,
lation JM TY SW GS SP ML TC TT SJ LN TN KP CS GA HL WC TS IC TE Northwestern TY 0.440 0.871 0.320 0.541 0.420 0.456 0.483 0.461 0.361 0.379 0.721 0.396 0.341 0.231 0.601 0.335 0.806 0.421 SW 12.11 0.893 0.588 0.250 0.343 0.733 0.566 0.398 0.375 0.201 0.673 0.572 0.587 0.850 1.279 0.602 0.523 0.775 GS 54.24 18.47 0.772 0.720 0.878 1.321 0.989 0.905 1.348 0.635 1.166 1.849 0.827 1.480 1.454 1.452 1.028 0.910 SP 3.52 13.02 39.95 0.889 0.760 0.611 0.475 0.427 0.434 0.448 0.658 0.786 0.547 0.450 0.966 0.480 0.565 0.598 Central ML 40.24 11.97 26.86 43.50 0.458 1.123 0.836 0.597 0.620 0.447 0.743 0.894 0.463 0.852 1.042 0.916 1.160 1.043 TC 4.82 4.90 31.37 2.50 25.04 0.177 0.179 0.261 0.327 0.265 0.555 0.327 0.460 0.626 0.763 0.645 0.489 0.310 TT 5.75 25.37 62.69 2.56 63.25 6.67 0.181 0.333 0.680 0.535 0.838 0.278 0.570 0.638 0.895 0.558 0.516 0.435 SJ 14.77 17.86 42.02 8.08 39.17 1.59 8.37 0.149 0.519 0.496 0.559 0.271 0.616 0.664 0.655 0.542 0.512 0.456 LN 11.89 5.32 28.44 9.37 22.74 −0.31 14.35 4.04 0.402 0.320 0.482 0.338 0.482 0.707 0.873 0.530 0.460 0.595 Southwestern TN 6.19 9.69 44.67 7.26 28.13 −0.57 8.71 3.17 1.97 0.245 0.589 0.516 0.498 0.426 0.691 0.542 0.421 0.362 KP 4.32 0.32 22.16 2.72 20.74 0.05 12.00 10.22 3.54 5.17 0.627 0.646 0.515 0.615 1.203 0.592 0.557 0.512 Southeastern CS 12.92 8.46 21.80 4.47 25.78 0.97 12.33 4.41 4.25 6.85 2.42 0.384 0.656 0.656 0.929 0.592 0.596 0.867 GA 11.50 18.61 35.19 1.33 54.97 6.93 5.12 10.48 13.38 14.60 6.79 5.40 0.564 0.597 0.865 0.514 0.614 0.802 HL 12.87 2.95 18.18 8.29 27.50 2.97 16.77 11.16 3.39 8.97 1.14 5.40 8.88 0.446 0.926 0.677 0.578 0.484 Northeastern WC 0.24 13.87 48.75 0.83 40.33 6.10 5.89 15.79 15.46 9.85 3.37 9.39 7.24 14.71 0.332 0.486 0.763 0.338 TS 13.26 25.98 79.23 21.12 41.29 10.39 17.02 12.71 13.69 4.64 21.14 23.87 34.38 28.13 20.64 0.568 1.174 0.414 IC 0.66 17.51 65.50 6.53 50.91 8.02 5.43 17.44 14.92 8.31 9.28 18.74 14.51 16.45 4.43 13.13 0.587 0.613 TE 5.02 6.89 41.28 6.61 31.38 −0.36 8.15 5.56 1.31 0.32 3.50 8.22 12.39 4.73 10.59 8.37 5.65 0.557 Offshore JM 18.59 31.34 72.45 17.92 45.36 9.29 13.91 5.70 13.59 5.94 22.65 16.03 26.74 29.15 20.97 4.43 20.28 12.57 aBoth D
s(above diagonal) and (ðµ)2(below diagonal) are given. Dsand (ðµ)2values more than 1 and 40 are shown bold-faced. TY: Tayuan; SW: Hsinwu; GS: Guanshi; SP: Hsinpu; ML: Miaoli; TC: Taichung; TT: Tsaotun; SJ: Shijou; LN: Linnei; TN: Tainan; KP: Kaoping; CS: Chrshang; GA: Guanshan; HL: Haulien; WC: Wuchiai; TS: Tungshan; IC: Ilan City; TE: Toucheng; JM: Jinmen.
D17Mit33 0.4476∗∗∗ 0.1226∗∗∗ 0.0968∗∗ 0.1005∗∗∗ 0.1823∗∗∗ 0.1767∗∗∗ 0.1728∗∗∗ D17Mit21 0.2210∗∗∗ 0.0417∗∗∗ 0.0135∗ 0.1249∗∗∗ 0.2390∗∗∗ 0.1412∗∗∗ 0.1403∗∗∗ D17Mit28 0.3365∗∗∗ 0.0539∗∗∗ 0.0589∗∗∗ 0.2025∗∗∗ 0.0901∗∗∗ 0.1539∗∗∗ 0.1482∗∗∗ D17Mit124 0.1570∗∗∗ 0.0779∗∗∗ 0.0394∗ 0.2609∗∗∗ 0.1365∗∗∗ 0.1285∗∗∗ 0.1261∗∗∗ D17Mit233 0.2718∗∗∗ 0.0514∗∗∗ 0.0064 0.1383∗∗∗ 0.1537∗∗∗ 0.1664∗∗∗ 0.1631∗∗∗ Overall 0.3209∗∗∗ 0.0791∗∗∗ 0.0495 0.1773∗∗∗ 0.2029∗∗∗ 0.1764∗∗∗ 0.1722∗∗∗ aAll populations were combined except offshore island Jinmen.
∗P <0.05 from permutation tests in FSTAT program. ∗∗P <0.01 from permutation tests in FSTAT program. ∗∗∗P <0.001 from permutation tests in FSTAT program.
cases, most of which showed heterozygote deficit (P < 0.05) except one (Hwalien at D17Mit28). The distribution of the 28 cases was somewhat clustered by three populations and two loci. For individual pop-ulations, Guanshi, Shijou, and Linnei had four loci that showed heterozygote deficiencies. For individual loci, D17Mit21 and D17Mit28 showed significant de-viations from HWE in 9 and 8 out of the 19 pop-ulations, respectively. If the data from the five loci and the 19 populations were combined, the over-all genotype frequencies deviated significantly from HWE (P < 0.001).
Genetic distances and differentiation among regions
Two genetic distance measures, allele-frequency-based genetic distance, Ds, and repeat-size-based
genetic distance (ðµ)2were estimated (Table 3). Two aspects are noteworthy. First, the high values of ge-netic distance for Guanshi and Miaoli from other pop-ulations indicate that they were quite different from the other mouse populations in Taiwan. The (ðµ)2values between Guanshi and other populations averaged more than 40 ( ¯X = 40.05, SD = 17.68; n = 17), while
the pairwise average between all 18 populations was 15.20 (SD= 15.11; n = 153). The averaged (ðµ)2 between Miaoli and others was 34.93 (SD= 13.43;
n = 17). The same pattern was observed for Ds.
Second, the population of the offshore island, Jinmen, was particularly similar to those in central and south-western Taiwan (Taichung, Tsaotun, Shijou, Linnei, Tainan, and Kaoping; Figure 1), except Miaoli. The mean values of (ðµ)2(11.85) and Ds(0.445) between
Jinmen and these six populations were significantly less than those between Jinmen and the remaining 12
Taiwanese mouse populations ((ðµ)2 = 26.32 and
Ds = 0.652) (P = 0.024 for (ðµ)2, and P = 0.016
for Ds).
Wright’s FST (Weir & Cockerham, 1984) for
pairs of populations was calculated as a measure of population differentiation. We found no signifi-cant correlation between population differentiation and geographic distance (P= 0.365), using Mantel’s (Mantel, 1967).
Moreover, we examined inter-population genetic differentiation within different geographic regions (Table 4). For the five MHC loci, all FST values
for each region were highly significant (P < 0.001) except for southwestern populations, indicating that there was substantial isolation among populations in each region. The northwestern region had the most dif-ferentiated populations (FST = 0.3209), followed by
Table 5. Summarized result of Watterson’s neutrality test
apply-ing to the five MHC microsatellite loci in 19 Taiwanese mouse populations
Locus Neutrality Fobsb,c Fexpc,d Ratioc rejected (%)a D17Mit33 31.6 (6/19) 0.3938 0.4410 0.8708 D17Mit21 31.6 (6/19) 0.2509 0.2801 0.8675 D17Mit28 15.8 (3/19) 0.3198 0.3604 0.8846 D17Mit124 31.6 (6/19) 0.3770 0.4086 0.8799 D17Mit233 10.5 (2/19) 0.3068 0.3443 0.8722 aNumber of populations in which neutrality was rejected (P < 0.1). bF
obs: sum of the squares of all allele frequencies. cF
obs, Fexp, and ratio (= Fobs/Fexp) were weighted values aver-aged over all populations.
dF
exp: expected F-values under neutrality obtained by computer simulations.
209
Table 6. Comparisons of allele number, variance in repeat number, heterozygosity, FSTvalue, and proportion of populations in which neutrality was rejected between the MHC and the non-MHC microsatellite loci
MHCa Non-MHCa
D17Mit33 D17Mit21 D17Mit28 D17Mit124 D17Mit233 Mean D6Mit138 D10Mit20 D15Mit16 34 105 150 Mean
Total no. of allele 14 21 17 13 15 16.0 12 9 12 9 7 12 10.2
Average no. of allele 6.0 11.4 8.9 5.9 7.6 8.0 7.6 6.5 6.8 6.8 4.1 6.9 6.5
per populationb Logarithm of 3.513 3.256 3.180 0.771 2.352 2.614 0.642 1.164 1.225 0.832 0.633 1.349 0.974 variance in repeat number Variance in repeat 1.771 0.739 0.972 0.106 0.443 0.806 0.088 0.165 0.161 0.122 0.211 0.146 0.149 number (Vlength-corrected) Observed 0.615 0.686 0.540 0.561 0.674 0.615 0.579 0.705 0.711 0.590 0.366 0.467 0.570 heterozygosity (Ho)b Expected 0.679 0.810 0.686 0.640 0.675 0.698 0.690 0.753 0.739 0.664 0.490 0.546 0.647 heterozygosity (He)b FSTvalue 0.199 0.171 0.183 0.175 0.215 0.193 0.137 0.132 0.147 0.216 0.198 0.269 0.201 % neutrality rejected 57.1 42.9 14.3 28.6 14.3 31.4 14.3 57.1 28.6 14.3 0 14.3 21.4 aAll values are averaged for the seven populations that have been surveyed for both of the MHC and non-MHC loci.
(Table 4).
Test of neutrality
Although MHC-linked microsatellite loci are not di-rectly involved in the function of MHC molecules, strong selection acting on the MHC coding regions may have effects extended to these microsatellite loci that are otherwise evolving neutrally. Since an even distribution in allele frequency characterizes the highly polymorphic MHC loci, we applied the ho-mozygosity test (Watterson, 1978) to assess whether the neutral MHC-linked microsatellite loci give sim-ilar evidence of non-neutrality, serving as markers for adjacent sites under selection. The results showed heterogeneity as to rejection of neutrality across popu-lations (data not shown) and across loci (Table 5). The proportion of populations for which neutrality was re-jected at the 5 or 10% significance level varied among loci. The loci at which the largest proportion of pop-ulations (31.6%) rejected neutrality were D17Mit33,
D17Mit21, and D17Mit124. At all five loci, the
av-erage ratios of observed to expected homozygosity (Fobs/Fexp) across all populations were less than 1
(i.e., Fobssmaller than Fexp), and similar to each other
(0.8675–0.8846).
Comparison of genetic variation and FST values
between MHC and non-MHC microsatellite loci
To evaluate the neutral and selective forces influencing the MHC-linked microsatellite polymorphism, we fur-ther compared its extent and pattern with that of six non-MHC-linked microsatellite loci, in seven popula-tions that have been surveyed for both categories of markers (Table 6). The total number of alleles for the MHC loci was significantly greater than that for the non-MHC loci (t-test; P = 0.010), ranging from 13 to 21 for the MHC loci ( ¯X= 16.0, SD = 3.2), and from
7 to 12 for the non-MHC loci ( ¯X= 10.2, SD = 2.1).
Variance in the number of repeats, another mea-sure of diversity, was ln-transformed such that it was normally distributed. For the MHC-linked mi-crosatellite loci, the natural logarithms of variance in repeat number were significantly higher than for the non-MHC microsatellite loci (t-test; P= 0.03) (Table 6). Length-corrected variances in repeat num-ber (Vlength-corrected) were also compared between two
loci (0.088–0.211).
However, the other estimators, such as the averaged number of allele per population, the heterozygosity, and the FST value, were similar and
not significantly different between both categories of markers (Table 6). Finally, from the result of Watterson’s neutrality test, the proportions of pop-ulations for which neutrality was rejected at the five MHC loci ranged from 14.3 to 57.1% ( ¯X= 31.4%);
whereas those at the six non-MHC loci ranged from 0 to 57.1% ( ¯X= 21.4%) (Table 6).
Discussion
Level of genetic variation in MHC
MHC polymorphism in humans (known as HLA) is one of the few cases of adaptive evolution well-documented at the molecular level. In the present study, we employed five MHC-linked microsatellite loci to assay genetic variation in the MHC region of southeast Asian house mouse, M. m. castaneus, and found there were some differences in levels of allelic variation among the five microsatellite loci, as measured by the total number of alleles and the expec-ted heterozygosity (He). Of those five microsatellites,
D17Mit21, located within the intron 3 of the A-beta-2
gene, and D17Mit28, in the promoter region of the K gene, exhibited higher degrees of polymorphism than the others. This result was in agreement with the fact that the Aβgene in the class II family and the K gene in the class I family are among the most polymorphic MHC genes (Hood, Steinmetz & Malissen, 1983). By contrast, D17Mit33, located within the class III C4/Slp
hybrid 3 gene, showed less polymorphism than the
others. It was consistent with the DNA sequence and allozyme data from previous surveys showing low ge-netic variation at the class III genes, in spite of their being closely linked to highly polymorphic class I and II genes (Hood, Steinmetz & Malissen, 1983; Klitz, Thomson & Baur, 1986; Satta, Li & Takahata, 1998). The fact that some genes located within the MHC re-gion exhibit extensive polymorphism, whereas other do not, is an interesting feature that may reflect func-tional requirements for diversity in the polymorphic loci (Hood, Steinmetz & Malissen, 1983).
211
Comparing genetic variation between MHC and non-MHC microsatellite loci
We compared the genetic variability of the MHC-linked with that of the non-MHC-MHC-linked microsatellite loci, attempting to assess the influence of strong link-age disequilibrium extending across the MHC region. Here we presumed the six non-coding, non-MHC-linked microsatellite loci to be selectively neutral. Although there are a few notable exceptions (e.g., tri-nucleotide expansions) (Thornton et al., 1997; Koob et al., 1999; Rubinsztein, 1999), most microsatel-lites appear to conform to the expectations for neutral loci (Watkins, Bamshad & Jorde, 1995; Hambuch & Lacey, 2000). Variation at these loci should be inde-pendent of selection, but instead should reflect demo-graphic parameters such as effective population size. Therefore, analyses of the non-MHC microsatellites provide a potentially valuable complement to stud-ies of the microsatellites closely linked to functional MHC loci.
The greater genetic variation at the five MHC-linked than the six non-MHC-MHC-linked microsatellite loci (Table 6) was revealed by the total number of alleles and the variance in repeat number. It was con-sistent with the result from computer simulations that showed that for a microsatellite locus closely linked to a locus under overdominant selection, their variance in the number of repeats increased (Slatkin, 1995), reflecting the longer-than-expected coalescence times (Hudson & Kaplan, 1988; Kaplan, Darden & Hudson, 1988).
Higher diversity at the MHC-linked microsatel-lite loci also agrees with allozyme data, showing increased variation at MHC-linked allozyme genes (Nadeau, Collins & Klein, 1982). This might be attrib-uted to the ‘linkage’ of neutral microsatellite alleles to selectively maintained MHC alleles (Maynard Smith & Haigh, 1974; Kaplan, Hudson & Langley, 1989; Ellegren, Davies & Andersson, 1993). In sheep, the length variation of one microsatellite immediately ad-jacent to a class II gene DRB was found highly correlated with the sequence polymorphism at DRB (Paterson, 1998). In humans, it has been shown that balancing selection operating at a class IA locus can strongly influence polymorphism within some 100 kb (Satta & Takahata, 2000). The extent of polymor-phism in a linked neutral sub-region is sensitive to its recombination rate with the nearest selected locus (Satta, 1997). Other similar examples associated with close linkage to a locus under balancing selection
in-clude the increased silent variation in the S alleles of plants (Clark & Kao, 1991; Richman, Uyenoyama & Kohn, 1996) and at the MHC loci in humans and mice (Hughes & Nei, 1989; Grimsley, Mather & Ober, 1998; O’hUigin et al., 2000).
A high degree of microsatellite variation in the MHC region may also be caused by high underly-ing mutation rates. Although we cannot exclude this possibility completely, it was not supported by the finding of Melvold, Wang and Kohn (1997) that the spontaneous mutation rate of H2 class I genes, except for a single allele at one locus (H2-Kb), appears to be equivalent to that found for non-H2 histocompatibility genes and comparable to rates reported for a variety of other mouse genes. Also, studies by Satta et al. (1993) on nucleotide substitutions in primates indicated that the mutation rate at the MHC loci is no higher than that of other non-MHC loci despite their extraordinary polymorphism.
Based on the assumption that the extent and pat-terns of variation for the MHC and the non-MHC loci would differ greatly if the selective forces acting on MHC are of significantly greater magnitude than the non-selective forces, we further compared the allele frequency distributions against neutral expectations for these two categories of microsatellite loci. This result, however, did not provide strong evidence for selection acting on the MHC region (Table 6). For the five MHC-linked and the six non-MHC-linked loci, the proportions of populations for which neutrality was rejected did not differ significantly (P = 0.4099). Considering the low power of the homozygosity test, we applied a modified Lewontin–Krakauer (L–K) test comparing the mean FST values for two classes
of loci with various types of polymorphism (Lewontin & Krakauer, 1973; Pogson, Mesa & Boutilier, 1995; Barker et al., 1997), to the seven mouse populations that have been surveyed for both categories of loci. A simulation study showed that, if there is population subdivision, balancing selection leads to decreased expected FST values for neutral sites linked to the
selected locus (Schierup, Charlesworth & Vekemans, 2000b). With balancing selection, two causes can lead to lower degree of genetic differentiation among prox-imate populations. First, since alleles are kept in more equal frequencies than for a neutral locus, balancing selection increases within-deme diversity HS, relative
to total diversity, thus decreasing the numerator in the expression for FST(= (HT − HS)/HT). Second, an
incoming migrant allele new to a deme will be selected for and thereby maintained, increasing its chance of
reduction in FST values at overdominant loci among
populations. A comparison of the FSTvalues of
over-dominant loci with those of neutral loci should provide evidence of balancing selection.
The result of the L–K test, however, did not show a significant indication of selection (FST(MHC)/FST(non-MHC)= 0.193/0.201). In
ad-dition to low power of the statistic test, there are two possibilities that might help explain failure to detect balancing selection. First, although a locus closely linked to a locus under balancing selection will show higher polymorphism than a neutral locus (Nei & Li, 1980; Strobeck, 1983), the polymor-phism seen at such a linked locus will be a function of the strength of linkage to the selected locus (Hughes & Yeager, 1998). In addition, even though the mi-crosatellite loci we used are within selected genes, they mostly reside in the non-coding region or the introns of genes, inasmuch as their mutation would disrupt any open reading frame (Schlötterer, Amos & Tautz, 1991). However, balancing selection acts primarily on exons, particularly those encoding the PBR, rather than on introns. Data from human class I loci HLA-A, -B, and -C have shown that nucleotide diversity is generally lower in introns than in exons (Hughes & Yeager, 1998; Hughes, 2000). In addition, despite being linked to exon sequences to some ex-tent, intron sequences are subject to homogenization by interallelic recombination, subsequent genetic drift, and loss of polymorphism (Hughes & Yeager, 1998; Meyer & Blasczyk, 2000).
Secondly, other evolutionary forces, such as ge-netic drift, may have a profound effect in shaping MHC diversity in the populations we surveyed. If that is the case, the extent and pattern of variation at the two types of microsatellite markers may not differ much; the overall FST values estimated from
MHC- and non-MHC-linked microsatellite loci did not differ significantly. The similar population struc-ture estimated from the two classes of loci might in-dicate that drift and limited migration have been more important than selection in shaping inter-population differentiation at the MHC-linked microsatellite loci. Comparable results have been found in bighorn sheep (Boyce et al., 1997) and in Atlantic salmon (Landry & Bernatchez, 2001), in which MHC genes and mi-crosatellites outside the MHC region gave similar FST
estimates across regions. In addition, recent
discover-fects of genetic drift (Belich et al., 1992; Watkins et al., 1992; Titus-Trachtenberg, Bugawan & Erlich, 1994). As these mouse populations inhabiting grana-ries were typically ephemeral and unstable due to regular turnover of grain within a 2–3-year period and occasional applications of poisons (Chou et al., 1998), the influence of genetic drift may have a major effect although our results also indicate that selection con-tinues to play a role in shaping the pattern of MHC variation.
There may be other problems with using mi-crosatellite loci in this context. For example, the mutation rate and/or mutation process may actually vary among microsatellite loci (Harr et al., 1998). Several parameters including the number of repeats, repeat motif length, and repeat motif composition, have been considered to influence microsatellite vari-ability (Schlötterer & Tautz, 1992; Schlötterer et al., 1998; Schug et al., 1998; Estoup & Cornuet, 1999; Bachtrog et al., 2000). Furthermore, the mutation pro-cess of microsatellites might not be as simple as first assumed for microsatellites and may follow a two-phase model. The prevalence of different mechanisms varies according to the structure of the repeat unit it-self (Shriver et al., 1993; Estoup & Cornuet, 1999), confounding the interpretation of the data. However, even if the microsatellites cannot provide conclusive evidence of the effects of linked selected loci, their abundance in genomes and ease of survey are still be-lieved to complement and test hypotheses based on comparisons of DNA sequences.
Genetic differentiation among populations and regions
Our results suggest that the seven populations in the central and southwestern regions of Taiwan might be part of a metapopulation structure. Within this re-gion (central–southwestern), the mean FST value was
relatively small both for the MHC (0.091) and for non-MHC microsatellite loci (0.104) (data not shown), suggesting that there may be high gene flow between these populations. By contrast, the northwestern re-gion appeared to be very different from these two regions. From Tables 3 and 4, it appears that the sharpness of the differentiation between the central– southwestern and the northwestern regions was at-tributable to two populations (Guanshi and Miaoli).
213 If these two populations were excluded, the mean
FST value of the western part of Taiwan
(north-western+ central + southwestern regions) would be lowered from 0.161 to 0.108 for the five MHC mi-crosatellite loci. The frequent gene exchange, implied by the low genetic differentiation among the pop-ulations in the western part of Taiwan, may occur through some pockets of feral populations (Chou et al., 1998), or through long-distance dispersal associated with human activities (Yu & Peng, 2002).
Among the 19 populations, there are some lines of evidence suggesting the Guanshi might be a pop-ulation established by a few founders, or subject to extinction–recolonization events. First, it was the least polymorphic population of all, with the lowest het-erozygosity (Table 2) and effective number of alleles (ne) (data not shown). Second, this population was
genetically distinct from most other populations, and in particular, it was quite different from its neigh-boring population (10.5 km apart only) of Shinpu (Table 3). This gave northwestern Taiwan the greatest inter-population differentiation (Table 4). Third, most of the homozygosity values for Guanshi were higher than the expected neutral value (i.e., Fobs/Fexp >1),
either at the five MHC loci (Fobs/Fexp = 1.4055,
1.9524, 1.6686, and 1.5796 for D17Mit33, D17Mit21,
D17Mit28, and D17Mit233), or at the six non-MHC
loci (Fobs/Fexp= 1.4948, 1.1544, 1.2377, and 1.1827
for D6Mit138, D15Mit16, 34, and 150). In general, a value of homozygosity higher than the expected neu-tral value could reflect founder effects and population bottlenecks, genetic drift, or positive directional se-lection for an allele (Mack et al., 2000). As a result, the bottleneck/founder effect appeared to be the expla-nation most consistent with the low genetic variation found in Guanshi.
On the other hand, the genetic diversity of Jinmen’s mouse population was the greatest among all the pop-ulations studied (Table 2, Appendix A), even though this offshore island has an area of only 132 km2. Fur-thermore, the population of Tainan, where the early settlement of Han Chinese people began, was the one most similar to that of Jinmen (Table 3) in ad-dition to Taichung. Interestingly, Tainan was also the only population with a genetic diversity comparable to Jinmen (Table 2, Appendix A). Therefore, we pre-sume that the mice from Jinmen, only 1–2 km away from Mainland China, are representative of mouse ge-nomes from Southeast China whence certain founders of modern Taiwanese mouse populations might have originated. And the close genetic relatedness between
Jinmen’s and Tainan’s populations, separated by the Taiwan Strait, indicates that there must have been human-mediated gene flow in M. m. castaneus popula-tions due to historical agriculture expansion and recent frequent traffic (Yu & Peng, 2002).
Acknowledgements
We thank Hwei-Yu Chang, Wen-Hsiung Li, Manyuan Long, Janice Spofford, Christian Schlötterer and four anonymous reviewers for their comments on earlier drafts of the manuscript. The comments help to im-prove the manuscript. This research was supported by the National Science Council of the Republic of China.
References
Ausubel, F.M., R. Brent, R.E. Kingston, D.D. Moore, J.G. Seidman, J.A. Smith & K. Struhl, 1995. Short Protocols in Molecular Biology – A Compendium of Methods from Current Protocols in Molecular Biology, Wiley, Canada, 3rd edn.
Bachtrog, D., M. Agis, M. Imhof & C. Schlötterer, 2000. Microsatellite variability differs between dinucleotide repeat motifs-evidence from Drosophila melanogaster. Mol. Biol. Evol. 17: 1277–1285.
Barker, J.S.F., S.S. Moore, D.J.S. Hetzel, D. Evans, S.G. Tan & K. Byrne, 1997. Genetic diversity of Asian water buffalo (Bubalus
bubalis): microsatellite variation and a comparison with
protein-coding loci. Anim. Genet. 28: 103–115.
Belich, M.P., J.A. Madrigal, W.H. Hildebrand, J. Zemmour, R.C. Williams, R. Luz, M.L. Petzl-Erler & P. Parham, 1992. Unusual
HLA-B alleles in two tribes of Brazilian Indians. Nature 257:
326–329.
Boyce, W.M., P.W. Hedrick, N.E. Muggli-Cockett, S. Kalinowski, M.C.T. Penedo & R.R. Ramey, 1997. Genetic variation of major histocompatibility complex and microsatellite loci: a comparison in bighorn sheep. Genetics 145: 421–433.
Chou, C.-W., P.-F. Lee, K.-H. Lu & H.-T. Yu, 1998. A population study of house mice (Mus musculus castaneus) inhabiting rice granaries in Taiwan. Zool. Stud. 37(3): 201–212.
Clark, A. & T. Kao, 1991. Excess nonsynonymous substitutions at shared polymorphic sites among self-incompatibility alleles of Solanaceae. Proc. Natl. Acad. Sci. USA 88: 9823–9827. Dallas, J.F., B. Dod, P. Boursot, E.M. Prager & F. Bonhomme, 1995.
Population subdivision and gene flow in Danish house mice. Mol. Ecol. 4: 311–320.
Dietrich, W.F., J. Miller, R. Steen, M.A. Merchant, D. Damron-Boles, Z. Husain, R. Dredge, M.J. Daly, K.A. Ingalls, T.J. O’Connor, C.A. Evans, M.M. DeAngelis, D.M. Levinson, L. Kruglyak, N. Goodman, N.G. Copeland, N.A. Jenkins, T.L. Hawkins, L. Stein, D.C. Page & E.S. Lander, 1996. A com-prehensive genetic map of the mouse genome. Nature 380: 149–152.
Ellegren, H., C.J. Davies & L. Andersson, 1993. Strong associ-ation between polymorphisms in an intronic microsatellite and in the coding sequence of the BoLA-DRB3 gene: implications for
ences from population data, pp. 49–65 in Microsatellites: Evolu-tion and ApplicaEvolu-tion, edited by D.B. Goldstein & C. Schlötterer. Oxford University Press, New York.
Ewens, W.J., 1972. The sampling theory of selectively neutral alleles. Theor. Pop. Biol. 3: 87–112.
Goldstein, D.B., A.R. Linares, L.L. Cavalli-Sforza & M.W. Feldman, 1995. Genetic absolute dating based on microsatellites and the origin of modem humans. Proc. Natl. Acad. Sci. USA 92: 6723–6727.
Goudet, J., 1995. FSTAT (version 1.2): a computer program to calculate F-statistics. J. Hered. 86: 485–486.
Grimsley, C., K.A. Mather & C. Ober, 1998. HLA-H: a pseudo-gene with increased variation due to balancing selection at neighboring loci. Mol. Biol. Evol. 15(12): 1581–1588. Guo, S.W. & E.A. Thompson, 1992. Performing the exact test of
Hardy–Weinberg proportion for multiple alleles. Biometrics 48: 361–372.
Hambuch, T.M. & E.A. Lacey, 2000. Contrasting patterns of MHC and microsatellite diversity in social and solitary tuco-tucos (Rodentia: Ctenomyidae), pp. 542–554 in Major Histocompat-ibility Complex: Evolution, Structure, and Function, edited by M. Kasahara. Springer, Tokyo.
Harr, B., B. Zangerl, G. Brem & C. Schlötterer, 1998. Conservation of locus specific microsatellite variability across species: a com-parison of two Drosophila sibling species D. melanogaster and
D. simulans. Mol. Biol. Evol. 15: 176–184.
Hood, L., M. Steinmetz & B. Malissen, 1983. Genes of the major histocompatibility complex of the mouse. Ann. Rev. Immunol. 1: 529–568.
Hudson, R.R. & N.L. Kaplan, 1988. The coalescent process in models with selection and recombination. Genetics 120: 831–840.
Hughes, A.L., 2000. Balancing selection: the major histocompat-ibility complex, pp. 54–89 in Adaptive Evolution of Genes and Genomes. Oxford University Press, New York.
Hughes, A.L. & M. Nei, 1989. Nucleotide substitution at major his-tocompatibility complex class II loci: evidence for overdominant selection. Proc. Natl. Acad. Sci. USA 86: 958–962.
Hughes, A.L. & M. Yeager, 1998. Natural selection at major histo-compatibility complex loci of vertebrates. Ann. Rev. Genet. 32: 415–435.
Huttley, G.A., M.W. Smith, M. Carrington & S.J. O’Brein, 1999. A scan for linkage disequilibrium across the human genome. Genetics 152: 1711–1722.
Kaplan, N.L., T. Darden & R.R. Hudson, 1988. The coalescent process in models with selection. Genetics 120: 819–829. Kaplan, N.L., R.R. Hudson & C.H. Langley, 1989. The hitchhiking
effect revisited. Genetics 123(4): 887–899.
Karl, S.A. & J.C. Avise, 1993. Balancing selection at allozyme loci in oysters: implications from nuclear RFLPs. Science 256: 100–102.
Klitz, W., G. Thomson & M.P. Baur, 1986. Contrasting evolutionary histories among tightly linked HLA loci. Am. J. Hum. Genet. 39: 340–349.
Kohn, M.H., H.-J. Pelz & R.K. Wayne, 2000. Natural selection mapping of the warfarin-resistance gene. Proc. Natl. Acad. Sci. USA 97(14): 7911–7915.
Koob, M.D., M.L. Moseley, L.J. Schut, K.A. Benzow, T.D. Bird, J.W. Day & L.P.W. Ranum, 1999. An untranslated CTG expan-sion causes a novel form of spinocerebellar ataxia (SCA8). Nat. Genet. 21(4): 379–384.
Atlantic salmon (Salmo salar). Mol. Ecol. 10: 2525–2539. Lewontin, R.C. & J. Krakauer, 1973. Distribution of gene frequency
as a test of the theory of selective neutrality of polymorphisms. Genetics 74: 175–195.
Lynch, M., M. Pfrender, K. Spitze, N. Lehman, J. Hicks, D. Allen, L. Latta, M. Ottene, F. Bogue & J. Colbourne, 1999. The quanti-tative and molecular genetic architecture of a subdivided species. Evolution 53: 100–110.
Mack, S.J., T.L. Bugawan, M. Stoneking, M. Saha, H.-P. Beck & H.A. Erlich, 2000. HLA class I and class II loci in Pacific/Asian populations, pp. 427–445 in Major Histocompatibility Complex: Evolution, Structure, and Function, edited by M. Kasahara. Springer, Tokyo.
Mantel, N., 1967. The detection of disease clustering and a general-ized regression approach. Cancer Res. 27: 209–220.
Maynard Smith, J. & J. Haigh, 1974. The hitch-hiking effect of a favorable gene. Genet. Res. Camb. 23: 23–35.
Meagher, S. & W.K. Potts, 1997. A microsatellite-based MHC gen-otyping system for house mice (Mus domesticus). Hereditas 127: 75–82.
Melvold, R.W., K. Wang & H.I. Kohn, 1997. Histocompatibility gene mutation rates in the mouse: a 25-year review. Immuno-genetics 47: 44–54.
Meyer, D. & R. Blasczyk, 2000. The effect of mutation, recombin-ation and selection on HLA non-coding sequences, pp. 398–411 in Major Histocompatibility Complex: Evolution, Structure, and Function, edited by M. Kasahara. Springer, Tokyo.
Minch, E., A. Ruiz-Linares, D. Goldstein, M. Feldman & L.L. Cavalli-Sforza, 1995. Microsat (version 1.4d): a computer pro-gram for calculating various statistics on microsatellite allele data.
Nadeau, J.H., R.L. Collins & J. Klein, 1982. Organization and evo-lution of the mammalian genome. I. Polymorphism of H-2 linked loci. Genetics 102: 583–598.
Nei, M., 1972. Genetics distance between populations. Am. Nat. 106: 283–292.
Nei, M., 1978. Estimation of average heterozygosity and ge-netic distance from a small number of individuals. Gege-netics 89: 583–590.
Nei, M. & W.-H. Li, 1980. Non-random association between elec-tromorphs and inversion chromosomes in finite populations. Genet. Res. Camb. 35: 65–83.
Nei, M. & A.L. Hughes, 1991. Polymorphism and evolution of the Major histocompatibility complex loci in mammals, pp. 222–247 in Evolution at the Molecular Level, edited by R.K. Selander, A.G. Clark & T.S. Whittam. Sinauer Associates Inc., Sunderland, MA.
O’hUigin, C., Y. Satta, A. Hausmann, R.L. Dawkins & J. Klein, 2000. The implications of intergenic polymorphism for major histocompatibility complex evolution. Genetics 156(2): 867–877.
Paterson, S., 1998. Evidence for balancing selection at the major histocompatibility complex in a free-living ruminant. J. Hered. 89: 289–294.
Pogson, G.H., K.A. Mesa & R.G. Boutilier, 1995. Genetic popula-tion structure and gene flow in the Atlantic cod Gadus morhua: a comparison of allozyme and nuclear RFLP loci. Genetics 139: 376–385.
Potts, W.K. & E.K. Wakeland, 1993. Evolution of MHC genetic di-versity: a tale of incest, pestilence and sexual preference. Trends Genet. 9(12): 408–412.
215
Raymond, M. & F. Rousset, 1995. GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. J. Heredity 86: 248–249.
Richman, R., M. Uyenoyama & J. Kohn, 1996. Allelic diversity and gene genealogy at the self-incompatibility locus in the Solanaceae. Science 273: 1212–1216.
Rousset, F., 1997. Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics 145: 1219–1228.
Rubinsztein, D.C., 1999. Trinucleotide expansion mutation cause diseases which do not conform to classical Mendelian expecta-tion, pp. 80–97 in Microsatellites – Evolution and Applications, edited by D.B. Goldstein & C. Schlötterer. Oxford University Press, New York.
Satta, Y., 1997. Effects of intra-locus recombination of HLA poly-morphism. Hereditas 127: 105–112.
Satta, Y., Y.-J. Li & N. Takahata, 1998. The neutral theory and natural selection in the HLA region. Frontier Biosci. 3: d459–d467.
Satta, Y. & N. Takahata, 2000. Polymorphism in the HLA class I region, pp. 398–411 in Major Histocompatibility Complex: Evolution, Structure, and Function, edited by M. Kasahara. Springer, Tokyo.
Satta, Y., C. O’hUigin, N. Takahata & J. Klein, 1993. The synony-mous substitution rate of the major histocompatibility complex loci in primates. Proc. Natl. Acad. Sci. USA 90: 7480–7484. Satta, Y., C. O’hUigin, N. Takahata & J. Klein, 1994. Intensity of
natural selection at the major histocompatibility complex loci. Proc. Natl. Acad. Sci. USA 91: 7184–7188.
Schierup, M.H., X. Vekemans & D. Charlesworth, 2000a. The effect of subdivision on variation at multi-allelic loci under balancing selection. Genet. Res. Camb. 76: 51–62.
Schierup, M.H., D. Charlesworth & X. Vekemans, 2000b. The effect of hitch-hiking on genes linked to a balanced polymorphism in a subdivided population. Genet. Res. Camb. 76: 63–73.
Schlötterer, C. & D. Tautz, 1992. Slippage synthesis of simple sequence DNA. Nucl. Acid Res. 20: 211–215.
Schlötterer, C. & T. Wiehe, 1999. Microsatellites, a neutral marker to infer selective sweeps, in Microsatellites: Evolution and Application, edited by D.B. Goldstein & C. Schlötterer. Oxford University Press, New York.
Schlötterer, C., B. Amos & D. Tautz, 1991. Conservation of poly-morphic simple sequence loci in cetacean species. Nature 354: 63–65.
Schlötterer, C., C. Vogl & D. Tautz, 1997. Polymorphism and locus-specific effects on polymorphism at microsatellite loci in natural
Drosophila melanogaster populations. Genetics 146: 309–320.
Schlötterer, C., R. Ritter, B. Harr & G. Brem, 1998. High mutation rates of a long microsatellite allele in Drosophila melanogaster provide evidence for allele-specific mutation rates. Mol. Biol. Evol. 15: 1269–1274.
Schug, M.D., C.M. Hutter, K.A. Wetterstrand, M.S. Gaudette, T.F. Mackay & C.F. Aquadro, 1998. The mutation rates of di-, tri-, and tetranucleotide repeats in Drosophila melanogaster. Mol. Biol. Evol. 15: 1751–1760.
Shriver, M.D., L. Jin, R. Chakraborty & E. Boerwinkle, 1993. VNTR allele frequency distributions under the SMM: A com-puter simulation approach. Genetics 134: 983–993.
Slatkin, M., 1995. Hitchhiking and associative overdominance at a microsatellite locus. Mol. Biol. Evol. 12(3): 473–480.
Spitze, K., 1993. Population structure in Daphnia obtuse: quantita-tive genetic and allozymic variation. Genetics 135: 367–374. Strobeck, C., 1983. Expected linkage disequilibrium for a
neu-tral locus linked to a chromosomal arrangement. Genetics 103: 545–555.
Takahata, N. & M. Nei, 1990. Allelic genealogy under overdo-minant and frequency-dependent selection and polymorphism of major histocompatibility complex loci. Genetics 124: 967–978. Takahata, N., Y. Satta & J. Klein, 1992. Polymorphism and
balanc-ing selection at major histocompatibility complex loci. Genetics 130: 925–938.
Thornton, C.A., J.P. Wymer, Z. Simmons, C. McClain & R.T. Moxley III, 1997. Expansion of the myotonic dystrophy CTG repeat reduces expression of the flanking DMAHP gene. Nat. Genet. 16: 407–409.
Titus-Trachtenberg, E.A., T.L. Bugawan & H.A. Erlich, 1994. Identification of a novel DQB1 allele (∗0609) segregating in an Ashkenazi Jewish family: implications for DQB1 typing systems. Tissue Antigens 44(2): 120–124.
Watkins, D.I., S.N. McAdam, X. Liu, C.R. Strang, E.L. Milford, C.G. Levine, T.L. Garber, A.L. Dogon, C.I. Lord, S.H. Ghim, G.M. Troup, A.L. Hughes & N.L. Letvin, 1992. New recom-binant HLA-B alleles in a tribe of South American Amerindi-ans indicate rapid evolution of MHC class I loci. Nature 357: 329–333.
Watkins, W.S., M. Bamshad & L.B. Jorde, 1995. Population ge-netics of trinucleotide repeat polymorphism. Hum. Mol. Genet. 4(9): 1485–1491.
Watterson, G.A., 1978. The homozygosity test of neutrality. Genet-ics 88: 405–417.
Weir, B.S. & C.C. Cockerham, 1984. Estimating F-statistics for the analysis of population structure. Evolution 38: 1358–1370. Yu, H.T. & Y.H. Peng, 2002. Population differentiation and gene
flow revealed by microsatellite DNA markers in the house mouse (Mus musculus castaneus) in Taiwan. Zool. Sci. 19(3): 475–483.
Appendix A.
Allele frequencies at the five MHC microsatellite loci in 19 Southeast Asian mouse populations grouped into six geographic regions in Taiwan. Populations were grouped into six geographical regions by locations. Alleles are in sizes (bp). Private alleles in each pop-ulation are shown bold-faced. See Table 2 for sample sizes and Table 3 for population abbreviations.
TY SW GS SP ML TC TT SJ LN TN KP CS GA HL WC TS IC TE D17Mit33 158 – – – – – – – – – 0.037 – – – – – – – – – 160 – – – – – – – 0.008 0.005 – – – – – – – – – – 172 0.139 0.467 0.929 – 0.900 0.250 0.028 0.239 0.379 0.037 0.313 0.278 0.115 0.224 0.042 0.094 – – – 174 – – – – – – – – – – 0.063 – – – – – – – – 176 – – 0.048 – – 0.063 – – – – – – – 0.197 – – – – 0.105 192 – – – – – – – – – 0.074 – 0.056 0.077 0.092 – – – – – 194 0.028 0.400 0.028 0.200 – 0.188 – 0.069 0.005 0.315 0.250 0.167 0.077 0.013 0.208 0.063 0.027 0.528 0.132 196 – – – – 0.100 – – – – – – 0.028 – 0.184 0.042 – 0.014 0.222 0.132 198 – – – 0.450 – – – 0.100 0.303 0.296 – 0.278 0.077 0.026 0.042 0.156 0.230 0.194 – 200 – – – – – – – – – – – – – – – – – 0.028 – 202 – – – – – – – – – – – – – – – – 0.014 – 0.053 206 0.694 0.100 – 0.350 – 0.313 0.556 0.485 0.303 0.185 0.125 0.194 0.654 0.263 0.667 0.656 0.446 0.028 0.316 208 – – – – – – – – – 0.037 – – – – – 0.031 0.135 – – 210 0.139 0.033 – – – 0.188 0.417 0.100 0.005 0.019 0.250 – – – – – 0.135 – 0.158 212 – – – – – – – – – – – – – – – – – – 0.105 D17Mit21 106 – – – – – – – – – – – – – – – – – – 0.026 110 – 0.033 – – – 0.063 0.167 0.262 0.187 0.185 0.056 – – – – – – 0.321 0.395 112 – 0.033 – – – – – 0.056 0.025 – – – – – – – 0.015 – – 114 – – – – – – – 0.008 0.005 0.130 – – – – – – – – – 120 – – – – 0.300 – – 0.071 0.035 0.093 – 0.056 – 0.066 0.182 0.800 0.030 – 0.132 122 – – – – – – – – – – – – – – – – – – 0.105 124 – – – – 0.100 0.125 0.028 0.032 – – – – – – – – – – – 126 – – – – – – – – – 0.019 – – – – – – – – 0.053 128 – – – – – 0.063 – – – – 0.056 0.333 – – – – – – – 130 0.026 – 0.048 – – 0.063 – – – – – – – 0.197 0.182 0.100 0.076 – 0.105 132 0.079 – – – – 0.188 – 0.024 0.096 – – – – – 0.227 – 0.015 – – 134 0.026 – – 0.050 – – – 0.103 0.202 – – 0.167 – 0.105 0.046 – – 0.036 0.026 136 0.316 0.133 – 0.150 0.100 0.063 0.083 0.119 0.106 0.167 0.167 – 0.375 – 0.091 0.100 0.242 0.107 0.026 138 0.105 0.400 – 0.400 0.300 – 0.111 0.071 0.005 0.074 – – – 0.013 – – 0.318 – – 140 0.026 0.033 0.024 0.150 – 0.188 – 0.024 0.081 0.148 0.222 – – 0.105 – – 0.015 – 0.105 142 0.211 0.033 0.048 – 0.200 0.250 0.139 0.048 0.071 0.056 0.056 0.306 0.125 0.224 – – – – – 144 0.026 – 0.024 – – – 0.472 0.127 0.126 0.037 0.111 – 0.208 – 0.091 – 0.015 0.036 0.026 146 0.184 0.167 0.833 0.200 – – – 0.016 0.015 0.019 0.167 – 0.042 0.013 – – 0.182 0.393 – 148 – 0.133 0.024 – – – – – – – – 0.139 0.042 – – – 0.091 0.107 – 150 – – – 0.050 – – – 0.032 0.025 – – – 0.083 0.026 – – – – –
217 152 – – – – – – – – 0.020 0.037 0.167 – 0.083 0.237 – – – – – 154 – – – – – – – – – – – – 0.042 – – – – – – 156 – – – – – – – 0.008 – – – – – – – – – – – 158 – – – – – – – – – 0.037 – – – 0.013 – – – – – 162 – 0.033 – – – – – – – – – – – – – – – – – 180 – – – – – – – – – – – – – – 0.091 – – – – 204 – – – – – – – – – – – – – – 0.091 – – – – D17Mit28 86 0.079 – – – – 0.028 0.008 0.030 0.093 0.111 0.278 – 0.013 0.167 0.267 0.371 0.412 0.026 90 0.553 0.700 0.071 0.300 0.500 0.313 0.306 0.192 0.460 0.333 0.389 – 0.308 0.368 0.125 0.167 0.214 0.265 0.184 92 0.105 – – – – – – 0.008 0.005 0.204 – – – – – – – – 0.053 94 – – – – 0.300 0.125 – – – – – – – – – – 0.171 – 96 – – – – – – – 0.062 0.030 0.037 – – – 0.026 0.292 0.533 0.043 – 0.184 102 0.026 0.033 – – – – – – – – – – – 0.132 – – – – – 104 – – – – – 0.063 – 0.062 0.030 0.111 – – – – – 0.033 – – – 106 – – – – 0.100 0.063 0.028 0.139 0.005 – – – – 0.316 – – 0.114 – 0.026 108 0.079 0.167 0.024 – – 0.250 0.167 0.077 0.085 0.074 0.111 – – – 0.250 – – 0.029 0.184 110 0.132 0.067 0.833 0.700 – 0.125 0.250 0.339 0.305 – 0.333 0.139 – 0.053 – – 0.014 0.147 0.158 112 – – 0.024 – – – 0.222 0.008 – – – – 0.115 – – – 0.014 – 0.026 114 – – – – 0.100 0.063 – – – 0.019 0.056 – – – – – 0.043 0.118 0.053 116 – 0.033 0.048 – – – – – 0.035 – – 0.111 0.115 0.092 – – – – – 118 – – – – – – – – – 0.037 – 0.472 0.423 – 0.083 – 0.014 – 0.105 122 – – – – – – – – – – – – – – 0.042 – – – – 124 – – – – – – – – – 0.074 – – – – 0.042 – – 0.029 – 126 – – – – – – – – 0.005 – – – – – – – – – – 128 – – – – – – – 0.085 – 0.019 – – – – – – – – – 130 0.026 – – – – – – 0.023 0.010 – – – 0.039 – – – – – – D17Mit124 147 – 0.133 0.024 – – – – – – – – – – – – – – – 0.053 149 – – – – – 0.063 0.111 0.092 0.235 0.019 0.222 0.083 0.115 – – – 0.200 – – 150 – – – – – – – – – – – – – – – – – 0.056 – 151 0.079 0.033 – – 0.100 – – – – – – – – – – – 0.217 0.111 0.053 152 – – – – – – – – – 0.037 – 0.028 – – – – – – – 153 0.105 0.267 0.024 – 0.300 0.188 – 0.185 0.065 0.111 – 0.361 0.308 0.027 – 0.125 0.083 – – 154 – – – – – – – – – – – – – 0.014 – 0.042 – – – 155 0.658 0.367 0.500 0.800 0.600 0.250 0.472 0.200 0.315 0.426 0.389 0.417 0.346 0.892 0.792 0.250 0.300 0.611 0.395 156 – – – – – – – – – – – – – – – – – – 0.053 157 – – 0.452 – – 0.438 0.278 0.154 0.060 0.148 – – – – – 0.375 0.017 0.222 0.342 158 0.053 – – 0.200 – – 0.139 0.254 0.055 – – – – – 0.042 0.125 0.033 – – 159 – – – – – – – 0.008 0.205 0.074 – – – 0.068 0.042 0.083 0.150 – 0.105 160 – – – – – – – – – – – 0.028 0.231 – – – – – – (continued)
Locus Allele Northwestern Central Southwestern Southeastern Northeastern Offshore, JM TY SW GS SP ML TC TT SJ LN TN KP CS GA HL WC TS IC TE 161 – 0.167 – – – 0.063 – 0.008 0.065 0.167 0.389 0.056 – – – – – – – 162 0.105 0.033 – – – – – – – – – – – – – – – – – 166 – – – – – – – – – – – – – – 0.083 – – – – 167 – – – – – – – 0.085 – 0.019 – 0.028 – – 0.042 – – – – 169 – – – – – – – 0.015 – – – – – – – – – – – D17Mit233 102 – – 0.048 – – 0.063 – 0.069 0.035 0.074 0.056 – – 0.197 – 0.071 0.063 – 0.105 104 – 0.033 – – – – – – – – – – – – – 0.036 0.016 – – 106 0.026 0.300 – 0.150 0.375 0.639 0.546 0.585 0.148 0.111 0.361 0.654 0.329 – 0.071 0.203 0.769 0.053 108 – 0.100 0.024 0.100 – – – 0.008 0.010 – 0.056 – – – – – 0.016 – – 110 0.237 – 0.810 – – – – – – 0.093 – – – 0.303 0.050 0.179 0.078 – 0.158 112 – – – – – – – – – 0.056 0.056 – – – – – – – 0.026 114 0.026 – – 0.250 – – – – – – – – – – – – – – 0.053 116 0.132 0.167 0.050 0.300 0.186 0.115 0.120 0.389 0.333 0.083 0.192 0.066 0.150 – – – 0.053 118 – – – – 0.100 0.125 0.083 0.015 – – 0.056 0.028 – 0.079 – 0.036 0.141 0.039 0.053 120 0.316 0.033 0.048 0.450 – – – 0.131 0.190 0.148 0.056 0.250 – 0.013 0.350 0.250 0.188 0.039 0.105 122 0.053 0.233 0.071 – – – – 0.008 0.035 0.019 0.222 0.250 – – 0.100 0.143 0.266 – 0.158 124 – – – – 0.100 0.250 0.278 0.108 0.025 – – 0.028 0.154 – – – 0.031 – – 126 0.211 – – – 0.100 – – – – 0.037 – – – 0.013 – 0.071 – – 0.053 128 – – – – 0.100 – – – – 0.037 0.056 – – – 0.350 0.143 – 0.115 0.026 130 – 0.133 – – – – – – – – – – – – – – – 0.039 0.079 132 – – – – – – – – – – – – – – – – – – 0.079 136 – – – – 0.300 – – – – – – – – – – – – – –
Mean allele no./locus 6.4 6.2 4.6 3.6 4.0 6.0 4.4 9.8 9.0 9.4 6.0 5.6 5.0 5.0 6.0 5.4 8.4 5.2 9.2 H–W exact testa ∗ nsb ∗∗∗ ns ns ns ns ∗∗∗ ∗∗∗ ∗∗∗ ns ∗∗∗ ∗∗∗ ∗ ∗∗ ns ∗∗∗ ∗∗∗ ns aSignificant levels of H–W probability test.
bP >0.05. ∗0.01 < P < 0.05. ∗∗0.001 < P < 0.01. ∗∗∗P <0.001.
