Mutation rate heterogeneity and the generation of allele diversity at the human minisatellite MS205 (D16S309)
Mutation rate heterogeneity and the generation of allele diversity at the human minisatellite MS205 ( D16S309 )Celia A. May*, Alec J. Jeffreys and John A. L. Armour+
Department of Genetics, University of Leicester, Leicester LE1 7RH, UK
Received July 3, 1996;Revised and Accepted August 21, 1996
Many tandemly repeated minisatellite loci display extreme levels of length variation as a consequence of high rates of spontaneous germline mutation altering repeat copy number. Direct screening for new allele lengths by small-pool PCR has shown that instability at the human minisatellite locus MS205 (D16S309) is largely germline specific and usually results in the gain or loss of just a few repeat units. Structural analysis of the order of variant repeats has shown that these events occur preferentially at one end of the tandem array and can result in complex rearrangements including the inter-allelic transfer of repeat units. In contrast, putative mutants recovered from somatic DNA occur at a substantially lower rate and are simple and non-polar in nature. Germline mutation rates vary considerably between alleles, consistent with regulation occurring in cis. Although examination of DNA sequence polymorphisms immediately flanking the minisatellite reveals no definitive associations with germline mutation rate variation, differences in rate may be paralleled by changes in mutation spectrum. These findings help to explain the diversity of MS205 allele structures in modern humans and suggest a common mutation pathway with some other minisatellites.
Minisatellite loci have become firmly established as highly informative genetic markers because of their frequently extreme length polymorphism (1 -5 ). Allelic variation results from high rates of spontaneous germline mutation that change tandem repeat copy number. For the most variable loci, mutation rates to new length alleles can be directly measured in pedigrees, and estimates as high as 13% per gamete have been recorded in humans (6 ,7 ).
All human minisatellites examined to date also exhibit variation in the interspersion pattern of subtle repeat unit sequence variants along the tandem array. This additional source of variability can reveal levels of allelic diversity far greater than those resolved by allele length measurement. This variation has been successfully exploited using MVR-PCR (Minisatellite Variant Repeat mapping by PCR) (8 -13 ) to provide a powerful approach to analysing minisatellite mutation. Comparisons of allele structures defined by MVR-PCR can give insights into the general rules governing minisatellite instability (8 ,13 -16 ), whilst comparisons between de novo mutants and their progenitors can help define the processes involved (9 ,17 ). However, screening human pedigrees for new-length alleles is an inefficient method of identifying germline mutations and gives no information on the mutation rate of the individual. We therefore developed an alternative approach, termed small-pool PCR (SP-PCR), capable of detecting unlimited numbers of minisatellite mutants directly in germline (sperm) DNA (17 ).
In SP-PCR, locus-specific primers external to the tandem array are used to amplify minisatellite molecules in dilute aliquots of sperm DNA. The `pool' size of each aliquot (usually between 40 and 160 haploid genomes) is such that a single new-length molecule represents a sufficient proportion of the total to be identified after PCR amplification by gel electrophoresis and Southern blot hybridization. To screen large numbers of gametes, multiple aliquots are amplified for each DNA.
This strategy was originally developed for minisatellite MS32 (D1S8) and to date over 1000 MS32 length mutants have been identified, compared with only 11 found in pedigrees (A.J.Jeffreys, unpublished data). The mean mutation rate has been estimated at 0.8% per sperm, comparable with that seen in pedigrees, with a strong bias towards gains of repeat units (17 ). Most sperm mutations are germline-specific and involve small inter-allelic gene conversion-like changes at one end of the minisatellite array. However, attempts to establish general mutation principles from a single locus should be treated with caution, not least because MS32 has an atypical interstitial chromosomal location (18 -20 ), but also because our current data are restricted to the shortest alleles at this locus (17 ).
To gain further insights into tandem repeat instability, we have extended SP-PCR to another well-characterized minisatellite, MS205 (D16S309). This locus is located in the proterminal DNA of the short arm of chromosome 16 (21 ) and has an estimated heterozygosity of 99.7% in Europeans. MS205 has been extensively used in paternity case work and has a sex-averaged germline mutation rate of ~0.4% per gamete but with a strong paternal mutation bias (11/12 mutants, giving a male mutation rate of 0.74% per sperm). These features, plus relatively small alleles (8-87 repeats of a 45-54 bp repeat unit) (14 ), make MS205 ideal for mutation study by SP-PCR. All alleles can be amplified by PCR and the analysis of sperm DNA alone should provide a representative view of germline turnover at this locus. Furthermore, MVR-PCR has already been developed at MS205 and used to determine the complete internal structures of more than 300 alleles from around the world (14 ). This diversity survey has revealed instances of alleles with unusually high population frequencies, suggesting possible allele-specific suppression of instability.
We now present an analysis of MS205 mutation using SP-PCR. In conjunction with MVR-PCR, this approach has provided detailed insights into the mutation dynamics of this locus. The results are discussed in relation to allele diversity studies and mutation processes at other minisatellite loci.
Initial attempts to transfer the SP-PCR conditions used for MS32 to MS205 resulted in marked pool to pool variation in amplification efficiency. This was overcome by supplementing the PCR reactions with glycerol (22 ). This co-solvent may facilitate denaturation of the locus which is located in a highly G/C rich domain and is composed of very G/C rich repeats (8 ).
Sperm DNA from 10 donors was screened for MS205 length changes using the modified SP-PCR procedure (see Materials and Methods; Fig. 1 a). In addition to strong signals from the progenitor alleles, a number of putative length-changed mutants were also detected (Fig. 1 b). The signals from these latter PCR products were comparable with those of the progenitor alleles at single molecule dilutions, suggesting that they had amplified from the first cycle of PCR. As with the MS32 SP-PCR assay, prolonged autoradiography failed to detect any additional products as might be expected from PCR artefacts arising at various stages during amplification. Further, the frequency of mutant products was proportional to sperm DNA input over the range of 40-160 molecules per pool. These findings suggest that most or all new-length products are derived from authentic de novo mutants in sperm.
Most mutation events detected both in pedigrees (17 ) and in single sperm involve small length changes. Sperm mutants from each donor were therefore assigned to the progenitor allele closer in size; subsequent structural analysis by MVR-PCR proved that most of these assignments were correct (see below). The combined mutation spectrum so deduced from all 10 sperm donors is shown in Figure 2 a.Changes of +-1 repeat unit could only be scored for alleles <35 repeats in length. Overall, slightly more events resulted in the gain than loss of repeats but this did not represent a significant bias ([chi]2 = 0.484, P >0.05 with 1 d.f.). In contrast, both events scored in blood resulted in loss of repeat units (Fig. 2 b). Deletions also predominated amongst the low frequency events detected by SP-PCR of 13 lymphoblastoid DNAs (Fig. 2 c) during a separate study to assess the use of minisatellites as a reporter system for environmental monitoring (C.A.May and J.A.L.Armour, unpublished data). At present, it is not possible to be certain that these rare events detected in somatic (blood, lymphoblastoid) DNA represent authentic mutants rather than PCR artefacts.
Mutation events were further characterized by comparing the internal structures of mutant and progenitor alleles by MVR-PCR mapping (8 ,14 ). The mapping procedure discriminates between two alternatives at a single base position, allowing repeat units to be classified as either `A'- or `T'-type. Sequence analysis has shown that this is not the only polymorphic site and, in support of this, reproducible intensity differences are observed for both repeat types on mapping autoradiographs. These intensity data were used to enhance the information content of the `two-state' system and in many cases allowed more definitive conclusions to be drawn about the position and type of mutational change (Fig. 3 a).
Mutation rates varied between donors, with rates ranging from 0.13% to 1.05% per sperm per individual,excluding changes of +-1 repeat since these could not be scored for all allele lengths (see above). Reproducible estimates were obtained from a given semen DNA sample on different occasions, and also from two samples from the same donor, provided and prepared for analysis 4 years apart. Nine of the sperm donors carried at least one MS32 allele suitable for SP-PCR analysis, allowing a rough comparison to be made between the mutation rate estimates for these two minisatellite loci. Although the same individual exhibited the highest mutation rate at both loci, no other correlation of mutation rates between these two loci was apparent (data not shown).
For the nine sperm donors heterozygous for array length at MS205, it was possible to examine mutation rates at the level of individual alleles. Considerable rate heterogeneity was noted [G (William's) = 85.76, P <0.001 with 17 d.f.], with estimates ranging from 0.05 to 0.66% per sperm. Significantly different mutation rates were noted for the two alleles carried by the same sperm donor for four of the nine individuals (including one individual sampled twice over a 4 year period, Figure 1 b), suggesting that mutation rates at MS205 may be modulated in cis (Fig. 5 a). No association between mutation rate and minisatellite array length was found over the range of allele sizes screened (22-62 repeats) (Fig. 5 b). Similarly, no obvious relationship existed with the two-state internal structures of the alleles (Fig. 6 ), in terms either of their 5' structure, which can be used to establish different allele lineages, or their complexity as measured by the frequency of switching from one repeat unit to another along the array (14 ). Given that additional sites of variation exist within the repeat units, it remains possible that other aspects of tandem array structure may correlate with mutation rate variation.
Four of the sperm allele structures had previously been observed in our survey of MS205 MVR maps, in which most alleles were seen only once (14 ) (Fig. 6 ). Consistent with the predictions of neutral mutation and random drift in allele frequency, low mutation rates were associated with three of these alleles (s, t and v). Allele t from a northern European, with a mutation rate almost an order of magnitude lower than the mean (0.05%), has been seen at very high frequency (0.21-0.75) in Saami (i.e. Lapps), Surui (Lupi speakers from western Amazonia) and Japanese (14 ). Allele s, also from a northern European, and with the second lowest recorded mutation rate of 0.11%, has been seen three times amongst 106 North European alleles from the Centre d'Etude du Polymorphisme Humain (CEPH) panel and once amongst 18 Basque alleles. Allele v (northern European) has been observed only once before, in another North European, and has a mutation rate comparable with that of allele t (i.e. 0.05%). However, allele k, now seen twice amongst 42 Zimbabwean alleles, was found to have a mutation rate rather closer to the observed mean (0.23%). These associations between low mutation rates and high population frequencies provide strong evidence that mutation rate heterogeneity (Fig. 5 ) cannot be wholly attributed to fluctuating rates within individuals. Instead, at least part of the heterogeneity results from allele-specific differences in mutation rate and implies the existence of cis-acting regulators of MS205 instability.
A subset of MS32 alleles shows a profoundly suppressed mutation rate associated with, and probably caused by, a single base transversion in the DNA flanking the unstable end of the tandem repeat array (15 ). Although the flanking regions of MS32 and MS205 show no significant similarity in their DNA sequence (A.J.Jeffreys and J.A.L.Armour, unpublished data), the sperm donors were nonetheless typed for 10 polymorphic sites close to the MS205 array [(8 ) and C.A.May, unpublished data] to test for possible haplotype-specific mutation rate variation (Figs 1 a and 6 ). Seven variants, at -998, -680, -575, -565, -175, +10 and +942, showed little or no variation amongst the sperm alleles. The -8 variant showed no correlation with mutation rate. The G variant at the more informative +266 site tended to be associated with lower mutation rates, though this was not significant (P = 0.07, Kruskal-Wallis test). At the -382 site, the rarer C variant was linked to allele c with the highest mutation rate, and was also present in the heterozygous state in the individual displaying the highest mutation rate. Unfortunately, this individual is homozygous for alleles of the same length (a and b) and it has not proved possible to develop reliable allele-specific SP-PCR from the -382 site needed to assign mutation rates to the individual alleles.
The most unstable MS205 allele (c) showed an unusual spectrum of mutants. The 11 mutants typed contained three instances of the same mutant structure each detected 2-3 times; at least two of these had apparently involved polarised inter-allelic transfers (Fig. 7 a). In contrast, only two instances of pairs of isolates of the same mutant were observed amongst all other alleles analysed. This suggests a possible shift in mutation process at allele c, either towards pre-meiotic inter-allelic recombination between homologous chromosomes to generate true mutational mosaicism in sperm DNA, or towards highly polarised and canalised inter-allelic conversion resulting in the recurrent generation of the same mutant structure, possibly at meiosis.
Putative mutants from the relatively stable allele s were also unusual (Fig. 7 b). In contrast with mutants from the other MS205 allele of this individual (f in Fig. 3 aiii), these low frequency events were evenly distributed along the tandem array. For two, s1 and s2, typical complex rearrangements had occurred suggesting that they are authentic mutants. The remaining involved simple deletion events similar to those seen in blood and lymphoblastoid DNA and may represent PCR artefacts. If so, then the true mutation rate at this allele may be as low as 0.02% per sperm. Even though mutation rate is suppressed, this allele can still apparently act as a donor of repeat units to the other unstable allele in this individual (mutants f11 and possibly f3, Fig. 3 aiii).
Although hypervariable minisatellites have been recognised as highly informative genetic markers for over a decade and are widely exploited in DNA typing systems (2 -4 ,11 ), it is only now that we are beginning to understand the mutation processes underlying the variation seen at these loci (8 ,9 ,14 ,15 ,17 ). The development of MVR-PCR has been fundamental to this work, but until recently most studies have been restricted to the small number of de novo mutation events identified in pedigrees. These limitations were overcome at the human minisatellite MS32 by the direct screening of sperm DNA for new-length alleles by SP-PCR, enabling individual mutation rates to be determined (15 ,17 ). We have now successfully used this technique at a second minisatellite, MS205, and have evidence that this approach is applicable to further loci (unpublished data).
Two features of the locus make MS205 particularly suitable for SP-PCR analysis. First, the strong paternal bias in mutation rate observed in pedigrees means that the analysis of sperm DNA alone provides a comprehensive view of MS205 turnover. In contrast, no parental bias has been noted for MS32, and from our limited pedigree data, we cannot exclude the possibility of qualitative differences in mutation process between the sexes (17 ); unfortunately, there is no practical way of accessing the human female germline for SP-PCR analysis. Second, given the size distribution of allele lengths at MS205, it is possible to carry out SP-PCR analysis on most or all DNA samples. Indeed, the alleles currently assayed by SP-PCR at MS205 cover over 85% of the range of known allele lengths. At MS32, most alleles are too long (>6 kb) for analysis by SP-PCR (17 ).
Several lines of evidence suggest that the majority of new-length products detected at MS205 by SP-PCR of sperm DNA are true germline mutants. They show the appropriate intensity expected for products of a single molecule, their frequency is proportional to DNA input, the mean mutation rate in sperm is similar to estimates from pedigrees, and the rate is substantially higher than that seen in blood and lymphoblastoid DNA. Further, significant differences exist between the size distributions and internal structures of putative mutants obtained from sperm and somatic DNA. The majority of germline mutants exhibit changes at the 3' end of the tandem repeat array, as predicted from polarity in allelic variation (8 ,14 ) and as seen for all pedigree mutants characterized to date (17 ), and many involve complex rearrangements of repeat units. These sperm-specific features cannot be explained by PCR artefact processes.
The structural analysis of relatively large numbers of MS205 mutants isolated by SP-PCR allows more detailed comparisons of mutation to be made with other loci, in particular with MS32 (Table 1 ). Despite the differences in chromosomal location, repeat unit lengths, allele size ranges and parental bias in mutation rate, sperm mutation events at both loci show some remarkable similarities. For both loci, male mutation rate is largely germline specific with rates in sperm apparently modulated in cis but independent of array length. Most changes result in the gain or loss of just a few repeat units and there is an overall tendency towards gain events. Inter-allelic transfer of repeat units has been noted for both MS32 and MS205, although it is difficult to directly compare the frequency of such events as this requires similarly informative MVR mapping systems for the two loci. Nonetheless, the polar and often complex rearrangements of repeat units seen at MS205 are also typical of MS32 and have been seen amongst MS31A (D7S21) pedigree mutants (17 ).Polarity at all three loci for both intra- and inter-allelic events suggests not only a common mode of mutation at these loci, but also a common pathway for intra- and inter-allelic mutation.Less striking parallels can be drawn with the highly unstable minisatellite CEB1; although inter-allelic events do occur and tend to be clustered at one end of the tandem array, the predominating intra-allelic events are less obviously polar, possibly reflecting the action of additional or modified mutation processes (9 ).
We have previously proposed a model for minisatellite mutation based on current ideas for the initiation of meiotic recombination in yeast (17 ). This model satisfactorily accounts for the clustering of mutation events, and the apparent length-independence of mutational activity observed for MS32 and MS205 alleles. It is envisaged that a cis-acting mutation initiator in the flanking DNA directs a double-strand break (DSB) towards one end of the recipient tandem array; this break then expands to form a gap extending into the repeats and the gap is subsequently repaired using either the sister chromatid or homologous chromosome as a template. Additional breaks introduced into the mutation intermediates might account for the more complex rearrangements.
At all three `polar' minisatellites, inter-allelic exchange tends to involve the gain of repeat units from the corresponding region of the donor, suggesting that donor and recipient arrays are paired in register over the ultravariable end prior to transfer. This pairing of homologues might also promote exchange of markers upstream of the unstable end of the tandem array. Indeed, comparisons of haplotypes of flanking polymorphisms between closely related alleles suggests a high frequency of exchange near the 3' end of MS205 indirectly estimated at 0.07% per gamete or more (8 ). However, the present study has failed to reveal direct evidence of exchange of closely linked 3' flanking markers accompanying minisatellite length mutation, and sets an upper limit of 0.04% per sperm for such events. It therefore remains possible that haplotype switching arises from events primarily restricted to the flanking DNA.
The discovery of a G to C transversion 48 bp upstream of the MS32 hotspot which apparently suppresses mutation in cis provides direct evidence for the role of flanking DNA in minisatellite mutation (15 ). Although the C variant (`O1C') disrupts an 18 bp same-strand palindrome, the significance of this remains unclear as sequence comparisons, now of the order of several kilobases, have failed to identify this or any other specific DNA element in the vicinity of either MS205 or MS31A (J.A.L.Armour, A.J.Jeffreys and D.Neil, unpublisheddata). Thus, it is possible that minisatellite instability is regulated by the surrounding DNA in a non-sequence directed fashion. Consistent with this, studies in yeast have recently shown that the DSBs associated with specific meiotic recombination hotspots do not occur at consensus sequence sites but that their position and frequency may be influenced by local higher-order chromatin structure (24 -26 ). Under this model, recombination occurs preferentially in `open' regions of the chromatin since these are the most accessible to the recombination machinery. Interestingly, this `accessibility' model has also been proposed to account for both imprinting and sex-specific differences in recombination frequency in humans (27 ), and by analogy might explain the sex-bias in mutation rate observed at some human minisatellites (28 ).
MS32 O1C alleles and the MS205 allele s are relatively stable in the germline, consistent with their high frequencies within and between some populations. Although both types of allele can act as donors in inter-allelic exchange, the putative mutants derived from each show less sign of mutational polarity than their unstable counterparts. For MS32 O1C alleles, this loss of polarity is supported by population diversity studies (15 ). MS205 allele s is linked to the rare T form at the +10 site, raising the possibility that the +10T variant may be functionally equivalent to O1C, perhaps by blocking mutation initiation but not subsequent pairing and gap-repair. This can only be tested by characterizing other +10T alleles on different haplotypic backgrounds.
The existence of unusually stable MS205 alleles helps to explain some of the patterns of allele diversity seen at this locus. We have previously argued that the geographical distribution of the most common MS205 allele, CE1 (identical in structure to allele t in this study), may be indicative of recent divergence and expansion of the Japanese, Saami and Surui populations (14 ). However, SP-PCR has shown that CE1 is relatively stable in the germline, a situation that may have allowed the allele to reach the observed high frequencies in several populations by drift alone without being disrupted by mutation.
We recently showed that non-African populations carry limited sets of the MS205 allele lineages seen in African populations, consistent with a recent African origin for modern humans. Using the mean mutation rate estimate from pedigree analysis and the observed level of allele-sharing between African and non-African populations, we estimated a much more recent split than previous studies (14 ). To account for the discrepancy, we suggested that the shared alleles may exhibit an unusually low mutation rate. The present study shows that mutation rates do indeed vary over an order of magnitude or more. Future direct analysis of mutation rates at relevant alleles should allow a more precise estimate of the divergence time between African and non-African populations.
male germline mutation rate estimate per sperm ([mu]):
pedigrees
1.00%
0.74%
SP PCR
0.81%
0.50%
somatic mutant frequency per haploid genome3
<= 0.06%
<= 0.025%
mean size change in sperm (number of repeats)
+2.9
-0.5
ratio of gains to losses in sperm
3:1
1.1:1
position of sperm mutation events
polar
polar
frequency of inter-allelic transfer of repeat units in sperm
>80%
>16%
length changes accompanied by exchange of flanking markers
0-0.015%
0-0.04%
(95% C.I. per sperm)
sperm [mu] variation at level of individual alleles (range)
0.00-0.88%4
0.05-0.66%5
correlation of sperm [mu] with allele length
none observed6
none observed7
correlation of low sperm [mu] with high population frequencies
yes
yes
correlation of sperm [mu] variation with allele structure
none observed
none observed
correlation of sperm [mu] variation with flanking DNA sequences
yes
none observed
1As determined from Southern blot profiles of the CEPH panel.2As determined from pedigree data.3As determined by SP-PCR of male blood DNA.4Alleles 22-164 repeats, gains of 3-20 repeats only.5Alleles 22-62 repeats, excludes changes of +-1 repeat unit.6Allele range of 22-164 repeats.7Allele range of 22-62 repeats.
Sperm samples from 10 donors were analysed in this study, including seven North Europeans, two Pakistanis and a Zimbabwean. DNA was prepared from sperm, and where appropriate from lymphocytes, as described previously (17 ).
For the nine length heterozygotes, the progenitor alleles were amplified to levels detectable by ethidium bromide staining and recovered from agarose gels by standard procedures (8 ). For the length homozygote, limiting DNA dilutions were used to recover single molecule PCR products derived from each progenitor (29 ). All structures were internally mapped by MVR-PCR in both forward and reverse directions (8 ,14 ).
Nine polymorphisms in the DNA flanking the MS205 array had been previously identified (8 ). Genotyping of the 10 sperm donors for sites -998, -382, -175, +10, +266 and +942 was carried out as described. For heterozygous individuals, the association of particular minisatellite alleles and flanking alleles was established by allele-specific PCR (30 ). Sites -680 and -575 were genotyped by amplifying the DNA segment between 205H and 205J and digesting with AvaII and NcoI, respectively (8 ). Phase between these rare sites and the minisatellite alleles carried by one sperm donor (i.e. alleles j and r) was established by PCR amplification with the allele-specific primers for the +266 site and 205H, followed by reamplification with 205J and digestion of the resulting single allele segments with AvaII and NcoI. [First round PCR: 100 ng genomic DNA in a 10 [mu]l volume supplemented to 6% (w/v) glycerol, primers at 0.2 [mu]M, at 95oC 1 min, 73oC 1.5 min for 30 cycles. Second round: PCR products diluted 1 in 100, 1 [mu]l used to seed 20 [mu]l reaction with primers 205J and 205H at 0.25 [mu]M for 20 cycles as before.] Status and phase at the -565 site were established by allele-specific oligonucleotide typing (31 ) of diploid and haploid 205H-205J amplicons respectively using ASO565C (5'-GGCACATTCACCGGCCA-3') and ASO565T (5'-GGCACATTTACCGGCCA-3').
5' and 3' flanking DNA were scanned for further polymorphic sites by Taq cycle sequencing (32 ) of amplicons spanning bases from -1 to -406 (205A-205TAGRT and 205J-205H) and from +1 to +343 (205TAGA-205GRUN, 205T-205LA2/LG, 205X-205Y) for 14 of the 20 alleles in the survey (205GRUN = 5'-TGCAGAGTCCCCCCCCCCGCCGAG-3'; 205X = 5'-TAAGCTCCGGATAACCGGTGGGTG-3'; 205Y = 5'-AGCGTGTGACCAGACCACG TTCCG-3'). A single additional polymorphic site at -8 was identified; the rare C form creates an NlaIV site, providing a PCR-RFLP assay on amplicon 205A-205TAGRT.
Multiple (40-255) 7 [mu]l reactions each consisting of the PCR buffer described previously (10 ) plus primers 205A and 205B (8 ) at a final concentration of 0.2 [mu]M, glycerol at 6% (w/v) and 0.7 U Taq polymerase (Advanced Biotechnologies) were used to amplify on average 110 haploid genomes (330 pg) of MboI digested sperm DNA per reaction. Reactions were overlaid with paraffin oil and amplified on a GeneAmp PCR system 9600 thermal cycler (Perkin Elmer Cetus) at 96oC 50 s, 65oC 45 s, 72oC 5 min for 25 cycles, followed by a chase at 65oC 45 s, 72oC 10 min. Aliquots of each SP-PCR reaction (2.7 [mu]l) were electrophoresed through 40 cm horizontal 0.8% SeaKem LE agarose gels in 0.5* TBE (44 mM Tris-borate pH 8.3, 1 mM EDTA) until the 872 bp [Phi]X174.HaeIII marker was at the end of the gel, then blotted onto Hybond Nfp (Amersham) and hybridized with a 32P-oligo-labelled pMS205.2 probe (8 ).
The frequency of mutants per progenitor molecule was determined after estimating the efficiency of single molecule amplification. Two cocktails were made, each consisting of five sperm donor DNAs; each DNA was diluted to 20 pg/[mu]l in 5 mM Tris-HCl pH 7.5 plus 0.1 [mu]M carrier primer 205A. For each cocktail, 40* 7 [mu]l reactions each containing 5.14 pg of each of the DNAs were amplified with primers 205A and 205B as above. The products were resolved on agarose gels and detected by Southern hybridization. The mean number of amplifiable molecules (m) per 5.14 pg input was estimated from the Poisson distribution, expressed as z = e-m , where z is the frequency of negative PCR reactions for a given allele. Assuming 3 pg per haploid genome, the mean single molecule PCR efficiency was found to be 61%. This efficiency was apparently array length independent and comparable across all alleles tested.
Pools containing mutant molecules were diluted 1 in 20 with distilled water and 1 [mu]l of this used as input material for hemi-nested PCR. Primers 205B and 205C (8 ) were used at a final concentration of 0.25 [mu]M in a 20 [mu]l volume consisting of the standard PCR buffer supplemented with glycerol to 6% (w/v). Amplification was carried out at 96oC 50 s, 70oC 45 s, 72oC 5 min for 12 cycles. PCR products were electrophoresed as before and gel slices spanning the region of each mutant were excised and crushed in 150 [mu]l of distilled water. The DNA was released into solution by freeze/thaw treatment and 1 [mu]l aliquots were reamplified for 25 cycles as above. Mutant bands, now detectable by ethidium bromide staining after gel electrophoresis, were recovered and aliquots corresponding to ~20 pg DNA were used in MVR reactions (see above).
Mutant and progenitor MVR maps were aligned by eye, assuming the simplest of changes where possible. The location of each event along the progenitor allele was determined and for comparative purposes, each was expressed as a fraction of the total number of positions such a change could theoretically occupy: scores of 0 were indicative of events at the extreme 5' end of an allele, whilst scores of 1 indicated events at the extreme 3' end.
For inter-allelic events, the analysis was in terms of the `recipient' allele, which was taken as the progenitor allele that had contributed most significantly to the mutant structure. In the case of tandem duplications it was not possible to distinguish between the original and duplicated version of the motif, so the event was assumed to have occurred at the midpoint of the range. Similarly, the precise location of deletion events could not always be determined and in such cases the midpoint of the most central of all possible positions was used.
Insertion events could theoretically occur anywhere along the entire length of the progenitor allele, but the range of possible positions for deletion events was constrained by the size of the event concerned. For example, if the midpoint of an 11 repeat deletion was placed at repeat 31 along a 43 repeat unit long progenitor, then the relative location would be expressed as 31/32 since the midpoint of a deletion of this size could only fall between repeat numbers 6 and 38 (i.e. a range of just 32 repeats).
Mutants characterized by MVR-PCR were not a fully random selection of events detected by SP-PCR; alleles displaying mutation rates at the extreme ends of the range were preferentially targeted, resulting in alleles with average mutation rates being under-represented. This was taken into account in Figure 4 a by making an adjustment based on the relative frequency and mutation rates of `high', `average' and `low' mutating alleles (details available on request).
We thank Jane Blower for semen samples, Jane Cole for lymphoblastoid cell line DNA, Jon Wetton, John Brookfield, Yuri Dubrova, Nicola Royle and members of A.J.J.'s laboratory for helpful advice and comments. This work was made possible by grants from the CEC (EV5V-CT910585) and the Wellcome Trust (038225/Z/93/Z).The work of A.J.J. is supported in part by an International Research Scholar's award from the Howard Hughes Medical Institute and by grants from the Medical Research Council and Royal Society.
1 Jeffreys, A.J., Wilson, V. and Thein, S.L. (1985) Hypervariable `minisatellite' regions in human DNA.Nature314, 67-73.MEDLINE Abstract
2 Jeffreys, A.J., Wilson, V. and Thein, S.L. (1985) Individual-specific `fingerprints' of human DNA.Nature316, 76-79.MEDLINE Abstract
3 Jeffreys, A.J., Turner, M. and Debenham, P. (1991) The efficiency of multilocus DNA fingerprint probes for individualisation and establishment of family relationships, determined from extensive casework.Am. J. Hum. Genet. 48, 824-840.MEDLINE Abstract
4 Monckton, D.G. and Jeffreys, A.J. (1993) DNA profiling. Curr. Opin. Biotech.4, 660-664.
5 Nakamura, Y., Leppert, M., O'Connell, P., Wolff, R., Holm, T., Culver, M., Martin, C., Fujimoto, E., Hoff, M., Kumlin, E. and White, R. (1987) Variable number of tandem repeats (VNTR) markers for human gene mapping. Science235, 1616-1622.MEDLINE Abstract
6 Jeffreys, A.J., Royle, N.J., Wilson, V. and Wong, Z. (1988) Spontaneous mutation rates to new length alleles at tandem-repetitive hypervariable loci in human DNA. Nature332, 278-281.MEDLINE Abstract
7 Vergnaud, G., Mariat, F., Apiou, F., Aurias, A., Lathrop, M. and Lauthier, V. (1991) The use of synthetic tandem repeats to isolate new VNTR loci: cloning of a human hypermutable sequence.Genomics11, 135-144.MEDLINE Abstract
8 Armour, J.A.L., Harris, P.C. and Jeffreys, A.J. (1993) Allelic diversity at minisatellite MS205 (D16S309): evidence for polarized variability.Hum. Mol. Genet.2, 1137-1145.
9 Buard, J. and Vergnaud, G. (1994) Complex recombination events at the hypermutable minisatellite CEB1 (D2S90). EMBO J. 13, 3203-3210.MEDLINE Abstract
10 Jeffreys, A.J., Neumann, R. and Wilson, V. (1990) Repeat unit sequence variation in minisatellites: a novel source of DNA polymorphism for studying variation and mutation by single molecule analysis. Cell60, 473-485.MEDLINE Abstract
11 Jeffreys, A.J., MacLeod, A., Tamaki, K., Neil, D.L. and Monckton, D.G. (1991) Minisatellite repeat coding as a digital approach to DNA typing. Nature354, 204-209.MEDLINE Abstract
12 Tamaki, K., Monckton, D.G., MacLeod, A., Allen, M. and Jeffreys, A.J. (1993) Four-state MVR-PCR: increased discrimination of digital DNA typing by simultaneous analysis of two polymorphic sites within minisatellite variant repeats at D1S8. Hum. Mol. Genet.2, 1629-1632.MEDLINE Abstract
13 Neil, D.L. and Jeffreys, A.J. (1993) Digital DNA typing at a second hypervariable locus by minisatellite variant repeat mapping. Hum. Mol. Genet. 2, 1129-1135.MEDLINE Abstract
14 Armour, J.A.L., Anttinen, T., May, C.A., Vega, E.E., Sajantila, A., Kidd, J.R.,Kidd, K.K., Betranpetit, J., Pääbo, S. and Jeffreys, A.J. (1996) Minisatellite diversity supports a recent African origin for modern humans.Nature Genet.13, 154-160.
15 Monckton, D.G., Neumann, R., Guram, T., Fretwell, N., Tamaki, K., MacLeod, A. and Jeffreys, A.J. (1994) Minisatellite mutation rate variation associated with a flanking DNA sequence polymorphism.Nature Genet. 8, 162-170.MEDLINE Abstract
16 Huang, X-L., Tamaki, K., Yamamoto, T., Suzuki, K., Nozawa, H., Uchihi, R., Katsumata, Y. and Neil, D.L. (1996) Analysis of allelic structures at the D7S21 (MS31A) locus in the Japanese, using minisatellite variant repeat mapping by PCR (MVR-PCR). Ann. Hum. Genet. 60, 271-279.MEDLINE Abstract
17 Jeffreys, A.J., Tamaki, K., MacLeod, A., Monckton, D.G., Neil, D.L. and Armour, J.A.L. (1994) Complex gene conversion events in germline mutation at human minisatellites. Nature Genet. 6, 136-145.MEDLINE Abstract
18 Armour, J.A.L., Wong, Z., Wilson, V., Royle, N.J. and Jeffreys, A.J. (1989) Sequences flanking the repeat arrays of human minisatellites: association with tandem and dispersed repeat elements. Nucleic Acids Res. 17, 4925-4935.
19 Royle, N.J., Clarkson, R.E., Wong, Z. and Jeffreys, A.J. (1988) Clustering of hypervariable minisatellites in the proterminal regions of human autosomes. Genomics3, 352-360.MEDLINE Abstract
20 Wong, Z., Wilson, V., Patel, I., Povey, S. and Jeffreys, A.J. (1987) Characterization of a panel of highly variable minisatellites cloned from human DNA. Ann. Hum. Genet. 51, 269-288.MEDLINE Abstract
21 Royle, N.J., Armour, J.A.L., Webb, M., Thomas, A. and Jeffreys, A.J. (1992) A hypervariable locus D16S309 located at the distal end of 16p. Nucleic Acids Res. 20, 1164.MEDLINE Abstract
22 Barnes, W.G. (1994) PCR amplification of up to 35kb DNA with high fidelity and high yield from [lambda] bacteriophage templates.Proc. Natl Acad. Sci. USA 91, 2216-2220.
23 Jeffreys, A.J., Allen, M.J., Armour, J.A.L., Collick, A., Dubrova, Y., Fretwell, N., Guram, T., Jobling, M., May, C.A., Neil, D.L. and Neumann, R. (1995) Mutation processes at human minisatellites. Electrophoresis16, 1577-1585.MEDLINE Abstract
24 Lui, J., Wu, T. and Lichten, M. (1995) The location and structure of double-strand DNA breaks induced during yeast meiosis: evidence for a covalently linked DNA-protein intermediate. EMBO J.14, 4599-4608.
25 Wu, T. and Lichten, M. (1994) Meiosis-induced double-strand break sites determined by yeast chromatin structure. Science 263, 515-518.MEDLINE Abstract
26 Ohta, K., Shibata, T. and Nicolas, A. (1994) Changes in chromatin structure at recombination initiation sites during yeast meiosis. EMBO J. 13, 5754-5763.MEDLINE Abstract
27 Paldi, A., Gyapay, G. and Jami, J. (1995) Imprinted chromosomal regions of the human genome display sex-specific meiotic recombination frequencies. Curr. Biol. 5, 1030-1035.MEDLINE Abstract
28 Henke, J. and Henke, L. (1995) Recent observations in human DNA-minisatellite mutations. Int. J. Legal Med. 107, 204-208.MEDLINE Abstract
29 Monckton, D.G. and Jeffreys, A.J. (1991) Minisatellite `isoallele' discrimination in pseudohomozygotes by single molecule PCR and variant repeat mapping. Genomics11, 465-467.MEDLINE Abstract
30 Monckton, D.G., Tamaki, K., MacLeod, A., Neil, D.L. and Jeffreys, A.J. (1993) Allele-specific MVR-PCR analysis at minisatellite D1S8. Hum. Mol. Genet. 2, 513-519.MEDLINE Abstract
31 Handelin, B. and Shuber, A.P. (1995) In Dracopoli, N, et. al. (eds),Current Protocols in Human Genetics. John Wiley and Sons, pp 9.4.1-9.4.5.
32 Slatko, B.E., Albright, L.M. and Tabor, S. (1994) In Ausubel, F.M. et. al. (eds), Current Protocols in Molecular Biology. John Wiley and Sons pp 7.4.11.
*To whom correspondence should be addressed+Present address: Department of Genetics, University of Nottingham, Nottingham NG7 2UH, UK
This page is maintained by OUP admin. Last updated Thu Oct 31 15:28:51 GMT 1996. Part of the OUP Journals World Wide Web service.Copyright Oxford University Press, 1996
