Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (7)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Langdon, J. A.
Right arrow Articles by Armour, J. A.L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Langdon, J. A.
Right arrow Articles by Armour, J. A.L.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Human Molecular Genetics, 2003, Vol. 12, No. 8 891-900
DOI: 10.1093/hmg/ddg105
© 2003 Oxford University Press

Evolution and population genetics of the H-ras minisatellite and cancer predisposition

Jacqueline A. Langdon and John A.L. Armour*

Institute of Genetics, University of Nottingham, Queen's Medical Centre, Nottingham NG7 2UH, UK

Received January 6, 2003; Accepted February 19, 2003


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS AND DISCUSSION
 CONCLUSIONS
 MATERIALS AND METHODS
 REFERENCES
 
Case–control studies have implicated rare length H-ras minisatellite alleles in cancer risk. In Europeans, this locus has four common alleles, and a larger number of rare alleles; possession of a rare allele has been identified as a risk factor responsible for 5–10% of some cancers. This unusual model of predisposition has been controversial in case–control studies, but also makes characteristic predictions about the population genetics of the locus, which we examine in this study. Using minisatellite variant repeat (‘MVR’) mapping, and compound haplotypes composed of the minisatellite and surrounding substitutional polymorphisms, we have reconstructed the main steps in the evolution of this locus in human populations. MVR-calibrated measurements of allele length yield rare allele frequencies significantly higher than most previous studies, and show that most other analyses have not distinguished two common alleles of 84 and 85 repeat units. Alleles classified as ‘rare’ in European populations predominate (70%) in the African sample studied. Small-pool PCR (SP–PCR) analysis on common alleles in sperm DNA gives an estimate for the germline minisatellite mutation rate of about 0.05% (95% confidence upper limit 0.15%). Overall, our results do not reflect a locus subject to frequent mutation and strong selection, and are difficult to reconcile with the proposed cancer predisposition. Restricted variation at this locus is most simply explained by low mutation rate, and although definitive case–control studies can only be performed using methods capable of reproducibly distinguishing rare and common alleles, our work suggests that most studies to date have not resolved alleles adequately.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS AND DISCUSSION
 CONCLUSIONS
 MATERIALS AND METHODS
 REFERENCES
 
The H-ras minisatellite is located approximately 1 kb downstream from the stop codon of the H-ras gene (1). Alleles at this minisatellite vary between ~1 and 3 kb in length, being composed of 30–100 repeat units of 28 bp (2). Population studies have reported that ~93% of alleles in Europeans consist of four common alleles, a1, a2, a3 and a4, with the remaining alleles being designated as rare (3). Rare alleles cluster in size around the four common alleles. Rare alleles at the H-ras minisatellite have been reported to be associated with cancer, with several studies finding a significantly higher frequency of rare alleles in cancer patients compared with non-cancer controls (4–8). The results of other studies, however, do not support such an association (9–14) and the link between the H-ras minisatellite and cancer remains uncertain. Most early studies used Southern blot hybridization of restriction fragments to determine allele length, and may not have resolved small (especially single-repeat) differences between alleles fully. It is only more recently that methods combining PCR amplification with polyacrylamide gel resolution have been applied (5–8,14).

There are no published studies on the phylogenetic relationships between alleles at the H-ras minisatellite. Studies of diversity at other minisatellite loci have provided insights into the evolutionary dynamics of those loci (15–19), with polymorphisms within the DNA flanking the minisatellite allowing the resolution of evolutionary relationships at (for example) MS205 (20). Allelic variation detectable at the H-ras minisatellite is increased by the distribution of variant repeats along the repeat array (due to variants at positions 7 and 15 within the 28 bp repeat unit). Studies of the variation in internal structure of H-ras alleles have revealed that each of the four common alleles displays a characteristic internal structure, and in the majority of cases rare alleles seem to be derived from the common allele closest in size (21–23). While it has been possible to separate alleles into groups according to length and internal structural similarity, the reconstruction of phylogenetic relationships between these groups of alleles has proved more difficult.

In this study haplotypes of substitutional polymorphisms, identified within the DNA flanking the repeat array, have been used to reconstruct the relationships between alleles at the H-ras minisatellite. MVR mapping has been used to calibrate analyses of allele lengths, in which we find a significantly higher frequency of ‘rare’ alleles than in other studies. We also use small-pool PCR (SP–PCR) analysis of sperm DNA to obtain a first estimate of the mutation rate to new length alleles at the H-ras minisatellite locus. Overall, our results are not consistent with the established characterization of the H-ras minisatellite as a locus subject to purifying selection against rare alleles, mediated by cancer predisposition. This and evidence that allelic states have not been adequately resolved in other studies cause us to raise serious questions about the evidence for the association between rare H-ras alleles and cancer risk.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS AND DISCUSSION
CONCLUSIONS
 MATERIALS AND METHODS
 REFERENCES
 
Relationships of H-ras haplotypes
In a sequence survey of a region of 923 bp upstream of the H-ras minisatellite, and a 774 bp region downstream, 11 single-nucleotide polymorphisms were detected. Eight of these positions were identified in an initial survey of five unrelated UK samples, and three further polymorphisms identified on sequencing 10 African Pygmy DNA samples. Haplotypes of these substitutional polymorphisms were determined in Europeans (20 Castilian DNA samples) and Africans [10 African Pygmy (Biaka/Mbuti) DNA samples]. The lengths of the minisatellite alleles associated with each haplotype in each of the 20 European and 10 African samples were determined by sizing of PCR products on agarose gels and verified by MVR mapping to give allele lengths defined as a precise number of repeat units. Since individuals heterozygous at flanking sites were always heterozygous for minisatellite length, haplotypes could be resolved using allele-specific amplification across the minisatellite.

Among the 60 chromosomes analysed, nine distinct flanking haplotypes were observed (Table 1). There were four common flanking haplotypes (H1, H2, H3 and H4), corresponding to the common length minisatellite allele classes (a1–a4). Each flanking haplotype was also associated with rare length alleles clustered around the respective common length allele, implying that most rare alleles are related to the common allele closest in size (21,23). The remaining five flanking haplotypes were each observed once within the 60 chromosomes analysed.


View this table:
[in this window]
[in a new window]
  		Table 1. Flanking haplotypes and associated minisatellite allele lengths. The number of Castilian and African Pygmy alleles detected with each flanking haplotype is shown, along with the total number of each flanking haplotype. The inferred ancestral human flanking haplotype is also shown. Positions in the present day haplotypes that differ from the inferred ancestral state are highlighted in bold. Position 4473 is C in all these haplotypes, but is shown because a single T variant was found in initial sequencing of a UK sample but not among the Pygmy or Castilian samples. Similarly, 5749G was found on initial sequencing but not among the Pygmy or Castilian samples. For haplotypes H1–H4, the numbers in brackets show the numbers of each of the corresponding common alleles (a1–a4 respectively) found
 
To infer the ancestral state of each human flanking polymorphism (Table 1), the regions flanking the H-ras minisatellite were amplified and sequenced from orang-utan (Pongo pygmaeus), gorilla (Gorilla gorilla) and chimpanzee (Pan troglodytes) DNA samples (accession numbers AJ535307–AJ535312). To identify individual bases, the numbering used here follows that of accession number J00277. For all but three of the polymorphic positions, the ancestral state could be simply deduced from sequence comparisons, as all non-human species had the same base. However, at position 5990 [also polymorphic (G/A) in chimpanzees], the ancestral state was inferred as G from association with haplotypes of other sites, and for position 5749 (deleted in chimpanzee DNA) the ancestral state was inferred as A from its association with ancestral states at other sites, and the association of the G allele with a restricted set of MVR maps similar to structures associated with the (derived) haplotype H2. Position 4661 was found to be polymorphic (T/C) in humans, T in the chimpanzee sample analysed, and C in the gorilla and orang-utan DNA samples analysed. The state of the great ape common ancestor could therefore be inferred as C. The ancestral human state could be either C or T, depending on the mutation events that have occurred at this position in the primate lineages: either a C to T change in the human/chimpanzee common ancestor, followed by a reversion to C along the human lineage, or two C to T changes occurring independently along the chimpanzee and human lineages. Position 4661 is part of a CG doublet, making two independent C to T changes more parsimonious. The ancestral human state of position 4661 was therefore inferred as C.

The flanking haplotypes were assembled into an allele tree for this locus starting from the inferred ancestral haplotype (Fig. 1). All the haplotypes observed in the populations studied could be joined by simple substitutional steps without recombination, except for haplotype P8U, which seems to have been created by a recombination event and is omitted from the tree. Haplotype H3 represents the ancestral human haplotype, possessing the inferred ancestral state at each of the 11 polymorphic positions identified. The tree then branches with the common haplotype H4 evolving down a separate lineage to the common haplotypes H1 and H2.



View larger version (18K):
[in this window]
[in a new window]
  		Figure 1. Allele tree of H-ras minisatellite flanking haplotypes, showing the evolution from the inferred ancestral haplotype to haplotypes observed in the Castilian and African Pygmy populations studied. The four commonest haplotypes (H1–H4) are shown in bold, and the lengths of associated minisatellite alleles in italics. Haplotype P8U (see Table 1) has not been placed on this tree and is probably derived from a recombinant between haplotypes H3 and H4.

 
Minisatellite variant repeat variability
A further level of variability at the H-ras minisatellite comes from variation in repeat unit sequence and the distribution of these minisatellite variant repeats (MVRs) along the array. In the absence of recombination we would predict that minisatellites associated with the same flanking haplotype have descended from a common ancestor and thus possess similar, related MVR structures. Four-state MVR mapping was developed to produce full-length MVR maps of the interspersion pattern of variant repeats for all 60 H-ras minisatellite alleles from the Castilian and African Pygmy DNA samples. Four repeat-specific primers were used, distinguishing five main repeat types: 1, 2, 3, 4 and 0 (repeat types designated ‘0’ contain mutations which prevent primer binding). As MVR-mapping was carried out from each end of the array it was possible to distinguish additional repeat types by comparing the forward and reverse maps. For example, repeat type 5 corresponds to a type 1 signal with forward MVR mapping and a type 0 with reverse mapping, and type 6 corresponds to a forward type 4 signal and a reverse type 0 signal. The positions of repeat types 5 and 6 within the repeat array were found to correspond to the positions of these repeats detected by sequence analysis (23).

Alignment of full-length MVR maps of all 40 Castilian and 20 African Pygmy alleles (full data available via www.nottingham.ac.uk/~pdzjala/hras/hras.htm) does not show any clear polarity of mutation, but instead shows that structural changes are spread throughout the array. Each of the common length allele groups has a characteristic interspersion pattern of repeats, clearly distinct from the maps of the other common alleles. Previous sequence analysis (23) divides a1 alleles into two sub-classes, a1A and a1B, differing only by the insertion of a type 2 repeat at position 17 and the deletion of a type 1 repeat between positions 27 and 28 in a1B. Among the 40 Castilian alleles 15 a1A alleles and four a1B alleles were detected. In our Castilian sample three a2 alleles were detected with identical MVR maps, while only one a3 allele was detected with an MVR map that differs from the ‘a3A’ sequence of Ding et al. (23) at position 42. Although three a4 length alleles were found among the Castilian alleles, none of these have the ‘typical’ a4 type pattern (23) (based on the sequencing of three a4 alleles).

Relationships between the MVR maps associated with each haplotype can be inferred from the relationships between flanking haplotypes. Representative MVR maps associated with each of the flanking haplotypes have been superimposed on the allele tree for this locus, shown in Figure 2 (all the MVR maps, grouped according to associated flanking haplotype, are available via the web page, www.nottingham.ac.uk/~pdzjala/hras/hras.htm). Each haplotype is associated with a distinctive subset of MVR maps. Apart from the simple relationships evident between the groups of shorter alleles (haplotypes H5, H1 and H2), the MVR maps in different haplotypes do not retain sufficient similarity to allow the unambiguous reconstruction of their evolution; some possible links between structures found in different haplotypes can be seen via the web site above. Low levels of variation were observed between the MVR maps within each group, even between the African and European populations studied. Indeed, on the background of haplotype H2, a substitutional variant has arisen (5749G) which is still associated with 46-repeat minisatellite alleles (Figs 1 and 2). These low levels of variation suggest that a low mutation rate operates at this locus compared to hypervariable minisatellites (with heterozygosities >90% and mutation rates of the order of 1%); the alternative explanation (4) is that selection is acting at this locus to limit minisatellite allele variation associated with each haplotype.



View larger version (38K):
[in this window]
[in a new window]
  		Figure 2. MVR maps of representative alleles associated with each flanking haplotype superimposed on the allele tree (Fig. 1). Key: cross, repeat type 0; solid circles, repeat type 1; open circles repeat type 2; open squares, repeat type 3; solid squares, repeat type 4; asterisks, repeat type 5; plus, repeat type 6; the symbol represents a (non-5, non-6) repeat at which a discrepancy was observed between the forward and reverse maps. The probable recombinant haplotype P8U is shown separately from the tree.

 
Flanking haplotype H5, which is associated with a rare-length allele, is found at an internal position within the tree at an intermediate position between haplotypes H3 and H1. If rare-length alleles were selected against because of association with cancer, they would be predicted to be transient in the population and found at the tips of the allele tree. Three additional examples of haplotype H5 were identified from a total of 204 DNA samples from individuals sampled from Nottinghamshire, UK. Each of the examples of haplotype H5 was associated with a 32 repeat length minisatellite allele with closely similar MVR maps (shown on the web page, see above for URL). The similarity between MVR maps suggests that the haplotype H5 alleles represent a true family derived from a single common ancestor. This, along with the internal location of haplotype H5 in the allele tree, is inconsistent with the hypothesis that all rare-length alleles are selected against. It is possible that only a subset of rare alleles is selected against, but if so the effects of cancer predisposition and selection should be concentrated on a smaller number of alleles, and hence be of greater relative magnitude.

MVR and allele length measurement
MVR mapping allows the precise sizing of minisatellite alleles in numbers of repeats, and was used to determine the exact sizes of all 40 Castilian and 20 African Pygmy alleles. A larger survey of allele lengths was carried out on 204 DNA samples (from Nottinghamshire, UK) to obtain an accurate estimate of the rare allele frequency in a control population (Fig. 3). MVR-mapped marker alleles were included on each gel to enable an accurate comparison of allele size, and accuracy in sizing large (a3 and a4) alleles was enhanced by resolving alleles of similar size on an extended gel run to allow direct ‘side by side’ comparison. Furthermore, the sizes of some alleles were checked by MVR mapping. Using these methods all 408 alleles were accurately sized in numbers of repeats (available via web page, see above for URL).



View larger version (12K):
[in this window]
[in a new window]
  		Figure 3. MVR-calibrated allele size/frequency distribution for the H-ras minisatellite in 408 alleles (204 individuals) sampled from a UK population. Arrows indicate the four common allele size classes: a1 (30 repeats), a2 (46 repeats), a3 (68 repeats) and a4 (84/85 repeats).

 
Four common allele classes were detected, classified as 30 (a1), 46 (a2), 68 (a3) and 84/85 (a4) repeats in length, occurring at frequencies of 54, 9.6, 9.1 and 11%, respectively. The a4 allele class (45 alleles) was composed of two distinct alleles; 29 alleles contained 84 repeats, and 16 alleles were 85 repeats long. While most previous studies treat a4 alleles as homogeneous in length (4), they are classified by simple length estimation, while only Ding et al. (23) give allele sizes in exact numbers of repeats: they detected four 84 repeat alleles and two 85 repeat alleles. It is likely that previous studies were unable to detect the heterogeneity that exists within the a4 allele class, due to incomplete resolution of single repeat differences.

Alleles which differ by one repeat or more from one of the four common alleles are classified as ‘rare’ (4). If alleles of both 84 and 85 repeats are classified as common, within this sample of 408 UK alleles, 68 rare alleles were found, giving a frequency of 16.7% (95% confidence interval 13.9–20.0%). If only 84-repeat alleles are classified as a4, and 85-repeat alleles are therefore ‘rare’, there are 84 rare alleles and a frequency of 20.6% (17.6–24.2%). Rare alleles cluster around the four common alleles, with the majority close in size to the a1 allele. Using either interpretation of the 84/85 repeat unit alleles, the rare allele frequency detected in this study is significantly greater than reported by Krontiris et al. (3,4) and in the majority of other studies (5–8). Since we have used additional resolution to distinguish closely spaced large alleles, and have verified the resolution using MVR mapping, the difference is more likely to be due to improved resolution than population heterogeneity. Firgaira et al. (14) detected a rare allele frequency of 16.9% in controls, sizing PCR amplified alleles on an automated DNA sequencer. With the exception of the 84/85 repeat distinction, which may be very hard to resolve on a system designed for much smaller fragments, the frequencies in our study agree very closely with those found by Firgaira et al. (14), and support their contention (24) that the resolution of alleles in previous studies has been inadequate.

The distribution of haplotypes at the H-ras minisatellite between the African and non-African populations differs from that observed at other loci. Studies at other loci have generally indicated that genetic diversity is much greater within African populations, with lineages in non-African populations making up a subset of those found in Africa (25). In contrast, the H-ras minisatellite locus has similar numbers of distinct haplotypes and allele lengths observed in the European and African populations, and the major branches of the allele tree are represented in both the African and non-African populations. Comparison of the estimated heterozygosity rates for the Castilian and Biaka/Mbuti DNA samples revealed that the flanking haplotype heterozygosity rate did not differ significantly between these populations (P>0.05), while the minisatellite allele length heterozygosity in Africans was found to be significantly higher than in Castilians (P<0.01). The frequency of ‘rare’ alleles in the African Pygmy sample studied was 70% compared with 25% in Castilians (alleles of repeat length 30, 46, 68 and 84 classed as common). Thus, although the number of different allele lengths observed is the same in each of the samples, the estimated heterozygosity rate for allele length is much greater in the (smaller) African Pygmy sample, with a significantly higher frequency of alleles that would be ‘rare’ in Europe (P<0.01). Allele length diversity therefore seems to be higher in the African Pygmy sample, but the variation observed in Europe does not simply represent a subset of that seen in Africa. The common European haplotypes H2 and H3 (associated with a2 and a3 alleles, respectively) were absent from the African Pygmy sample studied. This could be due to sampling bias, compounded by the limited size of the sample studied (20 alleles), but a previous survey of 226 alleles from Japanese individuals (11) also reported an absence of common-length a2 alleles, suggesting that this common-length allele class may be limited to European populations. The different pattern of haplotype diversity observed at this locus, compared to that observed at other regions, could be due to balancing selection acting to preserve distinct lineages at this locus.

The action of selection on this locus was investigated by carrying out neutrality tests on sequence data from the DNA flanking the H-ras minisatellite. Tajima's D (26), Fu and Li's D* and F* (27), and Fu's Fs (28) neutrality tests were carried out on 1400 bp of flanking DNA (822 bp upstream of the minisatellite and 578 bp downstream) from 60 chromosomes (40 Castilian and 20 African Pygmy), and in no case differed significantly from zero. This may indicate that the conditions of the null hypothesis (neutrality, constant population size and panmixia) are being fulfilled at this locus, but may also result from insufficient power of the data set. However, the actions of selection and population growth on a genealogy are very difficult to distinguish. As the genome-wide evidence for population growth is strong (29), it is also possible that at the H-ras minisatellite balancing selection may counteract the effects of population expansion. Any selection detected at the H-ras minisatellite may not be acting directly at this locus, but may instead be due to selection at a nearby linked locus, for example the dopamine receptor gene DRD4.

Estimating mutation rates by SP–PCR
The H-ras minisatellite has been characterised as a highly unstable locus at which four common alleles are maintained due to purifying selection against rare alleles which predispose to cancer (30); the observed frequencies of rare alleles result from a balance between mutation creating new rare alleles and selection removing them from the population. However, no studies to date have attempted to accurately measure the mutation rate at the H-ras minisatellite (23,31). We have developed SP–PCR analysis at the H-ras minisatellite to estimate the male germline mutation rate directly in sperm DNA.

SP–PCR analysis involves the amplification of minisatellite alleles from dilute aliquots of DNA, with mutant alleles being detected as variant-length PCR products (32–36). In SP–PCR analysis of two sperm donors, N1 (heterozygous a1/a2) and N22 (heterozygous a1/a4), two putative mutant alleles were detected in 4080 sperm equivalents tested. This corresponds to a mutation rate of 0.049%, with 95% confidence intervals of 0.02–0.15%. Given that even the two putative mutant bands observed may be PCR artefacts, a sperm mutation rate of between 0 and 0.15% can be inferred, placing an upper bound on the mutation rate. This compares to sperm mutation rates of 0.8% at MS32 (32), 0.5% at MS205 (33) and 9.3% at CEB1 (35). If the frequency of rare H-ras alleles is attributed to an equilibrium between mutation and (dominant) selection, a mutation rate as low as 0.05% predicts an equilibrium rare allele frequency of only 1%, assuming a selection coefficient of 0.05 against rare allele carriers. To be consistent with the observed frequencies, selection against rare H-ras alleles would have to be very weak. While predisposition to (usually postreproductive) cancer could indeed lead to weak selection, our results show that the shaping of the allele distribution by selection could not possibly take the form of strong selection against carriers, or of frequent ‘prenatal lethals’ (30).

Estimating lineage ages
The observed minisatellite mutation rate was used to estimate the age of some H-ras lineages based on their associated minisatellite variation, using the method developed by Slatkin and Rannala (37). For example, within the sample of 40 Castilian chromosomes 25 examples of flanking haplotype H1 were identified, including eight distinct MVR maps, corresponding to four different allele lengths.

Conditions for the simulations are described briefly in the Materials and Methods, and further details can be found through the web site www.nottingham.ac.uk/~pdzjala/hras/hras.htm. Based on the Castilian sample of 40 chromosomes, the maximum likelihood estimates for the age of haplotype H1 ranged from 52 600 to 81 000 years for simulations considering varying population size and mutation rate, but with a constant population growth rate of 0.005. Using data from the 408 Nottinghamshire alleles sampled, the age of the haplotype H1 lineage was estimated as 61 000 years (95% confidence interval 55 600–76 200 years). A maximum likelihood estimate for the age of haplotype H2 of 55 400 years was obtained from the Nottinghamshire sample data; this younger date is consistent with the origin of haplotype H2 as derived from haplotype H1. Haplotype H2 was absent from the African population studied, and may represent a lineage that has diverged and expanded outside Africa.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS AND DISCUSSION
 CONCLUSIONS
MATERIALS AND METHODS
 REFERENCES
 
Several observations in this study are not easily reconciled with the hypothesis that the H-ras minisatellite is a highly unstable locus at which purifying selection acts against rare alleles to maintain just four common alleles in the population (4,30). The low mutation rate found on SP–PCR analysis of sperm DNA suggests that this locus is not highly unstable. The low levels of MVR map diversity associated with each flanking haplotype are also consistent with a low mutation rate. The low levels of minisatellite diversity observed can therefore most simply be attributed not to selection, but to the relative inability of infrequent mutation to counteract the effect of drift. The higher length heterozygosity and rare allele frequency in our African sample may therefore reflect the higher effective sizes typical of other African populations. The detection of a class of rare length alleles associated with flanking haplotype H5, which is located at an internal position within the allele tree, is similarly inconsistent with the hypothesis that all rare alleles are deleterious. The rare allele frequency detected in our study (16.7%) is significantly higher than reported in the majority of previous studies on similar populations, but more consistent with the study of Firgaira et al. (14), which used high-resolution methods, and which detected no association between rare alleles and cancer. Assuming that the populations examined are comparable, our work and that of Firgaira et al. (14) both suggest that initial work using Southern blot hybridization (3,4) may have underestimated the rare allele frequency in controls by as much as 2–3-fold, and that even some more recent PCR-based studies (5–8), reporting control rare allele frequencies in the range 12–14%, may also involve some systematic underestimation. Furthermore, two relatively frequent alleles (84/85 repeats) have not hitherto been distinguished, also suggesting that resolution of allele sizes was limited in previous studies. Therefore previous studies, including those on which the proposed association with cancer risk has been established (4,5,7,8,38), may have been incapable of reliably distinguishing small length differences at the H-ras minisatellite, the very differences said to be the basis of cancer predisposition. Because of those studies, the H-ras minisatellite is widely regarded as a locus of great significance in cancer predisposition. However, our results are not consistent with rare allele frequencies being determined by a balance between high mutation rates and strong selection, and suggest that a study using methods capable of accurately distinguishing minisatellite alleles, and using samples of sufficient power, should be undertaken to assess thoroughly the reported association between rare H-ras minisatellite alleles and cancer predisposition.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS AND DISCUSSION
 CONCLUSIONS
 MATERIALS AND METHODS
REFERENCES
 
Genomic DNA samples
British, Castilian (17), Biaka Pygmy and Mbuti Pygmy (Coriell Cell Repository) genomic DNA samples were used in this study. The UK H-ras allele size survey was carried out on DNA samples from 204 individuals from (non-urban) Nottinghamshire, UK, with a mean age of 57 years. British sperm DNA samples were used for small-pool PCR analysis. Primate DNA samples were obtained from ECACC (European Collection of Animal Cell Cultures).

DNA amplification and sequencing
All PCR used a final concentration of 50 mM Tris–HCl pH 8.8, 12 mM (NH4)2SO4, 5 mM MgCl2, 7.4 mM 2-mercaptoethanol, 125 µg/ml BSA, and 1.1 mM each dNTP (modified from 39). The sequences of all primers used are shown in Table 2. DNA flanking the 5' end of the H-ras minisatellite was amplified and sequenced with primers Ras1 and Ras3, and sequencing completed with Ras7. DNA flanking the 3' end of the minisatellite was amplified and sequenced with primers Ras4 and Ras6. The same primers were used to amplify and sequence the flanking regions in chimpanzee, gorilla and orang-utan. However, as primer Ras7 does not anneal in the orang-utan, primer Ras9 was designed to anneal in all primates to allow full sequencing of the 5' flanking region. Sequencing was carried out using the Big Dye cycle sequencing system (ABI Prism; PE Biosystems).


View this table:
[in this window]
[in a new window]
  		Table 2. Sequences of all primers used. Annealing temperatures are also shown for allele-specific primers
 
Substitutional polymorphisms and haplotypes
Eleven single nucleotide polymorphisms were located within the H-ras minisatellite flanking sequence at positions 3904, 3973, 4334, 4414, 4472, 4473, 4482, 4546, 4661, 5749 and 5990 (the numbering relates to the position within the sequence in GenBank accession number J00277). Restriction tests were used to genotype polymorphisms 3973 (AluI), 4414 (MnlI), 4472/3 (AccI and RsaI), 4482 (NlaIV), 4546 (AluI) and 5990 (MspA1I). Amplification with allele-specific primers was used to genotype polymorphisms 3904, 4334, 4661 and 5749. To resolve the haplotype of flanking polymorphisms on each chromosome, allele-specific primers were used to amplify specifically from each allele of each heterozygous position, across the minisatellite to a common primer on the opposite side of the repeat array.

Amplification and sizing of H-ras minisatellite alleles
Minisatellite alleles were amplified in 20 µl reactions using primers Ras2 and Ras5 with final concentrations of 0.5 µM for each primer, 0.05 U/µl of Taq polymerase/Pfu polymerase mix (40 units Taq:1 unit Pfu), 6% glycerol, 34 mM Tris base and 1xPCR buffer system (as above) to amplify 10 ng of input DNA. Thermal cycling conditions were: 95°C for 1 min, 67°C for 1 min and 70°C for 10 min, for 21 cycles followed by one cycle of 65°C for 45 s and 72°C for 10 min. As markers for common-length alleles (a1, a2, a3 and a4), the minisatellite alleles from Castilian sample 5 (genotype a2/a3) and sample 10 (genotype a1/a4) were amplified, pooled and loaded on each gel. H-ras minisatellite alleles were resolved on long (30 cm) 1.6% agarose gels, in the absence of ethidium bromide. Marker common length alleles and 100/500 bp marker-ladder DNA were also loaded on each gel, to allow accurate allele sizing. Southern blotting and hybridization with a probe for the H-ras repeats and separate probes for the 100 and 500 bp marker DNA was carried out, and the results visualized by autoradiography. To enable more accurate sizing of longer alleles (>60 repeats), selected samples with large alleles of similar length were run on a second gel for direct, ‘side-by-side’ comparison.

Minisatellite variant repeat mapping
The majority of alleles to be MVR mapped were first amplified preparatively (32 cycles) using primers Ras2 and Ras5. Long alleles (a3 and a4) were selectively amplified using primers Ras5 and the allele-specific primer 4472-GC, reducing competition from shorter alleles. Following separation by agarose gel electrophoresis minisatellite DNA was released from gel blocks by freeze/thawing. From the supernatant from this procedure, 1 µl (assumed to contain about 20 pg DNA) was used as input in MVR–PCR reactions.

Minisatellite variant repeat (MVR) interspersion patterns were obtained in a similar way to (22), distinguishing four repeat-unit types. The bases discriminating the repeat types are shown in bold in Table 2. Mapping was carried out from each end of the array; primer Ras2 was used as the 5' fixed flanking primer and Ras5 as the 3' flanking primer. Twenty-microlitre MVR–PCR reactions were carried out with final concentrations of 0.05 U/µl of Taq polymerase, 1 µM flanking primer, 1 µM EXT, 5 nM repeat-specific primer and 1xPCR buffer system (as above). PCR cycling conditions employed varied between alleles. Cycling conditions were: 96°C for 1.3 min and 72°C for 6 min cycled between 10 and 23 times, depending on the allele size and the detection method used, then 1 cycle of 70°C for 10 min. PCR products were run on 30 cm 1.4% agarose gels, and detected by Southern blotting and hybridization with radioactively labelled H-ras repeat probe.

In individuals homozygous for allele length but with heterozygous MVR structure, the mixed alleles from these individuals were MVR mapped to preparative levels from both directions, and the largest heterozygous MVR bands recovered and individually MVR mapped in both directions, producing MVR maps of the separated alleles from which each original allele structure could be reconstructed.

Small-pool PCR analysis
Small-pool PCR analysis was carried out largely as described by May et al. (33) for MS205. Primers Ras2 and Ras12 were used to amplify minisatellite alleles from dilute aliquots (15 diploid genome equivalents) of sperm DNA in 20 µl SP–PCR reactions. PCR efficiency was estimated by carrying out PCR reactions with approximately single molecule inputs of DNA, with Poisson analysis being used to calculate the number of amplifiable molecules in each reaction. All DNA samples analysed by SP–PCR were diluted with 10 mM Tris–HCl (pH 7.5) and 1 ng/µl herring sperm DNA (34). Fifteen microlitres of the products from each SP–PCR reaction were run on 30 cm 1.4% agarose gels, and detected by Southern blotting and hybridization with a radioactively labelled H-ras minisatellite probe.

Tests of neutrality
To test for departure from a standard neutral model DNASP 3.5 (www.ub.es/dnasp/) (40) was used to perform Tajima's D (26), Fu's Fs (28), and Fu and Li's D* and F* (27) tests on the flanking sequence data. These tests compare the observed pattern of sequence variation against that predicted from a standard neutral model. As these tests assume the infinite sites model and no recombination, position 4661 was omitted from the analysis and a total of 1399 bp of sequence analysed in the neutrality tests. DnaSP was also used to estimate a P-value for each of these test statistics by computer simulations using the coalescent theory. These P-values represent the estimated probability of obtaining values lower than (or equal to) the ones observed. Each of these simulations was carried out using 10 000 replicates, and a recombination parameter C=4Nec (where c is the recombination rate per generation) with a value of 10 was used.

Dating lineages
The program BDMC21 (www.rannala.org/) (37) was used to obtain maximum likelihood estimates for the ages of lineages. Maximum likelihood simulations of lineage age were carried out using a minisatellite mutation rate of 0.05%, as well as the 95% confidence range (0.02–0.15%). A population growth rate parameter of 0.005 per generation was used (37), appropriate for recent rapid growth in European populations. Simulations were also carried out using growth rate parameters of 0.002 and 0.008 (41).


   ACKNOWLEDGEMENTS
 
We thank John Brookfield, Santos Alonso, Ed Hollox, Nigel Gillson, Celia Burgoyne, Pari Datta and Tamsin Majerus for helpful discussions and comments on the manuscript. This work was supported by a University Research Scholarship (to J.A.L.) from the University of Nottingham.


   FOOTNOTES
 
* To whom correspondence should be addressed. Tel: +44 1159249924; Fax: +44 1159709906; Email: john.armour{at}nottingham.ac.uk Back


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS AND DISCUSSION
 CONCLUSIONS
 MATERIALS AND METHODS
 REFERENCES
 

  1. Capon, D.J., Chen, E.Y., Levinson, A.D., Seeburg, P.H. and Goeddel, D.V. (1983) Complete nucleotide sequences of the T24 human bladder carcinoma oncogene and its normal homologue. Nature, 302, 33–37.[CrossRef][Medline]

  2. Krontiris, T.G., DiMartino, N.A., Colb, M., Mitcheson, H.D. and Parkinson, D.R. (1986) Human restriction-fragment-length-polymorphisms and cancer risk assessment. J. Cell. Biochem., 30, 319–329.[CrossRef][Web of Science][Medline]

  3. Krontiris, T.G., DiMartino, N.A., Colb, M. and Parkinson, D.R. (1985) Unique allelic restriction fragments of the human Ha-ras locus in leukocyte and tumour DNAs of cancer patients. Nature, 313, 369–374.[CrossRef][Medline]

  4. Krontiris, T.G., Devlin, B., Karp, D.D., Robert, N.J. and Risch, N. (1993) An association between the risk of cancer and mutations in the HRAS1 minisatellite locus. New Engl. J. Med., 329, 517–523.[Abstract/Free Full Text]

  5. Gosse-Brun, S., Sauvaigo, S., Daver, A., Page, M., Lortholary, A., Larra, F., Bignon, Y.J. and Bernard-Gallon, D. (1999) Specific H-ras minisatellite alleles in breast cancer susceptibility. Anticancer Res., 19, 5191–5196.[Web of Science][Medline]

  6. Lindstedt, B.A., Ryberg, D., Zienolddiny, S., Khan, H. and Haugen, A. (1999) Hras1 VNTR alleles as susceptibility markers for lung cancer: relationship to microsatellite instability in tumors. Anticancer Res., 19, 5523–5527.[Web of Science][Medline]

  7. Rossell, R., Calvo, R., Sanchez, J.J., Maurel, J., Guillot, M., Monzo, M., Nunez, L. and Barnadas, A. (1999) Genetic susceptibility associated with rare HRAS1 variable number of tandem repeats alleles in Spanish non-small cell lung cancer patients. Clin. Cancer Res., 5, 1849–1854.[Abstract/Free Full Text]

  8. Weitzel, J.N., Ding, S.F., Larson, G.P., Nelson, R.A., Goodman, A., Grendys, E.C., Ball, H.G. and Krontiris, T.G. (2000) The HRAS1 minisatellite locus and risk of ovarian cancer. Cancer Res., 60, 259–261.[Abstract/Free Full Text]

  9. Thein, S.L., Oscier, D.G., Flint, J. and Wainscoat, J.S. (1986) Ha-ras hypervariable alleles in myelodysplasia. Nature, 321, 84–85.[CrossRef][Medline]

  10. Sutherland, C., Shaw, H.M., Roberts, C., Grace, J., Stewart, M.M., McCarthy, W.H. and Kefford, R.H. (1986) Harvey–ras oncogene restriction fragment alleles in familial melanoma kindreds. Br. J. Cancer, 54, 787–790.[Web of Science][Medline]

  11. Ishikawa, J., Maeda, S., Takahashi, R., Kamidono, S. and Sugiyama, T. (1987) Lack of correlation between rare Ha-ras alleles and urothelial cancer in Japan. Int. J. Cancer, 40, 474–478.[Web of Science][Medline]

  12. Gerhard, D.S., Dracopoli, N.C., Bale, S.J., Houghton, A.N., Watkins, P., Payne, C.E., Greene, M.H. and Housman, D.E. (1987) Evidence against Ha-ras-1 involvement in sporadic and familial melanoma. Nature, 325, 73–75.[CrossRef][Medline]

  13. Mackay, J., Elder, P.A., Porteous, D.J., Steel, C.M., Hawkins, R.A., Going, J.J. and Chetty, U. (1988) Partial detection of chromosome 11p in breast-cancer correlates with size of primary tumor and estrogen-receptor level. Br. J. Cancer, 58, 710–714.[Web of Science][Medline]

  14. Firgaira, F.A., Seshadri, R., McEvoy, C.R.E., Dite, G.S., Giles, G.G., McCredie, M.R.E., Southey, M.C., Venter, D.J. and Hopper, J.L. (1999) HRAS1 rare minisatellite alleles and breast cancer in Australian women under age forty years. J. Natl Cancer Inst., 91, 2107–2111.[Abstract/Free Full Text]

  15. Gray, I.C. and Jeffreys, A.J. (1991) Evolutionary transience of hypervariable minisatellites in man and the primates. Proc. R. Soc. B, 243, 241–253.[Medline]

  16. Armour, J.A.L., Anttinen, T., May, C.A., Vega, E.E., Sajantila, A., Kidd, J.R., Kidd, K.K., Bertranpetit, J., Paabo, S. and Jeffreys, A.J. (1996) Minisatellite diversity supports a recent African origin for modern humans. Nat. Genet., 13, 154–160.[CrossRef][Web of Science][Medline]

  17. Alonso, S. and Armour, J.A.L. (1998) MS205 minisatellite diversity in Basques: evidence for a pre-neolithic component. Genome Res., 8, 1289–1298.[Abstract/Free Full Text]

  18. Jobling, M.A., Bouzekri, N. and Taylor, P.G. (1998) Hypervariable digital DNA codes for human paternal lineages: MVR-PCR at the Y-specific minisatellite, MSY1 (DYF155S1). Hum. Mol. Genet., 7, 643–653.[Abstract/Free Full Text]

  19. Stead, J.D.H., Buard, J., Todd, J.A. and Jeffreys, A.J. (2000) Influence of allele lineage on the role of the insulin minisatellite in susceptibility to type 1 diabetes. Hum. Mol. Genet., 9, 2929–2935.[Abstract/Free Full Text]

  20. Rogers, E.J., Shone, A.C., Alonso, S., May, C.A. and Armour, J.A.L. (2000) Integrated analysis of sequence evolution and population history using hypervariable compound haplotypes. Hum. Mol. Genet., 9, 2675–2681.[Abstract/Free Full Text]

  21. Kasperczyk, A., DiMartino, N.A. and Krontiris, T.G. (1990) Minisatellite allele diversification: The origin of rare alleles at the HRAS1 locus. Am. J. Hum. Genet., 47, 854–859.[Web of Science][Medline]

  22. Conway, K., Edmiston, S.N., Hulka, B.S., Garrett, P.A. and Liu, E.T. (1996) Internal sequence variations in the Ha-ras variable number tandem repeat rare and common alleles identified by minisatellite variant repeat polymerase chain reaction. Cancer Res., 56, 4773–4777.[Abstract/Free Full Text]

  23. Ding, S., Larson, G.P., Foldenauer, K., Zhang, G. and Krontiris, T.G. (1999) Distinct mutation patterns of breast cancer-associated alleles of the H-RAS1 minisatellite locus. Hum. Mol. Genet., 8, 515–521.[Abstract/Free Full Text]

  24. Hopper, J.L., Firgaira, F.A., Dite, G.S., Giles, G.G., McCredie, M.R.E., Southey, M.C., Venter, D.J., Seshadri, R. and McEvoy, C.R.E. (2000) Re: HRAS1 rare minisatellite alleles and breast cancer in Australian women under age forty years—response. J. Natl Cancer Inst., 92, 756–757.[Free Full Text]

  25. Tishkoff, S.A. and Williams, S.M. (2002) Genetic analysis of African populations: human evolution and complex disease. Nat. Rev. Genet., 3, 611–621.[CrossRef][Web of Science][Medline]

  26. Tajima, F. (1989) Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics, 123, 585–595.[Abstract/Free Full Text]

  27. Fu, Y.X. and Li, W.H. (1993) Statistical tests of neutrality of mutations. Genetics, 133, 693–709.[Abstract]

  28. Fu, Y.X. (1997) Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics, 147, 915–925.[Abstract]

  29. Rogers, A.R. (2001) Order emerging from chaos in human evolutionary genetics. Proc. Natl Acad. Sci. USA, 98, 779–780.[Free Full Text]

  30. Krontiris, T.G. (1995) Minisatellites and human disease. Science, 269, 1682–1683.[Free Full Text]

  31. Larson, G.P., Ding, S., Lafreniere, R.G., Rouleau, G.A. and Krontiris, T.G. (1999) Instability of the EPM1 minisatellite. Hum. Mol. Genet., 8, 1985–1988.[Abstract/Free Full Text]

  32. 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. Nat. Genet., 6, 136–145.[CrossRef][Web of Science][Medline]

  33. May, C.A., Jeffreys, A.J. and Armour, J.A.L. (1996) Mutation rate heterogeneity and the generation of allele diversity at the human minisatellite MS205 (D16S309). Hum. Mol. Genet., 5, 1823–1833.[Abstract/Free Full Text]

  34. Jeffreys, A.J. and Neumann, R. (1997) Somatic mutation processes at a human minisatellite. Hum. Mol. Genet., 6, 129–136.[Abstract/Free Full Text]

  35. Buard, J., Bourdet, A., Yardley, J., Dubrova, Y. and Jeffreys, A.J. (1998) Influences of array size and homogeneity on minisatellite mutation. EMBO J., 17, 3495–3502.[CrossRef][Web of Science][Medline]

  36. Tamaki, K., May, C.A., Dubrova, Y.E. and Jeffreys, A.J. (1999) Extremely complex repeat shuffling during germline mutation at human minisatellite B6.7. Hum. Mol. Genet., 8, 879–888.[Abstract/Free Full Text]

  37. Slatkin, M. and Rannala, B. (1997) Estimating the age of alleles by use of intraallelic variability. Am. J. Hum. Genet., 60, 447–458.[Web of Science][Medline]

  38. Gosse-Brun, S., Sauvaigo, S., Daver, A., Larra, F., Kwiatkowski, F., Bignon, Y.J. and Bernard-Gallon, D. (1998) Association between H-ras minisatellite and colorectal cancer risk. Anticancer Res., 18, 2611–2616.[Web of Science][Medline]

  39. 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. Cell, 60, 473–485.[CrossRef][Web of Science][Medline]

  40. Rozas, J. and Rozas, R. (1999) DnaSP version 3: an integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics, 15, 174–175.[Abstract/Free Full Text]

  41. Pritchard, J.K., Seielstad, M.T., Perez-Lezaun, A. and Feldman, M.W. (1999) Population growth of human Y chromosomes: a study of Y chromosome microsatellites. Mol. Biol. Evol., 16, 1791–1798.[Abstract]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
R. M. Tamimi, S. E. Hankinson, S. Ding, V. Gagalang, G. P. Larson, D. Spiegelman, G. A. Colditz, T. G. Krontiris, and D. J. Hunter
The HRAS1 Variable Number of Tandem Repeats and Risk of Breast Cancer
Cancer Epidemiol. Biomarkers Prev., December 1, 2003; 12(12): 1528 - 1530.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (7)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Langdon, J. A.
Right arrow Articles by Armour, J. A.L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Langdon, J. A.
Right arrow Articles by Armour, J. A.L.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?