- Split View
-
Views
-
Cite
Cite
Neil Howell, Corinna Herrnstadt, David A. Mackey, Different Patterns of Expansion/Contraction During the Evolution of an mtDNA Simple Repeat, Molecular Biology and Evolution, Volume 18, Issue 8, August 2001, Pages 1593–1596, https://doi.org/10.1093/oxfordjournals.molbev.a003946
- Share Icon Share
The human mtDNA noncoding control region contains simple homopolymeric repeat sequences that undergo expansion and contraction (Torroni et al. 1994 ; Bendall and Sykes 1995 ; Marchington et al. 1997 ; Howell and Bogolin Smejkal 2000 ). In most instances, the expansion/contraction process is sufficiently rapid that individuals are heteroplasmic and carry multiple repeat length variants. Torroni et al. (1994) were the first group to report an expansion of the C6 repeat at nucleotides 568–573 (all mtDNA sequences presented here are those of the rCRS L-strand; Andrews et al. 1999 ). They reported that mtDNAs of European haplogroup I were expanded in this repeat and carried an additional 2–6 residues. We report here that this repeat can undergo both “small” and “large” expansions (defined below) but that all such expansions appear to be unstable and prone to further expansion/contraction. Furthermore, large expansions have occurred at least twice during human evolution. The distributions of length variants for these two expansions differ substantially, and we suggest that this difference is due to a relatively slow evolution toward a stable equilibrium of expansion/contraction.
We identified five individuals and one matrilineal pedigree of eight members among our DNA samples from approximately 300 Europeans, whose mtDNA belongs to haplogroup I. The pedigree carries a LHON (Leber hereditary optic neuropathy) mutation at nucleotide 14482, and it has been described elsewhere (Howell et al. 1998 ). DNA sequence for the control region segment that spans the 568 repeat has been obtained for all 13 haplogroup I individuals. Torroni et al. (1994; see their table 2) reported that the expanded 568 repeat was homoplasmic in their subjects, although they noted that repeat length could vary among tissues from the same individual. However, we observed that all of our haplogroup I individuals and family members were heteroplasmic for 568 repeat length variants. A total of 366 M13 clones with the appropriate mtDNA insert were sequenced, and the distribution of pooled repeat tract lengths at nucleotide 568 is shown in figure 1A. Repeat length varied from 7 to 14 residues with a symmetrical distribution about a modal length of 11 residues (“large” expansion).
The distributions of repeat lengths are approximately the same in all individuals (data not shown). For example, we separated the data into two groups: the five unrelated haplogroup I individuals and the eight members of the Turkish LHON family. The distributions of the 568 repeat lengths for these two groups were the same according to the results of a Kolmogorov-Smirnov (KS) test (P ≈ 1.0). Although heteroplasmy of mtDNA repeat lengths has been studied for more than 15 years, it was important to verify that the distributions of repeat lengths reported here reflected those in situ and that they were not biased by the procedures used for nucleotide sequencing, especially PCR amplification. Our PCR amplifications routinely utilized Taq polymerase, which has a relatively high rate of substitution errors. Therefore, we carried out an independent analysis of one of the haplogroup I individuals with Pfu polymerase, which has an error rate that is approximately sixfold lower (Strategene, Inc.), using the manufacturer's specifications. A KS test showed that the two distributions were not significantly different (P ≈ 0.66). We have also shown that the distribution of repeat lengths at a different mtDNA control region repeat was not biased by the cloning and sequencing methods (Howell and Bogolin Smejkal2000) .
Torroni et al. (1994) observed a second 568 expanded repeat in a European haplogroup H mtDNA. We reported previously that the mtDNA of the VIC1 3460 LHON pedigree had an expanded 568 repeat, and it is clear that this mtDNA also belongs to haplogroup H on the basis of the alleles at nucleotides 73 and 14766 (Howell et al. 1995 ). We analyzed the 568 repeat expansion in 14 VIC1 family members from four generations (1 member of the first generation, 7 members of the second, 2 members of the third, and 4 members of the fourth), and the cumulative results are shown in figure 1B. The distribution of repeat lengths is markedly different from that of the haplogroup I repeat (the P value from a KS test is <0.0001). The modal repeat length for the pooled results from the members of the H haplogroup pedigree is 9 residues (large expansion), ranging from 6 to 17 residues, and the asymmetric distribution is clearly skewed toward longer lengths. There is insufficient phylogenetic information to conclude with certainty that the H haplogroup expansions observed by us and by Torroni et al. (1994) are due to the same event, or whether they represent independent expansions. However, we have accumulated sequence for ∼100 haplogroup H control regions within the UTMB collection of DNA samples (unpublished observations), and we have observed the 568 expansion only in the VIC1 pedigree. Among the MitoKor collection of DNA samples, control region sequence is available for ∼180 H haplogroup mtDNAs, and none of these carries a large expansion at nucleotide 568 (unpublished observations). Therefore, it is more likely that there has been only a single large expansion event during the evolution of haplogroup H mtDNAs.
The distributions of repeat lengths of the VIC1 LHON pedigree did not show significant differences between unaffected and affected family members (the P value of the KS test was ∼0.1). However, unlike the haplogroup I mtDNAs, one family member had a significantly different distribution of repeat lengths. The single member of the first generation, a maternal aunt to all members of the second generation, had a modal repeat length of 8 residues, whereas for all others it was 9 residues (data not shown). This difference was significant in KS tests versus the members of the second generation or versus the members of the fourth generation (P < 0.0001 in both tests). In contrast, there was no difference in repeat length distributions when the KS test was used to compare the members of the second and fourth generations (P ≈ 0.65) or the members of the third and fourth generations (P ≈ 0.65). The difference for this one family member may simply represent the rapid drift that can occur during transmission of heteroplasmic mtDNA molecules (Howell et al. 2000 ).
Previous studies have shown that plasmid clones that carry mtDNA repeat sequences, when transfected into a bacterial host, provide a simple model system in which expansion and contraction of the repeat sequences also occurs (Hauswirth et al. 1984 ; Howell and Bogolin Smejkal 2000 ). Our preliminary results for the 568 repeat are summarized in table 1 . After passage in bacteria, we analyzed 568 repeats whose starting lengths varied from 9 to 17 residues. Overall, the results show that repeats of 9 residues tend to expand, whereas repeats with 12 or more residues tend to contract during passage in the bacterial host (see below).
We have also observed a third type of 568 expansion. In addition to the haplogroup I mtDNAs and the haplogroup H LHON pedigree, we obtained control region sequences for another 133 genealogically unrelated mtDNA lineages within the UTMB DNA collection (normal controls, patients with nonmitochondrial neurological disorders, and patients with mitochondrial disorders including LHON). Of these mtDNAs, 130 (98%) had 568 repeats of 6 residues in length (among a total of 871 clones, 1 clone had a repeat length of 8 residues), and 3 (2%) had repeat lengths of 7 residues. The latter three mtDNAs included those from two individuals in whom the repeat length had increased to 7 residues (small expansion). These mtDNAs belonged to different sub-branches of haplogroup J, and we surmise that the two expansion events were independent. The third lineage with an expansion to 7 residues comprised the four members of a MELAS (mitochondrial encephalopathy with lactic acidosis and stroke-like episodes) pedigree (see Chinnery et al. [1999] for additional details on this pedigree). On the basis of characteristic polymorphic sites in the coding region, the mtDNA in this pedigree belongs to European haplogroup U (data not shown) and thus represents an independent expansion event. The distribution of repeat lengths from the pooled analyses of the four family members is narrow, with a pronounced mode at 7 residues (fig. 1C ). For one family member, repeat lengths were determined for mtDNA from WBC/platelets and for DNA isolated from muscle. The distributions were the same in the two tissues, a result in accord with previous findings for other mtDNA simple repeats (e.g., Howell and Bogolin Smejkal2000) . Among the MitoKor collection of DNA samples (and excluding the haplogroup I mtDNAs with their large expansions), it has been independently observed that 9 control region sequences among a total of 451 (2%) have 568 repeat lengths of 7–8 residues (small expansions; unpublished observations).
The difference between the distributions for the haplogroup I and haplogroup H mtDNAs is striking (figs. 1A and B ). Because of the number of individuals from the two haplogroups that were analyzed, and because of the inheritance data available from pedigrees, both sampling effects and rapid genetic drift through the developmental bottleneck (the effective number of mitochondrial genomes is reduced to a relatively small number in early oogenesis; reviewed in Howell et al. 2000 ) can be excluded as explanations for this difference. As a working hypothesis to explain this difference, we therefore propose the following. For the haplogroup I mtDNAs, the 568 repeat length distribution appears to have achieved some sort of balanced and dynamic equilibrium, with a maximum “stability” at 11 residues. In contrast, the distribution of repeat length variants has not yet reached equilibrium for the haplogroup H 568 repeat, as evidenced by the modal length of 9 residues and by the marked skew toward longer repeat lengths. Considering only the haplogroup H distribution of repeat lengths, one might surmise that repeats of less than 9 residues are selected against. However, that alternative seems unlikely on the basis of the prevalence of small expansions in which the modal length had increased from 6 to 7–8 residues.
There is a general tendency for repeat sequences to expand (see the discussion in Goldstein and Pollock 1997 ), but—in many cases—there is a point, or boundary, where longer repeats compromise genome stability or transmission (e.g., Farrall and Weeks 1998 ; Falush and Iwasa 1999 ). The imposition of an effective boundary on the length of the 568 repeat would explain the symmetrical distribution found for the haplogroup I mtDNAs, because expansions of shorter variants are presumably balanced with contraction and/or segregational loss of longer variants that have decreased fitness at the molecular level. It thus appears that expansion of the 568 repeat beyond the ancestral, or “ground,” state of 6 residues results in an unstable condition in which there is net expansion until a new stability, or equilibrium, is attained with a modal length of 11 residues. In this regard, expanded 568 repeats are associated with a subpopulation of deleted mtDNA molecules in which there has been apparent intramolecular recombination between the 568 repeat and the simple C-repeat that starts at nucleotide 303 (Torroni et al. 1994 and references therein). It is possible that repeat variants >11 residues become progressively more prone to deletion or gross rearrangements of the mtDNA molecules, and thus to loss through failure to replicate or segregate. There is also evidence that the control region repeat at nucleotide 16189 has an upper boundary on length (Bendall and Sykes 1995 ; Howell and Bogolin Smejkal 2000 ).
We further suggest that the time required for the 568 repeat to reach the expansion equilibrium is relatively long in terms of human evolution. Although we cannot make a precise estimate of their ages, the haplogroup I expansion must be much older in terms of human mtDNA evolution than the expansion event that occurred in the haplogroup H mtDNAs. Haplogroup I is one of the oldest European mtDNA haplogroups (Torroni et al. 1994 ), and the 568 expansion must have occurred very early, because it is present in all haplogroup I mtDNAs that have been analyzed. In contrast, haplogroup H has a more recent evolutionary origin (e.g., Richards et al. 1998 ), and it is only within a small subgroup of haplogroup H mtDNAs that the 568 expansion has occurred (we can estimate that <1% of haplogroup H mtDNAs carry the 568 large expansion). One may speculate that the haplogroup H distribution of 568 repeat lengths will require several thousand years to resemble that of the haplogroup I mtDNAs. That is, we suggest that the different haplogroup H and I repeat length distributions are two different “time” points on a single evolutionary trajectory.
Expansion and contraction of simple repeat sequences, including those in mtDNA, occur through polymerase slippage (e.g., Schlötterer and Tautz 1992 ). However, such slippage predominantly occurs in increments of one or two repeat units (reviewed in Goldstein and Pollock 1997 ). It is possible that the small expansions—which also appear to have occurred more frequently—are relatively very recent and that the distribution of these repeat lengths will eventually resemble those of the large expansions. However, one may also have to consider the occurrence of more extensive polymerase slippage events or the possibility that the large 568 repeats were generated through some other mechanism, such as intramolecular recombination.
Ross Crozier, Reviewing Editor
Abbreviation: rCRS, revised Cambridge Reference Sequence for human mtDNA.
Keywords: mitochondrial DNA mitochondrial genetics repeat sequences pedigree analysis human genetics Homo sapiens.
Address for correspondence and reprints: Neil Howell, Biology Division 0656, Room 3.348, Gail Borden Building, University of Texas Medical Branch, Galveston, Texas 77555-0656. nhowell@utmb.edu.
We gratefully acknowledge the technical assistance of Monica Toth and Iwona Kubacka. The DNA samples from the English MELAS pedigree were generously supplied by Drs. Douglass Turnbull and Mark Walker (University of Newcastle). This research was supported by grants from the National Science Foundation (BCS-9910871) and the John Sealy Endowment Fund awarded to N.H., and by a grant from the Ophthalmic Research Institute of Australia awarded to D.A.M.
References
Andrews R. M., I. Kubacka, P. F. Chinnery, R. N. Lightowlers, D. M. Turnbull, N. Howell,
Bendall K. E., B. C. Sykes,
Chinnery P. F., P. J. G. Zwijnenburg, M. Walker, N. Howell, R. W. Taylor, R. N. Lightowlers, L. Bindoff, D. M. Turnbull,
Falush D., Y. Iwasa,
Farrall M., D. E. Weeks,
Goldstein D. B., D. D. Pollock,
Hauswirth W. W., M. J. Van De Walle, P. J. Laipis, P. D. Olivo,
Howell N., C. Bogolin Smejkal,
Howell N., C. Bogolin, R. Jamieson, D. R. Marenda, D. A. Mackey,
Howell N., P. F. Chinnery, S. S. Ghosh, E. Fahy, D. M. Turnbull,
Howell N., I. Kubacka, S. Halvorson, B. Howell, D. A. McCullough, D. Mackey,
Marchington D. R., G. M. Hartshorne, D. Barlow, J. Poulton,
Richards M. B., V. A. Macauly, H.-J. Bandelt, B. C. Sykes,
Schlötterer C., D. Tautz,