- Email: [email protected]
The rate of mitochondrial mutagenesis is faster in mice than humans
Mutation Research 377 Ž1997. 157–166
Accelerated publication
The rate of mitochondrial mutagenesis is faster in mice than humans Endi Wang, Alice Wong, Gino Cortopassi
)
Department of Molecular Biosciences, UniÕersity of California, DaÕis, CA 95616, USA Received 20 January 1997; revised 14 March 1997; accepted 17 March 1997
Abstract We have investigated mitochondrial DNA ŽmtDNA. mutagenesis in the laboratory mouse. Using a nested PCR method for quantification, the absolute frequency, tissue distribution and rate of increase of mitochondrial deletion mutations was determined. Multiple deletions arise in brain, cardiac muscle and kidney tissues; deletions occur most frequently at regions of directly repeated mtDNA homology. Deletion frequencies rose by 2.5 = 10 5, 6300- and 4000-fold in heart, brain and kidney, respectively, between young and old mice. The rates of mtDNA mutation accumulation in mouse and human hearts are modeled well by exponential equations, with r-values of 0.96 and 0.97, and mutations rose much faster in mouse than human mtDNA per unit time. Thus, maintenance of the human mitochondrial genome is much better than that of mice, consistent with the higher rate and final extent of total DNA repair in humans than mice, that has been observed by others and consistent with the predictions of the disposable soma model of aging. A comparison of mtDNA mutagenesis from cardiocytes vs. whole heart tissue was undertaken. Deletion mutations were observed to be 100-fold lower in DNA prepared from isolated cardiocytes than from whole heart homogenates, consistent with a model of uneven mtDNA mutation accumulation. q 1997 Elsevier Science B.V. Keywords: Mitochondrial; Mutagenesis; Aging; Mouse; DNA repair; Comparative
1. Introduction The mitochondrial genomes of humans w1–6x, monkeys w7x, rats w8,9x, mice w10–12x, and nematodes w13x accumulate mutations with age. Some of the deletion mutations that accumulate in humans with age are identical to those that occur in human mitochondrial genetic disease. In humans, mtDNA mutations accumulate preferentially in postmitotic tissues that are oxidatively active w2x; and in the aging brain, cells of the substantia nigra accumulate the highest levels of mitochondrial mutation w14x. Although the )
Corresponding author.
rise in mtDNA mutagenesis has been confirmed by several laboratories, the causes of this mutagenesis, and its relevance for cellular aging have yet to be determined. We have studied the occurrence, mutational spectra, cell-specific distribution, and rate of increase of mtDNA deletions in a short-lived experimental organism, the laboratory mouse. 2. Materials and methods 2.1. Animals and maintenance One- and 15- to 20-month-old female and male C57BLr6J mice were purchased from Jackson Lab-
0027-5107r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S 0 0 2 7 - 5 1 0 7 Ž 9 7 . 0 0 0 9 1 - 2
158
E. Wang et al.r Mutation Research 377 (1997) 157–166
oratory ŽBar Harbor, ME.. Very old C57BLr6J mice Ž31–35 months old. were purchased from the National Institute on Aging’s colony maintained by Charles River Laboratories ŽWilmington, DE.. At these facilities, mice were fed Purina Mouse Chow and were exposed to 12 h of light and 12 h of dark. Once received, the mice were housed at Animal Resource Services at the University of California, Davis. Mice were fed Purina Mouse Chow at ad libitum levels and were exposed to 12-h light–dark cycles. 2.2. DNA preparation, PCR strategy and DNA sequencing Total DNA was prepared from chopped mouse brain, heart and kidney by proteinase K digestion and phenolrchloroform extraction w15x; DNA yields were estimated spectrophotometrically. Oligonucleotide primers were synthesized ŽOperon Technologies Inc., San Pablo, CA. to anneal to DNA segments flanking three direct repeats in the regions 8884–9566 and 12956–13357 w16x. The sequences of primers are: F1, gcattagcagtccggcttac Ž8389– 8408.; F2, taattcaagcctacgtattc Ž8549–8569.; F3, caagtccatgaccattaactgg Ž8644–8665.; F4, catgatctaggaggctgctgacctc Ž8934–8958.; F5, aggaattttcctactggtccg Ž12500–12520.; R1, gggatgtttttaggcttagg Ž13364–13383.; R2, gattttatgggtgtaatgcg Ž13338– 13357.; primers F2, F3 and R1 are similar to those in w10x ŽFig. 1.. PCR conditions were initial denaturation at 948C for 4 min, followed by 30 cycles of denaturation, annealing and extension at 948C for 20 s, 558C for 20 s and 728C for 20 s, respectively, and last extension at 728C for 4 min; buffers were as described in w1,17x. Primers F2 and R1 were used as outer primers for the first 30 cycles, then 2 ml of the
first reaction were transferred to the nested PCR with inner primers F3 and R2 for the second 30 cycles. Products were electrophoresed through 1.5% agarose gels and stained with ethidium bromide. Primers F4 and R1 were used to estimate the relative quantity of undeleted mtDNA in each sample. For DNA sequence analysis, single deletion mutant PCR products were isolated by amplification of dilute samples. Because 60 cycles of amplification are sufficient to amplify a single deleted mtDNA molecule, dilution of samples to the point at which single molecules are amplified is equivalent to cloning in bacterial plasmids, as molecules pass through a population bottleneck of one molecule, which was verified by gel electrophoresis, that gave a single amplified product, and DNA sequencing, that gave a single sequence. Products were purified from contaminants by ultrafiltration on a Centricon 100 membrane, and sequenced on an automated 373A Sequencer ŽApplied Biosystems, Foster City, CA.. 2.3. A single deleted mtDNA molecule is detected by nested PCR The sensitivity of the assay of deleted mtDNA genomes was reconstructed by dilution of a known number of deleted mtDNA molecules. The 989-bp product was generated by amplification with primer F2rR1 and F3rR2 using DNA from mouse brain extract. The product was gel purified, and the number of 989-bp product molecules was estimated spectrophotometrically by absorbance at 260 nm. After calculating the number of moleculesrml of sample, 1, 10, 100, and 1000 molecules of PCR product were diluted into 500 ng of human DNA, in quadruplicate. Fully nested PCR, using primers F4 and R2, was performed with the same conditions as mentioned
Fig. 1. Location of mouse and human PCR primers. The primer positions for both mouse and human are shown. Positions of the directly repeated sequences in mouse mtDNA are also shown.
E. Wang et al.r Mutation Research 377 (1997) 157–166
159
2.5. Human DNA preparation and PCR strategy
Fig. 2. Reconstruction of the sensitivity of an assay for deleted mitochondrial genomes. A known number of molecules of a synthetic mitochondrial deletion were serially diluted and mixed with 500 ng of human genomic DNA, in quadruplicate. The numbers on top of the gel indicate the average number of target molecules in each PCR. M, size marker fX174DNArHaeIII fragments of 1353, 1078, 872, 603, 310, 271r281 and 234 bp are visible in order from the top of the lane. Detection of the target signal became stochastic at a concentration of approximately a single template molecule, consistent with Poisson expectations.
above; the predicted product is 699 bp in length. Detection of the mutants became stochastic at a concentration of 1 deletion molecule ŽFig. 2., consistent with Poisson expectations. 2.4. Confirmation of detected deletion products by direct DNA sequencing PCR products of length 989, 846, and 476 bp Ži.e., the most frequently detected products. were directly sequenced, as were the less frequently observed products of size 1251, 753, 658, 337 and 111 bp ŽTable 1.. The product size, junction sites and repeat length of these three most commonly observed deletions were: 989, 9553–13279, 14 bp; 846, 9089–12956, 15 bp; and 476, 8884–13120, 13 bp.
The most frequent age-related mtDNA deletion in human heart tissue erases 4977 bp of DNA w1–6x. DNA was prepared from cardiac tissue from newborn to 79-year-old humans taken at autopsy at Los Angeles County Hospital as described above. mtDNA deletions were detected with primers mt1rmt4 and mt1rmt3 of w1x: mt1, 8220–8247; mt2, 13176– 13198; mt3, 13476–13496; mt4, 13707–13728 ŽFig. 1.. Primer mt2 and primer mt4 were used as an internal control, for the normalization of mtDNA quantity. PCR conditions were: denaturation for 20 s at 948C; combined annealing and extension for 20 s at 608C, for 30 cycles. Afterwards, 2 ml of the PCR product was subjected to an additional 30 cycles of PCR using primers mt1 and mt3 Žnested PCR., under the same conditions described. Limiting dilution PCRs were performed to determine the frequency of mtDNA deletion in each individual sample as described below. 2.6. QuantitatiÕe determination of mtDNA deletion frequency in mouse and human tissues The relative concentration of deleted mtDNA genomes was determined by limiting serial dilution. Each mouse and human DNA sample was serially diluted to a point Ž Dcrit . at which no mtDNA deletion product could be detected by nested PCR. To convert to deleted mtDNAs per microgram input
Table 1 Junction structure of mouse mtDNA deletions Product size Žbp.
Deletion size Žbp.
Breakpoints
Junction structure
476 989 846 658 111 753 337 1251
4236 3723 3867 4056 4603 3961 4377 3463
8883r13120 9554r13278 9088r12956 8992r13049 8697r13301 9262r13224 8956r13334 9627r13091
ttttTCTTTGCAGGATTcttc . . . agcaTCTTTGCAGGATTtgtc tcagAAGCAAATCCATATgaat . . . aaatAAAGCAAATCCATATtcat atcaAGCCCTACTAATTACcatt . . . cctgAGCCCTACTAATTACacta aagtCCCacta . . . ttttCCCcccc gtctagtt . . . ccttctca gcctacta . . . attaggat gacctcca . . . ttattcac ttggtagc . . . taaaccca
Deletions were amplified and sequenced as described in Section 2: Materials and methods. The first four deletions were at direct repeats. The repeated sequences are shown in capital letters and lower-case letters indicate surrounding sequences. The last four deletions Ž111, 753, 337, and 1251 bp. did not contain direct repeats.
160
E. Wang et al.r Mutation Research 377 (1997) 157–166
DNA, it was assumed that a single deleted genome is detectable, an assumption consistent with our nested PCR data presented above and in Fig. 1. Thus 1rDcrit was used as the quantitative measure of mutation frequency. For comparison of mutation frequencies between DNA samples, the base-10 logarithm of 1rDcrit was used of 5 or more animals in any tissue or age group. Geometric means and geometric standard deviations of each determination were used to calculate withingroup and between-group variation w18x. 2.7. Isolation of dispersed mouse cardiocytes Twenty-month-old mice were sacrificed by cervical dislocation, hearts were excised and washed with Hanks balanced salt solution, and chopped with a razor blade. Minced hearts were resuspended in 2 ml of Krebs–Henseleit buffer and 0.1% trypsin at 378C for 40 min, a treatment which produced maximum yields of cardiocytes. Cardiocyte preparations were clearly single-cell suspensions and binucleated cells characteristic of cardiocytes were visible. Cardiocyte concentration was determined hemocytometrically, and cells were aliquoted in pools. Cell preparations were pelleted at 3500 rpm for 5 min and frozen at y208C. Cells were lysed and DNA prepared by application of 20 ml GeneReleaser ŽBioVentures, Inc. Murfreesboro, TN.. DNA thus prepared was subjected to PCR analysis as above. 3. Results 3.1. Mitochondrial DNA mutation frequency rises with age in mice If mitochondrial mutations increase with age, then one would expect that a smaller amount of input DNA would be required to detect mutations in old animals than young ones. The results of amplification of 10 and 500 ng brain DNA from young Ž4–7 weeks. and old Ž15–20 months. mice appears in Fig. 3. Multiple deletion products are detected in 10 ng brain DNA from old mice, but no products are detected in the same amount of DNA from young mice ŽFig. 3, upper panel.. These deletion products were from 476 to 1251 bp in length. At 500 ng input DNA, deletion products were amplified in both young
Fig. 3. Detection of multiple mtDNA deletions in brain tissue from 5 independent young and 5 old mice. Upper panel: 10 ng input. Lower panel: 500 ng input. Upper panel: lanes 1–5, mice 1–5, 1 month of age; lanes 6–10, mice 6–10, 15–20 months of age. Lower panel: lanes 1–5, mice 1–5; lanes 6–10, mice 6–10. Numbers at right indicate the size of the expected product. M sf174 DNA digested with HaeIII enzyme, fragments of size 1353, 1078, 872 and 603 are visible in order from the top of the lane.
and old brain DNA. However, every sample from old brain DNA shows multiple bands, whereas the number of bands is less in young mice ŽFig. 3, lower panel.. Deletion product sizes of 989-, 846-, 658-, and 476-bp deletions were observed most often in aging brain, heart and kidney. These most common amplification products occur between directly repeated sequences in the mtDNA ŽTable 1., and as observed in w10,11x. The 658-bp fragment is the product of a deletion that has occurred between runs of cytosines which have been observed by others to be relatively unstable in mtDNA w19–23x. We observed the 976-, 846-, 658-, and 476-bp fragments most often in aged tissues. Other deletion products, such as the 753-, 111-, 337-, and 1251-bp products occur less frequently on a per-site basis, and are not associated with direct repeats ŽTable 1.. The 989 and 846 products were observed about 3–4 times more often than the 476 and 658 products in dilute samples, and are always the major bands observed in concentrated samples, consistent with a either a higher intrinsic rate of deletion mutagenesis at repeats, or faster rate of replication once they occur. 3.2. Quantification of age-related mtDNA mutagenesis Mitochondrial mutation frequencies were determined in young animals of 1 month of age, in older
E. Wang et al.r Mutation Research 377 (1997) 157–166
161
Fig. 4. Frequency of mtDNA deletions in heart, brain and kidney of mice. mtDNA mutation frequencies were determined as in Section 2: Materials and methods. Geometric means and geometric standard errors of 1rDcrit for each group were calculated by logarithmic transformation, and transformed back antilogarithmically. Bar heights represent mean values from 5 mice, error bars are geometric standard errors.
animals of 15–20 months of age, and in much older animals of 31–35 months of age, by the 1rDcrit method. Deletions rose by 2.5 = 10 5-fold in heart, 6300-fold in brain and 4000-fold in kidney between
young and extremely old mice Ž31–35 months of age.. The smaller rise in brain is a result of a higher mutational load at a young age ŽFig. 4.. The concentration of deleted mtDNA genomes in young and old
Table 2 Deletion mutation frequency in mice Age
Tissue
Dcrit
a
mg DNA
Žmonths. 1 1 1 15–20 15–20 15–20 31–35 31–35 31–35
Mutation frequency mtDNAs
H B K H B K H B K
2.5 40 1.6 10 000 3 900 2 500 6.3 = 10 5 1.6 = 10 5 5 000
0.4 0.025 0.6 0.0001 0.0003 0.0004 1.6 = 10y6 4.0 = 10y6 1.6 = 10y4
8
1.2 = 10 7.5 = 10 6 1.8 = 10 8 3.0 = 10 4 9.0 = 10 4 1.2 = 10 5 4.8 = 10 2 1.2 = 10 3 4.8 = 10 4
b
Lower bound y9
8.3 = 10 1.3 = 10y7 5.6 = 10y9 3.3 = 10y5 1.1 = 10y5 8.3 = 10y6 2.1 = 10y3 8.4 = 10y4 2.1 = 10y5
c
Mean
d
In cells y8
4.2 = 10 6.5 = 10y7 2.8 = 10y8 1.7 = 10y4 5.5 = 10y5 4.2 = 10y5 1.1 = 10y2 4.2 = 10y3 1.1 = 10y4
e
8.4 = 10y5 1.3 = 10y3 5.6 = 10y5 0.333 0.110 0.083 21.0 8.4 0.210
Tissues H, B, K are heart, brain, and kidney, respectively. a Dcrit is the geometric mean of Dcrit from 5 or more separate animals. b It was assumed that in 1 mg of DNA there are 150 000 diploid nuclear genomes, and 2000 mtDNAs per cell, i.e., 3 = 10 8 mtDNAsrmg total DNA. c The lowest bound of mutation frequency per mitochondrial genome, i.e., the inverse of column 5. d Mean mutation frequency per mitochondrial genome, corrected. It was assumed that if all mutation signals disappeared at a particular 10-fold dilution, that the best estimate of the actual mutation frequency was 5 = higher than that concentration Ži.e., midway between 10-fold serial dilutions.. e Mutant molecules per cells, using the assumption there are 2000 mtDNAsrcell.
162
E. Wang et al.r Mutation Research 377 (1997) 157–166
mutation frequency was observed when cardiocytes are used as the DNA source as opposed to DNA from tissue homogenates. 3.4. The rise in mtDNA deletion mutations follows exponential kinetics
Fig. 5. Mutagenesis frequencies in mtDNA from mouse cardiocytes and heart homogenates. Hearts from 20-month-old C57BLr6 mice were bisected, one half was prepared for homogenates, the other half for cell dispersal. PCR was performed in triplicate for each sample for the specified number of cells. Upper panel: PCR of DNA extracted from homogenates. Lower panel: PCR of DNA extracted from heart cardiocytes. The estimated number of cells are indicated on the top of each gel. q, positive control; M s f 174 DNA digested with HaeIII enzyme, fragments of size 1353, 1078, 872 and 603 are visible in order from the top of the lane.
mice was determined ŽTable 2.. The fraction of deleted mtDNA per total mtDNA was highest in very old mice. Heart DNA deletions were 1.0 = 10y2 of total mtDNA from 31- to 35-month-old mice DNA. Brain DNA deletions were 4.2 = 10y3 of total mtDNA, followed by kidney DNA at 1.1 = 10y4 . mtDNA from young mice had much smaller fractions of deleted mtDNA, with heart, brain, and kidney at 4.2 = 10y8 , 6.5 = 10y7 and 2.8 = 10y8 , respectively.
mtDNA deletion mutations were amplified by PCR from both human and mouse heart DNA. The most common mtDNA deletion found in humans is the 4977-bp deletion w1–6x. We have previously used the same PCR conditions described in this study to detect the common deletion Ž4977 bp. in adult heart tissue w1x. In the current study, cardiac tissue from newborn to 79-year-old humans was used to study the kinetics of mtDNA deletions. A 360-bp fragment was amplified if the 4977-bp deletion was present. In mouse hearts, PCR detected the same mtDNA deletions as shown in Fig. 3. mtDNA deletion mutations rose exponentially in mouse and human cardiac tissues, which were fitted to an exponential equation ŽFig. 6.. The rate of rise of mutations can be modeled by the equation y s a = 10 b x , in which y s mitochondrial mutation frequency, a s the y intercept, b s a constant that describes rate of growth and x s time in months. For mouse mutations, the curve fitting program determined that b s 0.2; and the fit
3.3. Analysis of distribution of somatic mtDNA mutations into cells To determine the distribution of mtDNA deletions in different cell populations from the same organ, total DNA from cardiocytes and heart homogenates were prepared from 20-month-old C57BLr6 mice. By using the same PCR conditions that were used to amplify the deletions seen in Fig. 2, we detected the same size deletion products in mouse homogenates and cardiocytes ŽFig. 5.. mtDNA deletions in mtDNA were detected at 3000 cardiocytes-worth of DNA and higher. By comparison, only 30 cells-worth of heart homogenate DNA was required to detect a mutation signal. Thus a 100-fold lower apparent
Fig. 6. Kinetics of mitochondrial mutagenesis and tumorigenesis with age in mice and humans. Mitochondrial mutations in mouse and human heart DNAs were amplified by PCR as described in Section 2: Materials and methods. The fraction of mutant to total mtDNA was determined by the Dcrit method, assuming 2000 mtDNAs per cardiocyte ŽTable 2 legend.. Each open circle represents an individual determination for humans, the filled circles represent individual mice.
E. Wang et al.r Mutation Research 377 (1997) 157–166
of the data to the line was good, r s 0.96. For the human data, the best-fit curve described an equation in which b s 0.005; and r was 0.97. The time required for human mutations to double was 5 years, whereas it was 1.5 months in mice.
4. Discussion 4.1. Mitochondrial mutagenesis increases in heart, brain and kidney of aged mice We observe, as have others w10–12x that mtDNA deletion mutations accumulate with age in laboratory mice. The assay with which deletion mutations are detected is quantitative and sensitive, with the ability to detect a single mutant genome in a background of approximately 150 000 000 contaminating mitochondrial genomes ŽFig. 2.. mtDNA mutations rise by factors of 2.5 = 10 5-fold in heart, 6300-fold in brain and 4000-fold in kidney between young and very old mice ŽFig. 4.. 4.2. mtDNA deletions mutations occur more frequently at direct repeats than elsewhere We observe that in aging mice, mtDNA deletions occur most frequently at directly repeated sequences ŽFig. 3 and Table 1.. In dilute samples, the 989- and 846-bp products, which are associated with direct repeats, occurred about 3–4 times more often than the 658- and the 476-bp products. Other mutations occurred much less frequently on a per-site basis, consistent with a lower intrinsic rate of mutagenesis. These results are also consistent with observations in aging humans and in Kearns–Sayre syndrome, in which deletions occur most frequently at directly repeated elements w1–6x. Direct repeat sequences are considered hotspots for large-scale deletions w10,11,19x. At 500 ng in the older animals, the direct repeat associated deletion fragments are present in multiple copies, and dominate the PCR product spectra. The 658-bp deletion product occurs more frequently than deletions at non-repeated sites, consistent with observations by others that runs of cytosines in mtDNA are prone to rearrangement w20,21x. The presence of homopurine or homopyrimidine re-
163
gions may alter mtDNA structure and increase the possibility of DNA bending and strand breakage w22,23x. 4.3. Data are consistent with an uneÕen distribution of mtDNA mutations into cells Most previous estimates of age-related mitochondrial mutagenesis have relied on DNA prepared from homogenized tissue. However, the use of isolated cells to estimate mitochondrial mutagenesis may be appropriate if mitochondrial mutations are distributed completely evenly with respect to cells. As a result, mutation frequency data resulting from tissue homogenates should be identical to mutation data resulting from pools of isolated cells. However if mitochondrial mutagenesis occurs as somatic mutational ‘jackpots’, then some cells will harbor more mtDNA mutations than others. Thus, if cells were aliquoted into pools, the mutation frequency determined from these cells should be lower than deletion frequency determined from homogenates. We observe a 100-fold lower frequency of mutations in isolated cells than tissue homogenates ŽFig. 5.. The mtDNA from randomly pooled isolated cells is not representative of the cells with mutant mtDNA. The simplest explanation of this observation is that agerelated mtDNA mutations are grouped, consistent with the results of Muller-Hocker et al. w24x and Brierly et al. Žpersonal communication.. 4.4. The burden of mtDNA mutation in aging mice Large increases in mtDNA mutation occur with age in mice. However, the mean burden of deletions that we observe in a mouse of 2.7 years is about 10 mutations per cell in heart tissue ŽTable 2.. Thus, if these mtDNA mutations were evenly distributed and occurred with age, they might not be likely to cause severe mitochondrial dysfunction. However, we do observe support for an uneven distribution of mtDNA mutations, and also note that the assay utilized likely underestimates the total deletion mutation frequency, since we only examined a third of the mitochondrial genome. In addition, this method will not detect point mutations, which also may accumulate with age w25,26x.
164
E. Wang et al.r Mutation Research 377 (1997) 157–166
4.5. The molecular maintenance of mouse mtDNA per unit time is much worse than of human mtDNA Our quantitative data on mtDNA mutagenesis in mouse and human hearts fit exponential equations well, with r-values of 0.96 and 0.97 ŽFig. 6.. The ratio of the exponents Ž bmouserb human. is 40; thus we observe a 40-fold faster rate of mtDNA mutagenesis in mice per unit time than humans. This indicates that maintenance of the mouse mitochondrial genome is at least 40-fold worse per unit time than that of humans. These results in mice are consistent with the results of Edris et al. who demonstrated in rats that mtDNA mutation accumulation is correlated with longevity rather than time w9x. The exponential nature of both mouse and human mutagenesis curves may indicate the occurrence of a non-linear mutational process; one potential explanation for non-linearity is selective replication of deleted genomes. 4.6. Why are mtDNA mutagenesis rates so different between mice and humans? The very large differences in mtDNA maintenance between mice and humans could be explained by two types of complementary models.
mans w29x. However, quantitative comparisons of the rate of mtDNA repair in mice and humans have not yet been performed. What has been demonstrated is that in humans, rats, and hamsters, some mtDNA lesions are repaired rapidly, some slowly, and some not at all. Oxidized bases, abasic sites and single-strand breaks are repaired efficiently in rat and human cells exposed to alloxan, 69 and 100% of the damage was repaired in 6 h, respectively w30,31x. A group of slowly repaired lesions in human, rat and hamster cells as defined by Shen et al. w32x include methylated purines and intrastrand X-links, with 70–77% of the damage repaired in 24 h w33–35x. A third class of bulky lesions, UV-induced pyrimidine dimers and complex alkylation products are not repaired in human or hamster or rat mtDNA w34,36x. Therefore, there is evidence that some lesions are efficiently repaired from human, rat and hamster mtDNA; however the relative rates of ‘global’ mtDNA repair have not yet been compared between mice and humans. Thus, two possibly complementary explanations of the high rate of mouse mtDNA mutagenesis are that mice are subject to a higher burden of oxidative damage, and secondly that mice may be partially deficient in repair of mtDNA lesions. 4.9. ComparatiÕe mtDNA mutagenesis rates
4.7. Mouse mtDNA may be exposed to a higher dose of damage Mice consume 8-fold more molecular oxygen per gram of body weight than do humans w27x. Molecular oxygen is metabolized primarily in mitochondria, and its metabolism can produce reactive oxygen species ŽROS., which are known to damage DNA w28x. Thus, their mitochondrial genomes may receive a much higher dose of reactive oxygen species than do human mitochondrial genomes. 4.8. Mouse mtDNA may be repaired less efficiently than human mtDNA A lower rate of repair may underlie the rapid rate of mouse mtDNA mutagenesis. Species comparisons of DNA repair and unscheduled DNA synthesis have consistently demonstrated that global measures of repair in mice are about 5-fold slower than in hu-
Data presented here indicate that humans protect their mitochondrial genomes much better per unit time than do mice. There is a 252 000-fold increase in mitochondrial mutations detectable by the method between the ages of 1 and 33 months in mouse heart, and rises of 6300-fold in brain and 4000-fold in kidney. By contrast there is little or no increase in human mtDNA mutagenesis in this time frame. Thus, while the high rate of mouse mitochondrial mutagenesis relative to humans does not necessarily indicate that mtDNA mutagenesis is a molecular barrier that itself defines lifespan, the relatively high rate of somatic mitochondrial mutagenesis in mice is inconsistent with a lifespan of 10 years, a time to which most humans survive without overt mitochondrial pathology. A relatively lower investment of metabolic energy into DNA repair in species which experience high rates of predation is a prediction of the disposable soma theory of aging which is reviewed in w29x.
E. Wang et al.r Mutation Research 377 (1997) 157–166
Acknowledgements We thank Daryl Davies, Janet Huang, Timothy K. Gallaher and the NIH ŽAG11967. for assistance and research support.
w13x
w14x
w15x
References w16x w1x G.A. Cortopassi, N. Arnheim, Detection of a specific mitochondrial DNA deletion in tissues of older humans, Nucleic Acids Res. 18 Ž1990. 6927–6933. w2x G.A. Cortopassi, D. Shibata, N.-W. Soong, N. Arnheim, A pattern of accumulation of a somatic deletion of mitochondrial DNA in aging human tissues, Proc. Natl. Acad. Sci. USA 89 Ž1992. 7370–7374. w3x S. Ikebe, M. Tanaka, K. Ohno, W. Sato, K. Hattori, T. Kondo, Y. Mizuno, T. Ozawa, Increase of deleted mitochondrial DNA in the striatum in Parkinson’s disease and senescence, Biochem. Biophys. Res. Commun. 170 Ž1990. 1044– 1048. w4x A.W. Linnane, A. Baumer, R.J. Maxwell, H. Preston, C.F. Zhang, S. Marzuki, Mitochondrial gene mutation: the ageing process and degenerative diseases, Biochem. Int. 22 Ž1990. 1067–1076. w5x M. Corral-Debrinski, T. Horton, M.T. Lott, J.M. Shoffner, M.F. Beal, D.C. Wallace, Mitochondrial DNA deletions in human brain: regional variability and increase with advanced age, Nature Genet. 2 Ž1992. 324–329. w6x S. Simonetti, X. Chen, S. DiMauro, E.A. Schon, Accumulation of deletions in human mitochondrial DNA during normal aging: analysis by quantitative PCR, Biochim. Biophys. Acta 1180 Ž1992. 113–122. w7x C.M. Lee, S.S. Chung, J.M. Kaczkowski, R. Weindruch, J.M. Aiken, Multiple mitochondrial DNA deletions associated with age in skeletal muscle of rhesus monkeys, J. Gerontol. 48 Ž1993. B201–B205. w8x M.N. Gadaleta, V. Petruzzella, L. Daddabbo, C. Olivieri, F. Fracasso, P. Loguercio-Polosa, P. Cantatore, Mitochondrial DNA transcription and translation in aged rat. Effect of acetyl-L-carnitine, Ann. New York Acad. Sci. 717 Ž1994. 150–160. w9x W. Edris, B. Burgett, O.C. Stine, C.R. Filburn, Detection and quantitation by competitive PCR Of an age-associated increase in a 4.8-kb deletion in rat mitochondrial DNA, Mutation Res. 316 Ž1994. 69–78. w10x J.-Y. Brossas, E. Barreau, Y. Courtois, J. Treton, Multiple deletions in mitochondrial DNA are present in senescent mouse brain, Biochem. Biophys. Res. Commun. 202 Ž1994. 654–659. w11x S.S. Chung, R. Weindruch, S.R. Schwarze, D.I. Mckenzie, J.M. Aiken, Multiple age-associated mitochondrial DNA deletions in skeletal muscle of mice, Aging 6 Ž1994. 193–199. w12x S.M. Tanhauser, P.J. Laipis, Multiple deletions are detectable
w17x
w18x w19x
w20x
w21x
w22x
w23x
w24x
w25x
w26x
w27x
w28x
165
in mitochondrial DNA of aging mice, J. Biol. Chem. 270 Ž1995. 24769–24775. S. Melov, G.Z. Hertz, G.D. Stormo, T.E. Johnson, Detection of deletions in the mitochondrial genome of Caenorhabditis elegans, Nucleic Acids Res. 22 Ž1994. 1075–1078. N.-W. Soong, D.R. Hinton, G. Cortopassi, N. Arnheim, Mosaicism for a specific somatic mitochondrial DNA mutation in adult human brain, Nature Genet. 2 Ž1992. 318–323. J. Sambrook, T. Maniatis, E.F. Fritsch, Molecular Cloning: A Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989. M.J. Bibb, R.A. Van Etten, C.T. Wright, M.W. Walberg, D.A. Clayton, Sequence and gene organization of mouse mitochondrial DNA, Cell 26 Ž1981. 167–180. E. Wang, G.A. Cortopassi, Mice with duplications and deletions at the Tme locus have altered MnSOD activity, J. Biol. Chem. 269 Ž1994. 22463–22465. G.W. Snedecor, W.G. Cochran, Statistical Methods, Iowa State University Press, Ames, IA, 1981, pp. 291–292. E.A. Schon, R. Rizzuto, C.T. Moraes, H. Nakase, M. Zeviani, S. Dimauro, A direct repeat is a hotspot for large-scale deletion of human mitochondrial DNA. Evolution of ageing, Science 244 Ž1989. 346–349. A. Torroni, M.T. Lott, M.F. Cabell, Y.-S. Chen, L. Lavergne, D.C. Wallace, mtDNA and the origin of caucasians: identification of ancient caucasian-specific haplogroups, one of which is prone to a recurrent somatic duplicaton in the d-loop region, Am. J. Genet. 55 Ž1994. 760–776. G. Manfredi, S. Servidei, E. Bonilla, S. Shanske, E.A. Schon, S. DiMauro, C.T. Moraes, High levels of mitochondrial DNA with an unstable 260-bp duplication in a patient with a mitochondrial myopathy, Neurology 45 Ž1995. 762–768. A. Larsen, H. Weintraub, An altered DNA conformation detected by S1 nuclease occurs at specific regions in active chick globin chromatin, Cell 29 Ž1982. 609–622. R.D. Wells, D.A. Collier, J.C. Hanvey, M. Shimizu, F. Wohlrab, The chemistry and biology of unusual DNA structures adopted by oligopurine.oligopyrimidine sequences, FASEB J. 2 Ž1988. 2939–2949. J. Muller-Hocker, P. Seibel, K. Schneiderbanger, B. Kadenbach, Different in situ hybridization patterns of mitochondrial DNA in cytochrome c oxidase-deficient extraocular muscle fibers in the elderly, Virchows Arch. 422 Ž1993. 7–15. J. Lauber, C. Marsac, B. Kadenbach, P. Seibel, Mutations in mitochondrial tRNA genes: a frequent cause of neuromuscular diseases, Nucleic Acids Res. 19 Ž1991. 1393–1397. A.W. Linnane, C. Zhang, A. Baumer, P. Nagley, Mitochondrial DNA mutation and the ageing process: bioenergy and pharmacological intervention, Mutation Res. 275 Ž1992. 195–208. R.W. Hart, R.B. Setlow, Correlation between deoxyribonucleic acid excision-repair and life-span in a number of mammalian species, Proc. Natl. Acad. Sci. USA 71 Ž1974. 2169– 2173. B.N. Ames, M.K. Shigenaga, T.M. Hagen, Oxidants, antioxidants and the degenerative diseases of aging, Proc. Natl. Acad. Sci. USA 90 Ž1993. 7915–7922.
166
E. Wang et al.r Mutation Research 377 (1997) 157–166
w29x G.A. Cortopassi, E. Wang, There is substantial agreement among interspecies estimates of DNA repair activity, Mech. Ageing Dev. 91 Ž1996. 211–218. w30x W.J. Driggers, S.P. LeDoux, G.L. Wilson, Repair of oxidative damage within the mitochondrial DNA of RINr 38 cells, J. Biol. Chem. 268 Ž1993. 22042–22045. w31x W.J. Driggers, V.I. Grishko, S.P. LeDoux, G.L. Wilson, Defective repair of oxidative damage in the mitochondrial DNA of a xeroderma pigmentosum group A cell line, Cancer Res. 56 Ž1996. 1262–1266. w32x C.-C. Shen, W. Wertelecki, W.J. Driggers, S.P. LeDoux, G.L. Wilson, Repair of mitochondrial DNA damage induced by bleomycin in human cells, Mutation Res. 337 Ž1995. 19–23. w33x C.C. Pettepher, S.P. LeDoux, V.A. Bohr, G.L. Wilson, Re-
pair of alkali-labile sites within the mitochondrial DNA of RINr 38 cells after exposure to the nitrosourea Streptozotocin, J. Biol. Chem. 266 Ž1991. 3113–3117. w34x S.P. LeDoux, G.L. Wilson, E.J. Beecham, T. Stevnsner, K. Wassermann, V.A. Bohr, Repair of mitochondrial DNA after various types of DNA damage in Chinese hamster ovary cells, Carcinogenesis 13 Ž1992. 1967–1973. w35x S.P. LeDoux, N.J. Patton, L.J. Avery, G.L. Wilson, Repair of N-methylpurines in the mitochondria DNA of xeroderma pigmentosum complementation group D cells, Carcinogenesis 14 Ž1993. 913–917. w36x D.A. Clayton, J.N. Doda, E.C. Friedberg, The absence of a pyrimidine dimer repair mechanism in mammalian mitochondria, Proc. Natl. Acad. Sci. USA 71 Ž1974. 2777–2781.
Accelerated publication
The rate of mitochondrial mutagenesis is faster in mice than humans Endi Wang, Alice Wong, Gino Cortopassi
)
Department of Molecular Biosciences, UniÕersity of California, DaÕis, CA 95616, USA Received 20 January 1997; revised 14 March 1997; accepted 17 March 1997
Abstract We have investigated mitochondrial DNA ŽmtDNA. mutagenesis in the laboratory mouse. Using a nested PCR method for quantification, the absolute frequency, tissue distribution and rate of increase of mitochondrial deletion mutations was determined. Multiple deletions arise in brain, cardiac muscle and kidney tissues; deletions occur most frequently at regions of directly repeated mtDNA homology. Deletion frequencies rose by 2.5 = 10 5, 6300- and 4000-fold in heart, brain and kidney, respectively, between young and old mice. The rates of mtDNA mutation accumulation in mouse and human hearts are modeled well by exponential equations, with r-values of 0.96 and 0.97, and mutations rose much faster in mouse than human mtDNA per unit time. Thus, maintenance of the human mitochondrial genome is much better than that of mice, consistent with the higher rate and final extent of total DNA repair in humans than mice, that has been observed by others and consistent with the predictions of the disposable soma model of aging. A comparison of mtDNA mutagenesis from cardiocytes vs. whole heart tissue was undertaken. Deletion mutations were observed to be 100-fold lower in DNA prepared from isolated cardiocytes than from whole heart homogenates, consistent with a model of uneven mtDNA mutation accumulation. q 1997 Elsevier Science B.V. Keywords: Mitochondrial; Mutagenesis; Aging; Mouse; DNA repair; Comparative
1. Introduction The mitochondrial genomes of humans w1–6x, monkeys w7x, rats w8,9x, mice w10–12x, and nematodes w13x accumulate mutations with age. Some of the deletion mutations that accumulate in humans with age are identical to those that occur in human mitochondrial genetic disease. In humans, mtDNA mutations accumulate preferentially in postmitotic tissues that are oxidatively active w2x; and in the aging brain, cells of the substantia nigra accumulate the highest levels of mitochondrial mutation w14x. Although the )
Corresponding author.
rise in mtDNA mutagenesis has been confirmed by several laboratories, the causes of this mutagenesis, and its relevance for cellular aging have yet to be determined. We have studied the occurrence, mutational spectra, cell-specific distribution, and rate of increase of mtDNA deletions in a short-lived experimental organism, the laboratory mouse. 2. Materials and methods 2.1. Animals and maintenance One- and 15- to 20-month-old female and male C57BLr6J mice were purchased from Jackson Lab-
0027-5107r97r$17.00 q 1997 Elsevier Science B.V. All rights reserved. PII S 0 0 2 7 - 5 1 0 7 Ž 9 7 . 0 0 0 9 1 - 2
158
E. Wang et al.r Mutation Research 377 (1997) 157–166
oratory ŽBar Harbor, ME.. Very old C57BLr6J mice Ž31–35 months old. were purchased from the National Institute on Aging’s colony maintained by Charles River Laboratories ŽWilmington, DE.. At these facilities, mice were fed Purina Mouse Chow and were exposed to 12 h of light and 12 h of dark. Once received, the mice were housed at Animal Resource Services at the University of California, Davis. Mice were fed Purina Mouse Chow at ad libitum levels and were exposed to 12-h light–dark cycles. 2.2. DNA preparation, PCR strategy and DNA sequencing Total DNA was prepared from chopped mouse brain, heart and kidney by proteinase K digestion and phenolrchloroform extraction w15x; DNA yields were estimated spectrophotometrically. Oligonucleotide primers were synthesized ŽOperon Technologies Inc., San Pablo, CA. to anneal to DNA segments flanking three direct repeats in the regions 8884–9566 and 12956–13357 w16x. The sequences of primers are: F1, gcattagcagtccggcttac Ž8389– 8408.; F2, taattcaagcctacgtattc Ž8549–8569.; F3, caagtccatgaccattaactgg Ž8644–8665.; F4, catgatctaggaggctgctgacctc Ž8934–8958.; F5, aggaattttcctactggtccg Ž12500–12520.; R1, gggatgtttttaggcttagg Ž13364–13383.; R2, gattttatgggtgtaatgcg Ž13338– 13357.; primers F2, F3 and R1 are similar to those in w10x ŽFig. 1.. PCR conditions were initial denaturation at 948C for 4 min, followed by 30 cycles of denaturation, annealing and extension at 948C for 20 s, 558C for 20 s and 728C for 20 s, respectively, and last extension at 728C for 4 min; buffers were as described in w1,17x. Primers F2 and R1 were used as outer primers for the first 30 cycles, then 2 ml of the
first reaction were transferred to the nested PCR with inner primers F3 and R2 for the second 30 cycles. Products were electrophoresed through 1.5% agarose gels and stained with ethidium bromide. Primers F4 and R1 were used to estimate the relative quantity of undeleted mtDNA in each sample. For DNA sequence analysis, single deletion mutant PCR products were isolated by amplification of dilute samples. Because 60 cycles of amplification are sufficient to amplify a single deleted mtDNA molecule, dilution of samples to the point at which single molecules are amplified is equivalent to cloning in bacterial plasmids, as molecules pass through a population bottleneck of one molecule, which was verified by gel electrophoresis, that gave a single amplified product, and DNA sequencing, that gave a single sequence. Products were purified from contaminants by ultrafiltration on a Centricon 100 membrane, and sequenced on an automated 373A Sequencer ŽApplied Biosystems, Foster City, CA.. 2.3. A single deleted mtDNA molecule is detected by nested PCR The sensitivity of the assay of deleted mtDNA genomes was reconstructed by dilution of a known number of deleted mtDNA molecules. The 989-bp product was generated by amplification with primer F2rR1 and F3rR2 using DNA from mouse brain extract. The product was gel purified, and the number of 989-bp product molecules was estimated spectrophotometrically by absorbance at 260 nm. After calculating the number of moleculesrml of sample, 1, 10, 100, and 1000 molecules of PCR product were diluted into 500 ng of human DNA, in quadruplicate. Fully nested PCR, using primers F4 and R2, was performed with the same conditions as mentioned
Fig. 1. Location of mouse and human PCR primers. The primer positions for both mouse and human are shown. Positions of the directly repeated sequences in mouse mtDNA are also shown.
E. Wang et al.r Mutation Research 377 (1997) 157–166
159
2.5. Human DNA preparation and PCR strategy
Fig. 2. Reconstruction of the sensitivity of an assay for deleted mitochondrial genomes. A known number of molecules of a synthetic mitochondrial deletion were serially diluted and mixed with 500 ng of human genomic DNA, in quadruplicate. The numbers on top of the gel indicate the average number of target molecules in each PCR. M, size marker fX174DNArHaeIII fragments of 1353, 1078, 872, 603, 310, 271r281 and 234 bp are visible in order from the top of the lane. Detection of the target signal became stochastic at a concentration of approximately a single template molecule, consistent with Poisson expectations.
above; the predicted product is 699 bp in length. Detection of the mutants became stochastic at a concentration of 1 deletion molecule ŽFig. 2., consistent with Poisson expectations. 2.4. Confirmation of detected deletion products by direct DNA sequencing PCR products of length 989, 846, and 476 bp Ži.e., the most frequently detected products. were directly sequenced, as were the less frequently observed products of size 1251, 753, 658, 337 and 111 bp ŽTable 1.. The product size, junction sites and repeat length of these three most commonly observed deletions were: 989, 9553–13279, 14 bp; 846, 9089–12956, 15 bp; and 476, 8884–13120, 13 bp.
The most frequent age-related mtDNA deletion in human heart tissue erases 4977 bp of DNA w1–6x. DNA was prepared from cardiac tissue from newborn to 79-year-old humans taken at autopsy at Los Angeles County Hospital as described above. mtDNA deletions were detected with primers mt1rmt4 and mt1rmt3 of w1x: mt1, 8220–8247; mt2, 13176– 13198; mt3, 13476–13496; mt4, 13707–13728 ŽFig. 1.. Primer mt2 and primer mt4 were used as an internal control, for the normalization of mtDNA quantity. PCR conditions were: denaturation for 20 s at 948C; combined annealing and extension for 20 s at 608C, for 30 cycles. Afterwards, 2 ml of the PCR product was subjected to an additional 30 cycles of PCR using primers mt1 and mt3 Žnested PCR., under the same conditions described. Limiting dilution PCRs were performed to determine the frequency of mtDNA deletion in each individual sample as described below. 2.6. QuantitatiÕe determination of mtDNA deletion frequency in mouse and human tissues The relative concentration of deleted mtDNA genomes was determined by limiting serial dilution. Each mouse and human DNA sample was serially diluted to a point Ž Dcrit . at which no mtDNA deletion product could be detected by nested PCR. To convert to deleted mtDNAs per microgram input
Table 1 Junction structure of mouse mtDNA deletions Product size Žbp.
Deletion size Žbp.
Breakpoints
Junction structure
476 989 846 658 111 753 337 1251
4236 3723 3867 4056 4603 3961 4377 3463
8883r13120 9554r13278 9088r12956 8992r13049 8697r13301 9262r13224 8956r13334 9627r13091
ttttTCTTTGCAGGATTcttc . . . agcaTCTTTGCAGGATTtgtc tcagAAGCAAATCCATATgaat . . . aaatAAAGCAAATCCATATtcat atcaAGCCCTACTAATTACcatt . . . cctgAGCCCTACTAATTACacta aagtCCCacta . . . ttttCCCcccc gtctagtt . . . ccttctca gcctacta . . . attaggat gacctcca . . . ttattcac ttggtagc . . . taaaccca
Deletions were amplified and sequenced as described in Section 2: Materials and methods. The first four deletions were at direct repeats. The repeated sequences are shown in capital letters and lower-case letters indicate surrounding sequences. The last four deletions Ž111, 753, 337, and 1251 bp. did not contain direct repeats.
160
E. Wang et al.r Mutation Research 377 (1997) 157–166
DNA, it was assumed that a single deleted genome is detectable, an assumption consistent with our nested PCR data presented above and in Fig. 1. Thus 1rDcrit was used as the quantitative measure of mutation frequency. For comparison of mutation frequencies between DNA samples, the base-10 logarithm of 1rDcrit was used of 5 or more animals in any tissue or age group. Geometric means and geometric standard deviations of each determination were used to calculate withingroup and between-group variation w18x. 2.7. Isolation of dispersed mouse cardiocytes Twenty-month-old mice were sacrificed by cervical dislocation, hearts were excised and washed with Hanks balanced salt solution, and chopped with a razor blade. Minced hearts were resuspended in 2 ml of Krebs–Henseleit buffer and 0.1% trypsin at 378C for 40 min, a treatment which produced maximum yields of cardiocytes. Cardiocyte preparations were clearly single-cell suspensions and binucleated cells characteristic of cardiocytes were visible. Cardiocyte concentration was determined hemocytometrically, and cells were aliquoted in pools. Cell preparations were pelleted at 3500 rpm for 5 min and frozen at y208C. Cells were lysed and DNA prepared by application of 20 ml GeneReleaser ŽBioVentures, Inc. Murfreesboro, TN.. DNA thus prepared was subjected to PCR analysis as above. 3. Results 3.1. Mitochondrial DNA mutation frequency rises with age in mice If mitochondrial mutations increase with age, then one would expect that a smaller amount of input DNA would be required to detect mutations in old animals than young ones. The results of amplification of 10 and 500 ng brain DNA from young Ž4–7 weeks. and old Ž15–20 months. mice appears in Fig. 3. Multiple deletion products are detected in 10 ng brain DNA from old mice, but no products are detected in the same amount of DNA from young mice ŽFig. 3, upper panel.. These deletion products were from 476 to 1251 bp in length. At 500 ng input DNA, deletion products were amplified in both young
Fig. 3. Detection of multiple mtDNA deletions in brain tissue from 5 independent young and 5 old mice. Upper panel: 10 ng input. Lower panel: 500 ng input. Upper panel: lanes 1–5, mice 1–5, 1 month of age; lanes 6–10, mice 6–10, 15–20 months of age. Lower panel: lanes 1–5, mice 1–5; lanes 6–10, mice 6–10. Numbers at right indicate the size of the expected product. M sf174 DNA digested with HaeIII enzyme, fragments of size 1353, 1078, 872 and 603 are visible in order from the top of the lane.
and old brain DNA. However, every sample from old brain DNA shows multiple bands, whereas the number of bands is less in young mice ŽFig. 3, lower panel.. Deletion product sizes of 989-, 846-, 658-, and 476-bp deletions were observed most often in aging brain, heart and kidney. These most common amplification products occur between directly repeated sequences in the mtDNA ŽTable 1., and as observed in w10,11x. The 658-bp fragment is the product of a deletion that has occurred between runs of cytosines which have been observed by others to be relatively unstable in mtDNA w19–23x. We observed the 976-, 846-, 658-, and 476-bp fragments most often in aged tissues. Other deletion products, such as the 753-, 111-, 337-, and 1251-bp products occur less frequently on a per-site basis, and are not associated with direct repeats ŽTable 1.. The 989 and 846 products were observed about 3–4 times more often than the 476 and 658 products in dilute samples, and are always the major bands observed in concentrated samples, consistent with a either a higher intrinsic rate of deletion mutagenesis at repeats, or faster rate of replication once they occur. 3.2. Quantification of age-related mtDNA mutagenesis Mitochondrial mutation frequencies were determined in young animals of 1 month of age, in older
E. Wang et al.r Mutation Research 377 (1997) 157–166
161
Fig. 4. Frequency of mtDNA deletions in heart, brain and kidney of mice. mtDNA mutation frequencies were determined as in Section 2: Materials and methods. Geometric means and geometric standard errors of 1rDcrit for each group were calculated by logarithmic transformation, and transformed back antilogarithmically. Bar heights represent mean values from 5 mice, error bars are geometric standard errors.
animals of 15–20 months of age, and in much older animals of 31–35 months of age, by the 1rDcrit method. Deletions rose by 2.5 = 10 5-fold in heart, 6300-fold in brain and 4000-fold in kidney between
young and extremely old mice Ž31–35 months of age.. The smaller rise in brain is a result of a higher mutational load at a young age ŽFig. 4.. The concentration of deleted mtDNA genomes in young and old
Table 2 Deletion mutation frequency in mice Age
Tissue
Dcrit
a
mg DNA
Žmonths. 1 1 1 15–20 15–20 15–20 31–35 31–35 31–35
Mutation frequency mtDNAs
H B K H B K H B K
2.5 40 1.6 10 000 3 900 2 500 6.3 = 10 5 1.6 = 10 5 5 000
0.4 0.025 0.6 0.0001 0.0003 0.0004 1.6 = 10y6 4.0 = 10y6 1.6 = 10y4
8
1.2 = 10 7.5 = 10 6 1.8 = 10 8 3.0 = 10 4 9.0 = 10 4 1.2 = 10 5 4.8 = 10 2 1.2 = 10 3 4.8 = 10 4
b
Lower bound y9
8.3 = 10 1.3 = 10y7 5.6 = 10y9 3.3 = 10y5 1.1 = 10y5 8.3 = 10y6 2.1 = 10y3 8.4 = 10y4 2.1 = 10y5
c
Mean
d
In cells y8
4.2 = 10 6.5 = 10y7 2.8 = 10y8 1.7 = 10y4 5.5 = 10y5 4.2 = 10y5 1.1 = 10y2 4.2 = 10y3 1.1 = 10y4
e
8.4 = 10y5 1.3 = 10y3 5.6 = 10y5 0.333 0.110 0.083 21.0 8.4 0.210
Tissues H, B, K are heart, brain, and kidney, respectively. a Dcrit is the geometric mean of Dcrit from 5 or more separate animals. b It was assumed that in 1 mg of DNA there are 150 000 diploid nuclear genomes, and 2000 mtDNAs per cell, i.e., 3 = 10 8 mtDNAsrmg total DNA. c The lowest bound of mutation frequency per mitochondrial genome, i.e., the inverse of column 5. d Mean mutation frequency per mitochondrial genome, corrected. It was assumed that if all mutation signals disappeared at a particular 10-fold dilution, that the best estimate of the actual mutation frequency was 5 = higher than that concentration Ži.e., midway between 10-fold serial dilutions.. e Mutant molecules per cells, using the assumption there are 2000 mtDNAsrcell.
162
E. Wang et al.r Mutation Research 377 (1997) 157–166
mutation frequency was observed when cardiocytes are used as the DNA source as opposed to DNA from tissue homogenates. 3.4. The rise in mtDNA deletion mutations follows exponential kinetics
Fig. 5. Mutagenesis frequencies in mtDNA from mouse cardiocytes and heart homogenates. Hearts from 20-month-old C57BLr6 mice were bisected, one half was prepared for homogenates, the other half for cell dispersal. PCR was performed in triplicate for each sample for the specified number of cells. Upper panel: PCR of DNA extracted from homogenates. Lower panel: PCR of DNA extracted from heart cardiocytes. The estimated number of cells are indicated on the top of each gel. q, positive control; M s f 174 DNA digested with HaeIII enzyme, fragments of size 1353, 1078, 872 and 603 are visible in order from the top of the lane.
mice was determined ŽTable 2.. The fraction of deleted mtDNA per total mtDNA was highest in very old mice. Heart DNA deletions were 1.0 = 10y2 of total mtDNA from 31- to 35-month-old mice DNA. Brain DNA deletions were 4.2 = 10y3 of total mtDNA, followed by kidney DNA at 1.1 = 10y4 . mtDNA from young mice had much smaller fractions of deleted mtDNA, with heart, brain, and kidney at 4.2 = 10y8 , 6.5 = 10y7 and 2.8 = 10y8 , respectively.
mtDNA deletion mutations were amplified by PCR from both human and mouse heart DNA. The most common mtDNA deletion found in humans is the 4977-bp deletion w1–6x. We have previously used the same PCR conditions described in this study to detect the common deletion Ž4977 bp. in adult heart tissue w1x. In the current study, cardiac tissue from newborn to 79-year-old humans was used to study the kinetics of mtDNA deletions. A 360-bp fragment was amplified if the 4977-bp deletion was present. In mouse hearts, PCR detected the same mtDNA deletions as shown in Fig. 3. mtDNA deletion mutations rose exponentially in mouse and human cardiac tissues, which were fitted to an exponential equation ŽFig. 6.. The rate of rise of mutations can be modeled by the equation y s a = 10 b x , in which y s mitochondrial mutation frequency, a s the y intercept, b s a constant that describes rate of growth and x s time in months. For mouse mutations, the curve fitting program determined that b s 0.2; and the fit
3.3. Analysis of distribution of somatic mtDNA mutations into cells To determine the distribution of mtDNA deletions in different cell populations from the same organ, total DNA from cardiocytes and heart homogenates were prepared from 20-month-old C57BLr6 mice. By using the same PCR conditions that were used to amplify the deletions seen in Fig. 2, we detected the same size deletion products in mouse homogenates and cardiocytes ŽFig. 5.. mtDNA deletions in mtDNA were detected at 3000 cardiocytes-worth of DNA and higher. By comparison, only 30 cells-worth of heart homogenate DNA was required to detect a mutation signal. Thus a 100-fold lower apparent
Fig. 6. Kinetics of mitochondrial mutagenesis and tumorigenesis with age in mice and humans. Mitochondrial mutations in mouse and human heart DNAs were amplified by PCR as described in Section 2: Materials and methods. The fraction of mutant to total mtDNA was determined by the Dcrit method, assuming 2000 mtDNAs per cardiocyte ŽTable 2 legend.. Each open circle represents an individual determination for humans, the filled circles represent individual mice.
E. Wang et al.r Mutation Research 377 (1997) 157–166
of the data to the line was good, r s 0.96. For the human data, the best-fit curve described an equation in which b s 0.005; and r was 0.97. The time required for human mutations to double was 5 years, whereas it was 1.5 months in mice.
4. Discussion 4.1. Mitochondrial mutagenesis increases in heart, brain and kidney of aged mice We observe, as have others w10–12x that mtDNA deletion mutations accumulate with age in laboratory mice. The assay with which deletion mutations are detected is quantitative and sensitive, with the ability to detect a single mutant genome in a background of approximately 150 000 000 contaminating mitochondrial genomes ŽFig. 2.. mtDNA mutations rise by factors of 2.5 = 10 5-fold in heart, 6300-fold in brain and 4000-fold in kidney between young and very old mice ŽFig. 4.. 4.2. mtDNA deletions mutations occur more frequently at direct repeats than elsewhere We observe that in aging mice, mtDNA deletions occur most frequently at directly repeated sequences ŽFig. 3 and Table 1.. In dilute samples, the 989- and 846-bp products, which are associated with direct repeats, occurred about 3–4 times more often than the 658- and the 476-bp products. Other mutations occurred much less frequently on a per-site basis, consistent with a lower intrinsic rate of mutagenesis. These results are also consistent with observations in aging humans and in Kearns–Sayre syndrome, in which deletions occur most frequently at directly repeated elements w1–6x. Direct repeat sequences are considered hotspots for large-scale deletions w10,11,19x. At 500 ng in the older animals, the direct repeat associated deletion fragments are present in multiple copies, and dominate the PCR product spectra. The 658-bp deletion product occurs more frequently than deletions at non-repeated sites, consistent with observations by others that runs of cytosines in mtDNA are prone to rearrangement w20,21x. The presence of homopurine or homopyrimidine re-
163
gions may alter mtDNA structure and increase the possibility of DNA bending and strand breakage w22,23x. 4.3. Data are consistent with an uneÕen distribution of mtDNA mutations into cells Most previous estimates of age-related mitochondrial mutagenesis have relied on DNA prepared from homogenized tissue. However, the use of isolated cells to estimate mitochondrial mutagenesis may be appropriate if mitochondrial mutations are distributed completely evenly with respect to cells. As a result, mutation frequency data resulting from tissue homogenates should be identical to mutation data resulting from pools of isolated cells. However if mitochondrial mutagenesis occurs as somatic mutational ‘jackpots’, then some cells will harbor more mtDNA mutations than others. Thus, if cells were aliquoted into pools, the mutation frequency determined from these cells should be lower than deletion frequency determined from homogenates. We observe a 100-fold lower frequency of mutations in isolated cells than tissue homogenates ŽFig. 5.. The mtDNA from randomly pooled isolated cells is not representative of the cells with mutant mtDNA. The simplest explanation of this observation is that agerelated mtDNA mutations are grouped, consistent with the results of Muller-Hocker et al. w24x and Brierly et al. Žpersonal communication.. 4.4. The burden of mtDNA mutation in aging mice Large increases in mtDNA mutation occur with age in mice. However, the mean burden of deletions that we observe in a mouse of 2.7 years is about 10 mutations per cell in heart tissue ŽTable 2.. Thus, if these mtDNA mutations were evenly distributed and occurred with age, they might not be likely to cause severe mitochondrial dysfunction. However, we do observe support for an uneven distribution of mtDNA mutations, and also note that the assay utilized likely underestimates the total deletion mutation frequency, since we only examined a third of the mitochondrial genome. In addition, this method will not detect point mutations, which also may accumulate with age w25,26x.
164
E. Wang et al.r Mutation Research 377 (1997) 157–166
4.5. The molecular maintenance of mouse mtDNA per unit time is much worse than of human mtDNA Our quantitative data on mtDNA mutagenesis in mouse and human hearts fit exponential equations well, with r-values of 0.96 and 0.97 ŽFig. 6.. The ratio of the exponents Ž bmouserb human. is 40; thus we observe a 40-fold faster rate of mtDNA mutagenesis in mice per unit time than humans. This indicates that maintenance of the mouse mitochondrial genome is at least 40-fold worse per unit time than that of humans. These results in mice are consistent with the results of Edris et al. who demonstrated in rats that mtDNA mutation accumulation is correlated with longevity rather than time w9x. The exponential nature of both mouse and human mutagenesis curves may indicate the occurrence of a non-linear mutational process; one potential explanation for non-linearity is selective replication of deleted genomes. 4.6. Why are mtDNA mutagenesis rates so different between mice and humans? The very large differences in mtDNA maintenance between mice and humans could be explained by two types of complementary models.
mans w29x. However, quantitative comparisons of the rate of mtDNA repair in mice and humans have not yet been performed. What has been demonstrated is that in humans, rats, and hamsters, some mtDNA lesions are repaired rapidly, some slowly, and some not at all. Oxidized bases, abasic sites and single-strand breaks are repaired efficiently in rat and human cells exposed to alloxan, 69 and 100% of the damage was repaired in 6 h, respectively w30,31x. A group of slowly repaired lesions in human, rat and hamster cells as defined by Shen et al. w32x include methylated purines and intrastrand X-links, with 70–77% of the damage repaired in 24 h w33–35x. A third class of bulky lesions, UV-induced pyrimidine dimers and complex alkylation products are not repaired in human or hamster or rat mtDNA w34,36x. Therefore, there is evidence that some lesions are efficiently repaired from human, rat and hamster mtDNA; however the relative rates of ‘global’ mtDNA repair have not yet been compared between mice and humans. Thus, two possibly complementary explanations of the high rate of mouse mtDNA mutagenesis are that mice are subject to a higher burden of oxidative damage, and secondly that mice may be partially deficient in repair of mtDNA lesions. 4.9. ComparatiÕe mtDNA mutagenesis rates
4.7. Mouse mtDNA may be exposed to a higher dose of damage Mice consume 8-fold more molecular oxygen per gram of body weight than do humans w27x. Molecular oxygen is metabolized primarily in mitochondria, and its metabolism can produce reactive oxygen species ŽROS., which are known to damage DNA w28x. Thus, their mitochondrial genomes may receive a much higher dose of reactive oxygen species than do human mitochondrial genomes. 4.8. Mouse mtDNA may be repaired less efficiently than human mtDNA A lower rate of repair may underlie the rapid rate of mouse mtDNA mutagenesis. Species comparisons of DNA repair and unscheduled DNA synthesis have consistently demonstrated that global measures of repair in mice are about 5-fold slower than in hu-
Data presented here indicate that humans protect their mitochondrial genomes much better per unit time than do mice. There is a 252 000-fold increase in mitochondrial mutations detectable by the method between the ages of 1 and 33 months in mouse heart, and rises of 6300-fold in brain and 4000-fold in kidney. By contrast there is little or no increase in human mtDNA mutagenesis in this time frame. Thus, while the high rate of mouse mitochondrial mutagenesis relative to humans does not necessarily indicate that mtDNA mutagenesis is a molecular barrier that itself defines lifespan, the relatively high rate of somatic mitochondrial mutagenesis in mice is inconsistent with a lifespan of 10 years, a time to which most humans survive without overt mitochondrial pathology. A relatively lower investment of metabolic energy into DNA repair in species which experience high rates of predation is a prediction of the disposable soma theory of aging which is reviewed in w29x.
E. Wang et al.r Mutation Research 377 (1997) 157–166
Acknowledgements We thank Daryl Davies, Janet Huang, Timothy K. Gallaher and the NIH ŽAG11967. for assistance and research support.
w13x
w14x
w15x
References w16x w1x G.A. Cortopassi, N. Arnheim, Detection of a specific mitochondrial DNA deletion in tissues of older humans, Nucleic Acids Res. 18 Ž1990. 6927–6933. w2x G.A. Cortopassi, D. Shibata, N.-W. Soong, N. Arnheim, A pattern of accumulation of a somatic deletion of mitochondrial DNA in aging human tissues, Proc. Natl. Acad. Sci. USA 89 Ž1992. 7370–7374. w3x S. Ikebe, M. Tanaka, K. Ohno, W. Sato, K. Hattori, T. Kondo, Y. Mizuno, T. Ozawa, Increase of deleted mitochondrial DNA in the striatum in Parkinson’s disease and senescence, Biochem. Biophys. Res. Commun. 170 Ž1990. 1044– 1048. w4x A.W. Linnane, A. Baumer, R.J. Maxwell, H. Preston, C.F. Zhang, S. Marzuki, Mitochondrial gene mutation: the ageing process and degenerative diseases, Biochem. Int. 22 Ž1990. 1067–1076. w5x M. Corral-Debrinski, T. Horton, M.T. Lott, J.M. Shoffner, M.F. Beal, D.C. Wallace, Mitochondrial DNA deletions in human brain: regional variability and increase with advanced age, Nature Genet. 2 Ž1992. 324–329. w6x S. Simonetti, X. Chen, S. DiMauro, E.A. Schon, Accumulation of deletions in human mitochondrial DNA during normal aging: analysis by quantitative PCR, Biochim. Biophys. Acta 1180 Ž1992. 113–122. w7x C.M. Lee, S.S. Chung, J.M. Kaczkowski, R. Weindruch, J.M. Aiken, Multiple mitochondrial DNA deletions associated with age in skeletal muscle of rhesus monkeys, J. Gerontol. 48 Ž1993. B201–B205. w8x M.N. Gadaleta, V. Petruzzella, L. Daddabbo, C. Olivieri, F. Fracasso, P. Loguercio-Polosa, P. Cantatore, Mitochondrial DNA transcription and translation in aged rat. Effect of acetyl-L-carnitine, Ann. New York Acad. Sci. 717 Ž1994. 150–160. w9x W. Edris, B. Burgett, O.C. Stine, C.R. Filburn, Detection and quantitation by competitive PCR Of an age-associated increase in a 4.8-kb deletion in rat mitochondrial DNA, Mutation Res. 316 Ž1994. 69–78. w10x J.-Y. Brossas, E. Barreau, Y. Courtois, J. Treton, Multiple deletions in mitochondrial DNA are present in senescent mouse brain, Biochem. Biophys. Res. Commun. 202 Ž1994. 654–659. w11x S.S. Chung, R. Weindruch, S.R. Schwarze, D.I. Mckenzie, J.M. Aiken, Multiple age-associated mitochondrial DNA deletions in skeletal muscle of mice, Aging 6 Ž1994. 193–199. w12x S.M. Tanhauser, P.J. Laipis, Multiple deletions are detectable
w17x
w18x w19x
w20x
w21x
w22x
w23x
w24x
w25x
w26x
w27x
w28x
165
in mitochondrial DNA of aging mice, J. Biol. Chem. 270 Ž1995. 24769–24775. S. Melov, G.Z. Hertz, G.D. Stormo, T.E. Johnson, Detection of deletions in the mitochondrial genome of Caenorhabditis elegans, Nucleic Acids Res. 22 Ž1994. 1075–1078. N.-W. Soong, D.R. Hinton, G. Cortopassi, N. Arnheim, Mosaicism for a specific somatic mitochondrial DNA mutation in adult human brain, Nature Genet. 2 Ž1992. 318–323. J. Sambrook, T. Maniatis, E.F. Fritsch, Molecular Cloning: A Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1989. M.J. Bibb, R.A. Van Etten, C.T. Wright, M.W. Walberg, D.A. Clayton, Sequence and gene organization of mouse mitochondrial DNA, Cell 26 Ž1981. 167–180. E. Wang, G.A. Cortopassi, Mice with duplications and deletions at the Tme locus have altered MnSOD activity, J. Biol. Chem. 269 Ž1994. 22463–22465. G.W. Snedecor, W.G. Cochran, Statistical Methods, Iowa State University Press, Ames, IA, 1981, pp. 291–292. E.A. Schon, R. Rizzuto, C.T. Moraes, H. Nakase, M. Zeviani, S. Dimauro, A direct repeat is a hotspot for large-scale deletion of human mitochondrial DNA. Evolution of ageing, Science 244 Ž1989. 346–349. A. Torroni, M.T. Lott, M.F. Cabell, Y.-S. Chen, L. Lavergne, D.C. Wallace, mtDNA and the origin of caucasians: identification of ancient caucasian-specific haplogroups, one of which is prone to a recurrent somatic duplicaton in the d-loop region, Am. J. Genet. 55 Ž1994. 760–776. G. Manfredi, S. Servidei, E. Bonilla, S. Shanske, E.A. Schon, S. DiMauro, C.T. Moraes, High levels of mitochondrial DNA with an unstable 260-bp duplication in a patient with a mitochondrial myopathy, Neurology 45 Ž1995. 762–768. A. Larsen, H. Weintraub, An altered DNA conformation detected by S1 nuclease occurs at specific regions in active chick globin chromatin, Cell 29 Ž1982. 609–622. R.D. Wells, D.A. Collier, J.C. Hanvey, M. Shimizu, F. Wohlrab, The chemistry and biology of unusual DNA structures adopted by oligopurine.oligopyrimidine sequences, FASEB J. 2 Ž1988. 2939–2949. J. Muller-Hocker, P. Seibel, K. Schneiderbanger, B. Kadenbach, Different in situ hybridization patterns of mitochondrial DNA in cytochrome c oxidase-deficient extraocular muscle fibers in the elderly, Virchows Arch. 422 Ž1993. 7–15. J. Lauber, C. Marsac, B. Kadenbach, P. Seibel, Mutations in mitochondrial tRNA genes: a frequent cause of neuromuscular diseases, Nucleic Acids Res. 19 Ž1991. 1393–1397. A.W. Linnane, C. Zhang, A. Baumer, P. Nagley, Mitochondrial DNA mutation and the ageing process: bioenergy and pharmacological intervention, Mutation Res. 275 Ž1992. 195–208. R.W. Hart, R.B. Setlow, Correlation between deoxyribonucleic acid excision-repair and life-span in a number of mammalian species, Proc. Natl. Acad. Sci. USA 71 Ž1974. 2169– 2173. B.N. Ames, M.K. Shigenaga, T.M. Hagen, Oxidants, antioxidants and the degenerative diseases of aging, Proc. Natl. Acad. Sci. USA 90 Ž1993. 7915–7922.
166
E. Wang et al.r Mutation Research 377 (1997) 157–166
w29x G.A. Cortopassi, E. Wang, There is substantial agreement among interspecies estimates of DNA repair activity, Mech. Ageing Dev. 91 Ž1996. 211–218. w30x W.J. Driggers, S.P. LeDoux, G.L. Wilson, Repair of oxidative damage within the mitochondrial DNA of RINr 38 cells, J. Biol. Chem. 268 Ž1993. 22042–22045. w31x W.J. Driggers, V.I. Grishko, S.P. LeDoux, G.L. Wilson, Defective repair of oxidative damage in the mitochondrial DNA of a xeroderma pigmentosum group A cell line, Cancer Res. 56 Ž1996. 1262–1266. w32x C.-C. Shen, W. Wertelecki, W.J. Driggers, S.P. LeDoux, G.L. Wilson, Repair of mitochondrial DNA damage induced by bleomycin in human cells, Mutation Res. 337 Ž1995. 19–23. w33x C.C. Pettepher, S.P. LeDoux, V.A. Bohr, G.L. Wilson, Re-
pair of alkali-labile sites within the mitochondrial DNA of RINr 38 cells after exposure to the nitrosourea Streptozotocin, J. Biol. Chem. 266 Ž1991. 3113–3117. w34x S.P. LeDoux, G.L. Wilson, E.J. Beecham, T. Stevnsner, K. Wassermann, V.A. Bohr, Repair of mitochondrial DNA after various types of DNA damage in Chinese hamster ovary cells, Carcinogenesis 13 Ž1992. 1967–1973. w35x S.P. LeDoux, N.J. Patton, L.J. Avery, G.L. Wilson, Repair of N-methylpurines in the mitochondria DNA of xeroderma pigmentosum complementation group D cells, Carcinogenesis 14 Ž1993. 913–917. w36x D.A. Clayton, J.N. Doda, E.C. Friedberg, The absence of a pyrimidine dimer repair mechanism in mammalian mitochondria, Proc. Natl. Acad. Sci. USA 71 Ž1974. 2777–2781.