Aging Aging

Mitochondrial DNA Mutations 275that the digestion has gone to completion. For the T10010C mutation analysis,there are two RsaI recognition sites in the wild type product, which generatefragments of 263, 53, and 34 basepairs. In the presence of the mutation, a site islost, leaving two fragments of 297 and 53 basepairs. For quantitation, the 263-basepairfragment is normalized to the 297-basepair fragment for deoxycytosine content, andthe level of heteroplasmy calculated as a percentage of the amount of radiolabelin the 263-basepair fragment relative to the combined amount in the 263 and297-basepair fragments (Fig. 1) (see also Notes 5–7).4. Notes1. Although dye primer cycle sequencing typically provides an even peak heightmaking it easier to detect heteroplasmic base changes, peak height is not directlyproportional to the level of mtDNA heteroplasmy within a sample. Consequently,proportions of mutant or wild-type mtDNA below 30% are unlikely to bedetected. If a novel change from the Cambridge sequence (8) is detected thatdoes not appear to be a recognized polymorphism, last hot cycle PCR-RFLPanalysis should be performed to exclude or confirm the presence of heteroplasmyand its level within the DNA sample.2. As previously mentioned, we recommend that both PCR primers are synthesizedwith the universal sequencing primer sequences at their 5' ends, thus allowingeach PCR product to be sequenced bidirectionally and any base substitutionsconfirmed. Moreover, there are occasionally small lengths of sequence (usuallyless than 5 basepairs) that cannot be read in one direction, presumably due to thenature of the DNA secondary structure. Sequencing of these samples in theopposite direction resolves this problem.3. The complementary DNA strands of the mitochondrial genome have an asymmetricdistribution of G’s and C’s, generating a heavy purine-rich H-strand, and alight pyrimidine-rich L-strand. Sequence data from the H-strand typically showhigher levels of background noise and poorer base-calling than that obtained fromsequencing the L-strand.4. A sample of the undigested, labeled PCR product must always be run on the gelto ensure that the restriction digest is 100% efficient.5. The most critical part of this method is the design of the RFLP to analyze themutation. Ideally, the PCR-generated fragment should have two restriction sitesfor the enzyme used, one of which is unique to the mutation. This can be eithergain or loss of a site. If the putative mutation does not create or destroy a naturallyoccurring restriction site, it is possible to engineer a site that is specific forthe mutation into the PCR fragment using a mismatch primer (7,18).6. Instead of performing a last hot cycle, the use of radioactivity can be avoided bythe addition of fluorescent-labeled deoxyuridinetriphosphates (dUTPs) to thefinal cycle of PCR, separating the restriction products by nondenaturing polyacrylamidegel electrophoresis (PAGE) (ABI Model 373A automated DNAsequencer [Applied Biosystems]), and quantitating the level of heteroplasmyusing Genescan software (Applied Biosystems) (19).