High levels of mitochondrial DNA heteroplasmy in human hairs by Budowle et al.

Abstract

The small size of the mitochondrial (mt) genome and its presence in many copies per cell make it useful for human DNA typing when only small amounts of material are available or when the material is badly degraded [1]. Despite its small size, and the fact that it is normally transmitted only maternally and without recombination, substantial DNA sequence polymorphism has accumulated in the mt genome, especially in two segments of the control region: HVI and HVII. The mtDNA sequence recovered from an item of evidence can thus be used to exclude many individuals as potential sources of that evidence. Conversely, if an evidence sequence matches a known reference sequence, the frequency of unrelated individuals who match the evidence sequence can be estimated by reference to a human mtDNA population database. The high copy number and high mutation rate that make mtDNA useful for forensic DNA typing also lead to mtDNA sequence heteroplasmy, the presence of more than one mtDNA sequence in an individual. Initially thought not to occur in the mitochondria of normal individuals [2], heteroplasmy is now known to be widespread. Major unresolved issues include the molecular mechanisms responsible for the occurrence of heteroplasmy to different extents in different tissues, and the possibility that heteroplasmy levels in an individual might vary with age [3–6]. Hairs are often strongly affected, so that different individual hairs from one person can differ in mtDNA sequence by one or more bases, or can show mixed sequences [7–10]. The article by Budowle et al. in the 28 March edition of Forensic Science International 126 (2002) 30–33 raises two issues in the course of evaluating the recent observation by Grzybowski [11,12] of very high levels of heteroplasmy in human hairs. The first issue is that detection of heteroplasmy is dependent on the exact laboratory procedures used to analyze mtDNA. The second is that interpretation of observed instances of heteroplasmy is dependent on models of mtDNA mutation and of human population genetics. Considering the relationship between assay conditions and amounts and kinds of heteroplasmy found, Budowle et al. note that both the amount of template DNA (20–80 ng) and the number of PCR cycles (30 þ 32, in a nested PCR strategy) in Grzybowski’s studies [11,12] differ from those used by the United States FBI Laboratory (0.1 ng target amount of DNA; 36 PCR cycles [13]). However, in validation studies carried out by Allen et al. [14], accurate mtDNA sequence information was reliably obtained with up to 33 ng of template DNA (the largest amount tested) in a 25 þ 32cycle nested PCR strategy, and in studies carried out to validate the British FSS protocol, accurate mtDNA sequence information was obtained with up to 200 ng of template in a nested PCR strategy with up to 32 cycles in the first step and up to 32 in the second [15,16]. Grzybowski’s assay conditions thus fall within the range validated by several laboratories for forensic mtDNA typing. At the same time, detection of heteroplasmy clearly is assay dependent. Such dependency was described by Sullivan et al. [7]. Calloway et al. [3] observed that their ability to detect heteroplasmy varied both with assay conditions and with the particular mtDNA sequence variants under study. Indeed, a recent inter-laboratory exercise provides a striking demonstration of the range of results possible when current forensic mtDNA typing protocols are applied to genuinely heteroplasmic hairs [10]. Continuing their discussion of experimental conditions that might confound mtDNA typing, Budowle et al. suggest that nuclear pseudogenes might be the source of some of the DNA sequences observed by Grzybowski [11]. Amplification of nuclear DNA corresponding to the HVI sequence has been reported, albeit only in samples depleted of mtDNA or containing mtDNA mutated so as to reduce its affinity for mt HVI PCR primers [17]. In contrast, amplification of nuclear DNA corresponding to mtDNA coding sequences such as Forensic Science International 130 (2002) 63–67