1.3. Mitochondrial DNA: the tool
1.3.1. Mitochondrial DNA
The animal mitochondrial DNA is a circular molecule of 15-20 kb in length and in vertebrates it contains genes for 22 tRNAs, 2 rRNAs and 13 mRNAs coding for proteins involved in electron transport and oxidative phosphorylation. The only major noncoding area of the mtDNA is the control region, typically 1 kb, involved in the regulation and initiation of mtDNA replication and transcription (see 1.3.2). The use of mtDNA has become increasingly popular in phylogenetic and population genetic studies, first with the developments in methodology for mtDNA isolation and use of restriction enzymes to detect nucleotide differences (Lansman et al. 1981), and further with the development of PCR methodology and applicability of ‘universal’ primers (Kocher et al. 1989) for amplification of mtDNA. Much of the interest is related to the fast rate of substitutions in mtDNA. The approximate mutation rate in mtDNA is 10-8/site/year (Brown et al. 1979, Ferris et al. 1983, De Salle et al. 1987) compared to 10-9/site/year in nuclear genes. Most differences between mtDNA sequences are point mutations, with a strong bias for transitions over transversions (Brown et al. 1982).
The mtDNA is haploid and uniparentally inherited (with some exceptions, see below) and thus the variability is introduced by mutations alone. Compared to diploid nuclear autosomal genes with biparental transmission, the effective population size of mtDNA is one quarter of that for nuclear autosomal genes (Moore 1995). Therefore, a mtDNA tree is more likely to be congruent with a species tree due to a high probability of coalescence even when speciation events have occurred within short time-periods.
Mitochondrial genes are inherited as one linkage group in the absence of recombination (Hayashi et al. 1985, Hoech et al. 1991). Recently, the clonal nature of mtDNA has been questioned and the possibility of recombination has been advocated based mainly on linkage disequilibrium in human and chimpanzee mtDNA, excess homoplasy in human control region and the existence of a globally rare transitional mutation found in more than one well-supported mtDNA clade present in one population (Awadalla et al. 1999, Eyre-Walker et al. 1999, Hagelberg et al. 1999). Subsequently, the methodology used has been criticised (Ingman et al. 2000, Kumar et al. 2000) and alternative explanations have been preferred, such as the presence of hypervariable nucleotide positions or selection (Wallis 1999). The necessary enzymatic machinery for recombination does exist in mammalian mitochondria (Thyagarajan et al. 1996), and mtDNA recombination has so far been shown to occur in plants, fungi, protists (Gray 1989) and phytonematodes (Lunt & Hyman 1997). The prerequisite for recombination to create new variants is that different types of mitochondria are present in the same cell. Although individuals usually carry one type of mtDNA in their cells, heteroplasmy (existence of more than one extranuclear DNA sequence type in an organism) has also been reported. The most common cases are length variants in repetitive areas found within an individual (e.g. Berg et al. 1995), but it has been suggested that these could be created de novo within a lineage (Lunt et al. 1998). Contrary to common belief, the sperm mitochondria have been shown to enter the oocyte in most mammals (Ankel-Simons & Cummins 1996) and paternal leakage or biparental inheritance of mtDNA have been reported (Kondo et al. 1990, Gyllensten et al. 1991, Zouros et al. 1992). However, in mice it has been shown that the paternal mitochondria are usually eliminated although the process of elimination does not work for interspecific crosses (Kaneda et al. 1995). Then, if mtDNA recombination takes place, the hybrids of deeply diverged lineages could produce recombinant genomes e.g. in hybrid zones (Wallis 1999), but as such this has yet to be reported.
1.3.2. Mitochondrial control region
The control region is the main regulatory region and the only major non-coding area in animal mtDNA. It contains the heavy-strand origin of replication (Desjardins & Morais 1990) and the promoters for heavy and light strand transcription (L’Abbé et al. 1991). The control region is situated in between tRNAglu and tRNAphe in most avian species studied and between tRNAthr and tRNApro in Picidae, suboscine and at least one oscine Passeriformes, Falconiformes and Cuculidae (Mindell et al. 1998, Bench & Härlid 2000). In avian species, the length of the control region varies from 1028 bp in Struthio camelus (Härlid et al. 1997) to 1580 bp in Cyanoramphus auriceps (Boon et al. 2000), and in mammals approximately from 880 to 1400 bp (Sbisà et al. 1997). The variation in length has been attributed to variation in the tandem repeat number (Berg et al. 1995) and small insertions/deletions usually in the 5’ and 3’ ends of the control region.
Despite its functional importance, the control region is suggested to be the most variable part of the mtDNA. In the human control region, the estimates of the rate of substitution were found to range between 2.8 (Cann et al. 1984) to 5 times (Aquadro & Greenberg 1983) the rate of the rest of the mtDNA. Most of the studies in which control region sequences have been used have focused on intraspecific patterns of variability and phylogenetic relationships of closely related species, a prominent example being the study of human population history (see Cavalli-Sforza et al. 1994, and references therein). A high mutation rate also means that the phylogenetic utility of the control region sequences diminishes in deep divergences due to saturation and ambiguities in homology determination.
Based on the distribution of the variable nucleotide positions and differential nucleotide frequencies in different parts of the control region, it is divided into three domains (Brown et al. 1986). Domains I and III are rich in L-strand adenine, whereas the central domain II is low in adenine. Most of the variability, both nucleotide substitutions and deletions/insertions, is concentrated in domains I and III whereas domain II is more conservative. Based on sequence similarity, tens of conserved sequence blocks with putative functional importance have been described (e.g. Southern et al. 1988, Lee et al. 1995, Sbisà et al. 1997, Randi & Lucchini 1998). The general structure of the control region and an overview of the sequence blocks are depicted in Fig. 1.
Figure 1. General structure of the vertebrate mitochondrial control region. The arrows indicate the location of the H-strand replication origin and the bidirectional promoter for L- and H-strand transcription. TAS, termination associated sequence; F through B, conserved sequence boxes in the central domain; CSBs, conserved sequence blocks.