DNA extraction and amplification
A brief perusal through the Material and Methods sections of manuscripts on aDNA reveals that many different procedures are aimed at recovering, processing, and amplifying heavily damaged and extremely rare molecules (O' Rourke et al., 2000; Pusch et al., 2003).
Since none of these procedures can be considered a priori the most effective, the extraction protocol generally needs to be modified and adapted to the peculiar characteristics of the archeological site where the remains are found. For instance, in the case of the Taq polymerase inhibitors present in the burial soil, the extraction protocol must remove these substances through extensive phenol-chloroform extraction or through adsorption of DNA molecules on silica particles. Recent reviews give an exhaustive picture of the most important procedures and strategies. However a recent report worth citing (Yang et al., 2003) suggests an interesting experimental behavior in establishing PCR amplification protocols and in the interpretation of results. In particular, the authors focused their attention on strategies aimed at minimizing contamination, adjusting the numbers of cycles, and varying the type of Taq DNA polymerase.
Animal aDNA: archeological site authentication and phylogeny
It is worth noting that once an archeological site becomes the object of any aDNA studies, the analysis of aDNA extracted from animal bone remains found there is extremely important for a number of reasons. First of all, data on animal remains allow the authentication of the site itself in terms of suitable conditions for nucleic acid preservation. In fact, modern DNA contamination, one of the most serious problems in aDNA studies, becomes meaningless when analyzing these kinds of remains, since there is no possibility that they could derive from present-day animals, especially in those laboratories where no modern animal DNA is handled. For instance, studies on a young Barbary macaque (Bailey et al., 1999) kept in Pompeii at the “Terme del Sarno” collection, as well as on the five equids (Di Bernardo et al., 2002) found in the “Casti Amanti” house showed the presence of amplifiable aDNA. This finding supported the hypothesis that burial conditions in Pompeii were favorable for aDNA preservation already proposed for human remains (Cipollaro et al., 1998, 1999).
The low probability of contamination when animal remains are investigated also allows an in-depth study of basic problems strictly related to aDNA characteristics, such as enzymatic repair of damaged molecules. Several research groups have reported data from animal remains to establish the relationship between species or to test the hypotheses on the status of a given taxa. A recent report by Huynen et al. (2003) clarifies the number of Moa species, extinct ratite birds that showed extreme variation in size. This work illustrates the role that nuclear DNA sequences can play in testing previously intractable hypotheses about extinct organisms. Edwards et al. (2003) explored microsatellite markers in ancient cattle bones from a Viking age settlement in Dublin and suggested an Irish origin for these medieval cattles. Vila et al. (2001) favored a widespread origin of domestic horse lineage analyzing mtDNA control region of 191 domestic horses, while Lambert et al. (2002) showed the rate of evolution of hypervariable region I of mtDNA in well-preserved subfossil bones of Adelie penguins dating back 7,000 years BP. Further studies suggested a transition from wild to domestic status in Neolithic goats (Bar-Gal et al., 2002), while Leonard et al. (2002) demonstrated that native American dogs originated from multiple Old World lineages. Poinar et al. (2003) provided information about phylogeny of the Pleistocene ground sloth and the three-toed sloth.
mtDNA and nuclear genes
aDNA investigations are largely based on the amplification of mtDNA DNA. This small genome is more prone to give positive amplification, since it has a cell copy number higher than nuclear genes, having only two copies per cell. mtDNA is particularly useful in studying human evolution: first, it is maternally inherited and thus does not undergo recombination; second, the mtDNA evolution rate is much higher than that of nuclear genes. As a consequence, a substantial number of mtDNA mutations have accumulated sequentially along radiating maternal lineages that have diverged as human populations colonized different geographical regions of the world. The sequencing of the mtDNA control region hypervariable segments I and II, together with restriction fragment length polymorphism (RFLP) investigation, have revealed a number of stable polymorphic sites that define related groups of mtDNAs, called haplogroups. Most of the mutations observed both in mtDNA coding and control regions in modern human populations have occurred on these pre-existing haplogroups and they define the individual mtDNA types or haplotypes. The analysis of mtDNA in extant individuals allowed us to understand the phylogeny of Homo sapiens sapiens (Stringer and Andrews, 1988; Pusch et al., 2003) or to infer routes and times for human expansion out of Africa (Dayton, 2003).
In particular, Ingman and Gyllensten (2003) presented a report on the mitochondrial genome variation and evolutionary history of Australian and New Guinea Aborigines. They showed that the genetic diversity of the Australian mitochondrial sequences is remarkably high and is similar to that found across Asia, in contrast to the pattern seen in previously described Y-chromosome data, where a specific Australian haplotype was found at high frequency (Dayton, 2003). The predominance of an unique Y-chromosome haplotype, contrasting with the high mitochondrial diversity would be, according to the Ingman and Gyllenstein hypothesis, the result of a founder effect, since the population expansion started from a few hundred individuals (Kayser et al., 2001). If this is the case, the authors interpreted their data as the results of an inappropriate sampling across subpopulations, rather than within a single tribe. They concluded that additional studies of autosomal loci are necessary to obtain a balanced view of the evolutionary history of the peoples in this region.
Many reports based on human aDNA are focused also on mtDNA, such as those analyzing Neanderthal mtDNA sequences (Cooper et al., 1997; Krings et al., 1997, 1999, 2000; Ovchinnikov et al., 2000; Gutierrez et al., 2002). The report of Schmitz et al. (2002) described a further study on Neanderthal remains, whose sequences cluster with other Neanderthal sequences already published. The author also sustained the importance of an integrated biological approach for a complete and reliable perspective on fossil material with the collaboration of geneticists, morphologists, archeologists, and dating specialists.
Aside from mtDNA, the analysis of nuclear DNA and in particular of autosomal microsatellites (STR) are useful to study the genetic structure of human populations. A recent report (Rosenberg et al., 2002), exploring 377 autosomal microsatellite loci in 1056 extant individuals from 52 populations identified six main genetic clusters, five of which correspond to major geographic regions, and subclusters that often correspond to individual populations.
STR are also explored to identify historical families as reported in the Romanov family investigation, one of the first attempts at molecular identification of skeletal remains (Gill et al., 1994). Recently, however, data related to the Romanov family reconstruction have raised doubts about their molecular and forensic consistency. In fact, Knight et al. (2004) claimed the impossibility to amplify a 1,223-bp fragment in aDNA as Gill et al. had reported in their work. Knight et al. found that the mtDNA consensus haplotype of Elisabeth, sister of Empress Alexandra, differs from that reported for her sister Alexandra at four sites thus sustaining that samples analyzed in Gill study had been contaminated with non-degraded high molecular weight “fresh” DNA.
A more recent report by Keyser-Tracqui et al. (2003) is focused on the analysis of biparental, paternal, and maternal genetic systems to reconstruct genealogies in a protohistoric necropolis of Mongolia.
A particularly effective way to screen samples is to carry out multiplex analyses of autosomal STRs, Y-chromosomal STR markers, and sex typing marker like amelogenin locus (Lassen et al., 1996; Faerman et al., 1997; Cipollaro et al., 1998).
Recent reports have illustrated that also single-locus nuclear DNA sequences can be consistently recovered from ancient material (Huynen et al., 2003). This possibility is of great interest in aDNA investigation since it paves the way to analyze genes involved in genetic diseases. Filon et al. (1995) detected a beta-thalassemia mutation in the aDNA of skeletal remains from the archeological site of Akhziv, Israel. More recently, Rabino et al. (2002) detected a sickle cell anemia in three Egyptian mummies stored at the Anthropological and Ethnographic Museum of Turin, Italy. However, this latter finding is in contrast with data reported by Marota et al. (2002) in their study on Egyptian papyri varying in age from 1,300 to 3,200 years BP. They showed the complete loss of authentic aDNA also in more recent papyri dating from the 8th century AD. This is in agreement with the level of amino acid racemization found to be higher than 0.08, commonly indicated as the highest value compatible with the presence of aDNA. The authors, claiming that the amino acid racemization level in Egyptian bones and tissues is consistent with that found in papyri, have also raised doubts about the Egyptian aDNA data published so far.
Post-mortem aDNA damages and base modifications
aDNA is always represented by heavily damaged and highly fragmented molecules (Lindahl, 1993). Damage includes either base modification (oxidation and deamination) or base loss (apurinic and/or apyrimidinic sites) as well as single-strand DNA breaks, with either nicks, gaps, or protruding ends. Interstrand crosslinks have also been shown to exist at relatively high frequencies (Paabo, 1989). Such damages interfere to different extents with common PCR procedures used in aDNA studies.
Among base modifications, accumulating when an organism dies (Hansen et al., 2001; Hofreiter et al., 2001a; Gilbert et al., 2003a,b), the most often represented is the cytosine deamination, a miscoding modification that causes a G/C A/T substitution, producing artifacts during PCR procedure. Aside from misincorporations due to the damages present in the aDNA molecules, there is also the possibility of misincorporation by Taq polymerase itself.
To exclude the presence of post-mortem base modifications, some authors have sequenced a DNA fragment of the best preserved region in the genome, that is, 16S rDNA (Di Bernardo et al., 2002; Lambert et al., 2002). The rationale was that, since this DNA region exhibits a low rate of mutations, the finding of base substitutions at an increased rate could indicate either post-mortem base modifications or Taq DNA polymerase mistakes (Poinar et al., 1998). The absence of polymorphisms in this genome region infer either the absence of nucleotide base modifications also in other genome regions or that these events are negligible.
Another strategy is to set up an asymmetric PCR (Hofreiter et al., 2001). This strategy allows each strand of the target to be preferentially amplified using an unbalanced number of primers. Different nucleotides found at the same position when sequencing the product of the two unbalanced amplifications will confirm this hypothesis.
Pre-amplification and aDNA enzymatic repair
Different strategies have been set up to enhance the rate of successful amplifications or to repair aDNA molecules. Pre-amplification procedures, such as Degenerated Oligonucleotide PCR (DOP-PCR) (Pusch et al., 2000) and Primed Extension PCR (PEP-PCR) (Satoh et al., 1998), have been proposed. The authors suggest performing a preliminary amplification on aDNA extracts using degenerated primers, thus increasing the number of molecules available for a specific PCR. However, no significant report has been produced applying these techniques.
Some authors have proposed enzymatically repairing aDNA molecules in the attempt to increase the number of mtDNA molecules to be amplified (Pusch et al., 1998) or to also rescue nuclear gene sequences other than mtDNA molecules (Di Bernardo et al., 2002). aDNA enzymatic repair can be performed successfully only for molecules presenting unmodified 3′OH and/or 5′P termini. It can be terminally elongated by DNA polymerase I, sealed by T4 DNA ligase, or filled in and then sealed by the concerted action of these two enzymes.