Neandertal DNA and Modern Humans

The variation of mitochondrial DNA (mtDNA) between modern humans and Neandertal sequences lie outside the mtDNA sequence variation within modern humans. This variation has led several researchers to conclude that Neandertals did not contribute to modern human DNA and are a separate species that went extinct in Europe. It is feasible that DNA can be retrieved from specimens that died thousands of years ago, given the ideal preservation conditions and extraction protocols. However, DNA also decays as the organism decomposes. Spontaneous hydrolysis, oxidation, and nucleotide modifications are a few of the processes that cause DNA decay and likely interfere with reliably obtaining a mtDNA sequence that accurately reflects the Neandertal mtDNA sequence. In addition to DNA decay, contamination of samples is also apparent in published Neandertal mtDNA sequences. A comparison of conserved sequence block 2 (CSB2) in hypervariable region II (HVRII) between Neandertal mtDNA and modern man, primates, and other mammals indicate that excess thymine in CSB2 of published Neandertal mtDNA is likely the result of contamination. * Daniel Criswell, Ph.D., Institute for Creation Research, 1806 Royal Lane, Dallas, TX, Phone 214-615-8300, [email protected] Accepted for publication October 22, 2008 DNA (aDNA) from Neandertal remains. Eleven Neandertal mitochondrial DNA (mtDNA) sequences have been published (Krings et al., 1997; Krings et al., 1999; Krings et al., 2000; Ovchinnikov et al., 2000; Schmitz et al., 2002; LaluezaFox et al., 2006; Caramelli et al., 2006; Orlando et al., 2006; Krause et al., 2007), and recently over one million base pairs of Neandertal nuclear DNA (nuDNA) were sequenced (Green et al., 2006; Noonan et al., 2006). These sequences claim to be evidence that Neandertals are distinctly different from modern humans and likely did not contribute to the modern human genome (Krings et al., 1997; Serre et al., 2004; Currat and Excoffier, 2004). The average sequence difference between some of the published Neandertal mtDNA sequences and modern humans is about three times the number of average variation between modern humans. In fact, the putative mtDNA Volume 45, Spring 2009 247 variation between modern humans and Neandertals lies completely outside the range of variation within modern humans (Krings et al., 1997). If currently obtained Neandertal mtDNA sequences accurately represent their original sequence, it would provide strong evidence that the two groups did not exchange DNA and they could be classified as different species or subspecies (Homo sapiens and Homo sapiens neanderthalensis). The relatedness between modern humans and Neandertals using mtDNA (or nuclear DNA) is dependent, however, on retrieving uncontaminated DNA from Neandertal bones and teeth that has been sufficiently preserved to prevent significant DNA decay. Mitochondrial DNA and Contamination mtDNA provides the best opportunity for acquiring templates to sequence ancient DNA. It is a 16,569 base pair, circular DNA strand divided into two functional regions—the control region (1120 base pairs, 16,024–16569 and 0–576) and the coding region (the remaining 15,450 base pairs), where proteins, transfer RNA, and ribosomal RNA are transcribed (Cutticchia, 1995). The control region is where transcription is regulated and is taxonomically significant, meaning there is enough variation in this region to distinguish sequences between species, members of different ethnic groups within a species, or even one human family from another. The two hypervariable regions, hypervariable region I, or HVRI (16,024–16,383), and hypervariable region II, or HVRII (57–372), in the control region are the sources of Neandertal mtDNA. Nine Neandertal mtDNA sequences are from HVRI and two sequences were obtained from HVRII. An important factor in obtaining useful ancient DNA from mitochondria is the number of mtDNA molecules that can be extracted from a sample. Each sample from a living organism should have thousands of cells, 100-1000 mitochondria in each cell, and 5–10 copies of DNA in each mitochondrion, providing the >1000 mtDNA molecules required for efficient amplification (Krings et al., 1997). However, bone or tooth samples that provide mtDNA for sequencing will have far fewer intact cells, if any, and the DNA that is retrieved is likely bound to the hydroxyapatite of the bone or tooth. The loss of cells and DNA post-mortem damage emphasize the value of mtDNA with thousands of copies for sequencing targets in a living organism, and many successful aDNA sequences have started with under 100 templates (Hofreiter et al., 2001). Each copy of mtDNA provides a target for primers (short oligonucleotides) that initiate the polymerase chain reaction (PCR) in amplifying enough mtDNA for sequencing. It is assumed that enough undamaged Neandertal mtDNA molecules are still present to provide a template for the PCR. Theoretically, several PCR amplifications with extracted template mtDNA provide sufficient sequences to derive a consensus sequence and accurately represent the original Neandertal mtDNA sequence. The more template mtDNA that is available at the start of the amplification process, the less likely the PCR product will be contaminated by extraneous sources of DNA. Contaminating DNA can block the PCR amplification process, incorporate foreign DNA bases into targeted DNA, or result in the amplification of only nontargeted DNA (Pusch and Bachmann, 2004). Contamination is a constant obstacle for obtaining authentic ancient DNA sequences, and the evidence for contamination has been observed in ancient mtDNA samples as well as ancient nuclear DNA (Wall and Kim, 2007). Valid research requires adequate experimental protocols to prevent the incorporation of contaminated DNA (Cooper and Poinar, 2000; Gilbert et al., 2005; Willerslev and Cooper, 2005). These requirements include a facility reserved only for ancient DNA research, the independent verification of the putative sequence in another laboratory, treatment with enzymes to remove damaged bases, and a sufficient number of clones to derive a consensus sequence (Cooper and Poinar, 2000; Gilbert et al., 2005; Bower et al., 2005; Willerslev and Cooper, 2005). DNA Decay Precautions can be taken to minimize contamination in aDNA samples, but the problem of DNA decay poses more difficult problems to obtain authentic ancient mtDNA sequences. As soon as an organism dies, DNA begins to degrade, and the repair mechanisms that maintain DNA sequence fidelity in living systems no longer function. Spontaneous hydrolysis and oxidation result in double-strand breaks, abasic sites, and nucleotide modifications or miscoding lesions (Lindahl, 1993). Double-strand breaks and abasic sites (most commonly depurination) can prevent PCR amplification, while nucleotide modifications can be incorporated into the amplified PCR product, mimicking the expected evolutionary changes in the putative DNA sequence. Spontaneous hydrolysis results in approximately 2,000–10,000 depurination events in each human cell per day. This is due to the instability of the N-glycosyl bond between a purine (adenine and guanine) and the 2' carbon of the deoxyribose sugar (Lindahl, 1993). In living systems, endonucleases rapidly initiate repair processes to maintain the integrity of DNA (Lindahl, 1993). However, after the organism dies, depurination results in small fragments of DNA, which are more difficult to amplify than intact DNA from a living organism. Kinetic calculations predict that amplifiable fragments less than 400 base pairs will survive no longer than 10,000 years 248 Creation Research Society Quarterly at temperate conditions (Poinar et al., 1996). As the temperature decreases, the rate of DNA decay also decreases, making samples in permafrost the best candidates for aDNA sequencing (Smith et al., 2001; Willerslev and Cooper, 2005). All of the environmental factors that affect DNA decay are not completely understood, but higher temperatures and increased moisture appear to be two of the more significant factors that accelerate DNA decay. Whether or not aDNA sequences are damaged or accurately represent the original sequence of the living organism, the retrieval of any DNA from ancient organisms is evidence for a recent existence of less than 1 million years for permafrost specimens (mammoths and bacteria) and less than 10,000 years for temperate specimens, including Neandertals (Poinar et al., 1996; Smith et al., 2001; Willerslev and Cooper, 2005). The most common result of decay in DNA is the deamination of cytosine resulting in a base change to uracil (Hansen et al., 2001; Hofreiter et al., 2001). This change is easily defined by the chemical nomenclature of the two bases. Cytosine is 2–oxy–4–amino pyrimidine and uracil is 2–oxy–4–oxy pyrimidine. Post–mortem damage from cytosine deamination can accumulate fairly quickly in the context of aDNA, considering that the half–life of a cytosine residue is about 200 years under human physiological conditions (37°C and 7.4 pH) (Lindahl and Nyberg, 1974). Cytosine deamination occurs an estimated 100–500 times a day in a living cell (Lindahl, 1993), where accompanying repair mechanisms are able to correct the damage. Uracil is a base not normally incorporated into the DNA sequence of any organism; consequently many organisms, including humans, have an enzyme, uracil–N–glycosylase (UNG), to remove deaminated cytosine (uracil) when it is incorporated into DNA. When an organism dies, spontaneous cytosine deamination can occur through a hydrolysis reaction that removes the amine group converting cytosine to uracil (Lindahl, 1993). Without UNG to repair the damaged base, any postmortem sequencing reaction will identify the deaminated cytosine as uracil and pair it with adenine on the complementary strand. In mtDNA, if the deaminated cytosine is on the heavy strand (H strand), an adenine will be incorporated on the complementary light strand (L strand) in place of the original guanine. Deamination of adenine to hypoxanthine and guanine to xanthine also occurs, but at less than 2–3% of the rate of cytosine deamination (Lindahl, 1993), making them less likely to cause major changes in aDNA sequences. Cytosine to thymine (C→T) and guanine to adenine (G→A) transitions are classified as Type 2 transitions in the context of aDNA sequencing, representing the deamination of cytosine to uracil (Hansen et al., 2001). Type 1 transitions represent the possible deamination of adenine→ hypoxanthine, which results in an A→G and T→C transition on complementary DNA strands (Table I). Obviously, scientists who sequence aDNA are aware of the problems that deamination can cause; one of the required protocols is to treat the samples with UNG to remove deaminated cytosine residues that could produce erroneous sequences from PCR amplification. UNG is believed to remove all deaminated cytosines eliminating Type 2 transitions from DNA (Hofreiter et al., 2001). However, treatment with UNG before amplification is not without problems. The removal of deaminated cytosine leaves an abasic site, creating strand nicks that can prevent amplification of the aDNA strand (Hofreiter et al., 2001). This is particularly worrisome when there might only be 100 templates available for amplification. Reducing the number of templates increases the risk of incorporating contaminating extraneous DNA into the targeted sequence, or completely sequencing contaminants (Pusch and Bachmann, 2004). Evidence exists that cytosine deamination is not the only source of Type 2 transitions in aDNA. Gilbert et al. (2007) and Hoss et al. (1996) both found that UNG treatment left behind half of the C→T Type 2 transitions that were identified from damaged sequences in controlled experiments. Presumably, the C→T transition resulted from the decay of guanine to adenine on the complementary strand. Gilbert et al. (2003; 2007) noticed this, at first speculating that UNG did not remove all deaminated cytosines and was successful mainly on longer templates. Apparently, these C→T transitions might occur from an (as yet) unidentified degradation of guanine to adenine and be as frequent as cytosine deamination in nonliving samples (Gilbert et al., 2007). The C→T transition would result from sequencing the complementary strand opposite the G→A deamination event. This observation has prompted Gilbert et al. (2007) to observe that most of the knowledge Table I. Type 1 and Type 2 Transitions. DNA decay will result in transitions on both strands when sequenced. Cytosine deamination is the most common form DNA damage in living systems. Transition Type H Strand L Strand Composite Change Type 1 T→C A→G TA→CG Type 2 C→T G→A CG→TA Volume 45, Spring 2009 249 about DNA damage comes from living systems and not from aDNA, leaving the possibility that post-mortem DNA is damaged from processes that are not yet understood (Gilbert et al., 2007). One possible explanation for postmortem conversion of G→A is the co-extraction of divalent metal ions that have been shown to damage DNA (Pusch and Bachmann, 2004). Concentrations of some divalent metal ions increase several times (up to 5,000 times) in aDNA when compared to contemporary samples, raising the possibility of mutagenic effects when aDNA is extracted and amplified through PCR (VelascoVazquez et al., 1997). This is particularly true of manganese. The G→A transition frequently occurs at site 1138 of the human FGFR3 gene in the presence of 0.25mM MnCl2. The same mutation occurs when human template DNA is spiked with aDNA, indicating that aDNA is the source for the mutagenic manganese (Pusch et al., 2004). It also explains why independent sequencing of putative aDNA samples would incorporate the same Type 2 transitions; the source of the mutagenesis (manganese) is in the bone or tooth material supplying the aDNA. The possibility that mutagenic effects from divalent metal ions cause sequence divergence between Neandertals and modern humans casts serious doubt on aDNA sequences being validated by independent laboratories.