Archaeological Chemistry - ACS Publications - American Chemical


Archaeological Chemistry - ACS Publications - American Chemical...

0 downloads 81 Views 2MB Size

Chapter 4

Downloaded by UNIV OF MICHIGAN ANN ARBOR on February 18, 2015 | http://pubs.acs.org Publication Date: August 16, 2007 | doi: 10.1021/bk-2007-0968.ch004

Extraction and Analysis of DNA from Archaeological Specimens 1

2

1

Brian M. Kemp , Cara Monroe , and David Glenn Smith 1

Department of Anthropology, University of California, Davis, CA 95616 Department of Anthropology, University of California, Santa Barbara, CA 93106 2

The study of D N A extracted from archaeological specimens is an exciting new avenue of research that can provide unique evidence for addressing archaeological questions. Here we give an overview of notable case studies, some of which were performed by the authors, that used genetic data retrieved from archaeological specimens to make interpretations about the past. Additionally we describe how ancient D N A (aDNA) differs from modern D N A and, thus, why only a few specific genomic markers are usually targeted and why protocols have been developed explicitly for the study of aDNA. We detail an aDNA extraction protocol that we have been developing over the past few years, and a description of genetic screening we routinely perform on D N A extracted from ancient human samples.

Introduction D N A extracted from archaeological specimens, ancient D N A (aDNA), provides the researcher a window into prehistory, allowing one to say something absolute about the genetic characteristics of someone or something that lived in the past. While modern D N A studies can make predictions about the past based

78

© 2007 American Chemical Society

In Archaeological Chemistry; Glascock, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Downloaded by UNIV OF MICHIGAN ANN ARBOR on February 18, 2015 | http://pubs.acs.org Publication Date: August 16, 2007 | doi: 10.1021/bk-2007-0968.ch004

79 on genetic theory, only aDNA studies can directly test these hypotheses by examining genetic variation at specific temporal and geographic axes. Through the analysis of genetic patterns found in archaeological assemblages, these data can be used, for example, to study human migration, to address the relationships between prehistoric and contemporary populations, to study kinship patterns revealed in burial contexts, to reconstruct prehistoric diets, and to study human behavior. Furthermore, the study of aDNA allows for the observation of molecular evolution over time and, from this, its rate can be estimated. However, these potential benefits must be weighed against the methodological challenges presented by the study of aDNA. The study of aDNA is more difficult than the study of modern D N A primarily because aDNA tends to be degraded, chemically modified, and in low copy number (1-3). As such, particular care must be taken in the study of aDNA and a high level of expertise is required to ensure authentic results. Below, we review some studies that have addressed anthropological questions with aDNA data retrieved from a wide variety of archaeological specimens that span both time and geography. We then proceed to outline a protocol that has been developed over the past few years in David Glenn Smith's Molecular Anthropology Laboratory at the University of California, Davis. In particular, this protocol addresses two of the major complications associated with the study of aDNA: modern D N A contamination and polymerase chain reaction (PCR) inhibition.

Examples of Ancient DNA Applications to Topics in Archaeology The following eight examples are meant to provide a representative survey of the wide range of topics that have been addressed with aDNA evidence. The first three examples are ones that followed the methodology described later in this paper. The additional five examples are ones that we find to be particularly interesting and, in our opinion, were well executed.

Molecular Sex Determination and Human Behavior Recently, de la Cruz and colleagues (4) used aDNA techniques to determine the sex of sacrificial victims that were buried under Temple R, dedicated to the Aztec god of wind and rain, Ehecatl-Quetzalcoatl, at the archaeological site of Tlatelolco. It is believed that these individuals were sacrificed and buried in a single ceremony that signified the founding of the temple. In this case, traditional morphometric techniques were insufficient for sex determination, as most of the individuals were infants, sub-adults, and/or were represented by fragmentary

In Archaeological Chemistry; Glascock, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

80 remains. The results of the molecular analyses suggest that most, i f not all, of these sacrificial victims were males and provide valuable insight into Aztec ritualistic behaviour.

Downloaded by UNIV OF MICHIGAN ANN ARBOR on February 18, 2015 | http://pubs.acs.org Publication Date: August 16, 2007 | doi: 10.1021/bk-2007-0968.ch004

Coprolites and Population Movement Kemp and colleagues (5) successfully analyzed human mtDNA extracted from eight samples of 700-2,000 year old human coprolites excavated from Fish Slough Cave located in northern Owen's Valley, California. It was discovered that individuals belonging to a rare, derived branch of mitochondrial haplogroup C made five of the fecal samples. A survey of mtDNA variation of Native Americans revealed that the specific haplotype represented by these coprolites is only found today among the Northern Paiute and Washo, who reside in the Great Basin (6). This evidence provides a genetic connection between the Northern Paiute and prehistoric Owen's Valley populations and is significant as the Numic sub-branch (including Northern Paiute) of the Uto-Aztecan language family was thought to have spread into the Great Basin (7, 8) from Owen's Valley, around the time that the inhabitants of Fish Slough Cave produced the ancient coprolites.

Detection of an Ancient Coastal Migration by Native Americans Recently, Kemp and colleagues (9) analyzed mtDNA from 10,300 year old skeletal remains that were excavated from On Your Knees Cave on Prince of Wales Island, Alaska. The mtDNA of this individual was determined to belong to a newly recognized founding maternal lineage that was carried to the Americas from Asia. It was discovered that a mere 1.4% of Native Americans descend from this founding matriline and that these individuals today live primarily along the western edge of North and South America, being found in, for example the Chumash in Southern California, the Cayapa of Ecuador, and populations of Tierra del Fuego. The evidence is compatible with an ancient, rapid migration of humans along the west coast of the Americas.

Biological Relationships Determined from Burial Context Keyser-Tracqui and colleagues (JO) successfully extracted aDNA from 62 individuals from the Eygin Gol Necropolis located in Northern Mongolia, dating to the 3 century B . C . to the 2 century A . D . Variation detected in the mitochondrial, autosomal, and Y-chromosomal genomes allowed the researchers rd

nd

In Archaeological Chemistry; Glascock, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

81

Downloaded by UNIV OF MICHIGAN ANN ARBOR on February 18, 2015 | http://pubs.acs.org Publication Date: August 16, 2007 | doi: 10.1021/bk-2007-0968.ch004

to correlate biological relationships of individuals buried in the cemetery to both space and time. Some sections of the cemetery were comprised of related males, possibly indicating an elite patriline that continued over multiple generations. Parent-child relationships were also determined, in some cases indicating that particular portions of the cemetery represent familial groups. Later burials demonstrated a shift in the genetic signature of the population, which is thought to be the evidence of an entrance of Turks in the area.

The Study of Animal Population Demography Over Time Recently, Shapiro and colleagues (//) used Bayesian statistical techniques to model Asian and American bison demographics over a 60,000 year period using aDNA data retrieved from 442 bison remains. Their results support the hypothesis that bison populations began a dramatic decline approximately 37,000 years ago, likely as a result of environment change. Shapiro and colleagues suggest that, if other species were similarly affected, human hunting may not have been the major cause of the extinction of American megafauna.

Dietary Reconstruction Poinar and colleagues (12) successfully extracted D N A from three human coprolites made by Native Americans in Hinds Cave, Texas more than 2,000 years ago. Not only were they able to extract human mtDNA from the coprolites, which were determined to belong to known Native American mitochondrial types, but they also were able to analyze D N A from the plants and animals eaten by these occupants of Hinds Cave. Importantly, they discovered that some of the species of plants and animals identified by molecular techniques were not detected in the same feces by macroscopic techniques (and some macroscopically identified species were not detected by the D N A analysis). The combination of the molecular and macroscopic evidence demonstrated that these three Native Americans had a diverse and well-balanced diet.

Study of Economic Stratification of Prehistoric Populations Using aDNA techniques, Speller and colleagues (13) were able to identify the species of-1,200 year old salmon remains excavated from various houses and structures at the archaeological site of Keatley Creek, British Columbia. Prior study of the site provided evidence of a socio-economic distinction among the residents of these dwellings based on the distribution of prestige goods and

In Archaeological Chemistry; Glascock, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

82

Downloaded by UNIV OF MICHIGAN ANN ARBOR on February 18, 2015 | http://pubs.acs.org Publication Date: August 16, 2007 | doi: 10.1021/bk-2007-0968.ch004

the density and variety of cultural and faunal remains excavated from the features. The researchers sought to test whether this socio-economic distinction resulted in differential access to preferred salmon species. Their results suggest that economic stratification had less effect on the distribution of preferred salmon species within the structures than previously thought. Further, they found no evidence of pink salmon (Oncorhynchus gorbuscha) at the site, despite the assumption that this species was a staple in the area.

Study of Disease Ancient D N A studies have also focused on determining the presence or absence of disease among human, as well as animal populations. Evidence from these studies has contributed significantly to osteological studies of signs of infection with ambiguous etiological origins. The study of ancient disease can also trace the evolutionary trajectory of pathogens and their dispersal through time. For example, Salo et al. (1994) identified tuberculosis (Mycobacterium tuberculosis) from a 1,000-year-old Peruvian mummy from the site of Chiribaya Alta, effectively dispelling the notion that tuberculosis was first introduced at the time of European contact. Additional studies of disease include: plagues (14, 15), leprosy (16), syphlis (17, 18), malaria, and gastrointestinal pathogens (19).

Problems and Properties of aDNA Sample Selection While the first successful study of aDNA was from the hide of an extinct quagga (20), the vast majority of aDNA studies since have been performed on bones and teeth, with fewer studies performed on preserved or mummified tissue (for an excellent review see 21). It is also possible to extract aDNA from coprolites (5, 22, 23), quids (chewed wads of plant fiber), and "aprons" (women's breechcloths) (24). Similarly, D N A has also been extracted from ancient pathogens, faunal remains, and botanicals. The choice of samples will obviously depend on the question that is being posed. For example, plant and animal D N A extracted from human coprolites (e.g., 22) or D N A extracted from the remains of animals found in house structures (e.g., 13) would be useful in the study of prehistoric dietary practices. However, sample choice also depends on the availability of material from times and places of interest. In this sense, where human skeletal samples are absent

In Archaeological Chemistry; Glascock, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

83 from the archaeological record, are in a poor state of preservation, and/or are simply restricted from study , coprolite or quid samples may represent the only material through which ancient population genetics can be studied (24). Some aspects of human behavior and prehistory may be inferred from the genetic patterns revealed in domesticates and, therefore, samples of plants and animals may suffice. Further consideration of sample selection is based on correlations between the type and physical state of the archaeological specimens and the probability of successfully extracting D N A from them. While much of this has been wellcovered by review papers (25-27), a few points are worth re-emphasizing here, with additional reference. In general, well-preserved remains are more likely to contain well-preserved D N A ; sampling material that exhibits signs of degradation due to microbial activity, burning, highly acidic or basic environments, cracked bones or teeth, or teeth with caries should be avoided, i f possible. Samples that cannot be radiocarbon dated are also poor candidates for D N A extraction, as D N A degrades in a similar fashion to other biomolecules, and a positive correlation between the preservation of collagen and D N A has been demonstrated in ancient remains (28). While, the age of the sample is negatively correlated with D N A preservation, with younger samples having a higher likelihood of containing preserved D N A , environmental conditions probably have a greater effect on D N A preservation than does time alone. That is, older samples from cold, dry climates might contain better preserved D N A than younger samples excavated from hot and/or wet environments. Additionally, D N A has been shown to preserve better in teeth than in bone (9, 29), and cortical bone (compact) is preferable over cancellous (spongy) bone for D N A extraction. Lastly, as for yet unknown reasons, D N A has been shown to preserve remarkably well in coprolites (5, 22, 23).

Downloaded by UNIV OF MICHIGAN ANN ARBOR on February 18, 2015 | http://pubs.acs.org Publication Date: August 16, 2007 | doi: 10.1021/bk-2007-0968.ch004

1

Molecular Markers Due to the high copy number of the mitochondrial genome per cell (> 1,000), in comparison to the nuclear genome (two copies or each marker per cell), this system has been the focus of most aDNA studies. In mammals, this genome is passed solely through the female line and, therefore, is only informative about female prehistory. The male analogue of mitochondrial D N A (mtDNA) is the Y-chromosome; markers on this chromosome and others from the nuclear genome are much more difficult to analyze from aDNA extracts due to low copy number. As such, molecular sex identification, addressing paternal relationships, and/or addressing male prehistory is generally not often as feasible, but may be worth considering depending on the archaeological context and the level of D N A preservation in the sample.

In Archaeological Chemistry; Glascock, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

84

Downloaded by UNIV OF MICHIGAN ANN ARBOR on February 18, 2015 | http://pubs.acs.org Publication Date: August 16, 2007 | doi: 10.1021/bk-2007-0968.ch004

Contamination Control As D N A extracted from ancient remains tends to be in low copy number and is generally degraded into fragments less than 200 base pairs (bp) in length (/, 2), aDNA extractions are highly susceptible to contamination originating from modern sources. This is particularly a problem when one attempts the study of ancient D N A from human sources. Modern contaminating D N A can be in higher copy number and more fully intact than the endogenous aDNA and, thus, can better compete than aDNA as a target for amplification during the polymerase chain reaction (PCR) (for those unfamiliar with the process, it is explained very well by 30). In fact, contamination can completely out-compete aDNA in P C R leading to false positives and aberrant results (31). The detection of contamination is dependent on the specificity of the PCR. For example, human studies are particularly prone to contamination because the D N A of any human that has come into contact with the remains can potentially be amplified. This problem is not, however, peculiar to human studies. Bacteria from the soil that are closely related to targeted D N A of pathogen can also contaminate samples studied for the presence or absence of pathogens. Contamination is less of a problem, for instance, i f salmon D N A is targeted while the sample is contaminated with human D N A because a properly designed P C R that targets salmon should not be able to amplify human D N A . In spite of this, we cannot stress enough that contamination is a serious concern and all samples, regardless of type, should be treated with the same level of care to minimize contamination. Ancient D N A extractions can become contaminated via two sources: surface contamination of the archaeological specimens through bare-hands handling of the material or later during D N A extraction and analysis in the D N A laboratory. The former source of contamination can originate at any step of an aDNA study from the time of excavation to the time of D N A extraction. Modern contamination can originate with anyone who has had direct contact with the material, including the archaeologists that excavated the specimens and/or researchers that analyzed (e.g., cataloging, measuring) the specimens. This type of contamination can be minimized by following the advice of Yang and Watt (32) for preparing archaeological specimens for D N A analysis. The surfaces of archaeological specimens are especially vulnerable to contamination, and it is crucial that any contamination be removed before D N A extraction begins. Contaminating D N A can be removed from the surfaces of bones and teeth with a treatment of high concentration bleach (NaOCl, sodium hypochlorite) because aDNA is more resistant to the oxidant, than is contaminating D N A (31, 33, 34). This is believed to result from the fact that ancient endogenous D N A is protected within crystal aggregates of the bone (34), an observation that is consistent with the idea that D N A that is bound to hydroxyapatite aids not only in its long-term preservation, but also protects it from oxidation by highly concentrated bleach

In Archaeological Chemistry; Glascock, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Downloaded by UNIV OF MICHIGAN ANN ARBOR on February 18, 2015 | http://pubs.acs.org Publication Date: August 16, 2007 | doi: 10.1021/bk-2007-0968.ch004

85 (37). A method for the decontamination of non-skeletal remains (e.g. coprolites and quids) has not been developed. While many researchers use ultraviolet (UV) irradiation to "destroy" contamination (see studies reviewed in 37, 33), no study has demonstrated that this technique is actually effective in removing contamination on archaeological specimens (33). One way to avoid potential contamination when extracting D N A from coprolites and quids is to take samples from the interior portions of the specimens. The latter source of contamination can originate from reagents, labware, P C R carryover, and/or D N A lab personnel. As such, procedures that reduce contamination should be implemented, including using of D N A free lab-ware and reagents, processing ancient materials in a laboratory separated from the one in which modern D N A is examined, and using negative controls in both D N A extraction and amplification to monitor contamination, i f present (following the advice of 35). Contamination can also be removed from P C R cocktails, prior to amplification, with DNAase (36).

PCR Inhibitors Another major problem associated with the extraction of D N A from archaeological specimens is that the procedure often co-extracts impurities that can later complicate, or prevent, the study of the extracted D N A by inhibiting P C R amplification (reviewed by 5). Commonly encountered inhibitory substances found in aDNA extracted from teeth, bones, mummified tissue, and coprolites include humic acids, fulvic acids, tannins, porphyrin products, phenolic compounds, hematin, and collagen type I (37-42). The formation of Maillard products, commonly encountered in coprolite samples, can also prevent P C R amplification by causing D N A to become inaccessibly trapped in these sugar-derived condensation products (12). As the negative results in many aDNA studies are attributed to the presence of P C R inhibitors, our extraction method outlined below pays particular attention to the problem and offers a simple test for the presence of P C R inhibitors in D N A extracts.

Ancient DNA Extraction and Analysis DNA Extraction

2

The following D N A extraction protocol is one that has been developed and reported in two of our studies (5, 43). Remove < 0.5 g from the whole sample. 3

In Archaeological Chemistry; Glascock, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Downloaded by UNIV OF MICHIGAN ANN ARBOR on February 18, 2015 | http://pubs.acs.org Publication Date: August 16, 2007 | doi: 10.1021/bk-2007-0968.ch004

86 With the sample held firmly by a vise, portions of bones and teeth can be carefully removed by sawing off portions with one-time use "disposable" hacksaw blades. This is a relatively cost-effective alternative to the use of a Dremmel tool, the blades for which are quite expensive in comparison to hacksaw blades. Moreover, using the hacksaw provides more control over the dispersal of bone and tooth powder during sawing. A n alternative method, that works particularly well with rib samples, is to gently snap pieces of bone off the whole, while being held in a closed Ziploc bag. Portions of coprolites and quids can be removed from the whole by using one-time use disposable X-acto blades. Interior portions of coprolites and quids should be sampled, i f possible, to avoid sampling any surface material that may have unintentionally become contaminated. Scissors and/or tweezers used to tease apart samples should be cleaned by submerging them in a 50% v/v bleach solution for 5-10 minutes, and then wiping them clean with paper towel. Over time this treatment of the tools can lead to rust, at which time the tools should be replaced. If the sample is of bone or tooth, submerge it in full strength Clorox bleach (6% NaOCl) for 15 min; otherwise add the sample directly to E D T A , as described below. Rinse the sample well with D N A free water (Gibco) to remove the bleach (1-2 times). In a 15 mL conical tube made of polypropylene (polystyrene will melt i f exposed to the mix of phenol:chloroform, below), submerge the sample in molecular grade ( D N A free) 0.5 M E D T A , pH 8.0 (Gibco) for > 48 hours. A n extraction control to which an equal volume of E D T A is added, but contains no sample, should accompany the extraction and be subjected to all of the following steps. To the sample add 3 mg of Proteinase Κ (Invitrogen, Fungal Proteinase K ) and incubate at 65° C for approximately 4 hours. Extract D N A from the digested sample using a three-step phenol/chloroform method. First, extract D N A by adding an equal volume of phenol:chloroform:isoamyl alcohol (25:24:1) to the E D T A containing the sample. Thoroughly mix the layers by inverting the tubes vigorously and then centrifuge the tubes at 3100 rpm for 5 min. Remove the aqueous (top) layer, which contains the D N A , and place it in a new tube. Add to this an equal volume of phenol:chloroform:isoamyl alcohol (25:24:1) and re-extract the aqueous phase as just described. Perform a third extraction by adding an equal volume of chloroform:isoamyl alcohol (24:1) to the aqueous phase just removed. Thoroughly mix the layers by inverting the tubes vigorously. Centrifuge the tubes this time at 3100 rpm for 3 min. Remove the aqueous phase and place it into a new tube. To facilitate removal of co-extracted P C R inhibitors (5, 44), precipitate D N A from the solution by adding one half volume of room temperature 5 M ammonium acetate and, to this combined volume, one volume of room temperature absolute isopropanol. Store the solution for a minimum of 7 hours at 4

In Archaeological Chemistry; Glascock, M., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2007.

Downloaded by UNIV OF MICHIGAN ANN ARBOR on February 18, 2015 | http://pubs.acs.org Publication Date: August 16, 2007 | doi: 10.1021/bk-2007-0968.ch004

87 room temperature (overnight storage at room temperature is acceptable). Centrifuge the tubes for 30 min at 3100 rpm to pellet the D N A . At this point there may or may not be a visible pellet, but this is not an indication or whether the sample contains preserved D N A or not. However, in most cases in which P C R inhibitors have been co-extracted, a pellet is visible as a brown colored smear. Carefully decant the isopropanol from the tube and allow the tubes to airdry, inverted, for 15 min. Wash the D N A pellet by adding 1 mL of 80% ethanol to the tubes followed by 30 s of vortexing, making sure to dislodge the pellet, i f visible, from the side of the tube. Centrifuge the tubes again for 30 min at 3100 rpm. Carefully decant the ethanol and again air-dry the tubes in an inverted state for 15 min. 5

Suspend the D N A in 300 of DNA-free double-distilled water (i.e., d d H 0 ) and perform a silica extraction (45) using the Wizard P C R Preps D N A Purification System (Promega), following the manufacturer's instructions except that: 1) the "Direct Purification Buffer" need not be added (as it is only used for purifying P C R product from reactions containing mineral oil) and 2) finally elute with 100 μί, DNA-free d d H 0 . We recommend the use of the Promega silica extraction kit because they ensure contamination free reagents and it is a relatively inexpensive and quick procedure. 2

2

6

PCR Amplification: Screening SNPs, Molecular Sex Identification, and Sequencing Reactions Even i f contamination is removed from the surface of archaeological specimens (in the case of bones and teeth) or is intentionally avoided by sampling in interior portions of samples (in the case of coprolites and quids), contamination arising during P C R amplification remains a serious problem. We have suggested that using increased amounts ofTaq polymerase to combat P C R inhibitors (46-48) is counterproductive, as it simultaneously increases the sensitivity of P C R reactions to contamination (5). In fact, we have countered the hypersensitive nature of P C R (49) by reducing the amount of Taq polymerase used in our P C R reactions. Additionally, we recommend that small volume (15 μΐ) P C R reactions be used, when possible, for screening single nucleotide polymorphisms (SNPs) and length polymorphisms, as smaller-sized reactions will ultimately exhaust less D N A extracted from archaeological specimens. We further advocate amplifying small (