Chapter 3
Kennewick Man Ancient DNA Analysis: Final Report Submitted to the Department of the Interior, National Park Service
By D. Andrew Merriwether, Graciela S. Cabana, and David M. Reed

On May 18, 2000, two bone samples from the Kennewick collection - a right eighth rib [Catalog 97.I.12d(13)] and a left third metacarpal (MC3) [Catalog 97.L.16(Mca)] - were transferred to the laboratory of D. Andrew Merriwether for ancient DNA analysis. Three people with extensive ancient DNA research experience were involved in the subsequent analysis of these samples: the principal investigator, D. Andrew Merriwether, a doctoral student, Graciela Cabana, and a laboratory technician, John McDonough. Ancient DNA analysis of the Kennewick samples included five separate extractions using four different extraction procedures and multiple DNA amplification attempts using Polymerase Chain Reaction (PCR) on these extractions. We were unable to successfully amplify any ancient mitochondrial DNA (mtDNA) using standard methods. We conclude that if any DNA remains in these particular samples, it is inaccessible via current standard methods, but may be accessible and analyzable in the future using novel methods.
Because we did not obtain positive results in this analysis, this report consists primarily of a detailed description of the methods used and explanation of the results obtained. We attach as appendices (1) a log of electrophoresis gel images documenting the results of PCRs on these samples (Appendix 1), (2) a photographic log of procedures described in the text of this report (Appendix 2), and electropherograms documenting the contaminating sequences (Appendix 3).

DNA Analysis: Methods

Below is a general description of the methods used for the ancient DNA analysis of these Kennewick bone samples.

Contamination Controls
Extraordinary care has been taken to control for exogenous human DNA contamination. Contamination in ancient DNA work, however, is inevitable and controls must be included at every step of the process to monitor its effects. We followed these preventative steps:

Decontamination of Bone Surfaces & Bone Extraction Preparation
The outside of Kennewick bone samples were treated to remove any potential exogenous human DNA contamination by first scrubbing the bone with a 10% bleach solution using a sterilized brush. All bone surfaces were then subjected to UV light for half an hour to cross-link any superficial DNA.

Bone extraction preparation varied according the type of bone sample. The MC3 was sampled by drilling a fine bone powder using a Dremel tool. The Dremel bit, parts and tool were cleaned with a 10% bleach solution and UV-irradiated prior to use. The rib was either scraped or ground because there was little actual bone accessible with a drilling tool, and the inside of the rib was filled with a tough soil matrix. For the first extraction of the rib (Extraction #3, below), the rib sample was soaked in UV/UF H₂O for an hour in an attempt to loosen the soil matrix from the bone. The soil matrix did not dissolve, so the cortex of the rib was then shaved off with a sterile razor blade and only these shavings were used in the subsequent extraction. This method had the advantage of not including potential PCR inhibitors in the soil matrix, but also discarded any internal rib bone pieces from the extraction. For the second and third rib extraction (Extraction #4 and #5, below), a piece of rib with its soil matrix was simply placed under a sterilized piece of bench paper and ground with a sterilized hammer. The entire piece, matrix and all, was used in the subsequent extraction to maximize the amount of ancient human bone in the sample.

Extraction Procedures
We used four different extraction procedures on five separate occasions. Twice we attempted DNA extraction using a phenol-chloroform method (e.g., Kaestle 1998), once with a phenol-chloroform-silica method (modified from Baron et al. 1996), once with the QiaQuick PCR Purification Kit (Qiagen) (Yang et al. 1998), and once with a guanidine isothiocyanate - silica method (modified from Hoss & Paabo 1993). The first and second phenol-chloroform-based methods are known to produce the highest DNA yields, but have the disadvantage of co-extracting humic acids and possibly other elements typically present in soils that react with reagents used during PCR. These may subsequently inhibit PCR reactions and yield false negative results. The other two methods produce cleaner results, meaning that inhibitory elements tend to be removed. We used a variety of methods and attempted one method twice because results do vary from one extraction to another, even with the same protocol. The extraction methods used for the Kennewick samples can be summarized as follows:

Phenol-Chloroform (e.g., Kaestle 1998)

Phenol-Chloroform-Silica (modified from Baron et al. 1996)

Silica-Based Extraction (Yang et al. 1998)

Guanidine Isothiocyanate - Silica (modified from Hoss & Paabo 1993)

Post-Extraction "Clean-Up" Procedure using Proteinase K
To further remove inhibitors from extracted samples, a Proteinase K clean-up method was used on Extractions #1 and #2. The method follows:

Polymerase Chain Reaction (PCR) Amplification Procedures

We performed PCR amplifications under varying conditions depending on the Taq manufacturer's specifications. In addition, we used two cycling protocols, a standard PCR and a touch-down PCR.

The standard PCR cycling conditions began with an initial 4-10 min denaturation step at 94 °C, followed by 40 to 45 cycles in which the DNA samples are denatured at 94 °C for 40 sec, annealed at 52 °C for 40 sec, and extended at 72 °C for 40 sec. This cycling procedure is followed by a final extension step at 72 °C for 2 min.

The touchdown PCR involved an initial 2 min denaturation step at 94 °C, followed by 50 cycles in which the DNA samples are denatured at 94 °C for 14 seconds, annealed in the first of the 50 cycles at 58 °C for 30 seconds with a decrease in 0.1 °C in each successive cycle, and extended at 72 °C for 15 seconds.

Most PCRs included a positive control of modern DNA that is always added outside the ancient DNA room. This is to check that the PCR reaction itself functioned, even in the case of no amplification of the ancient samples.

DNA Target
Our strategy was to begin by targeting DNA from the maternally inherited mitochondrion, or mtDNA, for two reasons. First, mtDNA is valuable for population studies, in that it allows for the reconstruction of maternal lineages that often are correlated with geographical groups. In the case of the New World, it is widely accepted that most modern Native Americans fall into one of five mtDNA-based maternal lineages, or haplogroups, labeled A, B, C, D and X (e.g., Smith et al. 1999; Merriwether et al. 1995; Schurr et al. 1990). Studies of ancient samples of ancient Native Americans confirm the presence of these maternal lineages before European contact (e.g., O'Rourke et al. 1999; Stone and Stoneking 1998). If DNA amplification had been successful in the case of the Kennewick sample, such a haplogroup assignment might have been possible.

Second, because mtDNA exists in higher copy numbers per cell relative to nuclear DNA, it is best-suited for ancient DNA analyses. Positive amplification of mtDNA is a strong indicator of the general presence of DNA in ancient samples. If no mtDNA amplifies, it is unlikely that nuclear DNA remains in an ancient sample.

Moreover, in our experience, DNA from the displacement loop, or D-loop, of the mitochondrion tends to amplify more readily than DNA outside the mtDNA D-loop. Thus, our initial work targeted portions of the D-loop and we designed our primer pairs used in PCR amplifications accordingly. If amplification of the D-loop had occurred, then tentative population-specific haplogroup assignments could have been made, since certain D-loop mutations are known to be closely associated with population-specific mtDNA haplogroups. At that point, the appropriate mutations outside the D-loop would have been targeted to confirm the haplogroup assignment based on D-loop sequences.

DNA Visualization
After PCR amplification, 10 µl of the PCR product along with approximately 1 µl of sucrose loading dye was loaded onto a 3 % NuSieve/1 % agarose gel stained with ethidium bromide. The gel images were "photographed" on thermal paper using a UV transilluminator attached to a digital imaging system. These gel images were subsequently scanned and appear in Appendix 1.

DNA Sequencing
When we successfully amplified DNA from PCRs that yielded negative PCR controls as well as extraction controls (see PCRs 16 and 17), we sequenced the resulting PCR products using the following method:

Summary of Extractions and Results
DAM = D. Andrew Merriwether
GSC = Graciela S. Cabana
JM = John McDonough

Extraction Log

Extraction #1: performed by DAM from 5/22/00 to 5/25/00. A bone powder sample of the MC3 was taken, and the DNA extracted using the Qiagen method, as above.

Extraction #2: performed by DAM from 5/29/00 to 5/31/00. A bone powder sample of the MC3 was taken, and the DNA extracted using the Phenol-Chloroform method, as above.

Extraction #3: performed by GSC from 6/14/00 to 6/16/00. Bone scrapings were taken from the rib bone, and the DNA extracted using the modified Guanidine Isothiocyanate - Silica method, as above.

Extraction #4: performed by GSC from 7/4/00 to 7/6/00. A ground bone sample was taken from the rib bone, and the DNA extracted using the Phenol-Chloroform method, as above.

Extraction #5: performed by GSC from 9/3/00 to 9/6/00. The remaining rib bone was ground, and the DNA extracted using the Phenol-Chloroform-silica method, as above.

PCR Log

This log complements the gel images shown in Appendix 1.
Most PCR gels have standard size 1 kilobase ladders flanking the samples in both the left-most (lane 0) and right-most lanes.

PCR 1 (GSC)
Primers 16192 to 16322 (mtDNA D-loop)
PCR of Extraction #1 (Qiagen method)

Result: Although the PCR negative controls did not show amplification, the Kennewick extract controls did show contamination. This does not mean, however, that the reagents used in the PCR did not contain exogenous human contamination. This is because if very low levels of exogenous DNA are in the PCR reagents, and there are more "samples" than PCR negative controls, chances are the contamination from the PCR reagents will show up in the samples. Thus, we typically wait to see the results of several PCRs before determining that an extraction product is itself contaminated. Indeed, subsequent PCRs on this extract suggested to us that the this extract was not contaminated.

PCR 2 (GSC)
Primers 16268 to 16375 (mtDNA D-loop)
PCR of Extraction #1 (Qiagen method) and Extraction #2 (Phenol-Chloroform method)

Result: No amplification except for positive control.

PCR 3 (GSC)
Primers 16112 to 16237 (mtDNA D-loop)
PCR of Extraction #1 (Qiagen method) and Extraction #2 (Phenol-Chloroform method)

Result: PCR contaminated by an unknown source. The PCR reagents were discarded and samples were re-run with new reagents (below).

PCR 4 (GSC)
Primers 16112 to 16237 (mtDNA D-loop)
PCR of 1:20 dilutions of Extraction #1 (Qiagen method) and Extraction #2 (Phenol-
Chloroform method)

Result: No amplification except for positive control.

PCR 5 (GSC)
Primers 16268 to 16375 (mtDNA D-loop)
PCR of 1:100 dilutions of Extraction #1 (Qiagen method) and Extraction #2 (Phenol-Chloroform method)

Result: No amplification except for positive control.

PCR 6 (GSC)
Primers 16268 to 16375 (mtDNA D-loop)
PCR of Extraction #3 (modified Guanidine Isothiocyanate - Silica method)

Lane Sample Amplification
status 1-14 Other ancient samples 14 Kennewick (Ext. #3) Negative
15 Lysis Extraction Buffer Negative 16 Extraction Buffer Negative
17 Extract Control Negative 18 PCR Negative Control (+ H₂O) Negative 19 PCR Negative Control Negative 20 PCR Positive Control Positive

Result: No amplification except for positive control.

PCR 7 (DAM)
Primers 8195 to 8317 (Haplogroup B)
PCR of Extraction #1 (Qiagen method), Extraction #2 (Phenol-Chloroform method) and Extraction #3 (modified Guanidine Isothiocyanate - Silica method)

Result: PCR contaminated by an unknown source. The PCR reagents were discarded and samples were re-run with new reagents (below).

PCR 8 (GSC)
Primers 8244 to 8313 (Haplogroup B)
PCR of a Post-Extraction Clean-up Procedure (Proteinase K) of Extraction #1 (Qiagen method), Extraction #2 (Phenol-Chloroform method) and Extraction #3 (modified Guanidine Isothiocyanate - Silica method)
Note: this gel was made by melting down a previously-used gel, hence the stray band in the gel image.

Result: No amplification except for positive control.

PCR 9 (DAM)
Primers 16192 to 16375 (mtDNA D-loop)
PCR of a Post-Extraction Clean-up Procedure (Proteinase K) of Extraction #1 (Qiagen method), Extraction #2 (Phenol-Chloroform method) and Extraction #3 (modified Guanidine Isothiocyanate - Silica method)  

Result: PCR contaminated.

PCR 10 (JM)
Primers 16112 to 16237 (mtDNA D-loop)
PCR of a Post-Extraction Clean-up Procedure (Proteinase K) of Extraction #1 (Qiagen method), Extraction #2 (Phenol-Chloroform method) and Extraction #3 (modified Guanidine Isothiocyanate - Silica method)

Result: No amplification except for positive control.

PCR 11 (JM)
Primers 16268 to 16375 (mtDNA D-loop)
PCR of a Post-Extraction Clean-up Procedure (Proteinase K) of Extraction #1 (Qiagen method), Extraction #2 (Phenol-Chloroform method) and Extraction #3 (modified Guanidine Isothiocyanate - Silica method)  

Result: No amplification except for positive control.

PCR 12 (JM)
Primers 16192 to 16322 (mtDNA D-loop)
PCR of a serial dilution of Extraction #1.  

Result: No amplification except for positive control.

PCR 13 (GSC)
Primers 642 to 708 (Haplogroup A)
PCR of a small fragment (66bp) of Extraction # 3 and #4.  

Result: One Kennewick DNA sample and 2 extraction controls amplified. PCR product was subsequently digested with the Hae III enzyme (below).

Digest 1 (GSC)
Product of PCR 13 digested the Hae III enzyme to detect the mutation defining Haplogroup A.  

Result: No Kennewick sample or extract control digested.

PCR 14 and 15 (GSC)
Primers 16112 to 16237 (PCR 14) and Primers 16192-16322 (PCR 15) (mtDNA D-loop).
PCRs of Extraction # 5 (modified Phenol-Chloroform-silica method).  

Result: The Kennewick samples did not yield specific amplification products in that no distinct bands at the appropriate levels on the gel appeared. However, a smear of non-specific DNA appeared in the Kennewick sample lanes that is typical when bacterial and fungal DNA are also present. This is not unlikely given that the rib sample included a soil matrix. The co-extraction of bacterial and fungal DNA also suggests that humic acids and other potential inhibitors present in soils might be affecting our ability to detect any human DNA. One way to lessen their effect is to dilute the extract with the hopes of diluting inhibitory elements. PCRs 16 and 17 (below) represent PCRs based on these dilutions.

PCR 16 (GSC)
Primers 16112 to 16237 (mtDNA D-loop).
PCRs of dilutions of Extraction # 5 (modified Phenol-Chloroform-silica method).  

Result: Positive amplification of Kennewick sample (Lane #6), while PCR and extraction negatives showed no DNA amplification. The resulting PCR product was sequenced, and found to match the sequence of JM.

PCR 17 (GSC)
Primers 16192 to 16322 (mtDNA D-loop).
PCRs of dilutions of Extraction # 5 (modified Phenol-Chloroform-silica method).  

Result: Positive amplification of Kennewick sample (Lanes #2 and 6), while PCR and extraction negatives showed no DNA amplification. The resulting PCR product was sequenced, and found to match the sequence of JM (see Appendix 3 which contains electropherograms of the sequences from PCRs 16 and 17, with the sites differing from the published reference sequence circled). Note that for the region sequenced, all three mutations match the sequence of JM and there are no positions that do not match JM. All are consistent with a European origin of the sequence. The observed sequence is not found in genebank or HVRbase, only in JM from the Merriwether lab. The 16145 mutation has not been observed before. Thus the best explanation for this sequence is contamination of the extract by unknown means by the DNA from JM.

Conclusion

We were unable to obtain reliable ancient DNA amplification results from the Kennewick samples. This means that either (1) no original DNA was preserved in the bone samples transferred to our laboratory, or (2) original DNA was preserved in the bone samples but we were unable to extract it. The latter could occur either because the extraction methods used in this analysis were inadequate, or because we were unable to overcome PCR inhibitors once the DNA was extracted. However, we used a variety of methods to overcome PCR inhibition, including an additional Proteinase K clean-up step to further break down proteins in the samples, dilutions of the extractions reduce the proportion of inhibitors relative to DNA, and the use of BSA in the PCR reaction (BSA may bind to inhibitors, thus inhibiting the inhibitors).

Moreover, any DNA amplification we obtained seems to have been a product of exogenous human contamination. This is immediately clear in two cases: (1) when the PCR negative controls are contaminated, and (2) when the PCR product is sequenced and matched to an investigator (e.g., an excavator or other analyst). The first case implies that at the very least the reagents used in the PCR are contaminated, and reagents become easily contaminated through multiple uses. In this case, we threw out the old PCR reagents and performed a new PCR with fresh reagents. The second case demonstrates that even under the most stringent conditions, modern contaminating DNA is problematic in ancient DNA analysis.

Finally, it is possible that an amplified DNA product is the result of exogenous human contamination despite the fact that negative controls yield no amplification. If very low levels of exogenous DNA exist in the PCR reagents, and there are more "samples" than PCR negative controls, chances are the contamination from the PCR reagents will show up in the samples, rather than in the PCR controls. For similar sampling reasons, low levels of exogenous DNA in DNA extracts and disposable pipette tips, tubes, etc., also will amplify sporadically, from one PCR to another. Thus, in addition to the use of controls, we monitor the results of positive amplifications by sequencing the amplification products. For example, PCRs 14, 15, 16, and 17 showed no positive amplification of PCR or extract controls, yet the sequences (shown in Appendix 3) were an exact match of our laboratory technician.

References Cited

Baron, H., S. Hummel, B. Herrmann
(1996)       Mycobacterium tuberculosis Complex DNA in Ancient Human Bones, Journal of Archaeological Science, 23:667-671

Hoss, M., and S. Paabo
(1993)       DNA Extraction from Pleistocene Bones by a Silica-Based Purification Method. Nucleic Acids Research 21(16): 3913-3914

Kaestle, F. A.
(1998)       Molecular Evidence for Prehistoric Native American Population Movement: The Numic Expansion. Ph.D., University of California, Davis.

Merriwether, D. A, F. Rothhammer, R. E. Ferrell
(1995)       Distribution of the Four Founding Lineage Haplogroups in Native-Americans Suggests a Single Wave of Migration for the New-World. American Journal of Physical Anthropology 98(4):411-430.

O'Rourke, D. H., R. L. Parr and S W. Carlyle
(1999)       Molecular Genetic Variation in Prehistoric Inhabitants of the Eastern Great Basin. In Understanding Prehistoric Lifeways in the Great Basin Wetlands: Bioarchaeological Reconstruction and Interpretation. Edited by B. E. Hemphill and C. S. Larsen. Salt Lake City: University of Utah Press, 84-102.

Schurr, T. G., S. W. Ballinger, Y. Gan, J. A. Hodge, D. A. Merriwether, D. N. Lawrence, W. C. Knowler, K. M. Weiss, and D. C. Wallace
(1990)       Amerindian Mitochondrial DNAs Have Rare Asian Mutations at High Frequencies, Suggesting They Derived from Four Primary Maternal Lineages. American Journal of Human Genetics 46:613-623.

Smith, D. G., R. S. Malhi, J. Eshleman, J. G. Lorenz, and F. A. Kaestle
(1999)       Distribution of mtDNA Haplogroup X among Native North Americans. American Journal of Physical Anthropology 110(3):271-284.

Stone, A. C. and M. Stoneking
(1998)       MtDNA Analysis of a Prehistoric Oneota Population: Implications for the Peopling of the New World. American Journal of Human Genetics 62:1153-1170.

Yang, D. Y., B. Eng, J. S. Waye, J. C. Dudar and S. R. Saunders
(1998)       Technical Note: Improved DNA Extraction from Ancient Bones Using Silica-Based Spin Columns. American Journal of Physical Anthropology 105:539-543

Kennewick Home