The use of mitochondrial DNA and short tandem repeat typing in the identification of air crash victims. Nucleic acids.

Transcription

1 Electrophoresis 1999, 20, 1707± William Goodwin 1 Adrian Linacre 2 Peter Vanezis 1 1 Human Identification Centre, Department of Forensic Medicine and Science, University of Glasgow, Glasgow, UK 2 Forensic Science Unit, Pure and Applied Chemistry, University of Strathclyde, Glasgow, UK 1 Introduction Following any incident that results in death there are humanitarian and legal obligations for the accurate identification of all the individuals involved. Disaster Victim Identification (DVI) has drawn on a variety of techniques to establish identification. However, the use of many techniques is limited when the victims have experienced environmental insult resulting in fragmentation of the body and/or severe degradation of the remains. DNA profiling has become a powerful and accepted tool in DVI [1±4]. Short tandem repeat (STR) loci can be amplified from degraded samples by the polymerase chain reaction allowing comparison to other such samples and living relatives [5, 6]. Mitochondrial DNA (mtdna) has certain advantages compared to STR analysis when analyzing degraded DNA due to its high copy number and greater stability [7±9]. mtdna is also advantageous when close relatives are not available to use as a reference. Due to its maternal inheritance it can be matched to any maternal relative. On the 2 nd of February 1998, Cebu Pacific flight 387 carrying 104 passengers and crew on an internal flight in the Correspondence: Dr. William Goodwin, Human Identification Centre, Department of Forensic Medicine and Science, University of Glasgow, Glasgow G12 8QQ, UK why2y@udcf.gla.ac Fax: Abbreviations: mtdna, mitochondrial DNA; DVI, disaster victim identification The use of mitochondrial DNA and short tandem repeat typing in the identification of air crash victims In February 1998, a civilian air plane crashed into a remote mountainside in the Philippines, killing all 104 passengers and crew. The victims were subjected to severe environmental insult, preventing conventional identification methods in most cases. As part of the identification process, samples were subjected to a combination of mitochondrial DNA (mtdna) and short tandem repeat (STR) DNA profiling. The DNA extracted from the victims remains was in all cases highly degraded. However, profiling using mtdna was still successful with 95% of the victims samples; this compared to a 50% success rate using three STR loci. The use of mtdna and STR profiling enabled 187 human fragments from the crash site to be placed into 80 distinct groups; when combined with postmortem data, the samples could be further separated into 95 distinct groups, thereby assisting in the identification process. Keywords: Disaster victim identification / Mass disaster / Mitochondrial DNA / Short tandem repeat / Human identification EL 3479 Philippines crashed into a remote mountainside, killing all on board. In addition to the mechanical and heat stress that the victims suffered during the crash, the recovery of the remains from the crash site was hampered, leading to additional environmental insult. Conventional identification methods left 46 persons unidentified. From the crash site 187 tissue and bone samples were recovered and sent to Glasgow for DNA analysis. There was a requirement to group these 187 samples, with reference to remains that had already been positively identified, and to further identify the victims by comparison to living relatives. 2 Materials and methods 2.1 DNA extraction DNA from muscle and blood samples was extracted using the Puregene Genomic DNA Extraction Kit (Gentra, Plymouth, MN, USA) in accordance with the manufacturer s instructions. DNA was extracted from bone samples by removing 0.1 g of bone material from the samples using a hand-held drill with a 2.5 mm bit. The material was transferred to a 2 ml screw cap tube with 1 ml of extraction buffer (0.5 M EDTA, 100 mg/ml proteinase K, 1% Triton X-100). The extractions were allowed to proceed in a rotary incubator at 50 o C for 18±24 h. The resulting solution was phenol/chloroform-extracted. The aqueous phase was concentrated and exchanged into approximately 100 ml of sterile water using Centricon-30 filters (Amicon, Beverly, MA, USA). All DNA extractions were quantified using the Quantiblot Kit (PE Applied Biosystems, Foster City, CA, USA) according to the manufacturer s instructions. Nucleic acids WILEY-VCH Verlag GmbH, Weinheim, /99/ $ /0

2 1708 W. Goodwin, A. Linacre and P. Vanezis Electrophoresis 1999, 20, 1707± STR analysis D3S1358, vwa, and FGA STR loci were amplified by use of the Perkin-Elmer AmpFISTR Blue Kit. Nine STR loci (D3S1358, vwa, FGA, D21S11, D8S1179, D18S51, D5S818, D13S317 and D7S820) and the amelogenin sex test were amplified using the Perkin-Elmer Profiler Plus Kit. Approximately 2 ng of DNA was added to each PCR and amplification was performed according to the manufacturer s instructions using a PE Applied Biosystems 2400 thermalcycler. The STR loci were separated and detected on a PE Applied Biosystems PRISM 377 DNA sequencer. The data were collected and analysed by the GeneScan computer software. 2.3 mtdna analysis The distal portion of hypervariable region I was amplified using the primer pair LB (CTCCACCATTAG- CACCCAAAGC) and H16431 (CGAGGAGAGTAG- CACTCTTG). The distal 200 bp portion of hypervariable region I was amplified using primer pair LB 16,208 (B- CCCCATGCTTACAAGCAAG) and H The PCR amplifications were performed in a 50 ml volume, containing 10 mm Tris-HCl, 2.5 mm MgCl 2, 200 mm of each dntp (Promega, Madison, WI, USA), 0.2 pm of each primer (Pharmacia, Milton Keynes, UK), 100 mg/ml BSA (Boehringer Mannheim, Lewes, UK) and 1.25 units of Taq DNA polymerase (Promega). PCRs were carried out using a Hybaid Omnigene PCR machine (Ashford, Middlesex, UK). Reactions were initiated for 5 min at 95 o C, 30 cycles consisted of 1 min at 95 o C, 45 s at 50 o C and 1 min at 72 o C. The PCR ended with 20 min at 72 o C. The PCR products were made single-stranded, retaining the biotinoylated strand for sequencing. Streptavidin-coated DynaBeads (M-280) (Dynal, Wirsal, Merseyside, UK) were used in accordance with the manufacturer s instructions. All sequencing reactions were carried out using the Sequenase II sequencing kit (Amersham, Little Chalfont, Bucks, UK), in accordance with the manufacturer s instructions. The reactions were analysed on 55 cm 6% polyacrylamide gels followed by exposure to Fuji XR X- ray film. 2.4 Matching analysis A small number of relatives provided blood samples in order to identify more of the victims. When maternal relatives were provided, an initial match was made using mtdna. To allow for mutations in the mtdna that have been observed at a relatively high frequency [10] one mismatch was permissible to reduce the possibility of false exclusions. Once a match had been made the STR profiles of the remains and the relatives were compared. 2.5 Statistical analysis The probability of randomly selecting two samples with the same mitochondrial profile was calculated using the equation p = Sc2, which is simply the sum of the products of the frequencies of each haplotype. The likelihood ratios for samples matched by STR analysis were calculated according to Balding and Nichols [11]. The significance of matches made to the relatives was calculated as described by Ballantyne [12]. 3 Results 3.1 DNA extraction and quantification The majority of samples did not contain enough DNA to be detected on an ethidium bromide stained gel. However, in a number of samples a visible smear was detected, in which a distinct band could be seen, including one at approximately 200 bp. Other samples showed up to five bands at regular 200 bp intervals. These bands were presumed to be the DNA that was directly associated with the histone core and therefore less prone to degradation than the DNA in the linker region. The amount of DNA detected using the Quantiblot Kit varied greatly; approximately 30% of sample contained over 10 ng/ml while 25% of the samples contained less than 0.5 ng/ml (nondetectable). The discrepancy in the recovery of DNA reflected the type of different samples that were submitted for testing; the muscle samples generally yielded a large amount of DNA while little or no DNA could be detected in the bone and skin samples. 3.2 mtdna analysis A 400 bp region spanning the hypervariable region I was initially amplified and analysed. The amplification of the 400 bp amplicon was problematic for a large number of samples because of the highly degraded nature of the samples. The distal 200 bp of hypervariable region I was therefore amplified and sufficient product for sequencing was obtained from 177 of the samples. Sequencing of this region revealed a high level of polymorphisms within the population of the plane. The samples could be separated into 55 distinct groups, representing potentially 104 victims. The numbers of samples in each of the 55 groups varied between 1 and 16. The probability of two randomly selected samples having the same mtdna profile was 3.77%. This was not an ideal way of predicting the frequency of different haplotypes, as multiple fragments could be present from one individual. However, as maternal relatives were not available for many of the passengers and there was no database for the Philippine popula-

3 Electrophoresis 1999, 20, 1707±1711 Identification of air crash victims by mtdna and STR 1709 Figure 1. Examples of samples amplified with AmpFISTR Blue. Loci D3S1358 (16, 18) and vwa (17, 18) have been detected. No PCR products were detected in the size window of FGA, indicating that DNA fragments of greater than 200 bp were not present in the samples. tion, it was the only means available to predict the frequencies of the different haplotypes within the population of the plane. 3.3 STR analysis All the samples were analysed using the AmpFISTR Blue Kit; and samples that were matched to relatives were profiled using the AmpFISTR Profiler Plus kit. The analysis was completely successful with the AmpFISTR Blue kit with 50% of the samples. After quantification of the DNA, approximately 70% of the samples had been expected to give results when they were analysed by STRs. The lower success rate was possibly caused by the higher amounts of quantifiable DNA relative to the amplifiable DNA. Seven of the samples that yielded PCR products did so for only two of the three loci. The dropout of the third allele was in all cases dramatic (Fig. 1). The dropout between the vwa and the FGA loci coincided with the increase in allele size to over 200 bp, indicating that this was the approximate size of the longest DNA fragments extracted. No dropout was observed between the D3S1358 and vwa loci for any of the samples. Care had to be taken when assessing the FGA locus in all samples where homozygotes were detected as this spans a considerable range (220±268 bp) allowing the possibility of allelic dropout. However, this did not seem to be a major problem as the observed homozygosity of 10% was close to the predicted value at this loci [13]. 3.4 Identification Figure 2. Flow diagram illustrating the process of identification. There were 104 victims on the plane, from which 187 samples were collected at the crash site. Using mtdna sequencing the 187 samples produced 55 different groups. STR analysis, using AmpFISTR Blue, divided the 55 groups into 80 groups. Some of the samples in the 80 groups were incompatible, such as two right hands. Using postmortem data the 80 groups were finally divided into 95 groups. Not all 104 victims were accounted for, which may be due to the original samples collected. The primary aim of the DNA analysis was to group samples together that came from the same individual. As shown in Fig. 2, 187 tissue samples were recovered from the 104 victims. After analysis with mtdna it was possible to split the remains into 55 distinct groups. STR analysis could subdivide the 55 groups further, which led to a total of 80 groups. The use of postmortem data to separate incompatible matches produced 95 distinct groups. The use of STR analysis allowed the determination of the significance of the matches to be calculated. The resulting likelihood ratios varied between 7000 and When matches could only be made using mtdna it was more problematic because of the low statistical significance of the matches. These matches could only act as a guide and needed additional postmortem information to be confirmed. When samples were matched to relatives (as shown in Fig. 3), where nine loci were analysed, the likelihood ratios could be extremely high.

4 1710 W. Goodwin, A. Linacre and P. Vanezis Electrophoresis 1999, 20, 1707±1711 Figure 3. Example of identification of a victim from their living relatives. In this case nine STRs were generated from a victim and potential mother and father (six are shown). A likelihood ratio for the identity of the victim as being the child of the parents was determined by dividing the probability of the two parents producing the genotype of the victim, with the frequency of the victim s genotype. In this case a likelihood ratio of 15 million was obtained. 4 Concluding remarks The combined use of mtdna and STR analysis used in this case had several advantages. Despite the extreme levels of degradation it was possible to successfully type 95% of the samples using mtdna. While the information content of mtdna is not as great as obtained using STRs and other nuclear markers it does provide information that can potentially be used in conjunction with other information available to the DVI team to determine the unequivocal identity of the remains. When matches could be made using mtdna, and confirmed using STRs, the significance of the matches were extremely high, especially considering that the population involved was only 104. DNA typing undoubtedly provides an extremely powerful tool for human identification and, with the increase in forensic DNA typing facilities worldwide, its use in DVI will continue to increase. This case demonstrated that while it is not possible to DNA profile every sample from a mass disaster, it should, with a combination of mtdna and STR analysis, be possible to improve the success rate of analyzing highly degraded remains. Currently the DVI teams have to assess the benefits of DNA typing weighed against the disadvantages, which are that it is time-consuming and expensive. While it is unlikely to be used as a primary tool in all DVI cases, increasing the success rate of analysis should make it a more attractive tool to be drawn on in a wide variety of cases. Received February 26, References [1] Clayton, T. M., Whitaker, J. P., Maguire, C. N., Forensic Sci. Int. 1995, 76, 7±15. [2] Clayton, T. M., Whitaker, J. P., Fisher, D. L., Lee, D. A., Holland, M. M., Weedn, V. W., Maguire, C. N., DiZinno, J. A., Kimpton, C. P., Gill, P., Forensic Sci. Int. 1995, 76, 17±25. [3] Whitaker, J. P., Clayton, T. M., Urquhart, A. J., Millican, E. S., Downes, T. J., Kimpton, C. P., Gill, P., BioTechniques 1995, 18, 670±677. [4] Olaisen, B., Stenersen, M., Mevag, B., Nature Genet. 1997, 15, 402±405. [5] Jeffreys, A. J., Allen, M. J., Hagelberg, E., Sonnberg, A., Forensic Sci. Int. 1992, 56, 65±76. [6] Gill, P., Ivanov, P. L., Kimpton, C., Piercy, R., Benson, N., Tully, G., Evett, I., Nature Genet. 1994, 6, 130±135. [7] Sullivan, K. M., Hopgood, R., Gill, P., Int. J. Leg. Med. 1992, 105, 83±85. [8] Bogenhagen, D., Clayton, D. A., J. Biol. Chem. 1974, 249, 7991±7995.

5 Electrophoresis 1999, 20, 1707±1711 Identification of air crash victims by mtdna and STR 1711 [9] Robin,E. D., Wong, R., J. Cell Physiol. 1988, 136, 507±513. [10] Parsons, T. J., Muniec, D. S., Sullivan, K., Woodyatt, N., Alliston Greiner, R., Wilson, M. R., Berry, D. L., Holland, K. A., Weedn, V. W., Gill, P., Holland, M. M., Nature Genet. 1997, 15, 363±368. [11] Balding, D. J., Nichols, R. A., Forensic Sci. Int. 1994, 64, 125±140. [12] Ballantyne, J., Nature Genet. 1997, 15, 329±330. [13] Oldroyd, N. J., Urquhart, A. J., Kimpton, C. P., Millican, E. S., Watson, S. K., Downes, T., Gill, P. D., Electrophoresis 1995, 16, 334±337.