Marie D. Ricciardone1, Ann M. Lins2, James W. Schumm2,
and Mitchell M. Holland1
1 Armed Forces DNA Identification Laboratory, Office of the Armed Forces
Medical Examiner, Armed Forces Institute of Pathology, Rockville, MD
2 Promega Corporation, 2800 Woods Hollow Road, Madison, WI
The Armed Forces DNA Identification Laboratory (AFDIL) uses DNA typing to identify remains of military service members fatally injured in battle, aircraft crashes, and other accidents. The recovered remains are often fragmented and/or partially incinerated and, therefore, difficult to identify by traditional methods such as fingerprinting and dental comparison. DNA typing by short tandem repeat (STR) analysis is a sensitive method of identification that can be used even when the DNA is degraded.
STR loci consist of tandemly repeated sequence motifs, 3-7 base pairs in length, that are widely distributed throughout the human genome (1,2,3,4). Since the number of repeat motifs is highly polymorphic, STRs are ideal genetic markers for personal identification. Individual STR profiles are generated by polymerase chain reaction (PCR) amplification of the STR loci using primers that flank the repeat sequence. Each amplified fragment has a discrete length that depends on the number of repeat motifs. The amplified fragments are separated by denaturing gel electrophoresis and subsequently detected by silver stain or fluorescence methods.
This study investigated the feasibility of using the fluorescent GenePrint™ STR systems (Promega Corporation, Madison, WI) with 373 DNA Sequencer analysis (Applied Biosystems, Perkin-Elmer, Sunnyvale, CA). The 373 DNA Sequencer detects fluorescent PCR fragments as they migrate through the gel past a scanning laser beam. The instrument uses GeneScan™ software to automatically size the PCR fragments. This method offers a precise, accurate, and sensitive method of DNA fragment detection that has been successfully employed with several STR systems for personal identification (5,6,7,8,9,10).
Eight STR systems (11,12) were evaluated as two fluorescent multiplexes: the CTTV multiplex (CSF1PO, TPOX, TH01, and VWF loci) and the FFFL multiplex (F13A01, FESFPS, F13B, and LPL loci). The study analyzed the precision of GenePrint™ fragment length measurement, the sensitivity of the GenePrint™ systems, and the application of the GenePrint™ systems in forensic casework. Sixty-three specimens representing 21 sets of human remains were analyzed with both the GenePrint™ multiplex systems and an STR quadruplex system (FFTV:FESFPS, F13A01, TH01, and VWF) developed by the Forensic Science Service (Birmingham, UK) (7, 13) and validated for casework in our laboratory (14).
The CTTV multiplex contained the HUMCSF1PO (291-327 bp), HUMTPOX (224-252 bp), HUMTH01 (179-203 bp), and HUMVWFA31 (131-171 bp) primer pairs (11,12). This multiplex was evaluated using both fluorescein (FL)-labeled and HEX- labeled CTTV primers. The FFFL multiplex contained the HUMF13A01 (281-331 bp), HUMFESFPS (222-250 bp), HUMF13B (169-189 bp), and HUMLIPOL (105-133 bp) primer pairs (11,12) labeled with fluorescein. The CTTV and FFFL allelic ladders (11,15,16) contained a mixture of many or all of the known alleles for the STR loci.
The FFTV multiplex contained the HUMFESFPS, HUMF13A01, HUMTH01, and HUMVWFA31 primer pairs (7). In this multiplex the size ranges for the HUMFESFPS and HUMF13A01 fragments overlap, as do the size ranges for the HUMTH01 and HUMVWFA31 fragments. In order to distinguish the alleles, the HUMFESFPS and HUMTH01 primers were labeled with 6-FAM and the HUMF13A01 and HUMVWFA31 primers were labeled with HEX.
Genomic DNA was extracted from blood and tissue specimens using Chelex resin (17).
Four STR loci were co-amplified using a mixture of locus-specific primer pairs, each with one fluorescent and one unlabeled primer. The concentration of template DNA was approximately 2 ng (except where noted).
GenePrint™ STR Systems: DNA and primer pairs were combined in 0.5 ml thin-walled GeneAmp tubes (Perkin-Elmer, Norwalk, CT) with 25µl of STR buffer (Promega Corporation, Madison, WI); one unit of Taq DNA polymerase (Perkin-Elmer, Norwalk, CT or Boehringer Mannheim Corporation, Indianapolis, IN); and one drop of mineral oil (Perkin-Elmer, Norwalk, CT). PCR cycling was performed in a GeneAmp 9600 System (Perkin-Elmer, Norwalk, CT) using the following parameters: initial denaturation at 96°C for 2 minutes; followed by 10 cycles of 94°C for 1 minute, 70°C for 1.5 minutes; followed by 20 cycles of 90°C for 1 minute, 60°C for 1 minute, 70°C for 1.5 minutes; followed by a final extension at 70°C for 10 minutes.
FFTV Multiplex: DNA was combined in 0.5 ml thin-walled GeneAmp tubes with 20µl of a multimix containing fluorescent primer pairs; Parr-Excellence™ buffer (10mM Tris-HCL pH 8.3, 50mM KCl, 1.5 mM MgCl2, 1% Triton X-100); and 200 µM each deoxyribonucleotide triphosphate (dNTP) (Forensic Science Service, Birmingham, UK). Each reaction contained 1.25 units Taq DNA polymerase (Perkin-Elmer, Norwalk, CT). Final reaction volume was 50µl. Each reaction was overlaid with one drop of mineral oil (Perkin-Elmer, Norwalk, CT) using the following parameters: 28 cycles of 95°C for 1 minute, 54°C for 1 minute with a ramp time of 2 minutes; 72°C for 1 minutes; followed by a final extension at 72°C for 10 minutes.
Detection and Analysis of Amplified Products
PCR-amplified STR fragments were separated by electrophoresis on a 373 DNA Sequencer equipped with GeneScan™ 672 fragment analysis software (Perkin-Elmer, Applied Biosystems Division, Foster City, CA). The denaturing gel, 0.4mm thick, contained 6% polyacrylamide (19 acrylamide: 1 bis-acrylamide; BioRad, Hercules, CA), 8 M urea (Mallinckrodt, Paris, KY), and 1X Tris-borate EDTA buffer (Amresco, Solon, OH). Two microliters of the amplification reaction mixture were combined with 3 fmol of the internal lane standard, GeneScan™-2500 [ROX] (Perkin-Elmer, Applied Biosystems Division, Foster City, CA), mixed with an equal volume of formamide. This mixture was heat denatured at 95°C for 2 minutes immediately prior to loading. Fragments were electrophoresed at 30 watts constant power for approximately 3.5 hours. The distance from the well to the laser excitation/fluorescence detection region was 12 cm. Fragment sizes were determined automatically by the GeneScan™ 672 analysis software using the local Southern method. Precision was calculated using the following formula:
STR Precision and Allele Assignment
The precision of STR fragment length determination on the 373 DNA Sequencer was evaluated by replicate analysis of allelic ladders consisting of fluorescently labeled DNA fragments representing most of the known alleles for each STR locus. Representative electropherogram tracings for the CTTV (FL), CTTV (HEX), and FFFL ladders are shown in Figure 1.
The allelic ladder fragments were automatically analyzed by the GeneScan™ software, which used the internal lane markers to construct a calibration curve and size the DNA fragments according to the local Southern method. Fragment sizing was very reproducible with inter-gel precision ranging from 99.87% to 99.95% (Table 1). The range of fragment lengths observed for individual alleles was less than 1 base pair (Table 1); this ensures unambiguous allele assignment.
Band size measurements were defined for each allele based on the mean fragment length observed in the precision studies ± 4 standard deviations. Statistically, approximately 99.99% of all observations should fall between these boundary values. Allele assignments for typed DNA samples were then made by comparing the fragment length calculated by the GeneScan™ software to the band size measurements defined for the individual alleles. In blind testing of twenty DNA samples, alleles assigned by AFDIL using GeneScan™ analysis agreed with those assigned by Promega Corporation using visual comparison with silver stained allelic ladders (18).
The sensitivity of the multiplexes was evaluated by amplification and gel analysis of twofold serial dilutions of template DNA. Amplified fragments from reaction mixes which contained 10 ng to 0.2 ng of template DNA were successfully typed with both the CTTV multiplex (Figure 2) and the FFFL multiplex (Figure 3). With 10 ng of template DNA, the intensity of fluorescence was quite high (greater than 6000 units) and many alleles appeared as doublet peaks. Although the measurement precision was reduced with broad allele peaks, the fragment lengths still fell into the previously defined band size measurements and the alleles were correctly typed. With 0.2 ng of template, DNA the intensity of fluorescence was lower. However, all peaks showed heights greater than the threshold value (50 units) and were correctly typed. Thus, both multiplex systems can be used to type samples over a 50-fold range of template DNA.
Sixty-three specimens, representing 21 individuals, were recovered from the crash sites of four separate accidents: a B1 bomber crash in Valentine, TX; a C21 jet crash in Alexander City, AL; a CH47D helicopter crash in Fort Hood, TX; and a C130 plane crash in Bliss, ID. In all these incidents, the high velocity impact, subsequent explosion, and post-crash fire caused extensive fragmentation, desiccation, and charring of the human remains. Twelve of the recovered specimens were positively identified by fingerprints, footprints, or odontology, and served as references for re-association of other body parts. Twelve additional blood reference samples were also obtained: seven bloodstain cards from the Department of Defense DNA Repository, and five blood samples from family members.
All specimens were analyzed with both the GenePrint™ CTTV multiplex and the previously validated FFTV multiplex. The TH01 and VWF alleles are amplified by both the CTTV and FFTV systems; this overlap allowed a direct comparison of the two systems loci. Complete profiles were obtained for 62 of the 63 submitted evidence specimens and for all 12 reference specimens. Identical TH01 and VWF profiles were obtained with both systems.
The one specimen that did not give full CTTV profile consisted of a small amount of blood-like material swabbed from a helicopter rotor blade several days after the incident. Investigators had collected this sample to confirm that the helicopter blade had indeed decapitated one of the victims. Analysis with both the CTTV and the FFTV multiplexes revealed a partial STR profile, consisting of the VWF and TH01 loci. These results are consistent with severe degradation of the sample (13). Only the decapitated victim had a matching VWF/TH01 profile.
CTTV analysis of one specimen revealed evidence of a mixture at the VWF, TPOX, and CSF1PO loci (Figure 4). This specimen also showed evidence of a mixture with the FFTV multiplex at the VWF locus (data not shown). The major alleles (VWF-15, VWF-16, TH01-9.3, TPOX-8 and CSF1PO-10) matched the STR profile of one reference specimen (Table 2). The minor alleles (VWF-19, TPOX-10, CSF1PO-11, and CSF1PO-12) were found in the STR profile of another second reference specimen (Table 2). Thus, STR analysis allowed identification of the major and minor contributors to this specimen.
Seven sets of remains were identified by comparing the STR profile of the evidence to the STR profile of the corresponding bloodstain card retrieved from the Department of Defense DNA Repository. Twelve pieces of tissue recovered from the crash site were identified by fingerprints, footprints, or odontology. The STR profiles of these tissues were used as references for the re-association of other human remains. Two individuals were identified by parentage analysis. In the first case, the STR profile of the victim was compared to the STR profiles of the two parents (Figure 5). The victim's profile contained one maternal and one paternal allele at each locus. The second individual was identified by comparing the STR profile with the STR profiles of his two children and spouse.
Twelve of the evidence samples were analyzed with both CTTV and FFFL multiplexes. Separate amplification reactions were performed using the HEX-labeled CTTV primers and the fluorescein-labeled FFFL primers. The amplification reactions were combined and electrophoresed in a single gel lane. Identification was made by comparison of the eight-locus STR profiles obtained from the tissue specimens with the eight-locus STR profiles obtained from the reference blood samples (Figure 6). The STR profiles of the evidence and reference were declared a match when the allele fragments detected in both tissues were within 0.5bp.
STR analysis using the GenePrint™ STR systems and the Applied BioSystems 373 DNA Sequencer is a precise, sensitive, and accurate method of human identification. Precision was greater than 99.85% for all alleles at the eight loci tested. The range of allele size measurements, based on the mean fragment length ± 4 standard deviations, was less than one base pair for each allele. The GenePrint™ multiplex systems successfully typed samples containing 0.2 to 10 ng of template DNA. Complete CTTV profiles were obtained on 62 of 63 human remains specimens recovered from four different crash sites. In parallel testing of these specimens, there was 100% concordance between the CTTV and FFTV multiplex systems at the VWF and TH01 loci. Simultaneous analysis of eight loci in a single gel was demonstrated using HEX-labeled CTTV and fluorescein-labeled FFFL multiplexes.
Multiplex |
Locus |
Precision (%) |
Range (bp) |
| CTTV* | CSF1PO | 99.95 |
0.68 |
| TPOX | 99.93 |
0.89 |
|
| THO1 | 99.94 |
0.47 |
|
| VWF | 99.87 |
0.76 |
|
| FFFL** | F13A01 | 99.93 |
0.87 |
| FESFPS | 99.93 |
0.69 |
|
| F13B | 99.95 |
0.40 |
|
| LPL | 99.89 |
0.56 |
* CTTV data represent 90 replicate samples and 12
electrophoretic runs.
** FFFL data represent 49 replicate samples and 8 electrophoretic runs.
Table 2. Reference Genotypes for Mixture
Reference |
VWF |
THO1 |
TPOX |
CSF1PO |
1 |
15, 16 |
9.3 |
8 |
10 |
2 |
15, 19 |
9.3 |
8, 10 |
11, 12 |
Figure 1. GeneScan™ Analysis of multiplex STR allelic ladders.
Electropherograms of STR Allelic ladders run on 373 DNA Sequencer and analyzed with GeneScan™ software. Values along x-axis represent base pair size of DNA fragments; values along y-axis represent intensity of fluorescence. (Top) Fluorescein-labeled CTTV multiplex with VWF, THO1, TPOX, and CSF1PO alleles; (Center) HEX-labeled CTTV multiplex with VWF, THO1, TPOX, and CSF1PO alleles; and (Bottom) Fluorescein-labeled FFFL multiplex with LPL, F13B, FESFPS, and F13A alleles.
CTTV profile using (Top) 10 ng of template DNA and (Bottom) 0.2 ng of template DNA in a standard PCR amplification reaction.
FFFL profile obtained using (top) 5 ng of template DNA and (bottom) 0.2 ng of template DNA in a standard PCR amplification process.
Figure 4. Mixture detection with the CTTV multiplex.
CTTV profile of DNA islolated from one crash site specimen showed major peaks and minor peaks (arrows) at the VWF, TPOX, and CSF1PO loci suggesting that the specimen consisted of commingled remains.
The CTTV profile of the victim (Center) was compared to the CTTV profiles of the mother (Top) and father (Bottom). The victim's profile contained one maternal and one paternal allele at each locus.
Figure 6. Human identification using eight-locus STR profile.
DNA was amplified, in separate reactions, with both the HEX-labeled CTTV multiplex and the fluorescein-labeled FFFL multiplex. The amplification reactions were combined and electrophoresed in a single gel lane. Comparative analysis of the eight locus STR profile of the victim (Top) and the eight-locus STR profile of the reference tissue (Bottom) determined revealed that these specimens originated from the same source.
Go to proceedings home page