- Email: [email protected]
A protocol for direct and rapid multiplex PCR amplification on forensically relevant samples
Forensic Science International: Genetics 6 (2012) 167–175
Contents lists available at ScienceDirect
Forensic Science International: Genetics journal homepage: www.elsevier.com/locate/fsig
A protocol for direct and rapid multiplex PCR amplification on forensically relevant samples Saskia Verheij b,1, Joyce Harteveld a,1, Titia Sijen a,* a b
Netherlands Forensic Institute, Laan van Ypenburg 6, The Hague 2497 GB, The Netherlands National crime squad, Hoofdstraat 54, 3972 LB Driebergen, The Netherlands
A R T I C L E I N F O
A B S T R A C T
Article history: Received 15 October 2010 Received in revised form 14 March 2011 Accepted 18 March 2011
Forensic DNA typing involves a multi-step workflow. Steps include DNA isolation, quantification, amplification of a set of short tandem repeat (STR) markers, separation of polymerase chain reaction (PCR) products by capillary electrophoresis (CE) and DNA profile analysis and interpretation. With that, the process takes around 10–12 h. For several scenarios it may be very valuable to speed up this process and obtain an interpretable DNA profile, suited to search a DNA database, within a few hours. For instance in cases of national security, abduction with danger of life, risk of repetition by a serial perpetrator or when custody time of suspects is limited. By a direct and rapid PCR approach we reduced the total DNA profiling time to 2–3 h after which genotyping information for the 10 STR markers plus the amelogenin (AMEL) marker present in the commercially available AmpF‘STR1 SGM PlusTM (SGM+) profiling kit is obtained. This reduction in time is achieved by using the following elements: (1) the inhibitor tolerant, highly processive Phusion1 Flash DNA polymerase; (2) a modified, non-adenylated allelic ladder; (3) the quick PIKO1 thermal cycler system with ultra-thin walled reaction tubes; (4) profile interpretation guidelines with an increased allele calling threshold, modified stutter ratios and marked low-level artefact peaks and (5) regulation of sample input by the use of mini-tapes that lift a limited amount of cell material from swabs or fabrics. The procedure is specifically effective for high level DNA, single source samples such as samples containing saliva, blood, semen and hair roots. Success rates, defined as a complete DNA profile, depend on stain type and surface. Due to the use of tape lifting as the sampling technique, the swab or fabric remains dry and intact and can be analyzed at a later stage using regular procedures. Validation experiments were performed which showed that the protocol effectively instructs researchers unfamiliar with the procedure. We have incorporated direct and rapid PCR in a ‘‘DNA-6 h’’ service that can assist police investigations by rapidly deriving DNA information from trace evidence left by a perpetrator, searching the STR profile against a DNA database and reporting the outcomes to police or prosecution. ß 2011 Elsevier Ireland Ltd. All rights reserved.
Key words: Forensic science Direct PCR Rapid PCR
1. Introduction Forensic DNA typing results are not only used as evidence in a court of law, but may also assist police investigations. Especially for this latter purpose, the time available for generating DNA information may be limited, for instance in cases of crimes by a serial perpetrator with a high risk of repetition, abduction with danger of life, time-restricted threats against persons or places, or when custody time of suspects is limited for legal reasons. Several approaches can reduce DNA profiling time: (1) the use of a rapid amplification protocol
* Corresponding author. Tel.: +31 708886888; fax: +31 708886555. E-mail address: t.sijen@nfi.minjus.nl (T. Sijen). 1 Both authors contributed equally. 1872-4973/$ – see front matter ß 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.fsigen.2011.03.014
[1–4]; (2) DNA profiling without prior DNA extraction, directly on a biological stain [5–7] or (3) DNA typing at the scene of crime or police station via automated, miniaturized devices [8– 10]. The first two approaches have the advantage that DNA typing is performed under the strict laboratory conditions befitting forensic casework, and that amplified products are separated at the nucleotide-level on validated genetic analyzers. The last approach has the advantage that no time is consumed by transport of samples from the scene of crime or police station to the forensic laboratories. Regarding rapid thermal cycling, excellent progress has been made by shortening various steps in the PCR protocol, using rapid PCR enzymes and fast thermal cyclers [1–4]. For these experiments pristine DNAs were used and profiles with strong signal intensity, good peak balance between loci and a limited amount of (adenylation) artefacts were obtained.
168
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
Direct PCR methods meet with several complications. First, the polymerase can be inhibited by components in a sample. Whole blood for instance contains several components such as heme, hemoglobin, lactoferrin and immunoglobinG that can inhibit the PCR. In addition, compounds from the surroundings such as soil, dirt or fabric dyes may pollute forensic samples. Second, the absence of a DNA quantification step complicates DNA profiling as the forensic short tandem repeat (STR) multiplexes allow a modest window of DNA input; below 200 pg of DNA the profile may become incomplete and above 1 or 2 ng of DNA input profiles tend to get overloaded while with even higher inputs amplification is restricted to smaller amplicons in the multiplex or even completely absent. Commercially available direct PCR methods are therefore specifically recommended for reference samples and come with clear instructions regarding the size of a (FTA) paper punch to be used as PCR input [7]. Promising direct PCR results were obtained for the analysis of touched items or fabrics for which overloading of the profiles is hardly an issue due to the limited amount of DNA present in these samples [5]. Integrated (microfluidic) systems for forensic DNA analysis at the scene of crime or police station should combine DNA extraction, amplification and analysis of amplified products. Liu et al. [8] have developed a portable microdevice that integrates PCR and capillary electrophoresis (CE). Besides, Bienvenue et al. [10] developed a microfluidic device that performs an on-chip extraction followed by multiplexed STR amplification. Amplified products are analyzed by normal bench-top CE instruments. Since the microdevice does not quantify the DNA extract, the system is more suited for samples with known origin like reference samples. The microfluidic system developed by Hopwood et al. [9] is also particularly suited for reference buccal samples. From a crude extract DNA is isolated using a preset amount of DNA-binding beads. Consequently, there is limited variation between different buccal scrapings in the amount of DNA transferred to the amplification chamber and balanced profiles can be obtained after analysis of the PCR products on the integrated micro capillary electrophoresis chip. In this study, we aimed to develop a simple, rapid and direct PCR protocol to genotype a wide variety of forensic crime scene stains at a forensic laboratory. Such a procedure has the potential to provide a service important for police investigations and tracing of suspects. In addition, it can be used to quickly screen a set of stains, for instance blood spatters, for the stains with a profile deviant from the major donor (e.g. the victim). We evaluated an engineered Pyrococcus-like proofreading DNA polymerase that gained robustness and enhanced processivity from its fusion to a double-stranded DNA-binding domain [11,12] (Phusion1 Flash polymerase (Phusion)) that tethers the polymerase to the duplex DNA. This assists in increased association constants and robust, inhibitor-tolerant, amplification reactions. The use of a proofreading polymerase omits the final soaking step, but the allelic ladder needs to be re-amplified to obtain nonadenylated alleles. To further increase the speed of the cycling process, the effect of a high-speed thermal cycling system was assayed. The quality of the DNA profiles was evaluated by studying base-line noise, heterozygote balance, inter-locus balance and amplification artefacts in order to develop interpretation guidelines that include detection threshold, stochastic threshold and locus-specific stutter ratios. Several approaches were tested to regulate input from forensically relevant samples under the condition that the integrity and quality of the biological stain was maintained and that the sample could be analyzed at a later stage by standard procedures. The aim was to develop a simple protocol that results in STR profiles in 2–3 h, with a good success rate also for first-time users of the protocol.
2. Materials and methods 2.1. DNA samples and fabrics For initial experiments pristine DNA 007 was used which is the positive control in the AmpF‘STR1 SGM PlusTM (SGM+) kit (Applied Biosystems (AB), Nieuwerkerk a/d IJssel, the Netherlands). Body fluids (blood, semen, saliva), hairs and items representing mock case work samples were obtained from volunteers with informed consent. Bone samples were derived from femurs from 5 year-old graves excavated with consent in Brabant (the Netherlands). For all donors the STR genotype was known from standard DNA typing and concordancy was checked in all DNA profiles. When body fluids were placed on fabrics, the fabrics were cleared beforehand from DNA contamination by irradiation with 254 nm UV light in a CL-1000 UV CrossLinker (UVP Inc., UK) at 0.9 J/cm2 during 45 min for each fabric side. Smooth-surface mock case work samples such as cups, bottles, watches, car door handles, gear-controls, steering wheels, plastic surfaces were generally sampled using cotton swabs moistened with water according to standard procedures. Porous-surface mock case work samples like clothing, shoe laces, cigarette buds, plasters, tissues, cotton were sampled using tape-lift stubs. In addition, chewing gums, hairs and bone powder were used as mock case work samples. 2.2. Stubs and tape lifts Tape-lift stubs are produced from double-sided sellotape (12 mm 33 m, Pritt, the Netherlands). During in-house validation experiments, sellotape was found to efficiently collect cell material and not interfere with DNA extraction. These experiments involved the sampling of neck regions of worn clothes of various volunteers with five types of tape and subjecting the tape-lifts to three DNA extraction methods. Sellotape outperformed the other tape types and combined well with our standard QIAamp DNA extraction procedure (Fleur Beemster and TS, personal communication, details available on request), and after extensive validation this method is the accredited standard in our laboratory to sample fabrics. Sellotape is placed either on a 12.7 mm2 sized aluminium short pin stub holder (G301F, Agar Scientific, UK) or on a 6.6 mm2 mini pin stub holder (16181, Bioconnect, the Netherlands). The normal sized tape-lift stubs were used in the experiments in the sections ‘tape lifting as an effective approach’ and ‘direct and rapid PCR on mock case work samples’ while the mini tape-lift stubs were used in the experiments in the sections ‘direct PCR by unfamiliar users’ and ‘comparative DNA research within 6 h. The containers for the 12.7 mm2 stub holders are SEM stubs (G3626, Agar Scientific, UK), the containers for 6.6 mm2 stub holders are prepared from mini-tip swabs (B22715, WA Products, UK) by removing the wooden swab-containing stick and cutting the plastic container to 3 cm length. DNA decontamination occurs similar to the decontamination of fabrics. When a set of tape-lift stubs is produced, an acceptance test is performed to check for complete decontamination. We regard tape-lift stubs as one-time use items, although the plastic parts can tolerate one additional round of UV light decontamination. 2.3. DNA polymerases and STR PCR set-up Phusion1 Flash polymerase (Phusion) (Bioke´, Leiden, the Netherlands) was tested for direct PCR. A typical reaction had a 20 mL reaction volume and consisted of 10 mL 2 enzymecontaining Master Mix, 4 mL of appropriate STR primer mix and
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
6 mL of water and DNA. Standard AmpliTaq Gold1 DNA Polymerase (AmpliTaq) (AB) amplification reactions were carried out in 25 mL reaction volumes. Set-up occurred as recommended by the manufacturer. The following autosomal STR kits were used: SGM+, AmpF‘STR1 Identifiler1 (Identifiler), AmpF‘STR1 SEfiler PlusTM (SEfiler + ) or AmpF‘STR1 NGMTM (NGM) (all AB).
3. Thermal cyclers and amplification protocols Amplification reactions were performed on a GeneAmp1 PCR System 9700 (9700 cycler) (AB) or a 24-PIKO1 Thermal Cycler (PIKO cycler) (Bioke´). In the PIKO cycler appropriate ultra-thin walled tubes were used (Bioke´). SGM+, Identifiler and SEfiler+ amplifications used 28 cycles, while NGM amplifications took 29 cycles. Standard reactions used amplification conditions as provided by the manufacturer. Phusion amplifications on the 9700 cycler consisted of 1 min 98 8C hotstart, 28 cycles of 5 s 98 8C denaturation, 20 s 59 8C primer annealing and 20 s 72 8C elongation with a final soak of 30 min 60 8C. Unless stated otherwise, rapid amplification reactions on the PIKO cycler employed an initial denaturation at 98 8C for 5 min, followed by the appropriate number of cycles at 98 8C for at least 5 s, 59 8C for at least 30 s, 72 8C for 10 s with a final extension at 72 8C for 1 min. 3.1. STR analysis and allelic ladder preparation Amplified products were analyzed by adding 1 mL to a mixture of 8.7 mL Hi-Di formamide and 0.3 mL internal size standard (all AB, size standard for SGM+ is ROX GS500, that for SEfiler+ is LIZ GS600 and the one for Identifiler and NGM is LIZ GS500). When a reduced amount of amplified product is analyzed, the PCR mixture is diluted 1:100 (in water) and 1 mL of this dilution is added to 8 mL Hi-Di formamide and 1 mL of 1:100 diluted (in Hi-Di formamide) appropriate internal size standard. Samples were electrokinetically injected at 3 kV for 15 s and separated on a 3130xl Genetic Analyzer using POP-4 polymer and a 36 cm capillary array (all AB). After data collection, genotyping was performed using GeneMapperID-X or GeneMapper v3.2 (AB) using the bins and panels provided by the manufacturer. For the analysis of non-adenylated PCR products, allelic ladders were re-amplified as described in [13]. Appropriate allelic ladders were diluted 1:1000 in water, and 1 mL of this dilution was amplified during 15 cycles in 20 mL Phusion reaction. The mixture of buffer, enzyme and primers was prepared under pre-PCR laboratory conditions after which the template allelic ladder was added in a post-PCR room. Since re-amplified allelic ladders can be stored up to 2 years at 20 8C or 2 months at 4 8C, a large (e.g. 200 mL) batch of ladder can be prepared that – after quality check (all alleles present with sufficient height) – can serve many CE injections.
4. Profile interpretation When determining the percentage of detected alleles or the average peak heights, homozygous alleles are counted as two alleles. Peak height ratios (PHR) at heterozygous loci are determined by dividing the lowest peak height by the highest peak height. Inter-locus balance is calculated as the total rfu value observed at a locus divided by the total rfu value of the profile (ideally each locus would show 0.091 of the total rfu’s of an 11locus profile).
169
5. Results and discussion 5.1. STR profiling directly from blood and other forensically relevant body fluids AmpliTaq, the standard DNA polymerase in commercial STR typing kits, is known to be sensitive for inhibition by whole blood components while Phusion is reported by its manufacturer to tolerate up to 25% blood [12]. To explore the possibilities of Phusion for forensic purposes, amplification reactions were prepared from Phusion mastermix, SGM+ primermix and 0.4 ng pristine DNA. A standard 9700 cycler was used. Full profiles were obtained and negative amplifications showed no allele calls. The Phusion PCR products were one nucleotide shorter than the AmpliTaq products due to the 30 to 50 exonuclease activity of the enzyme. Successful allele calling was achieved by using an allelic ladder re-amplified with Phusion. The average peak heights, peak height ratios and inter-locus balance for the Phusion amplifications were similar as those in the standard AmpliTaq amplifications (Table 1). Next, various amounts (200 nL, 100 nL and 20 nL) of whole blood (resulting in 1%, 0.5% and 0.1% whole blood in the final PCR respectively) were added to the amplification mixtures that in addition contained 0.4 ng of pristine DNA. With AmpliTaq, no amplification was observed while with Phusion two-donor profiles were obtained that corresponded to DNA007 and the blood donor (results not shown). Clearly, Phusion can amplify DNA templates in the presence of whole blood. For the inputs of 200 nL and 100 nL whole blood the profiles were over-amplified, which implies that the majority of the blood DNA is released (100 nL blood contains approximately 2–4 ng DNA [14], which would indeed lead to an overloaded STR profile). Due to their structure, sperm cells may need a more stringent protocol (longer 98 8C hotstart period) to release DNA [15]. Two semen samples were diluted so that approximately 100 cells were added to the amplification reactions. These two semen samples and 3 nL whole blood were directly amplified in 4-fold with different hotstart periods (Supplementary Table S1). Overall optimal results were observed using a 5 min hotstart period, and this protocol was used when examining the ability of Phusion to perform direct PCR on various inputs of blood, saliva and semen. Phusion enabled direct PCR amplification of these body fluids as full SGM+ profiles were obtained (Table 2). AmpliTaq only showed efficient direct amplification for saliva but when inputs increased over 0.5 mL, A artefact peaks were abundant. For Phusion, overamplification was evident with inputs of 50 nL whole blood, 2 mL saliva or 1000 sperm cells, which is consistent with the amount of DNA generally present in these inputs [14]. In Phusion reactions, over-amplification is not accompanied by the occurrence of A artefact peaks as Phusion produces blunt-ended PCR products due to its proofreading ability. 5.2. Rapid, direct STR typing and guidelines for reliable interpretation of profiles To reduce the time needed for direct amplification, we tested the PIKO cycler system that is recommended by the manufacturer
Table 1 Characteristics of SGM+ profiling of 400 pg pristine DNA using Phusion or AmpliTaq.
Phusion AmpliTaq a
Averagea peak height
Average PHR
Inter-locus balance
875 576 rfu 872 488 rfu
0.83 0.13 0.83 0.12
0.042 (FGA)–0.133 (vWA) 0.048 (FGA)–0.120 (vWA)
Average of 9 independent amplifications.
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
170
Table 2 Performance of Phusion for direct amplification of various body fluids. Body fluid
Input
Blood
8 nL 16 nL 50 nL 0.1 mL 0.5 mL 2 mL 6 mL 100 cells 1000 cells
Saliva
Semen a b c d
Phusion
AmpliTaq
% Detecteda
Average rfub
% Detected
Average rfu
100% 100% 100% 100% 100% 100% 99 2% 100% 100%
1099 176 2094 83 5852 621c 1046 163 4217 491 6438 68c 5766 527c 1889 585 4376 1244c
40 36% 7 8% 1 2% 100% 100%d 100%c,d 87 13%d 93 10% 39 5%
114 36 54 5 58 5 613 153 1498 236 3747 905 1133 543 370 105 148 35
Average of 4 concurrent amplifications. Average of detected alleles. Over-amplification occurred. Showing A artefacts peaks.
to further speed up the amplification process when using Phusion [16]; time periods of 1 s for denaturation, 5 s for primer annealing and 15 s/kb for elongation may suffice. The forensic SGM+ STR kit is an 11-loci multiplex. To enable efficient annealing of all primers, several annealing periods were tested using 0.6 ng pristine DNA. Optimal results were obtained using the 30 s annealing time (Supplementary Table S2). The total run time for this PCR is 47 min (Supplementary Table S3). The heterozygote and inter-locus balances for the Phusion amplifications are comparable to those found for the standard AmpliTaq reactions (Supplementary Table S2, Table 1). No artefacts are observed in amplification negatives (results not shown). The 47 min PIKO cycler protocol was used to amplify eight amounts (ranging from 20 nL to 0.16 nL) of stored or freshly collected blood of eight different donors. In addition, 20 amplification negatives were performed. In the 64 blood amplifications, the base line noise was increased which was not the case in the amplification negatives. We assume this raised noise is due to components in the body fluids affecting the PCR. When the detection threshold was increased from 50 rfu to 80 rfu, the number of baseline artefact peaks that were called was greatly reduced except for the 20 nL and 10 nL inputs that showed high noise levels and over-amplification. In the blood profiles, two types of artefact peaks were observed that were not due to increased base line noise: (1) peaks at 1 repeat unit stutter position, and (2) low level artefacts peaks at specific positions. Stutters at +1 repeat unit were seldom observed. For each locus, the average percentage of the 1 repeat unit stutter peaks was calculated together with the standard deviation. By taking the mean plus twice the standard deviation (two sigma rule) locus-specific 1 repeat unit stutter ratios were determined for direct and rapid PCR (Table 3, Supplementary Fig. S1). For eight of the ten STR loci these 1 repeat unit stutter ratios are higher than the ratios provided by the manufacturer for standard SGM+ typing; for two loci the ratios are Table 3 Locus-specific 1 repeat unit stutter percentages for direct and rapid amplification. Marker
Average %
Stdev
# Stutters
m + 2sa (%)
AB %b
D3S1358 vWA D16S539 D2S1338 D8S1179 D21S11 D18S51 D19S433 TH01 FGA
9.3 11.5 10.4 13.7 11.8 6.4 11.8 7.6 8.8 14.9
2.9 2.0 3.0 4.1 3.3 2.2 3.6 2.5 3.8 2.9
56 63 66 68 60 62 78 61 60 60
15.1 15.5 16.4 22.1 18.5 10.8 19.0 12.6 16.3 20.7
11.0 11.0 13.0 15.0 12.0 13.0 16.0 17.0 6.0 11.0
a b
Mean plus twice standard deviation. Stutter percentage provided by manufacturer (AB).
lower (Table 3, Supplementary Fig. S1). To harbour all stutter peaks, the 1 repeat unit stutter distance in the analysis settings was slightly increased (from 3.25–4.75 to 3.0–5.0nt). Common, low level (but regularly above detection threshold) artefacts peaks were seen for the following positions: D3S1358 allele positions 11 and 14, vWA allele positions 10, 11 and 14, D16S539 allele position 7, D8S1170 allele position 7, D21S11 allele position 33, THO1 allele positions 6 and 12. The use of a different size standard (Promega’s ILS600 [17] instead of AB’s ROX-500), the application of a higher annealing temperature or the use of DTR column purification did not prevent the occurrence of these peaks (results not shown). In single donor profiles with rfu heights over 500 rfu, artefact peaks can be readily distinguished from true alleles as the artefact peaks stay low level (Fig. 1). When these analysis settings (detection threshold 80 rfu, modified 1 repeat unit stutter distance and ratios, marked artefact peaks) were applied to the 64 whole blood amplifications, full STR profiles were observed down to 1.25 nL blood (Table 4). The profiles showed adequate intra- and inter-locus balance (Supplementary Table S4). Due to the fact that various donors were used and that the storage times of the blood samples varied from 2 years to freshly taken, quite some variation in profiling results and peak heights were obtained (Table 4). Allele drop-out started to occur when 2.5 nL blood was analyzed. Drop-out was not found at specific loci; although larger amplicons are more prone to allele drop-out as with standard STR typing. The height of single peaks at heterozygous loci was determined in order to obtain insight in the stochastic threshold, which warns for the possible occurrence of an allelic drop-out. In total 69 allelic drop-outs were observed. The heights of the concurrent single peaks varied from 80 to 762 rfu (Supplementary Fig. S2). To flag 95.7% of these single allele calls as possibly heterozygous, a stochastic threshold of 375 rfu is needed. This is a rather high value and accordingly, 30 of the 142 homozygous peaks also receive this flag. Alternative approaches would be to evaluate only profiles above for instance 400 rfu, and/ or perform multiple independent amplifications for replicate analysis. We have taken this advice into account when developing a protocol to analyze case work samples, as described onwards. 5.3. Tape lifting as an effective approach to regulate PCR input from swabs and various fabrics Evidently, regulation of PCR input is essential for successful direct and rapid amplification; with too much input profiles become over-amplified and show increased base line noise and with too little input incomplete profiles are obtained. Samples at forensic laboratories mainly consist of textiles or swabs obtained from sampling various evidentiary items. We set out to develop a sampling technique that: (1) results in full STR profiles upon direct
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
171
Fig. 1. Low level artefact peaks (marked by arrows) occurring at specific positions in rapid and direct amplifications; (A) artefact at vWA position 14, (B) artefact at D21S11 position 33, (C) artefact at THO1 position 6 (coinciding with 1 repeat unit stutter peak). The upper number is the allele call, the lower number corresponds to peak height (rfu).
and rapid PCR, (2) is practical and applicable to both swabs and fabrics, (3) comes with minimal handling and contamination risks, and (4) leaves the remainder of the swab or fabric intact (dry) for standard analysis at a later stage. Blood (3 mL) was applied on cotton swabs, and five methods were tested: (1) twist swab twice or (2) 20 times in PCR tube (to enable this approach mini-tip swabs were used), (3) take out a fibre from swab, (4) cut the tip of the swab and (5) stub the swab with a mini tape-lift. We tested these five methods with a replicate number between 4 and 16, and found that only two methods resulted in full profiles for all direct amplifications namely the fibre method and the tape lifting approach (results not shown). The tape lifting method proved much more practical: easy to perform with a minimal risk of contamination (the tapes are treated extensively with UV light in a crosslinker to remove DNA contamination) and the swab remains completely intact for further use. Tape lifting (Fig. 2) can either be performed using a tape on a 12.7 mm2 stub holder from which an area of 5 mm2 is cut by a scalpel knife or by using a smaller tape placed on a 6.6 mm2 mini stub holder. The tape lifting method was also tested for bloodstains (3 mL undiluted and 3 mL 1:10 diluted blood) on various types of fabrics (Table 5), and proved successful. The fabrics have different characteristics: some hardly adsorb the blood (nylon), others easily release blood-containing fibres (wool) and some take up the blood tightly and hardly release fibres (course cotton). Therefore, the tape lift sampling number (times the stain is touched with the tape-lift stub) varied for the various fabrics depending on the release of blood and fibres as determined by visual inspection (Table 5). Only for denim no genotyping data were obtained which is probably due to PCR inhibition from denim components. 5.4. Direct and rapid PCR on mock casework samples 149 mock casework samples, expected to contain single source cell material of known donors, were collected to study the success
rate of direct and rapid PCR. The tape lift procedure was used for all stains residing on swabs. For items sampled by tape-lift stubs, 5 mm2 areas were excised from the tape. Chewing gum, hairs and bone powder were added directly to the PCR mixture. Full STR profiles concordant with the donor of the sample were obtained for 60% of the samples representing classical forensic trace evidence types (Table 6). All blood samples and half of the saliva samples yielded full profiles. Hair roots give 65% of full profiles when pulled hairs are used, but the profiles appear overloaded and are best analyzed after injecting less (0.01 mL) PCR product. When hairs are analyzed no sample remains as the hair is fully used during direct and rapid PCR. When items containing contact DNA and bone samples are included, the overall success rate decreases (46%) (Table 6). The bone samples gave full profiles in standard DNA analysis, but no genotyping results using direct PCR. This difference is most likely due to the use of too little bone material for the direct PCR; approximately 10 mL of bone powder was used for DNA isolation of which 1/20th was used as PCR input (this corresponds to 500 mL bone powder), while for direct PCR only 10 mL bone powder was added. For 31 samples the remainders of the swabs or fabrics were analyzed using the regular procedure of DNA isolation, quantification and profiling. The 31 samples were of three categories namely samples giving a full, a partial or no DNA profile using rapid and direct PCR. A large part of the samples gave a similar result using the regular genotyping procedure (Table 7). For samples showing no profile with the direct procedure, standard genotyping can be successful and result in a partial of even complete profile (Table 7); apparently insufficient cell material was collected for the direct procedure. For only one sample better results were found with direct and rapid PCR (full profile) than with the standard procedure (no profile). This sample comprised a tape lift from a car door handle, and it is possible that the part of the tape lift containing cell material was consumed in the direct PCR. In the profiles obtained
Table 4 Rapid and direct amplification of various inputs of whole blood.
Full profile Average rfuc PHR # drop-ins a b c d
20 nL
10 nL
5 nL
2.5 nL
1.25 nL
0.63 nL
0.32 nL
0.16 nL
8/8 4862 621 0.89 0.10 15
8/8 3989 1229 0.82 0.13 7
8/8 2658 1260 0.73 0.18 7
6/8a 1143 654 0.71 0.19 5
5/8b 517 281 0.67 0.19 1
0/8 241 117 0.57 0.22 0
0/8 150 34 ndd 0
0/8 180 65 nd 0
Only allelic drop-outs were seen, 1 per profile, the two profiles both had an average height of approximately 350 rfu. In one of the incomplete profiles a locus drop-out occurred, in all 3 incomplete profiles an allelic drop-out was seen. Homozygous alleles were calculated as 2 alleles. Not determined, too few heterozygous loci detected.
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
172
Fig. 2. Tape lifting to regulate input for direct and rapid PCR. A mini tape-lift stub (A) is used to lift cell material from a blood-containing swab (B and C) after which the tape is removed from the stub holder (D and E) and placed in a PCR tube (F).
Table 5 Direct and rapid PCR of blood stains using tape-lift stubs.
Cotton Undiluted blood # Times with tape 1:10 blood # Times with tape a b
Cotton swab
Denim
Linen
Viscose
Nylon
Fleece
Wool
Fine cotton
Coarse
16/16a 2 ndb nd
0/4 20 0/4 20
4/4 5 3/4 10
4/4 1 4/4 20
3/4 1 4/4 1
4/4 1 4/4 1
4/4 1 2/4 1
4/4 5 2/4 10
4/4 10 1/4 30
Number of full STR profiles of total number of amplifications. Not determined.
Table 6 Direct and rapid PCR of mock casework samples. Type
Subtype
Full profile
Partial profile
No profile
Blood
Plaster or tissue Smooth surface Cup or bottle Chewing guma Cigarette budb Tissue, spit or buccal swab Smooth surface Pulled Fallen
7/7 6/6 11/26 5/9 5/10c 9/12 3/4 11/17 0/4
0/7 0/6 10/26 3/9 2/10 2/12 1/4 4/17 1/4
0/7 0/6 5/26 1/9 3/10 1/12 0/4 2/17d 3/4
Total
Classical trace evidence
57/95 (60%)
23/95 (24%)
15/95 (16%)
Contact DNA
Bone
Clothes Personal itemse Used itemsf Powdered
4/16 4/14 3/18 0/6g
5/16 5/14 1/18 0/6
7/16 5/14 14/18 6/6
Total
Non-classical trace evidence
11/54 (20%)
11/54 (20%)
32/54 (60%)
Overall total
All types of trace evidence
68/149 (46%)
34/149 (23%)
47/149 (31%)
Saliva
Semen Hair
a b c d e f g
PCR input not by tape lift but by cutting a small piece. Filter side is stubbed. Mixed profiles were obtained for 3 cigarettes. Both from one specific donor. Watches, shoe laces. Door handles, car items. Powder of 2 bones was used.
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175 Table 7 Genotyping results using both direct and rapid PCR and standard profiling on the same samples (n = 31). Results direct
Results regular Full
Partial
No profile
Full (11) Partial (9) No profile (11)
10 0 2
0 9 5
1 0 4
Total (31)
12
14
5
by regular genotyping, no DNA contamination due to the tape lift procedure was detected. 5.5. Rapid and direct PCR executed by unfamiliar users Based on the above described experiments, a set of instructions for direct and rapid PCR was prepared (Table 8) that describes for various sample types how many reactions should be performed, how the tape lifting or the sample cutting should be executed, and how much of the PCR mixture should be analyzed by CE. With direct and rapid PCR, samples can get over-amplified. Since the Phusion polymerase has proofreading ability, amplified products do not carry adenine overhangs at their 30 ends. Therefore no A artefacts occur and analyzing 0.01 mL of the amplified products is very effective when over-amplification occurred. The set of instructions results in multiple DNA profiles from a sample; two to three tape lifts are taken (except for hair roots) and two amounts of amplified products are analyzed (except for items containing contact DNA). Therefore for most samples several interpretable (not over-loaded and not incomplete) profiles are obtained, often from independent amplifications (which may be useful to accommodate the relative high stochastic threshold for Phusion profiles). We set out to test whether the set of instructions (Table 8) could effectively be carried out by first-time protocol users. We generated four identical cotton patches with 6 bloodstains (2 or 5 mL spot size). Four stains consisted of undiluted blood and two stains contained 5-times diluted blood. Two different blood donors were used and one of the stains of undiluted blood consisted of a 1:1 mixture of the 2 donors. Three researchers unfamiliar with the procedure and one experienced researcher performed direct and rapid PCR by following the set of instructions. All researchers obtained full profiles for all stains (a representative electropherogram is shown in Fig. 3). For the diluted bloodstains all researchers generated full profiles with standard CE injection, but for undiluted stains often CE with reduced input (1:100 dilution) was needed, depending on spot size and researcher (apparently tape lifting is performed slightly different by the various researchers). All researchers identified the donors of the blood stains correctly including the mixed stain. This shows that direct and rapid PCR has an additional application: when many bloodstains are present on an evidentiary item a screening can be performed to identify stains that do not match the DNA profile of the major donor of the bloodstains e.g. the victim. These stains can be genotyped in a next analysis round using the standard DNA typing procedure.
173
5.6. Comparative DNA research within 6 h In four experiments we challenged the time for comparative DNA research when DNA profiling is performed by direct and rapid PCR. Nineteen mock casework samples were processed in total. Time was calculated from intake of the samples at the Netherlands Forensic Institute to reporting of the results by phone or email. At intake, samples were labelled and registered according to our standard procedures. The samples were analyzed by direct and rapid PCR after which DNA results were doubled-checked by two analysts. Next, a DNA database search was performed against persons present in the Dutch DNA Database and a reporting officer summarized the results in a report. After checking of this report by a second reporting officer, it was send by email. In all four experiments the total time for this process was below 6 h. To achieve a high success rate for this fast comparative DNA research, the trace evidence needs to comply to the following aspects: (1) contain blood, saliva or semen as these body fluids contain relatively high levels of DNA; (2) correspond to cell material of a single donor as mixture de-convolution requires too much time and rapid and direct profiles show increased (base-line) noise and low level artefact peaks and (3) process a maximum number of three items simultaneously. We have decided that single donor profiles obtained by this DNA-6 h process can be used for DNA database searching but not for permanent DNA database entry. Because of the increased noise in the rapid and direct profiles, the DNA-6 h service is always accompanied with priority research in which the remainder of the trace evidence (swab or fabric remaining after tape lift procedure) is analyzed using standard procedures that include presumptive testing for stain type, mixture analysis (if applicable) and a comprehensive report. Profiles obtained by this second (standard priority) research can be used for permanent database entry and searching international databases under the Pru¨m treaty. Due to this set-up DNA-6 h comes with additional costs and extra hands-on time (costs include the tape-lift stub, SGM+ primer mix, Phusion mastermix, PIKO cycler tubes and CE run costs). Therefore DNA-6 h is especially appropriate for cases that may benefit substantially from extremely fast analysis (national security, danger of life, serial perpetrator). This service is offered to the Dutch police since 7th October 2010. In the first three months four cases have been addressed; all cases involved blood samples and for all samples single donor full profiles were obtained within 6 h. Importantly, the same genotyping results were obtained with the DNA-6 h protocol and during the subsequent standard analysis. 5.7. Rapid and direct PCR and other STR kits Currently, the SGM+ kit is the standard genotyping kit in our laboratory. Other forensic laboratories may utilize a different standard kit and we tested whether direct and rapid PCR would be compatible with other STR kits. Six nanoliters of whole blood were amplified in 4-fold with the SGM+, SEfiler+, Identifiler and NGM primer sets. Since these STR kits amplify 12–16 loci, both the standard annealing time of 30 s and an extended time of 50 s were tested. Using a 80 rfu detection threshold (as with SGM+), full
Table 8 Instructions for tape lifting for direct and rapid PCR. Stain
Tape-lift: 3 mm2, gentle pressure, max 1 s per touching
CE injection
Blood undiluted Blood diluted (swab) Saliva or semen spot Saliva drinking or cigarette Chewing gum Contact traces Hair
Duplicate: touch 1–5 times, a faint light red shade Duplicate: touch 3–10 times, a faint light red shade Triplicate: touch 1, 5 and 10 times Triplicate: touch 10, 25 and 50 times Duplicate: cut 1 mm2 and 0.5 mm2 piece Duplicate: until tape is saturated Single (hair is fully used): cut hair root
1 1 1 1 1 1 1
and and and and and
1:100 1:100 1:100 1:100 1:100
and 1:100
174
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
Fig. 3. Typical SGM+ DNA profile obtained after direct and rapid PCR of a bloodstain on fabric. CE was performed using 0.01 mL PCR mixture. The AF14 label in the vWA locus refers to a low level artefact peak, commonly found at this position. The upper number is the allele call, the lower number corresponds to peak height (rfu). Note that some bleed-through is observed mainly in the yellow channel originating from peaks in the green channel; this should not be regarded as increased base line noise.
SEfiler+ and Identifiler profiles were obtained already with an annealing time of 30 s (Supplementary Fig. S3AB), showing that also 12 or 16 loci (SEfiler+ or Identifiler respectively) can be amplified simultaneously via this rapid and direct PCR protocol. The NGM profiles were incomplete, and the loci D21S11 and D18S51 consistently dropped out or were very low (Supplementary Fig. S3C). Apparently the direct and rapid PCR conditions (annealing time, buffer composition) disfavour amplification of these loci when using the NGM primer mix. When a combined annealing elongation time of 3 min at 59 8C was used (which resulted in a 2 h in stead of a 47 min PCR protocol), these loci were amplified (Supplementary Fig. S3D). Evidently, to enable direct and rapid PCR via the NGM kit, further protocol optimization is required. Artefact peaks did also occur in the Identifiler, SEfiler and NGM profiles, and some of these artefact peaks were at the same positions as in the SGM+ profiles (e.g. allele 14 vWA, allele 33 D21S11). This is probably due to the use of primers with identical primer binding site positions in the various kits. 6. Concluding remarks A procedure for direct and rapid PCR was developed that builds on several components; direct PCR is facilitated by Phusion1 Flash
DNA polymerase, a highly processive, robust, proofreading heat stable polymerase. Rapid cycling is aided by the use of the PIKO cycler system. Regulated sample input is achieved by collecting biological material from body fluid-containing cotton swabs or fabrics on small tape lifts and adding these to the amplification reactions. A set of instructions was prepared to achieve a PCR input of cellular material for different stains and different substrates that results in an interpretable PCR profile. This set of instructions could be successfully used by researchers unfamiliar with the procedure. Analysis of 149 mock casework samples showed that success rates mainly depend on the type of stain. Blood stains are the most promising, probably because blood is a visible stain with a relatively high number of DNAcontaining cells. DNA profiles obtained from direct and rapid PCR need to be analyzed using a modified allelic ladder as the PCR products are non-adenylated. The profiles show increased base-line noise, increased stutters at some loci and specific low level artefact peaks. To minimize aberrant allele calling, modified profile interpretation guidelines are required that accommodate these noise effects. Therefore, direct and rapid PCR profiles appear less suited to deduce mixed profiles (although the method can detect presence of multiple donors). Trace evidence consisting of blood, saliva or
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
semen from one donor, suits this method best. We have incorporated rapid and direct PCR in a ‘DNA-6 h’ service that aims to assist police investigations and tracing of suspects in cases for which every minute counts. DNA-6 h rapidly derives DNA information from a single-donor biological stain that is left by a perpetrator which can be used to search a DNA database. In addition, direct and rapid PCR can be used to screen a set of stains on an evidentiary item for the stains that do not relate to for instance the victim but may belong to the perpetrator. In preliminary experiments direct and rapid PCR was found to be compatible with other STR kits, indicating that the method described here may serve as a framework protocol that can be adapted and optimized within different forensic laboratories. Role of funding This study was supported by a grant from the Netherlands Genomics Initiative/Netherlands Organization for Scientific Research (NWO) within the framework of the Forensic Genomics Consortium Netherlands. Acknowledgements We thank Corina Benschop, Bas de Jong and Ankie van Gorp for useful discussion, Alexander Lindenbergh, Bas de Jong, Lise De Strooper, Bryan Bhoelai for experimental assistance and Ankie van Gorp and Charissa van Kooten for critically reading the manuscript.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.fsigen.2011.03.014.
175
References [1] P.M. Vallone, C.R. Hill, J.M. Butler, Demonstration of rapid multiplex PCR amplification involving 16 genetic loci, Forensic Sci. Int. Genet. 3 (2008) 42–45. [2] P.M. Vallone, C.R. Hill, D. Podini, J.M. Butler, Rapid amplification of commercial STR typing kits, Forensic Sci. Int. Genet. Suppl. Series 2 (2009) 111–112. [3] D.Y. Wang, C.-W. Chang, L.K. Hennessy, Rapid STR analysis of single source DNA samples in 2 h, Forensic Sci. Int. Genet. Suppl. Series 2 (2009) 115–116. [4] H. Giese, R. Lam, R. Selden, E. Tan, Fast multiplexed polymerase chain reaction for conventional and microfluidic short tandem repeat analysis, J. Forensic Sci. 54 (2009) 1287–1296. [5] A. Linacre, V. Pekarek, Y.C. Swaran, S.S. Tobe, Generation of DNA profiles from fabrics without DNA extraction, Forensic Sci. Int. Genet. 4 (2010) 137–141. [6] A. Barbaro, P. Cormaci, A. Teatino, A. Barbaro, Use of ‘‘AnyDirect PCR buffer’’ for PCR amplification of washed bloodstains: a stain report, Forensic Sci. Int. Genet. Suppl. Series 1 (2008) 11–12. [7] D.Y. Wang, C.-W. Chan, N.J. Oldroyd, L.K. Hennessy, Direct amplification of STRs from blood or buccal cell samples, Forensic Sci. Int. Genet. Suppl. Series 2 (2009) 113–114. [8] P. Liu, S.H.I. Yeung, K.A. Crenshew, C.A. Crouse, J.R. Scherer, R.A. Mathies, Real-time forensic DNA analysis at a crime scene using a portable microchip analyzer, Forensic Sci. Int. Genet. 2 (2008) 301–309. [9] A.J. Hopwood, C. Hurth, J. Yan, Z. Cai, N. Moran, J.G. Lee-Edghill, A. Nordquist, R. Lenigk, M.D. Estes, J.P. Haley, C.R. McAlister, X. Chen, C. Brooks, S. Smith, K. Elliott, P. Koumi, F. Zenhausern, G.A.J. Tully, Integrated microfluidic system for rapid forensic DNA analysis: sample collection to DNA profile, Anal. Chem. 82 (2010) 6991–6999. [10] J.M. Bienvenue, L.A. Legendre, J.P. Ferrance, J.P. Landers, An integrated microfluidic device for DNA purification and PCR amplification of STR fragments, Forensic Sci. Int. Genet. 4 (2010) 178–186. [11] Y. Wang, D.E. Prosen, L. Mei, J.C. Sullivan, M. Finney, P. Vander Horn, A novel strategy to engineer DNA polymerases for enhanced processivity and improved performance in vitro, Nucleic Acids Res. 32 (2004) 1197–1207. [12] P. Yang, C. Andre´, Efficient PCR from whole blood using Finnzymes high performance PCR solution, Nat. Methods Appl. Notes (2007) an20–an21. [13] J.M. Butler, Y. Shen, B.R. McCord, The development of reduced size STR amplicons as tools for analysis of degraded DNA, J. Forensic Sci. 48 (2003) 1054–1064. [14] http://www.oglethorpe.edu/faculty/k_aufderheide/Forensic_Science/Web_D ocuments/Catalin_Andrei_Mitrasca.pdf. [15] A.M. Hossain, B. Rizk, A. Behzadadian, I.H. Thorneycroft, Modified guanidium thiocyanate method for human sperm DNA isolation, Mol. Hum. Reprod. 3 (1997) 953–956. [16] http://www.finnzymes.com/pdf/f548sl_phusion_flash_highfidelity_pcr_maste rmix_datasheet_1_3_low.pdf. [17] R.S. McLaren, M.G. Ensenberger, B. Budowle, D. Rabbach, P.M. Fulmer, C.J. Sprecher, J. Bessetti, T.M. Sundquist, D.R. Storts, Post-injection hybridization of complementary DNA strands on capillary electrophoresis platforms: a novel solution for dsDNA artifacts, Forensic Sci. Int. 2 (2008) 257–273.
Contents lists available at ScienceDirect
Forensic Science International: Genetics journal homepage: www.elsevier.com/locate/fsig
A protocol for direct and rapid multiplex PCR amplification on forensically relevant samples Saskia Verheij b,1, Joyce Harteveld a,1, Titia Sijen a,* a b
Netherlands Forensic Institute, Laan van Ypenburg 6, The Hague 2497 GB, The Netherlands National crime squad, Hoofdstraat 54, 3972 LB Driebergen, The Netherlands
A R T I C L E I N F O
A B S T R A C T
Article history: Received 15 October 2010 Received in revised form 14 March 2011 Accepted 18 March 2011
Forensic DNA typing involves a multi-step workflow. Steps include DNA isolation, quantification, amplification of a set of short tandem repeat (STR) markers, separation of polymerase chain reaction (PCR) products by capillary electrophoresis (CE) and DNA profile analysis and interpretation. With that, the process takes around 10–12 h. For several scenarios it may be very valuable to speed up this process and obtain an interpretable DNA profile, suited to search a DNA database, within a few hours. For instance in cases of national security, abduction with danger of life, risk of repetition by a serial perpetrator or when custody time of suspects is limited. By a direct and rapid PCR approach we reduced the total DNA profiling time to 2–3 h after which genotyping information for the 10 STR markers plus the amelogenin (AMEL) marker present in the commercially available AmpF‘STR1 SGM PlusTM (SGM+) profiling kit is obtained. This reduction in time is achieved by using the following elements: (1) the inhibitor tolerant, highly processive Phusion1 Flash DNA polymerase; (2) a modified, non-adenylated allelic ladder; (3) the quick PIKO1 thermal cycler system with ultra-thin walled reaction tubes; (4) profile interpretation guidelines with an increased allele calling threshold, modified stutter ratios and marked low-level artefact peaks and (5) regulation of sample input by the use of mini-tapes that lift a limited amount of cell material from swabs or fabrics. The procedure is specifically effective for high level DNA, single source samples such as samples containing saliva, blood, semen and hair roots. Success rates, defined as a complete DNA profile, depend on stain type and surface. Due to the use of tape lifting as the sampling technique, the swab or fabric remains dry and intact and can be analyzed at a later stage using regular procedures. Validation experiments were performed which showed that the protocol effectively instructs researchers unfamiliar with the procedure. We have incorporated direct and rapid PCR in a ‘‘DNA-6 h’’ service that can assist police investigations by rapidly deriving DNA information from trace evidence left by a perpetrator, searching the STR profile against a DNA database and reporting the outcomes to police or prosecution. ß 2011 Elsevier Ireland Ltd. All rights reserved.
Key words: Forensic science Direct PCR Rapid PCR
1. Introduction Forensic DNA typing results are not only used as evidence in a court of law, but may also assist police investigations. Especially for this latter purpose, the time available for generating DNA information may be limited, for instance in cases of crimes by a serial perpetrator with a high risk of repetition, abduction with danger of life, time-restricted threats against persons or places, or when custody time of suspects is limited for legal reasons. Several approaches can reduce DNA profiling time: (1) the use of a rapid amplification protocol
* Corresponding author. Tel.: +31 708886888; fax: +31 708886555. E-mail address: t.sijen@nfi.minjus.nl (T. Sijen). 1 Both authors contributed equally. 1872-4973/$ – see front matter ß 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.fsigen.2011.03.014
[1–4]; (2) DNA profiling without prior DNA extraction, directly on a biological stain [5–7] or (3) DNA typing at the scene of crime or police station via automated, miniaturized devices [8– 10]. The first two approaches have the advantage that DNA typing is performed under the strict laboratory conditions befitting forensic casework, and that amplified products are separated at the nucleotide-level on validated genetic analyzers. The last approach has the advantage that no time is consumed by transport of samples from the scene of crime or police station to the forensic laboratories. Regarding rapid thermal cycling, excellent progress has been made by shortening various steps in the PCR protocol, using rapid PCR enzymes and fast thermal cyclers [1–4]. For these experiments pristine DNAs were used and profiles with strong signal intensity, good peak balance between loci and a limited amount of (adenylation) artefacts were obtained.
168
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
Direct PCR methods meet with several complications. First, the polymerase can be inhibited by components in a sample. Whole blood for instance contains several components such as heme, hemoglobin, lactoferrin and immunoglobinG that can inhibit the PCR. In addition, compounds from the surroundings such as soil, dirt or fabric dyes may pollute forensic samples. Second, the absence of a DNA quantification step complicates DNA profiling as the forensic short tandem repeat (STR) multiplexes allow a modest window of DNA input; below 200 pg of DNA the profile may become incomplete and above 1 or 2 ng of DNA input profiles tend to get overloaded while with even higher inputs amplification is restricted to smaller amplicons in the multiplex or even completely absent. Commercially available direct PCR methods are therefore specifically recommended for reference samples and come with clear instructions regarding the size of a (FTA) paper punch to be used as PCR input [7]. Promising direct PCR results were obtained for the analysis of touched items or fabrics for which overloading of the profiles is hardly an issue due to the limited amount of DNA present in these samples [5]. Integrated (microfluidic) systems for forensic DNA analysis at the scene of crime or police station should combine DNA extraction, amplification and analysis of amplified products. Liu et al. [8] have developed a portable microdevice that integrates PCR and capillary electrophoresis (CE). Besides, Bienvenue et al. [10] developed a microfluidic device that performs an on-chip extraction followed by multiplexed STR amplification. Amplified products are analyzed by normal bench-top CE instruments. Since the microdevice does not quantify the DNA extract, the system is more suited for samples with known origin like reference samples. The microfluidic system developed by Hopwood et al. [9] is also particularly suited for reference buccal samples. From a crude extract DNA is isolated using a preset amount of DNA-binding beads. Consequently, there is limited variation between different buccal scrapings in the amount of DNA transferred to the amplification chamber and balanced profiles can be obtained after analysis of the PCR products on the integrated micro capillary electrophoresis chip. In this study, we aimed to develop a simple, rapid and direct PCR protocol to genotype a wide variety of forensic crime scene stains at a forensic laboratory. Such a procedure has the potential to provide a service important for police investigations and tracing of suspects. In addition, it can be used to quickly screen a set of stains, for instance blood spatters, for the stains with a profile deviant from the major donor (e.g. the victim). We evaluated an engineered Pyrococcus-like proofreading DNA polymerase that gained robustness and enhanced processivity from its fusion to a double-stranded DNA-binding domain [11,12] (Phusion1 Flash polymerase (Phusion)) that tethers the polymerase to the duplex DNA. This assists in increased association constants and robust, inhibitor-tolerant, amplification reactions. The use of a proofreading polymerase omits the final soaking step, but the allelic ladder needs to be re-amplified to obtain nonadenylated alleles. To further increase the speed of the cycling process, the effect of a high-speed thermal cycling system was assayed. The quality of the DNA profiles was evaluated by studying base-line noise, heterozygote balance, inter-locus balance and amplification artefacts in order to develop interpretation guidelines that include detection threshold, stochastic threshold and locus-specific stutter ratios. Several approaches were tested to regulate input from forensically relevant samples under the condition that the integrity and quality of the biological stain was maintained and that the sample could be analyzed at a later stage by standard procedures. The aim was to develop a simple protocol that results in STR profiles in 2–3 h, with a good success rate also for first-time users of the protocol.
2. Materials and methods 2.1. DNA samples and fabrics For initial experiments pristine DNA 007 was used which is the positive control in the AmpF‘STR1 SGM PlusTM (SGM+) kit (Applied Biosystems (AB), Nieuwerkerk a/d IJssel, the Netherlands). Body fluids (blood, semen, saliva), hairs and items representing mock case work samples were obtained from volunteers with informed consent. Bone samples were derived from femurs from 5 year-old graves excavated with consent in Brabant (the Netherlands). For all donors the STR genotype was known from standard DNA typing and concordancy was checked in all DNA profiles. When body fluids were placed on fabrics, the fabrics were cleared beforehand from DNA contamination by irradiation with 254 nm UV light in a CL-1000 UV CrossLinker (UVP Inc., UK) at 0.9 J/cm2 during 45 min for each fabric side. Smooth-surface mock case work samples such as cups, bottles, watches, car door handles, gear-controls, steering wheels, plastic surfaces were generally sampled using cotton swabs moistened with water according to standard procedures. Porous-surface mock case work samples like clothing, shoe laces, cigarette buds, plasters, tissues, cotton were sampled using tape-lift stubs. In addition, chewing gums, hairs and bone powder were used as mock case work samples. 2.2. Stubs and tape lifts Tape-lift stubs are produced from double-sided sellotape (12 mm 33 m, Pritt, the Netherlands). During in-house validation experiments, sellotape was found to efficiently collect cell material and not interfere with DNA extraction. These experiments involved the sampling of neck regions of worn clothes of various volunteers with five types of tape and subjecting the tape-lifts to three DNA extraction methods. Sellotape outperformed the other tape types and combined well with our standard QIAamp DNA extraction procedure (Fleur Beemster and TS, personal communication, details available on request), and after extensive validation this method is the accredited standard in our laboratory to sample fabrics. Sellotape is placed either on a 12.7 mm2 sized aluminium short pin stub holder (G301F, Agar Scientific, UK) or on a 6.6 mm2 mini pin stub holder (16181, Bioconnect, the Netherlands). The normal sized tape-lift stubs were used in the experiments in the sections ‘tape lifting as an effective approach’ and ‘direct and rapid PCR on mock case work samples’ while the mini tape-lift stubs were used in the experiments in the sections ‘direct PCR by unfamiliar users’ and ‘comparative DNA research within 6 h. The containers for the 12.7 mm2 stub holders are SEM stubs (G3626, Agar Scientific, UK), the containers for 6.6 mm2 stub holders are prepared from mini-tip swabs (B22715, WA Products, UK) by removing the wooden swab-containing stick and cutting the plastic container to 3 cm length. DNA decontamination occurs similar to the decontamination of fabrics. When a set of tape-lift stubs is produced, an acceptance test is performed to check for complete decontamination. We regard tape-lift stubs as one-time use items, although the plastic parts can tolerate one additional round of UV light decontamination. 2.3. DNA polymerases and STR PCR set-up Phusion1 Flash polymerase (Phusion) (Bioke´, Leiden, the Netherlands) was tested for direct PCR. A typical reaction had a 20 mL reaction volume and consisted of 10 mL 2 enzymecontaining Master Mix, 4 mL of appropriate STR primer mix and
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
6 mL of water and DNA. Standard AmpliTaq Gold1 DNA Polymerase (AmpliTaq) (AB) amplification reactions were carried out in 25 mL reaction volumes. Set-up occurred as recommended by the manufacturer. The following autosomal STR kits were used: SGM+, AmpF‘STR1 Identifiler1 (Identifiler), AmpF‘STR1 SEfiler PlusTM (SEfiler + ) or AmpF‘STR1 NGMTM (NGM) (all AB).
3. Thermal cyclers and amplification protocols Amplification reactions were performed on a GeneAmp1 PCR System 9700 (9700 cycler) (AB) or a 24-PIKO1 Thermal Cycler (PIKO cycler) (Bioke´). In the PIKO cycler appropriate ultra-thin walled tubes were used (Bioke´). SGM+, Identifiler and SEfiler+ amplifications used 28 cycles, while NGM amplifications took 29 cycles. Standard reactions used amplification conditions as provided by the manufacturer. Phusion amplifications on the 9700 cycler consisted of 1 min 98 8C hotstart, 28 cycles of 5 s 98 8C denaturation, 20 s 59 8C primer annealing and 20 s 72 8C elongation with a final soak of 30 min 60 8C. Unless stated otherwise, rapid amplification reactions on the PIKO cycler employed an initial denaturation at 98 8C for 5 min, followed by the appropriate number of cycles at 98 8C for at least 5 s, 59 8C for at least 30 s, 72 8C for 10 s with a final extension at 72 8C for 1 min. 3.1. STR analysis and allelic ladder preparation Amplified products were analyzed by adding 1 mL to a mixture of 8.7 mL Hi-Di formamide and 0.3 mL internal size standard (all AB, size standard for SGM+ is ROX GS500, that for SEfiler+ is LIZ GS600 and the one for Identifiler and NGM is LIZ GS500). When a reduced amount of amplified product is analyzed, the PCR mixture is diluted 1:100 (in water) and 1 mL of this dilution is added to 8 mL Hi-Di formamide and 1 mL of 1:100 diluted (in Hi-Di formamide) appropriate internal size standard. Samples were electrokinetically injected at 3 kV for 15 s and separated on a 3130xl Genetic Analyzer using POP-4 polymer and a 36 cm capillary array (all AB). After data collection, genotyping was performed using GeneMapperID-X or GeneMapper v3.2 (AB) using the bins and panels provided by the manufacturer. For the analysis of non-adenylated PCR products, allelic ladders were re-amplified as described in [13]. Appropriate allelic ladders were diluted 1:1000 in water, and 1 mL of this dilution was amplified during 15 cycles in 20 mL Phusion reaction. The mixture of buffer, enzyme and primers was prepared under pre-PCR laboratory conditions after which the template allelic ladder was added in a post-PCR room. Since re-amplified allelic ladders can be stored up to 2 years at 20 8C or 2 months at 4 8C, a large (e.g. 200 mL) batch of ladder can be prepared that – after quality check (all alleles present with sufficient height) – can serve many CE injections.
4. Profile interpretation When determining the percentage of detected alleles or the average peak heights, homozygous alleles are counted as two alleles. Peak height ratios (PHR) at heterozygous loci are determined by dividing the lowest peak height by the highest peak height. Inter-locus balance is calculated as the total rfu value observed at a locus divided by the total rfu value of the profile (ideally each locus would show 0.091 of the total rfu’s of an 11locus profile).
169
5. Results and discussion 5.1. STR profiling directly from blood and other forensically relevant body fluids AmpliTaq, the standard DNA polymerase in commercial STR typing kits, is known to be sensitive for inhibition by whole blood components while Phusion is reported by its manufacturer to tolerate up to 25% blood [12]. To explore the possibilities of Phusion for forensic purposes, amplification reactions were prepared from Phusion mastermix, SGM+ primermix and 0.4 ng pristine DNA. A standard 9700 cycler was used. Full profiles were obtained and negative amplifications showed no allele calls. The Phusion PCR products were one nucleotide shorter than the AmpliTaq products due to the 30 to 50 exonuclease activity of the enzyme. Successful allele calling was achieved by using an allelic ladder re-amplified with Phusion. The average peak heights, peak height ratios and inter-locus balance for the Phusion amplifications were similar as those in the standard AmpliTaq amplifications (Table 1). Next, various amounts (200 nL, 100 nL and 20 nL) of whole blood (resulting in 1%, 0.5% and 0.1% whole blood in the final PCR respectively) were added to the amplification mixtures that in addition contained 0.4 ng of pristine DNA. With AmpliTaq, no amplification was observed while with Phusion two-donor profiles were obtained that corresponded to DNA007 and the blood donor (results not shown). Clearly, Phusion can amplify DNA templates in the presence of whole blood. For the inputs of 200 nL and 100 nL whole blood the profiles were over-amplified, which implies that the majority of the blood DNA is released (100 nL blood contains approximately 2–4 ng DNA [14], which would indeed lead to an overloaded STR profile). Due to their structure, sperm cells may need a more stringent protocol (longer 98 8C hotstart period) to release DNA [15]. Two semen samples were diluted so that approximately 100 cells were added to the amplification reactions. These two semen samples and 3 nL whole blood were directly amplified in 4-fold with different hotstart periods (Supplementary Table S1). Overall optimal results were observed using a 5 min hotstart period, and this protocol was used when examining the ability of Phusion to perform direct PCR on various inputs of blood, saliva and semen. Phusion enabled direct PCR amplification of these body fluids as full SGM+ profiles were obtained (Table 2). AmpliTaq only showed efficient direct amplification for saliva but when inputs increased over 0.5 mL, A artefact peaks were abundant. For Phusion, overamplification was evident with inputs of 50 nL whole blood, 2 mL saliva or 1000 sperm cells, which is consistent with the amount of DNA generally present in these inputs [14]. In Phusion reactions, over-amplification is not accompanied by the occurrence of A artefact peaks as Phusion produces blunt-ended PCR products due to its proofreading ability. 5.2. Rapid, direct STR typing and guidelines for reliable interpretation of profiles To reduce the time needed for direct amplification, we tested the PIKO cycler system that is recommended by the manufacturer
Table 1 Characteristics of SGM+ profiling of 400 pg pristine DNA using Phusion or AmpliTaq.
Phusion AmpliTaq a
Averagea peak height
Average PHR
Inter-locus balance
875 576 rfu 872 488 rfu
0.83 0.13 0.83 0.12
0.042 (FGA)–0.133 (vWA) 0.048 (FGA)–0.120 (vWA)
Average of 9 independent amplifications.
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
170
Table 2 Performance of Phusion for direct amplification of various body fluids. Body fluid
Input
Blood
8 nL 16 nL 50 nL 0.1 mL 0.5 mL 2 mL 6 mL 100 cells 1000 cells
Saliva
Semen a b c d
Phusion
AmpliTaq
% Detecteda
Average rfub
% Detected
Average rfu
100% 100% 100% 100% 100% 100% 99 2% 100% 100%
1099 176 2094 83 5852 621c 1046 163 4217 491 6438 68c 5766 527c 1889 585 4376 1244c
40 36% 7 8% 1 2% 100% 100%d 100%c,d 87 13%d 93 10% 39 5%
114 36 54 5 58 5 613 153 1498 236 3747 905 1133 543 370 105 148 35
Average of 4 concurrent amplifications. Average of detected alleles. Over-amplification occurred. Showing A artefacts peaks.
to further speed up the amplification process when using Phusion [16]; time periods of 1 s for denaturation, 5 s for primer annealing and 15 s/kb for elongation may suffice. The forensic SGM+ STR kit is an 11-loci multiplex. To enable efficient annealing of all primers, several annealing periods were tested using 0.6 ng pristine DNA. Optimal results were obtained using the 30 s annealing time (Supplementary Table S2). The total run time for this PCR is 47 min (Supplementary Table S3). The heterozygote and inter-locus balances for the Phusion amplifications are comparable to those found for the standard AmpliTaq reactions (Supplementary Table S2, Table 1). No artefacts are observed in amplification negatives (results not shown). The 47 min PIKO cycler protocol was used to amplify eight amounts (ranging from 20 nL to 0.16 nL) of stored or freshly collected blood of eight different donors. In addition, 20 amplification negatives were performed. In the 64 blood amplifications, the base line noise was increased which was not the case in the amplification negatives. We assume this raised noise is due to components in the body fluids affecting the PCR. When the detection threshold was increased from 50 rfu to 80 rfu, the number of baseline artefact peaks that were called was greatly reduced except for the 20 nL and 10 nL inputs that showed high noise levels and over-amplification. In the blood profiles, two types of artefact peaks were observed that were not due to increased base line noise: (1) peaks at 1 repeat unit stutter position, and (2) low level artefacts peaks at specific positions. Stutters at +1 repeat unit were seldom observed. For each locus, the average percentage of the 1 repeat unit stutter peaks was calculated together with the standard deviation. By taking the mean plus twice the standard deviation (two sigma rule) locus-specific 1 repeat unit stutter ratios were determined for direct and rapid PCR (Table 3, Supplementary Fig. S1). For eight of the ten STR loci these 1 repeat unit stutter ratios are higher than the ratios provided by the manufacturer for standard SGM+ typing; for two loci the ratios are Table 3 Locus-specific 1 repeat unit stutter percentages for direct and rapid amplification. Marker
Average %
Stdev
# Stutters
m + 2sa (%)
AB %b
D3S1358 vWA D16S539 D2S1338 D8S1179 D21S11 D18S51 D19S433 TH01 FGA
9.3 11.5 10.4 13.7 11.8 6.4 11.8 7.6 8.8 14.9
2.9 2.0 3.0 4.1 3.3 2.2 3.6 2.5 3.8 2.9
56 63 66 68 60 62 78 61 60 60
15.1 15.5 16.4 22.1 18.5 10.8 19.0 12.6 16.3 20.7
11.0 11.0 13.0 15.0 12.0 13.0 16.0 17.0 6.0 11.0
a b
Mean plus twice standard deviation. Stutter percentage provided by manufacturer (AB).
lower (Table 3, Supplementary Fig. S1). To harbour all stutter peaks, the 1 repeat unit stutter distance in the analysis settings was slightly increased (from 3.25–4.75 to 3.0–5.0nt). Common, low level (but regularly above detection threshold) artefacts peaks were seen for the following positions: D3S1358 allele positions 11 and 14, vWA allele positions 10, 11 and 14, D16S539 allele position 7, D8S1170 allele position 7, D21S11 allele position 33, THO1 allele positions 6 and 12. The use of a different size standard (Promega’s ILS600 [17] instead of AB’s ROX-500), the application of a higher annealing temperature or the use of DTR column purification did not prevent the occurrence of these peaks (results not shown). In single donor profiles with rfu heights over 500 rfu, artefact peaks can be readily distinguished from true alleles as the artefact peaks stay low level (Fig. 1). When these analysis settings (detection threshold 80 rfu, modified 1 repeat unit stutter distance and ratios, marked artefact peaks) were applied to the 64 whole blood amplifications, full STR profiles were observed down to 1.25 nL blood (Table 4). The profiles showed adequate intra- and inter-locus balance (Supplementary Table S4). Due to the fact that various donors were used and that the storage times of the blood samples varied from 2 years to freshly taken, quite some variation in profiling results and peak heights were obtained (Table 4). Allele drop-out started to occur when 2.5 nL blood was analyzed. Drop-out was not found at specific loci; although larger amplicons are more prone to allele drop-out as with standard STR typing. The height of single peaks at heterozygous loci was determined in order to obtain insight in the stochastic threshold, which warns for the possible occurrence of an allelic drop-out. In total 69 allelic drop-outs were observed. The heights of the concurrent single peaks varied from 80 to 762 rfu (Supplementary Fig. S2). To flag 95.7% of these single allele calls as possibly heterozygous, a stochastic threshold of 375 rfu is needed. This is a rather high value and accordingly, 30 of the 142 homozygous peaks also receive this flag. Alternative approaches would be to evaluate only profiles above for instance 400 rfu, and/ or perform multiple independent amplifications for replicate analysis. We have taken this advice into account when developing a protocol to analyze case work samples, as described onwards. 5.3. Tape lifting as an effective approach to regulate PCR input from swabs and various fabrics Evidently, regulation of PCR input is essential for successful direct and rapid amplification; with too much input profiles become over-amplified and show increased base line noise and with too little input incomplete profiles are obtained. Samples at forensic laboratories mainly consist of textiles or swabs obtained from sampling various evidentiary items. We set out to develop a sampling technique that: (1) results in full STR profiles upon direct
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
171
Fig. 1. Low level artefact peaks (marked by arrows) occurring at specific positions in rapid and direct amplifications; (A) artefact at vWA position 14, (B) artefact at D21S11 position 33, (C) artefact at THO1 position 6 (coinciding with 1 repeat unit stutter peak). The upper number is the allele call, the lower number corresponds to peak height (rfu).
and rapid PCR, (2) is practical and applicable to both swabs and fabrics, (3) comes with minimal handling and contamination risks, and (4) leaves the remainder of the swab or fabric intact (dry) for standard analysis at a later stage. Blood (3 mL) was applied on cotton swabs, and five methods were tested: (1) twist swab twice or (2) 20 times in PCR tube (to enable this approach mini-tip swabs were used), (3) take out a fibre from swab, (4) cut the tip of the swab and (5) stub the swab with a mini tape-lift. We tested these five methods with a replicate number between 4 and 16, and found that only two methods resulted in full profiles for all direct amplifications namely the fibre method and the tape lifting approach (results not shown). The tape lifting method proved much more practical: easy to perform with a minimal risk of contamination (the tapes are treated extensively with UV light in a crosslinker to remove DNA contamination) and the swab remains completely intact for further use. Tape lifting (Fig. 2) can either be performed using a tape on a 12.7 mm2 stub holder from which an area of 5 mm2 is cut by a scalpel knife or by using a smaller tape placed on a 6.6 mm2 mini stub holder. The tape lifting method was also tested for bloodstains (3 mL undiluted and 3 mL 1:10 diluted blood) on various types of fabrics (Table 5), and proved successful. The fabrics have different characteristics: some hardly adsorb the blood (nylon), others easily release blood-containing fibres (wool) and some take up the blood tightly and hardly release fibres (course cotton). Therefore, the tape lift sampling number (times the stain is touched with the tape-lift stub) varied for the various fabrics depending on the release of blood and fibres as determined by visual inspection (Table 5). Only for denim no genotyping data were obtained which is probably due to PCR inhibition from denim components. 5.4. Direct and rapid PCR on mock casework samples 149 mock casework samples, expected to contain single source cell material of known donors, were collected to study the success
rate of direct and rapid PCR. The tape lift procedure was used for all stains residing on swabs. For items sampled by tape-lift stubs, 5 mm2 areas were excised from the tape. Chewing gum, hairs and bone powder were added directly to the PCR mixture. Full STR profiles concordant with the donor of the sample were obtained for 60% of the samples representing classical forensic trace evidence types (Table 6). All blood samples and half of the saliva samples yielded full profiles. Hair roots give 65% of full profiles when pulled hairs are used, but the profiles appear overloaded and are best analyzed after injecting less (0.01 mL) PCR product. When hairs are analyzed no sample remains as the hair is fully used during direct and rapid PCR. When items containing contact DNA and bone samples are included, the overall success rate decreases (46%) (Table 6). The bone samples gave full profiles in standard DNA analysis, but no genotyping results using direct PCR. This difference is most likely due to the use of too little bone material for the direct PCR; approximately 10 mL of bone powder was used for DNA isolation of which 1/20th was used as PCR input (this corresponds to 500 mL bone powder), while for direct PCR only 10 mL bone powder was added. For 31 samples the remainders of the swabs or fabrics were analyzed using the regular procedure of DNA isolation, quantification and profiling. The 31 samples were of three categories namely samples giving a full, a partial or no DNA profile using rapid and direct PCR. A large part of the samples gave a similar result using the regular genotyping procedure (Table 7). For samples showing no profile with the direct procedure, standard genotyping can be successful and result in a partial of even complete profile (Table 7); apparently insufficient cell material was collected for the direct procedure. For only one sample better results were found with direct and rapid PCR (full profile) than with the standard procedure (no profile). This sample comprised a tape lift from a car door handle, and it is possible that the part of the tape lift containing cell material was consumed in the direct PCR. In the profiles obtained
Table 4 Rapid and direct amplification of various inputs of whole blood.
Full profile Average rfuc PHR # drop-ins a b c d
20 nL
10 nL
5 nL
2.5 nL
1.25 nL
0.63 nL
0.32 nL
0.16 nL
8/8 4862 621 0.89 0.10 15
8/8 3989 1229 0.82 0.13 7
8/8 2658 1260 0.73 0.18 7
6/8a 1143 654 0.71 0.19 5
5/8b 517 281 0.67 0.19 1
0/8 241 117 0.57 0.22 0
0/8 150 34 ndd 0
0/8 180 65 nd 0
Only allelic drop-outs were seen, 1 per profile, the two profiles both had an average height of approximately 350 rfu. In one of the incomplete profiles a locus drop-out occurred, in all 3 incomplete profiles an allelic drop-out was seen. Homozygous alleles were calculated as 2 alleles. Not determined, too few heterozygous loci detected.
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
172
Fig. 2. Tape lifting to regulate input for direct and rapid PCR. A mini tape-lift stub (A) is used to lift cell material from a blood-containing swab (B and C) after which the tape is removed from the stub holder (D and E) and placed in a PCR tube (F).
Table 5 Direct and rapid PCR of blood stains using tape-lift stubs.
Cotton Undiluted blood # Times with tape 1:10 blood # Times with tape a b
Cotton swab
Denim
Linen
Viscose
Nylon
Fleece
Wool
Fine cotton
Coarse
16/16a 2 ndb nd
0/4 20 0/4 20
4/4 5 3/4 10
4/4 1 4/4 20
3/4 1 4/4 1
4/4 1 4/4 1
4/4 1 2/4 1
4/4 5 2/4 10
4/4 10 1/4 30
Number of full STR profiles of total number of amplifications. Not determined.
Table 6 Direct and rapid PCR of mock casework samples. Type
Subtype
Full profile
Partial profile
No profile
Blood
Plaster or tissue Smooth surface Cup or bottle Chewing guma Cigarette budb Tissue, spit or buccal swab Smooth surface Pulled Fallen
7/7 6/6 11/26 5/9 5/10c 9/12 3/4 11/17 0/4
0/7 0/6 10/26 3/9 2/10 2/12 1/4 4/17 1/4
0/7 0/6 5/26 1/9 3/10 1/12 0/4 2/17d 3/4
Total
Classical trace evidence
57/95 (60%)
23/95 (24%)
15/95 (16%)
Contact DNA
Bone
Clothes Personal itemse Used itemsf Powdered
4/16 4/14 3/18 0/6g
5/16 5/14 1/18 0/6
7/16 5/14 14/18 6/6
Total
Non-classical trace evidence
11/54 (20%)
11/54 (20%)
32/54 (60%)
Overall total
All types of trace evidence
68/149 (46%)
34/149 (23%)
47/149 (31%)
Saliva
Semen Hair
a b c d e f g
PCR input not by tape lift but by cutting a small piece. Filter side is stubbed. Mixed profiles were obtained for 3 cigarettes. Both from one specific donor. Watches, shoe laces. Door handles, car items. Powder of 2 bones was used.
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175 Table 7 Genotyping results using both direct and rapid PCR and standard profiling on the same samples (n = 31). Results direct
Results regular Full
Partial
No profile
Full (11) Partial (9) No profile (11)
10 0 2
0 9 5
1 0 4
Total (31)
12
14
5
by regular genotyping, no DNA contamination due to the tape lift procedure was detected. 5.5. Rapid and direct PCR executed by unfamiliar users Based on the above described experiments, a set of instructions for direct and rapid PCR was prepared (Table 8) that describes for various sample types how many reactions should be performed, how the tape lifting or the sample cutting should be executed, and how much of the PCR mixture should be analyzed by CE. With direct and rapid PCR, samples can get over-amplified. Since the Phusion polymerase has proofreading ability, amplified products do not carry adenine overhangs at their 30 ends. Therefore no A artefacts occur and analyzing 0.01 mL of the amplified products is very effective when over-amplification occurred. The set of instructions results in multiple DNA profiles from a sample; two to three tape lifts are taken (except for hair roots) and two amounts of amplified products are analyzed (except for items containing contact DNA). Therefore for most samples several interpretable (not over-loaded and not incomplete) profiles are obtained, often from independent amplifications (which may be useful to accommodate the relative high stochastic threshold for Phusion profiles). We set out to test whether the set of instructions (Table 8) could effectively be carried out by first-time protocol users. We generated four identical cotton patches with 6 bloodstains (2 or 5 mL spot size). Four stains consisted of undiluted blood and two stains contained 5-times diluted blood. Two different blood donors were used and one of the stains of undiluted blood consisted of a 1:1 mixture of the 2 donors. Three researchers unfamiliar with the procedure and one experienced researcher performed direct and rapid PCR by following the set of instructions. All researchers obtained full profiles for all stains (a representative electropherogram is shown in Fig. 3). For the diluted bloodstains all researchers generated full profiles with standard CE injection, but for undiluted stains often CE with reduced input (1:100 dilution) was needed, depending on spot size and researcher (apparently tape lifting is performed slightly different by the various researchers). All researchers identified the donors of the blood stains correctly including the mixed stain. This shows that direct and rapid PCR has an additional application: when many bloodstains are present on an evidentiary item a screening can be performed to identify stains that do not match the DNA profile of the major donor of the bloodstains e.g. the victim. These stains can be genotyped in a next analysis round using the standard DNA typing procedure.
173
5.6. Comparative DNA research within 6 h In four experiments we challenged the time for comparative DNA research when DNA profiling is performed by direct and rapid PCR. Nineteen mock casework samples were processed in total. Time was calculated from intake of the samples at the Netherlands Forensic Institute to reporting of the results by phone or email. At intake, samples were labelled and registered according to our standard procedures. The samples were analyzed by direct and rapid PCR after which DNA results were doubled-checked by two analysts. Next, a DNA database search was performed against persons present in the Dutch DNA Database and a reporting officer summarized the results in a report. After checking of this report by a second reporting officer, it was send by email. In all four experiments the total time for this process was below 6 h. To achieve a high success rate for this fast comparative DNA research, the trace evidence needs to comply to the following aspects: (1) contain blood, saliva or semen as these body fluids contain relatively high levels of DNA; (2) correspond to cell material of a single donor as mixture de-convolution requires too much time and rapid and direct profiles show increased (base-line) noise and low level artefact peaks and (3) process a maximum number of three items simultaneously. We have decided that single donor profiles obtained by this DNA-6 h process can be used for DNA database searching but not for permanent DNA database entry. Because of the increased noise in the rapid and direct profiles, the DNA-6 h service is always accompanied with priority research in which the remainder of the trace evidence (swab or fabric remaining after tape lift procedure) is analyzed using standard procedures that include presumptive testing for stain type, mixture analysis (if applicable) and a comprehensive report. Profiles obtained by this second (standard priority) research can be used for permanent database entry and searching international databases under the Pru¨m treaty. Due to this set-up DNA-6 h comes with additional costs and extra hands-on time (costs include the tape-lift stub, SGM+ primer mix, Phusion mastermix, PIKO cycler tubes and CE run costs). Therefore DNA-6 h is especially appropriate for cases that may benefit substantially from extremely fast analysis (national security, danger of life, serial perpetrator). This service is offered to the Dutch police since 7th October 2010. In the first three months four cases have been addressed; all cases involved blood samples and for all samples single donor full profiles were obtained within 6 h. Importantly, the same genotyping results were obtained with the DNA-6 h protocol and during the subsequent standard analysis. 5.7. Rapid and direct PCR and other STR kits Currently, the SGM+ kit is the standard genotyping kit in our laboratory. Other forensic laboratories may utilize a different standard kit and we tested whether direct and rapid PCR would be compatible with other STR kits. Six nanoliters of whole blood were amplified in 4-fold with the SGM+, SEfiler+, Identifiler and NGM primer sets. Since these STR kits amplify 12–16 loci, both the standard annealing time of 30 s and an extended time of 50 s were tested. Using a 80 rfu detection threshold (as with SGM+), full
Table 8 Instructions for tape lifting for direct and rapid PCR. Stain
Tape-lift: 3 mm2, gentle pressure, max 1 s per touching
CE injection
Blood undiluted Blood diluted (swab) Saliva or semen spot Saliva drinking or cigarette Chewing gum Contact traces Hair
Duplicate: touch 1–5 times, a faint light red shade Duplicate: touch 3–10 times, a faint light red shade Triplicate: touch 1, 5 and 10 times Triplicate: touch 10, 25 and 50 times Duplicate: cut 1 mm2 and 0.5 mm2 piece Duplicate: until tape is saturated Single (hair is fully used): cut hair root
1 1 1 1 1 1 1
and and and and and
1:100 1:100 1:100 1:100 1:100
and 1:100
174
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
Fig. 3. Typical SGM+ DNA profile obtained after direct and rapid PCR of a bloodstain on fabric. CE was performed using 0.01 mL PCR mixture. The AF14 label in the vWA locus refers to a low level artefact peak, commonly found at this position. The upper number is the allele call, the lower number corresponds to peak height (rfu). Note that some bleed-through is observed mainly in the yellow channel originating from peaks in the green channel; this should not be regarded as increased base line noise.
SEfiler+ and Identifiler profiles were obtained already with an annealing time of 30 s (Supplementary Fig. S3AB), showing that also 12 or 16 loci (SEfiler+ or Identifiler respectively) can be amplified simultaneously via this rapid and direct PCR protocol. The NGM profiles were incomplete, and the loci D21S11 and D18S51 consistently dropped out or were very low (Supplementary Fig. S3C). Apparently the direct and rapid PCR conditions (annealing time, buffer composition) disfavour amplification of these loci when using the NGM primer mix. When a combined annealing elongation time of 3 min at 59 8C was used (which resulted in a 2 h in stead of a 47 min PCR protocol), these loci were amplified (Supplementary Fig. S3D). Evidently, to enable direct and rapid PCR via the NGM kit, further protocol optimization is required. Artefact peaks did also occur in the Identifiler, SEfiler and NGM profiles, and some of these artefact peaks were at the same positions as in the SGM+ profiles (e.g. allele 14 vWA, allele 33 D21S11). This is probably due to the use of primers with identical primer binding site positions in the various kits. 6. Concluding remarks A procedure for direct and rapid PCR was developed that builds on several components; direct PCR is facilitated by Phusion1 Flash
DNA polymerase, a highly processive, robust, proofreading heat stable polymerase. Rapid cycling is aided by the use of the PIKO cycler system. Regulated sample input is achieved by collecting biological material from body fluid-containing cotton swabs or fabrics on small tape lifts and adding these to the amplification reactions. A set of instructions was prepared to achieve a PCR input of cellular material for different stains and different substrates that results in an interpretable PCR profile. This set of instructions could be successfully used by researchers unfamiliar with the procedure. Analysis of 149 mock casework samples showed that success rates mainly depend on the type of stain. Blood stains are the most promising, probably because blood is a visible stain with a relatively high number of DNAcontaining cells. DNA profiles obtained from direct and rapid PCR need to be analyzed using a modified allelic ladder as the PCR products are non-adenylated. The profiles show increased base-line noise, increased stutters at some loci and specific low level artefact peaks. To minimize aberrant allele calling, modified profile interpretation guidelines are required that accommodate these noise effects. Therefore, direct and rapid PCR profiles appear less suited to deduce mixed profiles (although the method can detect presence of multiple donors). Trace evidence consisting of blood, saliva or
S. Verheij et al. / Forensic Science International: Genetics 6 (2012) 167–175
semen from one donor, suits this method best. We have incorporated rapid and direct PCR in a ‘DNA-6 h’ service that aims to assist police investigations and tracing of suspects in cases for which every minute counts. DNA-6 h rapidly derives DNA information from a single-donor biological stain that is left by a perpetrator which can be used to search a DNA database. In addition, direct and rapid PCR can be used to screen a set of stains on an evidentiary item for the stains that do not relate to for instance the victim but may belong to the perpetrator. In preliminary experiments direct and rapid PCR was found to be compatible with other STR kits, indicating that the method described here may serve as a framework protocol that can be adapted and optimized within different forensic laboratories. Role of funding This study was supported by a grant from the Netherlands Genomics Initiative/Netherlands Organization for Scientific Research (NWO) within the framework of the Forensic Genomics Consortium Netherlands. Acknowledgements We thank Corina Benschop, Bas de Jong and Ankie van Gorp for useful discussion, Alexander Lindenbergh, Bas de Jong, Lise De Strooper, Bryan Bhoelai for experimental assistance and Ankie van Gorp and Charissa van Kooten for critically reading the manuscript.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.fsigen.2011.03.014.
175
References [1] P.M. Vallone, C.R. Hill, J.M. Butler, Demonstration of rapid multiplex PCR amplification involving 16 genetic loci, Forensic Sci. Int. Genet. 3 (2008) 42–45. [2] P.M. Vallone, C.R. Hill, D. Podini, J.M. Butler, Rapid amplification of commercial STR typing kits, Forensic Sci. Int. Genet. Suppl. Series 2 (2009) 111–112. [3] D.Y. Wang, C.-W. Chang, L.K. Hennessy, Rapid STR analysis of single source DNA samples in 2 h, Forensic Sci. Int. Genet. Suppl. Series 2 (2009) 115–116. [4] H. Giese, R. Lam, R. Selden, E. Tan, Fast multiplexed polymerase chain reaction for conventional and microfluidic short tandem repeat analysis, J. Forensic Sci. 54 (2009) 1287–1296. [5] A. Linacre, V. Pekarek, Y.C. Swaran, S.S. Tobe, Generation of DNA profiles from fabrics without DNA extraction, Forensic Sci. Int. Genet. 4 (2010) 137–141. [6] A. Barbaro, P. Cormaci, A. Teatino, A. Barbaro, Use of ‘‘AnyDirect PCR buffer’’ for PCR amplification of washed bloodstains: a stain report, Forensic Sci. Int. Genet. Suppl. Series 1 (2008) 11–12. [7] D.Y. Wang, C.-W. Chan, N.J. Oldroyd, L.K. Hennessy, Direct amplification of STRs from blood or buccal cell samples, Forensic Sci. Int. Genet. Suppl. Series 2 (2009) 113–114. [8] P. Liu, S.H.I. Yeung, K.A. Crenshew, C.A. Crouse, J.R. Scherer, R.A. Mathies, Real-time forensic DNA analysis at a crime scene using a portable microchip analyzer, Forensic Sci. Int. Genet. 2 (2008) 301–309. [9] A.J. Hopwood, C. Hurth, J. Yan, Z. Cai, N. Moran, J.G. Lee-Edghill, A. Nordquist, R. Lenigk, M.D. Estes, J.P. Haley, C.R. McAlister, X. Chen, C. Brooks, S. Smith, K. Elliott, P. Koumi, F. Zenhausern, G.A.J. Tully, Integrated microfluidic system for rapid forensic DNA analysis: sample collection to DNA profile, Anal. Chem. 82 (2010) 6991–6999. [10] J.M. Bienvenue, L.A. Legendre, J.P. Ferrance, J.P. Landers, An integrated microfluidic device for DNA purification and PCR amplification of STR fragments, Forensic Sci. Int. Genet. 4 (2010) 178–186. [11] Y. Wang, D.E. Prosen, L. Mei, J.C. Sullivan, M. Finney, P. Vander Horn, A novel strategy to engineer DNA polymerases for enhanced processivity and improved performance in vitro, Nucleic Acids Res. 32 (2004) 1197–1207. [12] P. Yang, C. Andre´, Efficient PCR from whole blood using Finnzymes high performance PCR solution, Nat. Methods Appl. Notes (2007) an20–an21. [13] J.M. Butler, Y. Shen, B.R. McCord, The development of reduced size STR amplicons as tools for analysis of degraded DNA, J. Forensic Sci. 48 (2003) 1054–1064. [14] http://www.oglethorpe.edu/faculty/k_aufderheide/Forensic_Science/Web_D ocuments/Catalin_Andrei_Mitrasca.pdf. [15] A.M. Hossain, B. Rizk, A. Behzadadian, I.H. Thorneycroft, Modified guanidium thiocyanate method for human sperm DNA isolation, Mol. Hum. Reprod. 3 (1997) 953–956. [16] http://www.finnzymes.com/pdf/f548sl_phusion_flash_highfidelity_pcr_maste rmix_datasheet_1_3_low.pdf. [17] R.S. McLaren, M.G. Ensenberger, B. Budowle, D. Rabbach, P.M. Fulmer, C.J. Sprecher, J. Bessetti, T.M. Sundquist, D.R. Storts, Post-injection hybridization of complementary DNA strands on capillary electrophoresis platforms: a novel solution for dsDNA artifacts, Forensic Sci. Int. 2 (2008) 257–273.