Introduction

Bite marks on human skin often result from assaults, particularly sexually motivated attacks and child abuse [1–3]. The elastic nature of skin and frequent lateral movement of the dentition during biting complicate forensic morphometric analyses and consequent conclusions as to the origins of bite marks often involve elements of subjectivity [4–6]. Furthermore, the premise that humans have anatomically unique and distinguishable dentitions has been challenged with consequent demand for development of alternative methods of bite marks analysis [4, 7–10]. DNA-based techniques have been successfully applied to other areas of forensic science [9] and hold promise for bite mark analysis [6, 7, 11]. However, recovery of human genomic material from bite marks and saliva stains may be compromised by salivary components that interfere with persistence and amplification of DNA [12–15].

The genus Streptococcus comprises a genotypically diverse group of bacteria that constitute the numerically dominant cultivable genus establishing on the supragingival sites of human teeth [16, 17]. Individuals select distinctive (possibly unique), relatively stable collections of streptococcal genotypes [18, 19] and those cultured from experimental bite marks can be matched to isolates from the teeth responsible [11, 20] providing an encouraging avenue for objective analysis of bite marks [21]. However, the requirement for culturing bacteria has practical difficulties for forensic laboratories and the purpose of this study was to assess the feasibility of direct amplification of streptococcal DNA from bite marks. This approach could also extend the time, following assault, over which bacterial DNA evidence could be gathered because oral streptococci remain cultivable for no more than 24 h on human skin [11].

Materials and Methods

Experimental Bite Marks and Analysis of Streptococcal DNA

Self-inflicted bite marks were generated and sampled essentially as described previously [11, 20]. Twenty-four healthy participants bit their own exposed upper arms, maintaining the bite with as much force as they could tolerate for 5 s. Three hours later, the bite marks were sampled by firmly rubbing a saline-moistened swab over the marks corresponding to the lower incisors. Unbitten control sites, adjacent to the bite marks, were similarly sampled. Swabs were also taken from the lingual surfaces and biting edges of the lower incisors and were immediately agitated in sterile saline (2 mL). The saline was centrifuged (11,000×g for 3 min, 4 °C) and the supernatant discarded. Pellet material was combined with 200 μL of InstaGene™ matrix (BioRad Laboratories, Hercules, CA. USA) and incubated at 56 °C (10 min). Tubes were then boiled for 8 min, cooled, and centrifuged (11,000×g, 2.5 min). Supernatant material containing extracted DNA was recovered and frozen until required for PCR.

Comparison of amplicons from bite mark, incisor, and skin samples was achieved by denaturing gradient gel electrophoresis (DGGE) essentially following the method of Rudney et al. [22]. The PCR primers (given below) flank a region of 16S ribosomal DNA (≈240 bp) that distinguishes streptococci at a subspecies level [22]. Using reagents supplied in the Hotmaster™ Taq DNA Polymerase kit (Eppendorf), PCR reaction mixtures (50 μL) contained: 10× buffer (5 μL); Taq polymerase (0.2 μL); dNTP mixture (1 mM); forward and reverse primers (0.1 μM); sample extract (5 μL); and the remaining amount was water. Forward primer (GC clamp underlined): 5′-CGC CCG CCG CGC GCG GCG GGG CGG GGC GGG GCG GGG GCA CGG GGG GGA GGT TGA TCA TGG CTC AG-3′, Reverse primer: 5′-ACA ACG CAG GTC CAT CAT-3′. The thermal cycler programme was initiated at 94 °C for 1 min, followed by 35 cycles involving a denaturation step at 94 °C for 30 s; annealing at 56 °C for 30 s; and extension at 72 °C for 30 s. PCR products were loaded into polyacrylamide gels (8 %) containing a urea/formamide gradient as described previously [22]. The relative migration distance of amplicons was determined by normalizing against reference amplicons. Attempts were also made to compare PCR products by terminal restriction fragment-length polymorphic analysis [23], but this approach was deemed unsatisfactory because of the lack of terminal sequence variability.

Excised fresh porcine skin was used to determine the length of time that oral streptococcal DNA (derived from humans) could be amplified from deteriorating epidermis. Porcine skin is regarded as an appropriate alternative to human epidermis for bite mark experiments [24]. The lower incisors of three human participants were vigorously rubbed with sterile swabs that were then applied to the dermal surface of segments (approximately 25 cm2) of freshly obtained porcine abdominal skin (from local supermarket). Inoculated skin segments were maintained for up to 14 days under the following conditions: (i) room temperature with ambient moisture; (ii) room temperature with enhanced humidity; and (iii) 4 °C with ambient moisture. To insure an accumulation of biofilm material on the teeth, the experiments were timed such that at least 3 h elapsed between obtaining consecutive incisor samples from the same donor. Each epidermis square was sampled only once and samples were analyzed by PCR/DGGE as described above.

This investigation was approved by the University of Otago Human Ethics Committee (Ref. 06/169).

Results and Discussion

The term “corresponding” is used for comparisons of amplicon profiles derived from sites (skin, teeth, and bite marks) from the same participant, and “non-corresponding” refers to inter-participant comparisons.

Incisor samples generated between 4 and 17 amplicons following resolution by DGGE which accords with the range of distinguishable streptococcal strains cultivable from human teeth [20]. With one exception, fewer amplicons were recovered from the bite marks than from the corresponding incisors (Table 1) probably because material recovered from bite marks involved two transfers (from incisors to skin and from skin to swab), whereas sampling the incisors involved a single transfer. Five bite marks failed to generate detectable PCR products despite the production of several amplicons from the corresponding incisors samples (Table 1). Streptococcal amplicons were generated from ten of the unbitten skin samples, but these sites produced fewer amplicons than either the incisor or bite mark samples (Table 1). Only three amplicons recovered from unbitten skin sites matched amplicons from the corresponding bite mark, thus contamination from skin organisms was minor and did not interfere with the analysis. Examples of PCR/DGGE amplicon profiles are shown in Fig. 1.

Table 1 PCR amplification of streptococcal 16S ribosomal DNA from experimental bite marks, unbitten skin sites, and lower incisors, from 24 participants
Full size table
Fig. 1
figure 1

Comparison of PCR amplicons generated from (S) skin, (T) incisors, and (B) bite marks from three individuals (1, 2, and 3, respectively) by DGGE. The extreme lanes (R) contain streptococcus reference amplicons from S. mutans ATCC 25175, S. sobrinus NIDR 6715, S. sanguinis ATCC 10556, S. gordonii DL1, S. mitis I18, S. criceti ATCC 19642, and S. oralis 34

Full size image

Amplicon patterns generated from bite marks with six or more DNA bands were compared with those from corresponding and non-corresponding incisors by cluster analysis (Cluster version 3.0; Stanford University and University of Tokyo [http://bonsai.ims.u-tokyo.ac.jp/~mdehoon/software/cluster/software.htm]). Eight (of 15) bite mark amplicon patterns were matched to the corresponding incisor samples by correlation coefficients greater than 0.70, with one pairing (24B, 24T) scoring 1.0 (Fig. 2). A further three bite marks were also matched most closely to the corresponding teeth, but with coefficients ranging between 0.5 and 0.69 (Fig. 2). However, the bite marks exhibiting closest similarity to non-corresponding incisor samples (23B, 13T and 13B, 23T) recorded coefficients of 0.68 (Fig. 2) though both of these were more closely correlated to their respective corresponding sites. The highest correlation between incisor amplicon profiles was 0.57, giving an indication of the level of co-incidental similarity between unrelated profiles. As an alternative method of analysis, amplicon profiles were compared by pair-wise similarity coefficient (Cs) determination [25] after discarding profiles with less than six amplicons. The Cs values for corresponding bite marks and incisor samples were significantly greater than those for non-corresponding comparisons (Table 2), but again the range of values from corresponding and non-corresponding samples overlapped slightly (as with the cluster analysis). Cs values for corresponding samples were not significantly related to the number of amplicons generated from the bite mark (p = 0.38); thus, matches did not result simply because there were greater numbers of genotypes present. Note that both methods of analysis penalize the absence of bands from the bite mark profile and are therefore quite stringent for this application because generating a perfect match (Cs = 1) requires that a bite mark produces the same number of amplicons as the corresponding tooth sample. Since bite marks almost invariably generated fewer amplicons than the corresponding teeth samples (due to the extra material transfer), perfect matches were unlikely.

Fig. 2
figure 2

Similarity of streptococcal amplicon profiles derived from 15 bite marks (B) and 22 incisors (T) samples. The dendrogram was constructed with gene-expression visualization software (version 1.1.3; Java Treeview, Alokito [http://jtreeview.sourceforge.net]). Profiles generating less than six amplicons were omitted

Full size image
Table 2 Comparison of amplicon profiles derived from bite mark and incisor samples by (unweighted) pair-wise similarity coefficient
Full size table

The overall findings of the direct amplification approach accord with results from the more labor-intensive method of genotyping individual bacterial colonies cultivated from bite marks [20]. Thus, the numbers of distinct strains identified in each sample by DGGE was very similar to those determined by the colony-genotyping approach, implying that the direct PCR method amplifies and distinguishes the majority of strains present. The discriminatory limitation (between samples), however, is imposed by the DGGE technique. Comparison of amplicons among non-corresponding samples indicated that DGGE analysis using these primers does not exclusively identify the teeth responsible for generating a bite mark. On the other hand, bite mark evidence is rarely presented in the absence of other information implicating a suspect, and therefore this method could provide meaningful information toward identifying the assailant among a restricted number of suspects, which is generally the situation with assaults. The 16S rDNA primers selected in this study specifically amplify streptococcal DNA though this is a relatively conserved locus [22]. Conceivably, more discriminatory primers might be identified that could be used alternatively or in combination with the 16S rDNA primers to increase resolution.

Prominent amplicons from incisor samples were generated from porcine skin for up to 48 h (Fig. 3), but beyond this time amplicons matching those from the incisors were not reliably generated even following incubation at 4 °C, the most favorable condition for bacterial DNA persistence (Fig. 3). Inability to amplify incisor-derived streptococcal DNA may have been due to a number of factors, particularly the inevitable growth of other organisms on the deteriorating skin. Approximately one-third of control porcine skin samples generated amplicons with the streptococcus-specific primers (Fig. 3) and these were often prominent in the DGGE profiles derived from deteriorating skin segments, suggesting significant background DNA. Living human skin is generally washed regularly and likely maintains a lower bacterial load than porcine epidermis. Thus, tooth-derived streptococcal DNA could be detectable for longer periods on deceased human skin. Furthermore, the application of the incisor biofilm material to the skin segments involved two transfers, whereas a bite mark involves only one transfer and probably results in deposition of greater numbers of bacteria. Amplification of streptococcal DNA from human skin suggests that streptococcal species traditionally regarded as oral may be common on human skin [26], but very few streptococci can be cultured from the (unbitten) skin of living humans [11, 20], and sequences amplified directly from the skin likely derive from non-viable bacteria.

Fig. 3
figure 3

DGGE of streptococcal DNA amplified from human incisor biofilm material (derived from two individuals) deposited on porcine skin. M streptococcal reference amplicons. S untreated skin sample. T incisor sample. Numbered lanes contained amplified samples recovered after 24, 48, and 96 h incubation at 4 °C

Full size image

Compared to the previously reported method involving analysis of cultured streptococci by arbitrarily-primed PCR (AP-PCR) [20], the DGGE approach had lower resolution with some overlap of non-corresponding profiles. AP-PCR unequivocally identified biters according to the genotypic profiles of multiple colonies but has the disadvantage of requiring sampling and culturing within hours of biting because oral streptococci remain viable on human skin for only 24 h [20]. Unexpectedly, the direct PCR approach did not extend this temporal window of opportunity to an extent that would offer a major practical advantage over the culturing method. Whereas the time limit might be increased by 24 h, there is a concomitant loss of resolution of tooth-derived amplicons and increase in spurious products. From a forensic viewpoint, an appealing aspect of both techniques is that they provide visually presentable results that would be readily comprehensible to a jury. This is an important consideration and a potential advantage over technologically more advanced and sensitive techniques, such as pyro-sequencing, which require considerable data processing and specialist interpretation [27].

In conclusion, the study provides support for a microbiologically based approach to the analysis of bite marks. Whereas recovery and analysis of human DNA (derived from the biter) should always be given priority, when this is unsuccessful the same sample (swabbed from the bite mark) also offers the opportunity for microbial analysis by amplifying streptococcal DNA for comparison with amplicons derived from the dentition of a suspect.