Forensic Science International: Genetics 12 (2014) 93–99
Contents lists available at ScienceDirect
Forensic Science International: Genetics journal homepage: www.elsevier.com/locate/fsig
A simple identification method for vaginal secretions using relative quantification of Lactobacillus DNA Masanori Doi a,*, Shinsuke Gamo a, Tatsuyuki Okiura a,b, Hiroaki Nishimukai b, Migiwa Asano b a b
Forensic Science Laboratory, Ehime Prefectural Police Headquarters, 2-2 Minamihoribatacho, Matsuyama, Ehime 790-8573, Japan Department of Legal Medicine, Ehime University Graduate School of Medicine, Shitsukawa, Toon, Ehime 791-0295, Japan
A R T I C L E I N F O
A B S T R A C T
Article history: Received 8 December 2013 Received in revised form 7 May 2014 Accepted 8 May 2014
In criminal investigations there are some cases in which identifying the presence of vaginal secretions provides crucial evidence in proving sexual assault. However, there are no methods for definitively identifying vaginal secretions. In the present study, we focused on Lactobacillus levels in vaginal secretions and developed a novel identification method for vaginal secretions by relative quantification based on real time PCR. We designed a Lactobacillus conserved region primer pair (LCP) by aligning 16S rRNA gene sequences from major vaginal Lactobacillus species (Lactobacillus crispatus, Lactobacillus gasseri, Lactobacillus iners and Lactobacillus jensenii), and selected the human specific primer pair (HSP) as an endogenous control for relative quantification. As a result, the DCt (DCt = Ct[LCP] Ct[HSP]) values of vaginal secretions (11 out of 12 samples) were significantly lower than those of saliva, semen and skin surface samples, and it was possible to discriminate between vaginal secretions and other body fluids. For the one remaining sample, it was confirmed that the predominant species in the microflora was not of the Lactobacillus genus. The DCt values in this study were calculated when the total DNA input used from the vaginal secretions was 10 pg or more. Additionally, the DCt values of samples up to 6-monthsold, which were kept at room temperature, remained unchanged. Thus, we concluded in this study that the simple DCt method by real time PCR is a useful tool for detecting the presence of vaginal secretions. ß 2014 Elsevier Ireland Ltd. All rights reserved.
Keywords: Vaginal secretions Body fluid identification Lactobacillus Real time PCR Microflora Human DNA
1. Introduction At a crime scene the biological substances found are examined for tissue origin, and DNA typing is performed. Most forensically relevant body fluids can be identified by conventional immunological and enzymatic tests, however, there are no methods for definitively identifying vaginal secretions. In recent years, body fluid identification methods using molecular biological techniques have been studied. Specifically, there are many reports about a new method based on tissue-specific expression called mRNA profiling [1–4]. MUC4 (mucin 4) and HBD1 (human b-defensin 1) have been used as vaginal secretion-specific markers in most of these studies, but MUC4 and HBD1 have also been found to be expressed in saliva [2,5–7]. Moreover, when recent genome-wide mRNA profiling experiments were performed to find specific mRNA markers for four types of body fluids (blood, saliva, semen
* Corresponding author. Tel.: +81 89 934 0110. E-mail address: [email protected] (M. Doi). http://dx.doi.org/10.1016/j.fsigen.2014.05.005 1872-4973/ß 2014 Elsevier Ireland Ltd. All rights reserved.
and vaginal secretions), vaginal secretions showed the most heterogeneous gene expression patterns [8]. Thus, highly specific biomarkers for vaginal fluid have not been discovered to date. If vaginal secretions could be identified, it may, in certain circumstances, provide important probative evidence of sexual assault. For example, the identification of vaginal secretions (not other body fluids) on a suspect’s penis could indicate vaginal sexual intercourse and, identifying vaginal secretions on the suspect’s hands and implements could provide crucial evidence for reconstructing a sexual assault. Recently in the forensic community, methods using bacteria as an indicator of the presence of specific body fluids have been studied [9–13]. The reported indicators for vaginal secretions were Lactobacillus species [11–13]. A woman’s vagina is kept healthy by Lactobacilli [14]. This is because Lactobacilli produce lactic acid, hydrogen peroxide, bacteriocins and other antimicrobial substances to inhibit pathogenic organisms in the vagina. From previous studies of the vaginal microbiome, it is understood that there are four major Lactobacillus species; Lactobacillus crispatus, Lactobacillus gasseri, Lactobacillus iners and Lactobacillus jensenii
94
M. Doi et al. / Forensic Science International: Genetics 12 (2014) 93–99
[15–17]. However, these Lactobacillus species are present not only in the vagina; Lactobacilli are also ubiquitous in the environment. L. crispatus are found in the oral cavity [18] and L. gasseri are used in commercial yogurt [19]. Additionally, there is variation in the microbiome between individuals. Hence, identifying vaginal secretions by detecting only a few Lactobacillus species is considered difficult [20]. In the present study, we focused on levels of Lactobacillus DNA from vaginal secretions. We designed a primer pair in the Lactobacillus conserved region in 16S rRNA gene sequences for real-time PCR measures and tried to discriminate between vaginal secretions and other body fluids by a relative quantification (DCt) method standardized using human DNA. The results showed that DCt values of vaginal secretions were lower than that of other body fluids, and that the DCt method could be used to identify vaginal secretions. 2. Materials and methods
Fig. 1. Sliding window plot of nucleotide diversity among 54 sequences of 16S rRNA gene from four major vaginal Lactobacillus species. This plot was obtained with a window length of 20 sites and a step size of 1 site. Arrows F and R indicate the positions of the Lactobacillus conserved region primer pair (LCP) in the present study.
and reverse primer 50 -GTTAGCTGCAGCACTGAGAG (Lactobacillus conserved region primer pair: LCP; amplicon size, 205 bp).
2.1. Samples 2.3. DNA extraction Vaginal secretions (n = 12), saliva (n = 19), semen (n = 8), urine (n = 4) and skin surface (n = 7) samples were collected from healthy volunteers. Informed consent was obtained from all the participants who provided samples. Vaginal secretions and saliva were collected by wiping with a cotton swab, and skin surface samples were obtained by soaking on sterilized gauze. Semen and urine samples were collected in paper cups, and then, a 3 cm 3 cm gauze soaked in each fluid sample was used for the analysis. Vaginal secretions were collected regardless of the phase of the menstrual cycle (except during menstruation), and saliva samples were collected at least 1 h after food had last been consumed and before participants had rinsed their mouths. Skin surface samples were collected from the sudation part of the body (forehead, neck or tip of the nose). All samples were dried and frozen at 80 8C until the extraction of DNA. All the vaginal secretion samples were numbered (VS1-12) for the following tests of sensitivity and stability. To test sensitivity, extracted DNA from vaginal secretions (n = 3) were serially diluted as follows: 0.5 ng/ml, 0.05 ng/ml, 5 pg/ ml and 0.5 pg/ml. To test sample stability over time, multiple vaginal secretions on cotton swabs (n = 1) were kept at room temperature in the laboratory for the following periods: 3 days, 1 week, 15 days, 1 month, 2 months, 3 months and 6 months. As a reference for mixture samples, we mixed DNA solutions from a vaginal secretion sample (n = 1, 1 ng/ml) and from a semen sample (n = 1, 1 ng/ml) with the following ratios: 9:1, 7:3, 5:5, 3:7 and 1:9. 2.2. Primer design We aligned a total of 54 sequences of the 16S rRNA gene from four major vaginal Lactobacillus species which were retrieved from the National Center for Biotechnology Information web site (17 L. crispatus, 23 L. gasseri, 4 L. iners and 10 L. jensenii isolates) used by CLUSTAL W implemented in MEGA version 4 software [21]. Then the average numbers of nucleotide differences per site (nucleotide diversity) were calculated by sliding window analysis using DnaSP v4.10.9 software [22] and conserved regions among aligned sequences (low nucleotide diversity) were identified (Fig. 1). As amplification DNA of variable Lactobacillus isolates, some primer pairs were designed in the conserved regions to be around 200 bp amplicon. In this study we used a primer pair which obtained the best results in a preliminary experiment with some vaginal samples (data not shown): forward primer 50 -AGAGGAGAGTGGAACTCCATG,
Total DNA were extracted from the body fluid samples. In each extraction, a 2 mm 2 mm cotton swab or a 5 mm 5 mm square of gauze was used. All body fluid samples were treated with lysozyme which lyses Gram-positive bacteria, such as Lactobacillus and then total DNA were extracted using the EZ1 DNA Investigator kit (Qiagen, Hilden, Germany). The procedure in detail is as follows: each sample was incubated with 100 ml of 10 mg/ml lysozyme (Sigma–Aldrich Japan, Tokyo) at 25 8C for 1 h, followed by treatment with 10 ml of Proteinase K (Qiagen) and 90 ml of Buffer GE at 568 C for 1 h (8 ml of 1 M DTT was added only to semen samples for lysing sperm). Total DNA was isolated using the EZ1 DNA Investigator kit on the EZ1 BioRobotTM (Qiagen). The DNA was eluted with 50 ml of TE buffer and its concentration was measured with a NanoDrop-1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA); if its concentration was >10 ng/ml, the DNA was diluted to 1 ng/ml. 2.4. Real-time PCR and data analysis Quantitative real-time PCR (qPCR) reactions in this study were performed with SYBR1 Premix Ex TaqTM (TAKARA, Tokyo, Japan) on a Smart Cycler II System (Cepheid, Sunnyvale, CA) using the following parameters: initial denaturation at 95 8C for 10 s, followed by 40 cycles of 95 8C for 5 s, 55 8C for 20 s and 72 8C for 15 s and a final melting step from 60 8C to 95 8C. In each qPCR reaction, 2 ml of extracted DNA was used as input. The cycle threshold (Ct) value which resulted from the real-time PCR was defined by the second derivative method on Smart cycler software. For testing amplification efficiency of the LCP, a L. crispatus strain (ATCC #33820) was obtained from ATCC (American Type Culture Collection) and used. The L. crispatus strain was cultured in MRS agar and DNA was extracted. Near 100% PCR efficiency of the LCP was verified by using a standard curve analysis of L. crispatus DNA dilutions (Suppl. Fig. 1). Comparative analyses between forensically relevant body fluids were done with the delta Ct (DCt) method. We selected human DNA as an endogenous control for relative quantification of target Lactobacillus DNA. The human DNA quantification was performed using the human specific primer pair (HSP); forward primer, 50 -TTTTGCAGGATCTACAAGTGGA, and reverse primer, 50 -AAGAGGTCTACATGTCCCCTTG (amplicon size, 207 bp) [23]. The target of HSP is D17Z1 which is known as a human specific DNA sequence. Ct values obtained from the
M. Doi et al. / Forensic Science International: Genetics 12 (2014) 93–99
LCP and the HSP were named Ct[LCP] and Ct[HSP], respectively; DCt = Ct[LCP] Ct[HSP]. If a DCt value is small, a high level of target Lactobacillus DNA relative to human DNA is conjectured. The amplification efficiency of HSP was verified to near 100% using serial dilutions of human genomic DNA (TAKARA), and Ct[HSP] > 29 were considered to be below the limit of detection since a Ct[HSP] of 29 represents DNA lower than genome DNA in a cell (6.4 pg) (Suppl. Fig. 2). 2.5. Analysis of the dominant bacteria species To investigate the dominant species in vaginal secretions, the microflora of six vaginal secretion samples were analyzed using the clone library method. Bacterial genomic DNA was extracted using a MORA-EXTRACT kit (AMR, Gifu, Japan) in which zirconium beads physically crush bacteria, according to the manufacturer’s protocol. PCR reactions were performed with TaKaRa LA Taq (TAKARA) according to the manufacturer’s specifications; however, amplification was 25 cycles. To amplify an almost full-length 16S rRNA gene, the bacteria universal primer pair conventionally called 8f (50 -AGAGTTTGATCMTGGCTCAG) and 1492r (50 -TACGGYTACCTTGTTACGACTT) was used. The 1.4-kb amplified products were checked with agarose gel electrophoresis and cloned into a plasmid vector using a TOPO-TA cloning kit, as described by the manufacturer (Invitrogen by Life Technologies, Carlsbad, CA), and the library was transformed into Escherichia coli cells. The plasmid inserts from individual colonies were PCR-amplified for sequencing using universal primers equivalent to plasmid vector sequences. The amplified plasmid inserts were sequenced from both ends using BigDye Terminator Cycle Sequencing Kit v1.1 (Applied Biosystems by Life Technologies, Foster City, CA) and Applied Biosystems 3130xl Genetic Analyzer (Applied Biosystems). In each vaginal sample we determined the 16S rRNA gene sequences of 10 clones, and the obtained sequences were compared with data in GenBank.
3. Results 3.1. Comparison of DCt values between vaginal secretions and other body fluids Both of the Ct values used by LCP and HSP (Ct[LCP] or Ct[HSP], respectively) were examined in vaginal secretions (n = 12), saliva (n = 19), semen (n = 8), skin surface samples (n = 7) and urine (n = 4), followed by calculation of DCt in each sample. The results of vaginal secretions and other body fluids are shown in Tables 1 and 2, respectively. The calculated DCt values for each of the samples are shown in Fig. 2. It was possible to calculate DCt values in all vaginal secretion samples (Table 1). However, the DCt values Table 1 Ct[LCP], Ct[HSP] and DCt obtained from real time PCR of vaginal secretions. Sample
Ct[LCP]
Ct[HSP]
DCt
VS1 VS2 VS3 VS4 VS5 VS6 VS7 VS8 VS9 VS10 VS11 VS12
21.83 24.78 21.11 24.7 26.06 33.83 20.3 17.09 22.97 23 19.5 20.5
13.78 20.14 20.14 21.75 19.88 16.08 17.75 13.77 18.65 21.96 16.08 17.21
8.05 4.64 0.97 2.95 6.18 17.75 2.55 3.32 4.32 1.04 3.42 3.29
LCP, Lactobacillus conserved region primer pair; HSP, human specific primer pair.
95
Table 2 Ct[LCP], Ct[HSP] and DCt values obtained from real time PCR in various body fluids. Sample
Sex
Ct[LCP]
Ct[HSP]
DCt
Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva
Female Female Female Female Female Male Male Male Male Male Female Female Female Female Female Female Female Male Female
37.35 35 34.31 ND 35.79 ND 36.23 34.9 34.89 36.87 36.36 ND 32.05 35.8 29.86 32.83 36.26 29.75 28.5
24.25 20.41 21.4 25.03 21.73 23.77 22.93 21 23.54 23.69 22.17 19.58 15.98 19.5 19.86 18.17 25.23 19.83 16.68
13.1 14.59 12.91 – 14.06 – 13.3 13.9 11.35 13.18 14.19 – 16.07 16.3 10 14.66 11.03 9.92 11.82
Male Male Male Male Male Male Male Male
ND ND ND 35.94 35.85 ND 31.94 35.27
20.72 22.53 24.55 23.25 20.5 20.19 16.21 17.75
– – – 12.69 15.35 – 15.73 17.52
Female Male Male Male Male Male Female
34.85 32.76 27 ND 34.42 31.75 33.78
25.14 19.86 17.73 25.6 22.23 21.75 20.25
9.71 12.9 9.27 – 12.19 10 13.53
Female Male Female Female
29.93 37.41 27.89 19.3
24.96 29.21(>29) 24.79 18.5
Male Female Female
ND 35.3 ND
24.3 19.5 ND
#1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15 #16 #17 #18 #19
Semen Semen Semen Semen Semen Semen Semen Semen Skin Skin Skin Skin Skin Skin Skin
#1 #2 #3 #4 #5 #6 #7 #8
surface surface surface surface surface surface surface
Urine Urine Urine Urine
#1 #2 #3 #4 #5 #6 #7
#1 #2 #3 #4
Additional samples Nasal secretions Menstrual blood Feces
4.97 – 3.1 0.8
– 15.82 –
ND, not detected; –, unmeasurable.
in 3 samples of saliva (16%), 4 samples of semen (50%) and 1 skin surface sample (14%) could not be calculated as the Ct[LCP] values of these samples were not detected (Table 2). We also were not able to calculate DCt in the one male urine sample taken from one male (the other three were from females), because human DNA in the sample was below the limit of detection (Ct[HSP] > 29). The samples in which there were no DCt values were removed from subsequent statistical analysis, although the numbers of these samples are noted in Fig. 2. As expected, DCt values in all vaginal secretions except one were lower than those of saliva, semen and skin surface samples. However the DCt values in female urine samples were also low as were vaginal secretions. This may be explained by contamination of vaginal secretions, as these urine samples are from women. The average DCt values were 4.87 (SD = 4.51) for vaginal secretions, 13.15 (SD = 1.92) for saliva, 15.32 (SD = 1.99) for semen, 11.27 (SD = 1.99) for skin surface samples and 2.96 (SD = 2.09) for urine. A t-test was performed for vaginal secretions and saliva samples, which had more than 10 DCt values. The result was t(26) = 6.61 and showed a significant difference (p < 0.01) between the two sample groups. Additionally, we established a boundary value between vaginal secretions and saliva by utilizing the relative cumulative frequency distribution
M. Doi et al. / Forensic Science International: Genetics 12 (2014) 93–99
96
Table 4 Ct[LCP], Ct[HSP] and DCt values obtained from real time PCR using different DNA input levels (n = 3; VS1, 3 and 5). The samples are the same as those shown in Table 1.
Fig. 2. DCt values in various body fluids. The numbers in square brackets ([]) indicate the number of samples in which Ct[LCP] was not detected (*: Ct[HSP] was not detected).
Vaginal sample
Total DNA input (pg)
Ct[LCP]
Ct[HSP]
DCt
VS1
1000 100 10 1
28.72 32.3 36.14 ND
20.33 23.94 27.58 ND
8.39 8.36 8.56 –
VS3
1000 100 10 1
22.2 25.89 29.62 33.11
21.43 25.06 28.5 ND
0.77 0.83 1.12 –
VS5
1000 100 10 1
26.93 30.42 34.5 ND
20.89 24.55 28.5 ND
6.04 5.87 6 –
ND, not detected; –, unmeasurable; LCP, Lactobacillus conserved region primer pair; HSP, human specific primer pair.
and 5); the results are shown in Table 4. In all three samples, the (RCD method) and defined 9 as the border for discriminating vaginal secretions and other body fluids in this study (Suppl. Fig. 3). This boundary value is lower than all of the DCt values in saliva, semen and skin surface samples. Thus, if the DCt value are 9 and the dominant species was not of the Lactobacillus genus. In vaginal microflora research, it has been found that some healthy women lack a Lactobacillus-dominant vaginal microflora, and instead other lacto-acid producing species have been identified as the dominant vaginal phylotype [27,28]. These women’s vaginal secretions may
97
also be DCt > 9. Thus, even if DCt is >9, it is impossible to prove the absence of vaginal secretions. Our analysis of the dominant species in each sample also provided additional information. Streptococcus salivarius was found in one vaginal secretion sample (Table 3), and this species has been used as an indicator of saliva presence by Nakanishi et al. [9]. They described that S. salivarius may be present not only in saliva but also in vaginal secretions. Despite this presence, it would actually be difficult to unambiguously identify body fluid by only detecting some specific bacteria species. Benschop et al. [20] comprehensively explored the vaginal microbiome using nextgeneration sequencing and showed that no candidate genera/ species were found to positively identify all vaginal samples. Hence, we believe that many forensic biologists will be interested in our novel approach as it uses the level of Lactobacillus to discriminate between vaginal secretions and other body fluids. Recently, body fluid identification based on the body-fluidspecific expression of mRNA or microRNA has been investigated, and its usefulness has been reported. However, some body-fluidspecific biomarkers were found by previous studies to be unclear as to why their function is expressed only in specific body fluids. Furthermore, these biomarkers have potentially differing expression levels between human populations. Park et al. [8] also reports that HBA1 and PMR1, which are known as body fluid specific mRNA markers (for blood and semen, respectively), did not show the body-fluid-specific expression pattern in Korean samples. In contrast, it is well known that a large number of Lactobacilli colonize most vaginas as beneficial bacteria. Furthermore, it has been reported that most women worldwide are colonized by three common Lactobacillus species [15]. Empirical support of a Lactobacillus-based identification method for vaginal secretions would be important in forensic science, particularly in cases where explanation to jurors in court is necessary. Ideal body fluid identification in forensic practice should be as easy as possible to use in addition to being highly sensitive and stable. In a collaborative exercise organized by multiple laboratories, each mRNA multiplex identification method for blood, saliva, or sperm proved to be reproducible and sensitive [29–31]. Therefore, the application of mRNA-based methods in forensic practice became technically possible and worthy of using in forensic science laboratories widely. However, mRNA multiplex systems require many steps: extraction, reverse transcription, quantification, PCR, and capillary electrophoresis. This technique is more difficult and complex than conventional serology-based methods. In contrast, the DCt method protocol in this study is a very easy process; it involves only adding lysozyme treatment before DNA extraction. Furthermore, the DCt method showed good reproducibility using only 10 pg of total DNA, and thus the method has high sensitivity. Moreover, the DCt values from samples kept at room temperature hardly changed over 6 months, showing good stability. We predicted that the DCt value would remain stable over time because the target of this method is Lactobacillus and human DNA which is known to be more stable than RNA in forensic science. Some research groups have detected microorganisms using DNA extracts obtained using only the human DNA extraction method [12,20]; however, when lysozyme treatment was not performed in our preliminary experiments, the reproducibility of the DCt value was not high (data not shown). The reason may be the unstable extraction rate of bacterial DNA when it is recovered using only the human DNA extraction method. Accordingly, we conclude that the lysozyme treatment is essential for our relative quantification method (not detection method). At first, we planned to amplify only Lactobacillus DNA, however, it was difficult to amplify the Lactobacillus-specific DNA because a Lactobacillus-specific gene has not been found in previous
98
M. Doi et al. / Forensic Science International: Genetics 12 (2014) 93–99
comparative genomics studies [32]. Thus, we selected the 16S rRNA gene for species identification. The LCP used in this study was designed based on the Lactobacillus conserved region of the 16S rRNA gene for the purpose of amplifying DNA from variable Lactobacillus strains in the vagina. Therefore, the LCP are not Lactobacillus-specific primers, so the amplicons obtained by LCP are not solely from Lactobacillus DNA. Nevertheless, based on our results, we conclude that the LCP can be used broadly to indicate the level of Lactobacillus DNA in a sample, and it has specificity for detecting the presence of vaginal secretions. The following situations are examples of where differing circumstances may impact the outcomes of using the DCt approach to indicate the possible presence of vaginal secretions. Although Ct[LCP] was detected in commercial yogurt, the DCt of the yogurt could not be calculated because Ct[HSP] was not detected. However, the saliva collected from a volunteer who had eaten the yogurt 1 h before saliva collection had DCt = 6.75 (
Contents lists available at ScienceDirect
Forensic Science International: Genetics journal homepage: www.elsevier.com/locate/fsig
A simple identification method for vaginal secretions using relative quantification of Lactobacillus DNA Masanori Doi a,*, Shinsuke Gamo a, Tatsuyuki Okiura a,b, Hiroaki Nishimukai b, Migiwa Asano b a b
Forensic Science Laboratory, Ehime Prefectural Police Headquarters, 2-2 Minamihoribatacho, Matsuyama, Ehime 790-8573, Japan Department of Legal Medicine, Ehime University Graduate School of Medicine, Shitsukawa, Toon, Ehime 791-0295, Japan
A R T I C L E I N F O
A B S T R A C T
Article history: Received 8 December 2013 Received in revised form 7 May 2014 Accepted 8 May 2014
In criminal investigations there are some cases in which identifying the presence of vaginal secretions provides crucial evidence in proving sexual assault. However, there are no methods for definitively identifying vaginal secretions. In the present study, we focused on Lactobacillus levels in vaginal secretions and developed a novel identification method for vaginal secretions by relative quantification based on real time PCR. We designed a Lactobacillus conserved region primer pair (LCP) by aligning 16S rRNA gene sequences from major vaginal Lactobacillus species (Lactobacillus crispatus, Lactobacillus gasseri, Lactobacillus iners and Lactobacillus jensenii), and selected the human specific primer pair (HSP) as an endogenous control for relative quantification. As a result, the DCt (DCt = Ct[LCP] Ct[HSP]) values of vaginal secretions (11 out of 12 samples) were significantly lower than those of saliva, semen and skin surface samples, and it was possible to discriminate between vaginal secretions and other body fluids. For the one remaining sample, it was confirmed that the predominant species in the microflora was not of the Lactobacillus genus. The DCt values in this study were calculated when the total DNA input used from the vaginal secretions was 10 pg or more. Additionally, the DCt values of samples up to 6-monthsold, which were kept at room temperature, remained unchanged. Thus, we concluded in this study that the simple DCt method by real time PCR is a useful tool for detecting the presence of vaginal secretions. ß 2014 Elsevier Ireland Ltd. All rights reserved.
Keywords: Vaginal secretions Body fluid identification Lactobacillus Real time PCR Microflora Human DNA
1. Introduction At a crime scene the biological substances found are examined for tissue origin, and DNA typing is performed. Most forensically relevant body fluids can be identified by conventional immunological and enzymatic tests, however, there are no methods for definitively identifying vaginal secretions. In recent years, body fluid identification methods using molecular biological techniques have been studied. Specifically, there are many reports about a new method based on tissue-specific expression called mRNA profiling [1–4]. MUC4 (mucin 4) and HBD1 (human b-defensin 1) have been used as vaginal secretion-specific markers in most of these studies, but MUC4 and HBD1 have also been found to be expressed in saliva [2,5–7]. Moreover, when recent genome-wide mRNA profiling experiments were performed to find specific mRNA markers for four types of body fluids (blood, saliva, semen
* Corresponding author. Tel.: +81 89 934 0110. E-mail address: [email protected] (M. Doi). http://dx.doi.org/10.1016/j.fsigen.2014.05.005 1872-4973/ß 2014 Elsevier Ireland Ltd. All rights reserved.
and vaginal secretions), vaginal secretions showed the most heterogeneous gene expression patterns [8]. Thus, highly specific biomarkers for vaginal fluid have not been discovered to date. If vaginal secretions could be identified, it may, in certain circumstances, provide important probative evidence of sexual assault. For example, the identification of vaginal secretions (not other body fluids) on a suspect’s penis could indicate vaginal sexual intercourse and, identifying vaginal secretions on the suspect’s hands and implements could provide crucial evidence for reconstructing a sexual assault. Recently in the forensic community, methods using bacteria as an indicator of the presence of specific body fluids have been studied [9–13]. The reported indicators for vaginal secretions were Lactobacillus species [11–13]. A woman’s vagina is kept healthy by Lactobacilli [14]. This is because Lactobacilli produce lactic acid, hydrogen peroxide, bacteriocins and other antimicrobial substances to inhibit pathogenic organisms in the vagina. From previous studies of the vaginal microbiome, it is understood that there are four major Lactobacillus species; Lactobacillus crispatus, Lactobacillus gasseri, Lactobacillus iners and Lactobacillus jensenii
94
M. Doi et al. / Forensic Science International: Genetics 12 (2014) 93–99
[15–17]. However, these Lactobacillus species are present not only in the vagina; Lactobacilli are also ubiquitous in the environment. L. crispatus are found in the oral cavity [18] and L. gasseri are used in commercial yogurt [19]. Additionally, there is variation in the microbiome between individuals. Hence, identifying vaginal secretions by detecting only a few Lactobacillus species is considered difficult [20]. In the present study, we focused on levels of Lactobacillus DNA from vaginal secretions. We designed a primer pair in the Lactobacillus conserved region in 16S rRNA gene sequences for real-time PCR measures and tried to discriminate between vaginal secretions and other body fluids by a relative quantification (DCt) method standardized using human DNA. The results showed that DCt values of vaginal secretions were lower than that of other body fluids, and that the DCt method could be used to identify vaginal secretions. 2. Materials and methods
Fig. 1. Sliding window plot of nucleotide diversity among 54 sequences of 16S rRNA gene from four major vaginal Lactobacillus species. This plot was obtained with a window length of 20 sites and a step size of 1 site. Arrows F and R indicate the positions of the Lactobacillus conserved region primer pair (LCP) in the present study.
and reverse primer 50 -GTTAGCTGCAGCACTGAGAG (Lactobacillus conserved region primer pair: LCP; amplicon size, 205 bp).
2.1. Samples 2.3. DNA extraction Vaginal secretions (n = 12), saliva (n = 19), semen (n = 8), urine (n = 4) and skin surface (n = 7) samples were collected from healthy volunteers. Informed consent was obtained from all the participants who provided samples. Vaginal secretions and saliva were collected by wiping with a cotton swab, and skin surface samples were obtained by soaking on sterilized gauze. Semen and urine samples were collected in paper cups, and then, a 3 cm 3 cm gauze soaked in each fluid sample was used for the analysis. Vaginal secretions were collected regardless of the phase of the menstrual cycle (except during menstruation), and saliva samples were collected at least 1 h after food had last been consumed and before participants had rinsed their mouths. Skin surface samples were collected from the sudation part of the body (forehead, neck or tip of the nose). All samples were dried and frozen at 80 8C until the extraction of DNA. All the vaginal secretion samples were numbered (VS1-12) for the following tests of sensitivity and stability. To test sensitivity, extracted DNA from vaginal secretions (n = 3) were serially diluted as follows: 0.5 ng/ml, 0.05 ng/ml, 5 pg/ ml and 0.5 pg/ml. To test sample stability over time, multiple vaginal secretions on cotton swabs (n = 1) were kept at room temperature in the laboratory for the following periods: 3 days, 1 week, 15 days, 1 month, 2 months, 3 months and 6 months. As a reference for mixture samples, we mixed DNA solutions from a vaginal secretion sample (n = 1, 1 ng/ml) and from a semen sample (n = 1, 1 ng/ml) with the following ratios: 9:1, 7:3, 5:5, 3:7 and 1:9. 2.2. Primer design We aligned a total of 54 sequences of the 16S rRNA gene from four major vaginal Lactobacillus species which were retrieved from the National Center for Biotechnology Information web site (17 L. crispatus, 23 L. gasseri, 4 L. iners and 10 L. jensenii isolates) used by CLUSTAL W implemented in MEGA version 4 software [21]. Then the average numbers of nucleotide differences per site (nucleotide diversity) were calculated by sliding window analysis using DnaSP v4.10.9 software [22] and conserved regions among aligned sequences (low nucleotide diversity) were identified (Fig. 1). As amplification DNA of variable Lactobacillus isolates, some primer pairs were designed in the conserved regions to be around 200 bp amplicon. In this study we used a primer pair which obtained the best results in a preliminary experiment with some vaginal samples (data not shown): forward primer 50 -AGAGGAGAGTGGAACTCCATG,
Total DNA were extracted from the body fluid samples. In each extraction, a 2 mm 2 mm cotton swab or a 5 mm 5 mm square of gauze was used. All body fluid samples were treated with lysozyme which lyses Gram-positive bacteria, such as Lactobacillus and then total DNA were extracted using the EZ1 DNA Investigator kit (Qiagen, Hilden, Germany). The procedure in detail is as follows: each sample was incubated with 100 ml of 10 mg/ml lysozyme (Sigma–Aldrich Japan, Tokyo) at 25 8C for 1 h, followed by treatment with 10 ml of Proteinase K (Qiagen) and 90 ml of Buffer GE at 568 C for 1 h (8 ml of 1 M DTT was added only to semen samples for lysing sperm). Total DNA was isolated using the EZ1 DNA Investigator kit on the EZ1 BioRobotTM (Qiagen). The DNA was eluted with 50 ml of TE buffer and its concentration was measured with a NanoDrop-1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA); if its concentration was >10 ng/ml, the DNA was diluted to 1 ng/ml. 2.4. Real-time PCR and data analysis Quantitative real-time PCR (qPCR) reactions in this study were performed with SYBR1 Premix Ex TaqTM (TAKARA, Tokyo, Japan) on a Smart Cycler II System (Cepheid, Sunnyvale, CA) using the following parameters: initial denaturation at 95 8C for 10 s, followed by 40 cycles of 95 8C for 5 s, 55 8C for 20 s and 72 8C for 15 s and a final melting step from 60 8C to 95 8C. In each qPCR reaction, 2 ml of extracted DNA was used as input. The cycle threshold (Ct) value which resulted from the real-time PCR was defined by the second derivative method on Smart cycler software. For testing amplification efficiency of the LCP, a L. crispatus strain (ATCC #33820) was obtained from ATCC (American Type Culture Collection) and used. The L. crispatus strain was cultured in MRS agar and DNA was extracted. Near 100% PCR efficiency of the LCP was verified by using a standard curve analysis of L. crispatus DNA dilutions (Suppl. Fig. 1). Comparative analyses between forensically relevant body fluids were done with the delta Ct (DCt) method. We selected human DNA as an endogenous control for relative quantification of target Lactobacillus DNA. The human DNA quantification was performed using the human specific primer pair (HSP); forward primer, 50 -TTTTGCAGGATCTACAAGTGGA, and reverse primer, 50 -AAGAGGTCTACATGTCCCCTTG (amplicon size, 207 bp) [23]. The target of HSP is D17Z1 which is known as a human specific DNA sequence. Ct values obtained from the
M. Doi et al. / Forensic Science International: Genetics 12 (2014) 93–99
LCP and the HSP were named Ct[LCP] and Ct[HSP], respectively; DCt = Ct[LCP] Ct[HSP]. If a DCt value is small, a high level of target Lactobacillus DNA relative to human DNA is conjectured. The amplification efficiency of HSP was verified to near 100% using serial dilutions of human genomic DNA (TAKARA), and Ct[HSP] > 29 were considered to be below the limit of detection since a Ct[HSP] of 29 represents DNA lower than genome DNA in a cell (6.4 pg) (Suppl. Fig. 2). 2.5. Analysis of the dominant bacteria species To investigate the dominant species in vaginal secretions, the microflora of six vaginal secretion samples were analyzed using the clone library method. Bacterial genomic DNA was extracted using a MORA-EXTRACT kit (AMR, Gifu, Japan) in which zirconium beads physically crush bacteria, according to the manufacturer’s protocol. PCR reactions were performed with TaKaRa LA Taq (TAKARA) according to the manufacturer’s specifications; however, amplification was 25 cycles. To amplify an almost full-length 16S rRNA gene, the bacteria universal primer pair conventionally called 8f (50 -AGAGTTTGATCMTGGCTCAG) and 1492r (50 -TACGGYTACCTTGTTACGACTT) was used. The 1.4-kb amplified products were checked with agarose gel electrophoresis and cloned into a plasmid vector using a TOPO-TA cloning kit, as described by the manufacturer (Invitrogen by Life Technologies, Carlsbad, CA), and the library was transformed into Escherichia coli cells. The plasmid inserts from individual colonies were PCR-amplified for sequencing using universal primers equivalent to plasmid vector sequences. The amplified plasmid inserts were sequenced from both ends using BigDye Terminator Cycle Sequencing Kit v1.1 (Applied Biosystems by Life Technologies, Foster City, CA) and Applied Biosystems 3130xl Genetic Analyzer (Applied Biosystems). In each vaginal sample we determined the 16S rRNA gene sequences of 10 clones, and the obtained sequences were compared with data in GenBank.
3. Results 3.1. Comparison of DCt values between vaginal secretions and other body fluids Both of the Ct values used by LCP and HSP (Ct[LCP] or Ct[HSP], respectively) were examined in vaginal secretions (n = 12), saliva (n = 19), semen (n = 8), skin surface samples (n = 7) and urine (n = 4), followed by calculation of DCt in each sample. The results of vaginal secretions and other body fluids are shown in Tables 1 and 2, respectively. The calculated DCt values for each of the samples are shown in Fig. 2. It was possible to calculate DCt values in all vaginal secretion samples (Table 1). However, the DCt values Table 1 Ct[LCP], Ct[HSP] and DCt obtained from real time PCR of vaginal secretions. Sample
Ct[LCP]
Ct[HSP]
DCt
VS1 VS2 VS3 VS4 VS5 VS6 VS7 VS8 VS9 VS10 VS11 VS12
21.83 24.78 21.11 24.7 26.06 33.83 20.3 17.09 22.97 23 19.5 20.5
13.78 20.14 20.14 21.75 19.88 16.08 17.75 13.77 18.65 21.96 16.08 17.21
8.05 4.64 0.97 2.95 6.18 17.75 2.55 3.32 4.32 1.04 3.42 3.29
LCP, Lactobacillus conserved region primer pair; HSP, human specific primer pair.
95
Table 2 Ct[LCP], Ct[HSP] and DCt values obtained from real time PCR in various body fluids. Sample
Sex
Ct[LCP]
Ct[HSP]
DCt
Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva Saliva
Female Female Female Female Female Male Male Male Male Male Female Female Female Female Female Female Female Male Female
37.35 35 34.31 ND 35.79 ND 36.23 34.9 34.89 36.87 36.36 ND 32.05 35.8 29.86 32.83 36.26 29.75 28.5
24.25 20.41 21.4 25.03 21.73 23.77 22.93 21 23.54 23.69 22.17 19.58 15.98 19.5 19.86 18.17 25.23 19.83 16.68
13.1 14.59 12.91 – 14.06 – 13.3 13.9 11.35 13.18 14.19 – 16.07 16.3 10 14.66 11.03 9.92 11.82
Male Male Male Male Male Male Male Male
ND ND ND 35.94 35.85 ND 31.94 35.27
20.72 22.53 24.55 23.25 20.5 20.19 16.21 17.75
– – – 12.69 15.35 – 15.73 17.52
Female Male Male Male Male Male Female
34.85 32.76 27 ND 34.42 31.75 33.78
25.14 19.86 17.73 25.6 22.23 21.75 20.25
9.71 12.9 9.27 – 12.19 10 13.53
Female Male Female Female
29.93 37.41 27.89 19.3
24.96 29.21(>29) 24.79 18.5
Male Female Female
ND 35.3 ND
24.3 19.5 ND
#1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15 #16 #17 #18 #19
Semen Semen Semen Semen Semen Semen Semen Semen Skin Skin Skin Skin Skin Skin Skin
#1 #2 #3 #4 #5 #6 #7 #8
surface surface surface surface surface surface surface
Urine Urine Urine Urine
#1 #2 #3 #4 #5 #6 #7
#1 #2 #3 #4
Additional samples Nasal secretions Menstrual blood Feces
4.97 – 3.1 0.8
– 15.82 –
ND, not detected; –, unmeasurable.
in 3 samples of saliva (16%), 4 samples of semen (50%) and 1 skin surface sample (14%) could not be calculated as the Ct[LCP] values of these samples were not detected (Table 2). We also were not able to calculate DCt in the one male urine sample taken from one male (the other three were from females), because human DNA in the sample was below the limit of detection (Ct[HSP] > 29). The samples in which there were no DCt values were removed from subsequent statistical analysis, although the numbers of these samples are noted in Fig. 2. As expected, DCt values in all vaginal secretions except one were lower than those of saliva, semen and skin surface samples. However the DCt values in female urine samples were also low as were vaginal secretions. This may be explained by contamination of vaginal secretions, as these urine samples are from women. The average DCt values were 4.87 (SD = 4.51) for vaginal secretions, 13.15 (SD = 1.92) for saliva, 15.32 (SD = 1.99) for semen, 11.27 (SD = 1.99) for skin surface samples and 2.96 (SD = 2.09) for urine. A t-test was performed for vaginal secretions and saliva samples, which had more than 10 DCt values. The result was t(26) = 6.61 and showed a significant difference (p < 0.01) between the two sample groups. Additionally, we established a boundary value between vaginal secretions and saliva by utilizing the relative cumulative frequency distribution
M. Doi et al. / Forensic Science International: Genetics 12 (2014) 93–99
96
Table 4 Ct[LCP], Ct[HSP] and DCt values obtained from real time PCR using different DNA input levels (n = 3; VS1, 3 and 5). The samples are the same as those shown in Table 1.
Fig. 2. DCt values in various body fluids. The numbers in square brackets ([]) indicate the number of samples in which Ct[LCP] was not detected (*: Ct[HSP] was not detected).
Vaginal sample
Total DNA input (pg)
Ct[LCP]
Ct[HSP]
DCt
VS1
1000 100 10 1
28.72 32.3 36.14 ND
20.33 23.94 27.58 ND
8.39 8.36 8.56 –
VS3
1000 100 10 1
22.2 25.89 29.62 33.11
21.43 25.06 28.5 ND
0.77 0.83 1.12 –
VS5
1000 100 10 1
26.93 30.42 34.5 ND
20.89 24.55 28.5 ND
6.04 5.87 6 –
ND, not detected; –, unmeasurable; LCP, Lactobacillus conserved region primer pair; HSP, human specific primer pair.
and 5); the results are shown in Table 4. In all three samples, the (RCD method) and defined 9 as the border for discriminating vaginal secretions and other body fluids in this study (Suppl. Fig. 3). This boundary value is lower than all of the DCt values in saliva, semen and skin surface samples. Thus, if the DCt value are 9 and the dominant species was not of the Lactobacillus genus. In vaginal microflora research, it has been found that some healthy women lack a Lactobacillus-dominant vaginal microflora, and instead other lacto-acid producing species have been identified as the dominant vaginal phylotype [27,28]. These women’s vaginal secretions may
97
also be DCt > 9. Thus, even if DCt is >9, it is impossible to prove the absence of vaginal secretions. Our analysis of the dominant species in each sample also provided additional information. Streptococcus salivarius was found in one vaginal secretion sample (Table 3), and this species has been used as an indicator of saliva presence by Nakanishi et al. [9]. They described that S. salivarius may be present not only in saliva but also in vaginal secretions. Despite this presence, it would actually be difficult to unambiguously identify body fluid by only detecting some specific bacteria species. Benschop et al. [20] comprehensively explored the vaginal microbiome using nextgeneration sequencing and showed that no candidate genera/ species were found to positively identify all vaginal samples. Hence, we believe that many forensic biologists will be interested in our novel approach as it uses the level of Lactobacillus to discriminate between vaginal secretions and other body fluids. Recently, body fluid identification based on the body-fluidspecific expression of mRNA or microRNA has been investigated, and its usefulness has been reported. However, some body-fluidspecific biomarkers were found by previous studies to be unclear as to why their function is expressed only in specific body fluids. Furthermore, these biomarkers have potentially differing expression levels between human populations. Park et al. [8] also reports that HBA1 and PMR1, which are known as body fluid specific mRNA markers (for blood and semen, respectively), did not show the body-fluid-specific expression pattern in Korean samples. In contrast, it is well known that a large number of Lactobacilli colonize most vaginas as beneficial bacteria. Furthermore, it has been reported that most women worldwide are colonized by three common Lactobacillus species [15]. Empirical support of a Lactobacillus-based identification method for vaginal secretions would be important in forensic science, particularly in cases where explanation to jurors in court is necessary. Ideal body fluid identification in forensic practice should be as easy as possible to use in addition to being highly sensitive and stable. In a collaborative exercise organized by multiple laboratories, each mRNA multiplex identification method for blood, saliva, or sperm proved to be reproducible and sensitive [29–31]. Therefore, the application of mRNA-based methods in forensic practice became technically possible and worthy of using in forensic science laboratories widely. However, mRNA multiplex systems require many steps: extraction, reverse transcription, quantification, PCR, and capillary electrophoresis. This technique is more difficult and complex than conventional serology-based methods. In contrast, the DCt method protocol in this study is a very easy process; it involves only adding lysozyme treatment before DNA extraction. Furthermore, the DCt method showed good reproducibility using only 10 pg of total DNA, and thus the method has high sensitivity. Moreover, the DCt values from samples kept at room temperature hardly changed over 6 months, showing good stability. We predicted that the DCt value would remain stable over time because the target of this method is Lactobacillus and human DNA which is known to be more stable than RNA in forensic science. Some research groups have detected microorganisms using DNA extracts obtained using only the human DNA extraction method [12,20]; however, when lysozyme treatment was not performed in our preliminary experiments, the reproducibility of the DCt value was not high (data not shown). The reason may be the unstable extraction rate of bacterial DNA when it is recovered using only the human DNA extraction method. Accordingly, we conclude that the lysozyme treatment is essential for our relative quantification method (not detection method). At first, we planned to amplify only Lactobacillus DNA, however, it was difficult to amplify the Lactobacillus-specific DNA because a Lactobacillus-specific gene has not been found in previous
98
M. Doi et al. / Forensic Science International: Genetics 12 (2014) 93–99
comparative genomics studies [32]. Thus, we selected the 16S rRNA gene for species identification. The LCP used in this study was designed based on the Lactobacillus conserved region of the 16S rRNA gene for the purpose of amplifying DNA from variable Lactobacillus strains in the vagina. Therefore, the LCP are not Lactobacillus-specific primers, so the amplicons obtained by LCP are not solely from Lactobacillus DNA. Nevertheless, based on our results, we conclude that the LCP can be used broadly to indicate the level of Lactobacillus DNA in a sample, and it has specificity for detecting the presence of vaginal secretions. The following situations are examples of where differing circumstances may impact the outcomes of using the DCt approach to indicate the possible presence of vaginal secretions. Although Ct[LCP] was detected in commercial yogurt, the DCt of the yogurt could not be calculated because Ct[HSP] was not detected. However, the saliva collected from a volunteer who had eaten the yogurt 1 h before saliva collection had DCt = 6.75 (
