Forensic Science International 183 (2009) 20–23

Contents lists available at ScienceDirect

Forensic Science International

journa l homepage: www.e lsevier .com/ locate / forsc i in t

A novel method for the identification of saliva by detecting oral streptococciusing PCR

Hiroaki Nakanishi a,d,*, Akira Kido b, Takeshi Ohmori c, Aya Takada d, Masaaki Hara d,Noboru Adachi b, Kazuyuki Saito d

a Forensic Science Laboratory of Yamanashi Prefectural Police H.Q., 312-4 Kubonakajima, Isawa, Fuefuki, Yamanashi 406-0036, Japanb Department of Legal Medicine, Faculty of Medicine, University of Yamanashi, 1110 Shimokato, Chuo, Yamanashi 409-3898, Japanc National Research Institute of Police Science, 6-3-1 Kashiwanoha, Kashiwa, Chiba 277-0882, Japand Department of Forensic Medicine, Saitama Medical University, 38 Morohongo, Moroyama, Saitama 350-0495, Japan

A R T I C L E I N F O

Article history:

Received 22 January 2008

Received in revised form 30 June 2008

Accepted 1 October 2008

Available online 4 November 2008

Keywords:

PCR

Saliva

Streptococcus salivarius

Streptococcus mutans

A B S T R A C T

We have used DNA amplification methods to detect common oral bacterial strains to test for the presence

of saliva in forensic samples. Streptococcus salivarius and Streptococcus mutans were detected in various

forms of saliva samples, whereas these streptococci were not detected in semen, urine, vaginal fluid, or on

skin surfaces. Therefore, we demonstrated that these streptococci are promising new marker for the

forensic identification of saliva. Our data indicated that S. salivarius is more reliable than S. mutans as an

indicator of saliva presence, because the detection rates for S. salivarius and S. mutans by this method

were 100% and 90%, respectively. Furthermore, S. salivarius was detected in all saliva stain samples,

whereas S. mutans was only identified in 60% of the stains. Finally, using this method we were able to

successfully detect S. salivarius and S. mutans in mock forensic samples. We therefore suggested that this

method is useful for the identification of saliva in forensic science.

� 2008 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Recent developments in forensic practices have contributedto the investigation of crimes and the verification of criminalidentity. The discrimination of body fluids in forensic examina-tions is important to determine the events that took place at acrime scene. The detection of saliva is particularly important forunderstanding the details of a crime. For example, the detectionof a criminal’s saliva left on a victim’s skin provides evidencethat the criminal had contact with a victim. Furthermore, thecriminal’s DNA may be isolated from the saliva, therebyverifying identity. The conventional method for testing for thepresence of saliva is still a test for the presence of a-amylase[1,2]. However, a-amylase may be present in other body fluidssuch as urine, and semen [3,4]. Therefore, it is necessary todevelop a new assay that can reliably discriminate saliva fromother body fluids. An RNA-based assay targeting saliva-specific

* Corresponding author at: Forensic Science Laboratory of Yamanashi Prefectural

Police H.Q., 312-4 Kubonakajima, Isawa, Fuefuki, Yamanashi 406-0036, Japan.

Tel.: +81 55 262 0082; fax: +81 55 262 0082.

E-mail address: nakanishi@384.jp (H. Nakanishi).

0379-0738/$ – see front matter � 2008 Elsevier Ireland Ltd. All rights reserved.

doi:10.1016/j.forsciint.2008.10.003

gene expression products was recently reported [5]. We alsoaimed to establish a further assay system for saliva that does notdepend on a-amylase.

Numerous bacteria exist in the oral cavity. For example, theaverage total microscopic count is approximately 750 million oralbacterial cells per milliliter of saliva [6]. Of these, streptococci arethe most abundant [7]. Specifically, Streptococcus salivarius is oneof the most common streptococci in oral bacteria [6], and S. mutans

is the prime causative bacterium for dental caries [8]. Thedetection and identification of oral bacteria has generally beenperformed by culturing samples on nutritive media. However,methods that incorporate the polymerase chain reaction (PCR)have recently been used to detect and identify oral bacteria such asS. salivarius or S. mutans, and have been applied in the diagnosis ofcaries risk [9–15].

With respect to forensics, the detection of oral streptococci hasonly been used to verify bite marks [16–18]. However, thesereports did not discuss the identification of saliva presence. If itwere possible to detect the presence of oral-specific bacteria byPCR from a forensic specimen, this could be used to verify thepresence of saliva. Therefore, we evaluated whether the detectionof S. salivarius and S. mutans by PCR is sufficient to confirm salivapresence.

Fig. 1. Sensitivity of the PCR method, examined using purified chromosomal DNA

from (a) S. salivarius and (b) S. mutans. The amount of template DNA used in each

lane was as follows: lane M, 100-bp molecular mass marker; lane 1, 1 ng; lane 2,

100 pg; lane 3, 50 pg; lane 4, 25 pg; lane 5, 10 pg; lane 6, 5 pg; lane 7, 2.5 pg; lane 8,

1 pg; lane 9, 500 fg.

H. Nakanishi et al. / Forensic Science International 183 (2009) 20–23 21

2. Materials and methods

2.1. Samples

Saliva, semen, and urine samples were collected from 20 healthy donors,

and saliva stain samples were made by licking filter paper. Skin bacteria

were collected from 20 healthy donors by wiping the skin with a wet cotton

swab. Vaginal fluid samples were collected from 9 healthy donors by cotton

swabbing. Informed consent was obtained from all participants who provided

samples.

The bacterial strains used in this study were S. salivarius ATCC 13419, S.

mutans ATCC 35668, S. mitis ATCC 6249, S. sanguinis ATCC 10556, Bacillus

subtilis ATCC 6633, Escherichia coli ATCC 35218, Pseudomonas aeruginosa

ATCC 27853, Serratia marcescens ATCC 8100, and Staphylococcus aureus

ATCC 25923. These strains were purchased from Microbiologics (St. Cloud,

MN). The streptococci were cultured in Mitis-Salivarius agar (BHI, Difco

Laboratories, Detroit, MI), and the remaining bacteria were cultured in Nutrient

agar (BHI).

2.2. DNA extractions

For the DNA extractions, we used 50 ml of body fluid, 5-mm � 5-mm filter

paper, or a swab in 50 ml of water. Control bacteria were suspended in 50 ml of

water. Samples were boiled at 98 8C in micro-centrifuge tubes. To lyse the

bacterial cells, samples were incubated with 100 ml of 200 U/ml mutanolysin

(Sigma, St. Louis, MO) and 20 ml of 100 mg/ml lysozyme (Sigma) at 50 8C for

60 min, followed by treatment with 20 ml of Proteinase K (Qiagen, Tokyo, Japan)

and 180 ml of Buffer ATL (Qiagen) at 56 8C for 60 min. We then added 200 ml of

Buffer AL (Qiagen) before incubating at 70 8C for 10 min and adding 200 ml of

ethanol. The DNA was purified using a QIAamp1 Mini Kit (Qiagen) according to

manufacturer’s protocol and eluted with 150 ml of water, which was later

concentrated to 30 ml.

2.3. PCR conditions

The oligonucleotide primers used in this study were designed by Hoshino

et al. [9] and are listed in Table 1. PCR was carried out in 20 ml reaction mixtures

containing 0.5 U AmpliTaqGold DNA polymerase (Applied Biosystems,

Foster city, CA), 0.5 mM oligonucleotide primers, template DNA, and Gold STAR

10� Buffer (Promega, Madison, WI). PCR amplification was performed in a

GeneAmp PCR system 9700 (9600 emulation mode; Applied Biosystems) pro-

grammed for 11 min at 95 8C and 35 cycles of 1 min at 94 8C for denaturation,

1 min at 65 8C for annealing, and 1 min at 72 8C for extension. Identical PCR

conditions were applied for both primer sets. PCR products were identified by

2.0% agarose gel electrophoresis after staining with SYBR Green I (TaKaRa,

Kyoto, Japan).

2.4. Conventional a-amylase detection

The starch-agarose method was used as a conventional form of a-amylase

detection. One tablet of the Neo-Amylase Test (Daiichi Pure Chemicals, Tokyo,

Japan) was crushed and suspended in 50 ml of 1.0% agarose gel solution. The

solution was then poured onto a glass and gelled by cooling. A specimen cut out 5-

mm � 5-mm was placed on the gel and incubated at 37 8C. A positive result was

defined as the detection of digested starch after 2 h.

2.5. Forensic specimens

We used following samples as mock forensic samples to test the forensic

application of our methods; 10 of cigarette butts, 10 of cotton gauzes wiping licked

skin, six of saliva stains stored for 6 years on filter paper, and a mixture stain of

semen and saliva (10:1 by volume) on filter paper. We also used saliva swabs from 3

dogs and one cat as common animals and examined whether S. salivarius and S.

mutans are detected in animals using our methods. We attempted to extract a DNA

from 2-mm � 2-mm and 5-mm � 5-mm of cigarette butt, and 5-mm � 5-mm of

gauze, swab and filter paper.

Table 1Primer sequences used in this study.

Target Name Sequence (50-30) Fragment

Size (bp)

S. salivalius MKK-F GTGTTGCCACATCTTCACTCGCTTCGG 544

MKK-R CGTTGATGTGCTTGAAAGGGCACCATT

S. mutans MKD-F GGCACCACAACATTGGGAAGCTCAGTT 433

MKD-R GGAATGGCCGCTAAGTCAACAGGAT

3. Results

3.1. Primer specificity

To evaluate the specificity of the two primers sets used in thisstudy (i.e., S. salivarius-specific primers and S. mutans-specificprimers), DNA obtained from five species of well-known environ-mental bacteria and four species of oral streptococci wereamplified by PCR using 1 ng of the template DNA. Each primerpair specifically amplified the DNA of the target bacteria, but didnot amplify DNA of other bacteria. The size of the PCR productobtained from S. salivarius and S. mutans (�500 bp) correspondedto the expected DNA fragment size reported by Hoshino et al.(544 bp and 433 bp, respectively) [9].

3.2. Sensitivity of detection

The sensitivity of this method was evaluated using the purifiedDNA from S. salivarius and S. mutans. Chromosomal DNA seriallydiluted from 1 ng to 500 fg was used as template DNA. The detectionlimits for S. salivarius and S. mutans were 10 pg (corresponding tothe amount of DNA in approximately 5.0 � 103 bacteria) and 1 pg(corresponding to the amount of DNA in approximately 4.5� 102

bacteria), respectively (Fig. 1).

3.3. Detection of S. salivarius and S. mutans from various body

samples

Species-specific PCR products were successfully identified inthe saliva samples, and no non-specific bands were observed. ThePCR results for S. salivarius and S. mutans identification fromvarious body fluids and skin swabs are shown in Table 2. S.

salivarius was identified in all saliva samples tested in this study,

Table 2The PCR results for S. salivarius and S. mutans identification from various body fluids

and skin swabs.

Sample n S. salivarius S. mutans

Detected Not detected Detected Not detected

Saliva 20 20 0 18 2

Semen 20 0 20 0 20

Urine 20 0 20 0 20

Vaginal fuild 9 0 9 0 9

Skin 20 0 20 0 20

Table 3The PCR results for S. salivarius and S. mutans identification from human saliva stain

samples.

Species n Detected Not detected

S. salivarius 20 20 0

S. mutans 20 12 8

Table 4The PCR results for S. salivarius and S. mutans identification from various mock

forensic samples.

Sample n S. salivarius S. mutans

Detected Not

detected

Detected Not

detected

Cigarette butt

(5-mm � 5-mm)

10 9 1 8 2

Cigarette butt

(2-mm � 2-mm)

10 8 2 5 5

Cotton gauze that

wiped licked skin

10 8 2 5 5

Saliva stain stored

for 6 years

6 5 1 4 2

H. Nakanishi et al. / Forensic Science International 183 (2009) 20–2322

whereas S. mutans was not detected in 10% of the saliva samples.Both species were detected in saliva but not in semen, urine,vaginal fluid, or skin.

3.4. Detection of S. salivarius and S. mutans from saliva stains

Table 3 shows the PCR results of S. salivarius and S. mutans

identification from saliva stain samples. S. salivarius was detectedin all saliva stain samples as well as liquid saliva samples, whereasS. mutans was only detected in 60% of saliva stain samples. Fig. 2shows representative results of the PCR products from the salivastain samples. Individual differences were observed in the degreeof detection (i.e., the intensity of the bands) of the PCR products forboth bacteria.

3.5. Forensic applications

Table 4 shows the PCR results of S. salivarius and S. mutans

identification from cigarette butts, cotton gauzes wiping lickedskin, and 6-years stored saliva stains. For 5-mm � 5-mm ofcigarette butt, S. salivarius was detected in 9 of 10 samples and S.

mutans was detected in 8 of 10 samples. Additionally, the numberof detected sample of S. salivarius and S. mutans from 2-mm � 2-mm of cigarette butt was 8 and 5, respectively. For cotton gauzeswiping licked skin, S. salivarius was detected in 8 of 10 samples andS. mutans was detected in 5 of 10 samples. Either streptococcus wasdetected in a mixture of semen and saliva. For aged saliva stains

Fig. 2. Representative results for the detection of (a) S. salivarius and (b) S. mutans from sa

No. 1; lane 2, donor No. 2; lane 3, donor No. 3; lane 4, donor No. 4; lane 5, donor No.

that had been stored for 6 years, S. salivarius was detected in 5 of 6samples and S. mutans was detected in 4 of 6 samples. All of thesesamples were also tested using the starch-agarose method, and allsamples gave positive results. On the other hand, neither S.

salivarius nor S. mutans were detected in the saliva of 3 dogs andone cat.

4. Discussion

We did not detect S. salivarius or S. mutans in semen, urine,vaginal fluid, or skin surface samples, but did verify their presencein saliva. Our data suggest that the PCR-based identification ofthese streptococci is sufficient to demonstrate the presence ofsaliva. Although there was variation in the amount of PCR productdetected, the sensitivity of this method was almost consistent withthe data reported by Hoshino et al. [9]. Nine vaginal fluid sampleswere examined and no false positive results were observed. Theinvestigation of vaginal flora, particularly streptococci, has beenpreviously reported [19,20]. Specifically, Rabe et al. tested vaginalflora from 487 pregnant women, and showed that the existencerates of S. salivarius and S. mutans were 2% and 0.2%, respectively[19]. Similarly, Egido et al. reported that the existence rate of S.

salivarius was 1.7% in 195 vaginal fluid samples [20]. Therefore, incases of mixed saliva and vaginal fluid, an examiner must carefullyevaluate the results.

Given that our detection rates for S. salivarius and S. mutans were100% and 90%, respectively, our results suggest that S. salivarius ismore reliable than S. mutans as a target species for the identificationof liquid saliva. Furthermore, the detection rate of S. mutans was only60% in the saliva stains. S. salivarius is mainly found on the dorsumlinguae, whereas S. mutans inhabits dental plaques [21]. Thedifference in detection rates is likely due to these differences in

liva stain samples using PCR. Lane M, 100-bp molecular mass marker; lane 1, donor

5.

H. Nakanishi et al. / Forensic Science International 183 (2009) 20–23 23

location. It is therefore expected that S. salivarius inhabiting thedorsum linguae would be especially useful for verifying the presenceof saliva in samples, e.g., when wiped from a victim’s skin with gauze.

We detected neither S. salivarius nor S. mutans in saliva samplesfrom dog or cat. In accordance with previous studies, neither ofbacteria was found in dog [22], cat [23], pig [23] and goat [23].However, Scannapieco et al. reported resembled S. salivarius wasisolated from a dog [23], and Fujita et al. reported S. mutans wasisolated from 2 dogs having caries [24]. That is, dog may have S.

salivarius and S. mutans in rare case. Moreover, oral streptococcimay be found in other animals. Thus, an examiner must carefullyevaluate the results when forensic saliva samples might have beencontaminated by animal saliva.

Using this method, we were able to detect the presence of S.

salivarius and S. mutans in sufficient number of mock forensicsamples in which a-amylase was also detected by conventionalstarch-agarose gel method. However, this method could not detectthese bacteria from some of mock samples. This may be due to anindividual variation of oral bacteria inhabiting. As described above,a variation in the amount of PCR product detected has beenobserved among the subjects.

Chromosomal DNA extracted from forensic materials is generallya very small amount and is occasionally degraded by factors such assepticity and ultraviolet rays. In this study, we could show that thesetwo streptococci are sufficient to indicate the presence of saliva butwe consider that this method may be improved by combining toother detecting technology or by designing new primer sets withsmaller amplicons that can be used in cases of extremely degradedsaliva samples.

Acknowledgement

The authors would like to express their gratitude to Dr.Yoshihito Fujinami (National Research Institute of Police Science)for his guidance and support on bacteria issue.

References

[1] G.M. Willott, An improved test for the detection of salivary amylase in stains, J.Forensic Sci. Soc. 14 (1974) 341–344.

[2] B.W. Schill, G.F.B. Schumaker, Radial diffusion in gels for micro determination ofenzymes, Anal. Biochem. 46 (1972) 502–533.

[3] H. Tsutsumi, K. Higashide, Y. Mizuno, K. Tamaki, Y. Katsumata, Identification ofsaliva stains by determination of the specific activity of amylase, Forensic Sci. Int.50 (1991) 37–42.

[4] L. Quarino, Q. Dang, J. Hartmann, N. Moynihan, An elisa method for the identifica-tion of salivary amylase, J. Forensic Sci. 50 (2005) 873–876.

[5] J. Juusola, J. Ballantyne, mRNA profiling for body fluid identification by multiplexquantitative RT-PCR, J. Forensic Sci. 52 (2007) 1252–1262.

[6] G.W. Burnett, H.W. Scherp, G.S. Schuster, Oral Microbiology and Infectious Dis-ease Fourth Edition, The Williams & Wilkins Company, Baltimore, 1976.

[7] S.S. Socransky, S.D. Manganiello, The oral microbiota of man from birth to senility,J. Periodontol. 42 (1971) 485–496.

[8] W.J. Loesche, Role of Streptococcus mutans in human dental decay, Microbiol. Rev.50 (1986) 353–380.

[9] T. Hoshino, M. Kawaguchi, N. Shimizu, N. Hoshino, T. Ooshima, T. Fujiwara, PCRdetection and identification of oral streptococci in saliva samples using gtf genes,Diagn. Microbiol. Infect. Dis. 48 (2004) 195–199.

[10] T. Sato, J. Matsuyama, T. Kumagai, G. Mayanagi, M. Yamamura, J. Washino, N.Takahashi, Nested PCR for detection of mutans streptococci in dental plaque, Lett.Appl. Microbiol. 37 (2003) 66–69.

[11] A. Yoshida, N. Suzuki, Y. Nakano, M. Kawada, T. Oho, T. Koga, Development of a 50

nuclease-based real-time PCR assay for quantitative detection of cariogenicdental pathogens streptococcus mutans and streptococcus sobrinus, J. Clin. Micro-biol. 41 (2003) 4438–4441.

[12] A. Yano, N. Kaneko, H. Ida, T. Yamaguchi, N. Hanada, Real-time PCR for quantifica-tion of Streptococcus mutans, FEMS Microbiol. Lett. 217 (2002) 23–30.

[13] T. Oho, Y. Yamashita, Y. Shimazaki, M. Kushiyama, T. Koga, Simple and rapiddetection of Streptococcus mutans and Streptococcus sobrinus in human saliva bypolymerase chain reaction, Oral Microbiol. Immunol. 15 (2000) 258–262.

[14] S. Rupf, K. Merte, K. Eschrich, Quantification of bacteria in oral samples bycompetitive polymerase chain reaction, J. Dent. Res. 74 (1999) 850–856.

[15] T. Igarashi, Y. Yano, A. Yamamoto, R. Sasa, N. Goto, Identification ofStreptococcus salivarius by PCR and DNA probe, Lett. Appl. Microbiol. 32 (2001)394–397.

[16] T.R. Elliot, A.H. Rogers, J.R. Haverkamp, D. Groothuis, Analytical pyrolysis ofStreptococcus salivarius as an aid to identification in bite-mark investigation,Forensic Sci. Int. 26 (1984) 131–137.

[17] K.A. Brown, T.R. Elliot, A.H. Rogers, J.C. Thonard, The survival of oral streptococcion human skin and its implication in bite-mark investigation, Forensic Sci. Int. 26(1984) 193–197.

[18] M. Rahimi, N.C. Heng, J.A. Kieser, G.R. Tompkins, Genotypic comparison ofbacteria recovered from human bite marks and teeth using arbitrarily primedPCR, J. Appl. Microbiol. 99 (2005) 1265–1270.

[19] L.K. Rabe, K.K. Witerscheid, S.L. Hiller, Association of viridans group streptococcifrom pregnant women with bacterial vaginosis and upper genital tract infection, J.Clin. Microbiol. 26 (1988) 1156–1160.

[20] J.M. Egido, J.R. Maestre, M.Y. Pena Izquierdo, The isolation of streptococcusmorbillorum from vaginal exudates, Rev. Soc. Bras. Med. Trop. 28 (1995) 117–122.

[21] L.W. Wannamaker, J.M. Matsen, Streptococci and Streptococcal Diseases, Aca-demic Press, New York, London, 1972.

[22] K. Takada, K. Hayashi, K. Sasaki, T. Sato, M. Hirasawa, Selectivity of Mitis Salivariusagar and a new selective medium for oral streptococci in dogs, J. Microbiol.Methods 66 (2006) 460–465.

[23] F.A. Scannapieco, L. Solomon, R.O. Wadenya, Emergence in human dental plaqueand host distribution of amylase-binding streptococci, J. Dent. Res. 73 (1994)1627–1635.

[24] K. Fujita, M. Tonokura, S. Hanada, H. Takayanagi, M. Mitsumura, A. Tsuchiya, Y.Sibasaki, K. Hirota, K. Takayama, R. Fujita, Two cases of dental caries on themaxillary first molar in small dogs, J. Anim. Clin. Med. 10 (2001) 23–26.