Forensic Science International: Genetics 3 (2009) 80–88

mRNA profiling for body fluid identification by reverse transcriptionendpoint PCR and realtime PCR

C. Haas a,*, B. Klesser a, C. Maake b, W. Bar a, A. Kratzer a

a Institute of Legal Medicine, Forensic Genetics, University of Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerlandb Institute of Anatomy, University of Zurich, Winterthurerstrasse 190, 8057 Zurich, Switzerland

A R T I C L E I N F O

Article history:

Received 13 June 2008

Received in revised form 24 October 2008

Accepted 14 November 2008

Keywords:

Forensic science

mRNA profiling

Body fluid identification

A B S T R A C T

mRNA profiling is a promising new method for the identification of body fluids from biological stains.

Major advantages of mRNA profiling are the possibility of detecting several body fluids in one multiplex

reaction and of simultaneously isolating DNA without loss of material. A reverse transcription endpoint

polymerase chain reaction (PCR) method and a realtime PCR assay were established for the identification

of blood, saliva, semen, vaginal secretions and menstrual blood, and were compared to conventional

enzymatic and immunologic tests. The results for specificity, sensitivity and suitability to biological

stains were satisfying and mRNA stability was demonstrated for up to 2-year-old stains. Two novel

multiplex assays were created with the endpoint PCR primers: multiplex 1 amplifies two markers for

each of the above mentioned body fluids and is suited for screening; multiplex 2 was designed for the

detection of blood, vaginal secretions and menstrual blood. The results demonstrate that both endpoint

PCR and realtime PCR are suitable for the identification of body fluids in forensic stains and represent an

effective alternative to conventional enzymatic and immunologic tests.

� 2008 Elsevier Ireland Ltd. All rights reserved.

Contents lists available at ScienceDirect

Forensic Science International: Genetics

journa l homepage: www.e lsev ier .com/ locate / fs ig

1. Introduction

Identification of the biological source of a stain is an importantissue in forensic casework, that may also help in predicting thesuccess of a DNA analysis and in interpreting DNA results. Inforensic practice biological stains are routinely preanalyzed withenzymatic or immunologic tests. However, there are someproblems associated with these tests. Most of them are notspecific, due, for example, to cross-reactions with other species ortissues. Furthermore, no such tests are available for the identifica-tion of vaginal secretions and menstrual blood. In particular, thedifferentiation between menstrual blood and blood due to traumacan be crucial in forensic casework, for example, in sexual assaultcases.

Recently, the analysis of cell-specific mRNA expression hasbeen proposed as a promising new technique for the identificationof body fluids in biological stains. A number of markers have beenidentified for the forensically most relevant body fluids: blood,saliva, semen, vaginal secretions and menstrual blood [1–11].

However, RNA is notorious for its rapid post-mortem and invitro decay, because of the ubiquitously present RNases. Quiteunexpectedly, however, several reports have pointed out a highstability of RNA under controlled conditions; for example, RNA

* Corresponding author. Tel.: +41 44 635 56 56; fax: +41 44 635 68 58.

E-mail address: cordula.haas@irm.uzh.ch (C. Haas).

1872-4973/$ – see front matter � 2008 Elsevier Ireland Ltd. All rights reserved.

doi:10.1016/j.fsigen.2008.11.003

could be isolated from 15-year-old dried blood stains [12]. Evenfrom stains that were exposed to a range of environmentalconditions for up to 547 days, RNA of sufficient quality andquantity could be isolated [13]. Since casework material is oftenlimited, an important advantage of body fluid identification bymRNA profiling is the possibility of simultaneously isolating RNAand DNA from the same piece of stain [14,15].

In the present study a reverse transcription endpoint PCR and arealtime PCR method were developed and compared to conven-tional enzymatic and immunologic tests for the identification ofbody fluids. Endpoint PCR allows the detection of a specifictranscript, when present in sufficient quantity. In contrast, realtimePCR allows the detection of relative gene expression levels indifferent samples and in comparison to an endogenous control. Thefollowing genes, which have been reported to be expressed in atissue-specific manner, were analyzed: porphobilinogen deaminase(PBGD) [16], b-spectrin (SPTB) [17] and hemoglobin beta (HBB) [18]for blood, statherin (STATH) [19] and histatin 3 (HTN3) [19,20] forsaliva, protamine 1 and 2 (PRM1 and PRM2) [3,21] for sperm, humanbeta-defensin 1 (HBD1) [22] and mucin 4 (MUC4) [23] for vaginalsecretions and matrix metalloproteinases 7 and 11 (MMP7 andMMP11) [2] for menstrual blood. The housekeeping genesglyceraldehyde-3-phosphate dehydrogenase (GAPDH) and 18SrRNA were used as endogenous controls. The endpoint PCR primerswere combined and tested in two multiplexes, allowing thesimultaneous amplification of several markers in one PCR reaction.Specificity, sensitivity and suitability to biological stains were

C. Haas et al. / Forensic Science International: Genetics 3 (2009) 80–88 81

analyzed and mRNA stability was recorded over time. The menstrualmarker expression was investigated for a whole menstrual cycle(days 1–28).

2. Materials and methods

2.1. Samples

Body fluids were collected from healthy volunteers. Threeindividuals donated blood samples, two individuals saliva samples,semen samples were prepared from a frozen aliquot of oneindividual, vaginal secretion samples were donated from threeindividualsand menstrualbloodsamplesfrom11individuals.Unlessotherwise indicated, 10 ml aliquots of fresh blood (without antic-oagulation treatment) or saliva or frozen semen were pipetted oncotton swabs and dried at room temperature. Vaginal secretions andmenstrual blood were collected from the vagina on sterile cottonswabs and dried at room temperature. The swabs were stored atroom temperature, in a dark place, without special humidity control,for up to 2 years. Forensic stains included 13 mock samples, fourstains from a German proficiency trial (GEDNAP 34) and 15 caseworksamples from 11 routine cases. The 13 mock samples comprisedblood on wood, cardboard, tissue, cleaning cloth and from a labbench; saliva from chewing gum, postage stamp, bottle, cigaretteand fabric; semen on cotton and fabric. GEDNAP 34 stains included10 ml saliva on a cigarette butt, 20 ml blood on a cosmetic pad, 10 mlblood mixture on cotton and 20 ml semen–saliva mixture on cotton.

2.2. RNA-extraction and reverse transcription

Special precautions were taken to avoid the degradation of RNAby the ubiquitously present RNases, namely a separate working

Table 1Endpoint PCR primer sequences and realtime PCR (*) assay IDs. References can be foun

Body fluid Gene Primer sequence/assay ID

Blood SPTB f: AGG ATG GCT TGG CCT TTA

r: ACT GCC AGC ACC TTC ATC

SPTB* Hs00165820_m1

PBGD f. TGG ATC CCT GAG GAG GG

r: TCT TGT CCC CTG TGG TGG

HBB f: GCA CGT GGA TCC TGA GA

r: ATG GGC CAG CAC ACA GA

HBB* Hs00747223_g1

Saliva STATH f: TTT GCC TTC ATC TTG GCT

r: CCC ATA ACC GAA TCT TCC

HTN3 f: GCA AAG AGA CAT CAT GG

r: GCC AGT CAA ACC TCC ATA

Semen PRM1 f: GCC AGG TAC AGA TGC TG

r: TTA GTG TCT TCT ACA TCT

PRM2 f: GTG AGG AGC CTG AGC GA

r: TTA GTG CCT TCT GCA TGT

PRM2* Hs00172518_m1

Vaginal secretions HBD1 f: CCC AGT TCC TGA AAT CCT

r: CAG GTG CCT TGA ATT TTG

MUC4 f: GGA CCA CAT TTT ATC AGG

r: TAG AGA AAC AGG GCA TA

Menstrual blood MMP7 f: TCA ACC ATA GGT CCA AGA

r: CAA AGA ATT TTT GCA TCT

MMP7* Hs00159163_m1

MMP11 f: GGT GCC CTC TGA GAT CGA

r: TCA CAG GGT CAA ACT TCC

MMP11* Hs00171829_m1

Housekeeping genes GAPDH f: TCT TCA CCA CCA CGG AGA

r: AGG GGG CAG AGA TGA TG

GAPDH* 4326317E

18SrRNA f: CTC AAC ACG GGA AAC CTC

r: CGC TCC ACC AAC TAA GAA

place only for RNA, separate pipettes, RNase-free plastic ware,cleaning of the whole working area with RNaseZap (Ambion/AppliedBiosystems, Rotkreuz, Switzerland) and the regular changing ofgloves. Unless otherwise indicated, the whole swabs were used forthe RNA extraction. Total RNA was extracted with the RNeasy Micro/Mini Kit or the AllPrep DNA/RNA Mini Kit (Qiagen, Hombrechtikon,Switzerland) according to the manufacturer’s protocol with thefollowing adaptation for both Mini Kits: the piece of stain was placedin 345 ml RLT-buffer + 5 ml Carrier-RNA in a DNA LoBind tube(Vaudaux-Eppendorf, Schonenbuch, Switzerland). The RNase-freeDNase Set (Qiagen) was used for on-column DNA digestion. RNA waseluted in 24–30 ml H2O, from which 22 ml were immediately reversetranscribed into cDNA. For the reverse transcription reaction,random primers and Superscript III reverse transcriptase (Invitro-gen, Basel, Switzerland) were used according to the manufacturer’sprotocol. Half of the RNA was used as RT minus control (withoutreverse transcriptase) to detect genomic DNA contamination. cDNAwas obtained in a final volume of 20 ml.

2.3. Endpoint PCR

Primers for the amplification of gene-specific sequences wereadopted from Juusola and Ballantyne [6], including PBGD, SPTB,HTN3, PRM1, PRM2, MUC4 and MMP7. The STATH- and HBD1-primers were redesigned for the multiplex, because of unsuitableannealing temperatures. Additional primer sets were designed forHBB, MMP11, 18S rRNA and GAPDH, using the Primer3 and theRoche Applied Science-Universal Probe Library Software (Table 1).It was ensured that all primers overlap exon–exon-junctions orspan an intron. The forward primers were 50-labelled with 6-FAM,VIC, HEX or NED (Table 1). 10� Buffer I and AmpliTaq GoldPolymerase (Applied Biosystems) were used for the singleplexes

d in Section 2.

Dye Size (bp) Reference

AT FAM 247 [6]

TT

FAM 61 AB

C AGA AG VIC 177 [6]

ACA TAG CAA T

A C FAM 61 Roche

C

FAM 106 AB

CT FAM 93 Primer3

AA

G TA FAM 134 [6]

ATC

T CGC AG NED 153 [6]

CGG TCT

A CGC FAM 294 [6]

TCT CTT C

FAM 89 AB

GA FAM 215 Primer3

GT

AA NED 235 [6]

G GA

AC VIC 240 [6]

CC

FAM 101 AB

C NED 92 Roche

AGT

FAM 66 AB

A HEX 72 Roche

A C

VIC 122 AB

AC HEX 110 Roche

CG

Table 2Body fluid specificity of the mRNA markers. Three separate experiments were

analyzed with endpoint PCR singleplexes and realtime PCR (*). Results in

parentheses represent 1–2 positive results out of 3 experiments. Gray shaded

areas contain the results we expect to be positive.

C. Haas et al. / Forensic Science International: Genetics 3 (2009) 80–8882

and the multiplex PCR Kit (Qiagen) for the multiplexes, each in atotal reaction volume of 25 ml. For the singleplexes, primerconcentrations were optimized in dilution series with cDNA fromdifferent body fluids and were used as follows: SPTB 0.8 mM, PBGD0.8 mM, HBB 0.04 mM, STATH 0.16 mM, HTN3 0.16 mM, PRM10.8 mM, PRM2 0.8 mM, HBD1 0.8 mM, MUC4 0.04 mM, MMP70.16 mM, MMP11 0.4 mM, GAPDH 0.4 mM, 18S rRNA 0.001 mM. Forthe multiplexes, primer concentrations were optimized by thevariation of the primer mixture using cDNA extracted and reversetranscribed from a body fluid mixture on a single swab. Thefollowing primer concentrations were used for multiplex 1: SPTB0.1 mM, PBGD 0.2 mM, STATH 0.2 mM, HTN3 0.2 mM, PRM10.2 mM, PRM2 0.2 mM, HBD1 0.3 mM, MUC4 0.1 mM, MMP70.3 mM, MMP11 0.2 mM and for multiplex 2: SPTB 0.1 mM, PBGD0.2 mM, HBB 0.01 mM, HBD1 0.1 mM, MUC4 0.05 mM, MMP70.3 mM, MMP11 0.1 mM. The cDNA input amount was 1–5 ml orthe same amount of H2O as non-template control. The amplifica-tion conditions were as follows on a GeneAmp PCR System 9700(Applied Biosystems): for the singleplexes the initial denaturationwas at 95 8C for 11 min, followed by 33–35 cycles of 94 8C 20 s,55 8C 30 s, 72 8C 40 s and the final elongation at 72 8C for 5 min. Forthe multiplexes the initial denaturation was at 95 8C for 15 min,followed by 35 cycles of 94 8C 30 s, 57–60 8C 90 s, 72 8C 60 s and afinal elongation at 72 8C for 10 min.

2.4. Capillary electrophoresis

PCR products were detected with a 310 or 3100 GeneticAnalyzer (Applied Biosystems). One microliter of the amplifiedsample was added to 12.5 ml Hi-Di formamide and 0.5 ml ofGeneScan-500 LIZ or ROX Size Standard (all from AppliedBiosystems). The samples were heated at 95 8C for 2 min andsnap cooled on ice for 3 min. The following electrophoresisconditions were used for the 310 Genetic Analyzer: 5 s injectiontime, 15 kV injection voltage, 15 kV run voltage, 60 8C, 28 min runtime, Dye Set D (6-FAM, HEX, NED, ROX) and the 3100 GeneticAnalyzer: 10 s injection time, 3 kV injection voltage, 15 kV runvoltage, 60 8C, 21 min run time, Dye Set G5 (6-FAM, VIC, NED, PET,LIZ). Raw data were analyzed with the Genescan Software (AppliedBiosystems). The threshold for a positive result was 50 rfu, basedon the sensitivity of the instruments.

2.5. Realtime PCR

cDNA was amplified using pre-designed TaqMan Gene expres-sion assays (Table 1) and the Taqman Universal PCR Master Mix(Applied Biosystems) in a total reaction volume of 25 ml. The cDNAinput amount was 1–5 ml or the same amount of H2O as non-template control. Human GAPDH was used as endogenous control.The thermal cycling conditions were 95 8C 10 min, followed by 40or 50 cycles of 95 8C 15 s, 60 8C 60 s. For data analysis the thresholdwas set manually to 0.200. The amplification products weredetected with a 7000 sequence detection system (AppliedBiosystems). Each sample was normalized to the endogenouscontrol, thus delta Ct (dCt) was calculated by subtracting thethreshold cycle (Ct) value of the endogenous control from the Ctvalue of the target gene. The Ct values of samples not yielding asignal (undetermined) were set equal to the highest cycle numberused in the assay for the calculation of dCt. A small dCt indicates ahigh expression of the respective target transcript, a large dCtindicates a low or not detectable expression level.

2.6. DNA-extraction, amplification, detection

DNA was co-extracted with the AllPrep DNA/RNA Mini Kit(Qiagen) and eluted in 100 ml EB buffer. Conventional Chelex

extraction was performed according to standard protocols,followed by ultrafiltration with Amicon Ultra-4 Centrifugal filterdevices (Millipore, Zug, Switzerland) to a final volume of around100 ml DNA. Two microliters DNA (or up to 10 ml, if the 2 ml resultwas weak) were amplified with the SGMplus multiplex Kit(Applied Biosystems) in a total reaction volume of 25 ml on aGeneAmp PCR System 9700 (Applied Biosystems) according to themanufacturer’s protocol. PCR products were detected with a 310Genetic Analyzer (Applied Biosystems). Raw data were analyzedusing the Genescan and Genotyper Softwares (Applied Biosys-tems).

2.7. Enzymatic/immunologic tests

The following enzymatic and immunologic tests were usedaccording to standard protocols: The tetramethylbenzidine test[24] and the Hexagon OBTI test (Human GmbH, Wiesbaden,Germany) [25] for the identification of blood, the a-amylase-testfor the identification of saliva (Ecoline S+ a-Amylase CNP-G3,DiaSys Diagnostic Systems GmbH, Holzheim, Germany), Baecchistaining and microscopy for the detection of spermatozoids [26]and the acid phosphatase test for the identification of seminalplasma [27].

3. Results

3.1. Body fluid specificity

The body fluid specificity of the markers was to be demon-strated by the presence of the candidate gene in the correspondingbody fluid and the absence of all others, using endpoint PCRsingleplexes and realtime PCR. Blood, saliva, semen (10 ml aliquotson cotton swabs), vaginal secretions and menstrual blood (vaginalswabs) were tested for 13 markers with endpoint PCR, and blood,semen and menstrual blood for six markers with realtime PCR,including the housekeeping genes (Table 2). Two microliters cDNAwas used for each PCR reaction. All experiments were repeatedthree times. mRNAs of the supposedly blood-specific markersSPTB, PBGD and HBB were detected in blood and to a lesser extentin menstrual blood, but undetectable in saliva and semen samples.Unexpectedly, small amounts of SPTB mRNA were found in somevaginal secretion samples, irrespective of the detection method.

Fig. 1. Sensitivity of the RNA profiling methods (endpoint PCR and realtime PCR *)

compared to conventional enzymatic tests for the identification of blood, saliva and

semen. Gray squares = 3 positive results out of 3 experiments, hatched squares = 1–

2 positive results out of 3 experiments, white squares = 3 negative results out of 3

experiments.

C. Haas et al. / Forensic Science International: Genetics 3 (2009) 80–88 83

mRNAs of the saliva-specific markers STATH and HTN3 wereexclusively found in saliva samples. In addition to the amplicon ofexpected size, a larger fragment (152 bp) was detected in salivasamples using the HTN3 primers. mRNAs of the sperm-specificmarkers PRM1 and PRM2 were only apparent in semen samples.HBD1 and MUC4 transcripts were present in vaginal secretion andmenstrual blood samples, but undetectable in the other bodyfluids. MMP7 and MMP11 mRNAs were found in menstrual bloodsamples and in some vaginal secretion samples. One out of threeblood samples showed a positive result for MMP11. mRNAs of thehousekeeping genes GAPDH and 18S rRNA were detected in allbody fluids, very weakly in saliva and semen samples, moderatelyin blood samples and abundantly in vaginal secretion andmenstrual blood samples. Most RT minus controls were negative,except for the high input samples analyzed with the PRM1 primerpair, which revealed a small 240 bp peak representing minorgenomic DNA contamination. Both housekeeping genes weredetected in genomic DNA from a buccal swab when analyzed withendpoint PCR.

3.2. Sensitivity

The sensitivity of the methods was tested with differentamounts of dried body fluids (5, 1, 0.1, 10�2, 10�3, 10�4, 10�5 mlblood, saliva, semen) on cotton swabs (Fig. 1). Three sets ofexperiments were analyzed with endpoint PCR singleplex,realtime PCR and enzymatic/immunologic tests. For the testedbody fluids, the endpoint PCR and the realtime PCR method wereas sensitive as conventional enzymatic tests. Actually, HBBexpression was detectable in 10�5 ml blood using realtime PCR.The tetramethylbenzidine test was more sensitive (10�3 mlblood) than the RNA profiling methods with the SPTB and PBGDmarkers (0.1 and 1 ml blood, respectively). Except for the 240 bppeak in the PRM1 samples, the endpoint PCR RT minus controlsand all non-template controls were negative. For the sensitivitytests, realtime PCR results were compared without normal-ization, since equally prepared dilution series of body fluidswere used. A Ct value <40 was interpreted as positive result,since – assuming a 100% PCR efficiency, which is stated by themanufacturer – 40 PCR cycles would imply an initial copynumber of only few DNA molecules, which is the ultimate levelof sensitivity. All realtime PCR RT minus controls and all non-template controls were higher than the highest cycle numbersused (40 cycles).

3.3. Multiplexes

For forensic casework, the simultaneous analysis of severalmarkers in one multiplex PCR reaction would be an importantadvantage. To this end, different endpoint PCR primer sets wereoptimized and tested in two different multiplexes. Multiplex 1amplifies two markers for each body fluid: SPTB, PBGD, STATH,HTN3, PRM1, PRM2, HBD1, MUC4, MMP7, MMP11. To demon-strate the ability of multiplex 1 to display all markers in one PCRreaction and to optimize the concentrations of the primers, bodyfluid mixtures prepared on single swabs (10 ml blood, 10 mlsaliva and 10 ml semen, added to a menstrual blood/vaginalswab) were analyzed. Fig. 2a shows an electropherogram of abody fluid mixture, detected with multiplex 1, revealing twopeaks for each body fluid. Multiplex 2, amplifying the markersSPTB, PBGD, HBB, HBD1, MUC4, MMP7and MMP11, was designedfor the detection of blood, vaginal secretions and menstrualblood stains. The electropherograms of three body fluid samplesdetected with multiplex 2 are shown in Fig. 2b–d. All expectedmarkers were amplified, only the blood-specific SPTB transcriptwas not detectable in the menstrual blood sample (Fig. 2d),

probably due to low expression of this gene compared to theother markers.

3.4. Menstrual blood

In order to get a more precise picture of the menstrual blood-specific marker expression in menstrual blood and vaginalsecretion samples, MMP7 and MMP11 mRNA levels wereinvestigated for a whole menstrual cycle (days 1–28). Menstrualblood samples (days 1–4) from four different individuals andvaginal swabs from two different individuals (days 6–28) wereanalyzed by endpoint PCR singleplexes MMP7, MMP11, 18S rRNA(Fig. 3a), endpoint PCR multiplex 2 (data not shown) and realtimePCR (Fig. 3b). All three methods provided similar results: Themenstrual blood samples (days 1–4) revealed high levels of MMP7and MMP11. In the vaginal secretion samples (days 6–28) MMP7and MMP11 signals were weak, but nonetheless present. Somevaginal samples (days 6–15) even show high MMP11 expression.With multiplex 2 (data not shown) the blood markers SPTB, PBGDand HBB were rather weak in the menstrual blood samples (days1–4) and in some cases not detectable at all. The SPTB marker waspresent in some vaginal samples. The Housekeeping gene 18S rRNA(singleplex, Fig. 3a) and the vaginal secretion markers HBD1 andMUC4 (multiplex 2, data not shown) were expressed quiteconstantly in all samples, demonstrating equal input amounts ofbody fluids.

Fig. 2. (a) Endpoint PCR multiplex 1 result of a blood-saliva-semen-vaginal secretion-menstrual blood mixture, revealing two peaks for each body fluid. Electropherograms of

blood (b), vaginal secretion (c) and menstrual blood (d) samples detected with endpoint PCR multiplex 2.

C. Haas et al. / Forensic Science International: Genetics 3 (2009) 80–8884

Fig. 3. Menstrual blood (days 1–4) and vaginal secretion samples (days 6–28) analyzed with endpoint PCR singleplexes MMP7, MMP11, 18S rRNA (a) and realtime PCR MMP7

and MMP11, normalized to the housekeeping gene GAPDH (b).

C. Haas et al. / Forensic Science International: Genetics 3 (2009) 80–88 85

3.5. Old stains

The main concern about the use of mRNA for forensic analysis isthe believed limited stability of RNA. To investigate the suitabilityof mRNA profiling for old samples, up to 2-year-old stains, stored ina dry dark place, were analyzed with endpoint PCR multiplex 1. Allbody fluids could be detected in up to 2-year-old stains (Fig. 4).There was no tendency towards decreasing peak heights in the oldsamples. In general, the blood markers showed the weakest signalswith most peak heights <1000 rfu.

3.6. Forensic stains

The analysis of mRNA originating from forensic samplesusing the methods detailed above is a true test of the assays.This was performed with the co-extraction of RNA and DNAfrom 13 mock samples, four GEDNAP stains and 15 caseworksamples. RNA analysis was performed with endpoint PCRmultiplex 1 and the results were compared to conventionalenzymatic tests (Table 3). The 13 mock samples and the fourGEDNAP stains were all identified correctly. The RNA results ofthe casework samples were concordant with the enzymatic testresults and could be assigned to a specific body fluid or a bodyfluid mixture. Only one RNA result differed from the enzymatictest results (condom outside), probably because two differentswabs were used, the first one was secured for routine analysis(enzymatic tests and Chelex extraction) and the second one forRNA/DNA co-extraction. Genomic DNA, which was co-extractedfrom the forensic samples, revealed full or partial DNA profiles,allowing the identification of the stain donor. A qualitativecomparison of the co-extracted DNA with conventional Chelex-extracted DNA revealed similar DNA profiles, in terms ofanalyzed loci.

4. Discussion

The aim of this study was to test two mRNA profiling methods,namely endpoint PCR and realtime PCR, for the identification ofbody fluids and to compare these methods to conventionalenzymatic and immunologic tests with regard to specificity and

sensitivity. Suitability to old samples and forensic stains wasinvestigated by endpoint PCR. Two novel endpoint PCR multi-plexes were created, which allow the detection of several bodyfluids in one PCR reaction and which are suited for forensiccasework.

A high level of specificity could be demonstrated for theanalyzed candidate genes, as the respective RNA markers weredetected in the corresponding body fluids and mostly notdetected in the others. Occasional expression of MMP7 andMMP11 and, to a lesser extent, of SPTB was found in vaginalsecretion samples, irrespective of the analysis method. Juusolaand Ballantyne [6,8] did not find any cross reactivity, however,this may be explained by the less sensitive gel electrophoresisdetection method used. Nussbaumer et al. [7] reported crossreactivity of MUC4 with some saliva and semen samples using arealtime PCR assay and different primers, which we cannotconfirm with our endpoint PCR assay and did not test withrealtime PCR. The additional larger amplicon (152 bp) detectedin saliva samples using the HTN3 primers corresponds toanother histatin isoform (HTN1), which shows 95% sequenceidentity to HTN3 [20]. Since we used a different primer set forthe detection of STATH, we did not see the smaller STATHisoform described by Juusola and Ballantyne [6]. The RT minuscontrols where mostly negative for endpoint PCR and com-pletely negative for Realtime PCR, either because the primersoverlap exon–exon-junctions and do not reveal DNA contam-ination, or because the primers span a large intron which resultsin fragments >500 bp outside the capillary electrophoresisanalysis range. Only the PRM1 primer pair, which spans a smallintron of 91 bp, reveals a small 240 bp fragment in high inputsamples, representing minor genomic DNA contamination. Sincenone of the body-fluid specific primer pairs showed positiveresults with genomic DNA, we can conclude that minor DNAcontamination does not influence RNA results.

We found that the housekeeping genes GAPDH and 18S rRNAshowed only weak expression in saliva and semen samples,compared to the other body fluids. A reduced housekeeping geneactivity is explicable for these body fluids, since spermatozoidscontain little cytoplasm and ribosomes and the desquamated cellsof the cheek mucosa have almost no cell metabolism.

Fig. 4. Endpoint PCR multiplex 1 results in up to 2-year-old stains: blood (a), saliva

(b), semen (c), vaginal secretions (d) and menstrual blood (e). Each point represents

a single result.

C. Haas et al. / Forensic Science International: Genetics 3 (2009) 80–8886

Interestingly, GAPDH and 18S rRNA were detectable in genomicDNA. Since both primers were designed to span an intron,amplification of genomic DNA would result in longer-sizedfragments. This implies the presence of processed pseudogenes

for these housekeeping genes within the genome, which areknown at least for GAPDH [4,28]. When analyzing samples withhigh input amounts, for example, vaginal swabs, on-column DNAdigestion was not always sufficient to remove genomic DNA.Therefore GAPDH and 18S rRNA results have to be interpretedcarefully, when used as endogenous control. For future studies,better primers or other housekeeping genes may be chosen. Harperet al. [29] described a method that can be used to avoid theamplification of contaminating pseudogenes through the design ofprimers that bind only to genuine GAPDH mRNA transcript.

To investigate the sensitivity of the methods, we analyzeddifferent amounts of body fluids, comparable to forensic stains.Juusola and Ballantyne [8] described a 5- to 40-fold increase insensitivity for their realtime PCR method compared to an endpointPCR/capillary electrophoresis system. In their study, semen geneswere reported to have the lowest detection limit (1 pg RNA), bloodand menstrual blood genes an intermediate detection limit (100–150 pg RNA), and saliva genes the highest detection limit (2 ngRNA). In our study, realtime PCR was no more sensitive thanendpoint PCR or enzymatic tests. Only the HBB marker performedbetter for the detection of blood when analyzed with the realtimePCR method and compared to the other tests. On the basis of oursensitivity studies, both endpoint PCR and realtime PCR aresuitable for forensic casework.

In the present study the menstrual blood-specific markersMMP7 and MMP11 were detected abundantly on days 1–4 ofmenstruation, irrespective of the method used.

With endpoint PCR, Bauer and Patzelt [2] found 62% of thesamples to express MMP11 on day 1 of menstruation, compared to90–94% on days 2 and 3. Overall, MMP11 expression was found tobe more constant and stronger than MMP7 expression duringmenstruation. We found no difference in MMP7 and MMP11expression levels between days 1–4 of menstruation. But wedetected a weak expression of the two markers in vaginal secretionsamples (days 6–28) as well, which has not been observedpreviously. MMP expression in vaginal secretion samples, albeitweak, complicates the interpretation of a positive MMP result. Thiscould be overcome by comparing the result to a housekeeping geneand/or the vaginal secretion markers HBD1 and MUC4, which areexpressed quite constantly and strongly in all vaginal samples(days 1–28). In contrast to Bauer’s findings [2], we find the MMP7marker to be more suitable than the MMP11 marker in delimitingthe menstrual period (days 1–4), since some vaginal secretionsamples (days 6–28) revealed high MMP11 expression. All bloodmarkers were rather weak in the menstrual blood samples. Wholeblood accounts for only 30–50% of the total flow in most women[30]; therefore, it is not to be expected that menstrual blood willshow the same expression levels as whole blood.

From realtime PCR results, Bauer and Patzelt [10] concludedthat MMP7 expression is strong evidence for the menstrual originof a sample, reporting a dCt value range of between�3 and 8 for 60menstrual blood samples. In our work, most of the menstrual bloodsamples (days 1–4) were within the suggested range, but none ofthe vaginal secretion samples (days 6–28) were within this range,which is in accordance with Bauer’s results.

An unexpectedly high stability of mRNA for up to 15 years wasobserved for dried bloodstains [12]. We confirm in this report therecoverability of mRNA from body fluids stored in a dry, dark place,for up to 2 years. There was no significant decrease in peak heightsover time. Setzer et al. [13] report on successful RNA isolation fromstains that were exposed to a range of environmental conditionsfor up to 547 days. Environmental samples, protected from rain,exhibited housekeeping- and tissue-specific mRNA recoverabilityfrom 7 to 180 days. Samples exposed to rain showed results for upto 7 days at the most. The unexpectedly high stability of mRNA,even in the rough conditions to which casework samples may be

Table 3Endpoint PCR multiplex 1 results of 32 forensic samples (gray shaded areas contain the results we expect to be positive), compared to conventional enzymatic tests

(tetramethylbenzidine test for blood, a-amylase test for saliva, Baecchi staining and microscopy for spermatozoids). The number of analyzed loci in DNA profiles from co-

extracted DNA (from the same piece of stain) and from conventional Chelex-extracted DNA (from a neighboring piece of stain) are shown. 10 loci represent a full SGM+ profile.

n.d. = not done.

C. Haas et al. / Forensic Science International: Genetics 3 (2009) 80–88 87

exposed, is promising. Furthermore, amplification of short frag-ments (all of our amplicons are <290 bp) allows the detection ofpartly degraded RNA, which we definitely encounter in forensicsamples.

We were able to predict the source of most forensic stains withendpoint PCR multiplex 1 and could confirm them with enzymatictests. Some of the casework samples were difficult stains, since theoriginal material comprised body fluid mixtures or only few cells.From co-extracted DNA, full or partial DNA profiles could beproduced, allowing the identification of the stain donor. DNAprofiles from co-extraction were comparable with DNA profilesfrom conventional Chelex-extraction in terms of analyzed loci. Onthe basis of these first experiences, mRNA profiling seems suitablefor the identification of body fluids in forensic stains and DNA ofsufficient quality can be co-extracted.

4.1. Conclusion and future perspectives

In this study, a reverse transcription endpoint PCR and arealtime PCR method were compared to conventional enzymaticand immunologic tests for the identification of blood, saliva,semen, vaginal secretions and menstrual blood. In our study,realtime PCR did not prove to be any more sensitive than endpointPCR, but both methods were shown to be suitable for theidentification of body fluids in forensic stains. Major advantagesof mRNA profiling are the possibility of detecting several body

fluids in one multiplex reaction and the simultaneous DNAisolation without loss of material. We created two novel endpointPCR multiplex assays, multiplex 1 for screening the forensicallymost relevant body fluids and multiplex 2 for the detection ofblood, vaginal secretions and menstrual blood, which provedsuitable for, for example, the differentiation between menstrualblood and blood due to trauma.

Any additional markers that are discovered in the future couldbe added to the analysis without any change to the basic protocol.Additional and more sensitive markers would contribute to themore reliable identification of a stain. The identification of otherbody fluids like urine or tissues like skin are further forensicallyinteresting questions which might be resolved by mRNA profiling.

Acknowledgements

The authors thank F. Jaeggi and M. LeHir for reviewing themanuscript, N. Giezendanner, C. Moser and T. Burtscher fortechnical assistance, and J. Dewhurst for proof-reading.

References

[1] M. Bauer, A. Kraus, D. Patzelt, Detection of epithelial cells in dried blood stains byreverse transcriptase-polymerase chain reaction, J. Forensic Sci. 44 (1999) 1232–1236.

[2] M. Bauer, D. Patzelt, Evaluation of mRNA markers for the identification ofmenstrual blood, J. Forensic Sci. 47 (2002) 1278–1282.

C. Haas et al. / Forensic Science International: Genetics 3 (2009) 80–8888

[3] M. Bauer, D. Patzelt, Protamine mRNA as molecular marker for spermatozoa insemen stains, Int. J. Legal Med. 117 (2003) 175–179.

[4] J. Juusola, J. Ballantyne, Messenger RNA profiling: a prototype method to supplantconventional methods for body fluid identification, Forensic Sci. Int. 135 (2003)85–96.

[5] G. Ferri, C. Bini, S. Ceccardi, S. Pelotti, Successful identification of two years oldmenstrual bloodstain by using MMP-11 shorter amplicons, J. Forensic Sci. 49(2004) 1387.

[6] J. Juusola, J. Ballantyne, Multiplex mRNA profiling for the identification of bodyfluids, Forensic Sci. Int. 152 (2005) 1–12.

[7] C. Nussbaumer, E. Gharehbaghi-Schnell, I. Korschineck, Messenger RNA profiling:a novel method for body fluid identification by real-time PCR, Forensic Sci. Int. 157(2006) 181–186.

[8] J. Juusola, J. Ballantyne, mRNA profiling for body fluid identification by multiplexquantitative RT-PCR, J. Forensic Sci. 52 (2007) 1251–1262.

[9] M. Bauer, RNA in forensic science, Forensic Sci. Int. Genet. 1 (2007) 69–74.[10] M. Bauer, D. Patzelt, Identification of menstrual blood by real time RT-PCR:

technical improvements and the practical value of negative results, ForensicSci. Int. 74 (2008) 54–58.

[11] D. Zubakov, E. Hanekamp, M. Kokshoorn, W. van Ijcken, M. Kayser, Stable RNAmarkers for identification of blood and saliva stains from whole genome expressionanalysis of time-wise degraded samples, Int. J. Legal Med. 122 (2008) 135–142.

[12] M. Bauer, S. Polzin, D. Patzelt, Quantification of RNA degradation by semi-quantitative duplex and competitive RT-PCR: a possible indicator of the age ofbloodstains? Forensic Sci. Int. 138 (2003) 94–103.

[13] M. Setzer, J. Juusola, J. Ballantyne, Recovery and stability of RNA in vaginal swabsand blood, semen and saliva stains, J. Forensic Sci. 53 (2008) 1–10.

[14] M. Bauer, D. Patzelt, A method for simultaneous RNA and DNA isolation fromdried blood and semen stains, Forensic Sci. Int. 136 (2003) 76–78.

[15] M. Alvarez, J. Juusola, J. Ballantyne, An mRNA and DNA co-isolation method forforensic casework samples, Anal. Biochem. 335 (2004) 289–298.

[16] A.N. Gubin, J.L. Miller, Human erythroid porphobilinogen deaminase exists in 2splice variants, Blood 97 (2001) 815–817.

[17] Z.-L. Chu, A. Wickrema, S.B. Krantz, J.C. Winkelmann, Erythroid-specific proces-sing of human b spectrin I pre-mRNA, Blood 84 (1994) 1992–1999.

[18] P.P. Levings, J. Bungert, The human b-globin locus control region, Eur. J. Biochem.269 (2002) 1589–1599.

[19] L.M. Sabatini, T. Ota, E.A. Azen, Nucleotide sequence analysis of the humansalivary protein genes HIS1 and HIS2, and evolution of the STATH/HIS genefamily, Mol. Biol. Evol. 10 (1993) 497–511.

[20] L.M. Sabatini, E.A. Azen, Histatins, a family of salivary histidine-rich proteins, areencoded by at least two loci, Biochem. Biophys. Res. Commun. 160 (1989) 495–502.

[21] K. Steger, K. Pauls, T. Klonisch, F.E. Franke, M. Bergmann, Expression of protamine-1 and -2 mRNA during human spermiogenesis, Mol. Hum. Reprod. 6 (2000) 219–225.

[22] E.V. Valore, C.H. Park, A.J. Quayle, K.R. Wiles, P.B. McCray, T. Ganz, Human b-defensin-1: an antimicrobial peptide of urogenital tissues, J. Clin. Invest. 101(1998) 1633–1642.

[23] I.K. Gipson, S. Spurr-Michaud, R. Moccia, Q. Zhan, N. Toribara, S.B. Ho, A.R.Gargiulo, J.A. Hill, MUC4 and MUC5B transcripts are the prevalent mucin mes-senger ribonucleic acids of the human endocervix, Biol. Reprod. 60 (1999) 58–64.

[24] D.D. Garner, K.M. Cano, R.S. Peimer, T.E. Yeshion, An evaluation of tetramethyl-benzidine as a presumptive test for blood, J. Forensic Sci. 9 (1976) 97–102.

[25] M.N. Hochmeister, B. Budowle, R. Sparkes, O. Rudin, C. Gehrig, M. Thali, L. Schmidt,A. Cordier, R. Dirnhofer, Validation study of an immunochromatographic 1-steptest for the forensic identification of human blood, J. Forensic Sci. 44 (1999) 597–602.

[26] B. Baecchi, Neue methode zum nachweis von spermatozoen in zeugflecken,Dtsch. Med. Wochenschr. 35 (1909) 1105–1106.

[27] S.S. Kind, The acid phosphatase test, Methods Forensic Sci. 3 (1965) 267–288.[28] Z. Zang, P.M. Harrison, Y. Liu, M. Gerstein, Millions of years of evolution preserved:

a comprehensive catalog of the processed pseudogenes in the human genome,Genome Res. 13 (2003) 2541–2558.

[29] L.V. Harper, A.C. Hilton, A.F. Jones, RT-PCR for pseudogene-free amplification ofthe glyceraldehyde-3-phophate dehydrogenase gene (gapd), Mol. Cell. Probes 17(2003) 261–265.

[30] I.S. Fraser, P. Warner, P.A. Marantos, Estimating menstrual blood loss in womenwith normal and excessive menstrual fluid volume, Obstet. Gynecol. 98 (2001)806–814.