Anal Bioanal Chem (2015) 407:71357144

DOI 10.1007/s00216-015-8877-x

RESEARCH PAPER

Identification and evaluation of potential forensic

marker proteins in vaginal fluid by liquid
chromatography/mass spectrometry
Akihisa Igoh 1 & Yusuke Doi 1,2 & Koichi Sakurada 3

Received: 2 March 2015 / Revised: 21 June 2015 / Accepted: 23 June 2015 / Published online: 12 July 2015
# Springer-Verlag Berlin Heidelberg 2015

Abstract Vaginal fluid is one of the most common body

fluids found at crime scenes. Discriminating vaginal fluid
from other body fluids is important in forensic science; however, few potential protein markers have been reported to date.
Proteomic methods for identifying protein markers have
gained attention, although few reports have applied this technology to forensic protein markers. Therefore, to identify
characteristic vaginal proteins, we examined various body
fluids (nasal secretions, saliva, urine, semen, vaginal fluids,
and sweat) using liquid chromatography/electrospray ionization time-of-flight mass spectrometry and peptide mass fingerprinting. We identified three components (average molecular
mass values 17,2372, 18,0632, and 15,0751) detectable
only in vaginal samples: two human small proline-rich protein 3 (SPRR3) isoforms and a human fatty acid-binding protein 5 (FABP5) with an acetylated (+42) N-terminal region
lacking the initiator methionine residue (131). Using ELIS
A, these yielded markedly high average values in vaginal
fluids. The mass spectra of these proteins were not detected
in infant saliva but were detected in the vaginal fluid
Electronic supplementary material The online version of this article
(doi:10.1007/s00216-015-8877-x) contains supplementary material,
which is available to authorized users.
* Akihisa Igoh
op-qphif108@pref.okayama.jp
1

Forensic Science Laboratory of Okayama Prefectural Police H.Q.,

1-3-2 Tondacho, Kita-ku, Okayama 700-0816, Japan

Department of Legal Medicine, Okayama University Graduate

School of Medicine, Dentistry and Pharmaceutical Science, 2-5-1
Shikatacho, Kita-ku, Okayama 700-8558, Japan

Department of Forensic Dentistry, Graduate School of Medical and

Dental Sciences, Tokyo Medical and Dental University, 1-5-45
Yushima, Bunkyo-ku, Tokyo 113-8510, Japan

throughout the menstrual cycle. The results of forensic analysis (detection limit, mixed body fluid samples, casework samples, and blind samples) suggest that these proteins are potential forensic markers. In conclusion, high SPRR3 and FABP5
expression levels, which may be used as potential markers for
vaginal fluid identification in forensic science, were detected
in vaginal fluids from healthy adults.
Keywords Vaginal fluid identification . Liquid
chromatography/electrospray ionization time-of-flight mass
spectrometry . Peptide mass fingerprinting . Forensic
science . Protein markers

Introduction
Forensic analysis is a rapidly growing area of bioanalytical
chemistry [1]. In particular, DNA typing has been considerably
developed as the most common method for identifying individuals. In forensic biomaterial analysis, body fluids are investigated to determine the biological sources of trace evidence before
DNA typing to identify the individuals involved in the crime
scene. Body fluid identification is important because it helps in
determining the events that occurred during a crime. The increasing popularity of DNA typing has also highlighted the
importance of identifying body fluids [2]. Further, blood, saliva,
semen, and vaginal fluid are the most common body fluids
found at crime scenes. In sexual assault cases, vaginal fluids
are often left on the suspects fingers and clothes or on contraceptive devices and are considered as trace evidence. Therefore,
vaginal fluids can facilitate suspect identification or provide important evidence that could be used for deciding punishments.
In forensic institutes, body fluids are currently identified by
immunological methods using protein markers. However, unlike hemoglobin in blood [3], prostate-specific antigen in

7136

A. Igoh et al.

semen [4], or statherin in saliva [5], no characteristic protein

marker has been demonstrated for vaginal fluids, and an optimal method for identifying vaginal fluids has not been adequately established [6]. A method detecting cornulin and
cornifin for discriminating vaginal fluids from other body
fluids has been recently reported; however, this method involves time-consuming sample preparation for its use in practical forensic investigation, wherein rapidity is important.
Moreover, the forensic validation studies remain unclear [7].
Therefore, a simple and reliable method and specific protein
markers are required for forensic vaginal fluid identification.
Vaginal fluids are an important diagnostic tool in laboratory
medicine. Several proteomic vaginal fluid studies have recently attempted to identify novel candidate markers for diagnostics [810]; however, these studies have not compared the
protein component expression levels between vaginal and other body fluids. The comparison study among body fluids is
essential for forensic identification. Although proteomic
methods by mass spectrometry can be useful in identifying
forensic markers, only one such study has been reported by
Van Steendam et al. [7]. Another proteomic study testing different methods for sample preparation, separation, and marker
selection may identify novel forensic protein markers. Moreover, vaginal fluid is particularly influenced by several biological factors such as menstrual cycle, age, infection, frequency of sexual intercourse, use of contraceptive agents, antibiotics, and pregnancy [11, 12].
Mass spectrometry is also used for other purposes in forensic
institutes, such as identifying drugs and poisons in forensic
toxicology [1], but not for the analysis of biomaterials in body
fluids. The potential application of mass spectrometry in forensic biomaterial analysis has been described only in a few recent
reports [7, 13, 14]. However, applications of mass spectrometry
for this forensic purpose are expected to increase in the future.
In the present study, to identify the candidate forensic vaginal fluid markers for forensic analysis, various body fluids
from healthy volunteers were investigated using liquid
chromatography/electrospray ionization time-of-flight mass
spectrometry (LC/ESI-TOF MS). Subsequently, several proteins detected only in vaginal fluid were identified by peptide
mass fingerprinting (PMF). Further, the protein expression
levels in various body fluids were determined using enzymelinked immunosorbent assay (ELISA). Finally, the potential of
identified proteins for forensic markers was evaluated using
LC/ESI-TOF MS.

17), and urine (n=26), were collected from volunteers aged

2559 years. Vaginal secretions were collected on sterile cotton swabs by wiping the vaginal wall. Eleven participants
provided vaginal secretions regardless of their menstrual cycle
stage, and one participant collected samples weekly for
4 weeks to evaluate possible relationships with the menstrual
cycle. Two menopausal participants and one pregnant participant provided vaginal secretions. Sweat drops, primarily
from the face and arms, were collected on filter paper or into
plastic tubes after the participants had exercised. Nasal secretion samples were collected by blowing the nose into plastic
tubes. Saliva, semen, and urine samples were collected in
plastic tubes. In addition, fresh pediatric saliva samples were
collected from 14 volunteers aged between 2 months and
12 years. All body fluid samples were stored at 80 C until
they were analyzed.
All procedures involving human volunteers were approved
by the Human Subjects Committee of the Japanese Association of Forensic Science and Technology.
Recombinant SPRR3 (National Center for Biotechnology
Information (NCBI) accession no. gi20138065, 2169 a.a.;
CUSABIO, Wuhan, China) and recombinant FABP5 (NCBI
accession no. gi4557581, 1135 a.a.; ATGen, Seongnam, Korea) were refined by LC. Then, the concentrations were quantified using the 2D-Quant kit (GE Healthcare, Amersham,
UK). The N-terminal regions of the proteins were not
acetylated.

Materials and methods

LC/ESI-TOF MS

Sample collection and treatment

LC/ESI-TOF MS analyses were performed on a Dionex Ultimate 3000 liquid chromatography system (Dionex, CA, USA)
connected to a micrOTOFII TOF-MS system (Bruker
Daltonics, MA, USA) through a photodiode array detector

Fresh body fluids, including vaginal secretions (n=14), sweat

(n=22), nasal secretions (n=22), saliva (n=30), semen (n=

Analysis of body fluids

Sample preparation
The samples for body fluid examination were prepared as
follows: sterile cotton swabs with vaginal fluid were cut into
approximately 55-mm square pieces, and extracts were obtained from these pieces using 100 L of ultra-pure water. The
filter papers used to collect sweat were cut into approximately
11-cm pieces, and extracts were obtained from these pieces
using 50 L of ultra-pure water.
Each sample and extract collected from adults was diluted
five times with solvent A (0.1 %v/v solution of formic acid
and 4 mM ammonium acetate in water). The solutions were
filtered through Ultrafree-MC Centrifugal Devices (Millipore,
MA, USA) with a pore size of 0.45 m. Subsequently, 25 L
of each filtered body fluid solution was used for LC/ESI-TOF
MS.

Identification and evaluation of potential forensic marker

(PDA). Separations were performed on a C4 Jupiter column

(Phenomenex, CA, USA) with a 5-m particle diameter (column dimensions 2.0150 mm). The column was maintained
at 40 C, and solvents A and B were used for the separations
[solvent B, 0.1 %v/v solution of formic acid and 4 mM ammonium acetate in acetonitrile/water 90:10 (v/v)]. A linear
gradient of 595 % solvent B was applied for 90 min at a flow
rate of 0.2 mL/min from PDA to the electrospray ionization
(ESI) source. The PDA was set in the range 200800 nm.
Mass spectra were collected in the positive ion mode with
an MS spray voltage of 4.5 kV.
Average ESI mass spectra deconvolution was automatically performed with Bruker Compass Data Analysis software
(version 4.0 SP2; Bruker Daltonics).
Protein identification
Protein digestion
A few dozen HPLC fractions that contained characteristic
vaginal proteins were manually collected, and 25 mL of the
pooled eluates were exchanged with ultra-pure water and concentrated to a final volume of 25 L using Amicon Ultra
Centrifugal Filters with a 3-kDa cutoff (Millipore). A concentrated protein solution (10.5 L) was subsequently subjected
to in-solution digestion with 3.1 L of 1 M ammonium bicarbonate (pH 8.5) and 1.5 L of 100 mM dithiothreitol. These
solutions were incubated at 95 C for 5 min and subsequently
allowed to cool. Then, 3 L of iodoacetamide was added to
each solution, followed by a 20-min incubation in the dark at
room temperature; 1 L of 0.1 mg/mL proteomics-grade trypsin (Sigma-Aldrich, MO, USA) was then added, followed by
an overnight incubation at 37 C. The reactions were stopped
by adding 1 L trifluoroacetic acid. These solutions were
diluted with solvent A in a 1:5 ratio and filtered using
Ultrafree-MC Centrifugal Devices (Millipore) with a
0.45-m pore size. Finally, the filtered solution (100 L)
was used for LC/ESI-TOF MS.
Peptide mass fingerprinting
Peptide mass fingerprinting was performed on a Dionex Ultimate 3000 liquid chromatography system (Dionex) connected
to a micrOTOFII TOF-MS system (Bruker Daltonics). Separations were performed on a Jupiter 4u proteo 90A column
(Phenomenex) with a 4-m particle diameter (column dimensions 2.0150 mm). A linear gradient of 550 % of solution B
was applied for 45 min. Other parameters were as described in
the BLC/ESI-TOF MS^ section.
Compound detection and monoisotopic [M+H]+ listing
from total ion chromatograms were automatically performed
with Bruker Compass Data Analysis software. The proteins
were identified by PMF. The Mascot search engine (Matrix

7137

Science) was used for protein identification in the NCBI protein database. The database search parameters were carbamidomethylation as a fixed modification and methionine oxidation as a variable modification. Up to two missed tryptic
cleavages were allowed, and the peptide mass tolerance was
0.5 kDa.
ELISA
Sterile cotton swabs were cut into approximately 55-mm
squares and extracted with 100 L of 50 mM bicarbonate
buffer (BCB, pH 9.6). Other body fluids and vaginal fluid
extracts were diluted 1:100 with BCB. The filter papers used
to collect sweat were cut into approximately 1010-mm
squares and extracted with 250 L of 50 mM BCB. Recombinant SPRR3 was diluted from 0.3 ng/mL to 30 g/mL and
recombinant FABP5 was diluted from 1.5 ng/mL to 15 g/mL
with BCB. The ELISA procedure was previously reported [5].
As primary antibodies, a monoclonal anti-human SPRR3 antibody (Abnova, Taipei, Taiwan) diluted 1:1000 with 0.05 %
Tween-20 in phosphate buffered saline (PBST) or a polyclonal anti-human FABP5 antibody (R&D Systems, MN, USA)
diluted 1:500 in PBST was used. As secondary antibodies, a
horseradish peroxidase (HRP)-conjugated goat anti-mouse
IgG (KPL, MD, USA) diluted 1:5000 with PBST or a HRPconjugated rabbit anti-goat IgG (Sigma-Aldrich) diluted
1:1000 with PBST was used. For color development, 50 L
TMB+Substrate Chromogen (Dako Cytomation, CA, USA)
was added, and then the reaction was stopped by adding
50 L of 1 M H2SO4. Absorbance at 450 nm was measured
on a Molecular Devices SPECTRAmax PLUS 384 plate reader (Molecular Devices, CA, USA). Each absorbance value
was normalized by subtracting the blank absorbance of the
primary antibody solution. Standard curves from SPRR3 and
FABP5 ELISAs were obtained using the recombinant
proteins.
Limit of detection and quantification for LC/ESI-TOF MS
LC/ESI-TOF MS was performed using recombinant SPRR3
(2.5, 10, 20, 40, and 80 g/mL) and FABP5 (2.5, 5, 10, and
20 g/mL). The experimental parameters were the same as
described in the BLC/ESI-TOF MS^ section. Standard curves
for the proteins were obtained using the extracted ion chromatograms (EIC) peak area (1561.00.2 for SPRR3, 1074.7
0.2 for FABP5).
Potential forensic analysis by LC/ESI-TOF MS
Saliva samples collected from children were diluted 2-fold
with solvent A.

7138

Menstrual cycle vaginal samples collected once per week

were prepared as described in the BSample preparation^
section.
The dilution limit samples were prepared as follows:
average vaginal elution was obtained by mixing equal volumes of eluted samples collected from five volunteers. The
samples were diluted 1:2 to 1:32 with solvent A before
filtration.
Mixed body fluid samples were prepared as follows: average vaginal elutions were diluted 1:2 to 1:128 with other body
fluids or elutions.
Highly concentrated body fluid samples were prepared as
follows: 1 mL of nasal secretion, saliva, semen, urine, or sweat
was dried using a Centrivap concentrator (Labconco, MO,
USA). The obtained pellets were dissolved in 10 L ultrapure water and 40 L solvent A.
Two simulated casework samples were prepared as follows: (1) the sterile cotton swabs with vaginal fluid were cut
into approximately 55-mm square pieces and air-dried at
room temperature for 3 days; (2) a condom was air-dried after
sexual intercourse at room temperature overnight, and the outside surface was wiped with sterile cotton swabs. The swabs
were cut into approximately 55-mm square pieces, and extracts were obtained from these pieces using 100 L ultra-pure
water.
Menstrual cycle, dilution, mixed body fluids, and casework
samples were diluted five times with solvent A. Highly concentrated samples were not diluted. All solutions were filtered
through Ultrafree-MC Centrifugal Devices (Millipore) with a
pore size of 0.45 m. The filtered solutions of highly concentrated nasal secretion and semen samples were collected in
only approximately 5 L because of the high viscosity of
these fluids. Therefore, 5 L of each filtered solution was used
for the highly concentrated nasal secretion and semen analysis. One hundred microliters of each filtered pediatric saliva
solution and 25 L of each of the other filtered solutions were
used for LC/ESI-TOF MS.
For blinding, the investigator had no knowledge of the
numbering of filtered body fluid samples (n=1; vaginal secretions, sweat, nasal secretions, saliva, semen, and urine).
Twenty-five microliters of each filtered solution was subsequently used for LC/ESI-TOF MS.

Data analysis
The Peptide Mass program, available at SwissProt (http://us.
expasy.org/tools), was used to compare the experimental mass
values with the average theoretical mass values.
Absorbance was statistically analyzed using a one-way
analysis of variance with Scheffs multiple-comparison test.
The flowchart in Fig. 1 provides an overview of the different experiments.

A. Igoh et al.

Searching for vaginal characteristic proteins by LC/TOF-MS

Step 1. Comparing common fraction among vaginal fluids
Step 2. Confirming of the character of the fraction in body fluids

Identification of the fraction

1. Peptide mass fingerprinting by LC/TOF-MS
2. Confirming the results of identification and specificity by ELISA

Forensic potential analysis by LC/TOF-MS

1. Menstrual cycle
2. Childrens saliva
3. Dilution limit
4. Mixed or concentrated body fluids
5. Casework samples
6. Blind test

Fig. 1 Flowchart of the experiments

Results
Analysis of adult body fluids
The following two-step screening procedure was conducted to
detect vaginal specific or characteristic proteins in adult body
fluid samples by LC/ESI TOF-MS.
First, an investigation was conducted for commonly detectable fractions among vaginal fluid samples based on comparison of chromatograms monitored by PDA at 220 nm and MS
spectra, because in total ion chromatograms peaks overlapped,
to identify potential fractions. Typical ultraviolet chromatograms of vaginal fluid samples are shown in Fig. 2.
Next, to determine whether the fractions commonly observed in vaginal fluid samples were uniquely detected in
vaginal fluids, EICs were selected at the high ion intensity
m/z values to compare various body fluid samples [MS spectra
and EIC profiles of three fractions (F1F3) are shown in
Fig. 3; 1015.00.2 for F1, 1063.50.2 for F2, and 887.7
0.2 for F3]. MS has higher sensitivity than PDA in terms of
detection and thus the comparison of EICs may be more reliable than the comparison of UV chromatograms. On the basis
of the results of this comparison, three EIC peaks were detected only in vaginal samples and not observed at the same retention times in the other body fluids.
Protein identification
The three collected solutions were analyzed using LC/ESITOF MS and PMF. A database search was performed with

Identification and evaluation of potential forensic marker

Fig. 2 Typical UV
chromatograms of vaginal
samples (a), recombinant SPRR3
(b), and recombinant FABP5 (c).
Arrows show the positions of the
peaks identified by peptide mass
fingerprinting as characteristic
proteins for vaginal fluids. The
retention times of identified peaks
did not correspond with those of
recombinant proteins because of
slight differences in the amino
acid sequences and modifications

7139
150

F1

F2

F3

100

Absorbance at 220 nm (mAU)

50
0
150

100
50
0
150

100
50
0
10

20

30

40

50

60

70

80

90

Time [min]

+17
1015.0

+19

+13 +12
+11

100

+15

+13
+10

+12

+20
+14

50

+13

+12 +11
+10

+9

+8

RT: 32.8

2.0
vaginal
semen
nasal
saliva
urine
sweat

f
RT: 40.5

2.0

vaginal
semen
nasal
saliva
urine
sweat

1.0
0

0
1000

1500

m/z

2000

4.0

Intensity ( 103)

500

RT: 32.5

vaginal
semen
nasal
saliva
urine
sweat

0
3.0

+17
887.7
+16

2.0

0
4.0

+18
1004.5
+17
+20
1063.5

50

100

+9

+10

Intensity ( 103)

Relative Abundance
Relative Abundance

EIC profile

+14

50

F3

+16

+20

F2

4.0

Intensity ( 103)

100

excluded): (1) a 329 protein score for a human SPRR3 (Mav

value of 17,328, NCBI accession no. gi685073); (2) a 324
protein score for Homo sapiens SPRR3 (Mav value of 18,
154, NCBI accession no. gi4885607); (3) Homo sapiens
SPRR3 (M av value of 18,140); and (4) Homo sapiens
esophagin (also called SPRR3, Mav value of 18,163). The

Intensity ( 103)

F1

Relative Abundance

the Mascot search engine (see Electronic Supplementary

Material (ESM) Tables S1S4). Finally, F1F3 were identified. Average molecular mass (Mav) values were obtained
by deconvoluting the F1F3 protein MS spectra (Table 1).
The Mascot searches for F1 and F2 provided the following
four of eight best matching results (non-human proteins were

g
RT: 39.7
RT: 31.4

3.0

2.0
1.0
FABP5
SPRR3

0
25

Fig. 3 Typical MS spectra of commonly detected constituents in vaginal

samples (ac), and EIC profiles of body fluids (df) and recombinant
proteins (g). The selected m/z values for EIC profiles were 1015.00.2
(d), 1063.50.2 (e), 887.70.2 (f), 1561.00.2 (g: for SPRR3), and

30

35
40
Time (min)

45

1074.70.2 (g: for FABP5). The EIC peaks detected in vaginal fluid
samples were not observed for other body fluids (sweat, urine, saliva,
nasal, and semen) at the same retention times. EIC extracted ion
chromatograms, RT retention time

7140
Table 1

A. Igoh et al.
Proteins identified in Mascot searches

Fraction Mav exp.

no.

Protein
NCBI
Mascot Sequence
identified accession No. score
coverage (%)

17,2372 SPRR3

gi685073

329

93

18,0632 SPRR3

gi4885607

324

88

15,0751 FABP5

gi4557581

111

86

Mascot scores describe the significance of the Mascot search engine

results with a significance threshold of 86. Sequence coverage is defined
as the ratio of the identified amino acids to the total number of amino
acids in the protein (%)
Mav average molecular mass

reported protein variant (NCBI accession no. gi4885607;

Mav =18,140) differed from the protein with an Mav value of
18,154 because of the Leu149Val substitution; the Gln19
His substitution yielded an Mav value of 18,163. The protein
with an Mav value of 17,328 had the Leu149Val substitution
plus the loss of an 8-amino acid repeat (GCTKVPEP) in the
central domain (57104).
None of these mass values corresponded with the experimental Mav values for the F1 and F2 proteins. The experimental Mav values were consistent with an acetylated (+42) Nterminal region lacking the initiator methionine residue
(131), as observed in F1 (agreement with an Mav value of
17,328) and F2 (agreement with an Mav value of 18,154).
These modifications had been previously reported for human
SPRR3 from the saliva of a preterm newborn [15]; the same
modifications had apparently occurred in adult vaginal fluids.
The Mascot search for F3 provided the following three of
six best matching results (non-human proteins were excluded): (1) a protein score of 111 for human fatty acid binding
protein 5 (FABP5, Mav value of 15,164, NCBI accession no.
gi4557581); (2) Homo sapiens Chain A, FABP5 complex
with inhibitor Bms-309413 (Mav value of 15,436); and (3) a
protein score of 98 for Homo sapiens Chain A, FABP5

ELISA detection of SPRR3 and FABP5

SPRR3 and FABP5 ELISAs were used to confirm the results
of identifications and expression levels in body fluid samples.
All types of collected adult body fluids were analyzed. As
shown in Fig. 4, SPRR3 and FABP5 were detected in all
vaginal samples. The SPRR3 and FABP5 absorbance values
were higher in vaginal fluids than in the other body fluids. The
average vaginal fluid absorbance values diluted 1:100 were
0.810.40 for SPRR3 (3.6 g/mL) and 0.820.36 for FABP5
(0.1 g/mL), and these were significantly higher than those of
the other body fluids (p<0.01). In the SPRR3 ELISA, a saliva
sample showed the second highest absorbance value (0.004;
75 ng/mL). In the FABP5 ELISA, a nasal sample showed the
second highest absorbance value (0.06; 1 ng/mL).
Limit of detection and quantification
Recombinant SPRR3 and FABP5 were used to clarify the
limits of detection and quantification (LODs and LOQs, respectively) for SPRR3 and FABP5 by LC/ESI-TOF MS. The
LODs were estimated to be 20 g/mL for SPRR3 and 5 g/
mL for FABP5, and the LOQs were estimated to be 40 g/mL
for SPRR3 and 10 g/mL for FABP5.
Potential forensic analysis
To clarify the potential for the use of the identified proteins as
forensic markers, we analyzed pediatric saliva, menstrual

1.5

Absorbance (450 nm)

Absorbance (450 nm)

complex with endocannabinoid anandamide (Mav value of

15,422). Similar to F1 and F2, none of these mass values
corresponded to the experimental Mav values of the F3 protein.
The experimental Mav value was consistent with that of an
acetylated N-terminal region lacking the initiator methionine
residue of FABP5 (NCBI accession no. gi4557581).

1.0

0.5

0.0

1.5

1.0

0.5

0.0

nasal

saliva

urine

semen vaginal sweat

Fig. 4 Anti-SPRR3 (a) and anti-FABP5 (b) activities against different

body fluids as determined by ELISA. Body fluids were diluted 1:100 with
bicarbonate buffer. Values are shown as meansstandard deviations for
22 nasal secretions, 30 saliva, 26 urine, 17 semen, 14 vaginal secretions,

nasal

saliva

urine

semen vaginal sweat

and 22 sweat samples. The absorbance values for SPRR3 and FABP5
detection in vaginal fluids were significantly higher than those in other
body fluids (*p<0.01) as determined by one-way ANOVA with Scheffs
multiple-comparison test

Identification and evaluation of potential forensic marker

7141

cycle, dilution limit, mixed body fluid, highly concentrated

body fluid, casework, and blind samples. LC/ESI-TOF MS
was used to examine the differences in SPRR3 isoforms detected in these samples.
SPRR3 is highly expressed in the saliva of preterm newborns [15]; therefore, the saliva of children may contain a
larger amount of SPRR3 than adult saliva. However, the three
peaks were not observed in the EIC profiles of pediatric saliva
samples, and the corresponding peaks in the MS spectra were
not observed at any retention time.
In the analysis of menstrual cycle samples, the three peaks
were commonly detected in the EIC profiles of vaginal samples collected weekly, and these features in the MS spectra
were confirmed (Table 2).
Forensic samples often are very small in volume and
thus the dilution limit for detection is useful information
for forensic investigations. The F1F3 peaks were detectable in 0.31 L of average vaginal elution (25 L of 1:80
dilution) for F1, 0.63 L (25 L of 1:40 dilution) for F2,
and 0.16 L (25 L of 1:160 dilution) for F3,
respectively.
Forensic samples often are collected as a mixed body fluid,
and the results for mixed samples are shown in Table 3. Three
peaks were observed in all of the EIC profiles of equal volume
mixed samples; however, they were not observed in all the
EIC profiles of samples diluted 1:16 to 1:128.
If low levels of SPRR3 and FABP5 are expressed in other
body fluids, in the forensic evaluation of vaginal fluid samples, a false positive may be caused by the concentration of
body fluids. The three peaks of F1F3 were not observed in
any of the EIC profiles of highly concentrated body fluid
samples. The three peaks were observed in all of the EIC
profiles for simulated casework samples.
The blindly analyzed samples included nasal secretions
(n=1), saliva (n=1), semen (n=1), vaginal fluid (n=1), urine
(n=1), and sweat (n=1). The three peaks were detected only
in one sample among the six blindly analyzed samples
(Table 4). Thus, the investigator could have identified the
vaginal sample without a false positive result for the other
body fluid samples.

Table 3
samples
Sample

Sample

F1

F2

F3

1
2
3
4

+
+
+
+

+
+
+
+

+
+
+
+

week
week
week
week

The selected m/z values for EIC profiles

were 1015.00.2 (F1), 1063.50.2 (F2),
and 887.70.2 (F3), respectively

Fraction

Vaginal + nasal

Vaginal + saliva

Vaginal + semen

Vaginal + urine

Vaginal + sweat

Dilution ratio
2

16

32

64

128

F1

ND

ND

ND

ND

ND

F2

ND

ND

ND

ND

F3
F1

+
+

+
+

ND
ND

ND
ND

ND
ND

ND
ND

F2
F3

+
+

+
+

ND
ND

ND
ND

ND
ND

ND
ND

F1

ND

ND

ND

ND

ND

F2
F3

+
+

ND
ND

ND
ND

ND
ND

ND
ND

ND
ND

F1
F2

+
+

ND
ND

ND
ND

ND
ND

ND
ND

ND
ND

F3
F1
F2

+
+
+

+
+
+

ND
ND
ND

ND
ND
ND

ND
ND
ND

ND
ND
ND

F3

ND

ND

ND

ND

The selected m/z values for EIC profiles were 1015.00.2 (F1), 1063.5
0.2 (F2), and 887.70.2 (F3), respectively. Vaginal fluids were diluted
with nasal secretion, saliva, semen, urine, or sweat sample
ND not detected

Discussion
This study aimed to identify specific or characteristic proteins
in vaginal fluids and to assess their potential in forensic analyses. Using LC/ESI-TOF MS and PMF, we identified three
characteristic proteins, including two SPRR3 isoforms and
one FABP5 (Fig. 2 and Table 1). One or both SPRR3 isoforms
were detected in all vaginal fluid samples, and FABP5 was
detected in all vaginal fluid samples. Moreover, these proteins
were detected at markedly high levels in all vaginal fluid samples (Fig. 4). Although the use of another proteinase for PMF
would have improved the accuracy of the identifications, only
ELISA was used to facilitate protein identification.
Table 4

Table 2 Detection of
EIC peaks of F1F3 in
vaginal samples
collected once per week

Detection of EIC peaks of F1F3 in mixed body fluids

Detection of EIC peaks of F1F3 in blindly analyzed samples

Blind sample number

Source

F1

F2

F3

1
2

Urine
Saliva

3
4
5
6

Nasal
Vaginal
Sweat
Semen

The selected m/z values for EIC profiles were 1015.00.2 (F1), 1063.5
0.2 (F2), and 887.70.2 (F3), respectively

7142

SPRR3 has not been detected in normal adult epidermis but

was reportedly abundant in the oral epithelium, including the
tongue and tonsils [16, 17], and was also detected in adult
human saliva [18]. However, under the experimental conditions of the present study, SPRR3 was not detected in adult
saliva samples, suggesting that this protein is generally not
expressed at high levels in the saliva of healthy adults
(Fig. 4). This discrepancy in SPRR3 detection in adult saliva
samples may be because of the differences in sample amounts
and/or the sensitivity of LC/MS. SPRR3 expression is strongly induced during human epidermal keratinocyte differentiation; therefore, it has been considered as a squamous epithelial
marker [19, 20]. Furthermore, SPRR3 has two mRNA splice
variants, SPRR3-v1 and -v2. SPRR-v1 is highly expressed in
the esophageal mucosa of healthy individuals, whereas
SPRR3-v2 expression is higher in the adjacent mucosal tissues in patients with esophageal squamous cell carcinoma
[21]. High SPRR3 expression also has been reported in colorectal cancer, breast cancer, and glioblastoma multiforme (the
most common primary brain tumor in adults); thus, this protein is a reported candidate biomarker for cancer diagnosis
[2224].
In contrast, FABP5 was first detected and cloned from
keratinocytes of patients with psoriasis [25]. FABP5 is
expressed at low levels in the normal epidermis but at high
levels in human cultured keratinocytes and at even higher
levels in psoriatic lesions [26]. It has been reported that
FABP5 expression is up to 4-fold higher in primary tumors
than in the corresponding metastatic tissues; therefore, it has
been considered as an important marker of metastasis in squamous cell carcinoma of the tongue [27]. High FABP5 RNA
expression has been reported in triple-negative breast cancers,
which do not express the three receptors targeted by most
chemotherapies and thus FABP5 was believed to be a novel
prognostic biomarker in estrogen/progesterone receptornegative and retinoic acid-resistant breast cancers [28].
In the present study, SPRR3 and FABP5 were detected in
all vaginal fluid samples but not in other fluid samples
(Fig. 4). The reported polymorphism hypothesis regarding
the expression of SPRR3 isoforms was based on protein identification in the saliva of a preterm newborn [15]. The same
polymorphism may occur in adult vaginal fluids. Meanwhile,
the MS spectra of the two SPRR3 isoforms and one FABP5
were detected, regardless of the menstrual cycle, although
samples from various cycle phases were collected from only
one volunteer. Moreover, in samples from two postmenopausal participants and one pregnant participant, the MS spectra
for SPRR3 and FABP5 were detected. These results suggest
the potential for using these proteins as forensic markers.
Moreover, because SPRR3 expression was reportedly higher
in the saliva of preterm human newborns than in that of adults
[15], we analyzed 14 saliva samples from children, including
infants, by LC/ESI-TOF MS using the same volume as that

A. Igoh et al.

previously reported [15] (10 the volume of the adult saliva

samples in the present study). SPRR3 was not detected in the
MS spectra of these samples, suggesting that SPRR3 is either
not expressed or expressed at extremely low levels in the
saliva of healthy children. This result indicates that the vaginal
fluid can possibly be discriminated from the saliva of children
in forensic investigations. The sample population used in the
present study was small; thus, further studies are required to
clarify the association between the proteins present in vaginal
fluids with the menstrual cycle, menopause, and pregnancy.
The present study demonstrated that SPRR3 and FABP5
were more abundantly expressed in vaginal fluids compared
with other body fluids. This finding suggests the active differentiation of human epidermal keratinocytes in the adult vagina, which may serve the physiological function of defense
against external pathogens. However, further studies are required to elucidate the functional details. Meanwhile, the expression of these proteins has more recently been associated
with cancer and accordingly these proteins have been reported
as potential candidate markers for various early cancers, particularly carcinoma [2124, 27, 28]. We believe that the present study provides important information regarding the evaluation of these proteins as cancer markers. Further studies are
required for determining the potential of these markers for
vaginal cancers.
The dilution limits for SPRR3 and FABP5 in vaginal fluid
samples may be useful in actual forensic samples. These proteins were detectable in equal volume mixed body fluid samples. Moreover, these proteins were not detectable in highly
concentrated body fluid samples. If these proteins were
expressed in the body fluids that were second highest in absorbance value in ELISAs, the estimated concentrations were
lower than the LODs in LC/ESI-TOF MS. Therefore, these
proteins may not be detected in other body fluids by LC/ESITOF MS even if expressed. The presented values for the
LODs may be useful as concentration thresholds for discriminating vaginal fluid by more sensitive MS. Blindly analyzed
samples were accurately identified using these proteins as
markers without false positive results. Moreover, the analysis
of simulated casework samples revealed that these proteins
may be useful in forensic investigations, specifically the identification of vaginal fluid. Thus, the present study has provided useful information regarding vaginal fluid identification in
forensic science applications. SPRR3 has been previously reported as a characteristic protein of vaginal fluid; however, its
potential as a marker for vaginal fluid identification has not
been examined [7]. The present study has initially proven its
potential as a marker. SPRR3 is commonly identified as a
forensic vaginal fluid marker via another approach dependent
on, e.g., race, sample preparation, and separation; this property may increase its potential for use as a marker because vaginal fluids can change according to the menstrual cycle, pregnancy, the use of contraceptive agents, and the frequency of

Identification and evaluation of potential forensic marker

sexual intercourse [11, 12]. FABP5 was initially identified as a

characteristic protein of vaginal fluid; it is also detected at low
levels in nasal secretions, saliva, and semen (Fig. 4), which
may present nonspecific absorbances in ELISAs. Notably,
such absorbances were observed regardless of the dilution
ratio of body fluid samples (data not shown). Using a cutoff
value, the examination of vaginal fluids may be possible because of the marked difference in the absorbance values in
body fluids, which is similar to that of statherin [5], Tamm
Horsfall protein [29], and dermcidin [30].
At present, MS has been rarely used for biological material
analysis in forensic institutes. Various biological markers are
required for forensic biological analysis, and MS is useful for
identifying a novel marker similar to the markers described in
other biological fields. Further, the potential of forensic investigatory methods has been described [14, 7, 13], and MS could
be a useful tool in forensic biomaterial analysis if it is possible
to shorten the sample preparation time and prove the advantages of its use.
In conclusion, the present study revealed high expression
levels of two SPRR3 isoforms and one FABP5 isoform in
vaginal fluid samples from healthy adults. In forensic science,
these proteins may serve as promising markers for the identification of vaginal fluids.
Acknowledgments We wish to thank the volunteers who kindly donated samples for this study. We also wish to thank Dr. Osamu Noguchi and
Mr. Atsushi Yokoyama, Forensic Science Laboratory of Okayama Prefectural Police H.Q., for technical advice. This study was partially supported by the Japan Society for the Promotion of Science (JSPS) KAKE
NHI Grant Number 25933002.

7143

8.

9.

10.

11.

12.

13.

14.

15.

16.

References
17.
1.

Brettell TA, Butler JM, Almirall JR (2011) Forensic science. Anal

Chem 83(12):45394556. doi:10.1021/ac201075e
2. Virkler K, Lednev IK (2010) Raman spectroscopic signature of
blood and its potential application to forensic body fluid identification. Anal Bioanal Chem 396(1):525534. doi:10.1007/s00216009-3207-9
3. Itoh Y, Matsuzawa S (1990) Detection of human hemoglobin A
(HbA) and human hemoglobin F (HbF) in biological stains by
microtiter latex agglutination-inhibition test. Forensic Sci Int
47(1):7989
4. Stowell LI, Sharman LE, Hamel K (1991) An enzyme-linked immunosorbent assay (ELISA) for prostate-specific antigen. Forensic
Sci Int 50(1):125138
5. Akutsu T, Watanabe K, Fujinami Y, Sakurada K (2010)
Applicability of ELISA detection of statherin for forensic identification of saliva. Int J Legal Med 124(5):493498. doi:10.1007/
s00414-009-0391-2
6. Doi M, Gamo S, Okiura T, Nishimukai H, Asano M (2014) A
simple identification method for vaginal secretions using relative
quantification of Lactobacillus DNA. Forensic Sci Int Genet 12C:
9399. doi:10.1016/j.fsigen.2014.05.005
7. Van Steendam K, De Ceuleneer M, Dhaenens M, Van Hoofstat D,
Deforce D (2013) Mass spectrometry-based proteomics as a tool to

18.

19.

20.

21.

identify biological matrices in forensic science. Int J Legal Med

127(2):287298. doi:10.1007/s00414-012-0747-x
Dasari S, Pereira L, Reddy AP, Michaels J-EA LX, Jacob T,
Thomas A, Rodland M, Charles T, Roberts J, Gravett MG,
Nagalla SR (2007) Comprehensive proteomic analysis of human
cervical-vaginal fluid. J Proteome Res 6:12581268
Pereira L, Reddy AP, Jacob T, Thomas A, Schneider KA, Dasari S,
Lapidus JA, Lu X, Rodland M, Charles T, Roberts J, Gravett MG,
Nagalla SR (2007) Identification of novel protein biomarkers of
preterm birth in human cervical-vaginal fluid. J Proteome Res 6:
12691276
Gravett MG, Thomas A, Schneider KA, Reddy AP, Dasari S, Jacob
T, Lu X, Rodland M, Pereira L, Sadowsky DW, Roberts CT Jr,
Novy MJ, Nagalla SR (2007) Proteomic analysis of cervicalvaginal fluid: identification of novel biomarkers for detection of
intra-amniotic infection. J Proteome Res 6(1):8996. doi:10.1021/
pr060149v
Zegels G, Van Raemdonck GA, Coen EP, Tjalma WA, Van Ostade
XW (2009) Comprehensive proteomic analysis of human cervicalvaginal fluid using colposcopy samples. Proteome Sci 7:17. doi:10.
1186/1477-5956-7-17
Schwebke JR, Richey CM, Weiss HL (1999) Correlation of behaviors with microbiological changes in vaginal flora. J Infect Dis
180(5):16321636. doi:10.1086/315065
Nakazono T, Kashimura S, Hayashiba Y, Hisatomi T, Hara K
(2002) Identification of human urine stains by HPLC analysis of
17-ketosteroid conjugates. J Forensic Sci 47(3):568572
Donfack J, Wiley A (2015) Mass spectrometry-based cDNA profiling as a potential tool for human body fluid identification.
Forensic Sci Int Genet 16C:112120. doi:10.1016/j.fsigen.2014.
12.008
Manconi B, Cabras T, Pisano E, Nemolato S, Inzitari R, Iavarone F,
Fanali C, Sanna MT, Tirone C, Vento G, Romagnoli C, Faa G,
Castagnola M, Messana I (2010) Characterization of two isoforms
of human SPRR3 from saliva of preterm human newborn and
autoptic fetal oral mucosa, parotid and submandibular gland samples. Biochem Biophys Res Commun 398(3):477481. doi:10.
1016/j.bbrc.2010.06.103
Chen BS, Wang MR, Cai Y, Xu X, Xu ZX, Han YL, Wu M (2000)
Decreased expression of SPRR3 in Chinese human oesophageal
cancer. Carcinogenesis 21(12):21472150
Abraham JM, Wang S, Suzuki H, Jiang HY, Rosenblum-Vos LS,
Yin J, Meltzer SJ (1996) Esophagin cDNA cloning and characterization: a tissue-specific member of the small proline-rich protein
family that is not expressed in esophageal tumors. Cell Growth
Differ 7(7):855860
Yan W, Apweiler R, Balgley BM, Boontheung P, Bundy JL, Cargile
BJ, Cole S, Fang X, Gonzalez-Begne M, Griffin TJ, Hagen F, Hu S,
Wolinsky LE, Lee CS, Malamud D, Melvin JE, Menon R, Mueller
M, Qiao R, Rhodus NL, Sevinsky JR, States D, Stephenson JL,
Than S, Yates JR, Yu W, Xie H, Xie Y, Omenn GS, Loo JA,
Wong DT (2009) Systematic comparison of the human saliva and
plasma proteomes. Proteomics Clin Appl 3(1):116134. doi:10.
1002/prca.200800140
Gibbs S, Fijneman R, Wiegant J, van Kessel AG, van De Putte P,
Backendorf C (1993) Molecular characterization and evolution of
the SPRR family of keratinocyte differentiation markers encoding
small proline-rich proteins. Genomics 16(3):630637. doi:10.1006/
geno.1993.1240
Zhang Y, Feng YB, Shen XM, Chen BS, Du XL, Luo ML, Cai Y,
Han YL, Xu X, Zhan QM, Wang MR (2008) Exogenous expression
of Esophagin/SPRR3 attenuates the tumorigenicity of esophageal
squamous cell carcinoma cells via promoting apoptosis. Int J
Cancer 122(2):260266. doi:10.1002/ijc.23104
de Simo AT, Souza-Santos PT, de Oliveira DS, Bernardo V, Lima
SC, Rapozo DC, Kruel CD, Faria PA, Ribeiro Pinto LF, Albano

7144

22.

23.

24.

25.

A. Igoh et al.
RM (2011) Quantitative evaluation of SPRR3 expression in esophageal squamous cell carcinoma by qPCR and its potential use as a
biomarker. Exp Mol Pathol 91(2):584589. doi:10.1016/j.yexmp.
2011.06.006
Cho DH, Jo YK, Roh SA, Na YS, Kim TW, Jang SJ, Kim YS, Kim
JC (2010) Upregulation of SPRR3 promotes colorectal tumorigenesis. Mol Med 16(7-8):271277. doi:10.2119/molmed.2009.00187
Kim JC, Yu JH, Cho YK, Jung CS, Ahn SH, Gong G, Kim YS, Cho
DH (2012) Expression of SPRR3 is associated with tumor cell
proliferation in less advanced stages of breast cancer. Breast
Cancer Res Treat 133(3):909916. doi:10.1007/s10549-011-18685
Liu Q, Zhang C, Ma G, Zhang Q (2014) Expression of SPRR3 is
associated with tumor cell proliferation and invasion in glioblastoma multiforme. Oncol Lett 7(2):427432. doi:10.3892/ol.2013.
1736
Madsen P, Rasmussen HH, Leffers H, Honore B, Celis JE (1992)
Molecular cloning and expression of a novel keratinocyte protein
(psoriasis-associated fatty acid-binding protein [PA-FABP]) that is
highly up-regulated in psoriatic skin and that shares similarity to
fatty acid-binding proteins. J Invest Dermatol 99(3):299305

26.

Siegenthaler G, Hotz R, Chatellard-Gruaz D, Jaconi S, Saurat JH

(1993) Characterization and expression of a novel human fatty acidbinding protein: the epidermal type (E-FABP). Biochem Biophys
Res Commun 190(2):482487
27. Uma RS, Naresh KN, DCruz AK, Mulherkar R, Borges AM
(2007) Metastasis of squamous cell carcinoma of the oral tongue
is associated with down-regulation of epidermal fatty acid binding
protein (E-FABP). Oral Oncol 43:2732. doi:10.1016/j.
oraloncology.2005.12.024
28. Liu RZ, Graham K, Glubrecht DD, Germain DR, Mackey JR,
Godbout R (2011) Association of FABP5 expression with poor
survival in triple-negative breast cancer: implication for retinoic
acid therapy. Am J Pathol 178(3):9971008. doi:10.1016/j.ajpath.
2010.11.075
29. Akutsu T, Ikegaya H, Watanabe K, Fukushima H, Motani H, Iwase
H, Sakurada K (2010) Evaluation of Tamm-Horsfall protein and
uroplakin III for forensic identification of urine. J Forensic Sci
55(3):742746. doi:10.1111/j.1556-4029.2010.01325.x
30. Sakurada K, Akutsu T, Fukushima H, Watanabe K, Yoshino M
(2010) Detection of dermcidin for sweat identification by realtime RT-PCR and ELISA. Forensic Sci Int 194(1-3):8084. doi:
10.1016/j.forsciint.2009.10.015