Method for the determination of vitamin K homologues in human plasma using high-performance liquid chromatography-tandem mass spectrometry

Article abstract of DOI:10.1021/ac0489667

We report here the development of a precise and sensitive method for the determination of vitamin K homologues including phylloquinone (PK), menaquinone-4 (MK-4), and menaquinone-7 (MK-7) in human plasma using HPLC-tandem mass-mass spectrometry with atmospheric pressure chemical ionization (LC-APCI-MS/MS). The method involves the use of stable isotope ¹⁸O-labeled internal standard compounds, which were synthesized in our laboratory, and the selection of a precursor and product ion with a MS/MS multiple reaction monitoring method. The average intraassay and interassay variation values for PK, MK-4, and MK-7 were <10%. Average spiked recoveries from authentic compounds added to normal human plasma samples for PK, MK-4, and MK-7 were 98-102%. Mean plasma concentrations of PK, MK-4, and MK-7 from healthy subjects (n = 20) were 1.22 ± 0.57, 0.39 ± 0.46, and 6.37 ± 7.45 ng/mL, respectively. We conclude that this novel LC-APCI-MS/MS method should be useful for the evaluation of vitamin K status in postmenopausal women and elderly subjects and provides useful information for the treatment and prevention of osteoporosis with vitamin K.

Full text of DOI:10.1021/ac0489667

Anal. Chem. 2005, 77, 757-763
Articles
Method for the Determination of Vitamin K
Homologues in Human Plasma Using
High-Performance Liquid Chromatography-Tandem
Mass Spectrometry
Yoshitomo Suhara, Maya Kamao, Naoko Tsugawa, and Toshio Okano*
Department of Hygienic Sciences, Kobe Pharmaceutical University, 4-19-1, Motoyamakita-machi, Higashinada-ku,
Kobe 658-8558, Japan
residues.4,5 Thus, several of the biological activities of vitamin K
homologues such as phylloquinone (PK, vitamin K1) isolated from
green plants, menaquinones (MK-n, vitamin K2) synthesized by
microorganisms, and menadione (vitamin K3) have been reported.
Recently, it was reported that menaquinone-4 (MK-4) was the most
potent analogue, and all vitamin K derivatives were converted to
MK-4 in vivo.6,7 In Japan, MK-4 has been given to osteoporotic
patients, and phylloquinone has been used as a therapeutic agent
for vitamin K-deficient syndromes such as hypoprothronmbinemia
in newborn babies and in antibiotic-treated patients. However,
information on the physiological and pharmacological roles of
vitamin K in vivo is still limited.8 One reason for this is that the
detection and monitoring of vitamin K homologues in plasma and
organs have been difficult on account of quite small concentrations
and many kinds of impurities even though measurement of the
vitamin K plasma concentration is essential to optimize therapy.
For pharmacokinetic and epidemiological purposes, specific,
accurate, and sensitive analytical methods are required that allow
assays at low (ng/mL) plasma levels.
Several assay techniques have been described for the mea-
surement of vitamin K concentrations in human plasma. Initially,
high-performance liquid chromatography (HPLC) with ultraviolet
(UV) detection was the first choice for measuring the individual
forms of vitamin K.9 This method offered more selectivity than
the traditional chick bioassay commonly used to measure vitamin
K status; however, its sensitivity was still insufficient. More recent
methods include electrochemical techniques, fluorescence detec-
We report here the development of a precise and sensitive
method for the determination of vitamin K homologues
including phylloquinone (PK), menaquinone-4 (MK-4),
and menaquinone-7 (MK-7) in human plasma using
HPLC-tandem mass-mass spectrometry with atmospheric
pressure chemical ionization (LC-APCI-MS/MS). The
method involves the use of stable isotope 18O-labeled
internal standard compounds, which were synthesized in
our laboratory, and the selection of a precursor and
product ion with a MS/MS multiple reaction monitoring
method. The average intraassay and interassay variation
values for PK, MK-4, and MK-7 were <10%. Average
spiked recoveries from authentic compounds added to
normal human plasma samples for PK, MK-4, and MK-7
were 98-102%. Mean plasma concentrations of PK, MK-
4, and MK-7 from healthy subjects (n ) 20) were 1.22 (
0.57, 0.39 ( 0.46, and 6.37 ( 7.45 ng/mL, respectively.
We conclude that this novel LC-APCI-MS/MS method
should be useful for the evaluation of vitamin K status in
postmenopausal women and elderly subjects and provides
useful information for the treatment and prevention of
osteoporosis with vitamin K.
There is growing interest in the role, biochemical function,
and metabolism of vitamin K in vivo. Vitamin K is a blood clotting
agent.1 It serves as an essential cofactor of the carboxylase
involved in the activation of the blood coagulation cascade
proteins.2,3 Recent investigations indicate that vitamin K is required
for the synthesis of another calcium-binding protein, osteocalcin,
which is important for mineralization in bone. To activate calcium-
binding proteins, vitamin K participates in the carboxylation of
glutamyl residues of osteocalcin to form γ-carboxyglutamyl
(4) Hauschka, P. V.; Lian, J. B.; Gallop, P. M. Proc. Natl. Acad. Sci. U.S.A. 1975,
72, 3925-3929.
(5) Price, P. A.; Otsuka, A. S.; Posner, J. W.; Kristaponis, J.; Rama, N. Proc.
Natl. Acad. Sci. U.S.A. 1976, 73, 1447-1451.
(6) Yamamoto, R.; Komai, M.; Kojima, K.; Furukawa, Y.; Kimura, S. J. Nutr.
Sci. Vitaminol. 1997, 43, 133-143.
(7) Davidson, R. T.; Foley, A. L.; Engelke, J. A.; Suttie, J. W. J. Nutr. 1998,
128, 220-223.
(8) Tadano, K.; Yuzuriha, T.; Sato, T.; Fujita, T.; Shimada, K.; Hashimoto, K.;
Satoh, T. J. Pharmacobio-Dyn. 1989, 12, 640-645.
(9) Lefevere, M. F.; De Leenheer, A. P.; Claeys, A. E. J. Chromatogr. 1979,
186, 749-762.
* To whom correspondence should be addressed. E-mail: t-okano@
kobepharma-u.ac.jp.
(1) Schurgers, L. J.; Dissel. P. E. P.; Spronk, H. M. H.; Soute, B. A. M.; Dhore,
C. R.; Cleutjens, J. P. M.; Vermeer, C. Z. Kardiol. 90: Suppl. 3 2001, III/
57-III/63.
(2) Olson, R. E. Annu. Rev. Nutr. 1984, 4, 281-337.
(3) Suttie, J. W. Annu. Rev. Biochem. 1985, 54, 459-477.
10.1021/ac0489667 CCC: $30.25 © 2005 American Chemical Society
Published on Web 12/23/2004
Analytical Chemistry, Vol. 77, No. 3, February 1, 2005 757

tion after postcolumn reduction,10-13 and gas chromatography-
mass spectrometry with HPLC.14 These techniques provide greater
selectivity and sensitivity than UV detection. The most common
and conventional method is HPLC-fluorescence detection. For the
measurement, naphthoquinone in vitamin K is converted to a
hydroquinone analogue by a platinum oxide catalyst or electro-
chemical reduction after separation by HPLC. Then the hydro-
quinone, which fluoresces when exposed to ultraviolet light of a
wavelength of 320 nm, can be detected with a fluorescence
detector with high sensitivity. Most studies about the quantitation
of vitamin K in human or rat plasma have been carried out with
this method.
Scheme 1. Synthetic Route to 18O-Labeled
Vitamin K Homologues 1-3 and Their Chemical
Yield at Each Step: 1, 18O-Labeled PK; 2,
18O-Labeled MK-4; 3, 18O-Labeled MK-7
We used HPLC-fluorescence detection to measure the con-
centrations of vitamin K homologues (PK, MK-4, menaquinone-7
(MK-7)) in human plasma and found that PK and MK-7 could be
clearly detected as a single peak. However, the peak of MK-4
overlapped with impurities in the plasma and especially appeared
in the shoulder of a large peak of impurity. These results indicate
that detection of the plasma MK-4 level with this method depends
on the threshold, and the accuracy is uncertain. However, it is
essential to accurately determine plasma concentrations of vitamin
K homologues. If human plasma vitamin K levels could be clearly
determined, the data would provide significant information related
to applications to clinical trials. For example, investigating the
relationship between plasma vitamin K and undercarboxylated
osteocalcin levels might elucidate a therapeutic effect for os-
teoporosis. Furthermore, analyzing the distribution of vitamin K
homologues in vivo and their metabolic pathway should provide
valuable information for the development of new drugs. Against
this background, we evaluated a new method of quantitating
plasma vitamin K levels using HPLC-tandem mass-mass spec-
trometry with atmospheric pressure chemical ionization (LC-APCI-
MS/MS) system. One of the advantages of this method is the
much greater sensitivity and selectivity in comparison with other
LC-MS techniques. In this report, we validated the accuracy and
sensitivity of the LC-APCI-MS/MS method using a multiple
reaction monitoring mode (MRM) for the determination of vitamin
K homologues in human plasma.
compounds was prepared in ethanol according to the solubility
of the solute and stored in the dark at -30 °C prior to use. For
the analytical curves, working solutions of the standard mixture,
ranging from 25 to 400 ng/mL, were prepared by dilution of the
stock solution with ethanol. Stock solutions of 10 µg/mL for 18O-
labeled internal standards (PK-18O, MK-4-18O, MK-7-18O) as shown
in Scheme 1 were prepared by dilution in ethanol and stored in
the dark at -30 °C prior to use. Dilution of the solution with
ethanol gave working internal standard solutions of 3.6 and 100
ng/mL, respectively. The 3.6 ng/mL mixture was used for
determination of vitamin K homologues in human plasma samples.
An equal amount of the standard solution of 25-400 ng/mL and
18O-labeled internal standard solution of 100 ng/mL gave the
solution used for the standard curve. The final concentration
ranged from 12.5 to 200 ng/mL in the case of the standard and
contained 50 ng/mL internal standards.
Sample Preparation. A liquid control serum (Wako Pure
Industries, Ltd., Lot. No. DG118) was used for validation of the
LC-APCI-MS/MS method. The serum or donated human plasma
(0.5 mL) in a brown screw-capped Pyrex tube was supplemented
with 3.6 ng (in 1 mL of ethanol) of PK-18O, MK-4-18O, and MK-
7-18O, respectively, as an internal standard. Extra ethanol (1 mL)
was then added to denature the protein and 3 mL of hexane was
added, followed by shaking for 5 min. The solution was centrifuged
at 3000 rpm for 5 min, and the upper layer was separated. The
supernatant was applied to Sep-Pak silica (Waters, USA), which
was washed with 10 mL of hexane, and then eluted with 5 mL of
hexane/diethyl ether (97:3). The eluate was evaporated under
reduced pressure. The dried sample was reconstituted in 60 µL
of ethanol and vortexed for 10 s. The solutions were transferred
to microvials, capped, and placed in a SIL-10AD vp autosampler
rack. Aliquots (30 µL) were automatically injected into the HPLC
system.
EXPERIMENTAL SECTION
Reagents and Chemicals. HPLC-grade solvents and reagents
for chemical synthesis were purchased from Nacalai Tesque, Inc.
(Kyoto, Japan) and Wako Pure Chemical Industries, Ltd. (Osaka,
Japan). The 18O-labeled vitamin K derivatives used as internal
standards were synthesized in our laboratory, and the chemical
identity of the products was confirmed by nuclear magnetic
resonance (NMR) spectrometry and high-resolution MS spectra
(HREIMS) (Scheme 1). Control human serum was purchased
from Wako Pure Chemical Industries, Ltd.. PK, MK-4, and MK-
7, as standard samples, were kindly donated by Eisai Co., Ltd.
(Tokyo, Japan).
Preparation of Standard Solutions. A standard mixture (PK,
MK-4, MK-7) stock solution of 100 µg/mL for the reference
(10) Lambert, W. E., De Leenheer, A. P.; Lefevere, M. F. J. Chromatogr. Sci.
1986, 24, 76-79.
(11) Shino, M. Analyst 1988, 113, 393-397.
(12) Hiraike, H.; Kimura, M.; Itokawa, Y. J. Chromatogr. 1988, 430, 143-148.
(13) Wakabayashi, H.; Onodera, K. Folia Pharmacol. Jpn. 2000, 116, 43-48.
(14) Dolnikowski, G. G.; Sun, Z.; Grusak, M. A.; Peterson, J. W.; Booth, S. L. J.
Nutr. Biochem. 2002, 13, 168-174.
758 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

samples were centrifuged at 2000g for 5 min at room temperature,
and the plasma was separated and stored at -30 °C with shading
until assay of vitamin K content.
Table 1. LC-APCI-MS/MS Parameters
precursor
collision
product
Rt
compound
ion (m/z)
energy (V)
ion(s)a (m/z)
(min)
Chemical Synthesis and Data Analysis. The synthetic
method and the data on 18O-labeled vitamin K homologues as
internal standards are shown in Scheme 1 and described as
follows. The data for compounds 5, 6a, 6b, and 6c were
previously reported.15,16 The 500-MHz 1H NMR spectra of the
synthetic compounds were measured on a Varian VXR-500. All
compounds were dissolved in deuterized chloroform (CDCl3)
(Merk). Chemical shifts are given in ppm (δ) using tetramethyl-
silane as the internal standard. Mass spectra were registered on
a JMS SX-102A instrument. Column chromatography was carried
out on silica gel 60 F254 (Merk). Unless otherwise noted, all
reagents were purchased from commercial suppliers and used
as received.
2-Methyl-1,4-naphthalenediol, 1-Acetate (5). Vitamin K4
monoacetyl derivative 5 was prepared by deacetylation of diacetyl
4 as the reported method.16, 17 In short, a suspension of diacetate
4 (25.0 g, 97 mmol), K2CO3 (4.7 g, 34 mmol), and Na2S2O4 (5.0 g,
29 mmol) in 15% aqueous MeOH (300 mL) was kept for 1 h at
30-40 °C with stirring and then poured into 1 L of cold water.
The solution was extracted with ethyl acetate (3 × 300 mL), and
the organic layers were combined, dried, and concentrated. The
residue was purified by silica gel column chromatography using
5:1 hexane/ethyl acetate as eluent and then recrystallized from
hexane and ethyl acetate to give 5 (18.8 g) in 90% yield.
4-Acetoxy-3-methyl-2-phytyl-1-naphthaleneol (6a). Phytol
(500 mg, 1.7 mmol) was added dropwise to a solution of 5 (250
mg, 1.2 mmol) and BF3‚Et2O (30 µL) in dry dioxane (300 µL)
and ethyl acetate (300 µL). Then the mixture was heated to 50
°C for 3 h. The brown reaction mixture was poured into ice-cold
water (30 mL) and extracted with ether (3 × 50 mL). The ether
extracts were combined and washed with water (25 mL), dried,
and concentrated. Purification by flash chromatography using 10:1
hexanes/ethyl acetate as eluent afforded 6a (266 mg, 47% yield)
as a pale yellow oil.
4-Acetoxy-2-geranylgeranyl-3-methyl-1-naphthaleneol (6b).
Alcohol 6b (360 mg, 53% yield) was obtained according to the
procedure described above for 6a.
4-Acetoxy-2-geranylgeranylfarnesyl-3-methyl-1-naphtha-
leneol (6c). Alcohol 6c (800 mg, 57% yield) was obtained as
described above for 6a.
2-Methyl-3-phytyl-1,4-dimethoxynaphthalene (7a). A solu-
tion of 6a (400 mg, 809 µmol) in THF (10 mL) was added, using
a cannula, to a suspension of 30% potassium hydride, dispersion
in mineral oil (128 mg, 3.20 mmol) in THF (20 mL), under argon
at 0 °C. An additional 3 mL of THF was used to ensure a complete
transfer. The dark green reaction mixture was warmed to room
temperature. After 20 min, methyl iodide (230 µL, 3.71 mmol) was
added and the mixture was stirred overnight. During this time a
white precipitate appeared. The reaction mixture was cooled to 0
°C and quenched by careful addition of saturated aqueous
ammonium chloride (10 mL), diluted with water (40 mL), and
MK-4-18O
MK-4
449.3
445.4
455.4
451.3
653.5
649.5
29
31
35
33
39
45
191.2
187.2
191.3
187.1
191.3
187.2
20.8
20.8
41.3
41.3
74.9
74.9
PK-18O
PK
MK-7-18O
MK-7
a Ion sequence according to descending abundance.
Apparatus and HPLC Conditions. The HPLC analyses were
conducted with a Shimadzu HPLC system (Simadzu, Kyoto, Japan)
consisting of a binary pump (LC-10AD liquid chromatography),
automatic solvent degasser (DGU-14A degasser), and autosampler
(SIL-10AD autoinjector). Separations were carried out using a
reversed-phase C18 analytical column (Capcell PAK C18 UG120, 5
µm; 4.6 mm i.d. × 250 mm) (Shiseido, Tokyo, Japan) with a
solvent system consisting of an isocratic solvent A (25 min) and
then a linear gradient from 0 to 50% ethanol (50 min). Solvent A
contained methanol/0.1% acetic acid aqueous (95:5, v/v) and was
delivered at 1.0 mL/min. This mobile phase was passed through
the column at 1.0 mL/min. The column was maintained at 40 °C
with a column oven (CTO-10AC column oven). The HPLC system
was controlled by a SCL-10A System Controller (Shimadzu). Acetic
acid both functioned as an ion pair reagent during reversed-phase
HPLC and facilitated formation of protonated vitamin K, [M +
H]+, in the positive ion mode with an APCI. The autosampler was
maintained at 25 °C.
Apparatus and Mass Spectrometry. Mass spectrometry was
performed with an API3000 LC-MS/MS System (Applied Bio-
systems, Foster City, CA), equipped with an APCI electrospray
interface. All MS data were collected in the positive ion mode,
The following settings were used: corona discharge needle
voltage, 5.5 kV; vaporizer temperature, 400 °C; sheath gas (high-
purity nitrogen) pressure, 50 psi; auxiliary gas, none; and transfer
capillary temperature, 220 °C. The electron multiplier voltage was
set at 850 eV. Identification and quantitation were based on MS/
MS-MRM. The range for the parent scan was 400-500 atomic
mass units (amu) in the case of PK, MK-4, and their 18O-labeled
compounds and 600-700 amu for MK-7 and MK-7-18O. A complete
overview of the MRM transitions, collision energy, retention time,
and corresponding segment used for each analyte is given in Table
1.
Quantitation. A quantitative analysis was carried out using
MS/MS-MRM of the precursor ion of vitamin K homologues (m/z
445 (MK-4), 449 (MK-4-18O), 451 (PK), 455 (PK-18O), 649 (MK-
7), 651 (MK-7-18O)), and their product ion (m/z 187 (natural
vitamin K analogues), m/z 191(18O-labeled vitamin K analogues))
with a dwell time of 500 ms. Calibration, using internal standard-
ization, was done by linear regression analysis using five different
concentrations, 12.5, 25, 50, 100, and 200 ng/mL. The points were
given by the calculated peak area ratio of standard and internal
standard.
Precision and Accuracy. Interassay and intraassay precision
and accuracy were evaluated using human control serum samples.
Sample Collection. Blood samples (10 mL) were collected
into heparinized tubes from a convenient forearm vein. The
(15) Snyder, C. D.; Rapoport, H. J. Am. Chem. Soc. 1974, 96, 8046-8054.
(16) Flowers, II, R. A.; Naganathan, S.; Dowd, P.; Arnett, E. M.; Ham, S. W. J.
Am. Chem. Soc. 1993, 115, 9409-9416.
(17) Jpn. Kokai Tokkyo Koho 1982; 57140743.
Analytical Chemistry, Vol. 77, No. 3, February 1, 2005 759

extracted with ether (4 × 25 mL). The ether layers were
combined, dried, filtered, and concentrated to yield 412 mg of 7a
as pale yellow oil. Purification by flash chromatography using 5%
ethyl acetate in hexane yielded 346 mg (89% yield) of 7a as a
pale yellow oil: 1H NMR (500 MHz, CDCl3) δ 0.81 (s, 3H), 0.82
(s, 3H), 0.85 (s, 3H), 0.87 (s, 3H), 1.00-1.38 (m, 16H), 1.81 (s,
3H), 1.94-1.98 (m, 2H), 2.38 (s, 3H), 3.57 (dd, J ) 1.0, 6.5 Hz,
2H), 3.86 (s, 3H), 3.88 (s, 3H), 5.08-5.11 (m, 1H), 7.44-7.47 (m,
2H), 8.04-8.06 (m, 2H); 13C NMR (125 MHz, CDCl3) δ 12.4, 16.3,
19.7, 22.6, 22.7, 24.5, 24.8, 25.4, 26.3, 28.0, 32.7, 32.8, 36.7, 37.3,
37.4, 39.4, 40.0, 61.3, 62.2, 122.1, 122.3, 122.6, 125.2, 125.4, 126.9,
127.3, 127.5, 131.0, 136.1, 149.7, 150.1; HREIMS calcd for C33H5216O2
(M+), 480.3967, found 480.3967.
2-Methyl-3-geranylgeranyl-1,4-di-methoxynaphthalene
(7b). The compound 6b (70 mg, 143 µmol) was converted to
dimethoxynaphthalene (7b) (51 mg) in 75% yield as a colorless
oil according to the procedure described above for obtaining 7a
from 6a: 1H NMR (500 MHz, CDCl3) δ 1.56 (s, 3H), 1.57 (s, 3H),
1.59 (s, 3H), 1.67 (s, 3H), 1.82 (s, 3H), 1.92-2.10 (m, 12H), 2.37
(s, 3H), 3.57 (dd, J ) 0.5, 6.0 Hz, 2H), 3.86 (s, 3H), 3.88 (s, 3H),
5.06-5.13 (m, 4H), 7.43-7.46 (m, 2H), 8.03-8.06 (m, 2H); 13C
NMR (125 MHz, CDCl3) δ 12.4, 15.9, 16.0, 16.4, 17.6, 25.7, 26.3,
26.5, 26.6, 26.7, 39.7, 61.3, 62.1, 122.1, 122.2, 122.8, 124.0, 124.2,
124.4, 125.2, 125.4, 126.9, 127.2, 127.5, 130.9, 131.2, 134.9, 135.1,
135.7, 149.7, 150.1; HREIMS calcd for C33H4816O2 (M+), 474.3498,
found 474.3479.
146.2, 184.5, 185.4; HREIMS calcd for C31H4618O2 (M+), 454.3583,
found 454.3588.
Menaquinone-4-18O (2). The compound 7b (70 mg, 147
µmol) was converted to menaquinone-4 (2) (40 mg) in 60% yield
as a pale yellow oil according to the procedure described above
for obtaining 1 from 7a:1H NMR (500 MHz, CDCl3) δ 1.55 (s,
3H), 1.56 (s, 3H), 1.59 (s, 3H), 1.67 (s, 3H), 1.79 (s, 3H), 1.91-
2.09 (m, 12H), 2.19 (s, 3H), 3.37 (d, J ) 7.5 Hz, 2H), 5.00-5.10
(m, 4H), 7.67-7.70 (m, 2H), 8.06-8.10 (m, 2H); 13C NMR (125
MHz, CDCl3) δ 12.7, 15.95, 16.01, 16.4, 17.7, 25.7, 26.0, 26.5, 26.6,
26.7, 39.7, 119.1, 123.9, 124.2, 124.4, 126.2, 126.3, 131.2, 132.2,
133.26, 133.32, 134.9, 135.2, 137.6, 143.3, 146.2, 184.5, 185.4;
HREIMS calcd for C31H4018O2 (M+), 448.3113, found 448.3111.
Menaquinone-7-18O (3). The compound 7c (80 mg, 118
µmol) was converted to menaquinone-4 (3) (42 mg) in 55% yield
as a pale yellow oil according to the procedure described above
for obtaining 1 from 7a: 1H NMR (500 MHz, CDCl3) δ 1.56 (s,
6H), 1.59 (s, 12H), 1.67 (s, 3H), 1.79 (s, 3H), 1.91-2.09 (m, 24H),
2.19 (s, 3H), 3.37 (d, J ) 7.0 Hz, 2H), 5.00-5.13 (m, 7H), 7.67-
7.70 (m, 2H), 8.07-8.09 (m, 2H); 13C NMR (125 MHz, CDCl3) δ
12.7, 15.95, 16.01, 16.4, 17.7, 25.7, 26.0, 26.5, 26.6, 26.7, 26.8, 39.7,
119.1, 123.9, 124.2, 124.4, 126.2, 126.3, 131.2, 132.2, 133.27, 133.28,
134.9, 135.2, 137.6, 143.4, 146.2, 184.5, 185.5; HREIMS calcd for
C46H6418O2 (M+), 652.4991, found 652.4984
RESULTS AND DISCUSSION
Method Development. The separation of vitamin K deriva-
2-Methyl-3-geranylgeranylfarnesyl-1,4-dimethoxynaph-
thalene (7c). The compound 6c (80 mg, 101 µmol) was converted
to dimethoxynaphthalene (7c) (58 mg) in 75% yield as a colorless
oil according to the procedure described above for obtaining 7a
from 6a: 1H NMR (500 MHz, CDCl3) δ 1.57 (s, 6H), 1.59 (s, 12H),
1.67 (s, 3H), 1.82 (s, 3H), 1.92-2.07 (m, 24H), 2.38 (s, 3H), 3.57
(d, J ) 7.0 Hz, 2H), 3.87 (s, 3H), 3.88 (s, 3H), 5.06-5.13 (m, 7H),
7.44-7.46 (m, 2H), 8.03-8.06 (m, 2H); 13C NMR (125 MHz,
CDCl3) δ 12.4, 16.0, 16.4, 17.7, 25.7, 26.3, 26.6, 26.7, 26.8, 39.7,
61.3, 62.2, 122.1, 122.2, 122.8, 124.0, 124.2, 124.4, 125.2, 125.4,
126.9, 127.2, 127.5, 130.9, 131.3, 134.9, 135.1, 135.7, 149.7, 150.1;
HREIMS calcd for C48H7016O2 (M+), 678.5376, found 678.5374.
Phylloquinone-18O (1). A solution of 7a (60 mg, 124 µmol)
in degassed acetonitrile (0.5 mL) and ether (0.1 mL) was added
using a cannula to a solution of ceric ammonium nitrate (205 mg,
0.374 mmol) in degassed H218O (0.1 mL, >95% 18O). An extra 0.3
mL of acetonitrile was used to ensure a complete transfer. After
20 min at room temperature, the reaction mixture was treated
with water (10 mL) and ether (10 mL). The aqueous layer was
extracted with ether (25 mL), and the ether layers were combined,
washed with water (4 × 15 mL), dried, filtered, and concentrated
to yield 35 mg (62% yield) of yellow oil. The crude material of 1
was purified by flash chromatography through silica gel using
5% ethyl acetate in hexane as the eluent. Analysis of LC-APCI-
MS/MS showed a 95.5% incorporation of 18O; 13C NMR showed
95.3% labeling: 1H NMR (500 MHz, CDCl3) δ 0.81 (s, 3H), 0.82
(s, 3H), 0.83 (s, 3H), 0.87 (s, 6H), 0.98-1.41 (m, 16H), 1.78 (s,
3H), 1.92-1.96 (m, 2H), 2.20 (s, 3H), 3.37 (d, J ) 7.0 Hz, 2H),
5.01 (dt, J ) 1.0, 7.0 Hz, 1H), 7.67-7.71 (m, 2H), 8.06-8.10 (m,
2H); 13C NMR (125 MHz, CDCl3) δ 12.7, 16.3, 19.7, 22.6, 22.7,
24.4, 24.5, 24.8, 25.3, 26.0, 28.0, 32.6, 32.8, 36.6, 37.3, 37.4, 39.4,
40.0, 118.8, 126.2, 126.3, 132.16, 132.20, 133.26, 133.32, 138.0, 143.3,
tives and the respective internal standards in human plasma was
achieved in 80 min. After a wash of the column and reequilibration
period of 40 min, the next sample was injected. The reliability of
the LC method was evaluated based on the variation in retention
times. The relative standard deviation (RSD), calculated from
retention times obtained from over 30 injections, proved to be
less than 1.0% for all compounds, indicating good chromatographic
stability.
The precursor and product ion(s) for each analyte of interest
was determined by the direct infusion of single-analyte solutions
(1 µg/mL in ethanol). After optimization of the separation process
and selection of a unique precursor-product ion combination for
each compound, a quantitative LC-APCI-MS/MS method was
developed based on MRM.
To ensure maximum sensitivity in the MS analysis, the
chromatographic run was divided into seven segments. Each
segment was optimized for the compounds of interest eluted
within a given time period. The following mass spectrometric
parameters were specified within each segment: transfer capillary
voltage, tube lens voltage, ion optic voltage, collision energy, and
MRM scan events. Table 1 shows an overview of the MS
parameters including MRM transitions, collision energy, and
retention time. The MRM chromatograms for the target analytes
were obtained from the injection of a standard mixture (10 ng).
The retention times of the standard and internal standard peaks
completely matched; namely, this result proved that the com-
pounds had the same chemical properties.
18O-Labeled Internal Standards. As a preliminary experi-
ment, we tried to detect PK and MK-n (n) 1-10) in several
human plasma samples using LC-APCI-MS/MS. PK, MK-4, and
MK-7 were found as major peaks. Therefore, we chose stable
isotope-labeled vitamins as internal standards to measure these
760 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

Figure 1. LC-APCI-MS/MS chromatograms of the analytes of vitamin K Homologues for the injection of (i) standard mixture and (ii) human
plasma sample with internal standards. PK, phylloquinone; MK-4, menaquinone-4; MK-7, menaquinone-7; PK-18O, 18O-labeled phylloquinone;
MK-4-18O, 18O-labeled menaquinone-4; MK-7-18O, 18O-labeled menaquinone-7.
three kinds of vitamin K. The stable isotopes are nonradioactive
forms of elements that occur naturally within the environment
and have applications for human research. There has been much
recent study using stable isotope-labeled vitamins and provitamins,
such as [2H7]MK-4 (deuterated MK-4) and [2H7]PK (deuterated
PK), to examine metabolic pathways in vitro and in vivo using
mass spectrometry.14,18 We used isotope 18O-labeled vitamin K
analogues as internal standards, namely, MK-4-18O, PK-18O, and
MK-7-18O.19,20 The efficiency of extraction can be adjusted ac-
curately since the chemical properties of the labeled analogues
are almost the same as those of the original substrates. The
synthesis of the requisite 18O-labeled vitamin K homologues was
carried out as shown in Scheme 1. These analogues are much
easier and more convenient to synthesize than, for example,
deuterated vitamin K. First, we chose monoacetate 4 to afford a
coupling intermediate with an alkyl side chain, though there are
many different schemes for the synthesis of vitamin K deriva-
tives.15,21 The compound 5 was prepared from 1,4-hydroquinone
diacetate 4 by selective hydrolysis of 4-O-acetate with sodium
hydrosulfite.16,17 Treatment of the monoacetate 5 with an alkyl
side chain alcohol (phytol, geranylgeraniol, geranygeranylfarnesol)
in the presence of boron trifluoride etherate yielded 6a-6c as
reported.16 To afford the dimethyl ether 7a-7c, the monocetate
5 was treated with an excess amount of potassium hydride
(18) Sano, Y.; Kikuchi, K.; Tadano, K.; Hoshi, K.; Koshihara, Y. Anal. Sci. 1997,
13, 67-73.
(19) Naganathan, S.; Hershline, R.; Ham, S. W.; Dowd, P. J. Am. Chem. Soc. 1994,
116, 9831-9839.
(20) Dowd, P.; Ham, S.-W.; Hershline, R. J. Am. Chem. Soc. 1992, 114, 7613-
7617.
(21) Naruta, Y. J. Org. Chem. 1980, 45, 4097-4104.
Analytical Chemistry, Vol. 77, No. 3, February 1, 2005 761

Table 2. Summary of Assay Method for Vitamin Ka
PK
MK-4
50
MK-7
80
quantitation limit (pg/mL)
40
recovery
mean ( SD (ng/mL)
RSD (%)
1.89 ( 0.05
2.65
0.43 ( 0.02
4.65
3.19 ( 0.20
6.27
recovery (%)
98 ( 3
102 ( 5
0.27 ( 0.01
4.79
0.21 ( 0.02
9.10
102 ( 6
1.44 ( 0.08
5.85
1.51 ( 0.04
2.97
intraassay control serum
interassay control serum
mean ( SD (ng/mL)
RSD (%)
mean ( SD (ng/mL)
RSD (%)
0.95 ( 0.04
6.21
1.03 ( 0.06
6.21
a Recovery, intraassay, and interassay: n ) 5.
followed by methyl iodide to give 7a-7c in good yield. Finally,
oxidation of the naphthohydroquinone methyl ether 5 with Ce-
(NH4)2(NO3)6 in H218O gave the desired 18O-labeled vitamin K
homologues 1-3 in good yield. Thus, three kinds of stable
isotope-labeled compounds were prepared for LC-APCI-MS/MS
detection.
Calibration and Validation. Calibration using internal stan-
dardization with standard samples and 18O-labeled analogues was
performed. Under the stated conditions, stock solutions proved
to be stable for at least 3 months. Analyte recoveries in the stability
experiments were within the variability range obtained for preci-
sion and accuracy. No significant loss or deterioration of any of
the compounds of interest was observed. Analytes were stable
during sample pretreatment at room temperature.
The method fulfilled our analytical standard criteria. MRM
provided high specificity for all of the compounds, and no cross-
talk interference with the 18O-labeled internal standards was
observed. The positive precursor ion APCI-MS of PK, MK-4, MK-
7, and their 18O-labeled forms showed base peaks at m/z 451, 445,
649, 455, 449, and 653 corresponding to protonated molecules,
respectively. While the product ion of PK, MK-4, and MK-7 was
m/z 187, that of their labeled analogues was m/z 191.22 Each of
the calibration curves could be drawn as linear through zero. All
values were calculated as ratios (intensity of analyte area)/(that
of internal standard area). Over the range of vitamin K concentra-
tions, 12.5-200 ng/mL, positive ion APCI produced a linear
response and the each correlation coefficient of the calibration
curves in PK, MK-4, and MK-7 were 0.9993, 0.9997, and 0.9999.
LC-APCI-MS/MS was used throughout this investigation because
of its wide dynamic range and linearity of detector response. Thus,
vitamin K could be directly detected without conversion to other
derivatives.
The data were converted to the concentration in 1 mL of plasma
according to the following equation: vitamin K in plasma
concentration (ng/mL) ) measured data × (6/5) × (60/500).
As shown in Table 2, the average intraassay and interassay
variation (RSD) for PK, MK-4, and MK-7 was less than 10%.
Average spiked recoveries from authentic compounds (PK, 0.95
( 0.04 ng/mL; MK-4, 0.27 ( 0.01 ng/mL; MK-7, 1.44 ( 0.08 ng/
mL) added to normal human pool serum samples for PK, MK-4,
and MK-7 were 98-102%. With this method, the lower quantitation
limits were less than 0.1 ng/mL (PK, 40 pg/mL; MK-4, 50 pg/
mL; MK-7, 80 pg/mL). The quantitation limits of plasma vitamin
K concentrations are different from species. Indeed, the signal-
to-noise ratio depends on impurities in plasma samples. As far as
we know, the rat plasma or serum vitamin K concentrations were
easy to determine in comparison with human plasma because
impurities were not observed (data are not shown). These results
proved our system was reliable and reproducible for the measure-
ment of plasma vitamin K.
Plasma Concentration Profile of Vitamin K Analogues.
Next we examined healthy subjects (n ) 20) using this method.
The accuracy of determination was improved according to the
internal standards. The mean plasma concentrations of PK, MK-
4, and MK-7 from the subjects were 1.22 ( 0.57, 0.39 ( 0.46, and
6.37 ( 7.45 ng/mL, respectively. While plasma levels of vitamin
K analogues in osteoporotic patients on a MK-4 supplement were
1.90 ( 0.78, 172.90 ( 138.56, and 7.03 ( 5.79 ng/mL.
To compare the conventional fluorescence method and LC-
APCI-MS/MS detection, we examined another 27 plasma samples
using both methods. The concentrations of PK and MK-7 cor-
related although the fluorescence method gave slightly higher
concentrations than the LC-APCI-MS/MS method as shown in
Figure 2C and D. However, the plasma levels of MK-4 were not
related at all. We also examined 17 plasma samples from
osteoporotic patients given an MK-4 supplement. The concentra-
tions of all vitamin K derivatives were correlated to some extent
(Figure 2B-D). This result suggested that the detection of MK-4
using these methods was not related to the low concentration
range (less than 1.0 ng/mL) (Figure 2A) but correlated with the
higher concentration range (Figure 2B). Presumably, impurities
in the plasma sample are at least partly responsible for the result.
Our LC-APCI-MS/MS system has much greater sensitivity and
selectivity in comparison with the conventional method. As we
described above, the conventional method has various disadvan-
tages such as interference from impurities in human plasma and
correction of extractive efficiency. Most conventional methods do
Figure 1 shows the MRM chromatograms for the target
analytes obtained from the injection of human plasma samples.
We examined recovery and accuracy using commercially available
pooled plasma. The injection samples were obtained according
to the procedure described above. All measurements were well
performed, and each peak of the vitamin K analogues was clearly
afforded as a single peak.
After the analytes were extracted from 0.5 mL of plasma sample
including 3.6 ng/mL internal standard, they were concentrated
to 60 µL in an ethanol solution. The concentrations of both internal
standards in the samples and calibration curves were 50 ng/mL.
(22) Di Mari, S. J.; Supple, J. H.; Rapoport, H. J. Am. Chem. Soc. 1966, 88, 1226-
1232.
762 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

Figure 2. Comparison of assayed values (ng/mL) of human plasma vitamin K homologues between the LC-APCI-MS/MS method (MS method)
and HPLC with fluorescence detection (fluorescence method). (A) MK-4 (normal subjects); (B) MK-4 (osteoporotic patients); (C) PK (normal
subjects and osteoporotic patients); (D) MK-7 (normal subjects and osteoporotic patients).
not use any internal standards or often use another kind of vitamin
K homologue. Extractive efficiency is important for accurate
determination of plasma concentration. In our method, extractive
efficiency from plasma can be completely adjusted using 18O-
labeled vitamin K analogues as internal standards, which have
the same chemical properties as the original vitamin K homo-
logues. Therefore, the accuracy of determination was remarkably
improved. Though the conventional method is certainly a common
system, our method is useful to make gold standards of plasma
vitamin K concentrations.
We conclude that this novel LC-APCI-MS/MS method using
18O-labeled internal standards should be convenient for the
evaluation of vitamin K status in human plasma; therefore, the
data will provide useful information for the treatment and preven-
tion of osteoporosis with vitamin K.
ACKNOWLEDGMENT
We thank Dr. Masataka Shiraki, the Research Institute and
Practice for Involutional Disease. We are also grateful to Eisai
Co., Ltd. for the gift of vitamin K standards. This work was
supported in part by a Grant-in-Aid for Comprehensive Research
on Cardiovascular Diseases and the Research on the Dietary
Reference Intakes in Japanese from the Ministry of Health, Labor
and Welfare of Japan.
CONCLUSION
This study shows that LC-APCI-MS/MS provides a rapid and
relatively easy-to-use approach to the quantitation of vitamin K
analogues in human plasma without compromising assay sensitiv-
ity. This approach overcame major disadvantages of previous
methods. For the analysis of both the analyte and internal
standard, the method has been thoroughly validated. The sensitiv-
ity has been shown to be excellent, with no interference from
impurities. The interday precision for the analyte was less than
10% RSD. Furthermore, throughout our study, the HPLC column
used remained stable.
SUPPORTING INFORMATION AVAILABLE
Calibration curves for vitamin D homologues. This material is
available free of charge via the Internet at http://pubs.acs.org.
Received for review July 14, 2004. Accepted October 8,
2004.
AC0489667
Analytical Chemistry, Vol. 77, No. 3, February 1, 2005 763