Conversion of Phylloquinone (Vitamin K₁) into Menaquinone-4 (Vitamin K₂) in Mice

Materials—PK and MK-4 were purchased from Wako Pure Chemical Industries, Ltd., Osaka, Japan. D-labeled menadione (K₃-d₈) was purchased from C/D/N Isotopes, Inc. (Quebec, Canada). PK epoxide, MK-4 epoxide, ¹⁸O-labeled PK (PK-¹⁸O), and ¹⁸O-labeled MK-4 (MK-4-¹⁸O) were respectively synthesized in our laboratory as reported previously (, ). D-labeled PK (PK-d₇ and PK-d₉), D-labeled PK epoxide (PK epoxide-d₇), D-labeled MK-4 (MK-4-d₇ and MK-4-d₉), and D-labeled MK-4 epoxide (MK-4 epoxide-d₇) were also synthesized in our laboratory. The synthesis and unambiguous physicochemical assignments of these D-labeled vitamin K compounds will be described elsewhere. Culture media and antibiotics were purchased from Nakalai Tesque, Kyoto Japan. D-labeled chloroform (CDCl₃, 99.8%, NMR analytical grade) was obtained from EURISO-TOP (Gif-Sur-Yvette, France). Organic solvents of HPLC grade were purchased from Wako Pure Chemical Industries, Ltd.

Animals and Diets—Male and Female C57BL/6 mice at the age of 7 weeks were obtained from Japan SLC, Inc., Hamamatsu, Japan and were fed a commercially available normal laboratory chow diet (F-2 diet, Oriental Yeast Co., Ltd., Tokyo, Japan) for 1 week. They were housed 5 per cage with a 12-h light-dark cycle under controlled environmental conditions (temperature: 20 ± 2 °C, humidity: 50 ± 5%). All animals were allowed free access to diet and deionized water for a period of 1 week. The protocols for the experiments were approved by the Guidelines for the Care and Use of Laboratory Animals of Kobe Pharmaceutical University.

Measurements of PK, MK-4, and Their Respective Epoxides in Tissues and Plasma of Mice—At 8 weeks of age, blood was withdrawn by heart puncture under light diethyl ether anesthesia, and plasma was obtained by centrifugation at 3000 rpm for 10 min. Immediately after death, tissues were excised, immersed and rinsed in ice-cold saline, and stored at –80 °C until the assays. Tissue concentrations of PK, MK-4, and their respective epoxides were assayed as previously described (, ). Because vitamin K is sensitive to light, all procedures were performed under a dark light. Briefly, tissues (wet weight, 1–2 g) were pulverized thoroughly with anhydrous sodium sulfate (1:10, w/v) and transferred into a brown-colored glass tube with a Teflon-lined screw cap. The homogenates were added to 0.1 ml of ethanol containing PK-¹⁸O and MK-4-¹⁸O as internal standards, 0.9 ml of ethanol, and 9 ml of acetone, then mixed thoroughly with a Voltex mixer for 3 min, and allowed to stand for 5 min. This procedure was repeated three times. The resulting mixture was centrifuged at 3000 rpm for 5 min at 4 °C, and the upper layer was transferred into a small brown-glass tube and evaporated dry under reduced pressure. The residue was dissolved in 2 ml of water and 6 ml of hexane, mixed thoroughly, and centrifuged at 3000 rpm for 5 min at 4 °C. The upper layer was loaded onto a Sep Pak Vac Silica cartridge column (Waters). The column was eluted with 5 ml of hexane to remove concomitants, and thereafter the vitamin K-containing fraction was eluted with 5 ml of hexane and diethyl ether (97:3) solution. The eluate was evaporated under reduced pressure, and the residue was dissolved in 60 μl of methanol. An aliquot of this solution was subjected to APCI3000 LC-MS/MS (Applied Biosystems, Foster City, CA). The HPLC analyses were conducted with a Shimadzu HPLC system (Shimadzu, 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 column (Capcell PAK C₁₈ UG120, 5 μm; 4.6 mm inner diameter × 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 35 °C with a column oven (CTO-10AC column oven). All MS data were collected in the positive ion mode with atmospheric pressure chemical ionization (APCI). The following settings were used: corona discharge needle voltage, 5.5 kV; vaporizer temperature, 400 °C; sheath gas (high purity nitrogen) pressure, 50 p.s.i.; and transfer capillary temperature, 220 °C. The electron multiplier voltage was set at 850 eV. Identification and quantification were based on MS/MS using a multiple reaction monitoring (MRM) mode. The range for the parent scan was 400–500 atomic mass units. MRM transitions (precursor ion and product ion, m/z) and retention time (min) for each analyte were as follows: MK-4: precursor ion, 445.4; product ion, 187.2; retention time, 20.8; MK-4-epoxide: precursor ion, 461.3; product ion, 161.0; retention time, 14.6; MK-4-¹⁸O: precursor ion, 449.3; product ion, 191.2; retention time, 20.8; PK: precursor ion, 451.3; product ion, 187.1; retention time, 41.3; PK epoxide: precursor ion, 467.4; product ion, 161.2; retention time, 27.2; and PK-¹⁸O: precursor ion, 455.4; product ion, 191.3; retention time, 41.3. Calibration, using internal standardization, was done by linear regression with five different concentrations, 12.5, 50, 200, 800, and 1600 ng/ml.

Isolation and Identification of MK-4-d7 in Cerebra of Mice Orally Given PK-d7—At 8 weeks of age, 200 female mice were orally given PK-d₇ as a single dose of 10 μmol/kg body weight, and their cerebra were collected at 24 h post-administration and stored at –80 °C for analysis. The cerebra (total wet weight, 60 g) were divided into 40 portions (each ∼1.5 g) for the purification of MK-4-d₇. MK-4-d₇ in each cerebral sample was extracted and purified with the same procedures as described in the measurement of MK-4-d₇ in tissues. Subsequently, the sample solution dissolved in 100 μl of methanol was subjected to purification using a Shimadzu HPLC system consisting of a C-R4A Chromatopac, SPD-6A UV spectrophotometric detector, and LC-6A liquid chromatograph. Separations were carried out using a reversed-phase C₁₈ column (COSMOSIL 5C₁₈ ARII, 10 mm inner diameter × 250 mm, Nakalai Tesque) with a mobile phase containing methanol/ethanol (95/5). This mobile phase was passed through the column at 4.0 ml/min. Samples with a peak corresponding to authentic MK-4 were collected and evaporated dry, then re-dissolved in the mobile phase. After further purification of the combined samples with the same HPLC system, the sample was pure enough for ¹H NMR, D NMR, and LC-MS/MS analyses. The 500-MHz H NMR and D NMR spectra of the putative MK-4 and MK-4-d₇ were measured on a Varian VNS-500 (H: 500 MHz, D: 77 MHz). The sample was first dissolved in 40 μl of CDCl₃ in a nanoprobe for ¹H NMR spectrometry, and then after evaporation with a stream of nitrogen gas, the residue was dissolved in CHCl₃ for D NMR spectrometry. The LC-MS/MS analysis was carried out with the same method as described above.

Comparison of MK-4-d7 Accumulation in Cerebra of Mice Administered PK-d7 or K3-d8 via Four Routes—At 8 weeks of age, 5 female mice in each group were given orally, intravenously, or enterally, either PK-d₇ or K₃-d₈ as a single dose of 10 μmol/kg body weight or intracerebroventricularly either PK-d₇ or K₃-d₇ as a single dose of 0.1 μmol/kg body weight. To examine whether the conversion of PK-d₇ or K₃-d₈ into MK-4-d₇ occurs in a dose-dependent manner, 5 female mice in each group were given orally either PK-d₇ or K₃-d₈ as a single dose of 0.1, 1.0, or 10 μmol/kg body weight. After 24 h, the mice were killed, and cerebra were removed and stored at –80 °C for analysis. The measurements of MK-4-d₇ and MK-4 epoxide-d₇ in cerebra were carried out with the same LC-APCI-MS/MS method as described above.

Comparison of the Conversion of PK-d7 or K3-d8 to MK-4-d7 in Mouse Cerebral Slice Culture and Embryonic Primary Culture—At 8 weeks of age, female mice were sacrificed by aortic exsanguination under light diethyl ether anesthesia. Cerebra were excised and cut into 1-mm-thick slices. Two slices were placed on a stainless steel mesh in a culture dish (60-mm inner diameter) and cultured in 6 ml of medium containing minimum essential medium (MEM): Hepes buffered-saline solution (2:1) with 25% horse serum (Invitrogen) in the presence or absence (vehicle; ethanol) of PK-d₇ or K₃-d₈ (10^–5 m) at 37°C in 5% CO₂ in a humidified atmosphere for 24 h. The slices were washed with cold Ca,Mg-free phosphate-buffered saline 2 times and homogenized using a Dounce-type homogenizer with 1 ml of MilliQ water. The homogenates (20 μl) were used to determine protein concentrations with a BCA protein assay kit (Pierce). Using other homogenates, the measurements of MK-4-d₇ and MK-4 epoxide-d₇ were carried out with the same LC-APCI-MS/MS method as described above. Primary cultures of mouse cerebral hemispheres on embryonic day (E) 14 were prepared as described previously with minor modifications (). Fetuses (E14) were removed from pregnant C57BL/6J mice. Embryonic cerebral hemispheres were collected and incubated in the L-15 dissociation medium (Sigma) containing trypsin (0.25 mg/ml) and DNase I (0.1 mg/ml). After treatment, cells were dispersed by repeated trituration and suspended in low glucose Dulbecco's modified Eagle's medium (L-DMEM) with 10% FCS. The suspended cells were plated at a density of 8 × 10⁶ cells/well on polyethyleneimine (Sigma)-coated 6-well tissue culture plates. The cells were cultured in L-DMEM with 10% FCS at 37 °C in 5% CO₂ in a humidified atmosphere for 2 days. Then, the cells were incubated with or without PK-d₇ or K₃-d₈ (10^–6 m) in the L-DMEM at 37 °C in 5% CO₂ in a humidified atmosphere for 24 h. The amounts of MK-4-d₇ and MK-4 epoxide-d₇ generated from PK-d₇ or K₃-d₈ by the cells were measured by the LC-APCI-MS/MS method described above. Embryonic cerebral cells were cultured in L-DMEM with 10% FCS for 2 days and thereafter divided into two sub-culture groups, one for neurons and another for astrocytes (, ). For the neurons, the primary cerebral cells were cultured in L-DMEM containing 10% FCS along with 50 μm 1-β-d-arabinofuranosylcytosine for 2 days. For the astrocytes, the primary cells were dissociated by incubation in 0.1% trypsin-EDTA solution and cultured in L-DMEM medium containing 10% FCS for 2 days. The cells were suspended in L-DMEM containing 10% FCS, plated at a density of 10⁷ cells/plate (60 mm), and cultured for 2 days. The cells that reached confluency were dissociated by incubation in 0.1% trypsin-EDTA solution. The cells were suspended in L-DMEM containing 10% FCS and plated onto 6-well tissue culture plates at a density of 10⁶ cells/well for 2 days to reach confluency. Both cells were treated with the culture medium containing PK-d₇ or K₃-d (10^–6 m) for 24 h. After incubation, the cells were collected and washed with cold Ca,Mg-free phosphate-buffered saline 3 times and then refrigerated at –30 °C. After being warmed to room temperature, cells were lysed in 1 ml of Ca,Mg-free phosphate-buffered saline. With this procedure, 20 μl of cell lysate was analyzed for protein determination. To the cell lysate in a brown screw-capped tube were added PK-¹⁸O and MK-4-¹⁸O as internal standards. The measurements of MK-4-d₇ and MK-4 epoxide-d₇ in cerebra were carried out with the same LC-APCI-MS/MS method as described above.

Concentrations of PK, MK-4, and Their Epoxides in Plasma and Tissues of Mice—As shown in Table 1, four species of K-vitamins were found in all tissues examined, although their levels varied greatly while only PK and MK-4 were detected, and at low levels, in plasma. Basically, tissue-distribution patterns of K-vitamins did not differ between male and female mice, although the levels of MK-4 and MK-4 epoxide were relatively higher in the females. Concentrations of MK-4 and its epoxide were much higher than those of PK and its epoxide in all tissues examined. Relatively high PK levels were found in thyroid gland, aorta, heart, and adrenal gland, but most other tissues contained <50 pmol/g tissue. In contrast, relatively high MK-4 levels were found in almost all the tissues except for heart, lung, liver, muscle, and bowel content in male mice, and liver and bowel content in female mice, where concentrations were <50 pmol/g tissue. There was no consistent correlation between the tissue levels of PK and MK-4 or PK epoxide and MK-4 epoxide in male or female mice. The laboratory chow given to the mice was assayed for PK and MK-4 by LC-APCI-MS/MS. PK and MK-4 concentrations were 212.2 ± 0.9 and 2.2 ± 0.5 pmol/g diet, respectively. Thus, it is not conceivable that the high concentrations of MK-4 and its epoxide in tissues of mice originated from the intake of laboratory chow.

Isolation and Identification of MK-4-d7 in Cerebra of Mice Orally Given PK-d7—We next examined whether orally given PK-d₇ accumulates in cerebra as a converted form of MK-4-d₇ in mice. 200 female mice were orally given PK-d₇ at a single dose of 10 μmol/kg body weight, and cerebra were collected at 24 h post-administration. We subjected an aliquot of the lipid extract from the cerebra to LC-APCI-MS/MS and confirmed that the peaks correspond to MK-4 epoxide, MK-4, PK epoxide, and PK on the MRM chromatogram (Fig. 1). If PK-d₇ was converted into MK-4-d₇ in the body, then the MK-4 peak shown in Fig. 1B should include both MK-4 and MK-4-d₇, because they gave the same retention time on the MRM chromatogram, and therefore, we decided to isolate and purify the peak. After purification by HPLC as noted under “Experimental Procedures,” we finally obtained the MK-4 fraction containing MK-4 and MK-4-d₇ in amounts of 3.8 and 2.0 μg, respectively. Because tissue MK-4 has not yet been identified on the basis of structural assignments in animals and humans, and no information was available about whether the purified MK-4-d₇ in the presence of endogenous MK-4 can be identified by D NMR, we first analyzed the MK-4 fraction by ¹H NMR spectroscopy. The ¹H NMR spectra of authentic MK-4 and the MK-4 fraction are shown in Fig. 2, A and B, respectively. Consequently, the values of resonance derived from the 2-methyl-1,4-naphthoquinone ring and the geranylgeranyl side chain of the MK-4 fraction entirely coincided with those of authentic MK-4. The D NMR spectra of authentic MK-4-d₇ and the MK-4 fraction are shown in Fig. 2, C and D, respectively. The values of resonance derived from the D-labeled 2-methyl-1,4-naphthoquinone ring of the MK-4 fraction exactly coincided with those of authentic MK-4-d₇. In LC-APCI-MS/MS analyses, authentic MK-4 gave a parent peak (Q1) at m/z 445.6 [M+H]⁺ and a product ion peak (Q3) at m/z 187.3 [2,3-dimethyl-1,4-naphthoquinone + H⁺]⁺ (Fig. 3A). The MK-4 fraction gave the same parent peak at m/z 445.4 and a product ion peak at m/z 187.1 (Fig. 3B). Authentic MK-4-d₇ gave a parent peak (Q1) at m/z 452.3 and a product ion peak (Q3) at m/z 194.3 [2,2,2,5,6,7,8-deuterated 2,3-dimethyl-1,4-naphthoquinone + H⁺]⁺ (Fig. 3C). The MK-4 fraction gave the same parent peak at m/z 452.0 and a product ion peak at m/z 194.4 (Fig. 3D). The LC-APCI-MS/MS MRM chromatograms and MS spectra of the MK-4 fraction were completely congruent with those of authentic MK-4 and MK-4-d₇. Taken together, based on the results of the ¹H NMR, D NMR, and LC-APCI-MS/MS analyses, it is evident that MK-4 exists in cerebra of mice and originates from the intake of PK.

Comparison of MK-4-d7 Accumulation in Cerebra of Mice Administered PK-d7 or K3-d8 via Four Routes—To obtain more insight into the metabolic sites where PK is converted into MK-4, mice were given either orally, enterally, or intravenously, PK-d₇ or K₃-d₈ as a single dose of 10 μmol/kg body weight or intracerebroventricularly at 0.1 μmol/kg body weight. After 24 h, the concentrations of MK-4-d₇ and its epoxide in cerebra were measured by LC-APCI-MS/MS. With respect to the oral route, both PK-d₇ and K₃-d₈ induced the accumulation of MK-4-d₇ or its epoxide, and the conversion of PK-d₇ to MK-4-d₇ was as efficient as that of K₃-d₈ to MK-4-d₇ (Fig. 4A). With respect to the enteral route, both PK-d₇ and K₃-d₈ induced the accumulation of MK-4-d₇ and its epoxide and, as expected, the efficiency with which PK-d₇ was converted to MK-4-d₇ was similar to that for the oral route (Fig. 4B). The result indicates that the release of K₃ from PK does not necessarily require the aid of stomach juice. With respect to the intravenous route, the administration of K₃-d₈ induced the accumulation of MK-4-d₇ and its epoxide at low levels, but the administration of PK-d₇ did not (Fig. 4C). With respect to the intracerebroventricular route, K₃-d₈ induced the accumulation of MK-4-d₇ and its epoxide at low levels, but PK-d₇ did not (Fig. 4D). To examine whether the conversion of PK-d₇ or K₃-d₈ to MK-4-d₇ occurs in a physiological state, we carried out a dose-response study by the oral route and confirmed that both PK-d₇ and K₃-d₈ were converted linearly to MK-4-d₇ and MK-4-d₇ epoxide (Fig. 4E).

Conversion of PK-d9 to MK-4-d7 in Mice—Besides the integral conversion of the phytyl side chain to the geranylgeranyl side chain, desaturation of the phytyl side chain to form the geranylgeranyl side chain may be possible. To examine this possibility, mice were given orally PK-d₉ as a single dose of 10 μmol/kg body weight. After 24 h, the concentrations of PK-d₉, MK-4-d₉, MK-4-d₇, and MK-4-d₇ epoxide in cerebra were measured by LC-APCI-MS/MS. As shown in Fig. 5, MK-4-d₇ in a major form, and its epoxide along with a substrate, PK-d₉, accumulated in the cerebra; however, MK-4-d₉ was not detected, suggesting that the desaturation of the phytyl side chain of PK cannot be ascribed to the conversion of PK to MK-4.

Conversion of PK-d7 or K3-d8 to MK-4-d7 in Mouse Cerebral Slice Culture and Primary Culture—To clarify whether PK and K₃ are converted directly to MK-4 without the aid of intestinal absorption and tissue distribution processes (), mice were sacrificed and cerebra were removed to make slice cultures. Sliced cerebra were incubated with either PK-d₇ or K₃-d₈ at 10^–5 m for 24 h, and the measurements of MK-4-d₇ and its epoxide were performed by LC-APCI-MS/MS. Surprisingly, both PK-d₇ and K₃-d₈ caused remarkable MK-4-d₇ accumulation in cerebra (Fig. 6A), suggesting that the release of K₃-d₇ from PK-d₇ and the geranylgeranylation of K₃ took place within cerebra. Taken together, it is obvious that the cerebrum itself has enzyme(s) or biological machinery for cleaving the phytyl side chain of PK and introducing the geranylgeranyl group to position 3 of the K₃ molecule. From this result, we next examined whether the conversion takes place in a cerebral primary culture. As shown in Fig. 6B, a small but appreciable amount of MK-4-d₇ was detected and the K₃-d₈ to MK-4-d₇ conversion was much more prominent, although both conversion efficiencies in this primary culture were apparently lower than those observed in the slice culture. Another interesting feature was that both PK-d₇ and K₃-d₈ generated substantial amounts of MK-4 epoxide-d₇, which exceeded the levels of MK-4-d₇, probably due to the enhanced turnover rate of the vitamin K cycle in the primary culture. To examine which types of neuronal cells are responsible for this conversion, we prepared primary cultures with a high density of neurons or astrocytes and carried out the same conversion experiment as noted above. As shown in Fig. 6, C and D, no conversion of PK-d₇ to MK-4-d₇ was found, whereas the conversion of K₃-d₈ to MK-4-d₇ occurred as efficiently in the primary culture as in these neuronal cell cultures. In the neuron or astrocyte culture, the conversion of K₃-d₈ to MK-4-d₇ occurred with high efficiency. These results suggest that both neurons and astrocytes lack the ability to convert PK to MK-4, but they have the ability to prenylate K₃ into MK-4. The efficiency with which K₃ was converted to MK-4 was remarkably higher in astrocytes than in neurons.

The present study provides unequivocal evidence that MK-4 exists in cerebra of mice and originates from PK absorbed from the intestine. To the best of our knowledge, this is the first confirmation based on structural assignments from NMR and LC-MS/MS analyses of the conversion of MK-4-d₇ from PK-d₇ or K₃-d₈ in an animal model. The conversion of PK-d₇ into MK-4-d₇ occurred in a dosedependent fashion from a physiological dose of 0.1 μmol/kg body weight to a pharmacological dose of 10 μmol/kg body weight (Fig. 4E). This unique conversion was further confirmed not only in organ slice cultures but in primary cell cultures. Irrespective of differences in the route of administration or differences of organ and cell culture conditions, K₃-d₈ was consistently transformed into MK-4-d₇ with high efficiency. From this finding, it seemed reasonable that K₃ is an intermediate during the conversion of PK to MK-4. The most exciting finding in our study was that, like K₃-d₈, PK-d₇ was also converted into MK-4-d₇ in cerebral slice cultures as well as cerebral primary cultures. To our knowledge, this is also the first evidence of the tissue- or cell-specific conversion of PK into MK-4 based on the usage of D-labeled compounds. The current concept of how PK is converted to MK-4 is that the phytyl side chain of PK is removed, probably during intestinal absorption, the enterocytes would be responsible for this catabolic activity, and the released K₃ is delivered via an unknown transport system to various tissues, where it is geranylgeranylated to form MK-4. However, K₃ itself has not yet been directly identified in enterocytes, the bloodstream, or tissues. Thijssen et al. () recently reported that the urine collected from healthy male volunteers who orally took pharmacological doses of PK contained increased amounts of K₃, in the form of glucuronides and sulfate conjugates, and urinary K₃ excretion mirrored dietary PK intake. Additionally, this excretion of K₃ was not enhanced by the subcutaneous administration of PK. Taken together, our findings clearly indicate that the MK-4 accumulated in the cerebrum, and possibly in other tissues as well, is, at least in part, a metabolic product of the K₃ released from the ingested PK.

The age- and gender-related tissue distribution of K vitamins has been extensively studied in rats and humans (, , ^{,  ,}  ). However, little information is available from mice. It will be necessary to elucidate the routes and functions of cerebral MK-4 accumulation to measure the tissue concentrations of K vitamins. In our study, MK-4 was found in all tissues tested at high concentrations and the levels remarkably exceeded those of PK (Table 1). Previous studies in rats have shown that the liver contains more PK than MK-4 (). The inconsistency between our results and those of others may be a species difference between mice and rats or a dietary influence by how much and how long animals were fed diets containing PK or K₃. From the literature, Thijssen et al. reported that the concentrations of PK, PK epoxide, and MK-4 in the liver of rats fed normal lab chow (PK content of ∼2 μg/g) were 23.7, 6.8, and 2.3 ng/g, respectively (). In our study, the concentrations of PK, PK epoxide, and MK-4 in the liver of female mice fed commercial lab chow (F2 diet, PK content of ∼0.1 μg/g) were 0.7 (1.5), 0.04 (0.1), and 15.6 (35.0) ng/g (pmol/g), respectively. This great difference may derive from the difference in the amount of PK in the diets used. Another explanation is that F2 diet contains a large amount of K₃ (60,000 pmol/g diet, added as an additive by the manufacturer) instead of PK, although it is unclear whether the normal lab chow used in Thijssen et al.'s study was supplemented with K₃ or not, and from our finding that K₃ is a substrate for MK-4 synthesis in tissues, it is most likely that K₃ in the diets could influence the concentrations of MK-4 and MK-4 epoxide in the liver of mice.

Among the tissues, the cerebrum was particularly unique, because it contained large amounts of MK-4 but little or no PK, probably due to the low efficiency with which lipoprotein-bound PK was incorporated into the cerebrum by virtue of the blood-brain barrier and other unknown systems. Additionally, it was shown in a PK supplementation study in rats that dietary PK intake caused MK-4 levels to rise most dramatically in the brain rather than other tissues (). This previous finding supports the current results indicating that the oral administration of a single dose of PK-d₇ at 10 μmol/kg body weight induced the accumulation of MK-4-d₇ in cerebra at concentrations of 82.9 ± 6.1 pmol/g tissue, which corresponded to nearly 30% of the concentrations (252.5 ± 10.3) of MK-4 in cerebra of age-matched female mice. Irrespective of the marked accumulation of MK-4, little is known about the role and functions of MK-4 in the brain. Based on this background, our attention was directed mostly to the conversion of PK into MK-4 in cerebra of mice. To investigate whether the biosynthesis of MK-4 occurs within the cerebrum, mice were administered intracerebroventricularly either PK-d₇ or K₃-d₈ and the accumulation of MK-4-d₇ in cerebra was examined at 24 h post-administration. In this experiment, the dosage of PK-d₇ and K₃-d₈ was lowered to 0.1 μmol/kg body weight because a high concentration of K₃ is likely to generate reactive oxygen species that affect cell survival and functions (). Consequently, despite the great amounts (2973.3 ± 220.3 pmol/g) of PK-d₇ found in the cerebra, MK-4-d₇ was undetectable. The reason why PK-d₇ directly injected into cerebra was not converted into MK-4-d₇ is not known, but it is postulated that, if the conversion is catalyzed by the enzyme(s) responsible for cleaving the side chain of PK and for prenylating the released K₃ to MK-4, these metabolic activities may occur in a site-specific and cell-specific manner. To obtain more insight into the tissue-specific MK-4 synthesis, cerebral slice culture and cerebral primary culture experiments were carried out. In these systems, the conversion of PK-d₇ and K₃-d₈ into MK-4-d₇ was confirmed and the conversion efficiency was ∼2-fold higher for K₃-d₈ than PK-d₇. To explore the neuronal cells involved in this conversion, a cerebral primary culture was prepared and examined to see whether MK-4-d₇ synthesis occurs in the presence or absence of PK-d₇ or K₃-d₈. As expected, both PK-d₇ and K₃-d₈ were converted into MK-4-d₇, and the conversion efficiency was much greater for K₃-d₈ than PK-d₇. Interestingly, MK-4 epoxide-d₇, an oxidized form of hydroxymenaquinone-4-d₇ (a reduced form of MK-4-d₇) generated by GGCX during the turnover of the vitamin K cycle, was found at higher concentrations than MK-4-d₇ (Fig. 6). This result suggests that, in the primary culture, the turnover of the vitamin K cycle was enhanced and MK-4-d₇ converted from PK-d₇ or K₃-d₈ was rapidly oxidized to MK-4-epoxide-d₇. We next examined the synthesis of MK-4-d₇ from PK-d₇ and K₃-d₈ in two primary cell cultures; cultures with a high density of neurons or astrocytes. Both neurons and astrocytes were able to convert K₃-d₈, but not PK-d₇, into MK-4-d₇ (Fig. 6). Compared with neurons, astrocytes were highly active in the conversion. We found large amounts of PK-d₇ in primary cultured neurons and astrocytes (data not shown), implying that both neurons and astrocytes were capable of absorbing PK-d₇ and possess the ability to convert K₃-d₈ into MK-4-d₇, but lack the ability to cleave the side chain of PK-d₇ to release K₃-d₇. Because the conversion of PK-d₇ into MK-4-d₇ obviously took place in both cerebral slice cultures and primary cell cultures, besides neurons and astrocytes, other neuronal cells such as oligodendrocytes may be involved in the cleavage process. Based on the available data, the release of K₃ from PK may occur in a limited number of tissues and the intestinal enterocytes may play a central role in this cleavage activity and in supplying K₃ to various tissues. Beyond this systemic conversion of PK to MK-4, the tissue-specific conversion of PK to MK-4 obviously exists as exemplified by our current study in the cerebra of mice (Fig. 6).

There are several limitations to this study. First, we were unable to detect K₃-d₇ in the plasma and cerebra of the mice orally given PK-d₇. Unfortunately, K₃-d₇ is not easily detected by LC-APCI-MS/MS because of difficulty in the detection of either the precursor ion or product ion derived from K₃. Several studies have shown that K₃, existing in the form of conjugates, glucuronides, or sulfates, of menadiol in urine, was successfully measured by HPLC using a reversed-phase column with fluorescence detection after oxidation of the conjugates, extraction with an organic solvent and purification using a mini-column (, ). However, the sensitivity and accuracy were not good enough to detect K₃ present in plasma and tissues at low levels. We are currently developing an efficient LC-MS/MS method to measure both K₃ and its D-labeled compound. If it becomes available, it will serve to explore in more detail not only the routes and functions of the above conversion process but also the comprehensive mechanisms underlying MK-4 synthesis in the brain. Second, it is postulated that the conversion of PK into MK-4 is a metabolic process involving the enzymes responsible for the cleavage of the side chain of PK and subsequent prenylation of K₃. There are conflicting reports in regard to the tissue-sites of K₃ release. Using HPLC analysis, Thijssen et al. () found that various cell lines were capable of converting K₃ into MK-4, but none of these cell lines were able to convert PK into MK-4, presumably due to a lack of K₃-releasing activity. On the other hand, Davidson et al. reported that HEK 293 cells converted PK into MK-4 epoxide (). In the current study, the PK-d₇ to MK-4-d₇ conversion indeed occurred in the cerebral slice culture as well as primary culture; however, neither neurons nor astrocyes converted PK-d₇ into MK-4-d₇. The reason for these conflicting results is as yet unclear, but one explanation may be that the PK to MK-4 conversion in cerebra occurs at selective sites via selective pathways. At present, we have no evidence related to these putative enzymes, and more studies are needed to define and characterize the enzymes.

In summary, the present study shows for the first time that MK-4-d₇ existing in cerebra of mice originates from PK-d₇ and/or K₃-d₈ intake, and we suggest that there are two routes of cerebral MK-4-d₇ accumulation, one is the release of K₃-d₇ from PK-d₇ in the intestine, followed by prenylation of K₃-d₇ into MK-4-d₇ in the tissue, and another is the release and prenylation of K₃-d₇ both within the cerebrum (Fig. 7). The present results, which are derived from structural data on the one hand and the biochemical characterization of animals, tissue cultures, or cell cultures on the other, may aid in the clarification of the physiological role of MK-4 in the brain and the development of new drugs for the treatment of neuronal diseases.

We thank Reimi Tatsumi, Yuji Takahashi, Akane Ishibuchi, Ayaka Toda, and Yoshie Okamoto for excellent technical assistance.