Usefulness of Derivatization in High-Performance Liquid Chromatography/Tandem Mass Spectrometry of Conjugated Vitamin D Metabolites

Transcription

1 1999 The Japan Society for Analytical Chemistry 619 Usefulness of Derivatization in High-Performance Liquid Chromatography/Tandem Mass Spectrometry of Conjugated Vitamin D Metabolites Original Papers Tatsuya HIGASHI, Kanako MIURA, Junichi KITAHORI and Kazutake SHIMADA Faculty of Pharmaceutical Sciences, Kanazawa University, 13-1 Takara-machi, Kanazawa , Japan The usefulness of derivatization in high-performance liquid chromatography/ion trap tandem mass spectrometry (electrospray ionization) of conjugated vitamin D metabolites was examined. 24,25-Dihydroxyvitamin D 3 [24,25(OH) 2D 3]-3- and -24-glucuronides (G) were converted to the adducts with 4-phenyl-1,2,4-triazoline-3,5-dione (PTAD), the methyl esters and the acetates. The combination of the derivatization with PTAD and methylation improved the detection limit in both the positive- and negative-ion modes. The PTAD adducts of 24,25(OH) 2D 3-3G and -24G also gave a characteristic product ion indicating the conjugation position by the positive-ms/ms/ms, which was not observed in the intact forms. 25-Hydroxyvitamin D 3 3-sulfate gave the deprotonated ion in the negative-ion mode, but no characteristic ion was observed in the positive-ion mode. On the contrary, its PTAD adduct gave a molecular-related ion in the positive-ion mode, the MS/MS/MS of which provided structural information, that is, the conjugation position and the existence of a sulfuric acid group. Keywords High-performance liquid chromatography/tandem mass spectrometry, derivatization, vitamin D, conjugates Although the metabolism of vitamin D (D), particularly the oxidation of its side chain, has been investigated by many groups, the conjugates of D metabolites still remain poorly understood. Axelson 1 and we 2 have reported that 25-hydroxyvitamin D 3 [25(OH)D 3 ] 3-sulfate (S) is a major circulating metabolite in humans. Furthermore, we identified the monoglucuronides (G) of D compounds {vitamin D 23, vitamin D 3 (D 3 ) 4, 25- hydroxyvitamin D 23, 25(OH)D 34, and 24,25-dihydroxyvitamin D 3 [24,25(OH) 2 D 3 ] 5 } as biliary metabolites from rats dosed with each compound in comparison with standard samples based on their chromatographic behavior during high-performance liquid chromatography (HPLC) and/or liquid chromatography (LC)/mass spectrometry (MS). 3,5 On the other hand, the structures of the other unknown conjugated D metabolites existing in small amounts in biological fluids should also be clarified. In this respect, it is necessary to establish a characterization method of metabolites whose standard samples are not available. Although LC/MS, especially LC/electrospray ionization (ESI)-MS, is the most promising analytical method for polar metabolites, such as these conjugates, this method gives fewer fragment ions; therefore, it is unsuitable for structure elucidation. Recently, much interest has been focused on LC/ion trap MS for the analysis of biological substances, because of its easy performance for multi-stage tandem MS (MS n ), which is highly specific and provides more structural information. Furthermore, it is thought that derivatization is also effective to obtain more abundant and reliable information on LC/MS. In addition, the derivatization is expected to improve the detection limit in LC/MS. In the present paper, the advantages of derivatization for LC/ion trap MS analysis of the conjugates of D metabolites shown in Fig. 1 were examined. First, the effect of the derivatization for detection responses was examined using 24,25(OH) 2 D 3-3G and -24G as model compounds. Second, the advantages of derivatization for the characterization of the structure were studied using some conjugated D metabolites. Experimental Materials and reagents 25(OH)D 3 3S 6, 25(OH)D 3 G (-3G and -25G) 7, and 24,25(OH) 2 D 3 G (-3G and -24G) 5 were synthesized in To whom correspondence should be addressed. Fig. 1 Structures of conjugated D metabolites.

2 620 ANALYTICAL SCIENCES JULY 1999, VOL. 15 our laboratories. 4-Phenyl-1,2,4-triazoline-3,5-dione (PTAD) was synthesized from 4-phenylurazole (Nacalai Tesque, Kyoto, Japan) and purified by sublimation prior to use. 8 The other reagents and solvents were commercially available and of analytical grade. Instruments LC/MS was performed using a Finnigan MAT LCQ (San Jose, CA, USA) connected to a JASCO PU-980 (Tokyo, Japan) chromatograph, and the following conditions were used for the analysis. ESI: the spray needle voltage was 5 kv; the heated capillary temperature, the sheath gas flow rate and the auxiliary gas were set at 200 C, 70 units and 20 units, respectively. The capillary voltage was 1 or 1 V and the tube lens offset was 10 or 10 V in the positive- or the negative-ion mode, respectively. A YMC-Pack Pro C18 (5 µm, cm i.d.) (YMC, Kyoto) column was used at a flow rate of 0.4 ml/min at 30 C. For MS n analysis, helium was used as the collision gas and the relative collision energy was set at 20%. Atmospheric pressure chemical ionization (APCI): the spray needle voltage was 6 kv; the heated capillary temperature and the sheath gas flow rate were set at 150 C and 80 units with a vaporizer temperature of 350 C. The capillary voltage and the tube lens offset were 1 and 10 V, respectively. Derivatization PTAD adduct: the mixture of 24,25(OH) 2 D 3-3G and -24G (each 280 pmol), or the other conjugated D metabolites (each 280 pmol), was dissolved in AcOEt (50 µl) containing PTAD (5 µg) and kept at room temperature for 1.5 h. MeOH (50 µl) was added to decompose any excess reagent, and the solvent was evaporated under a N 2 gas stream. Incidentally, the adducts of D compounds with a Cookson-type reagent, such as PTAD, usually consisted of 6S and 6R isomers, and the former was the main product 9,10, which was used for a LC/MS analysis. Methyl ester: a mixture of 24,25(OH) 2 D 3-3G and -24G (each 280 pmol) was dissolved in MeOH (50 µl), and a freshly prepared ethereal solution of diazomethane (1 ml) was added to this solution. The mixture was stored at room temperature for 1.5 h, and the solvent was evaporated under a N 2 gas stream. PTAD adduct methyl ester: the mixture of 24,25(OH) 2 D 3-3G and -24G (each 280 pmol) was subjected to methylation and then derivatized with PTAD as described above. Acetate: a mixture of 24,25(OH) 2 D 3-3G and -24G (each 280 pmol) was dissolved in pyridine Ac 2 O (2:1, v/v, 50 µl) and stored at room temperature for 1.5 h. MeOH (50 µl) was added to the mixture, and the solvent was evaporated under a N 2 gas stream. LC/ESI-MS n of conjugated D metabolites and their derivatives The conjugated D metabolites (each 280 pmol) and their derivatives were dissolved in EtOH (50 µl), onetenth of which (5 µl, each 28 pmol) was subjected to LC/ESI-MS n. The effect of derivatization for the detection responses was examined using 24,25(OH) 2 D 3-3G, -24G and their derivatives. The mobile phase used was a mixture of MeOH and 10 mm AcONH 4, whose ratio was changed so that 24,25(OH) 2 D 3 3G or its derivatives were eluted between 6 and 7 min. The base ion of each compound was monitored in the selected ion monitoring mode, and the signal to noise ratio (S/N) and the peak area of each derivative were compared with those of the intact compounds. In order to evaluate the advantages of derivatization for structural analysis, the intact conjugates and their derivatives were subjected to LC/ESI-MS n and their fragment patterns were examined. Results and Discussion Comparison of sensitivity between 24,25(OH) 2 D 3-3G, -24G and their derivatives Initially, efforts were directed to selecting a mobile phase suitable for LC/ESI-MS. We first used MeCN as an organic modifier, but methylated derivatives did not give any characteristic ions. On the contrary, when MeOH was used as an organic modifier, all of the examined compounds gave molecular-related ions with no fragment ion, which are listed in Table 1. In the negative-ion mode, the methyl esters gave adduct ions with acetic acid ([M+AcO] ) and the others gave deprotonated ions ([M H] ). The detection responses of each compound were evaluated by monitoring the intensity of these ions. The methylation decreased the intensity in 3G, but increased it in 24G. We have already reported that the increase in the intensity by methylation was also observed in 24,25(OH) 2 D 3 25G formed in the in vitro experiments. 5 These data indicated that this phenomenon was peculiar to compounds having a glucuronic acid on the side chain. Although the drivatization with PTAD was not always effective, the combination of it and methylation gave the strongest intensity; that is, the S/N was improved by about 6 10 times compared with intact 24,25(OH) 2 D 3-3G or -24G. Generally, although ESI is suitable for ionic compounds, such as glucuronide, these results showed that derivatization to a non-ionic compound was occasionally useful for improving the sensitivity. On the contrary, all of the above compounds gave no characteristic ions in the LC/APCI-MS. The speculation concerning these phenomena is as follows. In glucuronide, the deprotonated ion is directly produced from a glucuronide anion in the solution by desolvation in the ESI mode. On the other hand, compounds such as the methyl ester are ionized by the addition of ions derived from the solvent, such as AcO. This ionization mechanism is not the APCI mode, but close to this mode, because of their non-ionic character, and improves the sensitivity. Thus, the ionization mecha-

3 621 Table 1 Comparison of sensitivity in 24,25(OH) 2 D 3-3G, -24G and their derivatives Compound Mobile phase a Conjugation position Retention time/min Monitoring ion (m/z) Negative-ion mode S/N Relative peak area Monitoring ion (m/z) Positive-ion mode S/N Relative peak area G 7: [M H] [M+NH 4 ] G methyl ester 4: [M+AcO] [M+NH 4 ] G-PTAD 7: [M H] [M+NH 4 ] G-PTAD 3: [M+AcO] [M+NH 4 ] methyl ester G acetate 13: [M H] [M+NH 4 ] a. MeOH/10 mm AcONH 4 (v/v). Table 2 Fragment patterns of conjugated D metabolites and their derivatives in LC/positive-ESI-MS n Compound Precursor ions (m/z) Characteristic ions (m/z) a. MS/MS. b. MS/MS/MS. c. The product ion derived from the cleavage of the C6 7 bond as shown in Fig. 2. nisms are different in the glucuronide and its methyl ester, even if the same LC/MS conditions are used, and it is inferred that the ionization efficiency of the latter is superior to that of the former. Although the methylation gave good results in this study, it was not effective in the analysis of estriol glucuronides, in which the sensitivity did not change; if anything, it deteriorated (unpublished data to be reported elsewhere). The acetylated compounds responded with very weak intensity in the negative-ion mode. On the other hand, all of the examined compounds gave an adduct ion ([M+NH 4 ] + ) in the positive-ion mode. The responses were increased by all of the derivatization, and especially, the PTAD derivatives showed strong intensity whether they were methyl esters or not. These results indicated that the introduction of nitrogen and/or oxygen, which have high proton receptivity, improve the detection responses in the positive-ion mode. Usefulness for characterization of structure In a previous paper, we reported that LC/ESI-MS/MS

4 622 ANALYTICAL SCIENCES JULY 1999, VOL. 15 operating in the positive-ion mode was useful for the characterization of 24,25(OH) 2 D 3 G in rat bile, which gave fragment ions occurring from the loss of water and sugar moieties. 5 Based on these data, LC/MS n was performed in the positive-ion mode, and the fragment ions listed in Table 2 were obtained. The methyl esters and the acetates of 24,25(OH) 2 D 3 G showed fragment patterns similar to those observed for the intact forms, that is, the loss of water or acetic acid and sugar moiety. Although the relative intensities of the product ions were different between 3G and 24G derivatives, no remarkable difference was observed in their fragment patterns. These results suggested that, in the case where the authentic samples of the positional isomers are available, both are identified according to their chromatographic behavior and mass spectral data. If not, it is hard to distinguish them by only LC/MS n. It is well-known that most of the fragmentation observed in collision-induced dissociation with low energy as used in this study is the cleavage of carbonhetero atom bonds, such as ester, ether, and amido bonds, or loss of ROH (water, alcohol, and so on); but the cleavage of carbon carbon bonds hardly occur. 11 However, the PTAD adducts of D compounds were reported to give characteristic fragment ions caused by the cleavage of the C6 7 bond of the D skeleton in LC/MS. 12,13 Furthermore, we had applied this derivatization to LC/MS of D 3 fatty acid esters and obtained the molecular-related and characteristic fragment ions, which were helpful to identify the compound. 14 Based on these data, derivatization with PTAD was applied to the characterization of the structure of conjugated D metabolites. The adducts of 24,25(OH) 2 D 3-3G and -24G gave characteristic ions at m/z 474 and 298, respectively, in MS/MS/MS (Fig. 2). Similar results were also obtained in the case of 25(OH)D 3-3G and -25G (Table 2). These data indicate that the conjugation position, that is, whether the glucuronic acid is present on the A ring or the side chain can be determined by derivatization with PTAD followed by monitoring the product ion at m/z 474 in MS/MS/MS. As mentioned above, these PTAD adducts consist of the 6R- and 6S-epimers, but were easily separated by LC and did not differ in mass spectral patterns. Incidentally, the formation ratios were about 1:35 and 1:4.5 in 24,25(OH) 2 D 3-3G and -24G, respectively, which was affected by whether a bulky group such as glucuronic acid was present on the A-ring. The PTAD methyl ester derivatives of 24,25(OH) 2 D 3 G showed fragment patterns similar to those observed in the PTAD adducts. Although the intensity of the precursor ion of the PTAD methyl ester was much stronger than that of the PTAD adducts (Table 1), the intensity of the product ions derived from the cleavage of the C6 7 bond did not differ between these two derivatives. These data led us to the conclusion that additional methylation of the PTAD adduct is unnecessary for a structural analysis. 25(OH)D 3 3S showed a deprotonated ion ([M H] ) Fig. 2 Product ion mass spectra of PTAD adducts of 24,25(OH) 2D 3-3G (a) and -24G (b). Conditions: mobile phase, MeOH/10 mm AcONH 4 (7:3, v/v); precursor ions, m/z 785 [M+NH 4] + and then 768 [M+H] + ; relative collision energy, 20% each. with a strong intensity in the negative-ion mode (S/N=77 per 28 pmol injection), but showed no characteristic ions to elucidate its structure in the LC/MS spectrum operating in the positive-ion mode. On the contrary, its PTAD adduct gave an adduct ion ([M+NH 4 ] + ), though its intensity was not as strong (S/N=5), which made the MS n analysis possible. The MS/MS/MS analysis shown in Table 2 provided structural information, which was similar to that observed for adducts of glucuronides together with the loss of SO 3 H. In conclusion, the advantage of derivatization for the analysis of conjugated D metabolites using LC/ion trap MS n was examined. The derivatization with PTAD was very useful for not only to characterize the structure,

5 623 but also to increase the detection response in the positive-ion mode. The adduct gave a characteristic product ion indicating the conjugation position by MS n. Furthermore, when it was combined with methylation, a molecular-related ion with a strong intensity was obtained in the negative-ion mode. The data obtained in this study are expected to be useful in the identification and characterization of unknown conjugated D metabolites in biological fluids. Part of this work was supported by a grant from the Ministry of Education, Science, Sports and Culture of Japan. References 1. M. Axelson, FEBS Lett., 191, 171 (1985). 2. K. Shimada, K. Mitamura and N. Kitama, Biomed. Chromatogr., 9, 229 (1995). 3. K. Shimada, K. Mitamura and I. Nakatani, J. Chromatogr. B, 690, 348 (1997). 4. K. Shimada, I. Nakatani, K. Saito and K. Mitamura, Biol. Pharm. Bull., 19, 491 (1996). 5. T. Higashi, M. Horike, R. Kikuchi and K. Shimada, Steroids, in press. 6. G. Parmentier and H. Eyssen, Steroids, 30, 583 (1977). 7. K. Shimada, K. Sugaya, H. Kaji, I. Nakatani, K. Mitamura and N. Tsutsumi, Chem. Pharm. Bull., 43, 1379 (1995). 8. R. C. Cookson, S. S. Gupte, I. D. R. Stevens and C. T. Watts, Org. Synth., 51, 121 (1971). 9. M. Shimizu and S. Yamada, Vitamins (Japan), 68, 15 (1994). 10. K. Shimada, K. Mitamura, M. Miura and A. Miyamoto, J. Liq. Chromatogr., 18, 2885 (1995). 11. M. Ishigai, Y. Ishitani and K. Kumaki, J. Chromatogr. B, 704, 11 (1997). 12. B. Yeung, P. Vouros, M.-L. Siu-Caldera and G. S. Reddy, Biochem. Pharmacol., 49, 1099 (1995). 13. A. S. Weiskof, J. Cunniff, F. Bjorkling, E. Binderup, L. Binderup and P. Vouros, Vitamin D: Chemistry, Biology and Clinical Applications of the Steroid Hormone, ed. A. W. Norman, R. Bouillon and M. Thomasset, p. 745, Vitamin D Workshop, Inc., California, K. Mitamura, Y. Nambu, M. Tanaka, A. Kawanishi, J. Kitahori and K. Shimada, J. Liq. Chromatogr. Rel. Technol., 22, 367 (1999). (Received February 19, 1999) (Accepted April 19, 1999)