Enantiomeric Separation and Detection of 2-Arylpropionic Acids

Technical Notes Anal. Chem. 1997, 69, 1793-1799

Enantiomeric Separation and Detection of 2-Arylpropionic Acids Derivatized with [(N,N-Dimethylamino)sulfonyl]benzofurazan Reagents on a Modified Cellulose Stationary Phase by High-Performance Liquid Chromatography Takeshi Fukushima, Tomofumi Santa, Hiroshi Homma, Salma M. Al-kindy, and Kazuhiro Imai*

Faculty of Pharmaceutical Sciences, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113, Japan

(RS)-2-Arylpropionic acids (2-APAs) were derivatized with the fluorogenic reagents, 4-[(N,N-dimethylamino)sulfonyl]-7-piperazino-2,1,3-benzoxadiazole (DBD-PZ) and 4-[[(N-hydrazinoformyl)methyl]-N-methyl]amino-7-[N,N(dimethylamino)sulfonyl]-2,1,3-benzoxadiazole (DBDCOHz), and their enantiomeric separation by a chiral stationary phase high-performance liquid chromatography was investigated in the reversed-phase mode with H2O/ CH3CN or H2O/MeOH as the mobile phase on a column of cellulose tris(3,5-dimethylphenyl carbamate) coated on a silica gel support (Chiralcel OD-R). The derivatives with DBD-PZ were enantiomerically separated well under the elution condition of H2O/MeOH, based on the π-π interaction between the derivatives and the stationary phase. The rigid and bulky structure of DBD-PZ was demonstrated to be more effective as compared to the less rigid ones. The derivatives with DBD-COHz were more efficiently separated into each enantiomer with H2O/CH3CN as the eluent. The effective separation was based on hydrogen-bonding interaction between the acid hydrazide of the derivatives and the carbamoyl moiety of the stationary phase. There was a reversal in the elution order of the enantiomers between the two fluorescent derivatives. The detection limits obtained for each enantiomer were approximately 10-30 fmol on column. The derivatization with the reagent and the concomitant use of the reversedphase and chiral stationary-phase HPLC were demonstrated to be useful for the enantiomeric quantification in rat plasma after intravenous administration of flurbiprofen racemate, a representative of 2-APAs. The sensitive determination of drug enantiomers is increasingly a requisite in the field of clinical pharmacy, especially in the field of therapeutic drug monitoring, since many new optically active drugs have been approved and launched in the market. However, not so many are in clinical use. For example, 2-arylpropionic acid (2-APA)-type nonsteroidal antiinflammatory drugs (NSAID) are now used as racemates in the treatment of rheumatic and S0003-2700(96)01105-5 CCC: $14.00

© 1997 American Chemical Society

inflammatory diseases. Nonetheless the enantiomeric determination of NSAID is of extraordinary importance because, although the pharmacological activity of 2-APA resides in the S isomer, the nonactive R isomer is inverted into the S isomer1-3 via thioester formation with coenzyme A4-8 in the body. In order to keep an appropriate therapy, enantiomeric determination is essential.9,10 Numerous methods for the enantiomeric separation and detection of (RS)-2-APAs have been described, most of which were by highperformance liquid chromatography (HPLC) with ultraviolet (UV) detection because of its ease of operation.11-20 However, taking into account the invasive nature of withdrawing a blood sample from the patient, a highly sensitive quantification method that required less sample is preferred. For sensitive detection, a fluorescent diastereomeric21-24 or nonchiral derivatization tech(1) Hutt, A. J.; Caldwell, J. J. Pharm. Pharmacol. 1983, 35, 693-704. (2) Foster, R. T.; Jamali, F. Drug Metab. Dispos. 1988, 16, 623-634. (3) Cardwell, J.; Hutt, A. J.; Fournel-Gigleux, S. Biochem. Pharmacol. 1988, 37, 105-114. (4) Baillie, T. A.; Adams, W. J.; Kaiser, D. G.; Olanoff, L. S.; Halstead, G. W.; Harpootlian, H.; Van Giessen, G. J. J. Pharm. Exp. Ther. 1989, 249, 517523. (5) Chen, C.-S.; Shieh, W.-R.; Lu, P.-H.; Harriman, S.; Chen, C.-Y. Biochem. Biophis. Acta 1991, 1078, 411-417. (6) Sanins, S. M.; Adams, W. J.; Kaiser, D. G.; Halstead, G. W.; Hosley, J.; Barnes, H.; Baillie, T. A. Drug Metab. Dispos. 1990, 19, 405-410. (7) Shieh, W.-R.; Chen, C.-S. J. Biol. Chem. 1993, 268, 3487-3493. (8) Hall, S. D.; Xiaotao, Q. Chem.-Biol. Int. 1994, 90, 235-251. (9) Hutt, A. J.; Caldwell, J. Clin. Pharmacokinet. 1984, 9, 371-373. (10) Adams, S. S.; Bressloff, P.; Mason, C. G. J. Pharm. Pharmacol. 1976, 28, 256-257. (11) Wainer, I. W.; Doyle, T. D. J. Chromatogr. 1984, 284, 117-124. (12) Hutt, A. J.; Fournel, S.; Caldwell, J. J. Chromatogr. 1986, 378, 409-418. (13) McDaniel, D. M.; Snider, B. G. J. Chromatogr. 1987, 404, 123-132. (14) Okamoto, Y.; Aburatani, R.; Kaida, Y.; Hatada, K. Chem. Lett. 1988, 11251128. (15) Noctor, T. A. G.; Felix, G.; Wainer, I. W. Chromatographia 1991, 31, 5559. (16) Aboul-Enein, H. Y.; Bakr, S. A. J. Liq. Chromatogr. 1992, 15, 1983-1992. (17) Pirkle, W. H.; Welch, C. J. J. Liq. Chromatogr. 1992, 11, 1947-1954. (18) Oi, N.; Kitahara, H.; Aoki, F.; Kisu, N. J. Chromatogr., A 1995, 689, 195201. (19) Carr, R. A.; Caille, G.; Ngoc, A. H.; Foster, R. T. J. Chromatogr., B 1995, 668, 175-181. (20) Loulin, R.; Vakily, M.; Jamali, F. J. Chromatogr., B 1996, 679, 196-198. (21) Toko’oka, T.; Ishibashi, M.; Terao, T. Analyst 1992, 117, 727-733.

Analytical Chemistry, Vol. 69, No. 9, May 1, 1997 1793

nique with a chiral stationary phase25-27 is usually adopted in HPLC. The advantage of diastereomeric derivatization with chiral reagents is that it allows the use of a common and inexpensive reversed-phase HPLC column such as an octadecylsilica (ODS) column. However, the optical purity of the reagent is critical for getting a precise value of the enantiomer concentration. Thus, the racemization of the reagents during storage should always be taken into consideration. Another disadvantage is that, occasionally, each diastereomer has a different fluorescence quantum yield, thus requiring different calibration curves for the quantification. The nonchiral fluorogenic reagents having a 2,1,3-benzoxadiazole (benzofurazan)28-34 moiety having long excitation and emission wavelengths, which minimized interference by native fluorescence from endogenous substances so that highly selective and sensitive detection was achieved. In this work, we study the enantiomeric separation on a chiral stationary phase, a cellulose tris(3,5-dimethylphenyl carbamate)11,12,35-41 coated on a silica gel support (Chiralcel OD-R) and highly sensitive detection of 2-APAs derivatized with the two different types of nonchiral [(dimethylamino)sulfonyl]benzofurazan reagents for a carboxyl group, 4-[(N,N-dimethylamino)sulfonyl]-7-piperazinobenzofurazan (DBDPZ, Figure 1)31 and [4-(N-hydrazinoformyl)methyl](N-methylamino)]-7-[(N,N-dimethylamino)sulfonyl]-2,1,3-benzoxadiazole (DBD-COHz, Figure 1).34 The interaction between the derivatives and the stationary phase for the enantiomeric separation is discussed. The application of DBD-COHz and/or DBD-PZ is demonstrated by the enantiomeric quantification of flurbiprofen in rat plasma after intravenous administration of the racemate. EXPERIMENTAL SECTION Materials. DBD-PZ was purchased from Tokyo Kasei Co., Ltd. (Tokyo, Japan). DBD-COHz and 4-[(N,N-dimethylamino)(22) Kondo, J.; Suzuki, N.; Naganuma, H.; Imaoka, T.; Kawasaki, T.; Nakanishi, A.; Kawahara, Y. Biomed. Chromatogr. 1994, 8, 170-174. (23) Iwaki, K.; Bunrin, T.; Kameda, Y.; Yamazaki, M. J. Chromatogr., A 1994, 662, 87-93. (24) Yasaka, Y.; Matsumoto, T.; Tanaka, M. Anal. Sci. 1995, 11, 295-297. (25) Fukushima, T.; Kato, M.; Santa, T.; Imai, K. Biomed. Chromatogr. 1995, 9, 10-17. (26) Fukushima, T.; Santa, T.; Homma, H.; Nagatomo, R.; Imai, K. Biol. Pharm. Bull. 1995, 18, 1130-1132. (27) Fukushima, T.; Kato, M.; Santa, T.; Imai, K. Analyst 1994, 120, 381-383. (28) Imai, K.; Watanabe, Y. Anal. Chim. Acta 1981, 130, 377-383. (29) Imai, K.; Toyo’oka, T.; Watanabe, Y. Anal. Biochem. 1983, 128, 471-473. (30) Toyo’oka, T.; Imai, K. Anal. Chem. 1984, 56, 2461-2464. (31) Toyo’oka, T.; Ishibashi, M.; Takeda, Y.; Nakashima, K.; Akiyama, S.; Uzu, S.; Imai, K. J. Chromatogr. 1991, 588, 61-71. (32) Toyo’oka, T.; Suzuki, T.; Saito, Y.; Uzu, S.; Imai, K. Analyst 1989, 114, 413-419. (33) Imai, K.; Fukushima, T.; Yokosu, H. Biomed. Chromatogr. 1994, 8, 107113. (34) Santa, T.; Kimoto, K.; Fukushima, T.; Homma, H.; Imai, K. Biomed. Chromatogr. 1996, 10, 183-185. (35) Okamoto, Y.; Kawashima, M.; Aburatani, R.; Hatada, K.; Nishiyama, T.; Masuda, M. Chem. Lett. 1986, 1237-1240. (36) Okamoto, Y.; Kawashima, M.; Hatada, J. Chromatogr. 1986, 363, 173186. (37) Okamoto, Y.; Kaida, Y.; Hayashida, H.; Hatada, K. Chem. Lett. 1990, 909912. (38) Ishikawa, A.; Shibata, T. J. Liq. Chromatogr. 1993, 16, 859-878. (39) Okamoto, Y.; Kaida, Y. J. Chromatogr., A 1994, 666, 403-419. (40) Aboul-Enein, H. Y.; Serignese, V. Biomed. Chromatogr. 1994, 8, 22-25. (41) Yashima, E.; Okamoto, Y. Bull. Chem. Soc. Jpn. 1995, 68, 3289-3307.

1794 Analytical Chemistry, Vol. 69, No. 9, May 1, 1997

Figure 1. Chemical structures of 4-[(N,N-dimethylamino)sulfonyl]benzofurazan reagents and the derivatization scheme for 2-arylpropionic acid.

sulfonyl]-7-(N-ethylenediamino)-2,1,3-benzoxadiazole (DBD-ED)42 were synthesized in our laboratory. Chiralcel OD-R (a modified cellulose with a 3,5-dimethylphenylcarbamoylation, 250 × 4.6 mm i.d., 10 µm) was obtained from Daicel Co., Ltd. (Osaka, Japan). Ultron ES-phCD (a modified β-cyclodextrin with a phenylcarbamoylation, 150 × 6.0 mm i.d., 5 µm) was from Sinwa Chemical Industry Co. (Kyoto, Japan). TSK gel ODS-80 Ts (150 × 4.6 mm, i.d., 5 µm) was donated by Tosoh Co., Ltd. (Tokyo, Japan). (RS)and (S)-Phenylpropionic acid, triphenylphosphine (TPP), and 2,2′dipyridyl disulfide (DPDS) were from Tokyo Kasei Co., Ltd. (Tokyo, Japan). (RS)-Ibuprofen, (RS)-ketoprofen, trifluoroacetic acid (TFA) of amino acid analysis grade, and acetone of HPLC grade were purchased from Wako Pure Chemicals, (Osaka, Japan). H2O was used after purification by Milli-Q system (Millipore, Waltham, MA). (RS)-Flurbiprofen was purchased from Sigma (St. Louis, MO). (S)-Ibuprofen, -ketoprofen, and -flurbiprofen were from Nagase Industry Co. Ltd. (Hyogo, Japan). (RS)and (S)-Pranoprofen were from Yoshitomi Pharmaceutical Co., Ltd. (Fukuoka, Japan). Methanol (MeOH), dimethylformamide (DMF), toluene, and acetonitrile were of HPLC grade from Kanto Chemical Co. (Tokyo, Japan). Poly(ethylene glycol) (PEG) 400 was from Kanto Chemical Co. The polypropyrene vials (2.0 mL) used in this study were purchased from Iatron Inc. (Tokyo, Japan). (42) Pablo, P.; Fukushima, T.; Santa, T.; Homma, H.; Mori, S.; Yokosu, H.; Imai, K. Anal. Chim. Acta, in press.

Derivatization and Preparation of the Fluorescent Derivatives. A 50-µL sample of (RS)- or (S)-2-APA dissolved in CH3CN (5 mM) was added to 50 µL of DBD-PZ or DBD-ED in CH3CN (50 mM) or DBD-COHz in CH3CN/DMF (50/50, v/v, 50 mM) in the presence of 50 µL of 140 mM TPP and DPDS in CH3CN. The reaction mixture was vigorously mixed and allowed to stand at room temperature for 5 h. A 10-µL sample of the solution was diluted 100 times with 0.1% TFA in H2O/CH3CN (40/60, v/v), and 10 µL of the diluted solution was injected into the HPLC equipped with a TSKgel ODS-80Ts column (150 × 4.6 mm i.d., Tosoh, Tokyo, Japan) through a Model 7725i sample injector (Rheodyne, Cotati, CA) with a 10-µL sample loop. The HPLC used consisted of a L-7100 intelligent pump, a L-7480 fluorospectrometric detector, and a D-7500 chromato-integrator (Hitachi, Tokyo, Japan). The column temperature during operation was ambient. The mobile phase was 0.1% TFA in H2O/CH3CN (40/60, v/v) for DBD-PZ, (55/45, v/v) for DBD-ED, and (60/40, v/v) for DBD-COHz. The flow rate was 1.0 mL/min. The fluorescent peak of the (RS)-2APA derivative detected at 560 nm with 450 nm as an excitation wavelength was isolated into a vial (2.0 mL), and the fraction was concentrated in vacuo by a centrifugal evaporator, SPE-200 (Shimadzu, Tokyo, Japan). The residue was dissolved in 500 µL of CH3CN and filtered through a 0.5 µm filter (Column Guard LCR, Nihon Millipore, Tokyo, Japan); 10 µL of the filtrate was injected onto a column of Chiralcel OD-R or Ultron ES-phCD. The flow rate was 0.5 mL/min, and the column temperature was ambient. Fluorometric detection was performed at 560 nm with 450 nm as an excitation wavelength. The mobile-phase composition employed was H2O/CH3CN (40/60, v/v) or H2O/MeOH (0/ 100 or 10/90, v/v). In the case of an addition experiment with acetone, 5% acetone in H2O/MeOH (10/90, v/v) or H2O/CH3CN (40/60, v/v) was used for the mobile phase. Determination of Detection Limit. Fifty femtomoles of each enantiomer of the isolated fluorescent derivative was injected into the HPLC column (Chiralcel OD-R), and the signal (S) and noise (N) levels were measured. The amount of enantiomer of the derivative on-column showing S/N ) 3 was regarded as a detection limit and determined by the ratio of 50 fmol on-column to the S/N. Quantification of Flurbiprofen Enantiomer in Rat Plasma after Administration of Flurbiprofen Racemate. Male Wistar rats weighing 250-260 g (7 week old, Nihon Charles River, Kanagawa, Japan) were housed in an environmentally controlled room with a 12-h light/dark cycle. Diet and tap water were provided ad libitum. (RS)-Flurbiprofen dissolved in PEG 400 (10 mg/mL) was administered intravenously via a tail vein at a dose of 20 mg/kg of weight43 (n ) 3). Blood (0.1 mL) was withdrawn from the jugular vein with a heparinized syringe at the following time points: 5 min, 1 h, and 5 h after administration. The blood samples were immediately centrifuged at 4000g for 10 min, and the plasma samples obtained were stored at -20 °C until analysis. Extraction, Derivatization, Isolation, and Quantification. Plasma (10 µL) was mixed vigorously with 90 µL of CH3CN and centrifuged at 4000g for 5 min. The supernatant (50 µL) was transferred to another vial containing 100 µL of 0.5 N HCl and 200 µL of toluene. After mixing for 30 s, 80 µL each of the organic layer was transferred to the respective two reaction vials and evaporated to dryness by SPE-200. Each residue was dissolved in 20 µL of CH3CN and derivatized with DBD-PZ or DBD-COHz (43) Evrard, P. A.; Cumps, J.; Verbeeck, R. K. Pharm. Res. 1996, 13, 18-22.

Figure 2. Chemical structure of the chiral moiety in a modified cellulose chiral stationary phase (Chiralcel OD-R).

in a manner similar to that described above. The reaction mixture was diluted three times with the mobile phase, and 10 µL of the solution was subjected to reversed-phase HPLC. The quantification of flurbiprofen as a mixture of R and S forms (total flurbiprofen) was made on the chromatogram obtained using the calibration curves described below. Next, the peak fraction of the flurbiprofen derivative was collected (elution time between 11.5 and 12.5 min for DBD-PZ and 39.0 and 41.0 min for DBDCOHz) and evaporated to dryness by SPE-200, and then the residue was dissolved in 30 µL of CH3CN. An aliquot of the solution was subjected to HPLC with the chiral stationary phase. Then, the enantiomeric ratio of the flurbiprofen derivatives was calculated by the peak area ratio of (S)- to (R)-flurbiprofen on the chromatogram. (S)- or (R)-Flurbiprofen concentration was quantified by multipling the enantiomeric ratio to the total flurbiprofen concentration obtained by reversed-phase HPLC. The validity of this calculation stems from the fact that each enantiomer affords the same peak area on the chromatogram. Calibration Curves for Total Flurbiprofen Derivatives. To 10µL aliquots of a blank plasma was added 70 µL of CH3CN and 20 µL of racemic flurbiprofen solution (6.25, 12.5, 25.0, or 50.0 µg/ mL in CH3CN) to obtain calibration curves for the derivatives with DBD-PZ or DBD-COHz on the reversed-phase column. The solution was mixed vigorously and treated in a manner similar to that described above. The calibration curves were obtained for the peak areas vs the amounts of the spiked flurbiprofen. RESULTS AND DISCUSSION Enantiomeric Separation of the 2-APA Derivatives with DBD-PZ. We first tried the separation with a phenylcarbamoylated β-cyclodextrin bonded chiral stationary phase (Ultron ESphCD), since an effective enantiomeric separation (separation factor R ) 1.08 under the elution condition of H2O/MeOH/CH3CN, 50/40/10, v/v/v) of (RS)-mexiletine [1-(2,6-dimethylphenoxy)2-aminopropane] derivatized at the amino functional moiety with 4-fluoro-7-nitro-2,1,3-benzoxadiazole (NBD-F) has already been reported.27 However, the enantiomeric separation of 2-APAs derivatized at the carboxyl moiety with DBD-PZ was insufficient (R ) 1.00) and with broad peak shape. A similar result was obtained with the nitrobenzofurazan reagent, 4-nitro-7-piperazino2,1,3-benzoxadiazole (NBD-PZ).31 It may be caused by the difficulty in penetrating into the β-cyclodextrin cavity by the bulky structure of the 2-APA derivatives with the reagent as compared to that of mexiletine. Thus, we searched for other types of chiral stationary phases. The finally selected column was of a stationary phase of a cellulose tris(3,5-dimethylphenyl carbamate) coated on a silica gel support (Chiralcel OD-R, Figure 2), which has a highly ordered 3/2 left-handed helical conformation on the basis of X-ray44 (44) Vogt, U.; Zugenmaier, P. Ber. Bunsenges. Phys. Chem. 1985, 89, 12171224.

Analytical Chemistry, Vol. 69, No. 9, May 1, 1997

1795

Table 1. Capacity Factors (k′) and Separation Factors (r) of 2-APA Derivatives with DBD-PZ by Chiral Stationary-Phase HPLC Using H2O/CH3CN or H2O/MeOH as a Mobile Phasea

Table 2. Capacity Factors (k′) and Separation Factors (r) of 2-APA Derivatives with DBD-ED by Chiral Stationary-Phase HPLC Using H2O/CH3CN or H2O/MeOH as a Mobile Phasea

H2O/CH3CN (40/60) H2O/MeOH (0/100)

H2O/CH3CN (40/60) H2O/MeOH (10/90)

2-APA

k′(R)

k′(S)

R

k′(R)

k′(S)

R

2-APA

k′(R)

k′(S)

R

k′(R)

k′(S)

R

ibuprofen ketoprofen flurbiprofen pranoprofen 2-phenylpropionic acid

7.15 5.62 9.95 4.76 3.50

8.19 5.88 10.80 4.76 4.19

1.15 1.05 1.09 1.00 1.20

1.94 3.57 3.54 5.10 2.57

3.10 4.81 4.90 5.84 3.21

1.60 1.35 1.38 1.15 1.25


2.11 1.74 2.65 1.24 0.96

2.11 1.74 2.65 1.24 1.06

1.00 1.00 1.00 1.00 1.10

1.38 2.06 2.41 1.98 0.90

1.43 2.65 2.84 1.98 1.01

1.04 1.29 1.18 1.00 1.12

a

k′(R) and k′(S) are the capacity factors of the respective enantiomers.

Figure 3. Chromatograms of pranoprofen derivatives with DBDPZ (a) and DBD-COHz (b) by chiral stationary-phase HPLC. The mobile phases employed are H2O/MeOH (0/100) for DBD-PZ and H2O/CH3CN (40/60) for DBD-COHz. Each of the upper chromatograms is derived from the S enantiomer of pranoprofen and the lower from the racemic pranoprofen, respectively. The other HPLC conditions are described in the Experimental Section.

analysis, followed by computational graphics.41,45 The interactions between cellulose tris(3,5-dimethylphenyl carbamate) and the racemate to be separated have been reported to be hydrogenbonding and dipole-dipole interaction with the carbamoyl groups and a π-π interaction between the 3,5-dimethylphenyl ring and the aromatic groups of the racemate.41 The indicated interaction mode for the stationary phase seemed to be applicable to the enantiomeric separation of the derivatives we selected in this study. First, the enantiomeric separation of (RS)-ketoprofen derivative with DBD-PZ was investigated using H2O/CH3CN or H2O/MeOH as the mobile phase since, in the case of application to a biological specimen such as plasma or urine, the reversed-phase mode of separation is preferred. As a consequence, they were enantiomerically separated (R ) 1.05) using H2O/CH3CN (40/60, v/v), suggesting an appropriate fit between the structure of the derivative and the stereochemical space made by the stationary phase. The ability of the organic solvent in the mobile phase to achieve suitable enantiomeric separation of 2-APA derivatives with DBD-PZ, was investigated. Table 1 shows the capacity factors (k′) and R’s of the derivatives obtained with H2O/CH3CN or H2O/ MeOH as the mobile phase, respectively. (45) Yashima, E.; Yamada, M.; Kaida, Y.; Okamoto, Y. J. Chromatogr., A 1995, 694, 347-354.

1796


a

k′(R) and k′(S) are the capacity factors of the respective enantiomers.

With MeOH as an organic component in the mobile phase, greater R values were obtained. Pranoprofen derivatives, especially, could be well separated in MeOH, although the derivatives were not separated in H2O/CH3CN. Typical chromatograms of the derivatives using MeOH as the mobile phase are shown in Figure 3a. Considering that an alcoholic hydroxyl group is capable of interacting with an amido group through hydrogen bonding,36,41 MeOH in the mobile phase is likely to interact with the amido group to prevent hydrogen bonding between the enantiomeric derivatives and a carbamoyl group in the stationary phase. However, with 100% MeOH in the mobile phase, a satisfactory separation of the enantiomer of the derivatives was achieved with R values ranging from 1.15 to 1.60. These results indicate that the major interaction necessary for the enantiomeric separation of the derivatives seemed to be a π-π interaction between benzofurazan or an aromatic moiety of the derivatives and the 3,5-dimethylphenyl groups in the stationary phase, accompanied by a minor interaction caused by a hydrogenbonding or a dipole-dipole interaction with the carbamoyl moiety. It is considered that the diastereomeric complexes formed via π-π interaction in the stationary phase could have the most different energy between the enantiomers and give effective enantiomeric separation of the derivatives with DBD-PZ. Hydrophobic interaction between the derivatives and the stationary phase may not be crucial since the elution orders of the derivatives on the chiral stationary phase observed in this case were different from that observed in the conventional reversedphase mode of separation, where the elution orders of the derivatives were pranoprofen, 2-phenylpropionic acid, ketoprofen, flurbiprofen, and ibuprofen, whereas on the former stationary phase, the elution orders were not predictable by their hydrophobicities (Table 1). According to the data obtained, a mobile phase consisting of H2O/MeOH afforded an excellent enantiomeric separation. The detection limits for R and S enantiomers with 100% MeOH were approximately 40-60 fmol on-column. In regard to the elution order of the enantiomers, the R enantiomer of any 2-APA derivative eluted first as compared to the S enantiomer using either CH3CN or MeOH as an organic component of the mobile phase. Effect of the Chemical Structure of the Benzofurazan Reagent on the Enantiomeric Separation of 2-APA Derivatives. Following the success of the enantiomeric separation of the derivatives with DBD-PZ, we next investigated the effect of the chemical structure of the benzofurazan reagent on the enantiomeric separation. DBD-ED, which has an ethylamino

Table 3. Capacity Factors (k′) and Separation Factors (r) of 2-APA Derivatives with DBD-COHz by Chiral Stationary-Phase HPLC Using H2O/CH3CN or H2O/MeOH as a Mobile Phasea H2O/CH3CN (40/60) H2O/MeOH (0/100) 2-APA

k′(S)

k′(R)

R

k′(S)

k′(R)

R


1.41 1.10 1.95 0.83 0.68

2.97 2.17 3.42 1.26 2.42

2.11 1.97 1.75 1.52 3.56

0.30 0.51 0.51 0.80 0.36

0.40 0.65 0.60 0.88 0.62

1.33 1.27 1.18 1.10 1.72

a

k′(S) and k′(R) are the capacity factors of the respective enantiomers.

Table 4. Effect of the Addition of 5% Acetone to the Mobile Phases on the Enantiomeric Separation of 2-Arylpropionic Acid Derivatives with DBD-PZ, DBD-ED, and DBD-COHz by the Chiral Stationary Phase HPLC 5% acetone

without acetone

H2O/CH3CN (40/60)

DBDPZ

DBDED

DBDCOHz

DBDPZ

DBDED

DBDCOHz


1.15 1.05 1.09 1.00 1.20

1.00 1.00 1.00 1.00 1.11

2.06 1.92 1.70 1.51 3.44

1.15 1.05 1.09 1.00 1.20

1.00 1.00 1.00 1.00 1.10

2.11 1.97 1.75 1.52 3.56

H2O/MeOH (10/90)

DBDPZ

DBDED

DBDCOHz

DBDPZ

DBDED

DBDCOHz


1.43 1.16 1.28 1.16 1.32

1.00 1.27 1.16 1.00 1.12

1.66 1.60 1.47 1.24 2.77

1.44 1.16 1.28 1.17 1.32

1.04 1.29 1.18 1.00 1.12

1.70 1.63 1.50 1.24 2.81

group for a coupling site, was selected since it has the same number of methylene units between the N atom linked to the benzofurazan skeleton of DBD-PZ and the N atom linked to the carbonyl group of 2-APA (Figure 1). The k′s and R’s of the derivatives obtained are summarized in Table 2. It shows that the derivatives with the DBD-ED eluted faster with smaller R values than those with DBD-PZ. Considering that (1) the same elution orders of enantiomer were observed between both the derivatives and (2) the larger R values were obtained using MeOH as compared to CH3CN as a component in the mobile phase, indicating the interaction mode of the derivatives with DBD-ED and the stationary phase seemed to be almost similar to that of the derivatives with DBD-PZ, the data suggest that a rigid and bulky structure of the 2-APA derivatives is effective for the enantiomeric separation on this stationary phase. Enantiomeric Separation of 2-APA Derivatives with DBDCOHz. DBD-COHz is an acid hydrazide-type fluorescent reagent bearing the 4-[(N,N-dimethylamino)sulfonyl]benzofurazan (Figure 1). As hydrogen-bonding or dipole-dipole interaction between the acid hydrazide moiety and the carbamoyl group of the stationary phase could occur, we expected a difference in the interaction mode of the enantiomeric separation of the derivatives from those with DBD-PZ would be observed. Table 3 shows k′s and R’s of the 2-APA derivatives with DBD-COHz obtained with H2O/CH3CN or H2O/MeOH as the mobile phase, respectively.

Figure 4. Chromatograms of the rat plasma total flurbiprofen derivatives with DBD-PZ (a and b) and DBD-COHz (c and d) by reversed-phase HPLC: (a) blank, (b) spiked with flurbiprofen, (c) blank, and (d) spiked with flurbiprofen. The arrow indicates the peak of the flurbiprofen derivative. The other HPLC conditions are described in the Experimental Section.

The derivatives were well separated with H2O/CH3CN as the mobile phase, and larger R values were obtained than in the case of the derivatives with DBD-PZ. Typical chromatograms are shown in Figure 3b. It is preferable for fluorometric detection to use a high content of organic solvent in the mobile phase because an enhancement of fluorescence intensity takes place in the less polar organic solvent with a sharp peak shape. The detection limit of each enantiomer with H2O/CH3CN (10/90, v/v) was in the range of 10-30 fmol on-column. Thus, DBD-COHz is preferable for the enantiomeric separation as well as the highly sensitive detection of 2-APAs. Another phenomenon observed is that the S enantiomers of the derivatives with DBD-COHz eluted first. This elution order was opposite to that of derivatives with DBD-PZ and DBD-ED. As indicated above, both the fluorescent derivatives could interact with cellulose tris(3,5-dimethylphenyl carbamate) through a different interaction mode for enantiomeric separation in the chromatographic process. The reversal elution of an insect growth regulator (pyriproxyfen) by changing the composition of the mobile phases in the normal-phase mode was reported;47 however, the observed case was the first example in the reversedphase mode in which the elution order of enantiomers could be reversed by changing the chemical structure of the derivatives formed. Effect of Addition of 5% Acetone to the Mobile Phase on Enantiomeric Separation. In order to further investigate the interaction site of the solute (fluorescent derivatives) with the (46) Yashima, E.; Yamada, M.; Okamoto, Y. Chem. Lett. 1994, 579-582. (47) Okamoto, M.; Nakazawa, H. J. Chromatogr. 1991, 588, 177-180.


1797

Figure 5. Chromatograms of the rat plasma flurbiprofen enantiomer derivatives with DBD-PZ (a) and DBD-COHz (b) by chiral stationaryphase HPLC. The mobile phases employed are H2O/MeOH (0/100) for DBD-PZ and H2O/CH3CN (40/60) for DBD-COHz. The other HPLC conditions are described in the Experimental Section.

Figure 6. Relationship between the concentration of the rat plasma flurbiprofen enantiomer derivatives with DBD-PZ and DBD-COHz.

stationary phase, acetone, which interacts with amide proton via hydrogen bonding in the stationary phase as suggested by the data from HPLC36 and 1H NMR studies,41,46 was added to both the mobile phases (H2O/MeOH and H2O/CH3CN). The concentration was set at 5% on the basis of their reports.46 Table 4 shows the comparison of R’s for separable derivatives on the chiral stationary phase between the absence and presence of acetone in the mobile phase. As a consequence, R values were scarcely changed in derivatives with DBD-PZ, indicating that a hydrogenbonding or a dipole-dipole interaction derived from the polar carbamoyl group in the stationary phase does not contribute to the interaction with the derivatives to separate them into their enantiomers. On the other hand, in the case of derivatives with DBD-COHz, the addition of acetone to the mobile phase led to the decrease of R values, especially in H2O/CH3CN (Table 4). These data indicate that hydrogen-bonding and dipole-dipole interactions occurred between an acid hydrazide of the derivatives and a polar carbamoyl group, which contributed to the formation of the appropriate diastereomeric complex for successful enantiomeric separation. The R values of 2-APA derivatives with DBD-ED were slightly changed by the addition of 5% acetone to the mobile phase, indicating that less interaction occurred between the derivatives 1798


with DBD-ED and the stationary phase, similar to that observed for the derivatives with DBD-PZ. Quantification of Flurbiprofen Enantiomers in Rat Plasma after Administration of Flurbiprofen Racemate. Based on the successful enantiomeric separation of 2-APAs with the procedure described above, the enantiomeric quantification of flurbiprofen, a representative of 2-APA, in rat plasma was carried out after intravenous administration of flurbiprofen racemate. Since there are many interfering substances in rat plasma, liquid/liquid extraction of the drug with toluene was required prior to derivatization with the reagent, DBD-PZ or DBD-COHz. Then, we combined the isolation step with reversed-phase HPLC of the derivatives, adopted in the basal experiment, for the first quantification of the total (R plus S)-flurbiprofen derivatives. Thus, the last chromatograms of the derivatives obtained from the chiral stationary-phase HPLC were used only to obtain the enantiomeric ratios of the derivatives. The quantitative values of each derivative, (R)- or (S)-flurbiprofen, were easily obtained by calculation based on the total (R plus S) amounts and the ratio (R/S) of the derivatives. Figure 4 shows the chromatograms of the derivatives with DBD-PZ (a and b) and DBD-COHz (c and d) obtained from the blank rat plasma and the rat plasma spiked with the flurbiprofen racemate (12.5 µg/mL) on the reversed-phase column. No interference peaks were observed for quantifying flurbiprofen racemate (elution time between 11.5 and 12.5 min for DBD-PZ and 39.0 and 41.0 min for DBD-COHz). The calibration curves obtained for both the derivatives with DBD-PZ and DBD-COHz gave good correlation coefficients (r2 > 0.997), and the coefficients of variation (CV%, n ) 3) were in the range of 1.61-4.64. The total flurbiprofen concentration quantified in rat plasma at 5 min, 1 h, and 5 h after intravenous administration (20 mg/kg of weight) were 99.7 ( 8.15, 75.1 ( 4.57, and 15.7 ( 2.4 µg/mL, respectively, and consistent with the values reported.43 Subsequently, the enantiomeric ratio of the flurbiprofen derivatives isolated was determined on the chiral stationary phase, and the typical chromatograms of flurbiprofen derivatives with DBDPZ (a) or DBD-COHz (b) in rat plasma are shown in Figure 5. The peaks of (R)- and (S)-flurbiprofen derived from the derivatives with DBD-PZ (a) and those with DBD-COHz (b) were consistent in terms of the respective elution times with those of the standard (48) Berry, B. W.; Jamili, F. Pharm. Res. 1988, 5, 123-125.

flurbiprofen enantiomer. Both enantiomeric ratios determined in rat plasma at 5 min, 1 h, and 5 h after administration of the flurbiprofen racemate were the identical. As shown in Figure 6, the regression analysis of these data showed an excellent agreement between the results obtained by both the derivatives (r2 ) 0.995), indicating that either reagent is applicable to the quantification of the enantiomers. Furthermore, a combined use of both the reagents is also used for the identification of the enantiomers. As to the sample volumes, 10 µL of plasma sample was sufficient for the quantification of flurbiprofen enantiomer until 5 h whereas 100-500 µL of plasma was required by the method published previously.43,48 In conclusion, the derivatives of 2-APAs with two different types of the fluorescent reagents, DBD-PZ and DBD-COHz, were enantiomerically well separated on the chiral stationary phase and detected sensitively. It was demonstrated that the reagents were

applicable to the enantiomeric quantification of 2-APAs with the concomitant use of the reversed-phase and the chiral stationaryphase HPLC. ACKNOWLEDGMENT We are grateful to Dr. Chang-Kee Lim for his kind advice and suggestions for preparing the manuscript. Thanks are also due to Daicel Co. Ltd., Yoshitomi Pharmaceutical Co. Ltd., Nagase Sangyo Co. Ltd., and Tosoh Co. Ltd., for their generous gift of Chiralcel OD-R, pranoprofen, (S)-ibuprofen, -ketoprofen, and -flurbiprofen, and TSK gel ODS- 80Ts, respectively. Received for review October 29, 1996. Accepted February 18, 1997.X AC961105G X

Abstract published in Advance ACS Abstracts, April 1, 1997.


1799

Enantiomeric Separation and Detection of 2-Arylpropionic Acids

Recommend Documents