Anal. Chem. 2006, 78, 920-927
Selective Determination of Native Fluorescent Bioamines through Precolumn Derivatization and Liquid Chromatography Using Intramolecular Fluorescence Resonance Energy Transfer Detection Makoto Yoshitake,† Hitoshi Nohta,*,† Hideyuki Yoshida,† Takashi Yoshitake,‡ Kenichiro Todoroki,† and Masatoshi Yamaguchi†
Faculty of Pharmaceutical Sciences, Fukuoka University, Nanakuma, Johnan, Fukuoka 814-0180, Japan, and Department of Neuroscience, Karolinska Institutet, S-171 77 Stockholm, Sweden
In this paper, we introduce a novel approach for the highly selective and sensitive analysis of native fluorescent bioamines (indoleamines and catecholamines). This method is based on intramolecular fluorescence resonance energy transfer (FRET) detection in a liquid chromatography (LC) system following precolumn derivatization of the bioamines’ amino groups. In this detection process, we monitored the FRET from the native fluorescent moieties (donor) to the derivatized fluorophore (acceptor). From a screening study involving 15 fluorescent reagents, we found that o-phthalaldehyde (OPA) generated the FRET most effectively. The OPA derivatives of the native fluorescent bioamines emitted OPA fluorescence (445 nm) through an intermolecular FRET process when they were excited at the excitation maximum wavelengths of the native fluorescent bioamines (280 nm). The generation of FRET was confirmed through comparison with the analysis of a nonfluorescent amine (isoleucine) performed using LC and a three-dimensional fluorescence detection system. We were able to separate the OPA derivatives of the indoleamines and catecholamines when performing LC on an ODS column. The detection limits (signal-to-noise ratio, 3) for the indoleamines and catecholamines, at a 20-µL injection volume, were 17-120 and 28-200 fmol, respectively. The sensitivity of the intramolecular FRET-forming derivatization method is higher than those of systems that take advantage of both native fluorescence detection (i.e., without derivatization) and the conventional detection of OPA derivatives. Furthermore, this method provides enough selectivity and sensitivity for the determination of the indoleamines present in the urine of healthy humans. Native fluorescent bioamines, indoleamines and catecholamines, play a number of significant roles in biological systems. * To whom correspondence should be addressed. Fax: +81-92-863-0389. E-mail:
[email protected]. † Fukuoka University. ‡ Karolinska Institutet.
920 Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
One of the indoleamines, 5-hydroxytryptamine (serotonin, 5-HT), is a well-known neurotransmitter that controls a wide variety of physiological functions, such as sleep, thermoregulation, sexual activity, and psychopathological states such as depression, anxiety, and drug reinforcement.1 The biological levels of catecholamines are related to various diseases, such as neuroblastoma, pheochromocytoma, and neurological disorders, and thus, they are useful analytical tools for use during diagnoses.2,3 Moreover, these levels are closely regulated by the concentrations of biosynthetic and metabolic enzymes; thus, the simultaneous determination of bioamines and their related compounds (precursors and metabolites) can provide a great deal of information to clinical and biochemical researchers. A number of methods have been reported for the determination of these monoamines.4-7 The widely used methods are based on liquid chromatography (LC),8-23 capillary electrophoresis24 or (1) Jacobs, B. L.; Azmitia, E. C. Physiol. Rev. 1992, 72, 165-229. (2) Nakagawa, A.; Ikeda, K.; Tsuneyoshi, M.; Daimaru, Y.; Enjoji, M. Cancer 1985, 55, 2794-2798. (3) Rosano, T. G.; Swift, T. A.; Hayes, L. W. Clin. Chem. 1991, 37, 1854-1867. (4) Lunn, G.; Hellwig, L. C. Handbook of Derivatization Reactions for HPLC; Wiley: New York, 1998. (5) Yamaguchi, M.; Ishida, J. In Modern Derivatization Methods for Separation Sciences; Toyo’oka, T., Ed.; Wiley: Chichester, 1999; pp 99-165. (6) Paul, M. M.; Haard, V. J. Chromatogr. Rev. 1988, 429, 59-94. (7) Kagedal, B.; Goldstein, D. S. J. Chromatogr. Rev. 1988, 429, 177-233. (8) Bizzarri, M.; Catizone, A.; Pompei, M.; Chiappini, L.; Curini, L.; Lagana, A. Biomed. Chromatogr. 1990, 4, 24-27. (9) Kele, M.; Ohmacht, R. J. Chromatogr., A 1996, 730, 59-62. (10) Krstulovic, A. M.; Powell, A. M. J. Chromatogr. 1979, 171, 345-356. (11) Jackman, G. P.; Carson, V. J.; Bobic, A.; Skews, H. J. Chromatogr. 1980, 182, 277-284. (12) Wielders, J. P. M.; Mink, J. K. J. Chromatogr. 1984, 310, 379-385. (13) Gironi, A.; Seghieri, G.; Niccolai, M.; Mammini, P. Clin. Chem. 1988, 34, 2504-2506. (14) Bearcroft, C. P.; Farthing, M. J. G.; Perrett, D. Biomed. Chromatogr. 1995, 9, 23-27. (15) Lakshmana, M. K.; Raju, T. R. Anal. Biochem. 1997, 246, 166-170. (16) Wood, A. T.; Hall, M. R. J. Chromatogr., B 2000, 744, 221-225. (17) Anderson, G. M.; Young, J. G.; Batter, D. K. J. Chromatogr. 1981, 223, 315-320. (18) Chou, P. P.; Jaynes, P. K. J. Chromatogr. 1985, 341, 167-171. (19) Seegal, R. F.; Brosch, K. O.; Bush, B. J. Chromatogr. 1986, 377, 131-144. (20) Odink, J.; Sandman, H.; Schreurs, W. H. P. J. Chromatogr. 1986, 377, 145154. 10.1021/ac051414j CCC: $33.50
© 2006 American Chemical Society Published on Web 12/30/2005
Figure 1. FRET-inducing derivatization of tryptophan and dopamine.
capillary electrochromatography25 coupled with UV detection,8,9 native fluorescence detection,10-16,24 and electrochemical detection.5,17-23 UV detection, however, is not very sensitive and selective for monoamines, and therefore, they require complicated cleanup procedures, such as liquid-liquid and solid-phase extractions. Although electrochemical detection is used most widely for high sensitivity and selectivity, it tends to lack reproducibilitys mainly because of hysteretic degradation of the electrode. Although native fluorescence detection provides reproducible results and requires simpler analytical systems, it is neither sensitive nor selective. Only a few fluorogenic derivatization methods have been reported using amine-labeling reagent, ophthalaldehyde (OPA),26,27 dansyl chloride (DNS-Cl),28 or fluorescamine.29 These methods, however, give fluorescence with all amino compunds, so they are not selective for indoleamines and catecholamines. On the other hand, highly selective LC methods based on specific fluorescence derivatizations have been developed using benzylamine30 for 5-hydroxyindoleamines and trihydroxyindole,31,32 ethylenediamine,33 and 1,2-diphenylethylenediamine34,35 for catecholamines. Although these approaches are sensitive and specific for their respective target amines, they cannot be applied to the simultaneous determination of a number of bioamines. (21) Jouve, J.; Martineau, J.; Mariotte, N.; Barthelemy, C.; Muh, J. P.; Lelord, G. J. Chromatogr. 1986, 378, 437-443. (22) Keating, J.; Dratcu, L.; Lader, M.; Sherwood, R. A. J. Chromatogr. 1993, 615, 237-242. (23) Patel, B. A.; Arundell, M.; Parker, K. H.; Yeoman, M. S.; O’Hare. D. J. Chromatogr., B 2005, 818, 269-276. (24) Park, Y. H.; Zhang, X.; Rubakhin, S. S.; Sweedler, J. V. Anal. Chem. 1999, 71, 4997-5002. (25) Paxon, T. L.; Powell, P. R.; Lee, H.-G.; Han, K.-A.; Ewing, A. G. Anal. Chem. 2005, 77, 5349-5355. (26) Maickel, R. P.; Miller, F. P. Anal. Chem. 1966, 38, 1937-1938. (27) Froehlich, P. M.; Cunningham, T. D. Anal. Chim. Acta 1978, 97, 357363. (28) Tsuchiya, H.; Tatsumi, M.; Takagi, N.; Koike, T.; Yamaguchi, H.; Hayashi, T. Anal. Biochem. 1986, 155, 28-33. (29) Imai, K. J. Chromatogr. 1975, 105, 135-140. (30) Ishida, J.; Iizuka, R.; Yamaguchi, M. Analyst 1993, 118, 165-169. (31) Okamoto, K.; Ishida, Y.; Asai, K. J. Chromatogr. 1978, 167, 205-217. (32) Yui, Y.; Itokawa, Y.; Kawai, C. Anal. Biochem. 1980, 108, 11-15. (33) Mori, K.; Imai, K. Anal. Biochem. 1985, 146, 283-286. (34) Nohta, H.; Mitsui, A.; Umegae, Y.; Ohkura, Y. Biomed. Chromatogr. 1987, 2, 9-12. (35) Panholzer, T. J.; Beyer, J.; Lichtwald, K. Clin. Chem. 1999, 45, 262-268.
Fluorescence resonance energy transfer (FRET) is a nonradiative transfer of excited-state energy from an initially excited donor fluorophore to an acceptor fluorophore. Therefore, FRET results in quenching of the donor fluorescence and emission of the acceptor fluorescence. FRET occurs under the conditions that two fluorophores (donor and acceptor) exist in proximity and in a suitable orientation and that the emission spectrum of the donor overlaps well with the excitation spectrum of the acceptor. Thus, FRET has become a powerful tools for the monitoring of structural dynamics of nucleic acids36,37 and proteins.38-40 Recently, an internal FRET-based fluorescent dye, BigDye, has been applied effectively to the challenge of DNA sequencing.41 To date, however, FRET has not been applied to the direct measurement of low-molecular-weight biosubstances. In the present study, we applied FRET to the highly selective and sensitive measurement of native fluorescent bioamines (indoleamines and catecholamines). This method is based on the FRET-inducing derivatization of these amines and the subsequent detection of FRET from the native fluorescent bioamines to the derivatized fluorophore (Figure 1). To find the most effective reagent, we screened 15 amino-group-selective compounds for their induction of FRET in indoleamines and catecholamines in comparison with that of the nonfluorescent amino compound, isoleucine (ILE). As a result, we found that OPA was the most effective agent for derivatizing both of these bioamines. In this paper, we validate the present method and compare it with previous methods. Furthermore, we have applied this method to the analysis of human urine without the need for any pretreatment procedure. (36) Sei-Iida, Y.; Koshimoto, H.; Kondo, S.; Tsuji, A. Nucleic Acids Res. 2000, 28, e59. (37) Stu ¨ hmeier, F.; Hillisch, A.; Clegg, R. M.; Diekmann, S. J. Mol. Biol. 2000, 302, 1081-1100. (38) Kinoshita, A.; Whelan, C. M.; Smith, C. J.; Mikhailenko, I.; Rebeck, G. W.; Strickland, D. K.; Hyman, B. T. J. Neurosci. 2001, 21, 8354-8361. (39) Grant, S. A.; Xu, J.; Bergeron, E. J.; Mroz, J. Biosens. Bioelectron. 2001, 16, 231-237. (40) Immink, R. G.; Gadella, T. W., Jr.; Ferrario, S.; Busscher, M.; Angenent, G. C. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 2416-2421. (41) Lee, L. G.; Spurgeon, S. L.; Heiner, C. R.; Benson, S. C.; Rosenblum, B. B.; Menchen, S. M.; Graham, R. J.; Constantinescu, A.; Upadhya, K. G.; Cassel, J. M. Nucleic Acids Res. 1997, 25, 2816-2822.
Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
921
Table 1. Reagents Screened for Reactivity toward Amino Groups reagent
abbreviation
λex (nm)
λem (nm)
o-phthalaldehyde (+ 2-mercaptoethanol) dansyl chloride 4,5-dimethoxyphthalaldehyde (+ 2-mercaptoethanol) 4-dimethylaminonaphthyl-1-isothiocyanate 6-[(7-amino-4-methylcoumarin-3-acetyl)amino]hexanoic acid succinimidyl ester N-[4-(6-dimethylamino-2-benzofuranyl)phenyl]isothiocyanate anthracene-2-isothiocyanate pyrene-4-butanoyl chloride 4-(5,6-dimethoxy-1,3-dihydro-1-oxo-2H-isoindol-2-yl)benzenesulfonyl chloride fluorescamine 3-chlorocarbonyl-6,7-dimethoxy-1-methyl-2(1H)-quinoxalinone naphthalene-2,3-dicarboxaldehyde (+ sodium cyanide) fluorescein-5-isothiocyanate tetramethylrhodamine-5-isothiocyanate sulforhodamine 101 acid chloride
OPA DNS-Cl DMPA DANITC AMCA-X,SE DBPITC AITC PBC DPS-Cl FA DMEQ-COCl NDA FITC TRITC SR101
335 350 330 330 353 348 358 345 315 381 400 460 495 540 587
445 522 447 460 442 425 453 375 385 470 480 490 520 572 620
EXPERIMENTAL SECTION Reagents and Solutions. Tryptamine (TA), 5-HT, 5-hydroxytryptophan (5-HTP), 5-methoxytryptamine (5-MT), and normetanephrine (NM) were obtained from Sigma (St. Louis, MO). 3,4Dihydroxyphenylalanine (DOPA), dopamine (DA), and norepinephrine (NE) were obtained from Wako Pure Chemical Co. (Osaka, Japan). Tryptophan (TRP), tyrosine (TYR), and ILE were purchased from Kishida Chemicals (Osaka, Japan). All fluorescence derivatization reagents listed in Table 1, except for 4,5dimethoxyphthalaldehyde (DMPA), were obtained from several commercial sources and used without further purification. DMPA was synthesized according to Bhattaacharjee’s method.42 All organic solvents were of LC grade. Other chemicals were of the highest purity available and used as received. Since sodium cyanide is a dangerous toxic compound, precautions must be taken to prevent any accidental release. Other reagents used to fluorescence derivatization and the organic solvents are toxic if exposed to lungs or skin and, therefore, should be carefully handled in accordance with the most current material safety data sheets. Distilled water, further purified using a Milli-QII system (Millipore, Milford, MA), was used to prepare all aqueous solutions. Stock solutions (10 mM) of the indoleamines, catecholamines, and ILE were prepared in aqueous methanol (50%, v/v) and stored in plastic vials at 4 °C. These solutions were stable for at least a month and were diluted further with methanol to the required concentrations prior to use. The 10 mM solution of OPA dissolved in methanol was usable for at least 3 days when stored at -20 °C. The 5 M 2-mercaptoethanol (2-ME) was prepared in 0.2 M borate buffer (pH 11.0) and stored at 4 °C. Urine Samples. Urine samples were obtained from healthy volunteers (n ) 11) in our laboratory. Aliquots of ∼10 mL from single-morning samples of volunteers were stored and frozen at -20 °C until analyzed. Prior to analysis, the urine was diluted 10-fold with methanol and then centrifuged at 1000g for 10 min at room temperature. A sample (100 µL) of the supernatant was subjected to derivatization. Fluorescence Derivatization. (1) Screening Tests. TRP and DA were used as model indoleamines and catecholamines, respectively. ILE was used for comparison with a nonfluorescent (42) Bhattacharjee, D.; Popp, F. D. J. Pharm. Sci. 1980, 69, 120-121.
922 Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
amine. The 15 fluorescent reagents selective for amines that are listed in Table 1 were used in screening for the degree of FRET induced upon derivatization. The amines were derivatized with the reagents according to the reported procedures.4,5,28,29,43-51 (2) Optimized OPA Reaction Conditions. The 10 mM OPA (100 µL) and 5 M 2-ME (100 µL) were added to an aliquot (100 µL) of a standard solution of amines or a urine sample placed in a 1.5-mL Reacti-vial (Pierce, Rockford, IL). The mixture was left to stand at room temperature for 10-15 min. After the addition of 1 M acetic acid (25 µL), a portion (20 µL) of the resulting mixture was injected into the LC system. As a blank, methanol (100 µL) was replaced for the standard solution and subjected to the same procedure. General LC System and Conditions. (1) LC System. The LC system consisted of a Jasco (Tokyo, Japan) PU-980 chromatograph pump, a Rheodyne (Cotati, CA) model 7125 syringe-loading sample injector equipped with a 20-µL sample loop, a Jasco DG980-50 on-line degasser, a Jasco LG-980-02 low-pressure gradient unit, a reversed-phase YMC-Pack ODS-AQ (250 × 4.6 mm i.d.; particle size 5 µm; YMC, Kyoto, Japan), and a Hitachi (Tokyo, Japan) L-7485 spectrofluorometer fitted with a 12-µL flow cell. The slit widths of the excitation and emission monochromators were set at 18 nm. Chromatograms were recorded on a Hitachi D-2500 integrator, and automatically integrated peak areas were used for the quantification. (2) Conditions for Screening Tests. The derivatized TRP, DA, and ILE were separated using acetonitrile-methanol-buffer [0.1 M sodium acetate buffer (pH 3.0-5.5) or 20 mM sodium phosphate buffer (pH 6.0-7.0)] as the mobile phase under (43) Titus, J. A.; Haugland, R.; Sharrow, S. O.; Segal, D. M. J. Immunol. Methods 1982, 50, 193-204. (44) Muramoto, K.; Kamiya, H.; Kawauchi, H. Anal. Biochem. 1984, 141, 446450. (45) Miyano, H.; Nakajima, T.; Imai, K. Biomed. Chromatogr. 1987, 2, 139144. (46) Cheng, Y. F.; Dovichi, N. J. Science 1988, 242, 562-564. (47) Kawasaki, T.; Higuchi, T.; Imai, K.; Wong, O. S. Anal. Biochem. 1989, 180, 279-285. (48) Inoue, H.; Kohashi, K.; Tsuruta, Y. Anal. Chim. Acta 1998, 365, 219-226. (49) Nohta, H.; Satozono, H.; Koiso, K.; Yoshida, H.; Ishida, J.; Yamaguchi, M. Anal. Chem. 2000, 72, 4199-4204. (50) Yoshida, H.; Ohno, Y.; Todoroki, K.; Nohta, H.; Yamaguchi, M. Anal. Sci. 2003, 19, 317-319. (51) Yoshida, H.; Harada, H.; Nakano, Y.; Nohta, H.; Ishida, J.; Yamaguchi, M. Biomed. Chromatogr. 2004, 18, 687-693.
isocratic elution conditions. The flow rate of mobile phase was set at 1.0 mL/min, and the column temperature was ambient (23 ( 3 °C). The fluorescence detector was operated at an excitation wavelength of 280 nm; the emission wavelength was set to their maximum fluorescence emission of the fluorescent reagent (Table 1). For comparative studies, the conventional (normal) fluorescence of each derivative was monitored at its fluorescence excitation and emission maxima (Table 1). (3) Conditions for the Detection of OPA Derivatives. The OPA-derivatized standard samples of indoleamines and catecholamines were separated using 10 mM acetic acid-methanolacetonitrile (9:6:5, v/v) and 25 mM acetic acid-acetonitrilemethanol (11:5:4, v/v), respectively, as the mobile phases. The flow rate of the mobile phase was set at 1.0 mL/min; the column temperature was ambient. The fluorescence detector was operated at excitation and emission wavelengths of 280 and 445 nm, respectively. For comparative studies, the conventional fluorescence of OPA derivatives was monitored at excitation and emission wavelengths of 340 and 445 nm, respectively. (4) Conditions for Urine Assay. After derivatization with OPA, the indoleamines present in urine were separated using a gradient elution system of (A) 100 mM acetate buffer (pH 4.5) and (B) methanol-acetonitrile (1:1, v/v). Gradient elution began at 40% B, was increased to 50% over 30 min, kept constant at 50% for 10 min, increased to 70% over 5 min, kept constant at 70% for 12 min, and then reduced to 40% over 1 min. The total analysis time was 75 min. Other LC conditions were the same as those used for the analysis of the standards. (5) Conditions for Native Fluorescence Detection. Prior to derivatization, the indoleamine and catecholamine standards were separated using 100 mM acetic acid-acetonitrile (7:1, v/v) and 100 mM acetic acid-methanol (99:1, v/v), respectively, as the mobile phases. The fluorescence detector was operated at excitation and emission wavelengths of 280 and 330 nm, respectively. Three-Dimensional Fluorescence Detection System. An Agilent (Palo Alto, CA) 1100 series LC system was used to confirm the generation of FRET. It consisted of a binary pump, an on-line degasser, an autosampler, a column oven, and a programmable 3-D spectrofluorometer equipped with an 8-µL flow cell. Other conditions, except for the detection conditions, were the same as those described in the previous section. On-line fluorescence spectra were obtained using a reaction mixture comprising TRP and ILE. The excitation spectra were monitored at 230-420 nm when the detector was operated with an emission wavelength of 445 nm; the emission spectra were monitored at 375-550 nm with excitation at 280 (for FRET) or 335 nm (for conventional fluorescence of OPA derivatives). The slit widths of both the monochromators were set at 20 nm. RESULTS AND DISCUSSION Screening Tests for FRET-Forming Derivatization Reagent. As Figure 1 indicates, the native fluorescent moieties of the bioamines and the derivatized fluorophore are in proximity, and thus, the derivatives satisfy the minimum requirement for the induction of FRET. The efficiency of FRET, however, is also affected by many other factors, such as the spectral overlap of the emission spectrum of the donor with the absorption spectrum of acceptor, the quantum yields, the orientation of the fluoro-
Table 2. Screening of Reagents: Comparative Study of FRET Induction between the Fluorescence Derivatives of Tryptophan and Isoleucine upon LC Using FRET and Conventional Fluorescence Detection Methods tryptophan (fluorescent)
selectivity factor
peak area peak area [(A)/(B)]/ (A)a (B)b (A)/(B) (C)a (D)b (C)/(D) [(C)/(D)]
reagent OPA DNS-Cl DMPA DANITC AMCA-X,SE DBPITC AITC PBC DPS-Cl FA DMEQ-COCl NDA FITC TRITC SR101 a
isoleucine (nonfluorescent)
129 147 2528 153 238 118 3735 152 81 57 72 2149 635 1038 863
76 88 960 128 1441 236 468 137 98 23 436 1233 2535 3915 4277
1.70 1.67 2.63 1.20 0.17 0.50 7.98 1.11 0.83 2.48 0.17 1.74 0.25 0.27 0.20
36 55 2033 39 192 36 3376 134 212 168 64 2808 1055 1073 1479
72 73 1424 72 710 85 310 125 313 55 361 1550 2790 4148 4570
0.50 0.75 1.43 0.54 0.27 0.42 10.89 1.07 0.68 3.05 0.18 1.81 0.38 0.26 0.32
3.40 2.23 1.84 2.22 0.63 1.19 0.73 1.04 1.22 0.81 0.94 0.96 0.66 1.04 0.63
FRET detection. b Conventional fluorescence detection.
Table 3. Screening of Reagents: Comparative Study of FRET Induction between the Fluorescence Derivatives of Dopamine and Isoleucine upon LC Using FRET and Conventional Fluorescence Detection Methods dopamine (fluorescent) reagent
selectivity factor
peak area peak area [(A)/(B)]/ (A)a (B)b (A)/(B) (C)a (Db) (C)/(D) [(C)/(D)]
OPA 60 52 DMPA 552 302 DANITC 25 13 DNS-Cl 42 58 AITC 3188 307 PBC 31 34 DPS-Cl 72 111 FA 126 58 NDA 2027 1003 a
isoleucine (nonfluorescent)
1.15 1.83 1.92 0.72 10.38 0.91 0.65 2.17 2.02
36 72 2033 1424 39 72 55 73 3376 310 134 125 212 313 168 55 2808 1550
0.50 1.42 0.54 0.75 10.89 1.07 0.68 3.05 1.81
2.30 1.29 3.56 0.96 0.95 0.85 0.96 0.71 1.12
FRET detection. b Conventional fluorescence detection.
phores, and the fluorescence lifetimes. Thus, it is practically impossible to predict the optimal fluorescent reagent for this purpose. Therefore, we screened 15 and 9 fluorescent reagents for indoleamines and catecholamines, respectively. In these screening tests, we used TRP and DA as a model indoleamine and catecholamine, respectively and, for comparison, used ILE as a nonfluorescent amine. Tables 2 (TRP) and 3 (DA) provide the fluorescence intensities (peak area) of the bioamines (A) and of ILE (C) in FRET detection and in conventional fluorescence detection [(B) and (D), respectively]. From the data, we propose two factors, A/B and (A/B)/ (C/D) as indices of the sensitivity and selectivity, respectively. The former is the ratio of peak areas in the FRET detection relative to that of the conventional detection system; increasing this value implies that FRET detection provides a more intense signal than does conventional detection. The latter is the ratio of the selectivity factor of the bioamines with respect to that of isoleucine; an Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
923
Figure 2. Chromatograms obtained using a standard mixture of indoleamines (20 pmol each on the column). (A) FRET and (B) conventional fluorescence detections monitored at excitation wavelengths of 280 and 335 nm, respectively; emission wavelength, 445 nm. Peaks: 1, 5-HTP; 2, TRP; 3, 5-HT; 4, ILE; 5, 5-MT; 6, TA; 7, unknown.
Figure 3. Chromatograms obtained using a standard mixture of catecholamines (20 pmol each on column). (A) FRET and (B) conventional fluorescence detections monitored at excitation wavelengths of 280 and 335 nm, respectively; emission wavelength, 445 nm. Peaks: 1, DOPA; 2, NE; 3, TYR; 4, NM; 5, DA; 6, ILE; 7, unknown.
increasing value suggests that the bioamine provides a more intense signal than does ILE in the FRET detection. We used the value (selectivity factor) as a primary criterion for screening. As is clear in Tables 2 and 3, OPA, DMPA, and DANITC were effective FRET derivatization reagents for both TRP and DA, whereas DNS-Cl was useful only for TRP. Although DANITC afforded greater selectivity, especially to DA, it caused so many artifact peaks in the chromatogram that it could not be applied to simultaneous determination of indoleamines and catecholamines. Thus, we selected OPA as the most suitable intramolecular FRETinducing derivatization reagent for native fluorescent indoleamines and catecholamines. It was surprise for us that FRET excitation gave more intense fluorescence than direct excitation of the donor in the OPA derivatives, which means A/B in Tables 2 and 3 was greater than 1 even though the C/D value was 0.5. The exact reason reamined unknown, but under the right conditions, FRET will excite the donor more effectively to afford intense fluorescence as in the 924
Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
case with BigDye.41 The OPA derivative of TRP, in comparison with that of DA, gave more intense fluorescence (FRET) and greater selectivity factor. The reason should be that the native fluorescence of TRP is 2.5 times more intense in the mobile phase, and its emission spectrum (maximum ∼340 nm) overlaps better with the excitation spectrum (maximum 335 nm) of OPAfluorophore than that of DA (maximum 310 nm). LC Separation. The best separations of the OPA derivatives of the indoleamines (TA, TRP, 5-HT, 5-HTP, 5-MT), catecholaminerelated compounds (TYR, DOPA, DA, NE, NM), and ILE were achieved on a YMC-Pack reversed-phase column when using acetonitrile-methanol-diluted acetic acid as the mobile phase under isocratic elution conditions. Figures 2 and 3 illustrate the typical chromatograms that we obtained from the standard mixtures of indoleamines and catecholamines, respectively. The fluorescence intensities of indoleamines and catecholamines, except for those of TYR and NM, that we monitored at excitation/ emission wavelengths (nm) of 280/445 (FRET detection) were
Figure 4. Three-dimensional fluorescence excitation chromatogram obtained using the OPA-labeled TRP and ILE samples (400 pmol each on the column) at an emission wavelength at 445 nm. Peaks: 1, TRP; 2, ILE; 3, unknown.
more intense than those monitored at 335/445 (conventional fluorescence detection), as is the case with TRP and DA in Tables 2 and 3. On the other hand, the fluorescence intensities of ILE and other non-natively fluorescent amines in the sample and the working environment monitored at FRET detection were weaker than those of the conventional fluorescence detection. Thus, this intramolecular FRET-inducing derivatization method permits the highly selective and sensitive determination of indoleamines and catecholamines in samples that also contain nonfluorescent amines. Three-Dimensional Fluorescence Chromatograms. Figure 4 presents the three-dimensional fluorescence excitation chromatogram of TRP and ILE recorded at an emission wavelength of 445 nm. The excitation maximum of TRP was more intense than that of ILE at an excitation wavelength at 280 nm because FRET occurred between the natively fluorescent moiety of TRP and the labeled OPA. Figure 5 illustrates the three-dimensional fluorescence emission chromatograms of TRP and ILE recorded at excitation wavelengths of 280 and 335 nm. The emission intensity of TRP
was more intense than that of ILE at an excitation wavelength of 280 nm, but they were equally intense at 335 nm. The shape of the emission spectrum of TRP obtained at an excitation wavelength of 280 nm was the same as that obtained at 335 nm. Thus, FRET between the natively fluorescent moiety of TRP and the labeled OPA should occur when TRP is derivatized with OPA and the derivative is irradiated at the wavelength of the native fluorescent moiety. The blank peaks (peaks 3 in Figure 5), which are due to environmental amines, also became smaller upon FRET detection, as was the case with ILE. Optimum Conditions for OPA Derivatization. We performed optimization studies of the OPA derivatization process to maximize the FRET fluorescence peak areas for indoleamines and catecholamines. We obtained maximum and constant peak areas within the concentration ranges 10-200 mM for OPA and 5-10 M for 2-ME; thus, for further experiments we adopted concentrations of 10 mM and 5 M, respectively. Among the solvents we examined for the derivatization reaction (dimethyl sulfoxide, tetrahydrofuran, acetone, acetonitrile, methanol, ethanol, 2-propanol, N,N-dimethylformamide), methanol gave the most intense peak in the chromatogram. Therefore, for subsequent experiments, we prepared the sample and OPA solutions in methanol. We examined the effect that the pH of the reaction mixture had on the fluorescence development by using borate buffer. The optimum value of pH for the fluorescence-derivatization reactions of both amines was pH 11.0; thus, 0.2 M borate buffer (pH 11.0) was selected for the preparation of 2-ME. The fluorescence reactions of the indoleamines and catecholamines with OPA proceeded fairly rapidly, even at 0 °C, but at temperatures higher than 50 °C, the peak areas decreased upon prolonging the reaction time. We obtained maximum peak areas when the reaction mixture was left to stand for 10-15 min at room temperature; thus, we performed subsequent experiments using a 10-15-min standing time at room temperature to obtain reproducible results. The resulting optimum derivatization conditions described in Experimental Section were almost identical to those in the screening tests.4,5 After the derivatization reaction, the reaction mixture had to be neutralized because the OPA-derivatized
Figure 5. Three-dimensional fluorescence emission chromatogram obtained using the OPA-labeled TRP and ILE samples at excitation wavelengths of (A) 280 and (B) 335 nm. Peaks and amounts of samples were the same as those listed in Figure 4.
Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
925
Figure 6. Chromatograms obtained using a sample of urine from a healthy human. (A) FRET and (B) conventional fluorescence detections monitored at excitation wavelengths of 280 and 335 nm, respectively; emission wavelength, 445 nm. Peaks and concentrations (nmol/mL of urine): 1, TRP, (191); 2, 5-HT, (1.40); 3, TA, (0.70). The peak areas were integrated by valley-to-baseline or valley-to-valley method.
indoleamines and catecholamines are unstable in alkaline media: their peak areas decreased to ∼50% of their maximum when the reaction mixture was stored for 30 min without neutralization. After neutralization, the OPA-derivatized indoleamines and catecholamines in the final reaction mixture were stable for at least 3 h in the dark at room temperature. Method Validation. (1) Specificity. Some nonfluorescent amines reacted with OPA under the present conditions to give the corresponding fluorescence derivatives. However, the compounds that we tested, amino acids and ammonia, had retention times that were different from those of the indoleamines and catecholamines, and thus, they did not interfere with the analyses. Moreover, the signals of these nonfluorescent amines decreased by ∼50% in comparison with the signals obtained under conventional detection. The following biological compounds did not afford any peaks under the present conditions at a concentration of 10 nmol/mL: R-keto acids (R-ketoglutaric acid, phenylpyruvic acid), other acids (acetic acid, palmitic acid, n-caprylic acid, n-capric acid, palmitic acid, stearic acid, oleic acid, homovanillic acid, 5-hydroxyindole3-acetic acid), sugars (D-glucose, D-fructose, D-galactose, D-ribose, N-acetyl-D-glucosamine, maltose, sucrose), nucleic acid bases (adenine, guanine, thymine, cytosine, uracil), and other compounds (methanol, acetone, phenol, cholesterol, urea). (2) Linearity. The relationships between the amounts of individual monoamines and their peak areas were linear over the concentration ranges 0.12-620 (5-HTP, 5-HT, 5-MT, TA), 0.12120 (DOPA, DA, NE, NM), and 0.36-620 pmol (TRP, TYR) per 20-µL injection volume; these ranges correspond to 0.02-100, 0.02-20, and 0.06-100 nmol/mL, respectively, of the samples in the solutions. The linear correlation coefficients (n ) 3) were >0.999 and >0.998 for all of the indoleamines and catecholamines, respectively, that we examined. (3) Precision. We established the within-day precision through repeated determinations of standard mixtures of the monoamines (3 nmol/mL). The relative standard deviations (n 926
Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
Table 4. Detection Limits of FRET, Conventional, and Native Fluorescence Detection Methods OPA derivatives
compd
FRETa (fmol)
conventional fluorescencea (fmol)
native fluorescencea (fmol)
TRP 5-HTP 5-HT 5-MT TA TYR DOPA DA NE NM
120 61 17 34 43 200 110 93 28 32
230 110 43 68 82 190 160 120 40 30
740 290 440 880 1400 2000 870 2300 1200 980
a Defined as the amount of sample per injection volume (20 µL) that gave a signal-to-noise ratio of 3.
) 7) did not exceed 5.9 and 6.3% for all of the indoleamines and catecholamines, respectively, that we examined. (4) Detection Limit. Table 4 presents the detection limits we obtained for the OPA derivatives detected using FRET wavelengths and conventional wavelengths and at the underivatized native fluorescent wavelengths of the indoleamines and catecholamines. The detection limits (per 20-µL injection volume; signalto-noise ratio, 3) of the natively fluorescent bioamines when using FRET derivatization were in the range 17-200 fmol; this range is 1-3 times lower than that obtained when using conventional fluorescence detection. Thus, we achieved the highly sensitive determination of indoleamines and catecholamines upon their derivatization with OPA and subsequent FRET detection. Determination of Indoleamines in Human Urine. To investigate the practicality of the FRET derivatization method for use in biological analysis, we applied our present method to the determination of the indoleamines in human urine. Figure 6 presents a typical chromatogram obtained using a 10-fold-diluted
urine sample. We changed the LC conditions from isocratic to gradient elution because we could not separate the OPA derivatives of the indoleamines from the blank components in urine. In Figure 6, peaks 1-3 correspond to TRP, 5-HT, and TA, respectively, on the basis of (a) their retention times when compared with those of the standard compounds and (b) coeluting the standard and the sample using various gradient patterns. The fluorescence intensities of three indoleamines monitored at an excitation wavelength of 280 nm (Figure 6A) were more intense than those monitored at 335 nm (Figure 6B). On the other hand, the other amines present in urine gave weaker peaks at an excitation wavelength of 280 nm than they did at 335 nm. Thus, the FRET detection method was quite selective and also provided a simple chromatogram for biological analysis. The determinable concentrations in urine were 0.6 nmol/mL (380 fmol on the column) for TRP and 0.2 nmol/mL (130 fmol on the column) for 5-HT and TA; these levels are lower than those found normally in human urine.19,52,53 The concentrations (mean ( SD) of TRP, 5-HT, and TA that we measured using the proposed method were 83 ( 58, 0.81 ( 0.49, and 0.59 ( 0.20 nmol/mL of urine from healthy subjects (n ) 11). These values are similar to those reported as being obtained when using UV,8 electrochemical,52 and native fluorescence53 detections. The intramolecular FRET-inducing derivatization method permits the highly selective and sensitive determination of indoleamines and can be applied to the determination of TRP, 5-HT, and (52) Mashige, F.; Matsushima, Y.; Miyata, C.; Yamada, R.; Kanazawa, H.; Sakuma, I.; Takai, N.; Shinozuka, N.; Ohkubo, A.; Nakahara, K. Biomed. Chromatogr. 1995, 9, 221-225. (53) Tsuchiya, H.; Ohtani, S.; Takagi, N.; Hayashi, T. Biomed. Chromatogr. 1989, 3, 157-160.
TA in human urine without the need to apply any special pretreatment procedures. CONCLUSION We have successfully applied a unique detection techniques one that takes advantage of intramolecular FRET-induced derivatizationsto the highly selective determination of indoleamines and catecholamines. The derivatized amines afforded FRET between their natively fluorescent moieties and the OPA-labeled moieties. We observed that FRET detection provided more-intense signals than did the conventional fluorescence detection and native fluorescence detection in the absence of derivatization. In fact, this method is sufficiently selective and sensitive to allow the simple assaying of urine from healthy humans. Therefore, we believe that this present method will be useful for biological and clinical investigations of natively fluorescent bioamines. ACKNOWLEDGMENT This study was supported in part by Grant-in-Aids for Scientific Research (B) (17390013) and (C) (12672100) from the Ministry of Education, Culture, Sports, Science, and Technology of Japan, and by a fund (046001) from the Central Research Institute of Fukuoka University. We especially thank Professor A. Takadate and Dr. T. Masuda (Daiichi College of Pharmaceutical Sciences, Fukuoka, Japan) for their generous support during the synthesis of DMPA. We are also grateful to Ms. N. Sejima for providing technical assistance. Received for review August 8, 2005. Accepted November 26, 2005. AC051414J
Analytical Chemistry, Vol. 78, No. 3, February 1, 2006
927
