Highly Selective Fluorometric Determination of Polyamines Based on

The detection limits (signal-to-noise ratio of 3) for the polyamines were 1 (Put), ..... of glutaric and 3-hydroxyglutaric acids in urine of glutaric ...
1 downloads 0 Views 131KB Size
Anal. Chem. 2000, 72, 4199-4204

Highly Selective Fluorometric Determination of Polyamines Based on Intramolecular Excimer-Forming Derivatization with a Pyrene-Labeling Reagent Hitoshi Nohta,†,‡ Hiroshi Satozono,† Katsumi Koiso,‡ Hideyuki Yoshida,‡ Junichi Ishida,‡ and Masatoshi Yamaguchi*,‡

Laboratory of Molecular Biophotonics, 5000 Hirakuchi, Hamakita, Shizuoka 434-8555, Japan, and Faculty of Pharmaceutical Sciences, Fukuoka University, 8-19-1 Nanakuma, Johnan, Fukuoka 814-0180, Japan

We introduce a novel approach in highly selective and sensitive fluorescence derivatization of polyamines. This method is based on an intramolecular excimer-forming fluorescence derivatization with a pyrene reagent, 4-(1pyrene)butyric acid N-hydroxysuccinimide ester (PSE), followed by reversed-phase high-performance liquid chromatography (HPLC). Polyamines, having two to four amino moieties in a molecule, were converted to the corresponding dipyrene- to tetrapyrene-labeled derivatives by reaction (100 °C, 20 min) with PSE. The derivatives afforded intramolecular excimer fluorescence (450-520 nm), which can clearly be discriminated from the monomer (normal) fluorescence (360-420 nm) emitted from PSE, its hydrolysate and monopyrene-labeled derivatives of monoamines. The structures of the derivatives were confirmed by HPLC with mass spectrometry, and the emission of excimer fluorescence could be proved by spectrofluorometry and time-resolved fluorometry. The PSE derivatives of four polyamines [putrescine (Put), cadaverine (Cad), spermidine (Spd), and spermine (Spm)] could be separated by reversed-phase HPLC on a C8 column with linear gradient elution. The detection limits (signal-to-noise ratio of 3) for the polyamines were 1 (Put), 1 (Cad), 5 (Spd), and 8 (Spm) fmol on the column. Furthermore, the present method was so selective that biogenic monoamines gave no peak in the chromatogram. Polyamines, putrescine (Put), spermidine (Spd), and spermine (Spm), occur in all eukaryotic cells and play important roles in cell growth and differentiation.1,2 Their intracellular concentrations are highly regulated by biosynthetic enzymes and increase at an early stage of proliferation. Because the concentrations in neoplastic cells were found to be higher than those in normal cells,3 the extracellular concentrations have attracted much attention * Corresponding author. E-mail: [email protected]. Fax: +81-92863-0389. † Laboratory of Molecular Biophotonics. ‡ Fukuoka University. (1) Tabor, C. W.; Tabor, H. Annu. Rev. Biochem. 1984, 53, 749-790. (2) Pegg, A. E.; McCann, P. P. Am. J. Physiol. 1982, 243, C212-C222. (3) Pegg, A. E. Cancer Res. 1988, 48, 759-774. 10.1021/ac0002588 CCC: $19.00 Published on Web 08/01/2000

© 2000 American Chemical Society

from clinical investigators such as tumor markers.4,5 Furthermore, many inhibitors against polyamine-synthetic enzymes have been investigated as anticancer drugs and/or preventites.6 For these studies, many analytical methods for polyamines have been reported on the basis of thin-layer chromatography,7,8 gas chromatography,9,10 or high-performance liquid chromatography (HPLC).11-21 Among them, HPLC methods employing fluorescence12-20 or mass spectrometric (MS)21 detection are relatively sensitive and selective, so they have been most widely applied to the assay of biological samples. Although the MS method is sensitive and highly reliable, its apparatus and operating costs are too expensive for routine analyses. On the other hand, the fluorescence detection is fairly simple, reproducible, and highly sensitive, but it requires a fluorescence derivatization procedure to convert nonfluorescent polyamines into the fluorescent derivatives. Thus, many fluorogenic reagents for the amino moiety have been utilized: 5-dimethylaminonaphthalene-1-sulfonyl chloride,12,13 o-phthalaldehyde,14,15 fluorescamine,16,17 9-fluorenylmethyl chloroformate,18 2-(1-pyrenyl)ethyl chloroformate,19 N-hydroxysuccin(4) Moulinoux, J.-P.; Quemener, V.; Delcros, J.-G.; Cipolla, B. In Polyamines in Cancer: Basic Mechanisms and Clinical Approaches; Nishioka, K., Ed.; Springer: New York, 1996; Chapter 10, pp 233-249. (5) Mitchell, M. F.; Tortolero, L. G.; Lee, J. J.; Hittelman, W. K.; Lotan, R.; Wharton, J. T.; Hong, W. K.; Nishioka, K. J. Cell Biochem. Suppl. 1997 28/29, 125-132. (6) Seiler, N.; Atanassovo, C. L.; Raul, F. Int. J. Oncol. 1998, 13, 993-1006. (7) Seiler, N.; Kno ¨dgen, B. J. Chromatogr. 1979, 164, 155-168. (8) Seiler, N. Methods Enzymol. 1983, 94, 3-25. (9) Smith, R. G.; Davies, G. D., Jr. Biomed. Mass Spectrom. 1977, 4, 146-150. (10) Niitsu, M.; Samejima, K.; Matsuzaki, S.; Hamana, K. J. Chromatogr. 1993, 641, 115-123. (11) Redmond, J. W.; Tseng, A. J. Chromatogr. 1979, 170, 479-481. (12) Stefanelli, C.; Carati, D.; Rossi, C. J. Chromatogr. A 1986, 375, 49-55. (13) Kabra, P. M.; Lee, H. K. J. Chromatogr. 1986, 380, 19-32. (14) Seiler, N.; Kno ¨dgen, B. J. Chromatogr. 1985, 339, 45-57. (15) van Eijk, H. M. H.; Rooyakkers, D. R.; Deuts, N. E. P. J. Chromatogr. 1996, 730, 115-120. (16) Kai, M.; Ogata, T.; Haraguchi, K.; Ohkura, Y. J. Chromatogr. 1979, 163, 151-169. (17) Hunter, K. J.; Fairlamb, A. H. Methods Mol. Biol. 1998, 79, 125-130. (18) Price, J. R.; Metz, P. A.; Veening, H. Chromatographia 1987, 24, 795-799. (19) Cichy, M. A.; Stegmeier, D. L.; Veening, H.; Becker, H.-D. J. Chromatogr. 1993, 613, 15-21. (20) Weiss, T.; Bernhardt, G.; Buschauer, A.; Jauch, K.-W.; Zirngibl, H. Anal. Biochem. 1997, 247, 294-304. (21) Feistner, G. J. Biol. Mass Spectrom. 1994, 23, 784-792.

Analytical Chemistry, Vol. 72, No. 17, September 1, 2000 4199

Figure 1. Intramolecular excimer-forming fluorescence derivatization of polyamines with PSE.

imidyl 6-quinolinyl carbamate,20 and so forth. In these methods, monoamino compounds in samples and the working environment are also fluorescence-derivatized to afford the fluorescent derivatives having the same fluorescence properties as those of the polyamines’ derivatives, which cause interfering peaks in the chromatogram. Therefore, in the assay of biological samples, the methods require highly sophisticated HPLC separation14,15,20 and/ or sample cleanup such as cation-exchange chromatography,16 liquid-liquid extraction,12,17 or solid-phase extraction13 are inevitable to eliminate the interferences from endogenous and contaminated monoamino compounds. Furthermore, two or more fluorophores in a molecule sometimes cause quenching of fluorescence (self-quenching).22-24 On the other hand, 1,3-di(1-pyrenyl)propane and other dipyrenylalkanes are known to form intramolecular excimer; one excited pyrene can form an excited-state complex ()intramolecular excimer) with the other ground-state pyrene in the molecule.24-26 The excimer emits longer wavelength (450-500 nm) fluorescence than pyrene monomer (normal) fluorescence (350-400 nm). In the present study, we introduce first the excimer fluorescence into the fluorescence derivatization using polyamines [Put, Spd, Spm, and naturally occurring aliphatic diamines, cadaverine (Cad)] as target molecules and have developed a highly selective and sensitive determination method for polyamines. This is based on an intramolecular excimer-forming derivatization with a pyrenelabeling reagent, 4-(1-pyrene)butyric acid N-hydroxysuccinimide ester (PSE), as shown in Figure 1, followed by excimer fluorescence detection in HPLC. Polyamines were converted to the (22) Baeyens, W. R. G. In Molecular Luminescencee Scectroscopy, Methods and Application: Part 1; Schulman, S. G., Ed.; John Wiley & Sons: New York, 1985; Chapter 2 (section 2.1), pp 30-34. (23) Ichinose, N. In Fluorimetric Analysis in Biomedical Chemistry; Ichinose, N., Schwedt, G., Schnepel, F. M., Adachi, K., Eds.; John Wiley & Sons: New York, 1987; Chapter 2 (section 5.4), pp 16-17. (24) Valeur, B. In Molecular Luminescence Spectroscopy, Methods and Application: Part 3; Schulman, S. G., Ed.; John Wiley & Sons: New York, 1993; Chapter 2 (sections 2.3.2.3, 2.3.2.4), pp 51-56. (25) Zachariasse, K. A.; Ku ¨ hnle, W. Z. Phys. Chem. (Wiesbaden) 1976, 101, 267276. (26) Duportail, G.; Lianos, P. In Vesicles; Rosoff, M., Ed.; Marcel Dekker: New York, 1996; Chapter 8, pp 295-372.

4200

Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

corresponding dipyrene- to tetrapyrene-labeled derivatives by reaction (100 °C, 20 min) with PSE. The derivatives afforded intramolecular excimer fluorescence, which can clearly be discriminated from the monomer fluorescence from PSE, its hydrolysate [4-(1-pyrene)butyric acid] and monopyrene-labeled monoamines. With detection in the excimer fluorescence region, highly simple chromatograms were obtained in HPLC by comparison with that detected in the monomer fluorescence region. The structures of the derivatives were confirmed by HPLC with mass spectrometry, and the emission of excimer fluorescence could be proved by spectrofluorometry and time-resolved fluorometry. The present HPLC methods are validated with respect to sensitivity, reproducibility, and selectivity. EXPERIMENTAL SECTION Chemicals and Solutions. Distilled water, further purified with a Milli-Q II system (Millipore, Milford, MA), was used for all aqueous solutions. Organic solvents used were of HPLC-grade. The polyamines (Put dihydrochloride, Cad dihydrochloride, Spd trihydrochloride, and Spm tetrahydrochloride) and monoamines (n-decylamine and di-n-pentylamine) were obtained from SigmaAldrich (St. Louis, MO). Stock solutions (10 mM) of the amines were prepared in a mixture of tetrahydrofuran (THF), dimethyl sulfoxide (DMSO), and water (1:2:1, v/v) and stored at -20 °C. The solutions were stable for at least 3 months and diluted further with the solvent mixture to the required concentrations before use. PSE was purchased from Molecular Probes Inc. (Eugene, OR) and used without further purification. The 5 mM solution of PSE prepared in acetonitrile was usable for at least a week when stored at -20 °C. Other reagents were of the highest purity available and were used as received. PSE 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. Fluorescence Derivatization. To 200 µL of a polyamine standard solution placed in a 1-mL Reacti-vial (Pierce, Rockford, IL) were added 5 mM PSE solution (200 µL) and 1 M potassium carbonate (in water, 10 µL). The vial was tightly sealed and heated at 100 °C for 20 min in a block heater, Pierce Reacti-Therm (Model

18970). After being cooling in ice water, the reaction mixture was injected into an HPLC. To prepare the reagent blank, 200 µL of a mixture of THF, DMSO, and water (1:2:1, v/v) in place of 200 µL of the polyamine standard solution was carried through the procedure. Fluorescence Spectral Characterization. Conventional Fluorometry. Fluorescence spectral measurements were performed with a Hitachi (Tokyo, Japan) F-3010 spectrofluorometer in 10 × 10 mm quartz cells; spectral bandwidths of 5 nm were used for both the excitation and emission monochromators. Time-Resolved Fluorometry. The fluorescence decay curves were measured with a laboratory-assembled time-correlated single photon counting system. A N2 high-pressure flash Lamp NFL111A (Horiba, Kyoto, Japan) was used as a light source, which was operated with a 2.5 ns full width at half-maximum. The light was monochromated by passing it though a Shimadzu (Kyoto, Japan) U-340 filter. A Jasco (Tokyo, Japan) CT-10 monochromator was used for spectroscopic analysis and a Y-46 color glass filter (Shimadzu) was set in front of the photomultiplier upon excimer measurement, which cut monomer fluorescence and scattering excitation light off completely. The detector for single photon counting was an R2949 photomultiplier (Hamamatsu Photonics, Hamamatsu, Japan). The fluorescence signals were acquired on a personal computer equipped with an SPC-300 single photon counting PC-module (Edinburgh Instruments, Scotland, UK) that includes a discriminator, A-D converter, time analyzer, and multichannel analyzer. Purification of Pyrene-Labeled Polyamines by HPLC. A standard mixture of the polyamines (10 nmol/mL each) was treated according to the derivatization procedure, and the reaction mixture was applied to the HPLC. The eluates from the column, which corresponded to the respective pyrene-labeled polyamines, were collected, respectively. The eluates were evaporated to dryness in vacuo at room temperature. The dried residues were reconstituted in an appropriate solvent, and the resulting solutions were used for fluorometry. HPLC System and Conditions. The gradient HPLC system consisted of a Jasco PU-980 pump, a Jasco DG-980-50 line degasser, a Jasco LG-980-02 low-pressure gradient unit, a Rheodyne (Cotati, CA) 7125 manual sample injector equipped with a 20-µL sample loop, a reversed-phase C8 column TSK-Gel SuperOctyl (100 × 4.6 mm i.d., particle size 2 µm; Tosoh, Tokyo), a Jasco FP-920 spectrofluorometer detector equipped with a 16-µL flow cell, and a Hitachi D-2500 integrator. Linear gradient elution was performed from aqueous 50% acetonitrile to 80% in 20 min at a flow rate of 1.0 mL/min and then back to the initial conditions in 1 min. The column was reconditioned with the initial mobile phase for 14 min. The mobile phases were filtered through a 0.45µm membrane filter (Advantec Toyo, Tokyo). The fluorescence detector operated at the excitation and emission wavelengths of 345 and 475 nm, respectively, and the slit widths of both the monochromators were set at 18 nm. For comparative studies, monomer fluorescence was monitored at the wavelengths of 345 and 375 nm, respectively. HPLC-MS System and Conditions. A Finnigan (San Jose, CA) LCQ, ion-trap mass spectrometer equipped with an electrospray ionization (ESI) interface was used in place of the fluorescence detector. Ion chromatograms and mass spectra were

Figure 2. Fluorescence emission spectra (excitation 345 nm) of the reaction mixtures: (1) 100 nmol/mL Put, (2) 0 nmol/mL Put, and (3) 100 nmol/mL n-decylamine were treated according to the derivatization procedure, and then the reaction mixtures were 1000-times diluted with aqueous 50% acetonitrile, followed by conventional fluorometry. Each spectrum was normalized to the first peak at 375 nm.

recorded on a personal computer installed with a Finnigan LCQ Navigator software version 1.2. Linear gradient elution was performed from aqueous 50% acetonitrile to 95% in 30 min at a flow rate of 1.0 mL/min, of which the gradient is the same as that in the fluorescence method in the first 20 min. Other system components and separation conditions were the same as those in the fluorescence method. The effluent from the HPLC column was directly introduced to the HPLC-MS interface without splitting. The ion source voltage and temperature of the heated capillary were set at +4.5 kV and 275 °C, respectively. Nitrogen gas was used as both the sheath gas (85 psi) and auxiliary gas (20 psi). The scan range was set at m/z 150-1500. RESULTS AND DISCUSSION Fluorescence from Intramolecular Excimer. Pyrene is known to form intermolecular excimer between two molecules in the vicinity. The formation of intermolecular excimer is highly dependent on the concentration, and it was generally observed only in higher concentrations (>1 mM).24 On the other hand, the compounds having two pyrene probes in a molecule can form intramolecular excimers independently of their concentrations. In the reaction mixture of the present derivatization, the PSE concentration is much higher than that of polyamines; thus, most of PSE remained unreacted or partially hydrolyzed to 4-(1-pyrene)butyric acid. Therefore, the intramolecular excimer fluorescence from dipyrene- to tetrapyrene-labeled polyamines could not be directly observed in the reaction mixture due to the highly intense background fluorescence of intermolecular excimer formed among PSE molecules. On the other hand, 1000-times or more dilution of the reaction mixture allowed one to observe the fluorescence from intramolecular excimer, where the PSE concentration is lower than 2.5 µM and the formation of intermolecular excimer could be neglected (Figure 2). The emission spectrum appeared as a structureless and broad peak around 475 nm, and this is in good agreement with those of dipyrenylalkanes.25,26 Figure 3 shows normalized fluorescent emission spectra of the pyrenelabeled polyamines purified by HPLC and that of monopyrenelabeled n-decylamine (monoamine) for comparison. The derivative Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

4201

Figure 3. Fluorescence emission spectra (excitation 345 nm) of the pyrene-labeled polyamines and n-decylamine; a standard solutions (10 nmol/mL) of (1) Put, (2) Cad, (3) Spd, (4) Spm, and (5) n-decylamine were treated according to the derivatization procedure, and the resulting derivatives were purified by HPLC (see Experimental Section). Each spectrum was normalized to the first peak at 375 nm. Table 1. Effect of Solvent on the Excimer Fluorescence Intensity of Dipyrene-Labeled Puta solvent

relative intensityb

solvent

relative intensityb

ethyl acetate dioxane DMSO DMF THF acetonitrile ethanol methanol aq 90% THF aq 90% acetonitrile

23 54 67 47 78 67 62 57 100 83

aq 90% ethanol aq 90% methanol aq 50% THF aq 50% acetonitrile aq 50% ethanol aq 50% methanol aq 10% THF aq 10% acetonitrile aq 10% ethanol aq 10% methanol

72 64 98 94 85 88 58 55 52 47

a Dipyrene-labeled Put, purified by HPLC, was dissolved in the respective solvent, and then the excimer fluorescence in the solution was measured at excitation and emission wavelengths of 345 and 475 nm, respectively. b The excimer fluorescence intensity obtained in an aqueous 90% THF solution was taken as 100.

of n-decylamine did not fluoresce in the excimer region as is the case with the reagent blank (Figure 2), but all the pyrene-labeled polyamines afforded eximer fluorescence at maximum wavelength of 475 nm. The ratio of fluorescence intensity in the excimer region (475 nm) to that in the monomer region (375 nm), dependent on the derivatives, might be closely related to the formation rate of intramolecular excimer. The factors affecting the ratio could be as follows:27 distance between pyrene moieties in a labeled molecule, flexibility of the chain, steric hindrance to the rotational motion of pyrene moiety, and so forth. The shape of the spectrum did not change when diluted 10 times in each derivative. Solvents also affected excimer fluorescence intensity (Table 1). Polar solvents, aqueous 50-90% solutions of watersoluble organic solvents, afforded intense excimer fluorescence, where pyrene moieties in a molecule will associate most rapidly to form the excimer. The excimer fluorescence intensities decreased in more polar solvents (10% organic solvents) probably (27) Snare, M. J.; Thistlethwaite, P. J.; Ghiggino, K. P. J. Am. Chem. Soc. 1983, 105, 3328-3332.

4202 Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

Figure 4. Decay curves of the excimer fluorescence from the dipyrene-labeled Put (1) and pyrene (2) and that of monomer fluorescence from pyrene (3). Each curve was normalized to the maximum intensity.

due to insolubility of the pyrene-labeled polyamines. All these results have strongly supported that each polyamine is labeled with two or more pyrene moieties in a molecule and the derivative affords intramolecular excimer fluorescence. Time-Resolved Fluorometry. Figure 4 shows the decay curves of pyrene and the pyrene-labeled Put, which are normalized to the respective maximum intensities. Although the monomer fluorescence from pyrene (curve 3) simply decreased almost exponentially with times, the excimer fluorescence from pyrene (curve 2) increased up to 20 ns (rising) and then decreased. This rising time is due to the time required for the association of two pyrene molecules (excited- and ground-state pyrenes) and the successive formation of intermolecular excimer. On the other hand, the excimer fluorescence from the pyrene-labeled Put (curve 1) showed a much shorter rising time. Therefore, the fluorescence from the pyrene-labeled Put is ascribable to the intramolecular excimer, which can be formed rapidly by the association between adjacent pyrene moieties in a molecule. HPLC Separation. Reversed-phase HPLC was investigated for the separation of the pyrene-labeled polyamines for the following reasons. (1) The derivatives afford intense excimer fluorescence in aqueous organic solvents, which is widely used as a mobile phase for reversed-phase HPLC. (2) The reaction mixture containing water can be directly injected to the reversedphase HPLC. (3) The derivatives are highly hydrophobic and nonionic compounds, which are generally separated by the reversed-phase HPLC. Of the tested columns (C18, C8, C4, C1, and phenyl columns), only the C8 column resulted in a satisfactory separation using aqueous acetonitrile as a mobile phase. However, gradient elution was inevitable because the pyrene-labeled polyamines greatly differed in hydrophobicity. A good separation of the derivatives was achieved with a linear gradient elution from aqueous 50% acetonitrile to 80% in 20 min. All the polyamines gave the respective single peaks, and they were separated well from each other and from the peaks of PSE and 4-(1-pyrene)butyric acid (Figure 5A). Any peak for polyamines could not be observed when the monomer fluorescence was monitored at 375 nm (Figure 5B). This is because the pyrene-labeled polyamines give weaker

Figure 5. Chromatograms obtained with a standard mixture of the polyamines. (A) Excimer and (B) monomer fluorescence intensities were monitored at emission wavelengths of 375 and 475 nm, respectively, with an excitation wavelength of 345 nm. A standard mixture of (1) 0.1 nmol/mL Put, (2) 0.1 nmol/mL Cad, (3) 0.5 nmol/ mL Spd, and (4) 1.0 nmol/mL Spm was treated according to the derivatization procedure, followed by HPLC. Peak 7 was due to PSE and peaks 8 were due to other artifacts.

Figure 7. Ion chromatograms obtained with the standard mixture of the polyamines: a standard mixture of the polyamines (200 nmol/ mL each) was treated according to the derivatization procedure. The monitored ions were the corresponding molecular ions (m/z ) [M + H]+ ( 0.5): (A) dipyrene-labeled Put, 629.3-630.3; (B) dipyrenelabeled Cad, 643.2-644.2; (C) tripyrene-labeled Spd, 956.2-957.2; (D) tetrapyrene-labeled Spm, 1283.3-1284.3. The retention times of the peaks were almost identical to those in Figure 5.

Figure 6. Chromatograms obtained with a standard mixture of n-decylamine (peak 5) and di-n-pentylamine (peak 6): a standard mixture of the amines (1.0 nmol/mL each) was treated according to the derivatization procedure, and other conditions are the same as those in Figure 5.

fluorescence in the monomer region as shown in Figure 3 and, on the contrary, monopyrene compounds [4-(1-pyrene)butyric acid, impurities contaminated in PSE and the PSE derivatives of environmental monoamines] in the derivatization mixture cause much larger peaks (peaks 8 in Figure 5B). On the other hand, monoamines (n-decylamine and di-n-pentylamine) did not give any peak in the excimer fluorescence detection, but they could afford the respective monomer fluorescence peaks (Figure 6). Thus, this method permits highly selective determination of polyamines in the samples containing monoamines and amino acids. In a previous HPLC method,19 a pyrene-labeling technique was attempted for the determination of polyamines. But the fluorescence detection was performed in the monomer fluorescence regions, and thus the selectivity was not very high as is often the case with other fluorogenic reagents (see the introduction). Structural Analysis by HPLC-MS. The structures of the pyrene-labeled polyamines were determined by HPLC-MS. The separation conditions were almost the same as those in the fluorometric detection method. The selected ion chromatograms (Figure 7) suggest that dipyrene-labeled derivatives were formed

from Put and Cad (MW, 628.8 and 642.7, respectively), tripyrenelabeled derivative from Spd (MW, 955.7), and tetrapyrene-labeled derivative from Spm (MW, 1282.8). Mass spectra for the peaks also provided the corresponding quasi-molecular ions ([M + H]+ or [M + K]+) as base peaks (Figure 8). When detected at the m/z ([M + H]+) corresponding to the monopyrene derivatives of Put, Cad, Spd, and Spm (MW, 359.5, 373.5, 416.6, and 473.7, respectively), dipyrene derivatives of Spd and Spm (MW, 689.6 and 744.0, respectively), and tripyrene derivative of Spm (MW, 1014.3), any significant peak was not observed in the respective ion chromatograms. From these observations, primary and secondary amino moieties in the polyamines were all derivatized with PSE under the present derivatization conditions (Figure 1). Optimum Derivatization Conditions. Optimization studies were carried out by HPLC to maximize the excimer fluorescence peak area for polyamines. The derivatization reaction proceeded effectively in the presence of organic solvents, methanol, ethanol, acetonitrile, DMSO, THF, DMF (N,N-dimethylformamide), or their mixtures since the pyrene derivatives of polyamines, especially Spd and Spm, are hardly soluble in water. The best results were obtained when acetonitrile was used for the preparation of PSE solution and THF-DMSO-water (1:2:1, v/v) for the sample solution. Some pyrene-labeling reagents for amines are commercially available: PSE, 1-pyrenesulfonyl chloride, and 1-pyreneisothiocyanate. In the preliminary studies using Put as a model polyamine, only PSE afforded excimer fluorescence, probably due to its long fluorescence lifetime (more than 100 ns28), (28) Haugland, R. P. Handbook of Fluorescent Probes and Research Chemicals, 6th ed.; Molecular Probes Inc.; Eugene, OR, 1996; p 36.

Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

4203

Figure 8. Mass spectra for the peaks of pyrene-labeled polyamines (A-D, see Figure 7).

enough to form an excimer. A PSE concentration of 3-7 mM in the reagent solution provided almost maximum peak areas for all the polyamines; a 5-mM PSE solution was used as the optimum. It is well-known that the derivatization reactions of amines proceed under weakly alkaline conditions where the amino moiety occurs in a nonionic form. Of the tested bases (pyridine, triethylamine, sodium acetate, sodium hydrogen carbonate, sodium carbonate, and potassium carbonate), potassium carbonate gave maximum peak areas at 0.8-5 M and did not cause any interfering peak in the chromatogram or any damage of the HPLC column; 1.0 M potassium carbonate was selected. The derivatization reaction proceeded more rapidly with increasing reaction temperature in the ranges 20-100 °C, but at higher than 110 °C the peak areas decreased with prolonging of the reaction time. Almost maximum and constant peak areas were obtained when the reaction mixture was heated at 100 °C for 15-30 min. Heating at 100 °C for 20 min was selected for obtaining reproducible results. The pyrenelabeled polyamines in the final reaction mixture were stable for at least 6 h in the dark at room temperature. Method Validation. Linearity. The relationships between the amounts of individual polyamines and the peak areas were linear over concentration ranges from 10 fmol to 100 pmol (Put and Cad) and from 25 fmol to 100 pmol (Spd and Spm) per 20-µL injection volume, which corresponded to 1-10 000 and 2.5-10 000 pmol/ mL in a sample solution, respectively. The linear correlation coefficients were more than 0.999 (n ) 3) for all the polyamines. Precision. The within-day precision was established by repeated determinations (n ) 8) using the mixture of polyamines (concentrations: Put and Cad, 100 pmol/mL; Spd and Spm, 1 nmol/ mL); their relative standard deviations were 3.0%, 1.5%, 2.0%, and 4.7%, respectively.

4204

Analytical Chemistry, Vol. 72, No. 17, September 1, 2000

Detection Limit and Quantitation limit. The detection limits (fmol per 20-µL injection volume, signal-to-noise ratio of 3) for Put, Cad, Spd, and Spm were 1, 1, 5, and 3, respectively, and the quantitation limits (fmol per 20-µL injection volume, signal-to-noise ratio of 10) were 4, 4, 15, and 25, respectively. These limits are 2-5 times lower than those obtained by the most sensitive HPLC methods using 2-(1-pyrenyl)ethyl chloroformate19 and N-hydroxysuccinimidyl 6-quinolinyl carbamate20 as labeling reagents. Specificity. The following biological compounds having only one amino moiety or none in a molecule, at a concentration of 10 nmol/mL, did not afford any peak under the present conditions; the compounds tested were neutral amino acids, acidic amino acids, ammonia, serotonin, catecholamines, R-keto acids (Rketoglutaric acid and phenylpyruvic acid), other acids (acetic acid, palmitic acid, oxalic acid, uric acid, homovanilic acid, vanylmandelic acid, 5-hydroxyindolacetic acid, and L-ascorbic acid), sugars (D-glucose, D-fructose, D-galactose, D-ribose, N-acetyl-D-galactosamine, maltose, and sucrose), nucleic acid bases (adenine, guanine, thymine, cytocine, and uracil), and other compounds (cholesterol, creatine, creatinine, and urea). The tested diamino compounds [basic amino acids (L-arginine, L-ornithine, L-lysine, and L-histidine), histamine and glycyl-L-lysine] also did not give any peak in the chromatogram. Most of them reacted with PSE to afford the corresponding dipyrene-labeled products, but they eluted rapidly from the HPLC column, overlapping with PSE or 4-(1-pyrene)butyric acid.29 CONCLUSIONS By the present derivatization technique, an intramolecular excimer-forming derivatization, polyamines were found to be converted to the respective polypyrene-labeled derivatives. They afforded intramolecular excimer fluorescence (450-520 nm), which can clearly be discriminated from normal fluorescence (360-420 nm) of PSE and other pyrene concomitants. This unique property as well as prolonged Stokes shift allowed highly sensitive and specific detection of polyamines. Therefore, the present method should be applicable to biological and biomedical investigations of polyamines. The intramolecular excimer-forming derivatization allows highly selective determinations of not only diamino compounds but also other bifunctional compounds.30 ACKNOWLEDGMENT This work was supported in part by Grants-in-Aid for Scientific Research (C) (106722033) from the Ministry of Education, Science, Sports and Culture of Japan.

Received for review March 2, 2000. Accepted June 11, 2000. AC0002588 (29) Dipyrene-labeled basic amino acids and histamine could be separated from PSE and 4-(1-pyrene)butyric acid on a C18 column with isocratic elution. The optimization studies for their determination are still under investigation. (30) Dicarboxylic, diphenolic (polyphenolic), and disulfhydryl compounds could be converted to di(poly)pyrene-labeled derivatives using commercially available reagents, and the details will be described in a subsequent publication.