JOURNAL OF BIOLUMINESCENCE A N D CHEMILUMINESCENCE VOL. 5 13-23

(1990)

Optimization of an HPLC Peroxyoxalate Chemiluminescence Detection System for Some Dansyl Amino Acids W. Baeyens’, J. Bruggeman’, C. Dewaele2, B. LinlV3and K. Imai4 ’Laboratory of Pharmaceutical Chemistry and Drug Analysis, Faculty of Pharmaceutical Sciences, State University of Ghent, Harelbekestraat 72, 8-9000 Ghent, Belgium ’Laboratory of Organic Chemistry, Faculty of Sciences, State University of Ghent, Krijgslaan 281-S4, B-9000 Ghent, Belgium 3(On leave from) The Laboratory of Instrumental Techniques, Faculty of Pharmacy, Complutense University of Madrid, 28040 Madrid, Spain 4Branch Hospital Pharmacy, University of Tokyo, 3-28-6 Mejirodai, Bunkyo-ku, Tokyo 112, Japan

Bis(2.4.6-trichlorophenyl) oxalate (TCP0)-hydrogen-peroxide-generated chemiluminescence (CL) of four dansyl amino acids has been used as a model system for the optimization of a detection system in reversed-phase high-performance liquid chromatography. Dansylated alanine, glutamic acid, methionine, and norleucine were subjected t o peroxyoxalate induced CL in a static system and in a flow system under various conditions with respect t o TCPO (ethyl acetate) and hydrogen peroxide (acetone) concentrations, solvent composition and flow, using a two-pump or a one-pump post-column reagent system. From the CL-decay curve, the influence on the emission signal from the total flow rate in the detector was investigated. Special attention was focused on the mixing of the LC eluate and the reagent in order t o combine an efficient collection of the emitted light using a 741.11flow cell (originally 10 pI in the fluorescence detector) with minimal extra column band broadening. Therefore, a capillary fused-silica tubing of about 100 pm i.d. was inserted against the end-frit of the column and brought through a mixing tee, in which the solutions of TCPO and hydrogen peroxide were added. The column end tubing ended in the flow cell and the LC eluate and the reagents were mixed when entering the flow-cell. Average detection limits (SIN= 2) of 200fmol injected dansylated amino acid could be reached. A comparison is made between the use of TCPO and DNPO (bis (2, 4-dinitrophenyl) oxalate). Keywords: Chemiluminescence; chromatography; dansyl amino acids; bis(2,4,6-trichlorophenyl)oxalate; TCPO; bis(2,4-dinitrophenyI)oxalate; DNPO;

peroxyoxalates INTRODUCTION

Chemiluminescence (CL) reactions have been used for the analysis of minute amounts of various compounds (Imai et al., 1986). Fluorimetry, a useful technique to measure low amounts of 0884-3996/90/010013-11$05.50 @ 1990 by John Wiley & Sons, Ltd

substances showing native fluorescence or fluorescing after chemical derivatization, is often connected to high-performance liquid chromatography (HPLC) systems to sensitively detect compounds in the column eluate. A mercury lamp, xenon arc lamp, or laser light source is then

14

used for the excitation of the fluorophores. In conventional fluorescence detection systems, the stray radiation, Raman light scattering, and/or variation of intensity of the excitation light ascribed to the light source interfere with the fluorescence intensity and lower the sensitivity of the system. In order to overcome these problems, chemical excitation has been used increasingly in the last decade to excite fluorescent compounds instead of a light source (Imai et al., 1985). In thin layer chromatography (TLC), however, the quantitation of a fluorescent compound via CL is a problem since CL light emission decays so quickly that it is difficult to adjust the timing of the measurement of the decaying light on the thin layer. In contrast, the timing of detection in HPLC is not such a problem since all conditions can be accurately and mechanically adjusted. Stieg and Nieman (1978) designed a modular flow cell and described experimental and theoretical considerations for its use in analytical chemiluminescence. CL detection in HPLC is gaining increasingly importance for the sensitive determination of fluorophores (Baeyens et al., 1987 a, b; Kobayashi and Imai, 1980). Peroxyoxalate CL detection, in particular, has become the method of choice for the measurement of catecholamines in urine (Kobayashi et al., 198l), in urine and plasma (Sekino et al., 1981) and for their determination using a newly developed chemiluminescence detector (Xie et al., 1987); for the determination of acetylcholine and choline using immobilized enzymes (Honda et al., 1986; Van Zoonen et al., 1987); for tertiary amines (Kwakman et al., 1987); polycyclic aromatic amines (Sigvardson et al., 1584); polycyclic aromatic hydrocarbons (Sigvardson and Birks, 1983); nitro-polycyclic aromatic hydrocarbons (Sigvardson et al., 1984); steroids (Koziol et al., 1984; Nozaki et al., 1588) and carboxylic acids (Grayeski and De Vasto, 1987). Applications to microbore chromatography (Grayeski and Weber, 1984) and to packed capillary columns (Weber and Grayeski, 1587) have been described by Grayeski’s group. De Jong et al., (1986) described the dependence of the half-life (TM) of the CL signal in post-column peroxyoxalate detection systems on the pH of the mobile phase and on the composition of the final solvent, obtained after mixing the mobile phase, hydrogen peroxide and oxalate.

W. BAEYENS ETAL.

Although other CL detection systems have been cited in the literature, for example the oxidation of alkylamines by benzoylperoxide (Burguera and Townshend, 1979); the cobaltcatalysed oxidation of luminol by hydrogen peroxide and the fluorescein-sensitized oxidation of sulphide by sodium hypochlorite (Burguera and Townshend, 1980); the CL reaction of morphine in the presence of permanganate in acidic tetraphosphate medium (Abbott et al., 1987); the detection of lipid hydroperoxides and hydrogen peroxide by an HPLC assay with isoluminol chemiluminescence (Yamamoto and Ames, 1987); amongst several others (see Grayeski, 1987). Increasing interest in peroxyoxalate based systems is seen in the work of Imai’s group from Tokyo in synthesizing novel aryl oxalate esters (Imai er al., 1986, 1987). The use of dansyl chloride (5-(dimethylamino)l-naphthalenesulphonyl chloride, Dns-C1) as a highly fluorescent labelling reagent for proteins, peptides, amino acids and imino acids has led to the development of HPLC methods for the quantitative separation of Dns-amino acids (Wilkinson, 1984; Rogerson, 1987). Singh and Hinze (1982), reported the remarkable micellar enhancement of fluorescence intensity obtained from dansyl glycine in the presence of quaternary ammonium compounds. Imai and co-workers have generated chemiluminescence from dansyl amino acids with bis(2,4,6-trichlorophenyl) oxalate (TCPO) (Kobayashi and Imai, 1980) and with bis(2,4dinitrophenyl) oxalate (DNPO) (Honda et al., 1983) and hydrogen peroxide providing femtomole detection limits. De Jong and co-workers have described a TCPO and a DNPO based CL detection system for the measurement of the dansyl derivative of a secondary amine drug in serum samples providing 1-10 pg detection limits (De Jong et al., 1984); they used their system for packed capillary liquid chromatography purposes (De Jong et al., 1987). The present paper reports on the TCPO- and DNPO-hydrogen-peroxide-generated CL of four dansyl amino acids as a model system for the optimization of a detection system in RP 18 HPLC. Special attention is focused on the device used for mixing the LC eluate with the reagents in order to combine an efficient collection of the emitted light using a 74-pl flow cell with minimal extra column band broadening.

HPLC OF DANSYL AMINO ACIDS

EXPERIMENTAL

15

HPLC apparatus

Chemicals

The flow diagram of the HPLC apparatus is shown in Fig. 1. The apparatus consisted of a De-ionized, re-distilled water was used through- Waters Model 510 HPLC pump, a Valco CV6U out. TCPO and DNPO were prepared by the 6-port sample injection valve, a sample loop of method of Mohan and Turro (1974), their varying volumes (10, 20, 50, lOOpl), a Waters reference compounds were purchased from Wako (Millipore) post-column reaction system consistChemicals (Neuss, FRG). Dansyl-1-alanine, ing of two single piston cam-driven pumps with dansyl-dl-glutamic acid, dansyl glycine, dansyl- pulse dampeners and a mixing device (De Jong et 1-lysine, dansyl-dl-methionine, dansyl-dl- al., 1987) for the mixing of the mobile phase and norleucine and dansyl-dl-phenylalanine were the reagents. The column consisted of a 250 x 4.6 purchased from Sigma Chemical Company (Mis- mm stainless steel tube packed with 10pm souri, USA) and stored below 0°C. Hydrogen Lichrosorb RP-18 (Alltech RSL, Eke, Belgium). peroxide (30% H 2 0 2 )and all other chemicals and Unless specified otherwise, chromatography was solvents were of analytical grade and were carried out isocratically at 23°C k 1 "C. Integraobtained from Merck (Darmstadt, FRG), UCB tion of chromatograms was performed using an (Leuven, Belgium), Aldrich Chemie (Brussels, integrating computer Model C-R3A ChromatoBelgium) or Janssen Chimica (Beerse, Belgium). pac, Shimadzu. The solvents were used after de-oxygenation with nitrogen followed by ultrasonic treatment. When required, the solutions were adjusted to Detection system their apparent p H values with hydrochloric acid, sodium hydroxide or ammonia, as measured with A Waters Fluorescence Detector Model 420-Ci a standard laboratory pH meter, and degassed 420-E with light source turned off was employed before use. and was equipped with a home-made cell holder

i

ELUENT

MIXING SYSTEM

Figure 1. Schematic flow diagram of the chromatographic set-up D, damper, F, flow cell, I, injector, P. pump, PM. photomultiplier, R. recorder

16

W. BAEYENS ET AL.

containing a 74-yl glass flow cell, the emission filter being removed from the light path. The cell was surrounded by aluminium foil in order to maximize collection of emitted CL light and thereby achieving a 2- to 3-fold sensitivity increase. Procedure

Static system. Aqueous dansyl amino acid solutions were mixed with TCPO and H 2 0 2 , both dissolved in organic solvents (ethyl acetate and acetone, respectively), and mixed vigorously by hand shaking followed by measurement. CL intensity and spectra were determined with a 3mm over-all slit system using an AmincoBowman spectrofluorimeter with the light source switched off (American Instrument Co. , Silver Spring, MD) and fitted with two grating monochromators, 1 x 1cm quartz sample cells, a photomultiplier R446 operated at 0.7 kV voltage and an X-Y recorder. It was noticed that both CL

column

RP-18

1 0 p

-

and fluorescence spectra were similar, maximum emission being at 520 nm. More reproducible results were obtained by directly injecting the TCPO-H202 mixture at the bottom of the cuvette containing the dansyl derivatives using a 2 ml syringe. The time course of relative CL intensity was recorded following the addition of the reagents and the 7712 value was determined from the obtained curves. All the experiments were performed at room temperature (23 "C k 1"C).

Flow system. Fig. 1 shows the instrumental set-up for CL measurement in the LC eluate. Stainless steel tubing (0.5mm i.d.) was used in all flow lines except for the mixing system, shown in Fig. 2. The eluent was the mixture 0.05 mol/l Tris-HC1 buffer, pH 7.7, and acetonitrile (2 + 1). Ethyl acetate and acetone were used as the solvents for TCPO and H 2 0 2 , repectively. The eluate from the column and the reagent solutions were mixed by passing through the mixing system which was connected to the flow cell. The mixing of the three solutions was promoted by bringing together first ethyl acetate containing TCPO and acetone containing H202followed by injection of the eluate into this premixed reagent. Comparison with classical set-ups for post-column reagent addition, i.e. by using a four-channel mixing cross, showed that the system described above provides a very efficient way of mixing the three solutions. Optimization of hydrogen peroxide concentration and flow. In order to obtain the optimum H 2 0 2 level at the mixing cross, various concentrations in acetone were tested at different flow values for a fixed TCPO flow of 0.6ml min-' at 5mmol/l in ethyl acetate.

II Figure 2. Detailed scheme for the mixing of the column eluate with the reagents required for the generation of chemiluminescence. A capillary of about 100pm i.d. is inserted against the end-frit of the column and brought through a Valco mixing cross by which the TCPO and hydrogen peroxide solutions are added. The glass 74.~1flow cell is also fixed to the mixing cross, the column end tubing ends into the flow cell and the reagents and the LC eluate are premixed in the first part of the flow cell (from De Jong eta/., 1987)

Optimization of TCPO concentration and flow. Different TCPO flows and concentrations were used for a fixed H202-flow of 0.6ml min-' at a 0.89 moVl concentration in acetone. Only two TCPO concentrations were investigated (5 mmoVl and 7mmol/l in ethyl acetate) as preliminary experiments had demonstrated that 4 mmol/l and lower concentrations were not suited for the generation of CL and, secondly it is not advisable to use higher concentrations than 7 mmol/l because of solubility problems. Such precautions must be taken to prevent TCPO precipitation between the mixing cross and the flow cell.

17

HPLC OF DANSYL AMINO ACIDS

RESULTS AND DISCUSSION CL in the static system Preliminary trials in the use of a one-reagent addition system for the generation of CL in the static system gave satisfactory results. However, the combination of TCPO and H 2 0 2(one reagent pump only) did not prove to be practical as analytical work had to be performed within half a day because of reagent decomposition. As a result, the two-reagent addition system was subsequently used throughout the work. Excellent reproducibility in retention times of the dansyl derivatives were recorded. Fig. 3 shows a typical chromatogram of four dansylated amino acids. Dansyl glycine coincides in the present system with dansyl alanine, dansyl lysine eluting very close after dansyl methionine. The peak of

dansyl phenylalanine partly coincides with that for dansyl norleucine; therefore, the latter amino acid derivatives were not used in further experiments.

Y

Solvent effects

With respect to CL intensity, reagent stability and solubility, ethyl acetate was selected as the solvent for TCPO and acetone as the solvent for H202.An eluent consisting of 0.05 mol/l Tris-HC1 (pH 7.7) and acetonitrile (2:l v/v) was chosen for separation of the dansyl amino acids. Replacement of the Tris-HC1 buffer by phosphate buffer did not provide better results. It is worth noticing that an increase of acetonitrile in the eluent leads to increased peak areas although the chromatographic resolution is then seriously affected.

Reagent concentrations

On consideration of the data shown in Figs 4 and 5 it was decided to use an H202-concentration of 0.89mol/l in acetone at a flow speed of 0.6ml min-’. Fig. 6 illustrates the results of the TCPO influence. A flow of 0.6-0.7ml min-’ in a concentration of 5 mmol/l in ethyl acetate was indicated for further work.

Flow-cell volume

-0

TIME lminl

Figure 3. Chromatographic separation of four dansyl amino acids detected by their peroxyoxalate-induced CL emission. A mixture of 5prnol of each dansyl amino acid was injected onto the column. Eluent: 0.05 rnol/l Tris-HCI buffer (pH 7.7) acetonitrile: 2 1, 0.6 rnl min-’. Reagents: 5 mrnol/l TCPO in ethyl acetate, 0.6rnl min- ’; 0.9mol/l H202in acetone, 0.6ml rnin GLU, dansyl glutamic acid; ALA, dansyl alanine; MET, dansyl rnethionine; NOR-LEU, dansyl norleucine. Cell volume: 49pl

+

~’.

A standard SuprasilR 10-y1 quartz flow cell was compared with home-made glass cells of 49 pl(1 = lOmm, i.d. = 2.5mm), 59yl (1 = 12mm, i.d. = 2.5mm) and 7 4 ~ 1(1 = 15mm, i.d. = 2.5mm). After considering sensitivity and band broadening effects, the 74-y1 cell proved to be most suited, not only because it provided a sensitivity increase of about 70% with respect to the 49-@ cell (which in turn yields about four times higher values than the standard cell), but also because of minimal band broadening effects. An average detection limit (SIN = 2) of 150fmol (alanine), 200fmol (norleucine) and 250fmol (glutamic acid and methionine) (amounts injected; 100 yl loop) could be attained. It should be mentioned that these values were obtained using both the commercially available and the purified TCPO .

W. BAEYENS ET AL.

18 PEAK AREA

200-

150

-

100

-

50

-

0

I 0;s

I 0.60

I

I

I

075

0.90

1.05

I 1.20

I 1.35

FLOW H*O*

1

1.50

[rnl mi"-']

Figure 4. Influence of H2OZ-concentrationand flow on the CL of dansyl amino acids. A mixture of 250prnol of each dansyl amino acid was injected onto the column. Eluent: 0.05rnolil Tris-HCI buffer (pH 7.7)-acetonitriIe: 2 1, 0.6rnl mtn- ' ; TCPO-reagent: 5 rnrnolil in ethyl acetate, 0.6 rnl m1n-l. Continuous line: (H202)= 0.53 rnolil. Dashed line: (H202)= 0 71 rnolil. Dotted line: (Hz02)= 0.89 molil

+

Reproducibility

Repeated injections of a standard solution of the dansylated amino acids showed that the reproducibility is about 4% relative standard deviation ( n = 6) for the injection of 1OOpg. The average detection limit is about 200 fmol. N o influence on sensitivity is observed because of the high peak dilution caused by the reagent flow. Reproducibility of peak areas determined with the classical mixing cross for adding both reagents postcolumn, even when inserting a highly porous mixing improving frit following the mixing cross, gave about 8% relative standard deviation ( n = 9). Enhancement of the CL signal

The influence of enhancers was investigated using

a second reagent pump for a combined TCPOH202reagent. a-Cyclodextrin and p-iodophenol provided an average increase of about 5% of the CL signal, Triton X-100 (polyethylene glycol p-isooctylphenyl ether) caused an overall increase of about 10%. Brij (polyethylene glycol fatty alcohol ether), benzalkonium chloride, cetrimonium bromide and sodium dodecyl suphate, on the other hand, caused a quenching of the CL emission. Different results were obtained when applying the two-reagent addition system for generating post-column CL and including the enhancer in the eluent. Triton X-100 caused distinct and quite reproducible enhancement of the CL signals; increases of about 70% and of 40% respectively could be noted from 10% and 5% (w/v) concentrations in the eluent. However, increasing Triton X-100 concentrations provide broader peaks in the chromatograms, i.e. spectroscopic

HPLC OF DANSYL AMINO ACIDS

19

PEAK AREA

150

-

100

-

50-

-

-GLU

0

I 0.53

~7

I 0.89

0.71

CONC

H202

IM.1

Figure 5. Influence of H20,-concentration on the CL of dansyl amino acids Parameters as in Fig 4 Each point on the curves represents an average value of seven observations taken at 5 mmol/l TCPO flows of 0 4, 0 6 and 0 8 rnl min (H,O,-flow of about 0 6 rnl min-')

'

PEAK AREA

-

/'/---/ , , / ---'-

I

0.3

I

I 0.6

MET.?

I

I

I

0.9

FLOW

TCPO

I 1.2

!mi rnin-'l

Figure 6. Influence of TCPO flow and concentration on the CL of dansyl amino acids A mixture of 250prnol of each dansyl amino acid was injected onto the column Eluent 0 05rnolil Tris-HCI buffer (pH 7 7)-acetonitrile 2 + 1. 0 6rnl min ', H,O,-reagent 0 89 rnolil in acetone, 0 6 rnl rnin-' Higher TCPO-concentrations cannot be used because of solubility limitations in ethyl acetate Continuous line (TCPO) = 5 mrnolil Dashed line (TCPO) = 7 mrnol/l

improvement is accompanied by a decrease of chromatographic efficiency. Moreover, dansyl hydroxide hydrolysis peaks, though weak, may not separate well from the parent Dnsderivatives. Sodium dodecyl sulphate at 0.01 mol/ 1 concentration in the eluent enhanced the CL of

dansyl methionine and of dansyl norleucine for about 10% with respect to chromatograms from the untreated eluent. p-Iodophenol could not be used as including this reagent in the eluent caused phase-demixing; P-cyclodextrin caused a precipitate to form in the flow cell (ethyl acetate

20

W. BAEYENS ETAL. PEAK AREA

0'

I 04

I 05

I 06

I 07

I 08

ELUENT FLOW (ml m i n - ' )

Figure 7. Influence of eluent flow on peak areas of DNPO-H202 induced CL from dansyl arnlno aclds A mixture of 250 pmol of each compound was injected Eluent 0 05 mol/l Tris-HCI buffer (pH 7 7)-aceton1trde 2 + 1, H,O,-reagent 0 89 mol/l In acetone, 0 6rnl min DNPO reagent 0 01 mol/l in acetonitrile, 0 6ml rntn

',

influence) while benzalkonium chloride and cetrimonium bromide caused retention of all derivatives on the column. Further work on the enhancement of CL signals is in progress in this laboratory. DNPO experiments Similar experiments were performed using DNPO-H202-induced CL from the dansyl amino acids. As DNPO is much less stable in the final solvent (eluent plus both reagents solutions) than TCPO and as DNPO shows a shorter half-life (De Jong et al., 1986), the influence of eluent flow on peak areas was investigated (Fig. 7). Flows of 0.4 ml min-' provided intense signals. A DNPO concentration of 10mmol/l in acetonitrile provided highest emission signals. An eluent flow of 0.4ml min-' was chosen for further DNPO experiments. Lower flow rates were not tested as they could not provide suitable chromatographic detection. Eluents with varying acetonitrile and buffer contents were tried so as to optimize the chromatographic detection limits, as both constituents are known to influence the CL decay curve (De Jong et al., 1986). As in the case of TCPO experiments, an increased acetonitrile content leads to a distinct increase of peak areas together with lower retention times; however, acetonitrile concentrations higher than 50% do

'

not allow chromatographic separation of the dansyl derivatives. Similar observations have been made in this laboratory when dealing with fluorimetric analysis of various fluorophores. An eluent composition of Tris-HC1 buffer (0.05 mol/l pH 7.7) and acetonitrile: 5 3 at a flow rate of 0.3 ml min-' , provided suitable chromatograms. Fig. 8 illustrates the influence of the peroxide flow. An increased flow caused increased signals, but higher peroxide flows could not be attempted because of capacity limitations of the reagent pump. The influence of hydrogen peroxide concentration (in acetone) was checked for a fixed 1.5ml min-' flow and Fig. 9 shows the results. A concentration of about 0.9mol/l provides best signal response.

+

TCPO - DNPO comparison DNPO did not provide the limits of detection offered by TCPO (about five times less sensitive). Moreover, the DNPO reagent, which has skinstaining properties, proves to be less stable and causes higher background CL owing to dinitrophenol formation during the peroxide reaction. Trichlorophenol production from TCPO caused less interference. According to the literature (De Jong et al., 1986), DNPO or TCPO can both be used for conventional liquid chromatography with a relatively large flow-cell. DNPO is better suited

HPLC OF DANSYL AMINO ACIDS

21

PEAK AREA

150-

100

-

50

-

0'

I 03

I

I 09

06

1

I

12

15

FLOW H202 (ml rnin-')

Figure 8. Influence of H202-flow(0 89 molil in acetone) on the (DNPO) CL of dansyl amino acids A mixture of 250 pmol of each dansyl amino acid was injected onto the column Eluent 0 05 mol/l Tris-HCI buffer (pH 7 7)-acetonitrile 5 3, 0 3 ml min-', DNPO reagent 0 01 molil in acetonitrile, 0 6 ml min-'

+

PEAK AREA

zoo

/-\

150

-

MET

-

NOR-LEU

100-

50

0

I

l

I

I

I

C53

071

089

107

125

CONC

HzOz ( M

)

Figure 9. Influence of H,02-concentration on the (DNPO) CL of dansyl amino acids Peroxide flow 1 5ml m i n - ' Other parameters as in Fig 8

for use in miniaturized systems because of its shorter half-life (De Jong et al., 1987) giving less band broadening. CONCLUSION Dansylated amino acids separated by reversed-

phase HPLC using aqueous pH 7.7 Tris-HCI acetonitrile mixtures can be rapidly determined at the sub-picomole level by CL detection after applying the TCPO-hydrogen peroxide system for chemical excitation of the fluorophores. The use of a special post-column mixing device allows preliminary mixing of TCPO in ethyl acetate with

22

W. BAEYENS ETAL.

the hydrogen peroxide in acetone followed by injection of the eluate into the premixed reagent. For each specific amino acid sample to be measured it is advisable, though, to optimize eluent composition and flow together with the concentration and flows of the peroxyoxalate reagents as these factors influence CL intensity and decay. The present instrumental set-up opens perspectives for the HPLC determination of fluorescing compounds, susceptible to peroxyoxalate CL induction. Acknowledgement

Financial support from the National Fund for Scientific Research (Belgium) (N.F.W. 0.-F.N. R.S.) (S21.5LV.E6,’87-’88) is greatly acknowledged.

REFERENCES Abbott, R . W . , Townshend. A . and Gill, R. (1987). Determination of morphine in body fluids by highperformance liquid chromatography with chemiluminescence detection. Analyst, 112, 397-406. Baeyens, W.. De Keukeleire. D., Van Der Wcken, G . and De Moerloose, P. (1987a). Luminescentietechnieken voor sensitieve en selectieve analyses. Chemie Magazine (BeIgium), 13, 17-21 (in Dutch). Baeyens, W . , D e Keukeleire, D . , Van Der Weken, G . and De Moerloose. P. (1987b). Techniques de luminescence pour les analyses sensibles et selectives. Chimie Magazine (Belgium). 6 , 15-19 (in French). Burguera, J . L. and Townshend, A . (1979). Determination of amines by using a chemiluminescent reaction in solution. Tulanta, 26. 795-798. Burguera, J . L. and Townshend, A . (1980). Flow injection analysis for monitoring chemiluminescent reactions. Anal. Chim. Aria, 114, 209-214. De Jong, G . J., Lammers, N., Spruit, F. J . , Brinkman, U.A.Th. and Frei, R. W. (1984). Optimization of a peroxyoxalate chemiluminescence detection system for the liquid chromatographic determination of fluorescent compounds. Chromatographia. 18, 129-133. De Jong, G . J . , Lammers, N . , Spruit, F. J., Frei, R . W. and Brinkman, U.A.Th. (1986). Analytical implications of the half-life of the chemiluminescence signal in the peroxyoxalate detection system for liquid chromatography. J . Chromatogr., 353, 249-257. De Jong, G . J . , Lammers, N.. Spruit, F. J., Dewaele, C. and Verzele, M. (1987). Low-dispersion chemiluminescence detection for packed capillary liquid chromatography. Anal. Chem., 59. 1458-1461. Grayeski, M . L. (1987). Chemiluminescence analysis. Anal. Chem., 59, 1243A-12436. Grayeski, M. L. and De Vasto, J. K. (1987). Coumarin derivatizing agents for carboxylic acid detection using

peroxyoxalate chemiluminescence with liquid chromatography. Anal. Chem., 59, 1203-1206. Grayeski, M. L. and Weber, A J. (1984). Investigation of instrument parameters for the application of peroxyoxalate chemiluminescence detection to microbore chromatography. Anal. Lett., 17, 1539-1552. Honda, K . , Sekino, J. and Imai. K. (1983). Bis(2,4dinitropheny1)oxalate as a chemiluminescence reagent in determination of fluorescent compounds by flow injection analysis. Anal. Chem., 55, 940-943. Honda, K., Miyaguchi, K. and Imai, K. (1985). Evaluation of fluorescent compounds for peroxyoxalate chemiluminescence detection. Anal. Chim. Acta, 177, 111-120. Honda, K., Miyaguchi, K.. Nishino. H . , Tanaka, H . , Yao, T. and Imai, K. (1986). High-performance liquid chromatography followed by peroxyoxalate chemiluminescence detection of acetylcholine and choline utilizing immobilized enzymes. Anal. Biochem., 153, 50-53. Imai, K., Miyaguchi, K. and Honda, K. (1985). Highperformance liquid chromatography+hemiluminescence reaction detection system of fluorescent compounds using bis(2,4,6-trichlorophenyl)oxaiate (TCPO) and hydrogen peroxide. In Bioluminescence and Chemiliimiriescerice: Instrumental Applications, Van Dyke, K. (Ed.), CRC Press, Boca Raton, Florida, pp. 65-76. Imai, K., Nawa, H . , Tanaka, M. and Ogata. H . (1986). Novel aryl oxalate esters for peroxyoxalate chemiluminesccncc reactions. Analyst, 1 1 1 , 209-21 1. Imai, K., Matsunaga, Y . , Tsukamoto, Y . and Nishitani. A . (1987). Application of bis[J-nitro-2-(3.6,9-trioxadecyloxycarbonyl)-phcnyl]oxalateto post-column chemiluminescence detection in high-performance liquid chromatography. J . Chromatogr., 400. 169-176. Kobayashi. S. and Imai, K. (1980). Determination of fluorescent compounds by high performance liquid chromatography with chemiluminescence detection. AnuI. Chem., 52, 424-427. Kobayashi, S . , Sekino, J . , Honda. K. and Imai, K. (1981). Application of high-performance liquid chromatography with a chemiluminescence detection system to determine catecholamines in urine. Anal. Biochem., 122, 99-104. Koziol, T . , Grayeski, M . L. and Wcinberger, R . (1984). Determination of trace levels of steroids in blood plasma by liquid chromatography with peroxyoxalate chemiluminescence detection. J . Chromatogr., 317. 355-366. Kwakman, P. J . M . , Brinkman, U. A . Th., Frei, R. W . , De Jong, G . J . , Spruit, F. J . , Lammers, N. G . F. M. and Van Den Berg, J. H. M . (1987). A post-column extraction system for the determination of tertiary amines by liquid chromatography with chemiluminescence detection. Chromatographia, 24, 395-399. Mohan, A. G . and Turro, N . J . (1974). A facile and effective chemiluminescence demonstration experiment. J . Chem. Educ., 51, 528-529. Nozaki, 0.. Ohba, Y. and Imai, K. (1988). Determination of serum estradiol by normal-phase high-performance liquid Chromatography with peroxyoxalate chemiluminescence detection. Anal. Chim. Acta, 205, 255-260. Rogerson, G. (1987). Amino acid analysis by HPLC with precolumn derivatisation. International Analysl, 4 , 19-26. Sekino, J . , Honda, K. and Imai, K. (1981). Development and use of a high-performance liquid chromatography (HPLC)

HPLC OF DANSYL AMINO ACIDS detection system using chemiluminescence. Determination of catecholamines in urine and plasma. Rinsho Kagaku Shirnpojurnu, 21, 17-19. Sigvardson, K. W. and Birks, J . W. (1983). Peroxyoxalate chemiluminescence detection of polycyclic aromatic hydrocarbons in liquid chromatography. Anal. Chern., 55, 432-435. Sigvardson, K. W., Kennish, J . M . and Birks, J. W. (1984). Peroxyoxalate chemiluminescence detection of polycyclic aromatic amines in liquid chromatography. Anal. Chem., 56, 1096-1102. Sigvardson, K . W . and Birks, J . W . (1984). Detection of nitro-polycyclic aromatic hydrocarbons in liquid chromatography by zinc reduction and peroxyoxalate chemiluminescence. J . Chrornarogr., 316, 507-518. Singh, H. N. and Hinze, W . L. (1982). Micellar enhanced fluorimetric determination of 1-N,N-dimethylaminonaphthalene-5-sulphonylchloride and o-phthalaldehyde-2-mercaptoethanol derivatives of amino acids. Analyst, 107, 1073-1080. Stieg, S. and Nieman, T. A. (1978). Experimental and theoretical considerations of flow cell design in analytical

23 chemiluminescence. Anal. Chern., 50, 401-404. Van Zoonen, P., Gooijer, C . , Velthorst, N. H . and Frei, R. W . (1987). HPLC detection of choline and acetylcholine in serum and urine by an immobilized enzyme reactor followed by chemiluminescence detection. J . Pharrn. Biorned. Anal., 5 , 485-492. Weber, A. J . and Grayeski, M. L. (1987). Peroxyoxalate chemiluminescence detection with capillary liquid chromatography. Anal. Chern., 59, 1452-1457. Wilkinson, J . M. (1984). Dansyl amino acids. In CRC Handbook of HPLC for the Separation of Amino Acids, Peptides, and Proteins, Vol. 1, Hancock, W . S. (Ed.), C R C Press, Boca Raton, Florida. Xie, G., Dai, J., Kaiho, I., Tachi, R. and Kawai, M. (1987). Determination of catecholamines by high performance liquid chromatography with a newly developed chemiluminescence detector. In Proc. 2nd Betjing Conf. and Exhib. on Instrum. Published in Analysis, 1987, 931. Yamamoto, Y. and Ames, B. N. (1987). Detection o f lipid hydroperoxides and hydrogen peroxide at picomole levels by an HPLC and isoluminol chemiluminescence assay. Free Radical Biol. M e d . , 3, 359-361.