Analyst, January 1996, Vol. 121 (49-54) 49 Sensitive Peroxyoxalate Chem il uminescence Determination of Psychotropic lndole Derivatives Juana Cepas, Manuel Silva and Dolores Pkrez-Bendito Department of Analytical Chemistry, Faculty of Sciences, University of Chrdoba, E-14004 Cbrdoba, Spain The continuous-addition-of-reagent (CAR) technique was used for the determination of hallucinogenic alkaloids such as N-methyltryptamine, 5-methyltryptamine, a-methyltryptamine, bufotenine and psilocin by measuring the peroxyoxalate chemiluminescence produced on reaction with the bis(2,4,6-trichlorophenyl)oxalate- hydrogen peroxide system following derivatization with dansyl chloride. This is the first reported use of chemiluminescence for the determination of these compounds. The special ability of the CAR technique to suppress background emission in peroxyoxalate chemiluminescence reactions provides a highly sensitive means of determination (the response to indole alkaloids is linear over two orders of magnitude and detection limits are in the picomole range).The proposed method compares favourably with existing alternatives in terms of sensitivity and was validated in the determination of psilocin in mushroom samples. Key words: Chemiluminescence; peroxyoxalate, continuous addition; indole derivatives Introduction Hallucinogens make up a heterogeneous group of compounds primarily resulting in altered cognitive and perceptual states that cannot otherwise be experienced. 192 Psychotropic indole deriva- tives with a chemical structure derived from tryptamine are a major group of hallucinogenic alkaloids of considerable biological interest as they block the neural uptake of noradre- naline and have complex effects on reflex activity in the spinal cord.3 The psychogenic NJV-dimethylated tryptamines are of special interest as they have been shown to occur in urine from schizophrenic patients and autistic ~hildren.~.~ Several thin- layer and gas chromatographic methods for the separation and identification of these tryptamines have been rep0rted.~.6>7 Also, psilocin (5-hydroxy-NJV-dimethyltryptamine) and its ana- logues have been widely determined by high-performance liquid chromatography (HPLC), especially in mushroom sam- ples of the genus Psilocybe using UV (254 nm),8 variable- wavelength UV,9JO photodiode-array,ll fluorimetric8JO and voltammetric8.12 detection.These methods are not very sensi- tive; their limits of detection are typically several micrograms per millilitre, except for electroanalytical detection, which affords sub-microgram per millilitre concentrations. More sensitive methods are therefore required for the determination of these psychotropic compounds. This paper reports a simple, sensitive method for the determination of psychotropic indole derivatives following derivatization with dansyl chloride; the method uses the continuous-addition-of-reagent (CAR) techniquel3,14 and per- oxyoxalate chemiluminescence (PO-CL) detection. The dansy- lated derivatives exhibit a high CL intensity on reaction with the bis(2,4,6-trichlorophenyl)oxalate (TCP0)-hydrogen peroxide system.The CAR technique was used owing to its ability to suppress background emission15 on account of the special way in which the sample and reagents are mixed (the TCPO solution is added to the reaction vessel containing hydrogen peroxide and the dansylated derivative). Under these conditions, high hydrogen peroxide to TCPO concentration ratios can be used and background emission can thus be suppressed. The proposed method is the first attempt at the determination of these psychotropic indole derivatives using a PO-CL reaction. The high sensitivity achieved makes it a useful alternative to other HPLC detection systems. The method was validated in the determination of psilocin in mushroom samples. Experimental Reagents All chemicals used were of analytical-reagent grade.Psycho- tropic indole derivatives were obtained from Sigma (St. Louis, MO, USA) with the exception of a-methyltryptamine, which was purchased from Aldrich (Milwaukee, WI, USA). All were used as received. Standard solutions containing 5 X 10-3 moll-' of each drug were prepared by dissolving the required amount in acetone (chromatographic grade, Romil Chemicals), and stored at 4 "C in a refrigerator. A 10-2 mol 1-1 bis(2,4,6-trichloro- pheny1)oxalate solution was made by dissolving 449 mg of the chemical (Aldrich) in 100 ml of ethyl acetate. A 0.7 mg ml-1 5-dimethy laminonaphthalene- 1 -sulfonyl chloride (dansyl chlo- ride, Sigma) solution was prepared in acetone and stored refrigerated. A 4 X 10-2 mol 1-l tris(hydroxymethy1)methyl- amine (Merck, Elmsford, NY, USA) buffer solution was prepared by dissolving 121.1 mg of reagent in water and adding enough hydrochloric acid to adjust the pH to 9.0 in a final volume of 25 ml.A 9 X 10-2 moll-' sodium carbonate buffer solution was made by dissolving 954 mg of salt (Merck) in water and adjusting the pH to 10.5 with HC1 in a final volume of 100 ml. Apparatus Chemiluminescence was measured on a Perkin-Elmer 650- 10s spectrofluorimeter with its light source off, the sample holder of which was replaced with a small magnetic stirrer holding a cylindrical glass reaction vessel; in order to acquire as much emitted CL as possible, an Oriel 44132 (1 in diameter) mirror was placed in front of the photomultiplier tube. The TCPO solution was added from a Metrohm Dossimat 665 autoburette.Kinetic data were acquired and processed by a NEC/Multisync 2A 33 MHz compatible computer equipped with a PC-Multilab 8 12 PG analogue-to-digital converter and running a program written by the authors in QBasic V.4.0.50 Analyst, January 1996, Vol. 121 Derivatization of Psychotropic Indole derivatives A volume of 200 pl of standard psychotropic solution containing between about 0.6 and 140 nmol of the hallucinogen, depending on its dynamic range (see Table I), was derivatized by adding 200 pl of 9 X mol 1-l carbonate buffer (pH 10.5) and 65 pl of 0.7 mg ml-1 dansyl chloride over 15 min at room temperature. Then, 45 pl of 0.13 mol 1-1 proline solution were added and the mixture was heated at 55 "C for 20 min. Next, 450 pl of doubly distilled water were added and the mixture was allowed to cool to room temperature.The derivative was extracted into 250 pl of chloroform and an aliquot of the organic phase was used for the subsequent CL determination. PO-CL Determination of Psychotropic Indole Derivatives To the reaction vessel of the CAR system were added 6.0 pl of derivatized psychotrope solution in chloroform, 65 pl of concentrated hydrogen peroxide (33% v/v, aqueous solution), and 20 1-11 of 4 X 10-2 mol I-' TRIS buffer (pH 9.0). The mixture was accurately diluted to 1 .O ml with acetone and 1 .O X moll-' TCPO was added at 8 ml min-1 with continuous stirring. The CL signal was monitored throughout the reaction and kinetic data (relative CL intensity versus time) were acquired at a rate of 10 ms per point, the maximum reaction rate being measured within about 1.5 s.The observed maximum reaction rate was corrected by subtracting the maximum reaction rate measured using the same procedure for a reagent blank (containing no psychotrope). Preparation of Mushroom Samples An amount of between 10 and 50 mg (accurately weighed) of ground, dried mushrooms was allowed to macerate in 10 ml of dilute acetic acid (pH 4.0) for 1 h, after which the mixture was heated at 70 "C for 30 min. Once cool, the mixture was filtered through glass-wool and the filtrate was adjusted to pH 8 with concentrated ammonia solution. The solution was extracted with 15 ml of chloroform for 10 min and the organic phase was then evaporated to dryness. The residue was dissolved in 200-300 pl of acetone and a 50 pl aliquot was subjected to the derivatization and PO-CL determination of psilocin in the mushroom sample. Results and Discussion The CAR technique is an effective means for increasing the signal-to-noise ratio in peroxyoxalate chemiluminescence reac- t i o n ~ ~ ~ and hence to ensure the sensitive determination of analytes.In this work, it was used to develop a PO-CL method for the determination of psychotropic compounds following the formation of dansylated derivatives using the TCPO-hydrogen peroxide system. Five hallucinogens were used, namely N- methyl tryp tamine, 5 -me thyltryp tamine, a-me thy ltryp tamine, bufotenine (4-hydroxy-NJV-dimethyltryptamine) and psilocin (5-hydroxy-NJV dimethyltryptamine); the structural formulae of their dansylated derivatives and the CL reaction scheme are shown in Fig.1. Their CL versus times curves, recorded under suitable experimental conditions, are shown in Fig. 2. As can be seen, the maximum response was exhibited by N-methyl- tryptamine. Optimization of the Dansylation Reaction Primary and secondary amines and phenols are known to react quantitatively with dansyl chloride (DNS-CI) to yield sulfo- namides or phenolic esters that exhibit intense yellow fluores- cence.I6 The dansyl derivatives are formed under different experimental conditions depending on the particular starting compound. Thus, the derivatization reaction usually takes place in alkaline medium at room temperature for about 30-40 min and some heating is occasionally required. In this work, we optimized the experimental conditions for the derivatization reaction since the psychotropic indole compounds studied had never been dansylated previously.We chose N-methyltrypta- mine (NMT) as the psychotropic compound to be used to investigate the effect of experimental variables on the deriva- tization reaction and the ratio between the maximum reaction rates (MRR) measured in the presence and absence of hallucinogen (MRRsam,1JMRRbl~) as the measured analytical parameter. All concentrations quoted are referred to the final volume in the reaction medium (1 .O ml). The effect of DNS-C1 on the formation of dansylated hallucinogens is shown in Fig. 3A; as can be seen, about 170 nmol of DNS-C1 ([DNS-Cl]/[NMT] = 25) ensured the maximum difference between the observed maximum reaction rate for the sample and blank reactions.We chose this measured analytical parameter since the MRRSmple/MRRblank was in- creased by low DNS-C1 concentrations owing to the low MRRblank levels, and under these conditions the dansylation reaction was not quantitative. The decrease in the observed analytical signal at high concentrations of the labelling reagent can be ascribed to an increase in MRRblank that resulted in a virtually negligible difference between the sample and blank signals. From these results, a DNS-C1 amount of 168.8 nmol was chosen as optimum, which was obtained by placing 65 pl of 0.7 mg ml-1 dansyl chloride in the reaction solution. As in most reported procedures, the alkaline medium for the derivatization reaction was provided by carbonate buffer; the optimum buffer concentration and pH were 9 X 10-2 moll-' and 10.5, respectively (Fig.3B). These values resulted in a maximum MRRsamplJMRRblank ratio of approximately 1.2. The effects of temperature and time on the dansylation reaction were also tested in the light and in the dark. The temperature was varied between 20 and 60 "C and the time over the range 5-150 min, both in the light and in the dark. Based on the results, the dansylation reaction was quantitative when the psychotropic solution was treated with alkaline DNS-C1 for 15 min at room temperature in the light. At that point, the derivatization reaction was optimized; however, the ratio between the measured maximum reaction rates of sample and blank was not very high (approximately 1.2) Table 1 Figures of merit of the calibration plots for the PO-CL determination of psychotropic indole derivatives Correlation coefficient LOD/ s,+ Alkaloid Linear range/nmol Linear regression equation* ( n = 15) pmol (%) N-Methyltryptamine 0.6-40 MRR = (4.4 X + 3.1 X 10-3C 0.9994 138 1.53 5-Methyltryptamine 0.6-40 MRR = -(2.0 x 10-3) + 2.0 x 10-3c 0.9986 184 1.90 a-Methyltryptamine 3.0-70 MRR = -(2.3 X lop3) + 7.1 X 10-4C 0.9950 830 2.20 Bufotenine 0.640 MRR = -(4.2 X lop4) + 1.1 X 10-3C 0.9981 230 2.00 Psilocin 10-140 MRR = -(1.1 X + 9.0 X 1O-T 0.9935 2920 1.66 * MRR = maximum reaction rate; C in nmol.Analyst, January 1996, Vol.121 51 owing to the high contribution of excess of DNS-CI to the analytical signal. Two alternative procedures have been re- ported for removing unused DNS-Cl, namely the selective extraction of dansylated derivatives in an organic solvent such as toluene17J8 or benzene,19920 and the addition of proline before the extraction.21-22 On the basis of the experimental results, we used proline to remove excess of DNS-CI and chloroform for the selective extraction of dansylated halluci- nogens.Under these conditions, the MRRsmplJMRR~l~ ratio was considerably increased (to approximately 3 .O). Optimization of the PO-CL Reaction Although the PO-CL reaction between TCPO and hydrogen peroxide has previously been used by the authors for the determination of phenothiazine derivatives using the CAR technique,15 re-optimization was required on account of the different spectrofluorimetric features of the analytes in this work.All experiments were carried out in an acetone-water- chloroform (91.0 + 8.5 + 0.5) medium with a minimum water content (water strongly decreases the CL signal) that would be exclusively supplied by the hydrogen peroxide and TRIS solutions added to the reaction vessel; the chloroform solvent contained the dansylated psychotropic indole derivatives and the acetone acted as co-solvent to ensure miscibility in the reaction vessel (between water, chloroform and ethyl acetate, the last of which was provided by the TCPO solution added from the autoburette). The effect of the concentration and pH of the TRIS buffer is shown in Fig. 4A; as can be seen, the maximum analytical signal SO, I 0 H Dansylated psilocin CI CI Dansylated methyltryptarnines R, R2 R3 NMethyltryptamine CH, H H 5-hhthyltryptrmine H H CH, CI H Dan sy I ated bufotenin e 0 0 c - c II II I I 0 - 0 C1 CI L J TCPO l.2-Dioxeth.~-3.4-dio~ r o o l * * Dansylated + hv hallucinogen + 2 co, --+ Da nsy la t ed Da nsy lated + ha1 lucinogen - hallucinogen Lo -O_I Fig.1 Structures of the psychotropic dansylated derivatives assayed and CL reaction scheme.52 Analyst, January 1996, Vol. 121 was obtained with 4 X 10-2 mol 1-1 TRIS buffer (pH 9.0), a volume of 20 pl of which was used in the reaction vessel. The effect of the hydrogen peroxide concentration was studied over the range 0.13-0.7 mol 1-1; the maximum response was obtained with a 0.7 moll-' concentration (65 pl of concentrated 1 .o 0.0 0.5 1 .o 1.5 2.0 2.5 3.0 Time/s Fig.2 Relative PO-CL intensity versus time plots for the psychotropic indole derivatives studied: 1 N-methyltryptamine; 2,5-methyltryptamine; 3 psilocin; 4, bufotenine; and 5, or-methyltryptamine. All compounds at 3.0 nmol except psilocin (30 nmol). For conditions, see under Experimental. [DNS-Cl]/[NMT] 0 1 0 2 0 3 0 4 0 5 0 6 0 I 1 I I 1 1 1 A 0 1 0 0 2 0 0 3 0 0 4 0 0 DNS-Cllnmol hydrogen peroxide added to the reaction vessel). This concen- tration was a compromise between the increased PO-CL signal obtained by raising the hydrogen peroxide concentration and the decreased signals resulting from the increased amounts of water contained in the aqueous H202 solution. The effects of the TCPO concentration and its addition rate are illustrated in Fig. 4B. A 1 X 10-2 moll-' TCPO solution in ethyl acetate added at 8 ml min-1 from the autoburette was selected for subsequent experiments.Higher TCPO concentrations could not be used owing to the low solubility of this reagent in ethyl acetate. Determination of Psychotropic Indole Derivatives Under the selected working conditions, psychotropic indole derivatives were determined at the nanomole level by using the proposed PO-CL method. Data relevant to the calibration graphs for these hallucinogens are summarized in Table 1. The limits of detection (LODs) were calculated following IUPAC recommendations;23 the precision, as the relative standard deviation, was checked on 11 samples containing between 3.0 and 10 nmol of psychotropic indole derivative (psilocin was tested at 40 nmol owing to its lower sensitivity).The over-all time required to perform three replicate analyses (excluding that required for the derivatization reaction) and sample changeover in the CAR system was 45 s, so the throughput was about 80 samples h-1. The different sensitivities achieved depend on the structure of each hallucinogen, which can influence the CL signal of the dansylated derivative. Buffer pH 7 8 9 10 11 12 1.3 8 0.0 0.1 0.2 0.3 0.4 [ Carbonate buffer ]/mot I-' Fig. 3 hallucinogens. [NMT] = 7.0 nmol. Conditions for PO-CL reaction as described under Experimental. MRR, maximum reaction rate. Effect of, A, the DNS-Cl concentrations; and B the carbonate buffer concentration (filled circle) and pH (open circle) on the dansylation of the TRIS pH 7 8 9 10 11 12 Addition Rate/ml min-' 0 2 4 6 8 1 0 1 2 1 4 3.5 3.0 2.5 2.0 0.00 0.05 0.10 0.15 [TRIS]/mol I-' 3.5 3.0 25 2.0 Fig.4 circle) on the PO-CL determination of 7.0 nmol of dansylated NMT. Other conditions as described under Experimental. MRR, maximum reaction rate. A, Influence of the TRIS concentration (filled circle) and pH (open circle); and B of the TCPO concentration (filled circle) and addition rate (openAnalyst, January 1996, Vol. 121 53 At this point, it is of interest to compare the analytical features of the proposed PO-CL method for the determination of psilocin with existing alternatives. Because such alternatives are chromatographic (HPLC) and use various detection system, we chose LODs for comparison and assimilated the amount of psilocin injected into the chomatograph with the LOD of the proposed method, in addition to the amount of derivatized psilocin used for the PO-CL determination in the CAR system.Based on these considerations, the proposed method has an LOD of approximately 15 ng for psilocin, which is similar to the LODs for the cromatographic methods using voltametric d e t e c t i o n , g ~ ~ ~ , ~ ~ but one to two orders of magnitude lower than those using optical (spectrophotometric or spectrofluorimetric) detection.*,l0,25 In view of these results, the proposed method is an excellent choice for the sensitive determination of psilocin. Its use as a CL detection system in HPLC is bound to introduce significant improvements in the determination of these hallucin- ogens and related compounds. Determination of Psilocin in Fungi of the Genus Psilocibe The above results suggested that the proposed method could be applied to the determination of hallucinogenic alkaloids in real samples.Because the recreational use of hallucinogenic mush- rooms is a matter of growing concern in some countries, we applied the proposed method to the determination of psilocin in this type of sample; among hallucinogenic mushrooms, several Psilosyhe species contain indole alkaloids (mostly psilocybin and smaller amounts of psilocin).I0 The proposed method is useful for the determination of psilocin in the presence of psilocybin since the hydroxyl group at which dansylation takes place is phosphorylated in psilocybin, which therefore produces no CL signal. However, psilocybin can readily be converted into psilocin by a rapid dephosphorylation reaction with alkaline phosphatase,* so the proposed method can be used to determine both indole alkaloids in two sample aliquots (the Table 2 Recovery of psilocin added to mushrooms samples Sample Cap 1 Cap 2 Cap 3 Stem Psi 1 oc in/pg Recovery Found Added I-18 % 3.75 - 12.25 20.40 12.25 20.40 12.25 20.40 12.25 20.40 5.78 - 6.53 - 5.72 - - 16.10 24.15 17.72 26.83 19.03 26.88 18.12 25.99 - - - - 100.6 100.0 98.2 102.5 101.3 99.8 100.8 99.5 Mean: 100.3 - - - SD: 1.28 Table 3 Determination of psilocin in Psilocyhe Semilanceata (Fr.) Kumm.by use of the proposed PO-CL method Sample Amount/mg Psilocin (%)* Average recovery (%) Cap 1 49.0 0.046 f 0.001 100.3 Cap 2 12.1 0.195 f 0.015 100.3 Cap 3 35.4 0.11 1 k 0.003 100.6 Stem 48.9 0.070 f 0.002 100.1 * Average of three determinations k standard deviation.psilocybin concentration would obviously be obtained by difference). A fungal species of the genus Psilocybe, viz., Psilocybe semilanceata (Fr.) Kumm., was collected at Puerto de 10s Anclares, Lug0 (Spain), in 1994. Several samples of caps and stems were subjected to the procedure described under Experimental, which is based on the aqueous-organic extrac- tion method for the isolation of psilocin from hallucinogenic mushrooms proposed by Casale26 (chloroform was used instead of diethyl ether, however). In order to evaluate potential adverse effects from other components, the recovery of psilocin added to these samples was also determined by comparing the results obtained before and after adding the psilocin standard solution.The recoveries obtained are given in Table 2. Table 3 shows the results provided by the proposed method, which were consistent with reported contents of psilocin in hallucinogenic mushroom samples. lo The authors gratefully acknowledge financial support from the Direcci6n General Interministerial de Ciencia y Tecnologia (DIGICyT) for the realization of this work as part of Project PB9 1-0840, and Dr. F. Infante of the Departamento de Biologia Vegetal y Ecologia of the University of C6rdoba (Spain) for kindly supplying the mushroom specimens. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Haddad, L. D., and Winchester, J. F., Clinical Management of Poisoning and Drug Overdose, Saunders, Philadelphia, 2nd edn., 1990.Wingard, L. B., Jr., Brody, T. M., Lamer, J., and Schwartz, A., Human Pharmacology: Molecular-to-Clinical, Mosby Year Book, St. Louis, MO, 1991. Bowman, W. C., and Rand, M. J., Textbook of Pharmacology, Blackwell, Oxford, 2nd edn., 1980. Faurbye, A., and Pind, K., Nature (London), 1968, 220,489. Himwich, H. E., Jenkins, R. L., Fujirnori, M., Narasimhachari, N., and Ebersole, M., J. Autism Child. Schizophr., 1972, 2, 114. Baumann, P., and Narasirnhachari, N., J. Chromatogr., 1973, 86, 269. Narasimhachari, N., and Hirnwich, H. E., Life Sci., 1973, 12, 475. Christiansen, A. L., and Rasmussen, K. E., J. Chromatogr., 1983, 270, 293. Vanhaelen-FastrC, R., and Vanhaelen, M., J. Chromatogr., 1984,312, 467. Wurst, M., Semerdzieva, M., and Vokoun, J., J. Chromatogr., 1984, 286, 229. Wurst, M., Kysilka, R., and Koza, T., J. Chromatogr., 1992, 593, 201. Kysilka, R., and Wurst, M., J. Chromatogr., 1989, 464,434. Mhrquez, M., Silva, M., and PCrez-Bendito, D., Analyst, 1988, 113, 1733. Velasco, M., Silva, M., and Perez-Bendito, D., Anal. Chem., 1992,64, 2359. Cepas, J., Silva, M., and Ptrez-Bendito, D., Anal. Chem., 1994, 66, 4079. Del Castillo, B., Alvhrez-Builla, J., and Lerner, D. A., in Luminis- cence Techniques in Chemical and Biochemical Analysis, ed. Baeyens, W. R. G., De Keukeleire, D., and Korkidis, K., Marcel Dekker, New York, 1991, ch. 5 , pp. 99-100. Ibe, A., Saito, K., Nakazato, M., Kikuchi, Y., Fujinura, K., and Nishima, T., J. Assoc. Off. Anal. Chem., 1991, 74, 695. Barrett, D. A., Shaw, P. N., and Davis, S. S., J. Chromatogr., Biomed. Appl., 1991, 104, 135. Cann-Moisan, C., Caroff, J., and Girin, E., J. Chromatogr., Biomed. Appl., 1992, 112, 134. Desiderio, M. A., Davalli, P., and Perin, A., J. Chromatogr., Biomed. Appl., 1987, 63, 285. Lindsay-Smith, J. R., Smart, A. U., Hancock, F. E., and Twigg, M. V., J. Chromatogr., 1991, 547, 447.54 Analyst, January 1996, Vol. 121 22 Lindsay-Smith, J. R., Smart, A. U., Hancock, F. E., andTwigg, M. V., Chem. Ind. (London), 1989,11, 353. 23 Long, G. L., and Winefordner, J. D., Anal. Chem., 1983, 55, 712A. 24 Kysilka, R., J. Chromatogr., 1990, 534, 287. 25 Bomer, S., and Brenneisen, R., J. Chromatogr., 1987, 408, 402. 26 Casale, J. F., J. Forensic. Sci., 1985, 30, 247. Paper 5104356E Received July 5 , I995 Accepted September 12,1995