Introduction

Khat, the leaves or short tops of the evergreen plant Catha edulis Forsk. (Celastraceae), is well known in East Africa (Kenya, Ethiopia, Madagascar) and the Arabian peninsula (North Yemen) for being chewed habitually by many people. At present, khat has begun to be introduced to certain developed countries, e.g. Great Britain, were use of khat among East African and Yemeni immigrants is substantial¹. Shipments of khat — mostly by air transportation — have been observed too by customs authorities in France and the United States^1,2. The widespread use and popularity is due to its CNS-stimulating phenylpropylamines, especially to cathinone (α-aminopropiophenone, 'natural amphetamine') and to a much lesser extent to norpseudoephedrine ('cathine') and norephedrine^1-6. After the World Health Organisation recommended that cathinone be placed under international control, the amphetamine-like substance has been included in Schedule I of the UN Convention on Psychotropic Substances. The recently found phenylpentenylamines merucathinone, merucathine and pseudomerucathine^3,7-9 play only a minor role concerning the psychoactive effects of khat¹⁰.

In the guidelines for future investigations which were formulated by the UN Expert Group on the Botany and Chemistry of Khat in 1979¹¹, an 'urgent need for specific qualitative means of identification of khat material and quantitative analysis of the important plant constituents' was emphasized. Whereas a HPLC method for the quantitation of all known khatamines has already been proposed^3,12,13, to our knowledge no simple and reliable test for identification of khat samples has been published until now. The method of Röhm and Schmid¹⁴ is based on a very time consuming extraction (6-15 h) and is therefore not suited for a rapid identification. Furthermore the proposed thin layer chromatography (TLC) system results in inefficient separation of complex khat extracts (e.g. MeOH extracts). A macroscopic, microscopic and histochemical examination of khat leaves does not show enough characteristics and is inappropriate for forensic purposes. Nordal and Laane¹⁵ suggested a combination of morphological, histochemical and chromatographic (TLC and GC) methods. Therefore, our aim was to establish a specific and sensitive test for khat based on a rapid; simple and inexpensive TLC method. With minimum laboratory equipment it can be used for forensic and law enforcement applications. The proposed TLC method was developed and tested by repeated analysis of numerous Catha samples of different type, origin and age which have been screened before for khatamines by HPLC and GC profiling^8,12,13,19.

Material and Methods

Most of the khat samples (Catha edulis) were bought between 1982 and 1986 at the khat markets of Addis Ababa and Awedai/Harrar (Ethiopia), Nairobi and Mombasa (Kenya), Sanaa (North Yemen) and Anivorano (Madagascar). The specimens were deep-frozen within 24 h, transported by air in a cooler to the laboratory and stored at -20°C until used. Dried khat samples originating from South Africa and Catha transvaalensis were obtained through the Lowveld Botanic Garden, Transvaal and the Johannesburg Botanic Garden, Johannesburg (South Africa). Catha spinosa and Ephedra dystachia were cultivated at the Botanical Garden of the University of Berne. Duplicates of all analyzed specimens are being kept in the herbarium of the Institute of Pharmacy, University of Berne.

One leaf (medium size, corresponding to about 500 mg fresh or 150 mg dried plant material) of a khat sample and 2.5 ml MeOH were ground thoroughly in a mortar. The suspension was then transferred to a 2.5-ml vial with cap and sonicated for 10 min. After passing through a 0.2-µm cellulose filter (or through the tip of a Pasteur pipette filled with cotton wool) the clear and greenish solution was concentrated to about 100 µl under a stream of nitrogen and again filtered. If no ultrasonic bath is available the following alternative extraction procedure is recommended: one leaf of a khat sample and 5 ml MeOH were ground properly in a mortar. The suspension was then transferred to a test tube with stopper and shaken vigorously for 10 min. After filtration through a 5-cm paper filter the solution was concentrated to about 250 µl on a water bath (70°C) and again filtered through the tip of a Pasteur pipette filled with cotton wool.

Five microlitres of the filtrate were applied bandwise (5-mm line) to a TLC plate (Macherey-Nagel Alugram Sil G/UV₂₅₄ 0.25 mm, 40 x 80 mm or an equivalent product) and developed in EtOAc-MeOH-NH₃ 25% 85:10:5^15,16. A methanolic solution of about 1 mg/ml of cathinone hydrochloride (synthesized according to⁶) and norpseudoephedrine or norephedrine hydrochloride (Sigma Chemical Comp., St. Louis, U.S.A.) was used as standard. Prior to visualization the plate must be dried at room temperature or, more quickly, by short use of a hot air blower. The plate was observed under visible and UV light (254 nm), sprayed with fresh ninhydrin reagent (0.3 g in 100 ml n-butanol + 3 ml conc. acetic acid) and heated at 110°C for 2 min.

Results and Discussion

An extraction without partition and with MeOH only as solvent was chosen to get complex TLC patterns showing the khat sample zones of the typical khatamines and other constituents (Figs. 1a-c). Under visible light (Fig. 1a) khat samples show a brown-green zone of polar products (flavonoids) on the starting line and at Rf 0.06 a grey zone, at R_f 0.62 a light-yellow and at 0.72 a yellow zone, probably artifacts (decomposition products of the unstable cathinone) can be observed. At R_f 0.82 a green zone locates the colouring principle of the plant (chlorophyll).

Under UV₂₅₄ cathinone is visualized as a dark-blue zone at R_f 0.43 (Fig. 1b) and in higher concentrations shows a violet zone after spraying with ninhydrin reagent (Fig. 1c). Norpseudoephedrine and norephedrine are not or only weakly visible under UV₂₅₄, due to a hundred times lower &value than cathinone, but can be visualized with ninhydrin as pink zone at R_f 0.25-0.3 (not separated). The minimum detectable amount of cathinone and norpseudoephedrine is about 250 ng (UV₂₅₄) and 3 µg (ninhydrin), respectively. If 5 µl of the filtered extracts are used for TLC, this corresponds to a relative detection limit of 0.5 µg cathinone and 6 µg norpseudoephedrine/norephedrine/100 mg dried khat leaves. The sensitivity of the test is therefore sufficient to identify even one leaf of a khat sample with low phenylalkylamine content, e.g. khat from Yemen and Madagascar¹³. The phenylpentenylamines can not be separated from the corresponding phenylpropylamines and are always below the detection limit.

As cathinone has only been found in Catha edulis a positive identification of this compound is a specific indication for khat. Nevertheless, positive results should be confirmed by an alternative technique, such as high-performance liquid chromatography (HPLC), gas chromatography (GC) or gas chromatography/mass spectrometry (GC/MS)^3,8,12,13,16.

Catha spinosa (endemic in North Yemen¹⁷) and Catha transvaalensis (endemic in South Africa¹⁸) have a morphology similar to Catha edulis but produce different TLC patterns (Figs. 1a-c). A phytochemical study of these relatively unknown Catha species is in progress¹⁹.

The total content of khatamines and especially the concentration ratios of cathinone, norpseudoephedrine and norephedrine are dependent on the khat type, origin, time of harvesting, drying method, age and storage conditions^3,6,8,12,13. Dried and old khat samples usually contain only very small but still TLC detectable amounts of cathinone due to enzymatic reduction of cathinone to norpseudoephedrine and norephedrine^3,6,12. In such plant material norpseudoephedrine is the dominating khatamine^12,13. But in any case only a quantitative analysis of major and minor khatamines, for example by HPLC profiling^12,13, provides more information about the "history" and the psychotropic potency of a given khat sample. Other plants with norpseudoephedrine as constituent, such as Ephedra species, are morphologically different from khat and give not the same pattern (Figs. 1a-1c).

Rapid but unspecific colour tests (Marquis', modified Marquis', Froehde's, Simon's, Chen-Kao's or Mandelin's reagent^20-23) for the presumptive identification of phenylalkylamines do not give positive results with cathinone and norpseudoephedrine. Therefore the commercially available Merck colour test for amphetamine (part of the Drug Identification Kit no. 11850) or the UN colour test for amphetamines (part of the UN Drug Identification Kit), both use the modified Marquis' reagent, are not suitable for testing khat samples.