(allicin) in garlic

Analytica Chimica Acta 441 (2001) 37–43

Determination of diallyl thiosulfinate (allicin) in garlic (Allium sativum L.) by high-performance liquid chromatography with a post-column photochemical reactor P. Bocchini a,∗ , C. Andalò a , R. Pozzi a , G.C. Galletti a , A. Antonelli b a

Dipartimento di Chimica “G. Ciamician”, Università degli Studi di Bologna, Via F. Selmi 2, I-40126 Bologna, Italy b Istituto di Industrie Agrarie, Università degli Studi di Bologna, Via S. Giacomo 7, I-40126 Bologna, Italy Received 8 August 2000; received in revised form 19 March 2001; accepted 17 April 2001

Abstract An analytical method for the determination of allicin (diallyl thiosulfinate) in garlic (Allium sativum L.) samples using reversed-phase HPLC with both UV and electrochemical detection (ED) and on-line post-column photochemical reaction is reported. Standard allicin was synthesised and its behaviour at the chosen analytical conditions was tested. The post-column irradiation at 254 nm on the one hand decreased the response of allicin to UV detector, and on the other hand, allowed the determination of this compound, which is otherwise electrochemically inactive, using the electrochemical detector. Detection limits of 0.1 and of 0.01 mg/l using UV and ED detectors, respectively, and linearity of response in the range 1–8 mg/l were obtained. Samples of garlic extracts, obtained following two different extraction procedures, were analysed and the results obtained using both UV and electrochemical detectors were compared. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Allicin; Allium sativum L.; Liliaceae

1. Introduction Garlic (Allium sativum L.) is a perennial plant of Liliaceae family. The oblong bulb, made of several bulbils is the edible part. Garlic has been grown for many centuries in the Mediterranean area for its characteristic flavour and medicinal properties and it is widely cultivated in Italy. Many researchers investigated the nature of the volatile compounds in garlic which is known to possess many beneficial activity for human health [1,2]. The importance of thiosulfinates in the flavour of garlic distillates has been known since the studies of ∗ Corresponding author. Tel./fax: +39-051-2099450. E-mail address: [email protected] (P. Bocchini).

Cavallito and Bailey in 1944 [3]. Cavallito and Bailey discovered diallyl thiosulfinate (allicin) as the responsible of fresh garlic flavour, and Stoll and Seebeck [4] elucidated its biochemistry. The characteristic flavour, aroma, and active principles in garlic are present only after rupture of the cell membranes, which allows the enzyme alliinase to degrade S(+)-allyl-l-cysteine sulfoxide (alliin), producing allicin and other sulfur containing compounds. A simplified scheme of allicin formation in garlic is shown in Fig. 1. Block and O’Connor, in 1974 [5], summarised the complex chemistry of alkyl thiosulfinates and Block et al. [6] proposed a suitable HPLC method for the quantification of alkyl thiosulfinate in garlic as well as in other Liliaceae. More recently allicin and related thiosulfinates in garlic were analysed by HPLC

0003-2670/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 0 3 - 2 6 7 0 ( 0 1 ) 0 1 1 0 4 - 7

38

P. Bocchini et al. / Analytica Chimica Acta 441 (2001) 37–43

Fig. 1. Simplified mechanism of allicin formation.

with MS detector [7–9]. Gas chromatography is not used due to the easy formation of thermal decomposition products unless precautions to minimise the thermal degradation of sulfur compounds are taken [7,10–12]. Post-column, on-line photochemical reactions have stimulated more and more interest among the HPLC users due to the improvement in response, sensibility and chromatographic resolution obtained for some compounds. The term derivatisation is sometime used to define this process. In post-column photochemical reaction, it is possible that a photochemical degradation product or an excited species of the analyte is detected with greater selectivity and sensitivity than the original molecule [13–17]. Aromatic and sulfur containing compounds have been determined by HPLC with electrochemical detection after photochemical reaction. By continuous on-line UV irradiation, non-electrochemically active compounds such as phenylalanine and aspartame have shown an oxidative current and have been detected by electrochemical detector (ED) [18,19]. Consequently, the selectivity and sensitivity of such a detector and the possible differences in lamp on/lamp off chromatograms, can be exploited for such a class of compounds. A recent paper dealing with post-column photochemical reaction in high-performance liquid chromatography was published by Lores et al. [20]. The present paper reports on the determination of allicin by HPLC-UV and HPLC-ED with post-column photochemical reactor. Chromatograms obtained in lamp on/lamp off conditions, and quantitative results of the amount of allicin in a commercial garlic sample are reported. The use of ED after a photochemical

reaction is of importance, because it offers a further unequivocal identification tool and extends the range of application of such a detector.

2. Experimental conditions 2.1. HPLC conditions The HPLC system consisted of the following components: a Waters (Milford, MA, USA) Model 590 pump, a Rheodyne (Cotati, CA, USA) Model 7725i injector, a reversed-phase C18 column (150 mm ×4.6 mm i.d., 5 ␮m particle size) Spherisorb (Phase Separation, Clwyd, UK). The photochemical reactor was a ICT Beam Boost (ICT, Frankfurt, Germany) D-6808 equipped with a 254 nm UV lamp and a 20 m × 0.3 mm i.d. Teflon reaction coil. The detectors were a TSP UV2000 detector (Thermo Separation Products, San Jose, CA, USA) and an ESA (Bedford, MA, USA) Coulochem Model 5100A electrochemical detector. The column was operated in isocratic mode (50:50 MeOH:H2 O) at a flow rate of 0.5 ml/min, corresponding to 240 s of irradiation in the photochemical reaction unit. Unless otherwise specified, the UV detector was set at 254 nm and the electrochemical detector was operated with the first electrode (V1 ) set at +0.0 V, all the determinations being made on the second electrode (V2 ) set at +1.7 V. 2.2. Standard allicin The synthesis of allicin was based on a previously published method [21]. Briefly, an equimolar


amount of perbenzoic acid (PBA), dissolved in 20 ml of dichloromethane, was slowly added to a solution of allyl disulfide (1.46 g/100 ml in dichloromethane), under rapid magnetic agitation and cooled to −10◦ C. The reaction mixture was allowed to stand at room temperature for 1 h. The excess of acid was removed by washing the mixture with a sodium bicarbonate solution. The dichloromethane solution was rinsed with distilled water, dried over sodium sulfate and the solvent was removed by rotary-evaporation. The dried product was weighed and standard solutions in MeOH were prepared. Analysis by HPLC-DAD (Diode Array Detector, Spectrasystem UV6000LP) of the solution showed that allicin accounted for >99% of the standard.

39

2.3. Real samples The brief description of two extraction procedures were adapted from a previously published method [6]: 1. Non-purified extract: A weighed amount of store-purchased garlic was crushed for 1 min using a blender mixer in the presence of distilled water (10 times in weight). The aqueous extract was filtered and diluted to 500 ml with water. An aliquot of the diluted extract was filtered through a 0.22 ␮m pore size filter and injected as such. 2. Purified extract: The diluted extract (10 ml), prepared as described in the previous paragraph, was saturated with NaCl and extracted with 10 ml of

Fig. 2. UV spectra of standard allicin (trace a) and of standard allicin after irradiation (trace b).

40


dichloromethane (three times). The three organic fractions were collected, dried over anhydrous Na2 SO4 , and desiccated by rotary-evaporation. The dried sample was dissolved in 10 ml of MeOH and injected.

3. Results and discussion A preliminary analysis of standard allicin using a Diode Array Detector with spectral range 200–600 nm was made under the chromatographic conditions reported in Section 2. The UV spectra were recorded with the photochemical reactor off and on in order to obtain spectrometric data on pure allicin and on its photochemical degradation products (Fig. 2, traces a and b). In the chromatogram, only the peak of allicin was present and no other degradation component or oxidation product was detected. The UV spectra of allicin changed dramatically after irradiation, as shown in Fig. 2b, indicating that the molecule underwent a change in its chemical structure. The analytical conditions for the analysis of allicin by HPLC-UV were studied. In Fig. 3a, the chromatograms obtained by analysing a solution of standard allicin (4 mg/l) without (trace A) and with (trace B) post-column photochemical reactor are shown. The response was linear in the range 1–8 mg/l both with and without post-column photochemical derivatisation. The linear regression equation was: y = 9757x, (y = area in arbitrary units, x = concentration in mg/l), with R 2 = 1.00 when the photochemical reactor lamp was off and y = 1968x with R 2 = 0.95 when the lamp was on (Fig. 4). The UV detector response (254 nm), after postcolumn irradiation of allicin, resulted lower (Fig. 3a, Trace B) than that obtained when the lamp was off. The lower response was ascribed to changes in the absorption spectrum due to the presence of allicin degradation products (Fig. 2). Three replicate injections of a standard solution of allicin (8 mg/l) yielded a relative standard deviation of 2.7 and 2.1% (lamp off and on, respectively). The detection limit was 0.1 mg/l in both experimental conditions. Allicin showed no electrochemical activity when the photochemical reactor was off. When the photochemical reactor was on, a hydrodynamic voltammogram was recorded in order to check both the

Fig. 3. (a) Cromatograms of standard solutions of allicin (4 mg/l) as obtained by HPLC with UV detection. Trace A: no post-column irradiation, trace B: with post-column UV irradiation. (b) Cromatograms of standard solutions of allicin (4 mg/l) as obtained by HPLC with elecrochemical detection (V2 = 1.7 V). Trace A: no post-column irradiation, trace B: with post-column UV irradiation.


41

Fig. 4. Calibration curves of standard allicin as obtained by HPLC-UV analysis with and without photochemical derivatisation and by HPLC-ED with photochemical derivatisation.

electrochemical behaviour of allicin photochemical degradation products and to decide the optimum voltage for detection. Allicin was injected at V2 voltages in the range 0.2–1.9 V. A one-step curve was obtained with a slow increase in the detector response reaching a plateau at values ≥1.7 V. Fig. 3b shows the chromatograms of allicin injections with the photochemical reactor off (trace A) and on (trace B) at V2 = 1.7 V. According to the manufacturer, the irradiation lasts 240 s for a mobile phase flow-rate of 0.5 ml/min in a 20 m × 0.3 mm i.d. coil at room temperature. It is apparent that, under such conditions, allicin degraded into products having different redox properties. The photolitic derivatisation induced the formation of at least one oxidisable site in the analysed molecule. It is known that allicin is easily degraded forming by-products [5], and many thermal recombination products were reported. However, the UV spectrum after photochemical irradiation, shown in Fig. 1b, is not sufficient to determine which degradation product is responsible of the electrochemical activity. A V2 potential of +1.7 V was used throughout the rest of the experiments. Under such conditions, three replicate injections of a standard solution (8 mg/l) of

allicin yielded a relative standard deviation of 2.0%. The detection limit was 0.01 mg/l with a linear response in the concentration range 1–8 mg/l. The linear regression equation was: y = 975711x, where y is the area in arbitrary units and x the concentration in mg/l, (linear correlation coefficient R 2 = 1.00, Fig. 4). The sensibility of the electrochemical detector to allicin proved to be greater than that obtained by using the “classical” method for allicin analysis (i.e. UV detection at 254 nm). 3.1. Garlic samples Samples of store-purchased garlic, were analysed using both UV and electrochemical detector. Results on the amount of allicin in the analysed samples were obtained applying two different procedures of sample preparation. In Fig. 5a and b, the chromatograms of garlic extracts (non-purified, see Section 2.3, procedure no. 1), as obtained by using UV and electrochemical detector with (trace B) and without (Trace A) post-column irradiation, are shown. From this figure, it is apparent that other extracted compounds undergo some sort of change in consequence of the photochemical irradiation.

42

P. Bocchini et al. / Analytica Chimica Acta 441 (2001) 37–43 Table 1 Quantitative data of allicin (mg/g of fresh garlic, three samples, triplicate analysis) as obtained by analysing fresh garlic samples by HPLC-UV, HPLC-hν-UV and HPLC-hν-ED

Non-purified Purified

HPLC-hν-ED

HPLC-hν-UV

HPLC-UV

3.6 ± 0.2 2.1 ± 0.1

3.8 ± 0.1 1.7 ± 0.1

4.6 ± 0.1 1.84 ± 0.01

Quantitative data on the amount of allicin in garlic were obtained using the calibration curve previously reported and are summarised in Table 1 as milligrams of allicin per gram of fresh garlic±standard deviation. The HPLC-UV and HPLC-ED data obtained analysing non-purified extracts (photochemical reactor on) were very similar but lower when compared to the result obtained with UV detection and photochemical detector off. The higher value obtained using the “classical” UV method can be due to possible co-eluting compounds when direct injection of non-purified garlic extracts were made. The presence of the post-column photochemical reactor allowed the elimination of this interference without any previous sample work-up. Quantitative data obtained after purification of the aqueous extract were similar using ED and UV detector both with and without post-column photochemical reactor. However, these data were lower than those obtained without purification of the sample. Some allicin was probably lost in the rather complex purification steps required.

4. Conclusion

Fig. 5. (a) Chromatograms of allicin aqueous solutions after extraction of store-purchased garlic (non-purified extracts, see Section 2.3 for extraction conditions) by HPLC/UV analysis. Trace A: no post-column irradiation, trace B: post-column UV irradiation on. (b) Chromatograms of allicin aqueous solutions after extraction of store-purchased garlic (non-purified extracts, see Section 2.3 for extraction conditions) by HPLC/ED analysis. Trace A: no post-column irradiation, trace B: post-column UV irradiation on.

Allicin is a molecule so far determined using HPLC-UV detection. The use of a post-column photochemical reactor allowed to extend the selectivity and sensitivity of electrochemical detection to the analysis of this molecule. Using a post-column photochemical reactor allowed us to avoid time consuming sample preparation steps with any consequent recovery losses. Moreover, the different responses of allicin depending on reactor modes on or off are additional tools for the unequivocal identification of this compound and improvement of possibly critical chromatographic separations.


References [1] W.H. Briggs, H. Xiao, K.L. Parkin, C. Shen, I.L. Goldman, J. Agric. Food Chem. 48 (2000) 5731. [2] M.-C. Yin, W.-S. Cheng, J. Agric. Food Chem. 46 (1998) 4097. [3] C.J. Cavallito, J.H. Bailey, J. Am. Chem. Soc. 66 (1944) 1950. [4] A. Stoll, E. Seebeck, Adv. Enzymol. 11 (1951) 377. [5] E. Block, J. O’Connor, J. Am. Chem. Soc. 96 (1974) 3929. [6] E. Block, S. Naganathan, D. Putman, S.-H. Zhao, J. Agric. Food Chem. 40 (1992) 2418. [7] S. Ferary, J. Auger, J. Chromatogr. A 750 (1996) 63. [8] E.M. Calvey, J.E. Matusik, K.D. White, R. DeOrazio, D. Sha, E. Block, J. Agric. Food Chem. 45 (1997) 4406. [9] E.M. Calvey, K.D. White, J.E. Matusik, D. Sha, E. Block, Phytochemistry 49 (1998) 359.

43

[10] E. Block, D. Putman, S.-H. Zhao, J. Agric. Food Chem. 40 (1992) 2431. [11] R. Artacho Mart´ın-Lagos, M.F. Olea Serrano, M.D. Ruiz Lopez, Food Chem. 53 (1995) 91. [12] I. Arnault, N. Mondy, F. Cadoux, J. Auger, J. Chromatogr. A 896 (2000) 117. [13] U.A.Th. Brinkman, Chromatographia 24 (1987) 190. [14] C.M. Selevka, I.S. Krull, I.S. Lurie, J. Chromatogr. Sci. 23 (1985) 499. [15] D.E.G. Shuker, S.R. Tannenbaum, Anal. Chem. 55 (1983) 2152. [16] C.J. Miles, M. Zhou, J. Agric. Food Chem. 38 (1990) 986. [17] R.W. Frei, Chromatographia 15 (1982) 161. [18] L. Dou, I.S. Krull, Anal. Chem. 62 (1990) 2599. [19] G.C. Galletti, P. Bocchini, J. Chromatogr. A 729 (1996) 393. [20] M. Lores, O. Cabaleiro, R. Cela, Trends Anal. Chem. 18 (1999) 392. [21] D.L. Small, J.H. Bailey, C.J. Cavallito, J. Am. Chem. Soc. 69 (1947) 1710.