Quantitative determination of S-alk(en)ylcysteine-S-oxides by micellar electrokinetic capillary chromatography

Journal of Chromatography A, 1212 (2008) 154–157

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Short communication

Quantitative determination of S-alk(en)ylcysteine-S-oxides by micellar electrokinetic capillary chromatography Roman Kubec ∗ , Eva Dadáková ˇ Department of Applied Chemistry, University of South Bohemia, Braniˇsovská 31, 370 05 Ceské Budˇejovice, Czech Republic

a r t i c l e

i n f o

Article history: Received 14 August 2008 Received in revised form 6 October 2008 Accepted 8 October 2008 Available online 14 October 2008 Keywords: Allium Brassica Garlic Onion FMOC Alliin Isoalliin Methiin CE MEKC

a b s t r a c t A novel method for determination of S-alk(en)ylcysteine-S-oxides by capillary electrophoresis has been developed and validated. The method is based on extraction of these sulfur amino acids by methanol, their derivatization by fluorenylmethyl chloroformate and subsequent separation by micellar electrokinetic capillary chromatography. Main advantages of the new method are simplicity, sensitivity, high specificity and very low running costs, making it suitable for routine analysis of a large number of samples. Employing this method, the content of S-alk(en)ylcysteine-S-oxides was determined in 12 commonly consumed alliaceous and cruciferous vegetables (e.g. garlic, onion, leek, chive, cabbage, radish, cauliflower and broccoli). The total content of these amino acids in the Allium species evaluated varied between 0.59 and 12.3 mg g−1 fresh weight. Whereas alliin was found only in garlic, isoalliin was the major S-alk(en)ylcysteine-S-oxide in onion, leek, chive and shallot. On the other hand, the cruciferous species analyzed contained only methiin in the range of 0.06–2.45 mg g−1 fresh weight. © 2008 Elsevier B.V. All rights reserved.

1. Introduction S-Alk(en)ylcysteine-S-oxides are important secondary metabolites occurring in many families of plants, fungi and algae. These sulfur amino acids are precursors of an extraordinary variety of sensory-active and health-beneficial compounds of Allium and Brassica vegetables (e.g. garlic, onion, leek and cabbage, broccoli, kohlrabi, etc.). Six S-alk(en)ylcysteine-S-oxides have so far been found in alliaceous plants, namely S-methyl-, S-allyl-, (E)S-(1-propenyl)-, S-propyl-, S-ethyl- and S-butylcysteine-S-oxides (methiin, alliin, isoalliin, propiin, ethiin and butiin, respectively, 1–6) (Fig. 1) [1–9]. On the other hand, cruciferous plants typically contain only methiin (1) with traces of ethiin (5) [10–12]. Disruption of the plant tissue results in the release of a C-S lyase and subsequent enzymatic cleavage of S-alk(en)ylcysteine-S-oxides to form thiosulfinates [RS(O)SR ], the flavor principles of freshly comminuted Allium vegetables. Numerous methods for quantitative determination of Salk(en)ylcysteine-S-oxides (1–6) have been developed. A leading

∗ Corresponding author. Tel.: +420 387 772 664; fax: +420 385 310 405. E-mail addresses: [email protected], [email protected] (R. Kubec). 0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.10.024

role among these methods plays HPLC determination after precolumn derivatization, with ortho-phthaldialdehyde (OPA)/tertbutylthiol being the most frequently used derivatization reagent [6]. Alternatively, S-alk(en)ylcysteine-S-oxides can be quantified by GC after derivatization with ethyl chloroformate and reduction of the thermolabile sulfoxide group by sodium iodide [1,2,10]. The GC method allows very sensitive determination in combination with the convenient possibility to verify the identity of analytes by mass-spectrometry. Surprisingly, only one method has thus far been published for determination of methiin (1) and alliin (2) by capillary electrophoresis (CE) [11]. However, no information was given regarding the applicability of this CE procedure to the analysis of other very important cysteine derivatives, namely isoalliin (3) and propiin (4). Quantification of isoalliin in both onion and garlic is of particular importance, as this amino acid is the key compound directly affecting the pungency of onion and the tendency of garlic to undesirable discoloration [13]. Thus, the main aim of this study was to develop a simple method allowing rapid, sensitive and reproducible determination of the whole range of S-alk(en)ylcysteine-S-oxides by means of capillary electrophoresis. Such a method would be suitable not only for analysis of fresh vegetables but also for evaluation of various products made of garlic and onion (e.g. garlic-based supplements, spices,

R. Kubec, E. Dadáková / J. Chromatogr. A 1212 (2008) 154–157

155

2.4. Isolation and derivatization procedure

Fig. 1. S-Alk(en)ylcysteine-S-oxides in Allium species.

etc.). The emphasis was put on the simplicity of isolation, clean-up and derivatization steps to ensure low financial costs per sample and general applicability of the method. 2. Experimental 2.1. Reagents and materials Chemicals were obtained from the Sigma–Aldrich group (St. Louis, MO, USA) and Spolana (Neratovice, Czech Republic). HPLCgrade solvents (methanol and acetonitrile) were purchased from LabScan (Dublin, Ireland). Garlic (Allium sativum L., China), yellow onion (Allium cepa L., Netherlands), chive (Allium schoenoprasum L., Czech Republic), leek (Allium ampeloprasum var. porrum L., Netherlands), shallot (Allium ascalonicum auct., Netherlands), cabbage (Brassica oleracea L. convar. capitata var. alba, Czech Republic), Pekingese cabbage (Brassica pekinensis Lour., Czech Republic), broccoli (B. oleracea DC. var. asparagoides, Spain), kohlrabi (B. oleracea L. var. gongylodes, Czech Republic), cauliflower (B. oleracea L. var. botrytis, Italy), radish (Raphanus sativus L. var. radicula, Czech Republic) and white radish (R. sativus DC. subsp. niger var. albus, Italy) were purchased from a local market in June 2008. 2.2. Reference compounds S-Substituted cysteines (S-methyl-, S-ethyl-, S-propyl-, S-allyl-, S-butyl- and S-isobutyl-l-cysteines) and diastereomeric mixtures of the corresponding S-alk(en)yl-l-cysteine-S-oxides were synthesized by methods described in [1]. The naturally occurring (SS ,RC )-diastereomers of alliin (ACSO, 2) and propiin (PCSO, 4) were obtained by repeated recrystallizations from aqueous acetone or ethanol, respectively. Isoalliin (PeCSO, 3) was isolated from white onion according to the procedure of Carson et al. [14]. The identity and purity (≥98%) of the reference compounds were checked by 1 H and 13 C NMR, HPLC and TLC. 2.3. Apparatus and methods Analyses were carried out on a fully automated system Spectraphoresis 2000, equipped with a UV–Vis scanning detector (Thermo Separation Products, Fremont, CA, USA). Separations were performed using a fused-silica capillary (70 cm × 75 ␮m I.D., Supelco; the effective length to the detector was 67 cm). Injections were achieved by application of vacuum for 2 s. The detection wavelength was set as at 265 nm. The separation buffer (pH 9.2) consisted of 20 mM sodium tetraborate, 20 mM sodium dodecyl sulfate (SDS) and 10% (v/v) MeOH. The applied voltage of +20 kV resulted in an electrophoretic current of 30 ␮A. The temperature around the capillary was maintained constant at 25 ◦ C.

The amino acids were extracted from fresh vegetables by 90% aqueous methanol containing 10 mM HCl [1–3]. Typically, about 10 g of carefully peeled garlic cloves were homogenized in 150 ml of acidified methanol by using a tissue homogenizer. The homogenate was allowed to gently boil under reflux for 5 min, filtered and repeatedly extracted with another 150-ml portion of boiling methanol. The combined methanolic extracts were reduced (at 40 ◦ C) to approximately 10–15 ml and adjusted to 25 ml by 20 mM borate buffer (pH 9.2). The extract was filtered with a 0.45␮m nylon filter and an aliquot of 100 ␮l was mixed with 150 ␮l of fluorenylmethyl chloroformate (FMOC-Cl; 10 mM in MeCN) and 0.75 ml of the borate buffer. The mixture was briefly shaken, allowed to stand at room temperature for 5 min and extracted by 1 ml of pentane. After clearing the layers, the aqueous (the lower) one was analyzed. The quantification was done relative to the internal standard of S-isobutylcysteine-S-oxide (i-BCSO, 20 mg ml−1 ) which was added prior to sample homogenization. All samples were analyzed in triplicate. Calibration curves for the analytes were generated using solutions prepared from the synthesized/isolated standards.

3. Results and discussion 3.1. Method development Initially, we attempted to adopt the recently published CE procedure of Horie and Yamashita [11]. They developed a capillaryelectrophoretic method that is based on indirect detection of methiin (1) and alliin (2). Using this method, they determined the content of these two amino acids in garlic and a few other Allium and Brassica vegetables. After a few slight modifications of the original procedure, we achieved quite good separation of a mixture of the standards. However, this method proved to offer absolutely unsatisfactory results for analyzing real sample extracts. The capillary deteriorated rapidly after a few injections, resulting in loosing the separation capability, even if the capillary was rinsed thoroughly after every run. Moreover, significant migration time variations and peak overlapping were observed, rendering identification and quantification of individual peaks extremely unreliable. Therefore, we turned our attention to developing a completely novel method. We decided to test the applicability of the FMOC-Cl derivatization procedure for CE analysis of compounds 1–6. FMOC-Cl is known to readily form stable derivatives with compounds possessing primary and secondary amino groups in virtually quantitative yield [15]. The reaction proceeds in aqueous solutions within minutes with no time-consuming clean-up procedure required. Moreover, the derivatives exhibit very high extinction coefficients, allowing their sensitive and specific detection. Thus, it is not surprising that FMOC-Cl belongs to the most frequently used derivatization reagents for analysis of amino acids in combination with HPLC or CE. In fact, FMOC-Cl was already successfully employed for analysis of S-alk(en)ylcysteine-S-oxides in Allium plants by HPLC [4]. When looking for optimal separation conditions, we tested various concentrations of sodium dodecyl sulfate (SDS, 0–30 mM) and MeOH (0–15%) in the running borate buffer. The FMOC-tagged derivatives were found to be optimally resolved in a system consisting of 20 mM borate buffer (pH 9.2), 20 mM SDS and 10% MeOH. Under these conditions, all compounds of interest were satisfactorily resolved within 20 min. The individual diastereomers of synthetically prepared S-alk(en)yl-l-cysteine-S-oxides were easily separable, with the naturally occurring (+)-isomers migrating

156

R. Kubec, E. Dadáková / J. Chromatogr. A 1212 (2008) 154–157

an internal standard. However, this amino acid was subsequently found to occur abundantly in Allium siculum and at trace levels in onion [7]. Therefore, we decided to use S-isobutylcysteine-S-oxide (i-BCSO) which was chosen for a number of reasons. Although iBCSO is not commercially available, it can be easily prepared in high yield and purity. Its thermal stability, solubility and chemical behavior are very similar to those of the analytes. It can also serve as a substrate for alliinase, thus possible enzymatic cleavage of the analytes during the homogenization and extraction steps is partially corrected. Furthermore, the peaks of the individual i-BCSO diastereomers do not coincide with any other compounds usually found in the extracts.

3.2. Method validation

Fig. 2. Electropherograms of garlic (above) and shallot (below) extracts.

faster. Thus, the diastereomers of S-alk(en)yl-l-cysteine-S-oxides exhibit a reverse migration order compared to their elution from a C-18 HPLC column. A great deal of attention was also paid to finding a suitable internal standard which would reduce potential errors occurring during the extraction, derivatization and injection steps. In our previous studies [1,2,10], we used S-butylcysteine-S-oxide (BCSO, 6) as

The detector response was found to be linear (r2 ≥ 0.99) in the range of 0.02–2.0 mM for all analyzed compounds. The limit of detection (LOD) of FMOC-tagged alliin (calculated as a signalto-noise ratio of 2) was found to be approximately 0.2 pmol. Although LOD values of the other compounds were not determined, it can be assumed they would be similar to that of alliin. The stability of FMOC–alliin upon storage at room temperature was evaluated by repeated injections over a 48-h period. No statistically significant decrease in the peak area was observed after 24 h and only a 4%-decline was observed after 48 h. Thanks to this high stability of FMOC derivatives, large sets of samples can be prepared at once, placed in an autosampler and analyzed within 24 h without compromising the accuracy of results. The FMOC-Cl reagent was found to be stable upon storage in the refrigerator for at least 2 weeks, as no statistically significant differences were observed in derivatization of a standard sample of alliin. Recoveries of the analytes were determined by adding two different concentrations of each organosulfur compound (1–4) to garlic before homogenization. Recoveries of 96–101% were obtained, with a mean relative standard deviation (RSD) of 4.3%.

Table 1 The content of S-substituted cysteine derivatives in Allium species (in mg g−1 fresh weight). Species

Garlic

Onion

Shallot

Leek

Chive

a b c d e f

Part analyzed

Relative proportion (%)

Total content

Ref.

12.3 ± 0.2 5.3–12.2 3.65 11.8

b

1

2

3

4

10 6–11 17 5

81 89–94 83 84

9 trc tr 11

n.d.a tr. n.d. n.d.

18 13–16 7–15

n.d. n.d. n.d.

82 84–87 85–93

n.d. n.d. n.d.

0.59 ± 0.04 1.13–1.55 0.92–2.50

b

19 27–44 5

tr 1–5 n.d.

81 40–59 95

n.d. 7–11 n.d.

1.35 ± 0.11 1.56 2.27

b

Aerial part

12 19 27 8

tr tr n.d. n.d.

88 81 73 92

tr tr n.d. n.d.

1.53 ± 0.09 0.22 1.04 2.48

b

Leaves

22 44 48

tr 3 6

78 43 46

tr 9 n.d.

2.45 ± 0.08 0.72 1.42

b

Cloves

Bulbs

Bulbs

n.d., not detected. Present study. tr, traces. GC determination. HPLC/FMOC-Cl determination. HPLC/dansyl-Cl determination.

[1]d [4]e [9]f [4] [9] [1] [9] [1] [4] [9] [1] [9]

R. Kubec, E. Dadáková / J. Chromatogr. A 1212 (2008) 154–157 Table 2 The content of methiin (1) in Brassica and Raphanus species (in mg g−1 fresh weight). Species

White cabbage Pekingese cabbage Cauliflower Broccoli Kohlrabi Radish White radish a b c

(1)

1.17 0.45 2.85 1.67 1.49 0.45 0.39

RSD (%)

± ± ± ± ± ± ±

0.07 0.04 0.26 0.07 0.02 0.02 0.01

5.7 9.0 9.2 4.4 1.0 4.5 3.2

Literature data [1]a

[11]b

[15]c

0.67 0.21 2.45 2.08 0.80 0.064 0.33

1.29

1.0

1.35–1.85

1.22 1.98 0.27

GC determination. CE determination. Amino acid analyzer determination.

3.3. Analysis of Allium species Five different Allium species were analyzed in total (garlic, onion, leek, chive and shallot). These species represent the most commonly consumed alliaceous vegetables worldwide. Typical electropherograms of garlic and shallot extracts are shown in Fig. 2. The total content of S-alk(en)ylcysteine-S-oxides (1–4) varied considerably within a wide range of 0.59–12.3 mg g−1 fresh weight (Table 1). Relative standard deviations observed ranged between 1.6% and 8.1%, with a mean of 5.2%. Whereas alliin (2) was found only in garlic, isoalliin (3) was the major S-alk(en)ylcysteine-Soxide in onion, leek, chive and shallot. Methiin (1) was present in all five Allium species as a second most abundant derivative. On the other hand, propiin (S-propylcysteine-S-oxide, 4) was not detected in substantial amounts in any of the samples evaluated. Also ethiin (S-ethylcysteine-S-oxide, 5) and butiin (S-butylcysteineS-oxide, 6), minor Allium flavor precursors, could not be reliably quantified, even though they could have been present in some samples analyzed. In general, the values found in the present study are consistent with data reported previously [1,2,4,7,9–11,16] (Table 1). Some discrepancies observed between our results and the literature data can be primarily attributed to differences in variety, agricultural factors (climatic conditions, soil composition, irrigation, fertilization, harvest date, etc.) and post-harvest handling [5,16,17]. 3.4. Analysis of Brassicaceae species To prove the broad applicability of the method, we also employed it for analysis of S-substituted cysteine derivatives in plants of the Brassicaceae (formerly Cruciferae) family. Seven commonly consumed vegetables belonging to the genera Brassica (e.g. cabbage, cauliflower or broccoli) and Raphanus (e.g. radish) were analyzed in total. It is well known that the S-substituted cysteine derivative pool in cruciferous plants is much simpler compared with that in Allium species. The former species usually contain only methiin (1), occasionally accompanied by traces of ethiin (5) [10–12]. Indeed, our current data confirm the previous findings, as methiin was the only S-alk(en)ylcysteine-S-oxide present in significant quantities in the samples analyzed. As summarized in Table 2, the content of methiin varied between 0.39 and 2.85 mg g−1 fresh weight, being the highest in cauliflower. The SDs ranged between

157

1.0% and 9.2%, with a mean of 5.3%. The data obtained are in very good agreement with those reported in the literature previously (Table 2). Again, some differences observed between our results and the literature data can be primarily attributed to different agricultural conditions during sample growth. 4. Conclusions Compared with the previously published CE determination [11], our method is significantly more sensitive and specific, allowing simultaneous analysis of the whole range of S-alk(en)ylcysteine-Soxides occurring in Allium species. This procedure seems to be a viable alternative to the most commonly employed HPLC methods. The data obtained using the CE method are in very good agreement with those reported in the literature previously. Main advantages of this rapid method are its simplicity, high specificity, sensitivity and very low running costs, making it suitable for routine analysis of large numbers of samples. One run takes only 20 min compared with a typical 40–50 min HPLC or GC analysis. On the other hand, migration time variations belong to its major drawbacks (intraday RSD of migration times of 1–3 varied between 0.1% and 0.4%). If the CE system is not equipped with a MS detector, verification of peak identities should occasionally be made by addition of standards to the sample. The method is universally applicable not only for analysis of fresh vegetables but also for evaluation of various products made of garlic and onion (e.g. garlic-based supplements, spices, etc.). Acknowledgements Financial support provided by the Ministry of Education of the Czech Republic (grants OC123–COST 924 and OC126–COST 927) is greatly appreciated. The authors also thank Dr. Jan Schraml (Institute of Chemical Processes, Czech Academy of Sciences, Prague) for conducting NMR analyses. References [1] R. Kubec, M. Svobodová, J. Velíˇsek, J. Chromatogr. A 862 (1999) 85. [2] R. Kubec, M. Svobodová, J. Velíˇsek, J. Agric. Food Chem. 48 (2000) 428. [3] M. Ichikawa, N. Ide, J. Yoshida, H. Yamaguchi, K. Ono, J. Agric. Food Chem. 54 (2006) 1535. [4] D.J. Thomas, K.L. Parkin, J. Agric. Food Chem. 42 (1994) 1632. [5] W.M. Randle, J.E. Lancaster, M.L. Shaw, K.H. Sutton, R.L. Hay, M.L. Bussard, J. Am. Soc. Hort. Sci. 120 (1995) 1075. [6] I. Krest, J. Glodek, M. Keusgen, J. Agric. Food Chem. 48 (2000) 3753. [7] R. Kubec, S. Kim, D.M. McKeon, R.A. Musah, J. Nat. Prod. 65 (2002) 960. [8] I. Arnault, J.P. Christidès, N. Mandon, T. Haffner, R. Kahane, J. Auger, J. Chromatogr. A 991 (2003) 69. [9] K.S. Yoo, L.M. Pike, Sci. Hortic. 75 (1998) 1. [10] R. Kubec, M. Svobodová, J. Velíˇsek, Eur. Food Res. Technol. 213 (2001) 386. [11] H. Horie, K. Yamashita, J. Chromatogr. A 1132 (2006) 337. [12] Y.K. Nakamura, T. Matsuo, K. Shimoi, Y. Nakamura, I. Tomita, Biosci. Biotechnol. Biochem. 60 (1996) 1439. [13] R. Kubec, M. Hrbáˇcová, R.A. Musah, J. Velíˇsek, J. Agric. Food Chem. 52 (2004) 5089. [14] J.F. Carson, R.E. Lundin, T.M. Lukes, J. Org. Chem. 31 (1966) 1634. [15] D. Melucci, M. Xie, P. Reschiglian, G. Torsi, Chromatographia 49 (1999) 317. [16] D.E. Kopsell, W.M. Randle, M.A. Eiteman, J. Am. Soc. Hort. Sci. 124 (1999) 177. [17] M. Ichikawa, N. Ide, K. Ono, J. Agric. Food Chem. 54 (2006) 4848.