Selective and Sensitive LC-MS Determination of ... - Springer Link

Selective and Sensitive LC-MS Determination of Avenanthramides in Oats

2006, 63, 419–423

J. Jastrebova&, M. Skoglund, J. Nilsson, L. H. Dimberg Department of Food Science, Swedish University of Agricultural Sciences, P.O. Box 7051, 75007, Uppsala, Sweden; E-Mail: [email protected]

Received: 17 January 2006 / Revised: 3 March 2006 / Accepted: 6 March 2006 Online publication: 26 April 2006

of avenanthramides is clustered in oat groats especially in the outer parts, i.e. the bran and pearling fractions, whereas oat hulls contain much lower levels of these compounds [3, 4, 6, 16]. On the basis of chemical structure, avenanthramides represent amides of different cinnamic acids with different anthranilic acids. They can be divided into several groups according to which derivative of anthranilic acids (anthranilic, 5-hydroxy-anthranilic, 5-hydroxy4-methoxy-anthranilic or 4-hydroxyanthranilic acid) is included in the structure. Amides of different cinnamic acids with anthranilic acid comprise

group 1 and amides of cinnamic acids with 5-hydroxy-anthranilic acid comprise group 2. The chemical structure of avenanthramides relating to the group 2 is presented on Fig. 1. The abbreviations 2p, 2c and 2f are used for amides of 5-hydroxy-anthranilic acid (2) and p-coumaric (p), caffeic (c) and ferulic acid (f), respectively. These avenanthramides have been found to be the dominant avenanthramide forms in oat extracts [10]. Since the 1980s, several liquid chromatographic (HPLC) methods using both ion-exchange and reversed-phase chromatography with UV detection have been developed for separation and quantification of avenanthramides [2–5, 13, 15]. However, low concentrations of avenanthramides, poor chromatographic separation as well as interferences from matrices make this analysis very difficult. Despite continuous development of HPLC methods for determining avenanthramides, there is still a need for better detection limits and higher selectivity when analysing complex matrices containing small amounts of different avenanthramides. Furthermore, much remains to be done to isolate and identify unknown avenanthramide forms occurring in oats at trace levels. The similarity of the UV spectra often hinders accurate identification of these compounds. The use of mass spectrometric detection with higher selectivity and sensitivity can be one of the solutions to these problems. Recently, an LC-MS method has been used for determination of stable isotope

Chromatographia 2006, 63, May (No. 9/10)

419

Abstract A new liquid chromatographic – mass spectrometric (LC-MS) method suitable for the determination of avenanthramides in oats has been developed. The avenanthramides were detected using positive ion electrospray with selective ion monitoring of protonated ions [M+H]+. The separation of avenanthramides was achieved on Genesis C18 column by using acetonitrile – aqueous formic acid as mobile phase and gradient elution. The effects of buffer nature and concentration on separation, peak shape and intensity of mass spectrometric signal were investigated. The detection limits ranged between 10 and 30 ng mL)1 for standard solutions when using 1 lL injection. The developed LC-MS method was found to be superior over commonly used liquid chromatographic methods with UV detection in terms of selectivity and sensitivity. The applicability of the method to real samples (oat groats) was demonstrated.

Keywords Column liquid chromatography - mass spectrometry Avenanthramides in oats

Introduction Avenanthramides are a group of substituted N-cinnamoylanthranilic acids unique for oats among cereals [1–7]. They are related to fresh appearance of oats [8] and have been characterized as antioxidants in vitro [3, 9, 10] and in vivo [11, 12]. There are also indications in the literature that they exert other biological activities related to e.g. allergy and cardiovascular diseases [12–14]. Avenanthramides are present in oats in rather small amounts; total concentrations in the range 3–289 mg kg)1 have been reported [3–7, 10, 15]. A major part

Original DOI: 10.1365/s10337-006-0769-y 0009-5893/06/05

2006 Friedr. Vieweg & Sohn/GWV Fachverlage GmbH

HO

was extracted and analysed as described below.

O R

NH CO 2 H 5-hydroxyanthranilic acid

OH cinnamic acid

Avenanthramide Abbr N-(4´-hydroxy)-(E)-cinnamoyl-52p hydroxy-anthranilic acid N-(3´,4´-dihydroxy)-(E)-cinnamoyl-52c hydroxy-anthranilic acid N-(4´-hydroxy-3´-methoxy)-(E)2f cinnamoyl-5-hydroxy-anthranilic acid

R H OH OCH3

Fig. 1. Chemical structure of avenanthramides derived from 5-hydroxy-anthranilic acid

labelled avenanthramides in aqueous solutions [17]. The aim of this study was to develop an LC-MS method suitable for detection and quantification of avenanthramides in oat samples.

Experimental Chemicals and Materials Acetonitrile was of gradient grade for HPLC. Other reagents were of p.a. grade. Water was purified using a Milli-Q system (Millipore, USA). All other chemicals were purchased from Merck (Darmstadt, Germany). Synthetic avenanthramides, N-(3¢,4¢dihydroxy)-(E)-cinnamoyl-5-hydroxyanthranilic acid (2c), N-(4¢-hydroxy3¢-methoxy)-(E)-cinnamoyl-5-hydroxyanthranilic acid (2f) and N-(4¢-hydroxy)(E)-cinnamoyl-5-hydroxy-anthranilic acid (2p), were kindly provided by Dr. K. Sunnerheim (Department of Chemistry, Uppsala University, Sweden) and stored at ambient temperature until use. The synthesis of these compounds was performed according to Collins [1] and Bratt et al [10]. The stock solutions (100 lg mL)1) were prepared in methanol. Aliquots of the stock solutions were placed in separate tubes and stored below )20 C until use. The calibration solutions were prepared immediately before use by dilution of the stock solution with methanol. Oat groats were freeze-dried and stored at ambient temperature. Prior to analysis the samples were milled using an ultracentrifugal mill type ZM 1 (Retsch, Haan, Germany) to a particle size £ 0.5 mm. The obtained oat flour

420

Sample Preparation Avenanthramides were extracted from oat flour (1 g) at 30 C under vertical shaking by using ethanol. Extraction was performed three times by adding 10 ml of 80% aqueous ethanol on each step. Extracts were pooled and dried under vacuum at 40 C. The residuals were reconstituted in 20 ml methanol and centrifuged. The obtained samples were analyzed with LC-MS as described below.

LC-MS Analyses were performed using an Agilent 1100 LC/MSD system equipped with a gradient quaternary pump, a thermostated autosampler, a thermostated column compartment, a variable UV detector and a single quadrupole mass analyzer (G1946D). An Agilent Chemstation software was used for control of the HPLC system and data processing. The separation of avenanthramides was performed by using a reversed-phase column Genesis C18, 150 mm*4,6 mm i.d.; 4 lm (Jones Chromatography, UK) with a guard column Opti-guard C18, 1 mm (Optimize Technologies, INC, USA) at 23 C. The flow-rate was 0.4 mL min)1; the injection volume 1–2 lL; the temperature in the thermostated autosampler 8 C. The UV detector was set at 340 nm. The mobile phase used was acetonitrile - 10 mM aqueous formic acid under linear gradient elution conditions. The gradient started at 20% (v/v) acetonitrile with a lag of 4 min, then the gradient was raised linearly to 60% acetonitrile during 14 min; thereafter it was decreased linearly during 1 min to the initial acetonitrile concentration that was applied for 8 min in order to re-equilibrate the column. The column effluent was diverted to waste during the first 15 minutes after injection and then diverted to mass detector for the last 12 minutes of the run. When developing the method, other buffers (5 mM and 10 mM aqueous acetic acid) were tested for preparation of the mobile phase; for these tests the same acetonitrile gradient from 20 to 60% as described above was applied. Chromatographia 2006, 63, May (No. 9/10)

Electrospray (ES) ionization was operated in positive ion mode. For optimizing ES parameters flow injection analysis (FIA) was used. A capillary voltage was set at 3.5 kV with drying gas temperature at 350 C and a drying gas flow of 9 L min)1. Nebuliser pressure was 30 psi. Fragmentor potential was set to 90 V for all avenanthramides. Under method development scan mode in the range m/z 100–800 was used. For quantification of avenanthramides selected ion monitoring (SIM) of protonated molecular ions [M+H]+ was used at m/z 300, 316 and 330 for 2p, 2c and 2f, respectively. Automated tuning and mass calibration were performed using an Agilent ES tuning solution. Mass spectral data as well as the retention times of compounds were used for peak identification. Peak widths (xh) were calculated at half-heights. Peak asymmetry (As) was estimated at 10% of the peak height from the ratio of the widths of the rear and front sides of the peak.

Quantification Quantification was based on an external standard method. A multilevel calibration curve forced to the origin was used (n=7). The peak area was plotted against concentration and least-squares regression analysis was used to fit lines to the data. The limit of detection (LOD) was defined as the lowest analyte concentration yielding a signal-to-noise (S/N) ratio of 3. The limit of quantification (LOQ) was defined as the lowest analyte concentration yielding S/N = 10.

Results and Discussion Optimisation of LC-ESI-MS Most HPLC analyses of oat avenanthramides were performed using conventional C18 columns based on classical stationary phases with type A silica such as lBondapak C18 [18] and Hypersil ODS [13, 19]. However, ‘‘new generation’’ of HPLC columns based on ultra-pure type B silica can provide better selectivity and efficiency due to lower silica acidity, less silanol interferences and higher ligand concentration [20–22]. Therefore, in the present study a Genesis C18 column based Original

Original

1,5 8 1 4

Peak width

2

12 Asymmetry factor

on ultra-pure type B silica was used for separation of avenanthramides 2p, 2c and 2f, all derived from 5-hydroxy-anthranilic acid. Since avenanthramides are carboxylic acids, acidic mobile phases (phosphate or acetate buffered) are commonly used to suppress ionisation of carboxylic groups and to achieve good separation [13, 18– 19]. In the present study, acetic acid (5 and 10 mM) and formic acid (10 mM) were compared as aqueous LC-MS compatible buffers for separation of avenanthramides when using linear gradient of acetonitrile. The retention of avenanthramides was only slightly influenced by different buffers; the baseline separation was achieved in all cases when using linear acetonitrile gradient from 20 to 60%. The peak shape, however, was strongly dependent on buffer nature, pH and concentration. The weaker acetic acid (pKa=4.76 compared to pKa=3.75 for formic acid) affected adversely the peak shape for all avenanthramides, most notably for derivative 2c (Fig. 2). Acceptable peak shape for 2c (As = 1.9; xh = 0.25 min) could be obtained only when using 10 mM formic acid which has lower pH and higher ion strength in comparison with 5 or 10 mM acetic acid. This buffer provided also good peaks for derivatives 2p and 2f (As £ 1.5; xh £ 0.19 min). Therefore 10 mM formic acid was chosen as LC-MS compatible buffer in this study. However, the impact of different buffers on ionization processes in electrospray varied for different avenanthramides: acetic acid provided slightly higher MS signal intensity for 2p and 2f, whereas formic acid resulted in increase of MS signal intensity for 2c. To optimise the ionization of avenanthramides different operational parameters of ES (positive ion mode) were tested by using flow injection analysis. Because of rather low content of acetonitrile in the mobile phase (20% at the start of gradient) the optimisation of capillary voltage and drying gas temperature and flow was very important in order to provide high MS signal under stable and reproducible ES conditions. Moderate capillary voltage (3.5 kV) in combination with high temperature of drying gas (350 C) was found to provide high reproducible MS signal and stable baseline, whereas high capillary voltage (5.0 kV) in combination with lower temperature of drying gas (300 C) gave noisy baseline and greater variations in

0,5 0

0 A

B

C

Buffer

Fig. 2. The influence of buffer nature and concentration on peak shape of avenanthramide 2c (—n— peak asymmetry factor, –h– peak width). Buffer used for preparation of mobile phase: A – 5 mM acetic acid; B – 10 mM acetic acid; C – 10 mM formic acid. Linear acetonitrile gradient from 20 to 60%, see Experimentals, LC-MS for more details

Fig. 3. Positive ES ionization mass spectra of avenanthramides 2p, 2c and 2f obtained at 90 V. Mobile phase: acetonitrile – 10 mM formic acid, linear acetonitrile gradient from 20 to 60%. For more details about operational ES parameters and chromatographic conditions see Experimentals, LC-MS

Fig. 4. MS and UV chromatograms of standard solution containing 2p 1310 ng mL)1, 2c 950 ng mL)1 and 2f 1070 ng mL)1 in methanol. Injection volume 1 lL. Mobile phase: acetonitrile – 10 mM aqueous formic acid, linear acetonitrile gradient from 20 to 60%; see Experimentals, LCMS for more details

MS signal intensity. These marked differences in reproducibility could be explained by much better conditions for effective stable formation of charged droplets and their desolvation in the first case (350 C, 3.5 kV). Pronounced fragmentation of avenanthramides followed by decrease of MS signal was observed at fragmentor potentials higher than 90 V. On the other hand, lower fragmentor potentials (30–70 V) resulted in slightly decreased MS signal possibly due to insufficient declustering of analyte ions. Therefore, fragmentor potential 90 V was chosen as optimal. Mass spectra of avenanthramides at 90 V are presented in Fig. 3. Protonated

molecular ions [M+H]+ were found to be the most abundant for all three derivatives. These ions with m/z 300, 316 and 330 for 2p, 2c and 2f, respectively, were used for quantification. Sodium adducts [M+Na]+ were found in minor amounts and could be used for additional identification of derivatives.


421

Method Performance A typical chromatogram from a standard mixture is presented in Fig. 4. The MS detector response was linear over the concentration ranges tested for all three derivatives and calibration curves had a

Table 1. Linearity and sensitivity of the methoda Avenanthramide

2p 2c 2f a b

MS detection

UV detection

Linearity range (ng/ml)

Slope of calibration curve (area/concentration)

30–4000 100–2500 30–3500

250±43 130±33 299±53

b

LOD (ng mL)1)

LOQ (ng mL)1)

Correlation coefficient (R2)

10 30 10

30 100 30

0.9996 0.9983 0.9997

b

LOD (ng mL)1)

LOQ (ng mL)1)

30 140 30

100 400 100

Abbreviations: LOD – limit of detection, LOQ – limit of determination, for calculations see Experimental, Quantification. Value represents average for five calibration curves

the developed LC-MS method makes it suitable for analysis of avenanthramides in complex food matrices. Further improvements may be achieved by using LC-MS-MS and stable isotope labelled internal standards as well as sample purification procedures.

Acknowledgements The Swedish Agency for Innovation Systems (Vinnova) is gratefully acknowledged for financial support of this project. The Knut and Alice Wallenberg Foundation is gratefully acknowledged for financial support of the LC-MS equipment. Fig. 5. MS and UV chromatograms of oat extract containing 2p 1410 ng mL)1, 2c 1730 ng mL)1 and 2f 825 ng mL)1 in methanol. Injection volume 1 lL. Mobile phase: acetonitrile – 10 mM formic acid, linear acetonitrile gradient from 20 to 60%, see Experimentals, LC-MS for more details

correlation coefficient higher than 0.998. Linearity and sensitivity data are presented in Table 1. When using 1 lL injection and MS detection, the limits of quantification (LOQ) were 30 ng mL)1 for derivatives 2p and 2f, respectively; whereas for 2c LOQ was 100 ng mL)1 due to lower MS signal intensity and relatively broader peak (see Fig. 4). Comparison of these results with corresponding results for UV detection (Table 1) showed considerable improvement of LOQ (3–4 times) when using MS detection.

Method Applicability to Real Samples To evaluate the suitability of the developed method in measuring avenanthramides in real samples we applied this method to determination of avenanthramides in oat extract. The MS and UV chromatograms of extract of oat groats are presented in Fig. 5. As seen from this Figure, MS detection provided much better sensitivity compared to UV

422

detection and the use of selective ion monitoring for each individual avenanthramide improved selectivity of the method. All three avenanthramides (2p, 2c and 2f) were found in oat groats, the dominating forms were 2c and 2p (44 and 35% of total amount, respectively), whereas 2f occurred in a lesser proportion (21% of total amount). This was in agreement with our previous results for oat groats [6, 7, 10].

Conclusions The present study demonstrated advantages of liquid chromatography coupled with MS detection over routinely used HPLC methods with UV detection for the analysis of avenanthramides. The use of MS detection improved considerably the detection and identification of avenanthramides owing to stable and reproducible positive ion mass spectra for each derivative. The detection limits ranged from 10 to 30 ng mL)1 for standard solutions when using 1 lL injection volume. High sensitivity and selectivity of Chromatographia 2006, 63, May (No. 9/10)

References 1. Collins FW (1989) J Agr Food Chem 37:60–66 DOI 10.1021/jf00085a015 2. Collins FW, Mullin WJ (1988) J Chromatogr 445:363–370 DOI 10.1016/S00219673(01)84548-9 3. Dimberg LH, Theander O, Lingnert H (1993) Cereal Chem 70:637–641 4. Emmons CL, Peterson DM (1999) Cereal Chem 76:902–906 5. Matsukawa T, Isobe T, Ishihara A, Iwamura H (2000) Z Naturforsch C 55:30– 36 6. Bryngelsson S, Mannerstedt-Fogelfors B, Kamal-Eldin A, Andersson R, Dimberg LH (2002) J Sci Food Agric 82:606–614. DOI 10.1002/jsfa.1084 7. Dimberg LH, Molteberg EL, Solheim R, Frolich W (1996) J Cereal Sci 24:263–272. DOI 10.1006/jcrs.1996.0058 8. Molteberg EL, Solheim R, Dimberg LH, Frolich W (1996) J Cereal Sci 24:273–282. DOI 10.1006/jcrs.1996.0059 9. Peterson DM, Hahn MJ, Emmons CL (2002) Food Chem 79:473–478. DOI 10.1016/S0308-8146(02)00219-4 10. Bratt K, Sunnerheim K, Bryngelsson S, Fagerlund A, Engman L, Andersson RE, Dimberg LH (2003) J Agr Food Chem 51:594–600. DOI 10.1021/jf020544f 11. Chen CY, Milbury PE, Kwak HK, Collins FW, Samuel P, Blumberg JB (2004) J Nutr 134:1459–1466 12. Ji LL, Lay D, Chung E, Fu Y, Peterson DM (2003) Nutr Res 23:1579–1590 DOI:10.1016/S0271-5317(03)00165-9

Original

13. Liu LP, Zubik L, Collins FW, Marko M, Meydani M (2004) Atherosclerosis 175:39– 49. DOI 10.1016/j.atherosclerosis.2004.01. 044 14. Nie L, Wise ML, Peterson DM, Meydani M (2005) Atherosclerosis (published online). DOI 10.1016/j.atherosclerosis. 2005.07.027 15. Mattila, P, Pihlava JM, Hellstrom J (2005) J Agr Food Chem 53:8290–8295. DOI 10.1021/jf051437z

Original

16. Emmons CL, Peterson DM, Paul GL (1999) J Agr Food Chem 47:4894–4898. DOI 10.1021/jf990530i 17. Okazaki Y, Isobe T, Iwata Y, Matsukawa T, Matsuda F, Miyagawa H, Ishihara A, Nishioka T, Iwamura H (2004) Plant J 39:560–572. DOI 10.1111/j.1365-313X. 2004.02163.x 18. Peterson DM, Emmons CL, Hibbs AH (2001) J Cereal Sci 33:97–103. DOI 10.1006/jcrs.2000.0347


19. Bryngelsson S, Ishihara A, Dimberg LH (2003) Cereal Chem 80:356–360 20. McCalley DV (1999) J Chromatogr A 844:23–38. DOI 10.1016/S0021-9673(99) 00250-2 21. Gilroy JJ, Dolan JW, Carr PW, Snyder LR (2004) J Chromatogr A 1026:77–89. DOI 10.1016/j.chroma.2003.11.043 22. Gilroy JJ, Dolan JW, Snyder LR (2003) J Chromatogr A 1000:757–778. DOI 10.1016/S0021-9673(03)00512-0

423