(mepiquat) under Maillard reaction conditions

This article was downloaded by: [Dr Richard H. Stadler] On: 11 December 2014, At: 07:14 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Food Additives & Contaminants: Part A Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tfac20

Role of choline and glycine betaine in the formation of N,N-dimethylpiperidinium (mepiquat) under Maillard reaction conditions a

a

b

a

Thomas Bessaire , Adrienne Tarres , Richard H. Stadler & Thierry Delatour a

Nestlé Research Centre, Lausanne 26, Switzerland

b

Nestlé Corporate Quality Management, CO-QM, Vevey, Switzerland Accepted author version posted online: 21 Oct 2014.Published online: 17 Nov 2014.

To cite this article: Thomas Bessaire, Adrienne Tarres, Richard H. Stadler & Thierry Delatour (2014) Role of choline and glycine betaine in the formation of N,N-dimethylpiperidinium (mepiquat) under Maillard reaction conditions, Food Additives & Contaminants: Part A, 31:12, 1949-1958, DOI: 10.1080/19440049.2014.979371 To link to this article: http://dx.doi.org/10.1080/19440049.2014.979371

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions

Food Additives & Contaminants: Part A, 2014 Vol. 31, No. 12, 1949–1958, http://dx.doi.org/10.1080/19440049.2014.979371

Role of choline and glycine betaine in the formation of N,N-dimethylpiperidinium (mepiquat) under Maillard reaction conditions Thomas Bessairea, Adrienne Tarresa, Richard H. Stadlerb and Thierry Delatoura* a

Nestlé Research Centre, Lausanne 26, Switzerland; bNestlé Corporate Quality Management, CO-QM, Vevey, Switzerland

Downloaded by [Dr Richard H. Stadler] at 07:14 11 December 2014

(Received 19 August 2014; accepted 18 October 2014) This study is the first to examine the role of choline and glycine betaine, naturally present in some foods, in particular in cereal grains, to generate N,N-dimethylpiperidinium (mepiquat) under Maillard conditions via transmethylation reactions involving the nucleophile piperidine. The formation of mepiquat and its intermediates piperidine – formed by cyclisation of free lysine in the presence of reducing sugars – and N-methylpiperidine were monitored over time (240°C, up to 180 min) using high-resolution mass spectrometry in a model system comprised of a ternary mixture of lysine/fructose/alkylating agent (choline or betaine). The reaction yield was compared with data recently determined for trigonelline, a known methylation agent present naturally in coffee beans. The role of choline and glycine betaine in nucleophilic displacement reactions was further supported by experiments carried out with stable isotope-labelled precursors (13C- and deuteriumlabelled). The results unequivocally demonstrated that the piperidine ring of mepiquat originates from the carbon chain of lysine, and that either choline or glycine betaine furnishes the N-methyl groups. The kinetics of formation of the corresponding demethylated products of both choline and glycine betaine, N,N-demethyl-2-aminoethanol and N,Ndimethylglycine, respectively, were also determined using high-resolution mass spectrometry. Keywords: mepiquat; choline; glycine betaine; Maillard reaction; methylation; heat treatment; model system; cereals

Introduction Research over the past years has shown that Maillard reaction-type mechanisms under essentially dry conditions can lead to the formation of vinylogous compounds amongst which acrylamide, derived from the amino acid asparagine, has raised particular attention in the past years (Stadler, Blank, Varga, et al. 2002; Zyzak et al. 2003; Goldmann et al. 2009). Other amino acids such as, for example, glutamine, aspartic acid and phenylalanine, have also been studied with regard to their potential to furnish reactive intermediates mostly under dry Maillard reaction conditions (Stadler et al. 2003; Goldmann et al. 2009). In a recent study, Nikolov and Yaylayan have shown that lysine, when subjected to high temperature, can be converted to pent-4-ene-1-amine in a decarboxylative deamination reaction analogous to that leading to acrylamide. Subsequent intra-molecular cyclisation of pent-4-ene-1-amine affords the alkaloid piperidine (Nikolov & Yaylayan 2010; Hammel et al. 2014). Further interest in this lysine rearrangement product was prompted by the discovery that under thermal conditions comparable with those encountered during the roasting of coffee, piperidine can react further via methyl transfer reactions to afford mepiquat (Hammel et al. 2014; Wermann et al. 2014). Mepiquat (Figure 1) is a well-known plant protection product widely used as a plant growth regulator in agriculture, and acts by *Corresponding author. Email: [email protected] © 2014 Taylor & Francis

inhibiting the synthesis of gibberellic acid (Rademacher 2000). This reaction to mepiquat occurs during the roasting of coffee, albeit with a very low yield leading to amounts in most cases below 1 mg kg–1 in the roasted bean (Wermann et al. 2014). This very recent study also identified that trigonelline (1-methylnicotinic acid), an alkaloid that occurs in relatively high amounts in both Coffea arabica (7.7–14.8 mg kg–1) (Koshiro et al. 2006; Lang et al. 2008) and Coffea canephora (5.2–15.5 mg kg–1) (Campa et al. 2004; Koshiro et al. 2006), plays a pivotal role in the transfer of N-methyl substituents to available nucleophiles. In fact, trigonelline has been identified in earlier studies as a good methyl group donor in thermally driven methyl displacement reactions leading to alkylated pyridines (Stadler, Varga, Hau, et al. 2002; Stadler, Varga, Milo, et al. 2002). These earlier and very recent studies indicate that nucleophilic displacement reactions involving intrinsic food constituents may lead to hitherto undiscovered molecules in foods. The reaction pathways are usually minor with very low yield and depend on several factors such as thermal input (time and temperature), pH, moisture and the availability/amount of intrinsic precursor molecules, i.e. nucleophiles and alkylating agents. A preliminary study conducted in our laboratory indicates that mepiquat can be formed in barley under typical industrial roasting conditions, albeit in very low amounts


1950

T. Bessaire et al.

Figure 1. Chemical structures and theoretical exact masses (Mth) of methylation agents (upper line), piperidine derivatives (middle line) and decomposition products of choline and betaine (bottom line).

(Wermann et al. 2014). This is a surprising result, because the levels of trigonelline in barley (around 0.25 mg kg–1) are several orders of magnitude lower than in green coffee beans (Corol et al. 2012). We therefore decided to investigate other candidate molecules that may participate in methyl rearrangement reactions particularly in cereal grains. This report now describes the formation of mepiquat in model systems conducted under typical Maillard reaction conditions, following the kinetics of formation of mepiquat over time by liquid chromatography-high-resolution mass spectrometry (LC-HRMS). The abundances of the key intermediates piperidine and N-methylpiperidine, as well as the demethylated congeners of choline and glycine betaine, namely N,N-dimethyl-2-aminoethanol and N,Ndimethylglycine, respectively, are also monitored.

N,N-dimethylglycine (> 99%) and N,N-dimethyl-2-aminoethanol (> 99.5%) were purchased from Sigma Aldrich (Buchs, Switzerland). D(–)-fructose was obtained from VWR International (Dietikon, Switzerland). Isotope-labelled L-lysine-13C6-dihydrochloride (isotopic purity >99%), 2-hydroxy-N,N,N-d9-trimethylethanaminium chloride (d9-choline, isotopic purity >98%) and carboxy-N, N,N-d9-trimethyl-d2-methanaminium (d11-betaine, isotopic purity >98%) were from Cambridge Isotope Laboratories (Andover, MA, USA). N-d3-Methyl,N-methylpiperidinium iodide (d3-mepiquat, isotopic purity >98%) was purchased from LGC (Wesel, Germany). Acetonitrile (Optima™ LC/MS) and water (Optima™ LC/MS) were from Fisher Scientific (Reinach, Switzerland). Methanol (gradient grade for LC) and ammonium acetate were obtained from Merck (Darmstadt, Germany).

Materials and methods Chemicals and reagents

Internal standard solution

The amino acid L-lysine (> 98%), betaine hydrochloride (glycine betaine, > 99%), choline chloride (> 99%), trigonelline hydrochloride (> 98%), piperidine (> 99%), N,N-dimethylpiperidinium chloride (mepiquat chloride),

d3-Mepiquat, used as internal standard (IS), was available in ready-to-use ampoule at 100 µg ml–1 in deuterium oxide. A working solution was prepared in methanol at a concentration of 1 µg ml–1 and stored at –20°C. Before use, this solution was brought to RT and thoroughly shaken.

Food Additives & Contaminants: Part A


Model systems and sample preparation Typically, either 114 mg of trigonelline hydrochloride or 116 mg of choline chloride or 128 mg of betaine hydrochloride were added to 122 mg of lysine and 150 mg of fructose and mixed into 100 μl of water prior to being heated in a tightly closed pyrolysis tube. Tubes were heated at 240°C in a temperature-controlled oil bath from 0 to 180 min. Controls (thermal reaction of a single compound or by successive omission of a single compound from the ternary model mixtures) were heated at 240°C for 90 min. The stable isotope labelling study was performed at 240°C for 90 min. Typically, 107 mg of d11-betaine or 124 mg of d9-choline were incubated with 122 mg of lysine and 150 mg of fructose and mixed into 100 μl of water. Experiments performed with labelled lysine were scaled down as follows: 50 mg of 13C6-lysine, 40 mg of fructose and 34 mg of betaine hydrochloride or 31 mg of choline chloride. The 100 µl of water were replaced by 56 µl of sodium hydroxide (32%) to ensure the same pH than in the experiments carried out with non-labelled lysine. At the end of the reaction the pyrolysis tubes were cooled at RT and 10 ml of a mixture methanol/water (50/50, v/v) were added. After shaking on a vortex and sonication, a 10 µl aliquot of the resulting mixture was sampled and diluted into 990 µl of methanol. For quantification experiments, the 10 µl were diluted into 970 µl of methanol and 20 µl of mepiquat d3 (1 µg ml–1). The resulting mixture was filtered (Nylon, 0.22 µm, VWR) prior to analysis by LC-HRMS.

Quantification of mepiquat and reaction yield Quantification of mepiquat was performed by isotopic dilution using d3-labelled mepiquat as IS. A nine-point calibration curve was prepared in methanol with nonlabelled mepiquat concentrations ranging from 0 to 200 ng ml–1 and a fixed IS concentration at 20 ng ml–1 (82 µmol ml–1). The reaction yield was expressed based on the conversion of the methylation agent, assuming that two moles of the methyl donor (choline, betaine or trigonelline) afford one mole of mepiquat.

Liquid chromatography-high-resolution mass spectrometry (LC-HRMS) conditions HPLC analysis was performed on a hydrophilic–lipophilic (HILIC) column XbridgeTM BEH HILIC 2.5 μm, 2.1 × 100 mm (Waters, Dublin, Ireland) using a Transcend System binary pump device from Flux Instruments AG (Basel, Switzerland) and an autosampler HTS PAL System from CTC Analytics AG (Zwingen, Switzerland). The

1951

mobile phase was comprised of acetonitrile/ammonium acetate 10 mM, 95/5 (solvent A) and acetonitrile/ammonium acetate 10 mM, 50/50 (solvent B). The gradient applied was: from 0 to 6.0 min, a linear gradient from 100% solvent A to a mixture 50/50 (ratio A/B); from 6.0 to 8.0 min, a linear gradient from a solvent A/B mixture at 50/50 to 100% solvent B; from 8.0 to 8.5 min, mobile phase at 100% solvent B. The column was then equilibrated at 100% solvent A from 8.5 to 14.5 min. The column temperature was set at 40° C with a solvent flow rate of 500 μl min–1. A volume of 5 μl of sample was injected onto the column for LC-HRMS analysis. Under these conditions, the retention times of the compounds were: 5.0, 5.1 and 5.3 min for piperidine, Nmethylpiperidine and mepiquat, respectively. The retention times of trigonelline, betaine and choline were 4.6, 4.4 and 5.4 min, and those of N,N-dimethylglycine and N,Ndimethyl-2-aminoethanol were 4.4 and 5.1 min respectively. The detection was carried out on a Q-Exactive Orbitrap mass spectrometer from Thermo Fisher Scientific (Bremen, Germany) equipped with an electrospray ionisation source HESI-II (Heated ElectroSpray Ionisation generation II) operating in positive ionisation mode. The ionisation parameters were: voltage: 3.5 kV; sheath gas: 65 arbitrary units; auxiliary gas: 20 arbitrary units; capillary temperature: 380°C; heater temperature: 500°C; and S-lens RF level: 40. Acquisitions were recorded either in full scan (Full MS) or product ion (Target MS2) modes. The acquisition time was 9 min in the full-scan mode with a mass range defined at m/z 50– 500. The resolution was set at 35 000 and the automatic gain control (AGC) target set at 106. The product ion mode was applied in the time period 4–6 min with targeted precursors at m/z 90.091340 or m/z 104.070605, with a resolution at 17 500 and an isolation window at m/z 4. The AGC target was set at 2 × 105 and the normalised collision energy (NCE) at 35. The exact masses of the compounds under investigation were calculated with the following elemental and electron exact masses: 12C: 12.000000; 13C: 13.003355; 1 H: 1.007825; D (2H): 2.014102; 16O: 15.994915; 14N: 14.003074; and electron: 0.000549 (Peiser et al. 1984; Ferrer & Thurman 2007). The single-exact mass chromatographic profiles were extracted from the full-scan recorded data with a mass window set at 10 ppm. The instrument was calibrated with a mass accuracy < 2 ppm. Theoretical masses are reported with six digits after the decimal place, whilst four digits were reported for the measured masses. Results and discussion We chose choline and betaine, two quaternary ammonium compounds that are reported to occur at relatively high amounts in different cereal grains (Corol et al. 2012), to study nucleophilic displacement reactions under pyrolytic

1952

T. Bessaire et al.


conditions in model systems. In the case of wheat, particularly high amounts of betaine are found in whole grain cereals versus refined wheat (Ross et al. 2014). The other limiting compound in the chemical reaction pathway to mepiquat is free lysine, an essential amino acid for humans that is present in the free form typically in the range 5–50 mg/100 g (Hamlet al. 2007). Higher amounts of free lysine have been recorded in rye flours compared with wheat across several cultivars (Hamlet et al. 2007). Moreover, lysine is not stable and is rapidly depleted (10– 30%) during the baking process (Warthesen & Kramer 1978). Analytical conditions and identification of relevant compounds The chromatographic separation of the compounds under investigation was achieved with an HILIC stationary phase column that enabled adequate resolution of highly polar low molecular weight compounds. The mass spectrum acquisitions were carried out in full-scan mode in order to detect masses of particular interest corresponding to relevant intermediates in the transmethylation steps. Specifically, the presence of N-methylpiperidine (molecular mass at 99.104799 Da) was postulated based upon exact mass extraction at m/z 100.112075 (related to [M + H]+) from the full-scan spectra. The fragmentation pattern of mepiquat was comprehensively studied in previous work (Hammel et al. 2014), and typical product ions at m/z 98.0969 and m/z 59.0660 were observed and structurally assigned. Similarly, the presence of N,Ndimethylglycine and N,N-dimethyl-2-aminoethanol was corroborated by the extraction of the [M + H]+ masses at m/z 104.070605 and m/z 90.091340. The chemical identity of N,N-dimethylglycine and N,N-dimethyl-2-aminoethanol were further confirmed by MS/MS data, and comparison of the retention time with pure standards. Piperidine was confirmed based upon the presence of the signal [M + H]+ at m/z 86.096425, and equivalence of the retention time with a commercial standard. Heat-induced formation of mepiquat Heating ternary mixtures comprised of lysine/fructose in the presence of either choline or betaine lead to a prominent signal at m/z 114.127725 detected at the retention time of 5.3 min (Figure 2). The essential role of lysine in the formation of mepiquat was confirmed by the absence of the signal at m/z 114.127725 when lysine was omitted in the reaction mix. Additional controls supported the proposed mechanism as no or only background levels of mepiquat were detected in the controls, with exception of those mixtures containing lysine and one of the quaternary nitrogen compounds (Figure 3). Indeed, when either trigonelline, choline or betaine were thermally incubated in

the presence of lysine, a detectable signal of mepiquat was observed in the chromatographic trace at m/z 114.127725. This suggests that under sufficient heat lysine itself can undergo intra-molecular cyclisation to piperidine, subsequently methylated in the presence of trigonelline, choline or betaine to give rise to mepiquat. The heat-induced cyclisation of lysine was already mentioned by Nikolov and Yaylayan (2010). It is also feasible that reactive intermediates are procured from the degradation products of trigonelline, betaine and choline; for example, the aminoethanol moiety could be oxidised to an aldehyde that could accelerate the formation of piperidine. Further studies are required to confirm this hypothesis. In a previous study (Hammel et al. 2014) we demonstrated that trigonelline, an alkaloid present in green coffee beans at high levels of up to 15.5 mg g–1 dry weight (Koshiro et al. 2006) acts as a methylation agent during roasting and furnishes N,N-dimethylpiperidinium (mepiquat) from piperidine, the latter in turn generated by Maillard-type cyclisation of lysine in the presence of a reducing sugar (e.g. fructose). A similar mechanism is postulated now for betaine and choline based upon the signals detected in the chromatographic profiles recorded in heat-treated ternary mixtures of lysine/fructose in the presence of either choline or betaine. Indeed, prominent signals at m/z 86.096425 (retention time: 5.0 min) and m/z 100.112075 (retention time: 5.1 min) assigned to piperidine and N-methylpiperidine (single methylation), respectively, were detected as evidenced by the profiles depicted in Figure 2. The yield of mepiquat generated in the presence of either choline or betaine (yield expressed against the initial amount of methylation agent) was monitored 5–180 min (Figure 4), and compared with the yield obtained with a trigonelline-driven model system. Under the conditions applied in the current study, the highest yield obtained with trigonelline was 0.158% after a 2-h incubation at 240°C (120-min time point); a subsequent slight decrease in the yield was observed after 60-min additional thermal treatment (180-min time point). Likewise, a similar decrease in the yield was observed with betaine reaching maximum after 1-h incubation at 240°C (yield at 0.056%). However, choline portrayed a different profile over the time range studied, showing the highest yield (0.214%) after 180 min. Based on the data generated in this model system, choline appears to be the most effective methyl transfer reagent under the chosen experimental conditions.

Role of lysine, choline and betaine in the formation of mepiquat Additional experiments were conducted using stable isotope-labelled materials in order to corroborate the proposed reaction pathway.



1953

Figure 2. Extracted ion chromatograms (m/z window set at 10 ppm) of piperidine, N-methylpiperidine and mepiquat obtained with thermally treated choline or betaine (240°C, 2 h) in the presence of either fructose (A) or lysine plus fructose (B).

Figure 3.

Signal intensity of mepiquat (m/z 114.127725) in heat-treated controls and samples.


1954

T. Bessaire et al.

Figure 4. (colour online) Time-course formation of mepiquat in heat-treated (240°C) mixtures of lysine and fructose in the presence of either trigonelline, betaine or choline. The yield is expressed as conversion of the methylation agent on a molar basis. Entries are averages of quadruplicate determinations.

When fructose was heat-treated with 13C-lysine, the signal of mepiquat was observed with an exact mass at m/z 119.1448 ± 0.0001 irrespective of the methylation agent employed (Table 1). The up-mass shift was found at 5.0166 Da when compared with the theoretical exact mass of mepiquat at m/z 114.127725, corresponding to the substitution of five 12C atoms with five 13C atoms (5 × 1.003355 = 5.016775 Da). This is confirmed by the up-mass shift of piperidine, found at m/z 91.1137 ± 0.0001, and N-methylpiperidine experimentally determined at m/z 105.1293 ± 0.0001, which unequivocally demonstrates that lysine furnishes the carbon backbone

of the nitrogen heterocycles piperidine, N-methylpiperidine and mepiquat. Lysine/fructose mixtures heated in the presence of d9choline (H atoms of the N-methyl groups substituted by deuterium atoms) lead to piperidine, N-methylpiperidine and mepiquat with exact masses at m/z 86.0970, m/z 103.1314 and m/z 120.1657, respectively. As expected, no mass shift was observed for piperidine ([M + H]+ at 86.069425 Da), whilst the tertiary amine N-methylpiperidine displayed a ΔM of +3.0186 and mepiquat of +6.0373 Da, when compared with the measured exact mass of the non-labelled compounds. This observation

Table 1. Exact mass shifts of piperidine, N-methylpiperidine and mepiquat obtained by heat treatment (240°C) of ternary mixtures lysine/fructose/choline or betaine conducted with either 13C-lysine, d9-choline or d11-betaine. Isotopic-labelled standards Betaine

d11-Betaine 13

Choline

Trigonelline

C6-lysine

d9-Choline 13

C6-lysine

13

C6-lysine

Molecules

m/z labelled

m/z non-labelled

Δm/z

Number of labelled atoms incorporated (13C or D)

Piperidine N-methylpiperidine Mepiquat Piperidine N-methylpiperidine Mepiquat Piperidine N-methylpiperidine Mepiquat Piperidine N-methylpiperidine Mepiquat Piperidine N-methylpiperidine Mepiquat

86.0970 103.1314 120.1657 91.1138 105.1294 119.1449 86.0970 103.1314 120.1657 91.1138 105.1294 119.1449 91.1136 105.1292 119.1447

86.0972 100.1128 114.1284 86.0972 100.1128 114.1284 86.0972 100.1128 114.1284 86.0972 100.1128 114.1284 86.0969 100.1124 114.1279

−0.0002 3.0186 6.0373 5.0166 5.0166 5.0165 −0.0002 3.0186 6.0373 5.0166 5.0166 5.0165 5.0167 5.0168 5.0168

0 3 6 5 5 5 0 3 6 5 5 5 5 5 5



can be rationalised in terms of the addition of a single CD3 group onto the nitrogen atom of piperidine leading to d3N-methylpiperidine, and two CD3 groups onto d6mepiquat. Similar results were obtained using d11-betaine (nine deuterium atoms on the methyl groups, and two on the methylene group) as methylation agent, with signals at m/z 86.0970 (as expected piperidine showed no mass shift), m/ z 103.1314 (N-methylpiperidine) and m/z 120.1657 (mepiquat) observed in the full-scan mass chromatographic profile. The up-mass shift of N-methylpiperidine and mepiquat were ΔM = +3.0186 and +6.0373 Da, again corresponding to a single CD3 group, and two CD3 groups respectively. Decomposition of choline and betaine In the case of betaine and choline, there is a paucity of data in general on the effect of cooking/heating on these quaternary amines. Recently, Ross et al. (2014) reported significant losses (> 50%) of betaine in pasta and noodles after cooking (boiling), irrespective of whether the products were whole or refined grain products. Results across different cereal grains are, however, not consistent as in

1955

some cases betaine may increase after cooking (e.g. in oatmeal). To our knowledge, there are no reports on the impact of roasting on the stability of betaine and choline. Therefore, we chose to measure the formation of the corresponding demethylated congeners N,N-dimethyl-2aminoethanol and N,N-dimethylglycine in the reaction mixtures. This approach provided concrete evidence of the role of the quaternary N compounds in methyl rearrangement reactions. Both choline and betaine can in theory transfer Nmethyl substituents to available nucleophiles. The loss of a single methyl group from choline and betaine should theoretically lead to N,N-dimethyl-2-aminoethanol and N, N-dimethylglycine, respectively. Indeed, these two degradation compounds were detected in the heat-treated ternary mixtures of lysine/fructose/choline (or betaine) by exact mass extraction of m/z 90.091340 and m/z 104.070605, corresponding to the protonated molecule [M + H]+ of N,N-dimethyl-2-aminoethanol and N,Ndimethylglycine respectively. Both signals were detected, and their intensity increased over time as evidenced by the curves depicted in the Figure 5. The chemical structures of N,N-dimethyl-2-aminoethanol and N,N-dimethylglycine were confirmed by recording MS/MS data from the

Figure 5. (colour online) Time-course formation of N,N-dimethylglycine (A) and N,N-dimethyl-2-aminoethanol (B) in heat-treated (240°C) ternary mixtures of lysine/fructose/choline or betaine. Entries are averages of quadruplicate determinations.


1956

T. Bessaire et al.

Figure 6. Fragmentation patterns of the protonated molecule [M + H] + of N,N-dimethylglycine (A) and N,N-dimethyl-2aminethanol (B).

signals at m/z 90.091340 and m/z 104.070605 (Figure 6), and by comparison of the chromatographic behaviour with commercial standards of N,N-dimethyl-2-aminoethanol and N,N-dimethylglycine. The base peak in the product ion spectrum of the precursor ion at m/z 90.0918 (N,N-dimethyl-2-aminoethanol) was observed at m/z 72.0814, corresponding to a mass loss of ΔM = –18.0104 Da that fits with the molecular mass of water (theoretical mass at 18.010565 Da). The fragment at m/z 72.0814 was assigned to the ion resulting from the loss of water induced by the electron withdrawal from the O-C bond to the charged oxygen atom of the ion [M + H]+ followed by a rearrangement. Further details regarding product ions at m/z 70.0658, m/z 57.0581 and m/z 56.0503 are shown in Figure 6. A similar process was observed with the fragmentation of the precursor ion at m/z 104.070605 (N,N-dimethylglycine) that led to the loss of the functional group in the β-position of the nitrogen atom. In that particular case, the cleavage mediate the loss of formic acid (theoretical mass at

46.005480 Da); the calculation of the mass difference between the protonated molecule [M + H]+ (m/z 104.0709) and the single fragment ion observed in the spectrum at m/z 58.0659 is ΔM = 46.0050 Da. Thermally treated mixtures were co-injected with pure standards of either N,N-dimethyl-2-aminoethanol and N, N-dimethylglycine, leading to a perfect match in the retention times between the standard and the compound to be identified in the sample. Furthermore, the product ion spectra of both standards demonstrated the identity of the fragmentation pattern with the compounds observed at m/z 90.0918 or m/z 104.0709 in the samples. These observations are confirmed by the data generated with the stable isotope-labelled methylation agents, namely d9-choline and d11-betaine. With d9-choline, the signal of N,N-dimethyl-2-aminoethanol was observed with an exact mass at m/z 96.1299; ΔM is +6.0381 Da when compared with the non-labelled compound at m/z 90.0918 (substitution of six H atoms by D atoms with a theoretical up-mass shift at 6 × 1.006277 = 6.037662 Da).

Food Additives & Contaminants: Part A Similarly, with d11-betaine, the up-mass shift between the exact mass of the stable isotope-labelled and the nonlabelled N,N-dimethylglycine compounds (masses at m/z 112.1211 and 104.0709 Da respectively) was found at ΔM = +8.0502 Da. These observations demonstrate that both choline and betaine have decomposed by the loss of one methyl group during the heat treatment.


Conclusions This model study demonstrates for the first time the ability of either choline or betaine to act as methylation agents under dry thermal conditions. More specifically, both quaternary amines, abundant in many different cereal grains, can alkylate available nucleophiles such as piperidine and N-methylpiperidine, thereby furnishing the N-methylated analogues N-methylpiperidine (tertiary amine) and N,Ndimethylpiperidinium (quaternary amine), respectively. The role of lysine – that provides the carbon backbone of the nitrogen heterocycle – as well as choline or betaine – that furnish the N-methyl substituents of mepiquat – are clearly demonstrated. This was further supported by monitoring the formation of N,N-dimethyl-2-aminoethanol and N,N-dimethylglycine, the respective decomposition products of choline and betaine obtained by the loss of one N-methyl group. This study suggests the possible presence of traces of mepiquat in toasted/roasted foods that naturally contain significant amounts of choline and/or betaine (de Zwart et al. 2003; Zeisel et al. 2003a, 2003b; Slow et al. 2005), together with the availability of free lysine. Cereal-based products may be of particular interest at evaluating mepiquat levels as both choline and betaine are known to occur in various cereals in the range 0.2–1.1 mg g–1 for choline, and 0.4–12.9 mg g–1 for betaine (Likes et al. 2007; Corol et al. 2012). In a further study, we hope to understand better the possible role of these food components in the formation of mepiquat in roasted/toasted cereal-based foods. Based on the very low yield observed in these model systems, and considering the multiple steps in what can be considered minor chemical pathways, only amounts in the very low part per million range are expected in the final food products. References Campa C, Ballester JF, Doulbeau S, Dussert S, Hamon S, Noirot M. 2004. Trigonelline and sucrose diversity in wild Coffea species. Food Chem. 88:39–43. Corol D-I, Ravel C, Raksegi M, Bedo Z, Charmet G, Beale MH, Shewry PR, Ward JL. 2012. Effects of genotype and environment on the contents of betaine, choline, and trigonelline in cereal grains. J Agric Food Chem. 60:5471–5481. de Zwart FJ, Slow S, Payne RJ, Lever M, George PM, Gerrard JA, Chamber ST. 2003. Glycine betaine and glycine betaine analogues in common foods. Food Chem. 83:197–204.

1957

Ferrer I, Thurman EM. 2007. Importance of the electron mass in the calculations of exact mass by time-of-flight mass spectrometry. Rapid Commun Mass Spectrom. 21:2538–2539. Goldmann T, Davidek T, Gouezec E, Blank I, Bertholet MC, Stadler RH. 2009. Formation of styrene during Maillard reaction is negligible. Food Addit Contam: Part A. 26:583– 594. Hamlet CG, Sadd PA, Liang L. 2007. Raw material variability: a tool to minimize acrylamide formation in baked cereal products? Presentation at ACS Boston, USA, August 2007. Hammel Y-A, Dubois M, Delatour T, Stadler RH. 2014. N,Ndimethylpiperidinium (mepiquat): Part 1. Formation in model systems and relevance to roasted food products. Food Addit Contam: Part A. 31:226–233. Koshiro Y, Zheng X-Q, Wang M-L, Nagai C, Ashihara H. 2006. Changes in content and biosynthetic activity of caffeine and trigonelline during growth and ripening of Coffea arabica and Coffea canephora fruits. Plant Sci. 171:242–250. Lang R, Yagar EF, Eggers R, Hofmann T. 2008. Quantitative investigation of trigonelline, nicotinic acid, and nicotinamide in foods, urine, and plasma by means of LC-MS/MS and stable isotope dilution analysis. J Agric Food Chem. 56:11114–11121. Likes R, Madl RL, Zeisel SH, Craig SAS. 2007. The betaine and choline content of a whole wheat flour compared to other mill streams. J Cereal Sci. 46:93–95. Nikolov PY, Yaylayan VA. 2010. Formation of pent-4-en-1amine, the counterpart of acrylamide from lysine and its conversion into piperidine in lysine/glucose reaction mixtures. J Agric Food Chem. 58:4456–4462. Peiser HS, Holden NE, De Bièvre P, Barnes IL, Hagemann R, De Laeter JR, Murphy TJ, Roth E, Shima M, Thode HG. 1984. Element by element review of their atomic weights. Pure Appl Chem. 56:695–768. Rademacher W. 2000. Growth retardants: effects on gibberellin biosynthesis and other metabolic pathways. Annu Rev Plant Physiol Plant Mol Biol. 51:501–531. Ross AB, Zangger A, Guiraud SP. 2014. Cereal foods are the major source of betaine in the Western diet – Analysis of betaine and free choline in cereal foods and updated assessments of betaine intake. Food Chem. 145:859–865. Slow S, Donaggio M, Cressey PJ, Lever M, George PM, Chambers ST. 2005. The betaine content of New Zealand foods and estimated intake in the New Zealand diet. J Food Comp Anal. 18:473–485. Stadler RH, Blank I, Varga N, Robert F, Hau J, Guy PA, Robert M-C, Riediker S. 2002. Food chemistry: acrylamide from Maillard reaction products. Nature. 419:449–450. Stadler RH, Varga N, Hau J, Vera FA, Welti DH. 2002. Alkylpyridiniums. 1. Formation in model systems via thermal degradation of trigonelline. J Agric Food Chem. 50:1192–1199. Stadler RH, Varga N, Milo C, Schilter B, Vera FA, Welti DH. 2002. Alkylpyridiniums. 2. Isolation and quantification in roasted and ground coffees. J Agric Food Chem. 50:1200– 1206. Stadler RH, Verzegnassi L, Varga N, Grigorov M, Studer A, Riediker S, Schilter B. 2003. Formation of vinylogous compounds in model Maillard reaction systems. Chem Res Toxicol. 16:1242–1250. Warthesen JJ, Kramer PL. 1978. High-pressure liquid chromatographic method for evaluating free lysine stability in lysine-fortified wheat-flour products. Cereal Chem. 55:481–488.

1958

T. Bessaire et al.


Wermann S, Theurillat V, Verzegnassi L, Hofmann J, Kuchenbecker R, Constable A, Delatour T, Stadler RH. 2014. N,N-dimethylpiperidinium (mepiquat) Part 2. Formation in roasted coffee and barley during thermal processing. Food Addit Contam: Part A. 31:234–241. Zeisel SH, Mar M-H, Howe JC, Holden JM. 2003a. Concentrations of choline-containing compounds and betaine in common foods. J Nutr. 133:1302–1307.

Zeisel SH, Mar M-H, Howe JC, Holden JM. 2003b. Erratum Concentrations of choline-containing compounds and betaine in common foods. J Nutr. 133:2918–2919. Zyzak DV, Sanders RA, Stojanovic M, Tallmadge DH, Eberhart BL, Ewald DK, Gruber DC, Morsch TR, Strothers MA, Rizzi GP, Villagran MD, 2003. Acrylamide formation mechanism in heated foods. J Agric Food Chem. 51:4782– 4787.