Individual Maillard reaction products as indicators of heat treatment of pasta — A survey of commercial products

Journal of Food Composition and Analysis 72 (2018) 83–92

Contents lists available at ScienceDirect

Journal of Food Composition and Analysis journal homepage: www.elsevier.com/locate/jfca

Individual Maillard reaction products as indicators of heat treatment of pasta — A survey of commercial products Michael Hellwig, Lennart Kühn, Thomas Henle

T

⁎

Chair of Food Chemistry, Technische Universität Dresden, D-01062 Dresden, Germany

A R T I C LE I N FO

A B S T R A C T

Keywords: Maillard reaction Pasta Drying Advanced glycation end product (AGE) Amadori rearrangement product Pyrraline CML Maltosine Glucosyl isomaltol Food analysis Food composition

During pasta production, drying processes promote glycation reactions (non-enzymatic browning, Maillard reaction). Certain “traditionally produced” pasta products are often labelled with claims suggesting high quality due to a gentle heat treatment. The impact of heat treatment of food can be assessed by the measurement of Maillard reaction products (MRPs). In the present study, the MRPs N-ε-fructosyllysine, N-ε-maltulosyllysine, furosine, N-ε-carboxymethyllysine (CML), N-ε-carboxyethyllysine (CEL), pyrraline, formyline, maltosine, methylglyoxal-derived hydroimidazolone 1 (MG-H1), 3-deoxyglucosone, 3-deoxygalactosone, glucosyl isomaltol, and HMF were quantified in 31 pasta products. N-ε-Maltulosyllysine was found to be the predominating MRP (up to 2.4 g/100 g of protein). Between 4 and 28% of the essential amino acid lysine was found to be modified due to glycation reactions. Taking ten MRPs into account, a low-MRP and a high-MRP cluster could be distinguished by cluster analysis. Closer examination of the clusters revealed that the MRPs N-ε-maltulosyllysine, pyrraline, maltosine, HMF, and glucosyl isomaltol differ most strongly and should further be considered as potential heating markers. Lastly, a study on 5 pasta products with different shapes provided evidence that glycated amino acids are not degraded during cooking. Maltosine is formed additionally during cooking. The consumption of pasta contributes substantially to the daily intake of Amadori rearrangement products and CML.

1. Introduction Pasta products are traditionally prepared from only two ingredients—water and durum wheat semolina. After mixing and working of semolina and water, the dough is extruded, giving the pasta products their final form. Beneath the characteristics of the raw materials, the heat treatment during pasta drying is of high importance for the quality of the final product. A certain heat impact can already occur during the extrusion process. Local heating may rise above 60 °C, but the processing time is very short in this step (Sicignano et al., 2015). Drying of the extruded intermediate goods to ca. 12.5% water content needs to be performed, in order to gain products that are microbiologically stable during storage at room temperature (Cavazza et al., 2013; Sicignano et al., 2015; Mercier et al., 2016). Drying times can be reduced through application of higher drying temperatures. Thus, different courses of heat treatment are applied during industrial pasta production, the more “traditional” low-temperature processes (below 60 °C), then high-temperature processes (60–80 °C, sometimes up to 100 °C) and very-high temperature cycles (80–100 °C and even higher)

(Pagani et al., 1996; Giannetti et al., 2014; Sicignano et al., 2015). Multi-stage drying processes are often implemented, where different low- and high-temperature steps are combined: The drying temperature is changed during the drying process, and the relative humidity is regulated concomitantly (Zweifel et al., 2003; Cavazza et al., 2013; Mercier et al., 2016; Padalino et al., 2016). Beneath faster processing and reduced costs, it is a further advantage of higher drying temperatures, that the final products obtain better cooking performance (Zweifel et al., 2003; Padalino et al., 2016). On the other hand, thermal “damage”, such as unwanted discoloration and negative sensory properties (e.g., bread crust flavor), can occur in high-temperature heated pasta products (Pagani et al., 1996). Moreover, the digestibility of amino acids in these pasta products may be reduced (Stuknytė et al., 2014). There is currently no legislative regulation concerning the differentiation between low- and high-temperature treatments. Therefore, manufacturers are relatively free in utilizing descriptions pointing to “low-temperature treatment”, “long-time drying”, or “gentle preparation”, which is considered the traditional way of pasta drying.

Abbreviations: AGE, advanced glycation end product; ARP, Amadori rearrangement product; aw, water activity; CEL, N-ε-carboxyethyllysine; CML, N-ε-carboxymethyllysine; HMF, 5hydroxymethylfurfural; LOD, limit of detection; LOQ, limit of quantification; MG-H1, methylglyoxal-derived hydroimidazolone 1; MRM, multiple reaction monitoring; MRP, Maillard reaction product; NFPA, nonafluoropentanoic acid; RT, room temperature ⁎ Corresponding author. E-mail address: [email protected] (T. Henle). https://doi.org/10.1016/j.jfca.2018.06.009 Received 13 March 2018; Received in revised form 6 June 2018; Accepted 14 June 2018 Available online 18 June 2018 0889-1575/ © 2018 Elsevier Inc. All rights reserved.

Journal of Food Composition and Analysis 72 (2018) 83–92

M. Hellwig et al.

Fig. 1. Proposed pathways of sugar degradation and Maillard reaction in pasta products (Literature and abbreviations see text).

pasta has been determined, whereas methylglyoxal (MGO) was not detectable (Degen et al., 2012). Only traces of 5-hydroxymethylfurfural (HMF) were determined in pasta samples (Resmini et al., 1993), and it was not detectable in pasta products after cooking (Degen et al., 2012). The concentration of glucosyl isomaltol (up to 10 mg/kg), which is a degradation product of maltose, correlates well with the heat treatment of pasta (Resmini et al., 1993; Pagani et al., 1996). In the late stage of the Maillard reaction, dicarbonyl compounds react with protein-bound amino acid residues, such as the ε-amino group of lysine or the guanidino group of arginine, to the so-called “advanced glycation end products (AGEs, Fig. 1)”, whose formation in food is not reversible (Hellwig and Henle, 2014). Carboxyalkylation and pyrrole formation at the ε-amino group of lysine and hydroimidazolone formation at the guanidino group of arginine are the most prominent glycation events in the late Maillard reaction on proteins (Thornalley et al., 2003; Hellwig and Henle, 2014). Several AGEs, such as CML, CEL, pyrraline, formyline, maltosine, and MG-H1, have already been quantified in a limited number of cooked and uncooked pasta products (Resmini and Pellegrino, 1994; Hellwig and Henle, 2012; Hellwig et al., 2016a; Scheijen et al., 2016). Among these AGEs, only pyrraline has been suggested as a heating marker, whose concentrations increase especially at drying temperatures above 80 °C and pasta moisture contents below 15% (Resmini and Pellegrino, 1994). Different drying cycles are applied for different pasta shapes, and these differences significantly influence the Maillard reaction, as was assessed using color indices, furosine, and maltulose formation (Cavazza et al., 2013). Thus, it was the aim of the present study to assess the suitability of a broader spectrum of MRPs as indicators of heat treatment of pasta products and to perform a survey on the concentrations of different MRPs in commercial spaghetti samples. Thereby, we wanted to gain insight into how modern drying conditions affect the concentrations of MRPs in pasta. Moreover, the stability of MRPs during cooking of pasta and the contribution of pasta consumption to the total daily intake of MRPs should be addressed.

Consumers increasingly appear to demand gently dried pasta and be ready to pay higher prices for the respective products: On the German market, prices for 1 kg of spaghetti range from 2.38 € to 10.76 € for correspondingly labelled products and from 0.98 € to 3.18 € for products that are not labelled with such claims (Table S1). Depending on the drying conditions and the relative humidity during drying, reactions between reducing sugars and amino and imino groups of free amino acids, peptides and proteins can take place, which are collectively called Maillard reaction, glycation or “non-enzymatic browning”. The reaction can also take place during cooking (Hellwig and Henle, 2014). Due to α-amylase activity, an increase in the ratio between reducing sugars and reactive lysine residues is observed during the manufacturing process, which facilitates Maillard reactions during drying (Lintas and D’Appolonia, 1973; Resmini et al., 1993; Pagani et al., 1996; Testani et al., 2017). These reactions are furthermore supported by the reduction of the water activity during the drying process (final aw = 0.8, Giannetti et al., 2014). In the first step of this reaction, Amadori rearrangement products (ARPs) are formed, which can collectively be quantified as furosine after acid hydrolysis (Fig. 1). Concentrations between 40 and 880 mg furosine per 100 g protein have been reported in pasta products, indicating that up to 60% of lysine residues can be modified in this fashion (Resmini and Pellegrino, 1994; Henle et al., 1995; Pagani et al., 1996; Acquistucci, 2000; Cavazza et al., 2013). The exact nature of the ARPs formed has not yet been assessed, but from the sugar composition of dried pasta products (Cavazza et al., 2013; Lintas and D’Appolonia, 1973), it can be expected that N-ε-fructosyllysine and N-ε-maltulosyllysine should predominate. The use of furosine for the assessment of heat treatment of pasta products has been suggested (García-Baños et al., 2004), but this is precluded by the low stability of ARPs during more intense heat treatment (Resmini and Pellegrino, 1994; Anese et al., 1999; Cavazza et al., 2013). In the second step of the Maillard reaction, ARPs decompose under formation of 1,2-dicarbonyl compounds, such as 3-deoxyglucosone (3DG), methylglyoxal, and glyoxal. Up to 8.8 mg 3-DG per kg of cooked 84

Journal of Food Composition and Analysis 72 (2018) 83–92

M. Hellwig et al.

column (Eurospher-C18, 250 mm × 4.6 mm, 5 μm, 100 Å) with a guard column (5 × 4 mm) of the same material (Knauer) was used at room temperature. The eluent (5% acetonitrile in water) was pumped at a flow rate of 1 mL/min for 30 min. Of the supernatants, 20 μL aliquots were injected. The absorbance was read at 280 nm, and spectra (190–400 nm, increment 1 nm) were recorded simultaneously. External calibration was performed with HMF and a synthesized standard of glucosyl isomaltol.

2. Materials and methods 2.1. Chemicals Nonafluoropentanoic acid (NFPA), sodium borohydride, o-phenylenediamine (OPD), quinoxaline, 2-methylquinoxaline, pepsin (3839 U/mg protein), leucine aminopeptidase (18 U/mg protein), prolidase (553 U/mg protein) and pronase E (4000 PU/mg) were purchased from Sigma-Aldrich (Steinheim, Germany). HPLC gradient grade acetonitrile and hydrochloric acid were obtained from VWR (Darmstadt, Germany). HPLC-MS/MS grade acetonitrile was from Fisher Scientific (Loughborough, UK) and 5-hydroxymethylfurfural from Alfa Aesar (Karlsruhe, Germany). Zinc sulfate heptahydrate and boric acid were purchased from Merck (Darmstadt, Germany). Tris and potassium hexacyanoferrate(II) were obtained from Riedel-de-Haën (Seelze, Germany). [2H2]CML was from PolyPeptide (Strasbourg, France). The water used for the preparation of buffers and solutions was obtained by a Purelab Plus purification system (USF Elga, Ransbach-Baumbach, Germany). The water for HPLC-MS/MS eluents was double distilled (first over potassium permanganate). MRP standards such as N-ε-fructosyllysine (Krause et al., 2003), N-ε-maltulosyllysine (Hellwig et al., 2016b), CML, CEL, pyrraline, formyline, MG-H1 (Henle and Bachmann, 1996; Hellwig et al., 2011), maltosine (Geissler et al., 2011; Hellwig et al., 2016a), and glucosyl isomaltol (Goodwin, 1983; GuerraHernández et al., 2002), as well as the quinoxaline derivatives of 3-DG and 3-DGal (Degen et al., 2012) were prepared according to the protocols given in the respective publications. Identity and purity of these standards were confirmed by nuclear magnetic resonance spectroscopy, mass spectrometry, elemental analysis, and amino acid analysis. All standards met the characteristics reported in the respective literature.

2.4. Quantification of 1,2-dicarbonyl compounds The analysis of 1,2-dicarbonyl compounds was performed according to a previously published protocol (Degen et al., 2012) by extracting ground pasta samples with water, followed by protein precipitation with methanol and derivatization of the extracts with o-phenylenediamine (OPD). Calibration of the HPLC system was done by use of the independently synthesized quinoxaline standards derived from 3deoxyglucosone and 3-deoxygalactosone as well as commercially available quinoxaline (derivative of glyoxal) and 2-methylquinoxaline (derivative of methylglyoxal). 2.5. Quantification of glycated amino acids Enzymatic hydrolysis was performed as published previously (Hellwig et al., 2016a, b; Hellwig et al., 2017), by first adding 1 mL of 0.02 M hydrochloric acid to ca. 25 mg of ground pasta samples or 40–50 mg of cooked pasta samples, respectively. After addition of 50 μL of 0.02 M hydrochloric acid containing 1 FIP-U of pepsin, the samples were incubated for 24 h at 37 °C in a drying chamber. Then, 300 μL of 2 M Tris buffer, pH 8.2, containing 400 PU of pronase E were added, and the samples were again incubated for 24 h at 37 °C. Finally, 0.4 U of leucine aminopeptidase and 1 U of prolidase were added. After incubation (37 °C, 24 h), the hydrolyzates were directly employed for amino acid analysis and HPLC-MS/MS. Quantification of glycated amino acids was performed on an HPLCMS/MS system consisting of a binary pump (G1312 A), an online degasser (G1379B), an autosampler (G1329 A), a column thermostat (G1316 A), a diode array detector (G1315D), and a triple-quadrupole mass spectrometer (G6410 A; all from Agilent Technologies, Böblingen, Germany). Nitrogen was used as the nebulizing gas (gas flow, 11 L/min; gas temperature, 350 °C; nebulizer pressure, 35 psi), and the capillary voltage was 4000 V. The column Zorbax 100 SB-C18 (2.1 × 50 mm, 3.5 μm; Agilent) was used for separations. HPLC solvent A was a solution of 10 mM nonafluoropentanoic acid (NFPA) in water, and solvent B was a solution of 10 mM NFPA in acetonitrile. The solvents were pumped at a flow rate of 0.25 mL/min at a column temperature of 35 °C in gradient mode (0 min, 5% B; 15 min, 32% B; 16 min, 85% B; 20 min, 85% B; 21 min, 5% B; 28 min, 5% B). The injection volume was 2 μL. MRPs were quantified by standard addition. In the first sample, 100 μL of enzymatic hydrolyzate was mixed with 20 μL of water. In the second sample, 100 μL of hydrolyzate was mixed with 10 μL of water and 10 μL of a standard solution. In the third sample, 100 μL of hydrolyzate was mixed with 20 μL of the standard solution. All samples were centrifuged before injection (10,000 rpm, 10 min, RT). In the standard solution, N-ε-fructosyllysine (87.4 μg/mL), N-ε-maltulosyllysine (69.8 μg/mL), CEL (1.0 μg/mL), formyline (0.2 μg/mL), MG-H1 (0.5 μg/mL), maltosine (0.3 μg/mL), and pyrraline (2.0 μg/mL) were dissolved in water. The operating conditions for MRM measurement are compiled in Table S2.

2.2. Food samples Pasta samples were purchased in local retail stores or via the internet. Spaghetti samples with claims regarding gentle drying (13 items) and spaghetti samples without such claims (13 items) were included in the study (Table S1). The ingredients of all samples were water and durum wheat semolina. No sample contained whole wheat flour or any other additive. All samples were analyzed directly after acquisition. The samples were first ground to a powder with a laboratory mill (Grindomix GM 100; Retsch, Haan, Germany). The fraction passing through a 500-μm sieve was used for all analyses. Pasta samples of different shapes (5 items) were also analyzed before and after cooking. For cooking, 50 g of pasta samples were added to 1000 mL of boiling water containing 1% (w/v) of sodium chloride and cooking was interrupted after the time period recommended by the manufacturer by removing the pasta through a cooking sieve. After weighing, the cooked pasta was homogenized with a blender. 2.3. Quantification of 5-hydroxymethylfurfural (HMF) and glucosyl isomaltol According to a literature method (Resmini et al., 1993), 250 mg of ground pasta products were incubated with 1000 μL 0.1 M sodium borate buffer (pH 8.2) at 30 °C in a water bath for 30 min. After shaking (2 min), the samples were centrifuged (10,000 rpm, 10 min, RT). To 600 μL of the supernatants, 30 μL Carrez I solution (15% (w/v) potassium hexacyanoferrate(II) in water) and 30 μL Carrez II solution (30% (w/v) zinc sulfate heptahydrate in water) were added with intermediate shaking. The mixtures were incubated at 4 °C for 1 h, centrifuged (10,000 rpm, 10 min, RT), and the supernatant was analyzed by RP-HPLC on a low-pressure gradient system consisting of a solvent organizer (K-1500; Knauer, Berlin, Germany), an autosampler (Basic Marathon; Spark Holland, Emmen, The Netherlands) a pump (Smartline 1000, Knauer), an online degasser (Knauer), a column oven, and a diode array detector (DAD 2.1 L, Knauer). For the separation, an RP

2.6. Quantification of N-ε-carboxymethyllysine CML was quantified after reduction/acid hydrolysis as described in the literature (Loaëc et al., 2015; Helou et al., 2016). To an aliquot of ground pasta samples (75 mg), 1.5 mL 200 mM sodium borate buffer (pH 9.5) were added followed by 1 mL of 1 M NaBH4 in 0.1 M NaOH. 85

Journal of Food Composition and Analysis 72 (2018) 83–92

M. Hellwig et al.

limits may be of interest for the evaluation of heat treatment parameters. Correlations between the drying conditions and the MRP concentrations have been described in the literature only for furosine as a marker for Amadori rearrangement products (ARPs), pyrraline, and glucosyl isomaltol (Resmini and Pellegrino, 1994; Pagani et al., 1996; Anese et al., 1999). In the present study, we wanted to focus on the late stage of the Maillard reaction and explore the use of the respective glycated amino acids as integral markers of heat impact. In a first step, the concentrations of selected early, advanced and late Maillard reaction products were analyzed in twenty-six spaghetti samples. Thirteen of these samples were labelled with descriptions concerning the drying conditions (e.g., gentle drying, slow drying at low temperatures; sample group “L”), thirteen were not labelled in this way (sample group “S”). The samples of group “L” had a significantly higher price (median 4.78 €/kg vs.1.70 €/kg) and a higher diameter (median 1.95 mm vs. 1.65 mm). Some of the glycated amino acids are unstable during acid hydrolysis (e.g., pyrraline, formyline, ARPs). Therefore, all samples were first analyzed after enzymatic hydrolysis. A major drawback of enzymatic hydrolysis is the possibility of incomplete release of amino acids from proteins. Therefore, all samples were also analyzed by amino acid analysis after acid hydrolysis. Then, the release of amino acids during enzymatic hydrolysis was compared to the release of amino acids during acid hydrolysis. The degree of hydrolysis in the present study was over 90% for the aliphatic amino acids alanine, valine, leucine, and isoleucine as well as for phenylalanine, which is in the range of literature data (Wellner et al., 2012). A slight underestimation of the MRP concentrations after enzymatic hydrolysis cannot be excluded. Pyrraline and formyline can be analyzed by RP-HPLC with UV-detection in pasta samples (Hellwig and Henle, 2012). Other glycated amino acids are either not UV-active (e.g., CML, MG-H1) or their concentrations in pasta samples are too low for quantification by UV-detection (e.g., maltosine (Hellwig et al., 2016a)). Therefore, HPLC-MS/ MS in MRM mode was applied as in our previous works (Hellwig et al., 2016b, 2017). Quantification was based on standard addition, because (i) no isotopically labelled internal standards are available for all glycated amino acids, and (ii) standard addition does not suffer from interferences that may occur when calibration is performed in not matrixmatched solution, because each sample with its individual matrix is spiked for quantification. Limits of detection and quantification as well as reproducibility data are given in Table 1. Quantitative data for glycated amino acids are expressed as mg/ 100 g of protein (Table 2). In our study, N-ε-maltulosyllysine turned out to be the main ARP, predominating over N-ε-fructosyllysine by a factor between 2 and 11. This distribution could be expected, since in pasta production, mainly maltose is released during starch degradation, making this disaccharide the primary free sugar in pasta products (Lintas and D’Appolonia, 1973; Cavazza et al., 2013). The concentrations of N-ε-maltulosyllysine in our study ranged between 198 and 2400 mg/100 g protein, whereas those of N-ε-fructosyllysine ranged between 105 and 399 mg/100 g protein. This is equivalent to a lysine blockage between 2 and 25% by N-εmaltulosyllysine and 2 to 7% by N-ε-fructosyllysine. In order to see, to what extent the concentrations of N-ε-maltulosyllysine and N-ε-fructosyllysine in spaghetti may explain the total ARP-derived lysine modification, furosine was measured by amino acid analysis. Furosine is generated from ARPs in defined molar yields during acid hydrolysis (Krause et al., 2003). In the present study, a conversion factor of 2.9 mol parent ARP per mole of furosine was used for N ε-maltulosyllysine and a factor of 3.1 was used for N ε-fructosyllysine. The measured furosine concentrations in spaghetti samples matched the theoretical furosine concentrations by 101% on average. We conclude that N-ε-fructosyllysine and N-ε-maltulosyllysine can more or less completely explain the full ARP modification in pasta products. A large discrepancy between the concentration of total ARPs (measured as furosine) and the concentrations of directly measured individual ARPs was observed for the beer proteome (Hellwig et al., 2016b).

After 4 h of incubation at room temperature, 2.5 mL of 12 M HCl were added in small portions. The mixtures were hydrolyzed for 20 h at 110 °C in a drying chamber. An aliquot of the hydrolysis solution (300 μL) was evaporated by use of a vacuum concentrator and reconstituted in 950 μL of 10 mM NFPA. A portion of 50 μL of an internal standard solution containing 0.25 mg/mL [2H2]CML was added. After mixing, the solutions were centrifuged (10,000 rpm, 10 min, RT) and the supernatants were analyzed by HPLC-MS/MS with the same device, column and eluents as described above. As published previously (Hellwig et al., 2015), a shortened gradient was applied (0 min, 5% B; 8 min, 29% B; 9 min, 85% B; 13 min, 85% B; 14 min, 5% B; 20 min, 5% B). The injection volume was 5 μL. Between 5 and 9.5 min, the transitions 205→130 (100 V, 9 eV) and 207→130 (80 V, 5 eV) were recorded for quantification of the analytes CML and [2H2]CML, respectively. Simultaneously, the transitions 205→142 (100 V, 9 eV) and 207→144 (80 V, 12 eV) were recorded for confirming the presence of the analytes. Fragmentor voltage and collision energy are given in parentheses. The dwell time was 120 ms. Calibration was performed with a standard of CML at concentrations between 0.004 and 0.4 mg/L. 2.7. Amino acid analysis Proteinogenic amino acids and furosine were quantified with an amino acid analyzer (S 433; Sykam, Fürstenfeldbruck, Germany) using a PEEK column filled with the cation exchange resin LCA K07/Li (150 × 4.6 mm, 7 μm). All buffers used for separations were purchased from Sykam and employed for custom gradient programs as utilized previously (Hellwig et al., 2011). After post-column derivatization with ninhydrin, the detection of products was performed with an integrated two-channel photometer (λ = 440 nm, 570 nm). Furosine was quantified by an additional UV-detector at λ = 280 nm (Krause et al., 2003). External calibration was performed with an amino acid mixture (SigmaAldrich, Steinheim, Germany). The injection volume was between 50 and 100 μL. Prior to analysis, 100 μL of the hydrolyzate was diluted with 300 μL of 0.12 M lithium citrate buffer (pH 2.2) and membrane filtered (0.45 μm). The protein concentration was calculated as the sum of amino acids in acid hydrolyzates after subtraction of one mole of water per mole of amino acid (Mæhre et al., 2018) and correction for the degradation of Amadori products. Regenerated lysine and original Amadori product were calculated from the furosine concentrations (Krause et al., 2003). 2.8. Statistical treatment The limits of detection (LOD) and quantification (LOQ), respectively, are the concentrations of the analytes that are necessary to show peaks with signal-to-noise ratios of 3 and 10, respectively. All statistical tests were performed using the software PASW statistics 18. Parameters of cluster analysis were the following: two-step cluster analysis, hierarchical clusters, and squared Euclidian distance. The significance of differences between the medians of the HMRP (samples with higher concentrations of MRPs) and LMRP (samples with lower concentrations of MRPs) clusters was determined by the two-tailed Mann-Whitney U test. MRP concentrations between LOD and LOQ were considered with the concentration of the LOQ. When MRPs were not detectable, they were considered with the concentration of the LOD. All samples were analyzed at least in duplicate starting from two independent analytical workups (extraction of sugar derivatives or acid/enzymatic hydrolysis for protein-bound glycated amino acids). 3. Results and discussion 3.1. Quantification of Maillard reaction products in spaghetti Data on the occurrence of MRPs in pasta have been gained mostly on small numbers of samples in the literature; however, the variation 86

Journal of Food Composition and Analysis 72 (2018) 83–92

M. Hellwig et al.

Table 1 Performance Parameters of the HPLC-MS/MS and HPLC-UV Methods for the Measurement of MRPs in Pasta Samples. method

LODa [mg/100 g protein]

LOQa [mg/100 g protein]

reproducibilityc [%]

N-ε-maltulosyllysine N-ε-fructosyllysine CML CEL pyrraline formyline maltosine MG-H1

LC-MS/MS LC-MS/MS LC-MS/MSb LC-MS/MS LC-MS/MS LC-MS/MS LC-MS/MS LC-MS/MS HPLC-UV HPLC-UV HPLC-UV HPLC-UV

32.5 28.5 0.9 15.6 1.0 0.5 0.3 2.2 [mg/kg] 2.7 2.6 0.14 0.05

7.1 6.5 7.1 — 6.1 6.7 7.7 10.0

3-DG 3-DGal glucosyl isomaltol HMF

11.0 10.1 0.3 3.5 0.3 0.2 0.1 0.6 [mg/kg] 0.8 0.8 0.07 0.02

5.7 — 5.4 10.1

a LOD and LOQ are determined in calibration experiments, except for 3-DG and 3-DGal (cf. Degen et al., 2012). For glycated amino acids, calculation is based on a protein amount of 4 mg for enzymatic hydrolysis. b CML was determined with a different method after reduction/acid hydrolysis as described in the Experimental section. c Mean value of the variation coefficients for repeated determinations of analytes in all pasta samples with concentrations > LOD.

CML was measured after a separate sample pretreatment, which consisted of borohydride reduction followed by acid hydrolysis (Loaëc et al., 2015; Helou et al., 2016). In the literature, combinations of reduction/hydrolysis are applied in order to restrict an oxidative formation of CML from ARPs during hydrolysis, which might lead to an overestimation of CML. Our results show that CML is the main AGE in dried spaghetti with concentrations between 3.5 and 15.5 mg/100 g protein (median, 9.7 mg/100 g protein), which corroborates literature data (Scheijen et al., 2016). The concentrations of pyrraline and formyline also match literature values (Resmini and Pellegrino, 1994; Hellwig and Henle, 2012). Traces of CEL were found only in four samples. The contribution of AGEs to the total lysine blockage was low as compared to the contribution of the ARPs. Even at the highest concentrations of the most abundant AGEs CML and pyrraline, only up to 0.4% and 0.5%, respectively, of lysine derivatization can be explained. The concentration of HMF in the spaghetti products did not exceed 0.55 mg/kg in the present study, which is in accordance with literature values (Resmini et al., 1993). The 1,2-dicarbonyl compound 3-DG was present in concentrations up to 8.4 mg/kg. The mean ratio between 3DG and HMF was 55. It is known from the literature that the HMF

Fig. 2. RP-Phenyl-HPLC-UV of extracts of (a) a spaghetti sample from the HMRP cluster, (b) a spaghetti sample from the LMRP cluster, and (c) a standard mixture containing the quinoxalines of 3-DG and 3-DGal.

precursor 3-DG accumulates in food systems with low water activity (Degen et al., 2012). The quinoxalines of the 1,2-dicarbonyl compounds glyoxal, methylglyoxal, and 3-deoxygalactosone were not detectable. However, several other peaks emerged in the chromatograms of all pasta samples (Fig. 2) after OPD derivatization. Degradation of maltose and formation of maltose-specific dicarbonyl compounds, such as maltosone, 3-deoxymaltosone and 1-deoxymaltosone, may be responsible for these peaks, which we will consider in future works. The maltose degradation product glucosyl isomaltol yielded concentrations up to 7.9 mg/kg, which is also in the range of literature data (Pagani et al., 1996; Resmini et al., 1993). A further comparison was performed between the “L” samples and the “S” samples. Differences between the medians of the MRP concentrations of these two sample groups are quite small and never significant. The differences in the medians do not exceed a factor of 2.1

Table 2 Concentrations of Individual Maillard Reaction Products in Dried Spaghetti Samples and Comparison of HMRP and LMRP Clusters. compound

unit

range (median)

range (median)

range (median)

significance (p value)b

lysinec N-ε-maltulosyllysine N-ε-fructosyllysine furosine CML pyrraline formyline maltosine MG-H1 3-DG glucosyl isomaltol HMF

g/100 g protein mg/100 g protein mg/100 g protein mg/100 g protein mg/100 g protein mg/100 g protein mg/100 g protein mg/100 g protein mg/100 g protein mg/kg mg/kg mg/kg

all samplesa 2.3–2.8 (2.6) 198–2400 (1345) 105–399 (224) 87–438 (324) 3.5–15.5 (9.7) tr–24.0 (3.5) 0.7–4.3 (1.6) tr–3.4 (0.5) tr–7.7 (3.1) tr–8.4 (4.6) 0.2–7.9 (1.8) n.d.–0.55 (0.10)

HMRP clustera 2.3–2.7 (2.6) 1247–2400 (1501) 150–399 (217) 287–438 (346) 7.9–15.5 (12.5) 3.3–24.0 (5.3) 1.4–4.3 (2.2) 0.4–3.4 (0.8) tr–7.7 (4.9) 2.9–8.4 (5.4) 1.1–7.9 (2.6) n.d.–0.55 (0.15)

LMRP clustera 2.4–2.8 (2.7) 198–1345 (879) 105–261 (247) 87–271 (225) 3.5–8.3 (6.9) tr–2.7 (1.5) 0.7–1.5 (1.2) tr–0.4 (tr) tr–3.0 (2.7) tr–5.0 (3.0) 0.2–1.5 (0.7) n.d.–0.08 (tr)

HMRP vs. LMRP 0.121 (n.s.) 0.000 (*) 0.938 (n.s.) 0.000 (*) 0.000 (*) 0.000 (*) 0.000 (*) 0.000 (*) 0.000 (*) 0.001 (*) 0.000 (*) 0.000 (*)

a b c

all samples, n = 26; LMRP cluster, n = 10; HMRP cluster, n = 16; n.d., not detected; tr, trace amounts (between LOD and LOQ). Significance of differences between the medians of HMRP and LMRP clusters was determined by the two-tailed Mann-Whitney U test. Lysine was quantified after acid hydrolysis and the value is corrected for reconversion of lysine from Amadori products. 87

Journal of Food Composition and Analysis 72 (2018) 83–92

M. Hellwig et al.

(glucosyl isomaltol and maltosine, cf. Table S3). Some of the “L” samples show concentrations of MRPs in the range of “S” samples. We conclude that the labelling of pasta products does not give reliable information about the heat damage of proteins by the Maillard reaction.

(Fig. 3D). We conclude that pyrraline formation is more strongly intensified than formyline formation when the heat treatment is intensified. 3.3. Influence of cooking on the concentration of glycated amino acids

3.2. Cluster analysis and correlation analysis Pasta is dried at the end of the manufacturing process in order to obtain storable products, which need to be rehydrated by cooking. Cooking imposes a further heat impact on these products, which may influence the concentrations of MRPs. In order to assess whether MRP concentrations of dried pasta products can be taken as a measure for the intake of MRPs through cooked pasta, different products were analyzed before and after cooking. It is known from the literature that differently shaped pasta products can show significantly different concentrations of certain MRPs due to differences in the respective drying cycles. Spaghetti has higher concentrations of furosine than penne and farfalle and more maltulose than vermicelli or macaroni (García-Baños et al., 2004; Cavazza et al., 2013). In the present study, differently shaped pasta products were cooked according to the manufacturer’s instructions. These pasta products were not labelled with respect to a particular heat treatment. The amount of water absorbed by the different pasta formats was different, and changes in the concentrations of MRPs could not easily be compared. Therefore, MRP concentrations were normalized to the leucine concentrations of the same hydrolyzates and expressed as leucine equivalents as in a previous work (Hellwig et al., 2017). After glutamic acid and proline, leucine is the third most abundant amino acid in pasta products, and it is not modified by the Maillard reaction. Three of the shapes (twisted pipes, corkscrews, and gnocchi) had very similar concentrations of individual MRPs before and after cooking (Table 3). In these shapes, the same ranking of products was observed (N-ε-maltulosyllysine > N-ε-fructosyllysine > CML > pyrraline > MG-H1 > formyline > maltosine). This may be due to a very similar way of drying. In the farfalle format, all MRPs reached their lowest concentrations, maltosine was found only in traces. The ranking was diverse (N-ε-maltulosyllysine > N-ε-fructosyllysine > CML > MGH1 > pyrraline ≈ formyline > maltosine), corroborating the hypothesis that pyrraline concentrations disproportionally rise during intensification of heat treatment. The highest concentrations of MRPs were detected in the spaghetti format. The differences in MRP concentrations found for different shapes in this study confirm the above mentioned literature data on maltulose and furosine (García-Baños et al., 2004; Cavazza et al., 2013). No pronounced changes of the concentrations during cooking were observed for most of the MRPs. The concentrations of pyrraline and formyline were constant. The concentrations of N-ε-maltulosyllysine and MG-H1 tended to increase, whereas those of CML and N-ε-fructosyllysine tended to decrease during cooking. Changes were within a range of ± 20% for these MRPs. However, during cooking, the concentration of maltosine rose by 70% on average in the four formats that allowed quantification of the compound. The conditions applied during pasta cooking may enable maltosine formation from its precursors maltose and maltotriose, which are constituents of dried pasta (Lintas and D’Appolonia, 1973; Stuknytė et al., 2014). Especially maltotriose can react with the ε-amino group of lysine even under conditions of high water activity (Hellwig et al., 2016a). Glucosyl isomaltol as the most potent maltosine precursor is also available for reaction in pasta during cooking (Fig. 4). The molar ratio between glucosyl isomaltol and maltosine is between 2 and 4, indicating that the reaction of glucosyl isomaltol with lysine residues may significantly increase the maltosine concentrations. It may be expected that the reducing aldehyde group at the end of the starch polyglucose polymers can also react with the εamino group of lysine. Lastly, pyridinium betains as disaccharide-specific compounds (Kramhöller et al., 1993) may be formed from maltose during drying and decompose during cooking (Fig. 4). The instability of such compounds in aqueous solution at slightly acid pH has been

Cluster analysis was performed using the software PASW Statistics, in order to test whether spaghetti samples can reasonably be classified into homogeneous groups by their MRP patterns. Only the data for N-εmaltulosyllysine, N-ε-fructosyllysine, 3-deoxyglucosone, HMF, glucosyl isomaltol and the AGEs pyrraline, formyline, maltosine, CML, and MGH1 were regarded. Furosine was excluded, because its concentration is fully dependent on those of the precursor ARPs. CEL was excluded because it was not detectable in 85% of the spaghetti samples. For cluster analysis, the concentrations of MRPs were replaced by their ranks, because the values measured for each MRP were not normally distributed. When trace amounts were detected, the LOQ concentration was used, and when substances were not detectable, the LOD concentration was used. Two clusters were obtained. One cluster contained ten samples, which in most cases had a lower concentration of the MRPs (LMRP cluster), and the other cluster contained sixteen samples by trend containing higher amounts of MRPs (HMRP cluster). In the LMRP cluster, samples with descriptions concerning gentle heat treatment were accumulated (70% “L” samples, 30% “S” samples), while samples without descriptions were accumulated in the HMRP cluster (63% “S” samples). “L” samples in the LMRP cluster were not labelled with claims different from those in the HMRP cluster. The differences of the spaghetti diameters in the LMRP and HMRP cluster were not significant. This indicates only a minor influence of a possible longer drying time or of the higher ratio between surface and volume of spaghetti with a higher diameter on the accumulation of MRPs. Cluster analysis was used as a starting point for unraveling possible heating markers. The differences in the concentration ranges of individual MRPs in the two clusters are compiled in Table 2. Significant differences between the clusters were found for all MRPs except N-εfructosyllysine. This might mean that N-ε-fructosyllysine is already degraded at the temperatures applied for industrial pasta drying. Data ranges are overlapping between the clusters, with the exception of furosine, pyrraline and maltosine. Nevertheless, medians in the HMRP cluster are more than twice as high as those in the LMRP cluster for the products pyrraline, maltosine, HMF, and glucosyl isomaltol. The most pronounced differences were identified for glucosyl isomaltol. Regarding the whole set of samples, variation limits larger than one order of magnitude were measured for N-ε-maltulosyllysine, pyrraline, maltosine, HMF, and glucosyl isomaltol. All these products are formed during degradation of maltose and maltooligosaccharides (Hellwig et al., 2016a). Primarily these products need to be considered as potential heating markers in systematic drying studies. Another maltose degradation product, maltulose, has already been suggested as a heating marker for pasta products (García-Baños et al., 2004). Resmini and Pellegrino (1994) had observed a very strong increase of the pyrraline concentrations when furosine concentrations grew over 300 mg/ 100 g protein. This biphasic course was corroborated in the present study (Fig. 3A), and the same trend was observed for other MRPs, such as glucosyl isomaltol (Fig. 3B), maltosine, and HMF. Strong correlations between individual compounds, such as pyrraline and maltosine, were observed (Fig. 3C). The compounds behaved similarly in the beer proteome, and this might be due to the fact that both products are formed from maltose under the same heating conditions and eventually share the same intermediate precursors (Hellwig et al., 2016b). The use of ratios between individual MRPs might also prove valuable: The molar ratio between protein-bound pyrraline and formyline is above 1.6 in more than 90% of the samples in the HMRP cluster, but always lower than 1.6 in the samples of the LMRP cluster. The ratio increases when the pyrraline concentration increases 88

Journal of Food Composition and Analysis 72 (2018) 83–92

M. Hellwig et al.

Fig. 3. Analysis of correlations between (A) protein-bound furosine and protein-bound pyrraline, (B) protein-bound furosine and glucosyl isomaltol, (C) protein-bound pyrraline and protein-bound maltosine, and (D) the molar ratio between pyrraline and formyline and the concentration of protein-bound pyrraline in all pasta samples. Open signs, LMRP cluster samples; closed signs, HMRP cluster samples; squares, samples with descriptions concerning low heat treatment; circles, samples without respective descriptions.

Table 3 Concentrations of Glycated Amino Acids in Cooked and Uncooked Pasta Samples of Different Formatsa. pasta formatb

N-ε-maltulosyllysine

N-ε-fructosyllysine

CML

pyrraline

formyline

maltosine

MG-H1

twisted pipes uncooked

47.7 ± 2.5

5.9 ± 0.5

0.77 ± 0.03

cooked (8 min)

42.3 ± 1.8

5.3 ± 0.1

0.62 ± 0.03

0.24 ± 0.03 0.25 ± 0.01

0.10 ± 0.02 0.11 ± 0.01

0.042 ± 0.003 0.059 ± 0.012

0.19 ± 0.03 0.23 ± 0.03

corkscrew uncooked

32.8 ± 1.1

7.1 ± 0.7

cooked (9 min)

36.4 ± 3.1

6.3 ± 0.5

0.53 ± 0.04 0.55 ± 0.01

0.24 ± 0.01 0.23 ± 0.01*

0.09 ± 0.01 0.10 ± 0.01

0.037 ± 0.004 0.060 ± 0.010

0.19 ± 0.01 0.22 ± 0.02

butterflies (farfalle) uncooked

18.2 ± 0.4

3.7 ± 0.8

18.5 ± 1.1

3.2 ± 0.3

0.06 ± 0.02 0.04 ± 0.02

0.06 ± 0.01 0.06 ± 0.01*

tr

cooked (11 min)

0.27 ± 0.02 0.24 ± 0.03

0.07 ± 0.01 0.08 ± 0.01*

gnocchi uncooked

40.7 ± 8.4

5.7 ± 1.1

cooked (9 min)

46.0 ± 0.4

5.8 ± 0.5

0.63 ± 0.05 0.64 ± 0.01

0.36 ± 0.01 0.38 ± 0.09

0.12 ± 0.01 0.12 ± 0.01

0.048 ± 0.002 0.084 ± 0.010*

0.18 ± 0.03 0.25 ± 0.01

spaghetti uncooked

51.5 ± 0.8

9.4 ± 0.6

cooked (11 min)

53.1 ± 2.9

8.4 ± 0.6

1.35 ± 0.01 1.28 ± 0.05

0.60 ± 0.02 0.67 ± 0.01*

0.16 ± 0.02 0.16 ± 0.01

0.14 ± 0.01 0.28 ± 0.03*

0.50 ± 0.05 0.58 ± 0.07

tr

Asterisks (*) denote statistically significant differences (p < 0.05) as evaluated by a t test (cooked vs. uncooked). a Data are given as means ± S.D. (n = 2) and expressed as leucine equivalents [μmol/mmol leucine]; tr, trace amounts (between LOD and LOQ). b Cooking time followed the recommendations of the producer and is given in parentheses. 89

Journal of Food Composition and Analysis 72 (2018) 83–92

M. Hellwig et al.

Fig. 4. Proposed pathways of formation of protein-bound maltosine during the cooking process. Circles refer to a protein residue.

treatments for their formation. The contribution of pasta consumption to the dietary exposure to pyrraline (0.6 mg/portion vs. 20–40 mg daily intake; Hellwig and Henle, 2012), formyline (0.2 mg/portion vs. 2–3 mg daily intake; Hellwig and Henle, 2012), and maltosine (0.1 mg/portion vs. 1–2 mg daily intake; Hellwig et al., 2016a) is therefore lower. This also applies to MG-H1 (0.6 mg/portion vs. 15–29 mg daily intake; Scheijen et al., 2018). Protein-bound Maillard reaction products, such as fructosyllysine, pyrraline and formyline, can be released during intestinal digestion in the form of glycated peptides and free glycated amino acids that are available for absorption (Hellwig and Henle, 2013; Hellwig et al., 2014). The amino acids pyrraline, CML, and formyline—but not fructosyllysine—can be absorbed into enterocytes when they are bound in dipeptides (Hellwig et al., 2011). Due to possible interaction with targets in the human body, dietary MRPs are often referred to as a risk for human health. As yet, this has not been unequivocally proven by reliable structure-activity relationships (Poulsen et al., 2013). It is not possible to attribute a risk to the consumption of pasta products. However, future studies on health effects of MRPs should also include pasta consumption due to its quantitative importance. Disadvantageously, lysine as an essential amino acid is partly blocked during pasta production, which lowers the nutritional value of the pasta protein. This is important primarily for children during growth. However, as compared to earlier published degrees of lysine blockage of up to 60% in pasta (Resmini and Pellegrino, 1994; Henle et al., 1995; Pagani et al., 1996; Acquistucci, 2000; Cavazza et al., 2013), the drying conditions applied today obviously lead to a lower degree of lysine damage. The highest degree of lysine blockage was 27% in the present study, which is comparable to milk powder and infant formulas (Mehta and Deeth, 2016).

Table 4 Amounts (mg) of Individual Maillard Reaction Products Ingested by Consumption of a 150-g Portion of Spaghetti[a]. compound

range

median

N-ε-maltulosyllysine N-ε-fructosyllysine CML pyrraline formyline maltosine MG-H1 3-DG glucosyl isomaltol HMF

30-340 16-51 0.5-2.4 tr-3.1 0.1-0.6 tr-0.4 tr-1.1 0.4-1.3 0.03-1.2 n.d.-0.08

206 34 1.4 0.6 0.2 0.1 0.6 0.7 0.3 0.02

Data are given in mg, based on the median levels and ranges compiled in Table 2 for all investigated spaghetti samples.

reported (Kramhöller et al., 1993). We conclude that, possibly with the exception of maltosine, the concentrations of MRPs in uncooked pasta can be taken as a basis for the estimation of the dietary intake of MRPs. An overview of the amounts of glycated amino acids ingested with a portion of spaghetti is given in Table 4. These amounts are based on 150 g dried spaghetti per portion. The consumption of pasta contributes significantly to the total dietary exposure to MRPs, especially of ARPs. When a daily intake of ARPs of 500–1200 mg (calculated as N-ε-fructosyllysine) is taken as a basis (Henle, 2003), the ingestion of one portion of pasta supplies 20–40% of this amount. A significant contribution may also be discussed for CML, for which a daily intake between 2.2 and 11.3 mg was estimated (Birlouez-Aragon et al., 2010; Delgado-Andrade et al., 2012; Scheijen et al., 2018). Between 10 and 60% of these amounts may be taken up with pasta with a median content of 1.4 mg/portion. The further glycated amino acids analyzed in this study require higher heat 90

Journal of Food Composition and Analysis 72 (2018) 83–92

M. Hellwig et al.

4. Conclusions

Guerra-Hernández, E., Ramirez-Jiménez, A., García-Villanova, B., 2002. Glucosylisomaltol, a new indicator of browning reaction in baby cereals and bread. J. Agric. Food Chem. 50, 7282–7287. Hellwig, M., Henle, T., 2012. Quantification of the Maillard reaction product 6-(2-formyl1-pyrrolyl)-L-norleucine (formyline) in food. Eur. Food Res. Technol. 235, 99–106. Hellwig, M., Henle, T., 2013. Release of pyrraline in absorbable peptides during simulated digestion of casein glycated by 3-deoxyglucosone. Eur. Food Res. Technol. 237, 47–55. Hellwig, M., Henle, T., 2014. Baking, ageing, diabetes: a short history of the Maillard reaction. Angew. Chem. Int. Ed. Engl. 53, 10316–10329. Hellwig, M., Geissler, S., Matthes, R., Peto, A., Silow, C., Brandsch, M., Henle, T., 2011. Transport of free and peptide-bound glycated amino acids: synthesis, transepithelial flux at caco-2 cell monolayers, and interaction with apical membrane transport proteins. ChemBioChem 12, 1270–1279. Hellwig, M., Matthes, R., Peto, A., Löbner, J., Henle, T., 2014. N-ε-fructosyllysine and N-εcarboxymethyllysine, but not lysinoalanine, are available for absorption after simulated gastrointestinal digestion. Amino Acids 46, 289–299. Hellwig, M., Bunzel, D., Huch, M., Franz, C.M.A.P., Kulling, S.E., Henle, T., 2015. Stability of individual Maillard reaction products in the presence of the human colonic microbiota. J. Agric. Food Chem. 63, 6723–6730. Hellwig, M., Kiessling, M., Rother, S., Henle, T., 2016a. Quantification of the glycation compound 6-(3-hydroxy-4-oxo-2-methyl-4(1H)-pyridin-1-yl)-l-norleucine (maltosine) in model systems and food samples. Eur. Food Res. Technol. 242, 547–557. Hellwig, M., Witte, S., Henle, T., 2016b. Free and protein-bound Maillard reaction products in beer: method development and a survey of different beer types. J. Agric. Food Chem. 64, 7234–7243. Hellwig, M., Rückriemen, J., Sandner, D., Henle, T., 2017. Unique pattern of proteinbound Maillard reaction products in Manuka (Leptospermum scoparium) honey. J. Agric. Food Chem. 65, 3532–3540. Helou, C., Gadonna-Widehem, P., Robert, N., Branlard, G., Thebault, J., Librere, S., Jacquot, S., Mardon, J., Piquet-Pissaloux, A., Chapron, S., Chatillon, A., NiquetLéridon, C., Tessier, F.J., 2016. The impact of raw materials and baking conditions on Maillard reaction products, thiamine, folate, phytic acid and minerals in white bread. Food Funct. 7, 2498–2507. Henle, T., 2003. AGEs in food: Do they play a role in uremia? Kidney Int. 63, S145–S147. Henle, T., Bachmann, A., 1996. Synthesis of pyrraline reference material. Z. Lebensm. Unters. Forsch. 202, 72–74. Henle, T., Zehetner, G., Klostermeyer, H., 1995. Fast and sensitive determination of furosine. Z. Lebensm. Unters. Forsch. 200, 235–237. Kramhöller, B., Pischetsrieder, M., Severin, T., 1993. Maillard reactions of lactose and maltose. J. Agric. Food Chem. 41, 347–351. Krause, R., Knoll, K., Henle, T., 2003. Studies on the formation of furosine and pyridosine during acid hydrolysis of different Amadori products of lysine. Eur. Food Res. Technol. 216, 277–283. Lintas, C., D’Appolonia, B.L., 1973. Effect of spaghetti processing on semolina carbohydrates. Cereal Chem. 50, 563–570. Loaëc, G., Niquet-Léridon, C., Henry, N., Jacolot, P., Jouquand, C., Janssens, M., Hance, P., Cadalen, T., Hilbert, J.L., Desprez, B., Tessier, F.J., 2015. Impact of variety and agronomic factors on crude protein and total lysine in chicory; Nε-carboxymethyllysine-forming potential during drying and roasting. J. Agric. Food Chem. 63, 10295–10302. Mæhre, H.K., Dalheim, L., Edvinsen, G.K., Elvevoll, E.O., Jensen, I.-J., 2018. Protein determination—method matters. Foods 7, 5. Mehta, B.M., Deeth, H.C., 2016. Blocked lysine in dairy products: formation, occurrence, analysis, and nutritional implications. Comp. Rev. Food Sci. Food Saf. 15, 206–218. Mercier, S., Mondor, M., Moresoli, C., Villeneuve, S., Marcos, B., 2016. Drying of durum wheat pasta and enriched pasta: a review of modeling approaches. Crit. Rev. Food Sci. Nutr. 56, 1146–1168. Padalino, L., Caliandro, R., Chita, G., Conte, A., Del Nobile, M.A., 2016. Study of drying process on starch structural properties and their effect on semolina pasta sensory quality. Carbohydr. Polym. 153, 229–235. Pagani, M.A., De Noni, I., Resmini, P., Pellegrino, L., 1996. Filiera produttiva e danno termico della pasta alimentare secca. Tecnica Molitoria 47, 345–361. Poulsen, M.W., Hedegaard, R.V., Andersen, J.M., De Courten, B., Bügel, S., Nielsen, J., Skibsted, L.H., Dragsted, L.O., 2013. Advanced glycation end products in food and their effects on health. Food Chem. Toxicol. 60, 10–37. Resmini, P., Pellegrino, L., 1994. Occurrence of protein-bound lysylpyrrolaldehyde in dried pasta. Cereal Chem. 71, 254–262. Resmini, P., Pellegrino, L., Pagani, M.A., De Noni, I., 1993. Formation of 2-acetyl-3-Dglucopyranosylfuran (glucosylisomaltol) from nonenzymatic browning in pasta drying. Ital. J. Food Sci. 5, 116–129. Scheijen, J.L.J.M., Clevers, E., Engelen, L., Dahnelie, P.C., Brouns, F., Stehouwer, C.D.A., Schalkwijk, C.G., 2016. Analysis of advanced glycation endproducts in selected food items by ultra-performance liquid chromatography tandem mass spectrometry: presentation of a dietary AGE database. Food Chem. 190, 1145–1150. Scheijen, J.L.J.M., Hanssen, N.M.J., Van Greevenbroek, M.M., Van der Kallen, C.J., Feskens, E.J.M., Stehouwer, C.D.A., Schalkwijk, C.G., 2018. Dietary intake of advanced glycation endproducts is associated with higher levels of advanced glycation endproducts in plasma and urine: the CODAM study. Clin. Nutr. 37, 919–925. Sicignano, A., Di Monaco, R., Masi, P., Cavella, S., 2015. From raw material to dish: pasta quality step by step. J. Sci. Food Agric. 95, 2579–2587. Stuknytė, M., Cattaneo, S., Pagani, M.A., Marti, A., Micard, V., Hogenboom, J., De Noni, I., 2014. Spaghetti from durum wheat: effect of drying conditions on heat damage, ultrastructure and in vitro digestibility. Food Chem. 149, 40–46. Testani, E., Giannetti, V., Mariani, M.B., Mannino, P., 2017. Maillard reaction products as markers of the durum wheat pasta manufacturing process: a commodity

In the present study, HPLC-MS/MS and HPLC-UV techniques were applied for the examination of MRPs in differently dried and differently shaped pasta products. N-ε-Maltulosyllysine was identified as the main ARP, whereas CML and pyrraline were the major AGEs. The furosine concentrations can be fully explained by the presence of only the two ARPs N-ε-fructosyllysine and N-ε-maltulosyllysine. Due to large data variation limits, N-ε-maltulosyllysine, pyrraline, maltosine, HMF, and glucosyl isomaltol have the highest potential to serve as sensitive markers for heat treatment. In particular, the biphasic courses of maltose degradation products (maltosine and glucosyl isomaltol) when plotted against furosine, as well as molar ratios of individual MRPs, deserve further attention. Glycated amino acids are stable during the cooking process. For maltosine, a significant formation during the cooking process was determined. The consumption of pasta products contributes significantly to the daily intake of ARPs and CML. It is a major drawback of the present study that the heat treatment conditions applied during production of the pasta samples are unknown and that mislabeling cannot be excluded. The application of the presented analytical methods for reliable conclusions on the drying processes will require model drying experiments with differently shaped pasta products. It will also be necessary to analyze the kinetics of formation of individual MRPs in pasta products; ratios between different products (Fig. 3) will probably give information about critical temperature ranges that have been reached. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Acknowledgments The authors want to thank Karla Schlosser, Institute of Food Chemistry, for performing the amino acid analyses. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.jfca.2018.06.009. References Acquistucci, R., 2000. Influence of Maillard reaction on protein modification and colour development in pasta. Comparison of different drying conditions. Lebensm. Wiss. Technol. 33, 48–52. Anese, M., Nicoli, M.C., Massini, R., Lerici, C.R., 1999. Effects of drying processing on the Maillard reaction in pasta. Food Res. Int. 32, 193–199. Birlouez-Aragon, I., Saavedra, G., Tessier, F.J., Galinier, A., Ait-Ameur, L., Lacoste, F., Niamba, C.-N., Alt, N., Somoza, V., Lecerf, J.-M., 2010. A diet based on high-heattreated foods promotes risk factors for diabetes mellitus and cardiovascular diseases. Am. J. Clin. Nutr. 91, 1220–1226. Cavazza, A., Corradini, C., Rinaldi, M., Salvadeo, P., Borromei, C., Massini, R., 2013. Evaluation of pasta thermal treatment by determination of carbohydrates, furosine, and color indices. Food Bioprocess. Technol. 6, 2721–2731. Degen, J., Hellwig, M., Henle, T., 2012. 1,2-Dicarbonyl compounds in commonly consumed foods. J. Agric. Food Chem. 60, 7071–7079. Delgado-Andrade, C., Tessier, F.J., Niquet-Léridon, C., Seiquer, I., Navarro, M.P., 2012. Study of the urinary and faecal excretion on Nε-carboxymethyllysine in young human volunteers. Amino Acids 43, 595–602. García-Baños, J.L., Corzo, N., Sanz, M.L., Olano, A., 2004. Maltulose and furosine as indicators of quality of pasta products. Food Chem. 88, 35–38. Geissler, S., Hellwig, M., Markwardt, F., Henle, T., Brandsch, M., 2011. Synthesis and intestinal transport of the iron chelator maltosine in free and dipeptide form. Eur. J. Pharm. Biopharm. 78, 75–82. Giannetti, V., Boccacci Mariani, M., Mannino, P., Testani, E., 2014. Furosine and flavour compounds in durum wheat pasta produced under different manufacturing conditions: multivariate chemometric characterization. LWT Food Sci. Technol. 56, 15–20. Goodwin, J.C., 1983. Isolation of 3-O-α-D-gluco and 3-O-β-D-galacto-pyranosyloxy-2-furyl methyl ketones from nonenzymic browning of maltose and lactose with secondary amino acids. Carbohydr. Res. 115, 281–287.

91

Journal of Food Composition and Analysis 72 (2018) 83–92

M. Hellwig et al.

Wellner, A., Nußpickel, L., Henle, T., 2012. Glycation compounds in peanuts. Eur. Food Res. Technol. 243, 423–429. Zweifel, C., Handschin, S., Escher, F., Conde-Petit, B., 2003. Influence of high-temperature drying on structural and textural properties of durum wheat pasta. Cereal Chem. 80, 159–167.

investigation. Acta Aliment. 46, 267–274. Thornalley, P.J., Battah, S., Ahmed, N., Karachalias, N., Agalou, S., Babaei-Jadidi, R., Dawnay, A., 2003. Quantitative screening of advanced glycation endproducts in cellular and extracellular proteins by tandem mass spectrometry. Biochem. J. 375, 581–592.

92