Stability of Individual Maillard Reaction Products in the Presence of

Jul 17, 2015 - Free and Protein-Bound Maillard Reaction Products in Beer: Method Development and a Survey of Different Beer Types. Michael Hellwig ...
1 downloads 0 Views 562KB Size
Article pubs.acs.org/JAFC

Stability of Individual Maillard Reaction Products in the Presence of the Human Colonic Microbiota Michael Hellwig,†,§ Diana Bunzel,§ Melanie Huch,§ Charles M. A. P. Franz,§,# Sabine E. Kulling,§ and Thomas Henle*,† †

Institute of Food Chemistry, Technische Universität Dresden, D-01062 Dresden, Germany Department of Safety and Quality of Fruits and Vegetables, Max Rubner-Institut, Federal Research Institute of Nutrition and Food, Haid-und-Neu-Straße 9, 76131 Karlsruhe, Germany

§

S Supporting Information *

ABSTRACT: Maillard reaction products (MRPs) are taken up in substantial amounts with the daily diet, but the majority are not transported across the intestinal epithelium. The aim of this study was to obtain first insights into the stability of dietary MRPs in the presence of the intestinal microbiota. Four individual MRPs, namely, N-ε-fructosyllysine (FL), N-εcarboxymethyllysine (CML), pyrraline (PYR), and maltosine (MAL), were anaerobically incubated with fecal suspensions from eight human volunteers at 37 °C for up to 72 h. The stability of the MRPs was measured by HPLC with UV and MS/MS detections. The Amadori product FL could no longer be detected after 4 h of incubation. Marked interindividual differences were observed for CML metabolism: Depending on the individual, at least 40.7 ± 1.5% of CML was degraded after 24 h of incubation, and the subjects could thus be tentatively grouped into fast and slow metabolizers of this compound. PYR was degraded by 20.3 ± 4.4% during 24 h by all subjects. The concentration of MAL was not significantly lowered in the presence of fecal suspensions. In no case could metabolites be identified and quantified by different mass spectrometric techniques. This is the first study showing that the human colonic microbiota is able to degrade selected glycated amino acids and possibly use them as a source of energy, carbon, and/or nitrogen. KEYWORDS: Maillard reaction, Amadori product, advanced glycation end product, colonic microbiota, metabolism



INTRODUCTION

During processing and storage of foods, proteins are subject to reactions with reducing carbohydrates in the Maillard reaction (nonenzymatic browning, glycation).1,2 In the early stage of the reaction, Amadori rearrangement products (ARPs) are formed from reducing sugars and α- as well as ε-amino groups of free amino acids, peptides, and proteins, for example, N-εfructosyllysine (FL; Figure 1) from glucose and the ε-amino group of (peptide-bound) lysine. The ARPs are degraded in the advanced stage of the reaction, giving rise to vicinal dicarbonyl compounds such as 3-deoxyglucosone (3-DG), glyoxal, and methylglyoxal (MGO), depending on the environmental conditions (oxidative/nonoxidative, temperature, pH value).3,4 In the final stage of the reaction, these reactive compounds react with nucleophilic sites of free amino acids, peptides, and proteins, predominantly the side chains of lysine and arginine. N-ε-Carboxymethyllysine (CML; Figure 1) is generated during oxidative degradation of Amadori products or by reaction of lysine with glyoxal. Pyrraline (PYR; Figure 1) is a condensation product of lysine and 3-deoxyglucosone.5 Maltosine (MAL; Figure 1) is formed from lysine during disaccharide degradation.6 On the basis of reports on concentrations of defined MRPs in foods, a daily intake of 500−1000 mg of ARPs, 20 mg of CML, and 20−40 mg of PYR can be estimated especially from the consumption of heat-treated food items such as milk products, bakery products, and coffee.7−10 Quantitative data on the occurrence of maltosine in foods are © XXXX American Chemical Society

Figure 1. Chemical structures of the investigated products: 1, N-εfructosyllysine (FL); 2, N-ε-carboxymethyllysine (CML); 3, pyrraline (PYR); 4, maltosine (MAL).

limited to its qualitative identification in heated evaporated milk.11 Received: March 18, 2015 Revised: June 22, 2015 Accepted: July 17, 2015

A

DOI: 10.1021/acs.jafc.5b01391 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry The formation of these lysine derivatives (“lysine blockage”) implicates a decrease of the nutritional value of proteins.12,13 Moreover, the accessibility by intestinal peptidases is impaired; thus, the release also of other proteinogenic amino acids into absorbable units is hampered.14,15 However, at levels of lysine modification that are relevant in food (8−15%), the digestibility of casein is not significantly lowered.16,17 Individual Maillard reaction products (MRPs) can be released from glycated proteins as small peptides and free amino acids, similarly to the proteinogenic amino acids,16,17 but they cannot be transported across the intestinal membrane via amino acid carriers. Bound in small peptides, some MRPs can be translocated into epithelial cells via the di- and tripeptide transporter PEPT1. After intracellular peptidolysis, glycated amino acids with polar or charged side chains (e.g., CML) are “trapped” inside the cells, whereas those with unpolar side chains (e.g., PYR and MAL) can pass the basolateral membrane.18−20 It has been shown in several human studies that dietary FL can pass into the circulation only to a very small extent, whereas 50−100% of dietary PYR must be absorbed as it appears in the urine after oral administration.7,21,22 Unabsorbed amino acids and peptides, as well as intracellularly accumulated amino acids released during desquamation, will pass into the colon. A passage of approximately 13 g per day of protein (endogenous protein from intestinal peptidases and mucins, as well as recalcitrant dietary protein) has been estimated.23 The distal colonic microbiota can degrade large peptides and proteins, and all proteinogenic amino acids can then be fermented by anaerobic bacteria.24 The predominant products are ammonia, carbon dioxide, hydrogen, and short-chain fatty acids (SCFA; acetic, propionic, and butyric acids). Branched-chain amino acids are degraded to branched-chain fatty acids (e.g., isobutyrate, isovalerate, and 2methylbutyrate), which can serve as urinary markers of colonic protein fermentation. Products such as phenol, cresol, indole, and substituted acetic and propionic acids mainly result from the fermentation of aromatic amino acids. Due to the toxicological properties of some of these products, protein fermentation in the colon is often regarded as detrimental.24,23,25 Glycated proteins passing into the colon can have specific effects on the composition of the colonic microbiota: On incubation with glycated bovine serum albumin, the amount of clostridia, sulfate-reducing bacteria, and Bacteroides in the microbiota from humans suffering from ulcerative colitis, but not that from healthy subjects, increased, whereas the amount of eubacteria and bifidobacteria decreased.26 An inhibitory effect on the growth of lactobacilli was also found in the colonic microbiota of adolescents consuming a diet rich in MRPs (bread with crust, grilled meat), compared to a diet low in MRPs (bread without crust, baked meat).27 Protein glycated with different sugars (galactose, lactulose, and galactooligosaccharides), however, can induce a bifidogenic effect on the human microbiota in vitro.28,29 The structural basis of these findings remains unclear. In particular, the stability of individual MRPs in the colon has been only poorly investigated. FL is the only MRP that was found to be unstable in the presence of rat and human intestinal microbiota in vitro and in vivo.7,21,30 Several enzymatic degradation mechanisms have been described.31,32 It has also been postulated that CML is degraded by the human colonic microbiota, because the recovery of the substance in urine and feces after oral administration never reaches 100%, either in human or in rat studies.33,34

As the human colonic microbiota has the ability to metabolize all proteinogenic amino acids,24 and because it is continuously exposed to substantial amounts of unabsorbed dietary MRPs, we hypothesized that the human colonic microbiota is able to degrade MRPs. Four individual MRPs were synthesized and incubated in the presence of fecal suspensions from eight human volunteers. The stability of the MRPs was measured by HPLC with UV or MS/MS detection.



MATERIALS AND METHODS

Chemicals. Nonafluoropentanoic acid (NFPA), resazurin sodium salt, and cysteine hydrochloride monohydrate were from Sigma (Steinheim, Germany). Acetic acid for HPLC was obtained from Merck (Darmstadt, Germany). HPLC grade methanol and acetonitrile were from VWR Prolabo (Darmstadt, Germany). [2H2]CML was obtained from PolyPeptide (Strasbourg, France). The synthesis of the following MRPs was performed according to the literature stated: FL,35 [13C6,15N2]FL,36 CML,20 PYR,18 and MAL.19 Brain−heart infusion (BHI) medium was used for anaerobic cultivation and prepared as follows: 18.5 g of brain−heart infusion broth (Merck) and 0.28 g of cysteine hydrochloride monohydrate were dissolved in distilled water, and 50 μL of an aqueous resazurin solution (1%, w/v) was added. The pH was adjusted to 7.6 with 5 M sodium hydroxide, and water was added to a final volume of 500 mL. The required volume of medium was filled into Hungate tubes. The filled Hungate tubes were subjected to three cycles of first vacuum and then N2 overpressure application to remove oxygen. The media were autoclaved (121 °C, 15 min). The medium needed for the preparation of fecal suspensions was allowed to equilibrate overnight under oxygen-free conditions in an anaerobic workstation (MG 500, Don Whitley, Shipley, UK). Incubation of MRPs in the Presence or Absence of Human Fecal Inocula. Feces samples were obtained from eight healthy volunteers (6 males, 2 females) aged between 25 and 55 years who consumed a Western diet (no vegetarians). The volunteers had not taken antibiotics in the previous 3 months. One experiment with a feces sample from one volunteer was performed per week. The preparation of fecal suspensions was performed as described previously37 in an anaerobic workstation under an atmosphere consisting of N2/CO2/H2 (80:10:10, v/v/v). Feces samples were suspended in BHI medium to a concentration of 25% (w/v). The suspensions were shaken (10 min, 180 rpm) and then centrifuged (10 min, 300g). The supernatants were removed carefully and directly utilized for the incubations as the native fecal suspension. In addition, an aliquot (20 mL) of the suspension was autoclaved (121 °C, 20 min). Aqueous solutions (5 mM) of the MRPs were prepared and sterilefiltered (0.2 μm). The stability of the MRPs in the presence of BHI medium was assessed by adding 0.5 mL of the 5 mM MRP solutions to 9.5 mL of autoclaved oxygen-free BHI medium in a Hungate tube (V = 16 mL). The stability of the MRPs in the presence of an autoclaved feces sample was assessed by adding 0.5 mL of the 5 mM MRP solutions and 1.0 mL of autoclaved fecal suspension to 8.5 mL of autoclaved, oxygen-free BHI medium. These controls were done in a single determination in parallel to each experiment. The stability of the MRPs in the presence of native feces samples was analyzed by adding 0.5 mL of the 5 mM MRP solutions and 1.0 mL of native fecal suspension to 8.5 mL of autoclaved oxygen-free BHI medium. The latter incubation was performed in triplicate for each feces sample. All MRPs were incubated separately. The stability of medium components was analyzed by mixing 1.0 mL of native fecal suspension and 9.0 mL of autoclaved oxygen-free BHI medium. Samples were withdrawn directly after the addition of the fecal suspension (0 h) and then after 4, 24, 48, and 72 h. A portion of sterilized gas was injected, and 1.0 mL of the suspension was removed. All samples were then frozen and stored at −80 °C. Preparation of Samples for Analysis. For the analysis of PYR and MAL, 50 μL of the samples was mixed with 950 μL of 0.1% acetic acid in 5% aqueous methanol. The mixture was shaken vigorously and B

DOI: 10.1021/acs.jafc.5b01391 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry Table 1. Transitions Recorded during Measurement of FL and CML in the MRM Modea transition

fragmentor voltage (V)

collision energy (eV)

Q/qb

FL

m/z 309 → m/z 84 m/z 309 → m/z 225

125 125

36 13

Q q

[13C6,15N2]FL

m/z 317 → m/z 233 m/z 317 → m/z 129

70 70

13 15

Q q

CML

m/z 205 → m/z 130 m/z 205 → m/z 142

100 100

9 9

Q q

[2H2]CML

m/z 207 → m/z 130 m/z 207 → m/z 144

80 80

5 12

Q q

analyte

a General conditions: positive mode; dwell time, 120 ms. bQ, transition used for quantification; q, transition used to confirm the presence of the analyte.

centrifuged (10 min, 13000 rpm). The supernatant was transferred to an HPLC vial. For the analysis of FL and CML, 50 μL of the samples was mixed first with 50 μL of methanol and shaken vigorously. After 45 min, 900 μL of 0.1% acetic acid was added; the mixture was shaken and frozen. After the mixture had been thawed and centrifuged (10 min, 13000 rpm), 180 μL of the supernatant was mixed with 20 μL of the internal standard solution containing 0.121 mM [2H2]CML and 0.656 mM [13C6,15N2]FL in water. Pellet extraction followed a protocol employed for the extraction of water-soluble metabolites from Escherichia coli.38 An aliquot of 500 μL of the above samples was centrifuged (10 min, 13000 rpm) and the supernatant removed. A 300 μL aliquot of 5 mM NFPA in 80% aqueous acetonitrile was then added to the pellet. After vigorous shaking, the suspension was stored at −80 °C for 30 min. After the suspension had been thawed and centrifuged (2 min, 13000 rpm), the supernatant was transferred to another tube and the pellet was extracted two more times. The supernatants were combined, and 300 μL of water was added. From these solutions, 200 μL was mixed with 800 μL of a solution of 0.1% acetic acid in 5% aqueous methanol. The mixtures were centrifuged (10 min, 13000 rpm) and the supernatants transferred to HPLC vials. High-Pressure Liquid Chromatography with UV Detection. HPLC-DAD analyses were performed on a LaChrom Elite system consisting of a pump L-2130, an autosampler L-2200, a column oven L-2300, and a diode array detector L-2450 (all from VWR Hitachi, Darmstadt, Germany). A stainless steel column (150 mm × 4.6 mm, 5 μm) filled with Eurospher 100 C-18 material with an integrated guard column (5 × 4 mm) of the same material (Knauer, Berlin, Germany) was used for the analysis of PYR.18 For the analysis of MAL, a polymer-based RP-18-column (PLRP-S, 100 Å, 250 mm × 4.6 mm, 8 μm, Polymer Laboratories, Darmstadt, Germany) with a guard column of the same material (5 × 4 mm) was used.19 Eluent A was a solution of 5 mM NFPA in water; eluent B was a solution of 5 mM NFPA in a mixture of 80% methanol and 20% water. During the analysis of PYR, a linear gradient from 13% B to 100% B in 20 min was applied at a flow rate of 0.8 mL/min, whereas during the analysis of MAL, a linear gradient was mixed from 20 to 80% B in 20 min at a flow rate of 1.0 mL/min. During all analyses, the temperature of the autosampler was maintained at 10 °C, and the temperature of the column oven was set at 35 °C. The injection volume was 50 μL. UV spectra were recorded between 200 and 350 nm, and the quantification was performed with the chromatograms extracted at 280 nm (MAL) and 297 nm (PYR), respectively. External calibration was performed with the isolated MRP standards. High-Pressure Liquid Chromatography with Mass Spectrometric Detection. Chromatography was performed with the highpressure gradient system 1200 series (Agilent Technologies, Böblingen, Germany), consisting of a binary pump, an online degasser, a column oven, and an autosampler. A stainless steel column (Zorbax 300 SB-C18, 50 mm × 2.1 mm, 3.5 μm, Agilent Technologies) was used for the analysis of FL and CML. The eluents used were a solution

of 10 mM NFPA in bidistilled water (solvent A) and a solution of 10 mM NFPA in acetonitrile (solvent B). A linear gradient from 5 to 29% B in 8 min was applied at a flow rate of 0.25 mL/min. The column temperature was maintained at 35 °C. The injection volume was 10 μL. The chromatographic system was connected to the mass spectrometer 6410 triple-quad (Agilent Technologies), working in the positive mode with a capillary voltage of 4000 V and a source temperature of 350 °C. FL and CML were quantified by MS detection in the multiple reaction monitoring (MRM) mode using the stable isotope dilution technique. The transitions used for the quantification of the analytes (quantifier, Q) and for the confirmation of the presence of the analytes (qualifier, q) are compiled in Table 1. All transitions were recorded between 1 and 9 min after injection. Data acquisition and evaluation were performed with the software MassHunter B.02.00 (Agilent). The same HPLC system was also used during metabolite analysis. Selected samples were measured first in the scan mode (m/z 90−500) and then in the product ion scan mode. For the latter approach, protonated molecular ions ([M + H]+) of possible metabolites (e.g., biogenic amines, oxidized/reduced species) were stabilized in the first quadrupole and fragmented in the second quadrupole (fragmentor voltage, 100 V; collision energy, 20 eV), and spectra (m/z 50−250) were recorded with the third quadrupole. Statistical Treatment. Comparisons of mean MRP concentrations between inoculated and non-inoculated media were performed using Student’s t test. P values ≤0.05 were considered significant using the software PASW Statistics 18.



RESULTS AND DISCUSSION Incubation Conditions. Significant amounts of proteinbound glycated amino acids are taken up with the daily diet. Unabsorbed glycated amino acids or those positioned at recalcitrant sites of heat-damaged proteins and not released during gastrointestinal digestion can pass into the colon. Although the microbial metabolism of FL has been investigated in vitro and in vivo,7,13,21,22,30 no knowledge has yet been gained about the stability of other MRPs in the presence of the colonic microbiota, especially the human microbiota. Smith and Macfarlane have performed such investigations with the proteinogenic amino acids.24 These authors had worked with a minimal medium and a final feces concentration of 20%. In our study, we wanted to conserve the metabolic variability of the gut microbiota as much as possible, and therefore chose brain−heart infusion broth, which is recommended for fastidious microorganisms, as in earlier investigations.37 Unfortunately, due to the abundance of other amino acids from peptones in the medium, the analysis of SCFA as typical products of microbial amino acid metabolism was not possible. C

DOI: 10.1021/acs.jafc.5b01391 J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Article

Journal of Agricultural and Food Chemistry

Stability of MRPs in the Presence of a Viable Microbiota. The four MRPs at a concentration of 250 μM were then added to 2.5% (w/v) fecal suspensions obtained from eight healthy volunteers. The analysis of FL by HPLCMS/MS allowed quantification of at least 0.2 μM in the undiluted samples, equivalent to 0.08% of the initial FL concentration. Likewise, as low as 2.7 μM CML, corresponding to 1.1% of the initial amount, could be quantified in the undiluted samples. The recoveries of FL and CML, respectively, after addition to the non-inoculated medium, were 94.2 ± 5.1 and 92.1 ± 2.7%, respectively. For data evaluation, the MRP concentration measured in the noninoculated medium was regarded as 100%, and all other values were expressed as percentages of this concentration. As could be expected from the literature,7,21,30 the concentration of FL declined very quickly in the presence of the viable microbiota (Figure 3A; Table S1). In some cases, >80% of the initial amount had already disappeared in the sample taken directly after the addition to the fecal suspension. The time between sampling inside the anaerobic chamber and deep-freezing (approximately 30−45 min) must have been sufficient for the degradation of the substance. This short time could probably point to metabolism of the substance by extracellular enzymes, as already proposed by Griffiths and Pridham.41 After 4 h of incubation, FL was in each case not detectable anymore. The separation by centrifugation of the samples taken from the inocula and the independent analysis of supernatants and pellets after extraction did not lead to a higher recovery of FL; that is, at no time was FL detectable in the cell pellet extract. Our data therefore suggest that the low recovery of orally administered FL in urine and feces21,22 can at least in part be explained by metabolism of the compound by the intestinal microbiota. CML was also degraded in the presence of all fecal samples, but to a lesser extent than FL. After 4 and 24 h, average amounts of 84.6% (range of 45−98%) and 36.2% (range of 0− 59%) remained, respectively (Figure 3B; Table S2). Only about 1% of the initial CML amount was detectable in the pellet extract for samples taken at different incubation times. The interindividual differences of CML degradation were very pronounced. After 24 h of incubation, CML was degraded to