Habitual consumption of medium amounts of coffee over the whole life-span is hypothesized to reduce the risk to develop diabetes type 2 (DM2) and Alzheimer's disease (AD). To identify putative bioactive coffee-derived metabolites, first, pooled urine from coffee drinkers and non-coffee drinkers were screened by UPLC-HDMS. After statistical data analysis, trigonelline, dimethylxanthines and monomethylxanthines, and ferulic acid conjugates were identified as the major metabolites found after coffee consumption. For quantitative analysis of these markers in body fluids, targeted methods based on stable-isotope dilution and UPLC-MS/MS were developed and applied to plasma samples from a coffee intervention study (n = 13 volunteers) who consumed a single cup of caffeinated coffee brew after a 10-day washout period. Chlorogenic acid-derived metabolites were found to be separated into two groups showing different pharmacokinetic properties. The first group comprised, e.g., ferulic acid and feruloyl sulfate and showed early appearance in the plasma (~1 h). The second group contained particularly chlorogenic acid metabolites formed by the intestinal microflora, appearing late and persisting in the plasma (>6 h). Trigonelline appeared early but persisted with calculated half-life times ~5 h. The plasma levels of caffeine metabolites significantly and progressively increased 2-4 h after coffee consumption and did not reach c max within the time frame of the study. The pharmacokinetic profiles suggest that particularly trigonelline, caffeine, its metabolites, as well as late appearing dihydroferulic acid, feruloylglycine and dihydroferulic acid sulfate formed from chlorogenic acid by the intestinal microflora accumulate in the plasma due to their long half-life times during habitual consumption of several cups of coffee distributed over the day. Since some of these metabolites have been reported to show antioxidant effects in vivo, antioxidant-response-element activating potential, and neuroprotective properties, respectively, some of these key metabolites might account for the inflammation- and DM2/AD risk reducing effects reported for habitual life time consumption of coffee.

Proline betaine has been proposed as a candidate dietary biomarker for citrus intake. To validate its suitability as a dietary biomarker and to gain insight into the range of this per-methylated amino acid in foods and beverages, a quick and accurate stable isotope dilution assay was developed for quantitative high-throughput HILIC-MS/MS screening of proline betaine in foods and urine after solvent-mediated matrix precipitation. Quantitative analysis of a variety of foods confirmed substantial amounts of proline betaine in citrus juices (140-1100 mg/L) and revealed high abundance in tubers of the vegetable Stachys affinis, also known as Chinese artichocke (∼700 mg/kg). Seafood including clams, shrimp, and lobster contained limited amounts (1-95 mg/kg), whereas only traces were detected in fish, cuttlefish, fresh meat, dairy products, fresh vegetable (<3 mg/kg), coffee, tea, beer, and wine (<7 mg/L). The human excretion profiles of proline betaine in urine were comparable when common portions of orange juice or fried Stachys tubers were consumed. Neither mussels nor beer provided enough proline betaine to detect significant differences between morning urine samples collected before and after consumption. As Stachys is a rather rare vegetable and not part of peoples' daily diet, the data reported here will help to monitor the subject's compliance in future nutritional human studies on citrus products or the exclusion of citrus products in the wash-out phase of an intervention study. Moreover, proline betaine measurement can contribute to the establishment of a toolbox of valid dietary biomarkers reflecting wider aspects of diet to assess metabolic profiles as measures of dietary exposure and indicators of dietary patterns, dietary changes, or effectiveness of dietary interventions.

No abstract

Quantitative analysis of the bioactives trigonelline (1), N-methylpyridinium (2), caffeine (3), and caffeoylquinic acids (4) in a large set of roasted Arabica (total sample size n = 113) and Robusta coffees (total sample size n = 38) revealed that the concentrations of 1 and 4 significantly correlated with the roasting color (P < 0.001, two tailed), whereas that of 2 significantly correlated inversely with the color (P < 0.001, two tailed). As dark-roasted coffees were rich in N-methylpyridinium whereas light-roasted coffees were rich in trigonelline and caffeoylquinic acids, manufacturing of roast coffees rich in all four bioactives would therefore necessitate blending of two or even more coffees of different roasting colors. Additional experiments on the migration rates during coffee brewing showed that all four bioactives were nearly quantitatively extracted in the brew (>90%) when a water volume/coffee powder ratio of >16 was used.