Chemical composition and sensory profiling of coffees treated with asparaginase to decrease acrylamide formation during roasting
Ana CarolinaVieira-Porto,Sara C Cunha,Elaine C Rosa,Juliana DePaula,Adriano G Cruz,Otniel Freitas-Silva,José O Fernandes,Adriana Farah
DOI: https://doi.org/10.1016/j.foodres.2024.114333
Abstract:Acrylamide is an amide formed in the Maillard reaction, with asparagine as the primary amino acid precursor. The intake of large amounts of acrylamide has induced genotoxic and carcinogenic effects in hormone-sensitive tissues of animals. The enzime asparaginase is one of the most effective methods for lowering the formation of acrylamide in foods such as potatoes. However, the reported sensory outcomes for coffee have been unsatisfactory so far. This study aimed to produce coffees with reduced levels of acrylamide by treating them with asparaginase while retaining their original sensory and bioactive profiles. Three raw samples of Coffea arabica, including two specialty coffees, and one of Coffea canephora were treated with 1000, 2000, and 3000 ASNU of the enzyme. Asparagine and bioactive compounds (chlorogenic acids-CGA, caffeine, and trigonelline) were quantified in raw and roasted beans by HPLC and LC-MS, while the determination of acrylamide and volatile organic compounds was performed in roasted beans by CG-MS. Soluble solids, titratable acidity, and pH were also determined. Professional cupping by Q-graders and consumer sensory tests were also conducted. Results were analyzed by ANOVA-Fisher, MFA, PCA and Cluster analyses, with significance levels set at p ≤ 0.05. Steam treatment alone decreased acrylamide content by 18.4%, on average, and 6.1% in medium roasted arabica and canefora coffees. Average reductions of 32.5-56.0% in acrylamide formation were observed in medium roasted arabica beans when 1000-3000 ASNU were applied. In the canefora sample, 59.4-60.7% reductions were observed. However, steam treatment primarily caused 17.1-26.7% reduction of total CGA and lactones in medium roasted arabica samples and 13.9-22.0% in canefora sample, while changes in trigonelline, caffeine, and other evaluated chemical parameters, including the volatile profiles were minimal. Increasing enzyme loads slightly elevated acidity. The only sensory changes observed by Q-graders and or consumers in treated samples were a modest increase in acidity when 3000 ASNU was used in the sample with lower acidity, loss of mild off-notes in control samples, and increased perception of sensory descriptors. The former was selected given the similarity in chemical outcomes among beans treated with 2000 and 3000 ASNU loads.