Sensory and physicochemical properties of enzymatic protein hydrolysates : Influence of raw material, protease, and downstream processing

Abstract

The world aquaculture, fisheries and poultry sectors generate large amounts of residual raw materials, such as heads, backbones, and carcasses. Almost 85% of the residual materials from Norwegian aquaculture and fisheries were utilized in 2019, but over 150 000 metric tons were wasted. This is not compatible with the aim of a circular bioeconomic food production where all the biomass should be utilized. Furthermore, the majority of utilized raw materials were used as low-value feed ingredients. The residual raw materials are excellent food-grade sources of protein and have high potential for further upgrading. However, the materials are not directly applicable for human consumption, but through enzymatic protein hydrolysis, the proteins will be cleaved into more water-soluble peptides and made accessible for use and valorisation.

Enzymatic protein hydrolysates may be utilized within human consumption as protein enrichment of food products and/or as a functional ingredient. However, the sensory properties of protein hydrolysates are considered a major limitation for hydrolysate inclusion in foods. Peptides, free amino acids, minerals, and other water-soluble molecules will follow the hydrolysate phase and contribute to the overall sensory profile. Increased knowledge of the flavour development in protein hydrolysates is imperative when producing products destined for human consumption. Furthermore, the potential amphiphilicity of the peptides generates surface-active properties which is important to understand for their use as functional ingredients in food applications.

The main objective of this study was to assess hydrolysate properties, important for food formulations, of products based on species generating a substantial fraction of residual raw materials. Sensory profiles of the hydrolysates were evaluated using a trained panel and combined with metabolite composition, based on 1H NMR. Both enzyme specificity and new membrane filtration technology were assessed to reduce the sensory properties of the hydrolysates. Furthermore, the effects of hydrolysis parameters on important physicochemical properties, i.e. emulsion activity index (EAI), emulsion stability index (ESI) and critical micelle concentration (CMC) were evaluated.

In Paper I, the use of nuclear magnetic resonance (NMR) spectroscopy as a new tool in sensory assessment of protein hydrolysates were evaluated. Hydrolysates were produced based on muscle tissue from cod, salmon, and chicken with two different enzymes (Bromelain and FoodPro PNL) and hydrolysis times (10 and 50 min). Metabolite composition of the 12 hydrolysates were determined by NMR and the sensory profiles assessed by a trained sensory panel. The results showed that raw material had a major effect on attribute intensity and metabolite variation. The formation of bitter taste was not affected by raw material, indicating a comparable release of bitter peptides independent of substrate. Partial least squares regression on 1H NMR and sensory data provided models for 11 of the 17 evaluates attributes, and significant metaboliteattribute associations were identified based on the obtained models. The study confirmed a potential for prediction of sensory properties based on 1H NMR data.

In Paper II, the effect of hydrolysis parameters on emulsion and surface-active properties were assessed. Direct protein extracts from salmon and cod backbones were compared to hydrolysates based on two different enzymes (Bromelain and FoodPro PNL) with increasing hydrolysis time. EAI, ESI, and CMC were measured for all products. Protein hydrolysis was found to have a negative impact on ESI and CMC, while the ESI generally increased. The direct protein extracts had comparable EAI to that of the commercial emulsifier casein but considerably lower ESI values. The study emphasized the complexity of functional properties in protein hydrolysates and the challenges of achieving high protein yield simultaneously with high surface-activity.

In Paper III, the effect of membrane filtration on sensory properties were evaluated. Heads and backbones from cod and salmon were hydrolysed for 50 min with either Bromelain or FoodPro PNL. The hydrolysates were purified by microfiltration and further refined by nanofiltration and diafiltration. Sensory profiles and metabolite compositions were assessed prior to, and after each nanofiltration step. Metabolite composition were determined and quantified by 1H NMR and sensory profiles were evaluated by a trained sensory panel. The results showed a substantial reduction in metabolite concentration by nanofiltration, with a concomitant reduction in the intensity of several sensory attributes. Bitterness, however, increased as small peptides associated with bitter taste (MW range 0.5−2 kDa) were rejected by the membrane. About 19-24% of the raw material protein were recovered in the nanofiltered product and the main loss was attributed to the removal of bones and solids in the crude hydrolysates. Considerable amounts of protein were also retained in the microfiltration retentate, emphasizing the need for process optimization.

In Paper IV, the sensory, nutritional, and chemical quality properties of protein hydrolysates based on backbones, heads, and viscera from salmon and mackerel were assessed. The hydrolysates were produced using FoodPro PNL and hydrolysed for 50 min. All products were high in essential amino acids and had low biogenic amines content. The raw material fractions caused most of the variation in sensory properties, where viscera products had highest attribute intensities. Mackerel was perceived as the most taste intense of the species, mostly due to high ash content giving strong salty taste of the mackerel hydrolysates. This illustrated the importance of salt removal when producing products for human consumption.