Fish protein hydrolysates based on Atlantic salmon by-products. Enzyme cost-efficiency and characterization of sensory, surface-active and nutritional properties

Abstract

The world fisheries and fish farming industries generate large amounts of by-products after the primary processing of fish to edible products. In Norway alone, this accounted for almost 900,000 tons in 2014. Based on present industrial practice, most of the by- products are either discarded or used in the manufacture of low-value commodity products such as fish silage, fishmeal and oil. By-product material from the primary filleting process, such as heads and backbones, contain high-quality food grade proteins with a great potential for value creation. The production of water-soluble protein hydrolysates using exogenous proteases may give an increased valorization of the by- products for human consumption and offers a mild and efficient processing approach without prejudicing the nutritional value. Proteases act by cleaving proteins into smaller peptides and free amino acids that are more water-soluble and have altered sensory and surface-active properties compared to the intact protein. A major drawback in the production of commercial fish protein hydrolysates (FPHs) is the formation of bitter and unpalatable tastes due to exposure of hydrophobic amino acids and moieties during the hydrolysis process. Moreover, the cost of enzymes and high processing expenses may be a hindrance in a profitable production of FPHs for human consumption. This has led to a demand for new and improved knowledge of cost-efficiency of enzymes and the process conditions that influences the formation and reduction of bitter taste. Reduction of the bitter taste is of utmost importance in the production of FPHs, but also knowledge of the surface-active and nutritional properties of a hydrolysate may be important for its potential inclusion in food products. The main objective of this study has been to produce FPHs based on Atlantic salmon (Salmo salar) head and backbone products with low bitter taste, good surface-active properties and high nutritional value. The hydrolytic and cost efficiency of five commercial endopeptidases (Alcalase 2.4L, Corolase 7089, Neutrase 0.8L, Promod 671L and Protex 7L) have been evaluated and compared in the hydrolysis of the salmon substrate, based on the pH-STAT method. The sensory properties of the hydrolysates were assessed based on generic descriptive analysis by a trained sensory panel. The hydrolysate surface-active properties were evaluated based on critical micelle concentration (CMC) using 1H NMR. Nutritional properties have been evaluated based on calculations of protein efficiency ratio (PER), amino acid score (AAS), digestible indispensable AAS (DIAAS) and protein digestibility corrected (PDCAAS) using FAO recommendations of indispensable amino acids for small children (six months to three years). In Paper I, substrate specific numbers for nitrogen factor (fN = 5.23 g protein/g nitrogen) and total number of peptide bonds (htot = 9.3 meqv/g protein) were developed to enable more accurate calculations of hydrolysis parameters, such as protein recovery (PR) and degree of hydrolysis (DH). Based on the experimental pH-STAT data, response surface regression models were established to evaluate the combined effects of hydrolysis time and enzyme activity-to-substrate ratio (U/S) on DH and yield of solubilized proteins. The models were combined with activity-specific enzyme cost to estimate the cost efficiency of the individual enzymes, important in upscale and industrial applications. The study demonstrated that all enzymes were equally efficient in hydrolyzing the substrate at low U/S, however, Alcalase 2.4L, Protex 7L and Promod 671L gave higher final DH at high enzyme addition, compared to Corolase 7089 and Neutrase 0.8L. All enzymes were equally efficient in solubilizing the substrate. This may be explained by the enzymes preferentially cleaving peptides already solubilized at high enzyme dose, rather than dissolving new proteins. A linear correlation of DH determined by the pH- STAT and OPA methods was established, which permits the use of the regression models in upscaling of processes where pH-STAT is not applicable. The studies in Paper II confirmed that both molecular weight distribution and enzyme specificity were important for the formation of bitter taste and surface-active properties of the hydrolysates. High intensity of bitter, astringent and pungent attributes was associated with a high DH (≥25%) and peptides with molecular weight <2000 Da. Hydrolysates based on Alcalase 2.4L were significantly more bitter compared to Promod 671L and Protex 7L. Other relevant attributes tested (sweet, salt, umami, acidic, sea and fish) were separated based on the dilution gradient to reach an identical test protein concentration. This suggests that these attributes are related to components inherent to the raw material rather than peptides and amino acids formed during the hydrolysis process. The determined CMCs revealed higher values for all hydrolysates compared to conventionally used food surfactants. The measured CMC was dependent on DH and molecular weight distribution, where low DH gave lowest CMC for all enzyme products. Superior properties (i.e. low bitter taste and low CMC) could be achieved without enzyme hydrolysis and only heat denaturation of the raw material, albeit, at a low yield of solubilized proteins. In Paper III, the effects of exopeptidase activity and activated carbon (AC) adsorption in the debittering of moderately hydrolyzed FPH with DH ≤18% and broad molecular weight distribution were evaluated. Exopeptidase (Flavourzyme 1000L) treatment revealed high release of the hydrophobic amino acids leucine, isoleucine and valine compared to only endopeptidase activity. However, only minor and insignificant reduction of bitter taste was observed, possibly explained by high content of peptides >500 Da and the hydrolytic specificity of the used exopeptidase. Acidic ACs gave largest reduction of bitter taste when added at 1% on protein basis (p = 0.09). The reduction of bitter taste could not be explained by the adsorption of peptides and amino acids from the hydrolysates. Chemical analyses revealed a decrease in salt-free ash and an increase in crude protein on dry matter content in the hydrolysates after treatment with AC. This suggested that ACs adsorb non-protein constituents. The nutritional properties of the raw material and resulting hydrolysates revealed low levels of tryptophan, leucine, isoleucine and valine to meet dietary requirements of children under three years of age. Treatment with exopeptidase and AC did not influence the hydrolysates nutritional properties.