smaller when the enzymes were combined with 

 native fish enzymes in subsequent digestions of 

 raw whole hake. The microbial enzymes Rho- 

 zyme P-ll and Bacterial Protease Novo (BPN) 

 gave good yields with whole fish despite rela- 

 tively low activities as measured by the standard 

 test procedure. Pepsin, which had high rel- 

 ative activity in tests with the standard sub- 

 strate, gave relatively poor results with the whole 

 raw hake because the native fish enzymes are 

 inactive at pH 2 which is optimum for pepsin. 



Limited feeding trials with some early batches 

 of enzymatically hydrolysed hake gave generally 

 poor results. Two batches prepared from raw 

 hake with Rhozyme P-ll and Rhozyme 41 yield- 

 ed fairly good protein efficiency ratios (PER) , 

 but feed consumption and growth rates were 

 low. Subsequent trials of three batches pre- 

 pared from presterilized hake gave very poor 

 results. Tryptophan and phenylalanine concen- 

 trations in the precooked samples were low and 

 tyrosine was absent. Raw hake hydrolysates 

 contained higher concentrations of tryptophan 

 and the phenylalanine-tyrosine combination, but 

 histidine contents were lower. 



Feeding trials were suspended while a series 

 of smaller (4- to 5-liter) batches were prepared. 

 A dozen different enzyme preparations, includ- 

 ing several newly acquired ones, were employed 

 over a wide range of processing conditions. 

 Amino acid analyses of the soluble products were 

 collected and studied in order to estimate the 

 nutritive quality of various products obtained. 

 During the course of these runs, the spray dryer 

 contained in the model scale installation was put 

 into operation on a regular basis (Fig. 3). The 

 spray-dried products were definitely lighter in 

 color and a little less hygroscopic, on average, 

 than the vacuum drum dried products. 



Amino acid contents. — Most of the soluble 

 products which were prepared had amino acid 

 analyses which were critically low in either 

 tryptophan or histidine. Tryptophan recovery 

 was very poor at pH 2 (pepsin) but improved 

 with higher pH of hydrolysis and was essentially 

 complete at pH 8. Conversely, histidine recov- 

 ery was good under acid conditions but very 

 poor at pH 8. These correlations between pH 

 and amino acid concentrations are presented in 

 the experimental results section. 



In an attempt to recover adequate concentra- 

 tions of both tryptophan and histidine in a sol- 

 uble product, several 2-stage digestions were 

 carried out. After an initial digestion at acid pH, 

 the soluble phase was removed and the undigest- 

 ed solids were treated with a second enzyme 

 at slightly alkaline pH in an attempt to extract 

 more tryptophan. Supernates from the two 

 stages were blended prior to spray drying. His- 

 tidine recovery was improved but tryptophan re- 

 covery and total yield were poor. The inade- 

 quate degree of solubilization in the second stage 

 was probably due to removal of native fish en- 

 zymes in the first stage supernate. 



Alkaline bacterial proteases. — Alkaline pro- 

 teases of Bacillus subtiUs, not available for the 

 original testing program, yielded soluble hydrol- 

 ysates with a better balance of essential amino 

 acids than had been obtained with other enzymes. 

 A series of soluble products were prepared using 

 the experimental enzymes E56 and P58 from 

 Rohm and Haas and also ABP (Alkaline Bac- 

 terial Protease) and Alcalase' from Enzyme De- 

 velopment Corporation. These enzymes have 

 been developed for use in laundry detergents be- 

 cause of their activity at high pH and temper- 

 ature. Large-scale use in detergents will prob- 

 ably lead to further reductions in the cost of 

 these enzymes, which is already quite competi- 

 tive. 



Digestion of hake at pH 9 and above with the 

 alkaline proteases gave a better recovery of his- 

 tidine than had been obtained at pH 7-8.5 with 

 other enzymes. In initial runs at pH 10 and 60°C 

 yields and tryptophan contents were poor but 

 they were much improved by operation at pH 

 8.5-9.5 and 55°C. Most of the products prepared 

 with alkaline enzymes had fairly good recovery 

 of both histidine and tryptophan, and the prob- 

 able limiting amino acid became isoleucine. 



FPC solubilizations. — Several soluble products 

 were prepared by enzymatic hydrolysis of sol- 

 vent extracted FPC. Pancreatin and ficin as well 

 as the alkaline enzymes were employed. Amino 

 acid patterns of these soluble products were in 

 general superior to those of products prepared 

 from whole hake. Taste and odor problems were 



Manufactured by NOVO Industries, Denmark. 



