Table 4. — Results of autolysis experiments with raw and cooked whole hake and 



flesh-viscera mixtures. 



Flask # 



Contents 



Approximate % 

 solubilization ■ 



Total bacterial count/gram 

 24-hr incubation at; 



35°C 



55°C 



1 Raw whole hake 82.0 <300 <300 



2 Cooked whole hake 63.5 8.8 X 10« 1.6 X 10^ 



3 Raw flesh + cooked viscera 55.1 <300 <300 



4 Cooked flesh -I- raw viscera 80.1 6.7 X 10« 1.07 X 10^ 



raw flesh. Blends of flesh and viscera were in 

 the same proportions as were derived from the 

 beheaded whole fish. The experimental results 

 are presented in Table 4. These results also 

 indicate that autolysis is due to native enzymes, 

 principally in the viscera, rather than to bac- 

 terial growth. Both the cooked and raw mater- 

 ials were exposed to atmospheric contamination 

 during transfer to reaction flasks. The high 

 bacterial counts for the cooked hake and the 

 cooked flesh-raw viscera mixture agree with 

 previous experience that cooked fish, at least in 

 the case of red hake, are more susceptible to 

 bacterial spoilage than are raw fish. 



Yields, PER, compositions for various 

 enzymes. 



Yields. — Average percent yields of soluble sol- 

 ids are listed in Table 5 for groupings of the 

 principal types of proteolytic enzymes employed 

 to hydrolyse red hake. They range from 10% 

 for autolysis without added enzymes to about 

 14% when an alkaline protease or pancreatin 

 was used. Most of the runs that are included 

 lasted 20 hr or more and employed enzyme con- 

 centrations equivalent to a cost of 2 cents per 

 pound of dry solids. Three of the runs with 



Table 5. — Average yields of dry solids from red hake 

 solubilized by proteolytic enzymes. 



Number Yield, dry solubles/wet fish, 

 Enzyme(s) of percent — standard 



runs deviation 



Autolysis 5 10.0 ± 0.4 



Pepsin 1 10.9 



Papain 5 11.3 ± 0.8 



Bacterial Protease Novo 4 11.7 ± 0.10 



Bromelin, ficin 3 12.3 ^ 0.3 



Rhozyme P-11 3 12.5 ± 0.75 



Pancreatin 7 13.8 ± 1.3 



Alkaline proteases .... 5 14.3 — 1.0 



alkaline proteases were at an enzyme cost of 3 

 cents per pound and only one of the runs with 

 alkaline enzymes lasted over 6 hr. 



Amino acid compositions. — Low PER in feed- 

 ing trials with fish protein hydrolysates could be 

 largely attributed to an imbalance of essential 

 amino acids in the soluble solids fraction. The 

 concept of chemical score was studied as a means 

 for estimating product nutritive quality. The 

 chemical score concept is that nutritive value 

 of a protein is directly related to the concentra- 

 tion of the most limiting amino acid. This amino 

 acid is assumed to be that with the lowest con- 

 centration relative to a "perfect" reference pro- 

 tein. Chemical scores were calculated for sol- 

 uble products on the basis of the most commonly 

 used reference protein, whole egg, and also on 

 the basis of the diet experimentally determined 

 by Ramo Rao, Metta, and Johnson (1959) to 

 give maximum growth of rats. Average chem- 

 ical scores calculated for soluble products and 

 based on the Rao diet are listed in Table 6 for 

 groupings of similar enzyme type and range of 

 hydrolysis pH. Among the wide range of en- 

 zymes used, the limiting amino acids are tryp- 

 tophan under acid conditions, histidine at slight- 

 ly alkaline pH, and principally isoleucine with 

 the alkaline proteases at higher pH. 



Tryptophan concentration. — The concentra- 

 tions of tryptophan measured in soluble hydroly- 

 sates are plotted versus the pH of hydrolysis in 

 Figure 7. A positive correlation is apparent and 

 the tryptophan concentration at pH 8 and above 

 is about 3 times greater than it is in soluble 

 products prepared near pH 2. 



Material balances showed that although a lit- 

 tle tryptophan may have been destroyed at the 

 lower pH values, the primary problem was one 



16 



