RESULTS

In Table 1 are presented the data obtained in the present work. In addition to the data reported here, data on analysis of oils from individual menhaden reduction plants and data on time of catch of fish from which these oils were produced have been analyzed statistically by a variance technique at the Department of Experimental Statistics of North Carolina State College at Raleigh. At present, these analyses of the physical and chemical characteristics of the fish body oils are being correlated with the many processing variables associated with the different lots of oil and will be reported later.

WEST COAST SARDINE AND TUNA PURSE SEINER

In the Pacific Coast tuna fisheries, the purse seiners or netters are next to tuna clippers in importance. Purse seiners are not as large as the clippers and have a smaller cruising radius and smaller cargo capacity. They were originally designed for sardine and mackerel fishing and usually pursue these species during the fall. Larger purse seiners may fish tuna the year-around and generally catch the same species as the clippers.

The purse seiners use a large net to encircle the schools of fish. The nets generally are about 1,800 feet long and 180 feet deep and generally cost about $30,000 each.

CHEMICAL AND NUTRITIONAL STUDIES ON FISH OILS 1/

By O. S. Privett,* J. R. Chipault,* H. Schlenk, ** and W. O. Lundberg***

ABSTRACT

This paper reports progress on the following four projects: (1) determination of the structure of fish-oil fatty acids and development of analytical techniques for determination of the fatty-acid composition of fish oils, (2) study of the chemistry of odor-producing compounds in fish oils, (3) study of the chemical reactions of fish-oil fatty acids, and (4) study of the nutritional effects of fish oils.

INTRODUCTION

The Hormel Institute, a research branch of the University of Minnesota Graduate School, currently is conducting four projects in the research program established by the U. S. Bureau of Commercial Fisheries with funds made available by the Saltonstall-Kennedy Act.

The objectives and the early stages of the work in three of these projects were described previously (Lundberg 1957). They are concerned with (1) determination of the structure of fish-oil fatty acids and development of analytical techniques for the determination of the fatty-acid composition of fish oils, (2) study of the chemistry of odor-producing compounds in fish oils, and (3) study of chemical reactions of fish-oil fatty acids.

Since our last report was made, a fourth project, which is concerned with the nutritional effects of fish oils, has been started. Progress in each of these four projects since the time of the last publication in this journal is summarized in the following paragraphs.

STRUCTURE AND ANALYSIS OF FISH-OIL FATTY ACIDS

Work on the structure and analysis of fish-oil fatty acids in the past year has involved two main lines of endeavor: (1) the concentration and purification of the principal polyunsaturated fatty acids of tuna and menhaden oils and (2) the complete determination of structure of the main polyunsaturated fatty acids.

CONCENTRATION AND PURIFICATION: The more tedious task, the isolation of pure individual polyunsaturated fatty acids, is nearing completion. Methyl esters of tuna and menhaden fatty acids have been subjected repeatedly to fractional distillation, urea fractionation, and low-temperature crystallization to obtain concentrates of relatively pure samples of individual acids. Separation into various chain lengths has been achieved mainly by distillation through a spinning-band column. Fractionation of each chain length into concentrates of the more important polyunsaturated acids has been achieved by repeated urea fractionations and fractional crystallizations. By these means, concentrates or relatively pure samples of C22 hexaenoic, C20 pentaenoic, and C18 and C16 tetraenoic acids have been prepared. In the course of the separation of these acids, the materials have been characterized by alkali-isomerization and chromatographic analyses.

Although C22 pentaenoic acid appears to be another important polyunsaturated constituent of some fish oils, a suitable concentrate of this acid from either menhaden or tuna oil has not been obtained. Other fish oils, therefore, are being investigated as possible sources of this acid.

STRUCTURE ANALYSIS: Considerable effort has been devoted to developing a technique for determining the locations of double bonds in polyunsaturated fish-oil 1/ This study was supported by funds made available under the Saltonstall-Kennedy Act through a contract with the U. S. Fish and Wildlife Service, Bureau of Commercial Fisheries.

* Associate Professor

**Professor

***Director

Hormel Institute, University of Minnesota, Austin, Minn.

fatty acids. A permanganate-periodate oxidation technique, developed in another laboratory for less unsaturated fatty acids, has been modified satisfactorily for the structure analysis of highly unsaturated fatty acids of fish oils.

A portion of quite pure methyl docosahexaenoate prepared from tuna oil was subjected to such structure analysis. The iodine value of this sample was 441.5 (theoretical 444.7). The purity of the material was established further by chromatographic analysis. By means of the permanganate-periodate oxidation technique, it was demonstrated unequivocally that this is a 4, 7, 10, 13, 16, 19, docosahexaenoate. For the first time, it recently was reported by another laboratory that a docosahexaenoate of this structure is present in pilchard oil. The fact that this acid now has been shown also to be present in tuna oil is of interest from two points of view: first, apparently this is the only docosahexaenoic acid that occurs in appreciable quantities in ordinary fish oils; and second, the acid found in fish oils evidently is identical with that obtained from land animals.

Only a small amount of work was done on the phase of the project concerned with developing analytical techniques for the analysis of fatty acid composition of fish oils. It is planned, however, that in the near future, when additional quantities of the pure polyunsaturated fatty acids become available, the characteristics of the individual acids in alkali isomerization will be determined, and spectral constants for fish-oil analysis will be established. On the basis of work done during the past year, another possible technique for analysis of the fatty acid composition of oils has come to light, involving a combination of distillation, ozonolysis, and paper chromatography of the oxidation productions of ozonolysis. This method of analysis will also be investigated because, if successful, it will yield more complete results than does the alkali-isomerization method.

CHEMISTRY OF THE ODOR-PRODUCING COMPOUNDS IN FISH OILS

Much of the work done on odoriferous materials in fish oil has been concentrated on the isolation, separation, and identification of carbonyl compounds because such compounds previously have been found to be involved in the oxidative deterioration of fish oils and because they also have been reported to be the cause, at least in part, for the "reverted" odors of vegetable oils. The distillates obtained by vacuum-steam deodorization at 100-120 C. (212-248 F.) and also by flushing the oxidized oil with nitrogen at 35-40° C. (95°-104° F.) have been examined. Column and paper chromatography have been used extensively in the separation and identification of the 2,4-dinitrophenylhydrazine derivatives of the carbonyl compounds. These methods have been supplemented by ultraviolet and infra-red spectrometry.

MONOCARBONYL COMPOUNDS: Monocarbonyl compounds that have been shown to be present in the steam-volatile distillates are formaldehyde, acetaldehyde, propionaldehyde, butyraldehyde, valeraldehyde, and hexaldehyde. Unsaturated aldehydes also have been detected.

UNKNOWN 2,4-DINITROPHENYLHYDRAZONE DERIVATIVES: The monocarbonyl derivatives, however, represent a relatively small proportion of the total 2,4-dinitrophenylhydrazones obtained. Most of the hydrazones could not be chromatographed on paper by any of a number of methods that were tried. Most of the nonmigrating compounds gave blue to violet colors when treated with strong alkali, suggesting that they are derivatives of the Greek letter "ALPHA"-dicarbonyl compounds, but some of the nonmigrating fractions remain unchanged on treatment with alkali.

From ultraviolet spectra, it has been determined that the unknown 2,4-dinitrophenylhydrazone derivatives are quite different, in some cases, from those of known mono- and dicarbonyl derivatives. Efforts therefore are being made to

separate and identify such carbonyl derivatives in a different manner. The volatile compounds from fish oils have been neutralized with sodium hydroxide and subjected to steam distillation to remove nonacidic compounds. These, in turn, have been treated with silver oxide to oxidize the aldehydes to acids, which again have been separated from the rest of the compounds by neutralization and steam distillation. Preliminary examination by paper chromatography has indicated the probable presence of acetic, propionic, and suberic or azelaic acids in the acidic fraction. Some mono- and dibasic acids that have not been identified have been detected in the acid fraction formed by oxidation with silver oxide.

These studies show that a variety of volatile, saturated and unsaturated, monocarbonyl and dicarbonyl compounds contribute to the odor of oxidized fish oils. Earlier studies showed that nitrogen compounds also contribute to the odor. Although the steam distillates from the oxidized oils have strong, pungent, disagreeable odors, they do not resemble the odor of the oxidized oil, and attempts to reproduce the odor of fish oil by diluting the distillates with bland mineral oil have not been successful. Thus, the compounds causing the characteristic "fishy" odor are altered or destroyed during steam deodorization.

The volatile compounds removed from oxidized fish oil by bubbling nitrogen through the oil at room temperature have therefore been collected by various devices, including passage of the nitrogen and accompanying volatile compounds through cold traps, 2,4-dinitrophenylhydrazine solutions, and charcoal. It has not been possible to deodorize completely fish oils in this manner, and only small amounts of volatile compounds have been collected. The eluate from the charcoal trap, however, has been found to contain the characteristic fishy odor. The fact that not all of the odor is removed by passage of the gases through 2,4-dinitrophenylhydrazine suggests that other compounds may be as important as are carbonyl compounds in the characteristic odor of fish oil.

A major part of the characteristic fishy odor produced by oxidation, resides, in extremely small amount, in the more volatile fractions. Our future efforts will be directed toward the fractionation, isplation, and identification of this material.

CHEMICAL REACTIONS OF FISH-OIL FATTY ACIDS

The aim in the study of the chemical reactions of fish-oil fatty acids is to prepare new derivatives from the fatty acids occurring in fish oils, mainly by taking advantage of their unsaturation. The work has involved a study of thiourea reactions, halogen reactions, and fractionation by means of sulfur dioxide.

THIOUREA REACTIONS: The incidental finding mentioned in the previous report in this journal--that thiourea reacts with hydroperoxides of autoxidized fatty esters--has been investigated further. In menhaden oil, peroxide values up to 500 can be reduced to about 20 by treatment with a solution of thiourea in methanol at C. (32° F.) or at room temperature. By this simple procedure peroxides are converted into unsaturated hydroxy acids, while unsaturated components that have not been autoxidized remain unchanged.

By using purified hydroperoxides of methyl oleate and linoleate, we found that two peroxidic groups react with one thiourea molecule. In the course of the reaction, the latter is converted into amino-imino-methane-sulfinic acid. Under these mild conditions, between 90 and 95 percent of the peroxidic groups are reduced, but the residual peroxides are not eliminated by repeating the reaction. Apparently, different types of peroxides are present, or rearrangement takes place in the purified hydroperoxides to prevent part of them from reacting with thiourea. The resulting sylfinic acid, and to some extent thiourea itself, is not stable in solvents at 50° to 70° C. (122-158° F.). When the reaction is forced to completeness at elevated temperatures, sulfurized compounds are formed from the secondary products and from the highly unsaturated esters.