March-April 1970 Volume 35 : Number 2

Pages 1 0 3 -1 9 4

JOURNAL o f F O O D SCIENCE

103 Bacteria-induced biochemical changes in chicken skin 150 Serologic identification of species of origin of sausage stored at 5 °C —M. A d a m c ic a n d D . S. C la rk m eats—A. B. Karpas, W. L. Myers and D. Segre 106 Physical and chemical properties of epimysial acid-soluble 156 Comparative growth of Clostridium perfringens in carbon collagen from meats of varying tenderness— W. G. dioxide and nitrogen atmospheres-K. G.Parekh and M. Kruggel, R. A. Field and G. J. M iller S o lb e rg I'll Localization of myoglobin in pig muscle—S. M o rita , R . G. Biochemistry of tea fermentation: Conversion of amino Cassens, E. J. Briskey, R. G. Kauffman and L. L. acids to black tea aroma constituents—//. Co and G. W. Kastenschmidt Sanderson 113 Thermal behavior of porcine collagen as related to post- 155 Variations of relative retention times in open-tubular mortem time—R. A. Field, A. M. Pearson, D. E. Koch gas-chromatographic columns: Relevancy to fruit vol­ and R. A. Merkel atiles— T. H. Schultz, T. R. Mon and R. R. Forrey 116 Enzymic oxidation of simple diphenols and flavonoids by Effect in rats of partial replacement of cow’s milk protein H. Bruemmer and B. Roe orange juice extracts—/. by supplementary nitrogen— V. R. Young and A. 120 A method for estimating the functional capillary system in V illa rre a l skeletal muscle—/?. Dairymple, R. G. Cassens and E. J. 175 Low-temperature, long-time heating of bovine muscle. 1. B ris k e y Changes in tenderness, water-binding capacity, pH and 122 Effect of temperature on viscosity of fruit juices and amount of water-soluble components—E. Laakkonen, purees—G. D. Saravacos G. H. Wellington and J. W. Sherbon 126 Effect of processing and storage conditions on the micro- 178 Low-temperature, long-time heating of bovine muscle. 2. flora of Clostridium perfringens-inoculated frank­ Changes in electrophoretic patterns—E. Laakkonen, furters—W. Solberg and B. Elkind J. W. Sherbon and G. H. Wellington 130 Production and properties of Penicillium roqueforti 181 Low-temperature, long-time heating of bovine muscle. 3. lipase—/?. R. Eitenmiller, J. R. Vakil and K. M. Shanani Collagenolytic activity—/'. Laakkonen, J. W. Sherbon 134 Pyroglutamyl dipeptides in mushroom, Agaricus campestris— and G. H. Wellington M. R. Altamura, R. E. Andreotti, M. L. Bazinet and L. L. 184 Poultry product quality. 1. Compositional changes during L o n g Jr. cooking of turkey roasts—/. H . M a c N e il a n d P. S. D im ic k 140 Polyphenols of cashew kernel testa—A. G. Mathew and Poultry product quality. 2. Storage time-temperature H. A. B. Parpia effects on carbonyl composition of cooked turkey and 143 Post-mortem isometric tension changes and shortening iVr /- chicken skin fractions—/. S. D im ic k a n d J. El. M a c N e il turkey muscle strips held at various temperatures—/?'. -A.- - ' - 191 Poultry product quality. 3. Organoleptic evaluation of Jungk and W. W. Marion ‘ cooked chicken and turkey skin fractions as affected by 1 4 6 Benzo(a)pyrene in smoked meat products-K..S. Rhee e n d storage time and temperature-/. H . M a c N e il a n d P. S. L. J. Bratzler 1. . • .D im ick A PUBLICATION OF THE INSTITUTE OF FOOD TECHNOLOGISTS March-April 1970 Volume 35 Number 2

JOURNAL of FOOD SCIENCE

C o n te n ts...... Front Cover Abstracts: In this Issu e ...... iii Research Papers B egin...... 103 Page Charge N o tic e ...... 139

Scientific Editor □ Board of Editors Walter M. Urbain E. F. Binkerd M. A. Joslyn IFT Copy Editor E. J. Briskey Lloyd L. Kempe Lois M. Orr W. L. Brown Z. John Ordal Director of Publications B. F. Buchanan Rose Marie Pangborn John B. K1 is A. E. Denton James W. Pence Production Manager Le Roy Dugan Gerald Reed Jon Day S. A. Goldblith W. M. Roberts Advertising Manager M. Peter Johnstone Willis A. Gortner William H. Stahl Publisher R. H. Hartigan George F. Stewart Calvert L. Willey John A. Holston William L. Sulzbacher

©Copyright 1970 by Institute of Food Technologists. All rights reserved. JOURNAL OF FOOD SCIENCE (formerly Food Research) is published bimonthly by Institute of Food Technologists, Suite 2120, 221 N. LaSalle Street, Chicago, Illinois 60601 USA. Printed in the USA. Second class postage paid at Champaign, 111. and at additional points of entry.

•Address ALL correspondence and manuscripts to the appropriate person at Institute of Food Technologists 221 N. LaSalle St., Chicago, IL 60601 USA Manuscripts: Scientific Editor Subscriptions, Change of address, loss claims: Subscription Department, Journal of Food Science Orders for back issues and 100 or more reprints: Publications Office Advertising Information, contracts and insertion orders: Advertising Manager

Member Subscriptions^ 10 per year. Non-Member Subscriptions—Accepted only on a calendar year basis—no refunds. Rates include postage. Payment must accompany order. Domestic and Pan American Union—$20; all other destinations-$25. Reduced rates for 2- and 3-year subscriptions. Change of address notice, with old address label, is required 4 weeks before issue date. Claims for lost copies are allowed only if received within one month after publication (3 months for foreign subscribers). Single copies and available back issues: $5 each; remittance to accompany order. R e p rin ts: REQUEST SINGLE REPRINTS FROM AUTHORS. Price schedule for 100 or more reprints available upon request from Publications Office.

•NO RESPONSIBILITY is assumed by the Institute of Food Technologists for statements and opin­ ions expressed by the contributors to its publications. •MANUSCRIPTS should conform to the style used in IFT journals. Authors should request from the Publications Office the leaflet “Style Guide for Research Papers.” Journal of Food Science re­ serves the privilege of editing manuscripts to make them conform with the adopted style of the journal or returning them to authors for revision. Editing changes may be reviewed by authors before publication.

Gk AN OFFICIAL PUBLICATION OF THE INSTITUTE OF FOOD TECHNOLOGISTS » m W * * 1

l The Finnigan

M o d e l 3 0 0 0 Gas Chrom atograph/ Peak Identifier.

temperature programmers, capillary or packed column injectors. Q A computer system for mass scan control, data handling and display, ready for delivery! Q 500 amu with guaran­ sy ste m des teed m inim um resolution of better than 525 chromatograph users. Check these important at mass 350—a mass range that usually features. covers 99% of your future GC/MS applica­ Scan speed for mass spectra from 0.1 tions. Q] Easy to understand operation con­ second to 300 seconds— suitable for capillary trols, reliable proven designs and components or packed column work. Q Normal operating to minimize your downtime and maximize your pressure 10 s torr, maximum 10 4 torr—use cap­ analytical throughput. illa ry c o lu m n s w ith o u t separator a n d get m a x i­ All of these and other outstanding fea­ mum sensitivity. Q Standard mass marker— tures offered by the company that understands quickly, easily and accurately identifies the your G C and M S needs. Priced from $ 15,000. mass numbers of interest. Q Linear total ion W rite for a brochure now! current monitor/integrator automatically pro­ vides a chromatogram output. Q Built-in GC, finnigan with splitter or separators, manual or linear FINNIGAN CORPORATION 2631 Hanover Street, Palo Alto, California 94304 (415) 328-3730 Offices in Union, New Jersey; Chicago, Illinois; Hemel Hempstead, U. K.; Munich. Germany; Basel, Switzerland.

11 EFFECT OF TEMPERATURE ON VISCOSITY OF FRUIT JUICES AND ABSTRACTS: PUREES. G. D. SARAVACOS. J. Food Sci. 35, 122-125 (1 9 7 0 )-T h e vis­ IN THIS ISSUE cosities of selected fruit products were measured with a coaxial-cylinder viscometer at temperatures 20-70°C. Clear apple and grape juices were BACTERIA-INDUCED BIOCHEMICAL CHANGES IN CHICKEN SKIN Newtonian fluids and their viscosity decreased considerably at higher tem­ STORED AT 5°C. M. ADAMCIC & D. S. CLARK. J. F o o d Sci. 35, peratures. Cloudy apple and orange juices changed from Newtonian to pseu­ 103-106 (1970)-The effect of psychrotolerant bacteria on the water-hold­ doplastic fluids at higher concentrations and the effect of temperature was ing capacity, as measured by the extract release volume (ERV), pH and smaller than on clear juices. Fruit purees were pseudoplastic and their appar­ protein degradation in chicken skin stored at 5°C was determined using ent viscosities decreased slightly at higher temperatures. The activation ener­ strains of Achromobacter and pigmented and nonpigmented Pseudomonas. gy for flow increased with the juice concentration and decreased with the All organisms caused a rapid decrease in ERV to about 50% of the original presence of suspended particles in the product. The viscometric data can be value during the early log phase of growth before development of off-odor. utilized in the design and operation of efficient processing equipment. Changes in pH and content of extractable nitrogenous materials occurred during bacterial growth, but were not pronounced until after the count was EFFECT OF PROCESSING AND STORAGE CONDITIONS ON THE greater than 108/ cells/g of skin and a faint off-odor was detectable. The MICROFLORA OF Clostridium perfringens-INOCULATED FRANK­ content of extractable materials decreased during the early log phase, corre­ FURTERS. M. SOLBERG & B. ELKIND. J. Food Sci. 35, 126-129 sponding to the period of rapid increase in pH; later, during the late log or (1970)-The growth support potential of frankfurters for Clostridium stationary phases of growth, it rapidly or gradually increased, depending on perfringens was determined. Inoculated emulsions were stuffed into glass the organism. tubes and processed so that they reached an internal temperature of 68-69°C in 30-48 min. The tubes were sealed with agar plugs and rubber PHYSICAL AND CHEMICAL PROPERTIES OF EPIMYSIAL ACID-SOL­ stoppers to simulate vacuum-packaging conditions. Total aerobic and UBLE COLLAGEN FROM MEATS OF VARYING TENDERNESS. W. G. anaerobic bacterial counts in uninoculated raw emulsion were 1 to 3 log KRUGGEL, R. A. FIELD & G. J. MILLER. J. Food Sci. 35, 106-110 cycles higher when petri dishes were incubated at 23°C as opposed to 37°C. (1970)-Components of the molecular structure of epimysial acid-soluble The low-temperature-growing mesophiles and psychrotrophs were destroyed collagen (ASC) from bovine longissimus dorsi muscles of varying tenderness during processing. The 37°C counts were reduced 1 log cycle and the 23°C were studied. Results of sedimentation analyses and viscosity measurements counts, 2 to 3 log cycles. Clostridium perfringens reproduced rapidly in suggested that epimysial ASC from tender meat contains a lesser degree of frankfurters at 37 and 23°C, more slowly at 15 and 12°C, and not at all cross-links. Amino acid analyses indicated that lysine was significantly higher during 2 wk of storage at 10°C or 4 wk at 0 -5 cC. Clostridium perfringens in epimysial ASC from tough meat. Chemical estimation of aldehydic-type made up 2.5-10% of the total anaerobic count at 0 time but became the cross-links in epimysial ASC samples showed that epimysial ASC of tough dominant organism during storage at all temperatures permitting growth. meat contained significantly more aldehyde than that of tender meat. PRODUCTION AND PROPERTIES OF Pénicillium roqueforti LIPASE. R. LOCALIZATION OF MYOGLOBIN IN PIG MUSCLE. S. MORITA, R. G. R. EITENMILLER, J. R. VAKIL & K. M. SHAHANI. J. Food Sci. 35, CASSENS, E. J. BRISKEY, R. G. KAUFFMAN & L. L. KASTENSCHMIDT. 130-133 (1970)-A Pénicillium roqueforti strain produced maximal J. Food Sci. 35, 111-112 (1970)-Myoglobin is localized in specific fibers in amounts of lipase when grown in 0.5% casitone—1% Proflo broth, pH 5.5, at pig muscle. A number of muscle fibers are essentially myoglobin negative in 27°C. Addition of butteroil, com oil or olive oil to the growth medium pig longissimus muscle. The correlation between myoglobin positive fibers inhibited lipase production. Under pH stasis the partially purified lipase of P. and color score was low but the hypothesis was offered that the differential r o q u e fo r ti had an optimum pH of 8.0 and an optimum temperature of 37°C. concentration of myoglobin in various fibers may be related to the PSE Maximum lipolytic activity occurred with 5% butteroil emulsion as the sub­ condition. strate. Manganese chloride and magnesium chloride stimulated the enzyme activity. Calcium, sodium and potassium salts had no appreciable effect on THERMAL BEHAVIOR OF PORCINE COLLAGEN AS RELATED TO lipolysis; silver, mercury and zinc salts were inhibitory. The lipase was ther- POST-MORTEM TIME. R. A. FIELD, A. M. PEARSON, D. E. KOCH & R. molabile, being inactivated completely within 10 min at 50°C. The lipase A. MERKEL. J. Food Sci. 35,113-116 (1970)-Differential thermal analy­ hydrolyzed tributyr.n, tricaprylin, tricaprin, tripropionin and triolein in de­ ses of epimysial connective tissue from normal and low quality porcine creasing order. muscle were compared at 0 and 24 hr post-mortem. In addition, the melting characteristics of intramuscular collagen were determined at 0 hr post-mor­ PYROGLUTAMYL DIPEPTIDES IN MUSHROOM, Agaricus campestris. M. tem. In all tissues studied, collagen from low-quality muscle consistently R. ALTAMURA, R. E. ANDREOTTI, M. L. BASINET & L. LONG JR. gave slightly lower peak melting points than that from normal muscle. Epi­ J. Food Sci. 35, 134-139 (1970)-A fraction was isolated from the edible mysial collagen had a significantly lower peak melting point at 24 hr than at m ushroom, Agaricus campestris. Upon hydrolysis, it liberated much glutamic 0 hr post-mortem. Peak melting temperatures were slightly higher for intra­ acid, small quantities of other amino acids and a hexosamine. The hydrolytic muscular connective tissue than for epimysial tissue. products were separated, quantified and identified by an amino acid analyz­ ENZYMIC OXIDATION OF SIMPLE DIPHENOLS AND FLAVONOIDS er. Besides proline and pyroglutamic acid, evidence was obtained for the BY ORANGE JUICE EXTRACTS. J. H. BRUEMMER & B ROE. J. F o o d presence in the unhydrolyzed fraction of the pyroglutamyl dipeptides of threonine, aspartic acid, valine, leucine, citrulline, phenylalanine, glycine, , Sci. 35,116-119 (1970)-Extracts of orange juice vesicles oxidized p-hydro- quinone (HQ) and o-dihydroxyphenylalanine (DOPA) with Og. Most of the alanine, glutamic acid and proline and the amino acid sugar, N-pyroglutamyl- oxidase activity was associated with a particulate fraction that sedimented at glucosamine. Pyroglutamylcitrulline and N-pyroglutamylglucosamine were 100,000 X g. Sonic disruption of the particulates followed by chromatog­ tentatively identified. The first 6 dipeptides are reported for the first time. A raphy on DEAE-cellulose increased specific activities with both substrates. mechanism is proposed for the enzymic biosynthesis of the pyroglutamyl The partially purified enzyme oxidized numerous naturally occurring dipeptides in mushroom. o-diphenols and o-methoxyphenols. Acidic media below pH 4 and heating above 70°C destroyed most of the oxidase activity. KCN, diethyldithiocar- POLYPHENOLS OF CASHEW KERNEL TESTA. A. G. MATHEW & H. A. bamate and 8-hydroxyquinoline but not EDTA inhibited the oxidase. The B. PARPIA. J. Food Sci. 35, 140-143 (1970)-The major polyphenols of partially purified enzyme reduced benzoquinonc and oxidized the reduced cashew kernel testa have been identified as (+) catechin, (-) epicatechin and form of nicotinamide-dinucleotide (NADH). Ubiquinone, a benzoquinone polymeric proanthoeyanidins, making use of paper chromatography and re­ derivative, but not menadione or vitamin Ki, naphthoquinone derivatives, actions with specific reagents. Two monomeric leucopelargonidins and two could replace benzoquinone in the oxidation of NADH. Ubiquinone, benzo­ monomeric leucocyanidins have been noticed as minor components. The quinone and similar p-quinones may function in the orange as the oxidation- darkening occurring during the processing of nuts has been shown to be reduction couple between NADH-quinone reductase and diphenol oxidase. caused by interaction of polyphenols with iron.

A METHOD FOR ESTIMATING THE FUNCTIONAL CAPILLARY POST-MORTEM ISOMETRIC TENSION CHANGES AND SHORTENING SYSTEM IN SKELETAL MUSCLE. R. DALRYMPLE, R. G. CASSENS & E. IN TURKEY MUSCLE STRIPS HELD AT VARIOUS TEMPERATURES. R. J. BRISKEY. J. Food Sci. 35, 120-122 (1970)-The functional capillary A. JUNGK & W. W. MARION. / F o o d Sci. 35, 143-145 (1970)-A pattern system was investigated by injecting trypan blue dye into 8 market-weight of tension development and gradual relaxation has been demonstrated to porcine animals followed by extracting the dye from various muscles. The occur post-mortem in strips of turkey breast muscle held isometrically. The trapezius contained more dye than the biceps femoris, with the gluteus time to maximum pension development occurs in 3.85 + 0.19 hr and is not medius and longissimus dorsi being equal in dye content but containing less linearly related (P< .05) to temperature. The amount of maximum tension dye than the trapezius and biceps femoris muscles. This rank compared well developed averaged 25 g/cm 2 and was significantly (P < .05) related to with the fiber type distribution, with the trapezius being high in red fiber temperature. Relaxation to about 50% of maximum occurs in 18 hr. The content. The method presents possibilities for study of capillary function amount of shortening that occurs post-mortem is linearly related (P < .01) to but some problems were encountered with the injection procedure. temperature. No “cold shortening” of turkey breast muscle was evident. IN THIS ISSUE

BENZO(A)PYRENE IN SMOKED MEAT PRODUCTS. K. S. RHEE& L. J. of heating there were only minor changes in tenderness. The major decrease BRATZLER. /. F o o d Sci. 35, 146-149 (1970)-A,procedure was developed in shear values occurred between the 4th and 6 th hr, when the meat was for determining benzo(a)pyrene (BaP) in smoked meat products. Artificial warming from 50—60°C. After 6 hr of heating to 60°C there were still cellulose casings markedly reduced the BaP content of bologna smoked uncoagulated water-soluble proteins. The shrinkage of collagen in long-time, during processing. Penetration of BaP into the bologna was minimal low-temperature cooking is considered. m (1.4-1.6 mm). More than half the amount of BaP in smoked bacon was LOW-TEMPERATURE, LONG-TIME HEATING OF BOVINE MUSCLE. 2. found in the fat drippings following cooking. Changes in Electrophoretic Patterns. E. LAAKKONEN, J. W. SHERBON & SEROLOGIC IDENTIFICATION OF SPECIES OF ORIGIN OF SAUSAGE G. H. WELLINGTON./. F o o d Sci. 35, 178-180 (1970)-Polyacrylamide gel MEATS. A. B. KARPAS, W. L. MYERS & D. SEGRE. J. Food Sci. 35, electrophoresis was used to follow changes in the nature of the w ater-soluble 150-155 (1970)-A partially purified immunoglobulin G (IgG) solution pre­ proteins and juices of bovine muscle during low-temperature heating. The pared from the serum of a species to be tested was heated to the specifica­ slowest-moving anodic proteins were coagulated first. The myoglobins and tions for sausages. The resulting supernatant fluid was decanted and the myoalbumins were altered significantly only by holding the meat at 60°C. precipitate washed with saline and used to immunize rabbits. The superna­ The largest changes in tenderness and amount of water-soluble material in tant fluid was used to sensitize tanned sheep red blood cells. The immune the meat occurred at the same temperatures under which the slow-moving serum was rendered monospecific by absorptions with heterologous, heated proteins were denatured. The most heat-sensitive proteins detected were IgG precipitates. A sample of monospecific immune serum was absorbed denatured before there were significant changes in tenderness oi water-solu­ with a washed homogenate of sausage. Aliquots of the monospecific immune ble substance content. serum, both untreated and sausage absorbed, were tested with cells sensitized LOW-TEMPERATURE, LONG-TIME HEATING OF BOVINE MUSCLE. 3. with the homologous heated IgG supernatant fluid. A significant reduction Coilagenolytic Activity. E. LAAKKONEN, J. W. SHERBON & G. H. of titer by sausage absorption indicated that the sausages contained the meat WELLINGTON. /. F o o d Sci. 35, 181-183 (1970)-Naturally occurring coi­ homologous to the immune serum. lagenolytic activity was found in the water-soluble fraction of bovine muscle. COMPARATIVE GROWTH OF Clostridium perfringens IN CARBON DIOX­ General proteolytic activity determined with Azocoll indicated that this IDE AND NITROGEN ATMOSPHERES. K. G. PAREKH & M. SOLBERG. total activity was much greater than the collagenase activity specifically determined according to the method of Wiinsch and Heidrich. The collage­ J. Food Sci. 35, 156-159 (1970)-Relative growth rates of 8 food poisoning nase fraction was concentrated by polyacrylamide gel electrophoresis and strains of Clostridium perfringens in a 100% nitrogen atmosphere and in a 100% carbon dioxide atmosphere at 760-mm pressure were determined. Gen­ the activity of the enzyme was studied under various pH and temperature eration time in nitrogen atmosphere was from 12.9-16.9 min; in carbon conditions. This collagenase could remain active in the meat at cooking dioxide, 12.9-17.2 min. When comparing individual strains in carbon temperatures experienced in long-time, low-temperature cooking, < 60°C. dioxide vs. nitrogen, no significant difference was noticed. Significant differ­ With faster heating and higher internal temperatures, >70-80°C , the col­ ences were noticed in growth rates in both media when strains were com­ lagenase observed in this study is inactivated. pared with one another. POULTRY PRODUCT QUALITY. 1. Compositional Changes During Cook­ ing of Turkey Roasts. J. H. MacNEIL & P. S. DIMICK. J. Food Sci. 35, BIOCHEMISTRY OF TEA FERMENTATION: CONVERSION OF AMINO 184—186 (1970)-White and dark turkey roasts were made with meat taken ACIDS TO BLACK TEA AROMA CONSTITUENTS. H. CO & G. W. from selected samples of the various entries in the Pennsylvania Turkey SANDERSON. J. Food Sci. 35, 160-164 (1970)—It was shown that leucine, Random Sample Meat Test. Additional roasts were made from sample birds isoleucine, valine and phenylalanine were partially converted to aldehydes from the University research flock. Cooking was carried out in aTelkes oven via what appears to be a Strecker degradation during the course of tea fer­ and all roasts were cooked to an internal temperature of 170°F. There were mentation in a tea-leaf homogenate. In contrast, aspartic acid, glutamic acid, no sex differences in cooking losses except when skin was examined sepa­ glutamine, arginine, threonine, serine and theanine gave rise to no detectable rately. Cooking loss differences did exist between Bronze and White for volatile compounds under the same conditions. Identical results were ob­ breast meat but not thigh meat; cooking losses were higher for breast meat tained in a model tea fermentation system containing crude soluble tea than for thigh meat. There was an indication that size of bird was not a enzymes, epigallocatechin gallate and 14C-amino acids. Ascorbic acid inhib­ significant factor in determining percentage cooking losses of roasts made ited formation of aldehydes from amino acids, whereas dehydroascorbic acid from breast and thigh meat. When fat drippings from cooked skin were caused the conversion of amino acids into aldehydes. analyzed for carbonyl content high skin yielding males were characterized by VARIATIONS OF RELATIVE RETENTION TIMES IN OPEN-TUBULAR the high concentration of the 2-enals in relation to the methyl ketones. Low GAS-CHROMATOGRAPHIC COLUMNS: RELEVANCY TO FRUIT VOLA­ skin yield groups consisted mainly cf methyl ketones. TILES. T. H. SCHULTZ, T. R. MON & R. R. FORREY. J. Food Sci. 35, POULTRY PRODUCT QUALITY. 2. Storage Time-Temperature Effects on 165-169 (1970)-Open-tubular gas chromatographic columns made from Carbonyl Composition of Cooked Turkey and Chicken Skin Fractions. P. S. stainless steel tubing, coated with methyl silicone oil, give variable retention DIMICK & J. H. MacNEIL./. F o o d Sci. 35, 186-190 (1970)-The carbonyl times for individual alcohols, dependent on the size of sample. This diffi­ compounds in cooked turkey and chicken skin fractions after storage were culty is largely overcome by using a small proportion of adsorption-reducing isolated as their 2,4-dinitrophenylhydrazones. The monocarbonyl class was material with the methyl silicone. There are cases of reversal of peak order, separated into methyl ketones, 2-enals and 2,4-dienals and measured spectro- caused by differences in column temperature or minor differences in the photometrically. The turkey skin residue fraction contained higher concen- ■, coating. However, with care, relative retention times are nearly constant and trations of carbonyls than did the chicken samples. The oil extract from the ^ are helpful as an aid in identification of compounds. A procedure for precise skin of both groups was similar in carbonyl concentration. Lower storage comparison of the retention times of unknown fruit volatiles with known temperature dramatically lowered the development of carbonyls. Phospho­ compounds is described. Relative retention times of representative com­ lipid phosphorus determinations indicated the residue contained high levels pounds are given for twelve different stationary liquids. of polar lipid; whereas, negligible amounts were in the oil. Thin-layer chro­ EFFECTS IN RATS OF PARTIAL REPLACEMENT OF COW’S MILK matography of the carbonyl classes from the skin residue indicated mainly PROTEIN BY SUPPLEMENTARY NITROGEN. V. R. YOUNG & A. VIL­ C7-C9 2-enals and Cg, C9 2,4-dienals in the unsaturated aldehyde fractions. LARREAL./. F o o d Sci. 35, 170—174 (1970)—Studies were made on growth Changes in fatty acid composition of the residue polar lipids during storage of weanling rats and urinary nitrogen excretion in young adult rats when a suggested linoleic and arachidonic acids as the probable substrates in autox- mixture of nonessential L-amino acids or a mixture of diammonium citrate idative deterioration. and glycine (DAC-Gly) replaced cow’s milk protein to varying degrees. In weanling rats, 20% replacement caused a significant reduction in growth. POULTRY PRODUCT QUALITY. 3. Organoleptic Evaluation of Cooked This reduction was not entirely prevented by supplementation with the sul­ Chicken and Turkey Skin Fractions as Affected by Storage Time and Tem­ fur amino acids and other essential amino acids. Changes in urinary nitrogen perature. J. H. MacNEIL & P. S. DIMICK. /. F o o d Sci. 35, 1 9 1-194 excretion in young adult rats also revealed an alteration in nitrogen utiliza­ (1970)-Cooked turkey and chicken skin residue and separated drippings or tion as a result of protein replacement with DAC-Gly. The reduced dietary oil were stored at various temperatures then presented to a trained taste panel for flavor evaluation. Panel members were able to discriminate be­ nitrogen utilization appeared to be due to decreased utilization of the sulfur tween a control (unstored) and a sample of turkey residue after 3 wk of amino acids and possibly all essential amino acids in the diet. storage at 40°F. They were not able to differentiate between control and LOW-TEMPERATURE, LONG-TIME HEATING OF BOVINE MUSCLE. 1. stored oil even after 7 wk of storage When chicken skin residue ar.d oil were Changes in Tenderness, Water Binding Capacity, pH and Amount of evaluated after storage at 40°F the panel members could detect differences Water-Soluble Components. E. LAAKKONEN, G. H. WELLINGTON & J. W. between the residue samples at 3 wk, but unlike the turkey oil stored at the ■ SHERBON. /. F o o d Sci. 35, 175 — 177 (1970)—Relationships between the same temperature they indicated discriminatory ability after 1 wk of storage. tenderness of very slowly cooked meat and its water-holding capacity, pH When both cooked chicken and turkey skin fractions were presented to the and the amount of water-soluble components were studied. Muscle samples panel at the same time without a reference control (unstored) they were able were heated under fixed temperature programs. Samples were analyzed at to identify differences but could not indicate a clear preference for either 1-hr intervals between the 3rd and 10th hr of heating. During the first 4 hr one. M. A D A M C IC a a n d D. S. C L A R K Division of Biology, National Research Council of Canada Ottawa, Ontario, Canada

BACTERIA-INDUCED BIOCHEMICAL CHANGES IN CHICKEN SKIN STORED AT 5°C

SUMMARY—The effect of psychrotolerant bacteria on the water-holding capacity as measured by lated by the spray-chamber method (Clark, the extract release volume (ERV), pH and protein degradation in chicken skin stored at 5°C was 1963) to give about 200,000 cells/g. Preliminary determined using strains o f Achromobacter and pigmented and nonpigmented Pseudomonas. AH work showed that a 500.000-rad dosage was organisms caused a rapid decrease in ERV to about 50% of the original value during the early log sufficient to destroy all psychrotolerant bacte­ ria in the ground skin. The inoculated skin from phase of growth before development o f off-odor. Changes in pH and content of extractable all plates was pooled, mixed thoroughly to nitrogenous materials occurred during bacterial growth but were not pronounced until after the spread the inoculum uniformly and then dis­ c o u n t was greater than 10s cells/g of skin and a faint off-odor was detectable. The content of tributed in 6-g samples into Petri plates for extractable materials decreased during the early log phase, corresponding to the period of rapio incubation. One sample for each inoculum was increase in pH; later, during the late log or stationary phases o f growth, it rapidly or gradually analyzed immediately after inoculation; the increased, depending on the organism. others were incubated at 5°C in a water-saturat­ ed atmosphere together with controls consisting INTRODUCTION of irradiated but uninoculated skin. Incubated MATERIALS & METHODS samples w'ere removed for analysis at regular AT ABOVE-FREEZING temperatures SKIN WAS OBTAINED from 9-and 24-week- intervals over 21 days. the shelf-life of whole chicken is often old broiler-type chickens (Ottawa Meat Control In a few tests, analyses were also made with limited by the activity of psychrotolerant Strain) raised and processed in the laboratory. samples irradiated but not inoculated or incu­ bated. This was done to determine initial pH strains of Pseudomonas and Achromobac­ The birds were slaughtered by cutting the jugu­ lar vein and carotid arteries, scalded (59°C) for and composition differences between the two ter growing on the surface of the skin 1 min, plucked by hand and rubbed free of ages of skin and to provide a basis for compari­ (Lochhead et d., 1935; Ziegler et al., loose pieces of the epidermal layer of skin and son with inoculated skin. 1954). The activity involves the produc­ parts of feathers with a sterile damp cloth. The All samples were analyzed for bacterial tion of off-odor and off-flavor com­ skin was removed, cooled to about 5°C. count extract release volume, pH and content pounds, presumably resulting from the stretched on wooden blocks and the subcu­ of total extractable nitrogen. The extractable breakdown of free amino acids and pos­ taneous fat removed with a sharp knife. The nitrogen fraction was further analyzed for its sibly skin proteins. Biochemical changes defatted skin was then cut into small pieces content of nonprotein nitrogen and ninhydrin- leading to the production of off-odor and and frozen by immersion in liquid nitrogen. and phenol-reagent-positive materials. To deter­ off-flavor in poultry meat stored asepti- The frozen skin was ground to a coarse powder mine bacterial count, 0.5 g of skin was commi­ cally have been studied (van den Berg et by passing it twice through a hand-operated nuted with 100 ml o f 0.1% peptone in a Waring meat grinder continually cooled with liquid Blendor and the resulting suspension plated on al., 1963, 1964; Khan et al., 1964; Khan, nitrogen. The resulting slurry was placed in a SM agar using the surface method (Clark, 1967). 1965), but little is known about the bio­ large beaker and stirred thoroughly to provide For determination of ERV, 5 g of skin was chemical changes that occur in the skin or a homogeneous mixture of skin from the dif­ ground with 20 ml of glass-distilled water in a muscle resulting from bacterial growth. ferent birds. After removal of the nitrogen by Sorval Omni Mixer for 2 min (50-ml capacity This paper reports results of a study to evaporation the skin was stored in plastic bags cup, 5,000 rpm); the resultant mixture was determine biochemical changes in skin at-100°C until used. For each age. skin from filtered through Whatman No. 1 filter paper and the volume of filtrate collected during 15 min of proteins caused by psychrotolerant enough birds to provide the material required for all subsequent tests was processed at one filtration, as described by Jay (1964a), was strains of Pseudomonas and Achromobac­ time. measured. pH changes in the skin were mea­ te r. This included measurement of Tests were made with pure cultures of sured electrometrically on the ERV extracts. changes in the extractable protein and psychrotolerant bacteria isolated from the skin To determine changes in the content of ex­ nonprotein nitrogenous materials as well of commercially processed poultry (Clark et tractable nitrogenous materials, 5 g of skin was as in the water-holding capacity, as mea­ al., 1969). Included were one strain each of extracted with 35 ml of 0.2 M NaCl at 2-3°C sured by Extract Release Volume (ERV) pigmented Pseudomonas (strain 1), nonpig­ (Jackson et al., 1958) for 24 hr on a shaker. The according to Jay (1964a; Jay, 1964b), m ented Pseudomonas (strain 47) and A chro­ suspension was centrifuged at 10.000 rpm at and pH of skin stored at 5°C. The extent mobacter (strain 89), and a mixed inoculum of 2-3°C for 20 min and the residue washed with 10 ml of extractant and recentrifuged. The and rates of the changes for 9- and 10 strains o f Achromobacter. 5 strains o f non­ pigmented Pseudomonas and 2 strains of pig­ supernatant and washing were combined, made 24-week-old birds were compared. The m ented Pseudomonas. The single cultures were up to 50 ml and filtered (No. 1 Whatman). Pre­ two ages of skin were studied to deter­ judged to be typical of the three types that liminary tests comparing 4 extractants, 0.3 M mine whether possible differences in com­ cause deterioration of refrigerated poultry sodium phosphate buffer (Orekhovich et al., position known to occur in mammalian meat (Ayres et al.. 1950; Thornley ct al., I960; 1948), citrate buffer (Orekhovich, 1948), skin (Carmichael et al.. 1967) would af­ Barnes et al.. 1968). The relative proportion of KCl-borate buffer (Khan, 1962) and 0.2 M NaCl fect biochemical changes induced by each type in the mixed inoculum was the same (Jackson, 1958), showed that extraction of sol­ psychrotolerant bacteria. The work is as that found on commercially processed poul­ uble protein and nonprotein nitrogen materials part of a larger project dealing with bio­ try in Canada (Clark et al., 1969). 24-hr-old was highest with 0.2 M NaCl. The total nitrogen cells grown on Standard Methods agar (SM agar. content of the extract and the nonprotein ni­ chemical and quality changes in poultry Difco) at 20°C and prepared as described pre­ trogen content of the protein-free extract meat during storage. viously (Clark, 1968) were used to make the (treated with trichloracetic acid, 5% w/v final inoculum in all tests. concentration) were determined by a micro- Kjeldahl method. The contents of ninhydrin- aNRC Post-doctorate Fellow 1967-69. Pres­ In most tests, ground freshly thawed skin ent address: University of Zagreb, Zagreb, w'as irradiated (500,000 rads), spread thinly positive materials and phenol-reagent-positive Y ugoslavia. over the bottoms of Petri plates and inocu­ materials in TCA-treated extracts were mea-

Volume 35 119701-JOURNAL OF FOOD SCIENCE-103 VM -JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

the ERV decreased with incubation time to zero (Jay, 1964a). This difference in extent of decrease is no doubt a reflec­ tion of the types and relative amounts of the proteins involved. Bacteria probably increase the water-holding capacity of tis­ sues by enzymatically breaking the inter- or intramolecular linkages of proteins, PSEUDOMONAS / (• — •) thereby increasing the content of free PSEUDOMONAS 47 (a - a) end-groups, which are known to bind A C HRCM OBA C TER 8 9 (c — St water (Ling, 1965). MIXED INOCULUM (* — x) Figure 3 shows that pH changes began to occur in inoculated skin after the log period of growth, but that these changes were not pronounced until after the 0 2 4 6 8 I0 I2 I4 I6 I8 20 count had reached 108 per g and a faint INCUBATION TIME, DAYS but distinct off-odor was noticeable. The Fig. 7 —Growth of psychrotolerant bacteria in Fig. 2—Effect of bacteria on extract release pH then rose rapidly for all organisms, comminuted skin of 9-week-old chickens. volume of skin of 9-week-old chickens. particularly for Pseudomonas 47 and Achromobacter 89. This period of in­ crease corresponded to a rapid increase in the extent of off-odor, presumably sured by the methods of Rosen (1957) and that of adults (Behrendt et al., 1955; caused mostly by ammonia and other de­ Folin et al. (1927), respectively, and were ex­ Behrendt et al., 1964). pressed as tyrosine equivalents. composition products. The relatively The effect of bird age on the chemical slower rate of pH increase for the mixed composition of the skin, however, had no RESULTS & DISCUSSION inoculum compared to that for P se u d o ­ effect on the rate of bacterial growth m onas 47 and Achromobacter 89 w o u ld RESULTS OF analysis of uninoculated (Clark, 1968) nor on the pattern of bacte- appear to indicate a dominant effect of samples showed that skin from 9-week- ria-induced biochemical changes measured. the pigmented pseudomonads in the m ix­ old birds contained more extractable ni­ Therefore, only the results for skin of ture even though their proportion initial­ trogenous material (29—39% more, de­ 9-week-old birds (broiler age) are pre­ ly was low (about 12%). The small pH pending on the specific material mea­ sented in detail. Rates of growth of the change obtained before off-odor develop­ sured) and had lower ERV and pH values individual test organisms and of the ment (0.1—0.2 of a pH unit), similar to (8—10% lower) than skin from 25-week- mixed culture in the ground skin were that reported for skin of poultry contami­ old birds (Table 1). The differences in ex- about the same (Fig. 1) and led to off- nated naturally during processing tractability were probably related to the odor after about 4 days of incubation, (Fromm et al., 1965), indicates that a pH effect of age on the solubility of neutral- coinciding with a total count of about 108 test on skin is of doubtful value for as­ salt-soluble collagenous proteins. Evi­ cells per gram of skin. sessment of quality of whole poultry. dence obtained for bovine skin has shown Results of tests with inoculated skin Figures 4 and 5 show that the bacteria that the solubility of these proteins in a showed that all inocula caused a rapid de­ caused changes in the nitrogenous constit­ sodium chloride solution decreases with crease in the extract release volume dur­ uents of the skin but, as with the pH test, age (Carmichael et al., 1967). The higher ing the early log phase of growth (Fig. 2). the changes were not large until after 4 ERV value for older skin also indicates The decrease reached a maximum of days of incubation and development of that solubility of skin collagen decreased about 50% of the original value after 4 off-odcr. All three types of bacteria w ith b ird age. W ith decreased s o lu b ility days, corresponding to a bacterial count caused a decrease in the contents of total the amount of inter- and intramolecular o f about 108 per g. The test gave consist­ extractable nitrogen, nonprotein nitrogen cross-linking among the a chains of the ent and similar results for the relatively and phenol-reagent-positive- and ninhy- collagen molecule increases (Banga, 1966; small sample used [5 g compared to 25 g drin-positive materials during the log Carmichael et al., 1967); as a result, sites used by Jay (1964a)] and for all types of capable of binding water are reduced. The bacteria studied. Therefore, the results effect of age on pH is not understood but show that the ERV test indicates the bac­ the change was similar in extent to that terial quality of whole poultry as it has reported for human skin; the skin of chil­ been shown to do for ground beef (Jay, dren is about 0.5 of a pH unit lower than 1964b). It is noteworthy that with beef

Table 1 —Effect of age on pH, ERV and content of extractable nitrogenous material of uninoc­ ulated chicken skin. Skin from 9-week-old Skin from 25-week-old birds (14 samples) birds (10 samples)

Material Avg Range Avg Range Total extractable nitrogen (mg/g) 2.8 2 .5 -3 .0 1.8 1 .7 -2 .0 Nonprotein nitrogen (mg/g) 0.85 0 .7 8 -0 .9 5 .57 0.35 -0 .7 5 Ninhydrin-positive material (mg tyrosine/g) 6.8 5 .9 -8 .2 4.8 3 .6 -6 .5 Phenol-reagent-positive material Og tyrosine/g) 172 160-185 109 7 0 -1 8 0 KRV (ml/5 g) 10.8 9 .4 -1 1 .6 11.4 9.9 12.0 Fig. 3—effect of bacteria on pH of skin of pH of ERV extract 6.6 6 .5 -6 .7 7.2 7 .1 -7 .5 9 -w e e k -o ld ch icken s. BIOCHEMICAL CHANGES IN CHICKEN SKIN - 105 phase, particularly between the 4th and of the observed changes in composition duce off-odor more quickly on the skin 6th day of incubation, corresponding to could have resulted from the extractable than on other poultry tissues (Lochhead the period of rapid increase in pH. After metabolic end-products and autolyzed et al., 1935; Ziegler et al., 1954), a bac­ this period, values for Pseudomonas 47 bacterial cells present. However, prelimi­ teriological quality test involving the skin and Achromobacter 89 remained station­ nary tests in which 1-g quantities of cells of whole birds would in most circum­ ary or increased gradually to the original were extracted under the conditions used stances apply to the whole carcass. Fur­ values; those for Pseudomonas 1 increased in the tests with skin, showed that less ther work is required to determine wheth­ rapidly, giving maximum values after than 2% of the nonprotein nitrogen and er similar results for the ERV test are ob­ 12-16 days. These results suggest that the total extractable nitrogen in the tests tained for leg and breast muscle exposed organisms utilize the low-molecular- with inoculated skin could be accredited in cut-up poultry. weight nitrogenous compounds during to the bacterial cells. the period of rapid growth and subse­ REFERENCES quently replace such compounds through CONCLUSIONS Ayres, J.C., Ogilvy, W.S. and Stewart, G.F. breakdown of proteins, rapidly or gradu­ 1950. Post mortem changes in stored meat. THE RESULTS showed that psychrotol- ally, depending on the organism. Further Microorganisms associated with develop­ erant spoilage bacteria cause a large in­ ment of slime on eviscerated cut-up poultry. studies are presently under way to deter­ Food Technol. 4, 199. crease in the water-holding capacity cf mine the compounds affected. Banga, I. 1966. Structure and function of elas- chicken skin before development of off- tin and collagen. Akad. Kiadc, Budapest, The changes in pH in inoculated skin odor. This change appears to be inde­ Hungary. during incubation appeared not to affect Barnes, E.M. and Impey, C.S. 1968. Psychro- pendent of the type of organism and of philic spoilage bacteria of poultry. J. Appl. markedly the results of the various tests. their relative ability to degrade proteins. Bacteriol. 31, 97. Figures 2 and 3 show that most of the Behrendt, H. and Green, M. 1955. The relation­ Bacteria-induced changes in the pH and changes in ERV values occurred before a ship of skin pH pattern to sexual maturation content of nitrogenous materials in skin in boys. Am. J. Diseases Children 90, 164. measurable change in pH. Preliminary Behrendt, H., Green, M. and Bernard, C. 1964. begin as the organisms start to grow, but work in which skin samples were ex­ Relation of skin pH pattern to sexual matu­ are not large until the cell count is above ration in girls. Am. J. Diseases Children tracted (24 hr at 5°C) with buffered 0.2 M 108 per g and off-odor is produced. 108, 37. NaCl at various pH values between 6.5 Carmichael, D.J. and Lawrie, R.A. 1967. Bo­ Therefore, ERV appears to be the most vine collagen. I. Changes in collagen solubil­ and 8.0 showed that over this range the promising test among those studied for ity with animal age. J. Food Technol. 2, total extractable nitrogen content varied 299. assessment of the microbial quality of less than ± 4% from the average. Other Clark, D.S. 1963. Uniform inoculation of nutri­ whole poultry (skin intact). Since ent surfaces. Biotech. Bioeng. 5, 123. workers have also shown with other Clark, D.S. 1967. Comparison of pour and sur­ psychrotolerant bacteria grow and pro­ tissues that pH between the values 6.5 face plate methods for determination of and 8.0 has little effect on the solubility of protein and non protein nitrogenous materials in NaCl solutions (Dyer et al., 1950; Yasui et al., 1964). Since cells of the test organisms could not be separated from the skin samples before extraction and centrifugation, part

Fig. 4—Effect of bacteria on the content of ex­ Fig. 5—Bacteria-induced changes in the content of phenol-reagent­ tractable nitrogenous materials in skin of 9- positive- and ninhvdrin-positive materials in skin of 9-week-old chick­ week-old chickens. ens. 106 -JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

bacterial counts. Can. J. Microbiol. 13, Food Technol. 18, 1637. of chilled poultry. J. Appl. Bacteriol. 23, 1409. Khan, A.W. 1962. Extraction and fractionation 487. Clark, D.S. 1968. Growth of psychrotolerant of proteins in fresh chicken muscle. J. Food van den Berg, L., Khan, A.W. and Lentz, C.P. pseudomonads and achromobacter on chick­ ScL 27, 430. 1963. Biochemical and quality changes in en skin. Poultry ScL 47, 1575-1578. Khan, A.W. and van den Berg, L. 1964. Changes chicken meat during storage at above-freez­ Clark, D.S. and Lentz, C.P. 1969. Microbiolog­ in chicken muscle protein during aseptic ing temperatures. Food Technol. 17, 91. ical studies in poultry processing plants in storage at above-free zing temperatures. J. van den Berg, L., Lentz, C.P. and Khan, A.W. Canada. Can. Inst. Food Tech. J. 2, 33. Food Sci. 29, 49. 1964. Changes in quality and water-holding Dyer, W.J., French, H.V. and Snow, J.M. 1950. Khan, A.W. 1965. Objective evaluation of qual­ and ion-binding properties of chicken meat Proteins in fish muscle. I. Extraction of pro­ ity in poultry meat. Nature 208, 204. during above-freezing storage under aseptic tein fractions in fresh fish. J. Fish. Res. Ling, G.N. 1965. Physiology and anatomy of conditions. Food Technol. 18, 729. Board Can. 7, 585. cell membrane: the physical state of water Yasui, T., Sakanishi, M. and Hashimoto, Y. Folin, O. and Ciocalteu, V. 1927. On tyrosine in the living cell. Federation Proc. 24 1964. Effect of inorganic polyphosphates and tryptophane in proteins. J. Biol. Chem. (Suppl.), 5103. on the solubility and extractability of Myo­ 73, 627. Lochhead, A.G. and Landerkin, G.B. 1935. sin B. Agr. Food Chem. 12, 392. Fromm, D. and Monroe, R.J. 1965. Some ob­ Bacteriological studies of dressed poultry. I. Ziegler, F., Spencer, J.V. and Stadleman, W.J. servations concerning the hydrogen ion vari­ Preliminary investigations of bacterial action 1954. A rapid method for determining spoil­ ability of the chicken carcass surface during at chill tem peratures. ScL Agr. 15, 765. age in fresh poultry meat. Poultry Sci. 33, storage. Poultry ScL 44, 325. Orekhovich, V.N., Tustanovshii, A.A., Orekhov- 1253. Jackson, D.S., Leach, A.A. and Jacobs, S. 1958. ich, K.D. and Plotnihova, N.E. 1948. About Ms. received 3/10/69; revised 7/2/69; accepted The amino acid composition of the collagen procollagen of the skin. Biochem. Lenin­ 7/11/69. fractions of rabbit skin. Biochim. Biophys. grad 13, 55. Acta 27, 418. Rosen, H. 1957. A modified ninhydrin colori­ N.R.C.C. No. 11383. Jay, J.M. 1964a. Release of aqueous extracts by metric analysis for amino acid. Arch. Bio­ The authors acknowledge the cooperation beef homogenates and factors affecting re­ chem. Biophys. 67, 10. of the Commercial Products Branch of Atomic lease volume. Food Technol. 18, 1633. Thomley, M.J., Ingram, M. and Barnes, E.M. Energy of Canada Ltd. in the preparation of Jay, J.M. 1964b. Beef microbial quality deter­ 1960. The effects of antibiotics and irradia­ sample material. They also thank Mr. N.U. mined by extract release volume (ERV). tion on Pseudomonas-Achromobacter flora Cholette for technical assistance.

W IL L IA M G. K R U G G E L , a RA'S A. FIELD and GLENN J. MILLER Division of Biochemistry and Division o f Animal Science University of Wyoming, Laramie, Wyoming 82070

PHYSICAL AND CHEMICAL PROPERTIES OF EPIMYSIAL ACID-SOLUBLE COLLAGEN FROM MEATS OF VARYING TENDERNESS

SUMMARY—Components o f the molecular structure o f epimysial acid-soluble collagen (ASC) of the structural factors. The structural from meats o f varying tenderness were studied by several methods. The ASC was studied since it factors may be said to include the molec­ contains an appreciable amount of intramolecular cross-links but is still soluble. Though the ular structure of the connective tissue pro­ amount o f total collagen in epimysial tissue was found to have no correlation with meat tenderness, tein, collagen, and how it influences the the molecular studies indicated some correlation o f the type o f epimysial ASC with meat tender­ degree of tenderness of a muscle. ness. Sucrose density-gradient ultracentrifugation analyses o f denatured epimysial ASC suggested A few researchers have deduced from that this type of collagen contains a lesser degree o f cross-links when obtained from tender meat their investigations that tenderness or samples. Viscosity measurements were found to be correlated to tenderness o f meat in a manner toughness is concerned with molecular similar to the results obtained by ultracentrifugation, in that the intrinsic viscosity o f epimysial cross-links in the collagen of connective ASC from tender meat was louver than that from tough meat, indicating differences among the tissue. Goll et al. (1962) studied the rate relative sizes o f the collagen molecules from the different samples. Partial amino acid analyses o f of connective tissue solubilization the epimysial ASC samples via gas chromatography showed no differences in the levels o f aspartic, brought about by the action of collagen- serine or hydroxyproline, amino acids which may be involved in the ester type o f cross-links. ase in the biceps femoris muscle obtained Results o f chemical estimation o f the ester type o f cross-links indicated that the proportions of from aged cow, cow, steer and veal age esters in the various ASC samples were similar. The amount o f lysine was found to be significantly groups. The study indicated the occur­ h igh er IP < 0.051 in epimysial ASC from tough meat compared to tender meat, suggesting an rence of more frequent or stronger cross­ increased potential o f the aldehydic-type o f cross-link, which is formed via an aldol condensation links within and among the tropocollagen o f two aldehydes derived biosynthetically from lysine. This was strengthened by the results ob­ molecules of collagen from the older age tained in the chemical estimation o f the aldehydic-type o f cross-link, in that the epimysial ASC of group, since a decreasing rate of solubili­ tough meat contained significantly more aldehyde than did that o f tender meat (P<0.05l. zation occurred as age increased. Herring et al. (1967) found results similar to those of Goll et al. (1962), collagen solu­ INTRODUCTION tors may be divided into three broad bility decreasing significantly with ad­ classifications: ante-mortem, post-mor­ vancing maturity of beef even though the AS THE INQUIRY into the causes of tem and structural factors. Ante-mortem collagen content was not different. tenderness or toughness of meat contin­ factors include physiological factors, such Cover et al. (1962) suggested that a ues, investigators have concluded that this as age and inheritance, and feeding and tightening of the protein structure during is a complex problem and cannot be sim­ management practices; whereas, post­ heating (denaturation) occurs as new sta­ ply resolved by the study of a single mortem factors include temperature and ble intermolecular cross-links are formed moiety, since many factors contribute to length of storage time after slaughter, between the peptide chains of collagen. variations in beef tenderness. These fac- methods of trimming and cutting and An alternate way that this may occur is cooking methods. The structural factors for the tightening effect to occur intra- aPresent address: The Worcester Founda­ of tenderness are now being investigated molecularly along the peptide chains of tion for Experimental Biology, Shrewsburg, more thoroughly and this study is centered collagen, thereby making the chains more Massachusetts 01545. upon some of the molecular implications resistant to cleavage. MEAT TENDERNESS AND ASC PROPERTIES-107

Goll et al. (1964a) suggested that 71°C. W am er-Bratzler shear-force values were bonds present in the solution. Thus, the larger stronger or more extensive cross-links in obtained on 1-in. cores. 7 animals o f varying the difference in absorbance between the 2 the collagen from older animals occur, W am er-Bratzler shear-force values were divided wavelengths, the greater the protein concentra­ and that the number and strength of into 3 groups: Group I, the most tender ani­ tion. The accuracy o f the m ethod was checked these cross-links play an important role in mals, consisting o f 1 steer and 2 bulls; G roup II, by the Low ry m ethod o f protein determ ination anim als o f interm ediate tenderness, 1 bull and 1 (Lowry et al., 1951). The sw 2 o (sedimenta­ meat tenderness (Goll et al., 1964b). Hill steer; Group III, the toughest animals, 1 cow tion) values of the peaks obtained by the a (1966) found strong presumptive evi­ and 1 steer. The epim ysial layer of uncooked 215-225 mu method were determined by the dence to confirm the earlier work of Goll steaks from these ^lim als served as sources o f tables o f McEwen (1967).

et al. (1964a), and stated that increased collagen for m olecular studies. Determ ination o f viscosities toughness of bovine meat may be ex­ After the fat and adhering meat were dis­ Solutions for viscometry measurements sected from the epimysial layer, the samples plained by an increase in cross-links. Car­ were clarified by centrifugation at 30,000 rpm were sonicated fo r 10 sec and loosened particles michael et al. (1967) found an increased (105,000 x g) for 90 m in at 20°C in the ultra- gently scraped away. Each sample was then formation of dimer or higher aggregates centrifuge. Viscosities of epimysial ASC sliced in the cold in to strips approxim ately 1 in. of the a-chains of collagen with an in­ (0.015-0.530 mg protein /m l 0.2 N acetic acid) b y 1 / 4 in . crease in animal age. It was indicated that were measured at 20.0 ± 0.1 °C in an Ostwald there was a marked increase in the pro­ Total collagen determ ination viscometer w ith a flow time for water of 112 portion of components in bovine collagen The extraction of total collagen from the sec. The intrinsic viscosity o f each o f the sam­ having intramolecular cross-links between epim ysial samples was carried out according to ples was determ ined from a plot o f reduced vis­ the 5th and 9th month of gestation. In the procedure of Neuman et al. (1950), except cosity vs. concentration (M athews et al., 1954). more mature animals, there was some in­ that 10-mg samples in 12 ml demineralized Preparation o f samples for gas-liquid chroma­ water were autoclaved for 12 hr, instead of tographic analyses dication of an enhanced degree of inter- placing the samples into 4 ml of water and 1-mg samples o f epim ysial ASC were hydro­ molecular cross-links. Therefore, the mo­ autoclaving for 2, 3-hr periods. The extracted lyzed in sealed tubes under nitrogen in 5 m l 6 N lecular structure of the collagen present collagen was hydrolyzed in 6 N HC1 for 7 hr at HC1 for 24 hr at 110°C. A fter hydrolysis the may be of some importance. The purpose 104°C. After neutralization w ith 6 N sodium hydrolysates were concentrated to about 2 m l hydroxide, the samples were filtered into 50-m l of this investigation was to study compo­ on a rotatory evaporator, about 3 m l o f dem in­ volum etric flasks and made up to volum e w ith nents of the molecular structure of colla­ eralized water added, and again concentrated. demineralized water. A solution o f a standard gen to discover if there were variations in A fter a second addition o f dem ineralized water, hydroxyproline sample (0.3 m g/m l) was treated such components among beef muscles of the samples were taken to dryness. s i m i l a r l y . varying tenderness. M ethyl and butyl esters o f the hydrolysates H ydroxyproline in each o f the samples was were prepared according to the procedure of determined according to the m odified proce­ EXPERIMENTAL Gehrke et al. (1965), except that the m ethyl dure o f Dahl et al. (1963), using hydrogen per­ Chemicals and solutions esters were made by heating in capped tubes at oxide for the oxidation o f hydroxyproline to 105°C for 2 hr and the butyl esters by heating The am ino acid standards used in gas-liquid pyrrole. A fter color development the o.d. was in capped tubes at 150°C for 4 hr. Acylation o f chrom atographic analyses were purchased from measured at 550 rriM- Values obtained for hy­ the samples w ith trifluoroacetic anhydride at Mann Research Laboratories. Trifluoroacetic droxyproline content were used in determ ining 150°C to form the «-butyl N -trifluoroacetyl de­ anhydride was an Eastman organic chemical the total am ount of collagen in each sample rivatives was carried out according to the purchased from D istillation Products Industries. according to the calculations o f Neuman et al. sealed-tube acylation m ethod o f Stalling et al. Anhydrous hydrogen chloride gas was obtained ( 1 9 5 0 ) . (1966). Dichloromethane was added to each from Matheson Scientific, Inc. and hydrogen E xtraction o f acid-soluble collagen (ASC) sample so as to have a concentration of 2 peroxide (30%) from Matheson, Coleman and m g/m l. A standard m ixture o f am ino acids sim i­ Bell. The p-dimethylaminobenzaldehyde was The epim ysial strips were extracted 3 times lar in am ount to those present in collagen was purchased from Brinkmann Instruments Inc., w ith 0.45 M sodium chloride for 24 hr at 2°C subjected to the same conditions o f hydrolysis and Tris (hydroxym ethyl) aminomethane from (Gross, 1958) and the ASC extracted from the and derivative form ation to decrease prepara­ Fisher Scientific Com pany. A ll other chemicals rem aining residue and purified according to the tion error. A fter preparation of the n-butyl were “ Baker Analyzed” reagents and were used procedure o f Rubin et al. (1965). A fter precipi­ N -trifluoroacetyl esters, the derivatives were w ithout further purification, except the m etha­ tation and lyophilization, the epimysial ASC subjected im m ediately to gas-liquid chrom ato­ nol, which was dried according to the proce­ samples were stored at -20°C . graphic analyses. dure o f Vogel (1951). The epimysial ASC was chosen for the fol­ The 2.5 N hydrogen chloride in dry metha­ lowing analyses, since it contains appreciable Gas-liquid chrom atography o f am ino acid deriv­ nol was prepared by bubbling anhydrous hydro­ am ounts o f cross-links b u t is still soluble. a t iv e s

gen chloride gas into the m ethanol w ith vigor­ Am ino acid analyses o f the n-butyl N -triflu- Density-gradient ultracentrifugation o f epim ys­ ous stirring. The end point o f addition was de­ oroacetyl esters were carried out by gas-liquid term ined by titration o f 1-ml aliquots from the i a l A S C chrom atography em ploying the follow ing oper­ methanolic solution against standard sodium Sedimentation analyses of ASC solutions ating conditions: Colum n temperature, initial hydroxide to a phenolphthalein end point. The were perform ed by centrifugation in 12-ml line­ 100°C, final 260°C. Program rate, 4°/m in after 2 N hydrogen chloride in butanol was prepared ar gradients o f 4-17% (w /v) sucrose buffered 8 min initial delay. Detector temperature, in a sim ilar manner. w ith 0.15 M Tris buffer, pH 8.0. The gradients 230°C. Injector temperature, 140°C. Carrier were allowed to stand after preparation for 24 flow , N 2, 30 m l/m in. A ir flow to detectors, 175 Instrum ents hr at 2°C before being used for ultracentrifuga­ m l/m in. Hydrogen flow to detectors, 15—20 Gas-liquid chrom atographic analyses were tion. Solutions o f epim ysial ASC were prepared m l/m in. Columns: 10 ft by 1/8 in. i.d. Pyrex carried out in an Aerograph 200 Dual Colum n by dissolving 1 mg o f collagen in 0.2 m l o f 0.15 columns packed with 3% OV-17 on 100/120 Gas Chromatograph w ith flame ionization de­ M Tris buffer, pH 8.0. After denaturation at mesh Gas-Chrom Q. tectors. U ltracentrifugations were perform ed in 45°C and centrifugation at 2,600 rpm , the solu­ an International M odel B-60 ultracentrifuge us­ U nit detector response (peak height X reten­ tions were carefully applied to the top o f the ing the SB-283 rotor. A Hitachi Model 139 tion tim e) for each am ino acid was determined sucrose gradients and centrifuged at 9°C for 18 Spectrophotom eter was used for all colorim et­ from the standard am ino acid preparation. Re­ hr at 30,000 rpm (105,000 x g) in the ultracen­ ric analyses, and the Beckman DU Spectro­ producibility o f this method of determ ination trifuge. Each o f the ultracentrifuged epim ysial photom eter was used for the * 215—225 m ethod o f the am ount o f each am ino acid was found to ASC samples was fractionated into 5 drop o f protein determ ination. be 95%. The standard response o f each amino fractions collected in test tubes. A fte r addition acid was then used to determ ine the quantity o f Selection and preparation o f tissue o f 0.5 m l dem ineralized water to each tube, the each am ino acid in the epim ysial ASC samples. Steaks obtained from the ninth rib o f the optical densities were read at 215 and 225 mu. Ester content o f epim ysial ASC samples longissimus dorsi muscle 2 days after slaughter A plot o f tube num ber vs. the difference in o.d. at 215 and 225 m/a was made (Chaykin, 1966). Hydroxamates were prepared according to of the animals were trim m ed o f the epimysial The difference gives a measure o f the peptide the procedure o f Blum enfeld et al. (1962) and layer and cooked to an internal tem perature of 108-JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

Table 1 —Sum m ary o f percent total collagen o f epim ysial tissue from Table 2—Percentage o f the components ob­ the longissimus dorsI muscle o f beef animals o f different age, sex and tained upon sucrose density-gradient ultracen­ shear-force values. trifugations of ASC samples from steaks of varying tenderness.1 Total Warner-Bratzler shear collagen 3S 9S I5S 19S Animal1 Age (months) Sex force values (lb)2 (%)3 Animal2 (“) (P) (7) A 13 Steer 11.89 72 69 * B 13 Bull 12.69 48 A (4) 18 13 85 * C 13 Bull 12.97 57 B (d ) 7 8 67 * E 13 Bull 16.47 45 C (d) 15 18 53 25 F 13 Steer 18.36 55 E(6) 8 14 G 13 Steer 22.75 56 F (4) 50 22 17 11 H 48 Cow 29.86 56 G (4) 59 25 9 7 H (9) 43 24 16 17 1 Animals A, B, C were considered tender; E, F intermediate; G, H tough. 2 Increasing values indicate a decrease in tenderness. * Shoulder at 9S. 3 Represents average of 4 trials. 1 Values are percent of total peak area. 2 A, B, C tender samples; F., F intermediate; G, H tough.

assayed by the colorim etric method of Lip- epimysial tissues were found to be about mann et al. (1945), except that o.d. was deter­ equal in value, except for 1 animal that lagen molecule. The peak at 3S has been m ined at 505 mg. had an extremely high content of colla­ attributed to the a-component, that at 9S Aldehyde content of epimysial ASC samples gen, and 2 that had low values for total to the j3-component and that at 15S to the 7 -component. The peak appearing at Phenylhydrazones were prepared according collagen (Table 1). No positive relation­ to the procedure o f Levene (1962). The absorp­ ship was found between percent total col­ 19S is apparently attributable to a more tion of each phenylhydrazone solution was lagen in the epimysial tissue samples and highly cross-linked collagen molecule. measured at 390 m y. Warner-Bratzler shear-force values. Cover The percentage of each peak from the A ll o f the preceding analyses were duplicat­ et al. (1956); Parrish et al. (1962); ultracentrifuged epimysial ASC samples is ed; where appropriate, values obtained were Ritchey et al. (1963) and Kim et al. shown in Table 2. From this table it can subjected to analysis o f variance according to (1967) in working with biceps femoris be seen that the a-component makes up the m ethods o f Steel et al. (1960). and longissimus dorsi muscles found no the majority of the total in all cases. RESULTS positive relationship between amount of However, it is very evident that the ASC PERCENTAGES of total collagen in the total collagen and tenderness. of epimysial tissue from tender meat con­ It has long been felt that connective tains more a-component and less ^-com­ tissue plays a very distinctive role in ten­ ponent than does that from intermediate derness of beef, and since collagen is a or tough meat. Actually, the 13-compo­ structural element of the tissue this pro­ nent in that from tender meat was so tein has been studied with relation to ten­ small that the peak area could not be cal­ derness. Only recently has it been sug­ culated. but this component comprised gested that the toughness-tenderness fac­ approximately 25% of the total in that tor may be influenced by the number and from intermediate and tough meat. The strength of cross-links in the collagen percentage of the y-component and 19S molecule. The degree of cross-linking in component was very variable and there epimysial collagen was investigated in this were no apparent differences among sam­ research by the physical measurements of ples. These results suggest that ASC of sedimentation and viscosity and by the epimysial tissue from tough meat con­ chemical approximation of cross-links. tains a higher proportion of cross-links Figure 1 is an illustration of the su­ than does that from tender meat. crose density-gradient ultracentrifugation Viscosities of epimysial ASC solutions patterns obtained with denatured epimys­ were found to be correlated to the ten­ ial ASC from meat with varying degrees derness or toughness of meat in a manner of tenderness. Upon calculation of sedi­ similar to the sedimentation coefficients mentation coefficients, values of 3, 9, 15 (Table 3). As meat tenderness decreased, and 19S were obtained. All of the tender the intrinsic viscosities of the epimysial samples studied exhibited a shoulder at ASC increased. These differences in in­ 9S, whereas, intermediate and tough sam­ trinsic viscosity may be reflecting differ­ ples studied showed a definite peak at 9S. ences among the relative sizes of the col­ Even though sucrose density-gradient lagen molecules from the various samples. ultracentrifugation of denatured collagens The partial amino acid compositions has not been reported in the literature, of the various epimysial ASC samples are ultracentrifugations at about 60,000 rpm shown in Table 4. Due to the small per­ of ASC solutions in sodium formate buff­ centage of several amino acids present in

TUBE NUMBER er have been performed (Piez et al., 1961 ; collagen (histidine, hydroxylysine, iso­ Piez et al., 1963; Lewis et al., 1964a; Fig. 1—Representative sucrose density-gradient leucine, methionine and tyrosine), the ulracentrifugation patterns obtained from dena­ 1964b; Cooper et al., 1965). On the basis amount of each could not be adequately tured epimysial ASC from meat o f varying ten­ of the data from these papers the peaks determined. In addition, threonine and derness. Broken line, tender sample (C); solid obtained in the sucrose density-gradient leucine could not be resolved under the line, intermediate sample (F); dotted line, ultracentrifugations have been assigned to conditions employed. Aspartic acid, an tough sample (H). the various chain-components of the col­ acidic amino acid, has been implicated in MEAT TENDERNESS AND ASC PROPERTIES - 1 0 9

Table 3—Intrinsic viscosities o f epimysial Table 4—Partial am ino acid com position o f the various epim ysial ASC samples.1 ASC samples as a function o f decreasing tender­ ness o f meat. S a m p le s 2 A m in o S ta n d a rd ASC a c id m ix t u r e A ( d ) B ( d ) C ( d ) E ( d ) F ( d ) G ( d ) H ( 9 )

A n im a l1 [ t7 ] , d l/ g S er 0 .8 1.4 2 .1 2 .1 2 .1 0 .9 2 .0 1.5 A s p 0 .9 1 .6 1.5 1 .9 2 .3 2 0 1 .6 3 .1 A ( d ) 8 .5 H p r 2 .2 3 .7 2 .9 4 . 6 3 .7 4 .1 4 . 3 8 .5 B (< 3 ) 1 2 .0 G lu 1.5 1.8 1.9 1.7 1.9 2 .0 2 .3 E(d) 2 1 .0 2 .4 L y s 0 .5 0 .4 0 .5 0 .6 1 .3 F(d) 2 1 .5 1.1 1 .7 1.1 A r g 0 .8 1.1 1 .9 2 .0 3.1 G(d) 2 4 .5 2 .5 2 .0 2 .0 G ly 3 .2 3 .5 3 .3 3 .4 3 .4 3 .4 3 .4 3 . 6 H ( 9 ) 2 3 .0 P ro 3.1 4 .8 3 .7 6.1 5 .1 5 .0 5 .4 9.1 1 A, B tender samples; E, F interm ediate; A la 3 .3 4 .8 4 .2 4 .6 5 .8 4 . 0 5 .1 4 .7 G, H tough. V a l 0 .7 1 .2 1.1 0 .9 2 .0 0 .9 1.1 1 .0 P h e 0 .3 0 .5 0 .4 0 .5 1.1 0 .7 0 .4 1 .0

1 Values are 10 9moles/mg. 2 A, B, C tender samples; E, F interm ediate; G, H tough. cross-links via an esterlike link (Blumen- feld et al., 1962), with the hydroxyamino acids, hydroxyproline, serine and threo­ nine (Gallop et al., 1959). As can be seen in Table 4, the amounts of serine, aspartic cross-links, and the current investigation Table 5—Esterlike links and aldehyde con­ acid, hydroxyproline and glutamic acid appears to lend credence to this point of tent of epimysial ASC from meat o f varying were not significantly different among view. The sedimentation studies clearly tenderness. the various ASC samples (P < 0.05). Re­ indicated that the a-component is higher E s t e r lik e sults of chemical estimation of esterlike A ld e h y d e and the /3-component lower in epimysial A n im a l1 l i n k 2 c o n t e n t 3 cross-links in the epimysial ASC samples ASC from tender meat. This can be ex­ via ester content are shown in Table 5. A ( d ) 7 .8 1 .9 plained by a higher proportion of macro­ B ( d ) 7 .5 2 .0 There were no apparent differences in the molecules with lower amounts of intra­ ester content among the various samples. C ( d ) 7 .5 1 .9 molecular cross-links. Results of the vis­ E ( d ) 7 .6 1 .6 Lysine is thought to be involved in cosity studies point to the same conclu­ F ( d ) 7 .7 2 .9 cross-links via an aldehyde derivative. sion, since intrinsic viscosities were lower G ( d ) 7 .4 2 .4 This involvement appears to be through in epimysial ASC of tender meat, indicat­ H ( 9 ) 7 .5 3 .7 aldol condensation of 2 aldehydes derived ing smaller molecules. Both of these 1 Anim als A, B, C are considered tender; E, F biosynthetically from 2 lysine residues in measurements together suggest a differ­ interm ediate; G, H tough. collagen (Rojkind et al., 1968). Results ence in the amount of intramolecular 2 Values represent nmoles hydroxam ate/100 obtained by gas chromatographic analyses cross-links rather than any differences in mg collagen. Values represent equivalents aldehyde/m ole indicate a significantly higher level (P < intermolecular cross-links. However, since c o lla g e n . 0.05) of this amino acid in epimysial ASC the degree of intermolecular cross-links from tough meat samples (Table 4), and was not assayed, it cannot be stated this suggests a potential for more of the whether there is an association of tender­ aldehydic type of cross-link. Further­ ness with such cross-links. ASC of tough meat has more of the alde­ more, results of the chemical estimation The partial amino acid analyses of the hyde type of cross-link. This possibility of aldehydes present in epimysial ASC epimysial ASC samples do not indicate a was strengthened by the fact that such samples tended to confirm this suggestion definite trend in regard to levels of the epimysial ASC contained significantly (Table 5). A significant difference was various amino acids and the tenderness of more apparent aldehyde. found in the amount of aldehyde present meat. The levels of 3 of the amino acids in the tender, intermediate and tough reported as being involved in cross-links It must be re-emphasized that the col­ samples (P < 0.05). This indicates that via esterlike links, serine, aspartic acid lagen studied was epimysial collagen, epimysial ASC of tough meat contained and hydroxyproline, were not significant­ which is not that collagen interspersed in more aldehyde than did that from tender ly different among the various ASC sam­ the muscular areas. Certainly, this colla­ meat, which implies that as the meat be­ ples. Even though glutamic acid has not gen should have more direct influence comes tougher, the epimysial ASC has a been reported as playing a role in cross­ upon tenderness than the epimysial colla­ gen. However, all collagens in a tissue higher level of aldehyde taking part in the links via an esterlike link similar to that might be related, in that if a high degree formation of intramolecular cross-links. of aspartic acid, there is a possibility that it could be involved. However, the levels of cross-linking occurred in one then it DISCUSSION of glutamic acid in the ASC samples were might also occur in the others. If this is true, then the amounts of cross-linking in AS EXPECTED, the amount of total col­ not significantly different. Results of chemical estimation of ester in the epi­ the epimysial and intramuscular collagen lagen in epimysial tissue has no correla­ may be related. tion with meat tenderness. However, re­ mysial ASC support the amino acid data, since the proportions of esters in the ASC sults obtained by molecular studies have REFERENCES indicated that the type of epimysial ASC were similar. Blum enfeld, O.O. and Gallop, P.M. 1962. The present is of more importance than the The level of lysine was significantly participation of aspartyl residues in the quantity. Evidence was obtained indicat­ higher in epimysial ASC of tough meat hydroxylam ine- or hydrazine-sensitive samples compared to tender meat sam­ bonds of collagen. Biochemistry 1: ing that an increase in the number of 9 4 7 - 9 5 9 . cross-links in epimysial ASC accompanies ples. Since this amino acid is suggested as Carmichael, D.J. and Lawrie, R.A. 1967. Bo­ a decrease in meat tenderness. Hill (1966) being a precursor to aldehydes which can vine collagen. II. Electrophoresis of collagen fractions. J. Food Technol. 2 : 3 1 3 - 3 2 4 . has suggested that an increase in tough­ link peptide chains together via aldol con­ Chaykin, S. 1966. Methods for protein deter­ ness may be explained by an increase in densation, it is possible that epimysial m ination. In “ Biochemistry Laboratory 110-JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

Techniques.” p. 20. John W iley and Sons, tral salt extracts of norm al guinea pig con­ Parrish, F.C., Bailey, M.E. and Naumann, H.D. Inc., New York. nective tissue. J. Exptl. Med. 107: 247-263. 1962. H ydroxy proline as a measure of beef Cooper, D.R. and Davidson, R.J. 1965. The ef­ Herring, H.K., Cassens, R.G. and Briskey, E.J. tenderness. Food Technol. 16(2): 68-71. fect of ultraviolet irradiation on soluble col­ 1967. Factors affecting collagen solubility Piez, K.A., Eigner, E.A. and Lewis, M.S. 1963. lagen. Biochem . J. 97: 139-147. in bovine muscle. J. Food Sci. 32: 534-538. The chrom atographic separation and am ino Cover, S., Ritchey, S.J. and Hostetler, R.L. H ill, F. 1966. The solubility of intram uscular acid com position of the subunits of several 1962. Tenderness of beef. III. The muscle- collagen in m eat animals of various ages. J. collagens. Biochem istry 2: 58-66. fiber com ponent of tenderness. J. Food Sci. Food Sci. 31: 161-166. Piez, K.A ., Lewis, M.S., M artin, G.R. and Grass, 27: 483-488. Kim , C.W., Ho, G.P. and Ritchey, S.J. 1967. J. 1961. Subunits of the collagen m olecule. Cover, S. and Sm ith, W.H. 1956. The effect of Collagen content and subjective scores for Biochim . Biophys. Acta 53: 596-598. two methods of cooking on palatability tenderness o f connective tissue in animals of Ritchey, S.J., Cover, S. and Hostetler, R.L. scores, shear force values, and collagen con­ different ages. J. Food Sci. 32: 586-588. 1963. Collagen content and its relation to tent of two cuts of beef. Food Res. 21*. Levene, C.I. 1962. Studies on the mode of ac­ tenderness of connective tissue in tw o beef 3 1 2 - 3 2 1 . tion of lathyrogenic compounds. J. Exptl. muscles. Food Technol. 17: 194-197. Dahl, O. and Perrson, K.A. 1963. H ydroxypro- Med. 116: 119-130. R ojkind, M., Rhi, L. and Aquirre, M. 1968. Bio­ line, m ethodological studies of analysis. Lewis, M.S. and Piez, K.A. 1964a. The charac­ synthesis o f the intram olecular cross-links in Acta Chem. Scand. 17: 2499-2503. terization of collagen from the skin of the rat skin collagen. J. Biol. Chem. 243: G allop, P.M., Seifter, S. and M eilm an, E. 1959. dogfish shark, Squalus acanthias. J . B io l. 2 2 6 6 - 2 2 7 2 . Occurence of esterlike linkages in collagen. Chem. 239: 3336-3340. Rubin, A.L., Drake, M.P., Davison, P.F., Pfahl, Nature 183: 1659-1661. Lewis, M.S. and Piez, K.A. 1964b. Sedimenta- D., Speakman, P.T. and Schm itt, F.O. 1965. Gehrke, C.W., Lam kin, W .M., Stalling, D.L. and tion-eqilibrium studies of the molecular Effects of pepsin treatm ent on the interac­ Shahrokhi, F. 1965. Q uantitative gas chro­ weight of single and double chains from rat- tion properties of tropocollagen m acron de­ matography of amino acids. Biochem. Bio- skin collagen. Biochem istry 3: 1126-1131. cides. Biochem istry. 4: 181-190. phys. Res. Com m un. 19: 328-334. Lipm ann, F. and Tuttle, L.C. 1945. A specific Stalling, D.L. and Gehrke, C.W. 1966. Q uantita­ Goll, D.E., Hoekstra, W.G. and Bray, R.W. m icrom ethod for the determ ination of acyl tive analysis of am ino acids by gas chrom a­ 1962. Changes in the physicochem ical prop­ phosphates. J. Biol. Chem. 159! 21-28. tography: Acylation of arginine. Biochem. erties of bovine connective tissue with Low ry, O.H., Rosebrough, N.J., Farr, A.L. and Biophys. Res. Com m un. 22: 329-335. chronological age. J. Animal Sci. 21: Randall, R.J. 1951. Protein measurement Steel, R.G.D. and Torrie, J.H. 1960. Analysis of 9 8 1 - 9 8 2 . with the Folin phenol reagent. J. Biol. variance. I. The one-way classification. In Goll, D.E., Hoekstra, W.G. and Bray, R.W. Chem. 193: 265-275. “ Principles and Procedures of Statistics,” 1964a. Age-associated changes in bovine Mathews, M.B., Kulonen, E. and Dorfman, A. pp. 99-131. McGraw-Hill Book Company, muscle connective tissue. II. Exposure to in­ 1954. Studies in procollagen. II. Viscosity Inc., New York. creasing temperature. J. Food Sci. 29: and molecular weight. Arch. Biochem. Bio- Vogel, A.I. 1951. “Practical Organic Chem­ 6 1 5 - 6 2 1 . phys. 52t 247-260. istry,” 2nd ed., pp. 167-168. Longmans, Goll, D.E., Bray, R.W. and Hoekstra, W.G. M cEwen, C.R. 1967. Tables for estim ating sedi­ Green and Co., New York. 1964b. Age-associated changes in bovine m entation through linear concentration Ms. received 7/7/69; revised 8/28/69: accepted muscle connective tissue. III. Rate of solu­ gradients of sucrose solution. Anal. Bio­ 8 / 2 9 / 6 9 . ______bilization at 100°C. J. Food Sci. 29: chem. 20! 114-149. 6 2 2 - 6 2 8 . Neuman, R.E. and Logan, M .A. 1950. The de­ Journal Paper 389 of the W yoming Agri­ cultural Experim ent Station. Gross, J. 1958. Studies on the form ation of col­ term ination of collagen and elastin in tis­ lagen. I. Properties and fractionation of neu­ sues. J. Biol. Chem. 186: 549-556. S . MORITA , a R. G. CASSENS, E. J. BRISKEY, R. G. KAUFFM AN and L. L. KASTENSCHMIDT

Muscle Biology Laboratory, University o f Wisconsin, Madison 53706

A RESEARCH NOTE LOCALIZATION OF MYOGLOBIN IN PIG MUSCLE

PALE, soft, exudative (PSE) muscle is a pigs and 15 market weight Chester White problem to the food industry (Briskey, pigs was used. The histochemical proce­ 1964). Lawrie (1950) has shown that the dure is given in detail elsewhere (Morita concentration of myoglobin in muscle et al.,1969). The longissimus muscle was varies with age, species, anatomical loca­ scored subjectively for color (Forrest et tion and activity. It might be assumed al., 1963) 24 hr post-mortem. Score 1 is that the pale component of PSE muscle is considered extremely PSE and 5 is con­ closely associated with decreased con­ sidered very dark, firm and dry. Scores of centrations of myoglobin. However, the 2 and below are PSE. situation is not clear and some disagree­ Figure 1 illustrates typical histochem­ ment exists in the literature. Briskey et al. ical results. Three staining intensities are (1959) reported that the decreased in­ distinguishable, and are indicated on the tensity of visual color (comparing normal figure as positive, intermediate and neg­ and PSE muscle post-mortem) was not ative. We made photographic enlarge­ accompanied by significantly lower myo­ ments of the sections and evaluated per­ globin concentrations. Henry et al. cent of total fiber area contributed by (1958), and Scopes et al. (1963) attribute each of the three staining intensities wi:h the paleness largely to diminished myo­ a Zeiss particle size analyzer. The quanti­ globin whereas Wismer-Pedersen (1959) tative results for myoglobin positive fi­ and Goldspink et al. (1964) attribute it to bers (does not include intermediate) are denaturation of myoglobin. shown in Figure 2. The results are ex­ We used a histochemical technique for pressed as a correlation for group 1 and as myoglobin localization (James, 1968; a correlation for groups 1 and 2 com­ Morita et al., 1969) and compared the bined. Groups 1(11 Poland China and 12 cellular localization of myoglobin in mus­ Chester White) and group 2 (3 Poland Fig. 2—Scatter p lo t and correlation coefficients cle that ultimately became PSE to that in China and 3 Chester White) represent two for color score versus m yoglobin positive fiber a re a . muscle which retained a normal color and separate experiments conducted approx­ gross morphology. The longissimus mus­ imately 6 weeks apart. cle from 14 market weight Poland China Results from group 1 gave a significant correlation (r = 0.52, P < .05) but when the second group was added the correla­ 1969) has implicated the metabolism of aOn leave from Nippon Veterinary and Zoo- intermediate fibers in the development of technical College, 1-7-1, Kyonan-C ho, M usashino tion (r = 0.23, P < .05) was greatly di­ C ity, Tokyo, Japan. minished. Previous work (Cooper et al., the PSE condition, but we do not know the significance of the “intermediate” reaction for myoglobin. There was slight­ ly more area for intermediate fibers in PSE muscle (score 2 and below) than in normal muscle. If the fibers intermediate in myoglobin reaction are intermediate with histochemical techniques for phos- phorylase and ATPase, then they may behave as white fibers in terms of me­ tabolism post-mortem (Cooper et al., 1969) . When only group 1 represented in Fig­ ure 2 is considered, then the muscle with low color score had a small area of myo­ globin positive fibers. Our histochemical results show a distinct difference between fibers in respect to myoglobin concentra­ tion. Pig muscle has areas of myoglobin positive fibers surrounded by large areas of fibers negative for myoglobin. This is contrasted to beef muscle (Morita et al., 1970) which contains a larger fiber area of myoglobin positive fibers; the red fi­ Fig. 1—Section o f porcine longissimus muscle reacted for m yoglobin. Three levels o f reaction prod­ uct intensity are indicated as positive (P), negative (N) and interm ediate (II. The unlabelled arrow bers are distributed in a more random points out a positive reaction for hemoglobin given by a trapped erythrocyte. 3Op thick section. fashion about the bundle even though the 8 4 x . red fiber content of beef muscle is not

Volume 35 (19701-JOURNAL OF FOOD SCIENCE- 1 1 1 112-JOURNAL OF FOOD SCIENCE- Volume 35 (1970) much greater than pig muscle. The rapid regimen on certain chemical and physical M orita, S., Cassens, R.G. and Briskey, E.J. ham muscle characteristics. J. Anim al Sci. 1969. Localization of m yoglobin in striated post-mortem glycolysis associated with 1 8 : 1 7 3 . muscle of the dom estic pig : benzidine and the development of PSE muscle is not Cooper, C.C. Cassens, R.G. and Briskey, E.J. NADK 2—TR reactions. Stain Technol. 44: common in beef muscle. Therefore, the 1969. Capillary distribution and fiber char­ 2 8 3 . acteristics in skeletal m uscle o f stress-suscep­ M orita, S., Cassens, R.G. and Briskey, E.J. cellular concentration of myoglobin and tible animals. J. Food Sci. 34: 299. 1970. Histochemical localization of m yo­ the distribution of myoglobin positive Forrest, J.C., Gundlach, R.F. and Briskey, E.J. globin in skeletal muscle of rabbit, pig and fibers within the muscle bundle may be 1963. A prelim inary survey of variations in ox. J. Histochem . and Cytochem . (in press). certain pork ham muscle characteristics. Scopes, R.K. and Lawrie, R.A. 1963. Post­ an important property of pig muscle Proc. X V Res. Conf., Am . Meat Inst. Found. m ortem lability of skeletal muscle proteins. Goldspink, G. and McLoughlin, J.V. 1964. Nature 197: 1202. which allows the great range in rate of Studies on pig muscle. 3. The effect of tem ­ Wismer-Pedersen, J. 1959. Quality of pork in post-mortem glycolysis. perature on the solubility of the sarcoplas­ relation to rate of pH change post-m ortem . m ic proteins in relation to colour changes in Food Res. 24: 711. post-rigor muscle. Irish J. Agr. Res. 3: 9. Ms. received 11/26/69; accepted 1/28/70. REFERENCES Henry, M., Romani, J.D. and Joubert, L. 1958. Further studies on pork muscle. Rev. Path­ M anuscript No. 568 of the W isconsin Agri­ Briskey, E.J. 1964. Etiological Status and Asso­ ol. Gen. Phys. C lin. No. 696: 355. cultural Experiment Station, Department of ciated Studies of Pale, Soft, Exudative James, N.T. 1968. Histochemical demonstra­ Meat and Anim al Science. Porcine Musculature. Adv. Food Res. 13: tion of m yoglobin in skeletal muscle fibers Supported in part by Public Health Service 8 9 . and muscle spindles. Nature 219: 1174. Research Grant FD—00107—11, and in part by Briskey, E.J., Bray, R.W ., Hoekstra, W .G., Phil­ Lawrie, R.A. 1950. Some observations on fac­ grants from the Arm our Food Research Labora­ lips, P.H. and Grum m er, R.H. 1959. The ef­ tors affecting m yoglobin concentrations in tories and the Am erican Meat Institute Founda­ fect of exhaustive exercise and high sucrose m uscle. J. Agr. Sci. 40: 356. t io n . R. A. FIELD,3 A. M. PEARSON, D. E. KOCH and R. A. MERKEL

Department of Food Science

Michigan State University, East Lansing, Michigan 48823

THERMAL BEHAVIOR OF PORCINE COLLAGEN AS RELATED TO POST-MORTEM TIME

SUM M ARY—DTA (differential thermal analysis) thermograms o f epimysial connective tissue from nective tissue, w hich collected on the screen as normal and low quality porcine muscle were compared at 0 and 24 hr post-mortem. In addition, the 0-hr muscle samples were being shaken the m elting characteristics o f intram uscular collagen were determined at 0 hr post-mortem. In all through a 40-mesh screen at -20°C , was re­ moved and saved for subsequent DTA deter­ tissues studied, collagen from low quality muscle consistently gave slightly lower peak m elting m inations. Intramuscular connective tissue points than that from norm al muscle. Epimysial collagen had a significantly (P < .0 5 ) lower peak could not be separated by this m ethod for mus­ m elting point at 24 hr than at 0 hr post-m ortem (64.24 vs. 65.77°C). Using epim ysial collagen, a cle samples removed from the longissimus of significantly (P < .0 1 ) greater proportion o f the total m elting range occurred at lower temperatures the carcasses at 24 hr post-m ortem . However, at 24 hr post-m ortem as compared to 0 hr (41.1 vs. 30.9%). The therm al behavior o f intram uscular McClain (1969) has only recently reported the collagen at 0 hr post-mortem was sim ilar to that o f epimysial collagen at 0 hr, but peak m elting isolation o f intram uscular connective tissues, temperatures were slightly higher for intram uscular connective tissue. using techniques sim ilar to those reported here­ in .

INTRODUCTION EXPERIMENTAL Although the low quality carcasses were se­ THE SHRINKAGE temperature (Ts) of INTRAM USCULAR connective tissue and the lected as being pale, soft and exudative (PSE) collagen is defined by Rigby (1964) as epimysium from the longjssimus muscle of 16 on the basis o f appearance, quality scores and the temperature at which an uncon­ Yorkshire pigs approxim ately 5 m onths old fur­ transm ission values, the carcasses did not show strained sample in 0.9% NaCI solution un­ nished the collagen used in this study. W ithin 5 high temperatures and low pH within a few dergoes an abrupt dimensional change. m in after exsanguination, and before dehairing m inutes after death (Table 1) as is characteristic Another thermal transition occurs at the or evisceration: (1) the longissimus samples o f PSE muscle (Briskey, 1964). Thus, these car­ from the lum bar region o f the carcass were re­ casses were called low quality instead o f PSE. equilibrium melting temperature, Tm, the m oved; (2) the epim ysial tissues were scraped From among the 16 pigs on which the above temperature of equilibrium between free o f fat and m uscle; and (3) both muscle and procedures were follow ed, 10 were selected for crystalline and amorphous phases in a cleaned epim ysium were frozen in liquid nitro­ this study. Q uality param eters used in selecting partially melted sample. Shrinkage is due gen. These tissues are hereafter referred to as 5 low er quality and 5 norm al carcasses included to the melting of the crystalline phase 0-hr post-m ortem tissues. Adjacent longissimus transm ission values at 24 hr post-m ortem (Hart, when both the intramolecular forces samples anterior to the lum bar region were re­ 1962) and subjective scores for firmness and within the triple-stranded molecule and moved and frozen in liquid nitrogen 45 min color at 24 hr post-mortem (Forrest et al., the crystallization energy are overcome post-exsanguination. Muscle and cleaned epi­ 1 9 6 3 ) . mysial tissues were removed from the 7th to D TA is a technique for studying the therm al (Harrington et al., 1961). For mammalian 8th rib region of the longissimus at approxi­ behavior o f m aterials as they undergo physical collagen in 0.9% saline, Ts is 59-60°C m ately 24 hr post-m ortem , and frozen in liquid and chem ical changes during heating or cooling. and Tm is about 51°C (Rigby, 1964). n i t r o g e n . The DTA method has been reported to have In severe cases of pale, soft exudative The frozen muscle samples were removed excellent sensitivity and reproducibility when (PSE) pork, muscles lose their connective from liquid nitrogen and stored up to 72 hr at used to study hydrotherm al shrinkage of colla­ tissue attachments. Briskey (1963) hy­ -20°C , then powdered as described by Bcr- gen (McClain et al., 1968). DTA thermograms pothesized that this condition was due to chert et al. (1965). The fine intram uscular con- from the DuPont 900 Differential Thermal a high temperature (36-40°C) and a low pH (5.0-5.2) within the first few hours after death. He stated that thermal

shrinkage or hydrolysis, or both may Table 1 —Comparison o f means and standard deviations for characteristics o f longissimus mus­ occur in connective tissue. Therefore, cles from 5 norm al and 5 low quality pigs. studying shrinkage characteristics of con­ nective tissue may aid in better under­ N o r m a l Low quality M e a n SD M e a n SD standing the PSE syndrome, as well as in Muscle variable helping define the role of connective tis­ N u m b e r sue in tenderness. The present study uti­ lized differential thermal analysis (DTA) Subjective score1 to eludicate changes in thermal shrinkage C o lo r 4 . 3 0 ' 0 . 4 5 2 . 6 0 3 0 . 4 2 of collagen at 0 and 24 hr post-mortem. F ir m n e s s 4 . 1 0 ' 0 . 8 2 2 . 6 0 3 0 . 8 2

Differences in thermal shrinkage charac­ Transm ission value, 24 hr 1 6 . 3 0 ' 1 4 . 2 8 6 9 . 6 0 3 1 4 . 8 9 teristics of collagen from normal and low­ er quality muscle at the two different Postm ortem pH post-mortem times were compared. In O h r 6 . 1 6 0 . 0 6 6 . 1 6 0 . 0 6 addition, the melting characteristics of 4 5 m in 6 . 0 5 0 . 1 5 5 . 9 6 0 . 1 9 2 4 h r 5 . 2 0 0 . 0 5 5 . 1 9 0 . 0 7 intramuscular collagen at 0 hr post-mor­ tem were compared to those of epimysial Postm ortem tem perature, °C collagen at 0 hr post-mortem. 0 h r 3 9 . 3 9 0 . 5 2 4 0 . 0 0 0 . 1 5 4 5 m in 3 9 . 4 5 0 . 6 5 4 0 . 1 5 0 . 3 0

Present address: Department of Animal 'Subjective scores: 1 = pale, soft and exudative; 5 = dark, firm and dry. Science, University of Wyoming, Laramie, W yoming 82070. ’3Means on horizontal lines bearing different superscripts differ significantly (P < .05).

Volume 35 (19701-JOURNAL OF FOOD SCIENCE - 1 1 3 i M4—JOURNAL OF FOOD SCIENCE-Volume35 (1970)

Table 2—Means from thermograms of epimysial and intramuscular collagen from norm al and low quality muscle.1

Epimysial Intramuscular

0 hr post-m ortem 24 hr post-m ortem 0 hr post-m ortem

L o w L o w L o w

q u a l i t y N o r m a l q u a l i t y N o r m a l q u a l i t y N o r m a l

N u m b e r 5 5 5 5 5 5

O n s e t , ° C 5 9 . 4 8 5 9 . 2 8 5 5 . 0 8 5 4 . 1 8 5 9 . 5 8 6 0 . 9 2 Extrapolated onset, °C 6 1 . 2 0 6 1 . 5 6 5 8 . 5 8 5 9 . 0 6 6 2 . 5 8 6 3 . 3 8

P e a k , ° C 6 5 . 6 6 6 5 . 8 8 6 3 . 9 6 6 4 . 5 2 6 6 . 2 8 6 7 . 2 8

Extrapolated recovery, °C 7 5 . 8 2 7 5 . 4 2 7 2 . 0 8 7 1 . 9 6 7 1 . 6 8 7 2 . 7 8

Recovery, °C 8 0 . 2 8 8 0 . 9 2 7 5 . 9 2 7 6 . 2 2 7 5 . 9 8 7 6 . 0 8

1 Sec Figure 1 for a definition o f the therm ogram points.

Fig. 1.—Effect o f sample preparation on shape o f thermogram for epimysial collagen (18- m onth-old steer). T scale setting (7 7 = 1 (f 72.54 cm ; temperature differential between sample and reference ('''■ T) = 0.2° 72.54 cm; fibers (Gustavson, 1956). The saline solu­ and heating rate o f collagen = 10° /min. The Differences in the shape of the ther­ sharp peak (solid line) is for epim ysial collagen mograms in Figure 1 are due to differ­ tion per se would not be expected to mixed in 0.9% NaCI and held for 1 hr. The ences in sample preparation. The sharper cause changes in the thermal characteris­ broad peak (dotted line) is for collagen placed in 0.9% NaCI and run im m ediately. thermogram represented by the solid line tics of the fibers during a 1 -hr period. is connective tissue which was removed Means from thermograms of longissi- from the freezer, mixed in 0.9% NaCI and mus epimysial and intramuscular collagen allowed to stand 1 hr as previously de­ for the normal and low quality pigs are Analyzer were obtained on epimysial connec­ scribed (Experimental). The thermogram given in Table 2. Collagen from low quali­ tive tissue from the norm al and low quality por­ represented by the broader melting peak ty muscles consistently gave slightly low­ cine muscle at 0 and 24 hr post-mortem. In is from connective tissue removed from er peak melting points than that from addition, the m elting characteristics o f intra­ the freezer, placed in 0.9% NaCI and anal­ normal muscle. This was also true for muscular collagen were determined at 0 hr yzed immediately without wetting or thermal transitions occurring at the extra­ post-mortem. Approxim ately 30-mg samples mixing. Onset and extrapolated onset polated onset and the recovery. Onset o f epim ysial or intram uscular connective tissue were similar in both cases. However, the temperatures for the epimysial collagen at were placed in glass tubes 25 m m in length and 0 hr post-mortem did not follow this pat­ 4 mm in diameter. They were mixed w ith 50k sharper peak was chosen since it gave o f 0.9% NaCI in distilled water and allowed to melting points for collagen that were clos­ tern. However, onset temperatures are stand at room temperature 1 hr, w ith interm it­ er to the values reported in the literature. open tc some subjective interpretation tent m ixing w ith a straightened paper clip. In addition, the sharper peak was more and they show more variation in multiple Therm ocouples were placed into the center of reproducible. Mechanical working of runs on the same tissue than the other the 4-m m tubes containing the collagenous fi­ leather in water prior to determining temperature parameters. Differences in bers. Shrinkage temperatures of the samples DTA peak temperatures was also the means between the normal and low quali­ were recorded w ith the DTA apparatus. The T method employed by Witnauer et al., ty tissues shown in Table 2 are not signifi­ Scale was set at 10°/2.54 cm, the A j Scale at (1964). Each curve in Figure 1 represents cant (P < .05). Nevertheless, the differ­ 0.2° /2.54 cm and the heating rate at 10° /m in ences in peak melting points between low in an air atmosphere. Duplicate determ inations an average of 7 thermograms. Average were made on all samples. peak temperatures and standard devia­ quality and normal muscle collagen of tions were 64.08 and 0.63, respectively, 0.56°C at the 24-hr post-mortem period RESULTS & DISCUSSION for epimysial samples held 1 hr in 0.9% are similar to the highly significant (P < NaCI, compared to 68.26 and 1.02 for .01) difference of 0.63°C reported by THERMOGRAMS of epimysial tissue are those run immediately. The difference in McClain (1968). The difference in statis­ shown in Figure 1. The thermograms tical significance between this study and trate the type of curves obtained with peak temperatures is probably due to wetting of the collagen fibers by mixing the previous one may be due to smaller porcine connective tissue and define the quality differences in this study. Collagen location of the transition temperatures re­ them in 0.9% NaCI. The fibers became from the low quality muscle at 0 hr post­ ported in this paper as follows: more dispersed (less tightly bound or less concentrated) when compared to those mortem also melted at lower tempera­ Onset: The temperature at which the which were not mixed and air was re­ tures than that from the normal tissues. thermogram starts to depart from the moved from the sample in the DTA Although the small differences in base line. tubes. This supports the work of Lennox shrinkage temperatures between collagen Extrapolated onset: The temperature (1949), who stated that the shrinkage from low quality and normal muscle are corresponding to the intersection of ex­ temperature of collagen increases with in­ not stat.stically significant, it is possible trapolation of the base line and the long­ creasing concentration and decreasing that they are real. However, little has est straight-line section of the low- moisture content. It also seems probable been reported on the relationship of con­ temperature side of the peak. that the samples allowed to stand for 1 hr nective tissue characteristics to quality of Peak: The temperature of reversal. in 0.9% N ad lost some of their ground muscle substantiating this. If connective Extrapolated recovery: Same as extra­ substance into solution (Lowther et al., tissue differences at 0 hr post-mortem do polated onset, except it is on the high- 1967; McClain, 1968), which would also exist, these differences may be extremely temperature side of the peak. make the fibers less concentrated or more meaningful in further elucidating the PSE Recovery: The temperature at which disperse. Saline solutions have commonly condition. the thermogram returns to either the been used by researchers to prevent Means showing different points on same or a different base line. changes such as swelling in collagenous thermograms of the epimysium at 0 hr

116-JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

McClain, P.E. 1969. Isolation of intramuscular Verzar, F. 1964. Aging of the collagen fiber. In imino acid. Arch. Biochem. Biophys. 93: connective tissue. Nature 221: 181-182. “International Review of Connective Tissue 440-447. McClain, P.E., Pearson, A.M., Miller, E.R. and Research.” ed. Hall, D.A. Vol. 2, pp. Ms. received 3/25/69; revised 10/31/69; ac- Dugan, L.R. Jr. 1968. Application of differ^ 243-325. Academic Press, New York. cepted 11/11/69.______ential thermal analysis to the study of Witnauer, L.P. and Wisnewski, A. 1964. Abso­ hydrothermal shrinkage in epimysial and lute measurement of shrinkage temperature Michigan Agricultural Experiment Station corium collagen. Biochim. Biophys. Acta of leather by differential thermal analysis. J. Journal Article No. 4737. 168: 143-149. Amer. Leather Chem. Assoc. 59: 598—612. This investigation was supported in part by Rigby, B.J. 1964. The effect of mechanical ex­ Woessner, J.F. Jr. 1961. The determination of Public Health Service Research Grant No. tension upon the thermal stability of colla­ hydroxy proline in tissue and protein sam­ UI—00191 from the National Center for Urban gen. Biochim. Biophys. Acta 79: 634-636. ples containing small proportions of this and Industrial Health.

JOSEPH H. BRUEMMER and BONGWOO ROE Fruit and Vegetable Products Lab. USD A Southern Utilization R & D Div., ARS , Winter Haven, Florida 33880

ENZYMIC OXIDATION OF SIMPLE DIPHENOLS AND FLAVONOIDS BY ORANGE JUICE EXTRACTS

SUMMARY—Extracts o f orange juice vesicles oxidized p-hydroquinone (HQ) and o-dihydroxy- MATERIALS & METHODS phenylalanine (DOPA) with 0 2. Most o f the oxidase activity was associated with a particulate fraction that sedimented at 100,000 x g. Sonic disruption o f the particulates followed by chroma­ CHEMICALS were obtained as follows: DL-3, tography on DEAE-cellulose increased specific activities with both substrates. The partially puri­ 4-dihydroxyphenylaianine (DOPA), nicotin­ fied enzyme oxidized numerous naturally occurring o-diphenols and o-methoxyphenols. Acidic amide-adenine dinucleotide, reduced (N AD H ), nicotinam ide-adenine dinucleotide, phosphate, media below pH 4 and heating above 70°C destroyed most o f the oxidase activity. KCN, diethyl- reduced (NADPH), 8-hydroxyquinohne and dithiocarbamate and 8-hydroxyquinoline but not EDTA inhibited the oxidase. The partially puri­ rutin from Sigma Chemical Company; tris- fied enzyme reduced benzoquinone and oxidized the reduced form of nicotinamide-dinucleotide (hydroxymethyl)aminom ethane (Tris), ethyl- (NADH). Ubiquinone, a benzoquinone derivative, but not menadione or vitamin K¡, naphtho­ enediam inetetraacetic acid (ED TA), L-ascorbic quinone derivatives, could replace benzoquinone in the oxidation o f NADH. Ubiquinone, benzo­ acid, potassium cyanide, p-benzoquinone, resor­ quinone and similar p-quinones may function in the orange as the oxidation-reduction couple cinol, p-phenylenediam ine, 2-m ercapto- between NADH-quinone reductase and diphenol oxidase. benzothiazole, hydroquinone, sodium dieth- yldithiocarbam ate, ammonium sulfate, is saturated at relatively low partial pres­ pyrogallol and sulfosalicylic acid from Fisher INTRODUCTION Scientific Company; catechol and L-tyrosine 0 2 DIPHENOL oxidases (EC No. 1.10.3.1 sures of and therefore its activity is from N utritional Biochemicals Company; and EC No. 1.10.3.2), ascorbic acid oxi­ practically independent of 0 2 -tension. o-catechin, chlorogenic acid and ferulic acid dase (EC No. 1.10.3.3) and cytochrome Biale found that respiration in orange from K & K Laboratories; ubiquinone, DEAE oxidase (EC No. 1.9.3.1) have been con­ slices was a function of 0 2-tension in the cellulose, hom ovanillic acid, vitam in K t , mena­ sidered as terminal oxidases in plant respi­ range of 0—8% 0 2. He showed that C0 2 dione, quercetin, 3-methoxy, 4-hydroxy- ration, but most evidence supports cyto­ production by oranges decreased rapidly phenylethylam ine, hesperidin, hesperetin, quercetrin, naringin, naringenin and criodictyol chrome oxidase as the 0 2-activating agent at subatmospheric levels of 0 2. Below from Calbiochcm Company. Feruloylputrescine (Beevers, 1961). In citrus fruit, however, 2.5% O2 the respiratory quotient (R.Q.) rose rapidly, indicating that anaerobic re­ was provded by Dr. Adair Wheaton, Florida the evidence is very weak. The cyto­ Citrus Experim ent Station, Lake Alfred. chrome system was implicated in orange actions dominated. These results suggest­ ed that phenol oxidase and ascorbic acid Preparation and fractionation of juice cell fruit respiration by Hussein (1944). He extract showed that CO inhibited 20% of 0 2- oxidase might be involved in respiration because they require more Oj for satura­ Peeled, mature oranges var. Valencia were uptake by orange flavedo slices and that quick-frozen in liquid nitrogen and the frozen light completely reversed the inhibition. tion than does cytochrome oxidase. How­ vesicles separated from seeds and m em branous (CO inhibition of diphenol oxidase and ever, ascorbic acid oxidase was not found tissues. The vesicles were m ixed in sufficient 1 ascorbic acid oxidase is not reversed by in mature oranges (Baker et al., 1968) M Tris in frozen-bead form to neutralize the light.) However, Hussein neither identi­ and diphenol oxidase has not been report­ fruit acid, then powdered in a precooled fied the cytochromes nor demonstrated ed in citrus. blender. This powder was thawed and strained through four layers of cheese cloth. The neu­ their function in respiration. To determine whether diphenol oxi­ Biale (1961) questioned whether cyto­ tralized orange juice was then centrifuged at dase participates in citrus respiration, 1,000 x g for 10 m in in a Lourdes Bctafuge at chrome oxidase solely could account for orange juice cells were examined for oxi­ 5°C. The supernatant separated from '.he resi­ respiration in citrus. Cytochrome oxidase dation of o- and p-diphenols. This paper due was designated Juice Cell E xtract (JCE). describes the partial purification of The JCE was fractionated into 2 particulate diphenol oxidase and its characterization aPaper presented at the 29th Annual Meet­ fractions, R 10 a n d R1 0 o, a n d 1 soluble fraction ing of the Institute of Food Technologists in as a potential terminal oxidase in the by centrifugation. R 10 was obtained by centri­ Chicago. reoxidation of NADH. fuging the extract at 10,000 x g f o r 10 m i n . ENZYMIC OXIDATION OF DlPHENOLS AND FLAVONO!DS-WT

Ri oo was obtained from the supernatant of Protein determination tained suspended particles salted out by R io by centrifuging it at x g for 60 100,000 Protein of samples was determined by the saturating the extract to 30% maximum min. The supernatant from R ioo was desig­ sulfosalicylic acid method as described by with ammonium sulfate. All the oxidase nated as the soluble fraction. Both R ]o and Layne (1957). activity followed the suspended particles Rioo were resuspended in 0.05 M Tris, pH 7.0, Quinone reductase into the residue. The data in Table 1 to give a protein concentration o f 1 m g / m l. Reductase activity was measured by record­ show that most of the HQ oxidase activi­ Purification o f diphenol oxidase ing the rate o f NADH oxidation at 340 h i m a n d ty is concentrated in the fine particle The JCE was saturated to 90% w ith ammo­ also by recording the rate of reduction of fraction that can be sedimented between nium sulfate, adjusted to pH 7.0, equilibrated 1,4-benzoquinone at 290 mp. Reference and for 1 hr at 0°C and centrifuged at 5,000 x g for 10.000 and 100,000 x g. Most of the as­ sample cuvette contained 0.45 Mmole o f benzo- 10 min at 0°C. The clear supernatant was corbate-reduced cytochrome c oxidase quinone and 0.5 Mmole of NADH in 2.5 m l of decanted and the residue resuspended in a m ini­ and succinoxidase activities were associat­ 0.1 M potassium phosphate buffer, pH 6.5. Re­ m um o f 0.05 M Tris, pH 7.0, and sonicated for ed with the fraction that sedimented at action was initiated by adding the enzym e prep­ 30 sec at 0°C in a Sonifier Cell D isruptor. The 1 0 . 0 0 0 aration and 0.1 M buffer to the sample cuvette x g. sonicate was dialyzed for 16 hr against 0.005 M and 0.1 M buffer to the reference cuvette fer a Purification of HQ and DOPA oxidases Tris, pH 8.0, then centrifuged at 5,000 x g for total volume in each of 3 ml. Activity of the 10 m in. The clear supernatant was diluted w ith Specific activities of both HQ and preparation was calculated from the change in the buffer to 1 mg protein per m illiliter and 40 DOPA oxidases increased about 20-fold absorbance using 6.22 x 103 as molar extinc­ m l of it was applied to a 5°C-precooled DEAE after sonication and chromatography on tion coefficient. 1 unit o f activity was defined cellulose colum n (25 by 500 m m) equilibrated as oxidation of 1 Mmole NADH/m in/m g pro­ DEAE cellulose. Figure 1 shows that HQ w ith 0.005 M Tris, pH 8. The column was elu­ t e in . and DOPA oxidases and quinone reduc­ ted w ith a parabolic NaCl gradient as described tase (QR) eluted into the same fractions. by M oore et al. (1960). 3, 10-m l fractions were RESULTS Attempts were made to purify HQ and collected per hour. The colum n, eluting solu­ Distribution of hydroquinone oxidase in tion and eluates were m aintained at 5°C. DOPA oxidases with acetone fractiona­ juice cell extract (JCE) tion, temperature and pH selectivity. Assay o f oxidase JCE oxidized hydroquinone (HQ) and Both oxidase activities were lost on ex­ Oxidase activities were assayed by the fol­ dihydroxyphenylalanine (DOPA), respec­ posure to acetone, and temperature- and lowing methods: tively, at rates of 0.2 and 0.3 /iliter pH-stability did not differentiate the ac­ 1. Oxygen uptake (Gregory et al., 1966). 2. 2 Form ation of chromogen product, DOPA- 0 /min/mg protein. The cell extract con­ tive enzyme from other protein. chrom e (Constantinides et al., 1967). 3. O xida­ tion o f ascorbate (Ei-Boyoum i et al., 1957). 1. 0 2-uptakewas measured at 30°C with the YS1 Model 53 Biological Oxygen M onitor (Yellow Springs Instrum ent Company, Yellow Springs, Ohio). The reaction vessel contained the substrate in 0.1 M potassium phosphate buffer, pH 7.0, saturated w ith air and the en­ zyme preparation in a total volume of 4 ml. Rate o f 0 2-uptake was recorded on a Sargent recorder, Model-SR, calibrated in m icroliters of 0 2. Specific activities of preparations were expressed as m 1 02/m in/m g protein. 2. DOPAchrome form ation was measured at 30°C by recording the rate of increase in absorbance at 475 mu in a G ilford Spectropho­ tom eter, M odel 222. The sample and reference cuvettes contained 0.1 mmole of DOPA in 2.5 m l 0.1 M potassium phosphate buffer, pH 7.0. The reaction was initiated by adding the en­ zyme preparation and 0.1 M buffer to the sam­ ple cuvette and 0.1 M buffer to the reference cuvette for a total volum e in each o f 3 m l. Spe­ cific activities o f preparations were calculated from the change in absorbance and expressed as Fig. 1 —Localization of enzyme activities. Activities of HQ, hydro­

a A /m in/m g protein o f sample. quinone, and DOPA, dihydroxyIphenylalaine, are in gl iters 0 2/m in /m l, 3. O xidation o f ascorbate was measured at and QFt, quinone reductase, in gmoles NADH /m in /ml. 30°C by recording the rate of decrease in ab­ sorbance at 265 mp in the G ilford spectropho­ tom eter. Ascorbate is oxidized nonenzym ically by the oxidized substrate and is used here as the redox indicator of the rate o f oxidation of Table 1—Distribution of hydroquinone and cytochrome c oxidases the substrate. Reference and sample cuvette in juice cell extracts. contained 0.1 Mmole o f ascorbate, 0.1 Mmole o f EDTA to prevent 0 2-metallic oxidation of as­ M lite r s 0 2 /m in/m g protein 2 corbate and 0.6 Mmole o f substrate in 2.5 m l of Ascorbate plus Succinate plus 0.1 M potassium phosphate buffer. The reac­ Hydroquinone cytochrom e c cytochrom e c tion was initiated by adding the enzyme prepa­ F r a c t i o n 1 ration and 0.1 M phosphate buffer to the sam­ R10 0.2 7 .1 3 . 4 ple cuvette and 0.1 M phosphate buffer to the R100 2 .3 1 .3 0 reference cuvette for a total volum e in each o f S o lu b le 0 . 4 0 3 ml. Specific activities of the preparations 0

were calculated from change in absorbance and 1 R i 0 designates residue after centrifuging extract at 10,000 X g for expressed as a A/m in/m g protein of sample. 10 m in . R ioo designates residue after centrifuging supernatant from This value was converted to uliters 0 2 b y u s in g R io at 100,000 X g for 60 min. the m olar absorption coefficient, 1.5 X 104, and 11.2 uliters 0 2 per Mmole ascorbate. 2 A ll substrates were 10 2 M . 1ÌS-JOUFÌNAL OF FOOD SCIENCE-Volume 35 (1970)

Table 2—pH and temperature stability o f DOPA orange oxidase. 1 Table 3—Inhibition of orange DOPA oxidase. 1

pH stability relative Tem perature stability M cone % Inhibition

t o p H 7 relative to 0°C EDTA 1 .3 X 10~3 0 DOPA oxidase DOPA oxidase R e s o r c in o l 3 . 3 X 10-4 4 0 T e m p (% activity) p H (% activity) 8-H ydroxy quinoline 1 .3 X 10"3 6 0 9 9 4 0 100 Potassium cyanide 3 .3 X 10-4 100 8 100 3 0 9 4 Sodium diethyldithiocarbam ate 1 .7 X 10'3 100 7 100 5 0 5 5 1 Substances were added directly to DOPA-enzyme reaction m ixture 9 4 7 0 6 6 in the DOPAchrome assay. 5 8 7 7 5 0 4 2 5 3 12

'P artially purified enzyme, 1 mg protein/m l, was exposed to the t e s t p H or temperature for 5 :min. Then pH was adjusted to 7, tern- and the phytyl derivative, vitamin K1; parature to 30°C and activity determ ined by DO PAchrom e assay. were inactive. NADH was also oxidized by the enzyme preparation when benzoquin­ one was replaced by HQ, or by the o-di- phenols, DOPA and catechol. Homovanil- lic acid or other methoxyphenolic com­ pH and temperature stability was about 25% of the rate with NADH. pounds could not replace the diphenolics. Table 2 shows that pH stability drops Table 5 shows that QR activity is specific Rate of oxidation of NADH with benzo­ off rapidly below pH 5. Only 25% of the for benzoquinone. The 2-methyl deriva­ quinone was higher than with the di­ activity remained at pH 4. A substantial tive of 1,4-naphthoquinone (menadione) phenol forms. drop in activity occurred as the tempera­ ture increased from 50 to 70°C. Only 6 % of the activity remained after heating to 70°C for 5 min. Table 4--Substrate specificity o f orange diphenol oxidase. Inhibitors S u b s t r a t e p l i t e r s 02 /m in/m g protein EDTA had no effect on DOPA oxidase o-Dihydroxyl and was used to suppress autoxidation of DOPA 3-(3,4-Dihydroxyphenyl)alanine 7 . 0 ascorbic acid in the ascorbic acid assay P y r o g a l l o l 1 1,2,3-Trihy droxybenzene 4 . 6 for DOPA oxidase. Table 3 lists a number C a t e c h o l2 1,2-D ihydroxy benzene 4 . 0 of inhibitors of DOPA oxidation. Resorci­ nol, the meta-hydroquinone, and p-Dihydroxyl

8 -hydroxyquinoline did not inhibit com­ p-Hydroquinone 1,4-D ihydroxy benzene 4 . 5 pletely at 5 x 10 2 M. o-Hydroxyl-m ethoxy Substrate specificity The partially purified oxidase showed Hom ovanillic Acid 4-Hydroxy-3-M ethoxyphenylacetic Acid 1 .4 specificity for o- and p-diphenolics and 3-M ethoxytyram ine 4-Hydroxy-3-M ethoxyphenylethylam ine 0 . 3 4 Ferulic A cid 3 4-Hydroxy-3-M ethoxycinnam ic Acid 0 . 1 4 o-methoxyphenolics. Table 4 lists a num­ Feruloylputrescine 3 ber of simple diphenolic and flavonoid N-(4-Am inobutyl)-4-hydroxy-3-m ethoxycinnam ide 0.21 H e s p e r e t in 3 , 5, 7-Trihydroxy-4-m ethoxyflavanone 0 . 5 6 compounds that supported 0 2 uptake with the enzyme preparation. All of the H e s p e r id in Hesperitin-7 rutinoside 0 . 5 6 flavonoids and many of the simple di- o-Dihydroxyi phenolics have been reported in citrus C a t e c h in 3,5,7, 3 ,4 -Pentahydroxyflavan 2 . 7 (Horowitz, 1961). Feruloylputrescine was Chlorogenic Acid 3-(3,4-D ihydroxycinnam oyl)quinic acid 0 . 8 1 recently isolated from the orange (Wheat­ Eriodictyol 3 , 4 , 5 , 7-Tetrahydroxyflavanone 1.1 on et al., 1965). Rates for the o-and p-di­ R u t i n Q uercetin-3-rutinoside 0 . 5 6 phenolics were about equal and were higher than for the o-methoxyphenolics. Q u e r c e t in 5 , 7 , 3 , 4,' -Tetrahydroxyflavonol 1 .3 Rates for the o-methoxyphenolic com­ Q u e r c e t r in Quercetin-3-rham noside 1 .3

pounds decreased in order as the substi­ m -Dihydroxyl tuted group changed from acid to amine to a, /3-unsaturated acid. Quercetrin and R e s o r c in o l 1,3-D ihydroxy benzene 0 hesperidin supported the same rate of M ono-hydroxyl 0 2-uptake as did the aglycones, quercetin L - T y r o s in e 3-(4-Hydroxyphenyl)alanine 0 and hesperetin. Tyrosine, resorcinol and N a r in g e n in 4 '5 ,7-T rihy droxy flavanone 0 naringenin (4 ',5 ,7 -trihydro xyflavanone) N a r i n g in Naringenin-7-neohesperidoside 0 were inactive; they do not have the requi­ site o- or p-dihydroxy group. Vitamin Ks p - D i a m i n o ( 2 -m eth y l-4-a mi no- 1 -naphthol) and p-Phenylene diam ine 1,4-Diam inobenzene 0 p-phenylenediamine also were inactive. p-Am ino-hydroxyl Quinone reductase V i t a m i n K s The purified enzyme preparation oxi­ 4-Am ino-2-m ethyl-l-naphthol 0 dized NADH with p-benzoquinone; ’ Assayed at pH 5.6 instead o f pH 7. p-hydroquinone was formed in the reac­ 2Catechol 1 x 10 '8 M with ascorbate 1 X 10 '4 M and EDTA 1 X 10 '4 M . tion. The rate of oxidation of NADPH 3Ferulic acid or feruloylputrescene 3 x Iff5, ascorbate 1 X 10 '4 M and EDTA 1 X 10 '4 M . ENZYMIC OXIDATION OF DlPHENOLS AND FLAVONOIDS- 1 1 9

Table 5—Substrate specificity of orange The orange enzyme is different from Biale, J.B. 1961. Postharvest physiology and chemistry. In “The Orange—Its Biochemis­ quinone reductase■ these oxidases, in that methoxy but not try and Physiology,” ed. Sinclair, Walton B. pmoles NADH /min/mg protein amine can substitute for one of the p. 96. University of California, Berkeley. hydroxyl groups on the substrate. This Clayton, R.A. 1959. Properties of tobacco Benzoquinone 1.6 polyphenol oxidase. Arch. Biochem. capacity of the orange enzyme to oxidize Biophys. 81: 404—417. U b iq u i n o n e 0.6 o-methoxyphenolic compounds may be Constantinides, S.M. and Bedford, C.L. 1967. M e n a d io n e 0.0 Multiple forms of phenol oxidase. J. Food metabolically significant. However, the V i t a m i n K j 0.0 Sci. 32: 446—450. unknown oxidation product of the o- Eiger, I.Z. and Dawson, C.R. 1949. Sweet pota­ to phenolase, preparation, properties, and methoxyphenolic compounds did not determination of protein content. Arch. support the oxidation of NADH in the Biochem. 21: 194—209. El-Boyoumi, M.A. and Frieden, E. 1957. A QR assay. Therefore, the methoxy deriva­ spectrophotometric method for the deter­ DISCUSSION tives do not function as intermediates in mination of the catecholase activity of tyro­ sinase and some of its application. J. Am. THE DIPHENOL oxidase in oranges as in the reoxidation of NADH. Chem. Soc. 79: 4854—4858. Demonstration thato-and p-diphenolic Gregory, R.P.F. and Bendall, D.S. 1966. The apples (Harel et al., 1966) is concentrated purification and some properties of the in the particulate fraction. Diphenol oxi­ compounds support the oxidation of polyphenol oxidase from tea. Biochem. J. dase is also particulate-bound in tea NADH with 0 2 uptake by the purified 101: 569-581. preparation is evidence that orange juice Harel, E., Mayer, A.M. and Shain, Y. 1966. (Gregory et al., 1966), lettuce seedlings Catechol oxidases, endogenous substrates (Mayer, 1961), tobacco (McClendon, cells contain components of a NADH- and browning in developing apples. J. Sci. linked terminal oxidase. The lower rate of Food Agr. 17: 389—392. 1953) and sugar beet (Mayer et al., Horowitz, R.M. 1961. The citrus flavonoids. In 1960). Using cytochrome c oxidase as the oxidation with the diphenolics compared “ The Orange,” ed. Sinclair, Walton B. to benzoquinone indicates that the di­ p.334. Div. Agr. Sci., Univ. of Calif., Berke­ enzyme marker for mitochondria, the ley. data on distribution can be considered phenol oxidase is rate-limiting in the Hussein, A.A. 1944. Respiration in the orange. evidence for the concentration of orange over-all reaction. Contribution of this oxi­ A study of systems responsible for oxygen dase system to the total respiratory 0 2 uptake. J. Biol. Chem. 155: 201—211. diphenol oxidase in a particle smaller Layne, E. 1957. Spectophotometric and tur- than the mitochondria. This particulate uptake must be determined to assess its bidimetric methods for measuring proteins. significance to citrus metabolism. In “Methods in Enzymology,” eds. Colo- fraction has net been examined for other wick, S.P. and Kaplan, N.O. Vol. Ill, p. 447. enzyme functions, but it corresponds in The poor stability of diphenol oxidase Academic Press, New York. Mayer, A.M. 1961. Nature and location of phe­ sedimentation rate to microsomal frac­ below pH 4 indicates why diphenol oxi­ nolase in germinating lettuce. Physiol. Plan- tion of plant and animal cells. dase has not hitherto been detected in tarum 14:322—331. orange juice. This inactivation in acidic Mayer, A.M. and Friend, J. 1960. Localization In contrast to the 5,000-fold purifica­ and nature of phenolase in sugar-beet leaves. tion of diphenol oxidase from tea solution might explain why enzymic J. Exptl. Botany 11: 141—150. browning is not a problem in citrus pro­ McClendon, J.H. 1953. Intracellular localiza­ (Gregory et al., 1966), orange diphenol tion of enzymes in tobacco leaves (II). Cyto­ oxidase was purified only about 2 0 -fold. cessing and why ascorbic acid is stable in chrome oxidase, catalase and polyphenol The specific activity of the tea enzyme is the presence of the many phenolic com­ oxidase. Am. J. Botany 40: 260—266. Moore, B.W. and Lee, R.H. 1960. Chromatog­ 1 0 0 times higher than the orange enzyme. pounds in citrus juices. raphy of rat liver soluble protein and locali­ Even with limited purity, the response to zation of enzyme activities. J. Biol. Chem. REFERENCES 235: 1359—1364. inhibitors, pH and temperature stability Wheaton, T.A. and Stewart, I. 1965. Feruloyl- and substrate specificity suggests that Baker. R.A. and Bruemmer, J.H. 1968. Oxida­ putrescine: Isolation and identification from tion of ascorbic acid by enzyme preparation citrus leaves and fruit. Nature 206: orange diphenol oxidase is similar to the from orange. Proc. Florida State Hort. Soe. 620—621. enzymes purified from tea (Gregory et 81: 269—275. Ms. received 5/29/69; accepted 8/25/69. Beevers, II. 1961. “Respiratory Metabolism in al., 1966), tobacco (Clayton, 1959) and Plants,” p. 105. Row, Peterson and Co., References in this paper to specific commer­ sweet potato (Eiger et al., 1949). White Plains, New York. cial products do not constitute endorsement. R. DALRYMPLE, R. G. CASSENS and E. J. BRISKEY

Department o f Meat and Anim ai Science

University o f Wisconsin, Madison, Wisconsin 53706

A METHOD FOR ESTIMATING THE FUNCTIONAL CAPILLARY SYSTEM IN SKELETAL MUSCLE

SUM M ARY—The functional capillary system was investigated by injecting trypan blue dye into 8 m l of aqueous 25% pyridine to each tube and m arket-weight porcine animals followed by extracting the dye from various muscles. The trapezius allowing them to stand covered for 4 hr at contained more dye than the biceps femoris, w ith the gluteus medius and longissimus dorsi being room temperature accompanied by frequent m ixing. A more rapid m ethod of extraction in­ equal in dye content but containing less dye than the trapezius and biceps femoris muscles. This volved the use of a 70°C water bath for 2 hr, rank compared well with the fiber type distribution, with the trapezius being high in red fiber also w ith frequent m ixing o f the covered tubes. content. The m ethod presents possibilities for study o f capillary function but some problem s were The samples were centrifuged for 10 m in at encountered with the injection procedure. 1,500 x g and then brought to a volume of 5 ml. The concentration of trypan blue in each INTRODUCTION trogen and pulverized in a W aring Blendor to supernatant was then determined colorim etri- obtain a fine muscle powder, then stored for THE IMPORTANCE of the capillary net­ cally by reading at 610 mju in a B & L Spec- analysis. Samples of frozen muscle powder work in skeletal muscles is in the control tronic 20. The values obtained were compared weighing 200—400 mg were weighed into 15-ml of metabolic functions within the living w ith standards o f 0, 5, 10, 20, 30 and 40 Mg o f centrifuge tubes. The follow ing procedure was animal. The capillary function may also trypan blue in 5 m l of aqueous 25% pyridine. used to extract the bound trypan blue from the play a significant role in setting the stage samples: The muscle samples were m ixed w ith for post-mortem conversion of muscle to 5 m l o f cold 5% TCA and allowed to stand in meat. the cold for a m inim um o f 5 m in. A fter centri­

fuging at 1,500 X g (0°C) for 10 m in the super­ OTHER MUSCLES A method for estimating the capillary DYE C O N C E N T R A T IO N TRAPEZIUS function of skeletal muscle, expressed as natants were discarded and the interiors o f the permeability and extent of vasculariza­ tubes wiped with absorbent paper wrapped tion, has been developed by modifying around a glass rod. This procedure was found to the procedures of Judah et al. (1962) and be helpful in preventing turbidity in the final extract. To remove lipid m aterial, the precipi­ Hawk (1968). Cecil et al. (1966) and tate was extracted w ith 5 m l o f absolute etha­ Cooper et al. (1968) have used a similar nol at room temperature for 5 m in, and, after trypan blue dye method to quantitate the centrifugation at 1,500 X g for 5 m in, re-ex- capillary function of uterus tissue of rats tracted twice w ith absolute ethanol at 70°C for and ewes as it is affected by endocrine 5 min. Following centrifugation (5 min at and other physiological factors. With this 1,500 x g) the inside walls of the centrifuge in mind, we developed a procedure for tubes were again wiped w ith absorbent paper. using trypan blue dye to quantitate capil­ The trypan blue was either immediately ex­ lary function in porcine skeletal muscle. tracted from the precipitate or the dye protein Fig. 1—Relationship o f trypan blue concentra­ com plex was stored in a freezer for subsequent tion in the 4 muscles examined, expressed as Trypan blue is an azo dye compound; it analysis. Extraction of trypan blue from the ratio o f dye per gram muscle to dye per gram has relatively low toxicity when injected protein com plex was accomplished by adding 4 tra p e z iu s . into the circulatory system of living ani­ mals. The particulate dye is ingested com­ pletely by the fixed macrophages con­ tained in the connective tissue that is a structural part of the capillary wall (Bloom et al., 1968).

MATERIALS & METHODS A SOLUTION of 8% trypan blue and 0.5% NaCl approximately of body temperature (39°C) was injected into the jugular vein o f 8 heavy market-weight (100-120 kg) Chester W hite and Poland China pigs anesthetized w ith 50/50 dial with urethane and nembutal. The dye solution was injected at the level o f 3 m l per kg o f body weight. The trypan blue dye was allowed to circulate for 30 m in, whereupon the pigs were exsanguinated by severing the ante­ rio r vena cava. (This process allows the blood to flow from the tissues.) The biceps fem oris, glu­ teus m edius, longissimus and trapezius muscles were excised and trim m ed of exterior fat and connective tissue. Large samples (50-g) were obtained from standardized locations (medial and both distal ends) o f the entire longissimus muscle so as to be representative of the com- Fig. 2.—Fresh frozen sections (magnification 50 X ) representing the 4 muscles exam ined in the dye plete muscle. The trapezius was used in its en- experiment. Each section was reacted for DPNH-TR so Type / fibers stained positive and appear tirety. These samples were frozen in liquid ni- dark, whereas Type II fibers stain only slightly and appear light.

120-JOURNAL OF FOOD SCIENCE-Volume 35 (1970) ESTIMATING MUSCLE CAPILLARY SYSTEM- 121

RESULTS & DISCUSSION muscle to g g dye/g trapezius. Through slightly greater for the gluteus medius the use of this trypan blue method the than for the longissimus, indicating slight­ FIGURE 1 shows the :rypan blue concen­ longissimus and gluteus medius were ly more variation between animals in the tration in the 4 muscles examined. Mean found to have similar relative dye values, capillary function of this muscle. As seen values for the 8 pigs shown in this figure indicating similar vascular capabilities. in Figure 1, the biceps femoris had a are expressed as the ratio of g g dye/g The standard deviation of the mean was greater vascularization than the previous 2 muscles, but the relative value was only slightly higher than half that of the tra­ pezius. According to data presented in

Fig. 2—Concluded this figure the trapezius was a highly vas­ cularized muscle, as it had the highest dye concentration of all muscles examined. The standard deviation of the mean was quite small for all these relative values, indicating that this relationship between vascularization and dye binding was basi­ cally the same for all animals used, re­ gardless of breed. It has been shown by Romanul (1965) that the extent of vascularization in skeletal muscle is related to muscle fiber types. Type I or red fibers are aerobic and have been shown histologically to have a greater capillary to muscle fiber ratio than do Type II or white fibers which are anaerobic. It appears that at a muscle with a greater number of red fibers should have taken up a greater concen­ tration of trypan blue. To see if this relationship could be shown by our pro­ cedure, we made histochemical sections of the same 4 muscles in a control pig of similar age and weight. Figure 2 shows the sections stained with diphospho- pyridine nucleotide tetrazolium reductase (DPNH-TR). The fibers classified as Type I (or red) stain positive are the dark fibers. Type II (or white) fibers stain only lightly; thus appear light in these sections. The trapezius consists mainly of red fi­ bers and thus its circulation should be high. This supports the high trypan blue content found in the trapezius muscle as seen in Figure 1. The biceps femoris con­ tained fewer red fibers than the trapezius and this was also supported by the trypan blue data (Fig. 1). The longissimus and gluteus medius are seen in Figure 2 to be white muscles, due to their low content or red fibers. This indicates that their circulation and trypan blue content should also be low. This was definitely indicated from our trypan blue data in Figure 1. It can generally be concluded that the quantitative relationship of fiber types within a muscle is associated with the vascular capacity of that muscle; the trypan blue dye method can be used to determine these vascular capacities. An important aspect should be con­ sidered before the dye injection method is applied to projects to estimate capillary function. When it is desired to compare the concentration of a specific muscle from animal to animal, it is critical that the dye volume be constant per body weight. The controlled injection of large standard quantities of dye is a major dif­ ficulty in the application of this tech­ nique. 122-JOURNAL OF FOOD SCIENCE-Vo/ume 35 (1970)

REFERENCES Hawk, H.W. 1968. Personal communication. Published with the approval of the Director Bloom, W. and Fawcett, D.W. 1968. A Text­ Judah, J.D. and Willoughby, D.A. 1962. A of the Wisconsin Agricultural Experiment Sta­ book of Histology. 9th ed. W.B. Saunders quantitative method for the study of capil­ tion, Department of Meat and Animal Science, Co., Philadelphia. lary permeability: extraction and determina­ Manuscript No. 555. This research was sup­ Cecil, H.C., Hannum, J.A.Jr. and Bitman, J. tion of trypan blue in tissues. J. Pathol. ported in part by Public Health Service Re­ 1966. Quantitative characterization of uter­ Bacteriol. 83: 567-572. search Grant UI—00266—09 from the National ine vascular permeability changes with estro­ Romanul, F.C.A. 1965. Capillary supply and Center for Urban and Industrial Health, in part gen. Am. J. Physiol. 211: 1099-1102. metabolism of muscle fibers. Arch. Neurol. by Armour Food Research Laboratories and in Cooper, B.C. and Hawk, H.W. 1968. Vascular 12: 497-504. part by the American Meat Institute Founda­ function in the sheep uterus. J. Animal Sei. Ms. received 6/25/69; revised 10/8/69; accepted tion. 27: 1189. 10/10/69.

G. D. S A R A V A C O S New York State Agricultural Experiment Station Cornell University, Geneva, New York 14456

EFFECT OF TEMPERATURE ON VISCOSITY OF FRUIT JUICES AND PUREES

SUMMARY—The viscosities o f selected fru it juices and purees were measured with a coaxial-cylin­ Viscometric measurements der viscometer in the temperature range 20—70°C. Depectinized apple juice and Concord grape A ll measurements reported here were made juice were Newtonian fluids at all concentrations and their viscosity decreased considerably at w ith a coaxial-cylinder viscometer, the Con- higher temperatures. Cloudy apple and orange juices changed from Newtonian to pseudoplastic at traves Rheom at 15, m anufactured by Contraves concentrations higher than 50 and 20° Brix, respectively. Temperature had a smaller effect on AG , Zurich, Switzerland. It is the same instru­ viscosity o f cloudy juices than o f dear juices. The apparent viscosity o f fru it purees (pseudoplastic m ent as the Drage viscom eter described by van fluids) decreased slightly at higher temperatures. The activation energy for flow increased with the W azer et al. (1963). juice concentration and decreased with the presence o f suspended particles in the fru it product. Each liquid sample was placed in the appro­ The data and conclusions are useful in the design and operation o f efficient food-processing equipm ent. priate measuring system and brought to the de­ sired tem perature (20-70°C ) using a constant- INTRODUCTION effect of temperature on the composition tem perature water bath. Speeds o f the rotating of each. inner cylinder varied 5.6-352.0 rpm. The in­ VISCOSITY is usually considered an im­ EXPERIMENTAL strument can be operated at 15 different portant physical property related to the speeds, w hich are changed stepwise w ith a selec­ Fruit products quality of liquid food products. Viscomet­ tor switch. Readings of the instrum ents were Purees o f apple, pear and peach were pre­ ric data are also essential for the design converted to either viscosity (centipcise) or pared as described by Saravacos et al. (1967). and evaluation of food-processing equip­ shear stress, using tables supplied by die m anu­ The purees were stored in glass jars at 32°F for ment such as pumps, piping, heat ex­ facturer. The same tables contain the N ew toni­ 12 t o 20 months before measurements were changers, evaporators, sterilizers, filters an shear rates, which are a function of the and mixers. made. No visible changes o f consistency were rotational speed and the dimensions o f each noticed due to storage. Concord grape juice measuring system. Literature values of viscosity for many (64° Brix) was prepared commercially from F ru it purees and some juices are considered liquid foods have been reported at a sin­ New York State grapes and stored at 0°F for gle temperature and frequently no men­ 10 m onths before being used. M ost o f the tar­ pseudoplastic fluids, obeying the power-law tion is made of whether the material is a trates had been removed during concentration equation r = K 711, where r is the shear stress, Newtonian or a non-Newtonian fluid. of the grape juice. Commercial orange juice 7 is the shear rate, K is the flu id consistency co­ Such data are of limited usefulness in concentrate (44° Brix) from Florida was also efficient, and n is the flow behavior index. food processing, where temperatures and kept frozen before it was used. Juices o f lower The index n was obtained as the slope o f the straight line resulting when log r was plotted shear rates may be quite different from concentrations were prepared from the concen­ trates by dilution w ith distilled water. against log N, where N is the rotational speed the conditions of a single viscosity mea­ Tw o types o f juice were prepared in the pi­ (rpm). The shear rates ( 7) of pseudoplastic surement. lot plant from New York State apples. M ixed fluids were determ ined by dividing the tabulat­ Viscosities of water and sucrose solu­ varieties were ground in a ham m erm ill and the ed Newtonian shear rate by the compensation tions at various temperatures have been juice extracted in a continuous screwpress. factor C, calculated from the follow ing equa­ published (Perry, 1963). Harper (1960), Cloudy juice was prepared by filtration through tion (Contraves, 1966): and Saravacos et al. (1967) reported vis­ a rotary vacuum filter, operated w ith a Celite precoat. Clear, depectinized juice was prepared cometric constants of fruit purees at 2 temperatures. The effect of temperature by treating the extracted juice w ith a com mer­ on the apparent viscosity of pear puree cial pectic enzyme preparation and filtering through the precoat rotary filter. was investigated by Harper et al. (1962) Both the clear and the cloudy apple juices and the viscometry of tomato concen­ were concentrated under vacuum at 100°F in a C = trates reported by Harper et al. (1965). pilot falling-film evaporator. Samples o f juices Fruit juices and purees vary widely in at interm ediate concentrations were obtained their viscometric behavior and it is neces­ for viscom etric measurements. The highest con­ sary that each product be studied sepa­ centrations obtained were 75° Brix for the de­ rately. The purpose of this work was to pectinized apple juice and 65.5° Brix for the cloudy juice. Viscom etric measurements o f the where Ri is the radius of the rotating obtain the needed viscometric data on apple juices were made im m ediately after prep­ various fruit products and to relate the cylinder and Ra is the radius of the a r a t io n . measuring cup. VISCOSITY OF FRUIT JUICES AND PUREES-WZ

The fluid consistency coefficient K was ob­ tained graphically from a plot of log t vs. log y. The shear stress-shear rate diagram of each product was constructed by taking readings in the following order: Starting with the lowest speed, the ascending line was obtained first by increasing stepwise the speed of rotation. When the maximum speed was reached, the speed was decreased gradually and the descending line ob­ tained. In most products (Newtonian and pseudoplastic) both lines coincided. In thixo­ tropic fluids, the descending line fell below the corresponding ascending line.

RESULTS & DISCUSSION THE VISCOMETRIC constants of the various fruit products obtained with the coaxial-cylinder instrument were in gener­

al agreement with those obtained with JUICE CONCENTRATION. -BRIX the tube viscometer at a single tempera­ ture (27°C) by Saravacos (1968). Fig. 1—Viscosity o f apple juices (30°C). TEMPERATURE , *c Apple juices Depectinized (clear) apple juice be­ Fig. 2—Effect o f temperature on viscosity of haved as a simple Newtonian fluid at all depectinized apple juices. concentrations within the temperature range 20—70°C. For "his type of liquid, a single point measurement is sufficient to obtain the viscosity at the desired temper­ ature. The viscosity of apple juices in­ creased rapidly with concentration, par­ ticularly above 50° Brix (Fig. 1). Cloudy apple juice, which contained significant amounts of suspended and col­ loid particles, was more viscous than the depectinized juice. This type of juice was Newtonian below 50° Brix but at higher concentrations it behaved as a pseudo­ plastic fluid. Its flow behavior index (n) ranged from 0.85 at 50° Brix to 0.65 at 65.5° Brix. The apparent viscosities of cloudy apple juice, shown in Figures 1 and 3, were calculated for a shear rate of 100 sec 1. This shear rate was chosen as a moderate value for comparison of viscos­ ities among various fruit products. It should be noted that the apparent viscos­ ity of a pseudoplastic fluid increases at Fig. 3—Effect of temperature on viscosity of TEMPERATURE , °C lower shear rates (rotational speeds). cloudy apple juices. Fig. 4—Effect of temperature on viscosity of Viscosities of depectinized apple juices Concord grape juices. decreased considerably as the tempera­ ture was increased from 20 to 70°C (Fig. 2). An exponential relationship between temperature and viscosity is indicated by ic properties. Thixotropy was indicated the straight lines obtained when the data by a progressive decrease in the instru­ the effect of temperature was similar to are plotted on semilogarithmic paper. The ment reading at a given rotational speed that of depectinized apple juice. The and by formation of a thixotropic loop effect of temperature was more pro­ grape juice contained some suspended nounced at higher juice concentrations. when viscometric data of ascending and solids (mainly pectins and tartrates) These results are very similar to the viscos­ descending speeds were plotted. For pur­ which made the 64° Brix concentrate poses of analysis, thixotropic products ities of sucrose solutions at various tem­ slightly pseudoplastic (n = 0.90). were considered time-dependent pseudo­ peratures (Perry, 1 963). The commercial orange juice concen­ plastic fluids and their viscometric con­ Temperature had a less pronounced trate used in our experiments contained stants estimated by taking readings at significant amounts of suspended pulp of effect on the apparent vicosities of ascending speeds, 30 sec after the in­ cloudy apple juices (Fig. 3). The same various sizes. The 44 and 30° Brix con­ strument was brought to a particular centrates were pseudoplastic and their trend was noticed with concentration, speed. i.e., the more concentrated the juice, the flow behavior index (n) was 0.65 and steeper the decrease of viscosity upon Grape and orange juices 0.85, respectively. heating. Figure 4 shows the effect of tempera­ Diluted orange juices of concentra­ At temperatures lower than 50°C, the ture on the viscosity of Concord grape tions lower than 20° Brix were Newtoni­ concentrated cloudy apple juice juice. This juice was a Newtonian fluid at an. However, the determination of then- (65.5° Brix) exhibited definite thixotrop- concentrations lower than 55° Brix and viscosity with the available narrow-gap \2A—JOURNAL OF FOOD SCIENCE-Volurrle 35 (1970)

logi M = ------aE + B 2.3RT

where, aE is the activation energy for vis­ cous flow, R is the gas constant, and B is a constant. Activation energies (A E) of various fruit products were calculated from the slopes cf the plots of our experimental W data (see Table 1). For comparison, the ce _ activation energy of water between 2 0

l0.. _i__L_i__I__i__I__i__I__i__I__i__1__i__ depectinized apple juice and Concord to 20 30 4 0 5 0 60 70 80 grape juice are very close to those of su­ TEMPERATURE , " C crose and glucose solutions of the same Fig. 6—Effect of temperature on viscosity of concentration. As a rule, the higher the fru it purees. juice concentration, the higher the activa­ TEMPERATURE, °C tion energy. Fig. 5—Effect of temperature on viscosity of The activation energy of flow (A E) orange juices. has been related to some fundamental pends primarily on the concentration, thermodynamic properties of the New­ measuring system (MS-0) was inconsist­ size and shape of the suspended particles. tonian fluids. For example, A E has been ent, because of the presence of large pulp The contribution of the serum to viscos­ found to be approximately equal to 1 /3 particles which tended to settle and inter­ ity is relatively small, except in concen­ or 1/4 the heat of vaporization, depend­ fere with the rotating inner cylinder. At trated purees. Temperature affects the ing on the shape and bonding of liquid higher concentrations (° Brix), these par­ flow properties of suspended solids only molecules. Empirical equations have been ticles remained in suspension and, since slightly and, therefore, the over-all effect suggested for the estimation of the activa­ the gap of the measuring system was larg­ of temperature is rather small. The effect tion energy as a function of the viscosity er, no serious interference with the rotat­ of temperature becomes more significant and the temperature of various classes of ing cylinder was noticed. when the viscosity of the serum is rela­ liquids (van Wazer et al., 1963). Filtered orange juices of 18 and tively high, as in concentrated purees, or Activation energy decreased signifi­ 10° Brix were prepared by removing the when sugar has been added to the puree. cantly when suspended particles were present in the product, as in cloudy juices suspended pulp with a No. 1 Whatman Activation energies filter paper. Figure 5 shows the effect of and fruit purees. The highest activation When the logarithm of the viscosity of temperature on the viscosity of orange energy (14.2 kcal/g mole) was obtained a Newtonian fluid (log /r) is plotted vs. juice concentrates and diluted filtered with the 75° Brix clear apple juice and the reciprocal of absolute temperature juices. The apparent viscosities of the the lowest ( 1 . 2 kcal/g mole) with the (1/T, where, T is °Kelvin), a straight line apple sauce. In pseudoplastic fruit prod­ concentrates (at 1 0 0 sec'1) decreased only moderately with increasing tempera­ is usually obtained, according to the Ar­ ucts, the activation energy was directly ture, presumably because of the presence rhenius equation proportional to the flow behavior index of considerable amounts of suspended pulp. The filtered orange juices behaved more like clear apple juice and sucrose Table 1 —Activation energies for viscous flow o f fru it juices and purees. solutions and their viscosity decreased Flow behavior Activation energy more rapidly with temperature than did P r o d u c t “ B r ix in d e x , n AE, kcal/g mole that of the unfiltered concentrates. Depectinized apple juice 7 5 1.00 1 4 .2 Fruit purees 5 0 1.00 8 .4 3 0 1.00 6 .3 In contrast to various fruit juices, 15 1.00 5 .3 temperature had a rather small effect on the apparent viscosity of fruit purees C lo u d y a p p le ju ic e 6 5 .5 0 .6 5 9 .1 5 0 0 .8 5 6.1 (Fig. 6 ). The fluid consistency coefficient (K) decreased slightly at higher tempera­ 4 0 1.00 5 .8 3 0 1.00 5 .1 tures, while the flow behavior index (n) 1 0 .5 1.00 3 3 remained practically constant. For all three purees, the value of n was approxi­ C o n c o r d g ra p e ju ic e 6 4 0 . 9 0 11.2 mately equal to 0.30. The apparertt vis­ 5 0 1.00 6 .9 3 0 1.00 6.2 cosities shown in Figure 6 were estimated 15 1.00 5 .3 at a shear rate of 1 0 0 sec'1, and a similar effect of temperature can be expected at O ra n g e ju ic e 4 4 0 .6 5 5 .4 any other shear rate, i.e., rotational speed 3 0 0 .8 5 4 . 2 or flow rate. Apple sauce was thixotropic, Filtered orange juice 18 1.00 5 k due to the larger size of suspended parti­ 10 1.00 5 .3 cles, and its apparent viscosity showed Apple sauce 11.0 0 . 3 0 1.2 the least dependence on temperature. Peach puree 1 1 .7 0 3 0 1 .7 In fruit purees and pulps, viscosity de- P e a r p u re e 1 6 .0 0 .3 0 1 .9 VISCOSITY OF FRUIT JUICES AND PUREES- 1 2 5

(n), i.e., the more pseudoplastic the prod­ cosity of fruit products have direct appli­ through small-diameter tubes at high flow uct, the less the effect of temperature on cation to several food-processing opera­ rates. In film evaporators, high flow rates its apparent viscosity. tions involving fluid flow, heat transfer are desirable for improved heat transfer. Activation energies of the pseudoplas­ and mixing. The effects of temperature With pseudoplastic fluids, low flow rates tic fruit products reported on in Table 1 and shear rate on the apparent viscosity without agitation may result in fouling of were obtained at a constant shear rate of a product are of primary importance the heating surface. (100 sec'1). Different activation energies and depend largely on its composition would be obtained when the effect of and concentration. REFERENCES temperature is considered at a constant The fact that temperature has a strong Contraves AG, Zurich, Switzerland. 1966. shear stress. effect on the viscosity of clear, concen­ Measurement of Rheological Properties. The low activation energies of fruit trated juices can be utilized by perform­ Harper, J.C. 1960. Viscometric behavior in rela­ purees were similar to those obtained ing more efficient heat exchange opera­ tion to evaporation of fruit purees. Food Technol. 14: 557. from the data of Harper et al. (1962) on tions at high temperatures. For example, Harper, J.C. and Lebermann, K.W. 1962. pear puree and Harper et al. (1965) on evaporation of clear fruit purees will be Rheological behavior of pear purees. 1st In­ tern. Congr. Food Sci. and Technol., Lon­ tomato concentrates. The data on tomato more efficient at high temperatures, pro­ don. concentrates were obtained at a shear rate vided the quality of the particular juice is Harper, J.C. and El Sahrigi, A.F. 1965. Visco­ metric behavior of tomato concentrates. J. of 500 sec"1 and the absolute tempera­ not damaged. For the same reasons, higli- Food Sci. 30: 470. ture was expressed in ° Rankine. The re­ tem perature—short-time sterilization is in­ Perry’s Chemical Engineers’ Handbook. 1963. dicated for clear juices. 4th ed.. New York. sulting slope of the lines Saravacos, G.D. 1968. Tube viscometry of fruit a E High shear rates, achieved by agitation purees and juices. Food Technol. 22: 1585. or high flow rate, are effective in reducing Saravacos, G.D. and Moyer, J.C. 1967. Heating R rates of fruit products in an agitated kettle. the apparent viscosity of fruit purees ar.d Food Technol. 21: No. 3A, 54A. called viscosity-temperature coefficient, cloudy concentrated juices. For such van Wazer, J.R., Lyons, J.W., Kim, K.Y. and was equal to 1,710. The corresponding Colwell, R.E. 1963. “ Viscosity and Flow products, agitated heat exchanges, for ex­ Measurement.” Interscience Publishers, New activation energy (AE) in metric units is ample, wiped film evaporators (Harper, Y o rk . 1.9 kcal/g mole, which agrees closely with 1960) and scraped vessels (Saravacos et Ms. received 6/2/69; revised 6/25/69; accepted 7 /7 /6 9 . the values of Table 1 for the fruit purees. al., 1967), are very efficient. Improved Processing considerations Journal Paper No. 1713 of the New York heat transfer can also be achieved when State Agricultural Experiment Station, Cornell The results and conclusions on the vis­ pseudoplastic products are pumped University, Geneva, New York 14456. MYRON SOLBERG and BARBARA ELKIND Department of Food Science Rutgers University, The State University, New Brunswick, New Jersey 08903

EFFECT OF PROCESSING AND STORAGE CONDITIONS ON THE MICROFLORA OF CLOSTRIDIUM PERFRINGENS—INOCULATED FRANKFURTERS

SUMMARY—The growth support potential of frankfurters for Clostridium perfringens was deter­ for all platings o f Clostridium perfringens. mined. Inoculated emulsions were stuffed into glass tubes and processed so that they reached an Frankfurters were prepared using 6 lb of internal temperature o f 68—69°C in 30—48 min. The tubes were sealed with agar plugs and rubber frozen boneless cow meat, 4 lb of pork jowls, 1 stoppers to simulate vacuum-packaging conditions. Total aerobic and anaerobic bacterial counts in lb chilled water, %-lb salt, 34 g spice mix uninoculated raw emulsion were 1 to 3 log cycles higher when petri dishes were incubated at 23°C (BHSS-WP Presco Foods, Flemington, N.J.) and as opposed to 37°C. The low-temperature-growing mesophiles and psychrotrophs were destroyed 10 ml of cure solution made by dilution of 7 g of sodium nitrite in 100 ml of sterile distilled during processing. The 37°C counts were reduced 1 log cycle and the 23°C counts, 2 to 3 log water. Cow meat, water, salt, spice mix and cycles. Clostridium perfringens reproduced rapidly in frankfurters at 3 7 and 23°C, more slowly at cure were mixed together and chopped in a si­ 15 and 12°C, and not at all during 2 wk of storage at 10° or 4 wk at 0—5°C. Clostridium lent cutter (Hobart Model T215 GA) for 7 min. perfringens made up 2.5—10% of the total anaerobic count at 0 time but became the dominant Pork was added followed by an additional 5- organism during the storage at all temperatures perm itting growth. min chopping. Inoculum from the stored vege­ tative cell and spore suspension was added in a INTRODUCTION The effect of reduced process time on carrier of 100 ml of water and chopping contin­ the microbial population of frankfurters ued to a final emulsion temperature of 13°C. CLOSTRIDIUM perfringens was responsi­ has not been reported in the scientific lit­ All of the frankfurter studies were inoculated ble for more food poisoning cases than erature. Conventional processing cycles with Clostridium perfringens strain 1362. The any microorganism with the exception of requiring 60—120 min produced frank­ emulsion was then spread in a 0 .5 -1-in. layer in Salmonella in 1968 (Anon., 1969). Sever­ a tray, covered with Saran, and stored at 5°C furters of very low microbial counts and al food poisoning strains have been isolat­ 1-2 hr. The emulsion was then transferred to frankfurter spoilage by microorganisms ed from cured meat products. Strain 1,000 ml tall form beakers which were kept in has usually been the result of post-pro­ 1362 was isolated from salami and strain an ice bath. A 26-mm (i.d.) glass tube was then cessing contamination during peeling and S—88 from cold cuts (Hall et al., 1963). filled with emulsion by drawing approximately packaging operations. With the introduction and rapid increase 50 g of meat up to the tube from the beaker with a syringe. The tubes were then plugged at of vacuum-packaged meat products in the This study was designed to evaluate the bottom with a Saran-coated rubber plug market, the shelf life of many items has the effect of 30—48 min heating cycles and the top covered with a plastic foam plug. been extended considerably and the pre­ upon the total microbiological population Filled tubes were placed in a water bath at viously available indicators of product de­ of frankfurters and to determine the growth characteristics of Clostridium per­ 53°C and the bath adjusted to increase in tem­ terioration, mold growth and discolora­ perature at rates varying 0.3-0.6°C per minute. f r in g e n s in a cured meat environment. tion are no longer valid. The potential Frankfurters were processed until they reached public health hazard may have been in­ EXPERIMENTAL an internal temperature of 68-69°C, after creased by the new packaging procedure SPORES o f Clostridium perfringens (strain which they were rapidly cooled in tap water at in which anaerobes or facultative anaer­ 1362-heat sensitive and S-80-heat resistant, approximately 10°C until an internal tempera­ obes become the predominant microor­ obtained from Dr. H. E. Hall, Taft Sanitary En­ ture of 21°C was reached. ganisms in the microflora (Shank et ah, gineering Center, Ohio) were harvested follow­ The frankfurters, in tubes, were then cov­ 1963; Ogilvy et ah, 1953; Allen et ah, ing the procedure described by Duncan et al. ered with 1 in. of 1.5% agar (Difco) and placed in storage at various temperatures. Sample stor­ 1 9 6 0 ). Clostridium perfringens h as a re ­ (1968). Washed spores and vegetative cells were age was at 4, 10, 12, 15, 23 and 37°C. Samples ported minimal growth temperature of stored in distilled water at 4-5°C. were drawn regularly for microbial analysis. 15°C (Breed et ah, 1957). Therefore, vac­ Heat shocking of mixed vegetative cell and spore suspensions was carried out at 80°C for uum-packaged cured meat products Total aerobic and total anaerobic plate 20 min followed by rapid cooling in an iced would have to be poorly handled to result counts were obtained using Tryptone Glucose water bath. Extract (TGE) Agar. Anaerobic conditions were in establishment of a Clostridium perfrin­ SPS agar (Angelotti et al., 1962) was used obtained in Case Anaerobic Jars evacuated and g e n s food poisoning condition, since re­ frigeration equipment is expected to op­ erate between 5 and 10°C. Many home refrigerators are operated at temperatures Table 1 —Stability o f Clostridium perfringens spores and vegetative approaching 15°C. Vacuum-packaged cells during storage in distilled water at 5°C. luncheon meats left out of refrigeration for several hours or even days may show Storage time Vegetative cells S p o r e s no color change even after significant S tr a in ( d a y s ) (organisms/ml) (organisms/ml) microbial growth has occurred. 1 3 6 2 7 1 . 3 X 1 0 8 3 . 7 X 1 0 6

Frankfurters are typical of many 1 8 1 . 6 X 1 0 8 7 . 1 X 1 0 6 luncheon meats, since they are cooked to 6 2 8 . 2 X 1 0 7 — a ready-to-eat form. They are often man­ 1 1 9 7 . 4 X 1 0 1 — ufactured using meat mixtures and spice 1 6 9 8 . 4 X 1 0 7 4 . 5 X 1 0 6 mixtures identical to those used in bolo­ gna manufacture. Modern frankfurter S — 8 0 3 2 . 6 X 1 0 7 5 . 5 X 1 0 4 manufacturing procedures have shortened 7 1 . 3 X 1 0 " 4 . 7 X 1 0 4 the heating cycle to 30 to 50 min fol­ 1 3 1 . 5 X 1 0 " 4 . 9 X 1 0 4 lowed by rapid cooling. 8 3 3 . 0 X 1 0 4 3 . 3 X 1 0 4

126-JOURNAL OF FOOD SCIENCE-Volume 35 (1970) CLOSTRIDIUM PE RF RI NGENS—INOCULA TED FRANKFURTERS- 1 2 7

85 - 80 -

O — O — O wear Temp( internal • — • — • water bath temp.

_L J -----1------1___ I__I___ i____ 1___ L- _1. I I I I 5 10 15 20 25 30 35 40 45 50 55 60 65 70 Time CmmutesD

Fig. 2 —Heating and cooling profile for rapid- process frankfurters prepared under experimen­ ta l c on ditions.

Fig. 1—Total plate counts for raw frankfurter emulsion evaluated in Tryptone Glucose Extract Agar incubated at 37 and 23°C.

flushed with nitrogen 4 times. Petri dish incuba­ profile used in these experiments is pre­ tion was at 37 or 23 °C. sented (Fig. 2). Such a process profile is Frankfurters being sampled were pushed commercially attainable using continu­ from the glass tubes into a stainless steel Waring ous-processing equipment available to the Blendor cup and 200 ml of 0.1% peptone water added. The mixture was blended at high speed industry today. The formulation of the for 3 min. Serial dilutions were prepared and frankfurters prepared was designed for a appropriate petri dishes poured using TGE minimal shrink such as would be expect­ Agar or SPS-Agar. ed from a process within a glass tube. The Petri dishes incubated at 37°C were counted total water added was approximately 10% after 36-48 hr. Petri dishes incubated at 23°C of the meat. There was some slight fat were counted after 60-72 hr. separation in some of the experiments. Aseptic technique was used throughout the The frankfurters reached an internal tem­ experiments with all glassware, utensils, media perature of 49°C after 10—15 min and and diluents sterilized and all equipment thor­ then increased from 0.5 1.0°C per min­ oughly sanitized before use. ute until an internal temperature of 68 ^ RESULTS & DISCUSSION was reached. The frankfurters were then Fig. 3 —Survival o f aerobes and anaerobes in frankfurters processed to reach an internal tem­ rapidly cooled to 21 °C during a THE STABILITY of harvested spores and perature of 68°C at various rates. Total plate the accompanying vegetative cells was 15—20-min period. One experiment was counts were made in Tryptone Glucose Extract evaluated over long periods of time. Re­ terminated at an internal temperature of Agar incubated at 37 or 23°C. sults shown in Table 1 indicated that 6 5 .5 °C . spore counts remained essentially con­ The effect of the heating rate upon stant throughout extended storage at re­ survival of total aerobes and anaerobes is frigerated temperature. Vegetative cells presented in Figure 3; data for survival of appeared to decrease during the first 1 or Clostridium perfringens spores and vege­ the resulting destructive effect being 2 wk of storage after which they too re­ tative cells are in Table 2. The shonest g reater. mained constant throughout several process cycle, which raised the frankfurt­ The continually changing slopes evi­ m o n th s. ers to 68°C in 30 min, destroyed only dent in Figure 2 may be due to different The raw emulsions were sampled prior 10% of the anaerobic organisms and only microorganisms within the mixed flora to inoculation; data obtained for total 43% of the aerobic organisms growing at being inactivated at different time-tem­ aerobic and total anaerobic plate counts 37°C. The destruction of the more perature relationships. It appears that the at 23 and 37°C are shown (Fig. 1). Total psychrotrophic organisms which grew at survival was diminishing rapidly in the aerobic counts were 15 to 65 times great­ 23°C was 99.2% of the anaerobes and post-45-min period, indicating that con­ er on plates incubated at 23°C; total 95.6% of the aerobes. siderable advantage would be gained anaerobic counts were 40 to 275 times The organisms which grew at 37°C either from continued heating to higher greater at 23°C. were much more heat resistant than those temperatures or short holding periods at The higher aerobic and anaerobic total which grew at 23°C. The effect of de­ 68—69°C before cooling. The experimen­ plate counts at 23°C may be due to the creasing the rate of heating was to signifi­ tal data from the frankfurters processed facultative pseudomonad population of cantly reduce the total plate counts. The to 65.5°C internal showed slightly higher fresh meat. Many pseudomonads will outer surfaces of the frankfurter were ex­ counts than an equal time process to grow at 23 but not at 37°C. posed to heat treatment for longer times 6 8 ° C. These points are enclosed in A typical frankfurter heat-processing during the slower heating processes, with dashed squares in Figure 3. Ì28—JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

Table 2 —Effect o f frankfurter processing rate to a final internai temperature o f 68°C upon survival and recovery o f spores and vegetative cells o f Clostridium perfringens, strain 1362.

Post-processing Raw emulsion Inoculum recovery Inoculum recovery Process time (spores/g) (organisms/g) (vegetative cells/g) (organisms/g)

30 2.8 x 102 4.6 x 103 3.3 x 103 6.6 x 103 35 2.8 x 102 1.9 x 103 8.0 x 103 5.0 x 103 45 3.0 x 102 1.6 x 103 4.9 x 103 4.5 x 103 48 3.0 x 102 1.7 x 103 5.5 x 103 7.0 x 103

The vegetative cells of Clorstridium the spread between 23 and 37°C plate perfringens added to the emulsion were counts to remain noticeable, whereas the Tim« (0 » y ü recoverable after mixing as demonstrated short-time, normal-temperature process Fig. 5 —Exponential growth phase for Clos­ in Table 2. The process time affected the (Expt. 4, Fig. 4) reduced the psychro- tridium perfringens in frankfurters stored at var­ Clostridium perfringens recovery level troph-mesophile differential. The C lo s ­ ious temperatures. Plate counts were made in tridium perfringens plate counts were with the longer process times resulting in SPS Agar incubated at 37°C. greater destruction. The post-processing equivalent when the SPS Agar plates were survival of vegetative cells was of such incubated at 37°C for 24 hr or 23°C for magnitude that the spore inoculum repre­ 72 hr. The stationary phase data present­ ed in Figure 6 offered additional support sented less than 20% of the total cells. p h ase o f Clostridium perfringens o rg a ­ to this finding, since the 23 and 37°C Therefore, no statement can be made re­ nisms is illustrated in Figure 5. The growth storage temperatures yielded equivalent lating to the efficiency of the heating rate diminished as the temperature was concentration maxima in frankfurters. process as a spore-germinating mecha­ reduced from 37 to 12°C. At 10 and 5°C nism . It is also apparent from Figure 4 that no growth of Clostridium perfringens w as The similarity of total plate counts at the total number of anaerobes present im­ observed after 14 days’ storage, although 37 and 23°C indicated in Figure 4 further mediately after processing was generally the aerobic counts at 10°C had increased demonstrated the susceptibility of the greater than the number of Clostridium approximately 1 log cycle after 14 days psychrotrophic organisms to the heating perfringens. and the total aerobic count at 4°C had process. The lower-temperature, normal­ The effect of frankfurter storage tem­ increased almost 2 log cycles at the end time treatment (Expt. 1, Fig. 4) allowed perature upon the exponential growth of 21 days’ storage. Growth of C l o s t r i ­ dium perfringens at 12°C has not been reported in the literature. Many house­ hold refrigerators and supermarket refrig­ erated cases operate at temperatures of approximately 12°C although they should be kept below 10°C. Short-term exposures to room temperature could result in sig­ Cpe r f r ingens nificant increases in Clostridium perfrin­ □ grown at 37G g e n s populations which might not be de­ ra C.perfringens grown at 23C tectable on a basis of gas production in Total anaerobes the semirigid type of vacuum package grown at 37C now used for cured meat products. The BR Total anaerobes grown at 23 C growth rate at poor refrigeration tempera­ Total aerobes tures as represented by 15 and 12°C grown at 37C was quite rapid with counts increasing ■ Total aerobes from 103 to 106 in 3 days at 15°C and grown at 23C ■ in 5 days at 12°C. Subsequent growth proceeded at a slower rate until a maxi­ mum cell concentration was reached be­ tween 7 and 14 days. Figure 6 shows the maximum cell concentration reached by Clostridium perfringens, at the various frankfurter storage temperatures. No ex­ planation can be offered for the low max­ imum count at 15°C. The result was re­ produced in 4 different experiments. The maximum cell level reached in the stationary phase by Clostridium perfrin­ g e n s as compared to that attained by total anaerobes is illustrated in Figure 7. In the Experiment Number stationary growth phase there was no ap­ parent difference between total C l o s t r i d ­ Fig. 4 — The post-processing microbial contamination o f frankfurters as a function o f process time and temperature. Process conditions: Expt. ium perfringens and total anaerobes at 1: 66.5°C internai, 45 min. Expt. 2: 68°C internal, 48 min. Expt. 3: any of the storage temperatures or at any 66.5° internai, 47 min. Expt. 4: 68°C internai, 30 min. Clostridium of the petri dish incubation temperatures. perfringens cultivated on SPS Agar. Total plate counts on TGE Agar. A lth o u g h Clostridium perfringens m a d e CLOSTRIDIUM PERFRINGENS—INOCULA TED FRANKFURTERS-^

I I C.perfringens grown at 37C R53 c.perfringens 1X3 grown at 23C Total anaerobes grown at 37C ¡S3 Total anaerobes grown at 23C

Readings Sample incubation temperature *C 2-5 2-5 14-28 14-28 Time SpanCdayd Fig. 7 —A comparison o f the vaible cell population o f Clostridium per­ Fig. 6 —Maximum cell concentration reached and maintained by Clos­ fringens on SPS Agar with the total anaerobic plate count on TGE Agar tridium perfringens in frankfurters stored at various temperatures. during the stationary growth phase. (2—5 days at 37 or 23°C; 14—28 days at 15 or 12°C.) up 10% or less of the total anaerobic pop­ ulation immediately after processing in 3 furter storage at temperatures 12—37°C ing refrigerated storage. Food Res. 25: of the 4 experiments plotted, comparison 1 9 -2 5 . s u p p o r t e d Clostridium perfringens of Figure 7 with Figure 4 indicates that Angelotti, R., Hall, H.E., Foter, M.J. and Lewis, growth. Therefore, cured meat products K.H. 1962. Quantitation of Clostridium per­ the large differences disappeared during which may be represented by frankfurters fringens in foods. Appl. Microbiol. 10: storage at all temperatures permitting the 193—199. will not serve as a vector for the transmis­ Anonymous. 1969. Morbidity and Mortality g ro w th o f Clostridium perfringens. T h e re ­ sio n o f Clostridium perfringens fo o d p o i­ Weekly Report, 18 (12): 98. fo re, Clostridium perfringens m u st have Breed, R.S., Murray, E.G.D. and Smith, N.R. soning if stored at temperatures below become the dominant anaerobe in the 1957. “ Bergey’s Manual of Determinative 10°C, but could become a vector as a re­ Bacteriology.” 7th ed. The Williams and system. Further proof was gained by Wilkins Company, Baltimore. sult of storage at temperatures as low as overlaying the TGE agar with SPS agar Duncan, C.L. and Strong, D.H. 1968. Improved 12°C . medium for sporulation of Clostridium per­ after counting the colonies. The overlaid fringens. Appl. Microbiol. 16: 82—89. became the plates were reincubated for 24 hr and Clostridium perfringens Hall, H.E., Angelotti, R., Lewis, K.H. and dominant anaerobe in processed frank­ Foter, M.J. 1963. Characteristics of Clos­ black colonies were counted; all of the tridium perfringens strains associated with furters stored between 12 and 37°C even colonies were black. food and food-borne disease. J. Bacteriol. when the initial post-processing popula­ 85: 1094—1103. CONCLUSIONS Ogilvy, W.S. and Ayres, J.C. 1953. Post-mortem tion represented only 2.5% of the total changes in stored meats. V. Effect of carbon THE SURVIVAL of microorganisms in an aero b es. dioxide on microbial growth on stored frankfurters and characteristics of some frankfurters subjected to a simulated Therefore, any cooked, cured, com­ microorganisms isolated from them. Food high-speed continuous process was signifi­ minuted meat product of the frankfurter Res. 18: 121-130. cant. The development of such systems Shank, J.L. and Lundquist, B.R. 1963. The ef­ type subjected to improper storage condi­ fect of packaging conditions on the bacteri­ must be accompanied by microbiological tions could become a vector of Clostrid­ ology, color, and flavor of table-ready meats. Food Technol. 17: 1163—1166. studies to establish the necessary time- ium perfringens food poisoning if con­ temperature relationships to assure shelf Ms. received 6/2/69; revised 9/8/69; accepted taminated with a significant number of 9 /1 0 /6 9 . life and public health safety. organisms before Clostridium perfringens A paper of the Journal Series, New Jersey A heat-sensitive strain of Clostridium or during processing. Agricultural Experiment Station, Rutgers Uni­ perfringens inoculated into frankfurter versity, The State University, Food Science De­ emulsion at concentrations of 3,000 to REFERENCES partment, New Brunswick, New Jersey. This research was supported in part by a 8,000 organisms per gram survived pro­ Allen, J.R. and Foster, E.M. 1960. Spoilage of grant from the United States Public Health cessing to the extent of 20—50%. Frank­ vacuum packed sliced processed meats dur- Service UI 00312. R. R. EITENMILLER, J. R. VAKIL and K. M. SHAHANI Department of Food Science and Technology University o f Nebraska, Lincoln, Nebraska 68503

PRODUCTION AND PROPERTIES OF PENCILLIUM ROQUEFORTI LIPASE

S U M M A R Y —A Penicillium roqueforti strain produced maximal amounts o f lipase when grown in ration. Precipitates obtained after 30 min at the 0.5% casitone—1% Proflo broth, pH 5.5, at 27°C. Addition o f butteroil, corn oil or olive oil to the various saturation levels were collected by cen­ trifugation at 17,000 x g for 30 min. The pre­ growth medium inhibited the lipase production. Under pH stasis the partially purified lipase o f P. cipitates were dissolved in minimum cold dis­ roqueforti had an optimum pH of 8.0 and an optimum temperature of 37°C. Maximum lipolytic tilled water and dialyzed for 12 hr at 4°C activity occurred with 5% butteroil emulsion as the substrate. Manganese chloride and magnesium against 3 liters of distilled water charged twice chloride stimulated the enzyme activity. Calcium, sodium and potassium salts had no appreciable during dialysis. The various fractions thus ob­ effect on lipolysis; silver, mercury and zinc salts were inhibitory. The lipase was thermolabile, being tained were assayed immediately after dialysis inactivated completely within 10 min at 50°C. The lipase hydrolyzed tributyrin, tricaprylin, tricap- and the active fractions frozen. rin, tripropionin and triolein in decreasing order. Protein determination INTRODUCTION of the mold, the organism was grown by sta­ Protein concentration in the growth me­ LIPASES have been studied extensively tionary and shake culture techniques. When sta­ dium and in the (NH4 )2 S0 4 fractions was de­ tionary culture was used, 100—ml aliquots of in the dairy and food industry because termined by the Folin Ciocalteau reagent meth­ broth in 500-m l Erlenmeyer flasks were inoc­ they are responsible for the development od o f Lowry et al. (1951). ulated with 1 loopful of spores collected by of hydrolytic rancidity in various food scraping the surface of a Czapek agar slant. RESULTS & DISCUSSION products. Even though lipase action con­ When shake culture was used, the broth was Factors affecting lipase production by stitutes a serious economic problem of inoculated with pregerminated spore inoculum. Penicillium roqueforti the food industry, many lipases play a Pregerminated inoculum was prepared by inoc­ Composition of the medium. Initially, major role in developing characteristic fla­ ulating 50 ml of broth in 125-ml Erlenmeyer the mold was grown as a stationary cul­ vors in ripened cheeses. In this regard, the flasks with 1 ml of a previously standardized ture in several media to study the ability spore suspension containing approximately 2.0 lip ase o f Penicillium roqueforti is o f of each to support lipase production. In x 107 spores per milliliter in 0.01% polyvinyl particular importance as it is largely re­ Czapek broth and a Czapek modified sponsible for the development of the alcohol. The broth was incubated at 27°C for 24 hr on a New Brunswick rotary shaker set at broth, which contained 1% lactose in characteristic flavor of Blue cheese. In ad­ 170 rpm. For lipase production studies, 5—ml place of saccharose as the carbon source, dition, lipases from various sources are aliquots of the pregerminated inoculum were the mold did not produce any significant being used to develop, through controlled transferred aseptically to 55 ml of broth in amounts of lipase. These results agree lipolysis, partially lipolyzed products for 125-ml Erlenmeyer flasks. with the work of Thibodeau et al. (1942), use in the food industry as flavor modi­ Lipase assay who concluded that NaN03, the only ni­ fiers and enhancers. Two lipase assay methods were used. Lipase trogen source in Czapek medium, was a The objectives of the present study on assay of the crude broth was done by the silica very poor nitrogen source for growth and the lipase of P. roqueforti were to study gel chromatographic method of Harper et al. enzyme production by P. roqueforti. E s­ factors that affect lipase production by (1956). The enzyme assay of the partially puri­ sentially similar observations were made the organism, to partially purify the lip­ fied lipase obtained by (NHa^SOa fractiona­ by Morris et al. (1953) who fortified tion of the broth was carried out by the pH-stat ase and to study factors that affect the Czapek’s solution with peptone and but- technique (Parry et al., 1966; San Clemente et activity of the enzyme. It was believed terfat for lipase production by P. r o q u e ­ al., 1967). that such a study might provide informa­ f o r t i and by Alford et al. (1964) who Partial purification o f the lipase tion to control undesirable rancidity in used case peptone and yeast extract for dairy and food products. Also, comple­ After incubation, the broth was filtered the same purpose. tion of the objectives would appear to be through several layers of cheese cloth to remove Other media tested in this study were a logical step toward adapting the lipase the mycelia, then through Whatman No. 1 filter 3% casitone broth, 1% Proflo broth and o f P. roqueforti to the production of paper with the addition of celite filter-aid to 2% casitone—1% Proflo combination modified fat preparations for commercial remove any extraneous material. Solid (NH4 )2 SC>4 was added slowly with stirring to broth. Casitone is hydrolyzed casein and use. the clear filtrate at 4°C to give the desired satu­ Proflo is a high protein, partially defatted MATERIALS & METHODS Cultures The P. roqueforti strain used throughout this study was isolated from a commercial sam­ Table 1—Effect o f initial pH o f the medium upon lipase production. ple of Blue cheese. The organism was main­ tained with periodic transfers on Czapek solu­ mg Mycelium Lipase activity tion agar slants. pH Lipase activity1 ml Broth per mg mycelium Media 5.5 85.8 1.6 54 The following commercial media were used: 6.0 78.8 1.5 53 Proflo (Traders Oil Mill Co.), Bacto-Casitone 6.5 74.6 1.4 53 (Difco), Czapek Solution Agar (Difco) and 7.0 58.2 1.1 53 Czapek Dox Broth (Difco). All media except 7.5 54.6 1.0 those containing Proflo were autoclaved at 15 55 psi for 15 min. Proflo media were autoclaved at 8.0 45.6 0.9 51 15 psi for 20 min. 8.5 44.6 0.9 50 Culture methods 1 Lipase activity is expressed as pmoles of free fatty acids liberated To examine the lipase-producing capabilities in 1 hr by 1 ml of broth.

130-JOURNAL OF FOOD SCIENCE-Volume 35 <19701 PENICILLIUM ROQUEFORT! LIPASE-W

80

0.5 1.0 1.5 2.0 INCUBATION (DAYS) PROFLO (%) Fig. 3.—Effect o f incubation time upon lipase Fig. 1.—Lipase production in casitone or Profio Fig. 2.—Lipase production in casitone: Proflo production in stationary culture. Lipase activity broth. Lipase activity is expressed as pmo/es o f combination broth. Lipase activity is expressed is expressed as given in Figure 1. free fatty acids liberated from the substrate in 1 as given in Figure 1. hr by 1 ml of the broth as measured by the silica gel m ethod. tion temperature upon the lipase produc­ tion in stationary culture revealed that a broad temperature optimum for lipase Addition of oils to the medium. To production exists between 21 and 27°C study the effect of oils in the medium (Table 2). At 16 and 30°C, the organism cottonseed flour. To each of these media upon lipase production by P. roqueforti, produced the lowest amounts of lipase. were added 0.1% Na2HP04 and 0.01% 1% (v/v) sterilized butteroil, corn oil or At 16°C, while only a small amount of KC1, and the pH was adjusted to 6.8. Af­ olive oil was added to aliquots of the growth occurred, a larger amount of lip­ ter 7 days of incubation at 27°C, high casitone-Proflo broth. After inoculation ase was produced per milligram dry lipolytic activity was noted in the 3 me­ and incubation at 27°C for 7 days in sta­ weight mycelium. In contrast, at 30°C, dia, with the highest lipase production tionary culture, lipase activity in the while heavy growth occurred, only a being in the casitone-Proflo combination broths was determined. All of the oils small amount of lipase was produced per broth. These results indicated that P. added to the medium caused a decrease in unit weight. r o q u e f o r t i requires a complex organic ni­ lipase production. Olive oil inhibited lip­ Effect of incubation time. Studies con­ trogen source for high lipase production. ase production by only 5%, but butteroil cerning the lipase production with respect Following the initial experiments, lip­ and corn oil inhibited lipase production to incubation time when the mold was ase production was determined after 7 of the mold by 21 and 29%, respectively. grown by stationary culture in casitone- days’ incubation at 27 °C in media con­ Initial pH of the growth medium and Proflo combination broth, pH 6.8, re­ taining varying concentrations of casitone incubation temperature. Table 1 presents vealed that the maximum amount of lip­ and Proflo. As shown in Figure 1, in the data concerning the effect of initial pH of ase was produced at the end of 7 days, 0.5, 1,2 and 3% casitone broth, the lipase the growth medium, ranging between pH following which the lipase activity tended production was nearly the same. In the 5.5 and 8.5, upon the lipase production to decrease with time (Fig. 3). Proflo broth there was a sharp increase in in stationary culture. Values below pH Figure 4 shows lipase production by P. lipase production as the Proflo concentra­ 5.5 were not checked, as the medium was tion was increased from 0.25 to 1%. not stable at lower pH levels. Highest lip­ Above 1.5% Proflo, the enzyme produc­ ase production occurred at pH 5.5 and, as 160 tion seemed to level off. the pH was increased the lipase produc­ Figure 2 presents the lipase production tion decreased. It was also observed that in 0.5, 1 and 2% casitone broth in com­ decreased lipase production at higher pH bination with varying concentrations of levels closely followed decreased growth Proflo. The lipase production was highest of the mold. At all pH levels, the produc­ in 0.5% casitone—1% Proflo combination tion of lipase per milligram dry weight broth. This medium, therefore, was used mycelium remained nearly the same. for all the additional studies. Data concerning the effect of incuba­

Table 2—Effect o f incubation temperature upon Upase production. Temp mg Mycelium Lipase activity °C Lipase activity1 ml Broth per mg mycelium 16 16.6 0.26 65 21 66.2 1.3 51 25 66.8 1.4 48 INCUBATION (hr) 27 68.4 1.3 53 24 30 31.0 1.3 Fig. 4.— Effect o f incubation time upon lipase 1 Lipase activity is expressed as Mmoles of free fatty acids liberated production in shake culture. Lipase activity is in 1 hr by 1 ml of broth. expressed as given in Figure 1. 122-JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

mal lipolysis point. A similar observation was made by Khan et al. (1967) with 40 - Achromobacter lipolyticum lipase which showed decreased activity at concentra­ tions above 10% with synthetic glycer­ ides. Time of reaction. Figure 6 shows lip­ ase activity in terms of chart units taken directly from the pH-stat recorder as a function of reaction time. One chart unit equals 0.05 ¿zmole of fatty acid. The reac­ tion rate remained linear for approxi­ mately 5 to 6 min, after which it tended to decrease slightly. San Clemente et al. (1967) reported that the reaction rate of a partially purified Staphylococcal lipase 5 ¡0 ¡5 SUBSTRATE CONCENTRATION (% ) under pH stasis remained linear for 6—8 min and that lipase assay methods which Fig. 5.—Effect o f substrate concentration upon utilize long incubation periods give values lipase activity. Lipase activity is expressed as Fig. 6.—Effect of time upon reaction rate. 1 on the low side of the initial rate. nmoles o f free fatty acids liberated from the c h a rt u n it = 0.05 pm ole o f free fatty acid. substrate in 1 min by 1 ml of the enzyme as Heat stability. Because the suscepti­ measured by the pH-stat method. bility of enzymes to heat treatments is significant to the food industry, the sus­ ceptibility of the lipase to heat treat­ ers may well be due to the use of differ­ ments was determined in an aqueous solu­ r o q u e f o r t i grown by shake culture in ent substrates, different assay conditions tion at pH 6.5. At 37°C, the enzyme lost casitone-Proflo combination broth at pH and different strains of P. roqueforti. 65% of its activity in 60 min, and at 5.5 inoculated with pregerminated spores. Although the lipase demonstrated high 40°C, 90% of its activity (Fig. 7). At The shake culture method produced max­ activity over a wide range, 25—45°C, it 50°C, the lipase was inactivated com­ imal amounts of lipase in 48 hr, yielding gave maximal lipolysis at 37 °C, indicating pletely within 10 min. However, the en­ approximately 128% more lipase than its optimum temperature to be 37°C. zyme appeared to be fairly stable at low when the mold was grown by stationary Substrate concentration. Using but­ temperatures and could be held frozen culture at pH 6.8. teroil and olive oil as the substrate, the for long periods without loss of activity. In subsequent work, all lipase pro­ effect of the substrate concentration The inactivation temperature of 50°C duced was obtained by inoculating 0.5% upon the lipase activity was studied. In reported here compares well with inacti­ casitone—1% Proflo broth, pH 5.5, with general, butteroil was hydrolyzed more vation temperatures reported for other pregerminated inoculum and incubating rapidly than olive oil at all the concentra­ fungal lipases. Fukumoto et al. (1963; at 27°C on a New Brunswick rotary shak­ tions studied (Fig. 5). With butteroil as 1964) reported that the lipases of A s p e r ­ er operating at 170 rpm. the substrate, lipolysis was maximum at gillus tiiger a n d Rhizopus delemar w ere Partial purification of the lipase the 5% substrate concentration, whereas inactivated at 55 and 50°C, respectively. For characterization of the lipase, a with olive oil, maximal hydrolysis oc­ These lipases resemble milk and pancre­ partially purified preparation was ob­ curred at the 3% concentration. With atic lipases which are heat sensitive but tained by celite-aided filtration of the both substrates, lipolytic activity de­ differ markedly from some microbial lip­ broth followed by (NH4)2S04 fractiona­ creased at concentrations above the maxi­ ases such as Aspergillus lipolyticum lip - tion of the filtrate. The lipase was precipi­ tated out over a broad range of ammo­ nium sulfate saturation. However, a fraction with highest specific activity was Table 3—Effect o f salts on Upase activity. obtained between 30 and 50% saturation. Stimulation On a protein basis, this partially purified or fraction had increased specific activity of Concentration Relative activity (inhibition) about sevenfold over that of the original Salt (M) (%) (%) broth. This fraction was used in the fol­ Control - 100 - lowing characterization studies. MnCl2 10'2 135 35

Factors affecting lipase activity 10'3 100 ~ pH and temperature. The effect of pH MgCl2 10'2 112 12 upon lipolysis of butteroil substrate was 10’3 100 - determined with the pH-stat method be­ NaCl 10'2 103 3 tw e e n pH 6.5 a n d 10. T h e lipase "was 10’3 99 1 quite active over a pH range of 7.5—9.0 , KC1 10’2 102 2 with an optimum being at pH 8.0. This 10'3 99 1 value agrees with the optimum pH value CaCl2 10’2 102 2 o f P. roqueforti lipase reported by Fodor 10'3 101 1 et al. (1949), but disagrees with the opti­ ZnSC>4 10’4 65 (35) mum pH values o f5.0—5.5 reported by HgCl2 10"4 31 (69) Shipe (1951) and 6.5—6.8 reported by AgNC>3 10’4 24 (76) Morris et al. (1963). It is possible that the CoCl2 10’4 101 1 differences in the pH optimum observed FeS04 10’4 102 2 in this study and in those of other work­ C uS04 10’4 98 2 PENICILLIUM ROQUEFORT! LIPASE- 1 3 3

ase, which is very heat stable (Khan et al., Table 4—Effect of lipase upon various tri- 1967). glycerides. Lipase Relative Effect of salts. Since various salts gen­ 2 Triglyceride activity1 hydrolysis erally present in most food products have Tripropionin 9.0 32.2 been shown to affect lipase activity mark­ Tributyrin 28.0 100.0 edly, the effect of several salts on the ac­ Tricaprylin 26.3 93.2 tivity of the Penicillium lipase was stud­ Tricaprin 17.3 61.8 ied (Table 3). Rate of hydrolysis of Triolein 5.7 20.0 butteroil substrate was stimulated by MnCl2 and MgCl2. However, calcium, lipase activity is expressed as mmoles of sodium or potassium salts did not seem to free fatty acids liberated from the substrate in have any appreciable effect. This observa­ 1 min by 1 ml of enzyme. tion differs from that of Shipe (1951), 2 Hydrolysis of various triglycerides is re­ who reported that CaCl2 accelerated the corded on a relative basis, considering 100% for activity of A. niger and P. roqueforti lip ­ tributyrin. ases. Also, calcium salts have been shown to accelerate the activity of pancreatic lipases as well as that of several microbial lipases (Constantin et al., 1960; Iwai et Fig. 7.—Heat inactivation of the Upase. ties of the extracellular lipase of Achro- mobacter lipolyticum. Biochim. Biophys. al., 1964; Wills, 1961). The P. roqueforti Acta 132: 68. lipase was inhibited by silver, mercury REFERENCES Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. 1951. Protein measurement and zinc salts, but not by cobalt, iron and Alford, J.A. and Pierce, D.A. 1961. Lipolytic with the Folin phenol reagent. J. Biol. copper salts. Iwai (1965) found ferrous, activity of microorganisms at low and inter­ Chem. 193: 265. ferric, silver, mercury, chromium, copper mediate temperatures. III. Activity of mi­ Morris, H.A. and Jezeski, J.J. 1953. Actions of crobial lipases at temperatures below 0°C. J. microorganisms on fats. II. Some character­ and tin ions to be inhibitory toward the Food Sci. 26: 518. istics of the lipase system of Penicillium action of crystalline A . niger lipase. These Alford, J.A., Pierce, D.A. and Suggs, F.G. 1964. roqueforti. J. Dairy Sci. 36: 1285. Activity of microbial lipases on natural fats Parry, R.M., Chandan, R.C. and Shahani, K. M. results show that the Penicillium lipase is and synthetic triglycerides. J. Lipid Res. 5: 1966. Rapid and sensitive assay for m ilk lip­ affected by salts differently from other 390. ase. J. D a iry Sci. 49: 356. Constantin, M.J., Pasero. L. and Desnuelle, P. San Clemente, C.L. and Vadehra, D.V. 1967. lipases. 1960. Quelques Remarques complementaires Instrumental assay of microbial lipase at sur 1’ Hydrolyse des Triglycerides par la Lip­ constant pH. AppL Microbiol. 15: 110. ase pancreatique. Biochim. Biophys. Acta Shipe, W.F. Jr. 1951. A study of the relative Substrate specificity. To study the 43: 103. specificity of lipases produced by Penicil­ substrate specificity of the lipase, the ac­ Fodor, P.J. and Chari, A. 1949. The ester­ lium roqueforti and Aspergillus niger. Arch. hydrolyzing enzyme systems of Aspergillus Biochem. 30: 165. tivity of the enzyme against several tri­ niger and of Penicillium roqueforti. En- Thibodeau, R. and Macy, H. 1942. Growth and glycerides was determined. Results shown zymologia 13: 258. enzyme activity of Penicillium roqueforti. in Table 4 indicated that the Penicillium Fukumoto, J., Iwai, M. and Tsujisaka, Y. 1963. Minn. Agr. Expt. Sta., Tech. Bull. 152. Studies on lipase. I. Purification and crystal­ Wilcox, J.C., Nelson, W.O. and Wood, W.A. lipase was most specific fortributyrin. The lization of a lipase secreted by Aspergillus 1955. The selective release of volatile acids lipase hydrolyzed tributyrin, tricaprylin, niger. J. Gen. Appl. Microbiol. 9: 353. from butterfat by microbial lipases. J. Dairy Fukumoto, J., Iwai, M. and Tsujisaka, Y. 1964. Sci. 38: 775. tricaprin, tripropionin and triolein in de­ Studies on lipase. IV. Purification and prop­ Wills, E.D. 1961. Studies on the purification creasing order. However, the difference erties of a lipase secreted by Rhizopus and specificity of pancreatic lipase. The en­ dele mar. J. Gen. Appl. Microbiol. 10: 257. zyme of lipid metabolism. Proc. 6th Intern. between the rate of hydrolysis for tribu­ Harper, W.J., Schwartz, D.P. and El-Hagarawy, Conf. Biochem. Lipids. Desnuelle, p., ed. tyrin and tricaprylin was slight. Triolein I.S. 1956. A rapid silica-gel method for Pergamon Press, N ew Y o rk . and tripropionin were hydrolyzed at measuring total free fatty acids in milk. J. Ms. received 5/2/69; revised 7/28/69; accepted Dairy Sci. 39: 46. 7 /3 0 /6 9 . much lower rates than the other triglycer­ Iwai, M., Tsujisaka, Y. and Fukumoto, J. 1964. ides studied. These observations are in Studies on lipase. III. Effect of calcium ion Published as paper No. 2576, Journal Series, on the action of the crystalline lipase of Nebraska Agricultural Experiment Station, harmony with those of Alford et al. Aspergillus niger. J. Gen. Appl. Microbiol. L in c o ln . (1961). Also, Wilcox et al. (1955) report­ 10: 87. This work was supported in part by a grant Iwai, M. 1965. Studies on mold lipase. Bull. ed th a t P. ro q u e fo rti lipase preferentially from the American Dairy Association and by a Osaka Municipal Tech. Res. Inst., Vol. 39. fellowship awarded by the Dairy and Food In­ hydrolyzed short-chain fatty acids from Khan, I.M., Dill, C.W., Chandan, R.C. and dustries Supply Association to the senior butter fat. Shahani, K.M. 1967. Production and prcper- author (R. R. Eitenmiller). M. R. A LT AM UR A, R. E. ANDREOTTI, M. L. BAZINET and L. LONG JR.

Pioneering Research Laboratory U.S. A rm y Natick Laboratories, Natick, M assachusetts 01760

PYROGLUTAMYL DIPEPTIDES IN MUSHROOM, AGARICUS CAMPESTRIS

SUMMARY—A small fraction, containing evidence for the presence o f pyroglutamyI dipeptides from the characterization of a dipeptide- and an N-pyroglutamylhexosamine, was isolated from the edible, commercial mushroom, Agaricus containing fraction, isolated from the campestris. Upon hydrolysis, this isolate liberated a copious amount of glutamic acid, together fresh, white, edible and commercially cul­ with smaller quantities o f other amino acids and a hexosamine. Analyses o f the unhydrolyzed and tivated mushroom, Agaricus cam pestris L. hydrolyzed portions o f the fraction by an automatic amino acid analyzer revealed the identities ex Fr. (Agaricaceae), sometimes referred and quantities o f the compounds released. Additional chemical and physical methods o f analysis to as Agaricus bisporus or Psalliota cam­ gave supporting evidence for the presence in the mushroom in addition to proline and pyroglutam- p e s tr is . These results include evidence for ic acid o f the pyroglutamyl dipeptides o f threonine, aspartic acid, valine, leucine, citrulline, phenyl­ the identity and general structure of the alanine, glycine, alanine, glutamic acid and proline and also o f the amino acid sugar, N-pyrogluta- compounds present in the mushroom iso­ m ylglucosam ine. Pyroglutamylcitrulline and N-pyroglutamylglucosamine were tentatively late and detected for the first time in identified. The first 6 dipeptides are reported for the first time. A mechanism is proposed for the mushrooms. enzymic biosynthesis o f the pyroglutamyl dipeptides in the mushroom. Results of the present EXPERIMENTAL work shed additional light on the complex nature o f the higher basidiomycetes. Source of mushroom INTRODUCTION did not possess an active amino group re­ The commercially grown fresh mushrooms, matured tc a tight-button stage and harvested in IN GENERAL, early work on mushrooms quired for color formation (such as an July, were purchased from the Bay State Mush­ was concerned with the isolation, separa­ a-amino group), they would not appear room Co., Sudbury, Mass. tion and identification of the more com­ on the chromatograms and consequently mon and essential amino acids, both free would escape detection. Acid hydrolysis, Source of model compounds and other and bound, from simple extracts of mush­ followed by chromatography, paper or materials rooms selected from different taxonomi- column (as in an amino acid analyzer), The authentic compounds and other materi­ cal families, and from those closely relat­ has provided an indirect method of estab­ als were purchased as follows: trimethylsilyl (TMS)-L-proline (in hexane solution), TRI-S1L ed to Agaricus campestris, such as lishing the presence of dipeptides (or higher peptides) regardless of whether or (silylation reagent), Silyl- 8 (GLC column con­ Agaricus hortensis and Boletus edulis. Re­ ditioner) and PO-17 3% on chromosorb W (HP) cently, the trend has been to obtain a not they are colored by ninhydrin. For 80/100 mesh (GLC stationary phase) from deeper insight into the chemical composi­ example, with ninhydrin-inactive pyro­ Pierce Chemical Co., Rockford, 111.; L-pyroglu- tion of the higher fungi. Altamura et al. glutamyl dipeptides, a chromatographic tamic acid from Sigma Chemical Co., St. Louis, (1967) have reported a novel procedure spot or elution peak, as the case may be, Mo.; L-pyroglutamyl-L-proline, L-pyrogluta- which afforded the isolation of a number would appear for glutamic acid after myl-L-glutamic acid monohydrate, and y -L - of the less common amino acids and hydrolysis, due to the decyclization of glutamyl-L-glutamic acid from Cyclo Chemical Corp., Los Angeles, Calif. The model com­ other nitrogenous substances from A. liberated pyroglutamic acid (PCA, III). This indirect method has been reported pounds were used for comparative and identifi­ campestris. Lately, 2 new amino acids, cation purposes. L-2-amino-3-hydroxymethyl-3-pentanoic by previous investigators of mushrooms acid and L-2-amino-3-formyl-3-pentanoic (Touze-Soulet, 1961) and marine alga Isolation of mushroom fraction acid have been isolated (Doyle et al., (Dekker et al., 1949). The initial part of the isolation procedure, i.e., from the washing of the fresh mushrooms 1968) from the mushroom, Bankera fulig- Although dipeptides (and higher pep­ (3.2 kg) to the completion of the petroleum in e o a lb a . tides) containing the Pyr moiety in the structure have not previously been detect­ ether extraction step, is the same as that given Although almost all living tissues seem in the previous paper (Altamura et al., 1967). to contain quantities of low-molecular- ed in mushrooms, a number of these com­ pounds have been found, both free and The petroleum ether insoluble solid (66.7 g) weight peptides, the work reported in the was shaken (10 x 200 ml) with wet ether literature regarding these substances in bound, in other natural products. Three (diethyl ether saturated with water) and the mushrooms is very meager. This is partic­ recorded examples are: Pyroglutamyl- ether extract removed each time. The insoluble ularly true of the dipeptides. Touze- glutaminyl-glutamine from the marine substance was shaken with absolute ethanol Soulet (1961) detected 2 unidentified alga, Pelvetia fastigiata (Dekker et al., (150 ml), the solvent removed at 35-40° and ninhydrin-active peptides in an extract 1949), pyroglutamyl-asparaginyl-trypto- under part al vacuum and the residue dried from B . e d u lis . Two free dipeptides, car- phan, from snake’s (Agkistrodon halps (35—40° / 5 mm) to constant weight (65.8 g). The reddish-brown, resinous-like mass was ex­ nosine ((3-alanyl-L-histidine) and 7 -L- b l o m h o f f i ) venom (Kato et al., 1966) and glutamyl-S-methyl-L-cysteine, were iden­ the pyroglutamyl dipeptide, pyrogluta- tracted with dry pyridine (150, 4 X 50 ml), and myl-glutamine, from the hydrolysate of the insoluble part was dried (room tempera- tified (Altamura et al., 1967) in an isolate ture/4 mmi to constant weight (63.5 g). The from A. campestris. Both of these dipep­ the heavy chain of rabbit immunoglobin resultant powder was extracted with methanol tides were positive to ninhydrin. With re­ IgG (Wilkinson et al., 1966). (7 x 250 ml), tire extracts combined, the sol­ gard to the pyroglutamyl (Pyr; sometimes The presence of the amino acid sugar, vent removed (Rinco/25-30°/ca. 34 mm) and called pyroglutamoyl) dipeptides, how­ N-pyroglutamylglucosamine (XVII) in the residue dried in vacuo (4 mm). This proce­ ever, no evidence was found in the litera­ mushrooms and other natural products dure gave a light-reddish-brown, resinous-like ture of their presence in mushrooms. has also escaped detection in the past. Re­ solid (39.1 g). For convenience, 20.8 g of the Conceivably, one reason for the failure to lated compounds, however, such as solid was stirred with water (416 ml). detect these substances can be attributed N -( L-(3 -aspartyl)-2-acetamido-2-deo xy-j3- The colored solution was filtered; the fil­ to the past practice of visualizing the D-glucosamine, have been isolated (Marks trate concentrated (Rinco) cautiously (foam­ chromatographically separated mushroom et al., 1963) from a partial hydrolysate of ing) at 15-20° and in vacuo (ca. 15 mm grad­ hen’s-egg albumen. ually reduced to 8 mm) to 166.4 ml. The components with ninhydrin reagent. If concentrate was refrigerated (ca. 1°) overnight pyroglutamyl dipeptides were present and This paper presents results obtained and then filtered cold. The filtrate was divided

134-JOURNAL OF FOOD SCIENCE-Volume 35 (1970) PYROGLÜTAMYL Dl PEPTIDES IN MUSHROOM- 1 3 5 into 3 equal parts, and each part stirred with ace­ Table 1 —Composition o f mushroom isolate. tone (1,550 ml) to give complete precipitation of a heavy, red oil and a white, fluffy solid. The A m ount mixture was refrigerated overnight to separate present the phases. At room temperature, the superna­ Com ponent (%) tant liquid was removed and the bottom phase (oil and solid) stirred with 75% acetone in Composite of free amino acids1 1.68 water until the white solid dissolved. Acetone Pyroglutamylproline 3.50 was then added cautiously until the white solid Proline 5.77 just started to precipitate. At this point, the Composite of pyroglutamyl dipeptides of amino acids2 plus supernatant liquid was removed from the bot­ N-pyroglutamylglucosamine 11.90 tom oil, filtered and stirred with acetone (276 Pyroglutamic Acid 25.723 ml) for complete precipitation of the white Pyroglutamylglutamic Acid 51.423 solid. The 2 other portions of the above con­ centrate were treated in like manner, and the 3 1 Minor amino acids. 2 mixtures pooled. Bound amino acids. The combined mixture was refrigerated 3 Approximate value. overnight and filtered cold. The solvents (ace­ tone and water) were removed from the filtrate at 25-30° and under vacuum, so regulated as to distill first the acetone and then, very cau­ tiously, the water. Occasional addition of ethanol (ca. 5 ml) retarded foaming and facili­ tated removal of the last traces of water. This tions generally used for protein hydrolysis. The the acidic, neutral and basic amino acids. These afforded a dark, highly viscous, glass-like solid fraction (4 mg) and 6 NHC1 (4 ml) were heated were compared with the chromatograms from (1.2 g). The solid was then stirred with minimal in a sealed tube (previously evacuated and standard mixtures of amino acids, generally (10 ml) methanol and the mixture refrigerated flushed with nitrogen) at 110° for 20 hr. After found in peptide and protein hydrolysates. The overnight. After filtration in the cold and re­ hydrolysis, the contents of the tube were evap­ identity and concentration of each amino acid moval of the solvent the residue was taken to orated to dryness, the residue dissolved in buff­ were determined by conventional procedures er solution (4 ml) and an aliquot (1 ml) subject­ constant weight (1.1 g) under the usual condi­ and the application of the prescribed method of ed to analysis on the amino acid analyzer. tions. It was then shaken with ca. 7 ml of 40% calculation given in the instruction manual. Re­ methanol in isopropyl alcohol and subsequently Automatic amino acid analysis sults are given in Table 2. refrigerated overnight and centrifuged at room The components, present in the unhycro- Infrared and ultraviolet measurements temperature. lyzed and hydrolyzed portions o f the mushroom The infrared spectra of the mushroom frac­ The supernatant liquid, after filtration and fraction, were separated and quantified by ion- tion (film between NaCl discs) and the refer­ evaporation, gave 1.0 g of a colored material. exchange column chromatography in the accel­ ence compounds (in K Br pellets), PCA. pyro- Repetition of this step (minus refrigeration and erated automatic amino acid analyzer (Model glutamylproline and pyroglutamylglutamic centrifugation) with minimal (8 ml) absolute K -8000-C, Phoenix Precision Instrument Co., acid monohydrate, were recorded on an Infra­ ethanol gave 753.2 mg of an ethanol soluble Philadelphia, Pa.). The method of analysis used red Spectrometer (Model 137, Perkin-Elmer, substance, which was subsequently decolorized with this instrument is essentially the one de­ Norwood, Conn.). (to pale yellow) by two separate treatments of scribed by Spackman et al. (1958) as applied to The ultraviolet spectra of the above sub­ its methanolic (15 ml) solution with 1.6 g of protein hydrolysates in a 50° program, except stances (in aqueous solutions) were taken on activated carbon (Darco, G-60) at boiling that the analysis is carried out on an accelerated the Cary Recording Spectrophotometers methanol. The light-colored, now odorless scale. The procedure is described in the opera­ (Models 11 and 11MS, Applied Physics Corp., solid, was then stirred ( 5 x ) with 2 ml of 0.5% tional manual, Liquid Chromatography Hand­ Pasadena, Calif.) between 210-400 mg methanol in isopropyl alcohol; the mixture was book, published ca. 1966, and provided with centrifuged each time; the separated, superna­ the instrument by the Phoenix organization. Trimethylsilylation tant liquids were combined and after complete­ Elution chromatograms were obtained for For the analyses obtained from mass spec- ly removing the solvents, approximately 213.4 mg of a viscous product obtained. Final extrac­ tions (2x2 ml and 3x1 ml) with isopropanol, and following through as before (including dry­ ing to constant weight at room temperature/3 mm) , yielded 211.7 mg of a clear, light-colored, Table 2—Change in amino acid and glucosamine content o f mushroom fraction due to hydroly- odorless, viscous syrup. For 39.1 g of the above sis. original solid, the estimated weight of this Am ount2 A m ount2 A m ount2 mushroom fraction was 398.0 mg (0.01% of the C om pound1 (before hydrolysis) (after hydrolysis) (released) weight of fresh mushrooms used). The mushroom fraction gave a) copious evo­ Glucosamine 0 0.011 0.011 lution of carbon dioxide with sodium bicar­ Aspartic Acid (Asp) Trace3,4 0.018 0.010 bonate (presence of free carboxyl group), b) Threonine (Thr) 0.015 0.031s 0.016s positive (moderate intensity) Molisch test (pres­ Glutamic Acid (Glu) Trace 2.990 2.982 ence of sugar moiety), c) positive (moderate Proline (Pro) 0.500 0.580 0.080 intensity) ninhydrin test (presence of a-amino Citrulline (Cit)6 Trace 0.056 0.048 group), d) negative biuret test (absence of pep­ Glycine (Gly) 0.010 0.041 0.031 tides with 2 or more peptide bonds), e) nitro­ 0.049 0.091 0.042 gen content, 10.59% and f) solubility behavior Alanine (Ala) 0.055 0.020 as evident from the isolation procedure. The Valine (Val) 0.035 composition of the isolate is given in Table 1. Leucine (Leu) 0.032 0.041 0.009 0.028 0.018 Molecular weight measurement Phenylalanine (Phc) 0.010 The average molecular weight of the frac­ 'Compound listed according to sequence of elution from columns. tion was measured on a vapor pressure osmom­ 2MMoles/mg fraction. eter (Model 301A, Mechrolab, Inc., Moun- 3Less than 0.008 pmole/mg fraction for a trace amount. tainview, Calif.) using methanol and water as 4For purposes of calculation, the presence of at most 0.008 Mmole/mg fraction is assumed for solvents. a trace compound. Hydrolysis of the mushroom fraction 5 Approximate value. 6For brevity the authors suggest “Cit” as a symbol for citrulline. The isolate was hydrolyzed under condi­ 1 3 6 -JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

Table 3— Comparison of the abundance of Table 4—Comparison of the percent abundance of the fragment the fragment ion 84. ions, 99, 127 and 140. TMS derivative of: m/e 84 (% Abundance) TMS glutamyl­ TMS pyroglutamyl­ TMS mushroom Mass ion (m/e) glutamic acid1 glutamic acid2 fraction Pyroglutamic Acid 100.01 Mushroom fraction 56.0 99 6.9 0.8 3.3 Pyroglutamylglutamic Acid 4.5 127 3.5 0.8 0 Pyroglutamylproline 4.2 140 19.8 0.4 2.5 Control (TRI-SIL) 0.1 'Alkyl-type, straight-chain dipeptide. 1 Mass peak 84 is the m ost intense peak (base 2Dipeptide containing a cyclic, pyroglutamyl moiety. peak) for TMS pyroglutamic acid.

trometry (MS) and the combined gas-liquid 7 -lactam, pyrrolidinonyl ion (V) of mass unit bound with PCA as dipeptides. The pres­ chromatography (GLC) Time-of-Flight (TOF) 84; b) determining from the spectra the abun­ ent finding and the previously reported mass spectrometry, the trimethylsilyl (TMS) dance or deficiency of other significant mass presence of PCA in Agaricus campestris derivatives of the mushroom isolate and the ref­ ions which could be used to differentiate be­ (Le Roux, 1961; 1962; Latche et al., erence compounds, PCA, pyroglutamylproline, tween alkyl-type, straight-chain, glutamyl com­ 1968) and in other higher fungi (Touze- pyroglutamylglutamic acid monohydrate and pounds, e.g., glutamyl dipeptides, exemplified Soulet et al., 1957) give support to the glutamylglutamic acid were used. They were by glutamylglutamic acid (XVIII), and cyclic, prepared essentially by the procedure of pyroglutamyl-type compounds, e.g., pyrogluta­ occurrence of dipeptides derived from this Sweeley et al. (1963), except that TR1—SIL myl dipeptides, such as pyroglutamylglutamic amino acid in the mushroom under con­ (mixture of hexamethyldisilazane, trimethyl- acid (XI). The 3 typical fragments selected have sideration. chlorosilane and anhydrous pyridine) was em­ mass units 99, 127 and 140; c) furnishing the The probability that PCA is formed as ployed as the silylation reagent. It was not nec­ molecular ions (M+) or the molecular ions mi­ an artifact from glutamic acid during the essary to isolate the TMS derivatives from the nus some logical fragment ions (M+-R) of the work-up of natural products has been dis­ reaction mixtures (Sweeley et al., 1963). In TMS pyroglutamyl compounds present in the proved by a number of investigators TMS mushroom isolate. each case the clear supernatant liquid, carefully (Niwaguchi et al., 1965; Wilkinson et al., removed from the reaction solids and contain­ Fragmentation patterns of the reference ing the desired derivative, gave satisfactory re­ compounds were obtained to provide frag­ 1966; Piquemal et al., 1967; Johnson et sults. The presence of excess silylation reagent ment-structure information and these were al., 1968). The very mild conditions used did not interfere. The test liquid was either in­ found most helpful for the interpretations re­ in the present work preclude the presence jected directly into the GLC column of the quired in evaluation of the mass spectra of the of PCA as an artifact. The rather high per­ TOF-mass spectrometer or into the inlet system TMS mushroom fraction. The mass spectral centage of the pyroglutamyl dipeptides in of the mass spectrometer, if the latter instru­ data are presented in Tables 3, 4 and 5. the mushroom fraction indicates the ef­ ment was used alone. ficacy of the isolation procedure to con­ RESULTS & DISCUSSION centrate a fraction containing predomi­ Combined gas-liquid chromatography and TABLE 1 lists the more important com­ nantly this class of compounds. time-of-flight mass spectrometry (GLC-TOF- MS) ponents of the mushroom isolate and Results obtained from the hydrolysis their amounts. The 2 free, major amino For these analyses, a GLC coiled stainless of the mushroom fraction are shown in steel column (1/8 in. by 10 ft) was attached to acids are pro line (I) and PCA; the free Table 2. Of particular interest is the copi­ a modified Bendix Time-of-Flight mass spec­ minor acids, present in above-trace quan­ ous amount of glutamic acid released in trometer (Model 14), so that the effluent of the tities (Table 2), are Gly, Ala, Thr, Val, contrast to the quantity of the other column was directed into the ionization source Leu and Phe. The quantities of PCA (the amino acids. This is consistent with the (Merritt et al., 1964). The column, packed with 7 -lactam form of glutamic acid, and large quantities of PCA and pyrogluta­ Chromosorb W (HP) 80/100 mesh support, sometimes referred to as 5-oxo-proline, mylglutamic acid present in the unhydro­ coated with 3% PO-17 and conditioned with 5-oxo-2-pyrrolidine carboxylic acid and lyzed fraction (Table 1 ) which, per se, lib­ Silyl • 8, was operated at a wide temperature pyrrolidinone carboxylic acid, etc.), to­ erate solely glutamic acid upon hydrolysis. range program of 75—225°C (12°C/min), an in­ gether with pyroglutamylglutamic acid jection temperature of 200°C and a carrier gas The other pyroglutamyl compounds also (helium) flow rate of 10 cc/min at atmospheric (XI), pyroglutamyl dipeptides of the contribute some glutamic acid. The re­ pressure. bound amino acids (VII—XVI) and lease of large amounts of glutamic acid N-pyroglutamylglucosamine were calcu­ upon the acid hydrolysis of pyroglutamyl Mass spectrometry lated from the corresponding amino acids peptides was also observed by Dekker et The spectral data for the TMS derivatives of and 2-glucosamine released by hydrolysis al. (1949). Presumably, hydrolysis not the mushroom fraction and standard com­ (Table 2). Approximately 2/3 of the only splits the peptide linkage but also pounds were obtained on a Consolidated Elec­ mushroom fraction components are decyclizes the pyroglutamyl moiety (IV). trodynamics Corp., Model 21-110B, mass spec­ trometer, operated at 70 e V and ca. 200°C. The test samples in the form of the supernatant liquids from the trimethylsilylation reactions were injected directly into the inlet system of the instrument with a direct introduction Table 5—Molecular ions and molecular ions minus fragment ions o f some components in the probe. The purpose of obtaining the data was TMS mushroom fraction. to provide additional evidence for the presence Component (M+) (K + - R) Nature of R of pyroglutamyl-type compounds in the mush­ room fraction by a) selecting a specific frag­ TMS proline 187 - — ment ion producible from the electron bom­ TMS Pyroglutamic Acid 201 - — bardment of the pyroglutamyl compounds, TMS Pyroglutamylproline - 283 (M+ - 15) CH3 - + which would have sufficient stability for its de­ TMS Pyroglutamylglutamic Acid - 256 (M+ - 146) 2 (CH3) 3Si • + tection and for the measurement of its intensity TMS pyroglutamylglycine 258 - — in the mass spectra. Accordingly, the fragment ion selected was the sufficiently stable, cyclic, TMS pyrogiutamylalanine 272 - - PYROGLUTAMYL D!PEPTIDES IN MUSHROOM- 1 3 7

The hydrolytic products, which now have available active a-amino groups, color nin- hydrin and become detectable on the elu­ tion chromatogram of the analyzer. How­ H, H, H, H, H; H, ever, the moderate ninhydrin reaction of the unhydrolyzed fraction was presum­ H? ,H H ably due to the free proline and the free 'COOH H, minor amino acids. PCA gives no color Ö o COOH reaction with ninhydrin. H

Besides glutamic acid, Table 2 also P r o lin e 2- Pyrrolidinone Pyroglutam ic acid shows that aspartic acid and citrulline were practically all in the bound form be­ (I) (II) (III) fore hydrolysis. The absence of free glu­ cosamine in the unhydrolyzed fraction H, H2 H, H ,+ suggests that this amino sugar was bound to the Pyr moiety through the peptide H linkage. This would then classify the ,H amino acid sugar, N-pyroglutamylglu- O CO- O' 'N ■ cosamine, as an N-acyl type amino sugar. H Hughes et al. (1958) reported the pres­ Pyroglutam yl group Pyrrolidinonyl ion ence of glucosamine in an hydrolysate of (IV) (V) a fraction derived from dried A. cam- pestris. They considered that this sugar was originally bound to a protein. In con­ ______.Hj trast to the other amino acids, nearly all of the leucine was present as free leucine R = amino acid moiety attached to (IV) in the present unhydrolyzed fraction. through the peptide (amide) link. Only the structures of the more important The average molecular weight (206.3) moieties are indicated below. of the mushroom isolate indicates that H the components present possess relatively Pyroglutamyl dipeptide low molecular weights. This value also 2 1 suggests that the peptide components are (VII) . R = Gly (XI). R = Glu, (-NH-CH( COOH)-CH2-CH2-COOH) not larger than dipeptides. This interpre­ (VIII) , R s Ala ( XII), R = Val tation is consistent v/ith the negative reac­ tion of the isolate with biuret. (IX) , R = Thr (XIII), R = Leu The infrared spectrum of the fraction (X) , R = Asp (XIV), R = Cit Hz was well defined. It exhibited absorption frequencies for the significant structures ( X V ) , R = P r o present in the components of the frac­ . (v -h\ oc o , o h ) tion, namely the 7 -lactam structure, pep­ tide link, un-ionized carboxyl, hydroxyl ( X V I) , R = P h e j------O ------1 and hydrogen bonding. The high-frequen­ (XVII ), R = 2-Glucosam ine, HO- CH- CH. CH( OH) • CH( OH) • CH- CH2OH cy spectral region showed the general ab­ sorption characteristics of peptides, and - NH also resembled that of the reference di­ (XVIII), HOOC- CH ( NH2) • CH2. CH2- CO- NH- CH( COOH) • CH2- CH2- COOH peptide, pyrolglutamylglutamic acid )f-GIutamylglutamic acid monohydrate. The fingerprint region, however, resembled PCA. Important max­ ima in the spectrum were a) broad, in­ tense bands in the region 3330—3050 cm" 1 (NH stretching vibration), 1670 cm" 1 (C = O stretching mode, Amide I) violet spectrum which showed a smooth­ The GLC-TOF-MS examination of the and 1560 cm" 1 (Amide II). The band at rising absorption curve with 2 small in­ TMS mushroom fraction revealed the 3330-3050 cm" 1 showed a split charac­ flections (near 260 mju and 320 m/i) and presence of 2 compounds whose elution teristic of dipeptides, b) an NH stretching no maxima. The 260 m/J inflection pre­ characteristics and mass spectra matched band (partly overlapped) for the 7 -lactam sumably arises from the benzyl Chromo­ those of the model compounds, TMS pro­ structure at 3170 cm" 1 (the C = O phore of the minute amount of free line and TMS pyroglutamic acid. No stretching mode for this structure, how­ phenylalanine present; the 320 m/u other compounds were detected on the ever, was overlapped by the C = O fre­ shoulder is probably due to the corre­ GLC chromatogram. quency of the peptide group) and c) sponding dipeptide, pyroglutamylphenyl- The detection of TMS pyroglutamic bands for un-ionized COOH at 3000—2500 alanine. The mushroom spectrum com­ acid in the GLC-TOF-MS run confirms cm" 1 (OH stretching mode) partly over­ pared favorably with the spectral profiles the presence of PCA in the mushroom lapped at 1320—1210 cm" 1 (C —O stretch­ of the reference compounds, PCA, pyro- under consideration. ing, or OH deformation) and at 955 cm" 1 glutamylproline and pyroglutamylgluta- The mass spectrum of the TMS mush­ (OH deformation out of plane). The C = O mic acid, all of which gave featureless room fraction, like those of the reference vibration mode for this group was also spectra. The spectrum of the isolate is compounds, was well defined. Table 3 overlapped by the C = O stretching mode consistent with the chemical composition gives a comparison of the percent abun­ of the peptide group. of the isolate as revealed by the amino dance of the fragment ion of mass 84 The mushroom fraction gave an ultra- acid analyzer. (pyrrolidinonyl ion, V) for the TMS 138 -JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

mushroom fraction and its 3 major com­ amino acid, then pyroglutamylglutamic vivo), however, the released threonine ponents. As expected from the structure acid (XI) would be formed. would again exhibit its natural biological and fragmentation pattern of standard The other dipeptides disclosed in this function. The rate of its release could, of PCA, the abundance of its 84 peak would paper could also be accounted for in the course, influence the magnitude of its be large. By comparison, the relatively same manner. For N-pyroglutamylglucos- function. The liberated PCA would also high intensity of this fragment ion in the amine, the coupling would have to occur become available for biochemical reac­ mushroom fraction resulting from elec­ between the activated carboxyl of PCA tions. tron bombardment indicates the presence and the 2 -amino group of glucosamine. of components containing the pyrogluta- The dipeptide-synthetase enzymes needed REFERENCES myl moiety or, if preferred, the cyclic, to catalyze the above phosphorylation 7 -lactam structure. The assignment of the and coupling reactions are presently un­ Altamura, M.R., Robbins, F.M., Andreotti, m/e 84 to the pyrrolidinonyl ion is re­ R.E., Long, L. Jr. and Hasselstrom, T. 1967. known in mushrooms. Presumably they Mushroom ninhydrin-positive compounds. ported by Heyns et al. (1961), Biemann are present, since a) about 2 0 , and pos­ Amino acids, related compounds, and other (1962) and Martin (1965). sibly more, enzymes exhibiting a variety nitrogenous substances found in cultivated of catalytic activities have been detected mushroom, Agaricus campestris. J. Agr. Table 4 provides additional evidence Food Chem. 15 ( 6 ): 1040. in A. campestris and other mushrooms Biemann, K. 1962. “Mass Spectometry. Organic for the presence of pyroglutamyl com­ Chemical Applications”, p. 267, McGraw- pounds in the mushroom fraction. It is (Miller, 1929; Thoai, 1942; Voinovitch, Hill Book Co., Inc., New York. 1950; Erickson et al., 1960; Latche, Dekker, C.A., Stone, D. and Fruton, J.S. 1949. evident that the abundances of the frag­ A peptide from a marine alga. J. Biol. Chem. ments of mass units 99, 127 and 140 are 1963) and in most cases identified, b) the 181: 719. higher in the alkyl-type, straight-chain di­ existence of dipeptides has been reported Doyle, R.R. and Levenberg, B. 1968. Identifica­ tion of two new )3-unsaturated amino acids peptide, TMS glutamylglutamic acid, than (Altamura et al., 1967) in A. campestris in the mushroom, Bankera fuligineoalba. in the TMS pyroglutamylglutamic acid. and c) a dipeptide-synthetase enzyme, Biochemistry 7 (7): 2457. Erickson, R.E., Brown, K.S. Jr., Wolf, D.E. and 7 -glutamylcysteine synthetase, has been Values of the 3 mass ions for the TMS Folkers, K. 1960. Coenzyme Q. XX. Isola­ mushroom fraction lie between those of found (Snoke et al., 1952) in other tissue. tion of coenzymes Q9 and Q10 from two basidiomycetes. Arch. Biochem. Biophys. 2 The proposed mechanism is exempli­ the reference compounds, though closer 90: 314. to the TMS pyroglutamyl dipeptide. fied as follows, using glycine as the partic­ Heyns, K. and Grützmacher, H.F. 1961. Dei ipating amino acid: M assenspektren der N-Formyl-a- Amino- The molecular ion (M+), or (M+—R), is sâuremethylester. Z. Naturforsch. 16b: 293. Hughes, D.H., Lynch, D.L. and Somers, G.F. listed in Table 5 for each of 6 compo­ enzyme 1958. Mushroom analysis. Chromatographic nents (4 major and 2 minor) present in 1. Pyroglutamic acid + ATP « - ■*- identification of the amino acids and carbo­ the TMS mushroom fraction. Spectral Pyroglutamylphosphate + ADP hydrates in the cultivated mushroom. J. correlations with reference compounds Agr. Food Chem. 6 (11): 850. 2. Pyroglutamylphosphate + glycine Johnson, F.B. and Hammill, M.M. 1968. The were good. The parent masses of TMS non-volatile organic acids of some fresh enzyme „ , , , . fruits and vegetables. Can. Inst. Food Tech- pyroglutamylcitrulline (469) and TMS ■■ „ —»- Pyroglutamylglycme nol. J. 1 (1): 3. pyroglutamylglucosamine (578) could Kato, H., Iwanaga, S. and Suzuki, T. 1966. The (VII) + h 3 p o 4 not be detected. Values of the molecular isolation and amino acid sequence of new pyroglutamylpeptides from snake venoms. ions given in Table 5 are in agreement Experientia 22 (1): 49. with the calculated molecular weights for Another possible pathway for the en­ Kritskii, M.S. and Kulaev, I.S. 1963. Acid-solu­ zymatic biosynthesis of a pyroglutamyl ble nucleotides of Agaricus bisporus L. fruit­ the TMS compounds. The excess silyla- ing bodies. Biokhimiya (USSR) 28 (4): dipeptide would be via the 7 -lactamiza- tion reagent, present in the test samples, 694. did not interfere with the spectral mea­ tion of an alkyl-type, straight-chain a-glu- Latche, J.C. 1963. Sur la Biosynthèse, par tamyl dipeptide. Transmination, de l’Alanine, des Acides surements. aspartique et glutamique et sur l’Existence The wide distribution of PCA in na­ de la glutamique Decarboxylase chez Agari­ Reactions involved in the enzymatic ture (animals, plants, bacteria and fungi) cus hortensis. Compt. Rend. 257: 499. biosynthesis of a pyroglutamyl dipeptide Latché, J.C., Baldy, P. and Piquemal, M. 1968. suggests that it could act as a precursor in in mushrooms are unknown. However, Evolution des Teneurs en Acides organiques the biosynthesis of the pyroglutamyl di­ Non volatiles et en Acides cétoniques (tes plausible mechanism for the formation of peptides in mushroom. Some PCA, per se, Carpophores provenant des Volées succes­ this type of compound may be reason­ sives d’une même Culture d’ Agaricus cam­ may conceivably undergo enzymatic de­ pestris Fr., variété bisporus. Compt. Rend. ably postulated by taking into account a) carboxylation to 2-pyrrolidinone (II), fol­ 266 (3): Series D, 214. the presence of enzymes, free PCA and LeRoux, P. 1961. Quelques Données quantita­ lowed by enzymatic hydrolysis of the lat­ other amino acids in mushrooms, b) the tives sur les Acides organiques Stables du ter compound to 7 -aminobutyric acid, an Champignon de Couch (Agaricus cam­ formation of a peptide (amide) bond pestris) Compt. Rend. 252: 205. amino acid, found in substantial quantity when appropriate energy supply is avail­ LeRoux, P. 1962. Métabolisme des Acides or­ in A. campestris. ganiques dans les Carpophores d’ Agaricus able and c) the analogy that exists in the campestris. Ann. Physiol. Vegetale 4 (2): biological systems of other living tissues The physiological activities of the 149. (Niwaguchi et al., 1965). Accordingly, it Levenberg, B. 1962. Role of L-glutamine as pyroglutamyl dipeptides present in mush­ donor of carbarnyl nitrogen for the enzy­ is conceivable that the nucleotide, adeno­ rooms are as yet unknown. However, matic synthesis of citrulline in Agaricus sine triphosphate (ATP), a high-energy they may contribute in some way to the bisporus. J. Biol. Chem. 237 (8 ): 2590. com pound present in A. campestris Marks, G.S., Marshall, R.D. and Neuberger, A. taste or flavor characteristics of the mush­ 1963. Carbohydrate in protein. 6 . S tudies (Levenburg, 1962; Kritiskii et al., 1963) room, possibly as flavor potentiators. on th e 2 arbohydrate-peptide bond in hen’s- could activate (phosphorylate) the car­ egg albumen. Biochem. J. 87: 274. From the nutritional viewpoint, the bio­ Martin, M. 1965. An investigation of the mass boxyl group of PCA to give pyrogluta- chemical activity of a pyroglutamyl di­ spectra of twenty-two free amino acids. mylphosphate. This anhydride, by virtue Tech. R.ept. No. IRL—1035, Instrumenta­ peptide may be altogether different from tion Res. Lab., Dept, of Genetics, Stanford of its activated carboxyl group now being that of each of its consitituent amino Univ., Palo Alto, California. susceptible to nucleophilic attack, would acids, or it may reflect a suppressed activ­ McCoy, R.H., Meger, C.E. and Rose, W.C. 1935. Feeding experiments with mixtures of react with an amino acid via the a-amino ity of 1 or both of its components. For highly purified amino acids. VIII. Isolation group to form the peptide link and liber­ example, in pyroglutamylthreonine (IX) and identification of a new essential amino acid. J. Biol. Chem. 112: 283. ate phosphoric acid (an acid present in the growth function (McCoy et al., 1935) Merritt, C. Jr., Walsh, J.T., Forss, D.A., Ange- higher fungi). If glutamic acid (an impor­ of the constituent, threonine (an essential lini, P. and Swift, S.M. 1964. Wide-range tant and plentiful amino acid in mush­ amino acid), may be repressed or com­ programmed temperature gas chromatog­ raphy in the separation of very complex room) happens to be the participating pletely lost. Upon enzymic hydrolysis (in mixtures. Anal. Chem. 36 (8 ): 1502. PYROGLUTAMYL D!PEPTIDES IN MUSHROOM- 1 3 9

Miller, E.R. 1929. The chemical nature of en­ Sweeley, C.C., Bentley, R., Makita, M. and Wilkinson, J.M., Press, E.M. and Porter, R.R. zymes. Science 70: 356. Wells, W.W. 1963. Gas-liquid chromatog­ 1966. The N-terminal sequence of the heavy Niwaguchi, T., Notohashi, N. and Strecker, H.J. raphy of trimethylsilyl derivatives of sugars chain of rabbit immunoglobin IgG. Bio­ 1965. The metabolic conversion of L-gluta- and related substances. J. Am. Chem. Soc. chem. J. 100: 303. mate to pyrrolidone carboxylate in rat tis­ 85: 2497. Ms. received 6/23/69; revised 7/18/69; accepted sues. Biochem. Z. 342: 469. Thoai, N-van. 1942. Phosphomonoestérases et 7 /2 2 /6 9 . Piquemal, M., Latché, J.C. and Baldy, P. 1967. Pyrophosphatases des Basidiomycètes. Métabolisme de l’Acide glutamique 3, 4"14C Compt. Rend. 214: 643. Gratefully acknowledged are the uses of the amino acid analyzer and the mass spectrometers et de l’Acide 7 -aminobutyrique dans les Touzé-Soulet, J.M. 1961. &es Acides amines de Carpophores d’ Agaric us hortensis. Compt. Boletus edulis Fr. ex Bull, et de quelques made available by Drs. F.M. Robbins and C. Rend. 264 (2): Series D, 318. Espèces voisines. Compt. Rend. 252: 208. Merritt Jr., respectively; the technical advice on Touzé-Soulet, J.M. and Montant, M.C. 1957. the interpretation of mass spectra by Dr. P. Snoke, J.E. and Block, K. 1952. Formation and Les Acides organiques de quelques Basidio­ Angelini; the recording of the ultraviolet spec­ utilization of 7 -glutamylcysteine in gluta­ mycètes supérieurs. Compt. Rend. 245: tra by Dr. J. Weinstein and the mass spectral thione synthesis. J. Biol. Chem. 199: 407. 1825. measurements made by Mr. W. Yeomans, all of Spackman, D.H., Stein, W.H. and Moore, S. Voinovitch, I. 1950. Mise en Evidence d:une our Pioneering Research Laboratory; as are also 1958. Automatic recording apparatus for Succino-Deshydrogenase dans un Extrait the molecular weight determinations made by use in the chromatography of amino acids. oxydasique Purifé d’ Agaricus campestris. Mr. R. Sacher of our Clothing and Personal Life Anal. Chem. 30: 1190. Compt. Rend. 230: 2330. Equipment Laboratory.

New Subscription Rates NOTICE: Page Charges for Publication of Research Papers To IFT Members: $10 per year ($5 to Emeritus and student members). EJA T ITS March 1969 meeting, the Executive Com­ To Non-Member Subscribers-per year, includes mittee o f IFT voted to increase the page charge for postage research articles published in Food Technology or in USA & Pan-American U nion...... $20 the J o u r n a l o f F o o d Sc i e n c e . The page charge All Other D estinations...... $25 of $50 per printed page will be effective for research Subscriptions accepted only on a calendar year manuscripts received AFTER April 1, 1969. The page basis. No Refunds. Reduced rates for 2- and 3-year charge shall not constitute a bar to acceptance o f re­ subscriptions. search manuscripts because the author is unable to Single Copy or back issue price: $5. pay the charge. A. C. M ATHEW and H. A. B. PAR PI A

Central Food Technological Research Institute, M ysore, India

POLYPHENOLS OF CASHEW KERNEL TESTA

SUM M ARY—The chemical nature of the polyphenols of cashew kernel testa has been determined RESULTS by means o f 2-dimensional paper chromatography and specific reactions. Presence o f (+) catechin HOT EXTRACTION of the testa by wa­ and I-) epicatechin as the major polyphenols has been confirmed by co-chromatography with ter, 50% aqueous acetone and acetone authentic samples. Proanthocyanidins have been studied by treatment with hot acid and identifica­ showed that the best extraction, with re­ tion o f the resultant anthocyanidins. Leucocyanidins and leucopelargonidins have been noticed as spect to polyphenol content and maxi­ monomeric components, whereas polymeric proanthocyanidins contain in addition small quantities mum number of spots on chromatogra­ o f leucodelphinidin. Individual components have been quantitatively estimated colori metrically phy was achieved by 50% aqueous ace­ with Folin-Denis reagent, after separation by fractional extraction using solvents o f increasing tone. All the extracts upon conversion polarity and paper chromatography. The characteristic bluish-black discoloration noticed in cashew- into anthocyanidin (Govindarajan et al., nuts has been identified as an iron-polyphenol com plex form ed during processing. The polypnenols 1965) showed a broad absorption peak at were found to be derived from the testa. 545 m/j, suggesting that the leucoantho- INTRODUCTION Extraction and fractionation cyanidins are dominated by leucocyanidin Extraction of polyphenolic materials was and an inflection at 450 m/u, indicating a THE KERNEL of Anacardium occiden- carried out by 50% aqueous acetone under the hydroxyl group at 5-position in the antho­ tale, valued as cashewnut in trade, is cov­ conditions described in an earlier com munica­ cyanidin formed (Jurd, 1962). No free ered with a thin reddish-brown skin, or tion (Mathew et al., 1964). The extract on anthocyanidin was noticed in the testa. testa. The testa has been reported to be a freeze-drying appeared as brownish pow der and good source of the catechol type of tan­ showed little change on chrom atographic exam ­ nins (Madhavan Pillai et al., 1963). i n a t i o n . For fractionation of testa polyphenols, 0.2 g During the normal processing of cash­ of the freeze-dried extractives was taken in 75 ewnut in the moisture-conditioning step ml of water. This was extracted successively in © before roasting, a small percentage of the a 150-ml separating funnel with 3 lots of 25 ml nuts develop bluish-black patches on the of ether. The residual aqueous phase was simi­ surface. Such nuts do not find ready mar­ larly extracted successively with ether-ethyl ket, although the taste is unaffected. Pre­ acetate (1:1), ethyl acetate - ethyl alcohol (4:1) vention or removal of such darkening is to get different fractions such as 1) ether; 2) bound to help the industry, particularly ether-ethyl acetate; 3) ethyl acetate; 4) ethyl since India is facing a shortage of raw ma­ acetate alcohol and 5) aqueous residue. terial for processing. Because of the close Chrom atographic procedure contact of the testa with the nut, the Identification of anthocyanidins form ed was water-soluble constituents of the testa as described earlier (Lakshminarayana et al., may reach into the nut. In this paper we 1967). The 2-dimensional paper chrom atogram s present a systematic examination of the on No. 3 papers were developed using n-buta- polyphenols of the testa and the influ­ nol-acetic acid-water (4:1:2.2 v/v) (BAW) as ence of these on discoloration of the the first direction solvent and 2% aqueous ace­ tic acid as the second. The chromatograms edible kernel. were sprayed with specific spray reagents EXPERIMENTAL ( R o u x e t A., 1960; Govindarajan et al., 1963). The quantitative estim ation of individual com­ Fig. 1 -Two-dimensional chromatogram of SAMPLES OF testa and nuts were collected ponents was carried out after 2-dimcnsional cashew kernal testa polyphenols. from processing centers in Kerala State. The separation on No. 3 W hatman paper, marking dried testa showed 42.5% total extractable sol­ the areas from another paper sprayed with t First direction BA W, -*■ second direction ids and 32% total phenol by the Folin-Denis F e C H + K 3Fe (CN)g and reacting with Folin- 2% aqueous acetic acid. Material 0.001 g; spray method (Swain et al., 1959) when extracted Denis reagent (Govindarajan et al., 1963; - K 3 F e ( C N ) 6 + F e C I3. w ith 50% aqueous acetone. Mathew et al., 1969). For details■ o f the spots, see Table 2.

Table 1 —Analysis o f the anthocyanidin m ixture from cashew kernel testa polyphenols.

R f

F o r m ic Absorption Phenolic Acid B a n d F o r e s ta l A cid p e a k m p Shift in peak obtained on Approx. % in N o . s o lv e n t s o lv e n t C o lo r (Ethanol-HCl) with A1G3 mp alkali fusion I d e n t ity the m ixture

i 0 .3 8 0 .1 4 Violet-pink 5 6 0 + 3 0 Gallic Acid Delphinidin 3 2 0 .5 0 0 .2 4 P in k 5 4 0 +20 Protocatechuic Acid C v a n id in 7 7 3 0.68 0 .3 7 R e d -p in k 5 3 0 N il p-Hydroxy Benzoic Pelargonidin 15 A cid 4 0 .8 1 0 .4 2 P in k 5 4 5 + 2 5 - n-Butyl cyanidin 4 5 0 .9 1 0 .5 0 R e d -p in k 5 3 0 N il - n-Butyl pelargonidin 1

140—JOURNAL OF FOOD SCIENCE-Volume 35 (1970) POLYPHENOLS OF CASHEW KERNEL TESTA- 141

Identification of polyphenols cyanidin and pelargonidin, respectively catechin, respectively, by co-chromatog­ The anthocyanidin mixture obtained (Mathew, 1969). raphy w ith authentic samples from on treatment with hot n-butanol-HCl sep­ Figure 1 shows the 2-dimensional Acacia catechu and arecanut (Govin- arated into 5 bands. From the Rf values, chromatographic separation of the total darajan et al., 1963). Spots 1 and 6 color, spectral characteristics and alkaline cashew kernel testa polyphenols. There showed no mobility in 2 % aqueous acetic degradation products their identity could were 15 spots or areas, whose properties acid, indicating the planar nature of the be established as given in Table 1. From and identity are presented in Table 2. The molecules (Harborne, 1960). From the the known facts about the formation of ether-extractable spots 3 and 4 were the color reactions and extractability with additional anthocyanidin, bands 4 and 5 major mobile components. These were ether, these compounds appeared to be can be taken as n-butyl derivatives of confirmed as (+) catechin and (- ) epi- flavonol derivatives. The proanthocyanidin spots 8 to 11 showed good mobility on paper and sol­ ubility in ethyl acetate, indicating their monomeric leucoanthocyanidin nature. Since spots 8 and 9 yielded pelargonidin on acid treatment, they were identified as two optical isomers of monomeric leuco- o’ o * pelargonidin. Spots 10 and 11 gave O O1 cyanidin and hence identified as mono­ O 0 1 O °» meric leucocyanidins. Spots 12 to 15 yielded anthocyanidins on treatment with hot acid. They repre­ sent pro anthocyanidins of increasing polymeric stages as shown by decreasing mobility on paper. Spot 12 appeared to be a simple oligomer from its relatively higher mobility and extractability both with moist ethyl acetate and with ethyl acetate containing alcohol. It was com­ posed mainly of leucocyanidin and minor amounts of leucopelargonidin. Spots 13 and 14 were predominantly leuco­ cyanidin with spot 13 showing traces of leucopelargonidin and spot 14 traces of leucodelphinidin. Spot 15, probably a highly polymerized proanthocyanidin, showed no mobility in either direction. This was composed of leucocyanidin and small amounts of leucodelphinidin. But for the differences in their proantho­ Fig. 2—Chromatograms o f cashew kernel testa polyphenols after solvent fractionation. cyanidin makeup, spots 13, 14 and 15 A —Ether, B—ether-ethyl acetate, C—ethyl acetate, D—ethyl acetate-ethyl alcohol, and E - could have been taken as a single spot. aqueous residue. ^ First direction— BAW, -> second direction—2% aqueous acetic acid. Estimation as in Table 3.

Table 2—Characteristics o f different polyphenolic spots in cashew kernel testa.

Intensity with R f in 2% S p o t N o . K3Fe(CN)6 + R f Vanillin-HCl Leucoanthocyanidin (F ig - 1 ) F e C b in B A W Acetic acid te s t te s t I d e n t ity

1 F a in t 0 .8 0 0 .0 4 N e g . N eg . F la v o n o l 2 F a in t 0 .8 9 0 .1 6 N e g . N eg . Non-flavalphenol 3 In te n s e 0 .8 1 0 .2 3 N e g . N e g . Non-flavanphenol 4 Very intense 0 .7 2 0 .4 2 P o s . N eg . (+) Catechin 5 Very intense 0 .6 2 0 .3 7 P o s . N eg . (-) Epicatechin 6 In te n s e 0 .6 4 0 .0 0 N e g . N eg . F la v o n o l 7 F a in t 0 .7 3 0 .6 7 Faintly pos. N eg . F la v a n 8 In te n s e 0 .5 8 0 .6 2 P o s. P o s. Leucopelargonidin 9 I n te n s e 0 .4 8 0 .5 4 P o s . P o s. Leucopelargonidin 10 In te n s e 0 3 4 0 .4 1 P o s . P o s. Leucocyanidin 11 F a in t 0 .4 1 0 .2 7 P o s. P o s. Leucocyanidin 12 In te n s e 0 .4 1 0 .0 P o s . P o s. Low polymeric proanthocyanidin 13 I n te n s e 0 . 2 2 - 0 . 3 4 0.0 Slight P o s . P o s. Polymeric proanthocyanidin s tr e a k in g 1 4 I n te n s e 0 . 0 - 0 . 2 2 0 .0 P o s . P o s. Polymeric proanthocyanidin 15 Very intense 0 .0 0 .0 P o s . P o s. Polymeric proanthocyanidin

All spots reacted with 6/j-diazotized benzidine and Folin-Denis reagent. Pos. = positive, Neg. - negative. 14 2 -JOURNAL OF FOOD SCIENCE- Volume 35 (1970)

Quantitative estimation of individual Table 3—Chromatographic estimation of polyphenolic spots in cashew kernel testa after solvent components fractionation. The chromatograms of the different fractions obtained after fractionation Amount in grams out of 0.2 g of total polyphenols with solvents of increasing polarity, i.e., II I l l IV V ether, ether- ethyl acetate, ethyl acetate, S p o t N o . 1 Ether-ethyl E th y l E th y l A q u e o u s % O f ethyl acetate-ethyl alcohol and aqueous ( T a b le 2 ) E th e r a c e ta t e a c e ta t e acetate-alcohol re s id u e T o ta l to t a l residue, are represented in Figure 2. The 1 0 .0 0 3 0 _ _ _ — 0 .0 0 3 0 1 .2 0 individual spots in each of the fractions 2 0 .0 0 1 0 _ — - - 0 .0 0 1 0 0 .4 0 were then estimated by Folin-Denis rea­ 3 0 .0 0 6 7 0 .0 0 0 2 — - - 0 .0 0 6 9 2 .7 7 gent and calculated in terms of the total 4 0 .0 1 4 5 0 .0 2 4 5 0 .0 0 7 5 - - 0 .0 4 6 2 1 8 .5 3 phenol in that fraction (Table 3). Values 5 0 .0 1 7 1 0 .0 2 8 6 0 .0 1 3 2 - - 0 .0 5 8 9 2 3 .6 3 — - - 0 .0 0 7 8 '3 .1 3 for similar spots in different fractions 6 0 .0 0 2 7 0 .0 0 5 1 7 _ - 0 .0 0 1 2 - - 0 .0 0 1 2 0 .4 8 were pooled together for calculation of 8 _ — 0 .0 0 2 9 0 .0 0 2 4 - 0 .0 0 5 3 2 .1 3 the percentage in the total. It can be seen 9 _ _ 0 .0 0 3 6 0 .0 0 4 0 - 0 .0 0 7 6 3 .1 5 that by this method the estimation was 10 _ _ 0 .0 1 1 0 0 .0 0 4 0 — 0 .0 1 5 0 6 .0 2 more accurate, owing to reducing the er­ 11 _ - 0 .0 0 6 6 - - 0 .0 0 6 6 2 .6 5 rors caused by overlapping of spots and 12 _ — 0 .0 0 6 2 0 .0 1 0 0 - 0 .0 1 6 2 6 .5 0 avoiding differential elution from paper, 13 _ - - 0 .0 0 9 0 0 .0 0 8 2 0 .0 1 7 2 6 .9 0 as each fraction represented a simpler 14 _ - - 0 .0 0 6 6 0 .0 1 7 1 0 .0 2 3 7 9 .5 1 1 3 .1 2 mixture of compounds of more or less 15 - - - 0 .0 0 4 1 0 .0 2 8 7 0 .0 3 2 8 similar adsorption to cellulose. (+) Cate- chin, (- ) epicatechin and polymeric pro- anthocyanidins represented the major polyphenols of cashew kernel testa. Discoloration of polyphenols Preliminary examinations showed that leucopelargonidin in other plant materials been noticed in other cases also (Bate- testa polyphenols can form a dark- tested, will prove it is a much more com­ Smith, et al., 1958; Donath, 1962; Chand­ colored complex with iron. Thus, soak­ monly occurring plant constituent than ler, 1964). That the formed pigment is ing good kernels in polyphenolic extract has been supposed (Swain, 1962). easily washed away by dilute acid further followed by dipping in ferric chloride so­ indicates that the darkening in cashew lution resulted in discoloration similar to Chromatographic studies show that nuts is probably due to polyphenol-iron normal darkening. In normal darkening, the main constituents are (+) catechin complex. the discoloration is more against the veins and (- ) epicatechin, which account for about 6 and 7.5%, respectively, of the on the periphery of the kernel, where REFERENCES there is close contact of the nut with the dried cashew kernel testa. Together they represent more than 40% of the total Bate-Smith, E.C. 1954. Leucoanthocyanins. 1. testa. Thus, the nature of the surface of Detection and identification of antho- the nut also is important. More deformed polyphenols. The polymeric proantho- cyanidins formed from leucoanthocyanins and undeveloped nuts are more liable to cyanidins form a little less than 40%. The in plant tissues. Biochem. J. 58: 122. Bate-Smith, E.C., Hughes, J.C. and Swain, T. get darkened by the above mechanism monomeric proanthocyanidins are repre­ 1958. Aftei>cooking discoloration in pota­ than are normal, healthy nuts. Dipping sented by 2 leucocyanidins and 2 leuco- toes. Chem. & Ind., 627. pelargonidins, but no leucodelphinidin. Chandler, B.V. 1964. Discoloration in pro­ the kernels in iron solution alone did not cessed cauliflower. Food Preserv. Quart. 24: give discoloration, since kernels them­ Kantamoni (1965) had recently reported 1 1 . selves do not contain polyphenols. the presence of gallic, caffeic and quinic Donath, H. 1962. Discoloration phenomena in sterile cauliflower packs. Nahrung 6, 473. Washing with dilute acid solutions re­ acids apart from a catechin and a leuco­ Farrer, C.E. 1935. The determination of iron in cyanidin. Gallic acid if present would biological materials. J. Biol. Chem. 110: moved the bluish-black patches in nor­ 685. mally darkened nuts; treatment with alka­ have occupied the position of spot 4; Govindarajan, V.S. and Mathew, A.G. 1963. li resulted in a darker brown color. The however, this spot showed full reactivity Polyphenolic substances of arecanut. 1. with vanillin-HCl, suggesting that it is en­ Chromatographic analysis of fresh mature ether extract of the acid wash showed nuts. Phytochemistry 2: 321. polyphenolic characteristics on treatment tirely composed of (+) catechin. Further­ Govindarajan, V.S. and Mathew, A.G. 1965. more, no gallic acid was detected by the Anthocyanidins from leucoanthocyanidins. with different spray reagents. Among the Phytochemistry 4: 985. different acids tried for removing the dis­ bicarbonate extraction of the ether frac­ Harborne, J.B. 1960. Chromatography of tion. Tests for caffeic acid (fluorescence) Flavonoids. Chromatographic Reviews, ed. coloration, hydrochloric and citric acids Lederer, M. Vol. 2, p. 105. Elsevier Publish­ were found to be the most convenient, and quinic acid also were negative. ing Company, Amsterdam. though each left behind a taste of its From the studies it is clear that the Jurd, L. 1962. Spectral Properties of Flavonoid black discoloration is due to the complex Compounds. “The Chemistry of Flavonoid own. Estimation of iron content showed Compounds” ed. Geissman, T.A., p . 107, that batches of darkened nuts have an formation between polyphenol and iron. Pergamon Press, London. Kantamoni, S. 1965. Isolation and identifica­ iron content of approximately 4.5 mg % While the polyphenol is derived from the tion of the constituents of cashewnut husk. compared to 3 mg % of good nuts (Far- testa, the source of iron cannot be so Leather Sci. 12: 396. clearly established. It may either be from Lakshminarayana, S. and Mathew, A.G. 1967. rer, 1935). Leucoanthocyanidins of sapota fruit. J. the nut itself, since nuts as a class are Food Sci. 32: 451. DISCUSSION found to be fairly rich in iron, or from Madhaven Pillai, K.S., Kedlaya, K.J. and IN CASHEWNUT testa, as in most other Selvarangam, R. 1963. Cashew seed skin as a external sources during processing. The tanning material. Leather Sci. 10: 317. plant products, leucocyanidin was found discolored nuts show a higher iron con­ Mathew, A.G. 1969. Some characteristics of the to be the predominant leucoantho- tent, which may indicate either absorp­ additional anthocyanidins formed during conversion of leucoanthocyanidins into cyanidin (Bate-Smith, 1954). However, tion of iron from external sources or a anthocyanidins. Phytochemistry 8: 677. leucopelargonidin was found to occur in higher iron content in nuts susceptible to Mathew, A.G. and Govindarajan, V.S. 1964. Polyphenolic substances of arecanut. II. larger amounts than leucodelphinidin. darkening. Discoloration involving a Changes during maturation and ripening. This observation, also the presence of heavy metal-like iron and polyphenols has Phytochemistry 3: 657. POLYPHENOLS OF CASHEW KERNEL TESTA-WZ

Mathew, A.G. and Lakshminarayana, S. 1969. Swain, T. 1962. Economic Importance of ents. J. Sci. Food Agr. 10: 63. Polyphenols of immature sapota fruit. Flavonoid Compounds: Foodstuffs. “The Ms. received 3/24/69; revised 8/19/69; accepted Phytochemistry 8: 507. Chemistry of Flavonoid Compounds, ” ed. 9/18/69. ______Roux, D.G. and Maihs, A.E. 1960. Selective Geissman, T.A., p. 513, Pergamon Press spray reagents for the identification and es­ Authors are indebted to E. Sankaran London. Nambudiri, V.S. Govindarajan and C.P. timation of flavonoid compound associated Swain, T. and Hillis, W.E. 1959. The phenolic with condensed tannins. J. Chromatog. 4: Natarajan for useful discussions at different constituents of Prunus domestica. 1. The stages of the investigation. 65. quantitative analysis of phenolic constitu-

R. A. JUNG K and W. W. M ARIO N

D epartm ent o f P oultry Science

Iow a State U niversity, Am es 50010

POST-MORTEM ISOMETRIC TENSION CHANGES AND SHORTENING IN TURKEY MUSCLE STRIPS HELD AT VARIOUS TEMPERATURES

SU M M A R Y—A study o f the physical changes associated w ith rigor m ortis in breast m uscle was tached to the carcass are restrained from undertaken to assess the factors th at m ay influence ultim ate tenderness. Isom etric tension changes shortening and tenderize more rapidly and shortening were m easured at tem peratures 2 —37°C . These changes were m easured w hile hold­ when held at the same temperature. ing the m uscle strips in a phosphate buffer, pH 7.2. Isom etric tension was m easured by transducers Recent work on bovine and porcine and recorded on a physiograph. A pattern o f tension developm ent and gradual relaxation has been species has used the isometric tension pat­ dem onstrated to occur post-m ortem in strips o f turkey breast m uscle held isom etrically. The tim e tern to follow the sequence of physical to m axim um tension developm ent occurs in 3.85 ± 0.19 h r and is n o t linearly related (P < .0 5 1 to changes in rigor (Jungk et ah, 1967; tem perature. The am ount o f m axim um tension developed averaged 25 g/cm 2 and was significantly Busch et ah, 1967; Schmidt et ah, 1968). IP < .05) related to tem perature. Relaxation to about 50% o f m axim um occurs in 18 hr. The The pattern has been shown to closely am ount o f shortening that occurs post-m ortem is linearly related IP < .0 1 ) to tem perature. No parallel the tenderness cycle and illus­ "cold shortening" o f turkey breast m uscle was evident. trates the tendency of a muscle to short­ en. This study was undertaken to eluci­ INTRODUCTION ness and requires longer aging time date the physical changes occurring post­ RAPID processing with emphasis on con­ (Scholtyssek et al., 1967). The rapid mortem in turkey muscle and to assess tinuous chilling is the trend in the poultry process of continuous chilling and im­ the influence of temperature on these industry. After evisceration the birds are mediate freezing without an intermediate changes. placed in a series of mechanical chillers aging period is not sufficient to ensure that reduce carcass temperature to 40°F adequate tenderness (Marion et ah, EXPERIMENTAL if desired. The processing step of over­ 1967). night aging in slush ice is rapidly being The biophysical or structural state of Source and sample preparation omitted; the live bird is transformed into muscle is quite relevant to meat tender­ The birds used in this study were from the frozen meat in a few hours. There are ness.. Lowe (1948) observed that poultry University research farm . They included 15 ma­ ture Bronze hens from a breeder flock and 25 several advantages in handling the prod­ muscles severed soon after death had Large W hite males. W ithin each group, the birds uct quickly; namely, the maintenance of higher shear values than those aged intact had similar genetic backgrounds, had been quality from a microbiological standpoint on the carcass. Lowe also suggested an raised under sim ilar conditions and did not vary and the economy afforded in time, labor association between shortened or con­ appreciably in weight. Exsanguination was ac­ and space requirements. However, there tracted muscles and toughness of meat. com plished by severing the jugular vein and the has been some concern expressed about More recent work using the bovine spe­ carotid artery, the birds being inverted in a the effect of rapid chilling and freezing cies has firmly established this relation­ metal cone during this process to minimize on meat-quality factors, specifically ten­ ship (Marsh, 1964; Herring et ah, 1965; struggling. No prior stunning or anesthetic was derness. Traditionally, it was thought that Marsh et ah, 1966; Howard et ah, 1968). used. Im m ediately after exsanguination, the skin was incised and a large portion o f the pec- adequate aging of the carcass must pre­ Conversely, muscles aged in a restrained toralis m ajor removed. Then strips of muscle cede freezing to insure tenderness. Al­ or stretched condition are more tender were dissected parallel to the direction of the though additional muscle relaxation has (Partmann, 1963; Herring et ah, 1967; muscle fibers. These strips were used in the ten­ been demonstrated to occur upon thaw­ Buck et ah, 1967). Goll et ah (1964) sion and shortening studies. ing (Klose et al., 1959; 1961a; 1961b), demonstrated distinctive tenderness pat­ Shortening measurements the tenderness of continuously chilled terns for muscles excised and those at­ M uscle strips were measured and placed in a birds has been questioned. The problem is tached to the carcass. The primary dif­ phosphate buffer, pH 7.2, previously adjusted most apparent with young turkeys of the ferences are those attributable to the to the desired temperature. The tim e required fryer-roaster classification. The turkey shortening that takes place post-mortem to dissect and measure the strips lim ited the fryer exhibits more variability in tender­ in the excised muscle. Muscles left at­ num ber of temperatures to be studied when an 144- JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

individual bird was used. Usually 6 tempera­ tures w ith 5 strips per temperature were used for each. The m axim um tim e from start o f ex- sanguination to placem ent o f the muscle strips in the environm ental temperature was 15 m in. The tabulated data are expressed as a percent­ age o f shortening from resting length. Tension measurements

The initial isom etric tension measurements on 7 birds were made using a device called an isometer, previously described by Jungk et al. (1967). This instrum ent was designed to meas­ ure post-m ortem tension developm ent in a mus­ cle strip held at constant length, thereby sim u­ lating the conditions of muscles aging on the Fig. 1—A typical isom etric tension pattern Fig. 3—E ffect o f environm ental tem perature on carcass. The sensitivity and capacity of this showing: A, the tim e to m axim um ; B, m ax­ the am ount o f tension developed in strips o f m ethod of measuring rigor has been increased im um tension and C, the relaxation phase. turkey breast m uscle held isom etrically. through the use o f isom etric myographs in con­ junction w ith an E & M Physiograph. Isom etric tension patterns were run sum ultaneously on 5 excised strips of pectoralis m ajor from each bird; 30 turkeys were measured using the iso­ m etric myographs. The muscle strips were im ­ mersed in a phosphate buffer, pH 7.2, to pre­ vent spurious tension pattern results due to drying o f the strips in air.

RESULTS & DISCUSSION Isometric tension pattern A pattern, shown in Figure 1, of ten­ sion development and relaxation can be demonstrated while holding muscle strips isometrically. There are 3 characteristics of this pattern that warrant elaboration. TEMPERATURE (°C>

They are a) the time lapse between death Fig. 2—The lack o f effect o f environm ental Fig. 4 —E ffect o f environm ental tem perature on

and maximum tension, b) the amount of tem perature on tim e to m axim um rigor. tension at maximum and c) the relaxation shortening o f excised strips o f pectoralis m ajor. phase. The time lapse before maximum ten­ The amount of tension that develops phase that meat becomes more tender. sion is important, because during this pe­ post-mortem is small when compared This conclusion is supported by the paral­ riod the muscle is attempting to shorten. with normal contraction. The maximum lel pattern of tenderness changes and In the animal, since the muscles are at­ strength of tetanic contraction of a mus­ isometric tension decline (Busch et al., tached and prevented from shortening, cle operating at normal muscle length is 1967). the antagonistic forces produce stiffness about 3 kg/cm2 (Huxley, 1958). The ten­ in the carcass. This stiffness is a common sion developed post-mortem was only Shortening experiments attribute of post-mortem muscle. Soon 24.78 + 1.45 g/cm2, or about 1% of the 10 turkeys were used in this facet of after death, if the muscles are severed or maximum strength for normal tetanic the experiment. Muscle strips were ex­ completely excised, they will shorten. In contraction. Temperature of aging envi­ cised and placed in environmental tem­ light of the known association between ronment does influence the amount of peratures ranging 3-37°C. The percent­ shortening and tenderness, it would seem tension that develops. This significant age shortening at each temperature was that conditions conducive to minimum (P < .05) relationship is shown in Fig­ noted. From the 263 individual observa­ shortening would allow the most tender ure 3. It is further substantiated by a tions, a linear regression analysis was pre­ meat. Also, the long aging might be re­ similar temperature effect on shortening pared. Figure 4 presents results of this duced if it were possible to know the ear­ of excised strips, which will be discussed analysis. Special emphasis was made to en­ liest time at which the muscles could be further. Tension development reflects the sure adequate sampling in the tempera­ excised without adverse effects on tender­ attempted shortening, similar to that ture range 10-20°C. It is in this range ness. which would occur on an intact carcass. that a cold-shortening phenomenon has In observing 121 strips from 30 tur­ The relaxation phase is a characteristic been demonstrated in both the mamma­ keys, we found the average time to maxi­ of the post-mortem isometric tension pat­ lian (Locker et al., 1963) and the avian mum tension to be 3.85 + 0.19 hr. The tern demonstrable at all temperatures. species (Smith et al., 1969). Our data do experiments were conducted at various The relaxation phase was evaluated by not show a cold-shortening effect, but temperatures 2-37°C. From the regres­ selecting a common time post-mortem for rather a significant (P < .01) linear rela­ sion analysis as shown in Figure 2, no sig­ comparison of the isometric tension tion between extent of shortening and nificant (P < .05) linear relationship be­ curves. The tension values at 18 hr were temperature. This is in conflict with re­ tween temperature and time to maximum used for this purpose and expressed as a sults of Smith et al. (1969). Our experi­ was observed. This is unusual in that percentage of the maximum. The mean ments were conducted on excised strips many of the other parameters of rigor value was 51.94 + 2.60% of maximum. of muscle over the temperature range (e.g., extent of shortening, loss of exten­ No consistent relationship with tempera­ 4-35°C. The previously reported work sibility, degradation of ATP and decline ture was found. This phase of the iso­ was done at 0 and 16°C on intact sec­ in pH) are closely associated with envi­ metric tension pattern bears further scru­ tions of pectoralis major, measured and ronmental temperature. tiny since it is during this relaxation then excised. Excision of muscles soon POST-MORTEM TENSION CHANGES IN TURKEY MUSCLE - 1 45 after death stimulates some shortening. Busch, W.A., Parrish, F.C. and Goll, D.E. 1967. Klose, A.A., Campbell, A.A., Pool, M.F. and Molecular properties of post-mortem mus­ Hanson, H.L. 1961b. Turkey tenderness in This mechanically induced shortening was cle. 4. Effect of temperature on adenosine relation to holding in and rate of passage not included in our data. Such shortening triphosphate degradation, isometric tension through thawing range of temperature. Poul­ parameters and shear resistance of bovine try Sci. 40: 1633-1636. would not normally be imposed on an muscle. J. Food Sci. 32; 390-394. Locker, R.H. and Hagyard, C.J. 1963. A cold aging carcass, whereas the thermal stimu­ Goll, D.E., Henderson, D.W. and Kline, E.A. shortening effect in beef muscles. J. Sci. lus would be. 1964. Post-mortem changes in physical and Food Agr. 14; 787-793. chemical properties of bovine muscle. J. Lowe, B. 1948. Factors affecting the palata- Final shortening measurements were Food Sci. 29; 590-596. bility of poultry with emphasis on histo­ taken at 5 hr. Most of the shortening was Herring, H.K., Cassens, R.G. and Briskey, E.J. logical post-mortem changes. Advan. Food 1965. Sarcomere length of free and re­ Res. 1: 232-235. completed within 3-4 hr. This is consist­ strained bovine muscles at low temperature Marion, W.W. and Goodman, H.M. 1967. Influ­ ent with results from the isometric ten­ as related to tenderness. J. Sci. Food Agr. ence of continuous chilling on tenderness of sion pattern, showing time to maximum 16; 379-384. turkey. Food Technol. 21: 89-91. Herring, H.K., Cassens, R.G., Suess, G.G., Brun- Marsh, B.B. 1964. Meat quality and rigor mor­ tension to be 3.85 hr. gardt, V.H. and Briskey, E.J. 1967. Tender­ tis. In “Carcass Composition and Appraisal In summary, the concept of post­ ness and associated characteristics of of Meat Animals.” CSIRO, Melbourne. Pa­ stretched and contracted bovine muscles. J. per No. 12. mortem aging of the carcass to allow ten- Food Sci. 32; 317-323. Marsh, B.B. and Leet, N.G. 1966. Studies in derization to occur is being changed. The Howard, R.D. and Judge, M.D. 1968. Compari­ meat tenderness. III. The effects of cold interest is now in controlling the state of son of sarcomere length to other predictors shortening on tenderness. J. Food Sci. 31; of beef tenderness. J. Food Sci. 33; 450-459. contraction of the muscle rather than just 456-460. Partmann, W. 1963. Post-mortem changes in allowing sufficient time for it to relax. Huxley, H.E. 1958. The contraction of muscle. chilled and frozen muscle. J. Food Sci. 28: Sci. Am. 199: 66-72. 15-17. New processing methods are being de­ Jungk, R.A., Snyder, H.E., Goll, D.E. and Schmidt, G.R., Cassens, R.G. and Briskey, E.J. vised that will minimize muscle contrac­ McConnell, K.G. 1967. Isometric tension 1968. Development of an isotonic and iso­ tion and maximize relaxation. However, changes and shortening in muscle strips dur­ metric rigorometer. J. Food Sci. 33: ing post-mortem aging. J. Food Sci. 32: 239-241. they must be amalgamated with condi­ 158-161. Scholtyssek, S. and Klose, A.A. 1967. Sources Klose, A.A., Pool, M.F., Wiele, M.B., Hanson, of variability in turkey tenderness. Poultry tions maintaining the other quality fac­ Sci. 46: 936-938. tors of meat. H.L. and Lineweaver, H. 1959. Poultry ten­ Smith, M.C., Judge, M.D. and Stadelman, W.J. derness. I. Influence of processing on ten­ 1969. A “cold shortening” effect in avian REFERENCES derness of turkeys. Food Technol. 13: muscle. J. Food Sci. 34: 42-46. 20-24. Ms. received 9/12/69; accepted 10/16/69. Klose, A.A., Campbell, A.A., Hanson, H.L. and Buck, E.M. and Black, D.M. 1967. The effect of Lineweaver, H. 1961a. Effect of duration Journal Paper No. J-6352 of the Iowa Agri­ stretch tension during rigor on certain physi­ and type of chilling and thawing on tender­ culture and Home Economics Experiment Sta­ cal characteristics of bovine muscle. J. Food ness of frozen turkeys. Poultry Sci. 40; tion. Project 1696. Sci. 32; 539-543. 683-688. Kl SOON RHEE and L. J. BRATZLER Department of Food Science Michigan State University, East Lansing, Michigan 48823

BENZO(A)PYRENE IN SMOKED MEAT PRODUCTS

SUMMARY-A procedure was established for analyzing benzo(a)pyrene (BaP) in smoked meat was first stuffed into a casing and cooked in products, which involves extraction, solvent partition, column chromatography and spectropho- water in a steam-jacketed kettle u n til the inter­ tofluorometric measurement. Studies were then made as to the effect o f artificial casings on the nal temperature of bologna reached 145°F content of BaP, penetration o f BaP through a meat product in smoking, possible presence o f the (usual cooking tim e was 2.5 hr). The cooked bologna was showered w ith cold water to an hydrocarbon in unsmoked meats and distribution of the compound after cooking smoked meat. internal tem perature o f 90°F, and processed in Bologna was used in most experiments except for the one on cooking effect, in which bacon was a smokehouse after removing the casing from used. A rtificial cellulose casings caused a noticeable reduction in the concentration o f BaP. Regard­ the experim ental sticks. less o f whether the smoke was generated in a small-scale or in a largerscale generator, bologna The bologna, w ith or w ithout casings, was smoked in casings had much less BaP than that smoked without casings. BaP was concentrated in processed in an air-conditioned smokehouse the outer part o f bologna. In smoking with or without casings, the hydrocarbon did not penetrate having abcut 5 air changes per m inute. Smoke farther, in most cases, than 1.4—1.6 mm from the surface o f the bologna. Several unsmoked meats was generated from sm oldering dampened hard­ o f beef and pork were analyzed and none seemed to have endogenous BaP, since the detected wood (predom inantly maple) sawdust by the amount was similar in magnitude to that of the solvent control. More than half the amount o f BaP smokehouse generator or from dry sawdust by in smoked bacon was found in the fat drippings following cooking. an experim ental generator (Porter et al., 1965) connected to the smokehouse. A temperature INTRODUCTION smoked fish and of other polycyclic com­ o f 140°F and 23% relative hum idity were m ain­ pounds in several liquid-smoke prepara­ tained for the first 2 hr o f processing. The tem ­ PRESERVATIVE action, flavor and color tions and bacon. Although using an perature was then raised to 160°F w ith a rela­ have been considered as the major effects experimental smoke generator and dry tive hum idity o f 31% until the bologna reached of smoking food (Draudt, 1963). With sawdust, the authors have found BaP in an internal temperature of 128°F. The smoke­ the development of other methods of smoke of hardwood (predominantly house temperature was then maintained at preservation, foods now are smoked maple) sawdust (Rhee et al., 1968). 160°F by injecting high-pressure steam, w hich mainly for their sensory qualities. resulted in 69% relative hum idity until an inter­ The presence of carcinogenic poly­ The present study was undertaken to nal bologna temperature of 152°F was at­ cyclic hydrocarbons in wood smoke has explore some aspects of smoked meat tained. The bologna was showered w ith cold been recognized (Berankova et al., 1953; products in relation to the content of water u n tl the internal temperature was re­ BaP. BaP was chosen as an “arbitrary duced to 90°F, then stored in a 36°F cooler Hollenbeck et al., 1963; Tilgner, 1958; fo r a few days u n til sampled. Gorelova et al., 1961). During the last indicator” of the presence of carcinogenic polycyclic hydrocarbons based on the An extreme outside layer of 1.4-1 .6 m m decade, interest has been increasing in the was removed from the sticks o f bologna fo r BaP observation with cigarette smoke that analysis of smoke and smoked food prod­ analysis and this layer designated as the “ A ” there is a correlation between the concen­ ucts for carcinogenic hydrocarbons. Dif­ layer. For the experim ent in which penetration ferent levels of carcinogens have been tration of BaP and “tar” tumorigenicity o f the hydrocarbon was studied, an inner layer found in smoked foods, the amounts (Wynder et al., 1968). In addition most, ( 1 . 4 - 1 .6 mm) next to the “ A ” layer was also being related to many factors, including if not all, other polycyclic hydrocarbons removed and called the “B” layer. Bologna method of smoke generation, temper­ are pyrosynthesized in wood-smoke pro­ samples were ground several times through a ature of combustion and oxidation, air duction at a favorable generation temper­ 1.5-mm plate o f a meat grinder. ature. Bologna was used as a typical supply, length of smoke ducts and den­ BaP analysis sity and temperature of the smoking smoked-meat product in this study, unless otherwise indicated, due to its suit­ Extraction. A 150-g sample o f the ground operation (Tilgner, 1968). Because of the bologna was thoroughly mixed with 300 g of ability for our study and its homogeneity historical origin of the problem—the anhydrous Na S and extracted three times for sampling. 2 04 discovery of the first carcinogenic constit­ with benzene (300 m l each) and twice with uent of coal tar, benzo(a)pyrene (BaP) or EXPERIMENTAL methanol (300 ml each) by refluxing for 30 m in after each extraction. The aqueous phase 3, 4-benzopyrene-much attention has S o lv e n t s been focused on this polycyclic hydro­ o f the benzene extract, if any, was separated. carbon. A ll solvents except cyclohexane and nitro- The m ethanol extract was filtered over a suffi­ methane (practical grade) were reagent grade. cient am ount o f anhydrous Na 2 S O 4. B e n z e n e It has been an open question whether Each solvent was purified by treating w ith acti­ and m ethanol were removed from the extracts any carcinogenic hydrocarbon can be vated charcoal follow ed by distillation (Crosby under reduced pressure using a flash evaporator found in the smoke generated under pro­ et al., 1966). connected to an aspirator, at a w ater bath tem ­ perature of about 50°C. Concentration of a duction conditions in the United States. Preparation o f sm oked bologna The most common method currently solution or evaporation o f any solvent was as used in the United States for producing A bologna emulsion was prepared w ith the just described, unless otherwise indicated. smoke for food processing is burning follow ing ingredients, in pounds: regular pork Solvent partition. Polynuclear compounds dampened sawdust in a batch operation trimmings, 9.2. cutter beef trimmings, 9.2. were separated from the bulk o f aliphatic m ate­ nonfat dried m ilk, 0.9. and ice; 5.9. And in (Draudt, 1963). Genest et al. (1964) ana­ rial by solvent partition (Lijinsky et al., 1964). grams: sodium nitrate, 1.4. sodium nitrite, 1.4. lyzed smoked fish, frankfurters and cheese The fatty extract residue was dissolved in 100 sodium ascorbate, 4.0. sugar, 32.6. and salt m l o f hexane and shaken w ith 2, 100- m l p o r ­ for the presence of BaP by a detection (NaCl), 152.0. method they developed with a reported tions of nitrom ethane. The nitrom ethane layer The em ulsion was stuffed into No. 5 Tee-Pak was separated and concentrated to dryness (I). sensitivity of 10-50 ppb; BaP was not Clear Zip cellulose casing (Tee-Pak, Inc.). The extraction of polynuclear arom atic mate­ detected. Lijinsky et al. (1965) reported When precooking was necessary to smoke the rial from the hexane solution was completed the presence of trace amounts of BaP in bologna em ulsion w ithout casings, the em ulsion w ith 3, 40-m l portions of dim ethylsulfoxide.

146-JOURNAL OF FOOD SCIENCE-Volume 35 (1970) BENZO(A)PYRENE IN SMOKED MEAT PRODUCTS - 147

To this extract were added 150 m l of hexane and 280 ml of water; the hexane layer was separated, combined w ith the nitromethane extract residue (I) and concentrated to dryness (II). For samples of very high fat content such as bacon, the amount of solvent or aqueous m edium was doubled. Further removal of impurities by washing. The extract (II) was dissolved in 100 ml of cyclohexane and washed with 110 m l o f 90% methanol. The methanol phase was again shaken w ith 100 m l o f fresh cyclohexane and the 2 cyclohexane extracts com bined (G rim m er et al., 1965). The cyclohexane solution was washed w ith 2 N H2S04 (3 tim es), water, 2 N NaOH (3 tim es) and water (Cooper et al., 1955) and dried over anhydrous Na2S04. The washed cyclohexane solution was concentrated to dryness and made to a volume of 2 ml with cyclohexane for colum n chrom atography. Column chromatography. Silicic acid (Mal- linckrodt) was washed w ith acid and activated as described by Grimm er et al. (1965), then partially deactivated by adding 8% (v/w ) dis­ tilled water. The extract (ca. 2 m l) was applied Fig. 2 —Em ission spectra in cyclohexane at ex­ to a silicic acid column (12 g, 15 mm i.d.) citation wavelength 381 mg. Authentic BaP packed with cyclohexane. The column was (— ): 2 gg/m l, M M = 0.3; the cyclohexane ef­ eluted w ith cyclohexane. The first 300 m l efflu­ Fig. 1—Em ission spectra o f cyclohexane col­ fluent (300 m l) o f a bacon sample (100 g) ent was collected for BaP determination. um n effluent fractions o f 12 gg BaP at excita­ (— ); 1 m l final volum e, M M = 0.1. During elution the colum n was protected from tion wavelength 381 m g. A : 1st 50-m l fraction, li g h t . MM = 0.1; B: 2nd 100-m l fraction, M M = 0 . 1 ; C: 3rd 100-m l fraction, M M = 0.03; D: 4th

Fluorometric determination of BaP 100-m l fraction, M M = 0.01; E : 5th 70-m l frac­

The cyclohexane effluent was concentrated tion, M M = 0.01. The final volum e o f each frac­ to near dryness and quantitatively transferred tion fo r spectrum determ ination was 8 m l. with unknown impurities, presumably orig­ to a graduated test tube using cyclohexane and inating from silicic acid. Thus, about 300 the final volume reduced to 1-3 m l under a ml of cyclohexane effluent were collected stream o f nitrogen. The emission spectrum was determined in cyclohexane at the excitation from the column for BaP elution in all wavelength 381 m g . An Am inco-Bowm an spec- experiments. Figure 2 presents an emis­ trophotofluorom eter was used w ith the follow ­ sion spectrum of a smoked bacon sample ing settings: sensitivity 50, slit arrangem ent No. smoke, only anthracene has an excitation which passed through the above over-all 2, meter m ultiplier (M M ) 0.1-0.03 and photo­ wavelength beyond 370 m^z, but this procedure of BaP analysis. Recovery of tube IP 21. The concentration of BaP was compound is one of the relatively volatile added BaP through this procedure was calculated from the 403 mju peak by the “ base hydrocarbons and can be mostly vol­ satisfactory, resulting in about 80% when line” m ethod (Cooper, 1954). atilized, with the quantity present in 1.0-4.5 jug of the hydrocarbon were smoked foods, during the steps of evap­ added to 1 50 g of bologna emulsion. RESULTS & DISCUSSION oration to dryness. There is no drying toss Analysis of BaP in meat products of BaP (Grimmer et al., 1965). There­ BaP in unsmoked meats The procedure for analyzing BaP has fore, determination of a BaP emission Attempts were made to determine been modified from that used in the spectrum at excitation wavelength 381 whether the unsmoked bologna emulsion previous experiments of wood smoke m u is unlikely to be interfered with by or meat mixture itself contained BaP. A (Rhee et al., 1968). Fats were removed the other polycyclic hydrocarbons of meat mixture was prepared with half from the meat extract by solvent parti­ wood smoke. Taking these facts into pork and half beef trimmings as used for tion, but the TLC step for removing the consideration, the fractionation step of bologna formulation. In analyzing several oily substance(s) in wood smoke of alumina column chromatography could samples, neither meat mixture nor unknown identity and the final alumina also be eliminated. Thus, the over-all bologna emulsion appeared to have any column chromatography were omitted. procedure for determining BaP in a meat endogenous BaP, because the indicated Since the amount of smoke deposited on product includes extraction with benzene amount of 0.013—0.027 £tg/l 50 g sample the smoked product was relatively small and methanol; fat removal by solvent or 0.09—0.18 ppb (on wet weight basis) as compared with the smoke concentrate partition using hexane, nitromethane was similar in magnitude to that of the we had previously, the oily material and dimethylsulfoxide; washing the re­ solvent control (0.014 pzg). associated with smoke did not cause any sulting extract with methanol, acid and significant problem, making elimination alkali; silicic acid column chromatog­ Effect of casings on BaP permeation of the TLC step possible. The polycyclic raphy and finally fluorometric determi­ The effect of use of casings on the con­ hydrocarbons present in greater quan­ nation of the column effluent. With 1 2 pig tent of BaP in smoked bologna has been tities found in the hardwood sawdust of authentic BaP, collection of 250 ml of explored. For this study, bologna emul­ smoke are noticeably volatile as shown by cyclohexane from the time of sample sion was precooked in casings, in 2 or 4 evaporation to dryness and much can be introduction to the column has proved sticks at one trial, and the casings of 1 or lost during the many steps of evaporation sufficient for complete elution of the BaP 2 precooked bologna sticks were removed or concentration to dryness, at a bath as shown in Figure 1, since there was no before smoking. The other half of cooked temperature of 50°C or higher (Grimmer 403 mju peak from the 4th or “D” frac­ bologna was smoked with casings. Using et al., 1965; Rhee et al., 1968). In addi­ tion. The illustration also shows that fur­ artificial cellulose casings (No. 5 Tee-Pak tion, of the hydrocarbons found in wood ther elution can contaminate the effluent Clear Zip), it was found that the casing 148-JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

impeded the transfer of BaP in smoking BaP content from commercial bologna, BaP values of the two brands were bologna to a great extent (Tables 1 and since the processing conditions may vary. nearly similar for both the “A” layer and 2). The determination of BaP content of a whole slice as shown in Table 3. As Since much heavier smoke was pro­ commercial bologna should give some compared with our products in casings duced by the smokehouse generator than practical information and should allow (Experiments 1,2 and 3 of Table 1 ), these by the experimental generator, the for comparisons with our products. “A” layer values are much smaller. amount of hydrocarbon absorbed by the Two well-known brands of commercial Although the processing conditions of the “A” layers of bologna sticks differed ac­ bologna were chosen. The outside appear­ 2 commercial bands were not known, it is cordingly. In all cases, regardless of ance (casing) of the two brands indicated reasonable to assume from the results of whether the smoke was generated in a very little smoking and the sliced bologna BaP analysis that these commercial prod­ small-scale or in a larger-scale generator, did not have an easily detectable smoke ucts were processed with little, if any, bologna smoked in casings had much less aroma. BaP content of the “A” layer and smoking. BaP values of whole slices were BaP. The results expressed are based on of a whole slice was determined. actually similar to those of the solvent dry weight of bologna to balance the dif­ ferences in moisture content of the samples. The cellulose casing itself did not contain BaP.

It may be noted that whereas a-cellu- Table 1 —Benzo(a)pyrene content o f "A " layers o f bologna sm oked

lose is fundamentally crystalline, regen­ w ith and w itho ut casings. erated cellulose (used for artificial cellu­ lose casings) is fundamentally a gel which B a P in “ A ” l a y e r shares the properties of other gelatinous ( p p b , o n dry w t basis) materials which are hydrophilic (Tuwinner, Experim ents W ith casing W ithout casing 1962). As the permeability of membranes Experim ent 1 (smokehouse generator) 4 . 6 11.0 appears to be related directly to swelling Experim ent 2 (smokehouse generator) 4 . 5 12.1 and imbibition, it is reasonable to find Experim ent 3 (smokehouse generator)1 1.2 6.6 that BaP, which is hydrophobic and has a 1 .7 8 . 4

high molecular weight, still penetrates in Experim ent 4 (experim ental generator) 0 . 5 4 . 8 small quantities through the hydrophilic 0 . 9 5 . 2 cellulose casing. This is especially true when the BaP concentration outside the 'R esults o f Experim ent 3 are o f lower values because the thickness casing, i.e., in the smokehouse, is high. of each layer was accidentally 2.2-2.4 mm instead of 1.4—1.6 m m, Elevated temperatures and humidities the routine thickness. would also help the penetration by caus­ ing the cellulose casing to swell. Unlike BaP, Simon et al. (1966) reported that cellulose casings did not act as a barrier to smoke components such as acids, car­ Table 2-P enetration o f benzo(a)pyrene through bologna layers. bonyls and phenolic compounds which B a P contribute to smoked flavor and color. (ppb, on dry w t basis) Use o f casing

Depth of BaP penetration Experim ents i n s m o k in g “A” layer1 “B” layer

Since a preliminary study with a bacon Experim ent 1 (experim ental generator) W it h 0 .5 0 . 5 sample showed that whole bacon had 0 . 9 0 . 4 much less BaP than outer bacon layers (2 W i t h o u t 4 . 8 0 . 5 layers taken from flesh side), further 5 . 2 0 . 5 exploration was made with smoked Experim ent 2 (smokehouse generator) W i t h 1.2 0 . 4 bologna to determine the extent of BaP 1 .7 0 . 5 penetration through the product in smok­ W i t h o u t 6.6 0 . 9 ing. 8 . 4 0 . 9 The BaP content of the “A” layer (the 1 The data on “ A ” layer of this table and those o f Experim ents 3 and 4 o f Table 1 are identical. outer layer) of bologna was compared with that of the “B” layer (next to the “A” layer). After smoking with or with­ out casings and holding for a few days, there was practically no penetration of the hydrocarbon beyond the “A” layer (Table 2). This is not surprising for a Table 3 —Benzo(a)pyrene in com m ercial bologna. compound of this molecular weight. Phe­ B a P (ppb, on dry w t basis)1 nolic compounds in wood smoke have B o lo g n a

been found to penetrate a considerable S a m p le s Ingredients “ A ” la y e r W hole slice depth into bologna (Bratzler et al.; 1969). Although Tilgner (1968) indicated B r a n d A Beef and pork, water, nonfat dry m ilk, dex­ 0 . 6 5 0 . 2 6 that some 38% of BaP of smoked trose, salt, corn syrup, flavorings, sodium eryth- products diffused into the inner parts, orbate, sodium nitrate, sodium nitrite detailed information was not given. B r a n d B Beef and pork, water, beef hearts, salt, dried 0 . 7 3 0 . 3 0 BaP in commercial bologna corn syrup, dextrose, flavoring, sodium eryth- It was thought that the bologna pre­ orbate, sodium nitrate, sodium nitrite pared in our laboratory might differ in 1 Average o f tw o sticks. BENZO(A)PYRENE IN SMOKED MEAT PRODUCTS - 149 control if expressed on wet weight basis Table 4 —Benzo(a)pyrene in the cooked bacon vs. f a t d r ip .

(0.017—0.021 /ig/150 g sample); these W e ig h t figures were also very close to those of originating from M g B a P unsmoked meats mentioned previously. 500 g uncooked bacon in given weight Distribution of BaP after cooking bacon S a m p le s 1 ( g ) in the left colum n Cooked bacon 2 3 7 . 2 0 . 5 9 Since polycyclic hydrocarbons are Bacon-I anterior soluble in most fat solvents, and smoked Fat drippings 1 7 9 .4 0 6 . 3 Cooked bacon 2 5 3 . 8 0 . 4 2 products having a fatty flesh had a higher Bacon-I posterior content of BaP than did lean products Fat drippings 1 6 3 . 2 0 . 4 7 (Tilgner, 1968), tests were made to deter­ Uncooked bacon 5 0 0 . 0 1 . 8 5 mine the loss of BaP from bacon during Bacon-II posterior Cooked bacon 1 5 1 . 6 0 . 7 0 cooking. Cooking bacon without adding Fat drippings 2 4 4 . 6 0 . 9 4 any water, either in a fry pan or in an 1 Bacon—I: Bacon slab smoked in the laboratory for about 8 hr in the first sm oking and re­ oven, is actually a process of fat ren­ smoked for 3 hr; skin removed afterward for slicing. Bacon-II: Comm ercial bacon slab resmoked dering, and much BaP in smoked, in laboratory for 7.5 hr. uncooked bacon might be in the fat drip­ pings. In the first experiment, a bacon slab was smoked in our laboratory for about 8 hr and resmoked for an additional 3 hr maintained under conditions similar to cooking other foods. By discarding the Kohlenwasserstoffen. J. Chromatog. 20: 89. those for bologna processing, and the skin Hollenbeck, C.M. and Marinelli, L.J. 1963. removed. Thin slices (about 2-mm thick­ fat, the amount of BaP ingested from VIII. Some chemical constituents of smoke ness) of the anterior and of the posterior smoked bacon can be reduced at least flavoring. 15th Res. Conf. Res. Council American Meat Inst. Found., University of half of the smoked bacon slab were anal­ half. Chicago, p. 67. yzed. If the second experiment, a com­ Lijinsky, W. and Shubik, P. 1964. Benzo(a)- pyrene and other polynuclear hydrocarbons mercial skinless smoked bacon slab was REFERENCES in charcoal-broiled meat. Science 145‘. 53. obtained and resmoked in our laboratory Berankova, Z. and Sula, J. 1953. Isolation and Lijinsky, W. and Shubik, P. 1965. Polynuclear for 7.5 hr. Taking the posterior half, identification of 3,4-benzopyrene in the hydrocarbon carcinogens in cooked meat wood tar of a smokehouse. Casopis Lekaru and smoked food. Ind. Med. Surg. 34! 152. odd-numbered slices were used for cook­ Ceskych. 92: 195. [Chem. Abstr. 49: 592 c. Porter, R.W., Bratzler, L.J. and Pearson, A.M. ing and even-numbered slices for analyz­ 1955.] 1965. Fractionation and study of com­ Bratzier, L.J., Spooner, M.E., Weatherspoon, pounds in wood smoke. J. Food Sci. 3 0 .‘ ing uncooked bacon. Cooking was done J.B. and Maxey, J.A. 1969. Smoke flavor as 615. in a broiler pan at the broiling tempera­ related to phenol, carbonyl and acid content Rhee, K.S. and Bratzler, L.J. 1968. Polycyclic of bologna. J. Food Sei. 34: 146. hydrocarbon composition of wood smoke. ture of an electric oven until the bacon Cooper, R.L. 1954. The determination of poly­ J. Food Sci. 33: 626. became crisp. After cooking, the fat drip­ cyclic hydrocarbons in town air. Analyst Simon, S., Rypinski, A.A. and Tauber, F.W. pings were drained and the broiler pan 79: 573. 1966. Water-filled cellulose casings as model Cooper, R.L. and Lindsey, A.J. 1955. S,4- absorbents for wood smoke. Food Technol. rinsed with hexane. The hexane washing Benzopyrene and other polycyclic hydrocar­ 20: 1494. was combined with fat drippings and the bons in cigarette smoke. Brit. J. Cancer 9: Tilgner, D.J. 1958. Smoke-curing processes. 304. Fleischwirtschaft 10, 649. [Chem. Abstr. solvent evaporated by a flash evaporator. Crosby, D.G. and Aharonson, N. 1966. Fluores­ 53: 1577e. 1959.] Both the cooked bacon and fat drippings cent solvent impurities as chromatogram ar­ Tilgner, D.J. 1968. Carcinogens in food. Food tifacts. J. Chromatog. 25: 330. Manufacture, June, p. 37. were weighed. The crisp cooked bacon Draudt, H.N. 1963. The meat smoking process: Tuwinner, S.B. 1962. “Diffusion and Mem­ was thoroughly crushed and ground in a a review. Food Technol. 12; 1557. brane Technology.” Chap. 10. Reinhold Genest, C. and Smith, D.M. 1964. A simple Publishing Company, New York. mortar, since the meat grinder was inef­ method for the detection of benzo(a)pyrene Wynder, E.L. and Hoffmann, D. 1968. Experi­ fective in grinding cooked bacon. in smoked foods. J. Assoc. Offic. Agr. m ental tobacco carcinogenesis. Science Chemists 47: 894. 162: 862. Results are shown in Table 4. BaP in Gorelova, N.D. and Dikun, P.P. 1961. Content Ms. received 7/11/69; revised 9/24/69; accepted smoked bacon was found in both cooked of 3,4-benzopyrene in fish smoked with the 9/26/69. bacon and fat drippings. Starting with aid of different types of smoke generators. Vop. Onkol. 7: 71. [Chem. Abstr. 56: Journal article 4788 Michigan Agricultural 500 g of uncooked bacon, the fat drip­ 13307.] Experiment Station. pings had more BaP than cooked bacon. Grimmer, G. and Hildebrandt, A. 1965. Koh­ This investigation was supported in part by lenwasserstoffe in der Umgebung des Mens­ Public Health Service Research Grant No. UI This may have some practical implication. chen. I. Mitt, eine Methode zur simultanen 00130-06 from the National Center for Urban In many homes, bacon fat is reused for Bestimmung von dreizehn polycyclischen and Industrial Health. ARTHUR B. KARPAS, WALTER L. MYERS and DIEGO SEGRE Department of Veterinary Pathology and Hygiene College of Veterinary Medicine, University of Illinois. Urbana, Illinois 61801

SEROLOGIC IDENTIFICATION OF SPECIES OF ORIGIN OF SAUSAGE MEATS

SUMMARY—A partially purified immunoglobulin G (IgG) solution prepared from the serum of a of Agriculture. All-sheep meat and all-horse species to be tested was heated to the specifications for sausages. The resulting supernatant fluid meat sausages were prepared in our laboratory was decanted and the precipitate washed with saline and used to immunize rabbits. The superna­ and stored at 20°C until needed.

tant fluid was used to sensitize tanned sheep red blood cells. The immune serum was rendered Antigen preparation monospecific by absorptions with heterologous, heated IgG precipitates. A sample o f monospecific Crude immunoglobulin G (IgG) was pre­ immune serum was absorbed with a washed homogenate o f sausage. Aliquots o f the monospecific pared from the sera o f the 5 species. A 40-m l immune serum, both untreated and sausage absorbed, were tested with cells sensitized with the sample o f each serum was precipitated 3 times homologous heated IgG supernatant fluid. A significant reduction o f titer by sausage absorption in a 50% saturated am m onium sulfate solution. indicated that the sausages contained the meat homologous to the immune serum. The final precipitate was redissolved to 0.5 its original volum e in 0.85% NaCl and dialyzed at INTRODUCTION denatured horse globulin gave no excep­ 4°C against 2 changes of pH 7.2 borate-buff­ tional cross-reactions. Species-specific ered saline, followed by one change o f the ap­ IDENTIFICATION OF the species of ori­ anti-heated human gamma globulin has propriate buffer for chromatography. After gin of heated ground meat samples is an been produced by Henney et al. (1968) dialysis, the crude globulin was divided into 2 obvious necessity for effective control of and Hirose et al. (1967). Both groups equal aliquots and stored at -20°C until further such products as frankfurters. While agree with Grabar (1958) who found that fractionated. adulteration of unheated ground meats IgG was separated from the crude globulin denatured molecules achieve a new and by column chrom atography on DEAE cellulose can be easily detected by serologic meth­ characteristic specificity which is well- ods using species-specific immune sera, (W illiam s et al., 1968). The conditions o f chro­ defined rather than random as proposed m atography yielding the best balance o f purity application of similar methods to heated earlier. and yield for each globulin species were deter­ meats has been only partially successful. We made this investigation to demon­ m ined. Phosphate buffers used were: Serologic testing for adulteration of strate that a simple, purified, heat-dena­ For bovine and ovine IgG 0.0175 M pH 6.8 heated meats has been limited almost ex­ tured serum protein used in conjunction For avian IgG 0.15 M pH 6.8 clusively to the precipitin technique. Yet with the sensitive tanned cell hemaggluti­ For equine IgG 0.0175 M pH 7.8 meat proteins heated in sausage prepara­ nation (HA) test could be used to identi­ For porcine IgG 0.0175M pH 7.5 tion coagulate and become insoluble, so fy the species of origin of heated meats. The colum n effluent was m onitored w ith that they cannot be used in the precipitin Immunoglobulin G (IgG), the second a continuous flow recording ultraviolet (A = test. Instead, a saline extract of the heat­ most abundant serum protein, was found 280 nm) spectrophotometer (Gilson Medical ed meats has been used, usually in combi­ at a concentration of 0.16 mg per gram of Electronics Co., M iddleton, W isconsin). Only nation with an anti-heated whole serum. wet rabbit muscle tissue. Heating precipi­ the first m aterial to cause deflection of the re­ The homology of such a system is ques­ tates approximately 70% of the IgG, trap­ corder needle was collected and concentrated tionable because the immunogen is insol­ ping it within the heated meat sample. by pressure ultrafiltration (Am icon Corp., Cam­ uble heated proteins and the test antigen bridge, Massachusetts) to 10 to 20 m g/m l. The Approximately this amount (0.10 mg/g) concentration of the IgG solutions diluted as is the soluble portion of heated meat. of residue was presumed to be present in Further, the whole serum immunogen required ir 0.1N NaOH was determ ined by spec­ all heated meat samples and was consid­ trophotom etry at A = 280 nm using the extinc­ contains many protein types with varying ered sufficient to make heated meat sam­ tion coefficient o f 1.5 density units/m g/m l. The degrees of species specificity. These fac­ ple absorptions effective in reducing the extinction coefficient for the five species o f IgG

tors led to the failure of earlier specificity titer of an homologous anti-heated-IgG- was assumed to be the same as that for rabbit tests. precipitate. The antiserum was tested for IgG found by Steiner et al. (1966). Nuttall (1904) used the precipitin test titer reduction by micro-HA, using on blood samples representing 500 animal tanned red blood cells sensitized with the Tests on antigen species and found specificity to be rela­ homologous heated IgG supernatant Purity of the IgG preparations was tested tive rather than absolute. However, work­ fluid. A significant reduction of titer by by im m unoelectrophoresis on glass microscope ers continued to apply the precipitin slides (Scheidegger, 1955). The precipitin lines meat absorption indicated that the meat were developed with homologous rabbit anti­ techniques to the problem of species sample was homologous to the antiserum. whole serum. specificity until Boy den (1953) showed MATERIALS & METHODS For prelim inary tests of IgG interspecies more conclusively that such techniques cross-reactivity we employed im m unodiffusion were inadequate for measuring serological M a t e r ia ls in agar on glass m icroscope slides (H artm ann et correspondence. W hole serum samples were obtained from al., 1957). A central w ell and 6 peripheral wells Early studies on identification of heat­ the follow ing species: chicken (avian, designat­ were cut in the agar in a regular (hexagonal) ed meat, heated meat extracts or heated ed A ), cow (bovine, B), horse (equine, E), sheep pattern. The central well was filled either w ith serum proteins involving anti-native (ovine, O) and swine (porcine, P). W hole sera rabbit origin anti-native IgG or w ith rabbit ori­ whole serum were unsuccessful (Furth, and serum fractions were stored at —20°C until gin anti-heated IgG precipitate. C ircum ferential 1925; Pinto, 1961; Ginsberg, 1948). Tests u s e d . wells contained either native IgG or heated IgG using heated immunogens found greater Rabbits o f various breeds, ages and weights supernatant fluid o f the 5 species investigated. success. Furth (1925) reported that a se­ and o f both sexes were obtained from com mer­ Heat denaturation o f IgG rum sample heated to 100°C elicited an cial sources fo r production o f antisera. Imm unoglobulins from the 5 species were Sausages o f the following meat types: all heat-denatured according to the specifications antibody specific for heated serum. chicken, all beef, all pork, chicken and pork, for the m axim um tim e and tem perature used in Rothen et al. (cited by Landsteiner, 1945) beef and horse, and a beef, sheep and pork m ix­ sausage preparation. Globulin aliquots in found that an immune serum for heat- ture were obtained from the U.S. Department screw-cap tubes were placed in a water bath at

ISO-JOURNAL OF FOOD SCIENCE-Volume 35 (1970) SEROLOGIC IDENTIFICATION OF SAUSAGE MEAT ORIGIN - 1 5 1

I. ANTIGEN PREPARATION

PRECIPITATION DIALYSIS CHROMATOGRAPHY

Whole Serum

Qualitate, Quantitate

Quantitate, Dilute Fluid Globulin Centrifuge

Wash 3 Times Precipitated: > , Resuspend, Quantitate Globulin DENATURATION

n. ANTIBODY PRODUCTION W. PREPARATION OF MONOSPECIFIC SERUM ■ Test Serum by Microtiter to Determine Reactivity to Homologous and Heterologous Globulins S h e e p ensitize Tanned Sheep RBCs I ith Globulins of Various Species Absorb Serum with Heterologous —Globulins and Retest for Monospecificity PREPARATION OF SAUSAGES Y. SAUSAGE MEAT

ogenize Sausage Wash 5 Times TESTING ale in Soline Resuspend Residue ^ Absorb Monospecific Serum with Sausage Residue and Retest by Microtiter

ction of Specific Titer by Sausage Residue Indicates Presence of Meat Type in Question.

Fig. 1-Flow sheet for preparation o f monospecific sera and sausage meat testing.

70°C for 150 m in, then removed and centri­ tions of 10 mg of IgG without adjuvant were Antibody testing procedures fuged at 1,50G x g for 5 m in and the superna­ given, or 20 to 30 m l of blood were collected To determ ine the antibody titers and species tant fluids drawn o ff and saved. The precipitates on alternate weeks. Two different schedules specificities o f the various antisera we used the were washed 3 times in 0.85% NaCl solution were used to produce antisera to heated IgG m icro-technique m odification (Sever, 1962) of and finally suspended in a convenient volume precipitates. By the first m ethod, a volum e of the hemagglutination (HA) test of Boyden o f 0.85% NaCl solution. Protein concentrations IgG precipitate suspension containing 3.3 mg of (1951) as m odified by Stavitsky (1954). Sheep o f supernatant fluids and precipitates were de­ IgG was injected subcutaneously 3 tim es a week red blood cells were obtained from whole sheep term ined by U.V. spectrophotom etry as just de­ on alternate weeks for 2 such series. A week blood collected in sterile m odified Alsever’s so­ scribed, except that 1.0N NaOH was used to after the last injection, 20 to" 30 m l o f blood lution (Buhantz et al., 1946) and stored at 4°C dissolve and dilute the precipitates. The heated were collected by cardiac puncture, or booster for no longer than 21 days. 12.5 pg of heated globulins were designated Ah, Bh, Eh, Oh or Ph injections were given subcutaneously w ith 10 IgG supernatant flu id was used to sensitize each supernatant fluid or precipitate, respectively. mg of precipitate suspensions, on alternate m illiliter o f 2.5% tanned cells. A fter sensitiza­ These materials were stored at -20°C until weeks for up to 7 tim es. By the second m ethod, tion, the cells were diluted 1:5 to obtain a 0.5% used. Figure 1 ,1 summarizes the procedure for a volume of IgG precipitate suspension con­ suspension for use in the micro-technique. antigen preparation. taining 10 mg o f IgG was m ixed w ith an equal When specificity tests were perform ed, aliquots volume of Freund’s complete adjuvant. The of cells were sensitized simultaneously w ith Immunization schedules emulsion was injected subcutaneously into heated IgG supernatant fluid o f each species to Rabbits were im m unized as follows. A vol­ rabbits at several sites, and the injection was insure identical handling. Titers were expressed ume of 0.85% NaCl solution containing 10 mg repeated 2 wk later. 1 wk after the last as the num ber o f wells showing agglutination in o f purified native IgG was m ixed w ith an equal injection, 20 to 30 m l o f blood were collected a series of tw ofold serum dilution. The first volume of complete Freund’s adjuvant (Difco by cardiac puncture or the anim al was given a w ell contained a 1:2 dilution o f serum. Laboratories) and injected intraderm ally into booster injection intram uscularly w ith thrice- several sites on their backs and repeated 2 w k weekly injections, each of 5 mg on IgG precip­ Preparation of formalinized red blood cells later. 20 to 30 m l o f blood were collected by itate on alternate weeks. The serum was sep­ Form alinized sheep red blood cells were pre­ cardiac puncture 1 w k after the last injection. arated from whole blood samples and kept pared by the m ethod of Csizmas (1960) from Thereafter, either intraderm al booster injec­ frozen until used. sheep whole blood stored in Alsever’s solutions 152-JOURNAL OF FOOD SCIENCE-Volume 35 ( 1970)

gppi « 1 * : £ n c c c ^ ^ ^ ! a ^1 Q ’fH1 : r, G -ç C ^ ^ [<• ?S *♦ ü' V - r * Ç & ,C'<# .6 g c L i# fr 1* Sfr 3# f ' t" % - : G r® . '• * - G & r* ; •* :'i§ . ' ' ' f* \ •' » # /ç.'f* m '© : {#: W ff 4 ' -** * ' i ; :(p \ « ® €* S -I»: © P Î#:;j» .i|| i («- (% 9 « i * w W ■ %§ i f * if ( m i ’■M © f * ¿» « (* i « Û. ftli ;|| J§ ,# £ c* # \ •* A* e * «I v O ■ # € •* €■ % w- ..y .-• e 4 € k f !: SZss 'w.* ; £ ;t;-e *, (L 0 * » « % ^ -, € %■ € -,y .<• \ t è e ... *, j: e # « i : # *y&fr .#.'• . . : \ ! # . e e e £ C5 : V . J H i Wé W * 4 -:-4 ' A ■A B E O P 7 2 % B BOP Z^S*§* § m m 1 ® ###? .. -

F/g. 2—M icrotiter patterns o f anti-Ph-IgG serum Fig. 3—M icrotiter patterns o f anti-Ph-IgG serum Fig. 4—Microtiter patterns of monospecific No. 72—11—3 prior to absorption to mono- No. 72—11—3 following absorption to mono­ anti-Ph-IgG serum No. 72—11—3 before and specificity. Tanned sheep RBC's sensitized with specificity with a mixture o f avian, bovine, and after absorption with various sausage homog­ fluid-heated globulins: A(vian), B(ovine), equine heated globulin precipitates. A, B, E, O, enate resiaues. Monospecific serum absorbed E(quine), O(vine), P(orcine). Titers: A=10, B= P: See legend. Fig. 2. a, b, e, o, p: negative with: 1, no absorption; 2, pork sausage; 3, 7, E = 6, O = 6, P = 8. Well No. 1 o f each row controls for A, B, E, O, P. Titers: A, B, E, O = beef-horse sausage; 4, beef sausage; 5, horse does not contain sensitized RBC's. negative; P - 8. sausage. 6 = negative controls. Titers: 1 = 8; 2 = 3 ; 3 = 7; 4 = 7; 5 = 7.

no longer than 5 days. Formalinized cells were P reparation o f sausages homogenized (VirTis Company, Inc., Gardiner, stored at 25% concentration in phosphate buff­ All-sheep and all-horse sausages were pre­ New York) at approximately 22,500 rpm for 3 ered saline, pH 7.2, at 4 C for up to 6 months. pared from raw materials. Horse and sheep min. The homogenate was centrifuged at 1,500 Use of formalinized red blood cells in HA was meat and fat were obtained from animals given x g and the supernatant fluid recovered. The similar to that of fresh red blood cells except postmortem examinations at the University of residue was washed 3 to 5 times with 30-ml that the diluent was 1% bovine serum albumin Illinois College of Veterinary Medicine. The fol­ aliquots of 0.85% NaCl solution. The washings (BSA), rather than 1% normal rabbit serum, in lowing recipe was scaled down to suit the meat were repeated until no oil or fat layer formed at 0.85% NaCl solution, and the red blood cell sample sizes available (horse meat—174 g; sheep the meniscus. The washed residue was resus­ concentration for the micro-HA test was 1%. meat-301 g): meat-12 lb (including 20% fat); pended in 20 ml of 0.85% NaCl solution and ice-4 lb; salt-100 g; dextrose-28.4 g; white kept frozen until used. The fat layer was lifted Preparation o f monospecific sera p e p p er-7 .0 g; mace—3.0 g; coriander-3.0 g; off the first supernatant fluid with an-applica­ Species cross-reactive antisera were made sodium nitrate-1.0 g; sodium nitrite-1.0 g; as­ tor stick and the aqueous portion kept frozen monospecific by absorption with purified het­ corbate-4.0 g; nonfat dry milk solids-20.0 g; until used. erologous IgG precipitates. Antisera were first w ater-20.0 g. These materials were mixed in a tested by micro-HA with cells sensitized with Testing o f sausage blender and stuffed into frankfurter skins. The homologous and heterologous heated IgG 1 ml of sausage homogenate suspension was sausages were heated in a water bath at 70°C to supernatant fluids. If the homologous titer was placed in a 13- by 100-mm round-bottomed test an internal temperature of 50°C (approximate­ high enough, the serum was first absorbed with tube and centrifuged at 900 x g. The superna­ ly 40 min) and kept frozen until used. the IgG precipitate of the species giving the tant fluid was drawn off, leaving a residue vol­ greatest heterologous titer. If monospecificity Preparation o f sausage homogenates and super­ ume of 0.2 to 0.3 ml. 0.1 ml of monospecific was not obtained, the absorption was repeated natant fluids antiserum was added to the residue. The tube with the major cross-reacting IgG, usually up to 10-g aliquots of each sausage sample were was shaken and the supernatant fluid recovered 3 times. If necessary, the serum was then ab­ added to 20 ml of 0.85% NaCl solution and as described for absorption with heated IgG sorbed with the second and the third most cross-reactive IgG’s. An antiserum was consid­ ered monospecific only if it had a homologous Table 1.—Numbers o f homologous and heterologous lines obtained in testing by double dif­ titer of 5 wells or more, and a heterologous fusion each of 5 anti-native IgG and 5 anti-heated IgG sera against 5 native and 5 heated IgG titer not exceeding 1 well. Absorptions were preparations. usually performed on 1 ml of antiserum heated Cross-reactivity at 56°C for 30 min and absorbed with red Number of precipitin lines index blood cells. An aliquot of IgG precipitate sus­ pension containing 2 mg of protein was placed ______Heterologous in a 13- by 100-mm round-bottomed test tube Total Heterologous Homologous Homologous and centrifuged at 900 X g for 5 min and the Anti-Native IgG Sera supernatant fluid discarded. 1 ml of the pre­ pared serum was added to the residue and the vs. Native IgG 21 9 12 0.75 tube rotated at an angle at 190 rpm for 30 min vs. Heated IgG 18 11 7 1.57 on a rotary shaker. The tube was centrifuged at 900 x g and the supernatant fluid recovered Anti-heated IgG Sera and tested by micro-HA. Figures 2 and 3 illus­ vs. Native IgG 7 2 5 0.4 trate microtiter patterns of an antiserum before vs. Heated IgG 6 1 5 and after absorption to monospecificity. 0.2 SEROLOGIC IDENTIFICATION OF SAUSAGE MEAT ORIGIN- 1 5 3

Table 2.—Numbers and types of heated IgG Table 3—Homologous and heterologous HA titers of several antisera before and after absorp­ precipitate absorptions required to render sev­ tions with various numbers and types of heated IgG precipitates. eral antisera monospecific. Titer to cells Number of absorptions sensitized with IgG type Antisera and IgG type Antiserum type Number of absorptions and number and IgG type Ah Bh Eh Oh Ph Anti-Eha 1 X Ah Anti-Ah 1 X Bh Anti-Bha 102-8-28 None 0 12 0 0 0 Anti-Oh 3 X Bh + 1 X Eh Anti-Ph 115-8-28 None 10 9 10 8 10 Anti-Ph 2 x Eh + 2 x Bh + 2 x Ah 1 x Ah 6 6 5 4 5 1 x Ah + 1 x Bh 2 0 0 0 0 a Ah: heated avian IgG: Bh: heated bovine IgG; Anti-Ph 59-6-22 None 7 8 8 4 9 Eh: heated equine IgG; Oh: heated ovine IgG; 2 x Eh 7 3 3 2 8 Ph: heated porcine IgG. 2 x Eh + 2 x Bh 7 0 0 0 8 2 x Eh + 2 x Bh + 2 x Ah 0 0 0 0 8 aSee footnote, Table 2. precipitate. The sausage-absorbed monospecific serum thus prepared was tested by micro-HA with cells sensitized with homologous heated IgG supernatant fluid. The HA titer was com­ pared with that of untreated monospecific se­ Table 4—Homologous HA titers of four monospecific antisera after absorption with homog­ rum. Figure 4 illustrates the microtiter patterns enates of various sausage types. of sausage-absorbed monospecific serum and -f- Micro-HA titer Figure 1 summarizes the procedure for testing After absorption with sausage meat type sausages. antiserum to Prior to IgG species3 absorption C B + E B + O + P P P + A Ea Micro-HAI test Bhb 9 NDd ND 8 To test for the presence of avian meat in 0 0 ND Eh ND 0 ND 5 sausages, we used the micro-hemagglutination 7 5 0 Oh 8 ND 6 2 ND ND ND inhibition test (HAI) of Boyden (1951) as n modified by Stavitsky (1954) and Sever (1962). Ph 8 7 0 3 0 7 Sausage supernatant fluids served as the in­ See footnote, Table 2. hibiting agent. The monospecific anti-avian- Anti-Bh-IgG serum was monospecific without pretreatment. heated-IgG serum was prepared as for the HA B, bovine; E, equine; O, ovine; P, porcine; A, avian. test just described. The HA titer of the serum was 12 wells, corresponding to a dilution of Not done. 1:4096. One drop of serum diluted 1:500, corresponding to 8 HA units of antibody, was added to each of a series of twofold dilutions of anti-heated IgG serum and cells sensitized generally removed by absorption with sausage supernatant fluid. The first well of the homologous heated IgG precipitate. For series contained undiluted material. Red blood with heated IgG supernatant fluid. Be­ cells sensitized with heated Ah-IgG supernatant cause of the low titers displayed by this this reason the latter system was selected fluid were added and the test read as usual. system, serial antiserum dilutions were as the most specific. Appropriate controls indicated that the sausage started with undiluted material. To con­ The number and types of absorptions supernatant fluid caused red blood cells to serve antisera, we used the micro-HA test. with heated IgG precipitates required to agglutinate nonspecifically. The nonspecific Initially, we used formalinized red blood render an antiserum monospecific varied agglutinin was removed by shaking the sausage cells for HA testing because they can be and had to be determined for each indi­ supernatant fluid with an equal volume of stored for up to 6 months without deteri­ vidual antiserum (Table 2). Examples of chloroform and recovering the aqueous phase. oration and can be kept frozen or freeze- an antiserum monospecific without pre­ Treatment with chloroform was repeated 2 or 3 treatment, an antiserum rendered mono- times, until testing of the aqueous phase dried for even longer periods. This would showed absence of nonspecific agglutination. have allowed testing of various materials specific and an antiserum which could with a single preparation of cells over a not be rendered monospecific are given RESULTS long period. However, after 4 months of (Table 3). DOUBLE DIFFUSION studies were con­ successful use, the formalinized cell sys­ Of 12 sera tested, 3 could be made ducted on the interspecies cross-reactivity tem suddenly became inoperative; cause monospecific, 3 could not be and 6 were of gamma globulins using anti-native IgG of this failure could not be determined. monospecific, or almost so, without pre­ and anti-heated IgG sera. Each serum was Since a second batch of formalinized red treatment. Factors which influenced the tested against all 5 native IgG’s and all 5 blood cells was also found to be unusable, effectiveness of absorption included the heated IgG’s. The anti-native IgG sera testing with formalinized cells was aban­ initial titer of the serum, the use of were generally more potent than the doned. homologous or heterologous antigens and anti-heated IgG sera. Despite the bias Absorption of anti-native IgG sera their sequence, and the quantity and created by the stronger antisera’s ability with a great excess of homologous heated quality of heat-precipitated IgG used for to develop a greater number of precipitin IgG precipitate failed to remove all anti­ each absorption. lines, some general trends were discerni­ body detectable by red blood cells sensi­ Monospecific antisera were used to test ble. Results are summarized in Table 1. tized with native IgG. Similarly, absorp­ for the presence of a specific IgG in a sau­ Anti-native IgG sera revealed twice the tion of anti-heated IgG sera with either sage sample homogenate. Examples of cross-reactivity among heated IgG’s of homologous heated IgG precipitate or na­ tests on sausage samples with 4 mono- that among native IgG’s, while anti­ tive IgG failed to remove all antibody de­ specific sera are shown (Table 4). heated IgG sera completely reversed this tectable by micro-HA using red blood The monospecific anti-Bh-IgG serum trend by revealing twice as much cross­ cells sensitized with native IgG. However, behaved as expected, showing complete reactivity among native IgG’s as among antibody detected in anti-heated IgG sera removal of homologous titer by absorp­ heated IgG’s. by testing with red blood cells sensitized tions with sausage samples containing The HA testing system chosen was with heated IgG supernatant fluid was beef and no significant reduction of titer 154-JOURNAL OF FOOD SCIENCE-Volume 35 ( 1970)

may result in the production of antisera ence of impurities, it was possible to pro­ Table 5—H em agglutination inhibition titers o f Anti-Ah-IgG serum No. PH 68-5-24 with that can be rendered monospecific more duce sera which were monospecific or sausage supernatant fluids and Ah-igG super­ easily. could be made so by absorption with het­ Results of the immunodiffusion analy­ erologous IgG’s. Since there was great var­ natant fluid. sis (Table 1) suggest that heating of IgG iation in the quality of the antisera pro­ Micro-HAI causes an increase of species cross-reactiv­ duced by individual rabbits, it was found Inhibitor titer3 ity as well as the acquisition of new, spe­ expedient to immunize several rabbits Ah-IgGb-S.c 16 cies-specific antigens (Furth, 1925; Land- with IgG of each species. It was necessary Saline 0 steiner, 1945; Maurer, 1954; Zinsser, to test the rabbit sera at frequent inter­ Ad-S.e 2 1924; Henney et ak, 1968; Grabar, vals during the course of immunization. AP-S. 2 1958). Results support the tentative con­ Sera obtained during the early phase of BE-S. 0 clusion that anti-heated IgG sera are more immunization were often more specific BOP-S. 0 apt to reveal species specificity of heated than those obtained later from the same IgG’s than are the anti-native IgG sera. rabbit. The broadening of specificity dur­ 3 8 HA units of antiserum were used. This conclusion was confirmed in prelim­ ing the course of immunization is a well- b 6.9 mg/ml. inary trials in which the 2 types of sera recognized phenomenon (Nuttall, 1904; Supernatant fluid. were absorbed with heterologous heated Abbreviations used here are cited in Table 4. Proom, 1943; Landsteiner, 1945). It was Supernatant fluids were chloroform extracted Ig

D epartm ent o f Food Science

Rutgers U niversity, The State U niversity, N ew Brunsw ick, N ew Jersey 08903

COMPARATIVE GROWTH OF CLOSTRIDIUM PERFRINGENS IN CARBON DIOXIDE AND NITROGEN ATMOSPHERES

SU M M A R Y-R elative growth rates o f 8 food poisoning strains o f Clostridium perfringens i n a ferred to freshly prepared modified Noyes Veal Broth, incubated in a constant-temperature 100% nitrogen atm osphere and in a 100% carbon dioxide atm osphere at 760-m m pressure were water bath at 43°C for 24 hr and stored at determ ined. G row th was m easured by the pour plate m ethod. Substrate was fluid thioglycollate refrigerated temperature. m edium flushed w ith 100% nitrogen gas or 100% carbon dioxide gas at atm ospheric pressure and Preparation of inoculum incubation was at 43 °C. Headspace gas in the culture flask was analyzed during the grow th period.

The nitrogen was reduced at the end o f 5 hr bu t the carbon dioxide rem ained at approxim ately To prepare inocula, 1 ml from a stock cul­ ture tube was inoculated into 10 ml of freshly 100% during the 3.5-4-hr exponential grow th phase. Generation tim e o f the Clostridium per­ prepared fluid thioglycollate medium (BBL). fringens strains grow n in 100% nitrogen atm osphere varied from 12.9— 16.9 m in and in the 100% The inoculated tubes were incubated at 43°C carbon dioxide atm osphere from 12.9— 17.2 m in. W hen com paring individual strains in carbon for 24 hr. 1 ml of a 24-hr culture was diluted in dioxide versus nitrogen, no significant difference was noticed. W hen each o f the 8 strains was 9 ml of fluid thioglycollate medium and 0.1 ml com pared w ith one another in a nitrogen atm osphere, as w ell as in a carbon dioxide atm osphere, of this dilution inoculated into 100 ml of fluid significant differences w ere noticed in grow th rates. thioglycollate medium in a Nephloflask (Bellco Glass, Vineland, New Jersey) to initiate each INTRODUCTION perfringens. Angelotti et al. (1962) re­ ported that 90% nitrogen and 10% carbon growth study. DURING 1968, 56 outbreaks were re­ dioxide created anaerobiosis for growth To prepare a 100% nitrogen atmosphere, ported and 5,966 persons were affected sterile fluid thioglycollate medium was flushed of Clostridium perfringens in anaerobic with nitrogen gas (Fig. 1). A manometer was Clostridium perfringens by food poison­ jars. There is no information available used during the flushing of nitrogen gas to ing in the United States (Anon. 1969). concerning the growth of Clostridium maintain 760 mm Hg pressure in the flasks. The The major vehicles associated with C los­ perfringens in higher nitrogen and high 100% nitrogen atmospheric condition in the tridium perfringens food poisoning were carbon dioxide atmospheres. This paper Nephloflask was checked by injecting a l-ml turkey, beef, chicken and pork. C lo strid ­ presents a comparative study of growth headspace sample at room temperature into a ium perfringens food poisoning was sec­ for several food poisoning strains of C los­ gas chromatograph (A-90 P3, Aerograph) fit­ ted with a molecular sieve column 'A in. in di­ ond only to staphylococcus as the cause tridium perfringens in a 100% nitrogen at­ of food-borne outbreaks, but led all oth­ mosphere and in a 100% carbon dioxide ameter and 5 ft long. Analysis for both nitrogen ers in number of persons affected. The atmosphere in fluid thioglycollate medi­ and oxygen was carried out. When no oxygen wide distribution of this organism results peak was detectable, it was assumed the nitro­ um at 43°C. gen atmosphere was complete and flushing was in the contamination of many foods and terminated. Time required to reach the end food products consumed by humans. EXPERIMENTAL point varied from 1.5-2 hr. The survival of both vegetative and EIGHT strains of Clostridium perfringens were The 100% carbon dioxide atmosphere was spore forms of Clostridium perfringens used. Strains NCTC 8235, 8237 and 8247 were obtained in Nephloflasks by following the has been reported under conditions vary­ obtained from Dr. Betty Hobbs, Central Public flushing procedure previously described for ni­ ing from below-freezing temperatures to Health Laboratory, London, England; strains trogen. To maintain the pH of the medium at boiling-point temperature (Hall et al., 1362, S-80, S-88 and 6867 from Dr. H. E. 6.8, 5 ml of 1.0 N NaOH were added to the 1965; Hobbs, 1957). Hall, Robert A. Taft Sanitary Engineering Cen­ fluid thioglycollate medium before flushing Hall et al. (1965) reported that ex­ ter, Cincinnati, Ohio. with 100% carbon dioxide. tremely rapid growth of these organisms Stock cultures were carried in Angelotti’s Headspace analysis for both air and carbon modification of Noyes Veal Broth (Angelotti et dioxide was accomplished using a Porpak Q would occur in food at temperatures be­ al., 1962). Every 3 months cultures were trans­ tween 43.3 and 46.6°C. Smith (1963) in­ column. When the air peak detected became dicated that 35°C was optimal for growth of Clostridium perfringens. Hobbs (1957) found that temperatures in the range 39—49°C allowed rapid growth in meat slices and gravy. Hall et al. (1965) report­ ed that the excellent growth obtained at 35° was preceded by a lag of 2—4 hr, whereas no lag phase was noted at 46°C. Temperatures at 46°C and above have been used for isolation purposes (Chap­ man, 1928). Barnes et al. (1963) found an interesting relationship existing be­ tween pH, composition of medium and growth temperature. Using Robertson’s cooked meat broth, they found growth of Clostridium perfringens at 20°C variable at pH 5.8 but rapid at pH 7.2. In a study of the relationship between pH and gener­ ation time by Smith (1963) no growth was evident at pH 5.0 or 8.5. Fuchs et al. (1957) and Boyd et al. (1948) studied the nutritional requirements of Clostridium Fig. 1—Gas flushing system .

156-JOURNAL OF FOOD SCIENCE-Volume 35 ( 1970) CLOSTRIDIUM PERFRINGENS GROWTH IN VARIOUS ATMOSPHERES- 157

very small it was assumed that the carbon diox­ Table 1—Duration o f adjustment phase for Clostridium perfringens ide atmosphere was complete and flushing was in 100% nitrogen and 100% carbon dioxide atmospheres. terminated. The time required to reach the end point varied from 1.5—2 hr. Duration of adjustment phase Flushed Nephloflasks were inoculated with Nitrogen Carbon dioxide 0.1 ml of a Clostridium perfringens culture Initial level Time Initial level Time through the flask septum using a sterile tuber­ Strain (org/ml) (min) culin syringe (B + D disposable, 23-G needle, (org/ml) (min) 2.5 cc). The flasks were incubated in a con­ 6867 6.4 X 103 60 1.4 X 103 90 stant-temperature water bath at 43°C. An un­ 4.5 X 103 60 1.9 X 103 60 inoculated flushed control was carried in every 8235 X X experiment. 4.0 103 90 8.0 103 60 1.4 X 103 60 5.8 X 104 60 The growth rate of Clostridium perfringens 8247 1.0 X 103 60 1.3 X 103 30 was evaluated by the pour plate method. Every 1.1 X 103 60 1.8 X 103 60 30 min the culture flasks were shaken and 1 ml 1362 1.1 X 103 90 2.8 X of the growing culture withdrawn from the 10s 120 Nephloflask with a sterile disposable syringe (B 9.2 X 103 60 2.7 X 10s 90 & D tuberculin, 23-G needle, 2.5 cc). Serial 8797 3.4 X 102 60 1.4 X 103 60 dilutions were made in pH 7.2 phosphate buffer 2.0 X 103 60 2.4 X 103 60 1 (Standard Method for the Examination of GO OO oo 1.2 X 104 120 8.2 X 103 150 Dairy Products, 1967) and plated in Sulfite 1.1 X 103 150 2.2 X 104 90 Polymyxin Sulfadiazine (SPS) agar (Difco, 8237 1.0 X 103 30 1.1 X 103 60 BBL) (Angelotti et al., 1962). After the initial 1.3 X 103 30 1.5 X 103 60 pour plate had solidified, a few milliliters of S -8 0 4.2 X 103 120 1.0 X 10s 90 SPS agar were poured into each petri dish to 3.6 X 103 90 9.8 X 104 90 cover the entire previously poured surface. Af­ ter solidification the Mates were inverted and placed into Case anaerobic jars (Case Laborato­ ries, Inc., Chicago, 111.), under an atmosphere of 100% nitrogen gas. Nitrogen was used for all pour plate incubations regardless of the original fluid thioglycollate growth atmosphere, to per­ Detector temperature- 110°C. Column teir.per- pected to reduce the adjustment phase mit valid comparisons without complications ature-50°C. Carrier gas-Helium 50 ml/mir.. time. A similarly reduced adjustment due to changes in colony formation potential The data accumulated during the exponen­ time was observed for strain 8237 in and morphology. Anaerobiosis was accom­ tial growth phase of Clostridium perfringens for plished by connecting the jars to a vacuum line 100% nitrogen atmosphere compared to all strains in each environmental condition were the same strain in carbon dioxide. The to obtain a reading of 20-25 in. on the at­ subjected to regression analysis. The numerical tached gauge, then filling the jar to atmospheric value for the slope of the best straightline was time required to enter the exponential pressure with 100% nitrogen. This procedure determined. Slopes from replicated experiments growth phase for strain 8237 in nitrogen was repeated 4 times before incubation of the were subjected to a “t” test at the 95% confi­ was only 30 min compared to 60 min in a petri dishes. A duplicate series of plates was dence level. Average slope values were convert­ carbon dioxide atmosphere. incubated in a carbon dioxide atmosphere dur­ ed into generation times for each strain in each ing one complete experiment involving cultures There appeared to be some adjustment of the two growth conditions and were ex­ phase duration differences between at­ under both atmospheres. There was no signifi­ pressed in minutes per generation. cant difference in cell counts between the two mospheres and definite differences be­ Growth of all strains of Clostridium perfrin­ sets of plates. Therefore, all subsequent plates tween some strains. g en s in 100% nitrogen and 100% carbon diox­ were incubated under nitrogen atmosphere ide were compared with one another by analy­ Gladstone et al. (1935) stated that car­ throughout the experiments. The jars were then sis of variance of the average slopes. bon dioxide is an essential factor for cer­ placed into a 37°C incubator for 24-48 hr. Tukey’s multiple comparison test a: the tain bacteria including Clostridium per­ Black colonies of Clostridium perfringens were 95% confidence level was used to compare frin gen s counted with a Quebec colony counter (Dark- and that carbon dioxide must be growth rates of the various strains in each of field Quebec Colony Counter, American Opti­ present in all media before multiplication cal Co., N. Y.). two atmospheres. begins. If this were true, some inhibition should have been evident in the nitrogen The atmosphere within the flasks was moni­ RESULTS & DISCUSSION tored during the course of microbial growth. atmosphere systems. The data indicated 1-ml headspace samples were removed from the DURATION of the adjustment phase for some possibility that growth was initiated Nephloflasks at 1-hr intervals. The flasks were strains of Clostridium perfringens in both even earlier for some strains in the ab­ kept at atmospheric pressure by periodic release nitrogen and carbon dioxide atmospheres sence of carbon dioxide and that there of gases generated by the microorganisms. is presented in Table 1. Strains 6867, was some delay in growth initiation in a Either an A90-P3 or a G.C.5 (Beckman Instru­ 8235 and 8797 showed approximately carbon dioxide atmosphere. ment Company) gas chromatograph equipped 60-min adjustment phase durations in Table 2 represents the generation time with a thermal conductivity detector was used both nitrogen and carbon dioxide atmo­ of 8 strains of Clostridium perfringens in for the analysis. For the analysis of oxygen, nitrogen and carbon dioxide a G.C.5 gas chro­ spheres. Extended adjustment phase dura­ the two atmospheres. There was no signi­ matograph with dual columns was used. The tions were observed for strains S—88 and ficant difference in the growth rate at the conditions of operations were Column 1. Mo­ S—80. They ranged from 90 to 150 min, 95% confidence interval within any strain lecular sieve (80-100 mesh). Column size-Vi with S—88 averaging approximately 120 whether grown in a carbon dioxide or ni­ in. by 6 ft. Injector temperature-45°C. Detec­ min and S—80 approximately 105 min. trogen atmosphere. Generation time of tor temperature-250°C. Column tempera- There was no difference between the re­ Clostridium perfringens in the 100% ni­ ture-65°C. Carrier gas-Helium 40 ml/min. sponse in nitrogen or carbon dioxide. trogen atmosphere ranged from 12.9—16.9 Column 2-Porpak Q (150—200 mesh). Strain 1362 tended towards an extended min. In the 100% carbon dioxide atmo­ Operating conditions o f the A 90-P 3 gas adjustment phase duration in carbon sphere the generation time ranged from chromatograph for the analysis of carbon diox­ dioxide but not in the nitrogen atmo­ 12.9—17.2 min. Analysis of variance of ide and air as used in the 100% carbon dioxide sphere. This is accentuated further by the average slopes for all of the strains, shown atmospheric growth studies were as follows: Column-Porpak Q (20-40 mesh). Column higher initial level of cells in the carbon in Table 3, indicated there were signifi­ size in. by 6 ft. Injector temperature-90°C. dioxide atmosphere, which would be ex­ cant growth rate differences between

C ° 3 4 t a e s a h p h t w o r g l a i t n e n o p x e e h t n i s n e g . e r e h p s o m t a n e g o r t i n % 0 0 1 a n i - n i r f r e p m u i d i r t s o l C f o e v r u c h t w o r G — 2 . g i F 4 5 °r C e p o r t e dby S m i t h( 1 9 6 3 )w a s efi e s gen perfrin t h e t w o o t h e r s . um pefi e s gen erfrin p m iu t h i r d t y p e w h i c h g r o w s a t a r a t e b e t w e e n growing,8 2 3 7 a n d S —8 8 ; a s e c o n d w h i c h i s s l o w e r f a s t e r t h a n t h e o t h e r t w o , s t r a i n s i n e i t h e r a t m o s p h e r e s . T h i s may s t r a i n s dioxide. u s is n e g n ti Tukey’ s n t n h i et t s s h a r m e o e m8 gii u e ns d l f n t e o t a r r a i nb m a d p l e a F l i i t i e n z i s g o s e e u n c e d ro o nf e e m f x o p 3 i p a r r , o n i c n w s a e h F o r n i n i b t c o g h i n u ar r l e e g p r r o 2 e w , t ­ h p w h h a i s c h e f r o e r p r e s e n t s t h e t r o g e n a n d c a r b o n d i o x i d e a t m o s p h e r e s . t e s t w i t h 95% c o n f i d e n c e i n t e r v a l i n n i ­ I t a p p e a r s t h e r e a r e 3 t y p e s o f . s e r e h p s o m t a e d i x o i d n o b r a c % 0 0 1 lor, NUMBER OF ORGANISMS Generation tim e = e tim Generation

d n a n e g o r t i n % 0 0 1 n i s n e g n i r f r e p m u i d i r t s o l C r o f s e m i t n o i t a r e n e g d n a s e t a r h t w o r G — 2 e l b a T T h e g e n e r a t i o n t i m e f o r

F i g u r e 4 c o m p a r e s a v e r a g e s l o p e s f o r 8

Strains

in 0 -8 S 8237 8797 8235 6867

8247 1362 L A N R U JO 1

OO 00 i.e., i n b a c t e r i o l o g i c a l m e d i u m a t s t r a i n s 8 2 3 5 a n d 8 2 4 7 a n d a

OF s t r a i n s : One w h i c h g r o w s .23 ±0.0011 0.0233 0.0037 ± 0.0217 ±0.0031 0.0212 0.0005 ± 0.0197 0.0008 ± 0.0177 0.0025 ± 0.0183 0.0019 + 0.0193 .20±0.0025 ± 0.0230 Average slopes slopes Average

FOOD otii perfringens m lostridiu C ------Slope C.I. ±C 100% Nitrogen atmosphere Nitrogen 100%

.. t9% level. 95% at •C.I. ENCEVlme E-Volum C N IE C S otii m lostridiu C i.e., osrd­ strid lo C strains Generation time time Generation mn ± .I. ±C (min) 31 1.6 ± 13.1 2.4 + 14.2 .4 0 ± 15.3 2.6 ± 16.4 1.4 ± 15.6 29 1.1 ± 12.9 .9 2 ± 13.8 .8 0 ± 16.9

35

( o f med i u m u s e d w a s r e p o r t e d . about g a s fT l h u i s o h g i l n y g c o y l i l e a l t d e e d M e g d i r u o m w t w h a s c u u r s v e d e s w i w t i h t o h u t w e r e c p r r o e b a a se b x e t l e d y n d g ead d e d u n j e e u r s t a t o t m i e t o n h n tp e ht a i s m r e e e sa g s n u . diT l n h a ­ e r s e e d f i f s t e u c r t b s ­ . e r e h p s o m t a e d i x o i d n o b r a c % 0 0 1 a n i C ° 3 4 ­ n i r f r e p m u i d i r t s o l C f o e v r u c h t w o r G — 3 . g i F 1970 t a e s a h p h t w o r g l a i t n e n o p x e e h t g n i r u d s n e g P r e l i m i n a r ys t u d i e si nw h i c hF l u i d ) .27 0.0021 ± 0.0217 0.0015 ± 0.0188 0.0029 ± 0.0208 0.0020 ± 0.0222 .22 0.0012 ± 0.0232 0.0039 ± 0.0232 2 1 0 0.0 ± 0.0180 0.0011 ± 0.0175 18.0 Average slopes slopes Average 100% Carbon dioxide atmosphere dioxide Carbon 100% odto (C) Condition Total Error C) X (S Interaction (S) Strain m i n . N e i t h e r s t r a i n n o r t y p e C.I. ±C * Significantly different at 95% confidence level. confidence 95% at different * Significantly T I M E ( m i n u t e s ) Generation time time Generation . e l b a t e c n a i r a v f o s i s y l a n A — 3 e i b a T (min) ± C.I. ± (min) Degree of of Degree 60± 1.2 ± 16.0 .2 2 ± 14.4 .1 1 ± 17.2 29 1.3 ± 12.9 .6 2 ± 12.9 1.5 ± 13.8 1.2 ± 16.7 .3 1 ± 13.6 freedom 31 16 7 7 1 0.00003786 0.00001591 0.0000021 0.0001186 0.0001725 Sum of of Sum square

fringens. a s s e e n i n F i g u r e 5 . T h i s may b e d u e t o r a p h y . F u ch h y s d e r o t g e a ns l u . l f ( 1 i 9 d 5 e 7 ) , h d y o d c u r m o e g n e t n ea d n do t h e r nique.t h es e n s ic t h a i n v gi i e d t e i y n no t p f i r it f o h n d i e u e im c d e e t .Th a dd s i u e u a r r r e i iwa l n n no s g gn l t t o e e h t c v e i h g e ­ c r l e o a w t b o h l f w e e r c e a n r o b t o n d i o x i d e , p o u n d st b i y o 3 n g s a o t n f r i at c h h i y s e d n r u s o l u g e f t o n f u i r l sg i u a c o z l s m a p f e o t i u s n d i d w s e o e a r n nt e f h d r r s o o o e t m u f p g h c h a e t a o r s u r h a tt p b e e t r h o v s o e n eg d e ed d u r i r c o b ­ o w com­ a y t x g l h gs i a t a d n e u s i t d n l o c y o h e .So r r f v o m g e m a e a a t n l l o i g l r c ­ e m a a a i n f n d t e d e o c r r o ­ n t s h t e a n 3 t r d h r o f g r o w t h , i n f r o m t h i o s u l f a t e a n d c y s t i n e . O t h e r g a s e s t h e m i d d l e o f t h e e x p o n e n t i a l p h a s e . T h e i n F it g i u o r n e l 5 o . w e a r n T e h a d e l y t i z h n i e n i g n th e i i a t a d r s o l p g a e c n n e is a t l m r p e o l g v e e e s ni l n la , gs e o a m a v e s s e b l u t s hown b y e r ­ i c a cgens, i d w a s f o u n d . C a r b o n d i ­ c h r o m a t o g rt a h p e h . g Cd r a i o r w o b t x o h n i t sd d r t i e o o u i g x d n i e it d n e h e . s e p o i w r T x h e o n e r d i o e u d c l c d e ­ u e e p l t v r a e o e r t d m et l i u d h s n c ef e t d ip l o oh r b a f n y o i s l n k h w , a c i s s t ic d r p a n u o r a o r b g l t i o e i n n n f c g d e e i r a d o m n x d e d i n u d t c r e a a i t r a n i b n g o o d n n h t y p h d r e r o o d g e u n c t w s e r a e n d h e a d s p a c e i w n i t g h i o f n t t h h e e c c u u l l t t u u r r e e a f t l a e s a k c h s s a a d m d p e l d i n g t i m e . MODEL MODEL s y s t e m e v i d e n c e p r e s e n t e d i n d i - um pefi e s gen erfrin p m iu ( 1 9 6 1 )a n d H a u s c h i l d et t o a t l h . e ( 1 o 9 x 6 y g 7 e ) n r e i ­ n c oa r n c p e o r o a f t i t o h n e . a T n h a i e s r o w b a i s c o s b ­ y s t e m b y t h e m i x ­ growth of a f t e r g r o w t h o f a n u m b e r o f p a t h o g e n s i n c u r v et s a i f n oT e l h d li i o s n w e t i d h s i t o i sp n o s t r t h a t u e g e d r d y e sf e t t o m a e h r t n a t i t t h ow o n i s ma t ap e x h i r m r g u t m y r e o h s w p g e ts e r h h d o e n a a w r t s t t v h e a e . i d o oth n b f ­ r o t u h g e ht h m ei e d n i u d m i . c a t o rm a t e r i a l c o m p l e x m e db i r a e a . k d F o o w r n pefo r d o rgr d o u w c tfr h t o sca m f r o b r o m e h d y d f r r a o t m e s g . l u c o s e gens, F r s i e gen d e erfrin m p a n m ne tridiu ta l .( 1 9 4 1 )s t u d i e dt h e T h e l a r g e s u r f a c e v o l u m e a n d t h e l a r g e F u c h se ta l .( 1 9 5 7 ) ;C o l l e ee ta l . D u r i n g g r o w t h o f l a c t i c a c i d , a c e t i c a c i d , e t h y l a l c o ­ g a s p r o d u c t i o n w a s e v i d e n t . 0.000002273 0.0000021 0.00001694 0.000002366 H y d r o g e n s u l f i d e w a s p r o d u c e d Mean of of Mean square osrdi rrngns gen erfrin p m iu strid lo C CONCLUSIONS _ o c c u r r e d i n 4 . 5 —5 h r . o b t a i n s e n e r g y n e e d ­ 0.9607 0.8875 7.1597* osrdi rrn­ erfrin p m iu strid lo C osrdi fi ­ rfrin e p m iu strid lo C - F osrdi ­ r e p m iu strid lo C — osrd­ strid lo C in ni­ los­ C

CLOSTRIDIUM PERFRINGENS GROWTH IN VARIOUS ATMOSPHERES-^

0.03

0.02- 1

0.01 —I--- 1___ 1___ I___ I___ I___ I— 6867 8235 8247 8237 S-88 8797 S-80 1362 STRAINS TIME (hours) Fig. 4 —Average slopes representing the exponential grow th phase for

8 strains o f C lostridium perfringens in fluid thyoglycollate broth at Fig. 5 —C ulture flask headspace analysis during

43°C in atm ospheres o f 100% nitrogen and 100% carbon dioxide. The grow th o f C lostridium perfringens after prior

lim its exhibited were determ ined by application o f Tukey's test at the replacem ent o f a ir w ith carbon dioxide or n itro ­

95% confidence level. g e n .

cates that gas-flushed systems, whether strain of Clostridium welchii in beef. J. Hall, H.E. and Angelotti, R. 1965. Clostridium Appl. Bacteriol. 26: 415—427. perfringens in meat and meat products. carbon dioxide or nitrogen, will permit Boyd, M.J., Logan, M.A. and Tytell, A.A. 1948. Appl. Microbiol. 13: 353—357. growth of Clostridium perfringens, all The growth requirement of Clostridium per­ Hauschild, A.H.W., Milsheimer, R. and Thatch­ other conditions being favorable. Genera­ fringens BP6K. J. Biol. Chem. 174: er, F.S. 1967. Acid resistance and infectivity 1013—1035. of food poisoning Clostridium perfringens. tion time for Clostridium perfringens in a Chapman, G.H. 1928. The isolation and estima­ Can. J. Microbiol. 13: 1041—1047. 100% nitrogen atmosphere was tion of Clostridium welchii. J. Bacteriol. Hobbs. B.C. 1957. The growth of heat resist­ 12.9—16.9 min and in a 100% carbon di­ 16: 193—199. ant Clostridium welchii in cooked meat. Collee, J.G., Knowlden, J.A. and Hobbs, B.C. Bacteriol. Proc. 57: 14. oxide atmosphere 12.9—17.2 min. 1961. Studies on the growth, sporulation Smith, L.D.S. 1963. “Microbiological Quality and carriage of Clostridium welchii with of Foods: Clostridium perfringens Food special reference to food poisoning strains. Poisoning.” Ed. Slanetz, L.W., Chichester, REFERENCES J. Appl. Bacteriol. 24: 326—329. C.O., Gaufin, A.R. and Ordal, Z.J., pp. 77—83. Academic Press, New York and American Public Health Association. 1967. Friedemann, T.E. and Kmieciak, T.C. 1941. Metabolism of pathegenic clostridia in com­ London. “Standard Methods for the Examination of Ms. received 5/21/69; revised 8/25/69; accepted Dairy Products.” 12th ed. American Public plex carbohydrate rich culture media. Proc. 8/27/69. Health Association, Inc., New York. Soc. Exptl. Biol. Med. 47: 87. Angelotti, R., Hall, H.E., Foter, M.J. and Lewis, Fuchs, C.A. and Bonde, G.J. 1957. Nutritional A paper of the Journal Series, New Jersey K.H. 1962. Quantitation of Clostridium per­ requirement of Clostridium perfringens. J. Agricultural Experiment Station, Rutgers Uni­ fringens. Appl. Microbiol. 10: 193—199. Gen. Microbiol. 10: 317—329. versity, The State University, Food Science De­ Anonymous. 1969. Morbidity and Mortality Gladstone, G.P., Filders, P. and Richardson, partment, New Brunswick, New Jersey. Weekly Report. 18 (12): 98. G.M. 1935. Carbon dioxide as an essenrial This research was supported in part by a Barnes, E.M., Despaul, J.E. and Ingram, M. factor in the growth of bacteria. Brit. J. grant from the United States Public Health 1963. The behavior of a food poisoning Exptl. Pathol. 16: 335—336. Service UI00312. H. CO and C. W. SANDERSON

Thomas J. U pton, Inc., 800 Sylvan Avenue

Englewood C liffs, N ew Jersey 07632

BIOCHEMISTRY OF TEA FERMENTATION: CONVERSION OF AMINO ACIDS TO BLACK TEA AROMA CONSTITUENTS

S U M M A R Y -1*C -am ino acids were added to fresh tea-leaf hom ogenate undergoing conversion to frozen at —40°C until required for use. black tea. A fte r conversion (30 m in, 2 5 °C), the volatile com pounds present in the headspace over Preparation of flavanols the reaction m ixture were collected and analyzed by gas chrom atography. Results show ed that The method used was a modification of that leucine, isoleucine, valine and phenylalanine were pa rtially converted to the aldehydes expected reported by Vuataz et al. (1959). from a S trecker degradation. These aldehydes are constituents o f black tea arom a. F urther, drying a. Preparation of crude tea polyphenol o f the ferm ented m ixture caused an additional am ount o f the aldehydes to be form ed. In contrast, extract from fresh tea leaves. Four batches of no detectable volatile com pounds were form ed from aspartic acid, glutam ic acid, glutam ine, argi­ 500 g fresh-frozen tea leaves were macerated in nine, threonine, serine o r theanine under the sam e conditions. P roduction o f aldehydes from am ino a Waring Blendor for 2 min with 1,000 ml of acids was show n to be dependent on the enzym ic conversion process: Tea leaf w hich had been cold (0°C) acetone. The suspension was filtered on a Büchner funnel. The residue was macerat­ inactivated b y steam treatm ent was no t effective in causing form ation o f volatile aldehydes from ed 2 more times with 1 liter of 80% acetone the am ino acids. Identical results were obtained in a m odel tea ferm entation system com posed o f a and finally with 1 liter of absolute acetone. crude soluble enzym es extract from tea leaves, pu rifie d epigallocatechin gal late and 1 * C - a m i n o The acetone extracts were combined and acids. Ascorbic acid was found to inh ibit form ation o f aldehydes from am ino acids in this m odel mixed with one-half their combined volume of tea ferm entation system ; dehydroascorbic acid by itself was found to be effective in causing chloroform in a separatory funnel. Distilled form ation o f volatile aldehydes from am ino acids. water was added until 2 layers formed and the chloroform layer was extracted 3 times with INTRODUCTION by the tea fermentation system using ra­ diotracer techniques and to obtain addi­ 1/4 its volume of distilled water. The combined IT HAS been shown that the concentra­ aqueous layers containing most of the poly­ tions of free amino acids undergo appreci­ tional information about the mechanism phenols free of fat-soluble materials were satu­ able changes during the various stages of of this conversion. This is part of an on­ rated with NaCl and extracted 4 times with 1/3 conversion of fresh tea-shoot tips to the going investigation of the mechanism by their volume of ethyl acetate. The ethyl acetate black tea of commerce: Namely, during which black tea aroma is formed during extracts were combined, dried with anhydrous the conversion of fresh tea-shoot tips to Na2 SC>4 and taken to dryness under reduced the first (withering) stage of black tea black tea. pressure. The solids obtained were extracted manufacture there is an over-all increase twice with dry ethyl acetate to extract the un­ in the concentration of free amino acids EXPERIMENTAL oxidized tea flavanols from oxidized (colored) (Roberts et al., 1951; Bhatia et al., 1965); material which is left behind. Materials during the second (fermentation) stage The ethyl acetate extract was evaporated to the concentrations of leucine, isoleucine, Fresh tea leaf was picked at Upton’s Experi­ dryness and the polyphenols were dissolved in serine, glutamic acid, glutamine, threo­ mental Tea Garden near Charleston, South distilled water and freeze dried. The final prod­ Carolina, and air freighted to us so as to arrive uct was a light-tan powder. nine, phenylalanine and theanine (thea­ in our laboratory the same day. The leaf was nine = 5-N-ethyl glutamine) are appreci­ b. Preparation of the flavanol-cellulose placed in a freezer at -40°C on arrival in the charge. The freeze-dried crude tea polyphenols ably reduced whereas other amino acids laboratory, for use when required. were mixed thoroughly with an equal weight of undergo little change (Roberts et al., Radioactive (14C) amino acids, generally cellulose jndcr an atmosphere of nitrogen. To 1966); during the last (drying) stage there labelled, were purchased from Schwarz BioRe- this mixture, distilled water (1/2 the dry weight is a small general decrease in free amino search, Inc.; epigallocatechin gallate (EGCG) of cellulose used) was added in small incre­ acid concentration (Roberts et al., 1966). was prepared from fresh green tea leaves by ments. Solvent “A” [ethyl propionate: ligroine Wickremasinghe et al. (1965) also report­ procedures described below. Dehydroascorbic 66-75°C. (3:1) mixture saturated with water] ed a decrease of amino acids during fer­ acid was purchased from Pierce Chemical Com­ was added to thoroughly wet the flavanol-cel­ mentation. pany and theanine was synthesized in our labo­ lulose charge. ratory according to the procedure of Sakato et c. Operation of the column. The charge The over-all decrease which occurs in al. (1950). the level of free amino acids during fer­ was packed on a prepared cellulose column (80- mentation suggests they are being con­ Preparation of enzymes solution by 10-cm i.d.) and the column eluted overnight verted to other substances. Several re­ Fresh frozen green-tea leaf (100 g) in 500 with solvent “A” at a flow rate of 6.5-6.7 searchers (Bokuchava et al., 1954; ml of phosphate buffer (pH 7.4, 0.1 M) con­ ml/min. The eluate was collected in 340-ml taining 50 g of Polyclar AT was macerated in a fractions maintained at 4°C. Following elution Nakabayashi, 1958; Wickremasinghe et Waring Blendor for 5 min. The macerate was of epicatechin gallate (ECG) from the column, al., 1964; 1965) have suggested that the filtered through 2 layers of cheese cloth and the gradient elution was started by diluting solvent amino acids are being converted, at least leaf residue discarded. The filtrate was treated “A” (2,0C0 ml) with solvent “B” [ethyl propi­ in part, to volatile compounds likely to with ammonium sulfate and the 30-90% satu­ onate: ligroine (9:1) mixture saturated with be important constituents of tea aroma. ration precipitate collected. This precipitate water] to maintain a constant volume in the Bokuchava et al. (1954) reported that was dissolved in a minimum amount of phos­ mixing reservoir. When epigallocatechin gallate when certain amino acids are incubated phate buffer (pH 7.4, 0.1 M). The solution was (EGCG) eluted from the column, the elution with tea flavanols in hot aqueous solu­ dialyzed against a large volume of phosphate gradient was changed so that the mixing reser­ tions, pleasant aromas are formed which buffer (pH 7.4, .0.1 M), with 2 changes of buff­ voir contained 2,000 ml of solvent “B” which are different for each amino acid, and a er daily, until the outside solution was color­ was continuously diluted with solvent “C” (eth­ less. This required about 3 days of dialysis. yl acetate saturated with water) to maintain mechanism to explain these results was All of the above-mentioned steps were car­ a constant volume in the mixing reservoir. The proposed (Popov, 1956). The purpose of ried out at 4°C using cold reagents and equip­ elution was continued until epigallocatechin our investigation was to verify the conver­ ment. Finally, the soluble tea enzymes prepara­ (EGC) eluted from the column. The eluant was sion of amino acids to volatile aldehydes tion, a yellowish, cloudy solution, was stored monitored spectrophotometrically at 275 nm.

160-JOURNAL OF FOOD SCIENCE-Volume 35 ( 1970) BIOCHEMISTRY OF TEA FERMENTATION - 1 6 1

All fractions from the cellulose column con­ taining a single flavanol, as determ ined from the UV absorption curve and paper and thin-layer chromatography, were combined and taken to dryness under reduced pressure at 30°C. The solids obtained were dissolved in distilled w ater and extracted with methylene chloride to re­ move any residual ligroine. The m ethylene chlo­ ride-washed aqueous solutions were freeze dried, d rie d . d. Crystallization of the flavanols. T h e freeze-dried flavanols were twice crystallized from water, dried in an evacuated desiccator o v e r P2 O 5 and stored at -40°C in sealed evacu­ ated vials until required for use. Model tea fermentation system Oxidations were carried out in jacketed re­ action vessels (Fig. 1) at 25°C. The reaction Fig. 1—Schematic diagram o f headspace volatiles collection apparatus. mixture was made up as follows: 1 0 .0 m l s u b - strate/buffer solution containing 5.0 mM epi- gallocatechin gallate in citrate-phosphate buffer (0.1 M, pH 5.4), 10 uliters (1 jj c ) 14C -a m in o acid solution and 1.0 ml of the soluble tea enzymes preparation. The enzymic oxidation was continued for 30 min, during which time the reaction mixture was aerated by stirring vigorously with a magnetic stirrer. Since 14C-theanine was not available, a higher con­ RESPONSE FROM FLAME-IONIZATION DETECTOR centration of theanine (30 mM) was used in the model tea fermentation system to improve the chance of detecting any volatile compound formed therefrom. Enzymic oxidation o f fresh green-tea homog­ enate Fresh-frozen tea leaves (16 g) were mixed with 64 ml of cold citrate-phosphate buffer (0.1 M, pH 5.4) and macerated in a cold base (0°C) Waring Blendor for 3 min. The homog­ enates were transferred to the reaction vessel (Fig. 1) and brought to 25°C. Then 10 /^liters (1 pc) of the I 4 C-amino acid solution was add­ ed to the homogenate and oxidation (called fer­ mentation) allowed to continue for 30 min, during which tim e the solution was aerated con­ tinuously by stirring vigorously with a magnetic Figure 2—Gas-liquid chromatogram o f the headspace volatiles obtained s tirre r. from a green-tea-leaf homogenate fermented in the presence o f 1 4 C- Headspace volatile analysis by gas-liquid chro­ 14 14 valine, C-leucine and C-phenylalanine. matography At the end of the oxidation period, a head- space volatiles trap was attached to one of the side arms of the vessel and the vessel was other­ wise closed. Hot water (95°C) was circulated through the jacket. The sample was stirred con­ used were of stainless steel, 3/16 in. o.d. by 50 duced by the flame ionization detector tinuously with a magnetic stirrer. 3 min after ft long, packed with 10% Carbowax 20 M on shows the typical complex aroma pattern the hot water was turned on, the headspace vol­ Gas-chrom Q, 60-80 mesh. of fermented tea (Bondarovich et al., atiles in the vessel were flushed into the trap Part of the column effluent was diverted to 1967), whereas the chromatogram pro­ w ith N 2 gas flowing at a rate of 100 ml/min for a radioactivity m onitoring unit (Barber-Colman 5 min. The headspace volatiles trap consisted of duced by the radioactivity detector shows Model 5190 Radioactivity M onitoring System). stainless steel tubing, 1/4 in. o.d. by 7 in. long, only 4 radioactive peaks, identified by packed with 0.9 g of 5% Apiezon L on Gas- RESULTS comparison of retention times to be iso- chrom CL, 100-120 mesh. The trap was main­ FRESH green-tea-leaf homogenate v/as butanal, isovaleraldehyde, phenylacetal- tained at -80°C with dry ice. fermented in the presence of 1 4 C-amino dehyde and phenylethanol. These 4 com­ Gas chromatography acids at 25°C for 30 min. The headspace pounds are constituents of black tea A Barber-Colman Series 5000 Selectra volatiles over the fermented (oxidized) aroma (Bondarovich et al., 1967, and ref­ System with dual hydrogen flame detectors was mixture were collected and analyzed by erences contained therein) and they cor­ u s e d . gas-liquid chromatography as described in respond to the Strecker degradation prod­ The trap containing the headspace volatiles the Experimental section. A typical set of ucts of the 1 4 C-amino acids (except for was attached to the inlet of the GLC unit and gas-liquid chromatograms produced si­ phenylethanol) added to the fermenting heated to 220°C with a heating tape. multaneously by the flame ionization de­ tea-leaf homogenate. Development of the chromatogram was tector and the radioactivity detector are The effect of the fermentation system started when the temperature of the trap shown in Figure 2. In this case 1 4 C-va- composition on formation of volatile reached 100°C (carrier gas = He; flow rate = 65 ml/min). The column temperature was pro­ line, 14 C-leucine and 1 4 C-phenylalanine compounds from amino acids was studied grammed from 70 to 200°C at 2°C/min with were added to the fermenting tea-leaf both in whole tea-leaf homogenates and hold at 200°C for at least 0.5 hr. The columns homogenate. The chromatogram pro­ in model tea fermentation systems. 1Q2—JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

Table 1—Effect of reaction mixture composition on formation of C-isovaleraldehyde from reaction mixture consisting of EGCG, tea ' ^ C-leucine. enzymes preparation and 14 C-leucine, only a very small amount of 14 C-iso­ Color of reaction 14 C-isovaleraldehyde valeraldehyde was produced. When no Reaction mixture composition mixture at end of incubation recovered (dpm) ascorbic acid was present, the amount of Experiment I 14 C-isovaleraldehyde formed was more

14 C-Leucine + green-tea homogenate B ro w n 5 ,1 0 0 than 10 times greater. Dehydroascorbic acid was able to cause conversion of 14 C-Leucine + green-tea hom ogenate G r e e n 9 5 0 14 C-leucine to 14 C-isovaleraldehyde in (enzyme inactivated) the absence of EGCG and tea enzymes 4 C-Leucine + green-tea homogenate, dried 1 D a r k - b r o w n 9 ,1 0 0 preparation. Experim ent II DISCUSSION 14C -L e u c in e C o lo rle s s 2 0 0 ISOBUTANAL, isovaleraldehyde, 2- 14C-Leucine + enzymes Light yellow 2 2 0 0 methylbutanal and phenylacetaldehyde 14C-Leucine + EGCG C o lo rle s s 8 8 0 were formed from valine, leucine, isoleu­ 14 C-Leucine + EGCG + enzymes Yellow-brown 4 ,6 4 0 cine and phenylalanine, respectively, in 14C-Leucine + EGCG + enzymes, inactivated Light yellow 2 9 3 0 the presence of fermenting (oxidizing) (All reaction mixtures were allowed to oxidize for 30 min at 25°C prior to collection of the fresh tea-leaf homogenate. Since these headspace volatiles for analysis by GLC. The procedures used are described in the Experimental aldehydes are present in black-tea aroma S e c tio n .) (Bondarovich et al.,1967), the results of 'Taken to dryness on a steam bath after 30 min oxidation. our investigation clearly show that a por­ Due to color of enzyme solution. tion of the black-tea aroma components come from amino acids. Further, at least part of the decrease of free amine acids during the conversion of green tea to black tea (Wickremasinghe et al., 1965; The first experiment (i.e., Experiment were converted into their corresponding Roberts et al., 1966) must be accounted I in Table 1) clearly demonstrated that Strecker degradation aldehydes; namely, for by this transformation. 14 C-leucine is actively converted to 14 C- isobutanal, 2 -methylbutanal, isovaleralde­ Results of our experiments also show isovaleraldehyde in a fermenting tealeaf hyde and phenylacetaldehyde, respective­ that the production of aldehydes from system. These results also show that pre­ ly, in both of the tea fermentation sys­ amino acids is largely dependent on the vention of the tea fermentation process tems tried. It is noteworthy that all of tea fermentation process (Roberts, 1962; by inactivation of the enzymes present these aldehydes have been reported to be Sanderson, 1965): Tea leaf which has in the tea leaf markedly reduced the constituents of black-tea aroma (Bondar- been activated by steam treatment pro­ amount (rate) of isovaleraldehyde formed ovich et al., 1967). In contrast, aspartic duced only a small amount of volatile from leucine. Further, drying of the fer­ acid, glutamic acid, glutamine, arginine, aldehydes from amino acids. Further, our mented mixture caused an additional threonine, serine and theanine did not model tea fermentation system studies amount of 14 C-isovaleraldehyde to be give rise to any detectable volatile com­ clearly showed that oxidation of tea fla­ formed. pounds under these conditions. vanols by tea enzymes markedly increases The model systems studies (Experi­ The effect of ascorbic acid and de- the amount of aldehydes formed from ment II in Table 1) showed that the con­ hy droascorbic acid on conversion of amino acids. version of 14 C-leucine to 14 C-isovaleral­ 14 C-leucine to 14 C-isovaleraldehyde was These results are interpreted to be a dehyde takes place in the presence of a investigated. Results are shown in Table reflection of a rate effect; that is. they tea enzyme extract when it is oxidizing 3. When ascorbic acid was added to the reflect the rate of production of oxidized tea flavanols (EGCG was used in this set of experiments). Incubation of leucine with the tea enzymes preparation alone produced only a very small amount of isovaleraldehyde, no more than obtained Table 2—List of amino acids tested in model tea fermentation sys­ by incubating leucine by itself. Incuba­ tem and in fermenting green-tea-leaf homogenates and the volatile prod­ tion of leucine with EGCG by itself or ucts formed. with inactivated tea enzymes preparation produced a small amount of isovaleralde­ 14 C-Amino Acid V o la tile 1 4 C - p r o d u c ts hyde: This system appears to be analo­ V a lin e I s o b u ta n a l gous to that when leucine is incubated Is o le u c in e 2-M ethylbutanal with an inactivated green-tea homogenate L e u c in e Isovaleraldehyde (Experiment I, Table 1). Phenylalanine Phenylacetaldehyde The several amino acids reported to Aspartic Acid N o n e decrease most during the fermentation Glutamic Acid N o n e stage of black-tea manufacture (Roberts G lu ta m in e N o n e et al., 1966) were incubated with the tea A rg in in e N o n e flavanol EGCG and a soluble tea enzymes T h r e o n in e N o n e preparation to determine their ability to S e rin e N o n e form volatile aldehydes. Table 2 shows a T h e a n i n e 1 >2 N o n e list of the 1 4 C-amino acids tried, both in (The experimental methods are described in the Experimental Sec­ the model tea fermentation system and in ti o n .) an oxidizing green-tea homogenate and 5-N-Ethyl glutamine. the volatile aldehydes produced. Valine, This experiment was carried out with nonradioactive theanine isoleucine, leucine and phenylalanine o n ly . BIOCHEMISTRY OF TEA FERMENTATION-163

tea flavanols which are the oxidants caus­ Table 3-E ffect o f ascorbic acid and dehydroascorbic acid on formation of 14 C-isovaleralde­ ing the Strecker degradation of amino hyd e fro m 4 C-leucine. acids in the tea fermentation system. Only when the tea leaf enzyme system is A m o u n t o f Color of reaction C-isovaleraldehyde operating is there an appreciable amount mixture at end re c o v e r e d of volatile aldehydes produced from Reaction m ixture composition of incubation ( d p m ) amino acids in the period of time normal­ ly allowed for tea fermentation, i.e., 0.5 14C-Leucine + EGCG + enzym es + ascorbic acid Light yellow 1 4 4 0 14C-Leucine + EGCG + enzymes to 3 hr (Eden, 1965;Harler, 1963;Keegel, Yellow-brown 4 ,6 3 0 14 C-Leucine + dehydroascorbic acid 1958). The relatively slow nonenzymat- Light brown 1 ,5 4 0 ic conversion of amino acids to volatile [The experimental m ethods are describee in the Experim ental Section. Ascorbic (5.0 mM) and aldehydes in the presence of tea flavanols dehydroascorbic acid (5.0 mM) were added to the substrate/buffer solution.] is probably analogous to the situation Due to color of enzym e solution. existing in the development of chocolate aroma (Biehl, 1967; Rohan et al., 1967). It was mentioned above that the abili­ other hand, aspartic acid, glutamic acid, (Roberts, 1962; Brown et al., 1969). If ty of oxidized flavanols to catalyze the glutamine, arginine, threonine, serine and ascorbic acid is present, no colored oxida­ oxidation of amino acids has been shown theanine did not give rise to any detect­ tion products are formed until such time by several investigators (Bokuchava et al., able volatile aldehydes. The results ob­ as essentially all of the ascorbic acid pres­ 1954; Popov, 1956; Nakabayashi, 1958; tained with this latter group of amino ent is oxidized to dehydroascorbic acid: Wickremasinghe et al., 1964) and that acids could well be expected if one con­ The oxidation of ascorbic acid by oxi­ Popov (1956) has proposed a mechanism siders the low volatility or the reactivity, dized tea flavanols is favored over other for the oxidative deamination of amino or both, of their expected Strecker de­ reactions possible in the tea fermentation acids during tea fermentation. Popov’s gradation products. Similar negative re­ system. The quinones formed from fla­ (1956) scheme proposes that orthoqui- sults were reported by Yu et al. (1968b) vanols are strong oxidizing agents and are nones produced by the action of catechol for alanine and aspartic acid and by capable of converting amino acids to their oxidase on tea flavanols during the tea Rooney et al. (1967) for lysine, arginine, corresponding Strecker degradation prod­ fermentation process are required for the histidine, tryptophan, glutamic acid and ucts (i.e. aldehydes). Further, dehydro­ production of aldehydes from amino proline in their investigation of simLar ascorbic acid formed from ascorbic acid acids. systems. may also act as an oxidizing agent capable Popov (1956) did show that carbon di­ Popov (1966) suggested that dehydro­ of converting amino acids into their cor­ oxide, ammonia and aldehydes were ascorbic acid formed by nonenzymatic responding aldehydes. formed in model tea fermentation sys­ oxidation of ascorbic acid can cause for­ The interrelationship which exists tems consisting of amino acids, tea tannin mation of aldehydes from amino acids by among the several tea leaf constituents and acetone dried powder of tea leaf as oxidative deamination. Our experiments studied in this investigation, and shown in source of tea-leaf enzyme system. How­ confirm and extend these results by Figure 3, points out again (Sanderson, ever, the results of these experiments did showing that dehydroascorbic acid can in­ 1965) the importance of the chemical not exlude the possibility that specific en­ deed cause the formation of aldehydes composition of the tea leaf in determin­ zymes were causing the observed changes. from amino acids and that the aldehydes ing the organoleptic properties of the fin­ Results of our investigation have extend­ formed are those expected from a Streck­ ished black tea. Unfortunately, it must be ed these earlier findings by demonstrating er degradation reaction. stated that as yet all too little is known that a) specific enzymes do not appear to The presence of ascorbic acid itself in about what the composition of a tea leaf be operating during tea fermentation to the tea fermentation system prevented should be to enable the best black tea to bring about the conversion of amino acids both color formation and appreciable to their corresponding Strecker degrada­ conversion of amino acids to volatile alde­ tion aldehydes, i.e., the crude tea en­ hydes. Ascorbic acid is known to prevent zymes preparation alone was not capable enzymic browning caused by oxidation of of bringing about this change, and b) the phenolic compounds by reducing the oxidation of tea flavanols is sufficient in quinones formed back to their original itself to cause the conversion of amino oxidation level, thus preventing polymeri­ acids to aldehydes. zation. In the system investigated here, it It is noteworthy that enzyme systems is postulated that the quinone formed Catechol capable of converting amino acids to alde­ from EGCG is preferentially converted by Oxidase O x id iz e d ^ ^ C o lo r e d F la v a n o ls hydes have been extracted from tomato ascorbic acid back into EGCG, with the Flavanols Products fruit (Yu et al., 1968a; Yu et al., 1968b) result that no quinone is available to con­ <°2> (Theaflavins. Thearubigins. and microorganisms (MacLeod et al., vert amino acids to aldehydes. etc ) 1956; Sasaki, 1962), but that the likely The results of our investigation fit well presence of phenolic compounds in the in a scheme for tea fermentation pro­ test systems used raises the possibility posed previously (Sanderson, 1965). The that the mechanism in these systems is part of this scheme concerned with the the same as that operating in fermenting present investigation has been amplified tea leaf. to incorporate the results discussed above A study of the ability of different and is shown in Figure 3. Essential fea­ amino acids to be converted into their tures of this scheme may be summarized C o lo r e d Dehydroascorbic Acid corresponding aldehydes by the tea fer­ as follows: During the conversion of fresh P r o d u c t s mentation system showed that valine, leu­ green-tea leaf to black tea, flavanols are cine, isoleucine and phenylalanine were oxidized by catechol oxidase to quinones Fig. 3—Scheme showing relationship between converted into volatile aldehydes corre­ (or possibly to semiquinones), which tea flavanol oxidation and formation of alde­ sponding to their respective expected polymerize to form colored products hydes from amino acids during tea fermenta­ Strecker degradation products. On the such as theaflavins and thearubigins tion . 164- JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

be produced. And, of course, much re­ thearubigins as polymeric proanthocyani- ing the manufacture of black tea. J. Sci. dins. Nature 221, 742-744. Food Agr. 17, 182-188. mains to be learned about how the black- Eden, T. 1965. “Tea.” Longmans, Green and Rohan, T.A. and Stewart, T. 1967. Precursors tea manufacturing conditions (Eden, Company, Ltd., London. of chocolate aroma: Studies in the degrada­ Finot, P.A., Muggler-Chavan, F. and Vuataz, L. tion of amino acids during the roasting of 1965; Harler, 1963; Keegel, 1958) affect 1967. La phenylalanine précurseur de la Accra cocoa beans. J. Food Sci. 32, the organoleptic properties of the final phénylacétaldéhyde dans l’arôme de thé 625-629. product. It has also been shown (Finot et noir. Chimia 26-7. Rooney, L.W., Salem, A. and Johnson, J.A. Harler, C.R. 1963. “Tea Manufacture.” Oxford 1967. Studies of the carbonyl compounds al., 1967) that amino acids can give rise University Press, London. produced by sugar-amino acid reactions. I. to volatile aldehydes through interaction Keegel, E.L. 1958. “Tea Manufacture in Cey­ Model systems. Cereal Chem. 44, 539-550. lon.” Tea Research Institute, Talawakelle, Sakato, Y., Hashizume, T. and Kishinoto, Y. with black-tea solids under conditions Ceylon. 1950. Studies on the chemical constituents existing during the normal brewing of a Li, L.P. and Bonner, J. 1947. Experiments on of tea. Part V. Synthesis of theanine. J. Agr. cup of tea. Perhaps the method of brew­ the localization and nature of tea oxidase. Chem. Soc. Japan. 23, 269-71. Biochem. J. 41, 105-110. Sanderson, G.W. 1964. Extraction of soluble ing tea is another black-tea aroma deter­ MacLeod, P. and Morgan, M.E. 1956. Leucine catechol oxidase from tea shoot tps. Bio- minant. metabolism of Streptococcus lactis var. ehim. Biophys. Acta 92, 622-624. maltigenes. II. Transaminase and decarboxy­ Sanderson, G.W. 1965. On the chemical basis of REFERENCES lase activity of acetone powders. J. Dairy quality in black tea. Tea Quart. 36, Sci. 39, 1125-1133. 172-182. Bhatia, I.S. and Deb, S.B. 1965. Nitrogen me­ Nakabayashi, T. 1958. Formation mechanism Sasaki, S. 1962. On the decarboxylase operat­ tabolism of detached tea shoots. 1. Changes of black tea aroma. V. The precursor of the ing in the degradative pathway of L-leueine: in amino-acids and amides of tea shoots dur­ volatile carbonyl compounds. Nippon Nogei by Proteus vulgaris. J. Biochem. 51, ing withering. J. Sei. Food Agr. 16, Kaishi 32, 941-5. [Chem. Abstr. (1961), 335-344. 759-769. 55, 15772.] Vuataz, L., Brandenberger, H. and Egli, R.H. Biehl, B. 1967. Proteinhydrolyse während der Popov, V.R. 1956. Oxidation of amino acids in 1959. Plant phenols. I. Separation of the tea Kakao fermentation in Abhängigkeit von the presence of tannins and polyphenoloxi- leaf polyphenols by cellulose column chro­ Weckseiwirkungen mit Polyphenolen unter dase of tea. Biokhimiya 21, 383-387. matography. J. Chromatog. 2, 173-187. anaeroben und aeroben Bedingungen. Z. Popov, V.R. 1966. The role of ascorbic acid in Wickremasinghe, R.L. and Swain, T. 1964. The Lebensm.-Untersuch. -Forsch. 133, 145-158. the oxidation processes involved in the prep­ flavour of black tea. Chem. & Ind. Bokuchava, M.A. and Popov, V.R. 1954. The aration of tea. In “Biokhim. Progress Tekh- 1574-1575. significance of amino acids in the formation nol. Chai Proiz.,” ed. Bokuchava, M.A., pp# Wickremasinghe, R.L. and Swain, T. 1965. of tea aroma by interactions with tannin 128-132. Academy of Sciences of USSR, Studies of the quality and flavour of Ceylon substances under conditions of elevated Moscow. tea. J. Sci. Food Agr. 16, 57-64. temperature. Dokl. Akad. Nauk SSSR 99, Roberts, E.A.H. 1962. Economic importance of Yu, M.H., Olson, L.E. and Salunkke, D.K. 145-148. [Chem. Abstr. (1955) 49, 3439.] flavanoid substances: Tea fermentation. In 1968a. Precursors of volatile components in Bondarovich, H.A., Giammarino, A.S., Renner, “Chemistry of Flavanoid Compounds,” ed. tomato fruit. II. Enzymatic production of J.A., Shephard, F.W., Shingler, A.J. and Geissman, T.A., pp. 468-512. Academic carbonyl compounds. Phytochemistry 7, Gianturco, M.A. 1967. Some aspects of the Press, London. 555-560. chemistry of tea. A contribution to the Roberts, E.A.H. and Wood, D.J. 1951. The Yu, M.H., Salunkhe, D.K. and Olson, L.E. knowledge of the volatile constituents. Agr. amino acids and amides of fresh and with­ 1968b. Production of 3-methylbutanal from Food Chem. 15, 36-47. ered tea leaf. Current Sci. 20, 151-153. L-leucine by tomato extract. Plant Cell Brown, A.G., Eyton, W.B., Holmes, A. and Roberts, G.R. and Sanderson, G.W. 1966. Physiol. 9, 633-638. Ollis, W.D. 1969. Identification of the Changes undergone by free amino acids dur­ Ms. received 5/20/69; revised 9/8/69; accepted 9/10/69. T. H. S C H U L T Z , T. Ft. M O N a n d R. R. F O R R E Y USDA Western Utilization R&D Div., ARS, Albany, Calif. 94710

VARIATIONS OF RELATIVE RETENTION TIMES IN OPEN-TUBULAR GAS-CHROMATOGRAPHIC COLUMNS: RELEVANCY TO FRUIT VOLATILES

SUMMARY—Open-tubular gas chromatographic columns made from stainless steel tubing, coated ficiently high (e.g., 225 °C for sesquiterpenes) with methyl silicone oil, give variable retention times for individual alcohols, dependent on the size to give essentially instantaneous vaporization o f sample. This difficulty is largely overcome by using a small proportion o f adsorption-reducing and minimize adsorption, and thus avoid peak material with the methyl silicone. There are cases o f reversal of peak order, caused by differences broadening. Samples were injected with a 1-jul in column temperature or minor differences in the coating. However, with care, relative retention syringe (excepting the experiments for Figs. 1 and 2). Flame-ionization detectors made in this times are nearly constant and are helpful as an aid in identification o f compounds. A procedure for laboratory were used (Teranishi et al., 1962). precise comparison of the retention times o f unknown fruit volatiles with known compounds is The carrier gas was helium. described. Relative retention times o f representative compounds are given for twelve different stationary liquids. RESULTS & DISCUSSION INTRODUCTION stationary phase (James et al., 1952; Averill, 1962, 1965; Mon et al., 1967), by Undesired peak-order reversals ONE DESIRABLE feature of an analyti­ treatment of the solid support (Scholz et Figure 1 shows chromatograms of a so­ cal gas-chromatographic column is con­ al., 1962; Ottenstein, 1968), or by sat­ lution of 1-propanol and 3-methylpen- stancy of relative retention time of each urating the carrier gas with water vapor tane in ethanol. The column used was encountered solute, referred to a standard (Knight, 1958; Teranishi et al., 1962; coated with methyl silicone oil without compound or set of standards. Although Mon et al., 1966). additive. Although alcohols tailed badly, identification of compounds by relative An example of reversal of peak order this column was once regarded as one of retention times alone is dubious, good re­ at different column temperatures has the best for the study of orange essence. tention data are helpful along with other been presented by Littlewood (1962a), The three chromatograms were all run methods. In research on fruit volatiles an with a discussion of the thermodynamic under the same conditions of temperature individual compound may occur at wide­ principles involved. More recently a dis­ and linear velocity of helium. The only ly different concentrations in different cussion of the temperature effect and a difference was in the size of sample in­ preparations. Therefore, possible varia­ compilation of considerable pertinent re­ jected. tion of retention time with amount of so­ tention data on hydrocarbons have been The hydrocarbon (peak 2) shows the lute must be considered along with other presented by Ettre et al. (1967). same retention time in all three runs, re­ variables. A number of factors which af­ This paper deals only with open-tubu­ gardless of sample size, but the alcohol, fect constancy of relative retention times lar columns. It shows cases of peak-order 1-propanol, (peak 1 ) appears before peak have been described. reversal with variation in sample size, as a 2 when the sample is large but after peak Porter et al. (1956) discussed the ef­ result of adsorption, and the degree of 2 when the sample is smaller. fects of sample size, nonlinearity of va­ improvement observed when a surface ac­ For the experiment shown in Figure 2, por-pressure variation with concentration tive agent is included in the stationary benzene was substituted for the ethanol of the solute in the stationary phase, and phase. and other conditions remained the same. charging technique on peak size and The second part of this paper is con­ As before, the aliphatic hydrocarbon shape. Although sample size of a given cerned with desired changes in relative re­ shows the same retention time in all runs, solute affects the retention time to the tention times obtained with different sta­ but 1-propanol appears before peak 2 , be­ peak maximum, the time to the front of tionary liquids. Considerable data on this tween peaks 2 and 3, or after peak 3, the peak would appear to change very lit­ topic have been published for packed col­ depending on the sample size. A similar tle in most cases if other independent var­ umns (Tenney, 1958; Wehrli et al., 1959; movement of the 1 -propanol peak was iables are unchanged (and if adsorption is Littlewood, 1962b; Institute of Petro­ observed when heptane, which appears at not a factor). leum, 1966) and open-tubular columns 16 min, was substituted for the benzene. The effect of adsorption of the solute (Mon et al., 1966; Institute of Petroleum, The conclusion from these runs and by the solid support on relative retention 1966). others is that hydrocarbons show the times with packed columns was discussed same retention time regardless of peak in the original paper on gas chromatog­ EXPERIMENTAL size while low molecular weight alcohols, raphy by James et al. (1952) and more THE COLUMNS were made in this laboratory at least up to C6, show widely different recently by others (Eggertsen et al., 1958; from three sizes of stainless steel tubing, 0.01, retention times depending on the amount Averill, 1962; Mon et al., 1966). Scholz 0.02 and 0.03 in. ID. Except where noted present, on columns of this type. This ef­ et al. (1962) showed with a packed col­ otherwise, the tubing was rigorously cleared fect is believed to be due to adsorption of umn which exhibited adsorption that the with a series of liquids including nitric acid, am­ alcohols on the walls of the tubing. How­ relative retention times (to both the monia and solvents (Mon et al., 1967). The ever, above a certain limit, further in­ maximum and peak front) of low-molecu- columns were coated by the procedure de­ creases in sample size have only a slight lar-weight alcohols may be changed con­ scribed by Teranishi et al. (1964). With the ex­ effect on retention time. Only when the ception of the earliest experim ents (Figs. 1 and amount of alcohol is relatively high does siderably by variation of the sample size. 2), the columns were operated in altered Ther- Elimination or minimizing of adsorption m otrac ovens (Beckman Instruments, Inc., Ful­ it exceed the adsorptive capacity of the effects in both packed and open-tubular lerton, Calif.), w ith on-column inlets (Carle In­ column inner wall, and then the retention columns has been described. This was strum ents, Inc., Anaheim, Calif.) in which the time of the alcohol is controlled mainly accomplished by including small propor­ sample is vaporized in a 2-in. X 0.07-in. ID tube. by partitioning between the silicone oil tions of surface active agents in the The inlet temperature was maintained suf­ and the carrier gas.

Volume 35 119701-JOURNAL OF FOOD SCIENCE-165 166—JO URN A L OF FOOD SCIENCE-Volume 35 (1970)

* - 2 X 0 3

I S A M P L E ' INJECTION, /I V.SMALLX _ J |______J 12______> 15 ' ' 10 ' ' 5 ' ' ‘~0 TIME, MINUTES TIME,MINUTES

Fig. 1—Chromatograms o f a solution consisting Fig. 2—Chromatograms o f a solution consisting Fig. 3—Chromatograms o f a solution consisting of 65% ethanol, 34% 1-propanol, and 1% 3- of 65% benzene, 34% 1-propanol, and 1% 3- of 65% benzene, 34% 1-propanol, and 1% 2- methyI pentane. Column: 200-ft, 0.01-in. ID, methy!pentane. Column and splitter: Same as methylpentane. Column: 500-ft, 0.02-in. ID, coated with methyl silicone SF-96(50), no tail­ in Fig. 1. Column temperature: 40°C. Peaks: 1, coated with methyl silicone SF-96(50) plus 5% ing reducer, new tubing, cleaned with organic 1-propanol; 2, 3-methylpentane; 3, benzene. of Igepal CO-880. No stream splitter. Column solvents only before coating. Stream splitter at temperature: 40°C. Peaks: 1, 1-propanol; 2, column inlet: 1/200. Column temperature: 2-methylpentane; 3, benzene. 40°C. Peaks: 0, ethanol; 1, 1-propanol; 2, 3- methylpentane.

to the peak front, of 1-propanol, is con­ stant with sample size variation from large to intermediate and is only 0 . 2 min panol was constant within 0 .1 min, for later when the sample size is less than one peaks from large size down to the small­ Use of adsorption-reducing material tenth the smallest amount used in the est that could be observed. We included a small proportion of earlier column. The samples for the mid­ With the newer silicone columns and polar material such as Carbowax or Igepal dle and bottom chromatograms were too peaks of moderate size, one must still be with silicones and hydrocarbons used as small to measure but can be estimated careful to avoid reversal of peak order. stationary liquids (Averill, 1965; Mon et from the size of the benzene peaks. With Figure 4 shows portions of chromato­ al., 1967). The original purpose of the a still smaller amount, so small that the grams of two similar samples of apple es­ polar material was to eliminate or reduce 1-propanol peak could barely be seen, sence (Flath et al., 1967). The columns tailing of alcohols and other oxygenated with the detector set at five times higher were nominally the same: 0.02-in. ID, compounds and to give peaks that have sensitivity, the response came 0.7 min 500 ft long, coated with methyl silicone areas more nearly proportional to the late (chromatogram not shown). This is in SF-96(50) containing 5% of Igepal CO- amount of compound injected. This is ac­ contrast with a shift of 7 min with the 880. The only known difference is that complished by preferential adsorption of earlier column. Thus, the small proportion column A had been fully cleaned with ni­ the polar coating material on the column of Igepal applied with the stationary tric acid, ammonia, and solvents before walls. liquid acts not only to reduce tailing but coating, while column B was simply Figure 3 shows the behavior of 1-pro­ also to give nearly constant retention washed out with solvent and recoated. It panol in a column coated with methyl sil­ times for alcohols, except at very low had previously been coated with the same icone SF-96(50) plus 5% of Igepal levels. material. Apparently all the Igepal was CO-880. The test compounds were mixed In a similar experiment with a 0.02-in. not removed by the solvent and this re­ in the same proportions as before. (In­ column coated with Carbowax 20 M sulted in a higher concentration of Igepal stead of 3-methylpentane the 2-methyl only, operated at 75°C, and with a solu­ after recoating. The effect was that the isomer was used because it is more clearly tion containing only 1% of 1-propanol relative retention times of alcohols and separated from 1 -propanol on this col­ and the usual 1% of 2 -methylpentane in unsaturated aldehydes were increased umn.) In this column the retention time benzene, the retention time of the 1-pro- with respect to esters. In Figure 4 the VARIATIONS OF VOLATILES' RETENTION 77/tf£S-167

COLUMN A

» 38 40 42 TIME, MINUTES ' . 2

COLUMN B

7 5 “ C.

_l___ i i I 2 0 2 5 3 0 I5 20 TIME,MINUTES TIME, MINUTES

Fig. 4-Portions o f chromatograms o f samples of apple essence (Flath et Fig. 5-Chromatograms o f a mixture o f 75% 2-hexenal, trans and 25% ai, 1967). Column A: 500-ft, 0.02-in. ID, coated with methyl silicone ethyl 2-methylbutyrate. Column: 650-ft, 0.03-in. ID, coated with SF-96(50) plus 5% Igepal CO-880 after complete rigorous cleaning, methyl silicone SF-96(50). Carrier gas saturated with water vapor at Column B: Same specifications as column A, but coated once, then room temperature. Peaks: 1, 2-hexenal, trans; 2, ethyl 2-methylbuty- washed out with organic solvents only, and recoated. Column tempera- rate, ture: 50° during first 10 min, then programmed at 1° per min. Peaks: 31, butyl acetate; 33, 2-hexenal, trans; 34, ethyl 2-methylbutyrate; 35, 2-m ethylbutyl acetate.

by diluting the known compound further with noninterfering hydrocarbons so that the degree of enhancement will be small. three esters, peaks 31, 34 and 35 are in able effect on retention times from this about the same relative positions on col­ part of the process. Relative retention times with various sta­ tionary liquids umns A and B, but 2-hexenal, peak 33, is Because of the situation noted above, later on column B than on A, and its or­ peak fronts (the lower front face) are fre­ Table 1 presents relative retention der with respect to ethyl 2 -methylbuty­ quently more useful than peak maxima times of four compounds chosen from rate is reversed. Differences of this kind, for helping to identify compounds. Ir. a constituents found in fruit essences, rep­ but of lesser degree, are known to occur recent study on orange essence (Schultz resenting alcohols, esters, aldehydes and with columns of this type after long use, et al., 1967) by combined gas chromatog­ terpene hydrocarbons. (The individual compared to performance when the col­ raphy and mass spectrometry, supporting times were corrected by subtracting the umn was new. retention time data were obtained by the time for carrier gas to pass through the Figure 5 shows peak-order reversal as a peak enhancement (enrichment) proce­ column.) The retention times are given result of variation of column tempera­ dure. The column was the same as noted relative to that of decane. The stationary ture, with the same two compounds, in the paragraph above. A 0.30-/rl sample liquids are listed in order of increasing interaction with 1-hexanol. Data were 2 -hexenal and ethyl 2 -methylbutyrate. of the nonaqueous essence was co-inject- All three runs were made on the same ed with 0.01—0.02 yeti of a mixture of 5 taken from an earlier publication (Mon et column. At 50°C, the aldehyde appears known compounds of widely different al., 1966). first; at 75° there is no separation; at retention time. Methyl silicone has been the principal 1 0 0 ° the ester appears first. The chromatogram was superimposed stationary phase for studying fruit vola­ on a plain chromatogram of the essence, tiles at this laboratory, because of its Precise comparison of retention times with successive lining up on different good stability, low background when In actual practice with a methyl sili- sharp peaks near the peaks representing used with mass spectrometry and gener­ cone-5% Igepal, 0.02-in. ID x 500-ft col­ the added known compounds, respective­ ally good separation ability for the com­ umn, operated repeatedly with the same ly. Where there was peak enhancement, pounds present. Table 1 will be of assist­ temperature program and helium pres­ i.e., where the added compound and one ance when a different stationary liquid is sure, relative times to the peak front for of the essence constituents were the used to separate compounds of different fruit essence constituents vary only slight­ same, the enhanced peak was usually con­ chemical classes, which are close together ly, or not at all, with size of sample. Rela­ siderably larger, and its lower front face a on methyl silicone SF-96(50). tive retention times to the peak maxi­ little earlier than that of the simple peak. Carbowax 20 M is usually a good mum show considerably more variation in With hydrocarbons, esters and aldehydes, choice because of its large difference many cases. These latter variations appear the lower front face of the enhanced peak from methyl silicone. However, from this to be caused by nonlinearity of vapor was 0-0.05 min early (excepting benz- table it may be predicted that any of the pressure with concentration of solute in aldehyde, 0.15 min early) but with alco­ Apiezons, Oppanol B 10, polyphenol the stationary phase, as discussed by hols it was 0.02—0.25 min early (0.08 ether, silicone OV-225 or Tween 20 Porter et al. (1956). The small elongated min with 1-propanol; 0.25 min with 1- would do better than Carbowax 20 M in inlets (injectors) and the injection tech­ octanol). When there is any doubt, a resolving an ester and an aldehyde that nique used should preclude any appreci­ more precise matching may be attempted are not separated by methyl silicone. Of 168- JOURNAL OF FOOD SCIENCE-Volurne 35 (1970)

Table 1—Relative retention times of representative compounds. Open-tubular columns at 125°C. given by either the nonpolar Apiezon or the polar Igepal or Carbowax. There is no A m yl Stationary liquid 1-H exanol acetate O ctanal Lim onene Decane reversal of peak order. In high-resolution preparative work, a few mg of each pure Apiezon L 3’ 1 ’ 2 0 .2 8 0.35 0 .7 5 1.4 1.0 sesquiterpene were collected by use of a Apiezon C3’ 1,2 0 .2 9 0 .3 3 0 .7 5 1.3 1.0 1000-ft, 0.03-in. ID, open tubular, Igepal Apiezon J 3, 1,2 0 .3 0 0 .3 7 0 .7 8 1.4 1.0 column, with repeated 0.5-jUl injections of Oppanol B 104 ’ 1 ’ 2 0.35 0 .4 4 0.88 1.4 1.0 the binary mixture (Teranishi et al., Silicone SF-96(50)5’ 2 0 .4 4 0 .5 3 0 .9 4 1.2 1.0 1967). Silicone OV-176’ 2 0.9 1.1 2.0 2.0 1.0 Table 3 shows another example of the Silicone QF-17’ 2 1.2 1.7 3.3 1.6 1.0 separation of two sesquiterpene hydrocar­ Polyphenyl ether8 1.3 1.4 2.8 2.6 1.0 bons. This time there is a greater differ­ Silicone OV-2259’ 2 3.3 2.2 4 .9 2.6 1.0 ence in structure, but again the retention times with methyl silicone differ by only Igepal CO-88010 4.1 2.0 3.8 2.8 1.0 2 or 3% at 175° or 150°C, the tempera­ Carbowax 20 M11 4.5 2.2 4 .2 1.0 3 .0 tures we have used for preparative work Tw een 2 0 12 4 .9 2.0 4 .3 2 .9 1.0 on sesquiterpenes. In contrast to data in Purified with alumina (Mon et al., 1966). Table 2, there is considerable improve­ 2With small proportion of Igepal CO-880. ment in separation with Apiezon but 3Hydrocarbon, refined by molecular distillation. Associated Electrical Industries Ltd., Man- scarcely any separation with Carbowax, chester, England. and the latter reverses the peak order ex­ 4Polyisobutylene. Badische Anilin & Soda Fabrik AG, Ludwigshafen Am Rhein, Germany. cept at 175° where there is no separation. sMethyl silicone. General Electric, Waterford, N.Y. Apiezon and the silicones show consider­ 6Phenyl methyl silicone. Ohio Valley Specialty Chemical Co., Marietta, Ohio.. ably better separation when the column 7Trifluoropropyl methyl silicone. Dow Coming Corp., Midland, Michigan. temperature is lower. Separation with 8 “Six-ring.” Supplied by Applied Science Laboratories, Inc., State College, Pa. Carbowax improves somewhat at 125°, 9Cyanopropyl methyl-phenyl methyl silicone. Ohio Valley Specialty Chemical Co., Marietta, but is poorer again at 1 0 0 °, although the O hio. relative retention times are about the 10Nonylphenoxy polyoxyethylene ethanol. General Aniline and Film Corp., New York, N.Y. same at these temperatures. 1 'Polyoxyethylene glycol (Polyethylene glycol). Union Carbide Corp., New York, N.Y. Data from the same runs presented in 12Polyoxyethylene sorbitan monolaurate. Atlas Chemical Industries, Inc., Wilmington, Del. a different manner are shown in Table 4, which presents the resolution (ratio of horizontal distance between peak maxima to mean peak width at base). Of course, the resolution depends not only on the this group of liquids, silicone OV-225 or When two similar compounds of the relative retention times, but also on col­ Tween 20 would be expected to make a same chemical class are difficult to sepa­ umn efficiency and sample size. These good separation of ester, aldehyde and rate with a given stationary liquid, it can­ tables illustrate the degree of improve­ alcohol where the three compounds are not be predicted which other stationary ment of separation we have obtained by not separated by methyl silicone. These liquid would give the best separation. employing different stationary phases for predictions are based on the assumption This is illustrated by the following exam­ closely similar fruit volatiles. that the percent change in relative reten­ ples. Table 2 (Teranishi et al., 1967) pre­ In both examples of pairs of sesquiter­ tion time given by two different station­ sents data on the separation of two ses­ penes, Apiezon gave fairly good separa­ ary liquids is the same for all compounds quiterpene hydrocarbons which are quite tions. However, there are other cases of within a chemical class. This assumption similar. Ylangene is a geometrical isomer pairs of both mono- and sequiterpene may be useful as an approximation. of a-copaene. These compounds are sepa­ hydrocarbons which are separated poorly, Trifluoropropyl methyl silicone, QF-1, rated by SF-96(50) but only slightly, as if at all, on both methyl silicone and compared with methyl silicone, is the the retention times differ by only 2%. In Apiezon, but are resolved fairly well by only stationary phase considered here preparative work a 2 % difference is usual­ silicones OV—17 or OV—225 or by Car­ that substantially increases the relative re­ ly insufficient to enable ready separation bowax 20 M. This behavior illustrates tention times of amyl acetate and octanal and collection of the two components in again the necessity of employing several more than it does for hexanol. The sili­ a reasonable state of purity. From Table different stationary phases for a thorough cone with cyano groups, OV-225, gives a 2 it is evident that equal improvement is study of complex mixtures. high relative retention time for octanal. Relative retention times are similar with the three purified Apiezons L, C and J (Table 1) which differ in average molecular weight. For work with fruit Table 2—Relative retention times of a-copaene and ylangene (Tera­ volatiles J or L is preferable because C nishi eta!., 1967). Open-tubular columns at 150°C. bleeds from the column at a faster rate at the temperatures required. The melting Stationary liquid Ylangene tt-Copacne temperature of Lis near 50°C, so J should Apiezon C 1 ’ 2 1.00 1.05 be used for columns which are at times Oppanol B 101 ’ 2 1.00 1.04 operated near or below that temperature. SF-96(5 0)2 1.00 1.02 The last three lines of Table 1 show O V -172 1.00 1.03 similarity among stationary liquids which Q F -12 1.00 1.03 are mainly polyoxyethylene. The other Igepal CO-880 1.00 1.05 chemical groups present in Igepal CO-880 Carbowax 20 M 1.00 1.05 and Tween-20 appear to have minor ef­ 1 Purified with alumina (Mon et al., 1966). fects on the relative retention times. 2With small proportion of Igepal CO-880. VARIATIONS OF VOLATILES' RETENTION TIM ES -169

Table 3—Relative retention times of ß-copaene. Caryophyllene = Table 4—Resolution of ß-copaene and caryophyllene. 1.00. Column temperature, °C Column temperature, °C Stationary liquid 100 125 150 175 Stationary liquid 100 125 15 0 175 A piezon J 1 ’ 2 8 .4 6.6 4 .4 4 .0 Apiezon J 1 ’ 2 1.12 1.10 1.08 1.06 Silicone SF-96(50)2 3 .0 1.9 2.1 1.1 Silicone SF-96(50)2 1.06 1.05 1.03 1.02 Silicone OV-172 2.0 1.8 1.1 ~ 0 .3 Silicone OV-172 1.02 1.01 1.05 1.03 Carbowax 20 M ~ 0 .6 ~ 0 .7 ~ 0 .4 0 Carbowax 20 M 0 .9 8 0 .9 8 0 .9 9 1.00 1 Purified with alumina (Mon et al., 1966). Purified with alumina (Mon et al., 1966). 2With 5% of Igepal CO-880. 2With 5% of Igepal CO-880. Columns: 0.02 in. ID; efficiency in the order of 100,000 theoretical plates at 150°C. Sample size: 0.02 p \ of a 25% /3-copaene-75% caryophyllene mix­ REFERENCES ture. Linear velocity of helium: 31 cm per sec at 150°C. Averill, W. 1962. Columns with minimum liq­ uid phase concentration for use in gas-liquid chromatography. In “Gas Chromatog­ raphy,” eds. Brenner, N., Callen, J.E. and Weiss, M.D., pp. 1—6. Academic Press, New York and London. Averill, W. 1965. The analysis of citrus oils by gas chromatography. Presented at the 16th Littlewood, A.B. 1962a. “Gas Chromatog­ Tenney, H.M. 1958. Selectivity of various liq­ Pittsburgh Conference on Analytical Chem­ raphy,” pp. 72—81. Academic Press, New uid substrates used in gas chromatography. istry and Applied Spectroscopy, Pittsburgh, York and London. Anal. Chem. 30: 2. Pennsylvania. Littlewood, A.B. 1962b. “Gas Chromatog­ Teranishi, R., Buttery, R.G. and Lundin, R.E. Eggertsen, F.T.L. and Knight, H.S. 1958. Gas raphy,” p. 102. Academic Press, New York 1962. Gas chromatography. Direct vapor chromatography. Effect of type and amount and London. analyses of food products with programmed of solvent on analysis of saturated hydrocar­ Mon, T.R., Forrey, R.R. and Teranishi, R. temperature control of dual columns with bons. Anal. Chem. 30: 15. 1966. Relative retention times with open dual flame ionization detectors. Anal. Chem. Ettre, L.S. and Billeb, K. 1967. Considerations tubular and packed column gas chromatog­ 34: 1033. on the retention index concept. 1. Reten­ raphy. J. Gas Chromatog. 4: 176. Teranishi, R. and Mon, T.R. 1964. Large-bore tion index and column temperature. J. Mon, T.R., Forrey, R.R. and Teranishi, R. capillary and low-pressure drop packed col­ Chromatog. 30: 1. 1967. Effects of addition of adsorption- umns. Anal. Chem. 36: 1490. Flath, R.A., Black, D.R., Guadagni, D.G., reducing material with open tubular and Teranishi, R. and Heathcock, C.H. 1967. Isola­ McFadden, W.H. and Schultz, T.H. 1967. packed column gas chromatography. J. Gas tion of copaene and ylangene from various Identification and organoleptic evaluation Chromatog. 5: 497. essential oils. Presented at the 154th meet­ of compounds in Delicious apple essence. J. Ottenstein, D.M. 1968. Comparison of support ing of the American Chemical Society, Agr. Food Chem. 15: 29. deactivation in gas chromatography. J. Gas Chicago, 111. Institute of Petroleum. Gas Chromatography Chromatog. 6: 129. Wehrli, A. and Koväts, E. 1959. Gas-chromato- Discussion Group, Data Subcommittee. Porter, P.E., Deal, C.H. and Stross, F.H. 1956. graphische Charakterisierung organischer 1966. Retention indices of solutes on squa- The determination of partition coefficients Verbindungen. Helv. Chim. Acta 42: 2709. lane, dinonyl phthalate, and polyethylene from gas-liquid partition chromatography. J. Ms. received 9/23/69; accepted 11/11/69. glycol 400. J. Gas Chromatog. 4:1. Am. Chem. Soc. 78: 2999. James, A.T. and Martin, A.J.P. 1952. Gas-liquid Scholz, R.G. and Brandt, W.W. 1962. The ef­ The authors are grateful to Roy Teranishi partition chromatography: the separation fect of solid supports on retention volumes. for many helpful suggestions. They thank R.A. and micro-estimation of volatile fatty acids In “Gas Chromatography,” eds. Brenner, N., Flath for the chromatograms of apple essence from formic acid to dodecanoic acid. Bio- Callen, J.E. and Weiss, M.D., pp. 7—26. and D.C. Patterson for assistance. chem. J. 50: 679. Academic Press, New York and London. Reference to a company or product name Knight, H.S. 1958. Gas-liquid chromatography Schultz, T.H., Black, D.R., Bomben, J.L., Mon, does not imply approval or recommendation of of hydroxyl and amino compounds. Produc­ T.R. and Teranishi, R. 1967. Volatiles from the product by the U.S. Department of Agricul­ tion of symmetrical peaks. Anal. Chem. 30: oranges. 6. Constituents of the essence iden­ ture to the exclusion of others that may be 2030. tified by mass spectra. J. Food Sci. 32: 698. suitable. VERNON R. YOUNG and ABEL VILLARREAL a

Department of Nutrition and Food Science

M assachusetts Institute o f Technology, Cambridge, M assachusetts 02139

EFFECT IN RATS OF PARTIAL REPLACEMENT OF COW'S MILK PROTEIN BY SUPPLEMENTARY NITROGEN

SUMMARY—A number of experiments studied growth of weanling rats and urinary nitrogen 15% protein from unsupplemented skim milk. excretion in young adult rats when cow's milk protein was partially replaced by varying levels o f This concentration of dietary protein was used, supplementary nitrogen. A mixture o f nonessential L-amino acids (NEAA) or a mixture o f diam­ therefore, as a control level in the growth stud­ monium citrate and glycine (DAC-Gly) was used as the source o f nitrogen. Substitution o f the 15% ies. The net protein utilization (NPU) of skim milk protein by rat assay (Bender et al., 1957) milk protein diet to the extent of 10% slightly reduced growth; significant growth reduction was 71 ± 2 compared with 63 ± 2 for unsupple­ occurred with substitutions o f 20% and greater. Fortification o f the diets containing 10.5% milk mented casein as a standard. protein and the supplementary nitrogen sources with sulfur amino acids did not restore growth to Male weanling rats (Charles River Laborato­ the maximum rate obtained with the 15% milk protein diet. Additional supplementation with ries, North Wilmington, Mass., CD strain) were tryptophan further improved growth slightly but not to the maximum rate. Several other essential fed the experimental diets ad libitum for 14 amino acids, alone or in combination, had no apparent effect. Based on urinary nitrogen excretion, days. The basal diets contained: dried skim comparable results were obtained with young adult rats by substituting nitrogen for m ilk protein in milk (Dried Nonfat Skim Milk, Carnation Com­ the diets. The reduced performance following substitution of milk protein with supplementary pany, Van Nuys, California), 28.1 to 50.5% nitrogen may be due partly to decreased utilization o f sulfur amino acids and possibly to decreased (10-18% protein); corn oil, 10%; salt mixture, utilization o f all essential amino acids. 5% (Rogers et al., 1965); vitamin mixture, 0.5% (Rogers et al., 1965); choline chloride, 0.2%; INTRODUCTION the source of dietary protein. A mixture and dextrin isucrose (2:1 weight ratio) to make of nonessential L-amino acids (NEAA), a total of 100%. In all diets lactose was partially WHEN ESSENTIAL amino acids form patterned as in cow’s milk, or a mixture substituted for the dextrin:sucrose mixture to the sole dietary source of protein for rats, of diammonium citrate and glycine achieve the lactose level present in the control a lower growth rate is observed than diet (15% skim milk protein). Weekly weight (DAC-Gly) was used as supplementary when the same level of dietary N is sup­ gain and daily food intake were recorded. nitrogen to replace part of the protein plied by a well-balanced amino acid The essential amino acid content of the nitrogen. The influence of the protein ni­ mixture containing approximately equal basal diet, calculated from the data of Orr et al. trogen replacement on growth rate was amounts of essential and nonessential (1957), and the essential amino acid require­ studied. Some experiments evaluated the ments for rat growth as reported by Rama Rao amino acids (Rose et al., 1948; Stucki et effects of adding several essential amino et al. (1964) are given in Table 2. The diet ap­ al., 1962; Young et al., 1968). acids to the diets containing the supple­ pears to meet all requirements although it is Animal proteins of high biological val­ mentary N sources. borderline for the sulfur amino acids, methio­ ue may contain some o r a ll o f the essen­ The influence o f age was also evalu­ nine and cystine. In some experiments methio­ tial amino acids, expressed per unit of nine and cystine (all L-amino acids purchased ated because available data indicate that total protein nitrogen, in excess of re­ from General Biochemicals, Inc., Chagrin Falls, for both man (Rose, 1957; Holt et al., quired amounts for the rat. Bender Ohio) were added in a ratio by weight of 3 to 1 1965) and rats (Rama Rao et al., 1964; (1 9 6 5 ) suggested tha t egg pro te in con­ in order to reach a calculated sulfur amino acid Smith et al., 1967) the requirements for tains a surplus of 10—20% of all essential content of 0.65% in the diet. essential amino acids per g total nitrogen amino acids. In this case, essential amino Experiments 1 and 2 differ in young and adults. acids are catabolized to meet calorie Experiment 1 was designed to determine the needs or are used for synthesis of nones­ EXPERIMENTAL & RESULTS extent to which a mixture of NEAA could iso- sential amino acids and other N-contain- nitrogenously replace a 15% milk protein diet Methods ing compounds. Therefore, isonitroge- without affecting growth in weanling rats. Body nous replacement of the excess essential A preliminary experiment (Table 1) con­ weight gains are summarized in Table 3. A high­ amino acids with supplementary nitrogen firmed general experience that maximum ly significant (P < 0.01) decrease in growth, growth rates are achieved with a diet containing PER, and food efficiency was observed only for (from nonessential amino acids or other utilizable N sources) should not decrease dietary N utilization and animal perform­ ance. For man, studies in this laboratory Table 1—Effect o f dietary level o f skim m ilk protein on the 14-day (Scrimshaw et al., 1966; Huang et al., growth rate o f weanling rats. 1966) and in others (Snyderman et al., 1962, Kofranyi et al., 1964) suggested Dietary skim Food Weight that high quality protein N can be partial­ Group milk protein intake gain1 FE2 ly replaced with supplementary nitrogen No. % g/day g /14 days % PER3 without reducing nitrogen retention. 1 10 8.8 45.5 ± 3.3 36.9 3.69 In the present series of experiments, 2 12 8.7 56.4 ± 1.5 46.2 3.86 weanling rats were fed dried skim milk as 3 15 9.9 76.4 ± 4 .0 54.8 3.67 4 18 9.6 7 7 .9 + 2 .1 57.8 3.22 1 Beginning weight of rats was 45 ± 2 g. Weight gain is mean ± stan­ aPresent address: Department of Biochem- dard error for 8 rats per group. istry, School of Medicine, University of Panama, 2Food efficiency (FE) is (g gain per g food intake) X 100. Panama City, Panama. 3 Protein efficiency ratio (PER) is g gain per g protein consumed.

M O-JOURNAL OF FOOD SCIENCE-Volume 35 (1970) EFFECTS OF NITROGEN ON A 4 7 3 - 1 7 1

Table 2—Calculated % of essential amino acids in experimental diets compared with requirements for young growing rats. Experimental diets

% of milk protein N replaced Amino acids 2 Basal1 10 15 20 30 required Amino acid Isoleucine 0.96 0.86 0.82 0.77 0.67 0.55 Leucine 1.47 1.32 1.25 1.18 1.03 0.70 Lysine 1.17 1.05 0.99 0.94 0.82 0.90 S-amino acids (Methionine + cystine) 0.50 0.45 0.43 0.40 0.35 0.50 Aromatic amino acids (Phenylalanine + tyrosine) 1.49 1.34 1.27 1.19 1.04 0.72 Threonine 0.69 0.62 0.59 0.55 0.48 0.50 Tryptophan 0.20 0.19 0.18 0.17 0.15 0.11 Valine 1.03 0.93 0.88 0.83 0.72 0.55 Arginine 0.55 0.50 0.47 0.44 0.39 — Histidine 0.39 0.35 0.33 0.31 0.27 0.25 1 Calculated values from data of Orr et al. (1957). 2 From Rama Rao et al. (1964).

the 30% substitution in this experiment. ious essential amino acids to diets containing milk protein N while concentrations of 0.65% Therefore, a second experiment was con­ 30% of total dietary N from the NEAA mix­ sulfur amino acids and 0.21% tryptophan were ducted with larger groups of rats to evaluate the ture. The added amino acids achieved the con­ maintained as in the control diet. The results, effect of 20% and 30% replacements of milk centrations calculated to be present in the 15% summarized in Table 6, show that a 10% re­ protein N by two sources of supplementary ni­ milk protein diet. Arginine, lysine, or threonine placement of milk protein produced a small but trogen. The results are summarized in the lower addition to the diets (Expt. 4, Table 5) failed to nonsignificant (P > 0.05) decrease in growth half of Table 3. Significant (P < 0.01) decreases improve weight gain significantly. rate of weanling rats. A significant reduction (P in growth rate, FE, and PER occurred for both The effects of adding tryptophan, trypto­ < 0.01) in growth rate did occur when 20% of sources of nitrogen at the two substitution lev­ phan and threonine, or a combination of several the protein was replaced by DAC-Gly. Propor­ els. essential amino acids, were also studied (Expt. tionately greater reductions were observed with 5, Table 5). Tryptophan significantly (P < still higher rates of replacement by DAC-Gly Experiment 3 0.05) improved the growth rate, mainly as a despite the unchanged amounts of sulfur amino Because the sulfur amino acids, methionine consequence of higher food intake (group 4). acids and tryptophan in the diet. and cystine, are limiting in cow’s milk protein Addition of tryptophan and threonine together Experiment 7 for weanling rats, a third experiment was con­ or in combination with arginine and lysine did In young adult rats isonitrogenous replace­ ducted to determine the influence of DAC-Gly not result in a higher growth rate than that ment of milk protein with the DAC-Gly mix­ substitution of milk protein N in diets supple­ achieved with tryptophan alone. mented with sulfur amino acids. Table 4 sum­ ture was studied by monitoring changes in uri­ Experiment 6 marizes these results. Comparison of groups 5 nary nitrogen excretion. The basal diet con­ and 9 reveals that inclusion of DAC-Glv in the DAC-Gly replaced from 10-50% of skim tained 4% milk protein, while the remaining diet resulted in a lower weight gain (P < 0.05) than when the diet contained 10.5% milk pro­ tein alone. This effect appeared to be related to the reduced food intake, particularly because efficiency of food conversion was similar for Table 3—Fourteen-day growth of weanling rats during isonitrogenous replacement o f skim m ilk both groups. by nonessential amino acids or glycine and diammonium citrate. Growth rates were similar for diets contain­ ing 10.5% milk protein, with or without added Level of milk % total dietary Food Weight DAC-Gly, supplemented with 0.15-0.28% sul­ protein in diet Nitrogen N from added intake gain FE fur amino acids (compare group 6 vs. 10, 7 vs. % source N source g/day g /14 days % PER 11). The highest level of S-amino acid supple­ mentation reduced body weight gain signi­ Expt. 1 1 15.00 — 0 9.4 73.8 ± 4 .0 56.1 3.74 ficantly when the diet contained DAC-Gly 13.50 NEAA2 10 9.3 69.1 ± 2 .2 53.1 3.54 (group 8) but not when the only other dietary 12.75 NEAA 15 9.4 66.8 ± 1 .9 50.8 3.38 N was milk protein (groups 4 and 12). These 12.00 NEAA 20 10.0 67.0 ± 2 .3 47.9 3.19 findings suggest that the DAC-Gly mixture af­ 10.50 NEAA 30 9.3 56.0 ± 2 .4 43.0 2.87 fected utilization of sulfur amino acids at the Expt. 23 — 9.8 68.0 ± 2 .4 49.6 3.30 high and possibly at the low dietary sulfur 15.0 0 amino acid levels. 12.0 NEAA 20 9.3 53.4 ± 2 .2 41.0 2.73 Although addition of 0.15-0.28% S-amino 12.0 DAC-Gly4 20 9.9 56.8 ± 2.1 41.0 2.73 acids to 10.5% milk protein diets containing 10.5 NEAA 30 8.9 48.3 ± 2 .0 37.8 2.58 DAC-Gly allowed weight gains which were simi­ 10.5 DAC-Gly 30 9.5 49.5 ± 2.5 37.2 2.48 lar to the 10.5% milk protein diets without the 1 Beginning weight of rats was 36 ±2 g. Weight gain is mean ± standard error for 8 rats per group. added nitrogen, growth rate remained lower 2Amino acid (g) per 100 gmixture: L-alanine, 6.5; L-aspartic, 13.72; L-glutamic,44.03; glycine, than that on the 15% milk protein diet which 3.72; L-proline, 20.90; L-serine, 11.09. supplied the same level of total N and was sup­ beginning weight of rats was 48 ± 3 g. Weight gain is mean ± standard error for 14 rats per plemented with sulfur amino acids. group. Experiments 4 and 5 4Diammonium citrate and glycine mixture. Each compound supplied equal amounts of N per A study was conducted with addition of var­ unit weight of mixture. M 2—JOURNAL OF FOOD SCIENCE- Volume 35 (1970)

Table 4—Fourteen-day growth of weanling rats on diets containing skim m ilk protein supplemented by S-amino acids, with and without added glycine and diammonium citrate (Experiment 3).______Level of milk % total dietary % dietary Food Weight . 3 Group protein in diet N from added Methionine & methionine intake gam FE • 2 No. % DAC-Glycine cystine added1 & cystine g/day g /14 days % PER

1 15.0 — _ 0.50 10.3 73.3 ± 4 .5 50.8 3.49 2 15.0 - 0.15 0.65 10.6 79.8 ± 3 .2 53.8 3.55 3 15.0 - 0.28 0.77 10.6 81.1 ± 4 .1 54.6 3.58 4 15.0 - 0.43 0.92 10.7 81.3 ± 3.1 54.3 3.52 5 10.5 30 - 0.35 8.9 49.0 ± 4.1 39.3 2.62 6 10.5 30 0.15 0.50 9.8 62.7 ± 5.1 45.7 3.02 7 10.5 30 0.28 0.65 11.1 73.3 ± 3.2 47.2 3.09 8 10.5 30 0.43 0.77 10.7 64.6 ± 2.1 43.1 2.79 9 10.5 - - 0.35 11.2 61.0 ± 4.2 38.9 3.71 10 10.5 - 0.15 0.50 10.9 67.1 ± 3.5 44.0 4.13 11 10.5 - 0.28 0.65 11.3 73.7 ± 4.8 46.6 4.32 12 10.5 - 0.43 0.77 12.2 78.9 ± 5.1 46.2 4.23 'Methionine and cystine were added in a weight ratio of 3:1. 2 Calculated values. beginning weight of rats was 45 ± 3 g. Weight gain is mean ± standard error for 7 rats per group.

dietary components were similar to those in protein diet supplied 0.13%. However, with a 2, 3 and 4 were given diets in which the DAC- previous experiments. Lactose was maintained 30% DAC-Gly substitution, isoleucine and thre­ Gly mixture replaced 10, 15 and 20% of the at the basal (4% milk protein) level in all diets. onine also were probably deficient. milk protein, respectively. During the final 10 Except for sulfur amino acids, the basal diet The rats, housed individually in metabolic days (days 31-40, Period 3) all groups were re­ contained concentrations of essential amino cages, were given 10 g of the basal diet daily for turned to the basal 4% milk protein diet. acids in excess of those suggested by Smith et an initial 10-day adaptation period followed On day 11 of the experiment, daily urine al. (1967) for maintenance in young adult rats. by a 10-day control interval (Period 1). For 10 collections began. Two 5-day pooled samples These authors suggested 0.29% for the total sul­ days (Period 2) after the control phase, group 1 of urine were prepared for each period and ana­ fur amino acids, whereas the present 4% milk was maintained on the basal diet, while groups lyzed for nitrogen by the Kjeldahl method.

Table 5—Fourteen-day growth o f weanling rats on diets containing skim m ilk protein supplemented with sulfur amino acids, NEAA, and various essentia! amino acids. Level o f milk % total dietary Essential amino Food Weight Group protein in diet' N from added acid added intake gain3 FE No. % NEAA mixture2 % g/day g /14 days % Experiment 4

1 15.0 - - - 12.1 85.6 ± 3 .0 50.5 3.34 2 10.5 30 - - 12.1 68.4 ± 2 .3 40.4 2.64 3 10.5 30 L-Arg. 0.12 12.3 71.0 ±3.1 41.2 2.67 4 10.5 30 L-Lys. 0.13 11.9 69.0 ± 2 .2 41.1 2.68 5 10.5 30 L-Threo. 0.10 12.3 75.4 ± 2.7 43.8 2.84 6 10.5 - - - 12.2 70.4 ± 3 .1 41.2 3.81 7 10.5 - L-Arg. 0.12 13.4 74.9 ± 3.3 39.9 3.66 8 10.5 - L-Lys. 0.13 13.6 70.2 ± 2.5 36.9 3.37 9 10.5 - L-Threo. 0.10 12.2 71.8 ± 2 .3 42.0 3.86 Experiment 5

1 15.0 - - - 11.7 80.7 ± 2.7 49.2 3.25 2 10.5 30 - - 11.3 65.4 ± 2 .9 41.3 2.70 3 10.5 30 L-Threo. 0.10 11.1 63.0 ± 3 .3 40.5 2.63 4 10.5 30 L-Try. 0.06 12.5 74.0 ± 4 .8 42.3 2.75 5 10.5 30 L-Threo. 0.10 11.5 71.0 ± 4 .2 44.1 2.85 + L-Try. 0.06 6 10.5 30 L-Threo. 0.10 11.1 68.3 ± 4 .2 44.0 2.80 + L-Try. 0.06 + L-Arg. 0.12 + L-Lys. 0.13 115% milk protein diets were supplemented with 0.15% methionine and cystine (3:1) and the 10.5% milk protein diets with 0.3%. 2 Composition of the NEAA mixture is given in Table 3. J Beginning weight of rats was 45 ±3 g. Weight gain is mean ± standard error for 8 rats per group in Experiment 4 and 7 rats per group in Experi- m ent 5. EFFECTS OF NITROGEN ON RATS- 1 7 3

Table 6—Fourteen-day growth of weanling rats fed diets with different levels o f DAC-Glycine for days 25-30 of Period 2, and for the first replacement o f skim m ilk protein supplemented with sulfur amino acids and tryptophan (Experi­ five days (days 31—35) of Period 3. The results m e n ts ). show that even with the concentration of S- amino acids and tryptophan maintained at the Level of milk Amino acid % total dietary Food Weight level present in the basal diet, urinary nitrogen Group protein in diet added N from added intake gain1 FE excretion increased for both levels of substitu­ No. % % DAC-Glycine g/day g/14 days % PER tion with DAC-Gly. However, maintenance of the dietary S-amino acid and tryptophan levels 1 15.0 Met-Cys 0.15 0 11.2 84.8 ± 2 .4 54.1 3.57 appeared to result in a smaller urinary nitrogen 2 13.5 Met-Cys 0.20 10 11.1 81.2 ± 2 .5 52.3 3.43 increase following 20% replacement than was + Try 0.02 observed in the earlier experiment without sup­ 3 12.0 Met-Cys 0.25 20 10.3 68.4 ± 4.4 47.4 3.10 plemental dietary S-amino acids (Table 7). + Try 0.04 4 10.5 Met-Cys 0.30 30 9.7 61.0 ± 2 .3 44.9 2.92 + Try 0.06 DISCUSSION 5 7.5 Met-Cys 0.40 50 8.3 35.0 ± 2 .8 30.1 1.94 GROWTH depression resulted from isoni- + Try 0.10 trogenous substitution of cow’s milk pro­ 1 Beginning weight of rats was 55 ± 3 g. Weight gain is mean ± standard error for 7 rats per group. tein by supplementary nitrogen, either a mixture of nonessential L-amino acids, patterned as in cow’s milk, or DAC-Gly. Various investigators have examined the interrelationship between glycine and Rats were weighed at the beginning of the than for rats given the basal diet throughout methionine. Glycine is known to alleviate study (mean weight 200 g) and at the end of (group 1). the growth depressing effects of high lev­ each dietary period. They showed little change In contrast to these results, the reduced uri­ els of dietary methionine (Hardin et al., in weight during the entire 40-day span. nary N excretion following return to the basal 1951; Roth et al., 1950; Benevenga et al., Table 7 shows that 15 or 20% substitution diet was highly significant (P < 0.01) for all 1967; Wertz et al., 1968) and methionine of milk protein with DAC-Gly resulted in a sig­ groups given DAC-Gly. The control group nificant increase (P < 0.01) in levels of urinary (group 1), maintained throughout the experi­ protects against excess glycine (Wertz et nitrogen excretion during the first five days of ment on the basal diet, did not show a signifi­ al., 1968). Period 2 (days 21-25). During the last five cant change in urinary N excretion (P > 0.1) In the present studies, DAC-Gly was days of this period, urinary nitrogen excretion between Periods 2 and 3. utilized as a source of supplementary ni­ decreased for groups 2 and 3 (10 and 15% sub­ trogen. The additional dietary glycine Experim ent 8 stitution) but was still high for group 4 (20% may have limited the utilization of the substitution). Because the maintenance requirement of sulfur amino acids and resulted in the young adult rats for S-amino acids (methionine When the rats were returned to the basal lower growth rates achieved when DAC- diet (Period 3) a sharp reduction in urinary N and cystine) may be higher than the amount Gly was added to the 10.5% milk protein output occurred in groups 2-4, indicating im­ supplied by the basal diet (Smith et al., 1967), proved retention of dietary nitrogen on the the effect of partial replacement of milk pro­ diet. Growth rate was reduced to the basal diet as compared with partial replacement tein nitrogen by DAC-Gly was studied while a same extent following the 20% or 30% of skim milk protein N by supplementary nitro­ constant dietary level of S-amino acids and substitutions of milk protein nitrogen by gen. tryptophan was maintained. In this study, the both DAC-Gly and NEAA mixture. Urinary nitrogen excretion during the last basal diet and 20% and 30% levels of substitu­ Also, in rats given the methionine- five days of Period 2 was significantly greater (P tion were utilized. The experimental design was supplemented diets containing 10.5% similar to that of Experiment 7, except that six < 0.01) for rats fed the diet containing 20% of milk protein plus nitrogen, growth rates rats were used per dietary group. N from DAC-Gly compared with other groups. remained lower than the rate of rats on However, the level of urinary nitrogen excre­ Table 8 shows values of urinary nitrogen ex­ the 15% milk protein diets with supple­ tion was not significantly higher (P > 0.05) for cretion for the last five days (15—20) of Period mentary methionine and cystine. There­ rats fed diets with the 10 and 15% substitution 1 during feeding of the basal 4% protein diet, fore, in addition to possible alterations in sulfur amino acid metabolism, decreased utilization of all essential amino acids may have been responsible for the re-

Table 7—Effect of 40-day replacement of milk protein by DAC- Glycine on urinary nitrogen excretion in young adult rats (Experi­ m e n t 7).

% of Milk protein N replaced1 Table 8—Urinary nitrogen excretion in young adult rats fed milk protein diets, supplemented with sulfur amino acids and tryptophan, 0(1) 10(2) 15(3) 20(4) following partial replacement with diammonium citrate and glycine for 40 days (Experiment 8). Days 2 ------(Mg urine N per day)3 expt. % of Milk protein N replaced 40.4 ± 1.5 1 6 -2 0 40.3 ± 1.2 42.8 ± 2.0 42.4 ± 1.6 0 20 30 2 1 -2 5 40.0 ± 1.0 43.7 ± 2 .0 46.7 ± 1.6 45.9 ± 1 .3 Days of 2 6 -3 0 38.0 ± 1 .4 39.7 ± 1 .9 41.1 ± 1 .2 44.5 ± 1.1 ex p t.1 ------(Mg urine N per day)2 3------3 1 -3 5 37.0 ± 0 .9 35.6 ± 1 .0 36.0 ± 1 .1 36.0 ± 1 .3 16-20 44.0 ±1.1 42.3 ±1.6 42.5 ± 1.7 3 6 -4 0 38.7 ± 0 .8 35.6 ± 1.8 36.6 ± 1 .3 37.3 ± 1.0 25-30 40.9 ±2.1 45.5 ±2.3 47.8 ±1.2 1 Group number is in parentheses. 3 1 -3 5 ______39.5 ± 1.6______38.9 ± 2 .7 35.5 ± 1.8 2 During days 16-20 and 31—40 all rats were given the basal 4% milk during days 16-20 and 31-35 all rats were given the basal diet protein diet. During days 21 — 30 groups of rats were fed the basal diet and during days 25-30, rats were fed the basal diet or diets containing or a diet containing 10, 15 or 20% of milk protein N replaced with DAC- 20 and 30% of total dietary N from DAC-Glycine. All diets contained Glycine. 0.13% methionine and cystine and 0.06% tryptophan. 3Mean ± standard error for 10 rats per group. 2 Mean ± standard error for 6 rats per group. W 4—JOURNAL OF FOOD SCIENCE-Volume 35 (1970) duced growth rates following replacement with or without supplementary sulfur total nitrogen ratio for beef protein fed to young men. J. Nutr. 90: 416. of the milk protein by the nitrogen. amino acids, only slightly and nonsignifi- Kofranyi, E. and Jekat, F. 1964. Zur Bestim- The essential amino acid pattern of the cantly reduced the growth rate in the mung der biologisehem Wertigkeit von Nahrungsproteinen. IX. Der ersatz von test protein probably is an important fac­ present study. Proportionately greater, hochwertigem eiweiss durch nichtessen- tor modifying response to substitution of significant reductions were observed at tiellen stickstoff. Hoppe-Seyler’s Z. Physiol. higher levels o f substitution. These results Chem. 338: 154. that protein by supplementary nitrogen. Moran, E.T., Summers, J.D. and Pepper, W.F. Braham et al. (1968) observed that addi­ differ from those in children (Snyderman 1967. Effect of nonprotein nitrogen supple­ et al., 1962) and adult men (Kofranyi et mentation of low-protein rations on laying tion of nonessential amino acids or DAC hen performance with a note on essential to egg protein did not influence weight al., 1964; Scrimshaw et al., 1969), who amino acid requirements. Poultry Sci. 46: gain of rats. Addition of either or both of apparently are able to tolerate at least a 1134. Orr, M.L. and Watt, B.K. 1957. Amino acid these nitrogen sources to casein, soybean, 10—20% replacement of m ilk protein with content of foods. Home Econ. Res. Report or cottonseed protein decreased weight supplementary nitrogen. No. 4, USDA, Washington, D.C. Rama Rao, P.B., Norton, H.W. and Johnson, gain. These authors suggest that the re­ When urinary nitrogen excretion was B.C. 1964. The amino acid composition and sponse to nonessential amino acids or the major parameter used for determining nutritive value of proteins. V. Ammo acid the effects of milk protein substitution requirements as a pattern for protein evalua­ DAC depends largely on the essential tion. J. Nutr. 82: 88. amino acid pattern of the protein tested. by nitrogen in young adult rats, the re­ Rogers, Q.R. and Harper, A.E. 1965. Amino sults also indicated lessened N utilization acid diets and maximal growth in the rat. J. Moran et al. (1967) observed a depres­ Nutr. 87: 267. sion in egg production and weight gain of (Tables 7 and 8). The increase in urinary Rose, W.C. 1957. The amino acid requirem ents nitrogen excretion was greater with of adult man. Nutr. Abstr. Rev. 27: 631. hens when DAC was added to a 10% pro­ Rose, W.C., Oesterling, M.J. and Womack, M. tein diet of which half the protein was higher levels of substitution. This adverse 1948. Comparative growth on diets contain­ supplied by corn meal and half by soy­ effect on nitrogen balance was not pre­ ing ten and nineteen amino acids with fur­ vented by maintenance of the S-amino ther observations upon the role of glutamic bean meal. Recently Daniel et al. (1968) and aspartic acids. J. Biol. Chem. 176: 753. have reported that addition of L-glutamic acid and tryptophan content of the diets Roth, J.S., Allison, J.B. and Milch, L. J. 1950. at the level present in the control 4% milk The sulfur balance of rats fed excess DL- acid to diets containing 5.0 to 8.5% milk methionine plus glycine or DL-alanine. J. protein adversely affected growth rates of protein diet. Biol. Chem. 186: 113. rats. On the other hand, Venkat Rao et Substitution of the milk protein nitro­ Scrimshaw, N.S., Young, V.R., Schwartz, R., Piehe, M.L. and Das, J.B. 1966. Minimum al. (1964) showed that, although weight gen by supplementary nitrogen sources at dietary essential amino acid-to-total nitro­ gain was proportional to the quantity of the 10 or 15% levels, without additional gen ratio for whole egg protein fed to young sulfur amino acids, resulted in increased men. J. Nutr. 89: 9. egg protein in the diet, addition of L-glu­ Scrimshaw. N.S., Young. V.R., Huang, P.-C., tamic acid to the egg protein diets did not urinary nitrogen excretion. However, a Thanangkul, O. and Cholakos, B. 1969. Par­ adversely a ffe ct gro w th rates in th is case. partial adaptation to protein replacement tial dietary replacement of milk protein by non-soecifie nitrogen in young men. J. Nutr. The present findings showed that with DAC-Gly appeared to occur, as indi­ 98: 9 cated by the decreased rate of urinary ni­ Smith, E.B. and Johnson, B.C. 1967. Studies of tryptophan supplementation of the amino acid requirements of adult rats. Brit. 10.5% milk protein diet containing added trogen excretion during the latter part of J. Nutr. 21: 17. DAC-Gly and sulfur amino acids resulted the replacement period. These results Snyderman, S.E., Holt, L.E., Dancis, J., Roit- with young adult rats are, therefore, gen­ man, E., Boyer, A. and Ballis, M.E. 1962. in a small improvement of growth rate. Unessential nitrogen—a limiting factor for Addition of several other essential amino erally similar to results obtained with human growth. J. Nutr. 78: 57. weanling rats. Stucki, W.P. and Harper, A.E. 1962. Effects of acids suspected as being limiting in the altering the ratio of indispensable to dispen­ methionine-supplemented 10.5% milk REFERENCES sable amino acids in diets for rats. J. Nutr. protein diets, with or without added ni­ 78: 278. Bender, A.E. 1965. The balancing of amino Swendseid, M.E., Hickson, J.B. and Friedrich, trogen, failed to improve growth rates. acid mixtures and proteins. Proc. Nutr. Soc. W.B. 1962. Effect of nonessential nitrogen However, Swendseid et al. (1962), using Engl. Scotl. 24: 190. supplements on growth and on amino acid Bender, A.E. and Doell, B.H. 1957. Note on the content in plasma and muscle of weanling an 8% casein diet supplemented with me­ determination of net protein utilization by rats fed a low-protein diet. J. Nutr. 78: 115. thionine, found that the growth depres­ carcass analysis. Brit. J. Nutr. 11: 138* Venkat Rao, S., Daniel, V.A., Joseph, A.A., sion resulting from addition of several Benevenga, N.J. and Harper, A.E. 1967. Allevia­ Sankara, A.N. and Swaminathan, M. 1964. tion of methionine and homocystine toxic­ The pattern of amino acid requirements of sources of nonspecific nitrogen disap­ ity in the rat. J. Nutr. 93: 44. albino rat for optimal growth. J. Nutr. peared after addition of tryptophan and Braham, J.E. and Bressani, R. 1968. Dilution of Dietet. (Coimbatore, India) 1: 103. protein with nonessential amino acids and Wertz, J.M., Seward, C.R., Hove, E.L. and threonine, except when glycine was the nonprotein nitrogen. (Abstr. No. 204).Fed­ Adkins, J.S. 1968. Interrelationship of die­ source of supplementary nitrogen. eration Proc. 27: 252. tary glycine, methionine and vitamin E in Daniel, V.A., Desai, B.M.L., Venkat Rao, S., the rat. J. Nutr. 94: 129. In the present study the higher growth Swaminathan, M. and Parpia, H.A.B. 1968. Young, V.R. and Zamora, J. 1968. Effects of rates of control rats fed the 15% milk Dilution of cow’s milk and egg proteins with altering the proportions of essential to non- glutamic acid and the effect on the protein essential amino acids on growth and plasma protein diet, compared with the controls efficiency ratio. J. Food Sci. 33: 432. amino acid levels in the rat. J. Nutr. 96: 21. used by Swendseid et al. (1962), may ex­ Hardin, J.O. and Hove, E.L. 1951. Prevention Ms. received 11/22/68; revised 1/26/70; accept­ of DL-methionine toxicity in rats by vita­ ed 1/28/70. plain the failure of tryptophan and threo­ mins E, Bj 2 , folacin, glycine and arginine. nine supplementation of diets containing Proc. Soc.Exptl. Biol. Med. 78: 728. This work was supported in part by Con­ the nitrogen to restore growth to the Holt, L.E. and Snyderman, S. 1965. Protein tract No. AID/csd—1441 with the U. S. Agency and amino acid requirement of infants and for International Development and by Air maximum rate achieved with the 15% children. Nutr. Abstr. Rev. 35: 1. Force Contract No. AF 33 (615)—2924. This m ilk protein diet. Huang, P.C., Young, V.R., Cholakos, B. and manuscript is Contribution No. 1398 from the Scrimshaw, N.S. 1966. Determination of the Department of Nutrition and Food Science, A 10% substitution of milk protein N, minimum dietary essential amino acid-to- Massachusetts Institute of Technology. E. LAAKKONEN,3 G. H. WELLINGTON and J. 1V . S H E R B O N Departments of Animal Science and Food Science

______Cornell University, Ithaca, N ew York 14850

LOW-TEMPERATURE, LONG-TIME HEATING OF BOVINE MUSCLE 1. Changes in Tenderness, Water-Binding Capacity, pH and Amount of Water-Soluble Components

SUMMARY—Relationships between the tenderness o f very slowly cooked meat and its water­ either Prime or Choice and had Moderately holding capacity, pH and the amount of water-soluble components were studied. Beef muscle Abundant or Slightly Abundant marbling. After portions from the longissimus, semitendinosus and rectus femoris muscles were heated under fixed aging 5-7 days at 0 ± 1°C, the muscles were temperature programs with samples from each analyzed at 1-hr intervals between the 3rd and the separated and trimmed of fat-and epimysium. 10th hr of heating. Weight losses after holding at the final temperature to the 24th hr were They were cut into 2.5-cm-thick slices weighing 100-130 g and sealed under vacuum in Cryo- determined. During the first 4 hr o f heating there were only minor changes in tenderness. The vac bags. They were then returned to 0°C until major decrease in shear values occurred between the 4th and 6th hr, when the meat was warming each muscle had been aged for 2 wk from from 50—60°C. The weight losses increased rather linearly to the 7th hr and remained constant for slaughter. the longissimus and the semitendinosus muscle. The pH values gradually increased during heating. Heating rates During the first 3 hr of heating, up to 45°C, there was only a slight decrease in the amount o f the water-soluble fraction. During the following 3 hr, from 45 to 58°C, the water-soluble fraction A preliminary experiment showed that the decreased more rapidly and the decrease was only slight during a following 4-hr holding period. center temperature of a 15.7-kg steamship A fter 6 hr o f heating to 60°C there were still uncoagulated water-soluble proteins. These studies round roast increased about 0.1°C/min when indicate that the final temperature o f meat has great influence on tenderness and weight loss. The cooked at 121°C in an institutional gas-heated oven. Therefore, the experimental samples were significance o f the shrinkage o f collagen in long-time, low-temperature cooking is considered. cooked in the sealed plastic bags by submerging in a 30°C water bath and increasing the bath INTRODUCTION et al. (1966) used a much higher brown­ temperature 0.1°C/min until the bath reached ing temperature, 218°C, and the internal THE TENDERIZING effect of long-time, 60°C. This temperature was maintained until temperature of the roasts never reached the total cooking time was 10 hr. From the 3rd low-temperature cooking has been widely the collagen melting point and they were to the 10th hr samples were removed from the studied since Cover (1937) found that more raw and less tender than the control water bath every hour. In addition, samples well-done roasts were more tender when samples. Funk et al. (1966) also used a were removed from the water bath at 36 and cooked at 124 than at 225 °C. Bramblett higher browning temperature than Dymit, 44°C and heated in 2 other baths at 37 and et al. (1959; 1964) found that muscles 45°C until total cooking periods of 6 and 10 hr and also reported a decreased quality cooked at 68° C were more tender and were reached. One control sample from each using the delayed service method. It had slightly better appearance and flavor muscle was at first tempered 1 hr at 30°C in a would appear that the more severe than those cooked at 93°C, although water bath, then heated up to 80°C in 1 hr, and browning used by Gaines et al. and Funk kept at this temperature for 1 hr. This is similar the y were less ju ic y . These w orkers con­ et al. impeded subsequent heat penetra­ to the schedule used by Marsh et al. (1966). cluded that the length of holding between tion at the lower temperatures, probably The approximate meat temperature-time curves 57 and 60°C was closely related to due to a severely coagulated surface layer. and the sampling points are shown in Figure 1. increased tenderness. Marshall et al. Korschgen et al. (1963) roasted large cuts Meat evaluation (1960) found that during low-tempera­ of beef to an internal temperature of ture roasting, evaporation losses increased 43°C in a 149°C oven. The meat was Tenderness was measured as shear value in pounds using a Wamer-Bratzler apparatus. Test with the internal temperature of the then cooled and sliced. They then broiled meat, drip losses increased w ith increasing cores were 2.5 cm in diameter, and an attempt slices for 3 min on each side on a 204°C was made to keep the axis of the core parallel oven temperature and with increasing grill. This procedure resulted in good internal temperature and that the total tenderness even of the visible connective losses were greater in the lowest-temper- tissue. ature oven. Schoman (1960) found that juice losses and power consumption could It is evident from these studies that be reduced by roasting at 121 instead of the rate of heating can affect various 149°C in a moisture-tight forced-convec­ properties of the meat associated with tion oven. These results indicate that tenderness. Yet the extent and nature of low-temperature cooking affects the the changes are not clear, due primarily water-binding capacity of the meat. to the use of different heating rates by Dymit (1961) introduced the “ delayed different workers. In the present study, service” method of meat cookery. He well-defined conditions of low-tempera­ browned beef ribs for 1.5 hr at 178°C, ture cooking were followed. At regular then held them for 3—48 hr at 60°C. intervals, the water-holding capacity, pH, Flavor, tenderness and juiciness improved amount of water-soluble components and during the first 24 hr and remained tenderness were measured. The objective constant during the next 24 hr at 60 °C. was to detect relationships between ten­ Only 15% shrinkage was observed. Gaines derness and these parameters. EXPERIMENTAL Raw materials Fig. 1.—Approximate heating curves for mus­ cles cooked at different rates. The dots repre­ 3 Present address: c/o Institute of Food Longissimus, rectus femoris and semitendi­ Chemistry and Technology, University of Hel­ nosus muscles from both sides of 3 Hereford sent times and temperatures when samples were sinki, Helsinki 71, Finland. steer carcasses were used. The carcasses graded withdrawn for analysis.

Volume 35 (19701-JOURNAL OF FOOD SCIENCE - 175 176 -JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

WATER WATER

el- o 46 5 520 570 60.0 SURFACE 6 “ 45 0 50.5 56 0 585 CENTER . , _j L ,. _i I ■ . ■ 1 • ■ -*■ 1 ' ■ _ o 2 4 6 8 10 HEATING TIME (HR) Fig. 3.—Shear value readings, weight loss, pH Fig. 2.—Shear value readings, weight loss, pH and amount o f freeze-dried water-soluble frac­ Fig. 4.—Shear value readings, weight loss and and amount o f freeze-dried water-soluble frac­ tion in semintendinosus muscle heated at 0.1 °C / pH in rectus femoris muscle heated at 0.1 °C / tion in longissimus muscle heated at 0.1° C/min min to 60°C and held for 10 hr total heating min to 60°C and held for 10 hr total heating to 60° C and held for 10 hr total heating time. tim e. tim e.

to the muscle fiber. 3 such cores were taken at During the first 4 hr of heating there and 4.4%/hr for the longissimus, semi­ each sampling point and each core was first were only minor changes in the tender­ tendinosus and rectus femoris muscles, re­ sheared into halves and these were sheared ness. B o th the longissim us and rectus spectively. again, thus giving 9 shear values for each muscle femoris appeared to increase slightly in sample. The pH values gradually increased dur­ shear value between the 3rd and 4th hr, Water accounts for the major portion of the ing heating, although there were fluctua­ although the differences were within the weight loss during meat cooking, especially tions in the rate of rise. These fluctua­ standard deviations of the observations. with low-fat cuts. Therefore, the weight loss tions may have been due to pH variations Both muscles were significantly more reflects the water-binding ability of the meat between different parts of the same mus­ proteins. The samples were weighed before and tender (lower shear values) than the cle (Lawrie, 1966) as well as to the use of after heating and the weight loss reported as semitendinosus. The major decrease in samples from different animals. The larg­ percent of original weight. shear values occurred between the 4th est fluctuations and slightly higher aver­ The water-soluble material, largely protein, and 6th hr, when the meat was warming age p H o f the unheated muscle were was extracted according to the scheme sug­ from 50 to 60°C. The rectus femoris found in the rectus femoris. Since there gested by Maier et al. (1966). 15 g of meat were continued to show decreasing shear values cut into strips with scissors and 60 ml distilled were also fluctuations in tenderness and to the 8th hr, but the longissimus did not water added. The mixture was homogenized for weight loss of this muscle during heating, show further changes after the 6th hr. 30 sec in a Waring Blendor. The pH of the all these factors may reflect some special Shear values of the semitendinosus mus­ homogenate was determined. The homogenate properties of its proteins. was then centrifuged 30 min at 2,000 X g at 0 ± cle remained constant between the 6th 1°C. The supernatant was fdtered and freeze and 8th hr, then decreased further during During the first 3 hr of heating, up to the remainder of the heating schedule. dried. The dry matter was weighed and is 45°C internal temperature, there was reported herein as milligrams per gram un­ The final shear values of the semitendi­ only a slight decrease in the amount of cooked meat nosus muscle were close to values ob­ the water-soluble fraction. During the tained for the 2 other muscles after 3 hr RESULTS & DISCUSSION following 3 hr, 45.0-58.5°C internal tem­ of heating. perature, the water-soluble fraction de­ Relationships between factors studied Weight loss upon cooking is a usable, if creased more rapidly. The decrease was Values for tenderness, weight loss, pH imprecise, measure of the water-holding only slight during the following 4-hr and water-soluble fraction observed capacity of meat (Hamm, 1960). Each holding period. The final values were throughout cooking the longissimus are muscle showed a slightly different pattern higher than for the control samples, shown in Figure 2. Similar values for the in weight loss during cooking. The weight showing that there were still uncoagulated semitendinosus muscle and rectus femoris losses increased rather linearly to the 7th water-soluble proteins in the meat. The are shown in Figures 3 and 4, respectively. hr, then remained constant for the lon­ water-soluble fraction includes both pro­ Values shown are means of measurements gissimus and the semitendinosus muscles. teins and salts. It is estimated from data made on 3 different samples of each After 6 hr, when the temperature of the presented by Lawrie (1966) that about muscle, except for the amount of water- meat had just reached 60°C, the weight 2/3 of the water-soluble material are soluble material, which is the mean of 2 loss of the rectus femoris held constant sarcoplasmic proteins. As these proteins determinations. These values appear in for 1 hr, then increased more gradually to coagulate, they can no longer be ex­ Figure 4. The water-soluble material was the end of cooking. Between the 4th and tracted. It follows that changes due to not determined for rectus femoris. 6th hr, rates of weight loss were 5.3, 7.2 heating represent a measure of protein HEATING OF BOVINE MUSCLE. 7 . - 1 7 7

Accordingly, reduced coagulation in this obtain the benefits of holding meat no study resulted in juicier meat. During the hotter than 60°C. subsequent holding period, further co­ agulation must have occurred, yet the amount of water-soluble material did no: REFERENCES decrease nor did the meat express more Bendall, J.R. and Wismer-Pedersen, J. 1962. Some properties of the fibrillar proteins of juice. This does not preclude a shift in the normal and watery porK muscle. J. F o o d location of the juice within the micro- Sei. 27, 144. Bramblett, V.D., Hostetler, R.L., Vail, G.E. and structure of the meat. The retention o: Draudt, H.N. 1959. Qualities of beef as the meat juice, together with the shrink­ affected by cooking at very low tempera­ age o f the collagen, m ay exp la in the tures for long periods of time. F o o d T e c h - n o l. 13, 707. increased tenderness of meat cooked for a Bramblett, V.D. and Vail, G.E. 1964. Further long time at low temperatures. studies on the qualities of beef as affected by cooking at very low temperatures for Figure 5 shows the effect of heating long periods. Food Technol. 18, 245. temperature on the total weight losses for Cover, S. 1937. The effect of temperature and time of cooking on the tenderness of roasts. all muscles included in the study. All Texas Agr. Expt. Sta., Bull. 542. samples were heated at 0.1°C/min to the Dymit, J.M. 1961. Cook now, serve later. Institutions 49 (11), 174. holding temperatures of 37, 45 and 60°C. Fujimaki, M. and Deatherage, F.E. 1964. Chro­ The average of the control values is also matographic fractionation of sarcoplasmic HEiTING TIME (HR) shown. As the temperature increased, proteins of beef skeletal muscle on ion-ex­ change cellulose. J. Food Sei. 29, 316. weight loss increased. Both the 37 anc Funk, K., Aldrich, P.J. and Irmiter, T.F. 1966. Fig. 5.—Effect of heating temperature on the 45°C samples continued to lose weight Delayed service cookery of loin cuts of beef. average weight loss o f 3 muscles (longissimus, J. Am. Dietet. Assoc. 48, 210. after the holding period began. The 37°C Gaines, M.K., Perry, A.K. and Van Duyne, F.O. semitendinosus and rectus femoris) from 3 samples lost weight at a constant rate 1966. Preparing top round beef roasts. J. Hereford steers. The data represent means and Am. Dietet. Assoc. 48, 204. standard deviations o f 9 measurements, except throughout the entire holding period. The Hamm, R. 1956. Die postmortalen Veränderun­ 45°C samples lost weight more rapidly gen des Fleisches. Fleischwirtschaft 8, 539. when the number o f measurements is indicated Hamm, R. 1960. Biochemistry of meat hydra­ in parentheses. during heating and the early part of the tion. Advan. Food Res. 10, 355. holding period, but little weight was los: Korschgen, B.M., Baldwin, R.E., Mangel, M., Chisholm, J.C., Pudelkewicz, C., Gordon, after a total cooking time of 10 hr. The H., Naumann, H.D. and Braschler, C. 1963. 60° C samples lost weight even more An evaluation of a new cooking method for rapidly during the first 9 hr, but lost little shoulder clod and chuck roll cuts of beef. Missouri Univ. Agr. Expt. Sta., Res. Bull. coagulation. The total water-soluble frac­ during the last hour of cooking. 825. Kronman, M.J. and Winterbottom, R.J. 1960. tion of the unheated longissimus was Marsh (1952a; 1952b) and Hamm Post-mortem changes in the water-soluble 2.91%, considerably lower than the value (1956) concluded that the post-mortem proteins of bovine skeletal muscle during aging and freezing. J. Agr. Food Chem. 8, obtained from Lawrie (1966). One reason increase in expressible water is accompa­ 67. for this may be differences in extraction nied by a tightening of fibrous structure Lawrie, R.A. 1966. “Meat Science.” 1st ed. and shrinkage of the tissue. Marsh (1962) Pergamon Press, Oxford and New York. methods, but another reason is probably Maier, G.E. and Fischer, R.L. 1966. Acrylamide the influence of aging and freeze-drying has reported that shortening of muscle gel disc electrophoretic patterns and extract- of the samples (Fujimaki et al., 1964). fibers during rigor mortis is directly pro­ ability of chicken breast muscle proteins during post-mortem aging. J. Food Sei. 31, K ronm an et al. (1 9 6 0 ) rep orted 2.51 ± portional to temperature, up to 43°C. 482. 0.17% water-soluble protein in aged mus­ The effect of temperature on weight loss Marsh, B.B. 1952a. Observations on rigor mor­ tis in whale muscle. Biochim. Biophys. cle. during cooking may be related to similar A c ta 9, 127. After 6 hr of heating, the samples had changes in meat structure. Marsh, B.B. 1952b. The effects of adenosine triphosphate on the fibre volume of a reached approximately 60°C, and repre­ From these studies, it would seem that muscle homogenate. Biochim. Biophys. sented the conditions in rare meat. At the final temperature of the meat is A c ta 9, 247. this temperature the largest decrease in extremely critical in affecting tenderness Marsh, B.B. 1962. 4th Meat Industry Res. Conf., Hamilton, Meat Industry Res. Inst. shear value had just been completed. and weight loss. If the temperature is N. Z. Inc. Publ. No. 55, p. 32. In Lawrie, Authorities agree that the . degree of sol­ below the temperature at which collagen R.A. 1966. “Meat Science.” Pergamon Press, London. ubility of collagen increases with tem­ shrinks, the major decrease in tenderness Marsh, B.B., Woodhams, P.R. and Leet, N.G. perature and that at about 60°C collagen does not occur. If the temperature is 1966. Studies in meat tenderness. I. Sensory and objective assessments of tenderness. J. A shortens and is converted into collagen higher than the shrinkage temperature of F o o d Sei. 31, 262. B (Lawrie, 1966). There were still unco­ collagen, the more severe coagulation w ill Marshall, N., Wood, L. and Patton, M.B. 1960. Cooking choice grade, top round beef agulated water-soluble proteins, since cause a higher weight loss and more roasts. Effect of internal temperature on 12.22 mg/g water-soluble material was tig h tly packed, less tender tissue w ill be yield and cooking time. J. A m . D ie te t. obtained from the experimental samples formed. If the meat is heated to the A ssoc. 36, 341. Schoman, C.M. 1960. The effect of oven air of longissimus and only 6.45 mg/g was collagen shrinkage temperature, there w ill temperature, circulation and pressure on the obtained from the well-done controls. be less w eight loss, yet the m ajor increase roasting of top rounds of beef and on factors determining its acceptability. Ph.D. The weight loss for this muscle was only in tenderness will have occurred. If the thesis, Rutgers University, New Brunswick, 21.9% as compared with 38.7% for the meat is to be well done, it should be held New Jersey. Ms. received 7/14/69; accepted 10/18/69.______well-done controls. It is reported that the at that temperature rather than heated This research was carried out in part under combination of low pH and high temper­ further. Although not apparent from Contract No. 12-14-100-8903 (73) with the Ag­ ature precipitates sarcoplasmic proteins these data, the literature cited and the ricultural Research Service, U.S. Department of Agriculture, administered by the Eastern Utili­ on to the myofibril, lowering the water­ experiences of the authors indicate that zation Research and Development Division, Phil­ holding capacity (Bendali et ah, 1962). slow heating may be essential if one is to adelphia, Pennsylvania 19118. E. LAAKKONEN , a J. W. SHERBON and G. H. WELLINGTON Departm ents of Food Science and Animal Science Cornell University, Ithaca, N ew York 14850

LOW-TEMPERATURE, LONG-TIME HEATING OF BOVINE MUSCLE 2. Changes in Electrophoretic Patterns

SUMMARY—Polyacrylamide gel electrophoresis was used to follow changes in the nature o f the freeze-dried water-soluble fraction was dis­ water-soluble proteins and juices o f bovine muscle during low-temperature heating. The slowest- solved ir. 0.5 ml distilled water, the solution moving anodic proteins were coagulated first. The myoglobins and myoalbumins were altered saturated with sucrose and 2 drops of brom- significantly only by holding the meat at 60° C. The largest changes in tenderness and amount of phenol blue added. Electrophoresis was done at 6°C for 7 hr. The initial current was set to 50 water-soluble material in the meat occurred at the same temperatures under which the slow-moving mA and increased 5 mA every 15 min until the proteins were denatured. The most heat-sensitive proteins detected were denatured before there voltage reached 250 v. The voltage was then were significant changes in tenderness or water-soluble substance content. kept constant to the end of the run. The gels were stained with Amido Black 10B and de- INTRODUCTION and 60°C. Samples were withdrawn periodical­ stained with 7% acetic acid. They were stored ly during the heating program for the various in the destaining solution until photographed. RANDALL et al. (1967) found 3 ca­ measurements. The freeze-dried water extracts Histograms were constructed such that the thodic bands and 13 anodic bands using were used for the electrophoresis. Details on width o: the bar indicates the widti of the starch gel electrophoresis of the water- muscle preparation, heating program and ex­ band and the height indicates the estimated in­ soluble fraction of bovine skeletal muscle. traction of the water-soluble material were pre­ tensity. The distance migrated relative to the sented in the preceding paper (Laakkoncn et Scopes (1964) has shown that myo- front is indicated by the position of the bar on al., 1969). albumin is the fastest-moving anodic band the abscissa. in sarcoplasmic extracts. Quinn et al. The method for vertical polyacrylamide gel (1964a; 1964b) found at least 3 myo­ electrophoresis presented by Thompson et al. RESULTS (1964) was modified for this study. The gels globin bands, the slowest being the most were 6% in polyacrylamide and 4.5 M in urea. AS EXPECTED, the electrophoretic pat­ predominant. Randall et al. (1967) ob­ The electrophoresis cell was buffered to pH 8.2. terns of the water-soluble material from served 2 pigmented bands, the 6th and Gel formation was catalyzed by addition of raw meat shown at the top of Figure 1 7th from the anodic end, the faster being 0.2% ammonium persulfate. 40 mg of the are similar to those obtained by Randall the more intensely colored. Rowland et al. (1968) found 1 major myoglobin band and several more rapid forms using poly­ acrylamide gel electrophoresis. Most of the sarcoplasmic proteins co­ agulate when bovine muscle reaches START- 40—60°C (Hamm, 1966). Grau et al. TIME FINAL (1963) showed that the proteins migrat­ OF MEAT HEATIN6 TEMP. ing in an electric field with the greatest (HR) (C) 17 16 18 MISaillO • 6 76 64 S C 1 velocity were denatured the most easily, I m~h-n-i—i ri^hr-»-. 1—l-il r.t although the cathodic proteins were more stable (Lee et al., 1966). The juice from 1 m 1—-n-i n rT h i-, ~ n Id. cooked meat shows different electro­ phoretic patterns from the cooked juice 46.8 IT T h ------^ ^ rfTh—i „ n Id. of raw meat, which Lee et al. (1966) sug­ gested was due to an influence of the 82.0 rD —f—l rTh__. r.f m yofibrillar proteins. Quinn et al. (1964a; 1964b) found that the 3 myo­ 87.0 1----- 1------. rnx~h-— rfl r.f globins evident in their electrophoretic patterns were not affected by heating to 60.0 n-i 55 °C for 5 min. r.f The objective of this study was to re­ 60.0 r-r-H --- r .t late changes in the electrophoretic pat­ terns of water-soluble proteins and juices 6 0 j0 n— obtained from bovine muscle during low- r .t temperature cooking to changes observed 60.0 pH-. r .t in the tenderness, water-holding capacity, pH and amount of water-soluble protein. 10 60.0 ru ______n r . t EXPERIMENTAL LONGISSIMUS, rectus femoris and semitendi- sao n r .f nosus muscles from Hereford steers were heated 10 RELATIVE MOBILITY at 0.1 °C/min to holding temperatures of 37, 45 Fig. 1 —Electrophoretograms of freeze-dried water-soluble proteins 3 Present address: c/o Institute of Food from samples heated to 60°C at 0.1°C/min and held for 10 hr total Chemistry and Technology, University of Hel­ heating time. Muscles: l.d. = longissimus, r.f. = rectus femoris. Control: sinki, Helsinki 71, Finland. rectus femoris heated to 80°C.

178 - JOURNAL OF FOOD SCIENCE-Volume 35 (1970) HEATING OF BOVINE MUSCLE. 2 . - 1 7 9

TIME FINAL START — ► + OF MEAT START----- > + HEATING TEMP. TIME OF 17 16 15 14 1312 II 10 9 8 7 6 5 4 3 2 1 HEATING (HR) (HR) (C) [7 16 I5 I4 I3I2II 10 9 8 7 6 5 4 3 2 1 0 o 1 fTTTl-r-H — I r T h i—h i l l-TI r.f o 1 TTT~h-r>-f — i H h i—)—i i— l-TI r.f. 0 0 1 11— i i— i ri h - r - u . 3 r r > n r T h - r - i I.d. 6 37 i—1 1_____i_i— i J 4 rT h n — h i—i rT h r.f 6 45 I 5 1 1 L------. rl Th— i n r.f. 6 60 £ ri L i—^ ,

7 1------1 n ______r r r.f. 10 37 _L h------1 _ rl h ,----1

8 1------1------, f~ h r.f. 10 45 1 h — i H h n m

9 i----- 1--- Hi A-- 1 ,-- 1 rl--- 1 r.f. 10 60 ,___ _

IO n — -— , m r.f. 24 37 1----- 1____ ■=,______I~k r—<

2 4 45 i------i ___ m r-t i—i 80C r.f. 0 10 RELATIVE MOBILITY 2 80 j= i______0 I0 RELATIVE MOBILITY Fig. 2.—Electrophoretogram s o f freeze-dried m eat juices Id rip) exuded from samples heated to 60°C at 0.1°C /m in and held for 10 hr total Fig. 3.—Electrophoretogram s o f freeze-dried w ater-soluble proteins heating tim e. M uscles: I.d. = longissim us, r. f. = rectus fem oris. C ontrol: from samples heated to 37, 45 and 60°C at 0 .1°C/m in and 80°C at rectus fem oris heated to 80°C . The pattern fo r w ater-soluble proteins 0.8°C /m in. M uscle: longissim us (r.f. = rectus fem oris show n fo r refer­ from unheated rectus fem oris is included fo r reference. e n c e ) .

et al. (1967). Although 17 anodic bands band 2. At 60°C, this was reversed. Ex­ from the samples. Below 45°C, the are present, the higher pH used in the tracts from the control sample, heated to amount of juice was very small, as re­ present study may have changed some 80° C in 1 hr and held 1 hr, had only a ported in the preceding paper. The most proteins from a net positive to a net nega­ trace of protein-like material and that did heat-sensitive bands, 9 —1 1 , did not ap­ tive charge. The 5 slowest-moving bands not leave the origin. pear in these electrophoretograms. The were difficult to differentiate, but they Figure 2 shows the changes in the pro­ slowest group of proteins was more clear­ included 2 intensively stained zones, teins of the meat juice (drip) exuded ly defined and unchanged in the juice bands 14 and 16. Band 7 had the most intense meat color prior to staining, but band 8 was also colored. Bands 5 and 6 TIME FINAL START— *■ + were more weakly colored. This agrees OF MEAT HEATING TEMP. with the results of Quinn et al. (1964a; (HR) (C) |7 |6 15 14 13 12 II 109 8 7 6 5 4 3 2 1 1964b) and Rowland et al. (1968). In Water fact, it confirms Quinn’s postulation of 4 O Soluble i m i n — ■ i - m i i l . r.f. myoglobin bands. Sharp bands were not always seen, but stained areas are de­ 6 37 i— i------r n h n - , r - i n . r.f. picted in the figures since they disap­ peared with continued heating. 6 4 5 1----- 1------r f l r n - , „ n r Figure 1 shows the changes in the elec- tro p h o re tic patterns during heating. 6 6 0 n ---- 1------, r T h - , ______=OL_ Bands 9—11 were the most heat sensitive and could not be extracted from muscle 10 3 7 1 l~ r ~ h r - H _ n heated for 4 hr (50.5CC in the center and 52.0°C on the surface). Bands 14—17, 10 45 1------1-----,______rl Ith— i— ___ r T h the slowest-moving, started to blur and disappear from extracts of meat cooked 10 6 0 r.f. for 5 hr. The myoglobin bands, 5—8, ap­ _ r _ l_ __ peared to remain unchanged until the 24 37 2_____ — i______Q l holding temperature of 60° C had been reached. Following this, the amounts of 24 4 5 J____ 1------1 ---- r T h - n _ ____O r myoglobins decreased gradually for the IO rest of the holding time. The myoalbu- RELATIVE MOBILITY mins, bands 1—3, were not affected until Fig. 4.—Electrophoretogram s o f the freeze-dried juices (drip) exuded the meat reached 52°C. Band 3 was dif­ from samples heated to 37, 45 and 60°C at 0 .1°C/m in and 80°C at fuse in extracts of meat heated to 57°C, 0.8°C /m in. M uscles: r.f. = rectus fem oris, s = sem itendinosus. The pa t­ then disappeared. In raw and partially tern fo r w ater-soluble proteins from unheated rectus fem oris is included rare meat, band 1 was more intense than fo r reference. 180-JOURNAL OF FOOD SC/ENCE-Volume 35 (1970) than in the extracts. After 7 hr of cook­ ported in our first paper. During the Quinn, J.R., Pearson, A.M. and Brunner, J.R. 1964a. Detection and isolation of multiple ing, the myoglobin bands were less in­ 60° C holding period, shear values and myoglobins from beef muscle. J. Food Sci. tense and less clearly resolved in the juice content of water-soluble material re­ 29: 422. than in the water extracts. The juice ap­ mained essentially constant in spite of the Quinn, J.R. and Pearson, A.M. 1964b. Charac­ terization studies of three myoglobin frac­ peared to have a slightly higher concen­ apparent decrease in extractable myo­ tions from bovine muscle. J. Food Sci. 29: tration of the myoalbumin group. No globins. The weight loss of the meat does 429. not seem to be related to the electropho- Randall, C.J. and MacRae, H.F. 1967. Hydro­ mobile proteins were found in the juice lytic enzymes in bovine skeletal muscle. II. from the control sample. The difference retically observable changes. There did Proteolytic activity of the water-soluble pro­ between the electrophoretograms of meat seem to be fewer extractable proteins in teins separated by starch gel electrophoresis. J. Food Sci. 32: 182. juice and of water-soluble material from samples having the higher weight loss. Rowland, L.P., Dunne, P.B., Penn, A.S. and cooked meat is clear but small, as stated Maher, E. 1968. Myoglobin and muscular by Lee et al. (1966). REFERENCES dystrophy. Arch. Neurol. 18:141.[C.A. 68, (15): 6723d, 1968.] Figures 3 and 4 show the electropho­ Grau, R. and Lee, F.A. 1963. Uber den Einfluss der Temperatur auf das Verhalten der Scopes, R.K. 1964. The influence of post­ retograms of water-soluble proteins and Eiweisstoffe des Rindermuskels. Naturwis- mortem conditions on the solubilities of juices, respectively, of meat samples held senchaften 50: 379. muscle proteins. Biochem. J. 91: 201. Hamm, R. 1966. Heating of muscle systems. In Thompson, M.P., Kiddy, C.A., Johnston, J.O. at different temperatures. Band 11 could “The Physiology and Biochemistry of Mus­ and Weinberg, R.M. 1964. Genetic poly­ not be seen in any of the heated samples, cle as a Food,” eds. Briskey, E.J., Cassens, morphism in caseins of cows’ milk. II. Con­ but band 9 was detectable in the meat R.G. and Trautman, J.C. pp. 363-385. Uni­ firmation of the genetic control of (3 -casein. versity of Wisconsin Press, Madison. J. Dairy Sci. 47: 378. held at 37° C. The myoglobins and myo- Laakkonen, E., Wellington, G.H. and Sherbon, Ms. received 7/14/69; revised 10/31/69; ac­ albumins were little affected by holding J.W. 1970. Low-temperature long-time heat­ cepted 11/11/69. ing of bovine muscle. I. Changes in tender­ the meat at temperatures below 60° C. ness, water-binding capacity, pH and the This research was carried out in part under Loss of solubility of the slowest-mov­ amount of water-soluble components. J. Contract No. 12-14-100-8903 (73) with the Food Sci. 35: 175-177. Agricultural Research Service, U.S. Department ing proteins occurs at the same points in Lee, F. and Grau, R. 1966. Verhalten von of Agriculture, administered by the Eastern the heating schedule as do the largest Rindersarkoplasm beim Erhitzen. (On the Utilization Research and Development Division, influence of temperature on the behavior of Philadelphia, Pennsylvania 19118. The authors decreases in shear values and total soluble proteins of beef.) Fleischwirtschaft gratefully acknowledge Julie Jordan of Cornell amount of water-soluble material re­ 46: 1239. University for assistance with analysis. E. LAAKKONEN,a J. W. SHERBON and G. H. WELLINGTON Departments of Food Science and Animal Science ______Cornell University, Ithaca, New York

LOW-TEMPERATURE, LONG-TIME HEATING OF BOVINE MUSCLE 3. Collagenolytic Activity

S U M M A R Y —N aturally occurring collagenolytic activity was found in the water-soluble fraction c f tween each sample incubated at 37° C and an bovine m uscle. General proteolytic activity determ ined w ith A zocoll indicated that this total average sample incubated at 0°C was used to activity was m uch greater than the collagenase activity specifically determ ined according to the calculate collagenase activity in 7 collagenase/ m ethod o f W iinsch and H eidrich. The collagenase fraction was concentrated b y polyacrylam ide gel mg dried sample from a standard curve. This standard curve was established using 4-phenyl- electrophoresis and the activity o f the enzym e was studied under various pH and tem perature azo-benzyloxycarbonyl-prolyl-leucine in ethyl- conditions. This collagenase could rem ain active in the m eat at cooking tem peratures experienced acetate. The samples incubated at 0°C were in long-tim e, low -tem perature cooking, < 60° C. W ith faster heating and higher internal tem pera­ used as the reference because preliminary ex­ tures, > 7 0 -8 0 ° C, the collagenase observed in this stu dy is inactivated. periments showed that the ethylacetate also tral and alkaline pH from the granule extracted various amounts and types of meat INTRODUCTION colors, depending upon the sample and heat fraction of human granulocytic leuko­ treatment. BY DEFINITION collagenase is an en­ cytes. Woods et al. (1965) demonstrated zyme attacking native collagen at or near far larger collagenolytic activity in the Proteolytic activity physiological pH (Mandl, 1961). How­ bone of rat than in any other of the tis­ Azocoll (Calbiochem, Box 54282, Terminal ever, collagen is easily denatured and then sues studied; namely, kidney, leukocytes, Annex, Los Angeles, California) is an insoluble is susceptible to almost any unspecific brain and liver. but hydrophilic complex of collagen and an azo dye (Schubert et al., 1968). Proteolytic en­ proteolytic enzyme. The objective of this study is to de­ Proteolytic enzymes are present in zymes release the dye from the complex, the termine whether collagenolytic activity rate of release reflecting the proteolytic activity muscle (Whitaker, 1964) and the pH could be measured in meat and to relate of the sample. optima of some of them are within the any such activity to changes occurring 40 mg of the freeze-dried sample were dis­ range of meat. Koszalka et al. (1960) re­ in bovine muscle during low-temperature solved in 5 ml of 0.1 M sodium phosphate buf­ ported a proteolytic enzyme in rat skele­ cooking. fer, pH 7.5, at 0°C. Then 25 mg Azocoll were tal muscle having optimal activity at pH added and the mixture incubated for 15 min. 8.5—9.0 . Nogochi et al. (1966) found EXPERIMENTAL Triplicate preparations were incubated at both that this enzyme is located mainly in the 37 and 0°C. After incubation, each sample was LONG1SSIMUS, rectus femoris and semitendi- myofibrils. filtered and the absorbance of the filtrate at nosus muscles from Hereford steers were heated 580 mju was determined. The difference be­ Davey et al. (1968) describe the aging at 0.1°C/min to holding temperatures of 37, 45 tween the absorbance of each solution incu­ of meat as a loss of the tensile strength of and 60°C. Samples were withdrawn periodical­ bated at 37°C and the absorbance of an average the myofibrillar components of the mus­ ly for analysis. Both the exuded meat juice and solution incubated at 0°C, indicative of the dye cle cell brought about by disintegration the water-soluble fractions were freeze dried released by the enzymic activity of the sample, of the Z bands. This dissolution could and held frozen until used in the present study. was converted to collagenase activity using a lead to a change in the extractability of Details on muscle preparation, heating program standard curve supplied by the substrate manu­ proteins. and extraction of the water-soluble material facturer. The samples incubated at 0°C were were presented in a preceding paper (Laak- Solovyov et al. (1967) isolated muscu­ used as references since myoglobin, \ max = konen et al., 1969). 555 m p, and oxymyoglobin, X max = 575-585 lar cathepsins at pH 5.6. This complex of mgi, are both soluble under these conditions. enzymes possessed a relatively low but Collagenolytic activity Isolation and characterization of collagenolytic clearly pronounced elastase activity. Collagenolytic activity of both the water- activity When studying catheptic action on the soluble fraction and the drip from muscles giv­ pure elastin obtained from fresh warm en various heat treatments was determined The freeze-dried water-soluble fraction ob­ meat, they found more than a threefold quantitatively according to a modification of tained from rectus femoris heated at 0.1°C/min increase in the N-terminal residues of the the method by Wiinsch et al. (1963), using for 4 hr to 52°C was chosen for the attempts to 6 amino acids investigated. Simultane­ 4-phenyl-azo-benzyl ox y carbonyl-L-prolyl- characterize the collagenolytic activity. 40 mg ously, soluble products of elastin break­ L-leucyl-glycyl-L-prolyl-D-arginine (Mann Re­ of the dried material were dissolved in 250 pi of 0.1 M calcium acetate and stored at 0°C for 2 age appeared. There was a 6 8 % increase search Laboratories, Inc., 136 Liberty St., N.Y. 10006) as substrate. hr. The solution was then evenly distributed of N-terminal amino acids. The amount among 20 slots across the width of a poly­ of elastolysis-soluble products had a near­ 40 mg of the freeze-dried sample were dis­ solved in 0.5 ml of 0.01 M calcium acetate. 10 acrylamide gel and subjected to electrophoresis. ly twofold increase during meat aging. mg substrate was dissolved in 0.1 ml methanol Details of the electrophoresis have already been Gross et al. (1965) reported a collage­ and diluted to 10 ml with acetate-veronal buf­ described. The unstained gel was cut trans­ nolytic enzyme in anuran tadpole tissue fer, pH 7.7. 3 identical tubes were preincubated versely into 10-mm-wide segments, 11 alto­ cultures. This enzyme has a pH optimum at 37°C for 15 min, and 3 additional tubes kept gether. The strong myoglobin band reported previously (Laakkonen et al., 1969) appeared in of 6 —8 , loses activity upon heating for 1 0 at 0°C. The substrate solution was similarly min at 50—60°C and is inhibited by preincubated, then 2 ml was mixed with each cf the 6th segment from the origin. Each segment the protein solutions. After holding for 15 min, was soaked in 10 ml of 0.1 M calcium acetate at EDTA and cysteine. Lazarus et al. (1968) 0.5 ml of each reaction mixture was pipetted 0°C. After a 5-day soaking period, 0.5-ml por­ extracted a collagenase operative at neu- into 1 ml of 0.5% citric acid then 5 ml ethyl- tions of the solution were subjected to the acetate added. The mixture was shaken for 15 Wiinsch et al. (1963) collagenase assay. sec. The phases were allowed to separate and 4 Triplicate determinations of the collage­ a Present address: c/o Institute of Food ml of the ethylacetate phase were removed and nolytic activity of material dissolved from the Chemistry and Technology, University of Hel­ dried over 0.3 g of anhydrous sodium sulfate. 7th and 8th strips were made at pH 4, 6, 7, 8, 9 sinki, Helsinki 71, Finland. The difference in absorbance at 320 mg be­ and 10. The acetate-veronal buffer used was

Volume 35 (19701-JOURNAL OF FOOD SCIENCE—181 182—JO URN A L OF FOOD SCIENCE-Volume 35 (1970) adjusted to the desired pH values with 0.1 N HC1 or NaOH. 3 to 5 determinations at pH 7.7 of the col- lagenolytic activity dissolved from the same gel strips were made at each of 3 incubation tem­ peratures: 37,45 and 60°C. RESULTS & DISCUSSION COLLAGENOLYTIC activity, as shown in Figure 1, was detected in all 3 muscles examined and was not completely elimi­ UNHEATED MUSCLE^ DRIP MUSCLE DRIP MUSCLE_DRIP^ nated by heating the meat to 58—60 °C. X. MUSCLE ' 37C 45C ’ UP TO 59C "

The activity detected in the water-soluble Fig. 1 —A m ounts o f collagenase In the freeze- fraction decreased as heating progressed. dried w ater-soluble fraction and in the freeze- The longissimus lost collagenolytic activ­ Fig. 2 —Collagenase a ctivity a t 37°C a t diffe re n t dried exuded m eat juice (drip) from unheated ity more rapidly during the early stages of pH values o f the 0.1 M calcium acetate solution m uscles, and after 6 h r o f hearing at a rate o f heating, whereas the semitendinosus lost obtained by soaking collagenase-active po ly­ 0 .1°C /m in to 37, 45 and 59°C. L= longissim us; acrylam ide gel strips. A c tiv ity is expressed as y activity in the later stages of heating. The Ft = rectus fem oris; S = sem itendinosus. collagenase in 0.5 m l o f 0.1 M calcium acetate exuded meat juices (drip) had higher col­ solution. S m all vertical lines indicate the stand­ lagenolytic activities than the correspond­ ard deviations o f 3 to 5 determ inations. ing water-soluble fraction except for the longissimus heated to 58—60°C. One reason for the higher activities in the drip may be that in the cut surface of increase in extractable calcium after 9 ity, although not in all samples. Heating the muscle there may be a mechanism days of aging at 3.3°C. for 6 hr to a 58.5° C internal temperature similar to that which releases collagenase The relatively high amount of collage­ caused a clear decrease in activity. activity in healing wounds (Ross, 1968, nolytic activity after heating for 6 hr at The pH activity curve obtained for the and Grillo, 1967). Another reason may be 37 and 45°C, and the rapid decrease collagenolytic material concentrated by a greater increase in calcium ion in the upon heating to 60° C may explain Cov­ electrophoresis is shown in Figure 2. The drip than in the meat, as calcium ion is an er’s (1941) early finding that heating for pH optimum seemed to be about pH 7—8 activator of collagenase. Van den Berg et 23 hr in a 90° C oven gave shear values of and there was no activity at pH 4. These al. (1964) found increasing amounts of 1/2 to 1/3 those obtained by heating for results are consistent with those reported calcium in meat juices exuded during 3 hr in 90°C water. The slower penetra­ by Lazarus et al. (1968). The temperature cooking of poultry meat. tion of heat causes less loss of meat juice; activity curve is shown in Figure 3. Of the Houck et al. (1968) reported that therefore, more collagenolytic activity 3 temperatures actually studied, the en­ when whole normal skin excised from will be retained in the meat. zyme was most active at 37° C, but was rats had been incubated in sterile-organ Table 1 shows the results of the deter­ significantly active at 45°. culture at 56° C for 16 hr, this in vitro mination of proteolytic activities, re­ There is probably little collagenolytic burned tissue demonstrated a translation ported as 7 collagenase/mg dried sample. activity in meat during aging when the pH of 30% of the insoluble collagen to a These values are about 10 times those ob­ is around 5.5 and the temperature is close form extractable in 0.15 M NaCl. The tained using the specific collagenase sub­ to 0°C. This may be the reason collage­ normally bound, inactive collagenase is strate of Wünsch et al. (1963). This wide nase has not been reported in meat. converted to a free, active form. This con­ difference confirms the expectation that Results of the present study seem to version is associated with the release of proteolytic enzymes other than collage­ show that in aged meat there is a collage- proteolytic activities. A similar release of nase also are present. The total proteo­ collagenase may occur in muscle tissue lytic activity did not change greatly upon during aging and during low-temperature heating to 37° C for 6 or 10 hr. Heating heating. The calcium ion may play a role, for 4 hr to 50.5°C internal temperature since'Webb et al. (1967) reported a 3X seemed to increase the proteolytic activ­

Table 1 —P roteolytic activity in freeze-dried w ater-soluble extracts o f heated m uscle sam ples as m easured w ith A zocoll and expressed as gam m a quantities o f collagenase.

Collagenase activity present Heating per mg of freeze-dried Muscle Temperature Time water-soluble fraction sample (C) (hr) (gamma) Rectus femoris, right1 0 1.36 Rectus femoris, left1 0 1.38 Fig. 3 —Collagenase activity at pH 7.7 at d iffe r­ Rectus femoris2 37 1.24 6 ent tem oeratures o f the 0.1 M calcium acetate Rectus femoris* 37 10 1.54 solution obtained by soaking collagenase-active Rectus femoris* 52 4 2.46 Longissimus dorsi* 52 4 1.50 polyacrylam ide gel strips. A ctivity is expressed Semitendinosus* 52 4 1.60 a s 7 collagenase in 0.5 m l o f 0.1 M calcium Rectus femoris, drip* 60 6 .75 acetate solution. Sm all vertical lines indicate the standard deviations o f 3 to 5 determ ina­ 1 Steer A. 2 Steer B. t i o n s . HEATING OF BOVINE MUSCLE. 3 . - 1 8 3 nase-like enzyme capable of attacking the Cover, S. 1941. Comparative cooking time and bauende Fermente in Fleisch.” Fleisch- peptide chain L-prolyl-L-leucyl-glycyl-L- tenderness of meat cooked in water and in wirtschaft 48: 468. oven of the same temperature. J. Home Ross, R. 1968. The fibroblast and wound re­ prolyl between leucine and glycine. Econ. 33: 596. pair. Biol. Rev. (British) 43: 51. Neither trypsin, chymotrypsin, carboxy­ Davey, C.L. and Gilbert, K.V. 1968. Studies in Schubert, M. and Hamerman, D. 1968. A meat tenderness. The nature of myofibrillar Primer on Connective Tissue Biochemistry, peptidase A & B nor amidase has this po­ proteins extracted from meat during aging. p. 42. Lea & Febiger, Philadelphia. tential (Wünsch et al., 1963). In most J. Food Sci. 33: 343. Solovyov,---- and Karpova,----- . 1967. Elastin meat-cooking methods, however, the rate Grillo, H.C. 1967. Collagenolytic activity dur­ changes during meat aging and as affected ing mammalian wound repair. Healing Osse­ by some preparations of proteolytic en­ of rise in temperature is fast and when ous Tissue Workshop, Warrenton, Va. 1965. zymes. Rept. 13th European Meat Res. the final internal temperature of 70—80 °C pp. 143-149. [C.A. 1968, 68 (17): 76403h.] Conf. p. 29. Am. Meat Sci. Assoc., Natl. Gross, J. and Nagai, Y. 1965. Specific degrada­ Live Stock and Meat Board, Chicago, Illi­ is reached this enzyme is obviously inac­ tion of the collagen molecule by tadpole nois. tivated. If meat is heated slowly to an in­ collagenolytic enzyme. Proc. Natl. Acad. van den Berg, L. 1964. Physiochemical changes Sci. 54: 1197. in some frozen foods. J. Food Sci. 29: 540. ternal temperature of 60°C, this collage- Houck, J.C., Johnston, J. and Jacob, R.A. Whitaker, J.R. 1964. Post mortem changes af­ nase-like enzyme is probably capable of 1968. Chemical biology of in vivo and in fecting myofibrillar proteins. Proc. 17th vitro wounds. Biochem. Pharmacol., SuppL Ann. Recip. Meat Conf. pp. 153-160. Natl. producing tender meat. pp. 19-26. Live Stock and Meat Board, Chicago, Illi­ Once the triple helix of collagen is un­ Koszalka, T.R. and Miller, L.L. 1960. Proteo­ nois. lytic activity of rat skeletal muscle. 1. Evi­ Woods, J.F. and Nichols, G. Jr. 1965. Distribu­ folded, several other proteolytic enzymes dence for the existence of an enzyme active tion of collagenase in rat tissues. Nature besides collagenase may attack it. Results optima 113 at pH 8.5 to 9.0 J. Biol. Chem. 208: 1325. 235: 665. Wiinsch, E. and Heidrich, H.G. 1963. Zur quan- appearing in Table 1 indicate that proteo­ Laakkonen, E., Wellington, G.H. and Sherbor., titativen Bestimmung der Kollagenase. Z. lytic activity in addition to collagenase J.W. 19 70. Low-temperature, long-time Physiol. Chem. 333; 149. was much greater than the collagenase heating of bovine muscle. 1. Changes in Ms. received 7/14/69; accepted 10/18/69.____ tenderness, water-binding capacity, pH and This research was carried out in part under activity specifically determined according amount of water-soluble components. Z. Contract No. 12-14-100-8903 (73) with the to the method of Wünsch et al. Food Sci. 35: 175-177. Agricultural Research Service, U.S. Department Lazarus, G.S. 1968. Human granulocyte col­ of Agriculture, administered by the Eastern REFERENCES lagenase. Science 159: 1483. Utilization Research and Development Division, Mandl, I. 1961. Collagenases and elastases. Philadelphia, Pennsylvania 19118. The authors Clayson, D.H.F., Beesley, J.A. and Blood, R.M. Advan. Enzymol. 23, 163. E., Nord, F. F. gratefully acknowledge Julie Jordan of Cornell 1966. The chemistry of natural tenderiza- Interscience Publishers, Inc., New York. University for assistance with analysis and Sis­ tion of maturation of meat. J. Sci. Food Nogochi, T. and Kandatso, M. 1966. Agr. Biol. ter A. Cormier and Dr. Ralph Maurer for advice Agr. 17: 220. Chem. (Tokyo) 30: 199. In “Eiweissab- and consultation. J. H. MacNEIL and P. S. DIMICK

D ivision o f Food Science and Industry

Pennsylvania State U niversity, U niversity Park 16802

POULTRY PRODUCT QUALITY. 1. Compositional Changes During Cooking of Turkey Roasts

S U M M A R Y —W hite and dark turkey roasts, averaging 7.1 and 3.1 lb respectively, were m ade w ith were monitored and when any roast reached m eat taken from selected samples o f the various entries in the Pennsylvania Turkey Random 170°F it was quickly removed from the oven, placed on a drip rack for 1 min, then weighed. Sam ple M eat Test. A dditional roasts were m ade from sam ple birds from the U niversity research The meat and skin were separated to measure flock. Roasts w ere w rapped in alum inum fo il, then cooked in a Telkes oven. A ll roasts were cooked losses occurring in each component, to an internal tem perature o f 170°F. There were no sex differences in cooking losses except when breast area but not including the neck skin. skin was exam ined separately. D ifferences in cooking losses were observed in the breast m eat bu t Temperatures were monitored by using a wood­ not in the thigh m eat o f roasts prepared from Bronze and W hite turkeys. Losses were higher for en plug containing three thermocouples which breast m eat than for thigh m eat. There was an indication that size o f bird was no t a significant were inserted into the thickest portion of each fa ctor in determ ining percentage cooking losses fo r breast and thigh roasts. W hen fa t drippings from roast. The roasts with the attached thermo­ cooked skin were analyzed fo r carbonyl content high skin yielding m ales were characterized by the couples were wrapped in aluminum foil and

high concentration o f the 2-enals in relation to the m ethyl ketones. Low skin groups consisted held at 37 to 38°F for 24 hr. White m jat roasts

m ainly o f m ethyl ketones. averaged 7.1 lb while dark meat roasts averaged 3.11b. pounds in lipid systems has been recog­ INTRODUCTION Cooking trials were conducted using a nized as making a major contribution to Telkes cooking oven maintained a325°F. MacNEIL et al. (1968) reported that in flavor. Removal of these compounds After removal from the storage refrigerator and addition to differences in meat yield be­ from the volatile fraction of chicken meat having the thermocouple lead wires connected tween different strains of turkeys, thick­ slurries resulted in a loss of the chickeny to a 24 point automatic recording potentiom­ ness of the epidermal skin may also in­ flavor (Minor et al., 1965). As a result, eter, the roasts were placed on the lower oven fluence the amount of usable raw mate­ this class of compounds was studied ex­ plate. As cooking progressed, the temperatures were monitored and when any roast reached rial. In their material, the portion of tensively (Gadbois et al., 1967). Pippen et 170°F it was quickly removed from the oven, eviscerated carcass represented by skin al. (1965) reported flavor panel differ­ placed on a drip rack for 1 min, then weighed. varied from 5.8 to 12.4 percent in differ­ ences between leg and breast muscles of The meat and skin were separated to measure ent strains. Similarly the spongy mass of broilers. losses occurring in each component. neck skin covering the crop area varied The purpose of this investigation was During the cooking tests there appeared to from 2.4 to 6.3 percent. Hoke et al. to observe and report cooking losses of be some differences in odor among the samples (1967) pointed out that differences in turkey roasts prepared from different being evaluated. Since taste and flavor observa­ cooking losses exist between white and strains of turkeys and to initiate prelimi­ tions had not been included in the experimental dark meat of turkeys. Earlier Marquess et nary studies to examine those chemical design, an additional test was designed to ex­ plore areas of possible flavor difference be­ al. (1963) observed differences in the re­ components involved in off flavor devel­ tween cooked turkeys. The logical place to be­ sponse of white and dark meat to differ­ opment in cooked turkey products. gin such tests would be to examine high- and ent cooking systems. low-skin yield groups. Two such groups of com­ Martinsen et al. (1968) stated that to­ EXPERIMENTAL parable size and age were selected from the tal precooking losses of turkey roast were TURKEY ROASTS were prepared using white University flock. Immediately after slaughtering not significantly affected by the use of and dark meat from Bronze and White turkeys and picking, the skin was removed and placed foil. However, the relative amount of drip in the Pennsylvania Turkey Random Sample in a freezer for several minutes then ground in a and evaporation were changed because, as Meat Test as well as samples selected from the refrigerated grinder. Samples of the ground skin expected, the foil decreased the amount University flock. Breast and thighs were re­ were cooked in a beaker at 325°F for 15 min. of volatile material that was vaporized. moved from the turkey carcass and each com­ RESULTS & DISCUSSION ponent (white and dark meat) was made into a Further documentation of differences in DATA TREATED to analysis of variance cooking responses between white and roast using the skin which normally covers, in one case the two thighs, and the other the (Snedecor, 1956) showed no significant dark meat was presented by Wilkinson et plate. As cooking progressed, the temperatures sex differences in percent cooking losses al. (1967). They reported that the shear force of dark meat decreased constantly and significantly as the temperature of the cooked meat increased. Values for Table 1—Sex and variety com parisons o f cooking losses in turkey roasts. white meat declined about one-third as Cooking loss Cooking loss Cooking loss the temperature increased from 60 to breast meat thigh meat total skin 6 6 °C. N % % % Goertz et al. (1962) reported that Sex comparisons cooking times and cooking losses were Males (mean) 75 27.4 33.6 17.4 similar for Bronze and White turkeys Taut Females (mean) 75 27.7 34.5 20.2 * that there were sex differences in cooking Indication of significance — ___ losses in breast and thigh meat. In addi­ Variety comparisons Bronze (mean) 70 27.1 34.1 17.1 tion, breast meat had a shorter cooking White (mean) 80 28.0 33.9 20.3 time and resulted in smaller cooking * ** Indication of significance — losses than thigh meat. * Significant at the 5% level (Analysis of variance). The significance of carbonyl com­ * * Significant at the 1% level (Analysis of variance).

1 8 4 -JOURNAL OF FOOD SCIENCE- Volume 35 (1970) POULTRY PRODUCT QUALITY. 1.- 185

of either breast or thigh meat (Table 1). Table 2 —M ean cooking losses fo r turkey m eat and skin from 1966 random sam ple test males. Significant differences (P < .05) were Group* Breast Breast Thigh Thigh Raw Breast Raw Thigh found in skin cooking losses between Mean meat skin meat skin roast roast roast roast sexes. When cooking losses from roasts Eviscerated loss loss loss loss weight loss weight loss prepared from Bronze and White turkeys Weight (lb) % % % % (lb) % Ob) % were examined significant differences (P 21.2 31.7 34.1 37.0 34.4 7.22 31.9 3.28 36.9 < .05) were found to exist in breast meat 20.5 30.4 27.8 37.0 33.7 6.83 30.2 3.09 36.6 losses but not in the thigh meat. Skin 22.1 29.8 31.0 37.6 29.4 7.35 29.9 3.18 36.8 19.9 30.0 23.4 36.3 23.0 2 3 4 29.8 2.94 35.3 cooking losses were significantly different 21.9 27.2 28.0 35.5 32.8 731 27.3 3.33 35.2 among roasts prepared from the different 20.2 26.0 24.9 34.9 20.5 632 26.0 3.06 33.7 strains of turkeys. However, since an in­ 21.1 29.5 24.0 35.1 20.6 6.89 29.3 3.13 34.0 21.9 29.9 36.1 35.6 25.4 7.75 30.1 3.22 34.9 teraction was found between sex and va­ 21.1 29.8 24.6 35.6 20.8 730 29.6 3.02 34.4 riety (P < .10) for this variable, its signif­ 20.0 28.4 31.7 34.7 18.6 7.38 28.5 2.87 33.6 20.8 28.7 27.7 36.9 22.8 6.99 28.6 3.11 35.7 icance as a true indicator of cooked yield 18.7 27.0 26.2 36.5 26.4 6.02 26.9 2.82 35.5 differences is questionable. 19.4 27.8 25.5 36.5 21.2 6.72 27.7 2.91 35.2 Cooking losses from samples observed 20.2 26.5 2 6 5 37.3 29.5 7.00 26.5 3.02 36.7 21.7 28.1 28.0 35.3 27.0 7.54 28.1 3.35 34.7 during the 1966 tests (Table 2) supports 19.4 24.6 25.9 36.2 21.6 6.48 24:7 2.89 35.0 the data previously reported (Marquess et 21.2 28.2 30.2 36.6 29.8 6.89 28.2 3.32 36.1 al., 1963) on the difference in cooking Average 20.7 28.44 27.99 36.15 25.74 7.08 28.44 3.09 35.32 losses between light and dark meat. Val­ * Each sample group contains five birds. ues ranging from 24.6 to 31.7 percent were observed in cooking losses for breast meat; whereas, the range for thighs was from 34.7 to 37.6 percent. As expected these differences were also evident when calculated on a roast basis. Both breast Table 3 —C ooking losses1 o f turkey roasts m ade from turkeys taken from the U niversity flock. and thigh skin cooking losses were similar Breast Thigh Breast Thigh Raw Breast Raw Thigh in most cases. Data collected the previous Eviscerated Total meat meat skin skin breast roast thigh roast year (MacNeil, unpublished data) showed weight skin loss loss !oss loss roast loss roast loss breast and thigh cooking losses similar to N Ob) % % % % % Ob) % % % those reported above. Cooking losses of 65 20.5“ 6.4“ 31.0“ 37.3“ 10.7“ 14.7“ 6.58 30.0“ 3.22 35.8“ the skin varied greatly possibly due to 48 15.3b 8.0b 28.9b 34.7b 14.2b 22.3b 4.56 28.l b 2.38 33.6b problems in determining end point tem­ 65 12.8C 10.0° 30.5“ 37.5“ 17.I e 29.0e 3.95 29.6* 2.04 36.7“ peratures. Ave. 16.2 8.1 30.1 36.5 14.0 22.0 5.03 29.2 2.55 35.4 In another test, meat and skin were 1Means having the same letter are not significantly different from each other P < .05. taken from non-commercial strains main­ tained at The Pennsylvania State Univer­ sity. Three strains and weight classifica­ tion groups were represented. Analysis of Variance, Snedecor (1956) and Multiple were differences in the relative concentra­ broth-like odor; the 2,4-dienal had a Range Test, Duncan (1955) showed sig­ tion of the individual monocarbonyl strong, painty, nutmeg or spicy odor. The nificant differences between groups in all classes. High-skin yield males were char­ variability in amounts and concentrations of the yield variables shown (Table 3). acterized by the high concentration of of these classes were sufficiently high to The main point noted in this study was the 2 -enals in relation to the methyl make them detectable by taste (Day et that bird size did not play a role in de­ ketones and alkanals; whereas, the drip­ al., 1963) and as a result strongly impli­ termining cooking losses when breast and pings from the low-skin yield groups cates the differentiation in flavor re­ thigh meat were fabricated into a natural consisted mainly of methyl ketones. sponse between the high and low skin- roast type product. Cooking losses of To evaluate the possible role of the yielding turkeys. roasts containing breast and thigh meat individual monocarbonyl classes with re­ Studies are currently underway to ful­ were not significantly different for the gard to odor, the carbonyls were regen­ ly characterize the monocarbonyl com­ highest and lowest weight groups. erated (Bassette et al., 1960) with the pounds in cooked turkey and chicken fat The fat drippings from the skin were following odor responses (Table 5). The under various processing and storage analyzed for carbonyl content according methyl ketones were described as being conditions in order to improve the utili­ to the methods of Schwartz et al. (1963). oily or minty; the alkanals meaty or tur­ zation of this turkey component in pro­ The results (Table 4) indicate that there key-like; the 2 -enals had a strong oxidized cessed poultry products.

Table 4—Concentration o f different carbonyl com pounds in drip­

pings from high and low skin-yield turkeys. Table 5 —O dors detected in regenerated car-

bony' com pounds isolated in turkey skin drip­ Micromoles per 10 g fat pings from cooked roasts. % Total Mono­ Methyl Sample N Fa: carbonyl carbonyl ketone Alkanal 2-Enal 2, 4-Dienal Monocarbonyl Odor response <5 —heavy 3 57.9 84.2 7.2 0.9 1.7 2.2 2.6 Methyl ketone Oily, minty <5 —light 4 38.7 73.2 9.2 6.2 _a 1.3 1.7 Alkanal Meaty, turkey-like, C8 aldehyde 9 —heavy 3 67.2 57.0 5.8 4.0 - 0.8 03 2-Enal Strong, oxidized, broth-like, C8 9 -lig h t 3 66.1 44.4 6.9 4 3 0.9 03 0.2 and C9 enal 2,4-Dienal Strong, painty, nutmeg, spicy 'Nodetectable level. ^86-JOURNAL OF FOOD SCIENCE-Volume 35 (1970)

REFERENCES Hoke, I.M., McGeary, B.K. and Kleve, M.K. the cooked volatile fraction. J. Food Sci. 1967. Effect of internal and oven tempera­ 30: 686-696. Bassette, R. and Day, E.A. 1960. Regeneration tures on eating quality of light and dark Pippen, E.L., deFremery, D., Lineweaver, H. of carbonyl compounds from 2,4-dinitro- meat turkey roasts. Food Technol. 21: and Hanson, H.L. 1965. Chicken broth phenylhydrazones with sulfuric acid. J. 773-777. Am. Oil Chemists’Soc. 37: 482—483. flavor and pH. Poultry Sci. 44: 816 — 824. Day, E.A., Lillard, D.A. and Montgomery, M.W. MacNeil, J.H. and Buss, E.G. 1968. Skin and Schwartz, D.P., Haller, H.S. and Keeney, M. 1963. Autoxidation of milk lipids. III. Ef­ meat yields of turkeys as influenced by 1963. Direct quantitative isolation of mono­ fect on flavor of the additive interactions of strain. Poultry Sci. 47: 1566—1570. carbonyl compounds from fats and oils. carbonyl compounds at subthreshold con­ Martinsen, C.S. and Carlin, A.F. 1968. Rate of Anal. Chem. 35: 2191-2194. centrations. J. Dairy Sci. 46: 291 — 301. heating during precooking in foil and quali­ Snedecor, G.W. 1956. “Statistical Methods.” Duncan, D.R. 1955. Multiple range and multi­ ty of boneless turkey roasts stored at 0°F. Iowa State College Press, Ames, Iowa. ple F test. Biometrics 11: 1 — 42. Food Technol. 22: 223—226. Wilkinson, R.J. and Dawson, L.E. 1967. Ten­ Gadbois, D.F., Mendelsohn, J.M. and Ronsival- Marquess, C.G., Carlin, A.F. and Augustine, derness and juiciness of turkey roasts li, L.J. 1967. Effects of radiation, heating G.M. 1963. Factors affecting the quality of cooked to different temperatures. Poultry and storage on volatile carbonyl compounds roasted turkey rolls. Food Technol. 17: Sci. 46: 15-19. in clam meats. J. Food Sci. 32: 511—515. 110-115. Ms. received 4/9/69; revised 6/19/69; accepted Goertz, G.E., Hooper, A.S. and Fry, J.L. 1962. Minor, L.J., Pearson, A.H., Dawson, L.E. and 6/27/69.______Edible yield of roasted breasts and thighs of Schweigert, B.S. 1965. Chicken flavor: the Authorized for publication on April 1, 1969, bronze and white turkeys. Poultry Sci. 41: identification of some chemical components as paper No. 3576 in the journal series of the 1295-1303. and the importance of sulfur compounds in Pennsylvania Agricultural Experiment Station.

P. S. DIMICK and J. H. MAC NEIL

D ivision o f Food Science and Industry

The Pennsylvania State U niversity, U niversity Park, Pennsylvania 16802

POULTRY PRODUCT QUALITY. 2. Storage Time-Temperature Effects on Carbonyl Composition of Cooked Turkey and Chicken Skin Fractions

SU M M A R Y—The carbonyl com pounds in cooked turkey and chicken skin fractions after storage progeny produced from a three-way cross were were isolated as their 2,4-dinitrophenylhdrazones. The m onocarbonyl class was separated into selected. All turkeys were April-hatched, reared m ethyl ketones, 2-enals and 2,4-dienals and m easured spectrophotom etrically. The turkey skin under natural day-length conditions, fed a pel­ leted commercial ration and slaughtered from residue fraction contained higher concentrations o f carbonyls than d id the chicken sam ples. The o il 29 to 31 wk of age over a 2-wk period. 100 extract from the skin o f both groups was sim ilar in carbonyl concentration. Low er storage tem ­ ready-to-cook, 9-week-old broilers including perature dram atically low ered the developm ent o f carbonyls. P hospholipid phosphorus determ ina­ both males and females, were obtained from a tions indicated the residue contained high levels o f p o lar lip id ; whereas, negligible am ounts w ere in commercial processing plant. the oil. Thin-layer chrom atography o f the carbonyl classes from the skin residue indicated m ainly The skin was removed following slaughter, C 7 -C 9 2-enals and Cg, Cg 2, 4-dienals in the unsaturated aldehyde fractions. Changes In fa tty acid evisceration, placed on stainless steel trays and com position o f the residue polar lipids during storage suggested linoleic and arachidonic acids as frozen in an air-blast freezer at -30°F. While the probable substrates In autoxidative deterioration. still frozen, the skin was ground twice in a Hobart meat grinder equipped with a 1/8-in. INTRODUCTION aroma was thought due to the sulfur com­ p la te . 1-lb portions of ground skin were then THE PRODUCTION of further processed pounds. Mecchi et al. (1956) demon­ spread in a 1/4-in. layer on 9- by 13-in. trays and cooked in a Telkes oven at 325°F for 20 min. poultry products is steadily increasing. strated the inferior oxidative stability of turkey fat as compared to chicken fat. The cooked skin and oil were scraped into a The economic importance of proper uti­ double thickness of cheese cloth and manually lization of turkey skin for the expanding This stability difference was attributed to the more efficient deposition of the anti­ pressed to extract as much oil as possible. The poultry product area requires an assess­ extracted oil from each portion (approxim ately ment of its quality. MacNeil et al. (1968) oxidant, tocopherol, in the carcass fat of 200 ml) was stored in Erlenmeyer flasks cov­ the chicken. reported that skin yields varied from 6 to ered with aluminum foil. The corresponding skin portions, denoted as residues, were re- 1 2 % by weight of the eviscerated carcass Preliminary studies (MacNeil et al., in different strains of turkeys. This per­ 1969) indicated that the differentiation in ground, placed in Cryovac bags and air-evacu­ centage of raw material is too large to be flavor response between heavy and light ated for storage. These portions were held at temperatures of 0, 40 or 60°F and samples for considered as waste by-product and may skin yield turkeys may be due to the lev­ els of individual monocarbonyl classes analyses taken periodically during 3, 7,13 or 23 be a potential source of raw material for wk of storage. further processed products. isolated from cooked skin of the respec­ The actions of heat and oxygen on tive strains. Chemical analyses lipids during cooking are known to ac­ Our objectives were to compare the A measured amount of extracted oil (ap­ celerate autoxidation which may result in cqmposition and concentration of the proxim ately 10 g) was dissolved in 2 0 0 m l o f carbonyl formation and affect product highly flavorful carbonyls produced dur­ purified hexane. The skin residue (50 g) was quality. Lineweaver et al. (1961) identified ing storage in cooked turkey and chicken slurried in purified hexane in a Waring Blendor numerous carbonyl compounds in cooked skin fractions. and the resulting solution filtered for further chicken broth and adipose tissue. Minor analysis. Fisher Certified hexane was purified EXPERIMENTAL by refluxing w ith sulfuric acid, distilling in glass et al. (1965) attributed the “chickeny” over KOH pellets and rendering carbonyl-free Sample description aroma of cooked chicken muscle to the according to the method of Schwartz et al. volatile carbonyls; whereas, the “meaty” 24 bronze-colored male turkeys representing (1961). The hexane-lipid solutions were then POULTRY PRODUCT QUALITY. 2 - 187

Table 1—Amounts and types of carbonyl compounds isolated from cooked turkey skin residue (R) and extracted oil (O) stored at 40°F.

jiM oies/10 g fat

Total Methyl Storage carbonyls Monocarbonyls ketones 2-Enals 2,4-Dienals

(W k) R O R O R O R O R O

0 1 1 1 .2 6 0 .2 3 9 .6 3 .9 2 4 .3 2 .0 2 .0 0 .3 1 .4 a b 1 5 2 .6 3 .7 2 .4 0 .2 0.1 2 4 0 .0 6 .5 4 .7 0 .2 0 .2 3 5 4 .6 6 .7 2 .9 0 .6 0 .2 5 1 2 7 .6 5 4 .8 2 4 .4 4 .0 1 3 .4 2 .8 3 .2 0 .3 1 .2 0 .2 7 2 9 2 .0 5 6 .7 3 4 .6 4 .6 2 1 .2 3 .3 1 0 .4 0 .2 2 .6 0 .3

a Not detectable. b Not determined.

passed over a Celite column impregnated with to m onitor the eluant from the column aided in 7 wk of storage; whereas, the concentra­ 2,4-dinitrophenylhydrazine, H 3 PO 4 a n d H 2 0 determining the separation. The DNP-hydra­ tion in the oil remained relatively con­ as described by Schwartz et al. (1963) to con­ zone derivatives were evaporated to dryness, stant at below 60 Mmoles/10 g fat. vert the carbonyls to DNP-hydrazones. The pooled according to class, dissolved in known Interestingly, throughout the entire DNP-hydrazones were eluted from the column volume of chloroform and read spectrophoto­ with 200 ml of purified hexane. The total con­ metrically to determine concentration. Classes study no measurable amounts of alkanals centration of the carbonyl derivatives was de­ were established on the basis of the following were found either in the cooked turkey termined by reading the absorbency of this hex­ absorption maxima: methyl ketones, 365; sat­ or in the chicken samples. The 2-enal and ane solution at 340 mu and converting to urated aldehydes, 355; 2-enals, 373; 2,4-dienals, 2,4-dienal classes increased strikingly in Mmoles using E = 22,500. All readings were con­ 3 9 0 . the residue as compared to the extracted ducted on a Beckman Recording Spectropho­ Separation of the individual compounds oil. Observations of the odors regenerated tom eter, Model DBG. within a class was carried out on Kieselguhr G (Bassette et al., 1960) from these fractions The lipids were removed from the DNP- (Brinkmann Instrum ents, Inc.) thin-layer plates are presented in Table 2. Data reported hydrazones by passing the mixture over a impregnated with Carbowax-400 with methyl by Day et al. (1963) and Langler et al. Celite 545-Sea Sorb 43 (Fisher Scientific Co.) cyclohexane as a solvent (Badings et al., 1963). (1964) demonstrated that the flavor c o lu m n ( 1:1, w/w) using 2 0 g of packing mate­ Phospholipid phosphorus was determined threshold values of the unsaturated alde­ rial. 200 ml of hexane rendered the DNP- according to the procedure of Rouser et al. hydrazones free of lipids. The absorbed mono­ (1966) on the total lipids extracted from the oil hydes may be 10 times lower than the sat­ carbonyl derivatives were eluted from this col­ and the skin residue (Folch et al., 1957). Polar urated aldehydes and methyl ketones, umn with 140 ml (3:1, v/v) of chloroform -nitro- lipids were separated on silicic acid absorption impheating their influence on flavor re­ methane. Ketoglyceride derivatives were sepa­ columns (Hirsch et al., 1958) and methyl ester sponse of autoxidized lipids. rated from the monocarbonyl DNP-hydrazones analysis of their corresponding fatty acids was The carbonyl concentrations resulting on an alumina column. 40 g of neutral alumina determ ined by GLC (M etcalfe et al., 1966). from the chicken skin fractions stored at (80-100 mesh), activated by heating 24 hr at the same temperature (Table 3) reflect 150°C, partially deactivated by addition of 6 % RESULTS & DISCUSSION similar differences in the stability be­ (w/w) distilled H20 and allowed to equilibrate CONCENTRATIONS of the carbonyl 16 t o 2 0 hr, was found necessary for good sepa­ tween the residue and oil. However, much ration. The resulting m onocarbonyl DNP-hydra­ classes for the turkey skin oil and residue lower concentrations of the unsaturated zones were evaporated to dryness, dissolved in a samples stored at 40°F are presented aldehydes were evident in the residue known volume of chloroform and the concen­ (Table 1). Lack of data for the residue even after 13 wk of storage. In contrast, tration determined spectrophotometrically at samples in this study was because the ex­ the chicken oil contained levels of car­ 3 6 5 h im using E = 22,500. tracted oil was originally thought to be bonyls comparable with the turkey oil The DNP-hydrazone classes were isolated the most unstable fraction and most sus­ with the exception of the 2,4-dienal from the monocarbonyl fraction by the pro­ ceptible to autoxidation. However, as this class. Only 2 of the 7 test samples indi­ cedure of Schwartz et ai. (1962). A 20-g col­ initial study progressed, the analytical cated measurable amounts of these un­ umn was prepared using Magnesia 2665 (Fisher data demonstrated that the residue, saturated aldehydes. It is evident from Scientific Co.) and Celite 545 (1:1, w/w) and which contained 10% fat, yielded higher the 40°F studies that the chicken residue separation of the DNP-hydrazone classes ac­ complished with the following solvents: 150 ml concentrations of carbonyls expressed in samples are more stable than the turkey of 15% chloroform in hexane, 100 ml of 30% Mmoles/10 g of fat than the oil extract with reference to formation of total car­ chloroform in hexane, 100 ml of 60% chloro­ from the same sample. The concentration bonyls and unsaturated aldehydes. form in hexane and 150 ml of 100% chloroform. of total carbonyls in the turkey residue Figure 1 is a histogram representing the A fraction collector equipped with a UV source increased to 181 Mmoles/10 g fat during the concentrations of the monocarbonyl classes isolated from chicken and turkey skin fractions held at 60°F storage tem­ peratures. These data readily demonstrate the formation of higher concentrations of the unsaturated aldehydes in the residue Table 2—O dor descriptions for regenerated carbonyl com pounds as compared to the oil for both groups. isolated from cooked chicken and turkey skin residue sam ples. Dramatic increases in the total carbonyl C a r b o n y l Odor response concentration were evident in the residue M ethyl ketones M inty, fruity, sweet, oily during 3 wk of storage at this tempera­ 2 -E n a ls Oxidized, poultry-like, painty ture. The turkey residue increased from 2, 4-Dienals Poultry-like, oxidized, putrid, spicy 91.6 to 348.9 Mmoles/10 g fat; whereas, 188-JOURNAL OF FOOD SC! E N CE-Vo!urne 35 (1970)

Table 3—Amounts of carbonyl compounds isolated in cooked chicken skin residue and extracted oil stored at 40°F.

juMoles/10 g fat

T o t a l M e th y l c a r b o n y ls M onocarbonyls k e t o n e s 2 -E n a ls 2,4-Dienals S to r a g e (w k ) R O R O R O R O R O a 0 1 4 5 .4 5 5 .3 4 6 .0 4 .9 2 0 .1 2 .9 0 .2 - - 1 9 0 .5 7 4 .6 3 0 .4 2 1 .3 3 2 .2 1 7 .8 1 .9 0 .3 - 0 .1 3 1 0 1 .1 5 8 .8 2 9 .3 2 .9 1 6 .4 1 .3 1.9 0 .3 1.5 - 5 9 9 .2 5 6 .7 3 3 .7 5 .8 2 1 .2 3 .5 1 .0 0 .5 - - 7 9 7 .3 5 1 .9 2 6 .8 5 .6 1 4 .8 3 .6 1 .7 0 .5 - - 11 1 3 1 .0 5 8 .8 4 4 .9 1 1 .0 2 7 .1 8 .4 2.1 0 .6 0 .6 - 13 1 9 0 .1 6 6 .4 3 6 .2 7 .5 2 1 .0 5 .9 2 .5 0 .4 0 .6 0 .2

a Not detectable.

the chicken increased from 91.1 to 240.2 however, the ability of microorganisms to they remained considerably lower than jiimoles/10 g fat. Similarly, the concentra­ selectively attack and utilize carbonyl those for the residue samples. Figure 2 tion of the 2-enals in the turkey residue compounds (Smith et al., 1968) may be demonstrates the influence of tempera­ was three times greater than in the chick­ the causative factor for the differences in ture on the 2-enal formation for the tur­ en residue, being 22.3 and 6.7 /imoles/10 concentration. key residue samples. Tentative identifica­ g, respectively. The increase in the 2-enals The influence of storage temperature tion of the individual compounds within during storage throughout these studies on carbonyl development can be seen by this 2-enal class is presented in Figure 3. was indicative of an autoxidative break­ comparing these data with the carbonyl An increase in the level of 2-heptenal, down and the accumulation of unsatu­ data obtained for turkey skin fractions 2-octenal and 2-nonenal was evident dur­ rated end products. stored at 0°F (Table 4). During this ex­ ing the 60°F storage study, even though Variability in the levels of methyl tended storage period, concentrations of considerable detail was lost in the thin- ketones in both the oil and the residue the respective carbonyl classes in the oil layer chromatographic plates during re­ cannot be explained in these studies; increased rather consistently; however, production. Similarly, the 2,4-dienal class during storage was composed largely of 2,4-nor.dienal and 2,4-decadienals (data not shown). In addition to the tentative identification of these compounds • on 30 Kieselguhr absorbent, the Rf of these carbonyl derivatives also correspond to 40-1 40-. knowns when chromatographed using Mi­ cro Cel T-38 as the stationary phase and methyl ketone I ! pure hexane as the solvent (Schwartz 2 -e n a l E et al., 1968). 3 0 - 30- 2,4-dienal H The proposed origin of these highly TURKEY CHfCKEN EXTRACT EXTRACT unsaturated aldehydes is the unsaturated fatty acids, namely linoleic and arach- 20- idonic acids (Badings, 1960; Patton et al., 1959). The polar lipids, consisting primar-

0 10 2 0 30 STORAGE TIME - WEEKS WEEKS Fig. 2—Concentration of the 2-enals isolated Fiff. 1—Concentration o f carbonyl classes in cooked turkey and chicken skin fractions stored at from cooked turkey skin residue following stor- 60°F. age at 0. 40 and 60°F. POULTRY PRODUCT QUALITY. 2 . - 1 8 9

Turkey, 2-enals '

Fig. 3 —Thin-layer plates o f 2-enals as their 2,4-dinitrophenylhydrazones isolated from cooked turkey and chicken skin residue stored at 60° F. K —Reference alk-2-enals C$ through C\ 0, 1. 2 and 3 are weeks o f storage.

Table 4—Comparison of amounts of carbonyl compounds isolated in cooked turkey skin residue and extracted oil stored at 0°F.

pM oles/lO g fat

Total Methyl carbonyls Monocarbonyls ketones 2-Enals 2,4-Dienals s t o r a g e (w k ) R 0 R 0 R O R O R O _ a 0 1 3 0 .8 4 7 .6 3 8 .8 2 .4 2 3 .4 0 .6 1 .8 1 .6 - 5 1 2 4 .8 5 3 .0 4 2 .0 2 .4 2 7 .0 0 .7 2 .2 - 0 .6 - 9 2 2 7 .9 7 6 .6 5 5 .1 5 .1 2 9 .8 3 .8 7 .4 0 .3 2 .0 - 17 1 3 2 .1 5 7 .0 4 3 .2 4 .4 2 2 .9 2 .4 2 .5 0 .4 0 .4 0 .2 23 3 5 5 .2 9 4 .0 6 0 .5 8 .9 3 4 .2 6 .5 8 .3 1.3 3 .7 -

Not detectable.

ily of phospholipids and a highly poten­ to the skin residue. This large difference similarities are also reflected in the fatty tial source of these labile fatty acids, were could account for the variations in car­ acid compositional data for the polar determined on the turkey and chicken bonyl development between the two frac­ lipids (Table 6). residue samples (Table 5). tions. The similarity in phospholipid con­ Using stearic acid as a stable compo­ No significant change in the levels of tent between species indicates that the nent of the lipid, the ratios of weight per­ phospholipids was evident during the difference in stability is not due to the centage data for 18:2/18:0 and 20:4/18:0 storage period; however, the data demon­ amounts of substrate available. This sup­ indicate a greater loss of linoleic and strated the lack of phospholipids in the ports the earlier findings of Nutter et al. arachidonic acids in the turkey residue extracted oil of both groups as compared (1943) and Mecchi et al. (1956). These polar lipids as compared to the chicken residue. The apparent instability of the turkey samples could be due to the low levels of the antioxidant, tocopherol, present (Mecchi et al., 1956). However, the inferior stability of the residue sam­ Table 5—Phospholipid content o f cooked turkey and chicken skin fractions. ples for both groups may be attributed to Mg Phosphorus/g fat the possible extraction of this fat-soluble antioxidant into the oil before storage, R e s id u e O il thereby increasing carbonyl development

M e a n S.D n ]M ean S.D. n in the residue. The data illustrated that the degree of keeping quality of turkey T u r k e y 2 ,0 0 5 5 2 3 7 2 .8 0 .8 13 skin oil as separated and stored in this C h ic k e n 2 ,0 1 7 1 95 11 0 .9 0 .5 1 0 study is comparable to chicken skin oil. 190-JOURNAL OF FOOD SC/ENCE-Volume 35 (1970)

REFERENCES Table 6—Fatty acid composition of polar lipids of cooked turkey (T) and chicken (Cf skin residues. Badings, H.T. 1960. Principles of autoxidation processes in lipids with special regard to the Weeks of storage at 60°F development of autoxidation of off-flavors. F a t t y a c id Neth. Milk Dairy J. 14: 215-242. Badings, H.T. and Wassink, J.G. 1963. Sepa­ 0 1 2 3 ration and identification of aliphatic alde­ C hydes and ketones by thin-layer chromatog­ T C T C T C T raphy of the 2,4-dinitrophenylhydrazones. 1 4 :0 2 .2 0 .3 0 .7 0 .4 0 .3 0 .6 0 .6 0 .7 Neth. Milk Dairy J. 17: 132-149. Bassette, R. and Day, E.A. 1960. Regeneration 1 4 :1 3 .2 1 .7 2 .7 1 .8 1 .6 1 .3 1 .4 2 .1 of carbonyl compounds from 2,4-dinitro­ 1 6 :0 2 2 .1 3 2 .5 3 0 .6 2 7 .7 2 5 .9 3 1 .2 2 4 .0 2 1 .3 phenylhydrazones with sulfuric acid. J. Am. 1 6 :1 4 .7 0 .4 1 .0 0 .3 0 .3 0 .4 0 .6 0 .6 Oil Chemists’ Soc. 37: 482-483. Day, E.A., Lillard, D.A. and M ontgomery, M.W. 1 6 :2 5 .0 1 .9 3 .0 1 .2 0 .6 0 .4 0 .7 0 .3 1963. Autoxidation of milk lipids. III. Ef­ 1 8 :0 1 7 .9 2 5 .0 1 9 .3 2 2 .2 2 4 .4 2 2 .9 2 5 .0 2 3 .1 fect on flavor of the additive interactions of carbonyl compounds at subthreshold con­ 1 8 :1 1 3 .7 2 3 .1 1 6 .9 2 0 .6 1 9 .1 3 1 .2 2 0 .3 2 3 .3 centrations. J. Dairy Sci. 46: 291-301. 1 8 :2 1 7 .1 1 1 .9 1 4 .4 1 2 .6 1 8 .4 1 0 .7 1 8 .3 1 3 .5 Folch, J., Lees, M. and Sloane Stanley, G.H. 9 .1 1957. A simple m ethod for the isolation and 2 0 :4 1 4 .1 1 3 .1 1 1 .3 1 3 .3 9 .4 1 1 .4 1 5 .3 purification of total lipids from animal tis­ sue. J. Biol. Chem. 226; 497-509. 1 8 : 2 / 1 8 : 0 .9 6 .4 8 .7 5 .5 7 .7 5 .5 0 .7 3 .5 8 Hirsch, J. and Ahrens, E.H. Jr. 1958. The sepa­ ration of complex lipide m ixtures by the use 2 0 : 4 / 1 8 : 0 .7 9 .5 2 .5 9 .6 0 .3 9 .5 0 .3 6 .6 6 of silicic acid chromatography. J. Biol. Chem. 233: 311-320. Langler, J.E. and Day, E.A. 1964. Development and flavor properties of methyl ketones in milk fat. J. Dairy Sci. 47: 1291-1296. Lineweaver, H. and Pippen, E.L. 1961. Chicken flavor. Proc., “Flavor Chemistry Symposi­ um,” Campbell Soup Company, Camden, New Jersey. MacNeil, J.H. and Buss, E.G. 1968. Skin and meat yields of turkeys as influenced by Nutter, M.K., Lockhart, E.E. and Harris, R.S. drazone derivatives of aliphatic monocar­ strain. Poultry Sci. 47: 1566-1570. 1943. The chemical composition of depot bonyls into classes on magnesia. Anal. MacNeil, J.H. and Dimick, P.S. 1970. Poultry fats in chickens and turkeys. Oil and Soap Chem. 34: 669-671. product quality. I. Compositional changes 20: 231-234. Schwartz, D.P., Shamey, J., Brewington, C.R. during cooking of turkey roasts. J. Food Patton, S„ Barnes, I.J. and Evans, L.E. 1959. and Parks, O.W. 1968. M ethods for the iso­ Sci. 35: 184-186. n-Deca- 2,4-dienal, its origin from linoleate lation and characterization of constituents M ecchi, E.P., Pool, M.F., Behman, G.A., and flavor significance in fats. J. Am. Oil of natural products. X. New and improved Hamachi, M. and Klose, A.A. 1956. The role Chemists’ Soc. 36: 280-283. methods for the analysis of carbonyl 2,- of tocopherol content in the comparative Rouser, G., Siakotos, A.N. and Fleischer, S. 4-dinitrophenylhydrazones and 2,4-dinitro- stability of chicken and turkey fat. Poultry 1966. Quantitative analysis of phospholipids phenylosazones. M icrochem. J. 13: 407-417. Sci. 35: 1238-1246. by thin-layer chromatography and phospho­ Smith, J.L. and Alford, J.A. 1968. Action of Metcalfe, L.D., Schmitz, A.A. and Pelka, J.R. rous analysis of spots. Lipids 1: 85-86. microorganisms on the peroxides and car­ 1966. Rapid preparation of fatty acid esters Schwartz, D.P., Haller, H.S. and Keeney, M. bonyls of rancid fat. J. Food Sci. 33: 93-97. from lipid for gas chrom atographic analysis. 1963. Direct quantitative isolation of m ono­ Ms. received 4/25/69; revised 7/3/69; accepted Anal. Chem. 38: 514-515. carbonyl from fats and oils. Anal. Chem. 35: 7 / 1 1 / 6 9 .______Minor, L.J., Pearson, A.M., Dawson, L.E. and 2 1 9 1 - 2 1 9 4 . Schweigert, B.S. 1965. Chicken flavor: the Schwartz, D.P. and Parks, O.W. 1961. Prepara­ Paper No. 3555 in the Journal Series of The identification of some chemical components tion of carbonyl free solvents. Anal. Chem. Pennsylvania Agricultural Experim ent Station. and importance of sulfur compounds in the 33: 1396-1398. The authors thank Mrs. Helen M. Walker cooked volatile fraction. J. Food Sci. 30: Schwartz, D.P., Parks, O.W. and Keeney, M. and Dennis H. Flanigan for technical assistance 6 8 6 - 6 9 6 . 1962. Separation of 2,4-dinitrophenylhy- with this study. J. H. M a cN E IL a n d P. S. D IM IC K Division o f Food Science and Industry The Pennsylvania State University, University Park 16802

POULTRY PRODUCT QUALITY. 3. Organoleptic Evaluation of Cooked Chicken

and Turkey Skin Fractions as A ffected by Storage Tim e and Tem perature

SUMMARY—Cooked turkey and chicken skin residue and separated drippings or oil were stored at Both control and treated samples were heated various temperatures then presented to a trained taste panel for flavor evaluation. Panel members to 105°F in a water bath before being evalu­ were able to discriminate between a control (unstoredl and a sample o f residue after 3 wk of ated by panel members. storage at 40°F. They were not able to differentiate between control and treatment (stored) oil TBA values were determined according to even after 7 wk o f storage. When chicken skin residue and oil were evaluated after storage at 40°F the m ethod of Tarladgis et al. (1960) as a mea­ sure of the degree of oxidative rancidity. the panel members could detect differences between the residue samples at 3 wk, but unlike the To examine the long-term effects of storage turkey oil stored at the same temperature they indicated discriminatory ability after 1 wk of on turkey residue and oil a study was designed storage. When both cooked chicken and turkey skin fractions were presented to the panel at the to observe flavor changes in these m aterials over same time without a reference control (unstored) they were able to identify differences but could 35 wk at 0°F storage. not indicate a dear preference for either one. Panel selection Bradley (1964) suggested the use of a The panel consisted of 12 women, most of INTRODUCTION modified triangle test in which the taste whom had previously participated in organo­ THE STEADILY INCREASING number panel member selected the odd sample leptic determ inations with beef products. Train­ and volume of further processed poultry from three samples, one of which is a ing sessions were conducted using various cooked skin residue and oil samples that might products offered to consumers emphasize variant or odd, and two are standard or be similar to some of those anticipated in the the expanded use of these products. alike. In addition the panel members various experiments. All of the panel members Many of these newer products contain scored the degree of difference between were faculty wives from various departments poultry lipids. The carbonyl compounds the paired samples and the odd sample. within the College of Agriculture, selected be­ in lipid materials, as suggested by Minor This procedure enabled Bradley to utilize cause of their previous experience with organo­ et al. (1965), contribute to the flavor of information on the number of correct leptic tests or because of their interest and will­ foods. The economic importance of using selections in the triangle tests as well as ingness to participate. These women were from large amounts of skin from turkeys was the information contained in the degree 24 to 59 years of age, with 2 identified as s m o k e r s . cited by MacNeil et al. (1968) when they of difference scores. Brant et al. (1967), noted that the skin can amount to as Evaluations were made on both cooked skin using a dual-standard panel comparison, residue and the resulting drippings or oil using a much as 12% of the dressed carcass reported that panel members detected modified triangle and hedonic test. Members weight. individual bird differences within either were seated in separate cubicles numbered 1 While initial flavor quality is impor­ control or treatment samples. Stier et al. through 12. 6, 25-w red lights provided over­ tant, its stability (or shelf life) must be (1967) pointed out that the same concen­ head illumination. Panel members were en­ maintained at an adequate level during tration of flavor enhancers will not nec­ couraged to sit at a different position each storage and while in the consumer’s essarily result in the same flavor response week. Samples were prepared in an adjoining hands. Martinsen et al. (1968) could not when tested in different food products. room and carried to them. detect any increase in oxidative rancidity, Our purpose was to study the effects The triangle tests consisted of three parts: as measured by TBA values, for cooked of storage times and temperatures on the 1) Is there a difference? 2) If there is a differ­ ence, complete the triangle test in the usual turkey meat during storage. A trained organoleptic responses of a trained taste manner and 3) indicate a preference among the taste panel of 6 indicated a significant panel testing cooked turkey and chicken samples tested. This modified triangle test was decrease in flavor for these samples over a oil and residue. similar to that outlined by Bradley (1964) ex­ 7-month period. Cash et al. (1968) found cept that a preference determination was re­ that rancidity was significantly affected EXPERIMENTAL quired rather than a degree of difference. In by precooking. They reported that TBA TURKEY SKINS were obtained from 24 subsequent statistical analyses of preference values for fresh frozen controls were al­ bronze-colored male progeny produced from a data only the preferences associated with cor­ ways lower than for precooked stored three-way cross and processed in the manner rect triangle judgm ents were used. The hypoth­ samples. Hoke et al. (1968) showed a sig­ described by Dimick et al. (1970). 100 broilers esis used in the Chi-square analysis of the tri­ nificant continuous decrease in the flavor of both sexes, originating from one flock, were angle test was that if there were no differences obtained from a commercial dressing plant and between control and treatment or stored sam­ of cooked dark turkey meat. Pickett et al. brought to the University poultry products lab­ ples the expected probabilities would be these: (1968) presented evidence that oxidative oratory, where the skin was removed in the 1/3 would indicate no difference; 1/3 would rancidity is inhibited by implantation of manner previously cited (Dimick et al., 1970). make a correct judgment and 1/3 would select turkey with vitamin E. Turkeys treated For taste panel evaluation the extracted oil the incorrect sample pair. The expected proba­ with vitamin E and stored 2 months (turkey or chicken) and the corresponding res­ bilities used with the preference data were 1/2 showed less change in TBA values than idue were removed from the appropriate stor­ for a preference of treatm ent and 1/2 for pref­ did stored turkeys receiving no vitamin E. age cooler (0, 40 or 60°F) and allowed to reach erence of control, assuming no difference be­ Cook et al. (1949) detected rancidity in room temperature. At the same time a pound tween samples. Hedonic values were analyzed by the analysis of variance. flesh from birds fed a tocopherol-free of frozen raw control skin was removed from the freezer, thawed completely, cooked at Consistency scores were computed for all diet, tocopherol being the precursor of 325°F for 20 min, then the oil and residue col­ panel members. They were considered able to vitamin E. The apparent instability of lected in the same manner as the treated skins. measure the panel members’ ability to be con­ turkey fat, as suggested by Mecchi et al. In the long storage experiment lasting for 35 sistent in their judgm ent in the triangle test and (1956), possibly is due to its low level of wk the skin from a freshly slaughtered turkey the hedonic test, since identical samples were the antioxidant tocopherol. was used instead of the raw frozen control. u s e d .

Volume 35 (19701-JOURNAL OF FOOD SCIENCE -1 9 1 "[22—JOURNAL OF FOOD SCIENCE-Volume 35 11970)

RESULTS & DISCUSSION Table 1—Taste panel evaluation of cooked turkey residue and oil stored at 40°F for 7 wk. RESULTS OF PANEL evaluation of Preference Mean hedonic scores3 cooked turkey residue and oil stored at Storage for control ------■ 40°F for 7 wk are presented in Table 1. time Triangle test (unstored) Residue Oil The Chi-square analysis of the triangle (wk) Residue Oil Residue Oil Control Stored Control Stored tests showed that the panel members 1 NS NS NS NS 3.2 5.1 3.1 3.6 were not able to discriminate between 2 NS NS + NS 2.0 4.9 2.1 3.2 control and treated oil extract samples 3 * NS + NS 2.7 4.8 2.5 3.6 after storage for 7 wk. After 3 wk they 5 * NS + NS 2.8 4.1 2.4 3.6 ** could discriminate between control and 7 NS + NS 2.6 5.5 3.1 .3.9 treated residue. When asked to indicate a * Significant discrimination between control and treatment 5% level. preference for either the control or ** Significant discrimination between treatment and control 1% level. treated samples, after 1 wk of storage, the NS- No significant discrimination between treatment and control. + - Preference for control samples significant at 5% level. panel members had no difficulty m pick­ 3 -Hedonic scores from 1, Extremely Good, to 7, Extremely Poor. ing out the control residue sample as the one they preferred. This same discrimi­ natory ability was not evident when oil samples were tested. Hedonic values for Table 2—Analysis o f variance o f hedonic scores of cooked turkey all of the samples in this test show a dis­ residue and oil stored at 40°F for 7 wk. tinct separation of scores between the control and treated residue samples but Residue Oil Source of variation df Mean square Mean square only very slight differences between con­ trol and treated oil samples. That there Treatment (stored vs. unstored) 1 126.58** 2584** are similar differences between control Storage (wk) 4 3.19* 2.49* and treated hedonic scores for each of the Treatment x storage 4 2.15 0.45 storage test periods, and panel members Panel members 10 4.39** 7.22** Treatment x panel members 10 5.60** 1.62* did not indicate this difference until the Storage x panel members 40 1.51 1.62** 3rd wk, cannot be explained fully at this Error 1.17 0.68 time. We must mention that this was the * F value significant at 5% level. first of 4 experiments and although the ** F value significant at 1% level. panel had been subjected to training ses­ sions some of them could still have been in the learning process for these particular products. The analysis of variance (Table 2) of the hedonic scores indicates significant differences between treated and control Table 3—Taste panel evaluation o f cooked chicken residue and oil stored at 40°F for 13 wk. (stored and unstored) samples, storage Mean hedonic scores3 times, panel members, judgments, as well Storage Preference time Triangle test as an interaction between treatment and for control Residue Oil panel members in both residue and oil (wk) Residue Oil Residue Oil Control Stored Control Stored samples. A significant interaction be­ 1 NS ** ++ + 2.8 5.3 2.9 5.4 tween storage and panel member existed 3 ** * ++ NS 2.0 5.0 2.7 3.7 in the oil samples. If the significant in­ 5 ** NS ++ NS 2.0 3.9 4.2 4.8 teractions were used as the error term to 7 * NS + NS 2.7 4.4 2.9 3.6 10 ** NS ++ NS 2.1 6.1 2.9 3.7 test significance, the treatment effects ** would still be significant, but it would 11 NS ++* NS 3.7 6.0 3.7 3.7 13 ** ** ++ NS 2.6 5.8 4.5 3.7 invalidate the significance of storage dif­ i ferences for both residue and oil. The * Significant discrimination between treatment and control 5% level. ** Significant discrimination between treatment and control I% level. significant effect of panel member dif­ + Preference for control samples significant at 5% level. ferences with regard to residue samples ++ Preference for control samples significant at 1% level. also would be cancelled. NS No significant discrimination between treatment and control. 3 Hedonic scores from 1, Extremely Good to 7, Extremely Poor. Cooked chicken residue and oil were stored at 40°F for 13 wk with taste panel evaluations at various intervals. The tri­ angle test analysis by Chi-square (Table 3) showed that after 3 wk the panel could to distinguish it from the control sample, and 2) while similar conditions existed discriminate between control and treated which was not true for the turkey oil with chickens the degree of discrimina­ residue. When oil samples were evaluated samples. Panel members were presented tion in chicken oil samples was greater the panel indicated an ability to detect the chicken samples without any previous than in the turkey oil samples. The differences between control and treat­ exposure to tasting chicken skin and oil marked difference between control and ment for the first 2 storage periods of 1 and only after several sessions of evaluat­ treatment for both chicken and turkey and 3 wk. If a significant level of 10% ing turkey residue and oil. residue is portrayed graphically in Figure were selected, the only storage periods in From the two preceding experiments 1. The oil samples did not follow this which the panel members could not it would appear that 1) there are differ­ same pattern (Figure 2), with control and discriminate between control and treated ences in flavor response between turkey treated samples overlapping in some in- «■ oil were 7 and 11 wk. It would appear that residue and the corresponding oil fraction stances. This further supports results re­ the treated chicken oil contained com­ after several weeks of storage, the skin ported by Dimick et al. (1969) showing pounds which enabled the panel members having a decidedly poorer taste response, the high levels of 2-enals and 2,4-dienal POULTRY PRODUCT QUALITY. 3.- 1 9 3

Fig. 1—Hedonic scores for control and treat­ ment chicken and turkey residue samples. (1, Extremely Good; 7, Extremely Poor.) F'9' 3~TBA values reported as optical density for chicken and turkey samples o f skin residue and oil.

Table 4—Taste panel evaluation o f cooked chicken and turkey residue and oil after storage at 60°F for 3 wk.

Preference for Mean hedonic scores3 Storage chicken time Triangle test over turkey Residue Oil (wk) Residue Oil Residue Oil Chicken Turkey Chicken Turkey 0 ** * NS NS 4.6 43 3.6 3.4 1 ** NS NS NS 4.8 4.2 53 53 2 * ** NS + 4.7 4.4 2.6 5.4 3 ** NS NS NS 4.3 4.4 3.1 3.3 * Significant discrimination between chicken and turkey 5% level. ** Significant discrimination between chicken and turkey 1% level. + Preference for chicken over turkey significant at the 5% level. NS No significant discrimination between turkey and chicken. 3 Hedonic scores from 1, Extremely Good to 7, Extremely Poor.

0 2 4 6 8 10 12 WEEKS

Fig. 2—Hedonic scores for control and treat­ ment chicken and turkey oil samples. 11, Ex­ residue and oil. To speed up developm ent and chicken was betw een 2 and 3. It was tremely Good; 7. Extremely Poor.) of undesirable flavor com pounds the res­ possible that when samples of similar idues and oils were stored at 60°F. taste response were given to the panel The triangle tests analyses (Table 4) they had difficulty in establishing a point showed that the panel members made a of reference. significant discrim ination betw een chick­ Taste panel scores presented in Table en and turkey skin at 0 storage tim e and S show that panel m em bers distinguished carbonyl classes in residue samples as were able to distinguish between these betw een residue samples but not betw een compared to the much lower levels ob­ samples throughout the rem ainder of the oil samples in samples stored at 0°F for served in oil. The striking taste panel dif­ study. The panel did not significantly pre­ 35 wk. The preference data given in the ferences are believed due to the dem on­ fer chicken skin over turkey skin even table showed no significant preference for strated association of these com pounds though the form er was slightly favored. control either in the residue or in the oil, w ith disagreeable flavor. The hedonic scores indicated similarities the highest significance level being 10%. TBA values are presented in Figure 3. betw een chicken and turkey skin residue, Perhaps a higher significance level would The instability of the residue samples is but it was surprising that both 0 storage be justified, since only those preferences apparent and it should be pointed out time hedonic values were higher than associated with correct judgm ents were that the oxidative deterioration took those observed for either turkey or chick­ used. For exam ple, at the storage period place im m ediately upon heating as indi­ en control residue in the two preceding of 9 wk, 6 panel m embers preferred con­ cated by the high TBA values at 0 tim e. experim ents. In every storage period ex­ trol and 2 preferred the treated, but 2 W ith this in mind a third experim ent was cept 2 the hedonic value in these 2 exper­ recorded no differences betw een the sam ­ conducted com paring chicken and turkey im ents for the control residue of turkey ples, which made the preference differ- 194—JO URNA L OF FOOD SCIENCE-Volume 35(1970)

ence not significant. The hedonic scores Table 5—Taste panel evaluation of cooked turkey residue and oil stored at 0°F for 35 wk. m ust be taken into consideration to get a Mean hedonic scores2 realistic appraisal of the flavor response. Storage Preference for time Triangle test control Residue Oil The consistency scores for the panel Residue Residue Stored Control Stored m em bers are presented in Table 6. It is of (wk) Oil Oil Control interest that in m ost cases the panel m em ­ 5 * NS NS NS 3.8 5.2 2.8 3.6 ** * bers m aking the m ost incorrect judgm ents 9 t NS 2.4 4.4 2.2 23 17 ** ** NS + 3.3 3.9 23 4.1 were consistent through all experim ents. 23 ** NS NS NS 4.0 5.0 2.9 4.0 The test in which chicken and turkey 35 ** NS NS NS 1.9 4.8 3.0 4.0 were com pared—in which there was no * Significant discrimination between treatment and control 5% level. control sam ple—gave panel members the ** Significant discrimination between treatment and control 1% level. + Preference for control samples significant at 5% level. m ost difficulty. f Preference for control samples significant at 1% level. NS No significant discrimination between treatment and control. REFERENCES a Hedonic scores from 1, Extremely Good to 7, Extremely Poor. Bradley, R.A. 1964. Applications of the modi­ fied triangle test in sensory difference trials. J. Food Sci. 29: 668-672. Brant, A.W., Sadler, W.W. and Lewis, H. 1967. The influence of disease on sensory proper­ ties of poultry meat. Food Technol. 21: 2 8 9 - 2 9 1 . Cash, D.B. and Carlin, A.F. 1968. Quality of frozen boneless turkey roasts precooked to different internal temperatures. Food Technol. 22: 1477-1480. Table 6—Consistency scores of panel members. Cook, B.B., Morgan, A.F. and Smith, M.B. 1949. Thiamine, riboflavin and niacin con­ Panel tent of turkey tissues as affected by stor­ age and cooking. Food Res. 14: 449-458. member Percent consistency Dimick, P.S. and MacNeil, J.H. 1970. Poultry identification product quality. II. Storage time-tempera­ number Test 1 Test 2 Test 3 Test 4 Avg ture effects on carbonyl composition of cooked turkey and chicken skin fractions. J. 6 86.0 81.7 88.6 95.0 87.8 Food Sci. 35: 186-190. 5 92.0 76.7 85.7 90.0 86.1 Hoke, I.M., McGeary, B.K. and Lakshmanan, F. 12 100.0 80.0 91.4 70.0 85.4 1968. Muscle protein composition and eat­ 9 86.0 78.3 87.1 85.0 84.1 ing quality of fresh and frozen turkeys. J. 1 78.0 85.0 87.1 Food Sci. 33: 566-571. 77.5 81.9 MacNeil, J.H. and Buss, E.G. 1968. Skin and 3 78.0 63.3 91.4 87.5 80.0 meat yields of turkeys as influenced by 2 82.0 80.0 91.4 65.0 79.6 strain. Poultry Sci. 47; 1566-1570. 11 92.0 90.0 76.7 55.0 78.4 Martinsen, C.S. and Carlin, A.F. 1968. Rate of heating during precooking in foil and qual­ 10 80.0 63.3 74.3 80.0 74.4 ity of boneless turkey roasts stored at 0°F. 8 54.0 75.0 95.0 57.5 70.4 Food Technol. 22: 223-226. 4 52.0 66.0 78.3 75.0 67.8 M ecchi, E.P., Pool, M.F., Behman, G.A., 7 62.0 61.7 76.7 65.0 66.4 Hamachi, M. and Klose, A.A. 1956. The role of tocopherol content in comparative stabil­ ity of chicken and turkey fat. Poultry Sci. 35: 1238-1246. Minor, L.J., Pearson, A.M., Dawson, L.E. and Schweigert, B.S. 1965. Chicken flavor: the identification of some chemical components and the importance of sulfur compounds in nucleotides on processed foods. Food Authorized for publication on April 9, the cooked volatile fraction. J. Food Sci. Technol. 21: 1627-1630. 1969, as paper No. 3584 in the journal series of 30: 686-696. Tarladgis, B.G., Watts, B.M. and Younathan, the Pennsylvania Agricultural Experiment Sta­ Pickett, L.D., Miller, B.F. and Moreng, R.E. M.T. 1960. A distillation method for quan­ tio n . 1968. Carcass quality of turkeys as affected titative determ inations of malonaldehyde in by estradiol-17-monopalmitate and vitamin The authors thank Mrs. Ruth Hartswick for rancid foods. J. Am. Oil Chem ists’ Soc. 37: E. Poultry Sci. 47: 1493-1496. her assistance in com puter analysis, D.H. Flan­ 4 4 -4 8 . Stier, F.E., Sawyer, F.M. and Fergenson, P.E. igan for his technical assistance and Dr. John II. Ms. received 4/25/69; revised 7/3/69; accepted 1967. A comparison of methodology, used Ziegler for making the taste panel facilities 7 / 1 1 / 6 9 . in determining the flavor effect of 5 -ribo- a v a ila b le . New from Wisconsin

THE FAT-SOLUBLE VITAMINS Edited by H. F. D eluca and J. W. Suttie

D uring the past decade, there has been a rapid and sizeable accum ulation

of know ledge relating to the fat-soluble vitam ins. T hese proceedings

of a sym posium at the U niversity of W isconsin in 1969, in honor

of H arry Steenbock, com bine the m ost recent research advancem ents

w ith an overview of past findings. T hirty-three studies are included.

448 pages ISBN 0-299-05600-7 $10.00

THE PHYSIOLOGY AND BIOCHEMISTRY OF MUSCLE AS A FOOD, VOL. 2 Edited by E. J. Briskey, R. G. C assens, and B. B. M arsh

T he studies of m uscle as a living m achine and m eat as a dead foodstuff

w ere long regarded as distinct and separate fields of research.

H ow ever, it has now becom e obvious that the tw o are

intim ately related, and are best considered as tw o aspects

of the one science of m uscle biology. T hese proceedings of an

international sym posium , held at the U niversity of W isconsin in 1969,

are a gratifying result of this integrative approach.

Tw enty-nine studies are included.

850 pages; illus.. figs. ISBN 0-299-05680-5 $20.00

R ecently P ublished

THE PHYSIOLOGY AND BIOCHEMISTRY OF MUSCLE AS A FOOD, VOL 1 Edited by E. J. Briskey, R. G. Cassens, and J. C. Trautman

T he proceedings of an international sym posium sponsored by the

U niversity of W isconsin in July, 1965. “T he book am ply dem onstrates

how rew arding such a m arriage betw een fundam ental and applied science

can be ... all praise to the editors for a m ost w orthy enterprise.” — Nutrition Abstracts and Reviews

1966 454 pages, figs. ISBN 0-299-04110-7 $8.50

Wisconsin residents add 4% sales tax

University of Wisconsin Press Box 1379, Madison, Wisconsin 537Q1 1 S n a 251