1 o . _ 1 .. _ 1.1 1. 1 . ...1 c ..1 .. . ...1 . . .1 . u. . 1 . 1. ..1 .1.... . .. .1 . . 1. .12 . . 1 _ . . .. _ . . 1 .. ._ Q . .. . 1 1 . .. 1 1 .Z. ._ o. 1 1 1 .1. . . . . . c . 1.. 1 . . 1 O 1 11 1. o v . t 1 . . . . O _ . . .1 1 1 o 1 n _ I 1. 1 1 s 0 . 1 1 1 . 11.1 1 .1 1. 1 .. 1 1 , 1 . . .1 . 1 . . .n 1 . .. __ 1 on . _ .. . .1 o . 1 o. . .. .O . . . .‘ a 1 1 . 1 V . 1 1 . 1 1 .. . 1 v .1 1 .1 1 _ 1 1 .1 . . 1 1 . 1 . . 1 . 1 n 1 1 o . . 1 O O . a. . .1. o o . o .1 1. _ . . 1 V. D 1 . 1 1 1. a . _ .o’ . . . . an; BY Lama A “I CGNTEMT - SYSTM : . .IB-IC POW i a FRGTEE 1 (m 3? 3.091% mm M A, s A £ch a 3183“ F1 ”1" ”Q-””~w~~m’v wwwu. .— . v..-,,- ‘4‘I“S'Q1~.¢l~ o. ..-..o "‘ .."..°"'.9~_ o f I F O FEY M. 3. Degree of €05 {S {t 1.! 3e: [.1 .1153: “1.111133%“. “51:11.1 . 13 .. FetuW. .1 H11. .3131 H 1.1.1.111...“ . r .A .1 .... c. 11.11 11.11%“? . 11..» .t.d .1’1’. {.11. u. 111.. 1’. s. . . 1 1 1 1. . . . . . 1 . .1. . 1.1 . . . , 1 1 . . . 1 . ”'119H.1.:r.9111hw.1.bh1v1:.1_ 1 1 1 . 1 .. . a) 1 x . 1. . . . . 1. . 1. 1 . 1 1. 1. fiw uwfnuu... 1 . . .1 _ 1. 1 1. 1 V. o _ u C I 1 . 1 1 . . . . . 1 .. .1. . 1 . 1 1 1 11...}... .33.. .1111 , .1. 11 v 11.1.: .11.}.11. .1 .1 113.11.11.1Jw. 2.1.... 11.1 10 1 -111. . . . .1101 1... . h . . . . : 1 . , , 11' . ...r.o....?r.........a1 O . GAME 515:3. 1{ME UPEWER E: 7'5 11:1w.oc11—1...1.O.l-11I 1.1.1....1o'u4. , R? . “1‘9. O as .l..?... . ,.1111_.r11111.1o.c. .11...14I.1.p . 1 .1111 1 1... 1 11.11.11... 1. . .11: 4 11.1.1321‘: 1 . .11..1. 1.1.1.2113. T111. C 1 1 . . ._ . 1 .11....3 1 1. .r. . . .. . 1 . ,. £1.11. \ .n. . 1 . . lo. 1 1.1. v. 0.3.1" 1. .0." V we gs . my cm _- . . 11 1 11 1.11.211 u._ 1 . ...1 ... :21..- .1. .n... V111~ ..(.o..meta>para=benzoic acid). The presence of hydrogen donors in an ortho position conferred greater binding abilities than the same radicals in meta and para positions. The reduced binding in the latter positions was due to a hydration shell, as opposed to the internal hydrogen bonding possible only with ortho substitution. Two ortho hydroxyl groups resulted in extensive binding as did compounds with amino substitution for the hydroxyl. Phenol and cyclohexanol were not appreciably bound, confirming the primary importance of the carboxyl group (Lindenbaum and Schubert, 1956). Competition experiments (Davison and Smith, 1961) suggested that various cyclic acids probably were bound to the same site. Hexahydrosalicylic acid apparently did not compete with salicylic acid, indicating that the benzene ring may have been of some importance, or that the alcoholic hydroxyl group in the ortho position might have interfered. Biological methods for the study of interactions have owed their usefulness to the diminution of drug activity caused by the interaction with proteins (Goldstein, 1949). Goldstein (1949) cited some of the applications. Bursck, in 1906, reported the inhibitory effect of serum on photodynamic action and other toxic effects of certain dyes, using paramecia as test organisms. He showed that serum 18 not only interfered with dye action but also altered diffusibility, fluorescence or light absorption, solubility and other properties. These phenomena he attributed to the formation of dye—albumin complexes. A quantitative analysis by Fawaz and Farah (1944) on cardiac glucosides, used systolic arrest of the frog heart for the assay of free drug concentration in the presence and absence of various serum protein fractions. C. Starches Goudah and Guth (1965) studied the interaction of benzoic acid and some derivatives with potato starch and arrowroot starch. No significant difference was observed between the binding of the two starches: 3.31 moles benzoic acid/potato starch equivalent and 3.41 moles benzoic acid/ arrowroot starch equivalent. Mansour and Guth (1968) reported that amylose was the main complexing component of starch with the drugs. The complex formation was due to the entrapment of the drug in the alpha-helical structure of amylase with supplementary stabilization by dipole-dipole interactions. The support of the inclusion theory (Goudah and Guth, 1965) was the decrease in degree of interaction with increasing molecular weight of the drug. Caffeine showed no interaction probably due to its large stereo- chemical configuration which would be difficult to fit in the voids of the helical starch molecule. 19 Hydrogen bonding was another factor to be con- sidered as part of the complexation mechanism. The multi- plicity of hydroxyl and carboxyl groups on the starch molecule enabled interaction with the polar drugs. The relative degree of interaction (meta and para hydroxy- benzoic acids>>para-aminobenzoic acid>benzoic acid) could be explained by the electrophilic nature of the hydroxyl group contrasted with the weak electrophilic nature of the amino hydrogen and the lack of electrophilic nature of the benzene hydrogen (Goudah and Guth, 1965). MATERIALS AND METHODS I. Materials The sodium benzoate used was a Pfizer U.S.P. dense powder with 100% activity. The soy protein was the PromineR-D Isolated soy protein manufactured by Central Soya, Chicago, Ill. The corn oil used was Mazola 100% corn oil. 11. Methods A. Preparation of the MOdel System 1. Brines The brines were prepared following commercial specifications for a 50/50 pack out ratio (50% peppers, 50% brine by weight). The equilibrated brine must be 17-18 grains as acetic acid (1.7 to 1.8%) and 1.5 to 2.0% NaCl. The actual values for the brines prepared were 3.6% acetic acid and 3.0% NaCl to give the desired equilibrated concen- tration. A 10% sodium.benzoate solution was used to prepare brines with benzoate concentrations of 0.0, 0.1, 0.2, and 0.3%. The benzoate solution had to be added to the water prior to the addition of the acid, particularly at the 0.3% level, to prevent the crystallization of benzoic acid. When crystallization did occur, the brine was filtered through Whatman #41 filter paper to remove the crystals. The pH of the solutions ranged from 2.7 for the 0% benzoate 20 21 brine to 3.0 for the 0.3% benzoate brine. 2. Agar Blocks a. Fat Agar blocks were prepared with 0.0, 0.5, 1.0, 2.0, 3.0, 5.0, 10, 15, 20, 30, and 40% corn oil by weight. All samples contained 2.0% agar. Deionized water was added as required. The agar and water were steamed until the agar dissolved. The hot solution was transferred to a tared blender jar and when necessary, water was added to replace any lost during steaming. The appropriate weight of oil was added and the sample was blended at high speed for 3 minutes in a Sorvall Omni-Mixer. The homogenized sample was poured into ice cube trays and put into a refrigerator to solidify. The cubes were approximately 30 ml in volume. b. Protein Agar blocks were prepared with 0.0, 0.5, 1.0, 2.0, 3.0, 5.0, 10.0 and 15.0% by weight PromineR-D soy protein isolate and 2.0% agar. The procedure was the same as that used to prepare the fat samples. The agar solution was allowed to cool slightly as the protein tended to clump and gel when added to the hot solution. Blending was done at a slower speed and only long enough to dissolve the protein. This was to prevent excessive foaming. c. Fat-Protein Combinations Fat-protein combinations were prepared using food composition table values to simulate six food samples. The 22 combinations were (1) 0.4% fat, 5.4% protein (peas), (2) 0.5% fat, 7.8% protein (kidney beans), (3) 0.5% fat, 2.7% protein (mushrooms), (4) 1.0% fat, 3.3% protein (corn), (5) 2.1% fat, 16.4% protein (avocado), (6) 5.7% fat, 11.0% protein (soy beans). The procedure was the same as that used to prepare the protein blocks. 3. Packing of the Model System Two or three cubes were weighed in tared laboratory sample containers. An equal weight of brine was added. Two or three replicates were made of each combination of benzoate concentration and fat or protein concentration. In one experiment, brine was added equivalent to two times the weight of the protein or fat containing agar cubes. B. Preparation of Food Products Six food products were purchased for this study: peas, corn, mushrooms, soy beans, kidney beans and avocado. Four replicates were made for each product in the 0% benzoate brine and six replicates for the three other brines. Frozen peas and corn were allowed to thaw in the plastic bags. The peas were packed as they were, while the starchy liquid from the corn was discarded before packing. Raw mushrooms were steam blanched for 2-2k minutes depending on the size.- No test was made for enzyme in- activation. Soy beans and kidney beans were allowed to soak in warm distilled water overnight. The water was discarded and 23 replaced with fresh water. The beans were cooked over a steam table for 2% hours. Fresh avocados were cut into chunks and packed in the brine solutions. C. Benzoate Analyses 1. Spectrophotometric Method A Beckman DU spectrophotometer with a Gilford model 220 digital readout attachment was used for the determination of benzoate in the brines from the model sys- tem and food products. A scan from 250 to 285 nm.showed a peak for the benzoate with minima at 267.5 and 277.5 nm and a maximum.at 273 nm. Readings were takenmat the three wavelengths. The height of the absorbsnce peak was used to measure the sodium.benzoate concentration. A standard curve was made of benzoate dissolved in a ten-fold diluted equilibrated brine with no benzoate (dilution brine). The pH was near 3.0. The linear portion of the curve was at benzoate concentrations of 0.0 to 0.01% (Horwitz, 1965b). a. Brines The original brines added to the sample were analyzed spectrophotometrically. The 0% and 0.1% brines were diluted 20-fold in the dilution brine and the 0.2% and 0.3% brines were diluted 40-fold. The dilution allowed reading in the linear portion of the standard curve. The dilution brine was used as the blank. b. Fat Samples 24 No interference was observed in the absorption of the brines from the fat samples in the 267.5 to 277.5 nm range. Samples with 0.0 and 0.1% benzoate brines were diluted lO-fold and 0.2% and 0.3% brines were diluted 20- fold. c. Protein Samples Protein dissolved in the brine was found to absorb in the region of the benzoate peak. Furthermore, the solutions became turbid upon dilution. Sodium.benzoate standard curves were made for each protein concentration at each dilution (10 or 20 fold), using the appropriate 0% benzoate sample. The solutions were diluted in centrifuge tubes and centrifuged at 12,000 rpm (15,000 x g) in a Sorvall SS3 Automatic Superspeed centrifuge for 20 minutes prior to reading. The samples containing benzoate were analyzed by diluting, centrifuging and reading. The 0% benzoate sample of the appropriate protein concentration and dilution was used as the blank. Above 5.0% protein, all samples were diluted 20-fold. d. Fat-protein Samples and Food Samples The fat-protein samples and food samples were analyzed in the same manner as the protein samples. All samples were diluted 20-fold to reduce turbidity and decrease the number of standard curves. 2. A.O.A.C. Titrimetric Method The A.O.A.C. method for benzoate determination was used for analysis of commercial food samples (Horwitz, 1965a). 25 The brine and tissue were analyzed separately (noting the proportions of thetmo phases) for determination of benzoate partitioning. Modification of the A.O.A.C. method was necessary to assure complete extraction of benzoate from all food samples. Solid samples were chOpped, if necessary, to facilitate blending. They were diluted 1:1 (w/w) with saturated NaCl and homogenized in a Sorvall Omni-Mixer for 2 to 3 minutes, or until the slurry appeared homogenous. Aliquots (150 grams) of the undiluted liquid samples or homogenates from solid samples were weighed into 400 m1 beakers and 25-35 grams of solid NaCl were added. The solution was titrated to pH 8-9.5 with 10% NaOH. The alka- line sample was transferred to a 500 ml volumetric flask. Rinsing of the beaker was done with 10-15 m1 of 0.02 N NaOH. Saturated NaCl solution was added to the mark. The solution was allowed to stand for at least two hours with periodic shaking. The solution was then centrifuged for 20 minutes at 7500 rpm (6780 x g) in a Sorvall SS3 centrifuge and filtered through cheese cloth. Two hundred m1 of the solution were pipetted into a 400 m1 beaker and titrated to pH 1.0 with a 3.025 N HCl solution. To this solution 2 m1 of methyl red indicator were added (200 mg methyl red/liter 0.02 N NaOH). Four chloroform extractions were done (70, 50, 40 and 30 ml chloroform) in a 500 m1 separatory funnel. The chloroform was collected in a round-bottom flask. The methyl red was 26 extracted as an orange pigment in the chloroform. A clear chloroform phase resulting from the extraction of all the methyl red from the aqueous phase was an indication of the effectiveness of the extraction. Formation of an emulsion at the interface of the two phase system often occurred with food samples. The emulsion was easily broken by centrifuga- tion of the emulsion layer for 15 sec. at 2000 x g in a clinical centrifuge. Except in cases where the emulsion problem.was particularly severe, this procedure was only used after the final chloroform extraction. Evaporation to dryness was accomplished in 20-25 ‘minutes on a flash evaporator with the water bath kept at 4o-44°c. The residue was dissolved in 30 ml ethanol and titrated with 0.05 N NaOH to pH 8.2. One ml 0.05 N NaOH = 7.2 mg anhyd. sodium benzoate. D. Protein Determination The food samples were analyzed for protein content using the micro-Kjeldahl method described by Lillevik (1970) with some modifications. The digestion mixture consisted of 5.0 g CuSO4: 5 H20 and 5.0 g Se02 in 500 ml of concen- trated H2804. Approximately 15 mg of protein for each sample were digested with 4 m1 of the digestion mixture for one hour. After cooling, 1 ml of 30% H202 was added to each flask and digestion continued for an additional hour. The digested mixture was rinsed with 10 ml deionized water and 27 then neutralized with approximately 25 m1 of a 40% NaOH solution. The released ammonia was trapped in 15 ml of a 4% boric acid solution containing 5 dr0ps of the indicator. The distillation was continued until a final volume of 60 ml of the solution was collected in the receiving flask. E. Fat Determination The food samples and fat-agar samples were dried overnight at 10000. The dry samples were extracted with ether in Soxhlett extractors at a rate of 4-5 drOps/second for 8 hours. The extracted fat was collected in a tared flask and weighed following evaporation of the ether. F. Calculations 1. Percent benzoate in the tissue of a sample separated into the tissue (solid) and brine (non-solid) phases. Bt = % benzoate equilibrated in the tissue phase Bi = % benzoate in the initial brine used for packing. Measured spectrOphotometrically. tn I b r % benzoate in the brine phase of the sample after equi- libration. Measured spectrophotometrically. L = fraction of liquid in sample (non-solid fraction). T = fraction of tissue in sample (solid fraction). Bt = Bi - (Bb x L) T 2. Ratio of the percent benzoate in the tissue to the percent benzoate in the brine. 28 -R = % benzoate in the tissue/% benzoate in the brine R = Bt / Bb 3. The Partition Coefficient is the concentration of benzoate in the fat and/or protein divided by the concentra- tion of benzoate in the aqueous phase. P.C. 8 Partition Coefficient Cf,p = benzoate concentration in the fat and/or protein (in mg of benzoate in the fat and/or protein/grams of fat and/or protein in a 100 gram sample). C = benzoate concentration in the aqueous phase (in mg of benzoate in the total aqueous phase/total grams of liquid in the brine and in the tissue in a 100 gram sample). Be = mg of benzoate/100 g of the initial brine after equi- libration. Bf,p = benzoate (in mg) bound to the fat and/or protein in a 100 gram sample A = grams of aqueous in the brine + grams of aqueous in the tissue in a 100 gram sample F,P = grams of fat and/or protein in a 100 gram sample a) Be = Bi x 1000 x L b) Bf,p = Be - (Bb x 10 x A) c) Cf,p = Bf,p/F,P d) Ca = (Bb x 1000) A e) P.C. = C /C RESULTS 1. Model System A preliminary equilibration study with a 2% agar- cube model system indicated that equilibration of the benzoate between the cubes and the brine was attained after six days storage with no significant change in readings after 12 days. All samples were stored for at least three weeks to assure equilibration. Agar cubes which had been equilibrated in brine con- taining benzoate were transferred to fresh brine without benzoate to determine whether benzoate would transfer out of the blocks into brine as readily as it transferred in the other direction. The sets of samples used for this study were the protein-agar blocks, fat-protein agar blocks simulating food products (simulated foods), and laboratory- prepared food products (actual foods). With a knowledge of the benzoate concentration in the brine after the first equilibration, a prediction was made for the benzoate con- centration in the brine following the second equilibration. The expected values were determined with equivalence equations: % benzoate in the brine X (% benzoate in brine (first equil.) = (second equiIL) % Benzoate in the initial % benzoate in the cubes brine 29 30 A comparison of the expected values and actual values for the simulated foods and actual foods packed in an initial 0.2% benzoate brine is shown on Table 1. Table 1. A comparison of expected and actual benzoate con- centrations in the brines of samples after replacement of the equilibrated brine with a fresh brine and the establishment of a second equilibra- tion. Values are given for the simulated food products and actual foods packed in an initial 0.2% benzoate brine. SIMULATED FOODS ACTUAL FOODS PRODUCT Expected % Actual % Expected % Actual % benzoate benzoate benzoate benzoate in brine in brine in brine in brine Kidney a b Beans ' 0.047 0.049 0.044 0.043 Corna’b 0.049 0.045 0.048 0.045 Peas 0.045 0.100 0.047 0.040 Mushroomsa'b0.050 0.049 0.049 0.050 Soy Beansa 0.040 0.040 0.039 0.035 Avocadoa 0.042 0.037 -- -- 8No significant difference between expected and actual values of simulated foods at the 95% level of significance. Data used were for the three benzoate concentrations. bNo significant difference between expected and actual values of actual foods at the 95% level of significance. Data used were for the three benzoate concentrations. A one-way analysis of variance was done for each pro- duct using values at all benzoate concentrations to test for significance between the expected and actual values. Nota- tions were made for those products in which there was no 31 significant difference at the 95% level of significance. A comparison of expected and actual values for all simulated foods was significant, as was the same analysis for the food products. There was no significant difference when the simulated and actual foods were analyzed together. Another study was done in which the fat and agar cubes were packed in brines containing 0.1, 0.2, and 0.3% benzoate, but using brine twice the weight of the cubes (33% cube, 67% brine). This increased the amount of benzoate which could be bound by the fat or protein. The equilibrated concentration of benzoate available in the fat or protein increased, as did the concentration in the aqueous phase. The net result was a ratio very close to the ratio for samples with equal weights of brine and cubes (50% cube, 50% brine). The ratios for the 0.2% benzoate data appear on Table 2. To study the effect of the binding of benzoate by the agar, 1%, 2%, and 4% agar blocks were prepared. There was no significant difference among the three agar concen- trations at the three benzoate levels. The percent benzoate in the equilibrated brine was approximately 0.054 for all three agar concentrations in the 0.1% benzoate brine and 0.105 and 0.114 for the 0.2 and 0.3% brines, respectively. 32 Table 2. Ratios of benzoate distribution for two pack-out ratios (50-50 and 33-67) of fat and protein agar cubes in a 0.2% benzoate brine. R=% Benzoate in tissue/% Benzoate in brine % FAT OR PROTEIN FATa PROTEIN 50/50 33/67 50/50 33/67 0.0 1.09 1.30 1.29 1.36 0.5 1.09 1.04 1.36 1.58 1.0 1.33 1.25 1.36 1.52 2.0 1.24 1.30 1.47 1.52 3.0 1.37 1.22 1.52 1.76 5.0 1.62 1.54 2.00 2.07 10.0 2.11 2.84 2.26 2.49 15.0 2.57 2.49 2.52 2.88 20.0 2.75 2.95 30.0 3.53 3.73 40.0 3.73 4.91 aNo significant difference at the 95% level of significance between the two sets using values from.the three benzoate concentrations. 11. Effect of Fat Concentration on Benzoate Distribution Figure 1 is a plot of the concentration of benzoate in the aqueous phase vs. the percent fat (w/w) in agar blocks. Least squares exponential equations of the form y=aebx fit the data, but did not show the curvature suggested by the points. The solid lines follow the equation, and the broken lines follow the data points. 33 The partitioning of a preservative between two or more phases was expressed in terms of a partition co- efficient (von Schelhorn, 1964; Bean st 31. 1965; Patel and Romanowski, 1970; Bean, 1972). This is the ratio of the con- centration of benzoate in the fat to the concentration of benzoate in the aqueous phase. Figure 2 shows the data plotted as the partition coefficient vs. the percent fat. A power function equation (y=axb) was used. For the three benzoate concentrations, the partition coefficient at the lowest fat concentration (0.5%) was near 30. It decreased sharply to 12-13 for the 5% concentration and gradually decreased at higher fat concentrations. The final value at 40% fat was in the range of 6-8 for the three benzoate con- centrations. III. Effect of Protein Concentration on Benzoate Distribution The concentration of benzoate in the aqueous phase vs. the protein concentration was plotted in Figure 3. The data points were quite scattered, but a least squares exponential equation of the form y=aebx could be used though the correla- tion coefficients were -.802, -.948 and -.969 for the 0.1, 0.2, and 0.3% benzoate brines, respectively. There was no significant difference in the shape of the curves for the three benzoate concentrations. The percent decrease from equilibrated benzoate concentrations in the 0% samples to the equilibrated benzoate concentration in the 15% samples was shmilar for fat and protein. The values for protein were % BENZOATE IN THE AQUEOUS PHASE 34 0.3% BENZOATE BRINE v = 0.148 E-JW X -.979 .090 . .050 .1051) 0.27. BENZOATE BRINE v = 0.096 5"022 x R=-.95Ll .065 L IMS .050 0.1% BENZOATE BRINE O - 022 x Y = 0.050 E ' 'mo :- \\\ , R = "0977 a‘ ‘O~.. .030 .. ° ‘3.“ .120 L L 1 1 j J I ~ 5 10 15 20 w 25 30 35 40 ‘73 FAT (WW) FIGURE 1. PERCENT BENZOATE IN THE EQUILIBRATED BRINE vs. % FAT N AGAR CUBES FOR THREE INITIAL BENZOATE CONCENTRATIONS IN BRINE, HE SOLID LINES FOLLOW THE EQUATION, THE BROKEN LINES FOLLON THE DATA POINTS. (BENZOATE) IN FAT / (BENZOATE) IN AQUEOUS 35 0.3% BENZOATE BRINE Y = 20.854 {266 R = -0959 + i 0.2% BENZOATE BRINE 30 v = 21.699 {301 R = -.901 4‘ 4 0.17. BENZOATE BRINE ' 30 Y = 22.940 {352 = -0939 20b 10 L . ? l 1 j I l 1 4 0 5 10 15 20 25 30 35 40 Z FAT (W/W) FIGURE 2. PARTITION COEFFICIENTS (BENZOATE CONCENTRATION IN FAT/BENZOATE CONCENTRATION IN AQUEOUS) vs. FAT CONCENTRATION IN FAT-ACAR BLOCKS IN BRINES OF THREE BENZOATE CONCENTRATIONS. Z BENZOATE IN AQUEOUS PHASE 36 0.37.1 BENZOATE BRINE v = 0.144 EDS“ x R = -.969 0.2% BENZOATE BRINE 0.1% BENZOATE BRINE Y = 0.049 9028 x 0 o R=‘.8m 5 10 15 7. PROTEIN (WW) FIGURE 3. PERCENT BENZOATE IN EQUILIBRATEO BRINES vs. % PROTEIN IN AGAR Guess FOR Tl-REE INITIAL BENZOATE CONCENTRATIONS IN Hf BRINE. 37 35.0, 35.6, and 37.1% while those for fat were 37.7, 38.8, and 41.6% for 0.1, 0.2, and 0.3% benzoate brines, respectiv- ely. Figure 4 shows the effect of protein concentration on the partitioning coefficient. The decrease in the co- efficient values with an increase from 0.5% to 15% protein fit a power function equation (y=axb). There was some indi- cation of an asymptotic decrease if the protein concentration was further increased. IV. Effect of Binding of Benzoate on Antimicrobial Activity An important aspect of the binding of benzoate by the fat is the effect which it has on the preservative activity of the benzoate. An attempt was made to determine the effective benzoate concentration in the aqueous phase of the model system. Malt extract (2%) was incorporated into the agar cubes. The equilibrated samples were inoculated with 108 cells/m1 of Hansenula anomala or with brine from a spoiled jar of peppers known to contain yeast. Quantitation of growth was done by turbidimetric readings at 525 nm 0, 5, and 12 days after inoculation. A random sampling was done for plate counts. Agar block fragmentation interfered with turbidimetric measurements. The plate counts showed no evidence of any surviving cells in the pH 3 brine with 1.5% NaCl and no benzoate. The brine was replaced with water. A 30% sucrose solution was added to some replicates to give a final concentration of 1% sucrose. No growth was observed (BENZOATE) IN PROTEIN / (BENZOATE) IN AQUEOUS 38 1(1) I? 80 0.37. BENZOATE BRINE ”.603 60L Y = 49.9) x Fm ”0 I. R = -0972 g 20 4: _x- 80 0.2% BENZOATE BRINE L. . '056 60 Y = 41.99 x ’40 = —'951 20 .. EDP 0.1% BENZOATE BRINE o '0352 40 - v = 21.89 x 2° \88°\3__9 "516 Ft) 1 1 20 0 5 10 15 z PROTEIN (N/w) FIGURE 4. PARTITION OOEFFICIENTS (BENZOATE CONCENTRATION IN PROTEIN/BENZOATE CONCENTRATION IN AQUEOUS vs. PROTEIN CONCENTRATION IN PROTEIN AGAR BLOCKS IN BRINES OF THREE BENZOATE CONCENTRATIONS. 39 in any of the samples in two weeks. The inability to obtain growth of test organisms in the model system with no benzoate present made it impossible to show the decrease in preservative effectiveness which has been observed in other systems (Hibbott and Monks, 1961; Anderson and Chow, 1967). V. Effect of Combinations of Fat and Protein on Benzoate Distribution Figure 5 is a comparison of the partition co- efficients of benzoate in fat and protein. The 0.2% benzoate data were used. Figure 6 shows the same data as Figure 5' but plotted as the ratio of the percent benzoate in the tissue to the percent benzoate in the liquid. This curve followed the equation of a parabolic function (y=1.l92 + 0.1097 x -0 0011 x2). Table 3 contains the ratios of the percent benzoate in the tissue to the percent benzoate in the liquid phase using the 0.2% benzoate brine (equilibrated =0.1%). The systems studied were a group of food products and protein- fat agar blocks to simulate these foods. For all ratios at the three benzoate concentrations, the paired analysis of variance showed no significant difference at the 95% level of significance between the simulated food products and the actual food products. The parabolic curve of Figure 6 was used for predic— tion of the ratios for the simulated and actual food products packed in a 0.2% benzoate brine. These values appear in Table 4. 40 1.21% E323 Md 5. BEE 9.8.5 52 2E9: S E E 298.. S .2“. N .m> A8823 5 @3352; S. 21 5 Egg mEmBEEB zoEEé .m was: :13: 23.8% N mo 2“. N 8 3 cm on 3 . 4 o .. ........ x ....... i S 8 R A.x . u> ._. v 235E u 0 III a E? 903107 NI (flWBBO/NJ ED NEW NI (11170298) 8 41 $5: zEEE N + E“. N .szm ESszm «N5 81 m. was .299: N + E“. N .m> most 803% w: 5 E569 N NE» 9 3%; me. 2” Egg ME Go 2.3 .0 98¢ S R om S o q 4 d d ..H ax :85 - x 32.0 + a? I > x .3... E“. n u . x x o 25.5”: N O . m . S o x x I QM x o... msmarwm amen/mum 3110214302 42 Table 3. Comparison of benzoate ratios (R:% benzoate in the tissue to the % benzoate in the liquid) of protein- fat agar blocks simulating food products and the actual food products. PRODUCT % BENZOATE IN TISSUE / % BENZOATE IN THE LIQUID Agar Blocks Food Products Peas 2.31 2.24 Corn 1.94 2.14 Mushrooms 1.86 1.97 Soy beans 3.11 3.22 Kidney beans 2.15 2.62 Avocado 3.04 * *Avocado food samples were not analyzed. Table 4. Actual and predicted benzoate ratios (R) for simulated and actual foods based on equal and cumulative binding of benzoate by corn oil and soy protein (0.2% benzoate samples). AGAR BLOCKS FOODS PRODUCT % fat + % fat + % Predicted Actual Predicted Actual protein ratio ratio protein ratio ratio Peas 5.8 1.79 2.31 5.7 1.78 2.24 Kidney beans 8.3 2.02 2.15 9.9 2.17 2.62 Mushrooms 3.2 1.53 1.86 3.5 1.56 1.97 Corn 4.3 1.64 1.94 4.0 1.61 2.14 Soy beans 1627 2.70 2.91 21.1 3.00 3.22 2.83 3.04 --- --- --— Avocado 18.5 43 VI. Commercial Food Products All commercial food samples were separated into solid and liquid fractions and analyzed separately for benzoate content. The results were expressed as the ratio (R) of the percent benzoate in the tissue to the percent benzoate in the liquid. Table 5 contains data for different types of pepper packs. Table 5. Benzoate distribution between tissue and aqueous in different types of paper products. Benzoate Benzoate Benzoate PRODUCT <7.) (7.) (7.) in sample in tissue in brine R Mild peppers 0.082 0.103 0.074 1.39 Med.hot peppers 0.097 0.122 0.088 1.40 Hot pepper rings 0.152 0.178 0.135 1.32 Hot pepper rings 0.132 0.155 0.109 1.42 Mild pepperoncini 0.125 0.149 0.090 1.52 Sweet red peppers 0.080 0.099 0.059 1.69 The effect of pepper pigmentation on the benzoate distribution in hot mexi peppers and cherry peppers is shown in Table 6. Table 7 presents the benzoate distribution data for the com- ponents of a green cherry pepper. 44 Table 6. Effect of pepper pigmentation on the benzoate distribution between the tissue and brine Benzoate Benzoate Benzoate PRODUCT (7°) (7°) . (7.) in sample in tissue in brine R Hot mexi peppers a)red peppers 0.099 0.143 0.048 2.98 b)yellow peppers 0.093 0.130 0.048 2.72 Cherry peppers a)red peppers 0.073 0.086 0.050 1.71 b)green peppers 0.075 0.091 0.050 1.81 Table 7. Effect of cherry pepper components on the distri- bution of benzoate between the tissue and brine % benzoate % benzoate FRACTION in fraction in brine R Green pepper (whole 0.091 0.050 1.81 Green pepper w/o seeds 0.083 0.050 1.64 Seeds 0.171 0.050 3.40 Three jars of spoiled sweet banquet peppers and two jars of spoiled relish were analyzed for benzoate content and distribution. The results appear in Table 8. Spoilage in the sweet peppers was evident by clouded brines and some sediment at the bottom of the jars. Gas 45 escape was observed upon opening one of the three jars. The peppers were turgid and did not appear damaged. Wet mounts of the brine showed the presence of yeast. The relish samples had gas pockets along the side of the jar. No wet mount was done with these samples. Table 8. Benzoate partitioning between the solid and liquid phases in spoiled samples of peppers and relish Benzoate Benzoate —. Benzoate PRODUCT <7.) (7.) (7.) in sample in tissue in brine R Spoiled sweet banquet peppers 0.088 0.106 0.069 1.55 " 0.092 0.105 0.078 1.35 " 0.071 0.082 0.062 1.32 Spoiled relish 0.065 0.070 0.055 1.27 " 0.128 0.142 0.105 1.35 Six refrigerated food products which listed 0.1% benzoate on their ingredient declaration were purchased and analyzed for benzoate. The results were tabulated in Table 9. Italian salad dressing was analyzed in three phases: solid, aqueous and oil. The analysis results appear in Table 10. A comparison of partitioning ratios between commercial food products and food products prepared in the laboratory (0.1% equilibrated benzoate brine) appears in Table 11. 46 Table 9. Benzoate distribution between solid and liquid phases of commercial products preserved with 0.1% sodium benzoate Benzoate Benzoate Benzoate PRODUCT (7.) <7.) (7.) in sample in tissue in.brine R Kosher pickles 0.127 0.132 0.122 1.086 Gelatin fruit cocktail salad 0.062 0.080 0.059 1.36 Mild Mexican sauce 0.090 0.101 0.086 1.17 Pineapple chunks 0.071 0.073 0.069 1.06 Baked beans 0.083 0.059 0.176 0.33 Cole slaw 0.070 0.043 0.155 0.27 Table 10. Distribution of benzoate among the solid, aqueous, and oil phases of Italian salad dressing Benzoate Benzoate Benzoate Benzoate PRODUCT <7.) <7.) (2) (7.) in sample in solid in aqueous in oil Italian salad 0.174 0.064 0.043 0.286 dressing (Benzoate) solid (Benzoate) oil (Benzoate) aqueous (Benzoate) aqueous 1.488 6.65 47 Table 11. Comparison of benzoate distribution ratios for food products prepared in the laboratory and commercial food products LABORATORY PRODUCTS COMMERCIAL (0.1% Benz., equil.) R FOOD PRODUCTS R Peas 2.24 Hot mexi peppers 2.98 Corn 2.14 Gelatin fruit 1.36 salad Mushrooms 1.97 Mild Mexican sauce 1.17 Soy beans 3.22 Hot pepper rings 1.42 Kidney beans 2.62 Baked beans 0.33 DISCUSSION I. Model System The 2% agar-cube model system was satisfactory for the observation of benzoate partitioning as a function of lipid and protein content. It did not disintegrate in the pH 3 acetic acid brine, it attained equilibrium.re1atively rapidly (6 days), and the agar concentration had no effect on the benzoate which was bound. Two additional tests were conducted with the model system. A study was carried out to determine the ease of benzoate transfer out of agar cubes and food products into a brine containing no benzoate. The transfer out of the cubes and foods was proportional to the initial transfer from.the brine into the cubes and foods. Refer to Table 1. In the second test, cubes were packed in benzoate-containing brines weighing twice as much as the cubes. These results were compared in Table 2 to cubes packed in an equal weight of brine. The ratios were close enough to conclude that an increased amount of brine con- taining benzoate did not change the distribution ratio of benzoate in the cube-brine system. This indicated that there were no limiting conditions in restricting the study to a 50/50 pack-out ratio. Physical limitations of the agar blocks were (1) incorporation of more than 50% oil, which prevented 48 49 solidification and (2) incorporation of more than 15% pro- tein, which was too viscous for mixing and cube formation. II. Methods of Analysis The spectrophotometric method of benzoate analysis, whenever it could be used, was preferred: to the A.O.A.C. Official Titrimetric method. The spectrophotometric method required a knowledge of the components of the brine for the preparation of a standard curve and its use as a blank. It was not suitable for commercial food products because inter- ference from.food components was too great. The modified A.O.A.C. method was satisfactory for all the food products tested. The addition of methyl red to the extraction mixture provided a visual method to determine when benzoate extraction was completed. 'Methyl red was chosen because its solubility in chloroform was approximately equal to the solubility of benzoic acid in chloroform. At the concentration at which the methyl red was added, it had a negligible effect on the final titration with NaOH. Its presence, however, obscured the phenolphthalein end point, requiring the titration to be carried out to pH 8.2. Methyl red changed from.orange-red to yellow near pH 7.2, but the change was not sharp enough to be used as a titration end point. III. Binding of Benzoate by Fat A. Discussion of Binding Data Benzoic acid, being more soluble in lipid than in 50 water, was expected to partition into the lipid over a range of lipid concentrations. Figure 1 indicated an ex- ponential decrease in the benzoate concentration in the aqueous phase as the percent oil was increased. The greatest decrease was observed from 0 to 15% fat followed by a gradual decline at higher fat concentrations. This sharp decline followed by a slower decline at higher fat concentrations was observed by Bean (1972) with methyl paraben in soy oil (the partition coefficient = 80). His data were graphed on Figure 7. Figure 1 shows curves of the same shape for the three benzoate concentrations. The percent decreases in the benzoate present in the aqueous phase of 40% fat cubes com- pared to 0% fat cubes were 56.6%, 55.3%, and 64.1% for the 0.1, 0.2, and 0.3% benzoate brines, respectively. This was a decrease of over half the benzoate assumed to be active in a product. The fat-binding data were plotted in Figure 2 as the partition coefficient (the concentration of benzoate in the fat divided by the concentration of benzoate in the total aqueous)vs. the fat concentration. The curves were similar to those in Figure 1, but they were better described by a power function equation than by an exponential equation. The partition coefficients at the various fat levels agreed reasonably well with those of von Schelhorn (1967). He reported a partition coefficient of 8.26 for mayonnaise con- taining 51.3% fat and 37.7% water. Extrapolation of the I “—n . rower. . 4 51 5:; LE. .3 ..8 >8 to 2255". < 2 was: 88:3 m: z. 5.3%. .25: C do >8 N .N #3: o o: om om 9 0 a 3i. 3SVHd SIIHDV 3H1 NI (NBEIWVd 'MHIJD 52 curves in Figure 2 to 51.3% resulted in a range (the varia- bility among benzoate concentrations) frmm 6-8, slightly lower than 8.26. The 7.90 value for marzipan at 25.5% fat and 12.3% water fell in the 7.75-8.75 range from the graph. B. Preservative Activity Many investigators have concluded that the preserva- tive activity of benzoate is dependent upon the concentration of benzoate in the aqueous phase. Hibbott and Monks (1961) prepared eight creams, all containing 0.15% methyl-p- hydroxy benzoate. The amount of lipid was constant, but fats were chosen with a variety of partition coefficients, resulting in a range of preservative concentrations in the aqueous phase. All those creams with 18 to 27.8 mg of preservative in 100 ml of water did not spoil, whereas creams with 11.6, 9.4, and 7.1 mg of preservative per 100 ml were unable to prevent growth of mold. In addition, the inclusion in oil:_water systems of propylene glycol or glycerol (2 to 20%), which reduce the partition coefficients, increased the preservative available in the aqueous phase, thus increasing preservative activity. Von Schelhorn (1964) reported a decrease in the effective preservation of ethyl and propyl esters of p-hydroxy benzoic acid with increasing fat concentration. The decreases in effective preservation of 70-90% with an increase from 0 to 10% fat, reflect the highly lipid-soluble nature of these preservatives. Though it was not possible to study the preservative activity of benzoate as a function of fat and protein 53 composition with the model system developed for this study, it is reasonable to expect that similar declines in preservative activity related to bound preservative would be observed with this system if a suitable organism is found. The total preservative required in a 2-phase system can be determined knowing the usual concentration required in the aqueous phase (CH20)’ and knowing the volume in each phase (V011, VHZO) and the distribution coefficient (D) (Patel and Romanowski, 1970): Total preservative = CH20 VH20 + D CH20 Voil This is valid in a system such as the model system which was near pH 3, where essentially all of the benzoate was in the undissociated form. Table 12 shows the proportion of undissociated benzoic acid at pH's from 2-6 (Kostenbauder, 1962). Table 12. Proportion of benzoic acid undissociated at various pH levels (Kostenbauder, 1962) pH % Undissociated Benzoic Acid 2 99.4 3 94.3 4 82.5 5 13.7 6 1.6 pKa Benzoic Acid = 4.2 54 It is clear, that the undissociated benzoic acid in the aqueous phase must be considered in determining the activity of benzoate when it is used at pH 4 and above. The total preservative required at any pH value can be determined by the following equations (Kostenbauder, 1962). Ka = (Hf) (A'), Ka = the dissociation constant of the acid (HA) The fraction undissociated = (HA) (HA) + (A-) after substituting for HA - (Hf) (III) + Ka Undissociated acids = (Total) (Ht) (H?) + Ka Total required preservative = [Required cone. of undissoc.acig7 ((11+) + Ks) q. (H ) IV. Binding of Benzoate by Protein Figures 3 and 4, showing plots of the binding of benzoate by protein were very similar to Figures 1 and 2 for fat. Again, there was no significant difference in the shape of the curves among the three benzoate concentrations. The shape of the curves for Figures 2 and 4 was similar to the shape of a curve obtained by Davison and Smith (1961) in studying the binding of benzoate at pH 5.4 to bovine serum albumin. They suggested that the curvilinear relationship 55 obtained was an indication that at least two binding sites existed. Kazmi and Mitchell (1971) used this relationship to interpret preservative-surfactant interaction. They suggested that the binding sites within micelles did not behave independently of one another. It was possible that uptake of solute into the micelles progressively altered the interaction between the binding sites and the solute leading to a change in both the number of sites available and the association constant. Perhaps the binding of benzoic acid by fat can also be explained this way. Most of the protein-preservative interaction studies were carried out at pH 6-7. At this pH the benzoate ion was the form bound by the protein (Klotz, 1946; Meyer and Guttman, 1968). The present study showed that undis- sociated benzoic acid was bound by soy proteins at pH 3.0. It is probable that hydrophobic pockets exist in the protein molecules which are of the right size for binding of the hydrophobic benzoic acid molecules. Figures 5 and 6 show similar relationships between the benzoate distribution coefficient and the concentration of fat or protein, but soy protein does appear to bind slightly more benzoate on a weight basis than corn oil. However, distribution coefficients in protein and fat were combined to determine if a reasonable prediction of benzoate distribution in food products could be made using data obtained with a model system. 56 The fact that the combination of fat plus protein simulating the food (colum.3 in Table 4) was a better pre- diction of the benzoate distribution in the food products (column 6 in Table 4) than the combined data which assume equal binding by protein and fat (column 5 in Table 4) suggests that protein does bind more benzoate on a weight basis than fat. However, the data indicate that addition of 0.5 to the predicted ratio (column 5 in Table 4) will pro- vide a good estimate of the distribution between bound benzoate and soluble benzoate when the proximate composition of a food is known. V. Commercial Samples Commercial food products containing sodium.benzoate were separated into the brine or liquid phase and the solid or tissue phase. The modified A.O.A.C. method was used to determine the benzoate in each phase. The results were ex- pressed as the ratio of the percent benzoate in the tissue (or solid) to the percent benzoate in the brine (or liquid). Different types of pepper products were analyzed (Table 5). The data showed accumulation of benzoate in the edible portion of food products for two reasons. First, the average benzoate concentration in the product was abovethe 0.1% limit. Secondly, in products, such as peppers, where benzoate accumulates in the tissue and the brine is usually not consumed, the benzoate in the food as eaten can be 57 greater than 0.1% even though the average concentration in the jar is 0.1% or less. The ratio of benzoate in the tissue to brine was the highest for red peppers (1.69). The ratio for green peppers (pepperoncini) was also quite high (1.52). The rest of the peppers with ratios from 1.32 to 1.42 were yellow. Table 6 showed higher ratios for the red peppers (2.98) than the yellow peppers (2.72) of hot mexi peppers, but also higher ratios for dark green cherry peppers (1.81) than red cherry peppers (1.71). The pigment had some effect, but it did not account for the difference between hot mexi peppers and cherry peppers. The data in Table 7 showed a high benzoate concentration in seeds. The large quantity of seeds in mexi peppers undoubtedly contributed to the high ratio observed with these peppers. It would be of interest to know the spoilage rate in peppers with a high tissue benzoate concentration compared to similar products with a lower ratio. If previous studies of preservative effectiveness are correct, it is expected that more spoilage would occur with higher distribution ratios. Table 8 contains the benzoate distribution data for three jars of spoiled sweet banquet peppers and two jars of spoiled relish. The results did not show any cause for spoilage on the basis of benzoate content. The benzoate concentration in the brines coupled with a low pH of 3.2 and a salt concentration of 1.5% should have been sufficient to 58 inhibit yeast growth. Pitt and Richardson (1973) reported spoilage in beverages and tomato sauce, where the benzoate levels were as high as legally allowed. They isolated strains of Saccharomyces bailii, Pichia membranaefaciens, and Candida krusei. The yeast were capable of growth in 0.02-0.07% benzoic acid and/or 1.5% acetic acid (pH 3.5). The spoilage was sporadic and believed to be due to post- processing contamination or contamination from improperly cleaned fillers. Table 9 contains data for several refrigerated food’ products. The ratios for pickles, gelatin salad, Mexican sauce and pineapple chunks (close to one in most cases) were lower than those observed for peppers. The ratio of 0.274 for cole slaw was explained by the mayonnaise present in the liquid phase, which resulted in a low partitioning ratio. The ratio of the percent benzoate in the liquid/the percent benzoate in the solid was 3.6. The baked beans were diluted to aid in the separation of the beans and the sauce. The liquid was centrifuged and the supernatant and precipitate were analyzed separately. After accounting for the initial dilution, the benzoate concentration in the aqueous was 0.176%. Solubilization of proteins may account in part for this high benzoate concentration in the aqueous phase. Table 8 has results for Italian salad dressing, analyzed as a three-phase system. The concentrations in the solid and aqueous were similar to those of other products 59 analyzed. The ratio of 6.65 for oil and aqueous was higher than would have been predicted from Figure 6. The value for the benzoate in the oil phase is not reliable due to problems with the final titration. Table 11 compares the ratios for laboratory- prepared food products (0.1% benzoate after equilibration) and some of the commercial food products. The food products prepared in the laboratory had relatively high ratios (2-3), whereas the commercial products, on the average, stayed in the l—l.5 range. The discrepancy may be attributed to the nature of the liquid phase. The laboratory products were packed in a simple brine containing only acetic acid, NaCl, and benzoate. The commercial brines, in addition, contained combinations of oils, spices, emulsifiers, and other in- gredients. This would tend to favor more partitioning of benzoate into the liquid phase, thus reducing the partition— ing ratio. The increased benzoate concentration in the liquid would not necessarily indicate a greater preservative capacity, since the benzoate would not be free, but bound to the in- gredients in an inactive form. CONCLUSIONS In the present study of a model system.and selected food products, previous observations of benzoate uptake by fat were confirmed. In addition, it was found that protein also binds benzoate. Von Schelhorn (1964) and Hibbott and Mbnks (1961) reported a decrease in antimicrobial activity of preservatives accompanying a decrease of the preservative in the aqueous phase of an aqueous-lipid system. In the present study, increased fat and protein concentrations decreased the amount of benzoic acid in the aqueous phase. The benzoate present in the aqueous phase of a product con— taining 15% fat and/or protein was only two-thirds of that present with no fat or protein. In a similar manner, the benzoate concentration in the aqueous was decreased by one- half by increasing the fat concentration from 0 to 40%. Equal weights of corn oil and soy protein bound benzoic acid in the same manner, but the binding of benzoic acid by protein was shown to be greater than the binding by corn oil. The relationship between the benzoate distribution in a fat-protein model system and the quantity of fat and protein in that system can be used for the prediction of distributions in food products. With a relatively simple system, (peppers, cucumbers, or other fruits and vegetables 60 61 packed in a cover brine containing negligible amounts of fat or protein) the model system can be used to predict the ratio (R). In this type of system, the % benzoate in the liquid is equivalent to the effective benzoate concentration. The separation of the brine and tissue allows a measure of the benzoate concentration in the edible portion. In the case of a product such as cole slaw, (A product in which the fat and protein are in the nonsolid phase or in both the solid and nonsolid phases, and in which the phases are not normally separated for consumption), if the fat, protein and moisture contents are known, the model system.can be used to predict the effective benzoate concentration in the entire system. Analyses of commercially prepared peppers of differ- ent varieties, pickles, and Mexican sauce showed benzoate concentrations in the edible portion greater than 0.1%, which is the maximum legal limit. This was true even when the total benzoate concentration was below 0.1%. All the products analyzed were labeled as containing 0.1% sodium.benzoate, however, there existed variability in the actual amounts measured even between two jars of the same product. The range of benzoate concentration in the samples analyzed was 0.062 to 0.174%. It appears that often processors do not have sufficient control of benzoate addition to food products to assure that it will be within the legal limit. 62 The ratio of the percent benzoate in the solid/ percent benzoate in the liquid was decreased by the addition of ingredients such as oils, spices, emulsifiers in the liquid phase. This increase in the benzoate concentration in the aqueous phase, however, would not be expected to have an effect in increasing the preservative activity as the benzoate would be bound in an inactive form. This study focused on the binding of benzoic acid by fat and protein. There is still too little known about the nature of this binding. Information concerning the mechanism of binding between benzoic acid and proteins at low pH, especially in food systems, is unavailable. It is ex- pected that benzoate bound to protein will be unavailable as a preservative just as has been previously observed for benzoate in fat (Hibbott and MOnks, 1961; von Schelhorn, 1964). This expectation needs to be tested. The binding of benzoic acid by carbohydrates, particularly starches, and cell wall constituents should be studied to get a better picture of the binding of a pre- servative by these components of a food system. Since ethyl p-hydroxybenzoic acid, propyl p- hydroxybenzoic acid and sorbate are also lipid soluble, it is expected that in fodd products similar distribution phenomena would be observed. This possibility should be investigated and the quantitative relationships developed. B IBLIOGRAPHY BIBLIOGRAPHY Aalto, T.R., Firman, M.C., and Rigler, N.E. 1953. P-hydroxy- benzoic acid esters as preservatives 1. Uses, anti- bacterial and antifungal studies, properties and determination. J. Am. Pharm. Assoc. Sci. Ed. 42, 449-457. Anderson, R.A. and Chow, C.E. 1967. The distribution and activity of benzoic acid in some emulsified systems. J. Soc. Cosmetic Chemists 18, 207-214. Anonymous, 1969. Preservatives for the food industry. Pfizer Chemicals Division, N.Y. Banfield, F.H. 1952. Problems arising from the use of chemicals in food preservatives, including anti-mould agents. Chemistry and Industry, 114-119. Bean, H.S. 1972. Preservatives for pharmaceuticals. J. Soc. Cosmetic Chem. 23, 703-720. Bean, H.S., Heman-ackah, S.M., and Thomas, J. 1965. The activity of anti-bacterials in two-phase systems. J. Soc. Cosmetic Chem. 16, 15-30. Bedford, P.C.C. and Clarke, E.G.C. 1972. Experimental benzoic acid poisoning in the cat. The Veterinary Record 29, 53-58. Bosund, I. 1959. The bacteriostatic action of benzoic and salicylic acids I. The effect on the oxidation of glucose and pyruvic acid by Proteus vulgaris. Acta Chemica Scandinavica 13, 803-8I3. Bosund, I. 1960. The bacteriostatic action of benzoic and salicylic acids III. The effect on pyruvate and acetate oxidation by different organisms. Acta Chem. Scand. 14, 1231-1242. Bosund, I. 1962. The action of benzoic and salicylic acids on the metabolism of microorganisms. In "Advances in Food Research," Vol. 11, Chichester, C.O., Mrak, E.M. and Stewart, G.F., eds., pp. 331-353, Academic Press, N.Y. 63 64 Chichester, D.F. and Tanner, F.W. 1972. Antimicrobial Food Additives. In "Handbook of Food Additives," 2nd Ed., Furia, T.E., ed., Chemical Rubber Co., Cleveland. Chittenden, R.H., Long, J.H., and Herter, C.A. 1909. The influence of sodium.benzoate on the nutrition and health of man. U.S. Department of Agriculture Secretary Report No. 88. Clague, J.A. and Fellers, R.C. 1934. Relation of benzoic acid content and other constituents of cranberries to keeping quality. Plant Physiology 2, 631-636. Cruess, W.V. 1932. Hydrogen-ion concentration in preserv- ative action. Ind. and Eng. Chem. 24, 648-649. Cruess, W.V., Richert, F.H., and Irish, J.H. 1931. The Effect of hydrogen ion concentration on the toxicity of several preservatives to microorganisms. Hilgardia 6, 295-314. Davison, C. and Smith, P.R. 1961. The binding of salicylic acid and related substances to purified proteins. J. of Pharm. and Exp. Therap. 133, 161-170. de Navarre, M.G. and Bailey, H.E. 1956. The interference of nonionic emulsifiers with preservatives II. J. Soc. Cosmetic Chem. 1, 427-433. Deuel, H.J., Jr., Slater, R.A., Weil, 0.8., and Smyth, H.F.,Jr. 1954. Sorbic acid as a fungistatic agent for foods I. Harmlessness of sorbic acid as a dietary component. Food Research 19, 1-12. Entrekin, D.N. 1961. Relation of pH to preservative effectiveness I. Acid media. J. of Pharm. Sci. 50, 743-746. '__ Evans, W.P. and Dunbar, S.F. 1965. Effect of surfactants on germicides and preservatives. In "Surface Activity and the Microbial Cell," pp. 169-190, S.C.I. Monograph No. 19, Society of Chemical Industry, London. Faith, W.L., Keyes, D.B. and Clark, R.L. 1950. Benzoic Acid In "Industrial Chemicals," pp. 127-132, John Wiley and Sons, Inc., N.Y. Fanelli, G.M. and Halliday, S.L. 1963. Relative toxicity of chlortetracycline and sodium.benzoate after oral administration to rats. Arch. int. Pharmacodyn. 144, 120-125. 65 Fonda, M.L. 1972. Glutamate decarboxylase: Inhibition by monocarboxylic acids. Archives of Biochem. and Biophysics 153, 763-768. Goldstein, A. 1949. The interactions of drugs and plasma proteins. Pharmacological Reviews 3, 102-165. Goudah, MLW. and Guth, E.P. 1965. Complex interactions of starches with certain drug pharmaceuticals. J. of Pharm. Sci. 23, 298-301. Gould, G.W. 1964. Effect of food preservatives on the growth of bacteria from spores. In "Microbial Inhibitors in Food," 4th International Symposium on food microbiology, Goteborg, Sweden, Molin, N. ed., pp. 17-24. Griffith, W.H. 1929. Benzoylated amino acids in the animal organism IV. A method for the investigation of the origin of glycine. J. of Biol. Chem. 33, 415-427. Hager, G.P., Chapman, C.W., and Starkey, E.B. 1942. The toxicity of benzoic acid for white rats. J. Am- Pharm. Assoc. Sci. Ed. 33, 253-255. Haggman, J. and Nikkila, O.E. 1962. Chemical preservatives in foodstuffs V. The effect of benzoic acid on the sporulation of Bacillus megaterium. Maataloustieteellinen Aikakauskirja 33, 96-106. Harshbarger, K.E. 1942. Report of a study on the toxicity 3f segeralafood preserving agents. J. Dairy Science 5, 1 9-17 . Hibbott, H.W., and Monks, J. 1961. Preservation of emulsions; p-hydroxy benzoic ester partition coefficient. J. Soc. of Cosmetic Chem. 33, 2-10. Horwitz, W., ed. 1965a. Ch. 27. Preservatives and Artificial Sweeteners, Benzoic Acid, Titrimetric Method. In "Official Methods of Analysis of the A.O.A.C.," 10th Ed., pp. 450-451, A.O.A.C., Washington, D.C. Horwitz, W., ed. 1965b. Ch. 27. Preservatives and Artificial Sweeteners, Benzoic Acid, Spectrophotometric Method. In "Official Methods of Analysis of the A.O.A.C.," 10th Ed., p. 451, A.O.A.C., Washington, D.C. Kazmi, S.J.A. and Mitchell, A.G. 1971. Interaction of pre- servatives with cetomacrogol. J. Pharm. Pharmocol. ‘33, 482-489. 66 Klotz, I.M; 1946. Spectrophotometric investigations of the interactions of proteins with organic anions. J. Am, Chem. Soc. 6Q, 2299. Kostenbauder, H.B. 1962. Physical chemical aspects of preservative selection for pharmaceutical and cosmetic emulsions. In "Developments in Industrial Microbiology," Vol. 3, p. 286, Plenum.Press, N.Y. Kotaki, A., Harada, M. and Yagi, K. 1966. Protective action of benzoate on the inactivation of D-amino acid oxidase by lyoxal. The Journal of Biochem., Japan 69, 592-59fi. Kowalewski, K. 1960. Abnormal pattern in tissue phospho- lipids and potassium produced in rats by dietary sodium.benzoate. Protective action of glycine. Arch. int. Pharmacodyn. 124, 275-280. Lillevik, R.A. 1972. Ch. 20. The determination of total organic nitrogen. In "Methods in Food Analysis," 2nd Ed., pp. 611-612, Joslyn, Mr, ed., Academic Press, N.Y. Lindenbaum, A. and Schubert, J. 1956. Binding of organic anions by serum albumin. J. of Physical Chem" 69, 1663-65. Mansour, Z. and Guth, E.P. 1968. Complexing behavior of starches with certain pharmaceuticals. J. of Pharm. Sci. 21, 404-411. Meyer, M.G. and Guttman, D.B. 1968. Interactions of xanthine derivatives with bovine serum.albumin III. Inhibition of binding. J. of Pharm, Sci. 52, 245-249. Morse, K.E. 1951. Mbde of action of sodium.benzoate. Food Research 16, 1-9. Oka, S. 1960a. Studies on transfer of antiseptics to microbes and their toxic effect Part I. Accumulation of acid antiseptics in yeast cells. Bull. Agr. Chem. Soc. Japan, 24, 59-65. Oka, S. 1960b. Studies on transfer of antiseptics to microbes and their toxic effect Part II. Relation between adsorption of acid antiseptics on yeast cell 32d gggigagoxic effect. Bull. Agr. Chem. Soc. Japan Oka, S. 1964. ‘Mechanism.of antimicrobial effect of various food preservatives. In "Microbial Inhibitors in Food" 4th international symposium on food microb., Goteborg, Sweden, Molin, N., ed., pp. 3-16. 67 Patel, N.K. and Foss, N.E. 1965. Interaction of some pharmaceuticals with macromolecules 11. Binding of certain benzoic acid derivatives by polysorbate 80 12d cegomacrogol 1000. J. of Pharmac. Sci. 54, 95-9 . Patel, N.K. and Kostenbauder, H.B. 1958. Interaction of preservatives with macromolecules I. Binding of parahydroxy-benzoic acid esters by polyoxyethylene 20 sorbitan monooleate (Tween 80). J. Am. Pharm. Assoc. Sci. Ed. 42, 289-293. Patel, N.K. and Romanowski, J.M. 1970. Heterogenous systems II: Influence of partitioning and molecular inter- actions on in_vitro biologic activity of preservatives in emulsions. 3. of Pharm. Sci. 22, 372-376. Pisano, F.D. and Kostenbauder, H.B. 1959. Interaction of preservatives with macromolecules II. Correlation of binding data with required preservative concen- trations of p-hydroxybenzoates in the presence of Tween 80. J. Am. Pharm. Assoc. Sci. Ed. 48, 310-314. Pitt, J.I. and Richardson, R.C. 1973. Spoilage by pre- servative-resistant yeasts. CSIRO Food Research Quarterly 33, 80-85. Radin, M.J. 1914. A note on the quantity of benzoic acid contained in prunes and cranberries. The J. of Ind. and Eng. Chem. 6, 518. Rittenberg, D., and Schoenheimer, R. 1939. Studies in protein metabolism VI. Hippuric acid formation studied with the aid of the nitrogen isotope. J. of Biol. Chem. 121, 329-331. Schwartz, C.S. and Mandel, H.G. 1972. The selective inhibi- tion of microbial RNA synthesis by salicylate. Biochem. Pharmacol. 21, 771-785. Shatalova, A.A. and Meerov, G i. 1962. The possibility of using benzoic acid- C1 for determining the synthetic- antitoxic function of the liver in man. Med. Radiol. 7, 48-52. von Schelhorn, M.L. 1964. Investigations of the distribu- tion of preservatives between fat and water in foods. In "Microbial Inhibitors in Food" 4th international symposium on food microb., Goteborg, Sweden, Molin, N., ed., pp. 139-144. 68 von Schelhorn, M.L. 1967. Distribution of preservatives between fat and water. II. Relation between physical- chemical distribution and antimicrobial effectiveness of preservatives in fat-containing foods. Z. Lebensm. Unters. Forsch 111, 227-41. Wan, S.H. and Riegelman, S. 1972. Renal contribution to overall metabolism of drugs. 1. Conversion of benzoic acid to hippuric acid. J. Pharm. Sci. 61, 1278-84. Weddenburn, D.L. 1964. Preservation of emulsions against microbial attack. In "Advances in Pharmaceutical Sciences," Vol. 1, Bean, H.S., Beckett, A.H., and Carless, J.E., eds., pp. 195-268, Academic Press, N.Y. White, A. 1941. Growth-inhibition produced in rats by the oral administration of sodium benzoate. Effect of various dietary supplements. Yale Journal of Biol. and Medicine 11, 759-768. Wiley, H.W. 1908. Influence of food preservatives and artificial colors on digestion and health IV. Benzoic acid and benzoates. USDA Bureau of Chem. B§11etin No. 84, washington, Government printing 0 ice. Wiley, H.W. 1929. "The History of a Crime Against the Food Law," The De Vinne-Hallenbeck Co. Inc., N.Y. Williams, R.T. 1947. Ch. 7. The metabolism.of aromatic acids. In "Detoxication Mechanisms," John Wiley and Sons, Inc., New York. Winsley, B.E. and Walters, V. 1965. The influence of pH upon the antifungal activity of phenol and benzoic acid. J. Pharm. and Pharmacol. 11 suppl. 228-273. MICHIGAN STATE UNIVERSITY LIBRARIES 3 1293 0316! 1481