314 ASSOCIATION OF OFFICIAL AGRICULTURAL CHEMISTS [Vol. XVII, NoJ DETECTION OF IN CULTURED BUTTERMILK AND COTTAGE CHEESE By G. A. RICHARDSON and N. P. TARASSUK (Division of Dairy Industry, University of California University Farm, Davis, California) Downloaded from https://academic.oup.com/jaoac/article/17/2/314/5765365 by guest on 29 September 2021 The Stokes method1 for the detection of gelatin in and was accepted as official by the Association of Official Agricultural Chemists.2 It has been recognized, however, that this method gives erroneous results in the case of sour milk, sour cream, cultured buttermilk and cottage cheese. This is presumably due to the presence in these products of sub- stances resulting from the partial hydrolysis of the , such sub- stances resembling gelatin in their behavior to the reagents used in the Stokes test. The Seidenberg hot water modification3 was designed to make the Stokes test specific for gelatin, but it often fails clearly to dif- ferentiate gelatin from the hydrolytic products of the proteins, especially in cultured buttermilk and cottage cheese. For example, Mendelsohn4 reported that the modification failed in the case of cream to which rennet had been added, a positive test for gelatin being obtained on samples that were known not to contain gelatin. Preliminary studies in connection with cultured buttermilk showed that the Seidenberg modification yields fairly accurate results under well- controlled conditions. The test, however, was found to be too uncertain, especially from the standpoint of regulatory laboratories. If certain dis- criminating observations are made, the method of Stokes leads to fairly correct decisions in the detection of gelatin in cultured buttermilks whether fresh or old, with or without rennet. Thus, when picric acid is added to the filtrate resulting from the addition of mercuric nitrate, old samples and those containing gelatin or rennet will become cloudy im- mediately. On shaking and standing, precipitates will form in all samples of cultured buttermilks, more especially in old samples and those to which rennet has been added. The types of precipitates are different. Those re- sulting from compounds derived from proteolysis are flocculent, settle rapidly, are non-adhesive to the walls of the container, and are easily dis- turbed. The serum is clear. The precipitates of gelatin-picrate are more crystalline in nature, settle slowly, and adhere tenaciously to the bottom and sides of the container. The serum will remain opalescent for days. By making use of these observations during the past three years, three sepa- rate classes of upper division students were enabled to report correctly on heterogeneously arranged groups of unknown samples of cultured butter- milk. It is imperative, however, that for precise work the method be as inde- » Analyst, 22, 320 (1897). * Methods of Analysis, A.O.A.C, 1930, 223. »Ind. Eng. Chem., 5, 927 (1913;. « Analyst, 55, 567 (1930). 1984] EICHAEDSON, TAKASSUK! GELATIN IN BUTTERMILK AND CHEESE 315 pendent as possible from a procedure in which a comparison of types of precipitate is used in drawing final conclusions. In addition, the method should also be applicable to cottage cheese, with and without rennet. The procedure outlined above does not yield satisfactory results with cottage cheese. Downloaded from https://academic.oup.com/jaoac/article/17/2/314/5765365 by guest on 29 September 2021 Jacobs and Jaffe1 recently reviewed the literature on the detection of gelatin in dairy products and reported a new method. They use basic lead nitrate as the precipitant, and calcined charcoal as adsorbent for the "pseudo-" formed in the souring process. Picric acid or tannic acid is used as the final precipitant. This method, however, in the hands of the writers gave erroneous results with very old samples of cot- tage cheese and cultured bettermilk, and with those containing rennet. The successful use of trichloracetic acid, especially in high concentra- tions, for the precipitation of the proteins of milk suggested that this acid might aid in differentiating gelatin from the compounds resulting from the partial hydrolysis of the milk proteins of cultured buttermilk and cot- tage cheese. Sanders2 found that four parts of 10 per cent trichloracetic acid solution to one part of milk was very effective in preparing protein-free filtrates for mineral determinations. Moir3 used trichloracetic acid in his study of the distribution of the proteins of milk. Sanders4 recently reported that trichloracetic acid in sufficient concentration is al- most as efficient as tungstic acid as a precipitant for the nitrogen com- pound of milk. EXPERIMENTAL Cultured Buttermilk The procedure that was found satisfactory for cultured buttermilk is as follows: REAGENTS (a) Mercuric nitrate*—DissolveHg in twice its weight of HNO3 and dilute this solution to 25 times its volume with H2O. (It is imperative that the reagent be free from the mercurous ion.) (b) Picric acid.—Saturated aqueous solution. (c) Trichloracetic acid.—20% aqueous solution.

PREPARATION OF SAMPLE If the cultured buttermilk is cold, warm it to approximately 25°C. and stir thor- oughly. PROCEDURE To 10 cc. of the sample, add 10 cc. of the mercuric nitrate reagent. Shake the mixture, allow to stand 5 minutes, and filter through a retentive, medium-fast filter paper (Filtrate No. 1).

ilnd. Eng. Chem. Anal. Ed., 4, 418 (1932). »/. Bid. Chem., 90, 755 (1931). »Analyst, 56, 228 (1931). < This Journal, 16, 140 (1933). • Methods of Analysis, A.O.A.C, 1930, 223, 26. TABL EI.—Observations recorded in testing commercial cultured buttermilks for gelatin

SAMPLE 1 2 3 4 5 6 Filtrate from mercuric nitrate Fairly clear Fairly clear Very turbid Turbid Turbid Fairly clear (No. 1) Aliquot of filtrate No. 1 plus § Small amount Small amount Colloidal tur- Turbid, no Flocculent pre- Voluminous vol. of saturated picric acid of flocculent of flocculent bidity, no floes fiocs visible cipitate. flocculent pre- precipitate. precipitate. visible Serum col- cipitate. Serum clear Serum clear loidally tubid Serum clear Prediction No gelatin No gelatin Gelatin Gelatin Gelatin and No gelatin but rennet rennet Aliquot of filtrate No. 1 chilled in Almost clear" Slight opales- Turbid Slightly turbid Colloidal tur- Turbid, (red- ice water, plus § volume of cence bidity dish, Millon's 20% trichloracetic acid reaction)

After standing overnight at Small amount Small amount Dense - Colloidal tur- Flocculent pre- Flocculent pre- 8-10°C, with occasional shak- of flocculent of flocculent al turbidity bidity cipitate. cipitate ing precipitate. precipitate. Serum turbid Serum clear Serum clear Serum clear (reddish) (reddish) Filtrate from trichloracetic acid Clear Clear Dense colloidal Colloidal tur- Dense colloidal Clear mixture heated to 50°C. plus \ turbidity bidity turbidity vol. of saturated picric acid (50°C.) Conclusions No gelatin No gelatin Gelatin Gelatin Gelatin and No gelatin but rennet rennet Actual content Fresh cultured Old cultured Fresh cultured Fresh cultured Fresh cultured Fresh cultured buttermilk— buttermilk— buttermilk plus buttermilk plus buttermilk plus buttermilk plus no gelatin no gelatin 2 % gelatin 0.1% gelatin 0.15% gelatin rennet

plus rennet Downloaded from https://academic.oup.com/jaoac/article/17/2/314/5765365 by guest on 29 September 2021 September 29 on guest by https://academic.oup.com/jaoac/article/17/2/314/5765365 from Downloaded 1984] RICHAEDSON, TAEASSUK: GELATIN IN BUTTEKMILK AND CHEESE 317

To an aliquot of Filtrate No. 1, add § volume of saturated picric acid. Observe the mixture for clearness and type of precipitate. Chill the remainder of the filtrate (No. 1) in ice water, add £ volume of 20% trichloracetic acid, shake well, and allow to stand at 8-10°C. about 16 hours, with occasional shaking particularly during the early part of the period. Observe the filtrate-trichloracetic acid mixture for clearness and type of precipitate. Filter cold, using a medium-fast filter paper (Filtrate No. 2). Warm the filtrate to 50-55°C, hold at this temperature for at Downloaded from https://academic.oup.com/jaoac/article/17/2/314/5765365 by guest on 29 September 2021 least 5 minutes, and add \ volume of warm saturated picric acid (50°C). Observe while still warm. Table 1 illustrates what the analyst may expect to observe, and from which he may draw conclusions. It is very evident that the method of Stokes would have led to incorrect conclusions in the case of each sample. When the type of precipitate and the nature of the serum found in con- junction with the Stokes test are considered, the prediction was correct. The use of trichloracetic acid clearly substantiates the prediction. It should be mentioned that apparently no sample had undergone deteriora- tion. It is recognized that it is not the practice to add rennet to cultured buttermilk, but the possibility of protein hydrolysis occurring during the treatment and subsequent culturing procedure is not re- mote. To be able to distinguish the reactions of rennet from those of gela- tin gives the test greater reliability. The samples referred to in Table 1 and others were tested by the tech- nic of Jacobs and Jaffe.1 The test proved to be excellent for perfectly fresh samples of cultured buttermilk, but it yielded erroneous results in the case of old samples or those to which rennet had been added. When the filtrates from the basic lead nitrate and charcoal were subjected to the trichloracetic acid treatment prior to the addition of tannic or picric acid, the results were as definite as those shown in Table 1. The test so modified offers no advantages, however, and it is more laborious than the tri- chloracetic acid modification of the Stokes method. COTTAGE CHEESE The procedure outlined for cultured buttermilk was adapted to cottage cheese by altering the method of preparing the sample, and by a double precipitation with mercuric nitrate. The sample is prepared as follows: Thoroughly mix 5 grams of cottage cheese with 10 cc. of distilled water at 50- 60°C. (a rubber-tipped stirring rod is very satisfactory for this purpose). To this mixture add 5 cc. of mercuric nitrate prepared as indicated previously. Shake, let stand for about 5 minutes, and filter through a medium-fast, but retentive filter paper (No. 40 Whatman paper is satisfactory). To this filtrate add 5 cc. of the mer- curic nitrate solution, shake, allow to stand until the precipitate settles, and filter. From this point, proceed as directed previously. Table 2 illustrates the steps in the procedure, the observations to be made, and the conclusions that may be drawn.

J Loc. dt. TABL E2.—Observations recorded in testing commercial cottage cheese for gelatin

SAMPLE 1 2 3 4 5 First filtrate from mercuric nitrate Slightly turbid Turbid Turbid Turbid Turbid 2nd addition of mercuric ni- Fairly clear Flocculent ppt. Flocculent ppt. Turbid Turbid trate Serum, fairly clear Serum, turbid 2nd filtrate Filters rapidly. Fil- Filters rapidly. Fil- Filters slowly. Fil- Filters slowly. Fil- Filters slowly. Fil- trate fairly clear trate fairly clear trate, turbid trate turbid trate turbid Prediction No gelatin No gelatin, but Gelatin and rennet Gelatin Gelatin rennet | vol. of sat. picric acid Small amount of Flocculent ppt. Flocculent ppt. Very cloudy Cloudy added to aliquot of 2nd flocculent ppt. Serum clear Serum cloudy filtrate Serum clear Ppt. on standing Flocculent, non- Flocculent, non- Mixture of non-ad- Adhesive, sticky Adhesive, sticky adhering adhering hering floes and sediment sediment adhesive, sticky sediment 3rd filtrate Clear Clear Slightly cloudy Slightly cloudy Slightly cloudy Prediction No gelatin No gelatin but ren- Gelatin plus rennet Gelatin Gelatin. net £ vol. of 20% trichloracetic Small amount of Large amount of Large amount of Small amount of Small amount of acid added to remainder flocculent ppt. flocculent ppt. flocculent ppt. flocculent ppt. flocculent ppt. of 2nd filtrate, and allow- Serum fairly clear Serum more or less Serum turbid show- Serum very turbid Serum turbid ed to stand 16 hours at clear. Reddish col- ing Tyndall ef- 8-10°C. oration (Millon's) fect. Reddish coloration Filtrate (4th) heated to 50° C. Clear Clear Slight colloidal tur- Colloidal turbidity Colloidal turbidity bidity i vol. of sat. picric acid Clear Clear Colloidal turbidity Colloidal turbidity Colloidal turbidity (50°C.) added to 4th fil- trate at 50°C. Conclusions No gelatin No gelatin but ren- Gelatin and rennet Gelatin Gelatin net Actual content No gelatin No gelatin but ren- 0.2% gelatin plus 0.2% gelatin 0.1% gelatin net used in man- rennet

ufacture Downloaded from https://academic.oup.com/jaoac/article/17/2/314/5765365 by guest on 29 September 2021 September 29 on guest by https://academic.oup.com/jaoac/article/17/2/314/5765365 from Downloaded 1934] RICHARDSON, TARASSUK.' GELATIN IN BUTTERMILK AND CHEESE 319

It is evident that cottage cheese, and more especially that in which rennet has been used in manufacture, contains derived proteins which the mercuric nitrate fails to remove completely, but which are precipitated by picric acid. These precipitates differ in physical characteristics from

gelatin-picrate and thus can be differentiated from the latter. Trichlor- Downloaded from https://academic.oup.com/jaoac/article/17/2/314/5765365 by guest on 29 September 2021 acetic acid in the concentration used is effective in removing these deriva- tives, but not gelatin, with the result that the later addition of picric acid clearly distinguishes between samples that contain gelatin and those that do not contain it. In an attempt to adapt the method of Jacobs and Jaffe to the testing of cottage cheese, the cheese was mixed with water to give a concentra- tion of milk solids comparable to that of cultured buttermilk in order to observe the conditions specified for the test. The results showed that the method is satisfactory for fresh cottage cheese except when rennet is used in its manufacture. Supplementing the test by the use of trichloracetic acid failed to give satisfactory results. In conducting the modified Stokes test, as illustrated in Tables 1 and 2, it is highly recommended, although not absolutely necessary after ex- perience has been gained, that a control sample be included for compari- son. The method has also proved successful in connection with sour milk and sour cream in cases where the Stokes test is known to give erroneous results. SUMMARY In testing cultured buttermilk and cottage cheese for the presence of gelatin by the Stokes method, the addition of picric acid causes a precipi- tate to be formed even in the absence of gelatin. This is believed to be due to the presence of derived proteins, which the mercuric nitrate fails to remove. When rennet is used in the manufacture of cottage cheese, this precipitate is particularly voluminous. When one-half volume of 20 per cent trichloracetic acid is added to the filtrate obtained after precipitating the milk proteins with mercuric ni- trate, the derived proteins are caused to coagulate, and may be removed by filtration. When picric acid is added to the filtrate thus obtained, it will remain clear in the absence of gelatin, but will become cloudy in the presence of from 0.1 to 0.25 per cent gelatin. Greater amounts of gelatin result in the formation of a yellow precipitate. The modification of the Stokes method presented is particularly satis- factory for cream, cultured buttermilk and cottage cheese that contain products derived from the milk proteins by the action of heat, acidifica- tion, or primary proteolysis.