PHOTOMETRIC MICRODETERMINATION OF HUMAN GAMMA . I. DEVELOPMENT OF A QUANTITATIVE FLOCCULATION-NINHYDRIN PROCEDURE

Abraham Saifer, Michael C. Zymaris

J Clin Invest. 1952;31(1):1-11. https://doi.org/10.1172/JCI102566.

Research Article

Find the latest version: https://jci.me/102566/pdf PHOTOMETRIC MICRODETERMINATION OF HUMAN . I. DEVELOPMENT OF A QUANTITATIVE* FLOCCULATION-NINHYDRIN PROCEDURE1 By ABRAHAM SAIFER AND MICHAEL C. ZYMARIS (From the Biochemistry Department of the Division of Laboratories, Jewish Sanitarium and Hospital for Chronic Diseases, Brooklyn, N. Y.) (Submitted for publication April 23, 1951; accepted October 1, 1951) There are three general categories of procedures monium or sodium sulfate; and of then determin- employed for the quantitative determination of the ing its content with a colorimetric biuret gamma globulin content of biological fluids. These procedure. The method of Wolfson and his co- are (1) electrophoretic methods, (2) chemical workers (6) has been converted into an even procedures, usually based on salt or organic sol- simpler turbidimetric method by de la Huerga and vent precipitation, and (3) immunochemical Popper (8), and compared by them (9) to the methods. zinc sulfate and the electrophoretic procedures for A fourth class of procedures which are qualita- gamma globulin. However, recovery experiments tive or semi-quantitative in nature, are the "pro- with added gamma globulin show a range of 73%o tein flocculation reactions." These flocculation to 92% with average values of about 85%o, which procedures are widely used in clinical work to de- are unsatisfactory results for a quantitative pro- termine changes in protein composition of disease cedure (6, 8). sera, especially in gamma globulin, as compared to A number of experimenters have employed al- normal sera (1). cohol-water mixtures at low temperatures and The application of the electrophoretic technique controlled pH to separate out various protein frac- of Tiselius to biological and medical problems has tions from human plasma. Pillemer and Hutchin- been fully discussed in recent review articles by son (10) have employed methanol-water mix- Stern and Reiner (2), Luetscher (3) and Gutman tures with acetate buffers to obtain -globu- (4). From a clinical viewpoint electrophoretic lin ratios which conform closely to those obtained methods have the disadvantages of requiring rela- with electrophoretic techniques. The most widely tively large amounts of protein material, expensive used procedures of this kind are the ethanol- equipment, and skilled technical help and then per- buffer methods developed by Edsall (11) and by mit only a few individual determinations to be per- Cohn and his associates (12-15). Their later formed daily. However, they remain the standard procedures are applicable for work with a few mil- procedure against which other simpler methods liliters of serum (16, 17) and are especially useful are generally compared in the quantitative esti- in animal protein fractionation studies, e.g., rat mation of a protein fraction. (18), dog (19), and guinea pig (20), where elec- In recent years three separate salt fractionation trophoretic analyses are generally unsatisfactory. procedures have been proposed for the determina- From a clinical viewpoint such procedures are tion of gamma globulin in sera and have been used also unsatisfactory in that they require large in clinical investigative work. The methods of amounts of blood, i.e., 5 to 20 ml. of plasma, and Jager and Nickerson (5), Wolfson and his associ- additional expensive equipment, such as refrig- ates (6) and Kibrick and Blonstein (7) have erated centrifuge or water bath; recover less than the advantages of requiring relatively small 90%o of the gamma globulin present electrophoreti- amounts of serum, i.e., 0.5 to 1.0 ml.; of deter- cally; and necessitate involved and time-consum- mining the gamma globulin directly on the precipi- ing techniques for the isolation and quantitative tate obtained by salt fractionation with either am- measurement of any particular component, e.g., 1 Presented before the Division of Biological Chemistry gamma globulin. of The American Chemical Society at the 119th meeting, A number of investigators (21, 22) have pro- Boston, Mass., April 15, 1951. posed the use of immunochemical procedures for 1 2 ABRAHAM SAIFER AND MICHAEL C. ZYMARIS determining gamma globulin in human sera. Such termining the experimental conditions under which procedures are truly microprocedures in that they gamma globulin could be flocculated out of solu- can determine 30 to 1,000 ug. of protein accurately tion quantitatively and the protein content of the and are used to determine quantitatively the gamma complex measured by the ninhydrin reaction of globulin content of cerebrospinal fluid as well as Moore and Stein (29). The effect of the presence sera (22). of each of the other protein fractions, which to- The disadvantages of the immunochemical pro- gether constitute the whole plasma, on the floccu- cedure are the involved procedures needed for the lation method for gamma globulin was also investi- preparation and assay of the antisera used, the gated quantitatively. This permitted a separation use of a refrigerated centrifuge and the technical of the fractions into interfering and non-inter- skill and the time required for the involved analy- fering ones and, with the aid of the electrophoretic ses. In addition Jager and his coworkers (23) data, the working out of simple salt fractionation have found that the immunological method yields techniques for the elimination of the interfering abnormally high values for serum gamma globu- fractions in applying the method to the micro- lin as compared to electrophoretic results. determination of serum gamma globulin. It is because of the difficulties encountered in the aforementioned procedures for clinical work that EXPERIMENTAL Kunkel (24) attempted to use a protein floccula- A. Reagents and apparatus tion procedure, i.e., zinc sulfate turbidity, as a 1. Sodium chloride. 0.85% solution. quantitative measure of serum gamma globulin con- 2. Phosphate buffer (pH 7.00)-0.4 M. Dissolve 39.2 g. tent. In contrast to the results obtained by Kunkel anhydrous NaHPO4 and 20.5 g. KHPO, in a liter of (24), d-e la Huerga, Popper, and coworkers (9) solution. Adjust pH to 7.00 ± 0.02 with a pH meter. report that zinc sulfate turbidity values do not cor- 3. Hayem's solution, modified (29). Dissolve 10.0 g. NaCl, 11.0 g. NaSO4 (anhydrous) and 2.5 g. HgCl2 in a relate well with either electrophoresis or the protein liter of solution. content of the centrifuged zinc-protein precipitates. 4. Cephalin-cholesterol emulsion. Prepared from the They attribute this to the influence of albumin and commercial product supplied by Difco Laboratories, Inc., lipid intake on the zinc sulfate turbidity. These Detroit, Michigan, according to Hanger (27), as follows: findings were confirmed by Schmid (25) who Heat 35 volumes of distilled water in an Erlenmeyer flask to 65° C. Add to the water 1 volume of the ether found that there was no direct relationship be- solution of the cephalin-cholesterol solution, drop by drop tween the gamma globulin fraction and the zinc with constant agitation, until a homogeneous emulsion is turbidity value when the salt fractionation method formed. Continue heating the emulsion on a hot plate, of Jager and Nickerson (5) was used to measure at just below the boiling point with occasional shaking, the gamma globulin. The lack of correlation en- until 30 volumes of the emulsion remain. Cool to room in disease cases, where electrophoretic temperature before using. In these experiments a fresh countered batch of emulsion was prepared for each day's run. gamma globulin values are known to be high 5. Gamma globulin standard solution.2 Weigh out ap- (e.g., multiple myeloma), was even more striking. proximately 60 mg. of gamma globulin (Cohn's Fraction Saifer (26) applied the Hanger cephalin-cho- II) 3 into a wide-mouth 125 ml. glass bottle. Add 50 ml. lesterol flocculation reaction (27) to the semi- 2Electrophoretically pure gamma globulin for use as quantitative determination of the minute amounts a standard in this procedure may be easily prepared by the of increased gamma globulin present in cerebro- ammonium sulfate fractionation method of Kendall as spinal fluid in disease cases. In qualitative stud- described in the textbook, "Experimental Immunochem- ies with Cohn's protein fractions, it was shown istry," by Kabat, E. A., and Mayer, M. M.; Charles C Thomas, Springfield, Ill., 1948, p. 461. that flocculation occurs, with the cephalin-choles- 3 Obtained from the University Laboratory of Physical terol emulsion in the presence of Hayem's solu- Chemistry of the Harvard University Medical School tion (28), with some of the fractions (for instance, through the courtesy of Doctors J. T. Edsall and L. D. fibrinogen or gamma globulin) but not with others Wojick. Fractions Lilly II-1 and Squibb 652 (98-99% (e.g., albumin) under the stated experimental con- gamma globulin) were used in these runs. Fractions used in these experiments were prepared by the older fraction- ditions. ation procedure known as Method 6. Those prepared by The present paper deals with the quantitative the newer Method 10 (15) are apt to contain glycine as aspects of this reaction from the viewpoint of de- a contaminant and must be dialyzed before use. PHOTOMETRIC GAMMA GLOBULIN. I. METHOD 3 of 0.85% NaCI under vacuum conditions, slowly and with 14. Spectrophotometer, Model B Beckman or other constant agitation until most of the protein goes into solu- similar instrument adaptable for reading small test tubes. tion. Centrifuge off any insoluble material at about 3,000 15. Test tube, cuvettes, selected-Select 75 X 12 mm. r.p.m. and transfer the clear supernatant fluid to a glass o.d. test tubes, pyrex, without lip, for light absorption stoppered bottle. Keep in an icebox at 4° when not in use. characteristics in a manner similar to that described in The exact protein content of this solution was determined the paper by Moore and Stein (29). for each day's run by a micro-Kjeldahl digestion and distillation method (30) and diluted with saline to a con- B. Procedure for the quantitative determination of elec- centration of about 400 jFg. per ml. The optical density trophoretically pure gamma globulin (Cohn's Fraction values for the standard gamma globulin solution were II) obtained with 0.00, 0.25, 0.50, and 1.00 ml. aliquots, in 1. Pipette 0.00 ml., 0.25 ml., 0.50 ml., 0.75 ml. and 1.00 quadruplicate, without prior flocculation, by the photo- ml. of the standard gamma globulin solution (Fraction metric ninhydrin procedure described below, simultane- II) in quadruplicate, into the selected test tubes, 75 X 12 ously with each flocculation run so as to establish the level mm., and then add 0.85% NaCl so that the volume in at which the flocculation method gave quantitative results. each tube is 1.00 ml. Add 0.5 ml. of the phosphate buffer 6. Ether-reagent grade, anhydrous. (pH 7.00) to each tube, followed by 0.75 ml. of the 7. Sodium hydroxide solution-0.1 N. Hayem's solution and 0.75 ml. of the cephalin-cholesterol 8. Ninhydrin reagent. The ninhydrin used in these emulsion.5 Mix the contents of each tube by inverting experiments was obtained from the Dougherty Chemical twice, replace in the Kahn racks and place the racks in a Company, Jamaica, New York, and twice recrystallized as water bath at 37.5° for 45 minutes. Centrifuge for 15 min- described by Moore and Stein (29). utes at 3,000 r.p.m., using multiple carriers, pour off su- a. Citrate buffer (pH 5.0)-0.2 M. Dissolve 21.008 g. pernatant fluid, invert tubes in the racks and allow to citric acid, reagent, and 200 ml. of 1 N NaOH in 500 drain on to cotton gauze placed on the bottom of the ml. of solution in a volumetric flask. Add a few racks. Break up the precipitate by tapping the bottom crystals of thymol as a preservative. of the tube and add approximately 2 ml. of ether using a b. Ethylene glycol monomethyl ether (methyl cello- 10 or 20 ml. glass syringe. Mix well to wash the precipi- solve). It was generally found necessary to purify tate and centrifuge again at 3,000 r.p.m. for five minutes. the product obtained from commercial sources by Aspirate off ether layer carefully without disturbing the distillation under vacuum to prevent turbid solutions. protein layer with gentle suction, using a glass pipette c. Ninhydrin solution. Dissolve 68 mg. of anhydrous drawn out to about a ;is inch opening. Repeat washing, stannous chloride (Flaked Stannochlor,4 obtained centrifugation and aspiration once more. Remove all from Metal and Thermit Corporation, New York, traces of ether by placing racks with the tubes in the 37.5° N. in water bath for an hour or longer. Y.) 50 ml. of citrate buffer in a 100 ml. glass 2. determination stoppered graduated cylinder. Add to this solution Quantitative of protein content with 2.0 g. of recrystallized photometric ninhydrin procedure. Add to each tube 1.0 ninhydrin and then 50 ml. of ml. of 0.85% and the purified methyl cellosolve. Stopper and dissolve NaCl 0.2 ml. of 0.1 N NaOH and mix by shaking the rack with the tubes. Add 0.5 ml. of the by shaking until all the salt is dissolved. A play of freshly prepared colors occurs during the solution process but the final ninhydrin reagent to each tube, mix the reagent has a pale yellow color and is contents well by tapping and place the rack with the tubes water clear. in a If the solution remains turbid it can usually be clari- vigorously boiling water bath for exactly 30 minutes. fied by the Mix the contents of each tube by tapping at the end of placing graduate and its contents in warm each 10 water. A sufficient amount of this reagent was pre- minute interval during the heating process. After pared for each day's run. removal of racks from the water bath allow to cool to 9. Diluent solution. Mix equal volumes of water and room temperature. Add 1.5 ml. of the n-propanol-water n-propanol, reagent. diluent to each tube, mix the contents well and centri- 10. fuge at 2,000 r.p.m. for five minutes to remove any insolu- Water bath, constant temperature, 37.5°. ble material. Clean the outside surface of each tube 11. Kahn racks. well and measure the optical density values in a spectro- 12. Centrifuge. photometer at 570 my, directly in the test tubes, against 13. Pan, aluminum, large 12" X 18" X 4" deep, for boil- pure diluent as the blank. ing water bath. A Beckman Model B spectrophotometer and a Coleman 4When Stannochlor is used in the preparation of the Junior spectrophotometer with test tube adapters were ninhydrin solution in place of the usual reagent SnCl,- used in these experiments. The results obtained in this 2H20, and the solution prepared exactly as described run are shown graphically in Figure 1. above, it was found that very low and consistent blanks 5 Where large numbers of determinations are to be per- were obtained on day-to-day runs and that no dark colors formed we have found the Cornwall Automatic pipetting developed in the solution on prolonged standing even in units, as recommended in the paper by Dern and Pull- the presence of air. There is, however, a decrease in man (31), to be both rapid and accurate for the de- color density per unit weight of protein on long standing. livery of a constant volume of a solution. 4 ABRAHAM SAIFER AND MICHAEL C. ZYMARIS Gamma globulin subfractions II-1, 11-2 and II-3 were phate buffer, pH 7.00 ± .02 was employed in all subse- run in an, identical manner and these results are also quent runs. included in Figure 1. .2. Effect of variation of the amount of Hayem's solW tion. The variation of this component was studied in a C. Studies of factors which influence the quantitative similar manner as was the phosphate buffer described in flocculation of gamma globulin from solution C 1 above. The results obtained in this run are described graphically in Figure 3. 1. Effect of variation of amount of phosphate buffer It should be noted from these data that practically no (0.1 M). To a series of selected test tubes was added flocculation occurs in the absence of the Hayem's solu- a constant amount of a standardized gamma globulin solu- tion and in the presence of the cephalin-cholesterol emuls tion. To each set of four tubes then was added the amount sion. The method yields quantitative results in the pres- of 0.1 M phosphate buffer (pH 7.00 ± .02) given by each ence of 0.75 ml. or greater of the Hayem's solution. point in Figure 2. The gamma globulin content for the 3. Effect of the variation of the amount of cephalin- different buffer concentrations was then determined by cholesterol emulsion. The variation of this component the quantitative flocculation-ninhydrin procedure exactly was studied in the same manner as were the other two as described above. components described above. The results obtained in this The results obtained in this run are shown in Figure 2. run are shown graphically in Figure 4. It is clear that buffer concentration is an important factor A considerable amount of the gamma globulin is floc- in determining whether quantitative flocculation- of the culated out of solution even in the absence of the emul- gamma globulin takes place. An amount of buffer cor- sion but in the presence of the Hayem's solution. The responding to 1.5 ml. of 0.1 M phosphate buffer or greater best quantitative result is obtained at about 0.75 ml. of gave satisfactory quantitative results. In order to re- the emulsion with slightly decreasing gamma globulin duce the volume of solution used, 0.5 ml. of a 0.4 M phos- values when greater amounts of the emulsion are used.

11 6.00 / 5.50 - A// °,S/0

4.00

3.50 z / w 3.00 - a I$/ -i 4 2.50 - o FRACTION Ir F ,/ 0. * FRACTION Z- I 2D00 i. * FRACTION 1- 2 1.50 * FRACTION I- 3 .00' #*a/ LECTROPHORETIC DATA ALB. FB. EY FRACTION U 1% .0' le| t 99% FRACTION E- 2[ 1 2% 1 98-% AA I ic.- I' ~ I I A A 0 200 400 600 800 1000 1200 1400 PROTEIN - MICROGRAMS FIG. 1. PHOTOMETRIC DATA FOR GAMMA GLODuLIN (FRACTION II) AND rrs SUBMAC'TIONS (FtCTnoNs II-1, 1I-2 AND 1-3) PHOTOMETRIC GAMMA GLOBULIN. I. METHOD 5

220

210

2001 208 Y- MICRO-KJELDAHL 180Y 208 Y- MICRO-KJELDAHL

e) 150 4 180

0

0 120

0 1601 z 0 z -i 90 I- D m 0 44 1404 60 I- aC.4

120+ 30

0 .25 .50 .75 1.00 125 iSo 1.75

0 .25 .50 .75 1.00 125 150 1.75 2.0 HAYEM'S SOLUTION - ML

PIMSPHATE BUFFER (pH 7.00)-ML. FIG. 3. EFFECT OF VARIATION OF THE AMOUNT OF FIG. 2. EFrEcT OF VARIATION OF THE AMOUNT OF HAYEM'S SOLUTION ON THE QUANTITATIVE FLOCCULA- PHOSPHATE (0.1 M) ON THE QUANTITATIVE FLOCCULA- TION METHOD FOR GAMMA GLOBULIN TION METHOD FOR GAMMA GLOBULIN vestigation of the effect of pH variation on the floccula- 4. Effect of variation of the time of standing at 37.5°. tion of the other major protein fractions. As can be ob- From a large series of preliminary experiments it was served from the data in Table I, minimum interference found that the temperature at which the flocculation oc- from the other fractions also occurs at pH 7.00. curs determines to a large extent the rate of flocculation. other At room temperature (25°) the rate of flocculation is D. Studies of the effect of the presence of protein rather small so that little flocculation occurs within an fractions on the quantitative determination of gamma hour. This rate increases markedly with increasing globulin by the flocculation-ninhydrin procedure temperature so that at 37.5°, gamma globulin in the range The fractionation of human plasma by Cohn and his of 100 to 1000 fig. is completely flocculated after ap- collaborators (12-15) and Edsall (11) has made possible proximately 30 minutes standing or longer. Higher tem- the preparation and characterization of many of the pro- peratures were not employed to avoid the possibility of tein components of human plasma. With their original denaturation of various protein fractions. The results ethanol-water separation technique, human plasma was obtained for the effect of variation of time of standing at fractionated into eight major components. The propor- 37.5° on quantitative flocculation of the gamma globulin tion of each component present in the original plasma, and from solution is shown in Figure 5. the electrophoretic composition of each fraction, are de- 5. Effect of variation of pH of phosphate buffer on the scribed in an article by Oncley, Scatchard, and Brown quantitative flocculation of gamma globulin and other (32). It was these fractions, rather than those pre- protein fractions from solution. Employing the quanti- pared with a later fractionation procedure (15), that tative flocculation-ninhydrin procedure described above, a were investigated for their effect on the quantitative floc- large number of experiments were performed with stand- culation-ninhydrin method for gamma globulin. In ad- ard gamma globulin solutions to determine the effect of dition a number of subtractions of Fraction II, i.e., II-1, small variations of pH in the vicinity of the reported iso- II-2 and II-3, were also determined with this procedure. electric point (15, 23), i.e., pH 7.3. The values given in The electrophoretic data on the bottom of each table are Table I give the results obtained at 0.7 pH units from those actually determined on the particular fraction in- the isoelectric point. Although the best quantitative re- vestigated, or those from a similar fraction, as determined sults are obtained for pH 7.00, the values obtained for at the University Laboratory of Physical Chemistry of the pH 8.00 were sufficiently quantitative to warrant an in- Harvard University Medical School. 6 ABRAHAM SAIFER AND MICHAEL C. ZYMARIS

o 420 . / 432 Y -MCRO-KJELDAHL 0 420./ C I 1400-

-i

$380 0

4c a o 360/ 44 340

320 0 15 30 45 60 75 90 05 20 MINUTES- 37.5* CEPHALIN-CHOLESTEROL EMULSMON (ML.) FIG. 5. EFFECT OF THE VARIATION OF THE TIME OF FIG. 4. EFFEcr OF VARIATION OF THE AMOUNT OF THE STANDING AT 37.5° ON THE QUANTITATI FLOCCULATION CEPHALIN-CHOLESTEROL EMULSION ON THE QUANTITATIVE METHOD FOR GAMMA GLOBULIN FLOCCULATION METHOD FOR GAMMA GLOBULIN tested to the gamma globulin present was varied so as Employing the quantitative procedure for gamma globu- to include the in are lin described above, mixtures of gamma globulin with ratio which these components present each of the other protein fractions, i.e., V (albumin), in normal plasma as given in Table VIII. The results ob- IV-1, IV4, III-0, III and I (fibrinogen) were deter- tained in these runs are given in Tables II through VII mined as if the protein mixtures were pure gamma globu- inclusive. The data on the gamma globulin subfractions lin solutions. The ratio of the particular fraction being are included in Figure 1.

BLE I Effect of pH variation on flocculation of Cohn's protein fractions

1 2 3 4 5 6 7 8 9 Protein fraction pH pH pH pH' H p 6.60 6.60 % 7o 700 8.00 8.00 Presein found ractiovere present found recovered present found evered

gag. gag. A t g. Ug. g.- Fraction II 224 138 62 219 216 99 219 188 84 (Gamma globulin-99%) 448 290 65 432 432 100 432 406 91 Fraction I 125 25 20 62.5 6.3 10.1 125 37.5 30 (Fibrinogen [61%] and antihemophilic globulin) Fraction IV-1 189 8 4.2 189 3.7 2.0 189 91 48 (Alpha,, alpha2, beta,) I l Fraction IV-4 405 6 1.5 405 5.1 1.3 405 317 78.2 (Albumin, alpha,, alpha2, beta,) Fraction V 900 6 0.67 900 20 2.2 900 743 82.5 (Albumin-95%) I PHOTOMETRIC GAMMA GLOBULIN. I. METHOD 7

DISCUSSION OF RESULTS TABLE III Effect of Fraction IV-1 on quantitative flocculation method Study of the various factors which influence the for gamma globulin (Fraction II) quantitative flocculation-ninhydrin reaction for Esti- gamma globulin, as shown in Figures 2 to 5 in- Fraction Gamma Fraction IV-1 mated IV-1 globulin Gamma glob. gamma Error Remarks clusive, prove that once a critical minimum level present present present globulin of a factor is exceeded, a further increase in the recovered amount of that factor has little influence on the ac- 199 Micro-Kjel- curacy of the method. Except for the case of the dahl and cephalin-cholesterol emulsion, where a definite Ninhydrin 107 199 0.54 185 -7.0 Ninhydrin maximum is indicated, the experimental factor 214 199 1.08 206 + 1.0 Ninhydrin chosen was sufficiently in excess of this critical 322 199 1.62 215 +1.1 Ninhydrin 535 199 2.69 188 -5.5 Ninhydrin value to avoid any possibility of error from this 749 199 3.76 204 + 1.0 Ninhydrin source. Once these critical values were determined the Electrophoretic data-Fraction I V-1 application of the method to the quantitative deter- Alpha Aiphasl Betft mination of human gamma globulin (Cohn's Frac- 77% 17% 5% tion II) was a relatively simple matter. However, quantitative results cannot be obtained unless the by the ether washings. The exact quantitative precipitated complex is thoroughly washed with composition of this complex is as yet unknown and anhydrous ether to remove any excess emulsion. is being investigated not only for gamma globulin A minimum of at least two ether washings was but for the other protein fractions as well. It is found essential for this purpose. sufficient for analytical purposes that the gamma The data in Figures 3 and 4 indicate that the globulin is removed from solution quantitatively Hg" ion in Hayem's solution plays a more im- and in a form; that is, as part of a complex, which portant role in the flocculation of gamma globulin can be easily centrifuged, washed and analyzed all from solution than does the cephalin-cholesterol in the original tube. The overall accuracy of such emulsion alone. However, both are necessary for a procedure is ± 5% for 100 to 1,000 ,ug. of quantitative results under the given experimental gamma globulin and the optical density values conditions. over this range follow the Bouger-Lambert-Beer Analysis of the complex proves that cholesterol Law. is firmly bound to the protein and is not removed The method does not differentiate between gamma globulin (Fraction II) and its subtractions TABLE II (II-1, II-2 and II-3) and in that respect con- Eject of albumin (Fraction V) on quantitative flocculation firms the findings of Jager and his coworkers (23) method for gamma globulin (Fraction II) and of Kabat and Murray (33) who found these fractions to be immunochemically homogeneous Esti- Gamma Albumin mated although according to the former investigators Albumi globulin Gamma glob. gamma Error Remarks peetpresent present globulin these subtractions are inhomogeneous with the ul- recovered tra-centrifuge and contain particles of different pg.g. pgI.% electrophoretic mobility. Cann, Brown, Kirkwood 221 Micro-Kjel- dahl and and Hink (34) have separated human gamma Ninhydrin globulin into four electrophoretically unique frac- 244 221 1.10 213 -3.6 Ninhydrin 487 221 2.20 216 -2.3 Ninhydrin tions by their electrophoresis-convection technique. 732 221 3.30 226 +2.3 Ninhydrin In spite of such differences these fractions are all 1,220 221 5.52 236 +6.8 Ninhydrin 1,750 221 7.72 242 +9.5 Ninhydrin true gamma and any variation in their relative proportions will have no effect on the de- Electrophoretic data-albumin (Fraction V) termination of the total gamma globulin content of a biological fluid by the flocculation-ninhydrin Albumin Alphai Alpha, Betai 95.5% 2% 1% 0.5% procedure. 8 ABRAHAM SAIFER AND MICHAEL C. ZYMARIS

TABLE IV TABLE VI Effect of Fraction IV4 on quantitative flocculation method Efect of Fraction III on quantitative flocculation method for for gamma globulin (Fraction II) gamma globulin (Fraction II)

Esti- Esti- Fraction Gamma FractionIWA mated Fraction Gamma Fraction III mated IV4 globulin Gamma glob. gamma Error Remarks III globulin Gamma glob. gamma Error Remarks present present present globulin present present globulin |recovered recovered , pg. sg DLp. % sig. pg. pL- % 199 Micro-Kjel- 155 Micro-Kjel- dahl and dal and Ninhydrin Ninhydrin 322 199 1.62 183 -8.0 Ninhydrin 106 155 0.69 240 +54.8 Ninhydrin 644 199 3.24 212 +6.5 Ninhydrin 212 155 1.38 298 +92.5 Ninhydrin 966 199 4.85 238 + 19.5 Ninhydrin 318 155 2.07 346 +123 Ninhydrin 1,610 199 8.08 230 + 15.6 Ninhydrin 424 155 2.76 396 +155 Ninhydrin

Electrophoretic data-Fraction I V-4 Electrophoretic data-Fraction III* Albumin Alphai Alphas Betai Beta, Alpha Beta Fibrin. Gamma 26% 10% 28% 34% 3% 7.0% 55.6% 26.2% 11.2% * From data by Oncley, Scatchard, and Brown (32). From the data in Table I it is apparent that ac- curate control of pH by means of buffers is not tenance of the pH below 7.00 gives poorer re- only important for quantitative results with gamma coveries of gamma globulin but lessens the in- globulin itself, but is especially important where terference from albumin. In view of the fact that gamma globulin is to be determined in the presence the albumin fraction inhibits the cephalin-cho- of considerable amounts of the other protein frac- lesterol flocculation reaction with gamma globulin tions, e.g., biological fluids. This is especially ap- (1) its elimination or neutralization is a most im- parent in the case of albumin where a change of portant consideration. one pH unit, i.e., from 7.00 to 8.00, causes an in- The data in Tables II to VII deal with the effect crease in the amount of protein flocculated of some of all the other fractions, as separated by Cohn's 80%o as compared to that initially present. Main- Method 6, which constitute the whole plasma. The non-interfering fractions are Fractions V, TABLE V IV-1, IV-4 and VI. The latter fraction was not tested and is listed as a non-interfering fraction be- Efect.of Fraction III-O* on quantitative flocculation method for gamma globulin (Fraction II) cause it contains only a small amount of protein

Esti- TABLE VII Fr~ton Gamma Fraction 11- mated 11-90 globulIn Gamma glob, gamma Error Remarks present present present globulin Effect of Fraction I (fibrinogen and antihemophilic globulin) recovered on quantitative for gamma globulin (FractionI) eg. g. peg. % 155 Micro-Kiel- Esti- dahl and Fraction Gamma Fraction I mated Ninhydrin I globulin Gamma glob. gamma Error Remarks 53 155 0.34 213 +37.4 Ninhydrin present present present globulin 106 155 0.68 219 +41.3 Ninhydrin recovered 159 155 1.02 229 +47.7 Ninhydrin Pg. pA. Ag. % 212 155 1.36 238 +53.5 Ninhydrin 240 Micro-Kjel- dahl and Ninhydrin Electrophoretic data-Fraction III-Ot 176.5 240 0.32 266 + 10.8 Ninhydrin. 253 240 0.64 288 + 20.0 Ninhydrin Alpha Beta Gamma Fibrin. 330 240 0.96 320 +33.3 Ninhydrin 8.3% 78.5% 5.0% 8.3% 06 240 1.27 350 +46.0 Ninhydrin * According to personal communication from Dr. L. D. Wojcik of the Harvard University Laboratory of Physical Chemistry, the Fracdtion I1I-0 used in these experiments is Electrophoretic data-Fraction I composed chiefly of beta-lipoprotein with a small amount of . Albumin Alphas Betat Fibrin. Gamma t From data by Oncley, Scatchard, and Brown (32). 7% 8% 15% 61% 9% PHOTOMETRIC GAMMA GLOBULIN. I. METHOD 9 which is mostly albumin. From the electrophoretic in plasma and a summary of the predicted errors viewpoint the non-interfering fractions are either for plasma and sera as calculated from the pre- rich in albumin or alpha protein and contain less ceding tables, Tables II to VII inclusive. In mak- than 37% of beta globulin. At the ratio of the ing these calculations it was assumed that any particular protein fraction to gamma globulin gamma globulin shown to be present in the electro- found in normal plasma (see Table VIII), these phoretic data, given at the bottom of each table, fractions when run together with gamma globulin, would be included in the analysis of sera or plasma do not give values more than 5%o greater than that for gamma globulin with the flocculation-ninhydrin due to gamma globulin alone. method. In addition it was assumed that any The interfering fractions are I, III and III- fibrinogen present in plasma would be removed and the amounts of each present in normal plasma as an interfering factor during the blood-clotting and their ratio to gamma globulin are also given in process and would not be present in sera. On the Table VIII. These fractions are rich in fibrinogen basis of these assumptions it was calculated that or beta globulin (or beta lipoprotein) and intro- an error of about + 82%o would occur if plasma duce considerably positive errors when run to- were analyzed by the flocculation-ninhydrin gether with gamma globulin. As shown in Tables method for gamma globulin. The error for sera V-VII these errors decrease rapidly with decreas- was estimated to be about + 487%. ing protein fraction/gamma globulin ratio, and in- The elimination of these large positive errors dicate the possibility of eliminating such errors for serum gamma globulin determinations, the by reducing the concentrations of these components comparison of the quantitative flocculation-ninhy- by salt fractionation prior to the flocculation drin method with electrophoretic data on the same procedure. sera, and the evaluation of a new clinical entity, the Table VIII contains data for the amount of each (total serum flocculation-ninhydrin value) protein fraction present in normal plasma, the ra- (globulin clot flocculation-ninhydrin value) tio of each fraction to the gamma globulin present both normal and disease cases is discussed in the

TABLE VIII Summation of data on the effect of Cohn's protein fractions on the quantitative flocculation method for gamma globulin (Fraction II)

Estimated amount of protein fraction Prot fraction Predicted errort Plasma protein fraction in 1 ml. of a 1:50 _ * from previous Remarks diltn. of normal Gamma glob, data-plasma plasma* Fraction II (Gamma globulin) 120 -6 With calorimetric ninhydrin method. Fraction I +(5-17) Contains 9% gamma glob. fibrin- (Fibrinogen, etc.) 68 0.57 Sera=0 ogen removed during blood clotting. Fraction III +(34-43) Contains 11.2% gamma and (Fibrinogen, beta globulin, etc.) 80 0.67 Sera- +(17-25) 26.2% fibrinogen. Fraction III-0 +(42-54) Contains 5% gamma and 8.3% (Beta lipoprotein, beta globulin) 108 0.98 Sera + (31-43) fibrinogen. Fraction IV-1 0 Presence of these fractions, at in- (Alpha, and alpha, globulins) 104 0.87 dicated ratio, gives gamma globu- Fraction IV-4 0 lin values within experimental (Albumin, alphaand beta, 116 0.97 limits of error of method. globulins) Fraction V 0 (Albumin) 632 5.3 Fraction VI 0 Not tested. The only protein (Albumin, etc.) 18 0.15 present in significant amounts is albumin. * Calculated from data by Oncley, Scatchard, and Brown(32). t Predicted error: for Plasma - + 87 pg. (aver.) or +82.5%; for Sera - 58 jig. (aver.) or +48.3%. 10 ABRAHAM SAIFER AND MICHAEL C. ZYMARIS subsequent paper. The application of the general 3. Luetscher, J. A., Jr., Biological and medical applica- principles of quantitative protein flocculation for tions of electrophoresis. Physiol. Rev., 1947, 27, 621. the microdetermination of other electrophoretic 4. Gutman, A. B., Plasma in disease, in Ad- components besides gamma globulin, is presently vances in Protein Chemistry, edited by Anson, being investigated in this laboratory, e.g., albumin M. L., and Edsall, J. T. Academic Press, Inc., New and fibrinogen. York, 1948, Vol. IV, p. 155. 5. Jager, B. V., and Nickerson, M., Simple quantitative SUM MARY chemical method for estimating gamma globulin in 1. A simple accurate photometric micropro- human serum. J. Biol. Chem., 1948, 173, 683. cedure for the quantitative determination of elec- 6. Wolfson, W. Q., Cohn, C., Calvary, E., and Ichiba, F., trophoretically pure gamma globulin (Cohn's Studies in serum proteins; rapid procedure for es- timation of total protein, true albumin, total globu- Fraction II) in the range of 100 to 1,000 ug. of lin, alpha globulin, beta globulin and gamma globu- protein, has been described. lin, in 1.0 ml. of serum. Am. J. Clin. Path., 1948, 2. The principle of the method involves the 18, 723. quantitative flocculation from solution of the 7. Kibrick, A. C., and Blonstein, M., Fractionation of gamma globulin with a cephalin-cholesterol emul- serum into albumin and alpha, beta and gamma globulin by sodium sulphate. J. Biol. Chem., 1948, sion in the presence of Hayem's solution and de- 176, 983. termination of the protein content of the ether 8. de la Huerga, J., and Popper, H., Estimation of se- washed precipitate photometrically with the nin- rum gamma globulin concentration by turbidi- hydrin reaction, all in the same tube. metry. J. Lab. & Clin. Med., 1950, 35, 459. 3. The various factors which influence the quan- 9. de la Huerga, J., Popper, H., Franklin, M., and titative flocculation-ninhydrin reaction, e.g., pH, Routh, J. I., Comparison of the results of gamma globulin and zine sulphate turbidity test with elec- temperature, time, concentration of the reagents, trophoretic determination of the gamma globulins. etc., have been carefully studied and controlled. J. Lab. & Clin. Med., 1950, 35, 466. 4. Subfractions of gamma globulin gave results 10. Pillemer, L., and Hutchinson, M. C., The determina- which were identical with those obtained with the tion of the albumin and globulin contents of hu- original whole Fraction II. man serum by methanol precipitation. J. Biol. Chem., 1945, 158, 299. 5. The effect of the presence of each of Cohn's 11. Edsall, J. T., The plasma proteins and their fractiona- other protein fractions (Method 6) which consti- tion, in Advances in Protein Chemistry, edited by tute whole plasma, have been investigated for their Anson, M. L., and Edsall, J. T. Academic Press, effect on the quantitative determination of gamma Inc., New York, 1947, Vol. III, p. 383. globulin. It was found experimentally that only 12. Cohn, E. J., History of plasma fractionation in Ad- those fractions which are rich in fibrinogen, beta vances in Military Medicine. Little, Brown & Co., Boston, 1947, Vol. I, Chap. 28, p. 364. globulin or beta lipoprotein interfere. 13. Cohn, E. J., Strong, L. E., Hughes, W. L., Jr., Mul- ford, D. J., Ashworth, J. N., Melin, M., and Tay- ACKNOWLEDGMENTS lor, H. L., Preparation and properties of serum and The authors wish to acknowledge their gratitude to plasma proteins. IV. A system for the separation Dr. Joseph Steigman of the Department of Analytical into fractions of the protein and lipoprotein com- Chemistry of the Polytechnic Institute of Brooklyn and to ponents of biological tissues and fluids. J. Am. Dr. E. A. Kabat of the College of Physicians and Surgeons Chem. Soc., 1946, 68, 459. of the Columbia University Medical School for their help- 14. Cohn, E. J., Directions for the Preparation of Pro- ful advice and criticism during the course of this work. thrombin, Isoagglutinins and Serum Gamma Glob- We are grateful to Dr. Bruno W. Volk, Director of ulin (Human), Method 9. Mimeographed pamphlet, Laboratories of the Jewish Sanitarium and Hospital for Lab. Phys. Chem., Harvard U. Medical School, Chronic Diseases, for his aid and encouragement of this Boston, Revision of July 27, 1945. work, and to Mrs. Jennie Herbst for typing and editing 15. Cohn, E. J., Gurd, F. R. N., Surgenor, D. M., Barnes, of the manuscript. B. A., Brown, R. K., Derouaux, G., Gillespie, J. M., Kahnt, F. W., Lever, W. F., Liu, C. H., Mittel- REFERENCES man, D., Mouton, R. F., Schmid, K., and Uroma, 1. Maclagan, N. F., Flocculation tests; chemical and E., A system for the separation of the components clinical significance. Brit. Med. J., 1948, 2, 892. of human blood: quantitative procedure for the 2. Stern, K G., and Reiner, M., Electrophoresis in separation of the protein components of human medicine. Yale J. Biol. & Med., 1946, 19, 67. plasma. J. Am. Chem. Soc., 1950, 72, 465. PHOTOMETRIC GAMMA GLOBULIN. I. METHOD 11

16a. Pearsall, H. R., and Chanutin, A., Electrophoretic, 23. Jager, B. V., Smith, E. L., Nickerson, M., and nitrogen and lipide analyses of plasma and plasma Brown, D. M., Immunological and electrophoretic fractions of healthy young men. Am. J. Med., 1949, studies on human gamma globulins. J. Biol. Chem., 7, 297. 1948, 176, 1177. b. Pearsall, H. R., and Chanutin, A., Electrophoretic, 24. Kunkel, H. G., Estimation of alterations of serum nitrogen and lipide analyses of plasma and plasma gamma globulin by turbidimetric technique. Proc. fractions in disease. Ibid., 301. Soc. Exper. Biol. & Med., 1947, 66, 217. 17. Lever, W. F., Gurd, F. R. N., Uroma, E., Brown, R. 25. Schmid, R., The zinc turbidity test and its clinical K., Barnes, B. A., Schmid, K., and Schultz, E. L., application. J. Lab. & Clin. Med., 1950, 36, 52. Chemical, clinical and immunological studies on 26. Saifer, A., Estimation of increased gamma globulin the products of human plasma fractionation. XL. and fibrinogen in cerebrospinal fluid; a serial di- Quantitative separation and determination of the lution-flocculation method. J. Lab. & Clin. Med., protein components in small amounts of normal hu- 1950, 36, 130. man plasma. J. Clin. Invest., 1951, 30, 99. 27. Hanger, F. M., Flocculation of cephalin-cholesterol 18. Gjessing, E. C., and Chanutin, A., Electrophoretic emulsions by pathological sera. Tr. A. Am. Physi- study of plasma and plasma fractions of normal and cians, 1938, 53, 148. injured rats. J. Biol. Chem., 1947, 169, 657. 28. Mandel, E. E., and Paris, D. A., Evaluation of the 19. Gjessing, E. C., Ludewig, S., and Chanutin, A., Frac- flocculation test with Hayem's solution. J. Lab. tionation, electrophoresis and chemical studies of & Clin. Med., 1949, 34, 653. proteins in sera of control and injured dogs. J. 29. Moore, S., and Stein, W. H., Photometric ninhydrin Biol. Chem., 1947, 170, 551. method for use in the chromatography of amino 20. Ecker, E. E., and Likover, B., A micromethod for the acids. J. Biol. Chem., 1948, 176, 367. determination of the albumin-globulin ratio in 30. Markham, R., Steam distillation apparatus suitable guinea pig serum. J. Lab. & Clin. Med., 1947, 32, for micro-Kjeldahl analysis. Biochem. J., 1942, 36, 1500. 790. 21. Kendall, F. E., Studies on serum proteins. I. Identifi- 31. Dern, R. J., and Pullman, T. N., The accuracy, pre- cation of a single serum globulin by immunological cision and utility of the syringe as a pipetting de- means; its distribution in sera of normal individu- vice. J. Lab. & Clin. Med., 1950, 36, 494. als and of patients with cirrhosis of the liver and 32. Oncley, J. L., Scatchard, G., and Brown, A., Physical- with chronic glomerulonephritis. J. Clin. Invest., chemical characteristics of certain of the proteins 1937, 16, 921. of normal human plasma. J. Phys. & Coll. Chem., 22a. Kabat, E. A., Glusman, M., and Knaub, V., Quanti- 1947, 51, 184. tative estimation of albumin and gamma globulin in 33. Kabat, E. A., and Murray, J. P., A comparison of hu- normal and pathologic cerebrospinal fluid by im- man gamma globulins in their reactivity with rab- munochemical methods. Am. J. Med., 1948, 4, 653. bit anti-gamma globulin by the quantitative pre- b. Kabat, E. A., Glusman, M., and Knaub, V., Im- cipitin method. J. Biol. Chem., 1950, 182, 251. munochemical estimation of albumin and gamma 34. Cann, J. R., Brown, R. A., Kirkwood, J. G., and Hink, globulin in normal and pathologic cerebrospinal J. H., Jr., Fractionation of human immune gamma fluid. Federation Proc., 1948, 7, 306. globulin. J. Biol. Chem., 1950, 185, 663.