The Sulfhydryl Groups of Ovalbumin L. R. MacDonnell, R. B. Silva, and R. E. Feeney From the Western Regional Research Laboratory,’ Albany, Received February 19, 1951
In an investigation of the foaming properties of egg white and egg white proteins we became interested in the relationship between foaming properties and the rate of appearance of sulfhydryl groups. Hellerman’s reagent, p-chloromercuribenzoate (1,2) appeared ideally suited for our purposes and a study of the applicability of this reagent to egg white and egg white proteins was initiated. It was first attempted to add the reagent prior to foaming the solutions and then to determine unreacted reagent after the solutions were foamed by back titration with cysteine to a nitroprusside end point. However, it was found that a considerable amount of the reagent apparently reacted with native ovalbumin prior to any foaming treatments under our conditions. This reaction with native ovalbumin was unanticipated in view of the conclusions of Anson that this protein either did not bind the reagent or bound it very loosely (3). This paper reports our studies on the reaction of p-chloromercuribenzoate with the sulfhydryl groups of native and denatured ovalbumin and egg white. Data are presented that support the conclusion that p-chloromercuribenzoate reacts with and forms a relatively stable compound with sulfhydryl groups of native ovalbumin under mildly acidic conditions. MATERIALS
Ovalbumin samples were prepared essentially by the procedure of Kekwick and Cannan (4), with either Na&O, or (NHd)$301. All samples were recrystallized three to ten times, dialyzed salefree, dried from the frozen state, and stored at 2°C. The pchloromercuribenzoate was synthesized by E. F. Jansen of this laboratory 1 Bureau of Agricultural stration, U. S. Department
and Industrial of Agriculture.
by the procedure given in Organic Synthesis (5). The purified acid was dried at 105°C. and stored at room temperature. It was nonhygroscopic and by mercury content was at least 99% pure. Solutions (0.005 M) were prepared by dissolving a weighed amount of the acid in 0.01 M NaOH. They were stable at room temperature for at least a month. Cysteine hydrochloride was prepared by tin reduction of good-quality cystine (6). The recrystallized acid was dried to constant weight in an Abderhalden drying pistol at the boiling point of acetone and stored at 2°C. in a tightly stoppered bottle. The dry acid developed only a slight odor of HtS after a year’s storage. Fresh 0.005 M solutions were prepared for each experiment and standardized against the mercury reagent. Solutions were not kept for longer than 4 hr. Nitroprusside indicator was made by mixing 1 part nitroprusside and 2 parts NalCOa. The mixture was ground to a fine powder in a ball mill. This indicator is stable and more convenient to use than nitroprusside solutions. Nitrogen determinations were made by the micro-Kjeldahl method with a mercury catalyst. Mercury analyses were performed by E. F. Potter of this laboratory according to Laug and Nelson (7). Values of pH were obtained with a Beckman model G meter. The titration procedure as routinely applied was as follows: 70-140 mg. of dry protein, or 1 ml. of solution containing similar amounts of protein was measured into a 30-ml. beaker.2 Three to five ml. of 0.2 M pH 5.3 acetate buffer and 2-3 ml. of 0.005 M p-chloromercuribenzoate were added and the solutions were allowed to stand for 5-15 min. The excess mercury reagent was then titrated with 0.005 M cysteine to the appearance of the nitroprusside test. This test was performed by mixing 2 drops of the solution with 4-6 mg. of the reagent on a chemical spot plate. In all cases, a preliminary titration was made to approximate the amount of cysteine solution required. More precise titrations were finally obtained by making a single addition of approximately 10% less (usually 0.10-0.15 ml.) of this amount of cysteine and adding the residual cysteine necessary dropwise. Because a small amount of cysteine was also required to give the nitroprusside test, it was necessary to correct for this extra cysteine. Each operator had his own corrective factor to apply because the degree of red color desired at the end point is a subjective factor. The usual concentration of cysteine detectable in this study was 0.05-0.06 mM and the absolute amount for which a correction was applied was 5-10% of the total amount employed in a titration. For determinations on denatured ovalbumin, 1.1 g. guanidine hydrochloride/ml. of solution was added after the mercury reagent, and the solution was brought back to room temperature by immersion in a 35°C. water bath for several minutes. It was then titrated with cysteine in the same manner as was native protein. Titrations of native and denatured proteins were also performed in pH 3.2 acetate buffer and pH 7.0 phosphate buffer. With the pH 3.2 buffer, a small drop of 6 N NHaOH was added to the test solution in the spot plate before adding the nitroprusside. This method is referred to as the reverse titration method. Typical results of titrations of native and z The dry ovalbumin preparations were equilibrated with air, and moisture was determined by heating to constant weight at 105°C. All weighed samples were corrected for moisture. Where solutions of the protein were used, nitrogen was determined on aliquots and the protein concentration calculated on the basis of 15.7% nitrogen (9).
MACDONNELL, SILVA AND FEENEY TABLE Typical
of Native and Denatured
at pH 5.3
Native or denatured’
0 Five ml. protein solution in 0.2 M pH 5.3 buffer for each titration. Titrations of native protein were performed by adding the mercury reagent and then back titrating with standard cysteine to determine free reagent. b Cysteine and mercury reagent solutions were 0.005 M. c Corrected from standard curve based on 0.00006 M cysteine as minimum detectable concentration. * Titrations were performed as for native ovalbumin; 1.1 g. guanidine hydrochloride/ml. solution was added after the mercury reagent. After warming to room temperature the back titration was performed. *Direct titration of denatured ovalbumin performed by adding the guanidine hydrochloride to the protein solution and then warming the solution to room temperature prior to titration of the freed sulfhydryl groups with the mercury reagent. denatured samples of ovalbumin are given in Table I. Sulfhydryl contents are expressed as equivalents per mole of ovalbumin on the basis of a molecular weight of 46,000 (8). Some particularly pure guanidine hydrochloride was available during part of this work.3 With this preparation it was possible to denature the protein for varying times in the absence of the binding reagent and to titrate the liberated sulfhydryl groups directly with pchloromercuribenzoate (see Table I).
A. The Sulfhydryl
Content of Native and Denatured- Ovalbumin
The results of 25 titrations of several samples of native ovalbumin performed at pH 3.2 and 5.3 gave a value of 2.92 equiv./mole with an 8 Guanidine hydrochloride preparations generally contain iron [see A&on (3) for discussion of heavy metal impurities]. Thus, from this standpoint, a pure guanidine hydrochloride sample is one containing no heavy metal. With a pure denaturant, ovalbumin can be denatured and a constant value for sulfhydryl groups obtained by titration at various times thereafter. With a less pure denaturant, the sulfhydryl groups decrease with increasing time of denaturation, unless, of course, a binding reagent is present during denaturation.
average deviation of 0.18 ‘equiv. for the reactive or available sulfhydry1 content. The results of 20 titrations of guanidine-denatured ovalbumin performed at pH 3.2 and 5.3 gave a value of 4.06 equiv./mole with an average deviation of 0.15 equiv. for the sulfhydryl content. Four of these titrations of denatured protein were by direct determinations of the amount of mercury reagent necessary to cause the disappearance of the nitroprusside test given by the denatured protein. The remainder were by the reverse procedure as employed for the native protein. Calculated on the basis of percentage of cysteine the average values for native and denatured protein were, respectively, 0.77% and 1.07%. Titrations performed at pH 7.0 or above with either native or denatured protein gave erratic and low values. B. Crystallization
of the p-Chloromercuribenzoate
Derivative of Ovalbumin
The crystalline derivative of the mercury reagent and native ovalbumin was easily prepared. For one preparation, 216 mg. p-chlorqmercuribenzoate was dissolved in 3-4 ml. water with the aid of alkali, and the solution was added to 9.78 g. ovalbumin adjusted to a pH of 5.3 in a volume of 50 ml. The mixed solutions were then adjusted to pH 4.7 and brought to opalescence with ammonium sulfate as for the crystallization of the unreacted ovalbumin. Conditions for crystallization were easily obtained and two subsequent recrystallizations were performed in a similar fashion. The crystals, although not measured, appeared very similar or identical to crystals of the unreacted ovalbumin. The twice-recrystallized protein was then dissolved in water, dialyzed, and lyophilized. No reactive sulfhydryl groups were found in this preparation by the reverse titration method, which indicated that these groups were still bound by the reagent. That reagent was still bound was confirmed by a mercury analysis of the preparation, although the amount of mercury found was equivalent to only 2.18 equiv. of cysteine/mole rather than the 2.92 equiv./mole found by titration of unreacted protein. However, mercury analyses on small samples in this study were usually low, and we considered this analysis as well as those described below only as confirmatory. Similar results were obtained with a second preparation of twice-recrystallized derivative.
C. The Reversibility of the Reaction of p-Chloromercuribenzoate with Ovalbumin and Cysteine
The reverse titration determinations of native ovalbumin indicated that the addition of small amounts of cysteine did not dissociate the reagent-protein derivative to a significant degree at pH 5.3 and pH 3.2. The possibility of reversibility or dissociation was further examined by the following experiment: Solutions of ovalbumin in 0.2 M pH 5.3 acetate and pH 7.0 phosphate buffers were prepared. Both solutions were divided into aliquots, and p-chloromercuribenzoate in excess of that required to react with the native, available sulfhydryl groups was added to each aliquot. In one set of aliquots the excess mercury reagent was retained; in a second set it was titrated with cysteine; and in a TABLE II The Dissociation ofthe p-Chloromercuribenzoate Derivativeof Native Ovalbuminby Cysteine
= Mercury reagent still bound after dialysis
Conditions before dialysis~ By titrationa
6.6 5.6 5.5 4.9 7.9d
protein Equiv./ mole prot.
4.35 4.35 4.35 5.15 5.15 5.15 4.35
0 1.50 23 0 2.30 23 23
By Hg analysis
Equiv./ mole prot.
100 100 42 100 100 63 29
2.45 2.45 0.92 2.45 2.07 1.97 0.65
_2.85 2.85 1.12 2.85 2.85 1.80 0.83
86 86 32 86 72 69 23
* Each sample originally contained 10 ml. of 67% ovalbumin solution and either 0.2 M acetate or 0.2 M phosphate buffer. Excess p-chloromercuribenzoate was added to each aliquot. Samples 2 and 5 were titrated with standard cysteine 10 min. after adding the mercury reagent to remove the excess reagent, and a large excess of cysteine was added to samples 3, 6, and 7. All samples were dialyzed against running tap water for 6 days to remove reagent not bound to protein. b After dialysis, 0.2 M pH 5.3 buffer was added to portions of each protein solution, and p-chloromercuribenzoate titrations were performed in the usual manner. c This protein preparation, by titration, had 2.85 equiv. native reactive sulfhydryl/ mole dry weight (0.7% cysteine). d Similar to sample 3, pH adjusted before dialysis.
third set a large excess of cysteine was added. All samples were then dialyzed for 6 days against cold running tap water to remove any free mercury reagent and cysteine. Aliquots of the solutions were employed for nitrogen determinations and titration of reactive sulfhydryl groups by the reverse titration technique. Mercury contents were determined on dried aliquots. The following observations are apparent from the data of this experiment presented in Table II: (a) At both pH 5.6 and 6.9 (samples 4 and 1) the reagent remained bound to the protein after dialysis. (b) Reagent in excessof that necessary to react with the available sulfhydryl groups was removed by dialysis. (c) Dialysis in the presence of an approximately sixfold excessof cysteine dissociated part of the reagent from the protein, and the amount dissociated was related to the pH (samples 3, 6, and 7). As described above, the mercury analyses were considered as confirmatory. These results also indicated that the intact, native protein could be recovered from the derivative of the protein by dialysis against cysteine. This was accomplished in two separate experiments in which 0.3-g. samples of the above-described crystalline derivative were dissolved in 20 ml. water containing 6 mg. cysteine, and the solutions were dialyzed for 3 days at 2°C. against 500 ml. of 0.005 M cysteine buffered at pH 7.0 with phosphate. The solutions were then dialyzed against water and titrated by the reverse titration method. In both instances the original amount (2.92 equiv./mole) of available sulfhydry1 was found and the recovered protein was easily crystallized to give typical rosettes of needles. Attempts to demonstrate a dissociation of the reagent-cysteine complex by the addition of protein were unsuccessful. In such an experiment, 115 mg. of the protein was added to each of three solutions containing 0.005 mequiv., 0.01 mequiv., and 0.015 mequiv. of both cysteine and the mercury reagent in pH 5.3 buffer. No cysteine sulfhydry1 (as detectable by the nitroprusside test) was obtained after 30 min. at room temperature. In control solutions containing only the cysteine and buffer, all or nearly all, of the added cysteine was found by direct titration with the mercury reagent. D. The Sulfhydryl
Contents of Iodine-Oxidized Denatured Ovalbumin
To 1 ml. ovalbumin solution containing 100 mg. protein was added 3 ml. of pH 3.2 buffer. The protein was then treated with amounts of
294 iodine groups groups tion of mined
from just sufficient to destroy 5-10Y0 of the total sulfhydryl (1 ml. of 0.001 N 12) to sufficient to destroy all the sulfhydryl (slightly over 1 ml. of 0.01 N 12) (10). Ten minutes after addiiodine to the native protein, the sulfhydryl contents were deterby the reverse titration method on native and guanidine-dena-
EQUIV I2 ADDED PER. MOLE
FIG. 1. The relationship of iodine-oxidizable to pchloromercuribenzoate reactive sulfhydryl groups in native and denatured ovalbumin. Light circles, native protein; dark circles, guanidine-denatured protein. The different designs of circles denote different experiments. Five different preparations of ovalbumin were employed in these experiments.
tured aliquots. Starch indicator was employed to show the absence of free iodine prior to the titration. The results of this series of titrations are presented in Fig. 1, in which the sulfhydryl titration is plotted against the amount of iodine added. A linear relationship between the amount of iodine added and the disappearance of sulfhydryl groups with both the native and denatured protein is evident. The amount of
iodine necessary to destroy all the sulfhydryl groups may also be determined from Fig. 1. This amount was 5.06 equiv. iodine/mole protein, or’ 1.2 equiv./equiv. titratable cysteine in the denatured protein.4 The results presented in Fig. 1 indicated that the sulfhydryl groups in native ovalbumin which reacted with p-chloromercuribenzoate were at least partially more susceptible to iodine oxidation than the residual sulfhydryl groups appearing on denaturation. If this were not the case and random oxidation had occurred, the reactive sulfhydryl groups of the native and of the denatured protein should have disappeared at such a rate that the curves in Fig. 1 would be coincident at the abscissa. That the sulfhydryl groups in the native protein were more susceptible was further proved by experiments in which the sulfhydryl groups detectable in the native protein were destroyed with the calculated amount of iodine and additional sulfhydryl groups were found after denaturation. In one such experiment, an ovalbumin solution buffered at pH 3.2 was oxidized with 4.3 equiv. iodine per mole of protein. No sulfhydryl was then detected by titration with the mercury reagent. When the partially oxidized protein was denatured with guanidine, a titration equivalent to 0.77 equiv. sulfhydryl/mole was found. E. Iodine Oxidation of the p-Chloromercuribenzoate Derivatives of Ovalbumin When native or guanidine-denatured ovalbumin was reacted with sufficient p-chloromercuribenzoate to react completely with the available sulfhydryl groups and iodine then added, oxidation of the reacted groups occurred. In a typical experiment, approximately lOO-mg. samples were treated with sufficient mercury reagent (3 equiv./mole for native and 4 equiv./mole for denatured) to react with the sulfhydryl groups. Iodine sufficient to oxidize only a portion of the groups was then added; as shown by starch indicator, no free iodine was present. Free mercury reagents6 was then determined with standard cysteine in 4 This amount of iodine is higher than that reported by Anson and Stanley (11). These investigators, in a study of the iodine oxidation of the sulfhydryl groups of tobacco mosaic virus, reported that the reaction with iodine was stoichiometric with both tobacco mosaic virus and ovalbumin at low temperature and in the presence of potassium iodide. In this investigation the earlier and more convenient iodine oxidation technique of Anson (10) was employed, since the primary interest was in destroying the reactivity or the sulfhydryl groups for p-chloromercuribenzoate. 6 The compound determined after oxidation might have been either p-chloromercuribenzoate or p-iodomercuribenzoate or a mixture of both.
the usual manner. In all casesthe oxidations were approximately similar to those shown in Fig. 1. Iodine oxidation of sulfhydryl groups reacted with p-chloromercuribenzoate was not unexpected since Anson (3) showed a similar oxidation of the cysteine derivative of p-chloromercuribenzoate. F. The Suljhydryl
Content of Native and Denatured Egg White
Approximately l-g. samples of egg white were diluted with 5 ml. of 0.2 M pH 5.3 acetate buffer and reverse titrations were performed as for ovalbumin. Whole egg white, on a dry-weight basis, was found to have 0.44yc available sulfhydryl as cysteine in the native material and 0.62% after guanidine denaturation. If all the sulfhydryls were supplied by ovalbumin, which comprises about 60% of the total solids (12), the values found should have been 0.46% and 0.64y0 cysteine, respectively, for native and denatured whole egg white. This close agreement with both the native and denatured whole egg white is good evidence that ovalbumin furnishes most, if not all, of the sulfhydryl groups found in egg white. DISCUSSION
It was concluded that these results demonstrated that native ovalbumin contains 3 equiv. sulfhydryl/mole which react with p-chloromercuribenzoate to form a derivative stable in slightly acidic solutions. These conclusions were based on: (a) the crystallization of the derivative, (b) the stability of the derivative upon dialysis, (c) the confirmatory agreement between the mercury contents and sulfhydryl titrations of the partially dissociated derivative, (d) the recovery of the native protein from the derivative, and (e) the loss of reactivity to the mercury reagent upon iodine oxidation of the native protein. Although Anson (3) concluded that p-chloromercuribenzoate was either not bound or was bound very loosely by native ovalbumin, his studies were conducted under different conditions, i.e., at a higher pH (pH 7.0) and by a method involving oxidation of the sulfhydryl groups with ferricyanide. Although reactive sulfhydryl groups have been found in native ovalbumin by a variety of reagents, the effects of these reagents were either not demonstrable until the protein was denatured or the extents of reaction were variable (10,13,14). With the alkylating or oxidizing agents the amount of reactive sulfhydryl found was dependent upon the concentration of the reagent or upon the particular reagent em-
ployed. During the course of this investigation, Hughes (15) reported that methylmercury iodide reacted with ovalbumin, but quantitative data have not been made available. This reagent should react with ovalbumin in a manner similar to p-chloromercuribenzoate. Mercury reagents have long been employed for the crystallization of proteins. Warburg and Christian (16) crystallized the enzyme enolasefrom ammonium sulfate solution with mercuric sulfate. Utilizing the same technique, Kubowitz and Ott (17) crystallized an enzyme from tumors which catalyzed the conversion of pyruvic acid to lactic acid, and showed the reversible combination of the enzyme with mercury. Hughes (15,18) employed methylmercury iodide to crystallize serum albumin. That the 3 equiv./mole of sulfhydryl which are reactive in the native ovalbumin are different from the remaining 1 equiv. available only after denaturation was indicated both by the titrations with p-chloromercuribenzoate and the results with iodine oxidation. The data obtained did not warrant speculation as to the reasons for these differences. Greenstein (19) in 1938 reported what is commonly considered the first reliable analysis of the cysteine content of ovalbumin. He found a cysteine content of 1.24y0 by a porphyrindin titration of ovalbumin denatured with guanidine hydrochloride. Similar values have since been reported by various other investigators; most often cited are those reported by Anson (3) and Hellerman (2). In general, these and other reports agree that the cysteine content of denatured ovalbumin is 1.20-1.25%. However, many of these results have been obtained with oxidizing and alkylating agents and the possibilities of nonspecific reactions cannot be eliminated. The lower percentage (1.07%) obtained in this study might be more accurate, or it might be low because of the difficulties reported in properly denaturing ovalbumin for sulfhydryl determinations (13). Another possibility is that ovalbumin may vary with the source of egg white (20). However, values comparing not too unfavorably with those reported herein have been previously obtained. Mirsky (21) reported a cysteine content of 0.96% using Anson’s ferricyanide technique (10). More recently, Benesch and Benesch (22) employing an amperometric titration in ethanolic solution reported a content of 0.851.34yn cysteine; the particular value obtained depended on the ethanol concentration employed. However, these workers also reported 0.96% cysteine in ovalbumin denatured in guanidine hydrochloride by direct titration with p-chloromercuribenzoate. Finally, Larson and Jenness (23) very recently have reported a content of 1.12%
MACDONNELL, SILVA AND FEENEY
cysteine. These workers oxidized the sulfhydryl groups of native ovalbumin at pH 6.6-7 with o-iodosobenzoate and employed an amperometric technique. While the reverse titration method reported herein might be used to study sulfhydryls in other proteins, it is obvious that every protein is a problem in itself and a direct adaptation of the method may not be possible. SUMMARY
1. The reactivity of the sulfhydryl groups of ovalbumin has been investigated by iodine oxidation studies and by a titrimetric method with p-chloromercuribenzoate. The titrimetric method involved the addition of p-chloromercuribenzoate to solutions of the protein buffered at pH 5.3 and back titration of the excessor unreacted mercury reagent with cysteine, with nitroprusside as an outside indicator. 2. The derivative of native ovalbumin and p-chloromercuribenzoate was recrystallized. It was dissociated by dialysis against excesscysteine at pH 7.0 and the original protein could be recovered. 3. Native ovalbumin was found to contain approximately 3 equiv./ mole reactive sulfhydryl, and denatured ovalbumin was found to contain 4 equiv./mole. 4. Partial oxidation of native ovalbumin with iodine followed by p-chloromercuribenzoate titrations indicated that iodine preferentially oxidizes those sulfhydryl groups that react with p-chloromercuribenzoate in the native protein. 5. It was concluded from titrations of native and denatured whole egg white that the ovalbumin contains most, if not all, of the sulfhydryl groups in egg white. REFERENCES 1. HELLERMAN, L., Cold Spring Harbor Symposia Quad Biol. 7, 165 (1939). 2. HELLERMAN, L., CHINARD, F. P., AND DEITZ, V. R., J. Biol. Chem. 147, 443 (1943). 3. ANBON, M. L., J. Gen. Phytil. 24, 399 (1941). 4. KEKWJCK, R. A., AND CANNAN, R. K., Biochem. J. 30, 224 (1936). 5. GILIKAN, H., AND BLATT, A. H., (ed.) Organic Syntheses, Coll. Vol. 1, pp. 159, 494, 519, John Wiley and Sons, New York, 1941. 6. HALE, W. S., unpublished procedure. Enzyme Research Division, Western Regional Research Laboratory, Albany, Calif. 7. LAUQ, E. P., AND NEIBON, K. W., J. Assoc. Ojk. Agr. Chemists, 25, 399 (1942). 8. TAYLOR, G. .L., ADAIR, G. S., AND ADAIR, M. E., J. Hyg. 33, 340 (1932). 9. CHIBNALL, A. C., REES, M. W., AND WILLIAMS, E. F., Biochem. J. 37,354 (1943).
10. ANSON, M. L., J. Gen. Physiol. 23, 321 (1940). 11. ANBON, M. L., AND STANLEY, W. M., J. Gen. Physiol. 24, 679 (1941). 12. LON~SWORTH, L.-G., CANNAN, R. K., AND MACINNES, D. A., J. Am. Chem. SOC. 62, 2580 (1940). 13. ANSON, M. L., Advances in Protein Chem. 2, 361 (1945). 14. BACQ, Z. M., AND FISCHER, P., Bull. sot. chim. bid. 28, 234 (1946). 15. HUQHES, W. L., Cold Spring Harbor Symposia Quad. Bid. 14, 79 (1950). 16. WARBURO, O., AND CHRISTIAN, W., B&hem. 2.310, 384 (1942). 17. KUBOWITZ, F., AND OTT, P., Biochem. 2. 314, 94 (1943). 18. HUQWES, W. L., J. Am. Chem. Sot. 69, 1836 (1947). 19. GREENSTEIN, J. P., J. Bid. Chem. 126, 501 (1938). ‘20. CSONKA, F. A., DENTON, C. A., AND RINQEL, S. J., J. Biol. Cbm. 169,259 (1947). 21. MIRSKY, A. E., J. Gen. Physiol. 24, 709 (1941). 22. BENESCH, R., AND BENESCH, R. E., Arch. Biochem. 19, 35 (1948). 23. LARSON, B. L., AND JENNESS, R., J. Dairy Sci. 33, 890 (1950).