Elongation of the disulfide bonds of bovine pancreatic ribonuclease and the effect of the modification on the properties of the enzyme.

Abstract A method is described for the intramolecular pairing of the sulfhydryl groups in fully reduced bovine pancreatic ribonuclease (RNase) by reaction with mercuric ions and formation of S-Hg-S cross-links. The protein derivative obtained contains 4 mercury atoms per protein molecule, sediments in the ultracentrifuge as a single symmetrical peak (s20,w = 1.81), and migrates as a single sharp band on acrylamide gel electrophoresis. The circular dichroism spectrum of the mercury derivative is consistent with the supposition that the tyrosine residue which has been found to be normalized in the RNase-mercury derivative, is tyrosine-92, although other interpretations are not excluded. The RNase-mercury derivative retains about 5% and 25% of the specific hydrolytic activity of the native enzyme towards RNA and cytidine 2',3'-cyclic monophosphate, respectively. Unlike the native protein it is digested by trypsin, although at a lower rate than oxidized RNase, with the consequence of complete loss of the enzymic activity. Both tryptic digestion and the concomitant loss of catalytic activity of the mercury derivative follow first order kinetics (trypsin concentration 10 µg per ml; initial protein concentration 0.1 mg per ml) with similar rate constants. These findings indicate that the RNase-mercury derivative is homogeneous and free of contamination by the native enzyme. It is concluded that the cysteine residues are predominantly paired in the RNase-mercury derivative in the same pattern as the cystines in native RNase, and that elongation of the internal cross-links in RNase by 3 A each does not abolish, but quantitatively modifies, the activity of the enzyme.


Elongation
of the Disulfide Bonds of Bovine Pancreatic Ribonuclease and the Effect of the Modification on the Properties of the Enzyme* (Received for publication, September 10, 1970) RUTH SPERLING AND IZCHAK Z. STEINBERG From the Department of Chemical Physics, The Weizmann Institute of Science, Rehovot, Israel SUMMARY A method is described for the intramolecular pairing of the sulfhydryl groups in fully reduced bovine pancreatic ribonuclease (RNase) by reaction with mercuric ions and formation of S-Hg-S cross-links.
The protein derivative obtained contains 4 mercury atoms per protein molecule, sediments in the ultracentrifuge as a single symmetrical peak (~~0,~ = 1.81), and migrates as a single sharp band on acrylamide gel electrophoresis.
The circular dichroism spectrum of the mercury derivative is consistent with the supposition that the tyrosine residue which has been found to be normalized in the RNase-mercury derivative, is tyrosine-92, although other interpretations are not excluded. The RNase-mercury derivative retains about 5 % and 25 % of the specific hydrolytic activity of the native enzyme towards RNA and cytidine 2',3'-cyclic monophosphate, respectively.
Unlike the native protein it is digested by trypsin, although at a lower rate than oxidized RNase, with the consequence of complete loss of the enzymic activity.
Both tryptic digestion and the concomitant loss of catalytic activity of the mercury derivative follow first order kinetics (trypsin concentration 10 pg per ml; initial protein concentration 0.1 mg per ml) with similar rate constants. These findings indicate that the RNase-mercury derivative is homogeneous and free of contamination by the native enzyme. It is concluded that the cysteine residues are predominantly paired in the RNase-mercury derivative in the same pattern as the cystines in native RNase, and that elongation of the internal cross-links in RNase by 3 A each does not abolish, but quantitatively modifies, the activity of the enzyme. While disuliide internal cross-links play a decisive role in maintaining the three-dimensional conformation of proteins which possess such links (l), the requirements as to the number of essential cross-links are not necessarily stringent. Thus biologically active derivatives of RNase (2,3), lysozyme (4, 5), trypsin, trypsinogen (6), bovine trypsin inhibitor (7), and * This research was supported in part by United States Public Health Service Grant GM13637.
immunoglobulin G (8) in which some of the disulfide bridges have been cleaved could be prepared. Less is known about the de-, mands on the length and geometry of the internal cross-links for maintaining the biologically active conformation of proteins. By the introduction of a mercury atom between the sulfur atoms of a disulfide bond, a single bridge was lengthened in RNase (bond IV-V) (2), and in papain (bond 43-152) (9), without affecting the properties of these enzymes.
In a previous communication (10) we have shown that the 8 sulfhydryl groups in fully reduced RNase react quantitatively with 4 mercuric ions, resulting in the formation of -S-Hg-Sbonds, which are known to be linear (11). All four internal bridges are thus lengthened by about 3 A (12)(13)(14). When prepared under the proper conditions, a protein derivative, [RNase . 4Hg], could be obtained by this reaction which was monomeric and which showed some resemblance to the native protein. Thus, [RNase.4Hg] possessed 2 tyrosine residues which titrated abnormally with alkali and it cross-reacted with antibodies to native RNase. In the following it is shown that [RNase .4Hg] is a homogeneous material, and further details concerning the properties of [RNasee4Hg] are presented.

EXPERIMENTAL PROCEDURE
Materials-Bovine pancreatic ribonuclease-A (five times crystallized, type lA, Lot 95B-0330), cytidine 2',3'-cyclic monophosphate (Lot 94&0560), and p-hydroxymercuribenzoate sodium salt (Lot 64B-5170) were purchased from Sigma. Trypsin (lyophilized, two times crystallized, Lot TRL 6295) was obtained from Worthington Biochemicals. Water-insoluble polytyrosyl trypsin bound to a synthetic diazotizable resin S-MDA (15) was kindly donated by Dr. L. Goldstein. Ribonucleic acid (yeast nucleic acid Lot 6502) was purchased from Schwarz Bio-Research. Mercuric chloride labeled with 203Hg (catalog No. D-l) was obtained from the Israel Atomic Energy Commission (Yavne, Israel). Analytical reagent grade urea, obtained from BDH, Ltd. (Poole, England), was recrystallized from 95% ethanol; solutions of this compound were prepared immediately before use. All other chemicals were of analytical grade.
Methods-In the preparation of [RNasea4Hg], RNase was reduced as described (16). A solution of the reduced protein (7 to 10 X 1CF M in 0.1 M aqueous acetic acid) and a solution of HgCls (5.6 to 8 x lO-+ M) were forced in parallel and in stoichiometric proportions (molar ratio of RNase-HgCls, 1:4) into At the beginning of the experiment the reaction vessel contained 3 to 4 X 1W M of p-mercuribensoate.
During the reaction, which was run at room temperature, 20", a slow stream of nitrogen gas was passed through the reaction vessel and the pH was frequently adjusted manually to pH 4.6 by addition of NaOH solution (1 M) from a third micrometric syringe burette. The reaction was stopped after addition of 5 ml of protein solution to the reaction mixture in a period of 50 mm.
(This procedure ensured that the reaction between RNase and HgCls took place continuously at low concentration of reactants; the p-mercuribenzoate apparently facilitated interchanges between various sulfur pairs linked by the mercuric ions (lo).) At the end of the reaction the solution was 1 The reaction was followed in a pH-stat. The NaOH uptake is equivalent to the amount of product of the ensymic hydrolysis, p. For a first order reaction, p is given by pt = pl,,(l -edktk'), where pl,, is amount of product at t = m. : To obtain a linear plot we made use of the following relations pt+A; = p&,(1 -eekcttAt)); pt = p&, (1 -e+') %+Ar -Pt = ?bm e -k t (1 _ e-kA') kt lodPt+At -Pt) = -2.303 + log C', where c' is a constant.
Plotting log (pl+Al -pt) versus t yields a straight line for first order kinetics, without the necessity of obtaining p,,,. In Fig. 1 At was chosen to be 2 min. adjusted to pH 7.0 by NaOH solution and precipitated aggregates were removed by centrifugation for 15 min at 8CW rpm in a Sorvall centrifuge (model SS-1). The protein in the supernatant was found to be monomeric by sedimentation velocity experiments, with a sedimentation coefficient Szs+, of 1.81. The yields were typically about 40%. [RNasee4Hg] prepared at pH 4.6 was stable at this pH and could be kept in the cold for a few days.
However, after adjustment of the pH to 7.0 the monomeric derivative became somewhat unstable and tended to precipitate. It was therefore freshly prepared every time before each of the experiments described below.
[RNase.4Hg] was also prepared by the above procedure at pH 8.0, using 0.1 M T&chloride buffer (pH 8.0) as solvent instead of the acetate buffer used above.
The yield of monomeric mercurated protein at pH 8.0 (-20%) was, however, lower than that at pH 4.6; [RNasea4Hg] was therefore prepared routinely at pH 4.6. The material prepared at pH 8.0 was checked only for enzymic activity. The electrodes were the G2222B glass electrode and the K4112 calomel electrode (Radiometer).
The kinetics of the digestion of RNase and its derivatives by trypsin was followed in a pH-stat.
The concentration of the RNase derivatives was 0.2 mg per ml and that of trypsin was 10 pg per ml.
The digestion was performed at pH 8.5, 30", with constant stirring under nitrogen atmosphere. The reaction was followed by the uptake of alkali (NaOH, 10T3 M) at constant pH by means of the pH-stat.
Circular dichroism measurements were carried out in a Cary model 6001 spectropolarimeter.
Cells of 5 cm path length were used with protein concentration of approximately 0.02%. Circular dichroism data are presented as Ae, the difference between the molar absorption coefficients of left-and right-handed circularly polarized light.
Signal to noise ratios at band maxima were 15 :l and 50: 1 in the cases of [RNase .4Hg] and native RNase, respectively.

RESULTS
In order to confirm that the RNase derivative obtained contains stoichiometric amounts of mercury, the above procedure of preparation was repeated with [203HgJHgCls.
The product obtained was dialyzed against 0.1 M sodium acetate buffer, pH 4.6, for several hours to remove any free mercuric salt in solution.
An aliquot of the protein solution was then removed, its protein content was determined spectrophotometrically, and its The spectra are presented as A.E = EZ -+, where EZ and 6 are the molar extinction coefficients for left and right circularly polarized light,, respectively.
(The molecular weight of RNase is taken as 13,683, the formula weight.) than that of oxidized RNase.
Native RNase is not digested by trypsin under the same conditions.
The course of loss of the enzymic activity of [RNase .4Hg], toward the substrate cytidine 2', 3'-cyclic monophosphate, upon digestion of the modified RNase by trypsin is presented in Fig. 1. The activity at any instant was taken as the slope of the curve which described the optical density at 290 rnp as a function of time.
The rate of inactivation of [RNase .4Hg] closely follows first order kinetics.
When equal amounts of trypsin were used for digestion, the rate constants of inactivation (i.e. the slopes of the plots of the logarithm of the activity VISXLS time) were comparable with the rate constant of the tryptic digestion of the mercurated protein.
When half the concentration of trypsin was used for digestion of [RNase .4Hg], the slopes of inactivation were reduced by about 50% (Fig. 1).
IRNasee4Hgl was also digested by insoluble trypsin with the result of complete loss of enzymic activity of the modified RNase. Under the same conditions, oxidized RNase is digested about' twice as fast as [RNase.4Hg], whereas native RNase is not digested at all by the insoluble trypsin. Fig. 2 presents the circular dichroism spectrum of [RNase .4Hg] in the 250-to 300-rnp region.
For comparison, the spectra of native RNase and oxidized RNase are included.
As seen from this figure, the circular dichroism spectrum of [RNase*4Hg] in this region is much less intense than that of RNase.

Reduced
RNase reacts quantitatively with 4 mercuric ions. Under selected conditions of reaction, a monomeric product is obtained which retains, within limits, some of the properties of the native enzyme.
Thus, 2 of its tyrosine residues titrate abnormally (10) (compared with 3 tyrosines in the native protein (23, 24)), it cross-reacts with anti-RNase (lo), and retains an appreciable fraction of the enzymic activity. Contrary to the native enzyme, [RNase.4Hg] is digested by trypsin, although at a lower rate than oxidized RNase.
[RNase.4Hg] migrates electrophoretically on acrylamide gel as a single band at a rate that may be distinguished from that of RNase.
The ratio of the specific activities of [RNase.4Hg] toward RNA and toward cytidine 2',3'-cyclic phosphate is markedly different from that of the native enzyme. Furthermore, in contradistinction to RNase, [RNase .4Hg] is digested by trypsin with concomitant complete loss of enzymic activity. We may therefore dismiss the possibility that [RNase .4Hg] contains any amount of native RNase.
[RNase .4Hg] is a homogeneous preparation by several criteria. It sediments in the ultracentrifuge as a single symmetrical peak. As mentioned above, it yields a single band on acrylamide gel electrophoresis.
It is digested by trypsin according to strict first order kinetics with a single rate constant. Last but not least, the loss in enzymic activity of [RNase .4Hg] proceeds closely in parallel with the extent of its digestion by trypsin (see Fig. 1 oxygen, one obtains active enzyme in the pH range of 6.2 to 8.5, but no active material is obtained at pH 4.6 (25). It should be noted that at room temperature the native conformation of RNase is still stable at pH 4.6 and starts to melt out only below pH 3 (26) ; one may therefore expect proper pairing of the cysteine residues of reduced RNase also at pH 4.6, as has actually been found in this work when the pairing is brought about by mercuric ions in the presence of p-mercuribenzoate. The failure to do so when the pairing is brought about by oxidation is therefore probably due to inhibition of the oxidation reaction of sulfhydryl groups at acid pH (27).
The 275-rnp circular dichroism band of RNase has been studied by various authors and some controversy has arisen as to its origin. The band was attributed to the normal tyrosines of the protein (28), the abnormal tyrosines (29, 30), especially to tyrosine-92 (29), or to electronic transitions involving both cystines and tyrosines (31, 32). The 275 rnp circular dichroism band of [RNase .4Hg] is much less intense than the corresponding band of the native enzyme. Simpson and Vallee (29) have found by optical rotary dispersion measurements marked diminution of the intensity of the Cotton effect at 275 rnE.1 upon normalization of the abnormal tyrosine-92. In this connection it may be recalled that one of the abnormal tyrosines of RNase has been normalized in [RNase .4Hg] (10). Our circular dichroism data may thus be rationalized if we assumed that the tyrosine residue which has been normalized in @Nase -4Hg] is tyrosine-92. Indeed, inspection of the model of RNase (33) shows that this tyrosine residue, which is hydrogen bonded to aspartate-38 (34), is a good candidate for normalization upon conversion of the S-S bond of cystine II-VII into an S-Hg-S bond, which forces Residues 38 and 92 slightly apart. On the other hand, one cannot exclude the possibility of a contribution of the S-Hg-S bonds in [RNases4Hg] to the 275 rnE.1 circular dichroism band of this protein derivative, as evidenced from circular dichroism studies of low molecular weight model compounds. 2 The enzymic activity of [RNases4Hg] is about 5% and 25% (compared with native RNase) when tested on RNA and cytidine 2',3'-cyclic monophosphate, respectively.
It is pertinent to note that a change in the relative reactivity toward high and low molecular weight substrates had also been reported before for an RNase derivative in which the two disulfide bonds, IV-V and III-VIII, had been split by phosphorothioic acid (3). On the other hand, no change in enzymic activity had been detected when only one crosslink, the disulfide bond IV-V, had been lengthened by the introduction of a mercury atom into the disulfide bond IV-V (2). It may thus be concluded that lengthening the cystine crosslinks I-VI, II-VII, and III-VIII in RNase by about 3 A each, does not abolish the enzymic activity of the enzyme, although it changes it quantitatively. 2 R. Sperling and I. Z. Steinberg, in preparation. , , ,