Reprinted with Permission by the
U. S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE
Public Health Service

THE JOURNAL oF BIOLOGICAL CHEMISTRY
Vol. 236, No. 5, May 1961
Printed in U.S.A.

Studies on the Reduction and Re-formation
of Protein Disulfide Bonds

CHRISTIAN B. ANFINSEN AND EpGar HABER

From the Laboratory of Cellular Physiology and Metabolism, National Heart Institute, National Institutes of Health,
Public Health Service, United States Department of Health, Education and Welfare, Bethesda, Maryland

(Received for publication, December 29, 1960)

Previous communications have reported on the reduction of
the disulfide bonds of ribonuclease (1-4), ribonuclease-S-protein
(5), and lysozyme (6) by treatment of the native protein or
protein derivative in 8 mM urea solutions with thioglycolic acid or
mercaptoethanol. Reduction by this procedure has a signifi-
cant advantage over certain other techniques (e.g. borohydride
(7) or sulfite reduction (8)) in that it yields reduced molecules
that have suffered minimal, if any, covalent change other than
that involved in the cleavage of the disulfide bonds to sulfhydryl
groups. Thus, it has been observed’ that fully reduced ribonu-
clease contains no new NH--terminal amino acids, and that its
molecular weight and amino acid content (after alkylation of
the SH groups) are in accord with the expected values. It is
therefore possible to study the re-formation of disulfide bonds
and to examine the influence of chemical and environmental
modifications on the efficiency of reoxidation and reactivation.
It has been demonstrated that the eight sulfhydryl groups of
fully reduced ribonuclease and of ribonuclease-S-protein undergo
spontaneous oxidation to yield macromolecules that are ex-
tremely similar, and probably identical, to the parent molecule
in their physical and enzymatic properties (4, 5).

The methods of reduction and subsequent manipulation that
we have reported have, over the past year, been somewhat modi-
fied. We would like to summarize here some of the aspects of
reduction, sulfhydryl group stabilization, and alkylation that
appear to be of general use in the reduction of proteins, as well
as some observations on conditions for regeneration of disulfide
bonds in reduced ribonuclease.

EXPERIMENTAL PROCEDURE AND RESULTS

Reduction—In a typical experiment, 350 mg of native RNase
(Sigma Chemical Company, lot R60-B-204, chromatographic
grade) were dissolved in 10 ml of a freshly prepared 8 m solution
of recrystallized urea, adjusted to pH 8.6 with 5% methylamine.
Mercaptoethanol (Eastman-Kodak) was added at a level of 1 yl
per mg of protein, the container was flushed with nitrogen, and
the solution was allowed to stand for 43 hours at room tempera-
ture. Methylamine was employed for neutralization to insure
the decomposition of any thioglycollides (3) that might be pres-
ent in the reducing agent. After this period, the pH was ad-
justed to 3.5 with glacial acetic acid and the entire solution was
applied to a 2.5- x 35-cm column of crossed-linked dextran (9)
(Sephadex G-25, Pharmacia Company, Uppsala, Sweden) which
had been thoroughly equilibrated with 0.1 m acetic acid. The
column was developed with the same solvent. As shown in
Fig. 1, the bulk of the protein emerged considerably ahead of the

reagents in the original mixture. (In experiments using longer
columns or smaller applied volumes of reaction mixture, the
mercaptoethanol, urea, and salts appear in a fraction entirely
separated from the protein peak by approximately 20 to 25 ml
of effluent volume; see, for example, Fig. 2.) Titrations with
PCMB! (10) and estimations of SH content by radioactivity
determinations of C%-iodoacetate-treated material (5) from two
portions of the major part of the peak gave the same values (see
legend to Fig. 1). In samples from the trailing portion of the
peak, determinations with PCMB gave high results because of
slight contamination with mercaptoethanol. Values obtained
by the radioactive method on these latter samples were not so
affected, since the derivatives were submitted, after radioactive
alkylation, to a second passage through Sephadex G-25 before
comparison of their specific radioactivities with that of the C*-
iodoacetate reagent.

The method of reduction and purification described above has
also been applied to trypsinogen, chymotrypsinogen, and lyso-
zyme. Fully reduced products that were soluble in the dilute
acetic acid solvent were obtained from each of these proteins.

Reoxidatton—A reinvestigation of the methods used for the
oxidation of reduced RNase was undertaken in an attempt to
increase the yields, of regenerated native enzyme and, particu-
larly, to decrease the production of insoluble side products.
High yields of enzymatically active material and the absence of
precipitation are achieved only when oxidation is carried out at
low protein concentrations (Table I) and when conditions ‘of
surface denaturation (e.g. bubbling with air) are avoided. As
shown in Table I, efficient oxidation occurs when dilute solutions,
adjusted to pH 8.0 to 8.5, are allowed simply to stand in open
vessels at room temperature for approximately 20 hours. Under
these conditions, the yields of regenerated, active enzyme have
been uniformly between 80 and 100%.

Conversion of SH groups in reduced RNase to disulfide bonds
is extremely slow at the hydrogen ion concentration of the acetic
acid solution employed for the Sephadex columns. At room
temperature and pH 3.0, a sample of protein containing 7.8 SH
groups per mole contained, after 24, 72, and 96 hours of exposure
to atmospheric oxygen, respectively, 7.5, 6.9, and 4.5 SH groups
per mole. At icebox temperatures, the SH group content is
undiminished for 2 to 3 days.

Stabilization of Protein SH Groups after Reduction—In studies
of the sulfhydryl-disulfide relationships in proteins, SH groups
are generally stabilized by reaction with suitable alkylating re-

1 The abbreviation used is: PCMB, p-chloromercuribenzoate.

1361
1362

 

T T T I T T T | r | T

nun WD WA
wm oa wn a
| I ] ]

a
T

OPTICAL DENSITY AT 275 Mu
_ Mw
Oo oO
T |

un
l

 

 

 

0 | | ! | L | | ! |
0 4 +B {2 I6 20 24 28 32 36 40 44

TUBE NUMBER (5.0 ML./ TUBE)

Fic. 1. The separation of reduced ribonuclease from reagents.
After reduction, the solution of reduced protein was adjusted to
pH 3.5 with glacial acetic acid and applied, in a volume of 12 ml,
to a 2.5- X 25-em column of Sephadex G-25 which had previously
been equilibrated with 0.1 m acetic acid. Samples were taken
from tubes 12, 15, 18, and 20 for estimation of SH content by ti-
tration with PCMB (10) and by determination of ‘radioactivity
after alkylation with iodoacetic acid-1-C1 (5) (see the text). The
latter method yielded 7.5, 7.4, 7.7, and 7.7 moles of SH per mole of
RNase whereas the PCMB titrations gave 7.6, 7.7, 11.5, and 20
moles of SH per mole of protein for the same samples. This
difference in the results obtained by the two methods indicates
the presence, in the two tubes from the trailing portion of the
protein peak, of mercaptoethanol that was not fully resolved from
the reduced protein on this particular column.

TasLz I

Oxidation of reduced RNase at various protein concentrations

The reduced protein was prepared as described in the text and
freed from reagents on Sephadex G-25. The protein in the column
effluent, after appropriate dilution, was adjusted to pH 8.5 and
allowed to stand without stirring at room temperature for 20
hours, at which time the proportion of soluble protein and the
enzyme activity of this fraction were estimated.

 

 

 

 

Activity of soluble
Concentration of | % Yield of soluble |fraction; % of equiva-| % Regeneration of
reduced protein protein lent concentration of activity
native RNase

mg/ml |

7.0 27 31 8

4.8 42 70 29

2.3 87 75 65

0.9 100 77 77

0.35 100 94 94

 

 

agents such as iodoacetic acid, iodoacetamide, or N-ethylmaleim-
ide. In spite of care in the maintenance of proper pH values
and times of reaction, experience in various laboratories has
indicated that alkylation of groups other than SH graups may
occur in small but significant amounts (11) (e.g. methionine sulfur
to the sulfonium derivative, lysine e-amino groups to the N-alky]
derivative). For this reason we have recently employed an ex-
cess of mercaptoethanol to immobilize the iodoacetate at the
end of the alkylating period.

The reduction mixture is diluted before alkylation with 0.5 u
Tris-acetate buffer, pH 8.5, to reduce the urea concentration to

Studies on Protein Disulfide Bonds

Vol. 236, No. 5

approximately 2m. At this level of urea, nonspecific alkylations
are reduced to a level of 0.1 to 0.2 mole of alkyl groups per mole
of protein, as estimated with C'-iodoacetate (5). The diluted,
reduced protein is then treated with a 10-fold excess of iodoacetic
acid (calculated on the basis of the mercaptoethanol present in
the reduction mixture) at room temperature for 12 to 15 minutes.
The presence of buffer obviates the necessity of constant monitoy-
ing of pH during the alkylation. At the end of the reaction, a
10-fold excess of mercaptoethanol is added (calculated on the
basis of the added icdoacetic acid). The mixture is then allowed
to stand for 30 to 60 minutes to permit complete reaction of free
iodoacetate and is finally passed through Sephadex G-25 to
yield the reagent-free, salt-free protein derivative. This material
is now ready for radioactive counting for the determination of
S-carboxymethy] group content. Counting is most conveniently
done in a liquid scintillation counter, the protein solution being
mixed directly with a water-miscible phosphor solution (12).
Specific radioactivities have generally been calculated on the
basis of protein concentrations estimated by spectrophotometric
determinations.

To stabilize the SH groups in a reversible form, the reduced
protein emerging from the Sephadex column is adjusted to ap-
proximately 5 mg per ml with 0.1 m acetic acid and treated with
a 1.1-fold excess of sodium PCMB (baséd on the moles of SH
groups), dissolved in water, and adjusted to pH 8. Although
PCMB is quite insoluble in 0.1 m acetic acid, the added material
reacts rapidly with the free SH groups of the protein derivative,

 

 

 

 

 

T T T T
a
E LSr 4 5
@ z
“ 4
LOR 410 8
z >
o “4
S "
= 05+ 405 oO
x<
3
a =
I 1 i
25 50 75 100

EFFLUENT VOLUME (ml)

Fia. 2. Regeneration of reduced RNase from its compound
with PCMB. The PCMB compound (110 mg) was dissolved in
10 ml of 0.1 m acetic acid by the addition of 100 ul of mercapto-
ethanol. After standing for 30 minutes, the solution was passed
through a 2.5- & 30-cm column of Sephadex G-25 previously equi-
librated with 0.1 M acetic acid. The column was developed with
the same solvent, and fractions of 2.5 ml each were collected. The
odor of mercaptoethanol was first detected in the vicinity of 65
ml effluent volume. The tubes indicated in the figure were pooled
and the content of SH groups per mole of RNase was determined
by titration with PCMB (10). The value obtained (7.4 SH groups
per mole) agrees well with that found for the original reduced
RNase (7.5 SH groups per mole) which served as starting material
for the preparation of the PCMB derivative. The recovery of
absorbancy units (measured at 280 my in the pooled tubes) was
65%. Absorbancy was measured at 280 mu (O) to detect the
reduced protein, and at 255 mu (@) to detect the emergence of
mercaptoethanol and its PCMB derivative.
May 1961

and marked precipitation is first observed only when the point
of equivalence has been reached. The final mixture may be
dialyzed against 0.1 m phosphate buffer at pH 8 to remove excess
PCMB and, after further dialysis against water, the suspension
of PCMB-RNase is lyophilized for storage. The derivative (re-
duced RNase-8C;H,O.Hg; calculated Hg content , 9.9%;
found 9.1% (13), 10.4% (14)) is somewhat soluble in 0.1 m
phosphate buffer, pH 8.0 (approximately 0.5 mg per ml), and
very soluble in strong urea solutions. It may be stored in the
refrigerator for at least 2 months without detectable loss in its
ability to regenerate fully reduced RNase after removal of the
PCMB groups. Such removal is accomplished by suspending
the derivative in 0.1 m acetic acid (10 to 20 mg per ml), adding
1 ul of mercaptoethanol per mg of derivative (whereupon the
solution becomes clear), and passing the solution through the
Sephadex column as described above (Fig. 2). The resulting
reagent-free protein contains the sulfhydryl group content of the
original reduced material and may be converted to active enzyme
in high yield by oxidation.

The availability of the PCMB derivative of reduced RNase
and its ready oxidation to active enzyme after removal of the
masking groups make possible a study of the ability of the re-
duced chain to re-form enzymatically active material after modi-
fication by limited selective cleavage with proteolytic enzymes.
Experiments with the Jacobsem-Leonis autotitrator (15) (pH-
Stat) indicate that the PCMB derivative of reduced RNase is
digested with trypsin to an extent equivalent to that of reduced
carboxymethylated material (6). Thus, the number of peptide
bonds cleaved is that expected from the lysine and arginine
content of the chain (t.e. 12 bonds cleaved per mole, with two
bonds unreactive, the NH--terminal lysy] bond and the lysyl
bond in the sequence, Lys. Pro).

DISCUSSION

The main point to be emphasized in connection with the meth-
ods described above is the relatively gentle nature of the proce-
dures for reduction and for subsequent separation on columns
of the reduced protein derivative from reagents. The high
yield of active regenerated RNase (Table I) obtained when oxi-
dations of SH groups are carried out at low concentrations of
reduced protein is in strong contrast to the rather poor yields
obtained after reduction by a reagent such as sodium borohydride
(16) which is known to cause limited, but probably highly critical
cleavage of peptide bonds (17). Perhaps the best evidence for
the mildness of the methods comes from experiments, to be de-
scribed in detail in a separate communication, in which RNase
was subjected to two successive rounds of reduction and air
oxidation without significant loss of activity.

The stability of the bulk of the covalent structure of RNase
to conditions of reduction, combined with the ability of the re-

C. B. Anfinsen and E. Haber

1363

duced molecule to undergo spontaneous oxidation and re-forma-
tion of secondary and tertiary structure, permits an approach
to the preparation of fragments of the RNase chain that might
recombine, through disulfide bonds, to yield enzymatically active
compounds.

SUMMARY

Methods are described for the reduction of disulfide bonds in
ribonuclease and other proteins, for the separation of the reduced
derivatives on columns of cross-linked dextran (Sephadex), and
for the convenient carboxymethylation of SH groups produced.
The reductive technique employed does not appear to alter other
covalent features of ribonuclease, as evidenced by the oxidizabil-
ity of the reduced chain to yield a product with native enzymatic
and physical characteristics. Conditions are described for opti-
mal regeneration of activity by exposure of dilute solutions to
molecular oxygen. The preparation of a p-mercuribenzoic acid
derivative of the reduced protein is described. This material
serves as a stable form of the fully reduced chain for degradation
with proteolytic enzymes.

REFERENCES

1. Sera, M., Wuirs, F. H., JR., anD ANFINSEN, C. B., Biochim.
et Biophys. Acta, 81, 417 (1959).

. Waite, F. H., Jn., anp Anrinsen, C. B., Ann. N. Y. Acad.
Set., 81, 515 (1959).

. Waite, F. H., Jr., J. Biol. Chem., 236, 383 (1960).

- Waite, F. H., Jn., J. Biol. Chem., 286, 1353 (1961).

. Hapur, E., anp ANFINSEN, C. B., J. Biol. Chem., 236, 422
(1961).

. ANFINSEN, C. B., in A. NEuBERGER (Editor), Symposium on
protein structure, John Wiley and Sons, Inc., New York,
1958, p. 223. .

7. Moons, 8., Coz, R. D., Gunpiacu, H. G., anp Sten, W. H.,
Fourth international congress of biochemistry, 1958, Vienna,
Symposium VIII, No. 11, Pergamon Press, New York, 1959.

8. Swan, J. M., Nature (London), 180, 643 (1957).

9. Poratu, J., Biochim. et Biophys. Acta, 39, 193 (1960).

10. Boysr, P., J. Am. Chem. Soc., 76, 4331 (1954).

11. Gunpiacu, H. G., Sretn, W. H., anp Moors, 8., J. Biol.
Chem., 284, 1754 (1959).

12, Bray, G. A., Anal. Biochem., 1, 279 (1960).

13. Laue, E. P., anp Neuson, K. W., J. Assoc. Offic. Agr. Chem-
tsts, 25, 399 (1942), as quoted in E. B. Sanpeuu (Editor),
Colerimetric determination of traces of metals, Interscience
Publishers, Ine., New York, 1944.

14. Oso, A., AND Farrar, G. E. (Editors), The dispensatory of
the United States of America, 25th edition, J. B. Lippincott,
Company, Philadelphia, 1955, p. 824.

15. Jacopssn, C. F., anp Leonts, J., Compt. rend. trav. lab. Carls-
berg, Sér. chim., 27, 333 (1951). :

16. WetLaursrR, D. B., Abstr. 185th meeting of Am. Chem. Soc.,
Washington, 1959, p. 270.

17. Kimmy, J. R., anp Parceuis, A. J., Federation Proc., 19, 217
(1960).

Gum oO be

&