INTERACTION OF COPPER(II) WITH BOVINE SERUM ALBUMIN
Nov. 5 , 1958
Thionine absorbs quite strongly and, under the proper conditions, the quantum yield for bleaching approaches unity, so that most of the quanta falling on the system cause reaction. Past this point, however, one may see two possibilities of loss. Some of the energy of the single quantum is converted into the chemical energy of the semiferric pair, but i t is not known how much of this is lost and how much is converted. (This question will be discussed more fully in the next paper of this series although a t the present concentrations one may estimate a conversion of about l5Y6of the 2 e.v. quantum.) A more likely possibility of loss may be found in the electrode process. Since the spontaneous back reaction in the body of the solu-
-
[CONTRIBUTION FROM
THE SCHOOL O F
5673
tion competes with the electrode process, i t is a bit difficult to understand how any significant amount of electrical work may be taken from the system (indeed, this may be the situation; no power measurements have been published for photogalvanic cells). A means of separating the energetic species, of course, would eliminate most of this loss. Acknowledgments.-The author wishes to thank Professor H. S. Johnston and the members of the Stanford Chemistry Department for many helpful suggestions. The research was partially supported by a du Pont Summer Faculty Research Grant. STANFORD, CALIF.
CHEMISTRY O F
THE
UNIVERSITY
OF
MINNESOTA]
The Interaction of Copper(11) with Bovine Serum Albumin1 BY I. M. KOLTHOFF AND B. R. WILLEFORD, JR.~ RECEIVED APRIL 28, 1958 Native bovine serum albumin (BSA) can be titrated amperometrically with copper(I1) in an ammonia buffer using the rotated platinum wire or rotated dropping mercury electrode as indicator electrodes. At the end-point, one mole of copper is bound per mole of albumin. The reactive group is not sulfhydryl. Nickel(I1) forms a stable complex under the same experimental conditions as copper(I1) does. BSA denatured in 4 M guanidine hydrochloride contains the same reactive group as in the native state. However, an excess of copper oxidizes the sulfhydryl to disulfide whereas this reaction does not occur with native BSA. Sulfhydryl in native BSA can be oxidized by oxygen when one mole or more copper(I1) per mole BSA is present. The titration with copper is proposed as a rapid, specific method for the determination of albumin in blood serum. have been described previously.6 ?-Globulin was a sample obtained from Professor John T. Edsall of the Harvard University Medical School. Conductivity water was used in the preparation of all solutions. Metal ion solutions were prepared from reagent grade commercial chemicals and analyzed by standard methods. Solutions were deaerated with Linde nitrogen of 99.9% purity. All other chemicals used were reagent grade. Instrumentation.-Current-voltage curves were measured manually by a circuit similar to that described by Lingane and Kolthoff' and automatically with a Leeds and Northrup type E Electrochernograph. Amperometric titrations were carried out using the manual circuit. The platinum electrode was rotated a t a constant speed of 600 or 900 r.p.m. by a Bodine synchronous motor. The rotated dropping mercury electrode was of the type A described by Stricks and Kolthoffs; a rotation speed of 300 r.p.m. was used. All potentials were measured against the saturated calomel electrode. Measurements of pH were made with a Beckman Model H-2 pH meter. Titration Procedure.-The titration vessels were 125-ml. beakers fitted with rubber stoppers with holes for the indicator electrode, nitrogen inlet and outlet, salt bridge and buret. The proper quantity of buffer (usually 25 ml. for R.p.e. titrations, 60 ml. for R.d.m.e) was placed in the titraMaterials .-The bovine serum albumin was obtained tion cell and deaerated with a stream of nitrogen which had from Armour Laboratories, and the guanidine hydrochloride been passed previously through buffers of the same compo(GHC1) from Eastman. Details of the properties of the The pH of the solution after completion of the tiBSA, preparation of solutions and purification of the GHCl sition. tration was measured frequently and in no case was there any significant change. When deaeration was complete, (1) A preliminary report of this work appeared in TEIISJOURNAL, the BSA solution (air-free) was introduced and the flow of 79, 2656 (1957). nitrogen continued for several minutes. The native albu(2) On leave from Bucknell University, Lewisburg, Pennsylvania. min solutions foamed considerably, and care was taken to (3) I. M. Kolthoff and W. Stricks, THIS JOURNAL, 73, 1728 (1951); avoid any loss of protein through the nitrogen outlet tube. . 4 n d Chem., 23, 763 (1951). On the other hand, the denatured protein solutions showed (4) I. M . Klotz and H. G. Curme, THISJOURNAL, 70, 939 (1948): very little tendency to foam. In current-voltage curve H. A. Fiess and I. h'f. Klotz, ibid , 7 4 , 887 (1952); 1. M. Klotz, J. M. measurements, the stream of nitrogen was diverted to pass Urquhart and H. A. Fiess, ibid., 74, 5537 (1952); I. M. Klotz, J. M. over the surface of the solution, while in amperometric ti-
I n work carried out in this Laboratory, i t has been found that the sulfhydryl group in cysteine (RSH) is oxidized by copper(I1) in ammoniacal medium to cystine (RSSR)3according to the over2Cu(") RSSR 2Cu('). all reaction 2RSI n the presence of sulfite ion, the reaction is ~ C U ( " ) RSSOs" -+ ~ C U ( ' ) RSSOs-. I t was of interest to us to determine whether similar reactions occur with the sulfhydryl group in native and denatured bovine serum albumin (BSA). The copper complexes of a number of proteins including BSA have been investigated by Klotz, et al.4 In particular, i t was reported that copper is bound through the albumin sulfhydryl group under the experimental conditions chosen. The formation of complexes between cations and proteins recently has been reviewed by Gurd and Wilco~.~ Experimental
+
+ +
-
+
+
Urquhart, T. A. Klotz and J. Ayers, ibid., 71, 1919 (1955); I . M. Klotz, I. L. Faller and J. M. Urquhart, J. Phrs. Colloid Chem., 64, 18 (1950); I. M. Klotz and H. A. Fiess, i b i d . , 66, 101 (1951). ( 5 ) F. R. N. Gurd and P. E. Wilcox, Advances i n Prolein Chem., 11, 311 (1956).
( 6 ) I. Id.Kolthoff, A. Anastasi, W. Stricks, B. H. Tan and G . S . Deshmukh, THISJOURNAL, 79, 5102 (1957). (7) J. J. Lingane and I. M. Kolthoff, ibid.. 61, 825 (1939). (8) W. Stricks and I. M. Kolthoff, ibid., 78, 2085 (1956).
I. M. KOLTHOFF AND B. R. U'ILLEFORD, JR.
5674
Vol. 80
trations it was passed through the solution continuously. molar) were added from a Gilmont Titrants (10-I to ultramicroburet of 0.1-ml. total capacity.
I
I
I
Results and Discussion Current-Voltage Curves.-Figure 1 shows the current-voltage curves of copper(I1) a t the R.p.e. in a buffer 0.1 M in ammonia and 0.1 M in amI
25
20
,
I
=I
I
I
-0.2 -0.4 -0.6 -0.8
0
3
5
-1.0
Applied Potential, volts.
+
c
Fig. 2.-Current-voltage curves of Cu(I1); R.p.e. at 900 r.p.m.: A, 0.1 M NHa 0.1 M NHdNOa f 4 M KC1, PH 9.3,4 X 10" M Cu("); B, 0.1 iM NH3 0.1 d.I NH4N03 4 M GHCI, PH 9.0, 4 X M CU("); C, 0.1 ~l!f NH3 0.1 M NH4NOa 4 M GHCI, pH 9.0, 9.3 X lo-' mrnole BSA, 9.6 X lo-' rnmole Cu(") (3.8 X J 4 1 ; 11, 0.1 NH8 0.1 M "&OI 4 A.I GHCI, PH 9.0, 9 3 X lo-' mmole BSA, 2.4 X mmole CU(").
?
+
IO
+
+
+
5
+
+
0 0
-0.2
-0.4 -0.6 -0.8
-1.0
A p p l i e d Potential, v o l t s .
Fig. 1.-Curreut-voltage curves at R.p.e., 930 r.p.tn. 0.1 111 X-H8 0.1 AI IVH4N03,pH 9.2: A, residual; 13, 4.0 x 10-5 M CU;"); C , 7.3 x 1 0 - 6 M BS~I; I), 7.3 x 1 0 - 5 AT RSX 6.3 x 10-5 -11 C U ~ " ) ; E, 7.3 x I O - + 31 RSA 11.0 x 10-5 lid c U ( " j .
+
+
+
+
monium nitrate (curve B). The first wave corresponds to the reduction of copper(I1) to copper(1). No copper waves are observed when BSA is present in a molar concentration greater than that of copper (curve D). Curve E illustrates the reappearance of the copper waves when an excess of copper is present. These curves indicate that amperoinetric titration of BSA with copper(I1) should be possible a t -0.3 to -0.4 volt. Titration a t this potential produces copper(1) a i d coating of the electrode surface with a metallic film does not occur. Curve A in Fig. 2 shows the current-voltage curve of copper(I1) in the presence of 4 dd pctassium chloride. It can be seen that the copper(J1) wave is decreased in height and is spread out over a larger potential range than is the case in the absence of chloride. When guanidine liydrochloride (GHCI) is added, the copper(I1) wave height is depressed considerably more (curve R). Once again i t can be seen that addition of I3SA causes the copper waves to disappear (curve C) ; they reappear when a sufficient excess of copper(I1j is present (curve D). In all cases, amperometric titration a t - 0.4 volt appears feasible. Amperometric Titrations.-Sumerous titrations of native BSA with copper(I1) in ammoniacal buffer a t pH 9.2 indicate a rapid reaction in a mole rat.io of copper(I1) to BSA of 1.00 k 0.03. If, after addition of an excess of copper, the solution is allowed to stand, a slower reaction of the excess copper(I1) with BSA is indicated by a gradual decrease in the diffusion current. The titrations with copper(I1)
were continued after various times of standing. I n all cases, the slope of the excess reagent line was the same as the original titration line. Extrapolation to the residual current line gives a measure of the total amount of copper(I1) reacting (see Fig. 3). When 2 moles of copper(I1) per mole of BSA is present, 0.34 additional mole of copper(1
0
l
2
3
5
6
0 1 AI NIis
+
4
M o l e s iu"' p e r Mole B S A . Fig. 3.--Titratiotr of BSA with Cu(");
0 1 dd ?CHINO$,pH 9.2, lo-* minole BSA. Time of stantling a t point U, 21 hr., R.p.e. a t 900 r.p.m; A, blank; U. first titration; C, second titration.
reacts after 1 I n . and 0.57 mole after 21 hr. When four moles of ccpper(I1) is added originally, the additional copper(I1) reacting is 0.59 mole and 0.60 mole after 4 and 21 hr., respectively. I n all cases the total copper(I1) reacting seems to approach 1.GO moles per mole of BSA on long standing under the conditions of these experiments. This slow reaction is not one between copper(I1) and sulfhydryl, as is evident from the results in Table I discussed below. The results of a number of titrations of native RSA with copper(I1) followed by a second titration with silver(1) or mercury(I1) chloride are given in
Nov. 5, 1958
INTERACTION OF COPPER(II)WITH BOVINESERUM ALBUMIN
Table I. A typical titration curve is shown in Fig. 4. It can be seen that after 1 or 2 moles of copper(11) per mole of BSA has been added, 0.68 equivalent of sulfhydryl per mole of BSA is titrated with silver(1) or mercury(I1) chloride even after standing with an excess of copper for periods u p to 21 hr. The initial rapid reaction of 1 mole of copper(11) per mole of BSA is again indicated. Included also in Table I are the results of two experiments in which 0.68 mole of silver(1) or mercury(I1) chloride is added prior to titration with copper(I1); this does not prevent the reaction of the copper(I1) with the BSA. All these results indicate neither the fast nor the slow reaction of copper(I1) with BSA involves the sulfhydryl group. TABLEI MMOLE BSA IN 25 ML. BUFFER
TITRATION OF (0.1 M ;iHa 0.1 M NHdNOa, pH 9.2) with copper(I1) and silver(1) or mercury(I1) chloride. Applied potential -0.4 volt; R.p.e. a t 900 r.p.m.
+
Moles C.,l(II) added per mole
BSA
Time of 2nd standing,a Titrating min. agent, M
Reaction ratio
Cu/BSA
M/BSA
Ag(') (1.00) 0.67 Ag(') (1.00) .63 Ag(') 0.97 .67 Ag(') .. .66 HgCh .99 .70 ..d 5 Ag(') .92* ( .68) .. 5 HgC1z .98 ( .68) a Time between end of 1st titration and start of 2nd titration. Low result attributed to change in platinum electrode surface caused by plating out of silver. Titration with &(I1) after addition of 0.68 mole of Ag(') per Titration with Cu(I1)after addition of 0.68 mole of BSA. mole of HgClz per mole of BSA.
1.0 1.0 2.0 2.0 2.0
5 15 5 21 hr. 5
Effect of Oxygen.-The sulfhydryl group in native BSA is not oxidized by oxygen in ammoniacal buffer at PH 9.6 However, when solutions of BSA at PH 9 which contain an excess of copper(I1) are exposed to air, the amount of titratable sulfhydryl decreases significantly. For example, after 1, 3 and 21 hr., the sulfhydryl values were 0.41, 0.35 and 0.16, respectively. Experiments in which mmole BSA in ammoniacal buffer of PH 9 and containing varying amounts of copper(I1) was saturated with a stream of oxygen for 20-30 minutes, allowed to stand overnight in a closed vessel and then titrated with silver(1) after removal of the oxygen revealed that with 0.75 mole of copper(I1) or less per mole BSA all the sulfhydryl was titrated even after 25 hr., while no sulfhydryl was found after this period of time with 1.0 or more moles of copper per mole of albumin. I n a similar experiment with 2.0 moles of copper(I1) per mole of BSA saturated with nitrogen instead of oxygen, all the sulfhydryl was found. Since only the excess copper(I1) can promote the oxidation of the sulfhydryl groups, the previously reported observation that a trace of copper(I1) has no effect on the rate of oxidation of BSA while i t greatly accelerates the oxidation of cysteine to cystine6 is readily explained. Effect of PH and Buffer Composition.-Currentvoltage curves of copper(I1) solutions in buffers of total ammonia concentration to 2 M both in the
5675
L
0 I 2 Moles C U ' ~ ' / E S A .
I 2 M o l e s Ag"'
3 / BSA.
Fig. 4.-Titration of BSA with Cu(") followed by titrammole BS.4 in 25 ml. buffer, 0.1 hl tion by Ag('): NHa 0.1 M NHdNOs, p H 9.2, R.p.e. a t 900 r.p.m.
+
presence and absence of BSA indicate that amperometric titrations of copper(I1) a t - 0 4 volt in these media are feasible though the copper waves are somewhat ill defined a t the higher buffer concentrations. No titration in 0.01 M ammonia a t pH 8 was possible, for a curved excess reagent line was obtained. A stoichiometric reaction between copper(I1) and BSA was found in the pH range 8-10 in buffers with ammonia concentrations ranging between 0.03 and 0.5 hl and ammonium nitrate concentrations varying between 0.01 and 1M . Titrations of Denatured BSA.-As mentioned previously, current-voltage curves in the presence of 4 M GHCl (Fig. 2) indicate that amperometric titrations of BSA with copper(I1) a t -0.4 volt should be feasible even though the diffusion currents observed are not large. A number of such titrations indicate that there is a rapid reaction of 1.3 moles of copper(I1) with denatured BSA. This is followed by an indefinite slower reaction (more rapid however than the secondary reaction with native BSA) which depends upon both the amount of excess copper(I1) present and the time of standing. There is no apparent pattern to this further slow reaction. In a typical case, the first excess reagent line gave a Cu/BSA reaction ratio of 1.31. When the ratio of copper(I1) added to BSA present reached 4, the solution was allowed to stand for 20 minutes, during which time the diffusion current dropped, a t first rapidly, then more slowly, from 5.0 to 3.2 pa. More copper(I1) was then added. Extrapolation of this excess reagent line gave a Cu/BSA ratio of 2.36. Again, on standing for 20 minutes (copper(I1) added to BSA ratio of about 7) the diffusion current dropped from 8.8 to 7.8 pa. Further addition of copper(I1) yielded an excess reagent line which extrapolates to a Cu/ BSA reaction ratio of 3.05. Further uptake of ropper(I1) was indicated by continued decrease of the diffusion current. This slow but continuous copper(I1) consumption is attributed to a slow hydrolytic fission of the disulfide groups in the albumin, the sulfhydryl formed being oxidized by copper(I1). Copper(II), like silver(1) and mercury(II), promotes this hydrolytic fission.
I. 11.KOLTHOFF A N D B. R. 11-ILLEFORD, JR.
5676
A series of experiments was carried out in which varying amounts of copper(I1) were added to the protein prior to denaturation and then an amount of air-free solution of GHCl in buffer to give a final concentration of GHCl of 4 M was added. Addition of the GHCl solution caused the diffusion current of the excess copper(I1) to drop rapidly. When the diffusion cuxrent reached a steady value (usually after about 8 minutes), titration with copper(I1) was then resumed, and the excess reagent line extrapolated to the residual current line. Figure 5 shows one of these titrations. The results confirm the increase of Cu/BSX reaction ratio from 1.0 to 1.3 when the protein is denatured in 4 M GHC1.
0
Moles
2 Cu")
4
6
p e r Mole B S A .
Fig. 5.-Titration of BSAqT< it11 Cu(") before and aftcr mmole BSA; denaturation: R p e. at 900 r.p m.: A , 0.1 M h7H3at 0 1 Jf NH4h7O3;B, 0.1 If S H 3 0 1 X NHaS08 4 111GHCI.
+
+
A series of titrations was carried out in which the BSA was treated with 0.67 mole of silver(1) or mercury(I1) chloride prior to denaturation in 4 M GHCl. After five minutes denaturation, the solution was titrated with copper(I1). BSA treated with silver(1) gave normal titration lines indicating reaction of one mole of copper(I1) per mole of BSA. Titrations of the mercury(I1) chloride treated BSA were less satisfactory because the excess reagent lines were curved. However, they too indicate a drop in the Cu/BSA reaction ratio from the 1.3 observed with untreated denatured BSA to somewhat less than 1.0. The silver(1) or mercury(I1) chloride ties up the sulfhydryl group in the denatured protein and renders i t unavailable for reaction with copper(I1). Confirmatory evidence for this was obtained from an experiment in which the denatured protein was treated with ferricyanide. It has been found that the sulfhydryl group of native BS.4 is not oxidized by ferricyanide but that denaturation of the protein renders the sulfhydryl susceptible to rapid ferricyanide o ~ i d a t i o n . ~A sample of BSA in denaturing mixture was treated with ferricyanide and then titrated with copper(I1); a Cu, BSA reaction tatio of 1.01 was found. On the basis of these experiments, it is suggested that the sulfhydryl group is involved in the rapid reaction of 1.3 moles of copper(I1) per mole (9) I \ I Kolthoff a n d A Anaitaii THISTOURNAI
8 0 , 1118 (1938)
VOl.
so
of denatured albumin. The series of reactions (1), (2), (3) and (4), similar to those proposed for the oxidation of cysteine by ~ o p p e r ( I I )is , ~proposed
+ 2PSH +2Cu(I) + 2PS + 2 H C 2PS +PSSP 2Cu(I) + 2PSH -+ 2PSCu(') + 2 H C 2Cu(Ir) + 4PSH +PSSP + 2PSCu(I) + 4 H + 2Cu(Ir)
(1) (2) (3) (4)
One mole of albumin contains 0.67 mole of sulfhydryl which, according to equation 4, reacts with 0.34 mole of copper(I1). Presumably 1.0 mole of copper reacts a t the same site as in native BSA. Thus, the total copper consumption should be 1.34 moles per mole of denatured albumin, in good agreement with the average experimental value of 1.31. This interpretation is supported by viscosity measurements carried out by B. H. Tan in this Laboratory. The reduced viscosity of 1% albumin in 4 M GHCl in a buffer of p H 9 was found to be 0.19; after oxidation with 2 equivalents of ferricyanide (per sulfhydryl) or of 4 moles of copper(II), the reduced viscosity increased to 0.25 as a result of dimer formation. Effect of Other Metal Ions.-Titrations of mmole of BSA by copper(I1) in the presence of varying amounts of cobalt(II), nickel(I1) and zinc(11) ions have been carried out. Results are given in Table 11. I t can be seen that nickel(I1) and cobalt(I1) are effective in suppressing the reaction of copper(I1) with native BSA while zinc(II), even in very large excess, has practically no effect. TABLE I1 AND Zll(") ON REACTION OF cU(") WITH SATIVE BSA 25 nil. buffer (0.1 X NH3 0.1 M NH4X03, pH 9.2); mmole BSA (concn. = 4 X 10-5 M ) ; R.p.e. at 900 r.p.m., -0.4 volt DS. SCE. EFFECTO F
cO("),
Xi('')
+
Metal
Moles Mjmole BSA
(Blank)
co
co co CO
Xi Si Si Iii Xi Zn
0 10 3 2 1 44
5
3 2 1 47
Cu/BSA reaction ratio
1.02,1.00,0.99 0.51 .53 .64
.92 . 00 .07 ,05 .24 .36 .91
The presence of 3 or more moles of nickel(I1) per mole of BSA reduces the copper(I1)-BSA reaction to less than lOy0of its value in the absence of nickel(I1) ; lesser amounts of nickel(I1) permit more copper(I1) reaction. Cobalt(I1) is less effective than nickel(I1) in suppressing the copper(11) reaction. Presumably the nickel(I1) and to a lesser extent cobalt(I1) prevent the reactions of copper(I1) by blocking the site on the BSA molecules where the copper(I1) reaction takes place. The order of stabilities of complexes of these metals is Co < Ni < Cu.l0 I n agreement with this order nickel(I1) was found to be more effective than cobalt(I1) in preventing complex formation with copper(I1). (10) H I n m L and I
