REACTIVITY OF CARBOBENZOXY-L-PROLYL-L-HISTIDYLGLYCINAMIDE 295
Jan. 20, 1959
THE DEPARTMENT OF BIOCHEMISTRY, CORNELL UNIVERSITYMEDICALCOLLEGE AND MEDICAL RESEARCH, EQUITABLE LIFE ASSURANCE SOCIETY O F THE UNITED STATES]
[ COSTRIBUTION FROM
THE
BUREAUOF
Coordination Complexes and Catalytic Properties of Proteins and Related Substances. 11. The Reactivity of Carbobenzoxy-L-prolyl-L-histidylglycinamidel BY 'It-ALTER L. KOLTUN, RICHARD E. CLARK,^ RICHARD N. DEXTER,3PANAYOTIS G. KATSOYANNIS AND FRANK
R . N. GURD
RECEIVED JULY 7 , 1958 The reactivity of carbobenzoxy-L-prolyl-L-histidylglycinamide,Cbz-PHG-"2, was measured kinetically by its ability to catalyze the hydrolysis of p-nitrophenyl acetate, NPA, and to interact with Zn(I1) and Cu(I1) ions as determined from simultaneous equilibrium PH values. The two methods of measuring the concentratioii of basic peptide were shown to be is 6.40 1. mutually compatible. The specific second-order rate constant k2 for the hydrolysis of NPX by Cbz-PHG-"2, mole-' min.+ a t 25.3". This value is in accord with that expected from the Bronsted catalysis law for imidazole derivatives where only the neutral molecules exhibit catalytic activity. I t is concluded that the presence of the peptide bond confers no special reactivity on the side chain imidazole group with respect to its ability to split NPAL At 25.1" and ionic strength 0.16 the pK' is 6.42. The pH titrations in the presence of Zn(I1) and Cu(I1) showed that the number of molecules of Cbz-PHG-NHz bound to either of these ions, V , approached an upper limit of 4, the value for imidazole itself. At 25.1' the logarithm of the first association constant, log kl was found to be 2.16 for Zn(1I) and 3.28 for Cu(I1). The subsequent steps for the combination of Cbz-PHG-NHZ with ZnlII) follow the same general course as those of imidazole with Zn(I1). With Cu( 111, however, the peptide shows positive interactions whereas imidazole shows negative interactions. The absorption characteristics in the visible between 600-800 mp of Cu(II)-Cbz-PHG-KH2 and Cu(I1)-imidazole complexes are presented. In both cases as 73 increases the extinction coefficient also increases and the absorption maximum is shifted to shorter wave lengths. The spectra of the individual Cu(I1) complexes were computed. With both types of ligand the absorption maximum decreases approximately 50 m p for each ligand molecule added to the complex. Cu( 11)-Cbz-PHG-SHz complexes, ho\wver, absorb 30-40y0 more strongly than do Cu( 11)-imidazole complexes.
Assessing the reactivity of the polar groups in proteins is complicated by their overlappin intrinsic reactivities. By simultaneously combining kinetic and equilibrium methods the reactivity of these groups may be defined more thoroughly than by relying upon one method alone. A previous report4 described this combined approach on a model system containing imidazole, p-nitrophenyl acetate (NPA) and Zn(I1) or Cu(I1) ions. The reactivity of imidazole was measured kinetically by its ability to catalyze the hydrolysis of NPA and to interact with Zn(I1) and Cu(I1) as determined from simultaneous equilibrium p H values. The two methods of measuring the concentration of basic imidazole were shown to be mutually compatible. A similar study is reported here on the synthetic tripeptide derivative, carbobenzoxy-L-prolyl-L-histidylglycinamide (Cbz-PHG-NHt), as a further step toward studying the reactivity of the histidyl residue in proteins. This compound has been used for the synthesis of an analog of the natural vasopre~sins.~It was chosen for this study because the imidazole group is the only unprotected functional group. Materials and Methods Materials.-The preparation and physical properties of Cbz-PHG-NH2 have been reported previously.6 The nitrate salt, H+PN03-, was prepared by dissolving the peptide in an equivalent amount of standardized nitric acid and lyophilizing the solution. It was stored over PzOS. Prior to preparing solutions the last traces of HzO, approximately (1) This investigation was supported by research grant No. H-2739 from the National Heart Institute, U. S. Public Health Service. ( 2 ) Holder of a Summer Research Fellowship under a grant from the U. S. Public Health Service t o Cornell University Medical College, 1957. (3) Lederle Medical Student Research Fellow, Cornell University Medical College, 1957. ( 4 ) W. L. Koltun, R. N. Dexter, R. E. Clark and F. R. N. Gurd, THISJOURNAL, 80, 4188 (1958). ( 5 ) P. G. Katsoyannis and V dii Vigneaud, Arch. Biochem. Bioghys., in press.
2%, were removed by drying a t 56" in mcuo over PZOSfor G hr The pK' was determined a t 25.1' by titrating a solution of 0.0496 M HfPN03- in 0.11 M NaN03 with 0.4670 N NaOH. After each addition of base 2 minutes were allowed for equilibration. The pK' was computed to be 6.42 3= 0.01 in the pH range 5.00-7.30. That the salt was pure may be judged from the constancy of the pK' values for widely varying ratios of free base to conjugate acid, 0.058 to 7.59. Furthermore, 0.050 M solutions of HfPiVO3- gave pH values of 3.86-3.88 whereas the theoretical pH of such a solution is 3.86. I n certain experiments the chloride salt, H+PCl-, was used. Usually 0.100 M solutions, containing 0.155 g. of Cbz-PHG-hTH2in 3.50 ml., were prepared by suspending the peptide derivative in water and adjusting the pH to 3.71 by the addition of standardized HC1 solution. A solution of this material was subjected t o paper electrophoresis (vide infra) and was found to give no ninhydrin test but instead a strong pink-orange band 6 cm. from the point of application when sprayed for the Pauly test. Solutions of the chloride and nitrate salts of Cu(I1) and Zn(I1) were standardized as previously d e ~ c r i b e d . ~ The p-nitrophenyl acetate (XPA) was the same as in a previous s t ~ d y . ~ Measurement of PH.--iZll determinations were made with a Radiometer Type TTT l a PH meter as described prev i o ~ s l y . ~Some of the earlier measurements were carried out on solutions a t room temperature (28-29') ; however, most solutions were thermostated in a water-jacketed cell a t 25.1 2c 0.1'. Unless stated otherwise all measurements reported were made a t 25.1 d= 0.1". Kinetic Method.-The catalysis of the hydrolysis of NPA by Cbz-PHG-NH2 was determined in the presence and absence of phosphate buffer. In one series of experiments 1 part of 1.0 X M K P h in 1.90y0 ethanol-H20 (v./v.) was added to 9 parts of a solution of phosphate buffer of varying pH and ionic strength containing varying concentrations of the peptide nitrate. The kinetic measurements were performed a t 26.5 i 1.0" and the pH measured a t 2629' a t the conclusion of the kinetic run. I n another series of experiments solutions were prepared containing peptide, peptide chloride, metal salt (as required) and sufficient NaC1 to maintain a constant ionic strength of 0.$6 during the reaction. After the p H was measured a t 25.1 , 0.10 ml. of M N P A in 1.90% ethanol-Hz0 (v./v.) was added 1.0 X to 0.90 ml. of the solution. The smaller volumes were accommodated in the Beckman Model DU Spectrophotometer in fused silica cells of internal dimensions 3 X 10 X 25 mm. supplied by Pyrocell Manufacturing Co., New York, X. Y. These kinetic runs were carried out a t 25.3 zk 0.3 O . The p H was again measured a t 25.1" a t the end of the run. Details of the technique and computations have been described.4
.
296
W. L. KOLTUN,R. E. CLARK,R. N. DEXTER,P. G. KATSOYANNIS AND F. R.N. GURD
Paper Electrophoresis.-The Spinco Model R Paper Electrophoresis apparatus was used. The electrophoresis was performed using a buffer containing 0.02 M sodium acetate and 0.02 M acetic acid. Samples of 10 microliters of solutions containing about 0.2 micromole of peptide derivative were run on strips of Schleicher and Schull No. 2043 .4 paper. A current of 2.5 milliamperes was maintained for 6 hr. The ninhydrin reagent was composed of 180 ml. of nbutanol, 20 ml. of H,O and 0.5 nil. of glacial acetic acid. The Pauly test (which may be made on strips previously sprayed with ninhydrin) was carried out by spraying with 0.02 Af diazobenzene-p-sulfonic acid in 5 1%' HC1; after the paper had dried it was sprayed with 5% Na2C03.
Results Catalysis by Cbz-PHG-NH2.-The observed first-order rate constants were taken to be equal to kl k,, where kl is the first-order rate constant for catalysis by Cbz-PHG-NH2 and k, is the rate constant for hydrolysis in a solution a t the same pH,temperature and ionic strength in the absence of ~ a t a l y s t . ~The apparent second-order rate constants, kz', were obtained from the expression kz' = kl/(PO), where (Po) is the total molar concentration of peptide derivative. The specific secondorder rate constants were calculated from the expression kz = kl/(P) where (P) is the molar concentration of the basic form of the peptide derivative. The results of kinetic measurements a t 26.5 rt 0.5' in the presence of phosphate buffers and 0.2% ethanol are shown in Table I. The PH dependence of k2', column 4, for any fixed value of (Po) suggests that the only catalytic species is the basic form of the imidazole group. This assumption was made in the calculation of kz, shown in column 5. The value of pK' of 6.42 was used to compute (P). The constancy of the values of ks in Table I supports the interpretation that the basic form of imidazole is the catalytic species and also shows that the concentrations of phosphate that were present had little or no effect. The grand average value of kz a t 26.5' is 6.73 1. mole-' min-1.6 For purposes of standardization of the method as a measure of (P) in the systems containing Zn(I1) and Cu(II), a similar group of measurements of kz was made a t 25.3' a t ionic strength 0.16 (adjusted by adding NaC1). The concentration of ethanol was 0.2%. No buffers were added. The average of 5 determinations over a range of (P) between M and 3.39 X M gave kz = 0.87 X 6.401. mole-' min.-'. The slightly lower value found a t 25.3' compared with 26.5' is in accord with previous experience with i m i d a ~ o l e . ~ * ~ Measurement of Free Basic Cbz-PHG-NH2 in the Presence of Zn(I1) and Cu(I1): (A) Kinetic Measurements.-The finding of a constant value of kp a t constant temperature and ionic strength over a wide range of concentration of H + P and P indicates that H + P does not act as a catalyst and that kinetic measurements may be used to determine the concentration of P. Table IIA shows the results of kinetic measurements performed in the presence of Zn(I1) and Cu(I1) ions a t 25.3'. The
+
(0) T h e magnitude of the temperature dependence of p K ' observed with imidazole derivatives' indicates that only small errors are introduced by using a p k " value determined a t 3 5 . 1 O . (7) J. T. Edsall, G. Felsenfeld, D. S. Goodman and F. R.N. Gurd,
'his
JOIJKNAL, 76, 3054 (1954). (8) T.C. Bruice and G, I,. Schmir, i b i d . , 79, 1663 (lQ57).
Vol. 81
TABLE I TIfE OF
A N D C O N C E S T R A T I O S D E P E N D E N C E OF TIIE KATE
CARBOnENZOYY-PKOLYL-HISTIDYL-GLYCINAMIDE
LYZED
HYDROLYSIS OF P-NITROPHENYL ACETATE
A.
p
7.45 7.27 7.01 6.69 6.41 6.03
B. 7.55 7.34 7.06 6.73 6.43 6.05
p
kn'
(1. m o l e - '
phosphate 31.77 30.65 27.82 22.30 15.48 9.58
= 0.06,
(pol = 5.0 x
phosphate 19.29 18.46 16.67 13.56 9.45 5.97
= 0.09, (PO)= 3.0
1.26 1.15 0.83 .63 .43 .12 1.41 1.21 I . 03 0.84 .50 .l5
(1. m o l e - ' miu. -1)
min.-I)
6.36 f;. 13 5.56 4.46 3.10 1.92 6.43 ti. 15 5.58 4.58 3.15 1.99
CATA26.5'
12 2
k , X 108 (min.-I)"
k~ X 108 (mh-1)
D 13
AT
10-3
AI 6.95
7.00 6.98 6.86 6.27 6.60
X 10-3 M 6.91 6.89 6.83 6.75 6.25 6.65
C. I.( phosphate = 0.11, (Po) = 1 5 X 10-8 Af 1.57 7.63 9.80 6.54 6.94 1.44 6.36 7.40 9.55 7.01 8.53 1.21 5.69 7.11 6.84 0 . nn 6.75 6.44 4.29 6.30 6.46 4.93 .57 3.29 6.30 .19 2.13 6.89 6.07 3.19 a For these measurements made in the presence of buffers, the values of k, were obtained b y interpolating between values of the observed rates in a series of buffers of known OH and ionic strength, instead of by extrapolating t o zero concentration of buffer.
concentration of free basic peptide, expressed as the negative logarithm, was calculated using a value of kp of 6.40 and assuming that complexes with these cations do not act ~ a t a l y t i c a l l y . ~ (B) Equilibrium Measurements.-After the completion of the kinetic measurements the equilibrium p H values were determined a t 25.1'. The concentration of the free basic peptide in each solution was calculated froin the measured PH,' using the appropriate value for the concentration of H+PCl- and the PK' of 6.42. The results are shown in Table IIB. A comparison of the two independent methods used to obtain the values of (P) may be made by comparing columns 5 and 8. The close agreement indicates (1) the validity of the assumption that the metal-peptide complexes are not catalytic agents to any measurable extent, and (2) the lack of interference by the hydrolytic products of NPA a t the low concentrations present. Association of Zn(I1) and Cu(I1) with CbzPHG-NH2 : (A) Zn(I1) .-From the data presented thus far, the first association constant of Zn(I1) and Cbz-PHG-NH2 may be determined. In Table I1 are listed the values of ti, the average number of peptide molecules bound per metal ion present, which were computed by the method previously employed with i m i d a z ~ l e . ~The ~~-~ data in Table I1 were analyzed by the method of Scatchard' from plots of log Q versus g where Q = Y / ( N - ;)(P) (9) Y. Nozaki, 79, 3123 (1957).
F. R. N. Gurd, R. F. Chen and J. T. Edsall, i b i d . ,
Jan. 20, 1959
REACTIVITY OF
CATALYSIS OF HYDROLYSIS OF NPA
BY
297
CARBOBENZOXY-L-PROLYL-L-HISTIDYLGLYCINAMIDE
TABIEI1 CARBOBENZOXY-PROLYL-HISTIDYL-GLYCINAMIDE, (P). IN THE PRESENCE OF Zn(I1) AND
-Composition-Initial total molar concn. of------
-kt kl
x
Cu(I1)
-
102,
AT
25.3'
A. Kinetic results 6.40 1. mole-1 min.-'-
ZnClt
H +PC1-
P
0.0098 .0098 .0098" .00196 .00196 .00196
0.04566 .04180 ,03745 .04807 .04517 .04228
0.00434 ,00820 .01255 .00193 .00483 .00772
1.216 1.716 2.536 1.091 1.976 2.888
2.72 2.57 2.43 2.77 2.50 2.34
0.04848 0.00145 .04708 .00290 ,04516 .00483 ,04324 .00675 .04035 .00965 6.73 1. mole-' m h - 1 .
0.215 .456 .962 1.529 2.885
3.47 3.15 2.82 2.62 2.35
min. -1
-log (P)
-PK' OH
Y
B.
Equilibrium results = 6.42-log (P)
Y
1.61
4.98 5.26 5.44 4.90 5.28 5.48
2.80 2.54 2.41 2.84 2.48 2.31
,245 .791 1.46
0.562 1.11 1.68 2.20 2.60
4.26 4.61 4.91 5.17 5.46
3.49 3.14 2.85 2.61 2.35
0.569 1.10 1.73 2.18 2.65
0.249 .563 ,899 ,114 .868
0.282 ,542 .BO
CuClr
0.00198 .00198 .00198 .00198 .00198 = 26.6', ka =
T and N denotes the maximum number of sites available for binding the ligand. In the present system it was assumed that N = 4.'3 The results for Zn(I1) are plotted in Fig. 1, closed circles. Since the kinetic and equilibrium measurements were in close agreement, only the values obtained from equilibrium measurements are shown. As previously d e m o n ~ t r a t e dthe , ~ ~value ~ of log ~ 1the , first intrinsic association constant, is equal to the limiting value of log Q when B approaches zero. Extrapolation yields a value of log ~1 = 1.56. Using the statistical relation kl = 4 ~ 1 where , kl is the first association constant, a value of log kl = 2.16 is obtained. The first association constant in such a system is of greatest interest as a guide to the possible behavior of imidazole groups in proteins.1° The general course of the subsequent steps of the reaction was determined by the technique of stepwise titration to conserve the supply of the peptide derivative. For example, a solution containing 0.050 M H+P NOa-, 0.002 M Zn(NO& and 0.10 M NaN03 was titrated with standard
NaOH solution containing sufficient NaN03 to maintain the ionic strength a t 0.16. The NaOH was added from a microburet dipped below the surface in a titration vessel fitted with a magnetic stirrer. The same technique yielded results with imidazole itself that agreed well with previous measurements on separately prepared s o l ~ t i o n s . ~ The results with the peptide derivative are summarized in Table I11 and Fig. 1, open circles. Usu-
2.80
2.40
c;
1
c
TABLE I11 TITRATION OF Cbz-PHGNH2 IN THE PRESENCE OF Zn(N03)z AT 25.1' -Total P
molar concn. ofZn(N0s)r NaOH
pH
-log(P)
;
0.05049 0.00199 0.000909 4.50 3.23 0.173 .05044 .00199 ,00134 4.70 3 . 0 3 .214 .05037 .00199 ,00201 4.90 2.84 ,282 .05026 ,00198 .00303 5.10 2.65 .394 .05009 .00198 ,00459 5 . 3 0 2.46 .580 .04997 ,00197 ,00573 5 . 4 0 2 . 3 7 .765 .04982 ,00196 .00714 5.50 2.29 1 . 0 2 ,04965 ,00196 .00870 5 60 2 . 2 1 1.28 ,04944 ,00195 ,01063 5.70 2 . 1 3 1 . 6 6 .04922 ,00194 ,01268 5 . 8 0 2.06 2.02 .04896" .00193 .01508 5 . 9 0 1.99 2 . 5 1 ,04868 ,00192 ,01762 6.00 1 . 9 3 3.03 .04842 .00191 .02000 6 . 1 0 1.87 3.35 ,04815 .00190 .02253 6 . 2 0 1.81 3 . 7 3 .0478gb .00189 ,02448 6 . 3 0 1 . 7 5 3 . 5 6 Transient precipitate. Permanent precipitate. (10) F. R. N. Gurd and P. E. Wilcox, Adv. in Prolein Ckem.. 11,311 (1956).
I.4O
0
I .o
2 .o
3.0
4.0
3 Fig. 1.-Plot of log Q us. V for the binding of Zn(I1) to Cbz-PHG-NHZ at 25.1". The closed circles, 0 , represent points derived from p H measurements on individual solutions; the open circles, 0, represent points derived from stepwise titration.
ally p H measurements were made 3 minutes after the addition of NaOH to allow the system to conie to equilibrium. Above p H 5.80, corresponding to a-2, transient precipitation occurred upon the addition of NaOH, and p H measurements were
(B) Cu(I1).--Measurements un individually prepared solutions are described in the lower part of Table 11. The kinetic results (A) may be secn to be in good agreement with thc equilibrium results (B) showing as before that the complexes with Cu(11) ions do not act catalytically. The results of the equilibrium iiicasurernents from Table IIB have been used to conipute values of log Q which are plotted against D in Fig. 2, closcd circles. X stepwise titration also was performed. as shown in Table IV. The first part of tlie titration consisted of successive atiditions of NaOrI solution to the solution containing II+PNO:I-, Cu(lnj0,)Z aiid suficieiit NaCl to give an ionic strength of O . l ( j . The results are preseiitcd in Table IVA aiid are used to compute values of log Q which arc plotted against v ill Fig. 2 (opeii circles). -1 solution of 0.4300 ,IT S a O H was usctl in order to niiniiiiize dilution. The second part of the titration consistcd of atldiiig Cu(NO;J? solution (ionic strength 0.165) instead of NaOH (Table IVB) and had the effcct of carrying li downward. It constitutes, therefore, ii type of e back-titration. The correspontliiig vaIucs of lug 2 . 6 0 L ~ ~ plottedl against D in Fig. 3 (half-closed @ are 0 1.0 2.0 3.0 4 .O circles). In the course of thc forward titration transient 3. precipitation was observed a t D = 0.95 and above. Fig. 2.-I'loC uf log Q vs. B for tlic binding of Cu(I1) to No precipitation was observed during the backCbz-PHG-XHz a t 26.1". Points dcrivcd from PH measuretitration. The agreement between the results of iiiciits on individual solutiviis are shown in closed circles, .; the forward titration (open circles) aiid the backpoints derived froin stepwise titratioii with S a O R arc showii titration (half-closed circles) indicates that the i n opcii circles, 0; points derived froin back-titratioii with system remained close to equilibrium throughout Cu( SOa)a arc shown iii half-closed circles, 4). T h c curve is the stepwise titration. coinputed as tlcscribcd in the test. The curve shown in Fig. 2 is ctmiputec17~9 using , IC?, log K~ and log of L'.GS, niade oiily after the precipitate appcared to have values of log ~ 1 log redissolved. Permanent precipitation occurrcd 2.90, -3.43 and 3.25, respectively. The corrcspoiitling log k-valucs are 3.28, 3.08, 2.27 and 2.&5.4s7,9 above pH 6.20, D 3.7. Absorption Spectra in the Visible of the Cu(I1) As may be seen from Fig. 1, the results of the stepwise titration approach much the same lixnit- Complexes of Cbz-PHG-NH2 and Imidazole.-ing value of log Q at li = 0 as found with the In view of the absorption in the visible of Cu(I1) iiidividually preparcd solutions (closed circles). complexes with basic nitrogen atoms such as those i t wzts The open circles in Fig. 1 could be fitted quite well in arnnionia,l a iriiiclazole7 or proteins, by taking log k2 = log k, = log k4 = 2.00 and con- felt advisable to exmiiue the absorptioii cliaracterstructing a theoretical curve. These constants istics of Cu(II)-Cbz-PHG-IYH2 complexes. i l b give a curve slightly steeper than that needed to sorption spectra over the wave length range GOO-SO0 fit the results for coniplcx-formation between Zn- nip were measured for the conditions iiitlicated ill (11) ions and imidazole i t ~ e 1 . f .Because ~ of the Table 11. The pertinent results are show1 iii formation of transient precipitates during the Table VA. All readings were made against a stepwisc titration, i t seems wise to assume rather water blank and tlic observed optical densities wide limits of uncertainty for the constants other corrcctcd for the absorption due to free Cu(I1) ion. than k, and, perhaps, k?. Experimentally, i t The percentage distributioii of Cu(I1) bttweeii the appears that transient precipitation introduces a free ion and the various species of coniplexcs R';~s bias toward low value of log Q." It is possible calculated as a functioii of the concentration of that the stages of titration showii by the open basic ligand, expressed as -log (A),using thc SUCcircles in Fig. 1 represented only very close ap- cessive values of the association constants given proaches to equilibrium and that the values of above. The values showii in cofunilis 4-S corvalues in column 3. The log kS and log kq taken above may be as much as respond to the -log (-1) extinction coefficients E are for a 1 c ~ n light . path 0.2 or 0.3 too 10w.I~ and are expressed in terms of the total molar cow (11) Similar stepwise titrations emiiloying higher concentrations of
l
-
Z n ( I I ) were niarked b y onset of t r a n b i e n t precipitation almost from t h e i i c y i n n i n g ; i f the rcsiilts v e r e plotted in I i g . 1, they would fall bel o w lie open circles. Longer equilibration ?erve,l to bring tlie reanlts cluser to those shown in the open circles. (12) One other possible source of error could not be evaluated empirically. It ai observed that releasing saturated KCI solution f r o m the ca!,,niel electrode i n t o s d u t i o n s in which V was greater t h a n ,,I,c>II~ 2 c r i t i t ~ r v s t t l t ~ t i 11, I I V I I C L . : , r e c i f ~ > t , ~ t i :it < ~ rt ih e I i w n d a r y i > i t h c
KCI solution. This obserratiun was m a d e only i n t h e presence of Zil(11) iuns. Presumably such a. ],henumcnull could occIIr at t h c i n t e r f . L.L ~at the til, of t h e s d t bridge f r o m tlic ciiloinel electrode d u r i n g n o r mal operalion, and might anect t h e accuracy of t h e measurements. (13) 1. Bjerrum, "hletal Ammine I'orniation i n Aqueous S o l u t i ~ ~ n , " I'. Haase and Son, Copenhagen, l%ll. (14) 1. h1. K l o t ~in " T h e Proteins," I1 1-eiiratli and 6 . Vol. I . I ' i r t I i . . l c ' ~ c l c ~ ~I'Tc-\, iic 1nc S c , i i T c ) I L , Y 1 I
'L.
Jan. 20, 1959
REACTIVITY OF
CARBOBENZOXY-L-PROLYL-L-IIISTIDYLGLYCINAIVIIDE
TABLE IV TITRATION OF Cbz-PHGXH2 IN THE PRESENCE OF Cu(N03)*AT 25.1" -Total P
molar concn. ofC u ( N 0 8 ) ~ NaOH
pH
--log(P)
U
A. Forward titration-addition of 0.4670 N NaOH 0.04885 0.00213 0.000420 .04880 .00213 .000940 .04872 ,00212 .0016i .04868 .00212 .00206 ,04863" .00212 .00253 ,04867 ,00212 ,00309 ,04851 .00211 ,00366 ,04844 .00211 ,00437 ,04837 ,00211 ,00504 ,04828 ,00210 .00589 ,04820 ,00210 ,00669 ,04811 .00210 .00751 ,03800 ,00209 .00854 .04789 ,00209 ,00964 ,04775 .00208 ,01095 ,04757 ,00207 .01270 ,04744 ,00207 .01395 ,04724 ,00206 ,01579
3.91 4.11 4.31 4.40 4.50 4.60 4.70 4.80 4.90 5.00 5.10 5.20 5.30 5.40 5.50 5.60 5.70 5.80
B. Back titration-addition of 0.0549 0.00256 0.01563 5.73 5.68 .00306 ,01549 5.59 .00355 .01534 5.51 ,01520 ,00396 5.43 ,00449 .01506 5.38 ,01492 .00495 5.30 .00541 .01479 5.23 ,01465 .00585 ,01452 5.19 .00629 .01439 5.12 ,00671 .01427 5.08 ,00713 ,00755 ,01415 5.03 5.00 ,00795 ,01402 4.98 ,00835 ,01390 4.93 ,01379 .00875 a Transient precipitate above pH 4.50.
n.04679 ,04635 ,04591 ,04545 ,04507 ,04465 ,04425 ,04385 .04346 ,04308 ,04270 ,04233 ,04197 ,04161 ,04126
3.83 3.63 3.44 3.35 3.26 3.16 3.07 2.98 2.88 2.79 2.70 2.61 2.52 2.44 2.35 2.28 2.20 2.12
0.325 .369 ,636 .781 .945 1.15 1.34 1.58 1.77 2.04 2.24 2.42 2.65 2.57 3.14 3.58 3.66 4.00
M cu(iY0~)~ 2.18 2.26 2.35 2.43 2.51 2.57 2.65 2.73 2.77 2.84 2.89 2.94 2.97 3.00 3.05
3.49 3.23 3.05 2.90 2.67 2.59 2.34 2.18 2.04 1.93 1.82 1.72 1.68 1.54 2.48
centration of bound Cu(I1). It is clear that as B increases the extinction coefficient also increases (column 10) and the absorption maximum is shifted to shorter wave lengths (column 9). For purposes of comparison spectral results obtained with Cu(I1)-imidazole complexes are shown in Table VB. The percentage distribution of CU(11) among the various forms was calculated using the successive association constants previously r e p ~ r t e d . ~The characteristic trends of the wave lengths of maximum absorption and of niaxiinuin extinction coefficient again were observed.13 If spectral measurements are to be useful in determining the degree of complex formation between imidazole groups and Cu(I1) ion, a knowledge of the absorption spectra of the individual complexes is essential. The individual spectra were computed from the calculated composition and the experimentally determined absorption spectra of the solutions. The wave lengths of maximum absorption and the maximum extinction coefficients are shown in Table VI. The close parallel between X max for Cu (II)-Cbz- PHG -NHz and for Cu(I1)-imidazole is evident. In both cases
290
the Amax decreases approximately 50 mk for each ligand molecule added to the The former complexes, however, absorb 3 0 4 0 % more strongly than do the latter complexes. Because the solutions studied were relatively weakly absorbing, the extinction coefficients of the Cu(I1)-CbzPHG-NHz complexes are considered reliable to within 10%. The values for the Cu(I1)-imidazole complexes are probably reliable to within 5%. As a check on the validity of the procedures used the extinction of each solution was computed a t Xmax from the composition of the solution and the calculated extinction coefficients of .the individual complexes. The observed and computed values are in excellent agreement (Table V, columns 10 and 11). The extinction coefficients for Cu(Im)++ and Cu(Im)4++, 19 and 56 1. mole-l cni.-'. respectively, are in reasonable agreement with those reported earlier, 16.5 and 53.' The slight differences may reflect the use of chloride salts in the present study instead of the nitrate salts used previously. It is interesting that the addition of the third ligand to the metal ion is marked in both systems by a relatively small increase in extinction coefficient, even though the displacement of Xmax is equivalent to that which accompanies the addition of the second and fourth ligands. Stability of Cbz-PHG-NH2 in the Presence of Cu(I1) Ion.-Following the spectral measurements, the last three solutions described in Table VA were pooled and brought to p H 5 with acetate buffer. The Cu(I1) ion was removed by extraction with dithizone in chloroform and the aqueous solution subjected t o paper electrophoresis. The material behaved exactly as did the stock preparation of Cbz-PHG-"2, which indicates that the peptide derivative was stable in the presence of Cu(I1) ion. Discussion Catalytic Properties.-Thc value of kz, the specific second-order rate constant for hydrolysis of NPA by Cbz-PHG-NH2, is 6.40 1. mole-' iiiin.-l a t 25.3" in 0.27, ethanol and ionic strength 0.16. For the hydrolysis by imidazole under cornparable conditions, kz is 31.3 1. mole-' rnin.-'e4 The lower value is to be expected in view of the lesser basicity of the peptide. In a study of solutions of various imidazoles in which the principal catalytic species is the neutral molecule, Bruice and Schmir showed that catalytic activity could be related by the usual Bronsted expression, log hl = apK' b.I6 The relation between the values of kl for Cbz-PHG-NH2 (pK' = (3.42) and for imidazole (pK' = 7.08)4 is rather close to that found by Bruice and Schmir for a series of imidazoles under somewhat different conditions.* It appears, therefore, that the presence of the peptide bonds linking the histidyl residue with its neighbors in Cbz-PHG-NH2 confers no special reactivity on the side-chain imidazole group with respect to its ability to split NP.4. Complex Formation with Zn(1I) and Cu(I1) Ions.-Just as the imidazole group in Cbz-PHG-
+
(15) Bjerrum has reported similar displacements in the Cu(I1)ammonia systern.la (le) T . C . Rrriice a n d P.. L. S c h m i r , THISJ O V R N A I . , 8 0 , 148 (19581.
TABLEIJ THEABSORPTIOHCHARACTERISTICS I N THE VISIBLE REGIONOF Cu(I1) COMPLEXES OF Cbz-PHGNHm
4.26 4.61 4.91 5.17 5.46
0.57 1.10 1.73 2.18 2.65
3.49 3.14 2.85 2.61 2.35
4.49 4.92 5.30 5.62 5.97 6.30 6.53
0.50 0.99 1.48 1.95 2.46 2.90 3.16
4.29 3.86 3.48 3.16 2.81 2.48 2.25
A . Cbz-PHG-SHz; total CUCIZ= 0,0020 M 53.0 33. r) 13.3 0.7 0.0 26.0 36.0 32.0 4.3 1.0 10.2 26.8 44.2 12.0 7.0 3.9 14.3 41.5 19.0 22.0 0.5 5.0 26.8 22.5 46.2
I3. Imidazole; total CuC19 = 0.0099 111 42.0 0.3 5.5 53.5 20.0 2.5
52.0 25.0 8.5 2.5 0.3
.o 0
44.0 25.0 9.5 3.5 1.5
ih3SoRPTIoN CIIARACTERISTICS
9.0 24.0 44.0 48.0 46.0
IMIDAZOLE
735 720 700 685 650
37 35 43 47 53
32 36 42 47 53
730 725 715 700 670 655 625
21 23 26 30 35 38 41
21 23 26 30 35 38 40
other imidazole complexes with Cu(I1) ion^.^^^
TABLE VI THE
38.0 46,O 38.0 24.0 15.0
0.0 .I .3 2.0 8.7 24.5 37.5
AND
I N THE VISIBLE
REGION To interpret the course of these later steps it is
teiripting to postulate that the complexes are stabilized by interactions between parts of the peptide tlerivative other than the linkages between --Cbz-PHG-NH2--,--Imidazole--emrr, fmnx, imidazole groups formed by the Cu(I1) ion. Amax, 1. mole-' Amax, 1. mole-' mg ern.-' rnfi cm.-l ^Ipparently the configuration of the Cu(I1) coinCuA 735 27 735 19 pleses alloivs greater interaction between the ligand CUA~ 685 49 690 35 molecules than does that of the Zn(I1) complexes. CuA3 635 55 8:35 39 Study of molecular models shows that the inlidCuA4 Goo 71 A00 56 azoie groups of four niolecules of the peptide derivative may combine a t one t.iine with the metal NH2 appears to behave as an individual group in ion either in the tetrahedral configuration charits reaction with N P 4 , so also does it appear to be acteristic of ZnjlI) complexes or the square co-the sole ligand group in the binding of the peptide planar configuration characteristic of Cu(II) coniderivative t o Zn(I1) and Cu(I1) ions. This is piexes.9,10 It must be borne in mind that the shown by the agreement between kinetic and equi- complex bearing two ligand groups per Cu(1I) librium measurements (Table 11) arid by the fact ion exists in two possible forms and that difthat with both ions approaches 4, the value of ferences in distribution betweell these fornis may N for imidazole U'e may conclude that account for part of the differences in the progression stable complexes are not formed that involve the of K-values for the systems containing Cbzmetal ion simultane.ously with more than one PHG-KIL'H2 and imidazole, r e s p e ~ t i v e l y . ~ ~Con'~ ligand group in each molecule of the peptide deriva- firmation of the role of secondary interactions and tive. configurational effects in the stability of such comThe value of 2.1G for log kl for the binding of plexes must await studies a t other temperatures. Cbz-PHG-NHZ to Zn(I1) ions falls slightly below Studies 01i dipeptides such as glycylglycine have the corresponding values for iinidazole and 4- shown that Cu(I1) ion may combine with an ainethylirnidazole, 2.j24 and 2.44,9 respectively. amino group and then proceed to displace a hydroAs mentioned earlier, the increases in affinity be- gen ion from the peptide bond.18-20 The postween Zn(1I) and Cbz-PHG-NH2 with each suc- sibility that the imidazole group might be able to cessive step of complex formation are probably a t play the role of the a-amino group and permit an least as great as those shown by the Zn(I1)- analogous structure to form in the Cu(I1)-Cbzimidazole system. PHG-XH? system appears to be excluded. The The value of 3.28 for log kl for the binding of agreement between the kinetic and equilibrium reCbz-PHG-NH2 to Cu(I1) ions falls decidedly below sults described above allows little room for such a the corresponding values for imidazole and 4- competing reaction. Furthermore, the spectral remethylirnidazole, 4.204 and 4.13, respectively. sults fit the Cu(II)--iinidazolc system well but do 'The reason is not apparent at present. It is pos- not ccdoriii to the behavior of systems of the sible that steric hindrance is responsible, but we Cu(II)- glycylglyci~~c type.18.20 have !ittle inforination about the variability of Coninlent on Cbz-PHG-MH2 as a Model for the kl vl~luesfor Cu(1I) cornpiexes of substituted iini- Imidazole Group in Proteins.---A group comprising tlazoles. 0 1 1 the other hand the course of the sur-- the side-chain of an amino acid residue in a prorcssivt steps oi complex formation (log k 2 =: tcin niay rezt,t indi~;idually or may form part of a ?.
