Jan. 5 , 1960
R E A C T I O N OF GLYCINE-CONTAINING
glycosyl bond is considerably smaller in the Cz’exo and Ca’-exo conformation than in any other conformation which has been considered acceptable. Thus, intramolecular conversions from one conformation to another of the ribose ring will cause changes in the diameter of polynuclGtides and nucleic acids which may or may not be followed by changes in specific rotation, A systematic survey of -the stereo configuration of five-membered
DIPEPTIDES WITH
CUPRIC
IONS
233
rings and their derivatives by p.m.r. and optical rotation studies is planned, as well as studies on the implications of the furanose conformation on nucleic acid structure and on the specificity of enzvmes which attack these comDounds. Acknowledgment*-l
am very grateful to Dr*
J. T. Edsall for helpfiil criticism of the manuscript. MASS.
CAMBRIDGE,
[CONTRIBUTION FROM THE DEPARTMENT OF BIOCHEMISTRY, CORTELL UXIVERSITY MEDICAL COLLEGE, THE DEPARTMENT OF B I O C H E M I S T R Y . rTNIVERSITY O F FLORIDA SCHOOL OF MEDICINE, AND THE BUREAUO F M E D I C A L RESEARCH, EQUITABLE h F E ASSURANCESOCIETY O F THE USITED STATES]
Coordination Complexes and Catalytic Properties of Proteins and Related Substances. IV. Reactions of Glycine-containing Dipeptides with Cupric Ions and with p-Nitrophenyl Acetatel BY
WALTER
L. KOLTUN, MELVIN
F R I E D 2 s 3 AND
FRANK R . N. GURD2p4
RECEIVED J U N E 10, 1959 Equilibria between Cu( 11) ions and glycylglycine (CG), sarcosylglycine (GS), L-prolylglycine (PG), glycyl-L-valine (GV), L-valylglycine (VG), glycylsarcosine (GS) and glycyl-1,-proline (GP) have been measured by potentiometric titration. All the peptides except the last two form complexes with Cu( 11) in which the peptide hydrogen atom is displaced. The potentiometric results are correlated with spectral measurements. Analysis of the equilibria above pH 7 in Cu(I1)-dipeptide mixtures is coupled with the demonstration that a certain basic complex catalyzes the hydrolysis of p-nitrophenyl acetate ( “ P A ) The formation of a catalytically-inert olate complex is also explored. The rate of acetylation of the various dipeptides by NPA is correlated with their basicity. Lastly, a kinetic method is used to measure the formation of a mixed complex of Cu( 11) with both glycylglycine and imidazole.
Introduction peptide linkage as well as the a-amino group in the The preceding paper of this series5dealt with the formation of metal complexes means that a knowleffect of zinc and cupric ions on the reaction of edge of the reactivity of only the first residue in a glycylglycine (GG) with p-nitrophenyl acetate peptide sequence is insufficient to permit the predic(NPA). By a combination of kinetic and equi- tion of the behavior of the N-terminus of a peptide librium measurements similar to previous s t ~ d i e s ~or , ~protein. For this reason the present study was on systems containing imidazoles, it was possible undertaken to compare the reactivities of several to determine unambiguously the association con- dipeptides with Cu(I1) ions and with NPA. The dipeptides studied primarily are glycylglystant for the formation of the complex CuGG*-, according to equation 3, and to confirm the picture cine (GG), sarcosylglycine (SG), L-prolylglycine of successive equilibria described by the following (PG), glycyl-L-valine (GV) and L-valylglycine (VG), all of which follow the reaction sequence 1-3 sequence of reactionsa with Cu(I1) ions. The evidence of Datta and Rabing that glycylsarcosine (GS) and glycyl-Lproline (GP) follow equations 1 and 4 has been confirmed.
(3)
In reaction 2 a hydrogen ion is displaced from the peptide linkage. The participation of the (1) This investigation was supported by research grant No. H-2739 from the National Heart Institute, U. S. Public Health Service. (2) A preliminary report presented at the 132nd National Meeting, American Chemical Society, New York, iX.Y., Sept. 8-13. 1967, described work begun in the Department of Biochemistry, Washington University School of Medicine, St. Louis, Missouri. (3) Senior Postdoctoral Fellow of the United States Public Health Service. (4) John Simon Guggenheim Memorial Fellow and Helen Hay Whitney Fellow. Washington University, St. Louis, 1954-1955. ( 5 ) W. L. Koltun and F. R. N. Gurd, THIS JOURNAL,81,301 (1959). (8) W.L. Koltun, R. N. Dexter, R. E. Clark and F. R. N. Gurd, i b i d . , 80, 4188 (1958). (7) W. L. Koltun, R. E. Clark, R . N. Dexter, P. G. Katsnyannis and F. R. N. Gurd, ibid, 81, 295 (1959). ( 8 ) T h e nomenclature of the association constants is defined in ref. 5; see especially footnote 17.
The values for the separate formation constants for these various peptides are compared, and with their help the properties of the absorption spectra of the individual complexes are estimated. Equilibria in equimolar mixtures of Cu(I1) ion and dipeptide a t higher p H values have been explored and the processes described in equations 5, 6 and 7 are shown to be possible sequels to that in equation 2. CUGG
CuGGOH-
CuGGOH-
+ CuGG
CuGGOH-) + H + kI,- = (H+)( ( CuGG) (5)
(CUGG)~OH-
(9) S. P. Datta and B. R. Rabin. Trans. Faraday SOC.,62, 1117 (1958).
kI,, = (H')(CuGG(OH)z") -_ ( CUGGOH'-J-
(7)
The complex CuGGOH- is shown to catalyze the hydrolysis of NPA. The various peptides are also cornpared with respect to their rate of reaction with NPX, and kinetic nieasurements on some of the Cu(I1)-dipeptide systems are used to confirm the values of kII,adduced from titration behavior. Lastly, kinetic ineasurements are used to show that CuCTG may combine with imidazole in a reaction comparable to equation 3.
the results of Dobbie and KermackI4 (20") and of Datta and Rabing (25") is limited to the region of low and variable ionic strengths. Datta and Rabin reported values for log k+ ' of 3.20 and for log k*' of 8.23, in good agreement with the present values when allowance is made for the difference in temperature. Dobbie and Kermack reported values a t 20" of 3.12 and 8.37 for log k+' and log ki', respectively. The values of log k*' shown in Table I for ( X , SG and GP are somewhat below the values of S . 7 , 8.63 and 8.66, respectively, reported by Datta and I t a b i ~ ~These .~ peptides probably follow the same pattern as GG in terms of variations of log k,' with ionic strength.15
Materials and Methods All peptides except glycylsarcosine (GS) and sarcosylglycine (SG) were supplied by Mann Research Laboratories, Inc., S e w York, N.Y. The commercial materials were titrated with €IC1 and XaOH under the same standard conditions as were used in the presence of Cu(I1) ions and all gave association constants with hydrogen ions which showed no trends during the course of the titration. Gl~vJ'lglyCilie ( G G ) , glycyl-L-proline ( G P ) and L-prolylglycine ( P G ) were recrystallized from water a t 50-60' by the addition of etlia n d . SG was prepared according t o Levene, et al.lo The n1.p. was 196-198". G S was prepared by coupling carbobenzoxyglycine with sarcosyl ethyl ester by means of N,S'dicyc1ohexylcarbodiimide.ll The resulting ester was saponified and the product reduced with hydrogen in the presence of 5% Pd on charcoal. The GS was crystallized from water by the addition of ethanol. T h e m.p. was 201.5-202..5". Acetylglcyine ( AcG) and acetylglycylglycine ( AcGG) were prepared according t o the methods of Fischer12and Gordon. Martin and Synge.'3 All stock solutions of peptides were tested for purity by paper electrophoresis in the Spinco Model R apparatus. using Spinco B2 buffer adjusted to PH 8.2 by the addition of about 150 ml. of 0.1 M HCI per liter, ionic strength 0.075. Schleicher and Schiill No. 2043A paper was used. The duration of the run was either 6 hr. a t 2.5 mamp. or 3 hr. 3.t 5 mamp. The colors were developed by ninhydrin spray.s All peptides behaved as single components. The initial colors on spraying were yellow for GG, G P , G S and GY; orange for PG; and purple for VG. \'ery little color was given by SG. The sample of imidazole and the other reagents were ns used previo~sly.~-7 Kinetic and pH measurements were made as previously described6-' a t 25.0 zk 0.1'. Some of the earlier titrations were performed using the Beckman Model G pH Meter with external electrodes thermostated a t 20.0 f 0.1".
Results Association Constants for Hydrogen Ions.-The logarithms of the association constants for equilibria with hydrogen ions in the absence of metal ions are shown in Table I. The constants describing the association of a hydrogen ion with the carboxylate group are denoted by k+ ' to show the charge type of the peptide formed in the reaction; likewise, the association constants for the reactions with the amino group are denoted by k*'. In addition to the measurements a t ionic strength 0.16 the titrations of GG were performed in the absence of added electrolyte (ionic strength 0.0005-0.0045) and in the presence of 1 M KN03 a t 20". The values of log k+' were 3.29 and 3.57, respectively, and of log k+' 8.46 and 8.43, respectively. Comparison with (10) P. A. Levene, H. S. Simms and M. H.Pfaltz, J . B i d . Cltem., 61, 450 (1924). (11) J. C. Sheehan and G. P. Hess, THISJOURNAL, 77, 1007 (1955). (12) E.Fischer, Be*., 37, 2486 (190-4). (13) A. H. Gordon, A. J. P. Martin and R. L. M. Synse, Biochem. J . , 3 1 , 79 (1943).
TABLE I
VALUESOFLOGk+' AND L m k i 'FORTHE VARIOUSP w r r n E s Temp. 25.1"; ionic strength 0.16 Peptide
Glycylglycine Glycylsarcosine Snrcocylglycine Glj-cyl-L-proline L- Prcilylglycine Glycyl-L-valine L-Valylglj-cine
log k,
'
3.19 '3.98 3.1,: 2.97 3.19 3.15 3.23
log
k*
8.13 8.57 8.56 8.48 8.97 8.18 8.00
Formation of 1:1 Cu(I1)-Dipeptide Complexes at pH Values below 7.-The course of the titration of a mixture approximately 0.01 M with respect to both CuClz and GG is shown in curve 3 of Fig. 1 The evaluation of kI,+ and kI,oin reactions 1 and 2 is complicated by the fact that the two reactions occur in the same PH range, primarily between PH 4 and 6, and that significant quantities of the GG+ form are present a t the lower end of this p H range. On the other hand the reactions described in equations 3-7 may be neglected below p H 7 in the presence of equimolar mixtures of cupric salts and peptides of the type of GG, SG, PG, GV and VG. The description of reactions 3, 5 , 6 and 7 for these peptides is reserved for a later section. A method of successive approximation has been used for evaluating ki,+ and kI,o. Arbitrary values of kI,o are chosen and substituted into appropriate expressions until a value is found that leads to a value of kI,+ which is independent of PH over the appropriate range, The following expressions apply to the conditions under discussion (CUT)= (Cut') (CuGG+) (CUGG) (8)
+ + + (GG*) + ( G G - ) + (CuCG') + (CuGG) (CuGG+) + 2(CuGG) + A(GG-) -
(GGT) = ( G G t ) (SaOH)
=
A(GG+)
- A(Hf)
(9) (10)
Here the total concentrations of Cu(I1) and GG are denoted by the subscript T and (NaOH) represents the formal concentration of NaOH added. In equation 10, the symbol A indicates the difference in concentration of the substance in the presence of the Cu(I1) salt and NaOH and in the absence of the Cu(I1) salt and NaOH. In the absence of added Cu(I1) salt and NaOH the concentrations of GG-, GGf and H + are negligible. In the experimental (14) H. Dobbie and W. 0. Kermack, ibid., 69,246 (1955). (15) A. Neuberger, Proc. Roy. Soc. ( L o n d o n ) , A168, 68 (1937).
Jan. 5, 1930
1LEACTION O F
GLYCINE-CONTAINING DIPEPTIDES
1
12
5.20
-
5.10
-
4 00
0
2
P
5.00
3.90
a
- -+
4.50.
1
2
3
4
Moles NaOH reacted per mole Cu(I1). Fig. 1.-Titrations of mixtures of Cu(1I) and dipeptides: curve 1, CuC12 0 01 M , GS 0.01 N ; curve 2, CuC12 0.01 M , GS 0.02 M ; curve 3, CuC12 0.01 df, GG 0.01 M ; curve 4, CuC12 0.01 M , G G 0.02 M . The arrow on curve 1 indicates onset of precipitation. The points shown on curve 3 are computed as described in the text.
solutions containing Cu(I1) the concentration of GG- is also negligible. Accordingly, equation 10 reduces to ( S a O H ) = (CuGG+)
+ 2(CuGG) - (GG+) -
( H + ) (11)
where (GG+) and (Hf) now represent the concentrations present in the experimental solutions. Introducing the expressions
and
and (14)
into equations 8, 9 and 11, we obtain on substitution and rearrangement the following series of expressions
(GG-) =
b c(H+)k*' -I_
+ 7a
[( re)(GGT) 1
(NaOH)
- (H+)]
(15)
(CuGG") =
(17)
I n a given solution a t a given pH all parameters on the right side of equations 15-17 are known or
0
-
2
0
@
e
4.90
0
0
B
t
235
IONS
WITH CUPRIC
I
-
e
*
*
e
e
3.00
"--i-----u,_ I
I
1
_-
236
WALTERL. KOLTUN, MELL,IX FRIED .\KD FRANK li. 5 . GURD
X KNO,. A value for gkr,o of 4.60 was found to yield a value of log kI,+ of 5.64 which was independent of pH over the range 4.0-5.0. Equilibria of 1: 1 Complexes at Higher p H Values.-Curve 3 of Fig. 1 shows that a 1:I mixture of CuClz and GG reacts with four moles of NaOH, two moles of which react above pH 7. The curve could not be fitted satisfactorily by assuming two constants k1,- and kI,= according to equations 5 and 7. rl satisfactory fit of the curve could be made, however, by postulating the existence of the equilibrium described in equation 6, and by choosing an appropriate value of kzI,-. The points shown above p H 8 on curve 3 in Fig. 1 are computed, using values of pkI,-, kzI,- and pkI,= of 9.37, 200 and 12.20, respectively. The same values were used to compute the form of the corresponding curve in a mixture containing twice as high initial concentrations of CuCll and GG (0.02 M ) ; the agreement with experiment was again very good. Similar results for 1:l mixtures of CuCL with SG, GV and VG are shown in Table 11. For GV and VG, the values were again found to fit the titration curves obtained with 0.01 iland 0.02 M Cu(I1)-dipeptide solutions. The titration of a mixture of 0.02 A f CuC12 and 0.02 -11PG consumed only three equivalents of NaOH up to p H 12, so that the value of pkr,, for this peptide must be considerably higher than for the other peptides, if indeed the process described in equation 7 occurs a t all. Table I1 shows that the values of pk1,- and log k21,- for the Cu(I1)-PG system fall in the same range as the other values listed for these constants.
Vol. b2
tures for the kinetic experiments. Sufficient NaCl was also added to maintain a final ionic strength near 0.16. T h e observed pH values are listed in the third column. The initial concentration of NPX was 1 X lo--' JI, and the first-order rate constants were computed from the rate of formation of N P - 10x1 as measured a t 400 mp. The observed first-order rate constants were corrected for the rate of hydrolysis of NPA alone under comparable conditions of pH, temperature and ionic strength.6 The corrected first-order rate constants are listed in the fourth column of Table 111. Computed values for the equilibrium concentrations of the three species CuGG, CuGGOH- and (CuGG)2OH- are listed in columns 5 , 6 and 7 of Table 111. The computations were made by combining equations 5 and G with the expression for the sum of the concentrations of Cu(I1)-GG complexes present and solving the resulting quadratic in (CuGGOH-). The values of pk1,- and log k.1, were taken from Table I1 and the pH values from column 3 of Table 111. Since the PIX values of the mixtures were between 7.20 and 1003, the concentrations of the forms CuGG and CuGG(OH)2121, and were computed to be a t most 3 X these forms were neglected. The eighth columii of Table 111 shows the secondorder rate constants ki computed on the assumption that the sole reactive species of the Cu(I1)-GG system under these conditions is CuGGOH-. The values of k2 were obtained by dividing the values for the first-order rate constants (column 4) by the concentrations of CuGGOH- (column 6). The average value is 12 9 1. mole-' 11ii11.-~and the average deviation 1.6 1. mole-' min -I. By conTABLE I1 trast, a similar computation to test the dependence & Q C I L I U R I U M C O N S T A N T S FOR F O R b l A T I O S O F cU(11)-DIof K 1 on ((CUGG)~OK-),in which the values iii PEPTIDE COMPLEXES column 4 are divided by- those in column 7, gives Temperature 25.0' ; ionic strength 0.16 log log log log values which range from 0.6 to 56 1. mole-' min.-' Peptide ki,, pk1.o k n kI1,o pki,- kzr,- pkr,and also show a strong inverse relation with the Glycylglycine 4.96 3.90 3.07 a 9 . 3 7 2 . 3 0 1'2 2 total concentration of the Cu(11)--GG system. Glycylsarcosine 6.13 a 4.62 a a a Sarcosylglycine 1.39 3 . 4 5 2 . 4 2 a 9 . 1 9 1 48 11 9 The results of these tests lend support to the asGlycyl-r-proline 6 . 4 3 a a 5.03 a a sumption that CuGGOH- is a much more reactive L-Prolylglycine 6.39 3.93 2.66 a 9 34 1 . 8 8 species than (CuGG)?OH-. The powerful basic Glycyl-r-valine 5.62 4 . 7 5 2.94 a 0.30 2.18 1 2 . 0 catalysis by hydroxyl ion limits the usefulness of ~-Valylglycine 4.87 3 . e 3 2 . 9 0 a 0.1: 2 . 2 4 11.8 the present technique above pH 10. and no attempt a Reaction not detected. Catalytic Activity of CuGGOH-.-Either or both was made to detect a kinetic effect due to the the basic complexes, CuGGOH- and (CuGG)2- species CuGG(OH),= Except in the most dilute mixtures studied in the OH-, may be expected to show a measurable rate of nucleophilic attack on NPX. The rate of experiments reported in 'Fable 111, the proportioli splitting of NP=1 was measured in fifteen mixtures of NaOH added in the final preparation of the exof the compositions described in Table 111. The perimental solutioiis (cnlumn 2'1 was much less mixtures were prepared using a single stock solu- than that which went into the preparation of the tion of CuClp, GG and NaOH mixed in the rnolar stock CuGG solutioii (columrl 1). It was for this proportions 1:1:2. It was found that dilutions of reason that the single stock solution was used this mixture alone, in which the form CuGG is to minimize iandom errors Since the stock s o h almost exclusively present, caused negligible in- tion was prepared by iilixing standard solutions creases in the rate of hydrolysis of NPA compared of CuClp, GG and NaOR, any imbalance in the to control solutions of the same pH and ionic ingredients is introduwd into all experiments. strength. Table 111 shows, however, that the rate An estiniate of this inihalaiicc is given by a conio f hydrolysis was increased markedly by addition Inrisoii of the sziiii o f thc. coiiccntrdtions given in of further NaOH. The first column lists various coluttiris G and 7 of Table I11 wit11 that in columu concentrations of total CuGG representing dif- 2 . The iiiib,tlaiicc in,~ybe sccn to be thc equivalent ferent dilutions of the stock solution. The second of an error of betwccn 2 and 362 in the ineasurecoluiiin shows total concentrations of NaOH added ment of the NaOH used to prepare the stock s o h to the stock dilutions i n the preparation of the mix- tion. lL
Jan. 5, 1960
KEACTION OF
GLYCINE-CONTAINING
DIPEPTIDES WITH
237
CUPRIC IONS
TABLE I11 DECOMPOSITION OF NPA IN BASICCu(I1)-GG SOLUTIONS Total CuGG, mole/l.
Added NaOH, mole/l.
0.1010 ,1010 .lolo ,0505 ,0505 ,0505 ,0505 I0202 ,0202 .0202 .0202 ,0101 ,0101 .0101 .0101
0.00434
,00651 ,00868 ,00217 ,00434 ,00651 ,00868 .00217 ,00434 ,00651 ,00868 ,00217 .00434 .00651
.00868
x
kl
102,
fiH
min. -1
7.20 7.32 7.48 7.47 7.69 7.98 8.16 7.99 8.56 8.88 9.26 8.66 9.17 9.69 10.03
0.670 0.691 1.28 0.459 0.966 1.46 2.32 0.92 2.02 3.95 5,84 1.43 3.60 7.05 8.35
--Computed CuGG
x
101
82.1 78.4 72.8 41.4 37.8 32.3 28.5 15.5 11.0 7.66 5.70 6.05 3.88 2 09 1.28
equilibrium concn., mole/l.CuGGOH(CuGG)zOHx 103 X 10'
9 11 13 4 5 8 10 2 3 5
0.555 ,700 938 ,521 ,789 1.31 1.76 0.647 1.70 2.48 4.42 1.18 2.44 4.36 5.84
Ln, 1. mole-1 min.-l
16 0 6 31 96 46 1 01
12 1 9 9 13 7 8 8 12 2 11 1 13 2 14 1 11 9 16 0 13 2 12 1 14 8 1G 2 14 3 12 9
77
03 5 04 0 935 1 89 1 83 1 49
A V kz
That CuGGOH- acts primarily as a catalyst for the hydrolysis of NPX instead of being consumed in the reaction has been shown by measuring the rate of the reaction during the consumption of a clear excess of NPA. Because of limitations of solubility the reaction was carried out in 40y0 ethanol by volume and in four stages. I n the first stage the reaction was begun by mixing 2.0 ml. of a solution of 0.02 M NPX in absolute ethanol with 3.0 ml. of a solution made up from CuC12, GG, NaOH and NaCI. The QH of the latter solution before the addition of the NPA solution was 8.42. At the moment of mixing the composition of the solution was as follows: CuClz 0.0200 M , GG 0.0200 M , NaOH 0.04445 M and NPA 0.0080 Jf. The ionic strength a t all stages was maintained at about 0.16 by the inclusion of NaCl in the reagents. 0.1" The temperature was maintained at 25.0 and the titration vessel was blanketed with a stream of N2 bubbled through distilled water. Within one minutc after mixing the p H reading reached 8.48 i. 0.01 and was maintained thereafter a t that value by addition by hand from a microburet of a solution 0.1011 N with respect to NaOH and 0.06 M with respect to NaC1. Buret readings were recorded at intervals until most of the reaction had occurred. The reaction mixture was then removed from the vessel and stored tightly stoppered a t 25" overnight, On the succeeding day 3.0 ml. of the reaction mixture was withdrawn and returned to the titration vessel. The pH was adjusted from 7.50 to 8.42 with the NaOH solution before the addition of 2.0 ml of 0.020 M NPA in 40% ethanol containing 0.16 Ad NaC1. The p H of the new reaction mixture was again maintained a t 8.48 and the buret readings recorded. The same procedure was followed through two more stages in an identical fashion, with the total concentration of Cu(I1) and GG in each stage somewhat less than 60% of that in the preceding stage. Each stage was matched by a control experiment in which Cu(I1) and GG were omitted but NPX and N P - were added in the concentrations present in the experimental solutions. The buret readings in the control expcriineiits were subtracted from the values
*
TABLE IV CATALSSIS OF HPDROLYSIS OF SP,\BY CuGGOH Ethanol 4070, apparent PH 8.48, ionic strength 0.16, temp. 25.0"
1 2 3 4
0 0200 0109
00595 00324
0 800 49% 300 176
0 340
5 7 9 7
296 220 100
A\
3 5 2 1
k2
7 3
obtained with the Cu(11)-GG system to yield the values ascribable to the CuGGOH- catalysis proper. This procedure allowed for a possible small degree of catalysis due to the accumulation of XP- or any effects due to the absorption of traces of COz. The corrections were of the same order of magnitude in each stage; by the fourth stage they amounted to about of the total experimental readings. Since the total consumption of NPA in the four stages that was attributable to the action of the Cu(I1)-GG system was greater than the total quantity of the latter system present, it was concluded that the role of CuGGOH- was largely catalytic. A quantitative estimate of the catalytic activity was made from the initial slopes of the plots of the rate of NaOH consumption against time in minutes. The number of moles of NaOH consumed per mole of KPX hydrolyzed a t pH 8.48 was taken as 1.92, where 1.00 mole is ascribed to neutralization of acetic acid and 0.92 mole to neutralization of N P which was found to have a pK' of 7.44 in the solvent containing 40y0 ethanol and 0.16 M NaCl a t 25'. From this value the initial rate of hydrolysis of NPX was computed and set equal to k2(NPA)(CuCrGOH-). In each rase the initial value of NPX was taken as 0.0080 -11 and (CuGGOH-) was computed using values for pk1,of 9.66 and for k21,- of 400 which had been derived from titration data obtained in the same solvent. The results are shown in Table IV. The average value o f k2, 7 . 3 1. inole-' 1111n - I , agrees with the
238
WALTERL. KOLTUN, MELVINFRIED AND FRANK R. N. GUKD
Vol. s 2
value of 7.5 1. mole-' min.-l measured separately about 0.015, the value of log kI,+ for acetylglyin the same solvent by the procedure described cylglycinate (AcGG-) and Cu(I1) ions was found above for the measurements reported in Table 111. to be 2.07; a t ionic strength 1.0, the value was 1.41. Although the $H values obtained in 40% ethanol The corresponding values for acetylglycinate cannot be compared directly with those obtained (AcG-) and Cu(I1) ions are 2.14 and 1.71. Both in 0.2% ethanol, the computations of k:! made in acetyl derivatives, therefore, show much the same this solvent are internally consistent and are not affinity for Cu(I1) ions as acetate, and the amide dependent on such a comparison. The lower groups do not appear to contribute to the stability value of k2 in the ethanol-water mixture has a of these complexes. In the case of AcGG- or parallel in the imidazole system.6 AcG- reaction 4 was not detected, presumably beEffects of Substituents on Complex Formation.cause the concentrations of ions employed, in the The evaluation of kI1,- according to equation 3 range of 0.01 M , were too low. The respective from titration of 2 : l mixtures of dipeptide and values of the association constants a t 20" of AcGGCu(I1) ion (Fig. 1, curve 4) has been d e ~ c r i b e d . ~and AcG- for hydrogen ions were found to be 3 66 The value of log k11,- for GG given in Table I1 is and 3.84 a t ionic strength near 0.015 and 3.50 taken from the preceding publication of this series and 3.68 a t ionic strength near 1.0. and is derived from kinetic measurements as well as For comparison with the substituted peptides, titration5 Other values in Table I1 are derived the complex formation of sarcosinate with Cu(I1) from titrations; in the case of PG, kinetic measure- ions was measured a t 20". The values of log kr,+ ments were also made as described below and con- and log kII.0 were found to be 8.08 and G.70 at firm the titration values. ionic strength 0.015 and 7.84 and 6.50 a t ionic In these 2 : l complexes the second peptide mole- strength 1.0. The values of log 4 ' and log k*' for cule appears to retain its peptide hydrogen atom, sarcosine a t ionic strength near 0.015 were 2.63 and a t least below pH 9. The values for log kII,10.31, respectively, and a t ionic strength 1.0, in Table IT suggest that substituents on the a- 2.64 and 10.28, respectively. For the Cu(1I)carbon atom of the N-terminal residue of the pep- glycinate system Flood and Lor%s17reported values tide may hinder the formation of the 2:l complex. of 8.22 and 6.97 for log kI,+ and log ku,o a t 20" For example, log k11,- for GV almost equals that and ionic strength 0.5. for GG, whereas the value for VG is distinctly Visible Spectral Characteristics.-The absorplower. At the same time all three peptides have tion spectra of mixtures of dipeptides with CUCIZ almost the same value of k k ' and VG and GG have and varying proportions of NaOH were measured very similar values for log kI,+ and pk1,o. A simi- in the Beckman Model DU Spectrophotometer lar type of effect may be apparent in the low values under the same conditions of temperature and of log k11,- for SG and PG compared to GG. Here ionic strength as were used to measure complex the hindrance may be due to the presence of sub- formation. The spectra of the CuGG form of stituents on the terminal nitrogen atom. complex were obtained directly from measureThe lower values of log k11,-for VG and SG rela- ments on solutions of the appropriate dipeptides tive to GG may represent the partial hindrance of mixed with Cu(I1) ions and NaOH in the molar reactivity by substitution. Much more marked proportions 1:1:2. The known values of kI,+ and effects are shown by GS and G P in which a hy- k1,o and the observed pH values were used to drogen ion cannot be displaced from the peptide compute the concentrations of the CuGG+ and bond to yield a structure of the form CuGG. CuGG forms of complexes in equimolar mixtures The use of such substituents to establish unequiv- of peptide and Cu(1I) ions containing less than two ocally the nature of reaction 2 was initiated by equivalents of NaOH ; from these computations Datta and Rabing and, independently, by two of the spectra of the CuGG+ forms were obtained. u s i 6 These peptides do not follow reactions 2 and The spectra of the CuGG2- complexes were meas3, but reaction 1 is succeeded directly by reaction ured by preparing solutions containing molar 4. The course of the titrations of GS in the total proportions of Cu(I1) ions, dipeptide and NaOH molar proportions relative to Cu(II) ion of 1:1 of 1:2 :3, respectively. The approximate wave and 2:l is shown in Fig. 1, curves 1 and 2. The lengths of maximum absorption A,, and correaction occurs in only two steps. The successive responding molar extinction coefficients Emax, are constants, k I + and kIr,o,were computed from the listed in Table V. titrations according to the procedure described for TABLE V the Zn(I1)-GG system5 or by a variant of the pro- THEABSORPTIONCHARACTERISTICS I N THE VISIBLEREGION cedure of successive approximations used for the OF INDIVIDUAL CU(I1 )-DIPEPTIDECOMPLEXES Cu(I1)-GG system in the present study. The CuGG CuGG CuGG€ma., ernas, values of log kI,+ and log k I I o in Table I1 for GS 1. mole-1 1 mole-', 1. mole-' and G P agree fairly well with the respective values A,X. cm.-' Amax cm-1 Am&= cm.-l Peptide found a t lower ionic strengths by Datta and 625 82 84 65 635 735 GG Rabing: 6.50 and 5.24 for GS; 6.66 and 5.24 for 645 105 100 34 635 735 PG GP. 625 87 95 27 640 760 GV Acetylation of the a-amino group to form acetyl620 91 92 100 635 715 VG glycylglycine (AcGG) markedly reduced the afThe results in Table V show general similarities finity for Cu(I1) ions. At 20' and ionic strength in the values of Amax for the individual types of (16) M . Fried and P.R. N. Gurd, Abst. 132nd National Meeting, +
Crnil.,
American Chemical Society, New York. N. Y.,Sept. 8-13, 1957, p 29C.
(17) H. Flood and (1945).
V. Loras, l'cdsskr. K j e n i Bergvercn Met.,
6, 83
Jan. 5, 1960
239
REACTION OF GLYCINE-CONTAINING DIPEPTIDESWITH CUPRICIONS
complexes. The values of both Xmax and Emax for the type CuGG+ are the most difficult to define precisely because solutions containing a preponderance of this form of complex cannot be prepared. The values listed for the Cu(I1)-GG complexes agree in a general way with those reported previously by Dobbie and Kermack.14 These authors showed also that the addition of more NaOH to solutions of CuGG did not cause large changes in the absorption spectra of the solutions. Similar measurements were made on the systems Cu(II)-GS and Cu(I1)-GP. For CUGS' Xmax was found to be 720 mp and emax 32; for CuGP+ the corresponding values were 745 mp and 32. For CuGSz Xmax was found to be 675 mp and Emax 45; for CuGPz the corresponding values were 665 and 50. The system Cu(I1)-AcGG formed faintly colored solutions. A weak maximum was observed near 760 mp. Various mixtures of complexes in the Cu(11)GG and Cu(I1)-GS systems were studied for their absorption characteristics in the ultraviolet region down to 210 mp. In each case an increase in absorption set in a t about 320 mp and continued to the lower limit of wave lengths measured. Measured separately, the CuClz and dipeptide solutions both exhibited end absorption below 240 mp. The sum of the separate absorptions, however, was less than that of comparable solutions containing complexes. Reaction of Dipeptides with NPA.-It has been shown that GG- may be acetylated by NPA.6 In the presence of a large excess of peptide the reaction may be studied under pseudo first-order conditions, and the measurement of the rate of release of p-nitrophenolate gives a measure of the concentration of GG- present. This procedure has been applied5 already to the measurement of the equilibrium described in equation 3 for GG. The second-order rate constants kz for the reaction of NPA with PG, GP, GG, GV and VG have been measured a t 25.1' and ionic strength 0.16. Similar measurements for glycinate, G-, and glycine ethyl ester, GEE, have been made for comparison. That k~ is related in general to the basicity of the peptide is indicated in Fig. 3, in which log k&' is plotted against log kz in the usual Bronsted plot. All the compounds except VG adhere quite closely to the relation log k2 = 0.84 log k*' -5.87. The value of kz for VG is, however, about onetenth of that predicted from this relation. The low rate of reaction between VG- and NPA has a parallel in the great resistance of D,L-valylglycine to hydrolysis in acid or alkali.l* Measurements of the kinetics of the reaction of PG- with NPA were made in the presence of Cu(I1) ions. The results of measurements on six mixtures were applied to the computation of kI1.- as described previously! The average value of log kI1,- was 2.75, with an average deviation of 0.22. This value agrees well with the value of 2.71 obtained using the pH values which were measured as part of the kinetic studies rather than the kinetic measurements themselves. It is in fair agreement (18) R. L. M . Synge, Biochem. J . , 89, 351 (1945).
10.0
9
9.0
7.0
I
/kP I
1
G
,I
PG
1
t
I
/ 0
1.0 log
2.0
kp.
Fig. 3.-Relation between log kt and log k'*: VG, Lvalylglycine; GEE, glycine ethyl ester; GG, glycylglycine; GV, glycyl-L-valine; GP, glycyl-L-proline; PG, Lprolylglycine; G, glycine.
with the value of 2.66 obtained from the stepwise titration curve (Table 11). The agreement between the kinetic and equilibrium methods of measuring the concentration of PG- and k I I - , extends the demonstration of the compatibility of the two methods to an a-imino peptide. Mixed Complex with Imidazole.-A counterpart to reaction 3 was studied in the reaction between CuGG and imidazole, Im CuGG
+ Im
CuGGIm;
Mixtures of CuC12, GG, NaOH and Im were prepared to yield the concentrations listed in Table VI after admixture with NPA. Sufficient NaCl was included to maintain an ionic strength of 0.16. The solutions were mixed with NPA solution and the rate of release of NP- measured spectrophotometrically. The range of final pH was 5.89 to 7.08, in which the forms CuGG+ and CuGGOH- are relatively negligible for 1:l Cu(II)-GG mixtures (Fig. 1, curve 3; Table 11). The concentration of free basic imidazole, (Im), was computed from the kinetic measurements using the value of 31.3 1. mole-l min.-l for the specific second-order rate constant for imidazole under these conditions6 It was assumed that CuGG and CuGGIm do not react with NPA; the lack of participation of CuGG was confirmed experimentally. The results of the kinetic assay for according to (Im) were used to compute kCuGGlm equation 18. In this computation the value of (Im) was combined with the pH talue and that of pK' (7.08) to yield (HIm+). The value of (CuGGlm) was then taken as the difference between the total concentration of imidazole and the suiii of (Im) and (HInif). The value of (CuGG) in turn was obtained by difference. The results of these computations are shown in Table VI. The average value of log k C u G G I m is 3.85 with an average deviation of 0.10. The constancy of this value confirms the assumption that CuGGIm itself does not react with NPA a t a comparable rate. The affinity of CuGG for Im is so great that the lower limit of reliability of the kinetic measurement of (Im) was reached with nearly one-third of the
TABLE T.1 BETWEENCUGG A S D I M I D A Z O L E 'rota1 CUCIP0.0100 iM; total GG* 0.0100 M; temperature 25.1'; ionic strciigtli U.16 EQUILIBRIUM
Initial total molar concn. of Irn A-aOH
0 ,00504 0100s 01008 .0100s 01008 l)1008 (1201(i
.
0,0190 .0150 . 0170 1)18U Ul80 ,0190 I i100
iJ ti.40
hi
5.89 6.32 6,5U G 62 7 08 6.28
x
109,
min. -:
0 382 0.743 1.43 1.54 1.W
2.51 4.53
total CuGG in the form of CuGGIm. Xeasurernents at higher values of (Im) failed t o reveal any evidence for the formation of a coinplex containing two imidazole molecules, c.g., CuGG(1m)z.
I:ig
4 --MiJcw:Iar
niridel uf CuGG
The large affinity of CuGG for I m made possiblc the preparation of solutions in which essentially 311 the Cu(I1) was in the form CuGGIm. From measurements of the absorption spectra of such solutions, it was computed that the value of A,,, for the complex was 610 mp and of emax 94. Discussion The results of the analysis of the titration curves below pH 7 of eyuimolar mixtures of CuCI2 and
[Irn)
x
103
0.122 . "7 ,463 ,492 540 .S Y i
1.45
(Hlrn-) x 103
0 584 3 ti7
-') (j7 1,s; 1.56 1 79 '3 1-1
(CuGGIrn)
x 108 4 33
6.17 6.95 7.72
7 '3% 8 29 0 37
(CuGG)
x
103
3 67 3.83 3 03 2.28 -.>. 0" 1 71 0 13 .i~. log kcucc;im
I',K
kc.ocr,,,
3.80 3.83 3.69 3 s4 2 . xti
3.73 4.19 3.85
dipeptides confirm the occurrence of the reactions shown in equations 1 and 2 . Rabin'Q has pointed out that the values of log kI,+ for Cu(I1)-complexes with several compounds containing N - termin a1 glycine are linearly related to log kk'. Datta, Leberman and KabinZ0have extended these observations to series of compounds containing ITterminal leucyl and sarcosyl residues, For each family of compounds with the same N-terminal residue the linear relation between log kI,+ and log k*' holds, but results for the different families of compounds are not superimposable. In the present series the same conclusions appear to hold: a plot of log kr,+us. log ki' for GG, CS, G P and GV is linear, whereas the values for SG and PG do not obey the same relation. The values for VG appear to fit the N-terminal glycine series. The Cu(I1) atom in complexes of the type CuGG+ probably is associated with the a-aniino N atom and the carbonyl 0 atom of the peptide bond. It has been shown that the relation between log kI,+ and log k i ' for peptides containing N-termirial glycine can be extended to predict the observed value of log k ~ , +for glycine itself.i9 Similar extensions to leucine and sarcosine have been rcported.20 The successful inclusion of the conipounds GS and GP in the N-terminal glycine series gives further support to the suggestion that thc Cu(I1) ion is bonded to the carbonyl 0 atom of the peptide bond in complexes of the type CuGG+, particularly in view of the present dem:>nstratiori of the similarity of the absorption spectra of the forms CuGG+, CuGS+ and CuGP+. This similarity iii absorption spectra implies a comnion structure f o r both CuGG' and the complexes in which the peptide IT is substituted. By contrast, the failure of the Cu(1I)-GS and Cu(I1)-GP systems to form a complex of the type CuGG is stroiig evidence for the involvement of the peptide N in bonding to Cu(I1) in the latter type of complex. ?, molecular model of CuGG has been construct$ to a scale corresponding to a distance of 2.0 A . 2 i The N-Cu-N angle is close to 90" without distortion and the ring formed of c u , a-amino N , methylene C, peptide carbonyl C and peptide N is coplanar. Drawings of the model are shown in Fig. 4. A structure of this sort preserves resonance in the peptide bond. The stability (10) B. R . R a b i n , T Y ~ J FZuS r , ~ ' l / l ySoc., 52, 1131 (1 (20) S P. D a t t n , R . Lrberrnan and B. R . Rahin, h7af2tvc, 183, 71.j (145!1). . l C / ~t z' r ~ s t, 5 , 3!1!1 (l!l:b2); (21) .4, 1 1 . hI;itliiest~ria n d 11. li. \i'vl,li 1%.Scc,ulc,iidi, ibid., 6 , I,:l (19>:3).
Jan. 5, 1960
K E X C T I O N O F C\rLYCINE-CONTA41NINGDIPEPTIDES WITH
of the complex is remarkable, considering that the dissociation of a proton from the peptide bond is usually taken to be negligible at fiH values below 13 or 14. The present method of computing kr,+ and kr,n is equivalent to those employed by Dobbie and KermackI4and by Datta and R a b h 9 The values computed for Cu(I1)-GG at low ionic strength are in essential agreement with those of both these groups of authors. The present method of computation subjects the data to a rather severe and readily visualized test of internal consistency as illustrated in Fig. 2. The certainty with which the data of a given titration may be analyzed mathematically by this method is probably within f 0.05 unit in log kr,a. The over-all uncertainty is probably at least =t0.10 unit. In making these computations it is assumed that forms such as CuGGf+f and CuGGf + are not present in significant concentrations; these assumptions appear reasonable in view of the low values of kI+ found for Cu(11)-AcGG and the probable increase in acidity of the carboxyl group of GG when the complex CuGG+ is formed. The complex CuGGOH- attacks NPA to catalyze its hydrolysis. Numerically the rate of the hydrolysis is very close to the rate of the reaction in which copper-free GG- undergoes acetylation by NPA. Because of the great stability of the complexes in equimolar mixtures of Cu(I1) and GG, very little of the latter reaction appears to occur in the presence of the Cu(I1). The complex CuGGOH- may be written in several canonical forms, and i t will be interesting to compare its catalytic properties with those of other Cu(I1)-N chelates in the hydroxo form. The possible catalytic activities of other Cu(I1)-dipeptide complexes of the CuGGOH- form have not yet been measured. Certain similarities between the system studied here and those described by Courtney, et al., should be noted.22 The nucleophilic properties of CuGGOH- find expression as well in the combination with CuGG ( 2 2 ) K . C. Courtneq K L Gustafson, S J tvesterback, H Hyp tiainen, S. C . Chaberek, J r , and A E. Martell, THISJ O U R N A L , 79,3030
(1957)
CUPRIC IONS
241
described in equation 6. Presumably the two CuGG moieties are joined to a common hydroxide ion to form an olate complex. The preference of the dipeptide complexes to form a monolate complex instead of a diolate complex may reflect the tendency of the carboxylate group of the dipeptide to act as a ligand to the Cu(I1) ion. The possibility that the GG in CuGG should be looked upon as a tridentate ligand may explain the reluctance of a second GG to give up its peptide hydrogen atom when i t combines with CuGG. The studies reported here may serve to some degree as models for the behavior of the N-terminus of longer peptide chains such as are present in proteins. I t has been shown that the N-terminus n a y react with NP-4 either to form the N-acetyl derivative or, in the presence of Cu(I1) ion, to catalyze the hydrolysis of the NPA. Under appropriate conditions the N-terminus may be bridged by Cu(I1) ion to another N-terminus or to a sidechain imidazole group, Some of these properties have been shown to vary materially with the nature of the dipeptide model and it may be expected that N-termini in proteins may exhibit a similar range of behavior. Whether or not longer peptide sequences will differ markedly from dipeptides in these properties is under study. The high stability of complexes of the form. CuGG and their apparent tendency to combine preferentially with one additional ligand point to their probable usefulness as reagents which might form ternary complexes with proteins. The present model studies suggest that such complexes would bind both to N-terminal a-amino groups and to imidazole groups and would probably show a preference for the latter. Acknowledgments.-The authors are grateful for LOW . the advice and assistance of Dr. Barbara li; in preparing the molecular model. The technical assistance of Miss Reta Roth, Mr. George A. Reich, Mr. David Reifsnyder and Miss Shirley Light is gratefully acknowledged. SEW YORK,?;. Y. GAIXESVILLE, FLORIDA
