M. M. TAQUI KHANAND A, E. MARTELL
10
Vol. 66
can observe that for both L-PSt and H-PSt the plots for l/Mapp deviate from a straight line at, relatively low concentrations and increase sharply with further increase in co; this upward curvature co (g./IOO Mapp co fg./100 II!r&PP ml.) X 10-5 R.p.m. ml.) X 10-6 R.p.m. is stronger for H-PSt than for L-PSt. On the other L-PSt H-PSt hand, although the measurements were limited to 0.8574 0.78 6200 0.7515 1 29 6700 lower concentrations than in the sedimentation 0.97 6135 .6400 .6135 1.67 experiments, the light scattering plots show negli6276 6819 ,5120 2.38 .5222 1.07 6794 gible upward trend in the ranges studied. This dif6680 .3415 1.20 .4130 3.01 ference suggests that the third and higher terms in 6358 1.31 .1931 7025 .3074 3.41 6824 the expansion 19would be quite different from those .1763 1.32 7980 3.69 6530 .2515 in the expansion 20. It is intriguing to extend the 1.37 10080 ,0882 ,1988 3.92 6900 present theoretical treatment to higher terms, with .1495 4.01 6500 the hope of finding out the real cause of this difference. I n connection with this problem, one 4.30 7846 .lo20 may remark that Kegeles, et u L , ~have observed a Outer, et ~ 6 of 0. t h ~and~D e~s r e ~ xand , ~ ~of Oyama, similar and even stronger upward curvature in et a125 their l/Mappplot for the system polyvinyl chloride (M, = 4.8 X lo4)-tetrahydrofuran at 25'. If TABLE I1 the strong upward trend of l/Mapp as seen in these COMPARISON OF SEDIMENTATION AND LIGHT SCATTERING data is the general behavior, it will diminish to DATA^ some extent the practical value of the Archibald L-PSt H-PSt ultracentrifugal method for the study of polymer 1.43 X lo6 4.76 x 105 solutions, since such a property of the plot neceslM, 1.27 X lo6 4.90 X lo6 sarily leads to a less reliable determination of the 1.75 x 10-4 1.35 x 10-4 intercept and the initial slope. (c.g.s.) LS 1.92 x 10-4 1.34 x 10-4 &' Sed. Acknowledgments.-The authors wish to thank ,'/1 315 d. .. ., . Professor M.Horio, Department of Polymer Chemis(s2), = x-average square radius of gyration of solute. try, Kyoto University, Kyoto, for his encouragement and his continued interest in the course of this Finally, me wish to point out one important dif- investigation. Part of this work was supported ference between the plot for l/iWapp vs. co and that by- a grant-in-aid of research from the Ministry of for Kcolious. co. This is concerned with the upward Education. One of us (H.F.) is indebted to the curvatures of these plots. From Figs. 4 and 5 one National Science Foundation (Washington, D. C.) (23) P, Outer, C.I. Carr and B. H. Zimm, J . Chem. Phys., 18, 830 for a fund (G-12477) which made it possible for him (1950). to participate in this research project. The assis(24) J. 0 t h and V. Desreux, Bull. SOC. chim. Belges, 63,285 (1954). tance of Mr. A. Nakazawa in the sedimentation (25) T. Oyama, K. Kawahara and M. Ueda, J. Chem. Soc. J a p a n experiments is gratefully acknowledged. ( N i p p a n Kagaku Zasshi). 79, 727 (1958). DAT-4
TABLE I APPARENTRfOLECULAR WEIGHT FUNCTION OH'INITIAL CONCENTRATION
FOR
{ E. {
e..
AS A
.
(I
METAL CHELATES OF ADENOSINE TRIPHOSPHATEf BY M. M. TAQUI KHANA N D A. E. MARTELL Department of Chemistry, Clark University, Worcester, Mass. Received February IY, 196s
Cocz, NInt2, Zn+2, Mgc2, Caf2, Stability constants of the 1:l chelates of ATP with divalent metal ions, Cuf2, Sr+2 and Ba+2 are reported at 25' and 0.1ionic strength. The stabilities increase in the sequence: Ba < Sr < Ca < Rfg < Co < Mn < Zn < Ni < Cu. The first and second hydrolysis constants and dimerization constants of copper(I1)-ATP chelate are reported.
In view of the participation of metal ions in the biological functions of adenosine phosphates, it was decided to investigate the interaction of these ligands with various metal ions commonly found in biological systems. Recently several investig a t o r ~ ~ - *reported the formation constants of adenosine phosphates with biologically important metals, particularly Ca(I1) and Mg(II), by various techniques. The presence of normal and protonic
complexes of Ca(I1) and of Mg(I1) have been reported by Martell and Schwarzenbach6 and also by Smith and Alberty.6 Ion-exchange techniques mere used by Naninga? and Wallass for a number of metals. The important differences between the present and the previous work are the use of a new mathematical treatment of the data and extension of the measurements to metal ions less basic than those of the alkaline earths. In the
(1) This investigation was supported by a research grant, H-3246, from the National Heart Institute, Public Health Servioe. (2) K. Burton and H. A. Krebs, Biochemicol J., 65, 94 (1953). (3) V. DiStefano and W. F. Neuman, J . B i d . Chem., 800, 759
(5) -4. E.Martell and G. Schwarzenbach, Helu. Chim. Acta, 89, 653 (1956). ( 6 ) R. M. Smith and R. A. Alberty, J . Am. Chem. Soc., 78, 2376 (1956). (7) L. B. Naninga, J . Phys. Chem., 61, 1144 (1957). (8) E. Wallas, Acta Chem. Scand., 12, 528 (1958).
(1953). (4) N. Melohior, ibcd., 208, 615 (1954).
AJETAL CHELhTES OF L b E S O S I N E TRIPHOSPHATE
Jan., 1962
present paper the formation constants of normal and protoriic complexes of ATP” (adenosine triphosphate) are presented. The formation coilstants with ADP” (adenosine diphosphate) and adenosine 3- and 5-phosphates are now in progress and will be the subject of a future publication. Experimental The experimental method consisted of potentiometric titration of the disodium salt of ATP in the absence and presence of Ihe metal ion being investigated. The ionic strength was1 maintained approximately constant in a 1: 1 titration by use of a medium containing 0.1 M potassium nitrate and relatively low concentrations of ligand and metal ion. In 10:l titrations the ionic strength was adjusted t o 0.10 with KNOB. The experimental technique was essentially the same as the one reported previously.6 The electrode system was calibrated by direct titration of acetic acid, the obEerved pII meter reading being compared with the actual hydrogen ion concentration calculated from data tabulated by Harned and 0wen.Q The pH regions below 3.5 and above 10.5 were calibrated by measurements in HCl and KOlH solutions, respectively. Reagents.---A pure sample of ATP prepared by the Nutritional Biochemicals Corporation was employed in this work. I n order to avoid possible hydrolysis prior to potentiometric measurements, each sample was weighed out as a solid and added to the experimental solution just before the addition of base. The metal ion solutions were standardized volume1,rically by titration with the disodium salt of EDTA in the presence of suitable indicators as outlined by Schwarzenbach.*o Carbonate-free KOH was prepared by the method of Schwarzenbach and Biedermannll and was standardized by titration with potassium acid phthalate.
Calculations The acid dissociation constants for the disodium salt of ATP (HPA) were calculated by a direct algebraic method. The equilibria and the dissociation constants involved are
If only 1:1 mo.nonuclear complexes form during a titration of a 1 : l mixture of ligand and metal ion, the equilibria may be expressed as Wf+ HA3MHA=
Kz
[MHA-] [HA3-]
(3)
[MA2-] [hP+][A4-]
(4)
m2+]
=
Two other equilibria may be described as M2’
+ &A’-
RIHA-
MHA-
q? MA2-
[ MHA -1 [H +I
K1’ = [M2+][:H2A2--]
+H+
+ H+
(5) (6)
[MA2-] [H+] K2’ = [MHA-]
and K2 are related to K,’ and Kz’ by the simple relationship
K1
(9) H. S. Harned and B. B. Owen, “The Physical Chemistry of Electrolytic Solutions,” Reinhold Publ. Carp., New York, N. Y., 1958. (10) G. Schwarzenbach, “Complexometric Titrations,” Interscience Publishers, New York, N. Y., 1957. ( 1 1 ) G . Sohwarzenbaoh and W. Biedermann, Hela. Chim. Acta, 31, 331 (1948).
Kt’Kz’
Kt‘
Kt
=ki
11
Kz
kikz = ___
(7)
If T A represents the total concentration of the various ligand species, and Tn;r that of all the metal species, then TA = [HzA*-] [HAS-] [A4-] [ILIA~-I
+ + [RIHA-I + Tnr = [Mz+] + [MA2-] + [MHA-] +
The total amount of [H+] = [HA$-]
(8) (9)
titratable hydrogen
+ 2[a4-] + [MHA-] + 2[MA2-] -
+ [OH-]
UTA
(10)
where represents moles of base added per mole of ATP present. From equations 5 , 6, 8, 9 and 10 the unknowns [HzA2-], [HA3-] and [A4-] may be eliminated. [Mz+] may be expressed in terms of Kz’ as
(11)
+
where x = 2[H+I2/klk2 and y = [H+I2/klk2 [H+]/kz 1. Equation 11 has only two unthe values of TA, U T A , knowns, Kz’ and [Mz+]; [H+] and [OH]- being experimentally measured. A set of values for Kz‘ was assumed and the corwere calculated for a responding values of [&!Iz+] given point on a titration curve. These values of [&Iz++] then were substituted in (12) to get the corresponding values of K1’.
+
Plotting the various values of Kz‘ against the corresponding values of K1’ gave a curved line. For each point on the titration curve a different line is obtained. If there is a unique solution for the above equations, plots of Kz‘ versus KI’ give curved lines intersecting a t one point. For a particular metal ion the experiment was carried out a t three different concentrations and in each case practically ‘ pK1‘ were obtained. the same values of ~ K z and The average deviation in most of the cases was about 0.01 of a pK unit. The series of intersecting lines is shown in Fig. 1 for one of the three concentrations studied in the case of the magnesium(I1) ion. Similar sets of curves were obtained for other metals. From the values of pK1’ and pKz’ the formation constants K1 and KZwere calculated by means of the relationships given by equation 7. The method of calculation of the formation constants for solutions with ten parts of metal per part of ligand is the same as was carried out for the 1:1 titration. I n this case it was assumed that the concentration of free metal ion is the same as the concentration of total metal ion added. This assumption is justified by the fact that the formation constants of the chelate species generally is low, and in the presence of tenfold excess of the metal a very small percentage of metal ion is bound. However, a correction for the metal ion bound was applied with the use of the formation constants obtained from 1:1 titration curves. The various constants obtained are listed in Table I. The equation derived for the calculation of formation constants for the 10: 1 titrations is
12
M. M. TAQUI KHAXAND A. E. MARTELL 2.6
Vol. 66
tration of metal and the ligand ions, i t is probable that four species exist: a normal (I : 1) chelate [CuA2-], a monohydroxo compound [Cu(OH)A. J4- and Hz0I3-, a dihydroxo species [CU(OH)~A a dimer [Cu(OH)AI2". The solution equilibria may be defined by the equations Cue+ A4- J _ CuA2-
2.2
+
& 1..8 R
1.4
CuA2-
1.Q
2CuA'-
44 48 51 p&'. Fig. 1.-Sample calculation plot for Mg-ATP system (1:1) a t 5 X 10-3 M , 2 5 O , and p = 0.10 M (MNO,). 4
+ 2Hz0 I- CU(0H)zA'- + 2H + 2Hz0
(CU[OH]A)Z'-
+
+ 2H+
The amount and distribution of the various chelate species present under various conditions of pH and total concentration may be calculated from the above equilibria and from the relationship between the total copper content, the value of -log[H+] and the amount of hydroxide added. If TOHrepresents the hydroxide added (in concentration units) to a solution of CUA during the titration, then TOE [H+] - [OH-] = [Cu(OH)As-]
+
2[(Cu(OH]A)z'-j
+ + 2[Cu(OH)zA4-]
(20)
At neutralization values less than a = 2.5, pH(5.5-6.5), it was assumed that the concentration of the dihydroxo chelate is negligibly small. Equation 20 then may be simplified to
+ [H+] - [OH-]
TOE
1
2
3
4
a.
Fig. 2.-Potentiometric titration of Mg(I1) and Cu(I1) chelates of ATP in 0.1 M KNOs a t 25' with the following molar ratios of ligand to metal ion: A, ligand alone; B, Mg (1:l); C, Mg (1:lO); D, Cu (.l:l). E, Cu ( 1 : l O ) ; a = moles of base added per mole of ligand.
- [H+] + [OH-]]
K2'[(2 - U ) T A
IA4-1 =
+ +
-
[H+][(u - ~ ) T A [He] [H+l (Z - 2/) KZ'Z
- lOH-11
= [Cu(OH)A3-]
+
2 [(Cu[OH1A)z6-1 (21 1
The total concentration of the chelate species T c uis~ defined by the relationship T C ~ L [CuA2-]
+ [Cu(OB)A3-] + 2[(C~[0H]A)z'-l
(22)
Combination of equations 16, 18 and 21 gives the expression
Equation 23 is valid only when a dimer is formed, in which case the plot of [H+l (TOH [a+] [OH-])/[ CuA2-] versus [ CuA2-]/[ H + ] should The value of [Ad-], when substituted in the ex- give a straight line with a slope equal to ~ K ( [M OHIA), pression for K1', gives a relationship as before and the intercept a t [CuA2-]/[H+] = 0 equal between K1' and Kz' and the experimentally de- to I
