PHILIPBROOKSAND NORMAN DAVIDSON
211s
The absorption band of medium intensity a t 1308 cm.-' in the reagent, which is also probably due to C-0 vibrations, is affected to a small extent in the case of Pb(I1) and Cd(I1) chelates but is shifted to a marked degree in the chelates of the transition metals. This implies that the C-0 group vibrations in the metal chelates of 4-hydroxybenzothiazole are influenced to a slight extent by the mass of the metal in the chelate ring. The interaction of the C-0 group with the n-bonds in the chelate ring as well as with the delectrons of the metal atoms play a much more important part in determining the extent of the shifts that are observed in C-0 group vibrations as we go from the reagent to the metal
[COh"IR I U U r I O N S O .
2$YG F R O M
THE
Vol. 82
chelates. It was found that if the frequency of the band that appears between 1310 and 13% an.-' for the transition metal chelates of 4-hydroxybenzothiazole was plotted against the stability constant (log &) of the chelate, a linear relationship was obtained. This type of relationship has been found in a number of metal c ~ m p l e x e s ' ~and ~ ' ~confirms the assignments made for the C-0 vibrations in the 4-hydroxybenzothiazole chelates. Acknowledgment.-The authors are grateful for a grant from the U. S. Atomic Energy Commission in support of this work. (14) H. F. Holtzclaw, Jr., and J. P. Collman, THISJOURNAL,79, 3318 (1957).
PITTSBURGH 13, PENNWLVANIA
GATESAXD CRELLIN LABORATORIES O F CHEMISTRY, PASADENA, CALIFORSIA]
Mercury(I1) Complexes of Imidazole and Histidine BY
PHILIP
BROOKSAND NORMAN DAVIDSON
RECEIVED AUGUST 31, 1959 The association constants of Hg*I with imidazole and L-histidine have been calculated From potential measurements of a I-Ig, Hg" electrode in solutions containing the complexing agents. For imidazole, the main result is H g + + 2C3114N: Hg(C3H&il)zf+ = 1016.7 A{-2
+
solid substance, H g ( C 3 1 i a S ~ ) C 1 0 ~ ~ Hprecipitates :0, from solutions a t p H (C3H3SzHg-)n+". For histidine, the main results are: H g + + 2LHgL2 H g + + L HL HgL(HL)+ K = 1021.2 A I - 2 K = 1018.4 AJ-Z I\
+
+
> 4.
+
I t probably contains an infinite ch:iin:
Hg++
+ 2HL A'
Hg(HL)z++ 1015.0 A[-2
where L- = (C3H3N2)CHzCH( SH?)CO2- and I-IL = (C3H3N2)CHCH(NHa+)CO:-. Because of the tendency of H g + + to form two bonds in a linear configcration, chelate formation contributes only slightly to the ligand binding by Hg"+. There is also evidence for the complex HgL+, with a formation constant of M-' (from L-).
I niidazole and histidine complexes with metal ions are of intrinsic interest as a part of the chemis-
try of complex ions, as well as being of possible biological significance. Complex ion formation with a nuniber of metal ions has been studied.' There is evidence that the imidazole ring is an important binding site for metal ions in serum : ~ l b u r n i n . ~In~ ~ general, mercury(I1) has the greatest binding power for nitrogen ligands of any of the metal ions; the present investigation is concerned with the mercury(I1) complexes of imidazole and histidine. Experimental Imidazole, from the Eastman Rodak Co. (m.p. 88-90', Eastman grade) and the Aldrich Chemical Co., and L-histidine, cfp. grade (99.98y0 pure by nitrogen analysis) from the California Corporation for Biochemical Research were dissolved in doubly distilled water. Concentrations calculated from PH titrations and from the weight of solute taken agreed within 1yo. Imidazolium perchlorate solutions were prepared by adding standard perchloric acid t o imidazole solutions. Mercuric perchlorate solutions were prepared by dissolving reagent grade mercuric oxide in a known excess of 9 F perchloric acid and diluting. In all experiments sodium perchlorate was used to keep the ionic strength constant a t 0.15 M . (1) For a general review, see J. Bjerrum. G. Schwarzenbach, L. S i l k , "Stability Constants, P a r t I." Special Publication No. 6, T h e Chemical Society, T.ondon, 1957, Tables 26 and 207; see also t h e speciiic references in Table 111. (2) F. R. N. Gurd and D. S. Goodman, THISJ O U R N A L , 74, 070 (1932). ( 3 ) C . ' I a i i f ~ ~ ridb,d . , 7 4 , '21 I ( I X d ) .
The principal experimental data were the potentials of a mercury, mercury(I1) electrode as a function of ligand concentration and pH. The cell, in a water bath a t 27.0 & 0.1', contained a J-type mercury electrode in which the mcrcury surface could be renewed by overflow, a 1.5 F S a S O a salt bridge connecting t o a saturated calomel electrode and provision for titrating in reagents and for maintaining a nitrogen atmosphere. Potential differences were measured with a Leeds and Northrup IC-2 potentiometer and a 0.01 p amp ./mm. galvanometer; p H measurements were made with a Beckman model GS p H meter. In every case ample time was allowed for the system t o attain equilibrium before final measurements were made.
Results Imidazole Complexes.-As anticipated, the binding of mercury by imidazole is so great that, under practical conditions, the equilibrium Hg++
+ nIniH+ eHg(Im),++ + %€I+ (1)
lies far to the right except for rather acid solutions with p H 5 2 . Under these circumstances, the pH titration method for the study of complex ion formation is difficult. We have therefore used a potentiometric method. Insoluble mercury-imidazole compounds precipitate a t pH's > 4, so measurements were made in the pH 2-4 range. The predominant imidazole species is the imidazolium ion, ImH+. Equation 1 then represents the over-all chemical rexction. Tlic Nernst law e m bc writteti
MERCURY (11) COMPLEXES OF IMIDAZOLE AND HISTIDINE
May 5, 1960
For Eo', the potential of the hypothetical Hg, 1.00 M Hg++ electrode a t p = 0.15 M vs. the see. (saturated calomal electrode), we take - 0.607 v. It is inconvenient to measure this quantity because of the Hg, Hg++, Hgz++ equilibrium. Our choice is based on the values Eo (Hg++ vs. Hz) = -0.841 in 1 F HC104,4 Eo (Hg++ DS. H2) = - 0.854 v. ( p = O).6 It is t o be noted that the quantities E that we quote are oxidation-reduction potentials using the American convention; a Hg, Hg++ electrode is electrically positive with respect to s.c.e. For the equilibria Hg++
+ ImH+
HgIm++
+H+ =
Hg++
[ H + ][HgIm++] [Hg++][ImH+]
+ 2ImH+ J_ HgImp++ + 2 H + P2'
we have
Equation 2 can then be written as E = Eo'
RT
- -In 2T
[ZHg"]
+
Plots of the quantity (25/RT)E
+ In (I: Hg")
os. In [ImH+]/[H+J were linear with a slope of
1.95-2.00. This shows that the dominant term Kl'[ImH+]/[H+] B?'. in the power series 1 ( [ImH+]/[H+])2 is the quadratic term and the principal complex formed is Hg(Im)z++. Representative data analyzed in terms of the simple equation
+
+
E = Eo'
- 0.0298log
[Z Hg"]
TABLEI POTENTIAL MEASUREMENTS OF Hg", I m H + SOLUTIOSS
10-4 10-4 10-4 10-4
1.05 5.49 1.82 3.80 6.02
[Z ImH+] [ImH+] F M X 10- 0.0963 0.0055 X 10-9 .0960 ,0952 X 10-10 ,0954 .0946 X 10-1' ,0954 ,0946 X 10-1' ,0954 ,0946
2.53 2.46 2.54 2.58 2.50
X X X X
10-4 10-4 10-4 10-4
2.52 6.74 1.81 7.71
X X X X
10-7 10' 10-8 10-
,0160 ,0317 ,0642 .0972
2.55 2.54 2.53 2.51
X X X X
10-4 10-4 10-8 10-8
7.76 2.75 5.49 4.46
X X X X
10-8 10-8 1010-
,0147 .0696 .a692 .Of384 log 81' =
[X Hgf*]b Fc 2 . 0 3 3 . 8 5 X 10-4
-E,a v. 0.369 ,361 ,317 ,267 .243
2.21 2.91 3.77 4.17
3.84 3.82 3.82 3.82
X X X X
,411 ,394 ,377 .366
2.12 2.12 2.10 2.10
3.87 4.08 4.10 3.88
2.58 5.11 2.60 5.26 2.59 1.01 ,388 2 . 6 0 6 . 4 4
[H$++]
PH
,395 .352 ,361
Av.
[ H+]2[HgIrn2+'] = [Hg++][ImH+]*
2119
.0153 .0309 ,0634 ,0965
log 021
,0138 2 . 3 8 .0686 2 . 4 0 .0674 2 . 4 2 ,0567 2 . 4 6 2.50 f 0.06
As already emphasized, these are conventional oxidation-reduction potentials vs. s.c.e.; the electrode was electrically positive with respect to s.c.e. [ZHgII] is the total (formal) concentration of Hg", corrected for Hg2++. F and M arr formula weights per liter and moles per liter, respectively. a
tion and imidazole concentration shows that the principal species is Hg(Im)2++. On changing concentrations of reagents, the potentials usually were established to f 0.5 mv. within a few minutes, although occasionally the potentials drifted for 10-30 minutes before becoming constant. In general a good way to investigate the possible influence of the first complex, HgIm++ and the term K,'[ImH+]/[Hf] is to start from the equation
+
Plots of the function on the left of equation 9 vs. [ImH+]/ [H+]gave fairly good straight lines; how[ImH+] (7) ever, it was not possible to obtain points close are presented in Table I. In all cases from the enough to the origin to result in a reliable extrapoactual potential, we know that [ZHg"]/ [Hg++] 2 lation for K1'. The difficulty is essentially this. 400. The equilibrium constant for the reaction To emphasize the contribution to the binding by HgTm++, one should work a t low [ImH+]/[H+] Hg Hg++ Hgz++ (8) ratios. This decreases the ratio [I: Hg"]/[Hg++]; has been measured as 130.s We have assumed the amount of Hgsff according to equation 8 then that the system was a t equilibrium with respect becomes large relative to [I: Hg"]. Under these to this reaction, although we do not know that such circunlstances, the corrections involved in analyzing is the case. With this assumption, [HgIm2++! = the data are too great to permit reliable interpreta[ 2 Hg"] = [Z Hq] - [Hg,++] where [I: Hg] is tion. There was no evidence for the formation the total amount of mercury added as Hg(I1) in of Hg(Im)3++or Hg(Im)4++ even a t [TmH+] = the solution and [I: Hg"] is the total amount of O.lOMandpH4. Hg", !.e., after correction for the mercurous ion The pKa of imidazolium a t 27" and p = 0.15 14 formed. Then [ImH+] = TI: IrnH+] - 2[Hg- was measured by a pH titration to be 7.12. This Imz++]. The corrections for the amount of Hgz++ can be compared with Edsall's value7 of 7.11 a t present and for the amount of imidazole tied up as 23O, ,u = 0.16 M , and Tanford and Wagner's HgTml++ are small in all cases and negligible for valuesof 7.12 a t 25", p = 0.15 M . Since high ratios [ImH+]/[Z Hg], so errors in the corHg++ 2IrnH+ = HgImz++ + 2H+, rections are not very important. PZ' = 102." (10) The constancy of the calculated values of p2' in Table I with varying PH, mercury(I1) concentra- H g + + + 2Im = HgImp++, 0.0596 log {B2'
I
+
+
(4) E. H. Swift, "Introductory Quantitative Analysis," PrenticeHall, Inc., New York, N. Y.,1950, p. 509. (5) W. bf. Latimer, "Oxidation Potentials," Prentice-Hall, Inc., Englewood Cliffs. N . J., 1956, p. 179. (6) L. G. Sillen, A. Jonsson and I. Qvarfort, Acta Chem. Scand., 1, 461 (1947).
(7) J. T. Edsall, G. Felsenfeld, D. S. Goodman and F. R . N . Gurd, THISJOURNAL, 76, 3054 (1954). (8) C. Tanford and M. L. Wagner, ibid., 76, 434 (1953).
PHILIPBROOKSAND NORMAN DAVIDSON
2120
1701.
82
This corresponds approximately to the formula Hg(C104)(C3H3N2)*H20a L-Histidine.-The ionization equilibria of histidine are HpL+
~: -100O
HL
K,
+ H+
--f
fi,.-
1,-
+ 2H'
(12)
where
I
t
> w
-200
-400 2.0
'
I 3.0
H
I
'
'
4.0
' ' '
I 5.0
PH
6.0
" 7.0
'
8.0
I
'
3.0
Fig. 1.-Potential it$. PH for fixed ZHg" and histidine concentrations. These concentrations, [2HgTr] = 3.80 X F and [ZL] = 0.095 F, actually decreased slightly during the titration due to dilution; suitable corrections were made. A similar straight line, displaced as expected by the Nernst law, was obtained for [ZHgII] = 3.4 X F, [ZL] = 0.011 F. The solid line is calculated from equation 15, and the values log pzl) = 2.98, log 8 2 1 = 9.18, log /3zr = 14.99.
Insoluble Hg (C3H3N2) (C104).H20.-Both the imidazole and histidine complexes with mercury become insoluble as the fiH is raised. The imidazole mercury complex precipitates copiously a t pH's greater than 4.5, but the histidine complex is more soluble, solutions a t very alkaline pH's (ca. 12-13) being just slightly cloudy. To ca. 150 ml. of solution containing 10 mmoles of Ha", 45 mmoles of Im, 35 mmoles of H f , was added about 13 mmoles of NaOH. About half the mercury precipitated as a white precipitate and t.he p H was 6.1. The precipitate was dried in a vacuum desiccator over Cas04 for several days and then analyzed. The analysis for mercury was carried out in 6 N HCW4 with NaSCN by a Volhard titration. About 60% of the NaSCN was added prior to the titration to dissolve the complex. The analyses for carbon, hydrogen, nitrogen and chlorine were done commercially by Dr. A. Elek. Element
yo Present
Hg 51.12 c1 9.41 N 7.63 H 1.21 C 9.75 0 (by difference) 20.88
Relative molar amouut
1.00 1.04 2.13 4.71 3.18 5.08
H
Our measurements give pKa = 6.08, PKa' = 8.12; in good agreement with other results (AlbertJg 6.08, 9.20, 20°, I.L = 0.01 M ; Li and Manning.l0 6.05, 9.12,25',p = 0.15M). In a solution with constant ZHgl' concentration and a constant, rather large [ZL] (where [ZL] = [HzL+] [HL] [L-1) the potential was observed to be a linear function of pH over the p H range 2.5 to 9, with a slope of 0.060 v. per p H unit, as displayed in Fig. I. The linear relation is rather surprising since the transformation H2L+ F? H L H + has pKa = 6 0. I t is necessary to suppose therefore that the complexes also can be protonated and we propose the following equilibria. It is convenient for the analysis to write all equilibria terms of the species HL.
+
+
+
+ H L If HgL+ + 1% H g + + + HL HgHL-+ H g + + + 2HL I_ HgLz + 2H+ H g f C + 2HL I_ HgL(HL)+ + H + Hg++
{9,0
(1';)
$ 1
0.0 1391
4-2HL Hg(HL)s+ B.2 The measurements used for qudntitative interpretation were in the p H range 9-3, so that it is necessary to consider the three histidine species, L-, H L and H2Lf (We can neglect HaL++,which has a PK, of 1 8).g lye write 21, for the total uncomplexed histidine Hgf+
[ZL] = [L-1
+ IHLl + [H*L+]
understanding that any significant corrections for the amount complexed have been made (if necessary, by a successive approximation procedure). Then, from equation 12
(9) A. hlbert, Bioclrem. J . , S O , GSO (19.52). (IO) N. C . Li and R. A . l l a n n i n g . Tms J O U K X A L , 77, 5225 (1025).
MERCURY (11) COMPLEXES OF IMIDAZOLE AND HISTIDIKE
May 5, 1960
2121
TABLE I1 SLOPE AND INTERCEPT OF EQUATION 15b FOR EXPERIMENTS AT FIXED pH PH
Intercept' Slope*
5.00
6.08
5.20
7.00
6.20
4.93(& 1)X lo9 1.30(* 1.3) X 1Olo 6.90(&2) X 10" 4.87(&3) X 1011 2.42(& 1 . 5 ) X 1013 5.46(*0.1)X IO1* 1 . 8 4 ( f 0 . 1 ) X 1 0 1 * 1.42(1O.2)X1O16 2.46(&OO.2)X10'5 9.75(&0.1)X 10l6
Intercept = ' q ( p l o IH 1
+ Pll[H+]).
bSlope =
+ PzI[H+]4-P22[H+l2).
where the function f(H+) is defined by the second ments in the p H range 8-3; the data a t 9 and 2.5 equality. are also predicted fairly satisfactorily.) The total mercury(I1) concentration is given by the relation: [ZHgT1]= [Hg++] [HgL+] [HgHL++] [HgLzl [HgL(HL)+] [Hg4.0 k (HL)z++]. [HLl = [ZL]f(H+) from equation 14. The equilibria 13 then permit us to express the quantities [HgL+], [Hg(HL)z++], etc., in terms of [Hg++], [H+] and [ZL]; for example, o/o' [HgL(HL)+1 = Pzi [&?+I [~LIz(j"(H+))z/[H+l. By substitution of these relations into the equation for [ZHg"] we obtain
+
+
+
+ +
c
where and 6 2 are implicitly defined by comparison of (15a) and (15b). The term on the left of (15) was plotted os. [ZL] for experiments a t several fixed values of the PH with varying [EL]. One of the best plots is shown in Fig. 2. The values of the intercepts and slopes are collected in Table 11. Examination of the plots, such as Fig. 2, strongly suggssts that there is a significant contribution by the 81 term, but the intercept cannot be estimated accurately, as we have tried to suggest by the estimated uncertainties in the table. The two best intercepts fo_rp H 7 and 5 give 81 = 2.6 X lo6 a t pH 7.00 and j31 = 6.5 X l o e a t pH 5.00. According to our formulation, 81 should increase with decreasing pH. It is quite possible that there is another complex such as HgL(OH), but we believe the data are not sufficiently good to warrant such a conclusion. We provisionally conclude only that there is a complex with one histidine per mercury, HgL+, that probably Plo = lo6 and that in the 5 to 7 p H range there is not much of a contribution_by Hg(HL) ++. A plot of 8 2 vs. pH is shown in Fig. 3, I n addition to the results in Table 11, we have used the data of Fig. 1 where the fiH was varied a t a constant high Z L concentration; a t this high concentration, only the 8 2 term is important. The solid lines in Figs. 1 and 3 are calculated from the values p20 = 2.98 log Pzi(M-') 9.18 log P Z Z ( M - ~ ) 14.99
log
The good agreement justifies the interpretation given. Incidentally, since different EHg" concentrations were used in different experiments, the fact that the complexes are monomeric as regards mercury has been demonstrated. (The constants &, ~ ? Z Z were actually derived from the measure-
1.0
c
1
I
v
d o
i
1
I 1
2
3
4
5
6
7
[PL] x I03
Fig. 2.-Plot
pH
=
of eq. l5b for the determination of P1 and pz; 5.00 & 0.01, [ZHg"] = 1.4 X lo-* F.
As already remarked, the fact that the variation of potential with p H is quite linear with a slope of RT/S in the p H 3-8 range (Fig. 1) was not to be expected in general. Reference to equation 15a shows that this linearity requires that the function log {[fW+)I2(Pzo Pzi[H+I Pzi[H+I')J should be constant. For pH
