- Email: [email protected]
The origin of the intense absorption in azurin
BIOINoRGAiVIC-STRY
179
8,179-184 (1978)
The Origin of the Intense Absorption in Azurin
DAVID R. McMILLIN Department of Chemistry, Purdue University, W.Lafqette,
Indiana 47907
ABSTRACT The iatense visiile absorption band of native azurin and the corresponding transitions of the Ni(II) and the Co(R) derivatives were analyzed. The bands we= shown to be analogous ligand-to-metalcharge transfer transitions and the optical electronegativity of the &and involved is estimated to be 2.6. Comparisons with small molecule systems strongly indicate that the donor involved is cysteine sulfur.
INTRODUCTION The visible absorption spectrum of a native b&decopper protein is dominated by a rather intense absorption band with L,, in the neighborhood of 615 5000 iW-%mB1. Recent studies show that the transition is runandwithe” likely to be a low-lying ligand-to-metal charge transfer (LMCT) transition [I-S] . The most compelling evidence is (i) that more bands are observed in the visible and near infrared than can be accounted for as ligand field transitions [4, 61, (ii) that the transition is much more intense than would be expected for a ligand field transition, and (iii) that an analogous intense band is observed in the near UV for the Co(R) derivatives of several blue copper proteins [3, 5]_ More specifically, evidence is mounting that a ligand which is a sulfur donor is invokd in the LMCT transition. Both cysteine sulfur [l] and methionine sulfur [7] have been suggested as the donor involved. Chemical [8] and physical studies [9] of various blue proteins, model studies of small molecule systems of Cu(II) [lo] and Co(R) [ 10, 1 l] , and studies of an artificial copper protein [S] support the contention that cysteine sulfur is involved, although not aU experiments are in agreement [8c, 9b]. On the other hand, optical [7], redox [12], and resonance Raman [13] studies of complexes of sulfide donors suggest that metbionine sulfur may be involved. In this report the intense transition of the blue copper protein azurin and the corresponding transitions of several metal substituted derivatives of azurfn are analyzed in terms of optical electronegativities [ 141. The analysis places 0 ElsevierNorth-Holland. Inc, 1978
0006-3061-78-0008-0179301.25
D. R. McMILLIN
180
the LMCT assignment on a ftrmer basis and lends further support to the idea that a sulk
donor, probably from cysteine, is involved_
CALCUJLATlONS Equation (1) is an empirically arrived at expression which has been found to correlate the energies (in kK, 1kK = 1000 cm-l) of LhKT transitions for a variety of transition ion complexes [ 14]_ The parameters a and _a are, respectively, the optical electronegativity of the ligand and the
%orr=30(?&
- XalI)
(1)
optical efectronegativity of the metal ion involved in the excitation_ The expression is used to calculate the enerv of the (hypothetical) transition [Core] oL 2dQ + [Core] oL%iQ+l in which an electron is excited from a ligand orbital to a d orbital of the metal ion. The quantity i3,,,, is obtained from the experimentally observed transition energy Fob+ by correcting for certain interelectron repulsion effects and &and field effects [ 143 _The correction which is associated with the interelectron repulsions involves the spin-pairing energy, denoted SPE. The SPE reflects part of the interelectron forces acting among the d electrons on the metal and is basically characteristic of the particular d configuration and the net spin associated with the d electrons. It is only indirectly affected by the Iigands through B, the effective Racah parameter of the system [ 141, and as the ligands affect the spin state of the metal. The SPE for a dn configuration with spin S is given by Equation (2)
D=7B_
(2b)
To correct for the spin pairing energy, the change in the spin-pairing energy A(SPE) in going from the dq configuration of the ground state to the dQ+l configuration of the LMCT state is subtracted from ZObs- If the system has tetrahedral symmetry about the metal ion and the LMCT transition terminates in the e sublevel of the 3d orbitals, this is the only correction applied and is
corr =Vobs
-
A(SPE) .
If the transition terminates in the t2 sublevels, a correction splitting 4 of the 3d orbitals is applied [ 141 in which case At. &or* =Fobs---A(SPE)+
for the ligand field
SHORT COMMUNICATIONS
181 TABLE 1
QuantitiesQ Used to Calculate XL System clr(II)Az
Ni(IMz co(IDAz Mn(II)Az
%bs
A(SPE)
16.0 22.?d 32.3 f 42.6 h
0 3.3 6.8 14.3
B 0_704e 0.73og 0.765’
At -8_‘2c -59e 4.98 -
v,,,, 7.8 13.5 20.6 28.3
XMb
2.4 2.1 1.9 cl.8
a All numbers, except the XM, are in units of kK. b Data from [14] _ c J. W. Hare, Ph.D. Thesis, California Institute of Technology (1976). d Data from [ 16al_ e The ratio of the effective B parameter and that of the free Ni(II) ion was assumed to be the same as that calculated for Co(II)Az. At was estimated from the ratio of At’s found for Ni(II) and Co(H) in the tetrachioro complexes [ 151 and-the A, calculated for Cc(II)Az. For a justification of these empirical procedures see C. K. Jorgensen, Absorption Spectra and Chemical Bonding in Complexes. Pergamon Press, London (1962), p_ 138 and p. 113. f Data from [3] _ g E. I. Solomon, J. Rawlings, D_ R_ McMiIIin, P. J. Stephens, and H. B. Gray, J. Am. Chem. Sot., 98,8046 (1976). h Datafrom [16bl. ’ Estimated from the B used for Ni(II)At and the empirical relationship of Jorgensen, in C. K. Jorgensen. op. cit., p. 138.
In Table 1 the data that wiII be used in conjunction presented. RESULTS
with Equations (l-2)
are
-AND DISCUSSION
For a particular Iigand environment a definitive procedure that can be used to identify a suspected LMCT transition is to observe the way in which the energy of the band shifts as different metal ions are incorporated into the ligand environment. As Jorgensen has shown [14], if the transition is a LMCT transition, it should be possible to associate an optical electronegativity with the ligand(s) involved which wiIl, in conjunction with the optical electronegativities of the metals involved and Eq. (l-2), predict the energy of the transition observed for each of the different metal ion systems. To date this approach has been of limited utility for the blue copper protein systems since only the Co(I1) and Cu(l1) derivatives have been characterized
D. R. McMILLIN
‘182 TABLE
2
Calculated Optical Electronegativities of the Azurin Ligand
Chromophore
talc
XL
Cu( II)Az
2.66
Ni(II)Az
2.55
Co(II)Az
2.59
[2, 3]_ Recently, however, we have prepared a series of metal-substituted derivatives of azurin from Pseuciomonas aerugritosa [ 161. In all cases we find absorption bands which appear to be anaIogous to the intense absorption bands which dominate the visible spectrum of native azurin. The energies of the intense absorption bands which are observed for the various derivatives of azurin are presented in TabIe I- Assuming the transitions are LMCT transitions and using the data in Table I and Equations (l-2), estimates of the optical electronegativity for the I@nd involved have been calculated and are presented in TabIe 2. The agreement between the ~~~~~~values for the Cu(i1). Ni(I1) and Co(B) derivatives of azurin is seen to be very good. The average value is 2.60. The close agreement between the values in Table 2 provides strong support for assigning the transitions as LMCT transitions. Moreover, these results and other observations [16] strongly suggest that the different metal ions bind to a common site of apoazurin. It is of interest of compare the value of xLcaic with the values characteristic of a thioether donor, as in methionine, and a thiolate donor, as in cysteine. Based on studies of an Ir(IV) system, an optical electronegativity of 2-9 has been assigned to thioether [17]. From the spectral data reported for bis[& methyhnercaptoethylamine] copper (II) diperchlorate [ 181 we calculate that xL = 2.82 for the thioether ligand- The MLCT transition identified in the h(W) system apparently originates in the nonbonding lone nair of the sulfur and terminates in a drr level of the metal [ 191, whereas the xL = 282 value is calculated from a transition that originates in the sigma bonding lone pair of rhe sulfur. 1181 Since the intense transition in the azurin derivatives originates in a sigma bonding Ione pair [3], the due of 2.82 is probably the better value to colnpare with X&Cal=_ We can estimate xL for cysteine sulfur from the spectral data of oxidized spinach ferredoxin Assuming the transition at 540 m is Rs + Fe(U) and assuming B = 0.7 kK 1203, we calculate xL = 2.64 for cysteine sulfur_ The agreement between xLearc = 2.60 and the optical electronegativity of cysteine sulfur is Striking and suggests that cysteine sulfur is involved in the
183
SHORT COMMUNICATIONS
metal binding site however, two complications should be considered. One is that the XL value of a given ligand might be expected to vary somewhat from system to system because of nonbonded repulsions with other ligands, etc. The evidence is, however, that optical electronegativities successfully correlate data from a large number of quite different systems and appear to be characteristic properties of the moieties involved [14] _ A second complication is the fact that the systems under consideration do not involve rigorous tetrahedral symmetry [3,4]. It is difficult to generalize on this point, but in general one would expect these effects to be small. To cite an example, the LMCT transitions of CuCl,aare rather insensitive to changes in the coordination geometry 1213 _ On the basis of the results discussed above we conclude that the intense transition observed at 626 nm for azurin is a LMCT transition that probably originates from a cysteine sulfur, although methionine sulfur cannot be ruled out, unequivocally1 _ Moreover a corresponding transition has been identified in the Co(H) and Ni(II) derivatives of the protein A final point concerns the optical electronegativity of Mn(II), for which only an upper limit has been reported. If we assume the transition observed at 230 nm in Mn(II)Az is the analogous L.MCT transition that we see in the other derivatives of azurin, we can estimate the optical electro-negativity for Mn(I1) to be I-66. This value appears reasonable and is consistent with the upper limit of 1.8 established by Jorgensen [ 14]_ NOTE ADDED IN PROOF: The reduced-minus-oxidized difference absorption spectra of various azurin derivatives have been found to give maxima in the region of -40 kR, making a LMCT assignment for the band reported for Mn(II)Az rather tentative_ See also I. Pecht, 0. Farver, and M. Goldberg, Advances in Chemistry Series 162,179 (1977). This lvork was supported
by (I grant from
the National
Institutes
of Health number
GM2276402.
REFERENCES 1. R. J. P. WiJliams,lnorg Chin Acra Rev. 5,137 (1971). D. R. McMillin, R. A. Holwerda, and H. B. Gray, Proc. NarZ. Acad. Sci. U.S.A. 71,
2.
3. 4.
1339 (1974)_ D. R. M&fiUi~, R. C. Rosenberg, and H. 3. Gray, ROC. NULLAcad_ SCL U.S.A. 71, 4760 (1974). E. I. Solomon, 3. W. Hare, and H. B. Gray, Proc. NarL Acad. Sci U.S.A. 73, 1389 (1976).
a It should be noted that methionine could not be involved in the case of the blue copper protein stellacyanin since methionine was not found in its amino acid analysis [22] _
184
D. R McMILLIN
L. Morpurgo, A. Finazzi-Agr6, G. Rotilio, and B. Mondo$ Eur. J_ Biochem 64,453 (1976). 6. K. E_ Falk and B. ReXmmmar, Biochim. Biophys. Acta 285,84 (1972). 7. T. E_ Jones, D. B. Rorabacher,and L. A. Ochrymowya,J. Am Chem Sot-. 97,7485 (1975)_ 8a_ A. Finazzi-Agr6, C. Giovagnoli, L_ Avigliano, G. Rotilio, and B. Mondo$ Eur J_ Biixhem 34,20 (1973)_ 8b. S. Katoh and A. Takarniya,J. Biochem (Tokyo) 55,378 (1964). SC. C. Brivin8and J. Deinum, FEBS Lett. 51,43 (1975). 9a_ E. I. Solomon. P. J. Clendening,H. B. Gray, and F- J. Grunthaner,J. Am C%em. Sot. 97,387s (1975). 9b. J. Peeling,B. G_ HasIett,I. M. Evans, D. J_ Clark, and D. Boulter,J. Am them Sot.. 99,102s (1977). 10. Y. Suguira, Y_ Hirayama, H. Tanaka, and K. Scbizu, J. Am Chem Sot. 99, 1581 (1977). 11. _D_htastropaolo.J. A_ Which,J_ A_ Potenza, and H. J. Schugar,J_Am Chem Sot_. 99, 424 (1977). 12 T. E_ Jones, D_ B. Rorabacher,and L. A_ Ochrymowycz, J_ Am them sOc_ 98,4322 (1976)_ 13. W_ H. Woodruff, privatecommunication_ 14_ C_ K. Jorgensen,hog_ Itwrg_Gzem 12,101(1970). 15. P. Day and C. K. Jorgensen,J. C%em Sot. 1964,6226. 16a. D. L. Tennent and D. R. McMiUin,to appear16b. D. L. Tennentand D. R. M&ill& to appear_ 17_ G_ B_ Kauffman, J_ Hwa-San Tsai, R. C_ Fay, and C. K_ Jorgensen.Inorg_ C?zem_2, 1233 (1963)_ 18. V. M. Miskowski, J. A- Thich, R- SoIomon, and H_ J_ Schugar,J. A.m them Sot 98, 8344 (1976). 19. M. D. Rowe, A. J. McCaffery. R. Gale, and D. N. Copsey, Inow Chem. 11, 3090 (1972)_ 20. J. Rawlings,0. Siiman. and H. B- Gray.&oc. NatL Acad. Sci USA. 71,125 (1974). 21. A. B_ P_ Lever.1. Chem Educ. 51,612 (1974). 22. J- Peisach,W- G- Levine,and W. E. Blumberg,J_ BtiL Chem 242,2847 (1967). 5.
Receiwd June 5.1977
179
8,179-184 (1978)
The Origin of the Intense Absorption in Azurin
DAVID R. McMILLIN Department of Chemistry, Purdue University, W.Lafqette,
Indiana 47907
ABSTRACT The iatense visiile absorption band of native azurin and the corresponding transitions of the Ni(II) and the Co(R) derivatives were analyzed. The bands we= shown to be analogous ligand-to-metalcharge transfer transitions and the optical electronegativity of the &and involved is estimated to be 2.6. Comparisons with small molecule systems strongly indicate that the donor involved is cysteine sulfur.
INTRODUCTION The visible absorption spectrum of a native b&decopper protein is dominated by a rather intense absorption band with L,, in the neighborhood of 615 5000 iW-%mB1. Recent studies show that the transition is runandwithe” likely to be a low-lying ligand-to-metal charge transfer (LMCT) transition [I-S] . The most compelling evidence is (i) that more bands are observed in the visible and near infrared than can be accounted for as ligand field transitions [4, 61, (ii) that the transition is much more intense than would be expected for a ligand field transition, and (iii) that an analogous intense band is observed in the near UV for the Co(R) derivatives of several blue copper proteins [3, 5]_ More specifically, evidence is mounting that a ligand which is a sulfur donor is invokd in the LMCT transition. Both cysteine sulfur [l] and methionine sulfur [7] have been suggested as the donor involved. Chemical [8] and physical studies [9] of various blue proteins, model studies of small molecule systems of Cu(II) [lo] and Co(R) [ 10, 1 l] , and studies of an artificial copper protein [S] support the contention that cysteine sulfur is involved, although not aU experiments are in agreement [8c, 9b]. On the other hand, optical [7], redox [12], and resonance Raman [13] studies of complexes of sulfide donors suggest that metbionine sulfur may be involved. In this report the intense transition of the blue copper protein azurin and the corresponding transitions of several metal substituted derivatives of azurfn are analyzed in terms of optical electronegativities [ 141. The analysis places 0 ElsevierNorth-Holland. Inc, 1978
0006-3061-78-0008-0179301.25
D. R. McMILLIN
180
the LMCT assignment on a ftrmer basis and lends further support to the idea that a sulk
donor, probably from cysteine, is involved_
CALCUJLATlONS Equation (1) is an empirically arrived at expression which has been found to correlate the energies (in kK, 1kK = 1000 cm-l) of LhKT transitions for a variety of transition ion complexes [ 14]_ The parameters a and _a are, respectively, the optical electronegativity of the ligand and the
%orr=30(?&
- XalI)
(1)
optical efectronegativity of the metal ion involved in the excitation_ The expression is used to calculate the enerv of the (hypothetical) transition [Core] oL 2dQ + [Core] oL%iQ+l in which an electron is excited from a ligand orbital to a d orbital of the metal ion. The quantity i3,,,, is obtained from the experimentally observed transition energy Fob+ by correcting for certain interelectron repulsion effects and &and field effects [ 143 _The correction which is associated with the interelectron repulsions involves the spin-pairing energy, denoted SPE. The SPE reflects part of the interelectron forces acting among the d electrons on the metal and is basically characteristic of the particular d configuration and the net spin associated with the d electrons. It is only indirectly affected by the Iigands through B, the effective Racah parameter of the system [ 141, and as the ligands affect the spin state of the metal. The SPE for a dn configuration with spin S is given by Equation (2)
D=7B_
(2b)
To correct for the spin pairing energy, the change in the spin-pairing energy A(SPE) in going from the dq configuration of the ground state to the dQ+l configuration of the LMCT state is subtracted from ZObs- If the system has tetrahedral symmetry about the metal ion and the LMCT transition terminates in the e sublevel of the 3d orbitals, this is the only correction applied and is
corr =Vobs
-
A(SPE) .
If the transition terminates in the t2 sublevels, a correction splitting 4 of the 3d orbitals is applied [ 141 in which case At. &or* =Fobs---A(SPE)+
for the ligand field
SHORT COMMUNICATIONS
181 TABLE 1
QuantitiesQ Used to Calculate XL System clr(II)Az
Ni(IMz co(IDAz Mn(II)Az
%bs
A(SPE)
16.0 22.?d 32.3 f 42.6 h
0 3.3 6.8 14.3
B 0_704e 0.73og 0.765’
At -8_‘2c -59e 4.98 -
v,,,, 7.8 13.5 20.6 28.3
XMb
2.4 2.1 1.9 cl.8
a All numbers, except the XM, are in units of kK. b Data from [14] _ c J. W. Hare, Ph.D. Thesis, California Institute of Technology (1976). d Data from [ 16al_ e The ratio of the effective B parameter and that of the free Ni(II) ion was assumed to be the same as that calculated for Co(II)Az. At was estimated from the ratio of At’s found for Ni(II) and Co(H) in the tetrachioro complexes [ 151 and-the A, calculated for Cc(II)Az. For a justification of these empirical procedures see C. K. Jorgensen, Absorption Spectra and Chemical Bonding in Complexes. Pergamon Press, London (1962), p_ 138 and p. 113. f Data from [3] _ g E. I. Solomon, J. Rawlings, D_ R_ McMiIIin, P. J. Stephens, and H. B. Gray, J. Am. Chem. Sot., 98,8046 (1976). h Datafrom [16bl. ’ Estimated from the B used for Ni(II)At and the empirical relationship of Jorgensen, in C. K. Jorgensen. op. cit., p. 138.
In Table 1 the data that wiII be used in conjunction presented. RESULTS
with Equations (l-2)
are
-AND DISCUSSION
For a particular Iigand environment a definitive procedure that can be used to identify a suspected LMCT transition is to observe the way in which the energy of the band shifts as different metal ions are incorporated into the ligand environment. As Jorgensen has shown [14], if the transition is a LMCT transition, it should be possible to associate an optical electronegativity with the ligand(s) involved which wiIl, in conjunction with the optical electronegativities of the metals involved and Eq. (l-2), predict the energy of the transition observed for each of the different metal ion systems. To date this approach has been of limited utility for the blue copper protein systems since only the Co(I1) and Cu(l1) derivatives have been characterized
D. R. McMILLIN
‘182 TABLE
2
Calculated Optical Electronegativities of the Azurin Ligand
Chromophore
talc
XL
Cu( II)Az
2.66
Ni(II)Az
2.55
Co(II)Az
2.59
[2, 3]_ Recently, however, we have prepared a series of metal-substituted derivatives of azurin from Pseuciomonas aerugritosa [ 161. In all cases we find absorption bands which appear to be anaIogous to the intense absorption bands which dominate the visible spectrum of native azurin. The energies of the intense absorption bands which are observed for the various derivatives of azurin are presented in TabIe I- Assuming the transitions are LMCT transitions and using the data in Table I and Equations (l-2), estimates of the optical electronegativity for the I@nd involved have been calculated and are presented in TabIe 2. The agreement between the ~~~~~~values for the Cu(i1). Ni(I1) and Co(B) derivatives of azurin is seen to be very good. The average value is 2.60. The close agreement between the values in Table 2 provides strong support for assigning the transitions as LMCT transitions. Moreover, these results and other observations [16] strongly suggest that the different metal ions bind to a common site of apoazurin. It is of interest of compare the value of xLcaic with the values characteristic of a thioether donor, as in methionine, and a thiolate donor, as in cysteine. Based on studies of an Ir(IV) system, an optical electronegativity of 2-9 has been assigned to thioether [17]. From the spectral data reported for bis[& methyhnercaptoethylamine] copper (II) diperchlorate [ 181 we calculate that xL = 2.82 for the thioether ligand- The MLCT transition identified in the h(W) system apparently originates in the nonbonding lone nair of the sulfur and terminates in a drr level of the metal [ 191, whereas the xL = 282 value is calculated from a transition that originates in the sigma bonding lone pair of rhe sulfur. 1181 Since the intense transition in the azurin derivatives originates in a sigma bonding Ione pair [3], the due of 2.82 is probably the better value to colnpare with X&Cal=_ We can estimate xL for cysteine sulfur from the spectral data of oxidized spinach ferredoxin Assuming the transition at 540 m is Rs + Fe(U) and assuming B = 0.7 kK 1203, we calculate xL = 2.64 for cysteine sulfur_ The agreement between xLearc = 2.60 and the optical electronegativity of cysteine sulfur is Striking and suggests that cysteine sulfur is involved in the
183
SHORT COMMUNICATIONS
metal binding site however, two complications should be considered. One is that the XL value of a given ligand might be expected to vary somewhat from system to system because of nonbonded repulsions with other ligands, etc. The evidence is, however, that optical electronegativities successfully correlate data from a large number of quite different systems and appear to be characteristic properties of the moieties involved [14] _ A second complication is the fact that the systems under consideration do not involve rigorous tetrahedral symmetry [3,4]. It is difficult to generalize on this point, but in general one would expect these effects to be small. To cite an example, the LMCT transitions of CuCl,aare rather insensitive to changes in the coordination geometry 1213 _ On the basis of the results discussed above we conclude that the intense transition observed at 626 nm for azurin is a LMCT transition that probably originates from a cysteine sulfur, although methionine sulfur cannot be ruled out, unequivocally1 _ Moreover a corresponding transition has been identified in the Co(H) and Ni(II) derivatives of the protein A final point concerns the optical electronegativity of Mn(II), for which only an upper limit has been reported. If we assume the transition observed at 230 nm in Mn(II)Az is the analogous L.MCT transition that we see in the other derivatives of azurin, we can estimate the optical electro-negativity for Mn(I1) to be I-66. This value appears reasonable and is consistent with the upper limit of 1.8 established by Jorgensen [ 14]_ NOTE ADDED IN PROOF: The reduced-minus-oxidized difference absorption spectra of various azurin derivatives have been found to give maxima in the region of -40 kR, making a LMCT assignment for the band reported for Mn(II)Az rather tentative_ See also I. Pecht, 0. Farver, and M. Goldberg, Advances in Chemistry Series 162,179 (1977). This lvork was supported
by (I grant from
the National
Institutes
of Health number
GM2276402.
REFERENCES 1. R. J. P. WiJliams,lnorg Chin Acra Rev. 5,137 (1971). D. R. McMillin, R. A. Holwerda, and H. B. Gray, Proc. NarZ. Acad. Sci. U.S.A. 71,
2.
3. 4.
1339 (1974)_ D. R. M&fiUi~, R. C. Rosenberg, and H. 3. Gray, ROC. NULLAcad_ SCL U.S.A. 71, 4760 (1974). E. I. Solomon, 3. W. Hare, and H. B. Gray, Proc. NarL Acad. Sci U.S.A. 73, 1389 (1976).
a It should be noted that methionine could not be involved in the case of the blue copper protein stellacyanin since methionine was not found in its amino acid analysis [22] _
184
D. R McMILLIN
L. Morpurgo, A. Finazzi-Agr6, G. Rotilio, and B. Mondo$ Eur. J_ Biochem 64,453 (1976). 6. K. E_ Falk and B. ReXmmmar, Biochim. Biophys. Acta 285,84 (1972). 7. T. E_ Jones, D. B. Rorabacher,and L. A. Ochrymowya,J. Am Chem Sot-. 97,7485 (1975)_ 8a_ A. Finazzi-Agr6, C. Giovagnoli, L_ Avigliano, G. Rotilio, and B. Mondo$ Eur J_ Biixhem 34,20 (1973)_ 8b. S. Katoh and A. Takarniya,J. Biochem (Tokyo) 55,378 (1964). SC. C. Brivin8and J. Deinum, FEBS Lett. 51,43 (1975). 9a_ E. I. Solomon. P. J. Clendening,H. B. Gray, and F- J. Grunthaner,J. Am C%em. Sot. 97,387s (1975). 9b. J. Peeling,B. G_ HasIett,I. M. Evans, D. J_ Clark, and D. Boulter,J. Am them Sot.. 99,102s (1977). 10. Y. Suguira, Y_ Hirayama, H. Tanaka, and K. Scbizu, J. Am Chem Sot. 99, 1581 (1977). 11. _D_htastropaolo.J. A_ Which,J_ A_ Potenza, and H. J. Schugar,J_Am Chem Sot_. 99, 424 (1977). 12 T. E_ Jones, D_ B. Rorabacher,and L. A_ Ochrymowycz, J_ Am them sOc_ 98,4322 (1976)_ 13. W_ H. Woodruff, privatecommunication_ 14_ C_ K. Jorgensen,hog_ Itwrg_Gzem 12,101(1970). 15. P. Day and C. K. Jorgensen,J. C%em Sot. 1964,6226. 16a. D. L. Tennent and D. R. McMiUin,to appear16b. D. L. Tennentand D. R. M&ill& to appear_ 17_ G_ B_ Kauffman, J_ Hwa-San Tsai, R. C_ Fay, and C. K_ Jorgensen.Inorg_ C?zem_2, 1233 (1963)_ 18. V. M. Miskowski, J. A- Thich, R- SoIomon, and H_ J_ Schugar,J. A.m them Sot 98, 8344 (1976). 19. M. D. Rowe, A. J. McCaffery. R. Gale, and D. N. Copsey, Inow Chem. 11, 3090 (1972)_ 20. J. Rawlings,0. Siiman. and H. B- Gray.&oc. NatL Acad. Sci USA. 71,125 (1974). 21. A. B_ P_ Lever.1. Chem Educ. 51,612 (1974). 22. J- Peisach,W- G- Levine,and W. E. Blumberg,J_ BtiL Chem 242,2847 (1967). 5.
Receiwd June 5.1977