1896
Journal of the American Chemical Society
(4) Takayanagi, T.; Yamamoto, H.; Kwan, T. Bull. Cbem. SOC.Jpn. 1975, 48, 2618-2622. (5) Stynes, D. V.; Stynes, H. C.; Ibers, J. A.; James, B. R. J. Am. Chem. SOC. 1973, 95, 1142-1149. (6) Abbreviations: Por, dianion of a natural or synthetic porphyrin; B, various donor ligands; Cap, dianion of the capped porphyrin, 5,10,15,20-[pyromellitoyl(tetrakis-0-oxyethoxyphenyi)]porphyrin; HmCap, dianion of the homologous capped porphyrin, 5,10,15,20-[pyromellitoyl(tetrakis-c~ oxypropoxyphenyl)] porphyrin; TPP, dianion of mes0-tetraphenylporphine; T(pOCH3)PP, dianion of tetra-pmethoxy-mesetetraphenylporphine; PPIXDME, dianion of protoporphyrin IX dimethyl ester; Deut. dianion of deuteroporphyrin IX; TpivPP, dianion of "picket fence porphyrin", mesotetra(cY,cY,cY,cY-epivalamidophenyl)porphyrin;THF, tetrahydrofuran; I-Melm. Kmethylimidazole; 1,P-Meplm, 1,2-dimethyIimidazole;2-Melm, 2-methyiimidazole; Im, imidazole; &ff, effective magnetic moment; (n-Bu)sP, tri-n-butylphosphine; P(OEt)a, triethyl phosphite; BUSH, I-butanethiol; BUS butyl mercaptide potassium crown ether; DMF, N,Ndimethylformamide; 3-Clpy, 3- chloropyridine; py. pyridine; 3,4-Me2py,3,4-dimethylpyridine; sec-BuNHz, sec-butylamine; tBuNH2, terf-butylamine; BzNC, benzyl isocyanide; PrNH2, n-propylamine; en. ethylenediamine; Me& tetramethylsilane. (7) Brault, D.; Rougee, M. Biocbem. Biopbys. Res. Commun. 1974, 57, 654-659. (8)Collman. J. P.; Brauman, J. I.; Doxsee, K. M.; Halbert, T. R.; Suslick, K. S. Proc. Natl. Acad. Sci. U.S.A. 1978, 75,564-568. (9) Brault, D.; Rougee, M. Biochemistry 1974, 13,4591-4597. (10) Rougee, M.; Brauit, D. Biochemistry 1975, 14, 4100-4106. (11) Rougee, M.; Brault, D. Biochem. Biopbys. Res. Commun. 1973, 55, 1364-1369. (12) Wayland. B. 8.; Mehne, L. F.; Swartz, J. J. Am. Chem, SOC. 1978, 100. 2379-2383. (13) Almog, J.; Baldwin. J. E.; Dyer, R. L.; Peters, J. J. Am. Cbem. SOC. 1975, 97, 226-227. (14) Almog, J.; Baldwin, J. E.; Huff, J. J. Am. Cbem. SOC. 1975, 97, 227228. (15) Szymanski. T. Ph.D. Thesis, Northwestern University, Evanston, Ill., 1978.
0,
/
102:6
/
March 12, 1980
(16) Adler, A. D.; Longo, T. R.; Kampas, F.; Kim, J. J. horg. Nucl. Cbem. 1970, 32, 2443-2445. (17) HmCapH2was previously prepared by Professor Jack E. Baldwin and coworkers, but the synthesis is unpublished. A sample from Professor Baldwin's laboratory was found to be identical with the HmCapH2 prepared in this synthesis: (18) Barnett, G. H.; Hudson, M. F.; Smith, K M. Tetrahedron Lett. 1973, 2887-2888. (19) The mass spectrum was performed on Fe"'(HmCap)CI by the Midwest Center for Mass Spectrometry, University of Nebraska, Lincoln, Nebr. (20) Evans, D. F. J. Cbem. SOC.1959, 2003-2005. (21) Brault, D.; Rougee, M. Biochemistry 1974, 13, 4598-4602. (22) Figgis, B. N.; Lewis, J. In "Modern Coordination Chemistry", Lewis, J.. Wilkins, R. G., Eds.; interscience: New York, 1967; p 403. (23) Hill, A. V. J. Physiol(London) 1910, 40, iv-vii. (24) White, D. K.; Cannon, J. B.; Traylor, T. G. J. Am. Cbem. SOC. 1979, TOT, 2443-2454. (25) Scheidt, W. R. Acc. Cbem. Res. 1977, 10, 339-345. (26) Basolo, F.; Pearson, R. G. "Mechanisms of Inorganic Reactions"; Wiley: New York, 1967; p 26. (27) Ellis, P. E., Jr.; Jones, R. D.; Basolo, F., J. Chem. SOC.,Cbem. Commun., in press. (28) Stern, J. 0.; Peisach, J. J. Biol. Cbem. 1974, 249, 7495-7498. (29) Chang, C. K.; Dolphin, D. J. Am. Cbem. SOC. 1975, 97, 5948-5950. (30) Collman, J. P.; Sorrell, T. N. J. Am. Chem. SOC.1975, 97, 4133-4134. (31) Chang, C. K.: Dolphin, D. J. Am. Cbem. SOC. 1976, 98,1607-1609. (32) Chang, C. K.; Dolphin, D. Proc. Natl. Acad. Sci. U.S.A. 1976, 73,33383342. '33) Jameson, G. B.; Ibers, J. A,, J. Am, Cbem. SOC.,in press. (34) Comer. W. M.; Straub. D. K. lnorg. Cbem. 1977, 16, 491-494. (35) Linard, J. E.; Ellis, P. E.. Jr.; Budge, J. R.; Jones, R. D.; Basolo, F. J. Am. Chem. SOC.,following paper in this issue. (36) Perutz, M. F. Br. Med. Bull. 1976, 32, 195-208. (37) Weber, E.; Steigemann, W.; Jones, T. A.; Huber. H.J. Mol. Biol. 1978, 120. 327-336. (38) Ellis, P. E., Jr.; Jones, R. D.; Basolo, F. Proc. Natl. Acad. Sci. U.S.A., 1979, 76,5418-5420.
Oxygenation of Iron( 11) and Cobalt( 11) "Capped" Porphyrins Jack E. Linard, Paul E. Ellis, Jr., John R. Budge, Robert D. Jones, and Fred Basolo* Contribution from the Department of Chemistry, Northwestern University, Ecanston, Illinois 60201. Received August 22, 1979
Abstract: The 0 2 binding constants are presented for a series of cobalt and iron complexes of the "capped" porphyrins (CapHz and HmCapH2). The C a p complexes have larger 0 2 affinities than the corresponding H m C a p systems, despite little difference in the basic structures in the capped porphyrins. The thermodynamic values of some of these complexes are compared to those of other oxygen carriers. The 0 2 affinities of the capped porphyrin complexes are considerably smaller than for other corresponding porphyrin complexes, and the major contributor to this difference is reflected in the A H o values. The capped porphyrins appear to inhibit oxygenation by imposing steric constraints on the movement of the metal atom from a stable, out of porphyrin plane position to an in-plane position. Evidence for the existence of a novel pseudo-seven-coordinate dioxygen complex, Fe(HmCap)(B)2(02), is presented. One of the bases is thought to coordinate to the metal in a nonaxial manner through the d,, or d,, orbitals of the metal.
Introduction Model oxygen carriers are synthesized and studied in order to understand more fully the nature of dioxygen binding to myoglobin and hemoglobin. I The simplest model compound, Fe(TPP)(B)22 in an aprotic solvent, was shown to reversibly bind dioxygen at low temperatures (-45 "C), but the reaction involves the substitution of dioxygen for one of the axially ligated bases.3 At higher temperatures the reaction of Fe(TPP)(B)2 with dioxygen results in an irreversibly oxidized product, the p o x 0 dimer (Fe(ll1)-0-Fe(Il1)). This differs from the natural hemoprotein where the iron is five coordinate, and oxygenation merely involves the addition of dioxygen to the vacant coordination site on iron. Other model oxygen carriers are made by the attachment of iron porphyrins to polymer support^.^ The polymer supports serve to prevent di0002-7863/80/ 1502-1 896$01 .OO/O
merization, because two iron centers cannot get close enough to react. Also the polymer support system provides a fivecoordinate ferrous ion, which is analogous to deoxymyoglobin and deoxyhemoglobin. Much interest in recent years has centered on metal complexes of substituted tetraphenylporphyrins. These porphyrins are synthesized to meet specific requirements, in that substituents can be varied to provide different electronic or steric environments around the metal. The most extensively studied5 of these are the iron(I1) "picket-fence'' porphyrins, which can reversibly oxygenate in aprotic solution containing excess axial ligand a t room temperature. Even in excess concentrations of ligand, the metal centers of some of the complexes can exist in a state of five coordination. Unique model properties are also found6 with chelated protoheme, having a "proximal" imid-
0 1980 American Chemical Society
Basoio et ai.
Oxygenation of Iron(II) and Cobait(II) “Capped” Porphyrins
1897
azole covalently attached. The iron complexes have many characteristics remarkably similar to those of myoglobin and As part of our investigation of factors which influence the binding of dioxygen in myoglobin Qnd hemoglobin, we are studying the characteristics of another series of substituted tetraphenylporphyrins. These are the metal “capped” porphyrins, elegantly prepared by Baldwin and c o - w o r k e r ~ . ~ ~ ~ Both CapH2 and HmCapHz are ortho-substituted tetraphenylporphyrins, with the four substituents being joined together on the same side of the porphyrin plane with a tetrasubstituted benzene (Figure I ) . grevious communications report the ligation and oxygenation e q ~ i l i b r i a ~and ? ’ ~the 0 2 and CO stretching frequencies]] of these metal porphyrin Figure 1. Schematic representation of CapHz ( x = 2) and HmCapH2 (x complexes. These studies show that a seemingly small struc= 3). tural change in the capped porphyrins of cobalt(I1) or iron(I1) results in significant differences in chemical properties. There are also large differences between the capped porphyrin Co(Cap) and Co(HmCap) base solutions were prepared in a complexes and the corresponding picket-fence porphyrin manner analogous to that of the Fe(Cap) base solutions, except that complexes. These facts may be important in discussing the extra precautions were taken to ensure that all glassware was as dry as possible. At the low temperatures needed for the titrations of the mechanism of cooperativity in hemoglobin, because the difcobalt complexes, the solutions were clouded by water. ference between the R and T states of hemoglobin may be the The base concentrations were chosen to give greater than 99% o f result of a slight conformational change. Our results suggest the five-coordinate complex in solution. The amount of the base that such a small change might suffice to cause a lower 0 2 needed was determined from the equilibrium constant for the addition affinity in the T state than in the R state. of base to the four-coordinated species.I3 In this paper we report our data on the oxygen affinity of T h e Fe(HmCap) base solutions were prepared by first reducing iron(I1) and cobalt(I1) Cap and HmCap complexes, and 0.8-2.0 mg of Fe(HmCap)CI in toluene or benzene with aqueous compare these values with data in the literature on related NazSz04. The bright red solution was then transferred to an airlesssystems. The nature of 0 2 binding to Fe(HmCap)( 1-MeIm)z ware flask equipped with a magnetic stir bar. The solvent was removed by heating gently under vacuum. The red residue was finally dried for is also discussed in terms of a unique pseudo-seven-coordinate a minimum of 3 h,at 150 “C in vacuo. The base solution, prepared as iron dioxygen c ~ m p l e x . ~
Experimental Section Reagents. Toluene was reagent grade and distilled under N2 from sodium benzophenone ketyl prior to use. Pyridine and 3,4-lutidine were reagent grade and distilled under N2 over BaO before using. I-Melm was dried over K O H , vacuum distilled, and stored under N2. 1.2Me2lm was vacuum distilled and stored below 0 OC. The amine bases were dried over K O H and distilled under N2.3-Clpy (Aldrich, 99%) was distilled under N2. All bases except I,2-Me2Im were stored over molecular sieves. P(n-Bu)3 and P(OEt)3 were vacuum distilled under N2. Gaseous dioxygen was Air Products Ultra-Pure Carrier Gas (containing less than 1 .O ppm H 2 0 ) or Matheson Primary Standard 0.999% 0 2 in N2. When the temperature of the equilibrium experiment to be used was less than -70 OC, the oxygen gas was further dried by passing over molecular sieves at -130 OC. Syntheses. The capped prophyrin compounds CapH2, Fe(Cap)CI, and Fe(Cap) were prepared and characterized according to the method of Baldwin and co-workers.8x’2 The compounds HmCap, Co(Cap), Co(HmCap), and Fe(HmCap)(Cl) were prepared by literature method^.'^ The Fe(HmCap) was prepared in situ by reduction of Fe(HmCap)(Cl) with aqueous sodium d i t h i ~ n i t eThe . ~ compounds Fe(HmCap)( I-Melm) and Fe(HmCap)(prNH2)2 were also prepared in situ by adding 1 mL of I-Melm and I niL of PrNH2, respectively, to dried samples of Fe(HmCap). The excess base was then evaporated by heating gently and reducing the pressure with a vacuum pump. The flask was heated to no more than 50 OC during this process, as excess heat will drive off all coordinated base. This method may not give the exact base-metal stoichiometry indicated, but it was suitable for the experiments conducted. Procedure. Equilibrium constants for the oxygenation reactions were determined spectrophotometrically, using a 4.0-cm path length low-temperature visible ~ e 1 1 . A I ~Cary 14 spectrophotometer was used in all experiments. In order to exclude water and dioxygen, all solution transfers were done with cannulas or with gas-tight syringes equipped with Luer-Lock valves. Fe(Cap) (-1 mg) was added to the visible cell, which contained a small magnetic stir bar, and flushed with N2 for at least 1 h before the solvent solution was added. The base solution was prepared by adding a known mass of the appropriate base to a 25-mL volumetric flask, which had been stoppered with a serum cap and thoroughly flushed with N2. Toluene was then added to make a total of 25.0 m L of base solution. After mixing, the base solution was transferred to the visible cell.
described above for Fe(Cap), was added to the Fe(HmCap) residue, mixed, and transferred to the visible cell. Again the base concentrations were chosen to ensure more than 99% of the five-coordinated species for the respective bases 1,2-MezIm, t-BuNH2, P(n-Bu)3, and P(OEt)3. These bases do not form six-coordinate complexes with Fe(HmCap).14 The spectra of these iron(l1) complexes are similar to those of other five-coordinateco~nplexes.’~ When I-Melm, py, or PrNH2 were the bases, concentrations were chosen to ensure greater than 96% six coordination. Once the solution was in the visible cell. it was cooled to the chosen temperature with the appropriate slush bath. While cooling, the cell was attached to the vacuum line and degassed for a total of 1 h. The system was then shut off from the vacuum pump and allowed to equilibrate, with the pressure reading being the vapor pressure of the solution. This value was subtracted from the equilibrium pressures recorded during the titration. Once the vapor pressure was recorded. an initial spectrum was taken. Aliquots of 0 2 were then added from a gas manifoldI6 via a gas-tight valve to the system. When P112O2was less than 10 Torr, a mixtureof 0.999% 0 2 in N2 was used as the titrant. The solution was allowed to stir until equilibrium was attained. Equilibrium was reached more quickly if the cell was also agitated by hand. Use of stainless steel flexible vacuum tubing allowed the cell to be shaken. The spectra for a typical titration of Co(Cap)(B) are shown in Figure 2. The pressure of 0 2 above the solution was determined with M K S Baratron, Type 221 series, pressure gauges having a range of 0-100 and 0-1 000 Torr. Pressures were obtained in a digital format from a Data Precision digital voltmeter (Model 245). Reversibility was checked after the last 0 2 addition by vacuum removal of 0 2 to obtain the identical initial deoxy spectrum. When either Fe(HmCap)(l-Melm) or Fe(HmCap)(PrNHz)z was used, base solutions were not added. Instead, toluene alone was added to the flask containing the sample, mixed quickly, and transferred via a cannula to the visible cell. The cell, containing a small magnetic stir bar, had been flushed thoroughly with N2 and cooled to -63 OC prior to addition of solution. The dioxygen titration was conducted in the manner described above, except that fewer additions of 0 2 were made. After the addition of sufficient dioxygen to completely oxygenate the sample, 2 mL of neat oxygen-free base was added to the solution. After mixing, a final spectrum was taken. When I - M e l m was added, the solution had to be warmed slightly (above -60 “ C ) to prevent the I-Melm from freezing. The oxygenation of five-coordinate complexes is illustrated by the
1898
Journal of the American Chemical Society
WAVELENGTH ( n m ) Figure 2. Spectral changes occurring upon titration of a toluene solution ~ f - l O . - ~M in Co(Cap)(l-Melm), 0.03 M in 1-Melin, with the following pressures of dioxygen at -78 OC; 0, 20.14, 5 1.64.9 I .37,204.7 I , 505.95, and 995.05 Torr.
equation M(Por)(B)
3
+ 02-M(Por)(B)(02)
(1)
The data from the spectrophotometric titrations are treated in one of two ways. ( I ) When the absorbance of the completely (>99%) oxygenated product can be obtained, the data are fitted to the Hill equation’* log [Y/(l
- 4’)l = n log Po, - log (Pl/20’)
(2)
wherey equals the fraction of the total metal porphyrin binding 0 2 . Values for P,p02 were determine,d from the x intercept of the regression line for a plot of log [y/( I - v)] vs. log f 02.For the simple additions of 0 2 to five-coordinate complexes, the values of n equal 1 .00. Values for Pl.po2 were found to be independent of the wavelength used to monitor the extent of oxygenation. (2) If only partial oxygenation occurred at a pressure of 1000 Torr, the data were fitted to an equation employed by C o l h ”
Po2= [ M ( P o r ) ] ~ b A t ( P o ~ / A . 4-) P 1 j 2 ~ ’
Results Solid Fe(Cap), Co(Cap), and Co( HmCap) are moderately stable with respect to aerial oxidation, and transfers of the solid materials can be performed in contact with the atmosphere without special precautions. Because Fe(HmCap) is prepared in situ by reduction of Fe(HmCap)Cl, there is no reason to expose it to the atmosphere. It is always kept under nitrogen or in vacuo. Exposing solutions of Fe(Cap) or Fe(HmCap),
102:6
/ March
! , , , , I 500
550
600
650
550
6 00
12, 1980
,
I 650
WAVELENGTH (nm)
Figure 3. Spectral changes which occur upon additions of 0 2 or PrNH2 to a toluene solution of M in Fe(HmCap)(PrNH2)2 at -63 OC: (a) initial spectrum of Fe(HmCap)(PrNHz)z; (b) 8.12 Torr of 02; (c) 38.60 Torr of 0 2 ; (d) 602.1 Torr of 02; (e) addition of 2 mL of neat PrNH2 to the solution in (d). Absorbance scale is different from that for (a)(4.
in the absence of a large excess of coordinating base, to the atmosphere a t room temperature results in rapid, irreversible oxidation to the p-oxo dimers. With excess concentrations of coordinating base solutions, the complexes and their 0 2 adducts are stable for several hours in contact with the atmosphere a t room temperature. The addition of base solutions to Fe(Cap), Co(Cap), or Co( HmCap) yields solutions of the corresponding five-coordinate species. This fact is shown by the visible spectra and by estimates made using the base concentrations and equilibrium constants (KB)13for the equilibria represented by the equation
(3)
where [M(Por)]l- is the total metal porphyrin concentration, b is the cell path length, A6 is the difference in molar absorptivity of the oxy and deoxy complexes, and AA is the difference between the absorbance at a particular Pozand the absorbance of the deoxy complex a t the same wavelength. Since [M(Por)]bAt is a constant, a plot of Po2 vs. f o , / A A will givea straight line withy intercept equal to -f 1 1 2 ~ 2 . Plots were made for at least two wavelengths over a pressure range of 0- 1000 Torr. I n several instances eq 3 was used with data originally fit to eq 2. The P1p02’s obtained from each were the same within experimental error. All data were fitted using a nonweighted linear least-squares method. Values of P1po2 are reproducible to better than 5%. The data for the oxygenation of Fe(HmCap)(B)2, with excess B and with B = I-Melm or py, were fitted to equations described prev i o ~ s l yThe . ~ Plj?Oz is determined from eq 2 or 3 except that the reaction is an addition of 0 2 to a six-coordinate complex.
/
KB
M(Por)
+ B +M(Por)(B)
(4)
These three metal porphyrins show no optical spectral change indicative of adding a second base. Figure 2 shows the spectra from a titration of Co(Cap)( 1-MeIm) with 0 2 . The addition of solutions of t-BuNH2 or of 1,2-Mezlm t o Fe(HmCap) gives only five-coordinate species, even in high concentrations of base.13 When unhindered bases such as 1 MeIm or py are used, Fe(HmCap) forms a six-coordinate species: KgB
Fe(HmCap)(B)
+ B +Fe(HmCap)(B)*
(5)
Although for these two complexes K B is~smaller than K B , it is still too large to allow the oxygenation of a predominantly five-coordinate complex. A significant concentration of the easily oxidizable four-coordinate complex Fe(HmCap) is still present in the solution when the five-coordinate Fe(HmCap)(B) concentration is maximized. All 0 2 titrations of the py and 1-MeIm adducts of Fe(HmCap) were therefore performed on the six-coordinate complexes. Base concentrations were large enough to ensure that the amounts of four- and five-coordinate complexes were negligible. Spectrophotometric oxygen titrations of toluene solutions
Basolo et ai.
1899
Oxygenation of Iron(II) and Cobalt(II) "Capped" Porphyrins
Table I. Equilibrium Data for Oxygenation of Five-Coordinate Iron(1l) Capped Porphyrin Complexes in Toluene (eq 1)
P I1 2 ~ 2Torr , complex Fe(Cap)
Fe(HmCap)
base 1 -MeIm
1,2-Me2lm PY 3.4-Lut 3-Clpy sec-BuNH2 t-BuNH2 P(OEt)3 P(~I-Bu)~ I-MeIm 1.2-Me2lm I-BuNH~ see-BuNHz PrNH2
PK, (€3H +)
log K B a
7.25c 7.8SC 5.27e 6.46e 2.84e 10.56f 10.83f 3.50g 8.43' 7.25c 7.8SC 10.83f 10.56f 10.53f
2.90 3.06 2.17 2.26 1.53 2.16 2.50 3.24 4.66 3.311 3.61 2.23 2.79 3.40J
0 "C 4.5 930 26 34 32 12
-23 " C
-45 "C
0.78 162 4.2
0.10d
-63 " C
27 0.36d
0.27
h h 120-180k
880 575
I I
Equilibrium constant for eq 4 at 23 "C; values from ref 13. Values are reproducible within 5%, except when noted otherwise. Reference 19. Value reproducible within 10%. e Reference 20. J' Reference 21. Estimated from values in ref 22. N o detectable spectral changes up to 1 a t m (760 Torr). Reference 22. J Equilibrium constants from ref 13 for addition of second base (K& are 0.77 (B = I-Melm), 2.85 (B = see-BuNHz), and 4.05 (B = PrNH2). For discussion see text and ref 9. Value for oxygenation of the five-coordinate complex is not able to be determined directly because the complex is six coordinate, with two axially ligated bases, in solutions containing excess base. There a r e no detectable spectral changes up to 1 a t m when the six-coordinate species in excess base solution is titrated with 0 2 at 0 "C.
Table 11. Equilibrium Data for Oxygenation of Cobalt(1I) Capped Porphyrins in Toluene (eq 1) Torrb -78 "C
P112O2,
complex Co(Cap)
Co(HmCap)
base I-Melm 1,2-MezIm PY sec-BuNH2 I-Melm
-63 " C 7.2SC 7.85c 5.27e 10.569 7.25c
2.32 1.84
f f
2.28
790
140 2000-4000 I380 955 >so00
-91 "C
45 d
a Equilibrium constant for eq 4 at 23 "C in text; values from ref 13. Values a r e reproducible within 10%. Reference 19. No detectable spectral change up to 1000 Torr of 0 2 . e Reference 20. f Kd not determined; solutions were made, however, to ensure greater than 99% five coordination. g Reference 21.
of Fe(Cap)(B), Fe( HmCap)(B), Fe(HmCap)(B)z, Co(Cap)(B), and Co(HmCap)(B), all with excess base present, were done. Adequate isosbestic points were maintained in all titrations. When eq 2 could be used on the data, plots of log [ y / ( l - y ) ] vs. log Po2 gave straight lines with slopes of 1 .0-0.1. The value of the x intercept is equal to -log Ko2, where KO? = ( P 1 / 2 O 2 ) - 1. When eq 3 was employed, plots of Po2vs. Po2/A A gave straight lines with t h e y intercept being equal to -P1/2OZ. Values obtained for the oxygenation of the five-coordinate complexes are shown in Tables I and 11. Unexpectedly, it was found that Fe(HmCap)( 1-MeIm)z binds 0 2 in a manner that is independent of base concentrati01-1.~ This fact also appears to be true when py is the base, but unfortunately Fe( HmCap)(py)z binds dioxygen very poorly (P1/2O2 > 1000 Torr at 0 "C). The values obtained for the pyridine system were not precise. For Fe(HmCap)(PrNHz)z, K B >~ KB,I3 and it does not bind 0 2 when excess base concentrations are present. These properties are also characteristic of Fe(TPP)(py)z, which will bind 0 2 reversibly via a substitution reaction at temperatures below --45 0C.3 Note that such a solution contains only stoichiometric amounts of pyridine, since only solid Fe(TPP)( p y ) ~is dissolved in the pure solvent. However, the addition of base easily prevents oxygenation, because under such conditions pyridine successfully competes with 0 2 for a coordination site. Low temperatures are needed to slow the reactions which form the p-oxo dimer. In order to determine if the normal behavior of Fe(Por)(B)z is also shown by Fe(HmCap)(PrNH2)2, it was prepared as the solid with approximately the stoichiometry indicated. The results of 0 2 titrations of a toluene solution of Fe(HmCap)(PrNH2)2 are shown in Figure 3. The deoxy spectrum a t -63 OC is similar to those of other six-coordinate iron( 11) porphyrin complexes.24 The first 0 2
addition, 8.12 Torr, gives strong evidence that oxygenation is occurring. Oxygenation was almost complete at 38.60 Torr, and at 602 Torr of 0 2 it is clear that the system was completely oxygenated. The 620-Torr spectrum is similar to that of Fe(HmCap)(t-BuNH2)(02), which is definitely a six-coordinate complex. Upon addition of 2 m L of PrNH2 to the completely oxygenated solution, the spectrum obtained is that of deoxy species Fe(HmCap)(PrNHz)z. A similar experiment was performed using solid Fe( HmCap)( 1-MeIm)z. The deoxy spectrum (Figure 4) of its toluene solution, however, resembled that of the five-coordinate Fe(HmCap)(t-BuNH2). The compound in solution is fivecoordinate Fe(HmCap)( 1 -MeIm), in accordance with the small binding constant ( K B ~for ) the second base.9 At -63 OC complete oxygenation was obtained at about 20 Torr of 02, but to be certain of this 304.2 Torr was used (Figure 4). At this pressure of 02, a large excess of 1-Melm was added to the solution and the spectrum changed only slightly. The spectrum still showed complete oxygenation (Figure 4d). Removal of 0 2 by vacuum resulted in the six-coordinate spectrum of Fe(HmCap)( l-MeIm)2. The thermodynamic values for oxygenation of Fe(Cap)( I-MeIm), Fe(Cap)( 1,2-MezIm), Fe(Cap(py), and Co(Cap)( I-MeIm) were obtained by determining the P1/2O2's at three different temperatures. Values of AH" and ASo were then determined from least-squares fits to the van? Hoff equation. The values for these and other systems are given in Table 111. Discussion Oxygenation of Five-Coordinate Complexes. The information contained in Tables I and I1 has many interesting features. In accord with several iron and cobalt system^,^^^.^^ the use of
Journal of the American Chemical Society
1900
/
102:6
/ March 12, I980
Table 111. Equilibrium and Thermodynamic Values for Oxygenation of Iron(l1) and Cobalt(I1) Porphyrin Complexes in Toluene (eq 1 )
complex
P1p02,
Fe(Cap)( I-Melm) Fe(Cap)( 1,2-Me2lm) Fe(Cap) (PY) Fe(triPiv(4CIm)PP) Fe(TpivPP)( 1,2-Me*lm) Mb (sperm whale)e Hb (human, “T”)f Hg (human, “R”)f Co(Cap)(I-Melm) Co(T@-OCH3)PP)(1 -Melm) Co(PP1XDME)-( I -MeIm) Co(TpivPP)( 1-MeIm) CoMb (suerm whale)‘
Torra
23
4.0 x 103 c 1.8 X lo2 0.60 38 0.29
A H ” , kcal/mol
AS”, eub
ref
-10.5 f 0.3 -9.7 f 0.2 -11.8 f 0 . 4 -16.8
-41 f 1 -49 f I -50 f 2 -42
this workd this workd this workd 6 6 25, 26
-14.3
-42
-14.8
-47
26
28,29
0.17 1.4 x 105 c 1.0 x 104 c 6.1 x 103 c 7.0 X IO’ 3.4 x 10’
-7.8 f 0.4 -8.9 -9.7 - 12.2 -13.1
-50 f 2 -49 -51 -51 -53
28, 29 this workd 30 31 7
32.33
a 25 “C for iron complexes, 15 OC for cobalt complexes. Standard state is 1 Torr. Calculated from thermodynamic values. Error limits for A H ” and ASo are standard deviations from van’t Hoff plots. e Aqueous solution, pH 7 , O . l M phosphate ion. f Aqueous solution, pH 7.4, 0.1 M NaCI, 0.05 M bis-Tris buffer.
I ‘
500
550
600
650
550
600
650
W A V E L E N G T H (nm)
Figure 4. Spectral changes which occur upon additions of 0 2 or I-Melm to a toluene solution of M in F e ( H m C a p ) ( l - M e l m ) at -63 OC: (a) initial spectrumof Fe(HmCap)(l-Melm); (b) 19.78 Torr 0 2 ; (c) 304.2 Torr 0 2 ; (d) addition of 2 mL of neat I - M e l m to the solution in (c) (absorbance scale is different from that in (a)-(c)); (e) removal of 0 2 by vacuum from solution (d). Absorbance scale is different from those in (a)-(d).
hindered bases such as 1,2-MezIm and t-BuNH2 as axial ligands destabilizes the 0 2 adduct. The Plpo2 for Fe(Cap)(1,2-Me2lm) is 200 times larger than the P1,2~2Fe(Cap)(I-MeIm) at 0 O C . The accepted explanation for this behavior is that the hindered bases do not allow the ferrous ion to easily move toward the porphyrin plane because of steric interaction with the pyrrole group on the porphyrin ring.’5-34The high-spin iron atom, as it binds 0 2 , becomes low spin and moves into the plane of the porphyrin. The effect of changing from I-MeIm to 1,2-Me2Im is to decrease the 0 2 affinity by a factor of about 180, which is comparable to the average difference of values for T- and R-state iron(I1) porphyrins in human hemoglobin.29 The difference in 0 2 affinity for Co(Cap)(B) for B = 1MeIm vs. 1,2-MezIm is not as large as it is for Fe(Cap)(B), as was also observed for the picket-fence complexes. This may
be because cobalt(I1) is not as far out of the porphyrin plane in the deoxy state as is i r ~ n ( I I ) . ~ A ~ - hindered ~* base will therefore have less effect on the binding of 0 2 to cobalt porphyrins than it does on the binding of 0 2 to iron porphyrins. The pK, for a series of substituted pyridines, as axial bases, has little effect on the 0 2 affinity of Fe(Cap)(B). The P1p02’s for 3-Clpy (pK, = 2.84), py ( 5 . 2 7 ) ,and 3,4-lut (6.46) are all about 30 Torr at 0 “C. The more electron donating a base (the higher the pKa), the better the metal should bind 0 2 because it can more readily donate electron density to the 0 2 moiety. For and manganese40 porphyrins there is a rough correlation between the pK, of various substituted pyridines and dioxygen binding equilibria; however, the overall effect is small. A similar study of the effect of substituted pyridines on the 0 2 uptake by the Co(Cap)(B) was not done, since the 0 2 affinity of Co(Cap)(py) is too small for us to determine accurately. That Fe(Cap)( 1-MeIm) has the highest affinity for 0 2 is expected, because 1-MeIm is a good 7r-electron donor.4i The increased electron density on the metal center increases its affinity for 02, since charge transfer is an essential part of M-02 bonding. I n contrast, the low 0 2 affinity of these systems when the axial base is a phosphine or phosphite is a result of the r-electron acceptor characteristics of these ligand~.~~ W e originally believed that the small 0 2 affinity of M(Cap)(B) relative to M(Por)(B) might be due to the small “cap size” placing constraints on 0 2 inside the cap. In order to test this we prepared M(HmCap)(B), which we think has a larger cap size, and if this were the important factor then its 0 2 affinity should be larger than that for M(Cap)(B). The results (Tables I and 11) show that just exactly the opposite happens. The deoxy state of a flat cobalt porphyrin would be better solvated by nonpolar solvents such as toluene, especially at low temperature. A polar bond such as Co(III)-O2- would not be stabilized in such solvents. Since the 0 2 moiety of the picketfence analogues is effectively shielded from the solvent by virtue of the “pickets”, the solvation effect would be minimized. This would lead to the greater stability of the oxygenated species. Yet the results presented here appear to negate this solvation hypothesis, since the bound 0 2 in the capped porphyrin systems is even more protected from solvent influence than in the corresponding picket-fence porphyrin complexes. The highly shielded cobalt( 11) capped porphyrins, Co(HmCap)(B) and Co(Cap)(B), have a lower 0 2 affinity than the corresponding highly solvated flat porphyrin complexes, Co(PPIXDME)(B) and Co(T(p-OCH3)PP)(B). The same trend follows with iron porphyrins, where Fe(Cap)( l MeIm) binds 0 2 about 40 times less than Fe(triPiv(4CImP)PP). The conclusion is that the solvolysis may
Basolo et al.
/
Oxygenation of Iron(II) and Cobalt(II) “Capped” Porphyrins
1901
be valid in accounting for the differences in 0 2 affinity between Chothia28have pointed to a conformational difference between CoMb and Co(PPIXDME)( 1-MeIm)33and between the coR-state and T-state hemes in hemoglobin. I n R-state hemobalt(I1) picket-fence porphyrins and C O ( T ( ~ - O C H ~ ) P P ) ( B ) , ~globin, it was concluded28 that the imidazole of the proximal but for the capped porphyrin systems solvolysis must play a histidine is in a symmetric position relative to the heme. This minor role in determining 0 2 affinity. The invariance of 0 2 supposedly lessens the steric interactions with the porphyrin affinity to the pK, of the axially substituted pyridines also ring and allows the heme to easily oxygenate. In the T state the suggests that some other factor(s) predominates in determining imidazole of the proximal histidine is coordinated to the ferrous the 0 2 affinity of the capped metal porphyrins. ion in an asymmetric conformation which causes steric interThe low 0 2 affinity of the capped porphyrin systems could action between the imidazole and the porphyrin nitrogens. be the result of several factors: (1) unfavorable steric interThus, the 0 2 affinity is lowered considerably because steric actions between the bound dioxygen and the porphyrin cap, interaction prevents the iron atom from going into the por(2) an increase in the conformational strain energy of the phyrin plane. This is similar to the conformational strain effect capped porphyrin complexes upon oxygenation, and (3) elecwhich we propose is responsible for the low 0 2 affinity of the tronic substituent effects. In going from M(Cap)(B) to the capped porphyrin complexes. corresponding M(HmCap) (B) the steric interactions between The change in axial bases for porphyrin complexes from bound dioxygen and the porphyrin cap should decrease because 1-MeIm to 1,2-Me2Im produces decreases in 0 2 affinity which the cap size is larger. This predicts that M(HmCap)(B) should are similar to those for going from R-state to T-state hemohave a larger 0 2 affinity than M(Cap)(B), but since the conglobins6 and cobalt-reconstituted hemoglobin.51 This effect trary is found factor (1) is ruled out. Electronic effects (3) are of 1 , 2 - M e ~ I mis due to the steric interaction of the 2-methyl expected to be small from previous s t ~ d i e s ,and ~ ~ our , ~ studies ~ substituent which resists the metal entering the porphyrin plane of the capped porphyrin systems show that electronic changes on oxygenation. In going from 1-MeIm to 1,2-Me2Im for have little effect on the 0 2 affinity of the complexes. Therefore, Fe(Cap)(B) or Co(Cap)(B), the 0 2 affinity also decreases by of the three factors considered, it appears that the conformaamounts similar to that for changes from R- to T-state H b and tional strain energy (2) must play a major role. CoHb. For example (Table H I ) , the P1p02 for Fe(Cap)The low 0 2 affinities displayed by the capped porphyrins (1,2-MezIm) is about 180 times higher than that for suggest that the conformational strain energy in M(capped)(B) Fe(Cap)(l-MeIm). For hemoglobin, the P1/2O2 for the T state is enhanced upon oxygenation to M(capped)(B)(02). The is 150 times higher than for the R state. Thus, the steric inmetal is out of the plane of the porphyrin, which may be teractions which cause the lowering of 0 2 affinity in going from “domed”,46 in the deoxy species, and movement of the metal 1-Melm to 1,2-MezIm are virtually independent of those ininto the porphyrin plane upon oxygenation may be hindered teractions which cause the low 0 2 affinity of the capped porowing to steric restraints imposed by the porphyrin cap. Conphyrin complexes. sequently, the capped porphyrins would tend to “lock” the The thermodynamic values in Table 111show that the major metal into the out-of-plane position. Thus the major difference differences in binding 0 2 by the metal complexes are manibetween the various systems (Table 111) may be the energy fested in the enthalpy term. If the capped porphyrins are needed to undome the porphyrin and to bring the ferrous ion constrained to a five-coordinate configuration, the process of into the porphyrin plane. For natural systems and thevarious binding 0 2 to form a six-coordinate system would require more picket-fence porphyrin complexes, this energy would be smaller energy than if these constraints were smaller. The values of than that for the corresponding capped porphyrin complexes. A H o are higher for the more constrained capped porphyrin Since the 0 2 affinities for the HmCap systems are less than systems than for the other metalloporphyrins of lesser strain. for the corresponding Cap systems (and the smallest of any However, the values of ASoare all much the same because the system in Table 111) it follows that the former systems lock the major factor contributing to AS”, the loss of degrees of freemetal even more rigidly into the five-coordinated conformadom in binding 0 2 , is similar for all systems. It should also be tion. noted that the enthalpy terms reflect the added steric restraint This five-coordinate conformation of the capped porphyrins of the 1 , 2 - M e ~ I msystems compared with the corresponding need not necessarily have more strain than do other deoxy 1-MeIm systems. Thus Fe(Cap)( 1,2-Me2Im) has the highest metal porphyrin systems. Gelin and K a r p l ~ have s ~ ~found from A H o and the highest P112O2 of the iron systems in Table energy calculations that the unliganded, or deoxy, heme is not 111. subject to any significant steric strain. The iron atom is in the Oxygenation of Fe(HmCap)(B)z. I n an earlier communicaoptimal position for a five-coordinate high-spin ferrous ion, and tion9 it was suggested that the oxygenation of Fe(HmCap)this position is not significantly farther from the mean plane (1 -MeIm)z occurs by a simple addition reaction: of the porphyrin than in five-coordinate model compounds.49 Fe(HmCap)(l-Me1m)z 0 2 In the deoxy state there may be some doming of the porphyrin ring, but the energy difference between a domed and planar KB202 -Fe(HmCap)(I-Melm)2(02) (6) structure is slight, not enough to store considerable amounts of strain energy.47Even the porphyrin ring itself in hemoglobin This result requires the formation of a most unusual pseudois very flexible, since an expansion of the ring to accommodate seven-coordinate complex. The addition of dioxygen would be a high-spin ferrous ion in the plane is estimated28 to require dependent on the base concentration if it proceeded by the less than 0.5 kcal. The capped porphyrin complexes have K B usual substitution reaction, represented here by the oxygenvalues that are about ten times smaller than other correation of F e ( T p P ) ( p ~ ) 2 : ~ sponding metal porphyrins.I3 These data suggest that the deoxy capped porphyrin complexes have similar or slightly Fe(TPP)(py)2 + 0 2 Fe(TPP)(py)(02) + PY smaller conformational stabilities than do the other systems. The major difference between the capped porphyrins and other where K,, = ( K B O ~ ) ( K B ~ ) - ’ systems must then result from the actual oxygenation process. - [Fe(TPP)(PY)(02)1 [PYl (,) If the porphyrin caps, owing to steric interactions with atoms [Fe(TPP)(py)21Po2 in the porphyrin ring, prevent the porphyrin from undoming and/or the metal from entering the mean porphyrin plane, the From eq 7 the apparent P112O2 would be a function of base 0 2 affinity of the complex would be quite small. concentration, and the following relationship may be deEnergy calculations by WarshelS0 and later by Baldwin and rived:
+
5
1902
Journal of the American Chemical Society
CH3
Figure 5. Schematic representation of nonaxial base, I-Melm, in Fe(HmCap)( 1 -MeIm)2. Horizontally lined orbital is iron dp2, vertically lined orbitals are sp2 lobes of imidazole nitrogens, and solid orbital is a lobe of the iron d,, (or dxz) orbital.
PI p o 2 (apparent) = (1 K B ~ [ B ] ) Pwhere ~ ~ ~P1p02 ~ , = (Ke02)-I
+
(8)
The measurement of dioxygen affinities was made a t 0 O C , where K B =~8.3 f 0.7. If the base concentration is raised from 0.5 to 2.0 M I-MeIm (B = I-MeIm), the Plp02 (apparent) should increase by a factor of 3.4 according to eq 8. The results, however, show that Plp02 (apparent) is constant. Thus, the reaction of Fe(HmCap)(l-MeIm)z with 0 2 is independent of the base concentration and is most likely an addition reaction. The P1p02 for the addition reaction, eq 6, can be determined from eq 2 or 3. The P1/2O2 values for the formation of the pseudo-seven-coordinate complex are 200 f I O Torr a t 0 "C and 22 f 2 Torr a t -45 O C . In order to compare the oxygenation with that for the five-coordinate systems, however, an estimate of 120-1 80 Torr is made for the reaction of 0 2 to form Fe( HmCap)( 1-MeIm) ( 0 2 ) .9 Further work on this novel system gives added credence to the reaction represented by eq 6. Solid Fe(HmCap)( 1-MeIm) was dissolved in toluene and titrated with 0 2 a t -63 O C (Figure 4). The deoxy spectrum (Figure 4a) is typical of a five-coordinate iron(I1) porphyrin complex. The addition of 0 2 oxygenates more than half of the complex at 1 1.2 Torr and most of it a t 19.8 Torr. To ensure complete oxygenation, the solution spectrum was determined at greater than 300 Torr of 0 2 . This spectrum (Figure 4c) has a peak at 549 nm and resembles the spectrum of Fe(Cap)( l-MeIm)(O*). When a large excess of I-MeIm is added, the spectrum changes little (Figure 4d). The maximum is now at 546 nm, and this is believed to be the spectrum of Fe(HmCap)( l-MeIm)2(02). It is significant that the addition of excess I-MeIm does not replace thedioxygen from the molecule. However, when 0 2 is removed by vacuum, the reverse of eq 6 occurs and the spectrum of Fe(HmCap)( l-MeIm)z is obtained (Figure 4e). For the addition of I-MeIm to Fe(HmCap), K B
