Aza-crown-capped porphyrin models of myoglobin: Studies of the steric interactions of gas binding

Article abstract of DOI:10.1021/ja963945i

A series of myoglobin active site analogues (1-6) has been synthesized and characterized. These synthetic models differ in their cavity dimensions, and have been designed to demonstrate the effects of steric factors on O₂ and CO binding affinit

Full text of DOI:10.1021/ja963945i

J. Am. Chem. Soc. 1997, 119, 3481-3489
3481
Aza-Crown-Capped Porphyrin Models of Myoglobin: Studies
of the Steric Interactions of Gas Binding
James P. Collman,*,† Paul C. Herrmann,† Lei Fu,† Todd A. Eberspacher,†
Michael Eubanks,† Bernard Boitrel,† Pascal Hayoz, Xumu Zhang,†
John I. Brauman,† and Victor W. Day‡
Contribution from the Department of Chemistry, Stanford UniVersity,
Stanford, California 94305-5080, and Department of Chemistry, UniVersity of Nebraska,
Lincoln, Nebraska 68588-0304
ReceiVed NoVember 13, 1996X
Abstract: A series of myoglobin active site analogues (1-6) has been synthesized and characterized. These synthetic
models differ in their cavity dimensions, and have been designed to demonstrate the effects of steric factors on O2
and CO binding affinities. Quantitative gas titrations were employed to measure these affinities, yielding M values
that are strikingly lower than those reported for hemoglobin and myoglobin. The 1,4,7-triazacyclononane-capped
porphyrin 1 has about 1200 times the CO affinity but only about 10 times the O2 affinity of the cyclam-capped
porphyrin 2, suggesting a more open gas binding cavity for 1. The cavity dimensions and conformation of 2 were
determined by single-crystal X-ray structural analysis of the Zn analogue 7. This paper unequivocally demonstrates
that steric effects can control the ratio of O2/CO binding constants.
Introduction
fence porphyrin was ca. 30 times that of Hb.6c If Hb or Mb
had the same CO to O2 affinity ratio (M value)7 as the picket
fence porphyrin, mammals would suffocate from their own heme
catabolism. Results from synthetic models have been consistent
with the hypothesis that the protein moieties of Hb and Mb are
large determinants of the protein’s M value.
On the basis of picket fence models as well as Hb and Mb
X-ray crystal structures, it was proposed that the affinity of Hb
and Mb for carbon monoxide is decreased by a steric interaction
of the linearly bound CO with “distal” amino acid residues in
the vicinity of the binding site.8 Dioxygen, binding in a bent
fashion, should be free of such an interaction. Work on the
“pocket”,9ab “hybrid”,9c “capped” 9d-f porphyrins provided ad-
Hemoglobin (Hb) and myoglobin (Mb) are the proteins
responsible for mammalian dioxygen transport and storage. Both
Hb and Mb contain a heme axially bound to a histidine residue,
resulting in a 5-coordinate high-spin (S ) 2) Fe(II) complex.
In addition to binding O2, this 5-coordinate Fe(II) active site
binds CO tenaciously. Consequently both proteins are severely
inhibited by carbon monoxide.1
The in ViVo production of carbon monoxide during the
breakdown of heme by heme oxygenase further complicates
dioxygen transport and storage.2 Exactly one molecule of CO
is produced per molecule of heme catabolized, resulting in a
partial pressure of CO on the order of 1 × 10-3 Torr at the
cellular level.3 In fact, approximately 1% of a human’s Hb is
carbonylated due to this in ViVo CO production. Thus, the heme
proteins are designed to transport and store dioxygen in the
presence of an endogenous poison, CO.
The nature of the O2 interaction with Hb has been of interest
for some time. Pauling hypothesized in 1964 (and neutron
diffraction studies on oxyMb later confirmed) that bound
dioxygen is stabilized by a hydrogen bond from the distal
histidine residue to the terminal O atom of the bound O2
molecule.4 Recently, synthetic models of Hb and Mb have been
invaluable in unraveling the subtle complexities of reversible
O2 binding and inhibition by CO.5 The earliest structurally and
functionally sound iron porphyrin model of the Hb and Mb
active sites was the “picket fence” porphyrin.6ab Dioxygen
affinity studies of this model yielded values similar to those
exhibited by Hb and Mb. However, the CO affinity of picket
(5) For reviews see: (a) Collman, J. P. Acc. Chem. Res. 1977, 10, 265.
(b) Jones, R. D.; Summerville, D. A.; Basolo, F. Chem. ReV. 1979, 79,
139. (c) Collman, J. P.; Halbert, T. R.; Suslick, K. S. In Metal Ion ActiVation
of Dioxygen; Spiro, T. G., Ed.; Wiley: New York, 1980; Chapter 1. (d)
Traylor, T. G. Acc. Chem. Res. 1981, 14, 102. (f) Scheidt, W. R.; Reed, C.
A. Chem. ReV. 1981, 81, 543. (g) Jameson, G. B.; Ibers, J. A. Comments
Inorg. Chem. 1983, 2, 97. (h) Niederhoffer, E. C.; Timmons, J. H.; Martell,
A. E. Chem. ReV. 1984, 84, 7. (i) Baldwin, J. E.; Perlmutter, P. In Topics
in Current Chemistry; Boschke, F. L., Ed.; Springer: Berlin, 1984; p 181.
(l) Suslick, K. S.; Reinert, T. J. Chem. Educ. 1985, 62, 974. (m) Momenteau,
M. Pure Appl. Chem. 1986, 58, 1493. (n) Momenteau, M.; Reed, C. A.
Chem. ReV. 1994, 94, 659. (o) Jameson, G. B. In Metal-Containing
Polymeric Materials; Carraher, C., Zeldin, M., Sheats, J., Culbertson, B.,
Pittman, C. U., Jr., Eds.; Plenum: New York, 1996; p 421.
(6) (a) Collman, J. P.; Gagne, R. R.; Halbert, T. R.; Marchon, J. C.;
Reed, C. A. J. Am. Chem. Soc. 1973, 95, 7868. (b) Collman, J. P.; Gagne,
R. R.; Reed, C. A.; Halbert, T. R.; Lang, G.; Robinson, W. T. J. Am. Chem.
Soc. 1975, 97, 1427. (c) Collman, J. P.; Brauman, J. I.; Iverson, B. L.;
Sessler, J. L.; Morris, R. M.; Gibson, Q. H. J. Am. Chem. Soc. 1983, 105,
3052.
O2
CO
(7) M ) KCO/KO ) P1/2 /P1/2
.
2
† Stanford University.
(8) (a) Collman, J. P.; Brauman, J. I.; Halbert, T. R.; Suslick, K. S. Proc.
Natl. Acad. Sci. U.S.A. 1976, 73, 3333. (b) Collman, J. P.; Brauman, J. I.;
Doxsee, K. M. Proc. Natl. Acad. Sci. U.S.A. 1979, 76, 6035.
‡ University of Nebraska.
X Abstract published in AdVance ACS Abstracts, April 1, 1997.
(1) (a) Perutz, M. F. Nature 1970, 228, 726. (b) Perutz, M. F.; Fermi,
G.; Luisi, B.; Shaanan, B.; Liddington, R. C. Acc. Chem. Res. 1987, 20,
309. (c) Perutz, M. F. Annu. ReV. Biochem. 1979, 48, 327. (d) Perutz, M.
F. Br. Med. Bull. 1976, 32, 195.
(9) (a) Kim, K.; Fettinger, J.; Sessler, J. K.; Cyr, M.; Hugdahl, J.;
Collman, J. P.; Ibers, J. A. J. Am. Chem. Soc. 1989, 111, 403. (b) Collman,
J. P.; Brauman, J. I.; Iverson, B. L.; Sessler, J. L.; Morris, R. M.; Gibson,
Q. H. J. Am. Chem. Soc. 1983, 105, 3052. (c) Tetreau, C.; Lavalette, D.;
Momenteau, M.; Fischer, J.; Weiss, R. J. Am. Chem. Soc. 1994, 116, 11840.
(d) Slebodnick, C.; Duval, M. L.; Ibers, J. A. Inorg. Chem. 1996, 35, 3607.
(e) Hashimoto, T.; Dyer, R. T.; Crossley, M. J.; Baldwin, J. E.; Basolo, F.
J. Am. Chem. Soc. 1982, 104, 2101. (f) Ross, E.; Boitrel, B.; Quelquejeu,
M.; Kossanyi, A. Tetrahedron Lett. 1993, 34, 7267.
(2) Sjostrand, T. Acta Physiol. Scand. 1952, 26, 328.
(3) (a) Metz, G.; Sjostrand, T. Acta Physiol. Scand. 1954, 31, 334. (b)
Coburn, R. F. Ann. N. Y. Acad. Sci. 1970, 174, 11.
(4) (a) Pauling, L. Nature 1964, 203, 182. (b) Phillips, S. E. V.;
Shoenborn, B. P. Nature 1981, 292, 81.
S0002-7863(96)03945-5 CCC: $14.00 © 1997 American Chemical Society

3482 J. Am. Chem. Soc., Vol. 119, No. 15, 1997
Collman et al.
bar and a rubber stopper was charged with R,R,R,R-tetrakis(o-
aminophenyl)porphyrin (1.029 g, 1.53 mmol), CH2Cl2 (80 mL), and
triethylamine (10 mL). Also in the inert atmosphere box, acryloyl
chloride (0.6 mL) was dissolved in CH2Cl2 (40 mL) and loaded into a
50 mL syringe. The syringe and flask were removed from the box,
and the flask was placed under a positive pressure of dry nitrogen.
The acryloyl chloride solution was added to the flask via syringe pump
at 10 mL/h. The solution was stirred for 24 h, washed with saturated
aqueous sodium bicarbonate (4 × 75 mL) and saturated aqueous sodium
chloride (1 × 100 mL), dried over sodium sulfate, filtered, and loaded
directly onto a flash silica gel column which had been prepared as a
CH2Cl2 slurry. The product was eluted with 20% acetone/CH2Cl2. Upon
removal of the solvent, the desired product was obtained (620 mg, 46%).
1H NMR (CDCl3): δ 8.86 (s, 4H); 8.83 (m, 4H); 7.95 (d, J ) 7.2 Hz,
4H); 7.88 (t, J ) 7.3 Hz, 4H); 7.54 (t, J ) 7.4 Hz, 4H); 6.93 (s, 4H);
5.93 (d, J ) 16.6 Hz, 4H); 5.04 (m, 8H); 1.54 (s, 4H); 1.25 (br s, 2H)
water; -2.77 (s, 2H). MS: m/e ) 890.9 (M+) for C56H42N8O4
(LSIMS). Anal. Calcd for C56H42N8O4‚H2O (water identified by
NMR): C, 73.99; H, 4.88; N, 12.33. Found: C, 74.08; H, 4.90; N,
12.08. UV-vis (toluene): λmax 422 (Soret), 516, 548, 588, 644 nm.
Iron Insertion. Iron insertion was performed in an inert atmosphere
box. In a typical reaction, a 100 mL round bottom flask equipped
with a stir bar and 45 cm Vigreux column was charged with the Michael
acceptor porphyrin (140 mg, 0.2 mmol), tetrahydrofuran (50 mL), and
2,6-lutidine (40 drops). The reaction was heated to reflux, and
anhydrous FeBr2 (400 mg, 3 mmol) was added. After the reaction
was heated at reflux for 12 h, the solvent was removed in Vacuo and
the residue dissolved in tetrahydrofuran/benzene (1:10), filtered through
a coarse frit, and loaded directly onto an alumina column (activity I
neutral, 1 cm × 10 cm). The product was eluted with methanol/
tetrahydrofuran/benzene (1:1:10). Upon removal of solvent in Vacuo,
the solid product iron(II) Michael acceptor was obtained (130 mg, 87%).
MS: m/e ) 944.2 (M+) for C56H40N8O4Fe (LSIMS). UV-vis
(toluene): λmax 418, 422 (split Soret), 538 nm.
Four-Nitrogen Aza-crown-Capped Porphyrin Syntheses. The
Michael acceptor (free base or iron(II) derivative) was allowed to react
with cyclen or cyclam aza-crown ethers. The conditions under which
it was performed were identical, with the exception that the free base
reactions were carried out under nitrogen while the syntheses using
the Fe(II)-metalated porphyrin were done in a drybox. In a typical
reaction, Michael acceptor (0.15 mmol) is dissolved in CH3OH/CH2-
Cl2 (25:1, 25 mL) in a 50 mL round bottom flask with a stir bar, an
aza-crown ether (0.75 mmol) added, and the solution heated to reflux.
After 72 h the solution is cooled and the solvent removed in Vacuo.
The resulting solid is dissolved in a minimum amount of CH2Cl2 and
loaded directly onto an alumina column (activity I neutral, 1 cm × 10
cm). The product is eluted with 2% CH3OH/CH2Cl2. Evaporation of
the solvent yields the desired product.
Cyclam-Capped Porphyrin Free Base. Yield: 91%. 1H NMR
(CDCl3): δ 10.1 (s, 4H); 8.95 (d, J ) 8.4 Hz, 4H); 8.81 (s, 4H); 8.78
(s, 4H); 7.77 (t, J ) 7.3 Hz, 4H); 7.49 (d, J ) 6.5 Hz, 4H); 7.30 (t, J
) 7.4 Hz, 4H); 2.18 (br m, 8H); 2.05 (br s, 8H); 0.3 (br s, 8H); -0.2
(br, 4H); -0.5 (br, 4H); -2.2 (br, 2H); -2.5 (br, 2H); -2.76 (s, 2H).
13C NMR (CDCl3): δ 171; 138; 137; 133; 131; 130; 129; 122; 121;
117; 51; 49; 41; 34; 19. MS: m/e ) 1091.5 (M+) for C66H66N12O4
(LSIMS). Anal. Calcd for C66H66N12O4‚CH2Cl2: C, 68.46; N, 14.31;
H, 5.84. Found: C, 68.12; N, 14.02; H, 5.76. UV-vis (toluene): λmax
426 (Soret), 516, 550, 590, 646 nm.
Cyclen-Capped Porphyrin Free Base. Yield: 70%. 1H NMR
(CDCl3): δ 9.55 (s, 4H); 8.88 (d, J ) 8.1 Hz, 4H); 8.83 (s, 8H); 7.87
(d, J ) 7.3 Hz, 4H); 7.83 (t, J ) 7.4 Hz, 4H); 7.44 (t, J ) 7.3 Hz, 4H);
1.99 (s, br, 8H); 1.77 (s, br, 8H); 1.4-0.0 (br region, 8H); 1.25 (br,
6H) water; -0.5 to -1.5 (br peak, 4H); -2.73 (s, 2H); -3 to -3.6 (br
peak, 4H). MS: m/e ) 1063.6 (M+) for C64H62N12O4 (LSIMS). Anal.
Calcd for C64H62N12O4‚3H2O (water identified in NMR): C, 68.80; N,
15.04; H, 6.15. Found: C, 69.29; N, 15.06; H, 5.79. UV-vis
(toluene): λmax 422 (Soret), 514, 546, 590, 642 nm.
ditional evidence for the plausibility of this hypothesis. The
decreased size of the gas binding cavity in the pocket porphyrins
resulted in M values on the order of 200. Infrared spectral
studies of these compounds were consistent with a CO more
weakly bound than in the case of picket fence porphyrin.
Various X-ray crystallographic studies of CO bound Hb and
Mb purported to observe a steric interaction in the CO bound
proteins. The CO unit is reported to be tipped off axis, relative
to the heme plane normal, from 7° to 47°.10 Recent interpreta-
tion of the available structural data favors a smaller deviation
from the heme plane normal.9d These recently estimated angles
have spawned considerable debate as to whether steric or
electrostatic effects play the greater role in determining the CO
affinity of Hb and Mb. Spiro et al. have suggested that a polar
interaction between the terminal oxygen atom of bound CO and
the lone pair on the distal histidine plays a substantial role in
determining the CO affinity of Mb. They have argued that off-
axis bending of the Fe-CO unit, while drastically reducing the
strength of the Fe-C bond, is energetically very costly to the
protein moiety and hence unlikely.11 In addition Traylor has
demonstrated that the polarity of the environment surrounding
the gas binding pocket has a large effect on the M value of
model complexes. He argued the M value in his studies is
influenced by an increase in O2 affinity with negligible change
in CO affinity.12 The presence of water molecules in the binding
pocket has also been suggested as contributing to the relative
O2 and CO affinities. The effects of water in the distal pocket
have been evaluated in a series of E7 Mb mutants and have
been advanced as the water-displacement model.13 In the
present paper we demonstrate that steric effects alone can
drastically reduce the carbon monoxide affinity of an iron(II)
porphyrin complex and hence its M value.
Experimental Section
Materials and Methods. All oxygen sensitive work was performed
in a N2-filled drybox kept at or below 0.5 ppm O2. All chemicals were
purchased commercially (Aldrich, Acros) and used as received unless
otherwise noted. Solvents were distilled under a nitrogen or argon
atmosphere from the indicated reagents immediately prior to use:
methylene chloride (P2O5); benzene, tetrahydrofuran, and toluene
(sodium/benzophenone ketyl); methanol (Mg). R,R,R,R-Tetrakis(o-
aminophenyl)porphyrin was prepared according to the literature
procedure.14 1H NMR and 13C NMR spectra were obtained on a Nicolet
WB-300 or a Varian XL-400 spectrometer and referenced to residual
proton solvents. Mass spectra were taken at the University of
California, San Francisco, Mass Spectrometry Facility. UV-vis spectra
were recorded on a Hewlett-Packard 8452A diode array spectrometer.
Elemental analysis was performed by Midwest Microlabs (Indianapolis,
IN).
Synthesis and Characterization. Michael Acceptor. In an inert
atmosphere box, a 500 mL round bottom flask equipped with a stir
(10) (a) Heidner, E. J.; Ledner, R. C.; Perutz, M. F. J. Mol. Biol. 1976,
104, 707. (b) Kuriyan, J.; Wilz, S.; Karplus, M.; Petsko, G. A. J. Mol.
Biol. 1986, 192, 133. (c) Teng, T. -Y.; Srajer, V.; Moffat, K. Nat. Struct.
Biol. 1994, 1, 701. (d) Cheng, X.; Schoenborn, B. P. J. Mol. Biol. 1991,
220, 381. (e) Quillin, M. L.; Arduini, R. M.; Olson, J. S.; Phillips, Jr. J.
Mol. Biol. 1993, 234, 140. (f) Schlichting, I.; Berendzen, J.; Phillips Jr.;
Sweet, R. M. Nature 1994, 371, 808. (g) Ivanov, D.; Sage, J. T.; Keim,
M.; Powel, J. R.; Asher, S. A.; Champion, P. M. J. Am. Chem. Soc. 1994,
116, 4139. (h) Derenenda, Z.; Dodsong.; Elmsley, P.; Harris, D.; Nagai,
K.; Perutz, M. F.; Reanud, J. P. J. Mol. Biol. 1990, 211, 515. (i) Lim, M.;
Jackson, T. A.; Anfinrud, P. A. Science 1995, 269, 962.
(11) (a) Ray, G. B.; Li, X.-Y.; Ibers, J. A.; Sessler, J. L.; Spiro, T. G. J.
Am. Chem. Soc. 1994, 116, 162. (b) Spiro, T. G. Science 1995, 270, 221.
(12) (a) Traylor, T. G. Acc. Chem. Res. 1981, 14, 102. (b) Traylor, T.
G.; Koga, N.; Dearduff, L. A. J. Am. Chem. Soc. 1985, 107, 6504.
(13) Springer, B. A.; Sligar, S. G.; Olson, J. S.; Phillips, G. N. Chem.
ReV. 1994, 94, 699.
Three-Nitrogen Aza-crown Porphyrin Capping Synthesis. The
synthesis is identical to that of the four-nitrogen aza-crown porphyrin
capping synthesis described above, except 1.1 equiv of the three-
nitrogen aza-crown was used rather than 6 equiv as in the case of the
four-nitrogen models.
(14) (a) Linsey, I. J. Org. Chem. 1980, 45, 5215. (b) Elliot, C. M. Anal.
Chem. 1980, 52, 666.

Aza-Crown-Capped Porphyrin Models of Myoglobin
J. Am. Chem. Soc., Vol. 119, No. 15, 1997 3483
1,4,7-Triazacyclononane-Capped Porphyrin Free Base. Yield:
54%. 1H NMR (CDCl3): δ 10.2 (s, 2H); 8.88-8.73 (m, 8H); 8.50 (d,
J ) 8.1 Hz, 2H); 8.43 (d, J ) 8.1 Hz, 2H); 7.9-7.5 (m, 4H); 7.65 (t,
J ) 7.3 Hz, 2H); 7.6-7.2 (br region, 6H); 6.0 (br, 1H); 5.1 (br, 2H);
2.1 (s, 2H); 1.9 (s, 2H); 1.7-1.4 (br m, 8H); 1.2 (m, 6H); 0.8 (m, 2H);
0.4 (br, 2H); -1.2 (br, 2H); -2.4 (s, 2H); -2.8 (br, 2H). MS: m/e )
1020.5 (M+) for C62H57N11O4. Anal. Calcd for C62H57N11O4‚
CH2Cl2‚1H2O: C, 67.37; N, 13.72; H, 5.47. Found: C, 67.80; N, 13.67;
H, 5.40. UV-vis (toluene): λmax 424 (Soret), 518, 552, 592, 650 nm.
CuBr Insertion. The porphyrin was dissolved in acetonitrile and
heated at reflux. Twenty mole equivalents of CuBr was added, and
the solution was left at reflux for 24 h. The solution was then cooled
to room temperature, solvent removed under vacuum, and the resulting
solid suspended in CH2Cl2. The suspension was filtered through a fine
frit with a 2 cm Celite pad. The filtrate, now free of excess CuBr, was
pumped dry under vacuum.
Synthesis of Tailed 1,4,7-Triazacyclononane-Capped Porphyrin.
1,4,7-Triazacyclononane-capped porphyrin was dissolved in methanol,
3-(aminomethyl)pyridine (1.2 equiv) was added, and the solution was
heated at reflux for 24 h. The reaction mixture was cooled and the
solvent pumped off under vacuum. The resulting solid was dissolved
in benzene, loaded onto a neutral activity I alumina column, and eluted
with 2% methanol in benzene. The solvent was removed under
vacuum, yielding the desired compound.
Å, c ) 22.501(4) Å, â ) 103.821(7)°, V ) 12970(4) Å3, and Z ) 8
{dcalcd ) 1.369 g/cm3; µa(Mo KR) ) 0.45 mm-1}. A full hemisphere
of diffracted intensities (1270 10 s frames with an ω scan width of
0.30°) was measured to 0.90 Å resolution using graphite-monochro-
mated Mo KR radiation on a Siemens SMART CCD area detector.
X-rays were provided by a normal focus sealed X-ray tube operated at
50 kV and 40 mA. Lattice constants were determined with the Siemens
SMART software package using peak centers for 50 reflections. A
total of 24461 integrated reflection intensities having 2θ (Μï ΚR) <
46.51° were obtained using the Siemens program SAINT. A total of
9185 of these were independent and gave Rint ) 0.038. The Siemens
SHELXTL-PC software package was used to solve the structure using
the “heavy atom” technique and to conduct initial least-squares re-
finements using Fo data. The final stages of weighted full-matrix least-
squares refinement were conducted using Fo2 data and the SHELTXTL
Version 5 software package and converged to give R1(unweighted, based
on F) ) 0.052 for 7466 independent reflections having 2θ (Mo KR) <
46.51° and I > 2σ(I) and wR2(weighted, based on F2) ) 0.131 for
8894 independent reflections having 2θ (Mo KR) < 46.51° and I > 0.
The structural model incorporated anisotropic thermal parameters for
all non-hydrogen atoms and isotropic thermal parameters for all included
hydrogen atoms. Several of the water and methanol solvent molecules
of crystallization are disordered. The hydrogen atoms bonded to
carbons of the porphyrin ligand were included in the structural model
at idealized sp2- or sp3-hybridized positions using a C-H bond length
of 0.96 Å with their isotropic thermal parameters fixed at values 1.2
times the equivalent isotropic thermal parameters of the carbon atom
to which they are bonded. Hydrogen atoms bonded to the four amide
nitrogens of the porphyrin ligand and the hydrogens on the coordinated
water molecule were located from difference Fourier synthesis and
included in the least-squares refinement cycles as independent isotropic
atoms. Although hydrogens bonded to the oxygen (O2W) for one of
the lattice water molecules could be located from a difference Fourier
synthesis, they could not be satisfactorily refined as independent
isotropic atoms. These two hydrogens were therefore included in the
structural model with the coordinates fixed at the Fourier values and
isotropic thermal parameters fixed at values which were 1.2 times the
equivalent isotropic thermal parameter of oxygen atom O2W. Hydrogen
atoms on the remainder of the solvent molecules of crystallization were
not located or included in the structural model.
1,4,7-Triazacyclononane-Capped Pyridine-Tailed Porphyrin Free
Base. MS: m/e ) 1128.4 (MH+) for C68H63N13O4 (LSIMS).
Iron(II)-Metalated 1,4,7-Triazacyclononane-Capped Porphyrin
1. 1H NMR (pyridine--d5): δ 45-50 (m); characteristic of â-pyrrolic
protons on a paramagnetic 5-coordinate iron (S ) 2) porphyrin. MS:
m/e ) 1074.3 (M+) for C62H57N11O4Fe (LSIMS). UV-vis (toluene):
λ
max 420, 448 (split Soret), 538 nm. UV-vis (toluene, 1000 equiv of
1,2-dimethylimidazole): λmax 444 (Soret), 566 nm.
Iron(II)-Metalated Cyclam-Capped Porphyrin 2. 1H NMR
(pyridine-d5): δ 49 (s); 51 (s); characteristic of â-pyrrolic protons on
a paramagnetic 5-coordinate iron (S ) 2) porphyrin. MS: m/e ) 1145.5
(M+) for C66H64N12O4Fe (LSIMS). UV-vis (toluene): λmax 420, 448
(split Soret), 538 nm. UV-vis (toluene, 1000 equiv of 1,2-dimethyl-
imidazole): λmax 444 (Soret), 566 nm. The sample was exposed to
oxygen, and the resulting µ-oxo iron(III) dimer was precipitated from
benzene/methanol/THF and sent for elemental analysis. Anal. Calcd
for (FeC62H57N11O4)2O‚2benzene: C, 70.24; N, 13.65; H, 5.73.
Found: C, 69.92; N, 13.43; H, 5.68.
Results and Discussion
A series of porphyrins (1-6, Figure 1) were synthesized
according to the “congruent multiple Michael addition” method
previously reported.15 Carbon monoxide and dioxygen gas
binding to their Fe(II) derivatives was studied preliminarily by
1H NMR spectroscopy in pyridine-d5, which acts as both the
solvent and the axial ligand. After the spectrum indicative of
a 5-coordinate iron(II) species was recorded, the sample was
exposed to CO or O2 and a second spectrum recorded. The
5-coordinate porphyrin-bound iron(II) is high-spin and has an
NMR spectrum indicative of a paramagnetic S ) 2 ground state
as evidenced by â-pyrrolic proton signals at ca. 45-55 ppm
downfield relative to TMS. When CO or O2 binds to iron, the
resulting complexes are diamagnetic and the NMR spectra are
characteristic of an S ) 0 system.16 Typical spectra are shown
in Figure 2. The results lead to the important conclusion that
the cyclam-capped porphyrin 2 has no detectable affinity for
CO at pressures up to 1 atm!
Iron(II)-Metalated Cyclen-Capped Porphyrin 3. UV-vis
(toluene): λmax 420, 448 (split Soret), 538 nm. UV-vis (toluene, 1000
equiv of 1,2-dimethylimidazole): λmax 444 (Soret), 566 nm.
Iron(II)-Metalated CuBr 1,4,7-Triazacyclononane-Capped Por-
phyrin 4. 1H NMR (pyridine-d5): δ 45-50 (m); characteristic of
â-pyrrolic protons on a paramagnetic 5-coordinate iron (S ) 2)
porphyrin. MS: m/e ) 1217.3 (M+) for C62H57N11O4FeCuBrO2
(LSIMS). The spectra matched the calculated isotope distribution
exactly. UV-vis (toluene): λmax 420, 448 (split Soret), 538 nm. UV-
vis (toluene, 1000 equiv of 1,2-dimethylimidazole): λmax 444 (Soret),
566 nm.
Iron(II)-Metalated 1,4,7-Triazacyclononane-Capped Pyridine-
Tailed Porphyrin 5. MS: m/e ) 1182.6 (MH+) for C68H63N13O4Fe
(LSIMS). UV-vis (toluene): λmax 444 (Soret), 566 nm.
Iron(II)-Metalated CuBr 1,4,7-Triazacyclononane-Capped Pyr-
idine-Tailed Porphyrin 6. UV-vis (toluene): λmax 444 (Soret), 566
nm.
Zinc(II) Cyclam-Capped Porphyrin 7. A 100 mL flask is charged
with methanol (30 mL) and Michael acceptor porphyrin (100 mg) and
heated at 50 °C for 2 h. The solvent is removed in Vacuo and the
resulting solid dissolved in methylene chloride, placed directly onto a
silica column (1 cm × 15 cm), and eluted with chloroform/methanol
(20:1). The Zn(II) Michael acceptor is capped according to the
procedure described above. 1H NMR (CD2Cl2): δ 8.70 (d, 8 H); 8.75
(d, 4H); 7.70 (t, 4H); 7.50 (d, 4H); 7.270 (t, 4H); 1.95-2.05 (br m,
28H); 0.20 (br m, 8H); -0.5 (br s, 4H).
X-ray Crystal Structure Determination of Zinc Cyclam-Capped
Porphyrin. Single crystals of H2O/CH3OH/CH2Cl2-solvated Zn-
(cyclam-capped porphyrin)(OH2) (7) are, at 20 ( 1 °C, monoclinic,
space group C2/c-C2h6 (no. 15), with a ) 47.088(7) Å, b ) 12.606(2)
(15) Collman, J. P.; Zhang, X; Herrmann, P. C.; Uffelman, E. S.; Boitrel,
B; Straumanis, A; Brauman, J. I. J. Am. Chem. Soc. 1994, 116, 2681.
(16) (a) LaMar, G. N.; Eaton, G. R.; Holm, R. H.; Walker, F. A. J. Am.
Chem. Soc. 1973, 95, 63. (b) LaMar, G, N.; Walker, F. A. J. Am. Chem.
Soc. 1973, 95, 1782. (c) Goff, H.; LaMar, G. N.; Reed, C. A. J. Am. Chem.
Soc. 1977, 99, 3641. (d) Goff, H.; LaMar, G. N. J. Am. Chem. Soc. 1977,
99, 6599. (e) Collman, J. P.; Brauman, J. I.; Collins, T. J.; Iverson, B. L.;
Lang, G.; Pettman, R. B.; Sessler, J. L.; Walter, M. A. J. Am. Chem. Soc.
1983, 105, 3038. (f) A sharp diamagnetic 1H NMR spectrum at room
temperature is probably the quintessential criterion of a pure iron(II)
dioxygen complex: Collman, J. P.; Brauman, J. I.; Doxsee, K. M.; Halbert,
T. R.; Bunnenberg, E.; Linder, R. E.; LaMar, G. N.; Del Gaudio, J.; Lang,
G.; Spartalian, K. J. Am. Chem. Soc. 1980, 102, 4182.

3484 J. Am. Chem. Soc., Vol. 119, No. 15, 1997
Collman et al.
Figure 1. Chemical structures of the compounds referred to in the text. An R-state model results when L ) 1,5-dicyclohexylimidazole, and a
T-state model results from L ) 1,2-dimethylimidazole.
Typical data are shown in Figure 3. Ten micromolar porphyrin
solutions in toluene with 1000 molar equivalents of axial ligand
were titrated with pure O2 and CO. The resulting P1/2 and M
values are tabulated in Table 1 along with literature values for
Hb, Mb, and pertinent model porphyrin complexes. Carbon
monoxide pressures greater than 1 atm were not employed; thus,
the reported M values represent a lower limit for the cyclam
(2) and cyclen (3) capped porphyrins. Even though a very large
excess of axial ligand was used, the iron(II) remained 5-coor-
dinate until addition of O2 or CO, because the macrocycle-
porphyrin cavity excludes imidazole bases, and thus the
formation of a 6-coordinate Fe(II) complex. In addition, the
oxygenated (carbonylated) species are stable in solution, for over
10 days.
The M values reported in Table 1 are similar whether the
axial ligand is 1,5-dicyclohexylimidazole (1,5-DCIm), 1,2-
dimethylimidazole (1,2-Me2Im), or a covalently attached pyri-
dine tail. Pyridine has a different trans labilizing influence on
Fe than does imidazole.18 Hence, pyridine acts very differently
as an axial ligand than imidazole. Even though the affinity of
(17) (a) Rose, N.; Drago, R. S. J. Am. Chem. Soc. 1959, 81, 6138. (b)
Figure 2. 1H NMR spectra of O2 bound (a) and O2-free (b) complexes
Beugelsdijk, T. J.; Drago, R. S. J. Am. Chem. Soc. 1975, 97, 6466. (c)
Collman, J. P.; Brauman, J. I.; Suslick, K. S. J. Am. Chem. Soc. 1975, 97,
7185.
(18) (a) Rougee, M.; Brault, D. Biochemistry 1975, 18, 4100. (b)
Collman, J. P.; Brauman, J. I.; Doxsee, K. M.; Sessler, J. L.; Morris, R.
of compound 2 in pyridine-d5. Spectra are typical of compounds 1-6.
Note the absence of â-pyrrolic proton signals in (a).
Dioxygen and carbon monoxide binding affinity values were
determined using a tonometer and UV-vis spectroscopy.17
M.; Gibson, Q. H. Inorg. Chem. 1983, 22, 1427.

Aza-Crown-Capped Porphyrin Models of Myoglobin
J. Am. Chem. Soc., Vol. 119, No. 15, 1997 3485
Figure 3. UV-vis spectra of O2-free and O2 bound compound 1 with
1000 equiv of 1,2-dimethylimidazole in toluene. Spectra are typical of
compounds 1-6.
the Fe for the axial ligand is almost identical for pyridine and
imidazole, the observed insensitivity of the M value to the axial
ligand suggests that the M value is predominantly determined
by the porphyrin cavity structure.
Also interesting is the relatively small 7-fold increase in O2
binding affinity on changing from 1,2-Me2Im (T-state model)
to 1,5-DCIm (R-state model) as the axial ligand for compound
2. Most reported model complexes demonstrate a 10-100-fold
increase under the same conditions.19 The 10-fold effect
observed for picket fence porphyrin has been attributed to the
geometric relationship of the Fe to the porphyrin plane (see
Figure 4). Upon O2 binding in the T-state picket fence model,
the Fe atom is not permitted to become coplanar with the
porphyrin. Instead, the Fe atom is forced to remain slightly on
the proximal side of the porphyrin away from the bound O2.20
The R-state picket fence model on the other hand permits the
Fe atom to be in the plane of the porphyrin.
In the aza-crown-capped models, however, an R-state axial
ligand which permits the Fe to be coplanar with the porphyrin
would require the terminal end of the bound O2 molecule to
extend out, closer to the cap. While the Fe geometry of an
R-state model increases the affinity of the Fe for O2, the terminal
end of the bound gas molecule in these models would be forced
closer to the cap where it could be sterically destabilized.
Such steric interactions with the intrisically linear CO group
appeared to be the origin of the differing M values in Table 1.
Consequently, a more detailed structural understanding of cavity
dimensions for these aza-crown models was needed. Since zinc
Figure 4. Selected dimensions of R- and T-state picket fence porphyrin
complexes as determined by X-ray crystallography. See ref 19 for more
details.
porphyrins are easier to crystallize than iron porphyrins and
because zinc porphyrin complexes are air and moisture stable,
the Zn analogue of 2 (7) was synthesized. X-ray quality crystals
of 7 were obtained by vapor diffusion of isopropyl alcohol into
a saturated solution of 7 [CH2Cl2/CH3OH (3:1)] at room
temperature.
A single-crystal X-ray structural analysis of 7 (Figure 5)
revealed that it was composed of discrete molecules of the aquo
zinc analogue of 2 with H2O, MeOH, and/or dichloromethane
solvent molecules of crystallization surrounding it in the lattice.
The Zn2+ ion is 5-coordinate, with the pyrrole nitrogens of the
porphyrin macrocycle occupying the four basal sites and the
oxygen (O1W) of a coordinated water molecule occupying the
axial site of a square pyramid. The four pyrrole nitrogens are
coplanar to within 0.004 Å, and the Zn atom is displaced from
their least-squares mean plane by 0.336 Å toward the apical
oxygen which is on the unhindered (proximal) side of the
porphyrin. The cyclam cap is attached through its nitrogens to
the distal side of the porphyrin by N-phenylacrylamide bridges
to the meso-methine carbons of the porphyrin core. By linking
the two macrocycles in this fashion, one might expect to produce
an approximately staggered arrangement for the two sets of four
(pyrrole or amine) nitrogens. The conformation actually
observed is one midway between staggered and eclipsed. This
(19) Appleby, C. E.; Blumberg, W. E.; Bradbury, J. H.; Fuchsman, W.
H.; Peisach, J.; Wittenberg, B. A.; Wittenberg, J. B.; Wright, P. E. In
Hemoglobin and Oxygen Binding; Ho, C., Ed.; North-Holland: Amsterdam,
1982; p 435.
(20) (a) Collman, J. P.; Gagne, R. R.; Reed, C. A.; Robinson, W. T.;
Rodley, C. A. Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 1326. (b) Jameson,
G. B.; Molinaro, F.; Ibers, J. A.; Collman, J. P.; Brauman, J. I.; Rose, E.;
Suslick, K. S. J. Am. Chem. Soc. 1980, 102, 3224. (c) Jameson, G. B.;
Rodley, G. A.; Robinson, W. T.; Gagne, R. R.; Reed, C. A.; Collman, J. P.
Inorg. Chem. 1978, 17, 850. (d) Jameson, G. B.; Molinaro, F. S.; Ibers, J.
A.; Collman, J. P.; Brauman, J. I.; Rose, E.; Suslick, K. S. J. Am. Chem.
Soc. 1978, 100, 6769. (e) Collman, J. P.; Reed, C. A. J. Am. Chem. Soc.
1973, 95, 2048.

3486 J. Am. Chem. Soc., Vol. 119, No. 15, 1997
Collman et al.
Table 1. Gas Binding Data for Compounds 1-6 and Related Compounds of Interest from the Literaturea
P1/2(O2)
(Torr)
P1/2(CO)
(Torr)
M
bound CO stretching
(P1/2(O2)/P1/2(CO))
frequency (cm-1
)
T-State Model
0.3
0.0089
100
Hb(T)b
40
38
280
2.3
135c
Fe(picket fence) (1,2-Me2Im)
Fe(4-atom-linked cap) (1-MeIm)
1 (1,2-Me2Im)
4280d
2.8e
0.79
2.9
1956
1965
2 (1,2-Me2Im)
3 (1,2-Me2Im)
4 (1,2-Me2Im)
22
760
irreversible
>3500
>3500
5.9
<0.006
<0.22 at 4 °C
NA
R-State Model
0.0014
Hb(R)b
0.22
150c
Mbb
0.37-1
0.014-0.025
20-40f
1 (1,5-DCIm)
2 (1,5-DCIm)
3 (1,5-DCIm)
4 (1,5-DCIm)
<0.2
∼3
>3500
>3500
3.6
<0.003
<0.22
NA
∼760
irreversible
Pyridine-Tailed Compounds
5
6
3.7
irreversible
1.5
2.8
2.47
NA
a Note all data were obtained at 25 °C except the T-state compound 3 which was measured at 4 °C. b Measured in H2O at pH 7. c Sharma, V.
S.; Schmidt, M. R.; Ranney, H. M. J. Biol. Chem. 1976, 251, 4267. Steinmeier, R. C.; Parkhurst, L. J. Biochemistry 1975, 14, 1564. d Collman,
J. P.; Brauman, J. I.; Inverson, B. L.; Sessler, J. L.; Morris, R. M.; Gibson, Q. H. J. Am. Chem. Soc. 1983, 105, 3052. e Johnson, M. R.; Seok, W.
K.; Ibers, J. A. J. Am. Chem. Soc. 1991, 113, 3998. f LaGow, J.; Parkhurst, L. J. Biochemistry 1972, 11, 4520.
can be seen in Figure 5a and by the facts that the Zn, C10, C20,
N31, and N38 atoms are coplanar to within 0.03 Å but the Zn,
C5, C15, N28, and N35 atoms have the following displacements
from their least-squares mean plane: 0.01, -0.41, 0.40, 0.46,
and -0.45 Å. This arrangement has one diagonal of the
rectangular array of four cyclam nitrogens aligned parallel to
one diagonal of the square array of four pyrrole nitrogens and
reduces the number of symmetry elements for possibly relating
the non-hydrogen atoms of 7. Although the non-hydrogen
atoms of 7 could ideally be related by C2V symmetry, only C2
symmetry is approximated in the crystal, with the pseudo-C2
axis passing through the Zn atom, the oxygen atom of the
coordinated water molecule, and the centers of gravity for the
four coordinated pyrrole nitrogens and the four amine nitrogens
of the cyclam cap.
Although 7 does not bind O2 or CO, it should be a reasonable
structural analogue of the 5-coordinate deoxygenated high-spin
iron(II) cyclam-capped porphyrin. The average Zn-Np21 bond
length is 2.069 (5, 3, 3, 4) Å,22 the Zn-O bond length is
2.092(4) Å, and the Zn-Ct23 and Ct-Np distances are 0.336
and 2.041 (-, 3, 5, 4) Å, respectively. These values are close
to those observed for similar bonds in other 5-coordinate zinc(II)
porphyrin complexes,24 and to those of high-spin 5-coordinate
iron(II) porphyrin models of oxygen binding proteins.20b,c The
four N atoms of the cyclam cap are coplanar to within 0.12 Å,
and their least-squares mean plane is within 0.9° of being parallel
to the least-squares mean plane through the four pyrrole
nitrogens. The 24 carbon and nitrogen atoms of the porphyrin
core are coplanar to within 0.21 Å; this 24-atom core is pseudo-
S4-ruffled and within 0.5° of being parallel to the two N4
groupings of the porphyrin and cyclam moieties. The mean
planes of the 24-atom porphyrin core and the 14-atom cyclam
cap are separated by 4.05 Å and within 0.4° of being parallel.
The bond lengths and angles presented in Table 2 reveal the
presence of highly precise, but otherwise unexceptional, metrical
parameters for the capped porphyrin ligand of 7.
An examination of nitrogen-nitrogen separations for the
cyclam cap of 7 is informative. The two pairs of unique
nitrogen-nitrogen separations in the cyclam cap which cor-
respond to n values of 2 and 3 for the N-(CH2)n-N chains,
average 3.83 (4, 0, 0, 2) and 4.93 (6, 3, 3, 2) Å, respectively.
These values can be compared with values of 3.68 and 4.94 Å
which can be calculated for these same nitrogen-nitrogen
separations assuming totally extended N-(CH2)n-N chains in
the cap with tetrahedral carbons and C-N and C-C bond
lengths of 1.48 and 1.54 Å, respectively. These nitrogen-
nitrogen separations indicate a highly stretched conformation
for the cyclam cap which places the four nitrogens as far as
possible away from each other and as far as possible from the
pseudo-C2 axis of the molecule. The slightly longer observed
nitrogen-nitrogen separation for the n ) 2 chains is the result
of angles at the carbons which are slightly larger than the
idealized tetrahedral value.
The steric requirements of attaching the cyclam cap to the
four N-phenylacrylamide bridges from the methine carbons are
partially responsible for this stretched conformation, but another
factor appears to be the formation of hydrogen bonds (Table 3)
in these bridges. The amide N-H bond in each of the
acrylamide bridges is pointed at the lone pair of the nearest
cyclam nitrogen. The four (amide)N‚‚‚N(amine) separations
range from 2.80 to 3.01 Å, and the (amide)N-H‚‚‚N(amine)
angles range from 142° to 149°. Each amide nitrogen is
coplanar (to within 0.07 Å) with the three atoms covalently
bonded to it as well as the hydrogen-bonded cyclam nitrogen.
These hydrogen bonds produce a kink in each of the acrylamide
links to the cyclam cap. While hydrogen-bonded in this manner,
these acrylamide links are less flexible, and the distance which
they can span is reduced. The net effect of these hydrogen
bonds is to “stretch” the easily deformed cyclam cap radially
away from the pseudo-C2 axis, thereby providing maximal
separation between the “linked” amine nitrogens in the cap; they
(21) The symbol Np is used to designate a pyrrole nitrogen of the
porphyrin core.
(22) The first number in parentheses following an average value of a
bond length or angle is the root-mean-square estimated standard deviation
of an individual datum. The second and third numbers are the average and
maximum deviation from the average value, respectively. The fourth number
represents the number of individual measurements which are included in
the average value.
(23) The symbol Ct is used to represent the center of gravity for the
four coordinated pyrrole nitrogens of the porphyrin core.
(24) Glick, M. D.; Cohen, G. H.; Hoard, J. L. J. Am. Chem. Soc. 1967,
89, 1996.
(25) Pauling, L. The Nature of the Chemical Bond, 3rd ed.; Cornell
University: Ithaca, NY, 1960; p 260.

Aza-Crown-Capped Porphyrin Models of Myoglobin
J. Am. Chem. Soc., Vol. 119, No. 15, 1997 3487
Table 2. Selected Average Bond Lengths (Å) and Angles (deg)a
Involving Non-Hydrogen Atoms in Crystalline Zn(cyclam-capped
porphyrin)(OH2) (7)b,c
parameter
value
parameter
value
Bond Lengths (Å)
Zn-N21
Zn-O1W
C2-C3
C5-C1A
C6A-N1A
C7A-O1A
2.069 (3, 4, 7, 4)
2.092 (3)
C1-N21
1.374 (4, 4, 8, 8)
1.438 (5, 3, 5, 8)
1.401 (4, 6, 11, 8)
0.81 (4, 6, 12, 4)
1.348 (6, 14, 18, 4)
C1-C2
1.343 (5, 1, 1, 4)
1.506 (5, 6, 7, 4)
1.411 (5, 3, 7, 4)
1.224 (6, 8, 16, 4)
C4-C5
N1A-H1A
C7A-N1A
Bond Angles (deg)
N21ZnN22
N21ZnN23
N21C4C5
C3C4C5
88.5 (1, 9, 12, 4)
161.3 (1, 2, 2, 2)
125.5 (3, 2, 5, 8)
125.2 (3, 3, 6, 8)
107.4 (3, 2, 8, 8)
122.9 (3, 5, 6, 4)
N21ZnO1W 99.3 (1, 15, 17, 4)
ZnN21C4
N21C4C3
C4C5C6
126.2 (2, 11, 16, 8)
109.3 (3, 2, 4, 8)
125.5 (3, 8, 13, 4)
118.8 (3, 5, 9, 4)
C2C3C4
C2AC1AC5
C6AC1AC5
a Bond lengths and angles in the porphyrin core and N-phenylacryl-
amide links have been averaged according to the idealized C4V symmetry
of the “uncapped” ligand. Entries in the table are for one unique bond
length or angle of each set. b See ref 22 for an explanation of the
numbers in parentheses following each averaged value. c Atoms are
labeled in agreement with Figure 5.
accommodate a CO which is linearly-bonded to a nearly in-
plane Fe2+ ion can be demonstrated by making selective
alterations to the structure of 7: (1) place an Fe2+ ion at the
center of gravity (Ct) of the four pyrrole nitrogens; (2) place
the carbon and oxygen atoms of a CO ligand at distances26 of
1.77 and 2.89 Å, respectively, away from the Fe2+ ion along
the pseudo-C2 axis of the molecule on the distal side of the
porphyrin. Assuming no alteration in the cyclam-capped
porphyrin ligand, such an arrangement produces a series of very
short O‚‚‚H and O‚‚‚C contacts between the CO ligand and the
cyclam cap on the porphyrin: O‚‚‚H26a, 1.38 Å; O‚‚‚H29a, 1.98
Å; O‚‚‚H33a, 1.37 Å; O‚‚‚H36b, 1.59 Å; O‚‚‚C26, 2.30 Å; O‚‚‚C29,
2.75 Å; O‚‚‚C33, 2.29 Å; O‚‚‚C34, 2.93 Å; O‚‚‚C36, 2.39 Å; and
O‚‚‚C37, 2.80 Å. These can be compared with the van der Waals
values24 of 2.60, 3.10, and 3.40 Å for O‚‚‚H, O‚‚‚C, and
O‚‚‚methyl contacts, respectvely. Such short nonbonded con-
tacts would obviously preclude the presence of a CO ligand
which is linearly bonded to the Fe unless the conformation of
the porphyrin ligand changed dramatically. Significant struc-
tural changes in the porphyrin ligand could result from a
disruption of the hydrogen bonds present in the acrylamide links
followed by rotations about the appropriate C-C and C-N
single bonds in these links and the cyclam cap. It is not obvious
that such structural alterations would increase the size of the
gas binding cavity sufficiently to accommodate a CO ligand
which is linearly-bonded to a nearly in-plane Fe. This cavity
could obviously more easily accommodate a bent terminally-
bonded O2 ligand20b with an Fe-O bond length of 1.898 Å and
an Fe-O-O angle of 129°.
As noted above, the gas binding cavity in 7 contains four
nearly equivalent van der Waals pockets at the ends of its
diagonals near the amine nitrogens and amide protons. The
most favorable orientation for a bent O2 ligand would presum-
ably be one which directs the noncoordinated oxygen into one
of these pockets. This arrangement should produce the smallest
number of short O‚‚‚H and O‚‚‚C contacts between the O2 ligand
and the cyclam cap. Four such orientations exist for a bent O2
ligand about the pseudo-C2 axis of 7; each of these places the
Fe-O-O grouping near a plane which passes through opposite
pairs of methine carbons and is oriented perpendicular to the
porphyrin core. This also happens to be the orientation for a
Figure 5. (a, top) Perspective drawing of the solid-state structure for
the Zn(cyclam-capped porphyrin)(OH2) molecule (7) present in its
crystalline H2O/CH3OH/CH2Cl2 solvate. The Zn atom is represented
by medium-sized shaded sphere, and carbon and hydrogen atoms are
represented by medium-sized and small open spheres, respectively. The
molecule is viewed nearly parallel to the pseudo-C2 axis of the molecule
which passes through the Zn atom and the oxygen (O1W) of the
coordinated water molecule. (b, bottom) Perspective drawing of the
solid-state structure for the Zn(cyclam-capped porphyrin)(OH2) mol-
ecule (7) present in its crystalline H2O/CH3OH/CH2Cl2 solvate. All
atoms are represented as in (a). The molecule is viewed nearly
perpendicular to the pseudo-C2 axis of the molecule which passes
through the Zn atom and the oxygen (O1W) of the coordinated water
molecule.
also reduce the size of the gas binding cavity on the distal side
of the porphyrin by reducing the separation between the cyclam
cap and the porphyrin core.
The sterically congested nature of the interior surface of the
gas binding cavity in 7 can be seen from the space-filling
drawing in Figure 6. Examination of a similar space-filling
drawing for the interior of the gas binding cavity (with the Zn,
coordinated water, and porphyrin core deleted) shows it to be
roughly rectangular in shape with four “van der Waals pockets”
at the ends of its diagonals near the amine nitrogens and amide
protons. It also contains four cyclam methylene groups (C26,
C29, C33, and C36 and their hydrogens) clustered around the
pseudo-C2 axis of the molecule. One hydrogen from each of
these four methylenes protrudes into the cavity and makes H‚‚‚H
contacts of 2.20-2.51 Å with at least two of the other three.
These H‚‚‚H contacts are comparable to the 2.40 Å van der
Waals diameter of hydrogen25 and are probably responsible for
the reduction in molecular symmetry mentioned above. These
four methylene groups would presumably have unfavorable
steric interactions with a terminally-bonded CO ligand on the
distal side of the porphyrin. The inability of this cavity to
(26) Peng, S. -M.; Ibers, J. A. J. Am. Chem. Soc. 1976, 98, 8032.

3488 J. Am. Chem. Soc., Vol. 119, No. 15, 1997
Collman et al.
Table 3. Intramolecular Hydrogen-Bonding Interactions in N-Phenylacrylamide Links of Crystalline Zn(cyclam-cappped porphyrin)(OH2) (7)a
donor atom
(D)b
acceptor atom
(A)
distance (Å)
(D‚‚‚A)
distance (Å)
(H‚‚‚A)
angle (deg)
(D-H‚‚‚A)
angle (deg)
(H-D‚‚‚A)
angle (deg)
(H‚‚‚A-X)c
N1A-H1A
N1B-H1B
N1C-H1C
N1D-H1D
N28
N31
N35
N38
2.798
2.934
2.825
3.007
2.10
2.27
2.17
2.19
149
142
149
146
23
28
23
24
83 (C9A)
117 (C27)
114 (C29)
79 (C9B
125 (C30)
119 (C32)
)
84 (C9C
122 (C34)
112 (C36)
)
80 (C9D
119 (C25)
124 (C37)
)
a All atoms belong to the asymmetric unit for which fractional atomic coordinates are given in Tables S1 and S3 of the Supporting Information.
b Hydrogen atoms involved in the interaction are also included. c The symbol X is used to designate an atom which is covalently bonded to the
acceptor nitrogen atom A. The atom given in parentheses after the angle value is X.
specifically, disruption of the hydrogen bonds in the N-
phenylacrylamide link to N31, and presumably those to N28 and
N35 as well, would allow the cyclam cap to swing away from
the porphyrin core, thereby producing a substantially larger
pocket inside the gas binding cavity near N31 and the uncom-
plexed oxygen atom of the O2 ligand. Alleviating short
nonbonded interactions between a (nearly) linearly-bonded CO
ligand and the cyclam cap in the modified structure of 7 would,
on the other hand, require a substantial movement of the cyclam
cap not only away from the porphyrin mean plane but also away
from the pseudo-C2 axis. A linearly-bonded CO ligand would
have its bond nearly parallel to the pseudo-C2 axis and
perpendicular to the mean planes of both the porphyrin core
and the cyclam cap; its oxygen would also be pointed directly
Figure 6. Space-filling drawing of the solid-state structure for the
at the cyclam methylene hydrogens which protrude into the gas
Zn(cyclam-capped porphyrin)(OH2) molecule (7) present in its crystal-
binding cavity near the pseudo-C2 axis.
line H2O/CH3OH/CH2Cl2 solvate. The molecule is viewed as in Figure
Reducing the number of methylene groups in either the
N-phenylacrylamide links or the chains connecting amine
nitrogens within the cyclam cap (e.g., 3) would be expected to
reduce the size of a maximally expanded gas binding cavity.
Shorter acrylamide links to the cap would not be able to span
the same distance as longer links when hydrogen bonding is
disrupted. These shorter links would therefore produce a smaller
separation between the cap and the porphyrin core when the
gas binding cavity is fully expanded. Shorter methylene chains
between the four amine nitrogens in the cap would stretch the
N-phenylacrylamide links by pulling them closer to the C2 axis;
if these four links were already fully extended, the cap would
again be pulled closer to the porphyrin core, producing a smaller
fully expanded gas binding cavity. Either type of ligand
modification would be expected to reduce the size of the van
der Waals pocket(s) closest to that modification. This in turn
could reduce the ability of the gas binding cavity to accom-
modate an O2 ligand which is bonded in such a bent fashion to
5bsnearly perpendicular to the pseudo-C2 axis of the molecule which
passes through the Zn atom and oxygen (O1W) of the coordinated water
molecule.
bent O2 which minimizes nonbonded contacts between the
uncomplexed oxygen of the O2 ligand and the pyrrole nitrogens
of the porphyrin core and has been observed previously20b,c in
synthetic models of the O2 binding site in Hb and Mb.
The ability of the gas binding cavity in the cyclam-capped
porphyrin to accommodate a terminally-bonded bent O2 ligand
can also be assessed by making selective alterations to the
structure of 7: (1) place an Fe2+ ion at the center of gravity
(Ct) of the four pyrrole nitrogens; (2) place one oxygen atom
(O1x) of an O2 ligand at a distance20b of 1.898 Å away from the
Fe2+ ion along the pseudo-C2 axis of the molecule on the distal
side of the porphyrin; (3) place a second oxygen atom (O2x)
1.205 Å away from this (first) oxygen along the vector from
the Fe-bonded oxygen atom (O1x) to the cyclam nitrogen N31.
This arrangement produces an Fe-O-O angle of 123°, an
O2x‚‚‚N31 separation of 2.68 Å, and a series of short O‚‚‚H and
O‚‚‚C contacts between the O2 ligand and the cyclam cap:
a nearly in-plane Fe2+
.
The above structural data and arguments are indeed borne
out through experimental results, as seen revisiting Table 1.
Immediately note that the M values are strikingly low. Some
are, in fact, ca. 5 orders of magnitude lower than those reported
for Hb and Mb. This observation seems to truly indicate that
cavity dimensions are important. Specifically, comparison of
compounds 1, 2, and 3 shows a tremendous decrease in CO
affinity. 1 binds carbon monoxide with a P1/2(CO) ) 2.9 Torr
while 2 and 3 show no CO affinity up to 1 atm.
O‚‚‚H26a, 2.19 Å; O‚‚‚H29a, 1.45 Å; O‚‚‚H33a, 1.12 Å; O‚‚‚H33b
,
2.27 Å; O‚‚‚C26, 3.03 Å; O‚‚‚C29, 2.26 Å; O‚‚‚C30, 2.57 Å;
O‚‚‚C32, 2.80 Å; O‚‚‚C33, 1.99 Å; O‚‚‚C34, 3.13 Å; and O‚‚‚C36,
3.24 Å.
Although several of these contacts are considerably shorter
than those which would be expected from summing the
respective van der Waals radii,25 they could be lengthened with
considerably less ligand adjustment than those required for a
linearly-bonded CO. Disrupting the hydrogen bonds in one or
more of the N-phenylacrylamide links could cause the cyclam
cap to be tilted away from the porphyrin mean plane. More
Furthermore, the O2 affinity of 3 is drastically lower than
that of 2. While 1,2-Me2Im-ligated 2 yields a P1/2(O2) ) 22
Torr at 25 °C, 3 under the same conditions yields an immeasur-
ably low O2 affinity. Only after decreasing the temperature to

Aza-Crown-Capped Porphyrin Models of Myoglobin
J. Am. Chem. Soc., Vol. 119, No. 15, 1997 3489
4 °C was a P1/2(O2) ) 760 Torr determined. The only difference
in the two compounds is the cap size, the cap of 3 having two
-CH2- units less than the cap of 2 (n ) 2, 2, 2, 2 vs n ) 2,
3, 2, 3). In following with the trend observed for CO affinities,
1 yields an increased O2 affinity compared to 2, and of course
3. These trends correlate positively with increased cavity
dimensions, indicating a large cavity for 1 relative to 2 and 3.
The avid binding of O2 to 4 and 6 was demonstrated by the
lack of displacement of bound O2 under a continuous purge
with pure Ar. This is dramatically different from the reactivity
of the Cu-free compounds, which showed reversible O2 binding
under the same conditions. The difference in reactivity between
Cu(I) and Cu-free compounds demonstrates that the Cu(I) plays
a crucial role in dioxygen binding.
The possibility of an electrostatic interaction with the lone
pairs of the aza-crown ring does not seem likely since the lone
pairs point away from the cavity toward the amide H and appear
to be involved in hydrogen bonding according to the crystal
structure of 7. Comparison of compounds 1 and 5 with
compounds 4 and 6 is instructive. Carbon monoxide affinity
with and without Cu(I) provides a test as to the extent of the
lone pair effects on the binding. If the lone pairs were playing
a significant role in the inhibition of CO binding, the Cu-
containing compounds would have greater carbon monoxide
affinities, since the lone pairs of the ring are bound to a cationic
metal ion. Instead the presence of the Cu slightly decreases
the model’s affinity for CO. However, one cannot ignore the
potential steric implications of a Cu ion for the cavity. Without
structural evidence, it remains unclear whether or not the Cu
ion is held in a position that could add its own steric clash with
a bound CO. Future X-ray studies of 1 and 4 or the
corresponding Zn analogues will be required to settle this
question.
with greater CO π-back-bonding in the Cu-free complex. Back-
bonding is correlated with the strength of the Fe-C bond;10a
hence, the increased CO stretch in the Cu(I)-containing complex
4 is consistent with a more weakly bound CO molecule and a
lower CO affinity. The similar CO affinity of 4, compared with
1, is indirect evidence that the bulky bromide ligand does not
protrude into the cavity. If this were the case, the CO affinity
of 4 would be much lower.
Conclusion
In conclusion we have synthesized and characterized a class
of iron(II) porphyrins which show unusually low M values.
These models are the first reported to exhibit M values less than
1. The M values are predominantly influenced by CO affinity,
and experimental evidence indicates that steric interactions are
responsible for the effect. This work suggests that steric effects
cannot be ruled out as a major determinant of Hb and Mb CO
affinity.
Acknowledgment. We thank the NIH (Grant 5R37 GM-
17880-26) and the NSF (Grant CHE9123187-A4) for financial
support. P.C.H. thanks the Stanford Chemistry Department for
a Franklin Veatch Fellowship; B.B. thanks NATO and CNRS
for a postdoctoral fellowship. We thank Dr. Charles Campana
of Siemens Analytical X-ray for collection of the CCD X-ray
data for 7. We thank Dr. Miroslav Rapta for helpful discussion.
We also thank the Mass Spectrometry Facility, University of
California, San Francisco, supported by the NIH (Grants RR
04112 and RR 01614).
Supporting Information Available: Crystal structure report,
listings of positional and thermal parameters, listings of bond
lengths and angles, and perspective structural drawings for 7
(26 pages). See any current masthead for ordering and Internet
access instructions.
Infrared spectra were also examined (compounds 1 and 4,
Table 1). The Cu(I) derivative model 4 has a CO stretch 9
cm-1 higher than that of the Cu-free model. This is consistent
JA963945I