Solid-state NMR, crystallographic and density functional theory investigation of Fe-CO and Fe-CO analogue metalloporphyrins and metalloproteins

Article abstract of DOI:10.1021/ja9832818

We have synthesized and characterized the following four metalloporphyrins: Fe(OEP)(CO)(1-MeIm), Ru(OEP)(CO)(1-MeIm), Os(OEP)(CO)(1- MeIm), and Fe(TPP)(iPrNC)(1-MeIm), where OEP = 2,3,7,8,12,13,17,18- octaethylporphyrinate, TPP = 5,10,15,20-tetraphenylporphyrinate, and 1-MeIm = 1-methylimidazole, using single-crystal X-ray diffraction, solid-state nuclear magnetic resonance (NMR), and density functional theory (DFT) methods. Unlike the situation found with the Fe-, Ru-, Os(TPP)(CO)(1-MeIm) analogues, which have ruffled porphyrins, all four systems here have essentially planar porphyrin rings, and a rule is developed that successfully predicts the presence or absence of ring distortion in a broad range of metalloporphyrins. In each of the three CO complexes, the M-C-O bond is close to linear and untilted, but with the iPrNC adduct, there are noticeable ligand distortions supporting the idea that RNC groups (but not CO) may be distorted in metalloproteins. Solid-state ¹³C, ¹⁵N, and ¹⁷NMR shifts and shifts and tensors determined experimentally are in generally good agreement with those computed via DFT. For isocyanide binding to proteins, the experimental shifts are more deshielded than in the model system, and the effects which might contribute to this difference are explored theoretically. Unlike CO, electrostatic field effects are unlikely to make a major contribution to protein shielding. Neither are Fe-C-N tilt-bend distortions, although a bend at nitrogen is energetically feasible and also gives a large deshielding, as seen with proteins.

Full text of DOI:10.1021/ja9832818

3818
J. Am. Chem. Soc. 1999, 121, 3818-3828
Solid-State NMR, Crystallographic and Density Functional Theory
Investigation of Fe-CO and Fe-CO Analogue Metalloporphyrins and
Metalloproteins†
Renzo Salzmann,‡ Michael T. McMahon, Nathalie Godbout, Lori K. Sanders,
Mark Wojdelski,§ and Eric Oldfield*
Contribution from the Departments of Chemistry and Biophysics, UniVersity of Illinois at
Urbana-Champaign, 600 South Mathews AVenue, Urbana, Illinois 61801
ReceiVed September 14, 1998. ReVised Manuscript ReceiVed February 12, 1999
Abstract: We have synthesized and characterized the following four metalloporphyrins: Fe(OEP)(CO)-
(1-MeIm), Ru(OEP)(CO)(1-MeIm), Os(OEP)(CO)(1-MeIm), and Fe(TPP)(iPrNC)(1-MeIm), where OEP )
2,3,7,8,12,13,17,18-octaethylporphyrinate, TPP ) 5,10,15,20-tetraphenylporphyrinate, and 1-MeIm ) 1-meth-
ylimidazole, using single-crystal X-ray diffraction, solid-state nuclear magnetic resonance (NMR), and density
functional theory (DFT) methods. Unlike the situation found with the Fe-, Ru-, Os(TPP)(CO)(1-MeIm)
analogues, which have ruffled porphyrins, all four systems here have essentially planar porphyrin rings, and
a rule is developed that successfully predicts the presence or absence of ring distortion in a broad range of
metalloporphyrins. In each of the three CO complexes, the M-C-O bond is close to linear and untilted, but
with the iPrNC adduct, there are noticeable ligand distortions supporting the idea that RNC groups (but not
CO) may be distorted in metalloproteins. Solid-state 13C, 15N, and 17O NMR shifts and shift tensors determined
experimentally are in generally good agreement with those computed via DFT. For isocyanide binding to
proteins, the experimental shifts are more deshielded than in the model system, and the effects which might
contribute to this difference are explored theoretically. Unlike CO, electrostatic field effects are unlikely to
make a major contribution to protein shielding. Neither are Fe-C-N tilt-bend distortions, although a bend
at nitrogen is energetically feasible and also gives a large deshielding, as seen with proteins.
Introduction
catabolism. A variety of mechanisms have been postulated for
this discrimination, ranging from a protein-induced distortion
of the Fe-C-O bond11 to hydrogen-bonding stabilization of
bound O2 by the distal ligand12-14 and the closely related
hydrogen bonding/water binding model.10,14 In addition, met-
alloporphyrin distortions have been reported in some heme
proteins,15 and can also be expected to influence ligand binding.
However, the precise nature of these stabilizing/destabilizing
effects has been difficult to evaluate at the molecular level, since
the protein structures themselves are part of the debate.16 There
is, therefore, interest in employing spectroscopic methods to
study structure and bonding in metalloproteins, using well-
characterized model systems to help make the structure-
spectroscopic correlations, which should then form the basis
for further structure refinements of proteins themselves, both
in solution and in the crystalline solid-state. As an example of
this approach, we recently reported the synthesis and charac-
terization of a series of carbonmonoxymetalloporphyrins17 which
The preferential binding of O2 over CO by the metalloproteins
hemoglobin and myoglobin and the structural features which
contribute to this discrimination have been topics of debate for
some time.1-10 The interest arises from the fact that CO binds
much less strongly to these proteins than it does to unhindered
metalloporphyrin model compounds, a fortunate circumstance
since CO is produced in vivo as a product of porphyrin
† This work was supported by the United States Public Health Service
(National Institutes of Health grant HL-19481).
‡ Swiss National Science Foundation Postdoctoral Research Fellow,
1996-1997; American Heart Association, Inc., Illinois Affiliate, Post-
doctoral Research Fellow, 1997-1998.
National Institutes of Health Cellular and Molecular Biophysics
Training Grant Trainee (grant GM-08276).
§ Colgate Palmolive Scholar.
(1) Pauling, L.; Coryell, C. D. Proc. Natl. Acad. Sci. U.S.A. 1936, 22,
210-216.
(2) Pauling, L. Nature 1964, 203, 182-183.
(3) Weiss, J. J. Nature 1964, 203, 182-183.
(4) 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, 6769-6770.
Collman, J. P.; Gagne, R. R.; Reed, C. A.; Robinson, W. T.; Rodley, G. A.
Proc. Natl. Acad. Sci. U.S.A. 1974, 71, 1326-1329.
(5) Sage, J. T. J. Biol. Inorg. Chem. 1997, 2, 537-543.
(6) Slebodnick, C.; Ibers, J. A. J. Biol. Inorg. Chem. 1997, 2, 521-525.
(7) Lim, M.; Jackson, T. A.; Anfinrud, P. A. J. Biol. Inorg. Chem. 1997,
2, 531-536.
(8) Spiro, T. G.; Kozlowski, P. M. J. Biol. Inorg. Chem. 1997, 2, 516-
520.
(9) Spiro, T. G.; Kozlowski, P. M. J. Am. Chem. Soc. 1998, 120, 4524-
4525.
(10) Olson, J. S.; Phillips, G. N., Jr. J. Biol. Inorg. Chem. 1997, 2, 544-
552.
(11) Kuriyan, J.; Wilz, S.; Karplus, M.; Petsko, G. A. J. Mol. Biol. 1986,
192, 133-154. Stryer, L. Biochemistry; W. H. Freeman and Co.: New York,
1995.
(12) Phillips, S. E. V.; Schoenborn, B. P. Nature 1981, 292, 81-82.
(13) Nagai, K.; Luisi, B.; Shih, D.; Miyazaki, G.; Imai, K.; Poyart, C.;
DeYoung, A.; Kwiatkowsky, L.; Noble, R. W.; Lin, S.-H.; Yu, N.-T. Nature
1987, 329, 858-860.
(14) Springer, B. A.; Sligar, S. G.; Olson, J. S.; Phillips, G. N., Jr. Chem.
ReV. 1994, 94, 699-714.
(15) Shaanan, B. J. Mol. Biol. 1983, 171, 31-59.
(16) Ray, G. B.; Li, X.-Y.; Ibers, J. A.; Sessler, J. L.; Spiro, T. G. J.
Am. Chem. Soc. 1994, 116, 162-176.
(17) Salzmann, R.; Ziegler, C. J.; Godbout, N.; McMahon, M.; Suslick,
K. S.; Oldfield, E. J. Am. Chem. Soc. 1998, 120, 11323-11334.
10.1021/ja9832818 CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/08/1999

CO-Heme Analogue Structures
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3819
Experimental Section
proved to be useful in analyzing nuclear magnetic resonance
(NMR), Mo¨ssbauer, and infrared (IR) spectroscopic data on CO-
heme proteins.18-20 Interestingly, in that work17 we noted that
there were pronounced rufflings for each of the three metal-
loporphyrins containing 1-methylimidazole as an axial base,
while all three species containing axial pyridine ligands were
essentially planar.17
Synthetic Aspects. The synthesis of Fe(OEP)(CO)(1-MeIm), Ru-
(OEP)(CO)(1-MeIm), and Os(OEP)(CO)(1-MeIm) followed the basic
protocols outlined previously for the synthesis of Fe(TPP)(CO)-
(1-MeIm), Ru(TPP)(CO)(1-MeIm), and Os(TPP)(CO)(1-MeIm),17 and
will not be further elaborated on, except that the Fe(OEP)(CO)(1-MeIm)
crystals were grown from hexane/benzene while the Ru, Os derivatives
were from methylene chloride/methanol. Analytical data for Fe(OEP)-
(CO)(1-MeIm)•0.5hexane: Elemental Anal. Found (Calculated): C,
70.89 (71.23); H, 7.91 (7.75); N, 10.88 (11.33). IR (νCO in CH2Cl2):
In this work, we have extended the earlier study17 to
encompass different porphyrin ring substituents, as well as a
different (but isoelectronic) axial ligand, isopropylisocyanide
(iPrNC), to see experimentally what effects changes in ring and
axial substitutions have on structure. We have synthesized and
characterized four new compounds: the octaethylporphyrin
(2,3,7,8,12,13,17,18-octaethylporphyrinate ) OEP) adducts Fe-
(OEP)(CO)(1-MeIm), Ru(OEP)(CO)(1-MeIm), and Os(OEP)-
(CO)(1-MeIm), which unlike the TPP analogues we find to be
planar rather than ruffled, and the isocyanide adduct Fe(TPP)-
(iPrNC)(1-MeIm) (TPP ) 5,10,15,20-tetraphenylporphyrinate),
which unlike the CO TPP derivative is also found to contain a
planar porphyrinsalthough the RNC group is noticeably
distorted, as are several RNC-protein adducts.21,22 These new
compounds, together with the O2-analogue species described
in the following article,23 form an interesting series of com-
pounds with which to study how porphyrin ring substitutions
and axial ligands can influence porphyrin distortions, a topic
more typically limited to metal and porphyrin ring substitu-
tions,24 but which may also be of importance in metalloprotein
function. An empirical rule is developed which enables the
correct prediction of the presence or absence of porphyrin
ruffling for 16 out of 16 systems, containing TPP, OEP, CO,
CCl2, RNC, RNO, py, or 1-MeIm ligands. We also report and
investigate via density functional theory the solid-state NMR
of these four new compounds, together with an analysis of the
potential energy surfaces for Fe-RNC distortion. This provides
information on the extent to which the alkylisocyanides can be
distorted in proteins, analogous to previous work on CO-
containing metalloporphyrin model systems,18-20 and comple-
mentary to the work reported on heme protein isocyanide
systems by Mims et al.,25 but using a quantum chemical
approach.
1965 cm-1
Cl2: Elemental Anal. Found (Calculated): C, 60.73 (60.86); H, 6.31
(6.33); N, 9.97 (10.14). IR (νCO in CH2Cl2): 1924 cm-1 13C NMR
.
13C NMR (CO): 206.9 ppm. Ru(OEP)(CO)(1-MeIm)•CH2-
.
(CO): 182.8 ppm. Os(OEP)(CO)(1-MeIm)•CH2Cl2: Elemental Anal.
Found (Calculated): C, 54.99 (54.95); H, 5.67 (5.71); N, 8.97 (9.15).
IR (νCO in CH2Cl2): 1893 cm-1 13C NMR (CO): 141.2 ppm. For solid-
.
state NMR experiments, we used 13CO, C17O, [2-N13C] propane and
[2-15NC] propane axial ligands, the labeled isopropylisocyanides being
synthesized from 2-iodopropane and labeled AgCN basically as
described elsewhere.26
(2-Isocyanopropane)(1-methylimidazole)(5,10,15,20-tetra-
phenylporphyrinato)iron(II) and (2-Isocyanopropane)(pyridine)-
(5,10,15,20-tetraphenylporphyrinato)iron(II). The syntheses of both
of these isopropylisocyanide Fe TPP adducts were carried out with
use of the following procedure: Five milliliters of THF was degassed
with Ar for 10 min in a 15 mL Schlenk flask equipped with a magnetic
stirring bar, followed by 4 freeze-pump-thaw cycles. Fifty milligrams
of (octaethylporphyrin)FeIIICl was then added under Ar counterflow,
followed by ca. 2 equiv of NaBH4. The Schlenk was then closed and
stirred for 45 min on an open Ar line. 2.2 equivalents of pyridine or
1-MeIm were than added under Ar and the system closed and stirred
for 1 h. The bis(1-MeIm) complex gives a metallic purple precipitate
while the bis-py complex is orange and soluble. THF was then
completely removed under high vacuum. Three milliliters CHCl3
(previously degassed) was then added and the solutions transferred via
cannula to a 100 mL Schlenk flask, to which was added a 10-fold excess
of 2-isocyanopropane. The solutions were stirred for 20 h, then a 10-
fold excess of pentane (degassed) was added dropwise to form a 2-layer
system, from which the corresponding isocyanide adducts crystallized
after several days. Fe(TPP)(iPrNC)(1-MeIm)•0.5pentane: Elemental
Anal. Found (Calculated): C, 76.91 (76.48); H, 5.41 (5.54); N, 11.30
(11.45).
All compounds were fully characterized by field desorption mass
spectrometry, UV-visible absorption, and solution and solid-state NMR
spectroscopy. Elemental analyses were conducted in the University of
Illinois School of Chemical Sciences Microanalytical Laboratory. Field
desorption mass spectrometry measurements were carried out by using
a Finnigan-MAT (Bremen, Germany) Model 731 instrument. Porphyrin
UV-visible spectra were measured with use of a Hitachi Ltd. (Tokyo,
Japan) Model 3300 UV-visible double monochromator spectropho-
tometer. The single-crystal X-ray measurements were made on a
Siemens (Madison, WI) SMART diffractometer. Solid-state NMR
spectra were obtained on “home-built” 360 and 500 MHz spectrometers,
using Oxford magnets (Oxford Instruments, Oxford, UK), Tecmag
(Houston, TX) data systems, and Doty Scientific (Columbia, SC)
probes.
(18) Godbout, N.; Havlin, R.; Salzmann, R.; Debrunner, P. G.; Oldfield,
E. J. Phys. Chem. A 1998, 102, 2342-2350.
(19) Havlin, R. H.; Godbout, N.; Salzmann, R.; Wojdelski, M.; Arnold,
W.; Schulz, C. E.; Oldfield, E. J. Am. Chem. Soc. 1998, 120, 3144-3151.
(20) McMahon, M. T.; deDios, A. C.; Godbout, N.; Salzmann, R.; Laws,
D. D.; Le, H.; Havlin, R. H.; Oldfield, E. J. Am. Chem. Soc. 1998, 120,
4784-4797.
(21) Johnson, K. A.; Olson, J. S.; Phillips, G. N., Jr. J. Mol. Biol. 1989,
207, 459-463. Johnson, K. A. Thesis, Rice University, 1993. Eich, R. F.;
Li, T.; Lemon, D. D.; Doherty, D. H.; Curry, S. R.; Aitken, J. F.; Mathews,
A. J.; Johnson, K. A.; Smith, R. D.; Phillips, G. N., Jr.; Olson, J. S.
Biochemistry 1996, 35, 6976-6983.
(22) Johnson, K. A.; Olson, J. S.; Smith, R. D.; Phillips, G. N., Jr. Protein
Data Bank in Crystallographic DatabasessInformation Content, Software
Systems, Scientific Applications; Data Commission of the International Union
of Crystallography; file nos. 101m, 103m, 104m, 105m, 106m, 107m, 108m,
109m, 110m, 111m, 112m, 2hbc, 2hd, 2hbe, 2hbf, 2mya, 2myb, 2myc,
2myd, and 2mye.
Crystallographic Aspects. Single-crystal data for the three systems
described above were collected on a Bruker (Madison, WI) SMART/
CCD diffractometer using Mo KR radiation (λ ) 0.71073 Å). The
structures were solved by using the SHELXTL V5.0 (Bruker) system
and refined by full-matrix least squares on F2. Hydrogen atoms were
assigned idealized locations and given isotropic thermal parameters 1.2
times the thermal parameter of the atom to which they were attached.
The data were corrected for Lorentz and polarization effects, and an
empirical absorption correction was applied.
(23) Godbout, N.; Sanders, L. K.; Salzmann, R.; Havlin, R. H.; Wodjelski,
M.; Oldfield, E. J. Am. Chem. Soc. 1999, 121, 3829.
(24) Sparks, L. D.; Medforth, C. J.; Park, M.-S.; Chamberlain, J. R.;
Ondrias, M. R.; Senge, M. O.; Smith, K. M.; Shelnutt, J. A. J. Am. Chem.
Soc. 1993, 115, 581-592. Munro, O. Q.; Bradley, J. C.; Hancock, R. D.;
Marques, H. M.; Marsicano, F.; Wade, P. W. J. Am. Chem. Soc. 1992,
114, 7218-7230. Scheidt, W. R.; Lee, Y. J. Struct. Bonding (Berlin), 1987,
64, 1-70.
(26) Jackson, H. L.; McKusick, B. C. In Organic Synthesis; Cairns, T.
L., Cason, J., Johnson, W. S., Leonard, N. J., Newman, M. S., Price, C. C.,
Sheehan, J. C., Eds.; John Wiley & Sons: New York, 1955; Vol. 35, pp
62-64.
(25) Mims, M. P.; Olson, J. S.; Russu, I. M.; Miura, S.; Cede, T. E.;
Ho, C. J. Biol. Chem. 1983, 258, 6125-6134.

3820 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Salzmann et al.
Table 1. Crystallographic Data Summary
Fe(OEP)(CO)(1-MeIm)
Ru(OEP)(CO)(1-MeIm)
Os(OEP)(CO)(1-MeIm)
Fe(TPP)(iPrNC)(1-MeIm)
formula
C41H50N6OFe•
hexane
741.81
C41H50N6ORu•
methylene chloride
828.87
C41H50N6OOs•
methylene chloride
918.00
C53H41N7Fe•
0.5pentane
855.84
formula weight
color
red
red
dark red
black
crystal system
a (Å)
b (Å)
triclinic
11.872(2)
13.192(3)
14.305(3)
84.19(3)
69.72(3)
69.85(3)
1972.5(7)
2
triclinic
triclinic
triclinic
10.4568(2)
12.98060(10)
15.9683(2)
80.6950(10)
76.9820(10)
72.0180(10)
1998.65(5)
2
10.3677(3)
12.9871(4)
15.9564(5)
80.6350(10)
76.7310(10)
72.2190(10)
1981.25(10)
2
10.9981(7)
13.2076(8)
17.1650(11)
73.86
78.0500(10)
79.5640(10)
2322.9(3)
2
c (Å)
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
D
calc (g cm-3
)
1.249
1.377
1.539
1.224
space group
radiation,
P1h
P1h
P1h
P1h
Mo KRj, 0.71073 Å
Mo KRj, 0.71073 Å
Mo KRj, 0.71073 Å
Mo KRj, 0.71073 Å
wavelength (Å)
µ (mm-1
)
0.424
0.567
0.339
0.369
crystal size (mm)
temp (K)
0.23 × 0.16 × 0.08
0.12 × 0.12 × 0.10
0.11 × 0.09 × 0.01
0.26 × 0.17 × 0.08
198(2)
198(2)
198(2)
198(2)
diffractometer
no. of data points
collected
Bruker SMART/CCD
10326
Bruker SMART/CCD
13046
Bruker SMART/CCD
10705
Bruker SMART/CCD
11376
no. of data points
with I > 2σ(I)
abs min/max
R1a (obsd data)
wR2(a,b)b
6785
9023
6919
7147
0.847/0.930
0.059
0.128 (0.061, 0.924)
1.026
0.673/0.569
0.059
0.122 (0.037, 4.960)
1.142
0.656/0.959
0.065
0.123 (0.025, 17.037)
1.147
0.825/1.00
0.0663
0.160 (0.072, 3.968)
1.078
GOFc
a R1 ) ∑(||Fo| - |Fc||)/∑(|Fo|). b wR2 ) [∑[w(Fo2 - Fc2)2]/∑[w(Fo )2]]1/2 where w ) 1/[σ2(Fo ) + (a/P)2 + b/P] and P ) (Fo2 + 2Fc2)/3. c GOF
2
2
) S ) [∑[w(Fo - Fc2)2]/(n - p)] where n ) the number of reflections and p ) the total number of parameters refined.
2
The color and morphology of the crystals, the crystallographic
systems, and space groups, and other information related to the crystal
structure determinations, are summarized in Table 1. In general, from
10 326 to 13 046 data points were collected with the area detector,
and from 6785 to 9023 data points having I > 2σ(I) were used in the
refinements. The final R1 values varied from 0.059 to 0.066 and the
GOF values varied from 1.026 to 1.147, Table 1. (See also Supporting
Information.) Atomic coordinates, bond lengths, angles, and thermal
parameters have also been deposited with the Cambridge Crystal-
lographic Data Centre (CCDC). Any request to the CCDC for this
material should quote the full literature citation and the reference
number.
Computational Section. The computational methods used here are
very similar to the density functional theory (DFT) methods we have
used elsewhere,18-20 where numerous studies of the effects of functionals
and basis sets on metal and ligand shieldings, and on metal electric
field gradients (EFGs), were described. For investigating electrostatic
field effects on 13C, 15N shielding, we used the [FeCNMe](2e) species
shown in Figure 1A, which is basically that used by deDios and Earle27
in their study of electrostatic field perturbations of the [FeCO](2e)
species, where the minus signs (Figure 1A) represent 0.4e point charges,
incorporated to maintain a low-spin d6 Fe ground state.27 For calcula-
tions of the effects of ligand tilt and bend on the computed shieldings,
we used the Fe(bis(amidinato))(MeNC)(1-methylimidazole) model,
which is based on the CO derivative used previously in our study of
Fe-C-O distortion, and is essentially that used by several other
groups28-30 to deduce, e.g., tilt-bend energy surfaces for both Fe-
C-O and FeIII-C-N. In both sets of calculations, we used the sum-
over-states density functional perturbation theory in its LOC1 approxi-
Figure 1. Schematic illustrations of the various structures used in the
DFT calculations. (A) [FeCNMe](2e)(+) cluster used to evaluate the
electrostatic contributions to shielding, as discussed by de Dios and
Earle (ref 27). (B) Fe bis(amidinato)(CNMe)(Im) molecule used to
evaluate 13C, 15N axial ligand shifts and energetics as a function of tilt
(τ), C-bend (â), and N-bend (θ). (C) Basic M(P)(CO)(1-MeIm) structure
used for 13C, 17O shielding calculations, using the bond lengths and
bond angles derived crystallographically. (D) Fe(P)(iPrNC)(1-MeIm)
structure used for 13C, 15N shielding calculations, using the bond lengths
and bond angles derived crystallographically for Fe(TPP)(iPrNC)(1-
MeIm).
mation31 using individual gauges for localized orbitals32 as implemented
in the deMon program.33 We used Wachters’ all electron basis set for
(27) deDios, A. C.; Earle, E. M. J. Phys. Chem. A 1997, 101, 8132-
8134.
(28) Strich, A.; Veillard, A. Theor. Chim. Acta 1981, 60, 379-383.
(29) Jewsbury, P.; Yamamoto, S.; Minato, T.; Saito, M.; Kitagawa T. J.
Am. Chem. Soc. 1994, 116, 11586-11587. Jewsbury, P.; Yamamoto, S.;
Minato, T.; Saito, M.; Kitagawa, T. J. Phys. Chem. 1995, 99, 12677-12685.
(30) Vangberg, T.; Bocian, D. F.; Ghosh, A. J. Biol. Inorg. Chem. 1997,
2, 526-530. Ghosh, A.; Bocian, D. F. J. Phys. Chem. 1996, 100, 6363-
6367.
(31) Malkin, V. G.; Malkina, O. L.; Casida, M. E.; Salahub, D. R. J.
Am. Chem. Soc. 1994, 116, 5898-5908. Salahub, D. R. In Theoretical and
Computational Chemistry; Politzer, P., Seminario, J. M., Eds.; Elsevier:
New York, 1995.
(32) Kutzelnigg, W.; Fleischer, U.; Schindler M. In NMR-Basic
Principles and Progress; Springer: Heidelberg, 1990; Vol. 28, p 1965.

CO-Heme Analogue Structures
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3821
Figure 2. SHELXTL (Bruker, 1998) X-ray structures of the four compounds investigated. (A) Fe(2,3,7,8,12,13,17,18-octaethylporphyrinate)(CO)-
(1-methylimidazole); (B) Ru(2,3,7,8,12,13,17,18-octaethylporphyrinate)(CO)(1-methylimidazole); (C) Os(2,3,7,8,12,13,17,18-octaethylporphyrinate)-
(CO)(1-methylimidazole); and (D) Fe(5,10,15,20-tetraphenylporphyrinate)(2-isocyanopropane)(1-methylimidazole). The representations show 35%
probability ellipsoids for non-H atoms. H atoms were omitted for clarity. Ethyl group disorder is present in parts B and C.
iron (14s9p5d/8s5p5d),34,35 an IGLO-II basis for H,C,N, together with
the Perdew-Wang (PW91) gradient-corrected exchange-correlation
functional.36 The electrostatic field perturbations were generated by the
presence of point charges of varying magnitudes at 4.45 Å from the
isocyanide N, along the Fe-C-N axis. We also performed a second
set of calculations using the FeCO(2e) model and a point charge of
varying magnitude at 4.45 Å from the carbonmonoxy O, along the Fe-
C-O axis, for direct comparison with the isocyanide results.
metalloporphyrins, we employed the Gaussian-94/DFT/GIAO program37
as described previously.18-20 We used our experimental X-ray structures
but minus the porphyrin ring substituents (i.e., Ph, Et f H, refs 18-
20), which does not affect ligand shielding, together with a Wachters’
basis on Fe, 6-311++G(2d) on the attached porphyrin nitrogens, the
metal coordinated nitrogen of the axial base and the isocyanide C and
N, 6-31G* on all porphyrin carbons attached to nitrogen, and 3-21G*
elsewhere. The same scheme was used for the carbon monoxide
complexes. For the Ru and Os compounds, we employed the same
scheme as that described previously,17 in which the metals are
represented with effective or model core potentials (ECPs).37 This
locally dense38 approach has been used and tested previously18-20 and
gives good accord with experiment. The BPW91 exchange correlation
For the 13C, 15N, and 17O shielding calculations on the three
(33) Salahub, D. R.; Fournier, R.; Mlynarski, P.; Papai, I.; St.-Amant,
A.; Ushio, J. In Density Functional Methods in Chemistry; Labanowski, J.,
Andzelm, J., Eds.; Springer: New York, 1991. St.-Amant, A.; Salahub, D.
R. Chem. Phys. Lett. 1990, 169, 387-392.
(34) Wachters, A. J. H. J. Chem. Phys. 1970, 52, 1033-1036. Wachters,
A. J. H. IBM Technol. Rep. RJ584 1969.
(35) Basis sets were obtained from the Extensible Computational
Chemistry Environment Basis Set Database, Version 1.0, as developed and
distributed by the Molecular Science Computing Facility, Environmental
and Molecular Sciences Laboratory, which is part of the Pacific Northwest
National Laboratory, P.O. Box 999, Richland, WA 99352, and is funded
by the U.S. Department of Energy under contract DE-AC06-76RLO 1830.
Contact David Feller, Karen Schuchardt, or Don Jones for further informa-
tion.
(37) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Oritz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian94, Revision C.2; Gaussian,
Inc.: Pittsburgh, PA, 1995.
(38) Chesnut, D. B.; Moore, K. D. J. Comput. Chem. 1989, 10, 648-
659.
(36) Perdew, J. P.; Wang, Y. Phys. ReV. B 1992, 45, 13244-13249.

3822 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Salzmann et al.
Table 2. Structural Summary for OEP/TPP Porphyrinsa
Fe(OEP)(CO)(1-MeIm)
Ru(OEP)(CO)(1-MeIm)
Os(OEP)(CO)(1-MeIm)
Fe(TPP)(iPrNC)(1-MeIm)
M-N(1)
1.998(3)
1.998(3)
2.005(3)
2.000(3)
1.744(5)
2.077(3)
1.158(5)
90.03(14)
178.11(13)
89.86(14)
89.79(14)
178.31(13)
90.26(13)
89.8(2)
2.054(3)
2.060(3)
2.063(3)
2.060(3)
1.829(5)
2.192(4)
1.156(5)
90.17(12)
175.66(14)
89.75(13)
89.64(13)
175.6(2)
90.11(13)
91.4(2)
2.063(8)
2.058(8)
2.057(8)
2.061(8)
1.817(13)
2.177(9)
1.171(13)
89.9(3)
174.1(4)
90.2(3)
89.4(3)
173.9(4)
89.9(3)
94.2(4)
2.004(4)
1.994(4)
1.996(4)
2.000(4)
1.847(5)
2.041(4)
1.155(6)
89.89(14)
178.6(2)
89.9(2)
89.9(2)
178.5(2)
90.3(2)
89.9(2)
96.1(2)
M-N(2)
M-N(3)
M-N(4)
M-C
M-N(base)
C-O/C-N
N(1)-M-N(2)
N(1)-M-N(3)
N(1)-M-N(4)
N(2)-M-N(3)
N(2)-M-N(4)
N(3)-M-N(4)
C-M-N(1)
C-M-N(2)
C-M-N(3)
C-M-N(4)
C-M-N(base)
M-C-O/M-C-N
88.3(2)
92.0(2)
93.4(2)
176.8(2)
175.1(4)
90.9(2)
92.9(2)
93.5(2)
178.3(2)
177.3(4)
94.2(4)
91.7(4)
91.9(4)
177.9(4)
177.0(9)
91.4(2)
85.3(2)
174.1(2)
170.1(4)
angle between N(base) and
porphyrin N-N bond
vector
1°
4°
4°
24°
angle between N(base) ring
and the porphyrin plane
91°
87°
88°
85°
a M: metal center Fe, Os, or Ru as appropriate.
functional was used in each case.37 Calculations were carried out in
this laboratory on International Business Machines (Austin, TX) RS/
6000 computers (models 340, 350, 360, 365, and 3CT) and on an
8-processor Silicon Graphics/Cray Research (Mountain View, CA)
Origin-200 cluster, and on SGI Origin-2000 and Power Challenge
multiple processor machines located at the National Center for
Supercomputing Applications, located in Urbana, IL, using up to 16
processors.
Results and Discussion
We have synthesized and elucidated the structures of four
new porphyrin compounds: the Fe, Ru, and Os complexes of
OEP with 1-methylimidazole and CO as axial ligands, together
with the Fe(iPrNC)(1-MeIm) complex of TPP, which enable
us to begin to investigate in a systematic way how metallopor-
phyrins containing different ring and axial ligand substitutions
may influence porphyrin distortions, as well as serving as useful
model compounds with which to probe metalloprotein structure.
The SHELXTL structures of these four molecules are shown
in Figure 2, and selected metric details are presented in Table
2. The availability of these new structures now enables a
comparison to be made between the structures of the OEP and
TPP CO systems and between the CO and RNC TPP systems,
and these results, together with those for the OEP and TPP O2-
analogue systems to be described elsewhere,23 lead to new
insights into the likely occurrence of porphyrin (and ligand)
distortions in metalloporphyrins.
As may be seen from Figures 3 and 4, there are remarkable
differences, as well as remarkable similarities, between the TPP
and OEP systems. The largest differences occur in the extent
of porphyrin ruffling. In the cases of Fe(TPP)(CO)(1-MeIm),
Ru(TPP)(CO)(1-MeIm), and Os(TPP)(CO)(1-MeIm), there are
pronounced saddle distortions, with mean absolute deviations
Figure 3. Schematic illustrations showing atom deviations (in units
of 10-2Å) from the least-squares porphyrin plane in several CO
metalloporphyrins: (A, C, E) the Fe-, Ru-, Os(OEP)(CO)(1-MeIm)
systems described in this paper; (B, D, F) the Fe-, Ru-, Os(TPP)(CO)-
(1-MeIm) systems described elsewhere (refs 17 and 50). The TPP
derivatives are much more distorted than the OEP derivatives.
of Câ from the least-squares porphyrin plane of |Câ| ) 0.286 Å
(Fe TPP), |Câ| ) 0.196 Å (Ru TPP), and |Câ| ) 0.183 Å (Os
TPP), to be compared with |Câ| ) 0.023 Å (Fe OEP), |Câ| )
is almost no distortion, with |Câ| ) 0.014 Å, Figure 4.
Differences are also seen in the orientation of the 1-MeIm with
respect to the nearest N-N bond vector. The plane of the
0.031 Å (Ru OEP), and |Câ| ) 0.037 Å (Os OEP) in the OEP
case. In contrast, in the case of Fe(TPP)(iPrNC)(1-MeIm) there

CO-Heme Analogue Structures
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3823
Table 3. Comparison between Experimental Porphyrin Câ
Distortions and Prediction
mean Câ
(expt, Å)a predictionb
distortion
type
system
TPP/CO
Fe(TPP)(CO)(1-MeIm)
Fe(TPP)(CO)(py)
Ru(TPP)(CO)(1-MeIm)
Ru(TPP)(CO)(py)
Os(TPP)(CO)(1-MeIm)
Os(TPP)(CO)(py)
Fe(OEP)(CO)(1-MeIm)
Ru(OEP)(CO)(1-MeIm)
Os(OEP)(CO)(1-MeIm)
Fe(TPP)(iPrNC)(1-MeIm)
Fe(TPP)(PhNO)(1-MeIm)
FE(TPP)(PhNO)(py)
Fe(TPP)(NODMA)(py)
Fe(OEP)(PhNO)(1-MeIm)
0.286
0.049
0.196
0.04
+
-
+
-
+
-
-
-
-
-
-
+
+
-
-
+
0.183
0.055
0.023
0.031
0.037
0.014
0.032
0.288
0.156
0.038
0.047
0.235
OEP/CO
RNC
RNO
5-coordinate Co(OEP)(NO)
Fe(TPP)(CCl2)
Figure 4. Schematic illustration showing atom deviations (in units of
10-2Å) from the least-squares plane in Fe(TPP)(iPrNC)(1-MeIm).
a The mean Câ values were for the 14 structures determined in this
laboratory, together with the structures of Fe(TPP)(CO)(py) and
Ru(TPP)(CO)(py) reported in refs 49 and 50. b Predictions were made
by using the logic-based model described in the text: f(A,B,C) )
A•BxC, where A represents the presence of a ring phenyl group, B
indicates the presence of a large or “distorting” axial subsituent
(including 1-MeIm), and C is a small or “nondistorting” ligand. Note
the presence of two distorting axial ligands results in no net distortion.
Unlike the CO derivative, the iPrNC system is essentially planar (|Câ|
) 0.014 Å).
imidazole is oriented at ∼3° to the N-N bond vector in the
OEP derivatives, Table 2 and (schematically) Figure 3, to be
compared with a ∼30° orientation in the TPP CO derivatives,
Figure 3, and a ∼24° orientation in Fe(TPP)(iCN)(1-MeIm),
Figure 4. These values are similar to those seen in the OEP/
TPP-RNO adducts23 where the imidazole plane is oriented at
∼8° in Fe(OEP)(PhNO)(1-MeIm) but at ∼27° in Fe(TPP)-
(PhNO)(1-MeIm). The M-C-O bonds are generally close to
linear, although there is a ∼5° distortion in Fe(OEP)(CO)(1-
MeIm), Table 2.
Using a cutoff of |Câ| g 0.06 Å to indicate the presence of ruffling,
there is a 1:1 correlation between experiment and prediction.
(py), Fe(TPP)(NODMA)(py), and Fe(TPP)(CCl2), while the
three unruffled systems are Fe(TPP)(PhNO)(1-MeIm), Fe(OEP)-
(PhNO)(1-MeIm), and Co(OEP)(NO).23
While these results may at first all appear unrelated, we can
in fact now begin to develop a predictive model for ruffling,
based on the following observations: First, of the 16 compounds
considered here and elsewhere, ruffling is only seen in the TPP
series, although it is not always present, Table 3. There are 6
out of 16 compounds containing large distortions: the three
1-MeIm/CO adducts, the two py/RNO adducts, and Fe(TPP)-
(CCl2). The two 1-MeIm/RNO adducts, unlike the MeIm/CO
species, are undistorted, as is the other five-coordinate species,
Co(OEP)(NO). These results suggest the following rule: that
in order for there to be a porphyrin distortion there needs to be
one and only one repulsive interaction between a porphyrin ring
substituent and an axial ligand. This can be expressed symboli-
cally by the logical operation: f(A,B,C) ) A•B x C where A,
B, C represent the presence or absence of a particular group,
with A representing the ring substitution. The dot (•) represents
a logical AND statement and the x an exclusive OR statement.
Sixteen out of sixteen porphyrin distortions considered can be
predicted by dividing the axial/equatorial substituents into two
classes: 1 ) PhNO, NODMA, CCl2, 1-MeIm, ring-phenyl, and
0 ) CO, NO, pyridine, ring-ethyl. For example, for A ) Ph )
1, B ) CO ) 0, and C ) py ) 0, the operation A•B x C
yields f(A,B,C) ) 0, or no ring distortion. For A ) Ph ) 1, B
) CO, and C ) 1-MeIm, f(A,B,C) ) 1, ruffling is predicted.
For A ) Ph ) 1, B ) PhNO ) 1, and C ) py ) 0, f(A,B,C)
) 1, ruffling is predicted. For A ) Ph, B ) PhNO, and C )
1-MeIm, f(A,B,C) ) 0, no distortion, and so forth, as shown in
Table 3. To what extent these effects are purely inter-/
intramolecular steric or electronic is unknown, and will have
to be determined in the future by using quantum chemical
geometry optimization methods, possibly on both single mol-
ecules and with periodic boundary conditions, on crystal
lattices.40 But qualitatively, the picture that emerges from the
above analysis is very simple and is that ring distortions due to
The major structural differences observed between the various
systems therefore involve porphyrin ruffling and are quite
pronounced, and puzzling. If we invoke an electronic origin
due to porphyrin substitution, then the TPP CO species might
be distorted while the OEP derivatives are not, but the
isoelectronic TPP RNC derivative is quite undistorted, Figure
4. Also, we noted previously in the corresponding pyridine
adducts, Fe-, Ru-, Os(TPP)(CO)(py), that each system is
undistorted.17 It therefore appears to be necessary to consider
the axial base (py TPP CO adducts are planar, 1-MeIm TPP
CO adducts are distorted), the axial ligand (TPP CO 1-MeIm
is distorted, TPP iPrNC 1-MeIm is planar), as well as the
porphyrin substitution pattern, to arrive at a model that can
predict, at least qualitatively, the observed trends. While
quantum chemical geometry optimizations should in the future
permit a more detailed analysis of these observations, for now
we focus on the development of a plausible empirical model
that predicts the observed results with good accuracy.
An Analysis of Porphyrin Ruffling. As noted above, the
four new compounds we have synthesized have planar metal-
loporphyrins, even though the RNC derivative has a relatively
bulky isopropylisocyanide ligand, and phenyl substitutions. The
origins of the porphyrin ring distortions are thus not immediately
obvious. However, in other work,23 we have investigated the
structures of five additional metalloporphyrins, four of which
are O2-heme analogues, and we have also obtained a high-
resolution structure of an additional, distorted carbene system.39
Three of these six compounds are ruffled and three are not.
The three distorted systems are (using the shorthand notation
given above and noting that PhNO ≡ nitrosobenzene and
NODMA ≡ 4-nitroso-N,N-dimethylaniline) Fe(TPP)(PhNO)-
(39) Salzmann, R.; Ziegler, C.; Suslick, K.; Oldfield, E., unpublished
results.

3824 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Salzmann et al.
Figure 5. Typical CP-MAS NMR spectra: (A) 8.45 T 13C CP-MAS
NMR spectrum of Os(OEP)(13CO)(1-MeIm), 5.0 kHz spin speed, 4720
scans at a 3 s recycle time and (B) 15N MAS NMR spectrum of Fe-
(TPP)(iPr15NC)(1-MeIm), 2.5 kHz spin speed, 7040 scans at a 2 s
recycle time. The asterisks in part A arise from the OEP porphyrin
carbons.
axial ligation differences are only seen when the ligation pattern
involves one large or distorting ligand, interacting with a large
ring substituent. When two such axial ligands are present, the
distortions cancel, as indicated logically with the A•B x C
operation. The effect certainly appears steric, as evidenced by
the increased size of the A-type ligands listed above, with
1-MeIm being in the A-class due to its somewhat angular
substitution pattern. On the basis of the results shown in Table
3, the predictions are quite robust: for all six distortion
predictions (“+” in Table 3) the |Câ| value is on average 0.22
Å, to be compared with |Câ| ) 0.04 Å for the 10 nondistorted
predictions (“-” in Table 3).
Solid-State NMR and Quantum Chemistry. We next
carried out a 13C, 15N, and 17O nuclear magnetic resonance
spectroscopic investigation of 13CO, C17O, iPrN13C, and iPr-
15NC labeled metalloporphyrins, using “magic-angle” sample
spinning (MAS). We also used density functional theory to
predict these same parameters, which provides a test of our
ability to predict spectroscopic observables in relatively well
characterized materials. Testing the theoretical methods on heme
model compounds is of importance since it can give enhanced
confidence in the reliability of the calculations themselves, or
more precisely, in the values of properties such as the charge
density, (F(r)), the electrostatic potential (Φ(r)), and the electric
field gradient (∇•E), which can also be independently deduced
Figure 6. Graph showing experimental versus theoretical shifts and
shift tensor elements for the compounds investigated. (A) Graph
showing correlations for carbon-13 tensor elements (slope ) 1.18, R2
) 0.99): (O) TPP metalloporphyrin complexes from ref 17; (×) OEP
metalloporphyrin complexes (data from Table 4); (b), Fe(TPP)(iPrNC)-
(1-MeIm) (data from Table 4). (B) Graph showing correlations for
nitrogen-15 tensor elements (slope ) 1.01, R2 ) 1.0) for Fe(TPP)-
(iPrNC)(1-MeIm). (C) Graph showing correlations for oxygen-17
isotropic shifts (slope ) 1.32, R2 ) 0.99): (O) TPP complexes from
ref 17; (×) OEP complexes (Table 4). The line is drawn through all of
the data points.
2
in some cases from diffraction data,41-43 as well as ∇ F(r), of
interest in the context of chemical and hydrogen bonding44sa
long-term objective of this research.
We show in Figure 5 representative 13C and 15N CP-MAS
NMR spectra, in this case of Os(OEP)(13CO)(1-MeIm) and Fe-
(TPP)(iPr15NC)(1-MeIm). Additional spectra of these and the
other compounds were also obtained at different spinning speeds,
and the principal components of the chemical shift tensor, δii,
were derived by using the Bayesian probability/Herzfeld-Berger
method45 described elsewhere.46 A compilation of the experi-
mental results is given in Table 4. For the 17O spectra of the
(40) Marchi, M.; Hutter, J.; Parrinello, M. J. Am. Chem. Soc. 1996, 118,
7847-7848. Rovira, C.; Kune, K.; Hutter, J.; Ballone, P.; Parrinello, M. J.
Phys. Chem. A 1997, 101, 8914-8925.
(41) Koritsa´nszky, T.; Flaig, R.; Zobel, D.; Krane, H.-G.; Morgenroth,
W.; Luger, P. Science 1998, 279, 356-358.
(42) Flaig, R.; Koritsanszky, T.; Zobel, D.; Luger, P. J. Am. Chem. Soc.
1998, 120, 2227-2238.
(43) Coppens, P. X-ray Charge Densities and Chemical Bonding; Oxford
University Press: New York, 1997.
(44) Bader, R. F. W. Atoms in Molecules. A Quantum Theory; Clarendon
Press: Oxford, 1990.
(45) Herzfeld, J.; Berger, A. J. Chem. Phys. 1980, 73, 6021-6030.

CO-Heme Analogue Structures
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3825
Table 4. Comparison of Experimental NMR Chemical Shifts and Chemical Shift Tensor Elements with Theoretical Values, for CO and iPrNC
Model Systemsa,b
δiso
(ppm)
δ11
(ppm)
δ22
(ppm)
δ33
(ppm)
|δ33 - δ11|
system
(ppm)
Fe(OEP)(13CO)(1-MeIm)
expt
calc
expt
calc
expt
calc
expt
calc
expt
calc
expt
calc
expt
calc
expt
calc
expt
expt
expt
expt
expt
expt
expt
206.9
215
358
385
343
383
-83
441
508
-123
Fe(OEP)(C17O)(1-MeIm)
Ru(OEP)(13CO)(1-MeIm)
Ru(OEP)(C17O)(1-MeIm)
Os(OEP)(13CO)(1-MeIm)
Os(OEP)(C17O)(1-MeIm)
Fe(TPP)(iPrN13C)(1-MeIm)
Fe(TPP)(iPr15NC)(1-MeIm)
376
183
188
307
290
141
159
242
199
158
174
186
168
151
188
635
336
370
634
336
368
-142
-124
-174
777
460
544
558
296
352
557
296
350
-244
-169
-227
802
465
579
507
504
-416
923
325
346
331
306
330
312
-108
-118
-138
432
464
469
Fe(TPP)(iPrN13C)(py)c
Fe(TPP)(iPr15NC)(py)c
sperm whaled Mb•iPrN13C
adult humand Hb•iPrN13Cd site 1
site 2
357
332
-125
482
173.4
177.3
178.0
170.4
176.9
rabbit Hb•iPrN13C site 1
site 2
a The theoretical chemical shieldings (σ) were converted to theoretical chemical shifts (“calc” in the table) by using the following absolute
shieldings: 13C, δ(13C, TMS) ppm ) 186.5 - σ (ref 51); 15N, δ(15N, NH3) ppm ) 244.6 - σ (ref 52), and for 17O, δ(17O,H2O) ppm ) 306.7 -
σ (ref 53). b Shielding calculations were typically performed by using the G94/DFT/GIAO Fe Wachters’/6-311++G(2d)/6-31G//3-21G//BPW91
approach with crystallographic structures: see the text for more details. c The structure of this compound was not solved so only the experimental
shift results are shown. d Reference 48.
CO species and the 13C spectrum of Fe(TPP)(iPrN13C)(1-MeIm),
only the isotropic shifts are reported, due to weak signal-to-
noise ratios and quadrupolar effects. Our 13C isotropic shift for
the Fe(TPP)(iPrNC)(py) compound is essentially the same as
that determined previously by Morishima et al. for a (reported)
bis-isocyanide TPP complex,47 but we believe that their
compound was actually the mono-pyridine adduct, since it has
the same chemical shift as our Fe(TPP)(iPrNC)(py), and their
solvent was pyridine. For both the 13CO and R15NC shielding
tensors, we find good accord between theory and experiment
for each of the individual tensor elements, Table 4 and Figures
6A,B. This result was expected for the Fe, Ru, Os systems, based
on the results found previously for Fe-, Ru-, Os(TPP)(CO)(1-
MeIm),17 but no shielding calculations have been reported for
RNC-metalloporphyrins. In addition, C17O isotropic shieldings
are quite well predicted, Figure 6C. For the essentially linear
and untilted Fe-, Ru- and Os-13CO compounds, we find that
there is very close to axial symmetry in both the experimental
shielding tensor elements and the calculations, Table 4. More
surprisingly, the 15N shielding tensor, both experimentally and
theoretically, is also very close to being axially symmetric, with
σii ) 346, 330, and -118 ppm (expt) and σii ) 331, 312, and
-138 ppm (calc), with overall tensor spans of 464 (expt) and
469 ppm (calc). As discussed elsewhere, errors in the absolute
shielding in such systems (for 13C, 15N, and 17O) are expected
to be ∼10-20 ppm.17
major change from the small shielding differences seen previ-
ously between model 13CO-containing compounds and metal-
loproteins, which are typically only ∼2 ppm.17 One possible
explanation for this unusual finding might be that the RNC
group has been reported to display a large range of tilt (τ) and
bend (â) angles in proteins,21,22 where τ and â are as shown
below:
In addition, large distortions at nitrogen in the alkylisocyanide
adducts of myoglobin and hemoglobin have been reported,21,22
and might also influence shielding. For example, C′NR bond
angles (θ, above) from 145.6 to 96.0° have been noted and
(48) Mansuy, D.; Lallemand, J.-Y.; Chottard, J.-C.; Cendrier, B.; Gacon,
G.; Wajcman, H. Biochem. Biophys. Res. Commun. 1976, 70, 595-599.
Dill, K.; Satterlee, J. D.; Richards, J. H. Biochemistry 1978, 17, 4291-
4297. Stetzkowski, F.; Banerjee, R.; Lallemand, J.-Y.; Cendrier, B.; Mansuy,
D. Biochimie 1980, 62, 795-801.
(49) Peng, S.-M.; Ibers, J. A. J. Am. Chem. Soc. 1976, 98, 8032-8036.
(50) Little, R. G.; Ibers, J. A. J. Am. Chem. Soc. 1973, 95, 8583-8590.
Slebodnick, C.; Seok, W. K.; Kim, K.; Ibers, J. A. Inorg. Chim. Acta 1996,
243, 57-65.
(51) Jameson, K. J.; Jameson, C. J. Chem. Phys. Lett. 1987, 134, 461-
466.
In sperm whale myoglobin and in adult human and rabbit
hemoglobin RNC adducts, there are, however, considerable
shielding differences to those seen with the model compound.48
In particular, the iPrN13C shift in the proteins is some 15-20
ppm downfield from that observed in the model compound, a
(46) Havlin, R. H.; McMahon, M.; Srinivasan, R.; Le, H.; Oldfield, E.
J. Phys. Chem. 1997, 101, 8908-8913.
(47) Morishima, I.; Hayashi, T.; Inubushi, T.; Yonezawa, T.; Uemura,
S. J. Chem. Soc. Chem. Commun. 1979, 483-485.
(52) Jameson, C. J.; Jameson, A. K.; Oppusunggu, D.; Willie, S.; Burrell,
M.; Mason, J. J. Chem. Phys. 1981, 74, 81-88.
(53) Wasylishen, R. E.; Mooibroek, S.; Macdonald, J. B. J. Chem. Phys.
1984, 81, 1057-1059.

3826 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Salzmann et al.
Figure 7. Electrostatic field contributions to shielding in [FeCO](2e)
and [FeCNMe](2e) model systems. 2, 13CO; ∆, CH3N13C; b, C17O;
and O, CH315NC. The isotropic chemical shifts are shown as a function
of the size of the charge placed 4.45 Å distal to the terminal O, N
atoms, along the Fe-C bond axis. 13CO slope ) 4.6 ppm/au, R2 )
1.0; CH3N13C slope ) -9.0 ppm/au, R2 ) 0.99; C17O slope ) 25.4
ppm/au, R2 ) 1.0; CH315NC slope ) 22.1 ppm/au, R2 ) 1.0.
rationalized in terms of the existence of the Pauling valence
forms:
Figure 8. Ligand τ,â energy surfaces for three model metalloporphyrin
systems. (A) Fe(bis(amidinato))(MeNC)(Im), this work. (B and C)
Energy surfaces for the FeIIICN (B) and Fe(CO) (C) derivatives, based
on the work of Vangberg et al.30 The contour values are in kcal/mol.
The RNC derivative is most sensitive to both tilt and bend, CO the
least. The energy penalties at large τ,â with the isocyanides make such
distortions less likely in metalloproteins.
Unfortunately, it is not at present very realistic to expect to
reliably evaluate these protein shieldings, since there are too
many structural uncertainties. Indeed, as with our previous work
on 13C, 17O NMR of CO-heme proteins,18-20 use of the reported
X-ray structures does not give good agreement with experiment
(data not shown). It is, however, of interest to try to assess which
factors are most likely to influence isocyanide shielding and
energetics in proteins, since this approach could lead to ruling
out some explanations while favoring others.
experimentally. The τ,â energy surface for the Fe(bis(amidi-
nato))(MeNC)(Im) model is shown in Figure 8, together with
the corresponding FeIII(bis(amidinato))(CN-)(Im) and Fe(bis-
(amidinato))(CO)(Im) surfaces reported previously by Vangberg
and co-workers.30 As anticipated, ligand tilt (τ) plays the major
role in controlling energetics, since both σ and π* effects will
be strongly affected by moving the nearest neighbor C atom
off of the heme normal. That is, σ bonding will greatly decrease
as C moves away from the pz, dz2 orbitals, while bending (at C)
has a much less pronounced effect, as with CO.30 For τ j 15°,
the τ,â energy surfaces for all three systems are very similar,
but at high τ, the RNC energies are larger than those seen in
the CO, CN surfaces, making large τ,â inaccessible. Signifi-
cantly though, from the perspective of chemical shielding,
increasing τ,â causes a uniform increase in 13CNR shielding,
the same effect as that seen previously for the Fe-CO
containing systems,20 but the opposite of the model compound
f protein result seen experimentally. For example, for τ ) 30°,
â ) 0°, we find for the isocyanide adduct a 10 ppm increase in
shielding, which is to be compared with the 12.2 ppm result
found with the CO adduct. Neither tilt nor â-bend therefore
appear to be responsible for the large deshieldings seen in the
metalloproteins. The only other likely solution, therefore,
appears to be one in which a θ-bend (at N) is involved. We
first carried out a test calculation at τ ) â ) 0°, θ ) 128°, a
structure close to that of the bent Pauling valence form shown
above. We found a 46 ppm deshielding, a major effect that can
We therefore first investigated the sensitivity of both 13C and
15Nshieldingstoelectrostaticfieldeffects,usingthe[FeCNMe](2e)
cluster shown in Figure 1A, and the same basis sets as reported
previously by deDios and Earle.27 Since linear relationships
between charge and shift are expected,27 only a limited number
of points were evaluated, and these results are shown, together
with a corresponding set of results for [FeCO](2e), in Figure 7.
There is the expected anti-correlation between RN13C and R15-
NC shieldings (due to polarization) as a function of the
electrostatic field (the value of the external charge), with a close
correspondence between the results for the two systems. This
strongly suggests that it is most unlikely that electrostatic field
effects alone could be responsible for the 13C chemical shift
changes seen between the model compound and the metallo-
proteins, since in the presence of a similar local charge field to
that present in the CO-heme proteins,27 13C shift changes of
only a few ppm would be anticipated, not the 15-20 ppm seen
experimentally. More pronounced structural differences are,
therefore, indicated.
We consequently evaluated tilt-bend (τ,â) energy and
shielding surfaces for the model system shown in Figure 1B,
to see to what extent such distortions are permissible, and to
what extend they could contribute to the shielding changes seen

CO-Heme Analogue Structures
Table 5. Computed 13C, 15N Shieldings (ppm) and Energies (kcal) for Fe(bis(amidinato))(RNC)(Im) Model Compoundsa
13C 15N
σ22 σ22
J. Am. Chem. Soc., Vol. 121, No. 16, 1999 3827
ligand
MeNC
τ,â,θ (deg)b
σi
σ11
σ33
|σ33 - σ11|
σi
σ11
-3
σ33
|σ33 - σ11|
Ec
(0,0,180)
6
7
8
-117
-117
-116
-114
-117
-116
-115
-113
-114
-115
-114
-101
-98
-105
-105
-104
-102
-105
-102
-99
242
243
245
247
242
241
240
240
238
236
233
234
233
242
241
240
237
232
226
220
359
360
361
361
359
357
355
353
352
351
347
335
331
365
381
407
437
466
491
503
116
116
115
114
116
117
117
116
120
119
115
119
117
114
110
105
98
6
6
344
347
355
368
339
338
342
350
318
316
321
268
258
344
344
343
340
333
325
315
347
354
373
404
340
338
344
362
318
314
322
277
264
349
363
383
406
429
451
468
0.0
0.66
2.90
6.88
3.42
2.65
4.00
7.31
11.74
13.39
18.24
30.35
31.87
0.20
0.47
1.37
2.68
5.46
9.32
16.03
(0,10,180)
(0,20,180)
(0,30,180)
(10,0,180)
(10,10,180)
(10,20,180)
(10,30,180)
(20,10,180)
(20,20,180)
(20,30,180)
(30,0,180)
(30,10,180)
(0,0,170)
(0,0,160)
(0,0,150)
(0,0,140)
(0,0,130)
(0,0,120)
(0,0,110)
-7
-18
-36
-1
7
9
11
7
7
9
10
9
10
11
16
11
13
11
9
41
38
26
98
98
3
0
-2
-12
0
-94
-97
-90
2
-84
-1
-86
-9
16
3
-86
-6
-123
-140
-167
-200
-234
-265
-283
-109
-113
-118
-123
-128
-133
-136
-5
-4
-15
-28
-43
-57
-67
-19
-40
-66
-96
-126
-153
5
13
23
37
54
75
91
84
79
a Chemical shieldings were evaluated with the deMon/MASTER-CS program (refs 31-33), as described previously for the carbonmonoxy
system (ref 20). b The tilt (τ), C-bend (â), and N-bend (θ) distortions are in degrees and are as shown in the text. c The energies shown are energy
differences with respect to the τ ) â ) 0°, θ ) 180° (linear) conformer, in kcal/mol.
θ-bend distortion of ∼30° (θ ) 150°), corresponding to only a
∼1.5 kcal distortion, would be quite sufficient to explain the
protein shift seen experimentally.
Conclusions
The results we have presented above are of interest for a
number of reasons. First, they represent the first detailed
comparative study of the structures of six-coordinate CO-
liganded octaethyl- and tetraphenylporphyrins. The results show
that all three OEPs are planar, while the analogous TPP
derivatives are ruffled.17 Second, we have reported the first
example of an RNC/1-MeIm metalloporphyrin, a model for alkyl
isocyanides bound to heme proteins. The porphyrin here is
planar, even though TPP and 1-MeIm are present. Third, we
have presented a logic-based approach that successfully predicts
the presence of porphyrin distortions in 16 out of 16 CO, RNC,
NO, CCl2, RNO containing OEP/TPP metalloporphyrins. Fourth,
we have obtained 13C, 15N, and 17O solid-state NMR spectra of
the four new metalloporphyrins, and used DFT methods to
successfully reproduce the experimental spectra. An interesting
observation is that the 13C shift in iPrNC containing metallo-
proteins is ∼18 ppm deshielded from that observed in the
synthetic model compound. DFT calculations rule out electro-
static field effects or Fe-C-N tilt/bend, but a bend at nitrogen
in the proteins, such as is present in one of the Pauling valence
forms, would introduce carbene-like character, resulting in
deshielding at carbon, as observed experimentally.
Acknowledgment. We are grateful to B. J. Moreno, W. D.
Arnold, S. R. Wilson, and T. Prussak-Wiekowska for experi-
mental assistance, V. Malkin, O. Malkina, and D. Salahub for
Figure 9. 13C, 15N shielding and energy changes as a function of the
N-bend (θ) in the Fe(bis(amidinato))(MeNC)(Im) model compound:
(A) 13C shielding and (B) change in energy.
providing their deMon DFT/shielding program, and K. Suslick
and C. Ziegler for access to and help with experimental facilities.
This work was carried out in part by use of the SGI/Cray Origin
2000 and Power Challenge clusters at the National Center for
Supercomputing Applications (funded in part by the US National
Science Foundation, grant CHE-97002ON). The X-ray crystal-
lographic facilities were supported by the National Science
Foundation (grant CHE 95-03145). Solution NMR spectra were
obtained in the Varian-Oxford Instruments Center for Excellence
be qualitatively related to the carbene-like nature of the Fed
C-NR tautomer. For example, the 13C shielding in the true
carbene Fe(TPP)(CCl2) is at ∼225 ppm downfield from TMS,39
a very large change from that seen in the RNC model system.
This expectation is borne out by the more detailed results
presented in Table 5 and in Figure 9, which show the 13C
shielding and energetics as a function of the θ-bend angle. A

3828 J. Am. Chem. Soc., Vol. 121, No. 16, 1999
Salzmann et al.
in NMR Laboratory. Funding for this instrumentation was
provided in part from the W. M. Keck Foundation, the National
Institutes of Health (grant RR-10444), and the National Science
Foundation (grant CHE 96-10502). High-resolution mass spectra
were obtained in the Mass Spectrometry Laboratory, School of
Chemical Sciences, University of Illinois, supported in part by
a grant from the National Institute of General Medical Sciences
(grant GM-27029). The 70-VSE mass spectrometer was pur-
chased in part with a grant from the Division of Research
Resources, National Institutes of Health (grant RR-04648).
Supporting Information Available: Tables of crystal data,
structure solution and refinement, atomic coordinates, aniso-
tropic displacement parameters, hydrogen coordinates, bond
lengths and angles, and torsion angles (PDF). X-ray crystal-
lographic files, in CIF format, are also available through the
Internet. This material is available free of charge via the Internet
at http://pubs.acs.org.
JA9832818