Resonance Raman (RR) spectra are reported for structurally defined CO adducts of two sterically constrained Fe(II) porphyrins: PocPiv, in which three of the pivaloylamino pickets of ''picket fence'' porphyrin (5,10,15,20-tetrakis[o-(pivaloylamino)phenyl]porphyrin) are attached to a benzene cap by single methylene groups, and C2-Cap, in which a benzene cap is attached by carboxylate links and a pair of methylene groups to the four hydroxyl groups of 5,10,15,20-tetrakis(o-hydroxyphenyl)porphyrin. Although the X-ray crystal structures of the two adducts show very similar FeCO geometries, involving a small amount of bending and tilting, their vibrational frequencies and RR enhancement patterns are very different. Relative to unconstrained porphyrins, the C-0 and Fe-C stretching (nu(CO) and nu(FeC)) frequencies show increased Fe-->CO back-bonding for PocPiv but decreased back-bonding for C2-Cap. The Fe-C-O bending mode (deltaF(FeCO)) is activated for PocPiv but not for C2-Cap. These contrasting patterns can be attributed to the different polar interactions of the bound CO in the two adducts. In the PocPiv adduct, the 0 atom is in close contact with one of the amide NH groups; this positive polar interaction increases back-bonding. In the C2-Cap adduct, the linker arms have ester and ether, rather than amide, groups and there is a close contact between the CO and the benzene pi-cloud; this negative polar environment decreases back-bonding. In addition the cap interaction may compress the FeC bond in C2-Cap, as evidenced by a nu(FeC) frequency elevation from the value expected on the basis of the back-bonding decrease. The extensive vibrational data on CO adducts of heme proteins and on sterically constrained synthetic porphyrins are reexamined in the light of these results. Essentially all of the data are consistent with expected polar effects in the binding pocket. Positive distal polar interactions (1) increase nu(FeC) and decrease nu(CO) along a back-bonding correlation that depends only on the nature of the trans axial ligand, (2) diminish the intensity of the nu(FeC) RR band (attributed to a reduction in the Fe-C bond displacement between the ground and excited states), and (3) activate the FeCO bending mode (by perturbing the 4-fold electronic symmetry via off-axis interactions). A recent proposal that the RR band usually assigned to delta(FeCO) actually arises from the bending mode overtone is discussed and is rejected on the basis of comparisons with transition metal carbonyls for which thorough vibrational analyses are available. Elimination of positive polar interactions reverses all three of these effects. The energetics of the small FeCO distortions seen in the PocPiv and C2-Cap adducts are evaluated and are found to account for most of the CO affinity reductions of these constrained porphyrins. The polar interactions appear to contribute only a small part of the overall energy, even though they have a dominant effect on the vibrational frequency variation. It is argued that large sterically-induced FeCO distortions are precluded by the prohibitive energy cost, especially in proteins, where the available steric force is limited by the conformational flexibility of the polypeptide. Somewhat greater FeCO bending than has so far been observed in models is indicated for one of the substates, A3, of the CO adduct of myoglobin. This effect is suggested to arise from a donor interaction with an imidazole lone pair, consistent with the distal histidine tautomer determined by neutron diffraction. Even in this case, however, the IR frequency precludes FeCO angles as small as 120-140-degrees, values reported from early determinations of MbCO crystal structures containing disordered CO. The dominant MbCO substates in solution, A1,2, are indicated to have a significant polar interaction with the distal histidine, in the alternate tautomer to the one seen in the neutron structure, and probably have nearly linear FeCO units, consistent with the most recent X-ray structure determination of a nondisordered crystal form.
