Normally, the metal ion in the porphyrins is six-coordinate with four ligand atoms provided by the porphyrin and two axial ligands from the protein or the solution. Thus, a typical metal porphyrin complex, e.g. haemoglobin, contains about 90 atoms, which can be reduced to 50-60 atoms if the side-groups of the porphyrin ring are removed. This large number of atoms has precluded quantum chemical studies of these complexes, except for studies at a low level of theory (semiempirical or minimal-basis calculations) on complexes enforced to have a high symmetry. The development of fast computers and effective methods for the quantum chemical treatment of many-atom systems during the last few years have opened the possibility to study metal porphyrins with accurate high-level methods.
As our first project, we studied how myoglobin discriminates between O2 and CO. CO binds ~20 000 times stronger than O2 to haem in solution, but in the proteins this factor is only 25. Thus, myoglobin favours O2 before CO by about 17 kJ/mole. This is crucial to life, since about 1% of the haem units are poisoned by endogenously produced CO in healthy people. According to biochemical textbooks, the reason for this is that the nearby distal histidine residue forces CO to bind in a bent fashion (the natural conformation of O2) [10]. Although, the suggestion is supported by X-ray and neutron structures, recent spectroscopic experiments have questioned it. We have studied hydrogen bonds between histidine and a haem-bound CO or O2. The energy difference of hydrogen bonds to these two ligands, 24 kJ/mole, is close to the protein's discrimination between these two molecules [31]. Thus, our results indicate that the protein discriminates between the two ligands by electrostatic means rather than by strain. In addition we have studied the vibrational frequencies of CO and showed that the neutron structure of this protein, probably involves an unnatural conformation of the protein.
Secondly, we have studied the reorganisation energy of cytochromes [43]. In these, the haem group is the same as in myoglobin, but the axial ligands are different. The iron ion is invariably six-coordinate with two axial ligands, typically, histidine, methionine, tyrosine, or the amino terminal. Uncharged ligands give a very low reorganisation energy (around 8 kJ/mole), whereas charged ligands give a much higher energy, up to 50 kJ/mole. As expected, most electron-transfer sites in nature utilise uncharged ligands. The reorganisation energy of cytochromes is appreciably lower than for the the other electron carriers used in nature, blue copper proteins and iron-sulphur clusters [29,40,42,44].
Ferrochelatase is the enzyme that incorporates iron into the haem group. It has been crystallised and studied thoroughly at three experimental departments in Lund. We have studied the reaction of this enzyme to explain the experimental data. As a first project we have studied how much energy it costs to bend the porphyrin ring without or with different metal ions in the centre [?]. This study makes it easier to understand the crystal structures. We have also interpreted the effect of binding different metal ions to various sites of the protein [?].
Kasper P. Jensen, my third student, studies the reactions of vitamin B12-dependent enzymes. Vitamin B12 consists of a Co ion (Co+, Co2+, or Co3+) in a corrin ring (a highly saturated, modified porphyrin). The cobalt ion forms a unusual Co3+-C bond, which is an almost unique example of a metal-carbon bond i biology. This bond may be cleaved either homolytically (to Co2+ and a carbon radical) or heterolytically (to Co+ and a carbocation), depending on the enzyme. We study how the enzymes discriminate between these two pathways and how they may accelerate the rate of Co-C bond cleavage by a factor of 1012.
As a first project [46], we have shown that the most widely accepted mechanism for homolytic enzymes, the mechanochemical trigger mechanism is unlikely. This mechanism suggests that the ligand opposite to the Co-C bond may interact with the corrin ring, causing it to bend upwards by steric repulsion and destabilise the Co-C bond. We show that the trans axial Co-N bond is very flexible and that constraints in this bond has small influence on the structure of the corrin ring and on the strength of the Co-C bond.
Updated 2001-04-27