The problem with the last mechanism is simple:

• It doesn't correspond to the physical reality of the bonding in the reactant

• The reacting carbon is sp²; therefore the bond to carboxyl is a s-bond in the plane of the ring

• Therefore it cannot bind a proton as suggested by the picture

Furthermore, I did an ab initio calculation on orotidine (with a CH₃ in place of the ribose).

• I had to go four MOs below the HOMO, down 27.6 kcal/mol in energy, to find the MO corresponding to the bond to the carboxyl.

Several investigators suggested that instead of doing a nucleophilic attack at C-5, it should be protonated by Lys93:

I have modified the mechanism slightly - the authors write a carbene as the decarboxylation product.

This mechanism is easily testable. Just run the reaction in D₂O. At least some D should be found at C-5 in the product. So far, the experiment has not been done.

A final mechanism comes from a non-observation and a couple of observations.

• None of the crystal structures has OMP itself bound

• Two have BMP (with an OH); two have the product UMP; and one has 6-azaUMP - that is, no carboxyl is present in any of the molecules bound in the catalytic site

• A molecular mechanics calculation suggests significant steric hindrance to placing the carboxyl of OMP where the OH of BMP is found

• 2-thiaOMP (A below) binds but does not decarboxylate

Perhaps, therefore, the conformation of OMP when bound is as shown in B above.

• If so, then Lys93 could protonate the C-2 carbonyl

• The carbanion would then be protonated by water

So far, no crystallographic verification of this idea is available: the OMP decarboxylates so rapidly that it is gone before crystals can be grown.

Where does this leave the mechanism of ODCase?

We have at least eight mechnisms, each of which has some supporting evidence from experiment or calculation. Each of them also has defects revealed by experiment or calculation. Here they are, as summarized by Houk [J. Phys. Chem. B, 2007, 111, 12573], who used high-level MO calculations to evaluate the energetic of each. No calculated value for the activation energy matches the experimentally determined one.

No final, definitive mechanism is available. But I have my own favorite:

By analogy to electrophilic aromatic substitution.