ODCase active site models

Results. Lee and Houk were the first to model part of the ODCase active site when they calculated decarboxylation energetics for orotate in the presence of methylammonium ion as a mimic of the key active site lysine. Based on their conclusion that 4-protonation is an energetically favorable pathway (see above), they calculated the energy of reaction of orotate (la) plus Cf NHj to form a carbene-methylamine complex plus C02 (equation 1). 

The best low-energy models correspond to arrangements that are very unlikely given the crystallographically determined structure of the ODCase active site (Fig. 2). The calculated barriers never drop below 30 kcal mol-1, which is much higher than the experimentally observed barrier for decarboxylation by ODCase (AG = 15 kcal mol-1, Fig. I),1,6 prompting the authors to conclude that direct protonation is not a viable mechanism. 

Developing Active Site Models of ODCase— from Large Quantum Models to a QM/MM Approach... 

How can a theoretical method decide between proposed mechanisms, and how can the origin of the enzymatic power be identified This review will try to answer these questions for one particular theoretical approach, the one where an active site model is treated by accurate quantum mechanical (QM) methods. The main idea in the QM active site approach is to make sure that the computational results have the required accuracy. During the last decade the accuracy of density functional methods (DFT) has been dramatically improved, and in particular the hybrid B3LYP functional has achieved a remarkable accuracy [8, 9]. The use of DFT has also made it possible to treat dramatically larger molecular systems than can be done with conventional wave-function methods of similar accuracy. In spite of this important development, DFT models have usually been limited to 50-60 atoms, but more recently systems with more than 100 atoms have been treated efficiently. Still, even 100 atoms is a very small part of the total number of 8,300 atoms in yeast ODCase, not counting hydrogens or surrounding water molecules. Thus a very severe selection has to be made when the enzyme model is set up, and an important task is to select the residues required to solve the mechanism and to analyze all important contributions. 

Developing Active Site Models of ODCase—from Large Quantum Models 91... 

Nonetheless, there is still hope that quantum mechanical studies may play a key role in deducing the ODCase mechanism. What these studies have shown is that several mechanisms are energetically viable. They have also provided structural models of transition states and their complexes with active site groups that can be used to design experiments for distinguishing between the several mechanisms that remain in the running. One particularly promising experiment that has already been proposed is the measurement of the 1SN decarboxylation isotope effects for the N3 site of OMP. Phillips and Lee have made the computational prediction that while decarboxylation via 2-protonation and without pre-protonation should have normal isotope effects (1.0027 and 1.0014, respectively), the 4-protonation pathway should display an inverse IE of 0.9949.47 Thus, the combination of computationally predicted and experimentally measured IE values may ultimately lead to elucidation of the enzyme mechanism. 

The enzyme mechanism, however, remains elusive. Quantum mechanical models generally disfavor C6-protonation, but 02, 04, and C5-protonation mechanisms remain possibilities. Free energy computations also appear to indicate that C5-protonation is a feasible mechanism, as is direct decarboxylation without preprotonation O-protonation mechanisms have yet to be explored with these methods. Controversy remains, however, as to the roles of ground state destabilization, transition state stabilization, and dynamic effects. Because free energy models do take into account the entire enzyme active site, a comprehensive study of the relative energetics of pre-protonation and concerted protonation-decarboxylation at 02, 04, and C5 should be undertaken with such methods. In addition, quantum mechanical isotope effects are also likely to figure prominently in the ultimate identification of the operative ODCase mechanism. 

Kollman and coworkers apphed a variety of computational methods to this mechanistic problem—including quantum mechanics on small model systems, molecular dynamics simulations with the AMBER force field on the whole ODCase-substrate system, and MM-PBSA free energy calculations on ODCase with bound OMP [38]. Based on their results, they proposed a decarboxylation mechanism for ODCase that involves C5 protonation. Their calculations at the MP2/6-31+G //HF/6-31+G level showed that C5 has a greater intrinsic proton affinity than C6, 02, and even 04. This, coupled with the fact that Lys72 (M. thermoautotrophicum numbering see Table 2) is near C5 and C6 in the inhibitor-bound crystal structures, prompted the authors to embrace a C5 protonation mechanism. However, the authors themselves acknowledged the uncertainties of their calculations because of approximations employed in representing the enzyme active site. 

Fig. 3 Picture of the active site in ODCase showing only selected amino acids. The geometry of the active site is taken from the X-ray structure IDQX and the carboxylate group has been added to the inhibitor BMP. Selected residues have either been shown to be important for catalysis or play important roles in the computational models of this review. No hydrogens are included in the figure...

The present methodology, using accurate QM models that treat only a rather small part of the active site, has met considerable success during the last few years. The authors of the present review have used the method mainly for metalloenzymes [18, 19, 20] but have also applied the methodology to ODCase [21]. That study treated the concerted reaction mechanism and the base protonation mechanism with 02 as protonation site. The present review includes those results but also presents significant extensions to the modeling of these mechanisms. In addition, results from investigations of other base protonation mechanisms, and the mechanism where the C-C bond is cleaved prior to protonation are also presented. 

This contribution reviews computational results for three classes of reaction mechanisms proposed for ODCase. Firstly, the mechanism that assumes protonation of C6 concerted with decarboxylation is described. Secondly, the base protonation mechanisms are reviewed. Finally, a shorter treatment is given of a reaction mechanism where the C-C bond is broken before the proton attaches to the base. All values in the review are obtained by the use of QM models of the active site. Effects of different residues on the reaction barrier are analyzed when going from small to large QM models. A QM/MM treatment is applied to each mechanism to see whether this treatment has any major effect on the calculated results. The goal of the review is to provide information regarding the activity of ODCase and to shed light on the requirements on QM models that are applied to enzymatic systems. 

The above results, using extended quantum mechanical models for the active site of ODCase, indicate that the base protonation mechanism has a decarboxylation barrier that is too high to be compatible with the experi-... 

In analyzing the origin of enzyme catalysis, Warshel and others have advocated the importance of comparing the enzymatic reaction with a reference reaction in water [32]. In addition, it is also necessary to study the reference reaction in the gas phase in order to understand the intrinsic reactivity and the effect of solvation. Thus, to understand enzyme catalysis fully, we must compare results for the same reaction in the gas phase (intrinsic reactivity), in aqueous solution (solvation effects), and in the enzyme (catalysis). This is not possible when there is no model reaction for the uncatalyzed process in the gas phase and in water, or if the uncatalyzed reaction is a bimolecular process as opposed to a unimolecular reaction in the enzyme active site. None of these problems apply to the ODCase reaction. Furthermore, OMP decarboxylation is a unimolecular process, both in water and the enzyme, providing an excellent opportunity to compare directly the computed free energies of activation [1] this is the approach that we have undertaken [16]. Warshel et al. used an ammonium ion-orotate ion pair fixed at distances of 2.8 or 3.5 A as the reference reaction in water to mimic an active site lysine residue [32]. 

Lundberg, Blomberg, and Siegbahn show how modern quantum chemistry has been applied to the ODCase problem. They describe results from quantum mechanical calculations on large models of OMP and the active site residues in ODCase that surround it, and provide a critical evaluation of many of the proposed mechanisms for catalysis. 

---