A Case Study of the Glycoside Hydrolase Enzyme Mechanism Using an Automated QM-Cluster Model Building Toolkit

Abstract

Glycoside hydrolase enzymes are important for hydrolyzing the β-1,4 glycosidic bond in polysaccharides for deconstruction of carbohydrates. The two-step retaining reaction mechanism of Glycoside Hydrolase Family 7 (GH7) was explored with different sized QM-cluster models built by the Residue Interaction Network ResidUe Selector (RINRUS) software using both the wild-type protein and its E217Q mutant. The first step is the glycosylation, in which the acidic residue 217 donates a proton to the glycosidic oxygen leading to bond cleavage. In the subsequent deglycosylation step, one water molecule migrates into the active site and attacks the anomeric carbon. Residue interaction-based QM-cluster models lead to reliable structural and energetic results for proposed glycoside hydrolase mechanisms. The free energies of activation for glycosylation in the largest QM-cluster models were predicted to be 19.5 and 31.4 kcal mol−1 for the wild-type protein and its E217Q mutant, which agree with experimental trends that mutation of the acidic residue Glu217 to Gln will slow down the reaction; and are higher in free energy than the deglycosylation transition states (13.8 and 25.5 kcal mol−1 for the wild-type protein and its mutant, respectively). For the mutated protein, glycosylation led to a low-energy product. This thermodynamic sink may correspond to the intermediate state which was isolated in the X-ray crystal structure. Hence, the glycosylation is validated to be the rate-limiting step in both the wild-type and mutated enzyme.

Eliminating Imaginary Vibrational Frequencies in Quantum-Chemical Cluster Models of Enzymatic Active Sites

“…The most common way to apply electronic structure theory to study enzymatic reactions is to use a hybrid quantum mechanics/molecular mechanics (QM/MM) formalism combined with all-atom MD simulations. − Setup of QM/MM calculations requires considerable care, − and an underappreciated aspect is just how slowly thermochemical predictions converge with respect to the size of the QM region, typically requiring hundreds of QM atoms. − Some progress has been made toward automated selection of QM model regions. − …”

Eliminating imaginary vibrational frequencies in quantum-chemical cluster models of enzymatic active sites

“…[12][13][14][15][16][17][18][19][20][21][22][23][24] Some progress has been made toward automated selection of QM model regions. [22][23][24][25][26] An alternative to QM/MM-MD simulations to study mechanistic aspects of enzyme catalysis is to use limited "cluster" models of the active site. [27][28][29][30][31] This approach neglects any atomistic description of the larger protein environment, and is thus unable to describe chemical transformations that are driven by conformational changes of the protein, but for a limited set of problems the QMcluster approach has an important advantage of simplicity.…”

Fragment-Based Calculations of Enzymatic Thermochemistry Require Dielectric Boundary Conditions

“…(The use of smooth cutoffs in gradient calculations has already been demonstrated. 7 ) Network analysis can be used to build sensible (if sizable) models of the enzyme−substrate complex, 55,106,107 and then the protocols developed here can provide converged results for any given model. Together, these developments promise to make QC modeling of enzymatic reactions more robust and systematic.…”