Can Molecular Dynamics and QM/MM Solve the Penicillin Binding Protein Protonation Puzzle?

Abstract

Benzylpenicillin, a member of the β-lactam antibiotic class, has been widely used to combat bacterial infections since 1947. The general mechanism is well-known: a serine protease enzyme (i.e., DD-peptidase) forms a long lasting intermediate with the lactam ring of the antibiotic known as acylation, effectively preventing biosynthesis of the bacterial cell wall. Despite this overall mechanistic understanding, many details of binding and catalysis are unclear. Specifically, there is ongoing debate about active site protonation states and the role of general acids/bases in the reaction. Herein, a unique combination of MD simulations, QM/MM minimizations, and QM/MM orbital analyses is combined with systematic variation of active site residue protonation states. Critical interactions that maximize the stability of the bound inhibitor are examined and used as metrics. This approach was validated by examining cefoxitin interactions in the CTX-M β-lactamase from E. coli and compared to an ultra high-resolution (0.88 Å) crystal structure. Upon confirming the approach used, an investigation of the preacylated Streptomyces R61 active site with bound benzylpenicillin was performed, varying the protonation states of His298 and Lys65. We concluded that protonated His298 and deprotonated Lys65 are most likely to exist in the R61 active site.

Ligand-Induced Proton Transfer and Low-Barrier Hydrogen Bond Revealed by X-ray Crystallography

“…Moreover, the QM/MM calculations predicted that the charged (Lys73-NH 3 ) + -(Ser70-O) − pair will become neutral (Lys73-NH 2 )-(Ser70-OH) upon β-lactam substrate binding, and that the proton on Ser70 will transfer to Lys73 as Ser70 attacks the substrate 26 . We previously performed a QM/MM investigation of CTX-M with the non-covalent β-lactam substrate cefoxitin 29 . Two protonation state combinations were examined as a function of altering the Lys73 and Glu166 protonation state.…”

Acylation and deacylation mechanism and kinetics of penicillin G reaction with StreptomycesR61 DD‐peptidase

“…It should be noted that Hargis and Woodcock found that doubly protonated His298 (N ε and N δ ) might be required for facile acylation. [ 31 ] Therefore, two different models were generated by our in‐house program package Residue Interaction Network‐based ResidUe Selector (RINRUS) and were used to explore mechanistic differences arising from the protonation state of His298. The first model contains 277 atoms (His298 N ε protonated as pH = 6.8 for the experimental condition) with 17 C α atoms frozen and 5 C β atoms frozen (in Phe120, Tyr159, Asn161, Trp233, and Arg285).…”

“…Various mutational, enzyme kinetic, crystallographic, and computational studies on DD‐peptidases and class A and class C β‐lactamases (produced by bacteria and resistant to β‐lactam antibiotics, ABL and CBL in short) have devoted extensive discussion to proton transfer throughout the mechanism. [ 17,21–51 ] A proton must be removed from the active site serine hydroxyl group and another proton must be added to the amine leaving group; similar proton transfer must be facilitated during the deacylation mechanism. Although structural information is available on these enzymes and enzyme‐ligand complexes, illustration of the reaction mechanism and activity is still unclear due to multiple possible protonation states for some important residues.…”

Recommending Hartree–Fock Theory with London-Dispersion and Basis-Set-Superposition Corrections for the Optimization or Quantum Refinement of Protein Structures

“…However, we also notice that their rate of acceptance in the general community of "QM users" is low in the fields of protein optimization or quantum refinement. [15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33] This finding is somewhat surprising, for it has been shown that newer QM methods may offer specific advantages over the more established ones. Most of the recent QM and QM/MM optimizations have been carried out with Kohn-Sham density functional theory (DFT [72][73] ) methods in combination with small atomic-orbital (AO) basis sets.…”

“…Most of the recent QM and QM/MM optimizations have been carried out with Kohn-Sham density functional theory (DFT [72][73] ) methods in combination with small atomic-orbital (AO) basis sets. [15][16][17][18][19][20][21][22][23][24][25][26][27][28][29] Indeed, the vast majority of these applications used the BP86 [74][75] and B3LYP [76][77] functionals with basis sets such as the 6-31G* 78 Pople basis or the SV(P) 79 and def2-SV(P) 80 Ahlrichs basis sets. Some authors also used Hartree-Fock (HF) theory instead of DFT.…”