QM/MM simulations identify the determinants of catalytic activity differences between type II dehydroquinase enzymes† †Electronic supplementary information (ESI) available: Fig. S1–S5, Tables S1–S3 and extra details on umbrella sampling simulations. See DOI: 10.1039/c8ob00066b

Multiscale simulations pinpoint specific interactions responsible for differences in stabilization of key reacting species in two recognized targets for antibiotic development.

. Free energy profiles obtained using Umbrella Sampling simulations at the SCC-DFTB/ff03 level for the whole reaction for both enzymes [MtDHQ2 (blue), HpDHQ2 (magenta)]. For 1 st step in HpDHQ2 and 3 rd step in both enzymes, minimum free energy paths extracted from two-dimensional free energy surfaces are shown. S3 Figure S2. Detail view of the salt bridge between Arg20 and Asp18 in HpDHQ2 enzyme observed during MD simulation studies. Hydrogen bonding interactions involving residues 18-21 are also highlighted. Relevant side chain residues are shown and labeled. S4 Figure S3. Comparison of reaction energies for the formation of the enolate 2 from substrate 1 with phenolate as model for Tyr24/Tyr22 as the base. Energies were obtained by single point calculations in vacuum using several levels of theory. SCC-DFTB provides energy values for the reaction that are about 6−8 kcal mol −1 higher than more accurate methods such as MP2, B3LYP or MPW1K.  Figure S5. Detailed view of TS1 (A) and enolate intermediate 2 (B) for the 2 nd step of the HpDHQ2 mechanism. Geometries were taken from the potential energy surface (SCC-DFTB/ff03 QM/MM level, see Figure S4). Relevant residues and water molecules are shown and labeled. Key hydrogen bonding interactions and bonds broken/formed are indicated as red dashed lines. Note how the catalytic water molecule W2 interacts by hydrogen bonding with the neutral Asp89´ in both TS1 and enolate intermediate 2. Figure S6. Variation of the distance between the oxygen atom of the water molecule W2 and O3 atom in the enolate intermediate 2 (A) or the NH2 atom of Arg109 (B) during a 50 ps QM/MM simulation of the HpDHQ2/2 enzyme complex. Note how W2 (from QM region) is replaced by another water molecule from the bulk solvent (from MM region) during the simulation. The distances were calculated considering the position of W2 observed at the beginning of the simulation (blue trace) and of any water molecule that replaces W2 during the simulation (magenta trace). . All energies are relative to the energy for the substrate 1 (i.e. equivalent energy calculated for the substrate geometry was set to 0 kcal mol −1 ). b Energy contribution for W2 was calculated by subtracting the energy of (a) from (b). c Energy contribution for Asp88´ was calculated by subtracting the energy of (b) from (c). d Energy contribution for Arg17 was calculated by subtracting the energy of (b) from (d). e Energy contribution for W1 was calculated by subtracting the energy of (b) from (e). Note that positive values indicate a higher stabilization effect on 2 relative to 1, while negative values indicate the opposite.  The geometries for the studied species were extracted from the potential energy surface of the 2 nd step of the conversion catalysed by the H. pylori enzyme ( Figure S4). Energy values were calculated in vacuum using MPW1K/6-31+G(d,p) as a difference of the energy for both species together minus the sum of the energies for each one alone. b Geometries for the hydrogen bond were optimized at the same level of theory as before. Note how the hydrogen bond between a water molecule and phenolate is stronger than with the corresponding neutral phenol. A smaller difference was found for the respective interaction with an enolate vs a ketone group.

Reaction coordinates for Umbrella Samplings simulations with restraints (A) MtDHQ2 model
The reaction coordinates and the restraints included in each step are summarized in Table S4.
Because the reaction mechanism was evaluated in the backwards direction, starting from the product complex, the reaction steps are described below in that order (third, second and first step).
(i) For the third step, two reaction coordinates were used: r 1a that is the distance between C1 atom in 3 and the O atom of the manually docked product water molecule and r 1b represents the deprotonation of the water molecule by His101 residue. A 1D Umbrella Sampling simulation along reaction coordinate r 1a decreasing distance from 3.5 to 1.5 Å was initially performed, however only the results of the WHAM analysis between 2.3 and 3.5 Å were considered due to poor overlapping of the frecuency histograms of the reaction coordinate in the windows 1.7 to 2.0 Å. In order to achieve a better description of the formation of the C1−O bond as well as the water deprotonation, a 2D Umbrella Sampling was carried out between the distance values of 1.5 and 2.3 for r 1a and 1.0 to −1.0 Å for r 1b using the restraints included in Table S4. Restraint M30b was used to prevent the collapse of the active site caused by the attraction between Arg108 and Glu20 (it was previously used during equilibration (ii) For the second step, a single reaction coordinate (r 2 ), which represents the proton transfer from Tyr24 to C2 in 2, was used and evaluated from 1.1 to −1.0 Å using the restraints indicated in Table   S4. Restraints A05 and M06 were employed to avoid the protonation of the catalytic tyrosinate by Arg19 and the subsequent protonation of the resulting neutral guanidinium by Asp89'. Restraints M04 and M05 were used to ensure the correct binding of C1 region of the natural substrate. The final structure of the window for −0.8 Å was selected as a suitable starting point for the 1 st step.
(iii) For the first step, two reaction coordinates were employed: r 3a that is a linear combination of distances to break the Glu20−Asp89ʼ interaction and r 3b that corresponds to the deprotonation of neutral Asp89' by the catalytic tyrosinate. Using the restraints indicated in Table S4, reaction coordinate r 3a from −1.9 to 1.8 Å was evaluated. The final structure of the window for −0.1 Å was selected as a suitable starting point for r 3b . Reaction coordinate r 3b from −0.6 to 1.0 Å was then S12 followed. Restraints M12 and M09 were applied to prevent the regeneration of the enolate intermediate 2 and to avoid the formation of an enol intermediate, respectively. Restraints M07 and M08 were employed to keep Arg19 in the active site. Restraint M10 was used to prevent displacement of the water molecule W1 from its binding pocket. Finally, restraints M11a and M11b were included to ensure that Glu20 was not expelled from the active site.  Figure S7 for more details. b The type of restraint, the force constant and the atoms involved are indicated in Table   S6. Figure S7. Schematic representation of the atom names used in the definition of the reaction coordinates (Tables S4-S5) and restraints (Table S6). S13

(B) HpDHQ2 model
The reaction coordinates and the restraints included in each step are summarized in Table S5.
(i) For the third step, a similar procedure as for the MtDHQ2 model was used. Only the particular differences are explained. Thus, the r 1b reaction coordinate was evaluated from 1.0 to −0.6 Å using the restraints indicated in Table S5. The final structure of the window for 1.5 Å and −0.6 Å was selected as a suitable starting point for the 2 nd step. Restraints A02h and H03 were used to keep the Tyr24 in the active site and to avoid formation of an enol instead of enolate 2, respectively.
Restraints H01a and H02a were applied to avoid diffusion of the water molecule W2 into the bulk solvent. Restraints H04, H05 and H06 were introduced to maintain Arg17 in the active site.
(ii) For the second step, as for MtDHQ2 model a single reaction coordinate r 2 was employed. r 2 was evaluated from 1.5 to −1.2 Å using the restraints shown in Table S5. The final structure of the window for −0.9 Å was selected as a suitable starting point for the 1 st step. Restraint A05 was employed needed to circumvent the formation of an enol instead of the enolate 2. Restraints H01b and H02b were applied to avoid diffusion of the water W2 into the bulk solvent.
(iii) For the first step, two reaction coordinates were employed: r 3a that is the deprotonation of the catalytic water molecule W2 by the catalytic tyrosinate and r 3b that involves the abstraction by W2 of the proton from the neutral Asp88'. In both cases the reaction coordinate varied from 0.8 to −0.8 Å and no additional restraints were applied. --a See Figure S7 for more details. b The type of restraint, the force constant and the atoms involved are indicated in Table   S6.   Figure S7 for more details. b Distance in Å, angles in degrees. c Units (kcal mol −1 Å −2 ) or (kcal mol −1 rad −2 ). d LCOD is Linear Combination of Distances.