⁶⁷Zn NMR data collected at 18.8 T/10K for WT LpxC prepared at (a) pH 8.7 and (b) pH 6 where each 10 kHz block had a recycle delay of 1 min for 256 accumulations. Simulations using the parameters from Table 1 are in red and overlaid on the respective experimental data.

⁶⁷Zn NMR data collected at 18.8 T/10K for H265A LpxC prepared at (a) pH 8.7 and (b) pH 6 where each 10 kHz block had 512 accumulations with recycle delays of 30 and 60 s respectively. Identical simulations (in red) using a C_q of 9.55 MHz and an η_q of 0.87 are overlaid on both data sets.

Representation of the starting quantum region for the QM/MM calculations generated with MacPyMol 1.0r2 from PDB structure 1P42.

QM/MM optimized quantum regions of WT LpxC with water bound to the zinc and His265 protonated as (1a) doubly protonated (HIP), (1b) singly protonated at N_ε (HIE), and (1c) singly protonated at N_δ (HID). Water network, including zinc bound water, is represented as ball and stick and hydrogen bonds within 2.0 Å are drawn as red dashed lines.

QM/MM optimized quantum regions of WT LpxC with hydroxide bound to the zinc and His265 protonated as (2a) doubly protonated (HIP), (2b) singly protonated at N_ε (HIE), and (2c) singly protonated at N_δ (HID). Water network, including zinc bound OH, is represented as ball and stick and hydrogen bonds within 2.0 Å are drawn as red dashed lines.

QM/MM optimized quantum regions of H265A LpxC with (a) water bound to the zinc and (b) hydroxide coordinated to the metal ion with the respective predicted quadrupole coupling parameters. Shown in (c) is an expansion of (a) showing the result of a proton transfer of the zinc water to a solvent water which protonates Glu78.

QM/MM optimized quantum regions of WT LpxC with doubly protonated His265, water bound to zinc and a proton added to Glu78 O_ε1 and O_ε2 in (a) and (b) respectively and H265A LpxC with water bound to zinc and (c) Glu78 O_ε1H and (d) Glu78 O_ε2H.

Scheme 1 shows a representation of the residues that have been implicated to participate in the reaction mechanism for LpxC (using the E. coli numbering). This scheme depicts the protonation state of the resting enzyme at neutral pH based on structural9 and biochemical studies11,18 where the mechanism for LpxC is proposed to involve a doubly protonated His265 residue functioning as either a general acid or an electrostatic catalyst, and the side chain of Glu78 acting as a general base in the reaction. Mutagenesis experiments have identified side chains that are important for catalytic activity, including Glu78, Thr191, Lys239, Asp246, and His265.7,9-11 The H265A mutation in Aquifex aeolicus (Aa) LpxC reduces catalytic activity (k_cat/K_M) by ∼170-fold, while the E78A/H265A double mutation further decreases activity by an additional order of magnitude (1700-fold reduced).11 The LpxC-catalyzed reaction exhibits a bell-shaped dependence on pH with two apparent pK_a values of 5.8 and 7.9 for the wild type (WT) Aa-LpxC at 60 °C (the values shift to 6.2 and 9.2 for the E. coli enzyme at 30 °C).11,18 Results from mutagenesis experiments suggest that the pK_a1 value observed in activity measurements reflects ionization of Glu78.11,18 Based on its value, pK_a2 has been proposed to reflect ionization of His265 or the zinc-water. The value of pK_a2 varies with the identity of the active site metal ion (Zn, Ni, Co) consistent with pK_a2 reflecting ionization of the metal-water.11 Product affinity measurements suggest that the pK_a values of Glu78 and His265 (in Ec-LpxC) are 6.5 ± 0.1 and 7.4 ± 0.1, respectively.19 Furthermore, the pH dependence of NMR chemical shifts (imidazole; ¹H/¹³C HSQC) demonstrated that the pK_a value of the His265 residue is 7.6 ± 0.1 at 50°C in a mutant of Aa-LpxC (the peaks in the NMR spectrum were not resolved enough to make the determination in WT LpxC, therefore a variant containing 7 mutations, including H200Y, was used for these experiments).12 These measurements suggest that the pK_a of His265 is lower than the value of pK_a2 observed in the pH-rate profile.