The importance of a critical protonation state and the fate of the catalytic steps in class A beta-lactamases and penicillin-binding proteins

Abstract

Beta-lactamases and penicillin-binding proteins are bacterial enzymes involved in antibiotic resistance to beta-lactam antibiotics and biosynthetic assembly of cell wall, respectively. Members of these large families of enzymes all experience acylation by their respective substrates at an active site serine as the first step in their catalytic activities. A Ser-X-X-Lys sequence motif is seen in all these proteins, and crystal structures demonstrate that the side-chain functions of the serine and lysine are in contact with one another. Three independent methods were used in this report to address the question of the protonation state of this important lysine (Lys-73) in the TEM-1 beta-lactamase from Escherichia coli. These techniques included perturbation of the pK(a) of Lys-73 by the study of the gamma-thialysine-73 variant and the attendant kinetic analyses, investigation of the protonation state by titration of specifically labeled proteins by nuclear magnetic resonance, and by computational treatment using the thermodynamic integration method. All three methods indicated that the pK(a) of Lys-73 of this enzyme is attenuated to 8.0-8.5. It is argued herein that the unique ground-state ion pair of Glu-166 and Lys-73 of class A beta-lactamases has actually raised the pK(a) of the active site lysine to 8.0-8.5 from that of the parental penicillin-binding protein. Whereas we cannot rule out that Glu-166 might activate the active site water, which in turn promotes Ser-70 for the acylation event, such as proposed earlier, we would like to propose as a plausible alternative for the acylation step the possibility that the ion pair would reconfigure to the protonated Glu-166 and unprotonated Lys-73. As such, unprotonated Lys-73 could promote serine for acylation, a process that should be shared among all active-site serine beta-lactamases and penicillin-binding proteins.

Reaction 1

Fig. 1. The pH dependence of k_cat/K_m (A), k_cat (B), K_m (C) of the wild-type (closed circles; left y-axis) and of γ-thialysine (open circles; right y-axis) TEM-1 β-lactamases with mezlocillin

The error bars in the plots represent the standard deviations for each data point.

Fig. 2. ¹³C HSQC spectra for wild-type and K73A mutant TEM-1 β-lactamase, specifically ¹³C labeled at the C(ε) position, and the corresponding ¹H-¹³C(ε) chemical shifts from all lysines for the wild-type protein, obtained from the pH titration

Spectral features are labeled in panel A. The wild-type spectra at pH 6.0 (A) and pH 9.3 (C), and those of the K73A mutant protein at pH 6.0 (B) and pH 9.3 (D) are depicted. Red boxed areas designate the position (or lack thereof) for peak 4, attributed to the Lys-73 signal. Chemical shifts of proton (E) and carbon (F) as a function of pH for all lysine H₂C(ε) resonances. The profile for peak 4 is shown in red. Peak labels are as follows: Peaks 1a (●), 1b (○), 2 ([+]), 3a (▲), 3b (▼), 4 (■), 5 (×), 6 (□), 7 (◆), 8 (⊙),9 (¦), 10 (+), and 11 (\). G, the NMR chemical shifts of the proton and carbon for peak 4 (E and F) were fitted to a single-ionization model (Equation 2).

Fig. 3

Stereo view of the active site of the TEM-1 β-lactamase from the x-ray structure depicted as a Connolly surface.