Improved sensitivity in assays for binding of novel beta-lactam antibiotics to penicillin-binding proteins of Escherichia coli.

Tigemonam and temocillin, but not aztreonam, bound to penicillin-binding proteins (PBPs) 1a and 3 of Escherichia coli with apparent improved affinity when challenged with benzylpenicillin at lowered temperatures. Half times for deacylation of the tigemonam-PBP complexes were shorter than were those of the corresponding aztreonam-PBP complexes. The implications of the routine testing of PBP affinities are discussed. Images

Interactions of 3-lactam antibiotics with specific receptor proteins in bacteria are examined critically in many antibiotic development programs. Early studies by Spratt (11) were instrumental in establishing assay procedures by which characteristics of these penicillin-binding proteins (PBPs) could be defined, especially with respect to their affinities for penicillins and cephalosporins. However, the recent appearance of structurally novel r-lactam antibiotics has required a reexamination of the nature of these interactions and a reevaluation of assay procedures used routinely for the study of these interactions.
(This work was presented in part previously [K. Bush, S. A. Smith, and D. P. Bonner, Program Abstr. 26th Intersci. Conf. Antimicrob. Agents Chemother., abstr. no. 855,1986].) Tigemonam (Fig. 1), a new oral monobactam, exhibits excellent activity' against aerobic gram-negative bacteria (3,13) and is very similar to aztreonam in its microbiological properties. As with typica,l -lactam antibiotics, the mode of action for killing appears to be inhibition of cell wall synthesis by binding to PBPs. In members of the family Enterobacteriaceae, filamentation is the primary morphological response to monobactams such as aztreonam, tigemonam, and carumona,m when tested at the MIC, consistent with the observation that PBP 3 has been inhibited (5, 7, 11). Although MICs are identical for the three monobactams, aztreonam and carumonam bind to PBP 3 of Escherichia c oli at concentrations lower than 0.1 pLg/ml (2. 7), whereas significant binding of tigemonam to PBP 3 was observed at a concentration at least 10-fold higher. Similar observations have been reported for temocillin and SQ 81,377 ( Fig. 1), for which binding to PBP 3 also was not consistent with the observed biological activity (5, 8).
Because of these anomalies, we investigated the possibility that the standard conditions used to test the binding of 3-lactams to PBPs as described by Spratt (11) may result in an underestimation of actual affinities for PBPs. Based on the studies with temocillin by Labia et al. (8), assay conditions were developed in this laboratory that gave results * Corresponding author. more consistent with the observed biological activity of certain antibiotics.
Solubilized membranes were prepared from E. coli SC8294, a strain that produces no detectable 3-lactamase activity. Binding of 3-lactarn antibiotics was determined by procedures described previously (4, 11). Modifications of incubation conditions were developed from the studies of Labia et al. (8). Under standard assay conditions (12), mem'branes were incubated with monobactam for 10 min, followed by a 10-min incubation with 10 nmol (0.2 mM) of [14C]benzylpenicillin, all at 30°C. When altered conditions were used, the second incubation temperature was lowered to 0°C. Reaction mixtures were'analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (9), followed by fluorography (1, 10). X-ray films were analyzed with a laser densit'ometer (model 2212; LKB Instruments, Inc.).
For determination of deacylation rates, incubation times with ['4C]benzylpenicillin were varied from 2 to 30 min. B'inding of 13-lactam antibiotics to PBPs is determined after an antibiotic is equilibrated with solubilized membranes and then challenged with saturating levels of [14C]benzylpenicillin. Because this method is based on the competition between an antibiotic and benzylpenicillin for a PBP, transiently stable PBP-antibiotic complexes may not be detected; these complexes c'ould dissociate within the time of the assay and reform as the ['4C]benzylpenicillin-PBP complex. Therefore, it is necessary to select assay conditions that favor the antibiotic-PBP complex over the benzylpenicillin-PBP complex. In our altered assay conditions, a lower incubation temperature was used for the challenge with [14C]benzylpenicillin, thus allowing fewer kinetic events to occur with respect to complex degradation or formation.
Tigemonam appeared to bind much more tightly to PBPs la, 3, and 4 when lower incubation temperatures were used for the benzylpenicillin challenge (Fig. 2). The shifts in concentrations for 50% inhibition of binding of tigemonam and SQ 81,377 to PBPs (Table 1) were similar to observations reported previously for temocillin (8), which was included in our study as a reference. However, binding of aztreonam and the related monobactam, SQ 81,402, was not significantly affected by alterations in assay conditions, indicating differences in behavior within classes of 13-lactam antibiotics.
Because the effect of lowered temperature in the ['4C]benzylpenicillin challenge was quite marked for tigemonam, temocillin, and SQ 81,377, kinetic interactions between these antibiotics and various PBPs must havebeen affected.
In contrast, deacylation half times for aztreonam were >30 min for each of these PBPs. Acylation of PBP 3 occurred rapidly for both antibiotics, with completely acylated PBP 3 complexes detected after a 30-s incubation of monobactam with membranes. Therefore, the improved binding observed