A last line of defence against "superbugs" are the vancomycin group antibiotics. This review describes the determination of their mode of action, and a mechanism of resistance to them. Remarkably, this mechanism of resistance can be overcome without directly modifying the binding site of the antibiotics for the cell-wall precursors of pathogenic bacteria.

The mode of action of a semisynthetic glycopeptide active against vancomycin-resistant bacteria has been investigated. It is shown that the antibiotic, biphenylchloroeremomycin or LY307599, dimerizes strongly and anchors to membranes. It is hypothesized that these two locating devices, previously identified by us when acting separately, might combine to give enhanced binding at a cell surface. This hypothesis is tested experimentally by showing that glycopeptides can bind cell-wall precursor analogues from resistant bacteria (terminating in -D-lactate) in a similar manner to those from susceptible bacteria (terminating in -D-alanine) and by using model cell surfaces where the benefits of dimerization can be expressed and studied. These model systems use vesicles to represent the cell membrane, to which cell wall analogues are anchored Via a docosanoyl chain, so mimicking the arrangement encountered at the cell surface. Using 1 H NMR spectroscopy, we demonstrate enhanced binding due to dimerization and propose that this enhancement will act cooperatively with membrane anchoring in biphenylchloroeremomycin.

Dimerisation or the use of a membrane anchor enhances the binding of the glycopeptide antibiotics at the surface of a model cell membrane.

High-performance liquid chromatography-solid phase extraction-NMR spectroscopy (HPLC-SPE-NMR) has recently become commercially available and has been evaluated with regard to its applicability in a pharmaceutical environment. The addition of an automated SPE unit to an HPLC-NMR system for peak trapping results in an improved NMR signal-to-noise ratio (S/N) and also has other practical advantages. The trapping efficiency is shown to depend on compound polarity and is highest for compounds eluting late on reversed-phase HPLC systems. Multiple peak trapping further increases the S/N, again with the best results for less polar compounds. For polar compounds, multiple peak trapping resulted in no S/N gain as the amount of material retained on the SPE cartridge was equivalent to that from a single injection. When compared with conventional HPLC-NMR, a S/N gain of up to five-fold could be achieved for some compounds in a single trapping step. A major advantage of the technique is the independence of the chromatographic step from the NMR step, resulting in greater versatility than conventional HPLC-NMR in the HPLC solvents and NMR solvents that can be used. Practical applications from both drug metabolite and drug impurity identification are presented.

The 1H NMR spectra of 24 compounds containing the ester group are given and assigned. These data were used to investigate the effect of the ester group on the 1H chemical shifts in these molecules. These effects were analysed using the CHARGE model, which incorporates the electric field, magnetic anisotropy and steric effects of the functional group for long-range protons together with functions for the calculation of the two- and three-bond effects. The effect of the ester electric field was given by considering the partial atomic charges on the three atoms of the ester group. The anisotropy of the carbonyl group was reproduced with an asymmetric magnetic anisotropy acting at the midpoint of the carbonyl bond with values of Deltachi(parl) and Deltachi(perp) of 10.1 x 10(-30) and -17.1 x 10(-30) cm3 molecule(-1). An aromatic ring current (=0.3 times the benzene ring current) was found to be necessary for pyrone but none for maleic anhydride. This result was confirmed by GIAO calculations. The observed 1H chemical shifts in the above compounds were compared with those calculated by CHARGE and the ab initio GIAO method (B3LYP/6-31G**). For the 24 compounds investigated with 150 1H chemical shifts spanning a range of ca 10 ppm, the CHARGE model gave an excellent r.m.s. error (obs - calc) of <0.1 ppm. The GIAO calculations gave a very reasonable r.m.s. error of ca 0.2 ppm although larger deviations of ca 0.5 ppm were observed for protons near to the electronegative atoms. The accurate predictions of the 1H chemical shifts given by the CHARGE model were used in the conformational analysis of the vinyl esters methyl acrylate and methyl crotonate. An illustration of the use of the CHARGE model in the prediction of the 1H spectrum of a complex organic molecule (benzochromen-6-one) is also given.