Drug efflux proteins are widespread amongst microorganisms, including pathogens. They can contribute to both natural insensitivity to antibiotics and to emerging antibiotic resistance and so are potential targets for the development of new antibacterial drugs. The design of such drugs would be greatly facilitated by knowledge of the structures of these transport proteins, which are poorly understood, because of the difficulties of obtaining crystals of quality. We describe a structural genomics approach for the amplified expression, purification and characterisation of prokaryotic drug efflux proteins of the 'Major Facilitator Superfamily' (MFS) of transport proteins from Helicobacter pylori, Staphylococcus aureus, Escherichia coli, Enterococcus faecalis, Bacillus subtilis, Brucella melitensis, Campylobacter jejuni, Neisseria meningitides and Streptomyces coelicolor. The H. pylori putative drug resistance protein, HP1092, and the S. aureus QacA proteins are used as detailed examples. This strategy is an important step towards reproducible production of transport proteins for the screening of drug binding and for optimisation of crystallisation conditions to enable subsequent structure determination.

Receptor-dependent productive uptake of GLP1-conjugated antisense oligonucleotides occurs selectively in pancreatic β-cells.

The membrane topology of proton-pumping nicotinamide-nucleotide transhydrogenase from Escherichia coli was determined by site-specific chemical labeling. A His-tagged cysteine-free transhydrogenase was used to introduce unique cysteines in positions corresponding to potential membrane loops. The cysteines were reacted with fluorescent reagents, fluorescein 5-maleimide or 2-[(4 -maleimidyl)anilino]naphthalene-6-sulfonic acid, in both intact cells and inside-out vesicles. Labeled transhydrogenase was purified with a small-scale procedure using a metal affinity resin, and the amount of labeling was measured as fluorescence on UV-illuminated acrylamide gels. The difference in labeling between intact cells and inside-out vesicles was used to discriminate between a periplasmic and a cytosolic location of the residues. The membrane region was found to be composed of 13 helices (four in the ␣-subunit and nine in the ␤-subunit), with the C terminus of the ␣-subunit and the N terminus of the ␤-subunit facing the cytosolic and periplasmic sides, respectively. These results differ from previous models with regard to both number of helices and the relative location and orientation of certain helices. This study constitutes the first in which all transmembrane segments of transhydrogenase have been experimentally determined and provides an explanation for the different topologies of the mitochondrial and E. coli transhydrogenases.Nicotinamide-nucleotide transhydrogenases are integral membrane proteins that catalyze the reduction of NADP ϩ by NADH with a concomitant translocation of protons across the membrane, e.g. from the periplasm or intermembrane space to the cytosol or matrix in bacteria and mitochondria, respectively, according to the following:Transhydrogenases are conformationally driven proton pumps in which binding and release of NADP(H) are thought to cause the structural changes allowing hydride transfer as well as proton translocation to occur (1). It is generally concluded that the H ϩ /H Ϫ ratio of the transhydrogenase reaction is 1 (2-4). At present, transhydrogenases from 12 different species have been cloned and/or sequenced that show an overall amino acid sequence identity of 23%. All transhydrogenases possess the same structural organization, regardless of primary sequence arrangement, with three major domains: two hydrophilic domains (domains I and III) comprising a 400-residue-long NAD(H)-binding domain and a 200-residue-long NADP(H)-binding domain, respectively, and domain II, containing a 360-residue-long hydrophobic domain that constitutes the membrane-spanning part of the enzyme (for reviews, see Refs. 4 and 5). The hydrophilic domains have been shown to protrude into the cytosol or matrix in bacteria and mitochondria, respectively (6, 7).Escherichia coli transhydrogenase consists of two subunits, ␣ (510 residues) and ␤ (464 residues), assembled as an ␣ 2 ␤ 2 -tetramer. The membrane domain is composed of the 110 Cterminal residues of the ␣-subunit and the 260 N-terminal residues of the ␤-subunit and ...

Nicotinamide nucleotide transhydrogenase from Escherichia coli was investigated with respect to the roles of its cysteine residues. This enzyme contains seven cysteines, of which five are located in the alpha subunit and two are in the beta subunit. All cysteines were replaced by site-directed mutagenesis. The final construct (alphaC292T, alphaC339T, alphaC395S, alphaC397T, alphaC435S, betaC147S, betaC260S) was inserted normally in the membrane and underwent the normal NADPH-dependent conformational change of the beta subunit to a trypsin-sensitive state. Reduction of NADP+ by NADH driven by ATP hydrolysis or respiration was between 32% and 65% of the corresponding wild-type activities. Likewise, the catalytic and proton pumping activities of the purified cysteine-free enzyme were at least 30% of the purified wild-type enzyme activities. The H+/H- ratio for both enzymes was 0.5, although the cysteine-free enzyme appeared to be more stable than the wild-type enzyme in proteoliposomes. No bound NADP(H) was detected in the enzymes. Modification of transhydrogenase by diethyl pyrocarbonate and the subsequent inhibition of the enzyme were unaffected by removal of the cysteines, indicating a lack of involvement of cysteines in this process. Replacement of cysteine residues in the alpha subunit resulted in no or little change in activity, suggesting that the basis for the decreased activity was probably the modification of the conserved beta-subunit residue Cys-260 or (less likely) the non-conserved beta-subunit residue Cys-147. It is concluded that the cysteine-free transhydrogenase is structurally and mechanistically very similar to the wild-type enzyme, with minor modifications of the properties of the NADP(H) site, possibly mediated by the betaC260S mutation. The cysteine-free construct will be a valuable tool for studying structure-function relationships of transhydrogenases.

Scoring potency is a main challenge for structure based drug design. Inductive effects of subtle variations in the ligand are not possible to accurately predict by classical computational chemistry methods. In this study, the problem of predicting potency of ligands with electronic variations participating in key interactions with the protein was addressed. The potency was predicted for a large set of cyclic amidine and guanidine cores extracted from β-secretase (BACE-1) inhibitors. All cores were of similar size and had equal interaction motifs but were diverse with respect to electronic substitutions. A density functional theory approach, in combination with a representation of the active site of a protein using only key residues, was shown to be predictive. This computational approach was used to guide and support drug design, within the time frame of a normal drug discovery design cycle.