The AcrAB-TolC system exports a wide variety of drugs and toxic compounds, and confers intrinsic drug tolerance on Escherichia coli. The crystal structures suggested that AcrB and TolC directly dock with each other. However, biochemical and biophysical evidence of their interaction has been contradictory until recently. In this study, we examine the interaction sites by means of in vivo disulfide cross-linking between cysteine residues introduced by site-directed mutagenesis at the tops of the vertical hairpins of AcrB and the bottoms of the coiled coils of polyhistidine-tagged TolC molecules, which are structurally predicted docking sites. The AcrB-TolC complex formed through disulfide cross-linking was detected when a specific pair of mutants was coexpressed in E. coli. Our observations suggested that the AcrB-TolC complex may be formed through a two-step mechanism via transient tip-to-tip interaction of AcrB and TolC. The cross-linking was not affected by AcrA, the substrate, or a putative proton coupling site mutation.

A new toxin, named HsTX1, has been identified in the venom of Heterometrus spinnifer (Scorpionidae), on the basis of its ability to block the rat Kv1.3 channels expressed in Xenopus oocytes. HsTX1 has been purified and characterized as a 34-residue peptide reticulated by four disulphide bridges. HsTX1 shares 53% and 59% sequence identity with Pandinus imperator toxin1 (Pi1) and maurotoxin, two recently isolated four-disulphide-bridged toxins, whereas it is only 32-47% identical with the other scorpion K+ channel toxins, reticulated by three disulphide bridges. The amidated and carboxylated forms of HsTX1 were synthesized chemically, and identity between the natural and the synthetic amidated peptides was proved by mass spectrometry, co-elution on C18 HPLC and blocking activity on the rat Kv1.3 channels. The disulphide bridge pattern was studied by (1) limited reduction-alkylation at acidic pH and (2) enzymic cleavage on an immobilized trypsin cartridge, both followed by mass and sequence analyses. Three of the disulphide bonds are connected as in the three-disulphide-bridged scorpion toxins, and the two extra half-cystine residues of HsTX1 are cross-linked, as in Pi1. These results, together with those of CD analysis, suggest that HsTX1 probably adopts the same general folding as all scorpion K+ channel toxins. HsTX1 is a potent inhibitor of the rat Kv1.3 channels (IC50 approx. 12 pM). HsTX1 does not compete with 125I-apamin for binding to its receptor site on rat brain synaptosomal membranes, but competes efficiently with 125I-kaliotoxin for binding to the voltage-gated K+ channels on the same preparation (IC50 approx. 1 pM).

13C-NMR spectroscopy was used to estimate the p K a values for the Tyr(150) (Y150) residue in wild-type and mutant class C beta-lactamases. The tyrosine residues of the wild-type and mutant lactamases were replaced with (13)C-labelled L-tyrosine ([ phenol -4-(13)C]tyrosine) in order to observe the tyrosine residues selectively. Spectra of the wild-type and K67C mutant (Lys(67)-->Cys) enzyme were compared with the Y150C mutant lactamase spectra to identify the signal originating from Tyr(150). Titration experiments showed that the chemical shift of the Tyr(150) resonance in the wild-type enzyme is almost invariant in a range of 0.1 p.p.m. up to pH 11 and showed that the p K (a) of this residue is well above 11 in the substrate-free form. According to solvent accessibility calculations on X-ray-derived structures, the phenolic oxygen of Tyr(150), which is near the amino groups of Lys(315) and Lys(67), appears to have low solvent accessibility. These results suggest that, in the native enzyme, Tyr(150) in class C beta-lactamase of Citrobacter freundii GN346 is protonated and that when Tyr(150) loses a proton, a proton from Lys(67) would replace it. Consequently, Tyr(150) would be protonated during the entire titration.