A major factor in the folding of proteins is the burying of hydrophobic side chains. A specific example is the packing of alpha-helices on beta-sheets by interdigitation of nonpolar side chains. The contributions of these interactions to the energetics of protein stability may be measured by simple protein engineering experiments. We have used site-directed mutagenesis to truncate hydrophobic side chains at an alpha-helix/beta-sheet interface in the small ribonuclease from Bacillus amyloliquefaciens (barnase). The decreases in stability of the mutant proteins were measured by their susceptibility to urea denaturation. Creation of a cavity the size of a -CH2-group destabilizes the enzyme by 1.1 kcal mol-1, and a cavity the size of three such groups by 4.0 kcal mol-1.

The alpha-helix in proteins has a dipole moment resulting from the alignment of dipoles of the peptide bond which can perturb the pKas of ionizing groups. One of the two histidine residues (His18) in barnase, the small ribonuclease from Bacillus amyloliquefaciens, is located at the negatively charged end (C-terminal) of an alpha-helix. From NMR titrations of wild-type and engineered mutants we find that the pKa of His18 is 7.9 in wild-type enzyme, 1.6 units above the value in the urea-denatured enzyme and in model peptides. This implies that there is a favourable interaction between the protonated form of His18 and the alpha-helix that should stabilize the native structure at neutral pH by 2.1 kcal mol-1. Denaturation at various values of pH of wild-type and muant enzymes engineered at position 18 shows that this is so. The increase in stability of the enzyme as the pH changes from 8.5 to 6.3 is attributable to this interaction, and the pH-stability curve fits pKa values for His18 in native and urea-denatured enzymes that are consistent with the NMR data.

No abstract

Hepatitis C virus (HCV) protease NS3 and its protein activator NS4A participate in the processing of the viral polyprotein into its constituent nonstructural proteins. The NS3/4A complex is thus an attractive target for antiviral therapy against HCV. We expressed the full-length NS3 and NS4A in insect cells as a soluble fusion protein with an N-terminal polyhistidine tag and purified the two proteins to homogeneity. Cleavage at the junction between HisNS3 and NS4A occurs during expression, producing a noncovalent complex between HisNS3 and NS4A with a subnanomolar dissociation constant. We purified the HisNS3/4A complex by detergent extraction of cell lysate and by metal chelate chromatography. We removed the His tag by thrombin cleavage and then further purified the complex by gel filtration. The purified NS3/4A complex is active in a protease assay using a synthetic peptide substrate derived from the NS5A-NS5B junction, with kcat/K(m) of 3700 (+/- 600) M-1 s-1, an order of magnitude above those previously reported for NS3 expressed by other strategies. This high protease activity implies that the full-length sequences of NS3 and NS4A are required for optimal activity of the NS3 protease domain. We examined the dependence of the NS3/4A protease activity on buffer conditions, temperature, and the presence of detergents. We find that, under most conditions, NS3 protease activity is dependent on the aggregation state of the NS3/4A complex. The monodisperse, soluble form of the NS3/4A complex is associated with the highest protease activity.

Two-dimensional (2D) double-quantum-filtered correlation spectroscopy (DQF-COSY), total correlation spectroscopy (TOCSY), nuclear Overhauser effect spectroscopy (NOESY), and rotating-frame NOESY (ROESY) spectra were used to assign essentially all the protons in a 1:1 complex of Lactobacillus casei dihydrofolate reductase formed with an analogue of the antibacterial drug brodimoprim [2,4-diamino-5-(3',5'-dimethoxy-4'-bromobenzyl)pyrimidine]. The analogue has a 4,6-dicarboxylic acid side chain substituted on the 3'-O position designed to interact with the Arg 57 and His 28 residues in L. casei dihydrofolate reductase; it binds a factor of 10(3) more tightly to the enzyme than does the parent compound. Thirty-eight intermolecular and 11 intramolecular NOEs were measured involving the bound brodimoprim-4,6-dicarboxylic acid analogue. These provided the distance constraints used in conjunction with an energy minimization and simulated annealing protocol (using Discover from Biosym Ltd.) to dock the brodimoprim analogue into dihydrofolate reductase. In calculations where side chains and backbone fragments for binding-site residues were allowed flexibility, 90% of the 40 calculated structures had reasonable covalent geometry and none of them had NOE distance violations of greater than 0.36 A. The conformations of the aromatic rings in the bound ligand were well-defined in all the structures, with torsion angles tau 1 = -153 degrees +/- 4 degrees (C4-C5-C7-C1') and tau 2 = 53 degrees +/- 4 degrees (C5-C7-C1'-C2'): the aromatic rings of the ligand occupied essentially the same space in all the calculated structures (root mean square deviation value 1.83 A). Inclusion of the electrostatic interactions into the energy minimizations indicated that structures in which the 4,6-dicarboxylate group of the ligand interacts with the side chains of Arg 57 and His 28 are of low energy. Significant differences in side-chain and backbone conformations were detected between binding-site residues in the enzyme complexes with the brodimorpim analogue and methotrexate.