Engineering human MEK-1 for structural studies: A case study of combinatorial domain hunting.

Structural biology studies typically require large quantities of pure, soluble protein. Currently the most widely-used method for obtaining such protein involves the use of bioinformatics and experimental methods to design constructs of the target, which are cloned and expressed. Recently an alternative approach has emerged, which involves random fragmentation of the gene of interest and screening for well-expressing fragments. Here we describe the application of one such fragmentation method, combinatorial domain hunting (CDH), to a target which historically was difficult to express, human MEK-1. We show how CDH was used to identify a fragment which covers the kinase domain of MEK-1 and which expresses and crystallizes significantly better than designed expression constructs, and we report the crystal structure of this fragment which explains some of its superior properties. Gene fragmentation methods, such as CDH, thus hold great promise for tackling difficult-to-express target proteins.

IRAK-4 inhibitors. Part II: a structure-based assessment of imidazo[1,2-a]pyridine binding.

A potent IRAK-4 inhibitor was identified through routine project cross screening. The binding mode was inferred using a combination of in silico docking into an IRAK-4 homology model, surrogate crystal structure analysis and chemical analogue SAR.

Structure of an allosteric inhibitor of LFA-1 bound to the I-domain studied by crystallography, NMR, and calorimetry.

LFA-1 (lymphocyte function-associated antigen-1) plays a role in intercellular adhesion and lymphocyte trafficking and activation and is an attractive anti-inflammatory drug target. The alpha-subunit of LFA-1, in common with several other integrins, has an N-terminally inserted domain (I-domain) of approximately 200 amino acids that plays a central role in regulating ligand binding to LFA-1. An additional region, termed the I-domain allosteric site (IDAS), has been identified exclusively within the LFA-1 I-domain and shown to regulate the function of this protein. The IDAS is occupied by small molecule LFA-1 inhibitors when cocrystallized or analyzed by (15)N-(1)H HSQC (heteronuclear single-quantum coherence) NMR (nuclear magnetic resonance) titration experiments. We report here a novel arylthio inhibitor that binds the I-domain with a K(d) of 18.3 nM as determined by isothermal titration calorimetry (ITC). This value is in close agreement with the IC(50) (10.9 nM) derived from a biochemical competition assay (DELFIA) that measures the level of inhibition of binding of whole LFA-1 to its ligand, ICAM-1. Having established the strong affinity of the arylthio inhibitor for the isolated I-domain, we have used a range of techniques to further characterize the binding, including ITC, NMR, and X-ray crystallography. We have first developed an effective ITC binding assay for use with low-solubility inhibitors that avoids the need for ELISA-based assays. In addition, we utilized a fast NMR-based assay for the generation of I-domain-inhibitor models. This is based around the collection of HCCH-TOCSY spectra of LFA-1 in the bound form and the identification of a subset of side chain methyl groups that give chemical shift changes upon binding of LFA-1 inhibitors. This subset was used in two-dimensional (13)C-(15)N and (15)N-filtered and -edited two-dimensional NMR experiments to identify a minimal set of intraligand and ligand-protein NOEs, respectively (nuclear Overhauser enhancements). Models from the NMR data were assessed by comparison to an X-ray crystallographic structure of the complex, confirming that the method correctly predicted the essential features of the bound ligand.

Interactions of mutant and wild-type flap endonucleases with oligonucleotide substrates suggest an alternative model of DNA binding.

Previous structural studies on native T5 5' nuclease, a member of the flap endonuclease family of structure-specific nucleases, demonstrated that this enzyme possesses an unusual helical arch mounted on the enzyme's active site. Based on this structure, the protein's surface charge distribution, and biochemical analyses, a model of DNA binding was proposed in which single-stranded DNA threads through the archway. We investigated the kinetic and substrate-binding characteristics of wild-type and mutant nucleases in relation to the proposed model. Five basic residues R33, K215, K241, R172, and R216, are all implicated in binding branched DNA substrates. All these residues except R172 are involved in binding to duplex DNA carrying a 5' overhang. Replacement of either K215 or R216 with a neutral amino acid did not alter kcat appreciably. However, these mutant nucleases displayed significantly increased values for Kd and Km. A comparison of flap endonuclease binding to pseudoY substrates and duplexes with a single-stranded 5' overhang suggests a better model for 5' nuclease-DNA binding. We propose a major revision to the binding model consistent with these biophysical data.