7:30 am Registration

8:30 Opening Remarks
Margit Eder, Conference Director, Cambridge Healthtech Institute

8:35 Chairperson's Opening Remarks

9:25 Engineering Constitutively Active PXR Ligand Binding Domains for Inverse Structure-Based Drug Design
Wenyan Wang, Ph.D., Principal Scientist, Protein Engineering and Biochemistry Group, Structural Chemistry Department, Schering-Plough Research Institute
The nuclear xenobiotic receptor PXR is a ligand inducible transcriptional factor regulating drug-metabolizing enzymes and transporters, therefore potentially mediating dangerous drug-drug interactions. Upon ligand binding, the C-terminal ligand binding domain (LBD) of PXR complexes with a cellular coactivator (such as steroid receptor coactivator, or SRC1) and triggers downstream activation. In an effort to develop stable and effective PXR construct to design out PXR activity, several recombinant PXR-SRC1 complexes were engineered and evaluated for their stability and activity. A recombinant, polyhistidine-tagged PXR/SRC1 complex derived from a dual plasmid expression system reported in the literature [1-4] contained less than optimum PXR to SRC-1 ratio and was readily precipitated during purification and storage. A translationally-coupled, bicistronic expression system increased the expression level by 5-fold, improved protein solubility, and yielded PXR and SRC1 at ~ 1:1 ratio. However, the purified protein complex still required additional SRC peptide for stabilization. A single polypeptide chain encompassing PXR and a SRC-1 peptide connected with a peptidyl linker (sc-PXR-SRCp) was designed to form an intramolecular complex. This engineered protein (sc-PXR-SRCp) was expressed as a soluble protein which required no additional SRC-1 peptide for stabilization and showed the expected affinity to PXR ligands. The X-ray crystal structures of the protein, in the presence and absence of SR12813, have been determined to high resolution. In addition, a circular dichroism ­ based binding assay was developed to allow quick evaluation of small molecules for PXR liability, making this tethered protein a convenient and effective reagent for structure-based inverse drug design.

9:55 Coffee Break, Exhibit & Poster Viewing

10:30 Structural Characterization and Validation of a Novel Recognition Site in the C-Jun NH2 (JNK) Family of Protein Kinases
Celerino Abad-Zapatero, Ph.D., Associate Research Fellow, Protein Crystallography Laboratory, Abbott Laboratories
Using a combination of affinity selection, NMR and X-ray crystallography micromolar bi-aryl tetrazole leads, binding to a novel non-ATP competitive site, have been identified for the c-jun NH2 family of protein kinases (JNK kinases). In the JNK1-1 isoform, the compounds bind on a ledge of the protein surface exposed by the MAP insertion present in the CMGC family of protein kinases and distant from the active site and from the common docking (CD) site. The binding pocket and residues in its vicinity have been suggested to be important for the bipartite mechanism of ERK2 recognition by regulators and substrates and available data suggests that it is probably shared by all MAP kinases. The novel site is also part of a non-crystallographic dimer interface characterized in the crystallographic structure and probably extant in solution. This type of association is consistent with existing data for ERK2 and might be analogous for the JNK1/MKK7 activation. This novel class of compounds inhibits the activation of JNK1 by the upstream MKK7 kinase and certain members of this class also inhibit the phosphorylation of c-jun in cell assays. Developing more potent and selective analogs by structure-based drug design methods directed towards the JNKs or other MAP kinases might provide new biological insights and could be of therapeutic value.

11:00 Technology Spotlight (Sponsorship Available)

2:00 Chairperson's Opening Remarks

2:05 Roundtable Summary Presentations from Morning Program

2:30 Applied Chemogenomics: (Machine) Learning from the Past
Jeremy L. Jenkins, Ph.D., Research Investigator, Lead Discovery Center, Novartis Institutes for BioMedical Research Inc.
Despite our best efforts to rationally design drugs for a single target, most compounds have selectivity issues or undesirable off-target effects. However, given the vast institutional knowledge of structure-activity relationships, statistical modeling across thousands of target classes can give hints as to the probable targets, off-targets, and side effects of small molecules. Here we show case studies of using multi-target modeling to predict both primary and off-targets of orphan ligands or entire hitlists from cell-based screens, and to map this knowledge usefully onto compound structures. Further, a global "phylochemical" relationship between targets and adverse drug reactions is assessed.