Biologic-like In Vivo Efficacy with Small Molecule Inhibitors of TNFα Identified Using Scaffold Hopping and Structure-Based Drug Design Approaches.

Scaffold hopping and structure-based drug design were employed to identify substituted 4-aminoquinolines and 4-aminonaphthyridines as potent, small molecule inhibitors of tumor necrosis factor alpha (TNFα). Structure-activity relationships in both the quinoline and naphthyridine series leading to the identification of compound 42 with excellent potency and pharmacokinetic profile are discussed. X-ray co-crystal structure analysis and ultracentrifugation experiments clearly demonstrate that these inhibitors distort the TNFα trimer upon binding, leading to aberrant signaling when the trimer binds to TNF receptor 1 (TNFR1). Pharmacokinetic-pharmacodynamic activity of compound 42 in a TNF-induced IL-6 mouse model and in vivo activity in a collagen antibody-induced arthritis model, where it showed biologic-like in vivo efficacy, will be discussed.

Small molecule receptor protein tyrosine phosphatase γ (RPTPγ) ligands that inhibit phosphatase activity via perturbation of the tryptophan-proline-aspartate (WPD) loop.

Protein tyrosine phosphatases (PTPs) catalyze the dephosphorylation of tyrosine residues, a process that involves a conserved tryptophan-proline-aspartate (WPD) loop in catalysis. In previously determined structures of PTPs, the WPD-loop has been observed in either an "open" conformation or a "closed" conformation. In the current work, X-ray structures of the catalytic domain of receptor-like protein tyrosine phosphatase γ (RPTPγ) revealed a ligand-induced "superopen" conformation not previously reported for PTPs. In the superopen conformation, the ligand acts as an apparent competitive inhibitor and binds in a small hydrophobic pocket adjacent to, but distinct from, the active site. In the open and closed WPD-loop conformations of RPTPγ, the side chain of Trp1026 partially occupies this pocket. In the superopen conformation, Trp1026 is displaced allowing a 3,4-dichlorobenzyl substituent to occupy this site. The bound ligand prevents closure of the WPD-loop over the active site and disrupts the catalytic cycle of the enzyme.

Aroylguanidine-based factor Xa inhibitors: the discovery of BMS-344577.

We report the design and synthesis of a novel class of N,N'-disubstituted aroylguanidine-based lactam derivatives as potent and orally active FXa inhibitors. The structure-activity relationships (SAR) investigation led to the discovery of the nicotinoyl guanidine 22 as a potent FXa inhibitor (FXa IC(50)=4 nM, EC(2xPT)=7 microM). However, the potent CYP3A4 inhibition activity (IC(50)=0.3 microM) of 22 precluded its further development. Detailed analysis of the X-ray crystal structure of compound 22 bound to FXa indicated that the substituent at the 6-position of the nicotinoyl group of 22 would be solvent-exposed, suggesting that efforts to attenuate the unwanted CYP activity could focus at this position without affecting FXa potency significantly. Further SAR studies on the 6-substituted nicotinoyl guanidines resulted in the discovery of 6-(dimethylcarbamoyl) nicotinoyl guanidine 36 (BMS-344577, IC(50)=9 nM, EC(2xPT)=2.5 microM), which was found to be a selective, orally efficacious FXa inhibitor with an excellent in vitro liability profile, favorable pharmacokinetics and pharmacodynamics in animal models.

Assessing compound binding to the Eg5 motor domain using a thermal shift assay.

Eg5 is a kinesin whose inhibition leads to cycle arrest during mitosis, making it a potential therapeutic target in cancers. Circular dichroism and isothermal titration calorimetry of our pyrrolotriazine-4-one series of inhibitors with Eg5 motor domain revealed enhanced binding in the presence of adenosine 5'-diphosphate (ADP). Using this information, we studied the interaction of this series with ADP-Eg5 complexes using a thermal shift assay. We measured up to a 7 degrees C increase in the thermal melting (T(m)) of Eg5 for an inhibitor that produced IC(50) values of 60 and 130 nM in microtubule-dependent adenosine triphosphatase (ATPase) and cell-based cytotoxicity assays, respectively. In general, the inhibitor potency of the pyrrolotriazine-4-one series in in vitro biological assays correlated with the magnitude of the thermal stability enhancement of ADP-Eg5. The thermal shift assay also confirmed direct binding of Eg5 inhibitors identified in a high-throughput screen and demonstrated that the thermal shift assay is applicable to a range of chemotypes and can be useful in evaluating both potent (nM) and relatively weakly binding (microM) leads. Overall, the thermal shift assay was found to be an excellent biophysical method for evaluating direct binding of a large number of compounds to Eg5, and it complemented the catalytic assay screens by providing an alternative determination of inhibitor potency.

Cyanoguanidine-based lactam derivatives as a novel class of orally bioavailable factor Xa inhibitors.

The N,N'-disubstituted cyanoguanidine is an excellent bioisostere of the thiourea and ketene aminal functional groups. We report the design and synthesis of a novel class of cyanoguanidine-based lactam derivatives as potent and orally active FXa inhibitors. The SAR studies led to the discovery of compound 4 (BMS-269223, K(i)=6.5nM, EC(2xPT)=32muM) as a selective, orally bioavailable FXa inhibitor with an excellent in vitro liability profile, favorable pharmacokinetics and pharmacodynamics in animal models. The X-ray crystal structure of 4 bound in FXa is presented and key ligand-protein interactions are discussed.

Design, structure-activity relationships, X-ray crystal structure, and energetic contributions of a critical P1 pharmacophore: 3-chloroindole-7-yl-based factor Xa inhibitors.

An indole-based P1 moiety was incorporated into a previously established factor Xa inhibitor series. The indole group was designed to hydrogen-bond with the carbonyl of Gly218, while its 3-methyl or 3-chloro substituent was intended to interact with Tyr228. These interactions were subsequently observed in the X-ray crystal structure of compound 18. SAR studies led to the identification of compound 20 as the most potent FXa inhibitor in this series (IC(50) = 2.4 nM, EC(2xPT) = 1.2 microM). An in-depth energetic analysis suggests that the increased binding energy of 3-chloroindole-versus 3-methylindole-containing compounds in this series is due primarily to (a) the more hydrophobic nature of chloro- versus methyl-containing compounds and (b) an increased interaction of 3-chloroindole versus 3-methylindole with Gly218 backbone. The stronger hydrophobicity of chloro- versus methyl-substituted aromatics may partly explain the general preference for chloro- versus methyl-substituted P1 groups in FXa, which extends beyond the current series.

Involvement of DPP-IV catalytic residues in enzyme-saxagliptin complex formation.

The inhibition of DPP-IV by saxagliptin has been proposed to occur through formation of a covalent but reversible complex. To evaluate further the mechanism of inhibition, we determined the X-ray crystal structure of the DPP-IV:saxagliptin complex. This structure reveals covalent attachment between S630 and the inhibitor nitrile carbon (C-O distance <1.3 A). To investigate whether this serine addition is assisted by the catalytic His-Asp dyad, we generated two mutants of DPP-IV, S630A and H740Q, and assayed them for ability to bind inhibitor. DPP-IV H740Q bound saxagliptin with an approximately 1000-fold reduction in affinity relative to DPP-IV WT, while DPP-IV S630A showed no evidence for binding inhibitor. An analog of saxagliptin lacking the nitrile group showed unchanged binding properties to the both mutant proteins, highlighting the essential role S630 and H740 play in covalent bond formation between S630 and saxagliptin. Further supporting mechanism-based inhibition by saxagliptin, NMR spectra of enzyme-saxagliptin complexes revealed the presence of three downfield resonances with low fractionation factors characteristic of short and strong hydrogen bonds (SSHB). Comparison of the NMR spectra of various wild-type and mutant DPP-IV:ligand complexes enabled assignment of a resonance at approximately 14 ppm to H740. Two additional DPP-IV mutants, Y547F and Y547Q, generated to probe potential stabilization of the enzyme-inhibitor complex by this residue, did not show any differences in inhibitor binding either by ITC or NMR. Together with the previously published enzymatic data, the structural and binding data presented here strongly support a histidine-assisted covalent bond formation between S630 hydroxyl oxygen and the nitrile group of saxagliptin.

Molecular basis for increased susceptibility of isolates with atazanavir resistance-conferring substitution I50L to other protease inhibitors.

Protease inhibitors (PIs) are highly effective drugs against the human immunodeficiency virus (HIV), yet long-term therapeutic use is limited by emergence of HIV type 1 (HIV-1) protease substitutions that confer cross-resistance to multiple protease inhibitor drugs. Atazanavir is a highly potent HIV protease inhibitor with a distinct resistance profile that includes effectiveness against most HIV-1 isolates resistant to one or two PIs. The signature resistance substitution for atazanavir is I50L, and it is frequently (53%) accompanied by a compensatory A71V substitution that helps restore viability and increases atazanavir resistance levels. We measured the binding affinities of wild-type (WT) and I50L/A71V HIV-1 proteases to atazanavir and other currently approved PIs (ritonavir, lopinavir, saquinavir, nelfinavir, indinavir, and amprenavir) by isothermal titration calorimetry. Remarkably, we find that all of the PIs have 2- to 10-fold increased affinities for I50L/A71V protease, except for atazanavir. The results are also manifested by thermal stability measures of affinity for WT and I50L/A71V proteases. Additional biophysical and enzyme kinetics experiments show I50L/A71V protease is a stable enzyme with catalytic activity that is slightly reduced (34%) relative to the WT. Computational modeling reveals that the unique resistance phenotype of I50L/A71V protease likely originates from bulky tert-butyl groups at P2 and P2' (specific to atazanavir) that sterically clash with methyl groups on residue L50. The results of this study provide a molecular understanding of the novel hypersusceptibility of atazanavir-resistant I50L/A71V-containing clinical isolates to other currently approved PIs.

Structural and functional characterization of CFE88: evidence that a conserved and essential bacterial protein is a methyltransferase.

CFE88 is a conserved essential gene product from Streptococcus pneumoniae. This 227-residue protein has minimal sequence similarity to proteins of known 3D structure. Sequence alignment models and computational protein threading studies suggest that CFE88 is a methyltransferase. Characterization of the conformation and function of CFE88 has been performed by using several techniques. Backbone atom and limited side-chain atom NMR resonance assignments have been obtained. The data indicate that CFE88 has two domains: an N-terminal domain with 163 residues and a C-terminal domain with 64 residues. The C-terminal domain is primarily helical, while the N-terminal domain has a mixed helical/extended (Rossmann) fold. By aligning the experimentally observed elements of secondary structure, an initial unrefined model of CFE88 has been constructed based on the X-ray structure of ErmC' methyltransferase (Protein Data Bank entry 1QAN). NMR and biophysical studies demonstrate binding of S-adenosyl-L-homocysteine (SAH) to CFE88; these interactions have been localized by NMR to the predicted active site in the N-terminal domain. Mutants that target this predicted active site (H26W, E46R, and E46W) have been constructed and characterized. Overall, our results both indicate that CFE88 is a methyltransferase and further suggest that the methyltransferase activity is essential for bacterial survival.

Decrypting the biochemical function of an essential gene from Streptococcus pneumoniae using ThermoFluor technology.

The protein product of an essential gene of unknown function from Streptococcus pneumoniae was expressed and purified for screening in the ThermoFluor affinity screening assay. This assay can detect ligand binding to proteins of unknown function. The recombinant protein was found to be in a dimeric, native-like folded state and to unfold cooperatively. ThermoFluor was used to screen the protein against a library of 3000 compounds that were specifically selected to provide information about possible biological functions. The results of this screen identified pyridoxal phosphate and pyridoxamine phosphate as equilibrium binding ligands (K(d) approximately 50 pM, K(d) approximately 2.5 microM, respectively), consistent with an enzymatic cofactor function. Several nucleotides and nucleotide sugars were also identified as ligands of this protein. Sequence comparison with two enzymes of known structure but relatively low overall sequence homology established that several key residues directly involved in pyridoxal phosphate binding were strictly conserved. Screening a collection of generic drugs and natural products identified the antifungal compound canescin A as an irreversible covalent modifier of the enzyme. Our investigation of this protein indicates that its probable biological role is that of a nucleoside diphospho-keto-sugar aminotransferase, although the preferred keto-sugar substrate remains unknown. These experiments demonstrate the utility of a generic affinity-based ligand binding technology in decrypting possible biological functions of a protein, an approach that is both independent of and complementary to existing genomic and proteomic technologies.

Correlation of high-throughput pregnane X receptor (PXR) transactivation and binding assays.

Pregnane X receptor (PXR) transactivation and binding assays have been developed into high-throughput assays, which are robust and reproducible (Z' > 0.5). For most compounds, there was a good correlation between the results of the transactivation and binding assays. EC(50) values of compounds in the transactivation assay correlated reasonably well with their IC(50) values in the binding assay. However, there were discrepancies with some compounds showing high binding affinity in the binding assay translated into low transactivation. The most likely cause for these discrepancies was an agonist-dependent relationship between binding affinity and transactivation response. In general, compounds that bound to human PXR and transactivated PXR tended to be large hydrophobic molecules.

Comparative studies of active site-ligand interactions among various recombinant constructs of human beta-amyloid precursor protein cleaving enzyme.

Amyloid precursor protein (APP) cleaving enzyme (BACE) is the enzyme responsible for beta-site cleavage of APP, leading to the formation of the amyloid-beta peptide that is thought to be pathogenic in Alzheimer's disease (AD). Hence, BACE is an attractive pharmacological target, and numerous research groups have begun searching for potent and selective inhibitors of this enzyme as a potential mechanism for therapeutic intervention in AD. The mature enzyme is composed of a globular catalytic domain that is N-linked glycosylated in mammalian cells, a single transmembrane helix that anchors the enzyme to an intracellular membrane, and a short C-terminal domain that extends outside the phospholipid bilayer of the membrane. Here we have compared the substrate and active site-directed inhibitor binding properties of several recombinant constructs of human BACE. The constructs studied here address the importance of catalytic domain glycosylation state, inclusion of domains other than the catalytic domain, and incorporation into a membrane bilayer on the interactions of the enzyme active site with peptidic ligands. We find no significant differences in ligand binding properties among these various constructs. These data demonstrate that the nonglycosylated, soluble catalytic domain of BACE faithfully reflects the ligand binding properties of the full-length mature enzyme in its natural membrane environment. Thus, the use of the nonglycosylated, soluble catalytic domain of BACE is appropriate for studies aimed at understanding the determinants of ligand recognition by the enzyme active site.