Invention of MK-7845, a SARS-CoV-2 3CL Protease Inhibitor Employing a Novel Difluorinated Glutamine Mimic.

As SARS-CoV-2 continues to circulate, antiviral treatments are needed to complement vaccines. The virus's main protease, 3CLPro, is an attractive drug target in part because it recognizes a unique cleavage site, which features a glutamine residue at the P1 position and is not utilized by human proteases. Herein, we report the invention of MK-7845, a novel reversible covalent 3CLPro inhibitor. While most covalent inhibitors of SARS-CoV-2 3CLPro reported to date contain an amide as a Gln mimic at P1, MK-7845 bears a difluorobutyl substituent at this position. SAR analysis and X-ray crystallographic studies indicate that this group interacts with His163, the same residue that forms a hydrogen bond with the amide substituents typically found at P1. In addition to promising in vivo efficacy and an acceptable projected human dose with unboosted pharmacokinetics, MK-7845 exhibits favorable properties for both solubility and absorption that may be attributable to the unusual difluorobutyl substituent.

Structure Guided Discovery of Novel Pan Metallo-ß-Lactamase Inhibitors with Improved Gram-Negative Bacterial Cell Penetration.

The use of ß-lactam (BL) and ß-lactamase inhibitor combination to overcome BL antibiotic resistance has been validated through clinically approved drug products. However, unmet medical needs still exist for the treatment of infections caused by Gram-negative (GN) bacteria expressing metallo-ß-lactamases. Previously, we reported our effort to discover pan inhibitors of three main families in this class: IMP, VIM, and NDM. Herein, we describe our work to improve the GN coverage spectrum in combination with imipenem and relebactam. This was achieved through structure- and property-based optimization to tackle the GN cell penetration and efflux challenges. A significant discovery was made that inhibition of both VIM alleles, VIM-1 and VIM-2, is essential for broad GN coverage, especially against VIM-producing P. aeruginosa. In addition, pharmacokinetics and nonclinical safety profiles were investigated for select compounds. Key findings from this drug discovery campaign laid the foundation for further lead optimization toward identification of preclinical candidates.

Rapid Evolution of a Fragment-like Molecule to Pan-Metallo-Beta-Lactamase Inhibitors: Initial Leads toward Clinical Candidates.

With the emergence and rapid spreading of NDM-1 and existence of clinically relevant VIM-1 and IMP-1, discovery of pan inhibitors targeting metallo-beta-lactamases (MBLs) became critical in our battle against bacterial infection. Concurrent with our fragment and high-throughput screenings, we performed a knowledge-based search of known metallo-beta-lactamase inhibitors (MBLIs) to identify starting points for early engagement of medicinal chemistry. A class of compounds exemplified by 11, discovered earlier as B. fragilis metallo-beta-lactamase inhibitors, was selected for in silico virtual screening. From these efforts, compound 12 was identified with activity against NDM-1 only. Initial exploration on metal binding design followed by structure-guided optimization led to the discovery of a series of compounds represented by 23 with a pan MBL inhibition profile. In in vivo studies, compound 23 in combination with imipenem (IPM) robustly lowered the bacterial burden in a murine infection model and became the lead for the invention of MBLI clinical candidates.

Modulating the interaction between CDK2 and cyclin A with a quinoline-based inhibitor.

A new class of quinoline-based kinase inhibitors has been discovered that both disrupt cyclin dependent 2 (CDK2) interaction with its cyclin A subunit and act as ATP competitive inhibitors. The key strategy for discovering this class of protein-protein disrupter compounds was to screen the monomer CDK2 in an affinity-selection/mass spectrometry-based technique and to perform secondary assays that identified compounds that bound only to the inactive CDK2 monomer and not the active CDK2/cyclin A heterodimer. Through a series of chemical modifications the affinity (Kd) of the original hit improved from 1 to 0.005µM.

Fully activated MEK1 exhibits compromised affinity for binding of allosteric inhibitors U0126 and PD0325901.

Kinases catalyze the transfer of γ-phosphate from ATP to substrate protein residues triggering signaling pathways responsible for a plethora of cellular events. Isolation and production of homogeneous preparations of kinases in their fully active forms is important for accurate in vitro measurements of activity, stability, and ligand binding properties of these proteins. Previous studies have shown that MEK1 can be produced in its active phosphorylated form by coexpression with RAF1 in insect cells. In this study, using activated MEK1 produced by in vitro activation by RAF1 (pMEK1(in vitro)), we demonstrate that the simultaneous expression of RAF1 for production of activated MEK1 does not result in stoichiometric phosphorylation of MEK1. The pMEK1(in vitro) showed higher specific activity toward ERK2 protein substrate compared to the pMEK1 that was activated via coexpression with RAF1 (pMEK1(in situ)). The two pMEK1 preparations showed quantitative differences in the phosphorylation of T-loop residue serine 222 by Western blotting and mass spectrometry. Finally, pMEK1(in vitro) showed marked differences in the ligand binding properties compared to pMEK1(in situ). Contrary to previous findings, pMEK1(in vitro) bound allosteric inhibitors U0126 and PD0325901 with a significantly lower affinity than pMEK1(in situ) as well as its unphosphorylated counterpart (npMEK1) as demonstrated by thermal-shift, AS-MS, and calorimetric studies. The differences in inhibitor binding affinity provide direct evidence that unphosphorylated and RAF1-phosphorylated MEK1 form distinct inhibitor sites.

Crystal structures of MEK1 binary and ternary complexes with nucleotides and inhibitors.

MEK1 is a member of the MAPK signal transduction pathway that responds to growth factors and cytokines. We have determined that the kinase domain spans residues 35-382 by proteolytic cleavage. The complete kinase domain has been crystallized and its X-ray crystal structure as a complex with magnesium and ATP-gammaS determined at 2.1 A. Unlike crystals of a truncated kinase domain previously published, the crystals of the intact domain can be grown either as a binary complex with a nucleotide or as a ternary complex with a nucleotide and one of a multitude of allosteric inhibitors. Further, the crystals allow for the determination of costructures with ATP competitive inhibitors. We describe the structures of nonphosphorylated MEK1 (npMEK1) binary complexes with ADP and K252a, an ATP-competitive inhibitor (see Table 1), at 1.9 and 2.7 A resolution, respectively. Ternary complexes have also been solved between npMEK1, a nucleotide, and an allosteric non-ATP competitive inhibitor: ATP-gammaS with compound 1 and ADP with either U0126 or the MEK1 clinical candidate PD325089 at 1.8, 2.0, and 2.5 A, respectively. Compound 1 is structurally similar to PD325901. These structures illustrate fundamental differences among various mechanisms of inhibition at the molecular level. Residues 44-51 have previously been shown to play a negative regulatory role in MEK1 activity. The crystal structure of the integral kinase domain provides a structural rationale for the role of these residues. They form helix A and repress enzymatic activity by stabilizing an inactive conformation in which helix C is displaced from its active state position. Finally, the structure provides for the first time a molecular rationale that explains how mutations in MEK may lead to the cardio-facio-cutaneous syndrome.

Expression and purification of phosphorylated and non-phosphorylated human MEK1.

Kinases exist in either a high or low activity form depending on the phosphorylation state of the activating lip. These two different forms of the same kinase may adopt different conformations that affect not only activity but also inhibitor binding and the ability to crystallize the protein. Therefore, isolation of homogenous preparations of the phosphorylated and non-phosphorylated versions of a kinase is critical for accurate biophysical measurements of activity, stability and ligand binding as well as for protein crystallization. The aim of the present study is the expression, purification and characterization of recombinant human MEK1 protein in both the activated and low-activity states. A baculovirus co-expression system was developed for obtaining high levels of activated, phosphorylated human MEK1 kinase. High-Five cells were co-infected with human MEK1 virus and Raf-BXB, an untagged constitutively active version of Raf which is the activating kinase for MEK1. Unphosphorylated MEK1 was generated by treating MEK1 isolated from High-Five baculovirus expression with lambda-phosphatase. The proteins were characterized by SDS-PAGE, LC-MS, Western blotting, enzymatic activity, and circular dichroism. Previous reports of MEK1 expression and purification yielded lower levels of protein and purity. The yield using High-Five cells was 5mg/L for phosphorylated MEK1 and 10mg/L for unphosphorylated MEK1. For phosphorylated MEK1, the specific activity was 3530U/mg, the IC(50) values for the non-specific kinase inhibitors K252a and K252b were 8 and 47nM, respectively, and the IC(50) for the MEK1 non-ATP competitive inhibitor, PD0325901, was 43nM.

Ligand binding affinity determined by temperature-dependent circular dichroism: cyclin-dependent kinase 2 inhibitors.

To support drug discovery efforts for cyclin-dependent kinase 2 (CDK2), a moderate-throughput binding assay that can rank order or estimate the affinity of lead inhibitors has been developed. The method referred to as temperature-dependent circular dichroism (TdCD) uses the classical temperature-dependent unfolding of proteins by circular dichroism (CD) to measure the degree of protein unfolding in the absence and presence of potential inhibitors. The midpoint of unfolding is the Tm value. Rank ordering the affinity and predictions of the dissociation constant of compounds is obtained by measuring the increase in Tm for different protein-inhibitor complexes. This is the first time an extensive characterization of the TdCD method has been described for characterizing lead inhibitors in a drug discovery mode. The method has several favorable properties. Using the new six-cell Peltier temperature controller for the Jasco 810 spectropolarimeter, one can determine the affinity of 12-18 compounds per day. The method also requires only 20-40 microg protein per sample and can be used to estimate the affinity of compounds with dissociation constants of picomolar to micromolar. An important property of the method for lead discovery is that dissociation constants of approximately 5 microM can be estimated from a single experiment using a low concentration of compound such as 20 microM, which is generally low enough for most small molecules to be soluble for testing. In addition, the method does not require labeling the compound or protein. Although other methods such as isothermal titration calorimetry (ITC) can provide a full thermodynamic characterization of binding, ITC requires 1-2 mg protein per sample, cannot readily determine binding constants below nanomolar values, is most versatile with soluble compounds, and has a throughput of two to three experiments per day. The ITC method is not usually used in a high-throughput drug discovery mode; however, using the thermodynamic information from several ITC experiments can make the TdCD method very robust in determining reliable binding constants. Using the kinase inhibitors BMS-250595, purvalanol B, AG-12275, flavopiridol, and several other compounds, it is demonstrated that one can obtain excellent comparisons between the Kd values of binding to CDK2 obtained by TdCD and ITC.