Structure-Based Design and Identification of FT-2102 (Olutasidenib), a Potent Mutant-Selective IDH1 Inhibitor.

Inhibition of mutant IDH1 is being evaluated clinically as a treatment option for oncology. Here we describe the structure-based design and optimization of quinoline lead compounds to identify FT-2102, a potent, orally bioavailable, brain penetrant, and selective mIDH1 inhibitor. FT-2102 has excellent ADME/PK properties and reduces 2-hydroxyglutarate levels in an mIDH1 xenograft tumor model. This compound has been selected as a candidate for clinical development in hematologic malignancies, solid tumors, and gliomas with mIDH1.

Discovery and Optimization of Quinolinone Derivatives as Potent, Selective, and Orally Bioavailable Mutant Isocitrate Dehydrogenase 1 (mIDH1) Inhibitors.

Mutations at the arginine residue (R132) in isocitrate dehydrogenase 1 (IDH1) are frequently identified in various human cancers. Inhibition of mutant IDH1 (mIDH1) with small molecules has been clinically validated as a promising therapeutic treatment for acute myeloid leukemia and multiple solid tumors. Herein, we report the discovery and optimization of a series of quinolinones to provide potent and orally bioavailable mIDH1 inhibitors with selectivity over wild-type IDH1. The X-ray structure of an early lead 24 in complex with mIDH1-R132H shows that the inhibitor unexpectedly binds to an allosteric site. Efforts to improve the in vitro and in vivo absorption, distribution, metabolism, and excretion (ADME) properties of 24 yielded a preclinical candidate 63. The detailed preclinical ADME and pharmacology studies of 63 support further development of quinolinone-based mIDH1 inhibitors as therapeutic agents in human trials.

Identification and application of NEDD8 E1 inhibitors.

The NEDD8 conjugation pathway is initiated by the NEDD8 E1, also known as NEDD8 activating enzyme (NAE) or APPBP1/UBA3 (Gong, Yeh. J Biol Chem 274:12063-12042, 1999). The best described biological role for NEDD8 conjugation is to regulate the activity of the cullin RING ligase (CRL) family of ubiquitin E3 ligases (Gong, Yeh. J Biol Chem 274:12063-12042, 1999). In this way, the NEDD8 pathway regulates the turnover of a subset of ubiquitin proteasome system (UPS) substrates that are essential for cancer cell growth and survival (Soucy, Smith, Milhollen. Nature 458:732-737, 2009). We recently initiated clinical trials with a first-in-class small molecule inhibitor of NAE for the treatment of cancer (Soucy, Smith, Milhollen. Nature 458:732-737, 2009). Here we describe a biochemical and cell-based assay used to identify NAE inhibitors and monitor inhibition of the NEDD8 conjugation pathway.

Allosteric activation via kinetic control: potassium accelerates a conformational change in IMP dehydrogenase.

Allosteric activators are generally believed to shift the equilibrium distribution of enzyme conformations to favor a catalytically productive structure; the kinetics of conformational exchange is seldom addressed. Several observations suggested that the usual allosteric mechanism might not apply to the activation of IMP dehydrogenase (IMPDH) by monovalent cations. Therefore, we investigated the mechanism of K(+) activation in IMPDH by delineating the kinetic mechanism in the absence of monovalent cations. Surprisingly, the K(+) dependence of k(cat) derives from the rate of flap closure, which increases by ≥65-fold in the presence of K(+). We performed both alchemical free energy simulations and potential of mean force calculations using the orthogonal space random walk strategy to computationally analyze how K(+) accelerates this conformational change. The simulations recapitulate the preference of IMPDH for K(+), validating the computational models. When K(+) is replaced with a dummy ion, the residues of the K(+) binding site relax into ordered secondary structure, creating a barrier to conformational exchange. K(+) mobilizes these residues by providing alternate interactions for the main chain carbonyls. Potential of mean force calculations indicate that K(+) changes the shape of the energy well, shrinking the reaction coordinate by shifting the closed conformation toward the open state. This work suggests that allosteric regulation can be under kinetic as well as thermodynamic control.

The Cys319 loop modulates the transition between dehydrogenase and hydrolase conformations in inosine 5'-monophosphate dehydrogenase.

X-ray crystal structures of enzyme-ligand complexes are widely believed to mimic states in the catalytic cycle, but this presumption has seldom been carefully scrutinized. In the case of Tritrichomonas foetus inosine 5'-monophosphate dehydrogenase (IMPDH), 10 structures of various enzyme-substrate-inhibitor complexes have been determined. The Cys319 loop is found in at least three different conformations, suggesting that its conformation changes as the catalytic cycle progresses from the dehydrogenase step to the hydrolase reaction. Alternatively, only one conformation of the Cys319 loop may be catalytically relevant while the others are off-pathway. Here we differentiate between these two hypotheses by analyzing the effects of Ala substitutions at three residues of the Cys319 loop, Arg322, Glu323, and Gln324. These mutations have minimal effects on the value of k(cat) (≤5-fold) that obscure large effects (>10-fold) on the microscopic rate constants for individual steps. These substitutions increase the equilibrium constant for the dehydrogenase step but decrease the equilibrium between open and closed conformations of a mobile flap. More dramatic effects are observed when Arg322 is substituted with Glu, which decreases the rates of hydride transfer and hydrolysis by factors of 2000 and 130, respectively. These experiments suggest that the Cys319 loop does indeed have different conformations during the dehydrogenase and hydrolase reactions as suggested by the crystal structures. Importantly, these experiments reveal that the structure of the Cys319 loop modulates the closure of the mobile flap. This conformational change converts the enzyme from a dehydrogenase into hydrolase, suggesting that the conformation of the Cys319 loop may gate the catalytic cycle.

An enzymatic atavist revealed in dual pathways for water activation.

Inosine monophosphate dehydrogenase (IMPDH) catalyzes an essential step in the biosynthesis of guanine nucleotides. This reaction involves two different chemical transformations, an NAD-linked redox reaction and a hydrolase reaction, that utilize mutually exclusive protein conformations with distinct catalytic residues. How did Nature construct such a complicated catalyst? Here we employ a "Wang-Landau" metadynamics algorithm in hybrid quantum mechanical/molecular mechanical (QM/MM) simulations to investigate the mechanism of the hydrolase reaction. These simulations show that the lowest energy pathway utilizes Arg418 as the base that activates water, in remarkable agreement with previous experiments. Surprisingly, the simulations also reveal a second pathway for water activation involving a proton relay from Thr321 to Glu431. The energy barrier for the Thr321 pathway is similar to the barrier observed experimentally when Arg418 is removed by mutation. The Thr321 pathway dominates at low pH when Arg418 is protonated, which predicts that the substitution of Glu431 with Gln will shift the pH-rate profile to the right. This prediction is confirmed in subsequent experiments. Phylogenetic analysis suggests that the Thr321 pathway was present in the ancestral enzyme, but was lost when the eukaryotic lineage diverged. We propose that the primordial IMPDH utilized the Thr321 pathway exclusively, and that this mechanism became obsolete when the more sophisticated catalytic machinery of the Arg418 pathway was installed. Thus, our simulations provide an unanticipated window into the evolution of a complex enzyme.

Crystal structures of complexes of bacterial DD-peptidases with peptidoglycan-mimetic ligands: the substrate specificity puzzle.

The X-ray crystal structures of covalent complexes of the Actinomadura R39 dd-peptidase and Escherichia coli penicillin-binding protein (PBP) 5 with beta-lactams bearing peptidoglycan-mimetic side chains have been determined. The structure of the hydrolysis product of an analogous peptide bound noncovalently to the former enzyme has also been obtained. The R39 DD-peptidase structures reveal the presence of a specific binding site for the D-alpha-aminopimelyl side chain, characteristic of the stem peptide of Actinomadura R39. This binding site features a hydrophobic cleft for the pimelyl methylene groups and strong hydrogen bonding to the polar terminus. Both of these active site elements are provided by amino acid side chains from two separate domains of the protein. In contrast, no clear electron density corresponding to the terminus of the peptidoglycan-mimetic side chains is present when these beta-lactams are covalently bound to PBP5. There is, therefore, no indication of a specific side-chain binding site in this enzyme. These results are in agreement with those from kinetics studies published earlier and support the general prediction made at the time of a direct correlation between kinetics and structural evidence. The essential high-molecular-mass PBPs have demonstrated, to date, no specific reactivity with peptidoglycan-mimetic peptide substrates and beta-lactam inhibitors and, thus, probably do not possess a specific substrate-binding site of the type demonstrated here with the R39 DD-peptidase. This striking deficiency may represent a sophisticated defense mechanism against low-molecular-mass substrate-analogue inhibitors/antibiotics; its discovery should focus new inhibitor design.

A kinetic alignment of orthologous inosine-5'-monophosphate dehydrogenases.

IMP dehydrogenase (IMPDH) catalyzes two very different chemical transformations, a dehydrogenase reaction and a hydrolysis reaction. The enzyme toggles between the open conformation required for the dehydrogenase reaction and the closed conformation of the hydrolase reaction by moving a mobile flap into the NAD site. Despite these multiple functional constraints, the residues of the flap and NAD site are highly diverged, and the equilibrium between open and closed conformations ( K c ) varies widely. In order to understand how differences in the dynamic properties of the flap influence the catalytic cycle, we have delineated the kinetic mechanism of IMPDH from the pathogenic protozoan parasite Cryptosporidium parvum ( CpIMPDH), which was obtained from a bacterial source through horizontal gene transfer, and its host counterpart, human IMPDH type 2 (hIMPDH2). Interestingly, the intrinsic binding energy of NAD (+) differentially distributes across the dinucleotide binding sites of these two enzymes as well as in the previously characterized IMPDH from Tritrichomonas foetus ( TfIMPDH). Both the dehydrogenase and hydrolase reactions display significant differences in the host and parasite enzymes, in keeping with the phylogenetic and structural divergence of their active sites. Despite large differences in K c , the catalytic power of both the dehydrogenase and hydrolase conformations are similar in CpIMPDH and TfIMPDH. This observation suggests that the closure of the flap simply sets the stage for catalysis rather than plays a more active role in the chemical transformation. This work provides the essential mechanistic framework for drug discovery.

Reactions of peptidoglycan-mimetic beta-lactams with penicillin-binding proteins in vivo and in membranes.

The membrane-bound bacterial D-alanyl- D-alanine peptidases or penicillin-binding proteins (PBPs) catalyze the final transpeptidation reaction of bacterial cell wall biosynthesis and are the targets of beta-lactam antibiotics. Rather surprisingly, the substrate specificity of these enzymes is not well understood. In this paper, we present measurements of the reactivity of typical examples of these enzymes with peptidoglycan-mimetic beta-lactams under in vivo conditions. The minimum inhibitory concentrations of beta-lactams with Escherichia coli-specific side chains were determined against E. coli cells. Analogous measurements were made with Streptococcus pneumoniae R6. The reactivity of the relevant beta-lactams with E. coli PBPs in membrane preparations was also determined. The results show that under none of the above protocols were beta-lactams with peptidoglycan-mimetic side chains more reactive than generic analogues. This suggests that in vivo, as in vitro, these enzymes do not specifically recognize elements of peptidoglycan structure local to the reaction center. Substrate recognition must thus involve extended structure.

Reactivity of penicillin-binding proteins with peptidoglycan-mimetic beta-lactams: what's wrong with these enzymes?

Beta-lactams exert their antibiotic action through their inhibition of bacterial DD-peptidases (penicillin-binding proteins). Bacteria, in general, carry several such enzymes localized on the outside of their cell membrane to catalyze the final step in cell wall (peptidoglycan) synthesis. They have been classified into two major groups, one of high molecular weight, the other of low. Members of the former group act as transpeptidases in vivo, and their inhibition by beta-lactams leads to cessation of bacterial growth. The latter group consists of DD-carboxypeptidases, and their inhibition by beta-lactams is generally not fatal to bacteria. We have previously shown that representatives of the former group are ineffective at catalyzing the hydrolysis/aminolysis of peptidoglycan-mimetic peptides in vitro [Anderson et al. (2003) Biochem. J. 373, 949-955]. The theme of these experiments is expanded in the present paper where we describe the synthesis of a series of beta-lactams (penicillins and cephalosporins) containing peptidoglycan-mimetic side chains and the kinetics of their inhibition of a panel of penicillin-binding proteins spanning the major classes (Escherichia coli PBP 2 and PBP 5, Streptococcus pneumoniae PBP 1b, PBP 2x and PBP 3, the Actinomadura R39 DD-peptidase, and the Streptomyces R61 DD-peptidase). The results of these experiments mirror and expand the previous results with peptides. Neither peptides nor beta-lactams with appropriate peptidoglycan-mimetic side chains react with the solubilized constructs of membrane-bound penicillin binding proteins (the first five enzymes above) at rates exceeding those of generic analogues. Such peptides and beta-lactams do react at greatly enhanced rates with certain soluble low molecular weight enzymes (R61 and R39 DD-peptidases). The former result is unexpected and interesting. Why do the majority of penicillin-binding proteins not recognize elements of local peptidoglycan structure? Possible answers are discussed. That this question needs to be asked casts fascinating shadows on current studies of penicillin-binding proteins for new drug design.

Crystal structures of complexes between the R61 DD-peptidase and peptidoglycan-mimetic beta-lactams: a non-covalent complex with a "perfect penicillin".

The bacterial D-alanyl-D-alanine transpeptidases (DD-peptidases) are the killing targets of beta-lactams, the most important clinical defense against bacterial infections. However, due to the constant development of antibiotic-resistance mechanisms by bacteria, there is an ever-present need for new, more effective antimicrobial drugs. While enormous numbers of beta-lactam compounds have been tested for antibiotic activity in over 50 years of research, the success of a beta-lactam structure in terms of antibiotic activity remains unpredictable. Tipper and Strominger suggested long ago that beta-lactams inhibit DD-peptidases because they mimic the D-alanyl-D-alanine motif of the peptidoglycan substrate of these enzymes. They also predicted that beta-lactams having a peptidoglycan-mimetic side-chain might be better antibiotics than their non-specific counterparts, but decades of research have not provided any evidence for this. We have recently described two such novel beta-lactams. The first is a penicillin having the glycyl-L-alpha-amino-epsilon-pimelyl side-chain of Streptomyces strain R61 peptidoglycan, making it the "perfect penicillin" for this organism. The other is a cephalosporin with the same side-chain. Here, we describe the X-ray crystal structures of the perfect penicillin in non-covalent and covalent complexes with the Streptomyces R61 DD-peptidase. The structure of the non-covalent enzyme-inhibitor complex is the first such complex to be trapped crystallographically with a DD-peptidase. In addition, the covalent complex of the peptidyl-cephalosporin with the R61 DD-peptidase is described. Finally, two covalent complexes with the traditional beta-lactams benzylpenicillin and cephalosporin C were determined for comparison with the peptidyl beta-lactams. These structures, together with relevant kinetics data, support Tipper and Strominger's assertion that peptidoglycan-mimetic side-chains should improve beta-lactams as inhibitors of DD-peptidases.

The perfect penicillin? Inhibition of a bacterial DD-peptidase by peptidoglycan-mimetic beta-lactams.

6-(Glycyl-l-alpha-aminopimelyl)-aminopenicillanic acid and 7-(glycyl-l-alpha-aminopimelyl)-aminocephalosporanic acid have been synthesized as Streptomyces sp. peptidoglycan-mimetic beta-lactams. These compounds inactivate the Streptomyces R61 DD-peptidase with rate constants of 1.5 x 107 s-1 M-1 and 5.6 x 105 s-1 M-1, respectively. The former compound is thus the most effective beta-lactam inhibitor of a DD-peptidase yet described. The analogous d-alanyl-d-alanine peptide has previously been shown to react with this enzyme with comparable efficiency, kcat/Km = 8.7 x 106 s-1 M-1. These results show that, in this case at least, incorporation of a peptidoglycan-mimetic side chain into a beta-lactam greatly enhances its activity as a DD-peptidase inhibitor. This result has interesting implications for beta-lactam design.