Tuberculosis is the leading cause of death from an infectious disease, and is caused by Mycobacterium tuberculosis (M.tb). More than 1 billion people worldwide are thought to harbor an M.tb infection. The multidrug therapy that represents the current standard of care requires a minimum of four months of dosing and drug resistant Mycobacterium tuberculosis treatment regimens are significantly longer. Inosine-5'-monophosphate dehydrogenase (GuaB) is the enzyme that performs the rate-limiting step in de novo guanine nucleotide biosynthesis that is critical for growth and viability of bacteria including M.tb. The development of a novel antibiotic that inhibits GuaB could combine with existing therapies in novel ways and thereby contribute to effective therapeutic regimens for the treatment of tuberculosis. Here we describe the discovery of structurally distinct small molecule GuaB inhibitors that are potent against M.tb H37Ra and H37Rv strains and have desirable safety and ADME profiles.

Guanine nucleotides are required for growth and viability of cells due to their structural role in DNA and RNA, and their regulatory roles in translation, signal transduction, and cell division. The natural antibiotic mycophenolic acid (MPA) targets the rate-limiting step in de novo guanine nucleotide biosynthesis executed by inosine-5´-monophosphate dehydrogenase (IMPDH). MPA is used clinically as an immunosuppressant, but whether in vivo inhibition of bacterial IMPDH (GuaB) is a valid antibacterial strategy is controversial. Here, we describe the discovery of extremely potent small molecule GuaB inhibitors (GuaBi) specific to pathogenic bacteria with a low frequency of on-target spontaneous resistance and bactericidal efficacy in vivo against Acinetobacter baumannii mouse models of infection. The spectrum of GuaBi activity includes multidrug-resistant pathogens that are a critical priority of new antibiotic development. Co-crystal structures of A. baumannii, Staphylococcus aureus, and Escherichia coli GuaB proteins bound to inhibitors show comparable binding modes of GuaBi across species and identifies key binding site residues that are predictive of whole-cell activity across both Gram-positive and Gram-negative clades of Bacteria. The clear in vivo efficacy of these small molecule GuaB inhibitors in a model of A. baumannii infection validates GuaB as an essential antibiotic target. IMPORTANCE:The emergence of multidrug-resistant bacteria worldwide has renewed interest in discovering antibiotics with novel mechanism of action. For the first time ever, we demonstrate that pharmacological inhibition of de novo guanine biosynthesis is bactericidal in a mouse model of Acinetobacter baumannii infection. Structural analyses of novel inhibitors explain differences in biochemical and whole-cell activity across bacterial clades and underscore why this discovery may have broad translational impact on treatment of the most recalcitrant bacterial infections.

In drug discovery, partition and distribution coefficients, logP and logD for octanol/water, are widely used as metrics of the lipophilicity of molecules, which in turn have a strong influence on the bioactivity and bioavailability of potential drugs. There are a variety of established methods, mostly fragment or atom-based, to calculate logP while logD prediction generally relies on calculated logP and pKa for the estimation of neutral and ionized populations at a given pH. Algorithms such as ClogP have limitations generally leading to systematic errors for chemically related molecules while pKa estimation is generally more difficult due to the interplay of electronic, inductive and conjugation effects for ionizable moieties. We propose an integrated machine learning QSAR modeling approach to predict logD by training the model with experimental data while using ClogP and pKa predicted by commercial software as model descriptors. By optimizing the loss function for the ClogD calculated by the software, we build a correction model that incorporates both descriptors from the software and available experimental logD data. Additionally, we calculate logP from the logD model using the software predicted pKa's. Here, we have trained models using publicly or commercial available logD data to show that this approach can improve on commercial software predictions of lipophilicity. When applied to other logD data sets, this approach extends the domain of applicability of logD and logP predictions over commercial software. Performance of these models favorably compare with models built with a larger set of proprietary logD data.

BACKGROUND:There is a continued need for improvements in the efficiency of metabolite structure elucidation.OBJECTIVE:We propose to take LC Retention Time (RT) into consideration during the process of structure determination.METHODS:Herein, we develop a simple methodology that employs a Chromatographic Hydrophobicity Index (CHI) framework for standardizing LC conditions and introduce and utilize the concept of a predictable CHI change upon Phase 1 biotransformation (CHIbt). Through the analysis of literature examples, we offer a Quantitative Structure-Retention Relationship (QSRR) for several types of biotransformation (especially hydroxylation) using physicochemical properties (clogP, hydrogen bonding).RESULTS:The CHI system for retention indexing is shown to be practical and simple to implement. A database of CHIbt values has been created from re-incubation of 3 compounds and from analysis of an additional 17 datasets from the literature. Application of this database is illustrated.CONCLUSION:In our experience, this simple methodology allows complementing the discovery efforts that saves resources for in-depth characterization using NMR.

The optimization of the pharmacokinetic profile of a drug is one of the crucial aspects of medicinal chemistry campaigns. When efficacy is driven by a continuous coverage of the minimum efficacious plasma concentration, half-life must be optimized to achieve the optimal pharmacokinetic profile. The consensus in the field is that decreasing clearance, as opposed to increasing volume of distribution, is a better strategy to prolong half-life. While both the pharmacokinetic theory and the need for an optimal safety profile support this approach, this needs to be integrated with practical indications concerning the strategy to optimize clearance. This work presents an extensive analysis of Genentech's in vitro and in vivo rat pharmacokinetic data, which highlights how half-life optimization through simple modulation of lipophilicity is generally not a successful strategy. Decreasing lipophilicity without addressing a metabolic soft-spot will often lead to both lower clearance and lower volume of distribution without extending half-life.

Analyzing files containing chemical information is at the core of cheminformatics. Each analysis may require a unique workflow. This paper describes the chemalot and chemalot_knime open source packages. Chemalot is a set of command line programs with a wide range of functionalities for cheminformatics. The chemalot_knime package allows command line programs that read and write SD files from stdin and to stdout to be wrapped into KNIME nodes. The combination of chemalot and chemalot_knime not only facilitates the compilation and maintenance of sequences of command line programs but also allows KNIME workflows to take advantage of the compute power of a LINUX cluster.

With the development of ever-expanding synthetic methodologies, a medicinal chemist's toolkit continues to swell. However, with finite time and resources as well as a growing understanding of our field's environment impact, it is critical to refine what can be made to what should be made. This review seeks to highlight multiple cheminformatic approaches in drug discovery that can influence and triage design and execution impacting the likelihood of rapidly generating high-value molecules in a more sustainable manner. This strategy gives chemists the tools to design and refine vast libraries, stress "druglikeness", and rapidly identify SAR trends. Project success, i.e., identification of a clinical candidate, is then reached faster with fewer molecules with the farther-reaching ramification of using fewer resources and generating less waste, thereby helping "green" our field.

p21-activated kinase 1 (PAK1) has an important role in transducing signals in several oncogenic pathways. The concept of inhibiting this kinase has garnered significant interest over the past decade, particularly for targeting cancers associated with PAK1 amplification. Animal studies with the selective group I PAK (pan-PAK1, 2, 3) inhibitor G-5555 from the pyrido[2,3-d]pyrimidin-7-one class uncovered acute toxicity with a narrow therapeutic window. To attempt mitigating the toxicity, we introduced significant structural changes, culminating in the discovery of the potent pyridone side chain analogue G-9791. Mouse tolerability studies with this compound, other members of this series, and compounds from two structurally distinct classes revealed persistent toxicity and a correlation of minimum toxic concentrations and PAK1/2 mediated cellular potencies. Broad screening of selected PAK inhibitors revealed PAK1, 2, and 3 as the only overlapping targets. Our data suggest acute cardiovascular toxicity resulting from the inhibition of PAK2, which may be enhanced by PAK1 inhibition, and cautions against continued pursuit of pan-group I PAK inhibitors in drug discovery.

A series of 4-azaindole-containing p21-activated kinase-1 (PAK1) inhibitors was prepared with the goal of improving physicochemical properties relative to an indole starting point. Indole 1 represented an attractive, non-basic scaffold with good PAK1 affinity and cellular potency but was compromised by high lipophilicity (clogD=4.4). Azaindole 5 was designed as an indole surrogate with the goal of lowering logD and resulted in equipotent PAK1 inhibition with a 2-fold improvement in cellular potency over 1. Structure-activity relationship studies around 5 identified additional 4-azaindole analogs with superior PAK1 biochemical activity (Ki <10nM) and up to 24-fold selectivity for group I over group II PAKs. Compounds from this series showed enhanced permeability, improved aqueous solubility, and lower plasma protein binding over indole 1. The improvement in physicochemical properties translated to a 20-fold decrease in unbound clearance in mouse PK studies for azaindole 5 relative to indole 1.

Intestinal absorption in human is routinely predicted in drug discovery using in vitro assays such as permeability in the Madin-Darby canine kidney cell line. In silico models trained on these data are used in drug discovery efforts to prioritize novel chemical targets for synthesis; however, their proprietary nature and the limited validation available, which is usually restricted to predicting in vitro permeability, are barriers to widespread adoption. Because of the categorical nature of the in vitro permeability assay, intrinsic assay variability, and the challenges often encountered when translating in vitro data to an in vivo drug property, validation based solely on in vitro data might not be a good characterization of the usefulness of the in silico tool. In this work, we analyze the performance of three different in silico models in predicting the in vitro and in vivo permeability of 300 marketed drugs and 86 discovery compounds. The models differ in their approach (mechanistic vs quantitative structure-activity relationship) and the degree of complexity; one of them is a linear equation based on seven simple physicochemical descriptors and is presented for the first time in this work. Results show that in silico models can be successfully used to complement the discovery toolbox for characterizing in vivo intestinal permeability, defined using fraction of dose absorbed in human (Fa) and human jejunal permeability (Peff). While the in vitro permeability models outperformed the in silico approach at predicting each of the in vivo end points explored, the gap in predictivity between the in vitro and the in vivo data was generally comparable to the gap between in silico and in vitro data. The in vitro and in silico approaches shared many of the same outliers, which can often be explained by the route of drug absorption (paracellular vs transcellular, active vs passive). Data suggest that the discovery process can greatly benefit from an early adoption of in silico models for predicting permeability as well as from a careful analysis of the in silico to in vivo disconnects.

Structure- and property-based drug design is an integral part of modern drug discovery, enabling the design of compounds aimed at improving potency and selectivity. However, building molecules using desktop modeling tools can easily lead to poor designs that appear to form many favorable interactions with the protein's active site. Although a proposed molecule looks good on screen and appears to fit into the protein site X-ray crystal structure or pharmacophore model, doing so might require a high-energy small molecule conformation, which would likely be inactive. To help scientists make better design decisions, we have built integrated, easy-to-use, interactive software tools to perform docking experiments, de novo design, shape and pharmacophore based database searches, small molecule conformational analysis and molecular property calculations. Using a combination of these tools helps scientists in assessing the likelihood that a designed molecule will be active and have desirable drug metabolism and pharmacokinetic properties. Small molecule discovery success requires project teams to rapidly design and synthesize potent molecules with good ADME properties. Empowering scientists to evaluate ideas quickly and make better design decisions with easy-to-access and easy-to-understand software on their desktop is now a key part of our discovery process.

Signaling pathways intersecting with the p21-activated kinases (PAKs) play important roles in tumorigenesis and cancer progression. By recognizing that the limitations of FRAX1036 (1) were chiefly associated with the highly basic amine it contained, we devised a mitigation strategy to address several issues such as hERG activity. The 5-amino-1,3-dioxanyl moiety was identified as an effective means of reducing pK a and logP simultaneously. When positioned properly within the scaffold, this group conferred several benefits including potency, pharmacokinetics, and selectivity. Mouse xenograft PK/PD studies were carried out using an advanced compound, G-5555 (12), derived from this approach. These studies concluded that dose-dependent pathway modulation was achievable and paves the way for further in vivo investigations of PAK1 function in cancer and other diseases.

To increase kinase selectivity in an aminopyrazole-based PAK1 inhibitor series, analogues were designed to interact with the PAK1 deep-front pocket pre-DFG residue Thr-406, a residue that is hydrophobic in most kinases. This goal was achieved by installing lactam head groups to the aminopyrazole hinge binding moiety. The corresponding analogues represent the most kinase selective ATP-competitive Group I PAK inhibitors described to date. Hydrogen bonding with the Thr-406 side chain was demonstrated by X-ray crystallography, and inhibitory activities, particularly against kinases with hydrophobic pre-DFG residues, were mitigated. Leveraging hydrogen bonding side chain interactions with polar pre-DFG residues is unprecedented, and similar strategies should be applicable to other appropriate kinases.

Using data from the in vitro liver microsomes metabolic stability assay, we have developed QSAR models to predict in vitro human clearance. Models were trained using in house high-throughput assay data reported as the predicted human hepatic clearance by liver microsomes or pCL h . Machine learning regression methods were used to generate the models. Model output for a given molecule was reported as its probability of being metabolically stable, thus allowing for synthesis prioritization based on this prediction. Use of probability, instead of the regression value or categories, has been found to be an efficient way for both reporting and assessing predictions. Model performance is evaluated using prospective validation. These models have been integrated into a number of desktop tools, and the models are routinely used to prioritize the synthesis of compounds. We discuss two therapeutic projects at Genentech that exemplify the benefits of a probabilistic approach in applying the models. A three-year retrospective analysis of measured liver microsomes stability data on all registered compounds at Genentech reveals that the use of these models has resulted in an improved metabolic stability profile of synthesized compounds.

The p21-activated kinases (PAKs) play important roles in cytoskeletal organization, cellular morphogenesis, and survival and have generated significant attention as potential therapeutic targets for cancer. Following a high-throughput screen, we identified an aminopyrazole scaffold-based series that was optimized to yield group I selective PAK inhibitors. A structure-based design effort aimed at targeting the ribose pocket for both potency and selectivity led to much-improved group I vs II selectivity. Early lead compounds contained a basic primary amine, which was found to be a major metabolic soft spot with in vivo clearance proceeding predominantly via N-acetylation. We succeeded in identifying replacements with improved metabolic stability, leading to compounds with lower in vivo rodent clearance and excellent group I PAK selectivity.

It is known that the developability of drugs is related to their physicochemical and DMPK properties. Given the time and expense involved in discovering and developing new drugs, maximizing the chance of success by calculating properties ahead of chemical synthesis and testing, and only acting on those candidates whose properties fall into a desired range, would seem to make sense. This paper provides an overview of calculable physicochemical and DMPK properties, an assessment of their relative difficulty of their calculation and accuracy, and available software. Methods companies have employed to communicate results will be discussed, including the use of composite scoring functions and ranking schemes. Calculations do no good if chemists will not use them to prioritize synthesis decisions. Strategies are presented for facilitating model usage. An approach adopted at Genentech for presenting results that involves the close coupling of property calculations with 3D structure based drug design is described.