Retinitis Pigmentosa (RP) is an incurable inherited retinal degeneration affecting approximately 1 in 4,000 individuals globally. The aim of this dissertation was to identify drugs that can help patients suffering from the disease. To accomplish this goal, the zebrafish was utilized as a model for RP to perform <i>in vivo</i> drug screening. The zebrafish RP model expresses a human rhodopsin transgene which contains a premature stop codon at position 344 (<i>Tg</i>(<i>rho:Hsa.RH1_Q344X</i>)). This zebrafish model exhibits significant rod photoreceptor degeneration beginning at 7 days post fertilization (dpf). To assess the visual consequence of this rod degeneration the zebrafish behavior visual motor response (VMR) was assayed under scotopic conditions. The Q344X RP model larvae displayed a deficit in this VMR in response to a scotopic light offset. This deficit in behavior was utilized to perform a drug screen to identify compounds that could ameliorate the deficient Q344X VMR. The ENZO SCREEN-WELL® REDOX library was chosen to be screened since oxidative stress may increase RP progression in a non-specific manner. From this library, a β-blocker, carvedilol, was identified as a compound that improved the Q344X VMR behavior. This drug was also able to increase rod number in the Q344X retina. Carvedilol was shown to be capable of working directly on rods by demonstrating that the drug can signal through the adrenergic pathway in the rod-like human Y79 cell line. Since carvedilol is an FDA-approved drug, this screening paradigm was utilized to screen the Selleckchem FDA-approved library to identify more drugs that can potentially be repurposed to treat RP like carvedilol. Additionally, this scotopic VMR assay was used to demonstrate that it can identify behavioral deficits in the P23H RP model zebrafish<i> (Tg</i>(<i>rho:Hsa.RH1_P23H</i>)). This dissertation work provides a potential FDA-approved drug for RP treatment and sets the foundation for future drug screening to identify more drugs to treat different forms of RP.

Chemical science spans multiple scales, from a single proton to the collection of proteins that make up a proteome. Throughout my graduate research career, I have developed statistical and machine learning models to better understand chemistry at these different scales, including predicting molecular properties of molecules in analytical and synthetic chemistry to integrating experiments with chemo-proteomic based machine models for drug design. Starting with the proteome, I will discuss repurposing compounds for mental health indications and visualizing the relationships between these disorders. Moving to the cellular level, I will introduce the use of the negative binomial distribution to find biomarkers collected using MS/MS and machine learning models (ML) used to select potent, non-toxic, small molecules for the treatment of castration--resistant prostate cancer (CRPC). For the protein scale, I will introduce CANDOCK, a docking method to rapidly and accurately dock small molecules, an algorithm which was used to create the ML model for CRPC. Next, I will showcase a deep learning model to determine small-molecule functional groups using FTIR and MS spectra. This will be followed by a similar approach used to identify if a small molecule will undergo a diagnostic reaction using mass spectrometry using a chemically interpretable graph-based machine learning method. Finally, I will examine chemistry at the proton level and how quantum mechanics combined with machine learning can be used to understand chemical reactions. I believe that chemical machine learning models have the potential to accelerate several aspects of drug discovery including discovery, process, and analytical chemistry.

When applied to drug discovery, modern computational systems can provide insight into the highly complex systems underlying drug activity and predict compounds or targets of interest. Many tools have been developed for computer aided drug discovery (CADD), focusing on small molecule ligands, protein targets, or both. The aim of this thesis is the improvement of CADD tools for describing small molecule properties and application of CADD to several stages of drug discovery regarding two targets for the treatment of obesity and related diseases: the neuropeptide Y4 receptor (Y4R) and the melanocortin-4 receptor (MC4R). In the first chapter, the major categories of CADD are outlined, including descriptions for many of the popular tools and examples where these tools have directly contributed to the discovery of new drugs. Following the introduction, several improvements for encoding stereochemistry and signed property distribution are introduced and tested in scenarios meant to simulate applications in virtual high-throughput screening. Y4R and MC4R are both class A G-protein coupled receptors (GPCRs) with endogenous peptide ligands that play critical roles in the signaling of satiety and energy metabolism. So far, no structures from either receptor family have been experimentally elucidated. CADD was combined with high-throughput screening (HTS) to discover the first small molecule positive allosteric modulators (PAMs) of Y4R. Secondly, CADD techniques were used to model the interaction of Y4R and pancreatic polypeptide based on experimental results that elucidate specific binding contacts. Similar SB-CADD approaches were used to model the interaction of MC4R with its high affinity peptide agonist α-MSH. Due to its role in monogenic forms of obesity, these models were used to predict which residues directly participate in binding and correlate mutated residues with their potential role in the binding site.

Fragment-screening by X-ray crystallography (XFS) is an expensive and low throughput fragment drug discovery screening method, and it requires a lot of optimization for each protein target. The advantages with this screening method are that it is very sensitive, it directly gives the three-dimensional structure of the protein-fragment complexes, and false positives are rarely obtained. The aim of this project was to help Sprint Bioscience assess if the advantages with XFS outweigh the disadvantages, and if this method should be used as a complement to their differential scanning fluorimetry (DSF) screening method. An XFS campaign was run using the oncoprotein vaccinia related kinase 1 (VRK1) as a target protein to evaluate this screening method. During the development of the XFS campaign, a diverse fragment library was created which consisted of 298 fragments that were all soluble in DMSO at 1 M concentration. The crystallization of the protein VRK1 was also optimized in this project to get a robust, high throughput crystallization set up which generated crystals that diffracted at higher resolution than 2.0 Å when they were not soaked with fragments. The soaking protocol was also optimized in order to reduce both the steps during the screening procedure and mechanical stress caused to the crystals during handling. Lastly, the created fragment library was used in screening VRK1 at 87.5 mM concentration with XFS. 23 fragment hits could be obtained from the X-ray crystallography screening campaign, and the mean resolution of the crystal structures of the protein-fragment complexes was 1.87Å. 11 of the 23 fragment hits were not identified as hits when they were screened against VRK1 using DSF. XFS was deemed as a suitable and efficient screening method to complement DSF since the hit rate was high and fragments hits could be obtained with this method that could not be obtained with DSF. However, in order to use this screening method a lot of time needs to be spent in optimizing the crystal system so it becomes suitable for fragment screening. Sprint Bioscience would therefore need to evaluate the cost/benefit ratio of using this screening method for each new project.

<p>  </p> <p>Cyclic dinucleotides (CDNs) such as the bacterial CDNs (cyclic-di-AMP, cyclic-di-GMP and 3’3’cyclic GMP-AMP) and mammalian CDN, 2’3’-cGAMP, are essential immune response second messenger signaling molecules. These CDNs act via Stimulator of interferon genes (STING)-TANK Binding Kinase 1 (TBK1)-Interferon Regulatory Factor 3 (IRF3) pathway. However, data from our lab and others indicate that beyond the classical STING-TBK1-IRF3 pathway, CDNs also regulate other signaling axes related to both inflammatory and non-inflammatory pathways. But, a global view of how these CDNs affect signaling in diverse cells or through non-STING pathways is lacking. There is also paucity of data on CDN modulated kinases and no global assessment of phosphorylation events that follow cyclic GMP AMP synthase (cGAS)-STING axis stimulation in immune cells. Herein, I have used a proteomics approach to determine signaling pathways regulated by bacterial CDNs, c-di-GMP and c-di-AMP in human gingival fibroblasts such as pathways related to nucleotide excision repair (NER) which ordinarily do not channel through STING (Chapter 3). Additionally, with the use of phosphoproteomics and bioinformatics, this project accomplished a system-wide phosphorylation analyses of T cells treated with 2’3’cGAMP and showed that 2’3’cGAMP impact various, yet unreported critical kinases (E.g. LCK, ZAP70, ARG2) and signaling pathways important for T cell function (Chapter 4). Asides known interferon signaling, these differentially phosphorylated kinases were involved in T cell receptor (TCR) signaling, myeloid cell differentiation, cell cycle regulation, and regulation of double strand break repair. </p> <p>Another area of interest addressed by this project is the discovery of novel antibacterial agents against multi-drug resistant (MDR) bacteria. Thus, in Chapters 5 and 6, I show the identification, antibacterial activity and characterization of <strong>HSD1835</strong> and <strong>HSD1919 </strong>as novel SF5 and SCF-containing membrane active compounds, highly potent against preformed MDR biofilms with fast bactericidal activity against persister bacteria. Plus, an exciting addition to the fight against MDR bacteria in Chapter 7, the discovery of <strong>HSD1624</strong> and analogs, which are able to re-sensitize MDR and colistin resistant bacteria such as <em>Pseudomonas aeruginosa</em> from a colistin MIC of 1024 μg/mL to 0.03 μg/mL (64000-fold reduction). Ultimately, these compounds could be translated into anti-biofilm and, anti-MDR bacteria therapeutics, preventing repeated surgeries due to infections, and saving lives. </p>

Cystic fibrosis (CF) is a rare genetic disease caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR), an epithelial anion channel expressed in several vital organs. Absence of functional CFTR results in imbalanced osmotic equilibrium and subsequent mucus build up in the lungs-which increases the risk of infection and eventually causes death. CFTR is an ATP-binding cassette (ABC) transporter family protein composed of two transmembrane domains (TMDs), two nucleotide binding domains (NBDs), and an unstructured regulatory domain. The most prevalent patient mutation is the deletion of F508 (F508del), making F508del CFTR the primary target for current FDA approved CF therapies. However, no experimental multi-domain F508del CFTR structure has been determined and few studies have modeled F508del using multi-domain WT CFTR structures. Here, we used cryo-EM density data and Rosetta comparative modeling (RosettaCM) to compare a F508del model with published experimental data on CFTR NBD1 thermodynamics. We then apply this modeling method to generate multi-domain WT and F508del CFTR structural models. These models demonstrate the destabilizing effects of F508del on NBD1 and the NBD1/TMD interface in both the inactive and active conformation of CFTR. Furthermore, we modeled F508del/R1070W and F508del bound to the CFTR corrector VX-809. Our models reveal the stabilizing effects of VX-809 on multi-domain models of F508del CFTR and pave the way for rational design of additional drugs that target F508del CFTR for treatment of CF.

In drug development, molecular docking methods aim at characterizing the binding of a drug-like molecule to a protein. In a typical drug development process, a docking task is repeated millions of time, which makes optimization efforts essential. In particular, modern heterogeneous architectures, such as GPUs, allow for significant acceleration opportunities. AutoDock-GPU, a state-of-the-art GPU-accelerated molecular docking software, estimates the geometrical conformation of a docked ligand-protein complex by minimizing an energy-based scoring function. Our profiling results indicated that a reduction operation, which is performed several millions times in a single docking run, limits performance in AutoDock-GPU. Thus, we proposed a method to accelerate the block-level sum reduction of four-element vectors by using matrix operations. We implemented our method to make use of the high throughput capabilities offered by NVIDIA Tensor Cores to perform matrix operations. We evaluated our approach by designing a simple benchmark, and achieved a 4 to 7-fold runtime improvement compared to the original method. We then integrated our reduction operation into AutoDock-GPU and evaluated it on multiple chemical complexes on three GPUs. This evaluation allowed to assess the possibility to use half-precision reduction operations in parts of AutoDock-GPU code, without detrimental effects on the simulation result. In addition, our implementation achieved an average 27% improvement on the overall docking time during a real-world docking run. / Vid läkemedelsutveckling syftar molekylär dockningsmetoder till att karakterisera bindningen av en läkemedelsliknande molekyl till ett protein. I en typisk läkemedelsutvecklingsprocess upprepas en dockinguppgift miljontals gånger, vilket gör optimeringsinsatser nödvändiga. Framför allt moderna heterogena arkitekturer som GPU:er ger betydande accelerationsmöjligheter. AutoDock-GPU, en modern GPU-accelererad programvara för molekylär dockning, uppskattar den geometriska konformationen hos ett ligand-protein-komplex genom att minimera en energibaserad poängsättningsfunktion. Våra profileringsresultat visade att en reduktionsoperation, som utförs flera miljoner gånger i en enda dockningskörning, begränsar prestandan i AutoDock-GPU. Vi har därför föreslagit en metod för att accelerera summareduktionen på blocknivå av vektorer med fyra element med hjälp av matrisoperationer. Vi implementerade vår metod för att utnyttja den höga genomströmningskapacitet som erbjuds av NVIDIA Tensor Cores för att utföra matrisoperationer. Vi utvärderade vårt tillvägagångssätt genom att utforma ett enkelt testfall och uppnådde en 4- till 7-faldig förbättring av körtiden jämfört med den ursprungliga metoden. Vi integrerade sedan vår reduktionsoperation i AutoDock-GPU och utvärderade den på flera kemiska komplex på tre GPU:er. Denna utvärdering lät oss bedöma möjligheten att använda reduktionsoperationer med halvprecision i delar av AutoDock-GPU-koden, utan negativa effekter på simuleringsresultatet. Dessutom uppnådde vår version en genomsnittlig förbättring på 27% av den totala dockningstiden under en riktig dockningskörning.

No description available.

No description available.

The BioChemical Library (BCL) cheminformatics toolkit is an application-based academic open-source software package designed to integrate traditional small molecule cheminformatics tools with machine learning-based quantitative structure-activity/ property relationship (QSAR/QSPR) modeling. In this pedagogical article we provide a detailed introduction to core BCL cheminformatics functionality, showing how traditional tasks (e.g., computing chemical properties, estimating druglikeness) can be readily combined with machine learning. In addition, we have included multiple examples covering areas of advanced use, such as reaction-based library design. We anticipate that this manuscript will be a valuable resource for researchers in computer-aided drug discovery looking to integrate modular cheminformatics and machine learning tools into their pipelines.