De Novo Design Research Articles

Mixture-of-Experts Based Dissociation Kinetic Model for De Novo Design of HSP90 Inhibitors with Prolonged Residence Time.

The dissociation rate constant (koff) significantly impacts the drug potency and dosing frequency. This work proposes a powerful optimization-based framework for de novo drug design guided by koff. First, a comprehensive database containing 2,773 unique koff values is created. Based on the database, a novel generic dissociation kinetic model is developed with a mixture-of-experts architecture, enabling high-throughput predictions of koff with high accuracy. The developed model is then integrated with an optimization-based mathematical programming approach to design drug candidates with low koff. Finally, the τ-RAMD method is utilized to rigorously verify the designed potential drug candidates. In a case study, the framework successfully identified numerous new potential HSP90 inhibitor candidates, achieving a maximum 45.7% improvement in residence time (τ = 1/koff) compared to that of a known exceptional HSP90 inhibitor. These findings demonstrate the feasibility and effectiveness of the kinetics-guided optimization-based de novo drug design framework in designing drug candidates with prolonged τ.

The Role of Alkyl Chains in the Thermoresponse of Supramolecular Network.

Supramolecular materials provide a pathway for achieving precise, highly ordered structures while exhibiting remarkable response to external stimuli, a characteristic not commonly found in covalently bonded materials. The design of self-assembled materials, where properties could be predicted/design from chemical nature of the individual building blocks, hinges upon ourability to relate macroscopic properties to individual building blocks - a feat which has thus far remained elusive. Here, a design approach is demonstrated to chemically engineer the thermal expansion coefficient of 2D supramolecular networks by over an order of magnitude (\boldmath 120 to \boldmath 1000 × 10-6K-1). This systematic study provides a clear pathway on how to carefully design the thermal expansion coefficient of a 2D molecular assembly. Specifically, a linear relation has been identified between the length of decorating alkyl chains and the thermal expansion coefficient. Counter-intuitively, the shorter the chains the larger is the thermal expansion coefficient. This precise control over thermo-mechanical properties marks a significant leap forward in the de-novo design of advanced 2D materials. The possibility to chemically engineer their thermo-mechanical properties holds promise for innovations in sensors, actuators, and responsive materials across diversefields.

De Novo Design and Characterization of Novel Antimicrobial Peptides, Lk3 and Lk4, Containing Repeating Unit

De Novo Design and Characterization of Novel Antimicrobial Peptides, Lk3 and Lk4, Containing Repeating Unit

De Novo Design and Characterization of Amphiphilic Peptides with Basic Side Chains for Tailored Interfacial Chemistries.

Lysine-leucine (LK) peptides have been used as model systems and platforms for 2D material design for decades. LK peptides are amphiphilic sequences designed to bind and fold at hydrophobic surfaces through hydrophobic leucine side chains and hydrophilic lysine side chains extending into the aqueous subphase. The hydrophobic periodicity of the sequence dictates the secondary structure at the interface. This robust design makes them ideal candidates for controlling interfacial chemistry. This study presents the de novo design and characterization of two novel peptides: LRα14 and LHα14, which substitute lysine with arginine and histidine, respectively, in the helical LKα14 sequence. This modification is intended to expand the LK peptide platform to a new basic interfacial chemistry. We explore the stability of the new LRα14 and LHα14 designs with respect to changes in pH and salt concentration in bulk solution and at the interface using circular dichroism (UV-CD) and vibrational sum-frequency generation spectroscopy, respectively. Notably, the structural stability of the peptides remains unaffected across a wide range of pH and ionic strength values. At the same time, the variation of side-chain chemistry leads to a wide spectrum of interfacial water structures. By extension of the LK platform to include arginine and histidine, this study broadens the toolbox for designing tailored interfacial chemistries with applications in material and biomedical sciences.

De Novo Design of Peptide Binders to Conformationally Diverse Targets with Contrastive Language Modeling.

Designing binders to target undruggable proteins presents a formidable challenge in drug discovery, requiring innovative approaches to overcome the lack of putative binding sites. Recently, generative models have been trained to design binding proteins via three-dimensional structures of target proteins, but as a result, struggle to design binders to disordered or conformationally unstable targets. In this work, we provide a generalizable algorithmic framework to design short, target-binding linear peptides, requiring only the amino acid sequence of the target protein. To do this, we propose a process to generate naturalistic peptide candidates through Gaussian perturbation of the peptidic latent space of the ESM-2 protein language model, and subsequently screen these novel linear sequences for target-selective interaction activity via a CLIP-based contrastive learning architecture. By integrating these generative and discriminative steps, we create a Peptide Prioritization via CLIP (PepPrCLIP) pipeline and validate highly-ranked, target-specific peptides experimentally, both as inhibitory peptides and as fusions to E3 ubiquitin ligase domains, demonstrating functionally potent binding and degradation of conformationally diverse protein targets in vitro. Overall, our design strategy provides a modular toolkit for designing short binding linear peptides to any target protein without the reliance on stable and ordered tertiary structure, enabling generation of programmable modulators to undruggable and disordered proteins such as transcription factors and fusion oncoproteins.

De Novo Design of Inhibitors of DNA Methyltransferase 1: A Critical Comparison of Ligand- and Structure-Based Approaches.

Designing and developing inhibitors against the epigenetic target DNA methyltransferase (DNMT) is an attractive strategy in epigenetic drug discovery. DNMT1 is one of the epigenetic enzymes with significant clinical relevance. Structure-based de novo design is a drug discovery strategy that was used in combination with similarity searching to identify a novel DNMT inhibitor with a novel chemical scaffold and warrants further exploration. This study aimed to continue exploring the potential of de novo design to build epigenetic-focused libraries targeted toward DNMT1. Herein, we report the results of an in-depth and critical comparison of ligand- and structure-based de novo design of screening libraries focused on DNMT1. The newly designed chemical libraries focused on DNMT1 are freely available on GitHub.

De Novo Design and Facile Synthesis of Highly Crystalline 2D Conductive Metal-Organic Frameworks: A "Rotor-Stator" Strategy.

Two-dimensional conductive metal-organic frameworks (2D c-MOFs), which feature high electrical conductivity and large charge carrier mobility, hold great promise in electronics and optoelectronics. Nevertheless, the limited solubility of commonly used planar ligands inevitably brings challenges in synthesis and purification and causes laborious coordination conditions for screening. Moreover, most reported 2D c-MOFs are polycrystalline powders with relatively low crystallinity and irregular morphology, hindering the unveiling of the detailed structure-function relationship. Herein, we developed a "rotor-stator" molecular design strategy to construct 2D c-MOFs using a delicately designed nonplanar biscarbazole ligand (8OH-DCB). Benefiting from the special "rotor-stator" structure of the ligand, crystals of Cu-DCB-MOF were successfully prepared, allowing for the precise determination of their crystal structure. Interestingly, the crystals of Cu-DCB-MOF can be obtained in various organic solvents, indicating excellent solvent compatibility. The versatility of the "rotor-stator" molecular design strategy was further demonstrated by another two new ligands with a "rotor-stator" structure, and afford corresponding 2D c-MOF crystals (Cu-DCBT-MOF and Cu-DCBBT-MOF). The current work presents a facile approach toward the rational design and direct construction of highly crystalline 2D c-MOFs using nonplanar ligands.

PROTACable Is an Integrative Computational Pipeline of 3-D Modeling and Deep Learning To Automate the De Novo Design of PROTACs.

Proteolysis-targeting chimeras (PROTACs) that engage two biological targets at once are a promising technology in degrading clinically relevant protein targets. Since factors that influence the biological activities of PROTACs are more complex than those of a small molecule drug, we explored a combination of computational chemistry and deep learning strategies to forecast PROTAC activity and enable automated design. A new method named PROTACable was developed for the de novo design of PROTACs, which includes a robust 3-D modeling workflow to model PROTAC ternary complexes using a library of E3 ligase and linker and an SE(3)-equivariant graph transformer network to predict the activity of newly designed PROTACs. PROTACable is available at https://github.com/giaguaro/PROTACable/.

De Novo Design of a Highly Stable Ratiometric Probe for Long-Term Continuous Imaging of Endogenous HClO Burst.

Long-term continuous imaging of endogenous HClO burst is of great importance for the elucidation of various physiological or pathological processes. However, most of the currently reported HClO probes have failed to achieve this goal due to their insufficient photobleaching resistance under a laser source. Herein, a highly stable ratiometric probe, HFTC-HClO 1, which is capable of continuously monitoring endogenous HClO burst over a long period of time, has been judiciously developed. Briefly, the de novo development of HFTC-HClO 1 mainly involved three main steps: (1) novel coumarins (HFTC 1-5) were designed and synthesized; (2) the most stable scaffold, HFTC 3, was selected through dye screening and cell imaging validation; and (3) based on HFTC 3, three candidate HClO probes were constructed, and HFTC-HClO 1 was finally selected due to its superior sensing properties toward HClO. Furthermore, HFTC-HClO 1 can quantitatively measure HClO levels in various real samples with excellent recovery (>90.4%), and the use of HFTC-HClO 1-coated test strips for qualitative analysis of HClO in real samples was also achieved. In addition, the application of HFTC-HClO 1 for long-term continuous monitoring of intracellular HClO burst was successfully demonstrated. Significantly, HFTC-HClO 1 was able to visualize HClO generated in the rheumatoid arthritis mouse model.

De Novo Design of Near-Infrared Fluorescent Agents Activated by Peroxynitrite and Glutathione-Responsive Imaging for Diabetic Liver Disease.

Diabetes and its complications, such as diabetes liver disease, is a major problem puzzling people's health. The detection of redox states in its pathological process can effectively help us gain a deeper understanding of the disease. The pair of oxidation-reduction substances peroxynitrite (ONOO- ) and glutathione (GSH) is considered to be closely related to their occurrence and development. Thus, direct visualization of ONOO- and GSH in diabetes liver disease is critical to evaluate the disease at the molecular level. Herein, two activatable agents NTCF-ONOO- and NTCF-GSH are prepared for selectively detecting ONOO- and GSH through protection and deprotection strategies based on hydroxyl and amino groups of near-infrared fluorophore. Fluorescence imaging of exogenous and endogenous ONOO- and GSH changes in living cells and in vivo is observed. The ONOO- and GSH level in the diabetes liver disease cellular model are visualized and the possible redox imbalance mechanism related to the oxidized (NAD+ ) and reduced (NADH) nicotinamide adenine dinucleotides is explored in this process. Moreover, these probes can sensitively recognize ONOO- and GSH in the process of oxidative stress resulting from streptozotocin and streptozotocin/acetaminophen-induced complex diabetic liver disease in vivo. In addition, they can be applied for monitoring the clinical serum sample related with diabetic patients.

Artificial intelligence based de-novo design for novel Plasmodium falciparum plasmepsin (PM) X inhibitors

Plasmodium falciparum is the leading cause of malaria with 627,000 deaths annually. Invasion and egress are critical stages for successful infection of the host yet depend on proteins that are extensively pre-processed by various maturases. Plasmepsins (Plasmodium pepsins, abbreviated PM, I-X) are pepsin-like aspartic proteases that are involved in almost all stages of the life cycle. The goal of this study was to use de-novo generative modeling techniques to create novel potential PfPMX inhibitors. A total of 4325 compounds were virtually screened by structural-based docking methods. The obtained hits were utilized to refine a structure-based Ligand Neural Network (L-Net) generative model to generate related compounds. The obtained optimal L-Net Compounds with smina scores ≤ −5.00KCalmol-1 and QED ≥ 0.35 were further taken for amplification utilizing Ligand Based Transformer modeling using Deep generative learning (Drug Explorer/DrugEx). The resulting hits were then subjected to XP Glide conventional Molecular docking and QikProp ADMET screening; molecules with XP Docking score ≤ −7.00KCalmol-1 were retained. Based on their Glide ligand efficiency, originality, and uniqueness, 30 compounds were chosen for binding affinity and MM_GBSA energy determination. Following Induced Fit docking (IFD), 7 compounds were taken for 50 ns MD simulations and FEP/MD calculations. This study reported novel potential PfPMX inhibitors with acceptable ADMET profiles and reasonable synthetic accessibility scores, as well as sufficient docking scores against other PMs were generated. The PfPMX inhibitors reported in this article are promising antimalarials for the next stages of drug development, and the first of their kind to be investigated thoroughly. Communicated by Ramaswamy H. Sarma

Noninvasive Imaging of Tumor Glycolysis and Chemotherapeutic Resistance via De Novo Design of Molecular Afterglow Scaffold.

Chemotherapeutic resistance poses a significant challenge in cancer treatment, resulting in the reduced efficacy of standard chemotherapeutic agents. Abnormal metabolism, particularly increased anaerobic glycolysis, has been identified as a major contributing factor to chemotherapeutic resistance. To address this issue, noninvasive imaging techniques capable of visualizing tumor glycolysis are crucial. However, the currently available methods (such as PET, MRI, and fluorescence) possess limitations in terms of sensitivity, safety, dynamic imaging capability, and autofluorescence. Here, we present the de novo design of a unique afterglow molecular scaffold based on hemicyanine and rhodamine dyes, which holds promise for low-background optical imaging. In contrast to previous designs, this scaffold exhibits responsive "OFF-ON" afterglow signals through spirocyclization, thus enabling simultaneous control of photodynamic effects and luminescence efficacy. This leads to a larger dynamic range, broader detection range, higher signal enhancement ratio, and higher sensitivity. Furthermore, the integration of multiple functionalities simplifies probe design, eliminates the need for spectral overlap, and enhances reliability. Moreover, we have expanded the applications of this afterglow molecular scaffold by developing various probes for different molecular targets. Notably, we developed a water-soluble pH-responsive afterglow nanoprobe for visualizing glycolysis in living mice. This nanoprobe monitors the effects of glycolytic inhibitors or oxidative phosphorylation inhibitors on tumor glycolysis, providing a valuable tool for evaluating the tumor cell sensitivity to these inhibitors. Therefore, the new afterglow molecular scaffold presents a promising approach for understanding tumor metabolism, monitoring chemotherapeutic resistance, and guiding precision medicine in the future.

From De Novo Design to Redesign: Harnessing Computational Protein Design for Understanding SARS-CoV-2 Molecular Mechanisms and Developing Therapeutics.

The continuous emergence of novel SARS-CoV-2 variants and subvariants serves as compelling evidence that COVID-19 is an ongoing concern. The swift, well-coordinated response to the pandemic highlights how technological advancements can accelerate the detection, monitoring, and treatment of the disease. Robust surveillance systems have been established to understand the clinical characteristics of new variants, although the unpredictable nature of these variants presents significant challenges. Some variants have shown resistance to current treatments, but innovative technologies like computational protein design (CPD) offer promising solutions and versatile therapeutics against SARS-CoV-2. Advances in computing power, coupled with open-source platforms like AlphaFold and RFdiffusion (employing deep neural network and diffusion generative models), among many others, have accelerated the design of protein therapeutics with precise structures and intended functions. CPD has played a pivotal role in developing peptide inhibitors, mini proteins, protein mimics, decoy receptors, nanobodies, monoclonal antibodies, identifying drug-resistance mutations, and even redesigning native SARS-CoV-2 proteins. Pending regulatory approval, these designed therapies hold the potential for a lasting impact on human health and sustainability. As SARS-CoV-2 continues to evolve, use of such technologies enables the ongoing development of alternative strategies, thus equipping us for the "New Normal".

The rise of the sourdough: Genome-scale metabolic modeling-based approach to design sourdough starter communities with tailored-made properties

Sourdough starters harbor microbial consortia that benefit the final product's aroma and volume. The complex nature of these spontaneously developed communities raises challenges in predicting the fermentation phenotypes. Herein, we demonstrated for the first time in this field the potential of genome-scale metabolic modeling (GEMs) in the study of sourdough microbial communities. Broad in-silico modeling of microbial growth was applied on communities composed of yeast (Saccharomyces cerevisiae) and different Lactic Acid Bacteria (LAB) species, which mainly predominate in sourdough starters. Simulations of model-represented communities associated specific bacterial compositions with sourdough phenotypes. Based on ranking the phenotypic performances of different combinations, Pediococcus spp. – Lb. sakei group members were predicted to have an optimal effect considering the increase in S. cerevisiae growth abilities and overall CO2 secretion rates. Flux Balance Analysis (FBA) revealed mutual relationships between the Pediococcus spp. – Lb. sakei group members and S. cerevisiae through bidirectional nutrient dependencies, and further underlined that these bacteria compete with the yeast over nutrients to a lesser extent than the rest LAB species. Volatile compounds (VOCs) production was further modeled, identifying species-specific and community-related VOCs production profiles. The in-silico models' predictions were validated by experimentally building synthetic sourdough communities and assessing the fermentation phenotypes. The Pediococcus spp. – Lb. sakei group was indeed associated with increased yeast cell counts and fermentation rates, demonstrating a 25 % increase in the average leavening rates during the first 10 fermentation hours compared to communities with a lower representation of these group members. Overall, these results provide a possible novel strategy towards the de-novo design of sourdough starter communities with tailored-made characterizations, including a shortened leavening period.

De Novo Design and Development of a Nutrient-Rich Perfusate for Ex Vivo Lung Perfusion with Cell Culture Models.

Ex vivo lung perfusion (EVLP) has increased donor lung utilization through assessment of "marginal" lungs prior to transplantation. To develop it as a donor lung reconditioning platform, prolonged EVLP is necessary, and new perfusates are required to provide sufficient nutritional support. Human pulmonary microvascular endothelial cells and epithelial cells were used to test different formulas for basic cellular function. A selected formula was further tested on an EVLP cell culture model, and cell confluence, apoptosis, and GSH and HSP70 levels were measured. When a cell culture medium (DMEM) was mixed with a current EVLP perfusate-Steen solution, DMEM enhanced cell confluence and migration and reduced apoptosis in a dose-dependent manner. A new EVLP perfusate was designed and tested based on DMEM. The final formula contains 5 g/L Dextran-40 and 7% albumin and is named as D05D7A solution. It inhibited cold static storage and warm reperfusion-induced cell apoptosis, improved cell confluence, and enhanced GSH and HSP70 levels in human lung cells compared to Steen solution. DMEM-based nutrient-rich EVLP perfusate could be a promising formula to prolong EVLP and support donor lung repair, reconditioning and further improve donor lung quality and quantity for transplantation with better clinical outcome.

GENERA: A Combined Genetic/Deep-Learning Algorithm for Multiobjective Target-Oriented De Novo Design.

This study introduces a new de novo design algorithm called GENERA that combines the capabilities of a deep-learning algorithm for automated drug-like analogue design, called DeLA-Drug, with a genetic algorithm for generating molecules with desired target-oriented properties. Specifically, GENERA was applied to the angiotensin-converting enzyme 2 (ACE2) target, which is implicated in many pathological conditions, including COVID-19. The ability of GENERA to de novo design promising candidates for a specific target was assessed using two docking programs, PLANTS and GLIDE. A fitness function based on the Pareto dominance resulting from computed PLANTS and GLIDE scores was applied to demonstrate the algorithm's ability to perform multiobjective optimizations effectively. GENERA can quickly generate focused libraries that produce better scores compared to a starting set of known ACE-2 binders. This study is the first to utilize a DL-based algorithm designed for analogue generation as a mutational operator within a GA framework, representing an innovative approach to target-oriented de novo design.

MET-Targeting Anticancer Drugs—De Novo Design and Identification by Drug Repurposing

The Met protein is a cell surface receptor tyrosine kinase predominantly expressed in epithelial cells. Aberrant regulation of MET is manifested by numerous mechanisms including amplification, mutations, deletion, fusion of the MET proto-oncogene, and protein overexpression. They represent the common causes of drug resistance to conventional and targeted chemotherapy in numerous cancer types. There is also accumulating evidence that MET/HGF signaling drives an immunosuppressive tumor microenvironment and dampens the efficacy of cancer immunotherapy. Substantial research effort has been invested in designing Met-targeting drugs with different mechanisms of action. In this review, we summarized the current preclinical and clinical research about the development of Met-targeting drugs for cancer therapeutics. Early attempts to evaluate Met-targeted therapies in clinical trials without selecting the appropriate patient population did not produce satisfactory outcomes. In the era of personalized medicine, cancer patients harboring MET exon 14 alterations or MET amplification have been found to respond well to Met-inhibitor therapy. The application of Met inhibitors to overcome drug resistance in cancer patients is discussed in this paper. Given that kinases play critical roles in cancer development, numerous kinase-mediated signaling pathways are attractive targets for cancer therapy. Existing kinase inhibitors have also been repurposed to new kinase targets or new indications in cancer. On the other hand, non-oncology drugs have also been repurposed for treating cancer through kinase inhibition as one of their reported anticancer mechanisms.

De Novo Design of Polyhedral Protein Assemblies: Before and After the AI Revolution.

Self-assembling polyhedral protein biomaterials have gained attention as engineering targets owing to their naturally evolved sophisticated functions, ranging from protecting macromolecules from the environment to spatially controlling biochemical reactions. Precise computational design of de novo protein polyhedra is possible through two main types of approaches: methods from first principles, using physical and geometrical rules, and more recent data-driven methods based on artificial intelligence (AI), including deep learning (DL). Here, we retrospect first principle- and AI-based approaches for designing finite polyhedral protein assemblies, as well as advances in the structure prediction of such assemblies. We further highlight the possible applications of these materials and explore how the presented approaches can be combined to overcome current challenges and to advance the design of functional protein-based biomaterials.

De Novo Design of Aggregation‐Induced Emission Luminogen for Three‐Photon Fluorescence Imaging of Subcortical Structures Excited at Both NIR‐III and NIR‐IV Windows

AbstractThree‐photon fluorescence (3PF) imaging excited at 1700 nm window is an enabling technology for visualizing deep brain structures and dynamics. Recently, the 2200 nm window has emerged as the longest excitation window suitable for deep‐brain 3PF imaging. Bright fluorescent probes lay the material basis for deep‐brain 3PF imaging. Among various fluorescent probes, aggregation‐induced emission luminogens (AIEgens) have great potential in 3PF imaging excited at the 1700 nm window in vivo. However, to the best of knowledge, there is no AIEgens applicable to 3PF imaging excited at both the 1700 and 2200 nm windows. To readily fill this gap, here this study designs and synthesizes a novel AIEgen, namely TPE‐DPTT‐ICP, which generates bright 3PF signals excited at both 1700 and 2200 nm. The accordingly fabricated TPE‐DPTT‐ICP nanoparticles (NPs) possess excellent water dispersibility, colloidal stability, biocompatibility, photostability and large 3P action cross section, key to in vivo imaging. In mouse brain in vivo, TPE‐DPTT‐ICP NPs enable deep‐brain 3PF imaging of subcortical structures excited at both the two windows, reaching depths of 1640 and 880 µm below the brain surface, respectively. TPE‐DPTT‐ICP NPs are thus a versatile material simultaneously catering to the need at two infrared optical windows with deep tissue penetration.

De Novo Design of Molecules with Multiaction Potential from Differential Gene Expression using Variational Autoencoder.

The modulating effect of chemical compounds and therapeutics on gene transcription is well-reported and has been intensively studied for both clinical and research purposes. Emerging research points toward the utility of drug-induced transcriptional alterations in de novo molecular design and highlights the idea of phenotype-matching an expression signature of interest to the structures being designed. In this work, we build an autoencoder-based generative model, BiCEV, around this concept. Our generative autoencoder has demonstrably generated a set of new molecules from gene expression input with notable validity (96%), uniqueness (98%), and internal diversity (0.77). Further, we attempted to validate BiCEV by testing the model on gene-knockdown profiles and combined signatures of synergistic drug pairs. From these investigations, we found the designed structures to be consistently high in collective quality. However, when their similarities to the supposed functional equivalents as determined by shared targets were considered, the findings were somewhat mixed. In spite of this, we believe the generative model merits further development in conjunction with in vitro corroboration to lend itself to being an assistive tool for drug discovery experts, particularly to support the initial stages of hit identification and lead optimization.