Congbao Kang ,  Hung T. Nguyen ,  David E. Heppner ,  Bin Yu ,  Weijun Xu

Journal of Medicinal Chemistry

DOI: 10.1021/acs.jmedchem.6c00070

Over the past decade, RNA has emerged as an attractive and evolving landscape for small-molecule drug discovery. RNA functions as an intermediate macromolecule during gene expression and plays a diverse role in regulating cellular processes. (1) The promise in targeting RNA as therapeutic interventions arises from recent breakthroughs in structural biology, chemical biology, and computational modeling that collectively facilitate understanding of RNA biology. Multiple classes of RNA have been identified contributing to their functional diversity, including messenger RNA (mRNA), transfer RNA (tRNA), ribosomal RNA (rRNA), long noncoding RNA (lncRNA), microRNA (miRNA), and other noncoding RNAs (ncRNAs). While only a tiny fraction of 1.5% of human genome is translated into proteins, approximately 75% is transcribed into RNA, (2−4) making RNA a vastly larger reservoir of potential drug targets than the traditional protein targets. Moreover, many disease-associated pathways including viral replication, cancer progression, neurodegeneration, and immune regulation are regulated by RNA or through its interactions with different proteins, making RNA an increasingly relevant and strategic focus for next-generation drug discovery. (5) Several classes of small molecules define the landscape of RNA-targeting drug discovery (Figure 1), and many other modulators of RNA have been discovered for further development. (6) Small molecules targeting RNA can act through diverse mechanisms, including direct RNA binding to modulate RNA function, disruption of RNA interactions with RNA binding proteins, and stimulation RNA binding to RNA degrading enzyme for degradation. (7) The success in developing RNA drugs and RNA ligands prove the feasibility of RNA-focused drug discovery. (2) Patients with spinal muscular atrophy (SMA) lack enough survival motor neuron (SMN) protein to maintain adequate muscle function. The lack of SMN protein is believed to drive the pathophysiology of SMA. In August 2020, the FDA approved risdiplam (Evrysdi, RG7916, RO7034067) developed by Roche for the treatment of SMA in patients of all ages. (8) As a small-molecule modulator of SMN2 splicing, risdiplam represents the first FDA-approved RNA-targeting small-molecule drug. Mechanically, risdiplam acts as a “glue”-like compound that enhances RNA binding protein (RBP) to RNA to facilitate splicing and increases the production of full-length functional SMN protein. In addition to risdiplam, there are other drugs or clinical candidates that directly bind to RNA, including small molecules targeting RNA G-quadruplexes such as Quarfloxin and CX-5461, ribosome-targeting antibiotics like macrolides, and viral RNA polymerase nucleosides like sofosbuvir. (9) Together, these precedences highlight the growing potential of RNA as a therapeutic target and lay the foundation for the rational discovery of small-molecule RNA modulators. Figure 1. Small-molecule drugs that target RNAs. While strong incentives have been directed to drug RNA, several fundamental challenges make RNA-targeted drug discovery inherently more intractable than conventional protein-directed approaches. (10) Although proteins can undergo complicated conformational changes and post-translational modifications, RNA is even more dynamic and flexible, adopting distinct conformations that depend on its sequence, cellular environment, and interactions with other biomolecules. (11) Said another way, the highly dynamic nature of RNA implies that robust pockets may be impossible to bind small molecules like unstructured regions of proteins. Unlike proteins formed by 20 amino acids, RNA is composed of only four nucleotides, presenting challenges for specific and tight interactions with small molecules. Additional hurdles include the limited identification of clearly druggable RNA motifs, the intrinsic structural dynamics of RNA, and the difficulty in achieving selectivity across closely related RNA sequences. Additionally, traditional protein-directed drug discovery often relies on both polar and nonpolar interactions within well-defined hydrophobic and hydrophilic pockets. By contrast, RNA features overwhelmingly negatively charged backbones and typically lacks hydrophobic pockets. As a result, the number and structural diversity of druggable RNA binding sites are expected to be limited, which further complicates efforts to achieve selective modulation. As such, the structural and functional nature of RNA intrinsically favors chemical space dominated by highly polar molecules manifesting to at least two major issues: limited selectivity due to the abundance of similarly polar features across cellular RNAs and significant challenges in optimizing ADME properties, particularly lipophilicity. As with drug discovery of protein targets, hit identification via high-throughput screening (HTS), fragment-based screening, and virtual screening are commonly adopted. A major challenge for HTS against RNA targets is that most existing HTS libraries were designed for proteins and are poorly suited for recognizing the unique physicochemical and structural characteristics of RNA. Therefore, the development of RNA-focused libraries has become a key priority, which requires pharmacophore-based strategies through incorporating scaffolds known to interact with RNA motifs. This can be further complemented by the rise of DNA-encoded libraries (DELs) that have emerged as powerful source for hit identification for RNA targets. (4) For RNA to become a more tractable and widely druggable class of biomolecules, advances are needed in generating well-behaved biochemical tools and obtaining detailed structural information on unique or transient RNA conformations. Such innovations will be essential for enabling robust discovery platforms analogous to those that have long supported successful drug development for proteins and enzymes. Despite the above-mentioned challenges, RNA-targeted therapeutics continue to gain momentum across oncology, virology, and neurology. Recent advances in structural biology and computation, artificial intelligence, and multiscale modeling have begun to overcome these barriers and pave a rational foundation for RNA-targeted drug discovery. Structural studies of RNA using chemical biology and biophysical methods have made progress in recent years, providing detailed insights into secondary and tertiary motifs such as hairpins, internal loops, bulges, and pseudoknots, which are considered important structural architectures for developing small molecules. NMR spectroscopy remains a powerful technique for probing RNA structure and dynamics in solution. (12) Its ability to resolve conformational equilibria, detect transient states, and characterize ligand binding has made it indispensable for understanding RNA functional motions and identifying druggable conformations. X-ray crystallography, usually hindered by the intrinsic flexibility of RNA, has become increasingly feasible due to improved construct design, presence of stabilizing ligands, and the use of RBPs to facilitate crystallization. These innovations have enabled high-resolution structures of diverse RNA motifs and RNA–ligand complexes to be solved. (13) Cryo-EM has transformed the field through efficient determination of large RNA molecules and RNA–protein complexes that were previously inaccessible. (14) Complementing experimental approaches, computational methods have evolved rapidly and played a critical role in RNA structural biology and drug discovery. (15) Several algorithms and servers have been developed to predict RNA secondary and tertiary structures, model conformational ensembles, binding pocket identification and screen compound libraries against RNA targets. (16) Machine learning based methods that are supported by continuously expanding high-quality structural data, have improved accuracy in predicting RNA structures, RNA–RBP interactions and RNA–ligand interactions. Collectively, these methods provide a robust foundation for rational discovery of RNA modulators. Advances in physics-based molecular dynamics (MD) simulations are now mapping RNA conformational landscapes to reveal hidden and metastable pockets. Enhanced sampling MD, (17) coarse-grained (CG) modeling, (18) and Markov state models (MSMs) (19) allow the identification of transient states that are invisible to standard spectroscopy and experimental determination. (20) Coupled with ensemble-based docking against multiple conformers, hit rates may improve compared to static docking against RNA targets. Furthermore, the integration of hybrid QM/MM and machine-learning (ML) corrected scoring functions is providing a more rigorous treatment of stacking interactions, hydration shells, and ion-mediated electrostatics, which are critical for accurate RNA-ligand affinity prediction. These methods are increasingly used to derive druggability maps for riboswitches, repeat RNAs, viral elements, and structured motifs within long non-coding RNAs. (21) On the other hand, AI-driven prediction and generative design for RNA-targeted chemistry is on the rise. However, a central barrier is the limited availability of high-resolution RNA–ligand structures, which constrains traditional structure-based design. Emerging AI approaches are addressing this gap through multimodal learning frameworks that integrate RNA sequence, chemical features, SHAPE/DMS reactivity, and evolutionary covariation. Deep learning models improve secondary and tertiary structure inference, enabling more accurate identification of ligandable motifs and conformational states relevant for binding. (22) Recently, prediction of small-molecule–RNA interactions was achieved without the need for RNA tertiary structures as input. (23) However, their broader utility in RNA-focused medicinal chemistry remains to be fully established and will likely require larger data sets and more rigorous experimental validation. Moreover, many disease-relevant RNAs function through multivalent interactions with RBPs. Small molecules that restore normal RNA–RBP equilibrium have demonstrated proof-of-principle activity in correcting splicing, transcriptional regulation, and translation. (24) It remains to be seen whether disrupting RNA–protein binding interaction would bring clinical benefits. As data sets from CLIP–seq, RNP–MaP, and ligand–RNA cross-linking expand, predictive modeling of RBP selectivity will continue to improve. Last but not least, a rapidly growing therapeutic direction targets ribonucleoprotein condensates, that feature dynamic assemblies whose physicochemical properties (viscoelasticity, aging kinetics, etc.) encode key regulatory functions. (25) Aberrant condensate behavior is implicated in many diseases including neurodegenerative diseases and cancer. (25) Small molecules that alter condensate properties, such as softening, hardening, dissolving, or preventing gelation, represent a new modality of RNA-focused therapeutics. (26) Computational modeling is expected to play a central role capturing the structures of mesoscale organization and emergent behaviors of ribonucleoprotein condensates. These approaches collectively support rational design of condensate-modulating drugs, a modality fundamentally different from classical lock-and-key pharmacology. Overall, deep understanding of RNA targets, innovative screening and validation as a collaborative effort from chemical biology and medicinal chemistry is key to the success of RNA targeted drug discovery campaigns. With rapidly advancing techniques in structural biology and artificial intelligence are being developed, we may witness whether small molecule drug discovery targeting RNAs will meet its inflection point or remain a niche curiosity in the years to come. This article references 26 other publications. (acccessed 2024/10/22) PubMed (acccessed 2026/01/04) (acccessed 2026/01/04) This article has not yet been cited by other publications.

Xianfeng Li ,  Zehao Yin ,  Qiuyi Chen ,  Xinlong Hu ,  Gong Zhang ,  Xiaohong Fan ,  Yizhou Li

Organic Letters 

DOI: 10.1021/acs.orglett.6c00490

Abstract

The β-ketoamide motif represents both a privileged scaffold and a versatile synthetic intermediate in medicinal chemistry. Herein, we developed a DNA-compatible method for the efficient conversion of various DNA-conjugated amines into β-ketoamides. The resulting β-ketoamides facilitate rapid diversification into a panel of structurally diverse molecular scaffolds. Importantly, the synthetic route and subsequent derivatization steps were validated to be fully compatible with DNA encoding, offering a reliable and versatile platform for DNA-encoded library synthesis.

Yulong An ,  Ruolan Zhou ,  Xiang Li

ACS Medicinal Chemistry Letters 

DOI: 10.1021/acsmedchemlett.5c00738

Abstract

DNA-encoded library (DEL) technology has emerged as a transformative platform for discovering chemical inducers of proximity (CIPs), addressing challenges in both degrader and non-degrader CIP development. This Microperspective analyzes the results of recent DEL technology screens (2021–2025) to enable medicinal chemistry programs, focusing on CIP development including CIP-focused DELs, DEL-derived ligands for proteins of interest (POIs) and E3 ligase in rational CIP design, and directly functional CIP identification. Finally, we address current limitations of DEL technology in CIP research and outline future directions. This Microperspective underscores DEL’s pivotal role in advancing CIP discovery, providing actionable insights for addressing “undruggable” targets and accelerating translational research in chemical biology and medicinal chemistry.

Summary

This MicroPerspective reviews recent advances (2021–2025) in DNA-encoded library (DEL) technology for discovering chemical inducers of proximity (CIPs), spanning both degraders (e.g., PROTACs, molecular glue degraders) and non-degrader modalities (e.g., protein stabilization, subcellular relocalization, transcriptional activation). It synthesizes three key strategies: (1) CIP-focused DELs (CIP-DELs), enabling simultaneous dual-target (POI + E3 ligase) selection to directly identify cooperatively binding bifunctional compounds; (2) Conversion of DEL-derived POI/E3 ligands—leveraging well-defined DNA attachment sites as “exit vectors”—into functional CIPs; and (3) Discovery of non-degradative CIPs, including FKBP12-recruiting molecular glues and function-driven DEL screening (e.g., direct ubiquitination readout). DEL overcomes longstanding limitations of traditional HTS—including library size, cost, and scarcity of E3 ligands—thereby accelerating CIP development against “undruggable” targets.

Highlights

• Dual-Target CIP-DEL Screening: CRBN- or VHL-targeted DELs enable concurrent selection against POIs and E3 ligases, directly identifying ternary complex stabilizers with high cooperativity (e.g., BRD4/BRD2-selective PROTACs, BRD9 molecular glue).
• Ligand-to-CIP Conversion Paradigm: DEL-derived ligands for ERα, MAGE-A3, PIN1, DNPH1, and TRIM21 were optimized and converted into functional PROTACs or TrimTACs; the DNA attachment site serves as a built-in, precise “exit vector” for linker conjugation.
• Expansion to Non-Degradative Functions: An FKBP12-biased CIP-DEL identified a molecular glue that stabilizes the Crohn’s disease-associated ATG16L1 T300A variant; function-driven DEL screening (in presence of E1/E2/ATP) directly enriches ubiquitination-competent PROTACs, eliminating affinity-only false positives.
• New Frontier: RNA Targets: DEL screening against RNase L led to the design of RiboTACs targeting pre-miR-21—extending CIP therapeutics to RNA biology.

Conclusion

DEL technology has evolved from a single-target ligand discovery platform into a central engine driving the discovery of the full spectrum of CIPs—from degraders to non-degraders, and from proteins to RNA. Its core advantages lie in vast chemical space coverage, barcode-enabled precise hit identification, and intrinsic structural information (e.g., defined exit vectors). Future directions include integrating AI for POI–E3 interface–guided library design, developing robust in-cell DEL screening, expanding the repertoire of E3 ligase ligands, and strengthening functional phenotypic and preclinical translational studies—to fully unlock the therapeutic potential of CIPs against “undruggable” targets.

Qingao Xue ,  Ze Liang ,  Yi Zhang ,  Fei Wang ,  Fulian Wang ,  Lili Liu ,  Guang Yang ,  Lei Yan

Bioorganic Chemistry

DOI:10.1016/j.bioorg.2026.109627

Abstract

The membrane fusion process mediated by the SARS-CoV-2 spike protein is a key therapeutic target. Its heptad repeat 1 (HR1) domain forms a conserved trimeric groove critical for forming the fusion-competent six-helix bundle with HR2. We used DNA-encoded library screening to identify small molecules binding HR1. Hits including Rabeprazole-related compound E (Rab RCE), Omeprazole, Alvimopan, and Olmesartan were characterized. Biophysical assays confirmed binding, while computational simulations revealed distinct interaction modes, with Alvimopan showing high predicted affinity. Cell-cell fusion assays demonstrated potent inhibitory activity for Olmesartan and Rab RCE. Notably, Rabeprazole and Rab RCE showed partial antiviral activity against SARS-CoV-2 variants and HCoV-OC43, rescuing virus-induced apoptosis. Mechanistically, Rabeprazole competitively occupies the HR2-binding groove on HR1, blocking fusion. Our findings identify HR1-targeting molecules like Rabeprazole as promising leads for broad-spectrum coronaviral fusion inhibitors.

Highlights

• A DNA-encoded library (DEL) screening strategy was established to rapidly identify small-molecule binders targeting the conserved heptad repeat 1 (HR1) domain of the SARS-CoV-2 spike protein, enabling efficient mining of repurposed drug candidates from a ∼ 4 billion-compound chemical space.
• Four clinically approved drugs (alvimopan, olmesartan, rabeprazole sulfide, and omeprazole) were validated as HR1-targeting agents, sharing biaryl/heteroaryl cores and hydrogen-bond acceptor groups that mediate specific interactions with HR1 (binding affinities ranging from micromolar to millimolar).
• Two distinct inhibitory mechanisms were delineated: classical competitive occupancy of the HR1 hydrophobic groove (olmesartan, rabeprazole sulfide) and a novel ‘molecular wedge’ mode disrupting the trimeric HR1 interface (alvimopan), providing complementary strategies for targeting viral fusion.
• Olmesartan and rabeprazole sulfide exhibited potent inhibition of SARS-CoV-2-mediated cell-cell fusion, with efficacy comparable to the positive control Salsingle bondC, validating their potential as lead compounds for anti-COVID-19 therapeutics.
• This study establishes a robust pipeline integrating DEL screening, biophysical validation, molecular docking, and functional assays, offering valuable chemical scaffolds and mechanistic insights for developing broad-spectrum coronaviral fusion inhibitors.

Kejia Yan ,  Guilherme M. Lima ,  Tara Bahadur ,  Vincent Albert ,  Zoe O’Gara ,  Gary Bao ,  Christin Kossmann ,  William Kirby ,  Fernando B. Mejia ,  Matthew L. Michnik ,  Kristen Maiorana ,  Ratmir Derda

bioRxiv - Biochemistry

DOI: 10.64898/2026.02.14.705946

Abstract

Genetically encoded (GE) libraries enable identification of high-affinity ligands for diverse molecular targets through iterative in vitro selection and DNA sequencing or next-generation sequencing (NGS). Despite their impact in therapeutic development, a systematic framework for evaluating reproducibility in GE-molecular discoveries remains limited. To aid such analysis, we introduce the concept of baseline response, which reproducibly partitions active and inactive members of in vitro selection. The baseline response is provided by spiking a random DNA-barcoded population. We calibrated the baseline concept using Bioconductor EdgeR differential enrichment (DE) analysis of NGS of phage-displayed selection on oligosaccharide chitin and hepatitis virus NS3a* protease as model targets. We further show that mixing discovery campaigns also offers an effective baseline: chitin-enriched peptides serve as a baseline for DE-analysis of NS3a* selection and NS3a*-enriched peptides serve as a baseline for chitin binders. We applied baseline-stratified DE-analysis to 66 parallel selections performed in 3–5 replicates across 22 extracellular targets, including HER1-3, EpCAM, CAIX, PD-L1, and eight integrin receptors. Automated DE-analysis across hundreds of NGS files produced hits validated in a secondary screen and yielded synthetic macrocyclic ligands with mid-nanomolar affinity confirmed in 2–3 biophysical assays. For PD-L1, we further demonstrated how baseline-calibrated NGS data provide decision-enabling information for optimization of peptide macrocycles to yield potent single-digit nanomolar ligands for the cell-surface receptor. We anticipate that baseline-based analyses of NGS data from in vitro selection procedures will offer a scalable framework for reproducible hit discovery and standardized analysis across diverse in vitro selection campaigns.

Summary

This work introduces a universal baseline framework for in vitro selection of genetically encoded (GE) libraries—e.g., phage-displayed peptide libraries—to improve reproducibility, statistical rigor, and cross-target comparability. The core innovation is spiking a DNA-barcoded random peptide library (serving as an in situ or “cross-target” empirical baseline) into every selection round. This baseline mimics naïve library binding behavior and enables robust normalization and differential enrichment (DE) analysis using Bioconductor EdgeR on NGS data. Validation spanned 22–24 extracellular protein targets (including HER1–3, PD-L1, integrins, NS3a*, chitin) across 66 parallel selections. Baseline-stratified DE identified high-confidence hits, including synthetic macrocyclic ligands with mid- to single-digit nM affinity confirmed by biophysical assays. The method also supports functional benchmarking—e.g., revealing reduced infectivity in MBX-modified phage libraries—and replaces synthetic or computational baselines with empirically derived, target-agnostic mixtures.

Highlights

• Spiked DNA-barcoded random peptides serve as composition-agnostic, in situ baselines for normalization.
• Cross-target library mixing (e.g., chitin + NS3a*-selected peptides) yields effective empirical controls.
• EdgeR-based DE with TMM normalization and BH-FDR correction (α = 0.05) enables quantitative FC estimation binned by input abundance.
• Baseline depletion <1% after NS3a* selection confirms high selectivity.

Conclusion

The universal baseline standardizes hit discovery, improves enrichment fidelity assessment, and enables ML-ready, statistically benchmarked data generation without structural priors.

Yoojin Kim ,  Hayeon Kim ,  Jinhui Hong ,  Minseo Kang ,  Jaeyoung Bae ,  Sangyoon Ko ,  Minjae Kim ,  Byumseok Koh ,  Hakjoong Kim ,  Sang-Hee Shim ,  Kyubong Jo

bioRxiv - Bioengineering 

DOI: 10.64898/2026.02.15.706034

Abstract

DNA-encoded library (DEL) technology enables high-throughput small-molecule discovery but is typically performed using purified proteins under in vitro conditions that do not reflect native intracellular environments. Here, we present a microfluidic agarose microdroplet platform for cellular-context DEL screening. The porous hydrogel droplets provide mechanically stable yet permeable microenvironments that protect weak protein-ligand interactions while enabling efficient buffer exchange and ligand diffusion. Importantly, mild cell permeabilization within droplets selectively retained chromatin-associated proteins, allowing screening directly in a cellular context. Using BRD4 as a model target, we validated intracellular ligand engagement by fluorescence imaging and super-resolution microscopy. Small-scale DEL screening selectively enriched JQ1 in both bead-based and cell-based formats, and large-scale DEL screening across millions of encoded compounds successfully identified hit molecules by sequencing. This agarose microdroplet based strategy expands DEL technology toward biologically relevant and chromatin-associated targets under near-native conditions.

Summary

This work introduces a microfluidic agarose µ-droplet platform for DNA-encoded chemical library (DECL/DEL) screening against intracellular targets, with validation on BRD4—a nuclear epigenetic reader protein. The system encapsulates single cells or target-coated magnetic beads in monodisperse ~100 µm agarose droplets via flow-focusing microfluidics; the low-gelling-temperature (LGT) agarose forms a porous hydrogel upon cooling, permitting rapid diffusion of DNA-encoded small molecules while preserving intracellular architecture and protein–bead complexes. Permeabilization enables controlled probe access, and super-resolution Exchange-PAINT imaging confirms nanoscale colocalization of JQ1-BP with GFP-BRD4 in nuclear nanoclusters.

Highlights

• Agarose µ-droplets enable gravity-assisted washing, shear protection, and uniform molecular diffusion.
• Two-color Exchange-PAINT with orthogonal R2/R6 docking strands validates specific intracellular target engagement.
• DEL screening yields target-specific enrichment: JQ1 barcode is selectively enriched in BRD4-overexpressing HeLa droplets.
• Scalable to large DELs (96×96×96 combinatorial space across three scaffolds) with nanopore sequencing–based enrichment quantification.

Conclusion

The platform bridges functional intracellular DEL screening with high spatial fidelity and quantitative readouts—enabling both PCR- and sequencing-based hit identification while preserving native biomolecular context.