AI Insight
Researchers developed SPIRAL, a curated database of 1,098 RNA-small molecule crystal structures drawn from the Protein Data Bank, and analyzed them using a computational pipeline to characterize how small molecules bind to six functional categories of RNA. Unsupervised clustering of structural interaction fingerprints revealed six mechanistically distinct binding modes, with RNA functional class strongly determining which mode predominates. Among 275 affinity-characterized entries, C2'-endo sugar pucker frequency and buried contact surface area emerged as the strongest independent predictors of binding affinity, both enriched at junction loops, pseudoknots, and base multiplet networks.
Why it matters
RNA is an increasingly important therapeutic target, and identifying the structural features that govern potent, selective binding could guide the rational design of RNA-targeted small molecule drugs for diseases where current treatments are limited.
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
Small molecules that target structured RNA hold therapeutic promise across a wide range of diseases, yet the structural principles governing RNA-ligand recognition remain poorly defined. Here we present SPIRAL (Structural Pockets and Interacting RNA-Associated Ligands), a curated database of 1,098 RNA-small molecule structures from the Protein Data Bank covering 1,137 ligand-binding events across six functional RNA categories: riboswitches, ribozymes, synthetic aptamers, G-quadruplexes, ribosomal RNA, and regulatory RNA motifs. A customized pipeline built on DSSR (Dissecting the Spatial Structure of RNA) extracts structural interaction parameters from each complex, capturing stacking geometry, hydrogen-bond topology by RNA moiety, backbone contacts, groove engagement, and tertiary motif context. Unsupervised clustering of these fingerprints resolves six mechanistically distinct binding modes whose distribution is strongly governed by RNA functional class, demonstrating that different RNA categories engage small molecules through fundamentally different chemical strategies. To enable category-independent comparison of interaction quality across these mechanistically diverse modes, we introduce the Composite Binding Quality Score (CBQS), a seven-metric framework that ranks riboswitches highest and regulatory RNA motifs lowest among the six categories, while ribozymes, synthetic aptamers, and G-quadruplexes achieve statistically equivalent intermediate scores through three distinct recognition strategies. Analysis of 275 non-redundant affinity-characterized entries identifies C2′-endo sugar pucker count and total buried contact surface area as the dominant independent predictors of binding affinity. Both predictors are enriched at junction loops, pseudoknots, and base multiplet networks, the same tertiary structural sites most under engaged by current regulatory RNA motif binders, suggesting that ligands designed to contact these sites would improve both potency and selectivity simultaneously.