Nature Structural & Molecular Biology

Knowledge of protein structure and function underpins rational drug discovery, yet many targets lack known selectively druggable sites. Furthermore, the identification of secondary druggable sites offers a strategy to overcome drug resistance. Fragment-based drug discovery (FBDD) can identify new ligandable binding pockets, though how to triage those with the ability to exert biologically relevant effects can be unclear. Systematic approaches to identify novel functionally-important protein sites for therapeutic intervention, such as allosteric pockets or protein-protein interaction (PPI) interfaces, has the potential to accelerate drug discovery, particularly when combined with structure-based hit-finding modalities. The phosphoinositide-3 kinase (PI3K) signalling pathway is frequently altered in human cancer and resistance to approved inhibitors is an ongoing challenge. Here, we performed large-scale CRISPR base editing mutagenesis screens across 30 PI3K pathway proteins in three disease-relevant cancer cell models to systematically map functional residues. Integration of base editing data with structural information identified residues corresponding to known catalytic sites, fragment-binding pockets and PPI interfaces, providing validation for the approach. Additionally, we identified putative allosteric pockets near regions of undefined function. Together, these findings establish high-throughput base editing mutagenesis combined with structural analysis as a scalable strategy to delineate structure-function relationships and inform drug development.

O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=200 SRC="FIGDIR/small/741260v1_ufig1.gif" ALT="Figure 1"> View larger version (97K): org.highwire.dtl.DTLVardef@fd2b90org.highwire.dtl.DTLVardef@1371d4borg.highwire.dtl.DTLVardef@164661forg.highwire.dtl.DTLVardef@6cae26_HPS_FORMAT_FIGEXP M_FIG C_FIG Key findingsO_LICryo-EM reveals SDE2 as a novel factor stabilizing the spliceosome active center C_LIO_LISDE2 {Delta}N structures visualize a stalled C* state with impaired docking-factor engagement. C_LIO_LISDE2, Prp18, and FAM32A read a cis-code to stabilize weaker proximal 3'-ss. C_LIO_LISDE2 {Delta}N destabilizes 3'-ss docking and rescues BRCA1 and CFTR mis-splicing in vivo. C_LI Accurate alternative splicing requires discrimination between adjacent 3' splice sites (3'-ss) during catalysis and is disrupted by pathogenic AG-gain mutations that create competing 3'-ss. Here, we present cryo-EM structures of human spliceosomes assembled on native-sequence pre-mRNAs, revealing how the catalytic core controls alternative 3'-ss selection. SDE2 is a previously unrecognized active-center component that promotes a docking-competent spliceosome conformation. Machine learning, in vivo transcriptomics, and in vitro biochemistry show how SDE2 cooperates with FAM32A and Prp18 to act as readers of a cis-regulatory code that governs 3'-ss selection during catalysis. These factors promote weaker, proximal site use by counteracting an intrinsic distal bias generated by active-site interactions with the distal-site -4 nucleotide. Structural or genetic perturbation of these exon-ligation factors destabilizes proximal 3'-ss docking and restores canonical splicing in disease-relevant CFTR and BRCA1 AG-gain alleles. Our work establishes the spliceosome active center as a tunable regulatory hub for alternative splicing.

How the earliest structure within the unfolded state is formed during biosynthesis on the ribosome and whether it has any consequences for downstream folding remain open questions. Here, we combine 15N paramagnetic relaxation enhancement NMR with all-atom molecular dynamics simulations to characterise the unfolded state of a folding-competent immunoglobulin-like domain on the ribosome at the cusp of folding initiation. We identify three structurally distinct sub-ensembles that differ in compaction and ribosome interactions. Non-native contacts, together with ribosome interactions, likely delay folding, yet their persistence alongside early native-like contacts within a sparsely populated compact sub-ensemble suggests they may also facilitate the formation of a co-translational folding nucleus, whose contacts overlap with those of the downstream intermediates. From these findings we infer a mechanistic model of de novo folding initiation during biosynthesis and, by linking the folding nucleus to downstream partially structured intermediates and the native state, provide a complete atomistic description of a co-translational folding pathway.

G protein-coupled receptors (GPCRs) signal across a continuum of activation states, yet how ligands encode distinct signaling efficacies remains poorly understood. Here, we define the molecular mechanism underlying graded activation of the dopamine D1 receptor (D1R), a major therapeutic target for neuropsychiatric disorders. Functional analyses reveal that two widely used pharmacological tools, LE300 and SCH23390, possess intrinsic efficacy as an inverse agonist and a weak partial agonist, respectively, rather than the efficacy-silent neutral antagonists. Structural, molecular dynamics and mutagenesis analyses capture previously unrecognized inactive and intermediate receptor activation states that bridge known active conformations, and reveal that ligand efficacy is encoded through the progressive engagement of a conserved activation pathway centered on the W6.48 toggle switch. Guided by this mechanism, a single chemical modification markedly increases the agonist efficacy of SCH23390. Comparison with dopamine D2 receptor structures further reveals a conserved mechanism of inverse agonism despite distinct subtype-specific recognition. Together, these findings establish a structural framework for graded agonism at D1R and provide general principles for the rational design of efficacy-tuned therapeutics at dopamine receptors and related GPCRs.

The structure and function of eukaryotic chromatin have been extensively studied using conventional salt dialysis-based nucleosome assembly, which has provided fundamental structural insights into nucleosomes as well as a mechanistic understanding of their roles in DNA replication, repair and gene expression. Recently, we developed a labor-saving and time-efficient nucleosome assembly method using a wheat germ cell-free Expression and Assembly Coupled Technology (ExACT), which is emerging as a powerful tool for chromatin research. Here we report cryo-electron microscopy (cryo-EM) structures of human H3.1- and H3.3-containing nucleosomes assembled using this approach and validated by deep-learning-based amino acid-wise model quality (DAQ) scoring. The structures of H3.1- and H3.3-nucleosomes are nearly identical to previously reported models, confirming the structural fidelity of the method. In addition, we determined the previously unreported structure of the primate-specific H3.X-containing nucleosome. Together, these findings validate the cell-free co-expression nucleosome assembly platform and establish this method as a robust framework for biochemically investigating chromatin dynamics.

Cobalamin-dependent methionine synthase (MetH) is a highly dynamic enzyme that requires large-scale domain rearrangements to facilitate turnover and reactivation. Despite significant biochemical and structural efforts, a full picture of the conformational changes underlying the transition between MetH turnover and reactivation has remained elusive. Here, we leverage state-of-the-art cryo-electron microscopy (cryo-EM) methods to reanalyze an old dataset and gain new insights into these changes. By utilizing a neural-network particle picker, performing careful classification of particles, and incorporating high-resolution information to determine initial particle orientations, we uncovered and reconstructed a minor population of reactivation-state MetH ([~]4.5-5 [A] resolution) from a dataset that is dominated by resting-state MetH. This structure represents the first fully intact structure of the enzyme with a physiological cobalamin cofactor bound. These findings highlight the potential for improved cryo-EM methodology to uncover previously hidden details about enzyme function and deepen our understanding of structural ensembles.

Apicomplexan parasites, including Toxoplasma gondii and Plasmodium falciparum, are major human pathogens that cause toxoplasmosis and malaria, respectively. The existing structures of T. gondii translational machinery are from empty ribosomes that lack several key components, including ribosomal protein RACK1 (Receptor for Activated C Kinase 1). Here, we used cryo-electron microscopy (cryoEM) to determine high-resolution structures of T. gondii ribosomal complexes, including a translating 80S ribosome bound to mRNA and tRNA. These structures reveal that RACK1 occupies the conserved binding site on the 40S subunit observed in other eukaryotic ribosomes. We also determined the architecture of the ribosomal P-stalk and identified the ribosomal proteins uL10 and uL11, which were not observed in previous T. gondii ribosome structures. In addition, we determined structures of the 80S ribosome bound to mRNA, tRNA, and the translation inhibitor emetine in two distinct conformational states. These snapshots reveal two mechanisms by which emetine inhibits the translocation step of mRNA translation: either by dislodging the mRNA from the E-site of the ribosome or by acting as a molecular glue within the E-site, thereby stalling translocation. Together, these findings provide new insights into the molecular basis of protein synthesis in apicomplexan parasites and establish a structural framework for the development of future antiparasitic therapeutics. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=147 SRC="FIGDIR/small/740346v1_ufig1.gif" ALT="Figure 1"> View larger version (30K): org.highwire.dtl.DTLVardef@1997523org.highwire.dtl.DTLVardef@1d94da7org.highwire.dtl.DTLVardef@1790fe3org.highwire.dtl.DTLVardef@1023213_HPS_FORMAT_FIGEXP M_FIG C_FIG

Dihydroorotate dehydrogenase (DHODH), which catalyzes the rate-limiting step in de novo pyrimidine biosynthesis, is a validated therapeutic target in cancer, autoimmune disorders and infectious diseases. DHODH is hypothesized to utilize a ping-pong catalytic mechanism. However, available crystal structures of DHODH in complex with small molecule inhibitors are irreconcilable with this model, showing simultaneous occupancy of both substrate binding sites. To elucidate the structural basis for DHODH-mediated catalysis, we resolved the structures of two DHODH holoenzymes. These structures capture novel conformational states that show mutually exclusive substrate binding. The holoenzyme structures also suggest that conformational changes in the catalytic loop of DHODH play a critical role in regulating enzyme activity. To map the functional landscape underlying DHODH inhibitor resistance, we performed deep mutational scanning drug-resistance screens with two clinically-relevant structurally distinct DHODH inhibitors, BAY2402234 and brequinar. Resistance variants cluster in the inhibitor binding site and at a previously unappreciated surface pocket on DHODH that allosterically regulates ubiquinone binding. Together, our findings provide structural evidence for the ping pong model of catalysis, define the mutational landscape governing inhibitor resistance, and reveal novel structural vulnerabilities in DHODH that can be utilized for future drug development.

Cation-proton antiporters (CPAs) are vital for the maintenance of ionic homeostasis and normal physiology among diverse cell types. Despite recent insights into K+/H+ exchange transporters, the diversity in their structural organization and regulatory mechanisms of K+-specific CPAs are minimally understood. Here, we explore the architecture of an E. coli CPA1 K+/H+ antiporter, YcgO and its inhibition by the unphosphorylated form of PtsN, the terminal protein of a regulatory phosphorelay, using cryoEM structures at 3.4 [A] and 3.2 [A] resolution, respectively. Homodimeric YcgO bound to K+ ions in the occluded conformation, harbors additional linked cytosolic domains, RCK and CorC, to regulate the movement of the transport helices within the YcgO dimer. These domains are the sites of interaction and efflux inhibition by unphosphorylated PtsN, which interacts with the CorC domains with high affinity and allosterically augments inhibitory interactions of CorC with transport helices of YcgO. Inhibition is relieved leading to constitutive activation, upon disrupting the CorC-transport conduit interface. This study illuminates the structural basis of K+ efflux mediated through regulation of a K+/H+ antiporter in E. coli and related prokaryotes via a metabolic network involving a regulatory phosphorelay.

Nuclear organization is a fundamental feature of cell identity and cell fate determination. Compared with differentiated cells, pluripotent stem cells exhibit markedly distinct nuclear architecture at multiple levels, including chromosome folding, lamina association, histone modification landscapes and nuclear body organization. Although individual aspects of this reorganization have been characterized, how and in what order these features are remodeled as cells commit to a lineage remains poorly understood. Here, we temporally map nuclear reorganization across five stages of an in vitro differentiation system of human embryonic stem cells through definitive endoderm and hepatoblast intermediates into hepatocyte-like cells. By integrating Hi-C, RNA-seq, ATAC-seq, ChIP-seq, and CUT&RUN data, we establish a genome-wide framework linking structural, epigenetic, and transcriptional changes across differentiation. Combined with immunofluorescence imaging, chromosome painting, and liquid chromatin Hi-C (LC-Hi-C) we show that nuclear reorganization proceeds in a stepwise and temporally ordered manner through three major transitions. In the first transition, chromosomes condense into defined territories concurrent with transient anchoring of centromere-proximal regions to the nuclear periphery. In a second transition, active and inactive chromatin segregate, leading to stronger compartmentalization. This coincides with deposition and peripheralization of H3K9me2-marked chromatin and morphological changes in nuclear speckles, while chromatin conformation at speckle-associated regions stabilizes. In a third transition, after lineage-specific genes are activated, chromatin interactions globally stabilize, establishing a more stable nuclear architecture on top of these earlier large-scale structural rearrangements. Together, our results define a stepwise framework for nuclear reorganization during human embryonic stem cell differentiation and reveal that the transition from pluripotency to a differentiated state proceeds through coordinated, temporally ordered structural events.

AO_SCPLOWBSTRACTC_SCPLOWRabies virus phosphoprotein (P protein) antagonizes interferon signalling by selectively targeting activated STAT1 protein, but the structural basis for the selective recognition of activated STAT1 has remained unclear. Here we report cryo-EM structures of tyrosine-phosphorylated STAT1 (pY-STAT1) in complex with the C-terminal domain of P protein (P-CTD). P-CTD engages the pY-STAT1 tetramer through three discrete interfaces, contacting two DNA-binding domains within a pY-mediated STAT1 dimer and the N-terminal domain of a third STAT1 protomer. This binding mode explains how P-CTD selectively recognizes activated STAT1, occludes surfaces required for importin 5 and DNA binding, and stabilizes the compact tetrameric state, thereby restricting the conformational transition required for cooperative DNA binding. Cell-based assays further indicate that distinct P-CTD surfaces differentially contribute to antagonism of type I and type II interferon signalling. Further, by acting as an anchoring molecule, P-CTD also enabled structural elucidation of the complete STAT1 structures in the tetramer. Together, these data reveal critical mechanisms in viral immune evasion and pathogenesis, and fundamental innate immune signaling.

CRISPR technologies based on nuclease-deactivated Cas9 (dCas9) rely on programmable DNA binding rather than DNA cleavage, yet the intrinsic DNA-recognition properties that govern optimal guide RNA (gRNA) performance remain poorly understood. Existing approaches either measure genomic occupancy in cells or infer dCas9 behavior from cleavage-based Cas9 datasets, despite DNA binding being substantially more permissive than DNA cleavage. Here we introduce TANGO (Targeted Array-based Nucleic acid-Guided Occupancy), a high-density DNA-array platform that quantitatively profiles intrinsic dCas9:gRNA binding across tens of thousands of DNA targets in a cell-free system. TANGO captures established features of dCas9 target recognition, while providing substantially greater sensitivity than prior assays. Comparison with ChIP-seq data demonstrates that intrinsic DNA-binding specificity is a major driver of genomic occupancy and reveals that chromatin accessibility modulates the intrinsic binding affinity required for dCas9 recruitment. Across CRISPRi/a guides, TANGO identifies multiple independent biochemical determinants of guide performance--including on-target affinity, mismatch tolerance, and ribonucleoprotein assembly--and flags problematic and highly promiscuous guides overlooked by current specificity metrics. Unexpectedly, some guides retain substantial guide-directed DNA binding even in the absence of a protospacer-adjacent motif (PAM), revealing an additional dimension of dCas9 specificity. Together, these results establish intrinsic DNA recognition as a quantitative and experimentally accessible determinant of dCas9 function, providing a framework for improving guide selection and enhancing the precision of CRISPR technologies.

In the paper by Nguyen et al, published in Nature Communications in 2024, we describe structural polymorphism in fibrils extracted from patients with ATTR amyloidosis. ATTR amyloidosis is characterized by deposition of transthyretin amyloid fibrils across multiple organs, resulting in marked phenotypic variability. Previous studies on ATTR suggest its phenotypic variability may be linked to differences in fibril structure, as demonstrated in amyloid diseases such as AA amyloidosis, tauopathies, or synucleinopathies. Thus far, most ATTR fibrils adopt a strikingly similar fold across patients and affected organs. The most notable exception is described in the study by Nguyen et al., where we report the identification of two co-existing polymorphs in cardiac fibrils from three patients with the ATTRv-I84S variant. These cardiac polymorphs differ in conformation of the segment Gly57 to Gly67, termed the gate, which regulates access to a central polar channel within the fibril core. In one polymorph, the gate adopts an open conformation, exposing the channel, whereas in the other it is closed, shielding it. Whether these polymorphs extend beyond cardiac tissue remains unknown. Here, we address this question by extending our analysis to additional organs from a fourth patient with ATTRv-I84S amyloidosis.

The short arms of human acrocentric chromosomes are characterized by nucleolar organizer regions essential for ribosome biogenesis, but their highly repetitive nature has hindered genomic analysis. Leveraging the recently completed genomes of all major great ape lineages, we identified recurrent features of their acrocentrics, including enriched repeat classes, centromere repositioning by whole-arm inversion, interchromosomal sequence exchange, and birth-and-death evolution of multiple gene families under selection. Together, these processes have enabled the repeated amplification and diversification of the FRG1 gene family over 25 million years of ape evolution, and, in gorilla, the amplification of a novel IGSF3-GGT fusion gene. Similar evolutionary events also explain the distribution of segmental duplications and heterochromatin in the modern human genome, predisposing it to karyotypic abnormalities such as Robertsonian translocations. Our findings highlight acrocentric chromosomes as key drivers of evolution in the great apes, with implications for speciation, adaptation, and clinical genomics.

Targeted protein degradation (TPD) has emerged as an increasingly powerful approach for therapeutic development and biological discovery. TPD compounds including proteolysis-targeting chimeras (PROTACs), molecular glues, and tag-targeting protein degraders (tTPD) enable rapid, selective and reversible degradation of proteins through recruitment of the ubiquitin-proteasome system (UPS). However, genome-wide CRISPR screens performed with targeted protein degraders are frequently dominated by resistance mechanisms that disrupt degrader activity, including loss of recruited E3 ligase components and broader UPS regulators. The strong selective advantage conferred by these perturbations can obscure less penetrant, biological genetic interactions that operate downstream of target degradation. To overcome this limitation, through iterative genome-wide screening and manual curation, we developed a TPD-compatible CRISPR knockout library that retains near-genome-scale coverage while excluding a focused set of genes recurrently associated with degrader failure. Across multiple degrader screens, this library reduced the dominance of UPS-associated resistance mechanisms and improved the detection and prioritization of genetic interactions linked to target biology. Using the RBM39 molecular glue degrader indisulam as a model, we identified ZMAT2 loss as a resistance mechanism that preserves RBM39 degradation but attenuates the transcriptional and splicing consequences of target depletion. Together, our work establishes a TPD-compatible CRISPR screening framework that improves the biological resolution of degrader resistance screens and facilitates the discovery of genetic dependencies operating downstream of targeted protein degradation.

Cellular self-organisation counteracts entropy at the expenditure of energy. In case of the first level of nuclear DNA organisation, this relates to regular nucleosome arrays and interspaced nucleosome-depleted regions (NDRs), for example at promoters or replication origins1-5. The organisation of nucleosomes as building blocks of chromatin is orchestrated by the collective activities of ATP-dependent chromatin remodellers6-9. Yet, how remodellers achieve positional specificity, in particular regarding the promoter-proximal +1 nucleosomes, remains unclear. Here, we show that the S. cerevisiae chromatin remodeller INO80 unexpectedly distinguishes DNA sequence asymmetry within the +1 nucleosome of the SWH1 gene through distinct inhibited and active nucleosome-binding modes. In structural and biochemical analyses of INO80 on nucleosomes with the endogenous sequence, we identified an inhibited binding mode where the entire INO80 remodelling unit flipped on the +1 nucleosome. INO80 adopted this remodelling-incompetent binding mode when facing the promoter, but a remodelling-competent mode when facing the gene body. This directional read-out of intra-nucleosomal DNA sequence asymmetry, together with extranucleosomal NDR sequence features, prevented nucleosome sliding into the NDR while permitting array formation over the gene. Our work shows how DNA features contribute to ATP-dependent self-organisation of promoter chromatin by INO80.

Eukaryotic mRNAs carry a 5'-terminal 7-methylguanosine (m7G) cap that is recognized by multiple cellular proteins and is essential for mRNA self-discrimination, stability, and translation. The eukaryotic translation initiation factor 4E (eIF4E) binds the 5' cap and promotes recruitment of mRNAs to the ribosomal translation initiation complex, a critical step in cap-dependent translation. Beyond naturally occurring cap modifications that modulate mRNA interactions and fate, non-natural 5' cap modifications can be used to finely tune diverse aspects of mRNA biology. This has gained particular relevance with the emergence of in vitro transcribed mRNAs as a powerful therapeutic modality for protein replacement therapies and vaccines. Here, we report binding data for a series of synthetic 5' cap analogues interacting with eIF4E, together with X-ray crystal structures of eIF4E-cap complexes, providing detailed insights into the structural determinants of cap recognition by eIF4E. Furthermore, analysis of gene expression from mRNAs carrying different synthetic cap analogues in human cell lines and mice revealed a relationship between cap architecture and protein expression. Taken together, our structural, biochemical, and functional data establish structure-activity relationships for synthetic cap analogues and identify structural features associated with enhanced translational performance.

Variation alters inherent protein properties like thermodynamic stability and catalytic activity, and these intrinsic effects are shaped by the complex cellular environment. Signaling cascades typify this complexity, with tens or hundreds of proteins acting in concert through regulated states and interactions to transduce signals. Here, we measure the effect of 186,676 variants across 17 proteins in the RAS-MAPK pathway on protein abundance and pathway activity in human cells under diverse conditions, for a total of 503,099 variant effect measurements. We discovered 14,412 dominant negative variants across 12 proteins, nearly all previously unknown, that disrupt specific domains or interaction interfaces while leaving other functions intact and that are depleted from the human population. Measuring the abundance of 106,642 variants across 10 proteins under HSP90 inhibition showed that HSP90 shapes the effect of 67% of missense variants across the pathway, buffering destabilizing variants in both strong and weak clients. Our variant effect data, collected in the context of an intact proteome, enabled us to assess the functional importance of 395 experimentally and computationally modeled MAPK protein interactions. Using additional mutational scans, we resolved how the promiscuous switch II helix binding surface of KRAS accommodates overlapping interactions with LZTR1 and SOS1, revealing distinct allosteric networks that support each interaction. Thus, we establish a framework for measuring variant effects at unprecedented scale in human cells, illuminating fundamental MAPK pathway signaling mechanisms and demonstrating that cellular context profoundly shapes the effect of most variants.

Nucleosome positioning plays a fundamental role in regulating DNA accessibility and cellular activity, yet the precise rules by which mammalian sequences instruct this organization have remained incomplete. Prior work has established that intrinsic sequence preferences dictate nucleosome architecture in yeast, but in mammals, this sequence code is modulated by chromatin state and cell-type-specific factors. Here, we train a deep-learning model on genome-wide maps from mouse embryonic stem cells to predict nucleosome occupancy directly from the primary DNA sequence. Our model accurately predicts in vivo occupancy and captures shared biological signals while effectively denoising assay-specific technical biases. Through in silico mutagenesis, we identify specific nucleosome positioning regions comprised of short 5-to-20-base-pair motifs concentrated within dense regulatory windows. Furthermore, we demonstrate that repetitive DNA elements play a major role in establishing these highly organized nucleosome arrays on a genomic scale. Specifically, CCCTC-binding factor motifs embedded within short interspersed nuclear elements, zinc-finger-bound motifs within long interspersed nuclear elements, and microsatellite repeats act as potent organizers of nucleosome phasing. Together, our findings extend the concept of sequence-encoded nucleosome architecture to mammalian genomes and reveal that regularly spaced nucleosome islands are shaped by short, recurrent sets of DNA motifs that couple local sequence to chromatin structure.

Many chromatin-associated proteins have been found to bind RNA as a means of epigenetic regulation. Specifically, DNA methyltransferase 1 (DNMT1), which maintains cytosine methylation at CpG dinucleotides, is inhibited by RNA at transcribed DNA loci in cells. However, the mechanisms by which RNA binds DNMT1 and inhibits its activity remain unknown. Here, we determine a series of cryogenic electron microscopy (cryo-EM) structures of human DNMT1 bound to pUG-fold RNA, a non-canonical G-quadruplex previously observed to inhibit activity, revealing two distinct RNA-binding modes. The pUG-fold RNA binds the surface of DNMT1 in its autoinhibited conformation across a positively charged surface between the methyltransferase domain and the CXXC domain, and it binds directly in the active site of an open DNMT1 conformation. RNA binding is sterically incompatible with substrate DNA engagement in both states. Our 2.5 [A] structure captures the intricate network of hydrogen bonds and electrostatic interactions between amino acids in the methyltransferase domain and the tetrad layers of pUG-fold RNA. Metadynamics molecular dynamics simulations provide an orthogonal view of the conformational landscape of DNMT1, revealing the two distinct RNA-binding modes. Furthermore, our analysis of published DNMT1 RIP-seq and eCLIP-seq data confirms that DNMT1-interacting RNAs in cells exhibit a strong propensity to form non-canonical G-quadruplex RNA structures. Collectively, our study provides the first structural basis for pUG-fold RNA recognition by a protein and illustrates how cryo-EM and AI-based methods for protein and RNA structure prediction synergize to inform the mechanism of RNA-mediated regulation of DNMT1.