Structure

Disulfide-rich peptides possess exceptional potency and selectivity but are often limited by the instability and synthetic challenges associated with native disulfide bonds. Here, we report the design, synthesis, pharmacological evaluation, and structural characterisation of triazole-based peptidomimetics of the -GI conotoxin, a selective antagonist of the muscle-type nicotinic acetylcholine receptor (nAChR). A series of 1,4- and 1,5-disubstituted triazole analogues were prepared entirely on resin using CuAAC and RuAAC chemistry to replace the native Cys3/13 disulfide bridge. Functional evaluation against human muscle nAChRs revealed that 1,5-triazole analogues retained low-nanomolar potency, with the lead mimetic exhibiting activity comparable to native -GI. Cryo-electron microscopy of the lead compound bound to the muscle-type nAChR provided the first structure of a disulfide-isostere peptidomimetic in complex with a membrane receptor. The structure demonstrates that the 1,5-triazole reproduces the native peptide fold with high fidelity while contributing receptor-facing interactions not available to the native disulfide bridge. Molecular dynamics simulations further revealed conserved hydration networks and similar conformational sampling between the native peptide and lead mimetic. Together, these findings establish triazoles as effective disulfide surrogates and provide a structural framework for the rational design of stabilised conotoxin therapeutics.

The cluster of differentiation 36 (CD36) is a membrane protein with broad physiological roles in health and disease, and its function is regulated in part by phosphorylation. Experimental evidence shows that phosphorylation of Thr92 reduces CD36 affinity for thrombospondin-1 (TSP-1), binding of which initiates antiangiogenic signaling, whereas phosphorylation of Ser237 decreases CD36-mediated fatty acid uptake, with implications for energy metabolism. However, the only available crystal structure of CD36 lacks phosphorylation, and the molecular mechanisms by which phosphorylation regulates CD36 function remain largely unknown. This study provides an atomically detailed computational characterization of CD36 in unphosphorylated and dual phosphorylated states, using molecular dynamics simulations with a total sampling time of 30 microseconds in combination with Markov state models. We present, to our knowledge, the first evidence of a cryptic pocket on CD36 surface that is formed by phosphorylation. This cryptic surface pocket and a loop spanning residues 121-131 form a high affinity binding site for TSP-1 derived ligands, shifting their binding away from the canonical site. We propose that this altered binding provides a molecular basis for the disruption of antiangiogenic signaling upon CD36 phosphorylation. Additionally, our data indicate that, phosphorylation increases helicity and compaction within the helix-loop region spanning residues 296-331, narrowing one of the entrances to the internal cavity and reducing its overall volume. These conformational changes provide a potential mechanistic explanation for the decrease in fatty acid uptake upon CD36 phosphorylation. Our findings provide structural insights that may inform the future design of CD36 modulators and emphasize the importance of targeting phosphorylation induced CD36 conformations in angiogenic and metabolic diseases.

Initiator-methionine excision is among the earliest protein modifications, yet its relationship to assembly geometry is unknown. Burial of the mature first residue was measured across 7,246 deposited human biological assemblies (22,291 chain-level observations; 1,191 proteins). Among 1,143 analyzable proteins, termini in MetAP-permissive penultimate-residue sequence classes were less often interface-engaged than termini in MetAP-nonpermissive classes (37.4% versus 47.4%; adjusted odds ratio 0.65, p = 7.2e-4). Curated processing annotations did not show a corresponding burial difference, and correlated residue properties preclude attributing the sequence-class association specifically to iMet removal. The analysis identified 264 interface-engaged MetAP-permissive candidates concentrated in cellular machines. In a fully recomputed conformer scan of deeply buried proteasome positions, modeled methionine accommodation was less favorable than at observed-methionine controls (median overlap -0.30 versus -1.12 angstrom, p = 0.0049), although most scoreable sites permitted a nonoverlapping placement. The census therefore reveals a graded structural constraint - not universal steric failure - and prioritizes complexes in which altered packing, assembly kinetics, lipidation or N-terminal methylation can be tested.

Ammonia oxidation initiates nitrification and is closely linked to microbial N2O production. Ammonia monooxygenase (AMO) catalyzes the first and rate-limiting step of nitrification and is widespread across evolutionarily distinct ammonia-oxidizing archaea (AOA) and bacteria (AOB). The ocean is the largest biome for AOA and AOB, which have distinct ecological niches and markedly different sensitivities to nitrification inhibitors. However, the lack of archaeal AMO structures and inhibitor-bound AMO complexes has hindered mechanistic understanding of the architectural, catalytic, and inhibitory divergence between these two enzyme systems. Here, we report high-resolution cryo-electron microscopy (cryo-EM) structures of marine archaeal AMO captured in active and inactivated states within its native membrane environment, together with inhibitor-bound structures of estuarine bacterial AMO. Archaeal AMO forms an unexpected cup-shaped homotrimer composed of eight subunits per protomer and exhibits substantial architectural divergence from bacterial AMO. Integrated structural, biochemical, kinetic, and computational analyses reveal distinct periplasmic architectures, copper-center organization, and hydrophobic channels between archaeal and bacterial AMOs for ammonium acquisition, catalysis and inhibitor response. These findings provide a structural and mechanistic framework for understanding how archaeal and bacterial AMOs have diverged to distinct ammonia-oxidizing strategies and inhibitor susceptibilities across environmentally important ammonia oxidizers.

IgM is the most ancient antibody isotype, playing an important role in both circulatory and mucosal immune responses across vertebrates, yet structural characterization of its polymeric forms is limited outside of mammals. Here, we report the cryo-electron microscopy structure of mallard duck secretory (S) IgM at 3.37-[A] resolution. The structure revealed a pentameric core globally similar to human SIgM, supporting the view that pentameric IgM is subject to strong evolutionary constraints. However, compared to mammalian structures, we observed species-specific differences at molecular interfaces. Surface plasmon resonance binding assays characterizing secretory component (SC)-IgM interactions supported structural observations and, when compared to IgA binding, revealed isotype-specific contributions from the avian SC N-terminal extension. Together, these findings establish a comparative structural framework for polymeric IgM across vertebrates and provide insight into how avian SIgM-specific features may support mucosal immunity in birds.

Deep learning-based protein structure prediction methods that leverage evolutionary information from multiple sequence alignments (MSAs), exemplified by AlphaFold2, have achieved remarkable accuracy. However, existing methods still struggle to predict challenging proteins, particularly those with novel folds or limited evolutionary information, and to recover alternative conformational states. Here we show that structure prediction models trained under different MSA-depth distributions corresponding to different levels of evolutionary information exhibit complementary generalization behaviors, and that a model trained on a mixture of these distributions can combine their complementary generalization strengths. Building on this insight, we developed ProtMonomer, a deep learning framework trained on MSA-depth distributions representing a broad range of evolutionary information levels to improve structure prediction. Across benchmarks comprising CASP15 targets, non-redundant experimentally determined structures, orphan proteins, and short peptides, ProtMonomer performed comparably to or better than leading methods, including AlphaFold2 and AlphaFold3, with particularly strong performance on challenging targets. For fold-switching proteins, ProtMonomer also recovered alternative conformational states more accurately than AlphaFold2 and AlphaFold3 across diverse homologous sequence sampling strategies. In addition to improving predictive accuracy, ProtMonomer substantially reduced inference cost through an efficient architecture, enabling high-throughput applications. Together, these findings provide insights into the generalization of evolution-informed structure prediction models and support ProtMonomer as an accurate and efficient framework for protein structure prediction.

Human cytochrome P450 2C9 (CYP2C9) is a hepatic microsomal enzyme involved in the oxidative metabolism of clinically important drugs, but the structural organization of its oligomeric assemblies outside crystallographic packing environments remains poorly understood. Here, we report the cryo-EM structure of human CYP2C9 determined under aqueous, membrane-free conditions at 3.31 Angstrom resolution. The structure reveals a C2-symmetric hexameric assembly organized as a dimer of trimers. Individual protomers retain the conserved P450 fold and heme-binding architecture observed in previously reported crystal structures, indicating that assembly formation does not substantially perturb the catalytic core. The hexamer is stabilized by defined intra-trimer interfaces involving the N-terminal region and residues around Trp212 and Phe482, together with inter-trimer interfaces involving Leu71 and the 220-227 loop. These interfaces are distinct from the crystal packing contacts observed in CYP2C9 crystal structures, demonstrating that the assembly is not a simple recapitulation of crystallographic packing. Notably, the inter-trimer interface is located near the FG-loop-containing surface previously implicated in membrane association. This suggests that the observed hexamer may represent a membrane-free association of two trimers through membrane-related surfaces, whereas the trimeric arrangement itself may be compatible with membrane-associated organization. The structure therefore provides a framework for investigating how trimer formation, membrane interaction and local conformational changes in the FG-loop region may influence CYP2C9 function.

Cryo-electron tomography (cryo-ET) enables near-native visualization of subcellular architectures, yet applying it to moderately thick, multilayered tissues such as the retina is hampered by inadequate vitrification and inaccurate depth-targeting. Here, we developed PLCT, an integrated approach combining modified high-pressure freezing, cryo-ultramicrotome trimming, and plasma-based cryo-FIB milling to overcome these barriers. PLCT reliably vitrified <100 m retinal strips with minimal ice artifacts, navigates precisely to the outer plexiform layer using morphological landmarks, and produces high-quality lamellae suitable for high-resolution cryo-ET. Subtomogram averaging (STA) analysis identified microtubules at 16.33 [A] within retinal horizontal cell processes. Importantly, STA also resolved a 10-nm-diameter filamentous structure at 24.81 [A] in the same processes, featuring six peripheral strands surrounding an elongated central density with continuous intervening cavities, an architecture consistent with intermediate filaments. Together with its native localization and immunoreactivity, these features collectively identify the filaments as neurofilaments. Separately, 3D reconstruction of synaptic ribbons uncovered a previously unrecognized "mahjong tile"-like fine ultrastructure. These results demonstrate that PLCT-produced lamellae are of sufficient quality to support structural analysis in native tissue. Although demonstrated on retinal photoreceptor synapses as a proof-of-principle, PLCT is inherently generalizable, with its depth-navigation and vitrification strategies directly applicable to any multilayered tissues. This work establishes PLCT as a robust, reproducible platform for depth-resolved in situ cryo-ET of multilayered tissues.

Transcription factor Ets2 coalesce with the NF-{kappa}B pathway to regulate gene expression in specific signaling contexts. IKK2/{beta}-mediated phosphorylation events critically regulate the NF-{kappa}B pathway. However, any link between Ets2 and IKK2 remains elusive. Here we report Ets2 as a direct substrate of IKK2. In-vitro kinase assays using deletion constructs, high resolution MS-MS and site directed mutagenesis identified S295 as a prominent phosphorylation site distal to the DNA binding domain, substitution of which to phosphor-mimetic Glutamate triggers further phosphorylation of Ets2. MD simulations clearly indicate conformational constriction of the otherwise disordered N-terminal region and inhibition of DNA binding activity upon phosphorylation, which was further confirmed by Electrophoretic mobility shift assays. Our results uncover a phosphoregulatory connection between Ets2 and IKK2.

UCH-L1 is a monomeric deubiquitinating enzyme whose native structure embeds a shallow $5_2$ knot located near the N-terminus, placing the knotted topology in direct proximity to both the substrate-binding pocket and the catalytic site. While our previous work established that N-terminal integrity is critical for catalytic activity, the energetic cost of unknotting and its structural consequences remained unquantified. Here, we combine steered molecular dynamics with an umbrella sampling scheme to generate topologically modified variants of UCH-L1 and, for the first time, reconstruct the free-energy profile of UCH-L1 unknotting. The potential of mean force reveals a steep energetic barrier to knot disruption, consistent with knotting being a late, rate-limiting folding step that is effectively locked in once the native structure is established. Long unbiased MD simulations of fully unknotted variants in both apo and holo states show that knot removal increases local flexibility at the N-terminus without inducing significant global structural destabilization. Binding energy calculations indicate that the unknotted variant binds to ubiquitin less tightly than the wild-type ($\sim$-62~vs~$\sim$-76~kcal/mol), suggesting that topological integrity contributes to substrate affinity. Together, these results show that the $5_2$ knot in UCH-L1 is not a passive structural feature but a functional element that fine-tunes folding kinetics and contributes to substrate binding efficiency.

Quorum sensing in Gram-positive bacteria commonly relies on posttranslationally modified peptide pheromones. In Bacillus subtilis, the prenyltransferase ComQ catalyzes tryptophan prenylation of the quorum-sensing peptide ComX, but the structural basis of this unique peptide modification has remained unclear. Here we identified a previously uncharacterized ComQ homolog, StheQ, and its cognate peptide substrate, StheX, from Sphaerobacter thermophilus and investigated their structural and functional relationship. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis demonstrated that StheQ catalyzes prenylation of the tryptophan residue located second from the C-terminus of StheX. Crystal structures of apo StheQ and its complexes with a farnesyl pyrophosphate analog revealed that StheQ adopts the all--helical fold of the trans-isoprenyl diphosphate synthase (IPPS) superfamily while possessing an active-site architecture adapted for peptide-based indole prenylation. The structures identified a single Mg2+-binding site associated with the first aspartic acid-rich motif and showed no evidence for metal coordination at the pseudo-second aspartic acid-rich motif. Site-directed mutagenesis, complex formation assays, and docking analyses identified a peptide-binding pocket adjacent to the active site and suggested that N215 contributes to productive positioning of the acceptor tryptophan. These findings establish the structural basis for peptide prenylation by a ComQ-family enzyme, providing insight into the evolution of peptide-based indole prenylation within the IPPS superfamily, and support the view that ComQ-family enzymes constitute a distinct functional branch specialized for peptide modification.

Life began with the emergence of a molecule that could replicate its own genetic material, a task plausibly mediated by an RNA-dependent RNA polymerase ribozyme. Here, we present the structure of such a polymerase ribozyme, bound to RNA substrates comprising the template, primer, and nucleoside triphosphate (NTP) analog. The structure reveals how directed evolution shaped flanking elements around a highly conserved catalytic core derived from the ancestral class I ligase ribozyme. Each element serves as a functional module, positioning the primer-template duplex and incoming NTP within the active site of the enzyme. This emergent domain organization is remarkably similar to the "right hand" configuration of polymerase proteins, suggesting a common functional form for copying nucleic acids, regardless of biopolymer catalyst.

Linker chemistry and conformation are central determinants of PROTAC activity, shaping ternary-complex geometry, cooperativity, target-lysine presentation and cellular permeability. Existing linker generators often lack explicit control over linker flexibility, require predefined attachment sites and linker lengths, or produce structures that demand substantial geometric correction, limiting their utility in practical PROTAC design. Here we introduce FlexiTAC, a Bayesian flow network that jointly generates linker atom types and coordinates from the warhead and E3-ligase-ligand contexts. We also assemble PROTAC-3D, a quality-controlled collection of 63,554 component-resolved PROTAC structures for model training, and PROTAC-Bench, which covers molecular quality, fragment preservation, geometric fidelity, conformational stability, fragment awareness, rediscovery and sampling efficiency. Compared to the best 3D baseline models, FlexiTAC improves validity by 12.0-12.7% and achieves the highest PoseBusters pass rate of 79.5%-80.0%. A differentiable guidance module shifted generated linkers along a conformational ensemble-derived rigidity axis without retraining the generator. In silico case studies further show that the model can accept crystal-derived, redocked or predicted structural inputs. Together, FlexiTAC, PROTAC-3D and PROTAC-Bench establish an integrated and reproducible framework for data-driven PROTAC linker design, combining controllable structure-conditioned generation with standardized training data and evaluation protocols. This framework expands the linker chemical and conformational space accessible to computational exploration, provides a foundation for future method development and enables the systematic generation of structure-conditioned linker designs with tunable conformational flexibility.

Canonical McrBC enzymes are nucleotide-powered, motor-driven endonucleases that bind and cleave modified bacteriophage DNA. Non-canonical McrBC homologs like LlaI and BsuMI are distinguished by a unique three-gene organization and the ability to target DNA site-specifically. Here, we report the atomic-resolution crystal structures of the DNA-binding module LlaI.R1 and AAA+ motor LlaI.R2 from the Lactococcus lactis LlaI anti-phage defense system. The crystallized LlaI.R2 hexamer traps two distinct active site conformations that correlate to different states of the nucleotide hydrolysis cycle and reveal that the organization of the critical catalytic machinery present in canonical McrB homologs is also conserved in non-canonical R2 proteins. Although canonical McrB homologs are strictly GTP-specific, we find that the R2 proteins from LlaI and BsuMI do not discriminate between different nucleotides, even when in complex with their respective R1 partners. Using mutagenesis, we define surfaces on the LlaI.R1 structure that are critical for DNA-binding and interaction with LlaI.R2. These observations support computational modelling of the assembled LlaI restriction system bound to DNA. Together, our data provide new insights into the evolution of nucleotide specificity in McrBC restriction complexes and the molecular mechanisms governing McrBC-catalyzed DNA translocation and cleavage.

Pseudomonas aeruginosa is an opportunistic human pathogen capable of infecting a wide range of tissues and organs. Its persistence during chronic infection is strongly associated with biofilm formation, which depends on extracellular polysaccharides such as alginate. The HtrA-like periplasmic serine protease MucD is a key regulator of bacterial virulence, stress response, and alginate production, yet its molecular mechanism has remained largely unclear. Here, we discovered the alginate acetylation and export proteins AlgX and AlgK as MucD substrates, and characterized their degradation by mass spectrometry and bioinformatic analysis. We further determined the cryo-EM structure of MucD bound to an AlgK-derived substrate peptide, offering atomic insights into MucD oligomerization assembly, substrate recognition, and specificity. Together with structure-guided mutagenesis and biochemical assays, our results revealed that MucD proteolytic activity is governed by an equilibrium between a resting 12-mer and an active trimer. Crucially, we demonstrate that MucD represses alginate biosynthesis post-translationally, in addition to its previously implicated role in transcriptional regulation. These findings define a distinct activation mechanism and regulatory function for MucD and provide new insight into bacterial HtrA-like serine proteases.

Infectious bronchitis virus (IBV) is a member of the Gammacoronavirus genus responsible for respiratory illness and weakened eggshells in infected chickens, adversely impacting the poultry industry. Escaping innate immune detection during infection is crucial for coronavirus proliferation in the host. The production of double-stranded RNA during coronavirus replication triggers innate immune sensors to create an antiviral state within infected cells. To counter this response, coronaviruses employ nonstructural protein 15 (nsp15) endoribonuclease to degrade double-stranded RNA. Here, we use cryo-electron microscopy and biochemistry to characterize IBV nsp15 interactions with RNA. While the overall structure and active site of IBV nsp15 strongly resemble previous studies of nsp15 from other coronaviral genera, we note that double-stranded RNA contacts several non-conserved residues peripheral to the enzyme active site. Our data show that these residue positions can have strong impacts on RNA cleavage suggesting unique solutions for RNA engagement across coronavirus species. We also demonstrate a preference for IBV nsp15 to cleave double-stranded RNA over single-stranded RNA and observe nsp15 hexamers with two double-stranded RNAs bound simultaneously. This study reinforces the need to study diverse coronavirus species to identify distinct viral enzyme characteristics.

We present the solution structures of HIV-1 proteins NC(p7)1-55 corresponding to the full-length NC(p7) and mature p6. The studies were carried in water and, to mimic the membrane, in micellar DPC (Dodecylphosphocholine) conditions. Our results unravel for the first time the structure adopted by the N-terminal amino acids of the free NC(p7)1-55, with the formation of a small helix spanning residues F6 to R10. Our NMR and Fluorescence Anisotropy data disclose an interaction between NC(p7)1-55 and p6 both in water and DPC, with respective Kd of 2.5mM and 370 mM at 23{degrees}C. The interaction is thus strengthened in lipidic conditions. Protein p6 stabilizes the N-terminus of NC(p7)1-55 while increasing at the same time the dynamic of the first zinc finger. Although the entire p6 sequence is involved in the interaction, we show that its C-terminal region is particularly sensitive to the presence of NC(p7)1-55, with a propensity of forming a a helix ranging from amino acids S111 to F116. This study brings experimental evidence of a direct protein-protein interaction between p6 and the N-terminal region of NC(p7)1-55. We further show that such interaction is readily accommodated within the NC(p15) framework and hypothesize that it may facilitate the selective assembly of assembly of the viral genomic RNA (gRNA) in the cell.

Protein kinases function as dynamic, mechanically coupled nodes, yet the conformational drivers of multimeric activation remain unclear. Here, we present AlloQuant, a computational suite that translates AlphaFold3 structural ensembles into quantitative metrics of kinase regulation, including internal network rigidity, metastable-state populations, and sub-angstrom conformational drivers. Applying AlloQuant to CDK1, we demonstrate that binding of the Cyclin B1 (CCNB1) cofactor mechanically decouples a hyper-rigid inactive kinase core, allowing activating phosphorylation (pT161) to subsequently re-impose localized tension on the catalytic machinery. Conversely, the C-terminal Src kinase (CSK) faces a distinct conformational trap. While nucleotide-free monomeric CSK spontaneously samples a pre-active geometry, ATP binding excludes the active C-In conformation in all but 1 of 225 models. We show that docking partner engagement overcomes this blockade. Autophosphorylation of SRC at the activation loop (Y419) redistributes SRC conformational states without altering bulk rigidity. This redistribution is structurally coupled to the conformational state of CSK via the regulatory spine, not the catalytic machinery. Rather than mechanically deforming CSK, SRC engagement acts by conformational selection, committing roughly a quarter of CSK molecules to a fully active state. Thus, trans-allosteric kinase activation operates by defining the accessible conformational landscape of the receiver kinase. That control is exerted through mechanical remodeling in cofactor-dependent complexes and through conformational selection in transient kinase-kinase heterodimers. These findings establish AlloQuant as a general framework for quantifying how a binding partner reshapes a kinase's conformational landscape, applicable across the kinome because it assigns landmarks by profile-HMM alignment.

Gap junction (GJ) channels mediate direct intercellular communication by allowing the exchange of ions, metabolites, and signaling molecules between neighboring cells. Humans express 21 connexin (Cx) isoforms that can assemble into homotypic or heterotypic channels, creating a large potential interaction landscape that shapes tissue-specific communication networks. However, the rules governing which connexin isoforms can compatibly dock remain incompletely defined. Extracellular loop 2 (EL2) sequence features have been implicated in docking specificity and used to classify connexins into two canonical compatibility groups, K-N and H, but these assignments remain largely predictive. Most potential heterotypic connexin pairings have never been experimentally tested. This incomplete interaction map limits our ability to predict which connexin combinations can assemble, how isoform co-expression shapes intercellular communication, and how these relationships are altered or exploited in disease and engineered systems. Here, we used the FETCH (Flow Enabled Tracking of Connexosomes in HEK Cells) assay to evaluate docking compatibility across the complete human connexin family. To support family-wide compatibility mapping, we used literature-supported heterotypic interactions to define a data-driven FETCH score threshold for high-confidence interaction compatibility. Homotypic FETCH measurements varied substantially across the 21 connexin isoforms, with 15 producing mean scores above the empirical threshold. We then extended FETCH analysis to all 210 pairwise heterotypic isoform combinations. The resulting interaction landscape largely recapitulated expected motif-class relationships, including enrichment within the two canonical compatibility groups, but also identified neighboring-group interactions and unexpected cross-group pairings that represented clear exceptions to class-based predictions. Consistent with these findings, pairwise EL2 motif similarity was only modestly associated with threshold-based interaction classification, indicating that EL2 similarity alone was insufficient to predict compatibility outcomes. Together, these findings suggest that motif class provides a broad organizing framework for connexin compatibility, but that pairwise docking specificity also depends on yet-unresolved isoform-specific determinants that produce neighboring-group relationships and clear cross-group exceptions. Notably, Cx46, a lens Cx also associated with melanoma and breast cancers, emerged as a broadly permissive isoform capable of interacting with partners from both major compatibility groups and more than half of the connexin family. Together, these findings establish the first family-wide experimental atlas of human connexin docking compatibility, defining canonical interactions, previously unrecognized pairings, and exceptions to established compatibility rules. This atlas provides a foundation for defining the molecular determinants of connexin specificity, understanding how isoform diversity shapes intercellular communication, and designing gap junction channels with controlled docking behavior.

High-quality experimental datasets that link protein-ligand structures with binding affinity data are essential for developing and evaluating structure-based machine learning methods. To help address this need, we established OpenBind as an open-science initiative to generate large-scale experimental datasets for structure-based AI and molecular discovery. Here, we describe the first public OpenBind release, which, to the best of our knowledge, is the largest public single-target experimental structure-affinity dataset. The dataset focuses on enteroviral 2A protease, comprising 925 crystallographic binding events from 699 compounds and associated affinity measurements for 601 compounds. It combines structures from an initial fragment screen and follow-on molecules, together with affinity data, linking experimentally determined protein-ligand binding modes to biophysical measurements within a coherent antiviral discovery campaign. We used this dataset to evaluate protein-ligand structure prediction, binding-affinity prediction, and virtual screening using representative structure-based methods, including docking and cofolding. This exposed several challenges that are central to practical structure-based modelling: docking performance depends strongly on binding-pocket conformation, poses are difficult to rank, and structure-based affinity prediction remains challenging. Fine-tuning OpenFold3-p2 on the fragment-screen structures substantially improved pose prediction and virtual screening for related follow-on compounds, demonstrating how early-stage experimental structures can support target-specific model adaptation.