Enzymes and nucleic acid processing machines are among the most sophisticated molecules evolved by nature. These include several vital enzymes such as polymerases, nucleases, topoisomerases, and RNA enzymes. These enzymes exhibit complex structures and multicomponent assemblies that enable an exceptional efficiency and specificity in carrying out the fundamental chemical reactions required for life. Yet, these structures pose a challenge to our understanding of their catalytic precision, as this requires atomistic insight into how they operate on DNA or RNA. Indeed, the static or quasi-static view from experiments fails to fully explore the conformational space that these enzymes traverse during catalysis. Thus, advanced molecular simulation techniques are increasingly integrated with experimental results to overcome such sampling limitations and to explore the configurational space, thereby defining the functional dynamics of structural motions and rearrangements that would otherwise remain unclear. In this Account, I will examine how atomistic multiscale molecular simulations and free energy calculations─recently coupled with AI-guided enhanced sampling methods─have contributed to clarifying enzymatic mechanisms that could only be hypothesized by examining structures, as evidenced by increasingly complex structural biology results. This is a time when we can simulate ultralarge realistic model systems comprising proteins, nucleic acids, ions, and water molecules, all of which are critical components for enzymatic function. Remarkably, such modeling and simulations can nowadays enable the prospective exploration of increasingly structurally complex systems. In this scenario, there is today a relevant body of work that defines function through motion as a key element in these enzymes. My group has generated a broad collection of results on a range of nucleic acid processing enzymes, from polymerases to nucleases and RNA enzymes, which altogether highlight how dynamics define function, with specific residues─often evolutionarily conserved and localized in the second coordination shell of the catalytic core─that are strategically positioned to operate in a cooperative and highly coordinated fashion for specific enzymatic actions on nucleic acids. These coordinated residue motions, near the reaction center, appear distinct from large allosteric movements of distal domains. Instead, they resemble finely tuned, small-scale mechanisms that marry complexity with the precise, clocklike efficiency evolution has built into the catalytic core. How such elaborate enzymatic architectures move-to-function, captured and demonstrated by the most recent computational work and structural data, will therefore be the core topic discussed in this Account. While the concept of "function through motion" can obviously be extended to other catalytic systems, what is particularly remarkable about these enzymes─even if not unique─is that their functional complexity is handled with extraordinary accuracy through precise motions, despite the system's architectural complexity, on the move. These recent mechanistic findings also provide insights into how to engineer or modulate these enzymes, with implications for several scientific activities centered on nucleic acid chemistry.

We present a set of Grand Challenges for predictive modeling in small molecule drug discovery, with the goal of defining, prioritizing, and quantifying the areas where computation can have transformative impact. Rather than offering another broad survey of methods, this paper articulates specific scientific and technical problems that limit progress today and proposes measurable criteria by which advances can be judged. Our objective is to align researchers, investors, and industry leaders around the challenges that matter most for advancing real drug discovery programs. This work seeks to broaden participation in drug discovery by providing a clear roadmap for contributors from the growing areas of artificial intelligence, machine learning, robotics, high-performance computing, and quantum computing, where tools are rapidly advancing but are often disconnected from the practical realities of medicinal chemistry and pharmacology. The insights presented here draw on extensive discussions with experienced drug hunters, computational method developers, and thought leaders across biotech, pharma, software, and venture capital, as well as lessons learned from our own drug discovery efforts. While there is substantial enthusiasm (particularly around AI) for revolutionizing drug discovery, this moment demands sharper problem definition. Without clearly articulated challenges and performance metrics, computational innovation risks optimizing for benchmarks rather than for translational impact. We therefore identify Grand Challenges across four domains: Chemistry, Structure, Energy, and Pharmacology. For each domain, we present a common framework: a well-defined challenge, the underlying physical principles, its relevance to drug discovery decision-making, the current state of the field, and quantitative metrics that define progress. By grounding computational ambition in concrete scientific problems, we aim to catalyze sustained, measurable advances rather than episodic waves of enthusiasm.

CDC42 subfamily members of the RHOGTPases exist in either an active or inactive state, each with a distinct but highly flexible structure, making them challenging to target. Abnormal CDC42 activity is linked to various conditions, including cancer, neurological and ophthalmological disorders, skin diseases, and vascular conditions. This has driven increasing interest in developing CDC42 inhibitors, which have recently produced lead compounds demonstrating disease-modifying effects in vivo. In this review, we provide an overview of the physiological functions of CDC42 subfamily members and their emerging roles in disease. We also examine the latest advances in leveraging CDC42's structural features for selective inhibition, analyzing the mode of action and chemical structures of current inhibitors. These insights aim to inform the design of improved molecules with broad therapeutic potential.

Locally advanced head and neck carcinoma remains associated with high morbidity and long-term survival below 50%. Treatment commonly relies on cisplatin-based chemoradiotherapy, which is effective but frequently associated with significant acute and chronic systemic toxicities. Therefore, safer and more effective therapeutic strategies are urgently needed. FLASH radiotherapy (RT) has emerged as a promising irradiation modality because of its potential to reduce damage to healthy tissues while preserving antitumor efficacy. Here, we investigated the anticancer activity of the Topoisomerase (Topo)-II inhibitor ARN-24139, alone and combined with FLASHRT, in human papillomavirus-negative SCC-25 head and neck carcinoma biomodels. Antitumor activity was assessed in 2D cell cultures using viability, apoptosis, clonogenic, wound-healing, and γH2AX assays, as well as in SCC-25 3D spheroids and in chorioallantoic membrane (CAM) tumor models. ARN-24139 induced dose-dependent cytotoxicity in SCC-25 cells, with IC50 values of 7.3 ± 0.8 µM at 48 h and 7.2 ± 0.5 µM at 72 h, while showing limited toxicity in healthy HBEpC cells. Sequential low-dose FLASH-RT followed by ARN-24139 enhanced antitumor activity, reducing cell viability at 4 Gy after 8 days and decreasing tumor growth and Ki67 expression in CAM models. These proof-of-concept findings support further investigation in more clinically representative and mechanistically informative HNSCC models.

Abstract Neural network potentials (NNPs), trained on quantum mechanical (QM) data, can deliver near QM-level accuracy while being much faster than QM calculations. In this work, we introduce a new software tool, Torsionator, which currently supports MACE, OBIWAN, and UMA NNPs for energy minimization of small molecules. Torsionator can analyze a set of selected conformers to parameterize the dihedral angles. These dihedral parameters can then be used for molecular mechanics calculations and simulations. We showcase Torsionator by performing dihedral scanning and parameterization of two representative antiviral compounds: favipiravir defluoro analog T1105 and emivirine. To thoroughly evaluate its robustness and general applicability, we benchmarked Torsionator using MACE on the TorsionNet500 dataset, comprising 500 chemically diverse small molecules with reference QM torsional profiles. These results indicate Torsionator as a practical, efficient, and scalable software tool for dihedral parameterization, enabling its routine integration into molecular simulation workflows.

DNA polymerases (Pols) are essential enzymes for DNA replication within the cell. However, DNA lesions, such as cyclobutane pyrimidine dimers (CPDs) induced by ultraviolet (UV) radiation, can impair Pols's function and DNA replication. Therefore, the presence of CPDs can, for example, lead to xeroderma pigmentosum variant (XP-V), a rare genetic disorder characterized by an increased risk of skin cancer. Nonetheless, in these situations, specific translesion synthesis (TLS) Pols, such as human DNA polymerase η (Polη), can overcome such lesions, enabling DNA polymerization. That is, Polη prevents the pathological risks associated with CPD-caused DNA replication stalling. Here, we analyzed how a selected set of 8 Polη mutations perturbs its structure, DNA binding, and substrate translocation, thereby altering Polη function, preventing it from bypassing CPDs, thus making them XP-V pathogenic mutations. Leveraging recent structural and clinical data on these pathogenic Polη variants, we elucidated the mechanistic basis for their impairment of Polη's ability to bypass damage. We employed molecular dynamics simulations to examine their effects on Polη in pre- and post-translocation states. Although these residues vary in location and chemical nature, we found that all contributed to reducing DNA anchoring to Polη. In this way, we could identify a unified mechanistic framework for decoding how XP-V pathogenic mutations compromise Polη function by destabilizing the Polη-DNA complex.

Many RNAs rely on their 3D structures for function. While acquiring functional 3D structures, certain RNAs form misfolded, non-functional states ('kinetic traps'). Instead, other RNAs sequentially assemble into their functional conformations over pre-folded scaffolds. Elucidating the principles of RNA sequential assembly is thus important to understand how RNAs avoid the formation of misfolded, non-functional states. Integrating single-particle electron cryomicroscopy (cryo-EM), image processing, in solution small-angle X-ray scattering (SAXS), EM-driven molecular dynamics (MD) simulations, structure-based mutagenesis, and enzymatic assays, we have visualized the sequential multidomain assembly of a self-splicing ribozyme of biomedical and bioengineering significance. Our work reveals a distinct dynamic interplay of helical subdomains in the ribozyme's 5'-terminal scaffold, which acts as a gate to control the docking of 3'-terminal domains. We identify specific conserved and functionally important secondary structure motifs as the key players for orchestrating the energetically inexpensive conformational changes that lead to the productive formation of the catalytic pocket. Our work provides a near-atomic resolution molecular movie of a large multidomain RNA assembling into its functionally active conformation and establishes a basis for understanding how RNA avoids the formation of non-functional 'kinetic traps'.

Protein engineering of cutinases is a promising strategy for the biocatalytic degradation of non-natural polyesters. We report a mechanistic study addressing the hydrolysis of the aliphatic polyester poly(butylene succinate, or PBS) by the fungal Apergillus oryzae cutinase enzyme. Through atomistic molecular dynamics simulations and advanced alchemical transformations, we reveal how three units of a model PBS substrate fit the active site cleft of the enzyme, interacting with hydrophobic side chains. The substrate ester moiety approaches the Asp-His-Ser catalytic triad, displaying catalytically competent conformations. Acylation and deacylation hydrolytic reactions were modeled according to a canonical esterase mechanism using umbrella sampling simulations at the quantum mechanical/molecular mechanical DFT(B3LYP)/6-31G**/AMBERff level. The free energy profiles of both steps show a high-energy tetrahedral intermediate resulting from the nucleophilic attack on the ester's carboxylic carbon. The free energy barrier of the acylation step is higher (20.2 ± 0.6 kcal mol-1) than that of the deacylation step (13.6 ± 0.6 kcal mol-1). This is likely due to the interaction of the ester's carboxylic oxygen with the oxyanion hole in the reactive conformation of the deacylation step. In contrast, these interactions form as the reaction proceeds during the acylation step. The formation of an additional hydrogen bond interaction with the side chain of Ser48 is crucial to stabilizing the developing charge at the carboxylic oxygen, thus lowering the activation free energy barrier. These mechanistic insights will inform the design of enzyme variants with improved activity for plastic degradation.

Na+-K+-Cl- cotransporters functions as an anion importers, regulating trans-epithelial chloride secretion, cell volume, and renal salt reabsorption. Loop diuretics, including furosemide, bumetanide, and torsemide, antagonize both NKCC1 and NKCC2, and are first-line medicines for the treatment of edema and hypertension. NKCC1 activation by the molecular crowding sensing WNK kinases is critical if cells are to combat shrinkage during hypertonic stress; however, how phosphorylation accelerates NKCC1 ion transport remains unclear. Here, we present co-structures of phospho-activated NKCC1 bound with furosemide, bumetanide, or torsemide showing that furosemide and bumetanide utilize a carboxyl group to coordinate and co-occlude a K+, whereas torsemide encroaches and expels the K+ from the site. We also found that an amino-terminal segment of NKCC1, once phosphorylated, interacts with the carboxyl-terminal domain, and together, they engage with intracellular ion exit and appear to be poised to facilitate rapid ion translocation. Together, these findings enhance our understanding of NKCC-mediated epithelial ion transport and the molecular mechanisms of its inhibition by loop diuretics.

R132H IDH1 is an important therapeutic target for a variety of brain cancers, yet drug leads and radiotracers which selectively bind only to the mutant over the wild type are so far lacking. Here we have predicted the structural determinants of the Michaelis complex of this mutant using a QM/MM MD-based protocol. It shows some important differences with the X-ray structure, from the metal coordination to the positioning of key residues at the active site. In particular, one lysine residue (K212) emerges as a mostly likely proton donor in the key proton-transfer step of the R132H IDH1 catalytic reaction. Intriguingly, the same residue in its deprotonated state is likely to be involved in the reaction catalyzed by the wild-type enzyme (though the mechanisms are different). Our QM/MM protocol could also be used for other metal-based enzymes, which cannot be modelled easily by force field-based MD, like in this case.

Ligand-coated gold nanoparticles (AuNPs) can act as self-organized nanoreceptors capable of selectively recognizing small organic molecules (analytes) in solution. This ability can be applied in several fields, with NMR chemosensing being a notable example. To advance the rational design of such AuNP-based nanosensors, we present a data-driven scoring function to rapidly estimate AuNP-analyte binding affinities, thus allowing fast in silico prescreening of ligand-coated AuNP sensors. This scoring function implements chemical similarity, hydrophobicity, and charge complementarity as key molecular descriptors, demonstrating excellent predictive accuracy when validated against experimental data (R2 = 0.85, MAE = 0.45 kcal/mol). Enhanced sampling molecular dynamics on representative systems revealed that ligand flexibility, monolayer packing, and hydrogen bonding critically shape binding interactions, particularly for weak binding systems. Together, these data-driven and atomistic insights offer a robust framework for the rational design and optimization of AuNP-based nanosensors.

In the Big Data era, a change of paradigm in the use of molecular dynamics is required. Trajectories should be stored under FAIR (findable, accessible, interoperable and reusable) requirements to favor its reuse by the community under an open science paradigm.

Computational screening of a 100-tripeptide nanoparticle library ranked the nanoparticles according to their affinity for the neuroblastoma biomarker 3-methoxytyramine (3-MT). Synthesis and testing confirmed their ability to detect 3-MT at 25 μM.

Recent cryoelectron microscopy data have revealed significant conformational rearrangements of the megadalton spliceosome structure during splicing, an essential catalytic process for maturing most human transcripts. However, the molecular trigger of these structural rearrangements for splicing catalysis remains unclear at the atomic level. Here, by analyzing a consistent dataset of multicomponent spliceosome structures, we identified a minimal set of positively charged residues whose dynamic action remodels the spliceosome active site during the first step of splicing. Through equilibrium and enhanced sampling molecular dynamics simulations of multiple splicing intermediates (>2M atoms), we uncover how these residues dynamically operate in coordination with the transient, temporally ordered binding of key spliceosomal proteins (Prp11, Prp8, and Yju2) at the spliceosome core. Our findings reveal a molecular mechanism that ensures precise and timely spliceosome activation, opening promising directions to probe splicing regulation and potentially target its dysfunction in diseases, with broad implications for drug discovery and medicine.

Human topoisomerase II (topoII) is a well-known and validated target for cancer treatment. We previously reported a first set of 6-amino-tetrahydroquinazoline derivatives as novel human topoII inhibitors. Here, we report on the expansion and this molecular scaffold and present 17 additional analogs centered on the tetrahydropyrido[4,3-d]pyrimidine heterocycle. Some of these compounds exhibit promising topoII inhibitory and antiproliferative activities. Compound 24 (ARN21929) shows good in vitro potency, with an IC50 of 4.5 ± 1.0 µM, excellent kinetic and thermodynamic solubility, and good metabolic stability. Our results indicate that this new chemical class of topoII inhibitors deserves further exploration and represents a promising starting point for developing novel and potentially safer topoII-targeted anticancer drugs.

The translocation of DNA in polymerase (Pol) enzymes is a critical step for Pol-mediated nucleic acid polymerization, essential for storing and transmitting genetic information in all living organisms. During this process, the newly elongated double-stranded DNA has to shift along the Pol enzyme to recreate the initial configuration at the metal-aided reactive center, where nucleotide addition can occur recurrently at every catalytic cycle. Double-stranded DNA translocation, therefore, allows the enzyme to add one more nucleotide to the growing strand, complementary to the template strand, without the enzyme dissociating from the DNA. Yet, the dynamic mechanism by which the Pol·DNA complex accomplishes DNA translocation remains poorly understood at the atomistic level. Here, leveraging recent structural data on DNA polymerase η (Polη), we elucidate its translocation mechanism, which we show to occur via an enzyme motion where the shift of Polη is asynchronous along the two DNA strands. Through equilibrium molecular dynamics and deep-learning-guided enhanced sampling simulations, we found that such a mechanism relies precisely on a set of positively charged residues of the enzyme that operate in a coordinated way at the Polη·DNA interface. Moving like screen wipers, such a dynamic mechanism of these residues promotes DNA translocation. These findings now offer new avenues to comprehend further such a complex yet fundamental dynamic process for DNA polymerization.

The self-splicing group II introns are bacterial and organellar ancestors of the nuclear spliceosome and retro-transposable elements of pharmacological and biotechnological importance. Integrating enzymatic, crystallographic, and simulation studies, we demonstrate how these introns recognize small molecules through their conserved active site. These RNA-binding small molecules selectively inhibit the two steps of splicing by adopting distinctive poses at different stages of catalysis, and by preventing crucial active site conformational changes that are essential for splicing progression. Our data exemplify the enormous power of RNA binders to mechanistically probe vital cellular pathways. Most importantly, by proving that the evolutionarily-conserved RNA core of splicing machines can recognize small molecules specifically, our work provides a solid basis for the rational design of splicing modulators not only against bacterial and organellar introns, but also against the human spliceosome, which is a validated drug target for the treatment of congenital diseases and cancers. Splicing is a vital biological reaction and a druggable pathway to treat infection, genetic diseases and cancer. Here, the authors describe how splicing is modulated by small molecules that target the conserved splicing active site in the group II introns.