Acute myeloid leukemia (AML) is a hematologic malignancy characterized by impaired myeloid differentiation and uncontrolled proliferation. Central to its pathogenesis is the dysfunction of transcription factors that normally maintain hematopoietic homeostasis by directing lineage commitment and restricting cell growth. Although the genetic and epigenetic alterations that affect these factors have been extensively characterized, the biophysical mechanisms underlying their contribution to leukemogenesis remain incompletely understood. This review integrates insights from structural biology, chromatin profiling, transcriptome analyses, and pharmacological studies to delineate the biophysical principles by which key myeloid transcription factors-PU.1, IRF8, CEBPA, RUNX1, and GFI1-exert tumor-suppressive functions. We focus on how sequence-specific DNA recognition, cooperative complex assembly, chromatin accessibility, and dynamic protein–protein interactions govern transcriptional regulation in both normal and leukemic hematopoiesis. Disruption of these transcription factor networks may arise from mutations that alter the DNA-binding affinity, isoform switching that reshapes cofactor interactions, or the formation of aberrantly repressive complexes. These alterations converge on transcriptional reprogramming that favors proliferation over differentiation. Notably, the biophysical interfaces which mediate these interactions, such as the SNAG domain–lysine-specific demethylase 1 (LSD1) axis in GFI1, represent promising targets for therapeutic intervention. Elucidating these molecular mechanisms provides a framework to understand the disease heterogeneity and to develop differentiation-based therapies in AML.

The endoplasmic reticulum (ER) and mitochondria (MT) are two essential organelles within cells. Endoplasmic reticulum stress (ERS) and mitochondrial damage are key drivers of cell death in myocardial ischemia-reperfusion injury (MIRI), and their interplay also exerts important regulatory effects on the progression of cardiomyocyte injury. As a structural hub for physical coupling between the ER and mitochondria, mitochondria-associated endoplasmic reticulum membranes (MAMs) play a pivotal role in the pathogenesis and progression of MIRI. Disruption of MAMs’ architecture in cardiomyocytes can directly or indirectly trigger multiple pathological processes, including redox imbalance, ER stress, mitochondrial damage, energy depletion, and programmed cell death. This review first elaborates on the structural changes and functional consequences of the ER and mitochondria in MIRI, summarizes the structural features of MAMs and the regulatory functions and specific mechanisms of their enriched proteins in modulating calcium homeostasis, oxidative stress, and mitochondrial damage under MIRI pathology, and finally lists several bioactive components derived from traditional Chinese medicine that ameliorate MIRI by targeting MAMs-associated proteins.

Microtubules are dynamic biophysical polymers whose function is encoded by a combinatorial regulatory framework, the “tubulin code”, in which α/β-tubulin isotype composition, post-translational modifications (PTMs), and lattice conformation collectively determine effector recruitment and cellular behavior. Among these PTMs, α-tubulin detyrosination has become a particularly tractable model for biophysical investigation following the identification of its enzymatic machinery. In this review, we summarized progress across six thematic areas: Methodological advances enabling causal investigation of the tubulin code; the conceptual framework of isotype- PTM epistasis; the molecular mechanism and spatial regulation of detyrosination by the vasohibin- SVBP (VASH/SVBP) complex and MATCAP/TMCP family; microtubule lattice mechanics and plus- end dynamics; effector-mediated transport sorting; and pathophysiological consequences in neurological disease, cardiac dysfunction, and cancer. A central insight was that microtubule lattice conformation, expanded (GTP-like) versus compacted (GDP), constitutes a previously unrecognized third axis of the tubulin code. Lattice expansion enhances α-tubulin C-terminal tail (CTT) accessibility and selectively promotes VASH1/SVBP catalytic activity, mechanically gating detyrosination. This state-gated mechanism couples polymer biophysics to downstream transport selectivity, with consequences for axonal differentiation, chromosome segregation, ciliary organization, and cardiac mechanics. Neuronal axonal microtubules maintaining a GTP-like expanded lattice despite being GDP-bound further illustrates lattice conformation as an independent regulatory variable. Nevertheless, no single experimental system has demonstrated this complete causal cascade, from isotype composition through lattice mechanical state and PTM enzyme accessibility to reader recruitment, in an integrated manner. Establishing a unified reconstitution and imaging workflow to validate this cascade, and defining how its breakdown drives disease, stands as the central experimental challenge for tubulin biophysics. Microtubules are not merely structural scaffolds but a dynamic information- processing system whose biophysical logic is only beginning to be deciphered.

Voltage-gated T-type calcium channels (Cav3.1) are important therapeutic targets for neurological and cardiovascular disorders; the selective blocker Z944 is among the most potent inhibitors currently available. However, the molecular basis underlying the contribution of its halogen substituents to ligand recognition remains poorly understood. In this study, all-atom molecular dynamics simulations totaling 4.8 μs were performed to investigate the binding dynamics of the human Cav3.1 pore domain in the apo state and in complex with Z944 or a dehalogenated analogue lacking the chlorine and fluorine substituents (mZ944). Ligand binding preserved the overall pore architecture while locally stabilizing the blocker-binding cavity without inducing large-scale conformational changes. Relative MM/GBSA binding-energy calculations showed more favorable calculated binding energetics for Z944 than for mZ944, primarily because of greater dispersion interactions, with domain II residues, particularly Phe956, Asn952, Leu920, and Thr921, forming the dominant energetic contributions. Dynamic interaction analyses further revealed that Z944 maintains a broader protein–ligand interaction network, whereas mZ944 displayed greater positional and conformational variability within the binding cavity. Because mZ944 was initialized from the experimentally determined Z944 pose, its persistence within the same binding cavity over the simulated timescale does not exclude alternative binding modes. Moreover, because chlorine and fluorine were removed simultaneously, the observed differences reflect their combined removal rather than their individual contributions. Analysis of ligand conformational dynamics showed that the halogenated ligand occupies a more restricted conformational ensemble, favoring well-defined torsional states compatible with stable binding. Overall, these findings indicate that the presence of the chlorine and fluorine substituents is associated with more favorable calculated binding energetics, more favorable dispersion interactions, differences in direct and water-mediated interaction patterns, and reduced ligand conformational heterogeneity. This work provides atomistic insight into the molecular mechanism of Z944 recognition and establishes structural principles that may guide the rational design of next-generation Cav3.1 inhibitors.

RNA performs diverse biological functions that are often encoded in its three-dimensional structure. However, RNA-only structures represent a small fraction (~1%) of entries in the Protein Data Bank (wwPDB), limiting our knowledge of RNA structure–function relationships. The ribose-phosphate backbone of RNA is relatively flexible compared to proteins and typically forms weaker long-range tertiary interactions, resulting in structures that are less amenable to X-ray crystallography than folded proteins. Further, many functional RNAs range from 50 to 200 nt in size and are often too small for structure determination by cryo electron microscopy (cryo-EM). To facilitate RNA structure determination by cryo-EM, we present the procapsid RNA (pRNA)-assisted RNA imaging scheme (pARIS). In this approach, the RNA of interest is fused to the pRNA of bacteriophage Φ29 to enhance stability and reduce sample heterogeneity. The pRNA-linked RNA is then assembled onto the Φ29 procapsid to form a stable pentameric complex. This approach increases effective molecular mass, improves signal-to-noise ratio, and facilitates particle picking and alignments during cryo-EM imaging processing. Using pARIS, we determined a 4.8 Å structure of a 70 nt (~23 kDa) tRNA as proof of principle and a 6.7 Å structure of the liver-specific host microRNA-122 bound to its target site in the hepatitis C virus genome, demonstrating applicability to biologically relevant RNA targets.

A mathematical model of neuroinflammation is proposed that describes the interplay among the peripheral inflammatory response, microglial activation, neural tissue damage, changes in blood–brain barrier permeability, and the effects of preventive glucocorticoid administration. The model is formulated as a system of ordinary differential equations and is intended to investigate the conditions governing the transition from an acute inflammatory response to a stable chronic state. The analysis includes verification of the models’ well-posedness, identification of equilibria, local stability analysis, bifurcation analysis, and parameter sensitivity analysis. Three dynamic regimes were identified: rapid resolution of inflammation, delayed recovery accompanied by persistent changes in the central nervous system, and transition to chronic neuroinflammation. The development of chronic neuroinflammation was shown to be associated with strengthening of the positive feedback circuit between central inflammatory mediators and activated microglia. The parameters exerting the greatest influence on the duration of the inflammatory response, accumulation of neural tissue damage, and decline in the effectiveness of glucocorticoid-mediated regulation were identified. The model may be used to interpret experimental data, assess the contributions of individual components of the inflammatory response, and investigate the conditions underlying the development of chronic neuroinflammation.

Identifying protein binding sites is essential for mechanistic biology and structure-based drug discovery. However, single static structures often fail to capture functionally relevant sites that are shallow, transient, allosteric, or cryptic. This review focuses primarily on protein binding-site prediction and presents a unified biophysical framework emphasizing conformational dynamics, thermodynamics, and solvent-mediated recognition. We evaluated current experimental and computational approaches, with particular attention to molecular dynamics, enhanced sampling, Markov state models, and mixed-solvent molecular dynamics for identifying transient cavities and chemically favorable hotspots across conformational ensembles. Rather than relying on single-score predictions, we propose a multi-factor evidence framework that integrates pocket emergence, probe enrichment, local chemistry, reproducibility, and functional or ligand-based support. RNA binding regions and macromolecular interfaces are discussed as important extensions, where related principles of conformational heterogeneity, electrostatics, hydration, and molecular recognition also apply. Finally, we examine the emerging role of ensemble-aware artificial intelligence in integrating structural, dynamic, thermodynamic, and experimental information, and highlight the need for interpretable, uncertainty-aware, and prospectively validated prediction strategies.