Loop extrusion by structural maintenance of chromosome (SMC) complexes has emerged as a key mechanism of genome organization across domains of life, yet the rules that govern where these motors load, unload, and become enriched remain poorly understood. The canonical model treats cohesin as loading and unloading uniformly but stopping at extrusion barriers. Chromosome architecture is then shaped by where extruding complexes stop. Accumulating evidence broadens this view by uncovering cases of (i) targeted SMC loading across metazoa, plants, and fungi, (ii) sites of specific SMC release, and (iii) domains of crowded loop extrusion. Together, these new Rules of the Road shift the focus from where extruders stop to where they start, end, and dwell—with implications for development, gene regulation, and genome stability.
The spliceosome is a complex macromolecular machine that catalyses the essential cellular process of precursor mRNA (pre-mRNA) splicing. In the past decade, high-resolution structures for numerous stable spliceosome assembly intermediates became available, providing a structural framework for the basic splicing mechanism. However, the field still lags behind in mechanistic understanding of the precise trajectories of individual splicing factors, the regulation and organisation of splicing within the cellular context, and the mechanisms underlying splicing quality control and its links to human disease. In this review, we highlight how recent structural studies have bridged some of these knowledge gaps and brought forward our mechanistic understanding of pre-mRNA splicing. We discuss recent insights into the exon-definition problem, the role of assembly factors in small ribonucleoprotein particle (snRNP) biogenesis, spliceosome-associated quality control and the U12-dependent splicing pathway. We also provide our perspective on the future development of the field.
With recent advances in cryo-electron microscopy/cryo-electron tomography (cryoEM/cryoET), computational capabilities and machine learning, there has been an incredible increase in the speed of structure determination for large biomolecular assemblies. In order to connect these static descriptions with biological function, one needs reliable and accurate approaches to quantify the factors that regulate dynamics. For this, models that employ simplified energetics and/or coarse-grained representations have become indispensable tools for integrating structural insights, biochemical analyses and physical principles. Here we discuss examples where molecular simulations have provided a framework for understanding dynamics in assemblies of increasing size and complexity. We focus on three classes for which there have been notable advances in recent years: ribosomes, viral capsids and nuclear pore complexes. Each possesses unique physical attributes that have required distinct theoretical and computational strategies. Together, these studies provide a rich conceptual and technical foundation that can help shape our perspective on the machinery of the cell.
Molecular simulations have been central to our understanding of how enzymes work, and the ongoing explosion of artificial intelligence/ML (AI/ML) approaches is greatly expanding the scope of problems that can be addressed using simulations. The speed of development of the field is tremendous, with the near-constant development of new approaches and successful applications. This review aims to give a higher-level overview of four key areas where AI/ML approaches are significantly transforming computational enzymology (both understanding and design), specifically (1) enzyme discovery and functional annotation, (2) kinetic parameter prediction, (3) understanding of protein dynamics and enzyme reaction mechanisms, and (4) decoding protein allostery. We also discuss some of the outstanding technical challenges in this research space, and efforts being made to overcome them.
