Biophysical Society Thematic Meeting | Trieste 2024

Emerging Theoretical Approaches to Complement Single-Particle Cryo-EM

Poster Abstracts

8-POS Board 8 CLOSING OF THE SF3B SPLICING COMPLEX: INSIGHT FROM SIMULATIONS WITH MACHINE-LEARNED COLLECTIVE VARIABLES Pavel Janoš 1 ; Alessandra Magistrato 1 ; 1 CNR - Istituto Officina dei Materiali (IOM), Trieste, Italy RNA splicing is a crucial stage in gene expression wherein introns are removed from pre-RNA and exons are joined to form mature RNA, either protein-encoding mRNA or functional non coding RNA. The spliceosome, a complicated ribonuclear machine, plays a crucial role in precise pre-RNA splicing [1] and if deregulated can lead various diseases. [2] Spliceosome is composed of five small nuclear RNAs (snRNA) and tens of associated protein factors, which all undergo dynamical conformational changes. One of the important conformational transitions is the formation of the B act spliceosome associated with closing of the SF3b complex upon recognition of the branch point sequence, during which the pre-RNA is engulfed and the bulged branch point adenosine is positioned in a specific pocket [3]. Cryo-EM studies have made tremendous strides in shedding light into the splicing process [3]. However, they mostly provide static snapshots, while computational techniques can uncover the dynamical transformation between different stages in the splicing cycle and thus help fill-in the gaps between experimental results. The conformational change that the SF3b complex undergoes is essential for understanding key aspects of splicing, including the mechanism of small-molecule splicing modulators [5]. However simulating it is challenging due to the size and complexity of the system. To overcome this challenge, we apply the deepLDA machine learning approach [6] to obtain a collective variable that can describe such a complex conformational transformation. Set of distances and angles between distinct part of the SF3b complex [7] are used as the input parameters that get compressed by deepLDA into a single collective variable able to discriminate the open and closed states of SF3b complex. Using this collective variable with enhanced sampling methods we are able to simulate the opening/closing of the SF3b complex and obtain key mechanistic insight into spliceosome, including the effect of splicing modulators. [1] Wilkinson, M. E., Charenton, C., & Nagai, K. (2020). RNA splicing by the spliceosome. Annual review of biochemistry, 89, 359-388. [2] Borišek, J., Casalino, L., Saltalamacchia, A., Mays, S. G., Malcovati, L., & Magistrato, A. (2020). Atomic-level mechanism of pre-mRNA splicing in health and disease. Accounts of Chemical Research, 54(1), 144-154. [3] Sun, Chengfu. "The SF3b complex: splicing and beyond." Cellular and Molecular Life Sciences 77.18 (2020): 3583 3595 [4] Tholen, J., & Galej, W. P. (2022). Structural studies of the spliceosome: bridging the gaps. Current Opinion in Structural Biology, 77, 102461. [5] Rozza, R., Janoš, P., Spinello, A., & Magistrato, A. (2022). Role of computational and structural biology in the development of small-molecule modulators of the spliceosome. Expert Opinion on Drug Discovery, 17(10), 1095-1109. [6] Bonati, L., Rizzi, V., & Parrinello, M. (2020). Data-driven collective variables for enhanced sampling. The journal of physical chemistry letters, 11(8), 2998-3004. [7] Melo, Marcelo CR, et al. "Generalized correlation-based dynamical network analysis: a new high performance approach for identifying allosteric communications in molecular dynamics trajectories." The Journal of Chemical Physics 153.13 (2020): 134104.

58