PhD Scientific Days 2022

Molecular Sciences V.

Network analysis of liquid-liquid phase separation and membraneless organelles

Text of the abstract

Introduction
Network science is an interdisciplinary field, which can be used to examine complex systems, for example such intracellular mechanisms as liquid-liquid phase separation (LLPS). Eukaryotic cells consist of membrane-bound and membraneless organelles (MLOs). MLOs are mainly built of proteins and nucleic acids via LLPS. Although they have an effect on every cellular process; their exact biological function is unclear. MLOs may transition into pathological aggregates, contributing to several neurodegenerative diseases and many types of cancer where aberrant LLPS has been described. LLPS components can be classified into drivers (can phase separate on their own), regulators (activators of the process) and clients (form condensate with drivers).

Aims
Our aim is to build a network model to examine how LLPS regulates intracellular processes to understand its elements’ physiological and pathological function in different MLOs.

Methods
We took the SIGNOR signaling- and the HuRI protein-protein interaction network as a basis. The LLPS dataset was provided by our collaborating partner, Péter Tompa’s group. For the analysis we used NetworkX Python package among the ModuLand and EntOpt Cytoscape plug-ins.

Results
We have identified a new LLPS-client subgroup - the driver-neighbour-clients (DNCs), which have outstanding network topological properties. DNCs are enriched among network hubs, have high betweenness centrality and form local module cores. We made a special inquiry about the stress granule, which together with the P-body, regulates RNA storing, degradation and participation in translation. As they are overlapping structures, analyzing their strongly interacting modules may reveal the condensate-specific role of DNCs. DNCs appear to have a more important topological role than drivers themselves in stress granule networks.

Conclusion
Examination of LLPS using network science tools may help to clarify the role of LLPS components, such as DNCs in this process. Based on their network properties, DNCs may have an important role in MLO formation. With the help of this new point of view novel drug targets could be identified in LLPS-associated diseases.

Funding
Semmelweis 250+ Excellence Scholarship, NKFI 2020-4.1.1.-TKP2020 excellence program, TKP2021-EGA excellence program, K-13148 research grant of the Hungarian National Research and Development Office.