Nuclear speckle proteins form intrinsic and MALAT1-dependent microphases.

Abstract

Pre-mRNA processing components in nuclear speckles encompass one or more folded RNA recognition motifs (RRMs) and disordered regions with specific sequence grammars. Such proteins include serine/arginine-rich splicing factors (SRSFs) and transactive response DNA binding protein (TDP)-43. The SRSFs and TDP-43 are unique archetypes of block copolymers encoding specific patterns of inter-domain homotypic and heterotypic attractions and repulsions. The interplay of these interactions drives microphase separation and the formation of ordered, size-limited assemblies. Microphases of SRSFs and TDP-43 are 23-45 nm in diameter, each comprising tens of molecules. Sub-micron-scale assemblies of SRSFs in cells are consistent with being clusters of microphases. The speckle-associated regulatory long non-coding RNA (lncRNA) metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) binds specifically and preferentially to SRSF1 microphases, while destabilizing TDP-43 microphases. In protein mixtures, the interactions between microphases drive the formation of micron-scale double-emulsion structures with core-shell organization. Our findings show how interactions involving copolymers featuring folded domains and disordered regions drive the formation of microphases.