SARS-CoV-2 nucleocapsid protein undergoes liquid-liquid phase separation stimulated by RNA and partitions into phases of human ribonucleoproteins

Abstract

Tightly packed complexes of nucleocapsid protein and genomic RNA form the core of viruses and may assemble within viral factories, dynamic compartments formed within the host cells. Here, we examine the possibility that the multivalent RNA-binding nucleocapsid protein (N) from the severe acute respiratory syndrome coronavirus (SARS-CoV-2) compacts RNA via protein-RNA liquid-liquid phase separation (LLPS) and that N interactions with host RNA-binding proteins are mediated by phase separation. To this end, we created a construct expressing recombinant N fused to a N-terminal maltose binding protein tag which helps keep the oligomeric N soluble for purification. Using in vitro phase separation assays, we find that N is assembly-prone and phase separates avidly. Phase separation is modulated by addition of RNA and changes in pH and is disfavored at high concentrations of salt. Furthermore, N enters into in vitro phase separated condensates of full-length human hnRNPs (TDP-43, FUS, and hnRNPA2) and their low complexity domains (LCs). However, N partitioning into the LC of FUS, but not TDP-43 or hnRNPA2, requires cleavage of the solubilizing MBP fusion. Hence, LLPS may be an essential mechanism used for SARS-CoV-2 and other RNA viral genome packing and host protein co-opting, functions necessary for viral replication and hence infectivity.

Figure 1:. Domain structure, sequence comparison, and sequence composition of SARS-CoV-2 nucleocapsid protein.

A) SARS-CoV-2 N contains three putatively disordered regions, a globular N-terminal and a globular C-terminal oligomerization domain. Sequence alignment of N from SARS-CoV-2 and SARS-CoV showing 91% sequence identity. B) Sequence composition of the intrinsically disordered regions of SARS-CoV-2 N.

Figure 2:. MBP-CoV-2 N and cleaved CoV-2 N elute larger than their predicted molecular weights.

A) Gel filtration chromatogram of ~7ml of ~500 μM MBP tagged CoV-2 N (top) and ~300μl ~50 μM CoV-2 N (previously cleaved from MBP) (bottom). Vertical lines represent peak elution volumes of gel filtration protein calibration standards. Brackets represent range of gels lanes. Predicted molecular weight of the tagged and untagged N are 90 kDa and 46 kDa respectively, much smaller than their corresponding calibrated peak elution volumes, consistent with oligomerization. B) SDS-PAGE showing the peak fractions of the MBP-CoV-2 N chromatogram showing expected 90 kDa MW. C) SDS-PAGE showing the peak fractions of the cleaved CoV-2 N chromatogram showing expected 46 kDa MW. Circular feature in SDS-PAGE gels is a divet in plastic on which pictures were taken.

Figure 3:. SARS-CoV-2 nucleocapsid protein undergoes LLPS at physiological conditions.

A) Phase separation over time as monitored by turbidity of 50 μM MBP-N after addition of TEV protease in varying pH conditions. B-C) DIC micrographs of 50 μM MBP-N in 50 mM Tris 183 mM NaCl pH 7.4 or 20 mM MES 183 mM NaCl pH 5.5 with and without TEV protease (to cleave MBP from N) or 0.3 mg/mL desalted total torula yeast RNA. Scale bars represent 50 μm.

Figure 4:. SARS-CoV-2 N LLPS is modulated by salt and RNA.

A-B) Phase separation over time as monitored by turbidity of 50 μM MBP-N in 50 mM Tris pH 7.4 after addition of TEV protease with varying torula yeast RNA (at 100 mM sodium chloride) or varying sodium chloride concentrations. C-D) DIC micrographs of 50 μM MBP-N in 50 mM Tris NaCl pH 7.4 with varying torula yeast RNA or sodium chloride concentrations with and without TEV protease (to cleave MBP from N). Scale bars represent 50 μm.

Figure 5:. SARS-CoV-2 N partitions into liquid phases of hnRNPs.

A) N partitions into hnRNPA2 LC droplets even with the MBP tag attached. B) N partitions into TDP-43 CTD droplets even with the MBP tag attached. C) N partitions into FUS LC droplets only after cleavage of the MBP tag. D-F) N partitions into hnRNPA2 FL (D), TDP-43 FL (E), FUS FL (F) droplets.