Nuclear RNA binding regulates TDP-43 nuclear localization and passive nuclear export

Abstract

Nuclear clearance of the DNA/RNA-binding protein TDP-43 is a pathologic hallmark of amyotrophic lateral sclerosis and frontotemporal dementia that remains unexplained. Moreover, our current understanding of TDP-43 nucleocytoplasmic shuttling does not fully explain the predominantly nuclear localization of TDP-43 in healthy cells. Here, we used permeabilized and live-cell models to investigate TDP-43 nuclear export and the role of RNA in TDP-43 localization. We show that TDP-43 nuclear efflux occurs in low-ATP conditions and independent of active mRNA export, consistent with export by passive diffusion through nuclear pore channels. TDP-43 nuclear residence requires binding to GU-rich nuclear intronic pre-mRNAs, based on the induction of TDP-43 nuclear efflux by RNase and GU-rich oligomers and TDP-43 nuclear retention conferred by pre-mRNA splicing inhibitors. Mutation of TDP-43 RNA recognition motifs disrupts TDP-43 nuclear accumulation and abolishes transcriptional blockade-induced TDP-43 nuclear efflux, demonstrating strict dependence of TDP-43 nuclear localization on RNA binding. Thus, the nuclear abundance of GU-rich intronic pre-mRNAs, as dictated by the balance of transcription and pre-mRNA processing, regulates TDP-43 nuclear sequestration and availability for passive nuclear export.

Multivalent GU-rich oligonucleotides sequester TDP-43 in the nucleus by inducing high molecular weight RNP complexes

“…In ('AUG12')2 and Clip34nt-transfected cells, but not in (GU) 16 Only trace TDP-43 binding to (CA)16 was observed. As we previously reported, in (GU)6-transfected cells, SAV precipitated predominantly 43-45 kDa TDP-43 monomers, 28 whereas in (GU)16-transfected cells, SAVprecipitated numerous higher molecular weight bands that were not multiples of 43 kD. Clip34nt and ('AUG12')2 produced similar findings, although at slightly shifted molecular weights, likely as a result of the different sizes of the crosslinked oligos.…”

“…Previously, we showed that (GU)6 rapidly entered transfected HeLa cell nuclei, bound TDP-43, and promoted nuclear efflux, likely by competitive displacement of TDP-43 from endogenous nuclear RNAs, freeing the resulting ~45 kD complex for passive diffusion through nuclear pore channels. 28 Beyond shifts in nuclear/cytoplasmic localization, oligonucleotide-transfected cells displayed variable changes in the TDP-43 distribution pattern. SAV labeling showed that, in addition to diffuse signal, all four oligonucleotides formed scattered puncta in the cytoplasm suggesting a tendency to demix and self-associate.…”

“…The nuclear export of full-length TDP-43, initially thought to occur via exportin-1 (XPO1) binding to a putative nuclear export signal (NES) in the RRM2 domain, was subsequently shown to occur by passive diffusion through nuclear pore channels and is exquisitely size-dependent. [25][26][27][28] In multiple export assays, the addition of bulky tags markedly delayed TDP-43 nuclear export by raising its size above the permeability barrier of the nuclear pore complex (NPC), which increasingly excludes cargoes above 30-60 kDa in size. 29 An exception is the short TDP-43 isoform, which is upregulated by pathologic hyperexcitability and localizes to the cytoplasm via a distinct NES for XPO1 arising in the C-terminus.…”

“…30 We recently demonstrated that the nuclear localization of full-length TDP-43 critically depends on binding to nuclear RNAs, which sequester TDP-43 in the nucleus and restrict its availability for passive nuclear export. 28 In permeabilized and live cell nuclear export assays, depletion of nuclear RNA binding sites for TDP-43, via RNase degradation or inhibition of transcription, released TDP-43 for passive diffusion from the nucleus. RRM mutations or displacement of TDP-43 from endogenous nuclear RNAs with monovalent (GU)6 oligonucleotides also promoted cytoplasmic mislocalization.…”

Polyadenylated RNA and RNA-Binding Proteins Exhibit Unique Response to Hyperosmotic Stress

“…It is possible that the shifting of the RNA and the RBPs is linked, with one “piggybacking” into the cytoplasm with the other. This has been thoroughly demonstrated in the case of TDP-43, whose localization is tightly linked with its mRNA targets ( Duan et al, 2021 ). Whether or not the RBPs we present here are shifting under a similar mechanism is unknown.…”