TDP-43 and FUS are nuclear proteins with multiple functions in mRNA processing. They play key roles in ALS (amyotrophic lateral sclerosis) and FTD (frontotemporal dementia), where they are partially lost from the nucleus and aggregate in the cytoplasm of neurons and glial cells. Defects in nucleocytoplasmic transport contribute to this pathology, hence nuclear import of both proteins has been studied in detail. However, their nuclear export routes remain poorly characterized and it is unclear whether aberrant nuclear export contributes to TDP-43 or FUS pathology. Here we show that predicted nuclear export signals in TDP-43 and FUS are non-functional and that both proteins are exported independently of the export receptor CRM1/Exportin-1. Silencing of Exportin-5 or the mRNA export factor Aly/REF, as well as mutations that abrogate RNA-binding do not impair export of TDP-43 and FUS. However, artificially enlarging TDP-43 or FUS impairs their nuclear egress, suggesting that they could leave the nucleus by passive diffusion. Finally, we found that inhibition of transcription causes accelerated nuclear egress of TDP-43, suggesting that newly synthesized RNA retains TDP-43 in the nucleus, limiting its egress into the cytoplasm. Our findings implicate reduced nuclear retention as a possible factor contributing to mislocalization of TDP-43 in ALS/FTD.

Progeny capsids of herpesviruses leave the nucleus by budding through the nuclear envelope. Two viral proteins, the membrane protein pUL34 and the nucleo-phosphoprotein pUL31 form the nuclear egress complex that is required for capsid egress out of the nucleus. All pUL31 orthologs are composed of a diverse N-terminal domain with 1 to 3 basic patches and a conserved C-terminal domain. To decipher the functions of the N-terminal domain, we have generated several Herpes simplex virus mutants and show here that the N-terminal domain of pUL31 is essential with basic patches being critical for viral propagation. pUL31 and pUL34 entered the nucleus independently of each other via separate routes and the N-terminal domain of pUL31 was required to prevent their premature interaction in the cytoplasm. Unexpectedly, a classical bipartite nuclear localization signal embedded in this domain was not required for nuclear import of pUL31. In the nucleus, pUL31 associated with the nuclear envelope and newly formed capsids. Viral mutants lacking the N-terminal domain or with its basic patches neutralized still associated with nucleocapsids but were unable to translocate them to the nuclear envelope. Replacing the authentic basic patches with a novel artificial one resulted in HSV1(17+)Lox-UL31-hbpmp1mp2, that was viable but delayed in nuclear egress and compromised in viral production. Thus, while the C-terminal domain of pUL31 is sufficient for the interaction with nucleocapsids, the N-terminal domain was essential for capsid translocation to sites of nuclear egress and a coordinated interaction with pUL34. Our data indicate an orchestrated sequence of events with pUL31 binding to nucleocapsids and escorting them to the inner nuclear envelope. We propose a common mechanism for herpesviral nuclear egress: pUL31 is required for intranuclear translocation of nucleocapsids and subsequent interaction with pUL34 thereby coupling capsid maturation with primary envelopment.

Herpes simplex virus type 1 (HSV-1) glycoprotein M (gM/UL10) is a 473 aa type III transmembrane protein that resides in various membrane compartments. HSV-1 gM contains several putative trafficking motifs, but their functional relevance remains to be elucidated. We show here that transiently expressed gM 19-343 was sufficient for transport to the transGolgi network (TGN), whilst gM 133-473, where the first two transmembrane domains were deleted, and gM 1-342, which lacked the final residue of the last transmembrane domain, were retained in the endoplasmic reticulum (ER), indicating that all transmembrane domains are required for proper folding and ER exit. A series of bacterial artificial chromosome mutants revealed that in addition to the authentic start codon, translation of gM can be initiated at methionine 19 and 133/135. Whilst a protein lacking the first 18 residues supported WT-like growth, gM 133/135-473 resulted in reduced plaque diameters resembling a UL10 deletion mutant. An HSV-1 mutant encoding gM 1-342 showed similar growth characteristics and accumulated non-enveloped cytoplasmic particles, whilst gM 1-343 resulted in a gain of function, indicating that all transmembrane domains of the protein are important for viral growth. A C-terminal extension further supported viral propagation; however, the C-terminal trafficking motifs (residues 423-473) were completely dispensable. We propose a functional core within gM 19-343 comprised of all transmembrane domains that is sufficient to target the protein to the TGN, a favoured site for envelopment, and to support viral functions.

The herpes simplex virus type 1 (HSV-1) glycoprotein N (gN/UL49.5) is a type I transmembrane protein conserved throughout the herpesvirus family. gN is a resident of the endoplasmic reticulum that in the presence of gM is translocated to the trans Golgi network. gM and gN are covalently linked by a single disulphide bond formed between cysteine 46 of gN and cysteine 59 of gM. Exit of gN from the endoplasmic reticulum requires the N-terminal core of gM composed of eight transmembrane domains but is independent of the C-terminal extension of gM. Co-transport of gN and gM to the trans Golgi network also occurs upon replacement of conserved cysteines in gM and gN, suggesting that their physical interaction is mediated by covalent and non-covalent forces. Deletion of gN/UL49.5 using bacterial artificial chromosome (BAC) mutagenesis generated mutant viruses with wild-type growth behaviour, while full deletion of gM/UL10 resulted in an attenuated phenotype. Deletion of gN/UL49.5 in conjunction with various gM/UL10 mutants reduced average plaque sizes to the same extent as either single gM/UL10 mutant, indicating that gN is nonessential for the function performed by gM. We propose that gN functions in gM-dependent as well as gM-independent processes during which it is complemented by other viral factors.