The XPO6 Exportin Mediates Herpes Simplex Virus 1 gM Nuclear Release Late in Infection

Abstract

The glycoprotein M of herpes simplex virus 1 (HSV-1) is dynamically relocated from nuclear membranes to the trans-Golgi network (TGN) during infection, but molecular partners that promote this relocalization are unknown. Furthermore, while the presence of the virus is essential for this phenomenon, it is not clear if this is facilitated by viral or host proteins. Past attempts to characterize glycoprotein M (gM) interacting partners identified the viral protein gN by coimmunoprecipitation and the host protein E-Syt1 through a proteomics approach. Interestingly, both proteins modulate the activity of gM on the viral fusion machinery. However, neither protein is targeted to the nuclear membrane and consequently unlikely explains the dynamic regulation of gM nuclear localization. We thus reasoned that gM may transiently interact with other molecules. To resolve this issue, we opted for a proximity-dependent biotin identification (BioID) proteomics approach by tagging gM with a BirA* biotinylation enzyme and purifying BirA substrates on a streptavidin column followed by mass spectrometry analysis. The data identified gM and 170 other proteins that specifically and reproducibly were labeled by tagged gM at 4 or 12 h postinfection. Surprisingly, 35% of these cellular proteins are implicated in protein transport. Upon testing select candidate proteins, we discovered that XPO6, an exportin, is required for gM to be released from the nucleus toward the TGN. This is the first indication of a host or viral protein that modulates the presence of HSV-1 gM on nuclear membranes.IMPORTANCE The mechanisms that enable integral proteins to be targeted to the inner nuclear membrane are poorly understood. Herpes simplex virus 1 (HSV-1) glycoprotein M (gM) is an interesting candidate, as it is dynamically relocalized from nuclear envelopes to the trans-Golgi network (TGN) in a virus- and time-dependent fashion. However, it was, until now, unclear how gM was directed to the nucleus or evaded that compartment later on. Through a proteomic study relying on a proximity-ligation assay, we identified several novel gM interacting partners, many of which are involved in vesicular transport. Analysis of select proteins revealed that XPO6 is required for gM to leave the nuclear membranes late in the infection. This was unexpected, as XPO6 is an exportin specifically associated with actin/profilin nuclear export. This raises some very interesting questions about the interaction of HSV-1 with the exportin machinery and the cargo specificity of XPO6.

Keywords: BioID; INM; RANBP20; Ran; Ran-binding protein 20; UL10; XPO6; exportin; herpes; herpes simplex; herpesvirus; host-pathogen interactions; inner nuclear membrane; intracellular transport; nuclear egress.

FIG 1

The C-terminal tail (Ct) of gM is cytoplasmic. Control untransfected cells or expressing gM BirA*HA were monitored by flow cytometry to define the localization of gM’s Ct. The cells were mock treated (A and B) or permeabilized (C and D) to determine the presence of the antigenic epitopes within the cells or at their surface, respectively. (A and C) Control untransfected cells were unstained (red) or stained with an anti-CD81 antibody coupled to APC-Cy7-A (gray). (B and D) Both cell types were unstained (red) or stained with an anti-HA antibody (green), no first antibody (orange), and anti-γ-tubulin antibody (blue). CD81 is a surface marker in HEK293 cells, and γ-tubulin is a cytoplasmic marker. (E) Schematic representation of gM-BirA*HA based on the above results. HA, hemagglutinin tag; NFA, no first antibody. Results are representative of two independent experiments.

FIG 2

gM BirA*HA relocates from the nuclear membrane at 4 hpi toward the TGN at 12 hpi in transiently expressing cells. (A and B) 143B cells were seeded on coverslips and transiently transfected with pcDNA5FRT gM BirA*HA for 24 h. Cells were then infected with HSV-1ΔgM2 (to only detect exogenous gM) at an MOI of 5 for 4 h or 12 h to exclusively monitor the tagged gM. Cells were stained with anti-HA and anti-TGN antibodies (A) or with anti-NPC and anti-gM antibodies (B). In all cases, Hoechst was used to label the nuclei, and images were captured by confocal microscopy. (C) The expression level of each construct was determined by Western blotting using anti-HA or anti-gM antibodies. (D) Cells were incubated for 24 h with d-biotin (50 μM), total cell lysates were harvested, and BirA* activity was monitored by Western blotting using streptavidin-HRPO (Strep-HRPO). In panels C and D, 25 μg of materials was loaded on the gels. An immunoblot against γ-tubulin was used as a loading control. Results are representative of three independent experiments. NT, not transfected. The molecular masses of the markers are indicated to the left of the gels.

FIG 3

The BirA*HA and gM BirA*HA cell lines harbor a functional biotinylation enzyme. Total cell lysates were prepared from the HEK293 FLP-In, HEK293 BirA*HA, and HEK293 gM BirA*HA cell lines. For HEK293 NLS BirA*HA, the cell lysates were additionally subfractionated into cytoplasmic (Cyt) and nuclear (Nu) fractions to monitor the functionality of the NLS tag. (A) The expression of each construct was determined by Western blotting using anti-HA or anti-gM antibodies as indicated to the right of the blots. As control, GAPDH, which is cytoplasmic, was also monitored. (B) To confirm the enrichment of the cytoplasmic and nuclear fractions, they were probed for the presence of nuclear pore complexes (NPC) with a mixture of antibodies detecting p152, p110, and p62. (C) To evaluate if the BirA* constructs were enzymatically functional, the cell lines were incubated for 24 h with d-Biotin (50 μM), and BirA* activity was probed by Western blotting using streptavidin-HRPO (Strep-HRPO) to detect biotinylated proteins. In panels A and C, GAPDH was used as a loading control. Thirty micrograms was loaded in all lanes. NT, not transfected. Results are representative of three independent experiments. The molecular masses of the markers are indicated to the left of the gels.

FIG 4

gM BirA*HA shuttles as expected between the nuclear membrane and the TGN. HEK293 BirA*HA (A), HEK293 NLS-BirA*HA (B), and HEK293 gM BirA*HA (C) cell lines were mock treated or infected for 4 or 12 h with HSV-1ΔgM2. The TGN and nuclei were then labeled (with antibodies and Hoechst, respectively), and expression of the constructs was monitored with anti-HA antibodies. Scale bars, 10 μm. Results are representative of three independent experiments.

FIG 5

Biotinylated protein enrichment and MS analysis. (A) Schematic presentation of the strategy adopted for the d-biotin incubation and infection of the cell lines. Cells were incubated with d-biotin (50 μM, 24 h) and infected with HSV-1ΔgM2 for the last 4 or 12 h. (B and C) HEK293 BirA*HA, HEK293 NLS BirA*HA, and HEK293 gM BirA*HA preincubated with biotin were infected for 4 or 12 hpi, and total cell lysates were harvested. A 10% fraction of these total lysates (TL) was set aside, and remaining samples were captured on streptavidin columns. BB, bead bound. All were analyzed by Western blotting (2 μg loaded) using streptavidin-HRPO. (D) A Venn diagram shows the number of proteins identified at 4 hpi, 12 hpi, or at both times that were identified in all three independent experiments but absent from the BirA*HA and NLS BirA*HA controls.

FIG 6

Diversity of protein classes identified by BioID-MS that interact with HSV-1 gM. The 170 host proteins identified in our proteomics approach (Fig. 5D) were analyzed by the software PANTHER to delineate their protein classes and then plotted in Excel. PC numbers refer to the protein class as defined by PANTHER.

FIG 7

Inhibition by RNA interference of select targets implicated in vesicular transport. (A) 143B cells were transfected for 48 h with LipoJet alone or with dsiRNAs targeting SNX1, SNX2, SCAMP3, VPS33B, MTMR6, TMEM43, or XPO6. Their viability was measured with alamarBlue. (B) In parallel, the effect of the dsiRNAs on the relative gene expression (%) was measured by RT-qPCR. The mean values and SEMs from three independent experiments are shown. The data are normalized on the LipoJet alone (A) or nonspecific RNAi (dsiCTL) (B). (C) HSV-1 had minimal impact on the endogenous levels of the targets, as assessed by RT-qPCR. Cells were mock treated or infected with wild-type HSV-1 at an MOI of 5 for 4 or 12 hpi. (D) The dsiRNAs were equally efficient in mock and infected scenarios. The relative expression of the hits in mock-infected cells or cells infected with HSV-1 (MOI of 5, 4 hpi) were analyzed by RT-qPCR. The means values and SEMs from three independent experiments are shown. The data are normalized to the dsiCTL. Bilateral Student’s t tests were performed to detect significant hits compared to the reagent alone or dsiCTL. *, P < 0.05; **, P < 0,01; ***, P < 0,001; ****, P < 0,0001.

FIG 8

Inhibition of SNX1, SNX2, SCAMP3, VPS33B, MTMR6, and TMEM43 does not affect the localization of gM at 4 hpi. 143B cells seeded on coverslips were transfected for 48 h with LipoJet reagent alone or dsiRNA targeting the above-listed proteins and then infected for 4 hpi with wild-type HSV-1. Immunofluorescence was performed using antibodies recognizing gM or lamin B receptor (LBR), an inner nuclear membrane marker. Scale bars, 20 μm. Note that we previously showed that the punctate gM is associated with nuclear invaginations (11).

FIG 9

XPO6 is required for the nuclear export of gM at 12 hpi. 143B cells seeded on coverslips were transfected for 48 h with LipoJet reagent or dsiRNA targeting XPO6 and TMEM43 and then infected for 4 or 12 h with wild-type HSV-1. Immunofluorescence was performed as for Fig. 8 at either 4 (A) or 12 (B) hpi. Scale bars, 20 μm. NT, nontransfected cells.

FIG 10

XPO6 depletion does not impact viral yields. (A) Equal numbers of 143B cells were transfected for 48 h with nonspecific RNAi (dsiCTL) or dsiXPO6. Since none of the tested antibodies detected endogenous XPO6 proteins, we immunoprecipitated the protein then probed RNAi efficacy by Western blotting (top). As control, we probed γ-tubulin levels in the original total lysates used for the IP (bottom). (B) 143B cells were not treated (NT) or transfected with control RNAi or those targeting XPO6 for 48 h and then infected with HSV-1 at an MOI of 2. The supernatants were harvested at 18 hpi and viral titers were determined on Vero cells.