1. Immunology and Inflammation
  2. Microbiology and Infectious Disease

The interferon-inducible GTPase MxB promotes capsid disassembly and genome release of herpesviruses

  1. Manutea C Serrero
  2. Virginie Girault
  3. Sebastian Weigang
  4. Todd M Greco
  5. Ana Ramos-Nascimento
  6. Fenja Anderson
  7. Antonio Piras
  8. Ana Hickford Martinez
  9. Jonny Hertzog
  10. Anne Binz
  11. Anja Pohlmann
  12. Ute Prank
  13. Jan Rehwinkel
  14. Rudolf Bauerfeind
  15. Ileana M Cristea
  16. Andreas Pichlmair
  17. Georg Kochs
  18. Beate Sodeik  Is a corresponding author
  1. Institute of Virology, Hannover Medical School, Germany
  2. RESIST - Cluster of Excellence, Hannover Medical School, Germany
  3. Institute of Virology, Technical University Munich, Germany
  4. Institute of Virology, Freiburg University Medical Center, University of Freiburg, Germany
  5. Department of Molecular Biology, Princeton University, United States
  6. MRC Human Immunology Unit, MRC Weatherall Institute of Molecular Medicine, Radcliffe Department of Medicine, University of Oxford, United Kingdom
  7. German Center for Infection Research (DZIF), Hannover-Braunschweig Partner Site, Germany
  8. Research Core Unit Laser Microscopy, Hannover Medical School, Germany
  9. German Center for Infection Research (DZIF), Munich Partner site, Germany
https://doi.org/10.7554/eLife.76804
Share this article
Cite this article
  1. Manutea C Serrero
  2. Virginie Girault
  3. Sebastian Weigang
  4. Todd M Greco
  5. Ana Ramos-Nascimento
  6. Fenja Anderson
  7. Antonio Piras
  8. Ana Hickford Martinez
  9. Jonny Hertzog
  10. Anne Binz
  11. Anja Pohlmann
  12. Ute Prank
  13. Jan Rehwinkel
  14. Rudolf Bauerfeind
  15. Ileana M Cristea
  16. Andreas Pichlmair
  17. Georg Kochs
  18. Beate Sodeik
(2022)
The interferon-inducible GTPase MxB promotes capsid disassembly and genome release of herpesviruses
eLife 11:e76804.
https://doi.org/10.7554/eLife.76804
  2. Download BibTeX
  3. Download .RIS

Abstract

Host proteins sense viral products and induce defence mechanisms, particularly in immune cells. Using cell-free assays and quantitative mass spectrometry, we determined the interactome of capsid-host protein complexes of herpes simplex virus and identified the large dynamin-like GTPase myxovirus resistance protein B (MxB) as an interferon-inducible protein interacting with capsids. Electron microscopy analyses showed that cytosols containing MxB had the remarkable capability to disassemble the icosahedral capsids of herpes simplex viruses and varicella zoster virus into flat sheets of connected triangular faces. In contrast, capsids remained intact in cytosols with MxB mutants unable to hydrolyse GTP or to dimerize. Our data suggest that MxB senses herpesviral capsids, mediates their disassembly, and thereby restricts the efficiency of nuclear targeting of incoming capsids and/or the assembly of progeny capsids. The resulting premature release of viral genomes from capsids may enhance the activation of DNA sensors, and thereby amplify the innate immune responses.

Editor's evaluation

This paper uses an innovative cell-free protein-protein interaction system to identify factors that interact with HSV-1 capsids in infected cells. In addition to cataloging numerous capsid-interacting proteins, the manuscript probes the antiviral mechanism of one of these, MxB. The data provide strong support for an intriguing model in which MxB "punches" holes in HSV-1 capsids, releasing viral DNA and potentially triggering host DNA sensors. Moreover, the results suggest that viral proteins bind to and shield the capsids from MxB attack, offering a new perspective on how viruses might evade some host defenses.

https://doi.org/10.7554/eLife.76804.sa0

Introduction

Infections with human alphaherpesviruses are associated with painful and stigmatizing manifestations such as herpes labialis or herpes genitalis, but also cause life-threatening meningitis or encephalitis, potentially blinding eye infections, herpes zoster, and post-herpetic neuralgia, particularly in immunocompromised patients (Gershon et al., 2015; Whitley and Roizman, 2016; Whitley and Johnston, 2021). Herpes simplex viruses (HSV-1, HSV-2) and varicella zoster virus (VZV) productively infect epithelial and fibroblast cells of the skin and mucous membranes as well as neurons, but are restricted in immune cells. Macrophages, Langerhans cells, dendritic cells, and NK cells mount potent immune responses against alphaherpesviruses (Whitley and Roizman, 2016).

Intracellular DNA sensors are crucial to sense herpesvirus infections, and to induce caspase-1-mediated inflammation and type I IFN expression (Hertzog and Rehwinkel, 2020; Kurt-Jones et al., 2017; Lum and Cristea, 2021; Ma et al., 2018; Paludan et al., 2019; Stempel et al., 2019). During an unperturbed infection, capsid shells shield herpesviral genomes from cytosolic sensors during nuclear targeting as well as after nuclear genome packaging (Arvin and Abendroth, 2021; Döhner et al., 2021; Knipe et al., 2021). HSV-1 capsids can withstand compressive forces of up to 6 nN which is more than sufficient to endure the 18 atm repulsive pressure of the packaged viral DNA (Bauer et al., 2013; Roos et al., 2009). So far, it is unclear how cytosolic DNA sensors gain access to herpesviral genomes; either cytosolic host factors disassemble the sturdy herpesviral capsids during infection, or the nuclear envelopes become leaky.

HSV-1 virions contain an amorphous tegument layer that links the icosahedral capsids with a diameter of 125 nm to the viral envelope proteins (Crump, 2018; Dai and Zhou, 2018; Diefenbach, 2015). To identify cytosolic proteins that promote or restrict infection by interacting with HSV-1 capsids, we have developed cell-free methods to reconstitute capsid-host protein complexes using tegumented capsids from extracellular viral particles or tegument-free capsids from the nuclei of infected cells (Radtke et al., 2014). Intact capsids are incubated with cytosol prepared from tissues or cultured cells, and the capsid-host protein complexes are isolated, and characterized by mass spectrometry (MS), immunoblot, electron microscopy, and functional assays. We could show that HSV-1 capsids require inner tegument proteins to recruit microtubule motors, to move along microtubules, to dock at nuclear pore complexes (NPCs), to release viral genomes from capsids, and to import viral genomes into the nucleoplasm, and that capsids lacking tegument cannot move along microtubules, but still bind to nuclear pores (Anderson et al., 2014; Ojala et al., 2000; Radtke et al., 2010; Wolfstein et al., 2006).

Here, we searched for proteins that might contribute to sensing cytosolic capsids and thereby promote the detection of herpesviral genomes. Using extracts of matured THP-1 cells, a model system for human macrophages (Tsuchiya et al., 1980) we identified type I interferon (IFN) inducible proteins that bound specifically to HSV-1 capsids. Among them was the large dynamin-like GTPase myxovirus resistance protein B (MxB). MxB limits the infection of several herpesviruses, and can mediate almost 50% of the IFN-mediated restriction of HSV-1, although its mode of action has remained elusive so far (Crameri et al., 2018, Liu et al., 2012, Schilling et al., 2018, Jaguva Vasudevan et al., 2018). MxB has been first described for its potent inhibition of HIV infection (Goujon et al., 2013; Kane et al., 2013, Liu et al., 2013). The human MX2 gene codes for a full-length MxB (residues 1–715) and a smaller version (residues 26–715) that lacks an N-terminal extension (NTE), which both are highly expressed upon IFN induction (Melén et al., 1996). MxB likely operates as an anti-parallel dimer but can also form higher-order filaments; its N-terminal GTPase domain connects to a bundle signaling element that moves relative to the GTPase domain in response to nucleotide binding, and the C-terminal stalk domain is critical for MxB oligomerization (Alvarez et al., 2017; Chen et al., 2017; Fribourgh et al., 2014; Gao et al., 2011).

We show here that both, full-length MxB(1-715) and MxB(26-715) have the remarkable property to disassemble the capsids of the three human alphaherpesviruses HSV-1, HSV-2, and VZV, so that they can no longer transport nor shield the viral genomes. Capsid disassembly did not require proteases but depended on the ability of MxB to hydrolyse GTP and to dimerize. As the large tegument protein pUL36 links the capsid vertices to the other tegument proteins (Crump, 2018; Dai and Zhou, 2018; Diefenbach, 2015), and as an increasing amount of associated tegument proteins protected capsids against MxB-mediated disassembly, we propose that MxB attacks the capsids at their vertices. Our data suggest that MxB can bind to and disassemble incoming as well as progeny capsids, and thereby might increase the sensing of cytosolic and nuclear viral genomes. Therefore, the MxB GTPase might be the sought-after capsid destroyer that acts upstream of cytosolic or nuclear sensors to promote viral genome detection and induction of innate immune responses.

Results

IFN induction prevents HSV-1 infection of macrophages

Before investigating capsid interactions with macrophage proteins, we compared HSV-1 infection in human keratinocytes (HaCat), pigment epithelial cells (RPE), and THP-1 cells at low, moderate, or high multiplicity of infection (MOI). We stimulated monocyte THP-1 cells with phorbol 12-myristate 13-acetate to differentiate them into a macrophage-like phenotype, and used them either directly (Mφ) or after a resting period of 3 days (MφR). HSV-1 replicated productively in HaCat and RPE cells up to 20 hpi, while a pre-treatment with IFN delayed and reduced but did not prevent the production of infectious virions (Figure 1). Both Mφ and MφR released 10–100-fold less infectious HSV-1, and an IFN pre-treatment prevented infection at all MOIs. Thus, Mφ and MφR restricted HSV-1 infection efficiently, and the induction of IFN-stimulated genes (ISGs) prevented any productive infection.

Figure 1
Download asset Open asset
IFN restricts HSV-1 infection in keratinocytes, epithelial cells, and macrophages.

HaCat, RPE, Mφ, or MφR cells were mock-treated or treated with human IFN-α (1000 U/mL) for 16 hr and were infected with HSV-1(17+)Lox at 2.5 × 106 (MOI 5), 2.5 × 107 (MOI 50), or 5 × 107 PFU/mL (MOI 100), and the amount of cell-associated and extracellular virions was titrated on Vero cells. Each data point represents the mean of the three technical replicates of the combined cell-associated and extracellular titers. The error bars represent the standard deviation.

IFN-induced protein changes in the cytosol of macrophages

To identify cytosolic macrophage proteins that might foster or restrict HSV-1 capsid functions, we prepared extracts from MφR or IFN-induced MφIFN to reconstitute capsid-host protein complexes as they might assemble in macrophages (Figure 2—figure supplement 1). Using subcellular fractionation and subsequent dialysis (Figure 2—figure supplement 2A), we depleted the extracts of nuclei and mitochondria (Figure 2—figure supplement 2B; pellet P1), cytoplasmic membranes such as Golgi apparatus, endoplasmic reticulum and plasma membrane (P1, P2), and small metabolites (S2, S3, S4). Furthermore, most of the cytoskeletal tubulin and actin sedimented into the first pellet (P1), while glyceraldehyde 3-phosphate dehydrogenase (GAPDH), a bona-fide cytosolic protein, remained soluble in the supernatants (S1, S2, S2’, S3, S4). Next, we analyzed the proteomes of the MφR and IFN-induced MφIFN cytosols at low ATP/GTP concentration [ATP/GTPlow] by mass spectrometry (MS; Supplementary file 1). We detected 494 (Figure 2—figure supplement 2C; black circles) of more than 600 reported IFN-inducible proteins (Rusinova et al., 2013). Of those, a Fisher’s exact test identified the interferomeDB, and in particular GALM, COL1A1, LGALS3BP, NT5C3A, IFI44, IFIT2, IFIT3, GBP4, SRP9, IFIT5, DSP, and L3HYPDH as enriched by at least 2.8-fold (log2 1.5) in the MφIFN cytosol (Figure 2—figure supplement 2C; red). These changes might reflect IFN-induced transcriptional or translational regulation, post-translational modification, subcellular localization, or susceptibility to proteolysis, and show that the IFN induction had changed the cytosol proteome of the MφIFN.

HSV-1 capsids interact with specific cytosolic macrophage proteins

To search for cytosolic Mφ proteins whose interactions with HSV-1 capsids depend on their surface composition, we generated tegumented viral V0.1, V0.5, and V1 capsids as well as D capsids with a reduced tegumentation (Figure 2—figure supplement 1). For this, we lysed extracellular particles released from HSV-1 infected cells with non-ionic detergent to solubilize the envelope proteins and lipids, and in the presence of 0.1, 0.5, or 1 M KCl to modify intra-tegument protein-protein interactions (Anderson et al., 2014; Ojala et al., 2000; Radtke et al., 2010; Radtke et al., 2014; Wolfstein et al., 2006, Zhang and McKnight, 1993). Furthermore, we dissociated tegument from V0.1 capsids by a limited trypsin digestion to generate so-called D capsids. We then incubated similar amounts of different capsid types as calibrated by immunoblot for the major capsid protein VP5 (Figure 2—figure supplement 2D) with cytosol at ATP/GTPlow from MφR or IFN-induced MφIFN for 1 hr at 37 °C. The capsid-host protein complexes assembled in vitro were harvested by sedimentation, and their interactomes were determined by quantitative MS (Figure 2—figure supplement 1). As before (Radtke et al., 2010; Snijder et al., 2017), the protein intensities were normalized across samples to the abundance of the major capsid protein VP5 (Supplementary file 2, host; Supplementary file 3, viral).

Of 2,983 proteins identified (Supplementary file 2), we detected 1816 in at least three of the four replicates in any of the eight different capsid-host protein complexes. Of those, 598 host proteins bound differentially to one capsid type over another (Supplementary file 2; fold change ≥2 (log2 1); permutation-based FDR ≤ 0.05). The HSV-1 capsids had recruited specifically 279 proteins of MφR and 390 of MφIFN cytosol of which 71 were shared. Hierarchical clustering analyses of the associated MφR or MφIFN proteins identified four major classes; for example one enriched on V over D capsids (Figure 2—figure supplements 3–4, top green) and one enriched on D over V capsids (Figure 2—figure supplements 3–4, bottom violet). Therefore, we further compared the capsid-host interactions of D capsids directly to V0.1 (Figure 2A, D), V0.5 (Figure 2B and E), or V1 (Figure 2C and F) capsids, and identified 82 proteins of MφR (Figure 2A, B and C) and 141 of MφIFN (Figure 2D, E and F) with 35 being shared (Supplementary file 2; difference ≥2.83 fold (log2 1.5); FDR ≤ 0.01). The MφR capsid-host complexes included 12 and the ones of MφIFN 19 proteins listed in the interferome database (Rusinova et al., 2013; red in Figure 2). Gene ontology and pathway enrichment analyses showed that the identified 82 MφR (Figure 3—figure supplement 1) and 141 MφIFN (Figure 3) proteins included many proteins implicated in innate immunity, intracellular transport, nucleotide and protein metabolism, as well as intracellular signaling. Overall, the host proteomes of V0.1 (red) and D (gray) capsids were rather distinct, but more similar for V0.5 (blue) and V1 (green) capsids (Figure 3—figure supplement 1, Figure 3). For example, V0.1 capsids had recruited specifically the innate immunity proteins PIGR, IGHA1, BPIFA1 and DEFA3, but D capsids LRRFIP1, UFC, C3 and DCD from MφR cytosol. In MφIFN, the D capsids were enriched for C3, C6, IGBP1, UBA5, UBXN1, UBE3A, and RNF123. These data suggest that protein domains displayed on different capsids interacted with specific cytosolic MφR or MφIFN proteins.

Figure 2 with 4 supplements see all
Download asset Open asset
Cytosolic IFN-induced macrophage proteins binding to HSV-1 capsids.

Volcano plots of iBAQs counts of proteins identified in capsid-host protein complexes assembled in cytosol from resting THP-1 φ cells (A - C) or treated with interferon-α (D - F) using V0.1 (A, D), V0.5 (B, E), or V1 (C, F) capsids in comparison to D capsids. Proteins identified as highly specific interactions are indicated with larger symbols (log2 difference ≥1.5; Welch’s t-test, two-tailed, permutation-based FDR ≤ 0.01); those with a log2 difference ≥4 are annotated. ISGs (interferome.org) are indicated by filled black circles, and are annotated in red if significantly enriched (permutation-based FDR ≤ 0.05, and log2 difference ≥1.5). Proteins with a q-value = 0 were imputed to - log10 q-value = 3.1 (maximum of the graph), and were indicated with empty circles.

Figure 3 with 1 supplement see all
Download asset Open asset
Cytosolic proteins of IFN-induced macrophages binding to HSV-1 capsids.

Host proteins from cytosol of IFN-stimulated MφIFN (c.f. D, E, F; abundance log2 difference larger than 1.5; significance permutation-based FDR smaller than 0.01) interacting with V0.1, V0.5, V1, or D capsids were assembled into a functional interaction network of known protein-protein-interactions (gray lines; STRING database, confidence score of 0.7), and grouped according to their known functions (Gene Ontology, Pathway analysis). The Pie chart for each protein indicates its relative enrichment on V0.1 (red), V0.5 (blue), V1 (green), or D capsids (gray).

In these assays, the capsids interacted with several proteins already validated to promote or restrict HSV or VZV infection. Examples are the ESCRT-III co-factor VPS4 (Cabrera et al., 2019; Crump et al., 2007), EIF4H (Page and Read, 2010), the Kif2a subunit of kinesin-13 (Turan et al., 2019), the POLR1C subunit of RNA polymerase III (Carter-Timofte et al., 2018), the DNA protein kinase PRKDC (Justice et al., 2021), and DDX1 (Zhang et al., 2011). Moreover, the deubiquitinase USP7 (Rodríguez et al., 2020) and the ubiquitin ligases RNF123, TRIM72, UFC1 and UBE3A as well as the proteasome might regulate capsid functionality (Huffmaster et al., 2015; Schneider et al., 2021) or their degradation (Horan et al., 2013; Sun et al., 2019). These data show that HSV-1 capsids exposing a different tegument composition recruited specific cytosolic proteins from resting or IFN-induced macrophages.

HSV-1 capsids recruit specific proteins responding to or regulating type I IFN

We next analyzed the MφIFN samples in detail as IFN induction had prevented HSV-1 infection completely. We generated cluster maps for the 32 capsid-associated proteins belonging to the GO clusters Response to type I IFN or Regulation of type I IFN production (Supplementary file 2). V capsids recruited DHX9, HSPD1, and FLOT1 as well as proteins involved in the DNA damage response like PRKDC/DNA-PK, XRCC5, and XCCR6 from both, MφR and MφIFN cytosol (Figure 4). Interestingly, V capsids bound specifically to STAT1 in MφR, but to ADAR and IFIT2 in MφIFN cytosol. D capsids were enriched for IFI16, OAS2, POLR1C, STAT2, and MxB (gene Mx2) in MφIFN but not in MφR (Figure 4, Figure 4—figure supplement 1). Particularly interesting was the discovery of MxB in these capsid-host protein complexes. MxB was significantly enriched on HSV-1 D capsids in MφIFN cytosol, and the IFN treatment had the strongest impact on the interaction of MxB with capsids. Moreover, the calculated enrichment score for MxB on capsids was very high, although the MxB levels in the input cytosol were below the detection limit (undetected, Figure 2—figure supplement 2, Supplementary file 1). MxB but not its homolog MxA restricts infections of the herpesviruses HSV-1, HSV-2, MCMV, KSHV, and MHV-68, but its mode of action has not been elucidated (Crameri et al., 2018, Liu et al., 2012, Schilling et al., 2018, Jaguva Vasudevan et al., 2018). For these reasons, we investigated the interaction of human MxB with HSV-1 capsids further.

Figure 4 with 1 supplement see all
Download asset Open asset
HSV-1 capsids associate with proteins involved in type I IFN response.

Unbiased hierarchical clustered heat map showing the log2 fold changes of IFN-induced proteins (GO type-I IFN) identified from capsids-host protein sediments from cytosol of resting Mφ, or IFN-induced MφIFN macrophages. For each protein, the fold change was calculated based on their abundance (iBAQs) in V1, V0.5, and V0.1 capsids as compared to their abundance in D capsids, using a linear scale from violet being the lowest to dark green being the highest. (*) and (**) design the proteins with an FDR corrected p-value ≤ 0.05 and ≤ 0.01, respectively.

MxB binds to capsids

We first characterized the MxB fractionation behavior during the cytosol preparation (Figure 2—figure supplement 2). As reported (Goujon et al., 2013; Melén et al., 1996), MxB was upregulated in IFN-induced MφIFN. MxB sedimented with nuclei and mitochondria as reported before (Cao et al., 2020), and also with cytoplasmic membranes. Moreover, MxB can assemble into cytosolic filaments (Alvarez et al., 2017) which might have been sedimented on their own. Both, after the addition of ATP and GTP (ATP/GTPhigh) or the hydrolase apyrase (Pilla et al., 1996; ATP/GTPlow), a significant fraction of MxB remained soluble in the cytosol.

Next, we confirmed by immunoblotting that MxB co-sedimented with HSV-1 capsids which had been incubated in cytosols from MφR or MφIFN. In line with the MS results, MxB bound better to D than to V0.1, V0.5, or V1 capsids (Figure 5A). We next probed authentic nuclear capsids, namely empty A, scaffold-filled B, or DNA-filled C capsids, as well as tegumented V1, V0.5, V0.1 or D capsids with cytosol of A549-MxB(1-715) epithelial cells expressing MxB(1-715). Nuclear A and C as well as V1 and D capsids recruited MxB efficiently, while B, V0.1 and V0.5 capsids bound less MxB (Figure 5B). MxB did not sediment by itself, and also did not associate with agarose beads used as another sedimentation control (Figure 5A and B). These data indicate that MxB binds to specific structural features on the capsid surface.

Figure 5
Download asset Open asset
Tegumentation reduces MxB binding to HSV-1 capsids.

The binding of MxB to viral V0.1, V0.5, V1, or D, or to nuclear A, B, or C capsids was analyzed after incubation in 0.2 mg/mL cytosol prepared from (A; Figure 5—source data 1) THP-1 φ stimulated or not with IFN, or (B-C; Figure 5—source data 1; Figure 5—source data 1) A549 cells stably expressing MxA, MxB(1-715) full length, the short MxB(26-715), or MxB mutants defective in GTP-hydrolysis MxB(T151A), GTP-binding and hydrolysis MxB(K131A), or dimerization MxB(M574D). Sedimented capsid-host protein complexes were then analyzed by immunoblot for VP5 (capsid), MxB, MxA, and GAPDH as a loading control. As control cytosols were sedimented without capsids (A: sed), or with uncoated agarose beads (A, B: beads). The amounts of MxA/MxB found in the capsid-host protein complexes were quantified, and normalized to their respective VP5 levels. Error bars: SEM. summarized from three experiments. One asterisk denotes p < 0.05, two asterisks indicate p < 0.01 and three asterisks represent p < 0.001 as determined by Welch’s t-tests comparisons.

Figure 5—source data 1

Cytosolic MxB from THP-1 cells co-sediments with capsids in Figure 5A.

Tegument shields MxB binding sites on HSV-1 capsids in Figure 5B. MxB requires GTP binding, but no NTE, GTP hydrolysis or dimerization to bind capsids in Figure 5C.

https://cdn.elifesciences.org/articles/76804/elife-76804-fig5-data1-v2.zip

In cells, MxB-mediated restriction of herpesvirus replication depends on its N-terminal 25 amino acid residues (NTE), its GTPase activity, and its capacity to form dimers (Crameri et al., 2018; Schilling et al., 2018; Jaguva Vasudevan et al., 2018). We incubated capsids with cytosols containing MxA, MxB(1-715), MxB(26-715) (Melén et al., 1996; Melén and Julkunen, 1997), MxB(K131A) with reduced GTP binding, MxB(T151A) lacking the GTPase activity, or MxB(M574D) unable to dimerize (Alvarez et al., 2017; Fribourgh et al., 2014; King et al., 2004; Schilling et al., 2018). In contrast to MxA, MxB(1-175), MxB(26-715), and MxB(M574D) co-sedimented with capsids to a similar extent. Interestingly, MxB(K131A) did not bind to capsids, while MxB(T151A) bound even stronger (Figure 5C). These data suggest that conformational changes associated with GTP binding or hydrolysis contribute to MxB interaction with HSV-1 capsids.

MxB disassembles capsids of alphaherpesviruses

Next, we tested whether MxB might affect HSV-1 capsid stability. While the previous capsid sedimentation assays were performed at ATP/GTPlow, they suggested that the GTP/GDP state of MxB might modulate its interaction with capsids. To test this experimentally, we supplemented the cytosols with 1 mM GTP, 1 mM ATP, and 7.5 mM creatine phosphate to maintain high ATP/GTP levels [ATP/GTPhigh]. We resuspended sedimented capsid-host protein complexes and applied them onto EM grids (Figure 2—figure supplement 1), or we added isolated capsids directly onto EM grids and then placed them on a drop of cytosol to allow the formation of capsid-host protein complexes (Figure 6A). This direct on-grid assay required 50 times fewer capsids than the sedimentation-resuspension assay and allowed for time-course analyses. For both, we negatively contrasted the samples with uranyl acetate and analyzed them by electron microscopy.

Figure 6 with 1 supplement see all
Download asset Open asset
MxB induces disassembly of herpesviral capsids.

(A) Experimental design: Capsids were adsorbed onto hydrophilic enhanced carbon-coated EM grids for 20 min at RT. The capsids were incubated in cytosol with ATP/GTPhigh, and the incubation was stopped at different times by extensive washing. The samples were analyzed by EM after negative staining with uranyl acetate. (B–D) Capsids after incubation in cytosol derived from rested Mφ or IFN-induced MφIFN macrophages, or control or MxB(1-715) A549 expressing cells for 1 hr at 37 °C, and classified as (B) intact, (C) punched or (D) disassembled flat phenotypes. The number of capsomers per flat particle was counted, and is displayed at the bottom of each figures. (E) Nuclear VZV capsids remain intact (Ei) after incubation in the cytosol of A549 control cells, or but appear punched (Eii) or as flat shells (Eiii, Eiv) after incubation in the cytosol of A549 cells expressing MxB. Scale bar: 50 nm.

When capsids were incubated with cytosol from A549 control cells not containing MxB, we saw mostly intact capsids with an appropriate diameter of about 125 nm, and an intact icosahedral morphology characterized by pentons at the vertices and hexons on the triangular capsid faces (Figure 6B). The capsids contained genomic DNA as the uranyl acetate used for negative contrast staining had not or only partially entered the capsid lumen. But a treatment with cytosol from IFN-induced MφIFN or A549-MxB(1-715) cells dramatically impaired the capsid shell. Based on different MxB-induced morphological changes, we classified the capsid structures that we had identified by immunolabeling for capsid proteins (Figure 6—figure supplement 1) into three categories. Intact capsids (Figure 6B, Figure S6A) have an icosahedral morphology and include empty A, scaffold-filled B, and DNA-filled C capsids. Punched capsids are characterized by indentations on one or more vertices and an impaired icosahedral shape (Figure 6C, Figure 6—figure supplement 1B). Flat shells have completely lost their icosahedral shape (Figure 6D, Figure 6—figure supplement 1C). We estimated the number of capsomers on flat shells based on their area, and scored a structure with <100 capsomers as a half capsid and with ≥100 as one capsid (numbers in Figure 6D). Cytosols containing MxB(1-715) also disassembled capsids of HSV-2 (not shown) or VZV (Figure 6E) to punched capsids and flat shells. As MxB induced capsid disassembly of HSV-1, HSV-2, and VZV, these experiments suggest that MxB restricts the infection of herpesviruses by targeting their capsids.

MxB requires GTP hydrolysis and dimerization to attack herpesviral capsids

Next, we further characterized the capsid disassembly activity of MxB by quantitative electron microscopy. Cytosol from IFN-induced MφIFN disassembled more than 80% of the capsids within 1 hr while resting MφR disassembled only about 40% (Figure 7A). Cytosol derived from A549 control cells had a minor effect on capsids, while cytosol from A549-MxB(1-715) cells disassembled capsids almost as efficiently as cytosol from MφIFN. Spiking cytosol from A549 control cells with an increasing percentage of A549-MxB(1-715) cytosol led to an increasing capsid disassembly with a majority of punched capsids, at 50% or 66% MxB cytosol, while incubation in pure A549-MxB(1-715) cytosol lead to more than 95% disassembly to mostly flat shells within 1 hr of incubation (Figure 7B). We then asked whether MxB had activated other host proteins to mediate capsid disassembly, or whether it was directly responsible. We prepared cytosol from A549-MxB(1-715)-MxB(26-715) expressing both untagged MxB proteins, or from A549-MxB-FLAG expressing both MxB(1-715)-FLAG and MxB(26-715)-FLAG. Both cytosols promoted capsid disassembly (MxB; MxB-FLAG in Figure 7C). An immunodepletion with anti-FLAG antibodies removed only the FLAG-tagged MxB proteins (Figure 7—figure supplement 1), and accordingly the disassembly activity from the A549-MxB-FLAG cytosol (MxB-FLAG FT), but not from the A549-MxB(1-715)-MxB(26-715) cytosol (MxB FT) containing both untagged MxB proteins.

Figure 7 with 1 supplement see all
Download asset Open asset
MxB GTP hydrolysis and dimerization required for capsid disassembly and vDNA release of viral genomes.

HSV-1 (A–H), HSV-2 (G) or VZV capsids (G) were incubated with cytosol at ATP/GTPhigh for 1 hr or the indicated time (E,F) at 37 °C, and classified into intact, punched and flat capsids by electron microscopy (A–G), or the amount of released viral DNA was measured by qPCR (H). (A) Quantification of punched and flat D capsid shells in cytosol prepared from rested Mφ or IFN-induced MφIFN macrophages, or from control A549 (mock) or A549-MxB(1-715) cells. (B) Increasing amounts of MxB(1-715) [%] were added to control A549 cytosol, and the amounts of punched and flat capsids were quantified after incubation in these mixtures. (C) Cytosols of A549 cells expressing MxB(1-715) and Mx(25-715) or MxB(1-715)-FLAG and MxB(26-715)-FLAG were incubated with anti-FLAG antibodies coupled to magnetic beads, the flow-through fractions (FT) were harvested, capsids were treated with anti-FLAG treated or control cytosols, and the amount of punched and flat capsids were quantified. (D) Capsids were incubated in cytosols prepared from A549 cells expressing full-length (FL) MxB(1-715), MxB(26-715), MxB(K131A), MxB(T151A), or MxB(M574D) at ATP/GTPlow or ATP/GTPhigh levels. (E) Time-course of MxB-induced disassembly of capsids pre-adsorbed onto EM grids, incubated with cytosol from A549-MxB(1-715). (F) Analysis of D, V0.5, or V0.1 capsids treated with MxB(1-175) cytosol for broken (punched +flat) capsids after negative stain and EM as described for panel E. (G) Quantification of MxB cytosol disassembly of D capsids of HSV-1(17+)Lox, HSV-1(KOS), or HSV-2(333), or nuclear C capsids of VZV, after incubation in cytosol from A549-MxB(1-715) cells. (H) D capsids were incubated with different cytosols for 1 hr at 37 °C or treated with 1% SDS and 10% Tx-100 only, and the released DNA not protected by capsid shells was quantified by qPCR. Error bars: SEM from 100 capsids in three biological replicates. One symbol of *or § denotes p < 0.05, two p < 0.01, and three p < 0.001 as determined in One-way analysis of variance with a Bonferroni post-test, and comparing the relative amounts of (*) punched and (§) flat capsids, or indicating the differences with the mock-treated samples (*).

We next tested at ATP/GTPhigh the effect of various MxB mutants on HSV-1 capsid stability. While full-length MxB(1-715) induced capsid disassembly, the MxB mutants impaired in GTPase activity (T151A), GTP binding (K131A), or dimerization (M574D) as well as cytosol with MxB at ATP/GTPlow did not (Figure 7D). In contrast, the smaller MxB(26-715) protein lacking the NTE retained about 50% of the capsid disassembly activity. Furthermore, studying the stability of capsids pre-adsorbed on-grid in a time-course revealed a lag phase of about 30 min until broken capsids appeared with increasing rate (Figure 7E). The percentage of punched capsids reached a plateau at 50 min, while the amount of flat shells continued to increase (Figure 7E). Further experiments showed that MxB attacked D capsids more efficiently than tegumented V0.5 capsids, of which about 70% resisted the MxB attack (Figure 7F). In contrast, the V0.1 capsids seemed to be spared from MxB attack, since no broken capsids appeared within an 1 hr treatment. Since MxB restricts infection of several herpesviruses (Crameri et al., 2018, Liu et al., 2012, Schilling et al., 2018, Jaguva Vasudevan et al., 2018), we compared the impact of MxB on D capsids from HSV-1(17+)Lox, HSV-1(KOS), HSV-2(333), or on nuclear C capsids from VZV(rOka). Capsids of these human alphaherpesviruses were all susceptible to MxB attack (Figure 7G).

MxB attack leads to the release of viral genomes from capsids

Next, we determined how well the capsid shells protected the viral genomes against a DNA nuclease digestion. Capsids released three or two times more viral genomes in cytosols from MxB(1-715) or MxB-FLAG than from control or MxB(M574D) cells (Figure 7H). Together, these data indicate that the MxB GTPase disassembles the capsid shells and induces a release of viral DNA of several herpesviruses. Our experiments suggest that GTP binding and hydrolysis as well as dimerization contribute to MxB-mediated disassembly of alphaherpesvirus capsids. Its slow start with a lag of about 30 min indicates that the capsid attack might require some nucleating or cooperative reaction to assemble active MxB oligomers or an MxB-containing complex onto capsids.

Tegument proteins protect against MxB attack

As complete tegumentation shielded V0.1 capsids against destruction, while MxB bound to surface features exposed on V0.5, A, C and D capsids, we compared the proteomes of the V0.1, V0.5, V1, and D capsids. We calibrated the relative abundances of the 58 HSV-1 proteins detected to the normalized amounts of the major capsid protein VP5. The tegument compositions of V0.1, V0.5, and V1 capsids were similar to each other but different from D capsids (Figure 8). The bona-fide capsid proteins VP21, VP24, VP22a, VP19c, and VP23 varied little among all capsid types. However, D capsids contain a bit less capsid surface proteins; namely VP26, the capsid specific vertex components (CSVC) pUL17 and pUL25, and to some extent the portal pUL6, and less of the major tegument proteins VP22, VP13/14, VP16, VP11/12 as well as other tegument proteins with ICP0, pUL36, and pUL37 being most susceptible to the trypsin treatment. Overall, there were little differences in the relative tegument protein amounts among V0.5 and V1 capsids. In contrast, V0.1 capsids contained more tegument proteins, for example VP13/14, pUS3, and pUL16. All capsid preparations contained traces of membrane proteins and nuclear HSV-1 proteins contributing to DNA replication and packaging (Figure 8—figure supplement 1). These data further validated that a treatment with 0.5 or 1 M KCl during the detergent lysis of virions destabilized intra-tegument interactions. Furthermore, the limited trypsin digestion had reduced the capsid proteome further and increased the susceptibility to MxB attack.

Figure 8 with 1 supplement see all
Download asset Open asset
Structural and tegument characterization of V0.1, V0.5, V1, and D capsids.

The composition of HSV-1(17+)Lox derived V0.1 (red), V0.5 (blue), V1 (green), and D (gray) capsids was analyzed by quantitative mass spectrometry in four biological replica. The sum of all the peptides intensities (iBAQ, intensity-based absolute quantification) of each viral protein known to participate in the structure of the capsids was normalized to the one of VP5 and displayed in a bar plot for each viral protein.

Discussion

Cell-type-specific defence mechanisms shape the arms race between proteins restricting or promoting nuclear targeting of incoming viral capsids and viral genome release into the nucleoplasm. We have developed biochemical assays to investigate functional interactions of viral capsids with host cell structures (Radtke et al., 2014), and analyzed here HSV-1 capsid-host protein complexes assembled in cytosols from resting MφR or IFN-induced MφIFN cells. We show that the IFN-inducible MxB GTPase bound to alphaherpesviral capsids, most likely to structural features around the capsid vertices, and disassembled herpesvirus capsids in a GTP-dependent fashion, and so that they no longer shielded the viral genomes. Capsid disassembly by MxB could reduce nuclear targeting of incoming capsids and genomes, but stimulate the activation of cytosolic DNA sensors and innate immune responses.

Cytosolic IFN-induced macrophage proteins binding to HSV-1 capsids

IFN induction prevented HSV-1 infection of Mφ, and increased the cytosolic abundance of at least 12 proteins listed in the interferome database (Rusinova et al., 2013). Here, we assembled host protein-capsid complexes from HSV-1 capsids and cytosols of Mφ or MφIFN cells as they might also form in cells. While our MS analyses showed that V0.5 and V1 capsids recruited unique but also common proteins, the proteomes of V0.1 and D capsids were more distinct. These specific interactions are consistent with the notion that a treatment with 0.5 or 1 M KCl during the detergent lysis of virions destabilized intra-tegument interactions, that influenced, for example, the recruitment of dynactin, kinesin-1, and kinesin-2 from brain cytosol (Ojala et al., 2000; Radtke et al., 2010; Wolfstein et al., 2006). Moreover, these results are consistent with immunoelectron microscopy data showing that the surface of distinct V capsid types display different tegument epitopes (Radtke et al., 2010), and with cryoelectron tomography data revealing diminishing tegument densities from V0.1, V0.5, V1, capsids to C capsids (Anderson et al., 2014). Thus, the surface features of V0.1, V0.5, and V1 capsids differ as indicated by cryoelectron tomography, binding of anti-tegument antibodies, and the recruitment of distinct sets of cytosolic proteins from brain tissue (Radtke et al., 2010), or macrophages as shown here. Host proteins may bind to viral proteins in both states, when they are soluble in the cytosol or the nucleoplasm, or when they are associated with capsids. From host proteins shown here to bind to capsids, direct interactions with tegument proteins have already been reported; for example USP7 binding to ICP0 (Everett et al., 1997) or EIF4H binding to vhs (pUL41; Page and Read, 2010). Furthermore, proteins involved in intracellular trafficking or virus assembly associated particularly with tegumented V capsids. For example, importin α5 (KPNA1) might mediate capsid targeting to the nuclear pores (Döhner et al., 2018; Döhner et al., 2021), while RAB1B contributes to the envelopment of cytosolic HSV-1 capsids (Zenner et al., 2011).

MxB binding to alphaherpesviral capsids

In addition to MxB, the host-capsid complexes included other antiviral proteins which in turn might be counteracted by HSV-1 proteins. Several MφIFN proteins already know to restrict herpesviruses, for example STAT2, POLR1C, IFI16, DDX58 (RIG-I), and OAS2 (Kurt-Jones et al., 2017; Lum and Cristea, 2021; Ma et al., 2018), bound preferentially to D capsids. As it was not known how MxB might restrict herpesviral infection (Crameri et al., 2018; Schilling et al., 2018; Jaguva Vasudevan et al., 2018), we investigated its association with capsids further. B capsids are less sturdy and have not undergone the structural changes that stabilize the A and C capsids (Roos et al., 2009; Sae-Ueng et al., 2014; Snijder et al., 2017). Intriguingly, this stabilization depends on the CSVC proteins pUL17 and pUL25 (Sae-Ueng et al., 2014; Snijder et al., 2017), which are present on B, A, and C capsids (Anderson et al., 2014; Radtke et al., 2010; Snijder et al., 2017). As MxB bound to A, C and D, but not to B capsids, it might recognize surface features formed during capsid stabilization, e.g. matured CSVCs or portals, which are increasingly shielded on tegumented V1, V0.5, and V0.1 capsids.

MxA and MxB GTPases inhibit several viruses by blocking early steps of infection (Haller et al., 2015). MxB binding to HIV capsids depends on its N-terminal region (NTR) of about 90 residues and the GTPase domain (Betancor et al., 2019; Fricke et al., 2014; Smaga et al., 2019; Xie et al., 2021). Similarly, HSV-1 capsids bound MxB(1-715) and to a lesser extent MxB(26-715). But in contrast to HIV capsids (Betancor et al., 2019; Xie et al., 2021), HSV-1 capsids recruited also the GTPase deficient MxB(T151A) and the monomeric MxB(M574D). These data indicate that the interaction of MxB with HSV-1 capsids depends on the NTE of 25 residues, its GTP/GDP status, but not on its dimerization.

MxB induced disassembly of alphaherpesviral capsids

HSV-1 capsid disassembly did not require proteolysis as the cytosols contained protease inhibitors, but may be modulated by other host proteins as there was a considerable lag phase. MxB did not attack fully tegumented V0.1 capsids, while V0.5 or D capsids were more susceptible. The large tegument protein pUL36 links other tegument proteins to the capsids; it is tightly associated with pUL17 and pUL25 at the CSVCs at the pentons, and it extends toward the twofold symmetry axes connecting neighboring capsid faces (Coller et al., 2007, Liu et al., 2017, Newcomb and Brown, 1991, Roos et al., 2009, Schipke et al., 2012). Our electron microscopy data suggest that MxB attacked the fivefold symmetry axes as the punched capsids had dramatic dents on the capsid vertices. MxB might furthermore attack the portal cap, a cap of HSV1-pUL25 or its homologs in other herpesviruses, which seals the pUL6 portal after DNA packaging is completed (Liu et al., 2019, McElwee et al., 2018; Naniima et al., 2021). The high internal capsid pressure due to the negatively charged genome (Bauer et al., 2013; Roos et al., 2009) could support the MxB attack from the outside. The limited trypsin treatment might have primed the D capsids for disassembly, as they contained less pUL36, pUL17, pUL25, and pUL6 than the V capsids. However, MxB also attacked V0.5 capsids that resemble cytosolic capsids during nuclear targeting or after nuclear egress (Ojala et al., 2000; Wolfstein et al., 2006; Radtke et al., 2010; Anderson et al., 2014); just not as fast, and not as efficient. Altogether, these results suggest that increasing tegumentation protects incoming and newly assembled capsids, possibly by masking the MxB target structure, or by inhibiting its GTPase cycle.

The MxB-mediated capsid disassembly required its NTE(1-25), GTP hydrolysis, and dimerization. For the homologous MxA GTPase that limits infection of many RNA viruses (Haller et al., 2015), Gao et al., 2011 proposed a restriction mechanism that involves GTP hydrolysis and a mechano-chemical coupling within ring-like oligomers with the GTPase domains being exposed on their outer diameter (Gao et al., 2011). Similarly, MxB can also assemble into helical tubes with the NTE and the GTPase domain oriented outwards (Alvarez et al., 2017). Accordingly, MxB monomers and dimers might associate with the capsid vertices and insert between the hexons of neighboring capsid faces. A further oligomerization of MxB and/or conformational changes associated with GTP hydrolysis might then exert destabilizing forces onto the capsid shells, and ultimately push the capsid faces apart.

Does MxB induce capsid disassembly in cells?

Future studies need to investigate whether MxB also induces the disassembly of herpesviral capsids in cells. Upon docking of an incoming capsid to a NPC, the pUL25 portal cap is supposed to be displaced, the pUL6 portal to be opened, and the DNA to be ejected from the capsid into the nucleoplasm due to this intramolecular repulsion (Brandariz-Nuñez et al., 2019; Döhner et al., 2021; Ojala et al., 2000; Rode et al., 2011). In uninfected cells, there is a low amount of constitutively expressed MxB localized at the NPCs (Crameri et al., 2018; Kane et al., 2018; Melén and Julkunen, 1997), which might dislodge the portal cap and open the capsid portal on the incoming capsid to release the incoming genome into the nucleoplasm.

Crameri et al., proposed that the higher amounts of IFN-induced MxB may block cytosolic capsid transport, genome uncoating at the NPCs, and/or the release of viral genomes into the nucleoplasm, which is consistent with our biochemical data demonstrating MxB binding to HSV-1 capsids (Crameri et al., 2018). MxB-mediated disassembly of capsids that we report here would further reduce capsid targeting to the NPCs and genome release into the nucleoplasm. Accordingly, there are fewer HSV-1 capsid puncta in MxB expressing cells (Crameri et al., 2018). Consistent with our data on capsid disassembly with MxB(26-715), MxB(K131A), or MxB(M574D), restricting the infection of HSV-1, MCMV, and MHV68 also requires the NTE, GTP hydrolysis, and dimerization of MxB (Crameri et al., 2018; Schilling et al., 2018). Restriction of HIV infection depends also on MxB NTE and on MxB dimerization, while the role of its GTPase function requires further investigation (Buffone et al., 2015; Fricke et al., 2014; Goujon et al., 2014; Schulte et al., 2015; Xie et al., 2021). It will be interesting to determine whether MxB only competes for the binding of host factors required for HIV intracellular trafficking, such as microtubule motor adaptors (BICD2; FEZ-1) or nucleoporins (reviewed in Temple et al., 2020), or whether it also induces HIV capsid disassembly.

Our data together with Schilling et al., and Crameri et al., suggest that the IFN-inducible MxB restricts HSV-1, HSV-2, VZV, and possibly other herpesviruses, by promoting efficient capsid disassembly (Schilling et al., 2018; Crameri et al., 2018). We cannot exclude that a surplus of capsid- and NPC-associated MxB imposes further restrictions on intracellular transport and genome release into the nucleoplasm. However, if MxB(1-715) would disassemble viral capsids before they are oriented properly with their portal toward the NPCs, their genomes would end up in the cytosol and would not be delivered into the nucleoplasm. There are fewer incoming cytoplasmic capsids in cells expressing MxB (Crameri et al., 2018), and incoming VP5 is ubiquitinated and degraded by proteasomes in macrophages (Horan et al., 2013; Sun et al., 2019). Therefore, capsid disassembly intermediates might be degraded in cells, while we could characterize them in our biochemical cell-free assays in which proteases had been blocked.

The viral genomes exposed after MxB-induced capsid disassembly might be degraded by the DNase TREX1 (Sun et al., 2019), or stimulate the DNA sensors AIM2, cGAS, or IFI16, and the induction of antiviral host mechanisms. As an inoculation with destabilized HIV-1 capsids leads to an increased activation of the DNA sensor cGAS (Sumner et al., 2020), the IFN-induced increased MxB expression might lead to a similar outcome in cells infected with herpesviruses. Accordingly, MxB may not only restrict herpesviruses by capsid disassembly, but also increase the exposure of viral genomes to cytosolic DNA sensors, which in turn would induce an IFN response, inflammation as well as innate and adaptive immune responses. Thus, MxB could be the long sought-after capsid sensor that destroys the sturdy herpesvirus capsids, and possibly HIV cores and other viral capsids, to promote host viral genome sensing.

Materials and methods

Cells

All cells were maintained in a humidified incubator at 37 °C with 5% CO2, passaged twice per week, and were tested negative for mycoplasma contamination. BHK-21 (ATCC CCL-10) and Vero cells (ATCC CCL-81) were cultured in MEM Eagle with 1% NEAA (Cytogen, Wetzlar, Germany) and 10% or 7.5% (v/v) FBS, respectively (Good Forte; PAN-Biotech, Aidenbach, Germany). HaCat (Boukamp et al., 1988; kind gift from Detlef Neumann, Hannover Medical School, Hannover, Germany) and hTERT RPE-1 (RPE; CRL-2302, Clontech) were cultured in DMEM Gibco (Invitrogen) with 7.5% or 10% (v/v) FBS, respectively (Capricorn Scientific, Ebsdorfergrund, Germany). THP-1 cells (ATCC TIB-202; kind gift from Walther Mothes, Yale University, New Haven, USA) were cultured in RPMI Medium 1640 (Thermo Fisher Scientific, Waltham, Massachusetts, United States) with 10% FBS (Thermo Fisher Scientific, Waltham, Massachusetts, United States). THP-1 were stimulated with 100 nM phorbol 12-myristate 13-acetate (PMA; Sigma-Aldrich, Germany) for 48 hr and used immediately (Mφ) or after 3 days of rest (MφR). The cells were cultured with 1000 U/mL human type I IFN-α2a (MφIFN; R&D Systems, Minneapolis, Minnesota, USA) or left untreated for 16 hr.

A549-derived cells (ATCC CCL-185) were cultured in DMEM with 10% FCS. In addition to A549 control cells, we used A549 cell lines stably expressing MxB(1-715), MxB(1–715/K131A), MxB(1–715/T151A), MxB(1–715/M574D), MxB(26-715), or MxA(1-662) upon transduction with the respective pLVX vectors with an engineered Kozak sequence to favor expression of the MxB(1-715) over the MxB(26-715) proteins (Schilling et al., 2018). Furthermore, we generated A549-MxBFLAG cells expressing MxB(1-715)FLAG and MxB(26-715)FLAG, both tagged with the FLAG epitope (GACTACAAAGACGATGACGACAAG) at the C-terminus of MxB (GenBAnk NM_002463), and A549-MxB(1-715)-MxB(26-715) cells expressing untagged MxB(1-715) and MxB(26-715) using the pLKOD-Ires-Puro vector (Clontech Takara Bio, Mountain View, United States). MeWo cells (kind gift from Graham Ogg; University of Oxford, Oxford, UK) were cultured in MEM with 10% FCS, NEAA, and 1 mM sodium pyruvate. None of the cells used in this study were identified in the list of commonly misidentified cell lines (International Cell Line Authentication Committee; https://iclac.org).

Viruses

Virus stocks of HSV-1(17+)Lox (Sandbaumhüter et al., 2013), HSV-1 strain KOS (Warner et al., 1998; kind gift from Pat Spear, Northwestern Medical School, Chicago, USA), and HSV-2 strain 333 (Warner et al., 1998; kind gift from Helena Browne, Cambridge University, Cambridge, UK) were prepared as reported before (Döhner et al., 2006, Grosche et al., 2019). Extracellular particles were harvested from the supernatant of BHK-21 cells infected with 3–4 x 104 PFU/mL (MOI of 0.01 PFU/cell) for 2–3 days until the cells had detached from the culture flasks, and plaque-titrated on Vero cells. VZV rOka (kind gift from Jeffrey Cohen, NIH, Bethesda, US) was maintained in infected MeWo cells (Cohen and Seidel, 1993; Hertzog et al., 2020). After 2–4 days, the VZV-infected cells as indicated by cytopathic effects were harvested, mixed with naive MeWo cells at a ratio of 1:4 to 1:8 for continued culture. Aliquots of frozen infected cells were used to inoculate cultures used for capsid preparation.

HSV-1 infection

Request a detailed protocol

THP-1 were seeded at 2.5 × 105 cells per six-well, treated with 100 nM PMA (Sigma-Aldrich, Germany) for 48 hr, and used immediately (Mφ) or after 3 days of rest (MφR). The cells were then induced with 1000 U/mL of IFN-α (MφIFN) or left untreated for 16 hr. On the next day, they were inoculated with HSV-1(17+)Lox at 2.5 × 106, 2.5 × 107, or 5 × 107 PFU/mL (MOI of 5, 50, or 100 respectively) in CO2-independent medium (Gibco Life Technologies) supplemented with 0.1% (w/v) cell culture grade fatty-acid-free bovine serum albumin (BSA; PAA Laboratories GmbH) for 30 min, and then shifted to regular culture medium at 37 °C and 5% CO2. At the indicated times, the cells and the corresponding media were harvested separately and snap-frozen in liquid nitrogen. These samples as well as and HSV-1 and HSV-2 inocula were titrated on Vero cells (Döhner et al., 2006, Grosche et al., 2019).

Preparation of V0.1, V0.5, and V1 and D capsids

Request a detailed protocol

Extracellular HSV-1 or HSV-2 particles were harvested by sedimentation at 12,000 rpm for 90 min at 4 °C (Type 19 rotor, Beckman-Coulter) from the medium of BHK-21 cells (40 × 175 cm² flasks; 2–2.5 x 107 cells/flask) infected with 0.01 PFU/cell (2–6.7 x 104 PFU/mL) for 2.5 days. The resulting medium pellets (MP) were resuspended in 2 mL of MKT buffer (20 mM MES, 30 mM Tris-HCl, 100 mM KCl, pH 7.4), treated with 0.5 mg/mL trypsin (Sigma-Aldrich, Germany) at 37 °C for 1 hr which was then inactivated with 5 mg/mL trypsin inhibitor from soybean (SBTI; Fluka, Switzerland) for 10 min on ice (Ojala et al., 2000; Radtke et al., 2010; Radtke et al., 2014; Wolfstein et al., 2006. These samples were then mixed with an equal volume of 2-fold lysis buffer (2% TX-100, 20 mM MES, 30 mM Tris, pH 7.4, 20 mM DTT, 1 x protease inhibitor cocktail [PIs, Roche cOmplete] with 0.2 M, 1 M or 2 M KCl; Radtke et al., 2014). The samples were layered on top of 20% (w/v) sucrose cushions in 20 mM MES, 30 mM Tris, pH 7.4 with 10 mM DTT, PIs with the respective KCl concentration, and sedimented at 110,000 g for 20 min at 4 °C (TLA-120.2 rotor, Beckman-Coulter). The supernatants and the cushions containing solubilized viral envelope and tegument proteins were carefully removed. The pellets were resuspended in BRB80 (80 mM PIPES, pH 6.8, 12 mM MgCl2, 1 mM EGTA) with 10 mM DTT, PIs, 0.1 U/mL protease-free DNase I (Promega, USA), and 100 mg/mL protease-free RNase (Roth GmbH, Germany) for 1 hr at 37 °C and then overnight at 4 °C. The capsids were sedimented at 110,000 g for 15 min at 4 °C (TLA-120.2) and resuspended in capsid binding buffer (CBB: 5% [w/v] sucrose, 20 mM HEPES-KOH, pH 7.3, 80 mM K-acetate, 1 mM EGTA, 2 mM Mg-acetate, 10 mM DTT and PIs) by ultrasound tip sonication at 40 W for about 5 × 5 s on ice. Furthermore, we treated V0.1 capsids for 40 min at 37 °C with 10 µg/mL trypsin in CBB lacking PIs to generate D capsids by limited digestion. After the addition of 5 mg/mL SBTI for 10 min on ice to block the trypsin activity, the D capsids were sedimented at 110,000 x g and 4 °C for 15 min (TLA-120.2), and resuspended in CBB with PIs.

Preparation of nuclear A, B, and C capsids

Request a detailed protocol

HSV-1 nuclear capsids were prepared from 40 × 175 cm² flasks with BHK-21 cells infected with 0.01 PFU/cell (3–4 x 104 PFU/mL) for about 2.5 days (Anderson et al., 2014; Radtke et al., 2010; Radtke et al., 2014; Snijder et al., 2017; Wolfstein et al., 2006). VZV nuclear capsids were harvested from infected MeWo cells cultured in 5–10 x 175 cm2 flasks at maximum syncytia formation but before cell lysis. The cells were harvested, resuspended in MKT buffer (20 mM MES, 30 mM Tris, pH 7.4, 100 mM KCl), snap-frozen, and stored at –80 °C. Nuclear A, B, and C capsids were separated by sedimentation at 50,000 x g and 4 °C for 80 min (SW40Ti, Beckman Coulter) on linear 20% to 50% sucrose gradients in TKE buffer 20 mM Tris, pH 7.5, 500 mM KCl, 1 mM EDTA; diluted in three volumes of TKE supplemented with 2 mM DTT and PIs (Roche cOmplete). The capsids were sedimented in BSA-coated centrifuge tubes at 110,000 g at 4 °C for 20 min (TLA-120.2), resuspended in BRB80 buffer supplemented with 100 mg/mL RNase (Roth, Germany), 0.1 U/mL DNase I (M6101, Promega, USA), 10 mM DTT, and PIs, sedimented again, and resuspended in CBB with PIs.

Calibration of capsid concentration

Request a detailed protocol

To calibrate the amount of capsid equivalents (CAPeq) among different experiments, we compared all capsid preparations used in this study with a calibration curve generated from the same starting preparation. The capsids were suspended in sample buffer (1% [w/v] SDS, 50 mM Tris-HCl, pH 6.8, 1% [v/v] β-mercaptoethanol, 5% [v/v] glycerol, PIs [Roche cOmplete]), and adsorbed to nitrocellulose membranes (BioTrace, Pall Laboratory) using a 48-slot suction device (Bio-DOT-SF, Bio-Rad, Hercules, California, USA). The membranes were probed with a polyclonal rabbit serum raised against purified HSV-1 nuclear capsids (SY4563; Supplementary file 4; Döhner et al., 2018) followed by secondary antibodies conjugated to fluorescent infrared dyes (donkey-anti-rabbit IgG-IRDye1 800CW; Supplementary file 3), and documented with an Infrared Imaging System (Odyssey, Image Studio Lite Quantification Software, LI-COR Biosciences, Lincoln, Nebraska, USA). MPs harvested from one 175 cm² flasks of BHK-21 cells infected with HSV-1 contained about 0.5–1 x 109 PFU/mL, and 0.75–1.5 x 109 CAPeq/mL. A nuclear HSV-1 capsid fraction prepared from one 175 cm² flask contained about 0.5–1 x 107 CAPeq of A capsids, 1–2 x 107 CAPeq of B capsids, and 0.5–0.75 x 107 CAPeq of C capsids, and a nuclear VZV fraction from one 175 cm² flasks of MeWo cells 2–4 x 105 CAPeq of A capsids, 0.5–1 x 106 CAPeq of B capsids, and 0.8–1.6 x 107 CAPeq of C capsids. Capsid-host protein complexes were assembled in-solution using 7.5 × 108 CAPeq/condition for MS and immunoblot experiments, and for the on-grid electron microscopy assay 2 × 107 CAPeq/condition were used.

Preparation of cytosol

Request a detailed protocol

Cytosolic extracts were prepared as described before (Radtke et al., 2010; Radtke et al., 2014), dialyzed (7 K MW cut-off cassettes; Slide-A-Lyzer, Thermo Scientific), snap-frozen and stored at –80 °C. Prior to their use, the cytosols were supplemented with 1 mM ATP, 1 mM GTP, 7 mM creatine phosphate, 5 mM DTT, and PIs (Roche cOmplete), and centrifuged at 130,000 g for 30 min at 4 °C (TLA-120.2). We added nocodazole to 25 µM to the cytosols, and left them either untreated (ATP/GTPhigh) or supplemented them with 10 U/mL apyrase (Sigma; ATP/GTPlow) for 15 min at RT.

Assembly of capsid-host protein complexes in-solution

Request a detailed protocol

Capsids were resuspended in CBB and cytosol at a protein concentration of 0.2 mg/mL in an assay volume of 60 µL per sample on a rotating platform at 800 rpm for 1 hr at 37 °C (c.f. Figure 2—figure supplement 1). The capsid-host protein complexes were sedimented through a 30% sucrose cushion at 110,000 g for 20 min at 4 °C (TLA-100, Beckman-Coulter), resuspended in CBB by ultrasound tip sonication at 40 W for about 5 × 5 s on ice, and analyzed by mass spectrometry, immunoblot, or electron microscopy (Radtke et al., 2014).

SDS-PAGE and immunoblot

Request a detailed protocol

The samples were lysed in Laemmli buffer (1% [w/v] SDS, 50 mM Tris-HCl, pH 6.8, 1% [v/v] β-mercaptoethanol, 5% [v/v] glycerol, bromophenol blue, PIs [Roche cOmplete]). The proteins were separated on linear 7.5% to 12% or 10% to 15% SDS-PAGE, transferred to methanol-activated PVDF membranes, probed with rabbit or murine primary antibodies (Supplementary file 3) and secondary antibodies conjugated to fluorescent infrared dyes (anti-rabbit IgG-IRDye1 800CW; anti-mouse IgG-IRDye1 680RD; Supplementary file 3) and documented with an Infrared Imaging System (Odyssey, Image Studio Lite Quantification Software, LI-COR Biosciences, Lincoln, Nebraska, USA).

Mass spectrometry sample preparation and measurement

Request a detailed protocol

Capsid-host protein complexes were analyzed by liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) in four independent biological replicates. The samples were resuspended in hot Laemmli buffer and separated in NuPAGE 4% to 12% Bis-Tris protein gels (Invitrogen) before in-gel digestion. Briefly, proteins were fixed and stained by Coomassie solution (0.4% G250, 30% methanol, 10% acetic acid). Sample lanes were excised, destained (50% ethanol, 25 mM ammonium bi-carbonate), dehydrated with 100% ethanol and dried using a SpeedVac centrifuge (Eppendorf, Concentrator plus). Gel pieces were rehydrated in trypsin solution (1/50 [w/w] trypsin/protein) overnight at 37 °C. Tryptic peptides were extracted in extraction buffer (3% trifluoroacetic acid, 30% acetonitrile), dried using a SpeedVac centrifuge, resuspended in 2 M Tris-HCl buffer before reduction and alkylation using 10 mM Tris(2-carboxyethyl)phosphine, 40 mM 2-Chloroacetamide in 25 mM Tris-HCl pH 8.5. The peptides were purified, concentrated on StageTips with three C18 Empore filter discs (3 M), separated on a liquid chromatography instrument, and analyzed by mass spectrometry (EASY- nLC 1200 system on an LTQ-Orbitrap XL; Thermo Fisher Scientific) as described before (Hubel et al., 2019). Peptides were loaded on a 20 cm reverse-phase analytical column (75 μm column diameter; ReproSil-Pur C18-AQ 1.9 μm resin; Dr. Maisch) and separated using a 120 min acetonitrile gradient. The mass spectrometer was operated in Data-Dependent Analysis mode (DDA, XCalibur software v.3.0, Thermo Fisher).

Mass-spectrometry data analysis

Request a detailed protocol

Raw files were processed with MaxQuant using iBAQ quantification and Match Between Runs option, and the protein groups were filtered with Perseus for reverse identification, modification site only identification, and MaxQuant contaminant list (https://maxquant.net/maxquant/, v1.6.2.10; https://maxquant.net/perseus/, v1.6.5.0; Cox and Mann, 2008; Tyanova et al., 2016a; Tyanova et al., 2016b). The iBAQ intensities were normalized across all samples to the overall median intensity of the HSV-1 capsid protein VP5. Cytosol and beads incubated with cytosol samples were normalized to all proteins detected in at least three replicates in each condition. Significant differences between given conditions were determined by a two-sided Welch t-test on protein groups present in three replicates of at least one condition, followed by permutation-based FDR statistics (250 permutations), using an absolute log2 difference cut-off of 1 and an FDR cut-off of 0.05. To characterize the IFN induction, we annotated proteins reported as being induced by IFN type-I as ISGs proteins (InterferomeDB, > 2 x change; http://www.interferome.org/interferome/home.jspx; Rusinova et al., 2013). We used the Fisher’s exact test against ISGs proteins as well as all Gene Ontology terms (GO; Ashburner et al., 2000; Consortium, 2021; http://geneontology.org/) for enrichment analysis of proteins upregulated in IFN-induced MφIFN cytosol over MφR cytosol (log2 difference ≥1.5; permutation-based FDR ≤ 0.05). The data were summarized in volcano or bar plots (GraphPad Prism v5.0, https://www.graphpad.com/; Perseus v1.6.5.0; Tyanova et al., 2016b).

Interaction network assembly

Request a detailed protocol

We focused our analysis on proteins that showed specific differences from one capsid preparation to the other, within the same cytosol preparation, and considered host proteins with an enrichment higher than 1.5 log2 fold changes and a permutation-based FDR ≤ 0.01 as specifically enriched. To visualize enrichment among different capsid-host protein complexes, we generated integrative networks using Cytoscape (http://www.cytoscape.org/; v3.7.2; Shannon et al., 2003) and STRING (confidence score: 0.7; Szklarczyk et al., 2019). STRING uses a combination of databases on co-expression, conserved occurrences, GO terms and Kyoto Encyclopedia of Genes and Genomes (KEGG; https://www.genome.jp/kegg/; Kanehisa and Goto, 2000; Kanehisa, 2019; Kanehisa et al., 2021). To assemble pathway enrichments, we used DAVID, a Database for Annotation, Visualization and Integrated Discovery (https://david.ncifcrf.gov/home.jsp; v6.8; Huang et al., 2009a; Huang et al., 2009b) and the Cytoscape plug-ins ClueGO and CluePedia (http://apps.cytoscape.org/apps/cluego, v2.5.7; http://apps.cytoscape.org/apps/cluepedia, v1.5.7; Bindea et al., 2009; Bindea et al., 2013).

Electron microscopy

Request a detailed protocol

Capsid-host protein complexes were assembled at ATP/GTPhigh in solution, harvested by ultracentrifugation, resuspended in CBB, and adsorbed onto enhanced hydrophilicity-400 mesh formvar- and carbon-coated copper grids (Stork Veco, The Netherlands; Radtke et al., 2010; Roos et al., 2009). Moreover, capsids at a concentration of 1 × 107 CAPeq/mL were adsorbed directly for 20 min at RT onto the grids. The grids were incubated on a 10 µL drop of cytosol with a protein concentration of 0.2 mg/mL and ATP/GTPhigh in a humid chamber for 1 hr at 37 °C. The samples were left untreated or labeled with anti-VP5 (pAb NC-1) and protein-A gold (10 nm diameter; Cell Microscopy Centre, Utrecht School of Medicine, The Netherlands). For both protocols, the grids were washed with PBS and ddH2O, contrasted with 2% uranyl acetate at pH 4.4, air dried, and analyzed by transmission electron microscopy (Morgani or Tecnai; FEI, Einthoven, The Netherlands). The capsid morphology was evaluated for about 100 structures/assay from about 15 randomly selected images of 2.7 µm² of three biological replicates. We classified capsomer-containing structures as punched, if they lacked one or more of their vertices but still had an icosahedral shape, and as flat shells, if they lacked the icosahedral shape but contained capsomers, and scored them as one capsid equivalent structure if they contained more than 100 capsomers.

Capsid DNA uncoating assay

Request a detailed protocol

D capsids were incubated with cytosols from A549-control, A549-MxB(1-715), A549-MxB(M574D), or A549-MxB-FLAG for 1 hr at 37 °C or treated for 5 min with 1% SDS followed by 10 min with 10% TX-100 (Ojala et al., 2000). The viral genomes released during the assay were degraded by adding 50 U/mL of benzonase for 1 hr at 37 °C, and the remaining protected DNA was purified with the DNA Blood Mini Kit (Qiagen, Hilden, Germany) and quantified by real-time PCR on a qTower3 (Analytik Jena, Jena, Germany). The SYBR Green assay was performed with the Luna Universal qPCR Master Mix (NEB, Ipswich, MA, USA) according to the manufacturer’s instructions with primers specific for HSV-1 gB (UL27 gene) (HSV1_2 SYBR fwd: 5’-gtagccgtaaaacggggaca-3’ and HSV1_2 SYBR rev: 5’-ccgacctcaagtacaacccc-3’; Engelmann et al., 2008). Standards and samples were run in triplicates and results expressed as % released viral DNA with the SDS/Tx-100 treatment normalized to 100%.

Quantification and statistical analyses

Request a detailed protocol

We performed Welch’s t-testing, Kruskal-Wallis H-testing, Friedman and one-way analyses of variance with a Dunns or Bonferroni post-testing (GraphPad Prism v5.0; https://www.graphpad.com/).

Data availability

The raw datasets produced in this study are available at PRIDE (PXD028276; http://www.ebi.ac.uk/pride). The dataset analyses and the raw bottling images are included in the Supplementary Files 1-3 and in the Source Data folder, respectively.

The following data sets were generated
    1. Serrero MC
    2. Girault V
    3. Pichlmair A
    4. Sodeik B
    (2021) PRIDE
    ID PXD028276. The interferon-inducible antiviral GTPase MxB promotes capsid disassembly and genome release of herpesviruses.
    https://www.ebi.ac.uk/pride/archive/projects/PXD028276

References

  3. Book
    1. Arvin AM
    2. Abendroth A
    (2021)
    Varicella–Zoster Virus
    In: Howley PM, Knipe DM, Cohen JL, Damania BA, editors. Fields Virology: DNA Viruses. Wolters Kluwer Health/Lippincott Williams & Wilkins. pp. 445–488.
    1. Crump CM
    (2018) Virus Assembly and Egress of HSV
    Advances in Experimental Medicine and Biology 1045:23–44.
    https://doi.org/10.1007/978-981-10-7230-7_2
  56. Book
    1. Knipe DM
    2. Heldwein EE
    3. Mohr IJ
    4. Sodroski CN
    (2021)
    Herpes Simplex Viruses: Mechanisms of Lytic and Latent Infection
    In: Howley PM, Knipe DM, Cohen JL, Damania BA, editors. Fields Virology: DNA Viruses. Wolters Kluwer Health/Lippincott Williams & Wilkins. pp. 235–296.
  98. Book
    1. Whitley R
    2. Roizman B
    (2016) Herpes Simplex Viruses
    In: Richman DD, Whitley RJ, Hayden FJ, editors. Clinical Virology, Fourth Edition. ASM Press. pp. 415–445.
    https://doi.org/10.1128/9781555819439
  99. Book
    1. Whitley R
    2. Johnston C
    (2021)
    Herpes Simplex Viruses: Pathogenesis and Clinical Insights
    In: Howley PM, Knipe DM, Cohen JL, Damania BA, editors. In Fields Virology: DNA Viruses. Wolters Kluwer Health/Lippincott Williams & Wilkins. pp. 297–323.

Decision letter

  1. Adam P Geballe
    Reviewing Editor; Fred Hutchinson Cancer Research Center, United States
  2. Päivi M Ojala
    Senior Editor; University of Helsinki, Finland

Our editorial process produces two outputs: i) public reviews designed to be posted alongside the preprint for the benefit of readers; ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "The interferon-inducible GTPase MxB promotes capsid disassembly and genome release of herpesviruses" for consideration by eLife. Your article has been reviewed by 2 peer reviewers, one of whom is a member of our Board of Reviewing Editor, and the evaluation has been overseen by Päivi Ojala as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission. Both reviewers were impressed with the innovative experimental design, the extensive amount of data presented, and most importantly the new insights emerging from your experiments showing how MxB might function to inhibit HSV-1 and how HSV-1 might evade the MxB defense. The reviewers felt that parts of the manuscript need important clarifications but they do not feel additional experiments are necessary. Please address the concerns listed in the specific comments below.

Essential revisions:

1) The reviewers both ended up feeling that the initial studies cataloging the large amount of mass spectrometry data in the various samples did not logically lead to the very interesting in-depth focus of MxB. Can you explain more clearly your rationale for focusing on MxB from among all the factors identified?

2) The presentation and interpretation of the data in figure 8 should be clarified to address the concerns noted by reviewer 2.

Reviewer #1 (Recommendations for the authors):

I just have a few suggestions on possible follow-up studies with regards to the mechanism of MxB's capsid destruction.

1. Would SDS-PAGE or Westerns with antibodies to capsid structural proteins at different stages during the MxB induced capsid degradation reveal if any of the capsid structural proteins are altered?

2. In the cell-free assay would any of the MxB extracts degrade another DNA virus capsid such as adenovirus, polyomavirus, or a phage capsid? This would be another way of showing that specific capsid surface proteins are required for MxB capsid interaction.

3. Figure S6. Would an MxB antibody label capsids in these immuno-gold EM studies?

Reviewer #2 (Recommendations for the authors):

The link between the first, proteomics part of the paper, and the second part, (MxB mechanism) seems tenuous. Unless I am missing it, I don't see MxB appearing in Figure S2C, Figure S3, Figure 3, or the supplementary tables. In Figure 2, MxB appears only to be slightly (insignificantly?) preferentially associated with D capsids. Can the author make a stronger rationale for linking the two parts of the paper? The proteomics data are potentially useful for future studies, but since the authors have not independently validated the findings (except by repetition), the presentation of them could be more succinct and possibly even more of the data (e.g., Figure 2) could be presented only as supplemental figures.

Line 265-268, Figure 7F. Did the authors test V[1.0] capsids? That analysis would be useful, especially since the results shown in Figure 7 and Figure 8 are not exactly consistent with each other and with the model they propose. Figure 8 shows that the viral proteins associated with the various V capsid preparations are quite similar to each other. But in Figure 7, it seems that V[0.5] capsids are more sensitive than V[0.1] capsids to damage by MxB.

Lines 292-293. The text describing the similarities and differences in proteins binding to the various capsid preparations does not fit well with the data in Figure 8. For example, V[0.1] capsids do not really contain substantially more pUL41, especially compared to V[1] capsids; V[0.1] capsids contain similar amounts of pUL40 as V[1] capsid, but much more than V[0.5] capsids; etc. Perhaps the author could focus on differences that are statistically different. These data do not identify any major proteins that show a gradient from low in D capsids to higher in [0.5] to highest in [0.1] capsids that would fit with the gradient of effects shown in Figure 7F. The discussion of these results should reflect the limitation of these results.

https://doi.org/10.7554/eLife.76804.sa1

Author response

Essential revisions:

1) The reviewers both ended up feeling that the initial studies cataloging the large amount of mass spectrometry data in the various samples did not logically lead to the very interesting in-depth focus of MxB. Can you explain more clearly your rationale for focusing on MxB from among all the factors identified?

We modified the text accordingly (c.f. response to reviewer 2, comment #1).

2) The presentation and interpretation of the data in figure 8 should be clarified to address the concerns noted by reviewer 2.

We modified the text accordingly (c.f. response to reviewer 2, comment #3).

Reviewer #1 (Recommendations for the authors):

I just have a few suggestions on possible follow-up studies with regards to the mechanism of MxB's capsid destruction.

1. Would SDS-PAGE or Westerns with antibodies to capsid structural proteins at different stages during the MxB induced capsid degradation reveal if any of the capsid structural proteins are altered?

Thank you for this suggestion. We agree that immunoblot analyses of time-course experiments on MxB induced capsid disassembly might detect consecutive dissociation of vertex-associated proteins such as the tegument proteins pUL36 and pUL37 as well as the capsid proteins pUL17 and pUL25. Although such experiments would be very informative, we agree that the required workload would exceed this revision.

Nevertheless, some considerations from our side. We are quite satisfied that we could transfer our experiments from analyzing sedimented capsids and capsid disassembly intermediates to the on-grid analyses (c.f. Figure 6A; Figure 7E, F, G).

To detect several of the structural HSV-1 proteins in one lane in an immunoblot, we required capsids derived from about 1 to 2 x 175 cm2 cell culture flasks, for other HSV-1 proteins we would need even more. This is about 10 times as many capsids as for3 replicates in the on-grid EM assay (c.f. manuscript, lines 222-224). Thus, such immunoblot experiments are quite demanding in terms of working hours and consumables. Nevertheless, we plan to try the suggested experiments, either by immunoelectron microscopy or by SDS-PAGE.

Moreover, we treasure the “open-view” of the EM analysis. After we had discovered MxB-induced capsid disassembly, we became concerned whether disassembly intermediates might be less likely to sediment than intact capsids, and whether disassembly intermediates might be further damaged by ultracentrifugation and more importantly by the subsequent re-suspension.

2. In the cell-free assay would any of the MxB extracts degrade another DNA virus capsid such as adenovirus, polyomavirus, or a phage capsid? This would be another way of showing that specific capsid surface proteins are required for MxB capsid interaction.

Previous work from Georg Kochs et al. (Schilling et al. 2018) as well as Jovan Pavlovic et al. (Crameri et al. 2018) shows that MxB fails to restrict several RNA viruses, adenovirus Ad5, or vaccinia virus indicating that MxB shows antiviral activity against specific viruses.

Accordingly, we could not detect MxB binding to Ad5 capsids in our co-sedimentation assays (unpublished observations). In preliminary experiments with papillomavirus-like particles, we detected an MxB induced exposure of an internal epitope by immunoelectron microscopy (unpublished observations). Furthermore, we are expanding our studies to HCMV, MCMV and MHV68 capsids. While these experiments will be informative, we think that they require more work, and that they are beyond the scope of the present manuscript.

3. Figure S6. Would an MxB antibody label capsids in these immuno-gold EM studies?

Although such experiments would be informative, we think that they require more work, and that they are beyond the scope of the present manuscript.

So far, the commercially MxB antibodies that we tested did not label capsids treated with MxB cytosol in our immunoelectron microscopy trials. Therefore, we plan additional experiments with FLAG-tagged MxB and anti-FLAG antibodies.

Reviewer #2 (Recommendations for the authors):

The link between the first, proteomics part of the paper, and the second part, (MxB mechanism) seems tenuous. Can the author make a stronger rationale for linking the two parts of the paper?

In this study, we designed our cell-free capsid-host interaction assays (reviewed in Radtke et al. 2014) to identify IFN inducible proteins potentially restricting HSV-1 capsids, and therefore focused on proteins of the GO terms “Response to type I IFN” and “Regulation of type I IFN production” (c.f. Figure 4) binding to capsids.

To answer this comment, we have expanded the explanation for our rationale to focus on MxB (previous lines 179 – 183; in the revised manuscript lines 181 – 189):

“Particularly interesting was the discovery of MxB in these capsid-host protein complexes. MxB was significantly enriched on HSV-1 D capsids in MφIFN cytosol, and the IFN treatment had the strongest impact on the interaction of MxB with capsids. Moreover, the calculated enrichment score for MxB on capsids was very high although the MxB levels in the input cytosol were below the detection limit (undetected, Figure S2, Table S1). MxB but not its homolog MxA restricts infections of the herpesviruses HSV-1, HSV-2, MCMV, KSHV, and MHV-68, but its mode of action has not been elucidated (Crameri et al. 2018; Liu et al. 2012; Schilling et al. 2018; Vasudevan et al. 2018). For these reasons, we investigated the interaction of human MxB with HSV-1 capsids further.”

Unless I am missing it, I don't see MxB appearing in Figure S2C, Figure S3, Figure 3, or the supplementary tables. In Figure 2, MxB appears only to be slightly (insignificantly?) preferentially associated with D capsids.

In addition to line 85 of the introduction, we added to line 180: “…. were enriched for IFI16, OAS2, POLR1C, STAT2 and MxB (gene Mx2) in MφIFN but not in MφR (Figure 4, Figure S5).” to clearly indicate its protein name and its gene name.

Figure S2C reports on the changes in the protein composition of the cytosolic extracts upon IFN treatment. In these samples, the MxB induction was so low compared to the other IFN inducible proteins that it was not detected by mass spectrometry. In Figure S3B, MxB is listed under its gene name MX2 in bold in the 4th column as one protein of the 4th cluster; it is number 8 from the bottom of the top block of the 4th column. MxB is one of the many ISG proteins that we detected on the capsids, and that we indicated with the # symbol in this graph. In the supplementary tables, MxB is listed under its gene name MX2.

In Figure 2, we indicated in red those ISG proteins that were significantly enriched (permutation based FDR ≤ 0.05; log2 difference ≥ 1.5) on either capsid type in these binary comparisons, and MxB enriched on D capsids in comparison to the 3 other viral capsid types.

In our functional enrichment analyses we used a conservative significance cutoff of an enrichment higher than 1.5 log2 (2.83-fold change) and a permutation-based FDR ≤ 0.01 for the GO and the STRING analyses (Figure S4 and Figure 3). MxB did not reach these cut-offs and is therefore not represented in these graphs.

The proteomics data are potentially useful for future studies, but since the authors have not independently validated the findings (except by repetition), the presentation of them could be more succinct and possibly even more of the data (e.g., Figure 2) could be presented only as supplemental figures.

It is true that so far we have validated only the MxB interaction with capsids in depth. But Figure 2 provides a comparative overview on the complexity of the macrophage host protein-capsid complexes. We consider this dataset valuable and would prefer to present it without the need to consult supplementary figures and table.

Figure 2 shows that different protein sets bound to different viral capsids (0.1, 0.5 and 1 M) in comparison to the D capsids (compare A and B versus C as well as D and E versus F). As the tegumentation of the 0.1 M capsids was more complex, the set of associated host proteins was also more complex. Moreover, the capsids associate with other proteins in the IFN cytosol than in the control cytosol (compare left panels with respective right panels).

Therefore, we would like to keep Figure 2 among the main figures; also in light of comment 2.

Line 265-268, Figure 7F. Did the authors test V[1.0] capsids? That analysis would be useful, especially since the results shown in Figure 7 and Figure 8 are not exactly consistent with each other and with the model they propose. Figure 8 shows that the viral proteins associated with the various V capsid preparations are quite similar to each other. But in Figure 7, it seems that V[0.5] capsids are more sensitive than V[0.1] capsids to damage by MxB.

Unfortunately, we did not compare V[1.0] capsids in a time course disassembly experiment with V[0.5], V[0.1], or D capsids. We think that V[0.5] capsids are most similar to authentic cytosolic capsids, which unfortunately we cannot isolate in sufficient amounts for biochemical experiments.

V[0.5] capsids are the best substrates to recruit microtubule motors (Radtke et al. 2010), to move along microtubules in vitro (Wolfstein et al. 2006), dock to NPCs (Ojala et al. 2010; Anderson et al. 2014), and to release their genomes (Ojala et al. 2000). Thus, although the mass spectrometry analysis revealed only a few clear changes in the protein composition of V[1.0] capsids, V[0.5] capsids, and V[0.1] capsids, the functional assays suggest that the conformation of the HSV-1 proteins exposed on the capsid surfaces differ.

Based on these other studies, we therefore hypothesize that the differences in macrophage proteins associating with V[0.1] versus V[0.5] versus V[1] capsids were due to protein denaturation rather than protein extraction or removal from the capsids by increasing the KCl concentration from 0.1 to 0.5 to 1 M.

Lines 292-293. The text describing the similarities and differences in proteins binding to the various capsid preparations does not fit well with the data in Figure 8. For example, V[0.1] capsids do not really contain substantially more pUL41, especially compared to V[1] capsids; V[0.1] capsids contain similar amounts of pUL40 as V[1] capsid, but much more than V[0.5] capsids; etc. Perhaps the author could focus on differences that are statistically different. These data do not identify any major proteins that show a gradient from low in D capsids to higher in [0.5] to highest in [0.1] capsids that would fit with the gradient of effects shown in Figure 7F. The discussion of these results should reflect the limitation of these results.

We agree with the reviewer that we detected very few overall differences in the protein composition of V[0.5] when compared to V[1] capsids. The D capsids were most different from the 3 other capsid types. Furthermore, the major differences of V[0.1] versus V[0.5/1] capsids were in pUS3, VP13/14 and pUL16.

We therefore hypothesize that the differences in macrophage proteins associating with V[0.5] versus V[1] capsids were due to protein denaturation rather than protein extraction or removal from the capsids by increasing the KCl concentration from 0.5 to 1 M.

We have therefore modified lines 290-291 of the Results section:

“In contrast, V0.1 capsids contained more tegument proteins, e.g. pUS3, VP13/14 and pUL16. All capsid preparations contained traces of membrane proteins and nuclear HSV-1 proteins contributing to DNA replication and packaging (Figure S8).”

Furthermore, we expanded lines 312-314 (now lines 309-320) of the discussion:

“While our MS analyses showed that V0.5 and V1 capsids recruited unique but also common proteins, the proteomes of V0.1 and D capsids were more distinct. These specific interactions are consistent with the notion that a treatment with 0.5 or 1 M KCl during the detergent lysis of virions destabilized intra-tegument interactions that influenced, for example, the recruitment of dynactin, kinesin-1 and kinesin-2 from brain cytosol (Ojala et al., 2000; Radtke et al., 2010; Wolfstein et al., 2006).

Moreover, these results are consistent with immunoelectron microscopy data showing that the surface of distinct V capsid types display different tegument epitopes (Radtke et al., 2010), and with cryoelectron tomography data revealing diminishing tegument densities from V0.1, V0.5, V1 capsids to C capsids (Anderson et al., 2014). Thus, the surface features of V0.1, V0.5 and V1 capsids differ as indicated by cryoelectron tomography, binding of anti-tegument antibodies, and the recruitment of distinct sets of cytosolic proteins from brain tissue (Radtke et al., 2010), or macrophages as shown here.”

https://doi.org/10.7554/eLife.76804.sa2

Article and author information

Author details

  1. Manutea C Serrero

    1. Institute of Virology, Hannover Medical School, Hannover, Germany
    2. RESIST - Cluster of Excellence, Hannover Medical School, Hannover, Germany
    Contribution
    Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing - original draft, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8221-2725

  2. Virginie Girault

    Institute of Virology, Technical University Munich, Munich, Germany
    Contribution
    Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Writing – review and editing
    Competing interests
    No competing interests declared

  3. Sebastian Weigang

    Institute of Virology, Freiburg University Medical Center, University of Freiburg, Freiburg, Germany
    Contribution
    Data curation, Methodology, Resources, Validation, Writing – review and editing
    Competing interests
    No competing interests declared

  4. Todd M Greco

    Department of Molecular Biology, Princeton University, Princeton, United States
    Contribution
    Data curation, Formal analysis, Methodology, Software, Supervision, Writing – review and editing
    Competing interests
    No competing interests declared

  5. Ana Ramos-Nascimento

    Institute of Virology, Hannover Medical School, Hannover, Germany
    Contribution
    Data curation, Formal analysis, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  6. Fenja Anderson

    Institute of Virology, Hannover Medical School, Hannover, Germany
    Contribution
    Methodology, Writing – review and editing
    Competing interests
    No competing interests declared

  7. Antonio Piras

    Institute of Virology, Technical University Munich, Munich, Germany
    Contribution
    Data curation, Formal analysis, Methodology, Software, Writing – review and editing
    Competing interests
    No competing interests declared

  8. Ana Hickford Martinez

    Institute of Virology, Hannover Medical School, Hannover, Germany
    Contribution
    Data curation, Formal analysis, Investigation, Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared

  9. Jonny Hertzog

    MRC Human Immunology Unit, MRC Weatherall Institute of Molecular Medicine, Radcliffe Department of Medicine, University of Oxford, Oxford, United Kingdom
    Contribution
    Data curation, Investigation, Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-7089-982X

  10. Anne Binz

    1. Institute of Virology, Hannover Medical School, Hannover, Germany
    2. RESIST - Cluster of Excellence, Hannover Medical School, Hannover, Germany
    3. German Center for Infection Research (DZIF), Hannover-Braunschweig Partner Site, Hannover, Germany
    Contribution
    Data curation, Investigation, Methodology, Resources
    Competing interests
    No competing interests declared

  11. Anja Pohlmann

    1. Institute of Virology, Hannover Medical School, Hannover, Germany
    2. RESIST - Cluster of Excellence, Hannover Medical School, Hannover, Germany
    3. German Center for Infection Research (DZIF), Hannover-Braunschweig Partner Site, Hannover, Germany
    Contribution
    Investigation, Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared

  12. Ute Prank

    Institute of Virology, Hannover Medical School, Hannover, Germany
    Contribution
    Methodology, Resources
    Competing interests
    No competing interests declared

  13. Jan Rehwinkel

    MRC Human Immunology Unit, MRC Weatherall Institute of Molecular Medicine, Radcliffe Department of Medicine, University of Oxford, Oxford, United Kingdom
    Contribution
    Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing – review and editing
    Competing interests
    No competing interests declared

  14. Rudolf Bauerfeind

    Research Core Unit Laser Microscopy, Hannover Medical School, Hannover, Germany
    Contribution
    Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Supervision, Writing – review and editing
    Competing interests
    No competing interests declared

  15. Ileana M Cristea

    Department of Molecular Biology, Princeton University, Princeton, United States
    Contribution
    Conceptualization, Funding acquisition, Methodology, Resources, Software, Supervision, Writing – review and editing
    Competing interests
    No competing interests declared

  16. Andreas Pichlmair

    1. Institute of Virology, Technical University Munich, Munich, Germany
    2. German Center for Infection Research (DZIF), Munich Partner site, Munich, Germany
    Contribution
    Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Software, Supervision, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0166-1367

  17. Georg Kochs

    Institute of Virology, Freiburg University Medical Center, University of Freiburg, Freiburg, Germany
    Contribution
    Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review and editing
    Competing interests
    No competing interests declared

  18. Beate Sodeik

    1. Institute of Virology, Hannover Medical School, Hannover, Germany
    2. RESIST - Cluster of Excellence, Hannover Medical School, Hannover, Germany
    3. German Center for Infection Research (DZIF), Hannover-Braunschweig Partner Site, Hannover, Germany
    Contribution
    Conceptualization, Formal analysis, Funding acquisition, Investigation, Project administration, Resources, Supervision, Writing - original draft, Writing – review and editing
    For correspondence
    sodeik.beate@mh-hannover.de
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-4650-3036

Funding

Horizon 2020 Framework Programme (H2020-EU.1.3.1)

  • Jan Rehwinkel
  • Andreas Pichlmair
  • Beate Sodeik

Medical Research Council (MC_UU_00008/8)

  • Jan Rehwinkel

National Institutes of Health (NIGMS, GM114141)

  • Ileana M Cristea

European Research Council (ERC-CoG ProDAP 817798)

  • Andreas Pichlmair

Deutsche Forschungsgemeinschaft (PI1084/3)

  • Andreas Pichlmair

Deutsche Forschungsgemeinschaft (PI1084/4)

  • Andreas Pichlmair

Deutsche Forschungsgemeinschaft (PI1084/5)

  • Andreas Pichlmair

Deutsche Forschungsgemeinschaft (TRR179/TP11)

  • Andreas Pichlmair

Deutsche Forschungsgemeinschaft (TRR237/A07)

  • Andreas Pichlmair

Deutsche Forschungsgemeinschaft (KO1579/13-1)

  • Georg Kochs

Deutsche Forschungsgemeinschaft (CRC900 C2, 158989968)

  • Beate Sodeik

Deutsche Forschungsgemeinschaft (EXC62 REBIRTH, 24102914)

  • Beate Sodeik

Deutsche Forschungsgemeinschaft (EXC2155 RESIST, 390874280)

  • Beate Sodeik

Deutsche Forschungsgemeinschaft (SO403/6, 443889136)

  • Beate Sodeik

Deutsches Zentrum für Infektionsforschung (TTU 07.826_00)

  • Beate Sodeik

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Katinka Döhner and Franziska Hüsers (Institute of Virology, Hannover Medical School) as well as Miriam Schilling (University of Oxford, UK) for many constructive discussions and feedback on the manuscript, and Jasper Götting (Institute of Virology, Hannover Medical School) for support on bioinformatics analyses. We are grateful to Ari Helenius (ETH Zürich, Switzerland), Graham Ogg (University of Oxford, UK), Gary Cohen (University of Pennsylvania, USA), Helena Browne (Cambridge University, UK), Jay Brown (University of Virginia, USA), Jeffrey Cohen (NIH, Bethesda, USA), Pat Spear (Northwestern Medical School, USA), and Roselyn Eisenberg (University of Pennsylvania, USA) for their generous donation of virus strains and invaluable antibodies.

Our research was supported by the EU 7th framework (Marie-Curie Actions ITN-EDGE; https://ec.europa.eu/research/mariecurieactions/about/innovative-training-networks_en, H2020-EU.1.3.1, #675,278 to JR, AP, and BS), the UK MRC (core funding of the Medical Research Council Human Immunology Unit, MC_UU_00008/8 to JR), the NIH (NIGMS, GM114141 to IMC), an EU ERC consolidator grant (ERC-CoG ProDAP 817798 to AP), the German Research Foundation (http://www.dfg.de/; PI1084/3, PI1084/4, PI1084/5, TRR179, and TRR237 to AP; KO1579/13 to GK; CRC900 C2 158989968, EXC62 REBIRTH 24102914, EXC2155 RESIST 390874280, SO403/6 to BS) and the Deutsches Zentrum für Infektionsforschung (DZIF) (TTU 07.826_00 to BS). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Senior Editor

  1. Päivi M Ojala, University of Helsinki, Finland

Reviewing Editor

  1. Adam P Geballe, Fred Hutchinson Cancer Research Center, United States

Publication history

  1. Received: January 5, 2022
  2. Preprint posted: January 27, 2022 (view preprint)
  3. Accepted: April 22, 2022
  4. Accepted Manuscript published: April 27, 2022 (version 1)
  5. Version of Record published: May 30, 2022 (version 2)

Copyright

© 2022, Serrero et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,112
    Page views
  • 281
    Downloads
  • 4
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Manutea C Serrero
  2. Virginie Girault
  3. Sebastian Weigang
  4. Todd M Greco
  5. Ana Ramos-Nascimento
  6. Fenja Anderson
  7. Antonio Piras
  8. Ana Hickford Martinez
  9. Jonny Hertzog
  10. Anne Binz
  11. Anja Pohlmann
  12. Ute Prank
  13. Jan Rehwinkel
  14. Rudolf Bauerfeind
  15. Ileana M Cristea
  16. Andreas Pichlmair
  17. Georg Kochs
  18. Beate Sodeik
(2022)
The interferon-inducible GTPase MxB promotes capsid disassembly and genome release of herpesviruses
eLife 11:e76804.
https://doi.org/10.7554/eLife.76804

Categories and tags

Research organism

    1. Related to

    Defense Proteins: Exposing the invader

    Avraham N Bayer, Fred L Homa
  2. Further reading

Further reading

    1. Immunology and Inflammation
    2. Microbiology and Infectious Disease

    Defense Proteins: Exposing the invader

    Avraham N Bayer, Fred L Homa
    Insight

    A restriction factor induced by interferons blocks the replication of herpesviruses by disassembling the capsid proteins surrounding their genome.

    1. Immunology and Inflammation

    Regulatory T cells suppress the formation of potent KLRK1 and IL-7R expressing effector CD8 T cells by limiting IL-2

    Oksana Tsyklauri, Tereza Chadimova ... Ondrej Stepanek
    Research Article

    Regulatory T cells (Tregs) are indispensable for maintaining self-tolerance by suppressing conventional T cells. On the other hand, Tregs promote tumor growth by inhibiting anti-cancer immunity. In this study, we identified that Tregs increase the quorum of self-reactive CD8+ T cells required for the induction of experimental autoimmune diabetes in mice. Their major suppression mechanism is limiting available IL-2, an essential T-cell cytokine. Specifically, Tregs inhibit the formation of a previously uncharacterized subset of antigen-stimulated KLRK1+ IL7R+ (KILR) CD8+ effector T cells, which are distinct from conventional effector CD8+ T cells. KILR CD8+ T cells show a superior cell killing abilities in vivo. The administration of agonistic IL-2 immunocomplexes phenocopies the absence of Tregs, i.e., it induces KILR CD8+ T cells, promotes autoimmunity, and enhances anti-tumor responses in mice. Counterparts of KILR CD8+ T cells were found in the human blood, revealing them as a potential target for immunotherapy.

    1. Immunology and Inflammation
    2. Microbiology and Infectious Disease

    A hierarchy of cell death pathways confers layered resistance to shigellosis in mice

    Justin L Roncaioli, Janet Peace Babirye ... Russell E Vance
    Research Advance Updated

    Bacteria of the genus Shigella cause shigellosis, a severe gastrointestinal disease driven by bacterial colonization of colonic intestinal epithelial cells. Vertebrates have evolved programmed cell death pathways that sense invasive enteric pathogens and eliminate their intracellular niche. Previously we reported that genetic removal of one such pathway, the NAIP–NLRC4 inflammasome, is sufficient to convert mice from resistant to susceptible to oral Shigella flexneri challenge (Mitchell et al., 2020). Here, we investigate the protective role of additional cell death pathways during oral mouse Shigella infection. We find that the Caspase-11 inflammasome, which senses Shigella LPS, restricts Shigella colonization of the intestinal epithelium in the absence of NAIP–NLRC4. However, this protection is limited when Shigella expresses OspC3, an effector that antagonizes Caspase-11 activity. TNFα, a cytokine that activates Caspase-8-dependent apoptosis, also provides potent protection from Shigella colonization of the intestinal epithelium when mice lack both NAIP–NLRC4 and Caspase-11. The combined genetic removal of Caspases-1, -11, and -8 renders mice hyper-susceptible to oral Shigella infection. Our findings uncover a layered hierarchy of cell death pathways that limit the ability of an invasive gastrointestinal pathogen to cause disease.