(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



pUL21 is a viral phosphatase adaptor that promotes herpes simplex virus replication and spread

['Tomasz H. Benedyk', 'Department Of Pathology', 'University Of Cambridge', 'Cambridge', 'United Kingdom', 'Julia Muenzner', 'Viv Connor', 'Yue Han', 'Katherine Brown', 'Kaveesha J. Wijesinghe']

Date: 2021-08

The herpes simplex virus (HSV)-1 protein pUL21 is essential for efficient virus replication and dissemination. While pUL21 has been shown to promote multiple steps of virus assembly and spread, the molecular basis of its function remained unclear. Here we identify that pUL21 is a virus-encoded adaptor of protein phosphatase 1 (PP1). pUL21 directs the dephosphorylation of cellular and virus proteins, including components of the viral nuclear egress complex, and we define a conserved non-canonical linear motif in pUL21 that is essential for PP1 recruitment. In vitro evolution experiments reveal that pUL21 antagonises the activity of the virus-encoded kinase pUS3, with growth and spread of pUL21 PP1-binding mutant viruses being restored in adapted strains where pUS3 activity is disrupted. This study shows that virus-directed phosphatase activity is essential for efficient herpesvirus assembly and spread, highlighting the fine balance between kinase and phosphatase activity required for optimal virus replication.

Herpes simplex virus (HSV)-1 is a highly prevalent human virus that causes life-long infections. While the most common symptom of HSV-1 infection is orofacial lesions (‘cold sores’), HSV-1 infection can also cause fatal encephalitis and it is a leading cause of infectious blindness. The HSV-1 genome encodes many proteins that dramatically remodel the environment of infected cells to promote virus replication and spread, including enzymes that add phosphate groups (kinases) to cellular and viral proteins in order to fine-tune their function. Here we identify that pUL21 is an HSV-1 protein that binds directly to protein phosphatase 1 (PP1), a highly abundant cellular enzyme that removes phosphate groups from proteins. We demonstrate that pUL21 stimulates the specific dephosphorylation of both cellular and viral proteins, including a component of the viral nuclear egress complex that is essential for efficient assembly of new HSV-1 particles. Furthermore, our in vitro evolution experiments demonstrate that pUL21 antagonises the activity of the HSV-1 kinase pUS3. Our work highlights the precise control that herpesviruses exert upon the protein environment within infected cells, and specifically the careful balance of kinase and phosphatase activity that HSV-1 requires for optimal replication and spread.

Funding: JED is supported by a Senior Research Fellowship from the Wellcome Trust (219447/Z/19/Z). This work was supported by a Senior Research Fellowship from the Wellcome Trust to AEF (106207/Z/14/Z), a Biotechnology and Biological Sciences Research Council (BBSRC) Research Grant to CMC (BB/M021424/1) and a Sir Henry Dale Fellowship, jointly funded by the Wellcome Trust and the Royal Society, to SCG (098406/Z/12/B). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

We used quantitative proteomics to identify the cellular interaction partners of pUL21, revealing that pUL21 binds the catalytic subunit of protein phosphatase 1 (PP1) and the ceramide transport protein CERT, a protein that is regulated by phosphorylation. We show that pUL21 binds directly to PP1 and CERT, accelerating CERT dephosphorylation both in vitro and in cells, and identify a conserved motif in pUL21 required for PP1 binding. HSV-1 mutants where the ability of pUL21 to bind PP1 is abolished have impaired growth and replication, but these mutants rapidly adapt to cell culture via compensatory mutations in the viral kinase gene pUS3. We show that pUL21 and pUS3 regulate phosphorylation of multiple proteins, including components of the viral NEC, via opposing activities. While viral and cellular kinases have been intensely studied for many decades, the specificity and regulation of phosphatases is much less well defined. Our study demonstrates that a fine balance between kinase and phosphatase activity is required for herpesvirus replication and spread.

The diverse range of pUL21 functions and putative binding partners has complicated mechanistic dissection of its function, since abolishing pUL21 expression during infection impairs all functions of this multifaceted protein. Mechanistic insights into pUL21 are further hampered by its lack of homology to any well-characterised protein. HSV-1 pUL21 is a 58 kDa cytoplasmic protein [ 21 ] and crystallographic studies showed that pUL21 comprises two well-folded domains separated by a protease-sensitive linker region that is likely to be disordered [ 24 , 25 ]. Neither domain bears significant resemblance to other structurally characterised proteins, preventing inference of function by analogy. However, it was observed that the C-terminal domain of pUL21 co-purifies with RNA following recombinant expression in bacteria, suggesting that nucleic acid binding may represent yet another function of the enigmatic pUL21 protein [ 25 ].

The tegument protein pUL21 exemplifies the multifunctional nature of tegument proteins, being required for efficient capsid nuclear egress, secondary envelopment and virus spread. pUL21 is conserved across alphaherpesvirus [ 12 ] and is known to bind the conserved protein pUL16 [ 13 , 14 ], both proteins being known to promote transport of newly-assembled genome-containing virus capsids from the nucleus to the cytoplasm for subsequent envelopment [ 15 – 17 ]. However, the requirement for pUL16 and pUL21 to promote nuclear egress varies depending on the virus strain and cell line studied [ 16 ] and they function independently in this process [ 18 ]. Specifically, it has been shown that pUL21 is required for the correct ‘smooth’ distribution of the NEC around the nuclear envelope as viruses lacking pUL21 exhibiting large NEC punctae on the nuclear rim [ 18 ]. However, the molecular mechanisms by which pUL21 promotes viral nuclear egress remains uncharacterised. In addition to promoting capsid translocation to the cytoplasm, it has been reported that the pUL21:pUL16 complex may promote wrapping of cytoplasmic virions via interactions with the viral proteins pUL11 and gE [ 19 ], although the evidence for this is equivocal [ 8 ]. pUL21 has a poorly-characterised role in promoting virus spread, which may involve its interaction with gE [ 19 ]. Deletion of pUL21 leads to rapid outgrowth of HSV-1 harbouring mutations in gK that promote syncytia formation, where infected cells fuse with neighbouring cells, and pUL21 is required for maintenance of the syncytial phenotype in strains of HSV-1 where gB is mutated to induce syncytium formation [ 20 ]. pUL21 has also been observed to promote the formation of long cellular processes, potentially by regulating microtubule polymerisation [ 21 ]. Pseudorabies virus (PrV) pUL21 has been extensively studied as mutations in the gene encoding this protein contribute to the attenuation of the vaccine strain Bartha [ 22 ]. PrV pUL21 is required for neurovirulence and a recent study identified that pUL21 binds the dynein light chain protein Roadblock-1, promoting retrograde axonal transport [ 23 ].

Herpes simplex virus (HSV)-1 is a highly-prevalent human pathogen that establishes life-long latent infection, with reactivation of virus from peripheral neurons manifesting as orofacial or genital lesions or (occasionally) life-threatening viral encephalitis [ 1 – 3 ]. Like all herpesviruses, HSV-1 dramatically remodels the intracellular environment of infected cells to promote the production and dissemination of progeny virus particles [ 4 , 5 ]. After receiving their genomic cargo in the nucleus, HSV-1 capsids traverse the nuclear envelope via sequential envelopment and de-envelopment steps catalysed primarily by the viral nuclear egress complex (NEC), comprising pUL31 and pUL34 [ 6 ]. Cytoplasmic capsids associate with glycoprotein-studded membranes via a proteinaceous layer called tegument, budding into the lumen of post-Golgi intracellular vesicles (“secondary envelopment”) before being trafficked to and released at the plasma membrane [ 7 , 8 ]. Proteins of the HSV-1 tegument layer play roles in dampening the innate immune response to infection [ 9 , 10 ] and modulating host-cell morphology [ 11 ] in addition to their structural roles in virion assembly.

Results

To identify putative cellular binding partners, C-terminally GFP-tagged pUL21 (pUL21-GFP) was immunoprecipitated following ectopic expression in stable isotope labelling of amino acids in cell culture (SILAC)-labelled human embryonic kidney (HEK293T) cells. Quantitative mass spectrometry analysis identified that the catalytic subunit of protein phosphatase 1 (PP1) and the ceramide-transport protein CERT (a.k.a. Collagen type IV alpha-3-binding protein, COL4A3BP, or Goodpasture antigen binding protein, GPBP) were strongly enriched in the pUL21-GFP immunoprecipitation compared with a GFP control (Fig 1A and S1 Data). There are three isoforms of the PP1 catalytic subunit (α, β and γ) in human cells that share very high sequence identity (>85% amino acid identity overall) [26], complicating the definitive identification of co-precipitated isoforms by mass spectrometry, but immunoblots confirmed that all three isoforms co-immunoprecipitate with pUL21-GFP (Fig 1B). GST pull-down experiments using reagents purified following recombinant expression in Escherichia coli (pUL21 and PP1γ) and HEK293F cells (CERT) confirmed that pUL21 binds directly to both PP1 and CERT (Fig 1C). Published quantitative viromics analysis [27] confirms that the abundance of CERT and all three PP1 catalytic subunit isoforms is unchanged in keratinocytes over the course of HSV-1 infection.

PPT PowerPoint slide

PowerPoint slide PNG larger image

larger image TIFF original image Download: Fig 1. HSV-1 pUL21 directly binds host proteins CERT and PP1, promoting CERT dephosphorylation. (A) SILAC-labelled HEK293T cells were transfected with plasmids expressing GFP-tagged pUL21 or GFP alone. At 24 hours post-transfection the cells were lysed, subjected to immunoprecipitation (IP) using a GFP affinity resin, captured proteins were proteolytically digested and co-precipitated proteins identified using quantitative mass spectrometry. The horizontal axis shows average fold enrichment in IP of pUL21-GFP compared to GFP across three biological replicates and the vertical axis shows significance (two-sided t-test) across the three replicates. Proteins CERT and PP1 were identified as putative binding partners (blue). (B) Unlabelled HEK293T cells were transfected with plasmids expressing GFP-tagged pUL21 or GFP alone and subjected to IP as in (A). Captured proteins were subjected to SDS-PAGE and immunoblotting using the antibodies shown. GAPDH is used as a loading control and Ponceau S (Pon S) staining of the immunoblot membrane before blocking shows all proteins transferred. (C) Purified recombinant pUL21-GST or GST alone were immobilised on GSH resin and incubated with prey proteins PP1(7–300)-H 6 or Strep-CERT. After washing, the bound complexes were eluted and visualized by SDS-PAGE (Coomassie). (D) Lysates of HaCaT cells (parental or stably expressing pUL21) were analysed by SDS-PAGE and immunoblotting using the antibodies listed. The upper strip depicts SDS-PAGE where PhosTag reagent was added, retarding the migration of phosphorylated proteins to enhance separation of CERT that is hyper- (CERTP) or hypo-phosphorylated (CERTO). (E) HaCaT cells were infected at MOI = 5 with wild-type HSV-1 or a mutant lacking pUL21 expression (ΔpUL21). Lysates were harvested at 16 hours post-infection (hpi) and subjected to SDS-PAGE plus immunoblotting as in (D). The HSV-1 major capsid protein VP5 is used as a marker of infection. (F) Schematic representation of putative pUL21 activity, recruiting specific substrates for dephosphorylation by the PP1 catalytic subunit. https://doi.org/10.1371/journal.ppat.1009824.g001

PP1 is an abundant and highly active cellular phosphatase [26]. The catalytic subunit of PP1 has low intrinsic specificity and substrates are recruited for dephosphorylation via interactions with adaptor proteins known as regulatory subunits (also known as regulatory interactors of protein phosphatase one, RIPPOs) [26,28]. CERT mediates the non-vesicular transport of the lipid ceramide from the endoplasmic reticulum (ER) to the trans-Golgi, a crucial step in the synthesis of sphingomyelin and derived sphingolipids [29]. Interestingly, the activity of CERT is known to be regulated via phosphorylation: hyperphosphorylated CERT (CERTP) adopts an inactive conformation that is not membrane associated whereas hypophosphorylated CERT (CERTO) associates with ER and trans-Golgi membranes to promote ceramide exchange [30–32]. The ability of pUL21 to bind directly to PP1 and CERT suggested that pUL21 may be a novel viral PP1 regulatory subunit, recruiting PP1 to CERT to promote its dephosphorylation and hence activation. To test this hypothesis, the relative abundance of CERTP and CERTO was measured in human keratinocyte cells (HaCaT) where pUL21 was stably expressed and in HaCaT cells infected with either wild-type HSV-1 or a mutant where expression of pUL21 had been disrupted (ΔpUL21). In both cases, we observed that the CERTO predominates when pUL21 is present (Fig 1D and 1E), consistent with the identification of pUL21 as a viral regulatory subunit that promotes CERT dephosphorylation via direct recruitment of PP1 (Fig 1F).

Most cellular PP1 interacting proteins contain a conserved “RVxF” docking motif (consensus sequence [KR][KR][VI]ψ[FW], where ψ is any amino acid except FIMYDP) plus one or more ancillary motifs [33]. pUL21 does not contain any peptide sequences that conform to this consensus. It does contain two stretches of amino acids, 12RDVVF16 and 285RELWW289, that conform to a less stringent definition of the RVxF motif ([RK]x 0–1 [VI]{P}[FW], where x 0–1 is any residue, or no residue at all, and {P} is any residue except proline) [34]. Both of these sequences are in structured regions of the N- and C-terminal domains, respectively, and the key hydrophobic residues are buried in the core of the protein (S1 Fig). As such, these residues could not bind PP1 without significant structural rearrangement. The question thus arises: how does pUL21 associate with PP1?

Before dissecting the molecular determinants of pUL21 binding to PP1 and CERT we sought to determine the conformation of purified pUL21 in solution. Previous studies had shown that the two structured domains of HSV-1 pUL21 are joined by a linker region of 64 amino acids (Fig 2A) that is presumed to be disordered due to its susceptibility to proteolytic cleavage during recombinant purification [25]. We purified full-length C-terminally hexahistidine-tagged pUL21 (pUL21-H 6 ) following recombinant expression in E. coli (Fig 2B). While the C-terminal domain of pUL21 had previously been shown to co-purify with bacterial RNA [25] we did not observe nucleic acid co-purification with the full-length protein, pUL21-H 6 having a 260:280 nm absorbance ratio of ~0.58. Size-exclusion chromatography with inline multi-angle light scattering (SEC-MALS) confirmed that the protein was predominantly monodisperse and monomeric (Fig 2C). Small-angle X-ray scattering (SAXS) analysis of purified pUL21-H 6 (Fig 2D and S1 Table) showed that the protein does not adopt a single compact globular conformation: the frequency of real-space distances within the protein (p(r) profile, Fig 2E) is highly asymmetric with an extended tail of longer distance frequencies spanning the range 10–18 nm. The slow and steady decay of the p(r) profile, with no distinct peak evident at longer distance frequencies, indicates that pUL21-H 6 lacks a fixed distance between the N- and C-terminal domains. Moreover, the dimensionless Kratky plot (Fig 2F) lacks the distinct bell-shaped peak at sRg = √3 that is typical of globular proteins [35], indicating that pUL21-H 6 is a flexible and/or elongated particle. The ensemble optimisation method (EOM) [36] was therefore employed to characterise the conformational heterogeneity of pUL21. The fitted pUL21-H 6 ensemble (Fig 2G, 2H and 2I) samples a wide distribution of states that span compact (R g = 2.5–3.7 nm, 34% volume fraction), intermediate (R g = 3.7–5.0 nm, 50% volume fraction) and elongated (R g > 5 nm, 17% volume fraction) conformations. This is consistent with the pUL21 linker region being conformationally heterogeneous, with N- and C-terminal domains moving toward and away from each other in solution like two ends of an accordion (Fig 2J).

PPT PowerPoint slide

PowerPoint slide PNG larger image

larger image TIFF original image Download: Fig 2. The C-terminal domain of pUL21 binds CERT whereas the N-terminal domain and disordered linker bind PP1. (A) Schematic representation of pUL21, with crystal structures of the N- and C-terminal domains [24,25] shown in in light and dark purple, respectively. (B) Coomassie-stained SDS-PAGE of full-length pUL21-H 6 purified following bacterial expression. (C) SEC-MALS of purified pUL21-H 6 . Weight-averaged molar mass (thick line) is shown across the SEC elution profile (normalised differential refractive index, thin lines) with the expected molar mass for monomeric pUL21-H 6 shown as a dashed line. (D) SAXS profile (blue) measured from pUL21-H 6 and the corresponding Guinier plot (sR g < 1.15; inset). The Guinier plot is linear (yellow line), as expected for an aggregate-free system. The reciprocal-space fit of the p(r) profile to the SAXS data is shown a yellow trace. χ2, fit quality; p, Correlation Map (CorMap) probability of systematic deviations between the model fit and the scattering data [93]. (E) The real-space distance distribution function, p(r), computed from the experimental SAXS profile. (F) Dimensionless Kratky plot of the SAXS data. The expected maximum of the plot for a compact protein is shown as grey dotted lines (sR g = √3, (sR g )2I(s)/I(0) = 3e-1). (G) Fit to the SAXS profile of a refined pUL21-H 6 ensemble obtained by EOM. The ensemble multiple comprises conformational states of the pUL21 N- and C-terminal domains [24,25] joined by a flexible linker. (H, I) Comparison of the frequency distributions of R g (H) and D max (I) from an initially generated random pool of structures (grey) and the refined EOM ensemble (blue and violet, respectively) indicate that, in solution, pUL21 samples a distribution of states that encompasses compact, intermediate and extended conformations. (J) Selected representative models of the pUL21-H 6 ensemble. For each, the prevalence (volume fraction, %), R g and D max is shown. (K) Schematic representation of truncated pUL21 constructs used for immunoprecipitation experiments. (L) Immunoblots following immunoprecipitation from HEK293T cells transfected with GFP-tagged full-length pUL21 or truncations thereof. Cells were lysed 24 h post-transfection and incubated with anti-GFP resin to capture protein complexes before being subjected to SDS-PAGE and immunoblotting using the antibodies shown. Pon S, Ponceau S staining of the immunoblot membrane before blocking to show all proteins transferred. https://doi.org/10.1371/journal.ppat.1009824.g002

Having validated that the N- and C-terminal domains of pUL21 are joined by a mobile linker region, and can thus be considered as independent functional units, a series of GFP-tagged truncations mutants were designed to identify pUL21 regions that bind CERT and PP1 (Fig 2K). Immunoprecipitation experiments identified that PP1 binding requires both the N-terminal domain and flexible linker (Fig 2L). However, PP1 binding is lost when this region is N-terminally GFP tagged, suggesting that obstruction of the amino terminus of pUL21 via fusion with a bulky GFP domain (approximately the same size as the pUL21 N-terminal domain) could prevent PP1 binding via steric hindrance. Similarly, the C-terminal domain of pUL21 mediates CERT binding but only when the GFP tag is separated from this domain via the flexible linker or is attached to the carboxy terminus of the protein (Fig 2L), suggesting that residues near the beginning of the C-terminal domain are required for CERT binding and that binding is blocked by fusion of GFP to the start of this domain.

Mapping the sequence conservation of pUL21 across the family Alphaherpesvirinae identifies that the N- and C-terminal domains have more highly conserved sequences than the flexible linker (Fig 3A). However, there is a short stretch of conserved amino acids in the linker region between HSV-1 residues 239–248 [25], with the consensus sequence ϕSxFVQ[V/I][K/R]xI where ϕ is a hydrophobic residue and x is any residue. As the flexible linker region is required for PP1 binding we hypothesised that these conserved residues may contribute to the interaction. Site-directed mutagenesis was used to generate pUL21-GFP constructs where the conserved hydrophobic residues F242 and V243 were substituted with charged residues glutamate and aspartic acid, respectively, or a double mutant where both residues were substituted for alanine. Immunoprecipitation following transient expression in HEK293T cells showed that these pUL21 mutants retain the ability to bind CERT but no longer bind PP1 (Fig 3B). Immunoprecipitation following ΔpUL21 HSV-1 infection of HEK293T cells transiently expressing pUL21-GFP confirmed that the pUL21FV242AA double-mutant retains the ability to bind pUL16 (Fig 3C). Furthermore, differential scanning fluorimetry (a.k.a. Thermofluor) confirmed that wild-type and FV242AA pUL21-H 6 purified following recombinant expression in E. coli have similar melting temperatures (Fig 3D). Taken together, these results confirm that substitution of conserved residues in the flexible linker specifically disrupt the ability of pUL21 to recruit PP1 but do not interfere with the overall stability of the protein or its ability to interact with other cellular or viral partner proteins.

PPT PowerPoint slide

PowerPoint slide PNG larger image

larger image TIFF original image Download: Fig 3. pUL21 recruits PP1 via a conserved motif in the linker region to accelerate CERT dephosphorylation. (A). Conservation of pUL21 across Alphaverpesvirinae. The following sequences were aligned using ClustalW and conservation calculated using Jalview (Abbreviation and Uniprot ID are shown in parentheses): HSV-1 (HSV1, P10205), HSV-2 (HSV2, G9I242), cercopithecine herpesvirus 2 (CHV2, Q5Y0T2), saimiriine herpesvirus 1 (SHV1, E2IUE9), bovine alphaherpesvirus 1 (BHV1, Q65563), equine herpesvirus 1 (EHV1, P28972), pseudorabies virus (PRV, Q04532), anatid herpesvirus 1 (AHV1, A4GRJ2), varicella-zoster virus (VZV, Q6QCT9), turkey herpesvirus (MHV1, Q9DPR5). Alignment across the linker region (residues 217–280 of HSV-1 pUL21) is shown with conserved residues highlighted. (B) HEK293T cells were transfected with plasmids expressing GFP, wild-type (WT) pUL21-GFP or pUL21-GFP with amino acid substitutions in the conserved motif. At 24 hours post-transfection the cells were lysed, subjected to immunoprecipitation using a GFP affinity resin, and captured proteins were subjected to SDS-PAGE and immunoblotting using the listed antibodies. Ponceau S (Pon S) staining of the nitrocellulose membrane before blocking is shown, confirming efficient capture of GFP-tagged proteins. (C) Plasmids expressing wild-type or mutant pUL21-GFP, or GFP alone, were transfected into HEK293T cells. At 24 hours post-transfection cells were infected with ΔpUL21 HSV-1 (MOI = 5). Cells were lysed 16 hours post-infection and subjected to immunoprecipitation, SDS-PAGE and immunoblotting as in (B). (D) Differential scanning fluorimetry of WT (purple) and FV242AA substituted (blue) pUL21-H 6 . Representative curves are shown. Melting temperatures (T m ) is mean ± standard deviation (n = 3). Inset shows Coomassie-stained SDS-PAGE of the purified protein samples. (E) In vitro dephosphorylation assays using all-purified reagents. 0.5 μM CERT was incubated with varying concentrations of GST-PP1 (two-fold serial dilution from 100–3.1 nM) in the absence or presence of 2 μM pUL21-H 6 (WT or FV242AA) for 30 min at 30°C. Proteins were resolved using SDS-PAGE where PhosTag reagent was added to enhance separation of CERT that is hyper- (CERTP) or hypo-phosphorylated (CERTO) and gels were stained with Coomassie. Images are representative of three independent experiments. (F) Quantitation of pUL21-mediated stimulation of CERT dephosphorylation, as determined by densitometry. Ratio of CERTO to total CERT (CERTO + CERTP) for three independent experiments is shown (mean ± SEM). (G) 0.5 μM phosphorylated eIF2α (eIF2αP) was subjected to in vitro dephosphorylation using varying concentrations of GST-PP1 (two-fold serial dilution from 200–6.3 nM) in the absence or presence of 2 μM pUL21-H 6 as in (E). pUL21 does not enhance PP1-mediated dephosphorylation of eIF2α. (H) HEK293T cells were transfected with GFP, pUL21-GFP, the VZV homologue of pUL21 with a C-terminal GFP tag (ORF38-GFP), or with ORF38-GFP where amino acid in the conserved motif had been substituted with alanine. Cells were lysed at 24 hours post-transfection and subjected to IP, SDS-PAGE and immunoblotting as in (B). https://doi.org/10.1371/journal.ppat.1009824.g003

In vitro dephosphorylation assays with all-purified reagents were used to directly test the hypothesis that pUL21 is a novel viral PP1 regulatory subunit, promoting dephosphorylation of CERT by directly recruiting PP1 catalytic activity. CERT purified from mammalian suspension cell culture is predominantly hyperphosphorylated (CERTP, S2 Fig). The PP1γ catalytic subunit, purified from E. coli as a GST fusion to enhance protein solubility, is capable of dephosphorylating CERT when present at high concentrations (Fig 3E and 3F), consistent with low intrinsic specificity of PP1 catalytic subunits [26]. Addition of purified wild-type pUL21-H 6 dramatically lowers the concentration of GST-PP1γ required for efficient CERTP dephosphorylation but this effect is much less pronounced for pUL21FV242AA-H 6 (Fig 3E and 3F), consistent with specific pUL21-mediated recruitment of CERTP to PP1. Furthermore, addition of pUL21 did not enhance the dephosphorylation of an irrelevant substrate, phosphorylated eukaryotic initiation factor 2α (eIF2α, Fig 3G), confirming that pUL21 promotes substrate-specific enhancement of PP1 activity rather than acting as an allosteric activator of the phosphatase.

The conservation of the hydrophobic region within the pUL21 linker, which is required for PP1 recruitment, suggested that pUL21 homologues from other alphaherpesviruses may also bind PP1. Immunoprecipitation experiments confirmed that ORF38, the pUL21 homologue from varicella-zoster virus (VZV), also binds PP1 (Fig 3H). As with pUL21, mutation of the conserved hydrophobic residues in ORF38 to alanine (FV255AA) abolished the ability to co-precipitate PP1. Interestingly, neither wild-type nor mutant ORF38 co-immunoprecipitated with CERT, suggesting that the ability to bind and promote CERT dephosphorylation is specific to HSV-1 and that pUL21 and its homologues may have additional, conserved targets of PP1-mediated dephosphorylation.

To probe the specific role of PP1 binding during infection, mutant strains of HSV-1 were generated using two-step Red recombination [37] where the conserved residues F242 and V243 in pUL21 were mutated, either individually or in combination (pUL21F242E, pUL21V243D and pUL21FV242AA). For all mutants, initial virus stocks obtained following transfection of HSV-1 encoding bacterial artificial chromosome into Vero cells (P0 viruses) had impaired replication and/or spread when compared to the wild-type virus, yielding small viral plaques when infecting Vero or HaCaT cells (Figs 4A and S3). This defect was similar in magnitude to that seen when pUL21 expression was completely abolished (ΔpUL21) and the defect was rescued when plaque assays were performed on HaCaT cells constitutively expressing pUL21 (HaCaT pUL21 cells, S3 Fig), confirming that the defect did not arise from other mutations in the HSV-1 genome during virus generation. However, when these pUL21 mutant virus stocks were propagated on Vero cells for three additional generations we observed a rapid rescue of plaque size (Fig 4A): within one generation (P1) a mix of small and large plaques can be observed and by generation three (P3) the mutant virus plaques were as large as those formed by the wild-type virus (Figs 4A and S3). Sanger sequencing of the pUL21 gene amplified by PCR from P3 virus stocks of the three mutant strains revealed that the UL21 gene had not reverted to the wild-type sequence (Fig 4B). Furthermore, analysis of CERT phosphorylation following infection of Vero cells with the P3 mutant viruses confirmed that none of the three mutants had re-acquired the ability to promote CERT dephosphorylation (Fig 4C), strongly suggesting that the ability of pUL21 in these viruses to bind PP1 and recruit it to substrates like CERT remained impaired.

PPT PowerPoint slide

PowerPoint slide PNG larger image

larger image TIFF original image Download: Fig 4. Mutant HSV-1 where PP1 binding of pUL21 is abolished adapts rapidly to cell culture via compensatory mutations in the kinase gene US3. (A). Monolayers of Vero cells were infected with 100 pfu of wild-type (WT), ΔpUS3 and ΔpUL21 HSV-1 or with pUL21 point mutant viruses, harvested immediately following transfection of the recombinant BAC into Vero cells (P0) or following amplification for one (P1) or three (P3) generations in Vero cells. Infected cells were overlaid with medium containing 0.6% carboxymethyl cellulose and incubated for 48 h before fixing and immunostaining with chromogenic detection. Relative plaque areas (pixels) were measured using Fiji. Mean plaque sizes (red bars) were compared to WT using one-way ANOVA with Dunnett’s multiple comparisons test (n = 96–113; ns, non-significant; *, P > 0.05; **, P < 0.01; ***, P < 0.001) and images used for quantitation are shown in S3 Fig. (B) Sanger sequencing of pUL21 region, amplified by PCR from P3 stocks of virus, show that the introduced point mutations have not reverted to the wild-type sequence. (C) Vero cells were mock-infected or infected at MOI = 5 with WT HSV-1, ΔpUL21 virus or with P3 stocks of pUL21 point mutants shown. Lysates were harvested at 16 hpi and subjected to SDS-PAGE plus immunoblotting using the antibodies listed. The upper strip depicts SDS-PAGE where PhosTag reagent was added, retarding the migration of hyperphosphorylated CERT (CERTP) versus the hypophosphorylated protein (CERTO). The P3 stocks of all pUL21 point mutant viruses retain pUL21 expression but all lack the ability to promote CERT dephosphorylation, indicating that PP1 binding has not been restored. (D) Prevalence of sequence variants in P3 stocks of pUL21 point mutants and ΔpUL21 when compared to similarly passaged WT HSV-1 as assessed by next-generation sequencing. Top: Schematic representation of the HSV-1 genome is shown with alternating background colouring corresponding to HSV-1 genes, green denoting overlapping reading frames, and with selected genes labelled. Repeat regions excluded from the mapping of sequences (Mask) are shown in grey. U L and U S denote the unique long and unique short segments, respectively, R L /R L ′ are the inverted repeats bounding U L and R S /R S ′ are the inverted repeats bounding U S . Bottom: Non-synonymous variants (red) and insertions/deletions (purple) are shown across the HSV-1 coding sequences. Dotted grey line denotes 20% prevalence and selected high-prevalence variants are labelled. Mutations introduced into pUL21 by two-step Red recombination are not shown. (E) Model of the core kinase domain of pUS3 (residues 189–481) shown in cartoon representation with the catalytically important activation segment (grey) and glycine-rich loop (yellow) highlighted. Amino acid substitutions that are prevalent in adapted pUS21 point mutant viruses are shown as sticks (carbon atoms pink). The position of the catalytic metal ions (magenta spheres) and ATP (sticks, carbon atoms cyan) were modelled by superposition of pUS3 onto the structure of Mn2+ and ATP-bound protein kinase A [111]. https://doi.org/10.1371/journal.ppat.1009824.g004

As the rapid increase in spread of these mutant viruses could not be ascribed to reversion of the pUL21 mutation, we hypothesised that the gain of function arose from mutations elsewhere in the HSV-1 genome. Genomic DNA was extracted from the adapted (P3) mutant viruses, and from wild-type and ΔpUL21 viruses subjected to the same propagation strategy in Vero cells, and the genomes were sequenced by next-generation (Illumina) sequencing. Reads were mapped to an HSV-1 strain KOS genome [38] that had been updated to contain all the majority variants present in the wild-type sample, excluding all but one copy of the repeat regions from the analysis. This yielded 115,096 (WT), 248,145 (ΔpUL21), 9,858 (pUL21F242E), 46,399 (pUL21V243D) and 38,787 (pUL21FV242AA) read pairs aligned with high quality, corresponding to 64.7–89.4% coverage of the HSV-1 genome and 71.2–93.9% coverage of the protein coding sequences to a depth of at least ten reads (S4 Fig). Analysis of coding sequence variants between the WT genome and the point mutants identified a striking preponderance of high frequency missense mutations in the US3 open reading frame that encodes the serine/threonine kinase pUS3 (Fig 4D). The US3 variants most prevalent in the pUL21 point mutant viruses (with corresponding amino acid substitutions and frequency) were 135,698 G>T (D207Y; 34.9% in pUL21V243D), 135,702 G>A (S208N; 16.2% in pUL21V243D, 10.6% in pUL21F242E and 59.2% in pUL21FV242AA), 136,127 G>A (A350T; 67.1% in pUL21F242E) and 136,496 C>T (L473F; 29.7% in pUL21V243D). Mapping these mutations onto a structural model of the core pUS3 kinase domain (Fig 4E) generated using trRosetta [39] identified that these mutations lie in strand β2 overlaying the ATP binding pocket (D207Y and S208N), within the highly conserved AlaProGlu (APE) motif that forms part of and stabilises the kinase activation segment (A350T), and in helix αI of the C-terminal lobe (L473) that interacts with the APE motif to stabilise the activation loop [40,41]. These substitutions would all be likely to disrupt local folding of pUS3 and thereby alter its activity and/or stability. The ΔpUL21 P3 genome sequence also contained variants in US3, although at lower frequency than in the pUL21 point mutant viruses. The most prevalent was 136,523 T>C (31.2% prevalence), which would abolish the pUS3 stop codon and thereby append an additional 70 amino acids to the expressed protein. A number of low frequency missense mutations were observed in US8A (and the US8/US8A overlap region) in the ΔpUL21 P3 genome sequence. While poor read depth precluded analysis of this region in the pUL21F242E and pUL21V243D viruses, these mutations were not prevalent in the pUL21FV242AA P3 genome (Fig 4D). The missense mutation 12,006 C>G (L122F, 33.3%) present in pUL21F242E would be unlikely to affect the function of pUL4 as a phenylalanine is observed at this position in other alphaherpesviruses. Similarly, the missense mutations observed in pUL36 (77,646 A>C/T/G; 77,649 T>G/C), pUL37 (82,715 T>A/G/C) and pUL42 (93,123 T>A/G/C) are unlikely to have arisen as adaptation to disruption of pUL21 activity as these variants are also observed in the wild-type virus (Fig 4D).

The observation that the US3 kinase gene rapidly mutates to compensate for the inability of pUL21 mutants to recruit PP1 suggests that the two proteins have antagonistic activities, with pUL21 promoting PP1-mediated dephosphorylation of proteins that are phosphorylated by pUS3. The plaques formed by HSV-1 lacking pUS3 expression (ΔpUS3) are indistinguishable from wild-type plaques in Vero cells (Fig 4A). In HaCaT cells the size of ΔpUS3 plaques is greatly reduced (S3 Fig), possibly due to the role of pUS3 in evading antiviral responses [42,43] that are more robust in HaCaT than Vero cells [44,45]. Complementation of pUL21 in trans in HaCaT cells restores ΔpUL21 HSV-1 plaques to the size of the wild-type virus (S3 Fig). New stocks of wild-type, pUL21FV242AA and ΔpUL21 HSV-1 were thus prepared via amplification of the original BAC-derived (P0) virus for two generations in HaCaT pUL21 cells (H2 virus), on the basis that pUL21 trans-complementation and the selective pressure to maintain pUS3 function in HaCaT cells would synergise to delay the compensatory pUS3 adaptation. The replication of these viruses was assessed using a single-step growth assay where Vero and HaCaT cells were infected at a high multiplicity of infection (MOI) and the production of infectious progeny was measured across the 24 hour replication cycle of the virus (Fig 5A). There is no difference in the growth of wild-type HSV-1 propagated in Vero cell (P4) or HaCaT pUL21 cells (H2). The growth of both pUL21FV242AA H2 and ΔpUL21 H2 HSV-1 was approximately 10 to 100-fold reduced compared to wild-type virus. Furthermore, the plaques formed on Vero cells by the pUL21FV242AA H2 and ΔpUL21 H2 viruses were extremely small (Fig 5B), confirming that these viruses had not adapted to culture when propagated in HaCaT pUL21 cells. The similar growth defect of these two mutant viruses confirms that the ability to recruit PP1 is a critical function of pUL21.

PPT PowerPoint slide

PowerPoint slide PNG larger image

larger image TIFF original image Download: Fig 5. Abolition of pUL21 PP1 binding severely restricts HSV-1 replication and spread. (A). Single-step (high MOI) growth curve of WT and mutant HSV-1. Monolayers of Vero and HaCaT cells were infected (MOI = 5) with the indicated viruses, prepared following amplification of the original rescued stock for three or four generations in Vero cells (P3 and P4, respectively) or for two generations in HaCaT cells stably expressing pUL21 (H2). Samples were harvested at the indicated times and titers were determined by plaque assay using Vero cells. Data are presented as mean values of duplicates of one representative experiment. Error bars represent standard error of the mean (not shown where errors are smaller than the symbols). (B) Representative plaque assays used to titre the single-step growth curve shown in (A). Monolayers of Vero cells were infected with viruses harvested from titration, overlaid with medium containing 0.6% carboxymethyl cellulose, and then fixed and immunostained with chromogenic detection at 48 h post-infection. https://doi.org/10.1371/journal.ppat.1009824.g005

While the unadapted pUL21FV242AA mutant virus exhibited small plaques and reduced growth, the adapted pUL21FV242AA P3 virus has wild-type levels of growth in Vero and HaCaT cells (Fig 5A). This restoration of growth to wild-type levels is particularly striking because the replication of the pUL21FV242AA P3 virus on HaCaT cells dramatically exceeds that of the ΔpUS3 virus (Fig 5A). This suggests that abolishing kinase (ΔpUS3) or phosphatase (ΔpUL21 and non-adapted pUL21FV242AA) activity individually is more deleterious to virus replication than simultaneously modulating both counteracting activities.

The HSV-1 pUS3 kinase has a broad substrate specificity [46] that overlaps with those of the cellular kinases Akt [47] and protein kinase A (PKA) [48]. Immunoblotting confirmed an increase in phosphorylated Akt and PKA substrates when Vero cells were infected with wild-type HSV-1 but not when infected with ΔpUS3 HSV–1 (Fig 6A). The abundance of several phosphorylated substrates is further increased in cells infected with ΔpUL21 or unadapted (H2) pUL21FV242AA HSV-1, whereas the adapted pUL21FV242AA P3 virus exhibits an overall reduction in the phosphorylation of the pUS3 targets (Fig 6A). The same changes were also observed in HaCaT cells infected with the same panel of viruses (S5A Fig). This confirms that the mutations observed in the adapted virus impair pUS3 kinase activity and suggests that pUL21 may directly antagonise the pUS3 activity for some substrates.

PPT PowerPoint slide

PowerPoint slide PNG larger image

larger image TIFF original image Download: Fig 6. pUL21 antagonises pUS3-mediated protein phosphorylation and NEC distribution. (A). Vero cells were infected at MOI = 5 with wild-type (WT) or mutant HSV-1 that had been prepared following amplification of the original rescued stock for three or four generations in Vero cells (P3 and P4, respectively) or for two generations in HaCaT pUL21 cells (H2). Lysates were harvested at 16 hpi and subjected to SDS-PAGE plus immunoblotting. With the exception of the lysate from WT HSV-1 infected cells that was treated with lambda phosphatase (λ phos), all lysates were harvested in the presence of phosphatase inhibitors. Immunoblots were probed with the antibodies shown. Antibodies recognising phosphorylated Akt and PKA substrates (sub) illustrate the activity of pUS3, as the specificity of this viral kinase overlaps with those of cellular kinases Akt and PKA [47,48]. For the Akt sub immunoblot two exposures are shown, separated by a line, and phosphorylated Akt substrates more abundant in cells infected with virus lacking pUL21 that can recruit PP1 (ΔpUL21 and pUL21FV242AAH2) are marked with arrowheads. (B) Vero cells were infected at MOI = 1 with WT or mutant HSV-1, prepared as in (A). Cells were fixed at 10 hpi and stained using antibodies that recognise pUL34 (green) and pUL21 (magenta). The merge includes DAPI (blue) and the scale bar represents 10 μm. (C) Vero cells were infected at MOI = 5 with WT or mutant HSV-1 as listed. Lysates were harvested at 16 hpi and subjected to SDS-PAGE plus immunoblotting using the antibodies listed. The upper strips of the pUL31 and pUL34 blots depict SDS-PAGE where PhosTag reagent was added to enhance separation of hyperphosphorylated (pUL31P and pUL34P) and hypophosphorylated (pUL31O and pUL34O) forms of the proteins. Non-specific bands are indicated with an asterisk (*). Boxed region denotes pUL31 bands used for quantitation in (D). (D) Relative abundance of pUL31 phosphoforms. Two-dimensional intensity profiles of pUL31 immunoblots following PhosTag SDS-PAGE (C), calculated by averaging horizontal pixel intensities for each lane (arbitrary units) along the vertical axis, are shown as purple curves. The intensity profile for pUL21 WT P4 infected cells is shown for comparison as a pink filled curve. Prominent differences in the profiles for cells infected with virus lacking pUL21 that can recruit PP1 (ΔpUL21 and pUL21FV242AAH2) are denoted with arrowheads. https://doi.org/10.1371/journal.ppat.1009824.g006

One of the best characterised roles of pUS3 is in the regulation of herpesvirus nuclear egress. pUS3 helps reorganize the nuclear lamina by phosphorylating lamin A/C [46], promotes redistribution of the inner nuclear membrane associated protein emerin [49,50], and may facilitate fusion of the primary envelope with the outer nuclear membrane via phosphorylation of gB [51]. Furthermore, pUS3 has been shown to directly phosphorylate the NEC components pUL31 [52,53] and pUL34 [54,55] and thereby regulate nuclear egress of capsids. Recent evidence suggests that pUL21 may also regulate NEC function as nuclear egress is severely disrupted in HSV-2 where pUL21 expression is absent [12], and in both HSV-1 and HSV-2 infection the distribution of the NEC in the nuclear membrane was perturbed when either pUS3 or pUL21 expression was abolished [18]. Immunocytochemistry showed a smooth distribution of the NEC component pUL34 around the nuclear rim in wild-type HSV-1 infected Vero cells (Fig 6B). pUL21 partially co-localises with pUL34 at the nuclear rim, also being observed in the nucleoplasm, at cytoplasmic punctae and at the cell surface (Fig 6B). As previously identified [18], we observed pUL34 distributed as small punctae on the nuclear rim in cells infected with ΔpUS3 HSV-1 whereas pUL34 formed large punctae in cells infected with ΔpUL21 virus (Fig 6B). The distribution of pUL34 in cells infected with unadapted pUL21FV242AA H2 HSV-1 closely resembles the ΔpUL21 infection, with large punctate accumulations of pUL34 observed at the nuclei of infected cells (Fig 6B). The subcellular localisation of pUL21 was broadly similar between this mutant and the wild-type virus, confirming that the conserved motif is not required for recruitment of pUL21 to the nuclear rim, although the pUL21 cytoplasmic punctae are more prominent (Fig 6B). Strikingly, the distribution of pUL34 in cells infected with the adapted pUL21FV242AA P3 mutant resembles that of ΔpUS3 virus (Fig 6B), showing that the small-punctae phenotype arising from altered pUS3 kinase activity [53] is dominant over the large-punctae phenotype that may arise from an absence of pUL21-directed phosphatase activity. Similar changes to pUL34 localisation were also observed in HaCaT cells infected with the same panel of viruses (S5B Fig).

To directly assess the effects upon NEC phosphorylation, the level of phosphorylated pUL31 and pUL34 was assessed following infection of Vero cells. As observed previously [52,53,55], pUL31 and pUL34 are phosphorylated in cells infected with wild-type HSV-1 but not in cells infected with a ΔpUS3 mutant (Fig 6C). The extent of pUL31 hyperphosphorylation is increased when cells are infected with ΔpUL21 HSV-1 or the unadapted pUL21FV242AA H2 virus (Fig 6C and 6D). However, the extent of pUL34 phosphorylation remains unaltered, consistent with pUL21 specifically targeting pUL31 but not pUL34 for PP1-mediated dephosphorylation. The extent of pUL31 phosphorylation in cells infected with the adapted pUL21FV242AA P3 mutant virus is similar to wild-type infection and pUL34 phosphorylation is reduced compared to wild-type, consistent with impaired pUS3 kinase activity in the adapted virus. The abundance of pUS3 is decreased in pUL21FV242AA P3 infected cells compared with wild-type HSV-1 or other pUL21 mutants (Fig 6A and 6C), in accordance with the adaptive mutations reducing pUS3 stability in addition to altering its kinase activity. Similar changes to the phosphorylation of pUL31 and pUL34 are observed in infected HaCaT cells (S5C Fig). However, we note reduced abundance of viral proteins in HaCaT cells infected with ΔpUS3 HSV-1, presumably due to restricted virus infection caused by an absence of pUS3-mediated innate immune evasion [42,43]. The abundance of pUL31 is particularly diminished in the ΔpUS3 infection, suggesting that pUS3 may be required for pUL31 stability in HaCaT cells (S5C Fig).

[END]

[1] Url: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1009824

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.