Dysregulated nuclear export of the late herpes simplex virus 1 transcriptome through the vhs-VP22 axis uncouples virus cytopathic effect and virus production

Herpes simplex virus 1 (HSV1) expresses its genes in a classical cascade culminating in the production of large amounts of structural proteins to facilitate virus assembly. HSV1 lacking the virus protein VP22 (Δ22) exhibits late translational shutoff, a phenotype that has been attributed to the unrestrained activity of the virion host shutoff (vhs) protein, a virus-encoded endoribonuclease which induces mRNA degradation during infection. We have previously shown that vhs is also involved in regulating the nuclear-cytoplasmic compartmentalisation of the virus transcriptome, and in the absence of VP22 many virus transcripts are sequestered in the nucleus late in infection. Here we show that despite expressing minimal amounts of structural proteins and failing to plaque on human fibroblasts, the strain 17 Δ22 virus replicates and spreads as efficiently as Wt virus, but without causing cytopathic effect (CPE). Nonetheless, CPE-causing virus spontaneously appeared on Δ22-infected human fibroblasts, and four viruses isolated in this way had all acquired point mutations in vhs which rescued viral mRNA export and late protein translation. However, unlike a virus deleted for vhs, these viruses still induced the degradation of cellular mRNA, suggesting that vhs mutation in the absence of VP22 is necessary to overcome a disturbance in mRNA export rather than mRNA degradation. The ultimate outcome of secondary mutations in vhs is therefore the rescue of virus-induced CPE caused by late protein synthesis, and while there is a clear selective pressure on HSV1 to mutate vhs for optimal production of late structural proteins, the purpose of this is over and above that of virus production. Author Summary HSV is a human pathogen that lytically infects cells of the epidermis. Following viral genome replication, structural proteins are produced in abundance to enable the rapid assembly and release of large quantities of infectious progeny. Infected cells also exhibit cytopathic effect (CPE), morphological changes that are exemplified by cell rounding and the breakage of cell-to-cell contacts, facilitating virus dissemination. Here we show that HSV1 with a mutation that results in the nuclear retention of viral mRNA and concomitant shutdown of late protein synthesis, also fails to cause CPE. However, unexpectedly, we found that this virus is still able to release large numbers of infectious virus which can spread between cells without any evidence of cell damage. Nonetheless, despite efficient virus productivity, this virus spontaneously mutates to rescue late protein production and CPE, with mutations mapping to the process of mRNA export. There is therefore a clear selective pressure on HSV1 to optimize the synthesis of late structural proteins, but the purpose of this is over and above that of virus production, a result that has implications for why viruses in general express such large amounts of structural proteins.


Abstract 22
Herpes simplex virus 1 (HSV1) expresses its genes in a classical cascade culminating in 23 the production of large amounts of structural proteins to facilitate virus assembly. HSV1 24 lacking the virus protein VP22 (D22) exhibits late translational shutoff, a phenotype that has 25 been attributed to the unrestrained activity of the virion host shutoff (vhs) protein, a virus-26 encoded endoribonuclease which induces mRNA degradation during infection. We have 27 previously shown that vhs is also involved in regulating the nuclear-cytoplasmic 28 compartmentalisation of the virus transcriptome, and in the absence of VP22 many virus 29 transcripts are sequestered in the nucleus late in infection. Here we show that despite 30 expressing minimal amounts of structural proteins and failing to plaque on human 31 fibroblasts, the strain 17 D22 virus replicates and spreads as efficiently as Wt virus, but 32 without causing cytopathic effect (CPE). Nonetheless, CPE-causing virus spontaneously 33 appeared on D22-infected human fibroblasts, and four viruses isolated in this way had all 34 acquired point mutations in vhs which rescued viral mRNA export and late protein 35 translation. However, unlike a virus deleted for vhs, these viruses still induced the 36 degradation of cellular mRNA, suggesting that vhs mutation in the absence of VP22 is 37 necessary to overcome a disturbance in mRNA export rather than mRNA degradation. The 38 ultimate outcome of secondary mutations in vhs is therefore the rescue of virus-induced 39 CPE caused by late protein synthesis, and while there is a clear selective pressure on HSV1 40 to mutate vhs for optimal production of late structural proteins, the purpose of this is over 41 and above that of virus production. Herpes simplex virus type 1(HSV1) expresses its genes in a classical cascade of gene 62 expression during lytic infection, comprising immediate-early, early and late genes [1]. In 63 general, the late genes encode virus structural proteins and are transcribed predominantly 64 from replicated DNA genomes, leading to a large burst of late protein synthesis for optimal 65 virus assembly. Deletion of the HSV1 UL49 gene which encodes the tegument protein VP22 66 [2] results in a virus that exhibits translational shutoff of late protein synthesis [3,4], and in 67 many systems is detrimental to virus propagation [4][5][6]. This translational shutoff is not a 68 consequence of enhanced host responses such as the stress response kinase protein 69 kinase R. Rather, it correlates with the failure of late viral transcripts to be exported from 70 the nucleus of cells infected with the D22 virus, as demonstrated in our previous studies 71 using mRNA FISH, thereby preventing their translation in the cytoplasm [4]. 72 A clue to the mechanism by which viral transcripts are retained in the nucleus of D22-infected 73 cells came from the observation that spontaneous secondary mutations frequently arise in 74 the UL41 gene of the D22 genome [4, 6, 7], a gene which encodes the virion host shutoff 75 (vhs) protein [8]. These mutations rescue the deleterious effect of VP22 deletion on late 76 protein translation, restoring plaque formation [4,6,7]. The vhs protein is an 77 endoribonuclease which induces the degradation of cellular mRNA during HSV1 infection 78 through its endoribonuclease cleavage of cytoplasmic mRNAs followed by Xrn1 79 exonuclease degradation [9], and regulates the transition from IE to E and L gene 80 expression [10,11]. It was therefore originally proposed that VP22 is required to quench 81 vhs-induced mRNA degradation at later times in infection and that in the absence of VP22, 82 vhs endoribonuclease activity is lethal [6]. Nonetheless in our hands, infection of human 83 fibroblasts with a D22 virus did not result in unrestrained mRNA degradation compared to 84 Wt infection [4]. Moreover, in that study we also demonstrated that in Wt infection, IE and E 85 transcripts were retained in the nucleus at late times, but in cells infected with a Dvhs virus 86 all classes of transcripts were exported to the cytoplasm [4], suggesting that the vhs 87 endoribonuclease is involved in regulating mRNA export, and providing a link between 88 mRNA degradation in the cytoplasm and mRNA retention in the nucleus. The relative 89 compartmentalisation of the virus transcriptome was also mirrored by the localisation of the 90 polyA binding protein PABPC1, a protein that has a steady-state cytoplasmic localisation 91 but shuttles between the cytoplasm and nucleus to bind polyadenylated mRNAs ready for 92 export [12,13]. Once mRNA in the cytoplasm has been turned over, PABPC1 recycles back 93 to the nucleus, and in the presence of functional vhs, PABPC1 accumulates there in an 94 endoribonuclease dependent fashion [14,15]. 95 Here we report the unexpected result that despite translational shutoff and lack of plaque 96 formation in primary human fibroblasts, HSV1 lacking VP22 replicates, spreads and 97 produces as much infectious progeny virus as Wt virus without causing any cytopathic effect 98 (CPE) in these cells. Nonetheless, CPE-inducing virus rapidly appeared as plaques in D22-99 infected fibroblasts and these rescued viruses had all acquired mutations in vhs which 100 restored nuclear export of viral mRNA during infection rather than abrogating mRNA 101 degradation. This suggests that despite efficient virus propagation in the absence of VP22, 102 there is pressure on the virus to mutate vhs and restore late protein translation and 103 concomitant CPE, over and above what is required for virus production. These results have 104 implications for understanding why this and potentially other viruses express such large 105 amounts of late virus proteins. 106

HSV1 lacking VP22 replicates and spreads in primary human fibroblast cells without 108
causing cytopathic effect. Deletion of the VP22-encoding gene (UL49) has been shown 109 to be detrimental to HSV1, resulting in extreme late translational shutoff [4,6,7]. Our own 110 D22 virus based on strain 17 fails to plaque on primary human fibroblasts (HFFF) as late as 111 5 dpi (Fig 1A). The efficiency of viral DNA replication in D22 infection was measured by 112 harvesting at 2 or 16 h after infection and determining the relative viral DNA copy number 113 by qPCR of the virus gene UL48, to reflect input viral DNA (2 h) or viral DNA replication (16 114 h). Although there was less input DNA in D22 infected cells, the relative increase in genome 115 copies was similar in Wt and D22 infected cells at 16 h, indicating that the absence of VP22 116 has little effect on genome replication (Fig 1B), and that the block to virus production occurs 117 at a later stage. Western blotting of HFFF cells infected at high multiplicity confirmed that a 118 range of virus envelope proteins are poorly expressed (Fig 1C), in line with the previously 119 demonstrated translational shutoff in these cells [4], providing an obvious explanation for the 120 inability of this virus to form plaques in HFFF cells. Immunofluorescence of infected cells 121 revealed that the IE protein ICP4, which localises in a distinctive cytoplasmic punctate 122 pattern late in Wt infection, was restricted to the nucleus in D22 infected cells (Fig 1D, ICP4) 123 while the envelope protein glycoprotein E (gE) was concentrated in a juxtanuclear 124 compartment rather than progressing to the plasma membrane as it does in Wt infection 125 ( Fig 1D, gE). These results suggest there is a block to late protein trafficking in the absence 126 of VP22. However, despite these obvious defects in D22 infection, we found no significant 127 difference in the growth kinetics of D22 compared to Wt virus in HFFF in a one-step growth 128 curve ( Fig 1E). Moreover, low magnification imaging of HFFF cells infected with D22 (which 129 expresses GFP in place of VP22) showed that while all cells were GFP positive after 20 h, 130 there was no sign of the classical HSV1-induced cytopathic effect (CPE) of cell-rounding, 131 which was evident in cells infected with Wt or HSV1 expressing GFP fused to VP22 (GFP-132 22) infected cells (Fig 1F). By contrast, HSV1 expressing GFP in place of UL34, a protein 133 essential for nuclear egress [16], exhibited CPE similar to Wt and GFP-22 (Fig 1F), 134 indicating that even though this virus is unable to export capsids to the cytoplasm or 135 assemble progeny virions, it is still able to cause CPE. 136 Given that the D22 virus does not plaque on HFFF, we next investigated its ability to spread 137 in these cells. A multi-step growth curve was carried out by infecting HFFF cells at a 138 multiplicity of 0.01 and intriguingly, this also revealed little difference in the replication or 139 release of Wt and D22 viruses, in a scenario where optimal virus replication requires multiple 140 rounds of replication and spread to other cells in the monolayer (Fig 2A). GFP imaging of 141 cells infected at low multiplicity revealed that the entire monolayer of cells had become GFP 142 positive but without causing CPE, indicating that the D22 virus spreads efficiently without 143 affecting the integrity of the cells (Fig 2B). To further visualise the behaviour of the D22 virus Taken together with studies from other groups, which have described the rescue of D22 163 replication through spontaneous mutation of vhs [6,7], and single residue changes in vhs 164 having a profound effect on its activity [17,18], these results led us to initially hypothesize 165 that the A95T mutation had inactivated the vhs endoribonuclease activity, thereby rescuing 166 late protein synthesis and subsequent virus replication. We have now undertaken a more 167 extensive analysis of this and three additional rescued viruses that were isolated from 168 plaques on HFFF and which formed plaques approaching the size of Wt plaques (Fig  169   3A).They all express full-length vhs as demonstrated by Western blotting (Fig 3B), indicating 170 that no gross mutations had occurred, but metabolic labelling profiles confirmed that all 171 these viruses had rescued the extreme translational shutoff exhibited by the D22 virus, albeit 172 the PP13 virus recovering only slightly from the D22 base line (Fig 3C). 173 Sequencing of the UL41 gene in these rescue viruses revealed that they all had point-174 mutations in the vhs open reading frame ( Fig 3D). To determine if these variants were 175 present in our original D22 virus stock or had arisen during propagation on HFFF cells, we 176 carried out next generation sequencing of four amplicons covering the UL41 gene generated 177 from the genome of our D22 virus stock (Fig 3D), revealing that it already contained the 178 T102M and V271A variations at a rate of 33% and 35% respectively, with I223V at a much 179 lower rate of 2% ( Fig 3D). No V33A or A95T variations were found by deep sequencing this 180 virus suggesting that they may have arisen spontaneously during propagation on HFFF. 181 Interestingly, direct sequencing of the UL41 gene from 16 viruses isolated from plaques on 182 Vero cells in which this virus is able to plaque, or from nine non-CPE fluorescent foci on 183 HFFF such as those shown in Fig 2C, showed that all viruses contained either the T102M 184 or the V271A variation but none of them contained two mutations. This suggests that each 185 of these single variations in isolation was not sufficient to rescue CPE of this virus in HFFF,186 and that, with the exception of the A95T variation, a second point mutation was required. In uninfected cells, PABPC1 has a steady state cytoplasmic localisation, but shuttles 204 between the nucleus and the cytoplasm, binding the polyA tail of mRNAs in the nucleus and 205 being transported out on those tails. It then returns to the nucleus after mRNA turnover in 206 the cytoplasm to be exported again [12]. We have previously shown that in the absence of 207 VP22, the cellular polyA binding protein PABPC1 accumulates to high levels in the nucleus 208 [4, 14], a result we had postulated to be the consequence of the aforementioned nuclear 209 retention of late viral mRNA. We therefore examined the relative compartmentalisation of 210 PABPC1 in HFFF cells infected with Wt, Dvhs, D22 or rescue viruses at 10 and 16 hours 211 after infection. As shown before, PABPC1 had partially accumulated in the nucleus of Wt 212 infected cells at 16 h, but remained cytoplasmic in Dvhs infected cells throughout, confirming 213 the role that vhs plays in nuclear relocalisation of PAPBC1 (Fig 5). By contrast, in D22 214 infected cells, PABPC1 had already accumulated in nuclei by 10 h, and was almost entirely 215 nuclear by 16 h, correlating with the extensive accumulation of viral mRNA seen at this time 216 indicated that all rescued virus infections expressed L transcripts to a level much closer to 232 Wt than D22 infection, while D22* exhibited high levels of IE and E approaching those found 233 in Dvhs-infected cells ( Fig 6A). This suggests that the vhs point mutations in the D22 rescue 234 viruses had restored not only the export but also the level of virus transcripts. 235 To determine the relative effect of these additional vhs variants on vhs-induced mRNA 236 degradation, RNA samples were harvested 15 hours after infection and analysed by RT-237 qPCR for two cellular transcripts we have previously shown to be highly susceptible to vhs 238 degradation -MMP1 and MMP3 [4] (Fig 6B). Surprisingly, all rescue viruses maintained the 239 ability to reduce the levels of both these transcripts compared to Dvhs, with the PP13 and 240 PP15 viruses maintaining activity close to that seen in their parent D22 virus (Fig 6B). To 241 look in more detail at the behaviour of the weakest of these vhs variants present in D22*, we 242 carried out a time course of relative mRNA levels of MMP1 and MMP3 in comparison to Wt, 243 D22 and Dvhs infected HFFF. This confirmed that unlike the situation in Dvhs infected cells 244 where neither the MMP1 nor MMP3 transcript levels were altered during infection, the D22* 245 virus caused the gradual decline in these transcripts over time, which as in the D22 infection 246 began around 6 hpi and progressed through infection ( Fig 6C). 247 Taking these results together, we have shown that all viruses that were rescued from the 248 D22 virus retained vhs activity for mRNA degradation, suggesting that the rescue of D22 249 virus requires vhs mutation to restore late transcript export rather than simple inactivation of 250 vhs endoribonuclease activity. Moreover, this restoration of late transcript export has 251 restored late protein expression together with the ability of these viruses to cause CPE in 252 an environment where virus production had nonetheless not been compromised in the first 253 Primary human fibroblasts offer an excellent model for HSV1 -they are semi-permissive to 258 infection, having the capacity to restrict HSV1 at early stages, and they have a functioning 259 interferon pathway [19]. They also exhibit extreme CPE in response to HSV1 infection with 260 profound changes to cell architecture from long spindle-like to small rounded up cells (Fig  261   1E). It was therefore intriguing to discover that our mutant D22 HSV1 virus was able to 262 enter, replicate and spread within HFFF cells in a similar fashion to Wt virus but without 263 causing CPE. CPE is generally considered to be a combination of gross morphological 264 changes including cytoskeletal and membrane reorganisation, together with physiological 265 and biochemical changes caused by the virus hijacking cellular activities. The absence of 266 CPE and plaque formation in D22 infected HFFF cells, which we had originally assumed to 267 indicate attenuation, correlates with late translational shutoff suggesting that CPE is caused 268 by the high levels of late virus proteins, or specific proteins, made within the cell. Indeed, an 269 HSV1 deleted for ICP34.5 -a protein whose absence results in translational shutoff via the 270 PKR-eiF2a pathway -also fails to cause CPE in human fibroblasts [20]. Moreover, it has 271 recently been reported that HSV1 which fails to cause CPE in culture has been isolated from frame. These SNPs cluster within the N-terminal half of vhs (Fig 7), fitting with other 284 mutations found in previous D22 virus studies which cluster in conserved domain III of the 285 protein [18,22,23]. These have been interpreted previously as mutations that inactivate vhs 286 endoribonuclease activity [6,7,24] and in support of this, analysis of vhs sequences from 287 26 published strains of HSV1 showed that naturally occurring SNPs cluster within the C-288 terminus of the protein while the N-terminus is highly conserved (Fig 7). Moreover, in the 289 course of this study, the vhs gene from ten clinical isolates of HSV1 [25] were also 290 sequenced, with seven of them identical to strain 17, and three containing SNPs in the C-291 terminal half already identified in the published sequences. Nonetheless, our rescue viruses 292 retain the ability to degrade cellular mRNA -a readout for endoribonuclease activity -whilst 293 regaining the ability to export late mRNA to the cytoplasm for access to the translation 294 machinery. As such, these mutants have separated the contribution that vhs makes to 295 nuclear export and mRNA degradation. This is further emphasized by the early relocalisation 296 of PABPC1 to the nucleus of D22 infected cells at a time when mRNA degradation is reduced 297 compared to Wt infection (Fig 6B) [4], suggesting that mRNA and hence PABPC1 retention 298 in the nucleus can be uncoupled from mRNA degradation. Moreover, we have also recently 299 described a virus expressing vhs tagged with GFP, that fails to degrade cellular mRNA, is 300 unable to relocalise PABPC1 during infection, but can still cause nuclear retention of IE and 301 E transcripts [26]. 302 Although we do not yet understand the molecular basis of these complex phenotypes, the 303 broad range of SNPs across the N-terminus of vhs will allow us to further separate the 304 activities of mRNA degradation, PABPC1 relocalisation and mRNA nuclear retention. In 305 addition, it is important to note that although our D22 virus which is based on strain 17 does 306 not plaque in human fibroblasts, it is able to form small plaques, and hence CPE, on Vero 307 cells [4,27]. By contrast, other reports of HSV1 VP22 deletion viruses based on strain F 308 have indicated that these viruses are unable to plaque even on Vero cells [5,6]. The VP22 309 and vhs sequences in these two strains have several SNPs compared to strain 17, providing 310 scope for strain variation in the interplay between these proteins and the 311 machinery/pathways involved in regulating protein translation. 312 The work presented here adds further weight to the important roles that vhs and VP22 play 313 in the co-ordinated regulation of mRNA localisation and translation. HSV1 also expresses 314 the IE protein ICP27 which is known to be essential for late protein expression [28], and is 315 required for mRNA export from the nucleus by binding TAP/NXE1 to engage with the cellular 316 Aly/REF export pathway [29][30][31]. It is therefore paradoxical that one activity of vhs is to retain 317 virus transcripts in the nucleus, suggesting that vhs and ICP27 may work in opposition to 318 each other. Going forward, it will now be important to establish how vhs (and VP22) intersect 319 with the nuclear export activity of ICP27 to co-ordinate virus gene expression via mRNA 320

compartmentalisation. 321
The question remains as to why there is a selective pressure on the virus to restore late 322 translation/CPE through vhs mutation in the absence of VP22, if the virus is able to replicate 323 and spread as efficiently as Wt virus. One could imagine that a virus that propagates itself 324 without causing damage to its host cell might be able to survive "under the radar" of host 325 sensing mechanisms and antiviral measures. However, the rapid and reproducible 326 appearance of secondary mutations within vhs suggests that the production of a large 327 amount of CPE-causing structural proteins is advantageous to the virus, over and above the 328 requirement for virus assembly. It is therefore likely that structural proteins or a subset of 329 them are required to maintain a favourable environment for virus survival, and as suggested 330 elsewhere may reflect the ongoing battle between host and virus during virus infection [32]. 331 As such, the unexpected results described here may ultimately prove highly informative 332 about the range of mechanisms by which HSV1 overcomes host defences.

Cells and Viruses. HFFF and Vero cells (both obtained from European Collection of 335
Authenticated Cell Cultures -ECACC) were cultured in DMEM supplemented with 10% 336 foetal bovine serum (Invitrogen). Viruses were routinely propagated in Vero cells, with 337 titrations carried out in DMEM supplemented with 2% foetal bovine serum and 1% human 338 serum. HSV1 strain 17 (s17) was used routinely. The s17 derived VP22 deletion mutant 339 (D22) and the vhs knockout virus (Dvhs) have been described before [27,33]. The four D22 340 rescue viruses (D22*, PP12, PP13 and PP15) were isolated from plaques that appeared 341 spontaneously on HFFF cells. polyacrylamide gel electrophoresis. Following fixation in 50% v/v ethanol and 10% v/v acetic 360 acid, the gel was vacuum dried onto Whatman filter paper and exposed to X-ray film 361

overnight. 362
Quantitative RT-PCR (RT-qPCR). Total RNA was extracted from cells using Qiagen 363 RNeasy kit. Excess DNA was removed by incubation with DNase I (Invitrogen) for 15 min at 364 room temperature, followed by inactivation for 10 min at 65°C in 25 nM of EDTA. Superscript 365 III (Invitrogen) was used to synthesise cDNA using random primers according to 366 manufacturer's instructions. All qRT-PCR assays were carried out in 96-well plates using 367

MESA Blue qPCR MasterMix Plus for SYBR Assay (Eurogentec). Primers for viral genes 368
are shown in Table 1. Primers for cellular genes MMP1 and MMP3 were obtained from 369

Quantification of viral DNA. HFFF cells infected at MOI 3 were acid washed 1 h after 373
infection to inactivate unpenetrated virus, then harvested at 2 or 16 h after infection. DNA 374 was harvested using the DNeasy blood and tissue kit (Qiagen), and qPCR assays were 375 carried out in a LightCycler96 system (Roche), using MESA BLUE qPCR kit for SYBR assay 376 (Eurogentec) according to the manufacturer's instructions with primers for 18S (see Table  377 1) and HSV1 UL48 gene. 378 Amplicon Sequencing. The full-length UL41 gene was amplified in four fragments by PCR 379 from our virus submaster stock of the D22 virus, using the primers shown in Table 2. Each 380 PCR fragment was purified and subjected to amplicon next generation sequencing 381 (Genewiz) providing around 40,000 reads per amplicon. 382 Immunofluorescence. Cells for immunofluorescence were grown on coverslips and fixed 383 with 4% paraformaldehyde in PBS for 20 min at room temperature, followed by 384 permeabilisation with 0.5% Triton-X100 for 10 min. Fixed cells were blocked by incubation 385 in PBS with 10% newborn calf serum for 20 min, before the addition of primary antibody in 386 PBS with 10% serum, and a further 30-min incubation. After extensive washing with PBS, 387 the appropriate Alexafluor conjugated secondary antibody was added in PBS with 10% 388 serum and incubated for a further 15 min. The coverslips were washed extensively in PBS 389 and mounted in Mowiol containing DAPI to stain nuclei. Images were acquired using a Nikon 390 A1 confocal microscope and processed using ImageJ software [37]. 391 Mowiol containing DAPI to stain nuclei, and images acquired with a Nikon A2 inverted 402 confocal microscope and processed using Adobe Photoshop software. 403 Microscopy. Images were acquired on a Nikon A2 confocal microscope or CCD camera 404 system on an inverted Zeiss TV100 microscope and processed using Image J and Adobe 405

26.
Wise EL, Samolej J, Elliott G. Herpes simplex virus 1 expressing GFP-tagged virion 510 host shutoff (vhs) protein uncouples the activities of RNA degradation and differential 511   represented as Log2 fold change to uninfected (ΔΔCT), and the mean ± standard error for 628 n = 3 is shown. (C) HFFF cells were infected with Wt, Δ22, Δvhs or Δ22* viruses at a 629 multiplicity of 2, and total RNA was harvested at the indicated times (in hours). qRT-PCR 630 was carried out on MMP1 and MMP3 cellular transcripts with relative levels expressed as 631 log2 FC to uninfected (ΔΔCT) over time. The mean and ± standard error for n = 3 is shown. 632 Part of this data has been presented in a previous publication [4].  (Table  636 3); black -the well-characterised vhs1 mutation (T214I) that has been found to abrogate 637  with Wt s17 or D22 virus at a multiplicity of 3, acid washed at 1 hpi, then DNA was isolated 672 at 2 or 16 hpi. qPCR was performed for gene UL48 to determine the relative virus DNA copy 673 number represented as DCt to 18s rDNA (mean±SEM, n = 3). (C) HFFF cells infected with 674 Wt (s17) or D22 viruses at MOI 2 were harvested at 16 hpi and analysed by SDS-PAGE and 675 Western blotting with antibodies as indicated. (D) HFFF cells infected with Wt (s17) or D22 676 viruses at MOI 2 were fixed at 16 hpi and analysed by immunofluorescence with antibodies 677 to the IE protein ICP4 and the L protein glycoprotein E (gE), both in green. Nuclei were 678 approximately 20 pfu of D34 virus and representative brightfield and GFP images acquired 694 at days 1 and 3. Scale bar = 100 µm. 695