Histone deacetylase inhibitors reduce the number of herpes simplex virus-1 genomes initiating expression in individual cells

Although many viral particles can enter a single cell, the number of viral genomes per cell that establish infection is limited. However, mechanisms underlying this restriction were not explored in depth. For herpesviruses, one of the possible mechanisms suggested is chromatinization and silencing of the incoming genomes. To test this hypothesis, we followed infection with three herpes simplex virus 1 (HSV-1) fluorescence-expressing recombinants in the presence or absence of histone deacetylases inhibitors (HDACi’s). Unexpectedly, a lower number of viral genomes initiated expression in the presence of these inhibitors. This phenomenon was observed using several HDACi: Trichostatin A (TSA), Suberohydroxamic Acid (SBX), Valporic Acid (VPA) and Suberoylanilide Hydoxamic Acid (SAHA). We found that HDACi presence did not change the progeny outcome from the infected cells but did alter the kinetic of the infection. Different cell types (HFF, Vero and U2OS), which vary in their capability to activate intrinsic and innate immunity, show a cell specific basal average number of viral genomes establishing infection. Importantly, in all cell types, treatment with TSA reduced the number of viral genomes. ND10 nuclear bodies are known to interact with the incoming herpes genomes and repress viral replication. The viral immediate early protein, ICP0, is known to disassemble the ND10 bodies and to induce degradation of some of the host proteins in these domains. HDACi treated cells expressed higher levels of some of the host ND10 proteins (PML and ATRX), which may down regulate the number of viral genomes initiating expression per cell. Corroborating this hypothesis, infection with three HSV-1 recombinants carrying a deletion in the gene coding for ICP0, show a reduction in the number of genomes being expressed in U2OS cells. We suggest that alterations in the levels of host proteins involved in intrinsic antiviral defense may result in differences in the number of genomes that initiate expression.

that establish infection is limited. However, mechanisms underlying this restriction were 23 not explored in depth. For herpesviruses, one of the possible mechanisms suggested is 24 chromatinization and silencing of the incoming genomes. To test this hypothesis, we 25 followed infection with three herpes simplex virus 1 (HSV-1) fluorescence-expressing 26 recombinants in the presence or absence of histone deacetylases inhibitors (HDACi's). 27 Unexpectedly, a lower number of viral genomes initiated expression in the presence of 28 these inhibitors. This phenomenon was observed using several HDACi: Trichostatin A 29 (TSA), Suberohydroxamic Acid (SBX), Valporic Acid (VPA) and Suberoylanilide 30 Hydoxamic Acid (SAHA). We found that HDACi presence did not change the progeny 31 outcome from the infected cells but did alter the kinetic of the infection. Different cell types 32 (HFF, Vero and U2OS), which vary in their capability to activate intrinsic and innate 33 immunity, show a cell specific basal average number of viral genomes establishing 34 infection. Importantly, in all cell types, treatment with TSA reduced the number of viral 35 genomes. ND10 nuclear bodies are known to interact with the incoming herpes genomes 36 and repress viral replication. The viral immediate early protein, ICP0, is known to 37 disassemble the ND10 bodies and to induce degradation of some of the host proteins in 38 these domains. HDACi treated cells expressed higher levels of some of the host ND10 39 proteins (PML and ATRX), which may down regulate the number of viral genomes 40 initiating expression per cell. Corroborating this hypothesis, infection with three HSV-1 41 recombinants carrying a deletion in the gene coding for ICP0, show a reduction in the 42 number of genomes being expressed in U2OS cells. We suggest that alterations in the levels 43 6 single color recombinants with DICP0 phenotype. Each cross was plaque purified to 134 homogeneity on either Vero cells (for wild type recombinants) or U2OS strains (for DICP0 135 recombinants). Viral recombinants OK29, OK32 and OK40 carry a single fluorescent 136 protein (mCherry, mTurq2 and EYFP, respectively) in an ICP0 deletion background. The 137 new recombinants were tested by PCR, phenotype and growth curves ( figure 5A and B). 138 Viruses were grown and tittered on either Vero (for wild type recombinants) or U2OS (for 139 (or solvent only) was added to cells. The infected cells were incubated at 37 o C for 6 to 8 148 hours until images of the infected cells were taken. All inhibitors were tested for cell 149 toxicity after 12 hours of incubation. In working concentration, none of the HDACi's 150 induced more than 5% cell death (tested by Propidium iodide positive cells). In 4 fold 151 higher concentration, only Valporic Acid (VPA) had less than 80% viability. Anacardic 152 Acid (AA) working concentration resulted in ~10% cell death and up to 40% in 4 fold 153 higher concentration. 154

Image acquisition and analysis 155
To estimate the number of HSV-1 genomes expressed in each infected cell, we obtained 156 images (as described above) using a Nikon Eclipse Ti-E epifluorescence inverted 157 microscope. Each experimental condition (different cells, viruses, inhibitors and MOI) was 158 replicated in two wells, and the experiment was performed at least twice. From an 159 individual well, five random areas were imaged. From each image, 100 cells were analyzed 160 for their color content. To define the average number of incoming genomes being 161 expressed, we used the mathematical equation for estimating the most likely average 7 number of genomes expressed in each cell (l) according to the number of one-(r1), two-163 (r2), or three-color (r3) cells out of the number of cells analyzed (n), as was previously 164 developed (Kobiler et al., 2010): 165 l=-3ln(1-(r 1 +2r 2 +3r 3 )/3n) 166 The l was calculated for each well individually (based on 500 cells) and for each condition 167 the mean l and standard deviations were calculated. To predict the significance of the 168 difference between the conditions a two tailed student T-test was performed. 169

Live cell imaging 170
Vero cells plated in 8 well Nunc™ Lab-Tek™ II Chambered Coverglass were used. 171 Infection with OK11 was carried out in the presence or absence of 1.32µM TSA as 172 described above. Images were acquired using a Nikon Eclipse Ti-E epifluorescence 173 inverted microscope every 10 min with DAPI and RFP fluorescence at 37 o C in a 5% 174 (vol/vol) CO2 enriched atmosphere using a Chamlide TC stage top incubator system (Live 175 Cell Instrument). Two experiments were done with technical repeats (wells) per condition. 176 From each well, five frames were taken and analyzed. First, using the Imaris 8.1 (Bitplane) 177 image software we identified the individual cells according to the Hoechst DNA staining. 178 The level of red fluorescence at each time point was measured for each identified cell by 179 the software. We removed all cells in which fluorescence levels did not increase above 180 10% during the infection, as most of these cells were either dead or resistant to infection. 181  (Table 2). qPCR was carried out and 191 analyzed in StepOne TM (Applied Biosystems TM ). mRNA levels were normalized to the 192 expression of GAPDH gene. The results were collected from five independent repeats. We 193 performed an outlier identification using the extreme studentized deviate (EMD) method. To ensure that our results are specific to HDACi, we also tested the effect of the histone 218 acetyltransferases (HAT) inhibitor, Anacardic Acid (AA). In the presence of AA, no 9 significant change in the number of expressed viral genomes per cell was observed in all 220 MOI tested (figure1E). Our results show that all four HDACi tested, similarly, reduce the 221 number of viral genomes expressed while the HAT inhibitor did not; suggesting that this 222 phenomenon is dependent on the specific inhibition of the deacetylase activity. 223 The viral progeny levels are independent of the number of genomes replicating per 224

cell. 225
To test the effect of the number of genomes replicating on viral progeny, we carried out 226 single step growth curves in the presence or absence of the HDACi's. We found that 227

Differences in the number of viral genomes expressed in human cells. 242
As HSV-1 coevolved with its human host, it is likely that some host factors, that are 243 can initiate expression in HFF compared to Vero or U2OS, especially in the lowest MOI 254 tested (10, Figure 3A and Table 3). In U2OS cells more viral genomes per cell are capable 255 of establishing infection, with the biggest difference seen in the highest MOI 100 (figure 256 3A and Table 3). These results suggest that cellular factors are key regulators determining 257 the number of viral genomes being replicated per cell. 258

cells. 260
Several studies have shown that the effect of HDACi on herpes replication is cell type  (Table 3). Taken together, these results suggest that 273 the effect HDACi have on the viral initiation of replication is independent of cell type. 274

Treatment with HDACi increases the levels of antiviral ND10 proteins. 275
The ability of HDACi to decrease the number of viral genomes initiating expression per 276 cell is contra-intuitive to the predicted role of HDACi on viral genomes. As we observed 277 significant differences among the cell lines tested (figure 3A and table 3), we assume that 278 alterations in the levels of host factors influence the probability of viral genomes to initiate 279 expression and replication. We hypothesize that the treatment with HDACi results in higher 280 expression of cellular genes that are involved in inhibiting viral replication. To test this 281 hypothesis, we selected three genes, PML, ATRX and hDaxx, all part of the ND10 bodies 282 and known to be involved in the intrinsic response against HSV-1. We tested the mRNA 283 levels of these genes following three-hour incubation in the presence or absence of TSA.  Table 3). This decrease is almost two folds and 309 results in numbers similar to those obtained by wild type infection in HFF cells. (We were 310 unable to obtain titers for the ICP0 null recombinants that will be sufficient for HFF 311 infection at comparable MOIs). Our findings indicate that the ICP0 protein has a major role 312 in allowing viral genomes to initiate expression, even in U2OS cells, probably by 313 promoting the degradation of antiviral host factors.  We found that at MOI 10, while a small decrease is observed (in most cases), a statistically 334 significance is not obtained at this MOI. We previously suggested that at MOI 10 and lower  We identified significant differences between the three cell types we tested ( Figure 3A). 364 We found that in U2OS cells, more incoming viral genomes are expressed compared to 365 HFF cells. In Vero cells, the number of genomes was higher than those of HFF but 366 significantly lower than those of U2OS cells. It is important to note that Vero cells 367   repeats. *P < 0.05, ** P < 0.01; ***P < 0.001; by t test. 602 to U2OS cells is presented for each strain as marked. Each bar represents the ratio between 617 two biological repeats and error bars represent standard deviations between the repeats. (C) 618 U2OS cells were infected with a mixture of the three ICP0 deleted fluorophore expressing 619 HSV-1 strains (light green). The average number of viral genomes expressed was 620 compared to wild type infection of U2OS cells (Dark red). Images taken 6-8 hpi were used 621 for calculating the average number of genomes expressed per cell. Each bar represents six 622 biological repeats and error bars represent standard deviations between the repeats. ** P < 623 0.01; ***P < 0.001; by t test. 624 Table 1. List of inhibitors used in this work. 625 Table 2. List of primers used for qPCR. Both forward (f) and reverse (r) primers for 626 amplifying the specific genes are listed. 627