Compartmented neuronal cultures reveal two distinct mechanisms for alpha herpesvirus escape from genome silencing

The cell bodies that harbor silenced PRV genomes express LAT transcripts

In our previous report, we established quiescent infections by infecting axons in the axonal “N” compartment of modified Campenot tri-chambers with PRV 180 (which expresses an mRFP-VP26 fusion protein) at an MOI of 0.01 (100 plaque forming units per dish) [5]. At this very low MOI, no cell bodies express any mRFP-VP26 over a period of 3 weeks. Silent genomes can be reactivated after high MOI superinfection of cell bodies with UV treated PRV. This reactivated productive infection then spreads to all neurons in the chamber [5]. These observations raised several important questions including how many cell bodies initially get infected at this very low MOI, and do these silent genomes express latency associated transcripts (LATs), a hallmark of authentic latent infection. To estimate the number of neurons that harbored a quiescent PRV infection, we established quiescent infections using a trans-complemented gB null mutant, which expresses a diffusible GFP under CMV promoter (PRV 233) [11]. This virus is completely deficient in cell-cell spread—only those neurons infected via their axons with PRV 233 (at an MOI of 0.01) expressed the green fluorescent protein in the soma (Fig 1A). We expected the number of cell bodies harboring silent genomes would be small because quiescence was established with approximately 100 infectious particles. We counted on average 9.8 [±2.9 standard error of the mean (SEM)] GFP positive cell bodies in each S compartment after low MOI PRV 233 infection between 3 to 21 days post infection (dpi). We also assessed the number of cell bodies that send axons to the N compartment by adding a lipophilic dye (DiI) in the N compartment. We found that 49% (±4.7% SEM) of neurons were connected to the N compartment, and of these, 0.45% (±0.2% SEM) were infected with PRV 233 (Fig 1B).

Next, we assayed the viral transcripts from the cell bodies either from cultures with silenced PRV 180 genomes (no capsid expression was detected in cell bodies after 7 dpi) or from cultures that were productively infected with PRV 180 by direct infection of cell bodies at the same low MOI (red capsid expression was detected in all cell bodes at 7 dpi). After 7 days, the ratio of LAT to EP0 mRNA was 34.5 fold more in silenced cultures when compared to cultures with productive cell body infection (Fig 1C). In productively infected neurons, the EP0 transcript was 148.3 (±15 SEM) fold more abundant than the LAT transcript. This ratio went down to 4.8 (±0.65 SEM) in quiescently infected neurons at 7 dpi. These experiments showed that PRV infection of axons at an MOI of 0.01 results in a silenced infection in less than 0.5% of connected cell bodies in S compartments. These silenced genomes do not express late genes, indicated by the lack of mRFP-VP26 capsid protein fluorescence, but they do express LAT transcripts, as expected of authentic latent genomes.

The complementation assay: Activation of PKA by forskolin or dbcAMP in cell bodies results in escape from silencing

We established a system to study the mechanism of how low MOI PRV 180 retrograde infection which is destined to be silenced could be redirected to a productive infection. In these assays, we do not wait 3 weeks for the full establishment of quiescence. Instead, we infect axons with PRV 180 at an MOI of 0.01 as before, but we simultaneously treat or infect cell bodies in S compartments with drugs, inactive/mutant viruses or virus-like particles (Fig 1D). Over 7 days, we monitor mRFP-VP26 expression in the cell bodies to determine the extent of productive PRV 180 infection. In this complementation assay, we call a productive infection an ‘escape from silencing’. This assay is fundamentally distinct from `reactivation assays`that aim to investigate the reversal of repressive genome modifications after latency is established.

Because elevated cyclic adenosine monophosphate (cAMP) levels and consequent protein kinase A (PKA) activation have been shown to reactivate quiescent HSV-1 infections [12,13], we first tested the effect of treatments that cause elevated cAMP in our system. Treatments with forskolin or a cell-permeable dibutyryl cyclic AMP (dbcAMP) in the cell body compartment resulted in increased phosphorylation of PKA substrates, and this effect was blocked by additional treatment with a PKA inhibitor, H89 (Fig 2A). To quantitate escape from silencing, we measured the total fluorescence (mRFP-VP26 capsid protein) in chambered neuronal cultures after performing background subtraction and feature selection using Fiji/ImageJ v.1.48u [14,15] (Fig 2B). When applied to cell bodies during low-MOI axonal infection, both forskolin and dbcAMP promoted PRV 180 productive infection (Fig 2C). The mean value of control dishes, where no red capsid protein expression was observed is shown as the baseline. Virus infection spread throughout the S compartment after escape from silencing in approximately 7 days in the case of dbcAMP or forskolin treatment. We further confirmed that the observed effects of forskolin were due to PKA activation by including the PKA inhibitor H89 in the S compartment. Productive infection was blocked when PKA activity was inhibited (Fig 2C), indicating that forskolin-mediated escape from silencing requires PKA activation.

Forskolin-mediated escape from PRV silencing requires JNK activity

How latent alpha herpesvirus infections reactivate when no viral proteins are produced in the host cell to activate viral transcription is a topic of much research and debate. Recently, Cliffe et al., showed that cJun N-terminal kinase (JNK) activity and c-Jun phosphorylation are essential for HSV-1 reactivation [3]. It was proposed that stress or injury-related stimuli, including phosphatidylinositol 3-kinase (PI3K) inhibition, axotomy, and nerve growth factor (NGF) withdrawal converge on JNK activation and c-Jun phosphorylation for reactivation [3,16]. To understand whether forskolin mediated PRV 180 productive infection (which requires PKA activity) converges on the JNK pathway, we exposed forskolin treated cell bodies to 20 μM of the JNK inhibitor, JNKII, while simultaneously infecting their axons with low MOI PRV 180. In this condition, inhibiting JNK activity prevented the onset of PRV 180 productive infection (Fig 2C). The effect of this inhibitor was confirmed by monitoring the levels of total c-Jun (T-cJun) and phospho c-Jun (P-cJun). JNKII substantially reduced forskolin-stimulated c-Jun accumulation and completely blocked cJun activation (Fig 2D). We also checked c-Jun levels when the PKA inhibitor H89 was included in forskolin treated samples. We detected less c-Jun accumulation, but phosphorylation was not blocked. Also, there was a clear shift in the molecular weight of both total and activated forms of c-Jun. To confirm that chemical inhibitor treatment did not alter viability or virus production capacity of neurons, we monitored neuronal morphology and virus infection under conditions that did not allow PRV productive infection (forskolin+H89 and forskolin+JNKII). After 6 days, forskolin+H89 or forskolin+JNKII treated neurons displayed comparable morphology to untreated neurons, and PRV 180 infection, at an MOI of 5, proceeded comparably in all conditions (Fig 2E). Thus, when no viral components are delivered, PRV 180 escape from silencing occurs via a slow, PKA-dependent activation of the JNK pathway.

Infection of cell bodies with UV treated PRV stimulates rapid escape from silencing

We showed previously that superinfection of cell bodies with UV treated PRV is sufficient to reactivate quiescent PRV genomes [5]. Using our complementation assay, we asked whether virion components delivered to cell bodies can trigger escape from silencing similar to that observed with forskolin treatment. We infected cell bodies in S compartments with a high MOI of UV inactivated PRV 959 (mNeonGreen-VP26, a green fusion protein; UVPRV) [5], while simultaneously infecting axons with PRV 180 at low MOI. We chose to use green capsid PRV to be able to monitor the effect of UV inactivation. UV inactivated virus particles, as we have previously reported, are able to enter, undergo retrograde axonal transport in neurons, and deliver their defective genomes to the nuclei [5]. UVPRV does not exclude superinfection with a second PRV genome [17]. Simultaneous co-infection of cell bodies with high MOI UVPRV enabled efficient escape from silencing by PRV 180, as indicated by red capsid signal in the cell bodies (Fig 3A). Note that there is no green capsid signal in the cell bodies, confirming that the UV inactivated virus is unable to replicate or express late viral genes. Surprisingly, PRV 180 infection spread to all neurons in the S compartment in only 3 days following addition of UVPRV to the cell bodies. This is considerably faster than the 7 day period required for PKA-induced escape from silencing. Entry of UVPRV was required because escape from silencing was not observed with UV inactivated non-complemented entry deficient mutants (UVgBnull or UVgDnull) (Fig 3B). Interestingly, while the PKA inhibitor H89 efficiently blocked the activity of host PKA (Fig 3C), H89 had no effect on UVPRV-mediated escape from silencing (Fig 3B). We also confirmed that PRV 180 or UVPRV infection of cell bodies induces PKA activity early after infection (3 hpi), and this activity is also efficiently inhibited by the addition of H89 1 hpi (Fig 3C). These observations suggest that, upon cell body entry, components of UV treated virions trigger an accelerated escape from silencing through PKA-independent mechanisms.

We also monitored viral DNA replication and expression levels of the major capsid protein (VP5), the most abundant tegument protein (UL47) [18], and the lytic transactivator protein (EP0) in neuronal cultures infected by either PRV 180 alone in N compartments (3 day or 10 day post infection-dpi), UVPRV alone in S compartments (3 dpi), or both of them simultaneously as we did in the complementation assay (3 dpi). We were able to detect the major capsid protein VP5 only when PRV 180 escaped from silencing in the complementation assay (Fig 3D). Tegument protein UL47 was detectable in cell bodies infected with UVPRV but the levels were less than in the co-infected samples. We detected EP0 in UVPRV infected cell bodies, but much less than UL47 in the same samples. EP0 reached high levels during PRV 180 productive infection stimulated by UVPRV complementation (Fig 3D). We quantitated viral DNA amounts in cell bodies in these three conditions. In a silent infection (PRV 180 alone), DNA amounts were almost at the detection limit of our Q-PCR system (CT_mean = 33.76 ± 0.29 SEM at 3 dpi, CT_mean = 34.91 ± 0.08 SEM at 10 dpi). UVPRV infection of cell bodies yielded approximately 8x10³ fold more DNA than PRV 180 alone (CT_mean = 20.44 ± 0.12 SEM), and approximately 12.8 fold less DNA than the complemented samples (CT_mean = 16.76 ± 0.06 SEM) (Fig 3E). These results show that UV treated PRV particles deliver significant amounts of tegument proteins as well as replication deficient viral DNA. Clearly one or both of these components trigger escape from silencing in the complementation assay.

Light particles promote PRV 180 escape from silencing

We hypothesized that PRV quiescence is established after low MOI axonal infection, either due to insufficient tegument proteins delivered to the nuclei to transactivate viral gene expression, or insufficient genomes to saturate the silencing machinery. We were able to test these ideas using light particles (LP). We used cell supernatants from PRV UL25 null mutant infected cells. The UL25 protein is essential for proper genome encapsidation, capsid nuclear egress, and production of infectious progeny virions [19]. This mutant produces almost no infectious virions, but produces abundant virion-like light particles (LP) that contain tegument but do not contain capsids or genomes. We constructed PRV 495, a double recombinant that expresses two fluorescent fusion proteins: gM-pHluorin, a green fluorescent viral glycoprotein that is incorporated into light particles, and mRFP-VP26 (red capsid) [7,20]. The PRV 495 recombinant was prepared by infecting UL25-expressing helper PK15 cells. We prepared a capsid-free LP stock by infecting non-complementing PK15 cells with complemented PRV 495 at an MOI of 5. Varying amounts of this LP stock were run on SDS gels, and we assessed the presence of envelope proteins (gD, Us9), outer tegument (UL47 and EP0) and inner tegument proteins (UL36, Us3) (Fig 4A). Inner tegument proteins UL36 and Us3 were much less than the other tegument components UL47 and EP0. The lack of capsid proteins (VP26 and VP5) was confirmed by WB and by fluorescence microscopy of LP inoculum on axons and adhered to glass coverslips (Fig 4A and 4B). Many gM-pHluorin positive punctae were detected attached to axons at 1 h post-inoculation. We did not detect any dual color puncta on axons or coverslips, confirming that the PRV 495 LP preparation does not contain viral capsids. We noticed some rare single color red punctae that did not colocalize with the green gM signal, which possibly represents fluorescent debris from lysed PRV 495-infected cells (Fig 4B). We then inoculated cell bodies with the LP preparation to determine if LP could promote escape from PRV 180 silencing. PRV 495 LP promoted escape from silencing as fast as the UVPRV (Fig 4C). This effect was not dose dependent using the amounts we tested (10, 20 and 50 μl). Since there were no detectable capsids in our LP preparations, we concluded that viral genomes were not necessary for escape from silencing. To be sure that the LP-mediated escape from silencing is due to LP and not due to other soluble factors released from infected cells, we filtered the LP preparation through a 0.1 μm filter. This filtered LP preparation did not promote escape from silencing (Fig 4C). This indicates that the “escape from silencing” activity is particulate, and secreted viral/host proteins or smaller cellular exosomes are not able to promote escape from silencing. Moreover, similar to what was observed with UVPRV, the PKA inhibitor H89 was not able to block LP-induced escape from silencing (Fig 4C). LP infection of cell bodies induced phosphorylation of PKA targets comparable to what we found with UVPRV, and this induction was blocked by the addition of H89 (Fig 4D). These data confirmed that in the absence of capsids or genomes, viral proteins incorporated into particles are sufficient to induce an accelerated escape from silencing within 3 days, and this escape does not require PKA activity.

PRV tegument proteins Us3 and EP0 are not required for escape from silencing

We hypothesized that a tegument protein delivered by UVPRV or LP might trigger escape from silencing when PKA signaling is blocked. Specifically, we looked at PRV Us3, a potential analog of PKA [9], and PRV EP0, a functional homolog of HSV-1 ICP0, which contains a putative cAMP response element in its promoter [10]. To understand whether the PRV tegument components Us3 or EP0 play a major role in escape from silencing, we prepared stocks of viral mutants that do not express these proteins (Fig 5A). PRV 823 [21] contains a Us3 null mutation, and PRV EPO6 contains an EP0 null mutation (see Materials and Methods). We then performed the complementation assay by infecting cell bodies in the S compartment with UV inactivated EP0 null (UVEPO6) or Us3null (UV823) while simultaneously infecting axons with low MOI PRV180 (Fig 5B). The UV inactivated EPO null mutant promoted escape from silencing with comparable kinetics to UVPRV and LP in a PKA independent manner. Unexpectedly, we also found that the Us3null mutant promoted escape from silencing with or without H89 treatment (Fig 5B). Infection of neurons with either of these mutants induced PKA activity in 3 hours comparable to UVPRV and H89 blocked this activity when added 1 hour after infection with either PRV mutant (Fig 5C).

These data suggest that UVPRV- and LP-mediated escape from silencing involves multiple signaling pathways, which may be distinct from the forskolin- and dbcAMP-mediated PKA pathway that induces a slower escape from silencing. Alternatively, another viral kinase in the tegument, which is resistant to H89, may be able to activate productive infection by phosphorylating PKA targets (e.g. UL13).

We entertained the hypothesis that UVPRV and LP infections stimulated a general cell response to virus infection, which resulted in escape from silencing. We tested replication deficient adenovirus and baculovirus recombinants expressing diffusible GFP to determine if high MOI infection of cell bodies promoted PRV 180 escape from silencing. Although both DNA virions (one naked, one enveloped) efficiently transduced GFP expression in neuronal cell bodies, neither promoted PRV 180 escape from silencing (Fig 5B). We conclude that escape from silencing by superinfection with UVPRV or LP is not a generalized neuronal response to DNA virus infection. Rather, it is a herpesvirus-specific induction of productive cycle viral promoters by multiple tegument-host protein interactions or induction of injury response pathways.

UV inactivated PRV- and LP-mediated escape from silencing does not depend on JNK activity

Our findings that UVPRV and LP mediate an accelerated, PKA-independent escape from silencing support the hypothesis that the molecular mechanism(s) promoting quiescence or productive infection might differ from the mechanisms of de-repressing viral promoters during reactivation from latency. To test whether escape from silencing mediated by UVPRV or LP is regulated by JNK activity, we infected axons with PRV 180, simultaneously treated cell bodies with UVPRV or LP, and subsequently added a JNK inhibitor (JNKII) one hour post infection. When JNK activity was inhibited, neither UVPRV- nor LP-mediated escape from silencing was blocked, but the spread of infection among the S compartment neurons was delayed (Fig 6A and 6C). We observed red capsid accumulation starting at 3 dpi, but the full spread of infection took 6 days (Fig 6C). We also treated UVPRV or LP infected cell bodies with the PI3K inhibitor, LY294002. In this case, the onset and spread of PRV 180 productive infection was significantly delayed but not completely blocked (Fig 6A and 6C). Interestingly, we saw “strings” of infected neurons in S compartments (Fig 6C showing limited spread of infection), but full spread of infection was prevented particularly when PI3K activity was inhibited. These findings are consistent with the literature indicating a role for this kinase in herpesvirus infection (see [22] for review). The activity of the inhibitors also was tested in mock, UVPRV and LP treated neurons. Interestingly, treatment of neurons with UVPRV or LP increased total and phosphorylated c-Jun levels comparable to forskolin. JNKII blocked c-Jun phosphorylation in all of these samples (Fig 6B), but it interfered only with forskolin-mediated and not with UVPRV- or LP-induced escape from silencing. LY294002 effectively reduced activated AKT levels (pAKT) as expected, and induced c-Jun accumulation and phosphorylation in all of the conditions where the inhibitor was added (Fig 6B). These results further established that UVPRV- and LP-mediated escape from silencing proceeds through a distinct molecular mechanism from cAMP/PKA and JNK-mediated activation of productive infection.

Cell lines and viruses

Porcine kidney epithelial cells (PK15; ATCC) were used to produce and titer PRV. These cells were maintained in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum, 1% penicillin and streptomycin. PRV Becker is a wild-type laboratory strain [43]. PRV 180 expresses an mRFP-VP26 fusion protein in a PRV-Becker background [44]. PRV 959 expresses an mNeonGreen-VP26 fusion protein in a PRV Becker background [5]. PRV 233 is a gB-null mutant of PRV-Becker expressing diffusible GFP from the gG locus under the control of a cytomegalovirus (CMV) promoter [11]. PRV 233 stocks were propagated, and their titers were determined in PK15 cells stably transfected with LP64e3, a plasmid expressing PRV gB (constructed by Lisa Pomeranz in the Enquist lab). PRV 357 is a gD-null recombinant expressing diffusible GFP in a PRV Becker background [45]. PRV 357 is grown G5 cells that constitutively express PRV gD [46]. To obtain entry deficient virus stocks, non-complementing PK15 cells were infected with PRV 233 and PRV 357 at high MOI, and harvested at 24 hpi to obtain gB null and gD null virions, respectively. The titers of these stocks were determined based on the genomic DNA content by quantitative RT-PCR. Mutant stock titers were normalized using volumes of each stock corresponding to the amount of DNA in 10⁶ pfu of wt PRV. PRV 823 is Us3 null of PRV-Becker expressing mRFP-VP26 fusion protein [21] and PRV EP06 is EP0null expressing mNeonGreen-VP26 in a PRV-Becker background (constructed by Hao Oliver Huang in the Enquist lab). PRV 495 is UL25-null expressing mRFP-VP26 and gM-pHluorin fusion proteins [20]. Complementing cell line expressing PRV UL25 protein was a kind gift from G. Smith, Northwestern University [47].

Primary neuron cultures

Superior cervical ganglia (SCGs) were isolated from embryonic day 17 Sprague-Dawley rat embryos (Hilltop Labs). Primary neurons were cultured in tri-chamber dishes as previously described [48,49]. Multi-well or 35-mm plastic tissue culture dishes (Falcon) or optical plastic dishes (Ibidi) were coated with 500 μg/ml of poly-DL-ornithine (Sigma Aldrich) and 10 μg/ml of natural mouse laminin (Invitrogen). After coating, two sets of grooves were etched on the dishes. A silicone grease-coated tri-chamber (Tyler Research) was placed on top of a drop of 1% methylcellulose (in neuronal medium) covering the groves. Ganglia were trypsinized and triturated, and approximately 2/3 of an SCG was plated in the S compartment. Neurons were maintained in complete neuronal medium: Neurobasal medium (Gibco) supplemented with 100 ng/ml nerve growth factor 2.5S (Invitrogen), 2% B27 (Gibco) and 1% penicillin and streptomycin with 2 mM glutamine (Invitrogen). 2 to 3 days after seeding, 1 mM cytosine-D-arabinofuranoside (AraC; Sigma-Aldrich) was added for at least 2 days to eliminate non-neuronal cells. Neurons were cultured for 14–21 days prior to experiments.

Virus infection and drug treatment in compartmented neuronal cultures

Before adding the virus inoculum to N and/or S compartments in tri-chambers, 1% methylcellulose prepared in neuronal medium was added in the M (Middle) compartment to prevent any possible leakage from either of these compartments. Neuronal infections were done using the indicated amounts of recombinant PRV. DiI was added at 2.5 μg/mL (Life Technologies) to the N compartment. Reagents used in this study: Forskolin (Sigma), dibutyryl cAMP (Selleckchem), H89 (Selleckchem), LY294002 (Sigma) and JNK inhibitor II (Calbiochem).

Microscopy

Imaging was performed on a previously described Nikon Ti-E inverted epifluorescence microscope [50]. Tiled images of the entire S compartment were captured using Nikon NIS Elements software, a Cool Snap ES2 camera (Photometrics), and 4x magnification objective. To measure the total fluorescence in chambered neuronal cultures, we performed background subtraction and feature selection using Fiji/ImageJ v.1.48u [14]. Because the background fluorescence intensity is highly variable across tiled images, we first subtracted the local background by calculating a 20 pixel radius median filter and subtracting this median filtered image from the original. To select fluorescent cell bodies, we next performed granulometric filtering using the GranFilter plugin in Fiji/ImageJ [15]. GranFilter plugin settings were as follows: circular structure element, 3 pixel radius, 6 pixel step size. With these settings GranFilter selects fluorescent structures that are the approximate size and shape of fluorescent neuronal cell bodies. We then manually set a region of interest to exclude image artifacts caused by reflections off the chamber walls and grooves. Finally, we measured the integrated fluorescence intensity using the Measure function in Fiji/ImageJ and exported these measurements to Microsoft Excel. High magnification imaging of fluorescent particles was done using an Andor iXon Ultra EMCCD camera and 60x magnification objective. All images and movies were assembled for publication using NIS-Elements software (Nikon), Fiji/ImageJ, and Adobe Photoshop. For comparative analysis, fluorescence excitation intensity, exposure time, and other imaging parameters were consistent for all experimental conditions.

Western blot analysis

Dissociated neurons or cell bodies from the S compartment were lysed in radioimmunoprecipitation assay (RIPA) buffer supplemented with 1 mM dithiothreitol (DTT) and protease inhibitor cocktail (Sigma-Aldrich). Lysates were incubated on ice for 30 min, sonicated, and centrifuged at 11,000 rpm at 4°C. Supernatants were transferred into new tubes and mixed with 5x Laemmli buffer. Samples were heated at 90°C for 5 min before resolved by 4–12% NuPAGE BisTris gels (Invitrogen). Proteins were transferred to nitrocellulose membranes (Whatman) using semidry transfer (Biorad). For blocking, membranes were incubated in 5% non-fat dry milk in phosphate-buffered saline (PBS) solution for 1 h at room temperature. Immunoblots were performed using primary and secondary antibodies diluted in 1% milk PBS solution. Membranes were incubated with chemiluminescent substrates (Supersignal West Pico or Dura, Thermo scientific). Protein bands were visualized by exposure on HyBlot CL (Denville scientific) blue X-ray films. Primary antibodies used in this study: Mouse monoclonal antibody (mAb) anti-β-actin (Sigma), anti-Us9 mouse mAb IA8 clone [51], anti-VP5 mouse mAb (gift of H. J. Rziha, Federal Research Center for Viruses Diseases for Animals, Tubingen, Germany), anti-UL47, anti-UL35 and anti-UL36 polyclonal rabbit sera [gifts of T. Mettenleiter, Friedrich-Loeffler Institut [52–54]], anti-Us3 mouse mAb [21], rabbit polyclonal anti-EP0 (gift of H. Kida, Hokkaido University) [55], anti-gD polyclonal rabbit sera (gift of K. Bienkowska-Szewczyk University of Gdansk), anti-cJun (Cell Signaling), anti-phospho-cJun (Ser63; Cell Signaling) anti-AKT (Cell Signaling), anti-phospho-AKT (Ser473; Cell Signaling), anti-PKA substrates (P-S/T Kinase substrate Ab Sampler, Cell Signaling). Horseradish peroxidase-conjugated secondary mouse or rabbit antibodies (KPL) were used at 1:10000 dilution.

Quantitative RT-PCR

Q-RT-PCR was performed with Eppendorf Realplex Mastercycler. Reaction mixture was prepared using Kapa Syber Fast qPCR kit and samples were prepared as triplicates. Each experiment was done in duplicates. To determine genomic DNA amounts in neuronal soma, each S compartment was lysed in 10 μl of RIPA. 5 μl of this lysate was treated with proteinase K (New England Biolabs) for 50 min. at 55°C followed by 5 min at 95°C. To titer entry deficient PRV mutants (gBnull and gDnull), 90 μl of virus stock was first digested with 100 units of DNase I (Invitrogen) to remove contaminating DNA before proteinase K treatment. Viral genomic DNA was quantified by using UL54 specific primers as published [56]. PRV Becker nucleocapsid DNA and PRV Becker virus stock (1x10⁸pfu/ml) were used as standards to determine the amount of viral DNA corresponding to one plaque forming unit. To determine transcript amounts, total RNA was extracted from cell bodies in the S compartment using RNeasy Plus Mini kit (Qiagen). cDNA synthesis was performed using SuperScript III first-strand synthesis kit (Life Technologies) with Oligo-dT primers. LAT specific primers (fw: 5`-GATGCAGTCCAGACAG-3`, rev: 5`-GTAGTGGTCCCGAGTTGC-3`) amplifying a 141 bp fragment were used. EP0 transcript was quantitated using primers (fw: 5`-GGGTGTGAACTATATCGACACGTC-3`, rev: 5`-TCAGAGTCAGAGTGTGCCTCG-3`) to amplify a 150 bp region. Plotted values were calculated using the –ΔΔCt method and normalization to 28S rRNA (fw: 5`-GGGCCGAAACGATCTCAACC-3`, rev: 5`-GCCGGGCTTCTTACCCATT-3`). Fold changes were calculated using the comparative CT method (2^(C_TX^-C_TR⁾_control-^(C_TX^-C_TR⁾_test) where C_TX is the threshold cycle of the gene of interest and C_TR is the threshold cycle of the reference gene (28S rRNA) [57].

Statistical analysis

One-way analysis of variance (ANOVA) with Tukey’s posttest or Mann-Whitney test was performed using GraphPad Prism 5.0. Values in the text, graphs, and figure legends throughout the manuscript are means ± standard errors of the means (SEM).

Ethics statement

All animal work was performed in accordance with the Princeton Institutional Animal Care and Use Committee (protocols 1947–16). Princeton personnel are required to adhere to applicable federal, state, local and institutional laws and policies governing animal research, including the Animal Welfare Act and Regulations (AWA); the Public Health Service Policy on Humane Care and Use of Laboratory Animals; the Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research and Training; and the Health Research Extension Act of 1985.