Significance. 1 Present address: Sino-French Hoffman Institute of Immunology, Guangzhou Medical

Transcription

1 Role of activating transcription factor 3 in the synthesis of latency-associated transcript and maintenance of herpes simplex virus 1 in latent state in ganglia Minfeng Shu, Te Du, Grace Zhou 1, and Bernard Roizman 2 Marjorie B. Kovler Viral Oncology Laboratories, The University of Chicago, Chicago IL Contributed by Bernard Roizman, August 4, 2015 (sent for review July 17, 2015) A key property of herpes simplex viruses (HSVs) is their ability to establish latent infection in sensory or autonomic ganglia and to reactivate on physical, hormonal, or emotional stress. In latently infected ganglia, HSVs express a long noncoding RNA, a latencyassociated transcript (LAT), which plays a key role in maintaining latently infected neurons, but not viral proteins. To investigate the events leading to reactivation, we examined the use of ganglionic organ cultures that enable rapid reactivation in medium containing antibody to nerve growth factor (NGF) or delayed reactivation in medium containing NGF and epidermal growth factor (EGF). Here we report the discovery that activating transcription factor 3 (ATF3), a stress response protein, profoundly affects the interaction of HSV with its host. Specifically, (i) ATF3 is induced by stress, such as inhibition of protein synthesis or infection; (ii) in infected cells, ATF3 enhances the accumulation of LAT by acting on the response elements in the promoter of the LAT precursor RNA; (iii) ATF3 is induced nearly 100-fold in ganglionic organ cultures; and (iv) ATF3 plays a key role in the maintenance of the latent state, inasmuch as expression of ATF3 bereft of the C-terminal activation domain acts as a dominant negative factor, inducing HSV gene expression in ganglionic organ cultures harboring latent virus and incubated in medium containing NGF and EGF. Thus, ATF3 is a component of a cluster of cellular proteins that together with LAT maintain the integrity of the neurons harboring latent virus. stress response adenylate-uridylate rich mrna reactivation Following infection at a portal of entry, usually the mouth or genitals, herpes simplex virus (HSV)-1 or HSV-2 infect sensory nerve endings and are transported retrograde to the neuronal nucleus, where they establish a silent or latent infection (1). Characteristically, during the latent state, neurons accumulate a 1.5- or 2-kb noncoding RNA known as latency-associated transcript (LAT) (2 5). LAT is a stable intron derived from a much longer precursor RNA (plat) (3, 6). In a fraction of neurons infected with HSV-1, the virus reactivates and is transported anterograde to a site at or near the portal of entry, where it replicates and becomes available for transmission to another individual by physical tissue contact (7, 8). Knowledge is scant regarding the mechanisms by which HSV-1, a potentially virulent virus, can remain silent in neurons for the life of the host or reactivate from latency. The model organism closest to the human is the mouse, a model system that does not lend itself readily to molecular analyses of events occurring in the sensory ganglia. At the other extreme, neuronal cultures harboring HSV-1 silenced by drugs might not reflect the natural mechanisms by which the virus is silenced in ganglia. Moreover, in ganglia, sensory neurons are coated by satellite cells that affect the physiology of the neuron. In humans, mice, and rabbits, reactivation follows stimulation or a reversible or irreversible injury (1). In an attempt to probe the events occurring after injury, we developed a ganglionic organ culture model. In this model, trigeminal ganglia are harvested at 30 d after mouse corneal inoculation and incubated in medium containing anti-nerve growth factor (NGF) antibody to rapidly deplete NGF or in medium containing both NGF and epithelial growth factor (EGF) (9). In the former medium, expression of viral genes begins at 5 h after excision of the ganglia. In medium containing NGF and EGF, the virus remains silent for h. Studies on this model revealed that in ganglia maintained in medium containing NGF plus EGF, the virus is maintained in a dynamic equilibrium that is disrupted by inhibitors of HDAC-1 and -4, the PI3K-AKT-mTOR and STAT3 pathways, or proapoptotic drugs (10, 11). Reactivation in medium containing anti-ngf antibody is inhibited by the expression of REST or by a p300/cbp inhibitor (11, 12). Perhaps more significant for reactivation, in neurons all genes are expressed at once rather than sequentially in a coordinated fashion as is seen in infected cells at the portal of entry. The present study emanated from an attempt to identify the neuronal genes whose expression maintains the virus in a latent state. In this study, we compared by microarray analyses RNA transcripts in freshly excised ganglia with those of ganglia incubated for 5 h or 11 h in medium containing anti-ngf antibody and medium containing NGF plus EGF. The cellular gene that attracted our attention was that of ATF3, because its mrna increased nearly 100-fold both in ganglia incubated in medium containing anti-ngf and in medium containing NGF plus EGF. We found that ATF3 enhances the accumulation of LAT in infected cells in culture and plays a role in the maintenance of HSV-1 in a latent state. ATF3 is a member of the ATF/CREB family of transcriptional factors (13, 14). The current consensus holds that ATF3 is an immediate early-response gene, a regulator of stress response, and a factor in determining cell fate (15). ATF3 binds to a camp response element (CRE) with the canonical sequence 5 -TGAC- GTCA-3 in the form of homodimers or heterodimers with other members of the ATF/CREB family. HSV-1 DNA contains eight Significance A major unresolved issue confronting infectious diseases is the mechanism by which virulent, potentially lethal viruses remain silent (latent) in selected cells in the human body until induced to replicate and spread in response to stress. The establishment of latent infection by herpes simplex viruses in sensory or autonomic neurons is a common event in a large fraction of the human population. Here we report that activating transcription factor 3 (ATF3) is induced by stress both in infected cells in culture and in ganglia harboring latent virus. The function of ATF3 is to block the reactivation of virus induced by neuronal stress. Author contributions: M.S., T.D., G.Z., and B.R. designed research; M.S. and T.D. performed research; M.S., T.D., and B.R. analyzed data; and B.R. wrote the paper. The authors declare no conflict of interest. See Commentary on page Present address: Sino-French Hoffman Institute of Immunology, Guangzhou Medical University, Guangzhou , China. 2 To whom correspondence should be addressed. bernard.roizman@bsd.uchicago. edu. E5420 E5426 PNAS Published online August 24,

2 Fig. 1. HSV-1(F) induced ATF3 mrna is independent of new protein synthesis. (A) ATF3 mrna is induced by HSV-1(F) infection. Cells (HEK293, Vero, and Hep2) were infected with HSV-1(F) at a multiplicity of infection (MOI) of 10, and then harvested at the indicated time points. ATF3 mrna was detected by Northern blot analysis. (B) Effect of cycloheximide (CHX; 100 μg/ml) on HSV-1(F) induced ATF3 mrna level. Cells (HEK293, Vero, and Hep2) were infected with HSV-1(F) at an MOI of 10 in the absence or presence of CHX. 18s and 28s rrnas served as loading controls. CRE sites, but only two of these are in a promoter, specifically in the promoter 1 of LAT (LP1) (16). Mutagenesis of the CRE1 site (5 -CTGCGTCA-3 ) proximal to the plat TATA box results in diminished levels of LAT and reduced reactivation rates in a rabbit model (17, 18). Those results are consistent with results of other studies showing that LAT protects neurons from apoptosis (19, 20). Finally, the 181-residue protein has been shown to consist of five domains: an N-terminal activation domain, a repressor domain, a basic domain, a ZIP domain essential for dimerization, and a C-terminal activation domain (21, 22). Like many stress response genes, its mrna contains adenylate-uridylate (AU)-rich elements in the 3 UTR (23). Curiously, ATF3 is not subject to degradation by the viral host shutoff RNase. The other AU-rich RNAs spared from degradation play significant roles in the biology of HSV (24, 25). Results ATF3 Is Induced in Untreated HSV-1 Infected Cells and in Infected Cells Treated with Cycloheximide (100 μg/ml) at the Time of Infection. Replicate cultures of HEK293T, Vero, and HEp-2 cells were exposed to 10 PFU of HSV-1(F) per cell alone (Fig. 1A) orinthe presence of cycloheximide (Fig. 1B). The cells were harvested at the time of infection (0 h) and at the times shown in Fig. 1. The salient features of the results can be summarized as follows. All infected cells showed accumulation of ATF3 mrna. The patterns of accumulation varied. We noted a huge increase in ATF3 mrna in infected HEK 293T and HEp-2 cells, but a relatively modest accumulation in infected Vero cells. The presence of high molecular weight bands in HEK 293T and HEp-2 cells late in the course of infection suggested accumulation of unspliced ATF3 RNAs. In addition, the amounts of ATF3 mrna were higher in cells exposed to cycloheximide late in infection. These data are consistent with the conclusion that activation of ATF3 in infected cells does not require previous viral protein synthesis. The apparent increase in the amount of ATF3 mrnas could reflect either a more intense stress response to the presence of cycloheximide or inhibition of ATF3 mrna accumulation by a viral gene product produced relatively late after infection in untreated cells. Accumulation of LAT Is Enhanced in Cells Expressing Wild Type ATF3 or ATF3 Mutants Containing the C-Terminal Activation Domain. To follow up preliminary studies showing that ATF3 enhances the expression of LAT (26), we constructed a series of truncated ATF3 mutants, shown schematically in Fig. 2A. As noted in the introductory section, ATF3 consists of an N-terminal activation domain (A), followed by a repressor domain (R), a basic domain (B), a ZIP domain (Z), and a second C-terminal activation domain (A) (Fig. 2A). In these experiments, replicate cultures of HEK293T cells were transfected with 1 μg of pcdna or plasmids encoding intact or mutant ATF3 tagged at the N terminus with Flag. The cells were exposed to 1 PFU of HSV-1(F) per cell at 40 h after transfection and harvested at 20 h after infection. Northern blots of extracted LAT are shown in Fig. 2B. LAT was quantified by phosphoimaging (Fig. 2E) or by quantitative RT- PCR (qrt-pcr) (Fig. 2F). The amounts of wild type (WT) and mutant ATF3 accumulating in transfected cells are shown in Fig. 2D. The results show that the accumulation of LAT is enhanced by ATF3, and that the C-terminal activation domain is essential for the observed effect. ATF3 Binds to CRE1 and CRE2 Sites in the Promoter of LAT Precursor RNA. As noted above, only two of the eight CRE sites in HSV-1 DNA are contained in a promoter, specifically in the promoter of the LAT precursor RNA (LP1). The sequences of CRE1 (5 -CTGCGTCA-3 )andcre2(5 -TTACATCA-3 )differfromthe sequence of canonical CRE (5 -TGACGTCA-3 ). Previous studies have shown that mutagenesis of the CRE1 site results in decreased accumulation of LAT concurrent with decreased reactivation of HSV-1 in the rabbit model (17). To determine whether ATF3 binds to CRE1 or CRE2, 32 P-labeled oligonucleotides expressing CRE1 or CRE2 were reacted with nuclear extracts from cells transfected with plasmids encoding WT N-terminal Flag-tagged ATF3, M2, or M3. As described in detail in Materials and Methods, the mixtures were subjected to electrophoresis in nondenaturing gels,transferredtoacellulosesheet,andsubjectedtoautoradiography. To ensure that the CRE probes bound WT or mutant ATF3, the reaction mixture contained a competitor oligonucleotide or antibody to Flag. In mixtures in which the oligonucleotide containing the CRE1 or CRE2 sequence, ATF3, M2, or M3 antibody bound to the Flag tag retarded the electrophoretic mobility of the complex. Fig. 3C shows that transfected cells produced N-terminal Flag-tagged ATF3, M2, or M3 proteins. Fig. 3 A and B show the results obtained on CRE1 and CRE2, respectively. The results indicate that WT ATF3 bound both CRE1 (Fig. 3A,lane5) and CRE2 (Fig. 3B, lane 5), but neither M2 nor M3 bound CRE1 or CRE2. MICROBIOLOGY SEE COMMENTARY PNAS PLUS Shu et al. PNAS Published online August 24, 2015 E5421

3 A B C D E F were harvested at 2, 12, 24, or 48 h after exposure to 0.1 PFU of virus per cell. The accumulation of ATF3 in the transfected cells at the time of infection is shown in Fig. 5C. The results of titrations shown in Fig. 5B indicate that WT ATF3, M2, or M3 had no effect on viral replication. M2, but Not M3 or WT ATF3, Induced the Reactivation of Recombinant Virus in Ganglia That Were Excised and Incubated in Medium Containing NGFandEGFat30dAfterInfection.We constructed a series of recombinant viruses carrying WT ATF3, M2, or M3 driven by the cytomegalovirus (CMV) immediate early promoter and inserted between U L 3andU L 4 using procedures described elsewhere (27). Trigeminal ganglia of mice infected by corneal scarification were harvested at 30 d after infection and either analyzed immediately for viral gene expression or incubated for 24 h in medium containing NGF plus EGF or anti-ngf antibody. The harvested ganglia in groups of six were analyzed for the amounts of mrnas encoding representative α (ICP27), β (thymidine kinase), γ 1 (VP16), or γ 2 (U L 41) genes. The results, shown in Fig. 6, indicate that the recombinant virus encoding the M2 ATF3 mutant induced the expression of all representative viral gene groups to levels indistinguishable from those of ganglia incubated in medium containing anti-ngf antibody. The increases in the amounts of viral mrnas in ganglia harboring latent WT virus or recombinant viruses carrying WT ATF3 or M3 were significantly lower. Fig. 2. Effect of different ATF3 truncations on LAT expression. (A) Schematic diagram of different ATF3 truncations. A, activation domain; B, basic domain; R, repression domain; Z, ZIP domain. Both the basic domain and ZIP domain are required for binding to DNA. The truncations were named serially from M1 to M6. (B) Northern blot analysis showing the effect of transfection of ATF3 truncations on LAT (2.0-kb intron) expression. (C) The 18s and 28s rrnas served as loading controls. (D) Protein expression of different ATF3 truncations. (E) Intensities of the bands of various truncations transfection in B were quantified by phosphoimaging analysis. The value obtained was normalized with respect to 18s rrna and plotted relative to the RNA levels present in the pcdna-transfected cells. (F) Total RNA was extracted, reverse-transcribed, and used for quantification of LAT expression by qrt-pcr. The results are expressed as fold change compared with the RNA level present in the cells of pcdna transfection. Data represent mean ± SD. **P < 0.01; ns, not significant. WT ATF3 Interacts with ATF3, but Not with M2 or M3 Mutants. As noted above, ATF3 interacts with its response elements by forming homodimers or heterodimers with itself or other ATF/CREB proteins. The central question that we posed is whether WT ATF3 forms heterodimers with M2 lacking only the C-terminal activation domain or with M3, which lacks both the ZIP domain and the C-terminal activation domain. In these experiments, HEK 293T cells were transfected with plasmids encoding ATF3, M2, or M3 tagged at the N terminus with Flag. Soluble cell extracts prepared at 48 h after transfection (Fig. 4 A and B, lanes 1 4) were reacted with GST or GST-ATF3 bound to agarose beads (Fig. 4C). The proteins bound to the beads were solubilized, subjected to electrophoresis in denaturing gels, transferred to a nitrocellulose sheet, and reactedwithponceaustain(fig.4b) and then with antibody to Flag (Fig. 4A). The presence of ATF3 on the agarose beads is shown in Fig. 4C, lanes 3 and 4. The pull-down assays shows that the tagged ATF3 pulls down itself, but not M2 or M3 (Fig. 4A, lanes 8 13). Overexpression of ATF3, M2, or M3 Has No Effect on the Replication of HSV-1(F). We next asked whether overexpression of WT ATF3, M2, or M3 acting as a dominant negative factor affects virus replication. As illustrated in Fig. 5A, at 24 h before infection, replicate cultures of HEK 293T cells were transfected with 1 μg of plasmids encoding pcdna, WT-ATF3, M2, or M3. The cells WT and Recombinant Viruses Induce the Expression of ATF3 on Excision and Incubation in Medium Containing Anti-NGF Antibody and Medium Containing NGF and EGF. In these experiments, we used qrt-pcr to quantify ATF3 mrna in the extracts of the ganglia described above, that is, ganglia harboring latent WT or recombinant viruses encoding ATF3, M2, or M3 and prepared either immediately after excision or after incubation in medium containing anti-ngf antibody or medium containing both NGF and EGF. The results, shown in Fig. 7, indicate that ATF3 was induced and increased 100-fold after excision and incubation in medium containing anti-ngf antibody as well as in medium containing NGF plus EGF. Discussion The finding that prompted the studies described in this report was the observation that ATF3 is induced by nearly 100-fold in murine ganglia harboring latent HSV-1 and maintained in either medium containing anti-ngf antibody or medium containing NGF plus EGF (Fig. 7). The salient features of the results presented in this report are as follows. ATF3 is induced in cells productively infected with HSV-1. The HSV-1 DNA sequence contains eight CREs, of which only two are located in a promoter region, namely in LP1. Our results show that ATF3 binds to both CRE sites in LP1, forms a homodimer, and fosters increased accumulation of LAT. As noted above, ATF3 consists of several discrete domains. ATF3 mutant M2 bereft of the C-terminal activation domain and M3 mutant lacking both the ZIP and the C-terminal activation domains did not enhance LAT synthesis, bind to the CRE sites in LP1, or interact with WT ATF3. Neither WT ATF3 nor the mutants tested in this study affected the replication of HSV-1 in productively infected cells. A very different picture emerged from the studies on ganglia harboring latent viruses expressing WT ATF3, M2, or M3. The viruses carrying WT ATF3 or M3 could not be differentiated from HSV-1(F) with respect to viral gene expression on excision of the ganglia, on incubation in medium containing anti-ngf antibody, or in medium containing both NGF and EGF. In contrast, viral gene expression was activated in ganglia harboring the mutant virus M2 in medium containing either anti-ngf antibody or both NGF and EGF. E Shu et al.

4 A B C Fig. 3. ATF3 directly binds to the LAT promoter at both CRE1 and CRE2 sites. (A and B) EMSA and supershift assay showing that ATF3 interacts directly with both CRE1 (A) and CRE2 (B) sites. HEK 293T cells were transfected with 1 μg of pn-flag-atf3, pn-flag-m2, or pn-flag-m3 plasmid. The cells were collected at 48 h after transfection and fractionated into nuclear, membrane, and cytoplasmic fractions as described in Materials and Methods. Then 20 μg of nuclear protein extract was reacted with 32 P end-labeled DNA with the CRE, CRE1, or CRE2 consensus binding site. The gel supershift bands showing the specific binding of ATF3 to CRE1 or CRE2 were manifested in the presence of anti-flag antibody and indicated by a solid arrow. (C) Nuclear extracted ATF3, M2, and M3 protein expression from transfected cells. The results presented in this report suggest the following conclusions. ATF3 is expressed in response to stress and carries AU-rich elements in its 3 UTR characteristic of mrnas encoding stress response proteins (23). Activation of stress response genes is a common consequence of infection (28). Furthermore, the mrnas encoding stress response proteins are a primary target of VHS- RNase encoded by the U L 41 gene of HSV-1 (24, 25, 28, 29). The presence of two CREs in LP1 suggests that these sites have evolved to respond to cellular stress and raises the question of whether the synthesis of LAT is a stress response. Moreover, in natural infection, HSV-1 reactivates as a consequence of stress induced by physical injury to neurons [e.g., UV light exposure (30, 31)] or cytokines [e.g., IL6 (32, 33)] induced by hormonal or emotional stress (34, 35). In the model system used in these studies, the primary stress is excision of ganglia, and the induction of ATF3 appears to be at least one of the induced responses to stress. The observation that M2 induces viral gene expression in ganglia harboring latent virus and maintained in medium containing NGF plus EGF suggests that M2 acts as a dominant negative factor, possibly by interacting with a partner other than ATF3. As noted in the introductory section, ATF3 can form heterodimers with other members of ATF/CREB proteins (36, 37). M2 contains the ZIP domain required for the formation of homodimers or heterodimers, but it does not interact with WT ATF3. A relevant conclusion based on our results is that ATF3 plays a role in maintaining the integrity of the affected neuron and concurrently maintains viral DNA in a latent state. It is appropriate to place ATF3 in the context of other factors that play roles in maintaining HSV in a latent state. These known factors fall into one of three categories: constitutively active, activated on stress, or inducible on stress (Fig. 8). All share a common goal: maintenance of neuronal integrity. The constitutively active factors have been known for some time. Current consensus holds that the function of LAT and, most likely, of the mirnas is to maintain the integrity of the neuron to maintain the virus in a silent state (38, 39). In the absence of LAT, neurons harboring latent virus succumb to apoptosis (31). NGF is similarly required for the maintenance of neurons; withdrawal of NGF results in apoptosis (40, 41). In neurons, STAT3 resides in the cytoplasm and is translocated to the nucleus on neuronal stress (42). The function of STAT3 is to block neuronal death (43). Much of the information regarding the role of STAT3 is based on studies showing that inhibition of STAT3, PI3K, Akt, or mtor, or interference with the phosphorylation of tyrosine 705, results in activation of virus (44). ATF3 is the sole stress-inducible factor currently associated with suppression of activation of latent virus. Depending on the system, ATF3 acts to promote viability or death of the cell in which it is induced. In recorded instances, ATF3 acts with STAT3 to induce gene expression (36). The effect on LAT expression and the A B Fig. 4. ATF3 forms homodimerization, but M2 and M3 fail to dimerize with ATF3. (A) Dimerization was detected by GST pull-down assay. HEK-293T cells were transiently transfected with the indicated plasmids encoding pn-flag tagged full-length ATF3 (residues 1 181), M2 (residues 1 135), or M3 (residues 1 118). Soluble cell extracts were prepared at 48 h after transfection and reacted with 10 μl of glutathione-agarose beads bound to GST (lanes 6, 8, 10, and 12) or 80 μl of GST-ATF3 (lanes 7, 9, 11, and 13). (B) The proteins bound to the beads were solubilized and subjected to electrophoresis on a denaturing polyacrylamide gel, transferred to a nitrocellulose membrane, and stained with Ponceau stain, followed by reaction with antibody against Flag. Immunoreactivities of intact ATF3, M2, and M3 are shown in lanes 2, 3, and 4 indicated by arrows. The amounts loaded in each lane represent 10% (vol/vol) of the wholecell lysates. pcdna, pcdna 3.1(+). (C) Ponceau staining showing the GST and GST-ATF3 proteins bound to the beads used in the pull-down assay. Note that the amounts of beads in the reaction mixtures were determined on the basis of GST- ATF3 proteins bound to the beads. Shown are the amounts of GST and GST-ATF3 proteins solubilized from the beads subjected to electrophoresis in a denaturing gel and stained with Coomassie blue. C MICROBIOLOGY SEE COMMENTARY PNAS PLUS Shu et al. PNAS Published online August 24, 2015 E5423

5 Materials and Methods Cells and Viruses. HEp-2, Vero, and HEK 293T cell lines were obtained from American Type Culture Collection and cultured. The limited-passage HSV-1(F) is the prototype HSV-1 strain used in this laboratory (45). Antibodies. Mouse monoclonal antibodies against Flag and β-actin were purchased from Sigma-Aldrich. The antibodies were used at a dilution of 1:1,000. Plasmid Construction. The WT ATF3 (1 181) and truncated ATF3 (1 148), ATF3 (1 135), ATF3 (1 118), ATF3 (40 181), ATF3 (40 148), and ATF3 (1 39, ) fragments were obtained by PCR using the following primers: ATF3 (1 181)- N-Flag, forward: 5 -CGCGGATCCATGGATTACAAGGATGACGATGACAAGATGC- TTCAACACCCAGGCCAGGTC-3 ; ATF3 (1 181), reverse: 5 -CCGGAATTCTTAGC- TCTGCAATGTTCCTTC-3 ; ATF3(1 148), reverse: 5 -CCGGAATTCTTAAAGGTTGA- GCATGTATATCAAATGCTGC-3 ; ATF3 (1 135), reverse: 5 -CCGGAATTCTTA- GAGCTCCTCAATCTGAGCCTTCAGTTC-3 ; ATF3 (1 118), reverse: 5 -CCGGAAT- TCTTACGACTCTTTCTGCAGGCACTCCGTCTTC-3 ; ATF3 (40 181)-N-Flag, forward: Fig. 5. ATF3, M2, and M3 transient transfections have no effect on HSV-1(F) growth in vitro. (A) Schematic diagram of the experimental design. In brief, HEK293 T cells were transfected with 1 μg of pcdna, pn-flag tagged ATF3, M2, or M3 plasmids for 24 h. Cells were then exposed to HSV-1(F) at a multiplicity of 0.1 PFU per cell. The inocula were replaced at 2 h after exposure to virus. Cells were harvested at 2, 12, 24, and 48 h after infection. (B) Viral progeny were titrated on Vero cells. (C) Protein expression of transfected ATF3, M2, and M3. β-actin served as a loading control. high levels of expression in excised ganglia suggest that in the HSV- 1 infected cells, ATF3 functions primarily to maintain the integrity of neurons. Finally, it is noteworthy that each of the factors associated with the maintenance of HSV in a latent state is associated with blocking neuronal death. Conversely, deprivation of NGF, absence of LAT, or inhibition of STAT3 can cause virus reactivation and apoptosis. Is activation of apoptosis compatible with reactivation? There are no definite studies on this issue. The only available data are based on studies of ganglionic organ cultures reported elsewhere (10). In brief, exposure of ganglia harboring latent virus in medium containing NGF and EGF and two different proapoptotic drugs resulted in viral gene expression and viral DNA synthesis at levels compatible with those attained in medium containing anti-ngf antibody. In essence, the data indicate that proapoptotic stimuli could induce virus reactivation and may be the principal mechanism by which neuronal injury causes reactivation of latent virus. Fig. 6. Reactivation of HSV-1(F), recombinant viruses of HSV-1 expressing WT ATF3, M2, or M3. At 30 d postinoculation, TG excised from F, recombinant viruses expressing WT-ATF3, M2, or M3 were processed immediately (blue columns) or after 24 h incubation in medium containing anti-ngf antibody (red columns), or NGF plus EGF (green columns). Copy numbers of ICP27, TK, VP16, and U L 41 normalized to cellular RNA were plotted. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant. E Shu et al.

6 Fig. 7. ATF3 mrna was stimulated by HSV-1(F), recombinant viruses of HSV-1 expressing WT ATF3, M2, or M3 infection. Copy numbers of ATF3 normalized to cellular RNA were plotted as described in Fig CGCGGATCCATGGATTACAAGGATGACGATGACAAGTTTGTCAAGGAAGAGC- TGAGGTTTGCC-3 ; ATF3(1 39), reverse: 5 -GCTCTAGAGGGCGTCAGGTTAGCA- AAATCCTCAAAC-3 ; and ATF3 (85 181), forward: 5 -GCTCTAGAGAAGAT- GAAAGGAAAAAGAGGCGACGA-3. The WT and mutant ATF3 ORFs tagged with Flag at the N terminus were subcloned in the multiple cloning site of pcdna3.1(+).the resulting plasmids were named pn-flag-atf3 (WT, 1 181), M1 (1 148), M2 (1 135), M3 (1 118), M4 (40 181), M5 (40 148), and M6 (1 39, ), respectively. Plasmids containing ATF3 ORFs were generously provided by Shigetaka Kitajima (37) and Hsonwin Hai (22). HEK 293T Cell Transient Transfection. The procedure for transfection of HEK 293T cells has been described previously (24). The cells were harvested at 24 or 48 h after transfection and lysed according to the protocol for subsequent use. GST Pull-Down Assay. In brief, HEK 293T cells were transfected with 1 μg of pn-flag-atf3 (1 181), M2 (1 135), or M3 (1 118). Extracted total proteins were pulled down by beads bound to GST fused to the ATF3 protein, and then subjected to immunoblot analysis with the mouse anti-flag antibody. The pull-down procedures were performed as described previously (25). Immunoblot Analysis. Cells were collected at the indicated time points after infection. The procedures for harvesting, solubilization, protein quantification, SDS/PAGE, and transfer to nitrocellulose membranes were as described previously (25). The membrane was probed for Flag tag and β-actin with the antibodies listed above. Construction of Recombinant ATF3 and Mutant Viruses. The recombinant HSV-1 viruses expressing WT ATF3 or mutants were constructed using the BAC system, as described previously (27). In brief, ORFs of WT human ATF3 (46) or two mutants (M2 and M3) tagged with Flag at the N terminus were inserted between the genes encoding U L 3andU L 4 under CMV promoter, and the resulting viruses were designated R501 (WT-ATF3), R502 (M2), and R503 (M3). Northern Blot Analysis. The Northern blot analyses were carried out as described previously (47). In brief, 10 μg of total RNA was loaded onto denaturing formaldehyde gel, transferred, and probed with random hexanucleotideprimed 32 P-labeled fragments of the indicated genes (ATF3 and LAT). Electrophoretic Mobility Shift and Supershift Assays. For this experiment, 20 μg of nuclear extracted protein was added to a binding reaction mixture containing 0.5 ng of 32 P-labeled oligonucleotide, followed by electrophoresis and autoradiography. The Oligonucleotide labeling, nuclear protein fraction extraction, and electrophoretic mobility shift assay (EMSA) were done as described previously (47 49). The canonical CRE and CRE-like (CRE1 and CRE2) oligonucleotides were designed to the LAT promoter sequence HSV-1(F) corresponding to ATF3-binding sites. The two complementary singlestrain oligonucleotides were synthesized and annealed before labeling. The sequences were as follows: WT CRE1 ( ), forward: 5 -TGT- TTTTGCTGCGTCATCTG AGCCTT-3, reverse: 5 -AAGGCTCAGATGACGCAG- CAAAAACA-3 ; mutant CRE1, forward: 5 -TGTTTTTGCTGCGCA ATCTGAGCCTT-3, reverse: 5 -AAGGCTCAGATTGCGC AGCAAAAACA-3 ; WT CRE2 ( ), forward: 5 -AAAATAAAATTACATCACCTACCCAC-3, reverse: 5 -GTGGGTAGGT- GATGTAATTTTATTTT-3 ; mutant CRE2, forward: 5 -AAAATAAAATTACACA- ACCTACCCAC-3, reverse: 5 -GTGGGTAGGTTGTGTAATTTTATTTT-3 ; canonical CRE, forward: 5 AGAGATTGCCTGACGTCAGAGAGC TAG-3, reverse: 5 -CTAG- CTCTCTGACG TCAGGCAATCTCT-3. Murine Model of Virus Infection and Reactivation from Latency. The 4-wk-old inbred female CBA/J mice (The Jackson Laboratory) received unrestricted access to food and water. All animal studies were done in accordance with protocols approved by the Institutional Animal Care and Use Committee of the University of Chicago. Mice were infected by cornea scarification, and virus in the trigeminal ganglia was reactivated by a ganglionic organ culture model as described previously (9). RNA Extraction and RT-qPCR Analysis. The procedures for RNA extraction from cells or murine tissues and RT-qPCR analysis have been described previously (9). Viral mrna expression levels in ganglia immediately after excision, after a 24-h incubation in anti-ngf antibody, and after a 24-h incubation in NGF plus EGF were compared using the Student t test. A two-tailed P value of < 0.05 was considered statistically significant. MICROBIOLOGY SEE COMMENTARY PNAS PLUS Fig. 8. Schematic representation of the known factors maintaining HSV in latent form. The transition from the silent (latent) to activated state, shown here by a long black arrow, is irreversible. As indicated above the black arrow, the transition is blocked by constitutively present LAT, NGF, and chromatin-modifying enzymes (e.g., HDACs, methyl transferases) (50 53), activated factors exemplified by STAT3, and inducible factors such as ATF3. Inhibitors, dominant negative mutants, and other entities that support the roles of constitutively present, activated, or inducible factors are shown below the black arrow. Shu et al. PNAS Published online August 24, 2015 E5425

7 ACKNOWLEDGMENTS. We thank Lindsay Smith for technical assistance withtheanimalexperiments,anddr.shigetaka Kitajima (Tokyo Medical and Dental University) and Dr. Hsonwin Hai (Ohio State University College of Medicine) for the generous gifts of plasmids encoding ATF3. This study was supported by National Cancer Institute Grant CA Roizman B, Knipe DM, Whitley RJ (2013) Herpes simplex viruses. Fields Virology, eds Knipe DM, et al. (Wolters Kluwer/Lippincott-Williams and Wilkins, New York), 6th Ed, pp Kang W, Mukerjee R, Fraser NW (2003) Establishment and maintenance of HSV latent infection is mediated through correct splicing of the LAT primary transcript. Virology 312(1): Zabolotny JM, Krummenacher C, Fraser NW (1997) The herpes simplex virus type kilobase latency-associated transcript is a stable intron which branches at a guanosine. J Virol 71(6): Alvira MR, Goins WF, Cohen JB, Glorioso JC (1999) Genetic studies exposing the splicing events involved in herpes simplex virus type 1 latency-associated transcript production during lytic and latent infection. J Virol 73(5): Drolet BS, et al. (1998) The region of the herpes simplex virus type 1 LAT gene involved in spontaneous reactivation does not encode a functional protein. Virology 242(1): Farrell MJ, Dobson AT, Feldman LT (1991) Herpes simplex virus latency-associated transcript is a stable intron. Proc Natl Acad Sci USA 88(3): Preston CM, Efstathiou S (2009) Molecular basis of HSV latency and reactivation. Human Herpesviruses: Biology, Therapy, and Immunoprophylaxis, eds Arvin A, et al. (Cambridge Univ Press, Cambridge), 1st Ed, pp Roizman B, Whitley RJ (2013) An inquiry into the molecular basis of HSV latency and reactivation. Annu Rev Microbiol 67: Du T, Zhou G, Roizman B (2011) HSV-1 gene expression from reactivated ganglia is disordered and concurrent with suppression of latency-associated transcript and mirnas. Proc Natl Acad Sci USA 108(46): Du T, Zhou G, Roizman B (2012) Induction of apoptosis accelerates reactivation of latent HSV-1 in ganglionic organ cultures and replication in cell cultures. Proc Natl Acad Sci USA 109(36): Du T, Zhou G, Roizman B (2013) Modulation of reactivation of latent herpes simplex virus 1 in ganglionic organ cultures by p300/cbp and STAT3. Proc Natl Acad Sci USA 110(28):E2621 E Zhou G, Du T, Roizman B (2013) HSV carrying WT REST establishes latency but reactivates only if the synthesis of REST is suppressed. Proc Natl Acad Sci USA 110(6): E498 E Hunt D, Raivich G, Anderson PN (2012) Activating transcription factor 3 and the nervous system. Front Mol Neurosci 5: Tsujino H, et al. (2000) Activating transcription factor 3 (ATF3) induction by axotomy in sensory and motoneurons: A novel neuronal marker of nerve injury. Mol Cell Neurosci 15(2): Abe T, Oue N, Yasui W, Ryoji M (2003) Rapid and preferential induction of ATF3 transcription in response to low doses of UVA light. Biochem Biophys Res Commun 310(4): Ackland-Berglund CE, Davido DJ, Leib DA (1995) The roles of the camp-response element and TATA box in expression of the herpes simplex virus type 1 latency-associated transcripts. Virology 210(1): Bloom DC, Stevens JG, Hill JM, Tran RK (1997) Mutagenesis of a camp response element within the latency-associated transcript promoter of HSV-1 reduces adrenergic reactivation. Virology 236(1): Marquart ME, et al. (2001) A camp response element within the latency-associated transcript promoter of HSV-1 facilitates induced ocular reactivation in a mouse hyperthermia model. Virology 284(1): Perng GC, et al. (2000) Virus-induced neuronal apoptosis blocked by the herpes simplex virus latency-associated transcript. Science 287(5457): Peng W, et al. (2003) The gene that encodes the herpes simplex virus type 1 latencyassociated transcript influences the accumulation of transcripts (Bcl-xL and Bcl-xS) that encode apoptotic regulatory proteins. J Virol 77(19): Hai TW, Liu F, Coukos WJ, Green MR (1989) Transcription factor ATF cdna clones: An extensive family of leucine zipper proteins able to selectively form DNA-binding heterodimers. Genes Dev 3(12B): Chen BP, Liang G, Whelan J, Hai T (1994) ATF3 and ATF3 delta Zip: Transcriptional repression versus activation by alternatively spliced isoforms. J Biol Chem 269(22): Pan YX, Chen H, Kilberg MS (2005) Interaction of RNA-binding proteins HuR and AUF1 with the human ATF3 mrna 3 -untranslated region regulates its amino acid limitation-induced stabilization. J Biol Chem 280(41): Shu M, Taddeo B, Zhang W, Roizman B (2013) Selective degradation of mrnas by the HSV host shutoff RNase is regulated by the UL47 tegument protein. Proc Natl Acad Sci USA 110(18):E1669 E Shu M, Taddeo B, Roizman B (2015) Tristetraprolin recruits the herpes simplex virion host shutoff RNase to AU-rich elements in stress response mrnas to enable their cleavage. J Virol 89(10): Millhouse S, Kenny JJ, Quinn PG, Lee V, Wigdahl B (1998) ATF/CREB elements in the herpes simplex virus type 1 latency-associated transcript promoter interact with members of the ATF/CREB and AP-1 transcription factor families. JBiomedSci5(6): Gu H, Roizman B (2007) Herpes simplex virus-infected cell protein 0 blocks the silencing of viral DNA by dissociating histone deacetylases from the CoREST-REST complex. Proc Natl Acad Sci USA 104(43): Taddeo B, Esclatine A, Zhang W, Roizman B (2003) The stress-inducible immediateearly responsive gene IEX-1 is activated in cells infected with herpes simplex virus 1, but several viral mechanisms, including 3 degradation of its RNA, preclude expression of the gene. J Virol 77(11): Taddeo B, Zhang W, Roizman B (2009) The virion-packaged endoribonuclease of herpes simplex virus 1 cleaves mrna in polyribosomes. Proc Natl Acad Sci USA 106(29): Rooney JF, et al. (1991) Prevention of ultraviolet light-induced herpes labialis by sunscreen. Lancet 338(8780): BenMohamed L, et al. (2015) Decreased reactivation of a herpes simplex virus type 1 (HSV-1) latency-associated transcript (LAT) mutant using the in vivo mouse UV-B model of induced reactivation. J Neurovirol, /s Kriesel JD, et al. (1997) Anti interleukin-6 antibodies inhibit herpes simplex virus reactivation. J Infect Dis 175(4): Kriesel JD, Ricigliano J, Spruance SL, Garza HH, Jr, Hill JM (1997) Neuronal reactivation of herpes simplex virus may involve interleukin-6. J Neurovirol 3(6): Vicetti Miguel RD, et al. (2010) 17-beta estradiol promotion of herpes simplex virus type 1 reactivation is estrogen receptor-dependent. J Virol 84(1): Rivera L, Beuerman RW, Hill JM (1988) Corneal nerves contain intra-axonal HSV-1 after virus reactivation by epinephrine iontophoresis. Curr Eye Res 7(10): Kiryu-Seo S, et al. (2008) Neuronal injury-inducible gene is synergistically regulated by ATF3, c-jun, and STAT3 through the interaction with Sp1 in damaged neurons. J Biol Chem 283(11): Hua B, et al. (2006) A splice variant of stress response gene ATF3 counteracts NFkappaB dependent anti-apoptosis through inhibiting recruitment of CREB-binding protein/p300 coactivator. J Biol Chem 281(3): Jiang X, et al. (2011) The herpes simplex virus type 1 latency-associated transcript can protect neuron-derived C1300 and Neuro2A cells from granzyme B-induced apoptosis and CD8 T-cell killing. J Virol 85(5): Ahmed M, Lock M, Miller CG, Fraser NW (2002) Regions of the herpes simplex virus type 1 latency-associated transcript that protect cells from apoptosis in vitro and protect neuronal cells in vivo. J Virol 76(2): Wilcox BJ, Applegate MD, Portera-Cailliau C, Koliatsos VE (1995) Nerve growth factor prevents apoptotic cell death in injured central cholinergic neurons. J Comp Neurol 359(4): Wilcox CL, Johnson EM, Jr (1987) Nerve growth factor deprivation results in the reactivation of latent herpes simplex virus in vitro. J Virol 61(7): Dziennis S, Alkayed NJ (2008) Role of signal transducer and activator of transcription 3 in neuronal survival and regeneration. Rev Neurosci 19(4-5): Li GH, et al. (2009) STAT3 silencing with lentivirus inhibits growth and induces apoptosis and differentiation of U251 cells. J Neurooncol 91(2): Lee N, Neitzel KL, Devlin BK, MacLennan AJ (2004) STAT3 phosphorylation in injured axons before sensory and motor neuron nuclei: Potential role for STAT3 as a retrograde signaling transcription factor. J Comp Neurol 474(4): Ejercito PM, Kieff ED, Roizman B (1968) Characterization of herpes simplex virus strains differing in their effects on social behaviour of infected cells. J Gen Virol 2(3): Thompson RL, Sawtell NM (2010) Therapeutic implications of new insights into the critical role of VP16 in initiating the earliest stages of HSV reactivation from latency. Future Med Chem 2(7): Shu M, Taddeo B, Roizman B (2013) The nuclear-cytoplasmic shuttling of virion host shutoff RNase is enabled by pul47 and an embedded nuclear export signal and defines the sites of degradation of AU-rich and stable cellular mrnas. J Virol 87(24): Advani SJ, Weichselbaum RR, Roizman B (2000) E2F proteins are posttranslationally modified concomitantly with a reduction in nuclear binding activity in cells infected with herpes simplex virus 1. J Virol 74(17): Kristie TM, Roizman B (1986) Alpha 4, the major regulatory protein of herpes simplex virus type 1, is stably and specifically associated with promoter-regulatory domains of alpha genes and of selected other viral genes. Proc Natl Acad Sci USA 83(10): Liang Y, et al. (2013) A novel selective LSD1/KDM1A inhibitor epigenetically blocks herpes simplex virus lytic replication and reactivation from latency. MBio 4(1): e Liang Y, Vogel JL, Narayanan A, Peng H, Kristie TM (2009) Inhibition of the histone demethylase LSD1 blocks alpha-herpesvirus lytic replication and reactivation from latency. Nat Med 15(11): Knipe DM, Cliffe A (2008) Chromatin control of herpes simplex virus lytic and latent infection. Nat Rev Microbiol 6(3): Messer HG, Jacobs D, Dhummakupt A, Bloom DC (2015) Inhibition of H3K27me3- specific histone demethylases JMJD3 and UTX blocks reactivation of herpes simplex virus 1 in trigeminal ganglion neurons. J Virol 89(6): E Shu et al.