JOURNAL OF VIROLOGY, Oct. 1999, p. 8415–8426 Vol. 73, No. 10 0022-538X/99/$04.00ϩ0 Copyright © 1999, American Society for Microbiology. All Rights Reserved.

Activation of cJUN N-Terminal Kinase by Virus Type 1 Enhances Viral Replication

1 1,2 T. I. MCLEAN AND S. L. BACHENHEIMER * Curriculum in Genetics and Molecular Biology1 and Department of Microbiology and Immunology,2 University of North Carolina School of Medicine, Chapel Hill, North Carolina 27599

Received 19 February 1999/Accepted 19 July 1999

Signal transduction pathways convey signals generated at the cell surface into the cell nucleus in order to initiate a program of that is characteristic for particular stimuli. Here we present evidence that infection by herpes simplex virus type 1 activated the two terminal kinases, cJUN N-terminal kinase (JNK) and p38, of stress-activated signal transduction kinase cascades. By using a solid-phase kinase assay, a phospho- specific antibody, and extracts prepared from a variety of infected cell types, we determined that activation of both kinases began 3 to 4 h postinfection (p.i.) and remained elevated out to 14 h p.i. Through the use of UV-irradiated or antibody-neutralized wild-type virus and the temperature-sensitive mutant tsB7, the high level of JNK activation was shown to be dependent on viral gene expression. Activation of JNK following infection by vi13, an ICP4 mutant virus that does not express early or late genes, suggested that only virus entry and immediate-early gene expression were necessary for JNK activation. The activation of JNK and p38 correlated with increased chloramphenicol acetyltransferase (CAT) activity in reporter assays dependent upon the activity of cJUN and ATF2 trans-activation domains. Increased CAT activity dependent on TRE and CRE sites was also observed in response to herpes simplex virus infection. The activities of ERK and ERK-dependent transcription factors were unchanged or depressed following infection, showing that activation of JNK and p38 was a specific event. Finally, the activation of JNK was important for the efficiency of viral replication. The yield of virus in NIH 3T3 cells stably expressing JIP-1, an inhibitor of JNK translocation to the nucleus, was reduced 70% compared to that of control cells, in single-step growth experiments.

The cell’s ability to sense external stimuli and to react by are collectively referred to as mitogen-activated ki- initiating a program of gene expression often involves propa- nases (MAPKs). MEK1/2 now belongs to a class of gation of a cell surface-initiated signal along a specific path- known as MAPK kinases (MAPKKs), and RAF and related way(s) of protein kinases whose ultimate targets are nuclear- kinases are called MAPKK kinases (MAPKKKs). The most acting transcription factors. Conditions such as the type of well characterized of these analogous pathways are called the stimulus, duration of stimulation, and cell type can all play stress-activated protein kinase (SAPK) pathways, because roles in whether the transduced signals are interpreted as stresses such as UV exposure and high osmolarity lead to their growth stimulatory (proliferative), growth inhibitory (differen- activation (40, 52, 92). JNK1-3 (cJUN N-terminal kinase) and tiating), or apoptotic (death inducing) (40, 52, 75, 92). The p38, sometimes collectively referred to as SAPKs, are the prototypical signal transduction pathway is the mitogenic path- MAPKs for these pathways, and corresponding MAPKKs and way. This pathway is initiated when growth factors bind to and MAPKKKs have been characterized that specifically activate activate their cognate growth factor receptors on the surface of JNKs and/or p38 (40, 52, 92). The transcription factor ATF2 is the cell. Activation of the receptor leads to the membrane- a target for both p38 and JNKs, while the JNKs can also bound G-protein, RAS, adopting an active GTP-bound state phosphorylate cJUN (33, 55, 86). (18, 48, 75). RAS is critical for coordinating the activation of a Viruses are ultimately dependent upon the host cell for their serine/threonine kinase called RAF (5, 18). RAF then initiates replication. Because they reorganize or utilize various cellular a kinase cascade by phosphorylating and activating two highly functions, it seems likely that viruses would take advantage of related dual-specific kinases, MEK1 and MEK2, which in turn the preexisting signaling pathways to induce cellular and/or phosphorylate two highly related serine/threonine kinases, viral gene expression to promote viral replication. Indeed, ERK1 and ERK2 (11, 12, 15, 75). Upon activation, ERKs can studies characterizing the effects of infection by other viruses, migrate into the nucleus and activate transcription factors such such as bovine papillomavirus (60), virus (50), simian as cMYC and ELK1 by phosphorylating their trans-activation virus 40 (83), human immunodeficiency virus type 1 (41), and domains (TADs) (2, 7, 9, 28, 29, 32). These transcription fac- herpesvirus saimiri (44), have shown a dependence upon ERK tors can induce the expression of cdc25 and c-fos, respectively, cascade signaling for replication, and in some cases, viral pro- for example, which are important for promoting teins have been identified that induce activation of ERKs. progression into (3, 7, 25, 28, 47). Additionally, other studies have described activation of differ- Kinases analogous to the ERKs and pathways parallel to the ent MAPKs, but the purposes for their activation following RAF-MEK-ERK pathway have been described. The related infection remain unclear (20, 71, 81, 100). terminal serine/threonine kinases, which include the ERKs, During lytic infection by herpes simplex virus (HSV) in per- missive cells, there are increases in activities of AP-1 (42) and ␬ * Corresponding author. Mailing address: The Department of Mi- NF- B (30, 68, 77), both of which are downstream targets of crobiology and Immunology, 837 Jones, CB 7290, University of North SAPK pathways (53, 57, 92), suggesting that these pathways Carolina School of Medicine, Chapel Hill, NC 27599. Phone: (919) may be activated. Additionally, there is evidence that infection 966-2445. Fax: (919) 962-8103. E-mail: [email protected]. by HSV inhibits cell cycle progression, raising the possibility

8415 8416 MCLEAN AND BACHENHEIMER J. VIROL.

that ERK activity, which promotes progression, is inhibited Mem. Cells were refed with fresh medium at 5 and 18 h posttransfection and (14, 19). We present data here to support these predictions. were then infected and harvested as described above. Kinase assays. Cells were mock infected or infected with KOS (MOI ϭ 5) and We show that not only are SAPK pathways activated following harvested at the indicated times p.i. to produce whole-cell lysates. Kinase assays infection, but activation of one SAPK, JNK, promotes efficient were performed as described in reference 37. Cells were washed in phosphate- viral replication. ERK and ERK-dependent activities are not buffered saline and frozen in lysis buffer (25 mM HEPES [pH 7.7], 300 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.1% Triton X-100 plus inhibitors: 0.5 mM activated following infection. In agreement with our data is a ␤ ␮ dithiothreitol, 20 mM -glycerophosphate, 0.1 mM Na3VO4,2 g of leupeptin/ recent report which also showed the activation of JNK and p38 ml, 100 g of phenylmethylsulfonyl fluoride/ml, 1 mM NaF, and 1 ␮g of aprotinin/ following HSV infection (100). ml). Cells were allowed to thaw and lyse at 4°C for 4 to 5 h. Samples were then clarified by spinning at 13,000 ϫ g for 10 min at 4°C. For each sample, a volume of the resulting supernatant, representing an equivalent number of cells, was MATERIALS AND METHODS diluted (final concentrations: 20 mM HEPES, 75 mM NaCl, 2.5 mM MgCl2, 0.01 mM EDTA, 0.05% Triton X-100 plus inhibitors). To precipitate JNK, 10 ␮gof Cells and virus. NIH 3T3, U2OS, C33, and Vero cells were grown and main- glutathione S-transferase (GST)-cJUN (for JNK assays) was conjugated to GST- tained in Dulbecco’s modified Eagle’s media (DMEM) supplemented with 2 mM Sepharose beads and incubated overnight with lysates. In the assays utilizing ␮ L-glutamine. For growth of NIH 3T3 cells (originally obtained from Channing GST-ATF2, 6 g of that substrate was added to each sample. After extensive Der, UNC—Chapel Hill), the medium was supplemented with 10% bovine calf washing (wash buffer: 20 mM HEPES (pH 7.7), 50 mM NaCl, 2.5 mM MgCl2, 0.1 serum, while that for the Vero cells contained 5% bovine calf serum. DMEM for mM EDTA, 0.05% Triton X-100), the beads were incubated in kinase buffer (20 ␤ U2OS and C33 cells was supplemented with the addition of 10% fetal calf serum. mM HEPES [pH 7.6] and 20 mM MgCl2 plus 20 mM -glycerophosphate, 20 BHK-21 cells were grown and maintained in alpha minimum essential media mM p-nitrophenyl phosphate, 0.1 mM Na3VO4, 2 mM dithiothreitol, 1 mM NaF, ␮ ␥ 32 supplemented with L-glutamine as described above and 10% fetal calf serum. 20 M ATP) with 5 Ci of [ - P]ATP per reaction for 30 min at 30°C. Samples ϫ The G13.5 and G14.1 cell lines were established via stable transfection of NIH were diluted 1:2 in 2 sample buffer, boiled, and fractionated by sodium dodecyl 3T3 cells with either a pcDNA3 empty vector or a pcDNA3 JIP-1-expressing sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on 12% polyacrylamide vector, respectively. The JIP-1 vector (17) was provided by Roger Davis (Uni- gels. Experiments using the viral antibodies were performed after versity of Massachusetts Medical Center). These cells were selected and main- the virus had been incubated with the antibodies for1hat37°C prior to infecting tained in media supplemented with 600 ␮g of G418 (Geneticin; Gibco BRL) per the cells. For the ERK assays, cells were placed in 0.5% serum for 24 h prior to ml. All experiments were performed with the KOS strain of HSV-1, unless infection. The diluted extracts were first precleared with normal rabbit serum otherwise noted. tsB7, provided by David Knipe (Harvard University), is a and protein A beads. ERK1 antibody was added and allowed to incubate for at temperature-sensitive mutant derived from the HFEM strain of HSV-1 (4, 51). least 5 h, and then protein A beads were added for an overnight incubation. After ␥ 32 ␮ The virus mutant vi13, provided by Neal DeLuca (University of Pittsburgh), washing, the beads were resuspended in kinase buffer with [ - P]ATP and 10 g of myelin basic protein as a substrate and incubated for 30 min at 30°C. Samples encodes an ICP4 containing a small insertion at amino acid (aa) 338, which ϫ reduces ICP4’s affinity for DNA, prevents early and late gene expression, and were diluted 1:2 in 2 sample buffer, boiled, and fractionated by SDS-PAGE on results in overexpression of immediate-early (IE) proteins (reference 82 and this 15% polyacrylamide gels. report). Immunoblotting. Fractionated extracts were prepared as described by Hilton Antibodies. ERK1 (C-16) and JNK1 (FL) antibodies were purchased from et al. (39). Cell equivalent amounts of cellular lysates prepared for JNK kinase Santa Cruz Biotechnology, Inc. p38 (no. 9212) and phospho-p38 (no. 9211) assays were run on 10 or 12% polyacrylamide-SDS gels. Proteins were trans- antibodies (New England Biolabs) were a gift from Eng-Shang Huang. Mouse ferred to a polyvinylidene difluoride (PVDF) membrane and probed for JNK, monoclonal antibodies to viral IE proteins ICP0 (1083) and ICP4 (H943) were ERK, p38, or activated p38 or the indicated viral proteins. Membranes were obtained from Lenore Pereira (University of California at San Francisco). Rab- blocked in TBST (150 mM NaCl, 20 mM Tris [pH 7.5], 0.05% Tween 20) with bit polyclonal antibody to viral replication protein ICP8 was originally obtained 3% milk, and all probing and washing was done in TBST with 0.5% milk. The from Ken Powell (Burroughs Wellcome). Rabbit polyclonal antibody against secondary antibody was detected with the Renaissance chemiluminescence re- viral glycoprotein gC (Ab47) and monoclonal antibody against viral glycoprotein agent (DuPont NEN). gB (SS10) were obtained from Roselyn Eisenberg and Gary Cohen (University Single-step growth assays. NIH 3T3 cells (G13.5 and G14.1) were seeded (2.5 ϫ 105) into 60-mm-diameter dishes and infected the next day with HSV-1 of Pennsylvania). Monoclonal antibodies against viral gD (III 174- ϭ 1.2), gH (52S), and gC (VII 13-7) were provided by Patricia Spear (Northwestern KOS (MOI 5). Cells and supernatant were harvested at the indicated times p.i. University). To neutralize HSV, the amount of gB, gD, and gH antibody added and frozen and thawed four times. Virus titers were determined by plaque assays was enough to reduce virus titer by at least 2 orders of magnitude. Based on the on Vero cells. Each titration was done in triplicate, and each point represents relative quantity of antibody used for each of the neutralization experiments, an three independent experiments. equal amount of antibody protein to gC was used in control experiments. Anti- mouse and antirabbit secondary antibodies were obtained from Santa Cruz Biotechnology, Inc. RESULTS CAT Assays. Each effector plasmid used encodes a containing HSV induces the trans-activation function of cJUN and the yeast GAL4 DNA-binding domain and the TAD of the respective transcrip- tion factor. The 5ϫGAL-CAT reporter plasmid (54) encodes the bacterial chlor- ATF2 via a RAS-independent mechanism. As a first means of amphenicol acetyltransferase (CAT) protein and contains GAL4 DNA-binding investigating how HSV may be affecting cellular signal trans- sites in the promoter. The 2AP1CAT reporter plasmid (90) contains two con- duction pathways, we assessed the activities of the endpoints of sensus AP-1 sites upstream of the CAT gene. The 3ϫCRECAT reporter plasmid several of these pathways by determining the activity of various contains three copies of the CRE consensus sequence upstream of the adeno- virus E1b promoter and the CAT gene. transcription factors known to be phosphorylated and acti- For each assay, 5 ϫ 105 NIH 3T3 or BHK-21 cells in 60-mm-diameter dishes vated by different MAPKs. In these experiments, NIH 3T3 were transfected by the calcium phosphate precipitation procedure (31) with 2.5 cells were cotransfected with a CAT reporter plasmid contain- ␮gofa5ϫGAL-CAT reporter and the indicated amount of one of the following ing five GAL4 DNA-binding sites upstream of the CAT coding effectors: 150 ng of GAL-ELK1 (58), 500 ng of GAL-cFOS, 500 ng of GAL- cMYC (2), 500 ng of GAL-cJUN (37), 500 ng of GAL-ATF2, or 250 ng of sequence and one of a set of plasmids that encodes a fusion GAL-RELA (80). In experiments using the pZIP vectors, 250 ng of pZIP- protein containing the GAL4 DNA-binding domain and the RAS15A (8), 200 ng (unless stated otherwise) of CDC25X (72), or equivalent TAD of a transcription factor. This experimental design ne- amounts of the empty pZIP vector were cotransfected with 5ϫGAL-CAT and gates any differences in DNA binding and focuses solely on the either GAL-cJUN or GAL-ATF2. Other dishes were transfected with 2.5 ␮gof 2AP1CAT or 3ϫCRECAT reporter. Eighteen hours after transfection, the cells trans-activation ability of the various recombinant transcription were refed, and 12 h later, the cells were either mock infected or infected with factors. We utilized NIH 3T3 cells because the response of HSV KOS (multiplicity of infection [MOI] ϭ 5). Eighteen hours postinfection transcription factors to signals mediated by MAPK pathways (p.i.) the cells were harvested. CAT assays were performed (31) with equivalent and the methodology using these constructs in similar experi- amounts of protein from each sample. Results were quantitated on an AMBIS scanner and were the average of at least three separate experiments, except in ments to study these responses have been well characterized in Fig. 2A, which was performed only once. this cell line (10, 35, 49, 70, 90, 94). Figure 1A shows the results Lipofectamine reagent (GIBCO BRL) was also used to transfect cells in some of the CAT assays. Infection of transfected cells by HSV re- experiments as indicated. Opti-Mem (GIBCO BRL) (400 ␮l) was mixed with the pressed the level of CAT activity when the effector was GAL- previously described amount of DNA, and another 400 ␮l of Opti-Mem was mixed with 8 ␮l of Lipofectamine. The two solutions were mixed and allowed to cMYC, GAL-ELK1, or GAL-RELA. GAL-cFOS-dependent sit at room temperature for 30 min. An additional 2.2 ml of Opti-Mem was added CAT activity was relatively unaffected by infection. In contrast, to each solution prior to overlaying the cells that had been rinsed with Opti- CAT activity in extracts from HSV-infected cells transfected VOL. 73, 1999 ACTIVATION OF JNK BY HSV 8417

served increased GAL-ATF2-dependent CAT activity, CRE- dependent promoter activity was also measured with 3ϫCRE- CAT. In this case, we observed no change in CRE-dependent activity in NIH 3T3, where the constitutive level was high. In BHK-21 cells, however, where constitutive activity was low, we observed a two-and-a-half-fold increase in CRE-dependent CAT activity after infection (Fig. 1B). The presence of growth factors and mitogens upregulates RAS activity. Activated RAS in turn coordinates the induction of the RAF-MEK-ERK pathway, which upregulates the activ- ity of some transcription factors, such as cMYC and Ets (6, 87). Growth factors only weakly activate JNK, and hence cJUN, in a RAS-dependent manner (46, 95). cJUN can also be activated by pathways that are not RAS dependent (5, 6, 61, 62, 87). These pathways, however, are inhibited by activated RAS, since the absence of activated RAS results in modest increases in cJUN activity (5, 6, 61, 62, 87). To ascertain whether the activation of cJUN and ATF2 detected after infection was dependent upon RAS, we repeated the CAT assays and in- cluded an expression vector that encoded a dominant-negative version of RAS, RAS15A, or a constitutively active version of CDC25, CDC25X. RAS15A can still bind to guanine exchange factors (GEFs), but it is unable to release GDP in exchange for GTP (8). RAS15A, therefore, sequesters GEFs into nonpro- ductive complexes and inhibits the activation of wild-type RAS. CDC25X is a truncated, membrane-localized form of CDC25, a RAS GEF. Due to its membrane localization, it can constitutively interact with and activate RAS (72). In NIH 3T3 cells, cJUN can be regulated by a RAS-depen- dent pathway, since constitutively activated CDC25 increased GAL-cJUN-dependent CAT activity (Fig. 2A, lanes 1 to 4), while dominant-negative RAS (acting downstream of CDC25) reversed this increase (Fig. 2A, lanes 5 to 7). In HSV-infected cells, however, neither a dominant-negative RAS nor a consti- FIG. 1. HSV activates the TAD of cJUN and ATF2 and stimulates AP-1 activity. (A) NIH 3T3 cells (5 ϫ 105) were transfected with 2.5 ␮gofa5ϫGAL- tutively active CDC25 had any effect on the observed increase CAT reporter and 150 to 500 ng of the various effectors, each of which encodes in GAL-cJUN-dependent activity (Fig. 2B, lanes 2, 4, and 6), a fusion protein containing the GAL4 DNA-binding domain and the TAD of the while both these effectors caused an increase in GAL-cJUN respective transcription factors. Under each effector heading is the reported activity in mock-infected cells (Fig. 2B, lanes 1, 3, and 5). Thus MAPK, if known, that activates the TAD of that transcription factor. (B) NIH 3T3 or BHK-21 cells (5 ϫ 105) were transfected with 2.5 ␮g of 2AP1CAT or while cJUN can be weakly activated by a RAS-dependent 3ϫCRECAT reporter. 2AP1CAT contains two consensus AP-1 sites upstream of mechanism in uninfected NIH 3T3 cells, this pathway does not the CAT gene. 3ϫCRECAT contains three copies of the CRE in front of the significantly contribute to GAL-cJUN activation in HSV-in- adenovirus E1b promoter upstream of the CAT gene. For all CAT experiments, fected cells. Inhibition of RAS activity (e.g., expression of cells were refed 18 h after transfection, and 12 h later, the cells were either mock infected or infected with KOS (MOI ϭ 5). Eighteen hours p.i., the cells were RAS15A) can derepress pathways that also modestly activate harvested. CAT assays were performed with equivalent amounts of protein from cJUN as seen when comparing mock-infected pZIP- and pZIP- each sample. Results were quantitated on an AMBIS scanner and are the RAS15A-transfected cells (Fig. 2B, lanes 1 and 3). In contrast, average of at least three separate experiments. there is no difference between HSV-infected pZIP- and pZIPRAS15A-transfected cells (Fig. 2B, lanes 2 and 4). When RAS15A was cotransfected with CDC25X, the level of CAT with GAL-cJUN or GAL-ATF2 was increased 20-fold over activity was restored to that seen in cells transfected with the that detected in mock-infected extracts. (Note differences in reporter and GAL-cJUN alone (Fig. 2B, compare lanes 1 and scale among the assays for different effectors.) No CAT activity 2 to lanes 7 and 8). Expression of the dominant-negative RAS was detected in any extracts, mock or infected, in the absence had no effect on GAL-ATF2-dependent CAT activity in mock- of a transfected effector construct (negative data not shown). or virus-infected cells (Fig. 2B, lanes 9 to 12). Furthermore, It has been reported that HSV infection causes an increase none of the pZIP vectors had any effect on the level of activity in AP-1 activity (42). AP-1 is a dimeric transcription factor expressed from 5ϫGAL-CAT in the absence of GAL-cJUN or composed either of one JUN family member and one FOS GAL-ATF2, because these samples had no detectable CAT family member or of two JUN family members (3, 46). The activity (negative data not shown). Thus, even RAS-indepen- cognate binding site for AP-1, the tetradecanoyl phorbol ace- dent cellular pathways are not contributing to virus-induced tate (TPA)-responsive element (TRE), is TGACTCA. JUN cJUN activity. can also heterodimerize with CREB family members such as HSV activates SAPKs but not ERK. We next focused our ATF2, and these complexes bind to a related sequence called study on the possible activation of JNK and p38 as suggested the cyclic AMP-responsive element (CRE), TGACGTCA (3, by increased GAL-cJUN and GAL-ATF2 activity in the CAT 46). We confirmed that HSV infection increased AP-1 activity assays. We also assayed the activity of another MAPK, ERK, in both NIH 3T3 and BHK-21 cells by using a CAT reporter, that the CAT data suggested either decreased or remained 2AP1CAT, containing two consensus AP-1 binding sites (TRE unchanged following infection. We first performed in vitro sequences) in the upstream region (Fig. 1B). Since we ob- solid substrate and immune complex kinase assays for JNK and 8418 MCLEAN AND BACHENHEIMER J. VIROL.

FIG. 2. Activation of cJUN and ATF2 TADs is RAS independent. (A) NIH FIG. 3. HSV activates JNK but not ERK. (A) BHK-21, U2OS, NIH 3T3, ϫ 5 ␮ ϫ 3T3 cells (5 10 ) were transfected with 2.5 gofa5 GAL-CAT reporter, 500 Vero, and C33-A cells were mock infected, infected with KOS (MOI ϭ 5) for ng of GAL-c-JUN, and various amounts of pZIP-CDC25X, with or without 250 10 h, or treated with anisomycin (20 ␮g/ml) for 20 min. Cells were then harvested ng of pZIP-RAS15A. An adjusted amount of pZIP vector was added accordingly to produce lysates. Equivalent numbers of cells were assayed in each sample. The to equalize the amount of transfected DNA between the samples. (B) NIH 3T3 top panel for each cell type is the result of an in vitro solid substrate kinase assay ϫ 5 ␮ ϫ cells (5 10 ) were transfected with 2.5 gofa5 GAL-CAT reporter and 500 using 10 ␮g of GST-cJUN (aa 1 to 79) conjugated to GST-Sepharose beads, as ng of either GAL-c-JUN or GAL-ATF-2 and either 250 ng of pZIP-RAS15A, described in Materials and Methods. Two different GST-cJUN substrates were 200 ng of pZIP-CDC25X, or equivalent amounts of empty pZIP vector. For all used: a wild-type (WT) GST-cJUN and a mutant (AA) GST-cJUN. In the CAT experiments, cells were refed 18 h after transfection, and 12 h later, the mutant construct, the serines that are phosphorylated to cause cJUN trans- ϭ cells were either mock infected or infected with KOS (MOI 5). Eighteen hours activation (serines 63 and 73) have been mutated to alanines. The bottom panel p.i., the cells were harvested. CAT assays were performed with equivalent for each cell type is an immunoblot, probed with an antibody to JNK, performed amounts of protein from each sample. Results were quantitated on an AMBIS on a fraction of the pull down. (B) The same cell types used in panel A were used scanner and are the average of at least three separate experiments, except in to compare the level of ERK activity in mock-infected cells (M) versus infected panel A, which was performed only once. cells at 10 h p.i. (I). The top panel shows results from an in vitro immune complex kinase assay for ERK on extracts from serum-starved cells, as described in Materials and Methods. The bottom panel is an immunoblot from a fraction of the immunoprecipitation probed with an antibody against ERK. ERK, respectively. The results from five different cell types are shown in Fig. 3. We used NIH 3T3 cells to directly correlate any differences we saw in kinase activation with the CAT data (Fig. 1). We also chose to study Vero and BHK-21 cells be- prior to autoradiography showed that all lanes contained equal cause they are commonly used for studies of HSV replication, amounts of substrate (data not shown). and C33-A and U2OS cells to show that the effects also oc- The lower panel for each cell type set is an immunoblot for curred in human cells. Although there was some variability in JNK demonstrating the amount of JNK which was precipitated the degree to which JNK was activated, all cell types tested in each experiment was constant. These results provide evi- demonstrated upregulation of JNK following infection with dence that infection was activating preexisting JNK rather than HSV, as indicated by of a wild-type GST- inducing JNK expression. Because the kinase whose activity we cJUN substrate (Fig. 3A, upper panel of each cell type set). detected must be precipitable by binding to the amino termi- The induction was greater than 20-fold in all but NIH 3T3 cells nus of cJUN, and because the specificity of the phosphoryla- (6-fold), comparable to the level seen when the cells were tion of GST-cJUN in infected cells was identical to that seen in treated with the protein synthesis inhibitor, anisomycin, a po- anisomycin-induced control cells, we conclude that we are de- tent inducer of JNK activity (22, 36, 96). JNK phosphorylates tecting bona fide JNK activity. In support of this conclusion, we serines 63 and 73 in the amino terminus of cJUN (16, 37). have obtained identical results when using an antibody against When a GST-cJUN substrate was used in which Ser63 and JNK to immunoprecipitate the kinase for use in immune com- Ser73 were mutated to alanines (AA), phosphorylation of plex kinase assays with GST-cJUN as the substrate (data not GST-cJUN was abolished. Coomassie blue staining of the gels shown). Thus, activation of JNK upon HSV infection is a VOL. 73, 1999 ACTIVATION OF JNK BY HSV 8419

universal response, regardless of the species origin or cell type assayed. Similar assays to detect ERK activity indicated no change in the level of ERK activity as a result of infection (Fig. 3B). The experiment in Fig. 3B was performed on serum-starved cells to reduce the ERK kinase activity background so as to more easily detect any increases in activity that may have been in- duced by infection. In all cell types tested, no change in ERK activity (Fig. 3B, upper panel) was detected in comparing mock-infected cells (M) to cells infected with KOS for 10 h (I). In the few cases in which differences were detected, the change could be accounted for by variability in the amount of ERK protein that was brought down in individual immunoprecipi- tations (Fig. 3B, lower panel). Kinase assays detecting ERK activity were also performed with cells grown under normal culture conditions. In these experiments, the background ERK activity was high, and no increases in activity could be detected. In fact, slight decreases were often detected (data not shown). Thus, the CAT assay results were predictive of how HSV affected the activity of various MAPKs responsible for phos- phorylating and activating the TADs of the assayed transcrip- tion factors. Upregulation of JNK correlated well with in- creased GAL-cJUN-dependent CAT activity, while ERK was unaffected or slightly downregulated, as mirrored by the down- regulation, for example, of GAL-cMYC-dependent CAT ac- tivity. The data in Fig. 3 were from cells that had been infected for FIG. 4. Activation of JNK occurs with early kinetics. U2OS cells were mock ϭ 10 h. To determine the kinetics and magnitude of JNK activa- infected or infected with KOS (MOI 5) and harvested to produce cellular lysates at the indicated times p.i. An equivalent number of cells was assayed in tion as well as to detect any transient activation of the MAPKs, each sample. (A) The top two panels show results from an experiment using 10 we repeated the kinase assays over a time course of infection in ␮g of GST-cJUN (aa 1 to 79) as the precipitating agent and substrate in kinase U2OS cells (Fig. 4). Using GST-cJUN in solid substrate kinase reactions performed as described in Materials and Methods. The bottom two ␮ assays (top two panels of Fig. 4A), we detected little or no JNK panels are the results with 6 g of GST-ATF2 (aa 1 to 254) as the precipitating agent and substrate. (B) Cell-equivalent amounts of mock-infected and wild-type activity in mock-infected cells, but increased JNK activity was KOS-infected U2OS cellular lysates prepared for JNK kinase assays and col- detected between 3 and 5 h p.i. in virus-infected cells. The lected at the indicated times p.i. were run on 10% polyacrylamide gels. Proteins activity increased 20-fold by 9 h p.i. when it appeared to pla- were transferred to a PVDF membrane and probed for activated p38 (p-p38) or teau. It should be noted that great care had to be taken during total p38 (p38). One-sixth the amount of protein was loaded onto gels to be used for detecting total p38 as was loaded onto gels to be used for detecting activated infection to minimize cell exposure to air, because this situa- p38. (C) Samples of lysate were first precleared with normal rabbit serum and tion could induce a transient JNK activation (data not shown). protein A beads. ERK1 antibody was then added to precipitate the kinase. Repeating the assay with GST-ATF2 as the substrate, we Myelin basic protein (10 ␮g) was included in the kinase reaction as a substrate, saw similar kinetics for kinase activation, but the activation was as described in Materials and Methods. much reduced in magnitude (lower two panels of Fig. 4A). Both JNK and p38 are capable of binding to and phosphory- lating ATF2, although ATF2 is a better substrate for p38 than Immune complex kinase assays for ERK showed no change it is for JNK (33, 73). To determine if p38 was playing a role in in activity over the time course of infection (Fig. 4C). The the phosphorylation of GST-ATF2 or whether JNK is solely failure to detect an induction in ERK was not due to limiting activated, we decided to investigate p38’s activation state as a substrate or antibody in immunoprecipitation or kinase reac- result of infection. We infected U2OS cells over a similar time tions, because increases in both did not result in any higher course and harvested samples to use for immunoblotting. We ERK activity (data not shown). Identical results were obtained probed the immunoblots with a phospho-specific antibody that in time course experiments using BHK-21, Vero, and C33 cells recognizes p38 phosphorylated on threonine 180 and tyrosine (data not shown). 182. Phosphorylation of these two residues correlates with IE gene expression is necessary for JNK activation. The activation of p38 (73). The use of an antibody that recognizes absence of JNK activity at early times p.i. argued against the these two residues when dually phosphorylated, therefore, possibility that virions binding to cell surface receptors was should be equivalent to doing direct kinase assays for deter- initiating a receptor-mediated signal cascade. In order to fully mining p38 activation, as has been determined by other labo- negate the possibility of a delayed cell surface signal mediating ratories (64, 71). Figure 4B shows that over a time course, an the induction of JNK, we performed kinase assays using virus activated form of p38 was detected in the infected samples that had been neutralized with antibodies specific for different (HSV p-p38 panel) with similar kinetics as was seen for the rise glycoproteins. Antibodies which bind to glyco- in JNK activation (Fig. 4A, top HSV panel). No activation of proteins B, D, and H have been shown to neutralize virus by p38 was seen in the mock samples (Mock p-p38 panel). De- blocking the virions’ ability to fuse with the cellular membrane spite some fluctuations in total p38 in the mock samples of this following initial attachment (23, 24, 66, 67). The glycoprotein representative experiment (Mock p38 panels), infection by C monoclonal antibody, VII 13-7, does not prevent binding or HSV induced activation of p38 without affecting the level of penetration, so that infection proceeds just as it would for virus total p38 (HSV p-p38 and p38 panels). This result confirmed not incubated with neutralizing antibody (84). that both JNK and p38 are activated at relatively the same time The amount of each antibody (except that for gC) used to p.i., and their activities remain high out to 11 h p.i. neutralize virus reduced the titer by Ͼ2 logs as determined by 8420 MCLEAN AND BACHENHEIMER J. VIROL.

express detectable viral proteins upon infection due to thymine dimers in the viral DNA impeding transcription of the (38). We cannot, however, exclude the possibility that UV irradiation may also cross-link some virion proteins and that this action may be responsible for the inability of UV-irradi- ated virus to induce high levels of JNK activity. To address this concern, we used a temperature-sensitive viral mutant, tsB7. At 33°C, tsB7 grows normally, albeit more slowly than the parental wild-type strain, HFEM. At 39.5°C, tsB7 virions can penetrate cells, and the move to the nuclear pores, but the viral DNA is not released from the capsids into the nucleus (4). No viral genes, therefore, are expressed at the restrictive temper- ature (4, 51). tsB7 has been demonstrated to deliver functional VP16 to the nucleus of infected cells at the restrictive temper- ature (3a). The left panels of Fig. 4C show that at 33°C, mock- infected cells do not activate JNK, but tsB7- and HFEM-in- fected cells do with kinetics similar to that seen in KOS- infected cells. The right-hand panels show that at 39.5°C, only HFEM-infected cells significantly activate JNK. Activation ac- tually occurs earlier at the higher temperature, for reasons we do not understand at this time. tsB7-infected cells at 39.5°C only induce a minimal level of JNK activity. From this result, FIG. 5. Events prior to viral gene expression are not sufficient for JNK we conclude that de novo viral gene expression is necessary for ϭ activation. (A) U2OS cells were mock infected, infected with KOS (MOI 5), full JNK activation while virion proteins may activate JNK to or infected with KOS (MOI ϭ 5) previously incubated with an antibody specific for viral gB (ss10), gC (VII 13-7), gD (III 174-1.2), or gH (52S). Cells were low levels. harvested to produce cellular lysates at the indicated times p.i. Solid substrate A temporally regulated program of IE, early, and late viral kinase assays were performed with GST-cJUN as described in Materials and gene expression occurs following HSV infection (76). To de- ϭ Methods. (B) U2OS cells were mock infected, infected with KOS (MOI 5), or termine which virally encoded proteins were responsible for infected with KOS (MOI ϭ 5) that had been irradiated with UV light sufficiently to reduce the effective titer by 3 logs. Cellular lysates were prepared at the activating JNK, we first compared the timing of JNK activation indicated times p.i., and in vitro solid substrate kinase assays were performed with the expression and accumulation of representative pro- with GST-cJUN as described in Materials and Methods. (C) U2OS cells prein- teins of each of the viral kinetic classes of gene expression (Fig. ϭ cubated at 33 or 39.5°C were mock infected, infected with tsB7 (MOI 5 based 6A). By immunoblot analysis of the same extracts used in the on titer at 33°C), or infected with HFEM, the strain from which tsB7 was derived. Infections then proceeded at the indicated temperatures, and cellular lysates kinase assays, we could detect ICP4, an IE protein, within 1 h were prepared at the indicated times p.i. In vitro solid substrate kinase assays p.i. and a rapid increase in expression peaking at 5 h p.i. were performed with GST-cJUN as described in Materials and Methods. Expression of an , ICP8, was first detected at 5 h and increased out to 11 h. A , glycoprotein C, was not detected until 9 h p.i. Comparing the patterns of protein plaque reduction assays. A roughly equivalent protein amount, expression (Fig. 6A) and JNK activation (Fig. 4A) revealed a compared to monoclonal antibodies against gB, gD, and gH, rise in kinase activity following a significant accumulation of was used in experiments involving gC antibody. Additionally, ICP4, but concurrently with the accumulation of ICP8 (Fig. cells infected with virus which had been treated with the em- 6A). This result could be interpreted in two ways. First, the pirically derived amounts of gB, gD, and gH neutralizing an- expression of an early protein(s) leads to activation of JNK. tibody described above did not express viral proteins, while Second, one or more IE proteins are responsible for JNK cells infected with virus incubated with gC antibody expressed activation, so that the results of IE protein activity, i.e., early representative viral proteins (immunoblot data not shown). protein expression and JNK activation, occur simultaneously. Virus treated with gB, gD, and gH neutralizing antibodies To distinguish between the two possibilities presented prior to infection did not induce JNK activity in U2OS cells out above, we assayed JNK activity following infection by vi13, a to 11 h p.i. (Fig. 5A). Nor did incubation of virus with gC virus expressing an ICP4 with a 2-aa insertion at residue 338. antibody prevent JNK induction. We conclude that virus bind- The insertion reduces the DNA binding affinity of ICP4 and ing to the surface of the cell is not sufficient for JNK activation, prevents early and late gene expression, while causing overex- nor is JNK activation seen in nonneutralized viral infections pression of IE proteins (82) (Fig. 6C). Infection of U2OS cells due to a delayed signal cascade initiated at the cell surface. with vi13 resulted in a twofold increase in JNK activity com- Several virion proteins have additional functions upon en- pared to wild-type virus (Fig. 6B), indicating that expression of tering a newly infected cell. To determine if any of these early or late proteins was not necessary for JNK activation. proteins were responsible for inducing JNK, we compared Based on this result and the fact that the timing of JNK acti- mock-infected cell extracts to extracts from cells infected with vation was delayed relative to the binding and penetration of either wild-type virus or UV-irradiated virus over a time course the virus into the cell (Fig. 4A), the activation of JNK is likely of infection. The inoculum of irradiated virus used in this a result of the expression of one or more IE proteins. Addi- experiment had a titer that was reduced greater than 3 logs tionally, if one or more IE proteins are responsible for activa- relative to the control inoculum (data not shown). The exper- tion, then the increased JNK activation observed in vi13-in- iment presented in Fig. 5B shows that in U2OS cells, irradiated fected cells could be explained by the overexpression of IE virus, at best, minimally activated JNK. The magnitude of proteins. Immunoblot analysis confirmed that vi13-infected activation was much less than that seen in infections with cells did not express representative early or late proteins, but unirradiated virus. Identical results were seen in repeat exper- did overexpress IE proteins, e.g., ICP0 and ICP4 (Fig. 6C). iments using Vero and C33 cells (data not shown). It is be- Activation of JNK enhances viral replication. Having docu- lieved that UV-irradiated virus is impaired in its ability to mented activation of JNK following infection, we wished to VOL. 73, 1999 ACTIVATION OF JNK BY HSV 8421

FIG. 7. Preventing JNK from activating nuclear targets decreases the effi- ciency of viral replication. (A) Two stably transfected NIH 3T3 cell lines, G13.5 FIG. 6. JNK activation is dependent upon viral IE gene expression. (A) (control cell line) and G14.1 (JIP-1 overexpressing cell line), were isolated as Cell-equivalent amounts of wild-type KOS-infected U2OS cellular lysates pre- described in Materials and Methods and compared for their ability to support pared for JNK kinase assays were run on 12% polyacrylamide gels. Proteins were GAL-cJUN-dependent CAT activity. Cells were transfected with Lipofectamine transferred to a PVDF membrane and probed for viral proteins ICP4, ICP8, and reagent with 2.5 ␮gof5ϫGAL-CAT and 500 ng of GAL-cJUN and then infected gC. (B) An in vitro solid substrate kinase assay was performed with GST-cJUN with KOS (MOI ϭ 5) 24 h later. Eighteen hours p.i., the cells were harvested, and as described in Materials and Methods with lysates prepared at the indicated CAT assays were performed as described in Materials and Methods. (B) G13.5 ϭ times p.i. from wild-type KOS- and vi13-infected cells (MOI 5). The kinase and G14.1 cells were mock infected or infected with KOS (MOI ϭ 5) for 8 h. data points were derived by scanning the autorads on an LKB laser densitometer. Cells were harvested, and fractionated extracts were prepared. Equal amounts of (C) Immunoblot analysis of wild-type and vi13 ICPs was with the same antibodies protein from each fraction were run on 12% polyacrylamide gels. Proteins were as in panel A, with the addition of ICP0. transferred to a PVDF membrane and probed for JNK. MC, mock-infected cytoplasm; MN, mock-infected nucleus; IC, KOS-infected cytoplasm; IN, KOS- infected nucleus. (C) The two cell lines were infected with HSV at an MOI of 5. Cells and medium were harvested at the times indicated and frozen and thawed know if its activation was important for viral replication. To four times. Titers of aliquots were determined on Vero cell monolayers. address this question, we relied on the recently described in- hibitor of JNK, JIP-1 (JNK-interacting protein-1). JIP-1 acts as a cytoplasmic scaffolding protein that coordinates the trans- duction of a signal through the SAPK pathway from the level trol cells. In one such cell line that stably contains the of MAPKKKs to JNK (93). JIP-1 binds exclusively to JNK and pcDNA3-JIP-1 vector, G14.1, GAL-cJUN-dependent CAT ac- not to other MAPKs, and it binds with a 100-fold greater tivity was decreased in both mock (basal) and infected (in- affinity to JNK than JNK binds to any of its substrates (17). duced) extracts compared to that in extracts from an empty Overexpression of JIP-1 has been shown to effectively inhibit vector control cell line, G13.5 (Fig. 7A). To verify that JNK JNK from translocating into the nucleus (independent of was inhibited from translocating into the nucleus, we fraction- JNK’s activation state) which prevents JNK-mediated gene ated mock- and virus-infected G13.5 and G14.1 cells and ran expression (93). extracts on gels to be used in immunoblotting for JNK. JNK We created stable NIH 3T3 cell lines containing the was primarily found in the cytoplasm of mock-infected (MC) pcDNA3-JIP-1 expressing vector by G418 selection (see Ma- G13.5 and G14.1 cells, while very little JNK was detected in the terials and Methods). In the absence of an antibody to JIP-1 to nucleus (MN). Upon infection, JNK was found in both the determine the expression level of JIP-1, we performed two cytoplasm (IC) and nucleus (IN) of G13.5 cells, but JNK was functional assays to prove that JIP-1 was acting to inhibit JNK. still retained in the cytoplasm of G14.1 cells (Fig. 7B). Solid- First we performed GAL-cJUN-dependent CAT assays under phase GST-cJUN-dependent kinase assays of extracts from the premise that since JIP-1 inhibits JNK-mediated gene ex- G13.5 and G14.1 cells indicated that JNK activation occurred pression, GAL-cJUN-dependent CAT activity should be de- equally in both cell lines following infection (data not shown). creased in JIP-1-overexpressing cells compared to that in con- These results are consistent with the known properties of 8422 MCLEAN AND BACHENHEIMER J. VIROL.

JIP-1. They also confirm that G14.1 cells are expressing JIP-1 strated that TRE-binding proteins but not CRE-binding pro- so that JNK is unable to translocate into the nucleus and teins accumulated (100). The latter study also showed that activate its targets following infection. Single-step growth while TRE-binding complexes increased, the relative propor- curves for G13.5 and G14.1 cell lines demonstrated that inhib- tion of those complexes that contain cJUN decreased. These iting the ability of JNK to activate its targets decreased viral results are not in contradiction to our results, since they as- yield by approximately 70% (Fig. 7C). sayed DNA-binding activity of these transcription factors, while we assayed their trans-activation property. AP-1 is a DISCUSSION complex transcription factor—it can be composed of several different proteins that can bind to TREs and can differ in their HSV induces cJUN and ATF2 trans-activation. Utilizing a ability to activate transcription (46). Additionally, phosphory- GAL-CAT reporter system, we have shown that the TADs of lation at specific sites on these proteins can enhance trans- cJUN and ATF2 were substantially activated following infec- activation without changing DNA-binding or dimerization tion by HSV. This experimental approach eliminated any dif- properties (46). An important consideration in this regard is ferences in DNA binding that the transcription factors might that changes in DNA-binding activity do not mirror the regu- exhibit and allowed us to indirectly assess the activity of up- lation of transcriptional activity (46). In order to measure the stream kinases. cJUN and ATF2 TADs are activated by ability of AP-1 (and its constituents) to activate transcription, SAPKs—cJUN by JNK and ATF2 by a p38 family member the trans-activation ability of those complexes must be mea- (strongly) and by JNK (weakly) (33, 55, 73, 86). Although sured with an appropriate reporter construct (46). Therefore, ELK1 transcriptional activity can be induced by JNK or p38 as even though the population of cJUN-containing AP-1 com- well as ERK (7, 9, 28, 29, 32, 94, 96–98), it is clear that JNK is plexes may be decreasing, the remaining cJUN is being acti- not activating ELK1 under our experimental conditions. HSV vated by JNK. Likewise, even though CRE-binding proteins do may be activating pathways that specifically target certain JNK not increase after infection of BHK cells (100), the activity of and p38 isoforms which in turn activate specific TADs. Three these proteins can be induced (Fig. 1B). Additionally, we can- JNK genes have been identified, giving rise to at least 10 not rule out the possibility that other JUN family members, alternatively spliced mRNAs, and different JNK isoforms have such as JunB and JunD, which can bind to or be phosphory- different specificities and affinities for their respective sub- lated by JNK, respectively (45), are phosphorylated or mediate strates (34). Additionally, there are at least four different p38 the phosphorylation of dimerization partners after infection isoforms which are differentially activated by stress stimuli (13, and may be the true relevant targets in the infected cell. 21, 73, 88). The activities of all other cellular transcription RAS is a G-protein shown to be a key mediator of signaling factors we assayed decreased or were, at best, unchanged. The in many pathways (6, 87). There are pathways, however, that decrease in GAL-RELA-dependent CAT activity following in- do not involve RAS, and some downstream pathways are reg- fection agrees with a previous report that showed that NF-␬B, ulated by RAS only under certain conditions (5, 6, 61, 62, 87). while translocated to the nucleus following infection, did not cJUN-dependent transcription can be weakly upregulated ei- demonstrate an increased ability to induce trans-activation ther directly by activated RAS or indirectly by derepression of (68). At this time, we do not know if HSV is specifically inhib- pathways normally downregulated by active RAS, as seen in iting the activity of these factors or if the decreases are an Fig. 2B for the mock-infected extracts. The inability of either indirect effect of infection. dominant-negative RAS or constitutively active CDC25 to af- HSV infection has been shown to result in increased AP-1 fect the level of GAL-cJUN- or GAL-ATF2-dependent CAT activity (42). We used a CAT reporter containing two consen- activity in HSV-infected cells suggests that HSV is capable of sus AP-1 binding sites in the promoter region to confirm this stimulating cJUN and ATF2 TADs by a RAS-independent finding (Fig. 1B). AP-1 can be composed of several different mechanism(s). Bypassing any requirement for the upregulation JUN and FOS family members that all bind to consensus AP-1 of RAS is in line with a study which showed that activated RAS sites, called TREs (3, 46). Since cFOS is not activated follow- actually antagonized efficient viral replication. Infection of ing infection, but cJUN and ATF2 are, it is likely that increased RAS-transformed cells yielded less virus than nontransformed TRE-mediated trans-activation is a result of activated cJUN in cells of the same cell type (26). homodimers or in heterodimers with existing cFOS. Also of HSV activation of JNK is dependent on IE gene expression. note, the study by Tognon et al. (85) demonstrated downregu- To test the activities of different MAPKs, we performed im- lation of cFOS expression by ICP4 which would result in de- mune complex kinase assays to detect ERK activity and solid- creased JUN-cFOS activity in favor of JUN-JUN and JUN- phase kinase assays to detect JNK activity. For the latter, we ATF2 activity. Additionally, cJUN is capable of forming have used a GST-cJUN (1 to 79 aa) construct to bind endog- heterodimers with ATF2 (3, 46). This heterodimer binds to enous JNK and act as its substrate. Results from these exper- sites called CREs that differ from TREs by the addition of 1 bp iments indicated a 6- to 20-fold elevation, depending on the in the intervening sequence between the dyad repeats (3, 46). cell type, in JNK kinase activity compared to that of mock We observed a cell-type-specific increase in CRE-dependent samples (Fig. 3A). We found no evidence for induction of JNK CAT activity after HSV infection (Fig. 1B). at any time within the first hour of infection, corresponding to Activation of both cJUN and ATF2 could synergistically when virus is binding to cell surface receptors. Activation typ- activate to very high levels the expression of genes that contain ically began 3 to 4 h p.i., increased out to 8 h p.i., and remained CREs in their promoters, such as the c-jun gene itself. It has elevated out to at least 14 h p.i. (Fig. 4A, 5, and 6B). Since we been shown that c-jun transcripts are induced following infec- obtained identical results if we specifically immunoprecipitated tion consistent with cJUN-ATF2 activation (42, 47). In the for JNK, in addition to the fact that the responsible kinase study of Jang et al. (42) it was not possible to determine what must be able to bind to the N terminus of cJUN and only kind of probe was used to study AP-1 activation, but since most phosphorylate Ser63 and Ser73, we conclude that we were of the assays used sequences derived from the c-jun promoter detecting bona fide JNK activity and not another cellular ki- which contains CREs, we conclude that their gel shifts reflect nase or a viral kinase with JNK-like properties. increased CRE binding following infection. This interpreta- Using a GST-ATF2 construct in solid-phase kinase assays, tion, however, is in conflict with a recent report which demon- we detected kinase activity that occurred with identical kinetics VOL. 73, 1999 ACTIVATION OF JNK BY HSV 8423

but a lower magnitude of activity than that detected with GST- which can bind but not enter cells, argue against this mecha- cJUN as a substrate (Fig. 4A). Since the ATF2 TAD can be nism. phosphorylated by both JNK and p38, we attempted to deter- Infection with UV-irradiated virus failed to induce JNK to mine if p38 was activated by HSV. We took advantage of an high levels. We believe the inability to activate JNK was due to antibody that recognizes a phosphorylated form of p38 that the inability of irradiated virus to express viral genes; however, correlates with its being in an active state. We observed that we could not discount potential cross-linking of virion proteins p38 is activated with identical kinetics to that seen for the as being responsible for the lack of JNK activity. We therefore activation of JNK (compare Fig. 4B to 4A). MKK4 is activated turned to a temperature-sensitive mutant which infects cells at by HSV infection (100). Since MKK4 is a MAPKK that can the restrictive temperature, delivers functional VP16 to the activate both JNK and p38, it is possible that HSV is activating nucleus (3a), but does not release viral DNA from capsids into MKK4 to coordinate the activation of both SAPKs. the nucleus (4). Viral genes, therefore, are not expressed in a In contrast, over an 11-h-time-course experiment, we did not tsB7 infection at the restrictive temperature (4, 51), but tegu- detect any change in ERK activity (Fig. 4B). The binding of ment proteins are free to function normally. We failed to growth factors to their cognate receptors and other growth- detect high levels of JNK activity after infection by tsB7 at the promoting signals typically leads to ERK activation, which is restrictive temperature. We did note that in UV-irradiated virus extracts and tsB7 samples from 39.5°C, there was a slight critical for inducing the expression of certain G1 and G1/S phase-promoting proteins such as cyclin D (1a, 56, 67, 69, 75). elevation in JNK activity over mock samples (approximately There are many examples drawn from studies of DNA viruses, twofold). It is possible that this slight activation was due to the some of which are known to cause cell cycle progression, where function of a tegument protein, such as VP16 (100). Alterna- viral proteins cause activation of the ERK pathway. The E5 tively, it could reflect a low level of activation in the small protein of bovine papillomavirus inhibits epidermal growth population of cells infected by less-damaged virus in UV-irra- factor (EGF) receptor (EGFR) and colony-stimulating factor diated stocks or leaky tsB7 virus at the restrictive temperature, receptor down-regulation, leading to a persistence of activated mimicking an infection at a low MOI. We have noted that receptors at the cell surface (60). Vaccinia virus encodes a activation is MOI sensitive, so that at an MOI of 0.5, the level virus growth factor that mimics EGF’s ability to stimulate of JNK activation is very low (unpublished observation). These EGFR (50). ERK molecules are actually incorporated into the results suggest that in order for HSV to significantly induce human immunodeficiency virus type 1 virion, and ERK activity JNK activity, the virus must enter the cell and express viral is necessary for the disengagement of the reverse transcription genes. complex from the cell membrane and its translocation into the Recently, JNK and p38 were reported to be activated after HSV infection by a VP16-dependent mechanism (100). Com- nucleus (41). The herpesvirus saimiri protein STP-C488 asso- parisons of mock- and VP16 mutant in1814-infected cells in ciates with RAS (43) and increases its GTP/GDP ratio, leading that report were made difficult because of the very high back- to constitutive activation of ERK (44). Small T of ground (mock-infected) levels of JNK activity. While we inter- simian virus 40 binds to and prevents protein phosphatase 2A pret these results as at best arguing for a role for VP16 in JNK from dephosphorylating ERK2 and MEK, which prolongs activation, it is also possible that the defects in IE gene expres- their activities (83). The inability to detect ERK activation sion in in1814-infected cells may contribute to the failure to following HSV infection correlates with other evidence that detect JNK activation (1). Although the experiments with HSV infection fails to promote cell cycle progression (14, 19). in1814 were performed at a high MOI, it has not been directly Whether the failure to promote progression is related to the demonstrated that this condition of infection restores IE gene ability of HSV to encode most of its DNA replication machin- expression to normal levels. The delayed activation kinetics of ery, thus precluding the necessity to establish a true S phase, JNK (3 to 4 h p.i.) observed by us and Zachos et al. (100) also remains to be determined. argue for at best an indirect role for VP16 and more directly a Signals that lead to MAPK activation are often generated at role for one or more IE proteins in the mechanism of activa- the cell surface when receptors bind to their appropriate li- tion. An additional experimental approach by these authors, in gand. Some studies have shown that when virus binds to a cell, which transfection of wild-type VP16 and TAD-deleted VP16 it may do so by binding to a cell surface receptor, and that expression vectors resulted in elevated JNK activity, did not interaction may mimic normal ligand-receptor events to acti- take into account the possibility that overexpressed protein vate downstream pathways. Binding of simian immunodefi- may in and of itself have resulted in a cellular stress response. ciency virus (SIV) to cells activates ERK, JNK, and/or p38 We were able to determine that expression of one or more depending upon the cell type, and hence, the receptor to which IE proteins was necessary for JNK activation. Infection with SIV binds (71). Cytomegalovirus has been shown to generate vi13, an ICP4 mutant virus that only expresses IE proteins intracellular signals upon binding to the surface of permissive (reference 82 and this report), led to activation of JNK with ␬ cells that induce the translocation and activation of NF- B (79, identical kinetics to but greater magnitude than that of wild- 99). One reported cell surface receptor for HSV, HVEM or type virus, indicating that neither expression of early and late HveA, is a member of the tumor necrosis factor receptor genes nor viral DNA replication is necessary for activating (TNFR) family (63, 91). Activation of this receptor in overex- JNK. Supporting this conclusion was the observation that vi13- pression studies leads to the generation of a signal which up- infected cells overexpress IE proteins, which likely explains the regulates AP-1 activities (59). This upregulation, though not increased level of JNK activity compared to that of wild-type- directly shown, is presumably due to induction of JNK, as many infected cells. Some of our current studies are attempting to TNFR family members have been shown to induce JNK lead- determine the level of JNK activity following infection with ing to upregulated AP-1 activity (65, 74, 78). Even though the viruses that are deleted for each viral IE gene. Due to func- cells we used do not express HVEM, HSV does bind to other tional and physical interactions among different IE proteins, cellular receptors (27), and we predicted that binding to one or we may also resort to using viruses that have multiple deletions more of these receptors may lead to MAPK activation. How- of IE genes in order to determine all of the relevant viral ever, both the timing of JNK activation and the absence of proteins. induction following infection by antibody-neutralized HSV, The above result is somewhat at odds with a similar exper- 8424 MCLEAN AND BACHENHEIMER J. VIROL. iment reported by Zachos et al. (100) in which infection with demonstrated to serve an essential role in viral gene expression the ICP4 mutant tsK at the restrictive temperature did not (76). We also cannot rule out the possibility that JNK could be induce increased JNK activation over wild-type virus. The in- targeting viral trans-activators ICP0, -4, and -27 to regulate ability to detect a difference by using tsK may relate to the their roles in the temporal regulation of viral gene expression. small fold induction in JNK activity that these authors were Alternatively, activation of SAPKs could be targeting cellular able to detect with wild-type virus. While they noted a maxi- activities and upregulating cellular genes that would lead to a mum 4-fold increase in JNK activity over mock, we were able more conducive intracellular environment for viral replication. to detect a 20-fold induction. We have been careful to modify By blocking the ability of JNK to act on its nuclear sub- our infection and harvest protocols to minimize transient ac- strates, we showed that virus yield is decreased by 70% (Fig. tivation of JNK which could have the effect of increasing back- 7C), suggesting that HSV requires the action of this cellular ground levels. Regardless, small differences in the level of JNK activity for efficient replication. While the decrease in viral activation, such as the 2-fold increase by vi13 over KOS would yield in JIP-1 expressing G14.1 cells is modest, several points be much easier to detect when KOS is activating JNK 20-fold. should be noted. First, JIP-1 overexpression did not completely Activation of JNK provides a beneficial activity for viral abrogate the ability of JNK to activate cJUN-dependent trans- replication. HSV infection causes many changes to the in- activation (Fig. 7A). Additionally, the activation of JNK in fected cell, such as halting cellular macromolecular synthesis, NIH 3T3 cells following infection was weaker than in other cell reorganizing the nucleus, and causing cells to round up before types, so viral replication may not be as dependent upon JNK they are ultimately lysed as a result of the release of viral activity in NIH 3T3 cells as in other cell types, where JNK is progeny (76). In this report, we describe an additional event, activated to higher levels. Finally much of the viral genome is the activation of SAPKs. Neither the mechanism for activation devoted to increasing the efficiency of replication in different nor its role in the infectious process of HSV is currently known. cell types (76, 89), and we would not expect that inhibiting one The effect of JNK activation after infection could be complex cellular activity would completely block viral replication. A key and cell-type specific, since activation of JNK under different finding reported here, however, is that blocking a specific cel- conditions in different cell types correlates with proliferation, lular activity (the ability of JNK to activate its substrates) has oncogenic transformation, or apoptosis (40, 92). Previously, had a detrimental effect on replication. Patel et al. (68) re- only four reports have described induction of JNK activity ported that blocking the infection-mediated translocation into following virus infection, and the significance of these induc- the nucleus of another cellular transcription factor, NF-␬B, tions is also unknown (20, 71, 81, 100). reduced viral yield by 90%. The activity of E2F transcription JNK activation could represent a nonspecific or even an factors is reduced following infection by HSV as a result of an antiviral response by the cell, but several factors argue against increase in the association of these factors with regulatory this interpretation. First, our control experiments using mock- pocket proteins, such as pRB and p107 (39). Our findings in infected cells or cells infected with neutralized, UV-irradiated conjunction with these reports suggest that a number of dif- virus or tsB7 at the restrictive temperature demonstrate that ferent cellular activities which target transcription factors are JNK activation is not a nonspecific response to the mechanics being differentially manipulated in the infected cell and that of our infection procedure, due to growth factors or cytokines these activities, in total, are significant for the overall efficiency in the inoculum, or related to early events occurring in lytic of virus replication. infection. Additionally, the onset of JNK activation occurs 3 to 4 h p.i. By this time, virion components, such as VP16, have ACKNOWLEDGMENTS had ample time to perform their ascribed functions, and IE proteins have already reached high levels of accumulation. If T.I.M. was supported by NIGMS T32 GM 07092. These studies were the cells were responding to the presence, accumulation, or supported by PHS Program Project grant CA 19014 to S.L.B. We thank Albert Baldwin for the GAL-RELA expression construct; function of one or more of these proteins, a response would Shannon Kenney for the GAL-ATF2 expression construct and the likely have occurred sooner. The failure of UV-irradiated virus 3ϫCRECAT reporter construct; Laura Licota for GST-cJUN(AA); or tsB7 at the restrictive temperature to induce JNK activity to and Channing Der for all of the other GAL effector and CAT reporter high levels demonstrated that virion components are incapable constructs, GST-cJUN(WT), and GST-ATF2. We thank Roger Davis of, or at least not sufficient for, activating JNK (Fig. 5). It is for helpful discussions and the gift of the JIP-1-expressing construct. possible though that VP16, for example, could minimally in- We also thank Ginger L. Ehmann for critical review of the manuscript duce JNK, as seen in UV-irradiated samples and tsB7 samples and Robert Johnson for assistance in probing immunoblots for total at 39.5°C. Additionally, no early or late gene expression nor and activated p38. any activity involved in viral DNA replication is necessary for REFERENCES the activation of JNK in HSV-infected cells, as judged by the 1. Ace, C. I., T. A. McKee, J. M. Ryan, J. M. Cameron, and C. M. Preston. results with the ICP4 mutant, vi13 (Fig. 6B). 1989. Construction and characterization of a herpes simplex virus type 1 In principle, virally encoded proteins could directly target mutant unable to transinduce immediate-early gene expression. J. Virol. and thus alter the activity of cellular transcription factors (e.g., 63:2260–2269. cJUN and ATF2). However, it appears that HSV is targeting 1a.Aktas, H., H. Cai, and G. M. Cooper. 1997. Ras links growth factor signaling to the cell cycle machinery via regulation of cyclin D1 and the Cdk inhibitor the upstream regulators of these factors, since increases in p27KIP1. Mol. Cell. Biol. 17:3850–3857. their trans-activation function are accompanied by increased 2. Alvarez, E., I. C. Northwood, F. A. Gonzales, D. A. Latour, A. Seth, C. JNK and p38 activity. We propose that expression of one or Abate, T. Curran, and R. J. Davis. 1991. Pro-Leu-Ser/Thr-Pro is a consen- more IE proteins leads to the activation of the JNK pathway. sus primary sequence for substrate protein phosphorylation. J. Biol. Chem. 266:15277–15285. Once activated, JNK would phosphorylate cellular transcrip- 3. Angel, P., and M. Karin. 1991. The role of Jun, Fos, and the AP-1 complex tion factors that could enhance the expression of IE and pos- in cell-proliferation and transformation. Biochim. Biophys. Acta 1072:129– sibly early viral genes. Promoters of IE, early, and late genes 157. contain cognate binding sites for a variety of cellular transcrip- 3a.Batterson, W., and B. Roizman. 1983. Characterization of the herpes sim- ␬ plex virion-associated factor responsible for the induction of ␣ genes. J. Vi- tion factors, e.g., Oct-1, Sp1, AP-1, CAAT, and NF- B. Mu- rol. 46:371–377. tation of these sites affects the efficiency of promoter function, 4. Batterson, W., D. Furlong, and B. Roizman. 1983. Molecular genetics of although no cellular transcription factors per se have been herpes simplex virus. VIII. Further characterization of a temperature-sen- VOL. 73, 1999 ACTIVATION OF JNK BY HSV 8425

sitive mutant defective in release of viral DNA and in other stages of the 31. Gorman, C. M., L. F. Moffat, and B. H. Howard. 1982. Recombinant viral reproductive cycle. J. Virol. 45:397–407. which express chloramphenicol acetyltransferase in mammalian 5. Burgering, B. M. T., and J. L. Bos. 1995. Regulation of Ras-mediated cells. Mol. Cell. Biol. 2:1044–1051. signaling: more than one way to skin a cat. Trends Biochem. Sci. 20:18–22. 32. Gupta, S., and R. J. Davis. 1994. MAP kinase binds to the NH2-terminal 6. Campbell, S. L., R. Khosravi-Far, K. L. Rossman, G. J. Clark, and C. J. activation domain of c-Myc. FEBS Lett. 353:281–285. Der. 1998. Increasing complexity of Ras signaling. 17:1395–1413. 33. Gupta, S., D. Campbell, B. De´rijard, and R. J. Davis. 1995. Transcription 7. Cavigelli, M., F. Dolfi, F.-X. Claret, and M. Karin. 1995. Induction of c-fos factor ATF2 regulation by the JNK signal transduction pathway. Science expression through JNK-mediated TCF/Elk-1 phosphorylation. EMBO J. 267:389–393. 14:1785–1797. 34. Gupta, S., T. Barrett, A. J. Whitmarsh, J. Cavanagh, H. K. Sluss, B. 8. Chen, S. Y., S. Y. Huff, C. C. Lai, C. J. Der, and S. Powers. 1994. Ras-15A De´rijard, and R. J. Davis. 1996. Selective interaction of JNK protein kinase protein shares highly similar dominant-negative biological properties with isoforms with transcription factors. EMBO J. 15:2760–2770. Ras-17N and forms a stable, guanine-nucleotide resistant complex with 35. Hauser, C. A., J. K. Westwick, and L. A. Quilliam. 1995. RAS-mediated CDC25 exchange factor. Oncogene 9:2691–2698. transcription activation: analysis by transient cotransfection assays. Meth- 9. Chuang, C. F., and S. Y. Ng. 1994. Functional divergence of the MAP ods Enzymol. 255:412–426. kinase pathway, ERK1 and ERK2 activate specific transcription factors. 36. Hazzalin, C. A., R. Le Panse, G. Cano, and L. C. Mahadevan. 1998. Ani- FEBS Lett. 346:229–234. somycin selectively desensitizes signalling components involved in stress 10. Clark, G. J., J. K. Westwick, and C. J. Der. 1997. p120 GAP modulates RAS kinase activation and fos and jun induction. Mol. Cell. Biol. 18:1844–1854. activation of JUN kinases and transformation. J. Biol. Chem. 272:1677– 37. Hibi, M., A. Lin, T. Smeal, A. Minden, and M. Karin. 1993. Identification 1681. of an oncoprotein- and UV-responsive protein kinase that binds and po- 11. Crews, C. M., and R. L. Erickson. 1993. Extracellular signals and reversible tentiates the c-Jun activation domains. Genes Dev. 7:2135–2148. protein phosphorylation: what to Mek of it all. Cell 74:215–217. 38. Hill, T. M., J. R. Sadler, and J. L. Betz. 1985. Virion component of herpes 12. Crews, C. M., A. Alessandrini, and R. L. Erickson. 1992. The primary simplex virus type 1 KOS interferes with early shutoff of host protein structure of MEK, a protein kinase that phosphorylates the ERK gene synthesis induced by herpes simplex virus type 2 186. J. Virol. 56:312–316. product. Science 258:478–480. 39. Hilton, M. J., D. Mounghane, T. McLean, N. V. Contractor, J. O’Neil, K. 13. Cuenda, A., P. Cohen, V. Bue´e-Scherrer, and M. Goedert. 1997. Activation Carpenter, and S. L. Bachenheimer. 1995. Induction by herpes simplex of stress-activated protein kinase-3 (SAPK3) by cytokines and cellular virus of free and heteromeric forms of E2F transcription factor. Virology stresses is mediated via SAPKK3 (MKK6); comparison of the specificities 213:624–638. of SAPK3 and SAPK2 (RK/p38). EMBO J. 16:295–305. 40. Ip, Y. T., and R. J. Davis. 1998. Signal transduction by the c-Jun NH2- 14. De Bruyn Kops, A., and D. M. Knipe. 1988. Formation of DNA replication terminal kinase (JNK)—from inflammation to development. Curr. Opin. structures in herpes virus-infected cells requires a viral DNA binding pro- Cell Biol. 10:205–219. tein. Cell 55:857–868. 41. Jacque, J. M., A. Mann, H. Enslen, N. Sharuva, B. Brichacek, R. J. Davis, 15. Dent, P., W. Hasar, T. Haystead, L. Vincent, T. Roberts, and T. Sturgill. and M. Stevenson. 1998. Modulation of HIV-1 infectivity by MAPK, a 1992. Activation of mitogenic activated protein kinase kinase by v-raf in virion-associated kinase. EMBO J. 17:2607–2618. NIH3T3 cells and in vitro. Science 254:1404–1407. 42. Jang, K.-L., B. Pulverer, J. R. Woodgett, and D. S. Latchman. 1991. Acti- 16. De´rijard, B., M. Hibi, I.-H. Wu, T. Barrett, B. Su, T. Deng, M. Karin, and vation of the cellular transcription factor AP-1 in herpes simplex virus R. Davis. 1994. JNK1: a protein kinase stimulated by UV light and Ha-Ras infected cells is dependent on the viral immediate-early protein ICP0. that binds and phosphorylates the c-Jun activation domain. Cell 76:1025– Nucleic Acids Res. 19:4879–4883. 1037. 43. Jung, J. U., J. J. Trimble, N. W. King, B. Biesinger, B. W. Fleckenstein, and 17. Dickens, M., J. S. Rogers, J. Cavanagh, A. Raitano, Z. Xia, J. R. Halpern, R. C. Desrosiers. 1991. Identification of transforming genes of subgroup A M. E. Greenberg, C. L. Sawyers, and R. J. Davis. 1997. A cytoplasmic and C strains of herpesvirus saimiri. Proc. Natl. Acad. Sci. USA 88:7051– inhibitor of the JNK signal transduction pathway. Science 277:693–696. 7055. 18. Egan, S. E., and R. A. Weinberg. 1993. The pathway to signal achievement. 44. Jung, J. U., and R. C. Desrosiers. 1995. Association of viral oncoprotein, Nature 365:781–783. STP-C488, with cellular ras. Mol. Cell. Biol. 15:6506–6512. 19. Ehmann, G. L., T. I. McLean, S. Vidyarti, and S. L. Bachenheimer. Un- 45. Kallunki, T., T. Deng, M. Hibi, and M. Karin. 1996. c-Jun can recruit JNK published data. to phosphorylate dimerization partners via specific docking interactions. 20. Eliopoulos, A. G., and L. S. Young. 1998. Activation of cJun N-terminal Cell 87:929–939. kinase (JNK) by the Epstein-Barr virus-encoded latent membrane protein 46. Karin, M. 1996. The regulation of AP-1 activity by mitogen-activated pro- 1 (LMP1). Oncogene 16:1731–1742. tein kinases. Philos. Trans. R. Soc. Lond. B Biol. Sci. 351:127–134. 21. Enslen, H., J. Raingeaud, and R. J. Davis. 1998. Selective activation of p39 47. Karin, M., Z.-G. Liu, and E. Zandi. 1997. AP-1 function and regulation. mitogen-activated protein (MAP) kinase isoforms by the MAP kinase ki- Curr. Opin. Cell Biol. 9:240–246. nase MKK3 and MKK6. J. Biol. Chem. 273:1741–1748. 48. Khosravi-Far, R., and C. J. Der. 1994. The Ras signal transduction path- 22. Faris, M., N. Kokot, K. Latinis, S. Dasibhatla, D. R. Green, G. A. Koretzky, way. Cancer Metastasis Rev. 13:67–89. and A. Nel. 1998. The c-Jun N-terminal kinase cascade plays a role in 49. Khosravi-Far, R., M. A. White, J. K. Westwick, P. A. Solski, M. Chrza- stress-induced apoptosis in Jurkat cells by upregulating Fas ligand expres- nowska-Wodnicka, L. Van Aelst, M. H. Wigler, and C. J. Der. 1996. On- sion. J. Immunol. 160:134–144. cogenic Ras activation of Raf/mitogen-activated protein kinase-indepen- 23. Fuller, A. O., and P. G. Spear. 1987. Anti-glycoprotein D antibodies that dent pathways is sufficient to cause tumorigenic transformation. Mol. Cell. prevent adsorption but block infection by herpes simplex virus 1 prevent Biol. 16:3923–3933. virion-cell fusion at the cell surface. Proc. Natl. Acad. Sci. USA 84:5454– 50. King, C. S., J. A. Cooper, B. Moss, and D. R. Twardzik. 1986. Vaccinia virus 5458. growth factor stimulates tyrosine protein kinase activity of A431 cell epi- 24. Fuller, A. O., R. E. Santos, and P. G. Spear. 1989. Neutralizing antibodies dermal growth factor receptors. Mol. Cell. Biol. 6:332–336. specific for glycoprotein H of herpes simplex virus permit viral attachment 51. Knipe, D. M., W. Batterson, C. Nosal, B. Roizman, and A. Buchan. 1981. to cells but prevent penetration. J. Virol. 63:3435–3443. Molecular genetics of herpes simplex virus. VI. Characterization of a tem- 25. Galaktionov, K., X. Chen, and D. Beach. 1996. Cdc25 cell-cycle phospha- perature-sensitive mutant defective in the expression of all early viral gene tase as a target of c-myc. Nature 382:511–517. products. J. Virol. 38:539–547. 26. Garcin, D., Y. Michal, F. Jault, M. Lyon, G. Lenoir, and B. Jacquemont. 52. Kyriakis, J. M., J. R. Woodgett, and J. Avruch. 1995. The stress-activated 1990. Inhibition of HSV-1 multiplication in rat embryo fibroblasts consti- protein kinases. A novel ERK subfamily response to cellular stress and tutively expressing the EJ-ras oncogene. Virology 179:208–216. inflammatory cytokines. Ann. N. Y. Acad. Sci. 766:303–319. 27. Geraghty, R. J., C. Krummenacher, G. H. Cohen, R. J. Eisenberg, and R. G. 53. Lee, F. S., J. Hagler, Z. J. Chen, and T. Maniatis. 1997. Activation of the Spear. 1998. Entry of alphaherpesviruses mediated by poliovirus receptor- I␬B␣ kinase complex by MEKK1, a kinase of the JNK pathway. Cell related protein 1 and poliovirus receptor. Science 280:1618–1620. 88:213–222. 28. Gille, H., A. Sharrocks, and P. Shaw. 1992. Phosphorylation of p62TCF by 54. Lillie, J. W., and M. R. Green. 1989. Transcription activation by the ade- MAP kinase stimulates ternary complex factor formation at c-FOS pro- novirus E1A protein. Nature 338:39–44. moter. Nature 358:414–417. 55. Livingstone, C., G. Patel, and N. Jones. 1995. ATF-2 contains a phospho- 29. Gille, H., M. Kortenjann, O. Thamas, C. Moomaw, C. Slaughter, M. H. rylation-dependent transcriptional activation domain. EMBO J. 14:1785– Cobb, and P. E. Shaw. 1995. ERK phosphorylation potentiates Elk-1- 1797. mediated ternary complex formation and transactivation. EMBO J. 14:951– 56. Lukas, J., J. Bartkova, and J. Bartek. 1996. Convergence of mitogenic 962. signalling cascades from diverse classes of receptors at the cyclin D–cyclin- 30. Gimble, J. M., E. Duh, J. M. Ostrove, H. E. Gendelman, E. E. Max, and dependent kinase–pRb-controlled G1 checkpoint. Mol. Cell. Biol. 16:6917– A. B. Rabson. 1988. Activation of the human immunodeficiency virus long 6925. terminal repeat by herpes simplex virus type 1 is associated with induction 57. Malinin, N. L., M. P. Boldin, A. V. Kovalenko, and D. Wallach. 1997. of a nuclear factor that binds to the NF␬B/core enhancer sequence. J. Virol. MAP3K-related kinase involved in NF␬B induction by TNF, CD95 and 62:4104–4112. IL-1. Nature 385:540–544. 8426 MCLEAN AND BACHENHEIMER J. VIROL.

58. Marias, R., J. Wynne, and R. Treismann. 1993. The SRF accessory protein by CD40 on human B cells. J. Biol. Chem. 270:30823–30828. Elk-1 contains a growth factor regulated transcription domain. Cell 73:381– 79. Sambucetti, L. C., J. M. Cherrington, G. W. Wilkinson, and E. S. Mocarski. 393. 1989. NF␬B activation of the cytomegalovirus enhancer is mediated by a 59. Marsters, S. A., T. M. Ayers, M. Skubatch, C. L. Gray, M. Rothe, and A. viral transactivator and by T cell stimulation. EMBO J. 8:4251–4258. Ashkenazi. 1997. Herpesvirus entry mediator, a member of the tumor 80. Schmitz, M. L., M. A. dos Santos Silva, and P. A. Baererle. 1995. Trans- ␬ necrosis factor receptor (TNFR) family, interacts with members of the activation domain 2 (TA2) of p65 NF- B. J. Biol. Chem. 270:15576–15584. TNFR-associated factor family and activates the transcription factors 81. See, R. H., and Y. Shi. 1998. Adenovirus E1B 19,000-molecular-weight NF␬B and AP-1. J. Biol. Chem. 272:14029–14032. protein activates c-Jun N-terminal kinase and c-Jun-mediated transcription. 60. Martin, P., W. C. Vass, J. T. Schiller, D. R. Lowry, and T. J. Velu. 1989. The Mol. Cell. Biol. 18:4012–4022. bovine papillomavirus E5 transforming protein can stimulate the trans- 82. Shephard, A. A., and N. A. DeLuca. 1991. Activities of heterodimers com- forming activity of EGF and CSF-1 receptors. Cell 59:21–32. posed of DNA-binding- and transactivation-deficient subunits of the herpes 61. Minden, A., A. Lin, M. McMahon, C. Lange-Carter, B. De´rijard, R. J. simplex virus regulatory protein ICP4. J. Virol. 65:299–307. Davis, G. L. Johnson, and M. Karin. 1994. Differential activation of ERK 83. Sontag, E., S. Fedorov, C. Kamibayashi, D. Robbins, M. Cobb, and M. and JNK mitogen-activated protein kinases by Raf-1 and MEKK. Science Mumby. 1993. The interaction of SV40 small tumor antigen with protein 266:1719–1723. phosphatase 2A stimulates the MAP kinase pathway and induces cell pro- 62. Minden, A. J., A. Lin, F. X. Claret, A. Abo, and M. Karin. 1995. Selective liferation. Cell 75:887–897. activation of the JNK signaling cascade and c-Jun transcriptional activity by 84. Spear, P. Personal communication. the small GTPase Rac and Cdc42Hs. Cell 81:1147–1157. 85. Tognon, M., M. G. Romanelli, E. M. Cattozzo, C. Bovolenta, and E. Liboi. 63. Montgomery, R. I., M. S. Warner, B. J. Lum, and P. G. Spear. 1996. Herpes 1993. C-fos proto-oncogene transient transcription is negatively affected in simplex virus-1 entry into cells mediated by a novel member of the TNF/ the ELa4-2 transformed cell line. Pathobiology 61:288–292. NGF receptor family. Cell 87:427–436. 86. van Dam, H., D. Wilhelm, I. Herr, A. Steffen, P. Herrlich, and P. Angel. 64. Morooka, T., and E. Nishida. 1998. Requirement of p38 mitogen-activated 1995. ATF-2 is preferentially activated by stress-activated protein kinases to protein kinase for neuronal differentiation in PC12 cells. J. Biol. Chem. mediate c-jun induction in response to genotoxic agents. EMBO J. 14:1798– 273:24285–24288. 1811. 65. Natoli, G., A. Costanzo, A. Ianni, D. J. Templeton, J. R. Woodgett, C. 87. Vojtek, A. B., and C. J. Der. 1998. Increasing complexity of the RAS Balsano, and M. Levrerop. 1997. Activation of SAPK/JNK by TNF receptor signaling pathway. J. Biol. Chem. 273:19925–19928. 1 through a noncytotoxic TRAF2-dependent pathway. Science 275:200– 88. Wang, X. S., K. Deiner, C. L. Manthey, S.-W. Wang, B. Rosenzweig, J. Bray, 203. J. Delaney, C. N. Cole, P.-Y. Chan-Hui, N. Mantlo, H. S. Lichenstein, M. 66. Nicola, A. V., M. Ponce de Leon, R. Xu, W. Hou, J. C. Whitbeck, C. Zukowski, and Z. Yao. 1997. Molecular cloning and characterization of a Krummenacher, R. I. Montgomery, P. G. Spear, R. J. Eisenberg, and G. H. novel p38 mitogen-activated protein kinase. J. Biol. Chem. 272:23668– Cohen. 1998. Monoclonal antibodies to distinct sites on herpes simplex 23674. virus (HSV) glycoprotein D block HSV binding to HVEM. J. Virol. 72: 89. Ward, P. L., and B. Roizman. 1994. Herpes simplex genes: the blueprint of 3595–3601. a successful human pathogen. Trends Genet. 10:267–274. 67. Para, M. F., M. L. Parish, A. G. Noble, and P. G. Spear. 1985. Potent 90. Westwick, J. K., A. D. Cox, C. J. Der, M. H. Cobb, M. Hibi, M. Karin, and neutralizing activity associated with anti-glycoprotein D specificity among D. A. Brenner. 1994. Oncogenic RAS activates c-JUN via a separate path- monoclonal antibodies selected for binding to herpes simplex virions. J. Vi- way from the activation of extracellular signal-regulated kinases. Proc. Natl. rol. 55:483–488. Acad. Sci. USA 91:6030–6034. 68. Patel, A., J. Hanson, T. I. McLean, J. Olgiate, M. Hilton, W. E. Miller, and 91. Whitbeck, J. C., C. Peng, H. Lou, R. Xu, S. H. Willis, M. Ponce De Leon, T. S. L. Bachenheimer. 1998. Herpes simplex virus type 1 induction of persis- Peng, A. V. Nicola, R. I. Montgomery, M. S. Warner, A. M. Soulika, L. A. tent NF-␬B nuclear translocation increases the efficiency of virus replica- Spruce, W. T. Moore, J. D. Lambris, P. G. Spear, G. H. Cohen, and R. J. tion. Virology 247:212–222. Eisenberg. 1997. Glycoprotein D of herpes simplex virus (HSV) binds 69. Peeper, D. S., T. M. Upton, M. H. Ladha, E. Neuman, J. Zalvide, R. directly to HVEM, a member of the tumor necrosis factor receptor super- Bernards, J. A. DeCaprio, and M. E. Ewen. 1997. Ras signaling linked to the family and a mediator of HSV entry. J. Virol. 71:6083–6093. cell-cycle machinery by the . Nature 386:177–181. 92. Whitmarsh, A. J., and R. J. Davis. 1996. Transcription factor AP-1 regu- 70. Perona, R., S. Montaner, L. Sangier, I. Sa´nchez-Pe´rez, R. Bravo, and J. C. lation by mitogen-activated protein kinase signal transduction pathways. J. Lacal. 1997. Activation of the nuclear factor-␬B by Rho, CDC42, and Rac-1 Mol. Med. 74:589–607. proteins. Genes Dev. 11:463–475. 93. Whitmarsh, A. J., J. Cavanagh, C. Tournier, J. Yasuda, and R. J. Davis. 71. Popik, W., and P. M. Pitha. 1998. Early activation of mitogen-activated 1998. A mammalian scaffold complex that selectively mediates MAP kinase protein kinase kinase, extracellular signal-regulated kinase, p38 mitogen- activation. Science 281:1671–1674. activated protein kinase, and c-Jun N-terminal kinase in response to bind- 94. Whitmarsh, A. J., S.-H. Yang, M. S.-S. Su, A. D. Sharrocks, and R. J. Davis. ing of simian immunodeficiency virus to Jurkat T cells expressing CCR5 1997. Role of p38 and JNK mitogen-activated protein kinases in the acti- receptor. Virology 252:210–217. vation of ternary complex factors. Mol. Cell. Biol. 17:2360–2371. 72. Quilliam, L. A., S. Y. Huff, K. M. Rabun, W. Wei, W. Park, D. Brock, and 95. Woodgett, J. R., J. M. Kyriakis, J. Avruch, L. I. Zon, B. Zanke, and D. J. C. J. Der. 1994. Membrane-targeting potentiates guanine nucleotide ex- Templeton. 1996. Reconstitution of novel signaling cascades responding to change factor CDC25 and SOS1 activation of ras transforming activity. cellular stresses. Philos. Trans. R. Soc. Lond. B Biol. Sci. 351:135–142. Proc. Natl. Acad. Sci. USA 91:8512–8516. 96. Yang, D., C. Tournier, M. Wysk, H. T. Lu, J. Xu, R. J. Davis, and R. A. 73. Raingeaud, J., A. J. Whitmarsh, T. Barrett, B. De´rijard, and R. J. Davis. Flavell. 1997. Targeted disruption of the MKK4 gene causes embryonic 1996. MKK3- and MKK6-regulated gene expression is mediated by the p38 death, inhibition of c-Jun NH2-terminal kinase activation, and defects in mitogen-activated protein kinase signal transduction pathway. Mol. Cell. AP-1 transcriptional activity. Proc. Natl. Acad. Sci. USA 94:3004–3009. Biol. 16:1247–1255. 97. Yang, S. H., A. J. Whitmarsh, R. J. Davis, and A. D. Sharrocks. 1998. 74. Reinhard, C., B. Shamoon, V. Shyamala, and L. T. Williams. 1997. Tumor Differential targeting of MAP kinases to the ETS-domain of transcription necrosis factor-␣-induced activation of c-jun N-terminal kinase is mediated factor Elk-1. EMBO J. 17:1740–1749. by TRAF2. EMBO J. 16:1080–1092. 98. Yang, S.-H., P. R. Yates, A. J. Whitmarsh, R. J. Davis, and A. D. Sharrocks. 75. Robinson, M. J., and M. H. Cobb. 1997. Mitogen-activated protein kinase 1998. The Elk-1 ETS-domain transcription factor contains a mitogen-acti- pathways. Curr. Opin. Cell Biol. 9:180–186. vated protein kinase targeting motif. Mol. Cell. Biol. 18:710–720. 76. Roizman, B., and A. E. Sears. 1996. Herpes simplex viruses and their 99. Yurochko, A. D., E.-S. Hwang, L. Rasmussen, S. Keay, L. Pereira, and E.-S. replication, p. 2231–2295. In B. N. Fields, D. M. Knipe, and P. M. Howley Huang. 1997. The human cytomegalovirus UL55 (gB) and UL75 (gH) (ed.), Virology, 3rd ed. Lippincott-Raven Publishers, Philadelphia, Pa. glycoprotein ligands initiate the rapid activation of Sp1 and NF-␬B during 77. Rong, B. L., T. A. Libermann, K. Kogawa, S. Ghosh, L.-X. Cao, D. Pavan- infection. J. Virol. 71:5051–5059. Langston, and E. C. Dunkel. 1992. HSV-1 inducible proteins bind to NF␬B- 100. Zachos, G., B. Clements, and J. Conner. 1999. Herpes simplex virus type 1 like sites in the HSV-1 genome. Virology 189:750–756. infection stimulates p38/c-Jun N-terminal mitogen-activated protein kinase 78. Sakata, N., H. R. Patel, N. Terada, A. Aruffo, G. L. Johnson, and E. W. pathways and activates transcription factor AP-1. J. Biol. Chem. 274:5097– Gelfand. 1995. Selective activation of c-Jun kinase-activated protein kinase 5103.