We previously demonstrated that IFN-β transgene treatment protects mouse trigeminal ganglia (TG) cells from acute HSV-1 infection in vitro. However, IFN-α6 transgene treatment does not provide protection against acute HSV-1 infection in vitro, even though equivalent levels of IFN are expressed with both transgene treatments. In the present study we show that IFN-β transgene treatment before acute ocular HSV-1 infection protects mice from HSV-1-mediated mortality, whereas IFN-α6 transgene treatment does not reduce mortality. Treatment with the IFN-β and IFN-α6 transgenes was associated with increased expression of oligoadenylate synthetase (OAS)1a mRNA in the eye. However, protein kinase R mRNA was not up-regulated in the eye. In TG, only IFN-β transgene treatment reduced infectious virus levels. Furthermore, in the absence of a functional OAS pathway, corneal HSV-1 Ag expression was more widespread, and the ability of IFN-β transgene treatment to reduce infectious HSV-1 in eyes and TG was lost. Along with selective up-regulation of OAS1a mRNA expression in TG from IFN-β transgene-treated mice, we found increased levels of phospho-STAT1. Likewise, p38 MAPK phosphorylation was increased in TG from IFN-β transgene-treated mice, compared with both IFN-α6 and vector-treated mice. We also observed a time-dependent increase in JNK phosphorylation in TG from IFN-β transgene-treated vs IFN-α6 and vector-treated mice. Our results demonstrate that IFN-β is a potent antiviral cytokine that exerts protection against ocular HSV-1 infection via selective up-regulation of OAS1a mRNA in TG and by altering the phosphorylation of proteins in antiviral signaling cascades.
Epidemiological studies indicate that HSV-1 is the most frequent microbial cause of blindness in industrialized countries (1). Furthermore, it is estimated that 50% of U.S. citizens are presently infected with HSV-1 (2). One possible explanation for its prevalence is the capacity of the virus to interfere with endogenous antiviral pathways during its life cycle. Extensively studied examples include the dsRNA-dependent protein kinase R (PKR)3 and multienzyme 2′-5′ oligoadenylate synthetase (OAS) pathways (3). Type I IFN binding to the type I IFN receptor on the cell membrane induces a cascade of events involving changes in the phosphorylation/activation state of known and yet-to-be identified proteins. Downstream outcomes include decreased viral replication via regulation of IFN-stimulated gene expression (such as OAS and PKR) and subsequent induction of viral mRNA degradation, inhibition of viral protein synthesis, and activation of apoptosis in virally infected cells (3, 4, 5, 6, 7). Although all type I IFNs use the same receptor, events downstream of the type I receptor distinguish between subtypes of IFN species relative to the activation of genes ultimately responsible for establishment of the anti-HSV state (8). Understanding what role specific type I IFNs and IFN-inducible proteins play is a necessary prerequisite to developing an efficacious treatment for herpetic keratitis and associated inflammatory responses leading to irreversible corneal damage.
Ocular HSV-1 infection commences in host epithelial tissue in one of the three dermatomes innervated by the trigeminal ganglia (TG) (9). Specifically in corneal epithelium, viral replication leads to development of primary herpetic keratitis, which involves formation of epithelial lesions and may also be associated with stromal involvement, depending on the host species and strain (10, 11). HSV-1 enters sensory neuron terminals in basal corneal epithelium and traffics to TG via retrograde transport (12, 13, 14), where it undergoes further replication before establishing latency in a subpopulation of neurons (12).
The success of HSV-1 to replicate in eukaryotic cells is due to countermeasures developed by the virus against endogenous antiviral pathways (15). Multiple antiviral pathways have been shown to use JAK/STAT signal transduction pathways (3). Specifically, in response to virus or type I IFNs, JAKs are activated via phosphorylation. Activated JAKs subsequently phosphorylate STATs. Activation of STATs via phosphorylation causes homo or heterodimers of STAT (STAT1 and/or STAT2) to form and later translocate to the nucleus. Within the nucleus, STATs bind to gene regulatory elements and thereby control the transcription of IFN-stimulated genes, including enzymes in key antiviral pathways.
During the normal course of infection in cell types that use PKR antiviral pathways, IFN binding triggers the up-regulation of PKR. PKR then binds and phosphorylates an eukaryotic initiation factor substrate (16, 17), leading to inhibition of viral protein synthesis (18). In cell types that use OAS antiviral pathways, 2′-5′ oligoadenylates bind RNase L, resulting in the cleavage of viral mRNA and rRNA (19). Eukaryotic cells also respond to viruses via stress-activated signal transduction pathways. JNK and p38 MAPK are two involved protein kinases, which are activated within 3 h postinfection (p.i.) (20, 21, 22, 23). JNK is activated in response to viral protein peptides (20), whereas p38 MAPK, and sometimes NF-κB, are activated by virus binding to a cell surface receptor (24). The end result of IFN induction of antiviral pathways and signal transduction pathways is decreased production of viral progeny and decreased viral lytic gene expression (25, 26).
The current study is an extension of previous studies to compare antiviral activities of specific type I IFNs against HSV-1. Cumulative mouse survival for different IFN treatments was used as a measurement of effectiveness. Subsequently, IFN-β and IFN-α6 transgene treatments were used to begin to answer questions to determine why some type I IFNs inhibit HSV-1 pathogenesis in vivo whereas others are not effectual.
Materials and Methods
Animal treatment was consistent with the National Institutes of Health Guidelines on the Care and Use of Laboratory Animals (publication no. 85-23, revised 1996). All procedures were approved by the University of Oklahoma Health Sciences Center and Dean A. McGee Eye Institute Institutional Animal Care and Use Committees. ICR mice (Sprague Dawley) were obtained from Harlan Sprague Dawley; wild-type (wt)/C57BL/6 mice were obtained from The Jackson Laboratory. RNase L-null (RNase L−/−) and PKR-null (PKR−/−) mice were generated with a wt background as previously described (27, 28). Mice (ages 6–8 wk) were anesthetized by i.p. injection with xylazine (2 mg/ml; 6.6 mg/kg) and ketamine (30 mg/ml; 100 mg/kg). Corneas were subsequently scarified with a 5/8-inch 25-gauge needle, and tear films were blotted before topical application of murine IFN transgenes that were cloned as described (29). Plasmid constructs were transformed as described (29), grown in Terrific broth (Bio 101 Systems) containing 50 μg/ml kanamycin, and purified using a plasmid Maxiprep kit (Bio-Rad). Twenty-four hours posttransfection (p.t.), McKrae strain of HSV-1, prepared as described (30), was applied to the corneal epithelium (240 PFU/eye (ICR mice) or 1000 PFU/eye (C57BL/6 wt mice)). Mice were monitored over 30 days p.i., and mortality was recorded for each group of animals. For morphology studies and protein analysis, treated and untreated mice were euthanized by either cervical dislocation or CO2 asphyxiation.
All cell culture reagents were obtained from Invitrogen Life Technologies. Vero (African green monkey kidney) cells and L929 cells were obtained from the American Type Culture Collection, cultured in RPMI 1640 (Vero cells) or DMEM (L929 cells), supplemented with 10% FBS, 2% antibiotic/antimycotic, and 0.2% gentamicin. Cells were plated in 96-well flat-bottom plates (35,000 cells/well), and incubated at 37°C in an atmosphere of 5% CO2 and 95% humidity.
Determination of biologically active IFN
To determine the concentration of biologically active IFN following application of naked plasmid DNA constructs, corneal buttons were dissected under sterile conditions 24 h p.t. and cultured in 24-well plates containing DMEM, prepared as earlier described. Two days after establishment of corneal explant cultures, supernatants from transfected corneas were diluted (1/2 log dilutions) in DMEM and incubated with confluent L929 cells. After 24 h, supernatants were removed and replaced with DMEM containing vesicular stomatitis virus (multiplicity of infection = 0.05). Plates were examined for cytopathic effect 30–33 h p.t. Fifty percent inhibition of cytopathic effect was equivalent to 1 U/ml for IFN-α6 or IFN-β. A standard curve was generated using half-log dilutions of murine IFN-α (PBL Biomedical Laboratories). Data represent cumulative recordings from three mice for each treatment from two separate experiments (see Fig. 2).
Determination of HSV-1 viral titers
HSV-1 viral titers were examined to compare the quantity of recoverable HSV-1 from animals treated with pKCMV vector, pKCMV-α6, and pKCMV-β. Corneas from three mice per treatment were scarified, and 75–100 μg of plasmid DNA (pKCMV vector (control), pKCMV-α6, and pKCMV-β) was topically applied to corneas. Twenty-four hours p.t. HSV-1 was topically applied as described. At days 3 and 5 p.i., mice were euthanized and the eyes and TG were dissected under sterile conditions. Tissues from each mouse were homogenized in 500 μl of RPMI 1640 medium using a tissue miser (Fisher Scientific). Tissue homogenates were subsequently centrifuged at 10,000 × g for 1–2 min; resulting supernatants were serially diluted, and 100 μl of supernatant was transferred onto confluent Vero cells in 96-well culture plates. After 1 h incubation, transferred supernatants were discarded and replaced with 100 μl of 0.5% methyl cellulose, diluted in RPMI 1640. Cytopathic effect was observed after a 32-h incubation period and recorded as PFU per milliliter.
Western blot detection
Total cell protein was harvested from TG by homogenization with a tissue miser using the following lysis buffer: 1% Triton X-100 and 2 mM ethylene tetraacetic acid supplemented with a protease inhibitor mixture (Calbiochem) containing 500 μM AEBSF, hydrochloride, 150 nM aprotinin, 1 μM E-64 protease inhibitor, 0.5 mM disodium EDTA, and 1 μM hemisulfate leupeptin. Homogenates were centrifuged at 12,000 × g for 5 min to pellet insoluble detergent-resistant membranes and cytoskeletal material. Proteins in clarified supernatants were quantified with a bicinchoninic acid protein assay (Pierce) using a FL600 fluorescence plate reader (Bio-Tek Instruments). Supernatants for phospho-STAT Western detection were clarified by incubating with agarose-coupled protein G (Santa Cruz Biotechnology) by rotation at 4°C for 1 h. Total STAT in clarified supernatants (100 μg of protein extract diluted 1/1 in solubilization buffer (10 mM Tris-HCl (pH 7.4), 50 mM NaCl, 0.5 mM EDTA, and 1% Triton X-100)) was immunoprecipitated. Immunoprecipitates (for Western blot detection of p-STAT 8–12 h p.t. with IFN transgenes) or homogenates (for Western blot detection of STAT1, JAK1, and β-actin, 5 days p.i.) were diluted to a final concentration of 1× in sample buffer. Proteins were electrophoresed on 7.5% Tris-HCl Criterion precast gels (Bio-Rad) and semidry transferred to polyvinylidene difluoride membranes (Bio-Rad). Polyvinylidene difluoride membranes were blocked with 5% dry milk in 1× T-TBS overnight at 4°C. Western blot detection was performed by incubation for 1–2 h at room temperature on a shaking platform with the following primary Abs: rabbit polyclonal anti-mouse phospho-STAT1 (Y701; BioSource International); rabbit polyclonal anti-mouse anti-STAT1 p91/p84 Ab (Santa Cruz Biotechnology); rabbit polyclonal anti-β-actin JAK (Santa Cruz Biotechnology); or a rabbit polyclonal IgG directed against JAK1 (Santa Cruz Biotechnology). The primary JAK Ab was detected by a secondary incubation for 1 h at room temperature using anti-rabbit or anti-goat JAK coupled to HRP (Santa Cruz Biotechnology). HRP was detected using ECL Western blotting chemiluminescence detection reagents (Amersham Biosciences). Images were obtained by exposing polyvinylidene difluoride membranes to x-ray film. Film images were digitalized using the Bio-Rad 1000 image documentation system and software (Quantity One 4.0.3; Bio-Rad). To determine whether equivalent volumes of protein were loaded, membranes were stripped by submersion in stripping buffer (100 mM 2-ME, 2% SDS, 62.5 mM Tris-HCl, pH 6.7) and incubated at 50°C for a minimum of 30 min. Membranes were subsequently washed two times for 10 min with large volumes of 1× T-TBS, then blocked with 5% dry milk in 1× T-TBS, and subsequently a second Western blot detection was performed using Abs to either STAT1 (phospho-STAT Western blots), or β-actin and JAK1 (STAT1 Western blots).
In two separate experiments, three 10-wk-old wt mice per treatment and time point were transfected with 50–75 μg of plasmid or vector DNA as described. Mice were anesthetized with ketamine/xylazine at 1, 3, and 9 h p.t. and perfused with 1× PBS before TG were dissected. After rinsing TG with 1× PBS, they were frozen on dry ice. Total proteins were extracted from the TG by sonication in Bio-Plex cell lysis buffer (Bio-Rad). Equivalent amounts of protein (30 μg) were volumetrically normalized in cell lysis buffer, and phosphoprotein assays were performed for IκBα, JNK, and p38 MAPK according to the manufacturer’s instructions (Bio-Rad).
Real-time PCR, performed as previously described (31), was used to quantify gene expression of OAS and PKR in transfected mouse eyes and TG.
Immunohistochemical staining for HSV-1 Ags
To investigate phenotypic differences between wt and RNase L−/− mice, corneas were fixed in alcoholic Z-fix (Anatech) 48 h at room temperature, dehydrated, and embedded in parafilm. Five micron sections were deparaffinized, rehydrated, and stained for HSV-1 Ags as previously described (26).
Whole mount tissue preparation
Whole eyes were enucleated from infected wt and RNase L−/− mice and stored in fixative (4% paraformaldehyde; Sigma-Aldrich) in 1× PBS) at room temperature until corneas were dissected. Dissected corneas were postfixed overnight at room temperature, washed three times with wash buffer (1× PBS containing 1% Triton X-100 (Sigma-Aldrich)) and subsequently incubated in blocking buffer (10% horse serum in 1× PBS) for 1 h at room temperature. Corneas were then incubated with control (FITC-conjugated mouse IgG primary Abs) or FITC-conjugated primary Abs to HSV-1 Ags (DAKO), diluted 1/100 in 1× PBS, for 3 h in a dark, humid 37°C chamber. After Ag staining, corneas were washed three times with wash buffer and incubated overnight at 4°C in the dark with DAPI (4′,6′-diamidino-2-phenylindole; Vector Laboratories) mounting medium. Four equidistant relaxing incisions were made in each DAPI-stained cornea to facilitate flattening on microscope slides. Slides were stored at 4°C in the dark until analyzed.
ANOVA and Tukey’s t test were used to determine significant (p ≤ 0.05) differences between the transgene and vector-treated groups. Survival studies were analyzed for significance (p ≤ 0.05) by the Mann-Whitney U rank sum test. All statistical analysis was performed using the GBSTAT program (Dynamic Microsystems).
IFN-α1 and IFN-β transgene treatment reduced HSV-1-mediated mortality
Four type I IFN transgenes (IFN-α1, -α4, -α9, and -β) inhibited HSV-1 infection, with IFN-α1- and IFN-β-treated ICR mice, which are highly sensitive to the lethal effects of HSV-1 (32), showing enhanced survival (3-fold) compared with controls treated with saline or vector alone. IFN-α4 and IFN-α9 also increased mouse survival, in this case by 2.5-fold, compared with untreated controls (Fig. 1). In contrast, two IFN species (IFN-α5 and IFN-α6) were not effective. Specifically, IFN-α6 transgene treatment produced a negligible increase in survival. IFN-α5 transgene treatment had no affect on mouse survival; mortality rates at the end of the 30 days matched the baseline survival rates observed in both vector and saline-treated mice (Fig. 1).
Equivalent levels of IFN were detected in IFN-α6 and IFN-β transfected corneas
An effective (IFN-β) and noneffective (IFN-α6) transgene was chosen for subsequent studies to identify interactions with proteins in ocular anti-HSV-1 pathways. Initially, we determined the level of biologically active IFN generated in situ p.t. In both IFN-α6 and IFN-β transgene-treated corneas, equivalent levels of IFN were detected, whereas no IFN was detected in plasmid vector-treated corneas (Fig. 2).
Treatment with IFN-β transgenes reduced HSV-1 titers in eyes and TG
At day 3 p.i., both the IFN-α6 and IFN–β transgenes reduced viral titers in the eye compared with vector-treated control (Fig. 3). By day 5 p.i., only the IFN-β transgene-transfected eyes showed a significant reduction in viral titer, compared with both vector and IFN-α6 transgene-transfected animals. Consistent with this result, the TG from IFN-β transgene-transfected mice possessed significantly less HSV-1 compared with the IFN-α6 transgene or vector (Fig. 3). Taken together, the results suggest that the IFN-β transgene is capable of disrupting viral replication in the TG as well as the eye, whereas the IFN-α6 transgene suppresses viral replication only in the eye.
IFN-β transgene treatment causes up-regulation of phospho-STAT1 and STAT1 in TG
Phospho-STAT1 levels were found to be elevated at 8–12 h p.t. in TG from IFN-β transfected mice, compared with TG from vector and IFN-α6 transgene-treated mice (Fig. 4,A). However, similar levels of phospho-STAT were detected in eyes from vector, IFN-α6, and IFN-β transgene-treated mice. Similar levels of STAT1 were detected in the eyes and TG from vector, IFN-α6, and IFN-β transfected mice at 8–12 h p.t. (Fig. 4,A). The results depicted in Fig. 4 A represent a similar profile to what was seen in three of four independent experiments. There were no differences in phospho-STAT1 levels comparing the transgene- or vector-treated groups at other time points: 24, 72, or 120 h p.i. (data not shown).
Transfections with IFN-β transgenes also resulted in increased levels of total STAT1 at 5 days p.i. in TG, compared with TG from vector and IFN-α6 transgene-treated mice (Fig. 4,B). The results depicted in Fig. 4 B are representative of what was observed in five of five independent experiments. We did not identify consistent differences in total STAT1 levels comparing transgene treatments in TG at time points 24 or 72 h p.i. (data not shown).
The IFN-β transgene induces OAS1a but not PKR expression in TG
Two pathways associated with IFN-inducible anti-HSV activity include OAS and PKR (31, 33, 34, 35, 36). Therefore, we investigated the levels of OAS and PKR mRNA 24 h after transfection with IFN-α6 and IFN-β transgenes. OAS1a mRNA was up-regulated by both transgenes in eye compared with vector-treated (Fig. 5). However, in the TG IFN-β transgene selectively enhanced OAS1a mRNA levels compared with vector and IFN-α6 transgene treatments (Fig. 5). No significant differences in PKR were observed for either eyes or TG isolated from transfected animals (Fig. 5).
Antiviral activity of IFN-β is lost in the absence of a functional OAS pathway
RNase L, a downstream effector of the OAS pathway, provides resistance against viral infections by cleaving viral mRNA and rRNA (37, 38). Consequently, we evaluated resistance to HSV-1 in RNase L-deficient mice, treated with IFN-α6 and IFN-β transgenes. In the absence of RNase L, the antiviral effectiveness of the IFN-β transgene was lost. Nearly equivalent levels of HSV-1 were recovered in eyes and TG from vector and IFN-α6 transgene-transfected mice vs IFN-β-transfected animals (Fig. 6,A). In contrast, and similar to ICR mice, IFN-β transgene transfection of wt mice was associated with significantly reduced viral titers for eyes and TG (Fig. 6 A).
HSV-1 Ag expression is more widespread in RNase L-deficient mouse corneas
We found three of four corneas from RNase L-deficient mice displaying an increase in HSV-1 Ag expression (Fig. 6,B). Specifically, in RNase L−/− mouse cornea sections, HSV-1 Ag expression was visible in corneal epithelium, stroma, and endothelium compared with Ag expression only found in the epithelium of wt mice (Fig. 6,B). In wt mouse corneal flat mounts, HSV-1 Ag expression was visible only along scarification lines, but in RNase L−/− cornea flat mounts, HSV-1 Ag was dispersed in the surrounding tissue (Fig. 6 C).
Transgene treatment differentially affects IκBα, p38 MAPK, and JNK phosphorylation
To identify which signaling cascades are involved in the activity of IFN-α6 and IFN-β, phosphoprotein assays were performed for three select proteins in the TG (the site that distinguishes the antiviral activity of the type I IFNs). Corneal topical application of the IFN-α6 and IFN-β transgenes resulted in an increase in IκBα phosphorylation relative to vector at 9 h p.t. in TG (Fig. 7,A). IFN-β transgene treatment resulted in increased p38 MAPK phosphorylation, relative to both vector and IFN-α6 treatment at 9 h p.t. (Fig. 7,B). IFN-β transgene treatment resulted in increased JNK phosphorylation, relative to vector at 1 h p.t. and relative to IFN-α6 transgene treatment at 3 h p.t. (Fig. 7 C).
In this study we demonstrated that type I IFN transgene treatment dramatically reduces HSV-1 mediated mortality (Fig. 1). Specifically, the antiviral mechanisms induced by IFN-β led to a 3-fold reduction in mortality, indicating a significant systemic impact on the host. Other type I IFNs in this study and also in prior in vitro studies (40) were not equally effective. Although the IFN-α6 and IFN-β transgenes induced equal production of IFN in cornea (Fig. 2), only the IFN-β transgene provided protection against acute HSV-1 infection. This observation prompted us to perform further studies to identify the cause for this difference.
As previously demonstrated in vitro with both immortalized cells and primary TG cultures, IFN-β transgene treatment was successful at reducing viral titers in the eyes and TG of ICR and C57BL/6 mice (8, 9). Differences in the ability of IFN-β vs IFN–α6 to reduce viral titers were notable at day 5 when only IFN-β transgene treatment reduced viral titers in the eye and TG. The ability of the IFN-β transgene to impact viral load was ablated in the absence of a functional OAS pathway, as observed in RNase L knockout mice. This implicates a crucial role for OAS in the antiviral effectiveness of IFN-β transgene transfection. Although similar viral titers were detected in RNase L knockout and wt mice, stromal and endothelial HSV-1 Ag expression was detected only in RNase L-deficient mice. This observation is similar to the difference in primary vs recurrent herpetic stromal keratitis where no differences in HSV-1 titer are found, but Ag expression is more diffusely expressed in cornea during recurrent keratitis (41). Like humans but unlike HSV-1 susceptible mouse strains, wt mice rarely exhibit stromal involvement during acute HSV-1 infection (10, 11). However, RNase L-deficient mice exhibit stromal involvement, so genetic differences in OAS could possibly explain species and strain differences in the severity of primary herpetic keratitis as observed in this study and also by Zheng et al. (39).
Although both IFN transgenes induced antiviral gene expression in the eye, differences in the ability of IFN-β vs IFN-α6 transgenes to alter antiviral gene expression were notable in TG. Only IFN-β transgene transduction induced the expression of OAS in TG. We conclude that the greater antiviral activity of IFN-β transgene treatment is caused by site-specific changes in OAS expression.
p38 MAPK phosphorylation was also found to be increased in IFN-β-treated TG. p38 MAPK activation is associated with controlling cell death and proliferation (17) and is also activated in response to neurotoxic molecules (42). Relevant to type I IFN induction, in some cell types, p38 MAPK plays a role in the formation of trimeric complexes of STAT1/STAT2 heterodimers and p48 in the promoter region of IFN-stimulated genes (43). p38 MAPK also has a role in regulating phosphorylation of serine residues, essential for full transcriptional activation of STAT1 homodimers. In addition to the JAK-STAT mechanisms, some cells have also been shown to have p38 antiviral activities, such as initiating increased production of proinflammatory cytokines (44), which are independent of the JAK/STAT pathway (45). Taken together, we have identified two signaling cascades p38 MAPK and STAT1 that may function in unison or independently in promoting resistance to HSV-1. The importance of STAT1 activation to the host in resistance to HSV-1 is underscored by the findings that implicate targeting the JAK1/STAT1 pathway and induction of suppressor of cytokine signaling-3 by HSV-1 (46, 47). Along with our in vitro findings (40), it would appear genes responsive to the activation of STAT1 are the driving force behind the superior antiviral activity of IFN-β over other type I IFNs in controlling HSV-1 replication.
We thank Benitta John-Philip for technical assistance with the isolation of plasmid DNA, as well as Paula Pierce and Dr. John Ash for assistance with the immunohistochemical detection of HSV-1 Ag. We are also indebted to Dr. Juneann Murphy (Department of Microbiology and Immunology, University of Oklahoma of Health Sciences Center) for critical reading and discussion of the manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by a Research to Prevent Blindness Stein Research Professorship (to D.J.J.C.) and by National Institutes of Health Grants AI053108 (to D.J.J.C.), CA 44059 (to R.H.S.), and a National Eye Institute Core Grant EY12190.
Abbreviations used in this paper: PKR, protein kinase R; OAS, 2′-5′ oligoadenylate synthetase; TG, trigeminal ganglia; wt, wild type; p.t., posttransfection; p.i., postinfection.