EGR Proteins Mediate Interferon‐Independent Anti‐HSV‐1 Responses Through Viral and Host Targets
Shuaishuai Wang, Fujun Hou, Xunuo Jin, Yuhang Xiang, Dongli Pan
TL;DR
This paper shows that EGR proteins help fight HSV-1 infection in neurons without needing interferons, by targeting both the virus and host immune genes.
Contribution
The study reveals a novel interferon-independent antiviral mechanism involving EGR proteins during HSV-1 infection.
Findings
EGR1 represses HSV-1 replication in neuronal cells and mouse ganglia.
EGR1 increases LAT expression and activates host immune genes like IRF7 and ISG15.
EGR proteins mediate antiviral responses independently of interferons.
Abstract
Quick responses to viral infections, which are essential for controlling viral diseases, are typically mediated by interferons. Herpes simplex virus 1 (HSV‐1) switches between lytic and latent infections in neurons. Here we show that host early growth response (Egr) genes (including Egr1‐Egr4), which are not interferon‐stimulated genes, are generally upregulated in HSV‐1‐infected neuronal cells and acutely infected mouse ganglia. Surprisingly, Egr1 upregulation is independent of previously reported pathways upstream of Egr1 expression but dependent on viral protein ICP0. EGR1, in turn, represses HSV‐1 replication in neuronal cells. Recombinant HSV‐1 expressing EGR1 exhibits reduced replication in mouse ganglia and brainstems in vivo. Mechanistically, EGR1 binds to sites within the viral latency‐associated transcript (LAT) gene promoter to increase LAT expression, which is known to favor…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5
FIGURE 6
FIGURE 7- —National Key R&D Program of China10.13039/501100012166
- —National Natural Science Foundation of China10.13039/501100001809
- —Key Project of the National Natural Science Foundation of Zhejiang Province
- —State Key Laboratory for Diagnosis and Treatment of Infectious Diseases10.13039/501100011441
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsHerpesvirus Infections and Treatments · interferon and immune responses · Viral Infections and Immunology Research
Introduction
1
Herpes simplex virus 1 (HSV‐1) is a prevalent human pathogen that causes many diseases, such as cold sores, genital herpes, herpetic keratitis, and herpes simplex encephalitis. During HSV‐1 replication, entry of the viral DNA genome into the nucleus is followed by viral gene expression in an ordered cascade: immediate‐early (IE), early (E), and late (L) genes are sequentially activated, leading to abundant production of viral genomes and particles [1]. As the viral processes are ongoing, recognition of pathogen‐associated molecular patterns by host pattern recognition receptors signals through pathways to activate transcription factors such as IFN regulatory factor 3 (IRF3) and IRF7, which translocate into the nucleus to stimulate the expression of type I IFNs. After their release, the IFNs bind to their receptors on the plasma membrane, activating downstream pathways to induce the expression of interferon‐stimulated genes (ISGs) that encode antiviral proteins. At the same time, HSV‐1 encodes numerous immune‐evasion proteins to counteract the sophisticated defense system [2].
Despite abundant virus production during such productive (lytic) infection, most virus in peripheral tissues is eventually cleared by host immunity. However, some virus enters neurons of the peripheral nervous system to establish life‐long latent infection, where the only abundant gene products, including the latency‐associated transcript (LAT) intron and some microRNAs, are those derived from the LAT locus [3]. When stimulated by certain stresses, latent HSV‐1 reactivates [4, 5], leading to renewed lytic infection, disease recurrence, and virus spread. Sensory neurons of the trigeminal ganglia (TG) are the major sites of HSV‐1 latency. Establishment of latency entails repression of lytic infection, the mechanisms of which remain incompletely understood. It has been hypothesized that viral latency‐associated transcripts (LATs) and their microRNA derivatives [6, 7, 8, 9], as well as host neuron‐specific [10, 11, 12] and innate immune factors [13, 14], contribute to the repression. In particular, the LATs repress lytic gene expression [15, 16, 17, 18], increase heterochromatin modifications on the viral genome [19, 20], and promote neuronal survival [21, 22].
Host immediate‐early genes, such as c‐Fos, Arc, and early growth response (Egr) genes (including Egr1‐Egr4), are a class of genes that are activated rapidly and transiently in response to cellular stimuli [23, 24]. As zinc‐finger transcription factors, EGR proteins regulate their target genes by binding to the GCGKGGGCG (K = T or G) consensus sequence [25]. These proteins are implicated in a variety of biological processes. In the nervous system, EGR1 is abundantly expressed and regulates synaptic plasticity and neuronal activity [26]. The roles of EGR1 in herpesvirus infections have been reported [27]. For instance, EGR1 cooperates with CBP to activate the promoter of the Kaposi's sarcoma‐associated herpesvirus RTA gene [28]. EGR1 is induced by Epstein‐Barr virus LMP1 protein and may contribute to cancer cell survival [29]. EGR1 binds the human cytomegalovirus (HCMV) genome upstream of UL138 and promotes UL138 expression and latency establishment in CD34^+^ hematopoietic progenitor cells [30], while downregulation of EGR1 by HCMV miR‐US22 blocks proliferation of such cells to promote viral reactivation [31]. With regard to HSV‐1, induction of EGR1 expression by HSV‐1 has been observed in mouse brains and corneas in vivo as well as in multiple cell types in culture [32, 33, 34, 35]. Additionally, one study identified a GCGGGGGCG sequence in the HSV‐1 LAT promoter and confirmed EGR1 binding to this sequence by gel retardation and DNase I protection assays [36]. Other studies reported that EGR1 overexpression could increase the activities of ICP4, ICP27, and ICP22/47 promoters in human neuroblastoma SK‐N‐SH cells while EGR1 knockout or knockdown in mice reduced corneal disease and mortality [33, 34]. However, another study reported that EGR1 could repress ICP4 and ICP22 promoter activities in human embryonic kidney 293 cells [37]. Therefore, while previous studies consistently observed EGR1 upregulation by HSV‐1 in various contexts, the downstream effects of EGR proteins, especially in neuronal cells, require further clarification.
In this work, we found that during HSV‐1 infection of neuronal cells, most Egr genes were upregulated and they, in turn, functioned to suppress HSV‐1 replication. Notably, all Egr genes were upregulated in acutely infected mouse ganglia, and ectopically expressed EGR1 substantially reduced HSV‐1 acute replication in a mouse model. Mechanistically, EGR proteins bind to the LAT promoter to stimulate LAT expression, while also enhancing IRF7 expression to suppress HSV‐1 replication. Surprisingly, these effects of EGR1 and IRF7 were independent of IFNs. Therefore, this work reveals an IFN‐independent mechanism that quickly responds to viral infection to mitigate its effects.
Results
2
HSV‐1 Infection Induces Egr Gene Expression in Neuronal Cells in Culture and In Vivo
2.1
To identify host genes that quickly respond to HSV‐1 infection, we first conducted RNA‐seq analysis of mouse neuroblastoma Neuro‐2a cells infected with strain KOS at a multiplicity of infection (MOI) of 10 for 5 h. In mock‐infected cells, normalized read counts indicated that, among the four Egr genes, Egr1 expression was the highest, followed by Egr2, then by Egr3, while Egr4 mRNA was undetectable (Figure 1A). Relative to mock infection, Egr1 and Egr2 were top‐ranking host genes that were significantly upregulated by HSV‐1 infection, both with log_2_(fold change) values of ∼1.5 (Figure 1B and Table S1). Egr3 also showed a trend of upregulation with a log_2_(fold change) value of ∼1, but it was insignificant due to low expression and high variability. RT‐qPCR results showed that, when normalized to host GAPDH mRNA levels, Egr1, Egr2, and Egr3 mRNAs were upregulated by 4.9, 21.7, and 12.5 fold, respectively, at 4 h post‐infection (hpi) with no further upregulation at 8 hpi (Figure 1C), indicating early upregulation. Egr4 mRNA was still undetectable by RT‐qPCR. In human corneal epithelial cells (HCECs) and mouse macrophage RAW264.7 cells, Egr1 mRNA was also upregulated significantly within the first few hours of HSV‐1 infection (Figure S1A). Together with previous reports [32, 33, 34, 35], these data suggest that upregulation of EGR1 by HSV‐1 is not cell‐type or host‐species specific.
HSV‐1 lytic infection induces EGR1 expression in an ICP0‐dependent manner. (A) Read counts of Egr mRNAs in mock‐infected Neuro‐2a cells in the RNA‐seq data. (B) Volcano plot for differential gene expression between mock and HSV‐1 infected cells (MOI = 10, 5 hpi). Each dot represents one gene. Egr genes are indicated in red. (C) RT‐qPCR analysis of Egr mRNAs in mock‐infected and HSV‐1 infected Neuro‐2a cells (MOI = 10). (D) Neuro‐2a cells were pretreated with DMSO or 100 nM MG132 for 12 h, then infected with WT or ICP0‐null HSV‐1 (MOI = 10) in the presence of DMSO or 100 nM MG132 for the indicated times before Western blot analysis. (E) RT‐qPCR analysis of the indicated mRNAs in mouse TG at the indicated times after inoculation of mice with 2 x 105 PFU/eye. The n values represent the numbers of TG. (F) Neuro‐2a cells were transfected with 100 ng/mL of the indicated luciferase plasmid and 50 ng/mL of RL‐CMV plasmid (internal reference) for 24 h, then infected with HSV‐1 (MOI = 5) for the indicated times before luciferase assays. (G) Neuro‐2a cells were infected with indicated viruses (MOI = 5) for 10 h before Egr1 mRNA analysis by RT‐qPCR. (H) Same as F, but the indicated viruses were used for infection for 7 h. (I) Neuro‐2a cells were transfected with 400 ng/mL of pICP0 or pcDNA plasmid for 24 h before Egr1 mRNA analysis by RT‐qPCR. (J) Neuro‐2a cells were co‐transfected with 100 ng/mL of the luciferase plasmid with the Egr1 promoter, 50 ng/mL of RL‐CMV plasmid, and 300 ng/mL pcDNA or pICP0 for 24 h before luciferase assays. n = 3 (A, B, G, H, I, J), 4 (C, F) or 10‐14 (E) biologically independent samples. Data were analyzed by two‐tailed unpaired t‐tests (I, J) or one‐way ANOVA with Dunnett's multiple comparisons tests (C, E, F, G, H) and are presented as mean values ± standard deviations (SD).
In line with the mRNA results, EGR1 protein was upregulated at the very early time of 2 hpi in Neuro‐2a cells (Figure 1D). However, starting at 4 hpi, the original EGR1 band became weaker concomitant with the appearance of a lower‐motility band, suggesting that besides inducing EGR1 expression, HSV‐1 infection also induces EGR1 modifications, possibly accompanied by some degradation at later times. A proteosome inhibitor, MG132, somewhat intensified the bands indicative of some proteosomal degradation. Because of the well‐known function of ICP0 as a viral E3 ubiquitin ligase, we tested its involvement using an ICP0‐null (7134) virus. Unexpectedly, during ICP0‐null HSV‐1 infection, the original EGR1 band was not intensified at early times but rather became weaker over time, suggesting that EGR1 upregulation rather than its degradation was mediated by ICP0 (Figure 1D).
We then analyzed Egr gene expression in a mouse model of corneal inoculation in which the virus produced in the eyes is transported to TG, where acute replication is followed by latent infection. During acute infection, average Egr1, Egr2, and Egr4 mRNA levels increased by 0.83, 0.37, and 0.35 log, respectively, at 3 dpi, and by 0.52, 0.95, and 0.44 log at 5 dpi, with most of the increases being statistically significant (Figure 1E). Egr3 mRNA levels were unchanged at 3 dpi but increased by 0.34 log at 5 dpi. At 40 dpi, when latency was fully established, Egr1, Egr2, and Egr4 mRNA levels largely returned to the baselines, but Egr3 mRNA levels remained upregulated modestly relative to mock infection. Taken together, Egr genes are generally upregulated in HSV‐1‐infected neuronal cells in culture and acutely infected mouse ganglia in vivo.
Viral Protein ICP0 is Necessary and Sufficient for EGR1 Upregulation During HSV‐1 Infection
2.2
To understand the mechanism of Egr gene induction, we first tested the effects of inhibitors of different signaling pathways previously reported to be involved in Egr1 gene induction [38] (Table S2 and Figure S1B). In Neuro‐2a cells, the AKT and ATM inhibitors (MK‐2206 and KU‐55933) that block the PI3K‐AKT [30] and DNA damage [39] pathways, respectively, had little effect on Egr1 expression. The PI3K inhibitor (PI‐103) that also blocks the PI3K‐AKT pathway [40] even increased Egr1 expression. Although PDK1, MEK, and JNK inhibitors (BX‐795, Binimetinib, and SP600125) that block the PI3K‐ AKT [41], MEK/ERK [30], and JNK‐c‐Jun [42] pathways, respectively, decreased Egr1 expression, they could not reduce the extent to which Egr1 is upregulated by HSV‐1. Thus, different proteins in the PI3K‐AKT pathway appeared to have different effects on baseline EGR1 expression. Regardless, these results argue against HSV‐1 activating Egr1 expression through these pathways. Egr genes are not known as ISGs, and we confirmed that neither IFN‐α nor IFN‐β could increase Egr1 expression (Figure S1C). In reporter assays, HSV‐1 infection significantly increased luciferase expression from the Egr1 promoter, suggesting that EGR1 upregulation was due to promoter activation (Figure 1F).
To dissect the involvement of viral components, we treated Neuro‐2a cells with UV‐inactivated HSV‐1 or infected the cells with the ICP0‐null virus (7134) or its rescued derivative (7134R). Both luciferase assay and RT‐qPCR results showed that both UV inactivation and ICP0 deletion eliminated EGR1 upregulation by HSV‐1 (Figure 1G,H), indicating that the upregulation requires intact virions that express ICP0, consistent with the above Western blot results (Figure 1D). By contrast, treatment with acyclovir (ACV), an inhibitor of viral DNA synthesis, a process downstream of ICP0 expression, did not affect activation of the Egr1 promoter by HSV‐1 (Figure S1D). Furthermore, transfection of ICP0 alone into uninfected cells increased Egr1 mRNA levels and promoter activities (Figure 1I,J). Thus, ICP0 is necessary for HSV‐1 induction of Egr1 expression and sufficient for the induction when introduced alone.
EGR Proteins Suppress HSV‐1 Replication Through Interactions Between their Zinc Fingers and DNA
2.3
Next, we investigated the effects of EGR proteins on HSV‐1 replication. We first focused on EGR1, given its high expression in neuronal cells. In Neuro‐2a cells, two shRNAs that effectively knocked down EGR1 (Figure S2A) both increased HSV‐1 yields by ∼4 fold (Figure 2A). Two Egr1 siRNAs had similar effects (Figure S2B). Moreover, two independently derived EGR1 knockout cell lines both showed increased HSV‐1 yields, which were reduced again by transfected EGR1 (Figure 2B and Figure S2C), demonstrating that endogenous EGR1 suppresses HSV‐1 replication. When transfected, all human EGR proteins could suppress HSV‐1 replication by a range of 4 to 13 fold (Figure 2C and Figure S2C). The effects were not limited to neuronal cells since transfected EGR1 could suppress HSV‐1 replication in 293T and HCEC cells too (Figure S2D,E).
EGR proteins suppress HSV‐1 replication via DNA interactions with the zinc fingers. (A) Neuro‐2a cells were transfected with 400 ng/mL of plasmids expressing the indicated shRNAs for 24 h, then infected with HSV‐1 (MOI = 0.2) for 48 h before virus titration. (B) The indicated cells were transfected with 400 ng/mL of the indicated plasmids for 24 h, then infected with HSV‐1 (MOI = 0.2) for 48 h before virus titration. (C) Neuro‐2a cells were untreated, or mock‐transfected (with water), or transfected with the indicated plasmids for 24 h, then infected with HSV‐1 (MOI = 0.1) for 30 (left) or 48 (right) h before virus titration. (D) Schematic diagram of EGR proteins and EGR1 truncated mutants. The orange boxes represent zinc fingers. For the EGR1 truncated mutants, the numbers indicate amino acid positions. Expanded below are residues within the DBD with the mutated ones in red. (E) Neuro‐2a cells were transfected, infected and analyzed as in B. (F) The crystal structure of the EGR1 DBD complexed with DNA adapted by using the PyMOL software (PDB code 4R2D). The dotted lines represent hydrogen bonds whose distances are labeled. (G) Neuro‐2a cells were transfected, infected and analyzed as in B except that infection lasted for 42 h for the middle graph. n = 3 (A, B, C right, E left, G) or 4 (C left, E right) biologically independent samples. Data were analyzed by one‐way ANOVA with Dunnett's multiple comparisons tests (A, C, E, G) or two‐way ANOVA with Sidak's multiple comparisons tests (B) and are presented as mean ± SD.
EGR proteins have a conserved DNA‐binding domain (DBD) with three zinc fingers as well as other domains that are well‐defined for EGR1 but less so for other EGR proteins [24] (Figure 2D). Transfection‐infection experiments using plasmids expressing truncated EGR1 mutants showed that the DBD and N‐terminal activation domain (AD‐N) were required for virus suppression, while the inhibitory domain (ID) and C‐terminal activation domain (AD‐C) were dispensable (Figure 2E and Figure S2F). We then mutated each residue within the zinc fingers previously shown to contact DNA in a crystal structure [43] (Figure 2F). Mutations at R351, R357, R379, H382, R407 and R413 obliterated suppression of HSV‐1 replication by EGR1 (Figure 2G). These residues all make hydrogen bonds with DNA bases, suggesting that EGR1 suppresses HSV‐1 through sequence‐specific DNA interactions with the zinc fingers.
To further confirm the effects and to facilitate in vivo investigation (next section), we constructed recombinant HSV‐1 expressing EGR1 using the bacterial‐artificial chromosome (BAC) technology based on BAC‐derived WT HSV‐1 [44]. Thus, HSV1Egr1a and HSV1Egr1b from two independent BAC colonies were generated (Figure 3A,B). Instead of using WT virus as a control, we also constructed several control viruses, including HSV1GFP, HSV1Egr1(R379A), and HSV1Egr1(R407A) with the GFP gene or the corresponding mutated Egr1 gene inserted at the same location (Figure 3A,B). Furthermore, based on HSV1Egr1a, we replaced the inserted Egr1 with a kanamycin‐resistant gene to make HSV1Egr1rescue, which served as an additional control virus. Growth curves in Neuro‐2a, Vero and HCEC cells consistently showed that HSV1Egr1a and HSV1Egr1b replicated with kinetics much slower than HSV1GFP, HSV1Egr1rescue, HSV1Egr1(R397A) and HSV1Egr1(R407A) (Figure 3C and Figure S3A), confirming EGR1 suppression of HSV‐1 replication.
EGR1 ectopically expressed from HSV‐1 represses viral replication in mouse TG. (A) Schematic of recombinant viruses showing that different transgenes driven by the CMV promoter were inserted between US9 and US10 genes. The order of virus derivation is indicated by arrows connecting the virus names. (B) N2A‐EGR1‐KO1 were infected with the indicated recombinant viruses for 8 h before Western blot analysis. (C) Neuro‐2a cells were infected with the indicated viruses (MOI = 0.1) before growth curve analysis. n = 3 biologically independent samples. (D–F) Mice were infected with recombinant viruses (5 × 104 PFU/eye). Viral titers in eye swabs at the indicated times were determined (D). Viral titers (E) and genome levels (F) in TG at 5 days post‐infection (5 dpi) were quantified. (G) Mice were infected as in D for 5 days before titration of viruses in eyes (left) or qPCR analysis of the viral genome (right). (H) Mice were infected with either a low (5 × 104 PFU/eye) or high (5 × 105 PFU/eye) dose of the indicated recombinant viruses before determination of viral titers in eye swabs (left), or viral genome levels in TG (middle) or brainstems (right). (I) After infection as in D, the severity of mouse facial lesions was scored: 0, no lesion; 1, slight lesions in small areas; 2, nearly half of the face covered by lesions; 3, most of the face covered by lesions. (J) Viral genome levels in TG at 29 dpi were analyzed by qPCR. For mouse experiments, each point represents a value from one mouse (for eye swabs), TG or brainstem and the numbers of mice, TG or brainstems are labeled on the graphs. Data were analyzed by one‐way ANOVA with Dunnett's multiple comparisons tests (D, E, F, H) or two‐tailed unpaired t tests (G, I, J) and are presented as mean ± SD.
Recombinant HSV‐1 Expressing EGR1 Exhibited Reduced Replication in Mouse Ganglia In Vivo
2.4
We then used the recombinant viruses in our mouse model. After corneal inoculation, compared with both HSV1Egr1(R379A) and HSV1Egr1(R407A), average HSV1Egr1a titers were moderately lower at 1 and 3 dpi in eyes, but substantially lower at 5 dpi in both eyes and TG (Figure 3D,E), corresponding to greatly lower viral genome levels in TG at 5 dpi (Figure 3F). Accordingly, relative to HSV1GFP, average HSV1Egr1b titers were slightly or insignificantly lower in eyes at 1 and 3 dpi but significantly lower (by ∼1 log) in both eyes and TG at 5 dpi (Figure S3B,C). HSV1Egr1a titers in eyes and genome levels in TG were also greatly lower than HSV1Egr1rescue at 5 dpi (Figure 3G), demonstrating that the defects of HSV1Egr1a replication were due to EGR1 expression rather than unintended mutations. To examine whether the reduced titers in TG could be due to less input virus from eyes, we performed another experiment, in which, besides using the same dose of HSV1GFP and HSV1Egr1b, a group of mice was infected with a 10‐fold higher dose of HSV1Egr1b than HSV1GFP, such that HSV1Egr1b titers in eyes at 1 dpi were significantly higher than HSV1GFP titers. Even so, HSV1Egr1b titers in TG and brainstems at 5 dpi were still substantially lower than HSV1GFP (Figure 3H), suggesting that EGR1 mainly suppresses HSV‐1 replication in neuronal tissues. Furthermore, relative to both HSV1Egr1rescue and HSV1Egr1(R379A), HSV1Egr1a caused significantly less severe facial lesions at 9 or 10 dpi (Figure 3I) and lower latent genome levels at 29 dpi (Figure 3J). Therefore, EGR1 can strongly repress acute HSV‐1 replication in vivo, resulting in less severe disease and lower latent viral loads.
EGR1 Associates with Multiple Sites on the HSV‐1 Genome
2.5
To understand how EGR proteins regulate HSV‐1 replication, we first investigated EGR1 binding to the viral genome by performing chromatin immunoprecipitation‐sequencing (ChIP‐seq) using an EGR1 antibody to precipitate endogenous EGR1 in HSV‐1 infected Neuro‐2a cells. Peak calling analysis applying a threshold of adjusted P values < 0.05 identified 43 viral peaks (Figure 4A) and 923 host peaks. Motif search identified GCGKGGGCG as the most enriched motif in host peaks (Figure 4B), highly in line with the known consensus sequence of EGR1. Although the viral peaks were not numerous enough to identify a motif, 19 (44%) of them conform to the consensus (Figure 4A). Interestingly, the highest viral peak, one near the 3’ end of the ICP0 gene, contains no GCGKGGGCG but instead a run of nine GCGTGGGAG sequences, suggesting that variations of the consensus are permitted. The second‐highest peak covers five consecutive GCGGGGGCG sites (here referred to as 5sites) that are 978 nt upstream of the ICP4 transcription start site (TSS). Because 5sites is positioned at the edge of a repeat region, each of its two copies is also within introns of either the ICP22 or ICP47 gene, 281 nt downstream of the respective TSS (Figure 4C). The third‐highest peak is located within the LAT promoter. However, instead of being positioned on the GCGGGGGCG sequence (here named site1) 12 nt upstream of the LAT TSS as reported previously [36], the major peak is centered at a region containing two overlapping consensus sequences GCGGGGGCGTGGGCG 574‐588 bp upstream of the TSS (here named site2/3) that was unrecognized previously (Figure 4C). Other peaks were randomly distributed across the genome without noticeable patterns.
EGR1 stimulates LAT expression by binding to its promoter. (A) Neuro‐2a cells were infected with HSV‐1 (MOI = 10) for 5 h before ChIP‐seq analysis using an EGR1 antibody. Coverage plots along the HSV‐1 genome with terminal repeat regions deleted are generated by the IGV software. Signals from the input and IP samples are displayed in grey and blue respectively. Red vertical sticks represent peaks. Green vertical sticks represent positions of GCGKGGGCG or its complementary sequence. A region containing the LAT gene is expanded below. (B) The top motif in host peaks identified by the Homer motif search program. (C) Schematic representation of the sequences surrounding 5sites (upper) and the LAT upstream regulatory sequence (below) of strain KOS, as well as the constructed plasmids with the corresponding LAT upstream sequences inserted into the pGL3 vector. The consensus sequences are shown in blue and the red lines across them indicate deletion of the sequences. (D, E) Neuro‐2a cells were co‐transfected with 100 ng/mL of the indicated pGL3‐derived plasmids, 50 ng/mL of RL‐CMV plasmid, and 300 ng/mL of the EGR protein expressing plasmid for 24 h before luciferase assays. (F) The indicated cells were transfected with 100 ng/mL of the indicated plasmids for 24 before luciferase assays. (G) Neuro‐2a cells were transfected with 400 ng/mL of the indicated plasmids for 24 h, then infected with HSV‐1 (MOI = 10) for 5 h before ChIP‐qPCR analysis of the enrichment of transfected Flag‐tagged proteins on the indicated sites using a Flag antibody. (H, I) Viral LAT intron levels normalized to viral genome levels in TG at the indicated times after corneal inoculation (5 x 104 PFU/eye) of mice with the indicated viruses. n = 3 (D, E, F) or 6 (G) biologically independent samples. For H and I, the numbers of TG are labeled on the graphs. Data were analyzed by two‐tailed unpaired t tests (F, G, H right, I) or two‐way ANOVA with Sidak's multiple comparisons tests (D, E), one‐way ANOVA with Dunnett's multiple comparisons tests (H left) and are presented as mean values ± SD.
EGR1 Binds to the LAT Promoter to Stimulate LAT Expression
2.6
To determine whether the sites on the LAT promoter confer gene regulation by EGR1, we first used a previously described plasmid [45] (here named pGL3‐LATproL), whose luciferase expression is driven by a sequence between −876 and 36 bp relative to the TSS, which contains both site1 and site2/3. We also constructed a plasmid (pGL3‐LATproS) with a short LAT promoter (from ‐138 to 36 bp) containing only site1. Relative to the empty vector, proS, but not proL, strongly promoted luciferase expression, suggesting that there are cis elements between −139 and −876 bp that negatively regulate LAT expression (Figure 4D, E). Regardless, transfected EGR1 increased the proS activity by 2.4 fold and the proL activity by 12 fold, suggesting that EGR1 can promote LAT expression through both proS and proL sequences, with more contributions from proL. Furthermore, all EGR proteins could strongly stimulate the proL promoter activity (Figure 4E). Although deletion of site1 largely eliminated EGR1 stimulation of proS activity, disruption of site2/3 only partially reduced EGR1 stimulation of proL activity (Figure 4E), suggesting that site1, site2/3, and non‐canonical sites all contribute to proL activation by EGR1. Moreover, EGR1 knockout slightly reduced proS activity but markedly reduced proL activity (Figure 4F), indicating that endogenous EGR1 promotes LAT expression mainly through the long promoter. We then performed chromatin immunoprecipitation (ChIP)‐qPCR in Flag‐EGR1‐transfected and HSV‐1 infected N2AEGR1KO cells using a Flag antibody and observed significant enrichment of EGR1 on both site1 and site2/3, with higher enrichment on site 2/3 (Figure 4G), suggesting that EGR1 indeed associates with these sites, particularly site2/3.
LAT is known to be poorly expressed in cell culture. As expected, we could not detect the LAT intron in infected Neuro‐2a cells. Therefore, we assessed the effects of EGR1 on LAT expression in the mouse model using the recombinant viruses. Given the strong effects of EGR1 on viral replication, we normalized LAT intron levels to viral genome levels such that effects on gene expression were analyzed independent of DNA replication. At 5 dpi, HSV1Egr1a expressed ∼one log more LAT intron per genome than both HSV1Egr1(R379A) and HSV1Egr1(R407A) (Figure 4H). The differences were recapitulated by comparing HSV1Egr1b and HSV1GFP at 5 dpi (Figure 4H), and comparing HSV1Egr1a with both HSV1Egr1(R379A) and HSV1Egr1rescue at 29 dpi (Figure 4I). These results suggest that EGR1 can increase LAT expression in vivo. Although LAT is implicated in repressing lytic gene expression in vivo [15, 16, 17, 18], its low expression in cell culture argues against its involvement in EGR1 function there. Indeed, a LAT‐deletion virus (KdlLAT) and its rescued virus (KFSLAT) [16] were similarly suppressed by EGR1 in both Neuro‐2a and 293T cells (Figure S4), suggesting that other mechanisms account for repression of HSV‐1 replication by EGR1 in cell culture.
EGR1 Activates Certain IE Genes at Early Times but Globally Represses Lytic Genes at Late Times Post‐Infection
2.7
To understand how EGR1 repressed HSV‐1 replication in culture, we performed RNA‐seq comparing EGR1‐transfected and vector‐transfected N2A‐EGR1‐KO1 cells that were subsequently infected. Cells were harvested at 3 and 10 hpi, corresponding to IE and L times, respectively. The ratios of total viral reads to total reads were unaffected at 3 hpi but greatly reduced at 10 hpi by EGR1 (Figure 5A). Interestingly, total viral IE mRNA reads were even modestly increased by EGR1 at 3 hpi (Figure S5A). Heatmaps display global repression of viral gene expression by EGR1 at 10 but not 3 hpi (Figure 5B). Accordingly, RT‐qPCR data showed that EGR1 had no effect on the E gene TK and L gene gC (Figure S5B), and even slightly upregulated IE genes ICP0, ICP4, and ICP22 at 3 hpi (Figure 5C). However, all these genes were significantly repressed by EGR1 at 10 hpi. ACV had little effect at 3 hpi, but abrogated the repression by EGR1 at 10 hpi (Figure 5C, S5B). A protein synthesis inhibitor, cycloheximide (CHX), also abolished the repression of ICP0 and ICP4 by EGR1 at 9 hpi (Figure S5C). Consistently, Western blots showed that IE proteins ICP0 and ICP4 were modestly upregulated by EGR1 at 3 hpi, but all viral proteins tested were downregulated at 10 and 15 hpi (Figure 5D). Therefore, EGR1 plays a dual role in HSV‐1 gene expression in moderately upregulating certain IE genes at early times, but globally repressing viral genes at late times in a manner dependent on viral DNA synthesis and de novo protein synthesis.
Effects of EGR1 on HSV‐1 lytic gene expression. (A‐B) N2A‐EGR1‐KO1 cells were transfected with 400 ng/mL of pcDNA or pEGR1 for 24 h, then infected with HSV‐1 (MOI = 0.5) for 3 or 10 h before RNA‐seq analysis. A: The ratios of total viral reads to total reads. B: A heatmap showing the normalized transcripts per million (TPM) value for each viral transcript. (C) N2A‐EGR1‐KO1 cells were transfected with 400 ng/mL of the indicated plasmids for 24 h, then infected with HSV‐1 (MOI = 0.5) in media containing 100 µM ACV or DMSO for 3 or 10 h before RT‐qPCR analysis. (D) Neuro‐2a cells were transfected with 400 ng/mL of the indicated plasmids for 40 h, then infected with HSV‐1 (MOI = 5) for the indicated times before Western blot analysis. (E) Viral mRNA levels normalized to viral genome levels in TG at day 5 after corneal inoculation of mice (5 x 104 PFU/eye) with the indicated viruses. (F) Neuro‐2a cells were co‐transfected with 100 ng/mL of the indicated pGL3‐derived plasmids, 50 ng/mL of the RL‐CMV plasmid and 300 ng/mL of pcDNA or pEGR1 before luciferase assays. n = 3 (A, B, C, F) biologically independent samples. For E, the numbers of TG are labeled on the graphs. Data were analyzed by two‐way ANOVA with Sidak's multiple comparisons tests (A, C, F), one‐way ANOVA with Dunnett's multiple comparisons tests (E) and are presented as mean values ± SD.
Regarding gene regulation in vivo, in the mouse samples mentioned above for analysis of LAT expression, we also analyzed lytic transcripts. Relative to HSV1Egr1(R379A) and HSV1Egr1(R407A), HSV1Egr1a expressed similar amounts of TK and gC mRNAs (Figure S5D), and slightly more ICP4 (with differences of 0.3–0.5 log) and possibly ICP0 (the difference from one of the two control viruses was significant) mRNAs per viral genome (Figure 5E), suggesting that EGR1 might be able to upregulate certain IE genes in vivo too in the background of strong repression of viral replication.
To understand how EGR1 regulates lytic genes, we considered possible contributions from 5sites, given its location near multiple IE genes and the previous report that EGR1 could repress gene expression through this sequence [37]. However, we found that EGR1 elevated luciferase gene expression from a promoter containing 5sites by 2.6 fold after co‐transfection in Neuro‐2a cells (Figure 5F). This result is consistent with the upregulation of certain IE genes by EGR1, but also indicates that EGR1 represses viral replication by other mechanisms.
EGR Proteins can Upregulate Immune Regulators IRF7 and ISG15
2.8
Next, we explored the possibility that EGR1 represses HSV‐1 replication indirectly through host genes. Given that the repression required an EGR1 activation domain (Figure 2E), we focused on host genes activated by EGR1. The above RNA‐seq data identified 344 host genes that were upregulated by EGR1 at both 3 and 10 hpi (Figure 6A). Intersection of these genes with those containing putative host binding sites in the ChIP‐seq data resulted in 33 genes. However, although these genes may represent direct EGR1 targets, KEGG pathway analysis identified few related to infection or immunity (Figure S6). Concerned about potentially overlooking important targets due to false negatives in the ChIP‐seq data, we considered all the 344 Egr1‐responsive genes and discovered an overrepresentation of genes related to the immune system and viral infectious diseases in KEGG analysis (Figure 6A). Notably, nineteen genes are related to both immune responses and viral infections, including those involved in inflammatory or adaptive immune responses, such as CD4, IL6, Cx3cl1, and Tgfb1. To understand the effects in cell culture, we focused on innate immune‐related genes among them, particularly IRF7 and ISG15, which were upregulated by EGR1 by 4.9 and 2.9 fold, respectively, in the RNA‐seq data. RT‐qPCR analyses showed stimulation of these genes by all four EGR proteins (Figure 6B,C).
EGR1 can upregulate immune genes IRF7 and ISG15. (A) Left, Venn diagram for EGR1 activated genes at 3 and 10 hpi according to the RNA‐seq data. The following criteria were used to screen the genes: reads (pEGR1 group) > 20 transcripts per million, log2(fold‐change, pEGR1/pcDNA) > 1, adjusted P < 0.05. Right, KEGG pathway classification of the 344 genes with the arrow pointing to the list of genes related to viral infections and the immune system. (B) N2A‐EGR1‐KO1 cells were transfected with 400 ng/mL of pcDNA, pEGR1 or pEGR1(R407A) for 24 h, then infected with HSV‐1 (MOI = 0.5) for the indicated times before RT‐qPCR analysis. (C) N2A‐EGR1‐KO1 cells were transfected with 400 ng/mL of the indicated plasmids for 40 h before RT‐qPCR analysis. (D) Mice were infected with KOS (5 × 104 PFU/eye) for the indicated times before RT‐qPCR analysis of the indicated mRNAs in TG. (E) Correlation between the indicated mRNAs in HSV‐1 infected TG at 3 dpi. (F‐I) Mice were infected with either a low (5 × 104 PFU/eye) or high (5 × 105 PFU/eye) dose of the indicated recombinant viruses before RT‐qPCR analysis of TG. Shown are the relative IRF7 mRNA copy numbers per TG (F), per viral genome (G), per IFN‐α mRNA copy (I), and the relative IFN‐α mRNA copy numbers per viral genome (H). (J) Neuro‐2a cells were co‐transfected with 100 ng/mL of the plasmids containing the indicated promoters, 50 ng/mL of RL‐CMV plasmid and 300 ng/mL of pcDNA or pEGR1 for 24 h before luciferase assays. n = 3 (B, C, J) biologically independent samples. For D‐I, the numbers of TG are labeled on the graphs. Data were analyzed by one‐way ANOVA with Dunnett's multiple comparisons tests (C, D, F, G, H, I), two‐way ANOVA with Sidak's multiple comparisons tests (B, J) or simple linear regression (E) and are presented as mean ± SD.
In the mouse model, IRF7 and ISG15 mRNAs were both strongly upregulated during acute (days 3 and 5) but not latent (day 40) infection (Figure 6D). Notably, at 3 dpi, mRNA levels of either Egr1 or Egr2 were strongly correlated with those of either IRF7 or ISG15 in HSV‐1‐infected but not in mock‐infected TG samples (Figure 6E and Figure S7A). RT‐qPCR analysis of RNA purified in parallel with DNA from the HSV1GFP and HSV1Egr1b infected TG samples mentioned above (Figure 3H) showed that despite the great differences in viral genome levels, HSV1Egr1b‐infected TG expressed IRF7 at levels similar to HSV1GFP‐infected TG (Figure 6F), such that HSV1Egr1b‐infected TG expressed substantially more IRF7 per viral genome (by 1.1–1.5 logs) (Figure 6G). In contrast, IFN‐α expression was largely coupled with viral genome levels, with HSV1Egr1b‐infected TG expressing slightly more IFN‐α mRNA/genome (by <0.5 log) than HSV1GFP (Figure 6H), such that the IRF7/IFN‐α ratios were also significantly higher in HSV1Egr1b‐infected TG (Figure 6I). Similar results were obtained with ISG15 (Figure S7B). Additionally, brainstems of mice infected with the higher dose of HSV1Egr1b expressed more IRF7 and ISG15 per viral genome than HSV1GFP, with the lower dose showing the same trend without significance (Figure S7C). Taken together, these results suggest that EGR1 can upregulate IRF7 and ISG15 both in cell culture and in vivo.
Despite such effects, neither IRF7 nor ISG15 contains an EGR1 consensus sequence, and we did not detect EGR1 binding peaks in IRF7 or ISG15 promoters in our ChIP‐seq data. Nevertheless, in luciferase assays, co‐transfected EGR1 significantly increased luciferase expression from reporter plasmids containing either an IRF7 or ISG15 promoter (Figure 6J), suggesting that EGR1 may activate their promoters indirectly or in a consensus‐sequence‐independent manner.
EGR1 Suppresses HSV‐1 Replication by Stimulating IRF7 Expression
2.9
Both IRF7 and ISG15 are ISGs (https://isg.data.cvr.ac.uk). The anti‐HSV‐1 functions of IRF7 and ISG15 have been reported in mouse models of herpes simplex encephalitis [46, 47]. Accordingly, knockdown of either IRF7 or ISG15 increased HSV‐1 yields in Neuro‐2a cells (Figure 7A and Figure S8A). However, only the siRNA against IRF7, when co‐transfected with the EGR1 expressing plasmid, could eliminate repression of HSV‐1 replication by EGR1 (Figure 7B). Moreover, transfected IRF7, but not ISG15, significantly reduced HSV‐1 yields (Figure 7C). IRF7 expression was upregulated by HSV‐1 at late times post‐infection (Figure 7D), consistent with the late repressive effects of EGR1, whereas ISG15 was modestly upregulated only at early times (Figure S8B). These results indicate that IRF7 is the major factor mediating EGR1 suppression of HSV‐1 replication at least in Neuro‐2a cells. Consistent with its known role in type I IFN response [48], IRF7 overexpression increased IFN‐stimulated response elements (ISRE) activity by 62 fold (Figure 7E). Accordingly, EGR1 could also increase ISRE activity, but, unexpectedly, this effect was independent of IFN‐α treatment (Figure 7F). Furthermore, an inhibitor of the type I IFN receptor (IRFAR1) had little effect on suppression of viral replication by either EGR1 or IRF7 (Figure 7G), and both EGR1 and IRF7 could increase ISG15 expression in the presence of the IRFAR1 inhibitor (Figure S8C). These results suggest that IRF7 and EGR1 can boost innate immune response against HSV‐1 in an IFN‐independent fashion.
EGR1 suppresses HSV‐1 replication by enhancing IRF7 expression. (A) Neuro‐2a cells were transfected with the indicated siRNAs for 24 h, then infected with HSV‐1 (MOI = 0.2) for 48 h before virus titration. (B) Left, Neuro‐2a cells were transfected with the indicated siRNAs for 18 h, then transfected with the indicated plasmids for 20 h, then infected with HSV‐1 (MOI = 0.2) for 44 h before virus titration. Right, Neuro‐2a cells were transfected with the indicated siRNAs for 16 h, then transfected with the indicated plasmids for 22 h, then infected with HSV‐1 (MOI = 0.5) for 24 h before virus titration. (C) Left, Neuro‐2a cells were transfected with 400 ng/mL of indicated plasmids expressing flag‐tagged proteins for 24 h before Western blot analysis using a flag antibody. Right, Neuro‐2a cells were transfected with the indicated plasmids for 24 h, then infected with HSV‐1 (MOI = 0.2) for 48 h before virus titration. (D) RT‐qPCR analysis of IRF7 mRNA at the indicted times after HSV‐1 infection of Neuro‐2a cells (MOI = 10). (E) N2A‐EGR1‐KO1 cells were co‐transfected with 300 ng/mL of pcDNA or pIRF7, 100 ng/mL of the ISRE‐luc plasmid and 50 ng/mL of the RL‐CMV plasmid for 24 h before luciferase assays. (F) N2A‐EGR1‐KO1 cells were untreated or pretreated with IFN‐α (500 IU/mL) for 12 h, then co‐transfected with 300 ng/mL of pcDNA or pEGR1, 100 ng/mL of ISRE‐luc, 50 ng/mL of RL‐CMV in the absence (left) or presence (right) of IFN‐α for 24 h before luciferase assays. (G) Neuro‐2a cells were transfected with the indicated plasmids for 22 h, then treated with DMSO or 2 µM IFNAR1 inhibitor (IFN alpha‐IFNAR‐IN‐1) for 16 h, and then infected with HSV‐1 (MOI = 0.2) in the presence of DMSO or the inhibitor for 40 h before virus titration. n = 3 biologically independent samples for all panels. Data were analyzed by one‐way ANOVA with Dunnett's multiple comparisons tests (A, C, D) or two‐way ANOVA with Sidak's multiple comparisons tests (B, G), two‐tailed unpaired t tests (E, F) and are presented as mean ± SD.
Since EGR1 can also suppress HSV‐1 replication in 293T cells, we performed similar experiments in such cells, and observed upregulation of IRF7 and ISG15 by EGR1 too, albeit modestly (Figure S8D). Overexpression of IRF7, but not ISG15, markedly reduced HSV‐1 yields (Figure S8E). The IRFAR1 inhibitor had little effect on suppression of HSV‐1 replication by EGR1 or IRF7 (Figure S8F). Therefore, at least some aspects of the mechanisms by which EGR1 suppresses HSV‐1 replication may also function in non‐neuronal cells.
Discussion
3
Quick responses to viral infections are typically mediated by IFN‐dependent innate immunity. However, this work indicates that some host proteins that normally respond to physiological stimuli can be efficiently upregulated to suppress lytic infection independent of IFNs. Our results, along with previous reports [32, 33, 34, 35], consistently demonstrate induction of EGR1 expression by HSV‐1 under various conditions. Unexpectedly, we found that such induction is mediated by a viral E3 ubiquitin ligase (ICP0) rather than previously identified pathways activated by extracellular signals [27]. Although ICP0 is not a direct activator of gene expression, there are precedents of ICP0 inducing host genes, including Dux4 [49] and IFNs [50]. We speculate that the mechanism may be related to ICP0's chromatin remodeling function, as it has been reported that ICP0 can perturb host centromeres and telomeres [51, 52, 53].
Previous studies have yielded conflicting conclusions regarding the role of EGR1 in HSV‐1 infection [33, 37]. We note that the different previous studies and our present study used different cell lines and experimental conditions. Our results support a predominantly suppressive role during lytic infection in neuronal cells, while also hinting at a dual role because EGR1 can increase the expression and promoter activities of certain IE genes under certain conditions. In vivo, although previous reports showed that EGR1 knockout caused decreased corneal disease, brain infection, and mortality of mice [33, 34], those studies employed constitutive knockout, and the strongest effects were observed in corneas, the sites of inoculation, so the effects on mortality and brain infection could be secondary. Notably, their results showed that despite 155‐fold higher HSV‐1 titers in the eyes of Egr1^+/+^ than Egr1^−/−^ mice at 7 dpi, the titers in brains of Egr1^+/+^ and Egr1^−/−^ mice were only 10 fold different and the titers in TG were even not significantly different [33], which would be consistent with a suppressive role of EGR1 in neuronal tissues. Using the recombinant virus approach, we observed strong suppression of HSV‐1 replication by ectopically expressed EGR1, particularly in mouse TG, suggesting that EGR1 can robustly repress HSV‐1 replication in the peripheral nervous system.
Our studies support the previous observation that EGR1 binds to site1 within the LAT promoter [36] and additionally discovered site2/3 that contributes even more to LAT expression stimulated by EGR1. However, contrary to the repression of the LAT promoter by EGR2 and no effect of EGR1 and EGR3 in human neuroblastoma SY5Y cells in the previous report, we observed significant activation of the LAT promoter by all EGR proteins. However, our results argue against the possibility that EGR proteins are solely responsible for LAT expression since Egr gene expression largely dropped to baseline during latency, as LAT expression remains high. Nevertheless, our result that EGR1 knockout resulted in reduced LAT promoter activities indicates that endogenous EGR1 contributes to the expression of LAT, which is known to repress lytic gene expression in mouse TG in vivo [15, 16, 17].
According to our results, the major mechanism by which EGR1 suppresses HSV‐1 replication in cell culture is stimulation of IRF7 expression. Given the importance of IRF7 in multiple aspects of immunity [48], its upregulation is likely to be consequential in vivo too. Surprisingly, the effects of IRF7 and EGR1 on HSV‐1 replication were independent of the type I IFN receptor. IFN‐independent antiviral responses mediated by IRF7 have been reported previously and hypothesized to involve direct recognition of ISREs by IRF7 [54, 55]. Although ISG15 contributed less than IRF7 to EGR1 suppression of HSV‐1 replication in Neuro‐2a cells, it may contribute more in other contexts given its known antiviral functions. Additionally, our RNA‐seq data identified many other innate and adaptive immune genes upregulated by EGR1 that may contribute to the suppression in vivo (Figure 6A). Thus, despite not being an ISG itself, EGR1 can nonetheless boost immune responses to influence the outcomes of infections.
Some limitations and unanswered questions arising from this work warrant further investigation in the future. One limitation of using mutants with altered EGR1 functions is that the effects could be indirect through host pathways. To investigate the specific interactions between EGR proteins and the viral genome, mutant viruses with altered DNA binding sites may be needed. Another limitation pertains to our reliance on recombinant viruses expressing EGR1 for in vivo investigation. Although this approach was helpful for exploring potential effects of increased EGR1 expression, further studies are required to elucidate the role of endogenous EGR1 using approaches such as gene knockout. Moreover, the different viral loads between EGR1‐expressing and control viruses in latently infected ganglia prevented assessment of the impact of EGR1 on reactivation using such ganglia. Thus, the role of EGR1 in reactivation remains an important unresolved question. Furthermore, since multiple genes upregulated by EGR1 are involved in host defense against viral infections more broadly, it would be intriguing to investigate whether these mechanisms extend to other viruses as well.
Experimental Section
4
Cells
4.1
African green money kidney Vero cells (ATCC Cat# CCL‐81, RRID:CVCL_0059), mouse brain neuroblastoma Neuro‐2a cells (ATCC Cat# CCL‐131, RRID:CVCL_0470), human embryonic kidney 293T cells (ATCC Cat# CRL‐3216, RRID:CVCL_0063) and mouse macrophage Raw264.7 cells (ATCC Cat# TIB‐71, RRID:CVCL_0493) were maintained and used as previously described [56, 57]. SV40‐immortalized human corneal epithelial cells (BCRJ Cat# 0349, RRID:CVCL_1272) were a gift from Riken Institute of Physical and Chemical Research, Japan. These cells were used at fewer than 30 passages after they were originally obtained from ATCC. For all cell lines, routine mycoplasma tests were performed to guarantee that the cell lines were contamination‐free. For the construction of N2A‐EGR1KO cells, DNA sequences targeting tattaccgccgctgccctct (N2A‐EGR1‐KO1) or ctgcagatctctgacccgtt (N2A‐EGR1‐KO2) sequences of the mouse EGR1 coding region were synthesized and cloned into the PX459 plasmid vector. The resulting plasmids were transfected into Neuro‐2a cells (200 ng/well in a 24‐well plate). 48 h later, the supernatant was replaced with fresh medium with 1 µg/mL puromycin. When cells reached nearly 100% confluency, they were transferred into a 96‐well plate for single colony selection. Single‐cell colonies were transferred to 24‐well plates. After reaching confluency again, a fraction of the cells was used for genomic DNA extraction, PCR amplification, sequencing, and Western blot analysis of EGR1 protein expression to confirm knockout. The rest of the cells were propagated further in 1 µg/mL puromycin until the cell lines were stable.
Viruses
4.2
HSV‐1 strain KOS, ICP0‐null mutant 7134 virus, and its rescuant 7134R virus were obtained, propagated as described previously [57]. KdlLAT and KFSLAT viruses were generous gifts from Donald M. Coen at Harvard Medical School [16]. The recombinant HSV‐1 viruses that express human EGR1 were constructed using BAC technology based on HSV‐1 strain KOS as described previously [11]. Briefly, to construct HSV1Egr1a, HSV1Egr1b, HSV1Egr1(R379A), and HSV1Egr1(R407A), the kanamycin‐resistance (Kan‐r) gene plus I‐SceI site was inserted into the NheI site of pcDNAEGR1human plasmid, or a plasmid expressing the corresponding EGR1 mutant. EGR1‐BACf and EGR1‐BACr primers, with ∼40 bp homologous arms between US9 and US10, were used to amplify the EGR1‐Kan‐r cassette, each containing a CMV promoter and poly(A) signal sequence. The cassettes were then transformed into bacteria according to the two‐step red‐mediated recombination protocol [58] to generate BAC genomic DNA, which was used for virus generation in Vero cells. HSV1GFP was constructed in the same way except that an EGFP gene was inserted instead of the Egr1 gene. HSV1Egr1rescue was constructed by replacing the Egr1 gene of HSV1Egr1a with the Kan‐r gene. Virus titration by plaque assays and infection experiments was performed as described previously [59].
Plasmids
4.3
Human Egr (hEgr1) genes were PCR amplified from cDNA derived from 293T cells. The human Egr1 gene was inserted between the EcoRI and HindIII sites of FLAG‐HA‐pcDNA3.1‐ vector (Addgene, 52535, here referred to as pcDNA) to make pEGR1. Human Egr2, Egr3, and Egr4 genes were each inserted between the BamHI and EcoRI sites of FLAG‐HA‐pcDNA3.1 to make pEGR2, pEGR3, and pEGR4. The plasmids expressing EGR1 truncated mutants were constructed based on the pLVX‐EF1α‐IRES‐mCherry vector (Biofeng, 631987, here referred to as pLVX). The human Egr1 gene was inserted between the EcoRI and NotI sites of pLVX to make pLVX‐EGR1. Most mutations were introduced into pLVX‐EGR1 using a Q5 Site‐Directed Mutagenesis Kit (NEB, #E0552S). For example, using pLVX‐EGR1 as the template, PCR was performed using EGR1Δ(AD‐N)‐F and EGR1Δ(AD‐N)‐R primers to generate linear EGR1Δ(AD‐N). Then the linear products were incubated with the Kinase‐Ligase‐DpnI enzyme mix, which circularized the DNA product and removed the template. E354A and T385A mutations were introduced into pLVX‐EGR1 using the overlap PCR technique [60]. To construct the EGR1 shRNA plasmids, we selected target sequences 5'‐ATGCGTAACTTCAGTCGTA‐3' (sh1) and 5'‐CAGGACTTAAAGGCTCTTAAT‐3' (sh2). The sequences were annealed and inserted into the pLKO.1 vector (Addgene #128073). pIRF7(human) and pISG15(human) plasmids were obtained from MiaoLingBio and were verified by sequencing and Western blots. The ICP0‐expressing plasmid (pICP0) was constructed previously as described [61]. pGL3‐LATproLong was described previously and generously provided by Clinton Jones [45, 62]. The EGR1 promoter sequence was amplified using total DNA extracted from Neuro‐2a cells as a template, and EGR1pro‐lucF and EGR1pro‐lucR primers, and inserted between HindIII and NcoI sites of the pGL3 vector (Addgene #212936) to make pGL3‐EGR1pro. The other luciferase‐expressing plasmids, including pGL3‐LATproS, pGL3‐LATproS‐Mut, pGL3‐LATproL‐Mut, pGL3‐IRF7pro, pGL3‐IRF5pro, and pGL3‐ISG15pro, were constructed by inserting the corresponding promoter sequences, synthesized by TsingKeBio, between KpnI and HindIII sites of the pGL3 vector. pRL‐CMV was from Promega (#E2261). Primer sequences used for amplification are provided in Table S3.
Chemical Inhibitors
4.4
The following inhibitors were used at the following final concentrations: PDK1 inhibitor (BX‐795, 1 µm, MedChemExpress), PI3K inhibitor (PI‐103, 500 nm, MedChemExpress), MEK1/2 inhibitor (Binimetinib, 1 µm, MedChemExpress), AKT inhibitor (MK‐2206, 1.25 µm, MedChemExpress), ATM inhibitor (KU‐55933, 20 µm, MedChemExpress), FNAR1 inhibitor (IFN alpha‐IFNAR‐IN‐1, HY‐12836A, 2 µm, MedChemExpress), and JNK inhibitor (SP600125, 10 µm, MedChemExpress). The chemical inhibitors were dissolved in DMSO. The DMSO volume was equal to the volume of the inhibitors and less than 0.1% of the final cell culture medium. Cells were treated with the inhibitors from 6 h before infection until the cells were harvested after infection.
Transfection
4.5
Plasmids and siRNAs were transfected using Lipomaster 3000 transfection reagent (Vazyme Biotech Co., TL301‐01) according to the manufacturer's instructions. Synthetic siRNA mimics were purchased from Ribobio. The target sequences of siRNAs are listed in Table S4. The concentrations of plasmids and siRNA are specified in the figure legends.
Luciferase Assays
4.6
Luciferase activities were measured using a Dual Luciferase Reporter Gene Assay Kit (catalogue no. 11402ES60; Yeasen Biotechnology) according to the manufacturer's instructions. A Varioskan Flash Spectral Scanning Multimode Reader (Thermo Scientific) was used for quantification of signals.
qPCR and RT–qPCR
4.7
To quantify transcripts, total RNA was extracted using an Eastep Super total RNA extraction kit (catalogue no. LS1040m, Promega). To simultaneously quantify the viral genome and transcripts, DNA and RNA were isolated using a DNA/RNA Isolation Kit (TIANGEN). RNA was reverse transcribed using a HiScript II Q Select RT supermix kit (catalogue no. R233‐01; Vazyme Biotech) to make cDNA. Both cDNA and viral genomic DNA were analyzed by qPCR using a ChamQ Universal SYBR qPCR kit (catalogue no. Q711‐02/03; Vazyme Biotech). Transcript levels were normalized to GAPDH mRNA levels, and viral genome levels were normalized to Adipsin gene levels. Serially diluted total DNA or RNA was used to generate standard curves. qPCR primer sequences are listed in Table S5.
Western Blots
4.8
Western blotting was performed as previously described [44]. The following primary antibodies and dilutions were used: rabbit anti‐EGR1 antibody (1:1000, catalogue no. MA5‐15009; Invitrogen); mouse anti‐FLAG antibody (1:2000, catalogue no. F1804, SigmaAldrich); mouse anti‐ICP0 antibody (1:5,000, catalogue no. ab6513; Abcam), mouse anti‐ICP4 antibody (1:5,000, catalogue no. ab6514; Abcam); goat anti‐thymidine kinase antibody (1:500, catalogue no. sc‐28037; Santa Cruz Biotechnology); mouse anti‐gC antibody (1:1,000, catalogue no. 10‐H25A; Fitzgerald); rabbit anti‐β‐actin antibody (1:20,000, catalogue no. AC026, ABclonal). Secondary antibodies: HRP‐conjugated rabbit anti‐goat (1:5000, SouthernBiotech #6160‐05); goat anti‐mouse (1:5000, SouthernBiotech #1030‐05) and goat anti‐rabbit antibodies (1:5000, SouthernBiotech #4030‐05).
Mouse Procedures
4.9
All mouse housing and experimental procedures were in compliance with national ethical guidelines and approved by the Laboratory Animal Welfare and Ethics Committee of Zhejiang University with an approval code of ZJU20230321. Six‐week‐old male Institute for Cancer Research (ICR) mice were purchased from the Shanghai Laboratory Animals Center. For HSV‐1 infection, mice were anaesthetized by intraperitoneal injection of 0.4 mL of a mixture containing 4mg/mL pentobarbital sodium (Solarbio) and 500 µg/mL xylazine hydrochloride (catalogue no. X1251; Sigma–Aldrich) in sterile saline. Then, 3 µL of PBS containing HSV‐1 was dropped onto each scarified cornea. To collect eye swabs, mice were anaesthetized with 3% isoflurane (RWD Life Science) in oxygen with a flow rate of 0.5 mL/min using a V1 Table Top anaesthesia machine (Colonial Medical Supply). One eye of each live mouse was swabbed with one cotton‐tipped applicator, then the two applicators were placed into one glass bottle containing 1 mL of culture medium. For TG and brainstem acquisition, mice were euthanized by cervical dislocation, and TG and brainstems were removed and stored in Eppendorf tubes containing the appropriate solution (lysis buffer for DNA/RNA purification and culture medium for titer determination) at −80 °C before TG and brainstems were homogenized and processed for DNA/RNA purification or titer determination.
RNA‐Seq
4.10
To investigate the effects of HSV‐1 infection on the host transcriptome, Neuro‐2a cells were mock‐infected or infected with HSV‐1 for 5 h before being harvested for RNA purification. For assessing the effects of EGR1 on viral and host transcriptomes, N2A‐EGR1‐KO1 cells were transfected with 400 ng/mL of pcDNA or pEGR1 plasmid for 24 h and then infected with HSV‐1 (MOI = 0.5) for 3 h or 10 h before being harvested. RNA was purified using an ls1040 Eastep Super total RNA extraction kit (Promega). Library construction and sequencing were performed on the BGIseq500 platform (BGI‐Shenzhen). Raw data were filtered with SOAPnuke (v2.2.1) (https://github.com/BGI‐flexlab/SOAPnuke). Clean data were mapped to the mm10 mouse genome and KOSnorepeat HSV‐1 genome [11] by Bowtie2 (version 2.4.5) (http://bowtiebio.sourceforge.net/%2520Bowtie2%2520/index.shtml). Reads were calculated by RSEM (version 1.3.1) (https://github.com/deweylab/RSEM). Differential expression analysis was performed using DESeq2 (version 1.34.0) (http://www.bioconductor.org/packages/release/bioc/html/). The raw data files have been uploaded to the NCBI GEO repository under GEO access numbers GSE293057 and GSE300560.
ChIP‐Seq
4.11
A total of 5 x 10^6^ Neuro‐2a cells were cultured in each 100 mm plate and infected with KOS at an MOI of 10 for 5 h. Cells were crosslinked with 1% formaldehyde for 8 min and neutralized with 0.125 m glycine for 5 min, followed by two washes with 5 mL of ice‐chilled PBS. Cells were collected into 15 mL conical tubes using 5 mL of cold PBS and centrifuged at 2000 rpm for 5 min at 4°C. The resulting cell pellets were lysed in 500 µL of a lysis buffer (0.4% SDS, 10 mm EDTA, 50 mm Tris‐HCl, pH 8.0) for 30 min before being transferred to 1.5 mL Eppendorf tubes. The lysates were sonicated on ice using a SCIENTZ‐IID ultrasonic homogenizer for 8 cycles of 1 min each, with a 1 sec on and 1 sec off interval, at a power output of 90 watts. A 10 µL aliquot of the sonicated lysate was used for analyzing DNA size distribution. This aliquot was incubated with 110 µL of water containing 0.2 m NaCl at 65°C for 3–6 h to reverse crosslinking, followed by treatment with 0.2 mg/mL RNase A (Transgenic Biotechnology #GE101‐01) for 1 h at 37°C, then with 0.5 mg/mL proteinase K (Transgenic Biotechnology #GE201‐01) for 1 h at 55°C. DNA was purified using a HiPure Gel DNA Mini Kit (Magen #D2111‐03). DNA size distribution was analyzed by agarose gel electrophoresis. This DNA sample also served as an input control. After an optimal DNA fragment size range of 100–500 bp was achieved, immunoprecipitation reaction was carried out by incubating a mixture containing 1.7 mL of ChIP dilution buffer, 200 µL of the sonicated cell lysate, 5 µL of an anti‐EGR1 antibody (Invitrogen, catalogue no. MA5‐15009), 40 µL of Dynabeads Protein A (catalogue no. 10001; Invitrogen), and protease inhibitors (1 tablet per 50 mL of buffer, Roche #K562) with rotation overnight at 4°C. The beads were washed sequentially with an ice‐cold low‐salt wash buffer, high‐salt wash buffer, lithium chloride wash buffer, and Tris‐EDTA buffer. The compositions of these buffers and the ChIP dilution buffer were described previously [12]. For the immunoprecipitated sample, crosslinking was reversed, and DNA was extracted as above for the input sample. The purified input and immunoprecipitated DNA samples were submitted to Novogene for ChIP‐seq high‐throughput sequencing, followed by data analysis using protocols described previously [12]. Raw data have been uploaded to the NCBI GEO repository under the GEO access number GSE310510.
ChIP‐qPCR
4.12
Neuro‐2a cells in 100 mm plates were transfected and infected as described in the figure legends. The procedures for crosslinking, sonication, and DNA purification were the same as described above for ChIP‐seq. For immunoprecipitation using a Flag antibody, each reaction was conducted using 280 µL of sonicated cell lysate and 40 µL of Flag Affinity Gels (Merck Millipore) in the ChIP dilution buffer, supplemented with protease inhibitors for 3 h at 4 °C. After washing as described above for ChIP‐seq, the crosslinked DNA‐protein complexes were eluted by incubation with 90 µg of 3XFlag peptide (Beyotime) at 4 °C for 1 h. The purified input and immunoprecipitated DNA samples were analyzed by qPCR using the primers listed in Table S5.
Statistical Analyses
4.13
Statistical analyses were performed using Prism 8.01 for Windows (GraphPad Software, www.graphpad.com). All measurements were taken from distinct samples. All multiple comparisons have been corrected before adjusted P values are presented. The tests used are indicated in the figure legends. Basically, for comparisons between two groups, unpaired two‐tailed t‐tests were used. For comparisons among three or more groups, one‐way ANOVA with Dunnett's multiple comparisons test was employed. For multi‐condition group comparisons, two‐way ANOVA with Sidak's multiple comparisons test was used. Data are expressed as mean ± SD. A value of P < 0.05 was set to indicate statistical significance.
Data Availability
4.14
Raw RNA‐seq and ChIP‐seq data have been uploaded to the NCBI High‐throughput Gene Expression Omnibus repository (https://www.ncbi.nlm.nih.gov/geo/) with accession numbers of GSE293057 (RNA‐seq), GSE300560 (RNA‐seq), and GSE310510 (ChIP‐seq). The genomic sequence and the corresponding annotation files for HSV‐1 strain KOS were obtained from GenBank under accession number JQ673480.1. The terminal repeat sequences (TRL and TRS) were removed as described previously so that only one copy of the repeat sequence was retained (KOSnorepeat) [11]. Mouse (mm10) genomic sequence and annotation files were downloaded from the UCSC Genome Browser (https://genome.ucsc. edu/cgi‐bin/hgGateway?db=mm10). To model the EGR1 DBD‐DNA complex structure, the PDB entry (4R2D) was visualized in PyMOL, with key binding sites annotated. For analyzing protein signaling pathways, KEGG pathway classification analysis was conducted via Dr.TOM II resources (BGI). Some graphic elements were obtained from Server Medical Art (SMART) (https://smart.servier.com).
Author Contributions
D.P. conceived and supervised the study, contributed to some experiments in cell culture and in vivo, and prepared the draft of the manuscript. S.S.W. performed most of the molecular cloning, animal and cell culture experiments, as well as most of the data maintenance and analysis, and helped with writing the draft of the manuscript. F.J.H. constructed N2A‐EGR1KO cells as well as HSV1GFP and HSV1Egr1b viruses and contributed to some experiments in cell culture and in vivo. X.N.J. helped with some molecular cloning, Western blotting, and qRT‐PCR experiments. Y.H.X. helped with ChIP‐qPCR experiments. All authors revised and approved the manuscript.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supporting File 1: advs73512‐sup‐0001‐SuppMat.docx.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1D. M. Knipe , E. F. Heldwein , I. J. Mohr , and C. N. Sodroski , “Herpes Simplex Virus: Mechanisms of Lytic and Latent Infection,” Fields Virology Seventh Edition, eds. P. M Howley , D. M. Knipe , J. I. Cohen , B. A. Damania , (Lippincott Williams & Wilkins, 2021): 235–296.
- 2H. Zhu and C. Zheng , “The Race between Host Antiviral Innate Immunity and the Immune Evasion Strategies of Herpes Simplex Virus 1,” Microbiology and molecular biology reviews: MMBR 84, no. 4 (2020): 10–128.10.1128/MMBR.00099-20PMC 752861932998978 · doi ↗ · pubmed ↗
- 3H. Fu and D. Pan , “Mechanisms of HSV Gene Regulation during Latency and Reactivation,” Virology 602 (2025): 110324.39626607 10.1016/j.virol.2024.110324 · doi ↗ · pubmed ↗
- 4C. Jones , “Human Alpha‐herpesvirus 1 (HSV‐1) Viral Replication and Reactivation from Latency Are Expedited by the Glucocorticoid Receptor,” Journal of Virology 99, no. 4 (2025): e 00303–325.40145740 10.1128/jvi.00303-25PMC 11998515 · doi ↗ · pubmed ↗
- 5S. R. Cuddy and A. R. Cliffe , “The Intersection of Innate Immune Pathways with the Latent Herpes Simplex Virus Genome,” Journal of Virology 97, no. 5 (2023): e 0135222.37129520 10.1128/jvi.01352-22PMC 10231182 · doi ↗ · pubmed ↗
- 6J. L. Umbach , M. F. Kramer , I. Jurak , H. W. Karnowski , D. M. Coen , and B. R. Cullen , “Micro RN As Expressed by herpes Simplex Virus 1 during Latent Infection Regulate Viral m RN As,” Nature 454, no. 7205 (2008): 780–783.18596690 10.1038/nature 07103 PMC 2666538 · doi ↗ · pubmed ↗
- 7I. Jurak , M. F. Kramer , J. C. Mellor , et al., “Numerous Conserved and Divergent micro RN As Expressed by herpes Simplex Viruses 1 and 2,” Journal of Virology 84, no. 9 (2010): 4659–4672.20181707 10.1128/JVI.02725-09PMC 2863732 · doi ↗ · pubmed ↗
- 8D. Pan , J. M. Pesola , G. Li , S. Mc Carron , and D. M. Coen , “Mutations Inactivating herpes Simplex Virus 1 micro RNA mi R‐H 2 Do Not Detectably Increase ICP 0 Gene Expression in Infected Cultured Cells or Mouse Trigeminal Ganglia,” Journal of Virology 91, no. 2 (2017): 10–128.10.1128/JVI.02001-16PMC 521531927847363 · doi ↗ · pubmed ↗