BRLF1 suppresses RNA Pol III‐mediated RIG‐I inflammasome activation in the early EBV lytic lifecycle

Abstract

Latent infection with herpesviruses constitutively activates inflammasomes, while lytic replication suppresses their activation through distinct mechanisms. However, how Epstein–Barr virus (EBV) lytic replication inhibits the activation of inflammasomes remains unknown. Here, we reveal that the EBV immediate‐early protein BRLF1 inhibits inflammasome activation, and BRLF1 deficiency significantly increases the activation of inflammasomes and pyroptosis during early lytic lifecycle. BRLF1 interacts with RNA polymerase III subunits to suppress immunostimulatory small RNA transcription, RIG‐I inflammasome activation, and antiviral responses. Consequently, BRLF1‐deficient EBV primary infection induces robust T‐cell and NK cell activation and killing through IL‐1β and IL‐18. A BRLF1‐derived peptide that inhibits inflammasome activation is sufficient to suppress T‐cell and NK cell responses during BRLF1‐deficient EBV primary infection in lymphocytes. These results reveal a novel mechanism involved in the evasion of inflammasome activation and antiviral responses during EBV early lytic infection and provide a promising approach for the manipulation of inflammasomes against infection of oncogenic herpesviruses.

The Epstein‐Barr virus immediate‐early protein BRLF1 interacts with RNA polymerase III subunits and inhibits RIG‐I‐inflammasome activation. Primary infection with BRLF1‐deficient EBV induces T cell and NK cell activation and virus killing.

Introduction

Epstein–Barr virus (EBV) is a lymphotropic oncogenic herpesvirus and acts as a causative agent of several lymphoproliferative diseases and malignancies in humans (Hislop, 2015; Taylor et al, 2015). It primarily infects B cells and epithelial cells and also infects T and NK cells at low frequencies with a default latent lifecycle. In the presence of certain stimuli, latent EBV infection is reactivated to lytic replication, which leads to the expression of whole panel of viral genes and eventually produces infectious progeny virion particles (McKenzie & El‐Guindy, 2015). During EBV latent infection, the latent gene products activate multiple signaling pathways and induce the expression of a series of inflammatory factors to immortalize infected cells and produce the inflammation and senescence‐associated secretory phenotype (SASP); however, these phenomena are altered during the switch from the latent to the lytic lifecycle by immediate‐early proteins, such as BZLF1, to promote the optional lytic replication and attenuate SASP and the paracrine senescence of uninfected cells (Li et al, 2015; Long et al, 2016). Alternatively, the inflammasomes are constitutively activated through recognition by IFI16 of the viral DNA genome and recognition by RIG‐I of viral small RNA EBERs (Samanta et al, 2006; Ansari et al, 2013). However, whether and how inflammasomes are regulated during the EBV lytic lifecycle remain unknown.

Pattern recognition receptors (PRRs) sense endogenous or pathogenic DNA, RNA, or other products and intracellular damage signaling to induce immune responses and inflammasome activation (Brennan & Bowie, 2010; Hayward et al, 2018). Several PRRs have been characterized in organisms as simple as yeast and as complex as mammals, and inflammasome‐related PRRs include the DNA sensors AIM2 and IFI16, the RNA sensor RIG‐I and NOD‐like receptors (NLRs) that recognize microbial or stress‐related molecules. Inflammasomes are multiple protein complexes that function in caspase‐1 activation and the cleavage of IL‐1β, IL‐18, and other pro‐inflammatory factors. After their formation and activation are triggered by PRRs, inflammasome‐dependent cytokines are processed and released to the extracellular compartment to induce sterile and inflammatory responses; therefore, inflammasomes play important roles in immune responses, inflammation, and the pathogenesis of infectious and inflammatory diseases and many malignant tumors (Patel et al, 2017; Karki & Kanneganti, 2019).

Infection with herpesviruses is recognized by multiple sensors (Brennan & Bowie, 2010); IFI16 and cGAS recognize viral DNAs (Kerur et al, 2011; Ansari et al, 2013; Diner et al, 2016) and RIG‐I recognizes viral small RNAs or specific RNA fragments (Zhao et al, 2018). In addition, RNA polymerase III also acts as a DNA sensor to activate RIG‐I signaling through synthesizing 5ʹ‐pppRNA (Ablasser et al, 2009; Chiu et al, 2009), and it is also responsible for the transcription of viral small RNAs, such as HSV‐1 LAT and EBV EBERs. Moreover, 5ʹ‐pppRNA induces inflammasome activation through RIG‐I, which can bind to ASC and activate caspase‐1‐dependent inflammasomes (Poeck et al, 2010). Consequently, inflammasomes are constitutively activated during the latent infection of herpesviruses through IFI16, AIM2, or RIG‐I‐dependent pathways (Lupfer et al, 2015). These sensors activate both antiviral immune responses and inflammatory responses to inhibit viral lytic replication and establish latent infection.

For successful lytic replication, herpesviruses employ diverse strategies to escape from antiviral responses through the inhibition of interferon‐related responses and inflammatory responses and the attenuation of the activation and maturation of lymphocytes (Coscoy, 2007; Means et al, 2007; Biolatti et al, 2018). Similarly, they suppress the activation of inflammasomes and inflammasome‐related responses; for example, HSV‐1 VP22 and KSHV ORF63 inhibit AIM2 and NLRP1/3 inflammasome activation, respectively, to promote HSV‐1 and KSHV lytic replication (Gregory et al, 2011; Maruzuru et al, 2018).

Similar to infection with other herpesviruses, EBV latent infection induces inflammasome activation through the recognition of viral DNAs by IFI16 and the recognition of viral small RNA EBERs by RIG‐I (Chen et al, 2012; Ansari et al, 2013; Torii et al, 2017). Although an EBV miRNA can inhibit NLRP3 expression and consequent NLRP3 inflammasome activation (Haneklaus et al, 2012), the mechanism underlying the inhibition of inflammasome activation and inflammasome‐dependent responses during the EBV lytic lifecycle remains unknown. Here, we reveal that the EBV immediate‐early protein BRLF1 inhibits RIG‐I inflammasome activation and antiviral responses via RNA polymerase III during the early stages of primary infection and reactivation and then evades antiviral responses of T and NK cells through inflammasome‐dependent factors.

Results

BRLF1 suppresses the activation of inflammasomes during the early stages of herpesvirus infection

To investigate how EBV lytic replication inhibits the activation of inflammasomes, the systemic screening of an EBV lytic ORF‐expressing library was performed using a Gaussia luciferase‐based reporter of inflammasome activation, in which pro‐IL‐1β with a 31 aa N‐terminal truncation was fused with Gaussia luciferase (Fig EV1A). It is well‐known that HSV‐1 infection activates several kinds of inflammasomes (Johnson et al, 2013; Coulon et al, 2019); as such it was used as activator of inflammasomes in our studies. Following inflammasome activation stimulated by HSV‐1 infection (MOI = 1), we found that the lytic genes BHRF1 and BoRF2 significantly activate inflammasomes, whereas the lytic genes BRLF1 and BCRF2 significantly inhibit inflammasomes, and the immediate‐early gene BRLF1 exhibited the most significant inhibition (Fig EV1B). To confirm that BRLF1 inhibits inflammasome activation, BRLF1 was introduced into THP‐1 cells and Ramos cells using lentivirus‐based transduction, after which inflammasomes were activated by HSV‐1 infection. The levels of cleaved caspase‐1 (p20), mature IL‐1β, and mature IL‐18 were increased by HSV infection in both cell lysates and supernatants, and all three were inhibited by BRLF1 expression in both cell lines (Figs 1A and EV1C). The secretion of IL‐1β and IL‐18 was analyzed by enzyme‐linked immunosorbent assay (ELISA), both were similarly decreased by BRLF1 expression (Fig 1B and C).

Figure EV1. EBV ORF‐expressing library screening reveals BRLF1 as an inhibitor of inflammasomes.

1. The diagram showing the Gaussia luciferase‐based reporter of inflammasome activation. Pro‐IL‐1β with a 31 aa N‐terminal truncation was fused with Gaussia luciferase and expressed with the CMV promoter and was named pro‐IL‐1β‐DN reporter.
2. The pro‐IL‐1β‐DN reporter was transfected into A549 cells with empty vector or EBV lytic ORF‐expressing plasmids. Cells were infected with HSV‐1 (MOI = 1) for 12 h, and then, the supernatants were collected and measured using a reagent for Renilla luciferase activity. The values are shown as the mean ± standard deviation of triplicate analyses from three independent experiments. *P < 0.01. Tukey's multiple comparison test.
3. Ramos cells were transduced with BRLF1‐expressing or empty lentivirus for 24 h and then were left uninfected or infected with HSV‐1 (MOI = 1). Twelve hours after HSV‐1 infection, the levels of pro‐caspase‐1 (p45), cleaved caspase‐1 (p20), pro‐IL‐1β, mature IL‐1β, pro‐IL‐18, and mature IL‐18 in supernatants and whole‐cell lysates were detected by Western blotting as indicated. The release of mature IL‐1β and IL‐18 from Ramos cells was measured by ELISA. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, **P < 0.01, Tukey's multiple comparison test.

Source data are available online for this figure.


Figure 1. BRLF1 inhibited inflammasome activation induced by herpesviruses.

• A
  THP‐1 cells were transiently infected with BRLF1‐expressing or empty control lentivirus. Twenty‐four hours later, the control or BRLF1‐expressing THP‐1 cells were primed with 40 ng/ml TPA overnight and then were either uninfected or infected with HSV‐1 (MOI = 1) in serum‐free medium for 12 h. The supernatants and whole‐cell lysates were collected and analyzed by Western blotting to detect pro‐caspase‐1 (p45), cleaved caspase‐1 (p20), pro‐IL‐1β, mature IL‐1β, pro‐IL‐18, and mature IL‐18 as indicated.
• B, C
  The release of mature IL‐1β (B) and IL‐18 (C) from THP‐1 cells was measured by enzyme‐linked immunosorbent assay (ELISA) following HSV‐1 infection. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, **P < 0.01, Tukey's multiple comparison test.
• D–F
  Ramos cells were infected with wild‐type EBV‐WT‐ or BRLF1‐deficient EBV‐ΔBRLF1 virus at a high titer (MOI = 100). After being transduced with empty control or BZLF1‐expressing lentivirus, the supernatants and cell pellets of EBV‐WT‐ or EBV‐ΔBRLF1‐harboring Ramos cells were collected after 12 h in serum‐free culture, and then, the levels of EBNA1, BRLF1, BZLF1, pro‐caspase‐1 (p45), cleaved caspase‐1 (p20), pro‐IL‐1β, mature IL‐1β, pro‐IL‐18, mature IL‐18, GSDMD, and cleaved N‐GSDMD in the supernatants and whole‐cell lysates were analyzed as indicated (D). The release of mature IL‐1β (E) and IL‐18 (F) from Ramos cells was measured by enzyme‐linked immunosorbent assay (ELISA). Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, **P < 0.01, Tukey's multiple comparison test.
• G
  After being transduced with control or BZLF1‐expressing lentivirus for 48 h, EBV‐WT‐ and EBV‐ΔBRLF1‐infected Ramos cells were stained with PI and measured using fluorescent flow cytometry. Representative images of pyroptotic cells are shown, and the percentages of pyroptotic cells were calculated in duplicate.

Source data are available online for this figure.

To further investigate whether BRLF1 inhibits inflammasomes during the EBV lytic lifecycle, primary EBV infection was established in Ramos cells with wild‐type EBV‐WT‐ or BRLF1‐deficient EBV‐ΔBRLF1 virus. After the cells were transduced with control or BZLF1‐expressing lentivirus, the whole‐cell lysates and supernatants were analyzed for inflammasome activation. In the BZLF1‐transduced cells, the levels of cleaved caspase‐1 (p20), mature IL‐1β, and mature IL‐18 were increased in both cell lysates and supernatants from EBV‐ΔBRLF1‐infected cells compared with those from EBV‐WT cells (Fig 1D), and the secretion of both IL‐1β and IL‐18 was similarly increased in the EBV‐ΔBRLF1‐infected cells compared with the EBV‐WT‐infected cells (Fig 1E and F). The cleavage of the pyroptotic modulator GSDMD by caspase‐1 was elevated in EBV‐ΔBRLF1‐infected cells as well (Fig 1D). Studies have shown that the cleaved GSDMD isoform forms permeable pores in plasma membrane in cells undergoing pyroptosis, and staining with propidium iodide (PI) in living cells can detect cell pyroptosis (Wree et al, 2014; Gu et al, 2019). EBV‐ΔBRLF1‐infected cells exhibited an increased percentage of pyroptosis compared with wild‐type cells during the lytic stage via PI staining and flow cytometry analysis (Fig 1G). All these results confirm that BRLF1 inhibits the activation of inflammasomes during both HSV‐1 and EBV lytic infection.

To map the key BRLF1 region that is responsible for the inhibition of inflammasome activation, a series of BRLF1 mutants were constructed. Several key functional domains present in the BRLF1 protein, including the dimerization domain (1–232 aa), the DNA‐binding domain (1–280 aa), the nuclear localization sequence (NLS) (410–413 aa), and the transcriptional activation domain (TA domain) (415–605 aa), were subject to mutation (Appendix Fig S1A). The results of the mutagenesis indicated that neither the dimerization domain nor the DNA‐binding domain of BRLF1 exhibited an inhibitory effect. However, the NLS and the TA domain region in BRLF1 were required for inhibition. The requirement of NLS indicates that BRLF1 nuclear localization is important for its inhibition, which plays an important role in BRLF1 function by either direct DNA binding or other indirect mechanisms mediated by cellular components or transcriptional factors (Hsu et al, 2005). We then constructed a series of mutants with deletions and mutations in the TA domain to map the specific functional sites (Appendix Fig S1A–C). Deletion of the NLS or the 577–578 region or the L578A mutation abolished the inhibitory effects (Appendix Fig S1C and Fig EV2C). Interestingly, fusion of an 11 aa‐fragment from the BRLF1 572–582 region with the NLS was sufficient to inhibit inflammasome activation induced by HSV‐1 infection (Appendix Fig S1D and Fig EV2C). Therefore, a TAT‐driven cell‐permeable peptide was chemically synthesized, which was named TAT‐N572 (Fig EV2B). This peptide was able to inhibit the activation of caspase‐1 and the cleavage of pro‐IL‐1β in both HSV‐1‐infected THP‐1 cells and EBV‐positive p3HR‐1 lymphoma cells in a dose‐dependent manner (Fig EV2D and E). These results suggested that a short region in BRLF1 and its functional peptide were sufficient for inhibiting inflammasome activation induced by herpesviruses.

Figure EV2. Mapping of the domain of BRLF1 involved in the inhibition of inflammasomes.

• A
  The diagram of BRLF1 functional domains and sites.
• B
  The sequences of the TAT‐fused Flag and BRLF1 NLS‐fused 572–582 peptides, named TAT‐Flag and TAT‐N572, respectively.
• C
  Flag‐tagged BRLF1 wild type, ΔNLS, Δ577‐579, L578A, and NLS‐fused 572–582 constructs were transfected into THP‐1 cells, which were then infected with HSV‐1 (MOI = 1). Twelve hours post‐infection, the levels of cleaved caspase‐1 (p20) and mature IL‐1β in the supernatants were detected.
• D, E
  TAT‐fused cell‐permeable Flag peptide or BRLF1 N572 peptide were added to p3HR‐1 EBV‐positive cells for 12 h (D) or added to THP‐1 cells that were immediately infected with HSV‐1 (MOI = 1) for 12 h (E). The levels of pro‐caspase‐1, cleaved caspase‐1, pro‐IL‐1β, and mature IL‐1β in the whole‐cell lysates were detected.

Source data are available online for this figure.


BRLF1 interacts with RNA polymerase III subunits to downregulate small RNA transcription and 5ʹ‐pppRNA production

To reveal the mechanism underlying the inhibition of inflammasome activation by BRLF1, GST‐tagged BRLF1 or a control vector was transfected into A549 cells, which was followed by HSV‐1 stimulation, and then, the BRLF1‐binding proteins were isolated by affinity purification with GST‐agarose beads (Fig 2A) and sequenced by mass spectrometry. Several proteins, including hsp90, S100A7, RAC1, POLR3F, and POLR3G, were shown to bind to BRLF1; however, only POLR3F and POLR3G knockdown attenuated the inhibition of inflammasome activation in the presence of BRLF1 expression (Fig EV3A). When POLR3F and POLR3G expression were depleted by shRNAs in EBV‐ΔBRLF1‐infected HNE1 cells and HSV‐1‐infected A549 cells, BRLF1 no longer exhibited the inhibition of the cleavage of caspase‐1 or IL‐1β (Fig 2B), indicating that RNA polymerase III subunits play an essential role in the BRLF1‐mediated inhibition of inflammasome activation. When GFP‐tagged BRLF1 or BRLF1 L578A was coexpressed with GST‐tagged POLR3F or POLR3G in HEK293T cells, immunoprecipitation assays showed that BRLF1 interacted with both POLR3F and POLR3G, while BRLF1 L578A did not interact with either, which is consistent with the loss of inflammasome inhibition (Fig 2C).Further, GFP‐BRLF1 or GFP‐BRLF1 L578A was transfected in A549 cells followed by HSV‐1 infection or in HNE‐1‐EBV‐ΔBRLF1‐infected cells, after which cells were collected and lysed to detect the interaction of BRLF1 with endogenous sensors of inflammasomes. The immunoprecipitation assays showed that BRLF1 interacted with endogenous POLR3F but not with AIM2, IFI16, RIG‐I, or NLRP3, while BRLF1 L578A did not bind to any (Fig 2D). Moreover, immunoprecipitation assays with cell extracts undergoing EBV lytic replication showed that virus‐expressed BRLF1 interacts with endogenous POLR3F (Fig 2E). Predicting ligand–receptor interactions in depth provides a basis and direct understanding of the binding structure. We predicted the 3D structure of BRLF1, the BRLF1 TA domain, and the peptide NLS572‐582 in I‐TASSER (Yang et al, 2015); the results showed that four helixes of the BRLF1 TA domain were appropriately arranged in parallel within the predicted structure and that the peptide fragment in BRLF1 protruded from the BRLF1 molecule as a helix (Fig 2F, Datasets [Link], [Link], [Link]). We further analyzed the docking modes of this peptide with RNA Polymerase III open complex (PDB ID 6F40) (Vorländer et al, 2018) based on the CDOCKER analysis. Interestingly, the NLS572‐582 peptide or the peptide‐containing helix was found to bind with POLR3 in the DNA‐binding pocket, likely to interfere with DNA binding and influence the transcription function of POLR3 (Fig 2F, Datasets EV4 and EV5).

Figure 2. BRLF1 interacted with RNA polymerase III subunits.

1. GST‐tagged BRLF1‐expressing plasmid or control vector was transfected into A549 cells. After HSV‐1 infection (MOI = 1) for 12 h, the cells were collected, and the whole‐cell extracts were subjected to immunoprecipitation with GST‐affinity beads. The immunoprecipitated complexes were then separated by SDS‐PAGE, and the image obtained from silver staining is shown.
2. Scramble (sc) or shRNAs against POLR3F and POLR3G with empty control vector (ctr) or BRLF1‐expressing plasmid were transfected into EBV‐ΔBRLF1‐harboring HNE1 cells and A549 cells, followed by HSV‐1 infection. Thirty‐six hours after the transfection of the HNE1 cells and 12 h after HSV‐1 infection (MOI = 1) of the A549 cells, the cell pellets were collected and then the cleavage of caspase‐1 and IL‐1β were measured as indicated. The release of mature IL‐1β and IL‐18 from HNE‐1 and A549 cells was measured by enzyme‐linked immunosorbent assay (ELISA) after ultrafiltration and concentration with Amicon^® Ultra‐0.5 centrifugal filter devices. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, Tukey's multiple comparison test.
3. GFP‐tagged BRLF1 wild type, BRLF1 L578A, or empty vector was transfected into 293T cells with GST‐tagged POLR3F or POLR3G expressing plasmids for 48 h. Cells were collected, and immunoprecipitation with GST‐affinity beads was performed, and then, whole‐cell lysates and immunoprecipitated complexes were analyzed as indicated.
4. GFP‐tagged empty vector or GFP‐BRLF1‐ or GFP‐BRLF1 L578A‐expressing plasmids were transfected into A549 cells, after which the cells were then infected with HSV‐1, or transfected into HNE‐1‐EBV‐ΔBRLF1 cells. The cells were then collected and lysed. Immunoprecipitation with GST‐affinity beads was subsequently performed, and then, whole‐cell lysates and immunoprecipitated complexes were analyzed as indicated.
5. P3HR‐1 cells were induced by TPA plus NaB into lytic infection, and the cells were then collected. Immunoprecipitation with anti‐mouse IgG or BRLF1 antibody was performed, after which whole‐cell lysates and immunoprecipitated complexes were analyzed as indicated.
6. 3D structure of BRLF1, the BRLF1 TA domain, and the BRLF1 NLS572‐582 peptide were predicted using I‐TASSER. Receptor‐ligand interaction mode of the BRLF1 TA domain and the BRLF1 NLS572‐582 peptide with RNA Polymerase III open complex (PDB ID 6F40) was conducted using CDOCKER. The yellow sequence in top figure indicates the TA domain in BRLF1, and the red sequence represents the NLS572‐582 peptide. The yellow sequence in the middle figure represents the BRLF1 TA domain, and the red helix in middle and bottom figures represents the DNA double helix.

Source data are available online for this figure.

Figure EV3. Supplementary data for Fig 3 .

1. The pro‐IL‐1β‐DN reporter was transfected into A549 cells along with BRLF1 or empty vector in the presence of scramble or shRNAs against Hsp90, S100A7, RAC1, POLR3F, POLR3G, or POLR3F plus POLR3G. Twelve hours after HSV‐1 infection (MOI = 1), the cell media were collected, and the Gaussia luciferase activity in the supernatants was measured. Left results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, Tukey's multiple comparison test. The depletion of HSP90, S100A7, RAC1, and POLR3G was detected via real‐time PCR, and the relative mRNA levels are shown. Right results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, Student's t‐test. The depletion of POLR3F was detected by Western blotting analysis, and the results of which are shown.
2. Supplementary data for Fig 3B. Mean mRNA expression normalized to the housekeeping gene GAPDH. Results are presented as the mean ± SD, n = 3 biological replicates.
3. Supplementary data for Fig 3C. Mean mRNA expression normalized to the housekeeping gene GAPDH. Results are presented as the mean ± SD, n = 3 biological replicates.
4. EBV‐ΔBRLF1‐harboring HNE1 cells were transfected with control vector or BRLF1 in the absence or presence of shRNAs against the RNA polymerase subunits POLR3G plus POLR3F for 48 h. The total RNAs were extracted and subjected to real‐time PCR analysis. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05. **P < 0.01, Tukey's multiple comparison test. The depletion of POLR3F was detected by Western blotting analysis, and the results of which are shown.
5. A549 cells were transfected with control vector or BRLF1 along with scramble or shRNAs against POLR3G plus POLR3F for 36 h. After the cells were infected with HSV‐1 (MOI = 1) for 12 h, the total RNAs were extracted and subjected to real‐time PCR analysis. The depletion of POLR3F by shRNA is shown as indicated. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05. **P < 0.01, Tukey's multiple comparison test.

Source data are available online for this figure.


RNA polymerase III is responsible for small RNA transcription; therefore, a RNA‐seq analysis was performed in Ramos cells infected with EBV‐WT or EBV‐ΔBRLF1 virus at a high titer (MOI = 100). After classification of all kinds of RNAs, POLR3‐dependent RNA transcripts were selected and quantitated. We found that BRLF1 expression during EBV‐WT primary infection down‐regulated many POLR3‐dependent small RNA transcripts, while the BRLF1‐deficient EBV primary infection barely exhibited the differences, compared with that in uninfected cells. When the level of RNA expression in EBV‐WT and EBV‐ΔBRLF1 primary infection was compared, many POLR3‐dependent small RNAs were specifically down‐regulated for more than 5‐ to 10‐fold or greater by BRLF1 expression, including RN7SL1, 7SK, SNORD14E, SNORA53, SNORA3B, SNORA47, and U6 (Fig 3A). We also detected the level of cellular and viral small RNA transcription in EBV‐ and HSV‐1‐infected cells. In EBV‐ΔBRLF1‐infected HNE1 cells, BRLF1 expression inhibited both EBER1 and EBER2 expression regardless of the presence of latent or lytic infection, while the BRLF1 L578A construct failed to inhibit their expression (Fig 3B); in addition, the expression of the EBV miRNA BART15 and latent gene EBNA1 was barely affected. In HSV‐1‐infected cells, the expression of the viral small RNA LAT‐1 was also reduced by BRLF1 but not by BRLF1 L578A, while the expression of the viral lytic gene UL27 was not affected (Fig 3C). Interestingly, the expression of two cellular small RNAs, RN7SL1 and 7SK, which are transcribed by RNA polymerase III and play critical roles in inflammatory and innate immune responses (Nabet et al, 2017), was similarly decreased (Fig 3B and C); however, the expression of cellular inflammasome‐related genes, such as IL‐1β, caspase‐1 or ASC, that are transcribed by RNA polymerase II, was not changed (Fig EV3B and C). As expected, the BRLF1 inhibition of small RNA transcription was RNA polymerase III dependent, as it did not inhibit the transcription of these small RNAs when POLR3F and POLR3G were depleted by shRNAs, and the depletion of Pol‐III by these shRNAs resulted in a decreased level of transcription of ISG genes and cellular and viral small RNA rather than rescued BRLF1‐mediated inhibition (Fig EV3D and E). These results suggest that BRLF1 inhibits the transcription of immunostimulatory small RNAs through interacting with RNA polymerase III.

Figure 3. BRLF1 inhibited the transcription of viral and cellular immunostimulatory RNAs.

1. After infected with wild‐type EBV‐WT‐ or BRLF1‐deficient EBV‐ΔBRLF1 virus at a high titer (MOI = 100), POLR3‐dependent RNA transcripts in Ramos cells were detected by RNA‐seq analysis. The relative levels of cellular small RNA transcripts are shown as RNA fold changes in the heatmap in the mock (mock group vs. mock group), EBV‐WT (EBV‐WT group vs. mock group), and EBV‐ΔBRLF1 (EBV‐ΔBRLF1 group vs. mock group) comparisons.
2. EBV‐ΔBRLF1‐harboring HNE1 cells were transfected with control vector, BRLF1, or BRLF1 L578A in the absence or presence of BZLF1 for 48 h. The total RNAs were extracted and subjected to real‐time PCR analysis. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, **P < 0.01, Tukey's multiple comparison test.
3. A549 cells were transfected with control vector, BRLF1, or BRLF1 L578A for 36 h. After the cells were infected with HSV‐1(MOI = 1) for 12 h, the total RNAs were extracted and subjected to real‐time PCR analysis. The values are shown as the mean ± standard deviations of triplicate analyses from three independent experiments. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05. **P < 0.01, Student's t‐test.
4. EBER1 5ʹ‐pppRNA detection using splint‐ligation. EBV‐WT‐ or EBV‐ΔBRLF1‐harboring HNE1 cells were cotransfected with empty vector, BRLF1 or BRLF1 L578A, BRLF1 plus BZLF1, or BRLF1 L578A plus BZLF1 for 48 h. The total RNAs were extracted, and 5 μg of each RNA was subjected to splint‐ligation with FAM‐labeled probe to quantify the 5ʹ‐monophosphorylated EBER1 and total EBER1 RNA. The ligation products were separated using urea‐PAGE, visualized by a fluorescent scanner, and analyzed using ImageJ software. The ratio of 5ʹ‐pRNA density to total RNA density and the relative 5ʹ‐pppRNA level were calculated for three independent experiments. Representative images and the means of relative 5ʹ‐pppRNA levels are shown. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, Tukey's multiple comparison test.
5. EBV‐WT‐harboring HNE1 cells were transfected with control vector, BRLF1, or BRLF1 L578A‐expressing plasmid in the presence of GST‐tagged RNA polymerase subunit POLR3F for 48 h. The cells were harvested, and ChIP assay was performed with anti‐GST‐affinity beads. The POLR3F‐binding promoter fragments of BSRF1, BMLF1, EBER1, EBER2, EBNA1, 7SK, RN7SL1, and the GAPDH as negative control in the ChIP samples were detected by real‐time PCR assays. The relative levels of DNA binding were normalized to the input, and the means of three independent experiments are shown. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, Tukey's multiple comparison test.

RNA polymerase III complexes sense exogenous and viral dsDNA and produce 5ʹ‐triphosphate‐RNA (5ʹ‐pppRNA) and activate RIG‐I‐dependent signaling in response (Ablasser et al, 2009; Chiu et al, 2009). To further determine whether BRLF1 inhibits the RNA polymerase III‐dependent activation of inflammasomes, we measured the levels of EBER1 5ʹ‐pppRNA in EBV‐infected cells. Using a FAM‐labeled adapter DNA probe and an EBER1‐targeted bridge DNA probe, EBER1 5ʹ‐pRNA was directly ligated to the FAM adapter DNA probe while the 5ʹ‐pppRNA was not, which allowed the relative 5ʹ‐pppRNA level to be quantitated after the total EBER1 and 5ʹ‐pRNA EBER1 were detected (Appendix Fig S2). In both the wild‐type and BRLF1‐deficient EBV‐infected cells, 5ʹ‐pRNA EBER1 was consistently detected in both the latent and lytic stages; however, total EBER1 RNA was reduced by BRLF1 overexpression in both wild‐type and BRLF1‐deficient latent cells. The level of total EBER1 RNA was also decreased during BZLF1‐induced lytic replication by BRLF1 expression, whereas it was barely affected during lytic replication in the presence of BRLF1 deficiency; ectopic BRLF1 overexpression augmented inhibition during wild‐type lytic replication and rescued inhibition during lytic replication in EBV‐ΔBRLF1‐infected HNE1 cells, while BRLF1 L578A did not exhibit the inhibition in either the EBV‐WT‐ or EBV‐ΔBRLF1‐infected cells (Fig 3D). These results suggest that BRLF1 suppresses RNA polymerase III‐dependent 5ʹ‐pppRNA production during EBV lytic infection.

To further determine that BRLF1 down‐regulated 5ʹ‐pppRNA production through inhibiting POLR3 association with DNA promoters, ChIP assays were performed with GST‐tagged RNA polymerase subunit POLR3F in EBV‐WT‐harboring HNE1 cells with control vector, BRLF1, or BRLF1 L578A overexpression. The binding of viral promoters BSRF1, BMLF1, EBNA1, EBER1, and EBER2, and cellular promoters 7SK and RN7SL1 were detected by real‐time PCR assays, with Pol II‐dependent GAPDH as negative control. The strong interaction of GST‐POLR3F with EBER1, EBER2, 7SK, and RN7SL1 promoters was observed in the absence of BRLF1 expression, with the minimal interaction between POLR3F and BSRF1, BMLF1, EBNA1, or GAPDH promoters because they are transcribed by RNA polymerase II. Furthermore, BRLF1L578A completely lost the inhibition on all promoter binding with POLR3F (Fig 3E). These results suggest that BRLF1 suppresses 5ʹ‐pppRNA production by suppressing the binding of RNA polymerase III with DNA promoters.

BRLF1 inhibits RIG‐I‐dependent inflammasome activation and antiviral responses

To further confirm that the RNA polymerase III‐mediated RIG‐I inflammasome is inhibited by BRLF1 rather than other kinds of inflammasomes, 20 sensors of inflammasome activation were depleted by shRNAs and then the inhibitory effect of BRLF1 was detected. We found that RIG‐I knockdown significantly attenuated the inhibition of inflammasomes in the presence of BRLF1 expression (Fig EV4A). Unlike IFI16‐ and NLRP3‐dependent inflammasome activation, the activation of RIG‐I‐dependent inflammasome induced by HSV‐I infection was inhibited by BRLF1 (Fig 4A). When the expression of RIG‐I was depleted in EBV‐ΔBRLF1‐infected HNE1 cells, BRLF1 no longer inhibited the cleavage of caspase‐1 or IL‐1β (Fig 4B), confirming that BRLF1 inhibited RIG‐I‐mediated inflammasome activation. To confirm that BRLF1 inhibits RIG‐I inflammasome activation, the oligomerization of ASC was analyzed after DSS chemical cross‐linking of intracellular ASC was performed. In both HSV‐1‐infected 293T and EBV‐ΔBRLF1‐infected HNE1 cells, ASC oligomerization was inhibited by BRLF1 in the presence of RIG‐I overexpression, while it exhibited no inhibition in the presence of AIM2 overexpression (Fig 4C), suggesting that BRLF1 inhibits RIG‐I‐mediated formation and activation of inflammasome complexes. To further confirm that BRLF1 inhibits RIG‐I activation, the downstream immune signaling axis of the RIG‐I pathway was examined. The expression of the interferon‐related genes IFNβ, ISG15, IFIT1, and Mx1 was greatly inhibited by BRLF1 but not by BRLF1 L578A during both EBV and HSV‐1 infections (Fig 4D and E). The expression of these genes was decreased by the disruption of the RNA polymerase III subunits, after which BRLF1 no longer induced inhibition (Fig EV3C and D). These results suggest that BRLF1 inhibits RIG‐I‐dependent inflammasome activation and immune responses in a RNA polymerase III‐dependent manner.

Figure EV4. The requirement of other sensors and signaling proteins in BRLF1 inhibition in inflammasome and immune responses.

1. Supplementary data for Fig 4. The pro‐IL‐1β‐DN reporter was transfected into A549 cells with BRLF1 or empty vector in presence of scramble or shRNAs against 20 sensors of inflammasomes as indicated. Twelve hours after HSV‐1 infection (MOI = 1), the activity of Gaussia luciferase in the supernatants was measured. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, Tukey's multiple comparison test. The levels of AIM2, IFI16, RIG‐I, and NLRP3 by shRNA depletion as representative sensors were determined by Western blotting analysis, and the results of which are shown.
2. Supplementary data for Fig 1A. Panel A depicts the mean mRNA expression normalized to the housekeeping gene GAPDH. Results are presented as the mean ± SD, n = 3 biological replicates, **P < 0.01, Student's t‐test.
3. THP‐1 cells were transduced with BRLF1‐expressing or empty lentiviruses and with shTLR3‐, shMAVS‐, shSTING‐, and shcGAS‐expressing or empty lentiviruses for 24 h, after which the cells were infected with HSV‐1 (MOI = 1). Twelve hours after HSV‐1 infection, the levels of p‐TBK1, TBK1, pro‐IL‐1β, and mature IL‐1β in whole‐cell lysates were determined by Western blotting analysis as indicated. The total RNA was extracted and subjected to real‐time PCR analysis for the determination of IFN expression and the levels of shRNA depletion. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, Student's t‐test.

Source data are available online for this figure.


Figure 4. BRLF1 inhibits the activation of RIG‐I inflammasome and immune responses.

1. Stable 293T cells in the presence of pro‐caspase‐1, pro‐IL‐1β, and ASC overexpression were transfected with control vector or BRLF1 with control vector, IFI16, NLRP3, RIG‐I, or AIM2 for 36 h. After the cells were infected with HSV‐1 (MOI = 1) for 12 h, the cells were collected, and the cell lysates were analyzed by Western blotting to determine the levels of pro‐caspase‐1, cleaved caspase‐1, pro‐IL‐1β, and mature IL‐1β as indicated.
2. Scramble (sc) or shRNAs against RIG‐I and empty control vector (ctr), BRLF1‐ or BRLF1 L578A‐expressing plasmids were cotransfected into EBV‐ΔBRLF1‐harboring HNE1 cells. Thirty‐six hours after the transfection, cell pellets were collected, after which the cleavage of caspase‐1 and IL‐1β was measured as indicated.
3. Analysis of ASC oligomerization. ASC‐expressing plasmid was transfected into 293T cells or HNE1 cells with empty vector or GFP‐BRLF1 in the presence of AIM2‐ or RIG‐I‐expressing plasmid. Twelve hours after HSV‐1 infection (MOI = 1) in 293T cells or 48 h after EBV‐ΔBRLF1 infection (MOI = 1,000) in HNE1 cells, the cell pellets were collected, and cell lysates were treated with DSS to induce chemical cross‐linking and analyzed by Western blotting for ASC oligomerization.
4. EBV‐ΔBRLF1‐harboring HNE1 cells were transfected with control vector, BRLF1 or BRLF1 L578A in the absence or presence of BZLF1 for 48 h. The total RNAs were extracted and subjected to real‐time PCR analysis.
5. A549 cells were transfected with control vector, BRLF1, or BRLF1 L578A for 36 h. After the cells were infected with HSV‐1 (MOI = 1) for 12 h, the total RNAs were extracted and subjected to real‐time PCR analysis.

Data information: The values are shown as the mean ± standard deviations of triplicate determinations from three independent experiments. *P < 0.05. **P < 0.01. Tukey's multiple comparison test.

Source data are available online for this figure.

To determine whether BRLF1 could affect the signaling from other sensors and pathways such as the cGAS‐STING, RIG‐I‐MAVS, and TLR‐TRIF pathways, TLR3, MAVS, cGAS, and STING were depleted by shRNA to determine whether BRLF1 could affect signaling from these pathways. We found that impairment of these sensors or signaling proteins decreased the levels of TBK1 phosphorylation and IFN‐β expression, that BRLF1 still exhibited the inhibition under their depletion (with the exception of MAVS knockdown), and that knockdown of these proteins did not affect the inhibitory function of BRLF1 in inflammasome activation (Fig EV4B and C). These results indicate that BRLF1 suppresses inflammasome activation and immune responses not through the TLR3‐TRIF and cGAS‐STING pathways but mainly through RIG‐I‐dependent signaling cascades.

BRLF1 inhibits T‐ and NK cell activation during EBV primary infection through inflammasome‐dependent factors

Although BRLF1‐deficient EBV lytic infection induces pyroptosis in B cells (Fig 1D), we did not observe that the activation of intracellular NLRP3 inflammasomes upon adding LPS or the depletion of inflammasomes by caspase‐1 knockdown affected EBV lytic gene expression (Appendix Fig S3A). To investigate whether extracellular IL‐1β influences EBV infection or replication, the different amounts of IL‐1β cytokine were added to p3HR‐1 cells; neither significant cell death was observed, nor was viral gene expression affected when TPA plus NaB was added to induce lytic replication (Appendix Fig S3B). These results suggest that inflammasome activation and inflammasome‐dependent factors do not regulate EBV lytic replication or reactivation.

To investigate the role of the BRLF1‐mediated inhibition of inflammasome in EBV primary and lytic infection, the expression of immune and inflammatory cytokines was analyzed in Jurkat and NK‐92MI cells with EBV‐WT and EBV‐ΔBRLF1 primary infection (Fig 5A and B). Compared to EBV‐WT primary infection, EBV‐ΔBRLF1 primary infection induced the expression of IL‐1R1 and IL‐18R1 in both types of cells, suggesting that BRLF1 deficiency enhances their sensitivity to inflammasome‐dependent cytokines and EBV lytic replication blocks T‐ and NK cell responses to inflammasome activation through BRLF1 expression. Furthermore, the expression of the cell killing‐specific genes lymphotoxin, granzyme‐B, and perforin was induced in both types of cells infected with EBV‐ΔBRLF1 compared with these cells infected with EBV‐WT, as were the T‐cell‐activating IL‐2 receptor isoforms and NK cell‐activating receptors NKp30 and NKp46, indicating that BRLF1 suppresses T‐ and NK cell activation and cell killing during EBV primary infection.

Figure 5. BRLF1‐mediated inhibition of inflammasomes attenuates T‐ and NK cell activation during EBV primary infection.

• A, B
  Jurkat cells (A) and NK‐92MI cells (B) were infected with EBV‐WT or EBV‐ΔBRLF1 virus (MOI = 500) for 36 h. Then, the cells were collected, and the total RNAs were extracted, reverse‐transcribed, and analyzed by real‐time PCR. Color gradient depicts the mean mRNA expression normalized to the housekeeping gene GAPDH and compared to the mock group. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05. **P < 0.01, Student's t‐test.
• C–E
  Jurkat and NK‐92MI cells were cultured with conditional medium from EBV‐WT‐ or EBV‐ΔBRLF1‐infected Ramos cells (paracrine), or Jurkat and NK‐92MI cells were directly infected with EBV‐WT or EBV‐ΔBRLF1 virus (autocrine). (C) The cell pellets were collected, and the whole‐cell extracts were analyzed by Western blotting to detect TAK1, IKKα/β, and IκB phosphorylation. The cell killing activity of Jurkat cells (D) and NK‐92MI cells (E) was measured by the detection of the LDH release from the K562 target cells. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, **P < 0.01, Tukey's multiple comparison test.
• F–H
  Anti‐IL‐1β and/or anti‐IL‐18 neutralizing antibodies (NA), TAT‐Flag or TAT‐N572 peptide were added into EBV‐WT‐ or EBV‐ΔBRLF1‐infected Jurkat cells or NK‐92MI cells for 24 h. (F) TAK1, IKKα, and IκB phosphorylation was detected as indicated. In addition, the cell killing activity of Jurkat cells (G) and NK‐92MI cells (H) was measured. Results are presented as the mean ± SD, n = 3 biological replicates, *P < 0.05, **P < 0.01, Tukey's multiple comparison test.

Source data are available online for this figure.

To further confirm that BRLF1 expression inhibited T and NK activation and cell killing through inflammasome‐dependent factors, Jurkat cells and NK‐92MI cells were incubated with paracrine cytokines from Ramos cells with EBV‐WT or EBV‐ΔBRLF1 primary infection or directly infected with EBV‐WT or EBV‐ΔBRLF1 to produce autocrine cytokines. Both the EBV‐ΔBRLF1‐related paracrine and the autocrine cytokines induced the phosphorylation of TAK1, IKKα/β, and their substrate IκB; in comparison, these cytokines from EBV‐WT infection barely changed their phosphorylation levels (Fig 5C). Accordingly, LDH release from K562 target cells was greatly increased when K562 cells were incubated with Jurkat or NK‐92MI cells pretreated with EBV‐ΔBRLF1‐related paracrine or autocrine cytokines compared with those cytokines derived from EBV‐WT infection (Fig 5D and E). These results suggest that BRLF1 suppresses T‐ and NK cell activation and killing through the secretion of cytokines from EBV primary‐infected cells. By adding different amounts of IL‐1β or IL‐18 to Jurkat and NK‐92MI cells, we found that TAK1 and IKKα/β phosphorylation were induced by IL‐1β or IL‐18 in a dose‐dependent manner (Fig EV5A and B). To further define the roles of IL‐1β and/or IL‐18 in BRLF1‐mediated inflammasome inhibition, anti‐IL‐1β and anti‐IL‐18 neutralizing antibodies were used to deplete the activity of these cytokines. Anti‐IL‐1β neutralizing antibody inhibited IκB phosphorylation and anti‐IL‐18 neutralizing antibody resulted in slight inhibition in EBV‐ΔBRLF1‐infected Jurkat cells, while both anti‐IL‐1β and anti‐IL‐18 neutralizing antibodies resulted in the significant inhibition of IκB phosphorylation in NK‐92MI cells under the same conditions (Fig 5F). Consequently, treatment with an anti‐IL‐1β neutralizing antibody suppressed cell killing by EBV‐ΔBRLF1‐infected Jurkat cells (Fig 5G), and both anti‐IL‐1β and anti‐IL‐18 neutralizing antibody did so in NK‐92MI cells under same conditions (Fig 5H). These results suggest that IL‐1β and IL‐18 are responsible for Jurkat and NK‐92MI cell activation during EBV‐ΔBRLF1 primary infection. Notably, the inhibitory peptide TAT‐N572 was able to reduce cell killing when it was added to EBV‐ΔBRLF1‐infected Jurkat and NK‐92MI cells (Fig 5G and H). Thus, we can conclude that BRLF1 inhibits T‐ and NK cell activation through downregulating inflammasome‐dependent IL‐1β and IL‐18 secretion.

Figure EV5. Supplementary data for Figs 5 and 6.

• A, B
  IL‐1β and IL‐18 induce Jurkat and NK activation. Different amounts of IL‐1β and IL‐18 were added to Jurkat (A) and NK (B) cells for 24 h, and then, the cells were harvested and the cell lysates were analyzed for TAK1 and IKKα/β phosphorylation.
• C, D
  Supplementary data for Fig 6. PBMCs were infected with EBV‐WT or EBV‐ΔBRLF1 virus (MOI = 500). Thirty‐six hours after infection, the percentages of GFP‐positive cells were analyzed using FACS to measure the efficiency of EBV‐WT or EBV‐ΔBRLF1 viral infection in PBMCs (C). Thirty‐six hours after infection, the cells were harvested and the cell lysates were analyzed for TAK1 and IKKα/β phosphorylation (D).

Source data are available online for this figure.


BRLF1 is required for the evasion of T‐ and NK cell activation during EBV primary infection in human lymphocytes

To further confirm the inhibitory function of BRLF1 on inflammasome activation during EBV infection, PBMCs were isolated from healthy donors and primarily infected with EBV‐WT or EBV‐ΔBRLF1 virus; two viruses showed similar efficiency during EBV primary infection (Fig EV5C), indicating that BRLF1 deficiency and inflammasome activation did not affect EBV infection. As expected, the phosphorylation of TAK1 and IKKα/β were more greatly induced by EBV‐ΔBRLF1 infection compared to EBV‐WT infection (Fig EV5D). The expression of several genes involved in T‐ and NK cell activation was significantly increased by BRLF1‐deficient infection compared with EBV‐WT infection, including the IL‐18R1 receptor, the inflammatory factors IFNγ, IL‐17A, and IL‐22 (Fig 6A), the cell killing‐related genes perforin and granzyme‐B (Fig 6B), the T‐cell activation‐specific gene IL‐2 and its receptor (Fig 6C), and NK cell‐activating receptors NKp30, NKp46, and NKG2D (Fig 6D), indicating that BRLF1 deficiency enhanced T‐ and NK cell activation and immune responses during EBV primary infection. Therefore, T‐ and NK cell maturation and activation were analyzed by counting the CD3⁺CD25⁺ and CD56⁺CD69⁺ cells, respectively. In EBV‐ΔBRLF1‐infected PBMCs, the percentage of CD3⁺CD25⁺ and CD56⁺CD69⁺ cells were approximately 50 and 20%, respectively, compared with approximately 7 and 5% in EBV‐WT‐infected PBMCs, which was almost equivalent to the basal level of T‐ and NK cell maturation (Fig 6E–H); this suggested that BRLF1‐deficient EBV infection induces robust T‐ and NK cell maturation and activation, while wild‐type EBV infection successfully evades the induction of their maturation and activation. Interestingly, T‐ and NK cell maturation in EBV‐ΔBRLF1‐infected PBMCs was reduced by half by the administration of the inhibitory peptide TAT‐N572 (Fig 6E–H), indicating that the BRLF1 inhibitory peptide, which rescues inflammasome inhibition in the presence of BRLF1 deficiency, is capable of avoiding T‐ and NK cell activation during EBV primary and lytic infection. These results indicate that BRLF1‐mediated inflammasome inhibition is essential for the evasion of T‐ and NK cell‐mediated immune responses during EBV primary infection in lymphocytes.

Figure 6. BRLF1‐deficient EBV primary infection activates T and NK cells in human lymphocytes.

• A–D
  PBMCs were isolated and infected with EBV‐WT or EBV‐ΔBRLF1 virus (MOI = 500). Thirty‐six hours after infection, the cells were harvested and the total RNAs were extracted, reverse‐transcribed, and subjected to real‐time PCR analysis. The differential expression of receptors and cytokines (A), cytotoxic genes (B), T‐cell‐specific genes (C), and NK cell‐specific genes (D) is shown. The values are shown as the mean ± standard deviations of five analyses of PBMCs from five different health donors. Results are presented as the mean ± SD, n = 5 biological replicates, *P < 0.05. **P < 0.01, Student's t‐test.
• E–H
  FACS analysis of T‐ and NK cell activation in PBMCs in the presence of EBV‐WT or EBV‐ΔBRLF1 primary infection. The PBMCs were left uninfected or infected with EBV‐WT or EBV‐ΔBRLF1 virus for 36 h, and the EBV‐ΔBRLF1‐infected lymphocytes were incubated with additional TAT‐Flag or TAT‐N572 peptide for 12 h. Immunofluorescent PE‐CD3 and APC‐CD25 antibodies were used to detect T‐cell activation (E, G), and PE‐CD56 and APC‐CD69 were used to detect NK cell activation (F, H). Representative images of the FACS analysis of T‐cell and NK cell activation are shown (E, F), and the percentages of activated T cells and NK cells were calculated and are shown as the mean ± standard deviation (G, H) for lymphocytes from 5 different health donors; results are presented as the mean ± SD, n = 5 biological replicates, *P < 0.05. **P < 0.01, Student's t‐test.

Discussion

EBV infection constitutively activates IFI16 inflammasome through viral genomic DNA (Ansari et al, 2013) and induces the expression of inflammasome‐related genes through latent gene expression (Chen et al, 2012; Cai et al, 2017; Torii et al, 2017). Although NLRP3 inflammasome activation is prevented by the EBV‐encoded miRNA miR‐BART15 during latent infection (Haneklaus et al, 2012), the mechanism underlying the regulation of inflammasomes during lytic replication remains unknown. In the present study, we have revealed that the EBV immediate‐early protein BRLF1 suppresses RNA polymerase III‐mediated immunostimulatory RNA transcription and RIG‐I inflammasome activation and innate immune responses during early lytic replication and consequently prevents pyroptosis and inflammasome‐dependent T‐cell and NK cell activation during primary infection and the lytic lifecycle (Appendix Fig S4). Notably, a short BRLF1‐derived peptide is sufficient for the induction of this inhibition. Our studies reveal the distinct evasion of immune responses and inflammasome activation during EBV lytic replication and provide a novel cellular target and a promising strategy to disrupt viral evasion and prevent persistent infection.

Diverse sensors recognize viral DNA, RNA, and other products during the infection of herpesviruses and then constitutively activate immune responses and inflammasomes to control lytic replication and establish latency (Brennan & Bowie, 2010; Lupfer et al, 2015). Subsequently, lytic replication employs the distinct strategies to evade these immune and inflammatory responses. Studies have revealed the suppression of NLRP1/NLRP3 by KSHV ORF63 and of AIM2 inflammasome by HSV VP22 and HCMV pUL83 and IE86 (Gregory et al, 2011; Huang et al, 2017; Maruzuru et al, 2018; Botto et al, 2019). Similarly, EBV BRLF1 suppresses the activation of the RIG‐I inflammasome during early lytic infection, and BRLF1 deficiency during EBV primary infection induces the robust activation of inflammasome and pyroptosis, indicating that the suppression of inflammasome activation and pyroptosis is essential for EBV primary infection and lytic replication.

Interestingly, BRLF1 suppresses RIG‐I signaling in a RNA polymerase III‐dependent manner by interacting with the RNA polymerase III subunits and then disrupting the transcription of viral and cellular small RNAs. As transcription factors, BRLF1 and BZLF1 cooperatively activate cellular and viral lytic transcription through RNA polymerase II (Lu et al, 2006; Miller et al, 2007); however, BRLF1 suppresses the RNA polymerase III‐dependent transcription of viral small RNA, such as EBV EBERs and HSV‐1 LAT, through its interaction with the POLR3F and POLR3G subunits. The recognition of viral DNA by RNA polymerase III and the subsequent activation of RIG‐I signalosome are important for the production of type I interferon and inflammatory factors (Ablasser et al, 2009; Chiu et al, 2009; Zhao et al, 2018). RNA polymerase III transcribes the RIG‐I‐related immunostimulatory RNAs, and the deficiency of RNA polymerase III by POLR3A and POLR3C mutations results in defective IFN production and causes severe varicella zoster virus (VZV) infections (Ogunjimi et al, 2017; Carter‐Timofte et al, 2018), suggesting that this pathway is essential for innate defense against herpesvirus infection. Therefore, the lytic replication of herpesviruses employs multiple strategies to block RIG‐I activation, such as the deubiquitination and deamidation of RIG‐I (Inn et al, 2011; Zhao et al, 2016; Gupta et al, 2018). Alternatively, cellular DNA sensors such as cGAS and AIM2 are also blocked during the lytic infection of herpesviruses (Ma et al, 2015; Wu et al, 2015; Zhang et al, 2016; Huang et al, 2018; Zhang et al, 2018; Botto et al, 2019); however, whether and how the nuclear DNA sensors RNA polymerase III and IFI16 are inhibited remains unclear. Here, we reveal that EBV blocks the RNA polymerase III‐mediated transcription of immunostimulatory RNAs, which represents a novel strategy to inactivate the RIG‐I pathway during the early lytic lifecycle of herpesviruses. Although it is unknown whether this strategy is applicable to alpha‐ or beta‐herpesviruses, the BRLF1 homologue in gamma‐herpesviruses exhibits similar activity; therefore, we believe that the suppression of RNA polymerase III‐mediated immunostimulatory RNA transcription and the inactivation of RIG‐I pathway are common evasive tactics during the lytic replication of herpesviruses. Therefore, the disruption of the viral machinery that inhibits the sensing of viral DNA by RNA polymerase III may be promising strategy to combat herpesvirus infection.

Because of the inhibition of inflammasomes, EBV lytic infection is not sensitive to inflammasome‐dependent responses, and exogenous IL‐1β administration or the inactivation of inflammasomes by caspase‐1 depletion does not affect lytic gene expression. In addition, the levels of the receptors of IL‐1β and IL‐18 are also blocked by BRLF1 expression during EBV primary infection in T and NK cells, suggesting that BRLF1 not only inhibits the activation of inflammasome but also reduces the susceptibility to inflammasome‐dependent factors. In fact, EBV latent infection downregulates IL‐1 receptor 1 (IL‐1R1) through BHRF1‐2 miRNA to block IL‐1β responsiveness (Skinner et al, 2017). These studies indicate that EBV‐infected cells are resistant to inflammasome‐related responses regardless of the stage of the lifecycle. However, a low level of caspase‐1 activation is still required for late lytic replication (Gastaldello et al, 2013; Lv et al, 2018), which is consistent with our preliminary screening results that showed that late genes may activate inflammasomes and caspase‐1 during the late stage of infection (Fig EV1B).

In addition to B cells, T and/or NK cells are often involved in the development of chronic active EBV infection (CAEBV) and develop to T‐ or NK cell malignancies at a low frequency (Kimura & Cohen, 2017; Okuno et al, 2019). Studies have shown that the elevated levels of IL‐1β and IL‐18 can induce Th1 chemokines and retain activated T cells in peripheral tissues (Stojanov et al, 2011). In addition, IL‐12 and IL‐1β synergistically induce T cells to proliferate and produce IFN‐γ, after which IL‐12‐stimulated T cells are activated by IL‐18 or IL‐1β, leading to proliferation and IFN‐γ production (Tominaga et al, 2000). Additionally, NK cell proliferation and activation can be induced by IL‐18 (Fehniger et al, 1999; Lauwerys et al, 1999). These studies have shown that IL‐1β and IL‐18 are inducers of T‐ and NK cell proliferation and activation. Our results show that EBV infection in T and NK cells successfully evades T‐ and NK cell activation and killing through BRLF1‐mediated inflammasome inhibition. BRLF1 deficiency during EBV primary infection induces T‐ and NK cell activation and cell killing through inflammasome‐dependent paracrine and/or autocrine factors, suggesting that inflammasome activation is an important immune defense against EBV infection and EBV‐related diseases. Increased inflammasome activation has been observed in EBV‐positive nasopharyngeal carcinoma (NPC) and is correlated with the survival of NPC patients, in which the enhanced activation of inflammasomes and IL‐1β secretion inhibits tumorigenesis (Chen et al, 2012). BRLF1 is specifically detectable at high levels in NPC biopsies (Martel‐Renoir et al, 1995; Feng et al, 2000; Hu et al, 2016), indicating that the BRLF1‐mediated inhibition of inflammasomes and immune responses plays a critical role in NPC progression and survival. Since IL‐1β and IL‐18 and activated T and NK cells have multiple functions in antiviral and antitumor immune responses, BRLF1‐mediated inflammasome inhibition and IL‐1β/IL‐18 blockade certainly exert many potential effects on EBV persistent infection and diseases, including EBV‐related viral carcinogenesis, the progression of lymphoproliferative disorders, and autoinflammatory diseases. Our results indicate a promising strategy to prevent and cure EBV‐positive diseases through the induction of inflammasome activation or the disruption of inflammasome inhibition.

We have found that a small region of BRLF1 is responsible for the inhibition of RNA polymerase III‐mediated RIG‐I signalosome activation, and a short BRLF1‐derived peptide is sufficient for this inhibition during HSV‐1 infection and the rescue of inhibition during BRLF1‐deficient EBV infection. This finding indicates RNA polymerase III acts as a key modulator of the activation of the RIG‐I inflammasome and immune responses during the early lytic replication of herpesviruses and suggests a promising approach that could be used to develop small compounds or peptides to activate RNA polymerase III‐mediated immune responses and to disrupt the BRLF1‐RNA polymerase III association for the prevention of infection with EBV and other herpesvirus‐related diseases. Additionally, RNA polymerase III is activated by infection with several small DNA tumor viruses, and the abnormal activation of RNA polymerase III has been observed in tumor cells (Brown et al, 2000); therefore, the suppression of its activity by a BRLF1‐derived peptide and its influence on tumorigenesis are worth further investigation.

In conclusion, our current studies demonstrate that BRLF1 is a key suppressor of RNA polymerase III‐mediated RIG‐I inflammasome activation and innate immune responses during the EBV early lytic lifecycle. BRLF1 directly interacts with RNA polymerase III subunits to inhibit the transcription of immunostimulatory small RNAs and consequently suppresses pyroptosis and the inflammasome‐dependent activation of lymphocytes during EBV primary and lytic replication. This may represent a novel mechanism that underlies the evasion of inflammasome activation and innate immune responses during the lytic lifecycle of herpesviruses.

Materials and Methods

Cell culture

HEK293T, Ramos, HNE‐1, p3HR‐1, A549, K562, Jurkat, NK‐92MI, and THP‐1 cell lines were maintained in our laboratory. Adherent cells were cultured in DMEM supplied with 10% fetal bovine serum (FBS) and 1% antibiotics (penicillin and streptomycin). Lymphoma cells were cultured in suspension in RPMI 1640 containing 10% fetal bovine serum (FBS) and 1% antibiotics (penicillin and streptomycin). Whole blood from healthy donors was purchased from Guangzhou blood bank, and peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation using Lymphoprep (STEMCELL Technologies, Vancouver, BC) according to the manufacturer's instructions. The PBMCs were washed twice with PBS and resuspended in RPMI 1640 containing 10% fetal bovine serum (FBS) at a final concentration of 1 × 10⁶ cells/ml. All human blood work was approved by the medical ethics committee at Sun Yat‐sen University.

Human samples

All human blood work was approved by the medical ethics committee of the Sun Yat‐sen University. Donors were informed and gave signed consent to the Guangzhou Blood Bank that their blood may be used anonymized in scientific research to improve people's health before donating. The Institute of Human Virology, Zhongshan School of Medicine, Sun Yat‐Sen University, has an approved research cooperation with the Guangzhou Blood Bank and blood samples provided by Guangzhou Blood Bank under this agreement can be used by the virus research teams of Institute of Human Virology. We followed the ethics/consent rules of the ethics committee of the Sun Yat‐sen University; copies of the documents are available on request.

Plasmids and peptides

GFP‐tagged EBV lytic ORF expression plasmids were subcloned into the pEGFP vector as described previously (Li et al, 2015). The BRLF1 mutant‐expressing plasmids were constructed by QuickChange mutagenesis and subcloned into the pEGFP vector as well. BRLF1 and BRLF1 mutant constructs were subcloned into the pLVX‐IRES‐ZsGreen vector for lentiviral packaging. GST‐tagged BRLF1, POLR3F, and POLR3G were subcloned into the pEBG vector, and Flag‐tagged IFI16, RIG‐I, AIM2, and NLRP3 were cloned into the pCMV‐3tag vector. All shRNAs were expressed with the pLKO.1‐TRC vector. The peptides (> 90% purity) were synthesized by Royobiotech (Shanghai, China). The peptides were dissolved in double‐distilled H₂O, and small aliquots of the peptide solutions were stored at −80°C.

Immunoprecipitation and immunoblotting analysis

Immunoprecipitation and Western blotting analyses were performed as described previously. In brief, one 10‐cm dish of cells was transfected with 10–15 μg plasmid for 48 h, and then, the cells were collected and lysed in the presence of a protease inhibitor cocktail (Roche) and phosphatase inhibitors. For immunoprecipitation, the cell lysates were precleaned and incubated with antibodies at 4°C overnight, and then, the complexes were precipitated with protein G‐agarose or GST‐affinity beads. After washing five times, the immunoprecipitated complexes were subjected to immunoblotting analysis. For Western blotting, 40–60 μg protein of whole‐cell extracts per lane was separated by SDS–PAGE and transferred to membranes. Alternatively, the proteins in the supernatants were precipitated with a final concentration of 10% (m/v) TCA at 4°C. After washing two times with acetone and one time with methanol, the protein pellets were dissolved in SDS loading buffer and subjected to immunoblotting. The membranes were blocked in 5% dry milk and then incubated with primary antibodies at 4°C overnight, and then, they were subsequently incubated with species‐matched IRDye 680‐ or IRDye 800‐labeled secondary antibodies for 2 h at room temperature. The images were visualized using a LiCor Odyssey system.

Analysis of inflammasome activation

To detect inflammasome activation, cells were infected with HSV‐1 (MOI = 1) for 12 h, and then, the cleavage and processing of caspase‐1, IL‐1β, IL‐18, and GSDMD were analyzed by immunoblotting; the secretion of mature IL‐1β and IL‐18 was detected by ELISA.

Luciferase assays

A Gaussia luciferase‐based reporter for inflammasome activation was used to detect caspase‐1 activation and IL‐1β cleavage. Briefly, the pro‐IL‐1β‐DN reporter was transfected into A549 cells with either an empty vector or an EBV lytic ORF‐expressing plasmid in a 96‐well plate. Twenty‐four hours later, the cells were infected with HSV‐1 (MOI = 1) for 12 h, and then, the supernatants were collected and the secretion of Gaussia luciferase was measured using a reagent for the measurement of Renilla luciferase activity and a TriStar multimode reader. For each condition, three independent experiments were performed in triplicate.

Lentivirus packaging and infection

HEK293T cells in 10‐cm dishes were cotransfected with 10 μg of pLVX lentiviral expression plasmid or pLKO.1 shRNA plasmid, 5 μg psPAX2 packaging plasmid, and 5 μg pMD.2G envelope plasmid. The supernatants were harvested at 72 h post‐transfection and concentrated by ultracentrifugation at 100,000 × g for 1 h, and then, the pellets were suspended in PBS to prepare 100‐fold viral stocks. After preliminary titration, the lentiviral infections were performed following standard procedures.

EBV infection and the induction of viral lytic replication

The BRLF1‐KO and EBV‐WT bacterial artificial chromosome (BAC) DNAs were a kind gift from Dr. Henri‐Jacques Delecluse at the German Cancer Research Center (DKFZ), Heidelberg, Germany (Feederle et al, 2000). EBV BAC genomic DNA was transfected into 293T cells, and BAC‐harboring cells were selected with hygromycin for 2 weeks, after which the virions were produced and purified using ultracentrifugation. HNE‐1, Ramos, Jurkat, NK‐92MI, and PBMC cells were infected in the presence of 4 μg/ml polybrene with EBV viral stocks from 293‐BRLF1‐KO or 293‐EBV‐WT cells at a high titer as described previously (Wang et al, 2015), and then, the lytic program was induced by transfecting an expression plasmid encoding BZLF1 or BRLF1. P3HR‐1 cells were induced with 20 ng/ml TPA plus 0.3 mM NaB, and BRLF1‐KO and EBV‐WT cells were transfected with an expression plasmid encoding BZLF1 or BRLF1 to induce EBV lytic replication.

CDOCKER molecular docking analysis

The sequence and 2D structure of BRLF1 were downloaded from NCBI, and I‐TASSER (http://zhanglab.ccmb.med.umich.edu/I‐TASSER/download/) was used to model the 3D molecular structure of BRLF1, the BRLF1 TA domain, and the BRLF1 NLS572‐582 peptide. The RNA polymerase III open complex 3D model (PDB ID 6F40) was downloaded from Protein Data Bank (http://www.rcsb.org). For the molecular docking analysis, Discovery Studio software CDOCKER (Accelrys Software Inc., San Diego, CA, USA) was used to generate the best binding mode of 3D ligand–receptor interaction.

RNA sequencing

Three independent groups of Ramos cells were infected with EBV‐WT or EBV‐ΔBRLF1 virus at a high titer (MOI = 100) or left uninfected. Total RNAs were extracted with TRIzol following the manufacturer’s procedure. Approximately 10 μg of total RNAs were subjected to noncoding RNA sequencing. Library construction and sequencing were performed by Annoroad Gene Technology (Beijing, China). The libraries were sequenced on an Illumina HiSeq‐Xten platform, and 150‐bp paired‐end reads were generated. An average read depth of 12G reads per sample was collected and analyzed. Data analysis was performed with Molecule Annotation System 3.0 (MAS 3.0; Annoroad Gene Technology Co., Ltd, Beijing, China).

Mass spectrometry analysis

A549 cell was transfected with an empty GST‐tagged vector or a GST‐BRLF1‐expressing plasmid for 36 h and then infected with HSV‐1 (MOI = 1) for 12 h. Afterward, cells were collected, and BRLF1‐binding proteins were isolated with GST‐affinity beads. After the immunoprecipitated complexes were confirmed by SDS‐PAGE and silver staining, mass spectrometry analysis of the BRLF1‐binding complexes was performed by Novogene Biotech Co., Ltd. (Beijing, China). The nonspecific binding proteins and GST‐binding proteins were compared and removed.

Chromatin immunoprecipitation (ChIP)

The cells were fixed with 1% formaldehyde for 10 min and quenched with 125 mM glycine. After the cells were collected and washed twice with cold phosphate‐buffered saline (PBS), the cells (> 10⁷) were suspended in 1 ml of lysis buffer [150 mM NaCl, 50 mM Tris–HCl (pH 7.5), 5 mM EDTA, 0.5% NP‐40, 1% Triton X‐100, 1 mM phenylmethylsulfonyl fluoride (PMSF), 1× protease inhibitor cocktail] for 10 min on ice. After the lysates were subjected to sonication with the following condition: 5 s ON and 10 s OFF at power level of 75% for a total of 15 min on ice, the cells were centrifuged at 12,000 g at 4°C for 10 min. Precleaned cell lysates were mixed with GST‐affinity beads at 4°C overnight. The immunoprecipitated complexes were washed five times and eluted with 150 μl of Tris‐EDTA buffer containing 1% SDS. Cross‐linked complexes were reversed and digested with protease K (25 μg/ml) at 65°C overnight. Then, DNA was extracted and analyzed by real‐time PCR.

Real‐time PCR analysis

Total RNA was extracted, reverse‐transcribed into cDNA, and analyzed by real‐time PCR with SYBR Green. The primers for real‐time PCR were designed using PrimerBank (https://pga.mgh.harvard.edu/primerbank/), and the sequences of the primer pairs are listed in the supplementary materials.

Paracrine‐conditioned medium and cell culture

A total of 2 × 10⁵/ml Ramos cells were infected with wild‐type or BRLF1‐KO EBV virus at a multiplicity of infection (MOI) of 100. After infection with BZLF1‐ or BRLF1‐expressing lentivirus or control lentivirus for 8 h, the cells were washed twice with phosphate buffer and then incubated with fresh medium for an additional 24 h. The supernatants were then collected, filtered, and stored as conditioned medium. Then, Jurkat and NK‐92MI cells were cultured in conditioned medium mixed with 50% fresh medium (vol/vol) for 24 h.

Flow cytometry analysis

Pyroptotic cells were stained with PI and detected using flow cytometry. Briefly, 1 × 10⁶ cells were collected, gently washed twice with ice‐cold phosphate‐buffered saline (PBS), and suspended in 500 μl PBS. Then, 2.5 μl 1 mg/ml propidium iodide (PI) was added and incubated at 37°C for 30 min in the dark. After washing with PBS twice, the cells were suspended in 500 μl PBS and then analyzed using a BD LSRFortessa fluorescence‐activated cell sorter (FACS) system (BD Bioscience). T‐cell and NK cell activation in PBMCs was analyzed with flow cytometry. Briefly, 1 × 10⁶ EBV‐infected PBMC cells were collected, gently washed twice with ice‐cold phosphate‐buffered saline (PBS), and resuspended in 100 μl PBS. Then, 2 μl anti‐CD3‐PE and 2 μl anti‐CD25‐APC (for T cells) or 2 μl anti‐CD56‐PE and 2 μl anti‐CD69‐APC (for NK cells) were added and incubated on ice for 15–20 min in the dark, and then, the cells were washed with PBS twice, resuspended in 500 μl PBS, and analyzed using a BD LSRFortessa fluorescence‐activated cell sorter (FACS) system (BD Bioscience).

Antibody neutralization

After infection with EBV overnight, cells were washed twice and cultured in fresh medium, and then, 2 μg/ml anti‐IL‐1β, anti‐IL‐18, or an isotype‐matched control IgG (eBioscience) was directly added to the cell culture medium. The cells were cultured for an additional 24 h before being harvested and further analyzed.

ASC oligomerization assay

Ten‐centimeter dishes of cells were transfected with expression plasmids for 48 h. The cells were collected and lysed in immunoprecipitation lysis buffer. The cell lysates were centrifuged, and the supernatants were collected in a fresh tube. Equal amounts of total protein were cross‐linked with DSS at a final concentration of 2 mM at room temperature for 30 min, and then, the reaction was quenched by quenching buffer with a final concentration of 20–50 mM Tris–HCl (pH 7.5) at room temperature for 15 min. Finally, ASC oligomerization was detected by Western blotting with anti‐ASC antibody.

Splint‐ligation analysis

RNA was extracted with phenol:chloroform:isoamyl alcohol [25:24:1 (vol/vol)] followed by ethanol precipitation. To measure the total EBER1 RNA, 50 μg of total RNA was treated with alkaline phosphatase (calf intestine, CIAP) (Takara) for 1 h at 37°C according to the manufacturer’s recommendation, and then, 5 μg of CIAP‐treated RNA was phosphorylated by T4 polynucleotide kinase. The splint ligations were performed as previously described. For ligation, each reaction consisted of 100 fmol bridge oligonucleotide, 200 fmol FAM‐labeled ligation oligonucleotide, 5 μg untreated (for 5ʹ‐pRNA) or treated RNA (for total RNA), 1× T4 DNA ligase reaction buffer (Takara), and 10 units T4 DNA ligase (Takara). After the reaction mixture was denatured at 95°C for 1 min, cooled to 65°C for 5 min, and incubated at 37°C for 10 min, T4 DNA ligase was added to the reaction mixture and incubated at 30°C for 4 h. The reactions were terminated by heat inactivation at 75°C for 10 min and subsequently separated using denaturing 8 M urea 15% polyacrylamide gels. The images were obtained using a Bio‐Rad Molecular Imager^® Gel Doc XR System and analyzed using ImageJ software.

Analysis of Jurkat and NK‐92MI cell killing activity

K562 cells, Jurkat cells, or NK‐92MI cells were washed with PBS twice and then resuspended in RPMI 1640 medium at a final concentration of 4 × 10⁵/ml, 2 × 10⁷/ml, or 2 × 10⁷/ml, respectively. A total of 100 μl K562 cells and 100 μl Jurkat cells or NK‐92MI cells were seeded in each well of a 96‐well plate and cocultured for an additional 4 h at 37°C in an incubator with a 5% CO₂. Then, the supernatants were collected and the released amounts of LDH were analyzed. For each condition, three independent experiments were performed in triplicate.

Quantification and statistical analysis

The data were obtained from at least three independent experiments (n ≥ 3), and the representative data are presented as the mean ± standard deviation as indicated. The data were analyzed using an unpaired Student’s t‐test or Spearman rank correlation. GraphPad Prism 8 software was used to calculate the P‐values, and significance is depicted with asterisks as follows: *P < 0.05, **P < 0.01. Multiple comparison test was performed by SPSS software using 1‐way ANOVA and Tukey's multiple comparison test (α = 0.05), and significance is depicted with asterisks, as above.

Author contributions

XLo, XLi, and EK designed experiments and analyzed data. XLo, JY, XZ, ZY, YL, and FW performed experiments. EK wrote the paper.

Conflict of interest

A patent application for the BRLF1‐derived peptide for use in inflammasome activation and immune responses has been filed.

Supporting information

Appendix

Expanded View Figures PDF

Dataset EV1

Dataset EV2

Dataset EV3

Dataset EV4

Dataset EV5

Source Data for Expanded View and Appendix

Review Process File

Source Data for Figure 1

Source Data for Figure 2

Source Data for Figure 4

Source Data for Figure 5

EMBO Reports (2021) 22: e50714.

Data availability

The data and details of this study are available from the corresponding authors upon request. The raw RNA sequencing data of Ramos cells undergoing mock, EBV‐WT, or EBV‐ΔBRLF1 primary infection were deposited at the NCBI SRA archive under accession number GSE154786 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE154786). The mass spectrometry data from this publication have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the iProX partner repository (Ma et al, 2019) and assigned the identifier PXD021958.