NF-kappaB serves as a cellular sensor of Kaposi's sarcoma-associated herpesvirus latency and negatively regulates K-Rta by antagonizing the RBP-Jkappa coactivator

Abstract

Successful viral replication is dependent on a conducive cellular environment; thus, viruses must be sensitive to the state of their host cells. We examined the idea that an interplay between viral and cellular regulatory factors determines the switch from Kaposi's sarcoma-associated herpesvirus (KSHV) latency to lytic replication. The immediate-early gene product K-Rta is the first viral protein expressed and an essential factor in reactivation; accordingly, this viral protein is in a key position to serve as a viral sensor of cellular physiology. Our approach aimed to define a host transcription factor, i.e., host sensor, which modulates K-Rta activity on viral promoters. To this end, we developed a panel of reporter plasmids containing all 83 putative viral promoters for a comprehensive survey of the response to both K-Rta and cellular transcription factors. Interestingly, members of the NF-kappaB family were shown to be strong negative regulators of K-Rta transactivation for all but two viral promoters (Ori-RNA and K12). Recruitment of K-Rta to the ORF57 and K-bZIP promoters, but not the K12 promoter, was significantly impaired when NF-kappaB expression was induced. Many K-Rta-responsive promoters modulated by NF-kappaB contain the sequence of the RBP-Jkappa binding site, a major coactivator which anchors K-Rta to target promoters via consensus motifs which overlap with that of NF-kappaB. Gel shift assays demonstrated that NF-kappaB inhibits the binding of RBP-Jkappa and forms a complex with RBP-Jkappa. Our results support a model in which a balance between K-Rta/RBP-Jkappa and NF-kappaB activities determines KSHV reactivation. An important feature of this model is that the interplay between RBP-Jkappa and NF-kappaB on viral promoters controls viral gene expression mediated by K-Rta.

FIG. 1.

Identification of K-Rta-responsive KSHV promoters. Individual cultures of 293 cells were cotransfected with the K-Rta expression plasmid vector and a luciferase reporter plasmid containing 1 of each of 83 putative KSHV gene promoters. Fold activation over plasmid vector control is shown. Thirty-four KSHV promoters were activated over 10-fold.

FIG. 2.

NF-κB is a repressor of K-Rta. (A) 293 cells were cotransfected with K-Rta, K-Rta-responsive promoter, and NF-kB (RelA). Luciferase activity of K-Rta alone was normalized to a value of 1. (B) Effects of canonical and noncanonical pathways of NF-κB on K-Rta activity. 293 cells were cotransfected with different combinations of NF-κB proteins, K-Rta, and ORF57 promoter. (C) Enhancement of K-Rta activity by NF-κB inhibitor. 293 cells were cotransfected with K-Rta, ORF57 promoter, and the indicated plasmid. RelA significantly inhibits K-Rta transactivation, and IκBα enhances K-Rta activity. RLU, relative light units.

FIG. 3.

NF-κB inhibits KSHV reactivation. (A) Schematic diagram of the dual inducible cell lines. The cassette (for dual induction of IκBα and K-Rta or RelA and K-Rta) was transduced into TREx-BCBL-1 cells by cotransfecting with Flp recombinase expression vector. This cassette integrates at the identical site because of the design of the recombination system. Therefore, the copy number and the transcription levels remain the same. By adding doxycycline, K-Rta triggers reactivation, and the effects of NF-κB on KSHV reactivation can be examined. (B) RelA inhibits K-bZIP expression. After 12 h of induction, total cell protein was prepared and the indicated proteins were probed with specific antibodies. RelA reduced K-bZIP expression. W.B., Western blot. (C) RelA inhibits KSHV gene expression. Total RNA was prepared at 0, 12, and 24 h after induction. Viral gene expression was analyzed by measuring levels of transcripts by real-time qPCR. Cellular actin gene expression was used as an internal control. All values are expressed as the fold increase over the 0-h (uninduced) time point. Triplicate wells were used for each time point, and at least two independent experiments were conducted. (D) RelA inhibits KSHV replication. Viral DNA copy number in culture supernatant, as a measure of virion levels, was analyzed by real-time qPCR.

FIG. 4.

NF-κB inhibits K-Rta recruitment to the promoter. (A) Confirmation of KSHV reactivation. Total cell lysates were prepared at the indicated time points. For SDS-PAGE and immunoprobing, 20 μm of lysate per lane was used to test for the anti-K-Rta antibody, and 100 μg per lane was used for the K8.1 antibody. Comparable amounts of K-Rta expression were confirmed, and induction of KSHV lytic replication was assessed by measuring K8.1 expression. W.B., Western blot. (B) Time course analysis of K-Rta recruitment to selected KSHV promoters during reactivation. (a) Low-cycle PCR analysis was performed with ChIP DNA. (b) All values are expressed as fold increase over IgG control pull-down results, assayed by real-time PCR at 4, 12, and 24 h after doxycycline treatment. Each sample was tested in triplicate, and average values are shown in the figure. The ChIP elutions were assayed directly by real-time PCR after a 1:50 dilution. Two independent experiments were performed with similar results.

FIG. 5.

Competition of promoter binding between NF-κB and RBP-Jκ. (A) Purified proteins used in the gel shift assay. The RBP-Jκ, RelA, NF-κB1 (p50), and K-Rta proteins were produced in baculovirus expression systems. (B) NF-κB inhibits RBP-Jκ binding to the promoter. A gel shift assay was performed with the ORF57 promoter probe (gels a and b) or K12 promoter probe (gel c). The final concentration of RBP-Jκ was kept 100 nM (+). K-Rta was used at 400 nM, and NF-κB (RelA/p50 complex) was used at either a 100 nM (+) or 400 nM (++) final concentration. A supershift experiment was conducted with anti-RBP-Jκ andtibody (Ab). Specific (S) or nonspecific (NS; AP-1) competitor was used at a 100-fold excess of the labeled probe. Short, short exposure; long, long exposure.

FIG. 6.

RBP-Jκ interactions with Rel-A and functional consequences. (A) Association between RelA and RBP-Jκ. Purified RelA (0.1 μg) and RBP-Jκ (0.1 μg) were mixed in binding buffer, incubated at 4°C, and immunoprecipitated with anti-RelA antibody. The immunoprecipitate was probed with the indicated antibodies. The asterisk shows the position of the IgG heavy chain. W.B., Western blot. (B) Association between RelA and RBP-Jκ in transfected 293T cells. The indicated plasmids were cotransfected into 293T cells. Total cell lysates were immunoprecipitated with agarose beads conjugated with anti-Flag antibody. Immunoprecipitates were probed with the indicated antibodies. Endogenous RelA was also detected with anti-RelA rabbit IgG (right upper panel). (C) Suppression of NF-κB activation by RBP-Jκ. The NF-κB reporter was cotransfected with the RBP-Jκ expression plasmid. NF-κB was activated by treating the cultures with TNF-α at 24 h posttransfection. Forty-eight hours after transfection, cells were harvested, and levels of luciferase were determined. Fold activation over the control transfected value is shown. The amount of plasmid used is indicated at the bottom of the panel. (D) RelA expression reduced K-Rta binding to RBP-Jκ. Dual expression of IκBα and K-Rta or RelA and K-Rta was induced in BCBL-1 by adding doxycycline. After 12 h of induction, cells were harvested and subjected to immunoprecipitation with anti-RBP-Jκ antibody. Coimmunoprecipitated K-Rta was measured by immunoblotting with anti-K-Rta antibody. Due to similar molecular masses, the presence of rabbit IgG heavy chain obscured the detection of the immunoprecipitated RBP-Jκ.