Dengue virus co-opts UBR4 to degrade STAT2 and antagonize type I interferon signaling

Abstract

An estimated 50 million dengue virus (DENV) infections occur annually and more than forty percent of the human population is currently at risk of developing dengue fever (DF) or dengue hemorrhagic fever (DHF). Despite the prevalence and potential severity of DF and DHF, there are no approved vaccines or antiviral therapeutics available. An improved understanding of DENV immune evasion is pivotal for the rational development of anti-DENV therapeutics. Antagonism of type I interferon (IFN-I) signaling is a crucial mechanism of DENV immune evasion. DENV NS5 protein inhibits IFN-I signaling by mediating proteasome-dependent STAT2 degradation. Only proteolytically-processed NS5 can efficiently mediate STAT2 degradation, though both unprocessed and processed NS5 bind STAT2. Here we identify UBR4, a 600-kDa member of the N-recognin family, as an interacting partner of DENV NS5 that preferentially binds to processed NS5. Our results also demonstrate that DENV NS5 bridges STAT2 and UBR4. Furthermore, we show that UBR4 promotes DENV-mediated STAT2 degradation, and most importantly, that UBR4 is necessary for efficient viral replication in IFN-I competent cells. Our data underscore the importance of NS5-mediated STAT2 degradation in DENV replication and identify UBR4 as a host protein that is specifically exploited by DENV to inhibit IFN-I signaling via STAT2 degradation.

Figure 1. DENV NS5 binds UBR4.

(A) RFP-ubiquitin-NS5-TAP was expressed in 293T cells, in the presence or absence of human STAT2-FLAG, and then purified using the tandem affinity purification (TAP) method. The purified proteins were resolved on an SDS 4–15% gel that was then stained with silver. *NS5 and STAT2 run in similar positions in the gel. The bands highlighted by the arrows were analyzed by mass spectrometry at Taplin Biological Mass Spectrometry Facility, Harvard Medical School. (B and C) HA-tagged DENV2, DENV1, WNV and/or YFV NS5 proteins were immunoprecipitated from 293T cells using anti-HA beads. Immunoblot (IB) analysis of the immunoprecipitate (IP) and whole cell extract (WCE) was performed against UBR4, UBR5, STAT2, HA and/or β-tubulin.

Figure 2. The UBR4-NS5 interaction correlates with NS5-mediated STAT2 degradation.

(A) A representation of the DENV2 NS5 constructs used in this study. The E domain upstream of ubiquitin refers to the E protein of DENV2. (B and C) HA-tagged DENV NS5 constructs were immunoprecipitated from 293T cells using anti-HA beads. Immunoblot (IB) analysis of the immunoprecipitate (IP) and whole cell extract (WCE) was performed against UBR4, STAT2, HA and β-tubulin or GAPDH. (D) Vero cells expressing HA-tagged DENV NS5 constructs were fixed for immunofluorescence (IF) analysis of UBR4 (green) and NS5 (red). The nuclei were also stained with DAPI (blue). (E) An alignment of the first 10 or 11 amino acids of NS5 from several flaviviruses. Residues that are conserved in all examined flaviviruses are highlighted in green, while those residues that are only conserved amongst the four DENV serotypes are highlighted in red. (F) GFP-tagged DENV2 NS5 constructs and DENV2 NS5 point mutants were immunoprecipitated from 293T cells using anti-GFP antibody and protein G beads. Immunoblot (IB) analysis of the immunoprecipitate (IP) and whole cell extract (WCE) was performed against UBR4, STAT2, GFP and β-tubulin.

Figure 3. NS5 promotes STAT2 association with UBR4.

(A) Proteolytically-processed HA-tagged DENV2 NS5 constructs or an empty vector were expressed in human cells that naturally express human STAT2 (293T cells), human cells that do not express STAT2 (U6A cells), or murine cells that naturally express mouse STAT2 (Hepa1.6), and the lysates of these cells were subjected to anti-HA immunoprecipitation. Immunoblot (IB) analysis of the immunoprecipitate (IP) and whole cell extract (WCE) was performed against UBR4, human STAT2, mouse STAT2, HA and β-tubulin. (B) FLAG-tagged STAT constructs were overexpressed in 293T cells with or without proteolytically-processed HA-tagged DENV2 NS5. Immunoprecipitation was performed against the FLAG epitope, then IB analysis of the IP and WCE was performed against UBR4, FLAG, HA and GAPDH. hSTAT2 = human STAT2; mSTAT2 = mouse STAT2; hSTAT1 = human STAT1; h/mSTAT2 = a chimeric protein with the first 301 amino acids of mouse STAT2 replaced by the corresponding human STAT2 sequence; hSTAT2/1 = a chimeric protein with the first 316 of human STAT1 amino acids replaced by the corresponding human STAT2 sequence. (C) 293T cells and (D) monocyte-derived dendritic cells (MDDC) were mock infected or infected with DENV2 (Thailand/16681 strain) at an MOI of 3 for 24 hours then lysed for IP with anti-STAT2 antibody or control IgG. IB analysis of the IP and WCE was performed against UBR4, UBR5, STAT2, DENV NS5 and β-tubulin.

Figure 4. UBR4 promotes DENV-mediated STAT2 degradation.

(A) 293T cells stably expressing non-targeting shRNA (shControl) or shRNA against UBR4 (shUBR4) were infected with DENV2 (Thailand/16681 strain) at an MOI of 10. Each shUBR4 clone expressed shRNA targeting different sequences within the UBR4 gene. Lysates from mock-infected cells (M) and cells that been infected with DENV2 for 4, 8, 12 or 24 hours were subjected to immunoblot (IB) analysis against UBR4, UBR5, STAT2, DENV NS5 and β-tubulin. (B) Lysates from shControl and shUBR4 293T cells that had been mock infected or infected with DENV1, DENV3 or DENV4 at an MOI of 1 for 24 hours were subjected to IB analysis. (C) A proteolytically-processed HA-tagged DENV NS5 construct or empty vector were transfected into 293T cells stably expressing a V5-tagged construct of the first 2233 amino acids of UBR4 (UBR4-NT). The lysates of these cells were subjected to anti-HA immunoprecipitation. Immunoblot (IB) analysis of the immunoprecipitate (IP) and whole cell extract (WCE) was performed against UBR4, human STAT2, V5, HA and β-tubulin. (D) 293T cells expressing UBR4-NT were mock infected or infected with DENV at an MOI of 1. Lysates were subjected to IB analysis against STAT2, V5, NS5 and β-tubulin at 24 hours post infection.

Figure 5. UBR4 enhances the replication of DENV in an IFN-I-dependent manner.

(A) 293T cells stably expressing non-targeting shRNA (shControl) or shRNA against UBR4 (shUBR4 #2) were infected with DENV2 (Thailand/16681 strain), YFV (17D strain) or EMCV at the indicated multiplicities of infections (MOI). Cells and media were harvested at 24 hours post-infection and measured for virus by plaque assay on BHK-21 cells. (B) 293T cells were infected with DENV1, 2, 3 or 4 at an MOI of 1 and the virus levels were measured by plaque assay on BHK-21 cells 24 hours later. (C) Vero cells stably expressing non-targeting shRNA (shControl) or shRNA against UBR4 (shUBR4 #1) were infected with DENV2 (Thailand/16681 strain) at an MOI of 0.1. Cells were mock-treated or treated with IFN-I (1000units/ml IFNβ) at 6 hours post infection. Cells and media were harvested at 24 hours post-infection for plaque assay or immunoblotting (IB) with antibodies against UBR4, STAT2, NS5 and β-tubulin. (C) 293T cells were infected with DENV2 at an MOI of 1 and incubated with media containing 0.1 µg/ml control IgG antibody or 0.1 µg/ml IFNAR antibody. Cells and media were harvested at 24 hours post-infection and measured for virus by plaque assay on BHK-21 cells. (D) 293T clones were infected with DENV2 at an MOI = 1 and their RNA harvested at the indicated time post infection. The levels of ISG54 mRNA were measured by qPCR, normalized to 18s (a housekeeping gene) and represented as fold induction over mock samples.). Each graph represents the mean +/− the standard deviation of 6 experiments (A), 3 experiments (B, C and D), or 2 experiments (E). Statistical analyses were conducted using the unpaired t test function of Prism 4 for Macintosh (GraphPad Software, USA). *p<0.05; **p<0.01; *** p<0.001; ns = not statistically significant, where p>0.05.

Figure 6. UBR4 enhances the replication of DENV in monocyte-derived dendritic cells.

Monocytes from 5 donors were transduced with a non-targeting shRNA lentivirus (shControl) or an shRNA lentivirus against UBR4 (shUBR4 #2). Monocytes were differentiated for 5 days to monocyte-derived dendritic cells (MDDCs) before infecting with DENV2 (Thailand/16681 strain) at an MOI of 3. Cells were harvested at 12 hours post infection for measurement of (A) UBR4, (B) ISG15, (C) RIG-I, and (D) ISG54 mRNA levels, or (E) supernatants were harvested at 48 hours post infection for measurement of virus levels via plaque assay. RIG-I mRNA was undetectable in 1 donor. Statistical analyses were conducted using the two-way ANOVA function of Prism 4 for Macintosh (GraphPad Software, USA). p<0.001 for UBR4, ISG15, RIG-I and ISG54 levels and p<0.01 for DENV titers in shControl versus shUBR4 cells.

Figure 7. Model of DENV-mediated STAT2 degradation.

NS2B/3 protease cleaves NS4B away from NS5 to yield proteolytically-processed NS5. Proteolytically-processed NS5 brings STAT2 into contact with the degradation machinery of the cell by binding UBR4. UBR4 may function as an E3 ligase (A), but as it lacks a known E3 ligase catalytic domain such as the HECT or RING, it may instead act as the recognition component or scaffold of an E3 ligase complex (B). STAT2 degradation prevents the establishment of an antiviral state thereby allowing DENV to replicate optimally.