Histone H2B-IFI16 Recognition of Nuclear Herpesviral Genome Induces Cytoplasmic Interferon-β Responses

Abstract

IFI16 (gamma-interferon-inducible protein 16), a predominantly nuclear protein involved in transcriptional regulation, also functions as an innate immune response DNA sensor and induces the IL-1β and antiviral type-1 interferon-β (IFN-β) cytokines. We have shown that IFI16, in association with BRCA1, functions as a sequence independent nuclear sensor of episomal dsDNA genomes of KSHV, EBV and HSV-1. Recognition of these herpesvirus genomes resulted in IFI16 acetylation, BRCA1-IFI16-ASC-procaspase-1 inflammasome formation, cytoplasmic translocation, and IL-1β generation. Acetylated IFI16 also interacted with cytoplasmic STING and induced IFN-β. However, the identity of IFI16 associated nuclear proteins involved in STING activation and the mechanism is not known. Mass spectrometry of proteins precipitated by anti-IFI16 antibodies from uninfected endothelial cell nuclear lysate revealed that histone H2B interacts with IFI16. Single and double proximity ligation microscopy, immunoprecipitation, EdU-genome labeled virus infection, and chromatin immunoprecipitation studies demonstrated that H2B is associated with IFI16 and BRCA1 in the nucleus in physiological conditions. De novo KSHV and HSV-1 infection as well as latent KSHV and EBV infection induces the cytoplasmic distribution of H2B-IFI16, H2B-BRCA1 and IFI16-ASC complexes. Vaccinia virus (dsDNA) cytoplasmic replication didn't induce the redistribution of nuclear H2B-IFI16 or H2B into the cytoplasm. H2B is critical in KSHV and HSV-1 genome recognition by IFI16 during de novo infection. Viral genome sensing by IFI16-H2B-BRCA1 leads to BRCA1 dependent recruitment of p300, and acetylation of H2B and IFI16. BRCA1 knockdown or inhibition of p300 abrogated the acetylation of H2B-IFI16 or H2B. Ran-GTP protein mediated the translocation of acetylated H2B and IFI16 to the cytoplasm along with BRCA1 that is independent of IFI16-ASC inflammasome. ASC knockdown didn't affect the acetylation of H2B, its cytoplasmic transportation, and the association of STING with IFI16 and H2B during KSHV infection. Absence of H2B didn't affect IFI16-ASC association and cytoplasmic distribution and thus demonstrating that IFI16-H2B complex is independent of IFI16-ASC-procaspase-1-inflammasome complex formed during infection. The H2B-IFI16-BRCA1 complex interacted with cGAS and STING in the cytoplasm leading to TBK1 and IRF3 phosphorylation, nuclear translocation of pIRF3 and IFN-β production. Silencing of H2B, cGAS and STING inhibited IFN-β induction but not IL-1β secretion, and cGAMP activity is significantly reduced by H2B and IFI16 knockdown during infection. Silencing of ASC inhibited IL-1β secretion but not IFN-β secretion during de novo KSHV and HSV-1 infection. These studies identify H2B as an innate nuclear sensor mediating a novel extra chromosomal function, and reveal that two IFI16 complexes mediate KSHV and HSV-1 genome recognition responses, with recognition by the IFI16-BRCA1-H2B complex resulting in IFN-β responses and recognition by IFI16-BRCA1 resulting in inflammasome responses.

Fig 1. Demonstration of IFI16 interactions with H2B in the nuclear fraction of uninfected cells.

(A-D) Cytoplasmic and nuclear fractions from uninfected BJAB, HMVEC-d and HFF cells were immunoprecipitated (IP-ed) using anti-IFI16 and anti-H2B antibodies and immunoblotted for IFI16, H2B, BRCA1, H2A and ASC. The nuclear fraction from HMVEC-d cells was IP-ed with control IgG antibody for specificity control. (E and F) Uninfected cytoplasmic and nuclear fractions were western blotted with the above antibodies for equal inputs. TBP and Tubulin were used to monitor the purity of the nuclear and cytoplasmic fractions, respectively. (G and H) Uninfected BJAB, HMVEC-d and HFF cells were subjected to PLA reactions using anti-IFI16 and H2B antibodies as described in Materials and Methods. Boxed areas were enlarged and interactions of IFI16 with H2B are indicated by the red arrows. Bar diagrams represent quantitation of the average number of dots per cell in the cytoplasm and nucleus of uninfected cells. (I and J) The above cells were also tested by PLA using anti-IFI16, BRCA1 and H2B antibodies. Boxed areas were enlarged and localization of IFI16 with BRCA1 and H2B with BRCA1 are indicated by the red arrows. Nuclei were stained with DAPI.

Fig 2. Immunoprecipitation and PLA analysis demonstrating the interaction and redistribution of IFI16 and H2B during KSHV de novo infection in HMVEC-d cells.

(A-D) Nuclear and cytoplasmic fractions from uninfected and KSHV (30 DNA copies/cell) infected HMVEC-d cells at various time points (2, 4, 12, 24 h p.i.) were IP-ed using anti-IFI16 and anti-H2B antibodies and immunoblotted for IFI16, H2B, BRCA1, H2A and ASC. (E and F) Nuclear and cytoplasmic fractions were analyzed by WB for input controls, while tubulin and TBP WB showed the purity of cytoplasmic and nuclear fractions, respectively. (G and H) PLA analysis for the interaction of IFI16 with H2B at various time points of KSHV post-infection in HMVEC-d cells. Uninfected and KSHV (30 DNA copies/cell) infected for 4 h (G) and for 2, 12 and 24 h p.i. (H) HMVEC-d cells were subjected to PLA reactions using anti-IFI16 (mouse) and anti-H2B (rabbit) antibodies. Boxed areas were enlarged and the nuclear and cytoplasmic localization of IFI16-H2B are indicated by red and yellow arrows, respectively. (I-K) PLA analysis for the association and redistribution of H2B-IFI16 during HSV-1 de novo infection. Uninfected and HSV-1 infected (1 pfu/cell) (4 h) HFF cells were subjected to PLA reactions using anti-H2B, anti-IFI16, and anti-BRCA1 antibodies. Boxed areas were enlarged and H2B-IFI16, H2B-BRCA1 and IFI16-BRCA1 localization in the nucleus and cytoplasm are indicated by red and yellow arrows, respectively. Nuclei were stained with DAPI. Magnification: 40X.

Fig 3. Demonstration of KSHV latent infection induced IFI16-H2B interaction and redistribution.

(A and B) Cytoplasmic and nuclear extracts from BJAB and BCBL-1 cells were IP-ed using anti-IFI16 and anti-H2B antibodies, and immunoblotted with anti-IFI16, H2B, BRCA1, H2A and ASC antibodies. (C) Cytoplasmic and nuclear fractions were analyzed by WB for equal inputs, while tubulin and TBP confirmed the purity of cytoplasmic and nuclear extracts, respectively. (D-G) BJAB and BCBL-1 cells were tested by PLA using anti-IFI16 (mouse), anti-H2B (rabbit) and anti-H2A (rabbit) antibodies. Nuclear and cytoplasmic localization of IFI16-H2B and IFI16-H2A are indicated by white and red arrows, respectively (D and F). Nuclei were stained with DAPI. The average number of dots per cell in the nucleus and cytoplasm was quantitated and represented by bar graph (E and G). *, p<0.05; **, p<0.01; ***, p<0.001 nuclear vs. cytoplasmic dots.

Fig 4. Acetylation of H2B and IFI16 during KSHV de novo infection.

(A) Untreated HMVEC-d cells (UT) or cells preincubated with or without 1 μM of C646 for 2 h were infected with KSHV for 2 h, washed and incubated again for 2 h in the presence or absence of C646 and subjected to PLA analysis using anti-H2B (goat) and anti-acetyl lysine (rabbit) antibodies. White and red arrows indicate the nuclear and cytoplasmic localization of H2B-acetyl lysine, respectively. (B) Cytoplasmic fractions from uninfected and KSHV (KS) infected (4 h) cells treated with or without C646 (panel A) were western blotted with anti-IFI16 and anti-H2B antibodies. (C and D) Transport of H2B from the nucleus to the cytoplasm through Ran GTPase association during KSHV de novo infection. (C) Cells described in panel A were lysed in NETN-lysis buffer and whole cell lysates (WCL) were IP-ed using anti-Ran antibodies and immunoblotted for IFI16, H2B and Ran. The bottom panel shows the input controls for IFI16 and H2B. (D) Uninfected or KSHV infected cells treated with or without C646 from panel A were subjected to PLA reactions using anti-H2B (goat) and anti-Ran (rabbit) antibodies. Nuclear and cytoplasmic localization of H2B-Ran are indicated by white and red arrows, respectively. (E and F) Leptomycin B (LPT) blocks the nuclear export of IFI16-H2B. HMVEC-d cells preincubated in the presence or absence of LPT (50 nM) were uninfected or infected by KSHV for 2 h, washed, and incubated for 2 h in the presence or absence of LPT followed by PLA reactions using anti-H2B (goat or rabbit) and anti-IFI16 (mouse) antibodies. Nuclear and cytoplasmic localization of H2B-H2B (E) and H2B-IFI16 (F) are indicated by white and red arrows, respectively. (G) Cytoplasmic fractions from the above described cells (panels E and F) were immunoblotted with anti-H2B and anti-IFI16. (H) Effect of ASC on the acetylation of IFI16 and H2B. Cytoplasmic fractions from HMVEC-d cells electroporated with siC (control siRNA) and siASC for 48 h, infected with/without KSHV (4 h) were IP-ed with anti-acetyl lysine antibody and western blotted for IFI16 and H2B. The bottom panel shows the ASC knockdown efficiency.

Fig 5. Demonstration of IFI16 and H2B association with STING during KSHV and HSV-1 de novo infection.

(A, B and C) Cytoplasmic fractions from uninfected and KSHV infected HMVEC-d cells at 2, 4, 12, and 24 h p.i. were taken from Fig 2F and IP-ed using anti-IFI16, H2B and STING antibodies and western blotted for IFI16, H2B, and STING. (D) The cytoplasmic fraction was immunoblotted for STING for input control and tubulin and TBP for purity of the fraction. (E and F) Uninfected and KSHV infected (4 h) HMVEC-d were tested by PLA using anti-IFI16, H2B and STING antibodies. Boxed areas were enlarged. Red arrows represent the localization of IFI16-STING and H2B-STING in the cytoplasm. (G and H) PLA analysis demonstrating the association of IFI16-STING and H2B-STING during HSV-1 de novo infection. Uninfected and HSV-1 infected (1 pfu/cell) (4 h) HFF cells were subjected to PLA reactions. Boxed areas were enlarged and localization of IFI16-STING and H2B-STING in the cytoplasm is represented by red arrows. Magnification: 40X. (I, J and K) IFI16 and H2B association with STING during KSHV latent infection. Cytoplasmic and nuclear extracts from BJAB and BCBL-1 cells shown under Fig 3C experiments were IP-ed using anti-IFI16, H2B and STING antibodies and western blotted for IFI16, H2B, and STING. (L) Equal input of STING analyzed by WB. (M and N) BJAB and BCBL-1 cells were tested by PLA reactions using anti-IFI16 (mouse), H2B (goat) and STING (rabbit) antibodies. Red arrows represent the localization of IFI16-STING and H2B-STING in the cytoplasm. Nuclei were stained with DAPI. Magnification: 40X. (O) Association of cGAS with IFI16, H2B, STING and BRCA1 during KSHV de novo infection. Cytoplasmic fractions from uninfected and KSHV infected HMVEC-d cells at 2, 4, 12, and 24 h p.i. were IP-ed using anti-cGAS antibody and western blotted for IFI16, H2B, STING, BRCA1 and cGAS. (P) Input control for cGAS in the cytoplasmic fraction by WB. (Q and R) Association of cGAS with IFI16, H2B, STING and BRCA1 during KSHV latent infection. (Q) Cytoplasmic and nuclear fractions from BJAB and BCBL-1 cells were IP-ed with anti-cGAS antibody and western blotted for IFI16, H2B, STING, BRCA1 and cGAS. (R) Western blot showing input control for cGAS in cytoplasmic and nuclear fractions.

Fig 6. Association of IFI16, H2B, BRCA1, cGAS and STING during KSHV infection.

(A) For a double PLA reaction, two independent reactions were performed. In the first PLA reactions, uninfected and KSHV (KS) infected (4 h) HMVEC-d cells were immunostained using mouse (ms) anti-IFI16 and goat (g) anti-H2B antibodies, and detected by DUOLink green detection agent. Cells were washed, blocked and subjected to the second PLA reactions using goat anti-H2B and rabbit (rb) anti-STING antibodies and visualized with red detection agents. Green and red dots indicate the localization of IFI16-H2B and H2B-STING, respectively. (B-E) Similarly, double PLA for different combinations were performed using BRCA1 (ms) + H2B (g) and H2B (g) + STING (rb); IFI16 (ms) + H2B (rb) and H2B (rb) + cGAS (g); BRCA1 (ms) + H2B (rb) and BRCA1 (ms) + cGAS (g); and H2B (ms) + STING (rb) and STING (rb) + cGAS (g). Red arrows indicate yellow dots representing the associations of three different proteins together such as IFI16-H2B-STING (A), BRCA1-H2B-STING (B), IFI16-H2B-cGAS (C), BRCA1-H2B-cGAS (D) and H2B-STING-cGAS (E) in the cytoplasm of KSHV infected cells. Nuclei were stained with DAPI. (F and G) Effect of BRCA1, cGAS and H2B on the interaction of IFI16-STING during KSHV de novo infection. (F) WCL from HMVEC-d cells electroporated with siC, siBRCA1, sicGAS and siH2B for 48 h then infected or uninfected with KSHV for 4 h were western blotted with anti-BRCA1, cGAS and H2B antibodies. (G) The WCL lysates from panel F were IP-ed with anti-IFI16 antibody and immunoblotted for STING and H2B. The bottom panels showed H2B knockdown efficiency and STING as an input control. (H and I) Role of ASC in the association of STING with IFI16 and H2B during KSHV de novo infection. WCL from HMVEC-d cells electroporated with siC or siASC for 48 h followed by with and without KSHV infections (4 and 24 h) were IP-ed with anti-STING antibody and immunoblotted for IFI16, H2B and STING. The bottom panels show the knockdown efficiency of ASC along with equal inputs of IFI16, H2B and STING by WB.

Fig 7. Effects of IFI16, H2B, BRCA1, cGAS, STING and ASC knockdown on IFN-β during KSHV and HSV-1 infection.

(A) WCL from HMVEC-d cells electroporated with siC, siH2B, siIFI16, siSTING or siASC for 48 h followed by infection for 4 h with or without KSHV (KS) were western blotted with anti-IFI16, ASC, H2B and STING antibodies. (B) Phosphorylation of TBK1 and IRF3 in H2B, IFI16, BRCA1, cGAS, STING and ASC knockdown cells infected with KSHV. HMVEC-d cells electroporated with different siRNA (panel A and Fig 6F) for 48 h were followed by with or without KSHV infection (4 h), and immunoblotted using anti-pTBK1, tTBK1 (total), pIRF3 and tIRF3 antibodies. (C and D) Effect of IFI16, H2B, BRCA1, cGAS, STING and ASC knockdown on IFN-β and IL-1β secretion during KSHV de novo infection. (C) Cell culture supernatants from HMVEC-d cells electroporated with siC, siIFI16, siBRCA1, sicGAS, siSTING, siH2B or siASC for 48 h followed by infection (4 h) with or without KSHV (Fig 7A and Fig 6F), were collected and subjected to IFN-β ELISA. Results presented are means ± SD (* p<0.05, ** p<0.01 from siC vs. siIFI16, siBRCA1, sicGAS, siSTING, and siH2B with KSHV infection; NS: not significant). (D) The same supernatants from panel C were used to detect IL-1β secretion by ELISA. Results presented are means ± SD (* p<0.05, from siC vs. siIFI16, and siASC with KSHV infection; NS: not significant). (E) Effect of IFI16, H2B, BRCA1, cGAS, STING and ASC knockdown on IFN-β secretion during HSV-1 infection. Cell culture supernatants from HFF cells electroporated with siRNA (S6F Fig) for 48 h followed by with or without HSV-1 infection (4 h) were collected and IFN-β ELISA was performed. Results presented are means ± SD (*** p<0.001 from siC vs. siIFI16, siBRCA1, sicGAS, siSTING, or siH2B with HSV-1 infection; NS: not significant).

Fig 8. Effects of H2B knockdown on cGAMP production, IFI16-cGAS association, IFI16 acetylation, and IFI16-ASC cytoplasmic redistribution during KSHV de novo infection.

(A) THP-1-Lucia ISG cells (1X10⁵) were treated with pure cGAMP (0–2 μg) to measure the activity of cGAMP used as a positive control. (B1 and B2) HMVEC-d cells electroporated with siC, siIFI16 and siH2B were infected with KSHV for 4 h. Cells were lysed, treated with benzonase for 30 min at 37°C, and heat inactivated at 95°C for 5 min. 10 μl of heat inactivated lysates were used to measure the cGAMP production using THP-1-Lucia ISG and THP-1-Dual KO-STING cells as described in the Materials and Methods. Recombinant human IFN-β at 1000 IU/ml was used as a positive control (B2). Results presented are means ± SD of three independent experiments. * p<0.05 and ** p<0.01 of siC vs. siIFI16 and siH2B with KSHV infection. (C, E and F) HMVEC-d cells electroporated with siC and siH2B followed by KSHV infection for 4 h were subjected to a PLA reaction using anti-IFI16, acetylated lysine, cGAS, and ASC antibodies. Nuclear and cytoplasmic spots are depicted by white and red arrows, respectively. (D, G and H) Average PLA spots of cGAS-IFI16, IFI16-ASC and IFI16-Acetyl lysine associations were quantitated and presented by bar graph. Results shown are means ± SD of three independent experiments. * p<0.05 and ** p<0.01 of siC vs. siH2B with KSHV infection; NS is not significant.

Fig 9. Detection of EdU labeled KSHV or HSV-1 genome associated host cell proteins by chromatin pull down during de novo infection.

(A) HMVEC-d cells were infected by EdU labeled or unlabeled KSHV genome (200 DNA copies/cell) for 2 h and (B) HFF cells were infected by EdU labeled or unlabeled HSV-1 genome (10 PFU/cell) for 2 h. Protein-DNA cross-linking was performed, and biotin-TEG azide selectively linked to the reactive alkyne group of EdU containing DNA through a click reaction. DNA was sheared and short chromatin fragments captured on streptavidin beads. Pulled down proteins were analyzed by immunoblot using anti-H2B and H3 antibodies. (C-F) H2B is critical in KSHV or HSV-1 genome recognition by IFI16 during de novo infection. (C) HMVEC-d cells electroporated with siC, siH2B or siBRCA1 for 48 h were infected by EdU-labeled KSHV (30 DNA copies/cell) for 2 h and tested by PLA using mouse and rabbit anti-IFI16 (green dots). EdU labeled KSHV genome was detected by reaction with Alexa 594 labeled picolylazide (red dots). The presence of IFI16 and EdU labeled KSHV genome association (yellow color) was indicated by white arrows. Red arrows in the left panel indicate the distribution of IFI16 in the cytoplasm. (D) Bar graph showing the quantitation of IFI16-EdU-KSHV genome colocalization (yellow dots). ** p<0.01 and *** p<0.001 that represents siC vs. siH2B and siBRCA1 with EdU-KSHV infection. (E) HFF cells electroporated using siC, siH2B or siBRCA1 for 48 h were infected by EdU-labeled HSV-1 (KOS) with 1 PFU/cell for 2 h and tested for PLA as described in panel (C). The presence of IFI16 and EdU labeled HSV-1 genome association (yellow dots) is indicated by white arrows. Nuclei were stained with DAPI. (F) Average colocalized yellow dots per cell were quantitated and presented by bar graph. ** p<0.01 and *** p<0.001 that presents siC vs. siH2B and siBRCA1 with EdU-HSV-1 infection. (G) HFF cells electroporated with siC, siBRCA1 or siH2B for 48 h were infected by EdU-labeled or unlabeled HSV-1 genome (10 PFU/cell) for 2 h. Protein-DNA cross-linking was performed as described in panel B. Pulled down protein was analyzed by immunoblot using anti-H2B and anti-IFI16 antibodies. The bottom panel shows the input controls for IFI16, H2B, BRCA1 and H3. (H) Effect of BRCA1 on KSHV genome recognition by IFI16-H2B during de novo infection. HMVEC-d cells electroporated with siC or siBRCA1 for 48 h were infected by EdU-labeled KSHV (30 DNA copies/cell) for 2 h and tested by PLA using mouse anti-IFI16 and rabbit anti-H2B (green dots). EdU labeled KSHV genome was detected by red dots. Boxed areas were enlarged and the presence of IFI16-H2B and EdU labeled KSHV genome association (yellow color) was indicated by white arrows (bottom panel). Red arrows in the top panel indicate the redistribution of IFI16-H2B in the cytoplasm. (I) The average number of colocalized yellow PLA dots per cell as shown in Fig 9H was quantitated and presented in the bar graphs. **p<0.01 from siC vs. siBRCA1 with EdU KSHV infection.

Fig 10. Effect of BRCA1 on IFI16 and H2B acetylation during KSHV de novo infection.

(A) HMVEC-d cells electroporated with siC or siBRCA1 for 48 h were uninfected or infected by KSHV for 4 h and subjected to PLA using mouse anti-IFI16 and rabbit anti-acetyl lysine antibodies. The localization of IFI16-acetyl lysine (acetylation of IFI16) PLA spots are shown in red. (B) WCL from HMVEC-d cells electroporated with siC and siBRCA1 for 48 h followed by with or without KSHV infection for 30’ (30 min), 4 and 24 h were IP-ed using anti-IFI16 and anti-acetyl lysine antibodies and western blotted for IFI16, BRCA1, p300, and H2B. Bottom input control panels show the levels of BRCA1, IFI16, p300 and total protein acetylation. (C) Schematic model illustrating the essential role of histone H2B in IFI16-mediated viral DNA genome sensing and innate IFN-β production during KSHV or HSV-1 de novo infection. Soon after KSHV or HSV-1 DNA entry into the nucleus, IFI16 in complex with BRCA1-H2B or with BRCA1 recognizes the viral genome, leading into BRCA1 mediated p300 recruitment, interaction with IFI16, acetylation (Ac) of IFI16 and H2B by p300, and cytoplasmic transport of acetylated IFI16-H2B-BRCA1 via Ran GTP. C646 or LPT treatments abolish acetylation or cytoplasmic translocation of H2B and IFI16, respectively. The inflammasome (IFI16-BRCA1-ASC-procaspase-1) independent IFI16-H2B-BRCA1 complex in the cytoplasm associates with cGAS and STING to form a macromolecular complex leading into phosphorylation of TBK1 and IRF3, nuclear translocation of p-IRF3 and subsequent IFN-β production during KSHV or HSV-1 de novo infection. Viral genome recognition by IFI16-BRCA1 results in acetylation of IFI16 which interacts with ASC and procaspase-1 in the nucleus. This BRCA1-IFI16-ASC-procaspase-1 complex is transported to the cytoplasm via Ran-GTP resulting in caspase-1 formation, pro-IL-1β cleavage and IL-1β formation [1, 9]. These studies demonstrate that the innate nuclear foreign viral genome sensing responses are mediated by two IFI16 complexes, with recognition by the IFI16-H2B-BRCA1 complex resulting in IFN-β responses while recognition by the IFI16-BRCA1 complex results in inflammasome-IL-1β responses.