The small ubiquitin-like modifier-deconjugating enzyme sentrin-specific peptidase 1 switches IFN regulatory factor 8 from a repressor to an activator during macrophage activation

Abstract

Macrophages, when activated by IFN-γ and TLR signaling, elicit innate immune responses. IFN regulatory factor 8 (IRF8) is a transcription factor that facilitates macrophage activation and innate immunity. We show that, in resting macrophages, some IRF8 is conjugated to small ubiquitin-like modifiers (SUMO) 2/3 through the lysine residue 310. SUMO3-conjugated IRF8 failed to induce IL12p40 and other IRF8 target genes, consistent with SUMO-mediated transcriptional repression reported for other transcription factors. SUMO3-conjugated IRF8 showed reduced mobility in live nuclei and bound poorly to the IL12p40 gene. However, macrophage activation caused a sharp reduction in the amount of SUMOylated IRF8. This reduction coincided with the induction of a deSUMOylating enzyme, sentrin-specific peptidase 1 (SENP1), in activated macrophages. In transfection analysis, SENP1 removed SUMO3 from IRF8 and enhanced expression of IL12p40 and other target genes. Conversely, SENP1 knockdown repressed IRF8 target gene expression. In parallel with IRF8 deSUMOylation, macrophage activation led to the induction of proteins active in the SUMO pathway and caused a global shift in nuclear protein SUMOylation patterns. Together, the IRF8 SUMO conjugation/deconjugation switch is part of a larger transition in SUMO modifications that takes place upon macrophage activation, serving as a mechanism to trigger innate immune responses.

Figure 1

IRF8 SUMOylation status in resting and activated macrophages. (A) Left panels, Nuclear extracts (500 μg) from unstimulated RAW cells or stimulated with IFN-γ (150 U/ml) for 8 h or IFN-γ overnight, followed by CpG (150 ng/ml) for 6 h, were immunoprecipitated with Ab for IRF8 and blotted against SUMO2/3 or IRF8. Extracts were blotted for TFIIB for loading control. Right panels, Unstimulated RAW cells (0) or stimulated with IFN-γ overnight, followed by CpG for indicated times and nuclear extracts, were tested, as above. TFIIB was tested as a loading control. (B) 293T cells (1 × 10⁶) were transfected with Flag-IRF8 (1 μg), V5-SUMO1, SUMO2, or SUMO3 for 30 h, and extracts were precipitated with anti-Flag Ab and blotted with anti-V5 or anti-Flag Ab. The lower middle panel depicts results of a longer exposure revealing multiple SUMOylated IRF8 bands. The bottom panel indicates immunoblot of WCE with anti-V5 Ab. (C) NIH3T3 cells (1 × 10⁶) were transfected with 1 μg CFP-IRF8 and YFP-SUMO3 or SUMO3GA for 30 h and tested for a real-time interaction of IRF8 with SUMO3 by FRET analysis. Only SUMO3, not SUMO3GA, gave FRET signals (upper panel). In the lower panel, FRET signals were quantified by scoring fluorescent intensity of ∼200 cells, according to the modified Youvan method.

Figure 2

Identification of a SUMOylation site in IRF8. The 293T cells transfected with Flag-IRF8 mutants (K72R or K310R) and T7-SUMO1 or V5-SUMO3 were immunoprecipitated with anti-Flag Ab and blotted with anti-T7 and anti-V5 Ab (upper middle panel) or anti-Flag Ab (lower middle panel). Vector alone and wt Flag-IRF8 (wt) were tested as controls. The bottom panel indicates immunoblot analysis of WCE.

Figure 3

SUMO3 conjugation represses transactivation of IRF8 target genes. (A) RAW cells (1 × 10⁶) were transfected with 600 ng indicated luciferase reporters, 800 ng pcDNA-IRF8 vectors, and 60 ng pRL-TK vector for 30 h and stimulated with IFN-γ overnight and CpG for 6 h. Reporter activity was normalized by Renilla luciferase activity. Values represent the average of three assays ± SD. Statistical significance was tested by Student t test; *p < 0.01, **p < 0.005. (B) IRF8^−/− macrophages (CL2) were transduced with pMSCV vectors for wt IRF8, K310R mutant or SUMO3-IRF8, or IRF8-SUMO3 for 5 d and stimulated with (filled bars) or without (open bars) IFN-γ and CpG. Levels of Il12b and Ccl9 transcripts were measured by qRT-PCR. Values represent the average of three determinations ± SD. Statistical significance was tested by Student t test; *p < 0.01, **p < 0.005. (C) IRF8^−/− DCs were transduced with the above vectors for 5 d and stimulated with (filled bars) or without (open bars) CpG for 7 h and Il12b and Ifnb transcript levels were measured, as above. Statistical significance was tested by Student t test; **p < 0.005. (D) IRF8^−/− DCs transduced with indicated vectors were stimulated with or without CpG for 36 h, and expression of MHCII (I-A^b) was detected by flow cytometry. The gray lines indicate signals by control IgG.

Figure 4

SUMO3 conjugation reduces genome-wide mobility of IRF8 and target-binding activity. (A) NIH3T3 cells were transduced with pMSCV vectors for GFP alone, GFP-SUMO3, wt GFP-IRF8, or GFP-IRF8-SUMO3, and localization of each construct was viewed by live cell imaging. (B) In FRAP analysis, a small region within the nucleus was bleached at indicated time, and the recovery of fluorescence (FL) signals was plotted against time (sec). The recovery curves represent the average of 15 separate cells. See reduced recovery of GFP-IRF8-SUMO1 in Supplemental Fig. 3. (C) CL2 cells were transduced with indicated pMSCV vectors and stimulated with IFN-γ or IFN-γ/CpG. Binding of these constructs to the indicated regions of the Il12b gene was tested by ChIP using anti-GFP Ab. Values represent the average of three determinations ± SD. The bottom diagram indicates the exon-intron organization of Il12b. The arrow marks the TSS.

Figure 5

Macrophage activation triggers a global change in SUMOylated nuclear proteins, and the SENP1 enhances IRF8 transactivation. (A) RAW cells were stimulated with IFN-γ overnight, followed by stimulation by Newcastle disease virus, LPS, or CpG for indicated time. A total of 30 μg nuclear extracts was resolved in 4–12% Nu PAGE and blotted with anti-SUMO2/3 Ab. 0, indicates no treatment. (B) 293T cells were cotransfected with 2 μg pcDNA vector for Flag-IRF8 and 2 μg HA-Senp1 and 500 μg lysates were immunoprecipitated with anti-Flag Ab and blotted with anti-HA Ab (upper panel) or Flag (middle panel). In the lower panel, WCE (20 μg) were blotted with Ab for Flag. (C) 293T cells were transfected with 0.5 μg pcDNA vector for V5-SUMO3, 1 μg Flag-IRF8, and 1 μg HA-SENP1 or HA-C599S. A total of 500 μg lysates was immunoprecipitated with anti-Flag Ab and blotted with anti-V5 or anti-Flag Ab. In the two lower panels, WCE (20 μg) were immunoblotted with indicated Ab. (D) RAW cells were transfected with luciferase reporters for the Il12b or IFNb promoters, increasing amounts of pcDNA vectors for SENP1 or C599S (200, 400, 800 ng), IRF8 (800 ng), and pRL-TK and stimulated with IFN-γ/CpG, and luciferase activity was measured, as above. The values represent the average of three determinations ± SD. (E) CL2 macrophages were transduced with pMSCV vectors for IRF8 and SENP1 or the C599S mutant for 5 d and stimulated with or without IFN-γ/CpG. Expression of indicated transcripts was tested by qRT-PCR, as above. (F) IRF8^−/− BM DCs (10⁶ cells) were transduced with indicated vectors for 5 d and stimulated with CpG for 4 h, and expression of indicated transcripts was tested, as above. The values represent the average of three determinations ± SD. Statistical significance was tested by Student t test; *p < 0.01, **p < 0.005.

Figure 6

SENP1 knockdown inhibits IRF8 transactivation. (A) CL2 cells were transduced with pMSCV vector for wt IRF8, along with empty vector, control shRNA, or Senp1 shRNA pSuper vector, and cultured for 5 d, followed by stimulation with IFN-γ/CpG. Expression of indicated genes was measured by qRT-PCR, as above. (B) IRF8^−/− BM DCs (10⁶ cells) were transduced with viral vectors, as above, and stimulated with CpG for 4 h. The values represent the average of three determinations ± SD. Statistical significance was tested by Student t test; **p < 0.005. (C) SUMO conjugation-deconjugation switches IRF8 function: a model. SUMO-conjugated IRF8 present in resting macrophages acts as a repressor. Activation of macrophages with IFN-γ or IFN-γ/TLR stimulates SENP1 expression. SENP1 then removes SUMO from IRF8, converting it to a transcriptional activator. This SUMO switch is critical for the induction of IRF8-regulated genes.