UNC93B1 is essential for TLR11 activation and IL-12-dependent host resistance to Toxoplasma gondii

Abstract

Toll-like receptor (TLR) activation relies on biochemical recognition of microbial molecules and localization of the TLR within specific cellular compartments. Cell surface TLRs largely recognize bacterial membrane components, and intracellular TLRs are exclusively involved in sensing nucleic acids. Here we show that TLR11, an innate sensor for the Toxoplasma protein profilin, is an intracellular receptor that resides in the endoplasmic reticulum. The 12 membrane-spanning endoplasmic reticulum-resident protein UNC93B1 interacts directly with TLR11 and regulates the activation of dendritic cells in response to Toxoplasma gondii profilin and parasitic infection in vivo. A deficiency in functional UNC93B1 protein abolished TLR11-dependent IL-12 secretion by dendritic cells, attenuated Th1 responses against T. gondii, and dramatically enhanced susceptibility to the parasite. Our results reveal that the association with UNC93B1 and the intracellular localization of TLRs are not unique features of nucleic acid-sensing TLRs but is also essential for TLR11-dependent recognition of T. gondii profilin and for host protection against this parasite.

FIGURE 1.

TLR11 localizes intracellularly. A and B, HEK293 cells were plated on glass coverslips and co-transfected with TLR11-GFP (A) and the ER marker DsRed2-ER (B). Twenty-four hours later, the cells were stained with the cell-permeable fluorescent dye SYTO 62 (artificially colored in blue). C, representative co-localization image from six experiments is shown. Images were acquired using a Leica SPE microscope equipped with a 63× objective. D, RAW264.7 cells were transfected with TLR11-GFP, and their cell surface was labeled with cholera toxin subunit B-Alexa Fluor 647 conjugate (blue). E and F, RAW264.7 cells were transfected with TLR11-GFP as described above and were left untreated (E) or stimulated with 1 μg/ml T. gondii profilin for 1 h in the presence of the lysosomal marker LysoTracker Red DND-99 (F) (red). G, splenic DCs were pretreated for 1 h with the indicated concentrations of chloroquine or bafilomycin A and stimulated with 100 ng/ml T. gondii profilin (filled bars) or 1 μm CpG (open bars). IL-12/23p40 production was assayed by ELISA 20 h after stimulation. H, splenic DCs were activated by LPS (1 μg/ml) alone or in the presence of increasing amounts of chloroquine (filled bars) or bafilomycin A (open bars), and IL-12/23p40 secretion was measured as described above. A representative result from five experiments is shown. Error bars, S.D.

FIGURE 2.

Infection with live tachyzoites is not required for TLR11-mediated IL-12 production from DC. A, sort-purified CD11c⁺ splenic DCs were incubated with YFP-expressing tachyzoites alone or in the presence of the indicated concentrations of cytochalasin D for 20 h. The percentage of infected cells (CD11c⁺ YFP⁺) was assayed by flow cytometry gated on DCs. B, amount of IL-12/23p40 present in supernatants of DCs co-cultured with tachyzoites (as described in A) was measured by ELISA. The data shown are representative of six experiments. C, CD11c⁺ splenic DCs were incubated with YFP-expressing tachyzoites (2:1 ratio) for 20 h, and brefeldin A was added during the last 6 h of the culture. IL-12 production was analyzed by intracellular cytokine staining. As a positive control, DCs were stimulated with recombinant T. gondii profilin (1 μg/ml) as described above. D, mice were infected intraperitoneally with 10⁵ YFP-expressing tachyzoites. Twenty hours later spleen was isolated and the percentage of IL-12 producing and infected cells (YFP⁺) was analyzed by intracellular cytokine staining after incubation of splenocytes for 6 h in the presence of brefeldin A. E, splenocytes depleted of DCs (Sp) were placed in the bottom of a Transwell plate (separating membrane: 0.4-μm pores) and stimulated with live parental profilin-sufficient strain tachyzoites (Toxo, ΔTgPRFe/TgPRFi strain), recombinant profilin (1 μg/ml), or profilin-deficient strain tachyzoites (TgPRF KO Toxo, ΔTgPRFe/TgPRFi strain+ATc) in the presence or absence of DCs (upper portion of Transwell plate). As a positive control, DCs were cultivated with tachyzoites or recombinant profilin without a separating membrane. IL-12/23p40 present in supernatants was measured by ELISA. The data shown are representative of three performed experiments. Error bars, S.D.

FIGURE 3.

UNC93B1 associates and co-localizes with TLR11. A, HEK293 cells were co-transfected with empty vector, TLR9-GFP, TLR11-GFP, or TLR4-GFP together with Cherry-tagged UNC93B. Thirty-six hours after transfection, the cell lysates were immunoprecipitated (IP) with anti-GFP or anti-mCherry antibodies. Equal amounts of eluted protein were then analyzed by immunoblotting (IB) with anti-mCherry or anti-GFP antibodies to detect UNC93B1 or TLRs, respectively; ns denotes nonspecific bands. B–D, HEK293 cells were co-transfected with UNC93B1-mCherry (B) and TLR11-GFP (C), and co-localization (D) was analyzed by confocal microscopy. A representative result from five experiments is shown.

FIGURE 4.

UNC93B1 regulates TLR11-dependent DC cytokine responses to T. gondii profilin. WT (gray bars) and 3d (open bars) splenic DCs were stimulated with the indicated concentrations of T. gondii profilin (left panels), 1 μm CpG, 1 μg/ml LPS, or no treatment (right panels). IL-12/23p40 (A), TNF (B), and IL-6 (C) levels were assayed by ELISA 20 h after stimulation. A representative result from four experiments is shown. Error bars, S.D.

FIGURE 5.

UNC93B1 is essential for host resistance to T. gondii infection. A, WT, 3d, Myd88^−/−, and Tlr11^−/− mice (n = 5 for all groups except MyD88^−/−, n = 3) were injected intraperitoneally with 10 μg of T. gondii profilin and bled at the indicated time points. IL-12/23p40 responses were measured by ELISA. B, WT, 3d, Myd88^−/−, and Tlr11^−/− mice were infected with an average of 20 cysts/mouse of the ME49 strain of T. gondii, and serum IL-12/23p40 responses were measured 3 days (open bars) and 5 days (filled bars) later by ELISA. The data represent the mean ± S.D. (error bars) and are representative of four experiments. C, single-cell analysis of IFN-γ secretion by CD4⁺ and CD8⁺ T lymphocytes is shown. WT and 3d mice were infected with an average of 20 cysts/mouse of the ME49 strain of T. gondii, and IFN-γ production by T cells was analyzed on day 7 after infection. The data shown are representative of three experiments. D, WT (filled circles, n = 15) and 3d (filled squares, n = 14) mice were infected with 20 cysts/mouse of the ME49 strain, and their survival was monitored daily for up to 6 weeks after infection. In some conditions, 3d mice were treated with 150 ng/mouse IL-12p70 (filled triangles) or PBS (open triangles) during the 1st week of infection. The results from four separate experiments were pooled.