The interaction between the ER membrane protein UNC93B and TLR3, 7, and 9 is crucial for TLR signaling

Abstract

Toll-like receptors (TLRs) sense the presence of microbial and viral pathogens by signal transduction mechanisms that remain to be fully elucidated. A single point mutation (H412R) in the polytopic endoplasmic reticulum (ER)-resident membrane protein UNC93B abolishes signaling via TLR3, 7, and 9. We show that UNC93B specifically interacts with TLR3, 7, 9, and 13, whereas introduction of the point mutation H412R in UNC93B abolishes their interactions. We establish the physical interaction of the intracellular TLRs with UNC93B in splenocytes and bone marrow-derived dendritic cells. Further, by expressing chimeric TLRs, we show that TLR3 and 9 bind to UNC93B via their transmembrane domains. We propose that a physical association between UNC93B and TLRs in the ER is essential for proper TLR signaling.

Figure 1.

Characterization of wild-type and mutant UNC93B proteins. (A) Model of the UNC93B protein. The single point mutation of histidine residue 412 to arginine (H412R, 3d mutation) is located within transmembrane domain 9 (•). The two predicted N-linked glycosylation sites (N²⁵¹HT and N²⁷²KT) are indicated (*). Wild-type and mutant (H412R) UNC93B were fused at the C terminus with a Flag tag, followed by the TEV protease cleavage site and an HA-tag (designated as UNC93B-HA). (B) RAW macrophages stably expressing epitope-tagged wild-type UNC93B-HA (WT, left) or mutant UNC93B-HA (H412R, right) were metabolically labeled with [³⁵S]methionine/cysteine for 30 min (pulse) and lysed in RIPA buffer after 0, 2, 4, 8, and 12 h of incubation in normal medium (chase). UNC93B proteins were recovered by immunoprecipitation with an anti-HA antibody and resolved by SDS-PAGE. Wild-type and mutant UNC93B proteins show similar stability, but differ in their migration patterns; wild-type UNC93B migrates as heterodispersed material on SDS-PAGE, whereas the UNC93B mutant form migrates as a well-defined distinct polypeptide in addition to more diffuse material. (C) Endogenous and epitope-tagged UNC93B proteins were immunoprecipitated in 1% NP-40 lysis buffer from ³⁵S-labeled RAW macrophages that are nontransduced (–) or transduced with UNC93B-HA (WT) or UNC93B-HA (H412R) using a polyclonal antibody directed against the C terminus of UNC93B (anti–UNC-C). Reimmunoprecipitations were performed after mild denaturation of the initial immunoprecipitation with a polyclonal antiserum against the N terminus of UNC93B (anti–UNC-N) or antibodies to the HA or Flag epitopes. Recovered epitope-tagged and endogenous UNC93B proteins were resolved by SDS-PAGE. The distinct polypeptide observed in direct immunoprecipitations of mutant UNC93B is also observed after reimmunoprecipitation of mutant UNC93B, and is thus UNC93B itself. (D) BM-DCs from either wild-type C57BL/6 (B6) or UNC93B mutant (3d) mice were pulsed with [³⁵S]methionine/cysteine for 40 min and lysed in 1% digitonin lysis buffer after 0, 1, or 4 h of incubation in normal medium (chase). Endogenous UNC93B proteins were immunoprecipitated with the affinity purified anti–UNC-C antibody, subjected to digestion with glycosidases Endo H (H) or PNGase F (F) or not digested (–), and resolved by SDS-PAGE. Both wild-type and mutant UNC93B proteins retain full sensitivity to Endo H digestion. The polypeptides migrating at the range of 130–150 kD in nondigested samples and at ∼100 kD in Endo H/PNGase F-digested samples (arrowheads) are only present in wild-type samples, but not in the 3d samples (Fig. 3 A and Fig. 6 A).

Figure 2.

Maturation of MHC class II is not altered in UNC93B mutant mice. BM-DCs from either wild-type C57BL/6 (B6) or UNC93B mutant (3d) mice were pulsed with [³⁵S]methionine/cysteine for 40 min and lysed in 1% digitonin lysis buffer after 0, 1, 2, or 4 h of incubation in normal medium (chase). Endogenous MHC class II was immunoprecipitated with the anti-MHC class II antibody (N22). Immunoprecipitates were loaded under denaturing (boiled) or mildly denaturing (nonboiled) conditions and resolved by SDS-PAGE. SDS-stable class II MHCs are composed of the α/β heterodimer and antigenic peptide (αβpeptide). Ii, invariant chain.

Figure 3.

TLR3, 7, 9, and 13 bind to wild-type, but not mutant, UNC93B. (A) RAW macrophages (left) or RAW cells stably transduced with either mutant (middle) or wild-type (right) UNC93B-HA were labeled with [³⁵S]methionine/cysteine for 30 min. Cells were lysed in 1% digitonin lysis buffer after chase periods of 0, 90, or 180 min, and UNC93B-HA was immunoprecipitated with an anti-HA antibody. The samples were incubated with Endo H (H) or PNGase F (F) and resolved by SDS-PAGE. The polypeptide (∼130 kD in its glycosylated form; ∼100 kD in its deglycosylated form) that coimmunoprecipitated with wild-type UNC93B is indicated with arrowheads. (B) 4 billion RAW cells expressing wild-type (WT) or mutant (H412R) UNC93B-HA were lysed in 1% digitonin buffer and UNC93B-HA proteins were immunoprecipitated with an anti-HA antibody. After immunoprecipitation, UNC93B and UNC93B-associated proteins were released by incubation with TEV protease, resolved by SDS-PAGE, and visualized by silver staining. Polypeptides were excised from the gel and analyzed by LC/MS/MS after trypsin digestion. Peptides for UNC93B (black) were recovered in both samples, whereas peptides corresponding to sequences of TLR3, 7, 9, and 13 (gray) were identified only in the sample containing wild-type UNC93B. Peptide sequences identified by MS for UNC93B and TLRs are given in Tables S1 and S2, respectively. Tables S1 and S2 are available at http://www.jcb.org/cgi/content/full/jcb.200612056/DC1.

Figure 4.

Wild-type UNC93B associates with TLR3 and 9, but not with TLR4. HEK 293-T cells were cotransfected with empty vector or expression constructs for wild-type or mutant UNC93B-HA along with expression constructs for myc-tagged TLR3, 4, or 9 (Fig. 5 A). Cells were lysed with 1% digitonin lysis buffer and TLRs were immunoprecipitated with an anti-myc antibody. Samples were resolved by SDS-PAGE, and UNC93B proteins were detected by immunoblotting using an anti-HA antibody (top). Wild-type UNC93B was recovered by immunoprecipitation of TLR3 and 9, whereas the mutant UNC93B (H412R) did not coimmunoprecipitate with any of TLRs. Input lysates were analyzed by SDS-PAGE for expression levels of wild-type and mutant UNC93B-HA with an anti-HA antibody (bottom).

Figure 5.

TLR3 and 9 interact with UNC93B via their transmembrane segments. (A) Schematic presentation of myc-tagged TLR expression constructs used in this study. The myc-tag was fused to the C terminus of the TLRs, and it is indicated as a white rectangle. The TLR chimeras were generated such that the transmembrane segments were exchanged between TLR3, 4, and 9 to yield TLR3 and 9 with the transmembrane segment of TLR4 (TLR3-4-3 and TLR9-4-9), and TLR4 with the transmembrane segment of either TLR3 (TLR4-3-4) or TLR9 (TLR4-9-4). Binding capabilities of wild-type UNC93B to TLRs as shown in this study are indicated (–, no binding; +, binding). (B) HEK 293-T cells were cotransfected with an empty vector (–) or an expression construct for wild-type UNC93B-HA together with myc-tagged TLR3, 4 or 9 or the myc-tagged TLR chimeras TLR4-9-4, 9-4-9, 4-3-4, 3-4-3. Cells were metabolically labeled for 4 h with [³⁵S]methionine/cysteine and lysed in 1% digitonin lysis buffer. One fifth of the lysate was subjected to immunoprecipitation with an anti-myc antibody to assess the expression of the TLR constructs (middle). The rest of the lysate was subjected to immunoprecipitation with an anti-HA antibody and one tenth of the resulting UNC93B-HA immunoprecipitates was directly resolved by SDS-PAGE to show UNC93B-HA expression levels (bottom). The remaining UNC93B-HA immunoprecipitates were subjected to denaturation, and reimmunoprecipitations were performed with TLR-specific antibodies (anti-TLR3 for TLR3, TLR3-4-3, untransfected; anti-TLR4 for TLR4, TLR4-3-4, TLR4-9-4; anti-TLR9 for TLR9, TLR9-4-9) and resolved by SDS-PAGE (top). TLR3 and 9 were recovered by reimmunoprecipitation from UNC93B-HA immunoprecipitates, as well as the TLR chimeras TLR4-3-4 and TLR4-9-4, but not TLR4 and chimeras TLR9-4-9 and TLR3-4-3.

Figure 6.

Endogenous UNC93B and TLR7 associate in BM-DCs from wild-type, but not from UNC93B mutant mice. (A–C) Day 5 cultures of BM-DCs prepared from wild-type (C57BL/6, B6), UNC93B mutant (3d) or TLR7^−/− mice were metabolically labeled for 4 h with [³⁵S]methionine/cysteine and lysed in 1% digitonin lysis buffer. (A) The first immunoprecipitation was performed with the purified anti–UNC-C antibody (left). After mild denaturation of the initial immunoprecipitates, reimmunoprecipitations were performed with either an anti-TLR7 antibody (middle) or the anti–UNC-C antiserum (right). TLR7 was recovered with wild-type, but not with mutant UNC93B. Expression levels of wild-type and mutant UNC93B were comparable. (B) TLR7 expression levels in wild-type (B6), UNC93B mutant (3d), and TLR7 knockout (TLR7^−/−) mice were analyzed by direct immunoprecipitation of TLR7 with the TLR7 antibody. Endogenous TLR7 is present at equal levels in both wild-type and UNC93B mutant 3d mice, but absent in TLR7 knockout mice. (C) TLR7 was immunoprecipitated with the TLR7 antibody from digitonin lysates of BM-DCs from wild-type (B6) or UNC93B mutant (3d) mice. Immunoprecipitation with normal rabbit serum (NRS) served as a control for TLR7 immunoprecipitation. Reimmunoprecipitations with anti–UNC-C or anti-TLR7 antibodies were performed after mild denaturation of the initial TLR7 immunoprecipitates and resolved by SDS-PAGE. UNC93B was coimmunoprecipated with TLR7 in BM-DC lysates from wild-type, but not from UNC93B mutant, mice. Recovery of TLR7 from BM-DCs of both wild-type and mutant mice by immunoprecipitation and reimmunoprecipitation was equal. (D) Nontransduced A20 B cells (–) or A20 B cells stably transduced with wild-type (WT) or mutant UNC93B-HA (H412R) were metabolically labeled for 4 h with [³⁵S]methionine/cysteine (pulse) and stimulated with TLR7 agonists imiquimod (10 μM) or gardiquimod (1 μM) or the TLR9 agonist CpG DNA (1 μM) for 1 h during the final hour of the pulse. Cells were lysed in 1% digitonin lysis buffer, and immunoprecipitations were performed with an anti-HA antibody. Samples were not treated (–) or digested with Endo H (+) and resolved by SDS-PAGE. The polypeptides that associate only with wild-type UNC93B and have characteristics of TLRs are indicated by arrowheads. Interaction of these polypeptides with wild-type UNC93B was not considerably changed in stimulated cells compared with nonstimulated cells.