Fig 1.
TLR9 tyrosine 870 is conserved across TLRs and their adaptors.
Mouse TLRs, except TLR1, TLR6, and TLR12, have a tyrosine in the box 1 region of the TIR domain. N-terminal and C-terminal amino acid positions are noted to the left and to the right respectively. Note the aspartic acid residues in Mal that are underlined were found to be potentially involved in the dimerization surface in the crystal structure [15].
Fig 2.
Expression of HA-tagged TLR9 mutants in 293 T cells and BMDCs.
(A) 293T cells were transfected with 4μg of HA-tagged WT TLR9, Y870F, or Y870A. Cells were untreated (-) or treated for 18 h with 1 μg/ml CpG, and whole cell lysates were fractionated by SDS-PAGE and analyzed by Western blot for HA (TLR9) or Erk as a loading control. Blots were developed using Odyssey and analyzed by Licor software. The data shown are representative of at least 4 independent experiments. (B) Tlr9-/- bone marrow was transduced with retrovirus expressing HA-tagged WT TLR9, Y870F, or Y870A. BMDCs were harvested, permeabilized and stained with antibodies against CD11c and HA. HA expression (right panel) was determined for GFP+CD11c+ cells (left panels) by flow cytometry. Histograms are representative of at least 5 independent experiments. (C) Tlr9-/- BMDCs expressing WT TLR9, Y870F, or Y870A were left unstimulated (Untrx) or stimulated for 3 hours with CpG (1μg/ml). Supernatants were harvested, and TNFα and IL-6 were measured by ELISA. Bar graphs are representative of at least 4 independent experiments performed in triplicate. * indicates p-value < 0.05 by Bonferroni post-test.
Fig 3.
Expression of TLR9 Y870A inhibits ligand induced cytokine secretion by endogenous TLR9 in a dominant-negative manner.
(A) 293T cells were co-transfected with HA- and GFP-tagged forms of TLR9 as WT (WT) or Y870A (A) variants as indicated. HA-tagged TLR9 was immunoprecipitated using an antibody to HA, and interaction with GFP-tagged TLR9 was assessed by immunobloting for both HA and GFP in both the lysates (input) and the immunoprecipitates (IP: HA). Immunoblotting for Erk served as a control. Results are representative of at least 3 independent experiments. (B) WT (i.e., TLR9 sufficient) C57BL/6 bone marrow was transduced with retrovirus expressing HA-tagged WT or Y870A TLR9 or empty vector as a control, and then differentiated to DCs. BMDCs were stimulated for 3 hours with CpG, and then supernatants were harvested and TNFα and IL-6 were measured by ELISA. Bar graphs are representative of at least 3 independent experiments each performed in triplicates. * indicates p-value < 0.05 by Bonferroni post-test. (C) These same transduced BMDCs were stimulated with CpG for various times, and lysates were subjected to Western blot analysis for the presence of HA-tagged TLR9, phosho-Erk or total Erk. Blots are representative of 3 experiments. (D) Quantification from a representative experiment of phospho-Erk normalized to total Erk.
Fig 4.
Mutagenesis of TLR9 Tyr870 results in impaired receptor maturation.
(A, B) Tlr9-/- BMDCs expressing HA-tagged TLR9 WT, Y870F, or Y870A were untreated (-) or treated for 18 h with 1 mg/ml CpG. (A) Cells were lysed and analyzed by Western blotting for HA expression relative to Erk as a loading control. A representative blot from 3 experiments is shown at the left. Right, the 160 kDa and 80 kDa bands were quantified, and the amount of active receptor was calculated by normalizing the amount of processed receptor (80 kDa) to the amount of unprocessed receptor (160 kDa). Shown is the average from 3 experiments. (B) CpG-induced TNFα and IL-6 levels from WT or mutant TLR9-expressing BMDC cultures were normalized to the amounts of processed (active) receptor as determined in Fig 4A. N.D. indicates not determined, as there is no mature Y870A receptor to which cytokine can be normalized. Note, CpG-induced cytokine levels in Y870A-expressing cells were essentially no different from unstimulated samples, as in Fig 2C. * indicates p-value < 0.05 by Bonferroni post-test. (C) WT bone marrow was uninfected (left) or transduced with either WT or Y870A TLR9, and then differentiated to BMDCs. Cells were stimulated for 0, 60 or 180 min with CpG as in Fig 3B, and cleavage of overexpressed TLR9 to the 80kD form in these cells was assessed by immunoblotting for HA (relative to total Erk as a loading control) as in (A). Shown is a representative of 3 experiments.
Fig 5.
TLR9 Y870A interacts with ER resident proteins.
293T cells were transfected with 4μg each of FLAG-tagged UNC93B1 and GFP-tagged WT or Y870A TLR9 or a GFP vector control. Lysates were harvested and subjected to immunoprecipitation with anti-GFP-conjugated agarose beads. Aliquots of the lysate (input) and immunoprecipitates (IP: GFP) were analyzed by Western blotting for GFP (to detect ~160kDa TLR9 transgene or ~30 kDa GFP control), FLAG (to detect UNC93B1 transgene), GRP94 (to detect associated endogenous GRP94) or Erk as a control. Data are representative of 3 independent experiments.
Fig 6.
Defective egress of Y870A TLR9 from the ER results in an inability to generate active receptor.
(A) Tlr9-/- bone marrow was transduced with WT or Y870A TLR9 and differentiated to DCs. BMDCs were cultured at 37°C in the absence or presence of Baf A (0.5 μM) or Bref A (7 nM) for 4 hrs, or at 30°C for 18 hrs. Cell lysates were then assessed for TLR9 processing by immunoblotting for HA. Erk is shown as a loading control, and untransduced Tlr9-/- BMDCs (TLR9 KO) are shown as a negative control. Blot shown is representative of 3 independent experiments. Positions of bands representing the full length TLR9 precursor (160 kDa), the Golgi-modified full-length TLR9 (~180 kDa), or mature processed TLR9 (80 kDa) are indicated. B. Quantification of the mature 80 kDa TLR9 product normalized to Erk at 37 and 30 degrees conditions averaged from 3 experiments. * p < 0.001.
Fig 7.
Y870A TLR9 is not retained in the ER or Golgi.
Tlr9-/- bone marrow was transduced with doxycycline inducible WT or Y870A HA-tagged TLR9, and differentiated towards DCs. Three days later, protein expression was induced by doxycycline (0.5μg/ml) treatment, and BMDCs were fixed and permeabilized at 4 hours post-Doxycycline treatment. The 4 h post-doxycycline fixed cells were labeled for HA and either the late endosome/lysosome marker LAMP1 (A), the ER marker Calreticulin (B), or the Golgi marker GM130 (C). In all cases, nuclei were labeled by DAPI. Each panel shows an image of both WT and Y870A-expressing BMDCs for each individual label, a merged image (Merged), and a colocalization image in which areas of overlap of the two markers, after thresholding by the method of Costes [24], is indicated by the white areas (“co-loc. pixels”). HA colocalization with each organelle marker is expressed both as a Pearson’s correlation coefficient and by the percent of manually thresholded HA-labeled structures that overlap with the marker. *, p<0.01; **, p<0.005; ***, p<0.0001.
Fig 8.
Y870A TLR is consumed by autophagy.
Tlr9-/- bone marrow was transduced with doxycycline inducible WT or Y870A HA-tagged TLR9, and differentiated towards DCs. Three days later, protein expression was induced by doxycycline (0.5μg/ml) treatment, and BMDCs were fixed and permeabilized at 4- and 24- hours post-doxycycline treatment. The 4 h post-doxycycline fixed cells were labeled for HA and the autophagosome markers SQSTM1 / p62 (A) or LC3B (B), and then by species-specific Alexa568- and Alexa488-conjugated secondary antibodies. The 24 h post-doxycycline fixed cells were labeled for HA and LAMP1 (C). In all cases, nuclei were labeled by DAPI. Each panel shows an image of both WT and Y870A-expressing BMDCs for each individual label, a merged image (Merged), and a colocalization image in which areas of overlap of the two markers after thresholding by the method of Costes is indicated by the white areas (“co-loc. pixels”). HA colocalization with each organelle marker is expressed both as a Pearson’s correlation coefficient and by the percent of manually thresholded HA-labeled structures that overlap with the marker. *, p<0.01; **, p<0.005; ***, p<0.0001.