Volume 24, Issue 3, Pages 437-447 (March 2016) Reconstructing the TIR Side of the Myddosome: a Paradigm for TIR-TIR Interactions  Laurens Vyncke, Celia Bovijn, Ewald Pauwels, Tim Van Acker, Elien Ruyssinck, Elianne Burg, Jan Tavernier, Frank Peelman  Structure  Volume 24, Issue 3, Pages 437-447 (March 2016) DOI: 10.1016/j.str.2015.12.018 Copyright © 2016 Elsevier Ltd Terms and Conditions

Figure 1 MAPPIT Assays Reveal Four Binding Sites in MyD88 TIR (A–C) The MyD88 TIR crystal structure is shown in three different orientations. BS, binding site. (D) Ribbon presentations of the three different orientations. Residues are colored according to the effect of mutations in the MyD88 TIR-Mal and MyD88 TIR-MyD88 MAPPIT assays. The table below lists the relative activities (% of wild-type) corresponding to these colors. Non-mutated residues or residues which do not correspond to the table criteria are gray, and backbone atoms are black. See also Figure S1. Structure 2016 24, 437-447DOI: (10.1016/j.str.2015.12.018) Copyright © 2016 Elsevier Ltd Terms and Conditions

Figure 2 BS I, BS II, and BS III, but Not BS IV, Are Essential for NF-κB Activation (A–C) The MyD88 TIR crystal structure is shown in three different orientations. (D) Ribbon presentations of the three different orientations. Residues are colored according to the effect of mutations in the MyD88 TIR-Mal and MyD88 TIR-MyD88 MAPPIT assays, and in the NF-κB activation assay. The table below lists the relative activities (% of wild-type) corresponding to these colors. Non-mutated residues or residues which do not correspond to the table criteria are gray, and backbone atoms are black. See also Figures S2–S4 and Tables S1, S2 and S3. Structure 2016 24, 437-447DOI: (10.1016/j.str.2015.12.018) Copyright © 2016 Elsevier Ltd Terms and Conditions

Figure 3 BS IV Mutations in MyD88 Disrupt Mal Binding in a Co-immunoprecipitation Assay (A) We co-transfected FLAG-tagged wild-type or mutant MyD88 with E-tagged Mal. FLAG-tagged SVT was transfected as a negative control. All mutations in BS IV, indicated by asterisks, abolish the interaction with Mal (all p-values <0.01), while the S244A and S244D mutations do not disrupt Mal binding. The L252P mutation (p < 0.0001) also abolishes binding with Mal, but this could be due to lower expression of the mutant. (B) Densitometric analysis of the effect of MyD88 mutants on Mal binding. Data are represented as relative area from three to five independent experiments. Structure 2016 24, 437-447DOI: (10.1016/j.str.2015.12.018) Copyright © 2016 Elsevier Ltd Terms and Conditions

Figure 4 Electrostatic and Hydrophobic Complementarity Is in Agreement with the Site I-III and Site II-II Interactions (A and B) The site I-III interaction is electrostatically favorable. The electrostatic surface potential (kT/e−) is plotted on the crystal structure of MyD88 TIR, as indicated in the color scale. Electrostatic complementarity between BS I (A) and BS III (B) facilitates the site I-III interaction. (C) Hydrophobic complementarity is in line with the site II-II interaction. The core of BS II consists of a hydrophobic patch, which favors the site II-II interaction. The atomic hydrophobicity is plotted on the crystal structure of MyD88 TIR, as indicated by the color scale. Structure 2016 24, 437-447DOI: (10.1016/j.str.2015.12.018) Copyright © 2016 Elsevier Ltd Terms and Conditions

Figure 5 Detail of the Site I-III and Site II-II Interfaces, and the Position of S244 and L252 Near BS II and BS IV (A) In the site I-III docking models, the BB-loop of BS I interacts with the αE-helix of BS III. P200 of BS I interacts with W284 in BS III. D195 and D197 of BS I form salt bridges with R288 and K291. (B) In the site II-II models, a symmetrical binding interface is formed by the αC′- and αD-helices. The core of the hydrophobic interaction is mediated by L241, S266, and I267. Additional salt bridges are formed between D234 and R269. (C) S244 and L252 (spheres) are closely located to the αC′- and αD-helices. L252 contacts residues of the αC- and αD-helices, possibly affecting BS II (L241, I267, and R269) and BS IV (F270 and T272). See also Figure S8. Structure 2016 24, 437-447DOI: (10.1016/j.str.2015.12.018) Copyright © 2016 Elsevier Ltd Terms and Conditions

Figure 6 MyD88 TIR Dimerization via the Site II-II Interaction Results in a Large Mal-Specific Binding Platform, Consisting of BS IV and Residues S209 and L211 (A and B) Residues are colored according to the effect of mutations in the MyD88 TIR-Mal and MyD88 TIR-MyD88 MAPPIT assays (A), and in the NF-κB activation assay (B). The inset tables list the relative activities (% of wild-type) corresponding to these colors. The dashed line represents the site II-II interface. Non-mutated residues or residues which do not correspond to the table criteria are gray, and backbone atoms are black. (C) Ribbon presentation of the orientation of the MyD88 TIR dimer. Structure 2016 24, 437-447DOI: (10.1016/j.str.2015.12.018) Copyright © 2016 Elsevier Ltd Terms and Conditions

Figure 7 Alternating Site I-III and Site II-II Interactions Leads to a Helical MyD88 TIR Model that Fits the Helical Organization of the DDs in the Myddosome Crystal Structure (A) The left-handed helical MyD88 TIR model (top) closely resembles the organization of the DDs in the Myddosome crystal structure (bottom) (PDB: 3MOP). The backbone of the N-terminal M156 residue of each TIR domain is shown as spheres. The ID connections between TIRs and their corresponding DDs are presented as dashed lines. (B) Two alternative views of the MyD88 TIR helical model. Residues are colored according to the effect of the mutations in MyD88 TIR-Mal and MyD88 TIR-MyD88 MAPPIT assays, and in the NF-κB activation assay, as in Figure 6B. TIR 1-TIR 2, TIR 3-TIR 4 (not visible), and TIR 5-TIR 6 interact via site I and site III (top). TIR 2-TIR 3 and TIR 4-TIR 5 (not visible) interact symmetrically via their site II (bottom). This exposes a large Mal-specific binding site containing BS IV and residues S209 and L211. The helical assembly leads to additional TIR 1-TIR 5 and TIR 2-TIR 6 contacts, including residues E159 and R215. Structure 2016 24, 437-447DOI: (10.1016/j.str.2015.12.018) Copyright © 2016 Elsevier Ltd Terms and Conditions

Figure 8 Population Analysis on the Wild-Type MyD88 TIR and S244D and L252P Mutant Trajectories Reveals Differences in the Conformation of the Loop Containing Residues 243–248 The similarity of each trajectory frame to four representative cluster conformations, indicated as conf 1 to 4, for this loop was examined. The corresponding populations for wild-type (WT) MyD88 TIR and S244D and L252P mutants are shown. Wild-type MyD88 TIR preferentially adopts the conf 1 conformation. The S244D and L252P mutants preferentially adopt the conf 3 conformation. See also Figures S5–S7. Structure 2016 24, 437-447DOI: (10.1016/j.str.2015.12.018) Copyright © 2016 Elsevier Ltd Terms and Conditions