Functional insights from the crystal structure of the N-terminal domain of the prototypical toll receptor

Abstract

Drosophila melanogaster Toll is the founding member of an important family of pathogen-recognition receptors in humans, the Toll-like receptor (TLR) family. In contrast, the prototypical receptor is a cytokine-like receptor for Spätzle (Spz) protein and plays a dual role in both development and immunity. Here, we present the crystal structure of the N-terminal domain of the receptor that encompasses the first 201 amino acids at 2.4 Å resolution. To our knowledge, the cysteine-rich cap adopts a novel fold unique to Toll-1 orthologs in insects and that is not critical for ligand binding. However, we observed that an antibody directed against the first ten LRRs blocks Spz signaling in a Drosophila cell-based assay. Supplemented by point mutagenesis and deletion analysis, our data suggests that the region up to LRR 14 is involved in Spz binding. Comparison with mammalian TLRs reconciles previous contradictory findings about the mechanism of Toll activation.

Graphical abstract

Figure 1

Overall Topology of the Toll_N6-VLR Hybrid (A) View facing the concave surface of the leucine-rich fold. Glycans attached to asparagine residues 80, 140, and 175 are depicted with light-blue spheres. The Toll LRRNT cap is represented in marine, LRRs are in cyan, and the VLR LRRCT cap is in gray. (B) Left side view showing the curvature of the LRRs. Disulfide bonds are shown as yellow sticks. See also Figure S1.

Figure 2

Crystal Packing (A) Asymmetric unit with four molecules of native Toll_N6-VLR and two malonate ions (MLI). Chain A in green, B in cyan, C in magenta, and D in yellow with the VLR portion in a paler color. (B) View tilted by 90° showing the pseudo 2-fold axis, by which the A-C and B-D pairs are related to each other. (C) Molecular surfaces in the asymmetric unit. The VLR caps mediate most of the crystal contacts. A large glycan structure protrudes from one of the chains and is depicted in magenta sticks. (D) Close-up view on the complex glycan structure bound to Asn 140. See also Table S1 and Figure S3.

Figure 3

The N-Terminal Cap of Toll Adopts a Unique Fold (A) Sequence alignment of the N-terminal domain of Toll-1 receptor paralogues in insects. Drosophila melanogaster Toll (Dm-1); Drosophila sechellia (Ds-1); Aedes aegypti (Aa-1); Anopheles gambiae (Ag-1); Bombus terrestris (Bt-1); and Tribolium castaneum (Tc-1). (B–G) N-terminal domains of extracellular LRR proteins. (B) Toll, (C) TLR4, (D) glycoprotein Ib α, (E) Nogo receptor, (F) TLR1, and (G) CD14. LRR proteins are shown in the same orientation to highlight the structural diversity of their LRRNTs (in blue). LRRs are represented in cyan, and disulfide bonds are in yellow. See also Figure S2.

Figure 4

Conformational Diversity of the N-Terminal Leucine-Rich Repeats of Toll (A–F) Each Toll LRR is represented as a cross-section in a planar and a side view at 90°. (F) The hybrid construction in LRR6 is flat, allowing the complete burial of hydrophobic residues in the contiguous LRRs.

Figure 5

Protein-Protein Contacts (A–D) Areas of crystal contact are indicated on the molecular surfaces of Toll_N6-VLR in different views. (A) View facing the concave side, (B) right flank, (C) left flank, and (D) convex side. The first column delineates the structural areas in Toll_N6-VLR with LRRNT in dark blue, LRRs in marine, VLR-LRRCT in light cyan, and glycans in light blue. The second column shows the molecular surface of chain B, and the third is chain D. Contacts are color-coded according to the identity of the molecule that mediates them: chain A is in green, B in cyan, C in magenta, and D in yellow. Residues that have been targeted by site-directed mutagenesis are highlighted in red. See also Figure S3.

Figure 6

Binding Mode of Crystallization Molecules (A and B) Toll_N6-VLR binds malonate ions (MLI) in the native structure, and (C) I3C molecules in the derivative. (A) MLI bound at the pseudo 2-fold axis; (B) MLI bound to chain B in blue and C in magenta. (C) I3C binds to the concave side and interacts with residues of the beta-sheet of chain A and the left flank of the chain B. See also Table S2.

Figure 7

Spätzle Does Not Form a Stable Complex with Toll_N6-VLR (A) Size-exclusion profiles. Processed Spz C-106 and Toll_N6-VLR elute in very similar volumes during gel-filtration. (B) Analytical ultracentrifugation profiles. (C–E) Fit and residuals after fitting to a c(s) model in SEDFIT and the distribution of sedimentation coefficients are shown for (C) Spz C-106 at 422 μg.ml−1 (17.6 μM), (D) Toll_N6-VLR at 377 μg.ml−1 (10.5 μM), and (E) Toll_N6-VLR and C-106 in equimolar amounts (940 μg.ml−1).

Figure 8

Site-Directed Mutagenesis Confirms that Toll LRRNT Is Not Critical for Signaling HEK293ET cells were transfected with a Toll -TLR4 chimera and a NF-κB luciferase reporter. Luciferase production was measured 24 hr after Spz stimulation at a concentration of 10 nM. Data shown represents fold induction compared with stimulation with media only. Data are represented as means ± SEM. See also Table S3.

Figure 9

An Antibody that Recognizes the First Ten LRRs of Toll Blocks Spz Signaling (A) The anti-Toll antibody was added to a culture of S2 cells expressing Toll endogenously and which have been stably transformed with a luciferase reporter gene under a drosomycin promoter. Following a 2 hr incubation, 10 nM of cleaved Spz was added to the cells to test for activation and signaling. Controls included glycerol alone, as well as glycerol plus C-106, to ensure that glycerol played no part in the stimulation (when acting as a cryopreservant in the antibody solution), either by activating or inhibiting it. Data is displayed as fold induction compared to media control. Data are represented as means ± SEM. (B) Gel-filtration chromatography revealing that Toll_N13-VLR is monomeric and binds Spz C106 in a 1:1 complex.