An extracellular interactome of immunoglobulin and LRR proteins reveals receptor-ligand networks

Abstract

Extracellular domains of cell surface receptors and ligands mediate cell-cell communication, adhesion, and initiation of signaling events, but most existing protein-protein "interactome" data sets lack information for extracellular interactions. We probed interactions between receptor extracellular domains, focusing on a set of 202 proteins composed of the Drosophila melanogaster immunoglobulin superfamily (IgSF), fibronectin type III (FnIII), and leucine-rich repeat (LRR) families, which are known to be important in neuronal and developmental functions. Out of 20,503 candidate protein pairs tested, we observed 106 interactions, 83 of which were previously unknown. We "deorphanized" the 20 member subfamily of defective-in-proboscis-response IgSF proteins, showing that they selectively interact with an 11 member subfamily of previously uncharacterized IgSF proteins. Both subfamilies interact with a single common "orphan" LRR protein. We also observed interactions between Hedgehog and EGFR pathway components. Several of these interactions could be visualized in live-dissected embryos, demonstrating that this approach can identify physiologically relevant receptor-ligand pairs.

Figure 1. Schematic representation of the Extracellular Interactome

Schematic representation of the Extracellular Interactome in the following steps: (1) Annotation of all proteins containing extracellular IgSF, FnIII and LRR domains (see also Table S1), (2) Cloning of the ECDs of target genes with PCR from cDNA or RT-PCR from mRNA into Drosophila expression plasmids, (3) Expression of all cloned ECDs in bait and prey formats in Drosophila cell culture, (4) Application of the Extracellular Interactome Assay, (5) Collection and analysis of the assay results, and (6) Confirmation of hits via biophysical methods, such as Surface Plasmon Resonance. See Figure S1 for a detailed explanation of tags used for bait and prey.

Figure 2. Interactions detected by our Extracellular Interactome

(A) The matrix of data for 202×202 pairwise interactions measured using the Extracellular Interactome Assay (ECIA). Each row represents a single bait, while each column represents a single prey protein. The color scale maps to Absorbance values at 650 nm. See Figure S2 for the processed data matrix, and Table S2 for list of interactions detected. (B) Number of interactions observed (blue slices) out of all unique homophilic and heterophilic interactions tested (red circles). (C) Observed interactions are further classified as novel (green) vs. previously known (beige).

Figure 3. The Extracellular Interactome of the Drosophila IgSF, FnIII and LRR

The Extracellular Interactome of the Drosophila IgSF, FnIII and LRR, with the four IgSF subfamilies (Dpr, DIP, Side and Beat) delineated. See the inset for the classification of protein domain families and interactions. Uncharacterized gene names in the CGxxxx(x) format are shortened by removing the prefix CG. Homophilic interactions of boi, fred and robo3 were considered previously known based on the similar properties of their paralogs, ihog, ed and lea (robo2). * indicates interactions of the N-terminal fragment of Slit. The subsections highlight separate classes of proteins and interactions: (A) Dprs and the Dpr-DIP interaction network, (B) Beat and Side families, (C) proteins involved in IgSF–LRR interactions, (D) homologs of SYG and Netrin proteins, which are involved in synaptogenesis and axon guidance, (E) Vn and its interaction partners in the EGF and Hedgehog signaling pathways (See also Figure S4), (F) homophilic-only IgSF, including Dscams, (H) Fas2 and its interactions, and (G) others. For comparison, see Tables S3 and S4 for DroID reported interactions.

Figure 4. The Dpr-ome and Biophysical Validation

(A) Independent ECIA measurements of interactions between Dpr, DIP, DIPc and the negative control, Rst. (B) Same as A, but with bait and prey reversed. (C) The phylogenetic trees relating Dpr and DIP extracellular domain sequences, with interactions mapped between the Dprs and DIPs. Also see Figure S3 for the common domain topologies of the Dpr and DIP subfamilies. (D) Binding affinities and kinetics measured with purified recombinant monomeric ECDs using SPR for a selection of the Extracellular Interactome hits. See Figure S5 for details. * indicates kinetic constants that may be inaccurate due to fast kinetics; 0.1 s⁻¹ < k_off < 0.5 s⁻¹. † indicates that binding kinetics was too fast to be measured; k_off > 0.5 s⁻¹.

Figure 5. Fasciclin II-mediated homophilic adhesion can be detected by fusion protein binding to live embryos

Live-dissected stage 16 embryos incubated with Fas2-AP₅ were stained with anti-Fas2 mAb 1D4 (red) and rabbit anti-AP (green) antibodies. Three segments of the CNS and ventral periphery are shown in each image; anterior is up. (A, A1) In wild-type (wt) embryos, the pattern of Fas2-AP₅ binding is superimposable on the pattern of anti-Fas2 staining. Arrow, longitudinal tract; arrowhead, a motor axon bundle. (B, B1) Embryos heterozygous for the null Fas2^EB112 allele have only 50% of the normal amount of Fas2, and thus show less binding to Fas2-AP₅ and weaker staining with anti-Fas2 mAb. (C, C1) Homozygous Fas2^EB112 embryos, which have no Fas2, fail to stain with either Fas2-AP₅ or anti-Fas2 mAb. See Figure S6 for CNS patterning of these embryos. Scale bar, 10 μm.

Figure 6. Binding of Dpr-AP₅ and DIP-AP₅ fusion proteins to live embryos

Staining of the longitudinal tracts and motor axons was performed as in Figure 5, except that Dpr-AP₅ and DIP-AP₅ fusion proteins were used for staining. (A, A1) Dpr6-AP₅ stains a fuzzy ladder-like pattern (arrowhead) in the same focal plane as the longitudinal tracts. This may represent glial or neuronal cell surfaces. (B, B1) Dpr8-AP₅ weakly stains a segmentally repeated pattern of small puncta (arrowhead) within the borders of the longitudinal tracts. These puncta, and the puncta seen with Dpr12-AP₅, CG14521-AP₅, and CG42343-AP₅, are probably too small to represent entire cell bodies. For Dpr8-AP₅, Dpr12-AP₅, and CG42343-AP₅, the puncta are in the focal plane of the longitudinal tracts, and may represent cross-sections through dorsoventrally projecting neuronal processes and growth cones. (C, C1) The Dpr12-AP₅ puncta are slightly lateral to the longitudinal tracts. (D, D1) CG14521-AP₅ stains puncta (arrowhead) that are dorsal to the longitudinal tracts, as shown by the fact that longitudinal tracts are not visible in (D1), but motor axons, which are more dorsal, are in the focal plane. The dorsal localization of these puncta suggests that they may represent portions of the surfaces of longitudinal glia. (E, E1) CG42343-AP₅ stains puncta (arrowhead) and a fuzzy pattern along the longitudinal tracts. (F, F1) CG10824-AP₅, which binds to all Dprs and some DIPs, brightly stains most or all longitudinal axons (arrowhead). Scale bar, 10 μm. See Figure S6 Robo-AP₅ staining of live embryos.

Figure 7. Genetic evidence for a specific interaction between Dpr11 and CG14521

Staining was performed as in Figures 5 and 6, using Dpr11-AP₅ fusion proteins, anti-Fas2 mAb 1D4, anti-GFP, and anti-HRP (a marker that labels all Drosophila neurons). (A) In wild-type embryos, Dpr11-AP₅ (white) stains a segmentally repeated pattern of large dorsal cell bodies within the CNS (one cell per hemisegment; arrow). (B) In heterozygotes for the CG14521 MiMIC element, Dpr11-AP₅ (red) stains the same cell body pattern (arrow). Staining is weaker than in (A). (B₁) The same image as in B, but showing both the Dpr11-AP₅ (red) and anti-GFP (green) signals. Note that the green cells are ringed by red staining (arrow); one is shown at higher magnification in the inset. (C) Double-staining of a CG14521 MiMIC homozygote with anti-HRP (red) and anti-GFP (green) shows a regular pattern of axons, which is indistinguishable from wild-type (see Figure S6A₁ for anti-HRP staining of a wild-type embryo). (D) A CG14521 MiMIC homozygote stained with Dpr11-AP₅ (red). No staining of the cell bodies is observed. (D₁) The same image as in (D), but showing both the Dpr11-AP₅ (red) and anti-GFP (green) signals. Note that the large dorsal cell bodies are present (arrow), and are brighter than in the heterozygote (B₁), but there is no red staining associated with them. One cell body is shown at higher magnification in the inset. Note also that in the homozygote the brighter GFP expression reveals other cell bodies that stain more weakly with anti-GFP, implying that they express CG14521 at lower levels. Cell bodies at these positions are not detectably stained by Dpr11-AP₅ in (A) or (B). (E) A wild-type embryo double-stained with Dpr11-AP₅ and anti-Fas2 mAb, showing the focal plane of the muscles, which are dorsal to the CNS focal plane in a live-dissected embryo. No CNS staining is visible in this focal plane, and the muscles do not stain with Dpr11-AP₅. Motor axons are visible, however (red). The borders of the CNS are indicated by dotted lines. (F) An embryo in which CG14521 was overexpressed in muscles. Note that these embryos display bright staining of the outlines of the ventrolateral and ventral muscle fibers (arrowhead indicates the muscle 6/7 border). Arrow, a motor axon bundle. Scale bar, 10 μm.