(A) The current Dpr-ome. New interactions identified in this study are in red.
(B) Structure of Dpr6-D1D2 bound to DIP-α-D1. The complex is formed by interactions between the D1s only. The D1s are related to each other by a near-symmetrical two-fold axis (closed oval). N-linked sugars are shown as sticks.
(C) The same structure rotated to show the 130° angle between the interacting domains and the angle of approach of the two proteins.
(D) Comparison of the DPR-DIP complex structure with two similar IgSF CAM complexes. Side chains for residues at the centers of interface are shown as sticks.
(E) Superimposition of the Dpr6-DIP-α complex onto related D1 complexes. Dpr6 is in cyan and DIP-α in red. Note that the overall path of the main chains is almost identical to those in the SYG-1-SYG-2 complex. The Dpr6-DIP-α complex is less similar to JAML-CAR, and diverges most from the homophilic Nectin-2 complex.

(A) The interaction interface between Dpr6 (cyan) and DIP-α (red) is strongly hydrophobic and contains no salt bridges. The three side chain-to-side chain hydrogen bonds, including Dpr6 H114-DIP-α Q125, are labeled as dashes. The hydrophobic residues at the center of the interface, Dpr6 I115 and DIP-α I83, are also labeled. NAG represents the N-acetyl-DGlucosamine glycan residue on Dpr6 D1.
(B) Equilibrium and kinetic parameters measured for the interactions between wild-type and mutant versions of Dpr6 and DIP-α, and Dpr11 and DIP-γ using SPR. For several mutants, binding kinetics were too fast to measure. Raw binding curves are in .
(C) and (D) Rows of Dpr6 and DIP-α residues interdigitate, leading to shape complementarity between the two surfaces. Interestingly, Dpr6 Q156/158 and DIP-α K81 sidechains, while packing closely, do not form hydrogen bonds.

Projections of 3-6 confocal slices showing EGFP expression (green) and GAL4-driven nuclear dsRed expression (magenta) from: dpr11^{MiMIC MI02231 5’>EGFP} (A, B, E), DIP-γ^{MiMIC MI03222 5’>EGFP} (C, D), dpr6^{MiMIC MI04582 GT} (F, G), dpr10^{MiMIC MI03557 GT} (H, I), and DIP-α^{MiMIC MI02031 GT} (J, K). GT=gene trap. GAL4 drivers were: OK6 (motor neurons)(A, C, F, H, J); ChAT (cholinergic interneurons)(B, D, G, I, K); or DIP-γ^{MiMIC MI03222→GAL4} (E). Arrows: cells expressing both dsRed reporter and GFP at high levels. Carets: cells expressing dsRed and low levels of GFP. Double carets: cells that express only dsRed. Caret in (E): a cell that has GFP but only low levels of dsRed. Focal planes are different for OK6 and ChAT images, so different dpr/DIP cells are seen. One pair of OK6⁺ dorsal midline motor neurons expresses dpr11 and DIP-γ, the other pair expresses DIP-α at high levels and DIP-γ at low levels, and dpr6 and dpr10 are expressed by both pairs. Scale bar: 5 μm.

Panels A-I are muscle 4 NMJs.
(A, Ai) GFP driven by dpr11^GMR95G12-GAL4 (Janelia Farm) labels boutons (arrows). Note green puncta on the muscle surface. Magenta: anti-HRP. Scale bar: 5 μm.
(B, Bi) Dpr11-EGFP-Dpr11 outlines anti-HRP-labeled boutons, and is observed postsynaptically outside of bouton borders (arrowhead) and on the muscle surface. Scale bar: 2 μm.
(C, Ci): EGFP driven by DIP-γ^{MiMIC MI03222→GAL4} labels boutons (arrow). Note green puncta on the muscle surface. Scale bar: 5 μm.
(D-F) dpr11 and DIP-γ mutants have satellite bouton phenotypes (arrows indicate satellites). Anti-HRP staining. Insets: single 1b boutons. Scale bar: 5 μm.
(G): Quantitation of satellite bouton phenotypes. Experimental genotypes in black, controls in grey. p<.0001 (***) for differences between relevant genotypes.
(H-I): pMad is elevated at the NMJ in dpr11 mutants. Scale bar: 5μm.
(J): Tracings of spontaneous events. Recordings from muscles 6 and 7.
(K-L): Cumulative probability graphs of mEPSP amplitude (J) and mEPSP frequency (K). The curves for dpr11 and DIP-γ mutants (red and orange) are shifted to the right relative to all other genotypes. p<.0001 for mEPSP frequency change relative to controls, p<.001 for mEPSP amplitude (see ).

(A) The Me, Lo, and Lop areas of the OL. Top images show OL orientation relative to the fly head. Me neuropil layers are labeled and profiles of neurons of interest are superimposed onto the diagram. Adapted from ().
Panels B-F are projections of 3-4 confocal slices of 80 hr. APF medullas. nc82 (anti-Brp; magenta) was used to identify neuropil layers (marked on the sides of panels). Asterisks highlight labeled layers; layer numbers are indicated at the sides of the panels.
(B) dpr11^{MiMIC MI02231 5’>EGFP} labels M3, M6 (weak) and M8-10.
(C) DIP-γ^{MiMIC MI03222 5’>EGFP} labels M3, M6, M8, and M10.
(D) Dpr6-EGFP-Dpr6 labels M1, M4, M6, M8, and M10.
(E) dpr10^{MiMIC MI03557 GT} labels M1, M3, M4, M8 (weak) and M10.
(F) DIP-α^{MiMIC MI02031 GT} labels M1, M3, M6, and M8.
Scale bar: 10 μm for all panels.

(A) R7 growth cones in M6 at 47 hr. APF, labeled by dpr11^{MiMIC MI02231 5’>EGFP} (green) and 24B10 (magenta). A subset of R7s (arrows) express dpr11. Arrowheads, pR7s express only 24B10. Green layer is M3 (asterisk). Scale bars for (A-B): 5 μm.
(B) dpr11^{MiMIC MI02231 5’>EGFP} is selectively expressed by Rh4>tdTomato-expressing yR7s (magenta; arrows) in pupal retina at 80 hr. APF. Dotted circles outline a p (Rh3) ommatidium.
(C) DIP-γ^{MiMIC MI03222 5’>EGFP} (white/green) labels M6 (arrows) at 47 hr. APF and colocalizes with a mCD8-RFP reporter driven by a Dm8-specific split-GAL4 driver (magenta). Red asterisks, DIP-γ expressing Me layers. Scale bar: 10 μm.
(D) FLP-out labeling of a single DIP-γ^{MiMIC MI03222→GAL4} expressing Dm8 (green), in an adult. R7/R8 axons are labeled by 24B10 (magenta). Brackets, M6 arborization; arrow, cell body. Scale bar: 5 μm.
(E) FLP-out labeling of a single DIP-γ^{MiMIC MI03222→GAL4} expressing LPTC, in an adult. Cell body, arrow; main dendrite, arrowhead.
(F) dpr11 is expressed by T4 and T5 cells, as shown by double labeling with dpr11^{MiMIC MI02231 5’>EGFP} (green) and a T4/T5-specific GAL4, R42F06, driving mCD8-RFP (magenta). Arrow, T4 dendrites in M10; arrowhead, T5 dendrites in Lo1. Note that, although there are 4 layers of the Lop labeled by the driver (magenta/white in panel F; asterisks), only Lop1 and Lop2 have green GFP labeling in panel Fi (white asterisks).
(G) M10 labeling by dpr11^{MiMIC MI02231 5’>EGFP} reflects dpr11 expression by T4s, because T4 axons connecting M10 to the Lop are double-labeled (arrows).
(H) A subset of T4 and T5 cell bodies are labeled by dpr11^{MiMIC MI02231 5’>EGFP}. A T4 /T5 cell bodies that expresses dpr11 (arrow) and one that does not (arrowhead) are labeled.
(I) DIP-γ^{MiMIC MI03222 5’>EGFP} labels a zone in the Lop spanning Lop1/2 (brackets) at 47 hr. APF, most likely representing LPTC dendrites. Scale bars for (E-H): 10 μm.

(A) In dpr11^MiMIC/+ adults, Brp-short^mCherry labeled yR7 terminals (magenta; white asterisks) line up in M6 at a point defined by the dark/light demarcation line in the EGFP labeling (green). Scale bar for (A-D): 5 μm.
(B) In dpr11^MiMIC/Df adults, some yR7s overshoot M6 and grow into deeper layers (red asterisks). White asterisks: yR7s that do not overshoot. Note that the shapes of all terminals are irregular, and that they are not aligned along the vertical axis.
(C) In DIP-γ^MiMIC/+ adults, Brp-short^mCherry labeled yR7 terminals (magenta; white asterisks) are superimposed on Dm8 arborizations (green; arrow) labeled by EGFP. Red asterisk: a yR7 scored as an overshoot.
(D) In DIP-γ^MiMIC/Df adults, some yR7s overshoot and terminal shape and alignment are altered, as in dpr11 mutants. Red triangles indicate yR7 overshoots in gaps where no DIP-γ⁺ Dm8 is present, and white triangles highlight yR7 overshoots that grow through a DIP-γ⁺ Dm8 arbor (arrows).
(E) Quantitation of overshoot phenotypes. Overshoots tend to extend further in mutants (B, D) than in controls (C), but we have scored mutant and control overshoots as the same phenotype because we lack a quantitative way to distinguish them.
(F-G) Imaris surface renderings of yR7 terminals in dpr11^MiMIC/+ (F) and dpr11^MiMIC/Df (G). White asterisks, terminals at normal positions; red asterisks, overshoot terminals.
(H-I) R7 growth cones at 25 hr. APF in dpr11^MiMIC/+ (H) and dpr11^MiMIC/Df (I), visualized using PM181-GAL4 driving tdTomato. Asterisks indicate yR7s, identified by the dpr11^MiMIC GFP marker.
(J,K) In single confocal slices from DIP-γ^MiMIC/+ adults, the GFP-labeled M6 layer is continuous, but it has large gaps (red brackets) in DIP-γ^MiMIC/Df.
(L,M) M6 gaps (brackets) are also seen in DIP-γ mutants when Dm8 arborizations are labeled using a Dm8-specific split-GAL4 driver/reporter. Scale bars for (F-I): 10 μm.
(N,O) Quantitation of numbers of Ort^C2b+ Dm8s (N) and of Ort^C2b+ DIP-γ⁺ Dm8s (O) per OL in DIP-γ^MiMIC/+ and DIP-γ^MiMIC/Df.