The MIM I-BAR domain is essential but not sufficient for function and localization. (A) Schematic of the constructs expressed in transgenic lines. MIM^ΔIBAR—in-frame knock-in of an EGFP expression cassette into the I-BAR domain. MIM isoform C is illustrated, but all isoforms include the insertion. OE constructs of I-BAR and I-BAR-Emerald are under UAS control. We could not detect MIM^ΔIBAR EGFP signal in our imaging experiments, most likely because it is under our detection limit (over background). (B) Mean frequency (%) of stalling, full collapse (FC), kiss-and-run (KAR), and membrane crumpling (MC) in MIM^ΔIBAR/Def, MIM^ΔIBAR/+ and I-BAR^OE compared to MIM^+/Def and MIM^Null/Def (from Fig. 6 B, semi-transparent; included for convenience). MIM^ΔIBAR/Def phenocopied MIM^Null/Def with a high frequency of kiss-and-run events (68 ± 7% (****)). MIM^ΔIBAR/+ displayed a high frequency of full collapse events (50 ± 13% (**)), phenocopying MIM^KD. I-BAR^OE displayed a high frequency of kiss-and-run events (50 ± 7% (****)). SGs express Glue-GFP and LifeAct-Ruby. Error bars = SEM. Statistical significance with respect to MIM^+/Def for the MIM^ΔIBAR expressing SGs and to WT SGs for the I-BAR^OE SGs. N (SGs) ≥ 3, n (events) ≥ 150. P values for MIM^ΔIBAR/Def: ****(KAR) <0.000001, ****(MC) = 0.000012, for MIM^ΔIBAR/+: **(FC) = 0.003965, for I-BAR^OE: ****(KAR) = 0.000011, ****(MC) = 0.000015. Two-tailed, unpaired multiple t tests corrected using the Holm-Sidak method. (C) Time-lapse sequence (SRRF intensity-projection) of representative LSVs undergoing KAR w actomyosin and FC (complementary examples in Fig. S5 A) from MIM ^ΔIBAR/Def and MIM^ΔIBAR/+ SGs. Glue-GFP (green), LifeAct-Ruby (magenta). Fusion events (asterisks). Top: LSV (dashed line) that recruited actin, deformed, and moved away from the apical membrane, eventually fusing again at 11:12 (KAR w actomyosin). Bottom: Expanding pore (double arrowheads) leads to apparent vesicle integration into the apical membrane (FC). Scale bars = 1 µm. (D) Confocal intensity projection of a representative SG expressing Glue-DsRed (Costantino et al., 2008; green) and I-BAR-Emerald (cyan). Left: Overview of the SG showing the I-BAR-Emerald did not localize to cytoplasmic clusters. Top right: View of the apical surface showing homogenous I-BAR-Emerald localization on the apical membranes. Bottom right: Time-lapse sequence of a representative LSV (SRRF intensity-projection) during exocytosis, showing that I-BAR-Emerald did not localize specifically to the fusion pore (arrowheads). (E) Confocal intensity-projection of a representative SG ectopically expressing I-BAR-Emerald (cyan) and MIM-mScarlet (magenta). Left: Overview of the SG. Top right: Magnified views showing that the I-BAR-Emerald and MIM-mScarlet co-localized in cytoplasmic clusters and on the apical membrane. Bottom right: Time-lapse sequence of a representative LSV during exocytosis showing the IBAR-Emerald and MIM-mScarlet co-localizing at the fusion site and the fusion pore during secretion. UAS expression in MIM^+/Def, MIM^Null/Def, and MIM^ΔIBAR/Def SGs driven by fkh-GAL4, in I-BAR^OE (D and E) driven by c135-GAL4. In D and E, yellow squares mark the magnified area. Time mm:ss; relative to fusion in C–E. Scale bars = 20 µm (left), 10 µm (top right), 1 µm (bottom right).