The arm stabilizes at the leading edge.

(A) Green fluorescence indicates newly accumulating Arm, as only the GFP has had time to mature. Higher levels of Arm localize at the leading edge (the first row of epithelial cells of the dorsal gap). (B) Red fluorescence shows stable Arm protein with time for RFP to mature. (C) The leading edge exhibits bright yellow fluorescence, showing colocalization of mainly within dorsal closure.

Arm is required for dorsal closure.

(A) Arm-CRY2-mCh without blue-light activation localized at the leading edge of dorsal closure. The insertion of CRY2 and RFP did not perturb its localization or function. (A’) Light-induced oligomerization of Arm-CRY2-mCh through the application of blue light led to a disruption in dorsal closure.

Canonical Wnt activation does not affect dorsal closure, while inactivation of Wnt results in defective dorsal closure.

(A–A’’) Wingless (C381-Gal4 X UAS-Wg, UAS-RFP), the primary Wnt pathway ligand, is expressed perpendicular to the leading edge cells (Arm-GFP), and does not affect dorsal closure. (B–B’’) Wnt activation through uniform Wg overexpression (UAS Wg) does not affect dorsal closure. (C–C’’) Wnt signaling inactivation through Axin overexpression results in dorsal closure defect. (D–D’’) Expression of the gain-of-function allele of Arm, ΔArm resulted in a dorsal closure defect. (E–E’’) Gain of function allele ArmS56A also showed a dorsal closure defect.

E-cadherin, actin and α-catenin co-localize with Arm at the leading edge, and α-catenin is necessary for dorsal closure.

(A–A’’) E-cadherin (red) colocalizes with Arm (green) at junctions. (B, B’) Filamentous actin follows a similar pattern to adhesive junctions (B mKO tagged LifeActin, B’ projection using IMARIS software). (C) In the absence of blue light, α-catenin (red) localizes at the leading edge. (C, C’) Activation of CRY2 with blue light led to a loss of localization of α-catenin and a severe dorsal closure defect.

Functional dissection of point mutations in ArmTimer during dorsal closure.

(A–A’’) Wild-type ArmTimer control shows high production and stabilization at the leading edge, as indicated by green (GFP) and red (RFP) fluorescence. (B–B’’) ArmTimer-F1a shows a mild dorsal closure defect and reduced fluorescence, validating the allele-replacement strategy. (C–C’’) ArmTimer-FF (Y150F, Y667F) shows no overt dorsal closure defect but reduced GFP and RFP intensities. (D–D’’) ArmTimer-EE (Y150E, Y667E) shows no overt defect but further reduced GFP and RFP intensities. (F) Quantification of GFP/RFP intensities comparing wild type, –EE, and –FF. (G) Dorsal closure duration is increased in ArmTimer-EE.

ArmTimer-AA disrupts zippering dynamics, supporting the requirement for Arm–α-catenin interaction.

(A–A’’) ArmTimer-AA (T111A, T121A) shows asymmetric zippering and an uneven dorsal gap. (B) GFP comparison indicates lower Arm production/accumulation in ArmTimer-AA relative to wild type at the same time point. (B’) RFP comparison indicates similar stabilization signal to wild type. (C) Dorsal closure duration is significantly increased for ArmTimer-AA.

Dishevelled accumulates at the leading edge and the DEP domain is required for dorsal closure.

(A–A’’) Dsh visualized with frankenbody reporters accumulates at the leading edge (GFP, RFP, and overlap). (B–B’’) Membrane-tethered Dsh localizes to leading edge and amnioserosa and disrupts amnioserosa junction integrity. (C–C’’) DshΔDIX localizes to membrane and causes amnioserosa integrity defects without blocking closure. (D–D’’) DshΔPDZ is diffuse but dorsal closure proceeds. (E–E’’) DshΔDEP results in failure of germ-band retraction and dorsal closure.

JNK signaling regulates ArmTimer dynamics and is essential for proper dorsal closure.

(A–A’’) ArmTimer embryos with uniform JNK-GFP expression show normal closure but reduced Arm-RFP intensity. (B–B’’) ArmTimer × JNK RNAi shows disrupted dorsal closure and altered Arm leading-edge localization. (C) Quantification of Arm-RFP intensity across JNK-GFP, JNK RNAi, and wild-type ArmTimer.