Model of dynamin GG interface. (A) Crystal structure of rat dynamin GTPase domain (green) fused to myosin motor core (PDB: 2aka). A helix from the myosin (red) packs onto a hydrophobic groove comprised of the N- and C-termini (teal) of the GTPase domain. Hydrophobic side chains are shown in yellow. The myosin helix is predicted to mimic the GED docking arrangement. (B) Sequence alignments of the N_GTPase, C_GTPase, and C_GED regions. Hs, Homo sapiens; Rn, Rattus norvegicus; Dm, Drosophila melanogaster; Ce, Caenorhabditis elegans; Sc, S. cerevisiae; Dd, Dictyostelium discoideum; Dyn, dynamin. Black boxes indicate conserved hydrophobic residues. Green and cyan bars indicate the helical segments presumed to form the helical bundle that is modeled in C. Asterisk denotes the position of the kink in the C_GTPase helix. (C) Helical wheel diagrams of the GTPase and GED helices. Yellow, hydrophobic residues; green, GTPase residues; cyan, GED residues. The plots are arranged in the putative docking arrangement, with the conserved residues in B forming the buried hydrophobic core between the three helices.

Intramolecular cross-linking of GTPase and GED helices. (A) Residues selected for double cysteine mutagenesis in GG. Cysteine pairs included one substitution in the GTPase domain (R15C in N_GTPase or R297C in C_GTPase) and one substitution in the C_GED region (R730C, H733C, K736C, or S740C). (B) Structure and estimated distances of bifunctional methanthiosulfonate (MTS) cross-linkers used for analyzing interactions between GTPase and GED helices in GG. (C) MTS cross-linking of GG double cysteine mutants. Cross-linkers were incubated with GG and analyzed by SDS-PAGE and Coomassie staining as described in Materials and Methods.

Self-assembly properties of interface mutants. (A) Sedimentation of interface mutants (1 μM) in the absence of liposomes. After sedimentation, supernatant (S) and pellet (P) fractions were analyzed by SDS-PAGE. T, total amount of protein in the initial mixture. (B) Sedimentation of interface mutants (1 μM) in the presence of 0.1-μm PIP₂-containing liposomes (300 μM). (C) Representative electron micrographs of interface mutants assembled on 0.4-μm PIP₂-containing liposomes. Scale bar, 200 nm.

Class II interface mutants disrupt clathrin-mediated endocytosis in vivo. (A) Diagram of endocytosis intermediates and their accessibility to the avidin and MesNa probes. (B and C) Dynamin-2 knockout (Dyn2 KO) cells expressing either GFP (○) or the following GFP-tagged Dyn2 constructs: WT (•), L12N (■), F20N (▴), or A738N (♦) were generated as described in Materials and Methods. Endocytosis in these cells was quantified using biotinylated transferrin (BSS-Tfn) as the ligand and measuring its uptake into an avidin- (B) or MesNa- (C) inaccessible compartment. The amount of uptake is expressed as a percentage of the total surface-bound BSS-Tfn.

L12N increases the number of late endocytic intermediates in reconstituted Dyn2 KO cells. (A) Thin sections of Dyn2 KO cells reconstituted with GFP-tagged Dyn2 proteins (WT, L12N, and F20N and A738N, as indicated) were prepared and imaged as described in Materials and Methods. Scale bar, 200 nm. Endocytic intermediates were counted and scored as shallow (S), invaginated (I), or constricted (C) CCPs as indicated. (B) Quantitation of CCP intermediates expressed as % of total CCPs scored (n = 79, 90, 87, and 61 for WT, L12N, F20N, and A738N, respectively). **p < 0.01 in t test.

Design and functional characterization of a minimal GTPase-GED fusion (GG). (A) Design of GG fusion. Dashed line represents flexible linker connecting the GTPase and GED fragments. (B) Diagram depicting the putative folding of GG. The crystal structure of the rat dynamin-myosin fusion (PDB: 2aka), which was used as a model to guide the design, is shown below for comparison. (C) Solubility of human dynamin1 GTPase domain in the absence (left) or presence (right) of the C-terminal GED helix. Proteins were expressed in E. coli as MBP fusions. Black arrows indicate bands corresponding to MBP-GTPase and MBP-GG, respectively. (D) Purification of GG from E. coli. GG was expressed as an MBP fusion and purified by a three-column procedure as described in Materials and Methods. Lane 1, soluble fraction following cell lysis; lane 2: pooled amylose column elutions; lane 3, sample from lane 2 after cleavage of MBP tag by PreScission protease; lane 4, purified GG after anion exchange and size exclusion chromatography to remove the MBP. Arrows denote the bands corresponding to MBP-GG fusion, MBP, and GG. (E) GTPase activities of purified GG and GG S45N. Proteins were assayed at a concentration of 2 μM in buffer containing 20 mM HEPES, pH 7.5, 150 mM KCl, 1 mM MgCl₂ and 500 μM GTP. The activity of HA-tagged full-length dynamin-1 under the same reaction conditions is shown for comparison.

Basal and assembly-stimulated GTPase activities of interface mutants in vitro. Normalized activity for each mutant is expressed as a percentage of wild-type activity. All mutations were assayed in the context of full-length dynamin at a concentration of 0.5 μM in buffer containing 20 mM HEPES, pH 7.5, 150 mM KCl, 1 mM MgCl₂, and 500 μM GTP. Activity was measured on the soluble fraction of each sample. The data presented here derive from an initial screen aimed at identifying potentially interesting interface mutant phenotypes.

Membrane fission by class II interface mutants in vitro. Fluorescent SUPER templates were prepared as described in Materials and Methods. ■, fission by 0.5 μM dynamin in the absence of nucleotide; □, fission in the presence of 1 mM GTP.

L12N alters the morphology of clathrin-coated pits in reconstituted Dyn2 KO cells. Clathrin-coated pits were visualized by indirect immunofluorescence using the AP6 mAb directed against the α-adaptins of AP2. Dynamin-2 constructs (WT, L12N, F20N, or A738N) were localized using GFP fluorescence. White boxes in each merged image indicate sections magnified in the panels on the right.