Allosteric control of dynamin-related protein 1 through a disordered C-terminal Short Linear Motif

Article https://doi.org/10.1038/s41467-023-44413-6 Allosteric control of dynamin-related protein 1 through a disordered C-terminal Short Linear Motif Isabel Pérez-Jover 1,2,9 , Kristy Rochon 3,9 ,DiHu 4,9 , Mukesh Mahajan 4 , Pooja Madan Mohan 4 , Isaac Santos-Pérez 5 , Julene Ormaetxea Gisasola 1,2 , Juan Manuel Martinez Galvez 1,2 , Jon Agirre 6 ,XinQi 4,7 ,JasonA.Mears 3,7,8 , Anna V. Shnyrova 1,2 & Rajesh Ramachandran 4,8 The mechanochemical GTPase dynamin-related protein 1 (Drp1) catalyzes mitochondrial and peroxisomal fission, but the regulatory mechanisms remain ambiguous. Here we find that a conserved, intrinsically disordered, six-residue Short Linear Motif at the extreme Drp1 C-terminus, named CT-SLiM, constitutes a critical allosteric site that controls Drp1 structure and function in vitro and in vivo. Extension of the CT-SLiM by non-native residues, or its interaction with the protein partner GIPC-1, constrains Drp1 subunit conformational dynamics, alters self-assembly properties, and limits cooperative GTP hydrolysis, surprisingly leading to the fission of model membranes in vitro. In vivo, the involvement of the native CT-SLiM is critical for productive mitochondrial and peroxisomal fission, as both deletion and non-native extension of the CT-SLiM severely impair their progression. Thus, contrary to prevailing models, Drp1-catalyzed membrane fissionreliesonallostericcommunication mediated by the CT-SLiM, deceleration of GTPase activity, and coupled changes in subunit architecture and assembly-disassembly dynamics. Intrinsically disordered proteins (IDPs) and structured proteins that contain intrinsically disordered regions (IDRs) are ubiquitous comprising nearly half of the human proteome1–3. In contrast to the relatively stationary loops and turns that connect secondary structure elements in compactly folded protein domains, IDRs persist as a highly dynamic conformational ensemble, which in many instances undergoes a localized disorder-to-order structural transition upon partner interactions (withprotein, lipid, or nucleotide) and/or via various posttranslational modifications (PTMs)4–6. Consequently, IDRs function as regulatory nodes or hubs that govern host protein function by transcribing biological information from multiple interactions and modifications into discernible alterations in local protein fold, dynamics, and macromolecular assembly, including protein condensation via liquid-liquid phase separation (LLPS)7–10. This generalized description of IDR structure and function also pertains to dynamin-related protein 1 (Drp1), a self-assembling, multidomain GTPase that mechanochemically constricts tubularmembrane intermediates en route to mitochondrial fission11,12. Drp1 contains Received: 2 July 2023 Accepted: 7 December 2023 Check for updates 1 Department of Biochemistry and Molecular Biology, University of the Basque Country, 48940 Leioa, Spain. 2 Instituto Biofisika, CSIC, UPV/EHU, 48940 Leioa, Spain. 3 Department of Pharmacology, Case Western Reserve University School of Medicine, Cleveland, OH 44106, USA. 4 Department of Physiology and Biophysics, Case Western Reserve University School of Medicine, Cleveland, OH 44106, USA. 5 Electron Microscopy and Crystallography Center for Cooperative Research in Biosciences (CIC bioGUNE), Bizkaia Science and Technology, Park Bld 800, 48160-Derio Bizkaia, Spain. 6 York Structural Biology Laboratory, Department of Chemistry, University of York, Heslington, YO10 5DD York, UK. 7 Center for Mitochondrial Diseases, Case Western Reserve University School of Medicine, Cleveland, OH 44106, USA. 8 Cleveland Center for Membrane and Structural Biology, Case Western Reserve University School of Medicine, Cleveland, OH 44106, USA. 9 These authors contributed equally: Isabel Pérez-Jover, Kristy Rochon, Di Hu. e-mail: [email protected]; [email protected] Nature Communications | (2024) 15:52 1 1234567890():,; 1234567890():,; multiple IDRs, ranging up to 134 amino acid (aa) residues in length, which make up >20% of its overall sequence11,13,14 (Supplementary Fig. 1a). Yet, many of these IDRs are either absent or unresolved in any available Drp1 X-ray13 or cryo-EM structure15,16 to date (Fig. 1aand Supplementary Fig. 1b), obscuring further functional characterization. These include: (i) Molecular Recognition Features (MoRFs)8,10 of ~10–25 aa residues, such as MoRFs-1 and −2 embedded within the largely disordered variable domain (VD), enabling direct Drp1membrane interactions17 and (ii) Short Linear Motifs (SLiMs)8,10 of ~3–12 aa residues nested within highly structured domains, such as the G-domain ‘80-loop’and stalk ‘L1N’loop that direct protein-protein interactions during Drp1 self-assembly14. One such IDR is a unique stretchof ~6 aa residues at the Drp1 extreme C-terminus, which wecall the CT-SLiM (Fig. 1a and Supplementary Fig. 1a, b) that, unlike other Volume (mL) Molar mass (kDa) 0 100 200 300 400 500 11 12 13 14 15 16 17 WT CT+ 'CT4 'CT6 4-mer 2-mer SEC-MALS 5 PM protein at injection ~0.5 PM peak protein at the detector 0% 10% 20% 30% 40% 50% 60% Helical length (nm) WT 118.6 ± 39.0 (3.5) n=124 CT+ 192.3 ± 104.9 (9.1) n=134 Mean ± SD (SEM) b cd e + GMP-PCP WT CT+ 'CT4 'CT6 200 nm 50 nm a M1 (N-term) I693 (C-term) BSE GTPase (G) domain Drp1'VD crystal structure (PDB ID: 4BEJ) “CT-SLiM” Drp1 (iso3; 699 aa) BSE GTPase BSE Middle GED BSE NC VD RETHLW 699694 “CT-SLiM” Hinge 1 Hinge 1 R403 % of total 0% 10% 20% 30% 40% WT 46.9 ± 7.0 (0.7) n=84 CT+ 49.5 ± 7.0 (0.7) n=91 Mean ± SD (SEM) Ring diameter (nm) 0% 10% 20% 30% 40% 50% % of total 'CT4 29.2 ± 6.0 (0.5) n=110 'CT6 32.4 ± 8.0 (0.4) n=317 Ring diameter (nm) Article https://doi.org/10.1038/s41467-023-44413-6 Nature Communications | (2024) 15:52 2 IDRs in Drp1, is highly conserved among metazoans (Supplementary Fig. 1a). However, its function(s) remain largely unexplored. Recent studies have indicated that this CT-SLiM constitutes an atypical PDZ domain binding motif (PBM) that specifically interacts with the PDZ domain-containing adaptor protein GIPC-1 (GAIP interacting protein, C-terminus 1)18,19. CT-SLiM-bound GIPC-1, in turn, associates with the F-actin minus-end-directed motor myosin VI (MYO6) to guide Drp1 presumably to F-actin-pre-constricted mitochondrial division sites18–21. However, whether or how direct GIPC-1Drp1 interactions via the CT-SLiM influence Drp1 structure and/or function remains unknown. Here, using a comprehensive toolkit of structural, cell biological, and in vitro reconstitution approaches, we show that a deletion (ΔCT) or a non-native extension (CT + ) of the CT-SLiM distinctly alters Drp1 conformational dynamics, oligomerization propensity, self-assembly geometry, andcooperative GTPaseactivity,inaddition to differentially affecting Drp1 capacity to remodel target membranes. We demonstrate that whereas the ΔCT variants exhibit a predictable loss-offunction by either altering or inhibiting membrane fission both in vitro and in vivo, the CT+ variants display an aberrant gain-of-function by robustly catalyzing membrane fission in vitro, while remaining repressed in mediating mitochondrial fission in vivo. By contrast, WT Drp1, which is limited to constricting membranes on its own in vitro, remarkably progresses toward membrane fissionuponnativeCTSLiM-effected GIPC-1 interactions. Taken together, our data indicate a critical role for the native CT-SLiM in governing Drp1 structure, conformational dynamics, andmechanoenzymatic membrane remodeling activity. Furthermore, key protein partner interactions of Drp1, such as that of the CT-SLiM, emerge as an essential regulatory element in the allosteric control of Drp1 function during mitochondrial fission. Results CT-SLiM modifications alter Drp1 self-assembly propensity and geometry To discern the role of the CT-SLiM, we generated a host of Drp1 variants with either truncated or extended C-termini (Supplementary Fig. 1c). The truncated variants had the last four (ΔCT4) or six (ΔCT6) residues of the native CT-SLiM removed in order to separate the potential electrostatic (R694E695) and hydrophobic (L698W699)contributions of this segment to Drp1-partner protein interactions18.Conversely, the extended variants (CT + ) had non-native sequences of different lengths and composition, including affinity and epitope tags, appended to the CT-SLiM. These non-native extensions were introduced to isolate the backbone carboxylate moiety of the C-terminal residue, a requirement for high-affinity PDZ domain binding22, from the predicted PDZ domain recognition sequence in Drp1 (696THLW699)18. In addition, we sought to determine the influence of non-native C-terminal extensions on Drp1 structure and function, which in past studies have produced confounding and conflicting results23–25.Toenablepurification, WT Drp1 andselectCTvariants were modified at the N-terminus with a His 6 affinity tag (see Methods), which as previously shown26,27 did not affect Drp1 self-assembly or GTPase activity in vitro. Besides, N-terminally epitope (Myc)-tagged Drp1 effectively restored mitochondrial fission in Drp1 knockout (KO) cells17,26 indicating that these N-terminal modifications neither affect Drp1 function in vivo. Negative-stain electron microscopy (NS-EM) analysis revealed considerable alterations in the Drp1 oligomer structure due to the CTSLiM modifications. In the presence of the non-hydrolyzable GTP analogue, GMP-PCP, which mimics GTP binding and promotes Drp1 helical self-assembly in solution26, WT Drp1 characteristically formed a mixture of oligomeric rings and higher-order spirals of a consistent diameter and length (Fig. 1b–d). In contrast, the ΔCT4 and ΔCT6 variants failed to assemble into any such regular higher-order structures. Instead, the ΔCT4 and ΔCT6 variants predominantly constituted triangularly shaped nubs of much smaller dimensions with little to no indication of further higher-order self-assembly (Fig. 1b, c). Conversely, the CT+ variant formed consistently longer supramolecular helical assemblies, although similar in overall helical diameter to WT (Fig. 1b–d). Size-exclusion chromatography-coupled multi-angle light scattering (SEC-MALS) analyses of these variants in the nucleotide-free apo state at physiologically relevant concentrations in solution28 (~0.5µM at peak detection upon ~10-fold SEC dilution) revealed further differences in their oligomerization properties relative to WT (Fig. 1e). The ΔCT4 and ΔCT6 variants exhibited a sharp dimertetramer equilibrium similar to WT, albeit tending marginally toward minimal dimers under the conditions. In contrast, the extended CT+ variant largely favored higher-order oligomers consistent with its enhanced helical self-assembly in the presence of GMP-PCP (Fig. 1b, Supplementary Fig. 2a, b). This greater oligomerization propensity of the CT+ variant relative to WT was evident over a wide range of protein concentrations (Supplementary Fig. 2c). Besides, it was independent of the non-native CT+ sequence as this tendency was also manifest in a CT+*variant containing an extension of a different length (14 aa residues) and composition (Supplementary Fig. 2d). Shortening the non-native CT sequence of the CT+ variant from 24 to 9 aa residues by proteolytic cleavage (CT+sh) reduced its higher-order oligomerization propensity (Supplementary Fig. 2e), although this remained noticeably greater than that of WT Drp1. On the other hand, shortening the N-terminal tag sequence from 36 to 7 aa residues had no palpable effect on Drp1 oligomerization (Supplementary Fig. 2f). These data indicated thatthe disordered Drp1 CT-SLiM is a critical determinant of Drp1 self-assembly and helical propagation. CT-SLiM modifications alter Drp1 conformational dynamics and structure To gain insight into the molecular mechanisms underlying CT-SLiM function, we used NS-EM and performed 2D image classification to assess the impact of the various CT modifications on Drp1 subunit Fig. 1 | CT-SLiM modifications affect Drp1 oligomerization propensity and helical geometry. a The location and polypeptide sequence of the CT-SLiM in Drp1 isoform 3 primary structure is shown. The crystalstructure beneath corresponds to the Drp1ΔVD dimer with a color-coded representation of domain arrangement in a monomer. BSE (purple) is the bundle signaling element. The stalk comprises a fourhelical bundle composed of discontinuous middle (blue) and GED (GTPase effector domain; orange) regions. GTPase (G) domain is shown in green. Connecting black lines represent a few prominent IDRs in Drp1. The VD connects the middle and GED regions, whereas the 80-loop and LIN loops are nested within the G and stalk (middle) domains, respectively. The inset is a zoomed-in view of the BSE showing the well-resolved Drp1 N-terminal BSE helix (beginning from aa residue 1). The last six residues of the Drp1 C-terminus (R694-W699), an IDR which we call the CT-SLiM and represented here by a curved black line, remain disordered. I693, the last resolvedresidue of the C-terminal BSE helix is highlighted.bRepresentative NS-EM images of WT Drp1 and CT variants in the presence of the non-hydrolyzable GTP analogue, GMP-PCP. Scale bar, 200 nm. The left inset under each panel shows a zoomed-in view of the boxed region in the above micrograph, whereas 2D class averages of the predominant oligomer (ring) morphology are shown to their right. Insets scale bar, 50 nm. Data shown here are for mouse CT+ Drp1. Human CT+ Drp1 data are shown in Supplementary Fig. 2a. Histograms showing the distribution of assessed ring diameter (c) and helical polymer length (d) for WT Drp1 and CT variants. ΔCT4/6 Drp1 do not form helical polymers. Mean ± SD (SEM) is indicated. n is the number of particles. eSEC-MALS elution and molar mass profiles of human WT Drp1 and CT variants sieved through a Superose 6 10/300 GL column. When injected at 5 µM, Drp1 is diluted to ~0.5 µM peak concentration on arrival at the LS and dRI detectors. Horizontal lines indicate the theoretical masses of a Drp1 dimer (2-mer) and tetramer (4-mer). Article https://doi.org/10.1038/s41467-023-44413-6 Nature Communications | (2024) 15:52 3 conformational dynamics and self-assembly under different nucleotide-bound states (Fig. 2a, Supplementary Fig. 3a, b). In the apo state, we detected two different orientations for the WT Drp1 dimer—an S-shaped top-down (or bottom-up) orientation and a V-shaped side-on orientation with prominent densities evident for the dimeric stalk and the two individual GTPase (G) domains (Fig. 2a, Supplementary Fig. 3b, c). Notably, in the S-shaped orientation, the G domains were set far apart, whereas in the V-shaped orientation, the G domains appeared to be positioned in close proximity. Remarkably, the S-shaped structure was never found for the CT+ variant in the apo state and was relatively poorly sampled by the ΔCT4/6 variants (Fig. 2a, Supplementary Fig. 3b, c). As differential grid deposition or preferred orientations were unlikely to be influential factors owing to the minimal nature of the CT modifications involved (CT-SLiM deletion or a short non-native extension), we reasoned that these two orientations likely correspond to two different solution conformations of the Drp1 dimer that interconvert dynamically (Fig. 2a, Supplementary Fig. 3c). W90 only 330 W552 only 332 W699 only 337 WT 337 Omax (Trp) 0 0.2 0.4 0.6 0.8 1 320 340 360 380 400 Wavelength (nm) Normalized Trp emission intensity W699F 329 W552F 332 W90F 341 WT 337 Omax (Trp) 0 0.2 0.4 0.6 0.8 1 320 340 360 380 400 Wavelength (nm) CT+ 332 WT 337 Omax (Trp) 0 0.2 0.4 0.6 0.8 1 320 340 360 380 400 Wavelength (nm) a b d c WT ΔCT6 G G S G G S G G S S GG G SAXS-derived envelope of a Drp1 dimer (R403A) 90o 90o180o Side-on view Top-down view Bottom-up view Top Bottom G S G G S 0% 20% 40% 60% 80% 100% Extended Compact 0% 20% 40% 60% 80% 100% Extended Compact 0% 20% 40% 60% 80% 100% Extended Compact 0% 20% 40% 60% 80% 100% Open Compact Closed Compact 0% 20% 40% 60% 80% 100% Wide Rings Collapsed Rings 0% 20% 40% 60% 80% 100% Wide Rings Collapsed Rings 0% 20% 40% 60% 80% 100% Wide Rings Collapsed Rings 0% 20% 40% 60% 80% 100% Wide Rings Collapsed Rings 0% 20% 40% 60% 80% 100% Extended Compact 0% 20% 40% 60% 80% 100% Extended Compact 0% 20% 40% 60% 80% 100% Extended Compact 0% 20% 40% 60% 80% 100% Extended Compact * Extended Compact * Extended Compact * Extended Compact +GTP * 10 nm Extended Compact Dominant species WT * Extended Compact * Extended Compact * Open Compact Closed Compact 10 nm * Extended Compact apo ΔCT4 ΔCT6 CT+ Wide Ring Collapsed Ring * Wide Ring Collapsed Ring * Wide Ring Collapsed Ring Wide Ring Collapsed Ring 20 nm +GMP-PCP ~22 nm G G S Extended S GG Compact * * GG S GG S CT-SLiM Side-on view Side-on view * Article https://doi.org/10.1038/s41467-023-44413-6 Nature Communications | (2024) 15:52 4 For the apo WT Drp1 dimer, the extended S-shaped conformer was detected at a ~4-fold greater incidence than the compact Vshaped conformer, indicating a greater residence time for the native dimer in the extended conformation (Fig. 2a). These data indicated that the native CT-SLiM restricts Drp1 conformational dynamics in solution and retains Drp1 predominantly in the extended conformation, whereas its absence or non-native extension in the ΔCT4/6 and CT+ variants, respectively, differentially relieves this auto-inhibition favoring their conversion, to varying extents,to the alternate compact conformation. Moreover, in the presence of GMP-PCP, the oligomeric rings formed by the ΔCT4/6 and CT+ variants were largely irregular or poorly ordered (Fig. 2a) suggesting that the native CT-SLiM also functions as a spacer that sets the register and geometry of intersubunit interactions during nucleotide-dependent helical selfassembly. Furthermore, unlike WT, which reverted to the extended dimer conformation upon GTP hydrolysis, the CT+ and ΔCT4/6 variants largely remained in the compact conformation (Fig. 2a, Supplementary Fig. 3d). These data indicated that the compact CT+ Drp1 conformerlikely mimics an assembly-primed state based on itsgreater higher-order oligomerization propensity relative to WT both in absence and presence of nucleotide. 3D reconstruction from 2D class averages of the extended WT and compact ΔCT6 Drp1 conformations further allowed us to dock the available crystal structure of the ΔVD Drp1 dimer and examine the nature of the conformational rearrangements (Fig. 2b, Supplementary Fig. 3c). With the extended conformation, the G domains of the docked ΔVD Drp1dimer stretchedbeyondthe computed edgedensities.These data suggested that either the ΔVD variant is in an alternate conformation compared to WT14, with the WT G domains tucked in toward the stalk as recently indicated16, or that our reconstructed structure remainspartiallyunresolved owing toinherent dynamicsaroundhinge 1 at the BSE-stalk intersection (Fig. 1a). Fitting of the ΔVD Drp1 dimer structure into the computed 3D volume ofthe compact conformation, however,required a large-scale repositioning of the G domains around hinge 1 (Fig. 2b). Modeling of these two conformations using the crystal structure (Fig. 2b) and back projection of 2D class averages from the computed 3D volumes (Fig. 2b) revealed that the compact conformation sampled by ΔCT6 Drp1 is not observed in any of the projected 2D class averages for the extended WT Drp1 dimer. These data indicated that the observed extended and compact Drp1 forms are indeed conformationally distinct. Toconfirm the large-scale flexibility of the Drp1 dimer as indicated by the EM data, we mapped the conformational landscape of a minimal Drp1dimerinsolution using small-angle X-rayscattering(SAXS)29 as an orthogonal approach (Fig. 2c and Supplementary Fig. 4). The heterogenous mix of dimers, tetramers, and higher-order oligomers present in dynamic equilibrium for WT Drp1 and the CT variants is incompatible with SAXS and cannot be analyzed. Therefore, we employed a R403A mutation in Drp1 (R399A in Dyn130)(Fig.1a) that restricts Drp1 predominantly to a minimal dimer in solution (Supplementary Fig. 4a), similar to the minimal Dyn1 dimer previously assessed by SAXS31. Remarkably, ab initio reconstruction of the most probable lowresolution molecular envelope for the R403A Drp1 dimer revealed an overall shape that was compatible with both the extended and compact conformations (Fig. 2c), with sufficient volume present between the G domains and below the stalk of the overlaid ΔVD Drp1 dimer crystalstructure toaccommodate both shapes.Thus,the minimal Drp1 dimer in solution is highly dynamic and capable of interconversion between extended and compact states. To understand the molecular basis of the CT+ Drp1 variant’s distinctively compact conformation and gain-of-function in self-assembly, we used AlphaFold32 to predict the influence of the CT+ sequence extension on Drp1 structure. Remarkably, the computational data suggested that whereas the N-terminal His 6 affinity tag in our WT Drp1 was mostly disordered, the non-native CT extension in CT+ Drp1 propagated as a α-helixincloseappositiontothetopoftheGdomain, potentially constraining dynamics at the adjacent nucleotide-sensitive Gdomain-BSE interface (Supplementary Fig.5a).Consistentwiththis,a direct comparison of the top-ranked structures in isolation (Supplementary Fig. 5b) and upon superposition into the available Drp1 polymer cryo-EM structure (Supplementary Fig. 5c), revealed a slight inward buckling of the G domain toward the BSE in CT+ Drp1 compared to WT Drp1. In addition, given the proximity of the CT-SLiM to the stalk of the adjacent monomer in the Drp1 polymer (Supplementary Fig. 5c), the modeling data further indicated that the CT+ extension may influence Drp1 subunit-subunit interactions during higherorder helical self-assembly. We used intrinsic Tryptophan (Trp) Fluorescence Spectroscopy (iTFS)33,34 to experimentally validate these in silico predictions (Fig. 2d). Trp emission is highly sensitive to its microenvironment and therefore serves as an accurate probe of protein conformation or conformational changes25,26. When excited selectively at λ=295nm, the Trp emission spectrum is blue-shifted (peaking at shorter wavelengths) when present in a nonpolar environment, and red-shifted (peaking at longer wavelengths) when exposed to a polar or aqueous milieu. Drp1 contains three native Trp residues at positions 90, 552 and 699 (ubiquitous isoform 3 numbering; Fig. 1a). Of these, only W90 present in the G domain is structurally resolved13 (Supplementary Fig. 5d), whereas W552 and W699 are located in the disordered VD and CT-SLiM, respectively (Supplementary Fig. 1a). Using site-directed Drp1 mutants that retained only one of the three native Trp or that contained only a single native Trp-to-Phe substitution, we ascertained that Drp1 Trp emission primarily originates from W699, the terminal residue of the CT-SLiM. Consistent with the partial burial of W90 in the Drp1 G domain structure13 (Supplementary Fig. 5d), the W90-only mutant displayed a pronounced blue shift in Trp emission relative to WT (Fig. 2d). Similarly,theW552-onlymutantalso exhibitedasignificantblueshift, albeit less than that of the W90-only mutant, indicating that W552 is also partially occluded from solvent in the VD conformational ensemble (Fig. 2d). By contrast, the W699-only mutant was pronouncedly redshifted and was identical to WT in emission spectra (Fig. 2d). These data indicated that W699 in WT Drp1 is solvent accessible, and is the primary emitter largely owing to its location within Trp-Trp homoFig. 2 | CT-SLiM modifications alter Drp1 conformational dynamics. a NS-EM 2D class averages of dimers and oligomeric rings in the apo, GMP-PCP-bound,andGTP hydrolysis states for human WT Drp1 and CT variants. Drp1 dimer conformation in the apo and GTP hydrolysis states is classified as either extended or compact, with the latter classified further into open-compact or closed-compact states. Oligomeric ring morphology in the presence of GMP-PCP is classified into wide and collapsed ring states. Top-down and side-on views of the ΔVD Drp1 dimer crystal structure (PDB ID: 4BEJ) as well as color-coded cartoon illustrations of domain rearrangements between the extended and compact states are shown at the farright corner. G refers to the G domain, whereas S refers to the stalk. The number of particles in the extended (E) and compact (C) conformations in the apo and +GTP states, and displaying Wide Ring (WR) and Collapsed Ring(CR) morphologies in the +GMP-PCP state are shown in Methods. bi) WT and ΔCT6 Drp1 3D densities rendered from cryoSPARC homogeneous refinement processing, ii) WT and ΔCT6 Drp1 idealized models generated from the ΔVD Drp1 dimer crystal structure (PDB ID: 4BEJ) by thresholding the structure to 40Å for WT and by repositioning the G domains and thresholding to 40 Å for ΔCT6 Drp1, and iii) 2D class averages generated from the WT and ΔCT6 Drp1 idealized models are shown in rows. cViews of the SAXS-derived ab initio envelope of dimeric R403A Drp1 overlaid with the ΔVD Drp1 dimer crystal structure (PDB ID: 4BEJ). dNormalized Trp emission spectra of human WT Drp1 in comparison to single W-only mutants, single W-to-F mutants, and the CT+ Drp1 variant are shown, respectively, from left to right. The wavelength of maximum emission (λ max )isindicated. Article https://doi.org/10.1038/s41467-023-44413-6 Nature Communications | (2024) 15:52 5 FRET distance (~24 Å)33 of the high-energy FRET donor, W90 (Supplementary Fig. 5d). Consistent with this assessment, the W90F mutant wasred-shifted (by 11 nm) compared to the W90-only mutant, whereas the W699F mutant was blue-shifted (by 8 nm) relative to the W699only mutant (Fig. 2d). The W552F mutant, on the other hand, did not experience any such change (Fig. 2d). These data confirmed that W699 in the WT Drp1 CT-SLiM is exposed to water and is highly responsive to its local environment. Notably, by contrast to WT Drp1, CT+ Drp1 emission was blue-shifted (by 5nm) indicating that W699 in CT+ Drp1 is instead buried and relatively solvent inaccessible (Fig. 2d). No such change in Trp emission was observed when the 36 aa-residue Nterminal His 6 tag of WT Drp1 was replaced by a relatively short 7-aa residue overhang (Supplementary Fig. 5e), indicating that the difference in the environment of CT-SLiM in CT+ Drp1 is primarily due to the non-native CT extension. Thus, together with the cryo-EM data and AlphaFold predictions, the iTFS data demonstrated that non-native CT extension of CT+ Drp1 alters CT-SLiM microenvironment and overall Drp1 conformation. WT and CT+ Drp1 thus populate distinct conformational states. CT-SLiM modifications differentially affect Drp1 cooperative GTP hydrolysis We next determined the impact of the various CT-SLiM modifications on Drp1 GTPase activity under basal conditions in solution and upon helical self-assembly on CL-containing membranes. In the apo state, the CT variants retained the characteristic capacity of WT Drp1 to selfassemble on, and tubulate, large CL-containing liposomes to narrow diameters (Supplementary Fig. 6a, b). Similarly, the CT variants also assembled on highly curved and preformed galactosylceramide-laden CL-containing lipid nanotubes (GalCer CL-NTs) identically to WT (Supplementary Fig. 6c). These data indicated that the various CT modifications do not adversely affect stalk-mediated Drp1 selfassembly on membranes. Surprisingly, however, the ΔCT4 and ΔCT6 variants both exhibited a ~3-fold greater rate of GTP hydrolysis in solution compared to the CT+ variant and WT, which were similar in basal GTPase activity (Fig. 3a). By contrast, the CT+ variant displayed a ~2-fold lower activity compared to the ΔCT4/6 variants and WT when assayed on CLcontaining liposomes (Fig. 3b). Analysis of the pre-steady state ‘burst’ kinetics revealed that the ΔCT4/6 variants hydrolyzed GTP at a significantly faster rate than WT under both conditions (Fig. 3c). Similar trends also held up for the CT variants on GalCer CL-NTs on which ΔCT4/6Drp1hydrolyzedGTP atasignificantlyfasterratethanbothWT and CT+ Drp1 in the order: ΔCT6 > ΔCT4 > WT > CT+ (Supplementary Fig. 6d, e). Co-sedimentation analysis of the Drp1 variants on GalCer CL-NTs, in the absence and presence of GTP, demonstrated a greater steady-state association of the faster GTP-hydrolyzing ΔCT4/6 variants with the lipid templates than WT or CT+ Drp1 (Supplementary Fig. 6f). Consistent with this, NS-EM on GalCer CL-NTs in the presence of GTP revealed persistent self-assembly of the ΔCT6 variant on the lipid templates. Such phenotype was absent for both WT and CT+ Drp1, which showed widespread disassembly and membrane dissociation with GTP (Supplementary Fig.7). Interestingly,in thepresenceofGMPPCP, the ΔCT6 variant, relative to WT and CT+ Drp1, formed highly processive helical polymers, which in many instances extended beyond the lipid template (Supplementary Fig. 7). Collectively, these data indicated that in the absence of the CTSLiM, transition state-dependent inter-subunit G-domain dimerization, cooperative GTP hydrolysis, GDP/Pi release, and G-domain dimer disassembly critical for progressive rounds of GTP binding and hydrolysis are all significantly accelerated for the ΔCT4/6 variants, manifested in faster recycling and greater steady-state association with membranes. Conversely, for the same reasons, in the presence of a non-native CT extension that non-physiologically stabilizes inter-subunit Drp1 interactions and exaggeratedly promotes helical self-assembly in CT+ Drp1, GTP turnover and recycling on membranes appears to be decreased. Thus, faster dynamics in the absence of the CT-SLiM, and altered, slower dynamics in the presence of a non-native CT extension distinctively affect Drp1 cooperative GTPase activity relative to WT. Together with its impact on Drp1 dimer structure, these data therefore raised the intriguing prospect that the native CT-SLiM functionsasa‘kinetic timer’of Drp1’s GTP hydrolysis rate and coupled membrane remodeling activity. CT-SLiM controls Drp1-catalyzed membrane fission in vitro We therefore addressed whether the differential GTPase activity, and altered conformational and self-assembly dynamics of the CT variants relative to WT translated to distinct membrane remodeling abc 0 20 40 60 80 100 120 140 02468 10 12 Time (min) P released (PM) i WT 21.0 ± 0.5 CT+ 13.0 ± 0.6 'CT4 22.5 ± 0.1 'CT6 25.2 ± 1.1 Average (mink cat -1 ) CL-stimulated GTPase activity 0 20 40 60 80 100 120 140 0 102030405060 Basal GTPase activity 'CT4 5.1 ± 0.1 'CT6 5.2 ± 0.1 WT 2.0 ± 0.1 CT+ 1.8 ± 0.1 Average (mink cat -1 ) P released (PM) i Time (min) 0 10 20 30 40 50 WT CT+'CT4'CT6 (mink cat -1 ) Burst phase kinetics WT 2.4 ± 0.1 32.2 ± 0.5 CT+ 1.9 ± 0.0 18.8 ± 0.8 'CT4 7.3 ± 0.1 36.9 ± 0.3 'CT6 7.9 ± 0.1 47.2 ± 1.8 Basal CL-stimulated Fig. 3 | CT-SLiM modifications differentially affect Drp1 GTPase activity. Relative basal (a) and CL-stimulated GTPase activities (b) of WT Drp1 and CT variants at 0.5 µM each in the absence and presence of CL-containing liposomes (150µMtotal lipid). GTPase activities of the CT variants were measured in parallel in comparison to WT from two independent protein alone or protein-lipid mixture samples (n=2 measurements for each variant under either condition; n= 3 and 5 for WT basal and CL-stimulated activities, respectively). The concentration of inorganic phosphate (P i ) released is plotted against time. Individual data points and best fit traces for each time course are shown. In aand b, the average turnover number (k cat ) ± SEM derived from a linear regression analysis of both pre-steady state (burst phase; boxed regions) and steady-state kinetics data points is indicated above. ck cat for the burst-phase pre-steady-state kinetics only from each measurement. Individual data points are overlaid on bar plots representing the average. Burst-phase k cat ± SEM is indicated above. Article https://doi.org/10.1038/s41467-023-44413-6 Nature Communications | (2024) 15:52 6 phenotypes. To this end, we tested the efficacy of our CT variants in directing the scission of suspended lipid nanotubes (NTs) mimicking the mitochondrial outer membrane at pre-constricted mitochondrial division sites. NTs ranging from tens to hundreds of nanometers in diameter were formed between polymer micropillars in a microfluidic chamber35 (see Methods). WT Drp1 and CT variants at 0.5 µMfinal concentration, corresponding to the estimated cytosolic concentration of Drp128, as well as the concentration at which the catalytic activity (k cat ) of WT Drp1 nears saturation26, were then infused into the chamber in the presence of 1 mM GTP, while NT constriction and/or scission was monitored in real-time by fluorescence microscopy. As previously shown17,25, WT Drp1 did not effectively catalyze NT fission on its own (Fig. 4a, b). Instead, WT scaffold assembly on the NT resulted in NT constriction to a stable final radius of 14 ± 2 nm (measured at the membrane midplane) independent of the initial NT radii (Fig. 4a, Supplementary Movie 1). Surprisingly, both ΔCT4 and ΔCT6 Drp1 selectively mediated the fission of NTs with the highest initial curvatures (i.e., with radii < ~30 nm), with the ΔCT4 variant exhibiting the greater fission efficiency of the two (Fig. 4a, b, Supplementary Movies 2, 3). For both variants, the area of membrane constriction prior to fission appeared to be highly limited and narrow, being barely detectable by fluorescence microscopy (Fig. 4a). Thus, a partial or complete deletion of the CT-SLiM limits Drp1 scaffolding on membranes, probably due to the greater GTP hydrolysis rate of the ΔCT4/6 variants (Fig. 3and Supplementary Fig. 6d, e), causing rapid oligomer disassembly and recycling on membranes (Supplementary Fig. 7). In stark contrast to the ΔCT4/6 deletion variants, the CT+ extension variants elicited a robust constriction and fission of a broad range of initial NT curvatures (Fig. 4a, b, Supplementary Fig. 8a, Supplementary Movie 4). Notably, the scission efficiencies of these variants directly corresponded to their higher-order oligomerization propensities with CT+ ≈CT + * > CT+sh (Supplementary Figs. 8a, 2d, e). Thus, the markedly improved stability of the CT+ variants on membranes directly correlated with their improved membrane fission activities. Remarkably, fission efficiency was directly proportional to the preponderance of the assembly-primed, compact dimer conformer in solution in the presence of GTP, sampled almost exclusively by the CT+ variants, but not WT (Fig. 2a). Importantly, contrary to prevailing models, membrane fission activity was inversely correlated with the assembly-stimulated GTP hydrolysis rate on membranes, with CT+ variants of lower GTPase activity being more efficient in fission (Supplementary Fig. 8a, b). To further assess the impact of the CT-SLiM modifications and imposed structural alterations on membrane remodeling, we used cryo-EM to analyze the self-assembly of WT Drp1 and CT variants on preformed membrane NTs in the constant presence ofGTP (Fig. 4c). In agreementwiththereal-time fluorescence measurements, the cryo-EM data revealed that WT Drp1 formed organized helical polymers that constrictedtheNTsto aradius of ~15 nm.Bycontrast, CT+Drp1formed disorganized, fuzzy coats that further constricted the membranes to critical radii of <7 nm, frequently resulting in fission and consequent retraction of the cut NTs to the membrane reservoirs located on the EM grid. Interestingly, ΔCT4 Drp1 displayed helical polymers with highly variable diameter (Fig. 4c) consistent with a near complete loss of CT-SLiM-imposed inter-subunit helical register and polymer geometry. Notably, under these conditions, WT Drp1 polymers were observed on both highly curved and relatively flat membrane regions, whereas the CT+ variant was curvature-selective with an acute preference for binding the curved NTs (Fig. 4c). Together, these data indicate thatthe CT-SLiM governs both Drp1 polymer geometry and dynamics on membranes, and that CT modifications differentially affect membrane curvature selectivity and fission activity. c ab Fig. 4 | CT-SLiM modifications differentially affect membrane remodeling and fission. a Representative kymographsshowing the remodeling of freely suspended NTs upon addition of 0.5 μM of WT Drp1 or CT variants in the presence of 1 mM GTP. NT membrane fluorescence is displayed in cyan pseudocolor for clarity. Left images correspond to the initial frame of the kymographs. Right image sequences correspond to the framed region of the kymographs. Purple arrows indicate NT fission. bDistributions ofthe radiiof free-standing NTs thatunderwent fission (yes) orwereonlyconstricted(no) upon addition of0.5μM WT Drp1 or CT variants in the presenceof 1 mM GTP. The numbers on top of each box represent the total number of NTs for each condition. Error bars are SD, n= 3 independent experiments. Box plots indicate median (middle line), 25th and 75th percentile (box) and outliers (whiskers). cCryo-EM images showing WT Drp1, CT+ Drp1, and ΔCT4 Drp1 assembled on preformed NTs in the presence of 1mM GTP. White arrowheads indicate Drp1 scaffolds on highly curved NT membranes. Black arrowhead shows curvature-adaptable assembly of WT Drp1 also on relatively flat (low curvature) membrane regions, not observed with CT+ Drp1. Article https://doi.org/10.1038/s41467-023-44413-6 Nature Communications | (2024) 15:52 7 CT-SLiM interactions with GIPC-1 regulate Drp1-mediated membrane fission Next, we determined how Drp1-GIPC-1 interactions via the CT-SLiM affect Drp1 structure, assembly, dynamics, and function. GIPC-1 contains an N-terminal IDR in addition to a centrally located PDZ domain flanked by two unique GIPC homology domains (GH1 and GH2)36 (Supplementary Fig. 9a). The GH1 and PDZ domains are involved in GIPC-1 multimerization37, whereas GH2 binds MYO6. In the absence of the N-terminal IDR and a PBM (PDZ domain ligand), GIPC-1 forms an auto-inhibited, PDZ domain-swapped dimer that occludes both the PBM and MYO6 binding sites36 (Supplementary Fig. 9b). However, consistent with a previous report37, we found that in the unliganded state, and at relatively low concentrations in solution, purified full-length GIPC-1 exists in a fast, dynamic monomer-dimer equilibrium that largely favors monomers (Supplementary Fig. 9c, d). Conversely, at very high concentrations experienced during purification, GIPC-1 also formed long filamentous sedimentable polymers in solution (Supplementary Fig. 10) indicative of its capacity to multimerize when sequestered locally. Purified GIPC-1 at low bulk concentrations in solution however remains soluble (Supplementary Fig. 10) and does not associate with, or remodel, membranes (Supplementary Fig. 10). Together, these data indicated that Drp1 CT-SLiM binding may function to relieve GIPC-1 auto-inhibition, elicit GIPC-1 multimerization, and promote cooperative Drp1-GIPC-1 co-assembly on membranes. Consistent with this notion, multimeric GIPC-1 has previously been localized to membranes37 indicating a role for the GIPC-1 multimerization in ligand protein (e.g. Drp1) confinement at target membrane sites. GIPC-1 robustly inhibited the assembly-stimulated GTPase activity of WT Drp1 on CL-containing membranes in a concentrationdependent manner (Fig. 5a, b). ΔCT6 and CT+ Drp1, by contrast, were modestly inhibited, indicating weakened binding. The modest, but considerable, inhibition for these variants also indicated the presence of additional GIPC-1 interaction sites besides the CT-SLiM (Fig. 5a, b). GIPC-1 also potently inhibited the GMP-PCP-induced selfassembly of WT Drp1 into rings and spirals in solution in NS-EM experiments, indicating that GIPC-1 binding hinders the helical propagation of Drp1 (Fig. 5c). As expected, GIPC-1 did not considerably affect the GMP-PCP-induced helical self-assembly of the CT+ variant (Fig. 5c). Nevertheless, GIPC-1 still reduced the GMP-PCP-induced formation of triangular nubs by the ΔCT6 variant probably owing to the presence of additional binding sites (Fig. 5c). Likewise, GIPC-1 potently inhibited the WT Drp1-mediated tubulation of CL-containing liposomes (Supplementary Fig. 10b). ΔCT6 ab 10 nm WT (apo) + GIPC-1 (1:4) 0% 20% 40% 60% 80% 100% Open compact Closed compact ed f 6 Pm WT WT + GIPC-1 +GMP-PCP +GMP-PCP +GIPC-1 (1:4) WT 'CT6 CT+ c WT WT + GIPC-1 NT fission probability, % ** g 4 Pm 30 s WT + GIPC-1 + GTP 1 10 (mink cat -1 ) Drp1:GIPC-1 molar ratio 1:0 1:1 1:2 1:4 'CT6 WT CL-stimulated GTPase activity CT+ Drp1:GIPC-1 molar ratio % of maximum CL-stimulated GTPase activity 1 10 100 1:0 1:1 1:2 1:4 'CT6 WT CT+ Fig. 5 | CT-SLiM interactions with GIPC-1 potentiate membrane fission in vitro. aCL-stimulated GTPase activities of WT and ΔCT6 Drp1 with increasing concentrations of GIPC-1. k cat from two independent measurements are plotted versus Drp1:GIPC-1 molar ratio. bData in areplotted as % of maximum activity for each variant. cRepresentative NS-EMimages of WT Drp1 and CT variants incubated with GMP-PCP in the absence and presence of a 1:4 molar ratio of GIPC-1. Scale bar, 100 nm. dNS-EM 2D class averages of WT Drp1 inthe apo state in the presence of a 1:4 molar ratio of GIPC-1. Only the compact conformers (open and closed) were observed for WT-Drp1 in the presence of GIPC-1. eKymograph showing NT constriction and fission (arrowhead) byWT Drp1 inthe presence ofGIPC-1 at a1:1 molar ratio (0.5 μM each) in the presence of 1 mM GTP. NT membrane fluorescence is displayed in cyan pseudocolor for clarity. fPercentage of NTs that underwent fission upon addition of either 0.5 μM WT Drp1 alone or a 1:1 mixture of WT Drp1:GIPC-1 at 0.5 μM each in the presence of 1mM GTP. Each point represents a replicate. The number on top of each column represents the total number of NTs for each condition. Mean ± SD are shown. ** Statistically different at the 0.01 level (unpaired two sample t-test, equal variance assumed). gImages showing the rigid polymerization of WT Drp1alone versus the formation of much shorter scaffolds in the equimolar presence of WT Drp1 and GIPC-1 on NTs. Images shown were acquired approximately 2minutes after the addition of the proteins at 2 μMfinal concentration each in the presence of 1mM GTP. RhPE channel is shown. Pseudocolor is used for clarity. Article https://doi.org/10.1038/s41467-023-44413-6 Nature Communications | (2024) 15:52 8 Drp1 membrane remodeling activity, conversely, was not significantly affected (Supplementary Fig. 10c). Interestingly, at equimolar concentrations (1:1) under these conditions, WT Drp1 and GIPC-1 formed amorphous assemblies in solution, whereas at higher GIPC-1 ratios (1:4), linear and bundled filaments of assembled protein in solution, reminiscent of Drp1 co-assembly with the adaptor mitochondrial dynamics protein of 49 kDa or MiD4938 were evident (Supplementary Fig. 10b). Together, these data indicated that GIPC-1 interactions via the CT-SLiM alters Drp1 self-assembly geometry, with pronounced effects on membrane remodeling as determined by the lack of ordered helical self-assembly and resultant membrane tubulation. A similar inhibition of CL-stimulated GTPase activity and helical self-assembly was observed for WT Drp1 in the presence of GIPC-1 on GalCer CL-NTs indicating that GIPC-1 regulation of Drp1 activity does not vary with membrane curvature (Supplementary Fig. 11a, b). Surprisingly, NS-EM2Dclassificationof apoWT Drp1dimers inthe presence of GIPC-1 revealed the presence of the assembly-primed, compact Drp1 conformer in solution (Fig. 5c), in contrast to the autoinhibited, extended Drp1 conformer predominantly found in GIPC-1’s absence (Fig. 2a). Additional density representing GIPC-1, however, was not readily evident reflecting either a dynamic interaction ofGIPC1 with WT Drp1 in the apo state or a substantial overlap of GIPC-1 density with the closely spaced G domains of the compact WT Drp1 conformer. In the case of ΔCT6 and CT+ Drp1, however, various extra densities and altered subunit arrangements were observed attesting to the presence of additional GIPC-1 binding sites and alternate Drp1GIPC-1 interactions (Supplementary Fig. 11c, d). The reduced assembly-stimulated GTPase activity observed for Drp1 in the presence of GIPC-1 seemingly potentiates the membrane remodeling events leading to fission, as WT Drp1 in the presence of GIPC-1 catalyzed fission in >30% of free-standing NTs over a wide range of membrane curvatures (Fig. 5e, f, Supplementary Movie 5). Interestingly, at 0.5 μM Drp1 concentration and a 1:1 GIPC-1:Drp1 ratio, we detected the formationof short Drp1 scaffolds on the NTs immediately prior to fission. Besides, these scaffolds were highly mobile on the NT surface (as observed in the kymograph in Fig. 5e, Supplementary Movie 5), suggesting that in the presence of GIPC-1, Drp1 initially assembles into small pre-curved units, and not into complete rings, on the NT surface. Thus, GIPC-1 association with WT Drp1 via the CT-SLiM appears to disengage Drp1 inter-subunit interactions that promote higher-order Drp1 self-assembly. This differential behavior of WT Drp1 in the presence of GIPC-1 was better evidenced at higher protein concentrations (2 µM, Fig. 5g). Whereas WT Drp1 alone rapidly polymerized into long and rigid scaffolds, rendering kinks in the NTs that precluded membrane fission, in GIPC´s presence the growth of the WT Drp1 scaffolds on the NTs was comparable to that of CT+ Drp1 (compare Figs. 5gand4a), and resulted in NT fission at a similar fission rate to that detected at 0.5µM protein concentration (Fig. 5f). Thus, WT-Drp1 inthe presence of GIPC1 partially mimics CT+ Drp1, which exhibits reduced GTPase activity and altered CT-SLiM interactions. CT-SLiM regulation of Drp1 is critical for mitochondrial and peroxisomal fission in vivo Drp1 catalyzes both mitochondrial and peroxisomal fission12,39–41.Yet, GIPC-1 colocalizes with the mitochondria but not considerably with peroxisomes18. To determine whether the CT-SLiM is therefore differentiallyrequired for Drp1-catalyzed mitochondrial and peroxisomal fission in vivo, we examined and compared mitochondrial and peroxisomal morphology in Drp1 KO mouse embryonic fibroblasts (MEFs) overexpressing N-terminally Myc-tagged WT and ΔCT4/6 variants, and the C-terminally Myc/FLAG-tagged CT+ variant (Fig. 6a, Supplementary Figs. 12–14). Empty vector-transfected Drp1 KO MEFs displayed extensively hyperfused mitochondria or highly elongated peroxisomes in the absence of Drp1-catalyzed mitochondrial and peroxisomal fission (Fig. 6a, Supplementary Figs. 12–14). As expected, exogenous Myc-WT Drp1 overexpression effectively rescued and restored both mitochondrial and peroxisomal fission leading to the formation of highly fragmented mitochondria and distinctly punctiform (spherical) peroxisomes (Fig. 6a–d, Supplementary Figs. 12–14). By contrast, however, the overexpression of the ΔCT4 and ΔCT6 variants had no palpable effect on the initial morphology of either organelle (Fig. 6a–d), with the great majority of cells displaying a pronounced perinuclear clustering of hyperfused mitochondria and retaining highly elongated peroxisomes (Fig. 6a–d, Supplementary Figs. 12–14). More surprisingly, the CT+ variant containing the native CT-SLiM sequence was also significantly impaired in the fission of both organelles, albeit to a lesser degree than the ΔCT4/6 variants (Fig. 6a–d, Supplementary Figs. 12–S14). Thus, in spite of their apparent gain-of-function in effecting the fission of model membranes in vitro, the CT variants appeared nevertheless perturbed in effecting organellar fission in vivo. These data further reiterated that the altered self-assembly properties and dynamics of the CT variants as manifested in vitro, and consequent impairments in effector (e.g. GIPC-1) interactions downstream are likely responsible for their organellar fission defects in vivo. Consistent with this interpretation, when overexpressed in Drp1 KO MEFs, the CT+ variant, which forms supramolecular assemblies in vitro (Fig. 1b), constituted large granular puncta in the cytosol indicative ofaggregation. In contrast, the WT and ΔCT variants exhibited a more diffuse and homogeneous distribution (Supplementary Fig. 15a, b). In co-immunoprecipitation experiments, neither overexpressed WT Drp1 nor the CT variants co-precipitated withendogenous GIPC-1(SupplementaryFig. 15c, d)indicating a highly dynamic interaction that could not differentiate the CT variants from WT Drp1 in GIPC-1 binding. These data further suggested that perturbations in GIPC-1 interactions expected of the CT variants are likely secondary to their primary defects/alterations in self-assembly and conformational dynamics. From these collective data, we conclude thatthe native CT-SLiM is a critical structural and functional determinant of Drp1-catalyzed mitochondrial and peroxisomal fission, and that its perturbations influence Drp1 function both in vitro and in vivo. Discussion Structural and functional plasticity are two interlinked characteristics of IDRs10,42,43.Thisisbestexemplified by the longest and bestrecognized IDR in Drp1, the VD, which is involved in multiple protein-protein and protein-lipid interactions via various identified MoRFs and SLiMs14. Remarkably, the VD is auto-inhibitory to premature Drp1 self-assembly in solution14,44,45, while conversely promoting Drp1 self-assembly and function upon partner interactions, specifically with target lipids on mitochondrial membranes14,17,26,46, thus reflecting the VD’s duality and functional diversity. However, the VD and various other IDRs in Drp1 (e.g. the 80-loop) are relatively poorly conserved (Supplementary Fig. 1a) and are subject to extensive tissueand organism-specificalternativesplicing 47, indicating that some of their ascribed functions may not be entirely universal. Here, we demonstrate that the highly conserved CT-SLiM, previously implicatedinDrp1transport18,isyet anothercritical,multifunctional‘toggle’ that not only governs Drp1 conformational stability and dynamics, but also directs Drp1 self-assembly, assembly geometry, and cooperative GTP hydrolysis to facilitate partner protein-guided membrane constriction and fission. Our findings have major implications for the understanding of Drp1 function and regulation. We show that the CT-SLiM is critical for ordered Drp1 self-assembly, as oligomerization of the CT variants propagates out-of-register and eventually becomes disordered. The CT-SLiM also directly impacts Drp1 dimer conformational dynamics in solution, with the CT variants more readily sampling the ‘assemblyprimed’compact conformation in contrast to the CT-SLiM-imposed, Article https://doi.org/10.1038/s41467-023-44413-6 Nature Communications | (2024) 15:52 9 60. Barylko, B. et al. The proline/arginine-rich domain is a major determinant of dynamin self-activation. Biochemistry 49, 10592–10594 (2010). 61. Ferguson,S.M.&DeCamilli,P.Dynamin,amembrane-remodelling GTPase. Nat. Rev. Mol. Cell Biol. 13,75–88 (2012). 62. Zhang, R. et al. Dynamin regulates the dynamics and mechanical strength of the actin cytoskeleton as a multifilament actin-bundling protein. Nat. Cell Biol. 22,674–688 (2020). 63. Mattila,J.P.etal.Ahemi-fission intermediate links two mechanistically distinct stages of membrane fission. Nature 524, 109–113 (2015). 64. Bos, J. L., Rehmann, H. & Wittinghofer, A. GEFs and GAPs: critical elements in the control of small G proteins. Cell 129, 865–877 (2007). 65. Gasper, R., Meyer, S., Gotthardt, K., Sirajuddin, M. & Wittinghofer, A. It takes two to tango: regulation of G proteins by dimerization. Nat. Rev. Mol. Cell Biol. 10,423–429 (2009). 66. Ramachandran, R. & Schmid, S. L. The dynamin superfamily. Curr. Biol. 28,R411–R416 (2018). 67. Crosby, D. et al. Reconstitution of human atlastin fusion activity reveals autoinhibition by the C terminus. JCellBiol221, e202107070 (2022). 68. Krishna,S.&Ford,M.G.J.Theatlastinparalogs:Thecomplexityin the tails. J. Cell Biol. 222, e202305116 (2023). 69. Osellame, L. D. et al. Cooperative and independent roles of the Drp1 adaptors Mff, MiD49 and MiD51 in mitochondrial fission. J. Cell Sci. 129,2170–2181 (2016). 70. Kalia, R. & Frost, A. Open and cut: allosteric motion and membrane fission by dynamin superfamily proteins. Mol. Biol. Cell 30, 2097–2104 (2019). 71. Stepanyants, N., Macdonald, P. J., Madan Mohan, P. & Ramachandran, R. A Single Common Protocol for the Expression and Purification of Soluble Mammalian DSPs from Escherichia coli. Methods Mol. Biol. 2159,31–40 (2020). 72. Liu, R. & Chan, D. C. The mitochondrial fission receptor Mff selectively recruits oligomerized Drp1. Mol. Biol. Cell 26, 4466–4477 (2015). 73. Clinton, R. W., Bauer, B. L. & Mears, J. A. Affinity Purification and Functional Characterization of Dynamin-Related Protein 1. Methods Mol. Biol. 2159,41–53 (2020). 74. Wakabayashi, J. et al. The dynamin-related GTPase Drp1 is required for embryonic and brain development in mice. J. Cell Biol. 186, 805–816 (2009). 75. Leonard,M.,Song,B.D.,Ramachandran,R.&Schmid,S.L.Robust colorimetric assays for dynamin’s basal and stimulated GTPase activities. Methods Enzymol. 404, 490–503 (2005). 76. Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat. Methods 14,290–296 (2017). 77. Montecinos-Franjola, F., Bauer, B. L., Mears, J. A. & Ramachandran, R. GFP fluorescence tagging alters dynamin-related protein 1 oligomerization dynamics and creates disassembly-refractory puncta to mediate mitochondrial fission. Sci. Rep. 10,14777(2020). 78. Malaby, A. W. et al. Methods for analysis of size-exclusion chromatography-small-angle X-ray scattering and reconstruction of protein scattering. J. Appl. Crystallogr. 48, 1102–1113 (2015). 79. Nielsen,S.S.etal.BioXTASRAW,asoftwareprogramforhighthroughput automated small-angle X-ray scattering data reduction and preliminary analysis. J. Appl. Crystallogr. 42,959–964 (2009). 80. Manalastas-Cantos, K. et al. ATSAS 3.0: expanded functionality and new tools for small-angle scattering data analysis. J. Appl. Crystallogr. 54,343–355 (2021). 81. Hopkins,J.B.,Gillilan,R.E.&Skou,S.BioXTASRAW:improvements to a free open-source program for small-angle X-ray scattering data reduction and analysis. J. Appl. Crystallogr. 50,1545–1553 (2017). 82. Pettersen,E.F.etal.UCSFChimera–a visualization system for exploratory research and analysis. J. Comput Chem. 25, 1605–1612 (2004). 83. Goddard, T. D., Huang, C. C. & Ferrin, T. E. Visualizing density maps with UCSF Chimera. J. Struct. Biol. 157,281–287 (2007). 84. Jumper, J. & Hassabis, D. Protein structure predictions to atomic accuracy with AlphaFold. Nat. Methods 19,11–12 (2022). 85. Krissinel, E. Enhanced fold recognition using efficient short fragment clustering. J. Mol. Biochem. 1,76–85 (2012). 86. Agirre, J. et al. The CCP4 suite: integrative software for macromolecular crystallography. Acta Crystallogr. D. Struct. Biol. 79, 449–461 (2023). 87. Schindelin, J. et al. Fiji: an open-source platform for biologicalimage analysis. Nat. Methods 9,676–682 (2012). 88. Dar,S.,Kamerkar,S.C.&Pucadyil,T.J.Useofthesupported membrane tube assay system for real-time analysis of membrane fission reactions. Nat. Protoc. 12,390–400 (2017). 89. Arganda-Carreras, I., Fernandez-Gonzalez, R., Munoz-Barrutia, A. & Ortiz-De-Solorzano, C. 3D reconstruction of histological sections: Application to mammary gland tissue. Microsc Res. Tech. 73, 1019–1029 (2010). Acknowledgements We thank Ashutosh Prince and Shane Wyborny (both of CWRU) for technical assistance in protein production. NIH R01 grants GM121583 and GM125844 supported work in the R.R. and J.A.M. laboratories, respectively. Work in the A.V.S. laboratory was supported by the PGC2018-099971-B-I00 and PID2021-127844NB-I00 grants funded by MCIN/AEI/10.13039/501100011033/ and by “ERDF A way of making Europe”and by the Basque Government Grant IT1625-22. Jon Agirre is a Royal Society University Research Fellow (award codes UF160039 and URF\R\221006). I.P.J. acknowledges the predoctoral fellowship from the University of the Basque Country. We are grateful for computational support from the University of York High Performance Computing service, Viking and the Research Computing team, notably Jasper Grimm and Emma Barnes. The authors are grateful to the Electron Microscopy and Crystallography platform of the CIC bioGUNE and the Basque Resource for Electron Microscopy (BREM) of the Biofisika Institute for providing access to cryo EM sample preparation and analysis equipment. Molecular graphics images were produced using the UCSF Chimera package from the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco (supported by NIH P41 RR-01081). This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357. BioCAT was supported by grant P30 GM138395 from the National Institute of General Medical Sciences of the National Institutes of Health. Author contributions Conceptualization: A.V.S, and R.R.; Experimental methodology: A.V.S, R.R., J.A.M.; Experimental investigation and analysis: all authors; Writing —original draft: A.V.S, and R.R; Writing—review and editing: all authors; Funding acquisition: A.V.S and R.R.; Resources: J.A.M., A.V.S. and R.R.; Project supervision: A.V.S. and R.R.; Negative-stain EM data acquisition and Drp1 structure reconstructions: K.R. and J.A.M.; Cell biology experiments and peroxisomal circularity analysis: D.H. and X.Q.; SECSAXS experiments and data analyses: M.M.; Production of DNA constructs: P.M. and R.R.; Protein purification and GTPase assays: P.M., J.O.G., and R.R.; Alphafold predictions and structural overlays: J.A.; SECMALS experiments: R.R.; Analysis of mitochondrial fragmentation: R.R. and A.V.S; Preliminary data on Drp1-mediated NT fission: J.M.M.G.; NT fission assays and analyses: I.P.J. and A.V.S.; cryoEM analyses of preformed NT constriction: I.S.P. and A.V.S. Article https://doi.org/10.1038/s41467-023-44413-6 Nature Communications | (2024) 15:52 16 Competing interests The authors declare no competing interests Additional information Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-023-44413-6. Correspondence and requests for materials should be addressed to Anna V. Shnyrova or Rajesh Ramachandran. Peer review information Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available. Reprints and permissions information is available at http://www.nature.com/reprints Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/. © The Author(s) 2024 Article https://doi.org/10.1038/s41467-023-44413-6 Nature Communications | (2024) 15:52 17