Journal of Cell Biology

The Golgi-localized, {gamma}-ear containing, ADP-ribosylation factor binding proteins (GGAs) are a family of adaptor proteins that regulate transport of specific cargo receptors from the Golgi to endosomes. For many years it was assumed that GGAs transport cargo via interaction with the adaptor complex AP-1. However, recent findings suggest that GGA and AP-1 may have opposing roles, with GGAs facilitating forward transport between Golgi and endosomes, and AP-1 mediating the opposite trafficking step. To shed light on the functional connection of GGAs with AP-1, we combined CRISPR-Cas9 gene editing with live-cell imaging and TurboID-based proximity labelling. We find that GGAs localize not only to the Golgi apparatus but also, to a greater extent, to peripheral ARF1-positive compartments responsible for secretory trafficking and endocytic recycling. At both, the Golgi and peripheral sites, we observe distinct sorting domains containing either AP-1 or GGAs alone, as well as domains in which both adaptors are present. Interestingly, GGAs can recruit clathrin lattices independently of AP-1. Proximome mapping shows that AP-1 specific cargoes only localize to AP-1 domains in the absence of GGAs. These findings point to a regulatory role of GGAs in AP-1 transport. We speculate that GGAs prevent binding of AP-1 to its cargo clients to avoid premature retrieval and to modulate bi-directional trafficking between the Golgi and endosomes.

The autophagy core machinery mediates the enclosure of cytosolic cargo destined for degradation in the lysosome. The Atg9-Atg2-Atg18 complex coordinates phagophore expansion via directed lipid transfer until closure of the phagophore rim. Using an Atg2 variant (Atg2-PM4) as a model of decelerated autophagosome biogenesis, we visualized the morphological states prior to autophagosome closure by cryogenic correlative light and electron microscopy in S. cerevisiae. Using in situ cryo-electron tomography, we find an enlarged rim morphology of an expanding phagophore in Atg2-PM4 cells in comparison with Atg2 wildtype condition. Analysis of segmented rim membrane features as well as surrounding and attached vesicles suggest that the enlarged rims are a result of cytosolic vesicles fusing with the growing phagophore. High-resolution imaging in this study shows that, apart from the initial nucleation phase, vesicle fusion can also contribute to phagophore expansion during later stages of autophagosome biogenesis.

Autophagy involves the rapid growth of phagophores through membrane addition. This growth is triggered by vesicles containing the Atg9A protein. However, Atg9A is not incorporated into mature autophagosomes. We now demonstrate that Dynamin-2 (Dnm2) colocalizes with the BAR domain protein Endophilin-B1 (EndoB1/Bif-1/SH3GLB1) and other autophagy proteins when autophagy is induced. Our data suggest that Atg9A is retrieved from phagophores via fission, with Dnm2 acting as the membrane scission protein. Blocking Atg9A recycling, either by mutating Dnm2, using RNA interference, or applying chemical inhibitors, results in Atg9A remaining in autophagosomes and being degraded during autophagy. Overall, these findings provide new insights into the roles of membrane-scission proteins in autophagy.

At the early endosome, cargos are sorted into subdomains; receptors destined for recycling to the plasma membrane are sorted into tubulovesicular structures that undergo fission and release cargo-laden vesicles that traffic along microtubules. Although branched actin has been implicated in the establishment/maintenance of endosomal membrane subdomains, its role in cargo segregation, fission, and recycling has not been extensively studied. Using inhibitors of formin-and ARP2/3-mediated actin assembly, we show that branched actin, but not linear actin, is required for endosome fission and receptor recycling. To examine the spatial relationship between actin and cargo, we transfected cells with constitutively active RAB5 Q79L to generate enlarged endosomes and demonstrated that internalized transferrin localized to discrete endosomal regions adjacent to branched actin. ARP2/3 inhibition disrupted this organization, resulting in broader cargo distribution on the endosomal membrane and coalescence of degradative and retrieval subdomains. Consistent with impaired endosomal sorting and fission, branched actin inhibition led to cargo accumulation. Our findings identify ARP2/3-mediated branched actin as a key regulator of cargo segregation, subdomain maintenance, and fission at the early endosome.

Membrane contact sites (MCS) between intracellular organelles regulate lipid exchange, organelles dynamics and spatial organization of signaling pathways, yet their contribution to regulated exocytosis remains poorly understood. Here, we investigated the role of endoplasmic reticulum (ER) MCS in calcium-regulated exocytosis using primary bovine chromaffin cells. Combining electron microscopy on plasma membrane (PM) sheets with immunogold labeling, we identified ER structures contacting docked secretory granules and classified three types of MCS: ER-PM, ER-granule and tripartite ER-PM-granule contacts. These contacts are enriched at exocytic sites and contain Orai1 and STIM1, both known for mediating store-operated calcium release. Functional perturbation of the Orai/STIM pathway revealed that constitutive STIM activation or pharmacological inhibition of Orai1 reduced the number of exocytotic events, slowed catecholamine release and disrupted actin organization at granule docking sites. Together, our findings revealed a previously unrecognized role for ER MCS in organizing exocytic sites and controlling secretion efficiency in neuroendocrine cells.

Adaptor protein complex 3 (AP-3) mediates clathrin-independent transport to lysosomes, yet accessory factors supporting this pathway remain incompletely defined. In Saccharomyces cerevisiae, the C-terminal intrinsically disordered regions (IDRs) of both AP-3 large subunits ({delta} and {beta}3) serve as platforms for association with accessory factors. Through proteomic analysis of proteins associated with these IDRs, we identify the septin cytoskeleton as a candidate AP-3-associated factor. Bimolecular fluorescence complementation (BiFC) reveals a hierarchical pattern of association: AP-3 shows preferential proximity to core septin subunits (Cdc10, Cdc3, Cdc12) over terminal subunits (Cdc11 and Shs1). These terminal subunits serve as alternative caps of septin octamers, generating structurally distinct assemblies. Significantly, dysfunction of Cdc11 but not Shs1 selectively impairs AP-3-dependent cargo sorting without affecting the parallel vacuolar protein sorting (VPS) pathway to the vacuole (lysosome in yeast), providing genetic evidence for a specific functional connection between Cdc11-containing septin assemblies and AP-3-mediated transport.

Glucose transporter 4 (GLUT4) is sequestered intracellularly until the action of insulin causes its redistribution to the cell surface. How cells sequester GLUT4, generating and maintaining GLUT4 storage vesicles (GSVs), is still unclear. Intracellular nanovesicles (INVs) operate on various membrane trafficking pathways and show molecular similarity to GSVs. Here we show that GLUT4 and associated GSV cargo proteins (IRAP/LNPEP, cellugyrin/SYNGR2, sortilin/SORT1, and VAMP2) are trafficked by INVs, which we term GLUT4-flavor INVs. The majority of GLUT4 is carried in GLUT4-flavor INVs which represent a small fraction of the total INV pool. We use a method to capture specific vesicles and test their involvement in the redistribution of GLUT4 in response to insulin. Direct capture of defined GSVs ablated the response, however capture of INVs had a smaller effect. We show that the function of GLUT4-flavor INVs is in the sorting of GSV proteins to supply the insulin-responsive GSV pool and that interfering with INV function by TPD54/TPD52L2 depletion results in mistrafficking of GLUT4 to the plasma membrane. These findings establish GLUT4-flavor INVs as the precursor GSVs that are responsible for intracellular sequestration of GLUT4.

The ULK1 complex (ULK1C) and the class III phosphatidylinositol 3-kinase complex I (PI3KC3-C1) act together to initiate autophagy. Human ULK1C consists of ULK1 itself, FIP200, and the HORMA domain heterodimer ATG13:ATG101. PI3P generated by PI3KC3-C1 is essential to recruit and stabilize ULK1C on membranes for ULK1 to phosphorylate its membrane-associated substrates in autophagy induction, even though ULK1C subunits do not contain any PI3P-binding domains. Here we show that the ATG13:ATG101 dimer forms a tight complex with the PI3P-binding protein WIPI3, as well as with WIPI2. Bound to WIPI2-3, ATG13:ATG101 aligns with the membrane to insert its Trp-Phe (WF) finger into the membrane. Molecular dynamics simulations show that alignment of WIPIs and the ATG101 WF finger cooperatively stabilizes the complex on membranes, explaining the essential role of the WF residues in autophagy. Biochemical reconstitution and a cell-based assay show that WIPI3:ATG13 engagement is required for ATG16L1 phosphorylation by ULK1, ATG13 puncta formation, and bulk autophagic flux. We further showed that a kinase domain (KD)-proximal PVP motif within the ULK1 IDR docks onto the surface of the ATG13:ATG101 HORMA dimer and used molecular modeling to show how the ULK1 KD is brought close to the membrane surface. Biochemical reconstitution and cell-based assays show that the PVP motif is essential for in vitro ULK1 phosphorylation of ATG16L1 and important for starvation-induced autophagy and BNIP3/NIX-dependent mitophagy. These data establish a stepwise pathway for recruitment of the ULK1 KD to the vicinity of the membrane surface downstream of PI3KC3-C1.

Microtubules attach to kinetochores to facilitate chromosome movement to opposite spindle poles. Defective kinetochore-microtubule attachments lead to phosphoryation of the outer kinetochore protein KNL1 at conserved MELT motifs, which triggers spindle assembly checkpoint activation and recruitment of the fibrous corona. To identify additional phosphorylation sites that regulate kinetochores, we treated HEK 293T/17 cells with nocodazole, paclitaxel, or STLC to create defective kinetochore-microtubule attachment states. We then purified KNL1 and performed proteomics and identified 111 phosphorylation sites on KNL1, including several that may be attachment-state specific. These data demonstrate that KNL1 is extensively phosphoregulated in response to treatment with microtubule-disrupting compounds.

Following chromosome segregation, the nuclear envelope (NE) must be reassembled and holes in the nuclear membrane must be "sealed." During NE assembly, the NE-specific adaptor, Cmp7, recruits/activates ESCRT-III proteins to mediate NE sealing. However, recent evidence suggests the presence of alternative mechanisms. In a screen using the fission yeast, S. japonicus, we recently implicated the ESCRT adaptor, Alx1, and a conserved, but little studied protein, Vid27, in Cmp7-independent NE assembly. Here, we provide direct evidence that Alx1 functions in a Cmp7- and ESCRT-independent NE assembly pathway via positive regulation of Vid27. Consistent with a role in membrane remodeling, Vid27 localizes to sites of postmitotic NE sealing and is essential in S. japonicus. Alx1 and Vid27 form a complex and mutations disrupting their interaction abolish Alx1s enhancement of Vid27 function at the NE. These findings define components of a new Cmp7- and ESCRT-independent NE assembly pathway, advancing our understanding of the mechanisms crucial for maintaining the integrity of the nucleus.

Membrane lipids are heterogeneously distributed across the bilayers of cellular membranes. Cytosolic-facing pools of diacylglycerol (DAG) in the yeast Saccharomyces cerevisiae are enriched at both ends of the endomembrane system from the vacuolar membrane to the polarized plasma membrane (PM) of buds. However, how this distribution is maintained remains unknown. Using a genome-wide DAG biosensor screen in yeast, we identify regulators of DAG spatial distribution, enriched in proteins involved in vesicle or lipid transport and in phospholipid or sterol metabolism. A subset of mutants exhibited DAG mislocalization predominantly to the PM, with the most severe phenotype linked to a mutant of a predicted lipase we have named Drl1 (DAG redistribution lipase 1). Reversion of this phenotype required both enzymatic activity and the presence of an intrinsically disordered carboxy-terminal domain. Lipidomic analysis revealed that loss of Drl1 increased cellular lysophosphatidylcholine (LysoPC) levels. Remarkably, we find that supplementing cells with a non-metabolizable LysoPC analogue replicated the mutant DAG phenotype, implicating LysoPC as a novel spatial regulator of DAG. High-resolution imaging suggests that LysoPC reduces the PM sterol pool resulting in DAG expansion into new PM territories. More globally, our work expands the known interplay between various lipids and their co-regulation to maintain accurate membrane properties.

Faithful chromosome segregation requires the spatial and temporal remodeling of cell structures driven largely by cyclin-dependent kinase (Cdk) activity. In some experimental systems the timely elevation of one cyclin isoform is sufficient to support mitotic entry and progression. In human somatic cells, however, three cyclins - Cyclin A2, Cyclin B1, and Cyclin B2 - are present at mitotic entry, and their distinct contributions during mitosis remain largely unclear. We demonstrate that Cyclin A2 promotes the kinetochore localization of Cyclin B1, Cyclin B2 and the fibrous corona component CENPF, while suppressing recruitment of the kinetochore-microtubule (k-MT) stabilizer Astrin. Extending Cyclin A2 into metaphase by expressing a non-degradable mutant causes persistent kinetochore localization of Cyclin B1, Cyclin B2, and CENPF. Conversely, Cyclin B1 limits kinetochore localization of Cyclin B2 and CENPF, promotes Astrin recruitment, and stabilizes k-MT attachments and Cyclin B1 overexpression further reduces kinetochore localization of CENPF during prometaphase. These findings reveal an interdependence among mitotic cyclins in early mitosis and the countervailing activities of Cyclin A2 versus Cyclin B1/B2 in regulating key mitotic events. We propose that this circuitry enforces a switch-like transition from prometaphase - marked by corona assembly and high k-MT turnover - to metaphase - marked by corona disassembly, stabilization of end-on k-MT attachments, and eventual spindle assembly checkpoint satisfaction - to choreograph the structural changes required to ensure faithful chromosome segregation.

Neuronal extracellular vesicles (EVs) are released from synapses, and play roles in cellular communication, proteostasis, and the spread of toxic proteins in disease. The small GTPase Rab11 is required to maintain a reservoir of EV cargoes at presynaptic terminals, but how its diverse effector proteins contribute to this function and where Rab11 acts in neurons remains unclear. Using Drosophila motor neurons as a model, we show that EV cargoes redistribute from synapses to axons and cell bodies in rab11 mutants, concomitant with reduced release from synapses. We conducted a directed genetic screen of Rab11-associated factors and found that they have distinct roles in EV trafficking. Tethering and sorting factors are required to maintain levels of presynaptic EV precursors, supporting the hypothesis that Rab11 regulates EV cargo pools through recycling flux rather than by directly mediating EV release. Unexpectedly, we found that different classes of Rab11-associated proteins have opposite functions: the motor protein MyoV and the PI4KIII component Rbo sustain cargo levels at synapses, while the motor adaptor Nuf/Rab11FIP4 and the PI4KIII{beta} homolog Fwd restrict cargo levels. Together, these results indicate that Rab11 regulates multiple distinct organelle transport trajectories and PI(4)P populations to direct EV cargoes toward different cellular fates.

Quantitative organelle analysis is highly sensitive to image-processing choices, limiting reproducibility across microscopy studies. Here, we systematically compare automated, interactive machine learning, and deep learning-based pipelines for lipid droplet and mitochondrial quantification in live human osteosarcoma cells imaged by fluorescence microscopy and label-free holotomography. Using standardized downstream feature extraction, we evaluated script-based workflows (Fiji, Python), a modular platform (CellProfiler), interactive machine learning (ilastik), and pretrained deep learning models. Lipid droplet segmentation was qualitatively consistent across approaches; however, droplet counts, and size distributions varied substantially between pipelines and imaging modalities, with ilastik reducing background-driven detections and improving cross-modality agreement. In contrast, mitochondrial quantification proved highly sensitive to segmentation and skeletonization choices, particularly in holotomography where global intensity-threshold-based methods failed to capture network structure. Based on these cross-pipeline comparisons, we demonstrate how organelle- and modality-specific benchmarking can guide pipeline selection, illustrated by the analysis of metabolic perturbations affecting lipid droplets and mitochondria. Together, these results highlight modality- and morphology-dependent limitations in common analysis pipelines and provide practical guidance for selecting robust, reproducible strategies for quantitative organelle imaging.

The nuclear pore complex (NPC) is the only gateway that connects the nucleus with the cytoplasm in eukaryotic cells. Its nucleoplasmic face is decorated by the nuclear basket, a filamentous structure with important roles in mRNA export and chromatin organization. In contrast to major parts of the nuclear pore scaffold, the architecture and organization of the nuclear basket remain poorly defined. In this study, we investigate the interaction network required for formation of the nuclear basket in vivo using budding yeast. We demonstrate that the filamentous Mlp1 protein relies on coiled-coil segments outside its previously characterized NPC-binding region to stably interact with the NPC. Furthermore, our results reveal that Mlp1s paralogue, Mlp2, plays a central role in nuclear basket architecture. Specifically, Mlp2 associates with the NPC independently of Mlp1 and together with Mlp1 is essential for the efficient recruitment of Pml39 and additional Mlp1 subunits. Our findings allow us to propose a refined model of nuclear pore basket architecture and organization.

The tight junction (TJ) protein cingulin binds directly to nonmuscle myosin 2B (NM2B) through sequences in its C-terminal rod-tail region and recruits it to tight junctions (TJ) to control membrane cortex mechanics, epithelial morphogenesis and cingulin conformation. However, the minimal sequence required for cingulin-NM2B interaction and how this interaction is regulated is not known. Here we identify a 19-aminoacid sequence at the hinge between the cingulin rod and tail that is required for cingulin-NM2B interaction, and we investigate the role of phosphorylation of Ser residues within this region in regulating this interaction. Immunofluorescence microscopy localization of NM2B in cingulin-KO cells rescued with mutant cingulin constructs shows that phospho-mimetic but not dephospho-mimetic cingulin mutants inhibit NM2B recruitment to junctions and downstream regulation of cingulin conformation and TJ tortuosity, correlating with cingulin-NM2B interaction, as determined by GST pulldown analysis. In contrast, either phospo-or dephospho-mimetic mutants of Ser residues within the cingulin head domain do not affect either NM2B recruitment to TJ, or cingulin conformation and localization in cells, or TJ membrane tortuosity. Finally, Ser residues within the hinge display the consensus sequence for protein kinases CK1 and CK2, and, through in vitro phosphorylation, site mutation analysis and use of inhibitors, we identify a complex interplay between CGN phospho-sites, with a prominent negative role of Ser1162 phosphorylation in the regulation of cingulin-NM2B interaction. In summary, we show that cingulin-NM2B interaction is regulated by cingulin phosphorylation within the hinge and identify a potential role for CK1 and CK2 kinases in cingulin phosphorylation.

Microtubule networks are major determinants of cell architecture and logistics. Microtubule organization and density are regulated by severing enzymes, which cut microtubule lattices or affect their growth and shortening. These activities can lead to microtubule amplification or disassembly, depending on the presence of microtubule stabilizers or destabilizers, but the interplay between these factors is poorly understood. Here, we reconstituted in vitro the activity of microtubule severase katanin together with microtubule minus-end stabilizers CAMSAPs, their binding partner WDR47 and microtubule depolymerase kinesin-13/MCAK. We confirmed that katanin can amplify or destroy microtubules in a concentration-dependent manner. CAMSAPs recruit katanin to microtubules and reduce katanin concentration needed for both amplification and destruction, whereas kinesin-13 completely abolishes microtubule amplification. WDR47 binds to microtubules decorated by CAMSAPs and suppresses katanin binding and severing. In addition, both katanin and WDR47 inhibit polymerization of CAMSAP-decorated microtubule minus ends. These data explain how these proteins act together to fine-tune microtubule minus-end stability without strongly increasing microtubule abundance. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=169 SRC="FIGDIR/small/714132v1_ufig1.gif" ALT="Figure 1"> View larger version (20K): org.highwire.dtl.DTLVardef@746fe3org.highwire.dtl.DTLVardef@5dd5a8org.highwire.dtl.DTLVardef@762373org.highwire.dtl.DTLVardef@1192db_HPS_FORMAT_FIGEXP M_FIG Graphical abstract C_FIG

Wnts are secreted signalling molecules that regulate development and adult homeostasis. Most Wnts carry a lipid moiety that must be shielded from the aqueous environment. In the secretory pathway, this is achieved by a hydrophobic tunnel in Wntless, a multipass transmembrane protein. However, the Wnt lipid moiety must be released from Wntless before Wnts can engage with Frizzled receptors on receiving cells. Here we address the cell biological basis of Wnt-Wntless dissociation, using as a model the secretion of Drosophila Wingless in wing primordia. Super-resolution microscopy shows that Wingless first reaches the apical surface before being re-internalized to reach, without Wntless, specialized Rab7/Rab4-positive endosomes. From there Wingless traffics to the basolateral membrane where it can engage with glypicans to form a basolateral gradient. Acute inhibition of endocytosis, either with a temperature-sensitive dynamin mutant or a novel optogenetic means of inhibiting clathrin, leads to apical Wingless release in abnormal punctae devoid of Wntless, suggesting that Wingless-Wntless dissociation commences at the apical surface, perhaps because of a distinct lipid composition there. Indeed, similar looking punctae are produced upon genetic abrogation of the ceramide synthase Schlank, specifically in Wingless-producing cells. These punctae resemble insoluble aggregates that form in vitro upon detergent removal. Accordingly, punctae formation can be prevented by shielding the Wingless lipid, in vivo with excess Dally-like protein (Dlp) or in vitro with liposomes. Our results show that membrane lipid composition modulates the orderly transfer of Wingless lipid from Wntless to the inner endosomal surface thus preventing aggregation and ensuring seamless secretion in the basolateral space.

Lipid droplets (LDs) are dynamic organelles that store neutral lipids and form in the endoplasmic reticulum (ER) membrane. Formation of new LDs is a controlled process and requires proteins with specific functions to form and grow from the ER membrane without any defect. In vitro studies have suggested a role for membrane curvature in LD emergence from the ER. Here, we use the membrane-shaping protein Pex30 to investigate the impact of ER membrane curvature on LD biogenesis and morphology. We modified the reticulon homology domain (RHD) of Pex30, which is responsible for tubulating the ER membrane, by extending the short hairpin transmembrane domains (TMD). The Pex30 (TMD) mutants cannot tubulate the ER membrane and generate less local membrane curvature that WT Pex30. Additionally, these mutants are unable to restore delayed LD biogenesis observed in cells devoid of Pex30. Our results indicate that Pex30 RHD generates local membrane curvature at ER subdomains that drives formation of new LDs.

Microtubules, stiff rods built up from tubulin dimers, form a cytoskeletal network whose structure, behaviour, and function have been extensively investigated, mainly from a mechanical perspective. Here, we describe a role for tubulin in the cellular stress response. We overexpressed tubulin dimers in a controlled fashion in 293F cells. Despite the engagement of autoregulation, a mechanism that degrades tubulin-encoding mRNAs when tubulin levels are high, a surplus of tubulin and microtubules is detected in overexpressing cells. This leads to altered microtubule behaviour, mitotic problems, deregulation of the cell cycle, and replication stress. Surprisingly, we also observe proteostasis defects in tubulin overexpressing cells, which we attribute to mitochondrial stress-related translation attenuation. Conversely, tubulin and microtubules are downregulated as part of the response to oxygen or glutamine deprivation. Together, our data link tubulin levels, and hence autoregulation, to cellular quality control and proteostasis. We propose that competitive interactions with key partners, including the mitochondrial protein import and general translation machinery, underlie the tubulin-mediated control of cellular homeostasis.