Search

An outstanding question is how cells control the number and size of membrane organelles. The small GTPase Rab5 has been proposed to be a master regulator of endosome biogenesis. Here, to test this hypothesis, we developed a mathematical model of endosome dependency on Rab5 and validated it by titrating down all three Rab5 isoforms in adult mouse liver using state-of-the-art RNA interference technology. Unexpectedly, the endocytic system was resilient to depletion of Rab5 and collapsed only when Rab5 decreased to a critical level. Loss of Rab5 below this threshold caused a marked reduction in the number of early endosomes, late endosomes and lysosomes, associated with a block of low-density lipoprotein endocytosis. Loss of endosomes caused failure to deliver apical proteins to the bile canaliculi, suggesting a requirement for polarized cargo sorting. Our results demonstrate for the first time, to our knowledge, the role of Rab5 as an endosome organizer in vivo and reveal the resilience mechanisms of the endocytic system.

A genome-wide analysis of the human kinome — the sum total of protein, lipid and carbohydrate kinases — shows that the cell's signalling functions are intimately linked to endocytosis, the process by which material is imported into the cell inside a membrane vesicle. Many of the kinases involved in endocytosis are known to function in mitogenic signalling, stimulating cell division and lymphocyte transformation: this adds to growing evidence that endocytic transport and signal transduction are integrated processes. An accompanying manuscript looks at the role of some of the kinases identified in the kinome screen, revealing novel dynamic properties of caveola/raft-mediated endocytosis that are distinct from the current picture of membrane transport. A functional genomics approach has revealed that caveolae/raft-mediated endocytosis is subject to regulation by a large number of kinases1. Here we explore the role of some of these kinases in caveolae dynamics. We discover that caveolae operate using principles different from classical membrane trafficking. First, each caveolar coat contains a set number (one ‘quantum’) of caveolin-1 molecules. Second, caveolae are either stored as in stationary multi-caveolar structures at the plasma membrane, or undergo continuous cycles of fission and fusion with the plasma membrane in a small volume beneath the surface, without disassembling the caveolar coat. Third, a switch mechanism shifts caveolae from this localized cycle to long-range cytoplasmic transport. We have identified six kinases that regulate different steps of the caveolar cycle. Our observations reveal new principles in caveolae trafficking and suggest that the dynamic properties of caveolae and their transport competence are regulated by different kinases operating at several levels.

Endocytosis is a key cellular process, encompassing different entry routes and endocytic compartments. To what extent endocytosis is subjected to high-order regulation by the cellular signalling machinery remains unclear. Using high-throughput RNA interference and automated image analysis, we explored the function of human kinases in two principal types of endocytosis: clathrin- and caveolae/raft-mediated endocytosis. We monitored this through infection of vesicular stomatitis virus, simian virus 40 and transferrin trafficking, and also through cell proliferation and apoptosis assays. Here we show that a high number of kinases are involved in endocytosis, and that each endocytic route is regulated by a specific kinase subset. Notably, one group of kinases exerted opposite effects on the two endocytic routes, suggesting coordinate regulation. Our analysis demonstrates that signalling functions such as those controlling cell adhesion, growth and proliferation, are built into the machinery of endocytosis to a much higher degree than previously recognized.

GTPases and lipid kinases regulate membrane traffic along the endocytic pathway by mechanisms that are not completely understood. Fusion between early endosomes requires phosphatidylinositol-3-OH kinase (PI(3)K) activity as well as the small GTPase Rab5. Excess Rab5-GTP complex restores endosome fusion when PI(3)K is inhibited. Here we identify the early-endosomal autoantigen EEA1 which binds the PI(3)K product phosphatidylinositol-3-phosphate, as a new Rab5 effector that is required for endosome fusion. The association of EEA1 with the endosomal membrane requires Rab5-GTP and PI(3)K activity, and excess Rab5-GTP stabilizes the membrane association of EEA1 even when PI(3)K is inhibited. The identification of EEA1 as a direct Rab5 effector provides a molecular link between PI(3)K and Rab5, and its restricted distribution to early endosomes indicates that EEA1 may confer directionality to Rab5-dependent endocytic transport.

Vesicular transport is a dynamic process that requires coordinated interactions between membrane and cytoskeleton. The mechanisms and molecules integrating these interactions are unclear. A Rho protein, RhoD, might provide a molecular link between membrane traffic and the cytoskeleton. Activated RhoD causes rearrangements of the actin cytoskeleton and cell surface, and governs early endosome motility and distribution.

Endocytosis is a complex process fulfilling many cellular and developmental functions. Understanding how it is regulated and integrated with other cellular processes requires a comprehensive analysis of its molecular constituents and general design principles. Here, we developed a new strategy to phenotypically profile the human genome with respect to transferrin (TF) and epidermal growth factor (EGF) endocytosis by combining RNA interference, automated high-resolution confocal microscopy, quantitative multiparametric image analysis and high-performance computing. We identified several novel components of endocytic trafficking, including genes implicated in human diseases. We found that signalling pathways such as Wnt, integrin/cell adhesion, transforming growth factor (TGF)-beta and Notch regulate the endocytic system, and identified new genes involved in cargo sorting to a subset of signalling endosomes. A systems analysis by Bayesian networks further showed that the number, size, concentration of cargo and intracellular position of endosomes are not determined randomly but are subject to specific regulation, thus uncovering novel properties of the endocytic system.

Nature 464, 243–249 (2010); doi:10.1038/nature08779 Readers alerted us to a technical mistake (a shift of one row) made during the compilation of Supplementary Table 4, resulting in the assignment of 455 gene symbols to the wrong gene IDs. For the affected genes, the cluster group, the number of positive small interfering and endoribonuclease-prepared RNAs (si/esiRNAs) per gene and the Phenoscore values were assigned to the wrong gene symbol (although they were all assigned to the correct gene ID).

The GTPase cycle is a versatile regulatory mechanism directing many cell functions, and Rab family members use it to regulate intracellular transport. Current models propose that GTP hydrolysis by Rab proteins is either required for membrane fusion or occurs afterwards to allow recycling of the protein. To measure the GTPase activity of Rab5 in endocytic membrane fusion, we engineered a mutant that preferentially binds xanthosine 5'-triphosphate (XTP),Rab5(D136N) and monitored the kinetics of [alpha(32)P]-XTP hydrolysis in situ during endosome fusion in vitro. Surprisingly, nucleotide hydrolysis occurred even in the absence of membrane fusion, indicating that membrane-bound Rab5 undergoes futile cycles of GTP(XTP) binding and hydrolysis. Nucleotide triphosphate hydrolysis by Rab5 is not conditional on membrane fusion and is reduced by its effector Rabaptin-5. Our data reveal that the GTP cycle of Rab proteins differs from that of other GTPases (for example, EF-Tu) and indicate that GTP hydrolysis acts as a timer that determines the frequency of membrane docking/fusion events.

The Rab GTPases function as specific regulators of membrane transport. The GTP/GDP cycle is believed to control shuttling of Rab proteins between the cytosol and organelle membranes. In vitro, Rab proteins are removed from membranes by a protein that inhibits GDP dissociation (rabGDI), which leads to formation of a cytosolic complex of Rab with the inhibitor protein. Here we use a purified Rab5-rabGDI complex in a permeabilized cell system to investigate how the cytosolic complexed form of Rab reassociates with the membrane. We find that exogenous Rab5 is correctly targeted and induces the formation of enlarged early endosomes, demonstrating that it is functionally active. Binding of Rab5 to the acceptor membrane is accompanied by release of the rabGDI protein into the cytosol. A transient GDP-Rab5 intermediate was detected which was subsequently converted into the GTP-bound form. Our results indicate that there is a multistep mechanism for the insertion of Rab5 into the membrane which is mediated by a guanine-nucleotide-exchange factor.

MAMMALIAN cells express many ras-like low molecular mass GTP-binding proteins (rab proteins)1–6 that are highly homologous to the Yptl and Sec4 proteins involved in controlling secretion in yeast7-9. Owing to their structural similarity and to their variety, rab proteins have been postulated to act as specific regulators of membrance traffic in exocytosis and endocytosis10, and rab5 has been shown to be involved in early endosome fusion in vitro11. In agreement with their postulated functions, all rab proteins studied so far have been found in distinct subcompartments along the exocytic or endocytic pathways5,11–13. To define the region mediating their specific localization, we transiently expressed rab2, rab5 and rab7 hybrid proteins in BHK cells, and determined their intracellular localization by immunofluorescence confocal microscopy and subcellular fractionation. Here we present evidence that the highly variable C-terminal domain contains structural elements necessary for the association of rab proteins with their specific target membranes in the endocytic pathway.