James Pickett

Post-translational modification: UBE1, you're not alone

10.1038/nrm2218 Ubiquitin and other ubiquitin-like molecules (UBLs) are post-translationally attached to proteins through a sequential enzymatic cascade that uses an E1 activating enzyme, an E2 conjugating enzyme and an E3 ligase. It was presumed that a single E1 — UBE1 (ubiquitin-activating enzyme E1) — activates ubiquitin and that the substrate specificity of the cascade (that is, which proteins are modified with ubiquitin) was exclusively due to the diversity of E2s and E3s. Jin et al. show both of these assumptions to be incorrect by identifying a second E1 for ubiquitin that defines a separate ubiquitin-conjugating pathway. E1 enzymes are characterized by an adenylation domain that is composed of two ThiF-homology motifs that bind and adenylate (activate) the UBL, a catalytic cysteine domain (CCD) to which the adenylated UBL is transiently attached during the cascade and a C-terminal ubiquitin-fold domain (UFD) that recruits E2. Jin et al. searched the human genome for genes that encode a ThiF-homology domain. They identified all known E1s and a previously uncharacterized protein, UBA6 (also known as UBE1L2 (ubiquitin-activating enzyme E1-like 2)). UBA6 also contains a CCD and a UFD so, by domain composition alone, UBA6 closely resembles an E1 enzyme. The ability of UBA6 to activate different UBLs, including several for which the E1 is unknown, was tested in vitro. Surprisingly, UBA6 activated ubiquitin but not the other UBLs tested. This was confirmed in vivo: UBA6 that was conjugated to ubiquitin could be isolated from cells under conditions that stabilized the E1–ubiquitin intermediate. Next, the authors compared the specificity of UBA6 to transfer ubiquitin to different E2s with that of UBE1. Although some E2s were charged by UBA6 and UBE1 with equal efficiencies, the two E1s showed different specificities. Almost half of the E2s tested were charged by UBE1 only, whereas a previously uncharacterized E2 was identified as the first UBA6-specific E2 (USE1). The differential charging of E2s was conserved in vivo: small-interfering-RNA-mediated depletion of UBA6 selectively reduced the levels of charged USE1 but not those of a UBE1-specific E2. UBA6 and USE1 form part of a second ubiquitinconjugating cascade; however, the E3s and target substrates for USE1 remain unknown. The specificity of the different E1s is predominantly generated by the UFD because switching the UFDs between UBA6 and UBE1 reversed the specificity towards some (but not all) E2s. Together, these findings show an unanticipated level of regulation in the ubiquitin-conjugation cascade. Previously, the ubiquitin dependence of a process has often been investigated by inhibiting UBE1 and, therefore, a re-evaluation of true ubiquitin independence is now required. James Pickett

Dampening down destruction in dendritic cells

Dendritic cells are a key component of the bodys immunosurveillance squad. They internalize foreign proteins, tumour cells and pathogens into phagosomes, where they are partially degraded. The fragments are processed to antigenic peptides in the cytosol and presented on the plasma membrane as part of the major histocompatibility complex to cytotoxic T cells. In Nature Cell Biology, Jancic and colleagues report a molecular mechanism that ensures proteins are not entirely degraded in the proteolytic environment of the phagosome.

Dampening down destruction in dendritic cells: Membrane trafficking

Dendritic cells are a key component of the bodys immunosurveillance squad. They internalize foreign proteins, tumour cells and pathogens into phagosomes, where they are partially degraded. The fragments are processed to antigenic peptides in the cytosol and presented on the plasma membrane as part of the major histocompatibility complex to cytotoxic T cells. In Nature Cell Biology, Jancic and colleagues report a molecular mechanism that ensures proteins are not entirely degraded in the proteolytic environment of the phagosome.

Membrane trafficking: Dampening down destruction in dendritic cells

Dendritic cells are a key component of the bodys immunosurveillance squad. They internalize foreign proteins, tumour cells and pathogens into phagosomes, where they are partially degraded. The fragments are processed to antigenic peptides in the cytosol and presented on the plasma membrane as part of the major histocompatibility complex to cytotoxic T cells. In Nature Cell Biology, Jancic and colleagues report a molecular mechanism that ensures proteins are not entirely degraded in the proteolytic environment of the phagosome.

Cell signalling: New communication skills

Cell-to-cell communication can take numerous forms, including chemical or hormone-mediated signalling and direct cell-to-cell contacts. A study in Nature Cell Biology suggests that some cells may also signal to each other by secreting exosomes that are packaged full of mRNAs and microRNAs. Exosomes (not to be confused with the macromolecular machines that degrade RNAs) are small vesicles of endocytic origin. They are formed within larger multivesicular bodies that can fuse with the plasma membrane to release exosomes into the extracellular environment. The functions of exosomes are poorly characterized but it has been postulated that they interact with cells, into which they might deliver their contents. To identify potential cargoes for exosomes, Valadi et al. profiled the composition and contents of exosomes derived from mast cells and found them to be substantially enriched in small RNAs and mRNAs. The profile of RNAs found in exosomes was different to the profile found in the cytosol. For example, exosomes contained no ribosomal RNAs but contained large numbers of 19–22 nucleotide microRNAs, indicating that RNAs are packaged into exosomes by an active mechanism. Under the correct conditions, mRNAs contained in the exosome could be translated into proteins, which shows that these mRNAs are functional. To test whether exosomes could traffic mRNAs between cells, Valadi et al. incubated mouse-derived exosomes with human mast cells. After 24 hours, several mouse proteins were found to be expressed in the mast cells. It remains unknown how exosomes are taken up by recipient cells, but this process seems to be cell-type specific. The authors propose that exosomes can transport RNA, which they call ‘exosomal shuttle RNA’ (esRNA), between cells. Potentially, esRNA could have widespread effects on protein levels in recipient cells because microRNAs can downregulate the expression of large numbers of genes. However, the physiological significance of this mechanism is unclear; how exosome secretion is regulated and whether exosomes can regulate protein levels in recipient cells in vivo is unknown. James Pickett

Cell division: Protection for cyclists

Cdc14 http://db.yeastgenome.org/ cgi-bin/locus.pl?locus=cdc14 The unfolded protein response (UPR) helps the endoplasmic reticulum (ER) to cope with stressful situations — such as elevated levels of unfolded proteins — by increasing the functional capacity of the ER. Bicknell et al. now show that the UPR not only functions under conditions of extreme cellular stress, but that it is also activated by normal cell division, during which it protects cells against fluctuations in ER demand. The authors observed that the UPR was activated in unstressed budding yeast cells. To uncover the cause of this moderate level of UPR activation, Bicknell et al. studied ER-stressed cells and looked for cellular processes that were sensitive to ER stress. ER-stressed cells grew normally at early time points, but progressively contained larger amounts of DNA compared with unstressed cells and often had multibudded morphologies. This phenotype suggested that ER stress does not block DNA replication, but that it inhibits cell division — possibly during late mitosis or cytokinesis. The former possibility was ruled out by studying indicators of mitotic exit: the degradation of Clb2 and the cytoplasmic relocation of Cdc14 that mark mitotic exit were unaffected in ER-stressed cells, as was spindle disassembly. To confirm that cytokinesis requires UPR activation to progress, the authors studied a UPR-defective yeast mutant strain. Cytokinesis was inefficient and was delayed even in the absence of any external ER-stress inducers. Therefore, it seems that cytokinesis places a demand on the ER — perhaps because of a requirement for membrane vesicles or phospholipids from the secretory pathway that are used in the pinching process. The authors propose that cytokinesis may be one of several physiological processes that induce moderate levels of UPR signalling to fine-tune the capacity of the ER to the dynamic cellular environment. James Pickett

Post-translational modification: The importance of being inactive

10.1038/nrm2265 Protein stability is controlled by the formation of polyubiquitin chains on proteins, which target them for degradation by the proteasome. Ubiquitylation can be catalysed by HECT-E3 ligases that transfer ubiquitin (Ub) from a conserved Cys residue in the catalytic HECT domain onto target proteins, as well as onto other sites on the ligase itself. However, how the activity of HECTE3 ligases is controlled is poorly understood. In Cell, Wiesner et al. report that activity of the HECT-E3 ligase SMURF2 is regulated by an autoinhibitory mechanism that prevents futile cycles of protein production and degradation. SMURF2 is a modular protein, composed of an N-terminal C2 domain that mediates membrane localization, three WW domains involved in binding substrates and adaptor proteins, and the HECT domain. Using nuclear magnetic resonance (NMR) analysis, Wiesner et al. found that the C2 domain interacted with the HECT domain at a position that was close to its catalytic Cys residue. So, does this interaction have functional implications? The authors showed that deleting the C2 domain led to the increased formation of Ub linkages to the catalytic Cys of the HECT domain, and that mutating residues F29 and F30, which are required for the C2–HECT interaction, increased the autoubiquitylation of SMURF2. Therefore, the C2 domain has an inhibitory effect on activity of the HECT domain. Some ubiquitylating enzymes are thought to operate as dimers, suggesting that the HECT domain could be inhibited by the C2 domain within the same protein or from a separate protein. To distinguish between these possibilities, the authors measured levels of ubiquitylation of the constitutively active (F29A/F30A) mutant in the face of increasing levels of isolated C2 domain or catalytically dead (C716A), full-length SMURF2. In contrast to the isolated domain, the C2 domain from full-length C716A SMURF2 had no effect on SMURF2 (F29A/F30A) ubiquitylation, indicating that C2-mediated inhibition arises from an intramolecular interaction. The authors next examined the ubiquitylation of RhoA, which is a known substrate for the closely related HECT-E3 ligase SMURF1, but not SMURF2. However, SMURF2 (F29A/F30A) could ubiquitylate RhoA, which suggests that the C2– HECT interaction may play a part in determining substrate specificity. So, how might the autoinhibitory mechanism be relieved upon substrate recognition? The adaptor protein SMAD7 couples the HECT-domain activity of SMURF2 to substrate recruitment in vivo, and the authors showed that SMAD7 interfered with the inhibitory C2–HECT domain interaction. This elegant mechanism couples substrate recognition with the release of the autoinhibitory C2 interaction. The autoinhibitory mechanism was found to be conserved in several E3 ligases that have a C2– WW–HECT domain architecture. Therefore, this general mechanism maintains a set of E3 enzymes in an inactive state, which prevents these enzymes and their substrates from being ubiquitylated and degraded in an unregulated manner in vivo. James Pickett

Bending around the BAR

10.1038/nrm2206 The process of clathrin-mediated endocytosis (CME) is driven by the collective effort of proteins that cause membrane invagination and those that cause membrane scission. Shimada et al. now report the structure of the F-BAR (Bin–amphiphysin–Rvs) protein domain, which provides insight into how these crescent-shaped modules drive membrane remodelling during endocytosis. CME occurs in three steps. First, clathrin assembles on flat membranes and captures endocytic cargo to form hemispherical clathrin-coated pits. Second, the pit slowly invaginates with a narrowing neck region, and actin polymerizes around the nascent vesicle. Finally, scission proteins such as dynamin are recruited to the neck region where they sever the membrane. The neck region is shaped by the recruitment of N-BAR domain proteins, such as amphiphysins and endophilins, which self-organize into banana-shaped dimers and exert this shape onto membranes. However, these proteins are recruited at a late stage of CME and it is unclear what drives the initial invagination of the small clathrin-coated pit. Shimada et al. determined the structures of the related F-BAR domains from formin-binding protein-17 (FBP17) and Cdc42-interacting protein-4 (CIP4). Compared with the N-BAR domain, the F-BAR domain also forms a dimer and has a similar crescent-shaped architecture; however, the F-BAR domain has longer α-helices and shallower curvature. Both N-BAR and F-BAR domains caused spherical lipid vesicles to tubulate, although F-BAR domains were found to bind preferentially to larger liposomes and generated tubules of a larger diameter. F-BAR domains therefore both ‘sense’ and induce less membrane curvature than the related N-BAR domain. The authors show that, as well as enforcing their curved shape onto membranes, F-BAR domains oligomerize into filaments that wind around the membrane to induce curvature. Three pieces of evidence support this: F-BAR domains are arranged in filaments in the crystal lattice, striations of F-BAR protein decorate tubulated liposomes, and point mutations that block the formation of filaments also impair membrane tubulation. Furthermore, F-BAR domain proteins also appear to harness the membrane-deforming capabilities of actin filaments — the recruitment of FBP17 to clathrin pits was associated with activation of the actin-nucleating machinery and actin polymerization. The authors propose that F-BAR and N-BAR proteins are recruited at different time points during CME. F-BAR domains recognize the shallow curvature of clathrin-coated pits to drive membrane invagination, whereas N-BAR domains are recruited to regions of already high curvature, where they further constrict the membrane in preparation for scission. James Pickett

Endocytosis: Bending around the BAR

10.1038/nrm2206 The process of clathrin-mediated endocytosis (CME) is driven by the collective effort of proteins that cause membrane invagination and those that cause membrane scission. Shimada et al. now report the structure of the F-BAR (Bin–amphiphysin–Rvs) protein domain, which provides insight into how these crescent-shaped modules drive membrane remodelling during endocytosis. CME occurs in three steps. First, clathrin assembles on flat membranes and captures endocytic cargo to form hemispherical clathrin-coated pits. Second, the pit slowly invaginates with a narrowing neck region, and actin polymerizes around the nascent vesicle. Finally, scission proteins such as dynamin are recruited to the neck region where they sever the membrane. The neck region is shaped by the recruitment of N-BAR domain proteins, such as amphiphysins and endophilins, which self-organize into banana-shaped dimers and exert this shape onto membranes. However, these proteins are recruited at a late stage of CME and it is unclear what drives the initial invagination of the small clathrin-coated pit. Shimada et al. determined the structures of the related F-BAR domains from formin-binding protein-17 (FBP17) and Cdc42-interacting protein-4 (CIP4). Compared with the N-BAR domain, the F-BAR domain also forms a dimer and has a similar crescent-shaped architecture; however, the F-BAR domain has longer α-helices and shallower curvature. Both N-BAR and F-BAR domains caused spherical lipid vesicles to tubulate, although F-BAR domains were found to bind preferentially to larger liposomes and generated tubules of a larger diameter. F-BAR domains therefore both ‘sense’ and induce less membrane curvature than the related N-BAR domain. The authors show that, as well as enforcing their curved shape onto membranes, F-BAR domains oligomerize into filaments that wind around the membrane to induce curvature. Three pieces of evidence support this: F-BAR domains are arranged in filaments in the crystal lattice, striations of F-BAR protein decorate tubulated liposomes, and point mutations that block the formation of filaments also impair membrane tubulation. Furthermore, F-BAR domain proteins also appear to harness the membrane-deforming capabilities of actin filaments — the recruitment of FBP17 to clathrin pits was associated with activation of the actin-nucleating machinery and actin polymerization. The authors propose that F-BAR and N-BAR proteins are recruited at different time points during CME. F-BAR domains recognize the shallow curvature of clathrin-coated pits to drive membrane invagination, whereas N-BAR domains are recruited to regions of already high curvature, where they further constrict the membrane in preparation for scission. James Pickett

Bending around the BAR: Endocytosis

10.1038/nrm2206 The process of clathrin-mediated endocytosis (CME) is driven by the collective effort of proteins that cause membrane invagination and those that cause membrane scission. Shimada et al. now report the structure of the F-BAR (Bin–amphiphysin–Rvs) protein domain, which provides insight into how these crescent-shaped modules drive membrane remodelling during endocytosis. CME occurs in three steps. First, clathrin assembles on flat membranes and captures endocytic cargo to form hemispherical clathrin-coated pits. Second, the pit slowly invaginates with a narrowing neck region, and actin polymerizes around the nascent vesicle. Finally, scission proteins such as dynamin are recruited to the neck region where they sever the membrane. The neck region is shaped by the recruitment of N-BAR domain proteins, such as amphiphysins and endophilins, which self-organize into banana-shaped dimers and exert this shape onto membranes. However, these proteins are recruited at a late stage of CME and it is unclear what drives the initial invagination of the small clathrin-coated pit. Shimada et al. determined the structures of the related F-BAR domains from formin-binding protein-17 (FBP17) and Cdc42-interacting protein-4 (CIP4). Compared with the N-BAR domain, the F-BAR domain also forms a dimer and has a similar crescent-shaped architecture; however, the F-BAR domain has longer α-helices and shallower curvature. Both N-BAR and F-BAR domains caused spherical lipid vesicles to tubulate, although F-BAR domains were found to bind preferentially to larger liposomes and generated tubules of a larger diameter. F-BAR domains therefore both ‘sense’ and induce less membrane curvature than the related N-BAR domain. The authors show that, as well as enforcing their curved shape onto membranes, F-BAR domains oligomerize into filaments that wind around the membrane to induce curvature. Three pieces of evidence support this: F-BAR domains are arranged in filaments in the crystal lattice, striations of F-BAR protein decorate tubulated liposomes, and point mutations that block the formation of filaments also impair membrane tubulation. Furthermore, F-BAR domain proteins also appear to harness the membrane-deforming capabilities of actin filaments — the recruitment of FBP17 to clathrin pits was associated with activation of the actin-nucleating machinery and actin polymerization. The authors propose that F-BAR and N-BAR proteins are recruited at different time points during CME. F-BAR domains recognize the shallow curvature of clathrin-coated pits to drive membrane invagination, whereas N-BAR domains are recruited to regions of already high curvature, where they further constrict the membrane in preparation for scission. James Pickett