His domain protein tyrosine phosphatase and Rabaptin-5 couple endo-lysosomal sorting of EGFR with endosomal maturation.

His domain protein tyrosine phosphatase (HD-PTP; also known as PTPN23) collaborates with endosomal sorting complexes required for transport (ESCRTs) to sort endosomal cargo into intralumenal vesicles, forming the multivesicular body (MVB). Completion of MVB sorting is accompanied by maturation of the endosome into a late endosome, an event that requires inactivation of the early endosomal GTPase Rab5 (herein referring to generically to all isoforms). Here, we show that HD-PTP links ESCRT function with endosomal maturation. HD-PTP depletion prevents MVB sorting, while also blocking cargo from exiting Rab5-rich endosomes. HD-PTP-depleted cells contain hyperphosphorylated Rabaptin-5 (also known as RABEP1), a cofactor for the Rab5 guanine nucleotide exchange factor Rabex-5 (also known as RABGEF1), although HD-PTP is unlikely to directly dephosphorylate Rabaptin-5. In addition, HD-PTP-depleted cells exhibit Rabaptin-5-dependent hyperactivation of Rab5. HD-PTP binds directly to Rabaptin-5, between its Rabex-5- and Rab5-binding domains. This binding reaction involves the ESCRT-0/ESCRT-III binding site in HD-PTP, which is competed for by an ESCRT-III peptide. Jointly, these findings indicate that HD-PTP may alternatively scaffold ESCRTs and modulate Rabex-5-Rabaptin-5 activity, thereby helping to coordinate the completion of MVB sorting with endosomal maturation.

Endosomal recycling tubule scission and integrin recycling involve the membrane curvature-supporting protein LITAF.

Recycling to the cell surface requires the scission of tubular membrane intermediates emanating from endosomes. Here, we identify the monotopic membrane protein LPS-induced TNF-activating factor (LITAF) and the related protein cell death involved p53 target 1 (CDIP1) as novel membrane curvature proteins that contribute to recycling tubule scission. Recombinant LITAF supports high membrane curvature, shown by its ability to reduce proteoliposome size. The membrane domains of LITAF and CDIP1 partition strongly into ∼50 nm diameter tubules labelled with the recycling markers Pacsin2, ARF6 and SNX1, and the recycling cargoes MHC class I and CD59. Partitioning of LITAF into tubules is impaired by mutations linked to Charcot Marie Tooth disease type 1C. Meanwhile, co-depletion of LITAF and CDIP1 results in the expansion of tubular recycling compartments and stabilised Rab11 tubules, pointing to a function for LITAF and CDIP1 in membrane scission. Consistent with this, co-depletion of LITAF and CDIP1 impairs integrin recycling and cell migration.

A Pseudomonas aeruginosa TIR effector mediates immune evasion by targeting UBAP1 and TLR adaptors.

Bacterial pathogens often subvert the innate immune system to establish a successful infection. The direct inhibition of downstream components of innate immune pathways is particularly well documented but how bacteria interfere with receptor proximal events is far less well understood. Here, we describe a Toll/interleukin 1 receptor (TIR) domain-containing protein (PumA) of the multi-drug resistant Pseudomonas aeruginosa PA7 strain. We found that PumA is essential for virulence and inhibits NF-κB, a property transferable to non-PumA strain PA14, suggesting no additional factors are needed for PumA function. The TIR domain is able to interact with the Toll-like receptor (TLR) adaptors TIRAP and MyD88, as well as the ubiquitin-associated protein 1 (UBAP1), a component of the endosomal-sorting complex required for transport I (ESCRT-I). These interactions are not spatially exclusive as we show UBAP1 can associate with MyD88, enhancing its plasma membrane localization. Combined targeting of UBAP1 and TLR adaptors by PumA impedes both cytokine and TLR receptor signalling, highlighting a novel strategy for innate immune evasion.

Dissecting the role of His domain protein tyrosine phosphatase/PTPN23 and ESCRTs in sorting activated epidermal growth factor receptor to the multivesicular body.

Sorting of activated epidermal growth factor receptor (EGFR) into intraluminal vesicles (ILVs) within the multivesicular body (MVB) is an essential step during the down-regulation of the receptor. The machinery that drives EGFR sorting attaches to the cytoplasmic face of the endosome and generates vesicles that bud into the endosome lumen, but somehow escapes encapsulation itself. This machinery is termed the ESCRT (endosomal sorting complexes required for transport) pathway, a series of multi-protein complexes and accessory factors first identified in yeast. Here, we review the yeast ESCRT pathway and describe the corresponding components in mammalian cells that sort EGFR. One of these is His domain protein tyrosine phosphatase (HD-PTP/PTPN23), and we review the interactions involving HD-PTP and ESCRTs. Finally, we describe a working model for how this ESCRT pathway might overcome the intrinsic topographical problem of EGFR sorting to the MVB lumen.

ESCRT-0 marks an APPL1-independent transit route for EGFR between the cell surface and the EEA1-positive early endosome.

Endosomal sorting complexes required for transport (ESCRT)-0 sorts ubiquitylated EGFR within the early endosome so that the receptor can be incorporated into intralumenal vesicles. An important question is whether ESCRT-0 acts solely upon EGFR that has already entered the vacuolar early endosome (characterised by the presence of EEA1) or engages EGFR within earlier compartments. Here, we employ a suite of software to determine the localisation of ESCRT-0 at subpixel resolution and to perform particle-based colocalisation analysis with other endocytic markers. We demonstrate that although some of the ESCRT-0 subunit Hrs (also known as HGS) colocalises with the vacuolar early endosome marker EEA1, most localises to a population of peripheral EEA1-negative endosomes that act as intermediates in transporting EGFR from the cell surface to more central early endosomes. The peripheral Hrs-labelled endosomes are distinct from APPL1-containing endosomes, but co-label with the novel endocytic adaptor SNX15. In contrast to ESCRT-0, ESCRT-I is recruited to EGF-containing endosomes at later times as they move to more a central position, whereas ESCRT-III is also recruited more gradually. RNA silencing experiments show that both ESCRT-0 and ESCRT-I are important for the transit of EGF to EEA1 endosomes.

ESCRT-III on endosomes: new functions, new activation pathway.

The multivesicular body (MVB) pathway sorts ubiquitinated membrane cargo to intraluminal vesicles (ILVs) within the endosome, en route to the lysosomal lumen. The pathway involves the sequential action of conserved protein complexes [endosomal sorting complexes required for transport (ESCRTs)], culminating in the activation by ESCRT-II of ESCRT-III, a membrane-sculpting complex. Although this linear pathway of ESCRT activation is widely accepted, a study by Luzio and colleagues in a recent issue of the Biochemical Journal suggests that there is greater complexity in ESCRT-III activation, at least for some MVB cargoes. They show that ubiquitin-dependent sorting of major histocompatibility complex (MHC) class I to the MVB requires the central ESCRT-III complex but does not involve either ESCRT-II or functional links between ESCRT-II and ESCRT-III. Instead, they propose that MHC class I utilizes histidine-domain protein tyrosine phosphatase (HD-PTP), a non-canonical ESCRT interactor, to promote ESCRT-III activation.

The Charcot Marie Tooth disease protein LITAF is a zinc-binding monotopic membrane protein.

LITAF (LPS-induced TNF-activating factor) is an endosome-associated integral membrane protein important for multivesicular body sorting. Several mutations in LITAF cause autosomal-dominant Charcot Marie Tooth disease type 1C. These mutations map to a highly conserved C-terminal region, termed the LITAF domain, which includes a 22 residue hydrophobic sequence and flanking cysteine-rich regions that contain peptide motifs found in zinc fingers. Although the LITAF domain is thought to be responsible for membrane integration, the membrane topology of LITAF has not been established. Here, we have investigated whether LITAF is a tail-anchored (TA) membrane-spanning protein or monotopic membrane protein. When translated in vitro, LITAF integrates poorly into ER-derived microsomes compared with Sec61ß, a bona fide TA protein. Furthermore, introduction of N-linked glycosylation reporters shows that neither the N-terminal nor C-terminal domains of LITAF translocate into the ER lumen. Expression in cells of an LITAF construct containing C-terminal glycosylation sites confirms that LITAF is not a TA protein in cells. Finally, an immunofluorescence-based latency assay showed that both the N- and C-termini of LITAF are exposed to the cytoplasm. Recombinant LITAF contains 1 mol/mol zinc, while mutation of predicted zinc-binding residues disrupts LITAF membrane association. Hence, we conclude that LITAF is a monotopic membrane protein whose membrane integration is stabilised by a zinc finger. The related human protein, CDIP1 (cell death involved p53 target 1), displays identical membrane topology, suggesting that this mode of membrane integration is conserved in LITAF family proteins.

The role of the cytoskeleton and molecular motors in endosomal dynamics.

The endocytic pathway is essential for processes that define how cells interact with their environment, including receptor signalling, cell adhesion and migration, pathogen entry, membrane protein turnover and nutrient uptake. The spatial organisation of endocytic trafficking requires motor proteins that tether membranes or transport them along the actin and microtubule cytoskeletons. Microtubules, actin filaments and motor proteins also provide force to deform and assist in the scission of membranes, thereby facilitating endosomal sorting and the generation of transport intermediates.

Biosynthesis of the polymeric gel-forming mucin MUC5B.

MUC5B is a major polymeric mucin in the airway mucus gel and is an essential component of innate defense of the respiratory epithelium. Knowledge of the synthesis and intracellular processing of MUC5B is incomplete. We investigated the molecular details of MUC5B assembly in primary human bronchial epithelial cells (HBECs) grown at an air-liquid interface (ALI). Electrophoretic and centrifugal separations of intracellular forms of MUC5B probed with antibodies specific for non-O-glycosylated and O-glycosylated forms of the mucin identified three major intracellular populations of MUC5B (non-O-glycosylated monomer and dimer, and O-glycosylated polymers). Biophysical analysis of recombinant MUC5B COOH-terminus (CT5B; D4-B-C-CK) expressed in 293-EBNA cells showed that MUC5B dimerizes by disulfide linkage. Pulse-chase studies in the HBEC ALI cultures showed that non-O-glycosylated MUC5B was synthesized within 20 min of metabolic labeling and O-glycosylated, polymeric mucin within 2 h. Radiolabeled O-glycosylated mucin polymers were secreted within 2 h and the majority were released by 48 h. These data indicate that MUC5B follows a similar assembly to the related glycoprotein, von Willebrand factor (vWF); however, unlike vWF the MUC5B polypeptide shows no evidence of major proteolytic processing of D-domains during the production of the mature secreted polymeric mucin in normal and cystic fibrosis (CF) primary bronchial epithelial cells. In contrast, MUC5B D-domains were modified by neutrophil elastase, a protease commonly found in CF sputum, demonstrating that proteolytic degradation of MUC5B is an extracellular event in CF sputum. These results define the pathway for synthesis of MUC5B in primary human goblet cells.