Recent advances in clathrin-independent endocytosis

Dynamin-dependent clathrin-independent endocytic pathways

Endocytic pathways that do not use clathrin can either use the scission GTPase dynamin or be independent of it. They are known to be more sensitive to the dynamics of the cytoskeletal actin machinery for stabilization and scission². Recently, a role for BAR domain proteins has emerged in multiple clathrin-independent endocytosis (CIE) pathways. The members of the BAR domain protein endophilin—primarily endophilin-A2 and endophilin-A1—directly bind to and facilitate the endocytosis of selected G-protein-coupled receptors like β1-adrenergic receptor (β1-AR) and dopaminergic and acetylcholinergic receptors⁶. Endophilin and lamellipodin have previously been implicated in the clathrin-mediated endocytosis of epidermal growth factor receptor (EGFR)⁷. During the CIE of β1-AR, rapidly formed vesicles labeled by endophilin and its binding partners, synaptojanin and dynamin, internalized β1-AR from the leading edge of cells in culture and mediated the downregulation of its signaling. This novel endocytic pathway was named the fast endophilin-mediated endocytic pathway (FEME) and requires the lipid phosphatidylinositol 3,4-bisphosphate—PI(3,4)P2—as well as the protein lamellipodin, which binds both PI(3,4)P2 and the SH3 domain of endophilin for its recruitment to the leading edge. The class I PI(3)K pathway and associated phosphatases, including the 5′ phosphatase SHIP1/2, act near the plasma membrane to locally enhance the PI(3,4)P2 levels. This assists in the formation of FEME vesicles enabled by the recruitment of endophilin⁶.

Endophilin-A2 has also been recently implicated in the CIE of glycosphingolipid (GSL)-binding toxins Shiga (STxB) and cholera toxin B⁸. These toxins induce local membrane curvature upon binding to their receptor, following which the BAR domain-containing protein is recruited to the cholesterol-dependent curved membrane surfaces^8,9. Cortical actin dynamics were previously shown to be necessary for scission of STxB endosomes¹⁰. With reconstituted model membranes, endophilin-A2 was observed to stabilize the tube-shaped StxB vesicles. Actin, endophilin-A2, and dynamin independently and additively contribute to the scission of these vesicles, and the authors propose a model wherein endophilin-A2-stabilized invaginations are acted upon by the pulling forces of cytoskeletal motors (possibly microtubule-associated motors¹¹) to lead to scission at a controlled rate and length⁸.

Clathrin- and dynamin-independent endocytic pathways

The reliance on actin cytoskeletal machinery has also been studied in a clathrin- and dynamin-independent, CLIC/GEEC (CG) pathway. This is the major route of internalization of a large fraction of glycosylphosphatidylinositol-anchored proteins (GPI-APs) and fluid phase in several mammalian and Drosophila cell lines and tissues^1,12–16. The CG endocytic intermediates carrying GPI-APs and fluid phase markers were visualized by electron microscopy to reveal uncoated, tubular intermediates¹⁷ called clathrin-independent carriers (CLICs), which eventually fuse to form GPI-AP-enriched early endosomal compartments (GEECs)¹⁸. Arf1, its GEF (guanine exchange factor), GBF1/Garz (gartenzwerg), and membrane cholesterol are necessary for forming the CG endocytic vesicles in this pathway^1,12,15,19. In a screen in S2R⁺ insect cell lines to identify molecular players regulating this pathway, Arf1-COP1 machinery, BAR domain-containing proteins, vacuolar ATPase, lysosomal genes involved in vacuole biogenesis, and actin remodeling factors, including Slingshot, Coronin, Arcpc1, and Capping protein, were among the major hits¹⁹. Actin inhibitors used at concentrations that disrupt the cortical actin network without affecting the long-lived structures like stress fibers selectively perturb CG endocytosis while not affecting clathrin-dependent uptake¹². This acute sensitivity to actin dynamics also translates to the regulation of CG endocytosis by Cdc42, a modulator of actin polymerization.

Recently, the recruitment of these early molecular players at nascent endocytic vesicles was studied at high spatial and temporal resolution to understand the sequence of events leading to the formation of a CG endocytic vesicle²⁰. While reaffirming the role of ARF1-GBF1, the actin nucleator Arp2/3, and associated GTPase Cdc42 at the nascent sites of endocytosis, this study also identifies the role of a BAR domain protein, IRSp53, and an ARP2/3 inhibitor, PICK1. Imaging the direct recruitment of these molecular players also helps develop a mechanistic picture of how a GEEC is formed: initial recruitment of ARF1/GBF1, IRSp53, and Arp2/3 to the cell surface is followed by the arrival of Cdc42 that activates Arp2/3 and IRSp53 to catalyze the formation of F-actin and membrane buckling, leading to the formation of the CG endosome²⁰. The ARF1/GBF1 (Garz)-dependent CG pathway was also shown to be active in Drosophila wing disc tissue, wherein class I PI3-kinase activity is necessary for the recruitment of Garz to the cell surface and the initiation of the CG endocytosis¹⁶.

Besides the role of the actin machinery, membrane composition is a key component that determines the enrichment of clathrin-independent cargoes in endocytic pits². Perturbations in cholesterol and sphingolipid levels affect the distribution of GPI-APs, normally present as actively maintained nanoclusters on the cell surface²¹, and subsequently the formation of GEEC endosomes¹². The CIE of cholera and Shiga toxins too is initiated upon their binding to specific GSLs, which induces membrane bending and endosome formation^8,9,22. Endogenous galectins, which can bind glycan chains and GSLs, are postulated to use a similar mechanism for concentration and multimerization leading to membrane bending, potentially facilitating the internalization of a large variety of endogenous glycosylated proteins. This GSL-galectin-3-mediated mechanism (termed GL-Lect²²) can mediate the CIE of β1-integrin and CD44, dependent on their glycan chains²³. Recently, galectin-3 binding was also implicated in the endocytosis of the GPI-AP CD59 and major histocompatibility complex (MHC) class I protein, suggesting a more global utilization of this axis of interaction to lend specificity to CIE²⁴. Interestingly, the degree of glycan branching affects the internalization rates in contrasting ways. CD59 internalization is obstructed by a highly branched glycan lattice. This raises the possibility of clathrin-independent endocytic fluxes being selectively altered by factors affecting N-glycosylation, like metabolic substrates²⁴.