Cells communicate with their surrounding environments through interactions between the molecular components of the outer layer of the plasma membrane and external factors. For this reason, the composition of the membrane layers has to be tightly regulated. This kind of regulation is achieved by balancing two opposite, but equally important biological processes: exocytosis and endocytosis. Exocytosis ensures that newly synthesized or recycled membrane proteins and lipids are delivered to the external membrane while unnecessary and signaling chemicals are being expelled to the extracellular space. Contrary to exocytosis, endocytosis warrants the internalization of macromolecules and solutes into the cell. Here we briefly review the endocytosis process, two classification schemes for the known endocytic pathways and commonly used inhibitors and activators of these pathways.

While essential small molecules like amino acids, sugars and salts can cross the cellular membrane by diffusion or via specialized transport channels, macromolecules can only pass from membranes or the extracellular to the intracellular environment by endocytosis. This process involves the invagination of membrane-based vesicular carriers and the release of cargo for further processing, degradation or recycling in the intracellular space [1-3]. Among the internalized molecules are nutrients, growth factors, and plasma membrane proteins and lipids, including ion channels, transporters, and signaling ligands and receptors. Endocytosis is also the principal route of entry to cells for nanoparticle therapeutics [4].

Endocytosis is fundamental to cells. By controlling the composition of the plasma membrane, it controls essential cellular processes like nutrient intake, cell adhesion, junction formation, migration, cell polarity, and signal transduction. Endocytosis is also involved in specific cellular activities like neurotransmission [5], viral, bacterial and parasite infections [6-8], such as drug resistance in malaria [9], or clearing of apoptotic cells and cellular debris [10-13]. For example, dendritic cells and macrophages ingest the equivalent of their own cellular volume of extracellular fluid every 4 hours through pinocytosis [14]. Thus, it is essential to higher eukaryotic life and immunity but also exploited by pathogens that seek their internalization into host cells.

Endocytosis occurs by a variety of mechanisms that are being dictated by the diversity of cargoes and cargo-specific transport kinetics, with the involvement of many genes. Table 1 lists the commonly studied genes involved in endocytosis and their antibodies cited among the articles in Labome's Validated Antibody Database. The different endocytosis pathways have been classified based on either the morphologic features of the membranous structures or based on their requirement for certain key components. Among the characterized pathways are clathrin-mediated endocytosis, caveolar-type endocytosis, CLIC/GEEC-type endocytosis, phagocytosis, macropinocytosis [14], circular dorsal ruffles, and the putative flotillin-dependent endocytosis, IL2Rbeta endocytic pathway and Arf6-dependent endocytosis [59]. Membrane-associated periodic skeleton (MPS) may repress endocytosis in neurons [60], while pre-recruited dynamin 1xA can speed up synaptic vesicle recycling by orders of magnitude [30].

[enlarge]

Figure 1. Classification of endocytosis pathways based on clathrin and dynamin-involvement - schematic diagrams: A) Clathrin and dynamin-dependent endocytosis; B) Clathrin-independent dynamin-dependent endocytosis; C) Clathrin and dynamin-independent endocytosis.

The best-characterized endocytosis pathway is the clathrin-mediated pathway (CME). CME requires the formation of clathrin-coated membrane pits, and it is the primary endocytic route. It is constitutively active and supports all housekeeping cellular functions [61, 62], thus ensuring the cell’s homeostasis by balancing the secretory pathways. All other mechanisms are clathrin-independent (CIE) and are activated by specific stimuli (Figure 1). More recently an alternative classification, which takes into consideration the membrane lipid rafts, has been proposed [63] (Figure 2, see below).

[enlarge]

Figure 2. Lipid-rafts based classification of endocytosis pathways - schematic diagrams: A) Membrane domains without lipid rafts; B) Mixed membrane domains; C) Membrane domains in lipid rafts.

Clathrin-mediated endocytosis (CME), is also known as clathrin-dependent endocytosis (CDE) or receptor-mediated endocytosis. The pathway has been well studied in cell culture, yeast, and mammalian cells and shows a high degree of conservation [59, 64]. The mechanism involves several steps through which cell surface receptors are specifically recognized by adaptor proteins, concentrated and internalized into small (~50-100 nm diameter) clathrin-coated membrane vesicles [59, 64, 65]. The process was first described over 50 years ago and uses more than 50 proteins involved in different steps. Multiple recent studies reviewed by Kaksonen and Roux [62], revealed important insights into how endocytic vesicles are produced and released into the cytoplasm. Figure 1 indicates the main steps of vesicle traffic, from initiation, maturation, release from the plasma membrane and uncoating [66].

The sequence of these steps is similar in yeast and eukaryotic cells, but the requirement for key proteins differs [67]. In all cells, the process initiates by the adaptor and accessory protein-mediated coordination of clathrin at nucleation sites on the plasma membrane to form the clathrin-coated pits (CCPs); polymerization of clathrin and/or actin into curved lattice structures; followed by formation of vesicular necks. Further on, the necks are constricted to bring distant membrane regions nearby. In mammals, the scission protein dynamin, a large GTPase, forms a helical polymer around the neck and mediates the release of the vesicle from the plasma membrane [68]. This process requires GTP hydrolysis, is irreversible and triggers the release of clathrin from the vesicle [68]. In yeast, actin only is essential for scission [69, 70]. While clathrin is required for vesicle formation in mammalian cells, actin and dynamin are not. In contrast, the endocytic process in yeast requires actin for vesicle formation [71], while clathrin in nonessential [67, 72]. Also, dynamin is essential for vesicle release from the plasma membrane in mammalian cells, while it is not essential in yeast [67, 72].

Despite mechanistic differences, a large group of very similar endocytic accessory proteins (EAPs) are used in both mammalian and yeast cells. These are molecules that function as scaffolds, recruiters, sensors or regulators of the multiple steps in the endocytic process from clathrin-coated pits (CCPs) generation to membrane curving and internalization, vesicle fission, uncoating and recycling [67, 73].

As with most biochemical pathways the initial step in endocytosis, initiation of CCP formation, is strongly regulated. The main regulatory factor is the heterotetrameric adaptor protein AP2. AP2 interacts both with the 4,5-phosphatidylinositol biphosphate (PtdIns(4,5)P2) molecules concentrated at the plasma membrane and the cargo proteins. These interactions induce conformational changes in the body of the AP2 mediator that exposes a hidden clathrin-binding site. Clathrin is now able to assemble at the chosen membrane location, and other EAPs can be recruited to continue the process of CCP formation [73, 74]. Among the early EAPs participants, commonly referred to as “ endocytic pioneers” [73], are the scaffold proteins FCHo1/2, Eps15 and intersectin and adaptor proteins NECAp, CALM and Epsin. These factors participate in weak interactions among them and with AP2 to induce conformational changes on the AP2 mediator protein. These changes are translated into stable AP2 clusters, enhanced clathrin recruitment and CCP growth and stability [75-77]. In addition, AP2 is stabilized in its PtdIns(4,5)P2, cargo, and clathrin–binding conformation by phosphorylation events [78-81].

Even though clathrin and adaptor proteins were found to spontaneously form clathrin coats in solution [82, 83], in vivo the assembly of CCPs and their maturation into CCVs is regulated by multiple factors [73, 84] among which a series of lipid kinases and phosphatases that add or remove phosphate to convert the PtdIns(4,5)P2 into various versions of the polyphosphoinositide [85-87]. Also, endocytic cargoes, like signaling receptors, can control the rate of maturation of their carrier CCPs. One example is the ligand-bound G-protein coupled receptors (GPCR) which were shown to stimulate the process of CCP maturation upon ubiquitination and recruitment of the additional factor Epsin 1 [88]. Another example is the TRAIL (tumor necrosis factor α-related apoptosis-inducing ligand)-induced endocytosis of specific apoptosis-related receptors [89].

Membrane scission, the final step of endocytosis, is facilitated by the GTP-ase dynamin in mammals and by actin in yeast. In the case dynamin, two different isoforms, Dyn1 and Dyn2, can participate. While Dyn1 can release vesicles both from planar artificial lipid structures and from pre-formed membrane invaginations, Dyn2 senses the curvature but requires the assistance of other curvature-generating factors to catalyze fission [90, 91]. Being an important player in the process, dynamin is itself highly regulated. Regulation of Dyn1 is better understood that the regulation of Dyn2, but it is clear that both forms are regulated by allosteric [92, 93] or posttranslational mechanisms [94-96].

Once separated from the plasma membrane, the CCVs are rapidly losing their clathrin coat with the help of several other factors like the ATPase/chaperone Hsc70 [97-101] or the cyclin-G-associated kinase/auxilin2 (GAK/auxilin2) [100, 102, 103]. At the same time, PI4P is being regenerated, and the AP2-membrane interactions are lost [104].

In contrast to CME, clathrin-independent endocytosis (CIE) is not constitutively active and comes in many forms depending on a variety of stimuli and cargos. In spite of being studied for a shorter time, it has become generally accepted that CIE occurs and that it occurs via different mechanisms that can be either dynamin-dependent or independent [105].

Examples of clathrin-independent dynamin-dependent endocytosis pathways include the pathway responsible for uptake of IL-2 [106-109], pinching off of tubules induced by toxins like Shiga toxins [110-112], endocytosis of the epidermal growth factor (EGF) and of its receptor (EGFR) [113-115], and the most recently described fast endophilin-mediated endocytosis (FEME) [108-110, 116, 117]. For example, endophilin-A2 dependent VEGFR2 endocytosis is involved in angiogenesis [53]. Endophilin can transform into liquid-like condensates, which activate the generation of multi-protein assemblies [118]. This transformation can be induced by binding partners of endophilin such as the third intracellular loop (TIL) of the β1-AR and the C-terminal domain of lamellipodin (LPD) [118].

The caveolae-dependent endocytic pathway [119-122] is the second most studied endocytic route, is also clathrin-independent and requires dynamin. Caveolae are stable bulb-shaped plasma membrane structures [123-125] with a role in signaling [126, 127], internalisation of integrins and glycosphingolipids [128], and regulation of cellular adhesion [129]. Besides, caveolae are involved in the uptake of albumin [130] and toxins [123, 131]. However, toxins uptake can be done via multiple pathways, including clathrin-dependent endocytosis [132-134].

Biogenesis of caveolae is induced by the expression of a family of integral membrane proteins, the caveolin proteins [135, 136] that are essential for each caveola formation [137]. Caveolin-1, the principal structural protein of caveolae membranes, and caveolin-2 (Cav1 and Cav2) are expressed in a wide range of tissues [136, 138, 139], while caveolin-3 (Cav3) is tissue-specific as it is the only isoform expressed in skeletal and cardiac muscle [136, 140, 141]. The formation and function of caveolae is regulated by highly conserved components of the cavin protein complex, that includes: the polymerase I and transcript release factor (PTRF) or cavin-1, the serum deprivation response protein (SRD) or cavin-2, the serum deprivation response factor-related gene product that binds to C-kinase (SRBC) or cavin-3 and the muscle-restricted coiled-coil protein (MURC) or cavin-4. A few recent reviews describing the role of cavins in caveolae assembly and function are available [23-25]. The existence of caveolae with specialized functions has been hypothesized, based on the absence of certain key players like dynamin-2 in some caveolae [125].

In addition to dynamin, macropinocytosis of vesicles originating from circular ruffles [142, 143] and phagocytosis [144] may also require actin, thus inhibitors against cytoskeleton, such as cytochalasin D [145], can be used to block macropinocytosis. Viruses [146], protozoans [147] and other pathogens [148], as well as cancers [148, 149] and apoptotic cells [10, 11], and double-strand DNAs and the 2C10 anti-DNA antibody [145] are known to take advantage of these pathways. Detailed descriptions of the clathrin-independent, dynamin or actin dependent pathways and the most recent findings related to them can be found in these three reviews [105, 109, 112].

More and more endocytic pathways that are both clathrin and dynamin-independent are being described. The CLIC/GEEC pathway is Cdc42 dependent and [109, 150] is responsible for fluid phase intake. The pathway has been shown to take in GPI-anchored proteins [151], protein toxins [152, 153] and transmembrane proteins [154] and may be involved in cancer [155]. The ADP-ribosylation factor Arf1 and its guanine nucleotide exchange factors GEF1 are thought to have a role in this pathway [156, 157].

The nuclear co-repressors of transcription, C terminal-binding protein (CtBP) / brefeldin A-ribosylated substrate (BARS) was found to control fission in fluid-phase endocytosis in the absence of dynamin [158, 159]. CtBP/BARS was also shown to be involved in macropinocytosis [160] an endocytic pathway that can be either dynamin-dependent (see above) or independent [109, 161]. The acetyl-CoA-producing enzyme, ATP citrate lyase, is an important regulator of micropinocytosis [162]. Heterodimeric actin capping protein is involved in the actin remodeling processes, which regulate ruffling and macropinocytosis [162].

The cell membrane contains clusters of lipids that show a higher degree of organization that exists within the rest of the membrane lipid bilayer. These clusters are known as “lipid rafts” [163-165] and function as anchoring platforms for membrane proteins [166] that function in cell signaling, membrane trafficking, endocytosis, viral infections and apoptosis [167, 168].

The endocytic pathways can be classified based on the lipid rafts composition of the primary endocytic vesicle (Figure 2) [63]. Thus, there are endocytic pathways that do not involve lipid rafts in the endocytic vesicle, e.g., clathrin-mediated endocytosis (CME); endocytic pathways for vesicles that can contain both lipid rafts and non-raft membrane regions (phagocytosis and macropinocytosis); and endocytic pathways that take occur in lipid rafts. The latter include the clathrin-independent endocytic pathways presented in the previous sections and others [63].

In spite of the constant changes in the classification of the membrane internalization pathways, it is clear that at least three major pathways to generate endocytic carriers exist in eukaryotic cells: assembly of clathrin-coated vesicles, internalization via lipid raft plasma membrane domains followed by formation of large vacuoles mediating phagocytosis, and macropinocytosis. Regardless of its positions in a classification scheme, every endocytosis pathway is controlled by complex mechanisms that involve multiple proteins. Some of the components of these types of machinery are common to several pathways, while others are specific. Researchers often use methods in which a specific molecular component is selectively inhibited or activated to understand the biological roles of individual routes of endocytosis [169]. For example, Flaherty SE et al inhibited macropinocytosis, but not micropinocytosis, in bone marrow–derived macrophages with 100 uM LY294002 from MilliporeSigma for one hour [170]. Pharmacological inhibitors, RNA interference, and expression of dominant negative mutants are common approaches. Pathway classification is useful for the identification of such specific regulators among the common ones.

Although many commonly used compounds are less specific than initially thought, pharmacological inhibitors and activators remain popular with investigators [171, 172]. Tables 2 and 3 list commonly used inhibitors and activators of endocytosis. Bafilomycin A1 is also an inhibitor of autophagy process [173].

Intrabodies, or intracellular antibodies [212], have been recently used as alternatives to pharmacological inhibitors to target specific endocytic participants. For example, several intrabodies have been tested as inhibitors of β-arrestins [213], the scaffold proteins that support clathrin and the AP2 complex in promoting endocytosis of G protein-coupled receptors (GPCRs) [214]. Of these, an antibody fragment known as Fab5 was found to selectively disrupts the β arr2–clathrin interaction [213]. Moreover, the single-chain version of Fab5 - referred to as scFv5 - efficiently inhibited endocytosis of various GPCRs [213]. pHluorin, a conjugate of a pH-sensitive form of GFP and the VAMP luminal fragment, has been used to study the release and endocytosis of neurotransmitter. Awasthi A et al used pHluorin to study the endocytosis of Syt3 in hippocampal neurons [215]. Integrin cytoplasmic domain-associated protein-1 (ICAP-1) acts as an integrin inhibitor regulating mechanotransduction by interacting with nucleoside diphosphate kinase 2 to control CME of integrins [209]. ICAP-1 deficiency stimulates β3-integrin-mediated force generation. Wiskostatin and its analogues were found to function as inhibitors of helical dynamin GTPase activity and CME [210].

Parts of tables and a small portion of the text originally appeared in the article Activators and Inhibitors in Cell Biology Research.