Endocytosis is a process carried out by eukaryotic cells to internalize extracellular molecules, plasma membrane proteins and lipids (Doherty and McMahon, 2009). While several other pathways for endocytosis such as caveolin-mediated endocytosis, phagocytosis and macropinocytosis have been described, clathrin mediated endocytosis (CME) remains the major route for internalization of many membrane lipids and proteins (Kaksonen and Roux, 2018).

CME was first observed by Roth and Porter in 1964 where they found uptake of yolk-containing bristled-coated pits in the mosquito oocyte (Roth and Porter, 1964). Later these bristled-coated structures isolated from pig brain were identified as coat proteins and named ‘Clathrin’ by Barbara Pearse (1975, 1976). Since then, this process has been extensively studied and although we have a fairly good understanding of the process itself, many unanswered questions still remain about how over 50 molecules that take part in this molecular process (Haucke and Kozlov, 2018), come together in a highly coordinated manner.

CME is characterized by the recruitment of clathrin and its associated molecules to the plasma membrane allowing the formation of clathrin-coated vesicles (CCVs). The formation of CCVs involves the polymerization of ‘clathrin triskelia’, which are the basic building blocks of the clathrin coats (Kaksonen and Roux, 2018). A triskelion is composed of three clathrin heavy chains (CHC) (∼190 kDa) each of which is associated with a smaller clathrin light chain (CLC) (∼25 kDa). While the major role of the clathrin heavy chain is in intracellular trafficking, it is also involved in several other processes including chromosomal segregation during mitosis (Royle et al., 2005), regulation of basal NF-κB activity in epithelial cells (Kim et al., 2011), control of neuropeptide degradation and secretion during neuronal development (Nahorski et al., 2018), and maintenance of mouse embryonic stem cell pluripotency (Narayana et al., 2019; Mote et al., 2020).

Variations in the clathrin heavy and light chains alter the biophysical properties of the clathrin lattice, in turn affecting trafficking of receptors and thereby several physiological functions of the cell. The heavy chain is essential for triskelion assembly and for all clathrin-dependent endocytic events, with a number of excellent reviews highlighting the function of this protein (Kirchhausen, 2000; Brodsky, 2012; Kirchhausen et al., 2014; Brodsky, 2016; Kaksonen and Roux, 2018; Briant et al., 2020). In contrast, the role of the clathrin light chains remains relatively under-explored. In this review, we look at how the clathrin light chains affect clathrin polymerization, vesicle formation, receptor trafficking and cell signaling.

In metazoans, the clathrin heavy chain protein is encoded by a single gene, Cltc. In humans, due to large-scale gene duplications during chordate evolution, there are two CHC paralogs, CHC17 (encoded by Cltc) and CHC22 (encoded by Cltcl1) based on their location on chromosome 17 and 22, respectively. Although Cltcl1 is found in several other vertebrate species, it is functional only in humans. In mice, only a pseudogene for CHC22 is present (Wakeham et al., 2005). In yeast and invertebrates such as Drosophila and Caenorhabditis elegans, the CHC protein is encoded by a single gene. Plants have two genes for the clathrin heavy chain, CHC1 and CHC2 (Baisa et al., 2013).

In invertebrates, the clathrin light chain is encoded by a single gene. However, as a result of local gene duplication, higher eukaryotes have two light chains, CLCa and CLCb encoded by the genes Clta and Cltb, respectively (Wakeham et al., 2005). They both share 60% homology in their amino acid sequence but are expressed at different levels in various vertebrate tissues (Wu et al., 2016). Despite having considerable divergence in sequence, the single light chain from yeast shares various physical properties with mammalian light chains (Silveira et al., 1990). In plants, the three clathrin light chain genes CLC1, CLC2 and CLC3 (Scheele and Holstein, 2002; Baisa et al., 2013) share at least 30% sequence homology with mammalian CLCs (Wang et al., 2013).

Vertebrate CLCs contain a consensus region of 22 amino acids shared by both CLCa and CLCb. Additionally, they also include distinct domains for binding to calcium, clathrin heavy chain, calmodulin and a neuron-specific insertion sequence (Brodsky, 2012). CLCa contains a unique Hsc70 binding region (DeLuca-Flaherty et al., 1990). However, functions have been attributed to only some of these domains. A detailed representation of CLC domain organization can be found in Figure 1.

In mammals, at the N-terminus, a 22 amino acid conserved sequence is shared by CLCa (residues 28–49) and CLCb (residues 20–41) with the negatively charged residues, EED responsible for CLC binding to the CHC knee (Brodsky, 2012). This conserved sequence is also the binding site for the Huntington interacting protein (HIP) family (Chen and Brodsky, 2005), and plays a role in regulating clathrin self-assembly (Legendre-Guillemin et al., 2005; Ybe et al., 2007a).

CLCa has an Hsc70 binding sequence that was shown to stimulate uncoating in vitro (DeLuca-Flaherty et al., 1990). However later studies suggested that uncoating of vesicles could also be done in the absence of CLCs in vitro (Ungewickell et al., 1995) Both the light chains also have a calcium binding region (Nathke et al., 1990) and calmodulin binding domain (Pley et al., 1995) present at the centre and C-terminal, respectively. While these domains have been found to play a role in in vitro studies, no function has been attributed to them in vivo.

In Dictyostelium, overexpression of the C-terminal fragment of CLCa in clc null cells produced dynamic punctae distribution along the plasma membrane and within the cytoplasm, similar to full-length CLCa (Wang et al., 2006), indicating that its function can be attributed almost entirely to the C-terminal domain.

In vitro assembly of the clathrin hub, and CHC trimer stability is enhanced by the C-terminal domain of the light chain (Ybe et al., 2007b). Co-expressing the trimer-defective hub heavy chain mutant C1573A, along with the light chain C-terminal domain construct could achieve approximately 67% of wild-type clathrin assembly.

CLCs undergo alternate mRNA splicing in vertebrates, giving rise to four isoforms for CLCa and two for CLCb (Blue et al., 2018a). Exons 5 and 6, encoding 18 and 12 amino acids respectively, in the Clta gene are alternatively spliced resulting in four isoforms. These are: 1) neuronal CLCa (nCLCa) containing both 18 and 12 amino acid residue insertions; 2) a splice variant containing only the 18 residue-insertion found only in brain; 3) a splice variant containing only the 12 residue-insertion found in brain, heart and skeletal muscle; and 4) a splice variant without either insertion.

In vertebrates, the two splice variants for CLCb include an isoform having an insert of 18 residues present in neurons (nCLCb) and another that lacks the insert in non-neuronal tissues (Wong et al., 1990; Blue et al., 2018a). In rats, the CLCb gene contains six exons. The isoform containing all exons is brain-specific (LCB2), while the isoform lacking exon 5 (LCB3) is present in other tissues. LCB2 is predominantly present in primary rat neuronal cultures, whereas LCB3 is present in primary rat glial cultures (Stamm et al., 1992).

Due to the insertions mentioned above, neuronal splice variants have a higher molecular mass than CLCs in other cell types. Under oxidizing conditions, CLC isoforms can form internal disulfide bonds. Both brain-specific CLCa isoforms, contain the 12 residue insert in exon 5, allowing the formation of internal disulfide bonds between two cysteine residues in vitro, while the smallest CLCa isoform only has a single cysteine residue. Both CLCb isoforms have two cysteine residues present at the C-terminal. In vitro purification of CLCs in the presence or absence of thiols or alkylating agents caused an alteration in electrophoretic mobility depending on the formation of disulfide bonds (Parham et al., 1989). The electrophoretic mobility change of CLCs was also found to be species- and tissue-specific. This could be due to the presence of different isoforms (Ungewickell, 1983), disulfide bond formation (Parham et al., 1989), or other post-translational modifications such as phosphorylation (Ferreira et al., 2012).

At all developmental stages, the Clta transcript in the mouse heart does not include exon 5. However, due to alternative splicing, Clta exon 6 is included at a low level at birth with its inclusion increasing in adulthood (Giudice et al., 2014; Blue et al., 2018b). This suggests that expression of CLC splice variants is variable and can change with the stage of development, and in a tissue-specific manner.

Both the CLCs are developmentally and tissue-specifically regulated by alternative splicing but the physiological and functional implications of different splice variants of CLCs still remain unexplored. Although we still lack complete understanding of tissue-specific expression patterns of alternatively spliced variants of CLCs, one can speculate that the presence or absence of particular exons may lead to a change in interacting partners thereby influencing function.

Despite having 60% sequence similarity, the two light chains are diverse in nature due to alternative splicing, internal disulfide bond formation and tissue specific expression patterns, allowing speculation that their ability to perform distinct physiological and functional roles could be attributed to such differences.

The phosphorylation of clathrin light chains was initially identified in vitro in coated vesicles isolated from bovine brains (Usami et al., 1985). Using rat reticulocytes, Bar-Zvi et al. (1998), then demonstrated that unassembled pools of CLCb were highly phosphorylated. Further investigations revealed that CLCb underwent phosphorylation mediated by Casein Kinase 2 at the N-terminal serine residues at positions 11 and 13. These residues are unique to CLCb and absent in CLCa (Hill et al., 1988). However both CLCa and CLCb contain the phosphorylation site for G-Protein coupled receptor kinase 2 (GRK2) at Ser204. Mutation of all phosphorylation sites in CLCb impeded internalization of purinergic GPCRs, P2Y₁ and P2Y₁₂, with phosphorylation of Ser204 being specific for P2Y₁₂ uptake (Ferreira et al., 2012). Phosphorylation of CLCb at Ser204 was also required for lattice rearrangement and curvature generation by regulating clathrin exchange in a cargo-dependent manner (Maib et al., 2018). Together these findings indicate a role for phosphorylated forms of clathrin light chains in regulating the uptake of specific membrane-resident proteins.

Both CLCs can bind and regulate CHC17, but do not functionally interact with CHC22 (Towler et al., 2004). Previous studies have shown that CLCs bind to the proximal leg of the heavy chain via their central region (Kirchhausen et al., 1983; Brodsky et al., 1987; Jackson et al., 1987; Nathke et al., 1992; Liu et al., 1995). Using a yeast-two hybrid system it was shown that the core interaction occurs between CHC residues 1,267–1,522, and CLCb residues 90–157 (Chen et al., 2002). Mutations in the central region (residues 90–157) of CLCb disrupt the alpha-helical structure suggesting that this region is crucial for interaction with CHC. Cryo-EM based structural analysis revealed that two tryptophan residues (W105 and W127) were required for light chain binding to the heavy chain. Mutation of W105 to arginine disrupted CLC-CHC binding, but could be rescued by mutation of lysine to glutamate at residue position 1,326 of the heavy chain. Additionally, two helices present in the CLC trimerization domain (TxD) tended to form stable association with two heavy chain TxDs in trans conformation, connecting adjacent legs and forming the triskelion vertex (Morris et al., 2019).

The physiological importance of light chains has mostly been studied using knockdown or knockouts of the light chains, or through the use of mutant forms to study receptor trafficking (Huang et al., 2004; Poupon et al., 2008; Majeed et al., 2014; Wu et al., 2016; Redlingshöfer et al., 2020). Knockout of CLCa in mice hampered the internalization of Transforming growth factor β receptor2 (TGFβR2), affecting antibody isotype switching in B lymphocytes (Wu et al., 2016). Knockdown of both light chains in mammalian HeLa cells did not affect the internalization of β1 integrin but disrupted its recycling back to the plasma membrane (Majeed et al., 2014). CLC knockdown (KD) also altered the targeting of cation-independent mannose-6 phosphate receptor (CI-MPR) to the endosome, resulting in clustering of the receptor near the trans-Golgi network, leading to a delay in processing of the lysosomal hydrolase cathepsin D in HeLa and Cos7 cells (Poupon et al., 2008).

Internalization of GPCRs was also shown to be dependent on CLCb phosphorylation (Ferreira et al., 2012). Internalization of P2Y₁₂ receptor, a member of a family of purinergic GPCRs, in 1321N1 astrocytoma cells is regulated by phosphorylation of CLCb. Trafficking of low-density lipoprotein on the other hand, is not affected by the removal of light chains (Poupon et al., 2008). Furthermore, internalization of EGFR was also not affected by siRNA-mediated knockdown of both the light chains in HeLa cells (Huang et al., 2004). However, a study using single light chain-expressing H1299 cells, a non-small cell lung cancer cell line, showed accelerated internalization of EGFR in cells that expressed only CLCb in contrast to wild type and CLCa-only expressing cells (Chen et al., 2017). Similar observations have been made with respect to internalization of transferrin receptor (Tfr) (Huang et al., 2004; Chen et al., 2017). Tfr internalization can also be dependent on the phosphorylation status of CLCb. When transferrin receptor is clustered with other cargo, its uptake can become sensitive to the status of CLCb phosphorylation. In 1321N1 human astrocytoma cells, packaging of Tfr with P2Y₁₂ receptor resulted in delayed internalization of Tfr in presence of a phosphorylation-deficient mutant of CLCb (Maib et al., 2018), with similar results also observed in HeLa cells. Knockdown of both the light chains attenuated Tfr recycling in HeLa cells (Majeed et al., 2014), which remains unaffected in single light chain expressing H1299 cells (Chen et al., 2017).

From these studies one can infer that i) dependence of receptor trafficking on light chains is influenced by other factors such as the cell type and presence of other cargo; ii) CLCa and CLCb can differentially affect cargo uptake; and iii) phosphorylation status of CLCb can potentially be a method of regulating receptor trafficking.

Altered trafficking of receptors can compromise cell signaling, which is an important regulator of several physiological functions. The sections below discuss how light chains regulate important biological functions and pathological conditions. These are also summarized in Table 1.

The role of actin in endocytosis is well established. Actin is recruited to sites of endocytosis and helps the membrane to invaginate and form coated pits (Engqvist-Goldstein and Drubin, 2003; Mooren et al., 2012). Clathrin light chains can bind to Hip1/R proteins through a conserved domain present at their N-terminus, which in turn binds to actin (Chen and Brodsky, 2005). The light chains therefore act as a connecting link between the endocytic machinery and the cytoskeleton.

Hip1/R proteins can bind to actin through their THATCH domain independent of CLCs. Binding of the light chains to the coiled-coil domains of Hip1 and Hip1R reduce their actin-binding activity. This suggests that Hip proteins do not interact with actin while incorporated into the clathrin coat. Instead Hip proteins interact with actin at the neck of the budding vesicle or edge of the clathrin coat, promoting development of a budding vesicle (Wilbur et al., 2008; Boettner et al., 2011). Hip1 binding to CLC is necessary for its targeting to clathrin-coated pits (Legendre-Guillemin et al., 2005) and loss of CLCs result in mislocalization of Hip1R and overassembly of actin patches (Poupon et al., 2008), further emphasizing the point that light chains are essential for recruiting actin to sites of endocytosis by interacting with Hip1/R proteins.

In yeast, all clathrin-dependent endocytic events are actin-dependent and therefore, light chain-dependent (Chu et al., 1996). In mammalian cells however, the light chains and actin are not essential for CME to occur.

In what context is an endocytic event light chain-dependent or -independent? The factor that dictates the requirement of light chains is the amount of force that is required for the membrane to invaginate. Membrane tension opposes membrane deformation. Invagination of membranes with high tension require greater force. Boulant et al. (2011) showed that actin was recruited by the light chains to counteract membrane tension in polarized MDCK cells. It is plausible that in instances where clathrin polymerization does not produce enough energy to bend the membrane, the light chains recruit actin, which polymerizes and provides energy for membrane invagination. Membrane tension may differ between cell types. This explains why the uptake of the same receptor may be light chain dependent in one type of cell, and independent in another. Another factor that opposes membrane budding is turgor pressure. Turgor pressure of yeast is higher than that of mammalian cells (Aghamohammadzadeh and Ayscough, 2009), which may explain why all clathrin-dependent endocytic events in yeast are light chain and actin-dependent (Goode et al., 2015). It is important to note however, that in plants, which have a similarly high turgor pressure as yeast, actin is not required for endocytosis (Baisa et al., 2013), allowing speculation that other proteins may be involved in this process.

Recruiting actin to provide energy for membrane invagination is not the only way light chains help in vesicle formation. As mentioned above, light chains are also involved in lattice rearrangement which introduces membrane curvature in flat clathrin lattices as they transform into shallow pits (Maib et al., 2018). However, clathrin does not always assemble as a flat lattice first. This happens only when the constant area model of membrane invagination is followed. According to this model the clathrin coat assembles into a flat lattice of a given area. The lattice is then remodeled by inserting pentagons to introduce curvature without changing the area. Another proposed model is the constant curvature model for membrane invagination according to which clathrin assembles directly into a bud of constant curvature. As a spherical vesicle is formed from a shallow pit, the clathrin coated area increases (Figure 2). Both these models have been shown to exist in vitro (Scott et al., 2018). An increase in membrane tension also increases the number of flat clathrin lattices (Bucher et al., 2018). Based on these studies, Maib et al. (2018) hypothesized that in cases where the polymerization energy of the clathrin triskelia is not sufficient to deform the membrane directly, it will initially assemble as a flat lattice, whereas membranes that are easier to deform might follow the path of constant curvature, directly polymerizing into spherical vesicles.

To summarize, when there is low membrane tension, clathrin may directly polymerize onto the budding surface and light chains will not be required to rearrange the lattice and recruit actin. When there is high membrane tension clathrin may first assemble as a flat lattice which can then be rearranged with the help of light chains and actin to provide energy for membrane invagination.

This ability of the light chains to bind to Hip1/R protein and recruit actin is often exploited by bacteria and viruses to facilitate their entry into cells. Listeria monocytogenes for example, binds to cadherin through internalin, a protein which induces phosphorylation of the heavy chain. This phosphorylation recruits actin through the Hip1/R binding domain of the light chains to surround clathrin at the membrane and facilitate the entry of the pathogen into the cell (Bonazzi et al., 2011).

Five decades of research has provided a huge amount of insight into the complex process of CME (Kaksonen and Roux, 2018; Briant et al., 2020). While studies reveal the role of clathrin light chains in regulating clathrin assembly and several physiological processes, a number of open questions remain unanswered. For example, the specific roles of CLCa and CLCb and their splice variants are not completely understood. While recent studies have shed some light on their specific functions (Wu et al., 2016; Maib et al., 2018; Tsygankova and Keen, 2019; Redlingshöfer et al., 2020), we are only beginning to appreciate the role of each paralog. Apart from this, there is little clarity on why the requirement for light chains is different between different species and cell types. Other questions that need to be answered include the role of light chains in auxilin and GAK-mediated uncoating, mitosis, cell migration and neurodegeneration.

The molecular complexity and dynamic nature of endocytosis make it a difficult process to study. The presence of two paralogs further complicates the problem of elucidating the role of light chains. Small interfering RNA (siRNA)-mediated knockdown of CLCb is often compensated by increased expression of CLCa, whereas knockdown of CLCa is often accompanied by decrease in CHC expression (Majeed et al., 2014). This can lead to inconclusive and confounding results. Use of molecular techniques such as CRISPR-Cas9 genome editing can help overcome these problems by generating single isoform expressing cells. Spatio-temporal deletion of CLCs can further be instrumental in understanding their function in regulating physiological processes.