Endocytosis of the nicotinic acetylcholine receptor (AChR) is a proposed major mechanism of neuromodulation at neuromuscular junctions and in the pathology of synapses in the central nervous system. We show that binding of the competitive antagonist α-bungarotoxin (αBTX) or antibody-mediated cross-linking induces the internalization of cell surface AChR to late endosomes when expressed heterologously in Chinese hamster ovary cells or endogenously in C2C12 myocytes. Internalization occurs via sequestration of AChR–αBTX complexes in narrow, tubular, surface-connected compartments, which are indicated by differential surface accessibility of fluorescently tagged αBTX–AChR complexes to small and large molecules and real-time total internal reflection fluorescence imaging. Internalization occurs in the absence of clathrin, caveolin, or dynamin but requires actin polymerization. αBTX binding triggers c-Src phosphorylation and subsequently activates the Rho guanosine triphosphatase Rac1. Consequently, inhibition of c-Src kinase activity, Rac1 activity, or actin polymerization inhibits internalization via this unusual endocytic mechanism. This pathway may regulate AChR levels at ligand-gated synapses and in pathological conditions such as the autoimmune disease myasthenia gravis.
Communication at synapses requires the location and maintenance of receptors at specific sites. Factors controlling the distribution of receptors are critical determinants of the cell response to external signals. Agonist-induced endocytosis has been shown to operate in various structurally related ion channels, and this process may contribute to synaptic plasticity (Tehrani and Barnes, 1991; Ehlers, 2000; Man et al., 2000; Herring et al., 2003; Nong et al., 2003).
The acetylcholine receptor (AChR) is the best-characterized ligand-gated ion channel (for review see Karlin, 2002). This receptor is found at neuromuscular junctions (NMJs) and at the central nervous system (CNS). The AChR in skeletal muscle is a heterologous pentamer composed of four different but highly homologous subunits in the stoichiometry α2βγδ (embryonic receptor) or α2βεδ (adult receptor; Gotti et al., 2006). The binding of acetylcholine promotes transition of the receptor from a closed to an open state in which it is permeable to cations and subsequent depolarization of the postsynaptic membrane (for review see Karlin, 2002).
Blockage of activity, embryonic development (Drachman et al., 1978; Libby et al., 1980; Bursztajn et al., 1983; Akaaboune et al., 1999; Salpeter, 1999), agonist application (St John and Gordon, 2001), and pathological conditions such as myasthenia gravis (Barrantes, 1998) have been shown to affect AChR targeting and metabolic stability at the plasma membrane. The endocytic mechanism by which AChRs are internalized is not fully understood. At the same time, endocytic modulation of the AChR appears increasingly relevant for the understanding of synaptic plasticity at the CNS and NMJ (Salpeter, 1999; Sanes and Lichtman, 1999).
In this study, we characterize ligand- and antibody-induced internalization of the muscle adult-type AChR (α2βeδ) heterologously expressed in a CHO cell line (Roccamo et al., 1999) and endogenously expressed in the C2C12 muscle cell line. We find that the competitive antagonist α-bungarotoxin (αBTX) and antibody-mediated cross-linking induces down-regulation of cell surface AChR, occurring in two stages. The receptor is first removed from the surface via a surface sequestration mechanism, and then an endocytic process eventually traffics it to the late endosomes. The endocytic pathway of the αBTX–AChR complex differs from many of the well-characterized clathrin or caveolar pathways because internalization of the receptor is not blocked by inhibiting dynamin activity or membrane cholesterol removal (Conner and Schmid, 2003; Borroni et al., 2007; Mayor and Pagano, 2007). The αBTX-labeled receptor sequestration and internalization depends on the integrity of the cytoskeletal network and requires the activity of the Rho GTPase Rac1. This is stimulated by αBTX binding followed by induction of Src phosphorylation and activation.
αBTX binding to cell surface AChR causes receptor down-regulation
CHO-K1/A5 is a clonal cell line that expresses adult (α2βεδ) mouse AChR (Roccamo et al., 1999). Cell surface AChR can be detected using fluorescent derivatives of the competitive antagonist αBTX or with the specific monoclonal antibodies mAb210 or mAb35 (antibodies against an extracellular epitope of the α1 AChR subunit; Feng et al., 1998). To test whether αBTX binding affects AChR internalization, we monitored the levels of AChR on the cell surface before and after incubation with αBTX and upon chasing at 37°C. In the absence of αBTX, levels of surface AChR were similar at 0 and after 6 h of chase (Fig. 1 A, histogram; gray bars); incubation of CHO-K1/A5 cells for 6 h with a saturating concentration of αBTX resulted in a 40% reduction in surface AChR levels (Fig. 1 A). In the absence of αBTX, surface levels of AChR did not change even after treatment with cycloheximide for 6 h (unpublished data). This indicates that constitutive endocytosis and degradation of AChR are very slow processes in CHO-K1/A5 cells, and the contribution of biosynthetic pools to cell surface receptor levels is insignificant over this interval.
To test whether the internalization of AChR is induced by αBTX in a cell line endogenously expressing the receptor, we monitored AChR endocytosis in C2C12 cells. C2C12 is a myoblast cell line of murine origin that, upon differentiation, expresses the same AChR form (St John and Gordon, 2001) as the CHO-K1/A5 clonal cell line. In C2C12 cells treated with αBTX, surface levels of AChR were found to decrease significantly (Fig. 1 B), indicating that both C2C12 and CHO-K1/A5 cells exhibit αBTX-induced down-modulation of AChR.
Antibodies to AChR cross-link the receptors and increase their internalization and degradation rates (antigenic modulation), thus contributing to AChR loss in the autoimmune disease myasthenia gravis (Tzartos et al., 1986). To find out whether the heterologously expressed and endogenously expressed receptors responded to antibody-induced modulation, cells were treated with mAb35, an antibody against the main immunogenic region in the α subunit. Antibody-mediated cross-linking resulted in the reduction of cell surface levels of AChR in both CHO-K1/A5 and C2C12 cells (Fig. 2). These experiments suggest that AChR surface modulation in CHO-K1/A5 cells is similar to that observed in muscle cells and, therefore, validate use of the former cell line as a model system to study triggered AChR down-modulation.
AChR endocytic traffic en route to late endosomes/lysosomes
To assess the trafficking route of AChR upon internalization, we studied the kinetics of colocalization of αBTX-labeled AChR and several endomembrane markers. CHO-K1/A5 cells were labeled with Cy3αBTX for 1 h at 4°C and were chased for periods ranging from 0 to 12 h. Most of the label remained at the cell surface even after 2 h of chase (Fig. S1 A). Subsequently, a fraction of the fluorescently labeled AChR was found in punctate endosomal compartments. In live cells labeled with AlexaFluor488-αBTX, 4 h after labeling, endocytic carriers could be observed moving toward the perinuclear region (Video 1 and Fig. S2 B). The fraction of labeled AChR with a punctate intracellular distribution increased as a function of time (Fig. 3 A, 6 h; and Fig. S1 A). A significant number of such punctate compartments colocalized with lysosomal-associated membrane protein 1 (LAMP1), a late endosomal marker; colocalization with this marker increased upon further chase up to 12 h (unpublished data). Comparison of the time course of early endosomal antigen 1 (EEA1) and LAMP1 colocalization suggests that αBTX–AChR complexes are trafficked from EEA1-positive endosomes to late endosomes (compare colocalization of internalized AChR with EEA1 and LAMP1 in Fig. 3, A and B, respectively). Treatment of cells with 100 μM leupeptin (8 h), a lysosomal proteinase inhibitor, resulted in a higher amount of αBTX in late endosomes (Fig. S3), suggesting that the targeting of AChR to LAMP1-positive compartments leads to degradation of the receptor. In C2C12 cells, the intracellular distribution of AChR at various times after αBTX-induced internalization was similar to that observed in CHO-K1/A5 cells (Fig. 3 C), suggesting that the intracellular destination and trafficking of AChR in CHO-K1/A5 cells is not an artifact of exogenous expression. A significant fraction of αBTX-labeled AChR-positive punctate endosomes formed after 12 h in CHO-K1/A5 cells were accessible to a 20-min pulse but not to a 5-min pulse of fluorescein-labeled dextran, a marker of fluid-phase endocytosis (Fig. S4, top). This is consistent with their designation as fusion-competent late endocytic compartments.
To elucidate the nature of the punctate endosomal compartments accessed by AChR, we examined colocalization with markers of different endocytic routes. The small punctate endosomes that formed initially did not colocalize with endocytosed transferrin (Tf) bound to its receptor (Tf receptor [TfR]; Fig. S4, bottom), a marker of the clathrin-mediated endocytic sorting and recycling pathway, or with GFP–glycosyl-phosphatidylinositol (GPI; Fig. 3 D, left), a marker of the Cdc42-regulated GPI anchored protein–enriched early endosomal compartment pathway in CHO cells (Sabharanjak et al., 2002), or with peripheral caveolin-positive structures (Fig. 3 D, right), a marker of caveolar endocytosis (Parton and Richards, 2003). However, a fraction of the 2- and 4-h αBTX-labeled AChR-positive endosomes did colocalize with EEA1 (Fig. 3, A and B).
These results indicate that αBTX binding induces a slow internalization of AChR via an endocytic pathway that results in transient accumulation of the receptor in EEA1-positive endocytic compartments distinct from those derived from the clathrin-mediated caveolar pathways or the clathrin- and caveolin-independent Cdc42-regulated GPI anchored protein–enriched early endosomal compartment pathway.
Sequestration of AChR into a surface-connected compartment
To characterize the very initial steps in the AChR down-regulation process, we examined the time course of loss of surface AChR labeled with biotin-Cy3αBTX (b-Cy3αBTX). To measure the accessibility of surface AChR to small molecules, we used the membrane-impermeable quencher potassium iodide (KI). To estimate the quantitative change in quenching efficiency as a function of incubation time, the quenching efficiency, Q, at any time point was normalized to the maximum quenching efficiency attainable after labeling the cells on ice (see Materials and methods section Accessibility of αBTX-labeled receptors to small molecules). To measure the accessibility of large molecules to b-Cy3αBTX–labeled AChR, we measured surface binding of ∼50-kD Cy5-streptavidin (SA). Measurement of the Cy5/Cy3 fluorescence ratio yielded a quantitative estimate of the accessibility of Cy5-SA to b-Cy3αBTX per cell. We carried out similar accessibility experiments using an independent label on AChR, FITC-αBTX. In this case, KI was used as a low molecular weight probe and a quenching-competent antifluorescein antibody (∼150 kD; Blakeslee, 1976) as a high molecular weight probe. Both approaches yielded identical results (unpublished data); therefore, we only report experiments carried out with b-Cy3αBTX.
After incubation on ice, all labeled AChRs were accessible to both small and large molecular weight probes (Fig. 4 A); the quenching efficiency of the KI- and Cy5-SA–binding capacity (Cy3/Cy5 fluorescence ratio) of the probes was similar to that obtained for free-labeled αBTX (Fig. 4 C, 0 time point). After 2 h, <50% of b-Cy3αBTX–labeled AChRs were accessible to Cy5-SA, whereas almost all (∼90%) of the receptors remained accessible to the small molecular weight quencher (Fig. 4, B and C). Incubation for an additional 4 h resulted in a significant loss of accessibility of the labeled AChRs to both small and large molecular weight quenchers: ∼60% of the receptors became inaccessible to KI and Cy5-SA after 6-h chase (Fig. 4 C), and 70% of the labeled AChRs became KI inaccessible after 12 h. However, by 12 h, there appeared to be some loss in the total Cy3αBTX receptor staining (Fig. S1 B), rendering the surface accessibility assay insufficiently accurate for quantifying the intracellular pool.
Inspection of the fluorescence distribution of b-Cy3αBTX in cells labeled with b-Cy3αBTX on ice followed by incubation for 0 or 2 h at 37°C showed that almost all of the labeled probe was retained at the cell surface (Fig. 4, A and B; Cy3 images) and could be quenched with KI (Fig. 4, A and B, KI quenching; and C). Surprisingly, compared with the 0-h time point, after 2-h incubation at 37°C, a significant fraction of the b-Cy3αBTX probe (pseudocolored red in Fig. 4) quenched by KI could not be labeled by Cy5-SA (pseudocolored green in Fig. 4, A and B; and see merged images). Cy3-labeled regions devoid of the Cy5-SA signal could be resolved as small punctate spots (Fig. 4 B, arrows) often close to the optical resolution of the microscope setup used (∼300 nm).
The differential surface accessibility of the αBTX-labeled receptor to small and large molecular weight probes shows that the labeled receptor evolves from a fully surface-accessible state to a surface-connected but sequestered state upon incubation at 37°C. The rate of sequestration of labeled AChR may be derived from the rate of change of the Cy5-SA/b-Cy3αBTX ratio. This is described by a single exponential decay with a time constant, t1/2, of 1.98 ± 0.36 h. This result differs from those of the KI-quenching assay (Fig. 4 C), which is in agreement with the hypothesis of a sequestration process before loss of surface accessibility.
In an attempt to visualize these surface-connected but inaccessible structures, we utilized total internal reflection fluorescence (TIRF) microscopy, which reports on events in a 100–200-nm swathe of cytoplasm in proximity to the plasma membrane at the coverslip glass–water interface, where the evanescent field is generated (Axelrod, 1981). Alternating TIRF and wide-field imaging were carried out (see Materials and methods section Live cell and TIRF microscopy) to distinguish the structures being internalized beyond the TIRF field, leading to the disappearance of αBTX-labeled AChR from the cell surface. When observed live under TIRF microscopy, cells labeled with AlexaFluor488-αBTX after 2 h of chase exhibited punctate as well as tubular structures. The absence of wide-field correlates of most of the TIRF structures can be accounted for by the low contrast in wide-field illumination, especially near the cell surface (Fig. 4 D). Wide-field images showed tubular αBTX-labeled AChR-containing structures appearing as a collection of tubules and puncta in the TIRF images. These remained in the field of view for several frames but eventually disappeared from the TIRF field, although they were still detectable in the corresponding wide-field images. This behavior is consistent with the spatial movement of the tubular fragments out of the TIRF field as a result of their detachment from the surface (Video 2). Some structures were visible as spots, whereas the corresponding wide-field structures were tubular (Fig. S2 A and Video 3). Collectively, these observations suggest that αBTX–AChR complexes are endocytosed into tubular structures that reside in the proximity of the plasma membrane before detaching from the cell surface and being drawn inside the cell. Also, the occurrence of narrow tubular structures provides a likely explanation for the sequestration of AChR from high molecular weight quenchers but not KI.
Triggered AChR internalization is not blocked by inhibitors of dynamin function
Some ionotropic receptors have been shown to be internalized via a dynamin-dependent endocytic mechanism (Ehlers, 2000; Man et al., 2000; Herring et al., 2003; Nong et al., 2003). To investigate the mechanism operating in the case of the AChR, cells were transiently transfected with dynamin K44A–GFP, a GTPase-defective dominant-negative mutant of dynamin (Damke et al., 1994). We assessed endocytosis of αBTX-labeled AChR by monitoring the extent of loss of KI quenching of Cy3αBTX-labeled AChR at 6 h of uptake. The data showed that overexpression of the dominant-negative dynamin mutant did not affect internalization (Fig. 5, compare transfected cells in A and histogram in D) of the AChR, whereas it inhibited TfR uptake (Fig. 5 B). Independently, the sequestration of b-Cy3αBTX–labeled AChR was also unaffected by the expression of dynamin K44A (Fig. 5 C, compare Cy5-SA binding in transfected cells vs. untransfected cells; and Fig. 5 D, compare surface accessibility). Furthermore, overexpression of a deletion mutant of Eps15, EH21, a protein involved in endocytosis of a variety of receptors via clathrin-dependent (Benmerah et al., 1999) and -independent pathways (Mayor and Pagano, 2007), consistently blocked Tf endocytosis but had no detectable effect on AChR internalization induced by αBTX (Fig. 5 E) or mAb 35 (not depicted).
To confirm that dynamin is not required for triggered AChR endocytosis, we depleted cells of dynamin 2 using shRNA constructs as reported previously (Gomez et al., 2005). Expression of shRNAs resulted in a significant reduction in the rate of Tf endocytosis and in dynamin levels (Fig. 5, H–J). A CHO-K1–derived cell line that stably expresses human TfR, IA2.2, also showed similar defects in Tf endocytosis upon dynamin depletion (unpublished data). However, depletion of dynamin had no effect on AChR down-modulation by αBTX and mAb (Fig. 5, F and G) or by antibody alone (not depicted). These results confirm that AChR endocytosis is independent of the amount and activity of the dynamin 2 protein and that dynamin- and clathrin-dependent internalization machineries are not involved in AChR endocytosis.
AChR internalization depends on actin polymerization and activity of the Rho GTPase Rac1
Clathrin- and dynamin-independent endocytic processes appear to recruit the actin polymerization machinery to aid in endocytosis (Mayor and Pagano, 2007). Therefore, we investigated the role of actin polymerization in AChR endocytosis. We treated cells with cytochalasin D; cytochalasins inhibit actin polymerization by binding to the barbed end of the actin filament and by blocking monomer addition. Actin polymerization inhibitors have variable effects on clathrin-mediated endocytosis (Lamaze et al., 1997; Fujimoto et al., 2000). 2.5 μM cytochalasin D treatment for 2 h did not affect Tf uptake (Fig. 6, A and C; compare Tf uptake in treated vs. untreated cells), whereas it significantly decreased AChR sequestration (Fig. 6 B, compare Cy5-SA accessibility in treated vs. untreated cells). These results indicate that a dynamic rearrangement of the actin cytoskeleton is required for sequestration of the AChR. In addition, whereas untreated cells showed punctate endosomes 2 h after internalization, cytochalasin D–treated cells appeared devoid of internalized AChR puncta, showing that internalization of AChR is blocked by actin depolymerization.
The Rho family of GTPases regulates actin dynamics and is known to affect endocytosis of a variety of proteins and membrane lipids (Mayor and Pagano, 2007). The sensitivity of AChR internalization to actin polymerization indicated the possible involvement of a Rho family GTPase in AChR endocytosis. To address this possibility, cells were treated with Clostridium difficile toxin B, which specifically and irreversibly glucosylates members of the Rho family of GTPases (Sehr et al., 1998). Toxin B treatment resulted in inhibition of αBTX-induced AChR uptake (Fig. 6 D). To specifically investigate the identity of the Rho GTPase involved, cells were independently transfected with GTP exchange–deficient forms of Rho GTPases regulating distinct pathways of endocytosis (i.e., Rac1, Cdc42, and RhoA). Whereas overexpression of dominant-negative RhoA and Cdc42 had no effect on AChR uptake, dominant-negative Rac1 blocked AChR uptake to a significant degree (Fig. 6, D and E).
To assess whether αBTX binding to AChR induces Rac1 activation, we exploited the observation that upon activation, GTP-bound Rac1 binds tightly to p21-associated kinase (Pak; Bagrodia et al., 1995). αBTX labeling resulted in significant Rac1 activation, which was observed as an increase in the amount of Rac1 that coimmunoprecipitated with Pak (Fig. 6 F). These observations suggest that αBTX binding to AChR triggers Rac1 activation, which, in turn, regulates AChR endocytosis, possibly through modulation of the actin cytoskeleton.
Src phosphorylation and AChR endocytosis
Src activation has previously been shown to modify AChR anchoring and phosphorylation (Mohamed and Swope, 1999; Sadasivam et al., 2005). In CHO-K1/A5 cells, binding of αBTX to AChR leads to an increase in Src phosphorylation (Fig. 7 C). The phosphorylation of Src is sensitive to treatment with the Src kinase–specific inhibitor PP2 (Hanke et al., 1996), indicating this to be an autophosphorylation event. Inhibition of Src phosphorylation by PP2 also inhibited the activation of Rac1 (Fig. 6 F), an event critical for AChR internalization. Therefore, we tested the role of Src activation in αBTX-induced AChR endocytosis using pharmacological inhibition of Src kinase by PP2 as well as overexpression of a Src K297R kinase-dead mutant form. Treatment of cells with PP2 or Src K297R overexpression inhibited the entry of AChR into the cell (Fig. 7, A and B). These results confirm that αBTX-induced AChR internalization involves Src autophosphorylation and the downstream activation of Rac1.
Diffusion of AChRs to perijunctional areas and internalization has been shown to occur when neuromuscular transmission is blocked using αBTX, resulting in a shorter AChR half-life. Direct muscle stimulation inhibits this process (Akaaboune et al., 1999). Interestingly, these alterations in AChR density at the NMJ are equivalent to long-term potentiation and long-term depression in the CNS (Akaaboune et al., 1999). The rate of AChR endocytosis has also been found to increase in myotubes incubated in the continuous presence of cholinergic agonists (St John and Gordon, 2001) and by cross-linking with antibodies in human culture muscle cells (Tzartos et al., 1986) and in the BC3H-1 muscle cell line (Clementi and Sher, 1985). However, the endocytic mechanisms that lead to alterations in AChR levels at the cell surface have not been characterized.
In this study, we show that binding of the quasi-irreversible antagonist αBTX induces the internalization of the AChR via a relatively slow process. Internalization appears to proceed in two stages. First, the AChR seems to be sequestered in a surface-connected compartment that is not accessible to large molecular weight molecules such as SA or anti-AChR mAb. This process requires actin polymerization and Rac1 GTPase activation because treatment with cytochalasin D and toxin B and overexpression of a GDP exchange–deficient form of Rac1 result in a reduction of αBTX-triggered AChR sequestration. Second, αBTX-bound AChR is delivered to the lysosomal compartment via a distinct endocytic mechanism, as discussed below in this section. The relationship between these two stages is not clear, although it is likely that subsequent to sequestration, αBTX-liganded AChR is driven into the late compartments from surface-connected invaginations. The ability of small molecule quenchers (but not large molecular weight reagents) to access the nascent internalized receptors and their presence in surface-localized tubules visible in a TIRF field provide support for the proposed internalization model. However, these are indirect methods, and the ultrastructure of these surface-accessible structures remains uncharacterized. The sequestration in a surface-connected compartment (if reversible) may provide an alternative, faster mechanism for regulating the availability of AChRs to modulate signals accessible from the intersynaptic gap.
The mechanism behind surface sequestration resembles the sequestration mechanism of β-VLDL and other aggregated lipoproteins internalized via scavenger receptors into STEMs (surface tubule for entry into macrophages) in macrophages (Myers et al., 1993; Kruth et al., 1995; Zhang et al., 1997). Involvement of the actin cytoskeleton in both of these processes also suggests a similarity in mechanism; recruitment of the actin-modulating machinery to the site of sequestration appears necessary for both pathways. Surface sequestration and internalization are both insensitive to inhibition of dynamin function (via expression of a dominant-negative dynamin isoform or by expressing shRNA constructs), suggesting that this pathway is independent of the most extensively addressed vesicle-pinching mechanism (Conner and Schmid, 2003). The lack of a requirement for dynamin and Eps15, the absence of colocalization of AChR peripheral endosomes with caveolin, and the lack of colocalization with a well-characterized clathrin pathway marker, endocytosed Tf, indicate that this pathway is neither clathrin nor caveolin dependent. In addition, neither acute cholesterol depletion (Borroni et al., 2007) nor Cdc42 N17 expression inhibits internalization of αBTX-bound AChR. Endocytosed AChR also does not colocalize with GFP-GPI, further supporting the idea that this endocytic pathway represents an internalization mechanism distinct from the medley of clathrin-independent pathways available for membrane proteins at the surface of eukaryotic cells (Mayor and Pagano, 2007).
Once internalized, αBTX is trafficked to late endosomes via EEA1-positive endosomes, suggesting that the αBTX-bound receptor eventually accumulates in degradative compartments after fusing with early endosomes. However, at no stage do the endocytosed αBTX–AChR complexes colocalize with endocytosed Tf or a short pulse of fluid-phase marker. These results suggest that the trafficking of internalized αBTX–AChR complexes also represents an uncharacterized route to bonafide LAMP1-positive late endosomes.
At first glance, the data presented here appear to be in disagreement with a previous study (Trinidad and Cohen, 2004) showing the colocalization of internalized AChR and caveolin 3 in myotubes. This apparent discrepancy can be explained by the cell type and the scale of AChR organization during development: although in CHO cells, the AChR can be observed as submicrometer-sized puncta (Borroni et al., 2007) that we have interpreted as representative of preinnervation aggregates of AChRs and recently resolved as nanometer-sized AChR nanoclusters (Kellner et al., 2007), the observations of Trinidad and Cohen (2004) were carried out in C2C12 myotubes at a stage of differentiation in which AChRs are organized into large, several micrometer–sized clusters. In fact, the latter authors defined these AChR clusters as much larger structures (i.e., longer than 4 μm; Trinidad and Cohen, 2004). Thus, endocytosis of AChR in more mature clusters may proceed through caveolin-mediated mechanisms, although the colocalization with caveolin 3 has more recently been challenged (Stetzkowski-Marden et al., 2006).
Here, we show that binding of αBTX to AChR leads to the activation of Rac1, which is required for AChR internalization possibly via the modulation of actin dynamics, which is critical for AChR surface sequestration. In parallel, we observe that the concomitant autophosphorylation of Src is also required for AChR endocytosis. It will be interesting to investigate whether activated Src phosphorylates molecular regulators of AChR internalization like Rac1 or AChR, or both. A pivotal role of Src and Rac1 activation has also been recently observed in HIV-Nef–mediated down-modulation of immune costimulatory molecules CD80/CD86 (Chaudhry et al., 2005, 2007). Both AChR and CD80/CD86 are long-term cell surface–resident proteins in the cell systems studied. Thus, it is likely that this endocytic mechanism may be of general relevance in remodeling the molecular composition of the plasma membrane. The precise molecular events linking Src phosphorylation and Rac1 activation and their temporal regulation need to be explored in the context of AChR endocytosis.
Bruneau et al. (Bruneau et al., 2005; Bruneau and Akaaboune, 2006) have recently described the dynamics of recycled αBTX-labeled AChR at the living NMJ. Although the major fraction of AChRs appears to be targeted to late endosomes for degradation, in the two independent systems used here, we find that a small fraction of internalized αBTX–AChR colocalizes with Rab11, a recycling endosomal marker, at all time points of chase (unpublished data). Furthermore, the intrinsic mechanisms underlying the dynamics of AChR internalization described by Bruneau et al. (Bruneau et al., 2005; Bruneau and Akaaboune, 2006) share some similarities with the ones disclosed in this paper. First, these authors observed internalized αBTX–AChR in very stable EEA1-positive vesicles at the NMJ, the same as we do in CHO-K1/A5 and C2C12 cells. They also show that αBTX–AChR recycling to the cell surface is sensitive to Tyr phosphorylation/dephosphorylation; in this study, AChR endocytosis is shown to depend on the well-known Tyr kinase Src, which also regulates the stability of AChR clusters in vivo (Sadasivam et al., 2005).
AChR removal from the cell surface has important pathophysiological consequences in the autoimmune disease myasthenia gravis, in which receptor internalization is accelerated upon binding of anti-AChR antibodies (Barrantes, 1998). Thus, IgG from myasthenic patients causes antigenic modulation of muscle AChR in vivo and in vitro (Drachman et al., 1978). Our experiments show that monoclonal antibodies against the muscle-type AChR also induce its internalization. It is likely that the mechanism described here is similar to the one operating in the myasthenic muscle.
The αBTX-induced AChR endocytosis reported here appears to differ from the clathrin-dependent endocytosis of γ-aminobutyric acid, AMPA, and N-methyl-d-aspartate receptors (Man et al., 2000; Herring et al., 2003; Nong et al., 2003). It is important to note that all of these receptors are located at CNS synapses, whereas here, we deal with a peripheral muscle-type AChR. The rate of the endocytic mechanism operating in peripheral and central synapses also appears to differ, and this is mirrored by AChR in CHO cells and in the myocyte cell line C2C12. We calculated a t1/2 of 1.98 ± 0.36 h for endocytosis of the muscle-type AChR heterologously expressed in CHO-K1/A5 cells. The AMPA receptor in cultured neurons, on the other hand, exhibits a much faster process: t1/2 = 9.3 ± 2.3 min (Ehlers, 2000). It is possible that neuromuscular synapses regulate the availability of AChR and, consequently, synaptic strength via mechanisms different from those operating in CNS synapses.
Lynx1 is a GPI-anchored membrane-resident protein that possesses a structure notably similar to that of αBTX (Ibanez-Tallon et al., 2002). This neurotoxin-related protein was shown to directly interact with a neuronal subtype of AChR and to modulate its function (Ibanez-Tallon et al., 2002). A process similar to the slow down-regulation exerted by αBTX on muscle-type AChR may thus be operative in the CNS by the interaction between Lynx and neuronal AChR and may constitute a modulatory mechanism contributing to synaptic plasticity.
Materials And Methods
All chemicals were obtained from Sigma-Aldrich unless otherwise specified. Cy3 and Cy5 fluorochromes were purchased from GE Healthcare. AlexaFluor488, FITC, and AlexaFluor594-αBTX were purchased from Invitrogen. mAb 210 antibody against the AChR α subunit was a gift from J. Lindstrom (University of Pennsylvania Medical Center, Philadelphia, PA), and mAb 35 and anti-LAMP were purchased from Developmental Studies Hybridoma Bank. Anti-EEA1, Rac1, phospho-Src antibodies, and SA-phycoerythrin (PE) were purchased from BD Biosciences. PAK–Pak-binding domain (PBD)–coupled agarose (Chemicon) was used to pull down activated Rac. Antibodies against c-Src and pTyr418-Src were obtained from Cell Signaling Technology and Millipore, anti-Rac was obtained from BD Biosciences, and anti–lysobisphosphatidic acid antibody was a gift from T. Kobayashi (University of Geneva, Geneva, Switzerland). Secondary antibodies were obtained from Jackson ImmunoResearch Laboratories. Constructs encoding shRNA against dynamin were a gift from D. Billadeau (Mayo Clinic College of Medicine, Rochester, MN; Gomez et al., 2005), and a dominant-negative inhibitor of eps15, EH21-GFP, was provided by A. Benmerah (Institut National de la Santé et de la Recherche Médicale E9925, Faculté Necker-Enfants Malades, Paris, France; Benmerah et al., 1999).
Cell culture and transfections
CHO-K1/A5, a cell line derived from CHO-K1 cells (Roccamo et al., 1999), was grown in Ham's F-12 medium containing10% FBS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 0.5 mg/ml G-418. In some experiments, CHO-K1/A5 cells were transiently transfected with GFP-labeled caveolin, GFP-GPI, or GFP–dynamin K44A using Lipofectamine (Invitrogen) or Fugene (Roche) according to the protocol given by the manufacturer and were used for experiments 12 h after transfection exactly as described previously (Sabharanjak et al., 2002). The C2C12 murine myoblast cell line (provided by J. Dhawan, Centre for Cellular and Molecular Biology, Hyderabad, India) was grown in DME supplemented with 20% FBS at 37°C and 5% CO2. For differentiation, C2C12 cells were plated in phenol red–free DME (HiMEDIA Laboratories) containing 20% FBS and grown until 90% confluency. Cells were then shifted to medium containing 0.5% FBS for 24 h and used for endocytic assays.
Labeling of αBTX with fluorescent dyes
Native αBTX was dissolved in PBS (150 mM NaCl, 10 mM Na2HPO4, and 10 mM NaH2PO4, pH 7.5) to a final concentration of 125 μM. Conjugation was allowed to proceed for 1 h at room temperature with 10-fold molar excess dye. Labeled αBTX was subsequently separated from free dye by gel filtration on a Sephadex G-25 column. For biotin conjugation, an aliquot of Cy3αBTX was further incubated with biotin N-hydroxysuccinimide ester at a molar ratio of 1:5 for 30 min at room temperature. Excess biotin was separated by overnight dialysis. The dye/protein ratio was estimated, and conjugates with labeling efficiency less than one were used in experiments.
Accessibility of αBTX-labeled receptors to small molecules (KI quenching)
To measure the rate of sequestration and internalization of AChRs, CHO-K1/A5 cells were incubated with Cy3αBTX or FITCαBTX for 1 h on ice, rinsed with chilled M1, and shifted to 37°C for varying periods from 0 to 12 h. Cells were subsequently imaged both before and after adding the membrane-impermeable quencher KI at a concentration of 400 mM for b-Cy3αBTX and at 100 mM for FITCαBTX. The quenching efficiency was calculated with the equation Q(%) = 100 × (1 − FKI/F), where FKI and F are the fluorescence intensities of labeled αBTX cells immediately after adding the quencher or just before, respectively. The extent of quenching of the labeled receptors was normalized to the maximal quenching efficiency obtained after quenching receptors bound on ice. Maximal quenching efficiencies of the free probes were ∼80% for Cy3-labeled toxin and ∼60% for the FITC-labeled toxin at these concentrations of quencher, which is similar to those obtained on cell surface–bound receptors.
Surface accessibility to Cy5-SA of AChRs labeled with b-Cy3αBTX
To determine the extent of sequestration of AChRs, surface accessibility of labeled receptors was measured as described previously (Guha et al., 2003). In brief, CHO-K1/A5 cells were incubated with b-Cy3αBTX for 1 h on ice, rinsed with chilled M1, and shifted to 37°C for different periods (0–12 h). After the indicated time, cells were labeled with Cy5-SA for 1 h at 4°C, rinsed with chilled M1, and imaged by quantitative digital microscopy to measure fluorescence intensity from Cy3-labeled receptors and corresponding Cy5 signals from the same cells. The Cy5/Cy3 fluorescence ratio at different times provided an estimate of the accessibility of the biotin tag on the Cy3αBTX-labeled receptor to Cy5-labeled SA; the Cy5/Cy3 ratio of receptors labeled on ice was considered fully (100%) accessible.
Endocytosis and colocalization experiments
Unless otherwise specified, cells were incubated with saturating concentrations of AlexaFluor488-αBTX, Cy3αBTX, AlexaFluor594-αBTX, or other AChR probes for 1 h on ice and rinsed with chilled M1 medium (140 mM NaCl, 20 mM Hepes, 1 mM CaCl2, 1 mM MgCl2, and 5 mM KCl, pH 7.4). For different probes, saturation was achieved at ∼0.1–1 μM. For labeling and pulsing of AChR in the myocytes, differentiated C2C12 cells were incubated with either anti-AChR mAb or AlexaFluor488-αBTX for the indicated times, washed with M1 medium containing 0.1% BSA, 0.1% glucose, and 0.5% FBS, and imaged live in the same medium. Whenever required, protein biosynthesis was arrested by incubation of cells with 200 μg/ml cycloheximide for 6 h before labeling with fluorescent αBTX.
For endocytic trafficking of AChR, CHO-K1/A5 cells were incubated with Cy3αBTX for 1 h on ice, rinsed with chilled M1, and shifted to 37°C in growth medium for different periods (0–12 h). To assess colocalization with Cy5-Tf or FITC-dextran, these probes were added to the medium at different time points. After incubation, cells were washed for 10 min with chilled ascorbate buffer (140 mM sodium ascorbate, 65 mM ascorbic acid, 1 mM CaCl2, and 1 mM MgCl2, pH 4.5) to remove surface-bound Tf and were imaged. To test possible colocalization with the early endosomal marker EEA1 and with the late endosomal marker LAMP1, cells were labeled with Cy3αBTX and incubated at 37°C for 2, 4, 6, or 12 h. Cells were subsequently fixed, permeabilized, and immunostained with anti-EEA1 or anti-LAMP1 antibodies. For caveolin and GPI anchored protein colocalization studies, cells were transiently transfected with plasmids coding for caveolin-GFP or GPI-GFP as described previously (Sabharanjak et al., 2002), labeled with Cy3αBTX on ice for 1 h, chased at 37°C for 2 h, and imaged live. For quantitative assessment of colocalization, CHO-K1/A5 cells were labeled with Cy3αBTX on ice for 1 h, washed, chased at 37°C for 0, 2, 4, or 6 h, fixed, and processed for immunofluorescence for EEA1 or LAMP1. Cells were imaged using a spinning disc confocal microscope (CSU22; Yokagawa) mounted on a microscope (IX81S1F; Olympus) coupled to an electron multiplying CCD camera (Ixon; Andor Technology), and a single confocal slice from each plane was used for quantification. Each slice was processed for background correction using median filter in MetaMorph and spotting using Multispots to extract out spots or endosomes from whole cell images as described previously (Kalia et al., 2006). Spotted images of αBTX were compared with spotted images of EEA1 or LAMP1 using Multicolloc, a custom software designed to determine the extent of the colocalization of intensities of two probes via the generation of a colloc image. Regions were marked in the original images of cells so as to select internal fluorescence only, and these regions were extrapolated on spotted images to find total internalized intensities and extrapolated on the colloc images to find total colocalized intensities. These intensities were presented as a ratio (colloc image intensity/spotted image intensity) per cell.
For immunolocalization of organellar markers in C2C12 cells, cells were labeled with BTX, chased for the indicated time points, and fixed using 4% PFA for 20 min at room temperature. Subsequently, cells were permeabilized with 0.1% saponin for 20 min, washed, incubated with 2% goat serum for 1 h, washed, incubated with primary antibodies at 4°C overnight, washed, incubated with secondary antibody for 1 h, washed, and imaged. For washing C2C12 cells until fixation, serum-free growth medium at 37°C was used for washing instead of M1.
Cells were transiently transfected with dynamin K44A–GFP, a dominant-negative mutant of dynamin 2 (Altschuler et al., 1998). 12 h after transfection, AChR accessibility was determined as described in the Endocytosis and colocalization experiments section. As a positive control of dynamin inhibition, endocytosis of TfR was measured in CHO-K1/A5 cells by adding Cy5-Tf for 20 min. For dynamin 2 RNA interference studies, cells were transfected with the shRNA plasmids for 68–72 h. Transfected cells were identified by monitoring GFP fluorescence. The extent of Tf endocytosis in CHO-K1/A5 cells that express endogenous hamster TfRs was quantified by first determining the mean endocytic uptake of AlexaFluor546-Tf in GFP-expressing cells transfected with either vector control or dynamin-shRNA–containing vector by determining the amount of fluorescent Tf remaining in endocytic compartments after a pulse of 5 min, after removal of surface-bound Tf. These values were subsequently normalized to the mean level of surface TfRs by determining the extent that AlexaFluor546-Tf bound to the cells on ice in cells that were independently transfected with the corresponding constructs.
Surface accessibility of AChR was determined by incubating cells with mAb 35 in the presence or absence of αBTX at 4°C. Cells were then washed and either incubated with secondary antibody immediately or after a chase at 37°C for 6 h. Subsequently, cells were fixed and imaged.
To assess levels of dynamin 2 using Western blots, cells grown in 60-mm dishes were transfected with vector alone (control), shRNA m, or shRNA h for 72 h, harvested by scraping, washed, lysed, and analyzed by detecting proteins using nitrocellulose membranes by Western blotting. Lysate from ∼200,000 cells per lane were loaded on an 8% gel, and the approximate migration positions of dynamin 2 and actin were marked against prestained molecular weight markers after Ponceau staining. Actin- and dynamin-containing strips of membrane were processed and developed independently. For quantitation of protein levels, the developed films were scanned, and the dynamin 2 level was normalized to the respective actin level in the same lane.
Cytochalasin D treatment
Cells were labeled with b-Cy3αBTX for 1 h on ice and shifted to 37°C for 0, 2, and 4 h in complete growth medium containing 2.5 μM cytochalasin D or an equivalent amount of the solvent DMSO. 20 min before the incubation finished, AlexaFluor488-Tf was added to ascertain the specificity of cytochalasin D treatment. After incubation for specified periods, cells were washed for 10 min with ascorbate buffer, labeled with Cy5-SA for 1 h at 4°C, and imaged. The Cy5-SA/Cy3-biotin-αBTX ratio was calculated and normalized to 0 h, and the extent of TfR endocytosis was assessed by fluorescence quantification of acid wash–resistant intracellular AlexaFluor488-Tf.
Rho GTPase inhibition and toxin B treatment
CHO-K1/A5 cells were transfected with myc-RhoA N19 or myc-Rac1 N17, labeled with AlexaFluor594-αBTX on ice for 1 h, washed, and chased in complete medium at 37°C for 6 h. For toxin B treatment, the entire procedure was repeated in the presence of 1 μg/ml toxin B with a chase in complete medium for 1 h followed for a further 5 h in the absence of toxin B. Cells were then washed with chilled M1, labeled with mAb 210 followed with AlexaFluor647 secondary antibody on ice to detect the surface pool of AChR, fixed, and imaged.
Activation of c-Src and Rac1
Cells treated with αBTX for 2 h were lysed in radioimmunoprecipitation buffer (150 mM NaCl, 50 mM Tris-HCl, pH 7.5, 500 μM EDTA, 100 μM EGTA, 1.0% Triton X-100, and 1% sodium deoxycholate). For assessing Rac1 activation, lysates were precleared and affinity precipitated using PAK-PBD (Chemicon). The precipitate was resolved on a 10% SDS-PAGE and detected using anti-Rac1 antibody. To assess activation of c-Src, lysates of αBTX-treated CHO-K1/A5 cells were Western blotted with anti-Src (total c-Src) or antiphospho-–Tyr418-Src antibodies (activated c-Src). Protein content in the lysate was calculated using the microBCA kit (Thermo Fisher Scientific), and equal protein quantities were loaded.
Inhibition of c-Src activity
αBTX-treated CHO-K1/A5 cells were pretreated with 10 μM of the c-Src inhibitor PP2 for 1 h before assessing the effect of this treatment on Rac1 activation or inhibition of c-Src phosphorylation. Cells were labeled with biotin-αBTX on ice, washed extensively, and then shifted to 37°C followed by chase for 4 h. At the end of the chase, surface αBTX was removed using extreme buffer. The internalized label was detected by permeabilizing cells, staining them with SA-PE, and analysis on a flow cytometer (LSR; Becton Dickinson).
src gene mutants were made by cloning PCR-amplified versions of appropriate cDNAs into myc-p–containing pCDNA3.1 vector. Wild-type and kinase-dead (K297R) isoforms of human Src obtained from Millipore were cloned into a mammalian expression vector, pCDNA3.1(B-)MycHis+, using available restriction sites. The kinase-dead c-Src (Src K297R) was transfected into cells along with GFP expression vector in a ratio of 2:1 (wt/wt) with the maximum achievable efficiency of cotransfection.
Fluorescence microscopy and image analysis
Live cell and TIRF microscopy.
For live cell TIRF and wide-field imaging, an inverted system (TE 2000; Nikon) equipped with a 100× NA 1.45 oil objective (Nikon) coupled to a bank of lasers and a CCD camera (Cascade; Photometrics) was used. For dual-color live imaging, the growth medium was substituted with M1 containing 0.1% BSA and glucose, and a temperature-controlled stage maintained at 37°C was used. For TIRF imaging, an Argon ion 488-nm laser coupled to the aforementioned system was used at 10-mW output power.
Quantitative microscopy and image analysis.
For most of the quantitative fluorescence microscopy assays, imaging of cells was carried out on an inverted wide-field fluorescence microscope (TE 300; Nikon) using 20× NA 0.75 and 60× NA 1.4 objectives (Nikon) equipped with a sensitive digital imaging camera (MicroMax; Princeton Instruments) equipped with a sensor (TEK512 B; Roper Scientific). Images were collected using MetaMorph acquisition control software (MDS Analytical Technologies). Confocal microscopy was carried out on a spinning disc confocal imaging system (IX81; Olympus) equipped with a scanning device (CSU22; Yokogawa) and Ixon camera. For imaging, a 60× NA 1.42 oil objective (Olympus), Ixon camera, appropriate filter set dichroics, and lasers were used. Image acquisition was carried out using Revolution software (Andor Technologies) and further analysed using MetaMorph. Images were processed, and fluorescence was quantified by methods described in the figure legends and in greater detail in previous studies (Sabharanjak et al., 2002; Guha et al., 2003).
After staining with primary reagents (e.g., biotinylated αBTX), cells were stained with secondary reagents (e.g., SA-PE) on ice for 30 min or as appropriate. Stained cells were analyzed on a flow cytometer (LSR; BD Biosciences), and data from ∼100,000 cells were routinely acquired for each sample. Data were analyzed using FlowJo software (Tree Star, Inc.). All data shown are representative of at least three independent experiments.
Unless otherwise indicated, all of the values are weighted means, μx, obtained from at least two different dishes in each experiment calculated by the formula
where i is the number of coverslip dishes used in an experiment,
unless another indicator of the spread in the data such as SD was specified. Statistical significance was measured using an unpaired two-tail distribution t test, and the p-values are reported for each case. Each experiment was repeated two or more times with similar results unless otherwise stated.
Online supplemental material
Fig. S1 shows the intracellular distribution and variation in total αBTX levels during internalization. Fig. S2 demonstrates the snapshots from a time series of αBTX carrier movement close to the cell surface or inside the cell. Fig. S3 shows the effect of blocking lysosomal degradation on levels of endocytosed αBTX. Fig. S4 shows the relative intracellular distribution of αBTX with respect to fluid-phase marker or Tf. Videos 1–3 show morphology and movement of the AChR tubule moving away from the TIRF field (Videos 1 and 2) or the movement of these endosomes inside the cell (Video 3).
© 2008 Kumari et al. This article is distributed under the terms of an Attribution–Noncommercial–Share Alike–No Mirror Sites license for the first six months after the publication date (see http://www.jcb.org/misc/terms.shtml). After six months it is available under a Creative Commons License (Attribution–Noncommercial–Share Alike 3.0 Unported license, as described at http://creativecommons.org/licenses/by-nc-sa/3.0/).
S. Kumari and V. Borroni contributed equally to this paper.
A. Chaudhry's present address is Dept. of Immunology, University of Washington School of Medicine, Seattle, WA 98195.
B. Chanda's present address is Dept. of Physiology, University of Wisconsin, Madison, WI 53706.
R. Massol's present address is GI Cell Biology Laboratories, Children's Hospital Boston, Harvard Medical School, Boston, MA 02115.
Abbreviations used in this paper: αBTX, α-bungarotoxin; AChR, acetylcholine receptor; CNS, central nervous system; EEA1, early endosomal antigen 1; GPI, glycosyl-phosphatidylinositol; LAMP, lysosomal-associated membrane protein; NMJ, neuromuscular junction; Pak, p21-associated kinase; PBD, Pak-binding domain; PE, phycoerythrin; SA, streptavidin; Tf, transferrin; TfR, Tf receptor; TIRF, total internal reflection fluorescence.
We would like to thank Rahul Chadda and Abhijit Kale (National Centre for Biological Science [NCBS], Bangalore, India) for help with preparing labeled reagents and other members of the Mayor laboratory and NCBS for generous support and encouragement. We acknowledge the Wellcome Trust–aided imaging facility at NCBS under the able supervision of H. Krishnamurthy for help in confocal imaging. We are indebted to the Third World Academy of Sciences for associateships for S. Mayor and F.J. Barrantes and the Company of Biologists for travel grants to V. Borroni, R. Massoi, and B. Chanda. S. Mayor also thanks F.F. Bosphorus and K. Belur for inspiration.
F.J. Barrantes was supported by a 2002 award from the Sarojini Damodaran International Trust. S. Mayor was supported by the Wellcome Trust as an International Senior Research Fellow and by a Swarnajayanti Fellowship from the Department of Science and Technology (India). This work was supported in partnership by grants PICT01-12790 and 5-20155 from the Ministry of Science and Technology, grant PIP367 from the Argentinian Scientific Research Council (CONICET), by Philip Morris USA, Inc. and Philip Morris International, and by grant PGI 24/B135 from Universidad Nacional del Sur (Argentina) to F.J. Barrantes.