| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 35, 33583-33592, August 29, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¶ || || **
From the
Duke University School of Medicine,
Department of Neurobiology, Durham, North Carolina 27710, the
||University of Texas Health Science Center,
Department of Biochemistry, San Antonio, Texas 78229, and the
Marine Biological Laboratory, Woods Hole,
Massachusetts 02543
Received for publication, April 25, 2003 , and in revised form, June 4, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
100-amino acid domain that has
been found in a large number of proteins from yeast to man and named the
Eps15 homology
(EH)1 domain
(3). NMR studies show that the
structure of the central EH domain of Eps15 consists of a pair of EF hand
motifs, the second of which binds tightly to calcium
(4). EH domains are 60% similar
at the amino acid level, and the structurally critical residues are well
conserved, suggesting that their three-dimensional structure also is well
conserved. Binding partners for the EH domains of Eps15 and its homologues
include the clathrin assembly protein AP180
(5), epsin
(6,
7), and synaptojanin, an
inositol phosphatase involved in uncoating of clathrin-coated vesicles
(8). Most EH domains recognize
a well conserved Asp-Pro-Phe (NPF) motif found within their binding partners
(9). This NPF motif binds
within a hydrophobic pocket between two 
-helices of the central EH
domain of Eps15 (4). The
central domain of Eps15 contains a series of heptad repeats, and the
C-terminal domain of Eps15 contains 15 DPF repeats. This DPF-rich domain was
identified as one of the first of many binding partners that interact with the
-ear domain of AP-2, another clathrin assembly protein
(10–13).
Although its exact site of interaction is still unknown, Eps15 also interacts
both biochemically and genetically with dynamin, a GTPase believed to catalyze
the fission reaction as nascent coated vesicles are released from the membrane
(14). Given that many of the
binding partners of Eps15 are associated with clathrin, it is likely that
Eps15 might participate in some aspect of clathrin-mediated endocytosis.
Several studies have indeed implicated Eps15 in clathrin-mediated endocytosis. For example, Eps15 associates transiently with the growing edges of clathrin-coated pits, as identified by immunoelectron microscopy (15). Furthermore, clathrin-mediated endocytosis is inhibited in cells that overexpress Eps15 mutants lacking either the DPF-rich AP-2 binding domain (16) or the EH domains (17). Genetic disruption of the Caenorhabditis elegans homologue of Eps15, EHS-1, results in an uncoordinated phenotype and depletes synaptic vesicles, supporting a role for EHS-1 in SV recycling (14). In addition, Pan1p, a yeast homologue of Eps15, binds to and activates the actin-nucleating Arp2-Arp3 complex. This has led to the hypothesis that Eps15 plays an additional role in linking clathrin-coated pits to the actin cytoskeleton (5, 18).
Despite the wealth of evidence that Eps15 plays a role in clathrin-mediated endocytosis, the mechanism of action of this protein is not clear. Given that Eps15 is localized at the growing edges of clathrin-coated pits (15), it was hypothesized that Eps15 plays a role in the assembly of clathrin into coated pits and vesicles. We have evaluated this hypothesis by biochemically characterizing the effects of Eps15 on clathrin assembly. We have found that although Eps15 has no intrinsic clathrin assembly activity, it greatly stimulates the ability of AP180 to assemble clathrin. We have also identified two NPF-containing binding sites for Eps15 on sAP180, and we have shown that the Eps15-mediated stimulation of clathrin assembly is inhibited by peptides that mimic these Eps15-binding sites. To determine the physiological importance of these interactions, we utilized the squid giant synapse, a system well suited for study of the macromolecular interactions that underlie clathrin-mediated synaptic vesicle endocytosis (19–21). When injected into squid giant presynaptic nerve terminals, these NPF-containing peptides inhibit the formation of clathrin-coated pits and vesicles during synaptic vesicle endocytosis in vivo. These results indicate that interactions between Eps15 homology domains and NPF motifs are involved in clathrin-coated vesicle formation during synaptic vesicle recycling.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
.
Densitometry of SDS-PAGE gels was used to determine the concentrations of
sAP180, C45, and Eps15 utilizing the top band to quantify intact protein
relative to a clathrin standard. A peptide containing residues 387–405
of sAP180 was synthesized with an additional N-terminal cysteine (NPF1
peptide), and a peptide containing residues 433–450 of sAP180 was
synthesized with a C-terminal cysteine (NPF2 peptide). Mutant versions of each
peptide were also synthesized in which the NPF motifs were changed to AAA (Mut
NPF1 and Mut NPF2, respectively). All peptides were synthesized in the
University of Texas Health Science Center Institutional Protein Core
Laboratory (San Antonio, TX).
Surface Plasmon Resonance—All SPR experiments were done on a
BIAcore 3000 instrument using CM5 research grade sensor chips (BIAcore Inc.,
Piscataway, NJ) at 25 °C. A running buffer, HBS (HEPES-buffered saline
containing 10 mM HEPES, 150 mM NaCl, 3 mM
EDTA, and 0.005% surfactant P-20, pH 7.0) was used at a flow rate of 5
µl/min. Equivalent amounts of GST-sAP180, GST-hAP180-2, GST, and bovine IgG
(3000 response units (RUs), where 1000 RUs = a change of 1
ng/mm2 in surface protein concentration) were coupled to the sensor
chips utilizing an amine coupling kit (BIAcore Inc., Piscataway, NJ). Eps15
was passed over the surfaces for 6 min, followed by a 4-min injection of HBS.
The surfaces were then regenerated by a 3-min injection of 4 M
guanidine hydrochloride.
All peptides contained a cysteine at either their N-terminal or C-terminal
ends and were coupled to the sensor chip utilizing a thiol coupling kit
(BIAcore Inc., Piscataway, NJ). All the peptides were coupled to the sensor
chip at a similar level (800 RU). Eps15 was passed over the surfaces at
the indicated concentration for 6 min, followed by a 4-min injection of HBS
buffer. The surfaces were regenerated by a 3-min injection of 6 M
guanidine hydrochloride. The surfaces retained their activity for 20–50
experiments. All SPR studies were carried out in the University of Texas
Health Science Center for Macromolecular Interactions (San Antonio, TX).
Clathrin Assembly—The assembly of clathrin cages was
quantified as described previously
(20). Each assembly protein
was combined with 0.5 µM clathrin in the presence or absence of
Eps15. Clathrin assembly was initiated by adding 0.1 volume of 1 M
MES-NaOH, at the indicated pH. The mixture was incubated on ice for 45 min and
centrifuged at 400,000 x g at 4 °C for 6 min. The upper 80%
of the supernatant was removed and analyzed by SDS-PAGE and Coomassie Blue
staining. The gel bands were quantitated utilizing an Amersham Biosciences
Personal Densitometer and ImageQuant Software. Background amounts of clathrin
assembly were measured in the absence of added proteins; these values
constituted 2% of clathrin at pH 7.0, whereas at pH 6.7 and 6.5
3–5% of clathrin sedimented. The inhibition of clathrin assembly by
peptides was examined under the following conditions: 0.5 µM
clathrin, 1 µM C45, 1 µM Eps15, and 200–1000
µM peptide, 9 mM Tris-HCl, 0.1 M MES-NaOH,
pH 7.0. The amount of clathrin assembled was determined as described
above.
Solutions of assembled baskets were analyzed on 5–20% (w/v) glycerol gradients in 0.1 M MES-NaOH, pH 7.0. Typically 0.8 ml of samples were layered on 11 ml of gradient solution and centrifuged at 28,000 rpm for 90 min at 4 °C in an SW40 rotor (Beckman Instruments). One-ml fractions were collected from the bottom of the tubes by using a peristaltic pump and analyzed by SDS-PAGE.
Electrophysiology—Electrical measurements were made on giant synapses in isolated stellate ganglia of the squid, Loligo pealei (21). Ganglia were superfused with oxygenated physiological saline at 10–15 °C containing (in mM): 466 NaCl, 54 MgCl2, 11 CaCl2, 10 KCl, 3 NaHCO3, 10 HEPES, pH 7.2. An electrode filled with 3 M KCl was inserted into the presynaptic axon to inject current used to evoke presynaptic action potentials. Single action potentials were elicited by injecting current pulses (1 ms; 0.7–1.9 µA) every 30 s (0.03 Hz) into the presynaptic axon. A second electrode was inserted into the postsynaptic axon to measure PSPs resulting from transmitter released by the presynaptic action potentials. PSPs were quantified by measuring their initial rate of rise. A third electrode was inserted into the presynaptic terminal both to record presynaptic action potentials and to microinject peptide reagents. Electrical signals were detected with an Axoclamp-2A amplifier (Axon Instruments; Union City, CA), then acquired and analyzed with Axobasic programs written by F. E. Schweizer (UCLA).
For microinjection, NPF peptides and mutants were diluted in physiological solution containing (in mM): 250 potassium isothionate, 100 KCl, 100 taurine, 50 HEPES, pH 7.4, with NaOH. Peptides were injected from the microinjection pipette directly into presynaptic terminals by delivering pulses of positive pressure (10–80 ms; 10–100 pounds/square inch; N2 gas) from a Picospritzer injector (General Valve; Fairfield, NJ). Fluorescein isothiocyanate-dextran (Molecular Probes; 3 kDa; 100 µM) was co-injected in order to estimate the amount of peptide injected. Fluorescence was imaged with a Zeiss Axioskop upright microscope (x10, 0.25 NA objective), detected with a COHU SIT camera, and acquired with Image-1 software (Universal Imaging, Inc; Downingtown, PA).
Electron Microscopy—Following the injection of NPF peptides, when synaptic transmission was inhibited by greater than 95%, terminals were fixed for electron microscopy with 2.5% glutaraldehyde. Two terminals each were injected with NPF1 and Mut NPF1 peptides. Fixed terminals were processed as described (23). Briefly, an 80-nm section was taken every 50 µm through the entire length of the presynaptic terminal, resulting in 8–12 sections per injected terminal. Each section was magnified x12,000 and examined with a JEM-1200 ExII electron microscope (JEOL, Peabody, MA). Analysis was performed using Image-1 software.
For the electron microscopic analysis, distances between the active zone plasma membrane and individual synaptic vesicles were measured, as described (24). Because active zones are typically spaced about 1 µm apart, synaptic vesicles within a 500-nm radius surrounding the active zone were measured in order to ensure that each synaptic vesicle was counted only once. In addition, the total number of clathrin-coated pits and coated vesicles associated with active zones was counted in each section. Differences in the number and distribution of synaptic vesicles and coated vesicles were compared between NPF1 and Mut NPF1 peptide-injected terminals and analyzed using Origin 6.1 software.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
The amino acid sequence of sAP180 contains two NPF motifs, one at position 397–399 and a second at position 438–440. Because EH domains are known to interact with NPF motifs (9), we asked whether the two NPF motifs in sAP180 constitute Eps15-binding sites. Two 21-amino acid peptides (NPF1 and NPF2) representing the sequences surrounding these two NPF motifs were synthesized. Two additional peptides in which the NPF residues were mutated to AAA (Mut NPF1 and Mut NPF2) were used as controls. A cysteine residue was included at the end of each peptide, allowing disulfide interaction to attach these four peptides to adjacent surfaces of a Biacore sensor chip. There was significant binding of Eps15 to both the NPF1 and NPF2 peptides (Fig. 2A). This binding was concentration-dependent, because it was greater in the presence of 4 µM Eps15 (Fig. 2A) than in the presence of 1 µM Eps15 (Fig. 2B). At both concentrations, there was more binding of Eps15 to NPF1 than to NPF2 (Fig. 2, A and B), suggesting a higher affinity for NPF1 than for NPF2. In contrast, only background levels of binding were observed on the surfaces to which the control peptides (Mut NPF1 and Mut NPF2) were attached (Fig. 2, A and B). These results indicate that the two NPF motifs within sAP180, located at amino acids 397–399 and 438–440, constitute two separate Eps15-binding sites.
|
Eps15 Stimulates AP180-mediated Clathrin Assembly— Given that Eps15 binds to AP180, it is possible that Eps15 regulates the ability of AP180 to assemble clathrin. We examined this possibility by determining the effects of Eps15 on AP180-mediated clathrin assembly. For these experiments we used the 45-kDa C-terminal domain of sAP180 (amino acids 313–751; C45) because this domain is responsible for the ability of AP180 to assemble clathrin (21), contains the NPF motifs that mediate Eps15 binding, and is more stable than the full-length protein. The increased stability of C45 improved the quality of the quantitative analysis, although full-length sAP180 yielded comparable results. Because the ability of mammalian assembly proteins to polymerize clathrin depends upon pH (25), we first asked how pH influences the clathrin assembly activity of GST-C45 in vitro. As is the case for mammalian assembly proteins, C45 exhibited very little clathrin assembly activity at pH 7.0, but its activity dramatically increased at more acidic pH levels (Fig. 3). Thus, to allow easier detection of a possible regulatory influence of Eps15, we worked at a pH (7.0) where the endogenous activity of C45 was low.
|
We examined the effect of Eps15 on clathrin assembly. When clathrin polymerizes, it forms regular, polyhedral cages that pellet upon ultracentrifugation (25). Therefore, clathrin assembly can be measured as a loss of clathrin from the supernatant. When mixtures of Eps15 and clathrin were analyzed by SDS-PAGE, there was no loss of clathrin from the supernatant (Fig. 4A, 5th to 9th lanes) compared with the control condition (Fig. 4, A, 1st to 4th lanes, and B). Thus, Eps15 does not polymerize clathrin (Fig. 4B), which confirms a previous report (26). However, Eps15 dramatically stimulated clathrin assembly mediated by GST-C45, evident as a decrease in the amount of clathrin present in the supernatants (Fig. 4C). Increasing concentrations of GST-C45 (1–5 µM; 3rd to 7th lanes) produced very little polymerization of clathrin in comparison to the controls (1st and 2nd lanes). Yet adding 0.4 µM Eps15 yielded a dramatic reduction in the amount of clathrin and GST-C45 in the supernatants (8th to 12th lanes), indicating significant polymerization of clathrin into cages. Quantitative analysis of a series of experiments carried out at various concentrations of Eps15 indicated that Eps15 stimulated clathrin polymerization at concentrations as low as 0.1 µM (Fig. 4D). At the highest concentration of Eps15 tested (1 µM), clathrin assembly was comparable with that produced by C45 at pH 6.5 (Fig. 3). Therefore, it appears that the complex of Eps15 and C45 has a higher affinity for clathrin than does C45 alone. At a constant concentration of GST-C45, the activity of Eps15 saturated at 1 µM (Fig. 4, E and F). Human Eps15 also stimulates clathrin assembly mediated by mouse AP180, indicating the conservation of this effect among species.3 Thus, at physiological pH Eps15 potently stimulates clathrin assembly by AP180.
|
In the above studies, sedimentation was used as a measure of clathrin assembly. However, such measurements do not clearly distinguish between assembled clathrin baskets and large, nonspecific aggregates of clathrin. We considered the possibility of clathrin aggregation in three ways. First, clathrin aggregates should sediment easily after centrifugation at 14,000 x g for 10 min. There was no sedimentation of clathrin under these conditions (data not shown), suggesting that the clathrin was not aggregated. Second, we analyzed assembly reactions carried out with and without Eps15 by dynamic light scattering which can readily distinguish between aggregates and baskets. The particles assembled by C45 with and without Eps15 were indistinguishable and had an average diameter of 81 nm which is consistent with the sizes of clathrin cages that have been reported previously (27, 28). Third, we employed glycerol gradient centrifugation to determine the sedimentation coefficient of the assembled clathrin. As a control we first produced clathrin baskets with GST-C45 (2 µM) under standard conditions for assembly (pH 6.5; 1 µM clathrin; 0.1 M MES-NaOH) and then analyzed these on 5–20% glycerol gradients (Fig. 5A). Lanes 1–3 are from the top of the gradient and contained unpolymerized clathrin, which has a low sedimentation coefficient of 8 S (Fig. 5A). Lanes 6–8 contained assembled clathrin baskets, which have a sedimentation coefficient of 150 S (29). This profile with two distinct clathrin peaks is very different from that expected of clathrin aggregates, which would either create a smear of clathrin across the entire gradient or would cause clathrin to sediment at the bottom of the gradient. When this experiment was repeated at pH 7.0, where polymerization is minimal, <10% of the total clathrin was located in the fractions where baskets are found (Fig. 5B, lanes 6–8). Instead, most of the clathrin remained at the top of the gradient, indicating that it was in an unpolymerized state (Fig. 5B, lanes 1–3). When 1 µM Eps15 was included in the reaction, 60% of the clathrin was in the position corresponding to clathrin baskets (Fig. 5C, lanes 6–8), as expected if the clathrin was polymerized by the presence of Eps15. These results indicate that the structures formed in the presence of Eps15 are clathrin baskets rather than aggregates.
|
GST-C45 co-sedimented with assembled clathrin baskets (Fig. 5, A and C). The relative molar ratio of clathrin to GST-C45 in these fractions was nearly one to one, consistent with previous reports that AP180 assembles clathrin stoichiometrically (29). In contrast, very little Eps15 was found in the position of the baskets (Fig. 5C, lanes 6–8) but instead remained at the top of the gradient (Fig. 5C, lanes 1–3). This absence of Eps15 in the assembled clathrin fraction was more clearly seen when the peak fractions (6–8) were concentrated and analyzed by SDS-PAGE (Fig. 5D). The relative molar ratio of clathrin to GST-C45 was similar in both cases, whereas very little Eps15 was present in the assembled baskets (Fig. 5, C and D). These results reveal that Eps15 associates only transiently with the clathrin baskets. Furthermore, these results are consistent with the finding that Eps15 is present at the rims of clathrin-coated pits but does not associate with clathrin in mature coated vesicles (15).
We next wanted to determine whether the ability of Eps15 to stimulate AP180-mediated clathrin assembly is due to the formation of an Eps15-AP180 complex. If this stimulation is due to a transient interaction of Eps15 with AP180, then the stimulatory effect should be inhibited by the NPF peptides that mimic the Eps15-binding sites on C45 (see Fig. 2). In support of this hypothesis, both NPF1 and NPF2 peptides significantly reduced the ability of Eps15 to stimulate clathrin assembly in a concentration-dependent manner (Fig. 6, A and B). In contrast, neither Mut NPF1 nor Mut NPF2 had any significant effect on clathrin assembly (Fig. 6, A and B). These results indicate that the ability of Eps15 to influence clathrin assembly is due to its interaction with the NPF motifs of sAP180.
|
Collectively, these biochemical results demonstrate that Eps15 stimulates the ability of AP180 to assemble clathrin (Figs. 4 and 5) and that this stimulation is inhibited by NPF peptides (Fig. 6) in vitro. This led us to hypothesize that Eps15 regulates the efficiency of the clathrin-coated pit and coated vesicle formation by AP180 in vivo. We then set out to evaluate the predictions of this hypothesis in vivo. Because Eps15 and AP180 interact utilizing motifs that are found in many proteins, it was not possible to design inhibitors that prevent only the interaction of Eps15 with AP180. However, the NPF peptides from AP180 did allow us to determine what happens in vivo when the general interactions of EH domains and NPF motifs are inhibited. The prediction of our hypothesis is that the inhibition of EH domain-NPF motif interactions should result in a corresponding reduction in the number of clathrin-coated pits and vesicles that form during endocytosis. We chose to examine the effects of NPF peptides on clathrin-mediated synaptic vesicle endocytosis at the squid giant synapse. Clathrin-coated vesicle formation plays a crucial role in recycling of synaptic vesicles at this synapse, because reagents that interfere with clathrin assembly inhibit both neurotransmitter release and synaptic vesicle recycling (20, 21).
In order to evaluate whether EH domain-NPF motif interactions influence coated vesicle formation in vivo, we determined the effects of microinjected NPF1 on neurotransmitter release and presynaptic ultrastructure. Neurotransmitter release was measured by monitoring the postsynaptic responses to presynaptic action potentials elicited at a low frequency (0.03 Hz). When injected into the presynaptic terminal, NPF1 reduced the initial rate of rise (the slope) of the postsynaptic potentials (Fig. 7A). This indicates that this peptide inhibits neurotransmitter release from the presynaptic terminal. This inhibition of transmitter release occurred without altering the presynaptic action potentials (Fig. 7A) and was reversible, indicating that the inhibition was not due to microinjection damage (Fig. 7B). Mut NPF1, which had no effect on the ability of Eps15 to stimulate sAP180-mediated clathrin assembly (Fig. 6A), also had no effect on synaptic transmission (Fig. 7C). This demonstrated that the inhibitory effect of the NPF1 peptide was sequence-specific. The inhibition of transmitter release depended upon the concentration of NPF1 microinjected, whereas Mut NPF1 had no effect at any concentration tested (Fig. 7D). NPF2 also inhibited synaptic transmission in a concentration-dependent manner, although not as potently as NPF1 (Fig. 7D). On average, NPF1 inhibited synaptic transmission by 47%, and NPF2 inhibited synaptic transmission by 19% over the range of concentrations tested (Table I). This is consistent with the higher affinity of Eps15 for NPF1 compared with NPF2 (Fig. 2). These parallels between peptide inhibition of transmitter release and binding to Eps15 establish that interactions between EH domains and NPF motifs are important for maintaining neurotransmitter release.
|
|
If Eps15 regulates clathrin assembly by AP180, then one consequence of the inhibition of EH domain-NPF motif interactions should be a decrease in the number of clathrin-coated pits and vesicles that form during synaptic vesicle endocytosis. To determine the effects of NPF1 peptide on presynaptic ultrastructure, two nerve terminals were injected with NPF1 until synaptic transmission was inhibited by greater than 95% and then fixed and examined with electron microscopy. Mut NPF1 was injected into two other terminals as controls. There were no obvious differences in synaptic morphology in EM images of terminals injected with NPF1 or Mut NPF1 peptides (Fig. 8). A quantitative analysis was then performed to determine whether there were any subtle differences in the structure that could reveal the synaptic function of EH domain-NPF motif interactions. The number and spatial distribution of synaptic vesicles, as well as the number of coated pits and vesicles were measured in terminals injected with NPF1 and Mut NPF1. This analysis revealed only small differences in the spatial distribution of SVs in the terminals injected with NPF1 relative to the control terminals injected with Mut NPF1 (Fig. 9, A and B). NPF1 peptide resulted in a modest, but significant, 8% depletion of synaptic vesicles (Fig. 9C; p < 0.05). However, there was a dramatic reduction in both the number of coated pits (Fig. 9D) and the number of coated vesicles (Fig. 9E) in the terminals injected with NPF1 compared with Mut NPF1. The number of coated pits was reduced by 49%, and the number of coated vesicles was reduced by 34%; both of these decreases are statistically significant (p < 0.05). These results indicate that NPF1 inhibited coated vesicle formation from the plasma membrane and support a role for EH domain-NPF motif interactions in regulating clathrin-mediated endocytosis in nerve terminals.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Eps15 Interacts with AP180 and Facilitates Clathrin Coat Assembly—By using a sensitive surface plasmon resonance assay, we have shown that both sAP180 and hAP180-2 bind to Eps15 (Fig. 1). Furthermore, we have demonstrated that the two NPF motifs in sAP180 are Eps15-binding sites, because a direct interaction was measured between Eps15 and two NPF-containing peptides from sAP180, and this interaction was disrupted by mutating the NPF motif of these peptides (Fig. 2). The apparent KD values of 0.5 µM (Eps15-sAP180) and 0.7 µM (Eps15-hAP180-2) indicate that the interaction between AP180 and Eps15 is of moderate, but not high, affinity. This conclusion is consistent with experiments reported by others utilizing a variety of experimental approaches. The yeast Eps15 homologue, Pan1p, was demonstrated to be a binding partner for the yeast AP180 homologue based on yeast two-hybrid, GST pull-down, and co-immunoprecipitation experiments (5). Interactions between a recombinant fragment of hAP180-2 and Eps15 were detected in a far Western assay, although in this study hAP180-2-Eps15 interactions were not detected in GST pull-down or co-immunoprecipitation experiments (31). This may be due to the moderate affinity of the AP180-Eps15 interaction; such interactions may be disrupted during the extensive washing steps involved in column binding and co-immunoprecipitation experiments.
Whereas Eps15 had no intrinsic ability to assemble clathrin, we found that Eps15 greatly stimulated the clathrin assembly activity of AP180 under physiological conditions (Figs. 4 and 5). This stimulation required an interaction between Eps15 and AP180, because it was specifically inhibited by peptides representing a site of interaction between these two proteins (Fig. 6). Unlike the clathrin assembly protein sAP180 (or its C-terminal domain, C45), Eps15 was not incorporated in stoichiometric amounts into the clathrin coats (Fig. 5). This suggests that Eps15 either activates AP180 catalytically or acts stoichiometrically but rapidly dissociates once the clathrin coat is formed. The latter explanation appears attractive, given that Eps15 has been found to dissociate from AP-2 once clathrin coats form (26). This is also consistent with immunoelectron microscopy localization of Eps15 to the rims of coated pits (15).
Based on the localization of Eps15, it was hypothesized that dumbbell-shaped anti-parallel tetramers of Eps15 could span two vertices at the edge of a growing clathrin-coated pit (22). Such an arrangement could facilitate the recruitment of assembly proteins to the edges of the growing clathrin lattice and/or promote local curvature by facilitating the formation of pentagonal facets. Our finding that Eps15 promotes AP180-mediated clathrin assembly at physiological pH (Figs. 4 and 5), conditions where assembly of clathrin normally is poor, provides strong support for the idea that Eps15 serves as a physiological regulator of adaptor-mediated clathrin assembly. Furthermore, we suggest that the reason for AP-mediated assembly of clathrin to be inefficient at physiological pH is to allow assembly to be regulated by other proteins, such as Eps15. This is consistent with previous studies indicating that the AP180-AP2 interaction also serves to facilitate cooperative clathrin assembly at physiological pH (32).
Whereas the squid homologue of AP180 was primarily utilized in this study, other AP180 homologues, including those found in yeast, C. elegans, Drosophila, and human, also contain multiple copies of the NPF motif. Several observations indicate that our findings can be generalized to these other forms of AP180. First, we demonstrate here that human AP180-2, which contains two NPF motifs, binds to Eps15 (Fig. 1D). Second, surface plasmon resonance studies reveal that the C. elegans AP180 homologue unc-11C also binds to human Eps15.2 Third, Eps15 still is capable of stimulating clathrin assembly mediated by mouse AP180-1, albeit to a lesser extent than what we have observed with sAP180.3 Interestingly, although mouse AP180-1 does not contain any canonical "NPF" motifs, mouse AP180-1 has two DPF motifs that align precisely with the two NPF motifs of sAP180, as identified by analysis of the alignment of multiple sequences of AP180 homologues (21). Furthermore, there is evidence that DPF motifs may also interact with EH domains, although with lower affinity than NPF motifs (33). Indeed, it has been suggested that tetramerization of Eps15 may involve interaction of the N-terminal EH domains with DPF motifs found in the C-terminal domain (22, 34). Therefore, it is possible that DPF motifs in mammalian AP180 substitute for the NPF motifs found in the other homologues to provide an interaction interface for Eps15.
EH Domains Play a Regulatory Role in Synaptic Vesicle Endocytosis—When microinjected into living nerve terminals, the NPF peptides inhibit synaptic transmission (Fig. 7). Microinjection of NPF1 clearly affected synaptic vesicle endocytosis, because there was a substantial reduction in the number of clathrin-coated pits and coated vesicles following NPF1 injection (Figs. 8 and 9). Therefore, NPF1 seems to block an early stage of clathrin-mediated endocytosis, indicating that EH domain-NPF motif interactions are important for clathrin-coated pit and vesicle formation in vivo. Of the EH domain-NPF motif interactions that have been characterized, the two most likely to be important for formation of clathrin-coated pits are the interactions of Eps15 with AP180 and epsin. Both AP180 and epsin are known to promote clathrin assembly (7, 21, 35), and Eps15 stimulates the clathrin assembly activity of AP180 (Figs. 4 and 5). Although synaptojanin contains NPF motifs and is a binding partner of both Eps15 and intersectin, the major function of synaptojanin is to uncoat coated vesicles (36). Thus, it is unlikely that our results arise from a disruption of the uncoating function of synaptojanin because this would yield an increase in coated vesicles, rather than the depletion that we observed. It also is unlikely that NPF1 interfered with the interaction between Eps15 and AP-2, because AP-2 interacts with the DPF-rich C-terminal domain of Eps15 rather than at the N-terminal EH domains of Eps15 (10).
Synaptic vesicle endocytosis is severely impaired when clathrin assembly is prevented by peptides that prevent the interaction of clathrin with adaptor proteins such as AP180 and AP-2 (20). Under these conditions, there is a massive depletion of synaptic vesicles and coated vesicles, and a build up of plasma membrane that quantitatively parallels the amount of membrane lost from the synaptic vesicle pool (20). The effects of the NPF1 peptide were much less severe, consistent with the idea that interactions between the EH domains of Eps15 and AP180 serve a regulatory role, rather than an essential one, in the endocytotic process. Another possibility is that synaptic vesicles are not depleted because the NFP1 peptide causes a simultaneous inhibition of exocytosis (see next paragraph) that prevents vesicles from being depleted despite the inhibition of endocytosis. In contrast to our results, inhibition of Eps15 expression by RNA interference causes a massive depletion of synaptic vesicles in C. elegans (14). This severe phenotype may be caused by the fact that Eps15 interacts with a large number of endocytic proteins, and all of these interactions will be disrupted following loss of Eps15. Despite this difference, the two sets of results are complementary in supporting a role for Eps15 in synaptic vesicle endocytosis.
EH Domains May Be Involved in Synaptic Vesicle Exocytosis—Interfering with clathrin-based synaptic vesicle recycling inhibits synaptic transmission (20, 21). This inhibition of synaptic transmission is caused by depletion of SVs and/or accumulation of SV membrane in the plasma membrane. In contrast, it is remarkable that the NPF peptides inhibited synaptic transmission without a massive depletion of SVs. Following injection of NPF1 for ultrastructural examination, synaptic transmission was reduced by 95%, whereas the number of SVs was reduced by only 8% (Fig. 9). Clearly, in this case the inhibition of exocytosis cannot be due to loss of SVs. The mechanism involved in this inhibition of exocytosis is not clear. One possibility is that the NPF1 peptide could interfere with cargo selection during endocytosis, thereby depriving recycled SVs of a protein that is needed for vesicle fusion. Alternatively, it is possible that EH domain proteins function during exocytosis in addition to their well established role in endocytosis. For example, it is possible that the NPF1 peptide may interact with the EH domain of intersectin, a protein known to bind to the exocytic SNARE protein, SNAP-25 (37). More experiments will be required to understand the reason that exocytosis is impaired following disruption of EH domain-NPF motif interactions.
Eps15 and the Temporal Regulation of the Endocytic Reaction—Eps15 is a key protein in the web of interactions involving clathrin; indeed Eps15 may serve as a molecular link between the sequential steps of the endocytic reaction. Pan1p, a yeast Eps15 homologue, via interactions with Arp2–Arp3 complex, may contribute to a reorganization of the actin cytoskeleton that precedes coated pit formation (18). Eps15 contains a ubiquitin-interacting motif that may be involved in recruiting ubiquitinated cargo molecules into clathrin-coated pits (38). Further support for a role in cargo selection comes from the interaction of Eps15 with stonin, a synaptotagmin binding partner (39). The interactions of Eps15 with the clathrin assembly proteins AP180 (5), AP-2 (11), and epsin (6) have suggested a role for Eps15 in clathrin assembly. Our work here demonstrates that Eps15 is indeed an important regulator of AP180-mediated clathrin assembly. Eps15 has also been implicated in vesicle scission because of its interactions with dynamin (14) and in uncoating by virtue of its interactions with synaptojanin (8). By participating in cytoskeletal rearrangements, cargo selection, clathrin-coated pit formation, clathrin-coated vesicle scission, and clathrin-coated vesicle uncoating, Eps15 may direct the temporal linkage of these sequential reactions that underlie endocytosis.
|
FOOTNOTES |
|---|
¶ Present address: Yale University School of Medicine, Howard Hughes Medical
Institute, Cell Biology Dept., 295 Congress Ave., BCMM 236, New Haven, CT
06510.
** To whom correspondence should be addressed: Dept. of Biochemistry, MSC 7760, the University of Texas Health Science Center, 7703 Floyd Curl Dr., San Antonio, TX 78229-3900. Tel.: 210-567-3764; Fax: 210-567-6899; E-mail: Lafer{at}UTHSCSA.edu.
1 The abbreviations used are: EH, eps15 homology; sAP180, squid AP180; C45,
45-kDa C-terminal domain of sAP180; DPF, Asp-Pro-Phe; GST, glutathione
S-transferase; hAP180-2, human AP180-2 (CALM); NPF, Asn-Pro-Phe; PSP,
postsynaptic potential; RUs, response units; SPR, surface plasmon resonance;
SV, synaptic vesicle; MES, 4-morpholineethanesulfonic acid.
2 E. M. Lafer, K. Prasad, and A. Alfonso, unpublished observations.
3 K. Prasad and E. M. Lafer, unpublished observations.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
