SCAMP1 function in endocytosis.

Secretory carrier membrane proteins (SCAMPs) are ubiquitous components of recycling vesicles that shuttle between the plasma membrane, endosomes, and the trans-Golgi complex. SCAMPs contain multiple N-terminal NPF repeats and four highly conserved transmembrane regions. NPF repeats often interact with EH domain proteins that function in budding of transport vesicles from the plasma membrane or the Golgi complex. We now show that the NPF repeats of SCAMP1 bind to two EH domain proteins, intersectin 1, which is involved in endocytic budding at the plasma membrane, and gamma-synergin, which may mediate the budding of vesicles in the trans-Golgi complex. Expression of SCAMP1 lacking the N-terminal NPF repeats potently inhibited transferrin uptake by endocytosis. Our data suggest that one of the functions of SCAMPs is to participate in endocytosis via a mechanism which may involve the recruitment of clathrin coats to the plasma membrane and the trans-Golgi network.

SCAMPs 1 (secretory carrier membrane proteins) were discovered as major components of secretory vesicles in exocrine glands and later shown to be universally present in vesicles that recycle at the plasma membrane (1)(2)(3)(4). At least three SCAMPs are expressed in vertebrates, and all cells tested express at least one SCAMP isoform (5). 2 All SCAMPs are composed of four transmembrane regions with cytoplasmic Nand C-terminal regions. The function of SCAMPs is unknown. SCAMPs 1 and 3 are tyrosine-phosphorylated by the epidermal growth factor-receptor, suggesting that they are subject to regulation by phosphorylation (6). A knockout of SCAMP1, the most abundant SCAMP isoform, failed to display a major phenotype and exhibited only a moderate defect in membrane traffic in mast cells as analyzed by capacitance recordings (7). This defect could have been due to a change in the rate constants of either exocytosis, endocytosis, or both. The knockout indicates a function for SCAMPs in membrane traffic, with the mildness of the phenotype probably due to redundancy with the other SCAMP isoforms that appear to be co-expressed in most cells with SCAMP1.
Sequence analyses revealed that SCAMPs contain multiple NPF (asparagine-proline-phenylalanine) repeats in the cytoplasmic N terminus. NPF represents a binding motif for EH domains (8,9). EH domains are typically found in proteins involved in endocytosis, such as EPS15 (the origin of the term EH domain; see Refs. 10 -15) and intersectin 1/Ese 1/EHSH1 (16 -19). The presence of NPF repeats in the N terminus of SCAMPs suggests a potential role in binding to cytosolic EH domain proteins, possibly during endocytosis. Although several NPF ligands for EH domain proteins have been described, all of them are soluble cytosolic proteins. A role for transmembrane proteins, such as SCAMPs, as docking sites for cytosolic EH domain proteins is of potential significance because it would provide an intramembranous docking site for protein complexes in endocytosis. One of the major problems in understanding vesicular budding during endocytosis is how the endocytic clathrin coat is assembled on the membrane. It seems likely that specific binding sites for the adaptor protein complex AP-2 nucleate assembly of the clathrin coat on the plasma membrane (20 -22). Assembly of clathrin coats on intracellular membranes, e.g. the trans-Golgi apparatus, likely also involve recruitment of the appropriate adaptor complexes, such as AP-1, to the site of budding (23). The only currently available candidates for a nucleation site for endocytosis are synaptotagmins which bind to AP-2 with high affinity (24,25). However, synaptotagmins are primarily expressed in neurons, with extremely low levels of some isoforms outside of the brain. This makes it unlikely that synaptotagmins represent universal AP-2 assembly sites that are present in all eukaryotic cells. In the current study, we have pursued the hypothesis that SCAMPs are intrinsic membrane proteins of recycling vesicles which serve as assembly sites for clathrin coats on the plasma membrane and Golgi apparatus.
Yeast Two-hybrid Screens and Interaction Analyses-Screening of a rat brain cDNA library with pLexN-SCAMP1 as described previously * This work was supported by Grant RO1-GM55562 from the National Institutes of Health (to J. A.) and by a postdoctoral fellowship from the Spanish Ministry of Education and Culture and the Fulbright Commission (to R. F.-C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF242544 (for full-length rat synergin sequence). ** Present address: Dept. of Neurobiology and Anatomy, UT Houston Medical School, 6431 Fannin St., Houston TX 77030.
‡ ‡ To whom correspondence should be addressed. Tel.: 214-648-1876; Fax: 214-648-1879; E-mail: tsudho@mednet.swmed.edu. 1 The abbreviations used are: SCAMP, secretory carrier membrane protein; PMSF, phenylmethylsulfonyl fluoride; GST, glutathione Stransferase; MBP, maltose-binding protein. 2 R. Ferná ndez-Chacón and T. C. Sudhof, unpublished observation. (17,26,27) yielded ϳ170 positive colonies from 2.7 ϫ 10 Ϫ7 transformants. 16 clones with the strongest interactions based on ␤-galactosidase activation assays were sequenced. Four of these clones represented two independent isolates of ␥-synergin (28), two corresponded to BAT3, which is a large proline-rich protein of unknown function whose gene is located in the class III region of the human major histocompatibility complex (29), and the remaining clones were likely artifacts because they were transcription factors or proteins frequently isolated in our yeast two-hybrid screens independent of the bait. The two ␥-synergin prey clones encoded residues Ala 55 -Gly 512 and Gly 105 -Ser 465 of fulllength ␥-synergin, which extends 57 amino acids more N-terminally than the published sequence. Quantitative liquid ␤-galactosidase assays to estimate the strengths of yeast two-hybrid interactions were performed as described previously (26,27) with yeast L40 cells grown for 48 h in 5 ml of -UTL medium lacking uridine, tryptophan, and leucine. All measurements were performed in triplicates, normalized to the amount of protein determined in the same samples, and calculated as nanomoles of o-nitrophenol/min and milligrams of protein.
Affinity Chromatography-Rat brains were homogenized in 10 mM HEPES/NaOH, pH 7.4, 1 mM EGTA, 0.1 g/liter phenylmethylsulfonyl fluoride (PMSF) and other protease inhibitors (leupeptin, aprotinin, and pepstatin), extracted for 1 h at 4°C after addition of an equal volume of 10 mM HEPES/NaOH, pH 7.4, 1 mM EGTA, 0.1 g/liter PMSF, protease inhibitors, 0.2 M NaCl, 2% Triton X-100, and centrifuged at 100,000 ϫ g for 30 min at 4°C. For SCAMP1 batch affinity chromatography experiments, 0.1 ml of glutathione-agarose beads containing ϳ0.4 mg of GST-SCAMP1 or GST alone were incubated overnight at 4°C with 10.5 ml of total brain extract in buffer A (10 mM HEPES/ NaOH, pH 7.4, 1 mM EGTA, 0.1 g/liter PMSF, protease inhibitors, 0.1 M NaCl, 1% Triton X-100). The glutathione-agarose was recovered by centrifugation, washed three times with 5 ml of buffer A, resuspended in 0.4 ml of sample buffer, and 20 l of the eluted material was analyzed by SDS-polyacrylamide gel electrophoresis and immunoblots. MBP-Synergin affinity chromatography experiments were peformed in a similar manner but incubation was done in 1.5 ml and washed four times with 1.5 ml of buffer A.
Endocytosis Assay-COS cells were transfected with ϳ1 g of DNA of the indicated constructs using LipofectAMINE (Life Technologies, Inc.) on day 1, transferred to coverslips on day 2, and endocytosis was assayed on day 3. Cells were starved in serum-free medium for 30 -60 min, incubated with 20 g/ml of fluorescein isothiocyanate-transferrin for 15 min, rinsed with phosphate-buffered saline, and fixed with 3% paraformaldehyde in phosphate-buffered saline for 25 min at room temperature. Cells were permeabilized by immersion in Ϫ20°C acetone for 5 min, and stained with anti-SCAMP1 polyclonal antibodies (dilution ϭ 1:600). Images were acquired using a confocal microscope (Bio-Rad MRC1000 with an Axiovert 100 Zeiss microscope). Endocytosis was scored as the percentage of cells competent to internalize labeled transferrin, which was distributed in a punctate pattern after uptake (30). Cell were considered endocytosis-incompetent if they lacked fluorescent puncta. All transfections and analyses were performed by experimenters who did not know the nature of the plasmids used for a particular transfection.
Miscellaneous Procedures-␥-Synergin cDNA cloning was performed by screennig a rat brain cDNA library by standard procedures (17). SDS-polyacrylamide gel electrophoresis and immunoblotting were performed according to standard procedures (31,32). Monoclonal anti ␣-adaptin antibodies were obtained from Sigma; polyclonal anti-amphiphysin 1 antibodies were a gift from Synaptic Systems (Göttingen, Germany). Antibodies to SCAMP1 were described previously (7). For RNA blots, a rat multitissue RNA blot (CLONTECH) was probed with uniformly [␣-32 P]dCTP labeled 1.2-kilobase pair SalI cDNA fragment from the ␥-synergin prey clone. Prehybridizations and hybridization were performed at 42°C overnight in 50% formamide plus Denhardt's solution 0.1 mg/ml salmon sperm DNA. Filters were washed twice for 30 min at 65°C in 0.2 ϫ SSC and 0.5% SDS and exposed for 1-5 days. All figures and illustrations were produced in Adobe Photoshop and Adobe Illustrator.

RESULTS
Identification of ␥-Synergin as a SCAMP-binding Protein-We used data bank searches to study the evolutionary conservation and domain structure of SCAMPs. Highly conserved SCAMP sequences were detected in plants (Arabidopsis thaliana), nematodes (Caenorhabditis elegans), and insects (Drosophila melanogaster) in addition to vertebrates (data not shown). Sequence alignments demonstrated that all available full-length SCAMP structures contain multiple NPF repeats and a proline-rich sequence in the N-terminal cytoplasmic domain that precedes the four transmembrane regions. The presence of NPF repeats in SCAMPs was intriguing because NPF repeats represent binding sites for EH domains, autonomously folding protein modules that are primarily observed in proteins involved in endocytosis (8,9). Since SCAMPs appear to be universal membrane components of recycling vesicles in all cells tested (2,5), their interaction with cytosolic adaptor proteins in endocytosis could provide a mechanism to initiate the assembly of clathrin coats for endocytic budding. To test this idea, and to identify potential EH domain proteins which bind to SCAMP, we screened a rat brain cDNA library by yeast two-hybrid selection for interacting proteins with the N-terminal sequences of SCAMP1.
Sequencing of the prey clones obtained in the yeast twohybrid screen showed that ␥-synergin was the major protein isolated. Two independent clones for ␥-synergin were selected multiple times. ␥-Synergin is a recently described cytoplasmic protein that contains a central EH domain (28), suggesting that the EH domain could bind to the NPF repeats of SCAMP1. This hypothesis is supported by the fact that the two ␥-synergin clones isolated in the yeast two-hybrid screens overlap in the EH domain. The C terminus of ␥-synergin binds to ␣-adaptin, a component of the Golgi clathrin adaptor protein complex AP-1, but the ␥-synergin C terminus was not present in the yeast two-hybrid prey clones. To characterize ␥-synergin further, we isolated from a rat brain cDNA library multiple independent cDNA clones that covered the N-terminal two-thirds of the protein (data not shown). Sequencing of these clones confirmed the previously described structure of ␥-synergin except that at the N terminus, the rat ␥-synergin cDNA had an open reading frame that extends 57 residues beyond the reported N terminus of human ␥-synergin (28), suggesting that the N terminus of ␥-synergin is longer than previously thought. In addition, sequence analyses revealed multiple internal repeats in the Cterminal half of ␥-synergin, indicating that ␥-synergin is composed of distinct domains (data not shown). RNA blots showed that ␥-synergin is expressed in all tissues tested, with at least some tissues synthesizing multiple species of mRNA which appears to be a relatively abundant (data not shown).
SCAMP1 Binds Selectively to the EH Domains of ␥-Synergin and Intersectin 1-The yeast two-hybrid results suggested that the N-terminal NPF repeats of SCAMP1 might bind to the EH domain of ␥-synergin. To confirm this hypothesis, we further characterized the interaction of the two different ␥-synergin prey clones with the N-terminal region of SCAMP1. Two yeast two-hybrid assays were used, transactivation of histidine auxotrophy (Fig. 1A) or quantitation of ␤-galactosidase (Fig. 1B). Either the complete N-terminal sequence of SCAMP1 with the NPF repeats or a truncated sequence lacking the NPF repeats were tested. With both assays, SCAMP1 only interacted with ␥-synergin when the NPF repeats were present, providing further evidence that the N-terminal NPF repeats bind to the EH domain of ␥-synergin.
We next asked how specific the binding of ␥-synergin to SCAMPs is and if all EH domains or only a subset of EH domains bind to the NPF repeats of SCAMP1. GenBank TM searches revealed that the EH domain of ␥-synergin is most closely related to the two EH domains of intersectin 1, whereas the classical EH domain protein EPS15 is less homologous (data not shown). Therefore we tested in the same yeast twohybrid assays described above if the EH domains of intersectin 1 and Eps15 interact with SCAMP1 (Fig. 1). Again, a highly specific interaction of SCAMP1 with intersectin was observed. The SCAMP1/intersectin interaction depended on the presence of NPF repeats in SCAMP1 and of the EH domains in intersectin as would be expected for an EH domain-mediated binding reaction. In contrast to intersectin, Eps15 constructs containing one or two EH domains were inactive in both interaction assays. The Eps15 construct containing all three EH domains was borderline positive in the histidine auxotrophy assay but inactive in the ␤-galactosidase assay, suggesting a very weak interaction (Fig. 1). These data show that the NPF repeats of SCAMP1 specifically bind to a subset of EH domains, including those of ␥-synergin and intersectin 1.
Confirmation of the Interaction of SCAMP1 with ␥-Synergin and Intersectin 1 Using Pulldown Experiments-To confirm the yeast two-hybrids results with an independent protein-based assay, we performed pulldown experiments with rat brain proteins. First we used a MBP fusion with the EH domain of ␥-synergin to affinity purify rat brain proteins ( Fig. 2A). Immunoblotting revealed that SCAMP1 was bound only to the MBP-synergin fusion protein but not to MBP alone, suggesting that the EH domain of ␥-synergin can capture SCAMP1 from the brain extract. Next we used a fusion protein of GST with the N-terminal domain of SCAMP1 in similar pulldown experiments, with GST alone as a control. Immunoblotting revealed that intersectin is highly enriched in the bound protein frac-tion; in fact, it was the most enriched protein (Fig. 2B). Because we did not have an antibody to ␥-synergin, we could not test its binding in these experiments. However, immunoblotting also showed that ␣-adaptin of the AP-2 adaptor complex was enriched in the pellet. We currently do not know if AP-2 binds directly to a sequence in SCAMP1 or if it binds indirectly via intersectin. Furthermore, SNAP-25 (which we previously showed binds to intersectin (17)) was also specifically present in the bound proteins. Two controls indicate that these observations reflect specific binding reactions. First, GST alone was unable to capture either intersectin 1, AP-2, or SNAP-25 (lane 4, Fig. 2B). Second, synapsins and amphiphysin, which are relatively abundant and sticky proline-rich proteins, were not bound to GST-SCAMP1 (Fig. 2B).

FIG. 1. Analysis of yeast two-hybrid interactions of SCAMP1
with EH domain proteins. A, assay for histidine auxotrophy. Yeast L40 cells harboring the indicated combinations of bait and prey vectors were plated on selection medium lacking histidine (-THULL) to test the various vector combinations for induction of histidine auxotrophy. Control platings on selection medium containing histidine (-UTL) were performed to ensure that all vectors were present. The bait plasmids used encode the N-terminal cytoplasmic region of SCAMP1 with NPF repeats (SCAMP1) or without NPF repeats (SCAMP1-⌬NPF) and pLexN vector alone. Prey plasmids used include a vector-only control (pVP16) and vectors containing the following inserts: the two ␥-synergin prey vectors with the EH domain (synergin 1 and 2), intersectin 1 with the two N-terminal EH domains (ITSN-1) or without EH domains (ITSN-⌬EH), and EPS15 with only the first EH domain (Eps15 EH1), the first two EH domains (Eps15 EH1,2), or all three EH domains (Eps15). B, measurements of ␤-galactosidase activity. Yeast L40 cells harboring the same bait and prey vector combinations described above were harvested and their ␤-galactosidase activity induced by transactivation via the bait/prey protein interaction was measured by a quantitative liquid assay. Data shown are means Ϯ S.E. from triplicate determinations.

FIG. 2. Binding of ␥-synergin and intersectin 1 to SCAMP1 demonstrated by pulldown experiments.
A, SCAMP1 binding to ␥-synergin immobilized as a fusion protein with MBP-synergin. Detergent-solubilized rat-brain proteins (lane 1) were affinity-purified on immobilized MBP-synergin or MBP alone (MBP-control), and the flowthrough (lanes 2 and 3) and bound fractions (lanes 4 and 5) were analyzed by immunoblotting with polyclonal SCAMP1 antibodies. B, affinity purification of intersectin 1 (ITSN1), ␣-adaptin, and SNAP-25 on immobilized GST-SCAMP1. Experiments were performed with detergent solubilized rat brain proteins (lane 1) as in A, except that immobilized GST-SCAMP1 containing the N-terminal cytoplasmic sequences of SCAMP1 was used as an affinity matrix, with GST alone as a control. Flow-through (lanes 2 and 3) and bound proteins (lanes 4 and 5) were analyzed by immunoblotting for the proteins indicated on the right, with synapsins and amphiphysin which are very sticky proteins used as a control in the bottom panel. Positions of molecular weight markers are indicated on the left. SCAMP1 Functions in Endocytosis-Because SCAMPs are universally present in all cells tested, their binding to EH domain proteins could mediate the membrane recruitment of clathrin coats for budding, as occurs for example during receptor-mediated endocytosis. To test directly if SCAMP1 is involved in endocytosis, we transfected full-length SCAMP1, SCAMP1 with a deletion of the N-terminal NPF repeats, and the N-terminal NPF repeats of SCAMP1 without transmembrane regions into COS cells. Transfected cells were then incubated with fluorescently labeled transferrin, fixed, and stained for SCAMP1 to identify transfected cells. An observer who was blind to the nature of the transfected proteins quantitated the transfection results by scoring the percentage of cells that had taken up transferrin by endocytosis (Fig. 3).
COS cells transfected with full-length SCAMP1 display strong labeling of recycling vesicle compartments with SCAMP1 antibodies (cell examples identified by asterisks in panels A-C of Fig. 3); labeling is absent from adjacent nontransfected cells (example identified by a diamond). Endocytic transferrin uptake is normal in most nontransfected cells ( Fig.  3B; 92.0 Ϯ 0.7% of untransfected cells contain endocytosed transferrin (mean Ϯ S.E)). By contrast, cells expressing fulllength SCAMP1 exhibit a partial block of endocytosis (53.7 Ϯ 2.3% of transfected cells contain endocytosed transferrin (mean Ϯ S.E); see diamonds in panels A-C). In cells expressing an N-terminally truncated form of SCAMP1 lacking the NPF repeats (SCAMP1⌬-NPF, exemplary cells identified by asterisks in panels D-F of Fig. 3), endocytic transferrin uptake is almost completely inhibited (15.2 Ϯ 2.1% of transfected cells contain endocytosed transferrin (mean Ϯ S.E.)). Again, adjacent control cells are normal. In cells expressing only the Nterminal SCAMP1 fragment containing the NPF repeats (SCAMP1 N-terminal), the SCAMP1 fragment is diffusely localized throughout the cytoplasm (panels G-I), and no effect of the overexpression of the fragment on endocytosis is observed (89.0 Ϯ 5.2% of transfected cells contain endocytosed transferrin (mean Ϯ S.E.)). DISCUSSION Clathrin coats are assembled on plasma membranes, membranes of the trans-Golgi network, and other intracellular membranes in preparation to budding of transport vesicles from these membranes. Clathrin coats are assembled in all cells, at all times, in a highly regulated and localized manner. One of the unsolved questions surrounding the assembly of clathrin coats, for example during endocytosis, is how clathrincoat assembly is nucleated. Biochemical studies showed that a proteinaceous receptor on the plasma membrane may be involved in recruiting clathrin coats to specific plasma membrane domains that are destined to become endocytic vesicles (20,22). Since the receptor is endocytosed together with the clathrin coats, it is presumably a component of recycling vesicles. An alternative hypothesis is that the nucleation point is provided by the localized synthesis of certain lipids, such as phosphatidylinositol phosphates (see for example Refs. 33 and 34). The two hypotheses are not mutually exclusive but could potentially explain different characteristics of the clathrin assembly process. The proteinaceous receptor would provide a co-localization of the coat with the proteins of recycling vesicles, whereas lipid synthesis could explain the regulated assembly of the coat.
It has been suggested that synaptotagmin serves as an AP-2 recepor in endocytosis because it binds AP-2 with very high affinity (24,25). Indeed, based on a number of experiments it seems likely that synaptotagmin functions as an endocytotic receptor at synapses (e.g. see Ref. 35). However, in non-neuronal cells, the levels of the ubiquitous synaptotagmins (e.g. synaptotagmins III, VI, VII, and IX) are so low that they are barely detectable with sensitive methods, making it unlikely that these synaptotagmins represent the long sought after receptors for endocytosis (25). The situation is different for SCAMPs. These proteins, although also abundant components of synaptic vesicles, are universally expressed at relatively high levels in all cells tested where they are components of the recycling vesicle population (2,5). Thus based on their distribution, SCAMPs perform functions that are universally shared by all recycling vesicles.
We propose that SCAMPs function in endocytosis and may perform this function by directing the assembly of clathrin FIG. 3. Effect of overexpression of full-length or truncated SCAMP1 on endocytosis. Panels on top depict immunofluorescence pictures of COS cells transfected with full-length SCAMP1 (panels A-C), an N-terminally truncated form of SCAMP1 lacking the N-terminal NPF repeats (SCAMP1⌬-NPF, panels D-F), and a soluble N-terminal SCAMP1 fragment containing the NPF repeats but not transmembrane regions (SCAMP1 N-terminus; panels G-I). Transfected COS cells were incubated with fluorescein isothiocyanate-labeled transferrin, fixed, and stained with SCAMP1 antibodies followed by rhodamine-labeled secondary antibodies. Cells were viewed in a confocal microscope, and the SCAMP and transferrin signals shown in the left and middle panels were merged into the overlay pictures displayed in the right panels to reveal co-localization. Exemplary cells expressing or lacking the transfected proteins are identified by asterisks and diamonds, respectively. The percentages of transfected and nontransfected cells that endocytosed transferrin were quantified by an observer who was unaware of the construct with which the cells had been transfected. Results of the quantification are summarized in the bar diagram below the immunofluorescence panels as mean percentage of endocytosiscompetent cells in multiple transfection experiments (ϮS.E.). Total number of cells examined (n) is also given under each bar. coats at the plasma membrane via intersectin and possibly other interacting EH domain proteins and at the trans-Golgi network via ␥-synergin. The following evidence supports this hypothesis 1) SCAMPs contain N-terminal NPF repeats that are evolutionarily conserved. Since NPF repeats bind to EH domain proteins that in turn are often involved in clathrinrelated functions, the presence of NPF repeats is indicative of a function related to the assembly and budding of clathrin coats. 2) Transfection of SCAMP1 with a deletion in the N-terminal NPF repeats inhibits endocytosis, consistent with a function in endocytosis that involves coupling of the N-terminal sequences to the transmembrane regions. 3) Two proteins with a known function in clathrin coats are specifically bound to the N-terminal NPF repeats of SCAMP1, namely intersectin and ␥-synergin. For both proteins, a specific interaction was demonstrated by yeast-two hybrid assays and pulldowns. Furthermore, the pulldowns suggested that SCAMP1 forms a complex with AP-2 (Fig. 2). With these observations, SCAMPs are the only universally distributed transmembrane proteins that interact with components of the endocytic machinery.
The large number of proteins with putative functions in clathrin-mediated endocytosis is staggering (for reviews, see Refs. 36 -38). It appears that all of these proteins (by the last count more than 20) directly or indirectly interact with each other. Strikingly, all of these proteins are cytosolic, with no structural link to the membranes on which they are supposed to act. In this regard, the identification of SCAMPs as intrinsic membrane proteins that are capable of recruiting components of the clathrin-dependent membrane budding machinery is of potential significance. These results also suggest a possible explanation for the findings that in synaptotagmin I and SCAMP1 knockout studies, no obvious defects in synaptic vesicle endocytosis were detected (7,39). Because SCAMP1 and synaptotagmin I are both abundant in synaptic vesicles, they could potentially be mutually redundant in synaptic vesicle endocytosis.
In the transfection studies, overexpression of full-length SCAMP1 only had a mild inhibitory effect while the most severe inhibition was observed with truncated SCAMP1 from which the NPF repeats had been deleted (Fig. 3). This result suggests a model whereby the N-terminal NPF repeats of SCAMPs mediate the coupling of the transmembrane regions of SCAMP to the EH domain proteins intersectin and ␥-synergin. In the absence of the NPF repeats, this coupling is disrupted, and the rest of SCAMP is a dominant negative. This model implies that the highly conserved transmembrane regions of SCAMPs perform a separate additional function in membrane budding, which is currently unknown.