J Biol Chem, Vol. 274, Issue 37, 25963-25966, September 10, 1999

COMMUNICATION
The Calcineurin-Dynamin 1 Complex as a Calcium Sensor for Synaptic Vesicle Endocytosis^*

Michael M. Lai, Jenny J. Hong, Alicia M. Ruggiero, Patrick E. Burnett, Vladimir I. Slepnev^§, Pietro De Camilli^§, and Solomon H. Snyder^¶**

From the Departments of  Neuroscience, ^¶ Pharmacology and Molecular Sciences, and  Psychiatry, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205 and the ^§ Howard Hughes Medical Institute and Department of Cell Biology, Yale University School of Medicine, New Haven, Connecticut 06510

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Exocytosis of synaptic vesicles is calcium-dependent, with synaptotagmin serving as the calcium sensor. Endocytosis of synaptic vesicles has also been postulated as a calcium-dependent process; however, an endocytic calcium sensor has not been found. We now report a physical association between the calcium-dependent phosphatase calcineurin and dynamin 1, a component of the synaptic endocytic machinery. The calcineurin-dynamin 1 interaction is calcium-dependent, with an EC₅₀ for calcium in the range of 0.1-0.4 µM. Disruption of the calcineurin-dynamin 1 interaction inhibits clathrin-mediated endocytosis. Thus, the calcium-dependent formation of the calcineurin-dynamin 1 complex, delivered to the other endocytic coat proteins, provides a calcium-sensing mechanism that facilitates endocytosis.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Neurotransmitter release occurs by calcium-dependent exocytosis of synaptic vesicles from nerve terminals mediated by a complex of proteins whose interactions are calcium-dependent (1, 2). Endocytic recycling of released synaptic vesicles is initiated at the same time as exocytosis and involves a complex of endocytic proteins including clathrin, clathrin adapters, dynamin 1, amphiphysin, and synaptojanin (3-5). Dephosphorylation of endocytic proteins is required for their assembly into a functional complex (6). A possible link between calcium and dephosphorylation is suggested by evidence that the calcium-sensitive phosphatase calcineurin (Cn)¹ can dephosphorylate endocytic proteins (7-9) and that the drugs cyclosporin A and FK506, which inhibit Cn, impair endocytosis (10, 11). However, the evidence for calcium-dependence in endocytosis has been conflicting (11-17). We now provide direct evidence that Cn is physically linked to the endocytic machinery. We show that Cn binds to dynamin 1 independent of its catalytic activity and that the Cn-dynamin 1 complex combines with amphiphysin 1, the anchor protein of the endocytic complex. Moreover, we demonstrate that the Cn-dynamin 1 interaction is calcium-dependent, allowing this complex to act as a calcium sensor. Finally, we show that disruption of the Cn-dynamin 1 interaction leads to inhibition of clathrin-mediated endocytosis.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Affinity Purification of the Cn-Dynamin 1 Complex-- Adult rat brain was homogenized in lysis buffer (50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 2 mM CaCl₂, 2 mM MgCl₂, 0.2% Triton X-100, 0.5 mM -mercaptoethanol, 5 µg/ml aprotinin, 1 µg/ml leupeptin, 6 µg/ml chymostatin, 0.7 µg/ml pepstatin, 1 mM PMSF) and centrifuged at 20,000 × g for 20 min to remove insoluble materials. The resulting lysate was divided into equal aliquots and incubated with GST-FKBP12-Sepharose beads for 2 h at 4 °C. Where indicated, GST-FKBP12-Sepharose beads were pre-absorbed with FK506 for 1 h at 4 °C and then washed twice in lysis buffer prior to adding to the cell lysate. Following the incubation, the Sepharose beads were washed three times with lysis buffer and boiled in SDS sample buffer. Eluted proteins were separated by SDS-PAGE and stained with Coomassie Blue or transferred to nitrocellulose filters for immunoblotting.

Mass Spectroscopy-- The protein band was excised, tryptically digested, and analyzed by matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) (W. M. Keck Foundation). Protein 1 was identified as rat dynamin 1 with 0.015% error and a probability score of 1.0e + 00.

Yeast Two-Hybrid Screen-- The full-length human CnA open reading frame was cloned into yeast expression vector pPC97, containing the GAL4 DNA binding domain. This was used to screen a rat hippocampal cDNA library cloned into pPC86, containing the GAL4 transactivation domain as described in (18). Positive interactions were defined as clones that are both His⁺ and -galactosidase⁺.

Co-immunoprecipitation of Cn and Dynamin 1-- Rat brain lysate, prepared as described above, was incubated with 2 µg of the indicated antibody for 2 h at 4 °C following preclearing with mouse IgG and protein G-agarose for 1 h at 4 °C. Bound proteins were washed four times with lysis buffer, eluted in SDS sample buffer, and analyzed by immunoblotting. Where indicated, 2 mM EGTA was substituted for CaCl₂.

Determination of EC₅₀ for Calcium-- Cn and dynamin 1 were co-immunoprecipitated from rat brain lysate as before, using buffers containing varying concentrations of free calcium (50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 0.2% Triton X-100, 0.5 mM -mercaptoethanol, 2.5 mM EGTA, 0-2599 µM CaCl_2, 5 µg/ml aprotinin, 1 µg/ml leupeptin, 6 µg/ml chymostatin, 0.7 µg/ml pepstatin, 1 mM PMSF). Free calcium concentrations were calculated as described (19).

Transferrin Uptake Assay-- The CnA-H101Q or CnA-wild-type cDNA was subcloned into pRK5-HA vector and transiently transfected into PC12 cells, along with empty pEGFP vector (CLONTECH), using the LipofectAMINE Plus Reagent (Stratagene). After 40 h, cells were incubated in serum-free medium for 1 h before the addition of 1 mg/ml bovine serum albumin and 20 µg/ml Texas Red-conjugated transferrin (Molecular Probes) for 25 min. After a 10-min acid incubation (50 mM glycine, pH 3.0, 100 mM NaCl) at 4 °C, the cells were washed three times in phosphate-buffered saline (with Ca²⁺ and Mg²⁺) and fixed with 4% paraformaldehyde. Transferrin uptake was quantified by measuring the fluorescent intensities of internalized Texas Red labels in transfected and untransfected cells using a digital imaging system (Zeiss). Expression of transfected Cn constructs were independently confirmed by immunostaining with CnA antibody.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Using a GST-FKBP12 resin preincubated with FK506, we sought to immobilize Cn and Cn-associated proteins from a total rat brain extract. One major associated protein was detected as a 100-kDa band (Fig. 1A, protein 1 in lane 4) and subsequently identified by mass spectroscopy as rat dynamin 1. Its identity was further confirmed by Western immunoblotting. Independently, we also detected the Cn-dynamin 1 association in a yeast two-hybrid screen (data not shown). Using full-length CnA as a bait, positive interaction was seen with a peptide fragment corresponding to the C-terminal portion of dynamin 1 (dyn1C, amino acids 618-850). To address the possibility that dynamin 1 may bind to FKBP12/FK506 independent of Cn, we expressed dyn1C as a polyhistidine-tagged fusion protein in bacteria (18) and saw that it only bound the GST-FKBP12/FK506 resin in the presence of exogenously supplied Cn (Fig. 1B), supporting a direct interaction between dynamin 1 and Cn. Utilizing the yeast two-hybrid system, we localized this interaction to the last 135 amino acids of dynamin 1 (Fig. 1C).

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Cn associates with dynamin 1. A, total rat brain extract was incubated with GST-FKBP12 resin as described. In the presence of FK506, two bound proteins were visualized by Coomassie staining following SDS-PAGE. Mass spectroscopy and Western immunoblotting identified protein 1 as dynamin 1 and protein 2 as the A subunit of calcineurin (CnA). B, bacterially produced polyhistidine-tagged fragment of dynamin 1 (his-dyn1C) corresponding to amino acids 618 to 850 was incubated with GST-FKBP12 resin (bottom panels) plus varying concentrations of FK506 in the presence or absence of purified bovine Cn. His-dyn1C only attached to the resin in the presence of Cn. Bound proteins were analyzed by SDS-PAGE and immunoblotting. C, yeast two-hybrid analysis was performed between full-length CnA and deletion mutants of dynamin 1. Positive interactions (His+, -galactosidase+) were detected with C-terminal fragments containing amino acids 715 to 850. A shorter fragment containing only the proline-rich domain (PRD) showed a weak interaction. The shaded box denotes the fragment identified in the initial two-hybrid screen. The schematic drawing of the full-length dynamin 1 protein contains the N-terminal guanosine triphosphatase (GTPase) domain, a pleckstrin homology (PH) domain, and the proline-rich domain (PRD). All experiments were repeated three times with identical results.

Because phospho-dynamin 1 is a Cn substrate (7), we wondered whether the binding we observed reflected a substrate-enzyme interaction. This seems improbable, because the Cn-dynamin 1 interaction occurs in the presence of the Cn inhibitor FK506, and bacterially produced dynamin 1, which cannot be phosphorylated, bound well to Cn. We directly demonstrated the independence of phosphorylation from the Cn-dynamin 1 interactions by showing that dynamin 1 co-immunoprecipitates with a form of Cn in which the histidine at position 101 has been mutated to glutamine, a change which abolishes the catalytic activity of Cn (Fig. 2A) (20, 21). To demonstrate a physiologic association between dynamin 1 and Cn, we immunoprecipitated endogenous Cn from rat brain and observed a co-precipitation of dynamin 1 (Fig. 2B, top). Consistent with our previous findings, the addition of FK506 did not disrupt the Cn-dynamin 1 complex. Interestingly, we observed a strict dependence of this interaction on calcium, as addition of the calcium-chelator EGTA abolished the Cn-dynamin 1 association (Fig. 2B, bottom). To further characterize the calcium-dependence of the Cn-dynamin 1 interaction, we co-immunoprecipitated Cn and dynamin 1 from rat brain in the presence of varying calcium concentrations. Maximal interaction between Cn and dynamin 1 occurs with as little as 1 µM free calcium, with an estimated EC₅₀ for calcium in the range of 0.1-0.4 µM. This value falls within the physiologic range of intracellular [Ca²⁺]; and it is consistent with the reported affinity of Cn for calcium (K_d  1 µM) (22). In similar experiments, we also found that the presence or absence of Mg²⁺ does not affect the interaction between Cn and dynamin 1 (data not shown).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   The Cn-dynamin 1 interaction is catalytically independent, FK506-independent, and calcium-dependent. A, HEK293 cells were transfected with myc-tagged dyn1C and the indicated HA-tagged construct. Immunoprecipitation using monoclonal anti-myc antibody was performed as described (21). Immunoblot analysis with anti-HA antibody on the immunoprecipitated lysate revealed a similar level of association between dynamin 1 and either wild-type CnA (wt) or catalytically inactive CnA (H101Q) (top panel). Western immunoblotting using the input lysate demonstrated equal expression of all tagged proteins (middle and bottom panels). B, rat brain lysate was immunoprecipitated using the indicated antibody and subsequently analyzed by Western immunoblotting with anti-dynamin 1 and anti-CnA antibodies. Cn and dynamin 1 co-immunoprecipitated in the presence of calcium, unaffected by the addition of FK506 (top panel). Removal of calcium by the calcium-chelator EGTA abolished the Cn-dynamin 1 complex. C, Cn was immunoprecipitated from rat brain in the presence of varying concentrations of calcium. Immunoprecipitated Cn and co-immunoprecipitated dynamin 1 were visualized by SDS-PAGE and immunoblotting. Maximal association between Cn and dynamin 1 is seen at 1 µM calcium; and the estimated EC50 for calcium is in the range of 0.1-0.4 µM. Data were from typical experiments replicated three times with the same results.

Amphiphysin 1 is the anchor protein that brings dynamin 1, clathrin, synaptojanin, and other proteins together in the endocytic complex (5). We wondered whether dynamin 1 can deliver Cn to this complex. Accordingly, we applied a brain lysate to a GST-amphiphysin 1 column and demonstrated the binding of both dynamin 1 and Cn to amphiphysin 1 (Fig. 3A). The presence of EGTA again caused Cn to dissociate from this complex of proteins, whereas the association between dynamin 1 and amphiphysin 1 appears to be calcium-independent. To test our hypothesis that dynamin 1 links Cn to the endocytic protein complex, we attempted to disrupt the dynamin 1-amphiphysin 1 interaction by the addition of a peptide corresponding to the SH3 domain of amphiphysin 1 (amphSH3, amino acids 588-695) to our binding experiment. Because the SH3 domain of amphiphysin 1 mediates its binding to dynamin 1 (23), the addition of amphSH3 peptide caused both dynamin 1 and Cn to dissociate from the GST-amphiphysin 1 column (Fig. 3B), confirming that dynamin 1 provides a physical linkage between Cn and the endocytic protein complex. This experiment also showed that no direct interaction between Cn and amphiphysin 1 exists. Recent studies provided evidence that microinjection of the amphSH3 peptide inhibited synaptic vesicle endocytosis in living nerve terminals (24).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Dynamin 1 targets Cn to the endocytic coat complex. A, rat brain extract incubated with GST-amphiphysin 1 resin was analyzed as before. Western immunoblotting with anti-dynamin 1 and anti-CnA antibodies demonstrated the presence of both dynamin 1 and Cn in the complex of proteins bound to amphiphysin 1 (middle panel). Addition of EGTA caused the dissociation of Cn from this complex, but the binding of dynamin 1 to amphiphysin 1 remained unaffected (right panel). B, the addition of a peptide corresponding to the SH3 domain from amphiphysin 1 (amphSH3) caused a dose-dependent dissociation of dynamin 1 from the GST-amphiphysin 1 resin (top panel). The concomitant loss of Cn binding indicates that dynamin provides the physical linkage between Cn and the rest of the endocytic coat complex. Ponceau staining of the nitrocellulose filter shows the amount of GST-amphiphysin 1 in each condition (bottom panel). Data were from typical experiments replicated twice with the same results.

To assess the role of Cn in clathrin-mediated endocytosis, we disrupted the Cn-dynamin 1 interaction in vivo and assessed the endocytic uptake of transferrin molecules. Because CnA-H1010Q binds dynamin 1 but is devoid of catalytic activity (Fig. 2A), we reasoned that it could act as a dominant-negative mutant when overexpressed in PC12 cells, which contain both endogenous Cn and dynamin 1. Indeed, we observed that overexpression of CnA-H101Q caused complete inhibition of transferrin endocytosis in certain transfected cells (Fig. 4, A-C). Overall, cells overexpressing the CnA-H101Q mutant showed a decrease of approximately 35% in transferrin uptake (n = 46, p < 0.005, paired t test) (Fig. 4G). Overexpression of the wild-type CnA subunit did not alter the endocytosis of transferrin (Fig. 4, D-G).

View larger version (49K): [in this window] [in a new window]   Fig. 4.   Disruption of the Cn-dynamin 1 interaction leads to inhibition of clathrin-mediated endocytosis. PC12 cells were transiently transfected with either the catalytically inactive CnA (H101Q) subunit (A-C) or the wild-type CnA subunit (D-F) and assayed for their ability to endocytose transferrin as described. Transfected cells were identified by the presence of the Green Fluorescence Protein (GFP) (B and E), while internalized transferrin was visualized through its Texas Red conjugate (C and F). Expression of CnA-H101Q, but not wild-type CnA, inhibited transferrin uptake (transfected cells indicated by arrows). Pictures were taken from an experiment that was repeated three times with essentially identical results. Quantitation of a typical experiment showed that expression of CnA-H101Q caused an approximately 35% decrease in transferrin uptake (n = 46, p < 0.005, paired t test) (G).

The protein interactions we have described here are consistent with a model in which the interaction between Cn and dynamin 1 serves as a calcium sensor of endocytosis (Fig. 5). Under basal conditions, Cn and dynamin 1 are largely unassociated. A rise in intracellular calcium concentration, produced by either neuronal depolarization or extracellular stimuli, causes the association between Cn and dynamin 1, as well as the dephosphorylation of dynamin 1. The Cn-dynamin 1 complex binds to amphiphysin 1, and the delivery by dynamin 1 of Cn to the endocytic protein complex provides the proximity enabling Cn to initiate the dephosphorylation of other proteins in the complex. Consistent with previous studies, Cn, and possibly other phosphatases, allows the endocytic coat complex to remain assembled and functional.

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Proposed model of the role of Cn-dynamin 1 complex in assembly of endocytic coat complex. Please refer to the text for a discussion of the proposed model. adp, adaptin; amph1, amphiphysin 1; amph2, amphiphysin 2; Cn, calcineurin; cla, clathrin; dyn1, dynamin 1; syj, synaptojanin 1

Calcium is the crucial facilitator of both exocytosis and endocytosis of synaptic vesicles. Synaptotagmin has been implicated as the calcium sensor for exocytosis (2, 25, 26), whereas the nature of the calcium sensor for endocytosis has been unclear. Our findings provide a direct demonstration that Cn is a key component of the endocytic complex and is brought into the complex via its binding to dynamin 1. In addition, the calcium-dependent nature of the Cn-dynamin 1 association, which occurs in the presence of sub-micromolar calcium concentrations, provides a possible mechanism for sensing alterations in intracellular calcium concentration during endocytosis. Elucidating the precise molecular nature of the Cn-dynamin 1 interaction may provide valuable insights into the calcium-sensing mechanism of clathrin-mediated endocytosis.

	ACKNOWLEDGEMENTS

We thank Keqiang Ye for technical assistance, Herman Wolosker for help with determination of EC₅₀ for calcium, Jack Roos for suggestions, and Levante Egry for advice on the transferrin assay.

	FOOTNOTES

^* This work was supported by United States Public Health Service Grant MH-18501 and Research Scientist Award DA-00074 (to S. H. S.), National Institutes of Health Training Grant GM-07309 (to M. M. L. and P. E. B.), and National Institutes of Health Grant NS36251 (to V. I. S. and P. D. C.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^** To whom correspondence should be addressed: Departments of Neuroscience, Pharmacology and Molecular Sciences, and Psychiatry, Johns Hopkins University School of Medicine, 725 N. Wolfe St., Baltimore, MD 21205. Tel.: 410-955-3024; Fax: 410-614-6249; E-mail: ssnyder@bs.jhmi.edu.

	ABBREVIATIONS

The abbreviations used are: Cn, calcineurin; PMSF, phenylmethylsulfonyl fluoride; GST, glutathione S-transferase; PAGE, polyacrylamide gel electrophoresis; dyn1C, C-terminal portion of dynamin 1; amphSH3, SH3 domain of amphiphysin 1; HA, hemagglutinin.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES