Protein phosphorylation is required for ... - Wiley Online Library

Journal of Neurochemistry, 2001, 76, 105±116

Protein phosphorylation is required for endocytosis in nerve terminals: potential role for the dephosphins dynamin I and synaptojanin, but not AP180 or amphiphysin Michael A. Cousin,1 Timothy C. Tan and Phillip J. Robinson Cell Signalling Unit, Children's Medical Research Institute, Sydney, New South Wales, Australia

Abstract Dynamin I and at least ®ve other nerve terminal proteins, amphiphysins I and II, synaptojanin, epsin and eps15 (collectively called dephosphins), are coordinately dephosphorylated by calcineurin during endocytosis of synaptic vesicles. Here we have identi®ed a new dephosphin, the essential endocytic protein AP180. Blocking dephosphorylation of the dephosphins is known to inhibit endocytosis, but the role of phosphorylation has not been determined. We show that the protein kinase C (PKC) antagonists Ro 31-8220 and Go 7874 block the rephosphorylation of dynamin I and synaptojanin that occurs during recovery from an initial depolarizing stimulus (S1). The rephosphorylation of AP180 and amphiphysins 1 and 2, however, were unaffected by

Ro 31-8220. Although these dephosphins share a single phosphatase, different protein kinases phosphorylated them after nerve terminal stimulation. The inhibitors were used to selectively examine the role of dynamin I and/or synaptojanin phosphorylation in endocytosis. Ro 31-8220 and Go 7874 did not block the initial S1 cycle of endocytosis, but strongly inhibited endocytosis following a second stimulus (S2). Therefore, phosphorylation of a subset of dephosphins, which includes dynamin I and synaptojanin, is required for the next round of stimulated synaptic vesicle retrieval. Keywords: AP180, dynamin I, endocytosis, nerve terminal, phosphorylation, synaptojanin. J. Neurochem. (2001) 76, 105±116.

Phosphorylation and dephosphorylation of nerve terminal proteins by various protein kinases has been proposed to regulate synaptic vesicle recycling (Robinson et al. 1994; Turner et al. 1999). Of particular interest are the set of proteins that are phosphorylated in resting nerve terminals and that are coordinately dephosphorylated upon stimulation. These `dephosphins' include dynamin I, amphiphysins 1 and 2, synaptojanin, epsin and eps15 (Liu et al. 1994a; Bauerfeind et al. 1997; Chen et al. 1999). All have been directly implicated as essential for endocytosis (Robinson et al. 1993b; van der Bliek et al. 1993; Takei et al. 1995; McPherson et al. 1996; Shupliakov et al. 1997; Chen et al. 1998; Cremona et al. 1999). The ®rst characterized dephosphin was the GTPase dynamin I. Dynamin I is phosphorylated by protein kinase C (PKC) in vitro and probably by PKC in intact nerve terminals, whereas the protein kinase that phosphorylates the other dephosphins is still unknown (Robinson 1992; Robinson et al. 1993b; Liu et al. 1994b; Sontag et al. 1994). Dynamin I, amphiphysins 1 and 2 and synaptojanin are all dephosphorylated by the Ca21-dependent

phosphatase calcineurin in vitro and in nerve terminals (Liu et al. 1994a; Nichols et al. 1994; Bauerfeind et al. 1997; Marks and McMahon 1998). After dephosphorylation, dynamin I is fully rephosphorylated within 2 min (Robinson 1992; Robinson et al. 1994). Rephosphorylation can be inhibited with a range of low-speci®city PKC antagonists, suggesting that PKC is the dynamin I kinase in nerve terminals (Robinson 1992). Received April 26, 2000; revised manuscript received July 27, 2000; accepted August 11, 2000. Address correspondence and reprint requests to Phillip J. Robinson, Children's Medical Research Institute, Cell Signalling Unit, Locked Bag 23, Wentworthville, Sydney, NSW 2145, Australia. E-mail: [email protected] 1 Present address: Membrane Biology Group, Department of Biomedical Sciences, Teviot Place, University of Edinburgh, Edinburgh EH8 9AG, UK. Abbreviations used: MARCKS, myristoylated alanine-rich C kinase substrate; PKC, protein kinase C; PMA, phorbol 12-myristate 13-acetate.

q 2001 International Society for Neurochemistry, Journal of Neurochemistry, 76, 105±116

105

106 M. A. Cousin et al.

The highly unusual coordinated dephosphorylation of so many of the proteins that are essential for endocytosis in nerve terminals raises the possibility that other phosphoproteins essential for endocytosis may also be dephosphins. One such candidate is the clathrin assembly protein AP180. AP180 promotes the assembly of clathrin cages in vitro (Zhou et al. 1992; Ye and Lafer 1995), and disruption of this function inhibits synaptic transmission, increases plasma membrane surface area, reduces synaptic vesicle number and increases the size of the remaining synaptic vesicles in mammalian and invertebrate cells, but not in yeast (Zhang et al. 1998; Huang et al. 1999; Morgan et al. 1999; Nonet et al. 1999). Thus AP180 is required for endocytosis in nerve terminals (Morgan et al. 1999). Because it is also phosphorylated in resting nerve terminals (Zhou et al. 1992; Murphy et al. 1994) it has the potential to be another member of the dephosphin family. In this study we demonstrate that AP180 is a dephosphin. Phosphorylation has been proposed to control interactions between the proteins involved in endocytosis (Slepnev et al. 1998; Chen et al. 1999). For example, in vitro dephosphorylation of rat brain extracts apparently promotes the assembly of dynamin I, synaptojanin, amphiphysin, clathrin and AP-2 into complexes (Slepnev et al. 1998). Other studies have argued against a role for phosphorylation in regulation of vesicle recycling. For example, the GTPase activity of dynamin I, which is essential for synaptic vesicle ®ssion during endocytosis (Takei et al. 1995; Sweitzer and Hinshaw 1998; Stowell et al. 1999), is increased after phosphorylation by PKC in vitro (Robinson et al. 1993b). However, PKC cannot stimulate GTPase activity when dynamin I is at the neck of a budding vesicle, since dynamin I cannot be phosphorylated when associated with either the plasma membrane or the cytoskeleton (Liu et al. 1994b). Additionally, the kinetics of dynamin I phosphorylation in nerve terminals are too slow to be the trigger for the GTPase activity required for endocytosis (Robinson 1992; Robinson et al. 1994; Ryan 1996). Our aim was to determine whether the phosphorylation of dynamin I and the other dephosphins during recovery from a nerve terminal stimulation (which we refer to as `rephosphorylation') is essential for endocytosis. Stimulus-dependent protein phosphorylation is normally investigated with protein kinase inhibitors being applied shortly before the stimulus. No functional role for phosphorylation in endocytosis has been demonstrated using this approach in nerve terminals (Kraszewski et al. 1996; Klingauf et al. 1998; Cousin et al. 1999). This may have been because the resting level of phosphorylation of the dephosphins was already high when the antagonists were applied. Proteins such as the dephosphins that undergo stimulus-dependent dephosphorylation require a different pharmacological approach. Protein kinase inhibitors should exert their functional effects only after recovery from the ®rst depolarization stimulus (S1) and

application of a second stimulus (S2). Therefore, we used ¯uorescence-based assays in isolated nerve terminals of both exocytosis and endocytosis throughout three cycles of stimulus and recovery, with parallel monitoring of the phosphorylation status of the dephosphins. We demonstrate that the rephosphorylation of dynamin I and synaptojanin is blocked by PKC inhibitors, while the rephosphorylation of AP180 and amphiphysins 1 and 2 is not blocked. Under these conditions, endocytosis is normal after S1 but is arrested after S2. This demonstrates an obligatory role for the speci®c phosphorylation of a subset of the dephosphins that includes dynamin I and/or synaptojanin in the endocytosis of synaptic vesicles. Materials and methods Materials FM2-10 was purchased from Molecular Probes (Oregon, USA); Ro 31-8220, Go 7874 and cyclosporin A were purchased from Calbiochem/Novabiochem (Alexandria, Australia); and 32P-orthophosphate was purchased from NEN-Dupont (Sydney, Australia). Polyclonal antibodies against amphiphysins 1 and 2 were provided by Harvey McMahon (Cambridge, UK). The amphiphysin I GST± SH3 domain construct was provided by Pietro De Camilli (Yale, USA). Glutathione-sepharose beads and protein G-sepharose beads were from Amersham-Pharmacia Biotech (Sydney, Australia). All other reagents were obtained from Sigma Chemical Company (Poole, UK) or were of at least analytical reagent grade. Glutamate release assay Synaptosomes were prepared from rat cerebral cortex by centrifugation on discontinuous percoll gradients (Dunkley et al. 1986). The glutamate release assay was performed using enzyme-linked ¯uorescent detection of released glutamate (Nicholls and Sihra 1986; Cousin and Robinson 1998). Brie¯y, synaptosomes (0.6 mg in 2 mL) were resuspended in either plus (1.2 mm CaCl2) or minus (1 mm EGTA) Ca21 Krebs-like solution (118.5 mm NaCl, 4.7 mm KCl, 1.18 mm MgCl2, 0.1 mm K2HPO4, 20 mm Hepes, 10 mm glucose, pH 7.4) at 378C. For the ®rst stimulation (S1), synaptosomes were stimulated with 30 mm KCl after addition of 1 mm NADP1 and 50 U of glutamate dehydrogenase. For the second stimulation (S2), synaptosomes were stimulated with 30 mm KCl for 10 s in plus Ca21 solution then resuspended in either plus or minus Ca21 solution and stimulated identically to S1 stimulation. Increases in ¯uorescence due to production of NADPH were monitored in a Perkin-Elmer LS-50B spectro¯uorimeter at 340 nm excitation and 460 nm emission. Experiments were standardized by the addition of 4 nmol of glutamate. Data are presented as Ca21dependent glutamate release which is calculated as the difference between release in plus and minus Ca21 solutions. In experiments with inhibitors, synaptosomes were pre-incubated for 15 min with Ro 31-8220 and Go 7874 and for 5 min with cyclosporin A before stimulation. Endocytosis assay Endocytosis was measured using uptake of the ¯uorescent dye FM2-10, as previously described (Cousin and Robinson 1998). Synaptosomes (0.6 mg of protein in 2 mL) were incubated for


Phosphorylation is required for endocytosis 107

5 min at 378C in plus or minus Ca21 Krebs-like solution. Two related methods were used to load synaptosomes with FM2-10, which we refer to as S1/S2 or S2/S3 endocytosis (Fig. 1). For S1/S2 endocytosis, FM2-10 (100 mm) was added 1 min before stimulation with 30 mm KCl. After 2 min of stimulation to load FM2-10, synaptosomes were washed twice in plus Ca21 solution containing 1 mg/mL bovine serum albumin to remove non-internalized FM210 and the antagonists. Synaptosomes were resuspended in plus Ca21 solution at 378C and stimulated with a standard addition of 30 mm KCl. Endocytosis was measured as the loss of FM2-10 ¯uorescence (excitation 488 nm, emission 540 nm) during S2, since the standard KCl stimulation releases FM2-10 previously accumulated by endocytosis. For S2/S3 endocytosis, synaptosomes were stimulated for 10 s with 30 mm KCl in plus Ca21 solution (S1) before being resuspended in either plus or minus Ca21 solution. S2 loading of FM2-10 then proceeded as described for S1 loading, with dye release being measured at S3. Endocytosis was calculated as the total decrease in ¯uorescence stimulated by the standard KCl unloading as a percentage of total FM2-10 ¯uorescence. All displayed traces are representative and normalized to an arbitrary ¯uorescence value. No agent affected the ability of a standard pulse of KCl to evoke release of FM2-10-labelled vesicles, since Ca21-dependent glutamate release was unaffected (data not shown). Retrieval ef®ciency (Cousin and Robinson 1998; 2000) was calculated as endocytosis/exocytosis, where endocytosis is de®ned as above and exocytosis as Ca21-dependent glutamate release after 2 min of stimulation.

et al. 1993b) using 2 mL of polyclonal antibody Ra 1.2 (Marks and McMahon 1998) for each sample. Dynamin I and synaptojanin were af®nity puri®ed using the GST-amphiphysin I SH3 domain bound to glutathione-sepharose as described previously (Marks and McMahon 1998). Puri®cation and phosphorylation of dynamin I and AP180 Dynamin I was puri®ed from sheep brain as described previously (Robinson 1992; Robinson et al. 1993b). AP180 was copuri®ed with dynamin during the same procedure (see Results section). Phosphorylation of AP180 and dynamin I by PKC was performed as described previously (Robinson et al. 1993b).

Results

Phosphorylation status of nerve terminal proteins Labelling of intact synaptosomes with 32Pi, SDS polyacrylamide gel electrophoresis and autoradiography were as previously described (Robinson and Dunkley 1983; Robinson et al. 1993b). Acid-soluble nerve terminal proteins were extracted from 32Plabelled synaptosomes as described previously (Robinson et al. 1993a), with the exception that 2% SDS was used instead of 0.5% SDS to lyse the synaptosomes in order to improve recovery of the MARCKS protein and AP180. Immunoprecipitation of amphiphysins 1 and 2 was performed as described previously (Robinson

Distinct protein kinases phosphorylate the dephosphins Nerve terminal proteins undergo cycles of phosphorylation and dephosphorylation during cycles of stimulation and recovery. Thus, dynamin I and other dephosphins are primarily phosphorylated at rest, dephosphorylated upon stimulation and then rephosphorylated upon termination of the stimulus (Fig. 1). Other events occur within the nerve terminal with stimulation, such as Ca21 entry via voltagedependent Ca21 channels, fusion of synaptic vesicles to release neurotransmitter and the subsequent retrieval of vesicles by endocytosis. The rephosphorylation of dynamin I after dephosphorylation in nerve terminals was previously shown to be mediated by PKC, by the use of low-speci®city PKC inhibitors (Robinson 1992). However, the role of PKC in the phosphorylation of other dephosphins has not yet been examined. This is particularly relevant because most, if not all, dephosphins share the same phosphatase, calcineurin. The PKC antagonists Ro 318220 and Go 7874 were used, as they have a higher speci®cities for PKC. In the absence of antagonists, addition of

Fig. 1 Protocol for endocytosis assays. Isolated nerve terminals were stimulated for 2 min with 30 mM KCl up to three times (S1, S2 and S3). Upon stimulation with KCl dynamin I is dephosphorylated and then rephosphorylated upon removal of the stimulus (solid line: phospho-dyn). In the presence of PKC antagonists dynamin I cannot be rephosphorylated (Robinson 1992) and remains dephosphorylated

(dashed line: dephospho-dyn). Endocytosis was monitored by uptake of FM2-10 in the presence of antagonists of either phosphorylation or dephosphorylation. Endocytosis was sampled either at S1 or S2 and the extent of FM2-10 loading was assayed by unloading the dye at S2 and S3, respectively. These are termed S1/S2 and S2/S3 endocytosis.



Fig. 2 Ro 31-8220 and Go 7874 block dynamin I rephosphorylation in nerve terminals. Nerve terminals were labelled with 32Pi and then lysed either before stimulation (lanes 1±2, 9±10), after stimulation with 41 mM KCl for 10 s (S1, lanes 3±4, 11±12), 7 min after repolarization (lanes 5±6, 13±14) or 10 s after a second stimulation (S2) with 41 mM KCl (lanes 7±8, 15±16). Nerve terminals were pre-incubated for 15 min in the absence (lanes 1±8) or presence (lanes 9± 16) of either (a) 10 mM Ro 31-8220 or (b) 10 mM Go 7874. Samples were separated by SDS-PAGE and an autoradiograph is presented. Arrows indicate either dynamin I (Dyn) or synapsin I (Syn). Note that depolarization-stimulated synapsin I phosphorylation is largely unaffected by Ro 31-8220 and Go 7874. Results are in duplicate and are representative of at least three independent experiments. (c) Nerve terminals were prepared and labelled in an identical manner to that in (a) and (b), but after lysis dynamin I and synaptojanin were af®nity puri®ed using GST-amphiphysin I SH3 domain bound to glutathione sepharose. (d) Coomassie Blue-stained gel from the experiment shown in (c). Arrows indicate either dynamin I (Dyn) or synaptojanin I (SJ). Data are in duplicate and are representative of at least four independent experiments.

KCl (S1) stimulates dynamin I dephosphorylation and after the stimulus is removed dynamin I is rephosphorylated (Fig. 2). Subsequent stimulation (S2) evokes a second dephosphorylation of dynamin I. Pre-incubation with 10 mm Ro 31-8220 did not affect dynamin I dephosphorylation (Fig. 2a). However, Ro 31-8220 abolished the rephosphorylation of dynamin I after the removal of the S1

stimulus (Fig. 2a, lanes 13±14). Ro 31-8220 had little effect on the stimulus-dependent phosphorylation of synapsin I by Ca21-calmodulin-dependent protein kinase II at either S1 or S2, indicating some speci®city in the drug action. Go 7874 (10 mm) also inhibited dynamin I rephosphorylation, but not completely (Fig. 2b, lanes 13±14). These results support previous suggestions that PKC may be the protein kinase that phosphorylates dynamin I in nerve terminals. Because all the dephosphins share the same phosphatase, the role of PKC in the phosphorylation of other dephosphins was examined. To clearly resolve the effect of Ro 31-8220 on their phosphorylation status, dynamin I and synaptojanin were af®nity puri®ed from intact 32P-labelled synaptosomes using a GST fusion protein containing the SH3 domain of amphiphysin 1 bound to glutathione-sepharose (Marks and McMahon 1998). Both dynamin I and synaptojanin were dephosphorylated at S1 and rephosphorylated after removal of the stimulus in the absence of Ro 31-8220 (Fig. 2c). Upon treatment with 10 mm Ro 31-8220, the rephosphorylation of both synaptojanin and dynamin I was abolished (Fig. 2c). An equal amount of protein was recovered in all conditions (Fig. 2d). Thus, dynamin I and synaptojanin may share the same protein kinase. However, synaptojanin has not yet been shown to play a direct role in endocytosis, apart from a colocalization with clathrin pits and with dynamin I (McPherson et al. 1994; Haffner et al. 1997). Next, amphiphysins 1 and 2 were immunoprecipitated from intact 32 P-labelled synaptosomes. Both were dephosphorylated on stimulation and rephosphorylated after removal of the stimulus in the absence of Ro 31-8220 (Fig. 3). Their phosphorylation was unaffected by the presence of 10 mm Ro 31-8220 (Fig. 3). Thus, amphiphysins 1 and 2 are phosphorylated by a different protein kinase from dynamin I and synaptojanin. AP180 is a nerve terminal phosphoprotein that is required for endocytosis in nerve terminals (Zhang et al. 1998; Huang et al. 1999; Morgan et al. 1999). AP180 is known to be a highly acidic protein (isoelectric point 3.85) and like other acidic proteins or proline-rich proteins it exhibits anomalous migration on SDS gels (Zhou et al. 1992). AP180 is also known to be a phosphoprotein in intact synaptosomes (Zhou et al. 1992). Two other synaptic phosphoproteins, myristoylated alanine-rich C kinase substrate (MARCKS) and B50, are also both acidic and exhibit anomalous migration on SDS gels. These proteins are both heat and acid stable, and can rapidly be extracted from 32Plabelled nerve terminals on the basis of their solubility in acetic acid (Robinson et al. 1993a). AP180 can also be extracted from nerve terminals using acetic acid, by increasing the concentration of SDS in the extraction buffer fourfold from that previously reported for MARCKS extraction (Robinson et al. 1993a) (Fig. 4). The acid-soluble 180-kDa band was con®rmed to be AP180, because two tryptic fragments of this protein contained the amino acid



Fig. 3 Ro 31-8220 does not block rephosphorylation of amphiphysins 1 and 2. Immunoprecipitation of amphiphysins 1 and 2 from 32 P-labelled nerve terminals using polyclonal antibodies. Amphiphysins 1 and 2 were immunoprecipitated from nerve terminals either before stimulation, after stimulation with 41 mM KCl for 10 s (S1), 7 min after repolarization from stimulation or 10 s after a second stimulation (S2) with 41 mM KCl, and an autoradiograph is displayed. Nerve terminals were pre-incubated for 15 min in the presence or absence of 10 mM Ro 31-8220. Arrows indicate amphiphysins 1 and 2 (Amph1 and Amph2).

sequences VAEQVGIDK and GASPVPESSLTA, which were identical to that of rat AP180. We then investigated whether AP180 could be another member of the dephosphin family. When nerve terminals were depolarized AP180 was dephosphorylated and then rephosphorylated on removal of the stimulus, in the same manner as the other dephosphins (Fig. 4a). The dephosphorylation of AP180 was inhibited by the calcineurin antagonist cyclosporin A (Fig. 4b), indicating that it is also a calcineurin substrate. Both AP180 dephosphorylation and rephosphorylation in nerve terminals were insensitive to 10 mm Ro 31-8220 (Fig. 4a, lanes 13± 14). Ro 31-8220 abolished the stimulated phosphorylation of the PKC substrate MARCKS, con®rming that the drug was used at its optimal concentration in this study (Fig. 4a). Therefore, different protein kinases rephosphorylate the dephosphins. Dynamin I and synaptojanin are rephosphorylated by an Ro 31-8220-sensitive kinase, whereas AP180 and amphiphysins 1 and 2 are rephosphorylated by a distinct Ro 31-8220-insensitive protein kinase(s). To further investigate the protein kinase that phosphorylates AP180, we performed in vitro phosphorylation studies using PKC. During the puri®cation of dynamin I from brain (Robinson 1992; Robinson et al. 1993b), a protein of 180 kDa was noted to copurify through all the steps. This protein was ®nally separated from dynamin I by batch chromatography on phenyl-sepharose (Fig. 5a). Two tryptic fragments derived from this protein contained the amino acid sequences VAEQVGIDK and PGNNEGSGAPSPLSK, which were identical to that of AP180. To determine whether AP180 might be an in vitro substrate for PKC we attempted to phosphorylate it using PKC (Fig. 5b). AP180 was not detectably phosphorylated, while dynamin I, a minor contaminant, was phosphorylated. As their copuri®cation suggested that the two proteins might associate, further dynamin I was added to the AP180 fraction, but did not affect phosphorylation of AP180 by PKC. Therefore, native AP180 is not an in vitro PKC

Fig. 4 AP180 is acid soluble and dephosphorylated by calcineurin in nerve terminals. (a) Extraction of acid soluble proteins from 32 P-labelled nerve terminals was performed either before (lanes 1± 2, 9±10) or after stimulation with 41 mM KCl for 10 s (S1, lanes 3±4, 11±12), 7 min after repolarization (lanes 5±6, 13±14) or 10 s after a second stimulation (S2) with 41 mM KCl (lanes 7±8, 15±16). Ro 318220 (10 mM) was added 15 min prior to S1 (lanes 9±16). (b) Synaptosomes were pre-incubated with either 10 or 30 mM cyclosporin A (CysA) for 5 min before stimulation. AP180 and MARCKS were extracted either before (lanes 1±2 and 5±6) or after stimulation (S1, lanes 3±4 and 7±8) with 41 mM KCl for 10 s. Data are presented in duplicate and (a) is representative of four independent experiments.

substrate, as recently reported for recombinant AP180 (Hao et al. 1999). Rephosphorylation of dynamin I and/or synaptojanin is essential for endocytosis More than one protein kinase phosphorylates the dephosphins. Therefore it was now possible to selectively determine the role of phosphorylation in endocytosis of a speci®c subset of dephosphins, which includes dynamin I and synaptojanin. To determine whether their phosphorylation was essential for endocytosis, we monitored KCl-evoked exocytosis and endocytosis at two de®ned points. Endocytosis was measured ®rstly at S1, after inhibiting either dephosphorylation or phosphorylation (Fig. 1). This method is called S1/S2 endocytosis, because uptake of the ¯uorescent dye FM2-10 occurs during S1 and it is unloaded at S2. Secondly, rephosphorylation was inhibited by keeping the drug present throughout recovery and S2. The role of phosphorylation in endocytosis was then examined at S2, and this method is termed S2/S3 endocytosis. Dynamin I, synaptojanin, amphiphysins 1 and 2 (Liu et al. 1994a; Bauerfeind et al. 1997; Marks and McMahon 1998)



and AP180 are all dephosphorylated by calcineurin on nerve terminal stimulation. We examined the role of calcineurin in endocytosis by pre-incubating nerve terminals with the calcineurin antagonist cyclosporin A (40 mm). Cyclosporin A had little effect on Ca21-dependent glutamate release (reduced to 87.6 ^ 1.9% of control, Fig. 6a), but signi®cantly inhibited endocytosis (to 37.6 ^ 2.9% of control, Fig. 6b). To quantify endocytosis accurately, the preceding amount of exocytosis must also be considered, because retrieval of synaptic vesicles is partly dependent on the availability of fused vesicles. Measurements of endocytosis alone would therefore be affected by any agent that altered exocytosis. A parameter termed retrieval ef®ciency (Cousin and Robinson 1998, 2000) accounts for this, by dividing the amount of endocytosis by exocytosis for the same stimulus. If endocytosis was unaffected by cyclosporin A, retrieval ef®ciency should equal one (an arbitrary constant). Cyclosporin A reduced retrieval ef®ciency to 0.43 ^ 0.01 (Fig. 6c), demonstrating that calcineurin-mediated dephosphorylation is essential for endocytosis in nerve terminals (Marks and McMahon 1998). While dephosphorylation is clearly the trigger for endocytosis, this approach does not reveal which of the seven dephosphins is essential. Since Ro 31-8220 selectively abolished dynamin I and synaptojanin rephosphorylation, we examined whether the antagonist had an effect on either S1/S2 or S2/S3 endocytosis. Ro 31-8220 (10 mm) had no effect on S1 Ca21-dependent glutamate release (99.7 ^ 2.8% of control, Fig. 7a), as previously observed (Coffey et al. 1993). Go 7874 (10 mm) signi®cantly blocked glutamate release (47.1 ^ 6.0% of control, Fig. 7a). Both Ro 31-8220 and Go 7874 inhibited FM2-10 uptake by the S1/S2 method to differing extents (reduced to 78.3 ^ 0.4% of control for Ro 31-8220 and 46.1 ^ 4.1% for Go 7874, Fig. 7b). However, when retrieval ef®ciency was determined, the inhibition of endocytosis was small, with Go 7874 not blocking endocytosis and Ro 31-8220 blocking to 80% of control (Fig. 7c). Therefore stimulus-induced phosphorylation mediated by the protein kinase(s) sensitive to Ro 318220 or Go 7874 has a minor or no role in the immediate control of endocytosis. We next examined whether blocking rephosphorylation of dynamin I affected endocytosis at S2. Ro 31-8220 did not inhibit Ca21-dependent glutamate release at S2 (122.1 ^ 0.2% of control), while Go 7874 had a signi®cant inhibitory effect (49.6 ^ 0.2% of control, Fig. 8a). A much larger inhibition of S2/S3 endocytosis was observed for both Ro 31-8220 and Go 7874 than with S1/S2 stimulation (reduced to 37.3 ^ 3.9% and 22.0 ^ 3.0% of control, respectively, Fig. 8b). Retrieval ef®ciency also revealed that there was a large inhibition of endocytosis with both antagonists (Fig. 8c). The reduction in S2/S3 endocytosis was not due to a block of exocytosis during S3 by Ro 318220 or Go 7874, as glutamate release at S3 was not further

Fig. 5 AP180 is not a substrate for PKC. (a) Dynamin I was puri®ed from sheep brain by chromatography on Q-sepharose, S-sepharose (lanes 1±8) and phenyl-sepharose (lanes 10±15). A Coomassie Blue-stained gel shows that AP180 copuri®es with dynamin I until it is separated by the phenyl-sepharose step. Every second fraction from 29 to 41 from the S-sepharose column (from 60 fractions) is shown (lanes 1±8). AP180 was eluted from the phenyl-sepharose column by batch elution when the NaCl concentration was reduced to zero (lanes 10±12), and dynamin was eluted by elevated pH (8.8) in the presence of 1% tween 80 (lanes 13±15). Standards are shown in lane 9 (200, 116, 94 and 67 kDa). (b) AP180 is not an in vitro substrate for PKC. AP180 (0.1 mg, from lane 10 in panel a) was incubated with [g-32P]-ATP in the absence (lane 1) or presence (lane 2) of 20 ng of PKC. Dynamin I was added to the AP180 (0.4 mg, from lane 13 in panel a) and again incubated with [g-32P]ATP in the absence (lane 3) or presence of PKC (lane 4). A representative autoradiograph is shown (n 2).

affected after removal of either Ro 31-8220 or Go 7874 (data not shown). These results demonstrate that phosphorylation of a speci®c subset of nerve terminal proteins by an Ro 31-8220-sensitive kinase is required for endocytosis. Application of phorbol esters activates PKC in nerve terminals, but does not stimulate the phosphorylation of dynamin I (Robinson 1992). Therefore, we examined whether activation as well as inhibition of PKC could modulate endocytosis. Pre-incubation with 100 nm phorbol 12-myristate 13-acetate (PMA) had little effect on KClevoked Ca21-dependent glutamate release either in the S1 (90.1 ^ 2.0% of control) or S2 phase (87.5 ^ 0.4% of control) of stimulation (Figs 9a and b). Endocytosis was slightly affected in both the S1/S2 and the S2/S3 methods (89.6 ^ 3.0% and 81.6 ^ 0.8% of control, respectively, Figs 9c and d). Retrieval ef®ciency was unaffected (Fig. 9e). Therefore, activation of PKC by phorbol esters does not modulate endocytosis when the dephosphins are either already phosphorylated (S1) or fully rephosphorylated



Fig. 6 Cyclosporin A inhibits endocytosis in nerve terminals. (a) Ca21-dependent glutamate release evoked by 30 mM KCl is displayed either in the presence or absence of 40 mM cyclosporin A (CysA). Synaptosomes were pre-incubated with CysA 5 min before stimulation (n 3 ^ SEM). (b) CysA blocks endocytosis (S1/S2 method, Fig. 1). Synaptosomes were loaded with FM2-10 by stimulation with 30 mM KCl either in the presence or absence of Ca21. Synaptosomes

were pre-incubated for 5 min with CysA. The subsequent release of loaded FM2-10 by a standard stimulus of 30 mM KCl in the presence of Ca21 is displayed after subtraction of traces loaded in the absence of Ca21. Traces are representative and normalized to an arbitrary value. (c) The effect of CysA on retrieval ef®ciency, which is endocytosis/exocytosis (n 3).

(S2). This is consistent with the lack of effect of phorbol esters on dynamin I phosphorylation in vitro or in intact synaptosomes (Robinson 1992), and that primarily dynamin I and synaptojanin appear to be phosphorylated by an Ro 318220-sensitive protein kinase in nerve terminals. Alternatively, it indicates that the pool of phosphorylated dynamin I and/or synaptojanin required for endocytosis is already optimal.

The dephosphins are a set of neuronal phosphoproteins that are coordinately dephosphorylated on nerve terminal

depolarization by the Ca21-dependent phosphatase calcineurin (Liu et al. 1994a; Nichols et al. 1994; Bauerfeind et al. 1997; Marks and McMahon 1998). This serves as a coordinated Ca21-sensitive trigger for endocytosis (Marks and McMahon 1998; Cousin and Robinson 2000; this study). Since calcineurin dephosphorylates a whole set of proteins, all of which are implicated in endocytosis, it is not yet possible to deduce whether dephosphorylation of all or some of the dephosphins is essential for endocytosis. Prior to this investigation, there was no direct evidence for a role for the phosphorylation of any of these proteins in endocytosis. This study revealed four main ®ndings concerning endocytosis in nerve terminals.

Fig. 7 Inhibition of phosphorylation has noimmediate effect on endocytosis at S1. (a) Exocytosis is unaffected by Ro 31-8220 at S1. Ca21-dependent glutamate release evoked by 30 mM KCl (S1) is displayed, either in the presence or absence of 10 mM Ro 31-8220 or Go 7874. Synaptosomes were pre-incubated with both antagonists 15 min before stimulation (n 3 ^ SEM). (b) Weak effects of PKC inhibitors on endocytosis at S1. Synaptosomes were loaded with FM2-10 (S1) by stimulation with 30 mM KCl either in the

presence or absence of Ca21. Synaptosomes were pre-incubated for 15 min with either 10 mM Ro 31-8220 or Go 7874. The subsequent release of loaded FM2-10 by a standard stimulus of 30 mM KCl in the presence of Ca21 is displayed after subtraction of traces loaded in the absence of Ca21 (S1/S2 method, Fig. 1). Traces are representative and normalized to an arbitrary value (n 3±7). (c) KCl-evoked retrieval ef®ciency is displayed for Ro 31-8220 and Go 7874.

Discussion



Fig. 8 Inhibition of rephosphorylation blocks endocytosis at S2. (a) Exocytosis is unaffected by Ro 31-8220 at S2. Synaptosomes were pre-incubated with 10 mM Ro 31-8220 or Go 7874 for 15 min before stimulation with 30 mM KCl in the presence of Ca21. Synaptosomes were repolarized and Ca21-dependent glutamate release evoked by 30 mM KCl (S2) is displayed either in the continued presence or in the absence of the antagonists indicated (n 3 ^ SEM). (b) Endocytosis is blocked by PKC inhibitors at S2. Synaptosomes were pre-incubated with 10 mM Ro 31-8220 or Go 7874 15 min before stimulation with 30 mM KCl in the presence of Ca21. Synaptosomes were repolarized

and loaded with FM2-10 (S2) by stimulation with 30 mM KCl either in the presence or absence of Ca21. Synaptosomes were continually incubated with either 10 mM Ro 31-8220 or Go 7874 for all stages. The subsequent release of loaded FM2-10 by a standard stimulus of 30 mM KCl in the presence of Ca21 is displayed after subtraction of traces loaded in the absence of Ca21 (S2/S3 method, Fig. 1). Traces are representative and normalized to an arbitrary value (n 4). (c) KCl-evoked retrieval ef®ciency is displayed for Ro 318220 and Go 7874.

1 More than one protein kinase plays a role in the phosphorylation of different dephosphins, thus providing some degree of speci®city for phosphorylation events during endocytosis. 2 AP180 is a new member of the dephosphin family of proteins as it undergoes stimulus-dependent dephosphorylation by calcineurin in nerve terminals. 3 Phosphorylation of a subset of the dephosphins that includes dynamin I and/or synaptojanin, but not necessarily of AP180 or amphiphysin, is essential for endocytosis. Because phosphorylation is not required for immediate endocytosis, but for subsequent cycles, this shows that phosphorylation prepares the nerve terminal for the next round of stimulation. These results are the ®rst to demonstrate that the phosphorylation of a subset of nerve terminal proteins regulates endocytosis. 4 Only the small nerve terminal pool of phosphodynamin I may control endocytosis.

terminals has not yet been identi®ed but does not appear to be PKC. This study does not discount the possibility that the rephosphorylation of these dephosphins may also be required for endocytosis, but the question can be addressed only when a speci®c antagonist of the protein kinase(s) that phosphorylates AP180 and amphiphysins 1 and 2 is identi®ed. Phosphorylation of either or both dynamin I or synaptojanin (or an unidenti®ed dephosphin) could be essential for endocytosis, however it is more likely that dynamin I is the key target for the inhibitor. This is because synaptojanin has not been shown to directly participate in endocytosis, but rather in the uncoating of synaptic vesicles after endocytosis (Cremona et al. 1999). A block of synaptic vesicle uncoating renders vesicles unable to be released (Cremona et al. 1999), and this would have been detected in our assay by monitoring the resultant release of glutamate at the S3 phase of stimulation (Fig. 8). However no inhibition of S3 glutamate release was observed using Ro 31-8220 (data not shown). A possible role for synaptojanin in endocytosis has been predicted from circumstantial evidence, such as its binding to other proteins which are essential for endocytosis: eps15 (Haffner et al. 1997), amphiphysin (McPherson et al. 1996) and endophilin (De Heuvel et al. 1997; Ringstad et al. 1997). Synaptojanin is also highly enriched in brain and is concentrated in endocytic intermediate structures in nerve terminals (McPherson et al. 1994). However, overexpression of synaptojanin in non-neuronal cells leads to a massive rearrangement of the actin cytoskeleton, but has no effect on endocytosis (Sakisaka et al. 1997). The only demonstration of a possible direct role for synaptojanin comes from studies

Distinct protein kinases are involved in endocytosis The rephosphorylation of the dephosphins is not under the control of a single protein kinase, in contrast to their coordinated dephosphorylation by calcineurin. The PKC antagonist Ro 31-8220 abolished dynamin I and synaptojanin rephosphorylation, but not rephosphorylation of AP180 or amphiphysins 1 and 2. The protein kinase that phosphorylates dynamin I is likely to be PKC (Robinson 1992; Liu et al. 1994b). The present study suggests that synaptojanin might also be a substrate for this enzyme, however this remains to be determined. The protein kinase that phosphorylates AP180 and amphiphysins 1 and 2 in nerve



Fig. 9 Phorbol esters do not modulate endocytosis in nerve terminals. (a) and (b): Exocytosis is unaffected by PMA. Ca21-dependent glutamate release evoked by 30 mM KCl in (a) the S1 phase or (b) the S2 phase is displayed either in the presence or absence of 100 nM PMA. Synaptosomes were pre-incubated with PMA for 2 min before stimulation (n 3 ^ SEM). (c) and (d): Endocytosis is unaffected by PMA. Synaptosomes were loaded with FM2-10 by stimulation with 30 mM KCl in either (c) the S1 phase or (d) the S2 phase of stimulation in plus or minus Ca21 solution. Synaptosomes were pre-incubated for 2 min with 100 nM PMA where indicated. The subsequent release of loaded FM2-10 by a standard stimulus of 30 mM KCl in the presence of Ca21 is displayed after subtraction of traces loaded in the absence of Ca21. Traces are representative and normalized to an arbitrary value (n 3). (e) S1 and S2 retrieval ef®ciency for PMA-treated synaptosomes.

in yeast where double knockouts of synaptojanin-like genes had severe defects in receptor-mediated and ¯uid-phase endocytosis (Singer-KruÈger et al. 1998). It is therefore more likely that dynamin I phosphorylation, rather than synaptojanin phosphorylation, is the essential step in endocytosis revealed by PKC inhibitors, although this will need to be clari®ed by further studies. The lack of effect of phorbol esters on S2 endocytosis may appear to argue against a role for phosphorylation by PKC in endocytosis. However, dynamin I rephosphorylation is also unaffected by phorbol esters in nerve terminals (Robinson 1992). This raises the question as to whether PKC is the protein kinase that phosphorylates dynamin I and synaptojanin in nerve terminals. Ro 31-8220 is one of the most speci®c antagonists reported for PKC, however it is also exhibits similar in vitro potency against MAPKAP kinase1-b, p70-S6 kinase, cdc2 kinase and Akt-3 (Alessi 1997; Begemann et al. 1998; Masure et al. 1999). It is possible that Ro 31-8220 inhibits dynamin I and synaptojanin phosphorylation by blocking any of these protein kinases or an as yet unidenti®ed protein kinase. Thus it is not yet certain that dynamin I and synaptojanin share the same protein kinase. In vitro phosphorylation studies have shown that other protein kinases such as cdc2 kinase and MAP kinase can phosphorylate dynamin I (Hosoya et al. 1994; Earnest et al. 1996), and possible roles for such enzymes in nerve terminals cannot be eliminated.

AP180 is a new dephosphin We found that AP180 is a new member of the dephosphin family of nerve terminal phosphoproteins. Taking advantage of the extremely acidic nature of the protein AP180 (pI 3.85) we modi®ed a simple acetic acid extraction method (Robinson et al. 1993a) to isolate the phosphoprotein from 32 P-labelled synaptosomes. AP180 was identi®ed by amino acid sequencing and it underwent stimulus-dependent dephosphorylation mediated by calcineurin, as it was sensitive to cyclosporin A (Liu et al. 1994a; Bauerfeind et al. 1997). The protein kinase that rephosphorylates AP180 remains to be identi®ed. It seems certain that PKC is not the AP180 kinase because its phosphorylation in intact nerve terminals was not blocked by Ro 31-8220 and PKC could not phosphorylate either native (this study) or recombinant AP180 in vitro (Hao et al. 1999). AP180 is a known in vitro substrate for casein kinase II on at least three sites in its middle, acidic domain (Hao et al. 1999), but it is unclear if this is also the in vivo protein kinase. The neurone-speci®c AP180 (or AP180-1) has been shown by a variety of approaches to be essential for endocytosis of synaptic vesicles (Zhang et al. 1998; Morgan et al. 1999) and the ubiquitously expressed AP180-2 (or CALM) has been shown to be essential for endocytosis of EGF and transferrin receptors (Tebar et al. 1999). Our results raise the possibility that AP180 phosphorylation might play a role in synaptic vesicle endocytosis. AP180 and another clathrin assembly



protein AP-2 form a complex that assembles clathrin cages more ef®ciently than either AP180 or AP-2 alone (Hao et al. 1999). Phosphorylation of AP180 by casein kinase II reduces the binding of AP-2 and reduces the ability of the AP180/AP-2 complex to assemble clathrin. Therefore the stimulus-dependent dephosphorylation of AP180 observed in this study may promote its interaction with AP-2 and thus more ef®cient clathrin assembly. Obtaining pharmacological evidence for this hypothesis in intact nerve terminals is currently problematic because all of the dephosphins are dephosphorylated by calcineurin and speci®c cell-permeable casein kinase II antagonists are not available. Rephosphorylation of dynamin I is required for endocytosis Depolarization-stimulated phosphorylation of nerve terminal proteins is not essential for an immediate cycle of endocytosis, because Ro 31-8220, Go 7874 or phorbol esters did not affect endocytosis at S1. This is in agreement with previous studies using central neurones where no inhibition of endocytosis was found in primary cultures of cerebellar granule cells or hippocampal neurones incubated with the protein kinase antagonist staurosporine (Kraszewski et al. 1996; Cousin et al. 1999). Similarly, KCl-evoked glutamate release was not affected by Ro 31-8220 at either S1 or S2, in agreement with previous studies using nerve terminals and cultured neurones (Coffey et al. 1993; Cousin et al. 1999). Go 7874 partly inhibited glutamate release in S1 and S2, possibly via a non-speci®c block of a pre-synaptic Ca21 channel, thus reducing its utility for further studies of endocytosis. These results con®rm that PKC is not involved in an immediate round of endocytosis. Inhibition of protein kinase activity does not immediately affect the dephosphins, because they are already phosphorylated in nerve terminals prior to the stimulus. Previous studies examining the role of phosphorylation in endocytosis did not take this into account (Kraszewski et al. 1996; Cousin et al. 1999). The S2/S3 protocol employed here to study endocytosis ensured that the dephosphins were dephosphorylated when Ro 31-8220 and Go 7874 were present, so that the pools of phosphorylated dynamin I and synaptojanin were severely reduced prior to the next cycle of endocytosis. Using this paradigm we showed that the rephosphorylation of dynamin I, synaptojanin or other unidenti®ed dephosphins is required for a second or subsequent cycle of endocytosis, but not for exocytosis. The action of the PKC antagonists is likely to be phosphorylationspeci®c, as neither Ro 31-8220 nor Go 7874 block in vitro dynamin GTPase activity (J. Rusak, and PJR, unpublished). This suggests that nerve terminals cannot recover from a burst of endocytosis when dynamin I phosphorylation is inhibited. Therefore phosphorylation either terminates endocytosis or is a primary step required to set up the next cycle.

The results also highlight key issues in pharmacological approaches to inhibition of endocytosis. While Ro 31-8220 gave clear results that were speci®cally related to endocytosis, Go 7874 signi®cantly affected both exocytosis and endocytosis. Despite this effect on exocytosis a speci®c effect of Go 7874 on endocytosis was still evident through the use of the retrieval ef®ciency index (Cousin and Robinson 1998; 2000). Taking into account the effect of drugs on exocytosis reveals whether they have speci®c effects on endocytosis. This showed that the inhibition of S1 endocytosis by Go 7874 was primarily due to its effect on exocytosis. This emphasizes the need to always monitor exocytosis and endocytosis in parallel for the meaningful investigation of endocytosis in nerve terminals.

The phosphorylated pool of dynamin I controls endocytosis A ®nal conclusion from this study is that only the phosphorylated pool of dynamin I is required for endocytosis. More than 90% of dynamin I was previously shown to be membrane-associated and cannot be phosphorylated by PKC. The remaining 10% is cytosolic and includes all of the phosphorylated form (Liu et al. 1994b). Therefore only the phosphorylated pool of dynamin I located in the cytosol is required for endocytosis. The majority of dynamin I in resting nerve terminals might serve as a reserve pool in times of intense stimulation or it may play roles in addition to those involved in endocytosis (Kranenburg et al., 1999; Whistler and Von Zastrow 1999). The subcellular distribution of the other dephosphins with respect to their phosphorylation status remains to be determined. How does cycling of the phosphorylation status of dynamin I regulate endocytosis? One known consequence of phosphorylation of peripheral membrane proteins such as MARCKS, synapsin I and spectrin is an inhibition of their binding to phospholipids and the consequent release into the cytosol (Wang et al. 1989; Fowler and Adam 1992; Tarelli et al. 1992; Vorotnikov et al. 1992; Kim et al. 1994). Similarly dynamin I binding to phospholipid is abolished when it is phosphorylated by PKC (Powell et al. 2000). Therefore after completion of endocytosis, phosphorylation may remove dynamin I from the plasma membrane to a cytosolic compartment where it is appropriately localized to participate in the next cycle of endocytosis. Because membrane-associated dynamin I cannot be phosphorylated, a conformational change in dynamin I may be required to enable its phosphorylation and return to the cytosol. This may be a result of either dynamin I binding another protein, such as intersectin or amphiphysin, or a change in its conformation upon GTP hydrolysis (Stowell et al. 1999). In this model, the phosphorylation of membrane-associated dynamin I caused its redistribution back to the cytosol to set up the next cycle of endocytosis.



Acknowledgements We wish to thank Chandra Malladi (CMRI, Sydney) for puri®cation of AP180, and Peter Milburn (Australian National University, Canberra) and Amanda Hall (Newcastle Protein, Newcastle, UK) for the sequence analysis of AP180. We also wish to thank Harvey McMahon and Pietro De Camilli for providing antibodies and plasmids, and Peter Rowe for helpful comments on the manuscript. This work was supported by the National Health and Medical Research Council of Australia. MAC is a longterm fellow of the Human Frontiers of Science Program.

References Alessi D. R. (1997) The protein kinase C inhibitors Ro 318220 and GF 109203X are equally potent inhibitors of MAPKAP kinase-1beta (Rsk-2) and p70, S6 kinase. FEBS Lett. 402, 121±123. Bauerfeind R., Takei K. and De Camilli P. (1997) Amphiphysin I is associated with coated endocytic intermediates and undergoes stimulation-dependent dephosphorylation in nerve terminals. J. Biol. Chem. 272, 30984±30992. Begemann M., Kashimawo S. A., Lunn R. M., Delohery T., Choi Y. J., Kim S., Heitjan D. F., Santella R. M., Schiff P. B., Bruce J. N. and Weinstein I. B. (1998) Growth inhibition induced by Ro 31±8220 and calphostin C in human glioblastoma cell lines is associated with apoptosis and inhibition of CDC2 kinase. Anticancer Res. 18, 3139±3152. van der Bliek A. M., Redelmeier T. E., Damke H., Tisdale E. J., Meyerowitz E. M. and Schmid S. L. (1993) Mutations in human dynamin block an intermediate stage in coated vesicle formation. J. Cell Biol. 122, 553±563. Chen H., Fre S., Slepnev V. I., Capua M. R., Takei K., Butler M. H., Di Fiore P. P. and De Camilli P. (1998) Epsin is an EH-domainbinding protein implicated in clathrin-mediated endocytosis. Nature 394, 793±797. Chen H., Slepnev V. I., Di Fiore P. P. and De Camilli P. (1999) The interaction of Epsin and Eps15 with the clathrin adaptor AP-2 is inhibited by mitotic phosphorylation and enhanced by stimulationdependent dephosphorylation in nerve terminals. J. Biol. Chem. 274, 3257±3260. Coffey E. T., Sihra T. S. and Nicholls D. G. (1993) Protein kinase C and the regulation of glutamate exocytosis from cerebrocortical synaptosomes. J. Biol. Chem. 268, 21060±21065. Cousin M. A. and Robinson P. J. (1998) Ba21 does not support synaptic vesicle retrieval in rat isolated presynaptic nerve terminals. Neurosci. Lett. 253, 1±4. Cousin M. A. and Robinson P. J. (2000) Ca21 inhibition of dynamin arrests synaptic vesicle recycling at the active zone. J. Neurosci. 20, 949±957. Cousin M. A., McLaughlin M. and Nicholls D. G. (1999) Protein kinase C modulates ®eld-evoked transmitter release from cultured rat cerebellar granule cells via a dendrotoxin-sensitive K1 channel. Eur. J. Neurosci. 11, 101±109. Cremona O., Di Paolo G., Wenk M. R., Luthi A., Kim W. T., Takei K., Daniell L., Nemoto Y., Shears S. B., Flavell R. A., McCormick D. A. and De Camilli P. (1999) Essential role of phosphoinositide metabolism in synaptic vesicle recycling. Cell 99, 179±188. De Heuvel E., Bell A. W., Ramjaun A. R., Wong K., Sossin W. S. and McPherson P. S. (1997) Identi®cation of the major synaptojaninbinding proteins in brain. J. Biol. Chem. 272, 8710±8716. Dunkley P. R., Jarvie P. E., Heath J. W., Kidd G. J. and Rostas J. A. (1986) A rapid method for isolation of synaptosomes on Percoll gradients. Brain Res. 372, 115±129.

Earnest S., Khokhlatchev A., Albanesi J. P. and Barylko B. (1996) Phosphorylation of dynamin by ERK2 inhibits the dynamin± microtubule interaction. FEBS Lett. 396, 62±66. Fowler V. M. and Adam E. J. (1992) Spectrin redistributes to the cytosol and is phosphorylated during mitosis in cultured cells. J. Cell Biol. 119, 1559±1572. Haffner C., Takei K., Chen H., Ringstad N., Hudson A., Butler M. H., Salcini A. E., Di F. P. and De Camilli P. (1997) Synaptojanin 1: localization on coated endocytic intermediates in nerve terminals and interaction of its 170 kDa isoform with Eps15. FEBS Lett. 419, 175±180. Hao W., Luo Z., Zheng L., Prasad K. and Lafer E. M. (1999) AP180 and AP-2 interact directly in a complex that cooperatively assembles clathrin. J. Biol. Chem. 274, 22785±22794. Hosoya H., Komatsu S., Shimizu T., Inagaki M., Ikegami M. and Yazaki K. (1994) Phosphorylation of dynamin by cdc2 kinase. Biochem. Biophys. Res. Commun. 202, 1127±1133. Huang K. M., D'Hondt K., Riezman H. and Lemmon S. K. (1999) Clathrin functions in the absence of heterotetrameric adaptors and AP180-related proteins in yeast. EMBO J. 18, 3897±3908. Kim J., Shishido T., Jiang X., Aderem A. and McLaughlin S. (1994) Phosphorylation, high ionic strength, and calmodulin reverse the binding of MARCKS to phospholipid vesicles. J. Biol. Chem. 269, 28214±28219. Klingauf J., Kavalali E. T. and Tsien R. W. (1998) Kinetics and regulation of fast endocytosis at hippocampal synapses. Nature 394, 581±585. Kranenburg O., Verlaan I. and Moolenaar W. H. (1999) Dynamin is required for the activation of mitogen-activated protein (MAP) kinase by MAP kinase kinase. J. Biol. Chem. 274, 35301±35304. Kraszewski K., Daniell L., Mundigl O. and De Camilli P. (1996) Mobility of synaptic vesicles in nerve endings monitored by recovery from photobleaching of synaptic vesicle-associated ¯uorescence. J. Neurosci. 16, 5905±5913. Liu J. P., Sim A. T. R. and Robinson P. J. (1994a) Calcineurin inhibition of dynamin I GTPase activity coupled to nerve terminal depolarization. Science 265, 970±973. Liu J. P., Powell K. A., SuÈdhof T. C. and Robinson P. J. (1994b) Dynamin I is a Ca21-sensitive phospholipid-binding protein with very high af®nity for protein kinase C. J. Biol. Chem. 269, 21043±21050. Marks B. and McMahon H. T. (1998) Calcium triggers calcineurindependent synaptic vesicle recycling in mammalian nerve terminals. Curr. Biol. 8, 740±749. Masure S., Haefner B., Wesselink J. J., Hoefnagel E., Mortier E., Verhasselt P., Tuytelaars A., Gordon R. and Richardson A. (1999) Molecular cloning, expression and characterization of the human serine/threonine kinase Akt-3. Eur. J. Biochem. 265, 353±360. McPherson P. S., Takei K., Schmid S. L. and De Camilli P. (1994) p145, a major Grb2-binding protein in brain, is co-localized with dynamin in nerve terminals where it undergoes activity-dependent dephosphorylation. J. Biol. Chem. 269, 30132±30139. McPherson P. S., Garcia E. P., Slepnev V. I., David C., Zhang X. M., Grabs D., Sossin W. S., Bauerfeind R., Nemoto Y. and De Camilli P. (1996) A presynaptic inositol-5-phosphatase. Nature 379, 353±357. Morgan J. R., Zhao X., Womack M., Prasad K., Augustine G.J. and Lafer E. M. (1999) A role for the clathrin assembly domain of AP180 in synaptic vesicle endocytosis. J. Neurosci. 19, 10201±10212. Murphy J. E., Hanover J. A., Froehlich M., DuBois G. and Keen J. H. (1994) Clathrin assembly protein AP-3 is phosphorylated and glycosylated on the 50-kDa structural domain. J. Biol. Chem. 269, 21346±21352.



Nicholls D. G. and Sihra T. S. (1986) Synaptosomes possess an exocytotic pool of glutamate. Nature 321, 772±773. Nichols R. A., Suplick G. R. and Brown J. M. (1994) Calcineurin-mediated protein dephosphorylation in brain nerve terminals regulates the release of glutamate. J. Biol. Chem. 269, 23817±23823. Nonet M. L., Holgado A. M., Brewer F., Serpe C. J., Norbeck B. A., Holleran J., Wei L., Hartwieg E., Jorgensen E. M. and Alfonso A. (1999) UNC-11, a Caenorhabditis elegans AP180 homologue, regulates the size and protein composition of synaptic vesicles. Mol. Biol. Cell 10, 2343±2360. Powell K. A., Valova V. A., Malladi C. S., Jensen O. N., Larsen M. R. and Robinson P. J. (2000) Phosphorylation of dynamin I on Ser795 by protein kinase C blocks its association with phospholipids. J. Biol. Chem. 275, 11610±11617. Ringstad N., Nemoto Y. and De Camilli P. (1997) The SH3p4/Sh3p8/ SH3p13 protein family: binding partners for synaptojanin and dynamin via a Grb2-like Src homology 3 domain. Proc. Natl Acad. Sci. USA 94, 8569±8574. Robinson P. J. (1992) Differential stimulation of protein kinase C activity by phorbol ester or calcium/phosphatidylserine in vitro and in intact synaptosomes. J. Biol. Chem. 267, 21637±21644. Robinson P. J. and Dunkley P. R. (1983) Depolarisation-dependent protein phosphorylation in rat cortical synaptosomes: factors determining the magnitude of the response. J. Neurochem. 41, 909±918. Robinson P. J., Liu J. P., Chen W. and Wenzel T. (1993a) Activation of protein kinase C in vitro and in intact cells or synaptosomes determined by acetic acid extraction of MARCKS. Anal. Biochem. 210, 172±178. Robinson P. J., Sontag J.-M., Liu J. P., Fykse E. M., Slaughter C., McMahon H. T. and SuÈdhof T. C. (1993b) Dynamin GTPase regulated by protein kinase C phosphorylation in nerve terminals. Nature 365, 163±166. Robinson P. J., Liu J. P., Powell K. A., Fykse E. M. and SuÈdhof T. C. (1994) Phosphorylation of dynamin I and synaptic vesicle recycling. Trends Neurosci. 17, 348±353. Ryan T. A. (1996) Endocytosis at nerve terminals: timing is everything. Neuron 17, 1035±1037. Sakisaka T., Itoh T., Miura K. and Takenawa T. (1997) Phosphatidylinositol 4,5-bisphosphate phosphatase regulates the rearrangement of actin ®laments. Mol. Cell. Biol. 17, 3841±3849. Shupliakov O., Low P., Grabs D., Gad H., Chen H., David C., Takei K., De Camilli P. and Brodin L. (1997) Synaptic vesicle endocytosis impaired by disruption of dynamin±SH3 domain interactions. Science 276, 259±263. Singer-KruÈger B., Nemoto Y., Daniell L., Ferro-Novick S. and De Camilli P. (1998) Synaptojanin family members are implicated in endocytic membrane traf®c in yeast. J. Cell Sci. 111, 3347± 3356. Slepnev V. I., Ochoa G. C., Butler M. H., Grabs D. and DeCamilli P.

(1998) Role of phosphorylation in regulation of the assembly of endocytic coat complexes. Science 281, 821±824. Sontag J.-M., Fykse E. M., Ushkaryov Y., Liu J. P., Robinson P. J. and SuÈdhof T. C. (1994) Differential expression and regulation of multiple dynamins. J. Biol. Chem. 269, 4547±4554. Stowell M. H., Marks B., Wigge P. and McMahon H. T. (1999) Nucleotide-dependent conformational changes in dynamin: evidence for a mechanochemical molecular spring. Nat. Cell Biol. 1, 27±32. Sweitzer S. M. and Hinshaw J. E. (1998) Dynamin undergoes a GTPdependent conformational change causing vesiculation. Cell 93, 1021±1029. Takei K., McPherson P. S., Schmid S. L. and De Camilli P. (1995) Tubular membrane invaginations coated by dynamin rings are induced by GTP-gS in nerve terminals. Nature 374, 186±190. Tarelli F. T., Bossi M., Fesce R., Greengard P. and Valtorta F. (1992) Synapsin I partially dissociates from synaptic vesicles during exocytosis induced by electrical stimulation. Neuron 9, 1143±1153. Tebar F., Bohlander S. K. and Sorkin A. (1999) Clathrin assembly lymphoid myeloid leukemia (CALM) protein: localization in endocytic-coated pits, interactions with clathrin, and the impact of overexpression on clathrin-mediated traf®c. Mol. Biol. Cell 10, 2687±2702. Turner K. M., Burgoyne R. D. and Morgan A. (1999) Protein phosphorylation and the regulation of synaptic membrane traf®c. Trends Neurosci. 22, 459±464. Vorotnikov A. V., Bogatcheva N. V. and Gusev N. B. (1992) Caldesmon±phospholipid interaction. Effect of protein kinase C phosphorylation and sequence similarity with other phospholipidbinding proteins. Biochem. J. 284, 911±916. Wang J. K., Walaas S. I., Sihra T. S., Aderem A. A. and Greengard P. (1989) Phosphorylation and associated translocation of the 87-kDa protein, a major protein kinase C substrate, in isolated nerve terminals. Proc. Natl Acad. Sci. USA 86, 2253±2256. Whistler J. L. and Von Zastrow M. (1999) Dissociation of functional roles of dynamin in receptor-mediated endocytosis and mitogenic signal transduction. J. Biol. Chem. 274, 24575±24578. Ye W. and Lafer E. M. (1995) Bacterially expressed F1±20/AP-3 assembles clathrin into cages with a narrow size distribution: implications for the regulation of quantal size during neurotransmission. J. Neurosci. Res. 41, 15±26. Zhang B., Koh Y. H., Beckstead R. B., Budnik V., Ganetzky B. and Bellen H. J. (1998) Synaptic vesicle size and number are regulated by a clathrin adaptor protein required for endocytosis. Neuron 21, 1465±1475. Zhou S., Sousa R., Tannery N. H. and Lafer E. M. (1992) Characterization of a novel synapse-speci®c protein. II. cDNA cloning and sequence analysis of the F1±20 protein. J. Neurosci. 12, 2144±2155.
