Address correspondence and reprint requests to: Dr Seung R. Paik, Department of Biochemistry, College of Medicine, Inha University, 253 Yonghyun-Dong, Nam-Ku, Inchon 402–751, Korea (Republic of). E-mail: firstname.lastname@example.org
α-Synuclein, a pathological component of Parkinson's disease by constituting the Lewy bodies, has been suggested to be involved in membrane biogenesis via induction of amphipathic α-helices. Since the amphipathic α-helix is also known as a recognition signal of calmodulin for its target proteins, molecular interaction between α-synuclein and calmodulin has been investigated. By employing a chemical coupling reagent of N-(ethoxycarbonyl)-2-ethoxy-1,2-dihydroquinoline, α-synuclein has been shown to yield a heterodimeric 1 : 1 complex with calmodulin on sodium dodecyl sulfate–polyacrylamide gel electrophoresis in the presence and even absence of calcium, whereas β-synuclein was more dependent upon calcium for its calmodulin interaction. The selective calmodulin interaction of α-synuclein in the absence of calcium was also demonstrated with the aggregation kinetics of the synucleins in which only the α-synuclein aggregation was affected by calmodulin. A reversible binding assay confirmed that α-synuclein interacted with the Ca2+-free as well as the Ca2+-bound calmodulins with almost identical Kds of 0.35 µm and 0.31 µm, respectively, while β-synuclein preferentially recognized the Ca2+-bound form with a Kd of 0.68 µm. By using a C-terminally truncated α-synuclein of α-syn97, the calmodulin binding site(s) on α-synuclein was(were) shown to be located on the N-terminal region where the amphipathic α-helices have been suggested to be induced upon membrane interaction. By employing liposome and calmodulin in a state of being either soluble or immobilized on agarose, actual competition of α-synuclein between membranes and calmodulin was demonstrated with the observation that α-synuclein previously bound to the liposome was released upon specific interaction with the calmodulins. Taken together, these data may suggest that α-synuclein could act not only as a negative regulator for calmodulin in the presence and even absence of calcium, but it could also exert its activity at the interface between calmodulin and membranes.
sodium dodecyl sulfate–polyacrylamide gel electrophoresis.
Parkinson's disease (PD), characterized by resting tremor, slowness of initial movement, rigidity, and general postural instability, is one of the most prevailing neurodegenerative disorders among elderly population (Forno 1996). The disease is pathologically defined by loss of dopaminergic neurons in the substantia nigra and the presence of eosinophilic intraneuronal proteinaceous inclusions called Lewy bodies (Forno 1996; Gai et al. 2000). α-Synuclein is the major filamentous constituent of the Lewy bodies along with other components such as ubiquitin, neurofilaments, and lipids (Spillantini et al. 1997, 1998; Baba et al. 1998; Gai et al. 2000). In fact, α-synuclein became the first protein shown to be genetically linked to a few rare autosomal dominant familial cases of PD. Two independent missense mutations in its gene were identified, which resulted in substitution of alanine residues at positions of either 30 or 53 with proline and threonine (Ala30Pro and Ala53Thr), respectively (Polymeropoulos et al. 1997; Krüger et al. 1998). It was demonstrated that wild-type α-synuclein was fibrillated in vitro and the process was accelerated with the mutant forms (Conway et al. 1998, 2000; Narhi et al. 1999). The protein has been also pathologically related to other neurodegenerative disorders including Alzheimer's disease (AD), dementia with Lewy bodies (DLB), and multiple system atrophy (MSA) by participating in abnormal protein depositions (Goedert et al. 1998; Hardy and Gwinn-Hardy 1998). It has been generally accepted that the protein accumulation and resulting amyloid formation could play a significant role for neuronal cell death (Goedert et al. 1998). Thus, in order to assess the pathogenesis of PD, a large amount of research has been carried out from the aspect of protein aggregation. On the other hand, however, physiological function of α-synuclein whose disruption could also contribute to development of the disease has not been well established although the protein has been suggested to exhibit regulatory functions in vivo by interacting with proteins such as phospholipase D2 (Jenco et al. 1998), synphilin-1 (Engelender et al. 1999), tau (Jensen et al. 1999), G protein-coupled receptor kinase (Pronin et al. 2000), and 14-3-3 protein ligands including 14-3-3 proteins, protein kinase C, BAD, and extracellular regulated kinase (ERK) (Ostrerova et al. 1999).
Lipid and membranes have also drawn considerable attention as another specific intracellular target of α-synuclein interaction, which leads to suggest the protein being involved in regulating size of the presynaptic vesicular pool and ultimate synaptic plasticity (Clayton and George 1999; Murphy et al. 2000). Upon binding to membranes of acidic phospholipids, α-synuclein experienced dramatic structural transition from its ‘natively unfolded’ state (Weinreb et al. 1996; Kim 1997) to a structure containing around 80%α-helix (Davidson et al. 1998). The helical wheel analysis of α-synuclein predicted five potential amphipathic α-helical regions which occupied more than half of the molecule from its N-terminus (residues 1–93) (Davidson et al. 1998). In addition, several observations of self-association/oligomerization of α-synuclein upon membrane and fatty acid interaction were also reported (Jo et al. 2000; Leng et al. 2001; Perrin et al. 2001; Sharon et al. 2001). During subcellular fractionation experiments, however, α-synuclein has been isolated dominantly in its soluble form apart from at the membranes, suggesting that there might be a certain exchange mechanism of α-synuclein between membranes and cytosol (Davidson et al. 1998). In this respect, we have investigated calmodulin interaction of α-synuclein not only because calmodulin recognizes its target protein via the basic amphipathic α-helix (Sanyal et al. 1988; O'Neil et al. 1989; O'Neil and DeGrado 1990) which also serves as a common signal for membrane interaction of proteins, but also to examine the possible implication of α-synuclein in the calcium-dependent process of vesicular biogenesis (Bennett and Scheller 1994) which is a critical requirement for the proposed role of α-synuclein in membrane plasticity(Clayton and George 1998, 1999).
Materials and methods
Calmodulin, calmodulin–agarose, and ubiquitin–agarose were obtained from Sigma (St Louis, MO, USA). Calmodulin peptide inhibitor corresponding to the calmodulin binding domain of CaM kinase II Leu290-Ala309 and sephadex G-200 were obtained from Calbiochem (San Diego, CA, USA) and Pharmacia (Piscataway, NJ, USA), respectively. Protein assay kit employing bicinchoninic acid (BCA) and glycerol were from Pierce (Rockford, IL, USA) and Junsei Chemical Co. (Tokyo, Japan), respectively. Acetonitrile, ethanol, and methanol were from Fisher Scientific (Pittsburgh, PA, USA). Precast gels for 10–20% Tricine/sodium dodecyl sulfate (SDS)/polyacrylamide gel electrophoresis (PAGE) were provided from Novex (San Diego, CA, USA). Materials for chromatography and other reagents used in this study were obtained from Sigma, unless otherwise mentioned.
Preparation of recombinant α- and β-synucleins and the C-terminally truncated α-syn97
Recombinant α- and β-synucleins cloned in pRK172 (kindly provided by Dr R. Jakes) were overexpressed in Escherichia coli BL21(DE3). The α-synuclein was completely purified through heat treatment, DEAE–sephacel anion-exchange, sephacryl S-200 size-exclusion, and S-sepharose cation-exchange chromatography steps according to the previously described procedure (Paik et al. 1997). The C-terminally truncated α-synuclein, α-syn97, was prepared with endoproteinase Asp-N treatment as described elsewhere (Paik et al. 1999). For the purification of β-synuclein, IPTG-treated E. coli extract was subjected to heat treatment at 100°C for 20 min. The supernatant was purified with DEAE anion-exchange chromatography (2.5 cm × 10 cm) equilibrated with 20 mm Tris-Cl, pH 7.5. The bound proteins were eluted with a linear gradient between 0.1 m and 0.4 m NaCl at a flow rate of 1.5 mL/min. After combining the β-synuclein containing fractions as determined by SDS–PAGE, the sample was diluted three-fold with 20 mm Tris-Cl, pH 7.5, and subjected to Q-sepharose anion-exchange chromatography (1.6 cm × 7 cm). The protein was eluted with a linear gradient between 0.2 m and 0.5 m NaCl in the Tris buffer. The purified β-synuclein was dialyzed against 20 mm 2-(N-morpholino)ethanesulfonic acid (Mes), pH 6.5, with two times of buffer change for a total volume of 7.5 L during 16 h. The protein was quantitated with the BCA assay (Smith et al. 1985) and kept at a concentration of 2.14 mg/mL in 200 µL aliquots at −30°C.
Cross-linking between calmodulin and the synucleins
Interactions between calmodulin and the synucleins were analyzed with a chemical coupling reagent of N-(ethoxycarbonyl)-2-ethoxy-1,2-dihydroquinoline (EEDQ). The synucleins (100 pmol each) were separately preincubated with calmodulin at molar ratios of 1 : 0.1, 1 : 0.5, 1 : 1, and 1 : 3 (synucleins/calmodulin) in 20 mm Mes, pH 6.5, containing 1 mm EDTA in the presence and absence of 5 mm CaCl2 for 30 min at 37°C. Following preincubation, 0.3 mm EEDQ, originally stocked in DMSO, was added whilst keeping the final DMSO concentration less than 10% and further incubated for additional 1 h (Lee et al. 1998). The cross-linking reaction was stopped with Tricine/SDS/PAGE sample buffer composed of 8% (w/v) SDS, 24% (v/v) glycerol, 0.015% Coomassie blue G, and 0.005% phenol red in 0.9 m Tris/HCl, pH 8.45, and subjected to electrophoresis with the precast 10–20% Tricine/SDS/PAGE gel. The cross-linked products were visualized with silver staining (Morrissey 1981).
Aggregation kinetics of α- and β-synucleins
Protein aggregation of either α- or β-synuclein was monitored with turbidity by measuring absorbance at 405 nm. The synucleins at the final concentration of ∼ 2 mg/mL in 20 mm Mes, pH 6.5, were incubated at room temperature under continuous shaking with an orbit shaker at 150 r.p.m. (Red Rotor, Hoefer Scientific Inc.). The extent of aggregation at each time point was estimated with the increase in turbidity during more than 60 h of incubation. The aggregation kinetics were carried out in the presence and absence of calmodulin at a 1 : 10 molar ratio between calmodulin and the synucleins with or without 1 mm CaCl2 in a total volume of 1 mL.
Pull-down assay for α-synuclein from rat brain extract with calmodulin–agarose
Soluble brain extract was prepared by homogenizing rat brains in 50 mm Tris-Cl, pH 7.5, at a 1 : 2 (w/v) ratio and recovered as a supernatant from subsequent ultracentrifugation at 100 000 × g for 1 h at 4°C. The extract (100 µL) was incubated with either calmodulin–agarose containing 18 µg of the protein or the inhibitor-treated calmodulin–agarose in the presence and absence of 2 mm CaCl2 for 1 h at 37°C with continuous shaking at 1400 r.p.m. (Thermomixer Compact, Eppendorf, Germany). The calmodulin–inhibitor complex bound to agarose was prepared by preincubating the calmodulin–agarose with the calmodulin peptide inhibitor corresponding to calmodulin binding domain of CaM kinase II Leu290-Ala309 at a 1 : 1 molar ratio for 30 min at 37°C under the shaking condition. Following brief centrifugation of the reaction mixtures, the precipitates were collected and washed with 50 mm Tris-Cl, pH 7.5. α-Synuclein bound to either the calmodulin–agarose or the calmodulin–inhibitor complex on agarose was eluted by treating the precipitate with 1 m NaCl in the buffer for 10 min. After further centrifugation, the supernatant was subjected to 10–20% Tricine/SDS/PAGE and the presence of α-synuclein was identified with western blotting. The gel was electrically transferred to poly(vinylidene difluoride) membrane and blotted with synuclein-1 mAb (1 : 1000; Transduction Laboratory, San Diego, CA, USA) and HRP-anti-mouse secondary antibody (1 : 1000; Sigma). The protein was visualized with an enhanced chemiluminescence detection kit (Amersham Pharmacia Biotech, Piscataway, NJ, USA).
Dissociation constants between calmodulin and the synucleins
Synuclein interaction of calmodulin was examined according to the procedure (Slemmon et al. 1996) employing immobilized calmodulin with the following modifications. The α- and β-synucleins were incubated with 1.2 µm calmodulin immobilized on agarose (0.9 mg calmodulin/mL of gel) at various molar ratios in 20 mm Mes, pH 6.5, in the presence and absence of 5 mm CaCl2 for 1 h at 37°C. Free synucleins were separated from the bound proteins on the calmodulin–agarose with a centrifugal filter with pore size of 0.22 µm (Ultrafree®-MC from Millipore) at 250 gfor 1 min. After washing the calmodulin–agarose with 100 µL of 20 mm Mes, pH 6.5, the bound synucleins were eluted by treating the beads with 1 m NaCl in 20 mm Mes, pH 6.5, for 10 min at room temperature and recovered with a centrifugation at 250 gfor 1 min. The collected synucleins were quantified by measuring absorbance at 274 and 220 nm. Saturation curves drawn between the amount of bound synuclein versus the total protein initially treated were converted into double reciprocal plots between 1/[the amount of total protein] and 1/[the absorbance at 220 nm] to obtain dissociation constants between calmodulin and the synucleins. The absorbance at 220 nm was preferentially used instead of that at 274 nm in order to estimate the bound synucleins since the proteins were generally recovered in low quantities from the calmodulin–agarose.
Competitive interaction of α-synuclein between calmodulin and membranes
Liposome was prepared with a lipid mixture of phosphatidic acid (PA) and phosphatidylcholine (PC) at a mass ratio of 1 : 1. A total of 5 mg of the mixture was dissolved in 1 mL chloroform and dried under nitrogen gas with vortex to increase surface area of resulting lipid film. The film on a test tube stored at −30°C was transferred to 60°C for 30 s. To the tube, 1 mL of 20 mm phosphate, pH 7.2, and glass beads were added and vortexed vigorously. Following removal of the beads, the sample was sonicated for 10 min using a Branson 3200 sonicator. The resulting cloudy solution was subjected to repeating extrusion (15–17 times) through 0.1 µm membrane equipped in the mini-extruder (Avanti polar lipids Inc.) at 50–60°C. The resulting large unilamellarliposomewasseparatedfrom freelipidswithsephadexG-200size-exclusionchromatography.
Membrane interaction of α-synuclein was examined by incubating 500 µL of the liposome (1.67 mg lipid/mL) and 0.1 mg of α-synuclein (11.6 µm) in 20 mm sodium phosphate, pH 7.2, at 4°C for 3 h in a total volume of 600 µL. The liposome bound α-synuclein was separated from the free protein with sephadex G-200 (superfine) size-exclusion chromatography (0.9 cm × 14 cm) at a flow rate of 0.35 mL/min and an elution profile of the protein in the collected fractions was analyzed by using the precast 10–20% Tricine/SDS/PAGE. Competition of calmodulin during the interaction between α-synuclein and membrane was examined as follows. The liposome interaction experiment was performed with immobilized calmodulin–agarose at a 1 : 1 molar ratio with α-synuclein in the presence and absence of 5 mm calcium during 3 h of incubation at 4°C. The mixture was then subjected to a brief centrifugation to precipitate the calmodulin–agarose. α-Synuclein bound to theliposome or released from the membrane was analyzed with the supernatant by using the sephadex G-200 chromatography. On the other hand, the α-synuclein transferred to the calmodulin–agarose was recovered from the precipitate by treating the beads with 1 m NaCl in 20 mm phosphate buffer. The gel mixture was directly loaded on the top of the sephadex column. All the collected fractions were analyzed with the Tricine/SDS/PAGE and shifts in the elution positions of α-synuclein were compared with each other. As controls, Sepharose CL-6B and ubiquitin–agarose were employed during the competition experiment, which provided the same amount of either gel beads or number of moles of immobilized protein as the calmodulin–agarose, respectively.
Calmodulin-induced α-synuclein release from membranes
In order to examine calmodulin-induced α-synuclein release from its membrane interaction, the α-synuclein-bound liposome was treated with soluble calmodulin in either Ca2+-bound or Ca2+-free form and the distribution of α-synuclein between soluble and membrane fractions was evaluated with western blotting analysis. After the liposome prepared with total 5 mg lipids between PA andPCata1 : 1 mass ratio was incubated with 7 µmα-synuclein in 20 mm sodium phosphate, pH 7.2, for 4 h at 4°C in a final volume of 1 mL with continuous shaking, the liposome bound by α-synuclein was separated from the free protein by centrifugation at 100 000 × g for 20 min. Following resuspension of the precipitate with 500 µL of the phosphate buffer, the liposome (120 µL) was treated with 5.7 µm calmodulin in either Ca2+-bound or Ca2+-free form for 1 h at room temperature. Calcium–calmodulin complex was prepared via a desalting procedure of the preincubated mixture between 5.7 µm calmodulin and 1.6 mm CaCl2 in 20 µL phosphate buffer by employing spun column packed with sephadex G-25 and dehydrated at 500 gfor 1 min. After loading the mixture, the complex was recovered with the same centrifugation at 500 gfor 1 min. The samples incubated between the liposome and the calmodulins were subjected to ultracentrifugation at 100 000 × g for 1 h. Supernatants and precipitates were separately analyzed with 10–20% Tricine/SDS/PAGE and the western blotting with synuclein-1 mAb. For the experiment employing the calmodulin peptide inhibitor, the liposome (120 µL) was incubated with either 10 µm ApoCaM or Ca2+–CaM complex pretreated with or without the inhibitor for 1 h at room temperature. The calmodulin–inhibitor mixture was prepared by treating calmodulin with the peptide inhibitor at a 1 : 2 (inhibitor/calmodulin) molar ratio for 1 h at 37°C. The amounts of liposome-bound α-synuclein in the precipi- tates obtained in the presence and absence of either calmodulin or calmodulin–inhibitor complex were analyzed using western blotting.
Calmodulin interaction of the synucleins
α-Synuclein was predicted to form class A2 amphipathic α-helices found in the exchangeable apolipoproteins (Davidson et al. 1998) and actually shown to experience dramatic structural alteration upon membrane interaction to form 80% of α-helicity from its ‘natively unfolded’ structure (Weinreb et al. 1996; Kim 1997). Since calmodulin has been known to recognize its target protein through the basic amphipathic α-helix (Sanyal et al. 1988; O'Neil et al. 1989; O'Neil and DeGrado 1990), a physiological function of α-synuclein has been examined with respect to its calmodulin regulation and possible implications in cellular messenger system. The molecular interaction between calmodulin and α-synuclein was examined by using a chemical coupling reagent of EEDQ in the presence and absence of 5 mm calcium. Following the cross-linking step, the covalent heterodimeric complex formed between calmodulin and α-synuclein was visualized on a gel of 10–20% Tricine/SDS/PAGE with silver staining (Fig. 1a). Although α-synuclein itself showed a small amount of dimeric band (lane 1) as reported in other studies (Yoshimoto et al. 1995; Paik et al. 1998), the high molecular weight species with an approximate Mr of 36 kDa became apparent only when α-synuclein (calculated molecular weight of 14.5 kDa) and calmodulin (calculated molecular weight of 16.8 kDa) were combined and incubated with each other in the absence of calcium. The band intensity of the complex was increased proportionally as the calmodulin concentration was raised in the presence of the constant amount of α-synuclein (Fig. 1a, lanes 2–5). In addition, calmodulin itself did not form the higher molecular weight species with EEDQ at any concentrations employed in the experiments regardless of calcium although the protein appeared in doublet which was suspected to be due to intramolecular cross-linking by the coupling reagent (Fig. 1d). This fact demonstrated that the high molecular weight species with Mr of 36 kDa was indeed a heterodimeric 1 : 1 complex between α-synuclein and calmodulin. The heterodimeric complex formation, however, required the cross-linker since the high molecular weight species which would be located in between self-dimers of either α-synuclein and calmodulin was not observed on the silver-stained gel resolving the 1 : 1 mixture (200 pmol each) of the proteins incubated in the absence of EEDQ (data notshown). Interestingly, the calmodulin interaction of α-synuclein turned out to be independent of calcium because the 1 : 1 complex was also produced even in the absence of calcium (Fig. 1a, lanes 6–9). On the other hand, however, the β-synuclein–calmodulin interaction was more dependent upon the presence of calcium as shown in Fig. 1(b). The 1 : 1 heterodimer formed between β-synuclein and calmodulin was apparent in the presence of calcium whereas its level was significantly reduced in the absence of the cation (Fig. 1b). In order to find out whether this different calcium dependency for calmodulin interaction between the synucleins was attributable to the most variable acidic C-termini between the isoforms, a C-terminally truncated α-synuclein named α–syn97 (calculated molecular weight of 9.7 kDa) lacking the C-terminus from residue 98 was subjected to the cross-linking with calmodulin in the presence and absence of calcium. As shown in Fig. 1(c), calmodulin still formed a heterodimeric 1 : 1 complex (Mr of 25 kDa) with α-syn97 even in the absence of calcium (lanes 6–9). It was also noticed that some cross-linked adducts with sizes larger than the heterodimer were also observed from various lanes wherever the heterodimers were enriched. These data clearly indicated that the calmodulin binding of α-synuclein was localized on the region of residues 1–97 where the amphipathic α-helices were predicted and the interaction did not require calcium whereas the divalent cation was crucial for the effective β-synuclein interaction of calmodulin. It was not clear, however, whether the binding sites of calcium–calmodulin complex (Ca2+–CaM) and calcium-free calmodulin, so-called apocalmodulin (ApoCaM), on α-synuclein were shared with each other.
The calmodulin interaction of α-synuclein was reassessed by observing calmodulin effect on the protein aggregation processes of both α- and β-synucleins in the absence of calcium. The aggregation of α-synuclein (∼ 2 mg/mL) was clearly diminished by calmodulin at a 1:10 molar ratio between calmodulin and α-synuclein (Fig. 2a) while the kinetics of β-synuclein was not affected at all (Fig. 2b). This also demonstrated the selective interaction between α-synuclein and Apo-CaM. In the presence of calcium, however, calmodulin interaction of α-synuclein was not appropriately estimated because calcium itself stimulated the aggregation of α-synuclein (Nielsen et al. 2001). Although the augmented aggregation was suppressed to more than half by the calmodulin, it was ambiguous whether this inhibitory effect was solely due to its pure α-synuclein interaction or deprivation of the stimulatory calcium (data not shown).
In order to evaluate physiological relevance of the interaction between calmodulin and α-synuclein, rat brain extract was subjected to pull-down assay for α-synuclein with the calmodulin immobilized on agarose. Proteins bound to the calmodulin–agarose were recovered and analyzed using 10–20% Tricine/SDS/PAGE. The presence of α-synuclein in the proteins was examined with western blotting by using synuclein-1 mAb. As shown in Fig. 3, α-synuclein was demonstrated to exist in the pull-down fraction of the brain extract not only in the absence (lane 2) but also in the presence of calcium (lane 4). To verify the specific interaction between calmodulin and α-synuclein, the calmodulin peptide inhibitor corresponding to calmodulin binding domain of CaM kinase II (residues Leu290-Ala309) was employed to make the agarose-bound calmodulin–inhibitor complex. When this complex was used instead during the pull-down assay, the amount of α-synuclein recovered from the precipitate was significantly reduced in the absence (Fig. 3, lane 3) and presence (Fig. 3, lane 5) of calcium, indicating that the calmodulin interaction of α-synuclein in the brain extract was indeed a biochemically specific phenomenon with a possibility of its physiological relevance.
Dissociation constants between calmodulin and the synucleins
Binding affinities between calmodulin and the synucleins were examined with a reversible binding assay, which also reconfirmed the different calcium requirement for the calmodulin interaction between α- and β-synucleins. After the synucleins were incubated with the immobilized calmodulin on agarose, the free and the bound synucleins were separated with a centrifugal filter (Ultrafree®-MC from Millipore Corporation, Bedford, MA, USA). The bound protein was recovered from the calmodulin–agarose with 1 m NaCl and quantitated by measuring absorbance at both 274 and 220 nm. The saturation curves obtained by plotting the bound synuclein against the total protein were converted into the double reciprocal plots shown in Fig. 4. The dissociation constants between calmodulin and α-synuclein in the presence and absence of calcium were basically identical, being 0.31 µm and 0.35 µm, respectively (Fig. 4a). This fact again reflected that α-synuclein was independent of calcium for its calmodulin interaction. On the other hand, β-synuclein preferentially interacted with calmodulin in the presence of calcium as determined by the fact that the dissociation constants varied from 0.68 µm to 10.1 µm in the presence and absence of the cation, respectively (Fig. 4b).
Competitive interaction of α-synuclein between calmodulin and membranes
Since calmodulin and membranes have shared the amphipathic α-helix of their target proteins for specific recognition, it might be plausible to assume that α-synuclein shown to induce the amphipathic helix upon membrane interaction could have physiological implications by possibly shuttling between calmodulin and membranes. This possibility wasassessed by an in vitro competition experiment of α-synuclein between liposome and the immobilized calmodulin–agarose. α-Synuclein was incubated with liposome prepared with PA and PC at a ratio of 1 : 1 (w/w) for 3 h at 4°C in 20 mm sodium phosphate, pH 7.2. The mixture was then subjected to sephadex G-200 gel permeation chromatography. The collected fractions were analyzed with 10–20% Tricine/SDS/PAGE to determine the elution position of α-synuclein. In the absence of calmodulin, most α-synuclein was bound to the membrane since the protein was coeluted with the liposome at the void volume in fractions between 6 and 10 (Fig. 5a). For the competition experiment, the membrane interaction of α-synuclein was carried out in the presence of the calmodulin–agarose at 1 : 1 molar ratio between the proteins in the absence of calcium. After removal of the calmodulin–agarose as a precipitate with a brief centrifugation, the supernatant was analyzed with the gel permeation chromatography. Apparently, the majority of α-synuclein was eluted at the significantly shifted fractions between 11 and 15 where free α-synuclein usually came out although small amount of the protein was still observed at the void volume (Fig. 5b). As controls, the membrane interaction of α-synuclein was performed in the presence of either calmodulin-free sepharose CL-6B or ubiquitin–agarose instead of the calmodulin–agarose. α-Synuclein was neither dissociated from the liposome nor transferred to the gel beads (data not shown). In addition, intactness of the liposome during the competition experiment was evaluated with a transmission electron microscope (Hitachi, H7100) after negatively staining the liposome adsorbed on carbon-coated copper grid with 2% uranyl acetate. The liposomes were not disrupted by the calmodulin–agarose during the incubation for 3 h at 4°C although an average size of the liposomes was reduced by about 30–40% (data not shown). These facts indicated that α-synuclein was actually released from the liposome via the competitive action of the immobilized calmodulin although a residual amount of α-synuclein still remained on the liposome. The specific interaction between α-synuclein and the calmodulin–agarose in the presence of membranes was further supported by the observation that α-synuclein previously bound to the liposome was transferred and remained bound to the immobilized calmodulin (Fig. 5c). The calmodulin-bound α-synuclein was recovered from the precipitate by treating the beads with 1 m NaCl and subjected to the chromatography. The α-synuclein was obtained from the fractions where the free protein was eluted (Fig. 5c). In the presence of calcium, α-synuclein also exhibited identical competitive interaction between calmodulin and the liposome although α-synuclein analyzed with the chromatography appeared rather heterogeneous on the gel, which might be due to calcium-mediated self-interaction of α-synuclein (data not shown).
Calmodulin-induced α-synuclein release from its membrane interaction
Since α-synuclein could associate into higher molecular weight species upon the membrane interaction, the liposome interaction and its release in the presence of either ApoCaM or Ca2+–CaM complex were directly assessed with soluble calmodulin and analyzed using western blotting by eliminating the gel permeation chromatography step. The α-synuclein bound liposome was prepared and treated with the calmodulins for 1 h at room temperature. Following the ultracentrifugation, distribution of α-synuclein between soluble and membrane fractions was estimated by separately analyzing the protein levels in the supernatant and the precipitate, respectively. The α-synuclein found in the precipitate as the liposome-bound form (Fig. 6a, lane 1) was released into the supernatant upon ApoCaM treatment (lane 4) as the protein level in the precipitate was decreased (lane 2) although some of α-synuclein was still spontaneously released from the membrane even in the absence of the calmodulin (lane 3). In the presence of Ca2+–CaM complex, virtually an identical release pattern of α-synuclein was obtained (Fig. 6b, lanes 3 and 4). In order to verify this α-synuclein release as a calmodulin-specific phenomenon, the calmodulin peptide inhibitor of CaM kinase II Leu290-Ala309 was employed to make calmodulin–inhibitor complex in the presence and absence of calcium. The α-synuclein-containing liposome was treated with either calmodulin or its inhibitor complex and the protein remained bound to the liposome was examined within the precipitates of the reaction mixtures. As shown in Fig. 6(c), α-synuclein on the liposome (lane 1) was released by ApoCaM (lane 2) as determined by decreased band intensity, but its release was suppressed by the ApoCaM-inhibitor complex (lane 3). In the presence of calcium (Fig. 6d), the release of α-synuclein by Ca2+–CaM complex (lane 2) was also prevented by the Ca2+–CaM-inhibitor complex (lane 3). As another control, it was also confirmed that α-synuclein was not released from the liposome with non-specific proteins such as RNase A and bovine serum albumin (data not shown). These observations clearly indicate that the α-synuclein release depends on the specific interaction with calmodulin. Taken together, these facts lead us to hypothesize that α-synuclein could act at the interface between membranes and soluble calmodulin in the presence and even absence of calcium through competitive interaction toward the amphipathic α-helices.
α-Synuclein has been demonstrated in this report to interact with ApoCaM as well as Ca2+–CaM complex. Calmodulin is known to experience dramatic structural transition upon calcium binding. ApoCaM exists in compact structure with a hydrophobic core while Ca2+–CaM complex exhibits a relaxed structure with two exposed hydrophobic faces (Zhang and Yuan 1998; Jurado et al. 1999). The Ca2+–CaM complex recognizes not only α-synuclein but also β-synuclein, indicating that its interaction sites could be shared by both synucleins. When primary structures of the synucleins are compared, there are a few common potential calmodulin-binding sites of basic amphipathic α-helical structures composed of basic and hydrophobic amino acids (Fig. 7). In fact, five putative amphipathic α-helices in α-synuclein predicted to be induced upon membrane interaction have been suggested (Davidson et al. 1998). Although exact calmodulin-binding sites are not determined, the four helices distributed from residue 1–60 are almost completely identical between α- and β-synucleins except the six residues marked in black (Fig. 7). Hence, it is likely that Ca2+–CaM complex might recognize these segments as the common binding site(s). On the other hand, ApoCaM binds to both α-synuclein and α-syn97, but not β-synuclein, indicating that the fifth helix (residue 61–94) distinctive for each synuclein with a deleted segment of 11 amino acids from residue 74–84 in β-synuclein might be responsible for the selective ApoCaM interaction although any contribution of the acidic C-termini or differences in local structures, if any, to the interaction should not be neglected. The fifth helix, a hydrophobic central region of α-synuclein (residues 61–95), was also known as the non-Aβ component of AD amyloid (NAC) and was originally isolated from the senile plaques of AD as the second major component (Ueda et al. 1993). The binding sites of Ca2+–CaM and ApoCaM on α-synuclein therefore could be independent from each other, although a possibility that they may share the common site(s) cannot be completely excluded. Based on these considerations, another possible function of ApoCaM could be suggested as a protector against the protein aggregation of α-synuclein because the missing segment of 11 amino acids (residue 74–84) in β-synuclein was demonstrated to provide aggregative potential to α-synuclein (Giasson et al. 2001). As a matter of fact, we have shown the suppressive effect of ApoCaM during the aggregation of α-synuclein whereas the β-synuclein aggregation was not affected (Fig. 2).
The physiological or pathological significance of the ApoCaM interaction of α-synuclein in addition to the Ca2+–CaM interaction is currently elusive. However, the ApoCaM interaction should not be underestimated from the aspect of its physiological relevance. Calmodulin exerts its multiple functions by interacting with numerous target proteins in the presence and even absence of calcium (Jurado et al. 1999). Its diverse cellular functions include metabolism, cytoskeletal dynamics, cell proliferation, cell–cell interaction and development (Zhang and Yuan 1998; Jurado et al. 1999). These phenomena have been mediated by not only the Ca2+–CaM complex but also by ApoCaM. Although the Ca2+–CaM complex interacts with at least 30 different enzymes and proteins including protein kinases, phosphatases and proteins involved in second messenger system, cytoskeleton, muscle and metabolism, ApoCaM binds to a distinct set of proteins such as actin-binding proteins, cytoskeletal and membrane proteins, enzymes, receptors and ion channels (Zhang and Yuan 1998; Jurado et al. 1999). In fact, most calmodulin in unstimulated cells would exist in the state of ApoCaM because calmodulin level in eukaryotic cell is approximately 1–10 µm while free calcium concentration in unstimulated cell is 0.1 µm or less (Jurado et al. 1999). The relatively large proportion of ApoCaM could make the α-synuclein interaction plausible because α-synuclein is also known as another abundant protein which amounts to 0.1% of total protein from rat brain extract (Shibayama-Imazu et al. 1993). The point which needs to be clarified, however, is the fact that α-synuclein interacted with not only ApoCaM but also Ca2+–CaM complex while β-synuclein predominantly recognized Ca2+–CaM complex because most calmodulin binding proteins have exhibited preferential affinity for either ApoCaM or Ca2+–CaM, but not both. Nevertheless, this differential calmodulin binding properties between α- and β-synucleins may have physiological or eventual pathological implications, if any, which could provide a clue to explain the exclusive localization of α-synuclein in the Lewy bodies inside the neurons of a PD patient whereas both synucleins are present with comparable amounts in their soluble forms (Jakes et al. 1994).
Calmodulin activation in vivo could be regulated by either intracellular calcium concentration or the availability of the protein. In this respect, α-synuclein could be considered as a negative regulator of calmodulin or calmodulin-buffering protein in the presence and absence of calcium. In fact, it was already predicted that there might be many calmodulin regulators belonging to this group in the nervous system (Slemmon et al. 1996). The moderate affinity between α-synuclein and calmodulin with Kd in ∼ µm range further incites us to suggest that this reversible interaction could make α-synuclein be possibly involved in regulation of total pool size of calmodulin (Slemmon et al. 1996). As a matter of fact, the moderate affinity between α-synuclein and calmodulin in the absence of calcium was reassessed by employing surface plasmonresonancespectroscopic measurement (SPR670 from Nippon Laser & Electronics Laboratory), which gave rise to Kd of 1.17 µm with ka = 4.21 × 102m−1 s−1 and kd = 4.90 × 10−4 s−1. Although the value was not exactly identical to the dissociation constants obtained in Fig. 2, validity of the moderate affinity between α-synuclein and calmodulin was reconfirmed. Taken together, it might be speculated that α-synuclein could act in cellular signaling processes by interacting with calmodulin during the mediation of calcium signaling. Based on the observation we made in this study, it is also tempting to consider that α-synuclein could exhibit its function at the interface between calmodulin and membranes.
This work was supported by Korea Research Foundation Grant KRF-99–041-F00040.