Interaction of Calmodulin with Striatin, a WD-repeat Protein Present in Neuronal Dendritic Spines*

Rat striatin, a quantitatively minor protein belonging to the WD-repeat family of proteins, is a Ca2+/calmodulin-binding protein mostly expressed in the striatum and in the motor and olfactory systems (Castets, F., Bartoli, M., Barnier, J. V., Baillat, G., Salin, P., Moqrich, A., Bourgeois, J. P., Denizot, F., Rougon, G., Calothy, G., and Monneron, A. (1996) J. Cell. Biol. 134, 1051–1062). Generally associated with membranes, striatin is mostly found in dendritic spines where it is likely to play a role in Ca2+-signaling events. In this paper, we characterize its calmodulin-binding properties. By using deletion mapping and site-directed mutagenesis, we identified the sequence located between amino acids 149 and 166 as the main calmodulin-binding site. The predicted corresponding peptide is potentially able to form a basic amphiphilic helix, as is often the case for many known calmodulin-binding sites. Calmodulin binding to striatin is Ca2+-dependent, with half-maximal binding occurring around 0.5 μm free Ca2+. In the presence of Ca2+, the equilibrium dissociation constant of calmodulin/striatin fusion protein complex is 40 ± 5 nm. We also show that brain striatin subcellular localization, as studied by tissue fractionation, is Ca2+-dependent, this effect being probably mediated by calmodulin. Our results are in agreement with the hypothesis that striatin is a transducer involved in Ca2+signaling or an adapter protein involved in regulating macromolecular assemblies within dendritic spines.

Striatin is a quantitatively minor, Ca 2ϩ /calmodulin (Ca 2ϩ / CaM 1 )-binding protein of M r 86,225, mostly expressed in neurons of the motor and olfactory structures of the central nervous system (1,34). Cloning of the human homologous gene revealed that it is a highly conserved protein (96% identity, 98% similarity) (35). The C-terminal half of the protein sequence consists in a series of 8 WD repeats. WD repeats are found in a number of transducing proteins, the prototype of which is the ␤-subunit of heterotrimeric G proteins, and in proteins assisting in the assembly of multiprotein complexes (2). In synaptosomal preparations, striatin behaves as a mem-brane-associated protein, although its peptide sequence does not display hydrophobic stretches able to account for transmembrane domains nor signatures for myristoylation or farnesylation.
Striatin is exclusively expressed in the somatodendritic compartment of neurons, at the exclusion of axons. It is mostly found in dendritic spines, which are the postsynaptic compartments of excitatory synapses. Spines are tiny protrusions that stud the surface of dendrites (3). They are dynamic structures, their number and location varying with the ongoing neuronal activity (4,5). Ca 2ϩ plays a major role in these structures, which are considered to be independent subcellular Ca 2ϩ compartments (5-7). They contain glutamate N-methyl-D-aspartic acid receptors, which are major Ca 2ϩ ionophores (8). Not surprisingly, they also contain large amounts of CaM (9) and CaM-binding proteins, principally Ca 2ϩ /CaM kinase type II and, in significant amounts, calcineurin, adenylyl cyclases, and cAMP phosphodiesterases (reviewed in Refs. 3 and 10), as well as cytoskeletal proteins (myosin, fodrin 1, ␣-actinin) (11,12).
Striatin, a component of spines, has molecular features suggesting that it, too, is involved in Ca 2ϩ signaling events. We have shown previously that striatin directly interacts with CaM in the presence of Ca 2ϩ (1). In this study, we mapped the CaM-binding site of striatin and determined its affinity for CaM. We also show that Ca 2ϩ , most probably through its interaction with CaM, induces a partial subcellular redistribution of striatin. Our results are in agreement with the hypothesis that striatin is an adapter or transducer protein involved in Ca 2ϩ /CaM-dependent events within dendritic spines.

EXPERIMENTAL PROCEDURES
Construction of Expression Plasmids-Plasmids used to produce fragments of striatin fused to the C terminus of glutathione S-transferase (GST) were constructed using the pGEX-KT expression vector (13). First, plasmid pGEX-NcoI, with the new unique restriction sites NcoI, NdeI, and KpnI, was constructed by inserting an oligonucleotide (5Ј-G-ATCCATGGGACATATGGGTACCG-3Ј) and its complementary strand (5Ј-AATTCGGTACCCATATGTCCCATG-3Ј) between the BamHI and EcoRI sites of pGEX-KT. The 1.95-kilobase pair NcoI-EcoRI fragment encompassing codons 1-650 of striatin (obtained from plasmid pFCC, see Ref. 1) was inserted between the corresponding sites of pGEX-NcoI, yielding pGEX-striatin 1-650. Plasmid pGEX-striatin 1-427, which codes GST fused to the first 427 residues of striatin, was constructed from pGEX-striatin 1-650: the plasmid was digested by StuI and EcoRI, blunted with T4 DNA polymerase in the presence of the four dNTPs, and ligated (the EcoRI site is restored upon ligation). Plasmid pGEX-striatin 1-98, which codes GST fused to the first 98 residues of striatin, was constructed similarly by deleting the BglII-EcoRI fragment of pGEX-striatin 1-650. To produce a fusion protein encoding the whole striatin sequence, the NcoI to BamHI striatin fragment obtained from pFCC (in which a BamHI site immediately follows the STOP codon) was inserted into the NcoI-and BglII-digested plasmid pGEXstriatin 1-427. The resulting plasmid, called pGEX-striatin 1-780, codes GST fused for the entire striatin coding sequence (Fig. 1).
Expression and Purification of Fusion Proteins-The plasmids were transformed into the Escherichia coli XL-1 Blue strain. 200-ml cultures of bacterial cells were grown until 0.3 A 600 was reached. 0.1 mM isopropyl-␤-D-thiogalactopyranoside was then added, and incubation went on for 12 h at 18°C. Cells collected by centrifugation were resuspended in 4 ml of lysis buffer (20 mM Hepes, pH 7.2, 150 mM NaCl, 1 mM phenylmethylsulfonyl chloride). Cells were frozen at Ϫ70°C, thawed, and disrupted in a French press. The lysate was centrifuged at 12,000 ϫ g, at 4°C for 15 min. The 4-ml supernatant was mixed with 2 ml of a 50% glutathione-Sepharose 4B slurry (Amersham Pharmacia Biotech) previously equilibrated with lysis buffer and incubated for 2 h at 4°C. The suspension was centrifuged 5 min at 500 ϫ g and the unbound proteins discarded. The resin was washed twice with 20 ml of lysis buffer. Fusion proteins bound to the matrix were eluted with 3 ml of lysis buffer containing 10 mM reduced glutathione.
Site-directed Mutagenesis-To modify the first putative CaM-binding site of striatin 1-427, plasmid pGEX-striatin 1-427 was digested by BglII and ligated with the oligonucleotide GATCCG, resulting in the insertion of the dipeptide Pro-Asp between Asp-97 and Leu-98, in the middle of the putative CaM-binding site 1: 88 RKGQENKKDPDLVR-RIKML 107. The resulting protein is called striatin 1-427 mut 1. To modify the second putative CaM-binding site of striatin 1-427, a modified recombinant polymerase chain reaction was used for mutagenesis (14). Briefly, the primers 5Ј CAGCAGAACAGCCAGTTCATGG*GGAA-GCC*AGGCC 3Ј and 5Ј ATGAACTGGCTGTTCTGCTGTGGCTGCAC-CTCCG 3Ј were designed for recombinant polymerase chain reaction. The complementary regions are underlined. In these primers, striatin Trp-155 was changed to Gly and Gln-158 was changed to Pro (the corresponding codons are indicated by an asterisk). The resulting sequence reads: 149 QNSQFMGKPGRQLLRQYL 166. The obtained plasmid, pGEX-striatin 1-427 mut 2, contains a third BglI site (GCC-*AGGCCGGC), allowing to check for the expected mutations. To introduce this mutation in the whole fusion protein, we constructed plasmid pGEX-striatin 1-780 mut 2 by inserting the XbaI fragment of pGEXstriatin 1-780 in the XbaI-linearized pGEX-striatin 1-427 mut 2. The protein expressed by this plasmid is called GST-striatin 1-780 mut 2. All constructs were verified by sequencing.
Calmodulin Overlay-CaM was biotinylated using the immunoprobe TM biotinylation kit from Sigma. Purified fusion proteins were separated on SDS-polyacrylamide (7.5%) gels and transferred onto nitrocellulose. Nonspecific binding sites were blocked with 2% non-fat dry milk in TBS buffer (50 mM Tris-HCl, pH 8, 150 mM NaCl) for 1 h. The nitrocellulose sheets were then incubated in TBS buffer containing 1% dry milk and 50 nM biotinylated CaM (bCaM), for 12 h at 4°C (a 1% milk solution is about 3 mM Ca 2ϩ ). After three 5-min washes in the same buffer, the blots were incubated for 1 h at room temperature with a 1:5000 dilution of ExtrAvidin-alkaline phosphatase (Sigma) in the same buffer. After three washes, the blots were revealed by the nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate chromogenic system. The same procedure was achieved in the presence of 5 mM EGTA.
Purification of the GST-striatin Fusion Proteins by CaM-Sepharose Affinity Chromatography-E. coli extracts were prepared in lysis buffer containing 2 mM Ca 2ϩ . Each lysate was incubated with 100 l of CaM-Sepharose 4B slurry for 1 h at 4°C. The resins were washed with 1 ml of the same buffer. Bound proteins were eluted with 100 l of lysis buffer containing 5 mM EGTA. Aliquots of the lysates, unretained fractions, and eluates were run on SDS-polyacrylamide gels (7.5 and 9%) and transferred onto nitrocellulose. Western blots were revealed using a rabbit serum directed against a fusion protein containing the 50 first amino acids of striatin (serum 75) and the ECL procedure (Pierce).
Determination of the Affinity of GST-striatin 1-427 for Ca 2ϩ /CaM by a Microtitration Assay-To determine the dissociation constants (K D ) values of the complexes composed of the various striatins 1-427 and Ca 2ϩ /CaM, we used an adaptation of the competitive enzyme-linked immunosorbent assay method described by Friguet et al. (15), which allows to determine true equilibrium constants in solution. Biotinylated CaM, at a constant concentration, is incubated in solution with various amounts of a given ligand (GST-striatin in the present case) until the equilibrium is reached. The amount of free bCaM in the liquid phase is then monitored by an indirect binding assay, in which the liquid phase is transferred for a short period of time (15 min) into a microtitration plate coated with the catalytic domain of Bordetella pertussis adenylyl cyclase. Adenylyl cyclase binds CaM with high affinity (K D 0.1 nM) (16). bCaM bound to the plate is detected by addition of ExtrAvidin-peroxidase conjugate. To ensure the validity of the procedure, two conditions must be fulfilled (discussed in detail in Ref. 15). One is that the peroxidase activity measured is proportional to the amount of free bCaM in the liquid phase. We checked that it was indeed the case within the range of 0.03-4 nM bCaM. The other is that no readjustment of the equilibrium in the liquid phase occurs during the incubation in the adenylyl cyclase-coated wells. This is achieved by minimizing the incubation time. We checked that during a 15-min incubation time, less than 10% of total free bCaM bound the adenylyl cyclase-coated wells.
Practically, bCaM (1 nM) was incubated with various concentrations (0 -200 nM) of wild type or mutated GST-striatins 1-427 in buffer (20 mM Hepes, pH 7.2, 100 mM NaCl, 1 mg/ml trypsin inhibitor) containing either 100 M CaCl 2 (buffer B-CaCl 2 ) or 1 mM EGTA (buffer B-EGTA) for 15 h at room temperature. The mixtures were then transferred to enzyme-linked immunosorbent assay plates coated with 5 g/ml recombinant adenylyl cyclase solution (purified as in Ref. 16) and saturated with 10 mg/ml trypsin inhibitor. After 15-min incubation at room temperature, the wells were washed three times with the corresponding buffer. A 1:1000 dilution of ExtrAvidin-peroxidase (Sigma) was added to the wells and incubated 1 h at room temperature. After three washes, the adenylyl cyclase-bound bCaM was revealed by the o-phenylenediamine chromogenic system (Sigma) and the absorbance read on a Bio-Rad microplate reader model 550. The fraction of free bCaM in each sample is given by the ratio of absorbance measured for this sample to the absorbance measured in samples containing only bCaM. In these experiments, we checked that sequential 15-min incubations of the same mixture in different adenylyl cyclase-coated wells did not decrease the amount of the measured free bCaM by more than 10%.
Dependence upon Ca 2ϩ of the Formation of a Striatin⅐CaM Complex-Ca 2ϩ /EGTA buffers containing 100 mM Hepes, pH 7.20, at 20°C, 0.1% trypsin inhibitor, and various free Ca 2ϩ concentrations: 50, 10, 5, 2, 1, 0.5, and 0.1 M were prepared according to Tsien and Pozzan (17), using the fluorescent dye indicator fura-2 (Molecular Probes) and a Perkin-Elmer spectrofluorimeter. To set Ca 2ϩ concentrations above 10 M, known amounts of total Ca 2ϩ were added. The Ca 2ϩ dependence of the interaction of CaM with either brain cytosol striatin (see below) or striatin 1-427 was determined by using the procedure described by Warr et al. (18). 5 g of striatin 1-427 or 400 g of brain cytosol proteins, in buffer containing Ն50 M free Ca 2ϩ , were incubated with 50 l of CaM-Sepharose slurry for 1 h at 4°C. Unbound protein was removed by extensive washing in the same buffer. The resin was incubated, for 10 min at 4°C, twice with 1 ml of the same buffer and then step by step with Ca 2ϩ /EGTA buffers of decreasing Ca 2ϩ concentration, ending with buffer containing no Ca 2ϩ and 2 mM EGTA. The proteins eluted after incubation with each Ca 2ϩ /EGTA buffer were bound onto nitrocellulose by using a Schleicher and Schuell slot blotter Minifold. Striatin was revealed by Western blotting using an anti-striatin serum (see above) and ECL detection. Precise quantification of striatin was achieved by densitometric analysis of the immunoreactive bands, using a specially designed software (Biocom analysis system).
Subcellular Fractionation of Rat Brain-The brains of 2-week-old rats were homogenized, in the minute following decapitation, using a glass-Teflon Potter-Elvehjem homogenizer, in ice-cold 50 mM Tris-HCl buffer, pH 7.5, containing 150 mM NaCl, 5 mM MgCl 2 , inhibitors of proteases (as in Ref. 1), and either 1 mM EGTA (EGTA buffer) or 100 M Ca 2ϩ (Ca 2ϩ buffer). The homogenates were centrifuged at 100,000 ϫ g, at 4°C for 1 h. The supernatants, referred to as EGTA-cytosol or as Ca 2ϩ -cytosol, were saved. The pellets were solubilized as described (1) in 50 mM Tris-HCl buffer containing 0.9% Lubrol-PX (Lubrol/protein ratio of 2.5) and either 1 mM EGTA or 100 M Ca 2ϩ and centrifuged at 100,000 ϫ g for 30 min at 4°C. The EGTA-and Ca 2ϩ -Lubrol-soluble protein fractions were saved and the pellets resuspended by homogenization in the corresponding Lubrol buffers. Protein was determined according to Schaffner and Weissmann (19). The various fractions were analyzed on SDS-polyacrylamide (7.5%) gels and transferred to nitrocellulose. The blots were stained with Ponceau red, striatin was revealed by the ECL procedure, and the Western blots were scanned. The same experiments were repeated four times with the same results.

Deletion Mapping of the CaM-binding Site(s) of Striatin-In
a previous study, we showed that recombinant striatin expressed in E. coli bound CaM in a Ca 2ϩ -dependent manner (1). To localize the CaM-binding site(s), we produced in E. coli various fragments of striatin and determined their CaM-binding capabilities. To facilitate their purification, the various fragments were genetically fused to glutathione S-transferase. A schematic representation of the different constructs is shown in Fig. 1. The fusion proteins were produced in E. coli at a low growth temperature. In these conditions, the recombinant proteins were found in the soluble bacterial extracts and could be purified by a single chromatographic step on glutathione beads. Fig. 2A shows a SDS-polyacrylamide gel electrophoresis analysis of the purified preparations. The CaM binding properties of the various proteins were tested by a CaM-overlay procedure, in which the proteins were separated on SDS gels, transferred to nitrocellulose, and probed with bCaM. Fig. 2B indicates that bCaM bound in a Ca 2ϩ -dependent manner GSTstriatin 1-427 (lane 1) and GST-striatin 1-780 (lane 5) but not GST- striatin 1-98 (lane 4). In a different set of experiments, the various GST-striatin fusion proteins were tested for their ability to bind CaM-Sepharose in the presence of Ca 2ϩ . As shown in Fig. 3, GST-striatin 1-780 (box A) and GST-striatin 1-427 (box B) bound CaM-Sepharose in the presence of Ca 2ϩ and were eluted in the presence of EGTA. GST-striatin 1-98 (box F) did not bind CaM-Sepharose, whether in the presence or absence of Ca 2ϩ . Altogether, these results suggest that the main CaM-binding site(s) is (are) located in the central region of striatin, between residues 98 and 427.
Determination of the Striatin CaM-binding Site by Site-directed Mutagenesis-Many of the characterized CaM-binding sites consist in a basic amphiphilic helical peptide (20). Examination of the amino acid sequence of striatin reveals the presence of two such potential basic amphiphilic helices located from residues 88 to 105 (site 1) and residues 149 -166 (site 2). To determine whether one of these two regions is indeed involved in CaM binding, each site was modified so as to disrupt its potential ability to bind CaM. For site 1, the dipeptide Pro-Asp was inserted between Asp-97 and Leu-98, yielding GST-striatin 1-427 mut 1. This insertion was expected to both decrease the ability of this segment to form an ␣-helical structure and to abolish its amphipathicity. Two residues of site 2, Trp-155 and Gln-158, were substituted, respectively, by Gly and Pro, these substitutions being expected to both decrease the propensity of this segment to form an ␣-helical structure and to decrease the hydrophobic interactions involved in CaM binding. These substitutions were introduced in both GSTstriatin 1-427 and GST-striatin 1-780. The CaM-binding properties of the two modified proteins were analyzed by the CaMoverlay technique. As shown in Fig. 2B (lanes 3 and 6), GSTstriatin mut 2 did not bind bCaM in the presence of Ca 2ϩ , whereas GST-striatin 1-427 mut 1 (lane 2) bound bCaM in a Ca 2ϩ -dependent manner, just as did the wild-type GST-striatins. Binding assays on CaM-Sepharose confirmed that GSTstriatin mut 1 (box C) bound CaM in a Ca 2ϩ -dependent manner, whereas both GST-striatins mut 2 (boxes D and E) had lost the property to bind Ca 2ϩ /CaM. Since GST-striatin 1-780 mut 2 is unable to bind CaM in the presence of Ca 2ϩ , the peptide sequence 149 -166 represents the main CaM-binding site of striatin.
Determination of the Affinity and Ca 2ϩ Dependence of Striatin for Ca 2ϩ /CaM-The dissociation constants (K D ) of striatin 1-427 wild-type and striatins 1-427 mut 1 and mut 2 for CaM were measured by a microtitration assay, which allows the determination of the true affinity constant of a given complex in solution (see "Experimental Procedures"). By this method, the K D measured for the GST-striatin 1-427⅐CaM complex in the presence of Ca 2ϩ was 40 Ϯ 5 nM (Fig. 4A). The affinity of striatin 1-427 mut 1 for bCaM in the presence of Ca 2ϩ was identical to that of wild-type striatin (Fig. 4B), whereas that of striatin 1-427 mut 2 was drastically reduced, if not abolished (Fig. 4C).
The Ca 2ϩ dependence of the interactions of CaM with both GST-striatin 1-427 and cytosol brain striatin were determined. Both proteins bound CaM-Sepharose in the presence of 50 M of Ca 2ϩ . The Ca 2ϩ concentrations at which both brain cytosol striatin and striatin 1-427 dissociated from CaM-Sepharose occurred at a similar Ca 2ϩ concentration, around 0.5 M (Fig. 5).

Subcellular Fractionation of Brain Shows a Partial Translocation of Striatin as a Function of Ca 2ϩ
Concentration-In a number of cell types, certain sets of proteins are known to translocate from one subcellular compartment to another, often as a function of changes in distribution and concentration of second messengers. We tested whether the pattern of striatin distribution within brain subcellular fractions could be modified as a function of Ca 2ϩ /CaM concentration. Subcellular fractionation of rat brain homogenates was achieved in the presence or absence of Ca 2ϩ , yielding cytosol proteins, Lubrolsoluble proteins, and Lubrol-insoluble proteins (described under "Experimental Procedures") ( Fig. 6A). Whether the homogenization was performed in the presence of EGTA or of Ca 2ϩ , the cytosols contained 25 Ϯ 5% of the total protein, the Lubrol-soluble proteins 58 Ϯ 5%, and the Lubrol-insoluble pellets 17 Ϯ 3%. The relative amount of striatin in each fraction was assessed by scanning of Western blots (Fig. 6B). When the homogenate was prepared in the presence of 1 mM EGTA, striatin was present in highest amount in the Lubrol-soluble fraction (Fig. 6C, lane 2), but was also found in the cytosol fraction (lane 3) and in the Lubrol-insoluble fraction (lane 1). When the homogenate was made in the presence of 100 M Ca 2ϩ (lanes 4 -6), the Lubrol-insoluble fraction did not contain any striatin (lane 6). Densitometric analysis of the autoradiograms showed that in the latter case, the striatin content of the cytosol was almost twice that measured in the absence of Ca 2ϩ (Fig. 6C). The same experiments were repeated four times with the same results.

DISCUSSION
Striatin has been the first member of the WD-repeat family of proteins known to interact with the ubiquitous calcium sensor protein, CaM (1). The WD repeats have been identified in numerous signaling proteins, and it has been suggested that this motif is involved in their multiple interactions with target proteins, within a regulatory network. The function of striatin is currently unknown, but its ability to interact with CaM as well as the presence of WD repeats in its sequence suggest that this protein is involved in Ca 2ϩ -dependent signaling at the postsynaptic membrane level. Recently, the ␤-subunit of transducin, a heterotrimeric G protein, and the prototype of the WD-repeat family, was also shown by Liu and colleagues to directly bind Ca 2ϩ /CaM (21). The authors suggest that Ca 2ϩ / CaM-binding to G␤ differentially modulates the interaction of G␤ with its multiple effectors. We hypothesize that CaM binding to striatin could similarly affect its interaction with its targets.
The CaM-binding domain of striatin, called domain or site 2, has been defined in this study by directed mutagenesis and expression of fusion proteins. It includes residues 149 -166 and conforms to the frequently encountered CaM-binding motif consisting in a basic amphiphilic helix. In the CaM-binding peptide of myosin light chain kinase, which typifies this motif, two aromatic residues, separated by 12 amino acids, are re-sponsible for anchoring the peptide to the N-and C-terminal halves of CaM, respectively (22). In the case of striatin, it is likely that the aromatic residues Phe-153 and Tyr-165, separated by 11 amino acids, also anchor striatin to the N-and C-terminal halves of CaM. Some proteins have two or multiple CaM-binding domains, for example the Trpl ion channels (18), the ␣1-syntrophin (23), a CaM-stimulated phosphodiesterase (24), the NR1 subunit of N-methyl-D-aspartic acid receptors (25). The possibility that striatin possesses two CaM-binding domains appears unlikely since the GST-whole length striatin fusion protein no longer binds CaM when domain 2 is mutated. This mutation was designed in such a way as to decrease the ability of this region to form an ␣-helix and to prevent the hydrophobic interactions involved in CaM binding. In the case of the ␤-subunit of transducin, Liu and colleagues (21) determined that the CaM binding corresponded to residues 40 -63. Interestingly, residues 53-63 belong to the first WD repeat of the seven such repeats that constitute most of the protein. The CaM-binding domain of the ␤-subunit of transducin is there- fore part of the propeller-like structure revealed by crystallography (26,27). In striatin, on the contrary, the CaM-binding domain is away from the WD-repeat domain.
Striatin, a quantitatively minor, membrane-associated protein, is predominantly found in the dendritic spines of a few subsets of neurons (1). The Ca 2ϩ concentration requirement for CaM binding of the many postsynaptic CaM-binding proteins is variable, reflecting the fact that the intracellular concentration of Ca 2ϩ within dendritic spines varies in a highly dynamic way (6). Here we show that the affinity of the brain striatin⅐CaM complex for Ca 2ϩ is 0.1-0.5 M. Hence striatin reversibly binds Ca 2ϩ /CaM, depending upon the fluctuations of Ca 2ϩ concentration in spines. We measured the K D value of the complex made up of a GST-striatin fusion protein and of Ca 2ϩ /CaM. This K D is of the order of 40 nM, it falls in the same range as that measured for syntrophins (20 -100 nM) (23) and one of the two CaM-binding sites of the N-methyl-D-aspartic acid glutamate receptors (R1 subunit) (73 nM) (25).
It has been shown in some types of neurons that, following elevation of intracellular Ca 2ϩ and the ensuing binding of CaM to its targets, a certain percentage of CaM and of CaM-binding proteins is redistributed from membrane to cytosol and vice versa (28 -31). The translocating proteins would thereby cycle between various states of conformation and/or activity. Although the reality of these phenomena is beyond doubt, estimates of the percentage of translocated proteins under such circumstances are not quite reliable (29,32). In this study, we compared brain striatin subcellular localization in tissue extracts realized either in the presence or absence of Ca 2ϩ . In the absence of Ca 2ϩ , striatin was present in all subfractions, including the Lubrol-insoluble fraction, containing postsynaptic densities (10) and cytoskeletal elements. In the presence of 100 M Ca 2ϩ , on the contrary, striatin was no longer present in this fraction. Such an in vitro differential distribution of striatin is likely to be mediated by CaM, since striatin is not known to directly bind Ca 2ϩ . It is tempting to imagine that in vivo a shift in striatin distribution would also occur as a result of Ca 2ϩ fluctuations. Such a study could perhaps be attempted, using immunocytochemistry at the electron microscope level, on NT2 fully differentiated neurons, theoretically able to establish synapses (Stratagene). However, at the stage at which synapses are said to occur, we found that the cells were aggregated, the neurites completely entangled, and no conclusion could be drawn.
The expression of striatin in only a few well defined subsets of neurons renders the elucidation of the physiological role of striatin a difficult task. Techniques aiming at decreasing its expression in neuronal cells are currently being used. Future research can now take advantage of the possibility to study neurons expressing high levels of striatin whose CaM-binding site has been disrupted. Indeed, the regulation by Ca 2ϩ /CaM of the function of striatin will have to be appreciated, since this protein is expressed in structures in which neurodegenerative processes might well be related to Ca 2ϩ -dependent excitotoxicity (33).