- Email: [email protected]
Calcium- and myristoyl-dependent subcellular localization of the neuronal calcium-binding protein VILIP in transfected PC12 cells
Neuroscience Letters 225 (1997) 126–128
Calcium- and myristoyl-dependent subcellular localization of the neuronal calcium-binding protein VILIP in transfected PC12 cells Christina Spilker, Eckart D. Gundelfinger, Karl-Heinz Braunewell* Department of Molecular Biology and Neurochemistry, Federal Institute for Neurobiology, P.O. Box 1860, D-39008 Magdeburg, Germany Received 14 February 1997; accepted 6 March 1997
Abstract Wild-type neuronal calcium-binding protein VILIP (visinin-like protein), and a myristoylation mutant of VILIP which lacks the consensus sequence for N-terminal myristoylation, have been stably transfected in PC12 cells. Immunocytochemical studies of VILIPtransfected PC12 cells have revealed that wild-type VILIP is strongly concentrated at the cell membrane, particularly at cell-cell contact sites, but is also distributed throughout the cytosol at moderate levels. In contrast, myristoylation-mutant VILIP shows a more even distribution, with significantly less association at cell-cell contact sites. Western blot analysis of subcellular fractions has shown that wild-type VILIP associates in a calcium-dependent manner with membrane fractions, whereas the myristoylation mutant only weakly associates with this fraction. Therefore, a calcium-myristoyl switch seems to be a major, but not sole determinant for the association of VILIP with membranes in living cells. 1997 Elsevier Science Ireland Ltd. Keywords: Calcium-binding protein; Calcium-myristoyl switch; Cell transfection; Myristoyl mutant; PC12; Subcellular localization
Myristoylation of proteins has been proposed as a mechanism for the association of molecules with cell membranes, which is often required for their functional activity [8]. Several members of the family of neuronal calciumsensor proteins (NCS), including recoverin [3], hippocalcin [4], neurocalcin d [5] and VILIP (visinin-like protein) [2] have been shown to be N-terminally myristoylated. These NCS proteins are believed to associate with membrane structures via the mechanism of a calcium-myristoylswitch, where calcium binding of the molecule changes the protein conformation leading to an exposure of the hydrophobic myristoyl-residue. This in turn causes an interaction of the protein with the cell membrane [12]. For hippocalcin [4], neurocalcin d [5] and recoverin [4,9] a calcium-dependent membrane interaction has been reported in biochemical assays in vitro which is solely dependent on myristoylation of the molecule. In the case of other myristoylated proteins, such as the myristoylated alanin-rich C kinase substrate (MARCKS), myristoylation is sufficient for membrane association in vitro, but in transfected cells additional protein components are necessary for membrane * Corresponding author. Tel.: +49 391 6263203; fax: +49 391 6263229; e-mail: [email protected]
association [11]. We were therefore interested whether in the case of VILIP the myristioyl residue is necessary and sufficient to equip the molecule with a myristoyl-switch mechanism in vitro, and whether myristoylation is also sufficient to localize the molecule to membranes in living cells. To investigate this issue wild-type VILIP and a myristoylation mutant of VILIP were stably transfected in PC12 cells and the interaction of the two isoforms with membranes was investigated. Undifferentiated pheochromocytoma (PC12) cells, which do not express VILIP at detectable levels, were cultured in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum (Gibco), 2 mM glutamine and penicillin/streptomycin (100 mg/ml) in a humidified 95% air, 5% CO2 incubator. Wild-type VILIP or a non-myristoyable mutant VILIP (myristoylation consensus site, MGKQNSKLAP, [10], of VILIP was changed: glycin-2 to alanine-2) were amplified from the VILIP cDNA [6] by polymerase chain reaction using oligonucleotide primers with appropriate restriction sites [2]. The amplified DNAs were cloned into an isopropyl-b-d-thiogalactoside (IPTG) inducible mammalian expression system (pOPRSVI CAT, lac-operator containing vector; p3′SS, lac-repressor containing vector; Stratagene) [2]. The constructs were transfected into PC12
0304-3940/97/$17.00 1997 Elsevier Science Ireland Ltd. All rights reserved PII S0304-3940 (97 )0 0201-2
C. Spilker et al. / Neuroscience Letters 225 (1997) 126–128
cells using the calcium phosphate precipitation method. Stable transfectants were cultured in medium containing the selective markers geneticin and hygromycin (both 500 mg/ml). Several stably transfected PC12 cell clones were isolated. In this study the clones PC12.2 expressing wildtype VILIP and PC12M2 expressing myristoylation mutant VILIP were used. Other cell clones yielded identical results. PC12 cells were cultured on chamber slides and the expression of VILIP was induced overnight with IPTG (5 mM). The cells were fixed with 4% paraformaldehyde and 0.2% Triton X-100 in phosphate-buffered saline (PBS) and the distribution of VILIP was detected with a rabbit antiserum against VILIP in combination with a goat anti-rabbit dichlorotriazinyl-amino-fluorescein (DTAF)-conjugated secondary antibody (Dianova). Fluorescence was visualized with a Leica fluorescence microscope. In addition to a moderate cytosolic staining wild-type VILIP was detected at membrane structures of PC12 cells (Fig. 1A). A strong expression of VILIP was observed particularly at sites where the membranes of two adjacent cells adhere (see arrowheads). In contrast, the non-myristoylated mutant of VILIP showed a rather even distribution in the cells (Fig.
Fig. 1. Immunocytochemical localization of wild-type VILIP and myristoylation mutant VILIP in stably transfected PC12 cells. PC12.2 (A) and PC12M2 (B) cells grown on chamber-slides were analyzed by indirect fluorescein-immunofluorescence for wild-type VILIP (A) and for the myristoylation mutant of VILIP (B) using affinity-purified anti-VILIP antibodies. Arrowheads indicate sites of cell-cell contacts. Bar, 25 mm.
127
Fig. 2. Western blot analysis of the calcium-dependent subcellular localization of wild-type VILIP and myristoylation mutant VILIP in stably transfected PC12 cells. Subcellular fractionation of transfected PC12.2 and PC12M2 cells was performed in the presence (+) and absence (−) of calcium. Soluble (SN, lanes 1, 2, 5 and 6) and crude membrane (P, lanes 3, 4, 7 and 8) fractions of homogenates of PC12.2 (lanes 1–4) and PC12M2 (lanes 5–8) cells were prepared in the presence of 1 mM CaCl2 (lanes 1, 3, 5 and 7) or 2 mM EGTA (lanes 2, 4, 6 and 8). The proteins of the fractions were applied to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto nitrocellulose. Affinity-purified anti-VILIP antibodies were used to detect VILIP contents.
1B). However, at some cell-cell contacts (see arrowheads) a distinct staining was observed, which resembled the wildtype VILIP staining but was clearly less pronounced. In PC12 cells a differential subcellular distribution of wildtype VILIP and the non-myristoylated mutant was observed, suggesting that myristoylation is one major factor for the membrane association and subcellular localization of the protein in living cells. In order to investigate whether VILIP can calcium-dependently associate with the membrane in PC12 cells, we performed subcellular fractionation of transfected cells. Confluent cultures of stably transfected PC12 cells were induced (5 mM IPTG, overnight) to express wild-type VILIP or the myristoylation mutant VILIP. The cells were then washed once in 5 mM Tris–HCl (pH 7.5), 0.1 M NaCl and 1 mM phenylmethylsulfonyl fluoride and centrifuged at 800 × g for 5 min in a Heraeus centrifuge. For translocation assays, the buffer was supplemented with either 2 mM EGTA or 1 mM CaCl2, and the cells were homogenized by sonification in 1 ml buffer. The homogenate was centrifuged at 15 000 × g for 20 min in a Biofuge and the pellet was re-homogenized in 1 ml buffer containing 2 mM EGTA or 1 mM CaCl2. The supernatants (soluble protein fraction) and the pellet fractions (crude membrane fraction) were subjected to Western blot analysis. In the absence of calcium, wild-type VILIP was detected mainly in the soluble fraction of PC12 cells. A small amount of VILIP immunoreactivity was observed in the crude membrane fraction (Fig. 2, lanes 2 and 4). In the presence of 1 mM calcium, significant amounts of VILIP immunoreactivity shifted from the soluble to the crude membrane fraction (Fig. 2, lanes 1 and 3). These results suggest that calcium binding leads to a translocation of VILIP from the cytosol to the membrane fraction. In the case of the myristoylation mutant PC12M2, nearly all VILIP immunoreactivity was associated with the soluble fraction both in the presence and absence of calcium (Fig. 2, lanes 5–8). In this case, calcium did not induce a translocation of non-myristoylated VILIP from the cytosolic to the membrane fraction (Fig. 2, lanes 5 and 7). However, a
128
C. Spilker et al. / Neuroscience Letters 225 (1997) 126–128
still detectable amount of VILIP reactivity was associated with the crude membrane fraction both, in the presence and absence of calcium (Fig. 2, lanes 7 and 8). The data of this study, when taken together, indicate that N-terminal myristoylation is likely to anchor VILIP to the membrane. Myristoylation appears to be one major prerequisite for the localization of VILIP to cell membranes in living PC12 cells. This membrane interaction seems to be dependent on the calcium-binding of VILIP, as was shown by calcium-dependent subcellular localization, and therefore fulfills the criteria of the calcium-myristoyl-switch. In addition, the residual binding of myristoylated and non-myristoylated VILIP to cell membranes at cell contact sites in living cells as well as in vitro may be explained by an interaction of VILIP with specific membrane-associated binding partners. Immunocytochemical studies have revealed a co-localization of VILIP with the cortical actinbased cytoskeleton in PC12 cells [7]. The observed interaction of VILIP with actin in vitro and the co-localization of VILIP with the actin-based cytoskeleton [1,7] may explain the observed residual myristoyl-independent interaction of VILIP with the cell membrane and/or its associated cytoskeleton. The authors would like to thank Heidemarie Wickborn for expert technical assistance and to Dr. Denise ManahanVaughan for critical reading of the manuscript. This work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. [1] Braunewell, K.-H., Lenz, S.E. and Gundelfinger, E.D., VILIP (visinin-like protein). In M.R. Celio (Ed.), Guidebook to the CalciumBinding Proteins, Oxford University Press, Oxford, UK, 1996, pp. 107–109.
[2] Braunewell, K.-H., Spilker, C., Behnisch, T. and Gundelfinger, E.D., The neuronal calcium-sensor protein VILIP modulates cAMP accumulation in stably transfected C6 glioma cells: amino-terminal myristoylation determines functional activity, J. Neurochem., (1997) in press. [3] Dizhoor, A.M., Chen, C.-K., Olshevskaya, E., Sinelnikova, V.V., Phillipov, P. and Hurley, J.B., Role of the acylated amino terminus of recoverin in Ca2 + -dependent membrane interaction, Science, 259 (1993) 829–832. [4] Kobayashi, M., Takamatsu, K., Saitoh, S. and Nogushi, T., Myristoylation of hippocalcin is linked to its membrane association properties, J. Biol. Chem., 268 (1993) 18898–18904. [5] Ladant, D., Calcium and membrane binding properties of bovine neurocalcin delta expressed in Escherichia coli, J. Biol. Chem., 270 (1995) 3179–3185. [6] Lenz, S.E., Henschel, Y., Zopf, D., Voss, B. and Gundelfinger, E.D., VILIP, a cognate protein of the retinal calcium-binding proteins visinin and recoverin, is expressed in the developing chicken brain, Mol. Brain Res., 15 (1992) 133–140. [7] Lenz, S.E., Braunewell, K.-H., Weise, C., Nedlina-Chittka, A. and Gundelfinger, E.D., The neuronal EF-hand calcium-binding protein VILIP: interaction with cell membrane and actin-based cytoskeleton, Biochem. Biophys. Res. Commun., 225 (1996) 1078–1083. [8] McLaughlin, S. and Aderem, A., The myristoyl-electrostatic switch: a modulator of reversible protein-membrane interactions, Trends Biochem. Sci, 20 (1995) 272–276. [9] Tanaka, T., Ames, J.B., Harvey, T.S., Stryer, L. and Ikura, M., Sequestration of the membrane-targeting myristoyl group of recoverin in the calcium-free state, Nature, 376 (1995) 444–447. [10] Towler, D.A., Gordon, J.I., Adams, S.P. and Glaser, L., The biology and enzymology of eukaryotic protein acylation, Annu. Rev. Biochem., 57 (1988) 69–99. [11] Swierczynski, S.L. and Blackshear, P.J., Membrane association of the myristoylated alanine-rich C kinase substrate (MARCKS) protein, J. Biol. Chem., 270 (1995) 13436–13445. [12] Zozulya, S. and Stryer, L., Calcium-myristoyl protein switch, Proc, Natl. Acad. Sci. USA, 89 (1992) 11569–11573.
Calcium- and myristoyl-dependent subcellular localization of the neuronal calcium-binding protein VILIP in transfected PC12 cells Christina Spilker, Eckart D. Gundelfinger, Karl-Heinz Braunewell* Department of Molecular Biology and Neurochemistry, Federal Institute for Neurobiology, P.O. Box 1860, D-39008 Magdeburg, Germany Received 14 February 1997; accepted 6 March 1997
Abstract Wild-type neuronal calcium-binding protein VILIP (visinin-like protein), and a myristoylation mutant of VILIP which lacks the consensus sequence for N-terminal myristoylation, have been stably transfected in PC12 cells. Immunocytochemical studies of VILIPtransfected PC12 cells have revealed that wild-type VILIP is strongly concentrated at the cell membrane, particularly at cell-cell contact sites, but is also distributed throughout the cytosol at moderate levels. In contrast, myristoylation-mutant VILIP shows a more even distribution, with significantly less association at cell-cell contact sites. Western blot analysis of subcellular fractions has shown that wild-type VILIP associates in a calcium-dependent manner with membrane fractions, whereas the myristoylation mutant only weakly associates with this fraction. Therefore, a calcium-myristoyl switch seems to be a major, but not sole determinant for the association of VILIP with membranes in living cells. 1997 Elsevier Science Ireland Ltd. Keywords: Calcium-binding protein; Calcium-myristoyl switch; Cell transfection; Myristoyl mutant; PC12; Subcellular localization
Myristoylation of proteins has been proposed as a mechanism for the association of molecules with cell membranes, which is often required for their functional activity [8]. Several members of the family of neuronal calciumsensor proteins (NCS), including recoverin [3], hippocalcin [4], neurocalcin d [5] and VILIP (visinin-like protein) [2] have been shown to be N-terminally myristoylated. These NCS proteins are believed to associate with membrane structures via the mechanism of a calcium-myristoylswitch, where calcium binding of the molecule changes the protein conformation leading to an exposure of the hydrophobic myristoyl-residue. This in turn causes an interaction of the protein with the cell membrane [12]. For hippocalcin [4], neurocalcin d [5] and recoverin [4,9] a calcium-dependent membrane interaction has been reported in biochemical assays in vitro which is solely dependent on myristoylation of the molecule. In the case of other myristoylated proteins, such as the myristoylated alanin-rich C kinase substrate (MARCKS), myristoylation is sufficient for membrane association in vitro, but in transfected cells additional protein components are necessary for membrane * Corresponding author. Tel.: +49 391 6263203; fax: +49 391 6263229; e-mail: [email protected]
association [11]. We were therefore interested whether in the case of VILIP the myristioyl residue is necessary and sufficient to equip the molecule with a myristoyl-switch mechanism in vitro, and whether myristoylation is also sufficient to localize the molecule to membranes in living cells. To investigate this issue wild-type VILIP and a myristoylation mutant of VILIP were stably transfected in PC12 cells and the interaction of the two isoforms with membranes was investigated. Undifferentiated pheochromocytoma (PC12) cells, which do not express VILIP at detectable levels, were cultured in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum (Gibco), 2 mM glutamine and penicillin/streptomycin (100 mg/ml) in a humidified 95% air, 5% CO2 incubator. Wild-type VILIP or a non-myristoyable mutant VILIP (myristoylation consensus site, MGKQNSKLAP, [10], of VILIP was changed: glycin-2 to alanine-2) were amplified from the VILIP cDNA [6] by polymerase chain reaction using oligonucleotide primers with appropriate restriction sites [2]. The amplified DNAs were cloned into an isopropyl-b-d-thiogalactoside (IPTG) inducible mammalian expression system (pOPRSVI CAT, lac-operator containing vector; p3′SS, lac-repressor containing vector; Stratagene) [2]. The constructs were transfected into PC12
0304-3940/97/$17.00 1997 Elsevier Science Ireland Ltd. All rights reserved PII S0304-3940 (97 )0 0201-2
C. Spilker et al. / Neuroscience Letters 225 (1997) 126–128
cells using the calcium phosphate precipitation method. Stable transfectants were cultured in medium containing the selective markers geneticin and hygromycin (both 500 mg/ml). Several stably transfected PC12 cell clones were isolated. In this study the clones PC12.2 expressing wildtype VILIP and PC12M2 expressing myristoylation mutant VILIP were used. Other cell clones yielded identical results. PC12 cells were cultured on chamber slides and the expression of VILIP was induced overnight with IPTG (5 mM). The cells were fixed with 4% paraformaldehyde and 0.2% Triton X-100 in phosphate-buffered saline (PBS) and the distribution of VILIP was detected with a rabbit antiserum against VILIP in combination with a goat anti-rabbit dichlorotriazinyl-amino-fluorescein (DTAF)-conjugated secondary antibody (Dianova). Fluorescence was visualized with a Leica fluorescence microscope. In addition to a moderate cytosolic staining wild-type VILIP was detected at membrane structures of PC12 cells (Fig. 1A). A strong expression of VILIP was observed particularly at sites where the membranes of two adjacent cells adhere (see arrowheads). In contrast, the non-myristoylated mutant of VILIP showed a rather even distribution in the cells (Fig.
Fig. 1. Immunocytochemical localization of wild-type VILIP and myristoylation mutant VILIP in stably transfected PC12 cells. PC12.2 (A) and PC12M2 (B) cells grown on chamber-slides were analyzed by indirect fluorescein-immunofluorescence for wild-type VILIP (A) and for the myristoylation mutant of VILIP (B) using affinity-purified anti-VILIP antibodies. Arrowheads indicate sites of cell-cell contacts. Bar, 25 mm.
127
Fig. 2. Western blot analysis of the calcium-dependent subcellular localization of wild-type VILIP and myristoylation mutant VILIP in stably transfected PC12 cells. Subcellular fractionation of transfected PC12.2 and PC12M2 cells was performed in the presence (+) and absence (−) of calcium. Soluble (SN, lanes 1, 2, 5 and 6) and crude membrane (P, lanes 3, 4, 7 and 8) fractions of homogenates of PC12.2 (lanes 1–4) and PC12M2 (lanes 5–8) cells were prepared in the presence of 1 mM CaCl2 (lanes 1, 3, 5 and 7) or 2 mM EGTA (lanes 2, 4, 6 and 8). The proteins of the fractions were applied to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto nitrocellulose. Affinity-purified anti-VILIP antibodies were used to detect VILIP contents.
1B). However, at some cell-cell contacts (see arrowheads) a distinct staining was observed, which resembled the wildtype VILIP staining but was clearly less pronounced. In PC12 cells a differential subcellular distribution of wildtype VILIP and the non-myristoylated mutant was observed, suggesting that myristoylation is one major factor for the membrane association and subcellular localization of the protein in living cells. In order to investigate whether VILIP can calcium-dependently associate with the membrane in PC12 cells, we performed subcellular fractionation of transfected cells. Confluent cultures of stably transfected PC12 cells were induced (5 mM IPTG, overnight) to express wild-type VILIP or the myristoylation mutant VILIP. The cells were then washed once in 5 mM Tris–HCl (pH 7.5), 0.1 M NaCl and 1 mM phenylmethylsulfonyl fluoride and centrifuged at 800 × g for 5 min in a Heraeus centrifuge. For translocation assays, the buffer was supplemented with either 2 mM EGTA or 1 mM CaCl2, and the cells were homogenized by sonification in 1 ml buffer. The homogenate was centrifuged at 15 000 × g for 20 min in a Biofuge and the pellet was re-homogenized in 1 ml buffer containing 2 mM EGTA or 1 mM CaCl2. The supernatants (soluble protein fraction) and the pellet fractions (crude membrane fraction) were subjected to Western blot analysis. In the absence of calcium, wild-type VILIP was detected mainly in the soluble fraction of PC12 cells. A small amount of VILIP immunoreactivity was observed in the crude membrane fraction (Fig. 2, lanes 2 and 4). In the presence of 1 mM calcium, significant amounts of VILIP immunoreactivity shifted from the soluble to the crude membrane fraction (Fig. 2, lanes 1 and 3). These results suggest that calcium binding leads to a translocation of VILIP from the cytosol to the membrane fraction. In the case of the myristoylation mutant PC12M2, nearly all VILIP immunoreactivity was associated with the soluble fraction both in the presence and absence of calcium (Fig. 2, lanes 5–8). In this case, calcium did not induce a translocation of non-myristoylated VILIP from the cytosolic to the membrane fraction (Fig. 2, lanes 5 and 7). However, a
128
C. Spilker et al. / Neuroscience Letters 225 (1997) 126–128
still detectable amount of VILIP reactivity was associated with the crude membrane fraction both, in the presence and absence of calcium (Fig. 2, lanes 7 and 8). The data of this study, when taken together, indicate that N-terminal myristoylation is likely to anchor VILIP to the membrane. Myristoylation appears to be one major prerequisite for the localization of VILIP to cell membranes in living PC12 cells. This membrane interaction seems to be dependent on the calcium-binding of VILIP, as was shown by calcium-dependent subcellular localization, and therefore fulfills the criteria of the calcium-myristoyl-switch. In addition, the residual binding of myristoylated and non-myristoylated VILIP to cell membranes at cell contact sites in living cells as well as in vitro may be explained by an interaction of VILIP with specific membrane-associated binding partners. Immunocytochemical studies have revealed a co-localization of VILIP with the cortical actinbased cytoskeleton in PC12 cells [7]. The observed interaction of VILIP with actin in vitro and the co-localization of VILIP with the actin-based cytoskeleton [1,7] may explain the observed residual myristoyl-independent interaction of VILIP with the cell membrane and/or its associated cytoskeleton. The authors would like to thank Heidemarie Wickborn for expert technical assistance and to Dr. Denise ManahanVaughan for critical reading of the manuscript. This work was supported by the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. [1] Braunewell, K.-H., Lenz, S.E. and Gundelfinger, E.D., VILIP (visinin-like protein). In M.R. Celio (Ed.), Guidebook to the CalciumBinding Proteins, Oxford University Press, Oxford, UK, 1996, pp. 107–109.
[2] Braunewell, K.-H., Spilker, C., Behnisch, T. and Gundelfinger, E.D., The neuronal calcium-sensor protein VILIP modulates cAMP accumulation in stably transfected C6 glioma cells: amino-terminal myristoylation determines functional activity, J. Neurochem., (1997) in press. [3] Dizhoor, A.M., Chen, C.-K., Olshevskaya, E., Sinelnikova, V.V., Phillipov, P. and Hurley, J.B., Role of the acylated amino terminus of recoverin in Ca2 + -dependent membrane interaction, Science, 259 (1993) 829–832. [4] Kobayashi, M., Takamatsu, K., Saitoh, S. and Nogushi, T., Myristoylation of hippocalcin is linked to its membrane association properties, J. Biol. Chem., 268 (1993) 18898–18904. [5] Ladant, D., Calcium and membrane binding properties of bovine neurocalcin delta expressed in Escherichia coli, J. Biol. Chem., 270 (1995) 3179–3185. [6] Lenz, S.E., Henschel, Y., Zopf, D., Voss, B. and Gundelfinger, E.D., VILIP, a cognate protein of the retinal calcium-binding proteins visinin and recoverin, is expressed in the developing chicken brain, Mol. Brain Res., 15 (1992) 133–140. [7] Lenz, S.E., Braunewell, K.-H., Weise, C., Nedlina-Chittka, A. and Gundelfinger, E.D., The neuronal EF-hand calcium-binding protein VILIP: interaction with cell membrane and actin-based cytoskeleton, Biochem. Biophys. Res. Commun., 225 (1996) 1078–1083. [8] McLaughlin, S. and Aderem, A., The myristoyl-electrostatic switch: a modulator of reversible protein-membrane interactions, Trends Biochem. Sci, 20 (1995) 272–276. [9] Tanaka, T., Ames, J.B., Harvey, T.S., Stryer, L. and Ikura, M., Sequestration of the membrane-targeting myristoyl group of recoverin in the calcium-free state, Nature, 376 (1995) 444–447. [10] Towler, D.A., Gordon, J.I., Adams, S.P. and Glaser, L., The biology and enzymology of eukaryotic protein acylation, Annu. Rev. Biochem., 57 (1988) 69–99. [11] Swierczynski, S.L. and Blackshear, P.J., Membrane association of the myristoylated alanine-rich C kinase substrate (MARCKS) protein, J. Biol. Chem., 270 (1995) 13436–13445. [12] Zozulya, S. and Stryer, L., Calcium-myristoyl protein switch, Proc, Natl. Acad. Sci. USA, 89 (1992) 11569–11573.