(Received for publication, January 29, 1997, and in revised form, March 7, 1997)
From the Department of Biochemistry and Howard Hughes Medical Institute, P.O. Box 357370, University of Washington, Seattle, Washington 98195-7370
Guanylyl cyclase-activating protein 2 (GCAP-2) is
a recoverin-like calcium-binding protein that regulates photoreceptor
guanylyl cyclase (RetGC) (Dizhoor, A. M., and Hurley, J. B. (1996)
J. Biol. Chem. 271, 19346-19350). It was reported
that myristoylation of a related protein, GCAP-1, was critical for its
affinity for RetGC (Frins, S., Bonigk, W., Muller, F., Kellner, R., and
Koch, K.-W. (1996) J. Biol. Chem. 271, 8022-8027). We
demonstrate that the N terminus of GCAP-2, like those of other members
of the recoverin family of Ca2+-binding proteins, is fatty
acylated. However, unlike other proteins of this family, more GCAP-2 is
present in the membrane fraction at low Ca2+ than at high
Ca2+ concentrations. We investigated the role of the
N-terminal fatty acyl residue in the ability of GCAP-2 to regulate
RetGCs. Myristoylated or nonacylated GCAP-2 forms were expressed in
Escherichia coli. Wild-type GCAP-2 and the Gly2
Ala2 GCAP-2 mutant, which is unable to undergo
N-terminal myristoylation, were also expressed in mammalian HEK293
cells. We found that compartmentalization of GCAP-2 in photoreceptor
outer segment membranes is Ca2+- and ionic
strength-sensitive, but it does not require the presence of the fatty
acyl group and does not necessarily directly reflect GCAP-2 interaction
with RetGC. The lack of myristoylation does not significantly affect
the ability of GCAP-2 to stimulate RetGC. Nor does it affect the
ability of the Ca2+-loaded form of GCAP-2 to compete with
the GCAP-2 mutant that constitutively activates RetGC. We conclude that
while Ca2+ binding plays a major regulatory role in GCAP-2
function, it does not operate through a calcium-myristoyl switch
similar to the one found in recoverin.
Ca2+ enters outer segments (OS)1 of vertebrate photoreceptors through cGMP-gated Na+/Ca2+ channels in the plasma membranes (reviewed in Refs. 3-5). In darkness these channels allow Ca2+ influx, but light-induced closure of the channels lowers free intracellular Ca2+ because Ca2+ is continuously extruded from the OS by a light-independent Na+/K+, Ca2+ exchanger. The estimate of the magnitude of this effect is that light lowers free intracellular Ca2+ from a dark level of 500 nM to as low as 50 nM (6). The decrease in free Ca2+ concentration allows guanylyl cyclase activator proteins (GCAPs) (7-11) to stimulate a membrane guanylyl cyclase (RetGC) (12, 13). Two homologous GCAPs have been identified in the retina: GCAP-1 (7, 9) and GCAP-2 (8, 10). Based on immunoblot and immunocytochemical analysis, GCAP-1 (11) and GCAP-2 (10) are both specific for retina and present in photoreceptor cells. The relative abundance of GCAPs in rods and cones and their relative contributions to regulation of RetGC in vivo still remain to be determined. At least two different GCAP-regulated RetGCs have been identified in mammalian photoreceptors, RetGC-1 (8, 14-18) and RetGC-2 (19, 20) (or GC-E and GC-F, respectively). It has been demonstrated by independent groups that GCAP-1 and GCAP-2 can activate RetGCs in OS membranes in vitro within the range of free Ca2+ concentrations corresponding to the estimated physiological range (1, 7-12, 19). Unlike peptide ligands that regulate other known membrane guanylyl cyclases via the cyclase extracellular domains (reviewed in Refs. 21 and 22), both GCAPs interact with RetGC via the cyclase intracellular domain (23, 24). GCAP-1 is able to stimulate recombinant RetGC-1 (11, 24), while GCAP-2 stimulates both recombinant RetGC-1 and RetGC-2 (8, 19, 23). However, the question which of the two cyclases can be a target for any particular GCAP in vivo has not yet been properly addressed. Additional factors may also be involved in regulation of RetGCs, such as sodium concentration in the intracellular medium (25), phosphorylation (25, 26), ATP binding (23, 27), or actin binding (28). Also, an unidentified S100 protein-like factor ("CDGCAP" (29)) can stimulate RetGC in vitro at free Ca2+ concentrations that are significantly higher than estimated physiological range. Among all factors that might influence RetGC activity only GCAP-1 and GCAP-2 have been shown to stimulate RetGCs at low Ca2+ concentrations and to impart Ca2+ sensitivity to the cyclases in vitro within the physiological range of 50-500 nM free Ca2+.
Unlike many other known members of the EF-hand superfamily (with the exception of GCAP-1), GCAP-2 acquires its activating conformation only at low Ca2+ concentrations. At free Ca2+ concentrations similar to the levels in dark-adapted vertebrate photoreceptors, GCAP-2 is not only unable to activate RetGCs, but it also inhibits RetGC. Therefore, GCAP-2 can be considered as a RetGC regulator rather than activator protein (1). Inactivation of EF hands makes GCAP-2 a constitutive activator of RetGC unable to undergo an "activator-to-inhibitor" transition (1).
GCAPs belong to the family of recoverin-like proteins that also includes a variety of neuronal Ca2+-binding proteins such as neurocalcin, hippocalcin, and others (9, 10, 30). Unlike distantly related calmodulin, recoverin-like proteins are C14-fatty acylated at their N termini. For mammalian retinal proteins, this acylation is heterogeneous (C14:0, C14:1, C14:2, and C12:0) (31-33). While all C14 derivatives altogether can constitute up to 75% of the total amount of fatty acyl residues, myristoyl itself represents only as much as 25% of it (32). Fatty acylation has been shown to impart to recoverin and to several recoverin-like proteins the ability to compartmentalize to membranes in the presence of calcium (34-38). The affinity of recoverin-like proteins for membranes increases as a result of fatty acyl group exposure in response to the protein transition into its calcium-bound conformation (35, 36), a mechanism referred to as a "calcium myristoyl switch" (35). The calcium-myristoyl switch in recoverin was directly demonstrated by measuring the accessibility of the fatty acylated N terminus to proteolytic cleavage (36) and by NMR (39, 40). Both GCAP-1 and GCAP-2 have a consensus sequence for the N-myristoylation encoded in their cDNAs (9, 10, 41). It was reported that GCAP-1 is fatty acylated (9), and the N-terminal fatty acyl group may be important for RetGC regulation; nonacylated GCAP-1 was found to have dramatically lower affinity for RetGC and lower Ca2+ sensitivity than the fatty acylated GCAP-1 (2).
In this study we investigated whether or not the N-terminal fatty acylation of GCAP-2 is essential for RetGC regulation. We found that GCAP-2 is fatty acylated but that its affinity for membranes is affected by Ca2+ in a manner opposite to that of other recoverin-like proteins. It associates with membranes at low Ca2+ concentrations more efficiently than at high free Ca2+ concentrations. We also demonstrate that neither membrane association of GCAP-2 nor various aspects of RetGC regulation by GCAP-2 require fatty acylation. We therefore conclude that Ca2+ induces a conformational change in GCAP-2 protein but that the calcium-myristoyl switch found in other members of the recoverin family is not essential for the ability of GCAP-2 to regulate RetGCs. We also describe an efficient bacterial expression system for producing functional myristoylated and nonacylated GCAP-2.
GCAP-2 was isolated from a
heat-stable fraction of retinal proteins using immunoaffinity
chromatography on monospecific polyclonal Np24 antibodies coupled to
CNBr-activated Sepharose 4B as described previously (10). Purified
GCAP-2 had no detectable GCAP-1 in it as assayed by immunoblot.
GCAP-2 was expressed in HEK293 cells
transiently transfected with GCAP-2 cDNA-containing vector using
calcium phosphate precipitation. Protein extracts from expressing and
control (vector only) transfected cells were made as described
previously (1, 10). The G2A mutant was generated by introducing a
GGG/GCG substitution into the second codon of GCAP-2 cDNA (10) by
Pfu polymerase-catalyzed polymerase chain reaction using
the "splice by overlap extension" approach (42). The G2A GCAP-2
mutant and a wild-type GCAP-2 were expressed in cell cultures
transfected and harvested simultaneously. Protein extracts from
expressing and control (vector only) transfected cells were
prepared, and GCAP-2 expression was estimated by immunoblot using
Np24 antibody as described previously (10).
The GCAP-2 cDNA coding region (10) was inserted
into the NcoI/BamHI sites of the pET11d vector
(Novagen) and expressed under control of the
isopropyl--D-thiogalactopyranoside-regulated T7 promoter
in the BL21(DE3)pLysS E. coli strain (Novagen) carrying a
pBB131 plasmid encoding N-myristoyl transferase (NMT) (a
gift from Dr. J. Gordon). To produce nonacylated GCAP-2, we used the same strain of cells but without the pBB131 plasmid. Fresh overnight 5-ml cultures of expression cells were diluted in 500 ml of standard LB
media containing 50 µg/ml ampicillin or both 50 µg/ml kanamycin and
50 µg/ml ampicillin (for pBB131-containing strains). Bacteria were
grown at 37 °C until the culture reached 0.2-0.3 OD units at 600 nm. Free myristic acid (100 mg/ml ethanol solution) was added into the
suspension of bacterial cells to 50 µg/ml 30 min before the induction
with isopropyl-
-D-thiogalactopyranoside as described
earlier for recoverin (36).
Isopropyl-
-D-thiogalactopyranoside was added to a final
concentration of 1 mM, and typically in 3.5 h
bacterial pellets were harvested by centrifugation at 8,000 × g for 20 min at 4 °C. The cells were disrupted by three
cycles of ultrasonication of 30 s each. The expressed GCAP-2 and
its mutants were always found in the insoluble fraction. The insoluble material was washed three times with a 20 mM Tris-HCl
buffer (pH 7.5) containing 1 mM EDTA, 7 mM
2-mercaptoethanol, 100 µM PMSF, and 20 µg/ml leupeptin
(buffer A) by centrifugation at 20,000 × g for 10 min.
GCAP-2 was extracted from the pellet by homogenization in buffer A
containing 100 mM mercaptoethanol, 1 mM EDTA,
and 6 M freshly deionized urea for 30 min at 4 °C and
dialyzed twice against 300-1000 volumes of buffer A overnight at
4 °C. Precipitate was removed by centrifugation at 30,000 × g for 10 minutes. Recombinant protein was then purified by
gel filtration on a Sephacryl S100 column in 10 mM Tris-HCl
buffer (pH 7.5) and 10 mM mercaptoethanol. Fractions
containing GCAP-2 were combined and concentrated using Amicon YM10
membranes under nitrogen pressure to 2 mg/ml final concentration.
Concentrated protein was used either immediately or after being quickly
frozen in small aliquots and stored at
70 °C. Myristoylated and
nonacylated EF(2/3/4)
GCAP-2 mutant, which carried
substitutions E80Q/E116Q/D158N (1) was produced using the same
protocol. For their functional comparisons, different GCAP-2 forms were
always expressed simultaneously and purified in parallel.
ESI-MS of GCAP-2 was performed essentially as described previously for recoverin using a Sciex API III triple quadrupole instrument (31), except that purified retinal or recombinant GCAP-2 was injected into the mass spectrometer during its elution from a capillary reverse-phase high pressure liquid chromatography C18 column connected to the mass spectrometer.
Membrane BindingFor GCAP-2 extraction experiments, a
fraction of OS was isolated from frozen dark-adapted bovine retinas
using sucrose gradient centrifugation (43), and rhodopsin concentration
was determined by measuring absorbance of an aliquot diluted in 1%
Ammonix LO detergent at 500 nm before and after bleaching, using
500 = 40,000. OS typically diluted at 1 mg/ml of
rhodopsin were homogenized in 10 mM Tris-HCl buffer
containing 5 mM MgCl2, 10 mM
2-mercaptoethanol, 100 µM PMSF, and 10 µg/ml of
leupeptin and aprotinin (buffer B) containing 1 mM Ca-EGTA
buffers using a Dounce glass-to-glass homogenizer equipped with pestle
B. Only freshly isolated membranes could be used for these experiments.
Homogenate was centrifuged at 80,000 rpm for 15 min at 4 °C in a
Beckman T 100.1 rotor. The supernatant was aspirated, and the pellet
was resuspended in the same volume of buffer. Ten microliters of
soluble and membrane fractions were mixed with an equal volume of
Laemmli SDS sample buffer containing 2 mM EGTA, boiled, and
loaded onto 12.5% SDS-polyacrylamide gel. The addition of EGTA in
sample buffer prevents the appearance of bands of
Ca2+-bound GCAP-2, which have higher mobility than its
Ca2+-free form. After electrophoresis, GCAP-2 bands were
transferred to a nitrocellulose sheet, stained with a
Np24 antibody,
and developed using goat anti-rabbit peroxidase conjugate and an
Amersham ECL chemiluminescent reagent. Fluorograms of GCAP-2 bands
within the quasilinear range of density were scanned using a Bio-Rad model GS-670 imaging densitometer and quantified using Bio-Rad Molecular Analyst software. For reconstitution experiments, OS membranes were first washed six times with buffer B and then incubated typically at 2.5 mg/ml rhodopsin with 2 µM recombinant
GCAP-2 and 1 mM Ca-EGTA in 200 µl for 30 min at room
temperature in buffer B containing 50 mM Tris-HCl. The
mixture was centrifuged as indicated above, and the pellet was gently
rinsed with the corresponding incubation buffer, resuspended in such
buffer, and repelleted. The final pellet typically containing
approximately 10% of total protein was resuspended in the original
volume of water, and aliquots of membranes were analyzed in
SDS-polyacrylamide gel as described above. Washed membranes incubated
without added GCAP-2 were used in these experiments as controls to
ensure that no endogenous protein was left in the membranes.
Fresh mouse liver was homogenized in 10 volumes of ice-cold 0.32 M sucrose containing 20 mM Tris-HCl (pH 7.5), 5 mM 2-mercaptoethanol, and 0.2 mM PMSF and centrifuged at 5000 × g for 15 min. The supernatant was recentrifuged at 45,000 × g for 30 min. Membranes were resuspended in the same buffer without sucrose and washed three times by centrifugation. Protein concentration was determined using a Bio-Rad BCA protein assay kit.
Preparation of LiposomesBovine retinas were homogenized in a blender with chloroform/methanol solution for 5 min at the following volume proportions: 1.5 retinal tissue, 3.73 methanol, 1.87 chloroform. After homogenization, water (1.9 volume) and chloroform (1.9 volume) were added to the mixture. In 20 min, the organic phase was collected, further separated from the remainders of aqueous phase by centrifugation at 10,000 × g for 20 min in glass centrifuge tubes, and collected using Pasteur pipettes. Lipids were dried under a constant stream of argon at room temperature, dispersed in liposome buffer (10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM MgCl2, 10 mM 2-mercaptoethanol), flushed with argon, and sonicated for 30 s. A fraction of liposomes sedimenting at 20,000 × g was collected by three cycles of centrifugation followed by resuspension of the pellet in a fresh portion of liposome buffer. The final pellet was resuspended in 200 µl of 10 mM Tris-HCl (pH 7.5) preflushed with argon, and 30 µl were taken for reconstitution with recombinant GCAP-2 and immunoblot analysis to ensure the absence of RetGC-1, RetGC-2, and rhodopsin.
The RetGC AssayThe RetGC assay was performed using
[-32P]GTP as a substrate and [8-3H]cGMP
as an internal standard. The reaction was carried out under infrared
light and analyzed by thin layer chromatography essentially as
previously described in detail (10).
Ca-EGTA buffers were calculated as in Ref. 10 and made strictly according to Ref. 44. Free Ca2+ concentrations were verified by a Ca2+-selective electrode and Rhod-2 titration.
ImmunoblotProteins separated by SDS-PAGE were transferred
onto a Millipore nitrocellulose sheet overnight at 100 mA.
Anti-rhodopsin monoclonal antibody 4D2 (a gift from Robert Molday,
University of British Columbia), monospecific rabbit antibody Np24
against GCAP-2 (10), antiserum against the RetGC-1
Glu641-Trp657 peptide derived from its kinase
homology domain (23),2 and rabbit antiserum
against the RetGC-2 Val169-Arg180 peptide
derived from its extracellular domain (19) were used to probe the
immunoblot and developed using peroxidase conjugate and an Amersham ECL
substrate.
It has been demonstrated that several members of the recoverin family (S-modulin, recoverin, neurocalcin, and hippocalcin) acquire stronger affinity to membranes in the presence of Ca2+ (34-38). In each case the membrane binding depends on the N-terminal fatty acylation (myristoylation). It has been found that in recoverin the N-terminal myristoyl group becomes exposed as a result of conformational changes caused by calcium binding (36, 39, 40), a mechanism known as "calcium-myristoyl switch" (35, 47).
GCAPs are the most closely related members of the recoverin family
(9-11, 48). However, unlike other members of this family, GCAP-2
associates with membranes more efficiently when Ca2+
concentration is lowered below 1 µM (Fig.
1). The distribution of GCAP-2 between the soluble and
membrane fractions depends on free Ca2+ concentrations.
GCAP-2 binds more strongly to membranes when the free Ca2+
concentration is below 200 nM. An increase in free
Ca2+ concentrations promotes dissociation of this protein
(Fig. 1, A and B). This behavior is opposite to
that of recoverin or other recoverin-like proteins (Fig. 1B;
see also Refs. 34-38). The relative distribution of GCAP-2 between the
soluble and membrane fractions also depends on ionic strength; GCAP-2
is less soluble at 100 mM NaCl than in hypotonic buffer.
However, even at this ionic strength Ca2+ decreases the
amount of GCAP-2 associated with membranes (Fig. 1C). This
effect cannot be attributed to the nonspecific influence of divalent
cations because we varied the free Ca2+ concentration
within the micromolar range, whereas MgCl2 was present at 5 mM in all of those experiments.
GCAP-2 Is a Fatty Acylated Protein
The N terminus of GCAP-2 (MGQQFS ... ; Ref. 10), contains a consensus sequence recognized by NMT (45). Met1 is removed during protein synthesis, and Gly2, the actual amino-terminal residue, is acylated. To verify that such modification takes place in vivo, we used ESI-MS to evaluate the exact molecular mass of GCAP-2 purified from bovine retina by immunoaffinity chromatography. The results shown in Table I demonstrate that GCAP-2 is C14-fatty acylated. The accuracy of this analysis performed on the whole protein does not allow us to conclude at the moment whether or not GCAP-2 has predominantly tetradecanoyl, tetradecaenoyl, or tetradecadienoyl fatty residues.
|
We have developed two expression systems for producing functional
myristoylated and nonacylated forms of GCAP-2. One system is based on
coexpression of GCAP-2 with NMT in E. coli that bears a
plasmid encoding NMT in the presence of exogenous free myristic acid
(Fig. 2, A and B). The molecular
mass of GCAP-2 coexpressed with NMT equals the calculated molecular
mass of the myristoylated protein and is identical to the molecular
mass of retinal protein. GCAP-2 expressed in E. coli without
NMT shows the molecular mass of nonacylated protein (Table I). We have
also observed that the myristoylated recombinant GCAP-2 has slightly
higher electrophoretic mobility than the nonacylated GCAP-2 (Fig.
2C). Both retinal and recombinant myristoylated GCAP-2 have
the same mobility, while nonacylated recombinant GCAP-2 appears in an
SDS-polyacrylamide gel as a protein approximately 1 kDa larger in size
(despite the fact that its molecular mass is 210 Da lower than that of
the myristoylated protein). Often a trace amount of nonmyristoylated protein can be observed in preparations of the myristoylated
recombinant protein (Fig. 2C, lane b). A low
level of signal from nonmyristoylated protein can be also detected by
ESI-MS. Typically, at least 90% of GCAP-2 coexpressed with NMT in
E. coli is myristoylated.
Myristoylated and nonacylated GCAP-2 expressed in human HEK 293 cell culture also have different electrophoretic mobilities (Fig. 2D). Because mammalian cells have an endogenous N-myristoyltransferase, to prevent myristoylation of GCAP-2, we expressed it as a mutant with the amino-terminal Gly2 substituted for Ala (the N-terminal Ala is not a substrate for NMT (45)). Similar to what was found for nonacylated GCAP-2 expressed in E. coli, the G2A mutant expressed in HEK293 cells has a lower electrophoretic mobility as compared with the wild-type protein.
Myristoylation Is Not Essential for GCAP-2 Association with MembranesGCAP-2 activates its target enzyme, a membrane guanylyl
cyclase below 500 nM free Ca2+ concentration
(8, 10, 19). Lowering Ca2+ concentration below 500 nM also promotes GCAP-2 association with the membrane (Fig.
1, A and B). It would therefore be tempting to
imagine that Ca2+-sensitive association of GCAP-2 with the
membranes might directly reflect its interaction with the cyclase.
Nevertheless, we find that recombinant GCAP-2 can associate with OS
membranes in a Ca2+-sensitive manner even when RetGC-1 and
RetGC-2 are destroyed by trypsin. However, since the antibodies that we
used were specific for regions relatively remote from the putative
transmembrane region of RetGCs, there was still a possibility that some
fragments of RetGCs including their transmembrane regions remain
associated with the membrane after the proteolysis and that may solely
account for the observed association of GCAP-2 with the membrane. Our finding that GCAP-2 is also capable of binding in a
Ca2+-sensitive manner to liver membranes (Fig.
3B) makes such a possibility unlikely. To
further exclude this possibility, we prepared liposomes from retinal
lipids and found that GCAP-2 can associate with these liposomes in a
Ca2+-sensitive manner, although both RetGC-1 and RetGC-2
were undetectable in the artificial membranes by immunoblot (Fig.
3C). The liposomes used for reconstitution did not even show
traces of such an abundant integral membrane protein as rhodopsin. It
indicates that GCAP-2 is able to interact not only with the cyclase,
but also with the lipid layer itself. We were unable to see GCAP-2
binding to liposomes prepared from commercial crude soybean
phospholipids (data not shown). Therefore, it is unlikely that the
aggregation of GCAP-2 would account for the observed phenomenon. This
result also indicates that lipid composition of vesicles may be
important for GCAP-2 association with the membranes.
We found that the overall process of GCAP-2 translocation to the membrane fraction is Ca2+-sensitive; however, it does not require fatty acylation. When myristoylated and nonacylated GCAP-2 were compared for their abilities to associate with membranes, we found that myristoylation of GCAP-2 was not a necessary prerequisite for its ability to interact with the membranes in a Ca2+-sensitive manner (Fig. 3D).
The interaction with the membrane most likely plays an important role in providing compartmentalization of GCAP-2 to the membranes, where it can reach membrane RetGC via lateral diffusion (Fig. 3E). GCAP-2 apparently interacts with the cyclase both in low and in high intracellular Ca2+ concentrations based on its ability to either inhibit or activate RetGC, respectively (1). Its binding to RetGC is reversible, because GCAP-2 can be washed off the membrane (8, 13). GCAP-2 has the ability to associate with the membrane and to form a complex with the cyclase itself both at high (RetGC inhibition) and at low Ca2+ concentrations (RetGC activation). The affinity of GCAP-2 for membranes increases when GCAP-2 is in its "apo" form and decreases upon binding Ca2+. The difference is more apparent at low ionic strength that promotes the dissociation of GCAP-2 from the membrane. However, it is most likely that in vivo for both forms of GCAP-2 the distribution between cytosol and membrane must be shifted toward the membrane association, because at normal ionic strength GCAP-2 tends to bind to membranes more efficiently than in hypotonic conditions (Fig. 1C). Calcium-myristoyl switch apparently does not play an essential role in regulating Ca2+ sensitivity of GCAP-2 compartmentalization to membranes. Nevertheless, because of the complex nature of GCAP-2 interaction with membrane, this does not allow us to conclude that the fatty acylation of GCAP-2 is not essential for its ability to regulate RetGC. That issue requires a direct functional comparison between the fatty acylated and nonacylated forms of GCAP-2 in a RetGC activation assay.
Nonacylated GCAP-2 Regulates RetGCsWe found that the myristoylated and the nonacylated forms of recombinant GCAP-2 are almost equivalent in their abilities to regulate RetGCs in washed OS membranes.
Nonacylated GCAP-2 expressed in E. coli (Fig.
4A) as well as nonacylated G2A GCAP-2 mutant
produced in human cell culture (insert) stimulated RetGCs
with an efficiency of at least 75% compared with the fatty acylated
GCAP-2. Ca2+ sensitivities of RetGC regulation by
myristoylated and nonacylated GCAP-2 were also very similar to each
other (Fig. 4B).
We previously demonstrated that inactivation of EF hands made GCAP-2 a
stable Ca2+-insensitive activator of RetGC (1). Fig.
5 shows that the ability of Ca2+-insensitive
GCAP-2 mutant to become a constitutive activator of RetGC does not
require fatty acylation. Myristoylated and nonacylated GCAP-2
(EF(2/3/4)) both stimulate RetGC even at high
Ca2+ concentrations and with similar EC50, too.
The regulatory properties of GCAP-2 include not only the ability of its
"apo" form to stimulate RetGCs at low intracellular
Ca2+ concentrations but also the ability of its
Ca2+-loaded form to inhibit activation of RetGCs (1). We
therefore tested the potential influence of the fatty acylation on this inhibitory property of GCAP-2. We found that Ca2+-loaded
myristoylated and nonacylated forms of GCAP-2 inhibit the stimulation
of RetGCs produced by a Ca2+-insensitive myristoylated
(EF(2/3/4)
) mutant equally well (Fig. 5B).
Conversely, myristoylated and nonacylated GCAP-2
(EF(2/3/4)
) mutants demonstrate similar abilities to
outcompete an inhibitory effect of Ca2+-loaded wild-type
GCAP-2 (Fig. 5C). Therefore, fatty acylation of GCAP-2 is
not a critical element for RetGC regulation.
We conclude that myristoylation of GCAP-2 is not essential for its function as a regulator of RetGCs. It has been demonstrated that GCAP-2-related protein, recoverin, undergoes a conformational change called "calcium-myristoyl switch" (reviewed in Ref. 47) in which the fatty acyl group is exposed in response to calcium binding. Such a mechanism increases the affinity of recoverin and recoverin-related proteins for membranes. We find that Ca2+ sensitivity of GCAP-2-membrane interaction is opposite to that reported for recoverin and a number of other recoverin-like proteins (34-38). In the case of GCAP-2, we also find that all major aspects of its interaction with photoreceptor membranes and with its target RetGCs are apparently determined almost entirely by Ca2+-induced changes in the protein moiety of the molecule and are not significantly influenced by the myristoyl group. It has been observed (2, 41) that a nonacylated recombinant form of GCAP-1 had much lower (near 40-fold) affinity to RetGCs and lower Ca2+ sensitivity compared with fatty acylated GCAP-1 (2). GCAP-1 and GCAP-2 are most closely related to each other, both genetically (48) and functionally (1, 7-11). Given our observations, it is not immediately apparent why such highly homologous proteins as GCAP-1 and GCAP-2 would demonstrate such a difference in terms of their dependence on fatty acylation. The difference in dependence may reflect a subtle difference between GCAP-1 and GCAP-2 in the structures of their activating domains. However, since we demonstrate here that the fatty acyl group itself is not an essential element of RetGC regulation by a calcium-binding protein, it also seems possible that the fatty acylation may merely favor the proper folding of recombinant GCAP-1 rather than its interaction with RetGC.
Also, despite the fact that fatty acylation of GCAP-2 is not essential for RetGC regulation in vitro, we cannot exclude the possibility that it might be involved in some intracellular processes such as, for example, translocation of newly synthesized GCAP-2 to the proper cell compartment. Otherwise it would be difficult to explain why the myristoylation signal in GCAP-2 was not eliminated in the course of evolution of recoverin-like proteins.
We thank Rich Johnson and Greg Niemi for performing ESI MS analysis of GCAP-2, Kenneth Walsh for allowing us to use his mass spectrometry facilities and John Glomset for allowing us to use his argon stream evaporator.