Ca2+-myristoyl switch in the neuronal calcium sensor recoverin requires different functions of Ca2+-binding sites.

Abstract

Recoverin is an EF-hand Ca(2+)-binding protein that is suggested to control the activity of the G-protein-coupled receptor kinase GRK-1 or rhodopsin kinase in a Ca(2+)-dependent manner. It undergoes a Ca(2+)-myristoyl switch when Ca(2+) binds to EF-hand 2 and 3. We investigated the mechanism of this switch by the use of point mutations in EF-hand 2 (E85Q) and 3 (E121Q) that impair their Ca(2+) binding. EF-hand 2 and 3 display different properties and serve different functions. Binding of Ca(2+) to recoverin is a sequential process, wherein EF-hand 3 is occupied first followed by the filling of EF-hand 2. After EF-hand 3 bound Ca(2+), the subsequent filling of EF-hand 2 triggers the exposition of the myristoyl group and in turn binding of recoverin to membranes. In addition, EF-hand 2 controls the mean residence time of recoverin at membranes by decreasing the dissociation rate of recoverin from membranes by 10-fold. We discuss this mechanism as one critical step for inhibition of rhodopsin kinase by recoverin.

Full text

Page 1

Ca2?-Myristoyl Switch in the Neuronal Calcium Sensor Recoverin
Requires Different Functions of Ca2?-binding Sites*
Received for publication, May 3, 2002, and in revised form, September 26, 2002
Published, JBC Papers in Press, October 21, 2002, DOI 10.1074/jbc.M204338200
Ivan I. Senin‡, Torsten Fischer§, Konstantin E. Komolov‡, Dimitry V. Zinchenko¶,
Pavel P. Philippov‡ and Karl-Wilhelm Koch§?
From the ‡A. N. Belozersky Institute of Physico-Chemical Biology, Moscow State University, 119992 Moscow, Russia,
§Institut fu ¨r Biologische Informationsverarbeitung 1, Forschungszentrum Ju ¨lich, D-52425 Ju ¨lich, Germany, and ¶Branch
of Shemyakin and Ovchinnikov Institute of Bioorganic Chemistry, Puschino, 142292 Moscow, Russia
Recoverin is an EF-hand Ca2?-binding protein that is
suggested to control the activity of the G-protein-cou-
pled receptor kinase GRK-1 or rhodopsin kinase in a
Ca2?-dependent manner. It undergoes a Ca2?-myristoyl
switch when Ca2?binds to EF-hand 2 and 3. We investi-
gated the mechanism of this switch by the use of point
mutations in EF-hand 2 (E85Q) and 3 (E121Q) that im-
pair their Ca2?binding. EF-hand 2 and 3 display differ-
ent properties and serve different functions. Binding of
Ca2?to recoverin is a sequential process, wherein EF-
hand 3 is occupied first followed by the filling of EF-
hand 2. After EF-hand 3 bound Ca2?, the subsequent
filling of EF-hand 2 triggers the exposition of the myr-
istoyl group and in turn binding of recoverin to mem-
branes. In addition, EF-hand 2 controls the mean resi-
dence time of recoverin at membranes by decreasing the
dissociation rate of recoverin from membranes by 10-
fold. We discuss this mechanism as one critical step for
inhibition of rhodopsin kinase by recoverin.
G-protein-coupled receptor kinases provide desensitization
of G-protein-coupled receptors by phosphorylation of serine and
threonine residues at their cytoplasmic C terminus (1). A well-
known system represents the light absorbing pigment rhodop-
sin and the corresponding kinase, rhodopsin kinase, or GRK-1,
which phosphorylates (and thus desensitizes) photobleached
rhodopsin. Arrestin then binds to phosphorylated rhodopsin
and thereby stops any further activation of the G-protein trans-
ducin (2, 3). Illumination causes the decrease in the concentra-
tion of the intracellular transmitters of excitation and adap-
tion, cGMP and cytoplasmic [Ca2?], respectively. The decrease
in cytoplasmic [Ca2?] is sensed by Ca2?sensor proteins such as
recoverin (4–6; for a recent review, see Ref. 7). Recoverin or the
amphibian orthologue S-modulin (8, 9) inhibit rhodopsin ki-
nase at high levels of free [Ca2?], thereby relieving inhibition
when the Ca2?level decreases. The Ca2?-dependent control of
rhodopsin kinase via recoverin belongs to at least nine different
mechanisms in vertebrate photoreceptor cells that control ad-
aptation to background light intensities (10–13).
Of particular importance for recoverin as a Ca2?sensor of
rhodopsin kinase is its Ca2?-myristoyl switch (14, 15). A com-
bination of biochemical and structural
has deciphered some of the molecular details of this switch
mechanism; recoverin has a compact structure and consists of
two domains. Each domain harbors two EF-hand structures
but only EF-hand 2 and 3 can bind Ca2?ions. Retinal recoverin
is heterogeneously acylated (mainly myristoylated) at its N
terminus (16). The myristoyl group is sequestered within a
hydrophobic pocket when no Ca2?ions are bound to recoverin.
Extrusion of the myristoyl group is triggered by binding of
Ca2?, which then enables recoverin to associate with phospho-
lipid membranes (17–19).
Myristoyl switches are not restricted to recoverin and its
homologue neuronal calcium sensor proteins (reviewed in 20–
22); instead, myristoyl switches are also found in other signal
transduction proteins. For example, a Ca2?-myristoyl/palmi-
toyl switch mechanism operates in the protozoan Trypanosoma
cruzi and translocates a 24-kDa Ca2?-binding protein to the
flagellar membrane (23). Another type of myristoyl switch
works in ADP-ribosylation factors, which control the assembly
and disassembly of transport vesicle coats. These small GTP-
binding proteins undergo GDP/GTP-exchange, which triggers
the exposure of the myristoyl group (24, 25). A phosphorylation-
dependent myristoyl switch is present in myristoylated ala-
nine-rich C kinase substrates and Src kinases, where phospho-
rylation triggers the release of these proteins from the
membrane (26). Thus, myristoyl switches operate in a wide
range of signaling systems, but they share common themes in
their mechanisms: a trigger such as Ca2?ions, GDP/GTP ex-
change, or phosphorylation induces exposition of the fatty acyl
group and consequently translocation of the protein.
To understand the mechanism of the Ca2?-myristoyl switch
in greater molecular detail, we used recoverin mutants with
impaired EF-hand Ca2?-binding sites (27). For this purpose,
we introduced the substitutions E85Q (mutant RcE85Q) and
E121Q (mutant RcE121Q) in EF-hand 2 and EF-hand 3 of re-
coverin, respectively, because we know that substitutions in
the Z position of the loop of the EF-hand motif decrease the
affinity of the Ca2?-binding site for Ca2?by about 1000-fold
(28, 29). The structure of myristoylated RcE85Qwith one Ca2?
ion bound was recently published and represents an interme-
diate state between the structures of recoverin with zero and
two Ca2?ions bound (30). The myristoyl group in RcE85Qis
partially buried, and it was predicted that the mutant RcE85Q
would not exhibit any considerable binding to membranes (30).
The importance of the positions Glu-85 and Glu-121 was also
described for S-modulin in a study by Matsuda et al. (31). These
authors reported on sequential Ca2?binding and showed that
1H-NMR approaches
* This work was supported by grants from the Deutsche Forschungs-
gemeinschaft (to K.-W. K), a grant from the Forschungszentrum Ju ¨lich
for visiting scientists (to I. I. S., K. E. K., and P. P. P.), the Ludwig
Institute for Cancer Research (to P. P. P.), “International Projects” of
Ministry of Industry and Science, Russian Federation (to P. P. P.), and
Russian Foundation for Basic Research Grants 00-04-48332 (to
P. P. P.), 00-04-48332 (to I. I. S.), and 00-04-48332 (to K. E. K.). The
costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked “adver-
tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate
this fact.
? To whom correspondence should be addressed. Tel.: 49-2461-61-
3255; Fax: 49-2461-614216; E-mail: k.w.koch@fz-juelich.de.
THE JOURNAL OF BIOLOGICAL CHEMISTRY
© 2002 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 277, No. 52, Issue of December 27, pp. 50365–50372, 2002
Printed in U.S.A.
This paper is available on line at http://www.jbc.org
50365
by guest, on April 18, 2013
www.jbc.org
Downloaded from

Page 2

the S-modulin mutant E85M bound only one Ca2?, but it as-
sociated with membranes similar to the wild type. These data
were in contrast to the prediction based on the three-dimen-
sional structure of RcE85Q. We used the recoverin mutants
RcE85Qand RcE121Qto study the Ca2?-myristoyl switch in more
detail and measured binding of WT1-recoverin and the mutants
to rod outer segment (ROS) membranes and liposomes; in par-
ticular, we studied the dynamics of the binding to liposomes by
surface plasmon resonance (SPR) spectroscopy.
EXPERIMENTAL PROCEDURES
Materials—45CaCl2was purchased from Amersham Biosciences. All
other chemicals were obtained from Merck, Sigma, Amersham Bio-
sciences, and Fluka. Bovine brain phosphatidylcholine, bovine brain
phosphatidylserine and egg yolk phosphatidylethanolamine were from
Sigma and 98% pure. Polycarbonate filter membranes (type GTTP)
were from Millipore. All other reagents and buffers were at least ana-
lytical grade.
Preparation of ROS—Bovine ROS were prepared from fresh bovine
retinae and stored at ?80 °C as described previously (32). Urea-washed
ROS membranes were obtained by the homogenization of ROS in 5 M
urea and 20 mM Tris-HCl, pH 7.5, in the dark (33). Then ROS were
incubated on ice for 5 min and were centrifuged at 100,000 ? g for 40
min at 4 °C. The pellet was resuspended and washed three times with
buffer (20 mM Tris-HCl, pH 7.5, 2 mM MgCl2, 0.1 mM EDTA, 1 mM DTT,
and 1 mM phenylmethylsulfonyl fluoride) and stored at ?80 °C.
Preparation of Recoverin and Recoverin Mutants—Mutagenesis of
recoverin was described previously (27). Recombinant nonmyristoy-
lated and myristoylated forms of recoverin and its mutants were ex-
pressed in the E. coli strains pET11d rec/BL21 and pET11d rec/
pBB131/BL21C, respectively. Cells were lysed in buffer B (20 mM Tris-
HCl, pH 8.0, 100 mM NaCl, 1 mM DTT, 1 mM phenylmethylsulfonyl
fluoride, 0.1 mM EDTA, 0.2 mM EGTA) including 0.05 mg/ml lysozyme,
incubated on ice for 20 min and centrifuged at 15000g for 20 min at
4 °C. The supernatant was adjusted to 2 mM or 20 mM CaCl2for
WT-recoverin and RcE121Q, respectively, and was applied to a phenyl-
Sepharose column previously equilibrated with buffer B containing 2
mM or 20 mM CaCl2. Recoverin was eluted with buffer B containing 2
mM EGTA. In the case of mutant RcE85Q, a gel filtration on a Superose
12 (AmershamPhamacia Biotech) was used instead of chromatography
on a phenyl-Sepharose. The fraction containing recoverin was loaded on
a Mono-Q (5/5) column that was equilibrated with 20 mM Tris-HCl, pH
8.0, 1 mM DTT, and 1 mM phenylmethylsulfonyl fluoride. Nonmyristoy-
lated and myristoylated recoverin were eluted with a linear gradient of
0–400 mM NaCl in 20 mM Tris-HCl, pH 8.0. The purity of recoverin was
evaluated by SDS-polyacrylamide gel electrophoresis with Coomassie
Brilliant Blue staining.
The degree of myristoylation was determined by analytical high
performance liquid chromatography and was, in most cases, more than
95%. Only WT recoverin showed a lower degree of myristoylation (70–
80%) and was purified to a homogeneous myristoylated form by high
performance liquid chromatography. Denatured protein was refolded as
described in Ref. 34.
45Ca2?-binding Assay—All buffers and protein solutions used in
Ca2?titration experiments were passed over a Chelex column (Bio-Rad)
to remove residual amounts of Ca2?. Chelex resin was prepared and
equilibrated according to the manufacturer’s instructions. Binding of
45Ca2?was performed as described previously (17). Briefly, samples
(0.5 ml) that contained 50 or 100 ?M protein were dissolved in 20 mM
HEPES, pH 7.5, 100 mM NaCl, and 1 mM DTT and were transferred to
Centricon 10 devices (Amicon). A solution of 10 ?l of 0.2 mM45CaCl2
(0.4–0.5 ?Ci) solution was added and centrifuged for 1 min at 30 °C
(7000 rpm) in a tabletop centrifuge (model TJ-6; Beckman Coulter).
After centrifugation, the radioactivity in 15 ?l of the filtrate (free Ca2?)
and an equal volume of protein sample (total Ca2?) were determined by
liquid scintillation counting. On the next steps, nonradioactive CaCl2
was added and the centrifugation procedure above was repeated. Pro-
tein-bound Ca2?versus free Ca2?was determined from the excess Ca2?
in the protein sample over that present in the ultrafiltrate. The data
were analyzed as follows: Ca2?
radioactivity in the filtrate, Rpis radioactivity in protein sample, and
free? Rf/Rp? Ca2?
total, where Rfis the
Ca2?
respectively.
Preparation of Ca2?Buffer and Determination of Free Ca2?Concen-
tration—Free Ca2?concentrations of buffer solutions were calculated
using the program Webmax 2.0 (Stanford University). We used di-
bromo-BAPTA as Ca2?buffer in Ca2?titration experiments (35). The
free Ca2?concentration of buffer solutions was checked with a Ca2?-
sensitive electrode (World Precision Instruments, Inc.) using calibra-
tion standards in the range from 10?8to 10?1M free [Ca2?].
Equilibrium Centrifugation Assay—Binding of recoverin to ROS
membranes was performed according to a published procedure (14).
Briefly, 30 ?M recoverin or its mutants were mixed with bleached,
urea-washed ROS membranes (50 ?M rhodopsin) and incubated at
37 °C (Eppendorf Thermomixer 5436, 1000 rpm) for 15 min in 20 mM
HEPES, pH 7.5, 150 mM NaCl, 20 mM MgCl2, 1 mM DTT, 3 mM dibromo-
BAPTA, and 0–50 mM CaCl2(total volume 75 ?l). Membranes were
separated by centrifugation (15 min, 14,000 rpm; Eppendorf model 5415
tabletop centrifuge) and the supernatant was removed. The pellet was
analyzed by SDS-PAGE.
Protein Concentration and Electrophoresis—Protein concentrations
were determined by a Bradford assay (36) using a recoverin calibration
curve. SDS-PAGE was performed as described previously (37).
Preparation of Liposomes—Liposomes were prepared as described
previously (38) with slight modification. A mixture of 4 mg of lipids
(40% (w/w) phosphatidylethanolamine, 40% (w/w) phosphatidylcholine,
15% (w/w) phosphatidylserine; 5% (w/w) cholesterol corresponding to
the lipid composition in bovine ROS membranes in CHCl3was dried
down by vacuum in a SpeedVac concentrator. The sample was resus-
pended in 2 ml of degassed buffer (20 mM HEPES, pH 7.5, 150 mM KCl,
and 3 mM EGTA) and sonified for 2 ? 15 min (Branson B12; cup; 100
W). Large unilamellar vesicles (liposomes) were produced using the
extrusion technique. The suspension was soaked for 15–20 min and
extruded through polycarbonate filter with a pore diameter of 1 ?m
(first step) and 0.4 ?m (second step).
Surface Plasmon Resonance Measurements and Data Evaluation—
Binding of recoverin to liposomes and ROS membranes was monitored
by surface plasmon resonance (BIAcore) as described previously (38).
The running buffer contained 10 mM HEPES, pH 7.5, 150 mM KCl, 20
mM MgCl2, and Ca2?buffers as indicated. Liposomes were immobilized
on a sensor chip with a self-assembled monolayer of alkanethiols or,
alternatively, by binding on a hydrophobic sensor chip (pioneer L1;
BIAcore). Nonspecific binding was tested by application of bovine serum
albumin. Interaction of recoverin with immobilized liposomes or ROS
vesicles was tested by applying 2–75 ?M protein in running buffer at
different [Ca2?]. Free [Ca2?] was adjusted with dibromo-BAPTA as
Ca2?buffer. The flow rate was 5 ?l/min. Resonance signals at equil-
brium were determined and normalized to the amount of immobilized
phospholipids that allows the direct comparison of recordings obtained
with different surfaces. Data evaluation was performed with the soft-
ware BIAevaluation 3.1. Sensorgrams were evaluated by nonlinear
curve fitting according to a model of two parallel reactions. Using this
procedure, we determined relative kinetics of a fast and slow component
in the dissociation phase. The Ca2?-dependent development of the
slower component in the dissociation phase was analyzed by comparing
the relative response at 10 s after start of the dissociation process. The
amplitude of each sensorgram at that time was determined and was
assigned RUt. The maximal amplitude of the corresponding sensorgram
was RUmax.
totaland Ca2?
freeare the total and free Ca2?concentration,
RESULTS
Ca2?-dependent Binding of Myristoylated WT-Recoverin and
Recoverin Mutants to Rod Outer Segment Membranes—We
first tested the binding of
WT-recoverin and mutants RcE85Qand RcE121Q, which were
obtained as homogeneous protein preparations. At saturating
Ca2?concentrations (above 100 ?M), WT-recoverin bound two
Ca2?, RcE85Qbound 1.1 Ca2?and RcE121Qdid not bind Ca2?(Fig.
1). Binding of Ca2?to WT-recoverin was half-maximal at 17.6 ?M
(Hill coefficient, nH? 1.9). Ca2?-binding data for RcE85Qreached
a plateau (i.e. apparently saturated) above 100 ?M; however,
when we increased the Ca2?concentration to 2–10 mM, an addi-
tional increase in Ca2?binding was observed (Fig. 1). We inter-
pret the last finding to indicate binding of Ca2?to the mutated
second EF-hand with low affinity. Binding of Ca2?to RcE85Qwas
half-maximal at 34.6 ?M (Hill coefficient, nH? 1.0).
45Ca2?to myristoylated forms of
1The abbreviations used are: WT, wild type; ROS, rod outer segment;
DTT, dithiothreitol; BAPTA, 1,2-bis(2-aminophenoxy)ethane-
N,N,N?,N?-tetraacetic acid; RU, resonance units; SPR, surface plasmon
resonance.
Ca2?-Myristoyl Switch of Recoverin
50366
by guest, on April 18, 2013
www.jbc.org
Downloaded from

Page 3

Our results on Ca2?binding to WT-recoverin were in agree-
ment with previous work (17). The binding experiments with
the mutants showed that impairment of the second EF-hand
leads to binding of one Ca2?, whereas impairment of the third
EF-hand prevents binding of Ca2?to both sites. Initial charac-
terization of the mutants had proven that their general struc-
ture had not changed (39). Thus, mutants RcE85Qand RcE121Q
can serve as selective tools to study the Ca2?-myristoyl switch.
One approach to study the Ca2?-myristoyl switch of recov-
erin is to determine the amount of membrane-associated recov-
erin under varying Ca2?concentrations. Myristoylated forms
of WT-recoverin and the mutants were incubated with washed
ROS membranes at different free [Ca2?] from 117 nM to 50 mM.
Non-bound recoverin was removed by centrifugation, and the
amount of bound recoverin was determined by densitometric
scanning of Coomassie Blue-stained gels.
A representative example is displayed in Fig. 2A for WT-
recoverin and mutants RcE85Qand RcE121Q. The relative inten-
sity of recoverin bands is shown as a function of free [Ca2?] in
Fig. 2B. Half-maximal binding of WT-recoverin and RcE121Qto
ROS membranes was at 4.7 ?M and 5 mM, respectively. The
binding of RcE121Qto ROS membranes with an EC50of 5 mM
Ca2?(Fig. 2B) is far beyond the working range of the Ca2?-
myristoyl switch in WT-recoverin and might reflect the low-
affinity binding of Ca2?to the mutated EF-hand 3 (39), which
allows Ca2?to fill the native EF-hand 2 and to trigger binding
of the mutant to the membranes. A special observation was
made with RcE85Q. Binding of this mutant to ROS membranes
occurred in two steps, reaching a first plateau around 100 ?M
and full saturation above 10 mM (Fig. 2B). The data profile was
reminiscent of45Ca2?binding to nonmyristoylated recoverin
(i.e. binding to two independent sites; see Fig. 2A in Ref. 17).
Therefore, the data were well-fit according to a model of two
independent binding sites yielding two EC50values of half-
maximal binding of RcE85Qto ROS membranes equal to 9 ?M
and 2.6 mM.
Binding of Myristoylated WT-Recoverin to Liposomes Immo-
bilized on a Sensor Chip—In a second approach to studying the
Ca2?-myristoyl switch of recoverin, we applied SPR spectros-
copy to record binding of myristoylated WT-recoverin and mu-
tants RcE85Qand RcE121Qto phospolipid liposomes. SPR spec-
troscopy allows the quantification of the affinity of a binding
reaction and, in addition, can provide insight into the kinetics
of a binding reaction.
We immobilized phospholipid liposomes using a hydrophobic
sensor chip (pioneer L1; BIAcore) to capture phospholipids that
were injected over the flow cell. Immobilized liposomes were
stable over the time course of a typical experiment and could be
used to monitor association and dissociation of recoverin by
SPR-spectroscopy.
First, we tested whether the system was suitable for our
application. When myristoylated WT-recoverin was supplied in
the mobile phase, binding of recoverin to immobilized lipo-
somes on the sensor chip caused a change in resonance units
(RU) and was recorded as a sensorgram (Fig. 3A). The black bar
in Fig. 3A indicates the association phase of the sensorgram.
The association phase attained a plateau during injection of
recoverin, which indicates that the binding of recoverin to
liposomes had reached equilibrium. Dissociation of recoverin
from liposomes was triggered by flushing the flow cell with
running buffer without recoverin (Fig. 3, open bar). Complete
dissociation of recoverin from liposomes was achieved by a
short pulse of 10 mM Ca2?chelator EGTA during the dissoci-
ation phase (not shown). Sensorgrams were recorded at differ-
FIG. 1.
Myristoylated WT-recoverin (G), RcE85Q(E), and RcE121Q(?) were in-
cubated with increasing free [Ca2?]. Solid lines represent the best fit to
the Hill model (myristoylated WT-recoverin, KD? 17.6 ?M, n ? 1.9;
myristoylated RcE85Q, KD? 34.6 ?M and n ? 1.0).
45Ca2?binding to recoverin and recoverin mutants.
FIG. 2. Ca2?-dependent binding of WT-recoverin and recoverin
mutants to ROS membranes. A, Coomassie Blue stained SDS-poly-
acrylamide gels showing the presence of RcE121Q(top trace), RcE85Q
(middle trace), and WT-recoverin (bottom trace) at ROS membranes
after centrifugation. The free [Ca2?] is shown below the corresponding
protein bands for a few selected bands. The complete ranges are 117 nM,
570 nM, 200 ?M, 600 ?M, 1 mM, 2 mM, 3.5 mM, 5 mM, 7.5 mM, 10 mM, and
50 mM for RcE121Q; 117 nM, 3.4 ?M, 5.3 ?M, 9 ?M, 100 ?M, 200 ?M, 400
?M, 600 ?M, 800 ?M, 10 mM, and 50 mM for RcE85Q; and 117 nM, 570 nM,
980 nM, 1.7 ?M, 3.4 ?M, 5.3 ?M, 9 ?M, 28 ?M, 100 ?M, 1 mM, and 50 mM
for WT-recoverin. B, quantitative determination of bound WT-recoverin
(G), RcE85Q(E), and RcE121Q(?) to ROS membranes by densitometric
scanning of Coomassie Brilliant Blue-stained gels as shown in A, yield-
ing an EC50? 4.7 ?M and n ? 1.8 for WT-recoverin, EC1
EC2
points are from two to six different evaluations.
50? 9 ?M and
50? 2.6 mM for RcE85Q, and EC50? 3 mM, n ? 1.6 for RcE121Q. Data
Ca2?-Myristoyl Switch of Recoverin
50367
by guest, on April 18, 2013
www.jbc.org
Downloaded from

Page 4

ent concentrations of recoverin (Fig. 3A). At all concentrations
used, sensorgrams showed rapid association and dissociation of
recoverin. Thus, myristoylated recoverin did not permanently
stay on the membrane; instead, there was a rapid equilibrium
between the bound and soluble states.
Because a bulk refractive index change could lead to an
increase in RU without reflecting any true binding to immobi-
lized liposomes, we performed three control experiments to
exclude this possibility. 1) We applied 30 ?M calmodulin or
protein G to measure the change in bulk refractive index that
might be caused by a protein solution of comparable concentra-
tion (Fig. 3B). Injection of calmodulin and protein G both
caused a change of 70–80 RU or 8.5–9.8 RUnorm(normalized to
the amount of immmobilized lipids). 2) When we used a run-
ning buffer that contained the Ca2?chelator EGTA instead of
Ca2?, we observed signals with amplitude between 67 and 89
RU (data not shown). Amplitudes did not differ much when we
used either WT-recoverin or recoverin mutants in EGTA
buffer. 3) When nonmyristoylated WT-recoverin was injected,
the recorded change was 107–115 RU and 116–125 RU in the
presence and absence of Ca2?, respectively. Thus, nonmyris-
toylated WT-recoverin did not undergo a specific Ca2?-depend-
ent binding to membranes, which is consistent with previous
observations (14, 38). Together, these control recordings
showed that a change in bulk refractive index of a protein
solution is similar to the signal obtained with myristoylated
recoverin forms in the absence of Ca2?. Thus, a Ca2?-depend-
ent change in RU, which is below 70–80 RU, reflects unspecific
binding to immobilized lipids.
Half-maximal binding of myristoylated WT-recoverin to im-
mobilized lipids was determined from a plot of the normalized
amplitude at equilibrium (RUnorm) as a function of the recov-
erin concentration (Fig. 3C). We estimated an EC50of 26 ?M for
myristoylated WT-recoverin, which is similar to the value ob-
tained by Lange and Koch (38) with ROS liposomes and a
different immobilization strategy. Therefore, our experiments
showed that 1) a Ca2?-dependent association of recoverin with
membranes can be recorded on a sensor chip surface with
immobilized liposomes, 2) the association is reversible, and 3)
the binding proceeds in a recoverin concentration-dependent
manner with moderate affinity. These experimental observa-
tions therefore reflect the Ca2?-myristoyl switch of recoverin.
Binding of Myristoylated Recoverin Mutants to Liposomes
Recorded by SPR Spectroscopy—We next tested by SPR spec-
troscopy whether myristoylated recoverin mutants bind to li-
posomes immobilized on a sensor chip surface by the same
protocols as described above for WT-recoverin. First, binding
was tested at 100 ?M Ca2?, 1 mM Ca2?, and no Ca2?present
and second, in the case of binding, the apparent affinity of
mutants for membranes was determined. Myristoylated mu-
tant RcE121Qshowed small signals of the same amplitude at
100 ?M Ca2?, 1 mM Ca2?, and 0 Ca2?(data not shown). In the
presence of 2 mM Ca2?, injection of RcE121Qelicited a response
of ?80 RU. Because this maximal amplitude was identical to
changes in RU obtained in control recordings, we conclude that
RcE121Qdoes not show a Ca2?-myristoyl switch in the range of
free Ca2?from 0 to 2 mM. It was impossible to monitor Ca2?-
dependent binding of RcE121Qto immobilized liposomes by SPR
spectroscopy at higher Ca2?concentrations like it was tested in
the equilibrium centrifugation assay (above). We observed in
several independent experiments that recordings at high
[Ca2?] (?1–2 mM) were less accurate than when we used lower
concentrations of Ca2?. Amplitudes increased because of some
nonspecific matrix effects.
Myristoylated mutant RcE85Qexhibited properties different
from those of RcE121Q: binding amplitudes at 100 ?M Ca2?were
higher than in the absence of Ca2?, and the amplitude further
increased when Ca2?was 1 mM (for example, see Figs. 4 and
5A). The amplitudes (240–250 RU) were also significantly
higher than control recordings with calmodulin or protein G
(Fig. 3B). Thus, we conclude that in mutant RcE85Q, in contrast
to RcE121Q, the Ca2?-myristoyl switch is operative.
FIG. 3. Quantitative analysis of WT-recoverin binding to im-
mobilized liposomes by SPR spectroscopy. A, a representative
overlay of sensorgrams recorded with different concentrations of myr-
istoylated WT-recoverin is displayed. WT-recoverin at 2, 4, 5, 10, 15, 20,
30, 50, and 75 ?M was injected into the running buffer and association
(black bar) to immobilized liposomes was recorded. B, control recording
with 30 ?M protein G. C, maximal resonance signals at equilibrium
were determined, normalized, and are shown as a function of the
WT-recoverin concentration. Data are from two sets of recordings;
EC50? 26 ?M.
Ca2?-Myristoyl Switch of Recoverin
50368
by guest, on April 18, 2013
www.jbc.org
Downloaded from

Page 5

Binding of Myristoylated Forms of WT-Recoverin and Mu-
tant RcE85Qto Liposomes as a Function of Ca2?Concentra-
tion—A critical parameter of the Ca2?-myristoyl switch is the
free [Ca2?] at which the effect is half-maximal. Therefore, we
determined the free [Ca2?] at which binding of myristoylated
forms of WT-recoverin and mutant RcE85Qto immmobilized
liposomes is half-maximal (Fig. 4). The mutant RcE121Qwas not
further considered, because it did not show any Ca2?-depend-
ent binding to membranes within the range of 0.1 ?M to 2 mM.
Binding of WT-recoverin as a function of [Ca2?] showed a
sigmoidal dependence with an EC50of 2.2 ?M and a Hill coef-
ficient of 1.4. In contrast, binding of RcE85Qto membranes
differed in several aspects: the maximal amplitude at 1 and 2
mM Ca2?was less than half of the value observed with the
WT-recoverin, a first plateau of binding was visible around 100
?M Ca2?, and at higher Ca2?concentrations, the data points
did not reach a clear saturation. We did not extend the titration
to higher Ca2?concentrations, because nonspecific matrix ef-
fects of the sensor chip became prominent (see also above). The
arrow in Fig. 4 points to the value obtained from a control
measurement with protein G at 1 mM Ca2?. Amplitudes of
these control measurements include nonspecific binding as well
as nonspecific matrix effects. We can roughly calculate that any
nonspecific matrix effect at 1 mM Ca2?would constitute less
than 10% of the amplitude obtained with WT-recoverin and
RcE85Q.
Data in Fig. 4 seemed similar to data in Fig. 2B for RcE85Q
binding to ROS membranes (note plateau around 100 ?M Ca2?)
and thus could be fit by a model of two independent binding
sites (see above and Ref. 17). In so doing, we obtained two EC50
values: EC1
consistent with the values obtained by binding of RcE85Qto
ROS membranes (Fig. 2B). The binding at low Ca2?(1EC50?
2.2 ?M) made up 25% of the total binding signal, whereas the
binding at high Ca2?contributed to 75% of the total binding
signal. Binding at low Ca2?is similar to binding of WT-recov-
erin to membranes (Figs. 2B and 4). Therefore, we interpret the
small but significant binding of RcE85Qto membranes at low
Ca2?(EC1
mutant, when only one Ca2?ion is bound to EF-hand 3. In-
creasing the [Ca2?] led finally to a weak binding of Ca2?to the
50? 2.2 ?M and EC2
50? 681 ?M. The values were
50) as an operative Ca2?-myristoyl switch in the
mutated EF-hand 2 (Figs. 2B and 4), which led to a complete
exposure of the myristoyl group, but the half-maximal free
[Ca2?] at which the exposure took place was shifted to higher
[Ca2?].
Kinetics of Dissociation of Myristoylated Forms of WT-Recov-
erin and Mutant RcE85Qfrom Membranes—Sensorgrams re-
corded with myristoylated forms of WT-recoverin and RcE85Q
exhibited not only different maximal amplitudes but also dif-
ferent dissociation kinetics (Fig. 5). We compared sensorgrams
that were recorded at 1 ?M and 400 ?M Ca2?in the running
buffer using myristoylated WT-recoverin and RcE85Qas ana-
lytes. Association of WT-recoverin and RcE85Qto phospholipids
rapidly reached the maximal amplitude. Flushing the flow cell
with running buffer, thereby rapidly diluting the recoverin
concentration, triggered dissociation of both proteins from im-
mobilized phospholipids. Sensorgrams obtained with RcE85Q
showed a rectangular shape at all tested Ca2?concentrations
(see, for example, recordings at 1 ?M and 400 ?M Ca2?in Fig.
5A), which indicates rapid association and dissociation. Corre-
sponding dissociation rate constants were in the order of ?0.1
s?1(BIAcore Manual, 1993). In comparison, association and
dissociation of WT-recoverin at 1 ?M free [Ca2?] was also rapid,
and recordings of WT-recoverin and RcE85Qwere almost iden-
tical (Fig. 5A, bottom). However, sensorgrams recorded with
WT-recoverin became different when the free Ca2?concentra-
tion was raised to 400 ?M. In particular, the dissociation phase
became biphasic and contained a slower component (Fig. 5A).
Control recordings with protein G and calmodulin showed a
much smaller response (Fig. 5B), which reflected the rapid
change in bulk refractive index.
We analyzed this phenomenon by performing the experiment
at different free [Ca2?] within the range of 1 ?M to 2 mM using
WT-recoverin as analyte. An overlay of corresponding sensor-
grams is shown in Fig. 5C. To highlight the differences, only
the resonance signals at equilibrium and the dissociation
phases are displayed in Fig. 5C. Increasing free [Ca2?] caused
the development of a slower component in the dissociation
phase. We were seeking to derive a quantitative description of
this effect. First, we tried to fit the data according to a model of
two parallel reactions yielding two dissociation rate constants
(BIAevaluation software 3.1). By this approach, we obtained a
very fast rate constant kd1of ?0.1 s?1and a slower rate
constant of kd2?0.01 s?1, corresponding to the fast and slow
components of the dissociation phase. Because rapid dissocia-
tion rate constants on the order of ?0.1 s?1cannot be fitted
reliably by the BIAevaluation software, we consider these val-
ues only a semiquantitative description of the effect. Neverthe-
less, we conclude that the dissociation of recoverin from the
membranes at high free [Ca2?] is slowed by a factor of at least
10, which consequently prolongs the mean residence time of
recoverin at membranes.
Then we wanted to know at which free [Ca2?] the develop-
ment of the slower component was half-maximal. It is evident
from visual inspection of the sensorgrams that 10 s after start-
ing the dissociation, the slower component showed a clear Ca2?
dependence. Thus, we analyzed the sensorgrams by comparing
the relative response at 5, 10, and 20 s after start of the
dissociation process (see arrow indicating RUtin Fig. 5C). The
amplitude of each sensorgram at that time was determined and
normalized to the maximal amplitude of the corresponding
sensorgram yielding RUt/RUmax(see arrow indicating RUmax
in Fig. 5C). These “relative” amplitudes are shown as a function
of free [Ca2?] in Fig. 5D for a representative set of recordings.
Slowing of the dissociation kinetics is manifested as an in-
crease of relative response units RUt/RUmax. The effect was
half-maximal at 2–4 ?M free [Ca2?] (three sets of recordings),
FIG. 4. Ca2?titration of recoverin-liposome interaction. Bind-
ing of myristoylated WT recoverin (G) and myristoylated RcE85Q(E) to
liposomes was measured at different free [Ca2?] by SPR spectroscopy.
Maximal amplitudes of resonance signals at equilibrium were deter-
mined and normalized (Fig. 3). Normalized RU values are shown as a
function of free [Ca2?]. Values at 1 and 2 mM Ca2?were slightly above
the calculated curve because of some additional nonspecific matrix
effects. The arrow points to the value obtained in the control recordings
with calmodulin and protein G.
Ca2?-Myristoyl Switch of Recoverin
50369
by guest, on April 18, 2013
www.jbc.org
Downloaded from

Page 6

and the Ca2?dependence did not differ much when we ana-
lyzed the effect at different time points (Fig. 5D; differences in
RUt/RUmaxat saturation for different time points mirror the
decreasing phase of the dissociation signal).
Our data on dissociation of myristoylated WT-recoverin and
mutant RcE85Qfrom immobilized phospholipids by SPR spec-
troscopy allow us to conclude that slowing of the dissociation
kinetics by increasing free [Ca2?] occurs in WT-recoverin and is
conferred by the EF-hand 2.
DISCUSSION
A crystallographic study of myristoylated recoverin has
shown that EF-hand 3 has a classic “open conformation” that is
also found in prototypical EF-hand proteins such as calmodulin
and troponin C (40), whereas the structure of EF-hand 2 differs
from these prototypical EF-hand structures. Besides, EF-hand
2 and 3 differ in their affinity for binding Ca2?and might also
play different roles in the function of recoverin.
Structural differences between EF-hand 2 and 3 raise the
question: which site (or both) must bind Ca2?to expose the
myristoyl residue in the myristoyl switch mechanism? NMR
studies of WT-recoverin and fluorescence studies of mutants
RcE85Qand RcE121Qsuggest that the binding of Ca2?to recov-
erin is a sequential process: Ca2?binds first to EF-hand 3 and
then to EF-hand 2 (17, 31, 39). Our45Ca2?titration curves (Fig.
1) are in agreement with these findings. If EF-hand 3 is mu-
tated, Ca2?is unable to bind to the native EF-hand 2 (in
RcE121Q), whereas Ca2?can bind to EF-hand 3 when EF-hand
2 is impaired (in RcE85Q). We hypothesize that myristoylated
RcE121Qis in a locked conformation; because it cannot bind
FIG. 5. Difference in dissociation kinetics of WT-recoverin and RcE85Q. A, overlay of sensorgrams recorded at 400 ?M Ca2?(top) and 1 ?M
Ca2?(bottom) with WT-recoverin and RcE85Qas analyte in the mobile phase (black bar). B, control recordings with 30 ?M calmodulin (dotted line)
and protein G (line). C, overlay of sensorgrams recorded with 15 ?M WT recoverin at different free [Ca2?] (570 nM, 1 ?M, 1.7 ?M. 3.4 ?M, 5.3 ?M,
9 ?M, 28 ?M, 400 ?M, and 2 mM). Only a part of the association phase is shown. Arrows indicate maximal amplitude, RUmax, and the relative
amplitudes 5, 10, and 20 s after onset of dissociation (Rut). D, RUt/RUmaxas a function of free [Ca2?] for three different time points. Values of EC50
were 2.4 ?M ? 0.9 ?M (5 s), 4.1 ?M ? 1.3 ?M (10 s), and 4.4 ?M ? 2.3 ?M (20 s).
Ca2?-Myristoyl Switch of Recoverin
50370
by guest, on April 18, 2013
www.jbc.org
Downloaded from

Page 7

Ca2?in EF-hand 3, the myristoyl group remains clamped in
the hydrophobic cleft and a Ca2?-dependent membrane bind-
ing does not occur (Fig. 2).
EF-hand 2 and 3 are located in the two separate domains
that form recoverin’s compact structure (40). Ames et al. (17,
41) proposed that swiveling about glycine at position 96 alters
the interaction between EF-hand 2 and EF-hand 3 at the do-
main interface and thereby induces a structural change in
EF-hand 2. This change then increases the affinity for Ca2?in
EF-hand 2, which is a necessary step to further trigger the
exposure of the myristoyl group. In a recent study, Ames et al.
showed (30) that the mutant RcE85Qrepresents an intermedi-
ate state of recoverin along the pathway from the Ca2?-free
form (myristoyl group buried) to the fully liganded form (myr-
istoyl group exposed). The myristoyl group is partially un-
clamped in this intermediate state and has exposed its car-
bonyl end. We observed that myristoylated RcE85Qcan bind to
ROS membranes and immobilized liposomes by a two-step
mechanism. A small fraction of RcE85Qcan bind to membranes
at low [Ca2?] (2.2–9 ?M), whereas the major amount needs
millimolar [Ca2?] to fully associate with membranes. The bind-
ing of RcE85Qto membranes at low [Ca2?] probably reflects a
new equilibrium state, where the membrane stabilizes an in-
termediate state with an exposed myristoyl group. The Ca2?
binding affinity of the mutated EF-hand 2 is probably shifted in
the same way. Thus, we hypothesize the following sequence: in
the membrane-free system, the myristoyl group is partially
exposed, when one Ca2?is bound, the presence of membranes
then leads to a new equilibrium with a higher amount of RcE85Q
that has an exposed myristoyl group.
Our results differ significantly from the work of Matsuda et
al. (31), who found by the use of a different EF-hand 2 mutant
(E85M), that the Ca2?sensitivity of the switch of the E85M
mutant is almost identical to that of WT-recoverin. Because our
results with mutant RcE85Qare crucial for the assignment of a
specific function to EF-hand 2 (and EF-hand 3), we will discuss
possible reasons for these discrepancies. Matsuda et al. (31)
used a different mutation and a different protein, S-modulin,
the frog orthologue of bovine recoverin. So far, S-modulin has
exhibited the same biochemical properties as bovine recoverin;
therefore, it is unlikely that species differences account for the
different observations. Differences in the amino acid substitu-
tion at position E85, however, could well lead to different
properties of the mutants.
SPR recordings with WT-recoverin and RcE85Qallowed us
further to decipher another property of EF-hand 2. When EF-
hand 2 is impaired, recoverin can associate with membranes,
but the dissociation from the membranes is very fast both at
low and high [Ca2?] (see Fig. 5). This means that there is a
rapid equilibrium between the soluble and membrane-bound
states of RcE85Q. WT-recoverin showed a slower dissociation
from the membrane when [Ca2?] increased, which would in-
crease the mean residence time of recoverin at the membrane
by at least a factor of 10. This effect was half-maximal at 2 to
4 ?M [Ca2?] (see Fig. 5, C and D). Thus, it had Ca2?sensitivity
similar to that of the myristoyl switch and was similar to the
IC50of rhodopsin kinase inhibition (27, 39–46). EF-hand 2
seems to be critical for the control of rhodopsin kinase activity
(31). Alekseev et al. (27) found no inhibition of rhodopsin kinase
by myristoylated RcE85Qin the range of 1 to 100 ?M Ca2?.
However, a small fraction of RcE85Q(25%) can bind to mem-
branes. How then does the membrane-dependent control of
rhodopsin kinase activity work? Is, for example, the increase of
time in which recoverin remains associated with membranes at
high [Ca2?] important for inhibition of rhodopsin kinase? Work
from Satpaev et al. (47) indicates that this is a reasonable
assumption. These authors measured the apparent konand koff
rates of recoverin and rhodopsin kinase in a membrane-free
assay system using SPR spectroscopy. They report an apparent
konof 105M?1s?1and an apparent kdof 0.1 s?1. The mean
residence time of recoverin on the membrane should extend the
half-life of the recoverin-rhodopsin kinase complex to allow an
effective inhibition. The dissociation rate must therefore be
slower than 0.1 s?1, which is the case for WT-recoverin at high
[Ca2?] but not for the EF-2 mutant (see Fig. 5).
In summary, by combining our functional experimental data
with information on the structure of recoverin (18, 19, 30), we
propose the following model of the Ca2?-myristoyl switch of
recoverin. In the absence of Ca2?(19) and when EF-hand 3 is
mutated (Fig. 2B), the myristoyl group is clamped. According to
the sequential binding of Ca2?to recoverin (Fig. 1 and Ref. 31),
Ca2?binds first to EF-hand 3, which unlocks the clamped
myristoyl group and leads to a transition state (30). In this
transition state, recoverin has about one quarter of membrane
binding activity compared with WT-recoverin. The result-
ing binding of Ca2?to the second EF-hand increases mem-
brane-binding activity because of the complete exposition of the
myristoyl group (18). In addition, we emphasize that EF-hand
2 controls the mean residence time of recoverin at the mem-
brane and that this mechanism is critical for inhibition of
rhodopsin kinase by recoverin.
REFERENCES
1. Pitcher, J. A., Freedman, N. J., and Lefkowitz, R. J. (1998) Annu. Rev. Bio-
chem. 67, 653–692
2. Wilden, U., Hall, S. W., and Ku ¨hn, H. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,
1174–1178
3. Palczewski, K., and Saari, J. C. (1997) Curr. Opin. Neurobiol. 7, 500–504
4. Dizhoor, A. M., Nekrasova, E. R., and Philippov, P. P. (1991) Biokhimia 56,
225–228
5. Dizhoor, A. M., Ray, S., Kumar, S., Niemi, G., Spencer, M., Brolley, D., Walsh,
K. A., Philipov, P. P., Hurley, J. B., and Stryer, L. (1991) Science 251,
915–918
6. Lambrecht, H.-G., and Koch, K.-W. (1991) EMBO J. 10, 793–798
7. Senin, I. I., Koch, K.-W., Akhtar, M., and Philippov, P. P. (2002) Photoreceptors
and Calcium. Landes Bioscience, Georgetown, TX, in press.
8. Kawamura, S. (1993) Nature 362, 855–857
9. Kawamura, S., Hisatomi, O., Kayada, S., Tokunaga, F., and Kuo, C.-H. (1993)
J. Biol. Chem. 268, 14579–14582
10. Pugh, E. N., Jr., Nikonov, S., and Lamb, T. D. (1999) Curr. Opinion Neurobiol.
9, 410–418
11. Pugh, E. N., Jr., and Lamb, T. D. (2000) in Handbook of Biological Physics.
(Stavenga D. G., DeGrip W. J., and Pugh E. N., Jr., eds) pp 183–255,
Elsevier Science B.V., Amsterdam
12. Burns, M. E., and Baylor, D. A. (2001) Annu. Rev. Neurosci. 24, 779–805
13. Fain, G. L., Matthews, H. R., Cornwall, M. C., and Koutalos, Y. (2001) Physiol.
Rev. 81, 117–151
14. Zozulya, S., and Stryer, L. (1992) Proc. Natl. Acad. Sci. U. S. A. 89,
11569–11573
15. Dizhoor, A. M., Chen, C.-K., Olshevskaya, E., Sinelnikova, V. V., Phillipov, P.,
and Hurley, J. B. (1993) Science 259, 829–832
16. Dizhoor, A. M., Ericsson, L. H., Johnson, R. S., Kumar, S., Olshevskaya, E.,
Zozulya, S., Neubert, T. A., Stryer, L., Hurley, J. B., and Walsh, K. A. (1992)
J. Biol. Chem. 267, 16033–16036
17. Ames, J. B., Porumb, T., Tanaka, T., Ikura, M., and Stryer, L. (1995) J. Biol.
Chem. 270, 4526–4533
18. Ames, J. B., Ishima, R., Tanaka, T., Gordon, J. I., Stryer, L., and Ikura, M.
(1997) Nature 389, 198–202
19. Tanaka, T., Ames, J. B., Harvey, T. S., Stryer, L., and Ikura, M. (1995) Nature
376, 444–447
20. Braunewell, K.-H., and Gundelfinger, E. D. (1999) Cell Tissue Res. 295, 1–12
21. Ames, J. B., Tanaka, T., Stryer, L., and Ikura, M. (1996) Curr. Opin. Struct.
Biol. 6, 432–438
22. Burgoyne, R. D., and Weiss, J. L. (2001) Biochem. J. 353, 1–12
23. Godsel, L. M., and Engman, D. M. (1999) EMBO J. 18, 2057–2065
24. Franco, M., Chardin, P., Chabre, M., and Paris, S. (1996) J. Biol. Chem. 271,
1573–1578
25. Goldberg, J. (1998) Cell 95, 237–248
26. McLaughlin, S., and Aderem, A. (1995) Trends Biochem. Sci. 20, 272–276
27. Alekseev, A. M., Shulga-Morskoy, S. V., Zinchenko, D. V., Shulga-Morskaya,
S. A., Suchkov, D. V., Vaganova, S. A., Senin, I. I., Zargarov, A. A., Lipkin,
V. M., Akhtar, M., and Philippov, P. P. (1998) FEBS Lett. 440, 116–118
28. Strynadka, N. C. J., and James, M. N. G. (1989) Annu. Rev. Biochem. 58,
951–998
29. Moncrief, N. D., Kretsinger, R. H., and Goodman, M. (1990) J. Mol. Evol. 30,
522–562
30. Ames, J. B., Hamasaki, N., and Molchanova, T. (2002) Biochemistry 41,
5776–5787
Ca2?-Myristoyl Switch of Recoverin
50371
by guest, on April 18, 2013
www.jbc.org
Downloaded from

Page 8

31. Matsuda, S., Hisatomi, O., and Tokunaga, F. (1999) Biochemistry 38,
1310–1315
32. Koch, K.-W., Lambrecht, H.-G., Haberecht, M., Redburn, D., and Schmidt,
H. H. H. W. (1994) EMBO J. 13, 3312–3320
33. Shichi, H., and Somers, R. L. (1978) J. Biol. Chem. 253, 7040–7046
34. Neubert, T. A., Walsh, K. A., Hurley, J. B., and Johnson, R. S. (1997) Protein
Sci. 6, 843–850
35. Harrison, S. M., and Bers, D. M. (1987) Biochim. Biophys. Acta 925, 133–143
36. Bradford, M. M. (1976) Anal. Biochem. 72, 248–254
37. Laemmli, U. K. (1970) Nature 227, 680–685
38. Lange, C., and Koch, K.-W. (1997) Biochemistry 36, 12019–12026
39. Permyakov, S. E., Cherskaya, A. M., Senin, I. I., Zargarov, A. A., Shulga-
Morskoy, S. V., Alekseev, A. M., Zinchenko, D. V., Lipkin, V. M., Philippov,
P. P., Uversky, V. N., and Permyakov, E. A. (2000) Protein. Eng. 13,
783–790
40. Flaherty, K. M., Zozulya, S., Stryer, L., and McKay, D. B. (1993) Cell 75,
709–716
41. Ames, J. B., Ikura, M., and Stryer, L. (2000) Methods Enzymol. 316, 121–132
42. Gorodovikova, E. N., Gimelbrant, A. A., Senin, I. I., and Philippov, P. P. (1994)
FEBS Lett. 349, 187–190
43. Gorodovikova, E. N., Senin, I. I., and Philippov, P. P. (1994) FEBS Lett. 353,
171–172
44. Calvert, P. D., Klenchin, V. A., and Bownds, M. D. (1995) J. Biol. Chem. 270,
24127–24129
45. Chen, C.-K., Inglese, J., Lefkowitz, R. J., and Hurley, J. B. (1995) J. Biol.
Chem. 270, 18060–18066
46. Klenchin, V. A., Calvert, P. D., and Bownds, M. D. (1995) J. Biol. Chem. 270,
16147–16152
47. Satpaev, D. K., Chen, C.-K., Scotti, A., Simon, M. I., Hurley, J. B., and Slepak,
V. Z. (1998) Biochemistry 37, 10256–10262
Ca2?-Myristoyl Switch of Recoverin
50372
by guest, on April 18, 2013
www.jbc.org
Downloaded from