(Received for publication, January 14, 1997, and in revised form, February 10, 1997)
From the Division of Biology, 147-75, California Institute of Technology, Pasadena, California 91125
In vertebrate photoreceptor cells, transducin
mediates signaling between rhodopsin and cGMP phosphodiesterase by
transiently binding its subunit (PDE
). For the termination of
signaling GTP hydrolysis by the transducin
subunit (TD
) GTPase
is required. This reaction can be accelerated by GTPase-activating
proteins (GAPs), e.g. PDE
. Recently we identified a
second retinal GAP that interacts with TD
, RGS-r. Here we compare
the GAP function of RGS-r and PDE
. Both proteins stimulated single
turnover GTPase of TD
; however, RGS-r was more effective than
PDE
. When added together, PDE
competitively inhibited the
RGS-r-stimulated GTPase. In addition, the interaction of TD
in its
GTP-bound form (TD
GTP
S), the transition state
(TD
GDP*AMF) and the GDP-bound form
(TD
GDP) with RGS-r and PDE, respectively, was measured
by surface plasmon resonance. PDE
displayed highest affinity for
TD
GTP
S, weaker affinity for TD
GDP*AMF,
and weakest affinity for TD
GDP. RGS-r exhibited only a
comparable high affinity for TD
GDP*AMF. We conclude that
the observed competition between RGS-r and PDE
for TD
occurs when
the hydrolysis of GTP is initiated. By competing with PDE
and
removing it from TD
as well as increasing Pi release,
RGS-r apparently facilitates signal termination and TD
recycling.
Transducin (TD),1 the heterotrimeric
guanine nucleotide-binding protein (G protein) in vertebrate
photoreceptor cells mediates the signal between rhodopsin and the
effector enzyme cGMP phosphodiesterase (PDE). Photon absorption
activates rhodopsin which then catalyzes GDP/GTP-exchange with the subunit of TD. The GTP-bound transducin
subunit
(TD
GTP) activates PDE by binding to its inhibitory
subunit (PDE
). Activated PDE lowers cGMP levels, thus closing the
cGMP-gated cation channels and hyperpolarizing the photoreceptor cell
membrane (1-3).
Electrophysiological recordings of isolated photoreceptors revealed
that the entire cellular response to light occurs in less than a second
(4), suggesting that both the activation and inactivation of TD
occur on this time scale. In vitro studies have demonstrated
that TD
activation occurs in 100 ms (5). The subsequent interaction
between TD
and PDE occurs in less than 5 ms (6). The deactivation of
PDE in intact photoreceptor is also a rapid process (<2 s). To achieve
this, GTP hydrolysis by the intrinsic GTPase of TD
has to occur
fast. However, this reaction is a relatively slow process (>10 s)
in vitro, and therefore it has been postulated that a
GTPase-activating protein (GAP) exists to account for the rapid shutoff
of TD
in vivo. Several studies have shown that PDE
itself has GAP activity (7-11). The significance of this GAP activity
of PDE
is still a matter of debate (12-15). In addition, a thus
far, unpurified membrane-bound component with GAP activity for TD
has been described (14, 15). We have identified (16) a retinal specific
member of the RGS protein family (for review, see Refs. 17 and 18)
termed RGS-r. Like other RGS domains (19-22), RGS-r exhibits GAP
activity and, specifically, recognizes a conformation of TD
that
exists during the transition state in the hydrolysis of GTP by
transducin (16). In this report, we compared the GAP activities of
RGS-r and PDE
to get insight into their physiological relevance.
His6-RGS-r was expressed in the Escherichia coli strain BL21(DE3) and purified as described (16).
The cDNA for bovine PDE was subcloned into the NdeI
and BamHI sites of pET15B (Novagen). The amino-terminal
His6-tagged PDE
was expressed in the the E. coli strain BL21(DE3) and purified from bacterial cytosol as
described for RGS-r.
Bleached bovine ROS membranes were prepared from bovine retinae as
described (23). Transducin was eluted from the membranes by hypotonic
elution in the presence of 100 µM GTP or guanosine 5-(
-thio)triphosphate (GTP
S), and the subunits were separated by
affinity chromatography on blue Sepharose (Bio-Rad) (24). Urea-treated
ROS membranes were prepared from the eluted membranes as described
(25).
Transducin single turnover GTPase was
determined by a method described before (8) except that all
measurements were conducted at 0 °C in a buffer containing 50 mM Tris-HCl, pH 7.5, and 2 mM MgCl2. Briefly, the reaction was started by mixing 20 µl
of ROS membranes depleted of PDE (12 µM final rhodopsin
concentration) with 20 µl of [-32P]GTP (0.2 µCi,
0.25 µM final concentration) and incubated for the
indicated periods of time. Where indicated, RGS-r or PDE
was added
before mixing. The reaction was stopped by addition of perchloric acid
and inorganic 32P was measured by Cerenkov radiation in a
liquid scintillation counter. The GTPase rate constant was determined
by the exponential fit of the time course of inorganic 32P
release.
The multiple turnover GTPase of TD was determined in a reconstituted
system as described (16) in a reaction mixture (100 µl) containing 50 mM triethanolamine-HCl, pH 7.4, 100 mM NaCl, 5 mM MgCl2, 0.1 mg/ml bovine serum albumine, 0.2 µM TDGDP, 0.2 µM TD
,
10 µg of urea-treated ROS membranes, and the indicated concentrations
of RGS-r and/or PDE
. The reaction mixture was thermoequilibrated for
5 min at 30 °C, and reaction was initiated by addition of 10 µM [
-32P]GTP (0.2 µCi). After
incubation for 10 min at 30 °C, termination of reaction and
determination of inorganic 32P were performed as
described.
His6-RGS-r or His6-PDE were
tethered to a NTA-sensor chip (Pharmacia Biosensor) after priming the
chip with Ni2+. In detail, the chip was equilibrated at a
continuous flow rate of 5 µl/min with a buffer (BIA buffer)
containing 20 mM Hepes, pH 7.5, 100 mM NaCl, 2 mM MgCl2, 10 mM imidazole, and 0.01 mg/ml bovine serum albumine. First, the chip was washed with 10 µl of BIA buffer containing 50 mM EDTA. The chip was then primed
with 10 µl of 0.1 M NiCl2 and again
equilibrated with BIA buffer. 20 and 10 µl of 0.01 mg/ml RGS-r and
PDE
were applied, respectively, creating a stable surface of about
300-400 RU. For kinetic studies 35 µl of varying concentrations of
TD
GDP, TD
GDP*AMF, or TD
GTP
S were
injected at a flow rate of 5 µl/min. Each injection was followed by a
buffer flow for several minutes to monitor the dissociation of the
complex. The data were analyzed after subtraction of the background
signal (blank injections) with the BIAevaluation software (Pharmacia
Biosensor). The values of the dissociation constants (kd) were determined by exponential fitting of the
SPR signal decay after replacement of the analyte solution by BIA buffer. The values obtained and the applied concentration of the analyte were used to calculate the association constants
(ka) by exponential fitting of the SPR signal
increase after application of the analyte.
To compare
the two putative retinal specific GAPs: the RGS-r domain (16) and
PDE (8-11), we expressed and purified PDE
in a manner similar to
that described for RGS-r (16). The ability of the recombinant PDE
(10 µM) to interact with TD
GDP,
TD
GDP*AMF, or TD
GTP
S was studied by
the column trap assay used previously (16). TD
GDP bound
PDE
only weakly; however, activation of TD
GDP by AMF
significantly increased the binding to PDE
(data not shown). The
binding of TD
GDP*AMF to PDE
was clearly less than the
binding observed with the metabolically stable GTP analog GTP
S bound
TD
(TD
GTP
S). Using similar conditions, RGS-r trapped only TD
GDP*AMF but not TD
GDP or
TD
GTP
S. No binding of either form of TD
to the
matrix was observed in the absence of a trapping His-tagged protein
(16). The data therefore indicate that the recombinant PDE
was fully
functional with regard to the generally accepted scheme of effector-G
protein interaction, in which the interaction of these proteins is
dependent on the G protein activation state.
We next
compared the ability of both these proteins, RGS-r and PDE, to
enhance GTP hydrolysis by TD
in a single turnover assay with
bleached ROS membranes (8-11). In this assay the concentration of TD
(
1 µM, estimated by measurement of the rhodopsin
concentration) is several times higher than the concentration of
[
-32P]GTP (0.25 µM), and as
[
-32P]GTP is quickly bound to excess TD·rhodopsin
complex subsequent, formation of 32Pi reflects
a single synchronized turnover of the TD
-GTPase. As with other RGS
proteins (19-22), RGS-r was very effective in stimulating GTP
hydrolysis by TD
at room temperature (maximal hydrolysis of added
GTP occurred within seconds, data not shown). Therefore, the
temperature at which the reaction was performed was lowered to 0 °C.
Under these conditions we were able to obtain reliable measurements of
single turnover GTPase activity. Maximally 75% of the added GTP was
hydrolyzed (Fig. 1). A maximal stimulating concentration
(10) of PDE
(2 µM) was compared for its ability to
stimulate the GTPase with 200 nM RGS-r (Fig.
1A). The calculated rate constant for the basal GTPase
reaction was 7 ± 0.3 × 10
3 s
1.
In the presence of PDE
and RGS-r this rate increased to 14 ± 0.4 × 10
3 s
1 and 36 ± 1.8 × 10
3 s
1, indicating a 2- and 5-fold
acceleration of GTP hydrolysis, respectively. The RGS-r concentration
used was submaximal (Fig. 1B; EC50 and EC100 for mRGSr were about 0.1 and 1 µM,
respectively). The maximal increase in stimulation of the rate of GTP
hydrolysis was about 10-fold. When added together the GAP proteins did
not increase GTP hydrolysis in a synergistic manner. In contrast,
addition of PDE
inhibited the RGS-r-stimulated GTPase about 50%
(rate constant: 22 ± 0.5 × 10
3
s
1). These data were further confirmed when the influence
of both recombinant proteins on the rhodopsin-catalyzed multiple
turnover GTPase of TD (200 nM) was studied in a
reconstituted system described previously (16). RGS-r stimulated
Pi release maximally about 6-fold (Fig. 2).
Half-maximal and maximal stimulation of GTP hydrolysis occurred at 40 and 200 nM, respectively. In contrast, no stimulation of
Pi release by PDE
was observed in the multiple turnover
assay even when tested in concentrations up to 2 µM.
Furthermore, at concentrations of >1 µM, PDE
inhibited the rhodopsin-catalyzed steady-state GTPase (data not shown).
As suggested before (15), these data can be best explained by a complex
of TD
GDP with PDE
that is refractory to
reactivation.
When the inhibitory effect of PDE on RGS-r-stimulated GTPase was
titrated (Fig. 2A), we found that the extent of inhibition was dependent on the RGS-r concentration used to stimulate TD
GTPase
(Fig. 2A). As more RGS-r was present, more PDE
was
required to obtain the equivalent degree of inhibition. Conversely,
when the effect of increasing concentrations of RGS-r was studied in the absence and presence of 200 nM PDE
(Fig.
2B), the addition of PDE
caused a shift to the right in
the concentration dependence for RGS-r and the half-maximal stimulation
by RGS-r occurred at 40 and 200 nM in the absence and
presence of PDE
, respectively. The maximal stimulation achieved by
RGS-r was virtually not altered by PDE
. This suggest that RGS-r is
more effective than PDE
in stimulating both single and multiple
cycles of transducin activity and that the two proteins compete for
transducin at some stage in the activation cycle. This interpretation
was further supported by the finding that the extent of inhibition of
the RGS-r-stimulated GTP hydrolysis was dependent on the concentration
of TD used in the reconstitution (data not shown).
We studied the direct interaction of
TDGTP
S, TD
GDP*AMF, and
TD
GDP with RGS-r and PDE
, respectively, using SPR
(BIAcore). Either RGS-r or PDE
was tethered to a
Ni2+-NTA sensor chip via their His6
NH2 terminus as described under "Experimental
Procedures." Surfaces of 300-400 RU were created with either protein
and TD
GTP
S, TD
GDP*AMF, or
TD
GDP was used as analytes at various concentrations.
Examples of the measurements of the interaction of RGS-r with
TD
GDP*AMF and TD
GDP or of PDE
with
TD
GDP*AMF are shown in Fig. 3. A summary
of the results is presented in Table I. Binding of
TD
GDP*AMF to RGS-r was relatively fast and apparently
reached saturation at about 500 nM. The maximal increase in
SPR was about 350 RU on a 300-RU surface of RGS-r, which approximates
the value expected for one to one binding of a 39-kDa protein (TD
)
to 30-kDa protein (RGS-r). The dissociation constant
(KD) calculated from these measurements was 45 nM. As observed before by the column trap assay,
TD
GDP exhibited a relatively low affinity
(KD
2 µM), which was due to a
drastic reduction in the association rate. Saturation was not observed
under the conditions used. Binding of TD
GTP
S to RGS-r
was also very weak, and the maximal increase in SPR observed was in the
same range as observed with identical concentrations of
TD
GDP (data not shown), consistent with data most
recently reported for the RGS4-Go
interaction (20),
these data taken together indicate that the affinity of
TD
GTP
S for RGS-r is low. However, exact calculations
of kinetic constants for the TD
GTP
S-RGS-r interaction
were not possible as the binding data recorded were multiphasic and
could not be readily analyzed. PDE
displayed the highest affinity
for TD
in its GTP
S-liganded form. The KD value
obtained from our measurements (33 nM) was virtually
identical to that reported (10), with association (ka) and dissociation rate (kd)
constants within the same order of magnitude. Consistent with the data
obtained by the column trap assay, the affinity of PDE
for
TD
GDP*AMF was lower than for TD
GTP
S.
The difference in affinity was mainly due to an about 4.3-fold increase
in the dissociation rate compared with TD
GTP
S.
Interestingly, even TD
GDP displayed relatively high
affinity for PDE
(300 nM), which is almost an order of
magnitude higher than the affinity of RGS-r for TD
GDP. In summary, the data suggest that RGS-r has the ability to compete with
PDE
for TD
when the protein is in the transition state stabilized
by AMF.
|
When RGS-r and PDE were
compared for their GAP activity, it was evident that RGS-r was more
effective than PDE
. At maximally stimulating concentrations, the
increase in GTPase reaction velocity mediated by RGS-r was about 5-fold
higher than that achieved by PDE
. In fact PDE
inhibited the
RGS-r-stimulated GTPase in the single turnover GTPase as well as in the
multiple turnover GTPase. The inhibition by PDE
was dependent on the
concentration of RGS-r used to stimulate TD
GTPase and was fully
reversible when the concentration of RGS-r was several times higher
than the concentration of PDE
. The extent of inhibition was also
dependent on the concentration of TD. Thus, it is reasonable to suggest
that PDE
and RGS-r compete for TD
. This interpretation is further
supported by data obtained by measurement of the protein-protein
interaction using SPR. Consistent with recently reported data (16, 20,
21), TD
exhibited its highest affinity for RGS-r in the transition
state mimicked by AMF. The affinty of TD
GDP*AMF for
RGS-r was about 4-fold higher than for PDE
, while in the two other
nucleotide stabilized conformations of TD
, TD
GTP
S
and TD
GDP, the affinity of TD
for PDE
was much
higher than for RGS-r. Therefore, the data suggest that RGS-r competes
with PDE
for TD
only at the stage after GTP hydrolysis is
initiated.
Interestingly, in contrast to RGS-r, PDE was not
able to increase the rate of Pi release from TD
in the
multiple turnover GTPase assay, which mainly monitors the recycling of
TD
. At higher concentrations, PDE
(>1 µM)
inhibited the rhodopsin, and TD
catalyzed recycling of TD
itself, an effect that has been described previously by others (26,
27). This effect has been attributed to the formation of a complex of
TD
GDP with PDE
that is refractory to reactivation
(15) and is therefore sequestered from the further interaction with
TD
and rhodopsin (26). As already outlined above, RGS-r was able
to overcome the inhibition by PDE
and thus enhance again the
recycling of TD
. The SPR measurements revealed that RGS-r in
contrast to PDE
displayed very low affinity for TD
GDP, suggesting that after release of Pi,
TD
GDP will easily dissociate from RGS-r and thus is
capable of rebinding to TD
and rhodopsin. Our data are consistent
with the following reaction cycle (Fig. 4). After TD
is loaded with GTP by rhodopsin, it dissociates from the receptor and
TD
. The free TD
GTP then binds specifically to
PDE
and releases the inhibitory effect of this subunit from
PDE
(6, 28-30). When TD
reaches the state where GTP
hydrolysis is initiated, RGS-r effectively competes with PDE
and
thus releases TD
from PDE
, which then again inhibits the enzymatic activity of PDE
. The GAP activity of RGS-r,
i.e. its ability to stabilize a form of TD
that
facilitates GTP hydrolysis and Pi release, rapidly converts
it into its GDP-bound form, which dissociates easily from RGS-r and is
available for rebinding to TD
. It is known that TD
activated
by AMF is capable of stimulating the effector PDE in reconstituted
systems (3), suggesting that PDE is active as long as the
-phosphate
of GTP is not released from TD
. By competing with PDE
in the
stage when GTP hydrolysis is initiated and subsequently catalyzing the
complete hydrolysis, i.e. the release of
-phosphate,
RGS-r can apparently play a major role in PDE deactivation and
recycling of TD. Studies are currently in progress to address the
relationship between RGS-r and the membrane-bound GAP described by
Angleson and Wensel (14, 15).