Current Biology 20, 1327–1335, August 10, 2010 ª2010 Elsevier Ltd All rights reservedDOI 10.1016/j.cub.2010.06.013
Synergistic Activation of Phospholipase
C-b3 by Gaqand Gbg Describes
a Simple Two-State Coincidence Detector
Finly Philip,1Ganesh Kadamur,1,2Rosa Gonza ´lez Silos,1,3
Jimmy Woodson,1and Elliott M. Ross1,2,*
1Department of Pharmacology
2Molecular Biophysics Graduate Program
University of Texas Southwestern Medical Center, 6001 Forest
Park Road, Dallas, TX 75390-9041, USA
3Universidad de Valladolid, Departamento de Estadı ´stica
e Investigacio ´n Operativa, Facultad de Ciencias,
synergistically in cells to elicit cytosolic Ca2+transients that
are several-fold higher than the sum of those driven by each
receptor alone. Such synergism is commonly assumed to
be complex, requiring regulatory interaction between compo-
nents, multiple pathways, or multiple states of the target
Results: We show that cellular Gi-Gqsynergism derives from
direct supra-additive stimulation of phospholipase C-b3
(PLC-b3) by G protein subunits Gbg and Gaq, the relevant
components of the Giand Gqsignaling pathways. No addi-
tional pathway or proteins are required. Synergism is quantita-
tively explained by the classical and simple two-state (inacti-
ve4active) allosteric mechanism. We show generally that
synergistic activation of atwo-state enzymereflectsenhanced
conversion to the active state when both ligands are bound,
not merely the enhancement of ligand affinity predicted by
positive cooperativity.Thetwo-state mechanism alsoexplains
why synergism is unique to PLC-b3 among the four PLC-b iso-
forms and, in general, why one enzyme may respond synergis-
tically to two activators while another does not. Expression of
synergism demands that an enzyme display low basal activity
in the absence of ligand and becomes significant only when
basal activity is % 0.1% of maximal.
Conclusions: Synergism can be explained by a simple and
general mechanism, and such a mechanism sets parameters
for its occurrence. Any two-state enzyme is predicted to
respond synergistically to multiple activating ligands if, but
only if, its basal activity is strongly suppressed.
Cells integrate multiple incoming signals, and a response to
one signal can depend upon the presence or intensity
of others. Most often, acute responses to multiple signals
are simply additive, either positively or negatively. Occasion-
ally, however, the response to simultaneous stimuli is mark-
edly greater than the sum of the responses to each stimulus
alone. Such superadditive responses may be quantitatively
modest, but marked synergism can essentially create a
Boolean AND gate, or coincidence detector, with which a
cell responds significantly only when two signals are present
simultaneously. Superadditive responses are not frequent.
In a recent large-scale screen for signaling interactions in
macrophages, only about 1.5% of the ligand pairs that were
tested displayed significant synergism . In some cases,
mechanisms of cellular synergism are well understood. These
include multiple phosphorylation events, coactivation by tran-
scription factors, induction of synthesis of subsequently regu-
lated proteins, etc. Positively cooperative binding of activating
ligands can also create apparent synergism over a narrow
range of concentrations as each ligand increases the affinity
of the other [2–4]. Scaffolding proteins and membrane
surfaces potentiate signals essentially by this mechanism
[5–7]. For many acute superadditive cellular responses,
however, mechanisms of synergism involve multiple signaling
pathways, are otherwise complex [8, 9], or are unknown.
Here we use phospholipase C-b3 (PLC-b3) to elucidate
general mechanisms for creating synergism through allosteric
regulation, and we show that PLC-b3 regulation accounts for
a well-known set of superadditive responses in diverse cells.
It has been known for about 15 years that many animal cells
and primary cell lines display synergistic Ca2+responses to
simultaneous inputs from different G protein-coupled recep-
tors [10–19]. In these cells, synergism serves as a coincidence
detector, such that a robust Ca2+response and downstream
physiological regulation are only observed when both G
protein pathways are activated. Such synergism is physiolog-
ically important in platelets, neurons, and macrophages [10,
genesis in multiple cell types . In most of these cases, one
of the two receptors activates Gqand the other activates Gi,
and synergism does not depend on which Gi- or Gq-coupled
receptor initiates the signals. Gqand Giboth activate PLC-
b isoforms, and the PLC reaction product, inositol-trisphos-
phate (IP3), triggers Ca2+release from the endoplasmic
reticulum to the cytosol . Gqstimulates PLC-b via its Gaq
subunit, and Giacts via its Gbg subunit . Several studies
suggested that the mechanism of synergistic Ca2+signaling
directly involves PLC activation [10, 12, 16–19, 22–24], and
recent studies in macrophages and a macrophage-like cell
line argue that synergistic stimulation of Ca2+signaling
primarily requires the PLC-b3 isoform . However, other
work suggested that cellular Gi-Gqsynergism involves interac-
tionbetween theGproteins orthe IP3receptor ,and its
biochemical mechanism remained unknown.
We show here that purified PLC-b3 responds synergistically
to stimulation by Gaqand Gbg. Synergistic activation of PLC-
b3 can exceed ten times the sum of the responses to the indi-
vidual G protein subunits. Gbg-Gaqsynergism on PLC-b3 can
thus quantitatively account for synergistic Ca2+responses to
Giand Gqin cells, and its biochemical behavior is qualitatively
consistent with cellular events. Additional proteins or path-
ways are not required.
We also show that the synergistic response of PLC-b3 to
Gaqand Gbg can be explained quantitatively by a simple and
classical two-state allosteric model. Synergism does not
merely reflect positively cooperative effects of each subunit
on the binding affinity of the other, but results from increased
accumulation of the active form of PLC-b3. Synergism occurs
even when both Gaqand Gbg are tested at saturating concen-
The other PLC-b isoforms do not mediate synergistic Ca2+
responses in cells  or display synergism in vitro, even
though they are structurally homologous to PLC-b3 and
respond similarly to individual G proteins .
In general, why does one enzyme respond synergistically to
two activators while another does not? We show by modeling
and by analysis of PLC-b regulation that a superadditive
response by a single enzyme primarily depends on its having
very low activity in the absence of stimulating ligand. Maximal
attainable synergism by a simple two-state enzyme is approx-
imately proportional to its intrinsic bias for the inactive state.
A two-state enzyme whose intrinsic activation is R 1% of
maximal cannot display more than two-fold synergism, and it
can do so only with ligands that are fortuitously matched in
their efficacies and that are at near perfect concentrations. In
contrast, an enzyme with intrinsic activity % 0.1% will display
synergism tomost activatorsand will dosoover abroad range
of activator concentrations. Thus any allosteric enzyme with
a large dynamic range of regulation will display a synergistic
response to two or more activating ligands. Synergism, which
is widely assumed to be a complex phenomenon requiring
ligand-ligand interactions or multiple activity states, can be
described by a simple two-state allosteric equilibrium.
Gaqand Gbg Stimulate PLC-b3 Superadditively
In many cells, simultaneous stimulation of receptors coupled
to Giand Gqproduces a cytosolic Ca2+transient that is much
larger than the sum of the those elicited by the individual
receptors. The Ca2+signal presumably results from Ca2+
release from endoplasmic reticulum, which is triggered by
IP3that is produced by the activity of PLC-b. To see whether
the synergistic Ca2+response in cells reflects direct syner-
gistic activation of PLC-b3 by Gbg and Gaq, we measured
the activity of purified PLC-b3 at increasing concentrations
of GTPgS-activated Gaqand in the presence or absence of
Gb1g2 (Figure 1A). Together, Gaqand Gbg stimulated PLC-
b3 to an activity nearly ten times the sum of the activities
elicited by the two subunits added separately. We define
‘‘synergism’’ generally by this ratio: the activity of an enzyme
or signaling pathway in the presence of two regulatory ligands
(a and b) divided by the sum of the activities elicited by each
ligand (a or b) alone (Equation 1).
If two activities are merely additive, the ratio will be 1.0.
Synergism is described by a ratio substantially above 1, and
ratios above 10 approach an intuitive definition of coincidence
which approaches saturation. The extent of direct Gaq-Gbg
synergism on PLC-b3 can thus readily account for the 2- to
6-fold synergistic responses of cellular IP3-Ca2+pathways
that have been described for simultaneous stimulation by
Gq- and Gi-coupled receptors.
Superadditive stimulation of PLC-b3 by Gbg and Gaqalso re-
sembles cellular Gi-Gqsynergism qualitatively. Gbg mediates
Figure 1. Synergistic Activation of PLC-b3 by Gaqand Gbg
(A)PLC-b3activity was assayed at60nM[Ca2+]withincreasingconcentrations ofGaqintheabsence(open circles)orpresence of6nMGbg(closed circles).
The synergism ratio, the ratio of activities in the presence of both Gaqand Gbg to the sum of the activities in the presence of each subunit alone, is given at
(B) Gaq-Gbg synergism is independent of [Ca2+]. Lower panel: PLC-b3 activity was assayed at various Ca2+concentrations in the presence of 30 nM Gbg
(black triangles), 0.2 nM Gaq(white circles), or both 0.2 nM Gaqand 30 nM Gbg (black circles). Basal activity in the absence of G protein subunits was also
assayed, and is shown multiplied by 10 to distinguish it from baseline (open triangles). Zero Ca2+represents 5 mM EGTA with no added Ca2+. Upper panel:
synergism ratios at each Ca2+concentration. The ratio at zero Ca2+is not accurate because of relative errors in assaying such low activities. The range of
activities in this experiment is greater than 2000-fold. The maximum activity shown for the combination of Gaqand Gbg (both) is about one-third that in the
presence of an optimal concentration of Gbg. Error bars show standard deviation (SD).
Current Biology Vol 20 No 15
to terminate signaling after GTP hydrolysis [21, 27]. Similarly,
Gai-GDP blocked both stimulation of PLC-b3 by Gbg and
additive stimulation when added with Gaq(see Table S1, avail-
able online), consistent with the occurrence of synergistic
responses in diverse cell types. Other experiments used only
Gb1g2. Gbg-Gaqsynergism also requires activation of Gaqby
GTP or a nonhydrolyzable analog (GTPgS); Gaq-GDP neither
stimulates PLC-b3 nor potentiates stimulation by Gbg at the
iments shown here use Gaqthat has been activated by GTPgS.
Because Gaqactivated by GTPgS or GTP binds Gbg with rela-
tively low affinity , Gbg does not block its stimulation of
Gaq-Gbg synergism was independent of Ca2+concentra-
tion from well below that of resting cytosol (30 pM) to higher
than usually reported for stimulated cells (10 mM) (Figure 1B).
Responses to Gaqand Gbg should therefore be potentiative
continuously during a cytosolic Ca2+transient. Ca2+also had
a negligible effect on the EC50or Hill coefficient for either G
protein subunit. Because PLC activity with either or both G
protein subunits extrapolates to zero at low Ca2+, Ca2+
appears not to alter the G protein-driven activation4deacti-
vation equilibrium but simply to act as an amplifier of
PLC activity. We therefore used 60 nM Ca2+for all PLC-b3
experiments as a reasonable value for resting cytosolic
in assay conditions (mole fraction of PIP2, PLC concentration,
temperature, detergent, ionic strength, and lipid surface
composition; data not shown). Gq-Gisynergism can therefore
reasonably be expected in any cell that expresses PLC-b3.
vation of PLC-b3 by Gaqand Gbg can account for superaddi-
General Allosteric Mechanism for Synergistic Enzyme
valently, weaskedwhether asimpletwo-stateallostericmodel
for PLC-b3 activation can account for the markedly superaddi-
tive responses to these ligands. Such a mechanism, described
in Figure 2, demands only that (1) PLC-b3 exists in two confor-
mational states, active (P*) and inactive (P), in equilibrium
described by the constant J; (2) that both Gaqand Gbg bind
reversibly and independently to PLC-b3 in either conforma-
tional state; and (3) that both G protein subunits bind more
tightly to the P* conformation, as described by the bias
constants F and G. This model is classically used to describe
allosteric activation by individual ligands [2, 3]. Note that this
two-state allosteric mechanism is quite general: it neither
requires nor suggests any particular property of the P* state
that makes it more active than P, nor any biochemical mecha-
nism for the P4P* transition. Activation may represent
substantial subunit rearrangement, minor movement of resi-
dues at or near the active site, movement of an autoinhibitory
structure, altered interaction with the membrane surface,
some other event, or a combination of such changes.
We used a combination of fitting to experimental data and
numerical simulation to ask whether the allosteric mechanism
can quantitatively account for both the individual and the
synergistic activation of PLC-b3 by Gaqand Gbg. The activity
of PLC-b3 was measured over a wide range of concentrations
of activated Gaqand Gbg, covering almost a 600-fold range of
activities (Figure 3). These data were fitto an equilibrium equa-
PLC activity as the product of its maximal intrinsic-specific
activity, Z, and the fraction of PLC in the four active species
shown in Figure 2. The numerator sums each active species
and the denominator sums all species. Although this equation
is long, it contains few free parameters: binding constants for
Gaqand Gbg (defined for the less active state); an equilibrium
constant J that describes the inactive-active conformational
equilibrium in the absence of ligand; and two bias constants,
F and G, that describe the preference of Gaqand Gbg for
Table 1. Gai1-GDP Blocks Gbg-GaqSynergism
[Gaq-GTPgS] (nM)[ Gbg ] (nM) [ Gai1-GDP ] (nM) Synergism Ratio
Synergism ratios were determined at 0.2 nM GTPgS-activated Gaqand two
concentrations of Gbg, with or without a 3-fold molar excess of GDP-bound
Gai1. Controls contained Gai1that had been heated at 50?C for 60 min.
Results show means from two experiments, each with triplicate determina-
tions, and are representative of two additional experiments that did not
contain the heated Gaicontrol. Gai1-GDP also blocked stimulation by Gbg
alone (not shown).
Figure 2. A Two-State Allosteric Model for Synergistic Activation of PLC-b3
PLC can exist in one of two states, either relatively inactive (P) or highly
active (P*), with the intrinsic conformational equilibrium described by the
isomerization constant J. Gaqand Gbg can bind to either state at nonover-
lapping sites, with association constants Kqand Kbdefined for the inactive
(P) conformer. Gaqand Gbg, both allosteric activators, bind relatively more
bias constants F and G.
Coincidence Detection by Two-State Enzymes: PLC-b3
correct because maximal activity is more than 500-fold above
Theresponse of PLC-b3toa matrixof concentrations of Gaq
and Gbg was well fit by the allosteric model. Values of
constants displayed tolerable statistical errors (Table 2), and
overlay of the model-based simulation on the experimental
data was clear throughout the ranges of Gaqand Gbg concen-
tration (Figures 3 and 4). Values of maximum activities, EC50,
and Hill coefficient were all approximated well (Figure 3). Qual-
itatively similar fits were obtained for two additional similar
experiments (not shown). Experimental data are thus consis-
tent with the simple two-state model. To corroborate the
values for J, F, and G, we also estimated them from activities
measured in the presence of a single high concentration of
Gaq, Gbg, or both (Table 2). This method is independent of
Kb, Kq, and Z. Values for J and G were similar to those derived
from fitting the complete matrix of activities; the value of F was
somewhat higher but does not change maximal predicted
activation by Gaq because even the lower value predicts
The data of Figure 3 and Table 2 indicate that PLC-b3
resides w99.9% in the inactive state in the absence of G
protein under these assay conditions. (Fractional basal
activity = J / (1 + J).) Saturating Gaqstimulates w250-fold
and saturating Gbg stimulates about 50-fold. Combination of
saturating Gaq and Gbg together produced about 80% of
theoretical total activation (w600-fold) (Table 2). Each subunit
thus markedly potentiated PLC-b3 activation by the other.
Gbg and Gaq also each decreased the EC50 of the other
(Figure S1), indicating that each G protein subunit reciprocally
increases the other’s affinity for PLC-b3. Based on the param-
eters of Table 2, each subunit increases the affinity of the
other about 19-fold, representing DDG w 1.8 Kcal for the
binding interaction. Such positively cooperative binding is
also predicted by the basic allosteric model, which was devel-
oped to describe effects on ligand affinity [2, 3]. Note,
however, that synergism does not merely reflect the recip-
rocal increase in the affinity of each subunit by the other.
Synergism is above 7-fold at saturating concentrations of
Gbg and remains above 2-fold at the highest concentrations
of both subunits.
The extent and concentration dependence of Gaq-Gbg
synergism also agree well with simulation based on the allo-
steric model (Figure 4), and comparison of data and simula-
tion point out general aspects of allosteric synergism. The
synergism ratio displays a pronounced peak at intermediate
concentrations of both Gbg and Gaq, with a peak value of
10. The ratio falls off at high Gaqconcentrations but is signif-
icantly greater than 2.0 even at saturating concentrations of
Gaqand Gbg and remains above 1.0 at very low concentra-
tions where activation is minimal. The Gbg concentration did
not have a marked effect on the maximally synergistic
concentration of Gaq, nor did Gaqalter the maximally syner-
gistic concentration of Gbg. In all of these aspects, the
model-based simulation quantitatively mirrored the experi-
mental data. The two-state allosteric model can thus account
for both independent and synergistic regulation of PLC-b3 at
Figure 3. Coordinate Regulation of PLC-b3 by Gaqand Gbg
PLC-b3 activity was assayed at 60 nM Ca2+over a range of concentrations of Gaqand Gbg chosen to optimize fitting the data to the scheme shown in
averages of duplicates, with ranges, are shown out of a total of 115. Solid lines are simulations based on the scheme in Figure 2 and the parameter values in
J+J ? F ? Kq? ½Gaq?+J ? G ? Kb? ½Gbg?+J ? G ? Kb? ½Gbg? ? F ? Kq? ½Gaq?
J+J ? F ? Kq? ½Gaq?+J ? G ? Kb? ½Gbg?+J ? G ? Kb?½Gbg? ? F ? Kq?½Gaq?+1+Kq? ½Gaq?+Kb? ½Gbg?+Kb?½Gbg? ? Kq? ½Gaq?
Current Biology Vol 20 No 15
Other PLC-b Isoforms Do Not Display Gaq-Gbg Synergism
Seaman and coworkers  reported that only the PLC-b3 iso-
form produces synergistic responses to Gi- and Gq-coupled
receptors in macrophages, even though the four PLC-b
isoforms are structurally homologous and PLC-b1, -b2, and
-b3 are all individually stimulated by both Gaqand Gbg. We
surveyed activation of PLC-b1, PLC-b2, and PLC-b4 over
a wide range of concentrations of both subunits and under
diverse assay conditions but found that stimulation by Gbg
and Gaqwas always additive or less than additive for these
ratio never significantly exceeded 1.0. This negative finding is
If the simple model of Figure 2 quantitatively explains syner-
gistic stimulation of PLC-b3 by Gaq and Gbg, why do the
closely related PLC-b1 and PLC-b2 isoforms not give a syner-
gistic response? More generally, when will an enzyme that is
stimulated by noncovalent binding of two or more activating
ligands display a synergistic response? How is synergism
determined by the parameters of the model?
The simulations in Figure 5 show that the intrinsic isomeriza-
tion constant J determines both the maximal synergism that
can be attained by a two-state allosteric protein and the
sensitivity of synergism to the two bias constants F and G.
Decreasing J increases synergism, and maximum attainable
synergism is approximately inversely proportional to J
(Figure 5E). For an enzyme with more than 1% intrinsic activity
without ligand (J R 0.01), maximal synergism is at most 2.4-
fold (Figure 5B). Sensitivity to the values of F and G is also
very sharp, such that only perfectly matched F and G can yield
even slight synergism. J = 0.01 is thus the practical upper limit
At J = 0.001, about that of PLC-b3, maximal synergism is
increased to 10-fold, and the dependences on F and G are
far less strict (Figure 5C). Further, synergism is at least 3-fold
for almost all reasonable F-G combinations, similar to the
will display significant potentiative responses if J < 0.001. For
lower values of J, maximal synergism increases and depen-
dence on F and G broadens, such that J = 0.0001 can pro-
duce > 25-fold synergism over a wide range of F and G
bias constants (Figure 5) and on the concentrations of the
ligands relative to their intrinsic affinities for the target enzyme
(Figure 4). These two parameters are linked: the dependence
of synergism on ligand concentration varies with the bias
constants F and G at any fixed value of J (Figure 4 and
Figure 4. Gaq-Gbg Synergism Is Maximal at Intermediate G Protein Concentrations
(A) The data from the experiment shown in Figure 2, PLC-b3 activities assayed over a range of concentrations of Gaqand Gbg, are replotted as synergism
ratios, calculated as described in the legend to Figure 1. Each vertex on the surface represents a ratio calculated from the three assays (PLC-b3 plus Gaq,
Gbg, or both), each performed in duplicate. This plot is similar to those derived from two other similar experiments.
(B) Synergism ratios for the experiment in Figure 4A were simulated according to the allosteric model and the parameters shown in Table 2.
Table 2. Allosteric Model Parameters for PLC-b3
Matrix Fit4-Point Fit
5300 6 130 min21
0.00150 6 0.00047
0.220 6 0.042 nM21
434 6 154
0.0307 6 0.0056 nM21
45.9 6 9.0
0.00094 6 0.00002
1700 6 300
41 6 1.1
Values for the parameters of the allosteric model (Figure 2) were estimated
data from the experiment shown in Figure 3, which was performed at 60 nM
Ca2+. The complete experiment contained additional data points that were
included to improve the quality of the fit based on the results of pilot exper-
iments. Z is the maximum specific activity of the PLC under these assay
conditions and varies among assays according to the preparation of phos-
pholipid substrate vesicles. 4-point fit parameters (average of three exper-
iments, 6 SD) were calculated from activities obtained at saturating values
of Gaq, Gbg, both, or neither. The 4-point fit is independent of Z, Kq, and Kb.
Details are in the Supplemental Information.
Coincidence Detection by Two-State Enzymes: PLC-b3
Figure S4). When F and G are both high, the synergism ratio
displays a sharp dependence on ligand concentrations. When
both F and G are decreased, synergism is displayed over
a broad concentration range. Thus, for a given enzyme with
a suitable value of J, synergism is more likely for two ligands
that stimulate with bias constants on the order of 1/J. Further,
when the bias constant for only one ligand is high, its optimum
concentration is tightly defined but a wide range of concentra-
tions of the weaker activator can promote synergism.
Similarly, the synergism depends less on the precise values
of F and G if the concentrations of the two activating ligands
are both low (Figure S5). Lower concentrations allow
Figure 5. Predicted Effect of the Spontaneous Activation Constant J on Synergistic Activation by two allosteric regulators
The activity of an enzyme that is activated by two ligands according to the two-state model (Figure 2) was simulated over a range of values of the bias
constants F and G. For reference, J w 0.001 for PLC-b3 (Table 2). The graphs show calculated synergism ratios at four values of J: 0.1, 0.01, 0.001, and
0.0001. Note the scale differences among the synergism axes in each panel; the maximal synergism ratio for J = 0.1 is less than 1.0. The graph at the bottom
shows the nearly inverse relationship between the maximal synergism ratio and J, with a straight line of best fit drawn for reference. For the simulations, the
concentrations of the two activators were set equal to 1/Kqand 1/Kb. Changing the concentrations alters the location of the maximum synergism ratio in the
F-G plane but has no effect on its value over a wide range of concentrations (see Figures S4 and S5).
Current Biology Vol 20 No 15
synergism over a wide range of F and G, but saturating
concentrations of both ligands will produce superadditive
responses only for a limited range of F and G values. This is
the situation for PLC-b3 (Figure 4). In all cases, however, J is
the primary determinant of whether synergism will be
observed, its maximal extent, and the range of ligand concen-
trations over which it occurs.
Why PLC-b1, -b2, and -b4 Do Not Respond Synergistically
to Gaqand Gbg
The two-state allosteric model also allows us to explain why
only PLC-b3 of the four PLC-b isoforms responds synergisti-
cally to inputs from Giand Gq. PLC-b2 responds well to both
Gbg and Gaq. Its behavior was well fit by Equation 2
for the constants were strikingly different than those for PLC-
b3 (Table S3). Most important, the value of J was 0.15, which
precludes synergism (Figure 5E). The basal activity of PLC-
b2 is 140 6 45 min21under our assay conditions (six duplicate
assays), almost 20 times that of PLC-b3. Thus, PLC-b2 fails to
display synergism because its basal activity is too high,
placing a lid on any possible synergism. In the case of PLC-
b1, basal activity is low enough to permit synergism, with J
% 0.003, but PLC-b1 is not sufficiently sensitive to activation
by Gbg. PLC-b1 is stimulated less than 4-fold by Gbg over
a wide range of Ca2+concentrations, and it is known to be
less sensitive to Gbg than are the -b2 and -b3 isoforms .
For G % 4, simulations do not predict any synergism regard-
less of Gbg and Gaq concentrations, even for J w 0.001
(Figure 5, Figure S5). We saw no response of PLC-b4 to Gbg,
as reported previously . Therefore, G < 2 for PLC-b4, simi-
larly disallowing Gaq-Gbg synergism. The unique ability of
PLC-b3 to respond synergistically to Gaq and Gbg, even
though the other PLC-b isoforms do not, is thus explained by
the two-state model and the values of the isomerization and
bias constants for each enzyme.
Synergistic responses to multiple stimuli are relatively rare in
biology, but they are important because they allow cells to
respond distinctively to two simultaneous signals with novel
behaviors. Depending on the dynamics of the signaling
pathway, these novel behaviors can take several forms. If
each input elicits a minimal response alone and only simulta-
neousstimulation generatesanintracellularsignal,then syner-
gism creates a coincidence detector, or logical ‘‘AND’’ gate.
may be strong enough to initiate signaling on its own, and
synergism conveys information on context; each signal is
can be quantitative, more of the same cellular signal, but such
amplification can initiate qualitatively new outputs depending
on the response thresholds of downstream proteins.
This study shows that synergistic signaling by Gi- and Gq-
coupled receptors can be explained by the superadditive
response of PLC-b3 to stimulation by Gbg and Gaq. Gi-Gq
iologically important coincidence detector in diverse cells [10,
high Gbg concentration is required (Figures 3 and 4) and only
and release their Gbg adequately [21, 30]. Gis are the primary
source of Gbg for all signaling events, apparently for this
The 10-fold superadditive response of PLC-b3 to Gbg and
Gaq is quantitatively more than adequate to account for
cellular Gi-Gq synergism over the range of cytosolic Ca2+
concentrations. Only PLC-b3 among the PLC-b isoforms
displays this behavior, which agrees with the finding that
only PLC-b3 permits Gi-Gqsynergism in cells . PLC-b3 is
thus a sensitive cellular coincidence detector, one of few allo-
steric proteins that can act in this way. Gi-Gq synergism
requires no other cellular proteins or pathways. By expression
of this isoform, cells can switch between an additive response
to Giand Gqinputs and a coincidence detection mode.
Synergism demands that both Gaqand Gbg bind simulta-
neously to nonoverlapping sites on PLC-b, as suggested
previously . Because the relative spatial orientation of the
two binding sites is unknown [32, 33], it is unclear whether
Gaqand Gbg are in contact with each other when bound to
PLC-b3. When Gaq and Gbg bind to the RGS domain of
GRK2, the two subunits make no contact and lie on essentially
affinity of Gbg for GTPgS-activated Gaqis low enough that it
should not significantly sequester activated Gaq at the
concentrations used here . Does simultaneous binding of
Gaqand Gbg to PLC-b3 alter the conformation of either G
protein subunit? The ability of Gbg to inhibit the Gq GAP
activity of PLC-b [28, 35] might involve such contact, but
synergism between Gbg and GTPgS-activated Gaq shows
that synergism as such does not involve GAP inhibition.
General Mechanism for Synergistic Response by a Single
The synergistic response of PLC-b3 can be described quanti-
tatively by a simple two-state allosteric model that requires
only that PLC-b3 exist in two interconvertible states with low
and high intrinsic activities (Figure 2). It neither requires nor
predicts any particular physical property of the two states or
of the transition between them. Activation may reflect gross
domain rearrangement, movement of an autoinhibitory struc-
ture, minor motion of an active site residue, or, as suggested
for the PLCs , reorientation with respect to the membrane
bilayer. More broadly, a general two-state model can account
for synergism regardless of whether regulation is allosteric or
covalent. Noncovalent allosteric regulation of a protein that
is also stimulated by phosphorylation, for example, can be
described by the same conformational equilibria shown in
Figure 2. Similarly, the model is applicable to signaling
proteins that are not enzymes: transcription factors, channels,
of a protein’s dynamic structure, this model shows that syner-
gism can be attained without supposing distinct conforma-
tions favored by each ligand or their combination.
any direct interaction between the two ligands or any direct
effect of one ligand upon the binding of the other. In terms of
Figure 2, Gbg does not change F and Gaqdoes not change
G. Synergism occurs simply because the binding of both
ligands favors the active state. There is no ‘‘higher-order
coupling.’’ The two-state model was developed to deal with
cooperative ligand binding [2, 3] and obviously predicts posi-
tive cooperativity of binding of the two ligands (Figure S1).
Enhanced binding can result in physiologic synergism as one
ligand allows another to act at a lower concentration than it
Coincidence Detection by Two-State Enzymes: PLC-b3
would otherwise (e.g., [4, 36, 37]). However, the synergism
described here results from an increased population of the
active state of the enzyme rather than just increased affinity
for activating ligands.
Synergistic activation in a two-state system demands that
the enzyme strongly favor the inactive state in the absence
of ligand. J must be low, and this makes intuitive sense.
Binding of each ligand drives the enzyme to its more active
form with the free energy associated with its bias constant, F
or G. This is true regardless of J. However, a low value of J
provides a large enough dynamic range of activation that the
addition of these free energies can be expressed as a syner-
gistic response in net activity. Synergism therefore does not
bias for the active state. Each ligand contributes its own DDG
to the conformational equilibrium, but synergistic activation
does not require a ‘‘DDDG’’ for ligand-ligand interaction.
Such complex interactions surely occur for some enzymes,
glycogen phosphorylase for example [38, 39], but they
demand the explicit assumption of more and different stable
conformational states, which in general is unnecessary.
Why is synergism observed so rarely if the simplest and
most common model for allostery predicts it? Again, the
answer lies with the demand for a low value of J. Maximum
synergism and J are approximately inversely proportional
(Figure 5E). If an enzyme is even 1% active without ligands,
its capacity for a synergistic response will be slight, and it
will display no synergism at all unless the bias constants for
the activators and their concentrations are fortuitously well
matched. Most allosteric enzymes are stimulated less than
100-fold bytheir regulatory ligands,and far smaller stimulation
can be important for cellular regulation. Yet, these proteins will
not show detectable synergistic responses.
In contrast, decreasing intrinsic activation to 0.1% allows an
enzyme to respond with robust synergism, as is the case for
PLC-b3. Maximum synergism will exceed 8-fold and will be
observed for ligands that display a relatively broad range of
bias constants. The concentration optima for synergism will
depend on the bias constants, but high synergism will be
observed over a > 10-fold range of activator concentrations
and will be more than 2-fold for all relevant activator concen-
trations. This is the case for PLC-b3 (Figure 4). Values of
J < 0.001 further broaden both the extent of synergism and
the tolerance for divergent bias constants (Figure 5).
For the PLC-bs, this analysis explains why PLC-b3 responds
synergistically to Gaqand Gbg but PLC-b1 and PLC-b2 do not.
Although PLC-b2 responds well to both G protein subunits, its
intrinsic activity is too high, J = 0.15, and no combination of
concentrations or bias constants will allow synergism. For
PLC-b1, synergism is limited because its intrinsic response
to Gbg is too low, even though it responds to Gbg significantly
both in cells and after purification.
Using the basal activation set point to determine whether an
enzyme functions as a coincidence detector or merely as
a dual responder offers distinct evolutionary advantages.
Synergism can be acquired or lost by changing J only 10-
fold, while retaining the same fractional (‘‘-fold’’) responses
to each regulatory input. An enzyme with J = 0.01 can respond
to two ligands with almost a 100-fold dynamic range but
display essentially no synergism. Alternatively, for J = 0.001,
the protein will act as a sensitive coincidence detector in addi-
tion to providing a response to each ligand. An enzyme can
evolve between these two regimes without sacrificing under-
lying allosteric regulation. Even absolute signaling activity
can be retained with only minor changes in either catalytic
activity (kcat/Kmfor the active state) or level of expression. In
terms of cellular signaling, changing J in the range below
0.01 will have negligible practical effect on basal activity.
The general inverse dependence of synergism on an
enzyme’s basal level of activity suggests that any enzyme
that can be activated more than 500-fold (J < 0.002) is likely
to display synergism among its activators. Examples include
adenylyl cyclases , some protein kinase C isoforms ,
and the Rac exchange factor P-Rex1 . Novel synergisms
should be detectable by identifying other highly regulated
enzymes. Evaluating the behavior of these enzymes in cells
should drive discovery of new synergisms, coincidence detec-
tors, and biological AND gates.
Last, even though our data do not speak to the regulation of
synergism by additional inputs, the allosteric model argues
that synergism can be modulated best by controlling the value
of J, perhaps with an added benefit of reducing basal activity.
Modulation of J by other signaling mechanisms can thus
convert an enzyme that responds independently to stimuli
into a coincidence detector.
Detailed experimental procedures are in the Supplemental Information.
All proteins were purified essentially as described . Gaqand Gbg were
finally concentrated byadsorption toQ-Sepharose andelutionin 5mg/ml 3-
minimize detergent in the PLC assay. Gaqwas activated with GTPgS ,
but incubation was extended to 5 hr such that Gaqthat did not bind GTPgS
would be denatured and would not bind Gbg. Gb1g2 was used throughout
except in Table SI, where other Gbg isoforms were tested.
PLC activity was measured at 37?C by monitoring hydrolysis of [3H]PIP2
on the surface of large unilamellar vesicles composed of PE:PS:PIP2
(20:4:1 molar ratio), roughly similar to the inner monolayer of the plasma
membrane . Activities are reported as moles of IP3produced per min
per mole of PLC. The concentration of free Ca2+was adjusted with an
EGTA buffer and the program Bound and Determined  and was 60 nM
unless indicated otherwise. Because PLC-b3 can be activated more than
104-fold by combination of Ca2+, Gbg, and Gaq(see Figure 1), assay time
(2–40 min) and PLC-b3 concentration (10–4000 pM) were adjusted for
each assay to maintain linearity of activity with enzyme concentration,
obtain accurately measurable PIP2hydrolysis, prevent substrate depletion,
and control free concentrations of G protein subunits. CHAPS inhibits stim-
ulation of PLC-b3 with IC50= 100 mM. CHAPS was less than 20 mM in all
assays and was equalized among all samples in each assay.
Supplemental Information includes five figures, three tables, and Supple-
mental Experimental Procedures and can be found with this article online
We thank multiple colleagues at University of Texas Southwestern for
valuable discussion; William Seaman, Robert Rebres, and Tamara Roach
(University of California San Francisco) for sharing unpublished results;
Paul Sternweis (UT Southwestern) for the PLC-b2 baculovirus and PLC-
b3 cDNA; and Iain Frasier (National Institute of Allergy and Infectious
Diseases) for the PLC-b4 cDNA. This work was supported by National
Institutes of Health (NIH) grant R01GM030355 and an American Recovery
and Reinvestment Act competing supplement, and by Welch Foundation
Received: April 27, 2010
Revised: May 31, 2010
Accepted: June 1, 2010
Published online: June 24, 2010
Current Biology Vol 20 No 15
1. Natarajan, M., Lin, K.-M., Hsueh, R.C., Sternweis, P.C., and Rangana-
than, R. (2006). A global analysis of cross-talk in a mammalian cellular
signalling network. Nat. Cell Biol. 8, 571–580.
2. Weber, G. (1992). Protein Interactions (New York, NY: Chapman & Hall).
3. Wyman, J., and Gill, S.J. (1990). Binding and Linkage. Functional Chem-
istry of Biological Macromolecules (Mill Valley, CA: University Science
4. Prehoda, K.E., and Lim, W.A. (2002). How signaling proteins integrate
multiple inputs: A comparison of N-WASP and Cdk2. Curr. Opin. Cell
Biol. 14, 149–154.
5. Pawson, T. (2007). Dynamic control of signaling by modular adaptor
proteins. Curr. Opin. Cell Biol. 19, 112–116.
6. Winters, M.J., Lamson, R.E., Nakanishi, H., Neiman, A.M., and Pryciak,
P.M. (2005). A membrane binding domain in the Ste5 scaffold syner-
gizes with Gbg binding to control localization and signaling in phero-
mone response. Mol. Cell 20, 21–32.
7. Pu, M., Roberts, M.F., and Gershenson, A. (2009). Fluorescence corre-
lation spectroscopy of phosphatidylinositol-specific phospholipase C
monitorstheinterplayofsubstrate andactivatorlipidbinding. Biochem-
istry 48, 6835–6845.
8. Barrera, N.P., Morales, B., Torres, S., and Villalo ´n, M. (2005). Principles:
Mechanisms and modeling of synergism in cellular responses. Trends
Pharmacol. Sci. 26, 526–532.
9. Hlavacek, W.S., Faeder, J.R., Blinov, M.L., Perelson, A.S., and Gold-
stein, B. (2003). The complexity of complexes in signal transduction. Bi-
otechnol. Bioeng. 84, 783–794.
10. Roach, T.I.A., Rebres, R.A., Fraser, I.D.C., Decamp, D.L., Lin, K.M.,
Sternweis, P.C., Simon, M.I., and Seaman, W.E. (2008). Signaling and
cross-talk by C5a and UDP in macrophages selectively use PLCbeta3
to regulate intracellular free calcium. J. Biol. Chem. 283, 17351–17361.
11. Toms, N.J., and Roberts, P.J. (1999). Group 1 mGlu receptors elevate
[Ca2+]iin rat cultured cortical type 2 astrocytes: [Ca2+]isynergy with
adenosine A1receptors. Neuropharmacology 38, 1511–1517.
12. Okajima, F., Sato, K.,Sho, K., and Kondo, Y. (1989). Stimulation of aden-
osine receptor enhances a1-adrenergic receptor-mediated activation
of phospholipase C and Ca2+mobilization in a pertussis toxin-sensitive
manner in FRTL-5 thyroid cells. FEBS Lett. 248, 145–149.
13. Werry, T.D., Wilkinson, G.F., and Willars, G.B. (2003). Mechanisms of
cross-talk between G-protein-coupled receptors resulting in enhanced
release of intracellular Ca2+. Biochem. J. 374, 281–296.
14. Abrams, C.S.(2005). Intracellular signalinginplatelets.Curr. Opin. Hem-
atol. 12, 401–405.
15. Dickenson, J.M., and Hill, S.J. (1993). Coupling of histamine H1 and
adenosine A1 receptors to phospholipase C in DDT1MF-2 cells: Syner-
gistic interactions and regulation by cyclic AMP. Biochem. Soc. Trans.
16. Shah, B.H., Siddiqui, A., Qureshi, K.A., Khan, M., Rafi, S., Ujan, V.A.,
Yakoob, M.Y., Yaqub, Y., Rasheed, H., and Saeed, S.A. (1999). Co-acti-
vation of Gi and Gq proteins exerts synergistic effect on human platelet
aggregation through activation of phospholipase C and Ca2+signalling
pathways. Exp. Mol. Med. 31, 42–46.
17. Cilluffo, M.C., Esqueda, E., and Farahbakhsh, N.A. (2000). Multiple
receptor activation elicits synergistic IP formation in nonpigmented
18. Okajima, F., Tomura, H., and Kondo, Y. (1993). Enkephalin activates the
phospholipase C/Ca2+system through cross-talk between opioid
receptors and P2-purinergic or bradykinin receptors in NG 108-15 cells.
A permissive role for pertussis toxin-sensitive G-proteins. Biochem. J.
19. Werry, T.D., Wilkinson, G.F., and Willars, G.B. (2003). Cross talk
between P2Y2nucleotide receptors and CXC chemokine receptor 2 re-
sulting in enhanced Ca2+signaling involves enhancement of phospholi-
pase C activity and is enabled by incremental Ca2+release in human
embryonic kidney cells. J. Pharmacol. Exp. Ther. 307, 661–669.
20. Rozengurt, E. (2007). Mitogenic signaling pathways induced by G
protein-coupled receptors. J. Cell. Physiol. 213, 589–602.
21. Rebecchi, M.J., and Pentyala, S.N. (2000). Structure, function, and
control of phosphoinositide-specific phospholipase C. Physiol. Rev.
22. Carroll, R.C., Morielli, A.D., and Peralta, E.G. (1995). Coincidence detec-
tion at the level of phospholipase C activation mediated by the m4
muscarinic acetylcholine receptor. Curr. Biol. 5, 536–544.
23. Chan, J.S.C., Lee, J.W.M., Ho, M.K.C., and Wong, Y.H. (2000). Preacti-
vation permits subsequent stimulation of phospholipase C by Gi-
coupled receptors. Mol. Pharmacol. 57, 700–708.
24. Zhu, X., and Birnbaumer, L. (1996). G protein subunits and the stimula-
tion of phospholipase C by Gs-and Gi-coupled receptors: Lack of
receptor selectivity of Ga16and evidence for a synergic interaction
between Gbg and the a subunit of a receptor activated G protein.
Proc. Natl. Acad. Sci. USA 93, 2827–2831.
25. Quitterer, U., and Lohse, M.J. (1999). Crosstalk between Gai- and Gaq-
coupled receptors is mediated by Gbg exchange. Proc. Natl. Acad. Sci.
USA 96, 10626–10631.
26. Ogasawara, H. (2008). The calcium kinetics and inositol trisphosphate
receptor properties shape the asymmetric timing window of coinci-
dence detection. J. Neurosci. 28, 4293–4294.
chik, P. (1992). Isozyme-selective stimulation of phospholipase C-b2 by
G protein bg-subunits. Nature 360, 684–686.
28. Tang, W., Tu, Y., Nayak, S.K., Woodson, J., Jehl, M., and Ross, E.M.
(2006). Gbg inhibits Ga GTPase-activating proteins by inhibition of Ga-
GTP binding during stimulation by receptor. J. Biol. Chem. 281, 4746–
29. Lee, C.-W., Lee, K.-H., Lee, S.B., Park, D., and Rhee, S.G. (1994). Regu-
J. Biol. Chem. 269, 25335–25338.
30. Clapham, D.E., and Neer, E.J. (1997). G protein bg subunits. Annu. Rev.
Pharmacol. Toxicol. 37, 167–203.
31. Smrcka, A.V., and Sternweis, P.C. (1993). Regulation of purified
subtypes of phosphatidylinositol-specific phospholipase Cb by G
protein a and bg subunits. J. Biol. Chem. 268, 9667–9674.
32. Hicks, S.N., Jezyk, M.R., Gershburg, S., Seifert, J.P., Harden, T.K., and
Sondek, J. (2008).General and versatile autoinhibition of PLC isozymes.
Mol. Cell 31, 383–394.
33. Singer, A.U., Waldo, G.L., Harden, T.K., and Sondek, J. (2002). A unique
fold of phospholipase C-b mediates dimerization and interaction with
Gaq. Nat. Struct. Biol. 9, 32–36.
34. Tesmer, V.M., Kawano, T., Shankaranarayanan, A., Kozasa, T., and
Tesmer, J.J.G. (2005). Snapshot of activated G proteins at the
membrane: The Gaq-GRK2-Gbg complex. Science 310, 1686–1690.
35. Chidiac, P., and Ross, E.M. (1999). PLC-b1 directly accelerates GTP
hydrolysis by Gaqand acceleration is inhibited by Gbg subunits. J.
Biol. Chem. 274, 19639–19643.
36. Buck, M., Xu, W., and Rosen, M.K. (2004). A two-state allosteric model
for autoinhibition rationalizes WASP signal integration and targeting. J.
Mol. Biol. 338, 271–285.
37. Prehoda, K.E., Scott, J.A., Mullins, R.D., and Lim, W.A. (2000). Integra-
tion of multiple signals through cooperative regulation of the N-
WASP-Arp2/3 complex. Science 290, 801–806.
38. Sprang, S.R.,Acharya, K.R., Goldsmith, E.J., Stuart,D.I., Varvill, K.,Flet-
terick, R.J., Madsen, N.B., and Johnson, L.N. (1988). Structural changes
in glycogen phosphorylase induced by phosphorylation. Nature 336,
39. Barford, D., and Johnson, L.N. (1989). The allosteric transition of
glycogen phosphorylase. Nature 340, 609–616.
40. Sunahara, R.K.,Dessauer,C.W.,Whisnant, R.E.,Kleuss,C.,andGilman,
A.G. (1997). Interaction of Gsawith the cytosolic domains of mammalian
adenylyl cyclase. J. Biol. Chem. 272, 22265–22271.
41. Nishizuka, Y. (1992). Intracellular signaling by hydrolysis of phospho-
lipids and activation of protein kinase C. Science 258, 607–614.
42. Welch, H.C.E., Coadwell, W.J., Ellson, C.D., Ferguson, G.J., Andrews,
S.R., Erdjument-Bromage, H., Tempst, P., Hawkins, P.T., and Stephens,
L.R. (2002). P-Rex1, a PtdIns(3,4,5)P3- and Gbg-regulated guanine-
nucleotide exchange factor for Rac. Cell 108, 809–821.
43. Biddlecome, G.H., Berstein, G., and Ross, E.M. (1996). Regulation of
phospholipase C-b1 by Gqand m1 muscarinic cholinergic receptor.
Steady-state balance of receptor-mediated activation and GTPase-
44. Chidiac, P., Markin, V.S., and Ross, E.M. (1999). Kinetic control of
guanine nucleotide binding to soluble Gaq. Biochem. Pharmacol. 58,
45. Brooks, S.P.J., and Storey, K.B. (1992). Bound and determined: A
computer program for making buffers of defined ion concentrations.
Anal. Biochem. 201, 119–126.
Coincidence Detection by Two-State Enzymes: PLC-b3
Current Biology, Volume 20
Synergistic Activation of Phospholipase
C- 3 by G q and G Describes
a Simple Two-State Coincidence Detector
Finly Philip, Ganesh Kadamur, Rosa González Silos, Jimmy
Woodson, and Elliott M. Ross
5 supplemental figures with legends
3 supplemental tables
Supplemental Experimental Procedures
Table S1. Multiple G isoforms synergize with Gq
Gq G (min-1 ± SD)
+ --- 547 ± 4
- +G2 24 ± 3
+ +G 2250 ± 52
- +G 74 ± 1
+ +G 2840 ± 222
- +G 110 ± 9
+ +G 2970 ± 189
- +G 66 ± 12
+ +G 2370 ± 115
PLC-3 was assayed in the presence of four different G isoforms, 5 nM each, with and without
0.5 nM Gq. These G preparations were purified without the final anion exchange step and
therefore produced lower activities because of excess detergent in the assays. Similar results,
i.e. minimal difference in synergism ratios among the G preparations, were obtained in other
experiments using other concentrations of Gq and G.
Table S2. Gq must be activated to synergize with G.
Additions (min-1) ratio
None 36 ± 3
20 nM Gq-GDP 45 ± 3
20 nM G 178 ± 29
Both 242 ± 64
0.2 nM Gq-GTPS 235 ± 8
20 nM G 178 ± 29
Both 1630 ± 60 5.8
20 nM Gq-GDP 45 ± 3
30 nM G 238 ± 28
Both 263 ± 16
PLC-3 activity was measured in the presence of the G protein subunits shown. Gq was bound
either to GDP or GTPS prior to assay. Values are means, ± SD, of triplicate determinations
except for Gq-GDP, where n=6. Activity for 20 nM Gq-GDP alone is listed twice for clarity
in the table. Data are representative of results of 3 separate experiments.
Table S3. Model parameters for PLC-2 at 60 nM Ca2+
Z 2500 ± 130 min-1
J 0.15 ± 0.015
Kq 0.240 ± 0.062 nM-1
F 8.2 ± 1.4
Kb 0.00061 ± 0.00037 nM-1
G 410 ± 120
Values for the parameters, ± SE, of the allosteric model (Figure3) were estimated by fitting data
from the experiment shown in Figure S2, but with additional data not shown included to improve
the quality of the fit. Z is the maximum activity of the PLC-2 under these assay conditions,
with units of mol PIP2 hydrolyzed /min / (mol PLC).
Legends for supplemental figures
Figure S1. Gq and G each decrease the EC50 value of the other for activation of PLC-3.
Data are from the experiment shown in Figure 3. The EC50 values are derived from mid-points
(± 95% confidence limits) of individual concentration-activity curves for one G protein subunit
at selected concentrations of the other. Fits were based on a single-site saturation equation. The
solid lines are calculated using the parameters of Table 2, and show that each G protein subunit
decreases the EC50 of the other approximately 19-fold. This relative change is equal to the
calculated change in the equilibrium binding constant for one subunit that is caused by a
saturating amount of the other.
Figure S2. PLC-1, PLC-2 and PLC-4 do not display Gq-G synergism. Activities of
PLC-1, PLC-2 and PLC-4 were assayed in the presence of Gq, G or both, each at the
concentrations shown below the graphs. Three sets of data, ± SD, are shown for each PLC-
isoform, and the fourth bar in each set shows the sum of activities measured with Gq or G
alone. PLC-1 and PLC-4 were assayed at 1 M Ca2+ and PLC-2 was assayed at 60 nM Ca2+.
These results are representative of multiple experiments at diverse concentrations of PLC-,
Ca2+ and each G protein subunit. Unstimulated activities in these experiments were PLC-1, 72
min-1; PLC-2, 69 min-1; PLC-3, 1210 min-1. For reference, unstimulated activity for PLC-3
in the experiment of Figure 1B was 7 min-1 at 60 nM Ca2+ and 89 min-1 at 1 M Ca2+.
Figure S3. Independent (non-synergistic) regulation of PLC-2 by Gq and G. PLC-2
activity was assayed at 60 nM Ca2+ over a range of concentrations of Gq and G chosen to
optimize fitting the data to the allosteric model of Figure 2. Activities are plotted against the
concentration of Gq (panel A) and G (panel B) at various fixed concentrations of the other
subunit. The data sown in the graphs were selected from a set of 92 duplicate determinations.
Solid lines are simulations based on the model and the parameter values obtained in the fit (Table
Figure S4. Dependence of synergism on the concentrations of G protein subunits at varying
values of the bias constants F and G. To determine the interplay of G protein concentration and
the bias constants F and G, the dependence of synergism ratios on concentration was simulated
for four F,G pairs. Values of Kb, Kq and J = 0.0015 were from Table 2. Maximal synergism is
constant for all conditions, but the steepness of the peak and its symmetry vary according to
values of F and G. The upper left panel uses F,G values from the fitting to the data of Figure 3.
The right two use either the higher or lower value for both F and G, and the lower left panel has
F and G reversed from the experimentally determined values. For equal values of F and G (top
panels), synergism peaks symmetrically at an intermediate concentration of each subunit. At
high values of F and G, the peak is sharp. At lower values, the peak becomes a broad plateau,
such that significant synergism is displayed over a wide range of activator concentrations. For
unequal values of F and G (left panels), synergism remains high along a ridge that extends to
very high concentrations of the G protein subunit with the lower bias constant. The ridge falls off
at high concentrations of the subunit with the higher bias constant. The general dependence of
peak symmetry on F and G is independent of J, but peaks become lower and more sharp as J
increases (Figure 5).
Figure S5. Synergism is more pronounced and less sensitively dependent on F and G when G
protein concentrations are low. The synergism ratio for PLC-3 was simulated for varied
concentrations of Gq and G with J=0.0015 (Table 2). G protein concentrations are listed as
reduced concentrations, the product of the actual concentration and the association equilibrium
constant with reduced [Gq] abbreviated as Q and reduced [G] abbreviated as B. The upper
five panels, with the concentrations of Gq and G equal, show how increasing the
concentrations sharpens the peak and finally decreases the maximum synergism that can be
attained. However, at very low G protein concentrations, higher values of F and or G are needed
for synergism. Thus both axes are extended 8-fold in the upper left panel. When the
concentration of one G protein subunit is, as shown in the lower left two panels, the peak
becomes asymmetric. The effect is large; the G axis of the lower left panel is extended 5-fold.
Supplemental Experimental Procedures
N-terminally His6-tagged PLC-1 (rat), PLC-2 (mouse) and PLC-3 (mouse) were
purified as described for PLC-1 . PLC-4 (human) was purified similarly but with slight
modifications. PLC-4 was extracted from Sf9 membranes with 1 M NaCl in buffer A (20mM
NaHepes (pH 7.5), 1mM MgCl2, 5 mM 2-mercaptoethanol, 10 g/mL leupeptin, 1 g/ml
aprotinin and 0.1 mM PMSF). Extracted protein was stirred with Ni-NTA resin (GE) for 2 h in
the presence of 0.1% Lubrol and 5mM imidazole. The resin was washed sequentially with buffer
A plus 1 M NaCl, 0.1 % Lubrol and 5 mM imidazole and with buffer A plus 100 mM NaCl and
5 mM imidazole until A280 reached baseline. PLC was eluted with buffer A plus 100 mM NaCl
and 150 mM imidazole. The eluate was diluted in buffer B (20 mM NaMES pH 6.0, 20 %
glycerol, 1 mM DTT and 0.1mM EDTA) and loaded onto a Mono-S column that was
equilibrated with buffer B. After washing the column with 100 mM NaCl in buffer, PLC was
eluted by a gradient of 100-500 mM NaCl in Buffer B.
Gq was purified as described . To decrease the detergent:protein ratio, purified
Gq was diluted 5-fold in buffer C (20mM NaHepes (pH7.5), 0.1 mM EDTA, 1 mM DTT, 10
M GDP, 0.5 % CHAPS and 1 mM MgCl2), adsorbed to Mono-Q and washed with buffer C to
replace cholate with CHAPS. Gq was eluted with 300mM NaCl in buffer C, diluted 10-fold in
buffer C, adsorbed to a 0.1 ml column of Q-Sepharose and eluted with 300mM NaCl in buffer C.
For the experiment shown in Table S2, the Gq-GDP was this preparation with no additional
G was purified as described by Kozasa and Gilman , and then concentrated and
switched to a CHAPS-containing buffer as was done for Gq. G was used for all
experiments except those shown in Table S1.
Protein concentrations were determined by amido black binding .
Gq was activated before assay by incubation at 30 oC with 1mM GTPS in 50 mM
NaHepes (pH 7.5), 100 mM NaCl, 4 mM MgCl2, 1 mM DTT, 50 mM (NH4)2SO4, 0.1 mg/ml
BSA and 0.4 % CHAPS . Incubation was extended to 5 h such that any Gq that had not
bound GTPS would be denatured  and would be unable to chelate G in the PLC assay.
Gq was activated at the highest concentration possible to minimize detergent in the PLC assay.
The fraction of Gq bound to GTPS was determined by monitoring bound [35S]GTPS.
PLC assay, data analysis and fitting
PLC activity was measured by monitoring hydrolysis of [3H]PIP2 on the surface of large
unilamellar vesicles (PE:PS:PIP2, 20:4:1 molar ratio; 0.25 mM total phospholipid) [43,48].
Vesicles were prepared in 50 mM NaHepes (pH 7.5), 100 mM NaCl and 2 mM EGTA. PLC
activity is proportional to the concentration of PIP2 under these conditions. The assay times (2 -
40 min) and the concentration of PLC (10 - 4000 pM) were adjusted such that PIP2 hydrolysis
remained linear with time and PLC concentration. The concentration of PLC was also kept low
enough that it did not substantially deplete the total concentration of activated Gq in the assay.
Assays were initiated by the addition of PLC at 37 oC. The final concentration of free Ca2+ in the
assays was adjusted with an EGTA buffer that contained 2 mM EGTA and a concentration of
CaCl2 calculated according to the program Bound and Determined . Data are expressed as
moles of IP3 produced per min per mole PLC.
The specific activities for PLC-s under these assay conditions are relatively high,
45,000-60,000 min-1 at optimal [Ca2+], [Gq] and [G] ( Figure 1B). Because the PIP2
substrate is 4 mol % of total lipids and activity increases linearly with substrate concentration in
this range, higher activities would be observed with the higher PIP2 concentrations that are
frequently used. Cholate and CHAPS both inhibit PLC activity, with IC50's of 0.2 mM and 0.1
mM respectively under these conditions. The CHAPS concentration in the assays, derived from
added G protein subunits, did not exceed 0.02 mM and was maintained equal among all samples
within an assay.
Protein concentrations are reported in terms of the aqueous assay volume. However,
PLC-, Gq and G all bind to the surface of the substrate vesicles and occupy an annular shell
volume about 5000-fold smaller. In addition, rotational mobility of the proteins is limited to one
dimension. Values of Kb and Kq can be readjusted by factor to approximate actual protein-
protein binding affinities, although the loss of two dimensions of rotational freedom mortgages
thermodynamic comparisons with soluble reactions.
Assay data were fit using the Marquardt-Levenberg algorithm in either StatGraphic or
SigmaPlot. Values for the parameters of the allosteric model (Figure 2) that are shown in Table
2 as “Matrix Fit” were derived by fitting a single data set for a PLC assay in which
concentrations of Gq and G were varied as shown in Figure 3. These data are representative
of two other similar experiments. We were unable to fit data from all these experiments
simultaneously because absolute activities vary among preparations of substrate vesicles. EC50
values for the G protein subunits vary 2- to 3-fold among experiments, presumably because of
variable binding to the vesicles. Parameters for PLC-2 shown in Table S3 were calculated
similarly to those for PLC-3.
Parameters for PLC-3 shown as “4-point fits” (Table 2) were derived from three
separate experiments in which PLC activity (v in the equations below) was assayed in the
presence of saturating concentrations of Gq, G (3 determinations in each experiment), both (
6-8 determinations) or neither (4 determinations). Parameters were then calculated as Z = v for
the combination of Gq and G; J = v / Z without G protein; F = v/(J•(Z-v)) for Gq alone;
and G = v/(J•(Z-v)) for G alone. Data shown are the means of the values calculated from each
46. Kozasa T. and Gilman A.G. (1995). Purification of recombinant G proteins from Sf9
cells by hexahistidine tagging of associated subunits. Characterization of 12 and inhibition of
adenylyl cyclase by z. J. Biol. Chem. 270: 1734-1741.
47. Schaffner W. and Weissmann C. (1973). A rapid, sensitive, and specific method for the
determination of protein in dilute solution. Anal. Biochem. 56: 502-514.
proteins that activate the 1 isozyme of phosphoinositide-specific phospholipase C. J. Biol.
Chem. 266: 18206-18216.
Blank J.L., Ross A.H., and Exton J.H. (1991). Purification and characterization of two G-