Adrenergic modulation of NMDA receptors
in prefrontal cortex is differentially regulated
by RGS proteins and spinophilin
Wenhua Liu*, Eunice Y. Yuen*, Patrick B. Allen†, Jian Feng*, Paul Greengard‡, and Zhen Yan*§
*Department of Physiology and Biophysics, School of Medicine and Biomedical Sciences, State University of New York, Buffalo, NY 14214;†Department of
Psychiatry, Yale University School of Medicine, New Haven, CT 06508; and‡Laboratory of Molecular and Cellular Neuroscience, The Rockefeller University,
New York, NY 10021
Edited by Richard L. Huganir, Johns Hopkins University School of Medicine, Baltimore, MD, and approved October 6, 2006 (received for review June 2, 2006)
The noradrenergic system in the prefrontal cortex (PFC) is involved
ing memory and mood control. To understand the functions of the
noradrenergic system, we examined the regulation of NMDA
receptors (NMDARs), key players in cognition and emotion, by ?1-
and ?2-adrenergic receptors (?1-ARs, ?2-ARs) in PFC pyramidal
inhibitor reduced the amplitude but not paired-pulse ratio of
NMDAR-mediated excitatory postsynaptic currents (EPSC) in PFC
slices. Specific ?1-AR or ?2-AR agonists also decreased NMDAR-
EPSC amplitude and whole-cell NMDAR current amplitude in dis-
sociated PFC neurons. The ?1-AR effect depended on the phos-
pholipase C–inositol 1,4,5-trisphosphate–Ca2?pathway, whereas
the ?2-AR effect depended on protein kinase A and the microtu-
bule-based transport of NMDARs that is regulated by ERK signal-
ing. Furthermore, two members of the RGS family, RGS2 and RGS4,
were found to down-regulate the effect of ?1-AR on NMDAR
currents, whereas only RGS4 was involved in inhibiting ?2-AR
regulation of NMDAR currents. The regulating effects of RGS2/4 on
?1-AR signaling were lost in mutant mice lacking spinophilin,
which binds several RGS members and G protein-coupled recep-
tors, whereas the effect of RGS4 on ?2-AR signaling was not
altered in spinophilin-knockout mice. Our work suggests that
activation of ?1-ARs or ?2-ARs suppresses NMDAR currents in PFC
neurons by distinct mechanisms. The effect of ?1-ARs is modified
by RGS2/4 that are recruited to the receptor complex by spinophi-
lin, whereas the effect of ?2-ARs is modified by RGS4 independent
adrenergic receptor ? RGS4 ? neuropsychiatric diseases ? G protein-coupled
receptor ? microtubule
sequence information, receptor pharmacology, and signaling
mechanisms. Although ?-ARs are located mainly in the cardio-
vascular system, most ?1- and ?2-AR subtypes (except ?1Cand
?2B) are highly expressed in CNS regions, e.g., prefrontal cortex
(PFC), hippocampus, and brainstem. Norepinephrine, through
the action of ?1-ARs and ?2-ARs, has been implicated in many
key functions of PFC, including working memory and emotional
control (1–3). An aberrant noradrenergic system, complement-
ing altered serotonergic or dopaminergic signaling, contributes
significantly to the pathophysiology of a variety of neuropsychi-
atric diseases associated with PFC dysfunction, such as depres-
sion, anxiety, schizophrenia, and attention-deficit hyperactivity
disorder (4–7). Therefore, modifying noradrenergic signaling
has been considered one of the key therapeutic actions of many
antidepressants, anxiolytic drugs, and antipsychotics (8, 9). To
understand the functional role of ?1- and ?2-ARs, we need to
know their cellular targets that are important for cognition and
emotion. The NMDAR channel has been implicated in normal
cognitive processes and mental disorders (10–12), which makes
drenergic receptors (ARs) can be divided into three main
types: ?1 (?1A–D), ?2 (?2A–C), and ? (1–3), based on
it a potentially important target by which ?1- and ?2-ARs may
regulate PFC functioning.
Activation of ?1-ARs or ?2-ARs Reduces NMDAR-Mediated Currents in
PFC Pyramidal Neurons. We first examined the effect of the
noradrenergic system on NMDAR-mediated excitatory postsyn-
aptic currents (NMDAR-EPSC) evoked by stimulation of syn-
aptic NMDARs in PFC slices. As shown in Fig. 1A, application
of 100 ?M natural neurotransmitter norepinephrine induced a
strong and persistent reduction in the amplitude of NMDAR-
EPSC, whereas in parallel control measurements where no
norepinephrine was administered, NMDAR-EPSC remained
stable throughout the recording. Application of 20 ?M desipra-
mine, a norepinephrine transporter inhibitor, to activate endog-
enous noradrenergic receptors also caused a potent and long-
lasting reduction of the NMDAR-EPSC amplitude (Fig. 1B).
The specific ?1-AR agonist cirazoline (40 ?M) produced a
similar reducing effect, which was blocked by 40 ?M prazosin, a
specific ?1-AR antagonist (Fig. 1C). In a sample of neurons we
tested, norepinephrine, norepinephrine transporter inhibitors,
and ?1-AR agonist all significantly suppressed synaptic
NMDAR-mediated responses (norepinephrine: 31.5 ? 8.5%,
n ? 4; desipramine: 34.2 ? 2.7%, n ? 6; nisoxetine (an
norepinephrine transporter inhibitor, 50 ?M): 36.3 ? 5.3%, n ?
3; cirazoline: 36.2 ? 3.3%, n ? 4). We further examined the
noradrenergic effect on NMDAR-EPSC evoked by paired
pulses, a measure that is sensitive to changes in the probability
of transmitter release (13). As shown in Fig. 1D, application of
desipramine reduced the amplitudes of NMDAR-EPSC trig-
gered by both pulses, but it did not cause a significant change in
the ratio of the paired-pulse facilitation (control: 1.72 ? 0.1;
desipramine: 1.73 ? 0.1, n ? 5). This observation suggests that
activation of noradrenergic receptors in PFC pyramidal neurons
is likely to induce a change in postsynaptic NMDARs rather than
To verify the potential influence of ?1-ARs on NMDARs, we
examined the effect of ?1-ARs on NMDAR-mediated whole-
cell currents in dissociated PFC pyramidal neurons. As shown in
Fig. 1E, application of 40 ?M cirazoline caused a reversible
Author contributions: J.F. and Z.Y. designed research; W.L. and E.Y.Y. performed research;
The authors declare no conflict of interest.
This article is a PNAS direct submission.
Abbreviations: 2APB, 2-aminoethoxydiphenyl borane; AR, adrenergic receptor; EPSC, ex-
citatory postsynaptic currents; GPCR, G protein-coupled receptor; IP3, inositol 1,4,5-
trisphosphate; NMDAR, NMDA receptor; NR2B, NMDAR 2B; PFC, prefrontal cortex; PLC,
phospholipase C; RGS, regulators of G protein signaling.
§To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
© 2006 by The National Academy of Sciences of the USA
November 28, 2006 ?
vol. 103 ?
(18.2 ? 0.5%, n ? 130). This effect was significantly attenuated
by 40 ?M prazosin (Fig. 1F; 5.5 ? 0.9%, n ? 10), confirming that
?1-AR activation suppresses NMDAR currents.
We also examined the effect of ?2-ARs on NMDAR-
mediated synaptic and whole-cell currents. As shown in Fig. 2A,
application of 100 ?M specific ?2-AR agonist clonidine potently
reduced the amplitude of NMDAR-EPSC (38.3 ? 3.2%, n ? 5),
which was diminished by 40 ?M specific ?2-AR antagonist
idazoxan (9.0 ? 2.3%, n ? 5). The effect of the norepinephrine
transporter inhibitor desipramine on NMDAR-EPSC was par-
tially reduced by prazosin or idazoxan alone, but it was almost
abolished by coapplication of the two antagonists (Fig. 2 B and
C), suggesting the involvement of both ?1-ARs and ?2-ARs.
Clonidine also decreased NMDA-evoked currents in dissociated
PFC neurons (Fig. 2D; 15.6 ? 0.8%, n ? 30), which was
abolished by 60 ?M yohimbine (4.9 ? 0.3%, n ? 7), another
specific ?2-AR antagonist (Fig. 2E). As summarized in Fig. 2F,
the effects of different ?2-AR agonists on NMDAR currents
were all diminished by various ?2-AR antagonists, confirming
that ?2-AR activation also suppresses NMDAR currents.
Because both ?1-AR and ?2-AR reduce NMDAR currents, we
also examined whether their effects were additive to determine
As shown in Fig. 8A, which is published as supporting information
on the PNAS web site, in the presence of 100 ?M ?2-AR agonist
clonidine, 40 ?M ?1-AR agonist cirazoline caused further suppres-
sion of NMDAR currents (cirazoline: 18.5 ? 1.2%; clonidine:
whole-cell NMDAR currents. (A–C) Plot of peak NMDAR-EPSC in PFC slices
showing the effect of 100 ?M norepinephrine (A), 20 ?M norepinephrine
absence or presence of 40 ?M ?1-AR antagonist prazosin (C). (D) Plot of peak
NMDAR-EPSC evoked by double pulses (interstimuli interval, 100 ms) as a
function of time and 20 ?M desipramine application. (Insets A, B, and D)
Representative current traces (average of three trials) at time points denoted
by #. (Scale bars: 100 pA, 0.1 s.) (E and F) Plot of peak 100 ?M NMDA-evoked
currents in dissociated PFC pyramidal neurons showing the effect of 40 ?M
cirazoline (E) and its blockade by 40 ?M prazosin (F). (Inset E) Representative
current traces (at time points denoted by #). (Scale bars: 100 pA, 1 s.) (Inset F)
Cumulative data (mean ? SEM) summarizing the percentage reduction of
NMDAR currents by cirazoline in the absence or presence of prazosin.*, P ?
Activation of ?1-ARs reduces the amplitude of NMDAR-EPSC and
whole-cell NMDAR currents. (A) Plot of peak NMDAR-EPSC in PFC slices
showing the effect of 100 ?M ?2-AR agonist clonidine in the absence or
presence 40 ?M idazoxan, a specific ?2-AR antagonist. (Inset) Representative
current traces (at time points denoted by #). (Scale bars: 100 pA, 0.1 s.) (B) Plot
of peak NMDAR-EPSC showing that blocking both ?1 and ?2 receptors with
prazosin and idazoxan prevented the effect of norepinephrine transporter
inhibitor desipramine. (C) Cumulative data (mean ? SEM) summarizing the
percentage reduction of NMDAR-EPSC by desipramine in the absence or
presence of different antagonists.*, P ? 0.01, ANOVA. (D and E) Plot of peak
NMDAR currents showing the effect of 100 ?M clonidine (D) and its blockade
by 60 ?M yohimbine (E), a specific ?2-AR antagonist. (Insets D and E) Repre-
sentative current traces (at time points denoted by #). (Scale bars: 100 pA, 1 s.)
(F) Cumulative data (mean ? SEM) summarizing the percentage reduction of
NMDAR currents by different ?2-AR agonists (clonidine or guanfacine at 100
?M) in the absence or presence of different ?2-AR antagonists (yohimbine or
idazoxan).*, P ? 0.01, ANOVA.
Activation of ?2-ARs reduces the amplitude of NMDAR-EPSC and
Liu et al. PNAS ?
November 28, 2006 ?
vol. 103 ?
no. 48 ?
that the effects of ?1-AR and ?2-AR are additive. Moreover, the
NMDAR 2B (NR2B) antagonist ifenprodil (Fig. 8 B and C),
suggesting that ?1-AR and ?2-AR target different pools of
We further examined whether ?1-AR and ?2-AR agonists
may have direct effects on NMDARs. As shown in Fig. 9, which
is published as supporting information on the PNAS web site,
both cirazoline and clonidine produced similar effects on
NMDAR currents at different voltages (?60 mV, ?40 mV, ?20
mV), suggesting that these compounds do not cause a voltage-
dependent block of NMDAR channels. Moreover, both cirazo-
line and clonidine produced similar effects on NMDAR currents
in the presence of saturating (20 ?M) or subsaturating (1 ?M)
concentrations of glycine, suggesting that these compounds do
not reduce affinity of the NMDARs for glycine.
The ?1- and ?2-AR of NMDAR Currents Depend on Different Signaling
Pathways and Cellular Mechanisms. Next, we examined the signal-
ing mechanisms underlying the reduction of NMDAR currents
by ?1-ARs and ?2-ARs, which are coupled to Gq and Gi/o
proteins, respectively (14). As shown in Fig. 3A, application of
the 1 ?M phospholipase C (PLC) inhibitor U73122 significantly
blocked the cirazoline-induced decrease of NMDAR currents
(6.6 ? 0.9%, n ? 7, Fig. 3C), suggesting the involvement of the
PLC pathway in ?1-AR signaling. PLC activation catalyzes
the hydrolysis of membrane phosphoinositol lipids, leading to
the release of inositol 1,4,5-trisphosphate (IP3) and diacylglyc-
erol. 2-Aminoethoxydiphenyl borane (2APB; 15 ?M), a mem-
brane-permeable IP3receptor antagonist, substantially blocked
the reduction of NMDAR currents by cirazoline (Fig. 3B; 6.1 ?
0.8%, n ? 7; Fig. 3C). Dialysis with 10 mM 1,2-bis(2-
aminophenoxy)ethane-N,N,N?,N?-tetraacetate (BAPTA), a
Ca2?chelator, also diminished the effect of cirazoline (4.3 ?
0.7%, n ? 6; Fig. 3C), whereas 1 ?M selective PKC inhibitor
bisindolylmaleimide I failed to do so (16.2 ? 1.0%, n ? 6; Fig.
3C). On the other hand, clonidine lost the capability to suppress
NMDAR currents in the presence of 50 ?M PKA activator
cpt-cAMP (Fig. 3D; 4.4 ? 1.1%, n ? 7; Fig. 3F), and 10 ?M PKA
inhibitor H89 reduced NMDAR currents and occluded the
effect of clonidine (Fig. 3E; 2.7 ? 1.5%, n ? 5; Fig. 3F). These
results suggest that the PLC–IP3–Ca2?pathway is involved in
?1-AR regulation of NMDAR currents, whereas inhibition of
PKA signaling is required for ?2-ARs to regulate NMDAR
Recent studies have shown that NMDAR function can be
regulated through a mechanism dependent on microtubule/
motor protein-based dendritic transport of NMDARs that is
role of ERK and microtubules in adrenergic regulation of
U0126, a specific inhibitor of MEK (the kinase upstream of
ERK), the cirazoline-induced decrease of NMDAR currents was
not affected (17.3 ? 1.2%, n ? 5, Fig. 4C), whereas the effect of
6; Fig. 4F). Application of 30 ?M microtubule-depolarizing
agent colchicine did not prevent cirazoline from reducing
NMDAR currents (Fig. 4B; 17.5 ? 1.0%, n ? 5; Fig. 4C), but it
occluded the effect of subsequently applied clonidine (Fig. 4E;
3.5 ? 1.3%, n ? 6; Fig. 4F). The actin-depolarizing agent
latrunculin B (5 ?M) or 50 ?M dynamin-inhibitory peptide
failed to affect either ?1- or ?2-adrenergic regulation of
NMDAR currents (Fig. 4 C and F), suggesting the lack of
involvement of actin cytoskeleton or clathrin-mediated endocy-
tosis in the process. Taken together, these results suggest that
?2-AR but not ?1-AR regulation of NMDAR currents is
through a mechanism dependent on microtubule stability.
To confirm further that ?2-AR affects microtubule-based
NMDAR trafficking, we carried out a quantitative surface
immunostaining assay in PFC cultures transfected with GFP-
NR2B (GFP tag is placed at NR2B extracellular N terminus).
The surface distribution of recombinant NR2B was assessed by
immunostaining with anti-GFP primary antibody followed by
rhodamine-conjugated secondary antibody in nonpermeabilized
conditions (16, 17). As shown in Fig. 10, which is published as
supporting information on the PNAS web site, in neurons
treated with 100 ?M ?2-AR agonist clonidine for 10 min, the
fluorescent GFP-NR2B surface clusters on dendrites were sig-
nificantly reduced in both the density (control: 38.2 ? 1.6
clusters per 50 ?m, n ? 14; clonidine-treated neurons: 19.8 ?
0.95 clusters per 50 ?m, n ? 9; P ? 0.01, ANOVA) and the size
(control: 0.3 ? 0.03 ?m2, n ? 14; clonidine-treated neurons:
0.18 ? 0.02 ?m2; P ? 0.01, ANOVA), which was blocked by 10
?M microtubule stabilizer taxol (cluster density: 37.2 ? 3.1
clusters per 50 ?m; cluster size: 0.3 ? 0.02 ?m2, n ? 6). In
contrast, treatment with 100 ?M ?1-AR agonist cirazoline for 10
min caused little change on surface NR2B clusters. These results
suggest the involvement of a microtubule-dependent mechanism
in ?2-AR regulation of NMDARs.
regulation of NMDAR currents, respectively. (A and B) Plot of peak NMDAR
currents showing the effect of 40 ?M cirazoline in the absence or presence of
1 ?M PLC inhibitor U73122 (A) or 15 ?M IP3receptor antagonist 2APB (B). (D
and E) Plot of peak NMDAR currents showing the effect of 100 ?M clonidine
inhibitor H89 (E). (Insets A, B, D, and E) Representative current traces (at time
points denoted by #). (Scale bars: 100 pA, 1 s.) (C and F) Cumulative data
(mean ? SEM) showing the percentage reduction of NMDAR currents by
cirazoline (C) or clonidine (F) under various treatments.*, P ? 0.01, ANOVA.
The PLC–IP3–Ca2?or PKA pathway is involved in ?1- or ?2-adrenergic
www.pnas.org?cgi?doi?10.1073?pnas.0604560103Liu et al.
The ?1-AR and ?2-AR Effects on NMDAR Currents Are Modulated by
Different Regulators of G Protein Signaling (RGS) Proteins in PFC
Pyramidal Neurons. Because RGS proteins play an important role
in modulating G protein-coupled receptor (GPCR) signaling
(18), we examined their influence on ?1-AR and ?2-AR regu-
lation of NMDARs. To do so, we infused antibodies against
specific RGS family members into neurons through the record-
ing pipette to achieve an effective inhibition of endogenous RGS
function (19, 20), and then we tested whether these antibodies
could affect ?1 or ?2 regulation of NMDAR currents. We
focused on two RGS family members, RGS2 and RGS4, because
RGS2 is dynamically responsive to neuronal activity in a unique
fashion (21), whereas RGS4, which exhibits the highest expres-
sion in PFC (22), has recently been identified as a schizophrenia-
susceptibility gene (23–26).
As shown in Fig. 5A and B, bath application of 40 ?M
cirazoline reduced the amplitude of NMDAR currents, whereas
in neurons dialyzed with the anti-RGS2 (19) or anti-RGS4 (20)
antibody (4 ?g/ml), the reduction of NMDAR currents by
cirazoline was strongly augmented. In a sample of neurons we
tested (Fig. 5C), both RGS2 and RGS4 antibodies caused a
significant potentiation (?37% increase and ?26% increase,
respectively) in the effect of cirazoline on NMDAR currents
[control (?): 18.7 ? 0.4%, n ? 14; anti-RGS2: 25.6 ? 0.7%, n ?
RGS2 antibody failed to do so (18.8 ? 0.6%, n ? 6).
The role of RGS2 and RGS4 in ?2-AR signaling was also
examined. As shown in Fig. 5 D and E, the reduction of NMDAR
currents by clonidine was strongly potentiated (?43% increase)
by dialysis with the anti-RGS4 antibody [control (?): 15.2 ?
0.9%, n ? 9; anti-RGS4: 21.7 ? 0.5%, n ? 9; Fig. 5F], but not
the anti-RGS2 antibody (15.5 ? 0.6%, n ? 8; Fig. 5F). These
results suggest that both RGS2 and RGS4 are involved in
inhibiting ?1-AR signaling, whereas only RGS4 participates in
dampening ?2-AR regulation of NMDAR currents.
The Scaffold Protein Spinophilin Is Differentially Involved in RGS
Modulation of the ?1- or ?2-AR Effect on NMDAR Currents. It has
been shown that the multidomain scaffolding protein spinophilin
binds several GPCRs (27, 28) and RGS proteins (29), suggesting
that spinophilin may actively regulate GPCR signaling by re-
cruiting RGS proteins to the receptor complex. To test whether
spinophilin is involved in adrenergic regulation of NMDARs, we
examined the RGS modulation of ?1-AR and ?2-AR effects on
NMDAR currents in spinophilin-knockout mice (30).
As shown in Fig. 6A, both anti-RGS2 and anti-RGS4 anti-
through a mechanism involving ERK-regulated microtubule stability. (A, B, D,
and E) Plot of peak NMDAR currents showing that the effect of 40 ?M
by 20 ?M MEK/ERK inhibitor U0126 (A and D) or 30 ?M microtubule-
depolymerizing agent colchicine (B and E). (Insets) Representative current
traces (at time points denoted by #). (Scale bars: 100 pA, 1 s.) (C and F)
currents by cirazoline (C) or clonidine (F) in the absence or presence of various
agents.*, P ? 0.01, ANOVA.
The suppression of NMDAR currents by ?2-AR but not ?1-AR is
the effect of cirazoline (A) or clonidine (D) in neurons dialyzed with 4 ?g/ml
anti-RGS2 or anti-RGS4 antibody. (B and E) Representative current traces (at
time points denoted by #). (Scale bars: 100 pA, 1 s.) (C and F) Cumulative data
(mean ? SEM) showing the percentage reduction of NMDAR currents by
cirazoline (C) or clonidine (F) in the absence or presence of different RGS
antibodies.*, P ? 0.01, ANOVA.
Different RGS proteins modulate the effect of ?1-ARs or ?2-ARs on
Liu et al. PNAS ?
November 28, 2006 ?
vol. 103 ?
no. 48 ?
bodies significantly potentiated the effect of cirazoline on
NMDAR currents in neurons from wild-type mice (anti-RGS2:
38.9 ? 1.3% increase, n ? 10; anti-RGS4: 28.5 ? 2.5% increase,
n ? 6). However, in neurons from spinophilin?/?mice, dialysis
with the anti-RGS2 or anti-RGS4 antibody failed to augment the
effect of cirazoline on NMDAR currents (anti-RGS2: ?4.5 ?
1.1% increase, n ? 11; anti-RGS4: 6.3 ? 1.5% increase, n ? 7).
Representative examples from spinophilin?/?mice are shown in
Fig. 6B. The loss of RGS2/4-mediated attenuation of ?1-AR
signaling in spinophilin-deficient cells led to a 25.7 ? 0.5% (n ?
11) enhancement in the cirazoline effect on NMDAR currents.
In contrast, the effect of anti-RGS2 or anti-RGS4 on clonidine
regulation of NMDAR currents remained unaltered in neurons
from wild-type vs. spinophilin?/?mice (Fig. 6C; anti-RGS2:
wild-type, 1.3 ? 0.5% increase, n ? 8; knockout, 3.3 ? 0.3%
increase, n ? 6; anti-RGS4: wild-type, 40.5 ? 3.1% increase, n ?
6; knockout, 38.7 ? 2.9% increase, n ? 6). The intact RGS4
modulation of ?2-AR signaling in spinophilin-deficient cells led
to little change (2.0 ? 0.3%, n ? 11) in the clonidine effect on
NMDAR currents. These data suggest that the scaffold protein
spinophilin functions to facilitate the ability of RGS2/4 to inhibit
?1-adrenergic regulation of NMDAR currents, but it is not
involved in RGS4 modulation of the ?2-AR effect on NMDAR
In this work, we found that activating ?1-ARs or ?2-ARs inhibits
NMDAR currents through distinct mechanisms in PFC pyrami-
dal neurons, with the PLC–IP3–Ca2?pathway involved in the
?1-AR effect and PKA as well as the ERK-regulated microtu-
bule-based transport of NMDARs involved in the ?2-AR effect.
Moreover, the scaffold protein spinophilin selectively facilitates
the RGS2/4 modulation of the ?1-AR effect on NMDAR
currents (Fig. 7). Given the critical role of NMDA signaling in
controlling neuronal function, our results provide a potential
molecular and cellular mechanism for adrenergic regulation of
emotion and cognition subserved by PFC.
Previous studies suggested that NMDAR activity could be
regulated by serine/threonine and tyrosine phosphorylation by
GPCR-induced signaling pathways (31). Emerging evidence
indicates that another key factor in the regulation of NMDAR
channel functions is the trafficking of NMDARs, which involves
exiting from ER, transporting along dendritic microtubules,
delivery to actin-enriched postsynaptic density, internalization,
and lateral diffusion at synaptic and extrasynaptic sites (15).
Recent studies have revealed the different mechanisms under-
lying serotonin and dopamine receptor mediated-regulation of
NMDAR trafficking and function (16, 32, 33). The present
finding on adrenergic regulation of NMDARs has introduced
additional players in the process, such as RGS proteins and
of GPCR signaling by accelerating the GTPase activity of G?
subunits (34, 18). However, the mechanism by which RGS
proteins recognize GPCRs to confer signaling specificity re-
mains largely unknown. In vitro studies have found that RGS2
shows preference for Gqclass G?subunits (35), whereas RGS4
attenuates both Gi- and Gq-mediated signaling (36–38). Bio-
chemical evidence suggests that the divergent N-terminal do-
main of RGS proteins participates in GPCR recognition (39),
and signaling specificity in vivo is conferred by interaction of
RGS proteins with receptor complexes (40). The differential
involvement of endogenous RGS2 and RGS4 in the regulation
of NMDAR currents by ?1-ARs (Gq-coupled) or ?2-ARs (Gi-
coupled) suggests that RGS proteins act in a GPCR-specific
manner in native neurons.
A recent study has revealed that the scaffolding protein
spinophilin, which binds the third intracellular loop of ?-ARs
and the N-terminal domain of several RGS proteins, enhances
the ability of RGS2 to inhibit ?1B-AR-evoked Ca2?signaling
(29). Interestingly, we found that the RGS2/4-mediated atten-
currents is lost in spinophilin-knockout mice. (A and C) Cumulative data
(mean ? SEM) showing the percentage increase of the effect of cirazoline (A)
or clonidine (C) on NMDAR currents by 4 ?g/ml anti-RGS2 or anti-RGS4
effect of cirazoline in neurons dialyzed with an anti-RGS2 or anti-RGS4 anti-
body in PFC pyramidal neurons from spinophilin-knockout mice.
The RGS modulation of the ?1-AR but not ?2-AR effect on NMDAR
adrenergic regulation of NMDAR function in PFC pyramidal neurons. ?1-AR
(Gq-coupled) reduces NMDAR currents via the PLC-IP3 pathway, whereas
?2-AR (Gi/o-coupled) decreases NMDAR currents by interfering with the mi-
crotubule (MT)/motor protein-based transport of NMDARs that is regulated
signaling, whereas RGS4 is involved in both ?1-AR and ?2-AR signaling. The
scaffold protein spinophilin selectively facilitates the RGS2/4 modulation of
the ?1-AR effect on NMDAR currents.
Proposed model showing the mechanisms underlying ?1- and ?2-
www.pnas.org?cgi?doi?10.1073?pnas.0604560103Liu et al.