NMR analysis of G-protein βγ subunit complexes
reveals a dynamic Gα-Gβγ subunit interface
and multiple protein recognition modes
Alan V. Smrckaa,b,c,1, Nessim Kichika, Teresa Tarragóa, Michael Burroughsb, Min-Sun Parkc,d, Nathan K. Itogae, Harry A.
Sternc,d, Barry M. Willardsone, and Ernest Giralta,f,2
aInstitute for Research in Biomedicine, Barcelona Science Park, Barcelona, Spain;
Biochemistry and Biophysics, and
and Biochemistry, Brigham Young University, Provo, UT 84602; and
1. E-08028, Barcelona, Spain
bDepartment of Pharmacology and Physiology,
eDepartment of Chemistry
dDepartment of Chemistry, University of Rochester School of Medicine, Rochester, NY 14642;
fDepartment of Organic Chemistry, University of Barcelona. Martí Franqués,
Edited by Melvin I. Simon, California Institute of Technology, Pasadena, CA, and approved November 5, 2009 (received for review August 21, 2009)
G-protein βγ (Gβγ) subunits interact with a wide range of molecular
partners including: Gα subunits, effectors, peptides, and small mol-
ecule inhibitors. The molecular mechanisms underlying the ability
to accommodate this wide range of structurally distinct binding
partners are not well understood. To uncover the role of protein
flexibility and alterations in protein conformation in molecular rec-
ognition by Gβγ, a method for site-specific15N-labeling of Gβ-Trp
residue backbone and indole amines in insect cells was developed.
Transverse Relaxation Optimized Spectroscopy-Heteronuclear Sin-
gle-Quantum Coherence Nuclear Magnetic Resonance (TROSY-
HSQC NMR) analysis of15N-Trp Gβγ identified well-dispersed sig-
nals for the individual Trp residue side chain and amide positions.
Surprisingly, a wide range of signal intensities was observed in the
spectrum, likely representing a range of backbone and side chain
mobilities. The signal for GβW99 indole was very intense, suggest-
ing a high level of mobility on the protein surface and molecular
dynamics simulations indicate that GβW99 is highly mobile on the
nanosecond timescale in comparison with other Gβ tryptophans.
Binding of peptides and phosducin dramatically altered the mo-
bility of GβW99 and GβW332 in the binding site and the chemical
shifts at sites distant from the direct binding surface in distinct
ways. In contrast, binding of Gαi1-GDP to Gβγ had relatively little
effect on the spectrum and, most surprisingly, did not significantly
alter Trp mobility at the subunit interface. This suggests the inac-
tive heterotrimer in solution adopts a conformation with an open
subunit interface a large percentage of the time. Overall, these
data show that Gβγ subunits explore a range of conformations that
can be exploited during molecular recognition by diverse binding
Beta gamma subunits ∣ clam shell ∣ Hot Spots ∣ Molecular Recognition ∣
in cellular physiology through a series of highly regulated protein-
protein interactions (1–3). The array of functionally and struc-
turally diverse binding partners both upstream and downstream
of Gβγ is not consistent with a simple binding mechanism that
relies on well-defined structure or sequence modules (1). Rather,
Gβγ appears to have multiple binding modes for interacting with
receptors, G-protein α subunits, and downstream effectors (1, 4).
We have proposed that Gβγ subunits have a protein interaction
“hot spot” that mediates interactions between Gβγ and down-
stream signaling molecules and other binding ligands (5, 6).
Hot spots are subsets of amino acids in crystallographic protein–
protein interfaces that contribute the majority of the interaction
energy (7–10). These amino acids tend to be clustered at the cen-
ter of the interface and present diversity in chemistry that can
participate in multiple types of bonding interactions that can
be exploited by different binding partners. Additionally, where
he Gβγ subunit complex performs a central function trans-
ducing signals from G-protein-coupled receptors to changes
single protein binding sites interact with multiple diverse ligands,
hot spots are flexible and present binding epitopes of variable size
and shape, allowing recognition of diverse structures (8). It seems
that the Gβγ subunit hot spot must be flexible in order to mediate
its multiple diverse binding interactions.
G-protein βγ subunits are potential therapeutic targets based
on studies in animal models using protein-based inhibitors of Gβγ
functions (11, 12) or genetic deletion of downstream targets of
Gβγ signaling (13–15). More recently, small molecule inhibitors
of Gβγ subunit signaling have been identified that bind to the Gβγ
hot spot and inhibit Gβγ protein–protein interactions (16). These
have been used in cellular and animal models to further implicate
Gβγ signaling as a potential therapeutic target in pain (17), in-
flammation (18), and cancer (19). Understanding the flexible nat-
ure of this surface is important for developing approaches to
target Gβγ with “drug-like” molecules for treatment of disease.
To obtain information about Gβγ flexibility and its importance
in ligand binding, we developed an NMR spectroscopy protocol
to report on alterations in Gβγ structure. Whereas NMR can pro-
vide valuable information about protein structure and dynamics,
there are several limitations that must be overcome for success. It
is difficult to get informative spectra as the size of the protein
increases above 25 kDa because sensitivity decreases as a result
of relaxation-dependent line broadening, and the increased num-
ber of protons in the macromolecule results in complex spectra
with a high degree of spectral overlap (20, 21). To overcome
this issue, we prepared Gβγ, selectively labeled with15N-Trp,
to perform two-dimensional
ments. We have used a similar approach to study15N-indole-
labeled prolyl oligopeptidase (22). Using this method, we exam-
ined structural alterations that occurred upon formation of
complexes between Gβγ and multiple binding partners. The data
reveal that despite binding to similar sets of amino acids on Gβγ,
each partner produces unique alterations in the spectra, indicat-
ing that different protein conformations are involved in recogni-
tion of different binding partners.
Labeled Gβγ. Gβγ was labeled with 2-15N-Trp allowing us to obtain
1H-15N-TROSY-HSQC Analysis of
Author contributions: A.V.S., N.K., T.T., and E.G. designed research; A.V.S., N.K., M.B., and
M.P. performed research; A.V.S., M.B., N.K.I., B.M.W., and E.G. contributed new reagents/
analytic tools; A.V.S., N.K., T.T., M.P., H.A.S., B.M.W., and E.G. analyzed data; A.V.S., H.A.S.,
B.M.W., and E.G. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1To whom correspondence may be addressed. E-mail: Alan_Smrcka@urmc.rochester.edu.
2To whom correspondence may be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/cgi/content/full/
www.pnas.org/cgi/doi/10.1073/pnas.0909503107PNAS ∣ January 12, 2010 ∣ ol. 107 ∣
o. 2 ∣ 639–644n
dispersion (Fig. 1A). For all the peaks, those assigned as the
indole or amide are designated with i or a, respectively. There
β propeller of Gβγ, and at critical positions for protein–protein
interactions in the Gβγ hot spot (GβW99 and GβW332) (Fig. 1C).
Of the expected 16 peaks, 14 are observed. The two missing reso-
nances could be either exchange broadened, or overlapping with
other signals, as will be discussed, but the overall spectrum was
consistent with the expected results. To assign the peaks in the
spectrum, we systematically replaced each Trp residue with Phe
by site-directed mutagenesis. Two-dimensional1H-15N-TROSY-
HSQC spectra were then obtained for each of the purified
Gβγ-Trp mutants and compared with the wt-Gβγ-spectrum to
identify missing peaks. Details of the assignments are discussed
in SI Text and shown in Figs. S1 and S2. One peak was identified
as originating from unfolding protein (UP in Fig. 1A) because its
presence was inconsistent and tended to appear and increase as
the sample aged. A striking feature of the spectrum was the range
of peak intensities, with some low intensity signals (GβW99a,
GβW332a and GβW82i) and a very intense signal for GβW99i
1H-15N-TROSY-HSQC spectra with good
both corepositionsof the
of dynamics of individual Trp side-chain and backbone positions
on both fast and intermediate chemical exchange time scales, and
indicate significant flexibility of Gβ in the hot spot.
Molecular Dynamics Simulations. We hypothesize that the very
intense peak for GβW99i is the result of unusually high mobility
on very fast time scale on the protein surface. There are two
surface Trp residues, GβW332 and GβW99, but only GβW99i
gives this very intense signal. We examined the crystallographic
B factors for Trp residues in free Gβγ and only GβW99 had a
B factor that was significantly higher than the others (23). To
further examine the mobility of GβW99, we performed molecular
dynamics simulations. Four dihedral angles were monitored for
each Gβ-Trp during 10 independent 8 ns simulations. GβW99
showed significant alteration in these angles on this time scale
suggesting that GβW99 has the potential to be highly mobile
on the protein surface (Fig. 2 Left and Figs. S5 and S7). Interest-
ingly, the backbone ϕ and ψ bond angles in addition to the side
chain χ1and χ2bond angles, showed flexibility and were able to
adopt multiple states during these simulations. GβW332, on the
other hand, showed little change in dihedral angles (Fig. 2Right,
Figs. S6 and S8) and is similar to the other Trp residues (Figs. S7
and S8). Also examined was the propensity to form hydrogen
bonds that could restrict molecular motion. GβW99 was the only
tryptophan that did not participate in hydrogen bonds during the
simulations (Table S1). Thus, these simulations support the
notion that the strong NMR signal for GβW99 results from a high
level of mobility on the nanosecond time scale.
Effects of Peptide Ligand Binding on Resonance Intensities and
Chemical Shifts. As an initial experiment to evaluate how Gβγ
adapts to binding of ligands and protein partners, we examined
alterations in the two-dimensional1H-15N-TROSY-HSQC spec-
trum upon binding of two Gβγ binding peptides derived from ran-
dom peptide phage display (5). These peptides have no sequence
homology but bind to the hot spot on Gβγ and competitively
inhibit interactions with several downstream targets. We had pre-
viously shown that, whereas these peptides bind to an overlapping
surface, different subsets of amino acids on Gβ are required for
SIGKAFKILGYPDYD (SIGK) is a linear peptide that binds
Gβγ with an apparent affinity near 1 μM (5) and the structure of a
Gβγ-SIGK cocomplex has been solved (6). In Fig. 3A is a com-
parison of the spectrum of Gβγ alone with Gβγ in the presence of
two equivalents of SIGK. The most striking alteration is the dra-
matic decrease in intensity and change in chemical shift for the
GβW99i signal. This result is consistent with the three-dimen-
sional structure of the SIGK-Gβγ cocomplex where SIGK binding
would restrict the motion of GβW99 (6).
Also observed were significant increases in the intensity for the
GβW99a, a shift in position of the GβW332a resonance and
appearance of a new signal (Fig. 3A) corresponding to one of
dimensional1H-15N-TROSY-HSQC spectrum of 150 μM15N-Trp-Gβ1γ2ΔC ac-
quired for 12 h at 30°C on a Bruker 800 MHz spectrometer. Assignments were
made as discussed inthe text andin SupplementalMaterial. The numbering is
the amino acid number followed by (a) for amide resonances or (i) for indole
resonances. UP is unfolding protein. B.1H-projection of the data in A. Black
arrow indicates the very intense resonance for GβW99i and the red arrow in-
dicates an example of a very small signal for GβW99a. C. A three-dimensional
model of the Gβ1γ1derived from coordinates 1TBG (23). Red numerals indi-
cate the numbers of the blades of the propeller, according to ref.32, and in
black numbering are the Trp residues for which signals are identified in A.
Two-dimensional1H-15N-TROSY-HSQC spectrum of Gβ1γ2ΔC. A. Two-
8-ns trajectories for ϕ, ψ, χ1, and χ2 bond angles for GβW99 (Left) and
GβW332 (Right). Calculations were performed 10 times with different
starting points for each Trp residue as described in Methods.
Molecular dynamics simulations for Gβγ-Trp resonances. Example
www.pnas.org/cgi/doi/10.1073/pnas.0909503107Smrcka et al.
the missing peaks in the unbound spectrum that can be attributed
to GβW332i (Fig. S1F). GβW332 also directly contacts SIGK in
the three-dimensional structure (6). The most likely explanation
for these observations is that in the ligand-free state, the signal for
GβW332i is unobservable, and the GβW99a signal is very weak
due to line broadening resulting from chemical exchange between
different conformations on an intermediate time scale. Direct
binding of SIGK preferentially stabilizes one of these conforma-
tions, resulting in an increase in observable NMR signal. In ad-
dition to changes that occur due to direct interactions between
the peptide and the hot spot, a number of alterations occur at
amino acids at some distance from the peptide binding site in-
cluding: GβW297a, GβW82a, GβW211a, GβW82i, GβW63i
and GβW339i. Most of these changes are not large but they
are significant and indicate that binding of ligands to the hot spot
can transmit conformational information throughout the Gβ sub-
We also examined alterations in the spectrum that occurred
with a second peptide SCARFFGTPGCT (SCAR). This peptide
is very different in sequence from SIGK and is constrained in a
cyclic conformation by an internal disulfide bond, and is a com-
petitive inhibitor of some Gβγ-effector interactions with an ap-
parent affinity for Gβγ similar to SIGK (6). Like SIGK, SCAR
binding suppresses the intense GβW99i resonance but, in the case
of SCAR, there was no effect on GβW99a (Fig. 3C). There was
also a shift in GβW332a resonance but to a different position than
SIGK and no increase in GβW332i resonance was observed. At
positions distant from the binding sites for the peptides, there are
additional similarities and differences between the spectral
changes caused by the binding of the two peptides (Fig. 3D
G-Protein α Subunit Binding. G-protein α subunits bind to Gβγ
subunits with 1–10 nM affinity (24) and also interact with the
Gβγ hot spot. The switch II region of Gα subunit has high
structural homology with SIGK and interacts with all of the same
amino acids in the hot spot (Fig. 4A and Table S2). Gα has
additional contacts outside of the hot spot, as well (Fig. 4A
and Table S2). To assess the effects of Gα on Gβγ conformation,
we assembled15N-Trp-Gβγ subunit with unlabeled myristoylated-
Gαi1-GDP (mGαi1-GDP) subunits in a 1∶1.2 molar ratio and per-
formed1H-15N-TROSY-HSQC analysis. Assembly with mGαi1-
GDP subunits resulted in remarkably little change in the
15N-Trp-Gβγ spectrum(Fig.4B).Particularly striking was thelack
of effect of mGαi1-GDP binding on the apparent mobility of
GβW99i. Whereas the binding of SIGK peptide resulted in com-
plete suppression of the GβW99i resonance, assembly with
mGαi1-GDP resulted in only a partial 30–50% suppression of this
signal (Fig. 4C). This suggests that GβW99 retains a high level of
mobility in the assembled inactive heterotrimer. There is also a
loss of the GβW99a signal with mGαi1-GDP binding but this is
difficult to interpret because the signal is very weak in the absence
of mGαi1-GDP and the noise increases significantly upon forma-
tion of the large Gαβγ complex, which possibly leads to loss of low
intensity peaks. Upon addition of the G-protein activator, alumi-
num fluoride, the intensity of GβW99i is restored to the intensity
observed in the unbound state. The lack of increase in intensity of
other signals suggests that Gα remains bound to Gβγ because the
overall signal should increase due to the decrease in molecular
weight upon Gα dissociation (Fig. S4).
assembled with Gβγ. Upon assembly with mGαi1-GDP, a quality
spectrum was no longer achievable on the 600 MHz spectrometer
and acquisition of the spectrum required analysis on the 800 MHz
spectrometer that is indicative of a loss of sensitivity due to
relaxation-dependent line broadening that occurred upon forma-
ing analysis of the NMR sample indicated an approximate
doubling in the molecular weight upon assembly of Gβγ with
mGαi1-GDP. Only one molecular weight species corresponding
protein, free Gα, or Gβγ in the sample (Fig. S3).
Phosducin Binding. Phosducin is a protein found in the retina that
binds Gβγ and can inhibit its activity. It has been cocrystallized
with Gβγ and local conformational alterations in Gβγ are ob-
served in the complex (25, 26). To compare the extent and pattern
of structural alterations in solution with another Gβγ binding
and presence (Red) of 300 μM SIGK: SIGKAFKILGYPDYD. B.1H projection of the1H-15N-TROSY-HSQC spectra from A. C. Comparison of1H-15N-TROSY-HSQC
spectrum of15N-Trp-Gβ1γ2ΔC (150 μM) in the absence (Black) and presence (Blue) of 300 μM SCAR: SCARFFGTPGCT. D. Comparison of the downfield region of
the1H-15N-TROSY-HSQC spectra from Gβγ alone (Black) Gβγ with SIGK (Red) and Gβγ with SCAR (Blue). E. Comparison of the upfield region of the1H-15N-
TROSY-HSQC spectra from Gβγ alone (Black) Gβγ with SIGK (Red) and Gβγ with SCAR (Blue).
Peptide ligand effects on Gβ1γ2ΔC1H-15N-TROSY-HSQC spectrum. A. Comparison of HSQC spectrum of15N-Trp-Gβ1γ2ΔC (150 μM) in the absence (Black)
Smrcka et al.PNAS
January 12, 2010
partner, a complex between unlabeled phosducin (Pdc) and15N-
was analyzed by two-dimensional
HSQC (Fig. 5A). The results are in striking contrast to those
observed in the complex with Gαi1. Phosducin binding results
in significant alterations in peak intensities and/or chemical shifts
for virtually all the Trp residues. GβW99 and GβW332 in the hot
spot directly contact the N terminus of phosducin and their
signals are strongly altered. The resonance for GβW339i is also
altered and, whereas GβW339 does not directly interact with
phosducin, it is positioned to report on the local conformational
changes observed in crystal structures of the complex (25, 26) that
open the cavity between blades six and seven (see Fig. 1C) upon
Pdc-binding. More surprising are changes observed for Trp resi-
dues distant from the binding site of phosducin on Gβγ. These
other Trp residues are structural elements of the blades of Gβ
and the changes in chemical shift report overall structural
changes in the propeller in solution that are not obvious from
the three-dimensional crystal structures.
Phosducin consists of two discreet domains that interact with
the Gβ propeller, an N-terminal region that interacts with the hot
spot and a C-terminal region that interacts with blades six and
seven (see Fig. 1C). To more specifically determine the role of
binding at the hot spot in the observed conformational alter-
ations, a protein consisting of amino acids 1–107 from phosducin
(PdcN) was purified and complexed with15N-Trp-Gβγ. The two-
dimensional1H-15N-TROSY-HSQC spectrum (Fig. 5B) shows
many, but not all, of the alterations observed with full-length
phosducin. Thus binding of the N-terminal domain of phosducin
is sufficient to alter the global conformation of Gβ but the
spectrum of free Gβγ (Black) and Gβγ-phosducin complex (Red). Black arrows
indicate shifts of resonance position in the complex, Blue arrows indicate
resonances that disappear in the complex, and Green arrows indicate
resonances that appear in the complex but cannot be assigned. B. Compar-
ison of1H-15N-TROSY-HSQC spectrum of free Gβγ (Black) and Gβγ -phosducin
N terminus 1–107 complex (Red). Black arrows indicate changes that are
similar to the Gβγ full-length phosducin complex and Blue arrows indicate
differences from the full-length complex.
Phosducin binding to Gβγ. A. Comparison of1H-15N-TROSY-HSQC
Model of thethree-dimensional structure ofthe Gαi1Gβ1γ2heterotrimer from
coordinates 1GP2. Key Gβγ subunit Trp residues are numbered in Black. The
Dark Blue helix is the switch II region of Gα and the yellow helix is the
N-terminal helix of Gαi. B. Overlay of the1H-15N-TROSY-HSQC spectra of
the 1.2∶1 complex of Gα-15N-Trp-Gβ1γ2ΔC (Red) to15N-Trp-Gβ1γ2ΔC alone
(Black). Data were collected for 12 h for15N-Trp-Gβ1γ2ΔC alone and 24 h
for the complex by using an 800 MHz spectrometer for both. C.1H projection
of the data in B.
Gα subunit effects on Gβ1γ2ΔC1H-15N-TROSY-HSQC spectrum. A.
www.pnas.org/cgi/doi/10.1073/pnas.0909503107 Smrcka et al.
C terminus can modify the nature of these changes to some
The two-dimensional1H-15N-TROSY-HSQC spectrum for Gβγ
reveals a striking range of intensities of NMR signals in the
spectrum. We propose that this represents a range of dynamic
properties of each of these amino acids on Gβ on different
NMR time scales. Evidence that the diversity of peak intensities
represent conformational exchange on different time scales are:
(i) Molecular dynamics simulations that show that GβW99 has
the capacity to be uniquely mobile on a nanosecond timescale.
(ii) GβW99i peak intensity is greatly diminished by SIGK binding
consistent with the crystal structure showing that GβW99i-side
chain mobility would be severely restricted upon peptide binding
(6). (iii) The low intensity signals of both the GβW99a and
GβW332i increase upon SIGK binding to a level similar to the
mean for the remaining amino acids. This can be best explained
by chemical exchange between different conformations on an
intermediate time scale (μsec-msec) in the unbound protein, with
SIGK binding selecting and stabilizing a single conformation.
This hypothesis is supported by x-ray diffraction data showing
direct interactions of SIGK with GβW332 that would alter the
dynamic properties of this amino acid.
Aquantitative approach to measuring individual Trp side chain
and backbone dynamics would require rigorous chemical relaxa-
tion experiments (27–29). Several factors limit this approach for
this study, including long times (12 h) for two-dimensional
1H-15N-TROSY-HSQC data acquisition, and uncertain protein
stability for collecting multiple relaxation spectra. Additionally,
the intensities of some of the signals are extremely low or not
observable, making it difficult to follow their evolution over
the course of the relaxation experiment. Qualitatively, however,
the data supports the conclusion that many of the Trp residues in
unbound Gβγ undergo chemical exchange between different con-
formations on different time scales. We only observe Trp residues
in these spectra but it is likely that other amino acids in the hot
spot are dynamic, as well. This range of dynamics supports the
hypothesis that the hot spot of Gβ subunits explores a range
of conformations in the ligand-free state that present a range
of structures that can be selected for by a particular binding part-
ner. This concept could be important for explaining the ability of
Gβγ to bind to multiple protein and small-molecule ligands.
Of particular interest is the contrast in Gβγ-conformational
states bound by Gα, SIGK, and phosducin binding to Gβγ. Of
these proteins, mGαi1-GDP has the highest affinity for Gβγ with
a Kd of 1–10 nM (24), whereas phosducin and SIGK have Kd´s of
50 nM (30) and 1 μM (31), respectively. The studies presented
here were performed with Gβγ subunits missing the geranyl–
geranyl lipid that is important for high-affinity interactions
between Gβγ and various protein binding partners including:
Gα and phosducin. However; at high concentrations of Gβγ, such
as in the conditions of our NMR experiments (150 μM), Gβγ
binds to all these partners (32–34). SIGK is a structural mimic
of Gα switch II, interacts with the same amino acids as Gα
switch II, and has clear effects on the mobility and chemical shifts
of Trp residues inside and outside the hot spot. Striking differ-
ences in the NMR spectrum are also observed either in Pdc or
PdcN complexes compared to that of free-Gβγ. Thus, it is re-
markable that very little difference is observed in mGαi1-GDP
subunit bound spectrum of Gβγ compared to free-Gβγ. Most
surprising is the level of apparent mobility that is maintained
for both of the Trp residues in the hot spot, particularly GβW99i.
This would not be possible if the solution conformation of the
heterotrimer was the same as that reported in the crystal struc-
tures with GβW99i involved in hydrogen bond and van der Waals
interactions with Gα amino acids (32–34). These data strongly
suggest that the inactive heterotrimer is in an open conformation
exposing the Gβγ hot spot and Gα switch II interface a large per-
centage of the time allowing for increased mobility of amino acids
at this interface.
act with effectors or other molecules through either Gβγ, Gα,or
both surfaces in the absence of nucleotide exchange. Nucleotide
exchange could increase the availability of these surfaces or alter
the relative conformations of these surfaces to lead to effector
activation. The opening of this surface also provides a potential
mechanism for nucleotide exchange-independent G-protein
activation such as observed for some Activators of G protein
subunit dissociation through interaction with these surfaces in the
A general view in the field is that Gβγ subunits do not undergo
significant conformational changes primarily based on observa-
tions comparing the free-Gβγ structure with Gβγ cocomplexes.
In the studies presented here, SIGK, SCAR, Gα subunits, and
phosducin binding yield distinct changes in chemical shift at sites
within and outside the direct binding surface. For SIGK and
SCAR, the major changes in chemical shift and dynamics are
observed for GβW99i, GβW332i, and GβW99a; all directly in
the peptide binding site. These alterations may reflect local
alterations in conformation, but the additional, albeit subtle,
changes in chemical shift that occur outside the peptide binding
site indicate more global structural changes. For Gα subunits, any
changes outside the binding site were difficult to observe. In con-
trast, are the dramatic changes that occur throughout the protein
upon phosducin binding despite the 10-fold lower affinity of
phosducin than Gα for Gβγ. The data supports the idea that
the conformational change believed to bury the prenyl group
in a groove between β-propeller blades six and seven may be
more global than originally proposed from the crystal structures
(25, 26). The fact that similar structural changes in Gβγ were ob-
served with full length Pdc and its N-terminal domain is consis-
tent with biochemical studies showing that they both caused
dissociation of Gβγ from membranes despite the fact that only
the C-terminal domain sterically interferes with membrane bind-
ing (38). This finding led to the prediction that the N-terminal
domain could allosterically induce the conformation with the
prenyl group buried.
Together these data strongly support the hypothesis that Gβγ
subunits accommodate different binding partners through struc-
turally distinct conformations. To date, there are crystal struc-
tures of cocomplexes between Gβγ and phosducin (25, 26),
Gβγ and Gα subunits (32, 35), Gβγ and G-protein-coupled
receptor kinase 2 (39), and Gβγ and SIGK (6) or parathyroid
hormone receptor (40) peptides. There are close to 50 different
binding partners for Gβγ and our data suggest that there are
changes in conformation of Gβγ that occur in solution that are
not readily observed or interpretable from crystallographic struc-
tures. Thus it seems very likely that there are a wide range of
structural changes that can occur on Gβγ that depend on the
binding partner, and these conformational changes could be
important not only for molecular recognition but for modulation
of Gβγ signaling functions.
Materials and Methods
Expression and Purification of15N-Trp Gβγ. Gβ1(or Gβ1-Trp mutants) and
6His-tagged γ2truncated after F67 to remove the site of isoprenylation
and the distal amino acids in the Cysteine, aliphatic, aliphatic, X amino acids
(CAAX) box were expressed in High 5 insect cells grown in custom media
(Bioexpress 2000; Cambridge Isotope Laboratories) supplemented with un-
labeled amino acids, except for Trp isotopically labeled with15N at the
α-amine and ε-indole positions (Cambridge Isotope Laboratories). Protein
was purified, as has been described previously, and as in SI Materials and
Methods. The yield is 1–3 mg of pure15N labeled15N-6His-β1γ2ΔC. All protein
concentrations were estimated with a Bradford protein assay.
Smrcka et al. PNAS
January 12, 2010
NMR Analysis of Download full-text
NaHPO4, pH 7.0, 100 mM NaCl, and 10% D2O was introduced into a 3 mm
NMR tube. The analysis used 600 and 800 MHz Bruker Digital Avance
NMR instruments fitted with triple-resonance z-axis gradient cryoprobes.
The 600 MHz instrument was used for most of the experiments, although
the 800 MHz instrument was used where indicated. Measurements were
conducted by using a1H-15N-TROSY-HSQC pulse sequence and data were col-
lected for 8–12 h (unless otherwise indicated) at 30 °C. Data were processed
by using Topspin 2.0. To normalize for alterations in overall spectral intensi-
ties between samples and experiments, spectra were adjusted to the rela-
tively fixed signal of GβW63a. Thus discussions of spectral intensities refer
to relative intensities within a given spectrum. To assess sample integrity,
1-D proton NMR spectra were taken before and after each 2-D experiment.
No significant changes were observed in the protein spectra indicating a lack
of aggregation or protein loss over the course of the measurements.
15N Gβγ. 160–200 μl of sample in NMR buffer: 20 mM
Formation of Gβγ Complexes. For formation of Gβ1γ2-myristoylated-Gαi1GDP
complexes and 32.6 nmoles of Gβ1γ2ΔC was mixed with 40 nmoles of Gα sub-
units (0.9 mol∕mol GTPγS binding) in 2.5 ml buffer A: NMR buffer containing
25 μM GDP and 200 μM Dithiothreitol (DTT). The mixture was gel filtered
into 3.5 ml of buffer A, followed by centrifugal concentration to 200 μl. A
similar procedure was used to form phosducin complexes except GDP was
not included. For formation of peptide complexes, small aliquots of
10 mM SIGK or SCAR dissolved in NMR buffer were directly added to the con-
centrated Gβγ NMR sample. No DTTwas added to these complexes but a small
amount of 2-mercaptoethanol was present that was residual from the Gβγ
Simulations. All simulations were performed with the AMBER9 software
package by using the parm99 parameter set. The initial structure of the
G-protein was based on the crystallographic structure of Gβγ complex with
the SIGK peptide ligand (Protein Data Bank (PDB) code 1XHM). The SIGK
ligand was removed from the original complex. To create a set of initial
conformations, 1 ns simulations of the receptor or bound complex were
performed by using the Generalized Born implicit solvent model and
Langevin dynamics at 300 K. Configurations from these simulations were
saved and assigned to nine clusters on the basis of root-mean-square
deviation. For each cluster, the configuration closest to the cluster center
was used as the starting structure for an 8-ns molecular dynamics simula-
tion performed with explicit solvent. Simulations by using the crystal struc-
ture as an initial configuration were also performed, for a total of ten
independent 8-ns simulations. See Supporting Information Materials and
Methods for details.
ACKNOWLEDGEMENTS. This work was supported by National Institutes of
Health Grant GM081772 (to A.V.S.); MCYT-FEDER (Bio2008-00799); and the
Generalitat de Catalunya (X.R.B. and Grup Consolidat) (E.G.).
1. Smrcka AV (2008) G protein βγ subunits: central mediators of G protein-coupled
receptor signaling. Cell Mol Life Sci, 65:2191–2214.
2. Dupre DJ, Robitaille M, Rebois RV, Hebert TE (2009) The role of Gβγ subunits in
the organization, assembly, and function of GPCR signaling complexes. Annu Rev
Pharmacol Toxicol, 49:31–56.
3. Oldham WM, Hamm E (2006) Structural basis of function in heterotrimeric G proteins.
Q Rev Biophys, 39:117–166.
4. Ford CE, et al. (1998) Molecular basis for interactions of G protein βγ subunits with
effectors. Science, 280:1271–1274.
5. Scott JK, et al. (2001) Evidence that a protein-protein interaction “hot spot” on
heterotrimeric G protein βγ subunits is used for recognition of a subclass of effectors.
EMBO J, 20:767–776.
6. Davis T, Bonacci TM, Sprang SR, Smrcka AV (2005) Structural Definition of a Preferred
Protein InteractionSite intheG protein
7. Atwell S, Ultsch M, Devos AM, Wells JA (1997) Structural plasticity in a remodeled
protein-protein interface. Science, 278:1125–1128.
8. Delano WL, Ultsch MH, de Vos AM, Wells JA (2000) Convergent solutions to binding at
a protein-protein interface. Science, 287:1279–1283.
9. Wells JA, Devos AM (1996) Hematopoietic receptor complexes. Annu Rev Biochem,
10. Gordo Sl, Giralt E (2009) Knitting and untying the protein network: modulation of
protein ensembles as a therapeutic strategy. Protein Sci, 18:481–493.
11. Bookout AL, et al. (2003) Targeting Gβγ Signaling to inhibit prostate tumor formation
and growth. J Biol Chem, 278:37569–37573.
12. Koch WJ, et al. (1995) Cardiac function in mice overexpressing the β-adrenergic
receptor kinase or a βARK inhibitor. Science, 268:1350–1353.
13. Xie W, et al. (1999) Genetic alteration of phospholipase Cβ3 expression modulates be-
havioral and cellular responses to μ opioids. Proc Natl Acad Sci USA, 96:10385–10390.
14. Li Z, et al. (2000) Roles of PLC-2 and -3 and PI3K in Chemoattractant-Mediated Signal
Transduction. Science, 287:1046–1049.
15. Hirsch E, et al. (2000) Central role for G protein-coupled phosphoinositide 3-Kinase γ in
inflammation. Science, 287:1049–1053.
16. Bonacci TM, et al. (2006) Differential targeting of Gβγ-subunit signaling with small
molecules. Science, 312:443–446.
17. Mathews JL, Smrcka AV, Bidlack JM (2008) A novel Gβγ-subunit inhibitor selectively
modulates μ-opioid-dependent antinociception and attenuates acute morphine-
induced antinociceptive yolerance and dependence. J Neurosci, 28:12183–12189.
18. Lehmann DM, Seneviratne AMPB, Smrcka AV (2008) Small molecule disruption of
G protein βγ subunit signaling inhibits neutrophil chemotaxis and inflammation.
Mol Pharmacol, 73:410–418.
19. QinJ, et al.(2009) Upregulation of PIP3-dependentRac exchanger1(P-Rex1) promotes
prostate cancer metastasis. Oncogene, 28:1853–1863.
20. Pellecchia M, et al. (2008) Perspectives on NMR in drug discovery: A technique comes
of age. Nat Rev Drug Discovery, 7:738–745.
21. Tugarinov V, Hwang PM, Kay LE (2004) Nuclear magnetic resonance spectroscopy of
high-molecular-weight proteins. Annu Rev Biochem, 73:107–146.
22. Tarrago T, et al. (2009) A cost-effective labeling strategy for the NMR study of large
proteins: Selective15N-labeling of the tryptophan side chains of prolyl oligopeptidase.
23. Sondek J, et al. (1996) Crystal structure of a G-protein βγ dimer at 2.1 Å resolution.
24. Sarvazyan NA, Remmers AE, Neubig RR (1998) Determinants of Giα and βγ binding:
Measuring high affinity interactions in a lipid environment using flow cytometry.
J Biol Chem, 273:7934–7940.
25. Gaudet R, Bohm A, Sigler PB (1996) Crystal structure at 2.4 angstroms resolution of the
complex of transducin βγ and its regulator, phosducin. Cell, 87:577–588.
26. Loew A, Ho YK, Blundell T, Bax B (1998) Phosducin induces a structural change in
transducin βγ. Structure, 6:1007–1019.
27. Palmer AG, III (2001) NMR probes of molecular dynamics: Overview and comparison
with other techniques. Annu Rev Biophys Biomol Struct, 30:129–155.
28. Mittermaier A, Kay LE (2006) New tools providenew insights in NMR studies of protein
dynamics. Science, 312:224–228.
29. Boehr DD, Dyson HJ, Wright PE (2006) An NMR perspective on enzyme dynamics.
Chem Rev, 106:3055–3079.
30. Savage JR, et al. (2000) Functional roles of the two domains of phosducin and
phosducin-like protein. J Bio Chem, 275:30399–30407.
31. Ghosh M, Peterson YK, Lanier SM, Smrcka AV (2003) Receptor and nucleotide
exchange independent mechanisms for promoting G protein subunit dissociation.
J Biolchem, 273:34747–34750.
32. Wall MA, et al. (1995) The structure of the G protein heterotrimer Giα1β1γ2. Cell,
33. Sprang SR (1997) G protein mechanisms: Insights from structural analysis. Annu Rev
34. Lambright DG, et al. (1996) The 2.0 Å crystal structure of a heterotrimeric G protein.
35. Blumer JB, Cismowski MJ, Sato M, Lanier SM (2005) AGS proteins: receptor-
independent activators of G-protein signaling. Trends Pharmacol Sci, 26:470–476.
36. Blumer JB, Smrcka AV, Lanier SM (2007) Mechanistic pathways and biological roles for
receptor-independent activators of G-protein signaling. Pharmacol Ther, 113:488–506.
37. Yuan C, Sato M, Lanier SM, Smrcka AV (2007) Signaling by a non-dissociated complex
of G protein βγ and α subunits stimulated by a receptor-independent activator of
G protein signaling, AGS8. J Bio Chem , 282:19938–19947.
38. LukovGL, et al. (2004) Roleofthe isoprenylpocket ofthe G proteinβγ subunitcomplex
in the binding of phosducin and phosducin-like protein. Biochemistry, 43:5651–5660.
39. Lodowski DT, et al. (2003) Keeping G proteins at bay: A complex between
G protein-coupled receptor kinase 2 and Gβγ. Science, 300:1256–1262.
40. Johnston CA, Kimple AJ, Giguere PM, Siderovski DP (2008) Structure of the
parathyroid hormone receptor C terminus bound to the G-protein dimer Gβ1γ2.
www.pnas.org/cgi/doi/10.1073/pnas.0909503107Smrcka et al.