Biochem. J. (2013) 456, 241–251 (Printed in Great Britain)
241
doi:10.1042/BJ20130148
Dynamic conformational switching in the chemokine ligand is essential for
G-protein-coupled receptor activation
Prem Raj B. JOSEPH*, Kirti V. SAWANT*, Angela ISLEY†, Mesias PEDROZA*, Roberto P. GAROFALO‡§,
Ricardo M. RICHARDSON† and Krishna RAJARATHNAM*‡1
*Department of Biochemistry and Molecular Biology, The University of Texas Medical Branch, 301 University Boulevard, Galveston, TX 77555-1055, U.S.A., †Julius L. Chambers
Biomedical/Biotechnology Research Institute and the Department of Biology, North Carolina Central University, Durham, NC 27707, U.S.A., ‡Department of Microbiology and
Immunology, The University of Texas Medical Branch, 301 University Boulevard, Galveston, TX 77555-1055, U.S.A., §Department of Pediatrics, The University of Texas Medical Branch,
301 University Boulevard, Galveston, TX 77555-1055, U.S.A., and Sealy Center for Structural Biology and Molecular Biophysics, The University of Texas Medical Branch, 301
University Boulevard, Galveston, TX 77555-1055, U.S.A.
Chemokines mediate diverse functions from organogenesis
to mobilizing leucocytes, and are unusual agonists for
class-A GPCRs (G-protein-coupled receptors) because of
their large size and multi-domain structure. The current
model for receptor activation, which involves interactions
between chemokine N-loop and receptor N-terminal residues
(Site-I) and between chemokine N-terminal and receptor
extracellular loop/transmembrane residues (Site-II), fails to
describe differences in ligand/receptor selectivity and the
activation of multiple signalling pathways. In the present study,
we show in neutrophil-activating chemokine CXCL8 that the
highly conserved GP (glycine-proline) motif located distal to
both N-terminal and N-loop residues couples Site-I and Site-II
interactions. GP mutants showed large differences from native-
like to complete loss of function that could not be correlated with
the specific mutation, receptor affinity or subtype, or a specific
signalling pathway. NMR studies indicated that the GP motif
does not influence Site-I interactions, but molecular dynamics
simulations suggested that this motif dictates substates of the
CXCL8 conformational ensemble. We conclude that the GP
motif enables diverse receptor functions by controlling cross-talk
between Site-I and Site-II, and further propose that the repertoire
of chemokine functions is best described by a conformational
ensemble model in which a network of long-range coupled
indirect interactions mediate receptor activity.
INTRODUCTION
by directly binding receptors or indirectly influencing how Nterminal/N-loop residues bind and activate the receptors.
Chemokine CXCL8 [also known as IL-8 (interleukin-8)]
recruits neutrophils by activating the CXCR1 and CXCR2
receptors. Structure–function studies have shown that the Nterminal ELR and N-loop residues play important roles in defining
affinity and receptor activity [5,7,8]. Chemokine PF-4/CXCL4
has the GP motif and is not active for CXCR1/CXCR2 receptors,
but becomes active on introducing the N-terminal ELR motif [9].
The chemokine IP-10/CXCL10 does not have the GP motif and
is not active against CXCR1/CXCR2 receptors [10]. However, a
CXCL10/CXCL8 chimaera containing the 30s-loop, N-terminal
ELR and N-loop residues of CXCL8 is active, but a chimaera containing the 30s-loop and ELR residues is not active [7]. Therefore
the mere presence of 30s-loop GP motif or even the additional
presence of ELR residues are not sufficient for CXCL8 activity,
indicating that receptor activation is finely tuned and requires optimal cross-talk among N-terminal, N-loop and 30s-loop residues.
To address the role of the GP motif in CXCL8 function, we
have now characterized the structural and functional properties
of a panel of GP mutants (G31A, G31V, P32G and P32A).
The mutants show large differences in activity that could not be
correlated with the nature of the mutation, receptor type and/or a
specific signalling pathway. NMR studies show that the mutations
do not perturb the global fold, any local structural change could
Chemokines, a large family of ∼8–10 kDa proteins, play
fundamental roles from developmental biology to trafficking
immune cells and combating infection by activating receptors that
belong to the GPCR (G-protein-coupled receptor) class [1]. They
show a wide repertoire of receptor functions, as a given chemokine
can bind single or multiple receptor(s), and binding to any one
receptor can elicit multiple signalling pathways and different
activities [1,2]. According to the prevalent model, chemokine
receptor activation involves interactions between chemokine Nloop and receptor N-terminal residues (Site-I), and between
chemokine N-terminal and receptor extracellular/transmembrane
residues (Site-II) [2–6]. This model, although appealing in
its simplicity, fails to describe the large variation in receptor
selectivity, affinity and activity for a given chemokine and among
chemokines.
Sequence analysis of neutrophil-activating chemokines in
humans and across other species show that the GP (glycineproline) residues which form a β-turn in the 30s-loop are
highly conserved (Figure 1A). Structures reveal that the 30sloop is located distal to both the N-terminal and N-loop regions
(Figure 1B). High conservation implies that it must play some
crucial role for structure and/or function. Therefore the GP motif
could be essential for the structural fold and/or for receptor activity
Key words: chemokine, conformational ensemble, G-proteincoupled receptor, long-range coupling, signalling.
Abbreviations used: DMEM, Dulbecco’s modified Eagle’s medium; ERK, extracellular-signal-regulated kinase; GAG, glycosaminoglycan; GP, glycineproline; GPCR, G-protein-coupled receptor; HSQC, heteronuclear single-quantum coherence; MD, molecular dynamics; PGP, proline-glycine-proline; WT,
wild-type.
1
To whom correspondence should be addressed (email [email protected]).
c The Authors Journal compilation c 2013 Biochemical Society
242
Figure 1
P. R. B. Joseph and others
Structural characteristics of CXCL8
(A) Sequence alignment of human CXCL8 (hCXCL8) and related neutrophil-activating chemokines, and CXCL8 from other species. N-terminal ELR residues (in green) are absolutely conserved and
the 30s-loop ‘GP’ motif (in red) is highly conserved, except in CXCL7. The N-loop residues are indicated by a blue bar. (B) Tertiary structural components of the CXCL8 monomer. The 30s-loop,
N-loop (Site-I), ELR residues (Site-II) and disulfide bonds (black) are highlighted. The GP motif in the β-turn of the 30s-loop are shown as sticks, and the hydrogen bonds between Gly31 and Cys34
are shown as broken lines (see insert).
not be correlated with functional changes, and that they bind
the receptor N-domain like WT (wild-type) indicating that the
GP motif does not participate or influence Site-I interactions.
MD (molecular dynamics) simulations show that perturbing the
GP motif leads to conformational rearrangement and altered
dynamics resulting in access to non-native conformations, thus
providing a rationale for the stringent requirement of the GP
motif for native function. We propose that the GP motif functions
as a conformational switch and dictates the distribution and
selection of the functionally competent conformational substates
that activate various signalling pathways. We further propose
that the ‘conformational selection’ model can best describe the
spectrum of chemokine receptor functions [11,12].
MATERIALS AND METHODS
Cloning, expression and purification of CXCL8 GP mutants
The 30s-loop G31A, G31V, P32G and P32A mutants were
generated using the QuikChange® Site-Directed Mutagenesis kit
(Stratagene). Two versions of the mutant proteins, the full length
c The Authors Journal compilation c 2013 Biochemical Society
(residues 1–72) and a truncated monomeric form (residues 1–
66) were generated. The CXCL8-(1–66) deletion mutant is a
monomer, and its receptor affinity and activity are similar to those
of the trapped full-length monomer [8,13,14]. Both versions of
the mutants were expressed and purified as described previously
[13,15]. 15 N-labelled monomer mutants for NMR studies were
produced essentially as described previously, except that cells
were grown in minimal medium containing 15 NH4 Cl as the sole
nitrogen source [15]. The purity and molecular mass of the
mutants were confirmed using analytical HPLC and MALDI–MS.
Radioligand receptor-binding assays
Radioligand-binding assays were carried out using CXCR1and CXCR2-transfected RBL-2H3 cells as described previously
[16]. Briefly, RBL-2H3 cells sub-cultured overnight in 24-well
plates (0.5×106 cells) in growth medium were rinsed with
DMEM (Dulbecco’s modified Eagle’s medium) supplemented
with 20 mM Hepes (pH 7.4) and 10 mg/ml BSA and incubated
on ice for 2–4 h in the same medium (250 μl) containing 125 Ilabelled CXCL8 (0.1–1 nM) and GP mutants at concentrations
Long-range coupling modulates chemokine receptor function
from 0 to 1 μM. The reaction was stopped with 1 ml of icecold PBS containing 10 mg/ml BSA, and cells were washed three
times, solubilized with 200 μl of RIPA buffer and dried under
vacuum. Bound radioactivity was then measured in a γ -counter
and K i values was calculated as described previously [16]. Nonspecific-bound radioactivity was determined in the presence of
500 nM unlabelled WT CXCL8. Binding affinities are from at
least two independent experiments, and each experiment was
performed at least two times.
Cell culture
CXCR1 and CXCR2 stably transfected RBL-2H3 cells were
maintained as monolayer cultures as described previously [17].
Briefly, RBL-2H3 cells (1×107 cells) were transfected by
electroporation with 20 μg of pcDNA3 containing the receptor
cDNAs. Geneticin-resistant cells were selected by FACS analysis
and cloned into single cells. Cells were then maintained in DMEM
supplemented with 15 % FBS, 2 mM glutamine, 100 units/ml
penicillin and 100 mg/ml streptomycin (Gibco, Life Science
Technologies).
Intracellular Ca2 + mobilization
CXCR1- and CXCR2-transfected RBL-2H3 cells (5×106 ) were
washed with Hepes-buffered saline and loaded with 1 μM IndoI/AM (acetoxymethyl ester) in the presence of 1 μM pluronic
acid for 30 min at room temperature (23 ◦ C). The cells were
then washed and resuspended in 1.5 ml of Siriganian buffer, and
the intracellular Ca2 + mobilization activity of the GP mutants
was measured as described previously [18]. Curve fitting and
EC50 values were determined by non-linear regression analysis
of the dose–response curves generated using Prism 5 (GraphPad
Software).
ERK (extracellular-signal-regulated kinase) phosphorylation
CXCR1- and CXCR2-transfected RBL-2H3 cells (5×107
cells/ml) were stimulated with GP mutants (10 nM) for 1
and 5 min, and lysed with an equal volume of RIPA buffer
supplemented with proteinase and phosphatase inhibitors. The
sonicated cell lysates were electrophoresed at 150 V for 1 h using
SDS/PAGE (12 % gel), transferred on to a PVDF membrane at
100 V for 1 h, and blocked in 5 % non-fat dried skimmed milk
in TBST [10 mM Tris (pH 8.0) and 150 mM NaCl containing
0.05 % Tween 20] for 1 h at room temperature. Membranes were
immunoblotted with primary antibody overnight and then treated
with secondary antibody for 1 h. The protein bands were detected
using substrates from the Supersignal West Femto kit (Thermo
Scientific) and imaged using the ChemiDoc XRS system (BioRad Laboratories). The rabbit anti-phospho-ERK (Thr202 /Tyr204 )
and anti-(p44/p42 total ERK) antibodies were from Cell Signaling
Technology, and the secondary goat anti-rabbit IgG antibody was
from Santa Cruz Biotechnology.
Peritoneal neutrophil recruitment
Female BALB/c mice (8–10 weeks old) were purchased from
Harlan and housed under pathogen-free conditions in the animal
research facility, in accordance with the National Institutes of
Health and University guidelines for animal care. Under light
anaesthesia, mice were inoculated intraperitoneally with 0.25
or 10 μg of GP mutants (full-length 1–72 version) in D-PBS
(Dulbecco’s PBS). At 6 h post-inoculation, mice were killed
and peritoneal neutrophil levels were determined as described
243
previously [19]. The peritoneal cavity was flushed with 1.5 ml of
ice-cold HBSS (Hanks balanced salt solution) buffer to collect
the migrated neutrophils. Cytospin slides were prepared by
centrifuging 100 μl of peritoneal fluid for 5 min at 72 g in a
cytocentrifuge (Shandon, Thermo Electron). Slides were then
fixed and stained with Wright–Giemsa stain, and 200 cells were
counted to determine the percentage of neutrophils. For total
leucocyte determination, samples were diluted with Turk blood
diluting fluid (Ricca Chemicals), and the total cells were counted
using a haemocytometer.
NMR spectroscopy
NMR spectra were acquired at 30 ◦ C on a Varian Unity Plus 600
spectrometer equipped with field gradient accessories. The protein
concentrations were ∼0.1 mM in 50 mM sodium phosphate buffer
(pH 6.0) containing 1 mM sodium azide, 1 mM DSS (sodium
2,2-dimethyl-2-silapentane sulfonate) and 10 % (v/v) 2 H2 O. All
spectra were processed and analysed as described previously [20].
Binding of unlabelled CXCR1 N-terminal domain peptide to
15
N-labelled P32G and G31V mutants was characterized using
2D (1 H,15 N)-HSQC (heteronuclear single-quantum coherence)
NMR chemical-shift changes. The final peptide/protein molar
ratio for the P32G and G31V mutants was 5.2 and 5.8 respectively.
Apparent binding constants were determined by fitting the
binding-induced chemical shift changes as a function of
the peptide/protein molar ratios as described previously [20].
MD simulations
The CXCL8-(1–66) truncated monomer molecular model was
generated using the CXCL8 monomer structure (PDB code
1IKM) [20]. The GP variants were generated by mutating Gly31 to
alanine or valine, and Pro32 to glycine or alanine. The structures
were subjected to multiple cycles of constrained and free-energy
minimizations using the AMBER 12 suite to remove steric
hindrance introduced by the mutations [21,22]. The energyminimized structures were subjected to an equilibration protocol
in explicit solvent as described previously [22]. MD runs (200 ns)
were carried out using the PMEMD (particle mesh Ewald MD)
module on the Lonestar Dell Linux Cluster at the TACC (Texas
Advanced Computing Center, University of Texas, Austin, TX,
U.S.A.). Analysis of the trajectory was carried out using the
Ambertools12 and the VMD molecular visualization software
[21,23].
RESULTS
Cell-based functional assays
We characterized the activity of G31A, G31V, P32G and P32A
mutants by measuring binding affinity, Ca2 + release and ERK
phosphorylation using RBL-2H3 cells stably transfected with
CXCR1 or CXCR2 receptors (Table 1 and Figure 2). All of the
mutants bound CXCR2 with high affinity (Table 1). Although
G31V and P32A bound CXCR1 with high affinity, binding of
P32G and G31A could not be detected, suggesting that their
binding is significantly impaired. The higher affinity of the G31V
mutant compared with the WT for both receptors is particularly
interesting and unprecedented.
We measured EC50 values for Ca2 + release activity of all of
the mutants for both CXCR1 and CXCR2 receptors (Table 1 and
Figure 2A). The individual mutants showed a range of activity
from WT-like to being inactive. Whereas P32A and G31V were
4–12-fold less active, P32G and G31A were essentially inactive
for both of the receptors (Table 1). We characterized ERK1/2
c The Authors Journal compilation c 2013 Biochemical Society
244
P. R. B. Joseph and others
Figure 2
Activity of the GP mutants
2+
(A) Ca release activity of the GP mutants. Data are the average of three measurements and representative of three independent experiments. (B) ERK phosphorylation activity of the GP mutants.
Results are shown as a representative Western blot for pERK and ERK phosphorylation at 1 and 5 min time points. The blots are one of three independent experiments. The control (Cont) is stably
transfected CXCR1 or CXCR2 cells that have not been treated with chemokine.
Table 1 Binding affinity (K d ) and Ca2 + release activity (EC50 ) of the GP
mutants
Values are means +
− S.E.M. n.d., not detected.
CXCR1
CXCR2
Form
K d (nM)
EC50 (nM)
K d (nM)
EC50 (nM)
WT
G31V
G31A
P32G
P32A
4.1 +
− 1.2
0.3 +
− 0.1
n.d.
n.d.
4.9 +
− 1.2
2.5 +
−1
22.9 +
−3
n.d.
>1000
10.7 +
− 1.6
5.4 +
− 1.4
1.0 +
− 0.7
11.3 +
− 2.1
11.9 +
− 1.8
7.0 +
− 1.3
2.2 +
− 1.0
26.1 +
− 10
n.d.
>1000
21.7 +
− 10
phosphorylation activity of the mutants at 10 nM at 1 and 5 min
time points (Figure 2B). In general, the mutants were less active,
but there were differences among the mutants as a function of
receptor subtype and differences as a function of time point, which
suggest differences in phosphorylation kinetics. For CXCR2, all
of the mutants were essentially inactive at the 1 min time point
and only the WT was active. However, at the 5 min time point,
P32A and G31V were active. For CXCR1, all mutants showed
impaired activity, except P32A, which was as active as the WT at
the 5 min time point.
recruitment activity of the GP mutants at 0.25 and 10 μg doses
in a mouse peritoneum model (Figure 3). Interestingly, the G31V
mutant was more active than the WT at the 10 μg dose (P < 0.01)
and is as active as the WT at the 0.25 μg dose. All other mutants
were significantly less active than the WT (P < 0.01); only G31V
and P32A were active at the lower dose, and P32A showed
slightly better recruitment than the G31A and P32G mutants at the
higher dose. In vivo recruitment is dependent on CXCL8 binding
to CXCR2 on neutrophils and GAGs (glycosaminoglycans) on
endothelial cells. WT CXCL8 exists reversibly as both monomers
and dimers, and we have shown that the recruitment at the
0.25 μg dose is predominantly from a monomer, and from both
the monomer and dimer forms at the 10 μg dose [19]. As the GP
motif plays no role in GAG binding, reduced recruitment must be
due to reduced receptor function. The high activity of G31V is
particularly intriguing, as it also binds with higher affinity than the
WT to its receptors. As the mutants were bacterially expressed,
it could be argued that the higher activity of the G31V mutant is
due to spurious contaminants such as LPS (lipopolysaccharide).
We rule this out as the P32G mutant is essentially inactive,
and the R6K and CC-CXCL8 mutants purified using the same
protocol were also inactive [19,24]. It is possible that the higher
activity is due to differences between the human and mouse
CXCR2 receptor and/or that the conformational ensemble of
G31V enables receptor activation (see Discussion). Independent
of the actual mechanism of binding, these data convincingly show
that the glycine residue of the GP motif cannot be involved in
direct binding.
Neutrophil recruitment in a mouse model
Neutrophil recruitment in animal models best captures the
importance of all of the signalling pathways, and is the most
stringent to understand and characterize the consequence of
the different mutations for function. We measured neutrophil
c The Authors Journal compilation c 2013 Biochemical Society
Structural characterization of GP mutants
Proline and glycine residues are frequently observed in turn
regions that link secondary structural elements in proteins [25,26].
Long-range coupling modulates chemokine receptor function
245
Figure S1). The observation that G31V is more active than P32G
indicates that structural changes as a consequence of the mutation
do not negatively influence function. A previous NMR study of
the G31P mutant has also shown that its structure is like the WT
[31].
Structural characterization of CXCL8 receptor Site-I interaction
Figure 3
Neutrophil recruitment in a mouse peritoneum model
Data are means +
− S.E.M., and represent four to five mice/group. At the 10 μg dose, the higher
recruitment of G31V and lower recruitment of other mutants were significant compared with the
WT (∗∗∗ P < 0.01). At the 0.25 μg dose, all mutants, except G31V, were less active compared
with the WT (∗∗∗ P < 0.01). One-way ANOVA with Tukey’s post-hoc test was used to determine
statistical significance.
Proline is unique due to its cyclic pyrrolidine ring side chain that
restricts its φ/ψ conformational space in the Ramachandran plot,
and also because it cannot form a backbone hydrogen bond as
it lacks the amide proton. Glycine is also unique as it is not
conformationally constrained due to lack of a side chain and can
occupy regions of the φ/ψ space that are inaccessible to other
residues. In the WT CXCL8, Gly31 , Pro32 , His33 and Cys34 form
a tight type I β-turn, and the backbone of Gly31 and Cys34 are
hydrogen-bonded to each other (Figure 1B) [27–29].
The structural consequence of mutations in the GP motif was
characterized for the P32G and G31V mutants by solution NMR
spectroscopy. We chose these mutants for the reason that G31V
was the most active and P32G the least active. We produced
the mutants on the CXCL8-(1–66) monomer background, and
so will refer to the CXCL8-(1–66) monomer as the WT when
discussing NMR data. Chemical shifts are exquisitely sensitive to
secondary and tertiary structures, and therefore identical chemical shifts between the mutant and WT indicate identical
structures [30]. Chemical shifts indicate that both mutants adopt
a CXCL8-like fold, but also showed local structural changes
of the 30s-loop (Figure 4A and Supplementary Figure S1A at
http://www.biochemj.org/bj/456/bj4560241add.htm). Plots of the
backbone amide proton chemical shifts of the P32G mutant
compared with the WT showed that most residues fell on the
diagonal, indicating essentially a similar global folded structure
(Figures 4B and 4C). Cys34 has a characteristic up-field chemical
shift in the WT, and the chemical shifts of Gly31 and Cys34 in
the mutant are similar to those in the WT, indicating that the βturn structure is conserved. Chemical shifts of the G31V mutant
showed larger changes in the 30s-loop residues, indicating local
structural perturbation around the site of mutation (Supplementary
The structural basis of Site-I interactions can be characterized
outside the context of the intact receptor by studying binding to
receptor N-domain peptides using NMR spectroscopy [20,32].
We have previously shown that binding to the CXCR1 N-domain
induces significant chemical-shift perturbation of CXCL8 N-loop,
β 3 -strand and C-terminal helix residues, and that N-loop residues
are involved in direct binding and that the β 3 -strand and Cterminal helix residues indirectly influence the binding of N-loop
residues via coupling interactions [20].
Chemical-shift perturbation profiles of the P32G mutant on
binding to the CXCR1 N-domain peptide were essentially
identical with that observed for the WT (Figure 5). A number
of residues broaden out during the course of binding and
reappear at higher receptor concentrations, similar to what is
observed for the WT, indicating that the binding is in the
fast to intermediate exchange regime in the NMR time scale
(Figure 5A). Chemical shifts of the GP motif are not perturbed
either in the WT or the mutant, indicating that these residues
are not involved in direct binding to the receptor N-domain
(Figures 5C and 5D). Not surprisingly, we also observed that
the chemical-shift perturbation profile of the G31V mutant is
similar to that of WT (Supplementary Figures S2A and S2B at
http://www.biochemj.org/bj/456/bj4560241add.htm).
The binding affinity of the P32G mutant for the receptor Ndomain (K d = 11 +
− 1 μM) is identical with the WT (Figure 5B),
indicating that Pro32 does not participate or influence Site-I
interactions. The binding affinity of the G31V mutant for the
receptor N-domain (K d = 50 +
− 10 μM) is approximately ∼4-fold
weaker compared with the WT, but it actually binds the intact
receptor with higher affinity, indicating that weaker binding to
Site-I must be compensated by positive coupling interactions to
Site-II, and most importantly that Gly31 is also not critical for highaffinity receptor binding (Supplementary Figure S2C). Therefore
large variation in the activity of the mutants cannot be attributed
to Site-I interactions, and indicate that the GP motif influences
the receptor-activation step of the N-terminal residues at Site-II.
Conformational dynamics of GP mutants
NMR and X-ray structures can be viewed as an average
structure or a snapshot of a conformational ensemble in dynamic
equilibrium, and so cannot provide the detailed spatial and
temporal description of the relative motions between structural
modules. Computational MD simulation studies can provide this
critical knowledge. In order to characterize how the GP motif
influences structure and mediates cross-talk between N-terminal
and N-loop residues, we carried out 200 ns MD simulations on
the CXCL8-(1–66) monomer (WT), and each of the four GP
mutants (P32G, P32A, G31V and G31A) in the same monomer
background, using the AMBER 12 suite [21].
P32G mutant
MD simulations show two distinct interchanging conformations
GGC1 and GGC2 , and some additional transient conformers for
the 30s-loop (Figures 6A–6C and Supplementary Figure S3
c The Authors Journal compilation c 2013 Biochemical Society
246
Figure 4
P. R. B. Joseph and others
NMR structural characteristics of the P32G mutant
(A) (1 H,15 N)-HSQC spectrum of P32G (red) superimposed on the WT (black). Characteristic up-field- and down-field-shifted peaks as well as residues in the vicinity of the mutation are labelled.
Gln8 and His33 are broadened out in the P32G spectrum. (B and C) Diagonal plots of HN (B) and N (C) chemical shifts of P32G (y -axis) compared with WT (x -axis). The data show that the chemical
shifts are essentially the same, indicating that the structure of the mutant is essentially identical with the WT.
at http://www.biochemj.org/bj/456/bj4560241add.htm). Analysis
showed that the conformer GGC1 , observed over the course
of the simulation, retained type I β-turn architecture observed
in the WT (Figures 6A and Supplementary Figure S4 at
http://www.biochemj.org/bj/456/bj4560241add.htm). Conformer
GGC2 , which has the longest lifetime during the entire simulation,
results from a complete structural rearrangement of the 30s-loop,
arising from concerted conformational change in backbone and
side chains of Glu29 , Ser30 , Gly31 , Gly32 and Cys34 (Figures 6B
and 6C, and Supplementary Figure S3). This results in a transition
from a type I to type II’ β-turn conformation [25,26]. Overlay
of the GGC2 conformer with WT showed a large difference
in the structure of the 30s-loop, including a change in the
conformation of the Cys7 –Cys34 disulfide bridge (Figure 6B). The
rearrangement results in the loss of Gly31 (N) . . . Cys34 (O), but not
of the Gly34 (N) . . . Cys34 (O) hydrogen bond. It is to be noted that
this GGC2 conformation is not observed in the WT (Figure 6C).
P32A mutant
In stark contrast with the P32G mutant, the 30s-loop β-turn
retains the native WT-like type I conformation in the P32A mutant
(Figure 6D), and no additional distinct conformers were observed.
The less bulky alanine side chain compared with the pyrrolidine
ring results in increased flexibility/dynamics of the 30s-loop,
which is evident from backbone φ/ψ angle fluctuations (results
not shown). The relative orientation and modular motions of the
30s-loop in concert with the N-loop and N-terminus are similar
to those observed for the WT.
c The Authors Journal compilation c 2013 Biochemical Society
G31V mutant
Analysis of the trajectory showed that the bulky valine residue
introduces rigidity in the 30s-loop compared with the WT. The
Gly31 (N) . . . Cys34 (O) hydrogen bond was transient, whereas
the Cys34 (N) . . . Gly31 (O) hydrogen bond was intact throughout the simulation, and the β-turn retains the type I conformation.
The only perturbation observed was in the φ/ψ angles of Ser30
which flips from approximately − 50◦ /150◦ to approximately
50◦ /50◦ , which nevertheless does not introduce any perturbation
in the β 1 -strand. The overlay of structures at regular intervals over
the course of the trajectory show that the relative orientation of
the 30s-loop with the N-loop and N-terminal residues are fairly
similar to that seen in WT (Figure 7A).
G31A mutant
Analysis of the MD trajectory showed that there was increased
flexibility in the 30s-loop compared with G31V. We observed two
distinct conformers, APC1 (∼80 %) and APC2 (∼20 %), for the
30s-loop. The major APC1 conformer was similar to that observed
for G31V. The APC2 conformer showed a concerted rearrangement
in the 30s-loop (Ser30 , Gly31 , Pro32 and His33 ) resulting in a type II
β-turn and non-native 30s-loop conformer and altered orientation
of the 30s-loop, N-terminus and N-loop (Figure 7B).
The MD studies of the GP mutants have provided crucial
insights as to why the GP motif is stringent and essential
for optimal cross-talk and receptor activation. In particular,
the non-native conformers observed for P32G and G31A, and the
altered dynamics and the structural reorientation of the 30s-loop,
Long-range coupling modulates chemokine receptor function
Figure 5
247
Binding of the P32G mutant to the CXCR1 N-domain
(A) Sections of (1 H,15 N)-HSQC spectra showing binding-induced chemical-shift perturbations for the P32G mutant. Arrows indicate the direction of peak movement, and unbound peaks are in black
and the final bound peaks are in red. (B) The dissociation constants (K d ) were calculated from binding-induced chemical-shift changes from a subset of well-separated residues such as Ser44 and
+
Trp57 . Average K d values indicate that the P32G mutant binds with the same affinity as the WT (11 +
− 1 compared with 12 − 2 μM). (C) Histogram showing the binding-induced chemical-shift
change (δ) of the P32G mutant (red) compared with the WT (black). Data for Gln8 , Tyr13 , Lys15 , His18 , His33 and Thr37 are not shown due to line broadening (shown as *), and 16, 19, 32 and 53
are proline residues. (D) Surface plot showing the chemical-shift-perturbed residues (red) on CXCR1 N-domain binding (Site-I interaction), in relation to the 30s-loop (shown in orange).The ELR
motif is shown in green.
c The Authors Journal compilation c 2013 Biochemical Society
248
Figure 6
P. R. B. Joseph and others
MD simulations of the P32G mutant
Overlay of representative structures obtained from two distinctly interchanging populations, GGC1 (pink, A) and GGC2 (orange, B) identified over the course of a 200 ns MD simulation, with WT
(in grey). GGC1 is identical with the WT conformation. GGC2 is a non-native conformer with the longest lifetime, resulting from a type I to type II’ β-turn transition in the 30s-loop, due to the
conformational freedom introduced by Gly32 . (C) Comparison of φ/ψ plot for Gly32 (P32G mutant) and Pro32 (WT) showing that the non-native conformations (shown as an orange bar for the P32G
mutant) are not sampled in the WT simulations. Grey bars represent the GGC1 conformer. (D) MD simulations of the P32A mutant. Overlay of a representative structure (blue) from the MD simulation
with WT (grey). The 30s-loop conformation is similar to that of WT. The ELR motif is shown in green.
N-terminal and N-loop domains could not have been predicted
from the MD studies of the WT alone.
DISCUSSION
Protein ligands, such as chemokines, are unusual agonists for class
A GPCRs, as conventional agonists tend to be small molecules,
such as biogenic amines, peptides and those related to taste and
smell [33]. The small molecule agonists activate the receptor
by binding to the extracellular/transmembrane residues (SiteII). The larger size of the chemokines must confer functional
advantages that cannot be achieved by small molecule agonists,
and at the same time, chemokines must have evolved distinct
molecular mechanisms that exploit their large size and multimodular structure for receptor activation. Indeed, chemokines
also bind to the receptor N-terminal domain (Site-I) in addition to
Site-II. However, even this two-site mechanism fails to describe
the large differences in affinity, potency, receptor selectivity and
functional responses for any given chemokine and among different
c The Authors Journal compilation c 2013 Biochemical Society
chemokines. Clearly, other structural components of the protein
must regulate the binding interactions of the N-loop and Nterminal residues.
Our structural, functional and computational characterization
of CXCL8 GP mutants has now provided compelling evidence
that the GP motif couples Site-I and Site-II interactions. The
observation that the G31V mutant has higher affinity and could
be more active in vivo than the WT indicates that Gly31 does not
make direct receptor-binding interactions, as the steric bulk of
the isopropyl group [-CH-(CH3 )2 ] of the valine side chain will
interfere in the binding process. Intriguingly, an alanine mutation
results in lower activity, although its side chain methyl group
(-CH3 ) has smaller steric bulk compared with valine. Receptor
affinities of the Pro32 mutants were dependent on the nature of the
mutation and receptor subtype. The binding of P32A to both
receptors, and the binding of P32G to CXCR2, were similar
to that of WT, but P32G showed impaired binding to CXCR1.
Not surprisingly, P32A showed overall higher activity than the
P32G mutant. The large differences among the P32 mutants also
indicate that the side chain of Pro32 makes no direct receptor
Long-range coupling modulates chemokine receptor function
Figure 7
MD simulations of the G31V and G31A mutants
(A) Overlay of a representative structure of G31V (brown) with WT (grey). The 30s-loop
conformation is similar to the WT conformation except for flipping of the Ser30 φ/ψ angle
to gauge + /gauge + conformation resulting in local perturbation in the loop. (B) Overlay of a
non-native representative structure from the AGC2 population observed for the G31A mutant
(purple) with WT (grey). The ELR motif is shown in green.
interactions. NMR studies show that the local structural changes
as a consequence of mutation cannot be correlated with functional
changes. On the other hand, MD simulations indicate that the GP
mutants populate non-native conformations, which can dominate
and result in suboptimal cross-talk between modules that mediate
the receptor activation process. Other mutational studies have also
shown that the GP mutants bind like the WT [34,35]. These data
together indicate that the GP motif does not make direct receptor
contacts and must indirectly influence the receptor activation
process.
On the basis of our present and previous structure–function
studies, we propose that the ‘conformational-selection’ model
best describes how CXCL8 and all chemokines bind and
activate their receptors [11,12,36]. According to this model, all
protein conformations pre-exist in dynamic equilibrium, and on
binding to their partners, the conformational ensemble undergoes
a population shift, thereby redistributing the conformational
substates [37–41]. The existence of this phenomenon and its
importance in functional scenarios has been well established
through X-ray, NMR, MD, single-molecule fluorescence and
kinetics studies [42–45]. Also, several studies on protein dynamics
has led to the proposal that independent protein segments, such as
the 30s-loop whose dynamics are distinctly different from the rest
of the protein scaffold, can govern conformational transitions and
dictate specificity for protein–protein recognition and activation
of signalling cascades [46,47].
Previous structure–function studies on the conserved Nterminal residues and disulfides have also provided compelling
evidence on the importance of conformational dynamics and
function. CXCL8 structures reveal that the 30s-loop and the
N-terminal residues are linked via the Cys7 –Cys34 disulfide
bond and form a contiguous surface (Figure 1B). Cys7 –Cys34
disulfide is essential for function, as deleting or modifying
the disulfide by introducing non-natural cysteine analogues results
in substantial loss of activity [48,49]. The CXC motif in CXCL8
is also critical, as deleting the intervening Gln8 residue results
in significant loss of function [24]. Structures of Cys7 –Cys34
disulfide-modified mutants are indistinguishable from the WT
and the disulfides are buried, indicating that the disulfides and
CXC motif cannot be involved in direct receptor interactions
and that they indirectly influence the receptor activation process
[48]. Structures and characterization of dynamics from NMR
249
relaxation measurements also reveal that the N-terminal residues
are unstructured and highly mobile, and the N-loop and 30s-loop
residues and the disulfides are structured, but conformationally
flexible [27–29,50–52]. The structure of CXCR1 was recently
solved by solid-state NMR spectroscopy, but the N-terminal
domain in the structure could not be defined, indicating that it
is dynamic and conformationally flexible [53]. However, binding
studies of CXCL8 to the isolated CXCR1 N-domains have shown
that the N-domain adopts an extended and definite structure in the
bound form [54].
We propose that the N-terminus, N-loop and 30s-loop, and
the disulfide residues, exist as a conformational ensemble.
Chemokine receptors also exist as a conformational ensemble,
and it has been proposed that the ligands selectively bind the
active conformer, and/or that binding results in a redistribution
of the ensemble [55,56]. In essence, the receptor activation
involves a series of co-ordinated conditional and discrete binding
steps involving both partners, a process termed ‘interdependent
protein dance’ [57,58]. Therefore each binding step generates
a distinctly different conformational state dictated by the
conformational ensemble equilibrium, and a mutation could
result in a WT-like ensemble with no effect on affinity or
activity, a functionally impaired ensemble that has reduced
binding affinity and/or activity for some (but not all) signalling
pathways, or a functionally enhanced ensemble that has increased
binding affinity and/or activity. The proposed model is robust
and in excellent agreement with the generalized concept of
‘conformational selection and population shift’, and can account
for the large range in the observed changes both in affinity
and activity [11,12]. For instance, the new model can explain
the large differences in affinity and activity of the P32G
and G31V and of the N-terminal ELR and disulfide mutants, and
the antagonist activity of the N-terminal (such as R6K) and GP
(such as G31P) mutants [31,59]. Furthermore, our proposed model
can also explain how CXCL7/NAP-2 can activate neutrophils
despite having a threonine substitution for proline (Figure 1A);
its conformational ensemble, like in the G31V, enables receptor
activation at Site-II [60].
Temporally, the very first step involves binding interactions
between chemokine N-loop and receptor N-domain residues at
Site-I, and the very last step involves receptor activation by the
N-terminal residues at Site-II. The initial binding must induce
structural/dynamic changes in the ligand that facilitate activation
by the ELR residues. Disulfide bonds, especially the Cys7 –Cys34
link, clearly play an important role for activation at Site-II [24,48].
We propose that the penultimate step involves the conformational
switching by the GP motif of the N-terminal ELR residues for
receptor activation, resulting in downstream signalling events.
We further propose that a conformational ensemble model best
describes activation profiles and functional responses for all
chemokine/receptor pairs. A network of mosaic interactions in
the ensemble provides superb control in mediating the multistep
processes involved in multiple receptor signalling pathways, and
this shared property among chemokines enables diverse ligand
and receptor selectivity and different signalling activities.
Interestingly, a tripeptide PGP (proline-glycine-proline)
recruits neutrophils, and it has been proposed that it does so
by activating CXCR1/CXCR2 receptors [61]. These authors
propose that the PGP tripeptide binds CXCR1/CXCR2 receptors
by mimicking the 30s-loop GP residues on the basis that the
peptide adopts the conformation of the GP motif observed in
CXCL8 structures, and a previous study that showed mutating
GP to PG results in a significant loss of activity [7,61]. The PGP
peptide shows much lower potency (>1000-fold) and the authors
propose that this is because the peptide binds only part of the
c The Authors Journal compilation c 2013 Biochemical Society
250
P. R. B. Joseph and others
receptor. High binding affinities of the GP mutants, especially
of the G31V mutant, cannot account for activity of the PGP
peptides as due to mimicking the structure and function of the
GP motif. Our conclusion that the GP motif does not make direct
receptor interactions, but modulates the conformational ensemble,
indicates that any semblance of the tripeptide to the GP motif
structure is incidental. This provides compelling evidence that
even if PGP does bind CXCR1/CXCR2 receptors for neutrophil
recruitment, its mechanism of action is completely unrelated to
the CXCL8 mode of action [62].
AUTHOR CONTRIBUTION
Krishna Rajarathnam, Prem Joseph, Kirti Sawant, Ricardo Richardson and Roberto Garofalo
designed the research. Prem Joseph, Kirti Sawant, Mesias Pedroza and Angela Isley
performed the experiments, and all the authors analysed the data. Prem Joseph and
Krishna Rajarathnam wrote the paper.
ACKNOWLEDGEMENTS
We thank Ms Pavani Gangavarapu, Ms Meena Shanmugasundaram, Ms Renling Xu and
Dr Aishwarya Ravindran for helpful discussions and technical support. We thank Dr
Kathryn Cunningham and Dr Blake Rasmussen (The University of Texas Medical Branch,
Galveston, TX, U.S.A.) for access to the Flexstation plate reader and the Chemidoc, and
Dr Tianzhi Wang (The University of Texas Medical Branch, Galveston, TX, U.S.A.) for help
with the NMR instrumentation. We also thank the high-performance computing resource
provided by the TACC (Texas Advanced Computing Center) at University of Texas, Austin,
TX, U.S.A.
FUNDING
This work was supported by the National Institutes of Health [grant numbers R01 AI069152,
AI38910 and P01 HL107152].
REFERENCES
1 Bonecchi, R., Galliera, E., Borroni, E. M., Corsi, M. M., Locati, M. and Mantovani, A.
(2009) Chemokines and chemokine receptors: an overview. Front. Biosci. 14, 540–551
2 Rajagopalan, L. and Rajarathnam, K. (2006) Structural basis of chemokine receptor
function: a model for binding affinity and ligand selectivity. Biosci. Rep. 26, 325–339
3 Jarnagin, K., Grunberger, D., Mulkins, M., Wong, B., Hemmerich, S., Paavola, C., Bloom,
A., Bhakta, S., Diehl, F., Freedman, R. et al. (1999) Identification of surface residues of the
monocyte chemotactic protein 1 that affect signaling through the receptor CCR2.
Biochemistry 38, 16167–16177
4 Xanthou, G., Williams, T. J. and Pease, J. E. (2003) Molecular characterization of the
chemokine receptor CXCR3: evidence for the involvement of distinct extracellular
domains in a multi-step model of ligand binding and receptor activation. Eur. J. Immunol.
33, 2927–2936
5 Wells, T. N., Power, C. A., Lusti-Narasimhan, M., Hoogewerf, A. J., Cooke, R. M., Chung,
C. W., Peitsch, M. C. and Proudfoot, A. E. (1996) Selectivity and antagonism of
chemokine receptors. J. Leukocyte Biol. 59, 53–60
6 Crump, M. P., Gong, J. H., Loetscher, P., Rajarathnam, K., Amara, A., Arenzana-Seisdedos,
F., Virelizier, J. L., Baggiolini, M., Sykes, B. D. and Clark-Lewis, I. (1997) Solution
structure and basis for functional activity of stromal cell-derived factor-1; dissociation of
CXCR4 activation from binding and inhibition of HIV-1. EMBO J. 16, 6996–7007
7 Clark-Lewis, I., Dewald, B., Loetscher, M., Moser, B. and Baggiolini, M. (1994) Structural
requirements for interleukin-8 function identified by design of analogs and CXC
chemokine hybrids. J. Biol. Chem. 269, 16075–16081
8 Clark-Lewis, I., Schumacher, C., Baggiolini, M. and Moser, B. (1991) Structure–activity
relationships of interleukin-8 determined using chemically synthesized analogs. Critical
role of NH2 -terminal residues and evidence for uncoupling of neutrophil chemotaxis,
exocytosis, and receptor binding activities. J. Biol. Chem. 266, 23128–23134
9 Clark-Lewis, I., Dewald, B., Geiser, T., Moser, B. and Baggiolini, M. (1993) Platelet factor
4 binds to interleukin 8 receptors and activates neutrophils when its N terminus is
modified with Glu-Leu-Arg. Proc. Natl. Acad. Sci. U.S.A. 90, 3574–3577
10 Dewald, B., Moser, B., Barella, L., Schumacher, C., Baggiolini, M. and Clark-Lewis, I.
(1992) IP-10, a γ -interferon-inducible protein related to interleukin-8, lacks neutrophil
activating properties. Immunol. Lett. 32, 81–84
c The Authors Journal compilation c 2013 Biochemical Society
11 Ma, B., Kumar, S., Tsai, C. J. and Nussinov, R. (1999) Folding funnels and binding
mechanisms. Protein Eng. 12, 713–720
12 Tsai, C. J., Kumar, S., Ma, B. and Nussinov, R. (1999) Folding funnels, binding funnels,
and protein function. Protein Sci. 8, 1181–1190
13 Fernando, H., Nagle, G. T. and Rajarathnam, K. (2007) Thermodynamic characterization of
interleukin-8 monomer binding to CXCR1 receptor N-terminal domain. FEBS J. 274,
241–251
14 Rajarathnam, K., Sykes, B. D., Kay, C. M., Dewald, B., Geiser, T., Baggiolini, M. and
Clark-Lewis, I. (1994) Neutrophil activation by monomeric interleukin-8. Science 264,
90–92
15 Rajagopalan, L. and Rajarathnam, K. (2004) Ligand selectivity and affinity of chemokine
receptor CXCR1. Role of N-terminal domain. J. Biol. Chem. 279, 30000–30008
16 Nasser, M. W., Raghuwanshi, S. K., Grant, D. J., Jala, V. R., Rajarathnam, K. and
Richardson, R. M. (2009) Differential activation and regulation of CXCR1 and CXCR2 by
CXCL8 monomer and dimer. J. Immunol. 183, 3425–3432
17 Richardson, R. M., Pridgen, B. C., Haribabu, B., Ali, H. and Snyderman, R. (1998)
Differential cross-regulation of the human chemokine receptors CXCR1 and CXCR2.
Evidence for time-dependent signal generation. J. Biol. Chem. 273, 23830–23836
18 Nasser, M. W., Marjoram, R. J., Brown, S. L. and Richardson, R. M. (2005)
Cross-desensitization among CXCR1, CXCR2, and CCR5: role of protein kinase C-ε. J.
Immunol. 174, 6927–6933
19 Gangavarapu, P., Rajagopalan, L., Kolli, D., Guerrero-Plata, A., Garofalo, R. P. and
Rajarathnam, K. (2012) The monomer–dimer equilibrium and glycosaminoglycan
interactions of chemokine CXCL8 regulate tissue-specific neutrophil recruitment.
J. Leukocyte Biol. 91, 259–265
20 Ravindran, A., Joseph, P. R. and Rajarathnam, K. (2009) Structural basis for differential
binding of the interleukin-8 monomer and dimer to the CXCR1 N-domain: role of coupled
interactions and dynamics. Biochemistry 48, 8795–8805
21 Case, D. A., Cheatham, III, T. E., Simmerling, C. L., Wang, J., Duke, R. E., Luo, R., Walker,
R. C., Zhang, W., Merz, K. M., Roberts, B. et al. (2012) AMBER 12, University of
California, San Francisco
22 Lee, M. C., Deng, J., Briggs, J. M. and Duan, Y. (2005) Large-scale conformational
dynamics of the HIV-1 integrase core domain and its catalytic loop mutants. Biophys. J.
88, 3133–3146
23 Humphrey, W., Dalke, A. and Schulten, K. (1996) VMD: visual molecular dynamics.
J. Mol. Graphics 14, 33–38, 27–28
24 Joseph, P. R., Sarmiento, J. M., Mishra, A. K., Das, S. T., Garofalo, R. P., Navarro, J. and
Rajarathnam, K. (2010) Probing the role of CXC motif in chemokine CXCL8 for high
affinity binding and activation of CXCR1 and CXCR2 receptors. J. Biol. Chem. 285,
29262–29269
25 Wilmot, C. M. and Thornton, J. M. (1988) Analysis and prediction of the different types of
β-turn in proteins. J. Mol. Biol. 203, 221–232
26 Rose, G. D., Gierasch, L. M. and Smith, J. A. (1985) Turns in peptides and proteins. Adv.
Protein Chem. 37, 1–109
27 Clore, G. M., Appella, E., Yamada, M., Matsushima, K. and Gronenborn, A. M. (1990)
Three-dimensional structure of interleukin 8 in solution. Biochemistry 29, 1689–1696
28 Rajarathnam, K., Clark-Lewis, I. and Sykes, B. D. (1995) 1H NMR solution structure of an
active monomeric interleukin-8. Biochemistry 34, 12983–12990
29 Baldwin, E. T., Weber, I. T., St Charles, R., Xuan, J. C., Appella, E., Yamada, M.,
Matsushima, K., Edwards, B. F., Clore, G. M., Gronenborn, A. M. et al. (1991) Crystal
structure of interleukin 8: symbiosis of NMR and crystallography. Proc. Natl. Acad. Sci.
U.S.A. 88, 502–506
30 Wishart, D. S. (2011) Interpreting protein chemical shift data. Prog. Nucl. Magn. Reson.
Spectrosc. 58, 62–87
31 Cheng, H. T., Huang, K. C., Yu, H. Y., Gao, K. J., Zhao, X., Li, F., Town, J., Gordon, J. R.
and Cheng, J. W. (2008) A new protocol for high-yield purification of recombinant human
CXCL8(3–72)K11R/G31P expressed in Escherichia coli . Protein Expression Purif. 61,
65–72
32 Park, S. H., Casagrande, F., Das, B. B., Albrecht, L., Chu, M. and Opella, S. J. (2011)
Local and global dynamics of the G protein-coupled receptor CXCR1. Biochemistry 50,
2371–2380
33 Joost, P. and Methner, A. (2002) Phylogenetic analysis of 277 human G-protein-coupled
receptors as a tool for the prediction of orphan receptor ligands. Genome Biol. 3,
RESEARCH0063
34 Williams, G., Borkakoti, N., Bottomley, G. A., Cowan, I., Fallowfield, A. G., Jones, P. S.,
Kirtland, S. J., Price, G. J. and Price, L. (1996) Mutagenesis studies of interleukin-8.
Identification of a second epitope involved in receptor binding. J. Biol. Chem. 271,
9579–9586
35 Li, F., Zhang, X. and Gordon, J. R. (2002) CXCL8(3–73)K11R/G31P antagonizes ligand
binding to the neutrophil CXCR1 and CXCR2 receptors and cellular responses to
CXCL8/IL-8. Biochem. Biophys. Res. Commun. 293, 939–944
Long-range coupling modulates chemokine receptor function
36 Frauenfelder, H., Sligar, S. G. and Wolynes, P. G. (1991) The energy landscapes and
motions of proteins. Science 254, 1598–1603
37 Tsai, C. J., Del Sol, A. and Nussinov, R. (2009) Protein allostery, signal transmission
and dynamics: a classification scheme of allosteric mechanisms. Mol. BioSys. 5,
207–216
38 Tsai, C. J., del Sol, A. and Nussinov, R. (2008) Allostery: absence of a change in shape
does not imply that allostery is not at play. J. Mol. Biol. 378, 1–11
39 del Sol, A., Tsai, C. J., Ma, B. and Nussinov, R. (2009) The origin of allosteric functional
modulation: multiple pre-existing pathways. Structure 17, 1042–1050
40 Ma, B., Tsai, C. J., Pan, Y. and Nussinov, R. (2010) Why does binding of proteins
to DNA or proteins to proteins not necessarily spell function? ACS Chem. Biol. 5,
265–272
41 Ma, B., Shatsky, M., Wolfson, H. J. and Nussinov, R. (2002) Multiple diverse ligands
binding at a single protein site: a matter of pre-existing populations. Protein Sci. 11,
184–197
42 Lange, O. F., Lakomek, N. A., Fares, C., Schroder, G. F., Walter, K. F., Becker, S., Meiler, J.,
Grubmuller, H., Griesinger, C. and de Groot, B. L. (2008) Recognition dynamics up to
microseconds revealed from an RDC-derived ubiquitin ensemble in solution. Science
320, 1471–1475
43 Boehr, D. D., McElheny, D., Dyson, H. J. and Wright, P. E. (2006) The dynamic energy
landscape of dihydrofolate reductase catalysis. Science 313, 1638–1642
44 Greenleaf, W. J., Woodside, M. T. and Block, S. M. (2007) High-resolution,
single-molecule measurements of biomolecular motion. Annu. Rev. Biophys. Biomol.
Struct. 36, 171–190
45 Qin, H., Lim, L. and Song, J. (2012) Protein dynamics at Eph receptor–ligand interfaces
as revealed by crystallography, NMR and MD simulations. BMC Biophys. 5, 2
46 Csermely, P., Palotai, R. and Nussinov, R. (2010) Induced fit, conformational selection
and independent dynamic segments: an extended view of binding events. Trends
Biochem. Sci. 35, 539–546
47 Nussinov, R. and Ma, B. (2012) Protein dynamics and conformational selection in
bidirectional signal transduction. BMC Biol. 10, 2
48 Rajarathnam, K., Sykes, B. D., Dewald, B., Baggiolini, M. and Clark-Lewis, I. (1999)
Disulfide bridges in interleukin-8 probed using non-natural disulfide analogues:
dissociation of roles in structure from function. Biochemistry 38, 7653–7658
49 Rajagopalan, L., Chin, C. C. and Rajarathnam, K. (2007) Role of intramolecular disulfides
in stability and structure of a noncovalent homodimer. Biophys. J. 93, 2129–2134
251
50 Clore, G. M. and Gronenborn, A. M. (1991) Comparison of the solution nuclear magnetic
resonance and crystal structures of interleukin-8. Possible implications for the
mechanism of receptor binding. J. Mol. Biol. 217, 611–620
51 Grasberger, B. L., Gronenborn, A. M. and Clore, G. M. (1993) Analysis of the backbone
dynamics of interleukin-8 by 15 N relaxation measurements. J. Mol. Biol. 230, 364–
372
52 Bonvin, A. M. and Brunger, A. T. (1995) Conformational variability of solution nuclear
magnetic resonance structures. J. Mol. Biol. 250, 80–93
53 Park, S. H., Das, B. B., Casagrande, F., Tian, Y., Nothnagel, H. J., Chu, M., Kiefer, H.,
Maier, K., De Angelis, A. A., Marassi, F. M. and Opella, S. J. (2012) Structure of the
chemokine receptor CXCR1 in phospholipid bilayers. Nature 491, 779–783
54 Skelton, N. J., Quan, C., Reilly, D. and Lowman, H. (1999) Structure of a CXC
chemoline-receptor fragment in complex with interleukin-8. Structure 7, 157–168
55 Kobilka, B. K. and Deupi, X. (2007) Conformational complexity of G-protein-coupled
receptors. Trends Pharmacol. Sci. 28, 397–406
56 Vaidehi, N. and Kenakin, T. (2010) The role of conformational ensembles of seven
transmembrane receptors in functional selectivity. Curr. Opin. Pharmacol. 10, 775–781
57 Kovacs, I. A., Szalay, M. S. and Csermely, P. (2005) Water and molecular chaperones act
as weak links of protein folding networks: energy landscape and punctuated equilibrium
changes point towards a game theory of proteins. FEBS Lett. 579, 2254–2260
58 Antal, M. A., Bode, C. and Csermely, P. (2009) Perturbation waves in proteins and protein
networks: applications of percolation and game theories in signaling and drug design.
Curr. Protein Pept. Sci. 10, 161–172
59 Moser, B., Dewald, B., Barella, L., Schumacher, C., Baggiolini, M. and Clark-Lewis, I.
(1993) Interleukin-8 antagonists generated by N-terminal modification. J. Biol. Chem.
268, 7125–7128
60 Rajarathnam, K., Kay, C. M., Dewald, B., Wolf, M., Baggiolini, M., Clark-Lewis, I. and
Sykes, B. D. (1997) Neutrophil-activating peptide-2 and melanoma growth-stimulatory
activity are functional as monomers for neutrophil activation. J. Biol. Chem. 272,
1725–1729
61 Weathington, N. M., van Houwelingen, A. H., Noerager, B. D., Jackson, P. L., Kraneveld,
A. D., Galin, F. S., Folkerts, G., Nijkamp, F. P. and Blalock, J. E. (2006) A novel peptide
CXCR ligand derived from extracellular matrix degradation during airway inflammation.
Nat. Med. 12, 317–323
62 Rajarathnam, K. (2006) Reply to ‘A novel peptide CXCR ligand derived from extracellular
matrix degradation during airway inflammation’. Nat. Med. 12, 603–604
Received 28 January 2013/12 September 2013; accepted 16 September 2013
Published as BJ Immediate Publication 16 September 2013, doi:10.1042/BJ20130148
c The Authors Journal compilation c 2013 Biochemical Society
Biochem. J. (2013) 456, 241–251 (Printed in Great Britain)
doi:10.1042/BJ20130148
SUPPLEMENTARY ONLINE DATA
Dynamic conformational switching in the chemokine ligand is essential for
G-protein-coupled receptor activation
Prem Raj B. JOSEPH*, Kirti V. SAWANT*, Angela ISLEY†, Mesias PEDROZA*, Roberto P. GAROFALO‡§,
Ricardo M. RICHARDSON† and Krishna RAJARATHNAM*‡1
*Department of Biochemistry and Molecular Biology, The University of Texas Medical Branch, 301 University Boulevard, Galveston, TX 77555-1055, U.S.A., †Julius L. Chambers
Biomedical/Biotechnology Research Institute and the Department of Biology, North Carolina Central University, Durham, NC 27707, U.S.A., ‡Department of Microbiology and
Immunology, The University of Texas Medical Branch, 301 University Boulevard, Galveston, TX 77555-1055, U.S.A., §Department of Pediatrics, The University of Texas Medical Branch,
301 University Boulevard, Galveston, TX 77555-1055, U.S.A., and Sealy Center for Structural Biology and Molecular Biophysics, The University of Texas Medical Branch, 301
University Boulevard, Galveston, TX 77555-1055, U.S.A.
Figure S1
1
NMR structural characteristics of the G31V mutant
15
(A) ( H, N,)-HSQC spectrum of G31V (red) superimposed on the WT (black). Characteristic up-field- and down-field-shifted peaks as well as residues in the vicinity of the mutation are labelled.
Residues 8 and 33 are broadened out in the G31V spectrum. (B and C) Diagonal plots of HN (B) and N (C) chemical shifts of G31V (y -axis) compared with WT (x -axis). Data show that the chemical
shifts are essentially same, indicating that the structure of the mutant is essentially identical with the WT.
1
To whom correspondence should be addressed (email [email protected]).
c The Authors Journal compilation c 2013 Biochemical Society
P. R. B. Joseph and others
Figure S2
Binding of G31V mutants to the CXCR1 N-domain
(A) Histogram showing the binding-induced chemical-shift change (δ) of the G31V mutant (red) compared with WT (black). Data for Gln8 , Tyr13 , Lys15 , His18 , His33 and Thr37 are not shown due to
line broadening (shown as *), and 16, 19, 32 and 53 are proline residues. (B) Sections of (1 H,15 N,)-HSQC spectra showing binding-induced chemical-shift perturbations for the G31V mutant. Arrows
indicate the direction of peak movement, and unbound peaks are in black and the final bound peaks are in red. (C) The dissociation constants (K d ) were calculated from binding-induced chemical-shift
changes from a subset of well-separated residues such as Ser44 and Trp57 . Average K d indicates that the G31V mutant binds with ∼4-fold lower affinity compared with WT (50 +
− 10 μM compared
with 12 +
− 2 μM).
c The Authors Journal compilation c 2013 Biochemical Society
Long-range coupling modulates chemokine receptor function
Figure S3
30s-loop conformation in the P32G mutant
ϕ/ψ plots of 30s-loop residues Glu29 , Ser30 , Gly31 and Cys34 . Note the concerted conformational rearrangements effected during exchange between the native (GGC1 , grey bars) and non-native
(GGC2 , orange bars) 30s-loop conformations.
c The Authors Journal compilation c 2013 Biochemical Society
P. R. B. Joseph and others
Figure S4
30s-loop conformation in WT
ϕ/ψ plots of 30s-loop residues Glu29 , Ser30 , Gly31 and Cys34 . Note the two distinct conformational states sampled by residues Glu29 and Ser30 .
Received 28 January 2013/12 September 2013; accepted 16 September 2013
Published as BJ Immediate Publication 16 September 2013, doi:10.1042/BJ20130148
c The Authors Journal compilation c 2013 Biochemical Society