at why. CC chemokines do not bind to the. CXC receptors in several ways. First, ..... level of the various receptor types. It was therefore important to characterize ... [7]. No binding data has been shown to date for the recombinant receptor.
Selectivity
and antagonism
of chemokine
receptors
Timothy N.C. Wells,* Christine A. Power, Manjula Lusti-Narasimhan, Arlene J. Hoogewerf, Robert M. Cooke,t Chun-wa Chung,t Manuel C. Peitsch, and Amanda E. I. Proudfoot *Glaxo
Instiuuefor
Stevenage,
Abstract:
Molecular
United
The
Geneva,
Biology,
Switzerland
and
G1axo
Weilcome
Medicines
Research
Centre,
Kingdom
chemokine
superfamily
can
be
INTRODUCTION
sub-
divided
into two groups based on their amino terminat cysteine spacing. The CXC chemokines are primarily involved in neutrophil-mediated inflammation and, so far, two human cloned. The CC chemokines chronic
inflammation,
a fourth
leukocyte
recently
receptor
Understanding range of agonists potent vestigate
and
what makes is important
by looking
to the at the
CXC role
ligand,
and
at why
one
CC
solved
have been involved in have
receptor if we are
bind its develop
We have started of this receptor
to inselec-
ways.
do
not
bind
First, we looked structure of the
three-dimensional
struc-
for the protein-1, interface
flammatory ferent dimer interleukin-8
(IL-8).
of
tures
all
the
CC
chemokines
the are
monomer
very
and
in
bind
CC-chemokine
receptor-i
(CC-CKR-1)
troduces monocyte chemoattractant other mutations in this region, we interaction we have flue
with the found that
of HANTES
N-terminus modification
by addition
it into
an antagonist
together, receptor
this
with
data activation
activity. can show
in-
Using a direct
of one
amino
acid
a two-site selectivity
have
for CC
Biol.
distress
gic
This in
men.sional
structures
interleukin-8
the
CC
chemokine
and
characterization
areas
the
for
which such
arthritis,
of the
work
taken
to
receptors;
carried
and
out
clone
which
novel leads
between
finally,
of antagonists
and
structural
chemokines,
of specificity
there as aller-
three-dimensional of
control
clinically from acute respiratory
diseases
approaches
mutagenesis
of
and
re-
Chemok-
range acute
dermatitis,
three the
receptors; into
fatal
atopic
reviews
and
CXC
the
to
and
identification
to CC
chemokine
re-
ceptors.
i996.
Abbreviations: netic
resonance;
acids;
IL-B,
A
and
melanoma tractant tory
resonance
spectroscopy
-
three-di-
CC-CKR,
CC-chemokine
ENA-78,
epithelial
interleukin-8;
-B;
NAP-2, growth
IL-8RA
and
-1 and
-3;
requests: chemin
PCR,
normal
N.C.
Aulx,
interleukin-8 and
MIP-1,
polymerase
T cell
Timothy des
MCP-1 and
Wells,
receptors-
-3, monocyte
chain and Glaxo
mag-
78 amino
MGSA/Gro-a,
macrophage
expressed
1228
nuclear
factor,
polypeptide-2;
MIP-la
-1;
on actvation
NMR,
activating
IL-8RB,
stimulatingactivity;
proteins
14
receptor,
neutropil
neutrophil-activating
polypeptide-laand Reprint
magnetic
physiological
a variety
is often
psoriasis,
chemokine
in
As
are highly receptors at
range.
to chronic
laboratories:
insights
their
diseases. These diseases such as which
article
our
exert
implicated
treatment,
asthma,
Biology,
nuclear
been
syndrome,
regulated
Words:
and
on neutrophils.
chemokines bind their
subnanomolar-concentration
important inflammatory neutrophil-mediated
activation
types in inflaminvolved in the
are less important in the such as monocytes. CC profile in terms of cellular
or no effect
concentrations, in the
selective
Taken
model between
Leukoc.
Key
in the
variety of cell are principally
little
into
spacing of cysteine and on overall se-
makes
potency.
chemokines, with an initial receptor contact by the main body of the chemokine, and provided by the amino terminal region. J. 53-60;
a role
and
have
on the region,
molecular divided
.
and CXC provided activation
59:
play
of a large chemokines
are
ofCC-CKR1 Third, of the amino termi-
nanomolar
suggests and for
and
They
selectivity
studies
tion
based terminal
with
They
well as this selectivity of response, potent molecules, the majority
tations in IL-8,
kDa.
of neutrophils, and of other leukocytes, have the opposite
atheroma.
region, Leu-25 at this posimutant IL-8 to
of proteins
10
activation activation chemokines
physiological concentrations the proteins are likely to be monomeric. Second, by examining all the known CC and CXC chemokines, we have found a region that differs between the two subfamilies. Muof one of the residues in this to tyrosine (which is conserved CC chemokines) enables the
family
8 and
identity.
and recruitment mation. CXC
is no effective
at
a large
between
CC chemokines in the amino
residues
ines
struc-
similar,
and
CXC
sponse
chemokine macrophage inand it has a completely difto that of the CXC chemokine
However,
of
nanomolar
hire of RANTES using nuclear magnetic resonance spectroscopy. The structure is similar to that already determined
are
masses
quence
of ligands.
group
chemokines
the
cloned
to
receptors in several of the three-dimensional have
we
this
for
selective antagonists. the molecular basis
tivity
receptors tend to be
Chemokines
chemoatinflamma-
reaction;
RANTES,
secreted.
Institute
Plan-les-Ouates,
for
Molecular
Geneva,
Switzer-
land. Received
August
Journal
of
19,
Leukocyte
1995;
accepted
Biology
September
Volume
29,
59,
1995.
January
1996
53
CHEMOKINE RECEPTORS AS THE TARGETS OF ANTI-INFLAMMATORY Chemokines were ability to recruit inflammatory
Clinically
to be present at elevated diseases. Based on this hypothesis that interfering
specificities, adds to the role in disease, and thus least
27
human
chemokine-like
sequences
that more than for the recruitment
one
key
may
However, complexity
complexity identifying To
chemokines:
cloned genes, sequence
chemokine
have
been
there
are
me
the in
some
from
purified
1).
chemokine is going of any individual cell vary
from
disease
antagonism result
of
of inflam-
were
cloned,
These
were
as many
In 1991, and
two
shown
called
the
of the
chemok-
receptors
for CXC
to be present
interleukin-8
on neureceptor-A
own,
receptors,
have
used
egy, designing gions
to be responsible type, and that the to disease. degree of Our ulti-
of groups,’ chain
This initially produced for which ligands still
our
(PCR)
strat-
in studying
a large number of orphan have to be identified [2].
The
difficulty
due
to the
and The
the availability of labeled chemokines CC-CKR-1 originally isolated from
cells
including
reaction
degenerate primers based on conserved rein chemokine and chemotactic factor
found
receptors. receptors
likely
a variety
a polymerase
lack
these
receptors
of reproducibility
[2, 3] has
been
shown
has
in ligand
been
largely
binding for
HL6O
to be activated
assays analysis. or U937
by macrophage
Chemokines
IL- 8
SAKELRCQCIICTYSKPFHPKFIKELRVIESGPHCANTEI
T3 976 5 NAP - 2 ENA-
IVKLSD
TLSPVQGVLEVYYTSLRCRCVQESSVFIPRRFIDRIQILPRGNGCPRKEI AELR#{216}4CIKTFSG . IHPKNIQSLEVIGKGTHCNQVEVIATLKD
78
. SIVCVDPQAEWIQR*IEVLRKRSSSTLPV . GRKICLDPDAPRIKJCIVQKKLAGDESAD
. VHPKMISNLQVFAIGPQCSKVEVVASLKN
. GKEICLDPEAPFLKKVIQICILDGGNKEN
ASVATELRCQCLQTLQG
. IHPKNIQSVNVKSPGPNCAQTEVIATLKN
. GRKACLNPASPIVXKI
GRO-b
APLATELRCQCLQTLQG ASVVTELRCQCLQTLQG
. IHLKNIQSVKVKSPGPHCAQTEVIATLKN
. GQXACLNPASIk4VXXI
Gro-
AGPAAAVLRELRCVCLQTTQG
. GRELCLDPKENWVQRWEKFLKRAENS
IVWKKNK
GRO- a c
I P- 1 0 GCP NAP-
a particular
in a reduction
to characterize
as possible.
chemokine
proteins,
point of view, this not be a problem.
that
would
(IL-8R-A), which is specific for IL-8; and IL-8R-B which binds IL-8 and other CXC chemokines such as neutrophilactivating polypeptide-2 (NAP-2), epithelial neutrophil activating factor, 78 amino acids (ENA-78), and melanoma growth stimulating activity (MGSA/Gro-a) [1]. To find CC
of at
It is therefore
show
important
receptors
trophils.
and more recently a large number tag sequences that correspond to (Fig.
to
The molecular basis of selectivity is far more likely at the level of the various receptor types. It was
chemokines
their are
reports
is
receptor
therefore
shown
of understanding which chemokines
date
from a therapeutic and pleiotropy may
Cxc
mation. to lie
in a reduction in the level of the sheer number of chemokiso far, and their overlapping
targets.
some from of expressed
they
goal
chemokine
levels in a variety of inflammatory type of data, we can suggest with the chemokine cascade
diseased tissues will result inflammation. In this light, nes that have been discovered
therapeutic
mate
initially discovered on the basis of their and activate a variety of cell types in
conditions.
good
POTENTIAL THERAPY
-
2 4
SDF-
. IHLKNIQSVNVRSPGPHCAQTEVIATLKN
VPLSRTVRCTCISISNQPVNPRSLEKLEI GPVSAVLTELRCTCLRVTLR EAELQDLQV
1
PF4
. GKXACLNPASPMVQKI
IPASQPCPRVEI
IAThKXXGEKRCLNPESKAIKNLLKAVSXEMSKRSP
. VNPKTIGKLQVFPAGPQCSKVEVVASLKN
. GKQVCLDPEAPFLKXVIQKILDSGNK
. . . KTVXQVSPVHITSLEVDKAGR
. TPNCALQIVARLXNNN
GKPVSLSYRCPCRFFESH
. VAR.ANVKHLKILN
EAEEDGDLQCLCVKTTSQ
. VRPRHITSLEVIKAGPHCPTAQLIATLXN
MIG
IEXMLNSDKSN IEXMLKNGKSN IEKILNXGSTN
TPVVRKGRCSCISTNQGTIHLQSLKDLKQFAPSPSCEKIEI
IATLXN
sssss
ssssssss
sssss
ssssssss
. RQVCIDPKLKWIQEYLEKALNK
IK.XLLBS
. GRKICLDLQAPLYIUCI
. GVQTCLNPDSADVXELIKICWEKQVSQ
sssssss
HHHHHHIi14HHflffi
CC Chemokines RANTES
SPYSSDT
I 3 09 MIP - la
SKSMQVPFSRC ASLAADTPTAC
R83 915
. TPC
SFHFAADC
. RQVCANPBKXWVREYINSLEMS
. CFSFAEQEIPLRAILCYRNTSSI
. . CSNEGLI
. KEACALDTVGWVQRHRKMLRNCPSXRK
. CFSYTSRQI
. . CSKPGVI
PQNFIADYFETSSQ
TESSSRGPYHPSEC
. CPTY’I’TYKIPRQRIMDYYETNSQ
T58 84 7 MIPlb
QPKVPEVGEHPSTC APMGSDPPTAC
. CLXYYEKVLPRRLWGYRXALN
CSKPGVI
.
. . . CHLPAI
-
1
QPDAINAPVTC
. CYNFTNRKISVQRLASYRRITSSK.
MCP
-
2
QPDSVSI
MCP-
3
T642
62
. RQICADPNKKWVQICYISDLKLNA
RQVCAXPSGPGVQDQ4X1LKPYSI
. MSVCTNPSDKWVQDYIXIt4KEN
I FVTKRGTEEVCTNPNDDWQPTPACLPGNFVHG . KQVCADPSESWVQEYVYDLELN
. . CSQPAWFQTXRS
CLGYTDRILHPKFIVGFTRQLANEGCDINAI
MCP
. RQVCADPSEEWVQKYVSDLELSA
FLTKXG
. . CSKPGIVFITKRG
. CFSYTARKLPRNFVVDYYETSSL
SEAASNFDC
FKLKRG FLTKR.S
. . CPKLGVILLTKRG
. CTSYISQSIPCSLF4KSYFETSSB
HCC1
D3 1 06 5
HHHHHHHIIHHHNHH
. . CSNPAVVFVTRKN
. CLVYTXXQIPQKFIVDYSETSPQ
QVGTNXELC
R9 17 3 3
sssssss
. CPAYIARPLPRAHIKETFYTSGK
IFHTKXK
. LSVCANPXQTWVXYIVRLLRKICVKNt4
CPKEAVIFKTIVA
. KEICADPKQKWVQDSMDHLDKQTQTPKT
. CFNVINRKIPIQRLESYTRITNIQ
. CPKEAVIFXTKRG
. KEVCADPKERWVRDSMKHLDQIFQNLKP
QPVGINTS7I’C
. CYRFINKXI
. CPREAVIPXTXLD
. KEICADPTQKWVQDF4KHLDKKTQTPKL
QPDALNVPSTC
. CFTFSSKKISLQRLKSYVITTSR
PITC
PKQRLESYRRTFSSH
. . CPQKAVIFRTKLGQGDLCADPKEKWVQNYMXHLGAGKPT
C Chemokine LTN
Fig.
1
group
II- sheet a search
54
.
GVEVSDKRT
Multiple
sequence
corresponding and ofthe
Journal
alignment
to Leu-25
H representing NCBI
of
database
Leukocyte
in
of human
IL-8
a-helix.
Included
in September
Biology
is also
. CVSLTTQP.LPVSRIKTYTITEG
CXC,
CC,
shown in the
and
in bold. alignment
C chemokines. The are
1995.
Volume
59,
January
1996
secondary several
. .
The
. AVI
SLR
conserved
structure sequences
.
FITKRGLK
. VCADPQATWVRDVVRSMDRKSNTRNNMIQT
cysteines of the
from
are
shown
chemokines
expressed
sequence
The
iii bold.
is indicated, tag
...
conserved
hydrophobic
S representing
databases,
which
residues were
found
in a from
inflammatory
polypeptide-la
(MIP-lu)
though
is no
binding
there
published
and
RANTES
data
for
(al-
tivity
nd more recently to bind monocyte chemoattractant protein-3 (MCP-3) (B. Dougherty, unpublished observations). which was cloned from Monomac6 to bind MCP-1 and MCP-3, and
CC-CKR-2, been shown to be important
in monocyte
a third
receptor,
monocyte nophils
library has been and gives calcium
extravasation
CC-CKR-3, found fluxes
from
One
specificity gross differences ligands.
The
MIP-13, and RANTES [7]. No binding data has been shown to date for the recombinant receptor. Ourwork has identified a fourth CC-chemokine receptor in
three-dimensional
try-related anti-parallel 3). We have recently of the CC chemokine [13]. The overallfold CXC
C-terminal
helices
1 over35#{243}amino
ever,
topologies
360
shows
high
levels
of expression
in the thymus
and in peripheral blood leukocytes. Using FACS sorted cell populations we could also show that mRNA for CC-CKR-4 specifically
was
expressed
in T-cells,
cytes, as well as in platelets. Human detectable CC-CKR-4 expression. tion
for 15 mm
up-regulation CC-CKR-4
with
IL-5
(10
of receptor were initially
B-cells,
basophils However,
ng/ml)
there
and
mono-
ologically supported variant
a significant
mRNA expression. The ligands for determined to be MCP-1, MIP-la,
and RANTES from measurements ofCa2+activated currents in Xenopus laevLc oocytes injected with CC-CKR-4. confirmed lines(C.
The results by binding
of the genome. cluster
me
for MIP-la experiments
A. PowerandA.J.
The two groups
to chromosome receptors cluster
quence
comparisons,
unpublished
are located
four human
CC chemokine
the Duffy
antigen
that
binding or vice
or erythrocyte
and versa.
the
binding.
In our
initial
CC chemokine
suchas
receptor,
to understand
(Fig. CC
which this on the
we therefore studied chemokine, IL-8, and a
RANTES.
has
been
used
in these beta sheet
experiments, and with two symme-
are shorter
the low level although the
for the CC chemokines.
ofboth
RANTES
and
How-
MIP-1f
[14] are
relevant
not be important, since that IL-8 is monomeric
concentrations
[16,
by experiments with of IL-8, N-Methyl-Leu-25
17].
a chemically IL-8. This
functional selectivity
there is at physi-
This
is further
synthesised is unable
to
in a variety of cell- based is more likely, therefore, to
be due to the distribution of hydrophobic monomer surface. A theoretical analysis
sequences on the of hydrophobicity
shows that the distribution of hydrophobic regions is preserved in CC or CXC chemokines, providing some explanationforthelack of receptorcross-binding [19]. We therefore examined the sequences of CC and CXC for
regions
involved
One region that stood leucine, corresponding invariably 1). This
in
out contains to leucine
replaced is at the
receptor
selectivity.
a highly 25 in IL-8,
by a tyrosine dimer interface
conserved which
in CC chemokines in IL-8, but
is its
equivalent residue in CC chemokines is far away from the interface. To investigate the role of this residue in controlling cell and receptor specificity, we have made the mutant Leu25-*Tyr in IL-8 [10]. The resultant protein shows a 100-fold reduction in binding affinity to neutrophils, and this reduction is also seen in binding studies using recombinant IL-8R-A and -B expressed in HEK293 cells. The mutant calcium the
is also less mobilization
Leu25-Tyr
active in neutrophil assays. However,
mutation
introduces
chemotaxis and most importantly, monocyte
chemotac-
tic activity into IL-8, which is not present in the wild type protein (Fig. 4A). In addition, the mutant protein will compete with [‘25I]MIP-lcx for recombinant CC-CKR-1
STRUCTURE AND FUNCTION OF CHEMOKINES: MAKING CXC CHEMOKINES BIND TO A CC-CHEMOKINE RECEPTOR One ofthe initial chemokine-receptor
of IL-8
alpha helices over the top (see Fig. solved the three- dimensional structure RANTES by use of NMR spectroscopy ofthe monomers ofCC chemokines and
dimerise, but is fully assays [18]. Receptor
almost (Fig.
chemokine
identity of human
[9]. So far, however, a signaling response
attempts
molecular basis of receptor selectivity, the difference between a typical CXC
CC
structure
structure may data showing
chemokines all
signaling using CXC The exception to this is
chemokine
binds both CC and CXC chemokines receptor has not been shown to induce ligand
regions
receptors
3p2l-22, whereas the CXC chemokon chromosome 2q34-35. From seit is clear
receptors as ligands
been cell
results).
on different
receptors cluster together in terms ofsequence 2). However, there have been no reports chemokine chemokines
chloride cRNA for
and RANTES have using transfected
Hoogewerf,
ofreceptors The
quaternary experimental
showed barely after stimulawas
of
completely different from those ofthe CXC chemokines [11, 15]. For the CC chemokines, the dimer interface is formed by the N-terminal n-sheet of the protein (3O), rather than the second beta strand (termed p1). However, this difference in
can be clearly seen Northern blot analysis
of CC-CKR-4
dimer
their
is conferred by structure of the
chemokines appears to be similar, despite of sequence identity between the two proteins,
acids; 46% identity to the CC-CKR-2 (form amino acids; and 45% identity with CC-CKR-3
with
the molecularbasis
nuclearmagnetic resonance (NMR) and X-ray methods [1 1, 12]. The protein is dimeric at
the human basophilic cell line KU-812 [8]. We have called this K5.5,orCC-CKR-4 based on the otherrecently identified sequences. This receptor shows 49% identity with CC-CKRb) over
receptors
to explain
the millimolar concentrations forms a six-stranded antiparallel
activated
to be expressed on eosiin response to MIP-la,
over 356 amino acids. The similarities from the alignment in Figure 2 [8].
of chemokine
hypothesis
is that the receptor selectivity in the three-dimensional
solved by both crystallographic
Recently,
an
interaction
[10].
this
cells, has is presumed
[4-6].
cloned
of the
ligands
RANTES),
(RANTES/MIP-la with an affinity (Fig. 4B). To study
questions that we asked when we examined interactions was what controls the selec-
this
that
receptor) is only
effect
on selectivity
the Leu25-Cys mutation. to leucine so only a small
Wells
et al.
Selectivity
and
expressed 12-fold less
Antagonism
in HEK293 than MIP-lu
further,
we have
Cysteine is very similar change in the biological
of Chemokine
Receptors
cells, itself made in size activity
55
50 IL-8R-B
MESDSFEDFW
KGEDLSNYSY
IL-8R-A
MSNITDPQMW
DFDDLN.
CC-CKR-3
MTTS
CC-CKR-1
METP
CC-CKR-2
cc
-
-
NTTE.DYD IRNTNESGEE
MNPTDI
4
...
LDTVE.TFG
MLSTSRSRF
CKR
SSTLPPFLLD
ADTTLDES
IY
AAPCEPESLE
INKYFVVIIY
FTGMPPADED
YSPMLETET
LNKYVVIIAY
TTSYYD
LLCEK.ADTRA
LMAQFVPPLY
T1’TEFDYGDA
TPCQKVNERA
FGAQLLPPLY
V’VFFFDYDYG
APCHKFDVKQ
IGAQLLPPLY
SNYYLYESIP
KPCTKEGIKA
FGELFLPPLY
. DVG
151
200
IL-8R-B
LTQKRY.LVX
PICLSIWGLS
LLLALPVLLP
RRWYSSNVS
IL-8R-A
LTQKRH.LVX
FVCLGCWGLS
)OILSLPFFL?
RQAYHPNNSS
PVCYEVLGND
LRARTVTFGV
ITSIVTWGLA
VLAALPIFIF
YETEELFEET
LCSALYPEDT
CC-CKR-3 CC-CKR-1
cc cc
PACYEDMGNN
LRARTVTFGV
ITSIIDIALA
ILASMPGLYF
SKTQWEFTHH
TCSLHFPHES
-
CKR
-
2
LKAR’IVTFGV
VTSVITWLVA
VFASVPGIIF
TKCQKEDSVY
VCGPYFP
-
CKR
-
4
LRARTLTYGV
ITSLATWSVA TMIV
VFASLPGFLF
STCYTERNHT
YCKTKYSLNS
C
...
201
250
IL-8R-B
TANWRIILLRI
LPQSFGFIVP
LLICJFCYGF
TLRTLFKA}iM
GQK.HRAMRV
IL-8R-A
TAKWRXVLRI
LPHTFGFIVP
LFVMLFCYGF
TLRTLFKAHM
GQK.HRAMRV
CC-CKR-3
VYSWRHFHTL
RMTIFCLVLP
LLVMAICYTG
IIKTLLRCPS
KKK.YKAIRL
CC-CKR-1
LREWXLFQAL
KLNLFGLVLP
LLVMIICYTG
IIKILLRRPN
EKK.SKAVRL
. RGWNNFHTI
MRNILGLVLP
LLIMVICYSG
ILKTLLRCRN
EKKRHRAVRV
T.TWXVLSSL
EINILGLVIP
LGIMLFCYSM
IIRTLQHCKN
EKK.NKAVKJI
cc
-
CKR
-
2
CC-CKR-4
I’MV
300
251 IL-8R-B
IFAVVLIFLL
CWLPYNLVLL
ADTLMRTQVI
QETCERRNHI
DRALDATIIL
IL-8R-A
IFAVVLIFLL
QETCERRNNI
GRALDATEIL
IFVIMAVFFI
CWLPYNLVLL FWTPYNVAIL
ADTLMRTQVI
CC-CKR-3
LSSYQSILFG
ND.CERSKHL
DLVMLVTIVI
CC-CKR-1
IFVIMIIFFL
FWTPYNLTIL
ISVFQDFLFT
HE.CEQSRHL
DLAVQVTIVI
CC-CKR-2
IFTIMIVYFL
FWTPYNIVIL
LNTFQEFFGL
SN.CESTSQL
DQATQVTITL
CC-CKR-4
IFAVVVLFLG
FWTPYNIVLF
LITLVELEVL
QD.CTFERYL
DYAIQATKTL
TMVI