| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
triggers CXCR4 receptor dimerization and activates the JAK/STAT pathway
Department of Immunology and Oncology, Centro Nacional de Biotecnología, CSIC-Universidad Autónoma de Madrid, Campus de Cantoblanco, E-28049 Madrid, Spain
1Correspondence: Department of Immunology and Oncology, Centro Nacional de Biotecnología, CSIC, Universidad Autónoma de Madrid, Campus de Cantoblanco, E-28049 Madrid, Spain. E-mail cmartineza{at}cnb.uam.es
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
), the ligand for the
CXCR4 receptor, induces a wide variety of effects that include calcium
mobilization, chemotactic responses, bone marrow myelopoiesis, neuronal
patterning, and prevention of HIV-1 infection. Nonetheless, little is
known of the biochemical pathways required to achieve this variety of
responses triggered after receptor–chemokine interaction. We developed
a set of monoclonal antibodies that specifically recognize the CXCR4
receptor and used them to identify the signaling pathway activated
after SDF-1 binding in human T cell lines. Here we demonstrate that
SDF-1 activation promotes the physical association of
Gi with the CXCR4. Furthermore, within seconds of
SDF-1 activation, the CXCR4 receptor becomes tyrosine phosphorylated
through the activation and association with the receptor of JAK2 and
JAK3 kinases. After SDF-1 binding, JAK2 and JAK3 associate with
CXCR4 and are activated, probably by transphosphorylation, in a
Gi-independent manner. This activation enables the
recruitment and tyrosine phosphorylation of several members of the STAT
family of transcription factors. Finally, we have also observed
SDF-1-induced activation and association of the tyrosine phosphatase
Shp1 with the CXCR4 in a Gi-dependent manner. As occurs
with the cytokine receptors in response to cytokines, the CXCR4
undergoes receptor dimerization after SDF-1 binding and is a
critical step in triggering biological responses. We present compelling
evidence that the chemokines signal through mechanisms similar to those
activated by cytokines.—Vila-Coro, A. J.,
Rodríguez-Frade, J. M., Martín de Ana, A.,
Moreno-Ortíz, M. C., Martínez-A., C., Mellado, M.
The chemokine SDF-1 triggers CXCR4 receptor dimerization and
activates the JAK/STAT pathway.
Key Words: chemotactic response • tyrosine phosphorylation • Shp1 • G-proteins
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Chemokines have been classified into two
main categories, the CXC and the CC chemokines, according to the
position of their first two conserved cysteine residues. CXC
chemokines, including interleukin 8
(IL-8)2
, MGSA/gro, IP-10, Mig, and stromal cell-derived factor 1 (SDF-1),
have been classically described as neutrophil chemoattractants, whereas
CC chemokines, comprising monocyte chemoattractant protein-1 (MCP-1),
-2, -3, -4, macrophage-inflammatory protein-1 (MIP-1), MIP-1ß,
RANTES (regulated on activation, normal T expressed and excreted),
I-309, TARC, LARC, and others, mainly attract monocytes, eosinophils,
and basophils; members of both families have chemotactic activity on
different lymphocyte lineages (1
, 2)
. Two chemokines that
fall into neither category have recently been cloned. Lymphotactin is
the only member of the C subfamily (3)
, whereas
fractalkine is a CX3C chemokine with a long
mucin-like stalk and a transmembrane region (4)
.
Chemokine receptors are members of the seven-transmembrane domain
G-protein-coupled receptors, which have also been divided in two main
subfamilies (CXCR and CCR), depending on their chemokine specificity
(1
, 2
, 5)
. Chemokine receptor expression is heterogeneous
among different cells of the leukocyte lineage and is regulated at the
transcriptional level (6)
. Chemokines have also been the
focus of attention since they act as receptors for different HIV-1
strains. CCR5 has been shown to function as a receptor for
macrophage-tropic HIV-1 strains (7
, 8)
, whereas CXCR4 acts
as a receptor for T cell-tropic HIV-1 strains (9
, 10)
and
HIV-2 (11)
; HIV-1 strains have now been redefined as R5
and X4, depending on their ability to interact with CCR5 or CXCR4,
respectively.
Although most chemokine receptors bind several chemokines, CXCR1 and
CXCR4 are specific for only one physiological ligand. CXCR1 interacts
with IL-8, whereas the CXC-chemokine SDF-1 is the only ligand
identified so far for CXCR4 (1
, 2)
. This chemokine, cloned
by a gene trapping strategy, was first isolated from stromal cell
culture supernatants (12)
. Its chemotactic properties on
peripheral blood lymphocytes (13)
,
CD34+ progenitor cells (14)
, and
pre- and pro-B cell lines (15)
have been described.
Knockout mice lacking the SDF-1 protein show abnormalities in B cell
lymphopoiesis, bone marrow myelopoiesis, lack of blood vessel formation
in the gut, and severe ventricular septal defects (16)
. A
similar profile of deficiencies is found in CXCR4-deficient mice
(17
, 18)
.
The activation signals after SDF-1 stimulation are not well defined,
and various signal transduction pathways have been implicated through
the association of CXCR4 receptors with guanine nucleotide binding
protein (G-proteins) (19)
. For example, stimulation of
CXCR4 transfectants by SDF-1 results in increased phosphorylation of
focal adhesion components, including the related adhesion focal
tyrosine kinase (RAFTK/Pyk2), Crk, and paxillin. SDF-1-induced
activation of the p44/42 MAP kinases (Erk 1 and 2), PI3 kinase, and
NF-
B (20)
has also been described, implicating this
chemokine in modulation of signaling molecules and transcription
factors that mediate changes in the cytoskeletal apparatus and also
regulate cell growth.
Activation of some STAT transcription factor family members has been
described in T cells after RANTES or MIP-1 stimulation
(21)
. We recently demonstrated that MCP-1 and RANTES
trigger tyrosine phosphorylation of their receptors and activation of
the JAK/STAT pathway in a PTX-independent manner (22
, 23)
.
This JAK activation plays a critical downstream role for all other
signaling events, including the association between the receptor and
the Gi proteins (22)
. Here we
show that, as occurs with the cytokine receptors, CXCR4 undergoes
receptor dimerization after SDF-1 binding. We then analyze SDF-1
activation of the JAK/STAT pathway, demonstrating that both JAK3 and
JAK2 are activated and associated with the CXCR4. This activation
enables the recruitment and tyrosine phosphorylation of several members
of the STAT family of transcription factors. The phosphatase Shp1 is
also activated, translocated to the membrane, and associated with the
receptor as a consequence of SDF-1 activation. All together, these
data demonstrate that the activation of the JAK/STAT pathway is a
general signaling pathway for both CC and CXC chemokines and implicate
the Shp phosphate family in the shutdown responses triggered by
chemokines. Last, we present compelling evidence that the
chemoattractant cytokines (chemokines) as well as the cytokines signal
via mechanisms shared by these two protein families.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, -Gs,
anti-STAT1 to -STAT6, and anti-Pol II (Santa Cruz Biotechnology Inc.,
Santa Cruz, Calif.), and anti-ß-actin (Sigma, St. Louis, Mo.).
Preparation of monoclonal antibodies (mAb)
Peptides of CXCR4 receptor amino acid sequences 25–35 and
20–37 were synthesized, coupled to keyhole limpet hemocyanin
(Calbiochem, La Jolla, Calif.), and used as immunogens as described
(24)
. Spleen and/or lymph node cells from immunized mice
were fused with the P3X63-Ag8.653 myeloma cell line (ATCC CRL 1580)
following standard protocols. Supernatants were tested for antibodies
in enzyme-linked immunoassay (ELISA) and positive hybridomas cloned by
limiting dilution. Isotypes were determined in ELISA using PO-labeled,
subclass-specific anti-mouse immunoglobulin antibody (Southern
Biotechnologies Associates, Birmingham, Ala.). Monoclonal antibodies
were purified by ammonium sulfate precipitation (25)
and
chromatography on Sephacryl S-200 (Pharmacia Biotech, Uppsala, Sweden).
Antibody capture ELISA
ELISA was performed as described previously (26)
;
briefly, synthetic peptides [3 µg/ml in phosphate-buffered saline
(PBS), 100 µl/well] were adsorbed to microtiter plates (Maxi-Sorb,
Nunc, Denmark) after blocking, mAb were incubated, and the reaction was
developed using a PO-labeled goat anti-mouse immunoglobulin antibody
(GAM-PO; Tago Inc., Burlingame, Calif.) and OPD (Sigma).
Flow cytometry analysis
Cells were centrifuged (250 x g, 10 min, room
temperature), plated in V-bottom 96-well plates (2.5 x
105 cells/well), and incubated with a
biotin-labeled mAb (1 µg/50 µl/well for 60 min, 4°C) that had
been preincubated with the immunizing peptide CXCR4
(25
26
27
28
29
30
31
32
33
34
35)
, with the CCR2 (24
25
26
27
28
29
30
31
32
33
34
35
36
37
38)
control
peptide (10 µg/well, 60 min, 4°C), or with PBS. Cells were washed
twice in PBS with 2% bovine serum albumin (BSA) and 2% fetal calf
serum (FCS) and centrifuged (250 x g, 5 min, 4°C).
Fluorescein isothiocyanate-labeled streptavidin (Southern
Biotechnologies) was added, incubated (30 min, 4°C), and plates were
washed twice. Cell-bound fluorescence was determined in a Profile XL
flow cytometer at 525 nm (Coulter Electronics, Miami, Fla.).
Calcium determination
Changes in intracellular calcium concentration were monitored
using the fluorescent probe Fluo-3AM (Calbiochem). Cells (2.5 x
106 cells/ml), untreated or treated for 16 h
with cholera toxin (CTX; 0.4 µg/ml, Sigma), pertussis toxin (PTX; 0.1
µg/ml, Sigma), AG490, or tyrphostin A1 (both at 25 µM;
Calbiochem-Novabiochem, La Jolla, Calif.), were resuspended in RPMI
containing 10% FCS and 10 mM HEPES and incubated with Fluo-3AM (300
µM in DMSO, 10 µl/106 cells) for 15 min at
37°C. After incubation, cells were washed and resuspended in complete
medium containing 2 mM CaCl2 and maintained at
4°C until just before SDF-1 addition to minimize membrane
trafficking and eliminate spontaneous Ca2+ entry.
Calcium mobilization in response to 10 nM SDF-1 (PeproTech) was
determined at 37°C in an EPICS XL flow cytometer at 525 nm, and
includes background level stabilization and determination of the level
of probe loaded for each sample. Only samples with similar load, as
assessed by determination of Ca2+ mobilization
induced by a ionophore (ionomycin, Sigma, 5 µg/ml), are considered
acceptable.
Cell migration
MOLT4 cells, untreated or treated for 16 h with CTX, PTX,
AG490, or tyrphostin A1, as described above, were placed (0.25 x
106 cells in 0.1 ml) in the upper well of 24-well
transmigration chambers (5 µM pore size; Transwell; Costar Corp.,
Cambridge, Mass.), and 10 nM SDF-1 (in 0.6 ml RPMI containing 0.25%
BSA) was added to the lower well. Plates were incubated (120 min,
37°C) and the cells that had migrated to the lower chamber were
counted as described (27)
.
Immunoprecipitation, sodium dodecyl sulfate-polyacrylamide gel
electrophoresis (SDS-PAGE), and Western blot analysis
SDF-1-stimulated cells (20 x 106)
were lysed in a detergent buffer (20 mM triethanolamine pH 8.0, 300 mM
NaCl, 2 mM EDTA, 20% glycerol, 1% digitonin with 10 µM sodium
orthovanadate, 10 µg/ml leupeptin, and 10 µg/ml aprotinin) for 30
min at 4°C with continuous rocking, then centrifuged (15,000 x
g, 15 min). The amount of protein present in the cell
lysates was controlled before immunoprecipitation by using a protein
detection kit (Pierce, Rockford, Ill.). Immunoprecipitation was
performed essentially as described (22)
. Protein extracts
precleared by incubation with 20 µg of anti-mouse IgM-agarose (Sigma)
or protein A-Sepharose (60 min, 4°C) and centrifuged (15,000 x
g, 1 min) were immunoprecipitated with the appropriate
antibody (5 µg/sample, 120 min, 4°C), followed by anti-mouse
IgM-agarose or protein A-Sepharose if the first antibody was derived
from rabbit serum. Immunoprecipitates or protein extracts were
separated in SDS-PAGE and transferred to nitrocellulose membranes.
Western blot analysis was performed as described (22)
,
using 2% BSA in TBS as blocking agent for the anti-phosphotyrosine
analyses. When stripping was required, membranes were incubated (60
min, 60°C) with 62.5 mM Tris-HCl pH 7.8, containing 2% SDS, and
0.5% ß-mercaptoethanol. After washing with 0.1% Tween 20 in TBS for
2 h, membranes were reblocked, reprobed with the appropriate
antibody, and developed as above. In all cases, protein loading was
controlled by using a protein detection kit (Pierce) and, when
necessary, by reprobing the membrane with the immunoprecipitating
antibody.
Cross-linking
Serum-starved MOLT4 cells (15 x 106)
were either unstimulated or stimulated with 10 nM SDF-1 or RANTES
for 1 min at 37°C. The reaction was terminated by adding 1 ml of cold
PBS, then centrifuged (150 x g, 10 min). After washing
twice with cold PBS, 10 µl of 100 mM DSS (Pierce) was added for 10
min at 4°C, with continuous rocking. The reaction was terminated by
addition of 1 ml of cold PBS and washed three times. The pellet was
lysed for 60 min and immunoprecipitated as described above.
Nuclear extract preparation
Nuclear extracts were prepared from SDF-1-stimulated MOLT 4
cells (10 x 106). Where indicated,
pervanadate was prepared by incubating equal volumes of 12 mM
Na3VO4 (freshly dissolved
in water, Sigma) and 12 mM
H2O2 at room temperature
for 20 min, then added to cells during serum starvation (60 min, final
concentration of 60 µM). Briefly, cells were washed with ice-cold
PBS, resuspended in 1 ml of buffer A (50 mM NaCl, 0.5 M sucrose),
incubated 2 min at 4°C and pelleted (4500 x g, 3
min, 4°C). They were then resuspended in 1 ml of buffer B (50 mM
NaCl, 25% glycerol), pelleted (4500 x g, 3 min,
4°C), and incubated in 60 µl of buffer C (350 mM NaCl, 25%
glycerol) for 30 min at 4°C with continuous rocking. After
centrifugation (20,000 x g, 20 min, 4°C),
supernatants containing nuclear extracts were aliquoted and stored at
-80°C. All buffers contained 0.5 mM spermidine, 0.15 mM spermine,
0.1 mM EDTA, 10 mM HEPES pH 8, 0.5 mM PMSF, 2 µg/ml leupeptin, 3
µg/ml pepstatin, 0.2 IU/ml aprotinin, 1.75 mM ß-mercaptoethanol, 1
mM orthovanadate, and 10 mM NaF.
For Western blot analysis, 20 µg of nuclear extracts for each condition, prepared as before, were separated in SDS-PAGE and transferred to nitrocellulose membranes. Western blot analysis was performed as above and developed by using anti-STAT3 and anti-STAT5b antibodies (Santa Cruz).
Electrophoretic mobility shift assay (EMSA)
Nuclear extracts (10 µg) from untreated or chemokine-treated
cells were analyzed in binding reactions. Briefly, extracts were
incubated with 0.5 ng of 32P-end-labeled,
double-stranded oligodeoxynucleotides containing the sis-inducible
element (SIE) of the c-fos promoter sequence
[GGGGTGCATTTCCCGTAAATCTTGTCT] (wild type; wt-SIE); where indicated, a
mutant version of this sequence was used that is unable to bind STAT
proteins [GGGGTGCA TCCACCGTAAATCTTGTCT] (mut-SIE). The binding
reaction was carried out in EMSA buffer for 30 min at room temperature
(final volume 10 µl) with 1.5 µg of poly(dI-dC) and, where
indicated, a 20-fold molar excess of unlabeled SIE or nonspecific
oligonucleotide competitor. Protein–DNA complexes were resolved on a
4.5% polyacrylamide gel using 0.5x Tris borate-EDTA as running
buffer. EMSA buffer contained 10 mM Tris pH 7.5, 50 mM NaCl, 1 mM
dithiothreitol (DTT), 1 nM EDTA, and 5% glycerol. The amount of
protein present in the nuclear extracts was controlled before Western
blot or EMSA by using a protein detection kit (Pierce).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
induces functional responses in MOLT4 cells
A) and in Western blot analysis (Fig. 1B
). In both
flow cytometry and Western blot, specific recognition could be aborted
by preincubation of the mAb with the immunizing peptide, but not when a
peptide corresponding to the amino-terminal domain of the CCR2 (amino
acids 24–38) was used (Fig. 1A, B
).
|
In response to SDF-1, the MOLT4 cell line mobilizes calcium, which
is desensitized in response to a second stimulation (Fig. 2
A), and also triggers cell migration (Fig. 2B
).
PTX treatment abrogates both calcium release (Fig. 2A
) and
cell migration responses (Fig. 2B
), whereas no effect was
observed after incubation with CTX (not shown). This is consistent with
other studies showing that SDF-1 downstream signals are coupled to
PTX-sensitive G-proteins (28)
.
|
Using the CXCR4-specific mAb CXCR4–01, we analyzed the physical
association between the CXCR4 receptor and the
Gi subunit of the heterotrimeric
Gi protein in response to ligand binding. A CXCR4
receptor-associated Gi protein was
immunoprecipitated and detected in Western blot using an
anti-Gi-specific antibody (Fig. 2C
). This association is initiated within 60 s of
SDF-1 triggering and persists for 5 min after chemokine binding.
When the same assay was performed using
anti-Gs-specific antibody, however, no
association was found, indicating the specificity of association of the
Gi with the CXCR4 receptor. We therefore
conclude that the CXCR4 receptor regulates calcium release and couples
to the Gi protein in MOLT4 cells after
SDF-1 binding.
SDF-1 induces tyrosine phosphorylation of the CXCR4 receptor
When SDF-1-activated MOLT4 cell lysates were immunoprecipitated
with anti-PTyr and Western blots developed with anti-CXCR4 receptor
antibodies (Fig. 3
A, upper panel), or precipitated with anti-CXCR4 and
developed with anti-PTyr antibody (Fig. 3A
, lower panel),
the same phosphorylated 40 kDa band was observed in both cases,
confirming tyrosine phosphorylation of the CXCR4 on SDF-1
stimulation. No differences were observed in the amount of CXCR4 in
each lane (Fig. 3A
, lower panel) when the same membrane was
stripped and reblotted with anti-CXCR4 antibody. An increase in CXCR4
receptor phosphorylation is seen as early as 60 s after SDF-1
stimulation; phosphorylation persists for 5 min, decreasing thereafter.
We have observed residual phosphorylation of CXCR4 in untreated cells
(Fig. 3A
), the significance of which is discussed below.
|
SDF-1 induces rapid JAK activation and association with the
CXCR4 in a PTX-independent manner
As we described previously, the JANUS kinase family member
JAK2 is responsible for CCR2 receptor phosphorylation
(22)
. To identify the kinase responsible for the rapid Tyr
phosphorylation of the CXCR4 chemokine receptor, SDF-1-stimulated
MOLT4 cell lysates were immunoprecipitated with anti-CXCR4 or with
anti-ß2-microglobulin as an isotype-matched
antibody control. Anti-JAK2 antibodies identified a 130 kDa protein in
the anti-CXCR4 immunoprecipitate (Fig. 3B
). Whereas only
JAK2 associates with the CCR2 in MCP-1-stimulated Mono Mac 1 cells,
when MOLT4 cells were activated by SDF-1 not only JAK2, but also
JAK3, associated with CXCR4. JAK association with the CXCR4 receptor
occurs as early as 60 s after SDF-1 stimulation (Fig. 3B
); small amounts of both JAKs also associated with the
CXCR4 receptor in the absence of exogenous SDF-1, consistent with
the lower level of receptor phosphorylation in the absence of exogenous
ligand in MOLT4 cells. Results of immunoprecipitation of cell lysates
with isotype-matched control antibodies (Fig. 3B) rule out
nonspecific protein association with membrane components under these
experimental conditions. The rapid association of both of these JAKs
with the CXCR4 receptor suggests a role for this tyrosine kinase in
early receptor phosphorylation after ligand stimulation. Although JAK1
is present in MOLT4 cells, it was not found to be associated with the
CXCR4 receptor after chemokine activation (Fig. 3B
). JAK2
and JAK3 association correlates with their tyrosine phosphorylation
after SDF-1 stimulation (Fig. 3C
), again demonstrating
JAK pathway activation in SDF-1-mediated signaling.
To identify downstream events related to the JAK pathway, we analyzed
JAK kinase-activated STAT transcription factors in anti-CXCR4
immunoprecipitates, since STAT activation and association with CCR2 and
CCR5 has been described previously (21
22
23)
. STAT1, 2, 3,
and 5, but not STAT4 or 6, associate with the receptor complex in
response to SDF-1 (Fig. 4
A). The association of each STAT correlates in time with JAK
phosphorylation and binding. Furthermore, these STATs are tyrosine
phosphorylated, indicating their activation (Fig. 4B
). The
early SDF-1-triggered tyrosine phosphorylation of the CXCR4 receptor
is independent of PTX-sensitive G-proteins. When cells are PTX treated,
the CXCR4 receptor is continuously phosphorylated, as detected by
immunoprecipitation with either anti-PTyr or anti-CCR4 antibodies
(Fig. 5
A). Under the same conditions, both JAK2 and JAK3 associate
with SDF-1-activated CXCR4, much as occurs when cells are not PTX
treated (Fig. 5B
). Nevertheless, neither JAK2 nor JAK3
dissociate the receptor in PTX-treated cells (Fig. 5B
).
These data clearly resemble the results obtained after CCR2 receptor
stimulation by MCP-1, where PTX treatment blocks JAK kinase
dissociation from the receptor (22)
. This implies that
PTX-sensitive G-protein activation is important in the recycling of the
JAK/STAT receptor complex and assigns a potential role to
Gi pathway-related components in the control of
this recycling.
|
|
Despite the rapid and transient JAK-mediated CXCR4 tyrosine
phosphorylation, phosphorylated CXCR4 can be observed long after
maximum JAK association. This implies that other tyrosine kinases may
be implicated in Tyr phosphorylation of the chemokine receptors at
later times after chemokine binding; in fact, RANTES activation of
human T cells promotes the activation of p125FAK
and ZAP-70 (29)
.
To determine the functional significance of JAK/STAT pathway activation
by SDF-1, MOLT4 cells were treated overnight with the specific JANUS
kinase inhibitor AG490 (25 µM) (30)
prior to
determination of SDF-1-mediated Ca2+
mobilization and cell migration. This treatment completely blocks the
SDF-1-mediated CXCR4 effects (Fig. 6
A, B), as compared with cells treated with tyrphostin A1
(Calbiochem-Novabiochem), indicating a functional role for
ligand-induced JAK activation upstream of G-protein-related pathways.
Tyrphostin treatment was not toxic to MOLT4 cells, as shown by cell
cycle analysis using propidium iodide incorporation (not shown).
|
SDF-1-induced STAT translocation to the nucleus is blocked by
phosphotyrosine phosphatase activity
To further assess the role of the JAK/STAT pathway in chemokine
signaling, we analyzed the nuclear translocation of activated STAT3 and
STAT5 and their capacity to bind specific DNA sequences after chemokine
activation. Nuclear extracts of SDF-1-stimulated MOLT4 cells were
analyzed in Western blot using anti-STAT3 and anti-STAT5b antibodies,
and no nuclear translocation was observed at any of the time points
analyzed (Fig. 7
A). As expected, no SIE binding activity was detected,
measured by EMSA after chemokine treatment (Fig. 7B
).
Protein levels in the nuclear extracts were controlled using a protein
detection kit. Nuclear extracts were controlled in Western blot using
anti-Pol II antibody as nuclear marker and were free of cytoplasmic
contamination, confirmed by developing with an anti-ß-actin
cytoplasmic marker antibody (not shown).
|
In a growth hormone-stimulated female rat hepatocyte model, it
has recently been shown that a possible cause of the lack of STAT
translocation to the nucleus is the presence of enhanced
phosphotyrosine phosphatase activity. To rule out this possibility, we
tested the effect of the phosphotyrosine phosphatase inhibitor
pervanadate (31)
on SDF-1-stimulated MOLT4 cells.
SDF-1-mediated STAT3 and STAT5 translocation to the nucleus was
observed after pervanadate treatment, with a maximum effect at 30 min
after stimulation (Fig. 7A
); the same result was obtained
when SIE binding activity was evaluated (Fig. 7B
). Treatment
with pervanadate alone slightly increased basal STAT3 and STAT5
translocation levels (Fig. 7A, B
). This down-regulation of
STAT3 and STAT5 nuclear translocation and activation suggests that
phosphotyrosine phosphatase activity is involved in CXCR4/JAK/STAT
pathway regulation.
SDF-1 induces Shp1 activation and association with the CXCR4 in
a PTX-dependent manner
Protein phosphatases are critical in the regulation of
signal transduction pathways mediated by tyrosine phosphorylation
(32)
. Using SDF-1-activated MOLT4 cell lysates, we show
that Shp1, a cytoplasmic tyrosine phosphatase with two SH2 domains, is
activated and associated with CXCR4. We demonstrate physical
association between the CXCR4 receptor and the Shp1 phosphatase by
immunoprecipitating with the CXCR4–01 mAb and developing Western blots
with anti-Shp1 antibodies (Fig. 8
). This association is initiated within 5 min of SDF-1-triggering and
reaches a maximum at 15 min after chemokine binding.
Immunoprecipitation of cell lysates with isotype-matched control
antibodies to ß2-microglobulin or to other
membrane proteins did not reveal the presence of Shp1 (Fig. 8)
, ruling
out nonspecific protein association with membrane components under
these experimental conditions. SDF-1-induced association of Shp1 to
the CXCR4 receptor is blocked by PTX treatment, indicating that Shp1
association requires Gi activation and is
probably implicated in the control of JAK2/JAK3 activation (Fig. 8)
.
|
SDF-1 induces CXCR4 dimerization
Similar to growth factor-induced dimerization of tyrosine kinase
receptors, CC chemokine receptors undergo dimerization in response to
their ligands (23
, 33)
. We then tested whether SDF-1, a
CXC chemokine, triggers CXCR4 dimerization. DSS-mediated cross-linking
in MOLT4 cells was performed after SDF-1 and RANTES stimulation. In
CXCR4 immunoprecipitates of chemokine-stimulated cell lysates, we
observed a high molecular weight receptor species only when cells were
stimulated with SDF-1 (Fig. 9
). This band corresponds to the expected molecular weight of two CXCR4
molecules, as assessed by immunoprecipitation with anti-CXCR4
antibodies and Western blot developed with anti-CXCR4 antibodies. CXCR4
does not appear in RANTES-stimulated, cross-linked cell lysates
immunoprecipitated with the same anti-CXCR4 antibody (Fig. 9)
. These
results indicate that the CXC chemokine SDF-1, like the CC
chemokines RANTES and MCP-1 and similar to some members of the large
cytokine family, can trigger receptor dimerization, followed by
recruitment of tyrosine kinases that phosphorylate both the receptor
and the STAT transcription factors.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
helices. These
proteins have an extracellular amino-terminal portion and three
extracellular loops implicated in receptor–agonist interaction, as
well as a cytoplasmic carboxyl-terminal domain and three intracellular
loops that interact cooperatively to bind and activate G-proteins
(34)
. Most G-protein-coupled receptors have single
conserved cysteine residues in the extracellular loops, believed to
form a disulfide bond that stabilizes the functional protein structure.
Despite growing interest in and information on the chemokines and their
receptors, the complex rules that govern chemokine receptor triggering
remain unclear. In vivo, this process probably involves a
dynamic relationship including chemokine, receptor and the
receptor-linked cytoplasmic G-protein complex, although many of the
molecular events after chemokine receptor triggering remain to be
defined.
SDF-1 was first isolated from stromal cell culture supernatants. It
is a potent chemoattractant for several leukocyte populations and binds
specifically to the ubiquitous CXC receptor CXCR4. In CXCR4-expressing
cells, SDF-1 binding promotes Ca2+
mobilization and cell transmigration, processes that are blocked by PTX
treatment but not by CTX.
Here we describe the characterization of mAb specific for the CXCR4
chemokine receptor and their use in the identification of signaling
pathways activated after SDF-1 binding to the receptor in the MOLT4
human T cell line. Considerable data are available on SDF-1
activity, including its role in HIV-1 pathogenesis and its function in
the development of the immune, circulatory, and central nervous systems
as defined in SDF-1 and CXCR4 knockout mice. Nonetheless, little is
known about the signals required for the regulation of such a wide
array of activities. Several observations indicate the triggering of
multiple signaling pathways through chemokine receptors, including
G-protein-mediated pathways, MAP kinases, PI3K, tyrosine, and
serine/threonine kinases. CXCR4-mediated responses are PTX sensitive,
indicating a role for Gi in this signaling. Here
we demonstrate that after SDF-1 binding, Gi,
but not Gs associates rapidly (within 60 s)
to CXCR4. This confirms previous data indicating that other chemokine
receptors such as IL-8 receptors or CCR2, bind
Gi in response to their respective ligands,
IL-8 (35)
and MCP-1 (22)
.
It has recently been shown that RANTES induces activation and assembly
of macromolecular focal adhesion complexes and provokes T lymphocyte
homotypic adhesion by phosphorylation and association of
p125FAK and ZAP-70 (29)
. Stimulation
with SDF-1 also promotes increased phosphorylation of focal adhesion
components, including the p125FAK-related
adhesion focal tyrosine kinase (RAFTK/Pyk2), Crk, and paxillin
(36)
. Furthermore, we have recently shown the tyrosine
phosphorylation of two members of the CC receptor family, CCR2 and
CCR5, in response to their respective ligands (MCP-1 and RANTES) and
the critical role of JAK/STAT pathway activation in later chemokine
signaling events (22
, 23)
. The activation in T cells of
different STATs by RANTES and MIP-1 has also been reported
(21)
.
Here we extend these results and demonstrate the tyrosine
phosphorylation of a member of CXC receptor family, CXCR4. CXCR4 is Tyr
phosphorylated in response to SDF-1. Several assays performed to
identify the kinase responsible for CXCR4 receptor phosphorylation led
to the conclusion that the JANUS kinase family members JAK2 and JAK3,
but not JAK1, cause early receptor activation. Both JAK2 and JAK3
association with the CXCR4 occurs even in the presence of PTX,
indicating no Gi participation in this process.
Their dissociation was not observed in the presence of PTX, however,
suggesting active participation of Gi pathways in
uncoupling JAK members from the receptor. In response to SDF-1, both
JAK2 and JAK3 are phosphorylated as soon as 60 s after binding,
indicating that their activation is simultaneous with their aggregation
with CXCR4. This suggests differential usage of the JANUS kinase family
members, depending on the chemokine receptor and the cell line studied.
Several STAT transcriptional factors, including STAT1, -2, -3, and -5,
but not STAT4 or -6, are associated with the CXCR4 after SDF-1
activation, concurring with the role assigned to the JAK tyrosine
kinases in transducing signals from hematopoietic growth factor
receptors (37)
. In these receptors, the activation and
association of JAK kinase with the receptor creates docking sites for
SH2-containing proteins such as STAT, leading to their phosphorylation
and activation of gene transcription. This is not a unique feature of
chemokine receptors, since other GPCR also activate STAT transcription
factors. This is the case for a skeletal muscle serotonin
5-HT2A receptor (38)
, the
angiotensin II AT1 receptor (39)
,
and the cAR1 chemotactic receptor of Dictyostelium, in which
G-protein-independent STAT activation has been demonstrated
(40)
. This observation agrees with the fact that JAK/STAT
pathway activation by chemokines is not blocked by PTX pretreatment,
indicating a G-protein-independent pathway. Nevertheless, in both
Dictyostelium and chemokine receptors, the G-protein may
have a role in regulating JAK/STAT activation in a yet unknown way.
Together, these results indicate that chemokine-mediated activation of
G-protein-coupled receptors leads to signal transduction, which invokes
intracellular phosphorylation intermediates used by other cytokine
receptors.
SDF-1-induced G-protein-mediated physiological effects are abrogated
after treatment of MOLT4 cells with the JAK-specific inhibitor AG490.
This is not the case when cells are treated with other tyrphostins,
indicating that G-protein-mediated responses through this receptor are
dependent on JAK kinase activity. These results confirm, and extend to
the CXCR4, the data obtained for MCP-1-activated CCR2
(22)
, in which conformational changes in the receptor
promoted by both ligand interaction and JAK tyrosine kinase association
are required for Gi protein association with
its binding site.
Although receptor association and phosphorylation of STAT proteins are
detected after SDF-1 activation, neither translocation to the
nucleus nor SIE binding activity are observed, probably due the low
efficiency of STAT phosphorylation. Only weak phosphorylation is
detected after chemokine activation as compared with that observed
using cytokines, with which the JAK/STAT pathway has classically been
studied (41)
. These data suggest a control mechanism that
implicates phosphotyrosine phosphatase activity, as has been shown for
the down-regulation of liver JAK2-STAT5b signaling by continuous growth
hormone stimulation (42)
. In the case of
chemokine-activated STAT, this dephosphorylation does not occur in the
nucleus, since no STATs are detected in Western blot analysis of
SDF-1-stimulated MOLT4 nuclear extracts. The down-regulation of the
JAK/STAT pathway by dephosphorylation was confirmed when, prior to
SDF-1 stimulation, MOLT4 cells were treated with pervanadate, a
phosphotyrosine phosphatase inhibitor. Under these conditions, the
chemokine increases STAT translocation to the nucleus, implicating
phosphotyrosine phosphatase activity in the control of this pathway.
The implication of a phosphatase in JAK/STAT regulation and the data
indicating that JAK activation, but not its dissociation from the
CXCR4, is GI independent, led us to analyze the
proteins potentially involved in this regulation. We found that the
Shp1 phosphatase aggregates with CXCR4 after SDF-1 binding, with a
time course that coincides with JAK dissociation from the receptor.
When cells are PTX treated, Shp1 does not associate with the receptor.
Both JAK dissociation and Shp1 association with the receptor require
activation of the Gi pathway and suggest possible
Gi-dependent regulation of JAK/STAT pathways by
Shp1. The induction of Shp1 may therefore provide a negative feedback
mechanism that contributes to the termination of SDF-1-induced
JAK/STAT signaling. Tyrosine phosphatases have recently been implicated
in the control of the JAK/STAT pathway, which is activated after the
stimulation of a G-protein-coupled receptor (43)
.
Activation of Shp phosphatases has also been described, showing that
abrogation of Shp2 activation leads to elevated and prolonged STAT1 and
STAT3 activation (44)
. The identification of JAK/STAT
signaling and its regulation by Shp1 as a primary event triggered by
chemokines of the CC and CXC families have important implications,
since chemokines could use alternative pathways that would permit
precise responses to stimuli, depending on the cell type and
differentiation status.
We also demonstrate that, as is the case for the CC chemokine
receptors CCR2 and CCR5 after binding of their ligands, CXCR4 undergoes
dimerization after SDF-1 binding. We have observed chemokine-induced
dimerization of CCR receptors (23
, 33)
, and direct
interaction between two CCR5 molecules has also been demonstrated by
Benkirane et al. (45)
, even in the absence of ligand
stimulation. Our data now extend this observation and clearly show that
ligand activation also promotes dimerization in members of the CXCR
receptor family. The functional significance of dimerization was
suggested by Hebert et al. (46)
, who, using the epitope
tagging approach, showed that agonist stimulation of the
ß2-adrenergic receptor stabilized the dimeric state of the receptor.
Receptor clustering is known to occur during the initiation of
ligand-induced internalization, although we have recently described a
mutant CCR2 form, CCR2bY139F (33)
, which dimerizes in
response to MCP-1 but is not internalized. This CCR2b mutant acts as a
CCR2b dominant negative mutant, blocking chemokine responses by forming
nonproductive complexes with partners containing the functional domain,
and demonstrating the biological requirement for dimerization in
chemokine responses. This conclusion is also reached in the case of the
adrenergic receptor (46)
, in which dimerization is
observed using purified receptors, clearly indicating that dimerization
is an internalization-independent phenomenon.
This dimerization model provides a context in which to understand the
ability of different chemokine-like ligands to act as agonists or
antagonists. Our results identify a novel molecular mechanism that may
underlie chemokine responses, revealing new possibilities for the
characterization of transcriptional activation in early genes, as well
as having implications in HIV-1 prevention. Recent reports indicate
that heterodimerization between CCR5 and its mutant, CCR5
32, is a
molecular explanation of the delayed onset of AIDS in heterozygous
CCR5/CCR532 individuals (47)
. Our results also extend
the model of cytokine receptor signaling to the chemokines, which are
functionally related molecules, although they use structurally
unrelated receptors. This model of chemokine responses links the
pathways implicated in Tyr kinase activation with the G-proteins and
assigns to the latter an important role in control of the
former.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
, macrophage-inflammatory
protein-1; PBS, phosphate-buffered saline; PO, monoclonal
horseradish peroxidase; RANTES, regulated on activation, normal T
expressed and excreted; SDF, stromal cell-derived factor; SDS-PAGE,
sodium dodecyl sulfate-polyacrylamide gel electrophoresis; SIE,
sis-inducible element. 
Received for publication November 16, 1998. Revised for publication April 22, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, MIP-1ß receptor as a fusion cofactor for macrophage-tropic HIV-1. Science 272,1955-1958[Abstract]
-chemokine, stromal cell-derived factor-1a, binds to the transmembrane G-protein-coupled CXCR4 receptor and activates multiple signal transduction pathways. J. Biol. Chem. 273,23169-23175 activate STATs in T cells. J. Biol. Chem. 273,309-314 receptor, confers chemokine selectivity. Evidence for a two-step mechanism for MCP-1 receptor activation. J. Biol. Chem 271,19084-19092
