 |
INTRODUCTION |
Chemokines regulate both basal and inflammation-induced
trafficking of leukocytes. In non-immune cells such as neurons, smooth muscle, stromal, endothelial and epithelial cells, chemokines also
control other aspects of tissue homeostasis (1, 2). Most chemokines are
members of one of the two subfamilies, CXC or CC, depending on whether
or not the two conserved amino-terminal cysteines are spaced by an
extra (X) amino acid. The biological activity of chemokines
is regulated at both transcriptional and post-translational levels.
Thus, stimuli from injured tissues, which disturb cell homeostasis,
activate expression of inducible chemokines (3). At a
post-translational level, chemokines are in some cases regulated by
selective proteolysis after secretion (4-9). Regulation of leukocyte
migration would ultimately be determined by the interplay of chemokines
and proteinases (10, 11). Although natural proteolytic cleavage of
chemokines occurs either at carboxyl or amino terminus, most of
functionally relevant proteolysis takes place at the amino terminus,
which encompasses both receptor-binding and signaling domains. Discrete
amino-terminal proteolysis of chemokines changes their biological
functions. In some cases limited proteolysis produces fragments with
enhanced biological activity or active polypeptides from an inactive
precursor (5, 6, 12). In other cases, cleavage of the amino terminus leads to the generation of potent antagonists as exemplified by monocyte chemoattractant protein
(MCP)-2,1 MCP-3 (7, 13), and
regulated on activation normal T cell expressed and secreted (13).
Finally, deletion of amino-terminal residues can extend the specificity
of the chemokine to other chemokine receptors, as reported for
hemofiltrate CC chemokine-1 (14) and MCP-3 (9). Given the redundancy
and overlap of ligand/receptor specificities in the chemokine system,
one might expect that the more striking in vivo functional
consequences of selective proteolysis would occur among those
chemokines that maintain non-promiscuous partnerships with their
receptors. Such may be the case for the chemokine CXCL12 or stromal
cell-derived factor-1 (SDF-1) which, through interaction with its
unique receptor CXCR4, orchestrates cardiovascular (15-17), neuronal
(17, 18), and hematopoietic embryonic development (15-18) as well as
bone marrow (BM) stem cell engraftment (19-21), postnatal homeostasis
of several hematopoietic lineages (21), and regulates T or B lymphocyte
trafficking (21, 22). SDF-1 and CXCR4 are constitutively and broadly
expressed in tissues (23, 24). In addition to their physiological
properties, increasing evidence emphasizes the participation of
SDF-1
/CXCR4 in the pathogenesis of infectious and inflammatory
processes. CXCR4 plays an important role in HIV infection, because
CXCR4 (along with CCR5) is one of the two major HIV coreceptors (25),
and its unique ligand SDF-1 has the capacity to prevent cell entry
of CXCR4-dependent viral isolates (X4 virus) (26, 27) by
both occupying and promoting down-regulation of CXCR4 (28). Moreover,
new evidence implicates the SDF-1/CXCR4 pair in the pathogenesis of
allergic airway diseases (29), rheumatoid arthritis (30-32), and
through induction of platelet aggregation the development of
atherosclerosis and thrombo-occlusive diseases (33). Thus, although
SDF-1 and CXCR4 are involved in the regulation of homeostatic
processes, it becomes evident that they are also constituents of
adaptive responses by inflammatory cells. Proteolytic modification of
either of the partners may be a mechanism that ultimately regulates the
functional activity of the SDF-1/CXCR4 pair. Two proteolytic
activities, dipeptidyl peptidase IV (DPP IV) (34) and cathepsin G (CG)
(35), which are released by leukocytes, can cleave SDF-1 in
vitro. However, the large diversity of soluble or cell-bound
proteinases expressed by leukocytes and the possibility that a given
substrate may be the target of diverse enzymes suggest that other
enzymes could be involved in the regulation of SDF-1/CXCR4
interactions. Therefore, it becomes important to define the relative
contribution of putative proteolytic activities released by leukocytes
that regulate SDF-1 functions. Moreover, it is important to
determine whether the capacity of CXCR4 to interact with its ligand is
modified by proteolysis. In this study, we report that SDF-1 and its
receptor CXCR4 are modified by limited proteolysis in the presence of
secreted leukocyte proteinases. We identify leukocyte elastase (LE) as
the major enzyme accounting for this processing. Proteolysis occurs at
the amino terminus of both molecules and results in abrogation of CXCR4/SDF-1 interaction and function.
| |
EXPERIMENTAL PROCEDURES |
Reagents
Salt-free lyophilized human leukocyte elastase (LE) was
purchased from Calbiochem and reconstituted at 35 µM in
50 mM sodium acetate, 200 mM NaCl, pH 5.5. Hepes, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenoltetrazolium bromide
(MTT), Indo-1 AM probe, Triton X-100, EDTA, EGTA,
N-succinyl-Ala-Ala-Ala-p-nitroanilide (a
neutrophil elastase substrate), and
N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilide (a
neutrophil cathepsin G (CG) substrate) were purchased from Sigma. The
proteinase inhibitors 4-(2-aminoethyl)-benzenesulfonyl fluoride
hydrochloride (AEBSF) and
N-[N-(L-3-trans-carboxirane-2-carbonyl)-L-leucyl]-agmatine (E64) were purchased from Roche Molecular Biochemicals. Pepstatin, 1,10-phenanthroline, aprotinin, and eglin C were obtained from Sigma.
Specific and irreversible LE inhibitor
N-methoxysuccinyl-Ala-Ala-Pro-Val-chloromethyl ketone
(MeOSuc-AAPV-CMK) was obtained from Calbiochem. MeOSuc-AAPV-CMK, E64,
and 1,10-phenanthroline compounds were dissolved in dimethyl sulfoxide
(Me2SO) of which the final working concentration was less
than 0.1% v/v. The specific, reversible, and competitive inhibitor of
CD26, Lys-[Z(NO2)]pyrrolidine (36), was a kind gift from Dr. A. Hovanessian (Institut Pasteur, Paris, France). Cytochalasin B and N-formyl-methionyl-leucyl-phenylalanine
(fMLP) were purchased from Sigma. Bovine serum albumin (BSA) and
dextran were purchased from Sigma. Hanks' balance salt solution
(HBSS), without calcium and magnesium, was purchased from Invitrogen. Horseradish peroxidase (HRP)-conjugated streptavidin was purchased from
Zymed Laboratories Inc. (San Francisco, CA).
Antibodies
K15C is an anti-SDF-1 mouse monoclonal antibody (mAb) (IgG2a
)
that recognizes an epitope containing the three first amino acids of
the chemokine (37). Antibody 12G5 (IgG2a), used as a phycoerythrin
conjugate (BD PharMingen), is a mouse mAb directed against the second
extracellular loop of CXCR4, and mAb 6H8 (IgG1) recognizes
amino-terminal residues 22-25 of CXCR4 (38).
Chemokine Synthesis
Wild type SDF-1 (aa 1-67) and SDF-1-biotin (SDF-1-bt,
aa 1-68) were synthesized by the Merrifield solid phase method on a
fully automated peptide synthesizer (Pioneer, Applied Biosystems, Inc.,
Foster, CA). SDF-1-bt carries a single biotin molecule conjugated to
a carboxyl-terminal Lys residue (Lys68). The procedures
used for SDF-1 and SDF-1-bt synthesis were described previously
(37, 39). The concentration of each chemokine was determined by amino
acid analysis in a 6300 Beckman amino acid analyzer, after hydrolysis
for 20 h in 6 N HCl, 0.2% phenol in the presence of a
known amount of norleucine as internal standard. All chemicals for the
synthesis were purchased from Applied Biosystems, Inc. The capacity of
SDF-1-bt and wild type SDF-1 to bind to and activate CXCR4 was
compared and found to be identical (39). The synthetic SDF-1 lacking
the Lys1-Pro2-Val3 sequence
(SDF-1-(4-67)) (40) was generously provided by Dr. I. Clark-Lewis
(Biomedical Research Center, University of British Columbia, Vancouver, Canada).
Cells
The CXCR4-positive Jurkat lymphoblastoid T cell line was
obtained from the American Type Cell Collection (ATCC) (Manassas, VA).
The HeLa P4.2 cell clone is stably transfected with a human CD4
cDNA and an HIV-LTR-driven Escherichia coli
-galactosidase reporter gene (41). HeLa 243 cells co-expressing both
Tat and Env HIV-1 proteins (derived from the X4 pLai proviral molecular clone) were obtained from Dr. M. Alizon (Hôpital Cochin, Paris, France) (41). Cell cultures were maintained in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% heat-inactivated fetal calf serum (FCS), Glutamax, antibiotics (100 units/ml penicillin and 100 µg/ml streptomycin), and methotrexate (2 µM).
Baby hamster kidney cells (BHK-21) (ATCC CCL-10) were maintained in
Glasgow's modified Eagle's medium (Invitrogen) containing 5% FCS, 20 mM Hepes, and 10% tryptose phosphate.
Isolation of Human PBMC or PMN and Preparation of
Proteinase-enriched Supernatants
Human peripheral blood mononuclear cells (PBMC) were isolated
from healthy donors using Ficoll-Paque (Amersham Biosciences AB)
density gradient centrifugation. PBMCs (2 × 106
cells/ml) were cultured at 37 °C in AIM V culture medium
(Invitrogen) without FCS and supplemented with Glutamax and antibiotics
(100 units/ml penicillin and 100 µg/ml streptomycin). Supernatants of
PBMC cultures (PBMC-SN) were collected at various time points by
centrifugation (1,500 × g for 5 min at 25 °C) and
further cleared of cell debris by ultracentrifugation (18,000 × g for 20 min at 4 °C) before use in SDF-1 degradation
experiments. Polymorphonuclear neutrophil leukocytes (PMN) were
isolated at room temperature within 3 h of venipuncture using
differential sedimentations on dextran (Sigma) and Ficoll-Paque
(Amersham Biosciences) as described previously (42). Isolated cells
(>95% PMN) were resuspended (5 × 106 cells/ml) and
equilibrated at 37 °C for 5 min in HBSS supplemented with 1.3 mM CaCl2 and 1 mM
MgCl2. Proteinase-enriched supernatants (PMN-SN) were
obtained from activated and degranulated PMN. For this purpose, PMN
were incubated with 5 µg/ml cytochalasin B for 5 min at 37 °C,
followed by addition of 0.5 µM fMLP for 5 min with gentle
agitation. Cell suspensions were rapidly cooled to 4 °C and
sedimented by centrifugation (1,700 × g for 10 min at 4 °C). Thereafter, PMN-SN were collected and ultracentrifuged (18,000 × g, for 15 min at 4 °C) to remove cell
debris and stored at 
80 °C until use. Pretreatment with
cytochalasin B is necessary to disrupt cytoplasmic microfilaments and
to favor degranulation during PMN activation (43). Cytochalasin B
and/or fMLP was omitted from control samples, whereas in some
experiments, fMLP was replaced with human IL-8 at 12 nM or
SDF-1 over a range 0.01-1 µM. Enzymatic activities of
human neutrophil LE and CG were measured in PBMC-SN or PMN-SN by
hydrolysis of their specific synthetic substrates, i.e.
N-succinyl-Ala-Ala-Ala-p-nitroanilide and
N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilide, respectively, as described previously (44). As standards, we used
serial dilutions of each purified enzyme over a wide range of concentrations.
Exposure of SDF-1 to Leukocyte Proteinases
SDF-1-bt (1 µM) or SDF-1 (1 µM) were incubated for different times at 37 °C in
cell-free PBMC-SN or PMN-SN or with purified LE (diluted in AIM V
medium when needed). Proteinase inhibitors were added for 25 min at
37 °C prior to incubation with the chemokine. Proteolytic reactions
were carried out in a volume of 50 µl.
Western Blot Analysis of SDF-1 Proteolysis
Degradation of SDF-1-bt or SDF-1 exposed to leukocyte
proteinases was stopped with Laemmli buffer. Proteins were separated by
SDS-PAGE on 16% acrylamide Tricine gels (Invitrogen) at 4 °C for
4 h and at 25 mA per gel. Gels were transferred to polyvinylidene difluoride Immobilon-PSQ membranes (<10-kDa cut-off,
Millipore Corp., Bedford, MA) by semi-dry transfer for 1 h at 120 mA, and membranes were blotted with anti-SDF-1 mAb K15C (2 µg/ml) to
detect the amino-terminal part of the chemokine. After stripping in 0.1 M glycine, pH 2.3, and subsequent incubation in a
phosphate-buffered saline (PBS) solution containing 1 M
NaCl, pH 7.4, membranes were blotted with HRP-conjugated streptavidin
to identify the carboxyl terminus of the chemokine. Visualization and
quantification of protein bands were performed using an electronically
cooled LAS-1000 plus charge-coupled device (CCD) camera system, and
Image Gauge 3.4 software (Fuji Photo Film Co., Tokyo, Japan).
ESI-MS Analysis of SDF-1 Proteolysis
Immediately after exposure of SDF-1-bt or SDF-1 to
leukocyte proteinases, chemokine fragments were isolated by reverse
phase-high performance liquid chromatography (RP-HPLC), and samples
were further analyzed by electrospray ionization mass spectrometry (ESI-MS). Briefly, SDF-1-bt degradation mixtures (150 µl) were filtered (0.45 µm) and injected on a Nucleosil 5C18 300 Å semi-preparative column (250 × 10 mm) (Macherey-Nagel,
Düren, Germany). By using a 28-48% linear gradient of
acetonitrile in 0.08% aqueous trifluoroacetic acid for 20 min at a 6 ml/min flow rate, intact SDF-1-bt or SDF-1 and their respective
fragments were eluted within the 9-11-min zone. Moreover, no peaks
were detected in the 9-11-min elution zone upon injection of PMN-SN,
PBMC-SN, or purified LE blanks. All samples collected were lyophilized
and analyzed by ESI-MS on an API 365 triple-quadrupole mass
spectrometer (PerkinElmer Life Sciences).
Inhibition of HIV-1 Env-mediated Cell-to-Cell Fusion by
SDF-1
Co-culture of HeLa cells expressing HIV-1 Env and Tat proteins
(HeLa 243 cell clone) with CD4+/CXCR4+ HeLa P4.2 cells leads to HIV-1
Env-mediated cell fusion and Tat-dependent activation of an
HIV LTR-driven -galactosidase reporter gene. The enzymatic activity
of -galactosidase reflects the magnitude of the HIV-1 Env-mediated
cell fusion. To assess the capacity of proteolyzed SDF-1 to prevent
HIV-1 Env-mediated cell-to-cell fusion (45), co-culture of 243 and P4.2
HeLa cells was carried out in the presence of intact SDF-1 or
cleaved SDF-1 in PBMC-SN. -Galactosidase activity was measured in
the cell lysates according to the manufacturer's instructions using
the -galactosidase Reporter Gene Assay (Roche Molecular Biochemicals).
Analysis of CXCR4 Endocytosis Induced by
SDF-1--
Jurkat CXCR4+ cells (1 × 106 cells/ml) were
incubated with SDF-1 for 30 min at 37 °C in PBS-0.1% BSA or
medium. Cells were then incubated for 3 min at room temperature in an
acidic buffer (50 mM glycine, pH 2.3) that stops receptor
endocytosis and removes CXCR4-bound SDF-1 molecules that would mask
CXCR4 recognition by the mAb 12G5. Jurkat cells were washed twice with
ice-cold PBS, 0.1% BSA, before incubation with PE-conjugated mAb 12G5
(1:100) for 1 h at 4 °C. Samples were analyzed on a FACSCalibur
flow cytometer (BD PharMingen). Basal cell fluorescence intensity was
determined using cells stained with a PE-conjugated IgG2a isotype
control alone.
Chemotaxis Assay
Migration of CEMx174 cells, a human lymphoblastoid CXCR4+/CD4+ T
cell line, was assessed in 48-well chambers (Neuro Probe Inc., Cabin
John, MD) as described previously (37). Briefly, SDF-1 or
SDF-1-(4-67) was added to the lower well at different concentrations, in a total volume of 30 µl in 25 mM
Hepes-buffered Dulbecco's modified Eagle's medium at pH 7.4 (chemotaxis medium). The chemotaxis chamber was then assembled using
polyvinylpyrrolidone-free polycarbonate membranes with 8 µm pore size
(Costar, Cambridge, MA), and 50 µl of CEMx174 cells (1 × 106 cells/ml), in chemotaxis medium without chemokine, was
added to the upper well. After incubation for 4 h at 37 °C in a
5% CO2 humidified incubator, the chamber was disassembled,
and cells that migrated through to the lower wells were transferred to
a working 96-well plate. The migrated viable cells were then
quantitated by a sensitive and quantitative colorimetric assay using
MTT staining (46). Viable cells reduced MTT, and the degree of MTT
reduction, which corresponds to the relative cell number, was measured
automatically with an enzyme-linked immunosorbent assay reader reading
the absorbance at 590 nm.
Measurement of Cytosolic Free Calcium
Intracellular calcium levels were measured in a Wallac
VICTOR2 multilabel counter (EG & G Wallac, Turku,
Finland) using Indo-1 loaded CEMx174 cells. Briefly, cells (5 × 106 cells/ml) were loaded with 5 µM Indo-1/AM
in HBSS buffer (140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM MgSO4,
1.2 mM CaCl2, 10 mM Hepes, 5 mM glucose, 0.3 mM
KH2PO4, and 2 mM
Na2HPO4), pH 7.0, for 30 min at 37 °C.
Thereafter, an equal volume of HBSS, pH 7.4, containing 10%
heat-inactivated FCS, was added, and the cell suspension was incubated
for 30 min. After washing with HBSS containing 5% heat-inactivated
FCS, pH 7.2, cells were resuspended at 5 × 106
cells/ml and maintained at room temperature in the dark until use. For
calcium measurements, aliquots of this cell suspension were
preincubated for 5 min at 37 °C in a 96-well flat bottom plate, in a
total volume of 200 µl (1 × 106 cells/ml) in HBSS
supplemented with 5% heat-inactivated FCS, pH 7.4. Different
concentrations of SDF-1 or SDF-1-(4-67) were added at the
indicated times. Cell suspensions were excited at 355 nm, and calcium
levels were determined by monitoring fluorescence emissions at 405 nm
(Ca2+-bound dye) and 485 nm (Ca2+-free dye)
every 2 s at 37 °C. Free calcium concentration values were
calculated from the ratio of emission fluorescences (405/485 nm) using
the equation described by Grynkiewickz et al. (47), with a
Kd value of 250 nM for Indo-1. The
Rmax value was obtained by lysing the cells with
0.3% Triton X-100, followed by an addition of excess EGTA (300 mM in Tris-HCl, pH 7.4) for Rmin.
All experiments were performed in triplicate wells.
Analysis of CXCR4 Proteolysis
Immunoprecipitation and Electrophoretic Separation of
CXCR4--
BHK-21 cell monolayers (5 × 106 cells)
were infected (multiplicity of infection 50) with a recombinant
defective Semliki Forest virus (SFV) (48) encoding either full-length
sequences of -galactosidase or human CXCR4 fused to the TETSQVAPA
sequence (C9 peptide) encoded in the bovine rhodopsin (49). Fourteen
hours after infection, cells were pulse-labeled for 15 min with
[35S]methionine and [35S]cysteine (200 µCi/ml) and cultured for and additional 2 h. Thereafter, cells
were detached with PBS, 2 mM EDTA and exposed to PMN-SN for
2 h at 37 °C. CXCR4 proteolysis was stopped by adding 100 µl
of PBS containing 0.2 mg/ml eglin C. Cell pellets were lysed in
solubilization buffer made of 100 mM
(NH4)2SO4, 20 mM
Tris-HCl, pH 7.5, 10% glycerol, 1% (w/v)
n-dodecyl--D-maltoside and a proteinase inhibitor mixture (Roche Molecular Biochemicals). Lysates were incubated at 4 °C for 30 min, and insoluble debris was removed by
centrifugation (14,000 × g for 30 min). CXCR4 was
immunoprecipitated using either 12G5 (10 µg/ml) or 1D4 (10 µg/ml)
mAbs coupled to protein-G-Sepharose beads. The 1D4 antibody (National
Cell Culture Center, Minneapolis, MN) recognizes the C9 peptide (49).
Proteins were separated by SDS-PAGE on 10% acrylamide gels at 4 °C,
and radiolabeled bands were quantified in a Molecular Dynamics
PhosphorImager (Amersham Biosciences).
Flow Cytometry Analysis--
Jurkat or CXCR4-expressing BHK-21
cells were incubated in the presence of PMN-SN or purified LE in
serum-free AIM V medium for 2 h at 37 °C. Inhibition of LE
activity was achieved by incorporating into the samples either
MeOSuc-AAPV-CMK (3 µM) or eglin C (0.40 mg/ml) before
incubation with cells. Proteinase-exposed cells were washed with
ice-cold PBS, 0.1% BSA and labeled with either mAbs 6H8 or
PE-conjugated 12G5. Labeling by 6H8 was revealed by a secondary goat
PE-conjugated anti-mouse antibody (BD PharMingen). Antibody-labeled
cells were fixed in PBS, 1% formaldehyde and analyzed by flow cytometry.
Binding of SDF-1 to CXCR4
Jurkat cells were exposed to purified LE (3 µM)
for 2 h at 37 °C, and proteolysis was stopped by adding
MeOSuc-AAPV-CMK (3 µM). CXCR4 degradation was monitored
by flow cytometry analysis with either 12G5 or 6H8 mAbs. Cells were
incubated with 0.25 nM iodinated SDF-1 (specific
activity 2200 Ci/mmol; PerkinElmer Life Sciences) and competed with
unlabeled SDF-1 for 1 h at 4 °C in a final volume of 300 µl. Incubations were terminated by centrifugation at 4 °C. Cell
pellets were washed twice in ice-cold PBS. Nonspecific binding was
determined in the presence of 1 µM of unlabeled SDF-1.
Cell pellet-associated radioactivity was counted using a
microcomputer-controlled, 1272 CliniGamma counter (LKB-Wallac,
Stockholm, Sweden). Data were analyzed using GraphPad Prism 2.0 software (GraphPad Software, Inc. San Diego, CA).
| |
RESULTS |
Proteolysis of SDF-1 by Leukocyte Proteinases--
Modification
of SDF-1 structure and function was investigated following in
vitro exposure of the chemokine to human leukocytes. For this
purpose, synthetic SDF-1 carrying an additional Lys-conjugated biotin at the carboxyl terminus (SDF-1-bt, aa 1-68) was incubated with supernatants from blood mononucleated cells (PBMC-SN). SDF-1-bt was detected by Western blot analysis using either a monoclonal antibody (mAb) K15C recognizing an epitope encoded in the
amino-terminal part of the chemokine (37) or by reacting HRP-coupled
streptavidin with the carboxyl-terminal biotin. We observed that
incubation of SDF-1-bt with PBMC-SN obtained from cells cultured for
24 h led to the rapid loss of the amino-terminal epitope
recognized by the mAb K15C (Fig.
1a). Both the slight change on
the electrophoretic mobility and the intactness of the carboxyl
terminus region of SDF-1-bt, revealed by HRP-streptavidin labeling,
proved the discrete modification undergone by the chemokine (Fig.
1a, arrows). These findings indicate the
existence of a proteolytic mechanism that cleaves the amino-terminal
domain of SDF-1.
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 1.
Western blot analysis of
SDF-1 cleavage in PBMC-SN. a,
synthetic SDF-1-bt (1 µM) was incubated for up to
2 h in cell-free supernatants collected from 24-h PBMC cultures
(2 × 106 cells/ml). Proteins were separated on
Tricine/SDS-PAGE, transferred to polyvinylidene difluoride membranes,
and blotted with anti-SDF-1 mAb K15C to detect the amino-terminal
domain of the chemokine (N-ter. detection). After acid
stripping, membranes were blotted with HRP-streptavidin to identify the
carboxyl terminus of the chemokine (C-ter. detection).
Quantification of chemiluminescence signals (solid squares,
mAb K15C; solid circles, HRP-streptavidin) was performed
using an electronically cooled CCD camera. Arrows indicate
cleaved and uncleaved forms of the chemokine. b, SDF-1-bt
was incubated for 12 h in PMBC-SN in the absence (None)
or in the presence of the following proteinases inhibitors: 1 mM AEBSF; 10 µM aprotinin; 30 µM CD26 inhibitor (CD26-Inh); 2 µg/ml E64; 1 µM pepstatin; 5 mM 1,10-phenantroline
(1,10-Phen.); and 2.5 mM EDTA. c,
SDF-1-bt was incubated in PMBC-SN for 4 h in the absence ()
or in the presence of the specific elastase inhibitor, MeOSuc-AAPV-CMK.
b and c, amino-terminal and carboxyl-terminal
domains of the chemokine were detected by Western blot analysis as
described in a. The control (SDF-1-bt
lane) was generated by incubating SDF-1-bt with AIM V
medium instead of PBMC-SN. Illustrated is an experiment representative
of six, in which LE and CG concentrations in PBMC-SN were 23 and 24 nM, respectively.
|
|
To characterize the proteolytic activity accounting for the discrete
amino-terminal degradation of SDF-1-bt, chemokine was exposed to
PBMC-SN in the presence of various proteinase inhibitors (Fig.
1b). Aspartic, cysteine, or metalloproteinase inhibitors (pepstatin, E64, and 1,10-phenanthroline or EDTA, respectively) failed
to prevent degradation. A specific inhibitor of DPP IV, a proteolytic
enzyme expressed by lymphocytes (CD26) shown previously to selectively
cleave the amino-terminal part of SDF-1 (34), also failed to inhibit
degradation of the chemokine. In contrast, aprotinin and AEBSF
efficiently prevented proteolysis of SDF-1-bt thus indicating that
among leukocyte proteinases a serine proteinase accounts for
degradation of SDF-1.
To identify the SDF-1 residues cleaved by PBMC proteinases, we
analyzed SDF-1-bt degradation products by electrospray ionization mass spectrometry (ESI-MS). For this purpose, SDF-1-bt (aa 1-68, Fig. 2a, Control
spectrum) was incubated with PBMC-SN for 4 h at
37 °C and fractionated by RP-HPLC. Thereafter, eluted products were
analyzed by ESI-MS. A polypeptide of molecular mass 7,973 Da was
detected as the major molecular species generated upon cleavage and
corresponded to an SDF-1-bt fragment lacking the amino-terminal
Lys1-Pro2-Val3 sequence (Fig.
2a, PBMC-SN cleaved spectrum). This
finding confirms that the amino terminus of SDF-1 is the target of
the proteolytic attack by PBMC-secreted proteinases. To assess that the
incorporation of a biotinylated Lys at the carboxyl-terminal end of
the chemokine does not condition the pattern of proteolytic degradation
undergone by SDF-1-bt, chemokine fragments generated by PBMC
proteinases from either SDF-1-bt (8,298 Da) or a wild type,
unmodified SDF-1 (SDF-1; aa 1-67, 7,831 Da) were analyzed by
ESI-MS. The fragments were 7,973 Da and 7,507 Da, respectively, and
corresponded in both cases to polypeptides lacking the
Lys1-Pro2-Val3 sequence (Fig.
2b, PBMC-SN cleaved, SDF-1-bt-(4-68)
lower and SDF-1-(4-67) upper spectra). Thus,
we conclude that both SDF-1-bt and the unmodified SDF-1 show
identical susceptibility to the proteolytic attack by PBMC-secreted
proteinases.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 2.
ESI-MS analysis of
SDF-1-bt and SDF-1
degradation in PBMC-SN. a, synthetic SDF-1-bt
(aa 1-68, 8,298 Da) (1 µM) was incubated for 4 h at
37 °C in AIM V medium (Control spectrum), PBMC-SN
(PBMC-SN cleaved spectrum), or PBMC-SN containing 0.5 µM MeOSuc-AAPV-CMK inhibitor (MeOSuc-AAPV-CMK
inhibition spectrum). Thereafter, chemokine derivatives
were isolated by RP-HPLC and further analyzed by ESI-MS. b,
SDF-1 (aa 1-67, 7,831 Da) or SDF-1-bt (aa 1-68, 8,298 Da) (1 µM) were incubated for 2 h at 37 °C in PBMC-SN
and further analyzed by ESI-MS. a and b,
SDF-1-bt-(1-68) of 8,298 Da and SDF-1-(1-67) of 7,831 Da
represent the intact form of chemokines, whereas SDF-1-bt-(4-68) of
7,973 Da and SDF-1-(4-67) of 7,507 Da represent the corresponding
cleaved forms of the molecules. The peaks noted (+16)
correspond to the biotin-oxidized forms of SDF-1-bt that show an
increase of 16 Da as compared with the nonoxidized counterparts.
Illustrated is an experiment representative of six, in which LE and CG
concentrations in PBMC-SN were 23 and 24 nM,
respectively.
|
|
The pattern of SDF-1 cleavage by PBMC proteinases is compatible with
the enzymatic activity characteristic of LE. LE is a serine proteinase
that represents the major releasable enzyme in monocytes (50) and has
been demonstrated to have proteolytic activity on various chemokines
and cytokines (51). LE preferentially recognizes Pro-Val sequences and
cleaves Val-X peptide bonds (52). Thus it is reasonable to
postulate that LE may account for the removal of the
Lys1-Pro2-Val3 sequence of SDF-1
when the chemokine is exposed to PBMC-SN, which has an average LE
content of 18 nM ± 2 nM (mean ± S.E., n = 5). The cleavage of SDF-1-bt by purified human
LE at low enzyme concentration (1 nM) supported this
assumption (Fig. 3a), and
ESI-MS analysis revealed a single degradation fragment of 7,973 Da
(Fig. 3c), which corresponded to the SDF-1-bt-(4-68) polypeptide lacking the
Lys1-Pro2-Val3 sequence. Direct
evidence identifying LE among other PBMC proteinases as the enzymatic
activity hydrolyzing the SDF-1 amino terminus was obtained by
addition of the LE-specific inhibitor MeOSuc-AAPV-CMK (53) (Fig.
3b). Indeed, as shown by Western blot and ESI-MS analysis
(Fig. 1c and Fig. 2a, MeOSuc-AAPV-CMK
spectrum), degradation of SDF-1 by PBMC-secreted proteinases was
fully prevented in the presence of MeOSuc-AAPV-CMK. Using SDF-1 we
verified by ESI-MS analysis that biotinylation of the carboxyl terminus
of SDF-1 does not condition the susceptibility of the chemokine to
proteolysis by purified LE (data not shown). Finally, when SDF-1-bt
was directly exposed to PBMC suspensions (5 × 106
cells/ml), we observed that cleavage of the chemokine was identical to
that described above for cell-free PBMC-SN or purified LE, and
MeOSuc-AAPV-CMK completely inhibited the degradation of the chemokine
(data not shown). Together, these findings indicate that LE is the
major proteolytic activity among PBMC-secreted proteinases that cleaves
the amino-terminal end of SDF-1.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 3.
Western blot and ESI-MS analysis of
SDF-1-bt degradation by human purified
LE. a, SDF-1-bt (1 µM) was incubated
for 6 h at 37 °C with purified human LE at different
enzyme/chemokine molar ratios. b, SDF-1-bt (1 µM) was incubated for 6 h at 37 °C with purified
human LE at a 1:1000 (LE/SDF-1-bt) molar ratio in the absence ()
or in the presence of the specific LE inhibitor, MeOSuc-AAPV-CMK.
a and b, the chemokine degradation was analyzed
by Western blot as described in the legend to the Fig. 1, and the
SDF-1-bt lanes represent controls of intact
chemokine incubated in AIM V medium instead of LE. c,
SDF-1-bt-(1-68) (1 µM) was incubated for 6 h at
37 °C with purified human LE at a 1:100 (LE/SDF-1-bt) molar
ratio, and ESI-MS analysis was performed as described in the legend to
Fig. 2. SDF-1-bt-(4-68) of 7,973 Da represents the cleaved form of
the chemokine obtained upon exposure to LE. Results from one
representative experiment of three are shown.
|
|
Although LE can be released by monocytes, PMN represents among
leukocytes the major source of LE as well as of other serine proteinases such as cathepsin G (CG) and proteinase-3 (51). Stored in
the azurophilic granules of PMN, the amount of LE is roughly 20-fold
higher than in monocytes (50). PMN proteinases are released along with
other microbicidal products upon inflammatory responses (54). This
process may result in the proteolysis and regulation of the functional
activities of SDF-1/CXCR4 as is the case for various cytokines and
cytokine receptors (51). We thus examined the pattern of cleavage of
SDF-1 when exposed to PMN-secreted proteinases. PMNs were activated
by the bacterial agonist fMLP, which promotes macrophages and PMN
recruitment and activation at sites of inflammation (54). In keeping
with previous reports (42, 44), measurement of enzymatic activities
indicated that cell-free supernatants of fMLP-activated PMN (PMN-SN)
contained both active LE and CG (see legend of Figs.
4-6). When SDF-1-bt was incubated in
PMN-SN, Western blot analysis showed that cleavage of the chemokine at
its amino terminus (Fig. 4a) occurred within minutes, with
roughly 50% of the amino terminus of SDF-1-bt proteolyzed after 5 min of incubation, whereas the carboxyl terminus remained intact. As
compared with PBMC-SN, PMN-SN degrades SDF-1-bt more efficiently,
with kinetics that correlate with the amount of secreted LE accumulated
in the samples (compare Fig. 1a and 4a). ESI-MS analysis of isolated fragments identified a unique form showing a
fragment of 7,973 Da that, as observed previously with PBMC-SN or
purified LE, corresponds to a cleaved SDF-1 polypeptide lacking the
Lys1-Pro2-Val3 sequence (Fig.
4c, PMN-SN cleaved spectrum). As shown
by either Western blot or ESI-MS analysis, MeOSuc-AAPV-CMK fully
prevented degradation of SDF-1-bt (Fig. 4, b and
c, MeOSuc-AAPV-CMK inhibition spectrum), thus confirming the predominant role played by LE
among PMN-secreted proteinases in the proteolysis of the chemokine. Similar findings were observed when PMNs were induced by the
inflammatory CXC chemokine IL-8, which orchestrates leukocyte
trafficking during inflammatory responses and is known to induce PMN
secretion of microbicidal products, including proteinases (55). As in
the case of fMLP, the inhibitor MeOSuc-AAPV-CMK prevented proteolysis of SDF-1-bt exposed to PMN-SN from IL-8-induced PMN (data not shown). In keeping with previous reports showing that SDF-1, which
is not an inflammatory chemokine and fails to induce degranulation from
basophils (56), eosinophils (57), or mast cells (58), we observed that
SDF-1 did not induce proteinase release from PMN (data not
shown).
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 4.
Western blot and ESI-MS analysis of
SDF-1-bt degradation by PMN-SN derived from
fMLP-stimulated PMN. a, synthetic SDF-1-bt (1 µM) was incubated for up to 2 h in cell-free
supernatants from PMN (5 × 106 cells/ml) activated
with 0.5 µM fMLP in the presence of 5 µg/ml
cytochalasin B (PMN-SN). Chemokine degradation was analyzed by Western
blot, and protein bands were quantitated as described in the legend to
Fig. 1. Results are means ± S.E. of three independent
experiments. b, SDF-1-bt (1 µM) was
incubated for 2 h at 37 °C with PMN-SN in the absence () or
in the presence (+) of 1 µM MeOSuc-AAPV-CMK. The
SDF-1-bt lane represents control of intact
chemokine incubated in AIM V medium instead of PMN-SN. c,
SDF-1-bt (1 µM) was incubated for 4 h at 37 °C
with PMN-SN either in the absence (PMN-cleaved spectrum) or
in the presence of 1 µM MeOSuc-AAPV-CMK
(MeOSuc-AAPV-CMK inhibition spectrum). ESI-MS
analysis was performed as described in the legend to Fig. 2.
Illustrated is an experiment representative of three, in which LE and
CG concentrations in PMN-SN were 390 and 230 nM,
respectively.
|
|
The functional capacity of the LE-cleaved SDF-1 was next
investigated. The non-promiscuous interaction between SDF-1 and CXCR4 permits us to unambiguously assess the biological activity of the
chemokine through the analysis of either the ligand-induced endocytosis
or the HIV-coreceptor function of CXCR4 in the crude supernatants
obtained from PBMC cultures or isolated PMN. These SDF-1-mediated
functions require binding to CXCR4, are non-pertussis toxin-sensitive,
and do not require G-protein-mediated signaling in lymphocytes
(28).2 To this purpose, 1 µM SDF-1-bt was incubated with serial 2-fold dilutions
of either PBMC-SN or PMN-SN, and samples were assessed for their
capacity to promote CXCR4 endocytosis in Jurkat T cells. Expression of
CXCR4 at the cell surface was investigated using a specific mAb (12G5)
that recognizes a conformational epitope in the second extracellular
loop of the receptor (59). Because binding of SDF-1 to CXCR4
competes with the mAb 12G5 for CXCR4 occupation, residual chemokine
bound to its receptor was removed from the cell surface by acidic
washes before labeling with 12G5. Serial dilutions of PBMC-SN (Fig.
5c) or PMN-SN (Fig.
5d) resulted in a progressive increase in the capacity of
the samples to promote CXCR4 endocytosis, which correlated with the
amount of intact SDF-1-bt (see Western blots, top insets
in Fig. 5, c and d). The capacity of SDF-1 to
prevent HIV-1 Env-dependent cell-to-cell fusion largely
relies on the capacity of the chemokine to promote CXCR4 endocytosis
(Fig. 5f). Following incubation of SDF-1-bt with PBMC-SN
and in agreement with the impaired ability of the chemokine to induce
CXCR4 endocytosis, the capacity of SDF-1 to prevent HIV-1
Env-mediated cell fusion was severely reduced. It should be noted that
when PBMC-SNs were used as a source of proteinases, a discrete receptor
endocytosis and inhibition of HIV Env-dependent cell
membrane fusion were observed in some samples in the absence of
detectable intact SDF-1-bt by Western blot analysis, which is less
sensitive than the functional assays (Fig. 5, c and
f, see dilution 1/4). This phenomenon is likely
accounted by remnant, intact chemokine not degraded by the relatively
low accumulation of LE in PBMC-SN. This assumption was confirmed by the
complete lack of CXCR4 endocytosis and HIV inhibitory capacity shown by
the synthetic SDF-1-(4-67) (data not shown). Finally, exposure of 1 µM SDF-1 to 10 nM purified LE totally
abrogated the capacity of the chemokine to induce CXCR4 endocytosis
(Fig. 5e), and incubation of the undiluted PMN-SN with
MeOSuc-AAPV-CMK (Fig. 5d, AAPV) preserved the
CXCR4 endocytosis capacity of SDF-1-bt.
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 5.
Functional analysis of LE-cleaved
SDF-1 and synthetic
SDF-1-(4-67). a,
intracellular calcium mobilization. A comparative analysis of
SDF-1-(4-67) versus SDF-1 was performed using
Indo-1-loaded CEMx174 cells at the indicated concentrations of
chemokines. SDF-1-(4-67) did not induce calcium mobilization at any
concentration tested (ranging from 5 nM to 1 µM), and the result obtained at 1 µM is
shown. Results are means ± S.E. of three independent experiments.
b, chemotaxis assay. Migration of CEMx174 lymphoblastoid
cells was measured following stimulation with the indicated
concentrations of SDF-1 or SDF-1-(4-67) for 3 h at
37 °C, and migrated viable cells were quantitated using a sensitive
colorimetric MTT assay. Results are means ± S.E. obtained from
four independent experiments. c, d, and
e, CXCR4 endocytosis. Jurkat cells were exposed to
SDF-1-bt (1 µM) previously treated for 2 h at
37 °C in serial dilutions of 24-h cell-free PBMC-SN (c)
or serial dilutions of PMN-SN obtained from fMLP-activated PMN
(d) or with 10 nM purified LE in the presence
(+) or the absence () of MeOSuc-AAPV-CMK (0.05 µM)
(e). d, histogram labeled as AAPV
refers to SDF-1-bt (1 µM) treated by MeOSuc-AAPV-CMK
(1 µM) prior to incubation of the chemokine in PMN-SN
(1/1 dilution). c-e, CXCR4 expression was
analyzed by flow cytometry using a PE-conjugated 12G5 anti-CXCR4 mAb. A
PE-conjugated IgG2a isotype was used as a control. Results are
expressed as means ± S.E. of three independent experiments and
referred to CXCR4 expression in the absence of SDF-1-bt, taken as
100%. f, X4 HIV-1 Env-mediated cell fusion. Env- and
Tat-expressing HeLa 243 cells were coincubated with CD4+/CXCR4+ HeLa
P4.2 cells expressing a HIV LTR-driven -galactosidase reporter gene.
After fusion at a cell ratio of 1:1 for 16 h at 37 °C,
-galactosidase activity was measured in cell lysates. Results are
means ± S.E. of three independent
|
|
The capacity of the LE-cleaved SDF-1 to trigger pertussis
toxin-sensitive, G-protein-dependent intracellular
signaling was assessed in lymphocytes by measuring cytosolic calcium
mobilization and cell chemotaxis. Because of the potential presence of
unidentified ligands different from SDF-1 in the PBMC-SN or PMN-SN
capable of inducing cell signaling independently of CXCR4 activation, these experiments necessitated the use of a synthetic SDF-1
polypeptide lacking the
Lys1-Pro2-Val3 sequence cleaved by
LE from SDF-1. Our results indicate that, in contrast to the
unmodified SDF-1 (aa 1-67), the truncated SDF-1-(4-67) (kindly
provided by Dr. I. Clark-Lewis, University of British Columbia,
Vancouver, Canada) failed to mobilize calcium from
intracellular stores (Fig. 5a) and to promote chemotaxis of
CXCR4-expressing lymphoid cells (Fig. 5b).
We conclude that among proteinases expressed and secreted by human
leukocytes, a serine proteinase indistinguishable from LE is
responsible for the selective proteolysis of the amino terminus that
inactivates SDF-1.
Degradation of CXCR4 by Leukocyte Proteinases--
Similar to
other G-protein-coupled receptors, the amino-terminal part of CXCR4
represents an exposed domain susceptible to the attack by proteinases.
Proteinase-activated (60) and bradykinin B2 (61) receptors
are regulated through selective cleavage of their amino-terminal
domains by serine proteinases. Because the extracellular CXCR4 amino
terminus is a major determinant for high affinity binding of SDF-1,
we investigated whether CXCR4 is prone to cleavage by leukocyte proteinases.
The effect of PMN-derived proteinases on the cell surface expression of
CXCR4 was first assessed in BHK cells (Fig.
6a),
transiently expressing functional human CXCR4 from an SFV-derived
vector. Antibodies directed against specific regions of CXCR4 were used to look for molecular evidence of CXCR4 cleavage by leukocyte proteinases. The mAb 6H8 recognizes an epitope encompassing
amino-terminal residues 22-25 (38). Binding of mAb 6H8 was reduced by
80% after exposure of BHK cells to PMN-SN containing proteinases and
only to 50% when PMN-SN was diluted (Fig. 6a, 1/1 and
1/4 dilutions, respectively). In contrast, binding of mAb 12G5,
which recognizes a conformational epitope located in the second
extracellular loop, remained relatively unaffected. Like
MeOSuc-AAPV-CMK, eglin C blocks LE but displays a broader inhibitory
spectrum on other leukocyte serine proteinases (62). Preincubation of
PMN-SN with MeOSuc-AAPV-CMK or eglin C efficiently prevented loss of
mAb 6H8 binding to its amino-terminal epitope on CXCR4 (Fig.
6a), demonstrating involvement of a proteolytic mechanism in
6H8 epitope disappearance. In the presence of eglin C, maintenance of
amino-terminal CXCR4 expression was slightly better than that obtained
with MeOSuc-AAPV-CMK. Although the role of other serine proteinases
cannot be formally excluded, it can be assumed that a proteolytic
activity sensitive to MeOSuc-AAPV-CMK, i.e. LE, accounts for
most of the loss of the amino-terminal domain of CXCR4. Experiments
conducted in Jurkat T cells constitutively expressing CXCR4 led us to
similar conclusions (Fig. 6b). Although the putative CXCR4
degradation occurs preferentially at the amino terminus, longer
exposure (more than 2 h) of cells to undiluted PMN-SN resulted in
progressive loss of the epitope defined by the mAb 12G5 (data not
shown). The biological relevance of this phenomenon is uncertain and
may reflect the attack of CXCR4 at secondary cleavage sites either by
LE or other serine proteinases.
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 6.
Proteolysis of CXCR4 exposed to PMN-SN or
purified LE. For flow cytometry analysis, CXCR4-expressing BHK
(a) or Jurkat (b) cells (2 × 106 cells/ml) were exposed for 2 h at 37 °C to
dilutions of PMN-SN obtained from fMLP-activated PMN. PMN-SN were
treated (+) or not () either with 1 µM MeOSuc-AAPV-CMK
(AAPV) or 0.40 mg/ml eglin C (Eglin C) for 1 h at 37 °C prior to incubation with cells. The expression of CXCR4
was assessed with anti-CXCR4 mAbs 6H8 or PE-conjugated 12G5. In the
case of 6H8, cells were stained with goat PE-conjugated anti-mouse
immunoglobulins. Finally, cells were fixed and analyzed by flow
cytometry. For SDS-PAGE immunoprecipitation, BHK cells expressing
C9-tagged CXCR4 were 35S-labeled before being exposed to
PMN-SN (c and d) or to 1 µM
purified LE (c) for 2 h at 37 °C. PMN-SN were
treated (+) or not () with 0.2 mg/ml eglin C (c,
Eglin C) prior to incubation with cell cultures. CXCR4 from
cell lysates were either directly separated by SDS-PAGE (c)
or immunoprecipitated using either anti-CXCR4 mAb 12G5 or mAb 1D4
directed against the C9 tag (d). Irrelevant IgG2a were used
as control (d). c, 35S-labeled
bands were quantitated in a PhosphorImager device, and the LacZ
lane is a control obtained from cells infected with an SFV vector
expressing lacZ instead of CXCR4. The
asterisk indicates the band corresponding to the
lacZ product. c and d, the
full-length CXCR4 species are indicated by arrowheads and
correspond to a mass of 40 kDa. Illustrated is an experiment
representative of three, in which LE and CG concentrations in PMN-SN
were 542 and 334 nM, respectively. experiments. Values are represented as the percentage of
fusion inhibition referred to the intact SDF-1-bt. c-e,
control histograms labeled SDF-1-bt were obtained by
incubating SDF-1-bt (1 µM) in AIM V medium instead of
cell supernatants. After degradation of SDF-1-bt in PBMC-SN
(c and f), PMN-SN (d), or with
purified LE (e), LE and CG activities were blocked by
addition of 0.40 mg/ml of eglin C prior to incubation of the treated
chemokine with CXCR4+ cells. SDF-1-bt was detected by Western blot
using mAb K15C (c-f, top inserts) as described
in the legend to Fig. 1. LE and CG concentrations in experiments shown
in c, d, and f were 542 and 334 nM, respectively for PMN-SN, and 23 and 24 nM,
respectively for PBMC-SN.
|
|
More direct evidence supporting the amino-terminal degradation of CXCR4
by leukocyte proteinases was obtained by electrophoretic separation of
the metabolically radiolabeled receptor from BHK cells exposed to
PMN-SN. Expression of ectopic genes from infectious SFV vectors allows
robust and predominant expression of vector-encoded proteins, while
causing almost complete shut-off of host cell protein synthesis (48). A
protein migrating with a relative molecular mass of 40 kDa corresponded
to the expected size of full-length CXCR4 (25) (Fig. 6c).
The nature of the lower labeled band is unclear and may correspond to a
precursor of CXCR4 (Fig. 6c). Immunoprecipitation with
either mAb 12G5 or mAb 1D4 directed against the C9 tag at the carboxyl
terminus of the CXCR4 cytoplasmic tail confirmed that the two major
radiolabeled bands corresponded to CXCR4 (Fig. 6d).
Following exposure of cells to PMN proteinases, the amount of
full-length CXCR4 was drastically reduced in the total cell lysates
(Fig. 6c) and in 1D4 or 12G5 immunoprecipitates (Fig.
6d). Quantification of dried gels showed accumulation of a
truncated protein generated upon cleavage of CXCR4 by PMN proteinases at the level of the lower CXCR4 form (Fig. 6c,
histograms). Incubation in the presence of eglin C prevented
degradation of full-length CXCR4 by PMN proteinases and restored the
original ratio between the high 40-kDa full-length and the low
Mr forms (Fig. 6c). Finally, direct
incubation of CXCR4-expressing BHK cells with purified LE (1 µM) resulted in a similar disappearance of full-length
CXCR4 molecules and accumulation of the lower Mr
form (Fig. 6c). Based on both antibody epitope mapping and
biochemical profiles of CXCR4 degradation, it can be assumed that the
CXCR4 amino terminus bears a primary cleavage site for proteolysis by
the PMN serine proteinase elastase. Although our findings do not
formally rule out the existence of discrete, alternative cleavage
sites, they exclude the possibility of extensive CXCR4 degradation
following proteolysis of other extracellular domains of the receptor.
The consequences of CXCR4 degradation on the binding of SDF-1 were
next investigated in Jurkat cells. Cells were exposed to purified LE (3 µM), and cleavage of CXCR4 was monitored by antibody
labeling followed by flow cytometry analysis (Fig.
7a). As expected, LE-treated
Jurkat cells lost the amino-terminal 6H8 epitope, whereas the epitope
detected by mAb 12G5 was preserved (Fig. 7a). Cells were
then exposed to LE for 2 h at 37 °C, washed, and incubated with
MeOSuc-AAPV-CMK to inactivate any residual proteolytic activity before
addition of iodinated SDF-1. Our findings show that LE-cleaved CXCR4
is profoundly impaired in its capacity to interact with SDF-1 (Fig.
7b). Lack of specific binding by the digested receptor can
be explained by cleavage of the SDF-1-binding motif at the CXCR4
amino terminus. This conclusion is reinforced by maintenance of
SDF-1 binding to CXCR4 (Fig. 7b) and receptor recognition
by mAb 6H8 (Fig. 7a, solid square) when LE was
previously inactivated by MeOSuc-AAPV-CMK. This result excludes the
possibility that lack of high affinity interaction of SDF-1 with
cleaved CXCR4 was due to competition or steric hindrance mediated by
interaction of LE with CXCR4.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 7.
Binding of
125I-SDF-1 to LE-treated CXCR4+
Jurkat cells. a, Jurkat cells (2 × 106 cells/ml) were pretreated with different concentrations
of purified human LE for 2 h at 37 °C. Proteolytic reactions
were then stopped with 3 µM MeOSuc-AAPV-CMK before
addition of anti-CXCR4 mAbs 12G5 (open circles) or 6H8
(solid circles). The solid square corresponds to
cells exposed to LE pretreated with 3 µM MeOSuc-AAPV and
subsequently labeled with mAb 6H8. Flow cytometry was performed as
described in the legend to Fig. 6. Results are representative of three
independent experiments. b, Jurkat cells were either left
untreated (solid circles), exposed to 3 µM
purified LE for 2 h at 37 °C (open circles), or to
LE previously inhibited with 3 µM MeOSuc-AAPV-CMK
(open squares). Thereafter, cells were incubated with 0.25 nM 125I-SDF-1 and competed with unlabeled
SDF-1 for 1 h at 4 °C. Incubations were terminated by
centrifugation at 4 °C, and cell pellets were washed twice before
counting the associated radioactivity. Nonspecific binding was
determined in the presence of 1 µM of unlabeled SDF-1.
Results are means ± S.E. of six independent experiments.
Non-linear analysis of the competition curves yielded IC50
binding values as follows: untreated cells (solid squares),
6.6 ± 0.23 nM; MeOSuc-AAPV-CMK inhibition (open
squares), 5.7 ± 0.4 nM; LE-treated cells
(open circles), unfitted curves.
|
|
| |
DISCUSSION |
Our findings show an unprecedented proteolytic inactivation of
both a chemokine and its corresponding receptor. Our results demonstrate that limited degradation by leukocyte proteinases, restricted to the amino-terminal domains of both SDF-1 and CXCR4, abrogates their interaction and function.
Among leukocyte proteinases, we identify LE as the predominant
enzymatic activity responsible for limited proteolysis of the amino-terminal domains of both SDF-1 and CXCR4. As concerns
SDF-1, our conclusions are supported by manifold evidence. First,
incubation with purified human LE led to rapid degradation of the
SDF-1 amino terminus, detectable at LE concentrations in the low
nanomolar range. Second, as observed for the purified enzyme, the
effect of proteinases released by leukocytes is manifested rapidly and affects the amino terminus of SDF-1. The irreversible, highly specific leukocyte elastase inhibitor MeOSuc-AAPV-CMK (53), a
chloromethyl ketone peptide devoid of known activity on other proteinases, abolished degradation of SDF-1 exposed to leukocyte proteinases. Third, ESI-MS analysis of SDF-1 exposed to either PBMC-SN or PMN-SN revealed a major degradation product lacking the
three amino terminus residues
Lys1-Pro2-Val3, identical to that
observed after incubation of the chemokine with purified LE. Fourth,
only the uncleaved form of SDF-1 was identified by ESI-MS analysis
when the chemokine was exposed to leukocyte proteinases in the presence
of the LE inhibitor, MeOSuc-AAPV-CMK.
A previous report (35) identified CG as the enzyme responsible for the
cleavage of the amino-terminal domain of SDF-1 when exposed to
lymphocyte-associated proteinases. CG removes the five first residues
of the chemokine by hydrolyzing the peptide bond between
Leu5 and Ser6. However, in our hands, complete
inhibition of SDF-1 degradation by the specific LE inhibitor
together with ESI-MS analysis of the fragment generated from SDF-1,
i.e. SDF-1-(4-67), unambiguously identified LE among
leukocyte proteinases as the enzyme accounting for cleavage and
inactivation of SDF-1. The prominent role of CG reported by Delgado
et al. (35) could be due to the fact that the search for
SDF-1 degradation was restricted to detergent-solubilized proteins
extracted from fractionated lymphocyte membranes. The amino peptidase
DPP IV/CD26 isolated from lymphoblastoid T cells (63) has been shown to
remove Lys1-Pro2 residues from SDF-1.
However this cleavage was not observed in our experiments even when the
chemokine was exposed either to resting or phytohemagglutinin-blasted
PBMC (data not shown), a treatment known to reinforce the expression of
DPP IV/CD26 on PBMC (34). Low amounts of DPP IV/CD26 and/or the slow
degradation kinetics of SDF-1 shown by DPP IV/CD26 (63) may explain
the undetectable DPP IV/CD26-mediated degradation of SDF-1 when
exposed to primary leukocytes.
Purified matrix metalloproteinases (MMP) MMP-2 and MMP-9 have been
shown to cleave SDF-1 between Ser4 and Leu5
residues (64).3 Following PMN
degranulation induced by fMLP, we detected the presence of MMP-9 in the
PMN-SN (data not shown). Incubation of SDF-1 with either PBMC- or
PMN-secreted proteinases did not generate the expected SDF-1-(5-67)
fragment, and preincubation with 1,10-phenanthroline, a potent
inhibitor of MMP, failed to prevent the amino-terminal degradation of
SDF-1 (Fig. 1b).3 The predominant
accumulation of inactive, MMP-9 precursor forms in PMN-SN (data not
shown) and/or the presence of natural MMP inhibitors (65) may explain
the absence of SDF-1 proteolysis by leukocyte-secreted MMP.
In this report we show that the discrete proteolysis of the SDF-1
amino terminus by LE removes a domain encoding the critical Lys1-Pro2 residues involved in the activation
of CXCR4 (40) and generates an inactive chemokine. Cleavage of the
receptor activation motif of the chemokine is likely facilitated by the
exposed backbone of the disordered amino-terminal region, which
contains the binding site 2 (aa 1-8) for CXCR4. It has been
reported that a synthetic SDF-1 lacking the three amino-terminal
residues (Lys1-Pro2-Val3) removed
by LE proteolysis presented a 100-fold decrease affinity for CXCR4
(40). Thus, it can be concluded that LE generates an SDF-1 fragment
lacking both the key motifs for both binding and activation of CXCR4.
Besides the inability to trigger CXCR4 activation, LE-truncated
SDF-1 fails to prevent HIV-1 Env-mediated cell-to-cell fusion, a
phenomenon that requires both occupancy and endocytosis of CXCR4.
SDF-1 shows a constitutive expression in many tissues including
anatomical sites where HIV replication takes place (24, 66-68). It is
reasonable to postulate that LE-mediated inactivation of SDF-1 would
result in the damage of the natural SDF-1 barrier that likely oppose
spread of X4 HIV isolates.
Modification of G-protein coupled receptors by proteolysis has been
noted previously. The proteinase-activated receptors (PAR) are the
paradigm of this type of receptors (60). Typically, excision of part of
their amino-terminal regions by thrombin (69) and other serine
proteinases (70, 71) generates a tethered ligand with agonist capacity
that ultimately activates the receptor. However, if activation is the
physiological consequence of PAR proteolysis, proteinases can also
inactivate the PAR upon cleavage and deactivate the tethered ligand
(60). Interestingly, this is the case for LE that cleaves the thrombin
receptor PAR-1 at peptide bonds other than the thrombin-activating site
and removes the tethered ligand domain, thus preventing subsequent
activation of the receptor (72). Similarly, proteinases accumulated in PMN-SN cleaves the CXCR4 amino terminus as indicated by the
disappearance of both the epitope recognized by mAb 6H8 (aa 22-25)
(38) and the characteristic 40-kDa form of the protein corresponding to the mature receptor. A similar pattern of CXCR4 excision was obtained with purified human LE. Moreover, the blockade of the cleavage of CXCR4
by the specific inhibitor MeOSuc-AAPV-CMK confirmed the predominant
role played by the PMN-secreted LE in the inactivation of the receptor.
LE cleaves peptide bonds preferentially after Val and Ala residues,
although alternatively it can proteolyze Thr-X or
Ile-X bonds (52, 72, 73). Several putative sites for
LE-cleavage are encoded in the amino-terminal domain of the CXCR4
receptor among which hydrolysis at Ala34 or
Ile39 is compatible with the phenotypic pattern of
degradation that we observed. Cleavage of the
TOP
ABSTRACT
INTRODUCTION
