This information is current as
of June 16, 2017.
Structure and Function of the
Glycosaminoglycan Binding Site of
Chemokine Macrophage-Inflammatory
Protein-1 β
Witte Koopmann, Chandrika Ediriwickrema and Michael S.
Krangel
J Immunol 1999; 163:2120-2127; ;
http://www.jimmunol.org/content/163/4/2120
Subscription
Permissions
Email Alerts
This article cites 39 articles, 21 of which you can access for free at:
http://www.jimmunol.org/content/163/4/2120.full#ref-list-1
Information about subscribing to The Journal of Immunology is online at:
http://jimmunol.org/subscription
Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts
The Journal of Immunology is published twice each month by
The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 1999 by The American Association of
Immunologists All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
References
Structure and Function of the Glycosaminoglycan Binding
Site of Chemokine Macrophage-Inflammatory Protein-1b1
Witte Koopmann,2,3 Chandrika Ediriwickrema,2 and Michael S. Krangel4
T
he chemokines represent a family of cytokines that play
critical roles in orchestrating the inflammatory response
by promoting leukocyte adhesion, activation, and chemotaxis (1–3). All chemokines are related in amino acid sequence and
are thought to share the same basic tertiary structural framework,
in which a b-sheet composed of three antiparallel b-strands is
overlaid by a carboxyl-terminal a-helix. Four structurally distinct
subfamilies of chemokines are currently recognized. The largest of
these, the b or CC chemokines (including MIP-1a,5 MIP-1b,
RANTES, MCP-1, MCP-2, MCP-3, I-309, and others), are distinguished by the presence of two adjacent cysteine residues that
tether the amino terminus to the framework. In the a or CXC
chemokine family (including IL-8, gro, PF4, inflammatory protein10, and others), these two cysteines are separated by a single
amino acid. Two more recently defined chemokine subfamilies are
each currently composed of only a single member. The so-called C
family (lymphotactin) displays only one of the two amino-terminal
cysteines, whereas the CX3C family (fractalkine) displays three
amino acids separating the amino-terminal cysteines, and is tethered to the membrane via a long protein stalk.
Department of Immunology, Duke University Medical Center, Durham, NC 27710
Received for publication January 21, 1999. Accepted for publication May 27, 1999.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1
This work was supported by National Institutes of Health Grant AI37548.
2
W.K. and C.E. contributed equally to this work.
3
Current address: Department of Cell and Molecular Biology, Astra Draco AB, Lund,
Sweden.
4
Address correspondence and reprint requests to Dr. Michael Krangel, Department of
Immunology, Duke University Medical Center, P.O. Box 3010, Jones Building,
Room 313, Research Drive, Durham, NC 27710. E-mail address: krang001@
mc.duke.edu
5
Abbreviations used in this paper: MIP, macrophage-inflammatory protein; GAG,
glycosaminoglycan; hu, human; MCP, monocyte-chemotactic protein; PF4, platelet
factor 4; SEAP, secreted alkaline phosphatase.
Copyright © 1999 by The American Association of Immunologists
Chemokines interact with a large family of seven-transmembrane-spanning, G protein-coupled receptors on target cells (1, 3);
additionally, they interact with glycosaminoglycans (GAGs) on
cell surfaces, in the extracellular matrix, and in exocytic granules.
Chemokines have been shown to bind to purified subfractions of
heparin in vitro (4), as well as to naturally occurring GAGs such
as heparan sulfate and chondroitin sulfate in the extracellular matrix in vitro (5)6, and on the surface of endothelial cells both in
vitro (6 – 8) and in vivo (9). Furthermore, chemokines have been
found to be stored in and released from platelet a-granules (10, 11)
and T lymphocyte cytolytic granules (12) in association
with GAGs.
The ability of chemokines to bind to GAGs is thought to be
critical for chemokine biology. It has been proposed that the immobilization of chemokines by GAGs forms stable, solid-phase
chemokine foci and gradients necessary to direct leukocyte trafficking in vivo (13, 14). Consistent with this notion, immobilized
IL-8 has been shown to promote neutrophil migration both in vitro
(15) and in vivo (9). Furthermore, GAG-immobilized MIP-1b and
RANTES have been shown to promote leukocyte adhesion in vitro
(5, 14). GAG binding may also influence chemokine structure and
activity in other ways. For example, the binding of IL-8, MIP-1a,
RANTES, and MCP-1 to GAGs has been shown to promote their
multimerization in vitro (8). Binding of chemokines to cell surface
GAGs may also serve to increase their effective local concentration, and consequently increase their binding to cell surface receptors (8). Additionally, GAG binding could potentially influence chemokine t1/2 in vivo. As chemokines vary in the affinity (and perhaps
the specificity) with which they interact with GAGs (4), a detailed
understanding of chemokine-GAG interactions may be critical to appreciate functional distinctions among chemokines that may have
been judged by other criteria to be functionally redundant.
The molecular determinants of chemokine binding to GAGs
have been studied in several instances. The a chemokines carry a
highly basic, amphipathic carboxyl-terminal a-helix. In the cases
of IL-8 and PF4, this structure has been shown to be important for
binding to acidic GAGs (16, 17, 40). Additional basic residues
0022-1767/99/$02.00
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
The ability of chemokines to bind to glycosaminoglycans (GAGs) on cell surfaces and in the extracellular matrix is thought to play
a crucial role in chemokine function. We investigated the structural basis for chemokine binding to GAGs by using in vitro
mutagenesis to identify amino acids of chemokine macrophage-inflammatory protein-1b (MIP-1b) that contribute to its interaction with the model GAG heparin. Among six basic residues that are organized into a single basic domain in the folded MIP-1b
monomer, three (R18, K45, and R46) were found to contribute significantly to heparin binding. Of these, R46 was found to play
a dominant role, and proved essential for the interaction of MIP-1b with both heparin and heparan sulfate in physiological salt.
The results of this mutational analysis have implications for the structure of the MIP-1b-heparin complex, and a comparison of
these results with those obtained by mutational analysis of the MIP-1a-heparin interaction suggests a possible structural difference
between the MIP-1b-heparin and MIP-1a-heparin complexes. To determine whether GAG binding plays an important role in
receptor binding and cellular activation by MIP-1b, the activities of wild-type MIP-1b and R46-substituted MIP-1b were compared in assays of T lymphocyte chemotaxis. The two proteins proved equipotent in this assay, arguing that interaction of MIP-1b
with GAGs is not intrinsically required for functional interaction of MIP-1b with its receptor. The Journal of Immunology, 1999,
163: 2120 –2127.
The Journal of Immunology
Materials and Methods
Molecular cloning of wild-type huMIP-1b and huMIP-1b-SEAP
huMIP-1b full-length coding region cDNA was cloned from PHA/PMAstimulated HMC-1 human mast cells (22, 23) by RT-PCR using a primer
pair derived from the 59 and 39 ends of the published MIP-1b sequence
(24). Primers included EcoRI sites to facilitate cloning. The sequences of
the primers used are: 59-CCCGAATTCCACCAATACCATGAAGC-39
(sense) and 59-CACGAATTCAGCTCAGTTCAGTTCCAGG-39 (antisense). The cDNA, which included 11 bp of 59 untranslated sequence, was
subcloned into the mammalian expression vector pCAGGS (25, 26) and
sequenced in its entirety using Sequenase (United States Biochemical,
Cleveland, OH). The sequence obtained was identical to the published
huMIP-1b sequence Act-2 (24) with the exception of an S to G substitution
at position 70. This substitution has previously been reported for another
MIP-1b isoform (27). pDREF-hyg-MIP-1b-SEAP and pDREF-hygMIP-1b R46A-SEAP, which encode fusion proteins in which MIP-1b is
linked via a five-amino-acid carboxyl-terminal linker to secreted alkaline
phosphatase, were produced as described.6
MIP-1 b R46A antisense, 59-TTGCTTGCCTGCTTTGGTTTGG-39;
MIP-1b K48A sense, 59-CAAAAGAGGCGCACAAGTCTGCG-39;
MIP-1b K48A antisense, 59-CGCAGACTTGTGCGCCTCTTTTG-39.
PCR products were subcloned into pCAGGS and sequenced in their
entirety using Sequenase (Amersham, Arlington Heights, IL).
Transient transfection and metabolic labeling
Transient transfection of COS-P fibroblasts with wild-type and mutant
pCAGGS-MIP-1b constructs and metabolic labeling were performed
exactly as described (19). Transient transfection of 293/EBNA-1 cells
with pDREF-hyg-SEAP constructs was performed as previously described (29).
Heparin-Sepharose chromatography
Supernatants containing metabolically labeled wild-type or mutant MIP-1b
were mixed with supernatants containing MIP-1b-SEAP fusion protein and
chromatographed on 1 ml Hi-trap heparin columns (Pharmacia Biotech,
Piscataway, NJ), as described (19). The NaCl gradient used for elution was
0 – 800 mM; salt concentrations in individual fractions of a sample gradient
were determined using a conductivity meter (VWR Model 1050). Radioactivity in each fraction was determined by liquid scintillation counting
using an LKB 1209 RackBeta scintillation counter, and by SDS-PAGE
(30), followed by fluorography (31). Alkaline phosphatase activity was
determined by incubation with 1 mg/ml p-nitrophenyl phosphate at
37°C and measurement of absorbance at 405 nm using an Anthos
Labtech HT-2-2001 plate reader. For the experiment shown in Fig. 4,
radiolabeled chemokines were first partially purified by chromatography on Hi-trap Q-Sepharose columns (Pharmacia Biotech). Peak QSepharose eluate fractions were identified by SDS-PAGE and fluorography, pooled, and applied to 1 ml Hi-trap heparin columns in buffer
containing 150 mM NaCl. Columns were washed in the same buffer,
and bound chemokines were eluted with a linear 150 –500 mM NaCl
gradient. Flow-through and eluate fractions were pooled and analyzed
by SDS-PAGE and fluorography.
Purification of wild-type and mutant MIP-1b proteins
Unlabeled wild-type and mutant MIP-1b proteins were purified from supernatants of transiently transfected COS-P cells by sequential chromatography on Q-Sepharose and heparin-Sepharose columns. Following transfection, cells were cultured for 48 h, as described (19). The plates were then
washed twice in PBS and the medium was replaced with phenol red-free
DMEM (Life Technologies, Gaithersburg, MD). Cells were cultured for an
additional 24 h, and the supernatants were collected. Supernatants (200 ml)
were applied to 5 ml Hi-trap Q anion-exchange columns (Pharmacia Biotech) previously equilibrated in buffer B (50 mM Tris-HCl, pH 8, 5 mM
EDTA) at a flow rate of 0.5 ml/min. Columns were washed with 20-column
volumes of buffer B, and chemokines were eluted with a linear 0 –500 mM
NaCl gradient in buffer B. Peak chemokine-containing fractions were identified by SDS-PAGE and silver staining and pooled. The partially purified
chemokines were diluted to 50 mM NaCl in buffer B and applied to 1 ml
Hi-trap heparin columns. Columns were washed with 20-column volumes
of buffer, and bound chemokines were eluted with linear NaCl gradients, as
described above. Peak fractions were again identified by SDS-PAGE and
silver staining, pooled, dialyzed extensively against PBS at 4°C (3000
MWCO), aliquoted, and stored at 270°C. Chemokine concentrations were
determined as described (19). Chemokine purity was estimated to be
.95% by silver staining (32).
Site-directed mutagenesis
Solid-phase heparan sulfate-binding assay
Mutagenesis was performed by extension overlap amplification (28) using
400 ng of pCAGGS-MIP-1b as template. Note that the numbering refers to
amino acid residues of the mature, secreted protein. The sequences of the
mutagenic primers used are as follows: MIP-1b R18A sense, 59-TTACACC
GCGGCAAAGCTTCCTC-39; MIP-1b R18A antisense, 59-GAGGAAGCT
TTGCCGCGGTGTAA-39; MIP-1b K19A sense, 59-CACCGCGAGGGCA
CTTCCTCGCA-39; MIP-1b K19A antisense, 59-TGCGAGGAAGTG
CCCTCGCGGTG-39; MIP-1b R22A sense, 59-GAAGCTTCCTGCAAAC
TTTGTGG-39; MIP-1b R22A antisense, 59-CCACAAAGTTTGCAGG
AAGCTTC-39; MIP-1b K45A sense, 59-ATTCCAAACCGCAAGA
GGCAAG-39; MIP-1b K45A antisense, 59-CTTGCCTCTTGCGGTTTG
GAAT-39; MIP-1b R46A sense, 59-CCAAACCAAAGCAGGCAAGCAA-39;
Wild-type and R46A mutant MIP-1b-SEAP fusion proteins were partially
purified from supernatants (15 ml) of transiently transfected 293/EBNA-1
cells by chromatography over 1 ml Hi-trap Q anion-exchange columns
developed with a linear 0 –1 M NaCl gradient. Peak chemokine-containing
eluate fractions were identified by alkaline phosphatase assay and were
dialyzed against 20 mM Tris-HCl, pH 8. A solid-phase heparan sulfatebinding assay was modeled after a previously described solid-phase heparin-binding assay (33). The wells of a polystyrene 96-well plate (Corning,
Acton, MA) were coated with 50 ml of 125 mg/ml bovine kidney heparan
sulfate (Sigma H7640) or porcine kidney medulla heparan sulfate (34)
(kind gift of Dr. Robert Linhardt, University of Iowa, Iowa City, IA, and
Dr. Toshihiko Toida, Chiba University, Chiba, Japan) by incubation at
37°C for 16 h. The wells were rinsed twice with 20 mM Tris-HCl, pH 8,
containing 0.05% (w/v) BSA, followed by a third rinse that was allowed to
stand for 2 h at 23°C before aspiration. A total of 50 ml of 5 nM chemokine-SEAP fusion protein in 20 mM Tris-HCl, pH 8, and varying concentrations of NaCl were introduced into the wells for a 2-h incubation at
6
D. D. Patel, W. Koopmann, T. Imai, L. P. Whichard, O. Yoshie, and M. S. Krangel.
A subset of chemokines bind to extracellular matrix and mast cells in rheumatoid
arthritis synovium via electrostatic interactions with glycosaminoglycans. Submitted
for publication.
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
within the PF4 framework have been found to interact with GAGs
as well (18, 40). Among b chemokines, basic residues within the
MCP-1 a-helix were found to be important for GAG binding (41).
However, we (19) and others (20) previously noted that other b
chemokines do not have highly basic a-helices, and that other
regions of these molecules must be important determinants of
GAG binding. We used in vitro mutagenesis to test the roles of
individual basic amino acids of huMIP-1a, and identified three
noncontiguous amino acids (R18, R46, and R48) that were each
essential for heparin binding (19). A fourth residue, K45, was
found to contribute as well. Similarly, a double mutation of K45
and R46 was shown to abrogate heparin binding of murine MIP-1a
in another study (20). Based on a modeled MIP-1a structure, all
four basic residues are predicted to lie on one face of the molecule,
with the side chains of R18, R46, and R48 aligned and protruding
prominently from the surface.
The minimal GAG binding motif found in huMIP-1a is conserved in other b chemokines, but may be augmented by additional
basic residues. As an example, huMIP-1b contains basic residues
at positions 18, 45, 46, and 48, and, in addition, at positions 19 and
22. The solved MIP-1b structure (21) indicates that the side chains
of all of these residues lie in proximity on one face of the molecule,
leading to the prediction that MIP-1b uses an expanded binding
motif to bind more tightly to GAGs. Consistent with this idea, we
have found that MIP-1b interacts more tightly with heparin than
MIP-1a, as judged by the concentration of salt required to disrupt
the interaction. Nevertheless, mutagenesis experiments indicate
that only a subset of these residues contributes detectably to GAG
binding. Interestingly, functional analysis of one MIP-1b mutant
indicates that, as for MIP-1a, GAG binding is not required for
functional interaction of MIP-1b with its receptor.
2121
2122
A GLYCOSAMINOGLYCAN BINDING SITE IN CHEMOKINE MIP-1b
Chemotaxis assay
The ability of wild-type and mutant MIP-1b to stimulate chemotaxis of
activated human T lymphocytes was assessed using a 48-well Boyden
chamber assay. Human PBMC were isolated with LSM (Organon Teknika,
Durham, NC), washed four times in serum-free RPMI 1640 (Life Technologies), and resuspended in RPMI 1640 containing 10% FCS. Adherent
cells were removed by two rounds of adherence to plastic tissue culture
flasks for 1 h at 37°C. Nonadherent cells were cultured for 16 h at 37°C in
tissue culture plates coated with a culture supernatant containing anti-CD3
Ab OKT3 (kind gift of Dr. C. Doyle, Duke University, Durham, NC). Cells
were harvested, washed, and resuspended in RPMI 1640 supplemented to
20 mM HEPES-NaOH, pH 7.4, and 2% (w/v) BSA. Purified chemokines
in the same medium were added in triplicate to the bottom wells of the
chemotaxis chamber (NeuroProbe, Cabin John, MD) in a volume of 28 ml.
The wells were overlaid with a 5 mM polyvinylpyrrolidone-free polycarbonate chemotaxis filter (NeuroProbe) that had been precoated for 45
min at 37°C with 5 mg/ml type IV collagen (Sigma, St. Louis, MO), as
described (14). Cells (5 3 104/well) were added to the top wells and
allowed to migrate for 2 h at 37°C. Filters were stained with Diff-Quik
(Baxter Scientific Products, McGaw Park, IL), and cells migrated to the
underside of the filter were counted in five randomly selected high power
(3400) fields.
FIGURE 1. Basic amino acids of chemokine MIP-1b. A space-filling
model of the MIP-1b monomer was created using the program WebLab
Viewer (Molecular Simulations) from the coordinates of MIP-1b (21). Side
chains of the basic residues have a darker fill and are labeled.
23°C. Wells were rinsed three times with 20 mM Tris-HCl, pH 8, containing 0.05% (w/v) BSA, and bound fusion proteins were quantified by
alkaline phosphatase assay.
huMIP-1b contains six basic residues, at positions 18, 19, 22, 45,
46, and 48. Based on the MIP-1b structure (21), all lie in proximity
on one face of the molecule (Fig. 1). To determine which of the
basic amino acids contributes to the ability of MIP-1b to bind to
GAGs, we examined the interaction of wild-type MIP-1b and single amino acid mutants of MIP-1b with heparin-Sepharose. We
expressed wild-type MIP-1b and alanine substitution mutants
R18A, K19A, R22A, K45A, R46A, and K48A using a COS cell
transient expression system, and harvested the supernatants of metabolically labeled cells. SDS-PAGE analysis of unfractionated
COS cell supernatants revealed a 10- to 12-kDa species present in
the supernatants of transfected cells that was not present in the
FIGURE 2. Evaluation of huMIP-1b expression and heparin-Sepharose chromatography by SDS-PAGE. A, Supernatants of metabolically labeled
COS-P cells that had been mock transfected or transfected with wild-type and mutant MIP-1b expression constructs were analyzed. B, Fractions 3–14 of
the heparin-Sepharose chromatographic profile of wild-type MIP-1b shown in Fig. 3 were analyzed. C, Peak chemokine-containing fractions of each of the
heparin-Sepharose chromatographic profiles shown in Fig. 3 were analyzed. D, Fractions 3–14 of the heparin-Sepharose chromatographic profile of MIP-1b
R46A shown in Fig. 3 were analyzed. Analysis in each case was by 17% SDS-PAGE, followed by fluorography.
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
Results
The Journal of Immunology
2123
supernatants of mock-transfected cells (Fig. 2A). As observed previously in similar transient expression analysis of MIP-1a (19),
MIP-1b represented the most abundant radiolabeled species in
these supernatants. We also expressed wild-type MIP-1b in the
form of a nonradiolabeled fusion protein with SEAP by transient
transfection of 293/EBNA-1 cells (29),6 to be used as an internal
control for chromatography experiments.
Supernatants of metabolically labeled and unlabeled cells
were diluted to reduce the ionic strength, mixed, and applied to
heparin-Sepharose columns, as described previously (19).
Bound chemokines were then eluted by application of a linear
salt gradient, and were identified in eluate fractions by determination of total radioactivity, by SDS-PAGE followed by fluorography, and by measurement of alkaline phosphatase activity. Co-chromatography of radiolabeled wild-type MIP-1b with
unlabeled wild-type MIP-1b-SEAP fusion protein yielded
peaks that essentially coeluted at 400 mM NaCl (Fig. 3). SDSPAGE analysis across the profile confirmed that the majority of
radioactivity in the peak fractions represented MIP-1b (Fig.
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
FIGURE 3. Heparin-Sepharose chromatography of wild-type and mutant
MIP-1b proteins. Culture supernatants
containing metabolically labeled wildtype or mutant MIP-1b were mixed
with supernatants containing unlabeled
MIP-1b-SEAP fusion protein and were
analyzed by heparin-Sepharose chromatography. Eluate (but not initial flowthrough and wash) fractions are shown.
The radioactivity profile is indicated by
the solid line. Alkaline phosphatase activity is indicated by the bars. The salt
gradient is indicated by a dashed line in
the wild-type profile.
2B). This analysis confirmed that the SEAP moiety neither contributes to nor interferes with heparin binding by MIP-1b. By
comparison with MIP-1b, MIP-1a elutes at 250 mM NaCl (19).
The requirement for significantly higher salt concentrations to elute
MIP-1b is consistent with a stronger interaction, perhaps mediated by
additional basic residues unique to MIP-1b.
Each of the six alanine-substituted MIP-1b mutants was then
co-chromatographed on heparin-Sepharose along with MIP-1bSEAP as an internal control. Alanine substitutions in mutants
K19A, R22A, and K48A had no apparent effect on binding and
elution from heparin-Sepharose (Fig. 3). However, the substitutions in mutants R18A and K45A resulted in peaks of radioactivity
that eluted at salt concentrations (250 –300 mM) that were reduced
as compared with the wild-type fusion protein. For each of the
mutants, the peak radioactive fraction was confirmed to include
MIP-1b as a predominant component (Fig. 2C). The profile for
mutant R46A was complex in that two peaks of radioactivity were
identified, one eluting at very low salt concentrations (less than
100 mM), and one eluting at a position only slightly shifted from
2124
A GLYCOSAMINOGLYCAN BINDING SITE IN CHEMOKINE MIP-1b
the wild-type MIP-1b fusion protein. For mutant R46A, SDSPAGE analysis across the chromatographic profile (Fig. 2D) indicated that the first radioactive peak eluting from the column represented MIP-1b, whereas the second peak represented unrelated
background proteins that were also represented, albeit at lower
levels, in the wild-type elution profile (Fig. 2B). Additional experiments indicated that the relatively low ratio of MIP-1b R46A to
background proteins reflected the fact that a portion of MIP-1b
R46A flowed through the heparin-Sepharose column without binding (data not shown). We conclude that alanine substitutions at
positions R18 and K45 have small but readily detectable effects on
the interaction of MIP-1b with heparin, whereas an alanine substitution at position R46 has a dramatic effect on this interaction.
To further characterize heparin binding by MIP-1b R46A, we
generated highly enriched preparations of metabolically labeled
wild-type MIP-1b and MIP-1b R46A by ion-exchange chromatography on Q-Sepharose. Analysis of the Q-Sepharose eluates by
SDS-PAGE revealed that radioactivity was associated almost exclusively with the MIP-1b proteins (Fig. 4A). The highly enriched
preparations of MIP-1b and MIP-1b R46A were then applied to a
heparin-Sepharose column in a buffer containing 150 mM NaCl,
and the column was developed with a linear salt gradient. Under
these conditions, a small amount of radioactivity in the wild-type
preparation flowed through the column, but the vast majority
bound and was eluted at 400 mM NaCl (Fig. 4B). SDS-PAGE
analysis indicated that the eluate was exclusively MIP-1b, whereas
the flow-through consisted primarily of higher m.w. contaminants
(Fig. 4A). In contrast, MIP-1b R46A was found exclusively in the
flow-through (Fig. 4, A and B). Hence, MIP-1b R46A does not
stably interact with heparin when binding is assessed under conditions of physiological salt concentrations.
To ask whether the R46A mutation abrogates interaction between MIP-1b and GAGs that occur naturally on cell surfaces and
in the extracellular matrix, we assessed the binding of MIP-1b and
MIP-1b R46A to plastic-immobilized heparan sulfate in vitro (Fig.
5). Because sulfation levels vary among heparan sulfate chains
isolated from different sources, we analyzed binding to both bovine kidney heparan sulfate, which displays a relatively low degree
of sulfation, and porcine kidney medulla heparan sulfate, which
displays a high degree of sulfation (34). Partially purified preparations of MIP-1b-SEAP and MIP-1b R46A-SEAP fusion proteins
were incubated with the immobilized heparan sulfate preparations
at several salt concentrations, and binding was assessed by alkaline
phosphatase assay. In this binding assay, the signals detected in the
presence of 0.75 M NaCl represent nonspecific, salt-insensitive
binding. Relative to this binding, wild-type MIP-1b bound well to
both bovine kidney and porcine kidney medulla heparan sulfate in
subphysiological salt (30 mM), but only bound to the more highly
sulfated porcine kidney medulla heparan sulfate in physiological
salt. Importantly, binding of MIP-1b R46A was undetectable at
both physiological and subphysiological salt concentrations, indicating that the R46A mutation abrogates the ability of MIP-1b to
bind to a physiologically relevant GAG ligand.
To compare the in vitro biological activities of wild-type
MIP-1b and MIP-1b R46A, we purified unlabeled MIP-1b and
MIP-1b R46A from the supernatants of transiently transfected
cells by ion-exchange chromatography, followed by heparinSepharose chromatography with loading in low salt, as described
in Materials and Methods. The MIP-1b preparations obtained by
this approach were highly purified (Fig. 6A). The wild-type (as
well as a commercially available) MIP-1b preparation failed to
induce significant increases in cytoplasmic calcium concentration
in either monocytes or lymphocytes, and displayed a relatively
poor chemotactic index in assays of monocyte chemotaxis (data
not shown). Therefore, we assessed the ability of the wild-type and
mutant MIP-1b preparations to stimulate the chemotaxis of activated human peripheral blood T lymphocytes. Both preparations
displayed good chemotactic indices and yielded typical bellshaped curves with maximal migration at about 10 ng/ml MIP-1b
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
FIGURE 4. Heparin-Sepharose chromatography of wild-type MIP-1b
and MIP-1b R46A with loading in physiological salt. Metabolically labeled chemokines that were initially partially purified by Q-Sepharose
chromatography were subsequently analyzed by heparin-Sepharose chromatography. A, Q-Sepharose-purified chemokine loaded (L) onto heparinSepharose, pooled heparin-Sepharose flow-through (FT) fractions, and
pooled heparin-Sepharose eluate (E) fractions (equivalent fractions from
the wild-type and R46A profiles) were analyzed by 15% SDS-PAGE, followed by fluorography. B, Heparin-Sepharose chromatographic profiles for
wild-type MIP-1b (filled circles) and MIP-1b R46A (open circles) are
superimposed. The salt gradient is indicated by a dashed line.
FIGURE 5. Heparan sulfate binding by wild-type MIP-1b and MIP-1b
R46A. Partially purified chemokine-SEAP fusion proteins were incubated
with immobilized bovine kidney heparan sulfate (A) or porcine kidney
medulla heparan sulfate (B) at several salt concentrations. Binding was
assessed by alkaline phosphatase assay. Binding at 0.75 M salt reflects
salt-insensitive nonspecific binding, whereas the difference between binding at 0.75 M salt and binding at lower salt concentrations reflects saltsensitive, specific binding. The mean and SDs of triplicate determinations
are shown. Because the alkaline phosphatase assays were developed for
different lengths of time, A and B are not directly comparable with respect
to absolute levels of binding to the different forms of heparan sulfate.
The Journal of Immunology
(Fig. 6B). As the activities of the two preparations were indistinguishable by this functional assay (the slight difference between the profiles at the highest concentration tested was not
reproducible), we conclude that the ability of MIP-1b to bind to
GAGs is unnecessary for functional interaction of MIP-1b with
its receptor on T lymphocytes. We furthermore conclude that
the dramatic effect of the R46A mutation on GAG binding does
not result from any gross perturbation in the structure of the
MIP-1b protein.
Discussion
huMIP-1b includes six basic residues that are found in two segments of the linear amino acid sequence (R18, K19, R22 5 segment 1; K45, R46, K48 5 segment 2), and are organized into a
single basic domain in the folded MIP-1b monomer. Segment 2
conforms precisely to the BBXB pattern of basic residues implicated as a heparin-binding motif in other studies (35), whereas
segment 1 displays a related BBXXB pattern. Interestingly, the
basic domain in the folded MIP-1b structure takes a form that
appears related to the cationic cradle implicated in the binding of
lactoferrin (36) and fibronectin (37) to heparin. This structure consists of six basic residues arranged with three on each side of a
central cleft. For MIP-1b, residues 18 and 19 appear to line one
side of a cleft, and residues 48, 46, 45, and 22 the other side. This
organization suggests that all six residues could play significant
roles in mediating interactions between MIP-1b and heparin. Our
data indicate otherwise: R46 is apparently critical for the stable
interaction of MIP-1b with heparin, R18 and K45 play secondary
roles, and K19, R22, and K48 have roles that, if significant, are not
detectable in our assays. In contrast, mutations in all six residues
lining the cleft were found to influence the interaction of fibronectin module III-13 with heparin (37). Hence, the notion of a cationic
cradle may not be appropriate to describe the huMIP-1b heparin
binding site.
HuMIP-1a, which elutes from heparin-Sepharose at lower salt
concentrations than huMIP-1b, carries basic amino acids at positions 18, 45, 46, and 48, but lacks the basic amino acids at positions 19 and 22 that are contained within huMIP-1b. If huMIP-1b
residues K19 and R22 were responsible for the substantially different salt elution properties of huMIP-1b and huMIP-1a, we
should have detected an effect of mutations at these positions in
our assays. Furthermore, our data present an apparent paradox in
that the more avid interaction of MIP-1b with heparin seems to be
mediated by fewer basic residues (three rather than four). Recent
experiments by Hoogewerf et al. (8) may suggest an explanation.
These investigators found that several chemokines bound to immobilized heparin as tetramers, whereas MIP-1a bound as a dimer.
If MIP-1b (which was not examined in that study) were to bind as
a tetramer, binding of the MIP-1b tetramer to GAGs would have
contributions from 12 basic residues (3 per monomer), whereas the
MIP-1a dimer would have contributions from 8 basic residues (4
per monomer), providing an adequate explanation for the distinct
heparin-binding properties of the two molecules.
Why might MIP-1b residues R18, K45, and R46 be selectively
involved in the interaction with heparin? Residues R18 and R46
are notable in that their side chains are aligned with each other and
protrude directly out from the surface of the MIP-1b monomer in
a fashion that can easily be envisioned to interact with heparin
(Fig. 1). Although the side chain of K45 is not aligned similarly,
R18, K45, and R46 do form a relatively focused basic patch on the
surface of the molecule. Interestingly, in a MIP-1b dimer, the basic
patches of the two monomers face away from each other on opposite ends of the molecule. If MIP-1b binds to heparin as a multimer (either a dimer or tetramer), the positioning of the basic
patches would suggest that the heparin chain wraps around the
MIP-1b multimer.
Of the remaining basic residues in MIP-1b, R22 is probably not
involved in heparin binding because it extends away from the basic
patch formed by R18, R48, and K45. A clear explanation for the
apparent absence of involvement by K19 is less certain. On the
other hand, the absence of a significant role for K48 can probably
be best understood as a function of its orientation in the MIP-1b
dimer (21). Whereas R18, K45, and R46 create a pair of basic
patches on opposite ends of the dimer, the K48 residues of each
monomer face toward each other on the concave surface of the
dimer. As such, this surface of the dimer most likely plays no role
in heparin binding. Given this conclusion, it is striking that residue
R48 plays an important role in heparin binding by the highly homologous MIP-1a (19). This might be consistent with distinct quaternary structures for the MIP-1b-heparin and MIP-1a-heparin
complexes.
It has long been argued that chemokine binding to GAGs may
be important for stabilization of chemokines in the form of solidphase gradients for presentation to passing leukocytes (13, 14).
However, recent experiments have suggested that there is even
greater and more fundamental involvement of GAGs in chemokine
function. In support of this notion, enzymatic removal of cell surface GAGs results in a significant reduction in the apparent affinity
of RANTES for its cell surface receptor (8). Heparan sulfate synergizes with RANTES to inhibit HIV-1 replication in monocytes
(12). Furthermore, removal of cell surface GAGs has been shown
to block the inhibitory effect of RANTES on HIV-1 replication in
PM1 T cells (38) and to block intracellular Ca21 signaling in PBL
(39). GAG-induced RANTES oligomerization and a GAG-induced
increase in local RANTES concentration could be critical to potentiate the interaction of RANTES with its specific receptor (8), or
GAGs could represent a fundamental component of the receptorligand complex. In contrast, although an IL-8 mutant that binds
poorly to GAGs displays reduced activity in a neutrophil chemotaxis assay that is thought to depend on both receptor binding and
solid-phase immobilization, the same mutant displays normal activity in a neutrophil elastase release assay that is thought to depend on receptor binding only (9). Further, heparin binding mutants of MCP-1 stimulated normal Ca21 signaling and chemotaxis
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
FIGURE 6. Chemotactic activity of wild-type MIP-1b and MIP-1b
R46A. A, Purified wild-type MIP-1b and MIP-1b R46A were prepared as
described in Materials and Methods and analyzed by 15% SDS-PAGE,
followed by silver staining. B, Chemotaxis of activated human T lymphocytes was assessed using a Boyden chamber. The number of cells migrated
in five randomly selected high power (3400) fields is plotted. The unpaired
shaded bar represents background migration with no added chemokine.
The mean and SD of triplicate determinations are shown. Results are presented for one of two comparable experiments.
2125
2126
A GLYCOSAMINOGLYCAN BINDING SITE IN CHEMOKINE MIP-1b
Acknowledgments
We thank Dr. Dhavalkumar Patel for critical review of the manuscript, and
Drs. Robert Linhardt and Toshihiko Toida for their kind gift of porcine
kidney medulla heparan sulfate.
References
1. Rollins, B. 1997. Chemokines. Blood 90:909.
2. Baggiolini, M., B. Dewald, and B. Moser. 1997. Human chemokines: an update.
Annu. Rev. Immunol. 15:675.
3. Baggiolini, M. 1998. Chemokines and leukocyte traffic. Nature 392:565.
4. Witt, D. P., and A. D. Lander. 1994. Differential binding of chemokines to glycosaminoglycan subpopulations. Curr. Biol. 4:394.
5. Gilat, D., R. Hershkoviz, A. Mekori, I. Vlodavsky, and O. Lider. 1994. Regulation of adhesion of CD41 T lymphocytes to intact or heparinase-treated subendothelial extracellular matrix by diffusible or anchored RANTES and MIP-1b.
J. Immunol. 153:4899.
6. Rybak, M. E., M. A. Gimbrone, P. F. Davies, and R. I. Handin. 1989. Interaction of platelet factor four with cultured vascular endothelial cells. Blood
73:1534.
7. Luster, A. D., S. M. Greenberg, and P. Leder. 1995. The IP-10 chemokine binds
to a specific cell surface heparan sulfate site shared with platelet factor 4 and
inhibits endothelial cell proliferation. J. Exp. Med. 182:219.
8. Hoogewerf, A. J., G. S. V. Kuschert, A. E. I. Proudfoot, F. Borlat, I. Clark-Lewis,
C. A. Power, and T. N. C. Wells. 1997. Glycosaminoglycans mediate cell surface
oligomerization of chemokines. Biochemistry 36:13570.
9. Middleton, J., S. Neil, J. Wintle, I. Clark-Lewis, H. Moore, C. Lam, M. Auer,
E. Hub, and A. Rot. 1997. Transcytosis and surface presentation of IL-8 by
venular endothelial cells. Cell 91:385.
10. Kaplan, K. L., J. Broekman, A. Chernoff, G. R. Lesznik, and M. Drillings. 1979.
Platelet a-granule proteins: studies on release and subcellular localization. Blood
53:604.
11. Levine, S. P., L. K. Knieriem, and M. A. Rager. 1990. Platelet factor 4 and the
platelet secreted proteoglycan: immunologic characterization by crossed immunoelectrophoresis. Blood 75:902.
12. Wagner, L., O. O. Yang, E. A. Garcia-Zepeda, Y. Ge, S. A. Kalams,
B. D. Walker, M. S. Pasternack, and A. D. Luster. 1998. b-chemokines are
released from HIV-1-specific cytolytic T-cell granules complexed to proteoglycans. Nature 391:908.
13. Rot, A. 1992. Endothelial cell binding of NAP-1/IL-8: role in neutrophil emigration. Immunol. Today 13:291.
14. Tanaka, Y., D. H. Adams, S. Hubscher, H. Hirano, U. Siebenlist, and S. Shaw.
1993. T-cell adhesion induced by proteoglycan-immobilized cytokine MIP-1b.
Nature 361:79.
15. Rot, A. 1993. Neutrophil attractant/activation protein-1 (interleukin-8) induces in vitro neutrophil migration by haptotactic mechanism. Eur. J. Immunol. 23:303.
16. Maione, T. E., G. S. Gray, A. J. Hunt, and R. J. Sharpe. 1991. Inhibition of tumor
growth in mice by an analogue of platelet factor 4 that lacks affinity for heparin
and retains potent angiostatic activity. Cancer Res. 51:2077.
17. Webb, L. M. C., M. U. Ehrengruber, I. Clark-Lewis, M. Baggiolini, and A. Rot.
1993. Binding to heparan sulfate or heparin enhances neutrophil responses to
interleukin 8. Proc. Natl. Acad. Sci. USA 90:7158.
18. Mayo, K. H., E. Ilyina, V. Roongta, M. Dundas, J. Joseph, C. K. Lai,
T. E. Maione, and T. J. Daly. 1995. Heparin binding to platelet factor-4: an NMR
and site-directed mutagenesis study: arginine residues are crucial for binding.
Biochem. J. 312:357.
19. Koopmann, W., and M. S. Krangel. 1997. Identification of a glycosaminoglycanbinding site in chemokine macrophage inflammatory protein-1a. J. Biol. Chem.
272:10103.
20. Graham, G. J., P. C. Wilkinson, R. J. B. Nibbs, S. Lowe, S. O. Kolset, A. Parker,
M. G. Freshney, M. L.-S. Tsang, and I. B. Pragnell. 1996. Uncoupling of stem
cell inhibition from monocyte chemoattraction in MIP-1a by mutagenesis of the
proteoglycan binding site. EMBO J. 15:6506.
21. Lodi, P. J., D. S. Garrett, J. Kuszewski, M. L.-S. Tsang, J. A. Weatherbee,
W. J. Leonard, A. M. Gronenborn, and G. M. Clore. 1994. High-resolution solution structure of the b chemokine hMIP-1b by multidimensional NMR. Science
263:1762.
22. Butterfield, J. H., D. Weiler, G. Dewald, and G. J. Gleich. 1988. Establishment
of an immature mast cell line from a patient with mast cell leukemia. Leuk. Res.
12:345.
23. Selvan, R. S., J. H. Butterfield, and M. S. Krangel. 1994. Expression of multiple
chemokine genes by a human mast cell leukemia. J. Biol. Chem. 269:13893.
24. Lipes, M. A., M. Napolitano, K.-T. Jeang, N. T. Chang, and W. J. Leonard. 1988.
Identification, cloning, and characterization of an immune activation gene. Proc.
Natl. Acad. Sci. USA 85:9704.
25. Niwa, H., K. Yamamura, and J. Miyazaki. 1991. Efficient selection for highexpression transfectants with a novel eukaryotic vector. Gene 108:193.
26. Yoshimura, T. 1993. cDNA cloning of guinea pig monocyte chemoattractant
protein-1 and expression of the recombinant protein. J. Immunol. 150:5025.
27. Napolitano, M., W. S. Modi, S. J. Cevario, J. R. Gnarra, H. N. Seuanez, and
W. J. Leonard. 1991. The gene encoding the Act-2 cytokine: genomic structure,
HTLV-I/Tax responsiveness of 59 upstream sequences, and chromosomal localization. J. Biol. Chem. 266:17531.
28. Ho, S. N., H. D. Hunt, R. M. Horton, J. K. Pullen, and L. R. Pease. 1989.
Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77:51.
29. Imai, T., M. Baba, M. Nishimura, M. Kakizaki, S. Takagi, and O. Yoshie. 1997.
The T cell-directed CC chemokine TARC is a highly specific biological ligand
for CC chemokine receptor 4. J. Biol. Chem. 272:15036.
30. Laemmli, U. K. 1970. Cleavage of the structural proteins during the assembly of
the head of bacteriophage T4. Nature 227:680.
31. Bonner, W. M., and R. A. Lasky. 1974. A film detection method for tritiumlabelled proteins and nucleic acids in polyacrylamide gels. Eur. J. Biochem.
46:83.
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017
of THP-1 cells (41). Different chemokines may therefore be differentially dependent on GAG binding for efficient interaction with
their receptors.
We recently showed that a singly substituted MIP-1a mutant
that fails to interact with heparin displays wild-type binding to
CCR1 and wild-type Ca21 signaling (19). Although a two-aminoacid heparin-binding mutant of MIP-1a generated by Graham et al.
failed to bind to and signal through CCR1 (20), it seems likely that
in this case the mutation coincidentally disrupted either an adjacent
or distal receptor binding site. Importantly, both studies found
wild-type MIP-1a to bind normally to CCR1 on GAG-deficient
CHO cells. Thus, GAG binding is not critical for the functional
interaction of MIP-1a with its receptor.
The results presented in this work indicate that the same is true
for MIP-1b. Wild-type MIP-1b and MIP-1b R46A were indistinguishable in assays of T cell chemotaxis. From this we infer that
the two proteins interact indistinguishably with one or more cell
surface receptors on activated T cells, and that these interactions
do not depend on cell surface GAGs. Interestingly, Oravecz et al.
found that heparitinase digestion of PM1 T cells resulted in a dramatic inhibition of the antiviral effect of MIP-1b, a result that can
be interpreted to indicate that cell surface GAGs facilitate the interaction of MIP-1b with its receptor (38). Assuming that the T
cell chemotactic and antiviral effects of MIP-1b are both mediated
by CCR5, the mechanistic implications of the present study and
that of Oravecz et al. (38) are clearly distinct. It remains possible
that heparitinase treatment of cells modulates the antiviral activity
of MIP-1b by a mechanism that is more complex than that originally envisioned.
We note that our results do not necessarily imply that receptor
binding and activation by MIP-1b occur independently of GAGs
in all instances. These parameters may vary according to both the
biological assay and the receptor involved. In addition, because we
found that MIP-1b interacts preferentially with highly sulfated
GAGs, any dependence of GAGs for binding or activation may
vary according to tissue and/or cell type.
In conclusion, the work presented in this study dissects the heparin binding site of huMIP-1b, and provides evidence that the cationic cradle motif previously proposed to mediate the interaction
of GAGs with fibronectin and lactoferrin does not accurately describe the GAG binding site of MIP-1b. In addition, the description of a non-GAG-binding mutant MIP1b that stimulates chemotaxis in vitro extends our earlier findings with MIP-1a, in that the
capacity to bind to GAGs and to bind to and stimulate chemokine
receptors can be functionally uncoupled for both molecules. GAG
binding may nevertheless be critical for activities of these chemokines that depend specifically on their immobilization. Experiments designed to address the role(s) of GAG binding to chemokine biology in vivo are in progress.
The Journal of Immunology
32. Morrissey, J. H. 1981. Silver stain for proteins in polyacrylamide gels: a modified
procedure with enhanced uniform sensitivity. Anal. Biochem. 117:307.
33. Stephens, R. W., A. M. Bokman, H. T. Myohanen, T. Reisberg, H. Tapiovaara,
N. Pedersen, J. Grondahl-Hansen, M. Llinas, and A. Vaheri. 1992. Heparin binding to urokinase kringle domain. Biochemistry 31:7572.
34. Toida, T., H. Yoshida, H. Toyoda, I. Koshiishi, T. Imanari, R. E. Hileman,
J. R. Fromm, and R. J. Linhardt. 1997. Structural differences and the presence of
unsubstituted amino groups in heparan sulfates from different tissues and species.
Biochem. J. 322:499.
35. Cardin, A. D., and H. J. R. Weintraub. 1989. Molecular modeling of proteinglycosaminoglycan interactions. Arteriosclerosis 9:21.
36. Mann, D. M., E. Romm, and M. Migliorini. 1996. Delineation of the glycosaminoglycan-binding site in the human inflammatory response protein lactoferrin.
J. Biol. Chem. 269:23661.
37. Busby, T. F., W. S. Argraves, S. A. Brew, I. Pechik, G. L. Gilliland, and
K. C. Ingham. 1995. Heparin binding by fibronectin module III-13 involves six
2127
38.
39.
40.
41.
discontinuous basic residues brought together to form a cationic cradle. J. Biol.
Chem. 270:18558.
Oravecz, T., M. Pall, J. Wang, G. Roderiquez, M. Ditto, and M. Norcross. 1997.
Regulation of anti-HIV activity of RANTES by heparan sulfate proteoglycans.
J. Immunol. 159:4587.
Burns, J. M., R. C. Gallo, A. L. DeVico, and G. K. Lewis. 1998. A new monoclonal antibody, mAb 4A12, identifies a role for the glycosaminoglycan (GAG)
binding domain of RANTES in the antiviral effect against HIV-1 and intracellular
Ca21 signaling. J. Exp. Med. 188:1917.
Kuschert, G. S. V., A. J. Hoogewerf, A. E. I. Proudfoot, C. Chung, R. M. Cooke,
R. E. Hubbard, T. N. C. Wells, and P. N. Sanderson. 1998. Identification of a
glycosaminoglycan binding surface on human interleukin-8. Biochemistry 37:
11193.
Chakravarty, L., L. Rogers, T. Quach, S. Breckenridge, and P. E. Kollattukudy.
1998. Lysine 58 and histidine 66 at the C-terminal a-helix of monocyte chemoattractant protein-1 are essential for glycosaminoglycan binding. J. Biol.
Chem. 273:29641.
Downloaded from http://www.jimmunol.org/ by guest on June 16, 2017