(Received for publication, December 17, 1996, and in revised form, January 8, 1997)
From the The human immunodeficiency virus type
1 (HIV-1) requires the presence of specific chemokine receptors in
addition to CD4 to enter its target cell. The chemokine receptor CCR5
is used by macrophage-tropic strains of HIV-1, which predominate during
the asymptomatic stages of infection. Here we investigate whether the
ability of CCR5 to signal in response to its Human immunodeficiency virus (HIV-1)1
is the etiologic agent of AIDS, which results from the destruction of
CD4-positive lymphocytes in infected individuals (1-3). The entry of
HIV-1 into target cells is mediated by the viral envelope glycoproteins
(4, 5). The HIV-1 exterior glycoprotein, gp120, binds the cellular
receptor CD4 (6). CD4 expression on target cells is not sufficient for viral entry, however, and the chemokine receptors CXCR4 (previously known as HUMSTSR, LESTR, or fusin), CCR3, and CCR5 function as necessary co-receptors for the HIV-1 virus (7-12). Among these, CCR5
is thought to be especially important because primary viruses, which
infect both macrophages and T cells efficiently, use CCR5 (9).
Furthermore, individuals homozygous for a defect in CCR5 appear to be
protected from HIV-1 infection (13-15). The chemokine ligands for
CCR5, MIP-1 Chemokines are a family of small cytokines that share a common
structure containing four conserved cysteines, the first two of which
are adjacent (C-C or Here we describe mutants of CCR5 that fail to mobilize calcium
following chemokine ligation but that bind chemokine and support HIV-1
entry as well as wild-type CCR5. We also characterize a chimeric
receptor of CCR2 and CCR5 that binds MIP-1 The pHXBH10 CF2Th canine thymocytes (ATCC CRL 1430), Bing
(ATCC CRL 11554), and HEK293 cells were obtained from American Type
Culture Collection. Hela-CD4 cells were obtained from Dr. Bruce
Chesebro through the National Institutes of Health AIDS Research and
Reference Reagent Program. Cells were maintained as described
previously (9).
A single round of HIV-1 entry was
assayed as described previously (9), except that 25,000 cpm reverse
transcriptase activity of the recombinant viruses containing the ADA
and YU2 envelope glycoproteins were used per assay, and cells were
incubated with virus for 48 h. Briefly, HIV-1 virus with the
nef gene replaced by the CAT gene was used to infect cells
expressing CD4 and a chemokine receptor. Cells were lysed after
infection, and CAT activity was measured, indicating the level of
transcription from the integrated HIV-1 genome (5). In parallel to the
infection assays, anti-FLAG and anti-CCR5 antibody 5C7 were used to
quantify receptor expression by FACS analysis. 5C7 was generated
against the CCR5 receptor stably-expressed on a murine lymphocyte
line.2
HEK293 cells were transfected by the
calcium phosphate method (33) with 30 µg of plasmid DNA transiently
expressing the chemokine receptors. Cells were suspended in 10 ml of
buffer (Hanks' buffered saline solution, 25 mM HEPES, pH
7.2, 0.1% bovine serum albumin) per flask and incubated with 30 µg
of Fura-2/AM (Molecular Probes, Inc.) for 30 min at 37 °C. Cells
were then washed twice with phosphate-buffered saline and resuspended
in buffer. Calcium flux measurements in response to MIP-1 HEK293 or BING293 cells were transfected
by the calcium phosphate method with 30 µg of plasmid DNA transiently
expressing the chemokine receptors. In some cases, parallel
transfections were performed with a Changes in a
conserved aspartic residue in the second transmembrane domain have been
shown to block ligand-induced calcium mobilization by several seven
transmembrane-spanning receptors (34, 35). An analogous CCR5 mutant,
D76N, was made. Mutations affecting a highly conserved region of the
second intracellular loop have similarly blocked the coupling of other
seven membrane-spanning receptors to G-proteins (36-38), and we made
two constructs, R126N and D125N/R126N, that substituted asparagine for
conserved residues in this region of CCR5. These mutants were expressed
at or near wild-type levels in both HEK293 and CF2Th cells (Fig.
1 and 2 legends) but failed to mobilize
calcium in response to 500 ng/ml MIP-1
A chimeric molecule,
2M5, was also tested for responsiveness to MIP-1 Each of the CCR5 mutant
proteins was tested for its ability to bind MIP-1 We tested the
ability of each of the CCR5 mutants to support HIV-1 entry into
Hela-CD4 cells and CF2Th cells. Recombinant viruses containing the YU2
and ADA envelope glycoproteins infected Hela-CD4 cells expressing the
D76N mutant at levels comparable with that seen for cells expressing
wild-type CCR5. By contrast, both viruses inefficiently infected cells
expressing the 2M5 receptor, near the levels seen for cells expressing
the control receptor CCR4 (Fig. 3A). On CF2Th
cells cotransfected with CD4, each of the signaling defective mutants
D76N, D125N/R126N, and R126N supported efficient HIV-1 entry at a level
proportionate to their surface expression, as documented by FACS
analysis (Fig. 3B and its legend). Thus, the ability to
support HIV-1 infection is not significantly impaired in cells
expressing D76N, D125N/R126N, or R126N but is impaired in cells
expressing the 2M5 chimera.
In this work we asked whether there is a necessary relationship
between the signaling response of the chemokine receptor CCR5 to its
natural ligand and the role of CCR5 as a co-receptor for the HIV-1
virus. Although other investigators have attempted to probe this
relationship with pertussis toxin (39), a mutagenic approach was
necessary because chemokine receptors, in particular the closely
related CCR2, have been shown to be coupled to pertussis-insensitive as
well as pertussis-sensitive pathways (26). Both sets of pathways are
active in the natural target cells of HIV-1 (40). The D76N, D125N/R126N, and R126N mutants described here are incapable of efficiently mobilizing calcium in response to high levels of chemokine but are expressed well and bind MIP-1 The results with the 2M5 chimera demonstrate that the binding site for
MIP-1 Rucker et al. (42) have used constructs similar to 2M5 and
observed HIV fusion activity comparable with that of wild-type CCR5.
Several possibilities could account for this inconsistency with our
data. Unlike constructs used in the Rucker et al. report, 2M5 contains the first intracellular loop of CCR2 and is epitope-tagged at the N terminus. In addition, we used a single-step entry assay that
has a definite linear range and that may be more accurate than a
syncytium forming assay for quantifying the ability of different
receptors to support HIV-1 envelope-mediated membrane fusion. Other
data in Rucker et al. (42) indicate that HIV-1 envelope
glycoprotein-induced syncytium formation is sensitive to modifications
of the CCR5 N terminus. This conclusion is supported in this study with
a receptor whose expression and structural integrity are verified.
Ligands for many G-protein-coupled receptors, including chemokine
receptors, are thought to bind at least in part in a pocket formed by
the transmembrane helices and induce in the receptor a conformational
change that promotes guanine nucleotide exchange in G-proteins (32).
Chemokines are thought to bind this pocket at their N termini, and
chemokines with N-terminal truncations function as receptor
antagonists. Our data imply that the HIV-1 envelope need not induce an
activated conformation in CCR5 to enter and thus could bind away from
this pocket. Although chemokine inhibition of HIV-1 entry and gp120
binding might imply some overlap of the MIP1
Division of Human Retrovirology,
LeukoSite, Inc., Cambridge, Massachusetts 02142
-chemokine ligands is
necessary or sufficient for viral entry. Three CCR5 mutants with little
or no ability to mobilize calcium in response to macrophage inflammatory protein-1
could nonetheless support HIV-1 entry and the
early steps in the virus life cycle with efficiencies comparable with
those of wild-type CCR5. Conversely, a chimeric receptor with the N
terminus of CCR2 replacing that of CCR5 responded to macrophage
inflammatory protein-1
and MCP-1 but did not efficiently support
viral entry. These results demonstrate that chemokine signaling and
HIV-1 entry are separable functions of CCR5 and that only viral entry
requires the N-terminal domain of CCR5.
, MIP-1
, and RANTES (regulated on activation normal T
cell expressed and secreted), have been shown to inhibit the entry of
primary HIV-1 isolates (16) and to compete with gp120-CD4 complexes for
binding to CCR5 (17, 18).
chemokines) or separated by one intervening
residue (CXC or
chemokines) (19). Chemokines are believed to be
important in the trafficking of leukocytes in both basal and
inflammatory states (20). Chemokine receptors are G-protein-coupled,
seven transmembrane-spanning receptors (21, 22). Chemokine ligation of
receptor promotes the exchange of GDP for GTP in an associated
heterotrimeric G-protein, dissociation of G
from the G
and G
subunits, and numerous downstream effector functions, including
phospholipid hydrolysis and calcium mobilization (23). G-protein
subunits have been grouped in several classes based on sequence
similarity and common effector functions (24). Chemokine receptors have
been shown to be coupled to members of the Gi and the
Gq families (25-27). Signaling through Gi
proteins is inhibited by pertussis toxin, whereas Gq
signaling is not affected by pertussis toxin (24).
and mobilizes calcium in
response to MIP-1
and MCP-1, the ligand for CCR2 (28), but fails to
support efficient HIV-1 infection. These data demonstrate conclusively
that CCR5 coupling to G-proteins is not a requirement for efficient
HIV-1 entry. They also show that HIV-1 entry requires portions of the
CCR5 receptor not required for MIP-1
binding or signaling.
Plasmids
envCAT and pSVIIIenv plasmids used
to produce recombinant HIV-1 virions containing the envelope
glycoproteins from the primary, macrophage-tropic HIV-1 isolates ADA or
YU2 envelopes have been described previously (5, 9, 29, 30). The pCD4
plasmid used to express full-length CD4 in CF2Th cells has been
described (31). The cDNAs encoding epitope-tagged CCR5, CCR4, and
CXCR1 (IL8-RA) were cloned in a pcDNA3 vector (9). A pcDNA3
vector expressing FLAG epitope-tagged CCR2, was a generous gift of Dr.
Israel Charo (28). The FLAG epitope is DYKDDDDK (FLAG tag, IBI-Kodak)
inserted after the N-terminal methionine. Mutagenesis used to create
the expressor plasmids for the D76N, R126N and D125N/R126N mutants was
performed on CCR5 in a pcDNA3 vector using the QuikChangeTM method
of Stratagene, Inc., according to manufacturer's instructions. The 2M5
chimera was constructed by substituting the DNA encoding the
epitope-tagged CCR2 N terminus for the corresponding section of the
CCR5 gene in the pcDNA3 plasmid, using the common Msc-1 site as a
junction.
and MCP-1
(R & D Systems) were taken at excitation wavelengths 340 and 380 nm
and reported as a ratio of 340/380 nm. In parallel, an anti-CCR5
antibody, 5C7, was used to quantify receptor expression by FACS
analysis. Pertussis toxin-treated cells were incubated for 18 h
with 10 ng/ml pertussis toxin (CalBiochem).
-galactosidase expression
plasmid to assess transfection efficiency. Roughly 25-30% of the
cells were transfected. Cells were resuspended in binding buffer (50 mM HEPES, pH 7.5, 1 mM CaCl2, 5 mM MgCl2, and 0.5% bovine serum albumin).
Approximately 5 × 105 cells were mixed with 0.1 nM 125I-labeled MIP-1
(DuPont NEN) and
varying concentrations of unlabeled MIP-1
(R&D Systems) in a total
volume of 100 µl. Cells were shaken at 37 °C for 30 min,
centrifuged, resuspended in 0.6 ml of the same buffer containing 500 mM NaCl, and centrifuged again, and bound ligand was
quantitated by liquid scintillation counting. For affinity
measurements, nonspecific binding was determined in the presence of 200 nM Mip-1
and subtracted from all points.
Calcium Mobilization through CCR5 Mutants
(Fig. 1A).
Wild-type CCR5 responded strongly at 250 and 500 ng/ml (Fig.
1A and data not shown). When incubated 18 h with 10 ng/ml pertussis toxin, CCR2 and CCR5 expressing HEK293 cells responded
to 500 ng/ml MIP-1
with 50-60% of the peak values of the same
cells in the absence of pertussis toxin (Fig. 1C and data
not shown). We conclude that D76N, D125N/R126N, and R126N are expressed
at the cell surface but are not coupled to a signaling pathway leading
to calcium mobilization. We also conclude that, as previously reported
for CCR2 (26), CCR5 can couple to a signaling pathway that is
insensitive to pertussis toxin at high chemokine concentrations.
Fig. 1.
A, calcium mobilization in response to
MIP-1. Representative responses when wild-type CCR5, D76N, R126N,
and D125N/R126N were treated with 500 ng/ml MIP-1
at the time points
indicated by the arrows. Flux is displayed as a ratio of the
response at 340 nm to the response at 380 nm excitation wavelength. The
average mean fluorescence values of cells stained by anti-CCR5 antibody for CCR5, D76N, R126N, and D125N/R126N were 106 ± 50, 142 ± 19.5, 100 ± 10, and 52 ± 1, respectively. Background
staining observed with flourescein-conjugated second antibody only was
3.5 ± 0.1. For some experiments, lower amounts (20 µg/flask
rather than 30 µg/flask) of wild-type CCR5 DNA were used for
transfection to obtain expression levels comparable with those of the
CCR5 mutants. B, calcium mobilization of 2M5 chimera in
response to MIP-1
and MCP-1. Shown is a representative response of
2M5 when treated with 500 ng/ml MIP-1
or 1 µg/ml MCP-1 at the time
points indicated with the arrows. C, calcium
mobilization of CCR5 in the presence and the absence of pertussis toxin
treatment. CCR5-expressing HEK293 cells incubated with or without 10 ng/ml pertussis toxin (PTX) for 18 h before
measurements were taken. MIP-1
(500 ng/ml) was added at the time
points indicated with the arrows.
[View Larger Version of this Image (14K GIF file)]
Fig. 2.
A, competition binding of MIP-1 on
CCR5 mutants. Approximately 5 × 106 HEK293 cells
expressing CCR5 (squares), D76N (diamonds), and D125N/R126N (circles) were incubated with 0.1 nM
125I-labeled MIP-1
and varying concentrations of
unlabeled MIP-1
in duplicate. The results are expressed as the
percentage of counts bound to the same cells in the absence of cold
competitor. B, competition binding of MIP-1
on 2M5.
Approximately 5 × 106 Bing cells transfected with
plasmids expressing the chimeric 2M5 (circles) molecule or
wild-type CCR5 (squares) were incubated with 0.1 nM 125I-labeled MIP-1
and varying
concentrations of unlabeled MIP-1
, in triplicate. The results are
expressed as the percentage of counts bound to the same cells in the
absence of cold competitor.
[View Larger Version of this Image (14K GIF file)]
and the CCR2 ligand
MCP-1. The chimera was made by replacing the N terminus of CCR5 with
that of CCR2, with a junction in the second transmembrane domain. The
chimeric molecule responded like wild-type CCR5 to MIP-1
and also
gave an appreciable calcium flux to MCP-1 at 1 µg/ml, whereas
wild-type CCR5 responded only to MIP-1
and wild-type CCR2 responded
only to MCP-1 (Fig. 1B and data not shown). Thus the 2M5
chimera has retained the binding and signaling specificity of CCR5 and
has also acquired the ability to bind to and signal in response to
MCP-1.
Binding to CCR5 Variants
specifically.
Unlabeled MIP-1
competed for 125I-labeled MIP-1
binding to cells expressing wild-type and mutant proteins with very
similar efficiencies (Fig. 2A), yielding apparent dissociation constants of 6.8, 6.3, and 4.6 nM for
wild-type CCR5, D76N, and D125N/R126N, respectively. The 2M5 chimera
also bound MIP-1
at an affinity near that of wild-type CCR5 (Fig.
2B) with an apparent dissociation constants of 1.6 and 1.2 nM for the chimeric and wild-type proteins, respectively.
We conclude that 2M5, D76N, and D125N/R126N each bind MIP-1
with
affinities near that of wild-type CCR5.
Fig. 3.
A, infection of Hela-CD4 cells
expressing CCR5 mutants with recombinant HIV-1. A representative
experiment on Hela-CD4 cells expressing CCR5, D76N, 2M5, or control
receptor CCR4 is shown. HIV-1 viruses containing the ADA or YU2
envelope glycoproteins was were used to infect Hela-CD4 cells
expressing mutant receptors. Comparable results were obtained in other
experiments for D76N (n = 2) and 2M5 (n = 3). The results are expressed as the percentage of conversion of
chloramphenicol to acetylated forms. B, infection on CF2Th
canine thymocytes expressing CCR5 mutants by recombinant HIV-1. HIV-1
virus containing the ADA envelope glycoprotein was used to infect cells
expressing CD4 and D76N, R126N, D125N/R126N, wild-type CCR5, or control
receptor CXCR1. The results of two experiments are expressed as the
percentage of CAT activity relative to that observed for cells
expressing the wild-type CCR5 protein. Average mean fluorescence of
cells stained with the anti-CCR5 antibody: wild-type CCR5, 67.6; D76N,
39.1; R126N, 44.0; D125N/R126N, 32.8. Background staining was
3.4.
[View Larger Version of this Image (36K GIF file)]
with affinities close to that
of wild-type CCR5. The fact that these mutants support HIV-1 entry
excludes an obligate role for calcium mobilization and its sequelae in
promoting viral entry.
is distinct from that used by HIV-1 entry and that binding of
and efficient signaling through MIP-1
does not ensure a receptor
that supports HIV-1 entry. A second property of this chimera, the
ability to signal in response to the CCR2 ligand MCP-1, is consistent
with reports implying a strong requirement by MCP-1 for the N-terminal
domain of CCR2 (41). This contrasts with CCR1 and, as we have shown
here, CCR5, whose natural ligands are relatively insensitive to
perturbations in the N terminus of the receptor.
and gp120 binding sites
on CCR5, our data suggest that at least some of the elements of the
binding site are distinct. These differences may need to be considered
when designing strategies for therapeutic intervention. Further
understanding of the interaction of CCR5 with HIV-1 and with its
natural ligands could contribute to these efforts.
*
This work is supported by Grants AI24755 (to J. S.) and
AI/HL39759 (to C. G.) from the National Institutes of Health and by Center for AIDS Research Grant AI28691 to the Dana-Farber Cancer Institute. The Dana-Farber Cancer Institute is also the recipient of
Cancer Center Grant CA 06516 from the National Institutes of Health.
This work was made possible by gifts from the late William McCarty-Cooper, from the G. Harold and Leila Y. Mathers Charitable Foundation, from the Friends 10, from Douglas and Judi Krupp, and from
the Rubenstein/Cable Fund at the Perlmutter Laboratory.The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement" in
accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
§
These authors contributed equally to this work.
Supported by National Institutes of Health Grants HL51366 and
AI36162, as well as by the Rubenstein/Cable Fund at the Perlmutter Laboratory.
¶¶
To whom correspondence should be addressed.
1
The abbreviations used are: HIV, human
immunodeficiency virus; MCP, macrophage chemotactic protein; MIP,
macrophage inflammatory protein; CCR, CC chemokine receptor; CXCR, CXC
chemokine receptor; CAT, chloramphenicol acetyltransferase; FACS,
fluorescence-activated cell sorter.
2
L. Wu, W. A. Paxton, N. Kassam, J. Pudney, J. Rottman, D. J. Anderson, D. J. Ringler, J. Sodroski, W. Newman, R. A. Koup, and C. R. Mackay, submitted for publication.
©1997 by The American Society for Biochemistry and Molecular Biology, Inc.