(Received for publication, September 12, 1996, and in revised form, March 5, 1997)
From the Cancer Research Campaign Laboratories, The Beatson Institute for Cancer Research, Garscube Estate, Switchback Road, Bearsden, Glasgow G61 1BD, United Kingdom
The -chemokine macrophage inflammatory
protein-1
(MIP-1
) is chemotactic for many hemopoietic cell types
and can inhibit hemopoietic stem cell (HSC) proliferation, effects
mediated through G-protein coupled heptahelical receptors. We have
isolated cDNAs for seven chemokine receptors, CCR-1 to -5, MIP-1
RL1, and a novel cDNA, D6. Chinese hamster ovary cells
expressing CCR-1, -3, -5, and D6 bound 125I-murine
MIP-1
: the order of affinity was D6 > CCR-5 > CCR-1 > CCR-3. Each bound a distinct subset of other
-chemokines: the order of competition for 125I-murine MIP-1
on D6 was
murine MIP-1
> human and murine MIP-1
> human RANTES~JE > human MCP-3 > human MCP-1. Human MIP-1
and the
-chemokines did not compete. Like other chemokine receptors, D6
induced transient increases in [Ca2+] in HEK 293 cells
upon ligand binding. D6 mRNA was abundant in lung and detectable in
many other tissues. Bone marrow cell fractionation demonstrated T-cell
and macrophage/monocyte expression of D6, and CCR-1, -3, and -5. Moreover, we could detect expression of CCR-3, CCR-5, and to a greater
extent D6 in a cell population enriched for HSCs. Thus, we have
characterized four murine
chemokine receptors that are likely
involved in mediating the pro-inflammatory functions of MIP-1
and
other chemokines, and we present D6, CCR-3, and CCR-5 as candidate
receptors in MIP-1
-induced HSC inhibition.
Macrophage inflammatory protein-1
(MIP-1
)1 is a member of the chemokine
family, a group of almost 30 small polypeptides related functionally
with respect to their ability to act as stimulators of leukocyte
chemotaxis, and structurally due to the conservation of four
disulfide-forming cysteine residues (reviewed in Refs. 1-5). Variation
in this structural motif allows the family to be subdivided into two
main branches. The C-X-C or
chemokines, which include interleukin-8
(IL-8), macrophage inflammatory protein-2
and -
(MIP-2
and
-
), and KC, have a single amino acid residue separating the two most
amino-terminal cysteines. In the C-C or
branch these two cysteines
are juxtaposed: macrophage inflammatory protein-1
and -
(MIP-1
and -
), monocyte chemoattractant proteins 1, 2, 3, and 4 (MCP-1, 2, 3, and 4), C10, eotaxin and RANTES are all members of this subfamily.
Lymphotactin, a C chemokine, is unique in containing only a single
amino-terminal cysteine residue, and only two in the mature protein
(6). This also appears to represent a functional division as there is
no sharing of signaling receptors between the subfamilies and while
-chemokines are classically viewed as neutrophil chemotaxins, the
family stimulate monocyte and macrophage migration and function
(1-5). Eosinophils, basophils, and lymphocytes migrate in response to
certain members of both families, that in some cases induce more overt
inflammatory functions such as a respiratory burst or degranulation
(7-13). Thus, these proteins are likely to be crucial during leukocyte
trafficking and immune surveillance, and have been implicated in the
development of various allergic responses, such as asthma (14), and in
inflammatory diseases like rheumatoid arthritis (15) and chronic lung
inflammation (16). Mice lacking the MIP-1
gene do not exhibit a
normal immune response when infected with certain viruses, such as
cocksackie virus in which the myocarditis normally associated with this
virus is not observed (17). This phenotype attests to the in
vivo role of MIP-1
as a mediator of inflammatory response.
The chemokines have also been implicated in the control of the cellular
movement during angiogenesis (18-20), and in the control of cell
proliferation (21-27). Furthermore, mice lacking the chemokine
stromal cell-derived factor have severely reduced numbers of B-cell and
bone marrow myeloid progenitor cells (28), while mice lacking the
murine IL-8 type B receptor homologue have an increased number of
B-cells (29), implicating these two proteins in the control of blood
cell development. We have demonstrated that MIP-1
, and to a lesser
extent MIP-1
, are able to inhibit the proliferation of transiently
engrafting hemopoietic stem cells both in vivo and in
vitro (26). MIP-1
is also active as a reversible inhibitor of
clonogenic epidermal cell proliferation, with the intradermal source of
the chemokine being the epidermal Langerhans cells (27). Binding sites
have been identified on the surface of purified human hemopoietic stem
cells and on a number of immature myeloid progenitor cell lines that
appear to have a higher affinity for MIP-1
than those present on
more mature cell types (30-32). We have shown that the hemopoietic
stem cell receptor is distinct from the major monocyte receptor by
using a mutated form of MIP-1
, HepMut, which is still active in stem
cell inhibition assays but is unable to induce monocyte chemotaxis
(33).
A number of chemokine receptors, all members of the G-protein-coupled
heptahelical receptor superfamily, have now been cloned which are able
to bind selected subsets of chemokines from either the or
branches (reviewed in Refs. 4 and 5). Human CCR-1, binds MIP-1
,
MIP-1
, MCP-1, MCP-3, and RANTES and is expressed on nearly all
mature hemopoietic cell types (34-36). Human CCR-2, which is expressed
on monocytes and some T-cells, can be produced as two
alternatively-spliced forms both of which bind and signal in response
to MCP-1 and MCP-3 with the altered COOH terminus affecting G-protein
partner selection (36-38). Degenerate oligonucleotide-primed PCR has
resulted in the identification of human and murine CCR-3, -4, and -5 (4, 5). Human CCR-5, which acts as a high affinity site for MIP-1
,
-
, and RANTES, is a central component of the protein complex
required for the infection of macrophages and T-cells by many isolates
of the human immunodeficiency virus: CCR-2 and -3, and the stromal
cell-derived factor receptor fusin (CXCR-4) also act as co-receptors
for different viral isolates (39-44). The CCR-5 ligands are able to
prevent the infection of cells with human immunodeficiency virus (45),
revealing that these proteins may play an important role in controlling
the progression of AIDS. Furthermore, cells from individuals homozygous
for a natural mutant of CCR-5 are resistant to infection with certain human immunodeficiency virus isolates (46).
To further understand the mode of action of MIP-1, we have cloned
seven murine members of the
-chemokine receptor family, six
previously reported and one novel sequence, D6. Four of these genes,
including D6, produce proteins capable of binding to murine MIP-1
when expressed in Chinese hamster ovary (CHO) cells with dissociation
constants ranging from 110 pM to 12 nM. D6
appears to be the highest affinity murine MIP-1
receptor cloned to
date. The differential binding properties of murine and human MIP-1
, and murine HepMut, suggests that the interaction with each receptor is
mediated by different amino acid residues in the ligand. Also, the
spectrum of other chemokines that are able to interact with the four
MIP-1
receptors is highly variable and demonstrates the high degree
of receptor sharing that occurs within the
-chemokine family. We
have also obtained expression profiles for each of the receptors and
demonstrate their expression in many hemopoietic cell types. In
addition, we present evidence implicating a subset of these receptors
in the processes of stem cell inhibition.
Materials
All DNA modifying enzymes were purchased from Life Technologies, Inc. Ltd., Paisley, United Kingdom. Plasmid DNA and DNA fragments were purified using the appropriate products from Qiagen, Dorking, UK. Agarose gels were buffered with 1 × TAE (40 mM Tris acetate, 1 mM EDTA). PCR and RT-PCR was performed using AmpliTaqTM polymerase and buffers (Perkin-Elmer, Roche Molecular Systems, Inc., Branchburg, NJ) except where indicated. Oligonucleotides were synthesized on an Applied Biosystems 392 DNA/RNA synthesizer. DNA was sequenced on an Applied Biosystems 373A DNA sequencer.
Chemokines
All chemokines were purchased in 10-µg aliquots from R&D
Systems, Abingdon, UK, except human (h) lymphotactin and MIP-1, and
murine (m) eotaxin (Peprotech, London, UK). Due to the tendency of wild
type murine MIP-1
to aggregate, we have previously designed a number
of fully active non-aggregating variants of the molecule which are more
appropriate for binding studies of the type described here (47). One of
these mutants, PM2, is used throughout this work and will be referred
to as mMIP-1
hereafter.
Cell Culture
All cell culture reagents were purchased from Life Technologies, Inc. unless otherwise stated. CHO cells were maintained in special liquid medium supplemented with 4 mM glutamine and 10% fetal calf serum (FCS) (TCS Biologicals, Buckingham, UK). HEK 293 cells were maintained in Dulbecco's modified Eagle's medium with 10% FCS and 4 mM glutamine. XS52 Langerhan's cells were grown as described previously (48).
Cloning of Receptors
The following oligonucleotides were synthesized from regions of
similarity between human CCR-1 and -2, and cytomegalovirus US28 in the
case of TEFDY: sense GALPY: 5-GGGGCICA(A/G)CT(G/C)CT(G/C)CCICC-3
. Sense DIYLLN: 5
-AT(T/C)TA(T/C)CTICTIAACCT(C/G)GC-3
. Sense TEFDY: 5
-ACCAC(A/C)ITITTTGA(T/C)TATG-3
. Antisense FIILL:
5
-GTIAG(G/C)AGGATGAT(A/G)AA(A/G)AAAAT-3
. Antisense DRYLA:
5
-GACIAT(A/G)GCCAGGTACC(G/T)GTC-3
. Genomic DNA was isolated from the
homogenized spleen of a C3H mouse according to Sambrook et
al. (49). PCR was performed using variable amounts of
MgCl2 (from 1.25 to 2.25 mM), 0.3 mM of each dNTP, and 6 ng/µl of one sense and one
antisense oligonucleotide. Reactions were incubated for 35 cycles at
94 °C for 1 min, 50-52 °C for 1 min, and 72 °C for 2 min.
Products of expected size were cloned into pCRScript (Stratagene, La
Jolla, CA) and sequenced. 5
and 3
rapid amplification of cDNA
ends (RACE) was performed using RACE kits from Life Technologies, Inc.
according to their instructions on tissues or cell lines positive for
each receptor in RT-PCR reactions. No alternatively spliced forms of
the type seen with human CCR-2 (37) were detected. Pfu
polymerase (Stratagene) was used on cDNA or genomic DNA according
to the manufacturers' instructions, with reactions containing 10%
dimethyl sulfoxide and 5% glycerol and oligonucleotides specific to
the 5
and 3
ends of the gene. Three separate reactions were
performed, and all products were cloned into pCRScript and fully
sequenced in both orientations. Sequences were analyzed using GCG
software (50): phylogenic relationships were determined using the
Distances (Kimura method) and Growtree (Neighbor-joining method)
programs.
mRNA Analysis
Total RNA was extracted using Trizol (Life Technologies, Inc.).
Northern blots onto Hybond N+ (Amersham, Little Chalfont, UK) were
performed according to Nibbs et al. (51). DNA probes were
labeled with [-32P]dCTP (Amersham) using the
Ready-to-go kit (Pharmacia), unincorporated nucleotides were removed
using NICK Sephadex G-50 columns (Pharmacia) and the probe denatured
for 3 min at 100 °C prior to hybridization. Filters were exposed to
Kodak X-Omat x-ray film and stripped in 0.1% sodium dodecyl sulfate
(SDS) at 100 °C prior to reprobing. RT-PCR was performed using the
RNA PCR core kit (Perkin-Elmer). For each RNA sample reverse
transcription was done in two tubes, one with and one without reverse
transcriptase. 1 µg of heat-denatured total RNA was used per PCR
reaction, which had been previously treated with DNase I to remove
genomic DNA contaminants (according to Life Technologies, Inc. RACE
kits), followed by phenol/chloroform extraction, ethanol precipitation,
and 70% ethanol wash. Specific PCR primers used were (AS: antisense,
S: sense): D6; AS: 5
-GGAAGAGACAGTAATGAGTAAGGC-3
and S:
5
-GTGACAGAGAGCCTGGCCTTC-3
for tissue PCR (expected size, 345 base
pairs). PCR proceeded for 30 cycles of 94 °C for 1 min, 58 °C for
1 min, 72 °C for 1 min.
Generation of Stably Transfected Cell Lines
Full-length cDNAs were excised as BamHI/NotI fragments from the pCRScript vectors and cloned into pcDNA3 (Invitrogen, Abingdon, UK). These vectors were transfected into CHO cells and single cell clones selected in 1.6 mg/ml geneticin (Life Technologies, Inc.) as described previously (33). Total RNA was extracted from approximately 12 separate clones, Northern blots performed to assess expression levels and three of the highest expressors selected for ligand binding experiments. pcDNA3 containing D6 cDNA was transfected into HEK 293 cells using Transfectam (Promega, Southampton, UK) according to the manufacturers' instructions. Stably transfected pools of cells were selected in 700 µg/ml geneticin.
Receptor Binding Studies
5-µg aliquots of mMIP-1 were labeled on a regular basis
with Na125I (DuPont NEN) using IODO-GEN (Pierce) as
described previously (30). Stably transfected CHO cells were plated at
105 cells/well in 24-well plates and incubated overnight at
37 °C. After washing with warm PBS, the cells were incubated in 250 µl of binding buffer (special liquid medium plus 10% fetal calf
serum, 4 mM glutamine, 0.2% sodium azide, 25 mM HEPES (pH 7.4)) containing variable concentrations of
radioiodinated MIP-1
and cold competitor chemokine. Binding
proceeded for 90 min at 22 °C, then each well was washed three times
in ice-cold PBS and the cell monolayer was then solubilized by the
addition of 0.5 ml of 1% SDS. Lysed cells were transferred to a
counting vial and bound radioactivity counted for 1 min in a Beckman
Gamma 5500B counter. Each data point was assayed in triplicate and each
experiment performed at least twice for accuracy. Data was analyzed
using Scahot and Scafit programs in LIGAND software (52). Initially,
each transfected cell line, and the parental cell line, was tested for
binding at 40 and 4 nM 125I-mMIP-1
.
Transfected cells that exhibited a significant increase in binding
above background were subjected to a full Scatchard analysis with 12 data points in triplicate with varying concentrations of
125I-mMIP-1
. Kd values were then
confirmed by varying the amount of unlabeled mMIP-1
while
maintaining the 125I-mMIP-1
at approximately the
Kd value. Competition experiments with other
unlabeled chemokines were similarly performed.
Intracellular Ca2+ Measurements
HEK 293 cells stably transfected with D6 were harvested by
trypsinization, washed in HACM buffer (125 mM NaCl, 5 mM KCl, 1 mM MgCl2, 0.5 mM glucose, 0.025% bovine serum albumin, 20 mM
HEPES (pH 7.4)), and loaded in HACM containing 10 µM
Fura-2-AM (Sigma) for 60 min at 37 °C in the dark. Cells were washed
twice in HACM warmed to 37 °C and resuspended to 2-4 × 106 cells/ml in HACM. A 2-ml aliquot was placed in a
continuously stirred cuvette at 37 °C in a Perkin-Elmer LS50
Spectrometer, CaCl2 added to 100 nM and
fluorescence monitored at 340 nm (ex) and 500 nm
(
em). Data were collected every 100 ms. Chemokines were
added to a final concentration of 100 nM.
Generation of cDNA Fragments from Sorted Bone Marrow Cells
Preparation of cDNABone marrow cells from the femurs
of 20 mice were isolated into PBS containing 2% fetal calf serum
(PBS/FCS). 5 × 106 cells were used directly for
poly(A)+ mRNA preparation (see below). The remaining
cells were diluted to 2.5 × 107 cells/ml, layered
onto an equal volume of 1.077 g/ml Nycodenz solution (Nycomed, Oslo,
Norway) and spun at 1,000 × g for 30 min at 22 °C.
The low density cells at the interface were removed, spun for a second
time through the density gradient, washed twice in PBS/FCS, and
resuspended to 1 × 106 cells/ml in PBS/FCS. Different
cell types were then sorted sequentially in the following order by
addition of rat antibodies, Gr-1, B220, CD5/CD8, and Mac-1 (Pharmingen,
San Diego, CA, except Mac-1, Caltag, San Francisco, CA) followed by
magnetic separation with sheep anti-rat M-450 beads (Dynal, Oslo,
Norway), according to the manufacturers' instructions. After the four
antibody positive cell fractions were purified, remaining unbound cells
were designated lineage negative (lin). Poly(A)+ mRNA
was then isolated from the six cell populations (total bone marrow,
Gr-1+, B220+, CD5/CD8+, Mac-1+, and lin
) using the
Micro-FastTrackTM kit (Invitrogen) and cDNA prepared,
according to the suppliers instructions, with the Copykit (Invitrogen).
cDNA was digested with AluI, the enzyme was
heat-inactivated and the cDNA fragments were phenol/chloroform
extracted and ethanol precipitated.
Specific pairs of oligonucleotide
linkers, as shown, were used for each cDNA sample to minimize carry
over contamination: bone marrow: A,
5-ACAGTCCCATGGTACAAGTTCAGCA-3
and B, 5
-GAACTTGTACCATGGGACTGT-3
; granulocyte (Gr-1): A, 5
GAACTCGGATCCCATGTGAAGGTGT-3
and B,
5
-CTTCACATGGGATCCGAGTTC-3
; B-cells (B220): A,
5
-AACCGTAGATCTGCTGCACACCAAC-3
and B, 5
-GTGTGCAGCAGATCTACGGTT-3
; T-cells (CD5/CD8): A, 5
-TAGACCAAGCTTAGACTGACATCCT-3
and B,
5
-TGTCAGTCTAAGCTTGGTCTA-3
; monocytes/macrophages (Mac-1): A,
5
-AGCCATTCTAGACGTGTAACTGATA-3
and B, 5
-AGTTACACGTCTAGAATGGCT-3
;
lineage negative (lin
): A, 5
-TAGTCCGAATTCAAGCAAGAGCACA-3
and
B, 5
-CTCTTGCTTGAATTCGGACTA-3
. Each set of linker pairs (6 µg of
each oligonucleotide, heat denatured at 90 °C for 2 min and cooled
to 4 °C) were kinased in 1 × polynucleotide kinase buffer (NBL
Gene Sciences Ltd., Cramlington, UK) containing 1 mM dATP
and 1 µl of [
-32P]dATP (Amersham) with 20 units of
polynucleotide kinase (NBL) at 37 °C for 1 h. The reaction was
placed 70 °C for 5 min, allowed to cool over 2 h to 22 °C to
anneal the complementary linkers, and stored at
20 °C.
The duplexed, kinased linkers were ligated onto the corresponding cDNA according to Sambrook et al. (49), in 1 × ligase buffer (NBL) containing 1 mM hexamminecobalt chloride and 0.5 mM spermidine, at 16 °C for 20 h. Linkers were separated from cDNA by adding glycerol to 20% and briefly electrophoresing each sample on a 2% low melting point agarose TAE gel. DNA separate from the unincorporated linkers (~100-1000 base pairs) was excised in a gel volume of approximately 0.5 ml. PCR was performed directly on 5 µl of this slice melted at 70 °C in a reaction containing 0.4 mM of each dNTP, 16 ng/µl of the cell type-specific B oligonucleotide (see above), and 7.5 units of AmpliTaqTM. Reactions were incubated for 35 cycles at 94 °C for 1 min, 52 °C for 1 min, and 72 °C for 3 min.
Southern BlottingVisually estimated equal quantities
(approximately 5 µg) of the separate PCR products from each
population of sorted cells were separated on a 1.2% agarose gel that
was subsequently treated for 2 × 15 min with 1.5 M
NaCl, 0.5 M NaOH and then neutralized for 2 × 30 min
in 1.5 M NaCl, 0.5 M Tris (pH 7.5), 1 mM EDTA. DNA was transferred to Hybond N+, fixed,
prehybridized, hybridized, washed, and exposed as described for
Northern blots. Filters were stripped in 0.1% SDS at 100 °C prior
to reprobing. Full-length cDNA was used as a probe for each of the
chemokine receptors. Other probes used were murine T cell receptor C
cDNA and murine immunoglobulin H cDNA (provided by Dr. M. Stewart, University of Glasgow), murine lysozyme M (provided by Dr. M. Cross, CRC Paterson Institute, Manchester), v-fms (provided by Dr. A. Ford, LRF Chester Beatty Labs., London), murine CD34 cDNA (provided by Dr. G. May, LRF Chester Beatty Labs., London), and murine
-actin (D. Jarmin, Beatson Institute, Glasgow).
Utilizing combinations of the three 5 and two 3
primers
described above, we identified fragments of seven members of the murine
G protein-coupled receptor family which were subsequently used to
isolate full-length cDNAs (see "Experimental Procedures"). Six
of our sequences (Fig. 1) are essentially identical,
with a few amino acid changes, to previously cloned murine chemokine receptors, i.e. RN31 is CCR-1, FG15 is CCR-2 (homologous to
human CCR-2b form), G17 is CCR-3, D2 is CCR-4, RN3 is CCR-5, and ME is
the CCR-1-like orphan receptor MIP-1
RL1 (Refs. 53-57). In addition, we have also identified a novel receptor, D6, which shows homology to
these and other members of the subfamily of G-protein-coupled chemokine
receptors.
Characterization of the D6 cDNA, a Novel Seven Transmembrane Spanning Receptor
A full-length D6 cDNA, isolated from brain
RNA revealed a single open reading frame, encoding a protein of 378 amino acids (Fig. 2A). 5 RACE PCR performed
from heart, brain, and liver, gave three identical products that may
represent three potential transcriptional initiation sites for this
gene (indicated with a bracket in Fig. 2A): one
may initiate translation at the second methionine to produce an
amino-terminally truncated protein. This potential variant would be
similar to an alternatively spliced form of BLR-1, an orphan receptor
related to the IL-8 receptors, in which two distinct 5
exons result in
the generation of either full-length or amino-terminally truncated
forms of the receptor (58). No similar 5
RACE products starting prior
to the first ATG were seen with the six other receptors. However, we
cannot exclude the possibility that these D6 variants are the result of
PCR artifact. The full-length D6 protein contains the predicted seven-transmembrane spanning regions, the four conserved cysteine residues thought to maintain the receptor in a cylindrical structure and in addition has a single putative N-linked glycosylation
site at the amino terminus. In common with other chemokine receptors, the amino terminus is highly acidic although the first 13 amino acids
are predicted to form a hydrophobic domain that is not seen in the
related chemokine receptors. A phosphorylation site for cAMP-dependent protein kinase is present on the first
cytoplasmic loop and further phosphorylation sites for protein kinase C
and casein kinase II are found in the carboxyl-terminal tail which possesses a total of 13 serine and threonine residues. The strongly conserved DRYLAIV motif in the second intracellular domain which is
involved in G-protein docking is altered to DKYLEIV in D6, a change
which likely affects the selection of G-protein partners. Furthermore,
D6 lacks the putative G protein-binding site present in many
heptahelical receptors at the carboxyl-terminal end of the third
intracellular loop, i.e. BBXXB or BBXB (where B
is a basic amino acid and X is any amino acid) (59). It
does, however, contain a similar site at the amino-terminal end of the
long cytoplasmic tail which is not present in many of the other
chemokine receptors (Fig. 2A). Analysis of the homologies
between D6 and the other chemokine receptors indicate that D6 displays
some 30-37% identity and around 60% similarity to each of the cloned
chemokine receptors (data not shown). Phylogenic analysis suggests that
this receptor lies between the C-C and C-X-C receptor subfamilies, most
closely related to CCR-4 and two murine IL-8 receptor-like genes,
mIL-8R (29) and mIL-8RL2 (Fig.
2B). However, comparison of extracellular amino termini, believed to be important in ligand binding, shows that in this domain
D6 is more closely related to the C-C chemokine receptors, especially
CCR-1 (data not shown).
Initial analysis of D6 receptor expression patterns using tissue blots
(Fig. 3A) indicated that expression of a
single approximately 3.2-kilobase transcript for D6 was detectable in
lung, and at much lower levels in liver and spleen. In addition, RT-PCR
analysis revealed expression of D6 in heart, brain, thymus, ovary,
muscle, liver, and kidney, but not mammary gland or testes (Fig.
3B). Transcripts were also detectable by Northern blot
analysis in murine solid tissues for CCR-1 (spleen > lung,
heart > skeletal muscle), CCR-3 (spleen), CCR-4 (liver,
thymus > lung > heart, brain > striated muscle), and
CCR-5 (low levels in spleen, liver), but not CCR-2 or ME-MIP-1RL1
(data not shown). It must be emphasized, however, that the contribution
of resident leukocytes to the signal in many of these tissues cannot be
ruled out and definitive assessment of the cell type specific
expression patterns in each of these tissues awaits in situ
hybridization studies.
Ligand Binding Profiles of the Murine Chemokine Receptors
As
our primary interest lies in the identification of MIP-1 receptors,
each of the seven receptors described above was stably transfected into
CHO cells and three clones expressing each receptor were tested for
their ability to bind radioiodinated mMIP-1
. CCR-1, CCR-3, CCR-5,
and D6 bound 125I-mMIP-1
and we have thus concentrated
our subsequent binding studies on these four receptors. Murine CCR-4
has been reported to bind human MIP-1
(55) and our inability to
detect an interaction with murine MIP-1
is surprising and may imply
that CCR-4 can distinguish between MIP-1
from these two species.
However, we have not demonstrated cell surface expression of CCR-4 in
the CHO cell transfectants. The affinities of
125I-mMIP-1
for the receptors (expressed as dissociation
constants Kd, Table I) allows the
subdivision of the receptors into high affinity (D6 and CCR-5) and low
affinity (CCR-1 and CCR-3) groups. These results also present D6 as the
highest affinity receptor for MIP-1
identified to date. Displacement
curves (Fig. 4, A-D) and apparent
Kd values (Table II) indicate that these two subgroups can also be discriminated on the basis of their
ability to bind human MIP-1
and RANTES. CCR-5 and D6 display only
weak or no significant binding of hMIP-1
but interact well with
human RANTES (Kd values of 550 pM and 2 nM, respectively: Table II) whereas CCR-1 and CCR-3 both
bind human MIP-1
with an avidity similar to that seen with mMIP-1
but no interaction with hRANTES is observed.
|
|
All these four receptors bind murine and human MIP-1 with
Kd values slightly higher than those seen with
mMIP-1
. Other human MCP-1, murine MCP-1/JE, and human MCP-3
chemokines also bind to D6 with affinities in the nanomolar range,
while CCR-5 also interacts weakly with human and murine MCP-1 and human MCP-2 (Fig. 4, C and D; Table II). CCR-1 also
interacts with mMCP-1/JE, hMCP-2, and murine C10 when the ligand is
present at high concentrations (Fig. 4A). Thus, each of the
four MIP-1
receptors has a unique repertoire of potential ligands
although the physiological relevance of interactions which only occur
at very high concentrations, mC10 with CCR-1, for example, is
questionable.
None of the -chemokines tested (IL-8, KC and MIP-2
), nor
lymphotactin, are capable of displacing radiolabeled MIP-1
from any
of the cloned chemokine receptors indicating that these receptors are
specific for the
-chemokine family. This is particularly important
with respect to D6 given the apparently comparable degree of overall
similarity between this receptor and the
- and
-chemokine receptors. Eotaxin, which has been reported as a ligand for CCR-3 (60),
appeared to form aggregated complexes with the labeled mMIP-1
and
made interpretation of displacement experiments impossible (data not
shown).
Thus, D6 acts as a high affinity receptor for -chemokines when
it is expressed in CHO cells, binding to mMIP-1
> human and murine
MIP-1
> hRANTES, mMCP-1/JE > hMCP-3 > hMCP-1. The
affinity and specificity of this interaction most closely resembles
that seen with CCR-5, while CCR-1 and -3 bind with lower affinity to mMIP-1
and are somewhat more specific for human and murine MIP-1
and
.
Ca2+ mobilization upon ligand
stimulation is characteristic of chemokine receptors described to date.
Due to the inability in our hands of CHO cells expressing any of the
receptors to generate a Ca2+ flux upon chemokine treatment,
pools of HEK 293 cells stably transfected with D6 were generated; HEK
293 cells have been successfully used by others to study chemokine
receptor signaling. Treatment of transfectants with 100 nM
mMIP-1, mMIP-1
, or hRANTES generated a detectable increase in
intracellular Ca2+ demonstrating that this receptor is able
to signal upon ligand interaction; hMCP-3 gave a significantly weaker
response (Fig. 5). In addition, each of the ligands were
able to desensitize the cells to a subsequent treatment 100 s
later with a second D6 ligand (Fig. 5). Untransfected HEK 293 cells
were unresponsive to all the chemokines tested (data not shown).
Hemopoietic Subpopulation Expression Patterns of the Murine Chemokine Receptors
To more specifically examine the patterns of
receptor expression within the hemopoietic system, we have investigated
the expression of D6 and the other cloned MIP-1 receptors using
amplified cDNA fragments generated from sorted populations of
hemopoietic cells, including a lineage negative fraction containing
primitive cells (see "Experimental Procedures"). We have confirmed
the success of this sorting by probing Southern blots of the sorted
cell amplified cDNA fragments with lineage specific markers. The
results (Fig. 6A) suggest a high degree of
lineage separation and a marked enrichment of CD34+ primitive cells in
the lin
population. All bands were <500 base pairs due to the
AluI digestion of the cDNA prior to amplification and
blotting.
We have examined these lineage-specific cDNA populations for
evidence of expression of the four murine MIP-1 binding chemokine receptors, CCR-4 and MIP-1
RL1. The results from this analysis, shown
in Fig. 6B, reveal expression of all the receptors in the T
cell and monocyte/macrophage lineages. CCR-1, -3, -4, MIP1
RL1, and
possibly CCR-5 are also expressed in the granulocyte subpopulation. In
addition to macrophages and T-cells, we have also been able to detect
CCR-1 and D6 expression in the XS52 Langerhans cell line suggesting
that these receptors may be a normal component of the
Langerhans/dendritic cell receptor
population.3 Intriguingly, a number of the
MIP-1
receptors are also detectable in primitive lineage negative
hemopoietic cells suggesting a potential role for these genes in
MIP-1
-induced stem cell inhibition (Fig. 6B). D6 is
easily detected showing a strong enrichment in lin
cells relative to
total bone marrow to levels similar to that seen with CD34 (Fig.
6B). CCR-3, and possibly CCR-5, are also detectable and the
signals are significantly higher in the lin
cell population than in
unfractionated bone marrow and the sorted B-cells.
In the present study, we have cloned seven murine chemokine
receptors, including a novel cDNA which we have named D6. Four of
these cDNAs, including D6, produce proteins in CHO cells which will
bind to mMIP-1, and extensive displacement experiments have demonstrated that these receptors also bind distinct arrays of other
-chemokines. D6 is a signaling-competent receptor and can generate a
transient rise in intracellular Ca2+ upon interaction with
ligand. In addition, we have characterized the lineage-specific
expression of each of the receptors, demonstrating T-cell and
monocyte/macrophage expression of the four MIP-1
receptors, and
moreover, have shown the expression of D6, and possibly CCR-3 and -5 in
a stem cell-enriched bone marrow fraction.
Comparisons of the primary sequence of D6 and the other cloned
chemokine receptors suggests that it is equally related to the CCR and
CXCR families. Thus, D6 may structurally represent a member of a new
subset of chemokine receptors. This suggestion is supported by
alterations in motifs conserved throughout the majority of chemokine
receptors, such as the DRYLAIV motif at the amino-terminal end of the
second intracellular loop, that is changed to DKYLEIV in D6 altering
the charge of this domain and possibly the G-protein subunit selection
during signaling. However, it is clear from our binding and signaling
data that it functions as a receptor which preferentially binds
-chemokines. This may in part be mediated by the amino terminus of
D6 which bears significantly more homology to
-chemokine receptors,
particularly CCR-1, than
-chemokine receptors.
Full-length D6 cDNA expressed in CHO cells produces the highest
affinity murine MIP-1 receptor identified to date and shares many
binding characteristics with CCR-5. This protein also binds with high
affinity to human and murine MIP-1
and human RANTES but
interestingly, we were unable to displace mMIP-1
from D6 and CCR-5
with human MIP-1
. When the human and mouse MIP-1 proteins and human
RANTES are aligned a number of conserved residues are apparent that are
different in human MIP-1
. Two changes are of particular interest
(labeled with stars in Fig. 7), namely Arg to
Gln at position 22 (21 in mMIP-1
) and Glu to Lys at 61 (60 in
mMIP-1
) that alter the charge at these positions. We have previously
made a mutant of murine MIP-1
, PM3 (47), in which the Glu residue at
60 had been neutralized by alteration to Gln: this mutant exhibits wild
type activity in biological assays (47), and binds with comparable
affinity to D6 (data not shown). However, the charge reversal caused by
the presence of Lys at 61 in hMIP-1
may have a more powerful effect
on receptor-ligand interaction. The alteration from the bulky, basic
Arg residue at 21 to the smaller uncharged Gln residue at 22 in
hMIP-1
is a good candidate for being important in the difference in
binding to D6 seen with murine and human MIP-1
: the difference in
size and charge of the side chain may disrupt a potential
receptor-binding domain in this area of the protein. Interestingly,
based on the three-dimensional structures of
-chemokines (61, 62)
Glu-60 in the predicted carboxyl-terminal helix, and Arg-21, are likely
to be on the surface of the protein in close proximity and may act as a
D6-binding domain. However, Lys-44 and Arg-45 (labeled with a
cross in Fig. 7) are also involved in mMIP-1
binding to
D6 as neutralization of these residues in HepMut (33) reduces the
affinity of mMIP-1
for this receptor with the Kd
rising from 110 pM to 8 nM (data not shown).
These residues are not sufficient, as they are also found in hMIP-1
.
Thus, there may be several domains of mMIP-1
involved in mediating
high affinity binding to D6, and subtle changes in these domains may
exert quite profound effects on the ability of the ligand to interact.
Given the similarities in the binding profiles of D6 and CCR-5, it is
likely that the residues in mMIP-1
that mediate binding to CCR-5 may
be the same as those discussed above for D6.
D6 is also able to interact well with JE (mMCP-1) and human MCP-3,
demonstrating that this receptor is highly promiscuous with respect to
ligand selection. Recent data suggests that mMCP-4 can also bind to D6,
with a Kd in the low nanomolar range (data not
shown). In contrast, CCR-1 and CCR-3 appear to be more specific for
MIP-1 and -
but have significantly lower affinity for these
ligands than CCR-5 and D6. Interestingly, human RANTES did not displace
mMIP-1
from these two receptors, which is in conflict with results
previously published involving CCR-1 (53, 54) where this chemokine
displaced mMIP-1
with an apparent Kd of 4 nM (54). The reason for this discrepancy is puzzling
although there are a number of possible explanations. First, expression
in different cell types may affect receptor folding or G-protein
interaction that may alter ligand selection: we used CHO cells in this
study while COS cells (54) or K562 cells (53) were used previously.
Second, there are three amino acid changes between our sequence for
CCR-1 and published sequences (Fig. 1), and one of these, Gln instead
of His at 278, is in the third extracellular domain and may disrupt
hRANTES binding. Indeed, this region of CCR-1 appears to be important
in mediating competent ligand binding and receptor activation (56,
63).
CCR-3 binds human and murine MIP-1 and -
, and reportedly acts as
a receptor for human and murine eotaxin (60) although we have been
unable to demonstrate eotaxin binding due to the tendency of this
chemokine to aggregate with mMIP-1
. Gao and colleagues (60) were
unable to detect calcium ion flux in response to mMIP-1
and -
in
HEK 293 cells transfected with CCR-3, but eotaxin could signal into
these cells: it is still unclear whether the MIP-1 proteins are able to
transduce a signal upon interaction with CCR-3 or if they are able to
prevent eotaxin binding to CCR-3. Additional confusion comes from the
easy detection of this receptor in the T-cell and monocyte/macrophage
lineages, cells upon which eotaxin has no reported biological effect.
This may indicate that CCR-3 has cell-type specific signaling
properties permitting it to mediate eotaxin-induced chemotaxis in
eosinophils but potentially other functions in T-cells and
monocytes/macrophages in response to the MIP-1 proteins.
Finally, it is of note that the binding properties of these four murine receptors exhibit a number of differences compared with their presumed human homologues and highlights the potential discrepancy between structural and functional homology. Formal assignment of direct murine homologues therefore awaits more systematic generation of complete biological and functional data.
In conjunction with these detailed binding analyses we have
demonstrated that all the four MIP-1 receptors are expressed in
cells of the monocyte/macrophage and T-cell lineages. Which cell types
within these lineages express these proteins remains to be determined.
However, it seems possible that cells at distinct stages of
differentiation, or with certain biological functions, may present a
specific repertoire of MIP-1
receptors which could allow them to
respond only to certain concentrations of ligand or generate
receptor-specific biological responses.
We have previously shown that MIP-1 is able to inhibit the
proliferation of hemopoietic stem cells and clonogenic epidermal cells
(26, 27). While none of the four MIP-1
receptors is detected in
murine keratinocytes (data not shown), our bone marrow sorting
experiments have demonstrated that enrichment for hemopoietic stem
cells results in a concomitant enrichment for cells expressing D6, and
possibly CCR-3 and -5. These three receptors also bind to HepMut (data
not shown), a mMIP-1
mutant that inhibits stem cell proliferation
but is inactive in monocyte chemotaxis assays (33). However, to
unequivocally demonstrate any involvement of the receptor(s) in
hemopoietic stem cell inhibition it will be necessary to abrogate the
expression of the receptor(s) in the target cell to see if inhibition
by MIP-1
, and other chemokines, is lost. These experiments are
currently underway in our laboratory.
We thank M. Freshney for cell sorting expertise, R. MacFarlane for operation of the DNA sequencer, Drs. Stewart, Cross, Ford, and May for providing cDNA clones, Dr. Takashima for the XS52 cell line, and Dr. Czaplewski for the HEK 293 cells and advice on Ca2+ measurements. We also thank Professor Wyke for critically reading this manuscript. Dr. Nibbs also thanks Dr. Amanda Wilson for support services.