Jul 23, 1996 - and ,3 (CC) chemokines (Stoeckle and Barker, 1990;. Schall, 1991).We and others ... receptors belonging to the seven membrane-spanning. G-protein coupled ... binding to the human CCR-1 chemokine receptor. Our data further .... apparent affinity of wild-type MIP- I Cx for heparin is higher than that for ...
The EMBO Journal vol.15 no.23 pp.6506-6515, 1996
Uncoupling of stem cell inhibition from monocyte chemoattraction in MIP-la by mutagenesis of the proteoglycan binding site Gerard J.Graham', Peter C.Wilkinson2, Robert J.B.Nibbs, Sharon Lowe, Svein O.Kolset3, Anne Parker, Mary G.Freshney, Monica L.-S.Tsang4 and Ian B.Pragnell The Beatson Institute for Cancer Research, CRC Beatson Laboratories, Garscube Estate, Switchback Road, Bearsden, Glasgow G61 IBD, 2The Department of Immunology, University of Glasgow, Glasgow G12 8QQ, UK, 3Institute of Nutrition Research, University of Oslo, PO Box 1046, Blindern, 0316, Oslo, Norway and 4R & D Systems Inc., Minneapolis, MN 55413, USA 'Corresponding author
We have studied the role of proteoglycans in the function of Macrophage Inflammatory Protein-1 alpha (MIP-la), a member of the proteoglycan binding chemokine family. Sequence and peptide analysis has identified a basic region within MIP-la which appears to be the major determinant of proteoglycan binding and we have now produced a mutant of MIP-la lacking the basic charges on two of the amino acids within this proteoglycan binding site. This mutant (Hep Mut) appears to have lost the ability to bind to proteoglycans. Bioassay of Hep Mut indicates that it has retained stem cell inhibitory properties but has a compromised activity as a monocyte chemoattractant, thus suggesting uncoupling of these two properties of MIP-la. Receptor studies have indicated that the inactivity of Hep Mut on human monocytes correlates with its inability to bind to CCR1, a cloned human MIP-la receptor. In addition, studies using proteoglycan deficient cells transfected with CCRl have indicated that the proteoglycan binding site in MIP-la is a site that is also involved in the docking of MIP-la to the monocyte receptor. The site for interaction with the stem cell receptor must therefore be distinct, suggesting that MIP-la utilizes different receptors for these two different biological processes. Keywords: chemoattractant/inhibition/MIP- lc/ proteoglycan/stem cell
Introduction Chemokines are members of an expanding family of low molecular weight proteins (8-14 kDa) which have been shown to have diverse inflammatory properties both in vitro and in vivo. This family is defined by sequence homology and the presence of a conserved cysteine motif which allows the family to be subdivided into x (CXC) and ,3 (CC) chemokines (Stoeckle and Barker, 1990; Schall, 1991). We and others have been investigating the role of the , chemokine Macrophage Inflammatory Protein-1 alpha (MIP-loc), as an inhibitor of cellular
proliferation and have demonstrated the ability of MIPla to function as a reversible inhibitor of haemopoietic stem cells (Broxmeyer et al., 1990; Graham et al., 1990; Lord et al., 1992) and clonogenic epidermal cells (Parkinson et al., 1993). The in vitro activities of this molecule suggest that it has a therapeutic role as a myelosuppressive agent for use during chemotherapeutic treatments for cancer (Graham and Pragnell, 1991). To facilitate these studies we have also generated mutants of MIP-la which have a reduced tendency for self aggregation (Graham et al., 1994). Other members of the wider chemokine family also display proliferation regulatory functions ranging from stimulation of keratinocyte (Tuschil et al., 1992) or haemopoietic progenitor cell proliferation (Broxmeyer et al., 1990), to inhibition of endothelial cell (Gupta and Singh, 1994; Luster et al., 1995) or megakaryocyte proliferation (Gewirtz et al., 1995). In common with many other cytokines and with other members of the chemokine family (Witt and Lander, 1994), MIP-la binds to proteoglycans such as heparin, heparan sulfate and chondroitin sulfate, however the precise role of proteoglycan binding in chemokine function is not yet clear. Evidence has been presented suggesting a role for proteoglycan presentation of some ,-chemokines in adhesion of T cells to endothelium (Tanaka et al., 1993; Gilat et al., 1994) and it has also been reported that heparin potentiates the activity of the a-chemokine interleukin 8 in neutrophil migration assays (Webb et al., 1993). It is clear however from these studies that proteoglycan presentation is not an absolute requirement for IL8 activity. Furthermore, studies with other ca-chemokines such as IPIO and PF4 have suggested that proteoglycan binding may be directly responsible for the inhibitory functions of these peptides on endothelial cells (Luster et al., 1995). It is known that proteoglycan binding and presentation is an absolute requirement for productive interactions between other cytokines such as the fibroblast growth factors (FGFs) and their cognate receptors (Roghani et al., 1994). More recently, it has been demonstrated that this requirement reflects the need for FGF receptor dimerization prior to signal transduction and that the multivalent proteoglycans simultaneously present a number of ligand molecules, thus bringing receptors into close proximity, which in turn facilitates dimerization (Spivak-Kroizman et al., 1994). Chemokines are known to interact with cell-surface receptors belonging to the seven membrane-spanning G-protein coupled receptor family (Ahuja et al., 1994) such as the CCR- 1 human ,-chemokine receptor cloned by Neote and colleagues (Neote et al., 1993). These receptors are not believed to require dimerization for bioactivity, thus it is possible that any role for proteoglycans in chemokine function may relate more to ensuring localization or local presentation to receptors. A knowledge of the relevance of proteoglycans to chemokine function
6 6 6506 Oxford University Press
Uncoupling inhibition and chemoattraction in MIP-la
is therefore important as an aid to our overall understanding of the role of chemokines in the diverse biological systems in which they function. Studies on the FGFs and on other heparin binding proteins have shown that, in general, proteoglycan binding is defined by clusters of basic amino acid residues in the protein (Jackson et al., 1991). In the chemokine family, studies have been carried out demonstrating the importance of the basic amino acids, most notably within the carboxyterminal ox-helix, for binding of the ox-chemokines to proteoglycans. Mutant variants of PF4 have been generated in which the carboxy-terminal lysine residues have been modified to eliminate proteoglycan binding (Maione et al., 1991); furthermore, carboxy-terminal truncation of IL8 is also known to interfere with proteoglycan binding of this chemokine (Webb et al., 1993). In contrast, little is known about the nature of the interaction between proteoglycans and the 3-chemokines which lack the characteristic carboxy-terminal lysine residues seen in the a-chemokines. We have therefore set about attempting to address the question of the importance of proteoglycan binding to f-chemokine function using MIP- 1 x as a model molecule. Our approach has initially involved identification of the proteoglycan binding site within MIP-la and subsequent mutagenesis to generate a non-proteoglycan binding variant of MIP-la. Bioassays with this mutant indicate that proteoglycan binding has little impact on bioactivity, however the site identified as the proteoglycan binding site in MIP-loa is also coincidentally a site involved in binding to the human CCR-1 chemokine receptor. Our data further indicate that a biological consequence of this inability to bind to CCR- 1, which is known to be expressed on monocytes (Combadiere et al., 1995), is the uncoupling of stem cell inhibition from monocyte shape change and locomotion in the non-proteoglycan binding variant of MIP-Ia.
Results Identification of the proteoglycan binding site in MIP-1la As discussed above, proteoglycan binding sites generally comprise clusters of basic amino acid residues and as can be seen from Figure 1 a, there are two such clusters within the MIP- 1 Cx amino acid sequence. In an effort to examine the relative contributions of these two sites to MIP- 1o proteoglycan binding, we synthesized peptides spanning these two regions and examined their ability to bind to heparin-Sepharose affinity columns. Our results show (Figure lb) that whilst peptide 1 showed no significant binding to heparin affinity columns under physiological conditions with 1 00% of the peptide being eluted by 0.1 M NaCl, peptide 2 bound more strongly to such columns, requiring higher salt concentrations for complete elution. Other peptides spanning different regions over the entire length of the MIP-l1x molecule showed no significant affinity for heparin-Sepharose. It appears from these studies therefore, that the basic region encompassed by peptide 2 is likely to be the major determinant of proteoglycan binding in MIP-la. Generation of a non-proteoglycan binding variant of MIP-lia
Molecular modelling of PF4 and heparin has identified residues within PF4 that are likely to be involved in
a t
cc
loI
b x,)
+
-
-
IFSYSRKI RQFIVDY
VFLTKFThRQIQJ
Peptide 1
Peptide 2
t
-
-
-
;,
a
- &C
20
papiide*
2
1
t
2
2
_. ;7.
Fig. 1. Identification of the proteoglycan binding site in MIP-la. (a) Diagrammatic representation of the charge distribution in MIP-la and sequences of the peptides spanning the clusters of basic amino acid residues. (b) Assessment of the binding of peptides 1 and 2 to heparin affinity columns: peptides 1 and 2 were resuspended in 0.02 M Tris pH 7.6 at a concentration of 1 mg/ml and 200 tl of each were applied to heparin affinity columns. The columns were washed with 0.02 M Tris pH 7.6 to elute the non-binding fractions and residual binding peptides were subsequently eluted with 0.1 M and 2 M NaCl in 0.02 M Tris pH 7.6. Results are expressed as percentages of applied protein eluted with each wash.
proteoglycan interactions (Stuckey et al., 1992). Of the seven candidate residues, only one is conserved in the 3-chemokines and it corresponds to the first lysine of the ..KRNR.. motif in peptide region 2. We have therefore concentrated on this basic residue and its immediate neighbour in attempting to generate non-heparin binding variants of MIP- 1 cX. Accordingly, we have used techniques of overlap PCR to generate a mutated variant of MIP-lca, designated Hep Mut, in which we have neutralized these two basic amino acids in the proteoglycan binding cluster (...LTKRNRQI.. to LTNSNRQI...). To assess the impact of this mutation on the proteoglycan binding of the mutant we have passed this molecule down heparin affinity columns and examined its binding properties. The results show (Figure 2) that whilst the wild-type protein binds to heparin-affinity columns and requires up to 0.5 M NaCl for elution (Figure 2a), Hep Mut shows no significant affinity for heparin and is substantially eluted in salt-free Tris buffers (Figure 2b). It has been previously demonstrated (Kolset et al., 1996) that MIP-l1c can bind to serglycin, a major haemopoietic chondroitin sulfate-containing proteoglycan (Ruoslahti, 1988). Thus, to look more widely at the impact of the mutations in Hep Mut on proteoglycan binding by MIPoc, we have also examined the ability of Hep Mut to bind to chondroitin sulfate. The results from these studies indicate that whilst wild-type MIP- lIx binds to chondroitin sulfate-Sepharose affinity columns and requires up to 0.5 M NaCl for complete elution (Figure 2c), Hep Mut again shows no binding to this affinity matrix and is eluted in salt-free buffers (Figure 2d). It should be noted that the apparent affinity of wild-type MIP- I Cx for heparin is higher than that for chondroitin sulfate. These results suggest therefore that interference with two of the basic amino acid residues encompassed by peptide 2, results in a variant of MIP-l1c which is unable to bind to a range of proteoglycans.
6507
G.J.Graham et aL a
b
220 97 66
97
30
30
14
14
kDa
a
66
1
2
3
1 2 3 4
4
6
C
* X;
+x
220
d
220
8
97 66
97 66
30
30
'2
10
220
16
14
18
20
F RACTION NUMBER
20
15-0
4
1;
4
kDa
14
b 1 2 3 4
1 2 3 4
Fig. 2. Analysis of the glycosaminoglycan binding of Hep Mut and MIP-la. Western blots of MIP-la and Hep Mut outlining their binding properties on heparin (a and b) and chondroitin sulfate (c and d) affinity columns: 200 g1 of dimeric MIP-la (Graham et al., 1994) or Hep Mut at a concentration of 0.1 mg/ml in 0.02 M Tris pH 7.6 were applied to heparin affinity or chondroitin sulfate affinity columns and washed through with 0.02 M Tris pH 7.6 (1). Binding proteins were then eluted with 0.1 M (2), 0.5 M (3) and 2 M NaCl (4) in 0.02 M Tris pH 7.6. Fractions collected from each wash were resolved on 17.5% SDS gels, blotted and blots probed using murine MIP-la specific antibodies. Cross-reacting proteins were visualized by chemiluminescence.
Molecular weight analysis of Hep Mut In our previous study analysing the basis for the selfaggregation observed with MIP-la, we hypothesized that self-aggregation is a consequence of an interaction between carboxy-terminal acidic amino acid residues and internal basic residues such as those encompassed by peptide regions 1 and 2 (Graham et al., 1994). To attempt to examine the impact of the neutralizations of basic amino acids in Hep Mut on the native molecular weight of the molecule we have passed the molecule down a gel filtration column. As can be seen in Figure 3, whilst the wild-type MIP-la appears to be a dodecamer with a native mol. wt of ~100 kDa (Figure 3a) under the experimental conditions used, Hep Mut appears to elute from the column at a point corresponding to a native mol. wt of ~18-19 kDa (Figure 3b). Hep Mut appears therefore to be predominantly in the dimeric form; however, the elution profile of the protein from the S200 column is broad and would suggest that this dominant dimer exists in equilibrium with less abundant tetramers and even dodecamers. Given the predominance of the dimeric form however, we can conclude that interference with the charges on the two internal basic amino acid residues compromises the selfaggregation potential of this mutant. To control for any possible effects of the impaired aggregation of this protein on bioactivity, a non-aggregating heparin binding dimeric mutant variant of MIP-la with wild-type activity levels (henceforth referred to as dimeric wild-type MIP-la) was used as a control in all biochemical and functional assays (Graham et al., 1994) including the proteoglycan binding studies outlined in Figure 2.
Biological activity of Hep Mut To examine the impact of the mutations described above on the biological activity of Hep Mut, we have analysed
6508
6
8
10
12
14
16
18a
20
FRACTION NUMBER
Fig. 3. Gel filtration of wild-type MIP-la and Hep Mut. Two hundred microlitres of 1 mg/ml solutions of MIP-la (a) or Hep Mut (b) in PBS were applied to a column of Sephacryl S-200 in an HR 10/30 FPLC column and eluted with PBS. Fractions were collected and the point of elution of the proteins followed at 280 nm. Fractions were Western blotted to further confirm the point of elution of the proteins from the gel filtration column.
its activity in both the CFU-A stem cell inhibitory assay (Pragnell et al., 1988; Holyoake et al., 1993) and the monocyte shape change assay (Islam and Wilkinson, 1988). Results from the CFU-A assay, an in vitro correlate of the CFU-S assay (Lorimore et al., 1990) using either human or murine haemopoietic cells, are shown in Figure 4a and b. Data from the murine CFU-A assays (Figure 4a) indicate that in common with the dimeric wild-type molecule, the Hep Mut is active in inhibiting proliferation of CFU-A haemopoietic stem cells, both peptides showing half maximal inhibition at between 10 and 25 ng/ml. Although not obvious from Figure 4a, in repeated experiments, it appears that whilst both the dimeric wild-type and mutant peptides are active in inhibiting stem cell proliferation, overall, Hep Mut displays slightly reduced activity levels which in general appear to be -70% of those seen with the dimeric wild-type peptide. Similarly, in human CFU-A assays (Figure 4b), Hep Mut was seen to be active in inhibiting colony formation, showing a similar dose response to that seen with the dimeric wildtype protein with half maximal inhibition being seen at between 25 and 50 ng/ml. These results therefore suggest that an inability to bind to proteoglycans does not substantially impair biological activity of MIP- lax as an inhibitor of haemopoietic stem cell proliferation in vitro. Results from an analysis of Hep Mut activity in the monocyte shape change assays designed to examine the inflammatory properties of the molecule are shown in Figure 5. Typically, monocytes display a biphasic response to chemoattractants in these assay systems. Experiments on human peripheral blood monocytes demonstrate (Figure
Uncoupling inhibition and chemoattraction in MIP-la
a
a 50
T 1001l}0
**
T
T
-°
40
CD
30
-J w 0 0 W
:5
TT
40
0
X
20
z
T
H
20
w 0 cr w
10
a.
-
lOpg
lpg
1 pg
C:cntro)
2.
10]
5
25
F"-
-
Opg
ing
g
l0pg 100pg
1 ng
l0ng
50
_ I
g
I
1,ug
100 ng
concentration/ml
riLj -li.Pialpla
b -Ln U)
b
120
**
100
80)
IT
E
-T
90
.u
ICc
0
40
80
.m a
I C
I 2.5
I
25
50
T
70
200 60
ng!mI Fig. 4. Activities of MIP-laz and Hep Mut in the murine and human CFU-A assays. (a) Activities of dimeric MIP-la and Hep Mut in the murine CFU-A assay: MIP-la (filled bars) and Hep Mut (unfilled bars) were tested for activity in the murine CFU-A assay by direct addition (Graham et al., 1992) to a growth factor-containing underlayer. This layer was then overlaid with agar containing murine haemopoietic cells and the assay allowed to incubate for 11 days. The assays were then scored at this time with colonies of >2 mm diameter being scored as stem cell colonies (Lorimore et al., 1990). The data are expressed as means and standard errors and are representative of five separate experiments. (b) Activities of dimeric MIP-la (filled bars) and Hep Mut (unfilled bars) in the human CFU-A assay: assays were set up essentially as described above with the exception that human bone marrow was used as a source of stem cells and the assays allowed to incubate for up to 21 days (Holyoake et al., 1994). The data are expressed as means and standard errors and are representative of four separate experiments.
1
ng/ml
C Co
>
50 -
0 W
40
cr:
5
-
30-
z
"
.11
w
w
20-
a.
.L
5a), that Hep Mut is essentially inactive in inducing shape change in these cells over a wide range of concentrations at which the dimeric wild-type molecule is active (100 pg/ml to 10 ng/ml). There is some evidence of polarizing activity of Hep Mut which is coincident with a second wave of activity of the dimeric wild-type molecule seen at 100 ng/ml. The significance of this peak is uncertain, however, as it precedes cell damage due to peptide mediated toxicity (substantial cell damage is seen at 1 gg/ml) and may therefore not represent a physiological response to these agonists. Blood monocytes also showed no migration into filters in response to Hep Mut, although they responded well to dimeric wild-type MIP-loc (Figure 5b). This suggests therefore that the monocyte shape change and chemotactic functions have been substantially uncoupled from the stem cell inhibitory functions in Hep Mut, at least in human cells. To confirm that a similar functional uncoupling has occurred with respect to murine cells we have also examined the ability of Hep Mut to
10
--Ir
.1
I'll NN
).-
.1
-
-
-T, I
L
A.
I
--
1 pg
lOpg lOOpg
l_
_
1ng
N
N
N
N
l1 lOng 100ng 1pg
concentration/ml Fig. 5. Monocyte shape change and chemoattractant properties of MIP-la and Hep Mut. (a) Activities of dimeric MIP-la (solid line) and Hep Mut (broken line) on human blood monocyte polarization: human peripheral blood monocytes were exposed to increasing concentrations of chemokine, following which the cells were fixed with glutaraldehyde and polarization scored using phase contrast microscopy. Data points represent the mean and standard error obtained from three separate experiments (**P < 0.01). (b) Activities of dimeric MIP-icx (solid line) and Hep Mut (broken line) on human blood monocyte migration in a micropore filter assay. The data points represent the mean and standard error from triplicate filters with five fields being counted per filter (**P < 0.001). (c) Activities of dimeric MIP-lat (solid line) and Hep Mut (broken line) on murine J774 monocytic cells: J774 cells were activated with 10-3 M cAMP for 4 days to facilitate polarization, following which chemokines were added and polarization scored as described above. Data points represent the mean and standard error obtained from three separate
experiments (*P
- - -
-
2 3 1 2 3 1 2 3 0 C)
cn C
c:
-
1000
C)
:)
C
_
C)
1I
2
1
2
3
4I
5--
4
5
nano
b .f
1
2 t
G) >
D
2 3 1 2 3 1 C C) C) C)
s_
CA)
2 3 1 C) C A
2
C)i C)
3 1
2 3 0
C)
:n
Fig. 6. PCR analysis of CC chemokine expression by human and murine monocytes. (a) Human monocyte RNA was obtained and PCR analysis of CC chemokine receptor gene expression carried out as shown. For each gene: 1 = reverse transcriptase (RT) positive; 2 = RT negative control and 3 = genomic DNA control. (b) Murine J774 monocyte RNA was isolated using TRIZOL (Gibco BRL, Paisley, Scotland) and PCR analysis of CC chemokine receptor gene expression carried out as shown. Numbering is as outlined in (a) above with the exception that 3 represents plasmid controls.
polarize murine monocytic cells. Given the difficulties in obtaining primary murine monocytes in sufficient numbers and purity for the polarization experiments, these studies were carried out using the J774 murine monocytic cell line. Using these cells we have been able to demonstrate (Figure Sc) that, again, whilst the dimeric wild-type molecule shows the expected bell-shaped dose-response curve, with activity peaking at 1 ng/ml, Hep Mut appears to be essentially inactive on these cells. We have, however, noticed a low but consistent level of activity of Hep Mut at 1 ng/ml; however, this activity did not reach significance. Thus it appears that there is also substantial functional uncoupling of stem cell inhibition from monocyte shape change in the murine system.
Examination of monocyte chemokine receptor expression With a view to defining the subspecies of receptor that may be responsible for the lack of responsiveness of monocytes to Hep Mut, we have examined CC chemokine receptor expression in both human monocytes and in murine J774 cells, using PCR. Our results demonstrate (Figure 6a and b) that the predominant CC chemokine receptor expressed in human and murine monocytes is the CCR1 receptor. Of the cloned human receptors that are known to bind MIP-lIa, i.e. CCR4 (Power et al., 1995) and CCR5 (Samson et al., 1996), only very low levels of CCR5 expression were detected in the monocytes (Figure 6a). On murine cells (Figure 6b), we have detected lowlevel expression of CCR3 and of D6, a novel murine MIP-la binding receptor (R.J.B.Nibbs and G.J.Graham, 6510
7 6
7
8
Molar
Fig. 7. Binding of MIP-la and Hep Mut to CCRI-transfected CHO cells. Wild-type CHO cells transfected with a full-length CCR1 cDNA were plated at 2 X 105 per well in six well tissue culture plates. Sixteen hours later binding studies were performed using increasing concentrations of either dimeric wild-type MIP- la (broken line) or Hep Mut (solid line). c.p.m. represent specific binding counts obtained following subtraction of counts obtained from binding in the presence of a 100-fold excess of cold competing dimeric wild-type MIP-la or Hep Mut. This binding curve is representative of three replicate experiments.
manuscript in preparation). This D6 receptor, in contrast to CCR3, also binds Hep Mut and thus may be responsible for the low but reproducible activity level of Hep Mut on the murine J774 monocytic cell line. Receptor binding properties of Hep Mut To attempt to examine the role of CCR1, the dominant monocyte CC chemokine receptor, in the uncoupling of stem cell inhibition from monocyte chemoattraction, we have obtained the cDNA for the human CCR1 receptor (Neote et al., 1993) and have generated stable transfectants of Chinese hamster ovary (CHO) cells (which do not show cell-surface binding of MIP-la) expressing this receptor (KI-CCRl). As can be seen from Figure 7, binding studies on this KI-CCRl CHO cell line indicate that whilst dimeric wild-type MIP-la binds well to the CCR1 receptor, Hep Mut shows no detectable binding above background. It appears therefore, that this KlCCR1 CHO cell line recapitulates the biological situation seen with human monocytes, i.e. activity with the wildtype protein but not with the Hep Mut, and suggests that an inability to interact with CCR1 or its murine homologue may underlie the impairment of monocyte shape change activity observed with Hep Mut. There are two possible explanations for the inability of Hep Mut to bind to CCR1. Either, in a manner analogous to that seen with the FGFs (Roghani et al., 1994; SpivakKroizman et al., 1994), interaction with proteoglycans is required for binding of MIP-la to this receptor or alternatively, what appears to be the heparin binding site in MIP-la is also coincidentally a site involved in monocyte receptor binding. To attempt to distinguish between these two alternative explanations, we have generated stable transfectants of two mutant CHO cell lines that are compromised in their ability to generate cellsurface proteoglycans. Binding affinities obtained from Scatchard analysis of binding studies performed on these lines using the dimeric wild-type MIP-la indicate (Table I) that MIP-la can bind to CCR1 on transfected wild-
Uncoupling inhibition and chemoattraction in MIP-la Table I. Binding affinities of dimeric wild-type MIP-loa to proteoglycan mutant CHO cells stably transfected with CCR1 CHO cell line
KD (M)
Kl (wild type) 677 (no heparan sulfate) 745 (no proteoglycan)
3.7 x 1o-8 2.7 x I 0-8 1 .9x iO-
6000
5000
E
QO
3000
2000
.
1000
Contro.
2C
H
H
200 2000 .g;rrnl Hepariln
20 o. rr
2000 200 Heparanl SIiipZlaie
Fig. 8. Heparin and heparan sulfate competition for MIP-lI binding to CCRI-transfected CHO cells. Five nM 125I-labelled dimeric wild-type MIP-la was bound to CCRI-transfected wild-type CHO cells in the presence of increasing concentrations of heparin or heparan sulfate with or without a 100-fold excess of cold competing MIP-I1a. Counts represent specific counts obtained by subtracting counts obtained in the presence of cold competing MIP-la from total counts. Data are representative of three separate experiments.
type CHO cells (Ki) as well as to CCRl-transfected CHO-677 cells that are deficient in heparan sulfate (Lidholt et al., 1992) or to CCR1-transfected CHO-745 cells that lack any detectable cell-surface proteoglycans (Esko et al., 1985). In our hands the KD values measured for MIP-l1x binding to CCRl on the CHO cells are somewhat higher than previously published KDs. We assume that this is a consequence of the expression of this receptor in a heterologous cell type which may not have the appropriate G-protein complement for high affinity binding. In agreement with this suggestion, we have so far been unable to demonstrate calcium mobilization in the CCRI CHO cells following MIP-l1x binding.
Heparin does not compete for MIP-1la binding to CCR1 The above data suggest that proteoglycan presentation is not important for interaction of MIP-l1c and CCR1 and that the region identified as the proteoglycan binding site for MIP-l1u is also coincidentally involved in the interactions of MIP-l1x with CCRI. We have therefore examined the possible ability of heparin and heparan sulfate to inhibit binding of MIP-lu to this receptor in the KI-CCRI CHO cell transfectants by blocking the receptor binding site. Studies using a wide range of proteoglycan concentrations have failed to demonstrate any inhibitory potential of heparin or heparan sulfate to block MIP-loc binding to CCR1 (see Figure 8).
Discussion As described above, a number of cytokines and growth factors are dependent on proteoglycan binding for proper function (Roghani et al., 1994; Spivak-Kroizman et al., 1994) and in addition, proteoglycan binding may be important in the local binding, sequestration and presentation of a range of important regulatory factors (Gordon et al., 1987; Roberts et al., 1988). The proteoglycan binding properties of MIP-lu and other members of the chemokine family have thus led us to examine the role of such interactions in the in vitro biological functions of MIP- lc. The site identified as the proteoglycan binding site in MIP-loc (see Figure 1) represents a well conserved 3-chemokine motif which appears to have a high surface accessibility as demonstrated by NMR structural analysis of both human MIP-1,f (Lodi et al., 1994) and RANTES (Skelton et al., 1995). It should be pointed out that whilst the basic region encompassed by peptide 1 in Figure appears to show little proteoglycan binding, this does not rule out the possibility that it may, through structural orientation, contribute to the overall MIP- loc-proteoglycan interaction. Indeed in platelet factor 4, amino acids occupying a similar position appear to be involved in defining a ring of basic residues around the tetrameric PF4 molecule which together regulate interactions with proteoglycans (Zhang et al., 1994). Nevertheless it is clear from our data that interference with two of the basic residues in peptide region 2 is sufficient to disrupt proteoglycan binding in MIP- lo. Given the conservation of this region within the 13-chemokines, it is likely that disruption of this site will also interfere with proteoglycan binding by a wider range of chemokines. Functional analyses of the non-heparin binding mutant of MIP- luc indicate that it has retained stem cell inhibitory properties on both human and murine stem cells and that accordingly, proteoglycan interactions are not a prerequisite for the function of MIP- x in this assay system. In addition, to investigate the possibility that proteoglycans may enhance MIP- lu activity in the assay system, without being essential for activity, we have compared the potency of dimeric wild-type MIP- lu in murine CFU-A assays in the presence or absence of a range of proteoglycan concentrations. Our results indicate (unpublished data) that wild-type MIP- luc activity is unaffected by the presence or absence of heparin, heparan sulfate or chondroitin sulfate. We believe therefore, on the basis of the combined data from Hep Mut and the proteoglycan-containing CFU-A assays, that proteoglycan binding has little importance in the stem cell inhibitory function of MIP- lo. Analysis of the activity on monocyte shape change assays has, however, revealed a striking activity difference between Hep Mut and dimeric wild-type MIP- lu in that Hep Mut has an impaired ability to induce monocyte shape change and locomotion in these cells. This inactivity is seen on both human and murine monocytic cells; however, as shown in Figure 5, we have consistently observed a low level of residual Hep Mut polarizing activity towards the murine J774 cells. Whilst this activity did not reach levels of significance, it was observed on all occasions. Given the inability of Hep Mut to bind to CCRI, this suggests that alternative receptors may be
6511
G.J.Graham et aL
responsible for this low level of Hep Mut activity. Our PCR
data suggest that whilst CCR1 is clearly the predominant chemokine receptor expressed on J774 cells, these cells also express small amounts of both CCR3 and D6, a novel MIP- 1a receptor that has recently been cloned in our laboratory. Binding studies have indicated that Hep Mut does not bind to CCR3 but that it does show binding to D6 (R.J.B.Nibbs and G.J.Graham, manuscript in preparation). It is therefore possible, although it remains to be proven, that the low level Hep Mut activity on the J774 cells is a consequence of interaction with the D6 receptor, and that the overall activity of the wild-type molecule towards J774 cells is a combination of the interaction of this molecule with at least CCR1 and D6. It is of interest to note that human monocytes also express low levels of CCR5. However, the absence of Hep Mut activity on human monocytes would indicate either that Hep Mut does not bind to this receptor or that this receptor has no role to play in regulating the monocyte chemoattractant properties of MIP-la. Our studies on the murine homologue of CCR5 suggest that Hep Mut does indeed bind to this receptor, which may further suggest that this receptor indeed has no active role in monocyte chemoattraction. The high levels of CCR1 expression in the murine and human monocytes suggested that this may be the receptor that is largely responsible for signalling the monocyte migratory properties of MIP-lca. We have now been able to demonstrate that human CCR1 (Neote et al., 1993), when transfected into CHO cells, binds the wild-type MIP-lx but not the Hep Mut, thus recapitulating the situation seen on the monocytes (Figure 7). Further analysis has revealed that CCR1 transfectants of mutant CHO cells lacking cell-surface proteoglycans, bind wild-type MIP-la (Table I). This indicates that the lack of binding of Hep Mut to the CCRl-transfected CHO cells is not a consequence of its inability to interact with proteoglycans but relates more to the fact that the region in MIPla identified as the proteoglycan binding site is also coincidentally a site that is involved in the binding of MIP-lax to the monocyte receptor. This further suggests that the receptor responsible for articulating the stem cell inhibitory properties of MIP- lac is distinct from the CCR1 monocyte receptor and further that it requires regions of MIP- l a outside the heparin binding region for interaction. The inability of soluble glycosaminoglycans to inhibit binding of wild-type MIP-la to CCR1-expressing CHO cells (Figure 8) is intriguing and is in apparent contradiction with the above suggestion regarding the co-identity of the proteoglycan and monocyte receptor binding regions in MIP- 1 a. One possible explanation for this observation is that the region mutated in Hep Mut is not in itself directly involved in receptor binding but is a region of the molecule important as a structural determinant, without which the structure of the monocyte receptor binding site becomes sufficiently altered to block binding. Such a structural change is likely to be subtle, however, as Hep Mut retains stem cell inhibitory activity and we have previously demonstrated the requirement for structural integrity of MIP- 1 a for this function (Graham et al., 1993). Furthermore, our analysis of murine CC receptors indicates that Hep Mut has retained the ability to bind to CCR5 and D6, thus again arguing against dramatic struc-
6512
tural alterations in this mutant. A more plausible explanation lies in the likely differences in affinity of MIP-la for its receptor and for proteoglycans. In our hands, MIPla is a relatively weak proteoglycan binding peptide and thus is unlikely to compete to any significant extent with the receptor, for ligand binding. It may be, therefore, that concentrations of proteoglycan in excess of those outlined in Figure 8 are required for competition. If, as suggested by the present studies, the region mutated in Hep Mut is important in binding to the monocyte receptor, then identification of this region may allow us to design small peptide antagonists with the capacity to block the monocyte function of MIP-la and other j-chemokines sharing this conserved motif. Such antagonists may have a therapeutic role in the treatment of a range of inflammatory pathologies in which MIP-la is implicated and in which it utilizes this region of the molecule for receptor interaction. Studies are currently underway in our laboratory to test these possibilities.
Materials and methods Generation of the non-heparin binding mutant Mutation of two residues within the murine MIP-la heparin binding site was achieved using PCR-directed mutagenesis. The PCR primers used were as follows: (i) 5' GAC GCC CAT GGC GCC ATA TGG AGC TGA CAC CCC GAC TGC C; (ii) 5' ATA CGT GGA TCC TCA GGC ATT CAG TTC CAG GTC AGT GAT GTA TTC; (iii) 5' GAT CTG CCG GTT GCT GTT AGT CAG GAA AAT G; (iv) 5' CAT TTT CCT GAC TAA CAG CAA CCG GCA GAT C. The mutant cDNA was generated in three steps. First, a PCR was set up using wild-type MIP1 a as a template and using primers 1 and 3 to generate half of the mutant cDNA. The second half of the mutant cDNA was generated by PCR using primers 2 and 4. The products of these two reactions, which are complementary to each other in the overlapping regions coded for by primers 3 and 4, were visualized on 1.5% low melting point agarose gels and purified from gel fragments using Elutip (Schleicher & Schuell). A final round of PCR to generate the full-length mutant cDNA was therefore carried out using the products of the above reactions and primers 1 and 2 which encompass the entire cDNA sequence for the mature MIP-la peptide. All PCRs were performed for 30 cycles according to standard protocols. The final mutated cDNA was confirmed by sequencing and was then cloned into a bacterial expression vector at the NcoI-BamHI sites underlined in primers 1 and 2 above.
Production of mutant protein
As MIP-la and related mutants are non-glycosylated, we have been able to generate large quantities of wild-type (dimeric wild type: see below) and mutant MIP-la peptides using a bacterial expression system. The wild-type and mutant murine proteins were expressed as fusion proteins with a 30 kDa leader sequence in Escherichia coli. The recombinant fusion proteins are predominantly found as insoluble proteins in the inclusion bodies. Subsequent to solubilization and refolding of inclusion bodies using previously described techniques (Weir and Sparks, 1987), the fusion proteins were partially purified on a Fast Flow Q anion exchange column. Recombinant peptides were cleaved from the partially purified fusion protein by factor Xa treatment. Final purification of the peptides was achieved using a C4 reverse-phase HPLC column. Purified proteins were sterile filtered and stored as a solution in 30% acetonitrile, 0.1% trifluoroacetic acid. The identities of the wild-type and mutant proteins were confirmed by amino-terminal sequencing and amino acid analysis. For all the experiments described in the present report, the 'wild-type' MIP-la is a non-aggregating dimeric mutant of murine MIP-la which displays wildtype activity levels (Graham et al., 1994). This specific control was important in the present studies as Hep Mut appears to be predominantly in the dimeric form at the concentrations studied.
Chromatographic procedures Glycosaminoglycan affinity chromatography. Heparin-Sepharose (Pharmacia Biotech Ltd) or chondroitin sulfate-Sepharose was packed
Uncoupling inhibition and chemoattraction in MIP-la into an empty HR 5/5 FPLC column (1 ml column volume) and washed with five column volumes of 2 M NaCl-0.02 M Tris pH 7.6 and then equilibrated with a similar volume of 0.02 M Tris pH 7.6. For determination of the relative contributions of the two basic residues outlined in Figure 1 to proteoglycan binding, the two peptides shown were chemically synthesized (Alta Biosciences, Birmingham) and brought to a concentration of 1 mg/ml in 0.02 M Tris pH 7.6. Two hundred microlitres of each peptide was applied to the heparin affinity column and washed through with 5 ml of 0.02 M Tris buffer. Elution was performed in two steps. First, weakly binding peptide species were eluted with 0.1 M NaCl-Tris and finally more strongly binding peptides were eluted with 2 M NaClTris. The point of elution of the peptides from the affinity column was assessed by observing the A280 of the eluting buffers. For analysis of the effects of the PCR generated mutation of the basic residues in peptide region 2 on MIP-la binding to proteoglycan columns, 200 1l of either dimeric wild-type MIP-ltx (Graham et al., 1994) or Hep Mut at a concentration of 0.1 mg/ml in 0.02 M Tris pH 7.6 were loaded onto the heparin column or the chondroitin sulfate column and washed through with five column volumes of Tris buffer. Elution was performed in steps to 0.1 M NaCl-0.02 M Tris pH 7.6; 0.5 M NaCl-Tris and finally to 2 M NaCl-Tris. The presence of the MIP- 1 a or mutant proteins in the column fractions was assessed by Western blotting (see below). Gel filtration. The native molecular weight of MIP-laz (structural wild type) or Hep Mut was assessed by bringing the protein to 1 mg/ml in PBS (a concentration at which the wild-type molecule is known to have a native molecular weight of -100 kDa) and applying 100 lt of this solution to a Sephacryl S-200 matrix loaded into an HR1O/30 FPLC column. The protein was then eluted in the same buffer and the point of elution measured by both A280 and by Western blotting. Molecular weight standards were also run on the column for precise assignment of native molecular weight to the mutant peptide.
Western blotting Samples from the column chromatography runs were prepared for SDSPAGE by bringing to 2% SDS, 250 mM DTT and heating at 100°C for 2-3 min. Following addition of glycerol and bromophenol blue, the samples were loaded onto 15% SDS-polyacrylamide gels (Laemmli, 1970) and electrophoresis carried out until the bromophenol blue dye band reached the bottom of the gel. Resolved proteins in the gels were then blotted onto Immobilon P filters for 30 min at 50 mA using a semidry blotting apparatus. The blots were then washed for 30 min in three changes of Blotto (5% powdered milk/0.1% NP-40 in PBS) and placed in a solution of the primary anti-MIP-la antibody (R & D systems Inc., MN, USA) at a 1:1000 dilution in Blotto for 1 h. The blots were then washed three times in Blotto over 30 min following which they were placed in a solution of the secondary antibody (rabbit anti-goat HPD conjugated) again at a 1:1000 dilution in Blotto for 1 h. The blots were finally washed with multiple changes of PBS/0.1% Tween 20 over 30 min following which the antibody-antigen complexes were visualized using the ECL reagents supplied by Amersham International (Little Chalfont, Bucks, UK). In vitro bioassays The CFU-A assay. The murine CFU-A assay was performed as described previously (Pragnell et al., 1988; Lorimore et al., 1990; Graham et al., 1992). Briefly, normal murine bone-marrow cells (5 X 103/ml) were incubated in 0.3% agar in a-minimal Eagle's medium (a-MEM), 25% DHS on top of an underlayer of 0.6% agar/a-MEM/25% DHS with L929 and AF- 1 conditioned media as sources of macrophage-colony stimulating factor (M-CSF) and granulocyte macrophage-CSF (GMCSF) respectively. Assays were allowed to develop for up to 11 days in 5% 02, 10% CO2 and 85% N2, following which stem cell-derived colonies were scored as described previously, and the effects of exogenously added MIP- 1 a and associated mutants on stem cell colony formation assessed. The human CFU-A assay was carried out as described previously (Holyoake et al., 1993). Briefly, 5 X 103/ml human bone-marrow cells were resuspended in 0.3% agar/a-MEM/25% DHS and plated on top of a 0.6% agar/a-MEM underlayer containing 25% FCS and sources of GM-CS, M-CSF and stem cell factor (SCF). These assays were allowed to develop for up to 21 days in 5% 02, 10% CO2 and 85% N2, following which inhibition was scored as described previously (Graham et al., 1992).
The monocyte shape change assay. Mononuclear cells were prepared from human blood using lymphocyte-separating medium and washed in siliconized glass tubes to prevent loss of monocytes by adherence. Monocytes were then separated from the lymphocytes by differential
adhesion on fibronectin-coated surfaces using the method of Ackerman and Douglas (1978). The proportion of monocytes polarized by MIP-la or the mutated variants following a 30 min exposure was scored after glutaraldehyde fixation by phase contrast microscopy. Remaining lymphocytes may also be polarized by MIP-la. The two cell types were morphologically distinct but the monocyte specific marker CD14 was also used to distinguish spherical and polarized monocytes from lymphocytes by a modified APAAP method (Newman and Wilkinson, 1992). The J774 murine monocytic cell line direct from culture was poorly responsive to chemoattractants, but a motile phenotype was induced by culture for 4 days in dibutyryl cyclic AMP (l0-3 M). These cells were then tested in a polarization assay as above. The micropore filter assay. To check that shape change observed in the polarization assay reflected the locomotor properties of monocytes, the cells were allowed to migrate into nitrocellulose filters (8 ,tm pore size) for 2 h in response to MIP-la or Hep Mut placed below the filter. Following fixation and staining of the filters, migration was scored by measuring the distance migrated by the leading front of cells (Zigmond and Hirsch, 1973).
PCR analysis of the expression of chemokine receptors in human and murine monocytes Generation of monocyte RNA. Total RNA was generated from dibutyryl cyclic AMP activated J774 cells using TRIZOL according to the manufacturer's instructions (Gibco BRL, Paisley, Scotland). Human monocyte total RNA was a generous gift from Dr Dave Greaves (Sir William Dunn School of Pathology, University of Oxford, UK). PCR analysis. RNA was DNase treated for 15 min at room temperature to remove genomic DNA contamination, phenol/chloroform extracted, ethanol washed and resuspended at 1 tg/ml in water. RNA was heat denatured at 70°C for 5 min then cooled on ice for a further 5 min. Reverse transcription and PCR were carried out essentially according to instructions in the Perkin Elmer RNA PCR CORE kit. The PCR was
performed under the following conditions: 94°C for 2 min; 30 cycles of 94°C 1 min, 58°C 1 min and 72°C 1.5 min. This was followed by a final 10 min at 72°C prior to cooling and storing at 4°C. The primers used were as follows: human CCR 1: 5' TTCCTCACGAAAGCCTACGAG 3'; 3' AGTGCGTGTAGGCGATCACC 5' (product size 320 bp). Human CCR4: 5' TAGAGACCCTGGTGGAGCTAG 3'; 3' AAAGTTCATTGACTCTGCATTTC 5' (product size 330 bp). Human CCR5: 5' CTTCTTACTGTCCCCTTCTGG 3'; 3' GGCATAGATGATGGGGTTGATG 5' (product size 657 bp). Murine CCR1: 5' TCTTCTATTCTTCCTCCTCTGGA 3'; 3' AGAGCTCATGTTCTCCTGTGGA 5' (product size 318 bp). Murine CCR3: 5' TTGCAGGACTGGCAGCATTG 3'; 3' AAATAAGACGGATGGCCTTGTG 5' (product size 246 bp). Murine CCR4: 5' AAGAATGAGAAGAAGAACAGAGC 3'; 3' CTGGACATGTCAGCCGAGTAGAC 5' (product size 325 bp). Murine CCR5: 5' TCTTTACCAGATCTCAGAAAGAAGG 3'; 3' AGTCTCTGTTGCCTGCATGG 5' (product size 359). Murine D6: 5' GTTTTCTTCATGCTGTGGTTCC 3'; 3' TGTGCCCTGAACCCAGAGAC 5' (product size 694). Generation of CHO cell/CCR1 stable transfectants
All CHO cell lines used in the present study were generously supplied by Professor Jeffrey Esko (Department of Biochemistry, University of Alabama in Birmingham, Birmingham, AL 35294, USA). Wild-type CHO cells (KI) and the proteoglycan deficient variants pgsD-677 (lacks heparan sulfate; Lidholt et al., 1992) and pgsA-745 (makes little if any cell-surface proteoglycan; Esko et al., 1985) were maintained in Ham's F12 medium supplemented with 10% fetal calf serum with subculturing every 3-4 days. A full-length cDNA clone for the human CCR1 MIPla receptor (Neote et al., 1993) was kindly supplied by Dr Tom Schall (DNAX Research Institute, Palo Alto, CA, USA). Stable CCR1 transfectants were generated for each CHO cell line using the calcium phosphate transfection methodology of Chen and Okayama (1987). Transfected CHO cells were selected by culturing in 1.6 ,g/ml geneticin.
Receptor binding studies
Murine MIP-la was labelled with 125lodine as described previously (Graham et al. 1993). Briefly, 10 [tg of MIP-la or mutant in PBS was added to 20 ,ug of immobilized IODOGEN in an Eppendorf tube and the reaction initiated by addition of 37 MBq of Na[1251]. The reaction was allowed to proceed on ice for 15 min following which incorporated and unincorporated iodine were separated by desalting on GF5 disposable desalt columns. The peak of protein associated radioactivity was assessed
6513
G.J.Graham et al.
by y-counting and the labelled protein fractions were pooled and stored refrigerated for up to 1 month, after which time activity has been previously shown to decrease. For binding studies. 2x 105 CHO cells per well were plated in sixwell plates and allowed to adhere and proliferate for 16 h. The growth
medium was then removed and the cell monolayers washed three times with PBS. Increasing concentrations of radiolabelled MIP-loc or Hep Mut were then applied to the plates in binding buffer (Dulbecco's MEM/ 10% DHS/0.2% sodium azide) in the presence or absence of a 100-fold excess of cold competitor. Plates of cells were then incubated at room temperature for 60 min, following which the cell layers were washed three times with PBS and finally solubilized in 1% SDS and the cell associated counts measured in a y-counter. Scatchard analyses were performed using the LIGAND program of Munson and Robard (1980).
Acknowledgements The authors would like to thank Professor John Wyke and Dr Johann DeBono for their helpful comments on this manuscript. The generous gift of human monocyte RNA from Dr Dave Greaves is gratefully acknowledged. The authors would like to acknowledge the expert assistance of Amy Isaacson and Jagan Medicherla in the preparation of the wild-type and mutant peptides. The involvement of Drs Ian Hayes and Jane Mackenzie in the earlier parts of this study is also gratefully acknowledged. These studies were supported by grants from the Cancer Research Campaign to G.J.G. and I.B.P. S.O.K. is supported by grants from the Norwegian Cancer Society and P.C.W. by the Wellcome Trust.
References Ackerman,S.K. and Douglas,S.D. (1978) Purification of human monocytes on microexudate-coated surfaces. J. Im7nmunlol., 120, 1372-1374.
Ahuja,S.K., Gao,J.-L. and Murphy,P.M. (1994) Chemokine receptors and
molecular mimicry. Immunol. Today, 15, 281-287. Broxmeyer,H.E., Sherry,B., Lu,L., Cooper,S., Oh,K.,
Tekamp-Olson,P.,
Kwon,B.S. and Cerami,A. (1990) Enhancing and suppressing effects
of recombinant murine macrophage inflammatory proteins on colony formation in vitro by bone marrow myeloid progenitor cells. Blood, 76, 1110-1116. Chen,C. and Okayama,H. (1987) High efficiency transformation of mammalian cells by plasmid DNA. Mol. Cell. Biol., 7, 2745-2752. Combadiere,C., Ahuja,S.K., VanDamme,J., Tiffany,H.L., Gao,J.-L. and Murphy,P.M. (1995) Monocyte chemoattractant protein-3 is a functional ligand for CC chemokine receptors 1 and 2B. J. Biol. Chem., 270, 29671-29675. Esko,J.D., Stewart,T.E. and Taylor,W.H. (1985) Animal cell mutants defective in glycosaminoglycan biosynthesis. Proc. Natl Acad. Sci. USA, 82, 3197-3201. Gewirtz,A.M., Zhang,J., Ratajczak,J., Ratajczak,M., Park,K.S., Li,C., Yan,Z. and Poncz,M. (1995) Chemokine regulation of human megakaryocytopoiesis. Blood, 86, 2559-2567. Gilat,D., Hershkoviz,R., Makori,Y.A., Vlodavsky,I. and Lider,O. (1994) Regulation of adhesion of CD4+ T lymphocytes to intact or heparinase-treated subendothelial extracellular matrix by diffusible or anchored RANTES and MIP-1p. J. Imnuunol., 153, 4899-4906. Gordon,MY., Riley,G.P., Watt,S.M. and Greaves,M.F. (1987) Compartmentalization of a haematopoietic growth factor (GM-CSF) by glycosaminoglycans in the bone marrow microenvironment. Nature, 326, 463-466. Graham,G.J. and Pragnell,I.B. (1991) Treating cancer: the potential role of stem cell inhibitors. Eur J. Cancer, 27, 952-953. Graham,G.J., Wright,E.G., Hewick,R., Wolpe,S.D., Wilkie,N.M., Donaldson,D., Lorimore,S. and Pragnell,I.B. (1990) Identification and characterisation of an inhibitor of haemopoietic stem cell proliferation. Nature, 344, 442-444. Graham,G.J., Freshney,M.F., Donaldson,D. and Pragnell,I.B. (1992) Purification and biochemical characterisation of human and murine stem cell inhibitors (SCI). Growth Factors, 7, 151-160. Graham,G.J., Zhou,L., Weatherbee,J.A., Tsang,M.L.-S., Napolitano,M., Leonard,W.J. and Pragnell,I.B. (1993) Characterisation of a receptor for MIP-lo: and related proteins on human and murine cells. Cell Growth Differ, 4, 137-146. Graham,G.J. et al. (1994) Aggregation of the chemokine MIP-ltx is a
6514
dynamic and reversible phenomenon: Biochemical and biological analyses. J. Biol. Chem., 269, 4974-4978. Gupta,S.K. and Singh,J.P. (1994) Inhibition of endothelial cell proliferation by platelet factor-4 involves a unique action on S-phase progression. J. Cell Biol. 127, 1121-1127. Holyoake,T.L., Freshney,M.G., Konwalinka,G., Haun,M., Petzer,A., Fitzsimons,E., Lucie,N.P., Wright,E.G. and Pragnell,I.B. (1993) Mixed colony formation in vitro by the heterogeneous compartment of multipotential progenitors in human bone marrow. Leukaetnia, 7, 207-213.
Islam,L.N. and Wilkinson,P.C. (1988) Chemotactic factor-induced polarisation, receptor redistribution and locomotion of human blood monocytes. Immunology, 64, 501-507. Jackson,R.L., Busch,S.J. and Cardin,A.D. (1991) Glycosaminoglycans: Molecular properties, protein interactions and role in physiological processes. Physiol. Rev., 71, 481-539. Kolset,S.O., Mann,D.M., Uhlin-Hansen,L., Winberg,J.-O. and Ruoslahti,E. (1996) Serglycin-binding proteins in activated macrophages and platelets. J. Leukocyte Biol., 59, 545-554. Laemmli,U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680-685. Lidholt,K., Weinke,J.L., Kiser,C.S., Lugemwa,F.N., Bame,K.J., Cheifetz,S., Massague,J., Lindahl,U. and Esko,J.D. (1992) A single mutation affects both N-acetylglucosaminyltranferase and glucuronosyltranferase activities in a Chinese hamster ovary cell mutant defective in heparan sulfate biosynthesis. Proc. Natl Acad. Sci. USA, 89, 2267-2271. Lodi,P.J., Garrett,D.S., Kuszewski,J., Tsang,M.L.-S., Weatherbee,J.A., Leonard,W.J., Gronenborn,A.M. and Clore,G.M. (1994) High resolution solution structure of the f-chemokine hMIP- 1 A by multidimensional NMR. Science, 263, 1762-1767. Lord,B.I., Dexter,T.M., Clements,J.M., Hunter,M.A. and Gearing,A.J.H. (1992) Macrophage inflammatory protein protects multipotent haemopoietic cells from the cytotoxic effects of hydroxyurea in vivo. Blood, 79, 2605-2609. Lorimore,S.A., Pragnell,I.B., Eckman,L. and Wright,E.G. (1990) Synergistic interactions allow colony formation in vitro by murine haemopoietic stem cells. Leuk. Res., 14, 481-489. Luster,A.D., Greenberg,S.M. and Leder,P. (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. Med., 182, 219-231. Maione,T.E., Gray,G.S., Hunt,A.J. and Sharpe,R.J. (1991) Inhibition of tumour growth in mice by an analogue of platelet factor 4 that lacks affinity for heparin and retains potent angiostatic activity. Cancer Res., 51, 2077-2083. Munson,P.J. and Robard,D. (1980) A versatile computerised approach for the characterisation of ligand binding systems. Anal. Biochem., 107, 220-239. Neote,K., DiGregorio,D., Mak,J.Y, Horuk,R. and Schall,T.J. (1993) Molecular cloning, functional expression and signalling characteristics of a C-C chemokine receptor. Cell, 72, 415-425. Newman,I. and Wilkinson,P.C. (1992) Methods for phenotyping polarised and locomotive human lymphocytes. J. Immunol. Methods, 147,43-50. Parkinson,E.K., Graham,G.J., Daubersies,P., Burns,J.E., Heufler,C., Plumb,M., Schuler,G. and Pragnell,I.B. (1993) A haemopoietic stem cell inhibitor (SCIIMIP-Ic1) also inhibits clonogenic epidermal keratinocyte proliferation. J. Invest. Dermatol., 101, 113-117. Power,C.A., Meyer,A., Nemeth,K., Bacon,K.B., Hoogewerf,A.J., Proudfoot,A.E.I. and Wells,T.N.C. (1995) Molecular cloning and functional expression of a novel CC chemokine receptor cDNA from a human basophilic cell line. J. Biol. Chem., 270, 19495-19500. Pragnell,I.B. et al. (1988) The effect of stem cell proliferation regulators demonstrated with an in vitro assay. Blood, 72, 196-202. Roberts,R., Gallagher,J., Spooncer,E., Allen,T.D., Bloomfield,F. and Dexter,T.M. (1988) Heparan sulphate bound growth factors: a mechanism for stromal cell mediated haemopoiesis. Nature, 332, 376-379. Roghani,M., Mansukhani,A., Dell'Era,P., Bellosta,P., Basilico,C., Rifkin,D.B. and Mosactelli,D. (1994) Heparin increases the affinity of bFGF for its receptor, but is not required for binding. J. Biol. Chlemii., 269, 3976-3984. Ruoslahti,E. (1988) Structure and biology of proteoglycans. Annu. Rev: Cell Biol., 4, 229-255. Samson,M., Labbe,O., Mollereau,C., Vassart,G. and Parmentier,M. (1996) Molecular cloning and functional expression of a new human CC chemokine receptor gene. Biochemistry, 35, 3362-3367.
E.Yp.
Uncoupling inhibition and chemoattraction in MlP-la Schall,T.J. (1991) Biology of the RANTES/SIS cytokine family. Cytokine, 3, 165-183. Skelton,N.J., Aspiras,F., Ogez,J. and Schall,T.J. (1995) Proton NMR assignments and solution conformation of RANTES, a chemokine of the C-C type. Biochemistry, 34, 5329-5342. Spivak-Kroizman,T. et al. (1994) Heparin-induced oligomerisation of FGF molecules is responsible for FGF receptor dimerisation, activation and cell proliferation. Cell, 79, 1015-1024. Stoeckle,M.Y. and Barker,K.A. (1990) Two burgeoning families of platelet factor-4 related proteins: mediators of the inflammatory response. New Biol., 2, 313-323. Stuckey,J.A., St Charles,R. and Edwards,B.F.P. (1992) A model of the platelet factor 4 complex with heparin. Proteins: Structure Funct. Genet., 14, 277-287. Tanaka,Y., Adams,D.H., Hubscher,S., Hirnao,H., Siebenlist,U. and Shaw,S. (1993) T-cell adhesion induced by proteoglycan-immobilised cytokine MIP-1 beta. Nature, 361, 79-82. Tuschil,A., Lam,C., Haslberger,A. and Lindley,I. (1992) Interleukin-8 stimulates calcium transients and promotes epidermal cell proliferation. J. Invest. Dermatol., 99, 294-298. Webb,L.M.C., Ehrengruber,M.U., Clark-Lewis,I., Baggiolini,M. and Rot,A. (1993) Binding to heparin sulphate or heparin enhances neutrophil responses to interleukin 8. Proc. Natl Acad. Sci. USA, 90, 7158-7162. Weir,M.P. and Sparks,J. (1987), Purification and renaturation of recombinant interleukin-2. Biochem. J., 245, 85-91. Witt,D.P. and Lander,A.D. (1994) Differential binding of chemokines to glycosaminoglycan subpopulations. Curr. Biol., 4, 394-400 Zhang,X., Chen,L., Bancroft,D.P., Lai,C.K. and Maione,T.E. (1994) Crystal structure of recombinant platelet factor 4. Biochemistry, 33, 8361-8366. Zigmond,S.H. and Hirsch,J.G. (1973) Leukocyte locomotion and chemotaxis. New methods of evaluation and demonstration of a cellderived chemotactic factor. J. Exp. Med., 137, 387-410. Received on March 19, 1996; revised on July 23, 1996
6515
