Selective G protein coupling by C-C chemokine receptors.

The C-C chemokines are major mediators of chemotaxis of monocytes and some T cells in inflammatory reactions. The pathways by which the C-C chemokine receptors activate phospholipase C (PLC) were investigated in cotransfected COS-7 cells. The C-C chemokine receptor-1 (CKR-1), the MCP-1 receptor-A (MCP-1Ra), and MCP-1Rb can reconstitute ligand-induced accumulation of inositol phosphates with PLC beta2 in a pertussis toxin-sensitive manner, presumably through G beta gamma released from the Gi proteins. However, these three receptors demonstrated different specificity in coupling to the alpha subunits of the Gq class. While none of the receptors can couple to Galphaq/11, MCP-1Rb can couple to both Galpha14 and Galpha16, but its splicing variant, MCP-1Rb, cannot. Since MCP-1Ra and -b differ only in their C-terminal intracellular domains, the C-terminal ends of MCP-1Rs determine G protein coupling specificity. CKR-1 can couple to Galpha14 but not to Galpha16, suggesting some of the C-C chemokine receptors, unlike the C-X-C chemokine receptors, discriminate against Galpha16, a hematopoietic-specific Galpha subunit. The intriguing specificity in coupling of the Gq class of G proteins implies that the chemokines may be involved in some distinct functions in vivo. The commonality of the chemokine receptors in coupling to the Gi-Gbetagamma-PLC beta2 pathway provides a potential target for developing broad spectrum anti-inflammatory drugs.

Chemokines are a large family of small (8 -10 kDa), inducible, secreted, proinflammatory cytokines, which are produced by various cell types. Members of the chemokine family share 20 -90% homology in their amino acid sequences. The sequences usually have four conserved cysteine residues except lymphotactin. On the basis of the positions of the cysteine residues, the chemokine family can be divided into three subfamilies: the C-X-C or ␣ family, the C-C or ␤ family, and the C or ␥ family. The ␣ family includes IL-8, 1 GRO (growth-related oncogene), NAP-2, ENA-78, platelet factor 4, IP-10, and GCP-2, while the ␤ family includes macrophage chemotactic protein (MCP)-1, -2, and -3, RANTES (regulated upon activation, normal T cell expressed and secreted), macrophage inflammatory protein (MIP)-1␣ and -1␤, I309, and C10 (for reviews see Refs. [1][2][3]. The newly discovered ␥ family has only one member, lymphotactin. Lymphotactin, unlike other chemokines, has only two conserved cysteine residues (4). The exact physiological and pathophysiological functions of these factors are not yet clearly defined; however, it is generally believed that their main function is recruitment and activation of leukocytes at the site of inflammation.
CKR-1 and MCP receptors have typical structural characteristics of G protein-coupled receptors, and they induce cytosolic Ca 2ϩ efflux (7,9), presumably through activation of phospholipase C (PLC). Five cDNAs that encode the ␣ subunits of the G q class have been characterized, G␣ q , G␣11, G␣14, G␣15, and G␣16 (11), all of which can activate all isoforms of PLC ␤, PLC ␤1-4, to stimulate the release of inositol phosphates (IPs) (12)(13)(14)(15)(16)(17). COS-7 cells contain G␣ q and G␣11 but not G␣14, G␣15, or G␣16 (15). The expression of G␣15 and G␣16 was detected only in hematopoietic cells (G␣15 may be the mouse counterpart of human G␣16) (18 -20), while G␣14 is expressed in some lineage of hematopoietic cells as well as other cell types (19). Many receptors, including the IL-8 receptors, were found to couple to some of the ␣ subunits of the G q class to activate PLC. Recently, the G␤␥ subunits of G proteins were also found to activate specific isoforms of PLC ␤. The G␤␥-linked pathway may account for the PTX-sensitive activation of PLC mediated by the IL-8 receptors in mature leukocytes (6).
Since the C-C chemokines play important roles in chemotaxis of monocytes and some T cells, we characterized the G protein-coupled signal transduction pathways for the three C-C chemokine receptors by the cotransfection assay system in COS-7 cells. We found that the C-C chemokine receptors showed different specificity in coupling to the G␣ subunits of the G q class, while the receptors can all couple to the G i proteins to activate PLC ␤2 through G␤␥.

EXPERIMENTAL PROCEDURES
Cell Culture, Transfection, and Assay-COS-7 cells were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. For transfection, COS-7 cells (5 ϫ 10 4 cells/well) were seeded in 24-well plates the day before transfection. Plasmid DNAs (0.5 g/well) were premixed with 1.7 l of Lipofectamine (Life Technologies, Inc.) and added to cells. The cells were then labeled with [ 3 H]inositol (5 Ci/ml) 24 h later. The next day, ligand-induced accumulation of IPs was determined as described in Ref. 6. In brief, cells were lysed in 10% perchloric acid and neutralized with KOH. IPs were retained by AG1-X8 ion exchange resin and eluted with formic acid. Portions of eluted samples were counted in a scintillation counter. The basal IP level in COS-7 cells is about 2000 dpm. Transfection or cotransfection with cDNAs encoding the C-C chemokine receptors, G␣14, PLC ␤1, Lac Z, did not alter the basal level. Cells transfected with the G␣16 and PLC ␤2 cDNA increased the basal levels to 2500 and 3500 dpm, respectively. MCP-1 and MIP-1␣ were purchased from R&D Systems. All the assays were repeated at least three times. The representative ones were shown.
cDNA Cloning-The cDNAs encoding CKR-1, MCP-1Ra, and MCP-1Rb were cloned from human THP-1 monocytic cells by polymerase chain reaction using primers based on the published sequences (7,9). The sequences were verified by DNA sequencing. The cloned receptors were ligated into pcDNA/AMP (Invitrogen).
Ligand Binding Assay-Cos-7 transfectants were incubated with 125 I-labeled ligands (Amersham Corp., 2000 Ci/mmol) in Dulbecco's modified Eagle's medium containing 0.25% bovine serum albumin on ice for 1 h, and the cells were washed by phosphate-buffered saline containing 0.25% bovine serum albumin three times. Finally the cells were solubilized in 0.1 N NaOH, and aliquots were counted by a ␥-counter. The maximum binding sites and affinities were determined by Scatchard analyses.

RESULTS AND DISCUSSION
We tested in cotransfected COS-7 cells whether the newly cloned C-C chemokine receptors, including the MCP-1Ra, MCP-1Rb, and CKR-1, can couple to the ␣ subunits of the G q class of G proteins. We have previously shown that receptors that can couple to G␣ q or G␣11 gave ligand-induced accumulation of IPs in COS-7 cells transfected with the receptor cDNAs (21). Thus, to test whether these C-C chemokine receptors can couple to G␣ q or G␣11, we transfected the cDNAs corresponding to each of the C-C chemokine receptors into COS-7 cells and determined ligand-induced accumulation of IPs. There was little MIP-1␣-induced accumulation of IPs in cells expressing CKR-1, and neither was there MCP-1-induced accumulation of IPs in cells expressing MCP-1Ra or MCP-1Rb (Fig. 1A). These results indicate that these receptors cannot couple to endogenous G␣ q or G␣11. To test whether the receptors can couple to other members of the G q class, we cotransfected COS-7 cells with each of the receptor cDNAs and the cDNA corresponding to G␣14 or G␣16. We and others have previously found that all the receptors tested in the cotransfection assay, including ␣1A, -B, and -C (21), ␤2-adrenergic receptors (22), the m2-muscarinic receptor, D1-dopamine receptor, V2,V1a-vasopressin receptor, A2a-adenosine receptor, -opioid receptor, 5-1a, 1c/2c serotonin receptors, and thrombin receptor (17) can couple to G␣16. However, neither CKR-1 nor MCP-1Ra can couple to G␣16, since cells coexpressing G␣16 and CKR-1 or MCP-1Ra showed little ligand-induced accumulation of IPs (Fig. 1A). Interestingly, MCP-1Rb, the alternative splicing variant of MCP-1Ra, gave ligand-dependent release of IPs when coexpressed with G␣16 (Fig. 1A), suggesting that MCP-1Rb can couple to G␣16. The activation of G␣16 by MCP-1Rb was insensitive to PTX (Fig. 1B). Furthermore, these C-C chemokine receptors demonstrated different selectivity in coupling to G␣14; CKR-1 and MCP-1Rb can couple to G␣14, while MCP-1Ra cannot (Fig. 1A). The concentration-dependent responses to ligand indicate a mean effective concentration (EC 50 ) for MCP-1Rb-mediated activation of G␣16 and G␣14 of about 7 nM.
We have demonstrated in many of our reports (6,15,21,23) that coexpression of one protein does not significantly affect the expression of others. Nonetheless, we determined the expression of G␣16 and G␣14 in cells cotransfected with cDNA encoding CKR-1, MCP-1Ra, or MCP-1Rb to eliminate the possibility that the inabilities of CKR-1 and MCP-1Ra to couple to G␣16 or G␣14 were the results of lower expression levels of the proteins. As shown by Fig. 1, C and D

Selective G Protein Coupling by C-C Chemokine Receptors 3976
1Rb. Therefore, the inabilities of CKR-1 and MCP-1Ra to couple to G␣16 or G␣14 are not due to variations in expression levels.
The responses to the C-C chemokines (including MCP-1, MIP-1␣, and RANTES) in monocytic phagocytes were found to be mostly PTX-sensitive (1-3), yet the signal transduction pathways mediated by the ␣ subunits of the G q class are PTXresistant (11). PTX is a bacterial toxin, which modifies the C-terminal Cys residues of the G␣ i and G␣ o subunits. The modification prevents interactions between receptors and G proteins. Recently, we proposed a novel pathway to explain the PTX sensitivity; receptors interact with PTX-sensitive G proteins to release the G␤␥ subunits, which then activate PLC ␤2 (6,23). Since there are abundant G i proteins (predominantly G i 2 as well as some G i 3) (24,25) and the PLC ␤2 proteins in leukocytes (26), the G i -G␤␥-PLC ␤2 pathway is likely to occur in vivo. To test whether the C-C chemokine receptors can couple to endogenous PTX-sensitive G proteins of COS-7 cells to activate PLC ␤2, we transfected COS-7 cells with cDNA encoding PLC ␤2 and cDNA encoding each of the C-C chemokine receptors. COS-7 cells contain endogenous G␣ i 2 but not G␣ o proteins, and they contain endogenous PLC ␤1 but not PLC ␤2 as determined by specific antibodies (14). The accumulation of IPs in response to varying concentrations of MCP-1 or MIP-1␣ was determined. All three receptors can induce activation of PLC ␤2 with EC 50 of 0.5 nM for CKR-1 and 3 nM for MCP-1Ra and MCP-1Rb (Fig. 2). In addition, we found that the ligand-induced responses in cells coexpressing the receptors and PLC ␤2 were PTX-sensitive (Fig. 2). Thus, we conclude that all three C-C chemokine receptors can couple to endogenous PTX-sensitive G proteins, presumably the G i 2 protein, to activate PLC ␤2 via G␤␥ (the G␣ i subunits cannot directly activate PLC ␤ (14)). The finding that MCP receptors can inhibit adenylyl cyclase activity in A293 human kidney cells expressing the receptor confirms our notion that the receptor can couple to the G i proteins (27). The inability of the chemokine receptors to activate endogenous PLC ␤ (Fig. 1A) or recombinant PLC ␤1 (data not shown) is consistent with our previous observation that G␤␥ could not activate PLC ␤1 in the cotransfection system (6,21). In addition, we found that MCP-1 could not activate PLC in cells expressing CKR-1 and that MIP-1␣ could not induce IP formation in cells expressing MCP receptors.
Although the C-C chemokine receptors can couple to the G i -G␤␥-PLC ␤2 pathway, these receptors demonstrate interesting specificity in coupling to the ␣ subunits of the G q class. While none of the three receptors couples to G␣ q /11, MCP-1Rb can couple to both G␣16 and G␣14, but its splicing variant MCP-1Ra cannot couple to either G␣14 or G␣16. CKR-1 couples to G␣14 but not to G␣16. The differences between MCP-1Ra and MCP-1Rb in G protein coupling indicate that the C-terminal intracellular domains are critical in determining the G protein coupling specificity, since these two receptors differ only in the C-terminal ends (9). Moreover, the finding further supports our previous notion, drawn from our study of the ␣1B-adrenergic receptor, that different receptor sequences are required for activation of different G␣ subunits of the G q class (28). The study of the ␣1B-adrenergic receptor indicates that the ␣1B-adrenergic sequences required for activation of G␣14 are located in the third intracellular loop, whereas the sequences required for activation of G␣16 do not appear to be localized within the third inner loop. This report, however, points out that the sequences in the C-terminal intracellular domain are critical for activation of both G␣14 and G␣16. We interpret the apparent discrepancy to suggest that G proteininteracting sequences on different receptors may be located at different sites or that there exist multiple G protein-interacting sites on a receptor so that alteration of any one of them abolishes the ability of the receptor to couple to the G protein. In this report we did not account for the influences of different G␤␥ subunits on the coupling of these receptors to different G␣ subunits because there were no significant differences observed for different G␤␥ subunits in interaction with G␣ i 2 or in regulation of PLC ␤2 (15) or of adenylyl cyclase activities (29). Furthermore, the same system (COS-7 cells) was used in the studies; thus, the differences in the coupling of these chemokine receptors to different G␣ subunits cannot be attributed to G␤␥.
The physiological relevance of the pathways mediated by G␣ i , G␣16, and G␣14 is not clear. All these G␣ subunits were found in various hematopoietic cells. Although more systematic studies of the expression of these G␣ subunits are needed, previous studies suggest that there are very abundant G␣ i subunits with the majority of G␣ i 2 and some G␣ i 3 in leukocytes (24,25) and that the levels of the G␣ i subunits increase along with differentiation (18). G␣16 and PLC ␤2 was detected only in hematopoietic cells. G␣16 was detected in neutrophils, monocytes, lymphocytes, and erythrocytes as well as various hematopoietic progenitor cells, and its expression in HL-60 promyeloid cells decreases by 90% after differentiation (18). These results, in addition to the findings that responses to chemokines in mature leukocytes were mostly PTX-sensitive, suggest that the G i -linked pathway may be the predominant one in chemokine-mediated effects in mature leukocytes, such as chemotaxis and activation of leukocytes. If this hypothesis is correct, the activation of PLC ␤2 by G␤␥ would be an excellent target for developing broad spectrum anti-inflammatory drugs, because all the known chemokine receptors, including the C-C chemokine receptors, can couple to the G i -G␤␥-PLC ␤2 pathway. The fact that PLC ␤2 is expressed only in hematopoietic cells may limit potential side effects.
Recently, some evidence indicates that chemokines may be directly or indirectly involved in the regulation of hematopoiesis; MIP-1␣ inhibits proliferation of the hematopoietic stem cells (30), and the IL-8 receptor-null mice have expanded populations of neutrophils and B cells, in addition to their reduced abilities to respond to inflammatory stimuli (31). The G␣16linked pathway may play a role in hematopoiesis as well as in other hematopoietic functions, although there is a lack of evidence. Nevertheless, regulation of expression levels by differentiation, specificity in interactions between receptors and G proteins and between G proteins and effectors, and diversity of molecular natures of receptors, G proteins, and effectors in leukocytes underlie the molecular basis for the complex func-

Selective G Protein Coupling by C-C Chemokine Receptors 3977
tion of signal transduction networks in the hematopoietic system. Alternative splicing further expands the signal-processing capabilities of eukaryotic cells.