Two Basic Amino Acids in the Second Inner Loop of the Interleukin-8 Receptor Are Essential for Gα16 Coupling*

The involvement of basic residues of interleukin(IL)-8 receptors in coupling to the Gi and G16 proteins was investigated by using a series of IL-8 receptor mutants. Substitution of the basic amino acids in the third inner loop of the receptor does not alter the abilities of the receptor mutants to activate recombinant Gα16 or phosphoinositide-specific phospholipase C (PLC) β2 expressed in COS-7 cells. However, an IL-8 receptor mutant with double mutations at residues Lys158 and Arg159of the second inner loop loses its abilities to activate Gα16 but retains its ability to activate PLC β2. The activation of PLC β2 by an IL-8 receptor that is sensitive to pertussis toxin has been previously demonstrated to be mediated through Gβγ. Surprisingly, the IL-8 receptor mutants with substitution of Ala for either residue Lys158 or Arg159 can still activate Gα16, which suggests that either of the two basic residues in the second inner loop of the IL-8 receptor is sufficient for Gα16 coupling.

The involvement of basic residues of interleukin(IL)-8 receptors in coupling to the Gi and G16 proteins was investigated by using a series of IL-8 receptor mutants. Substitution of the basic amino acids in the third inner loop of the receptor does not alter the abilities of the receptor mutants to activate recombinant G␣16 or phosphoinositide-specific phospholipase C (PLC) ␤2 expressed in COS-7 cells. However, an IL-8 receptor mutant with double mutations at residues Lys 158 and Arg 159 of the second inner loop loses its abilities to activate G␣16 but retains its ability to activate PLC ␤2. The activation of PLC ␤2 by an IL-8 receptor that is sensitive to pertussis toxin has been previously demonstrated to be mediated through G␤␥. Surprisingly, the IL-8 receptor mutants with substitution of Ala for either residue Lys 158 or Arg 159 can still activate G␣16, which suggests that either of the two basic residues in the second inner loop of the IL-8 receptor is sufficient for G␣16 coupling.
Many biologically active molecules transduce their signals through specific cell-surface receptors. Some of the receptors interact with heterotrimeric GTP-binding proteins (G proteins) 1 (1,2). Molecular cloning has revealed the existence of genes encoding at least 20 G␣, 5 G␤, and 12 G␥ subunits in mammals (3). These subunits can form a variety of heterotrimers that serve to connect specific cell surface receptors to a large number of different effectors including at least 4 PLC ␤ isoforms and many adenylyl cyclases, as well as several specific ion channels (1-3). One of the intriguing questions posed by this apparent complexity is how signal transduction circuits are organized so that different kinds of receptors can be connected to effectors through various G proteins and coordinate a variety of responses in a large number of different cells. The specificity of some of the circuits is determined no doubt by developmental regulation of the expression of genes that encode the receptors, G proteins and effectors. In addition, subcellular localization may contribute to the specificity to a certain extent. However, the primary determinant for formation of a specific signal transduction circuit lies in specific proteinprotein interactions.
Work has been done to understand the molecular basis of the specificity in receptor-G protein interactions (4). Amino acid sequences that are involved in activation of G␣q have been mapped to the third cytoplasmic (inner) loops of the ␣1Badrenergic receptor, the m1 muscarinic receptor, and the glutamate receptors by using various chimeras (5)(6)(7)24). Although these sequences share no significant amino acid sequence homology, they appear to be different from the sequences involved in activating G␣s (8,9). Recently, we have found that different ␣1B-adrenergic receptor sequences are involved in coupling to different ␣ subunits of the Gq class (10). Furthermore, receptor sequences in other inner loops have also been implicated in the involvement of G protein coupling. Studies using receptorderived peptides have implicated that the second inner loop of the N-formyl peptide receptor may be involved in G protein interaction (11,12).
We have previously demonstrated that the IL-8 receptor (IL-8R), like many other chemoattractant receptors including the C5a and formyl-methionyl-leucyl-phenylalanine receptors, can couple to both G16 and Gi proteins (14). In this report, we will report our investigation of the IL-8R sequences involved in coupling to G16 but not to Gi by site-directed mutagenesis. Our results indicate that two basic amino acid residues in the second inner loop of the IL-8R are essential for coupling to G␣16 but not to Gi, whereas the basic residues in the third inner loop are not required for coupling to either Gi or G16.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-COS-7 cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum under 5% CO 2 at 37°C. The COS-7 cells were seeded the day before transfection into 24-well plates at a density of 1 ϫ 10 5 cell/ml. The medium was removed the next day, and 0.5 ml of Opti-MEM (Life Technologies, Inc.), which contained 5 g of lipofectamine (Life Technologies, Inc.) and 1 g of plasmid DNA, was added to each well. 5 h later, the transfection medium was replaced by the culture medium. The cells were labeled with 10 Ci/ml myo-[2-3 H]inositol the following day, and the levels of inositol phosphates (IPs) were determined one day later as previously described (13). All the cDNAs used in this study were constructed in the pCMV expression vector (13).

SDS-polyacrylamide Gel Electrophoresis and Western
Blot--Equal numbers of transfected cells were solubilized in the SDS sample buffer and loaded to 12% SDS-polyacrylamide gels. The proteins were then electroblotted onto nitrocellulose membranes and detected with antibodies indicated in the figure legends.
Receptor Binding Assays-COS-7 cells in 12-well plates were transfected with the cDNA encoding the IL-8R or its mutants. After 48 h, the cells were washed with phosphate-buffered saline and incubated with varying amounts of 125 I-IL-8 (3000 Ci/mmol, NEN Life Science Products) in phosphate-buffered saline containing 1 mg/ml bovine serum albumin for 1 h at 4°C. After washing three times with ice-cold phosphate-buffered saline containing bovine serum albumin, the cells were lysed in 0.5 ml of 0.2 N NaOH, and 0.1-ml aliquots were taken for counting in a scintillation counter. The nonspecific binding was determined by measuring binding of 125 I-IL-8 to nontransfected cells. The numbers of specific IL-8-binding sites (B max ) and dissociation constants (K d ) were determined by the Scatchard analysis (24).
Construction of IL-8R Mutants-All the IL-8R mutants listed in Fig.  1 were generated by polymerase chain reaction with the high fidelity DNA polymerase, pfu (Stratagene), and each of the mutations was confirmed by DNA sequencing.

RESULTS AND DISCUSSION
The IL-8 receptors were previously shown to couple to two G proteins, Gi and G16 (14). To investigate whether different receptor sequences are involved in coupling to these two G proteins, we have generated a series of mutated receptors as tabulated in Fig. 1. Since it was postulated that the BBXXB (B stands for basic amino acid, and X stands for any amino acid) motif might be responsible for Gi coupling (15), we first investigated whether the BBBXXB motif (residues Lys 247 to Arg 251 ) in the third intracellular loop of the human type B IL-8 receptor is involved in Gi coupling. We constructed the IL-8 receptor mutants, m1, m2, and m3, by substitution of Ala residues for the amino acids Lys 246 , His 247 , and Arg 248 , respectively. These mutants were tested for their abilities to couple to Gi and G16 in a previously established transient transfection assay (10, 14, 16 -19) to characterize the G protein-coupling specificity for the IL-8 receptors. The COS-7 cells used in the assay system do not contain endogenous IL-8 receptors, PLC ␤2, or G␣16, although they contain Gi2 and PLC ␤1 (13,14,17,20). Thus, IL-8 did not elicit any significant elevation of IP levels in cells expressing the IL-8 receptor and its mutants in the absence of G␣16 or PLC ␤2 (Fig. 2A). To test the G16 coupling of the IL-8 receptor mutants, we cotransfected COS-7 cells with cDNAs encoding G␣16 and the IL-8 receptor or its mutants, and IL-8-induced accumulation of IPs was determined. As shown in Fig. 2B, IL-8 induced marked PTx-resistant accumulation of IPs in cells coexpressing G␣16 and the IL-8 receptor or its mutants, m1, m2, or m3, which suggests that these three IL-8 receptor mutants, like the wild-type IL-8 receptor, can still couple to G␣16. To test the Gi coupling, we cotransfected COS-7 cells with the cDNAs encoding PLC ␤2 and the receptors. The IL-8 receptor was previously shown to couple to endogenous Gi proteins of COS-7 cells to release G␤␥, which then activates recombinant PLC ␤2 (14). As shown in Fig. 2C, there was clear IL-8-induced accumulation of IPs in cells coexpressing PLC ␤2 and the IL-8 receptor, m1, m2, or m3, and the ligand-induced responses were mostly PTx-sensitive. Therefore, these data indicate that the IL-8 receptor mutants can couple to both G16 and Gi in transfected COS-7 cells. To test further the importance of the triple basic amino acids in the third inner loop of the IL-8 receptor, these basic amino acids (Lys 246 -His 247 -Arg 248 ) were mutated to three alanine residues. As shown in Fig. 2, B and C, the IL-8R mutant can still couple to recombinant G␣16 and to PLC ␤2 via endogenous Gi proteins. Thus, it is clear that the BBBXXB (residues Lys 247 to Arg 251 ) motif at the N-terminal end of the third intracellular loop of the IL-8 receptor is by no means involved in the Gi coupling or the G16 coupling.
Another basic amino acid residue in the third inner loop, Lys 240 , was also investigated for its involvement in coupling to G16 or Gi. We constructed the mutant m5 by substitution of an Ala residue for the residue Lys 240 . The mutant m5 was subjected to the same tests as m1-4. The tests showed that m5, like the others, can couple to G16 and Gi. Thus, we conclude that the basic residues inside the third inner loop of the human type B IL-8 receptor are not involved in coupling to G16 or Gi.
Search of the IL-8 receptor sequence revealed a BBXXXB (Lys 158 -Lys 163 ) motif in the second inner loop of the receptor. To test whether the basic residue doublet (Lys 158 -Arg 159 ) is involved in the G protein coupling, we replaced the doublet with two Ala residues creating the mutant m8 (Fig. 1). By testing the mutant in the same cotransfection assay, we found that m8 can induce IP accumulation only in cells coexpressing PLC ␤2 (Fig. 3B) but not in those coexpressing G␣16 (Fig. 3A), which suggests that m8 can couple only to Gi but not to G␣16. Neither m6 nor m7, which have substitution of an Ala residue for one of the basic residue doublets, loses its ability to couple to G␣16 (Fig. 3). The ability of m8 to activate PLC ␤2 has eliminated the possibility that the mutations in m8 greatly changed the conformation of the receptor. Nevertheless, we also did the ligand-binding assay with 125 I-IL-8. The expression level of m8 and its affinity for IL-8 are similar to those of the wild-type IL-8 receptor, m6 and m7 (Fig. 1). In addition, we also determined the expression levels of G␣16 in cells coexpressing m8, m6, m7 and the wild-type IL-8 receptor. No major differences were noticed (Fig. 3C). Therefore, it is reasonable to conclude that either of the basic residues (Lys 158 and Arg 159 ) is apparently sufficient to retain the ability of the receptor to couple to G16 and that the presence of either of them is essential for the G16 coupling, although these two residues do not appear to play a significant role in the Gi coupling.
We have previously demonstrated that different ␣1-adrenergic receptor sequences are involved in coupling to G␣q/11 and G␣14. However, sequences involved in G␣16 coupling have not been elucidated. Recent reports (18,21) shows that G␣16 appears to be promiscuous in its coupling to various receptors. Almost all of the G protein-coupled receptors thus far tested, including Gq-, Gi-, and Gs-coupling receptors, can couple to G␣16 in transfected COS-7 cells (18,21). This coupling promiscuity suggests that most G protein-coupling receptors possess the sequence elements and/or conformation required for interaction with and activation of G␣16. Our results provide an insight into what the requirements are. The basic residues Lys 158 and Arg 159 may constitute the sequence that interacts with and/or activates G␣16 or may be critical for formation of the receptor conformation required for coupling with G␣16. More studies (knowledge of the three-dimensional structure of the receptor) are needed to understand exactly how these two basic residues are involved in G␣16 coupling. Our data also indicate that the BBXXB motif in the third loop of IL-8R is not essential for either G␣i or G␣16 coupling. These data are consistent with the observation that residue Met 241 in the third loop, as well as other non-charged amino acid residues in the second loop of IL-8R, are involved in coupling to G␣i2 (22).
Receptor consensus sequences for G protein-coupling were being pursued vigorously in the past. No such sequences have, however, been identified. Therefore, it is now generally believed that each individual receptor possesses specific receptor coupling elements, which were mostly found in the third inner loops of various receptors. G␣16 is an intriguing subunit. It lacks receptor coupling specificity; it couples to various G protein-coupling receptors ranging from Gs to Gi and Gq-coupling receptors. We have been looking for the receptor elements that are required for G␣16 coupling in both ␣1B-adrenergic receptors (10), but these elements have been eluding us until we identified the dual basic amino acids in the second loop of the IL-8 receptor. Although we did not identify consensus sequences for G16 coupling, our results are of great significance. 1) These results unequivocally prove that the second loop is involved in G protein-coupling specificity in contrast to most other studies, which usually only implicate the third inner loops. 2) This is the first time that G␣16-coupling elements have been identified. 3) The element required for G␣16 coupling is not required for G␣i coupling. 4) The basic residues in the second and third inner loops, which have been widely believed to be involved in Gi coupling, are not important for Gi coupling by the IL-8 receptor. Therefore, this work provides us with a better understanding of the specific interactions between receptors and G proteins. In addition, the receptor mutants that show limited yet defined G protein-coupling specificity would be useful in determining the specific in vivo function of signal transduction pathways mediated by specific receptors and G proteins.