Characterization of the Binding Site on the Formyl Peptide Receptor Using Three Receptor Mutants and Analogs of Met-Leu-Phe and Met-Met-Trp-Leu-Leu*

The formyl peptide receptor (FPR) is a chemotactic G protein-coupled receptor found on the surface of phagocytes. We have previously shown that the formyl peptide binding site maps to the membrane-spanning region (Miettinen, H. M., Mills, J. S., Gripentrog, J. M., Dratz, E. A., Granger, B. L., and Jesaitis, A. J. (1997) J. Immunol. 159, 4045–4054). Recent reports have indicated that non-formylated peptides, such as MMWLL can also activate this receptor (Chen, J., Bernstein, H. S., Chen, M., Wang, L., Ishi, M., Turck, C. W., and Coughlin, S. R. (1995) J. Biol. Chem. 270, 23398–23401.) Here we show that the selectivity for the binding of different NH2-terminal analogs of MMWLL or MLF can be markedly altered by mutating Asp-106 to asparagine or Arg-201 to alanine. Both D106N and R201A produced a similar change in ligand specificity, including an enhanced ability to bind the HIV-1 peptide DP178. In contrast, the mutation R205A exhibited altered specificity at the COOH terminus of fMLF, with R205A binding fMLF-O-butyl >fMLF-O-methyl > fMLF, whereas wt FPR bound fMLF >fMLF-O-methyl ∼fMLF-O-butyl. These data, taken together with our previous finding that the leucine side chain of fMLF is probably bound to FPR near FPR 93VRK95 (Mills, J. S., Miettinen, H. M., Barnidge, D., Vlases, M. J., Wimer-Mackin, S., Dratz, E. A., and Jesaitis, A. J. (1998)J. Biol. Chem. 273, 10428–10435.), indicate that the most likely positioning of fMLF in the binding pocket of FPR is approximately parallel to the fifth transmembrane helix with the formamide group of fMLF hydrogen-bonded to both Asp-106 and Arg-201, the leucine side chain pointing toward the second transmembrane region, and the COOH-terminal carboxyl group offMLF ion-paired with Arg-205.

The formyl peptide receptor (FPR) 1 is a chemoattractant G protein-coupled receptor found on the surface of phagocytes. It is thought to play an important role in allowing phagocytic cells to recognize the presence of bacteria (1), which are a source of formyl peptides (2,3). In addition, it recognizes and is activated by peptides derived from the GP-41 envelope protein of the human immunodeficiency virus type I (HIV-1) (4,5). Recent studies with FPR-deficient mice indicate that they exhibit an increased susceptibility to infection with Listeria monocytogenes and that neutrophils from these knockout mice fail to exhibit chemotaxis in response to fMLF (6).
The formyl peptide receptor was originally identified based on its ability to bind the formylated peptide, fMLF (1). Six different chemotactic peptides have been isolated from Escherichia coli, including fMLF (3), but the only similarity between them was an NH 2 -terminal formyl methionine, suggesting that this moiety is highly important in binding to FPR. Studies using NH 2 -terminal analogs of MLF had indicated that the formyl group had significant effects on the ligand binding affinity for neutrophil FPR. Free amino, desamino, and acetylated derivatives of MLF were all 3000-fold lower in affinity than fMLF (7). Substitution of the formyl group of fMLF with a tert-butyloxycarbonyl group made the ligand an antagonist of low affinity (8). However, the formyl group may be less essential than originally thought. N-Butyloxycarbonyl MLF exhibits agonist activity (9) with an affinity similar to fMLF, and phenyl and tolyl isourea derivatives of MLF exhibit activity similar to or greater than fMLF (10). On the other hand, most aliphatic isourea derivatives of MLF exhibit low affinity antagonist activity similar to what is observed with tert-butyloxycarbonyl-MLF (10). This difference indicates that FPR exhibits a high degree of specificity for NH 2 -terminal modifications of MLF, and that the specificity for the formyl group is not absolute. In addition, other reports have indicated that non-formylated pentapeptides can activate FPR. Both MNleLFF and MMWLL are effective activators of FPR (11,12), and acetyl-MNleLFF is more potent than fMLF (12), indicating that these pentapeptides exhibit somewhat different NH 2 -terminal specificities than does MLF.
Human FPR is one of three receptors in the human FPR family, which includes, FPR, the lipoxin A 4 receptor (15)(16)(17)(18) (also called FPRL1 and recently shown to bind and be activated by serum amyloid A (Ref. 19) and several other peptide ligands (Refs. 5, 20, and 21)), and FPRL2. Asp-106 is found in most species variants of FPR and FPRL1 receptors sequenced so far (15)(16)(22)(23)(24)(25). It is also found in FPRL2 of primates (24). Arg-201 and Arg-205 (or their equivalents) and all residues in between ( 201 RGIIR 205 ) are conserved in most species variants of both FPR and FPRL1. However, the FPRL2 receptors contain FLILH in these positions (24). The mouse does not have a homolog of FPRL2 but instead has six genes related to FPR and FPRL1 (25). Three of these genes (named Fpr1, Fpr-rs1, and Fpr-rs2) conserve residues Asp-106, Arg-201, and Arg-205, while Fpr-rs3, Fpr-rs4, and Fpr-rs5 have His or Asn in place of Asp-106; a Thr, Ser, or Ala in place of Arg-201; and a Ser or Asn in place of Arg-205. Thus, Asp-106, Arg-201, and Arg-205 might be important in determining ligand specificity in the FPR family.
Here we show that the selectivity for the binding of different NH 2 -terminal analogs of MMWLL and MLF can be markedly altered by mutating Asp-106 (III-8) or by mutating Arg-201 (V-2). Both D106N and R201A exhibit a very similar change in ligand specificity, including an enhanced ability to bind the HIV derived peptide, DP178. In contrast, the mutation R205A (V-2) resulted in altered specificity at the COOH terminus of fMLF.

MATERIALS AND METHODS
Peptide Synthesis-MMWLL and DP-178 (4) (HIV-1-strain B.FR.HXB2, GP-41 residues 643-678) were synthesized by Macromolecular Resources, Fort Collins, CO. fMLF, fMLF-O-methyl, and tertbutoloxycarbonyl-MLF were obtained from Sigma. Analogs of MMWLL or MLF were synthesized as follows. Peptide in 95% dimethylformamide, 5% triethylamine (ϳ2 mg) was added to a 3-fold molar excess of tolyl isocyanate, tolyl isothiocyanate, or acetic anhydride or formic acid with 3-fold molar excess (9-fold with respect to peptide) of diisopropylcarbodiimide, and incubated at 20°C for 1 h. fMLF esters were synthesized by incubating the free acid with 9-fold molar excess of diisopropylcarbodiimide in the presence of either 100% butanol or 60% dimethyaminoethanol, 40% dimethylformamide. N-Formyl-MMWLL butyl ester was synthesized by incubating the free acid with 9-fold molar excess of diisopropylcarbodiimide in the presence of 100% butanol. The peptides were separated by high performance liquid chromatography, and the products were identified using matrix-assisted time of flight mass spectrometry.
Guanine Nucleotide Binding Analysis-Stimulation of GTP␥S binding by fMMWLL of CHO membranes expressing wt FPR, D106N, R201A, or R205A was determined as described previously (14), using a buffer containing 5 mM HEPES, pH 7.4, 1 mM Mg 2ϩ , 3 M GDP (Buffer A). Effects of Na ϩ , Li ϩ , and K ϩ on GTP␥S binding to CHO cell membranes were determined using a buffer containing 75 mM Tris-Cl, pH 7.4, 12.5 mM Mg 2ϩ , 1 mM EDTA, 3 M GDP (Buffer B) to maintain the conditions previously described by Wenzel-Seifert et al. (26) and Gierschik et al. (27). Incubations were carried out at 30°C for 10 min, and GTP␥S binding never exceeded more than 10% of the total added to assure that binding was in the linear range.
Chemotaxis Assays-Chemotaxis assays were done as described previously by Miettinen et al. (28).

D106N FPR Mutant Localizes NH 2 -terminal Specificity-We
previously observed that the mutation D106N exhibited modest reductions in affinity to fNle-Leu-Phe-Nle-Tyr-Lys-fluorescein (35-fold) but was completely inactive when tested with fMLF for its ability to stimulate GTP␥S binding (13). Previous analysis of the thyrotropin-releasing hormone receptor, which binds the tripeptide pyroglutamic acid-histidine-proline-amide, indicated that a tyrosine in the analogous position to Asp-106 of FPR was the likely interaction site with the cyclic amide (between the side chain carboxyl and the amino terminus) at the NH 2 terminus of this tripeptide (29). To determine whether Asp-106 of FPR might likewise be an interaction site with the NH 2 termini of its ligands, we synthesized four NH 2 -terminal analogs of MMWLL, which is active in its free amino form, and two NH 2 -terminal analogs of MLF (the structures of these NH 2 -terminal modifications are shown in Fig. 1) and compared wt FPR's and D106NЈs ability to bind these analogs. Table I shows the effects of NH 2 -terminal modifications of MLF and MMWLL on ligand binding affinity. D106N exhibited markedly reduced affinity for fMLF, N-formyl-MMWLL (fMMWLL), and N-acetyl-MMWLL as compared with wt FPR, but D106N bound tolylisourea-MLF, tolylisourea-MMWLL, tolylisothiourea-MMWLL, and MMWLL better than did wt FPR. This behavior indicated that Asp-106 is important in NH 2 -terminal selectivity, agreeing with the above mentioned thyrotropin-releasing hormone receptor results (29).
Roles for Arg-201 and Arg-205 in FPR Function-Modeling of FPR suggests that both Arg-201 and Arg-205 are located sufficiently close to Asp-106 that either one of them might form an ion pair with Asp-106 (30). A direct interaction between helix III and helix V in FPR would be similar to what has been shown for rhodopsin, where Glu-122 (III-12) in helix III was shown to directly interact with His-211 (V-10) of helix V using Fourier transform infrared spectroscopy (31). Arg-205 (V-6) is at a membrane depth more similar to Asp-106 (III-8) than is Arg-201 (V-2) and was considered the more likely candidate. However, analog analysis on two G protein-coupled receptors, the C5a receptor and the angiotensin I receptor, which have an Arg and Lys in the position analogous to Arg-205 of FPR, respectively, indicated that this position was the likely interaction site between the COOH-terminal carboxylate of these peptide ligands and their respective receptors (32)(33). These observations, extended to FPR, suggest that Arg-205 might be the site of interaction between the COOH-terminal carboxylate of fMLF with FPR. ment of binding to wt FPR, a 300-fold enhancement of binding to R205A, but no increase in affinity with either D106N or R201A. This indicates that both Asp-106 and Arg-201 are necessary for recognition of the NH 2 -terminal formyl group while Arg-205 affects formyl group recognition only slightly. Acetylation of MMWLL enhanced binding to wt FPR and R205A by 100-fold, but reduced the affinity for D106N and R201A by 7-fold, again demonstrating the importance of Asp-106 and Arg-201 in NH 2 -terminal specificity.
Tolylisourea-MMWLL bound to R201A and D106N with very high affinity and to wt FPR and R205A with somewhat lower affinity. Changing the tolylisourea to a isothiourea reduced the affinity for wt FPR and R205A by 50-and 20-fold respectively, indicating that a carbonyl group on the NH 2 terminus of the peptide is important for binding to wt FPR and to a lesser extent with R205A. However, almost no change in binding affinity was observed for either D106N or R201A when the oxygen of the isourea was replaced by a sulfur, indicating that both Asp-106 and Arg-201 are necessary for recognition of the carbonyl group. Tolylisothiourea-MMWLL bound with much higher affinity to both D106N and R201A than to wt FPR, indicating that these mutations had altered the NH 2 -terminal ligand preference from a formyl moiety to an tolylisothiourea moiety. Moreover, the loss of affinity seen with R201A for the formylated peptides fMMWLL and fMLF, 900-and 400-fold, respectively, was approximately equal to the enhancement seen when MMWLL was formylated (1500-fold), indicating that the contribution of the formyl group to binding was the same whether it was disrupted by mutating Arg-201 or removing it from fMMWLL.
The above data indicate that Asp-106 and Arg-201 probably form an ion pair with each other and that this ion pair imparts formyl group preference to the NH 2 terminus, presumably by direct hydrogen bonding between Arg-201 and the formamide oxygen and between Asp-106 and the formamide NH ( Fig. 2A). Such an ion pair might be expected to exhibit a preference for oxygen over sulfur, since oxygen is more electronegative than sulfur, and because wt FPR binds tolylisourea-MMWLL 50 times better than tolylisothiourea-MMWLL. The lack of preference seen with either D106N or R201A strongly argues for a direct interaction between a Asp-106 -Arg-201 ion pair and the tolylisourea oxygen. Interestingly, the 50-fold preference for oxygen over sulfur exhibited by wt FPR for the NH 2 -terminal urea group is greater than the 3-fold reduction in ligand affinity seen with the thioformamide derivative of fMLF (34). This difference may be the result of orientation factors caused by the tolyl group. FPR has been shown to exhibit a high degree of specificity for the type of isourea on the NH 2 terminus of MLF, with tolylisourea-MLF binding much better than isourea-MLF (9). Replacement of the tolylisourea oxygen of tolylisourea-MMWLL with sulfur may alter the orientation of the tolyl group, since the larger sulfur atom may produce some steric hindrance not found with the smaller oxygen atom. In addition, it is likely the that tolylisourea group directly interacts with FPR at residues distinct from Arg-201 or Arg-205 since addition of the tolylisourea group at the NH 2 terminus of MMWLL enhanced binding to R205A and R201A to a similar or greater extent to that seen with wt FPR (300-, 170-, and 25-fold, respectively).
Carboxyl Modifications of fMLF Suggest Arg-205 Interacts with the Carboxyl Terminus of fMLF-Since analog analysis on the C5a receptor and the angiotensin I receptor indicate that position 6 of the fifth transmembrane-spanning region is the likely interaction site between the COOH-terminal carboxylate of these peptide ligands and their respective receptors (30,31), we attempted to ascertain whether Arg-205 (V-6) of FPR might interact with the COOH terminus of fMLF. We therefore synthesized several COOH-terminal analogs of fMLF and tested wt FPR, D106N, R201A, and R205A for their ability to bind these analogs. Table II shows the effects of COOH-terminal modifications of fMLF on ligand binding affinity. D106N, R201A, and R205A all exhibited markedly reduced affinity for fMLF, with a much greater reduction than was previously observed with fNle-Leu-Phe-Nle-Tyr-Lys-fluorescein (13). This effect was especially true for R205A, which exhibited only a 12-fold reduction of affinity for fNle-Leu-Phe-Nle-Tyr-Lys-fluorescein but a 500-fold reduction for fMLF. Substitution of the free carboxyl group of fMLF with a methyl ester enhanced its ability to bind to R205A (25-fold) and to R201A (7-fold). Moreover, substitution of a butyl ester for the free carboxyl group of fMLF improved the binding affinity for R205A to similar to that seen with wt FPR while having only a modest effect on either D106N or R201A. These effects suggest replacement of the hydrophilic, charged, and bulky arginine with the smaller hydrophobic alanine improves the binding of the bulkier more hydrophobic carboxyl substituents on fMLF. Thus, it appears that Arg-205 is the likely site of interaction with the carboxyl  group of fMLF bound to FPR. If the carboxyl group of fMLF interacts with Arg-205 (see Fig. 2B for proposed interaction), one would expect that substituting the negatively charged carboxyl group with a positive charge would markedly affect activity. We therefore synthesized the dimethylaminoethylester of fMLF and tested it for activity. The substitution of a positive charge for the carboxyl group abolished activity for both wt FPR and R205A. This result was expected for wt FPR, as juxtaposition of two positive charges near one another would be expected to destabilizing. The fact that R205A is unable to bind the positively charged ester group, and binds the negatively charged carboxylate group very poorly, probably indicates that the COOH terminus of the peptide binds in a relatively hydrophobic environment, which could not stabilize either a positive or a negative charge. Indeed, the 500-fold reduction in affinity of R205A for fMLF compared with wt FPR is similar to what would be expected if fMLF bound to R205A with its COOH group in its un-ionized form (assuming a pK a of ϳ4.8, 0.0025% would be unionized at pH 7.4). A hydrophobic environment with a low dielectric constant would be expected to enhance the binding energy between the fMLF COOH-terminal carboxylate and Arg-205 and contribute to high affinity binding of a small peptide.
The butyl ester of fMMWLL did not exhibit enhanced binding to R205A, indicating that enhanced binding was observed when the butyl group was added to the third but not the fifth residue.
D106N and R201A Bind a Peptide Derived from HIV (DP178) More Tightly than Does wt FPR-Recently, a peptide derived from the second heptad repeat region from HIV was shown to activate FPR (5). Since this peptide, like MMWLL, is active without having a formyl group on its NH 2 terminus (5, 12), our mutant analysis might suggest a mode of binding for this very different peptide. Therefore, we tested the ability of the NH 2 COOH terminally acetylated and NH 2 -terminally amidated form of this peptide to bind wt FPR, D106N, and R201A. DP178 bound to wt FPR, D106N, and R201A with K d values of 1200, 30, and 600 nM, respectively, indicating that D106N bound DP178 30-fold more tightly than wt FPR. In addition, R201A bound DP178 twice as tightly as wt FPR. This binding behavior is similar to that observed with MMWLL with relative affinity increases of 40-and 3-fold, respectively, for these two mutants. This is much different than that seen with acetyl-MMWLL where wt FPR bound acetyl-MMWLL 40-and 500fold more tightly than did D106N or R201A, respectively. This would indicate that the acetyl group on DP178 is situated away from the Asp-106 -Arg-201 ion pair in wt FPR. DP178, despite its much greater size compared with MMWLL, nonetheless appears to bind to FPR in a manner similar to MMWLL.
Consequences of FPR Mutation to FPR Function-Since our previous studies had indicated that 1 M fMLF was unable to stimulate GTP␥S binding in the mutants D106N, R201A, or R205A, we examined them for their ability to be stimulated by 1 M fMMWLL, which has 80 -500-fold higher affinity for these mutants (see Table I for respective binding constants). The results are shown in Fig. 3. R201A and R205A, unlike with fMLF (13), were stimulated to similar extents as wt FPR by fMMWLL. D106N, however, remained unresponsive. Interest-ingly, we also found that 1 M tertbutoloxycarbonyl-MLF, which is a known antagonist for FPR (8), stimulated R201A to a similar extent as was observed with 1 M fMMWLL (data not shown), again suggesting altered NH 2 -terminal specificity. 1 M tertbutoloxycarbonyl-MLF did not stimulate GTP␥S binding in CHO cells expressing wt FPR (data not shown).
Other studies have indicated that FPR exhibits a high level of constitutive activity that can be inhibited by Na ϩ and to a lesser degree by K ϩ (26 -27). It is believed that Na ϩ reduces FPR's affinity for fMLF by reducing the number of high affinity binding sites. This effect is cation-specific with NaCl Ͼ LiCl Ͼ KCl Ͼ choline chloride (27), suggesting that the ion pairing implied by our experiments might play a role in this inhibition. Therefore, we tested GTP␥S binding in CHO cells expressing wt FPR, D106N, R201A, and R205A, and compared their activity to that observed in CHO cells that did not express FPR. The results are shown in Fig. 4. wt FPR and the three mutants all exhibited high levels of constitutive activity, 5-8-fold greater GTP␥S binding than that seen with non-expressing CHO cell membranes. The constitutive activity of wt FPR, R201A, and R205A were readily inhibited by physiologic levels of Na ϩ (K1 ⁄2 of 50, 120, and 110 mM, respectively;), whereas the K1 ⁄2 for D106N was Ͼ1000 mM. In addition, wt FPR exhibited cation selectivity of Na ϩ Ͼ Li ϩ Ͼ K ϩ (80%, 63%, and 48% inhibition at 100 mM, respectively), whereas all three cations inhibited D106N similarly (ϳ50% inhibition at 100 mM for all three cations.) This result indicates that Asp-106, in addition to being important in NH 2 -terminal selectivity, may also be important for Na ϩ binding. Na ϩ competition with Arg-201 for interaction with Asp-106 might explain the Na ϩ regulation of ligand affinity observed by Giercheck et al. (27). However, we were unable to detect any Na ϩ -dependent changes in fNle-Leu-Phe-Nle-Tyr-Lys-fluorescein binding to CHO cells expressing wt FPR. No difference in binding was observed in the absence of Na ϩ , as compared with that seen in the presence of 100 mM Na ϩ , nor was there any difference in binding between 100 mM Na ϩ , 100 mM Li ϩ , K ϩ , or choline as reported previously by  Giercheck et al. (27). This result may reflect a difference between cation regulation of binding to intact cells versus that seen using membrane preparations (27). We also analyzed D106N for its ability to mediate chemotaxis toward either fMLF or MMWLL. Fig. 5 shows that D106N exhibited no chemotaxis toward fMLF, whereas MMWLL stimulated chemotaxis ϳ5-fold at 100 nM. D106N also exhibited a high degree of "random migration" in the absence of ligand. D106N exhibited a 4-fold increase (p ϭ 0.001) in random migration compared with wt FPR, again indicative of high constitutive activity.
Conclusions-Is FPR binding of f MLF similar to retinal binding in rhodopsin? We previously developed a model for fMLF binding to FPR (28) based on: 1) the structural similarities between fMLF and all-trans-retinal and the fact that mutation of Phe-291 (VII-11), which is analogous to the lysine that forms the Shiff base with retinal in rhodopsin, markedly affected binding (13); 2) the fact that several additional residues that affect fMLF binding are also known to be important in retinal binding (13); and 3) a photoaffinity label of fMLF with a photoaffinity label in the leucine position cross-linked to helix II near the extracellular space (14).
In rhodopsin, an ion pair exists between Glu-113 (III-3) and the Schiff base of retinal with Lys-293 (VII-11), and recent evidence indicates that this ion pair is indirect, i.e. mediated via a water molecule bound to the retinal Schiff base (35)(36)(37)(38). In addition, Glu-122 (III-12) of rhodopsin interacts with His-211 (V-10) (30,38). FPR Asp-106 at position eight of helix III is midway between these two ion-pairing residues, but one would expect FPR Asp-106 to be more closely aligned with rhodopsin Glu-122 (⌬ is 4 residues for Glu-122 or 400°rotation versus 5 residues or 500°for Glu-113). This position is consistent with our data above, which suggest that Asp-106 ion-pairs with Arg-201 of helix V. In the recent crystal structure of rhodopsin (38), Thr-118 (III-8) which is analogous to Asp-106 of FPR, and Met-207, which is analogous to Arg-205, both contact the rhodopsin ligand, retinal. Thr-118 is also in close proximity to helix V and is located within 6 Å of Phe-203 (determined by our distance measurements of the rhodopsin PDB file). Phe-203 is analogous to Arg-201 of FPR. Thus it is likely, given the overall similarity of G protein-coupled receptors (38), that Asp-106 and Arg-201 of FPR are sufficiently close to ion-pair with one another (the side chain of Asp is ϳ 1 Å longer than Thr and Arg ϳ1.5 Å longer than Phe, so that the Asp-Arg distance would be expected to be ϳ 3.5 Å.) In the crystal structure of rhodopsin, Phe-203 and Met-207, which are analogous to Arg-201 and Arg-205, respectively, are within 4 Å of one another. Our mutagenesis data suggest that Arg-201 and Arg-205 interact with the formyl group and the COOH terminus of fMLF, respec-tively. The expected distance between the formyl group and the carboxyl group is, of course, dependent on the peptide conformation. In a helical conformation, the carboxylate of fMLF would be hydrogen-bonded to the formamide nitrogen and the nitrogen to oxygen distance would be ϳ 3 Å. If fMLF bound in an extended conformation, the NH 2 and COOH termini could be up to 10 Å apart. Other conformations would be intermediate in separation distance. Thus it seems likely that Arg-201 and Arg-205 are positioned sufficiently close to one another to be able to interact with the formyl group and COOH terminus of fMLF, respectively.
In addition, our previous cross-linking data suggest that the leucine side chain of fMLF is probably located close to FPR 93 VRK 95 , which is at the COOH terminus of helix II (14).
Therefore, the most likely positioning of f MLF in the binding pocket of FPR is as follows.
1) The backbone of fMLF is approximately parallel to helix V with the formamide oxygen hydrogen-bonded to one of the guanidine nitrogens of Arg-201. One of the Asp-106 oxygen hydrogen-bonds with the formamide nitrogen, and the other carboxylate oxygen interacts with one of the guanidine nitrogens of Arg-201 (see Fig. 2A).
3) The leucine side chain is positioned toward helix II. This new placement of ligand parallel rather than perpendicular to the average helical axes of FPR does not rule out the possibility of remote global effects of the mutations or, for that matter, the existence of more than one position of the ligand. Clearly, the newly discovered promiscuity of FPR for MMWLL and DP-178 suggests the existence of multiple binding locations, one of which might still be compatible with our original retinal in rhodopsin-like placement of fMLF. Clearly, fMMWLL, fNle-Leu-Phe-Nle-Tyr-Lys-fluorescein, and other chemotactic peptides that are longer than three amino acids would not have a free carboxyl group in position three, and might take on a kinked conformation which might enable it to take advantage of the hydrophobic nature of the space between helices II and VII (30).
This positioning of fMLF in the binding pocket of FPR is similar to the proposed positioning of the tripeptide pyroglutamic acid-histidine-proline-amide (TRH) when bound to the TRH receptor (39) where the NH 2 terminus was proposed to interact with a tyrosine in the analogous position to Asp-106 (Tyr-106) and the COOH terminus was buried within the receptor. The proposed site for binding the COOH terminus of TRH, Arg-306 (VII-7), is different from what we propose for fMLF. Interestingly, the TRH receptor exhibits a marked preference for the COOH-terminal amide (the free acid is a metabolic breakdown product, which exhibits very low activity (Ref. 40)), and this receptor has an aspartic acid  in the analogous position to R205A, our proposed interaction site for the carboxyl terminus of fMLF. The effects of Asp-195 on TRH binding, however, have not been reported. Thus, there may be important motifs in ligand binding of peptides, and knowledge of these motifs may provide important insights to the structure of the ligand binding pockets of G protein coupled receptors.
The FPRL1 receptor is known to bind fMLF (16), albeit weakly, whereas no binding at all is observed with FPRL2 (17). The mouse formyl peptide receptor (Fpr1) also binds fMLF weakly (23). Our studies with Asp-106, Arg-201, and Arg-205 indicate that FPRL1, Fpr-rs1, and Fpr-rs2 might be expected to bind formylated peptides as well. Recently, Klein et al. (41), identified several peptides, including MMWLL, which could activate FPRL1. Formylation of either MMWLL, which exhibited Ͼ200-fold preference for FPR over FPRL1, or SLLWLT-CRPWEAM, which exhibited a Ͼ500-fold preference for FPRL1 over FPR, enhanced the binding to both receptors, although the effect of formylation was much greater for FPR than FPRL1. This would indicate that, although Asp-106 and Arg-201 are necessary for N-formyl specificity, other residues found only in FPR might be involved in maximizing formyl group binding. Studies are presently under way to identify such groups.