HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 280, Issue 24, 22951-22961, June 17, 2005

This Article Abstract Full Text (PDF) All Versions of this Article: 280/24/22951    most recent M502341200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ishino, T. Articles by Chaiken, I. Search for Related Content PubMed PubMed Citation Articles by Ishino, T. Articles by Chaiken, I. Social Bookmarking                      What's this?

Receptor Epitope Usage by an Interleukin-5 Mimetic Peptide^*

Tetsuya Ishino, Cecilia Urbina, Madhushree Bhattacharya, Dominick Panarello, and Irwin Chaiken

From the Department of Biochemistry and Molecular Biology and the A. J. Drexel Institute of Basic and Applied Protein Science, Drexel University College of Medicine, Philadelphia, Pennsylvania 19102

Received for publication, March 2, 2005 , and in revised form, April 4, 2005.

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
The cyclic peptide AF17121 is a library-derived antagonist for human interleukin-5 (IL5) receptor (IL5R) and inhibits IL5 activity. Our previous results have demonstrated that the sixth arginine residue of the peptide is crucial for the inhibitory effect and that several acidic residues in the N- and C-terminal regions also make a contribution, although to a lesser extent (Ruchala, P., Varadi, G., Ishino, T., Scibek, J., Bhattacharya, M., Urbina, C., Van Ryk, D., Uings, I., and Chaiken, I. (2004) Biopolymers 73, 556–568). However, the recognition mechanism of the receptor has remained unresolved. In this study, AF17121 was fused to thioredoxin by recombinant DNA techniques and examined for IL5R interaction using a surface plasmon resonance biosensor method. Kinetic analysis revealed that the dissociation rate of the peptide·receptor complex is comparable with that of the cytokine·receptor complex. The fusion peptide competed with IL5 for both biological function and interaction with IL5R, indicating that the binding sites on the receptor are shared by AF17121 and IL5. To define the epitope residues for AF17121, we defined its binding footprint on IL5R by alanine substitution of Asp⁵⁵, Asp⁵⁶, Glu⁵⁸, Lys¹⁸⁶, Arg¹⁸⁸, and Arg²⁹⁷ of the receptor. Marked effects on the interaction were observed in all three fibronectin type III domains of IL5R, in particular Asp⁵⁵, Arg¹⁸⁸, and Arg²⁹⁷ in the D1, D2, and D3 domains, respectively. This footprint represents a significant subset of that for IL5 binding. The fact that AF17121 mimics the receptor binding capability of IL5 but antagonizes biological function evokes several models for how IL5 induces activation of the multisubunit receptor system.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Asthma is an incurable disease characterized by eosinophil bronchial inflammation and tissue remodeling of the airway wall (1). Human interleukin (IL)¹-5 is a key cytokine that plays a critical role in the differentiation, proliferation, migration, and activation of eosinophils and has been implicated in the pathogenesis of eosinophil-associated allergic inflammation such as asthma (2, 3). Injection of IL5 increases eosinophil numbers in the blood, bone marrow, spleen, and peritoneal cavities (4, 5). This increase can be prevented by the passive administration of anti-IL5 or anti-IL5 receptor antibodies (5, 6). One distinguishing characteristic of IL5 versus other cytokines is that IL5 functions selectively to activate eosinophils (7). These data argue that IL5 plays a central role in the development of allergen-induced eosinophils and suggest that the ability to block IL5 activity evokes the potential for alternative treatment for asthma.

At the molecular level, IL5 leads to biological functions through recruitment of a cell-surface receptor composed of two polypeptide chains, and (8). The chain is IL5-specific and is called IL5 receptor (IL5R), whereas the chain is shared with IL3 and granulocyte/macrophage colony-stimulating factor (GM-CSF) (9–11) and is called the common chain (c). Despite a high degree of amino acid sequence similarity of the chains for IL5, IL3, and GM-CSF, their interaction with cognate cytokine is strictly specific, and no cross-reactivity is found among these cytokines. Previous observations argue that receptor subunit recruitment occurs stepwise, with initial formation of the IL5·IL5R complex required for c binding to induce cytoplasmic signal transduction (12). IL5R alone binds IL5 with an equilibrium dissociation constant of 0.8 nM when expressed in COS cells, and this binding affinity is increased by only 1.5-fold when and chains are coexpressed (13). Hence, IL5R provides most of the binding energy and specificity for IL5, whereas c is primarily responsible for signaling events.

Putting these data together, the IL5-IL5R interaction can be seen as a promising target for designing specific inhibitors aimed at suppressing eosinophil-related inflammation. Extensive mutational analyses have identified recognition epitopes on both IL5 and IL5R, and the interaction can be envisioned by relating structure to function of these epitopes. IL5 is a symmetric homodimer in which each helical bundle domain is composed of three helices (A–C) from one chain and one (D) from the other (14). The binding epitope on IL5 for IL5R is mapped on the structure, showing the importance of charged residues in helix B (His³⁸, Lys³⁹, and His⁴¹), the CD turn (Glu⁸⁸, Glu⁸⁹, Arg⁹⁰, and Arg⁹¹), and helix D (Glu¹¹⁰) for IL5R binding (15–17). For c binding and signal transduction, Glu¹³ in helix A is thought to be a key residue (17). IL5R comprises three fibronectin type III domains (D1, D2, and D3) in the extracellular region, and the membrane-proximal pair of D2 and D3 domains constitutes a cytokine recognition motif that generally can be recognized by a cytokine (18). Recently, we found that the IL5-binding residues are located in the D1 domain (Asp⁵⁵, Asp⁵⁶, and Glu⁵⁸) as well as in the D2 domain (Lys¹⁸⁶ and Arg¹⁸⁸) and the D3 domain (Arg²⁹⁷) of IL5R (19). The homology-deduced IL5R structure indicates that the binding interface of IL5R comprises a cluster of negatively charged residues from the D1 domain and a cluster of positively charged residues from the D2D3 tandem domain. From these observations, we have proposed the hypothesis that a pair of charge complementary interfaces plays an important role in the specific interaction between IL5 and IL5R.

Mimetic peptide methodology has led to the discovery of specific inhibitors for the IL5-IL5R interaction. A potent disulfide-cyclized 18-mer peptide, AF17121, inhibits the binding of IL5 to IL5R and blocks IL5-dependent eosinophil activation (20). Using mutational analysis of AF17121, we have recently found that the sixth arginine residue plays a crucial role in the antagonist effect and that N/C-terminal acidic residues make a finite but lesser contribution (21). Although AF17121 proved to be a specific antagonist of IL5R, showing the importance of its charged clusters for its antagonism, the binding mechanism for IL5R, including the binding epitopes on the receptor, has remained incomplete. The purpose of this study was to elucidate these mechanistic features of AF17121 by kinetic analysis and site-directed mutagenesis with IL5R. To enhance the solubility and binding signal of the peptide in surface plasmon resonance (SPR) assays, we employed a recombinant system to fuse the peptide to a highly soluble protein, thioredoxin. We used thioredoxin-fused AF17121 for kinetic interaction analysis with soluble IL5R (sIL5R; extracellular domain of IL5R) (22) as well as for biological inhibition assays. Alanine substitution of the two half-cystine residues in AF17121 completely impaired its binding to IL5R and its corresponding inhibitory action on cells. Furthermore, we used SPR binding analysis of sIL5R alanine-scanning mutational variants to explore the functional impact of each amino acid on the kinetic interaction of the receptor with AF17121. Despite the small size of the peptide, marked effects on the interaction were observed in the entire IL5R footprint for IL5. The revealed mechanism of cytokine mimicry simplifies the complexity of the IL5-IL5R interaction and evokes several models for how IL5 induces activation of the multisubunit receptor system.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Materials
Human IL5 and human sIL5R proteins were produced and purified as described previously (12). Mouse IL3 protein was purchased from Sigma. Anti-human IL5R monoclonal antibody 16 was a generous gift from Dr. Jan Tavernier (University of Ghent, Ghent, Belgium). All enzymes were purchased from New England Biolabs Inc. (Beverly, MA). The mouse BaF3 cell line was purchased from German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany). All DNA oligonucleotide primers, Drosophila S2 cells, cell culture medium, L-glutamine solution (200 mM), and RPMI 1640 medium were purchased from Invitrogen. For SPR measurements, the CM5 sensor chip, surfactant P20, N-ethyl-N'-(3-dimethylaminopropyl)carbodiimide, N-hydroxysuccinimide, 1 M ethanolamine (pH 8.5), and 10 mM glycine HCl (pH 1.5 and 2.0) were purchased from Biacore (Piscataway, NJ).

Construction of Expression Plasmids for Thioredoxin-fused AF17121
The DNA sequence of the AF17121 peptide (see Fig. 1A) was designed using the Escherichia coli K12 codon usage tabulated from the GenBank^TM Data Bank (CUTG) (23). The oligonucleotide primers used for this study were as follows: 5'-CATGGTTGACGAATGCTGGCGTATCATCGCTTCCCACACCTGGTTCTGCGCTGAAGAATGAA-3' and 5'-AGCTTTCATTCTTCAGCGCAGAACCAGGTGTGGGAAGCGATGATACGCCAGCATTCGTCACC-3'. Annealing was achieved with 10 nmol of sense and antisense oligonucleotides in 100 µl of annealing buffer (10 mM Tris-HCl, 100 mM NaCl, and 1 mM EDTA (pH 7.5)). The reaction mixture was incubated at 95 °C for 5 min and at 65 °C for 10 min and then was allowed to cool slowly at room temperature for 1 h. The double-stranded fragment was ligated to expression vector pET-32b (Novagen) that had been digested with NcoI and HindIII. The resulting plasmid was named pTRXaf17. In this construct, the gene for thioredoxin, a hexahistidine tag, the S·Tag peptide, an enterokinase cleavage site, and AF17121 sequence (from the N to C terminus) were placed behind the T7 promoter. The amino acid sequence of the final product (denoted thioredoxin (trx)-AF17121) is shown in Fig. 1B. For the Cys-to-Ala variant (denoted trx-AF17121(AA)) at positions 4 and 15 (numbering as in the AF17121 peptide), mutations were introduced into pTRXaf17 with a QuikChange site-directed mutagenesis kit (Stratagene) using the synthetic DNA oligonucleotide 5'-GTTGACGAAGCCTGGCGTATCATCGCTTCCCACACCTGGTTCGCCGCTGAAGAA-3' as the forward primer and its reverse complementary sequence as the reverse primer. The presence of the desired mutations was verified by DNA sequencing.

Expression and Purification of Fusion Peptides in E. coli
For production of trx-AF17121, plasmid pTRXaf17 was transformed into the host strain BL21-CodonPlus (DE3)-RIL (Stratagene). The transformed cells were grown in 1 liter of LB medium at 37 °C, and expression was initiated by addition of 1 mM isopropyl -D-thiogalactopyranoside to A_{600 nm} = 0.8. After 3 h, cells were harvested and frozen at –20 °C. The frozen cell paste was suspended in 10 ml of lysis buffer (100 mM Tris-HCl (pH 8.0), 100 mM NaCl, and 1 mM phenylmethanesulfonyl fluoride)/1 g of wet cells, followed by sonication for 20 min on ice. Insoluble material was removed by centrifugation at 20,000 xg for 30 min at 4 °C. The supernatant was then loaded onto a HiTrap chelating HP column (Amersham Biosciences) pre-equilibrated with lysis buffer. Any nonspecific binding proteins were washed away with buffer A (50 mM Tris-HCl (pH 7.4), 25 mM imidazole, and 300 mM NaCl), and the fusion peptide was eluted with a linear gradient from buffer A to buffer B (50 mM Tris-HCl (pH 7.4) and 300 mM imidazole). The protein fraction was diluted five times in redox buffer (1 mM reduced glutathione, 10 mM oxidized glutathione, and 100 mM Tris-HCl (pH 8.5)) and stirred to allow disulfide bond formation (24 h, 4 °C). The protein solution was dialyzed against buffer C (20 mM Tris-HCl (pH 8.0)) and then loaded onto a HiTrap Q HP column (Amersham Biosciences) pre-equilibrated with buffer C. The protein was eluted with a linear gradient of 0–1 M NaCl in buffer C. The fused protein was located in the 0.1–0.2 M NaCl fraction. The buffer of the purified protein fraction was exchanged with phosphate-buffered saline (1 mM KH₂PO₄, 10 mM Na₂HPO₄, 137 mM NaCl, and 2.7 mM KCl (pH 7.4)) using a HiPrep 26/10 desalting column (Amersham Biosciences), and the resulting protein solution was stored below –80 °C. The purified protein was analyzed by SDS-PAGE (4–20% linear gradient gel; Bio-Rad). The Cys mutant trx-AF17121(AA) and thioredoxin protein were produced and purified in the same manner as described for trx-AF17121, except for the oxidation process. The original expression vector (pET-32b) was used for production of thioredoxin.

Characterization of Fusion Peptides
The purity of the fusion peptides was evaluated by analytical reverse phase HPLC using a 4.6 x 250-mm 218 TP C18 silica gel-packed column (Vydac) and a linear gradient of 5–95% acetonitrile in 0.1% trifluoroacetic acid/H₂O solution over 50 min (flow rate of 1 ml/min). The integrity of the fusion peptides was certified by MALDI mass spectrometry (Wistar Institute, Philadelphia, PA). Protein quantitation was achieved by measuring the absorbance at 280 nm and calculating the concentration using the molar absorption coefficient (₂₈₀ = 25,230 M^–1 cm^–1) according to Pace et al. (24).

The amount of free cysteine residues in trx-AF17121 and trx-AF17121(AA) was measured using 5,5'-dithiobis(2-nitrobenzoic acid) (Ellman's reagent) (25). Protein (0.3 µM) and 5,5'-dithiobis(2-nitrobenzoic acid) (1 mM) were incubated for 20 min at room temperature in 100 mM sodium phosphate buffer (pH 7.3) containing 1 mM EDTA. The absorbance at 412 nm was measured, and the moles of sulfhydryl group were determined using the extinction coefficient of oxidized 5,5'-dithiobis(2-nitrobenzoic acid) (₄₁₂ = 13,600 M^–1 cm^–1). The reactivity of 5,5'-dithiobis(2-nitrobenzoic acid) with the buffer was subtracted from the results obtained. The number of sulfhydryl groups/molecule was then corrected using a standard curve obtained with reduced glutathione.

The apparent molecular masses and oligomeric states of trx-AF17121 proteins were evaluated using an HPLC system with a 7.8 x 300-mm BioSep-SEC-S-2000 size exclusion analytical column (Phenomenex) pre-equilibrated with phosphate-buffered saline. The column was calibrated with standards of known molecular mass (Amersham Biosciences). Linear regression of log₁₀(molecular mass) versus elution volume of the standard was used to extrapolate the apparent molecular masses of the fusion peptides.

S2 Cell Expression of IL5R and Its Mutational Variants
sIL5R and its mutational variants (D55A, D56A, E58A, K186A, R188A, and R297A) were transiently expressed in Drosophila S2 cells and secreted into serum-free medium as described previously (19). Cell-free supernatant was collected after 2 days and stored at –20 °C for the following binding analysis.

SPR Biosensor Binding Analysis
The kinetic interaction assay was carried out using an SPR biosensor (Biacore 3000, Biacore AB, Uppsala, Sweden). All experiments were conducted at 25 °C in phosphate-buffered saline containing 0.005% surfactant P20.

Noncovalent Immobilization of IL5R and Binding Assay of the Fusion Peptides—The conformation-sensitive anti-IL5R monoclonal antibody 16 (26) was covalently immobilized on a CM5 sensor chip, and sIL5R was captured by the antibody as described previously (19). Various concentrations of trx-AF17121 or trx-AF17121(AA) were passed over the antibody-captured sIL5R surface. To examine the interaction of sIL5R mutational variants with trx-AF17121, each receptor mutant was captured to a level of almost 200 resonance units (RU).

Covalent Immobilization of Fusion Peptides and Binding Assay of IL5R—Immobilization of trx-AF17121 and trx-AF17121(AA) on a CM5 sensor chip was carried out by the amine coupling method as recommended by Biacore. Briefly, 10 µM protein solution was diluted 50 times in 10 mM acetate (pH 4.0) and injected onto a biosensor surface that had been pre-activated by a 20-µl injection of a 1:1 mixture of 200 mM N-ethyl-N'-(3-dimethylaminopropyl)carbodiimide and 50 mM N-hydroxysuccinimide, followed by an injection of 1 M ethanolamine HCl (pH 8.5). The levels of immobilization were 200 RU. The real-time interaction was measured by injecting purified sIL5R onto these surfaces. The thioredoxin-coupled surface was used as a reference surface to correct for instrument and buffer artifacts. To regenerate chip surfaces, bound proteins were removed from the surfaces by an injection of 10 mM glycine HCl (pH 2.0) after each cycle. All procedures were automated to create repetitive cycles of injection of various concentrations of sIL5R (flow rate of 50 µl/min) and the regeneration buffer (flow rate of 100 µl/min).

Data Analysis—Nonlinear least-squares analysis was used to calculate the association and dissociation rate constants (k_on and k_off, respectively). Prior to the calculation, the binding data were corrected for nonspecific interaction by subtracting the reference surface data from the reaction surface data and were further corrected for buffer effect by subtracting the signal due to buffer injections from those due to protein sample injections (double referencing) (27). The interaction curves thus obtained were globally fit using a model for 1:1 Langmuir binding (BIAevaluation software, Biacore). Individual kinetic parameters were obtained from at least three separate experiments. The equilibrium dissociation constant (K_d) was calculated as K_d = k_off /k_on.

Control Experiments for the Mass Transfer Effect on the Biosensor—To assess the effect of the amount of immobilization on the apparent reaction rate constants, different injection times were used to vary the level of immobilization of sIL5R. Various amounts of sIL5R were captured by antibody 16 on the surface, and 40 nM trx-AF17121 solutions were passed over each amount of immobilized sIL5R. The binding data were collected at a flow rate of 10 µl/min. The effect of the flow rate on the apparent reaction rate constants was also assessed. The amount of sIL5R captured by antibody 16 on the sensor chip was fixed at 130 RU, and 40 nM trx-AF17121 solutions were passed over immobilized sIL5R. The binding data were collected at flow rates of 10, 20, 40, and 80 µl/min.

SPR Competition Assay
The effectiveness of trx-AF17121 and trx-AF17121(AA) in inhibiting the IL5-IL5R interaction was evaluated by two types of competition experiments. The first set of experiments measured the ability of the fusion peptides to compete with IL5 binding to sIL5R as described previously (21). For this competition assay, 4 nM sIL5R in the absence and presence of different concentrations of trx-AF17121 or trx-AF17121(AA) was passed over the IL5-captured surface (200 RU). The second set of experiments measured the ability of IL5 to compete with trx-AF17121 binding to sIL5R. In this experiment, trx-AF17121 was directly immobilized on a CM5 sensor surface as described above. sIL5R (80 nM) was passed over this surface in the absence and presence of different concentrations of IL5 or trx-AF17121. The concentration of sIL5R in each experiment was set close to the K_d value for the interaction of IL5R with IL5 or AF17121 (see below).

The maximum binding response (R_max) of the IL5-IL5R interaction under the influence of trx-AF17121 was calculated by globally fitting the association and dissociation phases of sensorgrams. Fits were to a 1:1 Langmuir model. Data analysis was conducted using the BIAevaluation program and Microsoft Excel. Inhibitory effects of the peptides resulting in decreases in R_max values are expressed as RU_eq; maximum inhibitory effects are expressed as RU_max; and peptide concentration is expressed as C.IC₅₀ values were calculated from the following equation: IC₅₀ = (C(RU_max – RU_eq))/RU_eq. Upon rearrangement, RU_eq/C = RU_max/IC₅₀ – RU_eq/IC₅₀. A plot of RU_eq/C against RU_eq at different peptide concentrations thus gave a straight line from which the IC₅₀ value was obtained.

Biological Assay
The assay was carried out using the human IL5-dependent erythroleukemia cell line TF1.28 and the mouse IL3-dependent pro-B lymphocyte cell line BaF3. Both TF1.28 and BaF3 cells were cultured in RPMI 1640 medium and 10% fetal calf serum (Hyclone Laboratories) supplemented with 10 nM human IL5 or 10 nM mouse IL3, respectively. Cell proliferation was measured using the tetrazolium salt WST-1 (4-(3-(4-iodophenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio)-1,3-benzene disulfonate; Roche Diagnostics). 96-Well plates were seeded at 5000 cells/well and incubated for 48 h at 37 °C in 5% CO₂ in the presence of 10 pM IL5 for the TF1.28 cells or 2.5 pM IL3 for the BaF3 cells with different concentrations of trx-AF17121 or trx-AF17121(AA). Following the 48-h incubation, 10 µl of soluble WST-1 was added to each well and incubated for 4 h at 37 °C. The plate was then read using a microplate reader at an absorbance of 450 nm with a 650-nm reference. Individual values were obtained from three experiments, and the average values were analyzed using Prism software (GraphPad Software).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Design, Production, and Characterization of Thioredoxinfused AF17121—AF17121 showed inhibitory effects on IL5 function and proved to be a promising lead for rational drug design as reported previously (20, 21). To elucidate the detailed mechanism of these effects, we first carried out an interaction assay using the AF17121 synthetic peptide by SPR-based technology. Because the binding signal in SPR biosensor analysis is proportional to the molecular mass of the interacting species, it is often difficult to detect signals from low molecular mass peptides. Enhancing the binding signal can be achieved by increasing the amounts of immobilized protein. However, large amounts of immobilization often cause non-ideal effects such as mass transfer, rebinding, crowding, and steric hindrance, which may result in deviations from ideal binding data. In the case of the AF17121 synthetic peptide, we could not observe its binding signal on a surface immobilized with a relevant amount of sIL5R (data not shown). To overcome the signal sensitivity limitation and to eliminate these artifacts in SPR-based biosensors, we used a recombinant system to fuse the peptide (2 kDa) to thioredoxin. Because the final product has a molecular mass of 19 kDa, 10-fold enhancement of the binding signal can be expected. The reason we chose thioredoxin as a fusion partner is that thioredoxin is known to be overexpressed in bacteria without formation of inclusion bodies (28). Thioredoxin is also advantageous for biophysical experiments because the protein is stable to heat and highly soluble.

We designed and produced a fusion peptide that consists of thioredoxin, hexahistidine, the S·Tag peptide, and AF17121 sequence from the N to C terminus (Fig. 1B). To investigate the importance of the disulfide-bonded cyclic structure in AF17121, we also designed mutant trx-AF17121(AA), in which Cys⁴ and Cys¹⁵ were replaced with alanines by site-directed mutagenesis. The recombinant protein was expressed in high yield, and the content of the protein was estimated to be half of the total cell protein judging from SDS-PAGE (Fig. 1C). After sonication and ultracentrifugation, the fusion protein was present not in inclusion bodies but rather in the soluble fraction. After purification by standard nickel chelating affinity column chromatography, free cysteine oxidation and disulfide bond formation were carried out using a redox equilibrium system. Oxidized trx-AF17121 contained no detectable free sulfhydryl group as judged using Ellman's reagent, confirming complete disulfide bond formation of the AF17121 domain (Table I). Even though we used a low concentration of protein solution during the oxidation process, bands corresponding to intermolecular disulfide bonded forms were detected by nonreducing SDS-PAGE. The reaction mixture was further purified by anion exchange column chromatography, yielding 5 mg of purified monomeric trx-AF17121 from 1 liter of culture. The purity of the protein was assessed by SDS-PAGE, revealing a single band between 20.7 and 28.8 kDa (Fig. 1C). The same molecular mass was observed upon nonreducing SDS-PAGE, demonstrating that there was no detectable intermolecular disulfide bond formation (data not shown). The final products were characterized by analytical HPLC for purity (usually >95%) and by MALDI mass spectrometry for molecular mass as shown in Table I.

View larger version (41K): [in this window] [in a new window]   FIG. 1. Recombinant construction of thioredoxin-fused AF17121. A, design of the sense DNA oligomer of the AF17121 peptide for the thioredoxin fusion. B, amino acid sequence of trx-AF17121. A His tag was used for purification and detection. The S·Tag and an enterokinase (Ek) cleavage site were part of the commercially obtained vector and were not utilized in this study. C, Coomassie Blue-stained SDS-polyacrylamide gel analysis of the process of trx-AF17121 purification. Lane 1, molecular mass markers (from top to bottom; 103, 77, 50, 34.3, 28.8, and 20.7 kDa); lane 2, whole cell; lane 3, supernatant after sonication; lane 4, fraction after nickel-nitrilotriacetic acid column chromatography; lane 5, fraction after oxidization and anion exchange column chromatography.

Kinetic Interaction of the Fusion Peptides with IL5R—It has been shown previously that the AF17121 synthetic peptide inhibits the IL5-IL5R interaction by binding to IL5R (20). Here, to validate the direct binding of AF17121 to IL5R and the binding stoichiometry, we employed the sandwich SPR-based binding assay, which we have established for cytokine-receptor interaction (15, 19). Fig. 2A shows whole real-time sensorgrams of sIL5R capture and the subsequent complex formation of IL5 or trx-AF17121 on the biosensor surface. Under saturating concentrations, bound trx-AF17121 produced a signal with nearly half the intensity of bound IL5. In contrast, no direct interaction between trx-AF17121 and IL5 was observed (data not shown). Because the biosensor signal is proportional to the mass, it was possible to estimate the binding stoichiometry of trx-AF17121 based on its mass difference with IL5. When 230 RU of sIL5R was captured on the surface (Fig. 2A), the R_max value for IL5 (glycosylated form, 34 kDa) was 105 RU and that for trx-AF17121 (19 kDa) was 49 RU. Using IL5 as a reference, the calculated molar ratio of IL5 and trx-AF17121 is 1 versus 0.86. Because IL5 has been demonstrated to interact with IL5R at a 1:1 stoichiometry (29), this finding demonstrates that a 1:1 complex is formed between AF17121 and IL5R.

Prior to determination of the kinetic constants, it is important to determine the optimum experimental conditions to eliminate artifacts in SPR biosensor assays. In particular, the mass transfer effect can be a major artifact when the association rate is fast or when a large amount of immobilization is required (30). Indeed, the binding of IL5 to sIL5R was compromised by the mass transfer effect even when using as low a level of immobilization as possible (19). The presence of the mass transfer effect is easily detected, as the binding signal will be dependent on the level of immobilization. Without the mass transfer effect, binding sensorgrams generated from various amounts of immobilized protein should be able to be superimposed by normalizing the R_max value. To assess the effect of the immobilization level on the kinetics, 40 nM trx-AF17121 solutions were injected onto various amounts of immobilized sIL5R. No significant difference in the binding signals was observed at varying levels of immobilization (Fig. 2B). We further examined the presence of the mass transfer effect by measuring the association rates at flow rates of 10–80 µl/min. Fig. 2C shows a slight increase in the association rate with increasing flow rate, indicative of the mass transfer effect at slower flow rates. Because the flow rates of 40 and 80 µl/min did not alter the measured association rates, further kinetic experiments were performed at a flow rate of 50 µl/min.

View larger version (16K): [in this window] [in a new window]   FIG. 2. Control AF17121 binding experiments for SPR biosensor assays. A, mass ratio of IL5 and trx-AF17121 bound to sIL5R on an SPR biosensor surface. The overlay of real-time sensorgrams shows sequential injections of the following solutions over the antibody 16-coupled surface: 50 nM sIL5R; running buffer (phosphate-buffered saline); 200 nM IL5 (blue line), 800 nM trx-AF17121 (red line), and running buffer (green line); running buffer; and regeneration buffer. Sharp spikes around 800 s were attributed to the very large difference in the bulk phase response signal between running buffer and regeneration buffer. B, effect of the amount of captured sIL5R on the kinetic interaction between sIL5R and trx-AF17121. The binding sensorgrams were measured with an sIL5R-captured surface of 410, 320, 220, or 130 RU (green, magenta, blue, or red line, respectively). For comparison, sensorgrams were normalized according to the Rmax values. C, effect of the flow rate on the association rate. The association rates were measured at flow rates of 10, 20, 40, and 80 ml/min (green, magenta, blue, and red lines, respectively).

View this table: [in this window] [in a new window]   TABLE II Global fitting analysis of the kinetic interaction of sIL5R and trx-AF17121 or IL5

Competitive Inhibition of the Fusion Peptide—Two types of SPR-based experiments were performed to examine inhibition of the IL5-IL5R interaction by AF17121. In the first experiment, IL5 was captured by anti-IL5 monoclonal antibody 4A6 immobilized on the biosensor chip (Fig. 4A). Soluble IL5R was preincubated for 1 h with different concentrations of the peptides and then injected onto the IL5-captured surface. trx-AF17121 showed inhibition of IL5R binding to IL5 in a dose-dependent manner (Fig. 4B), although inhibition was incomplete, as previously found in the competition assay for the AF17121 synthetic peptide (21). As expected, trx-AF17121(AA) did not inhibit the IL5-IL5R interaction at trx-AF17121(AA) concentrations up to 12 µM (Fig. 4C). This finding demonstrates the considerable importance of the disulfide bond bridge of AF17121 for its inhibitory activity. The inhibitory effect of trx-AF17121 was quantitated from the decrease in the R_max value depending on the peptide concentrations. A plot of RU_eq/C against RU_eq at different peptide concentrations gave an IC₅₀ value of 18 nM as the inverse of the slope (see "Experimental Procedures").

In the second experiment, sIL5R was preincubated for 1 h with different concentrations of IL5 and then injected onto the trx-AF17121-coupled surface (Fig. 4D). IL5 showed concentration-dependent inhibition of the peptide-receptor interaction and achieved complete inhibition at a concentration of 128 nM or higher (Fig. 4E). The IC₅₀ value for IL5 versus immobilized trx-AF17121 was 130 nM. To further exclude the possibility that the incompleteness of inhibition observed in the first experiment is due to aggregation of the peptide, we tested the autoinhibition of trx-AF17121 for receptor binding. Mixtures of sIL5R and different concentrations of trx-AF17121 were injected onto the trx-AF17121-coupled surface (Fig. 4F). Soluble trx-AF17121 was found to compete with immobilized trx-AF17121 binding to sIL5R, making it unlikely that a trivial reason such as trx-AF17121 aggregation could explain the incomplete inhibition shown in Fig. 4B.

View larger version (34K): [in this window] [in a new window]   FIG. 3. Kinetic analysis of trx-AF17121 and trx-AF17121(AA). A, the schematic diagram shows the results from a kinetic assay in which trx-AF17121 was injected onto the sIL5R-captured surface. B, the overlay of real-time sensorgrams shows sequential injections of 10, 20, 40, and 80 nM trx-AF17121 (yellow to orange lines) at 0 s, followed by injection of running buffer alone at 120 s. The rate constants were calculated by globally fitting the association phase (0–120 s) and the dissociation phase (120–360 s) to a model for 1:1 Langmuir binding. Black lines show calculated sensorgrams. C, the real-time sensorgram shows a single injection of 8 µM trx-AF17121(AA) (red line) over onto the sIL5R-captured surface. D, the schematic diagram shows the results from a kinetic assay in which sIL5R was injected onto the trx-AF17121-coupled surface. E, the overlay of real-time sensorgrams shows sequential injections of 10, 20, 40, and 80 nM sIL5R (yellow to orange lines) at 0 s, followed by injection of running buffer alone at 120 s. F, the real-time sensorgram shows a single injection of 1 µM sIL5R (red line) onto the trx-AF17121(AA)-coupled surface.

View larger version (40K): [in this window] [in a new window]   FIG. 4. Biosensor inhibition assays for the IL5-IL5R interaction. A, the schematic diagram shows the results of an inhibition assay for the IL5-IL5R interaction. B, various concentrations of trx-AF17121 (2, 8, 31, 125, 500, and 2000 nM; dark-blue to light-blue lines) were incubated with 4 nM IL5R and passed over the IL5-captured surface. Sensorgrams in the absence of the peptides were measured both at the beginning (dark-blue line) and the end (purple line) of each experiment. C, various concentrations of trx-AF17121(AA) (12, 47, 188, 750, 3000, and 12,000 nM; dark-blue to light-blue lines) were incubated with 4 nM IL5R and passed over the IL5-captured surface. D, the schematic diagram shows the results of an inhibition assay for the AF17121-IL5R interaction. E, various concentrations of IL5 (2, 8, 31, 125, 500, and 2000 nM; dark-blue to light-blue lines) were incubated with 80 nM IL5R and passed over the trx-AF17121-coupled surface. F, various concentrations of trx-AF17121 (2, 8, 31, 125, 500, and 2000 nM; dark-blue to light-blue lines) were incubated with 80 nM IL5R and passed over the trx-AF17121-coupled surface.

View larger version (19K): [in this window] [in a new window]   FIG. 5. Biological inhibition assays for proliferation of IL5- and IL3-dependent cells. A, proliferation of TF1.28 cells stimulated by 10 pM human IL5 in the presence of varying concentrations of peptides. The inhibitory activity of the peptides on TF1.28 cells was examined by incubating the cells with various concentrations (0.005, 0.020, 0.080, 0.320, 1, 5, 20, 81, 328, and 1300 nM) of trx-AF17121 () or trx-AF17121(AA) (). Activation by various concentrations (0.2 fM to 200 nM) of IL5 () is also shown as the standard. B, proliferation of BaF3 cells stimulated by 2.5 pM mouse IL3 in the presence of varying concentrations of peptides. The inhibitory activity of the peptides on BaF3 cells was examined by inhibition with various concentrations (0.005, 0.020, 0.080, 0.320, 1, 5, 20, 81, 328, and 1300 nM) of trx-AF17121 () or thioredoxin (). The standard curve was measured using 0.5 fM to 5 nM IL3 ().

Analysis of the Binding Epitope on IL5R—To elucidate the molecular recognition mechanism of AF17121, we performed kinetic interaction analysis of mutational variants of human IL5R. We previously reported that the key charged residues on IL5R, viz. Asp⁵⁵, Asp⁵⁶, Glu⁵⁸, Lys¹⁸⁶, Arg¹⁸⁸, and Arg²⁹⁷, are involved in high affinity binding to IL5 (19). Because trx-AF17121 competed with IL5 binding to the receptor, as found in the previous experiment, we assumed that the binding sites for the receptor could be partially or fully overlapped between IL5 and AF17121. In this study, we therefore chose these key residues for alanine site-directed mutagenesis to define the binding epitope for AF17121. We used the sandwich SPR biosensor method, in which each mutational variant of sIL5R is captured by antibody 16, providing an oriented receptor surface to be challenged with trx-AF17121 for the quantitative binding analyses (Fig. 3A). Mutants D55A, R188A, and R297A showed impaired binding to trx-AF17121 (Fig. 6). It is noteworthy that the sensitive residues are in all the fibronectin type III domains of IL5R, viz. in the D1 (D55A), D2 (R188A), and D3 (R297A) domains. In contrast, mutants D56A, E58A, and K186A showed a binding affinity similar to that of the wild-type receptor. For comparison, the relative values for k_on, k_off, and K_d are listed in Table III. We found that mutations such as D56A and E58A caused a slight increase in the binding affinity (a <2-fold decrease in the K_d value). As shown in Table III, these decreased K_d values were attributed to increases in k_on and decreases in k_off. Of note, Asp⁵⁶ and Glu⁵⁸ are negatively charged residues located in the D1 domain of IL5R. Because AF17121 is a highly acidic peptide (calculated pI of 4.6), this fractionally increased binding affinity may reflect the reduction of electrostatic repulsion between these negatively charged residues of IL5R and some negatively charged residues of the AF17121 peptide.

View this table: [in this window] [in a new window]   TABLE III Relative rate constants and dissociation constants for the interaction between trx-AF17121 and sIL5R mutants (given as ratio versus IL5R)

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Using the recombinant trx-AF17121 fusion enabled facile binding analysis of both surface-captured as well as sIL5R using an SPR-based technique (Fig. 3). By measuring association and dissociation rates, kinetic analysis can provide a deeper understanding of interaction dynamics. The receptor interaction of trx-AF17121 showed a 30-fold slower association rate and a 2-fold slower dissociation rate compared with those of IL5, leading to 15-fold weaker binding affinity (Table II). The association rate can be affected by factors such as electrostatic steering forces and conformational rearrangements for the final rigid complex. The latter could explain the slow association rate of the peptide because the effective conformational rearrangement could play a key role in the formation of the IL5·IL5R complex (19, 29). The measured dissociation rate of the peptide·receptor complex was comparable with that of the cytokine·receptor complex. AF17121 could achieve high affinity binding due to decreased dissociation. This is more likely because the peptide was affinity-matured by several rounds of selection (20). Because the slow dissociation rate is ascribed to the stability of the final complex, this observation indicates that AF17121 can employ similar interaction mechanisms compared with IL5. Receptor binding was completely eliminated by linearizing the cyclic conformation of the peptide (Fig. 3, C and F), suggesting that the full antagonism of AF17121 requires not only the binding residues (the sixth arginine residue and a cluster of N/C-terminal acidic residues), but also the proper spatial alignment of these residues. Although we cannot exclude the possibility that the disulfide bond bridge of AF17121 could be involved in the direct interaction with the receptor, the findings of the importance of the charge complementary interaction (see below) support the idea that the disulfide bond is critical for stabilizing a functional conformation rather than for direct interaction.

View larger version (43K): [in this window] [in a new window]   FIG. 6. Binding epitope analysis of trx-AF17121 and IL5R mutational variants. Six IL5R mutants (D55A, D56A, E58A, K186A, R188A, and R297A) were captured on the surface via antibody 16, and then various concentrations of trx-AF17121 were injected onto the surface. For D56A, E58A, and K186A, sequential injections of 10, 20, 40, and 80 nM trx-AF17121 (yellow to orange lines) are shown. Black lines show calculated sensorgrams. For D55A, R188A, and R297A, sensorgrams of a single injection of 8 µM trx-AF17121 (red lines) are shown.

View larger version (28K): [in this window] [in a new window]   FIG. 7. Maps of the binding epitopes for IL5 and AF17121 on the homology-modeled structure of IL5R. In this model, three fibronectin type III domains (D1, D2, and D3) are denoted from the N to C terminus (19). Epitope residues for IL5 binding, which have been found previously (19), are shown as Corey-Pauling-Koltun models. Residues that impaired the binding affinity of trx-AF17121 by alanine substitution are underlined. Oxygen (negatively charged) and nitrogen (positively charged) atoms are colored as red and blue, respectively. Loop regions are represented as coils and strands as ribbons. The molecular graphic figure was prepared with the program MOLMOL (38).

View larger version (68K): [in this window] [in a new window]   FIG. 8. Multiple sequence alignment of receptor for the IL5/IL3/GM-CSF family. Multiple sequence alignments were performed using ClustalW Version 1.60 (39) and adjusted further manually. Fully conserved amino acid residues (*), partially conserved amino acid residues (.), and conserved similar amino acid groups (:) are indicated. Groups of similar amino acid residues were categorized according to ClustalW Version 1.60. The strand regions estimated from the rat prolactin receptor as determined by its crystal structure (40) are underlined. Boxed sequences encompass residues involved in the IL5 interaction (19).

To investigate this specificity in more depth, we performed multiple alignments of receptor for IL5, IL3, and GM-CSF, which share a similarity in amino acid sequence (Fig. 8). Looking at this alignment, we noticed several features regarding variability versus commonality. First, the D1 domain is less conserved overall than the D2 and D3 domains in these receptors. The sequence alignment shows 21% similarity (4% identity) in the D1 domain, 44% similarity (24% identity) in the D2 domain, and 40% similarity (16% identity) in the D3 domain. More important, the positions of key IL5-binding residues in site I, all of which are in the D1 domain (Fig. 7), are hypervariable among these three receptors (Fig. 8). These observations suggest that site I is a primary determinant for ligand specificity. In contrast, site II shows a combination of variability (in the D2 domain) and commonality (in the D3 domain). Although specificity may also be driven through this site, one of the key IL5-binding residues in the D3 domain (Arg²⁹⁷) is conserved in all three receptor chains. This finding argues that the conserved arginine residue could be important for c recruitment because the receptor chains for IL3, IL5, and GM-CSF share the signal-transducing receptor. This speculation is favored by modeled structures of the three-component complexes of IL3·IL3 receptor ·c and GM-CSF·GM-CSF receptor ·c, in which the conserved arginine residues of receptor for IL3 and GM-CSF (Arg²⁵⁹ and Arg²⁸⁰, respectively) make close contacts with c (34).

Our finding that AF17121 recognizes both sites I and II suggests that both binding sites can represent major targets for structure-based design of IL5R antagonists. We surmise that targeting site I will be important to derive a specific antagonist for IL5R, whereas blocking site II offers the opportunity to interfere with IL5 biological activity induced by multisubunit receptor recruitment involving c (see below). This can be seen in the IL5 ligand side, as we previously found that the roles of the receptor-binding epitope on IL5 can be separated into receptor recognition residues such as Arg⁹¹ and receptor activation residues such as Glu¹¹⁰ (35, 36). In this regard, considering their charge complementarity, it would be reasonable to speculate that Arg⁹¹ and Glu¹¹⁰ of IL5 are more likely to be candidates for binding partners of sites I and II on IL5R, respectively. Correspondingly, Arg⁶ and at least a subset of the N/C-terminal acidic cluster of residues of AF17121 might mimic these types of residues in IL5 and therein recognize sites I and II, respectively, on the receptor. The distance between Arg⁹¹ and Glu¹¹⁰ in the crystal structure of IL5 is 24 Å, which appears to be far greater than the distance between Arg⁶ and the N/C-terminal acidic cluster in AF17121. Indeed, our previous modeling analysis of AF17121 has yielded a possible distance of 14 Å (21). This may argue the possibility that the hinge region between the D1 domain and the D2D3 tandem domain of IL5R might be flexible enough to allow a conformational rearrangement of the interdomain packing that brings sites I and II closer for AF17121 binding than for IL5 binding. This explanation seems to be in good agreement with the previous data from thermodynamic and kinetic analyses that considerable conformational rearrangement could take place in IL5R upon IL5 binding (19, 29).

Although AF17121, mimicking IL5, blocks not only the specific epitope (site I) but also the activating epitope (site II) on IL5R, no agonist activity for IL5R was observed (Fig. 5), as reported previously (20). Activation of the IL5 receptor depends on specific interaction of IL5 with IL5R, formation of oligomeric receptor complexes with c, disulfide bond formation between IL5R and c, and initiation of cytoplasmic phosphorylation events (37). The lack of agonist activity of AF17121 can be surmised to originate from the inability of the AF17121·IL5R complex to form an activated complex with c. This may be simply because AF17121 lacks a critical binding epitope for c, such as Glu¹³ in IL5 (17), so as to be incapable of full receptor oligomerization with c. Another possibility is that AF17121-IL5R complex formation leads to an alternative conformational state of IL5R that prevents effective c recruitment. This possibility is favored by the hypothesis that IL5 and AF17121 might induce different conformational states of IL5R, as discussed above. Because the free cysteine (Cys⁶⁶) in the D1 domain is thought to be involved in the disulfide bond formation between IL5R and c (37), the spatial alignment of the D1 domain and the D2D3 tandem domain of IL5R would be important for recruiting c to the IL5·IL5R complex. Such an alignment could be altered by an inappropriate IL5R conformation stabilized by AF17121.

In summary, we have demonstrated that Asp⁵⁵, Arg¹⁸⁸, and Arg²⁹⁷ are the critical amino acid residues in determining IL5R recognition by AF17121 and competition for IL5. This provides a strong argument that AF17121 mimicry of IL5 plays a key role in the mechanism of receptor antagonism. The results illuminate a simplifying concept of how IL5R can recruit an activating component (IL5) and an inhibitory component (AF17121) into the same binding site and provide a rationale for designing low molecular mass antagonists. We have proposed the notion that the IL5-binding epitope on IL5R can be divided into two functional sites: site I may recognize IL5 with high specificity, whereas site II may play a role in activating c. We believe that this notion has biological significance that can be applied to other cytokine receptor systems such as IL3 and GM-CSF.

	FOOTNOTES

 
^* This work was supported by National Institutes of Health Grants GM55648 and AI40462. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Dept. of Biochemistry and Molecular Biology, Drexel University College of Medicine, 11102 New College Bldg., 245 N. 15th St., Philadelphia, PA 19102. Tel.: 215-762-4197; Fax: 215-762-4452; E-mail: imc23{at}drexel.edu.

¹ The abbreviations used are: IL, interleukin; IL5R, interleukin-5 receptor ; GM-CSF, granulocyte/macrop