Binding Interactions of Human Interleukin 5 with Its Receptor α Subunit
1995; Elsevier BV; Volume: 270; Issue: 16 Linguagem: Inglês
10.1074/jbc.270.16.9459
ISSN1083-351X
AutoresKyung Johanson, Edward R. Appelbaum, Michael L. Doyle, Preston Hensley, Baoguang Zhao, Sherin S. Abdel‐Meguid, Peter R. Young, Richard M. Cook, Steven A. Carr, Rosalie Matico, Donna M. Cusimano, Edward Dul, Monica Angelichio, Ian Brooks, Evon Winborne, Peter McDonnell, Thomas A. Morton, Donald B. Bennett, Theodore D. Sokoloski, Dean E. McNulty, Martin Rosenberg, Irwin Chaiken,
Tópico(s)Immune Response and Inflammation
ResumoHuman interleukin 5 (hIL5) and soluble forms of its receptor α subunit were expressed in Drosophila cells and purified to homogeneity, allowing a detailed structural and functional analysis. B cell proliferation confirmed that the hIL5 was biologically active. Deglycosylated hIL5 remained active, while similarly deglycosylated receptor α subunit lost activity. The crystal structure of the deglycosylated hIL5 was determined to 2.6-Å resolution and found to be similar to that of the protein produced in Escherichia coli. Human IL5 was shown by analytical ultracentrifugation to form a 1:1 complex with the soluble domain of the hIL5 receptor α subunit (shIL5Rα). Additionally, the relative abundance of ligand and receptor in the hIL5·shIL5Rα complex was determined to be 1:1 by both titration calorimetry and SDS-polyacrylamide gel electrophoresis analysis of dissolved cocrystals of the complex. Titration microcalorimetry yielded equilibrium dissociation constants of 3.1 and 2.0 n M, respectively, for the binding of hIL5 to shIL5Rα and to a chimeric form of the receptor containing shIL5Rα fused to the immunoglobulin Fc domain (shIL5Rα-Fc). Analysis of the binding thermodynamics of IL5 and its soluble receptor indicates that conformational changes are coupled to the binding reaction. Kinetic analysis using surface plasmon resonance yielded data consistent with the Kdvalues from calorimetry and also with the possibility of conformational isomerization in the interaction of hIL5 with the receptor α subunit. Using a radioligand binding assay, the affinity of hIL5 with full-length hIL5Rα in Drosophila membranes was found to be 6 n M, in accord with the affinities measured for the soluble receptor forms. Hence, most of the binding energy of the α receptor is supplied by the soluble domain. Taken with other aspects of hIL5 structure and biological activity, the data obtained allow a prediction for how 1:1 stoichiometry and conformational change can lead to the formation of hIL5·receptor αβ complex and signal transduction. Human interleukin 5 (hIL5) and soluble forms of its receptor α subunit were expressed in Drosophila cells and purified to homogeneity, allowing a detailed structural and functional analysis. B cell proliferation confirmed that the hIL5 was biologically active. Deglycosylated hIL5 remained active, while similarly deglycosylated receptor α subunit lost activity. The crystal structure of the deglycosylated hIL5 was determined to 2.6-Å resolution and found to be similar to that of the protein produced in Escherichia coli. Human IL5 was shown by analytical ultracentrifugation to form a 1:1 complex with the soluble domain of the hIL5 receptor α subunit (shIL5Rα). Additionally, the relative abundance of ligand and receptor in the hIL5·shIL5Rα complex was determined to be 1:1 by both titration calorimetry and SDS-polyacrylamide gel electrophoresis analysis of dissolved cocrystals of the complex. Titration microcalorimetry yielded equilibrium dissociation constants of 3.1 and 2.0 n M, respectively, for the binding of hIL5 to shIL5Rα and to a chimeric form of the receptor containing shIL5Rα fused to the immunoglobulin Fc domain (shIL5Rα-Fc). Analysis of the binding thermodynamics of IL5 and its soluble receptor indicates that conformational changes are coupled to the binding reaction. Kinetic analysis using surface plasmon resonance yielded data consistent with the Kdvalues from calorimetry and also with the possibility of conformational isomerization in the interaction of hIL5 with the receptor α subunit. Using a radioligand binding assay, the affinity of hIL5 with full-length hIL5Rα in Drosophila membranes was found to be 6 n M, in accord with the affinities measured for the soluble receptor forms. Hence, most of the binding energy of the α receptor is supplied by the soluble domain. Taken with other aspects of hIL5 structure and biological activity, the data obtained allow a prediction for how 1:1 stoichiometry and conformational change can lead to the formation of hIL5·receptor αβ complex and signal transduction. INTRODUCTIONInterleukin 5 (IL5)1 1The abbreviations used are: IL5interleukin 5EDCN-ethyl- N′-(3-diethylaminopropyl)carbodiimideGM-CSFgranulocyte-macrophage colony stimulating factorhIL5human interleukin 5IgGimmunoglobulin GIL3interleukin 3MALD-MSmatrix-assisted laser desorption mass spectrometryNHSN-hydroxysuccinimidePAGEpolyacrylamide gel electrophoresisPEGpolyethylene glycolPNGaseN-glycanaseshIL5Rαsoluble human interleukin 5 receptor α chainshIL5Rα-Fcsoluble human interleukin 5 receptor α-Fc chimeratPAtissue plasminogen activator. is a disulfide-linked, glycosylated dimeric protein which plays a prominent role in the maturation, proliferation, and activation of eosinophils (Basten and Beeson, 1970Basten A. Beeson P.B. J. Exp. Med. 1970; 131: 1288-1305Google Scholar; Metcalf et al., 1974Metcalf D. Parker J.W. Chester H.M. Kincade P.W. J. Cell. Physiol. 1974; 84: 275-290Google Scholar; Warren and Sanderson, 1985Warren D.J. Sanderson C.J. Immunology. 1985; 54: 615-623Google Scholar; Sanderson et al., 1985Sanderson C.J. Warren D.J. Strath M. J. Exp. Med. 1985; 162: 60-74Google Scholar; McKenzie and Sanderson, 1992McKenzie A.J. Sanderson C.J. Kishimoto T. Interleukins: Molecular Biology and Immunology. Vol. 51. Karger, Basel1992: 135-152Google Scholar; Campbell et al., 1987Campbell H.D. Tucker W.Q. Hort Y. Martinson M.E. Mayo G. Clutterbuck E.J. Sanderson C.J. Young I.G. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 6629-6633Google Scholar; Clutterbuck et al., 1987Clutterbuck E.J. Shields J.G. Gordon J. Smith S.H. Boyd A. Callard R.E. Campbell H.D. Young I.G. Sanderson C.J. Eur. J. Immunol. 1987; 17: 1743-1750Google Scholar; Lopez et al., 1988Lopez A.F. Sanderson C.J. Gamble J.R. Campbell H.D. Young I.G. Vadas M.A. J. Exp. Med. 1988; 167: 219-224Google Scholar). IL5's central role in the control of eosinophilia has suggested that it is a major cause of tissue damage in asthma and other eosinophil-related disorders (Hamid et al., 1991Hamid Q. Azzawi M. Ying S. Moqbel R. Wardlaw A.J. Corrigan C.J. Bradley B. Durham S.R. Collins J.V. Jeffery P.K. Quint D.J. Kay A.B. J. Clin Invest. 1991; 87: 1541-1546Google Scholar; Bentley et al., 1992Bentley A.M. Menz G. Storz C. Robinson D.S. Bradley B. Jeffery P.K. Durham S.R. Kay A.B. Am. Rev. Respir. Dis. 1992; 146: 500-506Google Scholar; Corrigan and Kay, 1992Corrigan C.J. Kay A.B. Immunol. Today. 1992; 13: 501-507Google Scholar).Interleukin 5 exerts its biological effects via cell surface receptor proteins (Chihara et al., 1990Chihara J. Plumas J. Gruart V. Tavernier J. Prinm L. Capron A. Capron M. J. Exp. Med. 1990; 172: 1347-1351Google Scholar; Lopez et al., 1990Lopez A.F. Eglinton J.M. Lyons A.B. Tapley P.M. To L.B. Clark S.G. Vadas M.A. J. Cell. Physiol. 1990; 145: 69-77Google Scholar; Plaetinck et al., 1990Plaetinck G. Van der Heyden J. Tavernier J. Fache I. Tuypensm T. Fischkoff S. Fiers W. Devos R. J. Exp. Med. 1990; 172: 683-691Google Scholar; Migita et al., 1991Migita M. Yamaguchi N. Mita S. Higuchi S. Hitoshi Y. Yoshida Y. Tomonaga M. Matsuda I. Tominaga A. Takatsu K. Cell Immunol. 1991; 133: 484-497Google Scholar). The human IL5 receptor is composed of two types of subunits, denoted α and β (Tavernier et al.,1991). The cloned cDNA for the α subunit (Murata et al., 1992Murata Y. Takaki S. Migita M. Kikuchi Y. Tominagam A. Takatsu K. J. Exp. Med. 1992; 175: 341-351Google Scholar) encodes a glycoprotein of 420 amino acids with an amino-terminal hydrophobic region (signal sequence, 20 amino acid residues), a glycosylated extracellular domain (324 residues), a transmembrane domain (21 residues), and a cytoplasmic domain (55 residues). The α subunit expressed in a soluble form (the extracellular domain) can bind to IL5 without the β chain and is IL5-specific (Tavernier et al., 1991Tavernier J. Devos R. Cornelis S. Tuypens T. Van der Heyden J. Fiers W. Plaetinck G. Cell. 1991; 66: 1175-1184Google Scholar; Murata et al., 1992Murata Y. Takaki S. Migita M. Kikuchi Y. Tominagam A. Takatsu K. J. Exp. Med. 1992; 175: 341-351Google Scholar; Devos et al., 1993Devos R. Guisez Y. Cornelius S. Verhee A. Van der Heyden J. Manneberg M. Lahm H.W. Fiers W. Tavernier J. Plaetinck G. J. Biol. Chem. 1993; 268: 6581-6587Google Scholar). In contrast, the β chain is identical to the β chains of GM-CSF and IL3 receptors (Tavernier et al., 1991Tavernier J. Devos R. Cornelis S. Tuypens T. Van der Heyden J. Fiers W. Plaetinck G. Cell. 1991; 66: 1175-1184Google Scholar; Lopez et al., 1990Lopez A.F. Eglinton J.M. Lyons A.B. Tapley P.M. To L.B. Clark S.G. Vadas M.A. J. Cell. Physiol. 1990; 145: 69-77Google Scholar) and appears to be needed for signal transduction.The ligand binding of human IL5 receptor α chain (hIL5Rα) is of high affinity, with a Kdof 0.6-1 n M reported for the α subunit expressed in various heterologous cells (Tavernier et al., 1991Tavernier J. Devos R. Cornelis S. Tuypens T. Van der Heyden J. Fiers W. Plaetinck G. Cell. 1991; 66: 1175-1184Google Scholar, Tavernier et al., 1992Tavernier J. Tuypens T. Plaetinck G. Verhee A. Fiers W. Devos R. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 7041-7045Google Scholar; Takaki et al., 1993Takaki S. Murata Y. Kitamura T. Miyajima A. Tominaga A. Takatsu K. J. Exp. Med. 1993; 177: 1523-1529Google Scholar). The above studies showed that the affinity for α subunit alone was lower by a factor of 2-4-fold than that for cells transfected with both α and β subunits. Direct interaction of the β subunit with IL5 has been suggested by cross-linking (Devos et al., 1991Devos R. Plaetinck G. Van der Heyden J. Cornelis S. Vanderkerckhove J. Fiers W. Tavernier J. EMBO J. 1991; 10: 2133-2137Google Scholar; Tavernier et al., 1991Tavernier J. Devos R. Cornelis S. Tuypens T. Van der Heyden J. Fiers W. Plaetinck G. Cell. 1991; 66: 1175-1184Google Scholar, Tavernier et al., 1992Tavernier J. Tuypens T. Plaetinck G. Verhee A. Fiers W. Devos R. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 7041-7045Google Scholar; Takaki et al., 1993Takaki S. Murata Y. Kitamura T. Miyajima A. Tominaga A. Takatsu K. J. Exp. Med. 1993; 177: 1523-1529Google Scholar) and as such could contribute energetically to the IL5 receptor affinity.The stoichiometry of IL5 binding to α subunit has been suggested to be 1:1 (one α subunit/IL5 dimer). This was based on both cross-linking data as well as from quantitating the abundance of each component in a complex isolated by immunoaffinity capture followed by gel filtration chromatography (Devos et al., 1993Devos R. Guisez Y. Cornelius S. Verhee A. Van der Heyden J. Manneberg M. Lahm H.W. Fiers W. Tavernier J. Plaetinck G. J. Biol. Chem. 1993; 268: 6581-6587Google Scholar). A 1:1-sized cross-linked IL5-receptor α chain complex also has been observed indirectly (Mita et al., 1988Mita S. Harada N. Naomi S. Hitoshi Y. Sakamoto K. Akagi M. Tominaga A. Takatsu K. J. Exp. Med. 1988; 168: 863-878Google Scholar; Migita et al., 1991Migita M. Yamaguchi N. Mita S. Higuchi S. Hitoshi Y. Yoshida Y. Tomonaga M. Matsuda I. Tominaga A. Takatsu K. Cell Immunol. 1991; 133: 484-497Google Scholar; Devos et al., 1991Devos R. Plaetinck G. Van der Heyden J. Cornelis S. Vanderkerckhove J. Fiers W. Tavernier J. EMBO J. 1991; 10: 2133-2137Google Scholar; Tavernier et al., 1992Tavernier J. Tuypens T. Plaetinck G. Verhee A. Fiers W. Devos R. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 7041-7045Google Scholar). The above 1:1 stoichiometry is unexpected, since the IL5 dimer has two 4-helix bundle domains. Each of these domains resembles the 4-helix bundle domain of monomeric growth factor proteins such as growth hormone (Abdel-Meguid et al., 1987Abdel-Meguid S.S. Shieh H-S. Smith W.W. Dayringer H.E. Violand B.N. Bentle L.A. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 6434-6437Google Scholar; de Vos et al., 1992de Vos A.M. Ultsch M. Kossiakoff A.A. Science. 1992; 255: 306-312Google Scholar) and IL4 (Redfield et al., 1991Redfield C. Smith L.T. Boyd J. Lawrence G.M.P. Edwards R.G. Smith R.A.G. Dobson C.M. Biochemistry. 1991; 30: 11029-11035Google Scholar), and each of the latter can bind at least one molecule of receptor/molecule ligand. Each IL5 monomeric domain might therefore be expected to bind at least one receptor molecule. Interestingly, the molecular weight of the cross-linked product of IL5 with β chain also suggests a 1:1 stoichiometry (Devos et al., 1991Devos R. Plaetinck G. Van der Heyden J. Cornelis S. Vanderkerckhove J. Fiers W. Tavernier J. EMBO J. 1991; 10: 2133-2137Google Scholar; Tavernier et al., 1992Tavernier J. Tuypens T. Plaetinck G. Verhee A. Fiers W. Devos R. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 7041-7045Google Scholar). The structural basis for the 1:1 stoichiometry of IL5 and its α and β receptor subunits is as yet undetermined. The above methods used to define 1:1 stoichiometry, namely cross-linking and chromatographic isolation of complexes, likely would not detect weakly associating multimeric complexes. Nonetheless, if correct, 1:1 stoichiometry must be accounted for in explaining signal transduction.We have initiated an investigation of the quantitative and structural properties of the interaction of hIL5 with its receptor. In this report, we describe the expression and production of Drosophila-expressed hIL5 and both soluble and membrane-associated forms of its receptor α chain. We report the three-dimensional structure of deglycosylated Drosophila-expressed hIL5 and show that it is nearly identical to the protein produced in Escherichia coli (Milburn et al., 1993Milburn M.V. Hassell A.M. Lambert M.H. Jordan S.R. Proudfoot A.E.I. Graber P. Wells T.N.C. Nature. 1993; 363: 172-176Google Scholar). We further report the interaction properties of these functionally active proteins. The equilibrium binding parameters of soluble and full-length IL5Rα chain show that the soluble domain of the receptor accounts for most, if not all, of the binding energy of cell-bound receptor α subunit recognition. We provide rigorous confirmatory evidence for a stoichiometry of 1:1 for the hIL5·receptor α chain complex. Finally, we report new data on the thermodynamics and kinetics of the IL5·receptor interaction which argue that conformational change occurs in this process. Together, the data suggest a model for the interaction of hIL5 with its receptor α and β subunits leading to signal transduction.EXPERIMENTAL PROCEDURESMaterials For protein purification, Q-Sepharose fast flow, phenyl-Sepharose fast flow high substitution, Superose 12 and 6, and Superdex 75 were from Pharmacia LKB Biotechnol. Hydroxylapatite high resolution medium was from Calbiochem. ABX-plus ion exchange resin was from J. T. Baker Chemical Co. Lentil lectin-agarose and methyl D-mannosylpyranoside were from E-Y Lab (San Mateo, CA). For BIAcore kinetics measurements, the sensor chips CM5, surfactant P20, and the amine coupling kit containing N-hydroxysuccinimide (NHS), N-ethyl- N′-(3-diethylaminopropyl)carbodiimide (EDC), and ethanolamine hydrochloride all were from Pharmacia Biosensor. Murine IL3 was obtained from Genzyme (Cambridge, MA), and purified murine IL5 (specific activity 4.9 × 107units/mg) was prepared in a baculovirus expression system using Spodoptera frugiperda 21 cells (Mitchell et al., 1993Mitchell D.L. Young M.A. Entwistle C. Davies A.N. Cook R.M. Dodd I. Biochem. Soc. Trans. 1993; 21: 332SGoogle Scholar). The murine pre-B cell line, denoted B13, was obtained courtesy of R. Palacios, Basel Institute of Immunology, Switzerland. Expression of Recombinant IL5, shIL5Rα, and shIL5Rα-Fc Human IL5 and various forms of shIL5Rα were expressed in a Drosophila cell culture system (van der Straten et al., 1989van der Straten A. Johansen H. Rosenberg M. Sweet R.W. Curr. Methods Mol. Biol. 1989; 1: 1-8Google Scholar; Angelichio et al., 1991Angelichio M.L. Beck J.A. Johansen H. Ivey-Hoyle M. Nucleic Acids Res. 1991; 19: 5037-5043Google Scholar). Each gene to be expressed was cloned between a copper sulfate-inducible metallothionein promoter and an SV40 late polyadenylation site and introduced into Drosophila Schneider 2 (S2) cells by cotransfection with a vector carrying a hygromycin B resistance gene. Hygromycin B-resistant cells were selected as a stable polyclonal population, induced with copper sulfate, and examined for expression by Western blot analysis. Antisera used for Western blots were raised by immunization of rabbits with denatured fusion proteins (expressed in E. coli and gel-purified) carrying hIL5 or hIL5Rα sequences.hIL5 ExpressionHuman IL5 expression was achieved with a vector encoding a fusion of a human tissue plasminogen activator (tPA) secretion signal sequence to amino acids 4-115 of mature hIL5. Cleavage of the signal sequence during secretion leaves four amino acids from tPA fused to the NH2terminus of hIL5, as reported for other such fusions (Culp et al., 1991Culp J.S. Johansen H. Hellmig B. Beck J. Matthews T.J. Delers A. Rosenberg M. BioTechnology. 1991; 9: 174-178Google Scholar). Western blot analysis of medium from induced cultures revealed a protein with an apparent molecular mass of 15 kDa under reducing conditions and 30 kDa under nonreducing conditions, as expected for the disulfide-linked homodimer. The cell line was expanded to 30 liters to produce hIL5.shIL5Rα ExpressionTo express shIL5Rα in Drosophila, a cDNA fragment encoding the shIL5α protein precursor (signal sequence followed by mature shIL5Rα) was cloned from human eosinophil RNA by polymerase chain reaction and inserted into an expression vector (pMTAL) to yield pMTAL-shIL5Rα, a vector that does not contain the tPA sequence. DNA sequence analysis showed the shIL5Rα sequence to be identical to that reported by others (Tavernier et al., 1992Tavernier J. Tuypens T. Plaetinck G. Verhee A. Fiers W. Devos R. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 7041-7045Google Scholar). Western blot analysis of induced culture carrying this construct revealed a broad band with an apparent molecular mass of 40-43 kDa under both reducing and nonreducing conditions. The cell line was expanded to 4 liters in spinner flasks for production of shIL5Rα.shIL5Rα-Fc Chimera ExpressionA fusion of shIL5Rα to the CH2-CH3 region of human IgG1 was constructed and then expressed in Drosophila cells as described above for shIL5Rα. In this construct, the COOH terminus of shIL5Rα (Arg315of mature shIL5Rα) was linked to the NH2terminus of the IgG1 hinge region (Glu226, Kabat numbering) by the short sequence Ile-Glu-Gly-Arg, a Factor Xa cleavage site, allowing recovery of shIL5Rα by proteolytic cleavage of the purified chimera. The Fc region also had a mutation of Cys230to Ala. Western analysis using anti-shIL5Rα antiserum showed that this protein (70 kDa under reducing conditions, 140 kDa and hence dimeric under nonreducing conditions) was produced in Drosophila cell cultures at levels comparable to shIL5Rα lacking the Fc region. The cell line was expanded to 4 liters for production of the Fc receptor chimera. In the present study, we evaluated the binding properties of the uncleaved and unreduced chimera (molecular mass 140 kDa), which retains both sIL5Rα fragments/molecule. Expression of Full-length hIL5Rα in Drosophila Cells The transmembrane and cytoplasmic region of hIL5Rα was cloned by reverse transcriptase polymerase chain reaction from a butyrate-induced eosinophilic subline of human promyelocytic HL-60 cells and the intended DNA sequence confirmed. This fragment was cloned into pMTAL-shIL5Rα to create a new vector, pMTAL-hIL5Rα, containing the full-length receptor (20 amino acid signal sequence followed by 400 amino acid mature protein). Western blot analysis of induced cells carrying this construct revealed a band with an apparent molecular mass of 50 kDa. Protein Purification Expression levels of hIL5, shIL5Rα, and shIL5Rα-Fc in Drosophila media used for large scale purification were estimated to be 22, 17, and 10 mg/liter, respectively, from Western blot analysis and Coomassie Blue staining of SDS-PAGE. Western blot analyses of all three products in culture media suggested heterogeneity in the carbohydrate content. Protein concentration of the starting material and at various stages of purification was estimated by the Micro BCA Protein assay. Extinction coefficients at 280 nm calculated from tryptophan and tyrosine contents matched well with those measured by amino acid analysis: E280(cm - mg/ml)−1= 0.63 for hIL5, 1.9 for shIL5Rα, and 1.65 for shIL5Rα-Fc. Throughout this study, concentrations of purified proteins were based on absorbance at 280 nm.Purification of hIL528 liters of sterile-filtered Drosophila media were diluted 3-fold with water, adjusted to pH 8.0, and loaded onto a Q-Sepharose FF column (11.4 × 14.1 cm, Pharmacia) equilibrated with 20 m M Tris-HCl, pH 8.0. hIL5 in the flow-through fractions was pooled, adjusted to pH 7.4 with 6 M HCl, and loaded onto a hydroxylapatite column (HA) equilibrated with Buffer T (20 m M Tris-HCl, pH 7.4). The column was washed with Buffer T, and bound hIL5 was eluted with 0.25 M potassium phosphate, pH 7.5. Pooled fractions from the HA column were mixed with an equal volume of 3 M ammonium sulfate and applied to phenyl-Sepharose ff (5.7 × 7 cm) equilibrated with 1.5 M NH4SO4in Buffer P (20 m M sodium phosphate, pH 7.0). hIL5 was eluted with a linear gradient of 1.5-0 M NH4SO4in Buffer P. The pooled fraction from phenyl-Sepharose (600 ml) was concentrated to 80 ml, loaded onto Superose 12, and eluted with Buffer T + 0.15 M NaCl. The final yield of hIL5, approximately 500 mg, was >98% electrophoretically homogeneous based on silver staining of SDS-PAGE gels; nucleic acid and endotoxin were undetectable.Purification of shIL5Rα4.3 liters of sterile-filtered conditioned Drosophila media (MRD3) were diluted 2-fold with Buffer T, adjusted to pH 7.4, and loaded onto a Q-Sepharose column (5 × 8.8 cm, Pharmacia) equilibrated with Buffer T. Receptor in flow-through fractions (9 L) was concentrated to 1900 ml using a Miniset 10K cutoff membrane (Filtron, Northborough MA). The concentrated Q-Sepharose pool was applied to an ABX-plus column (5 × 7 cm, J. T. Baker Chemical Co.) equilibrated with Buffer A (20 m M sodium acetate, pH 5.5), and washed with Buffer A. Bound receptor was eluted with a linear gradient of 0-1 M NaCl in Buffer A. Fractions containing receptor were immediately neutralized with pH 7.4 1 M Tris-HCl buffer. This ABX pool (240 ml) was adjusted to 5 m M CaCl2, sterile filtered, and loaded onto two lentil lectin columns (2.5 × 5.8 cm) in tandem, equilibrated with Buffer C (20 m M Tris-HCl, pH 7.2, 0.15 M NaCl, and 5 m M CaCl2). Bound receptor was eluted with 0.25 M methyl D-mannosylpyranoside in Buffer C. The lentil column pool (88 ml) was concentrated to 12 ml on a YM-10 membrane (Amicon) and loaded onto a Superdex 75 preparation grade column (2.6 × 96.4 cm) equilibrated in 0.1 M HEPES, pH 7.5. Pooled fractions (29 ml) were stored at −70°C. The Superdex 75 column eluate contained 20 mg of receptor, with >98% electrophoretic homogeneity based on silver staining of SDS-PAGE gels.Purification of shIL5Rα-Fc Chimera4 liters of conditioned Drosophila media were diafiltered into 0.1 M Tris-HCl, pH 8.1, and concentrated to 600 ml. The diafiltered and concentrated media were loaded onto a Protein A-Sepharose ff column (2.5 × 5.8 cm, Pharmacia) equilibrated with 0.1 M Tris-HCl, pH 8.0, and washed with the same buffer. The Fc chimera was eluted with 0.1 M glycine HCl, pH 3.0, and fractions were collected in 1 M Trizma base to raise the final pH to 7.7. The chimera was fractionated on a Superose 6 column into 0.1 M HEPES, pH 7.5. 2.8 mg of Fc chimera (>95% electrophoretically homogeneous in SDS-PAGE) was obtained from 4 liters of media. Mass Spectroscopy Matrix-assisted laser desorption mass spectrometry (MALD-MS) data were obtained on a Vestec Research model mass spectrometer (Persceptive Biosystems, Boston, MA). Samples were prepared for analysis by mixing 1 μl of a 1-6 pmol/ml solution of protein in phosphate-buffered saline, pH 7.4, with 1 μl of the matrix, sinapinic acid, on the stainless steel target. The 337-nm line from a nitrogen laser (10 ns pulse width, 10 Hz repetition rate) was used for desorption/ionization of the sample. The spectra were the sum of ∼20-70 laser shots. Spectra were calibrated externally using the BSA (bovine serum albumin) and BSA dimer peaks.Electrospray mass spectra were recorded on a Sciex API-III triple quadrupole mass spectrometer fitted with a standard pneumatically assisted nebulization probe and an atmospheric pressure ionization source (Sciex, Ontario, Canada). Buffers and salts were removed by high performance liquid chromatography prior to mass spectrometry. Sample was loaded in aqueous buffer onto a 2.1 mm inner diameter C8 guard column, flushed with 0.1% trifluoroacetic acid for 5 min and then step eluted with 60% CH3CN, 40% H2O, 0.1% trifluoroacetic acid. An aliquot of the sample was concentrated to near dryness and brought to a final concentration of ∼20 pmol/μl with methanol/water (50:50 v/v) containing 0.2% formic acid. Sample was introduced into the mass spectrometer by infusion with a syringe pump (Harvard Instruments) at 2 μl/min. Approximately 2 min of data were recorded and ∼40 pmol of the sample consumed. Enzymatic Deglycosylation and Cross-linking to 125I-hIL5 hIL5 was iodinated with Bolton-Hunter reagent to a specific activity of >350 Ci/mmol using the manufacturer's protocol (DuPont NEN). 100 μg of shIL5Rα or 10 μg of 125I-hIL5 were enzymatically deglycosylated by treating with 5,000 units of PNGaseF (New England Biolabs) in 50 m M sodium phosphate, pH 7.5, for 12-16 h at 37°C in the presence of a mixture of protease inhibitors (1 m M phenylmethylsulfonyl fluoride, 2 μg/ml aprotinin, 20 μM leupeptin). Analysis of the shIL5Rα product by Western blot showed a reduction in molecular mass of ∼3-4 kDa relative to untreated receptor, consistent with loss of carbohydrate but absent proteolytic degradation (data not shown). The mobility was identical to shIL5Rα enzymatically deglycosylated in denaturing conditions (0.5% SDS, 1%β-mercaptoethanol), suggesting that deglycosylation was complete. Similarly, deglycosylation of hIL5 led to ∼1 kDa reduction in mobility on Western blots (data not shown).Cross-linking was performed by incubating 1 n M each of 125I-hIL5 and shIL5Rα together with 1 n M of the bifunctional cross-linker bis-suberimidate (Pierce) in 25 m M HEPES, pH 7.2, 0.1% BSA for 20-30 min at 4°C. Products were analyzed by 12.5% SDS-PAGE in reducing conditions, followed by gel drying and autoradiography. B13 Cell Proliferation Assay The murine IL5/IL3-dependent cell line LyH7.B13 was subcultured twice weekly in RPMI 1640 medium (Life Technologies, Inc.), supplemented with L-glutamine, non-essential amino acids, sodium pyruvate, penicillin-streptomycin (all Life Technologies, Inc.), plus 2-mercaptoethanol (5 × 10−5M, Sigma), 10% fetal bovine serum (Globepharm), and 1-10 units of recombinant murine IL5 (Mitchell et al., 1993Mitchell D.L. Young M.A. Entwistle C. Davies A.N. Cook R.M. Dodd I. Biochem. Soc. Trans. 1993; 21: 332SGoogle Scholar). For assays, cells were washed and maintained in medium without IL5 for a minimum of 2 h. They were cultured for 48 h in triplicate (5,000 cells/well) in 96-well round bottom plates in the presence of appropriately diluted test samples and pulsed with 0.5 μCi of [3H]thymidine (Amersham) for the final 4 h. They were processed for scintillation counting in a 1205 Betaplate (LKB Wallac). Results are presented as Stimulation Index ± standard deviation, relative to unstimulated signal. Crystallization Crystals of deglycosylated hIL5 (from PNGaseF treatment in 0.1 M HEPES, pH 7.4, 25°C, 48 h) and of the complex between glycosylated hIL5 and shIL5Rα were obtained using the hanging drop method of vapor diffusion (McPherson, 1976McPherson A.J. Methods Biochem. Anal. 1976; 23: 249-345Google Scholar). Deglycosylated hIL5 was concentrated to approximately 10 mg/ml in 100 m M HEPES buffer, pH 7.2. The complex was prepared by mixing equimolar amounts of hIL5 (dimer) and shIL5Rα in 100 m M HEPES, pH 7.2. Typically, 2 μl of either protein solution were mixed with 2 μl of various precipitant concentrations and vapor equilibrated at several temperatures against the solution added to the protein. At each temperature, a three-dimensional matrix was established for each precipitant, in which protein concentration, precipitant concentration, and pH were varied. Ammonium sulfate, sodium chloride, and various polyethylene glycols were used as precipitants. Determination of Molecular Ratio of hIL5 and shIL5Rα in Cocrystals Cocrystals were harvested under a microscope, washed twice in crystallization solution P (27% PEG 600 (v/v) in 100 m M sodium acetate, pH 5.0), solubilized in SDS-sample buffer, and analyzed by 12% SDS-PAGE. The gels were either stained using Coomassie Blue or silver-stained. Intensities of the bands of IL5 and shIL5Rα in the cocrystals were compared with those of controls (hIL5, shIL5Rα, or combinations of both proteins) by scanning with IS-1000 Digital Imaging System (Alpha Innotech Co, San Leandro CA), and the peak areas were integrated. Crystal Structure Determination of Deglycosylated hIL5 Crystals of deglycosylated hIL5 belong to the space group C2 with a = 118.3 Å, b = 24.3 Å, c = 43.8 Å, and β = 110°. Assuming one deglycosylated monomer/asymmetric unit, the VMvalue is 2.1 Å3/dalton, which is within the range found for most protein crystals (Matthews, 1968Matthews B.W. J. Mol. Biol. 1968; 33: 491-497Google Scholar). These crystals show diffraction beyond 2.6-Å resolution and are stable in the x-ray beam for several days.X-ray diffraction data were measured from a single crystal using a Siemens two-dimensional position-sensitive detector. Approximately 5776 reflections were measured in 2 days to give 3054 u
Referência(s)