The CryIA(c) Receptor Purified from Manduca sexta Displays Multiple Specificities
1995; Elsevier BV; Volume: 270; Issue: 35 Linguagem: Inglês
10.1074/jbc.270.35.20309
ISSN1083-351X
AutoresLuke Masson, Yang-jiang Lu, Alberto Mazza, Roland Brousseau, Michael J. Adang,
Tópico(s)Toxin Mechanisms and Immunotoxins
ResumoThe kinetic binding characteristics of four Bacillus thuringiensis CryI insecticidal crystal proteins to a Cry-binding protein, purified from Manduca sexta brush-border vesicles, were analyzed by an optical biosensor. This 120-kilodalton binding protein, previously determined to be aminopeptidase N, was converted to a 115-kilodalton water-soluble form by removing the attached glycosylphosphatidylinositol anchor with phospholipase C. The solubilized form recognized the three major subclasses of CryIA toxins but not CryIC even though all four CryI proteins are toxic to larvae of M. sexta. CryIA(a) and CryIA(b) toxins bound to a single site on the solubilized aminopeptidase N molecule whereas CryIA(c) bound to two distinct sites. Apparent kinetic rate constants were determined for each binding reaction. All three CryIA toxins exhibited moderately fast on rates (~10-5M-1 s-1) and a slow reversible off rate (~10-3 s-1). Although the second CryIA(c)-binding site retained a moderately fast association rate, it was characterized by a rate of dissociation from the aminopeptidase an order of magnitude faster than observed for the other CryIA-binding sites. CryIA(c) binding to both sites was strongly inhibited in the presence of N-acetylgalactosamine (IC50 = 5 mM) but not N-acetylglucosamine, mannose, or glucose. CryIA(a) and CryIA(b) binding were unaffected in the presence of the same sugars. Our results serve to illustrate both the complexity and the diverse nature of toxin interactions with Cry-binding proteins. The kinetic binding characteristics of four Bacillus thuringiensis CryI insecticidal crystal proteins to a Cry-binding protein, purified from Manduca sexta brush-border vesicles, were analyzed by an optical biosensor. This 120-kilodalton binding protein, previously determined to be aminopeptidase N, was converted to a 115-kilodalton water-soluble form by removing the attached glycosylphosphatidylinositol anchor with phospholipase C. The solubilized form recognized the three major subclasses of CryIA toxins but not CryIC even though all four CryI proteins are toxic to larvae of M. sexta. CryIA(a) and CryIA(b) toxins bound to a single site on the solubilized aminopeptidase N molecule whereas CryIA(c) bound to two distinct sites. Apparent kinetic rate constants were determined for each binding reaction. All three CryIA toxins exhibited moderately fast on rates (~10-5M-1 s-1) and a slow reversible off rate (~10-3 s-1). Although the second CryIA(c)-binding site retained a moderately fast association rate, it was characterized by a rate of dissociation from the aminopeptidase an order of magnitude faster than observed for the other CryIA-binding sites. CryIA(c) binding to both sites was strongly inhibited in the presence of N-acetylgalactosamine (IC50 = 5 mM) but not N-acetylglucosamine, mannose, or glucose. CryIA(a) and CryIA(b) binding were unaffected in the presence of the same sugars. Our results serve to illustrate both the complexity and the diverse nature of toxin interactions with Cry-binding proteins. INTRODUCTIONDuring sporulation the Gram-positive bacterium Bacillus thuringiensis produces a variety of intracellular insecticidal crystal proteins in the form of crystalline inclusions. These inclusions, when ingested by susceptible insects, are solubilized in the alkaline environment of the insect midgut where the protoxins undergo enzymatic conversion to the active toxin form by resident gut proteases(1Knowles B.H. Adv. Insect Physiol. 1994; 24: 275-308Crossref Scopus (311) Google Scholar). After activation, CryI toxins have been shown to bind to specific high affinity receptors on the surface of the midgut epithelial cell layer(2Hofmann C. Luthy P. Hutter R. Pliska V. Eur. J. Biochem. 1988; 173: 85-91Crossref PubMed Scopus (206) Google Scholar, 3van Rie J. Jansens S. Höfte H. Degheele D. van Mellaert H. Eur. J. Biochem. 1989; 186: 239-247Crossref PubMed Scopus (280) Google Scholar). Independent studies have shown that B. thuringiensis toxins are able to form ion channels in susceptible cultured insect cells at low concentrations (4Schwartz J.-L. Garneau L. Masson L. Brousseau R. Biochim. Biophys. Acta. 1991; 1065: 250-260Crossref PubMed Scopus (71) Google Scholar) or in artificial lipid bilayers in the complete absence of receptors at high toxin concentrations(5Slatin S.L. Abrams C.K. English L. Biochem. Biophys. Res. Commun. 1990; 169: 765-772Crossref PubMed Scopus (120) Google Scholar, 6Schwartz J.-L. Garneau L. Savaria D. Masson L. Brousseau R. Rousseau E. J. Membr. Biol. 1993; 132: 53-62Crossref PubMed Scopus (110) Google Scholar). Perturbation of the intracellular ionic homeostasis created by these ion channels is thought to ultimately result in cell death by lysis(7Knowles B.H. Ellar D.J. Biochim. Biophys. Acta. 1987; 924: 509-518Crossref Scopus (373) Google Scholar).Integral to our understanding of the toxin mode of action is the study of toxin interactions with their specific binding proteins. Most of the in vitro studies characterizing these proteins to date have utilized brush border membrane vesicles (BBMVs) 1The abbreviations used are: BBMVsbrush border membrane vesiclesSPRsurface plasmon resonanceHBSHEPES-buffered salineRUresonance unitsAPNaminopeptidase NCHAPS3-[(3-cholamidopropyl)dimethylammonio]-1-propane-sulfonatePIPLCphosphatidylinositol-specific phospholipase CPAGEpolyacrylamide gel electrophoresis. purified from the gut epithelium of susceptible insects (8Wolfersberger M. Luthy P. Mauer A. Parenti P. Sacchi V. Giordana B. Hanozet G. Comp. Biochem. Physiol. 1987; 86: 301-308Crossref Scopus (544) Google Scholar) or immunochemical staining of midgut sections(9Bravo A. Hendrickx K. Jansens S. Peferoen M. J. Invert. Pathol. 1992; 60: 247-253Crossref Scopus (93) Google Scholar). Characterization of toxin-binding sites on BBMVs from a variety of larvae has revealed highly complex patterns of toxin binding. The existence of separate distinct classes of toxin-binding proteins as well as single binding sites capable of recognizing multiple toxins has been clearly demonstrated by numerous competition and ligand blotting studies(3van Rie J. Jansens S. Höfte H. Degheele D. van Mellaert H. Eur. J. Biochem. 1989; 186: 239-247Crossref PubMed Scopus (280) Google Scholar, 10Hofmann C. Vanderbruggen H. Höfte H. van Rie J. Jansens S. van Mellaert H. Proc. Natl. Acad. Sci. U. S. A. 1988; 85: 7844-7848Crossref PubMed Scopus (342) Google Scholar, 11van Rie J. Jansens S. Höfte H. Degheele D. van Mellaert H. Appl. Env. Microbiol. 1990; 56: 1378-1385Crossref PubMed Google Scholar, 12van Rie J. McGaughey W.H. Johnson D.E. Barnett B.D. van Mellaert H. Science. 1990; 247: 72-74Crossref PubMed Scopus (284) Google Scholar, 13Wolfersberger M.G. Experientia. 1990; 46: 475-477Crossref PubMed Scopus (132) Google Scholar, 14Garczynski S.F. Crim J.W. Adang M.J. Appl. Environ. Microbiol. 1991; 57: 2816-2820Crossref PubMed Google Scholar, 15Ferré J. Real M.D. van Rie J. Jansens S. Peferoen M. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 5119-5123Crossref PubMed Scopus (299) Google Scholar, 16Oddou P. Hartmann H. Radecke F. Geiser M. Eur. J. Biochem. 1993; 212: 145-150Crossref PubMed Scopus (42) Google Scholar, 17Estada U. Ferre J. Appl. Env. Microbiol. 1994; 60: 3840-3846Crossref PubMed Google Scholar). However, the presence of multiple binding proteins on the surface of BBMVs, high levels of nonspecific binding, and inherent toxin integration into the membrane have made the interpretation of brush border binding results rather difficult and have led to some very complex interaction models(3van Rie J. Jansens S. Höfte H. Degheele D. van Mellaert H. Eur. J. Biochem. 1989; 186: 239-247Crossref PubMed Scopus (280) Google Scholar, 17Estada U. Ferre J. Appl. Env. Microbiol. 1994; 60: 3840-3846Crossref PubMed Google Scholar, 18MacIntosh S.C. Stone T.B. Jokerst R.S. Fuchs R.L. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 8930-8933Crossref PubMed Scopus (90) Google Scholar). In one notable case, a colony of Trichoplusia ni, having developed resistance to CryIA(b) by laboratory selection, failed to show resistance to CryIA(c) although in vitro binding studies demonstrated they share the same binding site(17Estada U. Ferre J. Appl. Env. Microbiol. 1994; 60: 3840-3846Crossref PubMed Google Scholar). These results suggest that binding site predictions based on BBMV studies do not necessarily correlate with in vivo toxicity. Clearly, studies on isolated toxin-binding proteins would be crucial in resolving the relationship between binding sites and binding proteins. Recently, three independent sources have described the purification of a 210-kilodalton (kDa) CryIA(b) (19Vadlamudi R.K. Ji T.H. Bulla Jr., L.A. J. Biol. Chem. 1993; 268: 12334-12340Abstract Full Text PDF PubMed Google Scholar, 20Vadlamudi R.K. Weber E. Ji I. Ji T.H. Bulla Jr., L.A. J. Biol. Chem. 1995; 270: 5490-5494Abstract Full Text Full Text PDF PubMed Scopus (290) Google Scholar) and a 120-kDa CryIA(c) (21Sangadala S. Walters F.W. English L.H. Adang M.J. J. Biol. Chem. 1994; 269: 10088-10092Abstract Full Text PDF PubMed Google Scholar, 22Knight P.J.K. Crickmore N. Ellar D.J. Mol. Microbiol. 1994; 11: 429-436Crossref PubMed Scopus (357) Google Scholar) toxin-binding protein from Manduca sexta. The CryIA(c)-binding protein was functionally determined to be aminopeptidase N (21Sangadala S. Walters F.W. English L.H. Adang M.J. J. Biol. Chem. 1994; 269: 10088-10092Abstract Full Text PDF PubMed Google Scholar, 22Knight P.J.K. Crickmore N. Ellar D.J. Mol. Microbiol. 1994; 11: 429-436Crossref PubMed Scopus (357) Google Scholar) whereas the CryIA(b)-binding protein was reported to share sequence similarity with the cadherin family of glycoproteins(20Vadlamudi R.K. Weber E. Ji I. Ji T.H. Bulla Jr., L.A. J. Biol. Chem. 1995; 270: 5490-5494Abstract Full Text Full Text PDF PubMed Scopus (290) Google Scholar).In this study, using surface plasmon resonance (SPR), we provide the first detailed kinetic analysis of the interaction between three B. thuringiensis CryIA toxin subclasses and a solubilized form of the 120-kDa CryIA(c)-binding protein purified from M. sexta. SPR is an optical detection technique which allows direct interaction analysis between a ligand immobilized on a modified dextran sensor chip and a specific analyte in a continuous flow system(23Jönsson U. Fägerstam L. Ivarsson B. Johnsson B. Karlsson R. Lundh K. Löf S. Persson B. Roos H. Rönnberg I. Sjölander S. Stenberg E. St R. Urbaniczky C. Ö stlin H. Malmqvist M. BioTechniques. 1991; 11: 620-627PubMed Google Scholar). The reactants are monitored in real time without the use of labels thus permitting the determinations of kinetic rate constants, binding affinities, and binding site characterization(24Karlsson R. Michalesson A. Mattsson L. J. Immunol. Methods. 1991; 145: 229-240Crossref PubMed Scopus (1008) Google Scholar, 25Fägerstam L.G. Frostell-Karlsson Å. Karlsson R. Persson B. Rönnberg I. J. Chromatogr. 1992; 597: 397-410Crossref PubMed Scopus (424) Google Scholar, 26O'Shannessy D.J. Brigham-Burke M. Soneson K.K. Hensley P. Brooks I. Anal. Biochem. 1993; 212: 457-468Crossref PubMed Scopus (513) Google Scholar). SPR has been used previously to determine the kinetics of CryIA binding to BBMVs from both the spruce budworm (Choristoneura fumiferana) and the diamondback moth (Plutella xylostella)(27Masson L. Mazza A. Brousseau R. Anal. Biochem. 1994; 218: 405-412Crossref PubMed Scopus (63) Google Scholar, 28Masson L. Mazza A. Brousseau R. Tabashnik B. J. Biol. Chem. 1995; 270: 11887-11896Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar). In this report, we present a detailed kinetic analysis of CryIA toxin binding interactions with a specific membrane protein purified from the brush border of the lepidopteran larva M. sexta. Detailed evidence is presented showing that two toxins, CryIA(a) and CryIA(b), recognize a single binding site on the purified 115-kDa protein and that CryIA(c) binds to two binding sites. Furthermore, we show that the latter binding is selectively inhibited by the amino sugar N-acetylgalactosamine.MATERIALS AND METHODSToxin PurificationThe three recombinant CryIA protoxins from B. thuringiensis kurstaki (strain NRD-12) were expressed in Escherichia coli as cytoplasmic inclusions and purified as described elsewhere(29Masson L. Préfontaine G. Péloquin L. Lau P.C.K. Brousseau R. Biochem. J. 1990; 269: 507-512Crossref PubMed Scopus (97) Google Scholar). A second CryIA(c) protoxin was isolated from the single gene strain of B. thuringiensis kurstaki (HD-73). All protoxins were activated by solubilization in 40 mM carbonate buffer pH = 10.5 and treated with trypsin (0.1% (w/v) final concentration) for 3 h at ambient temperature. The 65-kDa trypsin-resistant proteins were purified by ion-exchange liquid chromatography using either Mono Q or Q-Sepharose anion exchangers (Pharmacia LKB AB) buffered with 40 mM carbonate buffer pH = 10.5. Bound toxins were eluted using a 50-500 mM NaCl gradient, dialyzed against four changes of distilled water, and the precipitated protein recovered and stored as a concentrated stock in water. Protein levels were quantitated by the dye-binding method of Bradford (30Bradford M.M. Anal. Biochem. 1976; 72: 248-254Crossref PubMed Scopus (213347) Google Scholar) using bovine serum albumin as a standard. To minimize the formation of multimers, a small aliquot of precipitated toxin was resuspended in HBS (10 mM HEPES, pH 7.4, 150 mM NaCl, 3.4 mM EDTA) before each experiment and kept on ice(31Feng Q. Becktel W.J. Biochemistry. 1994; 33: 8521-8526Crossref PubMed Scopus (22) Google Scholar). CryIA(c) toxin for radiolabeling was purified by fast protein liquid chromatography and directly iodinated by the chloramine-T method (32Hunter W. Greenwood F. Nature. 1962; 194: 495-496Crossref PubMed Scopus (5839) Google Scholar) as described in Garczynski et al.(14Garczynski S.F. Crim J.W. Adang M.J. Appl. Environ. Microbiol. 1991; 57: 2816-2820Crossref PubMed Google Scholar). Specific activity of the labeled toxin was about 50 mCi/mg input toxin.Preparation of Brush Border Membrane VesiclesM. sexta eggs were obtained from the United States Department of Agriculture, Agricultural Research Services, Biosciences Research Laboratory (Fargo, ND) and larvae reared on artificial diet (Southland Products, Inc., Lake Village, AR). BBMVs were made from second day 5th instar larvae using the MgCl2 precipitation method of Wolfersberger et al.(8Wolfersberger M. Luthy P. Mauer A. Parenti P. Sacchi V. Giordana B. Hanozet G. Comp. Biochem. Physiol. 1987; 86: 301-308Crossref Scopus (544) Google Scholar), except that 1 mM phenylmethylsulfonyl fluoride was included in the final suspension buffer. BBMVs were stored at −80°C until needed.Purification of M. sexta 115-kDa ProteinBBMVs (10 mg of protein) were solubilized at 4°C for 30 min in 2 ml of Buffer A (20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride) containing 1% CHAPS (v/v). Undissolved material was removed by centrifugation at 100,000 × g for 1 h at 4°C. CHAPS-solubilized BBMVs were chromatographed in Buffer A plus 0.2% CHAPS on a Sephacryl 300 column (Pharmacia). Fractions were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (33Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (206016) Google Scholar) and ligand blotting. Fractions containing the 115-kDa toxin-binding protein were dialyzed against Buffer B (20 mM Tris-HCl, pH 7.4, 0.2% CHAPS) at 4°C overnight, loaded onto a Mono Q column pre-equilibrated with Buffer B, then the bound proteins eluted with 0-600 mM NaCl in Buffer B.Ligand Blot AnalysesBBMVs or purified proteins separated by SDS-PAGE were electrophoretically transferred to nitrocellulose filters. Filters were blocked with 5% (w/v) dry milk in TBS-T (20 mM Tris-HCl, pH 8.5, 150 mM NaCl, 0.1% Tween-20) for 1 h, followed by three washes with TBS-T. Filters were bathed with 125I-CryIA(c) (108 counts/min/10 ml TBS-T) for 3 h, washed three times in TBS-T, blotted dry, and exposed to x-ray film at −80°C.Instrumentation and ReagentsThe surface plasmon resonance detector (BIAcore™) system and CM5 sensor chips were obtained from Pharmacia Biosensor. All chemical immobilizations of the M. sexta 115-kDa receptor were done using the standard BIAcore amine coupling protocol provided with the Pharmacia coupling kit. Bovine serum albumin (fraction 5, radioimmunoassay grade) was purchased from Sigma. The buffers used with the BIAcore machine contained the following: HBS-P20 (HBS containing 0.05% BIAcore Surfactant P20), regeneration buffer (50 mM CAPS, pH 11.0, and 150 mM NaCl), carboxymethylated dextran activation solutions (A = 0.1 MN-hydroxysuccinimide, B = 0.1 MN-ethyl-N'-(3-diethylaminopropyl)carbodiimide), coupling buffer (20 mM ammonium acetate, pH 4.5), and deactivation solution (1 M ethanolamine, pH 8.5). Approximately 700-1500 RUs of the 115-kDa binding protein (dissolved as a 0.1 mg/ml stock solution in 20 mM ammonium acetate, pH 5) was amine-coupled to the dextran. A reagent flow rate of 5 μl/min was used during all experiments except for dissociation rates which used 15 μl/min.Binding AnalysesAll sensorgram data transformations and analyses were performed with BIAevaluation software version 2.1. using non-linear least-squares curve fitting. This method of kinetic binding analysis permits the determination of both the dissociation rate constant (kd) and the association rate constant (ka) for each binding experiment(26O'Shannessy D.J. Brigham-Burke M. Soneson K.K. Hensley P. Brooks I. Anal. Biochem. 1993; 212: 457-468Crossref PubMed Scopus (513) Google Scholar, 34O'Shannessy D.J. Curr. Opin. Biotechnol. 1994; 5: 65-71Crossref PubMed Scopus (131) Google Scholar). For all kinetic toxin binding experiments, kd was determined first by plotting the log of the drop in response, ln(R0/R), against the time interval (t), where R0 is the response at an arbitrarily chosen starting time soon after toxin flow is replaced by buffer, and Rand t are data points chosen every 0.5 s along the dissociation curve. The apparent kd value was used to constrain values when determining the ka constant. The maximum binding levels for each toxin was determined on surfaces containing approximately 700-1500 RUs (representing a surface of approximately 6.1-13.1 fmol/mm2) of immobilized 115-kDa receptor. Since the molecular mass of a protein is proportional to the SPR response (35Stenberg E. Persson B. Roos H. Urbaniczky C. J. Coll. Int. Sci. 1991; 143: 513-526Crossref Scopus (1002) Google Scholar), an immobilized surface, representing 1000 RU of the 115-kDa binding protein, could theoretically bind 565 RUs of a 65-kDa toxin analyte (A) in a simple one binding site model (A+B ⇔ AB) on ligand (B). To prevent any influence by toxin rebinding, the immobilized ligand was saturated with analyte and the start point arbitrarily chosen 5 s after the start of buffer flow where the ln(R0/R) against (t) curve initially becomes linear.To determine ka, eight different toxin concentrations ranging from 50 to 1500 nM were injected over immobilized APN. As with the kd determination, all data were fitted to either a one-site (A+B ⇔ AB) or a two-site model (A+B1+B2 ⇔ AB1+AB2). Whenever possible, fitted lines from either model were compared to each other using an F-test comparison. All models were verified by residual plots which calculate the difference between the observed and the fitted curves for each data point. To determine the goodness of fit of the data to the model, both association and dissociation rate constants were chosen from sensorgrams producing a chi2 value less than one. All models used in our kinetic evaluations were further evaluated by lag plots to determine the relationship of neighboring data points thus providing additional support for binding model selection (data not shown).Competition StudiesTwo types of competition studies were performed. For carbohydrate inhibition experiments, toxins were diluted in HBS-P20 buffer containing varying concentrations of either N-acetyl-D-galactosamine, N-acetyl-D-glucosamine (ICN Biomedicals, Aurora, OH), D-glucose (BDH, Inc., Toronto, Ontario), or D-mannose (Sigma) to a final concentration of 150 nM and injected over an immobilized receptor surface (approximately 1000 RUs). The highest response obtained for each sugar concentration injected over the receptor surface in the absence of toxin was subtracted from the highest RU value obtained in the presence of toxin and the results expressed as percent inhibition. For binding site competition experiments, saturating levels of toxin pairs were injected individually or co-injected over a surface of immobilized receptor.RESULTSPurification of 115-kDa CryIA(c)-binding APNThe 120-kDa APN that binds CryIA(c) toxin is cleaved to a 115 kDa form by treatment with PIPLC(36Garczynski S.F. Adang M.J. Insect Biochem. Mol. Biol. 1995; 25: 409-415Crossref Scopus (71) Google Scholar). The same 120-115 kDa conversion occurs in solubilized preparations of M. sexta BBMVs due to an endogenous PIPLC. 2Lu, Y. J., and Adang, M.(1995) Biochem. Mol. Biol., in press. The 115-kDa APN form has the same N terminus, lacks the lipid moiety on the glycosylphosphatidyl inositol membrane anchor, and still binds CryIA(c) toxin. M. sexta BBMVs were solubilized in 1% CHAPS then fractionated by S-300 gel filtration in the presence of 0.2% CHAPS. Protein blots were probed with 125I-CryIA(c) to identify fractions containing toxin-binding 115-kDa APN. Toxin-binding 115-kDa protein was further purified by a Mono Q chromatography step. Fig. 1shows a stained gel and ligand blot of CHAPS-solubilized BBMVs and purified 115-kDa protein.Stoichiometry and Kinetic AnalysesIncreasing concentrations of the different CryI toxin subclasses were injected over a low density surface of immobilized 115-kDa receptor to determine levels required for ligand saturation. As summarized in Fig. 2, CryIA(c) bound to almost a 2-fold higher level than either CryIA(a) or CryIA(b). Considering that the molecular masses of all three activated toxins are essentially the same (~65 kDa) and that the SPR response corresponds linearly to the surface protein concentration (35Stenberg E. Persson B. Roos H. Urbaniczky C. J. Coll. Int. Sci. 1991; 143: 513-526Crossref Scopus (1002) Google Scholar) these results indicate that the immobilized 115-kDa protein possesses a single binding site for either CryIA(a) or CryIA(b) but that CryIA(c) binds to two sites on the same molecule. Injection of a similar concentration of CryIC resulted in little or no specific binding demonstrating the specificity of the immobilized ligand for the three CryIA toxins.Figure 2:Stoichiometric analysis of CryIA binding. Saturating levels of CryIC, CryIA(a), or CryIA(b) (1500 nM) and of CryIA(c) (1000 nM) were injected at 5 μl/min over a surface of immobilized 115-kDa CryIA(c)-binding protein representing approximately 1000 RU. At the end of the injection, toxin flow was replaced by buffer alone and the sensorgram allowed to continue for an additional 100 s to demonstrate the rate of complex dissociation.View Large Image Figure ViewerDownload Hi-res image Download (PPT)To confirm the number of toxin-binding sites and determine the kinetic rate constants, different toxin concentrations were injected over a surface of immobilized 115-kDa ligand. The binding data from each curve were fitted by non-linear least-squares fitting to either a one- or two-site model. The model fitting further suggested that CryIA(a) and CryIA(b) bound to a single site on the immobilized ligand whereas CryIA(c) best fit a two-site model (F-test comparison between the two models gave a probability = 1). The possibility that the second observed CryIA(c) dissociation rate was caused by the rebinding of toxin to immobilized aminopeptidase was eliminated since a plot of the log of the drop in response against time interval produced a linear rather than a curved response, a normal indicator of rebinding. 3L. Masson, unpublished observations. Residual plots (i.e. a plot of the difference between the observed and the calculated response for each data point) of the dissociation segment of the response curves were created to verify the appropriateness of the binding model chosen. As shown in Fig. 3, A and B, the data point distribution is random around the x axis and the signal noise is no greater than background (±2 RU) thus indicating that the quality of fit was good for CryIA(a) and CryIA(b) to a single binding site model. The χ2 values, testing for the goodness-of-fit, for all the sensorgrams from the different toxin concentrations were 1.0. If the data are fitted to a model which accounts for two separate toxin-binding sites on the ligand, the point distribution becomes random indicating a good fit. Furthermore, all χ2 values obtained from the different sensorgrams fall below 1.0 providing additional support for the two-site hypothesis. Residual plots performed on data from the association area of the sensorgram were similar to those shown above for the dissociation segments (data not shown).Figure 3:Dissociation rate residual plots for CryIA(a) and CryIA(b). Residual plots, representing the randomness of data point distribution around a fitted curve, are shown for a typical binding data set taken 40 s after the start of complex dissociation for CryIA(a) (A) or CryIA(b) (B) when applied to a one-site model (A+B ⇔ AB). In the Response plot, the actual dissociation data points are represented by a solid line and the fitted curve by a dashed line. In the Residual plot the response differences (residuals) in RU of the fitted line around the dissociation data are represented by solid dots. A zero difference reference line was added to help visulize the randomness of point distribution.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 4:Dissociation rate residual plots for CryIA(c). A residual plot of a typical CryIA(c) binding data set applied to a one-site model (A+B ⇔ AB) is shown in panel A, and applied to a two-site model (A+B1+B2 ⇔ AB1+AB2) shown in panel B. As indicated in Fig. 3, the data set was taken 40 s after the start of complex dissociation. In the Response plot, the actual dissociation data points are represented by a solid line and the fitted curve by a dashed line. In the Residual plot the response differences (residuals) in RU of the fitted line around the dissociation data are represented by solid dots. A zero difference reference line was added to help visulize the randomness of point distribution.View Large Image Figure ViewerDownload Hi-res image Download (PPT)A summary of the apparent rate constants is shown in Table 1. Each toxin studied displayed a moderately fast association rate showing at most a 5-fold variation. The most interesting differences were found in the various dissociation rates. In general, all the E. coli produced NRD-12 toxins showed similar kd values; however, the second CryIA(c)-binding site demonstrated a much faster rate (≥ an order of magnitude) of toxin-receptor complex dissociation than that calculated for the other CryIA toxins. Furthermore, the B. thuringiensis produced CryIA(c) toxin, although showing ka rates indistinguishable from the E. coli produced CryIA(c), demonstrated a 2-fold faster dissociation rate for both sites than the E. coli produced toxin. Despite the observed variations in kinetic rates, the three CryIA toxins from NRD-12 essentially share the same affinity for the immobilized aminopeptidase with CryIA(c) also binding to a second site at a lower affinity.Tabled 1 Open table in a new tab Competition AnalysesIn order to verify whether the multiple binding sites were unique or shared among the three CryIA toxins, saturating levels of toxins were paired together and co-injected over a surface of immobilized ligand. If the toxins bind to separate sites, an additive effect should be observed when comparing the maximal binding response (Rmax) to single toxin injections. As shown in Fig. 5A, pairing of the CryIA(a) and CryIA(b) toxins produced a maximal binding level (Rmax) similar to either of the individual injections. This result agrees with the finding of van Rie et al.(3van Rie J. Jansens S. Höfte H. Degheele D. van Mellaert H. Eur. J. Biochem. 1989; 186: 239-247Crossref PubMed Scopus (280) Google Scholar) using BBMVs from M. sexta that these two toxins share the same site. When the CryIA(b) toxin is paired with the CryIA(c) toxin, an additive effect is again absent. The maximal binding level obtained with the co-injected toxins was similar to CryIA(c) alone suggesting that one of the two CryIA(c)-binding sites is shared with CryIA(b) and by extension, CryIA(a). Experiments using consecutive rather than simultaneous toxin injections were attempted, but the ability of toxins to bind to each other at high concentrations (28Masson L. Mazza A. Brousseau R. Tabashnik B. J. Biol. Chem. 1995; 270
Referência(s)