Insulin-binding Peptide DESIGN AND CHARACTERIZATION*

The design and characterization of a six-amino acid-containing peptide that binds insulin is described. The amino acid sequence of the insulin-binding peptide (IBP) was determined from the strand of DNA complementary to the strand of DNA coding for the insulin molecule in the domain of the insulin monomer believed to interact with the insulin receptor. The IBP (Cys-Val-Glu-Glu-Ala-Ser) binds specifically to insulin in a saturable manner with a Kd of 3 nM. This binding process is time dependent and slightly temperature dependent, and the peptide appears to interact with insulin near the carboxyl terminus of the B-chain of insulin. Incubation of insulin with the peptide de-creases insulin binding to the insulin receptor by 50%, with no effect on the affinity of insulin for the receptor and no effect on cellular insulin-stimulated deoxyglucose uptake. A polyelonal antibody produced against the IBP will inhibit specific insulin binding to intact cells by approximately 50%, with no effects on insulin-stimulated glucose uptake. From this data, we suggest that there are at least two domains of the insulin molecule through which it interacts with its receptor, the ”binding region” of insulin, which is the domain blocked by the IBP, and the “message region” of insulin, through which insulin not only binds to the receptor, but also generates the cellular signal. The amino acid sequence of this peptide was determined by the nucleotide sequence of the strand of DNA complementary to the strand of DNA coding for the ACTH (1-24) molecule. The results of indicated that this “complementary”


Victoria P. Knutson
From the Department of Phurmacobgy, The University of Texas Medical School at Houston, Houston, Texas 77225 The design and characterization of a six-amino acidcontaining peptide that binds insulin is described. The amino acid sequence of the insulin-binding peptide (IBP) was determined from the strand of DNA complementary to the strand of DNA coding for the insulin molecule in the domain of the insulin monomer believed to interact with the insulin receptor. The IBP (Cys-Val-Glu-Glu-Ala-Ser) binds specifically to insulin in a saturable manner with a Kd of 3 nM. This binding process is time dependent and slightly temperature dependent, and the peptide appears to interact with insulin near the carboxyl terminus of the B-chain of insulin. Incubation of insulin with the peptide decreases insulin binding to the insulin receptor by 50%, with no effect on the affinity of insulin for the receptor and no effect on cellular insulin-stimulated deoxyglucose uptake. A polyelonal antibody produced against the IBP will inhibit specific insulin binding to intact cells by approximately 50%, with no effects on insulinstimulated glucose uptake.
From this data, we suggest that there are at least two domains of the insulin molecule through which it interacts with its receptor, the "binding region" of insulin, which is the domain blocked by the IBP, and the "message region" of insulin, through which insulin not only binds to the receptor, but also generates the cellular signal.
A study was recently published in which a peptide was designed to mimic the ACTH receptor (1). The amino acid sequence of this peptide was determined by the nucleotide sequence of the strand of DNA complementary to the strand of DNA coding for the ACTH (1-24) molecule. The results of this study indicated that this "complementary" peptide bound to ACTH, and antibody generated against the complementary peptide appeared to interact with the ACTH receptor. In an attempt to generate an antibody against the insulin-binding domain of the insulin receptor, we decided to devise and synthesize a peptide according to the method described above (1). However, instead of synthesizing a peptide complementary to the whole insulin molecule, we restricted our search to only those residues of the insulin monomer which 1) are in a domain reported to interact with the insulin receptor (2-4) and 2) are in a linear sequence in the primary amino acid sequence of the insulin molecule. This report describes the design and characterization of this insulin-binding peptide. *This work was supported by Grants AM35397 and AM27685 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

MATERIALS AND METHODS
Design of the Insulin-binding Peptide-The various domains of the insulin molecule have been described by Blundell and Wood (Z), and others (3,4). Within the reputed receptor-binding domain is a linear string of 6 amino acid residues found near the carboxyl terminus of the insulin B-chain, in particular, B-chain residues 22-27. The sequence of the gene coding for the insulin molecule has been reported (5), and the sequence of the complementary strand of DNA (reading from the 5' to the 3' end) is shown in Fig. 1. Reading this strand of DNA from the 3' to the 5' direction (as would the RNA polymerase) would result in the generation of the messenger RNA species shown in Fig. 1, which codes for the polypeptide Cys-Val-Glu-Glu-Ala-Ser. This peptide was synthesized by OCS Laboratories, Inc. (Denton, TX).
Binding of Insulin to the Peptide-The IBP' was immobilized in the wells of flexible, U-bottom 96-well polyvinyl chloride microtiter plates (Falcon, Becton Dickinson Labware). To each well was added 100 ~1 of peptide, 5 pg/ml in 0.1 M sodium carbonate, pH 9.6, and the peptide was allowed to adsorb to the well overnight at 4 "C. Unreacted sites in the wells were blocked with 150 pl of 1% (w/v) ovalbumin (Sigma) in 0.1 M sodium carbonate, pH 9.6, for 1 h at room temperature, The wells were washed two times with PBS containing 0.1% bovine serum albumin (Fraction V, Sigma) and then incubated with '"1-insulin in PBS, 0.1% bovine serum albumin for 3 h at room temperature. The wells were washed three times with PBS, 0.1% bovine serum albumin and were then excised from the plate and counted in a gamma counter. Total binding of insulin to the IBP was determined with iodinated insulin alone. Nonspecific binding was determined in the presence of radiolabeled insulin plus 3 PM unlabeled insulin. Specific binding was calculated as the difference between total and nonspecific binding. All values are reported as specific binding, unless otherwise indicated. All determinations (both total and nonspecific) were performed in triplicate.
Insulin was iodinated by the method of chloramine T, as previously described (6).
In some experiments, the immobilized IBP was incubated with 10 mM N-ethylmaleimide (NEM, Aldrich) before the addition of 1251insulin to the wells.
The specificity of the binding of IBP to insulin was determined by incubating iodinated calcitonin, parathyroid hormone, or atrial naturiuretic peptide (kindly supplied by Dr. D. Scott Linthicum, University of Texas Medical School at Houston) with immobilized IBP. Further specificity was investigated by coincubating immobilized IBP with l2'I-insu1in and unlabeled proinsulin, desoctapeptide insulin (gifts of Dr. R. Chance, Lilly), insulin A-chain or insulin B-chain (Sigma). The specificity of the binding of iodinated insulin to immobilized peptides was determined by immobilizing peptides of various sizes in the microtiter wells, as described above. The peptides utilized were provided by Dr. D. Scott Linthicum, and have the following designations and amino acid sequences: S28, TTHYQKG; S79, AQGHRPQDEG; S81, YGSLPQKAQGHRPQDEG; S82, TH-YGSLPQKAQGHRPQDEG; EAE, FSWGAEGQR.
Binding data were analyzed and K d values determined through the use of the EBDA and LIGAND computer programs described by Munson and Rodbard (7).
The Generation and Characterization of Anti-IBP Antibodies-Three mg of hemocyanin (Behring Diagnostics) and 10 mg of IBP were dissolved in a total volume of 0.9 ml of PBS. After the addition The abbreviations used are: IBP, insulin-binding peptide; NEM, N-ethylmaleimide; PBS, phosphate-buffered saline. I N S U L I N : A r g -Cly -Phe -Phe -T y r -T h r cDNA: CGA ( 3 1 -5 ' ) cDNA: ACA of 100 pl of 0.5% glutaraldehyde, the solution was stirred overnight at room temperature. The mixture was then diluted to a total volume of 3 ml with PBS, resulting in a stock solution which was 1 mg/ml in carrier protein. The initial immunization of the rabbits was performed with complete Freund's adjuvant, containing 250 pg of immunogen (250 p1 of stock), 250 p1 of PBS, and 500 p1 of complete Freund's adjuvant. Each rabbit was injected subcutaneously on the back, at five separate sites, 200 p1 injected at each site. One month after the initial injection, the animal was boosted with incomplete adjuvant. Two weeks after the boost, the rabbits were bled from the ear vein to determine the titer of antibody against IBP. Boosting of the animals was continued at monthly intervals, bleeding the animals at 2-week intervals after each boost. Every third month, the animals were boosted with complete Freund's adjuvant.
Initial characterization of the antiserum was performed by assessing its ability to interact with the IBP. The peptide was immobilized onto nitrocellulose (0.45 micron, Bio-Rad) that was sandwiched in a 96-well "Biodot" vacuum apparatus (Bio-Rad) by adding 50 pl of IBP, 5 pg/ml in 0.1 M sodium carbonate buffer, pH 9.6, to each well of the apparatus. The peptide was allowed to adsorb to the nitrocellulose in the absence of an externally applied vacuum for 1 h. A vacuum was then gently applied to the Biodot apparatus in order to pull the remaining liquid through the nitrocellulose. The side of the membrane to which the peptide was applied was then marked, and the membrane was air-dried overnight. Unreacted sites on the membrane were blocked with "blot buffer" (50 mM Tris, 0.5 M NaCl, pH 7.5) containing 3% gelatin (Bio-Rad), by incubating the membrane in 100 ml of the buffer for 1 h at room temperature with constant agitation. The membrane was then blotted dry with filter paper before incubation with antibody. Dilutions of antisera, preimmune antisera, or normal rabbit IgG were made in blot buffer containing 1% gelatin. Replicate 100-pl aliquots of the antisera dilutions were added to the wells of polystyrene flat-bottomed 96-well microtiter plates, and the nitrocellulose membrane was positioned over the microtiter plate so as to align the immobilized IBP and the wells of the plate. A thin compressible plastic gasket was placed on top of the membrane, and the whole assembly was sandwiched between glass plates and clamped together. The sandwich was inverted, and the antiserum was allowed to incubate with the immobilized IBP for 2 h at room temperature. Following incubation, the sandwich was inverted to allow the antisera to flow back into the wells of the microtiter plate. The nitrocellulose membrane was then quickly washed twice with blot buffer containing 0.25% gelatin and 0.05% (w/v) Tween 20 (Sigma). The membrane was then incubated with 15 ml of lZ51-protein A (Amersham Corp.), 1.5 pCi/l5 ml, in blot buffer containing 1% gelatin, for 1 h at room temperature, with constant agitation. The nitrocellulose was quickly washed four times with blot buffer containing 0.25% gelatin and 0.05% Tween 20 and then washed two times with blot buffer alone. The membrane was blotted dry and subjected to autoradiography or the dots were cut from the sheet of nitrocellulose and counted in a gamma counter.
The effect of IBP or antisera on insulin binding to the receptor and insulin action were determined with confluent monolayers of 3T3-Ll adipocytes. These cells were seeded onto 24-well tissue culture dishes (Falcon, Becton Dickinson Labware) and induced to differentiate into adipocytes as previously described (8). The effect of IBP or antisera on insulin binding was determined by the coaddition of lZ5Iinsulin and IBP or antisera to the cell monolayers in the presence or absence of an excess of unlabeled insulin. After incubation for 18 h at 4 "C, the cells were washed with cold PBS to remove unbound insulin, the monolayers were solubilized in 1 N NaOH and counted to determine the level of specifically bound insulin. The ability of IBP or antisera to modify insulin action was determined through the cellular uptake of the glucose analog 2-deoxyglucose. Intact monolayers of 3T3-Ll adipocytes were washed with glucose-free Kreb's-Ringer phosphate buffer, and the cells were incubated in the presence or absence of antisera for 20 min at 37 "C. Following an additional 20-min incubation in the presence or absence of insulin, or insulin plus IBP, an aliquot of [3H]2-deoxyglucose (ICN Pharmaceuticals, Inc.) was added to the monolayers, and the uptake of radiolabel was assessed, as previously described (9).
All other reagents were of analytical grade or better and were obtained from common supply houses.

RESULTS
The time course and the temperature dependence of the binding of insulin to immobilized IBP is shown in Fig. 2. At 25 "C, the binding attains an equilibrium level within 1.5 h, with a half-time of approximately 20 min. At 4 " C , the binding process is slightly slower, requiring approximately 2 h to attain an equilibrium level of binding.
As demonstrated in Fig. 3A, the specific binding of insulin to IBP is a concentration-dependent, saturable process. Plotting of the binding isotherm data according to the method of Scatchard (lo), as in Fig 3B, generates a linear plot, with a K d of 3 nM.
We had some concern that this high affinity interaction between IBP and insulin was a result of thiol exchange between the cysteine sulfhydral group of IBP and a thiol group on insulin. To eliminate this possiblity, IBP was immobilized in the microtiter wells and then incubated with 10 mM NEM to alkylate the sulfhydral group of IBP. Blocking with ovalbumin and incubation with '251-insulin was performed as described previously. No change in the binding of insulin to IBP was observed as a result of incubation with NEM (data not shown). A number of variations of this experimental format were also performed IBP was dissolved in a buffer containing NEM, such that alkylation was achieved before immobilization of the peptide on plastic; NEM was incubated with IBP after the peptide was immobilized, but the alkylating agent was maintained in the buffers throughout the remainder of the experiment, including the incubation with '251-insulin. The results of these various protocols were the same. Sulfhydral alkylation of IBP has no effect on the binding of insulin to the peptide. The time and temperature dependence of insulin binding to IBP. The peptide was immobilized in the bottom of 96well microtiter plates, and at zero time, 2 nM '251-insulin, in the absence or presence of 3 p~ unlabeled insulin, was added to each of the wells. At the indicated times, the wells were washed with cold PBS, the wells were cut from the dish and counted, as described under "Materials and Methods." The indicated data points are average values of specifically bound insulin, and the error burs indicate the range of the data. The binding of insulin to IBP appears to be a specific process. As shown in Fig. 4, the specific binding of the low molecular weight polypeptide hormones calcitonin, parathyroid hormone, and atrial naturiuretic peptide is less than 1% of the level of the specific binding of insulin to IBP. The binding of insulin to various immobilized peptides, as shown in Fig. 5, demonstrates the specificity of the interaction between IBP and insulin.
In theory, IBP should be interacting with residues 22-27 of the insulin B-chain. If this is occurring, insulin analogs or subunits which contain these residues should compete with intact 1Z51-insulin for binding to IBP. That is, insulin B-chain, proinsulin, and unlabeled insulin should compete with lZ5Iinsulin for binding to immobilized IBP. Insulin A-chain and desoctapeptide insulin (which lacks the carboxyl-terminal eight amino acids of the B-chain), should not inhibit the binding of 1251-insulin to IBP. As shown in Fig. 6, proinsulin, insulin B-chain, and unlabeled insulin all compete equally effectively with '251-insulin for binding to IBP. The IC5o is approximately 50 nM. Desoctapeptide insulin and insulin Achain had no effect on '251-insulin binding to IBP. These results are consistent with the interpretation that IBP binds FIG. 6. The competition of unlabeled insulin and insulin analogs with 'asI-insulin for binding to immobilized IBP. IBP was immobilized in microtiter wells, as described under "Materials and Methods." To the wells was simultaneously added 2 nM '=Iinsulin and the indicated concentration of unlabeled insulin analog. After incubation at room temperature for 3 h, the wells were washed extensively of unbound radiolabel, cut from the plate, and counted in a gamma counter. The y axis represents the total amount of '=Iinsulin bound to the wells, with no correction for nonspecific binding. Each symbol is the average of triplicate determinations. 0, insulin; 0, proinsulin; A, insulin B-chain; A, IBP; W, desoctapeptide (DOP) insulin; 0, insulin A-chain.
time-and temperature-dependent interaction, with a high degree of specificity.
The data presented above address the interaction of insulin with immobilized "solid-phase" IBP. The interaction of insulin with solid-phase IBP might be expected to be different from the interaction of insulin with IBP that is free in solution. In an attempt to elucidate the relative affinity of insulin for "solution-phase" IBP compared to solid-phase IBP, '251-insulin was preincubated with various concentrations of soluble IBP before the addition of the 1251-insulin-IBP mixture to immobilized IBP. In theory, the solutionphase IBP would bind to the '251-insulin and prevent the subsequent binding of ''9-insulin to immobilized IBP. Solution-phase IBP is competing with solid-phase IBP for binding to 1251-insulin. The results of this experiment are shown in Fig. 6. Solution-phase IBP inhibits the binding of 1251-insulin to solid-phase IBP with an ICSo of 5 nM. This result indicates that the solution-phase affinity of IBP for insulin is similar to the solid-phase affinity of IBP for insulin.
The above data indicate that insulin and IBP interact with a high affinity. Since IBP was designed to act as a receptor analog, the next question to be addressed pertained to the ability of IBP to inhibit the binding of insulin to bona fide insulin receptor. In the experiment shown in Fig. 7, 1251insulin was incubated with intact 3T3-Ll adipocytes in the presence of increasing concentrations of either unlabeled insulin or unlabeled IBP. In the experiment shown in Fig. 7, unlabeled insulin effectively competed with 1261-insulin for binding to the cellular insulin receptor, but IBP inhibited insulin binding by approximately 30%. Repetition of this experiment demonstrated that IBP could never inhibit insulin binding by more than 50%. Maximal inhibition reproducibly occurred at approximately 20 nM IBP.
In order to determine if the inhibition of insulin binding to receptor induced by IBP, as shown in Fig. 7, is due to change in the affinity of insulin for the receptor or a change in the number of insulin molecules capable of binding to the recep- tor, insulin-binding isotherms were constructed in the absence or presence of a constant, high (250 nM) concentration of IBP. At this concentration of IBP, the peptide is in large excess over insulin, and based on the solution phase interaction between IBP and insulin shown in Fig. 6, the binding sites for IBP on insulin should be saturated. The binding isotherms are shown in Fig. 8. Plotting of the data of Fig. 8 according to the method of Scatchard (lo), and analysis of the data with the EBDA/LIGAND programs described by Rodbard and Munson (7), demonstrated no significant change in the high affinity Kd due to the presence of IBP, but a 2fold decrease in the number of high affinity binding sites as a result of IBP incubation (without IBP: & = 4.3 nM, BmaX = 0.46 nM; with I B P Kd = 3.1 nM, BmaX = 0.23 nM). Therefore, IBP did not appear to modify the affinity of insulin for receptor but instead inhibited the binding of 50% of the available insulin molecules.
The effect of IBP on insulin action was determined by quantitating 2-deoxyglucose uptake in the hormonally responsive 3T3-Ll adipocytes. Insulin concentrations ranging from 0 to 25 nM were incubated in the presence or absence of 250 nM IBP for 3 h to optimize the interaction between peptide and insulin. These solutions were then added to 3T3-Ll cells, and the assay was continued as described above. No effect of the peptide could be detected on either basal (without insulin) uptake, or insulin-stimulated uptake (data not shown), under conditions where 50% inhibition of insulin binding could be demonstrated.
The injection of hemocyanin-coupled IBP into rabbits resulted in the generation of antiserum that reacts with immobilized IBP. As shown in Fig. 9, antiserum from early bleed animals demonstrated a signal at dilutions as great as 1:400, compared to preimmune serum. Since this "dot-blot" assay is based on the binding of the antibody to protein A, the antibody would appear to be of an IgG class.
The effect of anti-IBP antibody on the binding of insulin to the insulin receptor in intact cell monolayers is shown in Fig. 10. Increasing concentrations of antiserum result in in-Dilutions I:I 00 It200 I :40 0 b800 1:1600 1~3200 It6400 FIG. 9. The detection of antibody against the IBP. The insulin-binding peptide was coupled to hemocyanin and injected into rabbits, as described. The antibody generated in the rabbits was titered by incubating antiserum (Immune) or preimmune serum (Preimrnune), at the indicated dilutions, with IBP immobilized on nitrocellulose. The antibody binding to the nitrocellulose was probed with '*'I-protein A, as described under "Materials and Methods," and the "dot-blot'' was subjected to autoradiography. Following an additional incubation with both antiserum and insulin, tritiated 2-deoxyglucose was added to the monolayers, and uptake was allowed to proceed for 20 min. The time-dependent uptake of radiolabel is linear over this 20-min period (data not shown). Uptake is terminated by the addition of ice-cold Kreb's-Ringer phosphate buffer containing 25 mM glucose, and cell-associated radioactivity is quantitated as previously described (6).
creasing inhibition of specific insulin binding to cells, such that at a 1:25 dilution of antiserum, insulin binding is inhibited by approximately 50%. Even utilizing undiluted serum, no more than 50% inhibition of insulin binding to intact cells could be demonstrated (data not shown). Corresponding dilutions of preimmune serum have no effect on insulin binding to the cells (data not shown). This antiserum was unable to immunoadsorb '2sI-insulin (data not shown). The effect of anti-IBP antiserum on cellular basal and insulin-stimulated deoxyglucose uptake is demonstrated in Fig. 11. At a 150 dilution of antiserum, where insulin binding to the cells is inhibited by 45%, no effect on either basal or insulin-stimulated uptake could be detected. That is, the antiserum itself did not stimulate deoxyglucose uptake nor did the antiserum inhibit the insulin-stimulated uptake of radiolabel. Treatment of the cells with preimmune sera or normal rabbit IgG was also without effect on cellular deoxyglucose uptake (data not shown).
Attempts to immunoadsorb insulin receptor out of a detergent extract of metabolically labeled cells with the anti-IBP antiserum were unsuccessful. No specific bands of radiolabeled protein could be detected by sodium dodecyl sulfatepolyacrylamide gel electrophoresis and autoradiography when comparing immune uersus preimmune antiserum. These experiments were performed under conditions which result in the clean and quantitative immunoadsorption of radiolabeled insulin receptor with an antibody directed against the j 3 subunit of the insulin receptor, and an antibody directed against phosphorylated tyrosine residues (data not shown). In addition, Western blot analysis of the insulin receptor with the anti-IBP antiserum and '251-protein A did not reveal the presence of any immunoreactive bands.

DISCUSSION
The methodology utilized in the design of the insulinbinding peptide described in this report was initially outlined by Bost et al. (1). This technique was based on the initial observation of Blalock and Smith (11) that for a given codon coding for a hydrophilic amino acid, the corresponding codon on the complementary strand of DNA will code for a hydro-phobic amino acid. In this way, two peptides synthesized from complementary strands of DNA would be "negative images" of each other, when the amino terminus of one peptide is lined up with the carboxyl terminus of the other peptide; hydrophobic regions of one peptide would be aligned with hydrophilic regions of the second peptide. Bost et al. (1) speculated that this degree of complementarity at the level of the amino acids should lead to binding interactions between the two peptides. For peptides long enough to have secondary and tertiary structure, one could envision how this might occur. A string of hydrophobic residues on one peptide might form a hydrophobic pocket. The complementary peptide would contain a string of hydrophilic residues in the corresponding region, which could form a solvent-accessible protrusion. The hydrophobic pocket might then interact with the hydrophilic region. While this scenario is highly speculative for describing the interaction of large peptides, it could clearly not be the case for describing the interaction between insulin and the insulin-binding peptide described in this report. The insulin-binding peptide is only 6 amino acid residues long. A single turn of an a helix requires approximately 3.6 residues. While the valine, glutamic acid, and alanine residues, which comprise 66% of IBP, are all believed to strongly participate in the formation of a helix (12), the peptide could, at most, generate a single helical turn in solution. Therefore, while IBP could have a small degree of secondary structure, there is no potential for tertiary structure. The question remains, then, as to the nature of the interaction between IBP and insulin. The solution-phase conformation of IBP and the nature of its interaction with insulin is currently under investigation with spectroscopic techniques. In addition, the interaction of IBP with synthetic fragments of insulin B-chain are also under investigation to ascertain the degree to which the intact insulin molecule stabilizes the interaction between insulin and IBP.
When immobilized in the bottom of a microtiter well, IBP interacts with insulin with a high affinity and an apparent high degree of specificity, an interaction characteristic of the binding of insulin to the insulin receptor. In addition, in order for this interaction to occur, the insulin derivative was required to contain that region of the molecule against which the peptide was designed (i.e. B-chain sequences 22-27). As such, neither insulin A-chain nor desoctapeptide insulin would inhibit insulin binding to IBP (Fig. 6). According to the hypothesis of Bost et al. (l), IBP should be considered a receptor "mimic." In many respects, this is definitely not the case. Insulin binding to bona fide insulin receptor is not inhibited by insulin B-chain or proinsulin, in contrast to the situation with insulin binding to IBP. However, the binding of insulin to IBP is remarkably high in affinity and specificity. A search for sequence homology between IBP and the insulin receptor was therefore performed to confirm any regions of identity. This search identified two regions, each with a 66% homology (4 of 6 residues). These homologous receptor sequences were found in the insulin-binding subunit, in the region postulated by Rutter et al. (13) to be the insulin-binding domain. We have synthesized one of these receptor fragments, and find that this receptor peptide also binds insulin with high affinity.' There is currently much controversy in the literature over the amino acid residues of the insulin molecule which interact with the receptor. Early work (3) indicated that the carboxyl * V. P. Knutson, unpublished results. terminus of the insulin B-chain was responsible for both binding activity and biological activity. More recent work (4) demonstrated that desoctapeptide insulin, which is lacking the terminal eight amino acids of the B-chain, would bind to the insulin receptor, but three orders of magnitude greater concentrations of ligand were required, compared to native insulin. It appeared clear, therefore, that this region of the insulin B-chain was involved in the binding interaction between insulin and the receptor, and that the binding domain map proposed by Blundell and Wood (2) might have some validity. The data presented here also support this region of the insulin molecule as important in binding to the receptor.
The finding that IBP would inhibit insulin binding to the receptor by 50% but have no effect on insulin action was consistent with the finding that the antiserum generated against the peptide would inhibit insulin binding by 50% but have no effect on glucose uptake. Regarding this discrepancy between insulin binding and insulin action, some data have been recently published indicating that insulin binding activity can be modulated independently of biological activity. Cutfield et aZ. (14) have demonstrated that the carboxyl terminus of insulin A-chain is involved in insulin bioactivity but not binding activity. More recently, in a series of papers (15-17), Chu et al. demonstrated that modification of AZ1asparagine led to a very slight reduction in binding activity but a drastic decrease in bioactivity. The conclusion from their studies was that the amide bond between residues 20 and 21 of insulin A-chain is part of the "message region" of the insulin molecule, and is quite distinct from the "binding region." Along these lines, then, IBP would appear to interact with the insulin molecule in a binding region, and not a message region. However, IBP and anti-IBP antiserum only inhibit 50% of the total insulin binding capacity. Therefore, another binding region may exist on insulin and on the receptor which has an affinity equal to the IBP-inhibitable binding region and is able to generate full biological response. Validation of this speculation awaits further study.