Binding of insulin to its receptor impairs recognition by monoclonal anti-insulin antibodies

The interaction between insulin and its receptor was investigated using both monoclonal and polyclonal anti-insulin antibodies. After covalent cross-linking of 125I-insulin to the insulin receptor on cultured human lymphocytes (IM-9 cells) using disuccinimidyl suberate, we inquired whether the insulin-receptor complex could be immunoprecipitated with anti-insulin antibodies. While a polyclonal guinea pig anti-insulin antiserum succeeded in immunoprecipitating receptor-bound 125I-insulin, binding to the receptor decreased the avidity of the antiserum for the insulin moiety by a factor of approximately 1000-fold. Sixteen distinct monoclonal murine anti-insulin antibodies were employed to immunoprecipitate receptor-bound 125I-insulin. Of these 16 monoclonal antibodies, only one (antibody 5.9F4) could be shown to recognize receptor-bound 125I-insulin. Moreover, even with antibody 5.9F4, binding of 125I-insulin to its receptor reduced the affinity of the antibody by a factor of 10- to 100-fold. These data strongly suggest that, when insulin binds to its receptor, the majority of the insulin molecule is unavailable for binding by anti-insulin antibodies. It seems likely that the hormone binding site on the receptor may be very large, thereby allowing for sequestration of the majority of the insulin molecule with relatively little of the hormone remaining exposed.     

Tissue-specific antibodies against the fibroblast insulin receptor in a patient with lupus nephritis and hypoglycemia

We recently reported that the serum from a patient with lupus nephritis, insulin resistance, and hypoglycemia contains multiple populations of antibodies directed at the human insulin receptor. In the present study, we found a subpopulation of antibodies (eluted from a protein A-Sepharose affinity column at pH 4.3) directed at the human fibroblast insulin receptor. When tested against human placental membranes, IM-9 lymphocytes, circulating monocytes and erythrocytes, and isolated adipocytes, the antibody subpopulation did not compete with 125I-insulin for binding to its receptor. In contrast, the antibody subpopulation competed with 125I-insulin for binding to the human fibroblast insulin receptor. This antibody subpopulation stimulated [3H]alpha-aminoisobutyric acid [( 3H]AIB) uptake to these cells. Unlike the effect of insulin, however, this regulation of transport was not antagonized by a mouse monoclonal antibody to the human insulin receptor that inhibits 125I-insulin binding. These studies indicate, therefore, that a tissue-specific antibody subpopulation can occur spontaneously in patients with antibodies to the human insulin receptor. Furthermore, they indicate the presence of anti-insulin receptor autoantibodies specifically directed against a tissue that is not primarily involved in glucose metabolism.     

An alteration in apparent molecular weight of the insulin receptor from the human monocyte cell line U-937

We have studied the structure of the insulin receptor from a human cultured monocyte cell line, U-937. The receptor is composed of alpha and beta subunits as seen in other insulin receptors, but these subunits are of greater apparent molecular weight (alpha 150,000 and beta 102,000) than in typical insulin receptors. Despite this, the U-937 insulin receptor appears to function normally. The alpha subunit binds insulin and the beta subunit is phosphorylated in response to insulin stimulation. Both subunits are expressed in the plasma membrane. Insulin binding isotherms are similar to those seen in IM-9 lymphocytes. Thus, the insulin receptor from U-937 monocytes appears functionally normal despite alterations in molecular weight of the subunits.     

The fate of insulin in rat hepatocytes. Evidence for the release of an immunologically active fragment

We have investigated the fate of 125I-insulin after binding by rat hepatocytes. Approximately 30% of the bound radioactivity dissociated from the cells as intact 125I-insulin; approximately 56% dissociated as 125I- and 125I-Tyr. The remaining radioactivity was recovered as peptides that we postulate are intermediate products of insulin metabolism. Experiments were performed in the presence of chloroquine (0.1 mM), an agent known to inhibit the intracellular processing of 125I-insulin. As expected, chloroquine increased the amount of radioactivity recovered as intact 125I-insulin (P less than 0.005) and decreased the amount of 125I- and 125I-Tyr (P less than 0.005). In addition, chloroquine decreased the amount of one of the insulin peptides (P less than 0.005), but increased the amount of the other (P less than 0.01). These data suggest the presence of two pathways of insulin metabolism in rat hepatocytes, one of which is inhibited by chloroquine. We have found a second pathway by which insulin is degraded due to the removal of several amino acids from the carboxy-terminus of the B-chain. The resulting fragment bound poorly to insulin receptors on IM-9 cultured human lymphocytes, and probably has little if any biologic activity. However, this fragment bound well to anti-insulin antibody and constituted about 20% of the immunoreactive radioactivity that dissociated from the hepatocytes.     

Processing of insulin by bovine endothelial cells in culture. Internalization without degradation

Insulin binding and processing was studied in monolayer cultures of bovine aortic endothelial cells. Specific 125I-insulin binding was both time and temperature dependent. Maximum binding at 37 degrees C occurred at 90 min, and was 3.8%/mg protein and, at 15 degrees C, 7%/mg protein at 4 h. 125I-insulin was crosslinked to its receptor using disuccinimidyl suberate (DSS), and the structure of the receptor complex was identified by SDS-polyacrylamide gel electrophoresis and autoradiography; a major band with Mr = 145,000 was identified, which corresponds to the alpha-subunit of the insulin receptor reported in other tissues. Receptor-bound insulin was internalized, and both the rate and the amount of internalization were temperature dependent. The rate of internalization was slowest at 4 degrees C, and fastest at 37 degrees C, and the maximum amount of 125I-insulin internalized in 120 min was 16% at 4 degrees C, 45% at 15 degrees C, and 81% at 37 degrees C. Despite the high rate of internalization, endothelial cells do not appear to degrade insulin significantly, as determined by gel chromatography and TCA solubility (7% at 4 h) of media-associated radioactivity. In addition, the majority of internalized insulin (75%) was released by 60 min, largely as intact insulin. Chloroquine treatment at high concentration did not exert any major effect on insulin binding or degradation within the first 60 min, but thereafter produced a marked increase in cell-associated radioactivity.(ABSTRACT TRUNCATED AT 250 WORDS)     

Simultaneous inhibition of insulin and somatomedin-C binding to cultured IM-9 lymphocytes by naturally occurring antireceptor antibodies

A 53-y4-old male patient with insulin-resistant diabetes was found to have circulating inhibitors of both insulin and somatomedin-C binding. Serum obtained from the patient at the time of initial presentation inhibited 50% of both 125I-insulin and 125I-SM-C binding to IM-9 lymphocytes at dilutions of 1:150. Spontaneous improvement in the patient's diabetic state was associated with a simultaneous and equal decrease in the serum inhibitory titers for both radioligands. Scatchard analysis indicated that the observed serum-induced decrease in both insulin and SM-C binding was due to decreased receptor affinity, with no alteration in receptor number. The serum inhibitors of both insulin and SM-C binding were precipitated equally by Staph-A and also by 40% ammonium sulfate, suggesting they were immunoglobulins. The observation of naturally occurring autoantibodies against both the insulin and SM-C receptors suggests a structural homology between the two receptors.     

New, simple insulin-receptor assay with universal application to solubilized insulin receptors and receptors in broken and intact cells

A new, simple insulin-receptor-binding assay has been devised. The assay is based on the separation of free and receptor-bound 125I-labeled insulin in 80% ethanol. It was found that the insulin-receptor complex was fully stable at this ethanol concentration, regardless of the source of the receptor employed. The assay has been evaluated with solubilized insulin receptors and membrane-bound receptors from human placenta and porcine liver as well as intact cells with the IM-9 cell line. The assay is simple, rapid, and has large capacity. Comparisons of the ethanol-based assay to the conventionally employed assays with polyethylene glycol or microfuge centrifugation for the separation of free and bound 125I-insulin revealed large discrepancies between the assays. The ethanol-based assay always appeared to provide a better separation. Microfuge centrifugation of placental membranes precipitated approximately 3% of the ethanol-precipitable insulin-receptor complex, while polyethylene glycol precipitation of solubilized insulin receptors varied between 40 and 80% of the ethanol precipitability, depending on the receptor concentration employed.     

Insulin receptors in cultured human fibroblasts.

In order to study human insulin resistance, we have first characterized the interaction of insulin with specific insulin receptors in cultures of normal human fibroblasts. 125 I-insulin bound rapidly to human fibroblasts in suspension at 15 degrees, achieving steady state between one and three hours. Insulin was not degraded during the binding assays. In competitive binding experiments, 2 ng/ml. (3.3 X 10(-10) M) of unlabeled insulin reduced 125 I-insulin binding by 50 per cent. Insulin analogues competed for binding in proportion to their biologic potencies. A curvilinear Scatchard plot was obtained, suggesting the existence of negatively cooperative site-site interactions among the insulin receptors. This was confirmed directly by studies of the dissociation kinetics. The high affinity, specificity, and negative cooperativity of the fibroblast insulin receptor closely resembles the properties of other human insulin receptors. The cultuted human fibroblast should prove a useful tissue for the study of insulin-resistant states in man.     

Insulin receptors in the heart muscle. Demonstration of specific binding sites and impairment of insulin binding in the plasma membrane of the obese hyperglycemic mouse.

The presence of insulin receptors in the heart muscle was investigated by measuring the binding of 125I-insulin to specific subcellular fractions of the rat and mouse myocardium. 125I-insulin bound to the plasma membrane fraction with a high degree of specificity and affinity. Insulin analogues competed with 125I-insulin in direct proportion to their biologic potency in vitro. Unlabeled insulin within the range of its concentrations in vivo inhibited 15 to 60 per cent of the 125I-insulin binding. The specific binding sites were finite in number and represented about 90 per cent of the total binding. The insulin-binding capacity of the plasma membrane fraction was twelve- to fifteenfold higher than that of the mitochondrial fraction. As in the liver, the binding was time- and temperature-dependent with a slower but higher binding achieved at a lower temperature. The binding sites appeared to be heterogeneous with respect to affinity. At 5 degrees C., the "higher-affinity" site had a K of about 2 times 10(9) M-1. No more than 10 per cent of the 125I-insulin was degraded by the heart plasma membranes after one hour at 30 degrees C. or twenty-two hours at 5 degrees C. Studies in the obese hyperglycemic (ob/ob) mouse revealed that the insulin binding is impaired in the heart muscle of this animal. Over a wide range of insulin concentrations, the plasma membrane fraction of ob/ob mice bound only 25 to 40 per cent as much insulin as did membranes of the thin littermates, suggesting that, as in the liver, the fat tissue, and the thymic lymphocyte, the number of insulin-binding sites is decreased in the heart of the ob/ob mouse. This defect selectively affected the plasma membrane fraction and could not be explained by differences in membrane purification or insulin-degrading activity. Heart and liver membranes of forty-hour fasted ob/ob mice bound two to three times as much insulin as did membranes of ob/ob mice fed ad libitum. These studies demonstrate and characterize the binding of insulin to heart muscle membranes; they extend to the heart muscle the insulin receptor defect also found in liver membranes and cells, in fat cell membranes, and in thymic lymphocytes of the ob/ob mouse.     

Evidence that two naturally occurring human insulin receptor alpha-subunit variants are immunologically distinct.

The IgG from a patient (Italy 2 [I2]) with hypoglycemia, due to autoantibodies to the insulin receptor, was purified on protein A Sepharose into two fractions that were tested in various human tissues and cells. The IgG fraction that bound protein A (absorbed IgG [IgGa]) nearly completely inhibited the binding of 125I-labeled insulin to various cells or tissues (placenta, IM-9, adipocytes, HEp-2-larynx cells, Epstein-Barr virus lymphocytes) but not greater than 50% of 125I-labeled insulin binding to human liver membranes. Conversely, both the IgG fraction from this patient, which did not bind protein A (flow-through IgG [IgGb]), and the IgGa fraction from a second similar patient (Italy 1 [I-1]) almost completely inhibited the binding of 125I-labeled insulin to liver membranes. The IgGa fraction from patient I-2 did not change receptor affinity because 50% inhibition of 125I-labeled insulin binding was not affected by either the presence or absence of these IgG fractions. Furthermore, liver binding data were not due to cross-reaction of 125I-labeled insulin to the insulinlike growth factor I receptor, and treatment of liver membranes with neuraminidase did not alter the inhibitory effect of the IgGa fraction from patient I-2 on 125I-labeled insulin binding to liver. Binding inhibition experiments performed with cells transfected with and overexpressing the -12 (human insulin receptor [HIR]-A) or the +12 (HIR-B) variant of HIR revealed that the IgGa fraction from patient I-2 inhibited 125I-labeled insulin binding to the HIR-A receptor but not to the HIR-B receptor.(ABSTRACT TRUNCATED AT 250 WORDS)     

Insulin and insulin receptor uptake into rat liver. Chloroquine action on receptor recycling

In the present study, the effect of chloroquine on both insulin and receptor distribution was examined in vivo. Insulin injection (25 nmol/100 g body wt) caused a marked accumulation of both insulin and its receptor in purified hepatic Golgi fractions by 15 min postinjection. Percoll fractionation of parent Golgi fractions resolved two endocytic components of low (rho = 1.040-1.050) and high (rho = 1.053-1.064) density in which the relative distribution of insulin binding sites was unaltered by chloroquine. Chloroquine significantly accumulated in the high-density region of the Percoll gradient consistent with this being a low pH compartment. 125I-insulin accumulated first in the low-density (1 min) and subsequently in the high-density region (5-10 min) of Percoll-subfractionated Golgi fractions. Chloroquine treatment caused marked accumulation of 125I-insulin in the high-density compartment with substantial retention of radiolabel therein at 20 min postinjection. 125I-insulin extracted from the Percoll fractions was comparably intact in control and chloroquine-treated rats. These data suggest that the chloroquine-accumulating, high-density compartment of hepatic Golgi fractions is the site of dissociation of internalized insulin-receptor complexes before degradation of the ligand and receptor recycling.     

Downregulation occurs normally in cultured Epstein-Barr virus-transformed lymphocytes from patients with extreme insulin resistance. Discrepancy between downregulation in vivo and in vitro

Levels of fasting plasma insulin are generally inversely correlated with 125I-insulin binding to circulating blood cells. In disease states associated with hyperinsulinemia (e.g., obesity and non-insulin-dependent diabetes mellitus), 125I-insulin binding is usually low. In contrast, 125I-insulin binding to circulating cells may be normal in patients with certain forms of extreme insulin resistance despite marked hyperinsulinemia. To explain this paradox, it has been proposed that postbinding defects in insulin action may give rise to defects in downregulation. We have employed cultured Epstein-Barr virus (EBV)-transformed lymphocytes from eight patients with extreme insulin resistance to address the question of whether there is a defect in the downregulation process in vitro. In this cell type, insulin leads to a decrease in the number of insulin receptors on the cell surface by accelerating the rate of degradation of insulin receptors. We could not detect any abnormality in in vitro down-regulation with cultured EBV-transformed lymphocytes from insulin-resistant patients. The apparent discrepancy between the in vivo and in vitro studies raises the possibility that some factor in the patient's internal milieu may prevent insulin-induced downregulation. An alternative possible explanation might be that the mechanism of downregulation in vitro differs from the mechanism whereby receptor number is regulated in vivo in insulin's target cells.     

Regulation of osteosarcoma EGF receptor affinity by phorbol ester and cyclic AMP

We studied the binding and degradation of 125I-labeled epidermal growth factor (EGF) by UMR-106 osteosarcoma cells and the regulation of EGF receptor affinity for EGF by the phorbol ester 12-O-tetradecanoyl-phorbol-13-acetate (TPA) and by treatments that raise intracellular levels of cyclic AMP. Cell surface binding of [125I]EGF to A431 cells reached a plateau after a 30 minute incubation at 37 degrees C but was undetectable in UMR-106 cells. Degradation of [125I]EGF proceeded at a 50-fold higher rate in A431 cells on a per cell basis, but receptor-bound [125I]EGF was internalized and degraded at a 3.5-fold higher rate by UMR-106 cells on a per receptor basis. At 4 degrees C, [125I]EGF labeled a single class of surface binding sites in the UMR-106 cell. Treatment with TPA at 37 degrees C reduced subsequent cell surface binding of [125I]EGF at 4 degrees C a maximum of 80% with an IC50 of 1.25 ng/ml. Maximal TPA reduction of [125I]EGF binding was observed within 5-15 minutes and was due to a reduction in the affinity of cell surface receptors of [125I]EGF without a change in receptor density. Pretreatment of the cells for 4 h with 30 microM forskolin, 1 mM isobutylmethylxanthine (IBMX) plus 30 microM forskolin, or 1 mM IBMX plus 100 ng/ml parathyroid hormone (PTH) attenuated the loss in [125I]EGF binding caused by a subsequent dose of 10 ng/ml of TPA by 17% (p less than 0.0005), 39% (p less than 0.0002), and 35% (p less than 0.002), respectively.(ABSTRACT TRUNCATED AT 250 WORDS)     

Characteristics of human erythrocyte insulin receptors

Highly specific insulin receptors have been identified on human erythrocytes. A modification of the monocyte insulin radioreceptor technique permitted distinct separation of human erythrocytes with their bound insulin from the free insulin. When incubated with 80 pg. per milliliter of 125I-insulin (pH 8.0, 3.5 hours, 15 degrees C.), erythrocytes from 17 normal volunteers specifically bound 10 per cent (+/- 1.450 S.D.) of the total 125I-insulin. Less than 15 per cent of the total 125I-insulin bound was nonspecific. Binding of 125I-insulin to human erythrocytes was dependent on pH and temperature. Less than 5 per cent of the insulin available to the plasma membrane was degraded. Both calcium and magnesium enhanced 125I-insulin binding by 100 per cent but had no synergistic effect when mixed in a 1:1 molar ratio. Scatchard analysis of the binding data resulted in a curvilinear plot with characteristics typical of negative cooperative interactions between receptor sites and with an unoccupied site affinity constant of 0.1 X 10(8) M-1. Human erythrocytes have 2,000 insulin binding sites per erythrocyte with 14 sites per square micrometer of surface area. The readily available human erythrocyte, thus, has both specific insulin binding sites and binding characteristics similar to other human cell types. These studies have provided the basis for further clinical investigation of polypeptide hormone receptors on human erythrocytes.     

Degradation of insulin by isolated mouse pancreatic acini. Evidence for cell surface protease activity

In the present study, we have used isolated mouse pancreatic acini to investigate the relationship between 125I-insulin binding and its degradation in order to probe the nature and cellular localization of the degradative process. In these cells, the proteolysis of 125I-insulin was dependent on time and cell concentration, and was saturated by unlabeled insulin with a Km of 290 nM. Since this value was much higher than the Kd for insulin binding to its receptor (1.1 nM), the data indicated that 125I-insulin degradation by acini occurred primarily via nonreceptor mechanisms. Several lines of evidence suggested that insulin was being degraded by the neutral thiol protease, insulin degrading enzyme (IDE). First, insulin degradation was inhibited by thiolreacting agents such as N-ethylmaleimide and p-chloromercuribenzoate. Second, the Km for degradation in acini was similar to the reported Km for IDE in other tissues. Third, the enzyme activity had a relative mol wt of approximately 130,000 by gel filtration, a value similar to that reported for purified IDE. Fourth, the degrading activity was removed with a specific antibody to IDE. Other lines of evidence suggested that enzymes located on the cell surface played a role in insulin degradation by acini. First, the nonpenetrating sulfhydryl reacting agent 5,5' dithiobis-2-nitrobenzoic acid blocked 125I-insulin degradation. Second, a specific antibody to IDE identified the presence of the enzyme on the cell surface. Third, chloroquine, leupeptin and antipain, agents that inhibit lysosomal function, did not influence 125I-insulin degradation. Fourth, highly purified pancreatic plasma membranes degraded 125I-insulin.     

Inducement of antibody that mimics insulin action on insulin receptor by insulin autoantibody directed at determinant at asparagine site on human insulin B chain.

We previously showed that purified human IgG1 insulin autoantibody (IAA) from the serum of male patient T.H. with insulin autoimmune syndrome is directed at a determinant at the asparagine site on the human insulin B chain. An anti-idiotypic antibody (anti-TH) that inhibited TH-IAA binding to human insulin was obtained by immunizing BALB/c mice with TH-IAA. Anti-TH bound to viable IM-9 cells and the purified insulin receptor from IM-9 cells. Anti-TH binding to IM-9 cells and the insulin receptor was inhibited by TH-IAA but not by human IgG. Moreover, incubation of HepG2 cells with anti-TH had an inhibitory effect on insulin binding to HepG2 cells. Anti-TH, like insulin, stimulated amino acid uptake in HepG2 cells. These findings indicate that the conformation of TH-IAA idiotope is a mirror image of the determinant on the insulin B chain, the binding site for TH-IAA on anti-TH is also related to the insulin binding site on the insulin receptor, and anti-TH mimics insulin action on the insulin receptor.     

Application of HPLC in disposition study of A14-125I-labeled insulin in mice

To describe quantitatively the in vivo distribution and elimination of insulin, high-performance liquid chromatography (HPLC) separation was applied to the pharmacokinetic study of human insulin labeled with 125I at tyrosine A14 (A14-125I-insulin) as a tracer. Intact A14-125I-insulin levels were determined by HPLC and trichloroacetic acid (TCA) precipitation in plasma and various tissues after its intravenous bolus injection into mice. TCA precipitation consistently overestimated the intactness of A14-125I-insulin compared with HPLC, possibly due to the presence of both a TCA-precipitable intermediate degradation product of labeled insulin found in HPLC elution profiles and reported high-molecular-weight forms of labeled insulin in plasma. Thus, TCA precipitation gave a considerably lower total plasma clearance (Cltot) value than HPLC. The half-life of A14-125I-insulin was prolonged by a simultaneous injection of 8 U/kg unlabeled insulin, and labeled insulin behaved similarly to [14C]inulin (an extracellular fluid marker). The concentration time profiles of HPLC-separated labeled insulin in plasma were analyzed by a noncompartmental moment method, and both Cltot and steady-state apparent volume distribution (VDss) of A14-125I-insulin were considerably decreased by unlabeled insulin coadministration. In particular, VDss of labeled insulin decreased by 79%, similar to that of inulin (181 ml/kg), suggesting that the nonspecific binding of labeled insulin to tissues was so small that VDss of labeled insulin was reduced to the extracellular fluid volume (approximately 20% of the body weight) when its receptor binding was blocked effectively by unlabeled insulin.(ABSTRACT TRUNCATED AT 250 WORDS)     

Early detection of degraded A14-125I-insulin in human fibroblasts by the use of high performance liquid chromatography

We studied the metabolism of A14-125I-insulin in intact human fibroblasts using high performance liquid chromatography (HPLC) to detect and separate its early degradation products. The high resolving power of HPLC enabled us to separate what has been considered "intact insulin" by Sephadex G-50 chromatography or TCA precipitability into two additional peaks that had decreased biochemical properties with respect to immunoprecipitability and receptor binding but not decreased TCA precipitability. We conclude that human fibroblast is capable of metabolizing insulin within 2 min at 37 degrees C into intermediate molecules that can be detected by HPLC but not by TCA precipitability or molecular sieve chromatography.     

Encapsulation of insulin for oral administration preserves interaction of the hormone with its receptor in vitro.

It has been shown that insulin associated with nanocapsules of isobutylcyanoacrylate retains biological activity after oral administration to diabetic rats from 6 to 21 days. Because part of this action is unexplained, we focused on the interaction of encapsulated insulin with the insulin receptor in vitro. We have shown that encapsulated insulin is able 1) to bind to insulin receptors both in rat liver plasma membranes and after solubilization from Chinese hamster ovary (CHO) cells transfected with the gene of human insulin receptor, 2) to accelerate 125I-labeled insulin dissociation from its receptor, and 3) to ensure transduction of a signal leading to stimulation of the beta-subunit phosphorylation, with parameters similar to those of native insulin. In addition, encapsulated 125I-insulin was rapidly internalized in transfected CHO cells. Analysis of cell-associated radioactivity showed that encapsulated insulin remained largely intact (greater than 80%) after 3 h, whereas native insulin was mostly degraded. These data indicate that encapsulated insulin fulfills all the earliest events at the receptor level leading to biological actions and suggests that encapsulation protects insulin against insulin degradation inside the cells.     

Degradation of insulin and insulin-like growth factors by enzyme purified from human erythrocytes. Comparison of degradation products observed with A14- and B26-[125I]monoiodoinsulin

An insulin-degrading enzyme has been purified from human erythrocytes. This enzyme degraded 125I-labeled insulin-like growth factor I (IGF-I) more slowly than 125I-IGF-II and degraded IGF-II more slowly than 125I-insulin. The time course of 125I-insulin degradation suggested the presence of intermediates, each of which was itself shown to be a substrate for the enzyme. One of these intermediates appeared to be made up entirely of B-chain residues and had HisB10 as its NH2-terminal. The final major radiolabeled degradation product of A14-[125I]monoiodoinsulin was a peptide with TyrA14 at the A-chain NH2 terminal. This peptide could be reduced with dithiothreitol, suggesting that it contained amino acid residues from both A- and B-chains. It was partially precipitated by trichloroacetic acid and anti-insulin antibody but bound poorly to IM-9 lymphocytes. The final major degradation product of B26-[125I]monoiodoinsulin was a peptide whose NH2-terminal was TyrB26 and could not be reduced by dithiothreitol. It was partially precipitated by anti-insulin antibody but was precipitated poorly, if at all, by trichloroacetic acid and bound poorly to IM-9 lymphocytes. The results show that this enzyme degraded insulin by sequential cleavage of peptide bonds on both A- and B-chains. We identified LeuA13-TyrA14, SerB9-HisB10, and PheB25-TyrB26 as three of the bonds that are cleaved.