Regulation of Insulin Degradation: Expression of an Evolutionarily Conserved Insulin-Degrading Enzyme Increases Degradation via an Intracellular Pathway
Wen-Liang Kuo, Barry D. Gehm, and Marsha Rich Rosner Ben May Institute Department of Pharmacological and Physiological Sciences University of Chicago Chicago, Illinois 60637
The insulin-degrading enzyme (IDE) is an evolutionarily conserved enzyme that has been implicated in cellular insulin degradation, but its site of action and importance in regulating insulin degradation have not been clearly established. We addressed this question by examining the effects of overexpressing IDE on insulin degradation in COS cells, using both human IDE (hIDE) and its Drosophila homolog (dIDE). The dIDE, which was recently cloned in our laboratory, has 46% amino acid identity with hIDE, degrades insulin with comparable efficiency, and is readily expressed in mammalian cells. Transient expression of dIDE or hIDE in COS monkey kidney cells led to a 5- to 7-fold increase in the rate of degradation of extracellular insulin, indicating that IDE can regulate cellular insulin degradation. Insulindegrading activity in the medium was very low and could not account for the difference between transfected and control cells. To further localize the site of IDE action, the fate of insulin after receptor binding was examined. The dIDE-transfected cells displayed increased degradation of prebound insulin compared to control cells. This increase in degradation was observed even when excess unlabeled insulin was added to block reuptake or extracellular degradation. These results indicate that IDE acts at least in part within the cell. The lysosomotropic agents chloroquine and NH4CI did not affect the increase in insulin degradation produced by transfection with dIDE, indicating that the lysosomal and IDE-mediated pathways of insulin degradation are independent. The results demonstrate that IDE can regulate the degradation of insulin by intact cells via an intracellular pathway. (Molecular Endocrinology 5: 1467-1476, 1991) 0888-8809/91 /1467-1476$03.00/0 Molecular Endocrinology Copyright © 1991 by The Endocrine Society
1467
INTRODUCTION
Insulin has pleiotropic effects on cellular metabolism and growth, including stimulation of glucose transport, lipid synthesis, and cell proliferation. Many studies have shown that the effects of insulin are initiated by the interaction of insulin with its receptor on the cell surface (reviewed in Ref. 1). Upon binding, insulin is internalized via receptor-mediated endocytosis, and internalized insulin eventually is degraded or shunted back to the outside of the cell through retroendocytosis (2). Although insulin degradation has been extensively studied, the site(s) of degradation, the specific enzyme(s) involved, the mechanism and the physiological importance of insulin degradation in cells are still not entirely clear. To date, substantial evidence suggests that the primary cellular insulin-degrading activity is the insulin-degrading enzyme (IDE) [reviewed by Duckworth (3)]. Since Mirsky and Broh-Kahn (4) first described a protease, termed insulinase, that degraded insulin with high specificity and low Km, extensive work has been performed to purify and characterize the enzyme(s). Insulin-degrading activity has been found in extracts of muscle, liver, kidney, fat cells, fibroblasts, placenta, brain, erythrocytes, and pancreas (5). The enzyme(s) primarily responsible for the degrading activity in several of these tissues has been purified. Various names, such as insulinase, insulin-specific protease (6), and IDE (7), have been given to these enzymes, and some differences among them have been reported, but further analysis of their substrate specificity and immunological cross-reactivity has suggested that these enzymes are identical (3, 8). The IDE is widely distributed and evolutionarily conserved. Homologs of IDE have been observed in humans, rodents, monkeys, Drosophila, and even bacteria (9-11). Comparison of the amino acid sequences of Drosophila (dIDE) and human (hIDE) IDEs shows 46% identity and 66.7% similarity (11). In addition, the degradation products generated by dIDE, hIDE, and rat
MOL ENDO1991 1468
IDE are very similar (8,12), and more importantly, they are similar to those found in vivo (13). Several lines of evidence suggest that IDE may play an important role in growth and development. In Drosophila, IDE is developmental^ regulated, being barely detectable in the embryo but elevated during the later stages. Studies using selective protease inhibitors have suggested that IDE, or a metalloprotease with a similar inhibitor profile, is required for muscle cell differentiation in mammals (14). Drosophila and rat liver IDEs were also shown to degrade transforming growth factor-a, a growth factor that binds to and activates epidermal growth factor receptors on mammalian cells (15). Thus, IDE is potentially also a selective growth factor-degrading enzyme and may have several functions in regulating cellular growth and differentiation. In the present experiments we examined the effect of overexpressing IDE on the degradation of insulin by COS cells. The recent cloning of hIDE and dIDE cDNAs and the ability to express active transfected IDE in cells enabled us to address two fundamental questions regarding IDE action. First, can IDE regulate the level of insulin degradation in a whole cell system? Second, if IDE does regulate insulin degradation, where does the degradation occur: within the cell after receptor-mediated endocytosis, or outside the cell by a secreted or peripherally associated enzyme? We now present evidence indicating that IDE can regulate insulin degradation intracellularly in COS cells.
RESULTS Expression of Enzymatically Active HIDE and dIDE in COS Cells The hIDE and dIDE cDNAs, which were both 3.4 kilobases (kb) long, were subcloned into an expression vector containing a cytomegalovirus promoter, pCMV0, as described in Materials and Methods; the constructed plasmids were designated pCMVhIDE and pCMVdIDE, respectively. Both plasmids contain an simian virus-40 (SV40) origin and can propagate as episomal vectors in COS cells, which were transformed with origin-defective SV40. Seventy-two hours after transfection of the COS cells with pCMVdIDE, pCMVhIDE, or pCMV0, cellular extracts were prepared, and insulin-degrading activity was measured, as described in Materials and Methods. The cellular extracts from cells transfected with pCMVhIDE and pCMVdIDE showed 2- to 2.5-fold higher insulin-degrading activity than those from the control cells (Fig. 1). The increased insulin-degrading activity in the transfected cells correlated well with increases in dIDE protein (11), and hIDE protein (data not shown), as measured by Western blotting. The insulin-degrading activity was about 27% higher in extracts from dIDE-transfected cells than in those from hIDE-transfected cells (Fig. 1). The results shown here are representative of at least three independent experiments. Degradation was insulin specific, since it could
Vol5No. 10
Control
hIDE
dIDE
TREATMENT Fig. 1. IDE Transfection of COS Cells Increases Insulin-Degrading Activity in Cell Extracts As described in Materials and Methods, COS cells were transfected with pCMVhIDE, pCMVdIDE, or pCMV0, and extracts of the cells were prepared and assayed for IDE activity. Percent degradation is shown as the mean ± SD of three determinations. Similar results were obtained in three separate experiments.
be substantially inhibited by the addition of excess unlabeled insulin. The absolute transfection efficiency in transient expression systems is typically on the order of a few percent (16). Thus, it is likely that a 2- to 2.5-fold increase in overall IDE activity reflects a much larger increase in activity in the cells expressing the transfected IDE. Differences in transfection efficiency might account for the higher IDE activity produced by pCMVdIDE compared to pCMVhIDE, so we used an internal control to compare relative transfection effiencies. The control enzyme, chloramphenicol acetyl-coenzyme-A transferase, was expressed from a cotransfected plasmid, pSV2-cat (17). In the experiment presented in Fig. 1, the transfection efficiencies for pCMVdIDE and pCMVhIDE were essentially identical. In some experiments transfection efficiencies for the two vectors differed, but after normalizing to control for these differences, extracts of dIDE-transfected cells consistently showed higher IDE activity than those of hIDE-transfected cells (data not shown). For this reason, pCMVdIDE was used in preference to pCMVhIDE in experiments requiring maximal expression of IDE activity. Expression of dIDE Regulates Insulin Degradation in COS Cells We have shown that IDEs could be expressed in an active form in mouse NIH3T3 cells (11) and monkey COS cells (Fig. 1) by comparing the insulin-degrading activity of cellular extracts from IDE-transfected and control cells. To determine whether IDE can regulate
Regulation of Insulin Degradation
1469
insulin degradation in intact cells, we examined the effect of IDE transfection on insulin degradation by COS cells. Seventy-two hours after transfection with pCMVdIDE or pCMVo, cells were assayed for degradation of extracellular 125l-insulin as described in Materials and Methods. The results shown in Fig. 2A indicate that dIDE-transfected cells degraded 125l-insulin ~7 times more rapidly than control cells during the 2-h assay, and that the degradation was insulin specific, as it could be almost totally inhibited by excess unlabeled insulin. Similar increases were obtained with hlDEtransfected cells (Fig. 2B). These results clearly demonstrate that IDE is able to regulate insulin degradation by cells, and that this property is evolutionary conserved.
insulin-degrading activity was found in the medium, so the degradation must have occurred at the cell surface or intracellularly. Although there have been some reports of insulin degradation at the cell surface (18,19), most work has focused on intracellular degradation in which insulin must follow a pathway of internalization, degradation, and release. To study this pathway in more detail, we analyzed the fate of 125l-insulin prebound to COS cells. Marshall (2) has performed similar experiments on isolated adipocytes and has proposed a dual pathway model in which some internalized insulin is degraded and some is released intact after traversing a more rapid nondegradative pathway, or shunt. We wished to learn whether the dual pathway model could be applied to COS cells, and how overexpression of insulin-degrading activity would affect the pathways. For analyses of internalized insulin, 125l-insulin was prebound to control and dIDE-transfected COS cells, as described in Materials and Methods. Briefly, the cells were incubated with 125l-insulin on ice to prevent internalization of ligand-receptor complexes (20). Unbound 125 l-insulin was then washed away, and the cells were shifted to 37 C to allow internalization. At specified times after the temperature shift, 125I was measured in three compartments: extracellular (released into medium), surface bound (cell associated, but released by acid washing), and intracellular (cell associated and resistant to acid washing). Acid washing has been widely used to distinguish surface-bound from internalized ligands, including insulin (21). The sum of the three categories was taken as the total 125l-insulin originally bound to the cells; the radioactivity in each category was expressed as a percentage of this total. Unlabeled insulin (1 HM) was included in the 37 C incubation in some experiments to inhibit any degradation by IDE in the medium and prevent possible reinternalization of released 125l-insulin.
One possible explanation for the observed results was that IDE expressed by the COS cells was released into the medium and degraded insulin outside of the cells. Since neither hIDE nor dIDE has obvious signal sequences, such release could be an artifact of cell stress. To test the possibility of IDE release, we collected conditioned medium (binding buffer incubated with cells for 2 h at 37 C) from IDE-transfected and control cells and assayed their insulin-degrading activity, as described in Materials and Methods. Almost no insulin-degrading activity was released from control cells, and very little activity (
Wen-Liang Kuo, Barry D. Gehm, and Marsha Rich Rosner Ben May Institute Department of Pharmacological and Physiological Sciences University of Chicago Chicago, Illinois 60637
The insulin-degrading enzyme (IDE) is an evolutionarily conserved enzyme that has been implicated in cellular insulin degradation, but its site of action and importance in regulating insulin degradation have not been clearly established. We addressed this question by examining the effects of overexpressing IDE on insulin degradation in COS cells, using both human IDE (hIDE) and its Drosophila homolog (dIDE). The dIDE, which was recently cloned in our laboratory, has 46% amino acid identity with hIDE, degrades insulin with comparable efficiency, and is readily expressed in mammalian cells. Transient expression of dIDE or hIDE in COS monkey kidney cells led to a 5- to 7-fold increase in the rate of degradation of extracellular insulin, indicating that IDE can regulate cellular insulin degradation. Insulindegrading activity in the medium was very low and could not account for the difference between transfected and control cells. To further localize the site of IDE action, the fate of insulin after receptor binding was examined. The dIDE-transfected cells displayed increased degradation of prebound insulin compared to control cells. This increase in degradation was observed even when excess unlabeled insulin was added to block reuptake or extracellular degradation. These results indicate that IDE acts at least in part within the cell. The lysosomotropic agents chloroquine and NH4CI did not affect the increase in insulin degradation produced by transfection with dIDE, indicating that the lysosomal and IDE-mediated pathways of insulin degradation are independent. The results demonstrate that IDE can regulate the degradation of insulin by intact cells via an intracellular pathway. (Molecular Endocrinology 5: 1467-1476, 1991) 0888-8809/91 /1467-1476$03.00/0 Molecular Endocrinology Copyright © 1991 by The Endocrine Society
1467
INTRODUCTION
Insulin has pleiotropic effects on cellular metabolism and growth, including stimulation of glucose transport, lipid synthesis, and cell proliferation. Many studies have shown that the effects of insulin are initiated by the interaction of insulin with its receptor on the cell surface (reviewed in Ref. 1). Upon binding, insulin is internalized via receptor-mediated endocytosis, and internalized insulin eventually is degraded or shunted back to the outside of the cell through retroendocytosis (2). Although insulin degradation has been extensively studied, the site(s) of degradation, the specific enzyme(s) involved, the mechanism and the physiological importance of insulin degradation in cells are still not entirely clear. To date, substantial evidence suggests that the primary cellular insulin-degrading activity is the insulin-degrading enzyme (IDE) [reviewed by Duckworth (3)]. Since Mirsky and Broh-Kahn (4) first described a protease, termed insulinase, that degraded insulin with high specificity and low Km, extensive work has been performed to purify and characterize the enzyme(s). Insulin-degrading activity has been found in extracts of muscle, liver, kidney, fat cells, fibroblasts, placenta, brain, erythrocytes, and pancreas (5). The enzyme(s) primarily responsible for the degrading activity in several of these tissues has been purified. Various names, such as insulinase, insulin-specific protease (6), and IDE (7), have been given to these enzymes, and some differences among them have been reported, but further analysis of their substrate specificity and immunological cross-reactivity has suggested that these enzymes are identical (3, 8). The IDE is widely distributed and evolutionarily conserved. Homologs of IDE have been observed in humans, rodents, monkeys, Drosophila, and even bacteria (9-11). Comparison of the amino acid sequences of Drosophila (dIDE) and human (hIDE) IDEs shows 46% identity and 66.7% similarity (11). In addition, the degradation products generated by dIDE, hIDE, and rat
MOL ENDO1991 1468
IDE are very similar (8,12), and more importantly, they are similar to those found in vivo (13). Several lines of evidence suggest that IDE may play an important role in growth and development. In Drosophila, IDE is developmental^ regulated, being barely detectable in the embryo but elevated during the later stages. Studies using selective protease inhibitors have suggested that IDE, or a metalloprotease with a similar inhibitor profile, is required for muscle cell differentiation in mammals (14). Drosophila and rat liver IDEs were also shown to degrade transforming growth factor-a, a growth factor that binds to and activates epidermal growth factor receptors on mammalian cells (15). Thus, IDE is potentially also a selective growth factor-degrading enzyme and may have several functions in regulating cellular growth and differentiation. In the present experiments we examined the effect of overexpressing IDE on the degradation of insulin by COS cells. The recent cloning of hIDE and dIDE cDNAs and the ability to express active transfected IDE in cells enabled us to address two fundamental questions regarding IDE action. First, can IDE regulate the level of insulin degradation in a whole cell system? Second, if IDE does regulate insulin degradation, where does the degradation occur: within the cell after receptor-mediated endocytosis, or outside the cell by a secreted or peripherally associated enzyme? We now present evidence indicating that IDE can regulate insulin degradation intracellularly in COS cells.
RESULTS Expression of Enzymatically Active HIDE and dIDE in COS Cells The hIDE and dIDE cDNAs, which were both 3.4 kilobases (kb) long, were subcloned into an expression vector containing a cytomegalovirus promoter, pCMV0, as described in Materials and Methods; the constructed plasmids were designated pCMVhIDE and pCMVdIDE, respectively. Both plasmids contain an simian virus-40 (SV40) origin and can propagate as episomal vectors in COS cells, which were transformed with origin-defective SV40. Seventy-two hours after transfection of the COS cells with pCMVdIDE, pCMVhIDE, or pCMV0, cellular extracts were prepared, and insulin-degrading activity was measured, as described in Materials and Methods. The cellular extracts from cells transfected with pCMVhIDE and pCMVdIDE showed 2- to 2.5-fold higher insulin-degrading activity than those from the control cells (Fig. 1). The increased insulin-degrading activity in the transfected cells correlated well with increases in dIDE protein (11), and hIDE protein (data not shown), as measured by Western blotting. The insulin-degrading activity was about 27% higher in extracts from dIDE-transfected cells than in those from hIDE-transfected cells (Fig. 1). The results shown here are representative of at least three independent experiments. Degradation was insulin specific, since it could
Vol5No. 10
Control
hIDE
dIDE
TREATMENT Fig. 1. IDE Transfection of COS Cells Increases Insulin-Degrading Activity in Cell Extracts As described in Materials and Methods, COS cells were transfected with pCMVhIDE, pCMVdIDE, or pCMV0, and extracts of the cells were prepared and assayed for IDE activity. Percent degradation is shown as the mean ± SD of three determinations. Similar results were obtained in three separate experiments.
be substantially inhibited by the addition of excess unlabeled insulin. The absolute transfection efficiency in transient expression systems is typically on the order of a few percent (16). Thus, it is likely that a 2- to 2.5-fold increase in overall IDE activity reflects a much larger increase in activity in the cells expressing the transfected IDE. Differences in transfection efficiency might account for the higher IDE activity produced by pCMVdIDE compared to pCMVhIDE, so we used an internal control to compare relative transfection effiencies. The control enzyme, chloramphenicol acetyl-coenzyme-A transferase, was expressed from a cotransfected plasmid, pSV2-cat (17). In the experiment presented in Fig. 1, the transfection efficiencies for pCMVdIDE and pCMVhIDE were essentially identical. In some experiments transfection efficiencies for the two vectors differed, but after normalizing to control for these differences, extracts of dIDE-transfected cells consistently showed higher IDE activity than those of hIDE-transfected cells (data not shown). For this reason, pCMVdIDE was used in preference to pCMVhIDE in experiments requiring maximal expression of IDE activity. Expression of dIDE Regulates Insulin Degradation in COS Cells We have shown that IDEs could be expressed in an active form in mouse NIH3T3 cells (11) and monkey COS cells (Fig. 1) by comparing the insulin-degrading activity of cellular extracts from IDE-transfected and control cells. To determine whether IDE can regulate
Regulation of Insulin Degradation
1469
insulin degradation in intact cells, we examined the effect of IDE transfection on insulin degradation by COS cells. Seventy-two hours after transfection with pCMVdIDE or pCMVo, cells were assayed for degradation of extracellular 125l-insulin as described in Materials and Methods. The results shown in Fig. 2A indicate that dIDE-transfected cells degraded 125l-insulin ~7 times more rapidly than control cells during the 2-h assay, and that the degradation was insulin specific, as it could be almost totally inhibited by excess unlabeled insulin. Similar increases were obtained with hlDEtransfected cells (Fig. 2B). These results clearly demonstrate that IDE is able to regulate insulin degradation by cells, and that this property is evolutionary conserved.
insulin-degrading activity was found in the medium, so the degradation must have occurred at the cell surface or intracellularly. Although there have been some reports of insulin degradation at the cell surface (18,19), most work has focused on intracellular degradation in which insulin must follow a pathway of internalization, degradation, and release. To study this pathway in more detail, we analyzed the fate of 125l-insulin prebound to COS cells. Marshall (2) has performed similar experiments on isolated adipocytes and has proposed a dual pathway model in which some internalized insulin is degraded and some is released intact after traversing a more rapid nondegradative pathway, or shunt. We wished to learn whether the dual pathway model could be applied to COS cells, and how overexpression of insulin-degrading activity would affect the pathways. For analyses of internalized insulin, 125l-insulin was prebound to control and dIDE-transfected COS cells, as described in Materials and Methods. Briefly, the cells were incubated with 125l-insulin on ice to prevent internalization of ligand-receptor complexes (20). Unbound 125 l-insulin was then washed away, and the cells were shifted to 37 C to allow internalization. At specified times after the temperature shift, 125I was measured in three compartments: extracellular (released into medium), surface bound (cell associated, but released by acid washing), and intracellular (cell associated and resistant to acid washing). Acid washing has been widely used to distinguish surface-bound from internalized ligands, including insulin (21). The sum of the three categories was taken as the total 125l-insulin originally bound to the cells; the radioactivity in each category was expressed as a percentage of this total. Unlabeled insulin (1 HM) was included in the 37 C incubation in some experiments to inhibit any degradation by IDE in the medium and prevent possible reinternalization of released 125l-insulin.
One possible explanation for the observed results was that IDE expressed by the COS cells was released into the medium and degraded insulin outside of the cells. Since neither hIDE nor dIDE has obvious signal sequences, such release could be an artifact of cell stress. To test the possibility of IDE release, we collected conditioned medium (binding buffer incubated with cells for 2 h at 37 C) from IDE-transfected and control cells and assayed their insulin-degrading activity, as described in Materials and Methods. Almost no insulin-degrading activity was released from control cells, and very little activity (
