- Email: [email protected]
Characterization of an intracellular insulin-degrading enzyme in human erythrocytes
BIOCHEMICAL
MEDICINE
AND
METABOLIC
BIOLOGY
39, 284-289 (1988)
Characterization of an Intracellular Insulin-Degrading Human Erythrocytes’
Enzyme in
K. K. GAMBHIR~ AND S. G. NERURKAR Molecular
Endocrinology
Laboratory, Department of Medicine, Howard University, College of Medicine, Washington, D.C. 20059 Received June 19, 1987
We showed that human erythrocytes have intracellular insulin-degrading activity (1,2). The nature of the activity was not specified in that report (2). The available information (3) showed that reduced glutathione can degrade insulin to some extent. Since reduced glutathione is present in human erythrocytes (1 mM in whole blood, none in plasma), it is possible to ascribe the insulin-degrading activity to reduced glutathione in erythrocytes. However, erythrocyte lysate degraded far more insulin than did physiologic and supraphysiologic concentrations of reduced glutathione. In this report we present data showing the insulin-degrading activity of human erythrocytes is due to a specific enzyme. MATERIALS
AND METHODS
Reduced glutathione, chloroquine, antipain, and bacitracin were obtained from Sigma Chemical Co. Trypsin was obtained from Microbiological Associates, Bethesda, Maryland, and ‘251-labeled mouse nerve growth factor (sp act 30 &i/kg) and epidermal growth factor (sp act 400 #Zi/pg) were gifts from Drs. David End and Gordon Guroff, Development Neurobiology, National Institute of Child Health and Human Development, Bethesda, Maryland. ‘251-labeled mouse prolactin (sp act 65 &i/pg) and human thyroid stimulating hormone (sp act 105 &i/pg) were obtained, respectively, from Dr. Anna-Lisa Borofsky, Department of Physiology and Radioimmunoassay Laboratories, Howard University, Washington, D.C. The other chemicals have been described in our earlier report (4). 1251labeled insulin was prepared following iodination of insulin with a sp act of 180 pCi/hg in our laboratory (4). Venous blood (30-50 ml) was drawn from non-obese, nondiabetic subjects, 25-40 years of age, after overnight fasting. They had no family history of diabetes. Erythrocytes were isolated by Hypaque-Ficoll gradient centrifugation as described ’ Parts of this paper were presented on July 5, 1984, at the Seventh International Endocrinology, Quebec City, Canada. * To whom all correspondence should be addressed. 0885-4505188 $3 .OO Copylight All rights
0 1988 by Academic Press, Inc. of reproduction in any form reserved.
284
Congress of
ERYTHROCYTE
INSULIN-DEGRADING
ENZYME
285
earlier (4). The following buffers were prepared from buffer G (4): For hemolysis of cells (buffer G,) buffer G (187 mM) without bovine serum albumin was diluted 37.4-fold with deionized glass distilled water and was adjusted to pH 7.4. BSA was added to achieve a final concentration of 0.1%. For enzyme purification (buffer GJ buffer G (without BSA) was diluted 37.4-fold, made up to 1 nM with reduced glutathione and was adjusted to pH 7.4. For the enzyme assay (buffer G,) BSA was added to buffer G2 to achieve a final concentration of 0.1%. METHODS Hormone Degradation Assay
Four hundred microliters of hemolysate (or an appropriate aliquot in 400 ~1 of buffer G,) was incubated with 50 pg of ‘251-labeled hormone (insulin, prolactin, hTSH,3 NGF, or EGF) in final volume of 425 ~1 at 37” for 25 min. The reaction was stopped by the addition of 400 ~1 bovine serum albumin (lOOg/liter) and 400 ~1 of 20% trichloroacetic acid. The resulting suspension was mixed and centrifuged, and radioactivity was measured in the supernatant and precipitant fractions. A control for every hormone containing 400 ~1 of buffer G3 instead of hemolysate was treated the same way. Percentage degradation of hormone was expressed as described earlier by Nerurkar and Gambhir (2). The enzyme activity was expressed as insulin degraded per minute either per milligram of protein or per milliliter of incubation mixture. (a) Preparation of cytosol fraction-isolated erythrocytes (13-20 ml) were hemolyzed using 20 vol of buffer G, at 23” for 15 min. Plasma membranes were removed by centrifuging the hemolysate at 35,000g at 4” in a Sorvall RC-5B refrigerated centrifuge using rotor SS-34. The resulting plasma membrane free membrane was used for further purification. (b) Ammonium sulfate fractionation-the hemolysate was saturated to 40% (243 g/liter) with ammonium sulfate at 4” and centrifuged. The precipitate was discarded and the supernate was further saturated to 60% (32 g/liter) with ammonium sulfate. The precipitate thus obtained was dissolved in 2 ml of buffer G2 and was dialyzed against the same buffer at 4” until free of ammonium sulfate. Ten microliters of the dialyzed solution was used for measuring insulin-degrading activity and protein content by the method of Lowry et al. (5). (c) Gel filtration of column chromatography-l ml of dialyzed solution was placed on a preequilibrated Sephadex G-100 column (1.8 x 90 cm) and eluted with buffer G2 at 4”. The eluate, in 2-ml fractions, read for absorbancy at 280 nm for protein and 419 nm for hemoglobin. An appropriate aliquot from each 2-ml fraction of eluate was used for the determination of insulin-degrading activity and protein. Properties of the Enzyme
(a) Specificity of the enzyme was tested by its action on various substrates and comparing it with that of trypsin. 3 Abbreviations used: hTSH, human thyroid stimulating hormone; NGF, nerve growth factor; EGF, epidermal growth factor; BSA, bovine serum albumin; EH, Eadie-Hofstee; LB, LineweaverBurk; IDE, insulin-degrading enzyme.
286
GAMBHIR
AND NERURKAR TABLE
Step Hemolysate (12 ml RBC) WH,),SO, (40-60% fraction dialyzed Sephadex G-100 (column eluate)
Volume (ml)
1
Total activity (w/min)
Total protein (mg)
1350
3938
322 280
Specific activity Wmin/mg)
Yield @)
Purification (-fold)
0.344
100
-
43.1
7.48
23.9
21.7
22.5
12.44
20.7
36.2
250 2 36
(b) Molecular weight of the enzyme was determined by passage of partially purified enzyme at 4” through a Sephadex G-200 (1.5 x 90 cm) column equilibrated with buffer GZ. The column was calibrated with standard proteins: catalase (232,000), aldolase (lSS,OOO), hemoglobin (64,000), and Dextran Blue 2000. (c) Determination of optimum pH was carried out for lysate using buffer G, and for partially purified enzyme using buffer G3. Series of buffers ranging in pH from 6 to 10 were prepared from buffers G, and G3 using 1 N HCl and 1 N NaOH . (d) The K,,, value for insulin was determined with partially purified enzyme (0.06 mg protein/tube). Using 100 pg of ‘251-insulin with O-5 pg of unlabeled insulin as a substrate, the initial velocity of insulin degradation was measured at 37” and pH 7.4 using the trichloroacetic acid precipitation technique. The velocity of insulin degradation was expressed as namograms/milliliter/minute. The data obtained in these experiments were analyzed by Lineweaver-Burk (l/v vs l/(S)) and Eadie-Hofstee (v against v/(S)) plots with the help of a HewlettPackard computer Model 9815 A and a linear regression data cartridge. These plots yielded K,,,, If,,,,, and coefficient of linear relationship (r). (e) Various inhibitors: bacitracin, N-ethylmaleimide, antipain, chloroquine, EDTA, and soybean trypsin inhibitor at different concentrations were used to find out the percentage inhibition of insulin-degrading activity. RESULTS
AND DISCUSSION
The erythrocyte lysate obtained after removal of plasma membranes contained most of the insulin-degrading activity. About 5% of the activity was precipitated in the O-40% ammonium sulfate saturated fraction. A major portion of the enzymatic activity was precipitated at 40-60% ammonium sulfate saturation (Table 1). This fraction (40-60%) was dissolved in 2 ml of buffer G2 and was dialyzed against the same at 4”, with frequent changes of the dialyzing buffer. Gel filtration column chromatography of the above fraction resulted in two major protein peaks. The first peak was eluted slightly after the void volume and contained most of the insulin-degrading activity. The second peak contained hemoglobin and had no insulin-degrading activity. Compared to trypsin, the insulin-degrading enzyme was very specific toward insulin (Table 2). Many proteases have been described in human erythrocytes.
ERYTHROCYTE
INSULIN-DEGRADING
Specificity of Erythrocytic Substrate
(100 w/ml) Insulin Prolactin hTSH NGF EGF Insulin + BSA (1 gm/liter) BSA (2 gm/liter) BSA (3 gm/liter)
TABLE 2 Insulin-Degrading Percentage trypsin 50 50 60 60 50 25 13 8
287
ENZYME
Enzyme (IDE) Degradation IDE
by
50 0 0 0 8 50 48 48
Our preparation can be similar to that reported by Brodal (6) and Kolb et al. (10). It appears that the enzymes described by others (7,9,11) are dissimilar to the enzyme reported in this communication. The enzyme preparation obtained by Kolb et al. (10) had the following differences with our findings. They carried out purification of the enzyme without a reducing agent, a procedure which we could not reproduce. Their claim of stability of enzyme at - 40” was not verifiable using our procedures. They have not reported a & value for insulin nor an optimum pH for the enzyme which we are reporting in this article. On the Sephadex G-200 column, the enzyme was eluted at the same volume as that of aldolase (158,000 Da). The approximate molecular weight was found to be 160,000 Da. The enzyme activity was evaluated over a wide range of pH. The optimal pH of enzymic activity in lysate and purified preparation was found to be in the range of physiological pH. At pH 6.0 and 9.0 enzyme activity was lowered by about 60% (data not shown). The Michaelis-Menten constant (ZC,) value for insulin was determined from the experimental data by two procedures. The first procedure was the LineweaverBurk plot and the second method involved the Eadie-Hofstee plot. On the basis of a K, value for insulin, optimum pH, and molecular weight, the enzyme from human erythrocytes was found to be comparable to those purified from rat adipocytes (12), liver (13), muscle (14), and especially from human muscles (16). On the other hand, placental (15) enzyme had optimum pH in the acidic range. Generally, the Lineweaver-Burk plot is employed to obtain a K, value for insulin. However, we analyzed the data from five different experiments by Lineweaver-Burk and Eadie-Hofstee plots (Table 3). It appears that linearity of the plot is more easily obtained in the former than in the latter. For example, the fourth experiment shows a very good linearity by the LB plot while the EH plot for the same data fails to show such linearity. On the other hand, the third experiment shows excellent linearity in the EH plot and concurrence in K, values by both plots. We are, however, reporting the average K, value
GAMBHIR
288
AND NERURKAR
TABLE 3 Kinetic Data for Insulin-Degrading
Enzyme
(LB) plot Expt. No.
Km (nM)
1
245.7 51.9 123.3 200.2 190.6 162.3
2 3 4 5 Mean
V,,X (ng/ml/min) 17.5 4.4 15.6 5.4 31.0 14.8
(EH) plot K, r”
0.99 0.97 0.99 0.97 0.97 0.98
hM)
179.3 85.1 118.4 73.3 149.8 133.2
V,,X (ng/ml/min) 14.8 5.8 16.5 3.6” 27.8 16.2
r
0.88 0.85 0.95 0.56h 0.84 0.88
a Coefficient of correlation. b These values were not included in averaging EH plot values.
obtained by LB plot to facilitate its comparison to the K, value for IDE from other tissues. The enzymatic activity was inhibited by bacitracin (500 units/ml) and Nethylmaleimide (1 mM), while antipain (1 mM), chloroquine (100 mM), < EDTA (50 mM), and soybean trypson inhibitor (5 mg/ml) failed to do so. These results are similar to those obtained with enzyme purified from other tissues (12-16) and emphasize the importance of -SH groups in maintenance of enzymatic activity. The presence of specific insulin receptors on human erythrocytes with characteristics and properties (17-19) similar to those on other cell types and their diagnostic importance (20) has been established. In this communication we partially purified cytosolic insulin-degrading enzyme from human erythrocytes. The enzyme is shown to have specificity and similar affinity (K, value) for insulin as in other cell types. Recently, we showed decreased insulin-degrading activity in uremics (12). Although these findings suggest that the enzyme has clinical significance, much work is needed to clarify this significance. Liver and kidney are the main organs which clear insulin from the circulation produced centrally by the pancreas. However, it is conceivable that due to the large number of RBCs (approximately 32 x 10”) in circulation, these cells might be responsible for maintaining the basal fasting level of this hormone. SUMMARY Using conventional techniques of ammonium sulfate fractionation and Sephadex gel column chromatography, insulin-degrading enzyme was partially purified from lysate of human erythrocytes. The enzymatic activity was measured by the trichloroacetic acid precipitation method. Compared to trypsin, the enzyme was highly specific for insulin. The apparent molecular weight of the enzyme was 160,000 Da, the optimum pH was the 7.4 to 7.8 range, and the K, value for insulin for the partially purified enzyme was 162 nM. Bacitracin and N-ethylmaleimide were potent inhibitors, while chloroquine, ethylenediaminetetraacetate, antipain, and soybean trypsin inhibitor failed to inhibit the activity of the enzyme.
ERYTHROCYTE
INSULIN-DEGRADING
ENZYME
289
Like most nucleated cells, human erythrocytes not only have the membranal insulin receptors, but also possess the cytosolic specific insulin-degrading enzyme. Insulin internalization and degradation are shown to be due to the receptor and the enzyme acting in concert as in many nucleated cells. Anucleated erythrocytes have both these entities for possible internalization and degradation of insulin. ACKNOWLEDGMENT These investigations were supported by grants from The Charles and Mary Latham Trust Fund, Washington, D.C., and The March of Dimes (Clinical Research Grant No. 6-262). REFERENCES 1. Gambhir, K. K., Nerurkar, S. G.. Das, P. D., Archer, J. A., and Henry, W. L., Jr., Endocrinology 109, 1787 (1981). 2. Nerurkar, S. G., and Gambhir, K. K., Clin. Chem. 27, 607 (1981). 3. Brush, J. S., and Heming, H., Endocrinology 104, 1639 (1981). 4. Gambhir, K. K., Archer, J. A., and Bradley, C. J., Diabetes 27, 701 (1978). 5. Lowery, 0. H., Rosebrough, N. F., Farr, A. L., and Randall, R. J., J. Biol. Chem. 193, 265 (1951). 6. Brodal, N. P., J. Biochem. 18, 201 (1971). 7. Tokes, Z. A., and Chambers, S. M., Biochim. Biophys. Acta, 389, 325 (1975). 8. Pontremoli, S., Salamino, F., Separatore. B., Melloni, E., Morelli, A.. Benatti. U., and DeFlora. A., Biochem. J. 181, 559 (1979). 9. Krischner, R. J., and Goldberg, A. L., J. Biol. Chem. 258, 967 (1980). 10. Kolb, H. J., and Standl, E., Hoppeseyler’s, Z. Physiol. Chem. 361, 1029 (1980). 11. Chandler, M. L., and Varandani, P. T., Diabetes 23, 232 (1974). 12. Goldstein, B. J., and Livingston, J. N., Endocrinology 108, 953 (1981). 13. Brughen, G. A., Kitabchi, A. E., and Brush, J. S., Endocrinology 91, 633 (1972). 14. Duckworth, W. C., and Kitabchi, A. E., Diabetes 23, 536 (1974). 15. Posner, B. I., Diabetes 22, 552 (1973). 16. Neal, G. W., and Kitabchi, A. E., Biochim. Biophys. Acta. 719, 259 (1982). 17. Peterson, S. W., Miller, A. L., Kelleher, R. S.. and Murray, E. F., J. Biol. Chem. 258, 9605 (1983). 18. lm, J. H., Meezan. E., Rockley, C. E., and Kim, H. D., J. Biol. Chem. 258, 5021 (1983). 19. Grigorescu, F., White, M. F., and Kahn, C. R., J. Biol. Chem. 258, 13708 (1983). 20. Gambhir. K. K., Nerurkar, S. G., Cruz, 1. A., and Hosten, A. O., Biochem. Med. 25, 62 (1981). 21. Nerurkar, S. G., Butterfield. S. G., Cruz, 1. A., Hosten, A. O., Dillard, M. G., and Gambhir, K. K., Clin. Res. 31, 393A (1983).
MEDICINE
AND
METABOLIC
BIOLOGY
39, 284-289 (1988)
Characterization of an Intracellular Insulin-Degrading Human Erythrocytes’
Enzyme in
K. K. GAMBHIR~ AND S. G. NERURKAR Molecular
Endocrinology
Laboratory, Department of Medicine, Howard University, College of Medicine, Washington, D.C. 20059 Received June 19, 1987
We showed that human erythrocytes have intracellular insulin-degrading activity (1,2). The nature of the activity was not specified in that report (2). The available information (3) showed that reduced glutathione can degrade insulin to some extent. Since reduced glutathione is present in human erythrocytes (1 mM in whole blood, none in plasma), it is possible to ascribe the insulin-degrading activity to reduced glutathione in erythrocytes. However, erythrocyte lysate degraded far more insulin than did physiologic and supraphysiologic concentrations of reduced glutathione. In this report we present data showing the insulin-degrading activity of human erythrocytes is due to a specific enzyme. MATERIALS
AND METHODS
Reduced glutathione, chloroquine, antipain, and bacitracin were obtained from Sigma Chemical Co. Trypsin was obtained from Microbiological Associates, Bethesda, Maryland, and ‘251-labeled mouse nerve growth factor (sp act 30 &i/kg) and epidermal growth factor (sp act 400 #Zi/pg) were gifts from Drs. David End and Gordon Guroff, Development Neurobiology, National Institute of Child Health and Human Development, Bethesda, Maryland. ‘251-labeled mouse prolactin (sp act 65 &i/pg) and human thyroid stimulating hormone (sp act 105 &i/pg) were obtained, respectively, from Dr. Anna-Lisa Borofsky, Department of Physiology and Radioimmunoassay Laboratories, Howard University, Washington, D.C. The other chemicals have been described in our earlier report (4). 1251labeled insulin was prepared following iodination of insulin with a sp act of 180 pCi/hg in our laboratory (4). Venous blood (30-50 ml) was drawn from non-obese, nondiabetic subjects, 25-40 years of age, after overnight fasting. They had no family history of diabetes. Erythrocytes were isolated by Hypaque-Ficoll gradient centrifugation as described ’ Parts of this paper were presented on July 5, 1984, at the Seventh International Endocrinology, Quebec City, Canada. * To whom all correspondence should be addressed. 0885-4505188 $3 .OO Copylight All rights
0 1988 by Academic Press, Inc. of reproduction in any form reserved.
284
Congress of
ERYTHROCYTE
INSULIN-DEGRADING
ENZYME
285
earlier (4). The following buffers were prepared from buffer G (4): For hemolysis of cells (buffer G,) buffer G (187 mM) without bovine serum albumin was diluted 37.4-fold with deionized glass distilled water and was adjusted to pH 7.4. BSA was added to achieve a final concentration of 0.1%. For enzyme purification (buffer GJ buffer G (without BSA) was diluted 37.4-fold, made up to 1 nM with reduced glutathione and was adjusted to pH 7.4. For the enzyme assay (buffer G,) BSA was added to buffer G2 to achieve a final concentration of 0.1%. METHODS Hormone Degradation Assay
Four hundred microliters of hemolysate (or an appropriate aliquot in 400 ~1 of buffer G,) was incubated with 50 pg of ‘251-labeled hormone (insulin, prolactin, hTSH,3 NGF, or EGF) in final volume of 425 ~1 at 37” for 25 min. The reaction was stopped by the addition of 400 ~1 bovine serum albumin (lOOg/liter) and 400 ~1 of 20% trichloroacetic acid. The resulting suspension was mixed and centrifuged, and radioactivity was measured in the supernatant and precipitant fractions. A control for every hormone containing 400 ~1 of buffer G3 instead of hemolysate was treated the same way. Percentage degradation of hormone was expressed as described earlier by Nerurkar and Gambhir (2). The enzyme activity was expressed as insulin degraded per minute either per milligram of protein or per milliliter of incubation mixture. (a) Preparation of cytosol fraction-isolated erythrocytes (13-20 ml) were hemolyzed using 20 vol of buffer G, at 23” for 15 min. Plasma membranes were removed by centrifuging the hemolysate at 35,000g at 4” in a Sorvall RC-5B refrigerated centrifuge using rotor SS-34. The resulting plasma membrane free membrane was used for further purification. (b) Ammonium sulfate fractionation-the hemolysate was saturated to 40% (243 g/liter) with ammonium sulfate at 4” and centrifuged. The precipitate was discarded and the supernate was further saturated to 60% (32 g/liter) with ammonium sulfate. The precipitate thus obtained was dissolved in 2 ml of buffer G2 and was dialyzed against the same buffer at 4” until free of ammonium sulfate. Ten microliters of the dialyzed solution was used for measuring insulin-degrading activity and protein content by the method of Lowry et al. (5). (c) Gel filtration of column chromatography-l ml of dialyzed solution was placed on a preequilibrated Sephadex G-100 column (1.8 x 90 cm) and eluted with buffer G2 at 4”. The eluate, in 2-ml fractions, read for absorbancy at 280 nm for protein and 419 nm for hemoglobin. An appropriate aliquot from each 2-ml fraction of eluate was used for the determination of insulin-degrading activity and protein. Properties of the Enzyme
(a) Specificity of the enzyme was tested by its action on various substrates and comparing it with that of trypsin. 3 Abbreviations used: hTSH, human thyroid stimulating hormone; NGF, nerve growth factor; EGF, epidermal growth factor; BSA, bovine serum albumin; EH, Eadie-Hofstee; LB, LineweaverBurk; IDE, insulin-degrading enzyme.
286
GAMBHIR
AND NERURKAR TABLE
Step Hemolysate (12 ml RBC) WH,),SO, (40-60% fraction dialyzed Sephadex G-100 (column eluate)
Volume (ml)
1
Total activity (w/min)
Total protein (mg)
1350
3938
322 280
Specific activity Wmin/mg)
Yield @)
Purification (-fold)
0.344
100
-
43.1
7.48
23.9
21.7
22.5
12.44
20.7
36.2
250 2 36
(b) Molecular weight of the enzyme was determined by passage of partially purified enzyme at 4” through a Sephadex G-200 (1.5 x 90 cm) column equilibrated with buffer GZ. The column was calibrated with standard proteins: catalase (232,000), aldolase (lSS,OOO), hemoglobin (64,000), and Dextran Blue 2000. (c) Determination of optimum pH was carried out for lysate using buffer G, and for partially purified enzyme using buffer G3. Series of buffers ranging in pH from 6 to 10 were prepared from buffers G, and G3 using 1 N HCl and 1 N NaOH . (d) The K,,, value for insulin was determined with partially purified enzyme (0.06 mg protein/tube). Using 100 pg of ‘251-insulin with O-5 pg of unlabeled insulin as a substrate, the initial velocity of insulin degradation was measured at 37” and pH 7.4 using the trichloroacetic acid precipitation technique. The velocity of insulin degradation was expressed as namograms/milliliter/minute. The data obtained in these experiments were analyzed by Lineweaver-Burk (l/v vs l/(S)) and Eadie-Hofstee (v against v/(S)) plots with the help of a HewlettPackard computer Model 9815 A and a linear regression data cartridge. These plots yielded K,,,, If,,,,, and coefficient of linear relationship (r). (e) Various inhibitors: bacitracin, N-ethylmaleimide, antipain, chloroquine, EDTA, and soybean trypsin inhibitor at different concentrations were used to find out the percentage inhibition of insulin-degrading activity. RESULTS
AND DISCUSSION
The erythrocyte lysate obtained after removal of plasma membranes contained most of the insulin-degrading activity. About 5% of the activity was precipitated in the O-40% ammonium sulfate saturated fraction. A major portion of the enzymatic activity was precipitated at 40-60% ammonium sulfate saturation (Table 1). This fraction (40-60%) was dissolved in 2 ml of buffer G2 and was dialyzed against the same at 4”, with frequent changes of the dialyzing buffer. Gel filtration column chromatography of the above fraction resulted in two major protein peaks. The first peak was eluted slightly after the void volume and contained most of the insulin-degrading activity. The second peak contained hemoglobin and had no insulin-degrading activity. Compared to trypsin, the insulin-degrading enzyme was very specific toward insulin (Table 2). Many proteases have been described in human erythrocytes.
ERYTHROCYTE
INSULIN-DEGRADING
Specificity of Erythrocytic Substrate
(100 w/ml) Insulin Prolactin hTSH NGF EGF Insulin + BSA (1 gm/liter) BSA (2 gm/liter) BSA (3 gm/liter)
TABLE 2 Insulin-Degrading Percentage trypsin 50 50 60 60 50 25 13 8
287
ENZYME
Enzyme (IDE) Degradation IDE
by
50 0 0 0 8 50 48 48
Our preparation can be similar to that reported by Brodal (6) and Kolb et al. (10). It appears that the enzymes described by others (7,9,11) are dissimilar to the enzyme reported in this communication. The enzyme preparation obtained by Kolb et al. (10) had the following differences with our findings. They carried out purification of the enzyme without a reducing agent, a procedure which we could not reproduce. Their claim of stability of enzyme at - 40” was not verifiable using our procedures. They have not reported a & value for insulin nor an optimum pH for the enzyme which we are reporting in this article. On the Sephadex G-200 column, the enzyme was eluted at the same volume as that of aldolase (158,000 Da). The approximate molecular weight was found to be 160,000 Da. The enzyme activity was evaluated over a wide range of pH. The optimal pH of enzymic activity in lysate and purified preparation was found to be in the range of physiological pH. At pH 6.0 and 9.0 enzyme activity was lowered by about 60% (data not shown). The Michaelis-Menten constant (ZC,) value for insulin was determined from the experimental data by two procedures. The first procedure was the LineweaverBurk plot and the second method involved the Eadie-Hofstee plot. On the basis of a K, value for insulin, optimum pH, and molecular weight, the enzyme from human erythrocytes was found to be comparable to those purified from rat adipocytes (12), liver (13), muscle (14), and especially from human muscles (16). On the other hand, placental (15) enzyme had optimum pH in the acidic range. Generally, the Lineweaver-Burk plot is employed to obtain a K, value for insulin. However, we analyzed the data from five different experiments by Lineweaver-Burk and Eadie-Hofstee plots (Table 3). It appears that linearity of the plot is more easily obtained in the former than in the latter. For example, the fourth experiment shows a very good linearity by the LB plot while the EH plot for the same data fails to show such linearity. On the other hand, the third experiment shows excellent linearity in the EH plot and concurrence in K, values by both plots. We are, however, reporting the average K, value
GAMBHIR
288
AND NERURKAR
TABLE 3 Kinetic Data for Insulin-Degrading
Enzyme
(LB) plot Expt. No.
Km (nM)
1
245.7 51.9 123.3 200.2 190.6 162.3
2 3 4 5 Mean
V,,X (ng/ml/min) 17.5 4.4 15.6 5.4 31.0 14.8
(EH) plot K, r”
0.99 0.97 0.99 0.97 0.97 0.98
hM)
179.3 85.1 118.4 73.3 149.8 133.2
V,,X (ng/ml/min) 14.8 5.8 16.5 3.6” 27.8 16.2
r
0.88 0.85 0.95 0.56h 0.84 0.88
a Coefficient of correlation. b These values were not included in averaging EH plot values.
obtained by LB plot to facilitate its comparison to the K, value for IDE from other tissues. The enzymatic activity was inhibited by bacitracin (500 units/ml) and Nethylmaleimide (1 mM), while antipain (1 mM), chloroquine (100 mM), < EDTA (50 mM), and soybean trypson inhibitor (5 mg/ml) failed to do so. These results are similar to those obtained with enzyme purified from other tissues (12-16) and emphasize the importance of -SH groups in maintenance of enzymatic activity. The presence of specific insulin receptors on human erythrocytes with characteristics and properties (17-19) similar to those on other cell types and their diagnostic importance (20) has been established. In this communication we partially purified cytosolic insulin-degrading enzyme from human erythrocytes. The enzyme is shown to have specificity and similar affinity (K, value) for insulin as in other cell types. Recently, we showed decreased insulin-degrading activity in uremics (12). Although these findings suggest that the enzyme has clinical significance, much work is needed to clarify this significance. Liver and kidney are the main organs which clear insulin from the circulation produced centrally by the pancreas. However, it is conceivable that due to the large number of RBCs (approximately 32 x 10”) in circulation, these cells might be responsible for maintaining the basal fasting level of this hormone. SUMMARY Using conventional techniques of ammonium sulfate fractionation and Sephadex gel column chromatography, insulin-degrading enzyme was partially purified from lysate of human erythrocytes. The enzymatic activity was measured by the trichloroacetic acid precipitation method. Compared to trypsin, the enzyme was highly specific for insulin. The apparent molecular weight of the enzyme was 160,000 Da, the optimum pH was the 7.4 to 7.8 range, and the K, value for insulin for the partially purified enzyme was 162 nM. Bacitracin and N-ethylmaleimide were potent inhibitors, while chloroquine, ethylenediaminetetraacetate, antipain, and soybean trypsin inhibitor failed to inhibit the activity of the enzyme.
ERYTHROCYTE
INSULIN-DEGRADING
ENZYME
289
Like most nucleated cells, human erythrocytes not only have the membranal insulin receptors, but also possess the cytosolic specific insulin-degrading enzyme. Insulin internalization and degradation are shown to be due to the receptor and the enzyme acting in concert as in many nucleated cells. Anucleated erythrocytes have both these entities for possible internalization and degradation of insulin. ACKNOWLEDGMENT These investigations were supported by grants from The Charles and Mary Latham Trust Fund, Washington, D.C., and The March of Dimes (Clinical Research Grant No. 6-262). REFERENCES 1. Gambhir, K. K., Nerurkar, S. G.. Das, P. D., Archer, J. A., and Henry, W. L., Jr., Endocrinology 109, 1787 (1981). 2. Nerurkar, S. G., and Gambhir, K. K., Clin. Chem. 27, 607 (1981). 3. Brush, J. S., and Heming, H., Endocrinology 104, 1639 (1981). 4. Gambhir, K. K., Archer, J. A., and Bradley, C. J., Diabetes 27, 701 (1978). 5. Lowery, 0. H., Rosebrough, N. F., Farr, A. L., and Randall, R. J., J. Biol. Chem. 193, 265 (1951). 6. Brodal, N. P., J. Biochem. 18, 201 (1971). 7. Tokes, Z. A., and Chambers, S. M., Biochim. Biophys. Acta, 389, 325 (1975). 8. Pontremoli, S., Salamino, F., Separatore. B., Melloni, E., Morelli, A.. Benatti. U., and DeFlora. A., Biochem. J. 181, 559 (1979). 9. Krischner, R. J., and Goldberg, A. L., J. Biol. Chem. 258, 967 (1980). 10. Kolb, H. J., and Standl, E., Hoppeseyler’s, Z. Physiol. Chem. 361, 1029 (1980). 11. Chandler, M. L., and Varandani, P. T., Diabetes 23, 232 (1974). 12. Goldstein, B. J., and Livingston, J. N., Endocrinology 108, 953 (1981). 13. Brughen, G. A., Kitabchi, A. E., and Brush, J. S., Endocrinology 91, 633 (1972). 14. Duckworth, W. C., and Kitabchi, A. E., Diabetes 23, 536 (1974). 15. Posner, B. I., Diabetes 22, 552 (1973). 16. Neal, G. W., and Kitabchi, A. E., Biochim. Biophys. Acta. 719, 259 (1982). 17. Peterson, S. W., Miller, A. L., Kelleher, R. S.. and Murray, E. F., J. Biol. Chem. 258, 9605 (1983). 18. lm, J. H., Meezan. E., Rockley, C. E., and Kim, H. D., J. Biol. Chem. 258, 5021 (1983). 19. Grigorescu, F., White, M. F., and Kahn, C. R., J. Biol. Chem. 258, 13708 (1983). 20. Gambhir. K. K., Nerurkar, S. G., Cruz, 1. A., and Hosten, A. O., Biochem. Med. 25, 62 (1981). 21. Nerurkar, S. G., Butterfield. S. G., Cruz, 1. A., Hosten, A. O., Dillard, M. G., and Gambhir, K. K., Clin. Res. 31, 393A (1983).