175
Molecular and Cellular Endocrinology, 72 (1990) 175-185 Elsevier Scientific Publishers Ireland. Ltd.
MOLCEL 02344
Characterization of insulin degradation products generated in liver endosomes: in vivo and in vitro studies Jean-Pierre
Clot ‘, Michel Janicot ‘, Franqoise Fouque ‘, Bernard Pierre-Yves Haumont 2 and Florence Lederer 2
Desbuquois
‘,
’ Uniti 30 INSERM, Hapita Necker Enfants Malades, 75015 Paris, France, and ’ Unit4 25 INSERM and WA 122 CNRS, Hapita
Necker Enfants Malades, 75015 Paris, France
(Received 23 January 1990; accepted 11 June 1990)
Key words: Insulin degradation;
Liver endosome
The degradation products generated from Al4 and B26 ‘251-labelled insulins in liver endosomes in vivo and in vitro have been isolated by high-performance liquid chromatography and cleavages in the B chain have been identified by automated radiosequence analysis. In rats sacrificed various times after injection of each of the ‘251-labelled insulins, two major degradation products slightly less hydrophobic than intact iodoinsulins were identified; these accounted, at 8 min, for about 45% (Al4 ‘251-labelled insulin) and 15% (B26 ‘251-labelled insulin) of the total radioactivity recovered, respectively. The products generated from Al4 ‘251-labelled insulin contained an intact A chain, whereas those generated from B26 ‘251-labelled insulin contained a B chain cleaved at the B16-B17 bond. With B26 125I-labelled insulin, two minor products, with cleavages at the B23-B24 and B24-B25 bonds, were also observed. In vivo chloroquine treatment did not alter the nature but caused a decrease in the amount of insulin degradation products associated with endosomes. When endosomal fractions isolated from iodoinsulin injected rats were incubated at 30°C in isotonic KCl, a rapid degradation of iodoinsulin, maximal at pH 6, was observed. With Al4 ‘251-labelled insulin, the two major degradation products identified in vivo were generated along with monoiodotyrosine, but with B26 ‘251-labelled insulin monoiodotyrosine was the main product formed. Addition of ATP, presumably by decreasing the endosomal pH, shifted the medium pH for maximal iodoinsulin degradation to about 7-8. These studies have allowed a direct identification of two previously suggested cleavage sites in the B chain. They have also shown that the degradation products generated in cell-free endosomes under conditions that promote endosomal acidification are similar to those identified in vivo.
Intraduction Address for correspondence: Dr. Bernard Desbuquois, Unite 30 INSERM, H6pitaI Necker Enfants Malades, 149 rue de S&res, 75015 Paris, France. Abbreviations: HPLC: high-performance Liquid chromatography; Mops: 3-( N-morphoIino)propane sulfonic acid; TFA: trifluoroacetic acid; TCA: trichloroacetic acid. Enzyme: Insulin protease, E.C. 3.4.22.11. 0303-7207/90/$03.50
During the past decade, much information has been gained on the mechanisms by which insulin undergoes degradation in the liver, one major site of insulin degradation in the body (Duckworth, 1988; Sonne, 1989). It is generally admitted that
0 1990 Elsevier Scientific Publishers Ireland, Ltd.
176
the initial step involved is the binding of insulin to specific receptors at the hepatocyte surface (Terris and Steiner, 1975). Then, based in part on the use of inhibitors acting at specific subcellular sites, two degradative pathways have been described. One pathway occurs while the insulin-receptor complex is still at the cell surface (Duckworth et al., 1981; Caro et al., 1982; Blackard et al., 1985; Fleig et al., 1986); the other pathway requires prior endocytosis of the complex and occurs intracellularly (Draznin et al., 1981; Duckworth et al., 1981; Caro et al., 1982; Blackard et al., 1986; Levy and Olefsky, 1987). Subcellular fractionation studies have shown that, although initially believed to occur in lysosomes, intracellular degradation occurs mainly in endosomes (Desbuquois et al., 1979; Posner et al., 1980; Pease et al., 1985, 1987; Juul et al., 1986; Hamel et al., 1988). Furthermore, evidence has been presented that the acidity of this organelle, in part by promoting the dissociation of the insulin-receptor complex, plays an essential role in the degradation of internalized ligand (Doherty et al., 1990; Desbuquois et al., manuscript submitted for publication). Using monoiodoinsulins labelled at the Al4 and B26 positions and HPLC techniques, Hamel et al. (1988) have recently isolated and characterized insulin degradation products generated in rat liver endosomes in vivo. Two major products with an intact A chain and a cleaved B chain were identified with the Al4 isomer, and three products with a cleaved B chain were identified with the B26 isomer; cleavages in the B chain were suggested to affect the 16-17, 24-25 and 25-26 bonds. However, the identification of these cleavage sites was based on the comparison of the elution profile of the products to that of the products generated by the enzyme insulin protease, and no direct sequence analysis was carried out. In addition, although insulin degradation has been demonstrated in cell-free endosomes (Pease et al., 1985, 1987; Doherty et al., 1990), the degradation products generated under these conditions have not been characterized. In the present studies, we have analyzed by reverse-phase HPLC the degradation products generated from Al4 and B26 ‘251-labelled insulins in liver endosomes in vivo and in vitro. In addition, using automated radiosequence analysis,
we have identified B chain occurring
the major in vivo.
cleavage
sites in the
Materials and methods Materials Porcine insulin and antiinsulin antibody were from Novo Research Industries (Copenhagen, Denmark). Carrier-free Na”‘I was from Amersham International (U.K.). Lactoperoxidase was from Calbiochem. Bacitracin, N-ethylmaleimide, l,lO-phenanthroline, Mops and ATP were from Sigma. Acetonitrile and TFA of HPLC grade were from Baker Chemical Co. Chloroquine sulfate was from Specia. All other chemicals were of analytical grade. HPLC separations were performed using a Waters model 600 liquid chromatograph equipped with a model U6K sample injector fitted with 1 ml loop, a guard column and an analytical micro bondapak C 18 (0.39 X 30 cm) column. Absorbance at 280 nm was monitored with an LC spectrophotometer, and radioactivity was recorded using a Berthold LB 504 gamma detector connected to an Apple IIe computer. Preparation, purification and characterization of ‘251-labelled insulins Al4 and B26 *251-labelled insulins were prepared by lactoperoxidase-catalyzed iodination in the absence or presence of urea, respectively, as described previously (Jorgensen and Larsen, 1980; Linde et al., 1981; Frank et al., 1983). Typically, 25 pg (3.8 nmol) of porcine insulin was reacted with 2 mCi of 125I (0.8 nmol) in buffer containing or not 6 M urea, in the presence of 2 pg of lactoperoxidase and 2.2 nmol of H202. Incorporaacid-precipitable tion of “‘1 into trichloroacetic radioactivity was about 85-95% for iodination in buffer alone and 50-60% for iodination in the presence of urea. was Purification of the ‘251-labelled insulins achieved by reverse-phase HPLC. After addition of 20 1 of a 20% (v/v) solution of TFA, the entire reaction mixture was injected in the chromatograph and eluted isocratically using a degassed solvent of 30% acetonitrile in 0.1% TFA, pumped at the rate of 1 ml/mm. Under these conditions, the elution times of unreacted insulin and of iodoinsulin isomers were: insulin and Al9 1251-
177
labelled insulin, 29 min; B26 ‘251-labelled insulin, 51 min; B16 ‘251-labelled insulin, 55 min; and Al4 ‘251-labelled insulin, 57 min (not shown). In the absence of urea, about 75% of the radioactivity incorporated into insulin was recovered as Al4 ‘251-labelled insulin, the remaining being recovered as Al9 and B26 isomers (about 15 and lo%, respectively). In the presence of urea, about 22, 28 and 50% of the radioactivity was recovered as A19, B26 and Al4 + B16 monoiodinated isomers, respectively. The identity of the monoiodoinsulin derivatives was established on the basis of previously described methods (Bailey and Cole, 1959; DeZoeten and Havinga, 1961; Jorgensen and Larsen, 1980). Briefly, the monoiodoinsulins were cleaved into the corresponding S-sulfo A and B chains by oxidative sulfitolysis (Bailey and Cole, 1959); sulfonated chains were then hydrolyzed by chymotrypsin and trypsin, respectively, and the hydrolyzates were fractionated by gel filtration on Sephadex G 25 (fine) in 1 M acetic acid. Based on the distribution of the radioactivity between peptides Al-14, A15-21, Bl-22 and B23-29, Al4 monoiodoinsulin was essentially free of contamination by other isomers, whereas B26 monoiodoinsulin contained about 10% of B16 isomer. Animals and injections Male Sprague-Dawley rats weighing 180-200 g were obtained from Charles River France and fasted for 16 h prior to sacrifice. Al4 or B26 ‘251-labelled insulin (20-100 X lo6 &pm) was diluted into 0.5 ml physiological saline containing 1 mg/ml bovine serum albumin and injected over 15 s into the penis vein under light ether anesthesia. Rats were sacrificed 90 s, 4 min and 8 min after injection. In some experiments chloroquine was given as two intraperitoneal injections (2 mg/lOO g body weight each) at 90 and 30 min prior to sacrifice. Subcellular fractionation Golgi-endosomal fractions were isolated from liver homogenates in 0.25 M sucrose according to a modification (Janicot and Desbuquois, 1987) of the method of Ehrenreich et al. (1973); in most cases, a ‘total’ Golgi-endosomal fraction (density, 1.03-1.16) was used. This fraction was about 30-
fold enriched in radioactivity and ATP-dependent acidification activity, and only in endosomal components was the radioactivity present (results not shown). For in vivo studies, 2.5 mM N-ethylmaleimide, 5 mM l,lO-phenanthroline and 1 mg/ml bacitracin were included in the homogenization medium. Each of these compounds is known to inhibit the enzyme insulin protease (Duckworth, 1988), and l,lO-phenantroline has been found to completely block the degradation of 1251iodoinsulin in cell-free endosomes (Doherty et al., 1990; Desbuquois et al., manuscript submitted for publication). For in vitro studies, subcellular fractionation was carried out in the absence of inhibitors. Golgi-endosomal fractions were used immediately after isolation (in vivo studies) or after incubation for 3-12 min at 30°C in 125 mM KCl, 5 mM MgSO, and 25 mM Mops/KOH, pH 5-9, in the presence or absence of 1 mM ATP (in vitro studies). Extraction and characterization of the radioactivity associated with Golgi-endosomal fractions Golgi-endosomal fractions were acidified with acetic acid (30% final concentration) and lyophilized. The dry residue was resuspended in 0.8 ml ice-cold 0.1% TFA (v/v) and 20% acetonitrile (v/v), and centrifuged for 15 min at 30,000 X g; at least 95% of the radioactivity was recovered in the supernatant. The extract was then injected into the column and eluted using as solvents 0.1% TFA (v/v) in water (A) and 0.1% TFA (v/v) in acetonitrile (B) at 2 ml/mm. The column was eluted isocratically for 5 min with 20% solvent B, then with a 50 min linear gradient to 45% solvent B. The major components in the eluate were collected in tubes containing 0.1 mg bovine serum-albumin and lyophilized. The dry residue was dissolved in 0.2 ml of 0.1 M Tris-HCl buffer, pH 7.4 and to this, 5 ~1 of 2-mercaptoethanol (2.5% final concentration) was added. After 30 min at room temperature, 10 ~1 of 10% TFA (v/v) was added, and samples were immediately reinjected into the chromatographic column and eluted as described above. Under these conditions, reduction of both Al4 and B26 monoiodoinsulins was essentially complete and no deiodination occurred. Some reduced degradation products were collected and lyophilized for sequence analysis.
178
In some experiments, the radioactivity associated with Golgi-endosomal fractions was extracted with 0.1 N HCl containing I% (w/v) bovine serum albumin and 0.1% bacitracin (w/v) as described previously (Posner et al., 1982). After centrifugation as above, the supematant was neutralized, and the extracted radioactivity was examined for precipitability by 5% TCA (w/v), binding to ~t~su~n antibody and binding to liver membrane receptors, by reference to control Al4 or B26 ‘251-labelled insulin. Radiosequence analysis of insulin degradation products by automated Edman degradation Automated Edman degradation was carried out using an Applied Biosystems sequenator model 470 A equipped with on-line identification of phenylthiohydantoin amino acids. At each cycle, both the radioactivity released from the insulin degradation products and the phenylthiohydantoin amino acid released from 1 nmol myoglobin added as carrier were monitored.
somal fractions rapidly decreased with time, confirming earlier reports (Desbuquois et al., 1979; Posner et al., 1980). At 4 min, ‘251-labelled insulin integrity relative to noninjected ligand was about 40-50% when estimated by binding to insulin receptors and antiinsulin antibody, and 70% when estimated by TCA precipitation. In the presence of protease inhibitors, the decrease in receptor and antibody binding ability was comparable to that observed in the absence of inhibitors, but the decrease in TCA precipitability was less marked (about 90%). Qualitatively similar results were obtained with B26 ‘251-labelled insulin although the loss of integrity of the latter was less rapid (results not shown). On HPLC analysis, a rapid decrease in the integrity of both Al4 and B26 ‘251-labelled insulins associated with endosomal fractions was also observed (Table 1). Representative elution profiles obtained at 4 min are shown in Fig. 1. In addition to intact i~oinsu~ns, which eluted at 37-38 min, a number of degradation products,
Results Uptake of Al 4 and B26 ‘251-labelled insulins in liver Golgi-endosomat fractions in vivo The uptake of Al4 and B26 ‘251-labelled insulins in Golgi endosomal fractions was kinetically comparable to that described earlier for less homogenous iodoinsulin preparations (Desbuquois et al., 1979; Posner et al., 1980), with a maximum occurring at 4 min. At this time, the radioactivity recovered ranged from 1.5 to 3% of the dose injected per mg protein (l-2% of the dose injected per g liver), and accounted for about 5510% of the radioactivity present in the homogenate. At 90 s, the uptake of B26 monoiodoinsulin exceeded by about 2-fold that of Al4 monoiodoinsulin, but at later times uptake was approximately the same for both isomers. Uptake was little affected by the presence of inhibitors of degradation in the homogenization medium (experiments not shown). Characterizatjon of the radioactivity associated with Golgi-endosomai fractions in Al4 and B26 “?‘Ilabelled insulin-injected rats In the absence of protease inhibitors in the homogenization medium, the integrity of Al4 1251-labelled insulin associated with Golgi-endo-
TABLE 1 RECOVERY OF INTACT INSULIN AND MAJOR INSULIN DEGRADATION PRODUCTS IN LIVER ENDGSOMES PREPARED VARIOUS TIMES AFTER INJECTION OF Al4 AND B26 “sI-LABELLED INSULINS Endosomal extracts were fractionated by HPLC as described in the Materials and Methods section (see Fig. 1). For each monoiodoinsulin isomer, the radioactivity recovered as intact insulin and the two major insulin degradation products was measured and expressed as percent of the radioacti~ty recovered in column eluates. Peaks A-I and A-II (elution times, 32 and 34 min, respectively) and peaks B-I and B-II (elution times, 32 and 33 min, respectively) designate the products generated from Al4 and B26 ‘251-labelled insulins, respectively. The results shown are the mean& SEM of eight (Al4 ‘251-labelled insulin) and three (B26 ‘251-labelled insulin) dete~nations.
‘251-labelled material
Al4-monoiodoinsulin Intact insulin Peak A-I Peak A-II B26-monoiodoinsulin Intact insulin Peak B-I Peak B-II
Time postinjection 90 s
4min
8min
74.7 f 1.9 5.8i0.6 8.1 * 1.2
51.0* 1.4 19.7IfrO.8 19.8 rtO.8
42.7 * 2.1 24.0+ 1.1 22.4& 1.4
81.2 f 2.6 2.5 + 0.3 5.320.5
75.9+ 1.2 3.7*0.4 8.8 * 0.7
14.3 + 2.9 5.2kO.8 11.5*0.9
179
recovered in peaks A-I and A-II, and, conversely, an increase in monoiodotyrosine; the percentage of intact iodoinsulin, though, was unaffected (experiments not shown).
1200
600
: 2 0
2 ‘; 800 .: 0 0 2 400
-
0
1 10
I 20 Elution
I 30 time
I 40
I 50
(min)
Fig. 1. HPLC elution profile of the ‘*‘I-labelled degradation products generated from Al4 and B26 ‘%abc11ed insulins in liver endosomes in vivo. Endosomes were isolated in rats sacrificed 4 min after injection of Al4 (panel A) or B26 (panel B) 1251-labelled insulin. Details of the procedures are given in the Materials and Methods section. Intact Al4 and B26 ‘251labelled insulins (peaks off scale) elute at 37 and 36 min, respectively. The profiles shown arc representative of three (B26 isomer) and eight (Al4 isomer) experiments.
most of which are less hydrophobic than iodoinsulins, were observed. With Al4 ‘251-labelled insulin, the main degradation products eluted as a doublet at 32 and 34 min (peaks A-I and A-II), just in front of intact iodoinsulin. With B26 1251labelled insulin, two peaks, eluting at 32 and 33 min (peaks B-I and B-II), were also prominent, although less abundant than the Al4 doublet; minor peaks eluted at 3 min, in the position of monoiodotyrosine, and at 8 (B-III) and 19 (B-IV) min. The percentage of radioactivity recovered as peaks A-I and A-II (with Al4 ‘*‘I-labelled insulin) and B-I and B-II (with B26 ‘251-labelled insulin) increased with time, reaching a maximum by 8 mm (about 45% and 15% of the total, respectively). With Al4 ‘251-labelled insulin, omission of inhibitors of degradation in the homogenization medium caused a decrease in the amount of radioactivity
Characterization and sequence analysis of the major insulin degradation products generated in Golgi-endosomal fractions In agreement with an earlier study (Hamel et al., 1988), all components identified after injection of Al4 ‘251-labelled insulin eluted exclusively, after disulpbide bond reduction and rechromatography, at the position of intact Al4 ‘251-labelled A chain (retention time, 42 mm). In contrast, among the components identified after injection of B26 1251labelled insulin, only the one corresponding to intact insulin eluted as intact B26 ‘*‘I-labelled B chain after reduction (retention time, 45 mm). Peaks B-I and B-II generated B chain fragments that both eluted at 24 min (results not shown); and the elution of peaks B-III (8 min) and B-IV (19 min) was unaffected by reduction, suggesting that these peptides were not disulphide-linked to any A chain fragment. The peptides derived from B26 ‘251-labelled insulin were submitted to automated Edman degradation to identify the cycle at which [1251]iodotyrosine was released (Fig. 2). Most of the radioactivity associated with reduced peaks B-I and B-II was released at the 10th cycle, indicating a cleavage at position B16-B17. However, about 5-10s of the radioactivity associated with these peptides was released at the 7th cycle. Although the latter observation is compatible with a cleavage at the B19-B20 position, the elution position of the resulting B20-B26+ fragment would not have been affected by reduction. Thus, a more likely explanation is a cleavage at the B9-BlO bond of contaminating B16 ‘251-labelled insulin. In this case, the BlO-B16+ and B17-B26+ peptides may be linked to the same A chain fragment, since before reduction the parent products cannot be resolved. Virtually all the radioactivity associated with peaks B-III and B-IV was released in the second and the third cycle, respectively, suggesting cleavages at the B24-B25 and B23-B24 bonds. Cleavages of contaminating B16 125~labelled insulin at the B14-B15 and B13-B14 bonds can be probably excluded, since the ex-
180
25
P al G z
20
% IO L >, ._ .:
0
: 25 0 G m 20
L
Q 0 : 10 0
*
2
0 Number
of
cycle
Time
Fig. 2. Radioactivity released from B26 ‘251-labe11ed degradation products (see Fig. 1, panel B) upon automated Edman degradation. Panel A, peptide obtained by reduction of the product which elutes at 33 min (peak B-II); panels B and C, products which elute at 8 (peak B-III) and 19 (peak B-IV) min, respectively. The results shown are representative of two experiments. Results identical to those shown in panel A have been obtained with the peptide obtained by reduction of the product which elutes at 32 min (peak B-I).
pected BlSB16 and B14-B16 have been eluted earlier.
4
after
6
injection
8 (
min
)
Fig. 3. Time course of appearance of the major ‘2SI-1abelled degradation products generated from Al4 ‘*sI-labelled insulin (see Fig. 1, panel A) in liver endosomes in control (0) and chloroquine-treated (0) rats. Pane1 A, product which elutes at 32 min (peak A-I); panel B, product which elutes at 34 min (peak A-II). The results shown are the meanf SEM of three experiments. * p < 0.05; * * p -c 0.01.
peptides would
Effect of chtoroquine treatment on the nature of radioactivity associated with Go&-endosomal fractions in iodoinsulin-injected rats Previous studies using gel-filtration and reeeptor binding techniques have shown that, in insulin injected rats, chloroquine treatment causes enhanced accumulation of undegraded insulin in liver endosomal fractions (Posner et al., 1982). To further assess the effects of this drug on the endosomal degradation of insulin, endosomal extracts prepared various times after injection of Al4 “‘Ilabelled insulin in control and chloroquine-treated rats were analyzed by HPLC. Chloroquine treatment did not qualitatively modify the elution profile of the radioactivity (experiments not shown), but caused a decrease in the rate of appearance of
5
6
7
8
95
6
7
8
9
PH Fig. 4. Degradation of A14- (panels A and B) and B26in isolated liver endo(panels C and D) ‘251-monoiodoinsulin somes in the absence (upper panels) or presence (lower panels) of ATP as a function of medium pH. Endosomal fractions were incubated isolated 90 s after injection of ‘251-iodoinsulin in buffered isotonic KC1 for 6 min at 30°C as described in the Materials and Methods section. Insulin integrity was assessed by TCA precipitation (O), binding to antiinsulin antibody (A) and binding to insulin receptors (0) by reference to control hormone. The results shown are the mean of three determinations.
181
the major degradation products, especially that of peak A-I (Fig. 3). In vitro degradation of ‘251-labelled insulins associated with endosomal fractions In agreement with the studies of Pease et al. (1985,1987), a time- and pH-dependent loss in the integrity of Al4 and B26 ‘Z51-monoiodoinsulins occurred when endosomal fractions containing these ligands were incubated at 30°C in buffered isotonic KCl. In time studies, this loss was more pronounced when estimated by receptor and antibody binding (half-life at pH 5.5, about 3 min for Al4 iodoinsulin and 5 min for B26 iodoinsulin) than by TCA precipitation (half-life, about 8 min for both iodoinsulins). In the absence of ATP, the loss of ‘251-iodoinsulin integrity was maximal at pH 5-6, but in the presence of ATP, which causes endosomal acidification (Doherty et al., 1990;
80
.L -
0
: 80
0
0
3
6 Time
9
12
(min)
Fig. 6. HPLC analysis of the t2’I-labelled material generated from B26 ‘2sI-labelled insulin in isolated liver endosomes as a function of time. Iodoinsulin degradation was measured at pH 5.5 in the absence of ATP (panel A) or at pH 8.5 in the presence of ATP (panel E). Symbols indicate recoveries of radioactivity as intact iodoinsulin (O), peaks B-I + B-II (A) and monoiodotyrosine (0). The results shown are the mean f SEM of three experiments.
0
9
6
3 Time
12
(min)
Fig. 5. HPLC analysis of the lZSI-labelled material generated from A14 ‘251-labelled insulin in isolated liver endosomes as a function of time. Endosomal fractions isolated 90 s after iodoinsulin injection were incubated in isotonic KC1 at pH 5.5 in the absence of ATP (panel A) or at pH 8.5 in the presence of ATP (panel B) at 30°C as described in the Materials and Methods section. Incubation mixtures were analyzed by HPLC and the radioactivity recovered as intact iodoinsulin (0). peaks A-I + A-II (A) and monoiodotyrosine (0) was measured. The results shown are the mean f SEM of three experiments.
Desbuquois et al., manuscript submitted for publication), this maximum was shifted to pH 7-8 (Fig. 4). The ATP-induced pH shift was somewhat greater with Al4 iodoinsulin than with B26 iodoinsulin, and was also greater when iodoinsulin integrity was estimated by TCA precipitation than by receptor and antibody binding. In order to characterize the insulin degradation products generated in this cell-free system, endosomal fractions containing Al4 and B26 iodoinsulins were incubated under similar time and pH conditions, extracted and analyzed by HPLC. Besides intact iodoinsulin, both peaks A-I and A-II (with Al4 1251-labelled insulin) and B-I and B-II (with B26 ‘251-labelled insulin) were identified in the eluates; an additional product, seen in vivo mainly with B26 ‘251-iodoinsulin, was monoiodotyrosine (experiments not shown). The distribution of the radioactivity between these various components was first studied as a function of time, both at pH 5.5 in the absence of
182
ATP and at pH 8.5 in the presence of ATP. With Al4 “‘I-iodoinsulin (Fig. 5) a time-dependent decrease in intact iodoinsulin, along with the appearance of peaks A-I and A-II and monoiodotyrosine, were observed; these changes occurred somewhat more rapidly at pH 5.5 in the absence of ATP than at pH 8.5 in the presence of ATP. Peaks A-I and A-II reached a maximum by 9 min (about 30% of the total at pH 5.5), whereas monoiodotyrosine increased linearly with time up to 12 min (about 3% of the total per min at pH 5.5). With B26 lz5I-iodoinsulin (Fig. 6), however, there was a time-dependent decrease in the amount of radioactivity initiahy associated with peaks B-I and B-II, and monoiodotyrosine was the main product generated. Incubations were then carried out for 6 min at various pHs in the absence and in the presence of ATP. With Al4 ‘251-iodoinsulin (Fig. 7), the decrease in intact iodoinsulin and the correlative A
5
6
7
8
9
80
.z c
0
: 80 0 ._ P m L
L
5
6
7
8
9
PH
Fig. 8. HPLC analysis of the ‘251-labelled material generated from B26 ‘251-labelled insulin in isolated liver endosomes as a function of medium pH. Iodoinsulin degradation was measured in the absence (panel A) or presence (panel 8) of ATP as described for A14 ‘251-labelled insulin in the legend to Fig. 6. Symbols indicate recoveries of intact iodoinsulin (0) peaks B-I+B-II (A) and monoiodotyrosine (0). The results shown are the mean f SEM of three experiments.
increase in monoiodotyrosine were maximal at pH 6, and the increase in peaks A-I and A-II was maximum at pH 5; addition of ATP shifted the corresponding pH values to about 7 and 6, respectively. With B26 ‘251-labelled insulin (Fig. 8) the pH dependence for insulin disappearance and monoiodotyrosine appearance was comparable to that observed with the Al4 isomer, and so was the ATP-induced shift of the corresponding optimum pH. However, the small amount of peaks B-I and B-II present was maximal at pH 5 (or below) and this maximum was little affected by ATP.
tJH Fig. 7. HPLC analysis of the ‘251-labelled material generated from Al4 ‘251-labelled insulin in isolated liver endosomes as a function of medium pH. Endosomal fractions isolated 90 s after iodoinsulin injection were incubated in buffered isotonic KC1 in the absence (panel A) or presence (panel B) of ATP for 6 mitt at 30°C as described in the Materials and Methods section. Incubation mixtures were analyzed by HPLC and recoveries of intact iodoinsulin (o), peaks A-I+A-II (A) and monoiodotyrosine (0) were measured. The results shown are the mean f SEM of three experiments.
Discussion
One first goal of these studies was to isolate and chemically characterize the degradation products generated from Al4 and B26 i2?-labelled insulins in liver endosomes in vivo. To achieve this, the approach devised by Hamel et al. (1988) was used with two major modifications. First, livers were homogenized in the presence of inhibi-
183
tom of proteases that, unlike those used by IIamel et al. (1988), effectively blocked insulin degradation during subcellular fractionation. Second, and most importantly, a direct identification of the products generated from B26 ‘*?-labelled insulin was achieved using automated Edman degradation. The HPLC elution pattern of iodoinsulin degradation products observed in this study was comparable to that described by Hamel et al. (1988), except for the absence of late eluting components with the Al4 isomer. With both iodoinsulins, two prominent products slightly less hydrophobic than insulin were identified; the products generated from the Al4 isomer eluted as a doublet and were more abundant than those generated from the B26 isomer. However, at comparable times after injection, the relative abundance of these products exceeded by at least 5-fold that observed by Hamel et al. (1988), and concurrently, much less iodotyrosine and small iodotyrosine peptides were present. This presumably reflects our use of l,lO-phen~t~o~e, which, unlike the inhibitors used by I-&me1 et al. (1988), effectively blocked the degradation of the iodoinsulin associated with endosomes during subcellular fractionation. In addition to the major products, two minor early eluted products were identified with B26 125I-labelled insulin. After reduction and rechromatography, all products generated from Al4 ‘251-labelled insulin eluted at the position of intact A chain. In contrast, with B26 ‘*?-labelled insulin, the two major products gave rise, after reduction, to a single peptide which eluted differently from intact B chain; the elution of the two minor products was unaffected by reduction and also differed from that of intact B chain. These results extend those reported by Hamel et al. (1988) and suggest the existence of at least three cleavages in the B chain, two of which occur ~boxy-te~n~ly to Cys-19. However, the apparent integrity of the A chain in studies using Al4 ‘251-labelled insulin obviously does not exclude cleavages in this chain as well. Indeed, monoiodotyrosine, one major product of A chain degradation in hepatocytes (Duckworth et al., 1988), might have rapidly diffused out of the endosomes once formed, and thus escaped detection. Furthermore, our studies with cell-free endo-
somes indicate that, in the absence of inhibitors of degradation, iodotyrosine is effectively generated. Using automated Edman degradation, we have found that the B chain peptide derived from the major degradation products of B26 ‘*‘I-labelled insulin is cleaved at the B16-B17 bond, and that the two minor products are cleaved at the B23-B24 and B24-B25 bonds, respectively. Since the two major products generate the same B16-B26+ peptide, they may differ either at the A chain or at the B chain sequence located amino-terminally to the B16 residue, in accordance with the sequence of proteolytic events suggested by Assoian and Tager (1982). Thus, as suggested from comparative HPLC elution profiles of the products generated in endosomes and of those generated by insulin protease (Hamel et al., 1988), two cleavages in the B chain affect the B16-B17 and B24-B25 bonds. However, contrary to expectation, the third cleavage was found to affect the B23-B24 and not the B25-B26 bond; the reasons for this discrepancy are not obvious. Chloroquine, an acidotropic drug, has previously been reported to cause enhanced accumulation of undegraded insulin in the endosomal compartment, whether injected intraperitoneally (Posner et al., 1982) or perfused intraportally (Smith et al., 1989). Our studies show that, in agreement with the studies of Smith et al. (1989), chloroquine treatment causes a decrease in the amount of the degradation products generated from Al4 and B26 12’I-labelled insulins, but does not qualitatively alter the elution of the products. At variance with these results, Hamel et al. (1987) have reported that, in isolated hepatocytes incubated with Al4 ‘251-labelled insulin, chloroquine treatment leads to the appearance of a degradation product more hydrophobic than the primary products found in control cells, while concomitantly blocking the formation of these products; all products contained an intact A chain. However, in their study hepatocytes were continuously exposed to insulin for 30 min, and total intracellular products rather than endosomal products were analyzed. Previous studies have shown that when endosomal fractions isolated from insulin-injected animals are incubated in vitro at 37°C under isotonic conditions, the insulin associated with
184
these fractions undergoes rapid degradation as judged from TCA precipitability, with a maximal at pH 6 (Pease et al., 1985, 1987). The present studies confirm these observations and show that insulin degradation occurs to a greater extent when estimated by binding to insulin receptors, binding to antiinsulin antibody and HPLC. These studies also show that the degradation products generated from the i~o~su~s in cell-free endosomes are qu~~tively similar to those identified in vivo. However, unlike observed in vivo, a large fraction of degraded iodoinsulin was recovered as monoiodotyrosine, especially with the B26 isomer. Differences between in viva and in vitro results may be explained by a rapid diffusion of monoiodotyrosine out of the endosomes in the in vivo studies. The present studies show that ATP, by decreasing the endosomal pH (Desbuquois et al., manuscript submitted for publication), elevates the medium pH required for optimum iodoinsulin degradation from 5-6 to 7-8. Although the ATPinduced pH shift was detectable regardless of the assay, pHs required for maximal degradation somewhat varied depending on the parameter measured. Thus, in HPLC studies, the pH values at which the generation of intermediate products was maximal were at least 1 unit lower than the pH at which the generation of iodotyrosine and the disappearance of insulin were maximal, whether ATP was present or not; this was especially marked for B26 ‘2sI-labelled insulin This suggests that several enzymes with different pH optima may be involved in insulin degradation; alternatively, the pH optimum for peptide bond cleavage by the insulin-degrading enzyme(s) may vary depending on the particular bond cleaved. In addition, since receptor-bound insulin can be degraded by insulin protease (Yonezawa et al., 1988) and since one consequence of lowering the endosomal pH is to promote the dissociation of insulin from its receptor (Doherty et al., 1990; Desbuquois et al., manuscript submitted for publication), the possibility that receptor-bound and dissociated insulin may be degraded differently, or by different enzymes, must also be considered. Based on the finding that the insulin degradation products generated from B26 ‘251-labelled insulin in liver endosomes (Hamel et al., 1988>,
like those generated by isolated hepatocytes (Duckworth et al., 1988), are similar to the products generated by insulin protease, it has been suggested that this enzyme may be responsible for insulin degradation in endosomes (Duckworth, 1988; Hamel et al., 1988). However, studies with the Al4 isomer have shown that while the insulin degradation products identified in hepatocyte media (Pell et al., 1986; Hamel et al., 1987) resemble those generated by insulin protease, this is not the case for the products identified in endosomes (Hamel et al., 1988) and hepatocyte extracts (Pell et al., 1986; Hamel et al., 1987). In addition, it is unknown whether insulin protease, an enzyme which is primarily cytosolic (Duckworth, 1988), is also present in endosomes. In summary, the insulin degradation products generated in liver endosomes in vivo have been isolated and characterized, and cleavages in the B chain at the B16-B17 and B24-B25 bonds, suggested to occur in a recent study (Hamel et al., 1988), have been firmly identified. In addition, the insulin degradation products generated in cell-free endosomes under conditions that promote endosomal acidification have been shown to be qualitatively similar to those identified in vivo. References Assoian, R.K. and Tager, H.S. (1982) J. Biol. Chem. 257, 9078-9085. Bailey, J.L. and Cole, R.D. (1959) J. Biol. Chem. 234, 17331739. Biackard, W.G., Ludeman, C. and Stillman, J. (1985) Am. J. Physiol. 248, El94E202. Blackard, W.G., Smith, R.M. and Jarett, L. (1986) Am. J. Physiol. 250, E148-E156. Caro, J.F., Muller, G. and Gennon, J.A. (1982) J. Biol. Chem. 257, 8459-8466. Desbuquois, B., Wiileput, J. and Huet de Froberville, A. (1979) FEBS Lett. 106, 338-344. DeZoeten, L.W. and Havinga, E. (1961) Rec. Trav. Chim. 80, 917-926. Doherty, II, J.J., Kay, D.G., Lai, W.H., Posner, B.I. and Bergeron, J.J.M. (1990) J. Cell Biol. 110, 35-42. Draznin, B., Solomons, CC., Toothaker, D.R. and Sussman, K.E. (1981) Endocrinology 1108, 8-17. Duckworth, W.C. (1988) Endocr. Rev. 9, 319-345. Duckworth, W.C., Runyan, K.R., Wright, R.K., Halban, P.A. and Solomon, S.S. (1981) Endocrinology 108, 1142-1147. Duckworth, W.C., Hamel, F.G., Peavy, DE., Liepnick, J.J., Ryan, M.P., Hem&son, M.A. and Frank, B.H. (1988) J. Biol. Chem. 263, 1826-1833.
185 Ehrenreich, J.H., Bergeron, J.J.M., Siekevitz, P. and Palade, G.E. (1973) J. Cell. Biol. 59, 45-72. Fieig, W.E., Hess, G., Nother-Fleig, G. and Ditschuneit, H. (1986) B&hem. J. 237, 99-104. Frank, B.H., Peavy, D.E., Hooker, C.S. and Duckworth, W.C. (1983) Diabetes 32, 705-711. Hamel, F.G., Peavy, D.E., Ryan, M.P. and Duckworth, W.C. (1987) Diabetes 36, 702-708. Hamel, F.G., Posner, B.I., Bergeron, J.J.M., Frank, B.H. and Duckworth, W.C. (1988) J. Biol. Chem. 263, 6703-6708. Janicot, M. and Desbuquois, B. (1987) Eur. J. B&hem. 163, 433-442. Jorgensen, K.H. and Larsen, U.D. (1980) Diabetologia 19, 546-554. Juul, S.M., Jones, R.H., Evans, J.L., Neffe, J., Sonksen, P.H. and Brandenburg, D. (1986) B&him. Biophys. Acta 856, 310-319. Levy, J.R. and Olefsky, J.M. (1987) Endocrinology 121, 20752086. Linde, S., Sonne, 0.. Hansen, B. and Ghemann, J. (1981) Hoppe-Seyler’s Z. Physiol. Chem. 362, 573-579.
Pease, R.J., Smith, G.D. and Peters, T.J. (1985) B&hem. J. 228, 137-146. Pease, R.J., Smith, G.D. and Peters, T.J. (1987) Eur. J. Biothem. 164, 251-257. Peh, M.E., Duckworth, W.C. and Peavy, D.E. (1986) B&hem. Biophys. Res. Commun. 137, 1034-1040. Posner, B.I., Patel, B.A., Verma, A.K. and Bergeron, J.J.M. (1980) J. Biol. Chem. 255, 735-741. Posner, B.I., Patel, B.A., Khan, M.N. and Bergeron, J.J.M. (1982) J. Biol. Chem. 257, 5789-5799. Smith, G.D., Christensen, J.R., Rideout, J.M. and Peters, T.J. (1989) Eur. J. B&hem. 181, 287-294. Sonne, 0. (1989) Physiol. Rev. 68, 1129-1196. Ten-is, S. and Steiner, D.F. (1975) J. Biol. Chem. 250, 83898398. Yonezawa, K., Yokono, K., Shii, K., Hari, J., Amano, K., Sakamoto, T., Kawase, Y., Akiyama, H., Nagata, M. and Baba, S. (1988) B&hem. Biophys. Res. Commun. 150, 605-614.
Molecular and Cellular Endocrinology, 72 (1990) 175-185 Elsevier Scientific Publishers Ireland. Ltd.
MOLCEL 02344
Characterization of insulin degradation products generated in liver endosomes: in vivo and in vitro studies Jean-Pierre
Clot ‘, Michel Janicot ‘, Franqoise Fouque ‘, Bernard Pierre-Yves Haumont 2 and Florence Lederer 2
Desbuquois
‘,
’ Uniti 30 INSERM, Hapita Necker Enfants Malades, 75015 Paris, France, and ’ Unit4 25 INSERM and WA 122 CNRS, Hapita
Necker Enfants Malades, 75015 Paris, France
(Received 23 January 1990; accepted 11 June 1990)
Key words: Insulin degradation;
Liver endosome
The degradation products generated from Al4 and B26 ‘251-labelled insulins in liver endosomes in vivo and in vitro have been isolated by high-performance liquid chromatography and cleavages in the B chain have been identified by automated radiosequence analysis. In rats sacrificed various times after injection of each of the ‘251-labelled insulins, two major degradation products slightly less hydrophobic than intact iodoinsulins were identified; these accounted, at 8 min, for about 45% (Al4 ‘251-labelled insulin) and 15% (B26 ‘251-labelled insulin) of the total radioactivity recovered, respectively. The products generated from Al4 ‘251-labelled insulin contained an intact A chain, whereas those generated from B26 ‘251-labelled insulin contained a B chain cleaved at the B16-B17 bond. With B26 125I-labelled insulin, two minor products, with cleavages at the B23-B24 and B24-B25 bonds, were also observed. In vivo chloroquine treatment did not alter the nature but caused a decrease in the amount of insulin degradation products associated with endosomes. When endosomal fractions isolated from iodoinsulin injected rats were incubated at 30°C in isotonic KCl, a rapid degradation of iodoinsulin, maximal at pH 6, was observed. With Al4 ‘251-labelled insulin, the two major degradation products identified in vivo were generated along with monoiodotyrosine, but with B26 ‘251-labelled insulin monoiodotyrosine was the main product formed. Addition of ATP, presumably by decreasing the endosomal pH, shifted the medium pH for maximal iodoinsulin degradation to about 7-8. These studies have allowed a direct identification of two previously suggested cleavage sites in the B chain. They have also shown that the degradation products generated in cell-free endosomes under conditions that promote endosomal acidification are similar to those identified in vivo.
Intraduction Address for correspondence: Dr. Bernard Desbuquois, Unite 30 INSERM, H6pitaI Necker Enfants Malades, 149 rue de S&res, 75015 Paris, France. Abbreviations: HPLC: high-performance Liquid chromatography; Mops: 3-( N-morphoIino)propane sulfonic acid; TFA: trifluoroacetic acid; TCA: trichloroacetic acid. Enzyme: Insulin protease, E.C. 3.4.22.11. 0303-7207/90/$03.50
During the past decade, much information has been gained on the mechanisms by which insulin undergoes degradation in the liver, one major site of insulin degradation in the body (Duckworth, 1988; Sonne, 1989). It is generally admitted that
0 1990 Elsevier Scientific Publishers Ireland, Ltd.
176
the initial step involved is the binding of insulin to specific receptors at the hepatocyte surface (Terris and Steiner, 1975). Then, based in part on the use of inhibitors acting at specific subcellular sites, two degradative pathways have been described. One pathway occurs while the insulin-receptor complex is still at the cell surface (Duckworth et al., 1981; Caro et al., 1982; Blackard et al., 1985; Fleig et al., 1986); the other pathway requires prior endocytosis of the complex and occurs intracellularly (Draznin et al., 1981; Duckworth et al., 1981; Caro et al., 1982; Blackard et al., 1986; Levy and Olefsky, 1987). Subcellular fractionation studies have shown that, although initially believed to occur in lysosomes, intracellular degradation occurs mainly in endosomes (Desbuquois et al., 1979; Posner et al., 1980; Pease et al., 1985, 1987; Juul et al., 1986; Hamel et al., 1988). Furthermore, evidence has been presented that the acidity of this organelle, in part by promoting the dissociation of the insulin-receptor complex, plays an essential role in the degradation of internalized ligand (Doherty et al., 1990; Desbuquois et al., manuscript submitted for publication). Using monoiodoinsulins labelled at the Al4 and B26 positions and HPLC techniques, Hamel et al. (1988) have recently isolated and characterized insulin degradation products generated in rat liver endosomes in vivo. Two major products with an intact A chain and a cleaved B chain were identified with the Al4 isomer, and three products with a cleaved B chain were identified with the B26 isomer; cleavages in the B chain were suggested to affect the 16-17, 24-25 and 25-26 bonds. However, the identification of these cleavage sites was based on the comparison of the elution profile of the products to that of the products generated by the enzyme insulin protease, and no direct sequence analysis was carried out. In addition, although insulin degradation has been demonstrated in cell-free endosomes (Pease et al., 1985, 1987; Doherty et al., 1990), the degradation products generated under these conditions have not been characterized. In the present studies, we have analyzed by reverse-phase HPLC the degradation products generated from Al4 and B26 ‘251-labelled insulins in liver endosomes in vivo and in vitro. In addition, using automated radiosequence analysis,
we have identified B chain occurring
the major in vivo.
cleavage
sites in the
Materials and methods Materials Porcine insulin and antiinsulin antibody were from Novo Research Industries (Copenhagen, Denmark). Carrier-free Na”‘I was from Amersham International (U.K.). Lactoperoxidase was from Calbiochem. Bacitracin, N-ethylmaleimide, l,lO-phenanthroline, Mops and ATP were from Sigma. Acetonitrile and TFA of HPLC grade were from Baker Chemical Co. Chloroquine sulfate was from Specia. All other chemicals were of analytical grade. HPLC separations were performed using a Waters model 600 liquid chromatograph equipped with a model U6K sample injector fitted with 1 ml loop, a guard column and an analytical micro bondapak C 18 (0.39 X 30 cm) column. Absorbance at 280 nm was monitored with an LC spectrophotometer, and radioactivity was recorded using a Berthold LB 504 gamma detector connected to an Apple IIe computer. Preparation, purification and characterization of ‘251-labelled insulins Al4 and B26 *251-labelled insulins were prepared by lactoperoxidase-catalyzed iodination in the absence or presence of urea, respectively, as described previously (Jorgensen and Larsen, 1980; Linde et al., 1981; Frank et al., 1983). Typically, 25 pg (3.8 nmol) of porcine insulin was reacted with 2 mCi of 125I (0.8 nmol) in buffer containing or not 6 M urea, in the presence of 2 pg of lactoperoxidase and 2.2 nmol of H202. Incorporaacid-precipitable tion of “‘1 into trichloroacetic radioactivity was about 85-95% for iodination in buffer alone and 50-60% for iodination in the presence of urea. was Purification of the ‘251-labelled insulins achieved by reverse-phase HPLC. After addition of 20 1 of a 20% (v/v) solution of TFA, the entire reaction mixture was injected in the chromatograph and eluted isocratically using a degassed solvent of 30% acetonitrile in 0.1% TFA, pumped at the rate of 1 ml/mm. Under these conditions, the elution times of unreacted insulin and of iodoinsulin isomers were: insulin and Al9 1251-
177
labelled insulin, 29 min; B26 ‘251-labelled insulin, 51 min; B16 ‘251-labelled insulin, 55 min; and Al4 ‘251-labelled insulin, 57 min (not shown). In the absence of urea, about 75% of the radioactivity incorporated into insulin was recovered as Al4 ‘251-labelled insulin, the remaining being recovered as Al9 and B26 isomers (about 15 and lo%, respectively). In the presence of urea, about 22, 28 and 50% of the radioactivity was recovered as A19, B26 and Al4 + B16 monoiodinated isomers, respectively. The identity of the monoiodoinsulin derivatives was established on the basis of previously described methods (Bailey and Cole, 1959; DeZoeten and Havinga, 1961; Jorgensen and Larsen, 1980). Briefly, the monoiodoinsulins were cleaved into the corresponding S-sulfo A and B chains by oxidative sulfitolysis (Bailey and Cole, 1959); sulfonated chains were then hydrolyzed by chymotrypsin and trypsin, respectively, and the hydrolyzates were fractionated by gel filtration on Sephadex G 25 (fine) in 1 M acetic acid. Based on the distribution of the radioactivity between peptides Al-14, A15-21, Bl-22 and B23-29, Al4 monoiodoinsulin was essentially free of contamination by other isomers, whereas B26 monoiodoinsulin contained about 10% of B16 isomer. Animals and injections Male Sprague-Dawley rats weighing 180-200 g were obtained from Charles River France and fasted for 16 h prior to sacrifice. Al4 or B26 ‘251-labelled insulin (20-100 X lo6 &pm) was diluted into 0.5 ml physiological saline containing 1 mg/ml bovine serum albumin and injected over 15 s into the penis vein under light ether anesthesia. Rats were sacrificed 90 s, 4 min and 8 min after injection. In some experiments chloroquine was given as two intraperitoneal injections (2 mg/lOO g body weight each) at 90 and 30 min prior to sacrifice. Subcellular fractionation Golgi-endosomal fractions were isolated from liver homogenates in 0.25 M sucrose according to a modification (Janicot and Desbuquois, 1987) of the method of Ehrenreich et al. (1973); in most cases, a ‘total’ Golgi-endosomal fraction (density, 1.03-1.16) was used. This fraction was about 30-
fold enriched in radioactivity and ATP-dependent acidification activity, and only in endosomal components was the radioactivity present (results not shown). For in vivo studies, 2.5 mM N-ethylmaleimide, 5 mM l,lO-phenanthroline and 1 mg/ml bacitracin were included in the homogenization medium. Each of these compounds is known to inhibit the enzyme insulin protease (Duckworth, 1988), and l,lO-phenantroline has been found to completely block the degradation of 1251iodoinsulin in cell-free endosomes (Doherty et al., 1990; Desbuquois et al., manuscript submitted for publication). For in vitro studies, subcellular fractionation was carried out in the absence of inhibitors. Golgi-endosomal fractions were used immediately after isolation (in vivo studies) or after incubation for 3-12 min at 30°C in 125 mM KCl, 5 mM MgSO, and 25 mM Mops/KOH, pH 5-9, in the presence or absence of 1 mM ATP (in vitro studies). Extraction and characterization of the radioactivity associated with Golgi-endosomal fractions Golgi-endosomal fractions were acidified with acetic acid (30% final concentration) and lyophilized. The dry residue was resuspended in 0.8 ml ice-cold 0.1% TFA (v/v) and 20% acetonitrile (v/v), and centrifuged for 15 min at 30,000 X g; at least 95% of the radioactivity was recovered in the supernatant. The extract was then injected into the column and eluted using as solvents 0.1% TFA (v/v) in water (A) and 0.1% TFA (v/v) in acetonitrile (B) at 2 ml/mm. The column was eluted isocratically for 5 min with 20% solvent B, then with a 50 min linear gradient to 45% solvent B. The major components in the eluate were collected in tubes containing 0.1 mg bovine serum-albumin and lyophilized. The dry residue was dissolved in 0.2 ml of 0.1 M Tris-HCl buffer, pH 7.4 and to this, 5 ~1 of 2-mercaptoethanol (2.5% final concentration) was added. After 30 min at room temperature, 10 ~1 of 10% TFA (v/v) was added, and samples were immediately reinjected into the chromatographic column and eluted as described above. Under these conditions, reduction of both Al4 and B26 monoiodoinsulins was essentially complete and no deiodination occurred. Some reduced degradation products were collected and lyophilized for sequence analysis.
178
In some experiments, the radioactivity associated with Golgi-endosomal fractions was extracted with 0.1 N HCl containing I% (w/v) bovine serum albumin and 0.1% bacitracin (w/v) as described previously (Posner et al., 1982). After centrifugation as above, the supematant was neutralized, and the extracted radioactivity was examined for precipitability by 5% TCA (w/v), binding to ~t~su~n antibody and binding to liver membrane receptors, by reference to control Al4 or B26 ‘251-labelled insulin. Radiosequence analysis of insulin degradation products by automated Edman degradation Automated Edman degradation was carried out using an Applied Biosystems sequenator model 470 A equipped with on-line identification of phenylthiohydantoin amino acids. At each cycle, both the radioactivity released from the insulin degradation products and the phenylthiohydantoin amino acid released from 1 nmol myoglobin added as carrier were monitored.
somal fractions rapidly decreased with time, confirming earlier reports (Desbuquois et al., 1979; Posner et al., 1980). At 4 min, ‘251-labelled insulin integrity relative to noninjected ligand was about 40-50% when estimated by binding to insulin receptors and antiinsulin antibody, and 70% when estimated by TCA precipitation. In the presence of protease inhibitors, the decrease in receptor and antibody binding ability was comparable to that observed in the absence of inhibitors, but the decrease in TCA precipitability was less marked (about 90%). Qualitatively similar results were obtained with B26 ‘251-labelled insulin although the loss of integrity of the latter was less rapid (results not shown). On HPLC analysis, a rapid decrease in the integrity of both Al4 and B26 ‘251-labelled insulins associated with endosomal fractions was also observed (Table 1). Representative elution profiles obtained at 4 min are shown in Fig. 1. In addition to intact i~oinsu~ns, which eluted at 37-38 min, a number of degradation products,
Results Uptake of Al 4 and B26 ‘251-labelled insulins in liver Golgi-endosomat fractions in vivo The uptake of Al4 and B26 ‘251-labelled insulins in Golgi endosomal fractions was kinetically comparable to that described earlier for less homogenous iodoinsulin preparations (Desbuquois et al., 1979; Posner et al., 1980), with a maximum occurring at 4 min. At this time, the radioactivity recovered ranged from 1.5 to 3% of the dose injected per mg protein (l-2% of the dose injected per g liver), and accounted for about 5510% of the radioactivity present in the homogenate. At 90 s, the uptake of B26 monoiodoinsulin exceeded by about 2-fold that of Al4 monoiodoinsulin, but at later times uptake was approximately the same for both isomers. Uptake was little affected by the presence of inhibitors of degradation in the homogenization medium (experiments not shown). Characterizatjon of the radioactivity associated with Golgi-endosomai fractions in Al4 and B26 “?‘Ilabelled insulin-injected rats In the absence of protease inhibitors in the homogenization medium, the integrity of Al4 1251-labelled insulin associated with Golgi-endo-
TABLE 1 RECOVERY OF INTACT INSULIN AND MAJOR INSULIN DEGRADATION PRODUCTS IN LIVER ENDGSOMES PREPARED VARIOUS TIMES AFTER INJECTION OF Al4 AND B26 “sI-LABELLED INSULINS Endosomal extracts were fractionated by HPLC as described in the Materials and Methods section (see Fig. 1). For each monoiodoinsulin isomer, the radioactivity recovered as intact insulin and the two major insulin degradation products was measured and expressed as percent of the radioacti~ty recovered in column eluates. Peaks A-I and A-II (elution times, 32 and 34 min, respectively) and peaks B-I and B-II (elution times, 32 and 33 min, respectively) designate the products generated from Al4 and B26 ‘251-labelled insulins, respectively. The results shown are the mean& SEM of eight (Al4 ‘251-labelled insulin) and three (B26 ‘251-labelled insulin) dete~nations.
‘251-labelled material
Al4-monoiodoinsulin Intact insulin Peak A-I Peak A-II B26-monoiodoinsulin Intact insulin Peak B-I Peak B-II
Time postinjection 90 s
4min
8min
74.7 f 1.9 5.8i0.6 8.1 * 1.2
51.0* 1.4 19.7IfrO.8 19.8 rtO.8
42.7 * 2.1 24.0+ 1.1 22.4& 1.4
81.2 f 2.6 2.5 + 0.3 5.320.5
75.9+ 1.2 3.7*0.4 8.8 * 0.7
14.3 + 2.9 5.2kO.8 11.5*0.9
179
recovered in peaks A-I and A-II, and, conversely, an increase in monoiodotyrosine; the percentage of intact iodoinsulin, though, was unaffected (experiments not shown).
1200
600
: 2 0
2 ‘; 800 .: 0 0 2 400
-
0
1 10
I 20 Elution
I 30 time
I 40
I 50
(min)
Fig. 1. HPLC elution profile of the ‘*‘I-labelled degradation products generated from Al4 and B26 ‘%abc11ed insulins in liver endosomes in vivo. Endosomes were isolated in rats sacrificed 4 min after injection of Al4 (panel A) or B26 (panel B) 1251-labelled insulin. Details of the procedures are given in the Materials and Methods section. Intact Al4 and B26 ‘251labelled insulins (peaks off scale) elute at 37 and 36 min, respectively. The profiles shown arc representative of three (B26 isomer) and eight (Al4 isomer) experiments.
most of which are less hydrophobic than iodoinsulins, were observed. With Al4 ‘251-labelled insulin, the main degradation products eluted as a doublet at 32 and 34 min (peaks A-I and A-II), just in front of intact iodoinsulin. With B26 1251labelled insulin, two peaks, eluting at 32 and 33 min (peaks B-I and B-II), were also prominent, although less abundant than the Al4 doublet; minor peaks eluted at 3 min, in the position of monoiodotyrosine, and at 8 (B-III) and 19 (B-IV) min. The percentage of radioactivity recovered as peaks A-I and A-II (with Al4 ‘*‘I-labelled insulin) and B-I and B-II (with B26 ‘251-labelled insulin) increased with time, reaching a maximum by 8 mm (about 45% and 15% of the total, respectively). With Al4 ‘251-labelled insulin, omission of inhibitors of degradation in the homogenization medium caused a decrease in the amount of radioactivity
Characterization and sequence analysis of the major insulin degradation products generated in Golgi-endosomal fractions In agreement with an earlier study (Hamel et al., 1988), all components identified after injection of Al4 ‘251-labelled insulin eluted exclusively, after disulpbide bond reduction and rechromatography, at the position of intact Al4 ‘251-labelled A chain (retention time, 42 mm). In contrast, among the components identified after injection of B26 1251labelled insulin, only the one corresponding to intact insulin eluted as intact B26 ‘*‘I-labelled B chain after reduction (retention time, 45 mm). Peaks B-I and B-II generated B chain fragments that both eluted at 24 min (results not shown); and the elution of peaks B-III (8 min) and B-IV (19 min) was unaffected by reduction, suggesting that these peptides were not disulphide-linked to any A chain fragment. The peptides derived from B26 ‘251-labelled insulin were submitted to automated Edman degradation to identify the cycle at which [1251]iodotyrosine was released (Fig. 2). Most of the radioactivity associated with reduced peaks B-I and B-II was released at the 10th cycle, indicating a cleavage at position B16-B17. However, about 5-10s of the radioactivity associated with these peptides was released at the 7th cycle. Although the latter observation is compatible with a cleavage at the B19-B20 position, the elution position of the resulting B20-B26+ fragment would not have been affected by reduction. Thus, a more likely explanation is a cleavage at the B9-BlO bond of contaminating B16 ‘251-labelled insulin. In this case, the BlO-B16+ and B17-B26+ peptides may be linked to the same A chain fragment, since before reduction the parent products cannot be resolved. Virtually all the radioactivity associated with peaks B-III and B-IV was released in the second and the third cycle, respectively, suggesting cleavages at the B24-B25 and B23-B24 bonds. Cleavages of contaminating B16 125~labelled insulin at the B14-B15 and B13-B14 bonds can be probably excluded, since the ex-
180
25
P al G z
20
% IO L >, ._ .:
0
: 25 0 G m 20
L
Q 0 : 10 0
*
2
0 Number
of
cycle
Time
Fig. 2. Radioactivity released from B26 ‘251-labe11ed degradation products (see Fig. 1, panel B) upon automated Edman degradation. Panel A, peptide obtained by reduction of the product which elutes at 33 min (peak B-II); panels B and C, products which elute at 8 (peak B-III) and 19 (peak B-IV) min, respectively. The results shown are representative of two experiments. Results identical to those shown in panel A have been obtained with the peptide obtained by reduction of the product which elutes at 32 min (peak B-I).
pected BlSB16 and B14-B16 have been eluted earlier.
4
after
6
injection
8 (
min
)
Fig. 3. Time course of appearance of the major ‘2SI-1abelled degradation products generated from Al4 ‘*sI-labelled insulin (see Fig. 1, panel A) in liver endosomes in control (0) and chloroquine-treated (0) rats. Pane1 A, product which elutes at 32 min (peak A-I); panel B, product which elutes at 34 min (peak A-II). The results shown are the meanf SEM of three experiments. * p < 0.05; * * p -c 0.01.
peptides would
Effect of chtoroquine treatment on the nature of radioactivity associated with Go&-endosomal fractions in iodoinsulin-injected rats Previous studies using gel-filtration and reeeptor binding techniques have shown that, in insulin injected rats, chloroquine treatment causes enhanced accumulation of undegraded insulin in liver endosomal fractions (Posner et al., 1982). To further assess the effects of this drug on the endosomal degradation of insulin, endosomal extracts prepared various times after injection of Al4 “‘Ilabelled insulin in control and chloroquine-treated rats were analyzed by HPLC. Chloroquine treatment did not qualitatively modify the elution profile of the radioactivity (experiments not shown), but caused a decrease in the rate of appearance of
5
6
7
8
95
6
7
8
9
PH Fig. 4. Degradation of A14- (panels A and B) and B26in isolated liver endo(panels C and D) ‘251-monoiodoinsulin somes in the absence (upper panels) or presence (lower panels) of ATP as a function of medium pH. Endosomal fractions were incubated isolated 90 s after injection of ‘251-iodoinsulin in buffered isotonic KC1 for 6 min at 30°C as described in the Materials and Methods section. Insulin integrity was assessed by TCA precipitation (O), binding to antiinsulin antibody (A) and binding to insulin receptors (0) by reference to control hormone. The results shown are the mean of three determinations.
181
the major degradation products, especially that of peak A-I (Fig. 3). In vitro degradation of ‘251-labelled insulins associated with endosomal fractions In agreement with the studies of Pease et al. (1985,1987), a time- and pH-dependent loss in the integrity of Al4 and B26 ‘Z51-monoiodoinsulins occurred when endosomal fractions containing these ligands were incubated at 30°C in buffered isotonic KCl. In time studies, this loss was more pronounced when estimated by receptor and antibody binding (half-life at pH 5.5, about 3 min for Al4 iodoinsulin and 5 min for B26 iodoinsulin) than by TCA precipitation (half-life, about 8 min for both iodoinsulins). In the absence of ATP, the loss of ‘251-iodoinsulin integrity was maximal at pH 5-6, but in the presence of ATP, which causes endosomal acidification (Doherty et al., 1990;
80
.L -
0
: 80
0
0
3
6 Time
9
12
(min)
Fig. 6. HPLC analysis of the t2’I-labelled material generated from B26 ‘2sI-labelled insulin in isolated liver endosomes as a function of time. Iodoinsulin degradation was measured at pH 5.5 in the absence of ATP (panel A) or at pH 8.5 in the presence of ATP (panel E). Symbols indicate recoveries of radioactivity as intact iodoinsulin (O), peaks B-I + B-II (A) and monoiodotyrosine (0). The results shown are the mean f SEM of three experiments.
0
9
6
3 Time
12
(min)
Fig. 5. HPLC analysis of the lZSI-labelled material generated from A14 ‘251-labelled insulin in isolated liver endosomes as a function of time. Endosomal fractions isolated 90 s after iodoinsulin injection were incubated in isotonic KC1 at pH 5.5 in the absence of ATP (panel A) or at pH 8.5 in the presence of ATP (panel B) at 30°C as described in the Materials and Methods section. Incubation mixtures were analyzed by HPLC and the radioactivity recovered as intact iodoinsulin (0). peaks A-I + A-II (A) and monoiodotyrosine (0) was measured. The results shown are the mean f SEM of three experiments.
Desbuquois et al., manuscript submitted for publication), this maximum was shifted to pH 7-8 (Fig. 4). The ATP-induced pH shift was somewhat greater with Al4 iodoinsulin than with B26 iodoinsulin, and was also greater when iodoinsulin integrity was estimated by TCA precipitation than by receptor and antibody binding. In order to characterize the insulin degradation products generated in this cell-free system, endosomal fractions containing Al4 and B26 iodoinsulins were incubated under similar time and pH conditions, extracted and analyzed by HPLC. Besides intact iodoinsulin, both peaks A-I and A-II (with Al4 1251-labelled insulin) and B-I and B-II (with B26 ‘251-labelled insulin) were identified in the eluates; an additional product, seen in vivo mainly with B26 ‘251-iodoinsulin, was monoiodotyrosine (experiments not shown). The distribution of the radioactivity between these various components was first studied as a function of time, both at pH 5.5 in the absence of
182
ATP and at pH 8.5 in the presence of ATP. With Al4 “‘I-iodoinsulin (Fig. 5) a time-dependent decrease in intact iodoinsulin, along with the appearance of peaks A-I and A-II and monoiodotyrosine, were observed; these changes occurred somewhat more rapidly at pH 5.5 in the absence of ATP than at pH 8.5 in the presence of ATP. Peaks A-I and A-II reached a maximum by 9 min (about 30% of the total at pH 5.5), whereas monoiodotyrosine increased linearly with time up to 12 min (about 3% of the total per min at pH 5.5). With B26 lz5I-iodoinsulin (Fig. 6), however, there was a time-dependent decrease in the amount of radioactivity initiahy associated with peaks B-I and B-II, and monoiodotyrosine was the main product generated. Incubations were then carried out for 6 min at various pHs in the absence and in the presence of ATP. With Al4 ‘251-iodoinsulin (Fig. 7), the decrease in intact iodoinsulin and the correlative A
5
6
7
8
9
80
.z c
0
: 80 0 ._ P m L
L
5
6
7
8
9
PH
Fig. 8. HPLC analysis of the ‘251-labelled material generated from B26 ‘251-labelled insulin in isolated liver endosomes as a function of medium pH. Iodoinsulin degradation was measured in the absence (panel A) or presence (panel 8) of ATP as described for A14 ‘251-labelled insulin in the legend to Fig. 6. Symbols indicate recoveries of intact iodoinsulin (0) peaks B-I+B-II (A) and monoiodotyrosine (0). The results shown are the mean f SEM of three experiments.
increase in monoiodotyrosine were maximal at pH 6, and the increase in peaks A-I and A-II was maximum at pH 5; addition of ATP shifted the corresponding pH values to about 7 and 6, respectively. With B26 ‘251-labelled insulin (Fig. 8) the pH dependence for insulin disappearance and monoiodotyrosine appearance was comparable to that observed with the Al4 isomer, and so was the ATP-induced shift of the corresponding optimum pH. However, the small amount of peaks B-I and B-II present was maximal at pH 5 (or below) and this maximum was little affected by ATP.
tJH Fig. 7. HPLC analysis of the ‘251-labelled material generated from Al4 ‘251-labelled insulin in isolated liver endosomes as a function of medium pH. Endosomal fractions isolated 90 s after iodoinsulin injection were incubated in buffered isotonic KC1 in the absence (panel A) or presence (panel B) of ATP for 6 mitt at 30°C as described in the Materials and Methods section. Incubation mixtures were analyzed by HPLC and recoveries of intact iodoinsulin (o), peaks A-I+A-II (A) and monoiodotyrosine (0) were measured. The results shown are the mean f SEM of three experiments.
Discussion
One first goal of these studies was to isolate and chemically characterize the degradation products generated from Al4 and B26 i2?-labelled insulins in liver endosomes in vivo. To achieve this, the approach devised by Hamel et al. (1988) was used with two major modifications. First, livers were homogenized in the presence of inhibi-
183
tom of proteases that, unlike those used by IIamel et al. (1988), effectively blocked insulin degradation during subcellular fractionation. Second, and most importantly, a direct identification of the products generated from B26 ‘*?-labelled insulin was achieved using automated Edman degradation. The HPLC elution pattern of iodoinsulin degradation products observed in this study was comparable to that described by Hamel et al. (1988), except for the absence of late eluting components with the Al4 isomer. With both iodoinsulins, two prominent products slightly less hydrophobic than insulin were identified; the products generated from the Al4 isomer eluted as a doublet and were more abundant than those generated from the B26 isomer. However, at comparable times after injection, the relative abundance of these products exceeded by at least 5-fold that observed by Hamel et al. (1988), and concurrently, much less iodotyrosine and small iodotyrosine peptides were present. This presumably reflects our use of l,lO-phen~t~o~e, which, unlike the inhibitors used by I-&me1 et al. (1988), effectively blocked the degradation of the iodoinsulin associated with endosomes during subcellular fractionation. In addition to the major products, two minor early eluted products were identified with B26 125I-labelled insulin. After reduction and rechromatography, all products generated from Al4 ‘251-labelled insulin eluted at the position of intact A chain. In contrast, with B26 ‘*?-labelled insulin, the two major products gave rise, after reduction, to a single peptide which eluted differently from intact B chain; the elution of the two minor products was unaffected by reduction and also differed from that of intact B chain. These results extend those reported by Hamel et al. (1988) and suggest the existence of at least three cleavages in the B chain, two of which occur ~boxy-te~n~ly to Cys-19. However, the apparent integrity of the A chain in studies using Al4 ‘251-labelled insulin obviously does not exclude cleavages in this chain as well. Indeed, monoiodotyrosine, one major product of A chain degradation in hepatocytes (Duckworth et al., 1988), might have rapidly diffused out of the endosomes once formed, and thus escaped detection. Furthermore, our studies with cell-free endo-
somes indicate that, in the absence of inhibitors of degradation, iodotyrosine is effectively generated. Using automated Edman degradation, we have found that the B chain peptide derived from the major degradation products of B26 ‘*‘I-labelled insulin is cleaved at the B16-B17 bond, and that the two minor products are cleaved at the B23-B24 and B24-B25 bonds, respectively. Since the two major products generate the same B16-B26+ peptide, they may differ either at the A chain or at the B chain sequence located amino-terminally to the B16 residue, in accordance with the sequence of proteolytic events suggested by Assoian and Tager (1982). Thus, as suggested from comparative HPLC elution profiles of the products generated in endosomes and of those generated by insulin protease (Hamel et al., 1988), two cleavages in the B chain affect the B16-B17 and B24-B25 bonds. However, contrary to expectation, the third cleavage was found to affect the B23-B24 and not the B25-B26 bond; the reasons for this discrepancy are not obvious. Chloroquine, an acidotropic drug, has previously been reported to cause enhanced accumulation of undegraded insulin in the endosomal compartment, whether injected intraperitoneally (Posner et al., 1982) or perfused intraportally (Smith et al., 1989). Our studies show that, in agreement with the studies of Smith et al. (1989), chloroquine treatment causes a decrease in the amount of the degradation products generated from Al4 and B26 12’I-labelled insulins, but does not qualitatively alter the elution of the products. At variance with these results, Hamel et al. (1987) have reported that, in isolated hepatocytes incubated with Al4 ‘251-labelled insulin, chloroquine treatment leads to the appearance of a degradation product more hydrophobic than the primary products found in control cells, while concomitantly blocking the formation of these products; all products contained an intact A chain. However, in their study hepatocytes were continuously exposed to insulin for 30 min, and total intracellular products rather than endosomal products were analyzed. Previous studies have shown that when endosomal fractions isolated from insulin-injected animals are incubated in vitro at 37°C under isotonic conditions, the insulin associated with
184
these fractions undergoes rapid degradation as judged from TCA precipitability, with a maximal at pH 6 (Pease et al., 1985, 1987). The present studies confirm these observations and show that insulin degradation occurs to a greater extent when estimated by binding to insulin receptors, binding to antiinsulin antibody and HPLC. These studies also show that the degradation products generated from the i~o~su~s in cell-free endosomes are qu~~tively similar to those identified in vivo. However, unlike observed in vivo, a large fraction of degraded iodoinsulin was recovered as monoiodotyrosine, especially with the B26 isomer. Differences between in viva and in vitro results may be explained by a rapid diffusion of monoiodotyrosine out of the endosomes in the in vivo studies. The present studies show that ATP, by decreasing the endosomal pH (Desbuquois et al., manuscript submitted for publication), elevates the medium pH required for optimum iodoinsulin degradation from 5-6 to 7-8. Although the ATPinduced pH shift was detectable regardless of the assay, pHs required for maximal degradation somewhat varied depending on the parameter measured. Thus, in HPLC studies, the pH values at which the generation of intermediate products was maximal were at least 1 unit lower than the pH at which the generation of iodotyrosine and the disappearance of insulin were maximal, whether ATP was present or not; this was especially marked for B26 ‘2sI-labelled insulin This suggests that several enzymes with different pH optima may be involved in insulin degradation; alternatively, the pH optimum for peptide bond cleavage by the insulin-degrading enzyme(s) may vary depending on the particular bond cleaved. In addition, since receptor-bound insulin can be degraded by insulin protease (Yonezawa et al., 1988) and since one consequence of lowering the endosomal pH is to promote the dissociation of insulin from its receptor (Doherty et al., 1990; Desbuquois et al., manuscript submitted for publication), the possibility that receptor-bound and dissociated insulin may be degraded differently, or by different enzymes, must also be considered. Based on the finding that the insulin degradation products generated from B26 ‘251-labelled insulin in liver endosomes (Hamel et al., 1988>,
like those generated by isolated hepatocytes (Duckworth et al., 1988), are similar to the products generated by insulin protease, it has been suggested that this enzyme may be responsible for insulin degradation in endosomes (Duckworth, 1988; Hamel et al., 1988). However, studies with the Al4 isomer have shown that while the insulin degradation products identified in hepatocyte media (Pell et al., 1986; Hamel et al., 1987) resemble those generated by insulin protease, this is not the case for the products identified in endosomes (Hamel et al., 1988) and hepatocyte extracts (Pell et al., 1986; Hamel et al., 1987). In addition, it is unknown whether insulin protease, an enzyme which is primarily cytosolic (Duckworth, 1988), is also present in endosomes. In summary, the insulin degradation products generated in liver endosomes in vivo have been isolated and characterized, and cleavages in the B chain at the B16-B17 and B24-B25 bonds, suggested to occur in a recent study (Hamel et al., 1988), have been firmly identified. In addition, the insulin degradation products generated in cell-free endosomes under conditions that promote endosomal acidification have been shown to be qualitatively similar to those identified in vivo. References Assoian, R.K. and Tager, H.S. (1982) J. Biol. Chem. 257, 9078-9085. Bailey, J.L. and Cole, R.D. (1959) J. Biol. Chem. 234, 17331739. Biackard, W.G., Ludeman, C. and Stillman, J. (1985) Am. J. Physiol. 248, El94E202. Blackard, W.G., Smith, R.M. and Jarett, L. (1986) Am. J. Physiol. 250, E148-E156. Caro, J.F., Muller, G. and Gennon, J.A. (1982) J. Biol. Chem. 257, 8459-8466. Desbuquois, B., Wiileput, J. and Huet de Froberville, A. (1979) FEBS Lett. 106, 338-344. DeZoeten, L.W. and Havinga, E. (1961) Rec. Trav. Chim. 80, 917-926. Doherty, II, J.J., Kay, D.G., Lai, W.H., Posner, B.I. and Bergeron, J.J.M. (1990) J. Cell Biol. 110, 35-42. Draznin, B., Solomons, CC., Toothaker, D.R. and Sussman, K.E. (1981) Endocrinology 1108, 8-17. Duckworth, W.C. (1988) Endocr. Rev. 9, 319-345. Duckworth, W.C., Runyan, K.R., Wright, R.K., Halban, P.A. and Solomon, S.S. (1981) Endocrinology 108, 1142-1147. Duckworth, W.C., Hamel, F.G., Peavy, DE., Liepnick, J.J., Ryan, M.P., Hem&son, M.A. and Frank, B.H. (1988) J. Biol. Chem. 263, 1826-1833.
185 Ehrenreich, J.H., Bergeron, J.J.M., Siekevitz, P. and Palade, G.E. (1973) J. Cell. Biol. 59, 45-72. Fieig, W.E., Hess, G., Nother-Fleig, G. and Ditschuneit, H. (1986) B&hem. J. 237, 99-104. Frank, B.H., Peavy, D.E., Hooker, C.S. and Duckworth, W.C. (1983) Diabetes 32, 705-711. Hamel, F.G., Peavy, D.E., Ryan, M.P. and Duckworth, W.C. (1987) Diabetes 36, 702-708. Hamel, F.G., Posner, B.I., Bergeron, J.J.M., Frank, B.H. and Duckworth, W.C. (1988) J. Biol. Chem. 263, 6703-6708. Janicot, M. and Desbuquois, B. (1987) Eur. J. B&hem. 163, 433-442. Jorgensen, K.H. and Larsen, U.D. (1980) Diabetologia 19, 546-554. Juul, S.M., Jones, R.H., Evans, J.L., Neffe, J., Sonksen, P.H. and Brandenburg, D. (1986) B&him. Biophys. Acta 856, 310-319. Levy, J.R. and Olefsky, J.M. (1987) Endocrinology 121, 20752086. Linde, S., Sonne, 0.. Hansen, B. and Ghemann, J. (1981) Hoppe-Seyler’s Z. Physiol. Chem. 362, 573-579.
Pease, R.J., Smith, G.D. and Peters, T.J. (1985) B&hem. J. 228, 137-146. Pease, R.J., Smith, G.D. and Peters, T.J. (1987) Eur. J. Biothem. 164, 251-257. Peh, M.E., Duckworth, W.C. and Peavy, D.E. (1986) B&hem. Biophys. Res. Commun. 137, 1034-1040. Posner, B.I., Patel, B.A., Verma, A.K. and Bergeron, J.J.M. (1980) J. Biol. Chem. 255, 735-741. Posner, B.I., Patel, B.A., Khan, M.N. and Bergeron, J.J.M. (1982) J. Biol. Chem. 257, 5789-5799. Smith, G.D., Christensen, J.R., Rideout, J.M. and Peters, T.J. (1989) Eur. J. B&hem. 181, 287-294. Sonne, 0. (1989) Physiol. Rev. 68, 1129-1196. Ten-is, S. and Steiner, D.F. (1975) J. Biol. Chem. 250, 83898398. Yonezawa, K., Yokono, K., Shii, K., Hari, J., Amano, K., Sakamoto, T., Kawase, Y., Akiyama, H., Nagata, M. and Baba, S. (1988) B&hem. Biophys. Res. Commun. 150, 605-614.
