| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Institut National de la Santé et de la Recherche Médicale U567 and Centre National de la Recherche Scientifique Unite Mixte de Recherche 8104 (B.D.), Département d’Endocrinologie, Institut Cochin, 75014 Paris, France; Institut National de la Santé et de la Recherche Médicale U530 (G.C.), Université René Descartes, Unité de Formation et de Recherche Biomédicale des Saints-Pères, 75007 Paris, France; Laboratoire de Spectrométrie de Masse (M.K.), Faculté de Médecine, 59000 Lille, France; and Institut National de la Santé et de la Recherche Médicale U510 (F.A.), Faculté de Pharmacie Paris XI, 92296 Châtenay-Malabry, France
Address all correspondence and requests for reprints to: Bernard Desbuquois, Département d’Endocrinologie, Institut Cochin, 24 rue du Faubourg Saint-Jacques, 75014 Paris, France. E-mail: desbuquois{at}cochin.inserm.fr.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Proinsulin binds to the insulin receptor and elicits acute insulin-like metabolic effects in adipocytes and hepatocytes, albeit with a 10- to 100-fold-lower binding affinity and biological potency (reviewed in Ref. 4). Remarkably, in vivo, the ability of proinsulin to suppress hepatic glucose production is less reduced than its ability to stimulate glucose disposal (5, 6). In addition, relative to insulin, in vivo effects of proinsulin develop more slowly and are more prolonged after hormone withdrawal; most long-lasting is suppression of hepatic glucose production (7). Proinsulin also exerts potent growth-promoting effects in chick embryo fibroblasts (8), IM-9 lymphoblasts (9), and small intestinal crypt cells (10). However, the latter effects are mediated by incompletely characterized receptors, which show a greater affinity for proinsulin than for insulin.
As expected from its low affinity for the liver insulin receptor (11, 12), proinsulin displays a reduced extraction by the isolated perfused liver (13, 14) and liver in vivo (15, 16), which accounts for its prolonged half-life in serum. Furthermore, after in vivo uptake into the liver, proinsulin is more slowly cleared than insulin, suggesting a slower intracellular processing (15, 16). However, although proinsulin has been shown to be less efficiently degraded than insulin by isolated liver subcellular fractions (11, 17) and purified liver insulin-degrading enzyme (IDE) (18), studies on proinsulin processing in intact cells have been limited to adipocytes (19) and fibroblasts overexpressing the insulin receptor (20). In the latter cells, proinsulin was found to be less efficiently degraded than insulin and, in part, released intact from the cell via a retroendocytotic pathway; this observation was suggested to account for the prolonged biological activity of proinsulin in vivo (20).
In the present study, the cell itinerary and metabolic fate of proinsulin in rat liver have been analyzed using cell fractionation, and the results have been compared with those obtained with insulin. Particular attention has been paid to the subcellular distribution of [125I]-proinsulin taken up by rat liver in vivo, the fate of in vivo internalized [125I]-proinsulin in liver cell-free systems, and the in vitro degradation of proinsulin by liver subcellular fractions. Proinsulin degradation intermediates, generated on incubation with soluble endosomal extracts, have been isolated by RP-HPLC and characterized by mass spectrometry. Finally, the abilities of native proinsulin and insulin to induce the endocytosis and tyrosine phosphorylation of the insulin receptor in vivo have been compared.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Preparation and characterization of [125I]-labeled peptides
Porcine and human [125I]-proinsulin and [125I]-insulin (about 0.6–2.2 mCi/nmol; 0.2–0.5 125I atom per molecule) were prepared using chloramine T and purified by gel-filtration on Sephadex G-50 (21, 22). To identify the tyrosine(s) residues iodinated, [125I]-proinsulin was quantitatively converted to [125I]-insulin by enzymatic digestion (19, 23), and both digested [125I]-proinsulin and [125I]-insulin were subjected to reverse-phase HPLC (RP-HPLC) (24, 25). Based on the elution profile of the radioactivity, iodination was found to occur mainly (70%) at tyrosines A14 in [125I]-insulin and A83 (the equivalent of tyrosine A14) in [125I]-proinsulin (results not shown). However, a minor population of molecules labeled at tyrosine A19 or equivalent was also identified. The ability of [125I]-proinsulin to bind to the liver insulin receptor at pH 7–8 was about 8–11% that of [125I]-insulin.
Animals and peptide injections
In vivo procedures were approved by the institutional committee for use and care of experimental animals. Male Sprague Dawley rats, weighing 200 ± 20 g, were obtained from Charles River (St. Aubin les Elbeufs, France) and Elevage Janvier (Le Genest-St-Isle, France) and were fasted for 16 h before killing. When indicated, porcine [125I]-labeled peptides (15–25 µCi; 20–30 pmol) or native human peptides (0.06–13 nmol), diluted into 0.5 ml 0.15-M NaCl, were rapidly injected into the penis vein under ether or pentobarbital anesthesia. At the time indicated, the liver was rapidly removed, minced in ice-cold 0.25 M sucrose, and immediately homogenized.
Liver subcellular fractionation
Livers were homogenized in 5 vol of 0.25-M sucrose using a Dounce homogenizer (loose-fitting pestle). When indicated, bacitracin (1 mg/ml), 1,10 phenanthroline (2.5 mM), and N-ethylmaleimide (1 mM) were included in the homogenization medium to reduce degradation of labeled peptides during fractionation. Nuclear (N), heavy mitochondrial (M), light mitochondrial (L), microsomal (P), and soluble (S) fractions and, when indicated, total particulate (NMLP) or combined LPS fractions, were isolated from homogenates by differential centrifugation. The plasma membrane (PM) fraction was isolated from the N fraction as described by Hubbard et al. (26). Two sets of three endosomal fractions were isolated from the L and P fractions, respectively, as described by Khan et al. (27); they were designated(by increasing rank order of densities) as Ll, Li, and Lh (L subfractions) and GEl, GEi, and GEh (P subfractions). In some experiments, the LPS, L, and P fractions were subfractionated by centrifugation on continuous Nycodenz density gradients (25, 28). Hypotonic treatment of endosomal fractions and isolation of soluble endosomal extracts were achieved as described previously (24, 25, 29).
Cell fractions, isolated after injection of [125I]-labeled peptides, were analyzed for peptide content and integrity; the latter was assessed by precipitation by 5% trichloroacetic acid (TCA) and gel-filtration on a Sephadex G-50 column (1 x 50 cm) equilibrated with 1 M acetic acid (21, 28). Association of the labeled peptides with the insulin receptor was assessed by cross-linking with disuccinimidylsuberate followed by SDS-PAGE and autoradiography (28). Cell fractions, isolated after injection of native peptides, were analyzed for peptide content after extraction by 0.1 N HCl containing 0.5% BSA and 0.05% bacitracin. in vitro insulin binding activity, and insulin receptor ß-subunit expression and phosphorylation (see below).
Fate of in vivo internalized [125I]-labeled and unlabeled peptides in liver cell-free systems
Degradation, extraendosomal release, and intraendosomal dissociation of in vivo internalized [125I]-labeled peptides in intact cell-free endosomes were studied as described previously (21, 24). Briefly, a total, P-derived endosomal fraction was isolated after injection of [125I]-insulin (2 min) or [125I]-proinsulin (4 min) and resuspended in 0.125 M KCl buffered with 25 mM citrate-phosphate, at the indicated pH, for various lengths of time at 37 C. Peptide degradation was assessed by TCA-precipitation; extraendosomal peptide release and intraendosomal peptide dissociation from the insulin receptor were assessed by polyethyleneglycol (PEG) precipitation in the absence and presence of 0.1% Brij-35, respectively.
Degradation of in vivo internalized native peptides in cell-free endosomes was studied in a comparable manner. Endosomes isolated after injection of insulin, proinsulin, des-[64, 65] proinsulin, or split-[32–33] proinsulin (1, 5, 1.66, and 3.33 nmol, respectively), were incubated in 0.125 mM KCl and 25 mM citrate-phosphate, pH 6. After various times at 37 C, peptides were extracted by 0.1 N HCl as described above, and their concentrations in the extracts were measured by RIA.
Cell-free endosome-lysosome transfer of in vivo internalized [125I]-labeled peptides was studied as described previously (25, 28). Aliquots (0.3–0.5 ml) of a LPS fraction, isolated 8 min after injection of [125I]-insulin or [125I]-proinsulin, were incubated at 4 C or 37 C with 1.5 mM ATP, 0.5 mg/ml creatine kinase, and 10 mM phosphocreatine. After cooling to 4 C, incubation mixtures were subjected to centrifugation on linear Nycodenz gradients; and the distribution of radioactivity and N-acetyl-ß-D-glucosaminidase activity among soluble, endosomal, and lysosomal components was determined.
Assays for in vitro degradation of [125I]-labeled and unlabeled peptides by liver subcellular fractions
125I-labeled (0.1 nM) and unlabeled (10 nM) peptides were incubated with endosomal, PM, and cytosolic fractions in 50 mM citrate phosphate buffer containing 0.5% (wt/vol) BSA at 37 C; specific conditions (time, medium pH, and concentration of cell fraction protein) are indicated in each particular experiment. Degradation of [125I]-labeled peptides was measured by TCA-precipitation and binding to excess antiinsulin antibody and degradation of native peptides by RIA.
Isolation and characterization of proinsulin degradation intermediates
Degradation intermediates, generated on incubation of native proinsulin with soluble endosomal extracts, were isolated by RP-HPLC exactly as described for insulin intermediates (25). Eluates were monitored on-line for absorbance at 214 nm, and resolved components were analyzed by ion spray mass spectrometry and HPLC coupled with ion spray ionization-mass spectrometry as described previously (25, 30).
Measurement of insulin binding activity and analysis of insulin receptor ß-subunit expression and tyrosine phosphorylation in liver cell fractions
Insulin binding activity of PM and endosomal fractions was measured using [125I]-insulin as a ligand (31).
Insulin receptor ß-subunit expression and tyrosine phosphorylation in cell fractions were analyzed by Western immunoblotting using antibodies against the ß-subunit of the insulin receptor and phosphotyrosine, respectively (24, 28). Cell fraction proteins, resolved by reducing SDS-PAGE electrophoresis, were electrotransferred to nitrocellulose membranes (0.45 µm) for 1 h at 380 mA in transfer buffer (25 mM Tris, 192 mM glycine, pH 8.3). After incubation for 3 h in blocking buffer (10 mM Tris-HCl, pH 7.5, 300 mM NaCl, 0.05% Tween-20) containing 5% skim milk (insulin receptor) or 2% BSA (phosphotyrosine), membranes were incubated with antibody to the insulin receptor ß-subunit (1:1000) or phosphotyrosine (1:2500) in the same buffer for 16 h at 4 C. Membranes were then washed three times in blocking buffer containing 0.5% skim milk (insulin receptor) or 0.2% BSA (phosphotyrosine) over a period of 1 h and probed using horseradish-peroxidase-conjugated goat antirabbit IgG (insulin receptor) or antimouse IgG (phosphotyrosine). Immune complexes were revealed using enhanced chemiluminescence reagents and autoradiography.
Measurement of insulin and proinsulin concentration
Concentrations of insulin, proinsulin, and proinsulin intermediates in incubation mixtures and acid extracts of liver cell fractions were measured by RIA using [125I]-insulin and an anti-insulin antibody that reacts identically with insulin and proinsulin. Proinsulin concentrations were also measured with a RIA that uses [125I]-proinsulin and a specific antiproinsulin antibody (22); the latter recognizes proinsulin and des-[31, 32] proinsulin but not des-[64, 65] proinsulin and insulin. Samples were assayed at three to four different dilutions against appropriate standard curves, and antibody-bound ligand was separated from free ligand by PEG precipitation (32).
Protein and enzyme assays
Protein (33, 34), galactosyltransferase (35), and N-acetyl-ß-D-glucosaminidase (35) activity in cell fractions were assayed using standard procedures.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Concurrently, a time-dependent decrease in TCA-precipitability of the peptides retained in the liver was observed (Fig. 1B). In agreement with a previous report (15), the fate of injected [125I]-proinsulin differed from that of [125I]-insulin in three respects: first, a 3-fold-lower uptake of radioactivity at early postinjection times (maximum, 0.008% and 0.026% of the injected dose per milligram of protein, corresponding to 1.6% and 5.2% of injected dose per gram of liver, respectively); second, a slower loss of radioactivity from the liver after 4 min (apparent half-lives, about 30 and 6 min, respectively); and third, a slower decrease in fractional TCA-precipitable radioactivity, which was most apparent before 8 min.
|
Subcellular distribution and characterization of [125I]-proinsulin and [125I]-insulin taken up into the liver
The distribution of [125I]-labeled peptides into the N, M, L, P, and S fractions was first examined, with results expressed as percent of total recovered radioactivity (Fig. 2). As shown previously (36), the P fraction was the locus of highest recovery of [125I]-insulin, with a maximum at 1.5 min (about 70%). However, at early times, some radioactivity was also present in the N fraction; and at later times, the recovery of radioactivity decreased in the P fraction while concurrently increasing in the S fraction and, to a slight extent, in the M and L fractions. Regardless of the fraction, the percentage of TCA-soluble radioactivity (shown in black) increased with time but did so more rapidly and to a greater extent in the S fraction than in sedimentable fractions.
|
The kinetics of uptake of [125I]-labeled peptides into PM and endosomal fractions was then studied (Fig. 3). In agreement with previous reports (24, 27, 36, 37), [125I]-insulin associated first with the plasma membrane and then with endosomes, with maximal enrichments of about 10-fold (plasma membrane), 40- to 80-fold (P-derived endosomes), and 100- to 300-fold (L-derived endosomes), respectively. In each of these fractions, the percentage of TCA-soluble radioactivity increased with time, especially in the PM and light endosomal fractions. Although qualitatively comparable, the distribution of [125I]-proinsulin differed from that of [125I]-insulin by a less marked and prolonged association with the PM fraction, a delayed and prolonged association with endosomal fractions, and a reduced and slower increase in TCA-soluble radioactivity in both fractions.
|
To confirm the involvement of the insulin receptor in the in vivo uptake of [125I]-proinsulin, PM and endosomal fractions were subjected to reducing SDS-PAGE analysis after cross-linking of ligand-receptor complexes. As with [125I]-insulin, a major band of molecular mass 135-kDa, the expected size of the 
-subunit of the insulin receptor, was identified (results not shown). In time studies, the intensity of this band paralleled the radioactivity content of the fractions; but at any given time, it was at least three times lower than with [125I]-insulin.
Gel filtration analysis of the radioactivity extracted from endosomal fractions showed that, in addition to the major proinsulin- and insulin-sized peaks, some radioactivity was eluted as high- and low-molecular-weight components (Fig. 4). At 8 min, the low-molecular-weight components accounted for about 35% of the total eluted radioactivity after [125I]-insulin injection but only 10% after [125I]-proinsulin injection. In addition, little or no radioactivity was eluted at the position of [125I]-insulin after [125I]-proinsulin injection.
|
). As [125I]-insulin, [125I]-proinsulin was degraded in a time- and pH-dependent manner, with a maximum at pH 6, and addition of ATP shifted the pH for maximal degradation to about 7.5. However, at the optimum pH, the rate of proinsulin degradation was about three times lower than that of [125I]-insulin degradation (Table 1). Studies with endosomes isolated after coinjection of increasing doses of native proinsulin (up to 13.3 nmol) with [125I]-proinsulin showed that, as with insulin (21), the fractional degradation of [125I]-proinsulin decreased with increasing endosomal ligand content (results not shown). However, regardless of ligand content, [125I]-proinsulin was always degraded about three times less rapidly than [125I]-insulin.
|
|
). Under these conditions, the rate of proinsulin degradation, as measured by RIA using an anti-insulin serum, was decreased about 6-fold, relative to that for insulin (2% and 11.7% per min, respectively). Two proinsulin conversion intermediates, more slowly cleared from the liver than insulin, des-[64, 65]proinsulin and split-[32–33]proinsulin, showed rates of endosomal degradation intermediate between those for insulin and proinsulin (8.3 and 4.7% per min, respectively). The rate of proinsulin degradation was also decreased when estimated by a proinsulin-specific RIA but to a lesser extent (3.8% per min; Table 1).
Because insulin degradation within endosomes requires its dissociation from the insulin receptor at the low endosomal pH, the possibility that the decreased degradation of [125I]-proinsulin resulted from its slower dissociation from the receptor was examined. To address this, the extraluminal release and intraluminal dissociation of [125I]-insulin and [125I]-proinsulin in intact cell-free endosomes, as a function of pH, were comparatively examined using a PEG-precipitation procedure (21, 24) (Fig. 6). As the generation of TCA-soluble radioactivity, the extraendosomal release of radioactivity (judged on the generation of PEG-soluble radioactivity in intact endosomes) was maximal at pH 6 but was about 2-fold lower for [125I]-proinsulin than for [125I]-insulin in the pH range 4–7 (Fig. 6A). This indicates that only TCA-soluble products diffuse out of endosomes. The intraendosomal dissociation of receptor-bound ligands (judged on the generation of PEG-soluble radioactivity in Brij-35-permeabilized endosomes) progressively decreased as the pH was increased, but was virtually identical for [125I]-proinsulin and [125I]-insulin (60–75%) in the pH range 4–6.5 (Fig. 6B). Dissociation of [125I]-proinsulin even slightly exceeded that of [125I]-insulin at pH 7 and above, consistent with a previous report (11). Thus, a slower dissociation of [125I]-proinsulin from the internalized insulin receptor does not account for its reduced degradation in endosomes.
|
).
|
). However, whereas the lysosomal association of [125I]-insulin was maximal at 15 min and did not exceed 16% of the total recovered, that of [125I]-proinsulin was maintained up to 30 min and reached 30% (Fig. 7B).
In vitro degradation of [125I]-labeled and native peptides by isolated liver subcellular fractions
The results described above suggest that the reduced degradation of proinsulin in endosomes results from its lower sensitivity to endosomal acidic proteases. To address this, the in vitro degradation of [125I]-insulin and [125I]-proinsulin by hypotonically disrupted endosomes as a function of pH was comparatively examined using the TCA-precipitation assay (Fig. 8B and Table 1). Both peptides were degraded with a pH optimum of 4.5; but throughout the pH range, [125I]-proinsulin degradation was at least 2-fold lower than [125I]-insulin degradation. Because about 90% of the insulin-degrading activity of hypotonically disrupted endosomes is soluble (29), results were similar whether a total extract or a soluble extract was used (results not shown). However, because of the selective extraction of endosomal proteases, the degrading activity in the soluble extract was concentrated by about 12- to 15-fold, when related to protein (Table 1). [125I]-proinsulin was also less degraded than [125I]-insulin when this was assessed by binding to excess anti-insulin antibody (Table 1).
|
and Table 1). Under these conditions, peptide degradation, as judged on the loss of insulin and proinsulin immunoreactivity, was maximal at pH 4; and within the pH range 3.5–4.5 proinsulin was degraded two to four times less rapidly than insulin. It is noteworthy that the rate of degradation of native peptides as measured by RIA was about twenty times higher than that of radiolabeled peptides using the TCA-precipitation assay. As with intact cell-free endosomes, the rates of degradation of des-[64, 65]proinsulin and split-[32–33]proinsulin by soluble endosomal extracts were intermediate between those of insulin and proinsulin (insulin > des-[64, 65]proinsulin > split-[32–33]proinsulin > intact proinsulin).
Because plasma membranes and cytosol degrade insulin in vitro and have been suggested to be potential sites of insulin degradation in intact cells (43, 44, 45), their ability to degrade proinsulin and proinsulin conversion intermediates was examined (Table 1). PM and cytosolic fractions degraded [125I]-proinsulin about 20–40 times less rapidly than [125I]-insulin; and with the cytosolic fraction, comparable results were observed when native peptides were used as substrates. As with endosomes, the rates of degradation of des-[64, 65]proinsulin and split-[32–33]proinsulin by cytosol were intermediate between those for insulin and intact proinsulin.
Selected protease inhibitors, including pepstatin-A (an aspartic protease inhibitor), leupeptin and E64 (two cysteine protease inhibitors), N-ethylmaleimide (a thiol blocking reagent), 1,10-phenanthroline (a metalloprotease inhibitor), and bacitracin (a nonspecific protease inhibitor) were tested for their ability to inhibit insulin and proinsulin degradation in isolated cell fractions. Consistent with previous reports (29, 30), endosomal degradation of [125I]-insulin as estimated by TCA-precipitation and binding to excess anti-insulin antibody, was totally inhibited by pepstatin-A (10 µg/ml) and bacitracin (5 mg/ml), partially inhibited by 1,10-phenanthroline (1 mM), but little affected by other inhibitors (results not shown). Pepstatin-A also totally blocked native insulin and proinsulin degradation by soluble endosomal extracts. In contrast, degradation of [125I]-labeled insulin by PM and cytosolic fractions was effectively suppressed by 1,10-phenanthroline, N-ethylmaleimide, and bacitracin but unaffected by pepstatin-A.
Characterization of proinsulin intermediates generated upon incubation with a soluble endosomal extract
Incubation of native insulin with a soluble endosomal extract has been shown to result in the generation of a number of degradation intermediates that can be resolved by RP-HPLC (25). Based on mass spectrometry analysis, the structure of these intermediates has been elucidated and cleavage sites in the A- and B-chains localized. We therefore used this strategy to isolate and characterize proinsulin degradation intermediates generated upon incubation with a soluble endosomal extract.
On RP-HPLC, human insulin and proinsulin eluted as single peaks with retention times of 34 min (results not shown) and 37 min (Fig. 9), respectively. Incubation of both peptides with a soluble endosomal extract at pH 4 and 5 resulted in the time-dependent disappearance of these peaks, along with the appearance of multiple, less hydrophobic peaks. Quantitation of peak areas confirmed that, at pH 4 and 37 C, human proinsulin was degraded less rapidly than human insulin (percent degraded at 30 and 60 min: proinsulin, 40% and 75%; insulin, 60 and 90%, respectively) (Fig. 9, and results not shown).
|
). Among these pools, five (1, 3, 5, 7, 8) were heterogeneous, and each contained two proinsulin intermediates. In addition to intact proinsulin, 12 proinsulin intermediates were identified. Eight intermediates (components 1–5) were single-chain peptides in which the C-terminal sequence Phe25-Thr30 of the B chain remained linked to N-terminal sequences of the C peptide of various lengths, three intermediates (pools 6 and 7) were two-chain peptides in which the N-terminal sequence Phe1-Phe24 of the B chain was connected via disulfide bridges to the A-chain linked to C-terminal sequences of the C peptide of various lengths, and one intermediate (pool 8b) was intact proinsulin with an internal peptide bond cleaved. From these data, it can be deduced that human proinsulin is cleaved by soluble endosomal proteases at nine distinct sites; of these, eight affect the C peptide (Ala34-Glu35, Asp36-Leu37, Gln38-Val39, Gly40-Gln41, Glu43-Leu44, Ala57-Leu58, Glu59-Gly60, and Gly60-Ser61) but only one the B-chain (Phe24-Phe25). In addition, because the generation of all intermediates involves a cleavage at the Phe24-Phe25 bond, the latter must be, as for insulin, the initial proteolytic event in proinsulin. Cleavages of bonds in the C-peptide region adjacent to the A-chain-C-peptide junction and of the Phe24-Phe25 bond would be expected to affect the recognition of proinsulin by, respectively, the antiproinsulin and anti-insulin antibodies used in this study.
|
and Fig. 10). In addition, the concentration of immunoreactive proinsulin and insulin in acidic extracts of these fractions was measured (Table 3).
|
|
). These doses had previously been shown to gradually inhibit the in vivo hepatic uptake of coinjected [125I]-insulin (data not shown). Although both peptides led to a dose-dependent increase in insulin receptor ß-subunit content and tyrosine phosphorylation, proinsulin was at least 10 times less potent than insulin in this respect; and at the highest dose of proinsulin tested (13.3 nmol), maximal effects were not achieved. These results are consistent with the 20-fold-lower in vivo affinity of proinsulin for the insulin receptor (dose required for 50% of inhibition of in vivo [125I]-insulin uptake, about 0.66 nmol insulin and 13 nmol proinsulin). Interestingly, based on endosomal ligand content (shown in legend to Fig. 10) proinsulin was also approximately 10–20 times less potent than insulin in increasing ß-subunit expression and autophosphorylation.
Plasma membrane and endosomal fractions were then isolated at various times after injection of 0.33 nmol insulin and 4.4 nmol of proinsulin (Fig. 10, B and C; and Table 3). At these doses, a comparable degree of insulin receptor occupancy was achieved, as judged on the inhibition of in vivo [125I]-insulin uptake (20–25%). Insulin treatment caused three major changes: first, a marked increase in insulin binding activity and expression of the ß-subunit of the insulin receptor in the endosomal fraction; second, a slight decrease in insulin binding activity and ß-subunit expression in the PM fraction; and third, an increase in tyrosine phosphorylation of the ß-subunit in both fractions. These changes achieved a maximum at 2 min but rapidly declined after 10 min, being no longer detectable after 20 min. Although of comparable magnitude, the changes induced by proinsulin differed from those induced by insulin in two respects: first, a slower and more prolonged increase in insulin binding activity, insulin receptor ß-subunit expression, and ß-subunit tyrosine phosphorylation in the endosomal fraction, which was maximal at 10 min and was still well detectable at 20 min (45–90 min for insulin binding activity); and second, a more prolonged increase in insulin receptor ß-subunit tyrosine phosphorylation in the PM fraction.
In agreement with results obtained with [125I]-labeled peptides, injected insulin and proinsulin were rapidly taken up into PM and endosomal fractions and were subsequently cleared from these fractions (Table 3). The association of both peptides with the plasma membrane was maximal at 2 min, but the translocation of proinsulin to endosomes was delayed, relative to that of insulin (maximum, 10 and 5 min, respectively). In addition, as expected from the slower clearance of proinsulin from blood plasma (4) and from the decreased processing of proinsulin in endosomes, injected proinsulin was cleared more slowly than insulin from both PM and endosomal fractions. Consequently, proinsulin concentrations achieved at 20 min and later in these fractions were at least 6- and 15-fold higher than insulin concentrations, respectively.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The patterns of in vivo liver uptake of injected proinsulin and insulin have been compared in two previous studies (15, 16). In one study, in which the liver was removed after injection of [131I]-labeled peptides, insulin uptake was maximal at 1 min and rapidly declined over 30 min, whereas proinsulin uptake was stable up to 10 min and declined more slowly; at maximum, 29% and 6% of the injected dose, respectively, were recovered in the liver (15). In another study, which involved external scintigraphic analysis of the liver after injection of peptides labeled with 123I at the A14 position, insulin uptake attained 32% of injected dose after 3 min and declined with a half-life of 6 min, whereas proinsulin uptake reached 18% at 1 min and decreased with a half-life of 45 min (16). Our estimates of liver uptake of [125I]-proinsulin and [125I]-insulin (1.6 and 5.2% of injected dose per gram of liver, respectively) and half-lives (28 and 6 min) compare favorably with these results.
Previous studies have shown that, on cell fractionation, [125I]-insulin taken up by rat liver in vivo associates first with plasma membranes sedimenting in the N and P fractions and then with endosomes sedimenting in the P and L fractions (27, 36, 37). In addition, although low or undetectable in vivo, a lysosomal association of internalized [125I]-insulin occurred in a cell-free system in the presence of cytosol and ATP (25, 28). Our results confirm these observations and show that the subcellular distribution of [125I]-proinsulin differs from that of [125I]-insulin in three respects: first, a reduced and slower association with the plasma membrane; second, a delayed and prolonged association with endosomes; and third, a greater and more sustained lysosomal association, best demonstrable in a cell-free system. Consistent with these observations, both the association of [125I]-proinsulin with cell surface insulin receptors and its internalization have been shown to be delayed in human fibroblasts overexpressing the insulin receptor (20).
[125I]-insulin taken up into PM and endosomal fractions in vivo has been shown to be converted to TCA-soluble and low-molecular-weight degradation products, which diffuse in the S fraction (24, 27, 36, 37). In addition, when endosomal fractions containing in vivo internalized insulin were incubated under isoosmotic conditions, the ligand underwent further degradation, with a pH optimum of 5–6; addition of ATP, by decreasing the internal endosomal pH, shifted this optimum to 7–8 (21, 24, 38, 39, 40, 41, 42). The current studies extend these observations to [125I]-proinsulin but show that the extent and rate at which TCA-soluble radioactivity is generated in vivo, as well as in intact cell-free endosomes, are about three times lower than for [125I]-insulin. Furthermore, a comparable and even-greater decrease in the rate of proinsulin degradation, as measured by RIA using anti-insulin and antiproinsulin antibodies, was observed in cell-free endosomes loaded with native peptides. The slower degradation of proinsulin in endosomes probably accounts for its prolonged retention in these organelles in vivo and may explain why, as other peptides or proteins which also show a low rate of endosomal degradation (25, 28, 47), internalized proinsulin undergoes a greater and more sustained lysosomal association in vivo and in the cell-free system.
Theoretically, the slower degradation of proinsulin in endosomes could result from a slower dissociation of the proinsulin-receptor complex at the low endosomal pH and/or from a decreased sensitivity of proinsulin to endosomal proteases. Our results do not support the former possibility and instead show that proinsulin degradation by hypotonically disrupted endosomes and soluble endosomal extracts, which occurs optimally at pH 4, is reduced by 3- to 5-fold relative to insulin degradation. This decrease is comparable to that observed in intact cell-free endosomes, supporting the view that the insulin- and proinsulin-degrading activity identified in soluble extracts is physiologically relevant (25, 29, 30). Furthermore, the relative degradation rates of insulin, intact proinsulin and proinsulin split conversion intermediates were similar in intact cell-free endosomes and soluble endosomal extracts (insulin > des-[64, 65]proinsulin > split-[32–33]proinsulin > intact proinsulin). The reasons why the pH optimum for insulin and proinsulin degradation is lower in disrupted endosomes (pH 4–4.5) (25, 29 and present study) than in intact cell-free endosomes (pH 5–6) (21, 24, 38, 39, 40, 41, 42 and present study) are not obvious. It is possible that, in freshly isolated endosomes incubated at low pH, endosomal integrity is not fully maintained, resulting in the extraluminal release of peptide and/or peptide-degrading activity.
In previous studies, degradation intermediates generated from [125I]-monoiodoinsulin in intact cell-free endosomes (42) and from native insulin in soluble endosomal extracts (25) have been isolated by RP-HPLC and characterized by radiosequencing and mass spectrometry, respectively. Based on these data, insulin degradation has been shown to proceed via an ordered sequential pathway, the PheB24-PheB25 bond being the major and earliest site of cleavage. Using the mass spectrometry approach, we show here that, when incubated with a soluble endosomal extract, human proinsulin is cleaved at eight bonds in the C peptide but only at one, the Phe 24-Phe25 bond, in the insulin moiety. Under closely similar pH, time, and substrate concentration conditions, insulin has been shown to be cleaved at eight bonds in the B chain and one in the A chain (25). This indicates that, despite similar conformations of proinsulin and insulin, the sensitivity of these bonds to endosomal proteases is reduced in proinsulin. As in insulin, cleavage of the Phe24-Phe25 bond in proinsulin is probably an early event, because it is involved in the production of the 12 intermediates identified. Although not detected here, additional cleavages in the insulin moiety of proinsulin would probably have occurred under more drastic conditions (longer incubation, lower substrate concentration), to account for the generation of low-molecular-weight TCA-soluble products from [125I]-proinsulin.
Endosomes have been shown to harbor several acidic proteases, including cathepsin-B, cathepsin-D, and ß-site amyloid precursor protein-cleaving enzyme. Based on a number of criteria, including the inhibitor profile of the degrading activity and the results of immunodepletion studies, the acidic protease cathepsin-D has been proposed as the enzyme that initiates insulin degradation in hepatic endosomes (30). That cathepsin-D may also initiate endosomal degradation of proinsulin is suggested by several observations. First, the Phe24-Phe25 bond cleaved in proinsulin is preferentially cleaved by this enzyme in insulin as well as in a number of synthetic substrates and proteins (reviewed in Ref. 25). Second, four of the eight bonds cleaved in C peptide, which contain an apolar residue on their carboxyl side, are also readily cleavable by cathepsin-D. Finally, like insulin degradation, proinsulin degradation by the soluble endosomal extract was totally inhibited by pepstatin-A, a cathepsin-D inhibitor. These observations do not exclude, however, the involvement of proteases other than cathepsin-D in the cleavage of bonds that do not contain apolar residues in C peptide, and/or in late cleavages in the A and B chains. The participation of IDE in endosomal proteolysis of proinsulin should be considered (see below), in view of the partial inhibition of the endosomal insulin-degrading activity by 1,10-phenanthroline and bacitracin, at least when using the TCA-precipitation assay (Refs. 29 and 43, 44, 45 and present study). Indeed, IDE has been implicated in the degradation of insulin in early endosomes before acidification (40, 43, 45).
In the present study, no insulin-sized components were identified in endosomes isolated after [125I]-proinsulin injection, nor were cleavages of the Arg31-Arg32 and Lys64-Arg65 bonds at the B–C and A–C junctions, respectively, detected in native proinsulin. However, although not converted to mature insulin when expressed in normal hepatocytes and hepatoma cell lines (48, 49), human proinsulin has been shown to be processed to des-[31, 32]proinsulin when incubated with Triton X-100 solubilized liver endosomal extracts (50) and stably expressed in rat Fao hepatoma cells (51). In these studies, proinsulin processing was attributed to furin, a subtilisin-like endoprotease which shows a preference for pairs of basic residues preceded by another basic residue at the -4 position and processes a number of proteins in liver (52). Our inability to show cleavage of proinsulin at the Arg31-Arg32 bond despite the presence of furin in liver Golgi-endosomal fractions (53) may be explained by two reasons. First, endosomes were disrupted by hypotonic shock, conditions under which furin should not be extracted because it is a membrane-associated enzyme; second, incubations were carried out in the absence of Ca2+, which is required for furin activity (52).
Taken together, the results of our in vivo and in vitro studies suggest that a reduced sensitivity of proinsulin to endosomal cathepsin-D may account, at least in part, for its slower hepatic clearance compared with insulin. Consistent with this, degradation rates of insulin, intact proinsulin, and split conversion intermediates in intact and hypotonically disrupted endosomes were found to vary in the same rank order as their hepatic clearances (16). However, because the ability of PM and cytosolic fractions to degrade proinsulin in vitro is also reduced (and, in fact, proportionally more than that of the endosomal fraction), a lower sensitivity of proinsulin to proteases at these loci must also be considered. One obvious candidate is IDE, an evolutionarily conserved neutral thiol-metalloendopeptidase, which seems to play a major role in cellular insulin degradation (43, 45). Indeed, IDE is recovered mainly in the cytosol but has also been identified in isolated plasma membranes and at surface of cells overexpressing (or not) IDE (43, 45, 54). In addition, in the present study, the sensitivity of proinsulin to the degrading activity associated with plasma membrane and cytosol was reduced to the same extent as that to purified IDE (18, 43, 44, 45).
The slower development and longer duration of proinsulin hepatic effects in vivo (5, 6, 7) led us to compare insulin and proinsulin effects on insulin receptor expression and tyrosine phosphorylation in liver subcellular fractions. As acute in vivo insulin treatment (24, 31, 46), proinsulin treatment led to a rapid and reversible endocytosis and tyrosine phosphorylation of the insulin receptor, but such effects differed with respect to dose- and time-dependence. Consistent with the lower affinity of proinsulin for the insulin receptor, on a molar basis, proinsulin was about 10–20 times less potent than insulin in its ability to increase endosomal receptor content and tyrosine phosphorylation. In time studies, administration of an inframaximal dose of proinsulin led, relative to insulin, to an increased residence time of the insulin receptor in endosomes as judged on insulin binding activity and receptor ß-subunit expression, and to a prolonged tyrosine phosphorylation of the ß-subunit (both at the plasma membrane and in endosomes). The most likely explanation of these observations is the slower clearance of proinsulin from blood plasma (4) and at these subcellular sites (present study), which, despite the low affinity of proinsulin for the insulin receptor, may keep the receptor occupied for a longer duration.
Genetically engineered insulin analogs, covering a wide range of receptor affinities, have been shown to differ from native insulin in their abilities to be taken up by and cleared from rat liver in vivo (55). We have previously shown that the subcellular distribution and metabolic fate of [HisA8, HisB4, GluB10, HisB27]insulin, a high-affinity analog slowly cleared from the liver, somewhat resemble those of proinsulin (24, 25). Thus, relative to insulin, this analog showed a prolonged accumulation and slower processing in endosomes in vivo and in vitro, it underwent a greater cell-free endosome-lysosome transfer, it was more slowly degraded by soluble endosomal proteases, and it induced in vivo a more prolonged increase in endosomal insulin receptor content and tyrosine phosphorylation. Along with the present results, these observations reinforce the concept that the endosomal processing of internalized ligands is an important determinant in ligand and receptor trafficking, as well as signal transduction, in liver cells.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received December 17, 2002.
Accepted for publication September 2, 2003.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
