0013.7227/92/1311-0447$03.00/0 Endocrinology Copyright 0 1992 by The Endocrine
Vol. 131, No. 1 Printed in U.S A.
Society
Ligand-Mediated Internalization in Intact Rat Liver FRANCOIS INSERM
AUTHIER*, lJ.30, H6pital
BERNARD
DESBUQUOIS,
des Enfants-Malades,
AND
of Glucagon BRIGITTE
Receptors
DE GALLE
75015 Paris, France
ABSTRACT The ligand-induced internalization of the hepatic glucagon receptor has been studied in rats in uiuo using cell fractionation. Injection of glucagon (11 nmol/lOO g BW) led to a 2. to d-fold increase in glucagonbinding activity in Golgi-endosomal (GE) fractions along with a lo20% decrease in binding activity in plasma membrane (PM) fractions. These changes were time and dose dependent, reaching a maximum by 12-24 min and undergoing reversal in 2 h. Glucagon injection also caused a 20% decrease in glucagon binding to the total particulate fraction, which did not occur when binding was measured in the presence of the detergent octylglucoside. The change in glucagonhinding activity in PM and GE fractions resulted mainly from a change in receptor number; affinity remained unaffected (apparent Kd, 0.5 and 5 nM, respectively). A 5- to lo-fold increase in the glucagon content of GE fractions was observed in glucagon-treated rats. Neither the distribution of PM and Golgi marker enzymes nor that of the asialoglycoprotein receptor was affected by glucagon treatment. Regardless of glucagon treatment, glucagon receptors in GE fractions were less sensitive to GTP than receptors in PM fractions with respect to both
inhibition of steady state binding and dissociation of prebound ligand. On sodium dodecyl sulfate-polyacrylamide gel electrophoresis, glucagon-receptor complexes formed in PM and GE fractions and subsequently cross-linked showed the same apparent mol wt (57 kilodaltons). In addition, they were identically sensitive to N-glycanase treatment, with two major species of lower mol wt generated. However, only crosslinked complexes associated with PM fractions showed detectable GTP sensitivity. GE fractions displayed a GTP-sensitive adenylate cyclase activity that was about 12 times lower than that of PM fractions. In both fractions, activity was stimulated by the addition of forskolin (8. fold) and, to a lesser extent, glucagon (a-fold). In uiuo glucagon treatment led to an increase in activity in GE, but not PM, fractions. These results are consistent with the view that upon acute occupancy, hepatic glucagon receptors are rapidly and specifically internalized along with their ligand. During this process, receptors retain structural integrity and uncouple, albeit partially, from other components of the adenylate cyclase system. (Endocrinology 131: 447-457, 1992)
R
that glucagon receptor number is reduced in liver plasma membranes (7-9) and intact liver cells (9, 10) of glucagontreated rats as well as in liver cells exposed to glucagon in vitro (11) suggeststhat the receptor may be internalized along with the ligand. To addressthis question, the effect of in uivo glucagon treatment on the subcellular distribution of the hepatic glucagon receptor has been examined in rats, using a biochemical approach similar to that used to study the fate of the radiolabeled ligand (5). In addition, some functional and structural properties of the internalized receptor have been characterized.
ECEPTOR -mediated endocytosis of glucagon in the hepatocyte, its major target cell, is well established. This event has been morphologically and/or biochemically documented with the use of ““I-labeled and colloidal goldlabeled ligands in isolated hepatocytes (l-3), isolated perfused liver (4), and intact liver in viva (5, 6). Studies with isolated hepatocytes have shown that, with time, cell-associated glucagon becomeslessdissociableby acid, suggesting its disappearance from the cell surface (3). On electron microscopy, internalized ligand has been shown to associate preferentially with lysosome-like structures in rat hepatocytes (1) and with both coated and noncoated vesicles in mousecells (2). Studies with intact rat liver have shown that, on cell fractionation, internalized glucagon accumulates mainly in low density components physically separablefrom Golgi elements and lysosomes, presumably endosomes (5, 6). Furthermore, these components have recently been identified as one major site at which internalized glucagon is degraded; ATP-dependent acidification is involved in this process(5, 6). Although suggesting that it is the glucagon-receptor complex that is internalized, the studies described above provide no direct information on the fate of the receptor. The finding Received November 18, 1991. Address all correspondence and requests for reprints to: Dr. Bernard Desbuquois, INSERM U.30, Hbpital des Enfants-Malades, tour Lavoisier, 6Pme &age, 149 rue de S&res, 75015 Paris, France. * Present address is: Department of Anatomy and Medicine, McGill University, Montreal, Quebec, Canada.
Materials
and Methods
Chemicals Porcine nlucagon, porcine insulin, and rabbit antisera against porcine glucagon (UK 4053 a;d K 44) were purchased from N&o Research Industries (Cooenhagen. Denmark). Galactosvlated BSA (Gal-BSA) was prepared f;om’ BSA vby ieductive lactosamination (12). Carrier-free ‘?, [c?P]ATP, and [3H]cAMP were obtained from the Radiochemical Center (Amersham, Aylesbury, Buckinghamshire, United Kingdom). N-OctylD-o-glucopyranoside (octylglucoside), ATP, GTP, creatine kinase, and phosphocreatine were purchased from Sigma (St. Louis, MO). 1,4-Difluoro-2,5-dinitrobenzene (DFDNB) and bis-[2-(succinimidooxycarbonyloxy)ethyl]sulfone (BSCOES) were obtained from Pierce Chemical Co. (Oud-Beijerland, The Netherlands). Glycanase (25,000 U/mg) was purchased from Boehringer (Mannheim, Germany). Chemicals for sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis, including mol wt markers, were purchased from Bio-Rad (Richmond, CA). Kodak XOmat AR-5 films were obtained from Eastman Kodak (Rochester, NY). The sources of other chemicals have been specified previously (5, 13).
447
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 11:53 For personal use only. No other uses without permission. . All rights reserved.
448 Animals
LIGAND-MEDIATED
INTERNALIZATION
Male Sprague-Dawley rats, weighing 180-200 g, were obtained from Charles River France (St.-Aubin-Les-Elbeuf, France) and fed ad libitum until death. Native glucagon (0.1-75 fig) or, in some cross-linking experiments, [“51]iodoglucagon (15-30 X lo6 cpm) was diluted into 0.5 ml 0.15 M NaCl containing 0.1% (wt/vol) BSA and injected over 15 set in the penis vein under light ether anesthesia. In studies with native glucagon, uninjected or saline-injected rats were used as controls. At the indicated times, the liver was removed from the abdominal cavity through a medium incision and immediately homogenized.
fractionation
This was performed using established procedures with minor modifications (5, 6, 13). A total particulate fraction was isolated from homogenates diluted to l-2 mg/ml protein in 5 rnM Tris-HC1 buffer, pH 7.4, by centrifugation at 40,000 X g for 30 min. A microsomal fraction and light, intermediate, and heavy Golgi-endosomal (GE) fractions (respectivedensities, 1.03-1.08, 1.08-1.11, and 1.11-1.18g/cm-3) were isolated from homogenates in 0.25 M sucrose according to the procedure of Ehrenreich et al. (14). In some experiments, a single total GE fraction (density, 1.03-1.14 g/cmm3) was prepared. Plasma membranes (PM) were prepared from homogenates in 1 mM bicarbonate according to the method of Neville up to step 11 (15). Subcellular fractions of control and glucagon-injected rats were stored in liquid nitrogen or at -80 C until used for biochemical determinations; only after the injection of [‘251]iodoglucagon were fractions immediately used for study.
Preparation assays
of radioiodinated
ligands and receptor binding
[“?]Iodoglucagon (at least 700 Ci/mmol) was prepared using chloramine-T (16) or lactoperoxidase (5,17) and purified by adsorption on talc (16) or reverse phase HPLC on a micro-Bondapak Cl8 column (5), respectively. [“sI]Iodo-Gal-BSA (2500 Ci/mmol) was prepared using lactoperoxidase and purified by gel filtration on Sephadex G-25. Incubation mixtures for measurements of glucagon binding to its receotor contained, in 0.2-0.3 ml 50 mM Tris-HC1 buffer, oH 7.4, O.Ol0.3dnM [‘251]iodoglucagon, 0.05 (PM) or 0.3 (GE) mg/mi cell fraction protein, 1% (wt/vol) BSA, and 0.1% (wt/vol) bacitracin. To measure nonspecific binding, parallel incubations were carried out in the presence of 5 F~M native glucagon. After 60 min at 20-23 C, membrane-bound [‘*?]iodoglucagon was precipitated by polyethylene glycol in the presence of bovine y-globulin, as described for precipitation of solubilized insulin-receptor complexes (13). The precipitate was counted, with results related to cell fraction protein and corrected for nonspecific binding (-6-8% of the total radioactivity added). In some experiments with PM fractions, membrane-bound glucagon was isolated by centrifugation at 10,000 X g for 10 min. However, only with the PM fractions were results comparable to those obtained using the polyethylene glycol precipitation method. Measurement of [iZ51]iodo-Gal-BSA binding to the asialoglycoprotein receptor was performed according to the method of Pricer and Ashwell (18) with slight modifications. Incubation mixtures contained, in 0.2 ml 50 mM Tris-HCI buffer, pH 7.4, 75 mM NaCI, 5 rnM CaC12, 0.1% (wt/ vol) Triton X-100, 0.5% (wt/vol) BSA, 0.5 nM [‘251]iodo-Gal-BSA, and lo-40 fig cell fraction protein. For measurement of nonspecific binding, 10 PM native Gal-BSA was also included. After 12 h at 4 C, receptorbound ligand was precipitated by polyethylene glycol, as described above, with results corrected for nonspecific binding (-5% of the total radioactivity added).
Extraction fractions
and measurement
of
glucagon
associated with GE
Glucagon associated with GE fractions was extracted by 0.1 N HCl containing 1% (wt/vol) BSA and 0.1% (wt/vol) bacitracin, as described previously for insulin extraction (13). The concentration of glucagon in the neutralized supernatants was determined by RIA, using porcine
Endo. Voll31.
RECEPTORS
glucagon as standard. Antibody-bound free hormone by polyethylene glycol
and injections
Liver subcellular
OF GLUCAGON
Cross-linking of receptor-bound subcellular fractions
glucagon precipitation,
1992 No 1
was separated from as described above.
~251]iodoglucagon
in liver
This was achieved using as cross-linkers DFDNB and BSCOES, basically as described by Sheetz and Tager (19). In the first set of experiments, [‘*51]iodoglucagon was allowed to bind to its receptor in vitro before the cross-linking step. PM or GE fractions (0.1 and 0.4 mg protein, respectively) were incubated under constant shaking with lo6 cpm [‘251]iodoglucagon in 1 ml 50 mM sodium phosphate buffer, pH 7.4, containing 0.1% (wt/vol) BSA and, when indicated, 2 PM native glucagon, 2 PM insulin, or 0.1 mM GTP. After 60 min at 20-22 C, 7 ml sodium phosphate buffer were added, and the membranes were sedimented by centrifugation at 40,000 X g for 15 min. The membranes were then resuspended in 1 ml buffer, after which 10 ~1 0.2 M DFDNB or 10 ~1 0.1 M BSCOES that had been freshly dissolved in dimethylsulfoxide were added. After 30 min at 20-22 C, incubation mixtures were diluted with 7 ml ice-cold 10 rnM Tris-HC1 buffer, pH 7.4, and the membranes were sedimented by centrifugation at 40,000 x g for 30 min. Routinely, about 30-40% of the [‘251]iodoglucagon bound to membranes before crosslinking was recovered in the final pellet. In the second set of experiments, [‘?]iodoglucagon-receptor complexes formed and internalized into liver endosomes in vivo were subjected to cross-linking. In this case, a total GE fraction was isolated 20 min after the injection of [‘251]iodoglucagon and suspended (0.4 mg/ml protein) into 1 ml 50 mM sodium phosphate buffer, pH 7.4. Membranes were allowed to react with 2 mM DFNDB, as described above, either immediately or after a 30.min incubation with 0.1 mM GTP. After 30 min at 20-22 C, incubation mixtures were diluted, and membrane material was sedimented by centrifugation, as described above.
Endoglycosidase complexes
digestion
of
cross-linked
glucagon-receptor
Subcellular fractions containing covalent [‘?]iodoglucagon-receptor complexes formed in vitro or in vivo were treated with N-glycanase according to the method of lwanij and Vincent (20) with minor modifications. PM or GE fractions (50 and 100 wg protein, respectively) were suspended in 0.10 ml 0.1 M sodium phosphate buffer containing 25 mM EDTA, 1% (wt/vol) NonidettP40, 0.2% (wt/vol) SDS, and 1% (wt/vol) 2-mercaptoethanol. These suspensions were incubated with 1 IJ Nglycanase for l-20 h at 37 C under constant shaking, after which the membranes were sedimented by centrifugation.
SDS-polyacrylamide
gel electrophoresis
and autoradiography
SDS-polyacrylamide slab gel electrophoresis was performed according to the method of Laemmli (21). Before electrophoresis, membrane pellets obtained as described above were solubilized in 1% (wt/vol) SDS containing 1% (vol/vol) 2-mercaptoethanol and boiled for 5 min. Samples (50-300 pg protein in a volume of 50-100 ~1) were applied on a 10% gel with a 3% stacking gel, and electrophoresis was carried out at 30 mamp for 5-6 h. Mol wt marker proteins (phosphorylase-b, 97,400; BSA, 66,200; ovalbumin, 42,699; carbonic anhydrase, 31,000; soybean trypsin inhibitor, 21,500; lysozyme, 14,400) were electrophoresed in one lane of the gel along with cell fractions. After electrophoresis, the gels were stained with Coomassie blue, destained, and dried. The dried gels were exposed to Kodak X-Omat AR-5 films in cassettes equipped with intensifying screens for 1-3 days at -80 C. After development of the autoradiograms, the mol wt of labeled components was deduced from their mobility by reference to the standard proteins.
Protein
and enzyme assays
The protein concentration was measured according to the method of Lowry et al. (22), using BSA as a standard. Adenylate cyclase activity was measured as described by Salomon et al. (23), with minor modifi-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 11:53 For personal use only. No other uses without permission. . All rights reserved.
LIGAND-MEDIATED
INTERNALIZATION
OF GLUCAGON
RECEPTORS
449
cations. Briefly, incubation mixtures contained, in a final volume of 0.1 ml, 25 rnM Tris-HCl buffer ( H 7.6), 5 rnM M&l,, 1 rnM EDTA, 0.5 mM [32P]ATP (-2 X lo6 cpm), [‘HIcAMP (-lo4 cpm), 0.1 rnM GTP, 20 pg creatine kinase, 0.5 mg phosphocreatin, and 30-300 pg cell fraction protein. After incubation for 10 min at 30 C, the reaction was stopped by boiling for 3 min, and CAMP was isolated by double column chromatography according to the method of Salomon et al. (23). S’-Nucleotidase (24), galactosyltransferase (25), and acid phosphatase (26) activities were measured according to procedures described previously.
ligand of asialoglycoprotein receptor; the latter binding activity was equally concentrated in the PM and GE fractions. These results suggestthat although someglucagon receptors may be intrinsic to Golgi elements and/or endosomes,most receptors are associatedwith the plasma membrane.
Statistical
The injection of 21 nmol glucagon, a dose that inhibits by 85% the hepatic uptake of [‘251]iodoglucagonin viuo (results not shown), caused an increase in glucagon-binding activity in GE fractions, along with a decreasein binding activity in the PM fraction. In time studies (Fig. l), both changes were detectable shortly after glucagon injection, achieved a maximum by 6-24 min, and were fully reversed in 2 h. Glucagonbinding activity increased somewhat earlier in the heavy and intermediate GE fractions than in the light fraction, as did uptake of [‘251]iodoglucagonin viva (5). Relative to those in control (noninjected) animals, the maximal changes in glucagon-binding activity were about 2- to 3-fold in GE fractions, but only 20% in the PM fraction. However, when expressed as a percentage of the total hepatic glucagonbinding activity (i.e. associated with the total particulate fraction), the relative changes in activity were of the same order of magnitude in PM and GE fractions (-l-2%). The injection of glucagon also resulted, at early times postinjection, in a 20% decrease in the glucagon-binding activity associatedwith the total particulate fraction (Fig. 2). However, this decrease was virtually abolished when the binding reaction was carried out in the presence of the detergent octylglucoside. These findings presumably reflect a reduced accessibility of the receptors to glucagon in the medium due to their sequestration in endosomes.They also suggestthat the increase in glucagon-binding activity in GE fractions of glucagon-treated rats may be greater than actually measured. The ability of glucagon treatment to increase glucagon-
analysis
Results were between groups
expressed as the mean + SEM. were performed using Student’s
Statistical t test.
comparisons
Results Subcellular liver
distribution
of
glucagon-binding
sites in normal
rat
The subcellular distribution of glucagon-binding activity in normal rat liver was first examined and compared to those of various membrane constituents (Table 1). PM and GE fractions prepared by standard procedures were selected for study becausethey were shown to sequentially concentrate [‘251]iodoglucagontaken up by liver in viva (5). At 0.3 nM glucagon, PM bound 8 times as much glucagon as the total particulate fraction, with a recovery of 7%. In contrast, light, intermediate, and heavy GE fractions showed no enrichment in glucagon-binding activity, with a summed recovery of only 0.5%. As shown in a previous study (27) and confirmed here (seeFig. 4 and associated text), the reduced glucagonbinding activity of GE fractions, relative to that of the PM fractions, resulted mainly from a lower receptor affinity. The distribution of glucagon-binding activity was comparable to that of adenylate cyclase and 5’-nucleotidase, two plasma membrane marker enzymes, except for a slight enrichment of GE fractions in the latter enzyme. In contrast, it markedly differed from the distribution of galactosyltransferase, a Golgi marker enzyme. The binding of glucagon to liver cell fractions also differed from that of Gal-BSA, a TABLE
1. Subcellular
distribution
of glucagon-binding
Subcellular fraction
sites and other
Glucagon binding
PM
RSA Recovery
8.4 + 0.9 7.9
GE light
RSA Recovery
0.74 + 0.10 0.04
GE intermediate
RSA Recovery
0.86 k 0.09 0.13
GE heavy
RSA Recovery
0.98 + 0.08 0.29
Effect of glucagon injection glucagon-binding activity
biochemical
5’-Nucleotidase
constituents Adenylate cyclase
9.4 f 0.8 8.8
4.9 + 0.2 4.6
2.39 1- 0.32” 0.94
0.38 + 0.05” 0.15
on the subcellular
in livers
of control Galactosyl transferase
distribution
of
rats Gal-BSA binding 2.85 f 0.12 2.65
20.4 k 2.8” 8.0
2.34 f 0.23” 0.92
A PM fraction and GE light, intermediate, and heavy fractions were isolated and assayed for ligand binding and enzymatic activities, as described in Materials and Methods. The concentration of glucagon in the assay medium for measurement of glucagon binding was 0.3 nM. Relative specific activities of constituents (RSA) were determined using as a reference the specific activities in the total particulate fraction, with results expressed as the mean + SEM of six to nine determinations. Specific activities of constituents in the total particulate fraction were, respectively: glucagon binding, 156 + 10 fmol/mg; Gal-BSA binding, 679 + 26 fmol/mg; adenylate cyclase, 16.5 & 1.9 pmol/min.mg; 5’-nucleotidase, 66 + 4 nmol/min mg; and galactosyltransferase, 40 + pmol/min . mg (mean of SEM of at least six determinations). Recoveries were calculated from relative specific activities and protein yields, with results expressed as percentages of activities in the total particulate fraction. ’ Measurement performed on a total GE fraction.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 11:53 For personal use only. No other uses without permission. . All rights reserved.
450
LIGAND-MEDIATED
INTERNALIZATION
OF GLUCAGON
0
RECEPTORS
10
20
Endo. Vol131.
30
40
Time of sacrifice glucagon injection
+--F--F+ 20
30
Time of sacrifice glucagon injection
after (min)
FIG. 1. Changes in glucagon-binding activity in liver PM and GE fractions at various times after glucagon injection. Rats were killed at the indicated times after the injection of 21.6 nmol glucagon/ZOO g BW; untreated rats were used as controls. PM and light (GEl), intermediate (GEi), and heavy (GEh) GE fractions were prepared and examined for glucagon-binding activity. The concentration of [“‘I] iodoglucagon in the assay mixture was 0.1-0.3 nM. The results are expressed as a percentage of the average glucagon-binding activity in control rats; they are the mean + SEM of four to eight determinations on subcellular fractions originating from separate livers. Binding activities achieved between 6-40 min after glucagon injection were statistically different from control values (P < 0.01).
1992 No 1
120 after (min)
FIG. 2. Changes in glucagon-binding activity in the total particulate fractions at various times after glucagon injection. Rats were killed at the indicated times after the injection of glucagon (21.6 nmol/200 g BW); untreated rats were used as controls. A total particulate fraction was prepared and examined for glucagon-binding activity in the absence (H) and presence (0) of 0.05% (wt/vol) p-octylglucoside. The concentration of [““Iliodoglucagon in the assay mixture was 0.3 nM. The results are expressed as a percentage of the average glucagon-binding activity in control rats; they are the mean & SEM of three to six determinations on subcellular fractions originating from separate livers. Octylglucoside treatment decreased by only 6.3 f. 0.6% glucagon binding to particulate fractions of control rats (not shown in the figure). Binding activities achieved 1.5-24 min after glucagon treatment were statistically different in the absence and presence of octylglucoside (1’ < 0.01).
x2 .z 5z
250:
binding activity in GE fractions was dose dependent (Fig. 3). At 20 min, this change was detectable at 30 pmol, halfmaximal at about 1 nmol, and maximal at 22 nmol. Glucagon content of liver cell fractions treated rats
in control
and glucagon-
Subcellular fractions of control rats contained low or undetectable amounts of immunoreactive glucagon (Table 2). Subcellular fractions of glucagon-treated rats displayed a marked increase in glucagon content, which sequentially affected the PM fraction, the heavy GE fraction, and the intermediate and light GE fractions (maximum, -1.5, 6, and 12 min, respectively). At peak uptake, more glucagon was associatedwith GE fractions (-1-2 pmol/mg protein) than with PM fractions (-0.3 pmol/mg protein). It was estimated that at the maximal concentrations achieved in cell fractions, endogenous glucagon led to a 20% underestimation of glucagon-binding activity in GE fractions, but had little effect on this activity in the PM fraction. Specificity of ligand-induced of glucagon-binding activity
changes in subcellular
distribution
Glucagon treatment caused little or no change in the specific activities of 5’-nucleotidase, galactosyltransferase, acid phosphatase, and Gal-BSA-binding activity in PM and
0.1
1
Dose of glucagon (nmo1/200 g body
10
100
injected weight)
FIG. 3. Dose response of change in glucagon-binding activity in GE fractions. Rats were killed 24 min after glucagon injection at the indicated dose. A total GE fraction was prepared and examined for glucagon-binding activity. The concentration of [““I]iodoglucagon in the assay mixture was between 0.1-0.3 nM. The results are expressed as a percentage of the binding activity in control rats; they are the mean + SEM of three or four determinations on cell fractions originating from separate livers. Injection of 57.4 nmol glucagon did not further increase glucagon-binding activity (not shown in the figure).
GE fractions (Table 3). Likewise, glucagon treatment little affected the recovery of protein in these fractions. At 20 min, the amounts of protein recovered in a total GE fraction were comparable in glucagon-injected and saline-injected rats (0.78 + 0.07% and 0.81 + 0.06% of total microsomal protein, respectively; mean f SEM of three to six determinations). These findings indicate that the glucagon-induced redistribution of glucagon-binding sitesis a fairly specific process.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 11:53 For personal use only. No other uses without permission. . All rights reserved.
LIGAND-MEDIATED TABLE
2. Glucagon Subcellular
content
fraction
PM GE light GE intermediate GE heavv
of liver Control
INTERNALIZATION
subcellular
fractions
of control Glucagon
rats
0.28 + 0.04 0.29 + 0.03 0.46 + 0.08
and glucagon-injected
content
1.5
Vol. 131, No. 1 Printed in U.S A.
Society
Ligand-Mediated Internalization in Intact Rat Liver FRANCOIS INSERM
AUTHIER*, lJ.30, H6pital
BERNARD
DESBUQUOIS,
des Enfants-Malades,
AND
of Glucagon BRIGITTE
Receptors
DE GALLE
75015 Paris, France
ABSTRACT The ligand-induced internalization of the hepatic glucagon receptor has been studied in rats in uiuo using cell fractionation. Injection of glucagon (11 nmol/lOO g BW) led to a 2. to d-fold increase in glucagonbinding activity in Golgi-endosomal (GE) fractions along with a lo20% decrease in binding activity in plasma membrane (PM) fractions. These changes were time and dose dependent, reaching a maximum by 12-24 min and undergoing reversal in 2 h. Glucagon injection also caused a 20% decrease in glucagon binding to the total particulate fraction, which did not occur when binding was measured in the presence of the detergent octylglucoside. The change in glucagonhinding activity in PM and GE fractions resulted mainly from a change in receptor number; affinity remained unaffected (apparent Kd, 0.5 and 5 nM, respectively). A 5- to lo-fold increase in the glucagon content of GE fractions was observed in glucagon-treated rats. Neither the distribution of PM and Golgi marker enzymes nor that of the asialoglycoprotein receptor was affected by glucagon treatment. Regardless of glucagon treatment, glucagon receptors in GE fractions were less sensitive to GTP than receptors in PM fractions with respect to both
inhibition of steady state binding and dissociation of prebound ligand. On sodium dodecyl sulfate-polyacrylamide gel electrophoresis, glucagon-receptor complexes formed in PM and GE fractions and subsequently cross-linked showed the same apparent mol wt (57 kilodaltons). In addition, they were identically sensitive to N-glycanase treatment, with two major species of lower mol wt generated. However, only crosslinked complexes associated with PM fractions showed detectable GTP sensitivity. GE fractions displayed a GTP-sensitive adenylate cyclase activity that was about 12 times lower than that of PM fractions. In both fractions, activity was stimulated by the addition of forskolin (8. fold) and, to a lesser extent, glucagon (a-fold). In uiuo glucagon treatment led to an increase in activity in GE, but not PM, fractions. These results are consistent with the view that upon acute occupancy, hepatic glucagon receptors are rapidly and specifically internalized along with their ligand. During this process, receptors retain structural integrity and uncouple, albeit partially, from other components of the adenylate cyclase system. (Endocrinology 131: 447-457, 1992)
R
that glucagon receptor number is reduced in liver plasma membranes (7-9) and intact liver cells (9, 10) of glucagontreated rats as well as in liver cells exposed to glucagon in vitro (11) suggeststhat the receptor may be internalized along with the ligand. To addressthis question, the effect of in uivo glucagon treatment on the subcellular distribution of the hepatic glucagon receptor has been examined in rats, using a biochemical approach similar to that used to study the fate of the radiolabeled ligand (5). In addition, some functional and structural properties of the internalized receptor have been characterized.
ECEPTOR -mediated endocytosis of glucagon in the hepatocyte, its major target cell, is well established. This event has been morphologically and/or biochemically documented with the use of ““I-labeled and colloidal goldlabeled ligands in isolated hepatocytes (l-3), isolated perfused liver (4), and intact liver in viva (5, 6). Studies with isolated hepatocytes have shown that, with time, cell-associated glucagon becomeslessdissociableby acid, suggesting its disappearance from the cell surface (3). On electron microscopy, internalized ligand has been shown to associate preferentially with lysosome-like structures in rat hepatocytes (1) and with both coated and noncoated vesicles in mousecells (2). Studies with intact rat liver have shown that, on cell fractionation, internalized glucagon accumulates mainly in low density components physically separablefrom Golgi elements and lysosomes, presumably endosomes (5, 6). Furthermore, these components have recently been identified as one major site at which internalized glucagon is degraded; ATP-dependent acidification is involved in this process(5, 6). Although suggesting that it is the glucagon-receptor complex that is internalized, the studies described above provide no direct information on the fate of the receptor. The finding Received November 18, 1991. Address all correspondence and requests for reprints to: Dr. Bernard Desbuquois, INSERM U.30, Hbpital des Enfants-Malades, tour Lavoisier, 6Pme &age, 149 rue de S&res, 75015 Paris, France. * Present address is: Department of Anatomy and Medicine, McGill University, Montreal, Quebec, Canada.
Materials
and Methods
Chemicals Porcine nlucagon, porcine insulin, and rabbit antisera against porcine glucagon (UK 4053 a;d K 44) were purchased from N&o Research Industries (Cooenhagen. Denmark). Galactosvlated BSA (Gal-BSA) was prepared f;om’ BSA vby ieductive lactosamination (12). Carrier-free ‘?, [c?P]ATP, and [3H]cAMP were obtained from the Radiochemical Center (Amersham, Aylesbury, Buckinghamshire, United Kingdom). N-OctylD-o-glucopyranoside (octylglucoside), ATP, GTP, creatine kinase, and phosphocreatine were purchased from Sigma (St. Louis, MO). 1,4-Difluoro-2,5-dinitrobenzene (DFDNB) and bis-[2-(succinimidooxycarbonyloxy)ethyl]sulfone (BSCOES) were obtained from Pierce Chemical Co. (Oud-Beijerland, The Netherlands). Glycanase (25,000 U/mg) was purchased from Boehringer (Mannheim, Germany). Chemicals for sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis, including mol wt markers, were purchased from Bio-Rad (Richmond, CA). Kodak XOmat AR-5 films were obtained from Eastman Kodak (Rochester, NY). The sources of other chemicals have been specified previously (5, 13).
447
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 11:53 For personal use only. No other uses without permission. . All rights reserved.
448 Animals
LIGAND-MEDIATED
INTERNALIZATION
Male Sprague-Dawley rats, weighing 180-200 g, were obtained from Charles River France (St.-Aubin-Les-Elbeuf, France) and fed ad libitum until death. Native glucagon (0.1-75 fig) or, in some cross-linking experiments, [“51]iodoglucagon (15-30 X lo6 cpm) was diluted into 0.5 ml 0.15 M NaCl containing 0.1% (wt/vol) BSA and injected over 15 set in the penis vein under light ether anesthesia. In studies with native glucagon, uninjected or saline-injected rats were used as controls. At the indicated times, the liver was removed from the abdominal cavity through a medium incision and immediately homogenized.
fractionation
This was performed using established procedures with minor modifications (5, 6, 13). A total particulate fraction was isolated from homogenates diluted to l-2 mg/ml protein in 5 rnM Tris-HC1 buffer, pH 7.4, by centrifugation at 40,000 X g for 30 min. A microsomal fraction and light, intermediate, and heavy Golgi-endosomal (GE) fractions (respectivedensities, 1.03-1.08, 1.08-1.11, and 1.11-1.18g/cm-3) were isolated from homogenates in 0.25 M sucrose according to the procedure of Ehrenreich et al. (14). In some experiments, a single total GE fraction (density, 1.03-1.14 g/cmm3) was prepared. Plasma membranes (PM) were prepared from homogenates in 1 mM bicarbonate according to the method of Neville up to step 11 (15). Subcellular fractions of control and glucagon-injected rats were stored in liquid nitrogen or at -80 C until used for biochemical determinations; only after the injection of [‘251]iodoglucagon were fractions immediately used for study.
Preparation assays
of radioiodinated
ligands and receptor binding
[“?]Iodoglucagon (at least 700 Ci/mmol) was prepared using chloramine-T (16) or lactoperoxidase (5,17) and purified by adsorption on talc (16) or reverse phase HPLC on a micro-Bondapak Cl8 column (5), respectively. [“sI]Iodo-Gal-BSA (2500 Ci/mmol) was prepared using lactoperoxidase and purified by gel filtration on Sephadex G-25. Incubation mixtures for measurements of glucagon binding to its receotor contained, in 0.2-0.3 ml 50 mM Tris-HC1 buffer, oH 7.4, O.Ol0.3dnM [‘251]iodoglucagon, 0.05 (PM) or 0.3 (GE) mg/mi cell fraction protein, 1% (wt/vol) BSA, and 0.1% (wt/vol) bacitracin. To measure nonspecific binding, parallel incubations were carried out in the presence of 5 F~M native glucagon. After 60 min at 20-23 C, membrane-bound [‘*?]iodoglucagon was precipitated by polyethylene glycol in the presence of bovine y-globulin, as described for precipitation of solubilized insulin-receptor complexes (13). The precipitate was counted, with results related to cell fraction protein and corrected for nonspecific binding (-6-8% of the total radioactivity added). In some experiments with PM fractions, membrane-bound glucagon was isolated by centrifugation at 10,000 X g for 10 min. However, only with the PM fractions were results comparable to those obtained using the polyethylene glycol precipitation method. Measurement of [iZ51]iodo-Gal-BSA binding to the asialoglycoprotein receptor was performed according to the method of Pricer and Ashwell (18) with slight modifications. Incubation mixtures contained, in 0.2 ml 50 mM Tris-HCI buffer, pH 7.4, 75 mM NaCI, 5 rnM CaC12, 0.1% (wt/ vol) Triton X-100, 0.5% (wt/vol) BSA, 0.5 nM [‘251]iodo-Gal-BSA, and lo-40 fig cell fraction protein. For measurement of nonspecific binding, 10 PM native Gal-BSA was also included. After 12 h at 4 C, receptorbound ligand was precipitated by polyethylene glycol, as described above, with results corrected for nonspecific binding (-5% of the total radioactivity added).
Extraction fractions
and measurement
of
glucagon
associated with GE
Glucagon associated with GE fractions was extracted by 0.1 N HCl containing 1% (wt/vol) BSA and 0.1% (wt/vol) bacitracin, as described previously for insulin extraction (13). The concentration of glucagon in the neutralized supernatants was determined by RIA, using porcine
Endo. Voll31.
RECEPTORS
glucagon as standard. Antibody-bound free hormone by polyethylene glycol
and injections
Liver subcellular
OF GLUCAGON
Cross-linking of receptor-bound subcellular fractions
glucagon precipitation,
1992 No 1
was separated from as described above.
~251]iodoglucagon
in liver
This was achieved using as cross-linkers DFDNB and BSCOES, basically as described by Sheetz and Tager (19). In the first set of experiments, [‘*51]iodoglucagon was allowed to bind to its receptor in vitro before the cross-linking step. PM or GE fractions (0.1 and 0.4 mg protein, respectively) were incubated under constant shaking with lo6 cpm [‘251]iodoglucagon in 1 ml 50 mM sodium phosphate buffer, pH 7.4, containing 0.1% (wt/vol) BSA and, when indicated, 2 PM native glucagon, 2 PM insulin, or 0.1 mM GTP. After 60 min at 20-22 C, 7 ml sodium phosphate buffer were added, and the membranes were sedimented by centrifugation at 40,000 X g for 15 min. The membranes were then resuspended in 1 ml buffer, after which 10 ~1 0.2 M DFDNB or 10 ~1 0.1 M BSCOES that had been freshly dissolved in dimethylsulfoxide were added. After 30 min at 20-22 C, incubation mixtures were diluted with 7 ml ice-cold 10 rnM Tris-HC1 buffer, pH 7.4, and the membranes were sedimented by centrifugation at 40,000 x g for 30 min. Routinely, about 30-40% of the [‘251]iodoglucagon bound to membranes before crosslinking was recovered in the final pellet. In the second set of experiments, [‘?]iodoglucagon-receptor complexes formed and internalized into liver endosomes in vivo were subjected to cross-linking. In this case, a total GE fraction was isolated 20 min after the injection of [‘251]iodoglucagon and suspended (0.4 mg/ml protein) into 1 ml 50 mM sodium phosphate buffer, pH 7.4. Membranes were allowed to react with 2 mM DFNDB, as described above, either immediately or after a 30.min incubation with 0.1 mM GTP. After 30 min at 20-22 C, incubation mixtures were diluted, and membrane material was sedimented by centrifugation, as described above.
Endoglycosidase complexes
digestion
of
cross-linked
glucagon-receptor
Subcellular fractions containing covalent [‘?]iodoglucagon-receptor complexes formed in vitro or in vivo were treated with N-glycanase according to the method of lwanij and Vincent (20) with minor modifications. PM or GE fractions (50 and 100 wg protein, respectively) were suspended in 0.10 ml 0.1 M sodium phosphate buffer containing 25 mM EDTA, 1% (wt/vol) NonidettP40, 0.2% (wt/vol) SDS, and 1% (wt/vol) 2-mercaptoethanol. These suspensions were incubated with 1 IJ Nglycanase for l-20 h at 37 C under constant shaking, after which the membranes were sedimented by centrifugation.
SDS-polyacrylamide
gel electrophoresis
and autoradiography
SDS-polyacrylamide slab gel electrophoresis was performed according to the method of Laemmli (21). Before electrophoresis, membrane pellets obtained as described above were solubilized in 1% (wt/vol) SDS containing 1% (vol/vol) 2-mercaptoethanol and boiled for 5 min. Samples (50-300 pg protein in a volume of 50-100 ~1) were applied on a 10% gel with a 3% stacking gel, and electrophoresis was carried out at 30 mamp for 5-6 h. Mol wt marker proteins (phosphorylase-b, 97,400; BSA, 66,200; ovalbumin, 42,699; carbonic anhydrase, 31,000; soybean trypsin inhibitor, 21,500; lysozyme, 14,400) were electrophoresed in one lane of the gel along with cell fractions. After electrophoresis, the gels were stained with Coomassie blue, destained, and dried. The dried gels were exposed to Kodak X-Omat AR-5 films in cassettes equipped with intensifying screens for 1-3 days at -80 C. After development of the autoradiograms, the mol wt of labeled components was deduced from their mobility by reference to the standard proteins.
Protein
and enzyme assays
The protein concentration was measured according to the method of Lowry et al. (22), using BSA as a standard. Adenylate cyclase activity was measured as described by Salomon et al. (23), with minor modifi-
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 11:53 For personal use only. No other uses without permission. . All rights reserved.
LIGAND-MEDIATED
INTERNALIZATION
OF GLUCAGON
RECEPTORS
449
cations. Briefly, incubation mixtures contained, in a final volume of 0.1 ml, 25 rnM Tris-HCl buffer ( H 7.6), 5 rnM M&l,, 1 rnM EDTA, 0.5 mM [32P]ATP (-2 X lo6 cpm), [‘HIcAMP (-lo4 cpm), 0.1 rnM GTP, 20 pg creatine kinase, 0.5 mg phosphocreatin, and 30-300 pg cell fraction protein. After incubation for 10 min at 30 C, the reaction was stopped by boiling for 3 min, and CAMP was isolated by double column chromatography according to the method of Salomon et al. (23). S’-Nucleotidase (24), galactosyltransferase (25), and acid phosphatase (26) activities were measured according to procedures described previously.
ligand of asialoglycoprotein receptor; the latter binding activity was equally concentrated in the PM and GE fractions. These results suggestthat although someglucagon receptors may be intrinsic to Golgi elements and/or endosomes,most receptors are associatedwith the plasma membrane.
Statistical
The injection of 21 nmol glucagon, a dose that inhibits by 85% the hepatic uptake of [‘251]iodoglucagonin viuo (results not shown), caused an increase in glucagon-binding activity in GE fractions, along with a decreasein binding activity in the PM fraction. In time studies (Fig. l), both changes were detectable shortly after glucagon injection, achieved a maximum by 6-24 min, and were fully reversed in 2 h. Glucagonbinding activity increased somewhat earlier in the heavy and intermediate GE fractions than in the light fraction, as did uptake of [‘251]iodoglucagonin viva (5). Relative to those in control (noninjected) animals, the maximal changes in glucagon-binding activity were about 2- to 3-fold in GE fractions, but only 20% in the PM fraction. However, when expressed as a percentage of the total hepatic glucagonbinding activity (i.e. associated with the total particulate fraction), the relative changes in activity were of the same order of magnitude in PM and GE fractions (-l-2%). The injection of glucagon also resulted, at early times postinjection, in a 20% decrease in the glucagon-binding activity associatedwith the total particulate fraction (Fig. 2). However, this decrease was virtually abolished when the binding reaction was carried out in the presence of the detergent octylglucoside. These findings presumably reflect a reduced accessibility of the receptors to glucagon in the medium due to their sequestration in endosomes.They also suggestthat the increase in glucagon-binding activity in GE fractions of glucagon-treated rats may be greater than actually measured. The ability of glucagon treatment to increase glucagon-
analysis
Results were between groups
expressed as the mean + SEM. were performed using Student’s
Statistical t test.
comparisons
Results Subcellular liver
distribution
of
glucagon-binding
sites in normal
rat
The subcellular distribution of glucagon-binding activity in normal rat liver was first examined and compared to those of various membrane constituents (Table 1). PM and GE fractions prepared by standard procedures were selected for study becausethey were shown to sequentially concentrate [‘251]iodoglucagontaken up by liver in viva (5). At 0.3 nM glucagon, PM bound 8 times as much glucagon as the total particulate fraction, with a recovery of 7%. In contrast, light, intermediate, and heavy GE fractions showed no enrichment in glucagon-binding activity, with a summed recovery of only 0.5%. As shown in a previous study (27) and confirmed here (seeFig. 4 and associated text), the reduced glucagonbinding activity of GE fractions, relative to that of the PM fractions, resulted mainly from a lower receptor affinity. The distribution of glucagon-binding activity was comparable to that of adenylate cyclase and 5’-nucleotidase, two plasma membrane marker enzymes, except for a slight enrichment of GE fractions in the latter enzyme. In contrast, it markedly differed from the distribution of galactosyltransferase, a Golgi marker enzyme. The binding of glucagon to liver cell fractions also differed from that of Gal-BSA, a TABLE
1. Subcellular
distribution
of glucagon-binding
Subcellular fraction
sites and other
Glucagon binding
PM
RSA Recovery
8.4 + 0.9 7.9
GE light
RSA Recovery
0.74 + 0.10 0.04
GE intermediate
RSA Recovery
0.86 k 0.09 0.13
GE heavy
RSA Recovery
0.98 + 0.08 0.29
Effect of glucagon injection glucagon-binding activity
biochemical
5’-Nucleotidase
constituents Adenylate cyclase
9.4 f 0.8 8.8
4.9 + 0.2 4.6
2.39 1- 0.32” 0.94
0.38 + 0.05” 0.15
on the subcellular
in livers
of control Galactosyl transferase
distribution
of
rats Gal-BSA binding 2.85 f 0.12 2.65
20.4 k 2.8” 8.0
2.34 f 0.23” 0.92
A PM fraction and GE light, intermediate, and heavy fractions were isolated and assayed for ligand binding and enzymatic activities, as described in Materials and Methods. The concentration of glucagon in the assay medium for measurement of glucagon binding was 0.3 nM. Relative specific activities of constituents (RSA) were determined using as a reference the specific activities in the total particulate fraction, with results expressed as the mean + SEM of six to nine determinations. Specific activities of constituents in the total particulate fraction were, respectively: glucagon binding, 156 + 10 fmol/mg; Gal-BSA binding, 679 + 26 fmol/mg; adenylate cyclase, 16.5 & 1.9 pmol/min.mg; 5’-nucleotidase, 66 + 4 nmol/min mg; and galactosyltransferase, 40 + pmol/min . mg (mean of SEM of at least six determinations). Recoveries were calculated from relative specific activities and protein yields, with results expressed as percentages of activities in the total particulate fraction. ’ Measurement performed on a total GE fraction.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 11:53 For personal use only. No other uses without permission. . All rights reserved.
450
LIGAND-MEDIATED
INTERNALIZATION
OF GLUCAGON
0
RECEPTORS
10
20
Endo. Vol131.
30
40
Time of sacrifice glucagon injection
+--F--F+ 20
30
Time of sacrifice glucagon injection
after (min)
FIG. 1. Changes in glucagon-binding activity in liver PM and GE fractions at various times after glucagon injection. Rats were killed at the indicated times after the injection of 21.6 nmol glucagon/ZOO g BW; untreated rats were used as controls. PM and light (GEl), intermediate (GEi), and heavy (GEh) GE fractions were prepared and examined for glucagon-binding activity. The concentration of [“‘I] iodoglucagon in the assay mixture was 0.1-0.3 nM. The results are expressed as a percentage of the average glucagon-binding activity in control rats; they are the mean + SEM of four to eight determinations on subcellular fractions originating from separate livers. Binding activities achieved between 6-40 min after glucagon injection were statistically different from control values (P < 0.01).
1992 No 1
120 after (min)
FIG. 2. Changes in glucagon-binding activity in the total particulate fractions at various times after glucagon injection. Rats were killed at the indicated times after the injection of glucagon (21.6 nmol/200 g BW); untreated rats were used as controls. A total particulate fraction was prepared and examined for glucagon-binding activity in the absence (H) and presence (0) of 0.05% (wt/vol) p-octylglucoside. The concentration of [““Iliodoglucagon in the assay mixture was 0.3 nM. The results are expressed as a percentage of the average glucagon-binding activity in control rats; they are the mean & SEM of three to six determinations on subcellular fractions originating from separate livers. Octylglucoside treatment decreased by only 6.3 f. 0.6% glucagon binding to particulate fractions of control rats (not shown in the figure). Binding activities achieved 1.5-24 min after glucagon treatment were statistically different in the absence and presence of octylglucoside (1’ < 0.01).
x2 .z 5z
250:
binding activity in GE fractions was dose dependent (Fig. 3). At 20 min, this change was detectable at 30 pmol, halfmaximal at about 1 nmol, and maximal at 22 nmol. Glucagon content of liver cell fractions treated rats
in control
and glucagon-
Subcellular fractions of control rats contained low or undetectable amounts of immunoreactive glucagon (Table 2). Subcellular fractions of glucagon-treated rats displayed a marked increase in glucagon content, which sequentially affected the PM fraction, the heavy GE fraction, and the intermediate and light GE fractions (maximum, -1.5, 6, and 12 min, respectively). At peak uptake, more glucagon was associatedwith GE fractions (-1-2 pmol/mg protein) than with PM fractions (-0.3 pmol/mg protein). It was estimated that at the maximal concentrations achieved in cell fractions, endogenous glucagon led to a 20% underestimation of glucagon-binding activity in GE fractions, but had little effect on this activity in the PM fraction. Specificity of ligand-induced of glucagon-binding activity
changes in subcellular
distribution
Glucagon treatment caused little or no change in the specific activities of 5’-nucleotidase, galactosyltransferase, acid phosphatase, and Gal-BSA-binding activity in PM and
0.1
1
Dose of glucagon (nmo1/200 g body
10
100
injected weight)
FIG. 3. Dose response of change in glucagon-binding activity in GE fractions. Rats were killed 24 min after glucagon injection at the indicated dose. A total GE fraction was prepared and examined for glucagon-binding activity. The concentration of [““I]iodoglucagon in the assay mixture was between 0.1-0.3 nM. The results are expressed as a percentage of the binding activity in control rats; they are the mean + SEM of three or four determinations on cell fractions originating from separate livers. Injection of 57.4 nmol glucagon did not further increase glucagon-binding activity (not shown in the figure).
GE fractions (Table 3). Likewise, glucagon treatment little affected the recovery of protein in these fractions. At 20 min, the amounts of protein recovered in a total GE fraction were comparable in glucagon-injected and saline-injected rats (0.78 + 0.07% and 0.81 + 0.06% of total microsomal protein, respectively; mean f SEM of three to six determinations). These findings indicate that the glucagon-induced redistribution of glucagon-binding sitesis a fairly specific process.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 11:53 For personal use only. No other uses without permission. . All rights reserved.
LIGAND-MEDIATED TABLE
2. Glucagon Subcellular
content
fraction
PM GE light GE intermediate GE heavv
of liver Control
INTERNALIZATION
subcellular
fractions
of control Glucagon
rats
0.28 + 0.04 0.29 + 0.03 0.46 + 0.08
and glucagon-injected
content
1.5
