0013-7227/90/1271-0278$02.00/0 Endocrinology Copyright © 1990 by The Endocrine Society
Vol. 127, No. 1 Printed in U.S.A.
The Rat Hepatic Corticosteroid-Binding Globulin Receptor: Distinction from the Asialoglycoprotein Receptor* UTPALENDU S. MAITRA, M. SAEED KHAN, XIN HUA ZHANG, AND WILLIAM ROSNER Department of Medicine, St. Luke's/Roosevelt Hospital Center, Columbia University College of Physicians and Surgeons, New York, New York 10019
ABSTRACT. This investigation was undertaken to ascertain whether rat liver cells contained a receptor for corticosteoidbinding globulin (CBG) that could be differentiated clearly from the asialoglycoprotein receptor. To do this, [125I]CBG, [125I] asialo-CBG, and [125I]asialofetuin were used as probes to differentiate the binding activities of the two receptors. On hepatic membranes, CBG bound to a single set of sites with a Kd of 0.74 ftM, asialofetuin bound to a single set of sites with a Kd of 0.018 jiM, asialo-CBG bound to two sets of sites with Kd values of 0.004 and 1.4 /*M, and in the presence of 1 juM asialofetuin,
asialo-CBG bound to a single set of sites with a Kd of 0.53 /xM, not different (P > 0.2) from the Kd of CBG. Cross-competition studies using the three 125I-labeled ligands and allowing each to compete with the three radioinert ligands indicated the existence of two separate receptors. Desialylation of hepatic membranes differentially affected the binding of CBG and asialofetuin. Finally, whole cells bound CBG specifically, but internalized it to only a minimal extent ( 0.2, by t test) from that for unmodified CBG (see Fig. 2). We found previously (13), when examining the interaction of CBG with rat splenic membranes, that there was no alteration in free [125I]CBG and a small degree of metabolism of bound [125I]CBG after 4 h of incubation. In these experiments after 4 h there was no alteration in either the free or bound [125I]CBG (Fig. 3).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
HEPATIC CBG RECEPTORS
280
I 2 3
FIG. 1. Purity of CBG and asialo-CBG. [125I]CBG and [125I]asialoCBG were subjected to PAGE on 8% polyacrylamide with 2.67% crosslinking, followed by autoradiography. The gels were deliberately overloaded to demonstrate the absence of CBG in the preparation of asialoCBG and vice versa. Lane 1, CBG; lane 2, mixture of CBG and asialoCBG; lane 3, asialo-CBG.
To extend these observations, a series of experiments was conducted in which all the permutations of competition among CBG, asialo-CBG, asialofetuin, and their 125 I-labeled analogs were examined (Fig. 4). These experiments demonstrated that asialo-CBG and asialofetuin bound more or less equally to the asialoglycoprotein receptor, and that asialo-CBG appeared to bind less tightly to the CBG receptor than did unmodified CBG. However, Fig. 4A must be interperted in the light of the fact that the activity of the added asialo-CBG is substantially reduced because of its tight binding to the asialoglycoprotein receptor (Fig. 2C). It is known that the asialoglycoprotein receptor is itself a sialoglycoprotein, and that treatment of it with neuraminadase markedly diminishes its binding for asialoglycoproteins (20, 21). Further, evidence has been adduced that the placental CBG receptor is also a sialoglycoprotein (8). Treatment of hepatic membranes with neuraminadase greatly reduces the binding of both asialofetuin and asialo-CBG (Fig. 5). Such treatment also diminishes the binding of CBG to its receptor, but the effect is substantially smaller (Fig. 5). That there is a loss of binding supports the observation that the receptor is a glycoprotein (8). That the decrease in binding is relatively (compared to the asialoglycoprotein receptor) small indicates only a moderate degree of influence of
Endo • 1990 Vol 127 • No 1
the terminal sialyl residues on binding. The differential decrease in binding between asialofetuin and asialoCBG, on the one hand, and CBG, on the other, further supports the hypothesis that they are binding to two different entities. Asialo-CBG binds to the CBG receptor with about the same affinity as CBG, but its binding to the asialoglycoprotein receptor is much tighter than its binding to the CBG receptor (Figs. 2 and 4). Thus, the similarity in the loss of binding between asialo-CBG and asialofetuin, and not between asialo-CBG and CBG (Fig. 5), was to be expected. In addition to binding to hepatic membranes, CBG bound specifically to whole cells (Fig. 6) and was internalized by them to only a minor degree (Fig. 7). We had previously examined the interaction of [125I]CBG with MCF-7 cells (22) using the well established method of washing with acetic acid (23) to determine internalization. For the transferrin receptor, it had been shown that the determination of internalization could be shown equally well with either acetic acid or pronase (24). However, it was subsequently shown that acetic acid washing resulted in an overestimate of the internalization of CBG by FAO cells (6). We, therefore, reexamined the question. Residual 125I, after washing hepatic cells with the same acetic acid procedure previously used for MCF-7 cells (22), showed that 88 ± 1.8% (±SE; n = 6) was acetic acid resistant. Using Kuhn's method for removing noninternalized [125I]CBG with pronase (1 mg/ ml; for 15 min at 37 C) (6), we found 39 ± 3.2% (n = 9) to be apparently pronase resistant, in reasonable agreement with his results (6). However, using a protocol more like that of Karin and Mintz (4 C for 60 min) (24), it became clear that at least 60 min of exposure to pronase were necessary to achieve a constant residual concentration of cell-associated, i.e. internalized, [125I]CBG (Fig. 7). As previously reported (6), microscopic examination of the cells after treatment with pronase showed them to be intact. Further, treatment of hepatic cells that had been incubated with [125I] asialofetuin for 1 h, and then with pronase (2.5 mg/ml) for 60 min at 4 C, showed 79% and 83% (two experiments in triplicate) of the cellassociated 125I to be pronase resistant. This positive control demonstrated that the hepatocytes were able to endocytose bound ligands, and that digestion of protein from the surface of the cells, rather than cell lysis, brought about the release of [125I]CBG. Heparin, which efficiently removes noninternalized low density lipoprotein from fibroblasts (25), was superior to acetic acid as a reagent, but not as good as pronase (Fig. 7).
Discussion Not very long after the first description (20) of the asialoglycoprotein receptor, the interaction of human
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
HEPATIC CBG RECEPTORS
1.0
CM O
• -
\ \
U. \ Q
UJ
•
] \ -
LU LU
281
z: o 0.5 m
Q
z:
-
O CD
c 0.0 2
1
4 6 CBG BOUND (nM)
1
1
1
2 3 ASCBG BOUND (nM)
1
1
4
0.2 -
1
2 ASF BOUND (nM)
3
0
20 40 ASCBG BOUND (nM)
FIG. 2. Scatchard analysis of the binding of CBG, asialo-CBG (ASCBG), and asialofetuin to hepatic membranes. The indicated 125I-labeled ligands were incubated with hepatic membranes with varying amounts of the homologous radioinert ligands at 37 C. The results are shown as Scatchard plots corrected for nonspecific binding. A, [12SI]CBG (142,000 cpm; 0.70 nM) was incubated with hepatic membranes (1.1 mg membrane protein/ ml) for 4 h. The apparent Kd was 540 nM, and the binding capacity was 5.4 pmol/mg membrane protein. From this and two other experiments (n = 3), the Kd was 737 ± 140 nM (±SE), and the binding capacity was 5.11 ± 0.63 pmol/mg membrane protein. Use of 0.1 the concentration of [125I] CBG, with appropriately lower additions of radioinert CBG, failed to indicate a second, higher affinity set of binding sites (data not shown). B, [126I]Asialofetuin (143,000 cpm; 0.21 nM) was incubated with hepatic membranes (0.6 mg membrane protein/ml) for 1 h. Preliminary experiments (data not shown) indicated that binding had reached steady state at 45 min. The apparent Kd was 14.4 nM, and the binding capacity was 4.1 pmol/ mg membrane protein. From this and two other experiments (n = 3) the Kd was 18.3 ± 6.80 nM, and the binding capacity was 6.3 ± 1.02 pmol/mg membrane protein. C, [125I]Asialo-CBG (14,500 cpm; 0.14 nM) was incubated with hepatic membranes (0.43 mg membrane protein/ml) for 4 h. Analysis of the data yielded two Kd values (4.0 and 1,400 nM). D, [125I]Asialo-CBG (65,000 cpm; 0.62 nM) plus 1 fiM asialofetuin and increasing concentrations of radioinert asialo-CBG were incubated with hepatic membranes (0.43 mg membrane protein/ml) for 4 h. The apparent Kd was 514 nM, and the binding capacity was 17 pmol/mg membrane protein. From this and two other experiments (n = 3), the Kd was 534 ± 113 nM, and the binding capacity was 25 ± 11 pmol/mg membrane protein.
CBG and human asialo-CBG with the perfused rat liver was examined (11). It was shown that desialylation resulted in the prompt clearance of asialo-CBG from the perfusate, and that the hepatic uptake of cortisol bound to asialo-CBG was greater than its uptake when bound to native CBG. These results were recently confirmed (26). Subsequent investigations of the interaction of hu-
man [125I]CBG and [125I]asialo-CBG with rat hepatocytes, and rat hepatic membranes revealed tight binding of asialo-CBG, which was effectively competed by asialofetuin (12, 27). Low level binding of [125I]CBG was noted, but apparently was felt to be unimportant, especially when compared to the substantial binding of [125I]asialoCBG. In those experiments, binding was evaluated after
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
HEPATIC CBG RECEPTORS
282
2
3
4
1.0
-I20K
o m
A
•
00
A A
•
•
)
A
A A
D
O
o
-92.5K
Endo• 1990 Voll27«Nol
-
•
0.5
O
-69K
O
A
• O
0
•O
0.0
-46K
-30K
FIG. 3. Sodium dodecyl sulfate-PAGE of [126I]CBG incubated with hepatic membranes. [125I]CBG (150,000 cpm) was incubated with hepatic membranes as described in Fig. 2. After 4 h, membrane-bound [125I]CBG (lane 2) was solubilized (13) and submitted to sodium dodecyl sulfate-PAGE (10% polyacrylamide; 2.6% cross-linking) along with free [125I]CBG (that in the supernatant; lane 3), standard [125I]CBG (lane 1), and mol wt standards (lane 4). Autoradiography of the gel was performed by drying the gel and exposing it to Kodak XAR film at —80 C for 18 h.
1 h of incubation, appropriate for asialoglycoproteins but, as we subsequently showed, giving misleadingly low binding for unmodified CBG (2, 13). We have shown now that there are separate receptors for CBG and asialoglycoproteins on the cell membrane of the rat hepatocyte. The binding of asialofetuin to the asialoglycoprotein receptor is about 40-fold tighter than the binding of CBG to its receptor. Asialo-CBG binds to the the asialoglycoprotein receptor with approximately the same affinity as asialofetuin and to the CBG receptor with approximately the same affinity as CBG. Thus, removal of sialic acid from CBG alters neither its ability to bind steroids (28), nor its ability to interact with its receptor. The relationship among the Kd values for the two receptors with the two forms of CBG is such that if asialo-CBG existed in the circulation, it would bind preferentially to the asialoglycoprotein receptor. This taken together with the fact that cortisol and asialo-CBG comigrate into hepatocytes (26, 29) makes it appear that asialo-CBG might serve to bring cortisol into the hepatic cell. There is, however, a caveat. The function of the asialoglycoprotein receptor in vivo is unclear (30). There is a lack of evidence that asialoglycoproteins escape from cells into the circulation. Specifically, there is no evi-
CM
10
9
8
7
COMPETITOR ( - l o g M) FIG. 4. Specificity of the CBG and asialoglycoprotein receptors. The indicated 125I-labeled ligands were incubated with hepatic membranes and varying amounts of the designated heterologous or homologous radioinert ligands at 37 C. Bo, Specific binding of 126I-labeled ligand in the absence of competitor; B, specific binding of 126I-labeled ligand in the presence of competitor. O, CBG; D, asialo-CBG (ASCBG); A, asialofetuin. A, [125I]CBG (127,000 cpm; 0.70 nM) was incubated with 1.1 mg/ml hepatic membranes. B, [125I]Asialo-CBG (155,000 cpm; 0.16 nM) was incubated with 0.60 mg/ml hepatic membranes. C, [126I] Asialofetuin (143,000 cpm; 0.21 nM) was incubated with 0.60 mg/ml hepatic membranes.
dence that asialo-CBG exists in plasma. Thus, the observations on cortisol transport by asialo-CBG, albeit interesting, seem to be without physiological significance. It is theoretically possible that CBG could be bound to yet another hepatic membrane receptor. CBG has extensive homologies with the serine protease inhibitors (SERPINS), ax-antiproteinase and arantichymotrypsin (31). It appears to have been derived from the same ancestral gene as these proteins (31). These homologies may have some biological significance. Human CBG has been shown to interact with elastase (a typical serine
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
283
HEPATIC CBG RECEPTORS CONTROL
100
CZI NEURAMINIDASE
^
100
75
i
a
50
ih
25
0 CBG
ASF
ASCBG
30
FIG. 5. The effect of neuraminadase treatment of membranes on binding. Hepatic membranes (4.5 mg protein) in 25 mM tris-HCl, pH 7.4, containing 10 mM CaCl2 and 0.2% BSA were incubated with 100 mU affinity-purified neuraminadase (type X) at 37 C for 30 min. The suspensions were centrifuged at 20,000 x g for 10 min at 4 C. The pellets were resuspended in the same buffer without neuraminadase and centrifuged as before. The final pellet was suspended in 1 ml of the same buffer and assayed for binding in triplicate. Incubations with [125I]asialo-CBG (ASCBG) and [125I]asialfetuin (ASF) took place for 1 h at 37 C, whereas [12SI]CBG was incubated for 4 h. The concentrations of the ligands approximated those in Fig. 2. The data are the mean ± range from two separate experiments, each in triplicate.
4 TIME (h)
6
8
FIG. 6. Binding of CBG to hepatocytes. Hepatocytes were prepared, and binding experiments were conducted as described in Materials and Methods. #, Total binding; A, nonspecific binding; O, specific binding.
protease), which results in the release of a 4K peptide from CBG's carboxy-terminus (32). Further, there is a hepatic membrane receptor for SERPIN-protease complexes (33). However, SERPINS alone do not bind to this receptor; only the complexes do (33). Thus, if CBG is preincubated with elastase, as has been done for other SERPIN-protease pairs, CBG might be shown to bind to the SERPIN-protease receptor. This could not have occurred in these experiments. After 4 h of incubation with hepatic membranes, CBG was unchanged; there was no evidence of a decrease in the mol wt of CBG. What of the interaction of CBG with the CBG receptor? In contrast to the results in this communication, our initial observations indicated that essentially all of
60
90
TIME (min)
FlG. 7. Internalization of CBG by hepatocytes. Hepatocytes were prepared and incubated with [125I]CBG (37 C) in petri dishes for 4 h and centrifuged, as described in the text, for the binding of CBG to hepatocytes in Materials and Methods. The pelleted cells were resuspended in 2.5 vol ice-cold Waymouth 752/1 medium containing 0.2% BSA. One-half-milliliter aliquots were distributed into polystyrene tubes and centrifuged as described above. After discarding the supernatant, the cells were divided into three groups. Controls (O) were suspended in 0.2 ml ice-cold 25 mM Tris-HCl, pH 7.4, containing 0.9% NaCl (TrisNaCl buffer); for pronase treatment (•) they were suspended in 0.2 ml Tris-NaCl buffer containing 2.5 mg pronase/ml; for heparin treatment (A) they were suspended in 0.2 ml 50 mg/ml heparin Tris-NaCl buffer. All subsequent procedures were performed at 4 C. At the indicated times, cells were centrifuged, and the cell-associated radioactivity was determined. The results are expressed as the percentage of the specifically bound [126I]CBG after washing with buffer at time zero. The figure represents two experiments, each performed in triplicate. The specifically bound [125I]CBG values at time zero were 1600 and 1400 cpm/106 cells. Changing the pronase concentration to 0.5 or 1.0 mg/ml (data not shown) yielded identical results.
the specifically bound CBG was internalized (22). This interpretation was based on the resistance of bound [125I] CBG to washing with acetic acid. Kuhn (6) subsequently confirmed, in the hepatic cell line FAO, that bound [125I] CBG was resistant to acetic acid washing. However, he also noted that about 50% was removed by pronase treatment (1 mg/ml) at 37 C for 15 min. We observed the same result under those conditions, but found that that protocol overestimated internalization. That so little internalization of CBG occurs simplifies the potential functions of the CBG-receptor system. Although a role for CBG in the movement of bound steroids into cells has been postulated (2, 6,9, 34,35), the data herein make it unlikely. Plasma proteins that function to bring small molecules into cells, e.g. low density lipoproteins and transferrin, do so by being endocytosed together with their cognate ligands (36, 37). In view of the small fractional internalization of CBG, such a mechanism probably would not have quantitative significance in vivo. Thus, the major function of CBG, in addition to regulating the concentration of free steroids in plasma, appears to be its role in the activation of adenylate cyclase activity (22).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
HEPATIC CBG RECEPTORS
284
Acknowledgment We thank Dr. Daniel J. Hryb for his helpful discussions.
References 1. Hryb DJ, Khan MS, Rosner W 1985 Testosterone-estradiol-binding globulin binds to human prostatic cell membranes. Biochem Biophys Res Commun 128:432 2. Hryb DJ, Khan MS, Romas NA, Rosner W 1986 Specific binding of human corticosteroid-binding globulin to cell membranes. Proc Natl Acad Sci USA 83:3253 3. Rosner W, Khan MS, Romas NA, Hryb DJ 1986 Interactions of plasma steroid-binding proteins with cell membranes. In: Forest MG, Pugeat M (eds) Binding Proteins of Steroid Hormones. Libbey Eurotext, London and Paris, p 567 4. Rosner W, Hryb DJ, Khan MS, Singer CJ, Nakhla AM 1988 Are corticosteroid-binding globulin and sex hormone-binding globulin hormones. Ann NY Acad Sci 538:137 5. Hsu BR-S, Siiteri PK, Kuhn RW 1986 Interactions between corticosteroid-binding globulin (CBG) and target tissues. In: Forest MG, Pugeat M (eds) Binding Proteins of Steroid Hormones. Libbey Eurotext, London and Paris, p 577 6. Kuhn RW 1988 Corticosteroid-binding globulin interactions with target cells and plasma membranes. Ann NY Acad Sci 538:146 7. Strel'chyonok OA, Awakumov GV, Survilo LI 1984 A recognition system for sex-hormone-binding protein-estradiol complex in human decidual endometrium plasma membranes. Biochim Biophys Acta 802:459 8. Awakumov GV, Krupenko SA, Strel'chyonok OA 1989 Study of the transcortin binding to human endometrium plasma membrane. Biochim Biophys Acta 984:143 9. Strel'chyonok OA, Awakumov GV 1983 Evidence for the presence of specific binding sites for transcortin in human liver plasma membranes. Biochim Biophys Acta 755:514 10. Ashwell G, Harford J 1982 Carbohydrate-specific receptors of the liver. Annu Rev Biochem 51:531 11. Van Baelen H, Mannaerts G 1974 The role of sialic acid in determining the survival of transcortin in the circulation. Arch Biochem Biophys 163:53 12. Hossner KL, Billiar RB 1979 Plasma clearance and liver metabolism of native and asialo-human transcortin in the rat. Biochim Biophys Acta 585:543 13. Singer CJ, Khan MS, Rosner W 1988 Characteristics of the binding of corticosteroid binding globulin to rat cell membranes. Endocrinology 122:89 14. Rosner W, Hochberg R 1972 Corticosteroid-binding globulin in the rat: isolation and studies of its influence on cortisol action in vivo. Endocrinology 91:626 15. Rosner W, Polimeni S, Khan MS 1983 A radioimmunoassay for human corticosteroid-binding globulin. Clin Chem 29:1389 16. Munson PT, Rodbard D 1980 LIGAND: a versatile computerized approach for ligand-binding systems. Anal Biochem 107:220 17. Bradford MM 1976 A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248 18. Warren L 1959 The thiobarbituric acid assay of sialic acids. J Biol
Endo • 1990 Vol 127 • No 1
Chem 234:1971 19. Scatchard G 1949 The attractions of proteins for small molecules and ions. Ann NY Acad Sci 51:660 20. Pricer WE, Ashwell G 1971 The binding of desialyated glycoproteins by plasma membranes of rat liver. J Biol Chem 246:4825 21. Weiss P, Ashwell G 1989 Ligand-induced modulation of the hepatic receptor for asialoglycoproteins. J Biol Chem 264:11572 22. Nakhla AM, Khan MS, Rosner W 1988 Induction of adenylate cyclase in a mammary carcinoma cell line by human corticosteroidbinding globulin. Biochem Biophys Res Commun 153:1012 23. Haigler HT, Maxfield FR, Willingham MC, Pastan 11980 Dansylcadaverine inhibits internalization of 125I-epidermal growth factor in BALB 3T3 cells. J Biol Chem 255:1239 24. Karin M, Mintz B 1980 Receptor-mediated endocytosis of transferrin in developmentally totipotent mouse teratocarcinoma cell lines. J Biol Chem 256:3245 25. Goldstein JL, Basu SK, Brunschede GY, Brown JL 1976 Release of low density lipoprotein from its cell surface receptor by sulfated glycosaminoglycans. Cell 7:85 26. Karaskova H, Bezouska K, Starka L, Hampl R, Pikulev AT, Sholukh MV, Taborsky O 1986 Transport and intracellular localization of cortisol-asialotranscortin complexes in rat liver. J Steroid Biochem 24:725 27. Hossner KL, Billiar RB 1981 Metabolism of asialotranscortin by hepatocytes and cortisol binding to hepatocytes. Endocr Res Commun 8:111 28. Burton RM, Westphal U 1972 Steroid hormone-binding proteins in blood plasma. Metabolism 21:253 29. Lindenbaum M, Chatterton RT 1981 Interaction of steroids with dexamethasone-binding receptor and corticosteroid-binding globulin in the mammary gland of the mouse in relation to lactation. Endocrinology 109:363 30. Steer CJ, Weiss P, Huber B, Wirth PJ, Thorgeirsson SS, Ashwell G 1987 Ligand-induced modulation of the hepatic receptor for asialogycoproteins in the human hepatoblastoma cell line, Hep G2. J Biol Chem 262:17524 31. Underhill DA, Hammond GL 1989 Organization of the human corticosteroid binding globulin gene and analysis of its 5'-flanking region. Mol Endocrinol 3:1448 32. Pemberton PA, Stein PE, Pepys MB, Potter JM, Carrell RW 1988 Hormone binding globulins undergo serpin conformational change in inflammation. Nature 336:257 33. Fuchs HE, Michalopoulos GK, Pizzo SV 1984 Hepatocyte uptake of aj-proteinase inhibitor-trypsin complexes in vitro: evidence for a shared uptake mechanism for proteinase complexes of ai-proteinase inhibitor and antithrombin III. J Cell Biochem 25:231 34. Siiteri PK, Murai JT, Hammond GL, Nisker JA, Raymoure WJ, Kuhn RW 1982 The serum transport of steroid hormones. Recent Prog Horm Res 38:457 35. Kuhn RW, Green AL, Raymoure WJ, Siiteri PK 1986 Immunocytochemical localization of corticosteroid-binding globulin in rat tissues. J Endocrinol 108:31 36. Brown MS, Anderson RGW, Goldstein JL 1983 Recycling receptors: the round-trip itinerary of migrant membrane proteins. Cell 43:374 37. Cuatrecasas P 1986 Hormone receptors, membrane phospholipids, and protein kinases. Harv Lect 80:89
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
Vol. 127, No. 1 Printed in U.S.A.
The Rat Hepatic Corticosteroid-Binding Globulin Receptor: Distinction from the Asialoglycoprotein Receptor* UTPALENDU S. MAITRA, M. SAEED KHAN, XIN HUA ZHANG, AND WILLIAM ROSNER Department of Medicine, St. Luke's/Roosevelt Hospital Center, Columbia University College of Physicians and Surgeons, New York, New York 10019
ABSTRACT. This investigation was undertaken to ascertain whether rat liver cells contained a receptor for corticosteoidbinding globulin (CBG) that could be differentiated clearly from the asialoglycoprotein receptor. To do this, [125I]CBG, [125I] asialo-CBG, and [125I]asialofetuin were used as probes to differentiate the binding activities of the two receptors. On hepatic membranes, CBG bound to a single set of sites with a Kd of 0.74 ftM, asialofetuin bound to a single set of sites with a Kd of 0.018 jiM, asialo-CBG bound to two sets of sites with Kd values of 0.004 and 1.4 /*M, and in the presence of 1 juM asialofetuin,
asialo-CBG bound to a single set of sites with a Kd of 0.53 /xM, not different (P > 0.2) from the Kd of CBG. Cross-competition studies using the three 125I-labeled ligands and allowing each to compete with the three radioinert ligands indicated the existence of two separate receptors. Desialylation of hepatic membranes differentially affected the binding of CBG and asialofetuin. Finally, whole cells bound CBG specifically, but internalized it to only a minimal extent ( 0.2, by t test) from that for unmodified CBG (see Fig. 2). We found previously (13), when examining the interaction of CBG with rat splenic membranes, that there was no alteration in free [125I]CBG and a small degree of metabolism of bound [125I]CBG after 4 h of incubation. In these experiments after 4 h there was no alteration in either the free or bound [125I]CBG (Fig. 3).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
HEPATIC CBG RECEPTORS
280
I 2 3
FIG. 1. Purity of CBG and asialo-CBG. [125I]CBG and [125I]asialoCBG were subjected to PAGE on 8% polyacrylamide with 2.67% crosslinking, followed by autoradiography. The gels were deliberately overloaded to demonstrate the absence of CBG in the preparation of asialoCBG and vice versa. Lane 1, CBG; lane 2, mixture of CBG and asialoCBG; lane 3, asialo-CBG.
To extend these observations, a series of experiments was conducted in which all the permutations of competition among CBG, asialo-CBG, asialofetuin, and their 125 I-labeled analogs were examined (Fig. 4). These experiments demonstrated that asialo-CBG and asialofetuin bound more or less equally to the asialoglycoprotein receptor, and that asialo-CBG appeared to bind less tightly to the CBG receptor than did unmodified CBG. However, Fig. 4A must be interperted in the light of the fact that the activity of the added asialo-CBG is substantially reduced because of its tight binding to the asialoglycoprotein receptor (Fig. 2C). It is known that the asialoglycoprotein receptor is itself a sialoglycoprotein, and that treatment of it with neuraminadase markedly diminishes its binding for asialoglycoproteins (20, 21). Further, evidence has been adduced that the placental CBG receptor is also a sialoglycoprotein (8). Treatment of hepatic membranes with neuraminadase greatly reduces the binding of both asialofetuin and asialo-CBG (Fig. 5). Such treatment also diminishes the binding of CBG to its receptor, but the effect is substantially smaller (Fig. 5). That there is a loss of binding supports the observation that the receptor is a glycoprotein (8). That the decrease in binding is relatively (compared to the asialoglycoprotein receptor) small indicates only a moderate degree of influence of
Endo • 1990 Vol 127 • No 1
the terminal sialyl residues on binding. The differential decrease in binding between asialofetuin and asialoCBG, on the one hand, and CBG, on the other, further supports the hypothesis that they are binding to two different entities. Asialo-CBG binds to the CBG receptor with about the same affinity as CBG, but its binding to the asialoglycoprotein receptor is much tighter than its binding to the CBG receptor (Figs. 2 and 4). Thus, the similarity in the loss of binding between asialo-CBG and asialofetuin, and not between asialo-CBG and CBG (Fig. 5), was to be expected. In addition to binding to hepatic membranes, CBG bound specifically to whole cells (Fig. 6) and was internalized by them to only a minor degree (Fig. 7). We had previously examined the interaction of [125I]CBG with MCF-7 cells (22) using the well established method of washing with acetic acid (23) to determine internalization. For the transferrin receptor, it had been shown that the determination of internalization could be shown equally well with either acetic acid or pronase (24). However, it was subsequently shown that acetic acid washing resulted in an overestimate of the internalization of CBG by FAO cells (6). We, therefore, reexamined the question. Residual 125I, after washing hepatic cells with the same acetic acid procedure previously used for MCF-7 cells (22), showed that 88 ± 1.8% (±SE; n = 6) was acetic acid resistant. Using Kuhn's method for removing noninternalized [125I]CBG with pronase (1 mg/ ml; for 15 min at 37 C) (6), we found 39 ± 3.2% (n = 9) to be apparently pronase resistant, in reasonable agreement with his results (6). However, using a protocol more like that of Karin and Mintz (4 C for 60 min) (24), it became clear that at least 60 min of exposure to pronase were necessary to achieve a constant residual concentration of cell-associated, i.e. internalized, [125I]CBG (Fig. 7). As previously reported (6), microscopic examination of the cells after treatment with pronase showed them to be intact. Further, treatment of hepatic cells that had been incubated with [125I] asialofetuin for 1 h, and then with pronase (2.5 mg/ml) for 60 min at 4 C, showed 79% and 83% (two experiments in triplicate) of the cellassociated 125I to be pronase resistant. This positive control demonstrated that the hepatocytes were able to endocytose bound ligands, and that digestion of protein from the surface of the cells, rather than cell lysis, brought about the release of [125I]CBG. Heparin, which efficiently removes noninternalized low density lipoprotein from fibroblasts (25), was superior to acetic acid as a reagent, but not as good as pronase (Fig. 7).
Discussion Not very long after the first description (20) of the asialoglycoprotein receptor, the interaction of human
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
HEPATIC CBG RECEPTORS
1.0
CM O
• -
\ \
U. \ Q
UJ
•
] \ -
LU LU
281
z: o 0.5 m
Q
z:
-
O CD
c 0.0 2
1
4 6 CBG BOUND (nM)
1
1
1
2 3 ASCBG BOUND (nM)
1
1
4
0.2 -
1
2 ASF BOUND (nM)
3
0
20 40 ASCBG BOUND (nM)
FIG. 2. Scatchard analysis of the binding of CBG, asialo-CBG (ASCBG), and asialofetuin to hepatic membranes. The indicated 125I-labeled ligands were incubated with hepatic membranes with varying amounts of the homologous radioinert ligands at 37 C. The results are shown as Scatchard plots corrected for nonspecific binding. A, [12SI]CBG (142,000 cpm; 0.70 nM) was incubated with hepatic membranes (1.1 mg membrane protein/ ml) for 4 h. The apparent Kd was 540 nM, and the binding capacity was 5.4 pmol/mg membrane protein. From this and two other experiments (n = 3), the Kd was 737 ± 140 nM (±SE), and the binding capacity was 5.11 ± 0.63 pmol/mg membrane protein. Use of 0.1 the concentration of [125I] CBG, with appropriately lower additions of radioinert CBG, failed to indicate a second, higher affinity set of binding sites (data not shown). B, [126I]Asialofetuin (143,000 cpm; 0.21 nM) was incubated with hepatic membranes (0.6 mg membrane protein/ml) for 1 h. Preliminary experiments (data not shown) indicated that binding had reached steady state at 45 min. The apparent Kd was 14.4 nM, and the binding capacity was 4.1 pmol/ mg membrane protein. From this and two other experiments (n = 3) the Kd was 18.3 ± 6.80 nM, and the binding capacity was 6.3 ± 1.02 pmol/mg membrane protein. C, [125I]Asialo-CBG (14,500 cpm; 0.14 nM) was incubated with hepatic membranes (0.43 mg membrane protein/ml) for 4 h. Analysis of the data yielded two Kd values (4.0 and 1,400 nM). D, [125I]Asialo-CBG (65,000 cpm; 0.62 nM) plus 1 fiM asialofetuin and increasing concentrations of radioinert asialo-CBG were incubated with hepatic membranes (0.43 mg membrane protein/ml) for 4 h. The apparent Kd was 514 nM, and the binding capacity was 17 pmol/mg membrane protein. From this and two other experiments (n = 3), the Kd was 534 ± 113 nM, and the binding capacity was 25 ± 11 pmol/mg membrane protein.
CBG and human asialo-CBG with the perfused rat liver was examined (11). It was shown that desialylation resulted in the prompt clearance of asialo-CBG from the perfusate, and that the hepatic uptake of cortisol bound to asialo-CBG was greater than its uptake when bound to native CBG. These results were recently confirmed (26). Subsequent investigations of the interaction of hu-
man [125I]CBG and [125I]asialo-CBG with rat hepatocytes, and rat hepatic membranes revealed tight binding of asialo-CBG, which was effectively competed by asialofetuin (12, 27). Low level binding of [125I]CBG was noted, but apparently was felt to be unimportant, especially when compared to the substantial binding of [125I]asialoCBG. In those experiments, binding was evaluated after
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
HEPATIC CBG RECEPTORS
282
2
3
4
1.0
-I20K
o m
A
•
00
A A
•
•
)
A
A A
D
O
o
-92.5K
Endo• 1990 Voll27«Nol
-
•
0.5
O
-69K
O
A
• O
0
•O
0.0
-46K
-30K
FIG. 3. Sodium dodecyl sulfate-PAGE of [126I]CBG incubated with hepatic membranes. [125I]CBG (150,000 cpm) was incubated with hepatic membranes as described in Fig. 2. After 4 h, membrane-bound [125I]CBG (lane 2) was solubilized (13) and submitted to sodium dodecyl sulfate-PAGE (10% polyacrylamide; 2.6% cross-linking) along with free [125I]CBG (that in the supernatant; lane 3), standard [125I]CBG (lane 1), and mol wt standards (lane 4). Autoradiography of the gel was performed by drying the gel and exposing it to Kodak XAR film at —80 C for 18 h.
1 h of incubation, appropriate for asialoglycoproteins but, as we subsequently showed, giving misleadingly low binding for unmodified CBG (2, 13). We have shown now that there are separate receptors for CBG and asialoglycoproteins on the cell membrane of the rat hepatocyte. The binding of asialofetuin to the asialoglycoprotein receptor is about 40-fold tighter than the binding of CBG to its receptor. Asialo-CBG binds to the the asialoglycoprotein receptor with approximately the same affinity as asialofetuin and to the CBG receptor with approximately the same affinity as CBG. Thus, removal of sialic acid from CBG alters neither its ability to bind steroids (28), nor its ability to interact with its receptor. The relationship among the Kd values for the two receptors with the two forms of CBG is such that if asialo-CBG existed in the circulation, it would bind preferentially to the asialoglycoprotein receptor. This taken together with the fact that cortisol and asialo-CBG comigrate into hepatocytes (26, 29) makes it appear that asialo-CBG might serve to bring cortisol into the hepatic cell. There is, however, a caveat. The function of the asialoglycoprotein receptor in vivo is unclear (30). There is a lack of evidence that asialoglycoproteins escape from cells into the circulation. Specifically, there is no evi-
CM
10
9
8
7
COMPETITOR ( - l o g M) FIG. 4. Specificity of the CBG and asialoglycoprotein receptors. The indicated 125I-labeled ligands were incubated with hepatic membranes and varying amounts of the designated heterologous or homologous radioinert ligands at 37 C. Bo, Specific binding of 126I-labeled ligand in the absence of competitor; B, specific binding of 126I-labeled ligand in the presence of competitor. O, CBG; D, asialo-CBG (ASCBG); A, asialofetuin. A, [125I]CBG (127,000 cpm; 0.70 nM) was incubated with 1.1 mg/ml hepatic membranes. B, [125I]Asialo-CBG (155,000 cpm; 0.16 nM) was incubated with 0.60 mg/ml hepatic membranes. C, [126I] Asialofetuin (143,000 cpm; 0.21 nM) was incubated with 0.60 mg/ml hepatic membranes.
dence that asialo-CBG exists in plasma. Thus, the observations on cortisol transport by asialo-CBG, albeit interesting, seem to be without physiological significance. It is theoretically possible that CBG could be bound to yet another hepatic membrane receptor. CBG has extensive homologies with the serine protease inhibitors (SERPINS), ax-antiproteinase and arantichymotrypsin (31). It appears to have been derived from the same ancestral gene as these proteins (31). These homologies may have some biological significance. Human CBG has been shown to interact with elastase (a typical serine
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
283
HEPATIC CBG RECEPTORS CONTROL
100
CZI NEURAMINIDASE
^
100
75
i
a
50
ih
25
0 CBG
ASF
ASCBG
30
FIG. 5. The effect of neuraminadase treatment of membranes on binding. Hepatic membranes (4.5 mg protein) in 25 mM tris-HCl, pH 7.4, containing 10 mM CaCl2 and 0.2% BSA were incubated with 100 mU affinity-purified neuraminadase (type X) at 37 C for 30 min. The suspensions were centrifuged at 20,000 x g for 10 min at 4 C. The pellets were resuspended in the same buffer without neuraminadase and centrifuged as before. The final pellet was suspended in 1 ml of the same buffer and assayed for binding in triplicate. Incubations with [125I]asialo-CBG (ASCBG) and [125I]asialfetuin (ASF) took place for 1 h at 37 C, whereas [12SI]CBG was incubated for 4 h. The concentrations of the ligands approximated those in Fig. 2. The data are the mean ± range from two separate experiments, each in triplicate.
4 TIME (h)
6
8
FIG. 6. Binding of CBG to hepatocytes. Hepatocytes were prepared, and binding experiments were conducted as described in Materials and Methods. #, Total binding; A, nonspecific binding; O, specific binding.
protease), which results in the release of a 4K peptide from CBG's carboxy-terminus (32). Further, there is a hepatic membrane receptor for SERPIN-protease complexes (33). However, SERPINS alone do not bind to this receptor; only the complexes do (33). Thus, if CBG is preincubated with elastase, as has been done for other SERPIN-protease pairs, CBG might be shown to bind to the SERPIN-protease receptor. This could not have occurred in these experiments. After 4 h of incubation with hepatic membranes, CBG was unchanged; there was no evidence of a decrease in the mol wt of CBG. What of the interaction of CBG with the CBG receptor? In contrast to the results in this communication, our initial observations indicated that essentially all of
60
90
TIME (min)
FlG. 7. Internalization of CBG by hepatocytes. Hepatocytes were prepared and incubated with [125I]CBG (37 C) in petri dishes for 4 h and centrifuged, as described in the text, for the binding of CBG to hepatocytes in Materials and Methods. The pelleted cells were resuspended in 2.5 vol ice-cold Waymouth 752/1 medium containing 0.2% BSA. One-half-milliliter aliquots were distributed into polystyrene tubes and centrifuged as described above. After discarding the supernatant, the cells were divided into three groups. Controls (O) were suspended in 0.2 ml ice-cold 25 mM Tris-HCl, pH 7.4, containing 0.9% NaCl (TrisNaCl buffer); for pronase treatment (•) they were suspended in 0.2 ml Tris-NaCl buffer containing 2.5 mg pronase/ml; for heparin treatment (A) they were suspended in 0.2 ml 50 mg/ml heparin Tris-NaCl buffer. All subsequent procedures were performed at 4 C. At the indicated times, cells were centrifuged, and the cell-associated radioactivity was determined. The results are expressed as the percentage of the specifically bound [126I]CBG after washing with buffer at time zero. The figure represents two experiments, each performed in triplicate. The specifically bound [125I]CBG values at time zero were 1600 and 1400 cpm/106 cells. Changing the pronase concentration to 0.5 or 1.0 mg/ml (data not shown) yielded identical results.
the specifically bound CBG was internalized (22). This interpretation was based on the resistance of bound [125I] CBG to washing with acetic acid. Kuhn (6) subsequently confirmed, in the hepatic cell line FAO, that bound [125I] CBG was resistant to acetic acid washing. However, he also noted that about 50% was removed by pronase treatment (1 mg/ml) at 37 C for 15 min. We observed the same result under those conditions, but found that that protocol overestimated internalization. That so little internalization of CBG occurs simplifies the potential functions of the CBG-receptor system. Although a role for CBG in the movement of bound steroids into cells has been postulated (2, 6,9, 34,35), the data herein make it unlikely. Plasma proteins that function to bring small molecules into cells, e.g. low density lipoproteins and transferrin, do so by being endocytosed together with their cognate ligands (36, 37). In view of the small fractional internalization of CBG, such a mechanism probably would not have quantitative significance in vivo. Thus, the major function of CBG, in addition to regulating the concentration of free steroids in plasma, appears to be its role in the activation of adenylate cyclase activity (22).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
HEPATIC CBG RECEPTORS
284
Acknowledgment We thank Dr. Daniel J. Hryb for his helpful discussions.
References 1. Hryb DJ, Khan MS, Rosner W 1985 Testosterone-estradiol-binding globulin binds to human prostatic cell membranes. Biochem Biophys Res Commun 128:432 2. Hryb DJ, Khan MS, Romas NA, Rosner W 1986 Specific binding of human corticosteroid-binding globulin to cell membranes. Proc Natl Acad Sci USA 83:3253 3. Rosner W, Khan MS, Romas NA, Hryb DJ 1986 Interactions of plasma steroid-binding proteins with cell membranes. In: Forest MG, Pugeat M (eds) Binding Proteins of Steroid Hormones. Libbey Eurotext, London and Paris, p 567 4. Rosner W, Hryb DJ, Khan MS, Singer CJ, Nakhla AM 1988 Are corticosteroid-binding globulin and sex hormone-binding globulin hormones. Ann NY Acad Sci 538:137 5. Hsu BR-S, Siiteri PK, Kuhn RW 1986 Interactions between corticosteroid-binding globulin (CBG) and target tissues. In: Forest MG, Pugeat M (eds) Binding Proteins of Steroid Hormones. Libbey Eurotext, London and Paris, p 577 6. Kuhn RW 1988 Corticosteroid-binding globulin interactions with target cells and plasma membranes. Ann NY Acad Sci 538:146 7. Strel'chyonok OA, Awakumov GV, Survilo LI 1984 A recognition system for sex-hormone-binding protein-estradiol complex in human decidual endometrium plasma membranes. Biochim Biophys Acta 802:459 8. Awakumov GV, Krupenko SA, Strel'chyonok OA 1989 Study of the transcortin binding to human endometrium plasma membrane. Biochim Biophys Acta 984:143 9. Strel'chyonok OA, Awakumov GV 1983 Evidence for the presence of specific binding sites for transcortin in human liver plasma membranes. Biochim Biophys Acta 755:514 10. Ashwell G, Harford J 1982 Carbohydrate-specific receptors of the liver. Annu Rev Biochem 51:531 11. Van Baelen H, Mannaerts G 1974 The role of sialic acid in determining the survival of transcortin in the circulation. Arch Biochem Biophys 163:53 12. Hossner KL, Billiar RB 1979 Plasma clearance and liver metabolism of native and asialo-human transcortin in the rat. Biochim Biophys Acta 585:543 13. Singer CJ, Khan MS, Rosner W 1988 Characteristics of the binding of corticosteroid binding globulin to rat cell membranes. Endocrinology 122:89 14. Rosner W, Hochberg R 1972 Corticosteroid-binding globulin in the rat: isolation and studies of its influence on cortisol action in vivo. Endocrinology 91:626 15. Rosner W, Polimeni S, Khan MS 1983 A radioimmunoassay for human corticosteroid-binding globulin. Clin Chem 29:1389 16. Munson PT, Rodbard D 1980 LIGAND: a versatile computerized approach for ligand-binding systems. Anal Biochem 107:220 17. Bradford MM 1976 A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248 18. Warren L 1959 The thiobarbituric acid assay of sialic acids. J Biol
Endo • 1990 Vol 127 • No 1
Chem 234:1971 19. Scatchard G 1949 The attractions of proteins for small molecules and ions. Ann NY Acad Sci 51:660 20. Pricer WE, Ashwell G 1971 The binding of desialyated glycoproteins by plasma membranes of rat liver. J Biol Chem 246:4825 21. Weiss P, Ashwell G 1989 Ligand-induced modulation of the hepatic receptor for asialoglycoproteins. J Biol Chem 264:11572 22. Nakhla AM, Khan MS, Rosner W 1988 Induction of adenylate cyclase in a mammary carcinoma cell line by human corticosteroidbinding globulin. Biochem Biophys Res Commun 153:1012 23. Haigler HT, Maxfield FR, Willingham MC, Pastan 11980 Dansylcadaverine inhibits internalization of 125I-epidermal growth factor in BALB 3T3 cells. J Biol Chem 255:1239 24. Karin M, Mintz B 1980 Receptor-mediated endocytosis of transferrin in developmentally totipotent mouse teratocarcinoma cell lines. J Biol Chem 256:3245 25. Goldstein JL, Basu SK, Brunschede GY, Brown JL 1976 Release of low density lipoprotein from its cell surface receptor by sulfated glycosaminoglycans. Cell 7:85 26. Karaskova H, Bezouska K, Starka L, Hampl R, Pikulev AT, Sholukh MV, Taborsky O 1986 Transport and intracellular localization of cortisol-asialotranscortin complexes in rat liver. J Steroid Biochem 24:725 27. Hossner KL, Billiar RB 1981 Metabolism of asialotranscortin by hepatocytes and cortisol binding to hepatocytes. Endocr Res Commun 8:111 28. Burton RM, Westphal U 1972 Steroid hormone-binding proteins in blood plasma. Metabolism 21:253 29. Lindenbaum M, Chatterton RT 1981 Interaction of steroids with dexamethasone-binding receptor and corticosteroid-binding globulin in the mammary gland of the mouse in relation to lactation. Endocrinology 109:363 30. Steer CJ, Weiss P, Huber B, Wirth PJ, Thorgeirsson SS, Ashwell G 1987 Ligand-induced modulation of the hepatic receptor for asialogycoproteins in the human hepatoblastoma cell line, Hep G2. J Biol Chem 262:17524 31. Underhill DA, Hammond GL 1989 Organization of the human corticosteroid binding globulin gene and analysis of its 5'-flanking region. Mol Endocrinol 3:1448 32. Pemberton PA, Stein PE, Pepys MB, Potter JM, Carrell RW 1988 Hormone binding globulins undergo serpin conformational change in inflammation. Nature 336:257 33. Fuchs HE, Michalopoulos GK, Pizzo SV 1984 Hepatocyte uptake of aj-proteinase inhibitor-trypsin complexes in vitro: evidence for a shared uptake mechanism for proteinase complexes of ai-proteinase inhibitor and antithrombin III. J Cell Biochem 25:231 34. Siiteri PK, Murai JT, Hammond GL, Nisker JA, Raymoure WJ, Kuhn RW 1982 The serum transport of steroid hormones. Recent Prog Horm Res 38:457 35. Kuhn RW, Green AL, Raymoure WJ, Siiteri PK 1986 Immunocytochemical localization of corticosteroid-binding globulin in rat tissues. J Endocrinol 108:31 36. Brown MS, Anderson RGW, Goldstein JL 1983 Recycling receptors: the round-trip itinerary of migrant membrane proteins. Cell 43:374 37. Cuatrecasas P 1986 Hormone receptors, membrane phospholipids, and protein kinases. Harv Lect 80:89
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 November 2015. at 21:41 For personal use only. No other uses without permission. . All rights reserved.
