Receptor-mediated endocytosis and degradation of bovine growth hormone in rat liver

13

Molecular and Cellular Endocrinology, 59 (1988) 13-25 Elsevier Scientific Publishers Ireland, Ltd.

MCE 01901

Receptor-mediated

endocytosis and degradation of bovine growth hormone in rat liver

Bolette Husman Departments

of ’ Medical Nutrition,

132,Jan-Ake

Gustafsson

’

and GBran Andersson

2*3

’ Pathology and ’ Oral Pathology, Karolinska Institute, Huddinge University Hospital, S-141 86 Huddinge, Sweden (Received 25 January 1988; accepted 16 May 1988)

Key words: Receptor;

Endocytosis;

Growth hormone;

Prolactin; (Degradation);

(Sex difference);

(Liver); (Rat)

Summary Receptor-mediated endocytosis of radiolabelled bovine growth hormone (“‘1-bGH) via somatogenic receptors in the liver was studied following in vivo intraportal injection. At different times after injection, subcellular membrane fractions involved in binding (plasma membranes), endocytosis (endocytic vesicles) and degradation (lysosomes) of peptide hormones were isolated by sucrose density gradient centrifugation. These fractions were evaluated for the time-course accumulation of radiolabelled bGH and for the ‘251-bGH-receptor complexes. These uptake studies indicate that after initial presence of internalized in two successive endocytic compartplasma membrane association of lz51-bGH the ligand is transported ments prior to arrival in lysosomes. The molecular weight of the somatogenic binders of male and female rat livers involved in internalization of ‘251-bGH was determined to 95 000, 64000, 55 000, 43 000 and 35 000, assuming a 1: 1 binding of the hormone to the binder. These binders were seen in both endosomes and lysosomes, which suggests that growth hormone is transported to the lysosomes in a complex with its receptor. Binding and uptake of ‘251-bGH was also compared in male and female rat livers, and endocytosis of The specific uptake of ‘251-bGH was compared to that of radiolabelled ovine prolactin ( “‘1-oPrl). ‘25I-bGH appeared not to be sexually differentiated in contrast to that of ‘251-oPrl which showed a 35-fold higher uptake in female rat liver. A distinct 15000 Da Degradation of ‘251-bGH was studied under in vitro binding assay conditions. fragment was generated by plasma membrane, endosomal and lysosomal fractions. Based on protease inhibitor studies, a non-trypsin-like serine protease is suggested to be involved in the degradation of bGH. The 15 000 Da proteolytic fragment of GH can be affinity cross-linked to somatogenic binders of similar molecular weights as those involved in the binding of intact GH.

Address for correspondence: Bolette Husman, Department of Medical Nutrition, Karolinska Institute, Huddinge University Hospital F69, S-141 86 Huddinge, Sweden. Abbreviations: PMSF, phenylmethylsulfonyl fluoride; TLCK, L-1-chloro-3-(4tosylamido)-7-amino-2-heptanone hydrochloride; BSA, bovine serum albumin; DSS, disuccinimidyl suberate; SDS, sodium dodecylsulfate; SDS-PAGE, sodium dodecylsulfate-polyacrylamide gel electrophoresis; Tris, tris (hydroxymethyl)aminomethane; bGH, bovine growth hormone; oPr1, ovine prolactin; M,, molecular weight; TIU, trypsin inhibitor units. 0303-7207/88/$03.50

0 1988 Elsevier Scientific Publishers Ireland, Ltd.

14

Introduction It is reasonable to assume that binding of GH to a specific somatogenic receptor protein present on target cells constitutes a primary event in the biological action of GH. Such specific somatogenie binding sites have been demonstrated by ligand affinity cross-linking on human IM-9 lymphocytes (Hughes et al., 1983) rat adipocytes (Carter-Su et al., 1984; Gorin and Goodman, 1984) and rat hepatocytes (Yamada and Donner, 1984; Husman et al., 1988). In addition, somatogenic receptors are highly enriched in low-density intracellular organelles isolated from rat liver (Husman et al., 1985). The significance of the enrichment of GH receptors in low-density intracellular membranes is unclear but may result at least in part from a recruitment of cell-surface receptors to endocytic compartments. Internalization of membrane receptor ligands is known to proceed via specialized membrane domains on the plasma membrane, the coated pits, to several distinct endocytic intracellular compartments and finally to lysosomes where at least the ligand is degraded (Farquhar, 1983; Helenius et al., 1983; Asakawa et al., 1985). The biological significance of polypeptide hormone endocytosis and intracellular degradation is still unclear. It has been demonstrated, however, that degradation of internalized proteins may be initiated in endocytic compartments (Diment and Stahl, 1985) and that proteolytic degradation of receptor-bound proteins may be required for endosome-lysosome fusion (Opresko and Karpf, 1987). Proteolytic cleavage of growth hormone has in several cases been associated with potentiation of somatomedin-like effects (Singh et al., 1974; Lewis et al., 1977; Liberti and Miller, 1978; Liberti and Durham, 1983). The site of proteolysis action on GH is not known, but it may be of relevance that a plasma membrane-enriched fraction from rabbit liver has been shown to contain serine protease activity which cleaves in the large disulfide loop of hGH and rGH. In this case, hormone receptor interaction was found not to be required for cleavage but the cleaved GH bound more readily than intact GH to rabbit liver somatogenic receptors (Schepper et al., 1984). Growth hormone appears to be an important

mediator of several sexually differentiated functions in rat liver (for a review, see Gustafsson et al., 1983). Prominent among these are regulation of sex steroid metabolism (Mode et al., 1981) and maintenance of high levels of prolactin receptors in the female rat liver (Norstedt et al., 1981, 1984; Baxter et al., 1984). A noticeable difference between male and female rats in relation to GH action resides in the circadian rhythm of GH secretion from the anterior pituitary (Eden, 1979). The male rat exhibits regular surges of GH every 3-4 h with low levels in between the peaks whereas in the female rat the pattern is more regular with lower peak heights and higher levels between the peaks. By mimicking the female pattern of GH blood levels by infusing human or rat GH through osmotic minipumps to either intact or hypophysectomized male rats, an up-regulation of female-specific activities such as the cytochrome P-450 isozyme catalyzing the 15&hydroxylation of Sa-androstane-3a,l7/3-diol 3,17-disulphate (McGeoch et al., 1984, 1985) and prolactin receptors (Norstedt et al., 1981, 1984) is triggered. In contrast, continuous GH infusion to intact male rats results in down-regulation to female levels of the male-specific cytochrome P-450 isozyme catalyzing the 16a-hydroxylation of 4-androstene-3,17dione and testosterone (Morgan et al., 1985). Materials and methods Animals Male and female rats (8 weeks; male rats, 280-300 g; female rats, 180-200 g) of the Sprague-Dawley strain (Alab, Stockholm, Sweden) were kept under controlled conditions (light on, 06.00-20.00 h, temperature, 23 &- lo C). Animals were starved for 18 h prior to experimentation which occurred between 08.00 a.m. and 10.00 a.m. Experiments were performed under hexobarbital anaesthesia (Evipan-Natrium, Bayer, F.R.G.) where female rate received 200 mg/lOO g body weight and male rats received 300 mg/lOO g body weight. Hormones bGH (recombinant bGH) was supplied from American Cyanamide, Piscataway, NJ, U.S.A. oPr1 (oPrl-17) and rGH (rGH-B-8) were obtained from

15

the NIAMDD, NIH (Bethesda, MD, U.S.A.). All hormones were iodinated using the iodogen technique as previously described (Husman et al., 1985). Specific activity of the iodinated hormones varied between 30 and 55 $Zi/pg. Subcellular fractionation The residual microsomes (RM), intermediate (I) and endosome I (E I) fractions were isolated essentially according to the sequential procedure described by Andersson et al. (1978). In short, perfused livers were homogenized with 0.25 M sucrose to give a 20% (w/v) homogenate which was centrifuged at 10000 X g for 20 min and the resulting supernatant was collected (Mic I). The pellet was resuspended and further centrifuged at 10000 X g for 10 min (Mic II). Mic I and Mic II fractions were concentrated by centrifugation (105 000 x g, 90 min), resuspended to a final concentration of 1.2 M (d = 1.156) sucrose, and subjected to density gradient centrifugation for the isolation of endosome I (mean density d< 1.147) fractions by flotation. The residual microsomes represents the load portion of the gradient, whereas the intermediate fraction consists of membranes in the volume between the load portion and the endosome I and endosome (see below) band. In some experiments, a modified technique was used for the isolation of endosomes. The livers were homogenized (20%, w/v) and centrifuged at 10000 x g for 10 min and the resulting supernatant was collected. The pellet was resuspended and centrifuged at 10000 X g for 10 min. The two supernatants were pooled and concentrated by centrifugation (105 000 X g, 90 min). The resulting pellet was resuspended to a final concentration of 1.125 M sucrose (d = 1.147) and subjected to density gradient centrifugation with 1.10 M and 0.25 M sucrose on top of the sample. The band between 1.10 M and 0.25 M sucrose is referred to as the endosome (E) fraction. Total particulate fraction (T) and cytosol (C) were prepared by centrifugation of the homogenate (H) at 105000 x g for 90 min, resulting in a pellet (T) and a supernatant (C). Plasma membranes (P), mainly consisting of the bile canalicular domain but also the blood-sinusoidal domain, were prepared by the method of Song et al. (1969). In short, livers were

perfused with 0.001 M NaHCO,, pH 7.5, followed by homogenization with a Dounce homogenizer. The homogenate was filtered and centrifuged at 1500 x g for 10 min. The resulting pellet was suspended and adjusted with 2 M sucrose to a final density of 1.22 which was then subjected to sucrose density gradient centrifugation. One band (A), which corresponds to the fraction located in the interface between 42.9% (w/v) and 48.6% (w/v) of sucrose was collected and used. The enrichment of fluoride-stimulated adenylate cyclase was about 20-fold compared to the homogenate. The endosome II and lysosome fractions (E II, L) were isolated from a piece (3 g) of the same liver as was used for isolation of microsomes. The method of Wattiaux et al. (1978) using a discontinuous metrizamide density gradient which separates lysosomes from a light mitochondrial-lysosomal fraction by flotation was used. In this gradient, the E II fraction floats on top of 20% metrizamide (d = 1.11) and the L fraction is recovered in the interphase between 20% and 24.5% (d = 1.135) metrizamide. Internalization of radiolabelled ligand Radiolabelled hormones in the absence or presence of unlabelled hormone were delivered in amounts indicated in the figure and table legends to the portal blood flow by injection into mesenteric veins. A proportional increase in the specific uptake of lz51-bGH was seen at least up to an injected dose of 25. lo6 cpm/lOO g body weight (data not shown). Internalization was terminated after different times by perfusing the liver via the caval vein with ice-cold 0.25 M sucrose. Experiments were performed on groups of six animals: three rats received labelled hormone alone to determine total uptake, and three rats received the same dose of labelled hormone together with a lOOO-fold excess unlabelled hormone to determine non-specific uptake. Specific uptake was calculated by subtraction of non-specific uptake from total uptake. Degradation of ligand in vitro lz51-bGH (300000 cpm) was incubated with membrane fractions for 18 h at 20° C in a total volume of 1 ml 50 mM potassium phosphate

16

buffer, pH 7.4, containing 0.1% (w/v) BSA. Incubations were performed in the absence or in the presence of a lOOO-fold excess of unlabelled bGH. The incubations were terminated by adding solubilizing buffer and boiled for 2 min. Samples were then subjected to SDS-PAGE. Cross-linking of iodinated hormone to membranes Freshly prepared DSS in dimethyl sulfoxide was added to ‘251-bGH-labelled membranes to give a final concentration of 0.5 mM. The incubation proceeded for 15 min at 4’ C and was stopped by adding Tris-HCl buffer, pH 7.5, to a final concentration of 10 mM. Samples were then subjected to analysis by SDS-PAGE. Electrophore~is and autoradiography Cross-linked samples (110 ~1) were mixed with 0.6 vol. of SDS-solubilizing buffer (125 mM Tris/HCl, pH 6.8; 20% (v/v) glycerol; 4% (w/v) SDS; 0.05% (w/v) bromophenol blue and 10% (v/v) ~-mercapt~thanol) and boiled for 2 min. SDS-PAGE was performed on 1.5 mm thick 7.5% or 15% polyacrylamide slab gels according to the method of Laemmli (1970). Gels were dried and subjected to autoradiography using Fuji RX X-ray film and DuPont Cronex Lighting-Plus enhancing screens. The following “C-methylated molecular weight standards were used: myosin (200000), phosphorylase b (92 500), BSA (69000), ovalbumin (46000) carbonic anhydrase (30000), trypsin inhibitor (soy bean) (21500), lysozyme (14 300) and cytochrome c (12 500). Protein concentration was measured by a modified Lowry method (Bensadoun and Weinstein, 1976) using BSA as a standard. Data are shown as the mean f SD or mean + SEM as stated in the legends to the figures. Results Receptor-mediated uptake in vivo of ‘2sI-bovine CH and ‘2SI-ovine Prl in livers from adult male and fernare rats In order to address the question of possible sex differences in the receptor-mediated endocytosis of a pure somatogenic receptor ligand in rat liver, adult male and female rats were given intramesen-

teric-intraportal injections of ‘251-bGH in the absence or presence of a lOOO-fold excess of unlabelled bGH. Since the 8-week-old rats used in this experiment differed considerably in body weight between the sexes and thus correspondingly in blood volume, special care was taken to relate the dose level of tracer injected to body weight (2.4. lo6 cpm/lOO g body weight or approximately 25 ng/lOO g body weight). Also, since peptide hormones (and other blood constituents) are cleared from the circulation not only by receptor-mediated endocytosis in their target tissues but also by a high-capacity, non-specific internalization mechanism (or fluid-phase endocytosis), the specific uptake/association of lz51-bGH was calculated by subtracting the liver-associated radioactivity obtained after coinjection of tracer and lOOO-fold excess of unlabelled hormone (nonspecific uptake) with the value obtained after injection of tracer alone (total uptake). Non-specific uptake was approximately 25% of the total uptake under the conditions used in this experiment. Following in vivo labelling with ‘251-bGH for 10 min and mild perfusion of the liver to remove unbound tracer (Table l), no significant differences could be found in the amount of radioactivity associated with either male or female rat liver homogenates or in the dist~bution to particulate and soluble subcellular fractions. In both sexes, 90% of the homogenate radioactivity at this time point was associated with the particulate fraction. Furthermore, this finding correlates with our previous data showing no sex difference in the number of somatogenic receptors (Husman et al., 1985). We also injected ‘2sI-oPrl in the absence or presence of unlabelled hormone as described above in order to substantiate the known sex difference in liver lactogenic receptor levels (Norstedt et al., 1984). In this case, a 35-fold higher amount of radioactivity in the female compared to male rat liver was noted, and also here around 90% of the homogenate radioactivity was found in the total membrane fraction. This sex difference in liver uptake of oPrl correlates well with the difference previously noted in lactogenic receptor levels (Norstedt et al., 1984). To check that the presence of 12’I-bGH radioactivity in the liver was actually due to active

17 TABLE UPTAKE FEMALE

1 OF 12’I-bGH RATS

AND

‘251-oPrl

IN

MALE

AND

Internalization of ‘251-bGH (2.4.106 cpm/lOO g body weight) and ‘251-oPrl (3.8. lo6 cpm/lOO g body weight) were terminated after 10 min by perfusion of the livers via the caval vein. Livers were removed followed by isolation of subcellular fractions. Radioactivity is expressed as cpm/g liver. Each group consisted of six animals where three rats received labelled hormone (total uptake) and three rats received labelled hormone together with a lOOO-fold excess of unlabelled hormone (nonspecific uptake). Specific uptake was calculated by subtracting non-specific uptake from total uptake. Each value represents the mean + SD of three animals. Sex

Fraction

Tracer

Specific radioactivity (cpm X lo-‘/ g liver)

Female

Homogenate Total particulate Fraction Cytosol Homogenate Total particulate Fraction Cytosol Homogenate Total particulate Fraction Cytosol Homogenate Total particulate Fraction Cytosol

12sI-bGH

426

+32

12sI-bGH 12sI-bGH 12’I-bGH

366 27 440

*46 + 5 +54

386 21 225

+35 f 3 *21

Male

Female

Male

“’ I-oPr1 ‘25 I-oPr1 ‘251-oPrl

206 k12 3.3* 1.1 6.5* 6.2 5.1+ NDa

4.9

’ ND, not detectable.

uptake and not solely to trapping in vesicles during homogenization, a control experiment was performed where ‘251-bGH with or without lOOO-fold excess of unlabelled bGH was added to the homogenization buffer followed by preparation of microsomes. This showed that only 4% of the radioactivity was trapped compared to 40% when 125I-bGH was administered and specifically internalized in vivo (data not shown). Transport of ‘251-bovine GH during endocytosis in male rat liver Since we could not demonstrate any sex differences in the specific liver uptake of ‘251-bGH in crude membrane and soluble fractions, the in-

tracellular movement of ‘251-bGH during endocytosis was further studied in isolated subcellular membrane fractions only from male rat liver. These fractions have been defined by marker enzyme analysis and ultrastructural morphology (Andersson et al., 1978; Andersson and Eriksson, 1981; Khan et al., 1981; Ferland et al., 1984; Eriksson et al., 1986). In brief, the residual microsomal fraction (RM) predominantly consists of vesicles derived from the rough and smooth endoplasmic reticulum and exhibits a 2- to 3-fold relative enrichment (on a protein basis) of the ER marker NADPH-cytochrome c reductase compared to a total particulate fraction. The enrichment values for F--stimulated adenylate cyclase (blood-sinusoidal plasma membrane), tartrate-sensitive acid phosphatase (lysosomes) and UDP-galactosyltransferase using N-acetylglucosamine as acceptor (Golgi complex) in the RM fraction are l.O-, l.O-, and 0.5-fold, respectively (Eriksson et al., 1986). The endosome I fraction (E I) is a pool of Golgi cisternae and vesicles in addition to secretory vacuoles and endocytic vesicles. This fraction exhibits a 40- to 50-fold enrichment of galactosyltransferase activity, over a lOO-fold enrichment of transferrin receptors (Eriksson et al., 1986) a 4to 5-fold increase in adenylate cyclase activity and no enrichment in acid phosphatase activity. The endosome II fraction contains low-density membranes (d s 1.11) and floats on top of a discontinuous metrizamide gradient developed for the isolation of mature or secondary lysosomes (L) from a crude mitochondrial-lysosomal fraction (Wattiaux et al., 1978). Morphological analysis of the E II fraction shows a large proportion of lipoprotein-filled vacuoles resembling secretory vacuoles, in addition to typical secondary lysosomes/dense bodies (Khan et al., 1981; Ferland et al., 1984). Marker enzyme analysis demonstrates a 3-, lo-, and 2.5fold enrichment of galactosyltransferase, acid phosphatase and adenylate cyclase activities, respectively. Finally, the L fraction is 20-fold enriched in acid phosphatase activity and 9-fold enriched in galactosyltransferase activity (Norstedt et al., 1984). No enrichment of adenylate cyclase activity was found (unpublished data). Fig. 1 shows the radioactive content of various subcellular membrane fractions at various time

18

Ii

RM

I

E

Eli

L

P

Fig. 1. Transport of ‘251-bGH during endocytosis in male rat liver. Male rats received an injection of ‘*‘I-bGH (6.3.106 cpm/lOO g body weight) into the portal blood. After 5 (open columns), 10 (filled columns), 20 (hatched columns) and 40 (cross-hatched columns) min of Iabelling, internalization was terminated by perfusing the livers with cold 0.25 M sucrose, Livers were removed and subcellular fractionation performed as described in Materials and Methods. The radioactive content, expressed as cpm/mg protein, was determined in the homogenate (H), residual microsomes (RM), intermediate fraction (I), endosome I fraction (E), endosome II fraction (E II), lysosome fraction (L) and plasma membranes (P). Each value represents mean + SEM of three experiments.

points following intramesenteric-intraportal injections of ‘251-bGH. The data represent specific uptake, i.e. non-specific uptake has been subtracted from total uptake. The labelling time was defined from time of injection (0 min) to termination of labelling by retrograde perfusion of the liver with cold homogenization buffer. The aim of liver perfusion was, in addition to termination of labelling, also to remove unbound ligand. In the homogenate (H) there was a rapid association with a peak at 5 min, followed by a gradual decrease during the labelling interval up to 40 min. A similar pattern of association and decay was also noted in the residual microsomal (RM) fraction and in the plasma membrane (P) fraction. Since the homogenate contains the total complement of membranes involved in bGH binding and endocytosis, the data indicate that a large fraction of ‘251-bGH is only temporarily associated with liver cells. In the endosome I fraction (E I), a relatively constant level of radioactivity was observed between 5 and 20 min. The enrichment

of radioactivity at the peak value was 20- to 25fold on a protein basis compared to the homogenate. In the E II fraction, the radioactive content, enriched lo- to 15-fold compared to the homogenate, was slightly shifted towards later time points compared to the E I fraction, i.e. lo-20 min. In the lysosome fraction (L), a distinct peak was noted at 20 min. Thus, ‘251-bGH appears to associate in a sequential manner first with plasma membrane-derived fragments (also recovered in the homogenate and residual microsomes), then in the endosome I fraction and later in the endosome II fraction. Those latter fractions probably contain different endocytic vesicles. Finally, bGH arrives in the lysosomal compartment (L fraction). Internalization of the somatogenic receptor in endocytic and lysosomal compartments Since we and others have shown that GH is internalized to intracellular compartments it was of interest to elucidate the fate of the somatogenic receptor during ligand internalization in male and female rats. For this purpose, affinity cross-linking of the internalized ligand to its receptor followed by SDS-PAGE and autoradiography was performed. ‘*‘I-bGH was injected in the absence or presence of a lOOO-fold excess of unlabelled hormone. Labelling proceeded for time points earlier shown as optimal for each fraction (see Fig. 1, i.e. 10 min for E, 20 min for E II and L). After membrane preparation, the ligand was affinity cross-linked to its receptor by DSS and aliquots of the labelled membranes were then subjected to SDS-PAGE. In the two endosomal fractions, i.e. E and E II, and in the lysosomal (L) fraction, strongly labelled species of M, 117 000, 86 000, 77 000 and 65 000, and a fainter species of M, 57000 were noted (Fig. 2). It may, however, be noted that in females, the band at M, 117 000 is more heavily labelled in the endosome and endosome II fraction as compared to males, where, on the other hand, the M, 77000 complex was more intensely labelled. All five bands were considered specific for the somatogenic receptor since lOOO-fold unlabelled, coinjetted bGH competed for the binding. This experiment therefore suggests that at least a sizeable fraction of somatogenic receptors is transported together with ligand in endocytic

19

vesicles and to lysosomes the ligand.

internalization

of

Comparison between the sub~e~~~~ar d~~ir~bul~on of “-‘l-ovine Prl and ‘“Sf-bavine GH jo 11o wing iabelling in vivo Since it is now well established that Prl is internalized in the liver by receptor-mediated endocytosis (Josefsberg et al., 1979; Pastel-Vinay et al., 1982), and the intracellular pathway of Prl internalization has been defined by radioautography at the EM level (Bergeron et al., 1983) and by subcellular membrane isolation techniques (Josefsberg et al., 1979; Postel-Vinay et al., 1982), it was of interest to compare liver uptake of bGH via the somatogenic receptor with the much better characterized Prl receptor uptake system. The result of this comparative study is shown in Table 2. Female rats were chosen for the study of ‘251-oPrl endocytosis due to their much higher content of hepatic Prl receptors compared to male rats. On the other hand, the relative subcellular distribution or enrichment of Prl receptors compared to a crude membrane fraction appears to be identical in both sexes (Norstedt et al., 1984). After 10 min of Iabelling in vivo, 83% of specifically associated/ internalized “‘1-oPrl radioactivity was associated with the subfractions isolated from the homogenate. A 95 and 73-fold enrichment on a protein basis over the homogenate was noted in the Golgi

Fig. 2. Affinity cross-Ii&@ of internalized somatogenic receptors in male and female rat livers. ‘25 I-bGH (10.5. lo6 cpm/lOO g body weight) was injected in vivo and Iabelling proceeded for 10 min (endosome) or 20 min (endosome II and lysosomes) followed by preparation of endosome (lanes A-D), endosome II (lanes E-H) and lysosome (lanes I-L) fractions. The subcelIular fractions were prepared from both male (lanes A, B, E, F, I and J) and female (lanes C, D, G, H, K and L), rats, respectively. In separate animals a lOOO-fold excess of unIabelled bGH was coinjected with tracer (lanes B, D, F, H, J and L). Subsequent to fractionation, the internalized tracer was cross-linked to the receptor using DSS (0.5 mM). Samples (E, 300 pg, E II, 700 gg and L, 1000 pg) were subjected to SDS-PAGE on 7.5% polyac~la~de gets under reducing conditions. The positions of “C-IabelIed molecular weight markers are indicated on the left side of autoradiogram. The positions of the major labelled species are indicated on the right side of the autoradiogram.

TABLE

during

2

COMPARISON

OF UPTAKE

OF ‘=I-bGH

AND

“‘1-oPrl

IN SUBCELLULAR

MEMBRANE

FRACTIONS

The subcellular distribution of lz51-bGH (1.106 cpm/lOO g body weight) and ‘251-oPrl (4.6.106 cpm/lOO g body weight) is are described in expressed as cpm/g liver and cpm/mg protein. The subcellular fractionation procedures and fraction terminology Materials and Methods. Female rats were used for injection of ‘251-oPrl and male rats for ‘251-bGH. RSU denotes relative specific uptake compared to the homogenate. Specific uptake is described as mean i SD. Fraction

Homogenate Residual microsomes Intermediate Endosomes Endosomes II Lysosomes

12’ I-0Prl

‘251-bGH cpm X 10e3/g

cpmX 10W3/mg

cpmX 10m3/g

Liver

RSU

Protein

RSU

Liver

RSU

Protein

RSU

(lO@ (16) (8) (16) (5) (1)

2+ 2* 12+ 56+13 36+ 13+

(1) (1) (6) (28) (18) (7)

197+ 39 32+ 9 3oi 7 49*13 46+14 8& 3

(100) (161 (15) (25) (23) (4)

I+ 0.5 2* 1 16k 2 95+15 73+ 7 14+ 2

(1) (2) (16) (95) (73) (14)

219f2l 36& 17* 34f 11* 34

6 2 4 5 0.5

0.5 0.1 2 8 4

cpm X 10W3/mg

and Ll fractions, respectively, suggesting that these fractions are enriched in endocytic vesicles. In the case of 12’I-bGH, however, only 46% of homogenate-associated radioactivity could be recovered in the subfractions and a correspondin~y lower enrichment was found, most evident in the endocytic fractions, i.e. E I and E II. These findings suggest that although Prl and GH share endocytic pathways leading to lysosomes and degradation, the kinetics of this intracellular transport appears to be different. This finding furthermore strengthens the notion that endocytosis of bGH and oPrI occurs via different cell-surface receptors.

In vitro degradation of f25f-bovine GH in various subcellular fractions from male rat liver In order to analyze the integrity of ‘251-bGH, particularly in membranes involved in growth hormone endocytosis, I25I-bGH was incubated with plasma membranes, different endosome-enriched fractions (E II and E) and lysosomes (L) in vitro (Fig. 3). The same conditions were used as for receptor binding analysis, i.e. pH 7.4 at 20 o C for 16 h, and the whole incubation mixture was then subjected to SDS-PAGE under reducing conditions. A fragment with &f, 15000 as well as intact ‘25I-bGH with i&f, 23 000 were seen in all fractions analyzed. Particularly, in the plasma membranes (P), lysosomes (L) and the ‘early’ endosomes (E), more &f,. 15000 fragments were generated than in the ‘iate’ endosome fraction E II. Addition of a lOOO-fold excess of unlabelled bGH during incubation led to decreased formation of the 15000 Da fragment, possibly suggesting that the degradation of “‘1-bGH is receptor dependent. When the iodinated bGH preparation was incubated in the absence of membranes, only the 23000 Da intact hormone was recovered (data not shown). When incubation of intact ‘251-bGH was carried out at a pH of 5.8 the &f, 15000 fragment appeared to be further degraded into smaller fragments in the presence of lysosomes, while endocytic vesicles (E and E II) only metabolized the M, 23 000 intact hormone to the M, 15 000 fragment (data not shown).

fracFig. 3. Degradation of ‘*‘I-bGH by different subcellular tions from male rat livers. 12’I-bGH (300000 cpm) was incubated with 50 gg plasma membrane (lanes A and B), 50 pg lysosome fraction (lanes C and D), 30 pg endosome II fraction (lanes E and F) and 30 fig endosome fraction (lanes G and H). Incubations proceeded in the absence (lanes A, C, E and G) or presence (lanes B, D, F and H) of 1000-fold excess of bGH for 18 h at 20 o C. Samples were subjected to SDS-PAGE on 15% polyacrylamide gels under reducing conditions. The positions of “C-labelled molecular weight markers are indicated on the left of the autoradiogram. The positions of the major labelled species are indicated on the right of the autoradiogram.

Effect of serine protease inhibitors on the formation of proteoIytic bGH fragment in plasma membranes In order to elucidate if the 15000 Da bGH fragment was generated by protease action and, if so, which type of protease that was involved, ‘251-bGH was incubated in vitro with plasma membranes in the absence or presence of different protease inhibitors (Fig. 4). The formation of the M, I5000 fragment was found to be inhibited by leupeptin, PMSF and aprotinin, while TLCK and cu,-antitrypsin inhibitor did not prevent the formation of the 15 000 Da fragment, suggesting that the protease present in

21

brane protein per ml buffer to generate the 15000 Da fragment. The experiment was performed to see if the 15 000 Da fragment binds to a defined species of somatogenic binders in male and female rat livers. Therefore, the proteolytic fragment obtained after incubating ‘251-bGH with 1 mg plasma membranes, was incubated with endosomes and affinity cross-linked to the receptor followed by analysis with SDS-PAGE and autoradiography. As

Fig. 4. Effect of protease inhibitors on “‘1-bGH proteolysis by plasma membranes from male livers. 12’I-bGH (300000 cpm) was incubated with 50 /.tg plasma membrane from male rat livers in the absence of protease inhibitors (lane A) or presence of 50 gg/ml leupeptin (lane B), 1 mM PMSF (lane C), I TIU/ml aprotinin (lane D), 5 pM TLCK (lane E) and 100 pg/ml a,-antitrypsin inhibitor (lane F). Incubations proceeded for 18 h at 20°C. Samples were subjected to SDS-PAGE on 15% polyacrylamide gels under reducing conditions. The positions of i4C-labelled molecular weight markers are indicated on the left side of the autoradiogram. The positions of the major labelled species are indicated on the right side of the

plasma membranes and active on bGH is a nontrypsin-like serine protease. Binding of M, I5 000 bGH fragment to somatogenic receptors The formation of the 15000 Da fragment was dependent on the amount of plasma membrane protein added to the incubation, and at 1 mg plasma membrane protein and 5~~ cpm 12’1bGI-I per ml incubation buffer, an almost complete conversion of intact hormone to fragment occurred (data not shown). Therefore, in the following experiment, we used 1 mg plasma mem-

Fig. 5. Affinity cross-linking of 15 000 Da lz51-bGH proteolytic fragment to endosome fractions (E) from male and female rat livers. The 15000 DA fragments was generated by incubating 1 mg plasma membrane protein with 5000@0 cpm ‘251-bGH for 18 h at 20°C. An aliquot of this 15000 Da fragment (approximately 4OOOB cpm) was then isolated and reincubated with fresh endosomes (150 ~.rg protein) from males (lanes C and D) and females (lanes G and H). Native hormone was incubated with fresh endosomes from males (lanes A and B) and females (lanes E and F). Incubations were carried out in the absence (lanes A, C, E and G) or presence (lanes B, D, F and H) of lOOO-fold excess of unlabelled bGH. Samples were subjected to SDS-PAGE on 7.5% polyacryiamide gels under reducing conditions. The positions of t4C-labelled molecular weight markers are indicated on the left of the autoradiogram. The positions of the major labelled species are indicated on the right of the autoradiogram.

22

shown in Fig. 5, binding of the 15000 Da fragment results in a receptor-ligand complex species with apparent molecular weights of 65000 and 117000 as with native intact hormone. One would, however, expect a receptor-aged complex with lower molecular weight when the M, 15000 fragment was incubated with the membranes compared to the M, 23000 intact hormone. The absence of such difference is possibly due to limitations in the resolution of this gel system. Discussion As a continuation of our previous studies on the mechanisms of regulation of sex-differentiated liver functions by growth hormone, and in particular regarding the role of intracellular somatogenic receptors related to GH action, we have in the present study examined receptor-mediated binding and uptake of “‘1-bGH, a specific ligand for the somatogenic receptor (Husman et al., 1985), in male and female rat liver following in vivo administration. Furthermore, we have aimed at defining the intracellular loci of GH degradation in vitro by an endogenous protease and at comparing the intracellular pathway of bGH endocytosis with that of the much better characterized lactogenic receptor system. It was initially established that the uptake of ‘251-bGH and ‘2sI-oPrl are receptor mediated and furthermore are well correlated with the number of receptors on the cell surface for each ligand. The uptake of ‘251-Prl was around 35-fold higher in females than in males, which matches the difference in lactogenic binding sites between the sexes (Norstedt et al., 1984). On the other hand, the uptake of 12’I-bGH was not sexually differentiated, again in agreement with our previous ligand binding data (Husman et al., 1985). This experiment also indicates that the uptake of ‘251-bGH results from binding to somatogenic receptors and not to lactogenic receptors. In a series of elegant studies regarding endocytosis via the lactogenic receptor system (Khan et al., 1981, 1986; Bergeron et al., 1983, 1986), the intracellular pathway of Prl receptor endocytosis in the female liver have been elucidated. Following initial binding to Prl receptors at the cell surface, Prl is internalized into several successive

endocytic compartments followed by deposition into lysosomes in the vicinity of bile canaliculi. We have previously demonstrated that a lowdensity membrane fraction isolated from crude microsomes is highly enriched in both lactogenic (Norstedt et al., 1984) and somatogenic (Husman et al., 1985) receptors. The kinetics of ‘251-bGH turnover in subcellular fractions isolated from male rats subsequent to in vivo administration of tracer (Fig. X), suggests that r2*I-bGH initially associates with membranes in the endosome I (E I) fraction, then in the endosome II (E II) fraction and is later transported into membranes residing in the mature lysosomal (L) fraction. Thus, ‘*‘I-bGH is endocytosed and sequentially transported in distinct endocytic ~mpartments prior to arrival into lysosomes, in analogy with endocytosis via the lactogenic receptor. The early endosome fraction E I contains in addition to endocytic vesicles also membranes derived from the Golgi complex, as evident by the enrichment of the trans-Go&i marker &-galactosyltransferase (Eriksson et al., 1986). However, since the Golgi complex does not appear to participate in receptor-mediated endocytic events (Goldstein et al., 1985), the presence of Golgi membranes in the endosome fractions is not likely to obviate the above inte~retations. To address the question whether the somatogenie receptor is transported together with its ligand during internalization, an experiment was performed where ligand was cross-linked to the receptor in isolated endocytic and lysosomal fractions after internalization to intracellular compartments in vivo. The affinity cross-linked receptorligand complex in these fractions was then analyzed by SDS-PAGE. Distinct somatogenic binders could be seen in the fractions and p~ticularly pro~nent binders were noted of M, 95 000 and 55 000 (after subtraction of M, 22000 of the ligand) (Fig. 2). The presence of several binders involved in ligand internalization may reflect receptor heterogeneity in the plasma membrane or that the receptor and/or ligand is proteoly~cally cleaved during internalization. In this respect, several studies have indicated that endocytic compartments harbour proteolytic capacity (Diment and Stahl, 1985). By cross-linking of GH to rat hepatocytes, it has been

23

demonstrated that GH binds to receptors of M, 280 000, 200000 and 100 000, assuming a 1: 1 binding of ligand to the receptor (Yamada and Donner, 1984). The reason why we find receptorligand complexes with lower molecular weight during internalization is not quite clear but by incubating ‘*%bGH with isolated subcellular membranes, e.g. with endosomes, binders of M, 95 000, 86 000, 5.5000 and 43 000 are found (Husman et al., 1988). These species are not affected by protease inhibitors either during membrane preparation or during incubation. The fact that we see the same receptor-ligand complexes in both endosomes (E and E II) and tysosomes (L) strongly suggests that at least a fraction of the receptor population is transported in endocytic vesicles and to lysosomes together with the ligand. This notion is supported by the work of Baxter (1985) where rate of disappearance of GH binding capacity in rat livers treated with cycloheximide showed a half-time of 30-40 min. We also find that maximal en~chment of internalized ‘251-bGH in isolated lysosomes is attained at around 30 min (Fig. 1). These lines of evidence indicate that down-regulation of somatogenie receptors is the result of transport of internalized receptor-ligand complexes to lysosomes. The biological significance of ligand and receptor internalization is not well understood but the prevailing view is that the ligand and/or receptor has to be degraded in the lysosomes to terminate its action. On the other hand, internalization may be of importance to generate second messengers which then mediate intra/intercellular events. The possibility that growth hormone is degraded to specific fragments during receptormediated internalization was considered to be of interest since several other investigations have shown that GH can be cleaved by various proteases like trypsin and plasmin and that some of the fragments formed seemed to have or to stimulate somatomedin-like activity (Mittra et al., 1984) indicating that proteolytic cleavage of GH may affect its biological activity. In this context, Schepper et al. (1984) have presented evidence that an isolated rat liver plasma membrane fraction upon incubation with ‘251-rGH produces a proteolytic fragment with a size of 15000 Da, which appears to bind to somatogenic receptors

with high affinity. In the present study, various subcellular fractions were examined for the presence of GH-protease activity. The putative endogenous protease was allowed to react with ‘251-bGH, and it was found that cleavage of GH to generate a 15000 Da fragment occurred in isolated lysosomes, plasma membranes and in endosome fractions (Fig. 3). This suggests that proteolysis of GH to generate a distinct 15 000 Da fragment is widely occurring in organelles involved in hormone endocytosis and degradation. A similar type of processing mechanism has been shown for flglucuronidase, which is internalized by mouse Lcells via the Man 6-P receptor to undergo a proteolytic cleavage in endocytic compartments on its way to the lysosomes (Gabel and Foster, 1987). Addition of unlabelled ligand in excess decreased the amount of radioactive 15000 Da fragment. This may either indicate that the proteolytic activity is more effective when hormone is bound to the receptor, or that GH is cleaved by a rather specific protease. The proteolysis was blocked by serine protease inhibitors like PMSF, aprotinin and leupeptin, but not by TLCK or a,-antitrypsin, suggesting that the protease is a non-trypsin-like serine protease, in agreement with the work of Schepper et al. (1984). The incubations with various membranes were performed at pH 7.4. When the same experiment was run under more acidic conditions, pH 5.8, the 15000 Da fragment was, only by the lysosomes, totally degraded to very small fragments, not resolved by the gel (data not shown). This indicates that the protease studied at pH 7.4 is probably not a classical lysosomal enzyme. It was furthermore shown, using affinity crosslinking, that the A4, 15000 proteolytic fragment specifically bound to similar binders as the native hormone, thereby suggesting that this fragment, when formed in vivo, can interact with somatogenie receptors. Several studies have indicated that specific proteolysis may in fact be a general event in the intracellular action of polypeptide hormones. This notion is emphasized by the demonstration of the existence of an intracellular insulin-degrading cysteine protease in many target tissues (Shii et al., 1985). It has furthermore been demonstrated that certain actions of insulin such as increases in glucose transport and lipogenesis

24

seem to involve a post-binding proteolytic reaction (Cherqui et al., 1985). It is conceivable that proteolytic fragments of internalized polypeptide hormones may be responsible for parts of the pleiotropic response often seen following hormone binding. In conclusion, in the present study we have found that GH is internalized to two endocytic compartments before arrival in the lysosomes. Using affinity cross-linking and SDS-PAGE we could demonstrate similar receptor-ligand complexes in both endosomes and lysosomes, suggesting that the somatogenic receptor in a complex with its ligand is transported to the lysosomes. By comparing the uptake of ‘251-oPrl with 1251bGH, which correlated well with its respective receptor number in males and females, it was possible to conclude that GH uptake was not mediated via lactogenic receptors. For instance, no sex difference in the uptake of GH could be shown. Finally, bGH was degraded to an M, 15000 fragment by a non-trypsin-like serine protease found in membranes and this fragment as well as intact hormone bound to defined somatogenic receptors. Acknowledgements We express our thanks to Marie-Louise Hildenborg for skilled technical assistance. This work was supported by grants from the Karolinska Institute, Kabi Vitrum AB and by grants 13 X-2819 and 12 X-7141 from the Swedish Medical Research Council. References Andersson, G.N. and Et&son, L.C. (1981) J. Biol. Chem. 256, 9633-9639. Andersson, G.N., Tomdal, U.-B. and Eriksson, L.C. (1978) B&him. Biophys. Acta 512, 539-549. Asakawa, K., Grunberger, G., McElduff, A. and Gordon, P. (1985) Endocrinology 117, 631-637. Baxter, R.C. (1985) Endocrinology 117, 650-655. Baxter, R.C., Zaltsman, Z. and Turtle, J.R. (1984) Endocrinology 114,1893-1901. Bensadoun, A. and Weinstein, D. (1976) Anal. B&hem. 70, 241-250. Bergeron, J.J.M., Resch, L., Rachubinski, R., Pate], B.A. and Posner, B.I. (1983) J. Cell Biol. 96, 875-886.

Bergeron, J.J.M., Searle, N., Khan, M.N. and Posner, B.1. (1986) Biochemistry 25, 1756-1764. Carter&, C., Schwartz, J. and Kikuchi, G. (1984) J. Biol. Chem. 259, 1099-1104. Cherqui, G., Caron, M., Capeau, J. and Picard, J. (1985) B&hem. J. 227, 137-147. Diment, S. and Stahl, P. (1985) J. Biol. Chem. 260, 15311-15317. Eden, S. (1979) Endocrinology 105, 555-560. Eriksson, L.C., Torndal, U.-B. and Andersson, G.N. (1986) Carcinogenesis 7, 1467-1474. Farquhar, M.G. (1983) Fed. Proc. 42, 2407-2413. Gabel, C.A. and Foster, S.A. (1987) J. Cell Biol. 105, 1561-1570. Goldstein, J.L., Brown, M.S., Anderson, R.G.W., Russel, D.W. and Schneider, W.J. (1985) Annu. Rev. Cell Biol. 1, l-39. Gorin, E. and Goodman, H.M. (1984) Endocrinology 114, 1279-1286. Gustafsson, J.-A., Mode, A., Norstedt, G. and Skett, P. (1983) Annu. Rev. Physiol. 45, 51-60. Hughes, J.P., Simpson, J.S.A. and Friesen, H.G. (1983) Endocrinology 112, 1980-1985. Husman, B., Andersson, G., Norstedt, G. and Gustafsson, J.-A. (1985) Endocrinology 116, 2605-2611. Husman, B., HaldosCn, L.-A., Andersson, G. and Gustafsson, J.-A. (1988) J. Biol. Chem. (in press). Josefsberg, Z., Posner, B.I., Pate& B. and Bergeron, J.J.M. (1979) J. Biol. Chem. 254, 209-214. Khan, M.N., Posner, B.I., Verma, A.K., Khan, R.J. and Bergeron, J.J.M. (1981) Proc. Natl. Acad. Sci. U.S.A. 78, 4980-4984. Khan, M.N., Savoie, S., Bergeron, J.J.M. and Posner, B.1. (1986) J. Biol. Chem. 261, 8462-8472. Laemmli, U.K. (1970) Nature 227, 680-685. Lewis, U.J., Singh, N.P., Vanderlaan, W.P. and Tutwiler, G.F. (1977) Endocrinology 101, 1587-1603. Liberti, J.P. and Durham, III, L.A. (1983) J. Endocrinol. 96, 195-199. Liberti, J.P. and Miller, M.S. (1978) Endocrinology 103, 29-34. McGeoch, C., Morgan, E.T., Halpert, J. and Gustafsson, J.-A. (1984) J. Biol. Chem. 259, 15433-15439. McGeoch, C., Morgan, E.T. and Gustafsson. J.-A. (1985) Endocrinology 117, 2085-2092. Mittra, I. (1984) Cell 38, 347-348. Mode, A., Norstedt, G., Simic, B., Eneroth, P. and Gustafsson, J.-A. (1981)Endocrinology 108, 2103-2108. Morgan, E.T., McGeoch, C. and Gustafsson, J.-A. (1985) J. Biol. Chem. 260, 11895-11898. Norstedt, G., Mode, A., Eneroth, P. and Gustafsson, J.-A. (1981) Endocrinology 108, 1855-1861. Norstedt, G., Andersson, G. and Gustafsson, J.-A. (1984) Endocrinology 115.672-680. Gpresko, L.K. and Karpf, R.A. (1987) Cell 51, 557-568. Postel-Vinay, M.-C., Kayser, C. and Desbuquois, B. (1982) Endocrinology 111, 244-251. Schepper, J.M., Hughes, E.F., Pastel-Vinay, M.-C. and Hughes, J.P. (1984) J. Biol. Chem. 259, 12945-12948. Shii, K., Baba, S., Yokono, K. and Roth, R.A. (1985) J. Biol. Chem. 260, 6503-6506.

25 Singh, R.N.P., Seavey, B.K., Rice, V.P., Lindsey, T.T. and Lewis, U.J. (1974) Endocrinology 94, 883-891. Smith, W.C. and Talamantes, F. (1987) J. Biol. Chem. 262, 2213-2219. Song, C.S., Rubin, W., Rifkind, A.B. and Kappas, A. (1969) J. Cell Biol. 41, 124-132.

Wattiaux, R., Wattiaux-De Coninck, S., Ronveux-Dupal, M.F. and Dubois, F. (1978) J. Cell Biol. 78, 349-368. Yamada, K. and Donner, D.B. (1984) B&hem. J. 220,361-369.