Characterization of Hepatic Lactogen Receptor SUBUNIT COMPOSITION AND HYDRODYNAMIC PROPERTIES*

The structure of the membrane-bound and Triton X- 100-solubilized female rat liver prolactin receptor has been studied by affinity cross-linkingfsodium dodecyl sulfate-polyacrylamide gel electrophoresis, gel filtration, and sucrose/HzO and sucrose/DzO density gra- dient centrifugation. Hydrodynamic characterization revealed that the ‘2SI-human growth hormone recep-tor-detergent complex reprysents a molecular species with a Stokes radius of 61 A, a sedimentation coeffi-cient of 5.0 s, and a calculated molecular weight of 158,000. The molecular weight of the receptor was calculated to be 92,000. Three lactogenic hormone- binding species with M, values of 87,000,40,000, and 35,000, respectively, were repeatedly found when de- tergent-solubilized preparations were analyzed using an affinity cross-linking technique. Estrogen treat- ment of female rats increased the intensity of these bands. Occasionally, an M, 165,000 hormone-binding species was also found. Two-dimensional sodium do- decyl sulfate-polyacrylamide gel electrophoresis studies (first dimension, nonreducing; second dimension, reducing) demonstrated that disulfide- and nondkul-fide-linked hormone-binding species with M, values of 40,000 and 35,000 are contained within the M, 87,000 species. It is concluded that the Triton X-100-solubi- lized female rat liver prolactin receptor has a molecular weight of about 90,000.

The structure of the membrane-bound and Triton X-100-solubilized female rat liver prolactin receptor has been studied by affinity cross-linkingfsodium dodecyl sulfate-polyacrylamide gel electrophoresis, gel filtration, and sucrose/HzO and sucrose/DzO density gradient centrifugation. Hydrodynamic characterization revealed that the '2SI-human growth hormone receptor-detergent complex reprysents a molecular species with a Stokes radius of 61 A, a sedimentation coefficient of 5.0 s, and a calculated molecular weight of 158,000. The molecular weight of the receptor was calculated to be 92,000. Three lactogenic hormonebinding species with M , values of 87,000,40,000, and 35,000, respectively, were repeatedly found when detergent-solubilized preparations were analyzed using an affinity cross-linking technique. Estrogen treatment of female rats increased the intensity of these bands. Occasionally, an M, 165,000 hormone-binding species was also found. Two-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis studies (first dimension, nonreducing; second dimension, reducing) demonstrated that disulfide-and nondkulfide-linked hormone-binding species with M, values of 40,000 and 35,000 are contained within the M, 87,000 species. It is concluded that the Triton X-100-solubilized female rat liver prolactin receptor has a molecular weight of about 90,000. This molecular species contains M, 40,000 and M, 35,000 hormone-binding subunits. It cannot be determined whether these subunits are combined with each other or with as yet undetected subunit(s) to make up the M. 90,000 species, or whether each one of these subunits is a proteolytic fragment of the M, 90,000 species.
Characterization of the prolactin-binding entities in rat liver has shown that they are membrane-bound glycoproteins (17,18). Structural data for the receptors have been obtained by the use of different techniques such as gel filtration and sucrose density gradient centrifugation as well as affinity Cross-linking of iodinated hormone and receptor combined with SDS-PAGE' and autoradiography. The latter method has revealed specific binding species with M , values ranging from 28,000 to 83,000 (19-23). No conclusive data have yet been presented ruling out a possible relation between the lower and higher molecular weight species. By hydrodynamic characterization of detergent-solubilized rat liver membrane preparations, molecular weights of 73,000 and 78,000 for the lactogenic receptor could be calculated (24,25).
Using affinity cross-linking, molecular characteristics for the lactogenic receptor in rat tissues other than the liver have been examined. In the ovary two specific binders with molecular weights of 80,000 and 40,000, respectively, were found (26). Results from experiments employing a two-dimensional electrophoretic technique indicated that an M , 40,000 binder is contained within the M , 80,000 binder. Affinity crosslinking has also been used to investigate the molecular characteristics of lactogenic binders in Leydig cells, mammary gland, and kidney (23). The molecular weights presented for specific lactogenic binders were: 91,000, 81,000, 37,000 and 31,000 (Leydig cells); 93,000,83,000,30,000, and 28,000 (mammary gland); and 65,000 and 30,000 (kidney). It was also shown that the lactogenic binders in Leydig cells are similar to the binders in the ovary with respect to an M, 37,000 species being contained within the M, 81,000 species. In lactogenic hormone target tissues in other animals, such as pig and rabbit mammary gland, prolactin-binding species with molecular weights ranging from 28,000 to 69,000 have been described (22,(27)(28)(29).
Earlier studies in our laboratory have shown that the sexually differentiated content of prolactin receptors in rat liver is regulated by the sexually differentiated secretory rhythm of growth hormone (30). In order to study this regulation at a molecular level, a knowledge of the structure of the prolactin receptor in liver is needed. This knowledge would also help in the purification of this receptor.
In an attempt to elucidate the structure of the lactogenic receptor in rat liver, we have undertaken an investigation involving analysis of '251-hGH-binding species by affinity cross-linking, gel filtration, and sucrose density gradient centrifugation.

RESULTS
In this study hGH (Crescormon) and biosynthetic methionyl somatotropin (Somatonorm) have been used for labeling specific lactogenic-binding sites in different preparations of female rat liver. No quantitative or qualitative differences were observed between these hormones. In the following text hGH will be used to designate both hormones.
Affinity Cross-linking Studies-The structure of the lactogenic receptor was studied by affinity cross-linking of lZ5I-hGH to binding species present in microsomal membranes or acid-precipitated, redissolved preparations of Triton X-100solubilized membranes, followed by SDS-PAGE and autoradiography.
The specificity of cross-linking lZ5I-hGH to acid-precipitated, redissolved preparations was determined by competition studies with different unlabeled hormones (Fig. 4). The specific bands at M , 109,000, 62,000, and 57,000 were completely reduced with hGH ( Fig. 4, lanes B and C ) . A reduction of these bands was also seen with the higher concentration of rat prolactin (Fig. 4, lanes D and E ) . Bovine growth hormone did not affect the appearance of these bands (Fig. 4, lanes F and G ) . A weak band at M, 187,000 was also seen in this experiment. It was reduced with hGH and a 500-fold excess of rat prolactin but not with bovine growth hormone. This band was only occasionally seen.
When a 1251-hGH-incubated and cross-linked microsomal membrane preparation was subjected to SDS-PAGE and autoradiography a single binding band at M , 62,000 was seen under both reducing and nonreducing conditions. The same analysis of Triton X-100-solubilized membranes revealed three lactogenic hormone-binding bands at M, 109,000, 62,000, and 57,000 and one somatogenic band at M, 54,000. The same lactogenic bands are also seen when microsomal membranes are treated with the detergent CHAPS.3 Due to the absence of the somatogenic binding species, the acidprecipitated preparation of Triton X-100-solubilized microsomal membranes was used in this study.
The cross-linking studies presented in this article were performed with preparations of livers from estrogen-treated female rats. To examine the effect of estrogen treatment on the intensities of the individual bands seen upon autoradiography, preparations were obtained from livers of treated and untreated female rats. These preparations, containing equal amounts of protein, were incubated with equal amounts of lZ5I-hGH, cross-linked with 0.5 mM DSS, and subjected to SDS-PAGE under reducing conditions (Fig. 5). The intensity of the bands at M , 109,000,62,000, and 57,000 were markedly increased when the preparation was taken from livers of estrogen-treated female rats (Fig. 5, lanes A and B ) .
A possible relationship between the different lactogenic hormone-binding species seen on autoradiograms was first investigated by comparing cross-linked samples run on SDS-PAGE under reducing or nonreducing conditions. It was shown that, in the absence of reductant, the intensities of the specific bands at M , 62,000 and 57,000 were reduced and that the intensities of the bands at M , 109,000 and 187,000 were increased as compared to the autoradiographic appearance of cross-linked samples run under reducing conditions (Fig. 6).  The radioactive spots to the right of the diagonal form represent an artifact produced by the cracking of the gel during the drying procedure.
specific bands with the two higher molecular weight bands.
To examine this further, the lZ5I-hGH-incubated and crosslinked acid-precipitated and redissolved preparation was investigated by two-dimensional SDS-PAGE in which the first and second dimensions were run in the absence or presence of reductant, respectively. As can be seen in the autoradiogram, the diagonal form contains three radioactive spots at MI 109,000, 62,000, and 57,000 (Fig. 7). Two spots with Mr 62,000 and 57,000 are also seen straight below the spot at M , 109,000, implying that they are derived from the higher molecular weight spot. Unfortunately, in several experiments, we were unable to detect any spot at M , 187,000.
The presence of a radioactive spot at MI 109,000 in the two-dimensional SDS-PAGE could be the result of intermolecular cross-linking of possible subunits of lactogenic receptor in addition to cross-linking of lZ5I-hGH with subunit(s). To test this hypothesis the following experiment was performed. A '251-hGH-incubated microsomal membrane preparation was cross-linked with DSS, and after Triton X-100 solubilization of membrane proteins the sample was divided into two portions. One of these was again subjected to crosslinking by DSS. SDS-PAGE was then run under reducing conditions. As can be seen in the autoradiogram (Fig. 8, lanes  A and B ) , cross-linking before Triton X-100 solubilization produced a specific band at M , 62,000. However, if crosslinking was also performed after detergent treatment, a specific band at M , 109,000 was seen in addition to the M , 62,000 band (Fig. 8, lanes C and D). It should be noted that no band at M , 57,000 was seen in the Triton X-100-solubilized membrane preparation in this experiment. These results suggest that the binder giving rise to the Mr 57,000 band is not exposed in the membrane preparation and thus cannot be cross-linked with '251-hGH. Treatment of membranes with detergents has been shown to expose a population of masked or cryptic lactogenic receptors for binding (39)(40)(41)(42).
To test this hypothesis, the '251-hGH-incubated microsomal membrane preparation was treated with Triton X-100 after removal of unbound hormone by centrifugation. This preparation was divided into two samples, one of which was reincubated with '251-hGH. Both samples were cross-linked with DSS and subjected to SDS-PAGE. The autoradiogram shows that reincubation of the Triton X-100-solubilized membrane preparation gives rise to an M , 57,000 band in addition to the M , 109,000 and 62,000 bands (Fig. 9, lane C). A specific band at M, 117,000 is also seen (Fig. 9, lane C). This band represents the labeling of a somatogenic binder? Prolonged exposure of the dried gel showed that the binders giving rise to the M, 117,000 and 109,000 bands were also present in the sample that was not reincubated (data not shown).

DISCUSSION
In this study basal characteristics such as precipitability, molecular weight, and subunit structure of the prolactin receptor in female rat liver have been studied.
When the pH of the Triton X-100-solubilized preparation was lowered, it was found that the solubilized proteins had a decreased solubility at pH values between 6 and 4. The maximal amount of protein precipitate and specific binding was found at a pH of 4.9. Chromatofocusing of unlabeled rat ovary prolactin receptors purified in Triton X-100 revealed three distinct peaks with PI values of 4.0, 5.0, and 5.3 (47). Isoelectric focusing of Triton X-100-solubilized unlabeled prolactin receptor from rabbit mammary gland gave two peaks with PI values of 5.0 and 5.9 (48). From these studies it might be concluded that the decreased solubility at pH values of 4-6 of the Triton X-100-solubilized specific binding seen in our study could be an isoelectric precipitation phenomenon. As judged from Scatchard analyses of '251-hGH binding in Triton X-100-solubilized and acid-precipitated redissolved preparations, an increased binding capacity was found in the latter preparation. This could either be due to enrichment of lacto-genic receptor, dissociation of endogenous hormone from the receptor, or a combination of both. Ovine prolactin specifically bound to rabbit mammary membrane receptor has been shown to dissociate in a pH-dependent manner with a pK of 4.7 (49).
When the redissolved precipitate was incubated with lz5I-hGH and subjected to cross-linking with DSS and run on SDS-PAGE, three lactogenic bands with M, 109,000, 62,000, and 57,000 were seen. Assuming a 1:l binding ratio and an M , of 22,000 for hGH, the M, values of the unoccupied binders would be 87,000,40,000, and 35,000, respectively (Fig. 12).
The finding of increased intensity of the M , 109,000 band and decreased intensity of the M, 62,000 and 57,000 bands in unreduced samples as compared to reduced samples suggested a direct relationship between the bands. In line with this notion, two-dimensional electrophoresis indicated that two binders with M, values of 40,000 and 35,000 are contained within the M, 87,000 binder (Fig. 12) and that these binders are disulfide-linked either to each other or to another protein present in the M, 87,000 binder. In addition, two labeled species with M, values of 62,000 and 57,000 were demonstrated along the diagonal seen in the autoradiogram, indicating that these binders with M , values of 40,000 and 35,000 are not disulfide-linked. From these results it might also be deduced that there exist two populations of M, 87,000 binders, one in which the M, 40,000 and 35,000 subunits are disulfidelinked and one population in which they are not.
When running lZ5I-hGH-incubated and cross-linked microsomal membrane preparations on SDS-PAGE, under both reducing or nonreducing conditions, only one specific binder with an M, of 40,000 is seen.3 Due to the presence of a binder with an M, of 40,000 that is disulfide-linked, it can be concluded that detergent treatment exposes a binder with this molecular weight in addition to the nondisulfide-linked M, 40,000 binder present in the membrane.
The two-dimensional electrophoresis experiment also indicated that there might exist an M, 87,000 binder from which no subunits can be released under reducing conditions, but as seen in Fig. 8, the existence of this "refractory" binder is probably due to intermolecular cross-linking between subunits. It is interesting to note that this intermolecular cross-  . ( H 2 0 ) , sucrose/ H20 density gradient centrifugation; S.D.G. (H20/D20),sucrose/Hz0 and sucrose/D20 density gradient centrifugation; 61A, fraction of Sephacryl S-400 chromatography (eluate that was further analyzed); 20, two-dimensional SDS-PAGE R.Z., reincubation with lZ5I-hGH L, lactogenic binding activity; S, somatogenic binding activity. linking does not occur when microsomal membrane preparations are used. Thus, detergent treatment either brings subunits into favorable position for cross-linking, exposes suitable sites for cross-linking due to disruption of membrane vesicles, or removes membrane lipids surrounding the proteins which might hinder cross-linking. The results in Fig. 8 also indicate that the M, 40,000 binder present in membrane preparation is associated with an unlabeled subunit to which it can be cross-linked only when Triton X-100-solubilized (Fig. 13a).
The possibility exists that the M, 35,000 binder could be a proteolytically modified or a less glycosylated form of the M , 40,000 binder and that each one of these binders is combined with an unlabeled subunit with an M , around 50,000 and not seen by cross-linking (Fig. 13, c and d). If this were the case, one would expect that this would generate two bands with M, values of around 110,000. To improve resolution of higher molecular weight species we have run SDS-PAGE of crosslinked samples on 10% polyacrylamide gel but only found one band (not shown). It is possible that the resolution obtained with 10% gels is not enough and that there exists two populations of higher molecular weight binding species.
It can be concluded, both from data in Fig. 9 and from the two-dimensional electrophoresis experiment, that the M, 35,000 binder is contained within an M, 87,000 species and that the 35,000 binder is not available for cross-linking with lZ5I-hGH in the membrane preparation but only in Triton X-100-solubilized preparation. These results and the findings presented in Fig. 8, showing the presence of the M , 87,000 and 40,000 binders only when cross-linking was performed before and after Triton X-100 solubilization of incubated microsomal membranes, give further support to the notion that each one of the M, 35,000 and 40,000 binders is combined with an unlabeled subunit or subunits to make up the M, 87,000 species (Fig. 13, c and d).
If the M , 87,000 binder is composed of a combination of the M, 40,000 and 35,000 species, one question concerns the localization of the actual binding site for lZ5I-hGH. Both the M, 40,000 and 35,000 binders interact with lZ5I-hGH and are cross-linked with DSS. It seems unlikely that the M , 87,000 binder contains binding sites for two molecules of lZ5I-hGH, as this would generate a labeled band at a higher M , than 109,000. Covalent binding of hGH to purified prolactin receptor from rat ovary and analysis on SDS-PAGE of the complexes showed that each one of the purified proteins of M, 41,000 and 88,000 bound one molecule of hGH (47). The M , 41,000 protein has earlier been shown to be contained within the M, 88,000 protein (26). An explanation for the labeling of both subunits could be that one of these, e.g. the M, 40,000 binder, contains the actual binding site and that the M, 35,000 binder is close enough to the M, 40,000 binder to allow contact with lZ5I-hGH in Triton X-100-solubilized preparations, such that one part of the M, 40,000-35,000 hGH complex will be cross-linked between the M, 40,000 binder and lZ5I-hGH and one part of the same complex cross-linked between the M , 35,000 binder and "'I-hGH, giving rise to M, 62,000 and 57,000 bands, respectively (Fig. 13b). The contention that the M, 40,000 binder contains the actual binding site is supported by the fact that its autoradiographic intensity is higher than that of the M , 35,000 binder. The difference in intensity could also be explained by the fact that the M , 40,000 binder is a subunit of a receptor population that is more abundant than a receptor population in which the M, 35,000 binder is present (Fig. 13, c and d).
Another possibility is that the M, 40,000 and 35,000 binders represent proteolytical fragments of an M , 87,000 species which dissociate when run on SDS-PAGE (Fig. 13, c and d). This would agree with the results of Mitani and Dufau (47), who observed M , 41,000 and M, 88,000 proteins in purified preparations of rat ovary prolactin receptors as analyzed by SDS-PAGE run under reducing conditions, although these authors included protease inhibitors in their purification procedure.
Analysis by two-dimensional electrophoresis of lZ51-hGHincubated and cross-linked Triton X-100-solubilized rat ovary and Leydig cell membranes has revealed lower molecular weight species in the M, 35,000-40,000 region that were contained within a higher molecular weight species with an M , of 80,000-90,000 (23, 26). Thus, it seems that, at least in ovary, Leydig cells, and liver, similarities exist between the rat lactogenic receptors, although the reported number and molecular weights of the binding species in the different tissues are not the same.
In order to describe the hydrodynamic properties and the amount of detergent bound to the lZ5I-hGH receptor-detergent complex and to determine-by another method than SDS-PAGE-a molecular weight for the lactogenic receptor, a method described by Clarke was used (35).
Gel permeation chromatography of a '251-hGH-incubated acid-precipitated redissolved preparation yielded a single peak of radioactivity, representing the hormone receptor-detergent complex, with a Stokes radius of 61 A. Cross-linking and SDS-PAGE under reducing conditions of this species revealed three bands with M, values of 109,000, 62,000, and 57,000 ( Fig. 12). On the basis of these results and the results from two-dimensional electrophoresis it can be concluded that the M, 87,000 binder precipitates in an intact form. Combining the data obtained by gel permeation chromatography and sucrose density gradient analysis the molecular weight of the solubilized receptor was calculated as 92,000.
The derived values for the frictional and axial ratios indicate that the hormone receptor-detergent complex has an elongated structure. This explains in part its behavior on gel chromatography where it co-elutes with ferritin, which has a molecular weight of 443,000 (50). Several investigators have used only gel chromatography to determine the M, of the Triton X-100-solubilized lactogenic receptor and obtained values ranging from 220,000 to 380,000 (18,51,52). In those studies the amount of detergent and the shape of the hormone receptor-detergent complex have not been taken into consideration.
The results presented here should be compared with the results of Jaffe (25) and Rae-Venter and Dao (24), who have used the same technique to characterize the Triton X-100solubilized lactogenic receptor from rat liver. They obtained sedimentation coefficients, partial specific volumes, and Stokes radii of 5.0 and 4.7 s, 0.791 and 0.766 ml/g, and 54 and 50 A, respectively. The molecular weights of the lactogenic receptor calculated from these data were 78,000 and 73,000.
In conclusion, the present study has shown that the Triton X-100-solubilized female rat liver prolactin receptor has a molecular weight of about 90,000. This species contains M, 40,000 and M, 35,000 hormone-binding subunits. It cannot be decided whether these subunits are combined with each other or with undetected subunit(s) to make up the M , 90,000 species, or whether each represents a proteolytic fragment of the M , 90,000 species. Results are presented which indicate that some of the M , 40,000 and 35,000 subunits are (also) disulfide-linked within the M, 90,000 species. Clearly, further studies including purification are needed to clarify the structure of the prolactin receptor.