Purification and Protein Sequence Analysis of Rat Liver Prolactin Receptor*

Prolactin receptors were purified from rat liver membranes by single-step immunoaffinity chromatography using a specific monoclonal antibody to the rat liver prolactin receptor. Scatchard analysis of human growth hormone binding to the purified receptor revealed two classes of specific binding sites with KO = 18.5 and 1.2 classes of binding sites are responsible for high affinity prolactin binding, the partially purified receptor preparation had a binding activity of 1.69 nmol/ protein, representing 1000-fold purification over microsomal receptors with a recovery of 52%. three separate purifications, 6 mg of partially purified prolactin receptor were obtained with a purity of -4 the use of monoclonal for affinity chromatography resulted in a large improvement of prolactin receptor purification compared to chromatography (300-fold 15% The purified receptor was on a of was by from corresponding Immunoblot analysis using a radiolabeled monoclonal two closely located bands of M, 42,000 and

labeling with l2'1-human growth hormone. Cross-linking of microsomes revealed a single band for the hormone-receptor complex with M, 62,000. On the other hand, cross-linking of Triton X-100-solubilized or partially purified receptor with labeled hormone resulted in the appearance of two bands with M. 62,000 and 102,000, suggesting the existence of a subunit structure of the prolactin receptor, or alternatively, the existence of two types of prolactin receptor.

lz'I-
Receptor was run on an sodium dodecyl sulfate gel, and a homogeneous radioactive preparation was obtained from gel slices of M, 42,000. When this preparation was stored for 3 weeks at -20 "C and analyzed on sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by autoradiography, a larger molecular weight form (M, 84,000) was identified in addition to the original M, 42,000 band, suggesting that dimerization of the prolactin receptor occurred. This also indicates that the larger hormone-receptor complex observed in cross-linking studies could represent a dimer of the M, 42,000 subunit, only one of which is able to bind prolactin.
Prolactin, a member of the lactogenic hormone family, is present in all vertebrates and is responsible for over 85 biological actions (1). The action of PRL' is mediated by a specific receptor located in membrane components and widely distributed in a number of tissues (2). In order to better understand the molecular mechanisms involved in the various biological actions induced by PRL, elucidation of the biochemical properties of the PRL receptor is of primary importance. The relative molecular mass (Mr) of lactogen binding subunits has been reported in the range of 32,000 -84,000 when analyzed on SDS-PAGE (3). A smaller form of the PRL receptor (Mr 40,000), which is not linked by disulfide bonds to itself or to other subunits, has been shown in rabbit mammary gland and rat liver (4)(5)(6)(7)(8). On the other hand, larger molecular weight forms of the receptor also have been reported in rat ovary (9), liver (lo), and rabbit mammary gland (ll), and one group (9) has suggested the existence of disulfide linkages between the subunits.

Prolactin Receptor Purification and
Sequence Analysis 5905 mammary glands using hGH-agarose chromatography. Using oPRL as an affinity ligand, receptors have been purified from mouse liver (13) and rabbit mammary gland (5,14). Moreover, Mitani and Dufau (9) reported purification from rat ovary using a two-step affinity purification: concanavalin A-Sepharose and hGH-agarose affinity chromatography. Although a high degree of purification has been reported in several tissues (9,14), many of the purification studies were at the analytical rather than preparative level. Rat liver, an organ rich in PRL receptor, has the highest receptor level during pregnancy (15) or after estradiol treatment (16). Purification of the rat liver PRL receptor using oPRL affinity chromatography has been reported (6), resulting in a 300-500-fold purification over microsomes, but with a relatively low recovery. To obtain a quantity of PRL receptor sufficient to permit further structural studies, we undertook the purification of rat liver PRL receptors by immunoaffinity chromatography, which resulted in a high yield of receptor protein. A homogeneous PRL receptor obtained by electroelution was digested with trypsin, and the resulting peptide fragments were analyzed for their amino acid sequences. Such partial amino acid sequence data enabled us to prepare oligonucleotide probes and identify cDNA clones for the rat liver receptor (17). In this report, we also describe the characterization of the PRL receptor using immunoblot and affinity labeling techniques.

EXPERIMENTAL PROCEDURES
Materials-oPRL (NIADDK-oPRL-16 30.5 IU/mg) and hGH (NIH hGH AFP-5180A 2.2 IU/mg) were kindly supplied by the National Hormone and Pituitary Program. Bromocriptine (CB-154) was generously provided by Sandoz, Basel, Switzerland. Triton X-100, CHAPS, BSA, phenylmethylsulfonyl fluoride, dithiothreitol, and TPCK-trypsin were purchased from Sigma. Sepharose and Sephadex were from Pharmacia LKB Biotechnology Inc. Affi-Gel 10, electrophoresis reagents, and protein molecular weight standards were from Bio-Rad. Prestained molecular weight standards were from Bethesda Research Laboratories. DSS was purchased from Pierce and '"I-Na, from ICN Biochemicals. Reagents for sequence analysis were from Applied Biosystems. All other reagents used were of analytical grade.
Purification of Prolactin Receptor-Livers from adult female Sprague-Dawley rats were used as the source of prolactin receptor. The animals were treated subcutaneously with estradiol valerate (0.5 mg/ rat) 9 days prior to killing and with bromocriptine (0.5 mg/rat/ injection) for 12-h intervals beginning 24 h prior to killing. Crude membrane fractions (microsomes) were prepared from liver as previously described (6). Microsomes were centrifuged at 200,000 X g in a 50.38 Sorvall rotor for 60 min at 4 "C, and pellets were resuspended in 25 mM Tris-HC1, pH 7.4, containing 1 mM phenylmethylsulfonyl fluoride up to a final protein concentration of 12 mg/ml. This preparation was solubilized with 1% (v/v) Triton X-100 by constant stirring for 30 min at room temperature, centrifuged at 200,000 X g for 60 min at 4 "C, and the clear supernatant fraction was collected (solubilized receptor).
A specific monoclonal antibody to the rat liver prolactin receptor, E21 (6), was purified using Protein A Affi-Gel (18), and 40 mg of E21 IgG was coupled to 25 ml of Affi-Gel 10 according to the manufacturer's description with a coupling efficiency of 90-100%. The Tritonsolubilized receptor (1500 mg of protein) was initially passed over a 25-ml benzamidine-Sepharose 6B precolumn in order to absorb serine proteases from the preparation and then applied to EP1-Affi-Gel 10 in a 3 X 30-cm column at a flow rate of 1 bed volume/h at 4 "C. The E21-Affi-Gel 10 column was washed with 5 bed volumes of 25 mM Tris-HCI, pH 7.4, containing 1 mM CHAPS (column buffer). After equilibrating the column to room temperature, the receptor was eluted with 1 bed volume of 5 M MgCl' including 10 mM CHAPS, followed by 4 bed volumes of column buffer. Twelve fractions (10 ml/fraction) were collected from the beginning of the elution, and the active fractions were pooled (50 ml) and applied to a Sephadex G-25 column (300 ml) previously equilibrated with column buffer including 15% glycerol to avoid receptor degradation (19). Active receptor fractions eluted in the void volume were combined and frozen until further use. Hormone affinity purification of the PRL receptor using oPRL as an affinity ligand was performed as described earlier (6).
Receptor Binding Assay-Microsomes or solubilized receptor (100 or 40 pg protein/tube, respectively) were incubated with various concentrations of unlabeled hGH diluted in 25 mM Tris-HC1, pH 7.4, containing 10 mM MgCIz and 0.1% BSA (assay buffer) in the presence of lZ5I-hGH (-35,000 cpm/tube) in a total volume of 500 pl for 17 h at room temperature. Partially purified receptor (0.1 pg/tube) was incubated as above, except that assay buffer including 4 mM CHAPS (final concentration) was used. The assays were terminated by the addition of 3 ml of assay buffer in the case of microsomes or by adding 0.5 ml of 0.1% bovine 7-globulin and 1 ml of 24% PEG 8000 and mixing in the case of solubilized or purified preparations. Bound lz5I-hGH was separated from free ligand by centrifugation at 3000 X g for 20 min at 4 "C, and pellets were counted for 1 min in a LKB 7counter with a counting efficiency of 60%. Binding data were analyzed using the LIGAND computer program (20). Binding assays for hormone affinity purification were done as above except that lZ5I-oPRL was used as a radioactive ligand, and 4 mM CHAPS was included to assay buffer for CHAPS-solubilized receptor as well as purified receptor.
Protein Determination-Protein concentrations of microsomes, solubilized receptor, and E21 IgG were determined according to the method of Lowry et al. (21) or the modified method (22) for the partially purified receptor, using BSA as standard.
lodination-oPRL, hGH, partially purified receptor, and E21 IgG were iodinated by the chloramine-T method (23). For labeling hormones, a concentration of 500 ng of chloramine T in a total reaction volume of 50 pl was used with 5 pg of protein and 750 pCi of Na'"I. For iodination of E21 IgG, 30 pg of IgG protein, 500 pCi of Na'"I, and 500 ng of chloramine T were used. Partially purified receptor (5 pg) was iodinated using 1000 ng of chloramine T in 50 pl volume, and 500 pCi of '"I-Na. Free iodine was separated by gel filtration (Sephadex G-75 for hormones, Sephadex G-25 for receptor, and Sepharose 6B for IgG). The columns were equilibrated and eluted with assay buffer containing 0.1% NaN3 for hormones and IgG, or 25 mM Tris-HCl, pH 7.4, containing 0.1% (v/v) Triton X-100, 0.1% BSA, and 0.1% NaN3 for purified receptor. Specific activities, which were calculated from recovery of radioactivity in the protein fraction, were 76-95 pCi/pg for oPRL, 90-105 pCi/pg for hGH, 4-14 pCi/pg for E21 IgG, and 43-50 pCi/pg for partially purified receptor.
Electroelution-Partially purified receptor (6 mg), extensively dialyzed against 2.5 mM Tris-HC1, pH 7.4, containing 0.1 mM CHAPS, was concentrated by lyophilization. After solubilization in 50 mM Tris-HC1, pH 8.1, 6 M guanidine hydrochloride, and 2 mM EDTA, a trace amount (300,000 cpm) of ''51-re~eptor was added, and the preparation was reduced and alkylated as described previously (24). Excess reagents were removed by dialysis against 1% acetic acid and the retentate lyophilized. The solubilized retentate was run on preparative SDS-PAGE. Bands ranging from M, 38,000 to 43,000 were cut from the gels according to the position of prestained standards, and the receptor protein was electroeluted from the gels using an Elutrap (Schleicher & Schuell) in 50 mM Tris-acetate, pH 8.0, containing 0.01% Triton X-100 for 6 h at 4 "C at 150 V. Eluted samples were pooled (6 ml) and concentrated on a Centricon-10 filter (Amicon) to a final volume of 400 pl and again concentrated to 50 pl in a Speed Vac (Savant) concentration. To assess purity and quantity, 1 pl of sample was run on a 10-15% SDS-PAGE Phast system (Pharmacia) and stained with silver according to the manufacturer's instructions. The intensity of silver staining obtained for the receptor was compared to that of various dilutions of standard proteins (Pharmacia low molecular weight kit).
Immunoblot Analysis of PRL Receptors-Microsomes, hormone affinity-purified receptor, immunoaffinity-purified receptor, and electroeluted purified receptor were analyzed by immunoblot technique using a specific monoclonal antibody' to the rat liver PRL receptor. Proteins were separated on SDS-PAGE and electroblotted onto nitrocellulose membranes at 4 "C overnight at 0.15 A and for l h at l A in 10 mM NaHC03, 3 mM Na2C03, pH 9.9, including 20% methanol (25). The blots were blocked with 1% ethanolamine, pH 7.4, containing 2% dry fat milk for 6 h at room temperature, and then incubated with '**I-labeled monoclonal antibody (lo7 cpm/50 ml in 10 mM phosphate-buffered saline, pH 7.4, with 1% BSA) overnight at room temperature. Membranes were washed extensively with phosphatebuffered saline, and subjected to autoradiography. *A monoclonal antibody to the rat liver PRL receptor, designated as U5. E21 was not used because this antibody failed to recognize reduced PRL receptors (6).
Protein Sequencing-Internal sequences of the prolactin receptor were obtained by gas-phase sequence analysis of peptides generated from digestion of the electroeluted receptor with TPCK-trypsin. The appropriate conditions of digestion were determined from the results of preliminary experiments in which electrophoretically purified lZ5Ilabeled receptor was subjected to various conditions of enzymatic digestion and analyzed on 5-20% SDS-PAGE followed by autoradiography (data not shown). Approximately 100 pg (estimated from silver-stained gel) of pure receptor was digested overnight at 37 "C with 1.6 pg of TPCK-trypsin in 25 mM Tris-HC1, pH 7.4.
The tryptic peptides were separated by reverse-phase HPLC on a Hypersil ODS 5 pm (100 X 2.1 mm) column using 0.1% ammonium bicarbonate and a gradient of acetonitrile. Chromatography was performed with a Hewlett Packard System Model 1090 instrument, equipped with a diode array detector. Fractions were collected by hand, and when necessary, further purified on the same HPLC system with a shallower gradient.
Automated Edman degradations were performed on a Model 470A Gas-Phase Sequencer equipped with an on-line Model 120A phenylthiohydantoin (PTH) analyzer (Applied Biosystems) employing the general protocol of Hewick et al. (26). Samples were applied to precycled filters, coated with 1.5 mg of Polybrene plus 0.1 mg of NaCl (Biobrene Plus), and standard programs (03 RPRE and 03 RPTH, Applied Biosystems) were employed for precycling and sequencing.
Characterization of PRL Receptor-Cross-linking studies were performed with microsomal, Triton X-100-solubilized, and immunoaffinity-purified receptors. The receptor preparations were bound with lZ5I-hGH under conditions of the "binding assay," and 50 pl of freshly prepared DSS (5.5 mM) in dimethyl sulfoxide was added to the samples to make the final concentration 0.5 mM. After 15 min on ice, the reaction was terminated by adding 0.15 volume of ice-cold 1 M Tris-HC1, pH 7.4. Aliquots of 50 pl each sample were subjected to SDS-PAGE followed by autoradiography.
Partially purified receptors were iodinated and subjected to preparative SDS-PAGE. A homogeneous radioactive sample was obtained by electroelution of the gel slices ranging from M, 38,000 to 43,000. The sample was reanalyzed on SDS-PAGE followed by autoradiography immediately or after 3 weeks of storage at -20 "C.

RESULTS
Purification-The rat liver PRL receptor was purified by Triton X-100 solubilization of microsomes followed by immunoaffinity chromatography. Human GH competitive assays were performed on each receptor preparation. The results of a representative purification are summarized in Table I. Scatchard analysis (Fig. 1) of binding results indicated one class of specific binding sites with K, = 2.3 nM" for microsomes, whereas solubilized and partially purified receptor had two classes of specific binding sites. The high affinity sites for solubilized receptor and partially purified receptor (Ka = 7.5 and 18.5 nM-l, respectively), were about 15-20-fold higher than for the low affinity sites (KO = 0.34 and 1.23 n"' ).
For subsequent calculations, both high and low affinity binding sites were considered to contribute to prolactin receptor binding activity.
Triton X-100 has been shown to cause aggregation of oPRL (12). Therefore, the Triton X-100 in the solubilized-purified fraction was exchanged with CHAPS, which does not affect the PRL molecule in the same way during the chromatographic steps (13). Binding data for both 1251-oPRL and '"I-hGH to the purified receptor were similar. Solubilization of the microsomal fraction with Triton X-100 resulted in 4-fold increase in binding capacity, as has been previously reported (4,6,8, E ) , probably due to exposure of cryptic sites. Almost 10,000 pmol of solubilized receptors were loaded onto a 40mg E21 coupled Affi-Gel 10 column, and 40% of receptor proteins was recovered in the 5 M MgC12 eluate. The binding capacity of purified receptor was increased to 1700 pmol/mg protein from 1.6 pmol/mgprotein of microsomes, representing -1000-fold purification over microsomes. Assuming that the molecular weight of the rat PRL receptor is 40,000-42,000 and that there is one binding site per receptor molecule, the purity of this preparation was estimated to be 6.5%. Consequently, 3.5 nmol of receptor were obtained from 165 g of rat liver by the present purification. Two other purifications were carried out, resulting in receptor preparations ranging in purity from 4 to 6%.
To assess the improvement of purification of the PRL receptor, a hormone affinity purification profile is shown in Table 11. A distinct difference between two purifications can be seen in the value of fold purification and binding capacity recovered. In the various separate purifications, the recovery of the PRL receptor from the affinity column was 2-4-fold greater for immunoaffinity than hormone affinity. The purity in final preparation also improved more than %fold by immunoaffinity chromatography. It should be noted that we also employed a benzamidine-Sepharose precolumn to remove serum proteases, thus the combined effect of removing proteases along with using a monoclonal antibody as a ligand for affinity chromatography results in a marked improvement in receptor purification and yield.
Electroelution-The affinity-purified receptors were reduced and alkylated and run on preparative SDS-PAGE. Values are calculated by combining both high and low affinity binding sites in solubilized and purified receptors. Binding capacity and affinity constants were determined from Scatchard analysis of lZ51-hGH competition experiments (Fig. 1).

Values are calculated by combining both high and low affinity binding sites in solubilized and purified receptors.
Homogeneous prolactin receptor was prepared by electroeluting the M , -38,000-43,000 region from the gel slices. During electroelution, SDS was omitted from the elution buffer, and a minimum concentration of Triton X-100 (0.01%) was used to facilitate the following sequencing steps. Receptor recovery after each step was estimated from the radioactivity in the preparations. The band of M, -38,000-43,000 contained -2% of the total radioactive counts applied on SDS-PAGE. After electroelution, 85-95% of radioactivity was recovered from the gel slices in a small volume. Approximately 100 Fg of homogeneous receptor was obtained from 6 mg of affinity purified receptors in a total volume of 6 ml. This value is in agreement with the amount of peptide detected by sequence analysis (Table 111), assuming a coupling efficiency of 60-80%. It was also similar to the estimate obtained by silver staining of the minigel (Phast system) (not shown).
Immunoblot Analysis of the PRL Receptor-By direct overlay of radiolabeled monoclonal antibody onto nitrocellulose followed by autoradiography, the detection sensitivity increased 50-fold as compared to the enzyme-substrate method. Thus, it became possible for microsomal prolactin receptor to be detected by immunoblot analysis. Analysis was performed on microsomal, hormone affinity-purified, immunoaffinitypurified, and electroeluted PRL receptor, and the autoradiogram revealed two separate but closely related bands of 42,000 and 40,000 for all preparations each with almost equal intensity (Fig. 2). Vastly different amounts of protein were loaded in each lane, which clearly reflects the degree of purity. Using this technique, it was calculated that there was 300fold purification between microsomal and hormone affinitypurified receptor, 1250-fold between microsomal and immunoaffinity purification, and 50,000-fold between microsomal and electroeluted receptor. These values are almost identical to those obtained from Scatchard analysis of each preparation (Tables I and 11). Although the relationship between two bands (M, 42,000 and 40,000) is still unclear, it has been suggested that isoforms of the PRL receptor may exist, and that the lower molecular weight form could be derived from the higher molecular form (6).
Protein Sequence Analysis-The pure prolactin receptor was extensively digested with TPCK-trypsin, and the resulting peptide fragments were separated from each other by Peptide sequences of the prolactin receptor Oligonucleotides for screening cDNA libraries were synthesized based on the underlined protein sequence data. The sequence of the oligonucleotide probes were: a 24-mixmer for the peak 13 sequence ( 5 ' -

C C A G T A G C C G T G G T C Y G C A -3 ' ) ,
A two probes that were prepared from the peak 21-5 sequence, a 45mer best predicted sequence (5'-CCTGAACTGGGTCTGGTGGCCGGTGAAGTGGATCTCCACTCCTC-3'), and a 23-mixmer (5'-GT AA TG AT TCCCACTCCTC-3'). Attempts to identify a positive recombinant were negative with the 24-mixmer, perhaps because for the Pro and Gly residues the codons that were selected (only two of the four possibilities were chosen) were incorrect; also, the failure with the 45-mer was certainly due to the fact that the selection of bases (based on a codon frequency table for the rat) was incorrect for 10 of the 14 choices.

93-124
PTH-carboxymethylcysteine (PTH-CMC) elutes about 0.1 min before PTH-glutamine (PTH-Q), however further evidence distinguishing Q from C in the sequence data is the coincident appearance of PTH-glutamic acid (PTH-E) due to deamidation of glutamine during sequencing. e PTH-aspartic acid (PTH-D) is masked by the PTH-ammonia (PTH-NH3) adduct under the PTH analysis system; however, assignment of D in the sequence was confirmed from the cDNA sequence.
Double sequence observed. Sequence data consistent with the two peptides within the predicted cDNA sequence. e Weak signal observed for tryptophan (W) residues probably due to oxidation during the electrophoretic purification steps. 'A blank cycle two positions before a serine (S) or threonine (T) residue is consistent with N-linked glycosylation of an asparagine (N) residue. This result is consistent with the predicted protein sequence deduced from the cDNA sequence.
reverse-phase HPLC (Fig. 3). NHn-terminal sequence analysis of several peaks shown in Fig. 3 provided essential information both for cloning the cDNA and for confirming the reading frame of the sequence data (17). The two blank cycles appearing two positions prior to serine (S) or threonine (T) residues were considered to be N-linked glycosylation sites of asparagine (N) residues (Table 111), which are consistent with the cDNA sequence data (Am-54 and Asn-99). The third potential glycosylation site (Asn-127) was not sequenced, therefore its glycosylation state has not been confirmed.
Peaks 21 and 25 in Fig. 3 were shown by protein sequence analysis to be mixtures containing more than one peptide, and in each case a major sequence could be derived from the data (Table 111). The observed sequences for the two peptides in the double-sequence cycles were found to be consistent with protein sequences deduced from the cDNA. One of these mixtures, peak 21, which resolved into a single sequence after six degradations, proved to be instrumental in cloning the prolactin cDNA (17). Slightly over 50% of the entire amino acid sequence of the prolactin receptor, as deduced from the cDNA sequence, was accurately determined by peptide sequence analysis.
Characterization of the PRL Receptor-Rat liver microsomes, Triton X-100-solubilized receptors, and partially purified receptors were affinity-labeled with lZ5I-hGH by DSS and analyzed on SDS-PAGE followed by autoradiography. Cross-linking of microsomes with lZ5I-hGH revealed a single M , 62,000 band on the autoradiogram (Fig. 4, lane 1 ). On the other hand, cross-linking of solubilized or purified receptors resulted in the appearance of an M, 102,000 band in addition to the M, 62,000 (Fig. 3, lanes 2 and 3). Both bands disappeared when excess unlabeled hormone was added to the incubation mixture, indicating specific labeling of the PRL binding subunit. By subtracting the molecular weight of hGH (22,000), molecular weights of 80,000 and 40,000 were estimated for the higher and the lower hormone binding species, respectively. Migration and density of the M, 102,000 and 62,000 bands were not affected by the use of reducing agents (not shown). Larger forms of the PRL receptor have been reported in rat ovary (9) and more recently rat liver (10) when detergent-solubilized receptors were analyzed by cross-linking. Although our results are in good agreement with these, we failed to observe any significant change in bands between reduced and nonreduced samples.
Radiolabeled receptor was reduced with 20 mM dithiothreitol and subjected to preparative SDS-PAGE. The homogeneous sample was electroeluted from gel slices ranging between M, 38,000 to 43,000 and analyzed on SDS-PAGE the following day (Fig. 5, lane 1). When this sample was stored at -20 "C for 3 weeks and then rerun on SDS-PAGE, a larger molecular weight form, M , 84,000 was identified on the autoradiogram in addition to the original M, 42,000 form (Fig. 5,  lane 2 ) . This clearly demonstrates that the M, 42,000 binding subunit dimerized into a larger form.

DISCUSSION
In order to obtain preparative amounts of receptor, it is necessary to begin purification with a tissue containing high column. Elution solvents were 0.1% am-r; monium bicarbonate (solvent A) and 100% acetonitrile (solvent B). Peptide fragments were eluted with a linear gradient of 5-55% solvent B in 60 min with a flow rate of 0.1 ml/min. The absorbance was measured at 210 nm. Arrows show the peptide peaks successfully sequenced (Table 111). 200 0 quantities and its isolation requires the proper detergent and highly specific methodologies. Affinity chromatography is the method of choice, since it bases the purification on the functional rather than on overall chemical and physical properties, which are similar among most detergent-solubilized membrane proteins. By using PRL affinity chromatography, we previously purified the PRL receptor from rat liver (6). Partially purified receptor preparations obtained by hormone affinity chromatography were useful for immunization and antibody production and for characterization of the PRL receptor molecule, while neither purity nor final recovery of receptor were satisfactory for protein sequencing.
Similar results using hormone affinity chromatography have been reported for the purification of rabbit (4, 12) and pig mammary gland (8) receptor. A possible explanation for the low recovery could be receptor degradation during the various purification steps and/or a strong interaction between receptor and ligand that does not permit separation by the dissociating agent. Recently, Berthon et al. (29) demonstrated an improvement in purification by immunoaffinity chromatography of pig mammary gland receptor.
Immunoaffinity chromatography employing a specific monoclonal antibody offers several advantages compared to hormone affinity chromatography. First, microsomes can be solubilized with Triton X-100, which is a more efficient detergent for solubilizing rat liver membranes, but is not compatible for use with oPRL affinity chromatography (12). Second, chromatography is performed at 4 "C rather than 20 "C, which should decrease degradation and/or loss of receptor binding activity. Finally, the binding characteristics between E21 and the prolactin receptor allow much larger amounts of receptor to be eluted from the affinity column by 5 M MgC12 without loss of specific binding activity in the eluate. Almost 40% of the receptor applied to the column was eluted, resulting in a 50% final receptor recovery over microsomes. Consequently 9 nmol of the PRL receptors were obtained from three purifications. Immunoaffinity chromatography was also superior to hormone affinity in terms of the purity of the final preparation. Assuming 1:l binding of PRL with the receptor, the purity was calculated from Scatchard analysis to be 4-6.5% and 0.5-1.5% for immunoaffinity-and hormone affinity-purified preparations, respectively. This estimation is supported by the results of immunoblot analysis (Fig. 2). However, the purity was still relatively low when compared to the purifications of other membrane receptors, probably due to nonspecific binding of proteins other than the PRL receptor protein to the chromatography matrix and spacer arms, or the degradation or aggregation of the receptor proteins during purification steps that resulted in a loss of binding activity. Two attempts to obtain NH2-terminal sequence data using intact PRL receptor were unsuccessful, presumably because the NH2 terminus of this protein is blocked. This agrees with the predicted NH2-terminal amino acid deduced from the cDNA sequence, which suggests that Gln may be present as a pyro-Glu residue. By microsequence analysis, two of the three potential N-linked glycosylation sites were confirmed in the receptor molecule. The biological significance of the oligosaccharide residues on the PRL receptor is far from clear. We are currently characterizing the type of oligosaccharide linkage found in the rat liver PRL receptor?
A larger molecular weight (Mr 80,000) binding species was found in addition to the M, 42,000 binding subunit when detergent-solubilized or partially purified receptors were analyzed by cross-linking (9, 10). Whether this larger form represents a different type of PRL receptor or is formed by association of the M, 42,000 subunits with each other or with another membrane component is still unclear. However, the appearance of the M, 84,000 form in the homogeneous receptor preparation (Fig. 5) suggests that the M , 42,000 form is able to dimerize. Moreover, when membranes from Chinese hamster ovary cells transfected with the cDNA clone encoding the M , 42,000 PRL receptor (17) were subjected to immunoblot analysis using several monoclonal antibodies to the PRL receptor, both the M , 42,000 and 84,000 form were seen (data not shown). Thus, the larger binding form in the cross-linking study (Fig. 4) probably represents a dimer of the smaller binding unit, only one of which is able to bind PRL. Interestingly, the larger form was not cross-linked with labeled hormone in the microsomal preparation. One possible explanation is that membrane lipids surrounding the receptor may interfere with cross-linking of lZ5I-hGH to the larger form. Receptor oligomerization has been reported for the epidermal growth factor receptor (30,31), the process being induced by binding of epidermal growth factor to its receptor, or can even occur in the absence of epidermal growth factor (31). The importance of microaggregation of receptors in the process of signal transmission has been demonstrated for epidermal growth factor (32), insulin (33), insulin-like growth factor-1 (34), and PRL (35) receptors. The present studies suggest that dimerization of the PRL receptor occurs naturally, and may be involved in the mechanism of action of prolactin.
In summary, preparative amounts of the PRL receptors were produced from rat liver by single-step immunoaffinity chromatography using a specific monoclonal antibody, which subsequently allowed us to determine internal amino acid sequences of the PRL receptor. Biochemical characterization indicates that the molecular weight of the rat liver PRL receptor is 42,000, but suggests that the receptor also appears naturally as a dimer within the membrane environment.