INTRODUCTION

PRL¹ plays a vital role in maintaining the function of the corpus luteum. It enhances luteinizing hormone receptor activity in the pregnant rat corpus luteum (1) by inducing luteinizing hormone receptor mRNA levels (2). It also sustains elevated levels of estrogen receptors in the corpus luteum (3, 4, 5). In addition, PRL stimulates luteal protein synthesis by causing the dephosphorylation of elongation factor 2 (6) and thus appears to be responsible for large luteal cell hypertrophy. Moreover, PRL sustains the high level of progesterone by inhibiting the expression of 20-hydroxysteroid dehydrogenase, thereby reducing the transformation of progesterone into inactive 20-hydroxyprogesterone (7).

Although the effects of PRL on the corpus luteum are well documented, the mechanism through which the PRL ligand-receptor interaction is transduced in the ovary remains completely uncharacterized. Unlike the mouse, in which there are multiple forms of the PRL receptors, three short forms and one long form (8, 9), in the rat, two types PRL receptors have been identified: the short form and long form. They share identical extracellular and transmembrane domains but differ in the length and sequence of their cytoplasmic domains (10, 11, 12). In addition, an intermediate form of the PRL receptor specific to Nb2 cells has been cloned and shown to be a mutant of the long form of the rat PRL receptor (13). Recently, both the long form and short form of the PRL receptor were found to be expressed in the whole rat ovary (11, 14, 15) and specifically in the rat corpus luteum (14).² It is unknown whether either or both of these receptor forms can transduce PRL signal in the corpus luteum. Lacking enzymatic activity, the PRL receptor ultimately relies on interactions with receptor-associated cellular proteins for the generation of second messengers (17, 18, 19, 20, 21, 22). Extensive studies carried out on Nb2 cells and mammary gland cells have shown that the long form and the intermediate form are fully capable of transducing a signal through an association with protein kinases (23, 24, 25, 26). As a result of PRL binding to the long form, a tyrosine kinase (JAK2) is activated, which subsequently phosphorylates a transcription factor (Stat5). Upon activation by phosphorylation, Stat5 binds to the PRL response element of an appropriate target gene, such as the -casein gene promoter, and induces -casein gene transcription (26, 27, 28, 29). Little information is available on the role of the short form of PRL receptor in the PRL signal transduction pathway, although a recent report indicates that one of the three short forms of the PRL receptor in the mouse can induce MAP kinase phosphorylation and cell proliferation in response to ligand binding (30).

To date, no proteins capable of binding to the cytoplasmic domain of the PRL receptor have been identified in the corpus luteum. Such proteins would presumably play an important role in the response of the corpus luteum to PRL during pregnancy. We have previously demonstrated that PRL and estradiol regulate the levels of a 32-kDa luteal phosphoprotein (31, 32). In the present study, we describe the cloning of this 32-kDa protein and show that this protein interacts with the short form of PRL receptor in the rat corpus luteum. We now name this protein PRAP for RL eceptor ssociated rotein.

EXPERIMENTAL PROCEDURES

Glutathione-agarose beads, glutathione, and mouse IgG₁ were from Sigma. Protein G-Sepharose 4 Fast Flow bead suspension was purchased from Pharmacia Biotech Inc. Mouse monoclonal antibody to rat PRL receptor (U6) was kindly provided by Dr. Paul A. Kelly (Nekar Hospital, Paris, France). Mouse monoclonal anti-phosphotyrosine was obtained from Upstate Biotechnology Inc. (Lake Placid, NY). Products for ECL (enhanced chemiluminescence) were from Amersham.

Corpora lutea from day 15 pregnant rats were used as a source of PRAP. Microsomal fractions from the corpora lutea were obtained by differential centrifugation (31). Protein samples for sequencing were purified by SDS-polyacrylamide gel electrophoresis (33). Separated proteins were transferred onto polyvinylidine difluoride membranes and stained with Ponceau S, and the 32-kDa band was excised. Protein sequencing was performed by The Protein Sequencing/Synthesis Laboratory at The University of Illinois at Chicago.

An oligo(dT)-primed rat cDNA expression library made from the mRNA of corpora lutea of day 15 pregnant rats was constructed in ZAP II (Stratagene). The peptide sequence MRKVVLIT at the N terminus of PRAP was used to design a degenerate 24-mer oligonucleotide. This oligonucleotide was labeled at the 5 end with [-³²P]ATP (Amersham) and used as a probe. The cDNA library was screened following standard methods (34). pBluescript SK() containing the insert was excised in vivo by ExAssist/SOLR system (Stratagene). Two of the 38 positive clones were selected for DNA sequencing using the dideoxy chain termination method (35) with a Sequenase Version 2.0 DNA-sequencing kit (U. S. Biochemical Corp.). Both strands of cDNA of the two clones were completely sequenced. The amino acid sequence predicted from the isolated cDNA was compared with sequences in PDB (Brookhaven Protein Data Bank), SwissProt (SWISS-PROT, release 31.0), PIR (Protein Identification Resources, release 45.0), and GenPept (CDS translations from GenBank, release 90.0) data bases, using the BLAST and FASTA software at the Research Resources Center of The University of Illinois at Chicago.

Amplification of 3 End of mRNA

Rapid amplification of cDNA ends (RACE) was done essentially as described by Frohman et al. (36). The first strand cDNA was synthesized using 1 µg of dT₁₆-adapter primer supplied with the 3-RACE System (Life Technologies, Inc./BRL, Life Technologies) and 1 µg of total RNA from the corpus luteum of day 15 pregnant rats as template. Polymerase chain reaction amplification was done using Universal Amplification Primer (UAP) from 3-RACE System and gene specific primer 1 (GSP1) (AGATCACCACAGCTGACA) and Oligo 3 primer (CTACAAGACCTTACT) derived from the 3 end sequence common to both clones. The products were analyzed on an agarose gel, and amplifications were repeated by extracting a 5-µl plug of agarose from the original amplification and adding it to 50 µl of a complete amplification reaction containing the gene specific primer 2 proceeded by four repeats of CUA (CUACUACUACUAACGTTCCCCAGAGCA), the sequence downstream of GSP1. After amplification, 3-RACE products was directly cloned into a vector using CLONEAMP pAMP10 system (Life Technologies, Inc./BRL, Life Technologies) (37) and transformed into DH5 competent cells. The cDNA was sequenced from both strands as described above.

Corpora lutea were obtained from day 13-15 intact pregnant rats and dissected from adhering ovarian follicles and interstitial tissues. Corpora lutea were kept in serum-free media and divided into four groups for treatment with PRL (NIDDK oPRL-18, 1 µg/ml) for 0, 20, 40, or 60 min. Corpora lutea were then homogenized in the homogenizing buffer (31) with or without 250 mM Na₃VO₄.

The cDNAs encoding the cytoplasmic domains of the long and short forms of the PRL receptor were inserted into the pGEX3X vector (8, 38) in-frame with the upstream sequences encoding glutathione S-transferase (GST). Expression of PRL-R fusion proteins was induced in Escherichia coli strain DH5 cells transformed with these constructs by addition of isopropyl-1-thio--D-galactopyranoside. To purify the fusion proteins, cells were harvested by centrifugation and suspended in 12.5 ml of 25% sucrose, 50 mM Tris, pH 8.0. Cells were lysed by treatment with 1 mg/ml lysozyme for 5 min, followed by addition of -mercaptoethanol to 20 mM, EDTA to 250 mM, and phenylmethylsulfonyl fluoride to 5 mM and agitation for 15 min at 4 °C. Samples were adjusted to 100 mM KCl, 0.2% Triton X-100, incubated for an additional 15 min at 4 °C, and centrifuged at 10,000 × g for 30 min at 4 °C. The supernatant was collected and the fusion proteins were recovered by binding to glutathione-agarose beads. The agarose beads were collected by centrifugation (1,000 × g) after a 60-min incubation at room temperature or overnight binding at 4 °C, and washed extensively with phosphate-buffered saline. The GST-PRL receptor-agarose beads were resuspended in an equal volume of phosphate-buffered saline. Fusion proteins were eluted with 50 mM free glutathione in 50 mM Tris, pH 8.0, and analyzed on SDS-PAGE (38).

Luteal protein extracts were incubated with glutathione-agarose beads for 60 min at room temperature to remove proteins that bound nonspecifically. Samples were centrifuged at 1,000 × g, and the supernatant was recovered and adjusted to 1% CHAPS. Fusion proteins containing the long (PRL-R_L) or short (PRL-R_S) form of the PRL receptor were isolated on glutathione-agarose beads, and the beads were then incubated overnight at 4 °C with 200 µg of luteal proteins. Following incubation, the beads were collected by centrifugation and washed in ice-cold phosphate-buffered saline. GST or PRL-R fusion proteins and their attached proteins were then eluted with either SDS-PAGE sample buffer or 500 µl of 50 mM glutathione with continuous mixing for 20 min at room temperature. Eluted proteins were precipitated in 10% trichloroacetic acid and analyzed by SDS-PAGE. After electrophoresis, proteins were transferred to nitrocellulose and stained. Immunoblotting was performed with PRAP antiserum at a 1:1000 dilution. Immunoreactive proteins were detected using the ECL system.

Luteal homogenates were solubilized in 1% Nonidet P-40 and 0.25% deoxycholate. 300 µg of solubilized luteal proteins were incubated overnight at 4 °C with shaking with PRL receptor monoclonal antibody or with nonspecific mouse monoclonal antibody of subtype IgG₁ at a concentration of 10 µg/ml. Polyclonal PRL receptor antiserum (9) and an anti-phosphotyrosine antibody were also used for immunoprecipitation. In these experiments, 300 µg of luteal proteins from corpora lutea treated with PRL were incubated overnight at 4 °C with polyclonal PRL receptor antiserum at a 1:500 dilution or with anti-phosphotyrosine antibody (10 µg/ml). 20 µl of protein G-Sepharose beads were added into the immune complexes and incubated at 4 °C or room temperature for 30 min. The beads were collected by centrifugation and washed extensively. Beads were then suspended in 15 µl of SDS-PAGE sample buffer and subjected to SDS-PAGE. Proteins were transferred to nitrocellulose membranes and immunoblotted with PRAP antiserum and polyclonal PRL receptor antiserum (1:500).

RESULTS

To explore the possibility that the PRL-inducible 32-kDa luteal phosphoprotein may be involved in the PRL signal transduction pathway, we first examined if this protein associates with the PRL receptor in the corpus luteum. Luteal cell extracts were immunoprecipitated with the U6 monoclonal antibody against the rat PRL receptor and then immunoblotted with the PRAP antiserum. As seen in Fig. 1, the 32-kDa protein was efficiently precipitated by the U6 antibody, but not by a control IgG. Since this result demonstrates that the 32-kDa protein is associated with the PRL receptor in luteal cell extracts, we now refer to this protein as PRL receptor associated protein, or PRAP. Parallel incubations of the luteal cell extracts with anti-phosphotyrosine antibody also resulted in the immunoprecipitation of PRAP (Fig. 1), consistent with PRAP being a substrate for PRL receptor-associated protein tyrosine kinases.

To determine if the association of PRAP with the PRL receptor is dependent upon PRL binding, corpora lutea obtained from day 13 pregnant rats were treated with PRL for 0, 20, 40, and 60 min. A polyclonal PRL receptor antiserum was utilized for this immunoprecipitation and was also able to co-precipitate PRAP (Fig. 2A) and the PRL receptor (Fig. 2B). Even though both PRL receptor forms were probably recovered by this procedure, only the long form was detected, presumably because of the greater abundance of that form in the ovary (11, 14, 15). The amount of PRAP associated with the receptor did not change in response to PRL treatment (Fig. 2A), indicating that PRAP is able to bind to the receptor in the basal state. Extracts from PRL-treated corpora lutea were also subjected to immunoprecipitation with the anti-phosphotyrosine antibody to determine if PRAP tyrosine phosphorylation is PRL-dependent. As shown in Fig. 3, PRL treatment did not affect the tyrosine phosphorylation status of this protein.

Since the corpus luteum contains both short and long PRL receptors, we next sought to determine which form of the receptor is capable of associating with PRAP. Prolactin receptor cytoplasmic domains were expressed in bacteria as GST fusion proteins. Luteal cell extracts from day 15 pregnant rats were applied to glutathione-Sepharose columns to which GST alone or GST linked to the cytoplasmic domain of either the long or short form of the PRL receptor was bound. Proteins retained on the column were recovered by addition of glutathione and then analyzed by SDS-PAGE and immunoblotting with the PRAP antiserum. Strikingly, PRAP was detected in association with the short form of the PRL receptor, but not with the long form of the PRL receptor nor GST (Fig. 4A). This experiment was repeated with fusion constructs containing either the whole cytoplasm domain of the long receptor form or just the unique cytoplasmic domain of the long receptor; PRAP failed to bind to either of these fusion proteins, but again was able to bind to the short receptor cytoplasmic domain (Fig. 4C). In the condition when GST and GST-PRLR were eluted with glutathione in the absence of luteal proteins, no signal was detected in any of the lanes except the last lane containing luteal proteins rich in PRAP (Fig. 4B).

The specific association of PRAP with the short PRL receptor is in marked contrast to the reported association of other factors with the long form of the receptor (18, 19, 20, 21, 22, 29, 39, 40). A more complete characterization of PRAP might therefore provide some insight into the role of the short receptor form in PRL action in the ovary. Purified PRAP was subjected to N-terminal protein sequencing, and a sequence for the first 18 amino acids was obtained (MRKVVLITGASSGIGLAL). A degenerate oligonucleotide based on the first eight amino acids was used to screen a rat corpus luteum cDNA library. Thirty-eight positive clones were isolated, and clones 13A5 and 14A1 were completely sequenced from both strands. The 1161-base pair sequence of clone 13A5 (Fig. 5) is identical to the 14A1 sequence except for an extra 94 base pairs at the 5 of clone 14A1 and differences in the 3-untranslated region.

Both clones contain a single identical open reading frame beginning with the 18 codons predicted from the N-terminal amino acid sequencing (Fig. 5, double underline), confirming that the isolated cDNA clone encodes PRAP. Two adjacent ATG triplets are found at the beginning of the open reading frame, but only one methionine was detected at the N terminus of the protein. Inspection of the two ATG codons reveals that the first ATG has a pyrimidine (T) at position 3, whereas the second ATG has a purine (A) at 3; thus, the second ATG is in a stronger context to act as a translation initiating codon (41). Furthermore, an in-frame stop codon (TGA) is present immediately 5 to the first ATG, thereby precluding the use of another, upstream methionine codon to initiate translation. A termination codon (TGA) is present at position 1017-1019, and thus the open reading frame encodes a protein of 334 amino acids with a calculated M_r of 37,372, slightly larger than the size of the protein detected by gel electrophoresis.

Based on the absence of a polyadenylation signal, the initial cDNA clones apparently lacked a complete 3-untranslated region. To obtain a cDNA clone from this region of the PRAP mRNA, 3-RACE was used with a (dT)₁₆-adapter primer for reverse transcription of total luteal mRNA from day 15 pregnant rats, and oligonucleotides from the 13A5 and 14A1 3-untranslated region and the antisense universal adapter primer as polymerase chain reaction primers. A second gene-specific primer was used for reamplification of the 3-RACE products, which were then cloned and sequenced (Fig. 5, with the gene-specific primer underlined). The sequences at the 5-end of the clone obtained by 3-RACE match the sequences at the 3-end of clone 13A5, but not clone 14A1. This analysis extended the 3-end of the 13A5 cDNA clone by 537 base pairs, a region that contains a putative polyadenylation signal AATTAAA (shown underlined in Fig. 5) and a residual poly(A) tail.

Comparison of the predicted amino acid sequence to the contents of protein data bases failed to identify any significant similarities to known proteins. However, sequence analysis revealed the presence of three potential N-glycosylation sites (positions 37, 178, and 229), two protein kinase C phosphorylation sites (170 and 195), and several casein kinase II phosphorylation sites (118, 125, 180, and 315) (Fig. 5). A PROSITE data base search did not indicate a tyrosine phosphorylation site. However, when compared to the sequence of tyrosine phosphorylation site of Stat1 and Stat5 (26), it appears that this protein contains a putative tyrosine phosphorylation site at position 296 which is presumably a target of JAK2. A hydropathy plot (42) identified a hydrophobic region at the N terminus (Figs. 5 and 6), suggesting the presence of a secretion signal peptide which is usually proteolytically removed during protein maturation. However, since the N-terminal peptide sequence predicted from the cDNA clone matches that of the purified protein, this region is not cleaved off. Another hydrophobic region spanning 21 amino acids is located between residues 230 and 250 (Fig. 6); this region may function as a transmembrane domain. If the N-terminal hydrophobic domain is used to initiate translocation across the endoplasmic reticulum membrane and the second domain acts as a stop-anchor, then mature PRAP at the cell surface would have the unusual structure of a membrane protein with an extracellular loop rather than an extracellular domain with a free end.

DISCUSSION

The corpus luteum is an important target tissue of PRL (43, 44, 45). PRL and PRL-like hormones of placental origin are essential for the survival of this endocrine gland and for its ability to produce progesterone during pregnancy (44, 45). Luteal cells are terminally differentiated and do not undergo cell division when stimulated by PRL, in contrast to the mitogenic effect of PRL on Nb2 lymphoma cells. PRL receptors, therefore, may mediate distinct signaling events in Nb2 cells, a standard model for PRL signal transduction, and the corpus luteum. Throughout pregnancy, both forms of the PRL receptor are present in the corpus luteum, with the long form more abundant than the short form (14).²

The long form of the PRL receptor is able to signal through JAK2 and Stat5 (17, 18, 19, 26, 28, 29). In the rat ovary, both JAK2 and Stat5 have recently been identified (46, 47). Preliminary data from our laboratory indicate that treatment of luteinized granulosa cells with PRL causes increased phosphorylation of both JAK2 and Stat5 (47). In addition, PRL induced the level of Stat5 mRNA and Stat5 binding to the cis-acting DNA elements of the ₂-macroglobulin gene (46), a gene known to be up-regulated by PRL in the rat ovary (48, 49). These data are consistent with a PRL signal being transduced through the JAK2-Stat5 system in the rat ovary, presumably via the long receptor form. However, inhibition of JAK2 activity with tyrosine kinase inhibitors did not affect the regulation of the 20-hydroxysteroid dehydrogenase gene by PRL, implying the involvement of other kinase systems or the presence of another PRL signaling pathway in the rat corpus luteum (47). Indeed, a serine/threonine kinase, Raf-1, has been reported to be complexed with the PRL receptor and activated by PRL (21), perhaps via recruitment of SHC and Ras to the receptor complex (40).

Although results from several studies have revealed that the short form is unlikely to be involved in milk gene transcription and cell proliferation, the short form was found to be expressed in many tissues, as ubiquitously as the long form (15). Thus, the short form may also play a particular role in signaling. The short form of the PRL receptor may regulate PRL responsiveness in target tissues by forming inactive heteromeric receptor complexes (25). Recent data, although, demonstrate that the short receptor can also signal and stimulate cell proliferation (30). Our results demonstrate the preferential association of PRAP to the cytoplasmic domain of the short form of the PRL receptor in luteal cells and further suggest that the short form may also be important in the luteotropic action of PRL. It is interesting to note that a 32-kDa protein, with a pI similar to that of PRAP, is expressed in HC11 cells and also associates with the short form of the PRL receptor (8), although this protein has not yet been identified. PRAP antiserum does recognize a 32-kDa protein in this mammary gland epithelial cell line (data not shown).

PRL treatment had no effect on the receptor association or the tyrosine phosphorylation of PRAP. Thus, phosphorylated PRAP appears to be associated with PRL receptor in the basal state. PRAP is tyrosine phosphorylated, and a putative tyrosine phosphorylation site at position 296 (GTNYVKGQ) is similar to the JAK2 phosphorylation sites on Stat1 (GTGYIKTE) and Stat5 (VDGYVKPQ) (26). Thus, one possible role for PRAP may be to occupy JAK2 and prevent it from acting on other targets in the absence of PRL. However, PRAP protein might also be a substrate of other tyrosine kinases, such as c-src, which has recently been reported to associate with the PRL receptor in the rat liver (39), a tissue in which the short form of the receptor is present at high levels (15).

The predicted amino acid sequence of the PRAP protein includes an N-terminal as well as an internal hydrophobic region that may function as a signal sequence and a transmembrane domain, respectively. In this model, the N-terminal 229 amino acids would serve as an extracellular domain containing all N-linked glycosylation sites, and the C-terminal 84 amino acids would represent an intracellular domain containing the potential tyrosine phosphorylation site. However, sequence analysis of the purified PRAP protein demonstrated that the N-terminal sequence is not cleaved off. Therefore, if PRAP is translocated to the cell surface, it may have the unusual structure of a transmembrane protein in which the extracellular domain is attached at both ends to the cell membrane to form an extracellular loop. If PRAP is a transmembrane protein, its association with the PRL receptor may be similar to the interaction of gp130 with other cytokine receptors (16, 50).

In summary, we have succeeded in cloning and characterizing a novel 32-kDa luteal protein which is the first identified protein shown to associate with the short form of the PRL receptor. It will be interesting to determine whether PRAP participates in PRL signal transduction and whether PRL inhibition of elongation factor 2 phosphorylation and/or inhibition of 20-hydroxysteroid dehydrogenase gene expression involve the short form of the PRL receptor-PRAP complex.