THE JOURNAL OF BIOLOGICAL CHEMISTRY (0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 5, Issue of February 15, pp. 3298-3305, 1993 Printed in U.S. A.

Heterologous Expression and Characterization of Prolactin-like Protein-A IDENTIFICATION OF SERUM BINDING PROTEINS*

(Received for publication, September 30, 1992)

Santanu Deb$, Gary P. HamlinS, Katherine F. Roby$Q, Simon C. M. Kwoknj, and Michael J. Soare&** From the Departments of $Physiology and YBiochemistry, University of Kansas Medical Center, Kansas City, Kansas 66160

In this report, we describe the heterologous expres- sion of prolactin-like protein-A (PLP-A) in Chinese hamster ovary (CHO) cells, the characterization of re- combinant PLP-A, and the identification of serum PLP-A-binding proteins CHO cell and native placental PLP-A showed similar immunoreactive characteristics and electrophoretic mobilities. N-terminal sequencing verified the identity and purity of the recombinant PLP-A species and the site of cleavage of the signal peptide from the mature secreted PLP-A species. Re- combinant PLP-A lacked activity in standardized pro- lactin and growth hormone in vitro bioassays. Anti- bodies generated to recombinant PLP-A facilitated the cellular localization of PLP-A and the identification of high molecular weight PLP-A complexes. Cross-link- ing analyses of radioiodinated PLP-A with serum har- vested from late gestation rats indicated the presence of two major cross-linked complexes migrating under reducing conditions at 130 and 250 kDa and two minor cross-linked complexes migrating at 70 and 110 kDa. Binding of PLP-A to serum proteins was specific for PLP-A and not effectively competed by other members of the prolactin/growth hormone family. The PLP-A binding species were also found in serum from non- pregnant female and male rats.

Prolactin-like protein-A (PLP-A)’ is one of at least six different structurally related proteins abundantly expressed by the rat placenta (Soares et al., 1991). PLP-A is a member of the prolactin (PRL)/growth hormone (GH) family. Genes for PLP-A, PRL, and relatives expressed in the placenta are localized to chromosome 17 of the rat genome whereas the GH gene is present on chromosome 10 (Cooke et al., 1986;

* This work was supported by Grants HD-20676 and HD-29036 from the National Institute of Child Health and Human Develop- ment. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

4 Supported in part by a Wesley Foundation postdoctoral fellow- ship and a National Research Service Award Postdoctoral Fellowship HD-07502. Present address: Dept. of Anatomy and Cell Biology, University of Kansas Medical Center, Kansas City, KS 66160.

11 Present address: Dept. of Obstetrics and Gynecology, Albert Einstein Medical Center, 5501 Old York Rd., Korman 303, Philadel- phia, PA 19141-3025.

** To whom correspondence should be addressed Dept. of Physi- ology, University of Kansas Medical Center, Kansas City, KS 66160. ’ The abbreviations used are: PLP-A, prolactin-like protein-A; PRL, prolactin; GH, growth hormone; PL, placental lactogen; CHO, Chinese hamster ovary; DMEM, Dulbecco’s modified Eagle’s me- dium.

Levan et al., 1991; Deb et al., 1991~). Unlike the pituitary- specific rat PRL protein, PLP-A is a glycoprotein and secreted in large amounts by trophoblast cells of the chorioallantoic placenta during the second half of gestation (Campbell et al., 1989; Deb et al., 198913). PLP-A is synthesized as a 25 kDa precursor that is glycosylated to 29 and 33 kDa species (Deb et al., 198913). In circulation, PLP-A exists as a high molecular weight complex (Deb and Soares, 1990) reminiscent of t.he behavior of GH in blood (Baumann, 1991). Although, the biological actions of PLP-A are presently unknown, the struc- tural relationship of PLP-A to PRL, GH, and other members of the cytokine superfamily (Bazan, 1990; Soares et al., 1991) suggests a possible role for PLP-A in coordinating metabolic adjustments between fetal and maternal compartments (see Soares et al., 1991, for review). As a first step in attempting to understand the role of PLP-A in pregnancy, we have generated recombinant PLP-A, investigated the activities of PLP-A in PRL and GH in vitro biological assays, and exam- ined the nature of circulating high molecular weight PLP-A complexes.

M A T E R I A L S A N D M E T H O D S

Reagents Reagents for polyacrylamide gel electrophoresis and Affi-Gel Blue

affinity chromatography resin were purchased from Bio-Rad. Adju- vant for immunizations was obtained from Organon Teknika Corp. (West Chester, PA). Methotrexate was purchased from Calbiochem (San Diego, CA). All restriction enzymes, polymerases, and DNA ligase were purchased from New England Biolabs (Beverly, MA). Nitrocellulose was obtained from Schleicher and Schuell. Human placental lactogen (PL) and rat and ovine PRL were obtained from the National Hormone and Pituitary Program (Baltimore, MD). Bovine GH was obtained from United States Department of Agricul- ture (Beltsville, MD). Antibodies to mouse cu2-macroglobulin were obtained from Nordic Immunological Laboratories (Capistrano Beach, CA). The cross-linking reagent, bis(sulfosuccinimidy1) suber- ate, and the iodination reagent, Iodo-Gen, were obtained from Pierce Chemical Co. Streptavidin-biotin immunoperoxidase kits for rabbit immunoglobulin G were obtained from Zymed Laboratories (South San Francisco, CA). Unless otherwise noted, all other chemicals and reagents were purchased from Sigma.

Animals and Tissue Preparation Holtzman rats were obtained from Harlan Sprague-Dawley (Indi-

anapolis, IN). The animals were housed in an environmentally con- trolled facility, with lights on from 0600-2000 h, and allowed free access to food and water. Timed pregnancies and tissue dissections were performed as previously described (Soares, 1987). The presence of a copulatory plug or sperm in the vaginal smear was designated day 0 of pregnancy. Blood was collected by decapitation from preg- nant rats (at different days of pregnancy), non-pregnant female rats, and male rats. Blood was centrifuged a t 3000 revolutions/minute for 15 min, and serum was collected. Placental cytosol was prepared as previously described (Deb et al., 1989a). Protein concentrations of

3298

PLP-A and PLP-A Binding Proteins 3299

sample preparations were estimated by the method of Bradford (1976). Protocols for the care and use of animals were approved by the University of Kansas Animal Care and Use Committee.

Generation of Recombinant PLP-A Construction of the PLP-A Expression Vector-The PLP-A cDNA

used in these experiments was obtained as a gift from Drs. Mary Lynn Duckworth and Henry G. Friesen (University of Manitoba, Manitoba, Canada; Duckworth et al., 1986). The expression vector, pMSXND, was originally constructed by Drs. S.-J. Lee and Daniel Nathans of Johns Hopkins University for the generation of recom- binant mouse proliferin (a member of the mouse placental PRL family; Lee and Nathans, 1988). This expression vector contains a metallothionein promoter, SV40 splicing, and polyadenylation sig- nals, a neomycin resistance gene, and a dihydrofolate reductase gene (Fig. 1). The protocol we have used is based largely on that previously described by Linzer and colleagues (Colosi et al., 1988a, 1988b).

A 825-base pair EcoRI fragment containing the entire coding sequence of PLP-A was isolated from plasmid pRP6-5 after electro- phoretic separation in a 5% polyacrylamide gel. Termini of the isolated fragment were filled with Klenow fragment of DNA polym- erase-I. XhoI linkers were added to the ends, enabling the PLP-A cDNA to be inserted into the unique XhoI site of the pMSXND vector. Subcloning at the XhoI site situates the PLP-A insert down- stream of the mouse metallothionein-I promoter and upstream of the SV40 splicing and polyadenylation signals. Recombinant plasmid DNA was purified from Escherichia coli JM105 cells by lysozyme and alkaline-SDS lysis of bacterial cultures, followed by two handings on CsC1-ethidium bromide density gradients (Sambrook et al., 1989). The orientation of the PLP-A cDNA in the expression vector was confirmed by Hind111 and BamHI digestion. The plasmid containing the correct PLP-A insert orientation was designated as pRPLPA.

Transfection, Selection, and Amplification-Chinese hamster ovary cells (CHO) obtained from the American Type Culture Collection (ATCC CCL 61, CHO-K1, Rockville, MD) were used as the host for the pRPLPA expression vector. CHO cells were routinely maintained in DMEM/MCDB-302 culture medium containing 1 mM proline, antibiotics, and 10% fetal bovine serum. Transfection was carried out by electroporation according to the procedure described by Bothwell e t al. (1990). Following electroporation, the cells were transferred to tissue culture flasks containing DMEM/MCDB-302 culture medium containing the supplements described above. Forty-eight h following electroporation, the selection process was initiated.

Selection was performed with the neomycin analog, Geneticin (G418). The cells were transferred to minimal essential medium containing proline, antibiotics, 10% dialyzed fetal bovine serum, and G418 a t a final concentration of 1 mg/ml. Once sizable colonies of cells were evident the amplification process was initiated.

Amplification takes advantage of the sensitivity of the dihydrofo- late reductase construct present in the pMXSND vector to metho- trexate. CHO-transfected and selected cells were exposed to succes- sive increases in methotrexate (in the G418 selection medium). The methotrexate concentrations were increased every 2-3 weeks in a step wise manner (starting with 0.1 p~ and concluding with 200 p ~ ) .

P L P - A

p R P L P A 9.6 kb

FIG. 1. Schematic representation of the PLP-A expression vector, pRPLPA. The filled box represents the location of the PLP- A cDNA in the plasmid. Abbreviations: SV40, splicing and polyade- nylation signals of SV40; mdhfr, mouse dihydrofolate reductase gene; (T;418', neomycin resistance gene; Amp', ampicillin resistance gene; Mtl, mouse metallothionein I promoter.

Expression of recombinant PLP-A protein was monitored by Western blot analysis.

Generation of Conditioned Medium from CHO Cells Transfected with the PLP-A Expression Vector-After selection and amplification the transfected CHO cells were grown to confluence in DMEM/ MCDB-302 medium containing 10% fetal bovine serum. The mono- layers were washed with DMEM/MCDB-302 medium without serum and cultured in the same serum-free medium supplemented with 100 nM cadmium chloride. The medium conditioned by the monolayers were collected and replaced with fresh medium at 24-h intervals for a t least 12 days. The first 24-h collection was discarded because of the presence of residual serum. The remainder of the conditioned medium was pooled and stored frozen a t -80 "C until further analysis.

Purification of Recombinant PLP-A-Approximately 8 liters of conditioned medium were processed in three different batches. Con- ditioned medium was thawed and centrifuged a t 19,000 X g for 20 min to remove debris. Zinc sulfate was added to the supernatant to a final concentration of 100 mM and stirred for 30 min. The resulting precipitates were collected by centrifugation a t 10,000 X g for 15 min. Pellets were then dissolved in 500 mM Na-EDTA, pH 7.5, and dialyzed against two changes of distilled water (6 liters each) and finally phosphate-buffered saline (PBS: 10 mM sodium phosphate, 150 mM NaCl, pH 7.4). The dialysate was then subjected to concanavalin A affinity chromatography as previously described (Deb et al., 1989b). Proteins specifically eluted from the concanavalin A-Sepharose col- umn with a-methyl mannoside (0.2 M) and N-acetylglucosamine (0.2 M) were concentrated by ultrafiltration and chromatographed on a Sephacryl S-200 column ( 3 X 92 cm), equilibrated with 20 mM sodium phosphate buffer, pH 7.4, containing 300 mM NaC1. During the chromatography the flow rate was maintained at 20 ml/h. Fractions collected from the gel filtration column were monitored for absorb- ance a t 280 nm and for PLP-A immunoreactivity by enzyme-linked immunoassay or by Western blot analysis (Deb et al., 1989a, 198913). All steps in the purification were performed at 4 "C. Separation of the 29- and 33-kDa PLP-A species was achieved by SDS-polyacryl- amide gel electrophoresis and electroelution (Deb and Soares, 1990).

Characterization of Recombinant PLP-A Amino-terminal Sequence Analysis-Amino acid sequences were

determined by the Edman degradation procedure, using a gas-phase protein sequencer as previously described (Deb et al., 1991a, 1991b). Approximately 20 pg of the 29- and 33-kDa PLP-A protein species were sequenced.

Generation of Antibodies to Recombinant PLP-A-Purified recom- binant 33-kDa PLP-A (approximately 100 pg) was used to immunize a New Zealand White rabbit (CAMM Research Laboratories, Wayne, NJ). Booster injections (50 pg each) were given at 2-week intervals. The remainder of the immunization protocol and sera collection were performed as previously described (Deb et al., 1989a).

Western Blot Analysis-Western blot analyses were performed as previously described (Soares et al., 1988 Deb et al., 1989a). Samples were separated by SDS-polyacrylamide gel electrophoresis in 7.5 or 10% gels under reducing or non-reducing conditions. Proteins from gels were electrophoretically transferred to nitrocellulose. Antibodies generated to synthetic peptides corresponding to three different re- gions of the PLP-A sequence (amino acids 101-114, 129-145, and 152-164; Deb et al., 198913) and to recombinant PLP-A were used in the Western blot analyses. In some experiments, preimmune serum, antibodies saturated with the respective purified antigen, or antibod- ies directed against amino acids 56-70 of rat PL-I1 (Deb et al., 1989a) were used as controls.

PRL and GH in Vitro Bioassays-PRL-like and GH-like bioactiv- ities of PLP-A were assessed by evaluating the effects of purified recombinant PLP-A on the proliferation of Nb2 rat lymphoma cells and on the differentiation of mouse 3T3-F442A cells, respectively. PRLs stimulate proliferation of Nb2 rat lymphoma cells (Gout et al., 1980), and GHs stimulate adipocyte differentiation of 3T3-F442A cells (Morikawa et al., 1982).

The Nb2 cell proliferation assay was conducted as described pre- viously (Tanaka et al., 1980) with minor modifications (Deb et al., 1991a).

The adipose conversion 3T3-F442A assay was performed according to the procedure of Green and co-workers (Morikawa et al., 1984; Djian et al., 1985). Briefly, exponentially growing 3T3-F442A cells were inoculated into 35-mm dishes a t a density of 600 cells/cm*. After 3-4 days (near confluence), growth medium (DMEM with 5% cat serum and 0.5% calf serum) was replaced with conversion medium

:3:300 PLP-A and PLP-A Binding Proteins

A B c

67

43

31

21

Excess peptide - + - +

Tissue placenta CHO-PLP-A

A B C D

67 - 43 -

31 -

21 -

101- 129- 152- PL-II 114 145 164 ab

PLP-A abs

A B C D

67 - 43 -

31 -

21 -

Cd++ - + - "

Day of culture 3 10

FIG. 2. Immunochemical analysis of PLP-A produced by engineered CHO cells. Top panel, examination of the specificity of the reaction of I'LP-A produced by the placenta (lanes A and R ) or hy CHO cells (lone C and D) with antipeptide antibodies to PLP-A. Placental cytosolic preparations and conditioned medium generated from CHO cells transfected with the pRPLPA expression vector were electrophoretically separated in 12.5% gels, transferred to nitrocel- lulose, and examined by Western blot analysis. Lanes A and C were incubated with antipeptide antibodies t.o amino acids 152-164 of the I'LP-A sequence. Lanes H and I ) were incubated with same antipep- t.ide antihodies saturated with antigen. M, standards (XlO-") are shown. Middle panel, examination of the reactivity of PLP-A pro- duced hy CHO cells with antibodies directed to various regions of the I'LP-A sequence (lanrs A-C) and to PL-I1 (lane I ) ) . Proteins were electrophoretically separated in 12.5% gels, transferred to nitrocel- lulose, and examined by Western hlot analysis. Antisera directed to amino acids 101-114 (lane A), 129-145 (lane R ) , and 152-164 (lane C ) of the PLP-A sequence, and amino acids 56-70 of the PL-I1 sequence (lane D ) were each used a t a final dilution of 1:500. M, standards (XIO-") are shown. Bottom panrl, stability of the respon- siveness of the engineered CHO cells t.o cadmium chloride. Proteins isolated from medium conditioned by engineered CHO cells after 3 (lanrs A and H ) or 10 (lanes C and I ) ) days of serum-free culture in the absence (lanes A and C) or presence (lanes H and D ) of 100 nM cadmium chloride were examined by Western hlot analysis using

(DMEM with 1.5% cat serum, 1% calf serum, 6 pg/ml bovine insulin. 5 pg/ml transferrin, 2 nM triiodothyronine, 1 pM biotin, 40 p M 2- mercaptoethanol, and 30 ng/ml epidermal growth factor) containing the test substance. After 10 days of additional culture, adipose con- version was monitored by measuringglycerophosphate dehydrogenase activitv according to the procedure of Wise and Green (1979).

A range of recombinant PLP-A concentrations were examined in each of the assays. Rat and ovine PRLs were used as standards in the Nh2 lymphoma cell assay and bovine GH was used as a standard in the 3T:3-F442A adipocyte conversion assay.

Lmmunoo'tochemistr-The distribution of PLP-A in CHO cells and in rat placental tissues was examined with antibodies generated to recornbinant PLP-A using previously described procedures (Hunt and Soares, 1988; Rohy et al., 1991). The PLP-A antiserum was used a t a final dilution of 1:200. Specificity of the immunoreactions was determined by comparing the reactivity of the antibodies with anti- bodies absorbed with excess antigen.

CHO cells transfected with the pRPLP-A expression vector were plated in 8-well chamber slides, exposed to 100 nM cadmium chloride for 3-5 days, rinsed in PBS, and fixed in ice-cold acetone for 5 min. Immunoreactive PLP-A complexes were localized by incubation with a fluorescein-conjugated goat anti-rabbit immunoglobulin G (154 final dilution).

Placentas from day 19 of gestation were fixed in Rouin's fixative, dehydrated, embedded in paraffin, sectioned at 8 pm and mounted on poly-1,-lysine-coated slides. Localization of PLP-A was determined using PLP-A antibodies and a streptavidin-biotin immunostaining kit for rabbit immunoglobulin G.

Identification of PLP-A Serum-binding Proteins

Evidence for the existence of PLP-A serum-binding proteins was derived from Western blotting experiments using antibodies to re- combinant PLP-A and from experiments using radiolabeled recom- binant PLP-A and the cross-linking reagent bis(sulfosuccinimidy1) suherat.e.

Enrichment and Wrstern Blot Analysis of Serum PLP-A High Molecular Weight Complexes-Serum (5 ml) collected from rats sac- rificed on day 19 pregnancy was subjected to concanavalin A and Affi-Gel Blue affinity chromatographic procedures. Proteins specifi- cally eluted from the concanavalin A-Sepharose column with c ~ -

methyl mannoside and N-acetylglucosamine were dialyzed exten- sively in sodium phosphate buffer (20 mM, pH 7.1) and then loaded on an Af'fi-Gel Blue (mesh size 50-100) column (1.5 X 10 cm) equilibrated with the same sodium phosphate buffer. The column was washed with 10 volumes of the loading buffer, and bound proteins were eluted with the same sodium phosphate buffer containing 2 M NaCI. Concanavalin A affinity chromatographically enriched serum proteins binding and not binding t.o Affi-Gel Blue were analyzed by Western blotting using antibodies to recombinant PLP-A.

Cross-linking Experiments with Radiolabeled Recombinant PLP- A-The identity of complexes formed between radioiodinated PLP- A and serum proteins was determined following cross-linking, elec- trophoresis, and autoradiography. Radioiodination was carried out using the iodination reagent Iodo-Gen as previously described (Soares et a/., 1988). The radiolabeled protein (2 X lo5 cpm) was incubated overnight with diluted rat serum (100 pl, 1:l ratio of serum to PRS) a t 4 "C in the absence or presence of excess recombinant PLP-A, ovine PRL, bovine GH, or human P L (100 pg each). Cross-linking reactions were carried out a t room temperature for 30 min in the presence of 1 mM his(sulfosuccinimidyl) suberate solut.ion in sodium phosphate buffer (50 mM, pH 7.4). The reaction was stopped by addition of the same sodium phosphate buffer containing 50 mM et.hanolamine and 20 mM N-ethylmaleimide. Cross-linked proteins were concentrated by acetone precipitation or concentrated and fur- ther enriched by immunoprecipitation with antibodies to recombinant PLP-A as previously reported (Deb et al., 1989b). Samples were then separated by SDS-polyacrylamide gel electrophoresis under reducing conditions, the gels dried, and the migration of radioactive proteins determined by autoradiography.

antibodies directed to amino acids 152-164 of the PLP-A sequence. Please note the stability of the response to cadmium chloride. M, standards (XlO-") are shown.

PLP-A and PLP-A Binding Proteins XI01

RESULTS placent,al PLP-A (Fig. 3). These results furt.her suggested a Heterologous Expression of PLP-A-Medium conditioned

by CHO cells transfected with the PLP-A expression vector (pRPLPA) contained prot.eins t.hat possessed immunochemi- cal and elect.rophoretic characteristics similar to native pla- cental PLP-A (Fig. 2, top and middle panels). Two major species were evident with sizes corresponding to 29 and 33 kDa. Minor species were also present possessing sizes either intermediate or less than the major species. These immuno- reactive proteins specifically reacted with antibodies directed t o three different domains of the PLP-A molecule but not to antipeptide antibodies specific to rat PL-I1 (Fig. 2, middle pnnel). CHO cell production of recombinant PLP-A was inducible by cadmium and maintained for a t least 10 days of culture under serum-free conditions (Fig. 2, bottom panel).

Comparison of the electrophoretic behavior of CHO cell- secreted PLP-A under reduced and nonreduced conditions revealed an interesting difference (Fig. 3). Unlike our previous analysis of nonreduced PLP-A from placental cyt,osol or serum which migrated as high molecular weight complexes (Deb and Soares, 1990), nonreduced CHO cell-PLP-A mi- grated as a monomer and faster than reduced PLP-A. The faster migration is suggestive of the presence of a large looped- disulfide bridge (see Southard and Talamantes, 1991). Ex- amination of placental cell-secreted PLP-A generated under similar serum-free culture conditions demonstrated that mon- omeric behavior was not unique to nonreduced CHO cell- expressed PLP-A but also a feature common to nonreduced

A B

96 - 67 - 43 - 31 - = 21 - 14 -

"..

2-mercaptoethand +

A B . .,

43 -

IrJ 3 1 - s

21 -

2-mercaptoethand + - FIG. 3 . Comparison of the electrophoretic behavior of PLP-

A produced by the placenta ( top panel) and by CHO cells (bottom panel) separated under reducing (lane A ) and non- reducing (lane B ) conditions. Samples were separated in 12..5r,; gel and examined by Western Mot analysis using antibodies to amino acids 152-164 of the PLP-A sequence. Please note the faster migra- tion of immunoreactive proteins under nonreducing conditions. M, standards (x10-9 are shown.

role for serum proteins in the generation of the high molecular weight PLP-A complexes.

Characterization of Purified Recombinant PIP-A-Sequen- tial steps in the purification of the 33-kDa recombinant PLP- A species are shown in Fig. 4. Steps in the isolation included: zinc sulfate precipitation, concanavalin A affinity chromatog- raphy, gel filtration chromatography, and SDS-polyacryl- amide gel electrophoresis with electroelution. The 29- and 33- kDa PLP-A isoforms could not be resolved by gel filtration chromatography. PLP-A-positive fractions obtained from the gel filtration column were divided int.0 two groups: fractions 50-65 and fractions 66-75. Fractions 50-65 contained most of t.he 33 kDa PLP-A species plus minor high molecular weight, contaminating non-PLP-A protein species (data not shown). Fractions 66-75 contained proportionately less of the 33-kDa PLP-A species but were devoid of cont.aminating non-PLP-A protein species (Fig. 4, lane C). The fraction 66-75 pool of PLP-A was used for assessing PLP-A biological and binding activities. PLP-A isoforms electroeluted following gel electro- phoresis were used for amino acid sequencing and antibody generation.

N-terminal sequence analysis of the purified 29- and 33- kDa proteins confirmed their identity with PLP-A. The first six N-terminal amino acids were the same for both the 29- and 33-kDa species and were identical to the N terminus predicted from the PLP-A nucleotide sequence (Duckworth et al., 1986) Met'-Arg'-Ala"-Lys4-Leu"-Leu". A minor PLP-A species (less than 20%) was also present in both the purified 29 and 33 kDa preparat,ions. The minor species represented a truncation of PLP-A with its N terminus beginning a t Leuh: Leu~-Leu~-Asn'-VaI~-His"-Asn'". Verification of the N termini of PLP-A species isolated from placental tissues has not been successful (Deb and Soares, 1990).

Recombinant PLP-A did not exhibit PRL-like activities in thc Nb2 lymphoma cell proliferation assay (Table I) nor GH- like activities in the 3T3-F442A adipocyt,e conversion assay (Table 11).

Antibodies generated to the 33-kDa recombinant PLP-A specifically recognized both 29- and 33-kDa PLP-A species secreted by placental explants (Fig. 5 ) . Furthermore, the antibodies direct,ed to recombinant PLP-A localized PLP-A predominantly to the cytoplasm of PLP-A-secreting CHO cells and t.o trophoblast cells within the junctional zone of the chorioallantoic placenta (Fig. 6). Antibodies made to the 38-

A B C D E

96 -. - 67

- 43

- 31

- .. - 21

-

0 -. -

m . . - -14

FIG. 4. Electrophoretic separation of protein samples from various steps in the isolation of PLP-A from CHO cell-condi- tioned medium. Samples were separated in 12.5:; gels under reduc- ing conditions and stained with Coomassie Blue. I,nnP A , zinc sulfate precipitation: lone H , concanavalin A chromatographv: Innc C, Seph- acryl S-200 gel filtration chromatographv (pool o t fractions 66-75): lnnr I), electroelution of SDS-polyacrylamide gel elect rophoretically separated :I:l-kDa PLP-A: lnnr b,', M, standards (XlO-") are shown.

3302 PLP-A and PLP-A Binding Proteins TABLE I

I’rolifrratilv rrsponses of Nb2 lymphoma cells to recombinant P I 2 - A and ovine PRI, (mean f S.E.)

Data in this table are from a representative experiment performed in triplicate. Approximate molecular weights: PLP-A, 29-33,000; PRL, 24,000. n:. mi.<“&-?Z

* ’ Treatment Cell no. (x io3) ’*

Control 6.0 f 0.6

PLP-A (ng/ml) 0.1 6.2 f 0.8

6.4 f 0.8 6.1 f 0.5 6.6 +. 0.9

1 .0 6.0 +. 0.6 -~ L ’ f” ~~ ~-

10.0 100.0

1000.0

PRL (ng/ml) 0.1 16.5 f 1.2 1 .0

10.0 28.6 f 2.0 32.2 f 1.8

100.0 1000.0

33.0 f 2.0 33.3 f 1.5

TARLE I1 I,’ffectc of rrcombinant PLP-A and bovine GH on adipocvtc

differentiation of 3T3-F442A cclls (mean f S.E.) Data in this table are from a representative experiment performed

in triplicate. Approximate molecular weights: PLP-A, 29,000-33,000; GH, 22,000. The 1000 ng/ml concentration of GH provides maximal stimulation in the assay.

Treatment CPDH activitf

Control 61 f 10

PLP-A (ng/ml) 1

1 0 100

1000

50 f 8 46 f 12 49 f 10 52 f 11

GH (ng/ml)

10 100

1 203 f 38 411 f 35 790 f 49

1000 1192 f 115 ” Glycerol phosphate dehydrogenase activity (units/mg protein).

One unit of enzyme activity was defined as the oxidation of 1 nmol of NADH/min.

A B C D

96 - 67 - 43 -

31 -

21 - 14 -

FIG. 5. Characterization of antibodies generated to recom- binant PLP-A. Medium conditioned by placental explants was concentrated and separated in 12.5% gels and examined by Western blot analysis using antibodies to amino acids 152-164 of the PLP-A sequence (lanes A and H ) or to recombinant PLP-A (lanes C and I)). Antibodies used in blots shown in lanes R and I1 were saturated with their respective antigens. M , standards (xlO-’) are shown.

FIG. 6. Immunocytochemical localization of PLP-A expres- sion in CHO cells and in placental tissues with antibodies generated to recombinant PLP-A. Toppanel, localization of PLP- A expressed in CHO cells. The CHO cells harboring the pRPLPA plasmid were grown, induced with cadmium chloride, and processed as described under “Materials and Methods.” Cells were incubated with PLP-A antibodies (1:200 dilution) and then with fluorescein- conjugated goat anti-rabbit immunoglobulin G (1:64 dilution). Micro- graph A , cells photographed under fluorescence microscopv; micro- graph H , cells photographed under phase-contrast microscopy. Note the predominantly cytoplasmic location of immunoreactive PLP-A. Magnification, X 200. Bottom panel, immunocytochemical localiza- tion of PLP-A in the rat chorioallantoic placenta with antibodies generated to recombinant PLP-A. Sections of the rat placenta were stained for the presence of PLP-A using a streptavidin-biotin im- munoperoxidase kit for rabbit immunoglobulin G. The antiserum was used at a final dilution of 1:200. Specific immunostaining was local- ized to spongiotrophoblast cells of the junctional zone of the placenta (micrographs A and C). Absorption of the antiserum with PLP-A resulted in a loss of immunostaining (micrographs R and D). Note the heterogeneity of spongiotrophoblast cell PLP-A immunostaining that is more evident. a t higher magnification (micrograph C). Abbre- viations: JZ , junctional zone; I,Z, labyrinth zone. Magnification X 100 (micrographs A and R ) ; X 200 (micrographs C and I ) ) .

kDa PLP-A protein exhibited a greater reactivity to PLP-A than did previously generated PLP-A antipeptide antibodies (Deb et al., 1989).

Identification of PLP-A Serum-binding Proteins-Serum PLP-A complexes are resistant to various perturbations such as denaturants, heat, salt concentration, pH (Deb and Soares, 1990). PLP-A monomers can be isolated from serum by the combined treatment with SDS, reducing agents, and heating (Deb and Soares, 1990). This strong association of serum PLP-A complexes permitted the use of SDS-gel electropho- resis under nonreducing conditions for the isolation of the PLP-A complexes. High molecular weight serum PLP-A com- plexes were enriched by concanavalin A and Affi-Gel Blue affinity chromatographies, electrophoretically separated un- der nonreducing conditions, transferred to nitrocellulose, and probed with antibodies generated to recombinant PLP-A.

PLP-A and PLP-A Binding Proteins 3303

Western blot analysis specifically identified two major PLP- A immunoreactive species possessing molecular masses of 110 and 230 kDa (Fig. 7).

In order to further determine the identity of the high molecular weight PLP-A complexes present in serum, cross- linking experiments were performed with radioiodinated re- combinant PLP-A and serum collected from rats at day 19 of pregnancy. The specificity of PLP-A associated with com- plexes possessing sizes of approximately 130 and 250 kDa was demonstrated (Fig. 8, top panel). The migration of these two major PLP-A-serum protein complexes and some minor com- plexes were further resolved following immunoprecipitation with antibodies to recombinant PLP-A (Fig. 8, middle panel). Electrophoretic migrations of the major complexes under reducing conditions were a t 130 and 250 kDa and the minor complexes at 70 and 110 kDa (Fig. 8, middle panel). Under nonreducing conditions the major complexes migrate faster, 110 and 230 kDa (data not shown), corresponding to the complexes identified by Western blot analysis. The specificity of PLP-A cross-linking with serum proteins was further de- termined by coincubation with excess ovine PRL, bovine GH, and human PL (Fig. 8, bottom panel). These hormones failed to interfere with the binding of radiolabeled recombinant PLP-A to serum proteins (Fig. 8, bottom panel) as did the addition of n2-macroglobulin and mannose 6-phosphate (data not shown). Furthermore, immunoprecipitation with antibod- ies to cu2-macroglobulin failed to precipitate the PLP-A com- plexes (data not shown). PLP-A-binding proteins were pres- en t in serum harvested from male and female rats (Fig. 9, top panel). The concentration of the 250-kDa PLP-A-binding protein appeared to increase as gestation progressed (Fig. 9, bottom panel).

DISCUSSION

In the present study, we have expressed PLP-A in CHO cells, characterized recombinant PLP-A proteins, and re- solved in part the nature of the high molecular weight PLP- A complex present in circulation.

PLP-A proteins expressed by engineered CHO cells pos- sessed immunochemical and biochemical characteristics sim- ilar to PLP-A produced by trophoblast cells of the rat cho-

200 ,

116 96 . 67 '

A 6

c 230

c 110

43 '

31

Excess - PLP-A

i-

FIG. 7. Immunoidentification of high molecular weight PLP-A complexes. PLP-A high molecular weight complexes were partially purified via concanavalin A and Affi-Gel Blue affinity chro- matographic procedures, separated by SDS-polyacrylamide gel elec- trophoresis under nonreducing conditions, and examined by Western hlot analysis using antibodies to recombinant PLP-A. Note the specificity of the immunoreactions with the two high molecular mass I'LP-A complexes a t 110 and 230 kDa. Under reducing conditions the two high molecular mass complexes migrate slower, 130 and 250 kDa (see Fig. 8). M , standard (xIO-") are shown.

67 - - . *

4 3 - 0 Excess PLP-A - +

- 250 200 - " - - 130

96 - - 70 c 110 116 -

67 -

200 . 116 . 96

67

A B C D E

I)-

FIG. 8. Specificity of cross-linking of radioiodinated PLP- A with rat serum proteins. Serum (100 PI) collected from day 19 pregnant rat was diluted with PRS (1:l v/v) and incuhated with radiolabeled PLP-A (2 X IO5 cpm) overnight a t 4 "C in the presence or absence of excess unlabeled PLP-A (top panel) or ovine PRL. bovine GH, and human PL (bottom panel). Cross-linking was per- formed with bis(sulfosuccinimidy1) suberate. Cross-linked products were either acetone precipitated (top panel) or immunoprecipitated with PLP-A antibodies (middle and bottom panels) and separated in 1.0 /A SDS-polyacrylamide gel electrophoresis under reducing condi- tions. Gels were then dried and processed for autoradiography. M , standard (XlO-") are shown. Note the specificity of the 130- and 250- kDa radiolabeled PLP-A cross-linked complexes (top panel). Immu- noprecipitation identified major PLP-A cross-linked complexes at 130 and 250 kDa and minor complexes a t 70 and 110 kDa (middle panel). Unlabeled PLP-A specifically competed for binding of radio- labeled PLP-A to serum proteins (top panel). However, the addition of excess ovine PRL (lane C), bovine GH (lane D), or human PL (lane E ) failed to interfere with the cross-linking of radiolabeled PLP-A to serum proteins (bottom panel). Please note that lane R of the bottom panel represents a control for the immunoprecipitation.

" -G

rioallantoic placenta. Analyses of recombinant PLP-A re- vealed new information about PLP-A. The N-terminal amino acids of the major secreted PLP-A isoforms (29 and 33 kDa) were identical. Consistent with our earlier prediction that the two molecular weight PLP-A species arose from differential

3304 PLP-A and PLP-A Binding Proteins

A B C

200 - '16 96 - - 1

":i - I P NP M A B C D

200 - m i k m

11 13 16 19 Day of Gestation

FIG. 9. Detection of PLP-A-binding proteins in serum from male and female rats (top panel) and from serum of female rats harvested on different days of gestation (bottom panel). Serum was diluted with PBS (1:l v/v) and incubatedwith radiolabeled PLP-A (2 X IO5 cpm) overnight at 4 "C. Cross-linking was performed with bis(sulfosuccinimidy1) suberate. Cross-linked products were im- munoprecipitated with PLP-A antibodies and separated in 7.5% SDS- polyacrylamide gel electrophoresis under reducing conditions. Gels were then dried and processed for autoradiography. M , standard (X lo-') are shown. Please note that the PLP-A-binding proteins were present in male and female sera (top panel) and that the 250-kDa PLP-A-binding protein complex appeared to increase as gestation progressed (bottom panel).

glycosylation of the PLP-A precursor and not from differen- tial proteolytic processing (Deb et al., 1989b). The information responsible for the generation of the two major PLP-A gly- coprotein species must be inherent in the PLP-A sequence and apparently independent of the cell type expressing PLP- A. In contrast, the intracellular distribution of PLP-A does not appear to be under similar regulation. PLP-A expressed by placental cells is secreted and directed to the nucleus (Deb and Soares, 1990), while CHO cell-expressed PLP-A is not found in appreciable amounts in the nucleus. Nuclear trans- port of PLP-A may be dependent upon factors uniquely present in trophoblast cells or may be the result of alternative transcripts not represented by the PLP-A cDNA used for the generation of the expression vector. The latter mechanism is utilized in the direction of platelet-derived growth factor to nuclear and extracellular compartments (Lee et al., 1987; Maher et al., 1989).

PLP-A is hypothesized to be part of the signaling arma- mentarium of the placenta. The information encoded in the PLP-A signal is yet to be resolved. Structural similarities of PLP-A with PRL and GH dictated an investigation of their functional similarities. PLP-A did not influence the prolifer- ation of Nb2 lymphoma cells, an action characteristic of proteins with PRL-like actions, and PLP-A did not influence the conversion of 3T3-F442A preadipocytes to adipocytes, a

GH-like action. These results suggest that PLP-A possesses biological actions differing from those associated with PRLs and GHs. Such a conclusion may not be entirely unexpected. Sequence identities between PLP-A and PRL or PLP-A and GH are modest, 30 and 22%, respectively (Duckworth et al., 1986). Similar to PLP-A, PL-I also shows a modest overall sequence identity with PRL (approximately 28%; Robertson et al., 1990). However, unlike PLP-A, PL-I shares similar biological actions with PRL (Colosi et al., 1988a). The signal within a regulatory protein dictating PRL or GH actions are not dependent upon each and every amino acid. Receptor binding and activation are dependent upon the precise orga- nization of selected amino acids within the protein (Cun- ningham et al., 1990, 1991; Cunningham and Wells, 1991). Apparently, PLP-A does not possess the basic skeletal ar- rangement required for activation of PRL or GH receptor signaling systems. The functional relationships of members of the rat PRL/GH family are possibly more analogous to the transforming growth factor-@ family which includes trans- forming growth factor-@s, activin, inhibin, and Mullerian inhibiting substance. Members of this family share the same basic structure and anywhere from 25 to 90% sequence iden- tity yet some of the individual members interact with distinct receptors and possess distinct biological activities (Massague, 1990, 1992).

Identification of the nature of the circulating high molecu- lar weight PLP-A complexes has provided some insight into the physiology of PLP-A. Circulating high molecular weight PLP-A complexes result from the specific association of PLP- A with proteins present in serum. The PLP-A complexes exist in multiple sizes suggesting the presence of different types of PLP-A-binding proteins or possibly that the different PLP- A complexes contain multiple copies of the same binding protein. Consistent with this latter point, considerable evi- dence is available indicating that ligands facilitate the oligo- merization of binding proteins or receptors (Cunningham et al., 1991). PLP-A-binding proteins are not related to proteins previously implicated in binding members of the PRL/GH family, a2-macroglobulin (Southard and Talamantes, 1989), mannose 6-phosphate receptor (Lee and Nathans, 1988), growth hormone-binding proteins (Sadeghi et al., 1990), or prolactin-binding proteins (Postel-Vinay et al., 1991). Addi- tionally, PLP-A-binding proteins are not restricted to preg- nant female rats. They are present in serum from both males and females suggesting a more elemental role for PLP-A. At this time, it is not known whether there are extraplacental sources of PLP-A or whether PLP-A interacts with signaling systems used by other regulatory molecules. Presumably, PLP-A-binding proteins direct the delivery of PLP-A to its appropriate target tissues. Relationships have been found between serum binding proteins and receptors (Smith et al., 1989; Baumbach et al., 1989), and this may also be evident for PLP-A-binding proteins and PLP-A receptors. Conse- quently, characterization of serum PLP-A-binding proteins may provide some insight into the signaling system utilized by PLP-A at its target tissues.

Acknowledgments-We gratefully acknowledge Drs. Mary Lynn Duckworth and Henry G. Friesen for generously supplying the PLP- A cDNA, Drs. Se-Jin Lee, Daniel Nathans, and Daniel I. H. Linzer for supplying the expression vector (pMSXND), Drs. Peter Gout and Charles Beer for providing the Nb2 cells, and Dr. Howard Green for providing the 3T3-F442A cells. We also acknowledge Douglas Larsen for technical assistance and James Rouse and Dr. Satya Yadav of the University of Kansas Medical Center Biotechnology Facility for assistance with the N-terminal amino acid sequencing of PLP-A.

PLP-A and PLP-A Binding Proteins 3305 REFERENCES

Baumbach, W. R., Horner, D. L., and Logan, J. S. (1989) Genes & Deu. 3,199-

Baumann, G. (1991) Endocr. Reu. 12,424-449 Bazan, J. F. (1990) Immunol. Today 11,350-354 Bothwell, A,, Yancopoulos, G. D., and Alt, F. W. (1990) Methods for Cloning

Bradford, M. M. (1976) Anal. Biochem. 72,248-254 Campbell, W. J., Deb, S., Kwok, S. C . M., Joslin, J., and Soares, M. J. (1989)

Colosi, P., Ogren, L., Southard, J. N., Thordarson, G., Linzer, D. I. H., and

Colosi, P., Swiergiel, J. J., Wilder, E. L., Oviedo, A,, and Linzer, D. I. H. (1988b)

Cooke, N. E., Szpirer, C., and Levan, G. (1986) Endocrinology 119,2451-2454 Cunningham, B. C. and Wells, J. A. (1991) Proc. Natl. Acad. Sci. U. S. A . 8 8 ,

Cunningbam, B. C., Henner, D. J., and Wells, J. A. (1990) Science 247,1461-

1205

Genes, pp. 145-148, Jones and Bartlett Publishers, Boston

Endocrinology 125, 1565-1574

Talamantes, F. (1988a) Endocrinology 123 , 2662-2667

Mol. Endocrinol. 2,579-586

3407-3411

1 Afi.5 Cunningham, B. C., Ultsch, M., De Vos, A. M., Mulkerrin, M. G., Clauser, K.

Deb, S., and Soares, M. J. (1990) Mol. Cell. Endocrinol. 74 , 163-172 Deb, S., Hashizume, K., Boone, K., Southard, J. N., Talamantes, F., Rawitch,

Deb, S., Youngblood, T., Rawitch, A,, and Soares, M. J. (1989b) J. Biol. Chem. A,, and Soares, M. J. (1989a) Mol. Cell. Endocrinol. 6 3 , 45-56

Deb, S., Faria, T. N., Roby, K. F., Larsen, D., Kwok, S. C. M., Talamantes, F., 264,14348-14353

Deb, S., Roby, K. F., Faria, T. N., Larsen, D., and Soares, M. J. (1991b) and Soares, M. J. (1991a) J. Biol. Chem. 2 6 6 , 1605-1610

Deb, S., Roby, K. F., Faria, T. N., Szpirer, C., Levan, G., Kwok, S. C. M., and Endocrinology 128,3066-3072

Djian, P., Phillips, M., and Green, H. (1985) J. Cell. Physiol. 124 , 554-556 Soares, M. J. (1991~) J . Biol. Chem. 266,23027-23032

Duckworth, M. L., Peden, L. M., and Friesen, H. G. (1986) J. Biol. Chem. 261 ,

R., and Wells, J. A. (1991) Science 254,821-825

10879-10884

Gout, P. M., Beer, C. T., Noble, R. L. (1980) Cancer Res. 4 0 , 2433-2441 Hunt, J. S. and Soares, M. J. (1988) Placenta 9 , 159-171 Lee, B. A,, Maher, D. W., Hannink, M., and Donoghue, D. J. (1987) Mol. Cell.

Lee, S.-J. and Nathans, D. (1988) J. Biol. Chem. 263 , 3521-3527 Levan, G., Szpirer, J., Szpirer, C., Klinga, K., Hanson, C., and Islam, M. Q.

Maher, D. W., Lee, B. A,, and Donoghue, D. J. (1989) Mol. Cell. Biol. 9 , 2251-

Biol. 7 , 3521- 3537

(1991) Genomccs 10,699-718

2253 Massague, J. (1990) Annu. Reu. Cell Biol. 6,597-641 Massague, J. (1992) Cell 69,1067-1070

Morikawa, M., Green, H., and Lewis, U. J. (1984) Mol. Cell. Biol. 4 , 228-231 Morikawa, M., Nixon, T., and Green, H. (1982) Cell 29,783-789

Postel-Vinay, M.-C., Belair, L., Kayser, C., Kelly, P. A., and Djiane, J. (1991) Proc. Natl. Acad. Sci. U. S. A. 88. 6687-6690

Robertson, M. C., Croze, F., Schioedter, I . C., and Friesen, H. G. (1990) Endocrinology 127 , 702-710

Roby, K. F., Larsen, D., Deb, S., and Soares, M. J. (1991) Mol. Cell. Endocrinol. 79,13-20

Sadeghi, H., Wang, B. S:, Lumanglas, A. L., Logan, J. S., and Baumbach, W. R. (1990) Mol. Endocrrnol. 4 , 1799-1805

Sambrook, J. Fritscb, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory, Cold Spring

Smith, W. C., Kuniyoshi, J. , Talamantes, F. (1989) Mol. Endocrinol. 3, 984- Harbor, NY aan

Soares, M. J. (1987) J. Re rod Fertil. 7 9 , 335-341 Soares, M. J., De, S. K., foster, B. A,, Julian, J. A,, and Glasser, S. R. (1988)

Soares, M. J., Faria, T. N., Roby, K. F., and Deb, S. (1991) Endocr. Reu. 12 ,

YY"

J. Endocrinol. 1 1 6 , 101-106

402-433 Southard, J. N., and Talamantes, F. (1989) Endocrinology 125, 791-800 Southard, J. N., and Talamantes, F. (1991) Mol. Cell. Endocnnol. 7 9 , C133-

Tanaka, T., Shiu, R. P. C., Gout, P. W., Beer, C. T., Noble, R. L., and Friesen,

Wise, L. S., and Green, H. (1979) J . Biol. Chem. 254 , 273-275

"_ "_

C140

H. G. (1980) J. Clin. Endocrinol. Metab. 51,1058-1062