Cloning and expression of genes encoding human pregnancy-specific glycoproteins.

The pregnancy-specific glycoproteins (PSGs) of the human placenta and the carcinoembryonic antigens comprise a subfamily within the immunoglobulin superfamily. There may be as many as 20 different PSG genes which are predominantly expressed in the placenta. As an initial step toward understanding the control of PSG expression, we isolated and characterized two nearly identical PSG genes, PSG1 and PSG1-I. PSG1, which lacks exon 1 (5'/L), but contains exons 2 (L/N), 3 (A1), 4 (A2), and 5 (B2-C), encodes five previously identified type I transcripts, PSG1a, 1b, 1c, 1d, and 1e in a L/N-A1-A2-B2-C domain arrangement. PSG1-I, which contains a complete transcriptional unit consisting of exons 5'/L, L/N, A1, and B2-C, encodes type II PSG transcripts in a L/N-A1-B2-C domain arrangement. The predicted PSG1-I-encoded proteins share nearly complete sequence identity with the PSG1-encoded members, except the latter contain extra A domains. Amplification by polymerase chain reaction of placental or hydatidiform mole cDNA demonstrates that PSG1-I is a functional type II PSG gene. Using transient expression assays, we demonstrated that the -834/-34 region upstream of the translational start site of the PSG1-I gene contained the PSG promoter elements and the -834 to -456 region contained negative control elements. Sodium butyrate, an inducer of PSG synthesis, greatly stimulated expression of all PSG1-I-chloramphenicol acetyltransferase (CAT) fusion gene constructs. However, butyrate was at least 2-fold more effective in stimulating CAT activity of fusion genes containing upstream sequences (-834 to -576) than those containing proximal sequences (-456 to -172), suggesting two regions in the PSG1-I gene that mediate the butyrate response.

Transcriptional regulation and the effects of sodium butyrate and glycosylation on catalytic activity of human germ cell alkaline phosphatase.

Human choriocarcinoma cells, the malignant trophoblasts, synthesize germ cell alkaline phosphatase (GCAP) which shares 98% sequence identity with the placental alkaline phosphatase (AP). The two isozymes are immunologically similar but react differentially toward inhibition by L-leucine or EDTA. Administration of sodium butyrate to choriocarcinoma cells greatly increased the transcription rate of the GCAP gene, resulting in an increase in mRNA expression and enzyme biosynthesis. The butyrate-modulated AP induction was blocked by cycloheximide, suggesting that a mediator protein may be involved. Protein sequence deduced from complementary DNA analysis suggests that GCAP contains two potential sites for asparagine (N)-linked glycosylation. The marked increase in GCAP expression by butyrate in choriocarcinoma cells allowed us to study the extent of N-linked glycosylation and its role on GCAP enzyme activity. After limited tunicamycin treatment, Mr 65,000 (fully processed), Mr 58,000 (nonglycosylated), and Mr 62,000 polypeptides were synthesized by these cells in the presence of butyrate. This suggests that the Mr 62,000 product may be the singly glycosylated GCAP monomer and that both sites are glycosylated in this phosphatase. The glycosylated and nonglycosylated GCAPs, synthesized by butyrate-treated choriocarcinoma cells in the absence or presence of tunicamycin, respectively, were similarly inhibited by L-leucine or EDTA. Moreover, the specific enzyme activity of glycosylated and nonglycosylated GCAP remained unchanged, indicating that AP lacking N-linked oligosaccharide side chains was catalytically active. This is supported by the finding that nonglycosylated GCAP incorporated inorganic phosphate which binds to the active site of AP. Since the active form of AP is a homodimer, our data indicate that the glycan moieties are not required for the dimerization and catalytic activity of GCAP.

Effects of sodium butyrate on the synthesis of human pregnancy-specific beta 1-glycoprotein.

Human pregnancy-specific beta 1-glycoprotein (PS beta G) is a polymorphic placental protein which shows strong sequence similarity with the oncofetal protein, carcinoembryonic antigen. To better understand the role of PS beta G in pregnancy, we examined its synthesis and regulation in placental fibroblasts, which had been shown to express the PS beta G gene. The major placental PS beta G is a 72-kDa glycoprotein, while the major fibroblast PS beta G is a 62-kDa species. Administration of sodium butyrate to these fibroblasts slightly stimulated the synthesis of the 62-kDa species but markedly increased the production of two additional PS beta Gs of 72 and 48 kDa. The similarity between the PS beta Gs synthesized by butyrate-treated fibroblasts and human placenta was confirmed by cell-free protein synthesis. Poly(A)+ RNA from butyrate-treated fibroblasts and placenta directed the synthesis of two polypeptides of 48 and 36 kDa, which form the polypeptide backbone of the 72- and 48-kDa glycoproteins. Moreover, the predicted molecular weights of PS beta Gs encoded by the two types of PS beta G cDNA clones were 48,000 and 36,000. Most PS beta G cDNAs identified to date, including the three cDNAs (PSG16, PSG93, and PSG95) isolated in this laboratory, share strong sequence similarity at the 5' region (designated PSG-5') but differ in sequences at their 3' regions. The PSG-5', PSG93-specific, PSG16/PSG93-3', and PSG95-3' probes, which identify the majority of PS beta G mRNAs, hybridized with three PS beta G mRNAs of 2.3, 2.2, and 1.7 kilobases from placental fibroblasts. Butyrate increased the steady-state levels of all three mRNAs. Ribonuclease protection analysis showed that butyrate increased the PS beta G mRNAs containing the PSG-5' or PSG93-specific sequence to approximately 20% of human placental levels. However, unlike human term placenta, which predominantly expressed PS beta G mRNAs with 3'-sequences similar to PSG16/PSG93, the butyrate-treated fibroblasts expressed roughly equal levels of PS beta G mRNAs with the PSG16/PSG93-3' and PSG95-3' ends. All PS beta G cDNAs identified encode proteins with distinct carboxyl termini, suggesting that the composition of the 72-kDa species in placenta and butyrate-treated fibroblasts is likely to be different. Placental fibroblasts provide a unique model for the study of the mechanisms responsible for the differential expression of the PS beta G gene.

Characterization of pregnancy-specific beta 1-glycoprotein synthesized by human placental fibroblasts.

We have previously demonstrated that human placental fibroblasts produce a pregnancy-specific beta 1-glycoprotein (PS beta G) immunologically indistinguishable from placental PS beta G. This was confirmed by the immunocytochemical localization of PS beta G in these fibroblasts. In addition, placental fibroblasts contain all three PS beta G mRNAs of 2.3, 2.2, and 1.7 kilobases which hybridize with the three PS beta G cDNAs (PSG16, PSG93, and PSG95) identified, although at 1.4-2.5% of the levels in human term placenta. The major PS beta G species synthesized by placental fibroblasts is a 62K glycopolypeptide formed from a 58K intracellular precursor polypeptide. However, the PS beta G species found in human placenta are one major glycoprotein of 72K and two minor ones of 64K and 54K. Poly(A)+ RNA from placental fibroblasts directed the synthesis of two polypeptides of 48K and 46K (major), whereas, poly(A)+ RNA from human placenta directed the synthesis of higher levels of four polypeptides of 50 K, 48 K (major), 46 K, and 36 K. Thus, the major PS beta G species found in fibroblasts and human placenta differ. The carbohydrate side-chains are essential for the stability of fibroblast PS beta G, because PS beta G synthesis in these fibroblasts could not be detected in the presence of tunicamycin, a protein glycosylation inhibitor which did not affect PS beta G mRNA expression. Our finding that a variant PS beta G species is produced in placental fibroblasts raises the possibility that the authentic placental PS beta G species may have different functions.

5-Bromo-2'-deoxyuridine induces placental alkaline phosphatase biosynthesis in cultured choriocarcinoma cells.

5-Bromo-2'-deoxyuridine (BrdUrd) stimulated the biosynthesis and hence increased the activity of placental alkaline phosphatase in choriocarcinoma cells. While BrdUrd had no effect on the rate of degradation or processing of placental alkaline phosphatase, it increased the rate of phosphatase synthesis. The stimulation of enzyme activity could be completely accounted for by the increase in alkaline phosphatase protein. Both control and BrdUrd-induced cells contained polypeptides of 61,500 and 64,500 Da, identified as the precursor and fully processed forms of placental alkaline phosphatase monomer. The half-life of this enzyme monomer in both control and BrdUrd-treated cells was estimated to be 36 h. BrdUrd induced a specific increase in the placental alkaline phosphatase mRNA leading to the observed enhancement of biosynthesis. The continued rise in alkaline phosphatase biosynthesis in BrdUrd-induced cells following BrdUrd removal indicated that this analog acted by incorporation into DNA.