The pregnancy-associated glycoproteins (PAG) constitute a large family of recently duplicated genes. They show structural resemblance to pepsin and related aspartic proteinases. A total of 21 bovine (bo) PAG and 9 ovine (ov) PAG cDNA have been identified. Phylogenetic analysis indicated that the PAG are divided into two main groupings that accurately reflect their tissue expression, as determined by in situ hybridization. In the first pattern, represented by ovPAG-2 and boPAG-2, -8, -10, and -11 (where the numbering is arbitrary and reflects order of discovery within species), expression occurred throughout the outer epithelial layer of the placenta (trophectoderm). The second pattern was predominant localization to binucleate cells. Ribonuclease protection assays, which allow discrimination between closely related transcripts, have shown that the expression of PAG varies in a temporal manner over pregnancy. Of those bovine PAG expressed predominantly in binucleate cells, boPAG-1, -6, and -7 are expressed weakly, if at all, by Day 25 placenta, but are present at the middle and end of pregnancy. Others, such as boPAG-4, -5, and -9, are expressed at Day 25 and at earlier stages. Although not among the earliest PAG produced by the trophoblast, boPAG-1 has been used for pregnancy diagnosis, particularly in dairy cows, where there is a major need for a sensitive method capable of detecting pregnancy within 1 mo of conception. It seems likely that some of the newly discovered PAG will be better candidates than PAG-1 for pregnancy diagnosis.

Pregnancy in cattle and sheep can be diagnosed by the presence of a conceptus-derived antigen in maternal serum that is secreted by trophoblast and placental tissue primarily as an acidic component of Mr. 67%000. Molecular doning of its cDNA reveals that the antigen belongs to the aspartic proteinase family and has >50% amino add sequence identity to pepsin, cathepsin D, and cathepsin E. The inferred sequences of the ovine and bovine polypeptides show -73% identity to each other. Critical amino acid substitutions at the active site regions suggest that both proteins are enzymatically inactive. The antigen is a product oftrophoblast binuceate cells that invade maternal endometrium at implantation sites.

No abstract

The pregnancy-associated glycoproteins (PAGs) are structurally related to the pepsins, thought to be restricted to the hooved (ungulate) mammals and characterized by being expressed specifically in the outer epithelial cell layer (chorion͞trophectoderm) of the placenta. At least some PAGs are catalytically inactive as proteinases, although each appears to possess a cleft capable of binding peptides. By cloning expressed genes from ovine and bovine placental cDNA libraries, by Southern genomic blotting, by screening genomic libraries, and by using PCR to amplify portions of PAG genes from genomic DNA, we estimate that cattle, sheep, and most probably all ruminant Artiodactyla possess many, possibly 100 or more, PAG genes, many of which are placentally expressed. The PAGs are highly diverse in sequence, with regions of hypervariability confined largely to surfaceexposed loops. Nonsynonymous (replacement) mutations in the regions of the genes coding for these hypervariable loop segments have accumulated at a higher rate than synonymous (silent) mutations. Construction of distance phylograms, based on comparisons of PAG and related aspartic proteinase amino acid sequences, suggests that much diversification of the PAG genes occurred after the divergence of the Artiodactyla and Perissodactyla, but that at least one gene is represented outside the hooved species. The results also suggest that positive selection of duplicated genes has acted to provide considerable functional diversity among the PAGs, whose presence at the interface between the placenta and endometrium and in the maternal circulation indicates involvement in fetal-maternal interactions.

The pregnancy-associated glycoproteins (PAGs) are secretory products synthesized by the outer epithelial cell layer (chorion) of the placentas of various ungulate species. The amino acid sequences of eight PAGs have been inferred from cloned cDNA of cattle and sheep, as well as of the non-ruminant pig and horse. We compare the PAG sequences and present results of the three-dimensional models of boPAG-1 and ovPAG-1 that were constructed on the basis of the crystal structures of homologous porcine pepsin and bovine chymosin using a rule-based comparative modelling approach. Further, we compare peptide binding subsites defined by interactions with pepstatin and a decapeptide inhibitor (CH-66) modelled on the basis of crystal structures of other aspartic proteinases. We have extended our analysis of the peptide binding subsites to the other PAG molecules of known sequence by aligning the PAG sequences to the structural template derived from the pepsin family and by making use of the three-dimensional models of the boPAG-1 and ovPAG-1. The residues that are likely to affect peptide binding in the boPAG-1, ovPAG-1 and other PAG molecules have been identified. Sequence comparisons reveal that all PAG molecules may have evolved from a pepsin-like progenitor molecule with the equine PAG most closely related to the pepsins. The presence of substitutions at the S1 and other subsites relative to pepsin make it unlikely that either bovine, ovine or the porcine PAG-1 have catalytic activity. Only two of the eight PAGs examined (porcine PAG-2 and equine PAG-1) retain features of active aspartic proteinases with pepsin-like activity. Our results indicate that in the PAGs so far characterized the peptide binding specificities differ significantly from each other and from pepsin, despite their high sequence identities. Analysis of the various peptide binding subsites demonstrates why both bovine and ovine PAG-1 are capable of binding pepstatin. The strong negative charge in the binding cleft of boPAG-1 and ovPAG-1 indicates a preference for lysine- or arginine-rich peptides. PAGs represent a family where the possible peptide binding function may be retained through their binding specificities, but where the catalytic activity may be lost in some cases, such as the boPAG-1, ovPAG-1 and the poPAG-1.

Pregnancy-associated glycoproteins (PAG) are members of the aspartyl proteinase gene family that were initially identified in cattle (bPAG) and sheep (oPAG) as placenta-specific antigens in maternal blood. The objective of this study was to determine whether PAG are expressed in pig trophoblast. A porcine conceptus cDNA library was screened with 32P-labeled ovine and bovine PAG cDNA. Of the approximately 10(4) plaques that were initially screened, a very high number (approximately 5.3%) were positive. Two distinct types were identified, and full-length clones representing each type (1371 bp, pPAG1; 1378 bp, pPAG2) were fully sequenced in both directions. Their open reading frames coded polypeptides of 389 and 387 amino acids, respectively, including 15 amino acid signal peptides. Each had several potentials sites for N-glycosylation. Both were members of the aspartic proteinase gene family, with approximately 50% amino acid sequence identity to porcine pepsinogen and 64% to each other. They were only distantly related to PAG of ruminant species (53% and 49% identify in amino acid sequence to oPAG1 and bPAG1, respectively). Interestingly, pPAG1 had amino acid substitutions within its catalytic center (Gly-->Ala81, domain 1; Thr-->Ser263, Thr-->Ser265, Ser-->Ala266, domain 2) that together were likely to render it enzymatically inactive, whereas pPAG2 retained sequences identical to pepsin in these regions. Western blotting of secretory products of porcine trophoblast with anti-oPAG1 and anti-bPAG1 antisera indicated that pPAG, like PAG from ruminants, had an unexpectedly high M(r)(approximately 70,000).(ABSTRACT TRUNCATED AT 250 WORDS)

Pregnancy-associated glycoprotein (PAG)-1 (PAG1) and pregnancy-specific protein B are either identical or closely related antigens released by trophoblast binucleate cells of placentas of cattle. Sheep and other ruminants produce similar products. There is evidence, however, that these antigens, which are related structurally to the pepsinogens and other aspartic proteinases, are not single gene products but members of an extensive family. Here, the sequential use of ammonium sulfate precipitation and Sepharose blue, anion-exchange, and cation-exchange chromatographies, as well as isoelectric elution from a Mono P column, has allowed several PAG1-related molecules to be purified from the medium after culture of explants from Day 100 sheep placentas. Each of these PAGs cross-reacted to a varying extent with a panel of three different anti-PAG1 antisera. Four of them, all of which were major secretory products of the placenta, were subjected to amino-terminal microsequencing. Although each was related to ovine (ov) PAG1, none was identical. Reverse transcription-polymerase chain reaction was then used to amplify PAG1-related cDNA from Day 100 placental RNA. Seven novel full-length cDNA, all distinct from ovPAG1, were identified from 25 cDNA selected for sequencing. Only two of these (ovPAG3 and ovPAG7) encoded polypeptides identical in sequence at their inferred amino termini to one of the PAGs (ovPAG65) purified from explant cultures. Even so, they were only 84% identical in overall sequence. The remaining five cDNA were unique. In situ hybridization analysis revealed that expression of ovPAG3 and ovPAG7, like that of ovPAG1, is confined to trophoblast binucleate cells. The data confirm that at Day 100 of pregnancy the ovine placenta produces many different PAGs, which differ considerably in sequence and immunological cross-reactivity.

Here we report the molecular cloning of several members of a family of novel proteins expressed by the ruminant trophoblast, known as the trophoblast Kunitz domain proteins (TKDPs). Each contains a carboxyl-terminal module of approximately 64 amino acids belonging to the Kunitz family of serine proteinase inhibitors. These Kunitz modules are preceded by one or more structurally related domains, each about 80 amino acids long. The function of these domains is unclear. The TKDPs differ considerably in sequence identity, with much of the diversity due to variability in the amino-terminal domains. However, nine of the ten Kunitz domains described here are themselves unique, ranging in amino acid sequence identity from 90% to 53% to each other and averaging only about 50% identity with bovine pancreatic trypsin inhibitor (BPTI). The "warhead" P1 residues, which govern specificity, are themselves variable and include some unusual amino acids, such as Asn, Thr, and Ile, as well as the more common Lys. The Kunitz domains of three of the TKDPs lack the conserved cysteines at positions 14 and 38 (BPTI numbering) that normally contribute to the orientation of the inhibitory loop. Northern blotting and reverse transcription-polymerase chain reaction (RT-PCR) demonstrated that the TKDP genes do not exhibit identical expression patterns during trophoblast development, although mostly are expressed maximally during early pregnancy. It is possible that the TKDPs provide a broad range of specificities against maternal proteinases that might be damaging to the trophoblast during pregnancy.