SCIENTIFIC CORRESPONDENCE Identity of subsequences of some pregnancy-associated proteins with SERPIN signature sites of some serine protease inhibitors and a carcinoma antigen

The foetus is an allograft in the uterus. Yet it is not rejected by the immunocompetent mother. Understanding of the exact mechanism of foetal survival is limited and the unequivocal evidence still seems elusive. It is proposed that one of the possible reasons for evasion of maternal lymphocyte-mediated lysis by the foetus could be local inactivation of serine proteases, secreted on the target trophoblast cells by cytotoxic lymphocytes during lethal hit, by some pregnancyassociated proteins which contain amino acid subsequences (and presumably function) identical to those of some serine protease inhibitors (SERPINs). In this study, the amino acid subsequences of three pregnancy-associated proteins – human pregnancy zone protein (H. PZP), sheep uterine milk protein (S. UTMP) and pig uteroferrin-associated protein (P. UFAP), were found to be identical to tri-, tetra-, penta-, hexaand heptapeptide sequences of SERPINs like human plasma serine protease (Protein C) inhibitor, and Rat SERPINs 1, 2.1 and 3, respectively. The overall identity of sequences of pregnancy-associated proteins with SERPINs ranged from 15.27% to 28.97%. Interestingly, among the identical sequences, one tripeptide (RPF) and one hexapeptide stretch (FNRPFL) of human plasma SERPIN, two tripeptide subsequences (FNR and RPF) of rat SERPIN 1 and a tetrapeptide (RPFL) and a pentapeptide subsequence (FDRPF) of rat SERPIN 3 were found to be parts of SERPIN signature sites of the respective inhibitors. Human squamous cell carcinoma antigen (H. SCCA) is a tumour marker. The amino acid sequences of S. UTMP and H. SCCA, H. PZP and H. SCCA as well as P. UFAP and H. PZP available in the Protein Sequence Database were aligned using the PC GENE software. Interestingly, the sequence of a pregnancy-associated protein, S. UTMP was found to have an identity of 20.51% with that of carcinoma antigen, H. SCCA whereas, the sequences of H. PZP and H. SCCA were only 9.74% identical to each other. Five tripeptides (Leu–Asp– Ala, Leu–Val–Asn, Phe–Lys–Gly, Met– Met–Arg and Pro–Phe–Leu) were found to be identical between S. UTMP and H. SCCA. The tripeptide sequence Pro– Phe–Leu happens to be a part of the SERPIN signature site of both, S. UTMP and H. SCCA. On the other hand, the sequence of P. UFAP, which is known to be similar to S. UTMP, was found to have only 9.35% identity with H. PZP even though both are pregnancy-associated proteins having SERPIN activity. In the present study, the overall identity of sequences of pregnancy-associated proteins with SERPINs and of a pregnancy-associated protein with a carcinoma antigen were quite significant. An unrelated protein c-myb oncoprotein used as a control, when compared with the pregnancy-associated proteins, was found to have identities of only 3.96%, 4.32% and 5.30% with sheep uterine milk protein, pig uteroferrin-associated protein and human pregnancy zone protein, respectively. This relationship is not only statistically significant but also important biologically. Similarities between pairs of protein sequences are potentially important because they may indicate some functional, structural or evolutionary relationship between the proteins. Biological significance of this kind does not necessarily imply a strong statistical significance. A low statistical significance does not imply that a similarity is not biologically important; on the other hand, a similarity which is very improbable does imply that sequences are related, even if this relationship is not yet understood. The main value of assessment of the statistical significance of a similarity is therefore to provide additional evidence that a biological relationship exists. The immunoglobulin superfamily