The human has two relaxins, termed H1 and H2, both of which are biologically active and coexpressed in the decidua, placenta and prostate; in the corpus luteum, the main source of circulating relaxin, only the H2 form is expressed. The reasons for this differential expression of the relaxin genes are unknown. The possibility that their 3 -untranslated regions (UTRs) contribute to this differential expression by affecting their mRNA stabilities was investigated. Thus the 3 -UTRs of both relaxin genes were isolated through a combined 3 -rapid amplification of cDNA ends-PCR (RACE-PCR) using poly (A) + RNA from human decidua, placenta, prostate and corpus luteum. The sequences obtained for each 3 -UTR were identical in the tissues examined, were AT-rich (72%) and showed 91% homology between relaxin H1 and H2 when maximally aligned to include several gaps, the significance of which is unknown. Relaxin H1 has two, and relaxin H2 has one, poly (A) + signal, in addition to one cytoplasmic polyadenylation element 30 nucleotides upstream of this. The mRNA levels of relaxin H1 and H2 in the prostate adenocarcinoma LNCaP.FGC cell line were determined by quantitative competitive RT-PCR. Relaxin H1 had a 10-fold greater number of molecules (2·5 10 7 ) per µg of total RNA than relaxin H2 (2·5 10 6 ). The stability of relaxin H1 and H2 mRNAs were compared in LNCaP cells treated with the transcription inhibitor actinomycin D (10 mM) for 0, 1, 2, 4, 8, 10, 14, or 24 h. Half-lives of 3·17 days for relaxin H1 mRNA and 11·4 h for relaxin H2 mRNA were obtained from semi-logarithmic plots. Thus both mRNAs are relatively stable; however, relaxin H1 mRNA is considerably more stable than relaxin H2, at least in LNCaP cells. This difference in their mRNA stability may partly explain the greater level of expression of relaxin H1 in these cells.

Relaxins are known endocrine and autocrine/ paracrine hormones that play a major role in reproduction. In the human there are two relaxin genes, H1 and H2 which share 90% sequence homology within their coding region. The biological and evolutionary significance of two highly homologous and biologically active human relaxins is unknown. In order to achieve a better understanding of the regulatory mechanisms involved in the differential expression of these two genes and to gain insight into their role(s) in the preterm premature rupture of the membranes, we have investigated the properties of their 5 -upstream regions and mapped them both by radiation hybrid and breakpoint mapping into the same chromosome 9p24·1 locus. The 5 ends of these relaxin genes could be divided into a proximal highly homologous segment and a distal non-homologous region. Within the proximal region are contained several putative regulatory elements common to both genes, suggesting a similar regulatory mechanism. The clustering of the relaxin genes within the same chromosomal locus suggests that these genes may be under a common regulation. On the other hand, a distinct gene-specific regulation may also exist for the individual relaxin genes since cis elements specific to each gene were identified at their 5 ends. Moreover, the observed divergence at the distal region of their 5 -upstream sequences may provide the structural features that act as gene-specific transcription regulators. Since the two genes are highly homologous in both their coding and flanking regions, the divergence at the distal region of their 5 ends may be important in the regulation of these genes and in their involvement in the pathology of preterm birth.