Early pregnancy factor (EPF) is a secreted protein with immunosuppressive and growth factor properties. It was first described as a factor that appeared in maternal serum within 24 hours of fertilisation, in all mammalian species tested. Subsequent studies have shown that it is required for successful establishment of pregnancy, and for proliferation of normal and neoplastic cells. Amino acid sequencing of EPF, isolated from human platelets, showed it is a homologue of chaperonin 10 (Cpn10), a heat shock protein associated with protein refolding in the mitochondrial matrix. These apparently identical proteins have quite different functions and sites of activity and this thesis explores the mechanisms which underlie such disparate biology.
The aim of the studies presented in this thesis was to determine if EPF and Cpn10 production is differentially regulated at the level of gene expression. Evidence was sought that these two homologous proteins are either the products of separate genes, or alternatively spliced products of the same gene. Two biological models, in which changes in the activity of EPF had been observed, were chosen to study Cpn10 related gene expression. These were liver regeneration following 2/3 partial hepatectomy and early pregnancy.
Cpn10 related gene expression in liver regeneration was examined using northern hybridisation with a probe to the Cpn10 open reading frame (ORF). Although RNAse H digestion, followed by northern hybridisation identified four distinct species of Cpn10 related mRNA, no real difference was detectable between samples from rats subjected to 2/3 liver removal and rats undergoing liver manipulation only (sham samples). Hence it seemed unlikely that any of the bands detected was related specifically to EPF appearance following partial hepatectomy.
Initial studies of Cpn10 gene related expression in 11.5 dpc mouse embryos, examined by wholemount in situ hybridisation with a Cpn10 ORF probe, showed a distinct pattern of expression which appeared to be restricted to specific regions. Intense staining was present in the distal part of the tail and in the fore and hind limb buds, with strong signal also evident within the somites and pharangeal arches. As this distinct pattern of hybridisation was seen in tissue regions known to be involved in rapid growth, it was thought to reflect EPF expression rather than Cpn10 mRNA, which was expected to be ubiquitous.
Because of this suggestion of specific EPF expression in the 11.5 dpc embryo, a cDNA library from that tissue source was screened with a Cpn10 ORF probe. Fifteen cDNA clones were isolated from the library and sequenced. Of these clones, six showed divergent 5' UTR or 3'UTR sequence. Expression of the divergent sequences was 1 investigated further by RT-PCR. These experiments failed to detect expression of these variant Cpn10-like messages in tissue from 11.5 dpc mouse embryos, or in mouse liver pre or 12 h post partial hepatectomy (PH), raising suspicion that these were chimeric sequences and thus artefacts of the library. A modified 5' RACE was also performed on RNA from 11.5 dpc embryos. Whilst this approach identified a further 53 nucleotides of Cpn10 sequence, 5' to the published mouse cDNA, it did not reveal any mRNA, distinct from the published mouse Cpn10 cDNA, which could feasibly be message from a separate EPF gene.
Subsequent to the initial embryo in situ hybridisation studies, a description of the structure of the rat Cpn10 /Cpn60 gene was published, in which it was shown that these two genes are transcribed from a bi-directional promoter. This information prompted a re-examination of Cpn10 expression in the 11.5 dpc embryo, with Cpn60 expression also examined for comparison. The restricted pattern of expression, which had been interpreted as EPF specific, was also seen for Cpn60, and indeed, for several other mitochondrial genes. Hence, the 11.5 dpc embryo was not particularly enriched for EPF expression, but rather was shown to have a high level of Cpn10 expression.
The abundant expression of Cpn10, noted in both liver and the 11.5dpc embryo, would hamper the detection of a rarer, Cpn10 related message. Studies switched from expression-based techniques, to genomic screening of a murine library for Cpn10 related sequences. In an initial screen of the genomic library, the Cpn10 ORF probe was used at low stringency to detect as many Cpn10 related sequences as possible. Reprobing the same filters at a higher stringency, with a probe containing exon one of the mouse Cpn60 gene, was then used to identify (and exclude) clones that contained both Cpn10 and Cpn60. These clones were considered likely to represent the mouse Cpn10 gene. The remaining clones should contain genomic variants of the Cpn10 gene, including the putative gene encoding EPF.
Results of the genomic library screen showed that the Cpn10 gene family consists of 8 distinct gene sequences. One gene family member, Cpn10-rs1, came close to satisfying the criterion that it could encode a protein which was essentially identical to Cpn10. Two other Cpn10 related sequences were identified that may encode functional variants of either Cpn10 or EPF, and the remaining gene family members, apart from Cpn10, were identified as pseudogenes.
Cpn10-rs1 is an intronless variant of Cpn10, and contains evidence that it is a retrotransposed cDNA sequence, including a remnant polyadenylated sequence. While most retroposons exist in the genome as processed pseudogenes, on rare occasions they acquire novel functions. This appears to be the case with Cpn10-rs1. Upstream of the translation start point, the Cpn10-rs1 and Cpn10 cDNA sequences diverge abruptly. Analysis of this 5' region of Cpn10-rs1 suggests that the retrotransposed gene has acquired a functional promoter, with basic features of this region including a TATA-like box, a predicted transcription initiation site 25 nucleotides downstream of this site and recognition sites for transcription factors in the upstream region. Primer extension studies were performed to identify the transcription start site for Cpn10-rsl. However, the results of these experiments were inconclusive.
To establish whether Cpn10-rs1 was a functional gene, it was imperative to determine if this genomic sequence was expressed. Furthermore, to determine if this was the gene encoding EPF in the mouse, the pattern of expression needed to be compared to that expected for EPF. EPF is known to be specifically induced during times of rapid protein synthesis, including embryogenesis, tumour growth, and liver regeneration, whilst its protein homologue, Cpn10, has been shown to be abundantly and ubiquitously expressed. Cpn10-rs1 expression was examined in several models of EPF production, including early pregnancy and liver regeneration.
A nested RT- PCR assay was developed to detect the putatively rare Cpn10-rs1 transcript in an abundant background of Cpn10. The final step of the Cpn10-rs1 assay consists of detection of a single nucleotide change to differentiate between the products of the two genes. Three different methods to detect a specific nucleotide change were examined - single nucleotide primer extension, oligonucleotide hybridisation, and an enzyme digest. The results obtained with all three methods were similar.
Cpn10-rs1 mRNA expression in murine liver post-PH was found to closely mirror the serum detection of EPF, with samples taken at 4 h, 6 h, 8 h and 12 h testing positive for Cpn10-rs1 whilst it was not detected in any pre hepatectomy samples tested. Cpn10-rs1 expression in murine ovaries during early pregnancy was also found to closely correlate with the protein expression pattern, with Cpn10-rs1 detectable in the very eariy stages of gestation (d 1-4), but not detected in ovaries from later stage pregnant animals (d 7.5, d 9.5). Interestingly, expression studies of the different stages of the mouse oestrus cycle showed that Cpn10-rs1 was also present in ovaries obtained from mice in early estrus and estrus, but not during diestrus. This differs from the pattern observed for protein activity and possible explanations for this phenomenon are explored in the thesis.
As well as examining expression of Cpn10-rs1 at the message level, the resulting protein product was examined to determine if it was capable of functioning as EPF. A recombinant Cpn10-rs1 protein was produced and its behaviour in an EPF bioassay examined, its performance compared favorably with that of native platelet derived EPF. As a preliminary experiment in studies to examine the mechanisms of EPF secretion, a recombinant Cpn10-rs1 protein fused to a green fluorescent protein was expressed in mammalian cells. Localisation of the protein within the cell was viewed by fluorescence microscopy. Accumulation of the protein in/on the mitochondria was seen.
In conclusion, this thesis identifies a putative murine gene for EPF. Expression and functional studies of Cpn10-rs1 suggest that the gene is expressed, and that it is likely to encode EPF. However, EPF is a protein detectable in all mammalian species investigated to date. Because of the highly conserved biological function of EPF, it would be expected that, if Cpn10-rs1 were the murine gene for EPF, highly homologous sequences would also be present in other species. Features of Cpn10-rs1 which differentiate it from Cpn10, such as a lack of introns and a divergent 5'UTR, should be detectable in other mammalian genomes. This thesis concludes by considering the results of in silico analysis of the human genome for a Cpn10-rs1 equivalent gene.