The D310 EC cell line is multipotent and differentiation can be directed along the neuronal pathway by addition of retinoic acid (RA). The correlation between the expression of tau protein and the development of stable microtubule arrays during neuronal differentiation was studied by indirect immunofluorescence, ELISA immunoblotting and Immunogold labelling techniques. By immunofluorescence staining tau protein is not detected in undifferentiated EC cells, but it starts to be expressed at day 3 of RA induction and shows a beaded morphological appearance. The expression of tau increases steadily during neuronal differentiation. Double immunofluorescence labelling demonstrates that the expression of tau protein correlates well with the development of colchicine-resistant microtubules. The levels of total cellular tubulin, total class III $\beta$-tubulin and total tau protein were measured by ELISA. The amount of class III $\beta$-tubulin and tau protein were expressed as a ratio of the relative amount of class III $\beta$-tubulin or tau protein/mg total tubulin. ELISA results support the observations of immunofluorescence staining. The relative amount of tau protein expressed after RA induction increases steadily and parallels the increased expression of class III $\beta$-tubulin, an indicator of neuronal differentiation. Immunoblotting shows that both juvenile and mature isoforms of tau protein are present in RA-induced EC cells. Two isoforms are detected in day 3 RA-induced cells and four isoforms are detected in day 6 RA cells. All four isoforms associate with colchicine-stable microtubules. A beaded morphological distribution of tau along neurite processes seen by immunofluorescence staining suggests that tau protein associates with some, but not all microtubules in neurite processes. The prediction that tau associates with colchicine-resistant microtubules was tested by immunogold labelling at the light and electron microscopic levels.

The objectives of the present study were: (i) to establish an immunization protocol for the preparation of monoclonal antibodies (mAbs) specific for epitopes that are polymorphic in humans, (ii) to use this protocol to prepare anti-apoE mAbs that are either apoE isoform-specific or are antibody mimetics of the lipoprotein (LDL) receptor. We immunized mice that carry a transgene encoding apoE of one isoform with a second isoform of human apoE. Two immunization protocols were used. In protocol 1, mice that carried an apoE3 transgene were immunized with either apoE4), a natural apoE variant apoE that is defective in binding to the LDL receptor, or both. We hoped to obtain anti-apoE4-specific mAbs. In protocol 2, apoE transgenic mice were immunized with apoE3 and apoE4. We anticipated that we would generate mAbs that recognize apoE3 but not apoE4 and mAbs that recognize the same conformational apoE epitope that is recognized by the LDL receptor. Mice showed a weak anti-apoE immune response that had the predicted isoform specificity. Only one of the nine fusions generated stable, apoE-specific hybridomas. Of the 7 anti-apoE mAbs obtained, none possessed the isoform-specificity predicted from protocol 1. Two of the mAbs, 9F3 and 7D4 reacted well with apoE3 and apoE4 but failed to react with apoE2. The epitopes for both mAbs were mapped to a region that coincides with the apoE LDL receptor-binding site. Other anti-apoE mAbs could effectively compete with 7D4 and 9F3 for binding to immobilized apoE. (Abstract shortened by UMI.)

Tumorigenesis results from the accumulation of genetic events which give rise to inappropriate behaviour of cells. Mutations which lead to cancer progression are generally classified in two major categories, that of oncogenes and tumor suppressor genes. The latter have been more difficult to identify due to their recessive nature. The application of somatic cell fusion and microcell fusion technologies presents a means to identify genes inactivated in tumor cells. We have used these methodologies to identify specific chromosomes that, once introduced into the DU 145 prostate cancer cell line, have the ability to suppress tumorigenicity. Chromosomes 8 and 12 were shown to encode tumor suppressor genes inactivated in this prostate cancer cell line. To isolate the tumor suppressor gene encoded on chromosome 12, a smaller candidate region was identified by the use of radiation hybrid technology. Molecular methods subsequently led to the identification of chromosome 12 genes displaying increased expression in suppressed hybrids compared to tumorigenic hybrids. Novel genes and genes encoded by other chromosomes also demonstrated increased expression. We suggest that these genes are involved in the cellular pathway that leads to suppression in DU 145 prostate cancer cells upon the introduction of chromosome 12. Further analysis of the genes isolated in this study have the potential to serve as prognostic markers of prostate carcinogenesis.

A mutation occurring outside the coding region in the 3$\sp\prime$UTR or the myotonic dystrophy kinase (DMK) suggests the formation of a defective gene product may not be involved in the disease process. Instead, the effect of the repeat expansion may be directed toward gene regulation or expression. In this study, investigation of the possibility or cellular factors capable of directly binding DNA or RNA revealed the presence of at least two factors in cytosolic extract. Mobility shift assays were done using an RNA probe comprising a portion of the 3$\sp\prime$UTR containing the repeat region and approximately 150bp of upstream and downstream flanking sequence. Probes with repeat sizes of 13 repeats, 45 repeats, and 90 repeats did not appear to vary greatly in gel mobility complex size. Clones using the CAT reporter gene (Gorman et al., 1982) with the DMK 3$\sp\prime$UTR containing repeat sizes from 5 to 90 repeats were constructed. Transient transfection of the constructs into TE32 cells and assay for CAT gene expression revealed increased CAT activity correlating with increasing repeat size. Deletion of portions of the 3$\sp\prime$UTR sequence, either up- or downstream of the repeat region abrogated any CAT activity from constructs containing these variants. The complete inactivity of the deletion clones suggests the repeat sequence must be presented in the context of the full 3$\sp\prime$UTR to impose any regulatory control. Taken together, these data suggest a role for the DMK 3$\sp\prime$UTR in the regulation of gene expression. (Abstract shortened by UMI.)

By screening a Trichomonas vaginalis $\lambda$gt11 cDNA library with anti-serum raised against a purified preparation of T. vaginalis cell-detaching factor (CDF), two clones were identified. Molecular analysis of the cDNA sequences indicated that the clones were distinct and could potentially represent two unique T. vaginalis genes. Further characterization of the CDF anti-serum used in the initial library screening indicated that the serum possessed reactivity to CDF and a variety of other T. vaginalis proteins. The multiple specificities of the anti-serum are likely to have contributed to the identification of the two distinct cDNA clones. While the relationship between the weaker of the two immunoreactive cDNA sequences (CDF-2) and the CDF protein was unclear, the results of Northern hybridization and sequence analysis suggested that the strongly reactive CDF-1 clone did not contain sequences representing a CDF gene. Since the CDF-1 cDNA clone contained an almost complete coding sequence and because the encoded polypeptide exhibited significant protein sequence similarity to bacterial N-acetylneuraminate (Neu5Ac) lyases, this T. vaginalis sequence was further analyzed. Using the cDNA sequence and cloned segments of overlapping genomic DNA, the complete coding and flanking regions of this T. vaginalis gene, designated TvnanA, were determined. TvnanA is a single copy gene that lacks introns and contains a 954 bp open-reading frame that is predicted to encode a 318 amino acid polypeptide with a calculated molecular mass of 35 kDa. The gene is transcribed as a single 1.1 kb polyadenylated mRNA with short 5$\sp\prime$ (17-18 nt) and 3$\sp\prime$ (28 nt) untranslated regions. A sequence resembling the T. vaginalis initiator element was identified in the region surrounding the transcriptional initiation site, however this Inr-like element did not direct accurate gene transcription. The TvnanA encoded polypeptide shows 35% identity and 53% similarity to the Neu5Ac lyase of Escherichia coli, and 73% identity to a predicted Neu5Ac lyase of Haemophilus influenzae. The homology includes significant similarity within a region corresponding to the putative active site pocket of the E. coli enzyme. A complete copy of the TvnanA gene was assembled and the gene product was expressed in E. coli as a glutathione S-transferase fusion protein. In the absence of detergents required for efficient solubilization, the T. vaginalis gene product exhibited Neu5Ac lyase activity that was 30 times greater than background activity in controls. Lyase activity was increased a further 5-fold by deletion of the N-terminal hydrophobic region of the T. vaginalis protein. While fusion proteins exhibited Neu5Ac lyase activity, they did not appear to be capable of producing cell-detaching activity when incubated in the presence of eukaryotic cell monolayers. Preliminary experiments to determine the size and location of the T. vaginalis Neu5Ac lyase suggests that the enzyme is present as an intracellular 35 kDa species. While questions remain concerning the regulation of the TvnanA gene and the role of the N-terminal region in the encoded polypeptide, this is the first reported characterization of a T. vaginalis gene product involved in sialic acid metabolism. The presence of this enzyme combined with the reports of a T. vaginalis neuraminidase suggests that the parasite may utilize host sialic acid as a nutrient source.

Following hormonal ablation in the rat ventral prostate (via castration) or rat mammary gland (after weaning) the epithelial cells of both glands are eliminated by active cell death during the regression of the glands. Several genes including TRPM-2, (testosterone repressed prostate message), RVP.1, fos, and myc, have been shown to be induced in the prostate during this process. We have investigated the expression of several other genes that may be associated with apoptosis, including tissue transglutaminase (tTGase), poly (ADP-ribose) polymerase (PARP), and heat shock protein 27 (Hsp27). Northern hybridization has been used to determine the steady-state mRNA levels of these genes in both the regressing prostate and mammary glands. The time course of induction has been compared to changes in the steady state levels of several control genes including prostate steroid binding protein (PSBP) (prostate), $\beta$-casein (mammary), $\alpha$-tubulin, and TRPM-2 (both prostate and mammary). A novel cross-screening approach first described by Wong et al., 1989 was used to clone and characterize genes other than those described above which were induced during the regression of both rat prostate and mammary gland. Two regression selected genes (RSG) refered to as RSG-2, and RSG-8, were isolated, and their role in the active cell death process has been investigated. RSG-2 is homologous to cathepsin B, a thiol protease that has been previously identified as one of the extracellular proteases that is activated in metastatic cells. RSG-8, is homologous to rat insulin like growth factor binding protein 5 (IGFBP-5), a member of a group of six related proteins that are known to mediate the effects of the growth factors IGF-I, and IGF-II. The steady state level of IGFBP-5 mRNA in the normal prostate is low but detectable. In the regressing prostate, the mRNA for IGFBP-5 mRNA is induced and its steady state peaks in the luminal epithelial cells of the prostate that undergo active cell death (ACD) between 3 and 4 days after castration. The gene is induced in a similar fashion in the regressing mammary gland following weaning. The expression of another IGFBP, namely IGFBP-2 contrasts that of IGFBP-5, with a high steady state level in both prostate and mammary gland, which increase only slightly during the regression process. The increase in mRNA steady-state levels for both IGFBP-2 and IGFBP-S were shown by in-situ hybridization to localize to the epithelial cells that are deleted during the regression process. The differential expression of IGFBP-2 and IGFBP-5 suggest that modulating the epithelial cells response to the IGF growth factors may influence the active cell death process in hormonally dependent tissues such as the prostate and mammary gland. (Abstract shortened by UMI.)

This work was first directed at studying plant mitochondrial gene expression during different early stages of plant development. Because of the approaches used, it also led to the isolation and characterization of new plant mitochondrial genes and in turn that led to the study of gene transfer from the mitochondrion to the nucleus during evolution. Mitochondrial cDNA clone libraries were constructed from mitoplast RNAs of 24 hour germinating wheat embryos and six-day etiolated seedlings. One such cDNA clone was found to share similarity with bacterial and chloroplast S7 ribosomal proteins. This protein is important for ribosome assembly and translation initiation. This single-copy gene (rps7) was found to be actively transcribed and the 5$\sp\prime$ termini of its major transcripts mapped immediately upstream of a repeat element, which also precedes the wheat mitochondrial genes for coxII, orf25 and atp6. The precursor rps7 RNA undergoes C-to-U type of RNA editing at three positions within the coding region and one site unexpectedly decreases amino acid similarity with homologues. This site and the second non-silent one show full editing in RNA from germinating embryos whereas some cDNA clones show incomplete editing for these or the silent site in RNA from etiolated seedlings. The patterns indicate no polarity to the RNA editing process. Sequences homologous to the wheat mitochondrial rps7 gene are present in the rice and pea mitochondrial genomes, but absent from that of soybean. Interestingly, the pea homologue lacks 119 bp of the 5$\sp\prime$ rps7 coding region and it is not present anywhere else in the pea mitochondrial genome nor is the 3$\sp\prime$ coding region transcribed. Therefore, it is a truncated pseudogene. Immediately upstream, there is an open reading frame of 248 amino acids, which is actively transcribed and homologous to the orf228 gene in liverwort mitochondria. The derived protein of this ORF shares similarity with those of the helC genes in Rhodobacter capsulatus and orf263 gene in Bradyrhizobium japonicum. This gene product is believed to be involved in the translocation of heme from the matrix to the intermembrane space in cytochrome c biogenesis. Southern blot analysis showed that there are rps7 homologous sequences present in pea and soybean nuclear DNAs. In order to study them further, pea nuclear DNA, pea cDNA and soybean nuclear DNA libraries obtained from Clontech were screened with the wheat mitochondrial rps7 probes. The potentially positive clones turned out to be yeast-bacterial shuttle vectors (in the case of the pea nuclear DNA library) and mitochondrial DNA contamination (in the pea cDNA library). The potentially positive clones from a soybean nuclear DNA library turned out to contain mitochondrial DNA of the orf228 gene plus a 92 bp DNA sequence homologous to the wheat mitochondrial rps7 gene, although the latter was undetected by Southern hybridization analysis. This work, in addition to identifying and characterizing new plant mitochondrial genes, provides evidence for the recent transfer of functional genetic information from the organelle to the nucleus.

The objective of this thesis was to determine the origin of the myotonic dystrophy (DM) mutation. I have used PCR-based assays of nine polymorphisms spanning a physical distance of 30 kb, within and immediately flanking the DM kinase gene, in order to examine patterns of allelic association with respect to the DM mutation. Four main haplotypes (A-D) were observed in the normal population using these nine markers at the DM locus. Significantly, DM is in complete association with haplotype A, the most common haplotype in the normal population. Our data suggest the presence of two founding chromosomes: one containing a stretch of five contiguous Alu elements (the progenitor for haplotype A) and the other in which three of these have been deleted (the progenitor for haplotypes B, C and D). Individuals with haplotype A displayed the full spectrum of (CTG)$\sb{\rm n}$ number, ranging from 5 to 35 repeats. The (CTG)$\sb5$ alleles as well as alleles with greater than 19 repeats are exclusively linked to haplotype A. In contrast, alleles in which the three Alu elements are deleted possess only (CTG)$\sb{11-14}$ repeats. Although the inference that the loss of the three Alu repeats may confer increased stability on the (CTG)$\sb{\rm n}$ repeat is speculative, the narrow size range of the (CTG)$\sb{\rm n}$ repeat on chromosomes in which the Alu elements have been deleted, relative to the variation seen on normal chromosomes with the DM haplotype (i.e. haplotype A), is striking. (Abstract shortened by UMI.)

Vertebrate distal-less homeobox genes are known to play a role in embryonic pattern formation. Six zebrafish distal-less homeobox genes have been isolated; dlx1, dlx2, dlx3, dlx4, dlx6, and dlx7. Their combinatorial embryonic expression patterns seen in the forebrain, visceral arches, and fins are suggestive of a new "homeobox code". Various doses of all-trans retinoic acid were used to disrupt distal-less expression during different visceral arch developmental time points. The abnormal visceral arch distal-less expression was further correlated with craniofacial cartilage development. Zebrafish embryos at various stages, treated with 10$\sp{-6}$ M RA, displayed a loss of all distal-less expression in the mandibular and hyoid arches. The embryos exhibiting a loss of distal-less expression, also developed with a loss of mandibular and hyoid arch derived craniofacial cartilage components. Embryos treated with 10$\sp{-7}$M RA displayed a stage dependent loss of distal-less expression, mainly in the branchial arches. The visceral arch craniofacial cartilage differentiated abnormally in embryos displaying a loss of distal-less expression. When distal-less expression was lost in the branchial arches, then only the branchial arch cartilage was lost. However, abnormal distal-less expression in any of the visceral arches could also be correlated with abnormal visceral arch cartilage development. These experiments suggest that dlx genes are part of the multi-step process leading to visceral arch cartilage development. Zebrafish genes involved in this process are sensitive to retinoic acid (10$\sp{-7}$M) until 24hpf, where RA does not seem to affect the process of craniofacial cartilage development. Accordingly, a loss in visceral arch dlx expression during any stage (up to 24hpf) of this process leads to abnormal craniofacial cartilage development. (Abstract shortened by UMI.)

Recently, the molecular basis of myotonic dystrophy(DM) has been characterized as an unstable trinucleotide CTG repeat amplification in the 3' untranslated region of a gene encoding a protien with serine/threonine kinase activity. As a first step towards understanding the molecular mechanisms underlying DM, we have analyzed the amplification of the CTG repeat and the DM kinase (DMK) mRNA steady state levels in tissues and cell lines obtained from normal and congenital DM individuals. We have raised polyclonal antibodies against human DMK fusion protein and undertook DMK protein expression analysis in freshly sampled muscle tissues from normal and DM individuals. Our antibody detected DMK protein isoforms of 72 and 84 kDa, for which the levels and distribution were not significantly altered in tissues from adult and congenital DM patients. In addition, we have demonstrated that the previously reported decrease in DMK mRNA expression in affected tissues may be the result of a significant loss of type I myofibers, which preferentially express DMK. In contrast to previous reports, our results also showed that the mutant DMK allele was clearly transcribed as a high molecular weight mRNA in muscle tissue of a severely affected patient. We have examined DMK expression during muscle differentiation in vitro and subsequently investigated the effect of DMK over-expression on the terminal differentiation of the murine myoblast cell line C2C12. We demonstrated that DMK is up-regulated 2 to 3-fold during skeletal myogenesis and that constitutive over-expression of DMK mRNA in myoblasts caused a marked inhibition of myoblast terminal differentiation. Surprisingly, this activity mapped to the 3'UTR of the DMK transcript.When the DMK 3'UTR was placed downstream of a hygromycin resistance gene, the same inhibition of myogenesis was observed. Over-expression of the DMK 3' UTR in NIH 3T3 fibroblasts did not have any effect on their proliferation, suggesting that the 3' UTR does not prevent cell cycle withdrawal and differentiation by promoting growth. Further characterization of the 3' UTR sequences mediating the observed inhibition of terminal differentiation mapped these elements to a 239 bp conserved segment of the 3' UTR located upsteam of the CTG repeat. Furthermore, the DMK 3' UTR did not have any significant effect on the activity of the myogenic regulator MyoD when co-transfected into 10T1/2 cells along with a reporter construct bearing a muscle specific enhancer element. This suggested that the 3' UTR did not interfere with the ability of MyoD to transactivate muscle-specific genes. However, when the mRNA levels for two early myogenic regulators were analysed, myoblast clones over-expressing the 3' UTR expressed normal levels of MEF-2C, but showed reduced myogenin mRNA levels compared to controls, following the induction of differentiation. In addition, in contrast to controls, myogenin protien levels were found to be unchanged during myogenesis in these clones. These data suggested that over-expression of the DMK 3' UTR may alter the expression of specific mRNAs leading to a delay in terminal differentiation. (Abstract shortened by UMI.)