Brachyury T is a gene required for mesoderm formation and is expressed early during gastrulation in presumptive mesoderm. We found that this gene is expressed transiently in P19 cells destined to differentiate into cells of the mesodermal lineages. The expression of Brachyury in P19 cells peaks two days after initiation of differentiation and is induced by cell aggregation. The formation of compact cell aggregates and the expression of Brachyury are both dependent on extracellular calcium. Activin A, bone morphogenetic proteins and cAMP all induced Brachyury expression. By contrast, leukaemia inhibitory factor significantly reduced Brachyury expression in aggregated P19 cells and inhibited differentiation. Overexpression of Brachyury in P19 cells leads to their spontaneous differentiation with low levels of Brachyury resulting in the development of skeletal muscle while higher levels result in the formation of neurons.

The gene of the ubiquitous starch degrading enzyme $\alpha$-amylase has been chosen to study evolutionary mechanisms and the relationships between several insect groups. In this project, the $\alpha$-amylase gene sequence of a representative of the order Lepidoptera, Bombyx mori, was determined. The coding region of Bombyx and the seven other insect genes are approximately the same size and have at least 60% identity with one another. There are various numbers of introns dispersed throughout the loci but often sites are shared between two or more species. There is evidence of differing codon biases among the genes with Drosophila and Anopheles being very GC rich and Choristoneura and Tribolium having virtually no bias. Biases caused discrepancies between the phylogenetic trees created by several different methods. Representatives of the same order always grouped together as predicted, but the order within the Lepidoptera varied with the method used. (Abstract shortened by UMI.)

A PCR strategy was used to amplify novel sequences residing between two degenerate protein tyrosine phosphatase (PTPase) -specific primers. Partial Class I and II clones were isolated from a library constructed with this P19 stem cell derived-cDNA. The predicted amino acid sequence of these PCR fragments revealed sufficient identity with the PTPase consensus to indicate that a region from two individual PTPases had been cloned. Northern analysis of the Class II or PTP2 PTPase revealed a 5 Kb mRNA in all mouse tissues and cell lines examined, while the mouse CAP-1 (Class I) gene was found to be expressed as 5 and 7 Kb mRNA transcripts in a variety of mouse tissues. CAP-1 PTPase cDNAs have been cloned from messenger RNA derived from mouse kidney and embryonal carcinoma cells. The composite cDNA contains an open reading frame which encodes 942 amino acids. The N-terminal portion of the predicted protein (M$\sb{\rm r}$ = 107,000) shares 30-37% identity with homologous regions in the cytoskeletal associated proteins, band 4.1, ezrin, moesin, radixin, and talin. This PTPase has been termed CAP-1 for cytoskeletal associated protein related PTPase. The C-terminal phosphatase domain is 33-41% identical to catalytic domains of known PTPases. The bacterially expressed GST/CAP-1 fusion protein preferentially removes phosphate moieties from tyrosine residues. CAP-1 exhibits a specific pattern of transcription in several tissues, that is, in leg muscle as well as heart and brain. In adult leg muscle, CAP-1 is expressed as two distinct transcripts which are approximately 500 bp shorter than that found in other murine tissues. The corresponding skeletal muscle cDNAs were cloned to elucidate the mechanism by which the different mRNAs arise. The composite leg muscle CAP-1 (LMCAP-1) cDNA, with leg muscle-specific sequence at the 5$\sp\prime$ end, predicts a truncated LMCAP-1 protein which lacks part of the protein 4.1 homology region. This CAP-1myc fusion was transiently expressed in COS-1 cells to elucidate the localization and possible function of CAP-1. A CAP-1myc specific protein close to the predicted molecular weight (110-120 kDa) was immunoprecipitated from these metabolically labeled cells. Concurrently, indirect immunofluorescence with an $\alpha$myc antibody revealed a punctate, non-nuclear pattern of exogenous CAP-1myc expression in transfected COS-1 cells. Morphological changes such as the "giant cells" observed in several CAP-1myc/pMT21- and CAP-1myc/pECE-generated NIH 3T3 clones, appear unique to CAP-1, suggesting that this phenomenon is a result of the dephosphorylation of a cytoskeleton/membrane protein found normally associated with this enzyme.

The cloning of the myotonic dystrophy (DM) was accomplished in a three part research plan: (1) the characterization of the DNA excision repair cross complementation (ERCC1) gene region by genetic and physical mapping to determine the location of the DM gene relative to this locus, (2) the undertaking of a chromosome walk from the ERCC1 region toward the DM gene to define a minimal area that is to contain the DM locus, and (3) characterization of the essential region containing the DM locus for CpG islands and DM associated abnormalities. Further work done in our laboratory and by others have shown the DM associated allelic expansion to be due to the amplification of a trinucleotide repeat, CTG. This repeat was mapped to the 3$\sp\prime$ untranslated region of a gene which based on sequence homology comparisons encodes a putative serine-threonine protein kinase. This is not unlike the allelic expansion found in fragile X syndrome that is due to an amplification of a CGG trinucleotide repeat mapping to the 5$\sp\prime$ untranslated region of a gene designated FMR-1. The size of the CTG repeats ascertained in 124 normal chromosomes was found to range from 5 to 30 with repeat numbers of 5 and 13 the most common. The repeat numbers in DM chromosomes was found to vary from a minimum of about 50 to over 2000. Only two out of 98 DM families did not show allelic expansion using Southern blot analysis or PCR protocols to ascertain the repeat numbers. These cases either have other mutations at or near this locus or they have another clinically similar disorder mapping elsewhere in the genome. In summary, it is evident from work presented in this thesis that the DM locus has been cloned and that the DM mutation has been identified. (Abstract shortened by UMI.)

The seed storage proteins of oat are mainly represented by the globulins (75%) and the prolamins (10%). These proteins are only found in the endosperm and accumulate within protein bodies during the maturation of the oat seeds. Three genomic sequences encoding globulin polypeptides and one avenin genomic sequence encoding an oat prolamin (avenin) have been isolated and characterized. The respective promoters of these genomic clones were fused to the coding sequence of the $\beta$-glucuronidase reporter gene (GUS). These constructs, together with the entire sequence of a globulin gene, were transferred into tobacco via Agrobacterium tumefaciens. Analysis of the transgenic tobacco seeds showed for the first time that the promoters of the oat globulin and avenin genes are able to regulate the expression of the GUS sequence in an endosperm-specific and developmentally controlled manner in a dicot plant, as in oat seeds with the original seed storage protein gene. One of the globulin promoters was shown to be probably inactive, whereas two other promoters appear to direct strong expression in the seeds of transgenic tobacco plants. A deletion analysis on one of the functional promoters demonstrated that a portion of the promoter upstream of nucleotide -259 (relative to the start of transcription as determined in oat) was required for expression. Sequence analysis of the globulin promoters showed the lack of conserved elements which are found in other storage protein gene promoters, and believed to play an important role in the regulation of seed storage protein genes. It was nonetheless demonstrated in this study that the absence of such elements did not prevent a correct functionality of the oat globulin promoters in transgenic tobacco seeds. The avenin promoter, when fused to the GUS sequence, also showed strong expression in the endosperm of transgenic tobacco seeds. Sequence analysis of the upstream region of the avenin gene confirmed the presence of a highly conserved 'prolamin box'. The possible role of this element was therefore further demonstrated in this work. This work has also showed for the first time a difference in the choice of transcriptional start sites of two monocot (oat) seed storage protein genes after transfer into a dicot species (tobacco).

Insulin autoantibodies (IAA) are, by definition, present in an individual who has not received exogenous insulin. The objectives of this project were: (1) to obtain monoclonal IAA-secreting cell lines, (2) to clone and characterize by sequence analysis the variable regions of the light and heavy chains (V$\sb{\rm L}$ and V$\sb{\rm H})$ of a human monoclonal IAA, and (3) to determine the degree of restriction of the V$\sb{\rm L}$ and the V$\sb{\rm H}$ gene usage at the onset of Type I DM. Type I DM patients seropositive for IAA were identified using a competitive insulin autoantibody assay. Their lymphocytes were isolated, immortalized with Epstein-Barr Virus and cloned. Two monoclonal insulin antibody-secreting cell lines were obtained: (1) 1-0.5G1 was obtained from a Type I DM patient treated with exogenous human insulin, and (2) 3-0.5F10 was obtained from a non-insulin treated Type I DM patient; the monoclonal antibody secreted by 3-0.5F10 is therefore an IAA. The V$\sb{\rm L}$ and V$\sb{\rm H}$ cDNAs of the monoclonal IAA secreted by the 3-0.5F10 cell line were cloned and sequenced. The genes were assigned to well defined germ-line elements, the influence of somatic hypermutation in the clonal evolution of IAA was evaluated and the sequences were compared with other antibody sequences. PCR amplification using IAA V$\sb{\rm H}$ and V$\sb{\rm L}$ specific primers followed by Southern analysis using IAA specific probes was performed. All of the Type I DM patients studied at the time of diagnosis had B lymphocytes which were producing antibodies with IAA-homologous V$\sb{\rm L}$ and V$\sb{\rm H}.$ Thus, the usage of IAA-homologous V$\sb{\rm L}$ and V$\sb{\rm H}$ elements in antibodies was found to be restricted in Type I DM patients at the time of diagnosis.

Epidermal Growth Factor (EGF) modulates cyclic AMP (cAMP) levels in non-neoplastic T51B rat liver epithelial cells induced by the $\beta$-adrenergic agonist isoproterenol (IPR) and by forskolin (FSK). In general the effect obtained depends upon whether pretreatment has taken place in the presence of serum (stimulation) or hepes (inhibition) on the duration of the pretreatment and on whether EGF was added before or after IPR or FSK. Activation of protein kinase C (PKC) by the tumor promoting phorbol ester, TPA, potentiated cAMP synthesis stimulated by FSK in both EGF pretreated cells and untreated cells although EGF pretreatment partially attenuated the effect of TPA. Cyclic AMP secretion is one method by which cells regulate the changes in intracellular cAMP. It was observed that cAMP secretion gradually increased with time when untreated and EGF pretreated cells were stimulated with FSK or IPR. The increase in cAMP secretion did not account for all the changes in intracellular cAMP. In fact, although all treatments resulted in some accumulation of external cAMP, there was no correlation between the level of internal cAMP obtained or the decrease in internal cAMP and the amount secreted. DNA synthesis was induced in serum deprived (0.2% BCS) cells by EGF or serum (10% BCS). This induction was inhibited by most cAMP elevating agents except isoproterenol and pertussis toxin. A combination of forskolin and the phosphodiesterase inhibitor, RO 201724 was most effective in inhibiting the DNA synthesis stimulated by EGF. Forskolin and RO 201724 inhibited DNA synthesis in a time dependent manner whereas maximum isoproterenol inhibition (approximately 22%) was observed regardless of time of isoproterenol addition. Genistein, a tyrosine kinase inhibitor, also inhibited EGF induced DNA synthesis. The inhibitory effect of genistein and cAMP elevating agents were not additive. These results suggest that EGF modulates cAMP levels and the mitogenic effect of serum and EGF is inhibited by cAMP elevating agents.

Protein tyrosine kinases (PYKs) play key regulatory roles in the control of cell growth and differentiation. Attempts to identify novel PYKs through expression cloning strategies have led to the identification of a novel family of protein kinases, referred to as dual specificity kinases (DSKs). In addition to their immunoreactivity with antiphosphotyrosine antibodies, DSK family members have the ability to phosphorylate serine, threonine as well as tyrosine residues. A novel protein kinase, Esk (Embryonal carcinoma Ser/thr/tyr Kinase), has been isolated from an embryonal carcinoma (EC) cell line using an expression cloning strategy. Sequence analysis of two independent cDNA clones (2.97 and 2.85 Kb) suggested the presence of two Esk isoforms in EC cells. The Esk-1 cDNA sequence predicted an 857 amino acid protein kinase with a putative membrane spanning domain, while the Esk-2 cDNA predicted an 831 amino acid kinase which lacked this domain. Genomic analysis revealed that the Esk transcripts could arise through alternative splicing of the same primary transcript to generate the cytoplasmic and transmembrane isoforms of the kinase. In adult mouse, Esk messenger RNA levels were highest in tissues possessing a high proliferation rate or a sizeable stem cell compartment suggesting that Esk may play some role in the control of cell proliferation or differentiation. As anticipated from the screening procedure, bacterial expression of the Esk kinase reacted with antiphosphotyrosine antibodies on immunoblots. Furthermore, in in vitro kinase assays Esk was shown to phosphorylate both itself and the exogenous substrate myelin basic protein on serine, threonine and tyrosine residues confirming that Esk is a novel member of the dual specificity family of protein kinases. An antibody raised to the Esk kinase revealed that the protein was subject to developmental regulation; being highly expressed in rapidly proliferating cells, and absent in terminally differentiated cells and in adult mouse tissues. Finally, the Esk kinase was found to associate with the 85 kDa subunit of phosphatidylinositol 3-kinase (PI3K) in proliferating stem cells. In vitro binding studies suggested that the interaction of the Esk kinase with the 85 kDa subunit of PI3K could be mediated via both the SH2 and SH3 domains of this protein. The results presented in this thesis suggest that the Esk dual specificity kinase may play a role in the control of cell growth and differentiation and that the effects of the kinase could be mediated by the regulation of PI3K activity. The interaction of the Esk kinase with an SH2 domain containing protein, is the first indication for the physiological function of the tyrosine phosphorylating activity of this kinase in mammalian cells.

The genetically obese mouse of the C57B1/6 strain displays hyperinsulinemia that is partly due to an exaggerated insulin-secretory responsiveness of its pancreatic $\beta$-cells. The cause of this exaggerated responsiveness is unknown. A comparative study of the behaviour of islets and cultured $\beta$-cells of lean and obese mice was undertaken to investigate the cause of the hyper-responsiveness of the $\beta$-cell of the obese mouse. Insulin secretion studies, electrophysiological studies and spectrofluorimetric studies using lean mouse islets and $\beta$-cells were conducted to study the regulation and modulation of the ionic events of the normal insulin secretory pathway. Such studies have never been conducted on lean mice of the C57B1/6 strain and hence provide a basis for comparison for work with the obese mouse and for future work on insulin secretion mechanisms in "normal" mouse $\beta$-cells. The results obtained in the lean mouse model were consistent with the currently accepted hypotheses for the role of K$\sp+$ and Ca$\sp{2+}$ channels in the control of insulin secretion. I hypothesized that the hypersecretion in the obese mouse $\beta$-cell was associated with an altered ionic channel. My research centered primarily on the ATP-sensitive K$\sp+$ channel and on voltage-dependent Ca$\sp{2+}$ influx, which depends on the activity of voltage-dependent Ca$\sp{2+}$ channels. Through the use of physiological and pharmacological agents, I found that the defect in the obese mouse $\beta$-cell could not be attributed to an alteration in the fundamental properties of the ATP-sensitive K$\sp+$ channel or in the regulation of voltage-dependent Ca$\sp{2+}$ influx. However, insulin secretion studies revealed that an apamin- and quinine-sensitive channel produced different secretory responses in lean and obese mouse islets. This type of inhibitor sensitivity could not be ascribed to any K$\sp+$ channel presently known to be involved in $\beta$-cell electrical activity. When modulatory mechanisms were compared in lean and obese mouse $\beta$-cells in spectrofluorimetric studies of (Ca$\rm\sp{2+}\rbrack\sb{i},$ it was found that there were differences in membrane repolarization involving a K$\sp+$ channel that also has a novel sensitivity to inhibitors (charybdotoxin, forskolin, TEA and tolbutamide). From the interpretation of the data the model proposed to explain hypersecretion in obese mouse $\beta$-cells is that in these $\beta$-cells there is the expression of an apamin-sensitive K$\sp+$ channel, which is not present in lean mouse $\beta$-cells, and that as a result of its altered structure or regulation, this apamin-sensitive K$\sp+$ channel is responsible for the altered behaviour of the obese mouse $\beta$-cell.

We were able to demonstrate that amylase genes carrying short promoter sequences are still fully functional and show their characteristic pattern of glucose repression in transformed larvae. Amy (amylase) promoter sequences linked to the coding sequences of unrelated genes, the luciferase gene of the firefly, or the Adh (alcohol dehydrogenase) gene of D. melanogaster, mediate a pattern of tissue-specificity and glucose repression typical of amylase. A reciprocal gene construct, in which Amy coding sequences are controlled by upstream sequences of the Adh gene (Adh-Amy hybrid construct) confirmed that Amy coding sequences do not contribute to glucose repression or tissue-specificity. Amylase promoter sequences were further analyzed in order to localize promoter elements mediating expression and glucose repression. A deletion analysis and site-directed mutagenesis of the Amy-1 proximal gene showed that 109 bp of upstream sequences are sufficient for full expression, and that a sequence element essential for gene expression is present between $-$109 and $-$82; no single region was found to be responsible for glucose repression in these tests indicating the possibility that multiple elements mediate the glucose response. The study of the distal amylase gene indicated that a short promoter sequence is also sufficient to control expression of this gene. A deletion analysis combined with DNA sequence comparisons revealed similarities between the proximal and distal promoters, although with some variation in the position of sequence elements important for gene expression. The DNA sequence comparison of the coding regions of proximal and distal amylase genes from the same chromosome allowed us to uncover unexpectedly high levels of nucleotide similarity, suggesting the occurrence of concerted evolution. (Abstract shortened by UMI.)