Genes encoding the starch-degrading enzyme, alpha-amylase, are found in all major groups of animals, plants and microbes. In this thesis, amylase-coding sequences have been chosen as a model system to investigate the patterns of molecular evolution in an enzyme-coding gene. Previous phylogenetic comparisons of amylase-coding sequences have shown high levels of primary sequence conservation over long evolutionary periods. The studies described here have concentrated on the evolution of these genes within the genus Drosophila. I have studied patterns of genetic variation within populations of a single species, Drosophila melanogaster, using two techniques: (i) allozyme variation and (ii) restriction fragment length polymorphisms (RFLPs). In order to get more precise information on the intraspecific patterns of variation, I have also isolated and partially sequenced the amylase genes from a wildtype strain of D. melanogaster. This sequence was compared to the known sequences which have already been described for laboratory strains. The work was extended to interspecific comparisons by studying amylase sequences from species that were closely-related to D. melanogaster (in this case D. erecta), and also a distantly-related species (D. virilis). This involved the isolation and sequencing of the amylase gene from a D. virilis genomic library. The results revealed several interesting patterns in the evolution of: (i) the primary gene sequences (e.g., gene conversion and codon bias), (ii) gene structure (e.g., changes in intron frequency and location) and (iii) gene organization (e.g., variation in the number of gene copies). These results, concerning both short-term and long-term patterns of molecular evolution within the genus Drosophila, are discussed in the context of what is currently known about patterns of molecular evolution in general, and about the molecular evolution of amylases in particular.

A chimeric gene consisting of the human cardiac actin promoter linked to the lac Z reporter gene encoding $\beta$-galactosidase ($\beta$-gal), was permanently transfected into P19 cells. Differentiation of these cultures into cardiac muscle resulted in high levels of $\beta$-gal activity, while differentiation along the neuroectodermal lineage did not lead to any increase in the expression of the actin-lacZ chimeric gene. The transcription of various muscle-specific genes can be activated in different nonmuscle cell lines by the expression of MyoD, a regulatory factor of skeletal myo-genesis. MyoD activates muscle-specific genes in cooperation with other factors which interact directly or indirectly at the various cis-acting regulatory sequences of the target gene. The role of MyoD in muscle differentiation was investigated in permanent transfections of P19 cells with the MyoD expression vector. MyoD transcripts were absent during the de novo differentiation of P19 cells into cardiac muscle. The myogenic differentiation program seems to be controlled by complex interactions among multiple upstream regulatory elements and myogenic factors that are functional only in the appropriate cellular context. (Abstract shortened by UMI.)

Myotonic dystrophy (DM) is the most common inherited neuromuscular disorder of adult life. The genetic defect for DM was identified as an unstable CTG trinucleotide repeat found in the 3' untranslated region (UTR) of a serine threonine p&barbelow;rotein k&barbelow;inase, DMPK. Normal individuals possess 5--35 CTG repeats, typical adult DM patients have repeat sizes ranging from 80 to 1000 while cDM patients have from 1000 to several thousand CTG repeats. This discovery provided a molecular basis to account for large variability of penetrance and age of onset in DM. Work in our laboratory progressed from ascertaining mRNA levels in patient tissues to testing the hypothesis that overexpression of DMPK might cause features of DM. Transgenic mice expressed the human DMPK mRNA and protein in the appropriate tissues and had many features of DM including type I fibre atrophy, central nuclei and ringed fibres. Induction of expression resulted in about three fold higher levels of CTG 99 mRNA over CTG 11 mRNA at 48 hours post induction. Levels of cell death were assayed following induction and CTG 99 cell lines showed a marked level of cell death while CTG 11 cell lines did not. The presence of CTG repeats within mRNA therefore appears to be very problematic for the cell. Furthermore, we found that patient amniocytes and myoblasts are susceptible to staurosporine induced cell death. Taken together, this data suggests that myoblasts expressing the DMPK 3' UTR are prone to cell death in an expression level and repeat length dependent manner. In addition, patient cells were found to be susceptible in much the same way. These experiments also revealed that myogenin levels in vivo were reduced in transgenic embryos compared with wild type embryos suggesting that expression of the DMPK 3' UTR in vivo inhibits accumulation of myogenin and perhaps myogenesis. In adult mice there was consistent muscle atrophy in CTG 91 but not in CTG 11 mice despite much lower expression levels of the CTG 91 transgene. Together, these results indicate that expression of the DMPK 3' UTR by itself may inhibit myogenesis in vivo and contribute to pathological features of DM-like muscle atrophy. (Abstract shortened by UMI.)

The removal of circulating androgens by castration has a marked effect on the physical size, protein, RNA and DNA contents of the rat ventral prostate. This androgen ablation also produces major changes in overall gene expression. Several castration induced mRNAs have previously been identified of which the major sequence is TRPM-2. The cDNA for the rat TRPM-2 has been previously cloned and its protein product characterized. This mRNA sequence is 1,640 bases in length. It codes for a protein of 70-80 kDa, that appears to play an important role in active cell death by blocking complement mediated cell lysis. TRPM-2 is present in a wide variety of species including rat and human. I have isolated genomic clones from two rat and human genomic libraries made from partial Mbo I digests in the EMBL3 vector. I have screened those libraries with the full length cDNA sequence corresponding to the rat TRPM-2, and have isolated overlapping clones which span the TRPM-2 locus in both species. The TRPM-2 gene has been characterized in both species by complete sequence analysis. The gene covers 13.750 bp in the rat and 16.570 bp in human, each having 9 exons. 5,700 bp of upstream sequences have also been sequenced for the rat gene and 1,300 bp for the human. The sequence similarity between the respective coding and 5$\sp\prime$ regions of the gene in both species is nearly 79%. There is also a striking exon/intron structural similarity between the two species. The rat TRPM-2 gene was also shown to be expressed in a wide variety of tissues at very different levels. Finally, the previously published sequence of SGP-2 cDNA is very similar to the TRPM-2 cDNA sequence except for their respective 5$\sp\prime$ ends. PCR analysis has clearly demonstrated that the SGP-2 leader is a cloning artifact.

Haloferax volcanii is a motile extreme halophile, and a detailed physical map of its genome exists. An effort is underway to locate more genes and to increase the detail of this map. Flagellated motile bacteria possess a number of filaments that function as cellular propellers. Motile bacteria swim outward in a low concentration agar medium (swarm medium) and form a swarm. I have isolated twenty-five independent motility-defective mutants and four independent super-motile mutants of Hf. volcanii WFD11. Some of these mutants were characterized by light and electron microscopy. The motility-defective mutants form three characteristic kinds of swarms on swarm medium: non-motile colonies, small fuzzy colonies and variable-size-colonies, Wild-type Hf. volcanii DS2 possesses 4-10 flagellar filaments which form a bundle. In stationary phase culture, some of these peritrichously flagellated cells become elongated and polarly flagellated. Light and electron microscopy of two mutants revealed that these cells that form non-motile colonies are non-motile, although they possess normal looking flagellar filaments. These mutants may have paralyzed filaments. Revertants of these mutants may produce dotted and/or scalloped-edge swarms. Mutants that form small fuzzy colonies were found to be cell cycle defective mutants which were not able to undergo normal cell division and thus continued growing until they formed very long spaghetti-like cells. These Hf. volcanii mutants have normal flagellar filaments and are motile when they are short but as they grow longer, they become less and less motile. The third group of mutants with variable size colonies, were normal looking motile cells under the light microscope. Super-motile mutants possess tens of flagellar filaments. Three independent Hf. volcanii DS2 genomic shot-gun libraries were constructed and amplified. These shot-gun libraries are presently being used to genetically complement Hf. volcanii mutants by a PEG-mediated transformation procedure. (Abstract shortened by UMI.)

The carAB operon from E. coli hybridized with the gonococcal clones carrying carA or carB genes under conditions of high stringency (i.e. detecting approximately 80% or greater similarity) suggesting that the nucleotide sequence of the carbamoylphosphate synthetase genes is very similar in these two organisms. Under these conditions for hybridization, the gonococcal clones carrying argB or argF genes did not hybridize with plasmids containing the corresponding E. coli gene. Hybridizations performed under conditions of lower stringency indicate that the nucleotide sequence of the argB gene is less conserved than that of argF in E. coli and N. gonorrhoeae. Co-complementation experiments established gene linkage only between carA and carB and between proA and proB gonococcal genes. Clones complementing a gene defect in argE were also able to complement an argA mutation. This suggests that the enzyme ornithine acetyltransferase from N. gonorrhoeae (encoded by argJ) may be able to complement both argA and argE mutations in E. coli. The prevalence of specific arginine biosynthesis gene defects was studied for 319 arginine-requiring clinical isolates of Neisseria gonorrhoeae by using the ability of the strains to utilize intermediates of arginine biosynthesis. Nearly 99% of the strains were defective either in the conversion of acetylornithine to ornithine (174 strains) or in the conversion of ornithine to citrulline (141 strains). Based on a nutritional requirement for carbamoyl phosphate, only 11% of the uracil-requiring strains defective in the carbamylation of ornithine to yield citrulline were apparently defective in carAB. Three argininosuccinate-requiring strains (i.e. probably defective in argG) of auxotype PAU were identified. A high polymorphism was observed in hybridization patterns of restricted genomic DNA from N. gonorrhoeae strains having the same auxotype and serotype with a gonococcal CPSase gene-specific probe suggesting that this probe may provide a useful epidemiological marker for N. gonorrhoeae. Some of the arginine auxotrophs of N. gonorrhoeae defective in carAB, argJ, argF or argG were complemented by genetic transformation with DNA from recombinant bacteriophages carrying characterized gonococcal arginine biosynthesis genes. Gene defects in proA (5 strains) and in proB (6 strains) were identified by gonococcal transformation assays with recombinant bacteriophages or plasmids carrying proline biosynthesis genes from N. gonorrhoeae. None of the eleven proline-requiring strains tested appears to be defective in proC. Polymerase chain reaction (PCR) amplifications using oligonucleotides specific to conserved areas of the E. coli carAB operon yielded amplified copies of various portions of the N. gonorrhoeae CPSase genes. Amplifications using primers specific to the duplicated region of the E. coli carB gene suggest that the gonococcal carB gene contains a similar duplication. (Abstract shortened by UMI.)

Estrogens exhibit potent anti-atherogenic effects through mechanisms that may involve direct effects on the artery wall. The existence of the classical estrogen receptor (ERalpha) in vascular tissues has been established. Recently a new estrogen receptor (ERbeta) has been discovered that represents a distinct gene product with homology to the classical ERalpha. The purpose of this thesis was to explore the expression pattern of ERbeta, in normal and diseased arteries. My experimental data demonstrate that ERbeta mRNA is also expressed in human vascular tissue. Using semi-quantitative RT-PCR, there was no relative difference in ERbeta mRNA expression in paired human vascular tissue from normal and diseased arteries. Southern blot and Northern blot analyses, as well as sequencing, were used to confirm these results. We also studied ERbeta protein expression in the normal and diseased coronary arteries using immunohistochemisty. The results further confirm that ERbeta protein is also expressed in both normal and diseased coronary arteries.

Statistical meta-analysis of large and diverse data sets has indicated that amphibians have been declining worldwide since the 1960s. Exposure to UV-B radiation and endocrine-disrupting chemicals (EDCs) have been considered as possible hypotheses to explain the observed declines. Newly-hatched leopard frog (Rana pipiens) tadpoles were exposed for ten days to one of three concentrations of octylphenol (OP) (10 muM, 1 muM, 1 nM) plus a 0.01% ethanol vehicle control, with and without two levels (7 muW/cm 2, 25 muW/cm2) of UV-B radiation. The LC50 for water borne OP was also determined. OP is an EDC which has been shown to elicit responses similar to 17 beta-estradiol (E2). The RNA-arbitrarily pruned PCR (RAP-PCR) differential display strategy was employed to isolate candidate genes differentially regulated in the tadpole diencephalon which were affected by various OP and UV-B treatments. A reverse Northern multiple-gene dot blot was used to verify expression patterns of specific cDNA transcripts cloned from differential display. Homology cloning was performed to obtain R. pipiens GAD65 and GAD67, enzymes responsible for gamma-aminobutyric acid (GABA) synthesis, and these were included on the dot blot. (Abstract shortened by UMI.)

This study addresses the problem of amino acid compositional bias in molecular phylogenetic reconstruction based on protein sequences. Compositional bias is a factor that has largely been left out of conventional phylogenetic analyses, but there is now reason to believe that it is a significant factor. It is shown that in animal mitochondria homologous genes that differ in AT/GC content code for proteins differing in amino acid content in a manner which reflects the AT and GC content of the codons. This relationship is also seen in nuclear and bacterial genes, and is significant even in the highly-conserved hsp70 and $ef 1\alpha / tu$ genes. This bias in the amino acid content of proteins can result in incorrect protein-based phylogenetic trees. A striking example is presented where common phylogenetic tools fail to recover the correct tree from animal mitochondrial protein sequences. The two taxa with the greatest compositional bias continually group together in these analyses, despite a lack of close biological relatedness. It is concluded that, while protein-based phylogenetic analyses have been favoured over analyses using biased DNA, even protein-based trees can be misleading. Current models of protein sequence evolution used for phylogenetic reconstruction assume that the amino acid composition is stationary. In order to accommodare evolution involving compositional change, directional mutation probability models are proposed and their properties described. In simulation studies with biased and non-biased branches on the same tree, the biased branches "attract" in subsequent analyses, paralleling results about using real mitochondrial sequences. Improved methodology is proposed which can use both stationary and directional models in phylogenetic reconstruction, and which can use different models on different branches of the same phylogenetic tree.

Genetic linkage analysis of two Simpson-Golabi-Behmel syndrome families in our lab has been conducted. The closest linkage to SGBS was observed for the chromosome Xq26 locus HPRT. SGBS-Xq marker recombinations map the disease locus to the Xqcen-DXS425-SGBS-DXS1123-Xqter interval on Xq25-q27. The close linkage between the disease locus and the HPRT locus strongly suggests that the causative gene for SGBS is located in the chromosome Xq26 region and is disrupted by the translocation. SGBS shows significant clinical overlap with another overgrowth disorder called Beckwith-Wiedemann syndrome (BWS). Both disorders demonstrate an increased risk for developing embryonal tumours. Two genes from the BWS linked chromosome 11p15 region, IGF2 and H19, are known to normally undergo parental imprinting. Relaxation of this imprinting has recently been demonstrated in some BWS cases as well as in some instances of isolated Wilms' tumour (WT). Since SGBS and BWS have been postulated to result from distinct defects in a common pathway, we have studied the methylation pattern and transcriptional activity of IGF2 and H19 in isolated SGBS +/$-$WT tissue. No consistent methylation abnormalities were observed in genomic DNA isolated from SGBS leucocytes, placenta, skin fibroblasts or cleft lip tissues. Genotyping of H19 cDNA polymorphisms showed biallelic expression in both SGBS and normal placentas and monoallelic expression in one SGBS fibroblast. However monoallelic expression of IGF2 was seen in control placenta and fibroblast but biallelic expression of IGF2 was shown in one SGBS placenta, and two SGBS fibroblasts tissues. It would thus appear that loss of genomic imprinting may play a role in the pathogenesis of SGBS. This is the first example that we are aware of a trans imprinting function for a human gene. (Abstract shortened by UMI.)