Genome-wide association studies identified a 5' APOC1 single nucleotide polymorphism (SNP), rs4420638 (minor allele frequency (MAF)=0.18) at 19q13.2, with a CAD risk odds ratio (OR) of 1.17 (1.08-1.28) in linkage disequilibrium with apoE4 risk SNP rs429358 (r 2=0.70; OR=1.06 (0.99-1.13), MAF=0.15). Differing OR and MAF led to the hypothesis that rs4420638 risk is partially independent to rs429358. Additional SNPs not in HapMap were genotyped by sequencing, however no strong linkage existed. Genotypically associated traits include serum apoE, as determined by ELISA (rs4420638 AA: 25.5+/-4.8; AB: 44.8+/-3.1; BB: 69.9+/-3.7ug/ml; p=7.27E-07, rs429358 AA: 14.8+/-3.4; AB: 40.1+/-3.1; BB: 67.0+/-3.4ug/ml; p=1.73E-08) and LDL (linear regression rs4420638 p=0.0007; rs429358 p=0.003) but not apoC1. Haplotype analysis indicated that rs429358 risk allele confers a greater CAD risk than rs4420638: rs4420638 risk allele alone is 6% in cases and 4% in controls; rs429358 risk allele is 3% in cases and 0.3% in controls. Likelihood ratio test confirms this conclusion.

The family Poecilopsettidae (Pleuronectiformes) occurs in the Pacific, Indian and Atlantic oceans. A taxonomic revision of all 19 species considered valid up to now was completed. Type specimens and additional material known for all these valid species were examined and measured. One new species is described (Poecilopsetta dorsialta), two species are synonyms of previously described species: Nematops chui is a junior synonym of N. macrochirus and Poecilopsetta megalepis is a junior synonym of P. plinthus. There are now 18 valid species in the family: two in the genus Marleyella, three in the genus Nematops and 13 in the genus Poecilopsetta. The genera Nematops and Marleyella are monophyletic. However, the monophyly of the genus Poecilopsetta remains doubtful because of the absence of unique characters for all 13 species of the genus. The family Rhombosoleidae (Pleuronectiformes) (sensu Norman, 1934) is a group of 19 species distributed in Southern Australia, New Zealand, in the Indo-Pacific Ocean off Western Australia and South America. A data matrix of 87 morphological characters was analysed using cladistic methodology to assess the monophyly of the Rhombosoleidae and to determine the interrelationships within this family. Achiropsettidae and Poecilopsettidae were used as outgroups. The results (16 equally parsimonious trees; 118 steps; CI = 0.669) found the family Rhombosoleidae to be monophyletic on the basis of eight apomorphies. The Rhombosoleidae was not monophyletic with the genera Azygopus, Oncopterus and Psammodiscus and they were removed from the family and added to a data matrix of of pleuronectiform families to assess their relationship within the order. Azygopus is placed in the Achiropsettidae. Psammodiscus and Oncopterus form a trichotomy with a large clade formed by Rhombosoleidae, Poecilopsettidae, Achiropsettidae, Samaridae, Achiridae, Soleidae and Cynoglossidae (6 equally parsimonious trees; 109 steps; CI = 0.477). The family Paralichthodidae (sensu Cooper and Chapleau, 1998a) is the sister-group of a large clade formed by the families exhibiting a bothoid type of caudal skeleton (Scophthalmidae, Bothidae, Paralichthyidae and Pleuronectidae) and by clade III (Psammodiscus, Oncopterus, Poecilopsettidae, Rhombosoleidae, Achiropsettidae, Samaridae, Achiridae, Soleidae and Cynoglossidae).

Neuroendocrine and reproductive systems are strictly regulated by a series of sex hormones, especially 17-beta estradiol (E2). Through specific membrane or nuclear receptors, E2 initiates a series of diverse signaling pathways that regulates the expression of target genes. This transcriptomic output shapes the specific spatial (cell or tissue level) and temporal (seasonal level) E2 actions. In this study, I attempted to define estrogen-related gene expression changes in the neuroendocrine and reproductive systems. Firstly, I targeted the physiological period of E2 action when gonad size is large just prior to spawning (March and April) in the goldfish ( Carassius auratus) model. The effect of the aromatase inhibitor fadrozole and the resultant decline in E2 on neuroendocrine gene expression and reproductive development was determined using microarray analysis. Several regulatory themes for physiological E2 action in fish brain have been revealed from these novel E2 regulated genes, including regulation of the calcium signaling pathway and auto-regulation of nuclear estrogen receptor action. Secondly, I aimed to define the seasonal gene expression characteristics that are associated with hormone profiles, typically E2 blood level change, during a breeding cycle in the goldfish. By using both theoretical and experimental strategies, I have identified a core set of genes in fish neuroendocrine brain that were differentially expressed between physiologically distinct stages including sexually mature prespawning, sexual regression, and early gonadal re-development. Moreover I demonstrated that gene expression changes between stages can be regulated by photoperiod. Thirdly, to further understand the mechanism underlying fadrozole effects on gonadal development, I used a frog model (Xenopus tropicalis) to show that a germline specific piRNA (piwi-interacting RNA) pathway may be involved in the E2/testosterone regulation of gonadal development in the tadpole. Here, I investigated the effect of fadrozole or finasteride (5 alpha-reductase inhibitor), which are known to influence gonadal development, on the gene expression of a piRNA-specific protein Maelstrom (MAEL) and showed both treatments increased MAEL mRNA expression. Moreover, since the specific function of MAEL is unknown, I conducted a bioinformatics analysis to infer its putative function and evolutionary history. This is one case study for our efforts to annotate some functionally unknown genes which are related to or regulated by E2 actions. In conclusion, the physiological stage or seasonal specific gene expression information defined in this study provides a series of new functional insights into regulatory mechanisms of E2 and related hormones in the vertebrate neuroendocrine and reproductive systems.

The yeast mitochondrial genome encodes seven hydrophobic subunits of oxidative phosphorylation enzymes and Var1p, an essential protein in the small ribosomal subunit. Expression of the membrane proteins is dependent on nuclear, mRNA-specific regulatory genes, several of which specify translational activators that recognize sites within the 5$\sp\prime$-untranslated leaders (UTLs) of their target mRNAs. At the onset of this work, it was not known if the expression of the Var1p also requires mRNA-specific regulatory genes. To investigate this and other aspects of Var1p synthesis and function, I developed a novel system in which Var1p is supplied from a recoded gene in the nucleus (VAR1$\sp{u})$ and the VAR1 coding sequence in mtDNA is replaced by a recoded nuclear gene for an arginine biosynthetic enzyme (Arg8p), thus creating a reporter gene designated $var{\it 1\/}{:}{:}ARG{\it 8\/}\sp{m}.$ This system has been used to address the following objectives: (1) Genetic screens were conducted to identify nuclear mutants defective in $var{\it 1\/}{:}{:}ARG{\it 8\/}\sp{m}$ expression. One such gene, SOV1, was identified and cloned. SOV1 is specifically required for the stable accumulation of VAR1 mRNA. (2) To determine whether the targeting information for mitochondrial membrane proteins is contained in UTLs of their mRNAs, I examined the ability of chimeric mRNAs containing the VAR1 UTL to direct expression of COX2 and COX3. Although cells expressing these chimeric mRNAs synthesized both proteins, they were deficient in the accumulation of Cox2p and Cox3p. These data suggest that translation of Var1p is different from that of the membrane proteins, and support the physiological importance of interactions between the translational activators and the 5$\sp\prime$-UTLs of the COX2 and COX3 mRNAs for localizing synthesis of hydrophobic proteins to the inner membrane. (3) Heretofore, it has not been possible to produce respiratory competent $\lbrack rho\sp+\rbrack$ diploids by mating two $\lbrack rho\sp-\rbrack$ haploid petites. An explanation for this lack of functional complementation is that $\lbrack rho\sp-\rbrack$ cells are devoid of small ribosomal subunits and translation cannot be restored without a source of Var1p. I have shown that respiratory competent diploids can be obtained in crosses between two complementing $\lbrack rho\sp-\rbrack$ strains, but only when Var1p is supplied from $VAR{\it 1\/}\sp{u}.$

Genetic and physical mapping of human and animal genomes has been greatly facilitated by the use of chromosome specific DNA libraries. Mapping with libraries specific to a chromosome or chromosomal region increases marker saturation by reducing the gaps resulting from a purely random shotgun approach. This study was undertaken to construct a genetic and physical map of microsatellites on the chicken Z chromosome. This chromosome is the fifth largest in the chicken genome, comprising about 8% of the total, yet very few microsatellites have been mapped to it. DNA originating from the chicken Z chromosome was previously isolated and reported. This was used to construct a small insert library in Lambda ZAP Express, representing 14 chromosome equivalents. This library was screened for microsatellites with an (AC)12 oligo, and positive clones were isolated. Confirmation of the presence of the microsatellite, as well as its approximate location in the insert was accomplished by PCR amplification. Clones with adequate flanking regions were sequenced, and primers for 19 microsatellites were developed. These primers were used to genotype individuals from the East Lansing poultry reference population and a linkage map was constructed. Thirteen markers were scorable and polymorphic in the population. These were combined with 64 existing markers, and the resulting map spans 220 cM with an average spacing of 2.7 cM between markers. The physical location of selected markers were established by fluorescent in situ hybridization (FISH.) Hybridization results enabled the anchoring and orientation of the linkage group along the length of the Z chromosome.

Homologous recombination is generally considered a major mechanism by which cells repair many types of DNA lesions and damaged replication forks. However, if this process is left unchecked, cells often show a hyper-recombination (hyper-rec) phenotype, and are susceptible to large deletions, duplications, or inversions of important genetic information. This dissertation describes two projects aimed at examining molecular mechanisms by which cells regulate homologous recombination. The first shows several 3'-5' exonucleases prevent RecA-GFP loading by destroying potential substrates. It is shown that two genetic pathways exist: one consisting of ExoIII and another comprised of ExoVII, ExoIX, ExoX, and ExoXI. ExoI acts upstream of both of these pathways. Although xthA cells have an increase in DSBs and recB-dependent loading of RecA-GFP, they are viable with a recB mutation and do not display a large increase in SOS expression. The increase in RecA-GFP is also independent of base excision repair (BER). These experiments uncovered that DNA in a population of wild type cells undergoes DSBs and is often repaired in a RecA-independent manner after processing by ExoI and ExoIII. The second project shows the helicase, UvrD limits the number and intensities of RecA-GFP foci. This activity is due to the ability of UvrD to remove RecA from DNA where it is loaded in a RecF pathway-dependent manner. This activity requires ATP binding by UvrD, suggesting that helicase/translocase activity is important for RecA-removal. The hyper-helicase mutation, uvrD303 confers UV sensitivity to cells. Epistasis analyses showed uvrD303 is defective in the recA pathway of UV repair and not in nucleotide excision repair (NER). Surprisingly, UvrD303 does not directly remove RecA after UV, as new RecA-GFP foci appear like in wild type cells. UvrD303 does, however, slightly inhibit SOS induction, and constitutively activating the SOS response restores UV resistance to these cells in a way that is independent of recA overexpression. Furthermore, uvrD303 was capable of suppressing the constitutive SOS phenotype of recA730. These experiments suggested that UvrD303 antagonizes the ability of RecA filaments to induce the SOS response, rendering cells UV sensitive.

During the course of normal DNA replication, replication forks are constantly encountering "housekeeping" types of routine damage to the DNA template that may cause the forks to stall or collapse. One product of this fork collapse is the induction of the SOS response, a coordinated global response to help pause the growth and replication of a cell while DNA damage is addressed and repaired. In E. coli, this response is activated by the formation of ssDNA, to which the RecA protein binds and forms a nucleoprotein filament, which acts as the activator for autocleavage of the LexA transcriptional repressor, which normally represses expression of SOS genes. Damage responses are crucial to maintaining genomic integrity, and are therefore essential to all forms of life, and this type of regulatory system is highly conserved. However, cells have mechanisms for tightly regulating induction of these responses, and can often repair routine damage to their chromosomes without the need to induce SOS. This is chiefly evidenced by the observation that more than 20% of cells in a population have RecA filaments, but less than 1% are induced for SOS. How cells make this decision to induce SOS is the subject of this work. This dissertation describes three projects aimed at examining molecular mechanisms by which cells regulate RecA filaments, and therefore the decision to induce the SOS response. The first examines the disparity between the formation of RecA filaments, as evidenced by RecA-GFP foci, and the induction of SOS in the absence of damage, using a psulA-gfp reporter system. It is shown that there are three independent factors that repress SOS expression in undamaged E. coli cells. These are radA, the amount of recA in the cell, and in some circumstances recX. The first two limit SOS in wild type cells in the absence of external damage, while the third is an additional factor required in xthA mutants, likely due to the fact there are more RecA loading events in these mutants. These factors are thought to change the character and reduce the half-life and persistence of RecA filaments in the cell. The second project shows that suppression of SOS through the use of recA4162 and uvrD303 mutants is substrate and situation-specific. This specificity is demonstrated by the fact that, while both recA4162 and uvrD303 can suppress SOS in the SOS constitutive mutant recA730, recA4162 can only suppress SOS when the signal occurs at replication forks and not at any other place on the chromosome, while uvrD303 appears to suppress SOS with less specificity, and can suppress after UV (shown previously), at induced DSBs, and other places not directly at the replication fork. Here mutants of different replication factors are used that uncouple the replisome and induce SOS to a high degree. The third project determines the factors necessary for loading RecA filaments at the replication fork versus other locations on the chromosome when SOS is induced in the absence of damage, and helps elucidate further mechanisms for induction of SOS at these substrates. It is shown that the sbcB and recJ exonucleases assist in inappropriate RecA filament formation by substrate processing exclusively at replication forks, but not other substrates, likely through mechanisms that are reliant on the activities of the RecA loading factors RecBCD and RecFOR.

Under the Dobzhansky-Muller model, postzygotic isolation results from incompatibility between interacting genes. Evidence points to regulatory networks as a rich source of incompatibilities that impact hybrid fitness. Pleiotropy is a natural feature of regulatory networks because regulatory elements generally have multiple targets. Both pleiotropy and hybrid incompatibility arise due to genetic interactions; therefore we can expect an intimate association between them. In the following chapters, I investigate the relationship between pleiotropy and hybrid incompatibility in the context of regulatory networks. In chapter one, I extend a general network-based study of hybrid incompatibility by incorporating a sequence-based thermodynamic model of transcriptional regulation. In the absence of pleiotropy, hybrid misregulation of a positively selected trait evolves quickly as a consequence of non-recognition or spurious binding in regulatory interactions across species boundaries. In a conserved trait, hybrid incompatibility evolves much slower as a product of compensatory drift. In chapter two, I show that pleiotropy can promote or constrain the evolution of hybrid incompatibility in a regulatory network depending on its fitness landscape, which emerges from the thermodynamic properties of molecular binding. Pleiotropy may promote hybrid incompatibility in accordance with the "selection, pleiotropy, and compensation model" of evolution, in which compensation for the pleiotropic side-effects of adaptation accelerates incompatibility in conserved traits. Pleiotropy can limit the evolution of hybrid incompatibility by constraining change in trans-acting regulatory elements in favor of adaptation at less pleiotropic downstream cis-regulatory targets. Without change in both interactors, incompatibility does not occur under the Dobzhansky-Muller model. In chapter three, I evaluate the hypothesis that pleiotropy facilitates the onset of hybrid incompatibility under antagonistic coevolution, an ubiquitous and persistent source of natural selection. When infectivity and resistance in a host-parasite system are determined epistatically by network interactions, reciprocal selective pressure results in a genotypic chase. This causes pleiotropic mutations to accumulate and be compensated over time, producing intrinsic hybrid incompatibility in both species independent of local adaptation. Thus, cyclical antagonistic coevolution eventually overcomes constraint on pleiotropic loci, facilitating the evolution of regulatory incompatibilities commonly observed in hybrids.

CSE1, an essential Saccharomyces cerevisiae gene was initially isolated in a screen for genes important for accurate chromosome segregation. cse1 mutants have pleiotropic phenotypes including defects in the ubiquitin-mediated degradation of mitotic cyclins and in cell cycle progression. SRP1, encodes a nuclear localization signal (NLS) receptor protein involved in nuclear protein import that is an allele specific dosage suppressor of cse1-1. CSE1 can rescue certain srp1 mutant phenotypes, indicating that Cse1p and Srp1p are functionally related or have roles in similar pathways. New mutant alleles of CSE1 were generated by linker insertion mutagenesis, including a temperature sensitive allele, cse1-2, that causes arrest in G2/M, chromosome missegregation and defective mitotic cyclin degradation. Analysis of CSE1 mRNA and Cse1p indicate that their levels do not change significantly during the cell cycle and that Cse1p is not phosphorylated. Cse1p is located in the nucleus and concentrated at the nuclear periphery, probably in association with nuclear pores. Current evidence supports the model that Cse1p is required for the export of Srp1p from the nucleus. Srp1p improperly accumulates in the nuclei of both cse1-1 and cse1-2 cells. Reporter proteins that contain NLS sequences accumulate in the cytoplasm of cse1-1 and cse1-2 cells indicating that Cse1p function is also necessary for nuclear protein import. Cse1p binds the nuclear transport protein Ran/Gsp1p-GTP via a conserved amino-terminal motif. In addition, the human protein CAS, which is an export factor for importin-$\alpha,$ is 60% similar to Cse1p. Although a physical interaction between Srp1p and Cse1p has not been shown, a two-hybrid screen identified two potential Cse1p-interacting proteins, Scj1p and Yma5p. Scj1p is a DnaJ homologue involved in protein translocation, folding, and complex assembly. Yma5p is a novel non-essential protein with an as yet unknown role in these important nuclear processes.

To study polyhydroxyalkanoate (PHA) accumulation and PHA regulation in Bacillus megaterium, transposon mutagenesis with Tn917-LTV1 was carried out and seventy two mutants that produce more or less PHA than normal were isolated and partially characterized. The chromosomal regions of B. megaterium flanking the Tn917-LTV1 insertions were cloned and sequenced from two PHA over-producing and six leaky mutants. The results showed that Tn917-LTV1 was less than ideal for generating a mutant bank due to the unavoidable occurrence of sibling transposants. Furthermore, the sequence data revealed that chromosomal deletions mediated by Tn917-LTV1 insertions were common. An alternative method was developed using a direct plating procedure and it was shown to circumvent these two problems. One PHA leaky mutant and one overproducing mutant were studied in more detail. The PHA leaky mutant, T4, had significantly reduced levels of PHA accumulation in all media tested and could not form spores. Chromosomal DNA, contiguous with the transposon was cloned and sequenced. Analysis of the sequence data showed Tn917-LTV1 inserted 24-bp upstream of an operon encoding a putative RNA helicase (deaD) gene and a nuclease (nucP) gene. A chromosomal target repeat 5$\sp\prime$-TATTT-3$\sp\prime$ was found on both sides of the insertion which indicated that no deletion was involved in the Tn917-LTV1 transposition. A plasmid carrying the intact RNA helicase operon was constructed and transformed into mutant T4. The transformant recovered the ability to form spores and accumulate wild-type levels of PHA in minimal media. Deletions in either DeaD or NucP did not complement the T4 mutant to the phenotype of VT1660. The results show that both DeaD and NucP are involved in PHA accumulation and are required for spore formation. The DeaD protein of B. megaterium is the first protein in the DEAD-box helicase family that is not essential for the organism. The fact that the growth rate of mutant T4 and VT1660 were similar in both LB and minimal media indicates that this DeaD protein may target specific mRNA, such as those of pha and spo genes. B001S, a PHA over-producing mutant, unlike its parental strain VT1660, produced large quantities of PHA in rich media. The chromosomal DNA flanking the transposon was cloned from both sides of the insertion. The 8.4 kilobase pairs of chromosomal sequence from left end (IR-L side) of Tn917-LTV1 coded for sixteen open reading frames (ORFs). Ten putative products of the sixteen ORFs shared sequence homology with known gas vesicle proteins (Gvp). The 8.4-kb fragment and its deletion derivatives were cloned into pBluescriptIISK and the plasmids were transformed into E. coli. Gas vesicles were formed and observed by phase contrast microscope, differential interference contrast microscopy and electron microscopy. The deletion analysis and sequence comparison with known gas vesicle proteins suggested 14 out of the 16 ORFs formed the gvp operon of B. megaterium VT1660. The deletion analysis showed that 11 genes are the maximum required for the gas vesicle formation in E. coli. The E. coli cells containing gas vesicles showed increased buoyancy. This is the first time that a functional organelle has been transferred to E. coli.