Bicoid morphogen is among the earliest triggers of differential spatial pattern of gene expression and subsequent cell fate determination in the embryonic development of Drosophila melanogaster. This maternally deposited morphogen, diffusing along the anterior-posterior axis of the embryo, establishes a concentration gradient which is sensed by target genes. In most computational model based analyses of this process, the translation of the bicoid mRNA is thought to take place at a fixed rate in the anterior pole of the embryo. Is this process of morphogen generation a passive one as assumed in the modelling literature so far, or would available data support an alternate hypothesis that the stability of the mRNA is regulated by active processes? This thesis demonstrates a Bicoid spatio-temporal model in which the stability of the maternal mRNA is regulated by being held constant for a length of time, followed by rapid exponential degradation. With the mRNA regulation, three computational models of spatial morphogen propagation along the anterior-posterior axis are analysed: (a) passive diffusion with a deterministic differential equation, (b) diffusion enhanced by a cytoplasmic flow term and (c) stochastic diffusion modelled by Gillespie simulation. Comparison of the parameter estimation in these models by matching to the publicly available data, FlyEx, suggests strong support for mRNA regulated stability. With a non-parametric Bayesian setting, we have applied Gaussian process regression to infer the mRNA regulation function as a posterior density. With synthetic data obtained from a linear spatio-temporal dynamical system and the experimental measurements (FlyEx), this approach is capable of inferring the driving input. Apart from confirming the validity of a regulated mRNA source, this work also demonstrates the applicability of a powerful non-parametric model of Gaussian processes in a spatio-temporal inference problem. In line with recent experimental works, we have also analysed this model with a spatial gradient of maternal mRNA, rather than being fixed at the anterior pole. Our final work is to analyse the dynamical topology of the gap gene network, which is the major developmental activity, taking place after the establishment and interpretation.

The self-association of guanine bases in a tetrameric square planar arrangement was first determined in the early 1960s. The tetramer, commonly termed the G-quartet, can stack upon other G-quartets to form four-stranded helices termed G-quadruplexes. Bioinformatics studies have revealed that guanine-rich sequences with the propensity to adopt these structures are found in telomeric DNA and throughout the human genome, particularly in gene promoter regions. It is thought that the location of these sequences is not a coincidence and that the folding potential of guanine-rich DNA in vivo may play an important role in biological events such as gene regulation. Repetitive guanine tracts of G-quadruplex-forming DNAs form highly polymorphic structures with parallel or antiparallel strand orientations, depending the on ionic condition and the length of the connecting loops, and can be assembled as inter- or intra-molecular complexes. While extensive research has demonstrated their formation in vitro, there is little direct evidence to support their formation in vivo. With the exception of the single-stranded telomeric DNA, all genomic guanine-rich sequences are always present in the duplex configuration. Therefore, these structures will need to compete with the duplex that is normally generated with the complementary cytosine-rich strand. In order for this to happen would first require the local dissociation of the strands. Negative supercoiling results from the unwinding of the DNA helix and is known to provide energy to facilitate the formation of a number of alternative DNA structures. The work described in this thesis therefore aims to investigate the formation of G-quadruplexes under negatively supercoiled conditions. This was examined by preparing plasmids that contained multiple copies of G-rich oligonucleotides, based on the sequences (G3T)n and (G3T4)n, cloned into the pUC19 vector. The formation of G-quadruplexes within these repeats has been assessed using the chemical probes dimethyl sulphate (DMS) and potassium permanganate, and the single-strand specific endonuclease S1. DMS probing revealed some evidence for G-quadruplex formation in (G3T)n sequences, though this was not affected by DNA supercoiling. However, probing with KMnO4 failed to detect exposed thymines in the loop regions, though there was some supercoil-dependent reactivity in the surrounding sequences, suggesting that this had been affected by the G-rich region. In contrast, the (G3T4)n sequences did not demonstrate protection from DMS, suggesting that G-quadruplex formation had not taken place. Surprisingly, the KMnO4 reactions identified structural alterations around, but not within, the inserted G-rich fragments. S1 nuclease digestions did not detect any structural perturbations in any of the sequences apart from a mutant plasmid containing an inverted quadruplex repeat at the 3’-end. Two-dimensional gel electrophoresis of DNA topoisomers was also conducted to detect any supercoil-dependent B-DNA to quadruplex transitions. Neither the (G3T)n nor (G3T4)n plasmids showed any such structural changes. However, the mutant plasmid did demonstrate some supercoil-dependent changes, though these may correspond to cruciform rather than G-quadruplex formation. These results do not support the suggestion that negative supercoiling can induce the formation of G-quadruplex structures.

The indiscriminate action of reactive oxygen species (ROS), if left unregulated, has long been considered contributory to a range of disease processes within the animal kingdom and is also a factor associated with ageing. Consequently modifying the molecular mechanisms that regulate ROS levels may prove therapeutic and could also positively affect longevity. One of the key components of this machinery is the superoxide dismutase (SOD) family of enzymes which regulate ROS levels by scavenging the ROS superoxide. Mammals have three distinct SOD enzymes each responsible for managing superoxide levels in different cellular compartments. In Drosophila homologues of two of the mammalian SODs, the intracellular (SOD1) and mitochondrial (SOD2) SODs, have been identified and studied extensively demonstrating a clear link between SOD and oxidative protection and survival. Recently the sequence of a third sod gene, homologous to both the relatively poorly characterised mammalian (sod3) and C. elegans (sod-4) extracellular sod, was identified in Drosophila and is also predicted to locate extracellularly (sod3). To date, no (published) work has been carried out to assess the role of sod3 within insects. This thesis reports the molecular and biochemical characteristics of sod3 in Drosophila. Detailed within are the steps taken to clone the sod3 gene which appears to be expressed as two gene products formed by alternative splicing. Furthermore, a combination of gene expression, proteomic and functional analysis of a number of sod mutants was used to: i) reveal sex specific sod gene expression; ii) validate a sod3 hypomorph mutant; iii) indicate a functional role for sod3 in protection against H2O2 induced oxidative stress; iv) suggest a SOD1-SOD3 co-dependency for maintaining Cu Zn SOD activity; v) demonstrate the appearance of genetic modifiers in the sod3 hypomorph. The findings of this report and further studies on the Drosophila sod3 gene should encourage the re-evaluation of the previous work concerning SOD’s influence on disease states and lifespan regulation.

The preimplantation mammalian embryo is sensitive to the environment in which they develop and grow, either in vivo or in vitro. Disturbance to embryo environmental conditions can affect embryo growth. Changes in gene expression is one of the embryo responses to such conditions that may influence developmental potential and phenotype during later gestation. The signalling network Myc/Max/Mxd function as a molecular switch that regulates cell growth, proliferation and differentiation by controlling a common set of genes. RING finger protein 17(Rnf17) enhance Myc activity by sequestering all four members of Mxd family and creating a "Mxd-null phenotype". The MXD family functions as MYC repressor proteins. Two isoforms were described for Rnf17, long (Rnf17L) and short (Rnf17S). Rnf17 is expressed in adult mouse testis and in the mouse preimplantation embryo. The Rnf17 proteins contain RING finger domain and varying copies of the Tudor repeat domain (Tdrd). The aim of my thesis was to study the potential for Rnf17 regulating the Myc/Max/Mxd network during preimplantation development on the phenotype of the blastocyst and foetal testes. First, F9 murine carcinoma cells were used to question whether Rnf17 can modulate Myc responses. Functional assays were developed for Rnf17 over-expression by expression plasmids or Rnf17 knockdown by RNAi construct. The specificity of Rnf17 E3 ligase was investigated by co-transfection of Flagg-tagged constructs for RNF17 and MXD1 in F9 cells. Second, expression of Rnf17 mRNA and subcellular localisation of RNF17 protein were examined in mouse preimplantation embryo and in E17.5 foetal testes under normal condition and in response to maternal protein diets (LPD; 9% casein, HPD; 30% casein and control NPD; 18% casein). Finally Rnf17 knockdown was performed in mouse embryo by siRNA microinjection to point out the role of Rnf17 in embryo development. RNF17 was able to degrade MXD1 by ubiquitinylation pathway in co-transfected Cos-1 cells. Rnf17-SV40 expression construct significantly induced Luciferase activity of c-Myc reporter construct (P≤0.05) when over-expressed in F9 cells, and knockdown of Rnf17 using Rnf17-RNAi construct significantly reduced Luciferase activity of Rnf17-sensor construct (P≤0.05). Rnf17 mRNA is expressed through preimplantation development, in E17.5 foetal testes and in F9 cells. Members of Myc/Max/Mxd are expressed in mouse preimplantation embryo. The expression pattern of RNF17 in F9 cells and mouse embryos was predominantly nuclear and also presented in the cytoplasm. RNF17 was only expressed in the nuclei of Sertoli cells with faint signal in prospermatogonia in E17.5 foetal testes. Knockdown of Rnf17 in microinjected embryos using Rnf17_siRNA reduced expression of Rnf17S, delayed blastocyst formation, and reduced RNF17 protein signal. Maternal diet differentially regulate expression of Rnf17 and c-Myc in mouse blastocyst and E17.5 testes. In conclusion, over-expression of Rnf17 can amplify Myc response in F9 cells. MXD is a target protein for Rnf17 E3 ligase. Knockdown of Rnf17 in mouse embryo delayed blastocyst development. This study implicates that Rnf17 involved in regulating Myc/Max/Mxd signalling network and indicates a role for Rnf17 in blastocyst proliferation and growth. Changes in Rnf17 expression have the capacity to modulate early embryonic development and may contribute to phenotypic changes occurring later in life.

The yellow fever mosquito Aedes aegypti is the main vector for several important arboviral diseases such as dengue, yellow fever, chikungunya and zika. With the advent of genetic control strategies, new species-specific tools have emerged for the control of Aedes aegypti. This thesis describes attempts at building different gene drive systems aiming for both population suppression or population replacement, as well as exploring the possibility of inserting exogenous sequences in the male locus of Aedes aegypti. An underdominance system, consisting of two mutually rescuing killers, was investigated in Drosophila melanogaster. It did not work as expected in the configuration tested. The chosen NIPP1 killer gene could not be upregulated by tTAV when under the control of hsp83, UAS and tetO. tetO and tTAV2 resulted in a lethal positive-feedback loop. Gal4Groucho and LexAGroucho fusion proteins, previously used as corepressors in the literature, were lethal when under the control of the tetO-tTAV system. Males showed the expected feminisation phenotypes, invovling male palp shortening and less feathery antennae, upon Nix knock-out using CRISPR-Cas9. However, subsequent homology directed repair into the Nix gene in the male locus of Aedes aegypti did not succeed even after reiterated injections. Setting out from the hypothesis of Act4 haploinsufficiency in Aedes aegypti, the building of two different gene drive systems was attempted; female-specific underdominance and RIDL with drive. A CRISPR-Cas9 driven act4 knock-in unexpectedly confirmed act4 haplosufficiency in Aedes aegypti. Whilst the initially devised gene drive systems could not function as such, act4 haplosufficiency marked the finding of a new female-specific recessive flightless target (effectively sterile) for use in future population suppression drive systems.

Multi-subunit complexes such as the testis-specific meiotic arrest complex (TMAC) and the dREAM complex are context specific gene regulators, controlling genes involved in spermatogenesis and the G2/M transition respectively. The TMAC and dREAM complexes largely consist of subunits that are the same or paralogous to one another. One shared subunit, chromatin assembly factor 1, is also a component of the nucleosome remodelling factor complex, which has a similar testis gene expression phenotype to a TMAC mutant when its testis specific isoform, NURF301, is mutated. Therefore both complexes are thought to control gene expression, at least in part, through modifying chromatin either directly or through associations with chromatin remodellers. To investigate this further I have employed an unbiased approach for determining the positions of DNA bound proteins in vivo called Chromatin Particle Spectrum Analysis (CPSA) which involves micrococcal nuclease digestion of native chromatin and paired-end mode Illumina sequencing. Strikingly, in the cells which have many genes activated by TMAC, the spermatocytes, the transcriptional start sites of TMAC target genes lack coherent nucleosome positioning, which is a robust indicator of high gene expression in somatic cells. Disruption of TMAC does not decidedly alter this structure, suggesting that TMAC does not influence nucleosome positioning surrounding testis specific transcriptional start sites. In contrast, when analysing dREAM subunit deficient S2R+ cells, dREAM is found to contribute to the depletion of a nucleosome sized particle at the mid-point between divergently transcribed genes. This phenotype is linked with the involvement of dREAM in both enhancer blocking between proximal genes and its interaction with the nucleosome remodelling and deacetylase complex. Overall, I uncover the unique chromatin structure of highly expressed genes in spermatocytes, and implicate dREAM as being involved in nucleosome removal between divergent gene pairs.

A Xenopus borealis genomic library was screened with the 5'-end of the Xenopus laevis N-Cadherin cDNA (DETRICK et a/., 1990). Four groups of clones were isolated that differed in restriction-enzyme digestion patterns. The sequencing of one of these clones, 3-9/4.8BS/pBS, has identified regions of DNA highly homologous to the X.laevis N-Cadherin gene. Accordingly, it is believed that the clone 3-9/4.8BS/pBS contains the 5'-end and promoter region of the X.boreal is N-Cadherin gene. A sequence analysis of this region has shown it to be GC-rich and has revealed consensus TATA-box, CCAAT-box and Spl binding sites. Other possible transcription-factor binding-sites have also been identified, as well as the first exon/intron boundary. A series of promoter-deletions were fused to the bacterial 3- galactosidase gene and micro-injected into X.laevis embryos. The 3- galactosidase staining patterns of whole embryos has visually shown that 1.3Kb of genomic DNA upstream of the translational start-site is sufficient to direct neural-specific transcription of this reporter gene.

The nucleotide sequences of the fusion (F) protein encoding gene of five isolates of subgroup A respiratory syncytial (RS) virus, representing the five lineages associated with current epidemics of RS virus-associated disease, were determined and compared with F gene sequences from laboratory strains. Overall natural variation in the fusion protein gene among subgroup A isolates is low. Amino acid identities of between 97% to 99.5% were deduced. The relationship of the five lineage isolates, based on the sequence of their fusion protein genes, agrees broadly with evolutionary relationships inferred from comparison of their G, SH and N protein genes. There is a region of high amino acid variation at the C-terminal end of the F2- subunit, specifically between residues #101 and #105. None of the apparent amino acid coding changes are located in known epitopes. Using predictive structural models it is suggested that the few amino acid changes observed may alter the fusion protein structure especially in the C-terminal domain of the F2-subunit. The secondary structure of the Fi-subunit is predicted to remain unaltered. It is hypothesised that amino acid variation in the F2-subunit may result in antigenic variation, by altering a potential conformational epitope formed by interaction between the N-terminal region of Fi and the C-terminal region of F2. Induced variation in the fusion protein gene of two candidate vaccine strains was investigated. Temperature-sensitive mutant MAI, derived by intensive mutagenesis from the A2 strain, has two distinctive phenotypes; temperature-sensitivity and a retarded fusion protein mobility in non-reducing gels. Previous analysis of the phenotypic characteristics of tsA1 indicated that the F gene might be the site of the ts mutation. The sequence data derived in this study suggest that the site of ts lesion is located at residue #66 (Glu —> Lys) and that the mobility phenotype is located at residue #102 (Pro —► Ser). Mutant tsA1 exhibits a complex pattern of reversion with two classes of revertant viruses observed; a fully revertant virus which has wild-type growth characteristics but still retains retarded fusion protein mobility; and a partially revertant virus that possesses near wild-type growth characteristics and wild-type mobility. In fully revertant viruses the correction of the ts phenotype has been identified as the reversion of amino acid #66 to the wild-type residue (Lys —> Glu) whilst the coding change at amino acid #102 is retained, resulting in a mobility phenotype similar to that found in mutant tsA1. In partially revertant viruses the coding changes at amino acid #66 and amino acid #102 are retained. In partially revertant viruses the coding change correcting for the mobility phenotype, and partially correcting for the ts phenotype, has been tentatively identified as additional coding changes at residues #103 (Thr —► Ala) and/or #105 (Asn —> Ser). In vitro expression of the fusion gene products of mutant, revertant and wild-type viruses in mammalian cells has confirmed that the mobility phenotype is solely a consequence of changes in the fusion protein gene. Temperature-sensitive mutant MIC is a triple ts mutant derived from the RSS-2 strain. The mutations detected in the fusion gene of mutant MIC are conservative in nature and are not located in known epitopes. Therefore it is unlikely that the coding changes observed in the fusion protein gene account for the ts-phenotype or viral attenuation. It is also thought that the induced mutations have not altered the antigenic properties of the virus.

Tumour necrosis factor receptor-associated periodic syndrome is an autoinflammatory disorder caused by mutations in the tumour necrosis factor receptor-type 1 (TNFR1) gene, leading to misfolding of the TNFR1 protein and a resultant chronic inflammatory condition. This study aimed to explore the whether a more physiologically relevant model of TRAPS could be made by gene editing the cell’s endogenous TNFR1 genes. We utilized two site-specific nucleases, TALEN and CRISPR-Cas9 as gene editing tools. These nucleases are able to introduce double-strand DNA breakage at a pre-determined DNA sequence, resulting in gene modification, by two mechanisms: non-homologous end joining (NHEJ) or homology directed repair (HDR). TALEN pairs were designed and constructed for targeting six sites on exon 2 of TNFR1 in the SK-Hep-1 Cell line. However, despite evidence of correct TALEN construction, gene editing at the target sites in TNFR1 was undetectable by the methods employed (Surveyor mutagenesis assay, DNA sequencing, Flow cytometry). This suggested that our TALEN pairs might be of very low efficiency or be non-functional in vivo. Consequently, the relatively simpler CRISPR-Cas9 system was used as an alternative gene editing tool to target the TNFR1 gene. Two plasmids expressing guide RNAs and Cas9 enzyme were designed to target exon 2 of TNFR1, and were functionally verified in vitro. Both were applied separately to induce gene editing in vivo, firstly by NHEJ, and subsequently by HDR. Analysis of DNA from pooled clones of transfected cells for NHEJ yielded little evidence of successful gene editing by Surveyor assay or pooled DNA sequencing. In silico analysis of exon 2 sequences using TIDE suggested a maximal efficiency of ≤3%. Cells co-transfected with the cas9-guideRNA plasmids and a HDR template, designed to insert the FLAG epitope sequence into TNFR1, was however more successful. PCR detection of FLAG sequence insertion into exon 2 of TNFR1 indicated successful editing had occurred. However, examination of transfected cells by flow cytometry staining for FLAG epitope expression indicated that gene editing was still very low efficiency, ≤ 1.37%, despite high efficiency of transfection. In conclusion, the CRISPR-Cas9 system, in our hands, shows evidence which supports it’s use in gene editing of TNFR1. However, the very poor efficiency of editing detected, suggests much further reagent and process/detection optimization is needed, or more fundamental issues, including accessibility of the TNFR1 gene within the SK-Hep1 cell line need to be addressed, before it’s application to generate disease specific mutations.

Myotonic dystrophy type 1 (DM1) is an extremely variable genetic disorder showing an autosomal dominant inheritance that is characterised by myotonia, insulin resistance, cardiac conduction defects and cataracts. It is caused by a trinucleotide repeat expansion of CTG sequence located in the 3’-untranslated region of the dystrophia myotonica-protein kinase (DMPK) gene on chromosome 19 at q13.3. The severity of symptoms ranges from mild adult onset to severe congenital form. A characteristic clinical feature of DM1 is anticipation phenomenon where disease severity increases and age of onset decreases over successive generations. The DM1 mutation is highly unstable in both the germline and soma, and showed to be an age-dependent, tissue-specific (skeletal muscles comprised the largest allele length of approximately thousand units) and expansion biased. The unaffected level of the repeat sequence falls between ~5-37 repeats whereas the disease associated range starts from ~50 repeats, reaching several thousand units. These properties account for the observed anticipation and contribute toward the tissue-specificity and progressive nature of the symptoms. The manifested phenotypes, symptoms severity and age at onset are extremely variable within and between families. This is mostly accounted for by the progenitor allele length (PAL) passed on from affected parents in addition to the level of somatic instability over time. Though, recent data have shown that additional sequence variations (CCG, CGG variant repeats) within the repeat and immediate flanking DNA are associated with additional symptomatic variation, modified stability and delayed age of onset. Furthermore, individual specific genetic factors have shown to be clustered within and between families as a heritable trait. Therefore, it has been verified that PAL, in addition to individual specific genetic variations are the main modifier of disease onset. More recently, it has been observed that mismatch repair (MMR) genes play a key role in modulating the dynamic of DM1 mutation, and subsequently impact on the age at onset. Therefore, these genes serve as powerful trans-acting modifiers of repeat instability and subsequent severity. Also, sequestration and up-regulation of RNA binding proteins (MBNL1, CELF1 respectively) against the trapped mutant transcripts are the hall mark of DM1 pathogenicity associated with alternative splicing defects that account for the variability of symptoms. Thus, sequence variations within these genes may underlie the genetic and phenotypic variability among DM1 patients. The current diagnostic test for DM1 only provides a qualitative value, and takes no account of the somatic instability and/or the presence of variations within or elsewhere in the genome. Thus, limited prognostic information is delivered to patients and their families. Although more elaborate genotyping approaches that measure the DM1 degree of instability was developed, they remain labour intensive, time consuming and are not suited to routine clinical diagnostics. In this project, we have evaluated the utility of more rapid and higher throughput next generation sequencing (NGS) technologies (Ion PGM and PacBio platforms) to simultaneously sequence the DM1 alleles of the Scottish patients, characterise the immediate flanking variants (5’-extra AAT and 5’-CCG variant repeats), elucidating the possible role of these variants on the DM1 instability, and finally sequencing the potential trans-acting modifiers in a massive customised panel (Ion AmpliSeq). Though, the accurate genotyping of the DM1 allele using NGS method remains challenging and cannot be used at the moment for accurate measurement of allele length. This is due to the sequencing biased nature towards shorter fragments resulting in differences of modal allele length measurement between PacBio and traditional SP-PCR methods. Additionally, Ion PGM platform was not successful at sequencing >20 CTG repeats. To correct for the sequencing biased distribution towards shorter alleles and distinguish between possible somatic variants from sequencing errors, safe sequencing (SafeSeq) method was conducted by tagging each original parental molecule with unique identifier (UID) sequences via PCR followed by sequencing using MiSeq platform. As the UID assignment was successful in tagging different population of repeats lengths, unfortunately we were not able to confidently differentiate between true somatic mutants from possible repeat slippage events in earlier cycles of PCR. Thus, it was decided to modify the incorporation of UID sequences using ligation based approach instead of PCR, and better optimise the method for more accurate results in the future. The identification of the immediate 5’-extra AAT flanking variant of the DM1 allele in a subset of the Scottish DM1 patients with and without CCG variant repeats has led us to speculate the possible presence of a new sub derived DM1 haplotype shared by a recent common ancestor in the Scottish population. In order to address this question, we were able to discriminate the normal allele haplotype of 11-13 repeats from >20 CTG haplotype among 18 DM1 patients whom were previously sequenced by Dr. Saeed Al ghamdi. These data have illustrated the most conserved haplotype around the DM1 allele. Therefore, the corresponding region was included in a customised Ion AmpliSeq sequencing panel for future larger scale haplotype analysis, in order to provide insights about future DM1 prevalence among the Scottish population. The data of this project highlighted the importance of using NGS technologies to characterise the structural pattern of the DM1 allele containing variants that may impact on symptom severity. It also showed the successful sequencing of trans-acting genetic modifiers in massive parallel fashion. Over larger scale analysis, these data could be used for better genotype-phenotype correlation and stratify patients in future clinical trials.