The work contained in this thesis was aimed at isolation of quiescent stem cells from both normal and CML samples and characterisation of the properties of these rare populations. To isolate these specific cell populations, enrichment of the stem cell fraction was achieved by selection methods based on the stem cell marker CD34. Positive or negative CD34 selection yielded up to 99% pure CD34+ cell populations. A fluorescent activated cell-sorting (FACS) strategy was developed to allow isolation of quiescent and cycling stem cells. Using Hoechst (HST) and Pyronin Y (Py) to stain the DNA and RNA respectively and gating on the viable, Propidium lodide- (PI)-, CD34+ population, it was possible to isolate viable stem cells in G0, G1 and S/G2/M phases of the cell cycle. Success of the sorting strategy was demonstrated by PCR analysis of differential cell cycle gene expression in the sorted populations. Further development of FACS methodology allowed the division history and stem cell status, as defined by CD34 expression, to be monitored under a range of conditions. Using this experimental strategy, it was shown that a number of CML stem cells were able to survive exposure to concentrations of ST1 in excess of those achievable in vivo, and that the surviving cells were undivided or quiescent. In an effort to further characterise quiescent stem cells, the HST/Py FACS strategy was used to isolate quiescent and cycling stem cell populations from both normal and CML samples, which were then processed for microarray studies. The results have shown a differential gene expression profile between normal G0 and cycling cells that fully validates the sorting strategy. The most significant different gene expression was in a set of chemokine genes, which were up-regulated in both normal and CML G0 cells compared to cycling cells. This has not been previously reported and represents a novel finding. Additional data analysis is ongoing and may, in time, yield therapeutic targets for CML disease eradication.

Rett syndrome (RTT), a disorder caused almost exclusively by mutations in the X-linked gene, MECP2, has a clinical phenotype thought to be primarily of neurological origin. Disruption of Mecp2 in mice results in a prominent RTT-like phenotype and Mecp2 knock-out animal models provide an excellent platform for investigating the role of MeCP2 in the brain development. In this thesis, I used the Mecp2-stop mouse model to address the effects of MeCP2 deficiency in the central nervous system. First, I assessed the genotype-phenotype relationship at the level of the whole organism by detailed temporal mapping of the RTT-like neurological signs. I also addressed the genotype-phenotype relationship at the level of neuronal networks by assessing alterations in neuronal (cortical) density including potential alterations across morphologically discrete cell subtypes. Finally, I also investigated the genotype-phenotype relationship at the level of the synapse by assessing RTT-related changed in synaptic plasticity. Whole organism phenotyping using observational scoring revealed the male hemizygous Mecp2-mutant mice (Mecp2stop/y) to show an early (from ~5 weeks) and aggressive onset of signs including locomotor and other general features. Correlated studies at the level of the synapse revealed the severity of gross organismal pathology to mirror a progressive decline in both short- and long-term forms of synaptic plasticity as measured in the CA1 region of the hippocampus. Specifically, extracellular field recordings in acute hippocampal slices from strongly symptomatic Mecp2-stop mice showed long-term plasticity (LTP) at the Schaffer-collateral-to-CA1 pyramidal synapse to be 40.2 ± 1.6 % of age-matched wild-type littermate controls. In addition, putative presynaptic short-term forms of plasticity (post-tetanic potentiation (PTP) and paired-pulse facilitation (PPF)) were also decreased in the Mecp2-stop hippocampus (45 ± 18.8 % and 78 ± 0.1 % of wild type for PTP and PPF respectively; all p<0.05). Moreover, the impairment in LTP was associated with symptom severity score whereby mice with a more ‘severe’ symptom score showed a more profound deficit in LTP. Refined axon stimulation protocols revealed evidence of pronounced LTP saturation in symptomatic Mecp2stop/y mice, suggesting an LTP ‘ceiling’ effect. I therefore assessed the action of the weak NMDA receptor blocker memantine, shown previously to reverse LTP saturation-related LTP deficits, in the hippocampus of control and Mecp2-stop mice. Application of memantine (1μM), resulted in substantial reversal of short-term plasticity deficits in the Mecp2-stop samples, without affecting plasticity in wild-type mice. However, systemic administration of memantine (30mg/Kg) in vivo did not have any observable effect on RTT-like phenotype at the organismal level including symptom onset, progression and survival. Utilising immunohistochemistry (NeuN) and histological staining (DAPI) to quantify cell density in layer five of the somatosensory cortex, I demonstrated that the symptomatic Mecp2-stop mouse had a higher cortical cell density compared to wild-type controls (1.28 times of whole cortical cell density and 1.41 times of neuronal cell density). Detailed analyses of distinct neuronal subpopulations (parvalbumin-, somatostatin-, calretinin- and calbindin-immunopositive cells) showed that calretinin (CR)- and somatostatin (SOM)- immunopositive cells had a lower cell density in Mecp2-stop mouse somatosensory cortex. However, the distribution patterns of different neuronal subtypes (using the same markers) in Mecp2-stop hippocampus were preserved and were grossly similar to those of WT brains. In summary, in this thesis, I demonstrated that the cell densities of CR- and SOM- positive neurons were altered in the somatosensory cortex in symptomatic Mecp2-stop mice. At the synaptic functional level, I showed both short-term and long-term plasticity deficits in the hippocampus of Mecp2-stop brains. Memantine, a clinically widely-used Alzheimer drug, partially restored the synaptic plasticity deficits in vitro. These data together supported that deficits in specific neuronal populations and progressive functional synaptic impairment may be key features in the RTT brain and also demonstrated the potential for the pharmacological restoration of synaptic plasticity function.

CHARGE association, or syndrome as it is now known, is a condition where a number of congenital malformations are non-randomly associated in a recognizable pattern. There are two sets of diagnostic criteria for CHARGE syndrome which are in common usage at present (Blake et al., 1998; Verloes, 2005). The etiology of CHARGE syndrome was unknown. We identified twin girls with CHARGE syndrome and a de novo apparently balanced chromosome translocation 46,XX,t(8;13)(q11.2;q22). By mapping the chromosome translocation breakpoints we found that the gene chromodomain-helicase-DNA-binding protein 7 (CHD7) located at 8q12 was disrupted in these girls. CHD7 has a genomic length of 188kb with 9000 coding bases over 37 exons. It has a putative function as a transcription factor which makes it a good candidate gene for a condition which affects multiple body systems. Concurrently with this study Vissers et al (2004) identified CHD7 as a cause of CHARGE syndrome. They found two individuals with CHARGE syndrome with overlapping microdeletions detected by array CGH. By sequencing the 9 genes in this region in a cohort of 17 cases they identified a mutation in CHD7 in 10 cases. We ascertained a cohort of 45 patients with a diagnosis of CHARGE syndrome or possible CHARGE syndrome by scrutinizing the clinical genetics databases in Glasgow and Sheffield. Part of the cohort was accessed by receipt of samples from clinical genetics departments elsewhere in the U.K. and in Lisbon. Clinical information was acquired on this cohort either by examination and review of the clinical notes by the author or by completion of a proforma by the referring clinician. Sequencing in this cohort of 45 patients was successful in 43 individuals. We identified 28 mutations; 16 nonsense, 10 frameshift and 2 splice site mutations. 20 of the mutations were novel, 8 had been reported in other studies. The mutations were found throughout the gene with no particular hotspots. No genotype/phenotype correlations were found either in relationship to the position of the mutation within the gene or with regards to the type of mutation. I have analyzed the phenotype in our cohort and compared it with the cases of CHARGE association reported prior to the availability of mutation analysis. I have also compared the phenotype in our mutation positive cases with those reported in other studies which were mutation-positive. We report two individuals with rare findings in CHARGE syndrome; one with a palsy of the twelfth cranial nerve reported anecdotally only once before (Blake et al., 2008), and another child with a limb reduction defect which has been reported in five other cases (Aramaki et al., 2006; Asamoah et al.,2004; Van de Laar et al., 2007). Our notes review ascertained an incidence of CHARGE syndrome of 1/10,000.

Chk1 is a key regulator of DNA damage response and genome stability in eukaryotes. To better understand how checkpoint proficiency affects cancer development particularly tumours induced by chemical carcinogens in murine skin, I investigated the effect of conditional genetic ablation of chk1. I found that complete deletion of chk1 immediately prior to carcinogen exposure strongly suppressed papilloma formation, and the few, small lesions that did form always retained Chk1 expression. Remarkably, chk1 deletion was accompanied by spontaneous cell proliferation followed by DNA damage and cell death within the hair follicle. This also affected and led to proliferation and ultimately depletion of label-retaining stem cells (LRCs) within the bulge region of hair follicles, the principal source for carcinogen-induced tumours. At later times, ablated skin became progressively repopulated by Chk1-expressing cells and normal sensitivity to tumour induction was restored if carcinogen treatment was delayed. In marked contrast, papillomas formed normally in chk1 hemizygous skin but showed an increased propensity to progress to carcinomas. I conclude that Chk1 is essential for the survival of incipient cancer cells but that partial loss of function (haploinsufficiency) fosters tumour progression.

To identify the full range of targets and the pathogenic consequences, we sought to mimic the pathogenesis of myotonic dystrophy type 1 with temporal and spatial control: temporal to reproduce the developmental pathogenesis of the congenital form, and spatial to isolate tissue specific pathology. To do this, we attempted to use the Cre-lox system for the conditional expression of an EGFP reporter-linked expanded CUG repeat RNA in the mouse. Expression of the transgene was controlled by Cre excision of a transcriptional stop, placed upstream of the EGFP-expanded repeat open reading frame. The transgenes were constructed and tested successfully, and a normal length repeat transgenic line was established. Unfortunately generation of the expanded repeat line was not successful. The constructs were used to generate cell-culture models of DM1, in both human and murine cells, which mimicked the nuclear foci formation and MBNL1 co-localisation seen in patient cells. Expression of exogenous MBNL1/GFP fusion protein in this model resulted in an increase in the size of foci, indicating that MBNL1 protein is limiting within the cell, and may possibly play a protective role. The murine DM1 cell-culture model was used to investigate the effects of expanded CUG repeat expression on splicing within the transcriptome. The differential effect between 5 and 250 repeat RNA expression using Affymetrix whole transcript and exon arrays was compared. Using whole genome arrays, 6 genes were down-regulated and 128 upregulated. With exon arrays, 58 genes showed alternative exon usage. Six genes were selected for further bioinformatics analysis: MtmR4, which has possible neuromuscular involvement; Kcnk4, Narg1, Ttyh1 and Bptf, potentially related to brain development; and Cacna1c, a promising candidate for heart conductance defects and sudden death.

The PDE 11 family of dual specificity phosphodiesterases was first identified in 2000, and has not been well characterised, although mutations in the gene have been linked to multiple disorders, including major depressive disorder, and cancer. DmPDE11 is a dual specificity phosphodiesterase, which shows 96% similarity with the catalytic domain of HsPDE11A, and around 40% similarity along the length of the protein. The focus of this project was to characterise this important enzyme using the model organism Drosophila melanogaster. The resources available to Drosophila researchers are unrivalled, and include a sequenced genome, unparalleled transgenic technology, of which stocks are freely available, and Homophila, a database of human disease genes and their Drosophila orthologues. Drosophila is genetically tractable to an extent not seen in any other multicellular organisms. The genetic dissection of gene function in Drosophila has allowed the identification and characterisation of numerous cell signalling genes. For example, mutations to Dunce were shown to affect olfactory learning. This allowed the identification and cloning of the mammalian dnc homologue PDE4. cAMP (and cGMP) were subsequently shown to modulate learning and memory in mammals. The 5.8 kb expressed sequence tag (EST) SD13096 had previously been shown to contain sequence present in the incomplete PDE11 RA ESTs previously released by Flybase, but also incorporating a 5’ UTR, and an in-frame start codon within two novel 5’ exons. A Northern blot of DmPDE11 RA produced one band of approximately 5.8kb; as this matches the size of the DmPDE11 RA ORF, was accepted that SD13096 encodes the entire PDE11 RA ORF (Day, unpublished). Expression of this EST in S2 cells revealed that the construct produced a protein of the accepted size, and the protein localised to the cytoplasm. However, PDE assays of S2 cell lysate revealed that the enzyme did not appear to encode an enzyme with either cA- or cG-PDE activity. DmPDE11 RA was replaced on Flybase by the new isoforms DmPDE11 RB and DmPDE11 RC, which had two key changes to the RA isoform. Both new isoforms had different N termini, sharing a second exon, with distinct first exons. Furthermore, exon 11 of the RA exon is not present in the newly predicted isoforms. These new isoforms were verified by reverse transcriptase- polymerase chain reaction analysis. In the course of this verification, two further novel isoforms were identified, which shared the novel N termini with the RB and RC isoforms, but include a novel exon/exon boundary within the original exon 19, which results in a truncated isoform. As such the four isoforms were named DmPDE11 RB long, DmPDE11 RB short, DmPDE11 RC long, and DmPDE11 RC short. The open reading frames of these isoforms were cloned from Drosophila cDNA using high-fidelity DNA polymerase and sequenced for fidelity. The open reading frames were tagged with YFP, and this tag was used to verify expression of these isoforms. Each isoform expressed a protein of the predicted size when expressed in Drosophila. DmPDE11 B and C proteins show distinct localisation in the Malpighian tubule, where the long and short isoforms of each isoform display indistinguishable localisations. DmPDE11 B localises to the apical and basolateral membranes, and DmPDE11 C localises to an unknown organelle, or to vesicles. All 4 isoforms were verified as dual specificity cA- and cG- PDEs. The previous finding (Day, unpublished) that DmPDE11 co-immunoprecipitates with cGMP dependent protein kinase activity, and that cGMP dependent protein kinases co-immunoprecipitate with cG-PDE activity, and thus that cG-PDE(s) interact with at least one cGMP dependent protein kinase, directly or indirectly, was investigated. DmPDE11 C long and short were co-transfected in Schneider 2 cells with the cGKs DG1, DG2P1 and DG2P2. Co-immunoprecipitation of these showed that both the long and short isoforms of DmPDE11 C interact with every cGK screened. Time did not permit the application of this protocol to screen DmPDE11 B interaction with the cGKs. Whether this interaction is direct or indirect was screened by peptide array. Peptide arrays were generated representing the sequence of DmPDE11, DG1, and DG2, and proteins were generated fusing fragments of these proteins with HIS6 and Glutathione-S-Transferase tags. These were expressed in E. coli, and verified by western blotting. HIS6 tagged protein expression was shown to be of higher quality, and was thus affinity purified, and used to overlay and probe the peptide arrays for putative direct interactions. When the PDE11 array was overlaid with tagged protein representing the C terminal half of DG1, and the N and C terminal halves of DG2, a putative direct interaction was identified between DG1 and PDE11 on two separate regions of the PDE11 array, which both fell within the sequence of PDE11 represented by the Middle-HIS6 fragment. As such, this was used to probe the PDE11 array. A reciprocal putative interaction was identified on three regions of the DG1 array, representing sequence in both DG1N-HIS6 and DG1C-HIS6 fragments. Unfortunately, although DG1-HIS6 was verified by western blotting at the analytical stage, attempts to affinity purify the protein failed. Time did not permit the probing of the array with DG1N-GST fusion protein, and so further putative interaction sites on PDE11 may remain. The generation of alanine substitution arrays, and subsequent mutagenesis analysis with yeast two hybrid or co-immunoprecipitation would be necessary to confirm this direct protein-protein interaction as bona-fide. The investigation into a putative direct interaction between PDE11 and DG2 did not yield conclusive data, and so further investigation is required. The role of DmPDE11 in immunity was investigated by the use of DmPDE11 RNAi and deletion lines. The DmPDE11 deletion line showed a qualitative reduction in survival in individual survival assays, but when these data were merged a significant decrease in survival compared to controls was seen. However, fly numbers did not permit the inclusion of all of the necessary controls, and so these assays should be repeated with these. However, upon immune challenge, progeny from a DmPDE11 RNAi (line 9) x Act5c (a ubiquitous GAL4 driver line) cross did not show a decrease in survival compared to parental lines. Transgenic Drosophila expressing H. sapiens PDE11A3 were generated. The protein localised to the nucleus at low levels of protein; increased expression led to nuclear exclusion, and localisation to the basolateral and especially apical membranes, with cytosolic localisation also. The work has provided the tools needed to further research PDE11. The implication of this gene as a tumour suppressor gene, and its role in other processes, means that it is of the utmost importance that this enzyme is further characterised.

Inborn errors of metabolism (IEMs) constitute a major class of genetic disorder. Most of IEMs are transmitted recessively, so consanguinity has a huge impact on disease prevalence, particularly in societies like Saudi Arabia, where consanguineous marriage is common. Understanding and treatment are very important in genetic diseases, and simple models would be helpful. Thus, the feasibility of applying the fruit fly, Drosophila melanogaster, as a model for a human renal genetic disease - xanthinuria - was investigated. Xanthinuria is a rare human genetic disease, caused by mutations in xanthine oxidase or molybdenum cofactor sulphurase; in Drosophila, the homologous genes are rosy (ry) and maroon-like (mal), respectively. The new Orbitrap technology of mass spectrometry has the potential to determine levels of many metabolites simultaneously by exact mass, and a major part of this thesis was to investigate the utility of Orbitrap technology in metabolomics of both wild-type and Drosophila mutant. Repeatable significant differences were identified between ry and wild-type flies, which recapitulated painstaking analytical biochemical determinations of the 1950s, but with greater precision. Additionally, completely novel impacts of the ry mutation (on pyrimidine metabolism, the urea cycle and osmolyte biosynthesis) were identified. As expected mal mutants showed more similar changes as ry, but with widespread metabolic perturbations. The online resource, FlyAtlas.org, provides detailed microarray-based expression data for multiple tissues and life-stages of Drosophila. Downstream genes, such as urate oxidase, are utterly tubule-specific. Accordingly, the utility of Orbitrap technology in elucidating tissue-specific metabolomes was also investigated. Additionally, genetic interventions using designed RNAi constructs were also made and validated by QPCR and metabolomics. As urate is a potent antioxidant, survival of urate oxidase knockdowns was tested in vivo, and a significant impact on survival identified. An Affymetrix microarray was performed, comparing ry506 mutant flies against wild-type and differences were identified in a second experiment, the anti-gout drug allopurinol was used to phenocopy the effects of ry. Overall, the thesis showed that Orbitrap technology was highly suitable for metabolomic analysis of both wild-type and mutant Drosophila, and had potential in the analysis of metabolomes of single tissues. The possibility of using Orbitrap-based metabolomicsin human diagnosis is discussed.

This study set out to analyse C. pseudotuberculosis within an environment relevant to that which would be encountered within its natural host. The impact of the availability of iron within the growth environment of numerous bacteria has been widely reported, and an equivalent investigation was conducted to determine whether the same was true of C. pseudotuberculosis. To this end, a novel chemically-defined medium was designed, which supported the growth of C. pseudotuberculosis, but in which the concentration of specific growth factors could be manipulated. Subsequently, iron was shown to be essential for C. pseudotuberculosis growth, and analysis of secreted protein profiles revealed differential expression between low- and high-iron growth conditions. Furthermore, growth experiments conducted in the defined medium revealed that C. pseudotuberculosis is capable of obtaining iron from the host iron-binding proteins, transferrin and lactoferrin. The results presented in this thesis confirm the importance of iron to C. pseudotuberculosis, and reveal the existence of an iron-dependent regulator which is involved in regulating the expression of multiple target genes.

Rapid advancements in experimental techniques have benefited molecular biology in many ways. The experiments once considered impossible due to the lack of resources can now be performed with relative ease in an acceptable time-span; monitoring simultaneous expressions of thousands of genes at a given time point is one of them. Microarray technology is the most popular method in biological sciences to observe the simultaneous expression levels of a large number of genes. The large amount of data produced by a microarray experiment requires considerable computational analysis before some biologically meaningful hypothesis can be drawn. In contrast to a single time-point microarray experiment, the temporal microarray experiments enable us to understand the dynamics of the underlying system. Such information, if properly utilized, can provide vital clues about the structure and functioning of the system under study. This dissertation introduces some new computational techniques to process temporal microarray data. We focus on three broad stages of microarray data analysis - normalization, clustering and inference of gene-regulatory networks. We explain our methods using various synthesized datasets and a real biological dataset, produced in-house, to monitor the leaf senescence process in Arabidopsis thaliana.

Synthetic molecules that target the major groove in a sequence-selective way are a major goal in molecular medicine. Recently a major step has been taken toward achieving this goal: a novel cylinder has been developed that binds strongly into the major groove of DNA. Experimental techniques have provided some information regarding the binding strength and preferred binding sites of the cylinder on DNA. From all the experimental data it is clear that the parent cylinder binds in the major groove and is able to induce dramatic conformational changes in the DNA; these are unprecedented effects with synthetic DNA binders. However, gaining molecular level information in such a macromolecular system is challenging. Molecular dynamics (MD) simulations can provide information at the molecular level that is complementary to experiment and therefore are an ideal way to get a better understanding of this system. In this work we present the results of various MD simulations designed to probe the DNA-cylinder system. We have studied the effect of using CHARMM22 and CHARMM27 as the force-field for the simulations. Results showed that uncomplexed DNA simulated with CHARMM22 was less stable in the B-form than the comparable strand of DNA simulated with CHARMM27. Investigations into the effects of the cylinders charge and shape are also reported. Multi-nanosecond simulations were performed using two related synthetic cylinders, one with two Fe(II) metal centers and the other with two Cu(I) centers, and DNA. Finally the role of DNA within the system was investigated by performing a series of simulations of the cylinders with d(ATATATATATAT)2, d(CGCGCGCGCGCG)2 and d(CGCGCATATACG). Simulations with these DNA strands has only produced one system (CCu2+ with d(ATATATATATAT)) where the cylinder causes a conformational change in the DNA.