RNA viruses are among the most virulent microorganisms that threaten the health of humans and livestock. Among the most socio-economically important of the known RNA viruses are those found in the family Orthomyxovirus. In this era of rapid low-cost genome sequencing and advancements in computational biology techniques, many previously difficult research questions relating to the molecular epidemiology and evolutionary dynamics of these viruses can now be answered with ease. Using sequence data together with associated meta-data, in chapter two of this dissertation I tested the hypothesis that the Influenza A/H1N1 2009 pandemic virus was introduced multiple times into Africa, and subsequently dispersed heterogeneously across the continent. I further tested to what degree factors such as road distances and air travel distances impacted the observed pattern of spread of this virus in Africa using a generalised linear modelbased approach. The results suggested that their were multiple simultaneous introductions of 2009 pandemic A/H1N1 into Africa, and geographical distance and human mobility through air travel played an important role towards dissemination. In chapter three, I set out to test two hypotheses: (1) that there is no difference in the frequency of reassortments among the segments that constitute influenza virus genomes; and (2) that there is epochal temporal reassortment among influenza viruses and that all geographical regions are equally likely sources of epidemiologically important influenza virus reassortant lineages. The findings suggested that surface segments are more frequently exchanges than internal genes and that North America/Asia, Oceania, and Asia could be the most likely source locations for reassortant Influenza A, B and C virus lineages respectively. In chapter four of this thesis, I explored the formation of RNA secondary structures within the genomes of orthomyxoviruses belonging to five genera: Influenza A, B and C, Infectious Salmon Anaemia Virus and Thogotovirus using in silico RNA folding predictions and additional molecular evolution and phylogenetic tests to show that structured regions may be biologically functional. The presence of some conserved structures across the five genera is likely a reflection of the biological importance of these structures, warranting further investigation regarding their role in the evolution and possible development of antiviral resistance. The studies herein demonstrate that pathogen genomics-based analytical approaches are useful both for understanding the mechanisms that drive the evolution and spread of rapidly evolving viral pathogens such as orthomyxoviruses, and for illuminating how these approaches could be leveraged to improve the management of these pathogens.

Background Outbreaks of community-acquired Pseudomonas aeruginosa are typically small and localized. We investigated an increase in P. aeruginosa clinical isolates in Cape Town, South Africa during a severe drought. Methods Cases were defined as P. aeruginosa isolated from any clinical sample, and “wild-type” as susceptibility to all antibiotics tested. Residential addresses of community-acquired wild-type cases were mapped. Whole genome sequencing and multi-locus sequence typing were used to determine clonality and identify virulence genes. A modified case-control study in a subset of patients with bloodstream infection compared demographic and clinical characteristics between sequence types. Results The outbreak lasted 10 months from December, 2016 to September, 2017 with 3,321 documented cases. At the peak, cases reached 2.3-fold baseline and the city's dams reached a nadir of 19% capacity. Cases were distributed widely across the city. Multi-locus sequence type (ST) 303 was found in 27 of 42 (64%) blood culture isolates of P. aeruginosa during the outbreak, one of 19 (5%) before, and none of 11 after. ST303 infection was independently associated with younger age, but not with co-morbidities nor increased mortality. Fifty-one virulence genes were differentially present in ST303 compared with other sequence types, including genes involved in biofilm formation, iron uptake, and gut penetration. Conclusion The investigation confirmed a citywide outbreak of P. aeruginosa coinciding with and potentially related to a severe drought. We identified a predominant outbreak-associated clone, ST303, which harboured genes that could contribute to virulence and survival in drought-related conditions. Enhanced surveillance for P. aeruginosa during periods of drought is recommended.

Aim: The purpose of this study is to verify the performance of the Roche Elecsys ® antinucleocapsid (qualitative) and anti-spike (quantitative) SARS-CoV-2 immunoassays to determine whether the performance of the assays is acceptable for diagnostic use in the Groote Schuur Hospital virology/chemistry laboratory, as well as other National Health Laboratory Service (NHLS) laboratories in South Africa. Methods: We performed a verification study using de-identified remnant serum or plasma samples. Standard verification experiments including sensitivity, specificity and precision were performed. Pre-pandemic samples were used to assess specificity. Samples with a linked positive SARS-CoV-2 polymerase chain reaction (PCR) result on a respiratory sample >10 days before the serum/plasma collection date were used to assess sensitivity. Additionally, postvaccine humoral response and other parameters was assessed in a cohort of laboratory staff. Results: For the anti-nucleocapsid antibody assay, specificity was 99.7% based on 316 samples and sensitivity 91.3% based on 404 samples. For the anti-spike antibody assay, the specificity based on 194 samples was 100%, and the sensitivity based on 384 samples was 93.8%. Both assays demonstrated acceptable precision. Furthermore, the anti-spike antibody assay sensitivity was >92% during the first three waves in South Africa, dominated by different SARS-CoV-2 variants. Post-vaccine seroconversion in 115 staff with no evidence of prior natural infection was 99% and hybrid immunity produced higher anti-spike antibody titres compared to vaccine-only participants. Conclusion: Both immunoassays met our acceptance criteria. Both assays can be used for seroprevalence studies. The anti-nucleocapsid immunoassay assay is valuable in confirming past natural infection in patients with previous asymptomatic infection, previous symptomatic infection where no PCR was done or PCR-negative patients who present to hospital with COVID-19 during the second week of illness or later. Most importantly, the antispike immunoassay can be used as a reliable, cheap, and easily accessible surrogate marker of post-vaccine humoral immune response and we recommend using this to confirm and monitor humoral immune response in patients with risk factors for non-seroconversion following vaccination and increased risk for morbidity and mortality following infection with SARS-CoV-2.

A biological method for peptide synthesis provides increased production capacity of inexpensive peptide pharmaceuticals with environmentally safe procedures relative to current chemical peptide synthesis. Most precursor peptides are readily produced from yeast and bacterial systems using recombinant DNA technologies but require C-terminal amidation for maximum biological activity. Peptidylglycine α-amidating monooxygenase (PAM) is the only enzyme that catalyses the C-terminal amidation of peptides in vivo through its two catalytic cores, peptidylglycine α-hydroxylating monooxygenase (PHM) and peptidylglycine αamidating lyase (PAL). The cost and limited quantities of the commercial PAM variants available have necessitated research into low cost, scalable quantities of PAM and peptide amidation to enable inexpensive biological peptide production. In the present study, an assay for measuring the product of PAM activity, glyoxylate, was developed based on a 2-aminobenzaldehyde-glycine-glyoxylate (AGG) absorbance assay. The AGG chromophore synthesised was identified with ultra-performance liquid chromatography mass-spectroscopy (UPLC-MS). PAM activity was measurable with glyoxylate between 25 µM and 1600 µM with the AGG assay. Furthermore, the activity of PHM alone was measured by the inclusion of an alkaline hydrolysis step to lyse glyoxylate as a substitute for PAL catalytic activity. Multiple candidate proteins and DNA sequences for PAM were identified by genetic sequence searches and a novel fungal PHM modelled in silico. Fourteen PAM, PHM, PAL and truncated constructs were expressed in the non-conventional yeast host, Yarrowia lipolytica. The novel fungal PHM's nutrient, temperature, and pH conditions were optimised to maximise protein expression. Enzyme purification was optimised with scalable industrial appropriate methods to purify milligram amounts of fungal PHM. The AGG assay was validated with a commercially obtained PAM, demonstrating a simple medium-throughput method to measure PAM activity. The novel fungal PHM was characterised with a pH optimum of 4.0 and maximum enzymatic activity at 45°C. Deglycosylation of fungal PHM enhanced enzyme activity by 1.83 fold, but lowered the temperature optimum to 37°C. The novel PHM and alkaline hydrolysis catalysed the conversion of the peptide pharmaceutical precursor for exenatide into its final bioactive form.

Includes abstract. / Includes bibliographical references (leaves 93-109). / Nearly 130 years have elapsed since the discovery of Mycobacterium tuberculosis (M.tb), the causative agent of tuberculosis, yet today it is estimated that 1 in every 3 of the worldʼs population is infected with this pathogen. In South Africa alone there were approximately 1000 new TB cases per 100 000 population in 2007, ranking the country second in incidence rate, globally. Hence research into new vaccine strategies to control the epidemic is vital. Current vaccines under development are prophylactic and designed to boost pre-existing immunity induced by the only licensed TB vaccine, BCG. A new approach is the development of a post-infection vaccine aimed at inducing an immune response that prevents progression to TB disease when administered to individuals latently infected with M.tb. This vaccine would have a dramatic impact on the worldwide TB burden. Our objective was to address 2 areas in TB vaccinology, firstly a novel postinfection TB vaccine strategy, and secondly, optimal measurement of vaccineinduced responses using a new immunological assay. The aim of the first study was to investigate human T cell responses to antigens that have been associated with M.tb latency. Rv2660 and Rv2659 were investigated, as these antigens are candidate antigens for a postinfection vaccine based on findings from in vitro models of M.tb suggesting preferential expression during latency in vivo. No information exists on the immune response to these antigens in M.tb infected or TB diseased individuals. Hence, we investigated the immune recognition of Rv2660 and Rv2659 in these 2 groups, and further characterised the nature of these antigen-specific T cell responses. We observed that (i) these antigens are significantly more likely to be recognised during M.tb infection compared with TB disease as shown by measurement of soluble IFN-γ in response to the specific antigens, (ii) M.tb infected persons had greater Rv2660- and Rv2659- specific CD4+ T cell proliferation and associated cytokine expression compared, with TB diseased persons. We propose that Rv2660 and Rv2659 may be candidates for incorporation into a post-infection vaccine.

Includes bibliographical references (leaves 75-91).

Lymphoblastic leukemia 1 (Lyl1) is a well-studied transcription factor known to exhibit oncogenic potential during various forms of leukemia. Since its discovery in 1989, many reports have been published describing its relationship with cancer as well as demonstrating its function during hematopoiesis. Lyl1 has been shown to serve a significant role during thymopoiesis by contributing to T-cell development. However, it has been recently reported that irrespective of its significance during T-cell development, mature comparable single positive T-cells are observed in mouse models. The use of murine models has been crucial in identifying potential targets for host-directed therapies (HDT) which has been shown to provide great potential in treating tuberculosis (TB). It is evident that Mycobacterium tuberculosis (Mtb), the causative agent for TB, is capable of developing resistance to various treatments that target the bacterium itself. Therefore, by designing therapies that directly target host factors could assist in circumventing Mtb resistance. By analyzing Mtb-infected bone marrow-derived macrophages (BMDM) that have been subjected to genome-wide transcriptional deep sequencing of total RNA using a single molecule sequencer in conjunction with the cap analysis gene expression (CAGE) technique, various differentially expressed genes were identified, including the oncogenic transcription factor, Lyl1. With the use of murine models, we investigated whether Lyl1 is important for various immunological responses at steady state, the regulation of Lyl1 in response to various immune stimulants including LPS and whether this transcription factor is relevant in bacterial infections including Listeria monocytogenes (Lm) and Mtb. The data in this thesis demonstrate comparable immunological responses, including cellular recruitment by means of flow cytometry and cytokine responses by means of ELISA, between naïve littermate control and Lyl1-deficient mice. Further evaluation of Lyl1 regulation revealed the influence of MAPk and NFκB signaling on Lyl1 expression upon LPS stimulation by significantly downregulating this transcription factor in immune stimulated macrophages. A role for Lyl1 during bacterial infections was observed in Lm-infected mice whereby Lyl1-/- mice succumbed earlier to listeriosis compared to the littermate controls. We further established a functional role for this transcription factor during Mtb infection in vitro and in vivo. The early surrender of Lyl1-deficient mice to Mtb HN878 infection, accompanied by increased bacterial burden during chronic Mtb infection, demonstrated enhanced susceptibility in the absence of Lyl1. We show that Lyl1-deficient host susceptibility is a consequence of enhanced inflammatory responses and increased bacterial growth. This is demonstrated by increased neutrophilic inflammation, pro-inflammatory cytokine and chemokine secretion along with a reduction in anti-inflammatory cytokine release during chronic Mtb infection. Here, we demonstrate the first non-leukemia role for Lyl1 by suggesting a role and requirement for this transcription factor during bacterial infections. Given the significant role during Mtb infection, our studies suggest the use of Lyl1 associated pathways as a potential HDT target for TB.

Southern Africa is disproportionately affected by HIV-1, with Botswana being among the most affected countries. The interindividual heterogeneity in susceptibility or resistance to HIV-1 and progression upon infection is attributable to, among other factors, host genetic variation. Characterisation of human genetic variations can contribute towards understanding the genetic aetiology of HIV-1 and foster development of novel preventive and treatment strategies against HIV-1. Despite the high burden of HIV-1 in Botswana, the population of Botswana is significantly underrepresentation in genomics studies of HIV-1. Furthermore, the bulk of previous genomics studies evaluated common human genetic variations, however, there is increasing evidence of the influence of rare variants in the outcome of diseases which may be uncovered by comprehensive complete and deep genome sequencing. This research aimed to characterise human genomic variations of Batswana in order to elucidate mutation burden, assess population structure and evaluate the role of these genomic variations in susceptibility to HIV-1 and progression through bioinformatics analyses. Whole genome sequences (WGS) of 265 HIV-1 positive and 125 were HIV-1 negative unrelated individuals from Botswana were computationally analysed. The sequences were mapped to the human reference genome GRCh38. Population joint variant calling was performed using Genome Analysis Tool Kit (GATK) and BCFTools. Variant characterisation was achieved by annotating the variants with a suite of databases in ANNOVAR. The genomic architecture of Botswana was assessed through principal component analysis and structure analysis and FST. Cumulative effects of rare variant sets on susceptibility to HIV-1 and progression (CD4+ T-cell decline) were determined with optimized Sequence Kernel Association Test (SKAT-O). Functional analysis of the prioritized variants was performed through gene-set enrichment using databases in GeneMANIA and Enrichr. Variant characterization revealed 24 damaging variants with the most damaging variants being ACTRT2 rs3795263, HOXD12 rs200302685, ABCB5 rs111647033, ATP8B4 rs77004004 and ABCC12 rs113496237. There was admixture of Khoe-San, Niger-Congo and European ancestries observed in the population of Botswana, however, there was no evidence of overall substructure among the HIV-1 positive/negative individuals of Botswana, indicating similar genetic exposure among HIV-1 samples. No variant set was significantly associated with susceptibility to HIV-1, while sets of novel rare-variants within the ANKRD39 (8.48 x 10- 8 ), LOC105378523 (7.45 x 10-7 ) and GTF3C3 (1.36 x 10-6 ) genes were significantly associated with HIV-1 progression. Functional analysis revealed that the variants affected several pathways including chemokine signalling, glycolysis, glycosylation, HIV-1 and host receptor glycoprotein biosynthesis, intracellular transport of molecules and transcription pathways. These findings highlight the significance of whole genome sequencing in pinpointing rare variants of clinical relevance. This PhD thesis unravelled novel genes and novel rare variants that are putatively linked to HIV-1 progression. The thesis contributes towards a deeper understanding of the host genetics HIV-1 and offers promise of population specific interventions against HIV-1.

Oculocerebrorenal or Lowe Syndrome (OMIM #309000) is an X-linked recessive condition characterized by a triad of congenital cataracts, proximal renal tubular dysfunction, and variable central nervous system involvement. Nearly all affected boys will be hemizygous for a pathogenic variant in the OCRL (NM_000276.4 c.2615delC) gene. We present a clinical and molecular characterization of an extended multiplex family of three affected boys with Lowe Syndrome and describe a novel variant, predicted to be pathogenic, in the OCRL gene. This is to the best of our knowledge the first description of its kind in South African patients and future research into more families with Lowe syndrome will be beneficial.

Includes abstract. / Includes bibliographical references. / Infectious diseases are endemic in Africa, especially tuberculosis (TB), malaria and human immunodefiency virus (HIV)/acquired immunodeficiency syndrome (AIDS). Genomics research has the potential to improve the health of Africans through identification of genetic markers associated with either disease susceptibility or therapeutic drug response. This project was set to investigate the genetic correlates for drugs associated with mitochondrial toxicity that are used as part of HIV therapy, especially nucleoside reverse transcriptase inhibitors (NRTIs). Toxicity from NRTIs manifests through metabolic diseases such as peripheral neuropathy, lipodystrophy, lactic acidosis and hyperlactatemia but show interpatient variability. Studying African populations is likely to open the door for the population to benefit from novel diagnostic tools and drugs developed on the basis of pharmacogenomics knowledge. In an effort to contribute to this knowledge, the role of variation in mitochondrial DNA (mtDNA) and polymerase gamma (POL-γ) on how patients respond to stavudine-containing antiretroviral therapy (ART) among adult Malawian HIV/AIDS patients was investigated.