The H9N2 avian influenza virus (AIV) has emerged, spread and established itself in poultry globally, in just under 30 years. During this time, multiple reassortants of H9N2 with increased zoonotic potential have been isolated in poultry and humans, causing a major threat to the economy and global health. Curiously, H9N2 appears to be compatible with multiple Hemagglutinin (HA) and Neuraminidase subtypes, in nature. Here, the aim was to investigate the HA reassortment dynamics of the poultry adapted H9N2 AIV, in a laboratory setting. Firstly, HA subtypes from wild bird isolates were cloned, before being co-transfected with the backbone of a chicken H9N2 AIV. The rescued H9N2 reassortants were titred on cells before the replication kinetics of a subset of the HA reassortants was assessed. The cDNA sequence of seven HA subtypes induced extensive recombination in E. coli, but ultimately ten out of eleven available HA subtypes were successfully cloned. Further, the chicken H9N2 AIV was compatible with all ten HA subtypes, producing infectious viral particles after co-transfection. However, all HA reassortants displayed decreased replicative fitness in MDCK-2 cells, compared to the wild-type virus. Interestingly, HA subtypes with similar genotypes cluster into distinct HA clades and groups, but these HA clades did not correlate with the replicative fitness of the reassortants. This study suggests that poultry adapted H9N2 AIV is compatible with many HA subtypes, highlighting the importance of reducing its spread in poultry, to reduce reassortment opportunities.

A ce jour, les données génétiques et moléculaires se rapportant aux virus influenza de type A (VIs) présents dans la population porcine au Québec sont relativement rares. Pourtant, ces informations sont essentielles pour la compréhension de de l'évolution des VIs à grande échelle de 2011 à 2015. Afin de remédier à ce manque de données, différents échantillons (pulmonaires, salivaires et nasaux) ont été prélevés à partir de 24 foyers dans lesquelles les animaux présentaient des signes cliniques. Ensuite, les souches virales ont été isolées en culture cellulaire (MDCK) ou sur oeufs embryonnés. Les 8 segments génomiques des VIs de 18 souches virales ont par la suite été séquencés et analysés intégralement. La résistance aux drogues antivirales telles que l’oseltamivir (GS4071) carboxylate, le zanamivir (GS167) et l’amantadine hydrochloride a également été évaluée par des tests d'inhibition de la neuraminidase (INAs) ainsi que par un test de réduction sur plaque. Deux sous-types viraux H3N2 et H1N1 ont été identifiés dans la population porcine au Québec. Douze souches des VIs de sous-type trH3N2 ont été génétiquement liées au Cluster IV, avec au moins 6 profils de réassortiment différents. D'autre part, 6 souches virales ont été trouvées génétiquement liées au virus pandémique A(H1N1)pdm09 avec au moins trois profils de réassortiment génétique différents. Le sous-type trH3N2 des VIs est le plus répandu dans la population porcine au Québec (66,7%). La cartographie d'épitope de la protéine HA de sous-type H3 a présenté la plus forte variabilité avec 21 substitutions d’acides aminés sur 5 sites antigéniques A (5), B (8), C (5), D (1), et E (2). Toutefois, la protéine HA du sous-type H1 avait seulement 5 substitutions d'aa sur les 3 sites antigéniques Sb (1), Ca1 (2) et Ca2 (2). Un isolat H1N1 (1/6 = 16,7%) et 1 autre trH3N2 (1/12 = 8,3%) ont été trouvés comme étant résistants à l'oseltamivir. En revanche, 2 isolats du H1N1 (2/6 = 33,3%) et 2 autres du trH3N2 (2/12 = 16,7%) ont révélé être résistants au zanamivir. Dans l'ensemble, le taux de résistance aux INAs et à l’amantadine était compris entre 33,3% et 100%. La présence des VIs résistants aux drogues antivirales chez les porcs ainsi que l'émergence possible de nouvelles souches virales constituent des préoccupations majeures en la santé publique et animale justifiant ainsi la surveillance continue des VIs dans la population porcine au Québec. / Data about genomic variability of swine influenza A viruses (SIV) in Quebec herds are scarce. Yet, this information is important for understanding virus evolution in Quebec from until 2015. Different clinical samples were obtained from 24 outbreaks of swine flu in which animals were experiencing respiratory disease. Samples including lung tissues, saliva and nasal swabs were collected and virus isolation was attempted in MDCK cells and embryonated eggs. All eight gene segments of the 18 isolated SIV strains were sequenced and analysed. Antiviral drugs resistance against oseltamivir carboxylate (GS4071), zanamivir (GS167) and amantadine hydrochloride was evaluated by neuraminidase inhibition assays (NAIs) and plaque reduction assay. Two subtypes of SIV, H3N2 and H1N1, were identified in Quebec pig herds. Twelve SIV strains were genetically related to trH3N2 Cluster IV and at least 6 different reassortment profiles were identified. On the other hand, 6 Quebec SIV strains were found to be genetically related to the pandemic virus A(H1N1)pdm09 and from which three reassortment profiles were identified. Overall, the trH3N2 was the most prevalent subtype (66.7%) found in Quebec swine herds. The epitope mapping of HA indicated that the H3 subtype was the most variable with a possibility of 21 amino acids (aa) substitutions within the 5 antigenic sites A(5), B(8), C(5), D(1) and E(2). However, the HA protein of the H1 subtype had only 5 aa substitutions within 3 antigenic sites Sb(1), Ca1(2) and Ca2(2). One H1N1 (1/6 = 16.7%) and one trH3N2 (1/12 = 8.3%) were identified as strains resistant against oseltamivir. In contrast, two H1N1 (2/6 = 33.3%) and two trH3N2 (2/12 = 16.7%) strains were found to be resistant against zanamivir. Overall, the SIV resistance against antiviral neuraminidase inhibitor drugs was (33.3%). All strains were resistant against the M2 inhibitor antiviral drug, amantadine. The presence of antiviral drug resistance in Quebec swine herds and the possible emergence of new SIVs strains are public health concerns supporting the surveillance of SIVs.

Dans le contexte de la menace pandémique liée au virus influenza A(H5N1), un projet «GRIPPE AVIAIRE ET GRIPPE PANDÉMIQUE » a émergé au sein de LyonBioPôle avec comme objectif le développement d’outils de caractérisation des virus influenza pour la production de vaccins. Pour étudier le réassortiment génétique entre virus influenza, nous avons développé 3 systèmes de génétique inverse : virus humain A(H3N2) et aviaires A(H5N2) et A(H5N1) et produit des virus réassortants de composition déterminée. Leurs capacités réplicatives ont été évaluées par cinétiques de croissance virale sur MDCK avec quantification de la production virale par qRT-PCR temps réel. L’émergence du virus influenza A(H1N1)2009 pose deux questions sur l’acquisition par réassortiment génétique, d’une résistance à l’oseltamivir d’une part ou de facteurs de virulence d’autre part. Nous avons donc développé un protocole de co-infection virale de cellules MDCK pour étudier les constellations de gènes des réassortants entre différents virus: A(H1N1)2009-A(H1N1) H275Y et A(H1N1)2009-A(H5N1). Nous montrons par deux approches différentes, génétique inverse et co-infections virales, que le réassortiment génétique entre souches aviaires et humaines et surtout aviaires et porcines est possible, en privilégiant certaines constellations. Nous rapportons que le virus pandémique peut acquérir la NA H275Y des virus A(H1N1) Brisbane-like résistants à l’oseltamivir sans que ses capacités de réplication ne soient altérées. De même nous montrons que son réassortiment avec un virus hautement pathogène A(H5N1) est possible. Ces observations renforcent la nécessité de promouvoir la vaccination afin de limiter les risques de co-infection virale chez un même individu. / In the context of A(H5N1) pandemics threat, an « avian flu and flu pandemics » project was proposed by LyonBioPole to develop influenza viruses characterization tools for vaccine production. To study genetic reassortment between influenza viruses, 3 reverse genetic systems of A(H3N2) human virus and A(H5N2) and A(H5N1) avian viruses were developed and reassortant viruses were produced. Their replicative capacities were evaluated using growth kinetics on MDCK cells with viral production quantification by real-time qRT-PCR. The A(H1N1)2009 emergence raises two questions about the acquisition by genetic reassortment of oseltamivir resistance and/or pathogenicity determinants. A co-infection protocol on MDCK cells was developed to study gene constellations of reassortant viruses like A(H1N1)2009-A(H1N1) H275Y and A(H1N1)2009-A(H5N1). We report here that genetic reassortment is possible between avian, human and swine strains using reverse genetic and viral co-infection and that some specific constellations emerged. We also report, that pandemic A(H1N1)2009 can acquire the H275Y mutated NA from seasonal oseltamivir resistant A(H1N1) viruses without any modifications on replicative capacities. This genetic reassortment is also possible with A(H5N1) viruses. These observations strenght the importance of vaccination against all these influenza strains to reduce the risk of one-individual viral co-infection.

Master of Science / Department of Diagnostic Medicine/Pathobiology / Wenjun Ma / With introduction of the 2009 pandemic H1N1 virus (pH1N1) into swine herds, reassortment between the pH1N1 and endemic swine influenza viruses (SIVs) has been reported worldwide. Recently, reassortant H3N2 and H1N2 variant SIVs that contain the M gene from pH1N1 virus and the remaining seven genes from North American triple-reassortant (TR) SIVs have emerged. These variant viruses have caused more than 300 cases of human infections and one death in the USA, creating a major public health concern. To date, the pathogenicity and transmissibility of H1N2 variant viruses in pigs has not been investigated. Through passive surveillance, we have isolated two genotypes of reassortant H1N2 viruses with pH1N1 genes from diseased pigs in Kansas. Full genome sequence and phylogenetic analysis showed that one is a swine H1N2 variant virus (swH1N2v) with the M gene from pH1N1; the other is a reassortant H1N2 virus (2+6 rH1N2) with six internal genes from pH1N1 and the two surface genes from endemic North American TR H1N2 SIVs. Furthermore, we determined the pathogenicity and transmissibility of the swH1N2v, a human H1N2 variant (huH1N2v), and the 2+6 rH1N2 in pigs using an endemic TR H1N2 SIV (eH1N2) isolated in 2011 as a control. All four viruses were able to infect pigs and replicate in the lungs. Both H1N2 variant viruses caused more severe lung lesions in infected pigs when compared to the eH1N2 and 2+6 rH1N2 viruses. Although all four viruses are transmissible in pigs and were detected in the lungs of contact animals, the swH1N2v shed more efficiently than the other three viruses in the respective sentinel animals. The huH1N2v displayed delayed and inefficient nasal shedding in sentinel animals. Taken together, the human and swine H1N2 variant viruses are more pathogenic and the swH1N2v more transmissible in pigs and could pose a threat to public and animal health.

Influenza A besitzt ein segmentiertes, achtsträngiges Genom in negativer Orientierung. Die einzelnen Segmente sind in virale Ribonukleoproteinkomplexe (vRNPs) verpackt. Genomische Segmentierung erlaubt es Influenza, zwischen verschiedenen Stämmen Reassortierung zu betreiben, was zur Entstehung von hochgradig virulenten und potentiell pandemischen neuen Stämmen führen kann. Die Existenz eines Packungsmechanismus wird vermutet, der sicherstellt dass exakt ein Segment jeden Typs in neu knospende Viren verpackt wird. Es gibt Indizien dafür, dass die vRNPs während ihres Wegs vom Nukleus zur Plasmamembran, wo die Knospung stattfindet, Multi-Segment-Komplexe ausbilden, die durch RNA-RNA-Interaktionen, sog. Packungssignale vermittelt werden. Dieser Prozess ist allerdings noch nicht hinreichend verstanden. In dieser Arbeit wurde eine neue RNA-FISH-Methode namens MuSeq-FISH entwickelt und angewendet, um die spektralen Limitierungen bisheriger Multiplexing-Ansätze zu überwinden und alle vRNA- und mRNA-Spezies vom humanen Stamm A/Panama des Influenza A Virus zu visualisieren. Außerdem wurde ein automatisierter Arbeitsablauf zur Registrierung/Ausrichtung, Punktdetektion, computergestützter Kolokalisationsanalyse und kombinatorischer Analyse der Mikroskopiebilder entwickelt, der auch große Datenmengen verarbeiten kann. Erstmalig wurde damit eine vollständige Kartographierung der Lokalisation und Häufigkeiten alle viralen RNAs in einzelnen Zellen vorgenommen. Aus diesen Daten konnten wir Erkenntnisse zu den Mechanismen und möglichen Hierarchien innerhalb des Packungsprozesses gewinnen. Dazu wurden Reaktionspfade und statistische Analysen von über 60 einzelnen Zellen und mehr als 105 einzelner vRNPs herangezogen. Es wurden auch Informationen über die vRNP-Häufigkeiten und deren Unterschiede zwischen Einzelzellen gewonnen, die zeigen dass sich Infektionsumgebungen auch in großer räumlicher Nähe stark unterscheiden und dadurch den Verpackungsmechanismus beeinflussen können. Weiterhin wurde eine Modellierung basierend auf bedingten Wahrscheinlichkeiten genutzt, um Reaktionskonstanten aus statischen FISH-Daten zu erhalten. Wir haben unsere Analysen zusätzlich auf den aviären Stamm A/Mallard und die reassortanten Stämme A/Pan-M, A/Pan-NS und A/Pan-NSM erweitert, die ein gemischtes Genom aus A/Panama und A/Mallard enthalten. Dabei konnte gezeigt werden, dass sich die Packungsdynamiken und -netzwerke auch zwischen nah verwandten Stämmen erheblich unterscheiden. Heterogene Verpackungsprozesse wurden für diese Stämme observiert, anhand welcher A/Pan-M und A/Pan-NS eher A/Mallard zugeordnet werden konnten. Ebenfalls wurden erste Schritte unternommen, um die Methode in verschiedener Hinsicht zu erweitern: es zeigte sich, dass MuSeq-FISH und STED-Mikroskopie im Prinzip kombinierbar sind, was auch durch gleichzeitige Detektion von drei vRNA-Segmenten gezeigt werden konnte. MuSeq-FISH wurde auch genutzt, um einzelne Virionen direkt nach deren Eintritt in die Zelle zu färben und auf deren genomischen Inhalt hin zu untersuchen. Dabei fiel auf, dass die Segmente 7 und 8 besonders häufig fehlten, wenn unvollständige Genome detektiert wurden. Außerdem wurde ein Plasmidsystem auf Basis des pHW2000-Vektors für fast alle Segmente von A/Panama umkloniert, welches nun die Expression von mRNA ohne die gleichzeitige Expression von vRNA ermöglicht. In einem ersten Experiment konnte die Funktionalität des Systems gezeigt werden, so dass es potentiell in Transfektionsexperimenten die Untersuchung vom Packungsmechanismus ermöglichen kann, und zwar unter infektionsähnlichen Bedingungen mit beliebig kombinierbaren vRNA-Sets. Wir erwarten, dass MuSeq-FISH zusammen mit dem automatisierten Arbeitsablauf auch eine nützliche Methode für andere biologische Fragestellungen darstellen wird, besonders wenn es um hochgradig kolokalisierte Untersuchungsobjekte geht. Fundiertes Wissen über den Packungsmechanismus von Influenzaviren kann helfen, die Entstehung von pandemischen Stämmen besser zu verstehen und kann Möglichkeiten aufzeigen, neue antivirale Medikamente zu entwickeln. / Influenza A has a segmented genome of eight single-stranded, negative-sense RNAs packed into ribonucleoproteins (vRNPs). This segmentation allows reassortment between different strains with the potential to create highly virulent, pandemic new strains. A packaging mechanism is supposed, ensuring the incorporation of one copy of each segment species into budding virions. En route from the nucleus to budding at the plasma membrane, the vRNPs are thought to form multisegment complexes via RNA-RNA and RNP-RNP interactions called packaging signals. This process is not yet completely understood. Here, a new RNA-FISH method (MuSeq-FISH) was introduced to overcome the spectral limits of multiplexing in order to visualize all IAV vRNA and mRNA targets of the human strain A/Panama. An image processing pipeline including image registration, spot detection, automated colocalization analysis and combinatorial analysis was developed, capable of high data throughput. For the first time, a complete map of the localization and abundance of all viral RNAs in individual cells has been generated. This data enabled detailed investigations about the mechanisms and potential hierarchies within the packaging process, which were inferred from pathways and statistical analysis of over 60 individual cells with more than 105 vRNP occurrences. We also gained information about the abundance and cell-to-cell heterogeneity of vRNPs among large sets of infected cells, unravelling that infection environments even in neighboring cells differ strongly in segment composition with an impact on packaging. In addition, conditional probability modelling was conducted to infer reaction constants from inherently static FISH data. We have extended this analysis to the avian strain A/Mallard and the reassortant strains A/Pan-M, A/Pan-NS and A/Pan-NSM, which contain reassorted genomes of A/Panama and A/Mallard. Here we have shown that packaging dynamics and networks differ widely, even among closely related strains. Packaging processes in these strains seemed to be very diverse, however we found A/Pan-M and A/Pan-NS to more closely resemble A/Mallard in terms of packaging. First steps have been taken to extend the method into different directions: combi- nation of MuSeq-FISH with STED imaging is in principle possible and has been applied for simultaneous detection of three vRNA species. MuSeq-FISH was also applied to single IAV virions directly after cell entry in order to study their genome content, where we found segments 7 and 8 to be lacking most frequently. In addition, a system of pHW2000-based plasmids expressing only mRNA has been created for almost all A/Panama segments. The functionality of this system was shown in a proof of concept, so that its use in transfection experiments can serve as a potential instrument to investigate vRNP packaging in artificial infection-like conditions with reduced vRNAs sets of choice. MuSeq-FISH together with its image analysis pipeline will be a useful tool also for other biological questions, especially concerning high-grade colocalization. Further understanding of the vRNP packaging in influenza can help us to understand the emergence of pandemic strains and open up paths to new antiviral medication.

La grippe, infection respiratoire virale fréquente, est due aux virus Influenza. Leur génome est constitué par huit molécules d’ARN de polarité négative retrouvés sous la forme de complexe ribonucléique (RNPv). Au cours du cycle viral, il a été démontré que les régions terminales des segments de gène étaient cruciales pour l’incorporation sélective des huit RNPv à l’intérieur des particules virales. Par des techniques d’interaction in vitro et de tomographie électronique, nous avons montré que les segments de gène du virus H3N2 interagissaient entre eux par des interactions ARN/ARN impliquant leurs régions de packaging. Nos résultats suggèrent que la mise en place de ce réseau permettrait la formation d’un complexe supra macromoléculaire multi-segmenté permettant l’incorporation d’un jeu complet des huit RNPv dans les particules virales néosynthetisées. En raison de la nature segmentée du génome viral, des phénomènes de réassortiment génétique peuvent avoir lieu lors d’une co-infection. Afin de définir les mécanismes responsables de la restriction observée lors de ce phénomène, nous avons évalué le taux de réassortiment génétique in vitro entre le virus humain H3N2 et le virus aviaire H5N2. Nos résultats suggèrent que le mécanisme gouvernant l’incorporation sélective des segments de gènes, régulerait le réassortiment génétique. Nous avons montré que la modulation de l’interaction ARN/ARN entre les segments de gènes HA et M permet d’augmenter le taux d’incorporation du segment de gène HA H5 dans le fond génétique du virus humain, prérequis pour l’émergence de virus pandémique / The Flu is a frequent viral infectious disease caused by the Influenza viruses. Their genomes are composed by eight negative single-stranded RNA organised as vRNPs. During the viral cycle, the terminal non-coding and coding regions of viral genome have been shown to be crucial for the selective incorporation of a complete set of the eight vRNPs into influenza viral particles. Band shift assay and electron tomography allowed us to show that all gene segments interact together by RNA/RNA interactions involving their packaging region. Our results suggest that the eight genomic vRNAs are selected and packaged as an organized supramolecular complex held together between identified packaging regions into neosynthesized virions. Due to genome segmented nature, genetic reassortment can occur during co-infection. In order to identify molecular mechanisms responsible for the observed restriction during the genetic reassortment, we have developed a new competitive reverse genetic strategy allowing us to evaluate the genetic reassortment between H3N2 and H5N2 viruses. Our results suggest that mechanism controlling the packaging should regulate genetic reassortment. We have shown that the modulation of RNA/RNA interaction between HA and M gene segment have allowed us to increase HA H5 gene segment incorporation rate into a viral human genetic background, prerequisite for pandemic virus emergency

Le génome des virus influenza de type A est composé de huit segments de gènes (ARNv) de polarité négative retrouvés sous la forme de complexes ribonucléiques (RNPv). L'incorporation sélective des huit RNPv dans les particules virales néosynthétisées se fait par un mécanisme moléculaire qui fait intervenir des signaux d'encapsidation dites « région de packaging ». Nous avons montré que les segments de gènes interagissaient entre eux via des interactions de type ARN/ARN permettant la formation d'un réseau d'interactions. Nous avons de plus montré que les régions de packaging décrites dans la littérature semblent héberger les régions impliquées dans la mise en place du réseau d'interactions. Cette étude a été réalisée pour le virus humain H3N2 et le virus aviaire H5N2. Le mécanisme d'incorporation sélective des segments de gènes semble également réguler le réassortiment génétique, processus génétique responsable de l'émergence de virus réassortants. Nous avons montré qu'une restriction génomique impliquant les régions de packaging semble être responsable du taux de réassortiment génétique faible observé in-vitro et in-vivo. La modulation du réseau d'interactions ARN/ARN semble être nécessaire pour l'incorporation de segments aviaire dans le fond génétique du virus humain. Pour finir, nous avons montré que la composition génomique des virus réassortants vaccinaux joue un rôle central dans la réplication virale et dans la production des antigènes vaccinaux. Par une stratégie de cryo-microscopie, nous avons montré que la protéine PB1 joue un rôle central dans l'optimisation de la production des antigènes de surface / The genome of the influenza A virus (IAV) comprises eight single-stranded negativesense RNA segments (vRNAs). All eight vRNAs are selectively packaged into each progeny virion via packaging signal sequences that are located at both ends of the vRNAs. How these signals ensure packaging of all eight vRNAs remains unclear. It was hypothesized that selective packaging might be driven by direct interactions between vRNAs. Combination of biochemical and reverse genetic approaches allowed us to identify short nucleotide regions on vRNAs interacting with each other in vitro. Here, we demonstrated the importance of these interactions in the packaging process of the human H3N2 and avian H5N2 viral genomes. Furthermore, our results suggest that the packaging process could regulate genetic reassortment. Indeed, we observed that the genetic reassortment between H3N2 and H5N2 viruses is restricted as the avian vRNA HA cannot be incorporated into the human genetic background. Our investigations indicated that (i) the packaging signals are crucial for genetic reassortment and (ii) the modulation of the vRNAs interaction network may be required for the incorporation of the avian HA gene into the human genetic background. Characterization of seed viruses showed that the genetic composition is important for both high growth ability and antigen production. Indeed, cryo-electronic microscopy observations of reassortant virus indicated that the PB1 gene can strongly influence the antigen glycoprotein spike density

Despite the global licensure of two live-attenuated rotavirus vaccines, Rotarix® and RotaTeq®, rotavirus remains the major cause of severe dehydrating diarrhoea in young mammals and the need for further development of additional rotavirus vaccines, especially vaccines effective against regional strains in developing country settings, is increasing. The design and formulation of new effective multivalent rotavirus vaccines is complicated by the wide rotavirus strain diversity. Novel rotavirus strains emerge periodically due to the propensity of rotaviruses to evolve using mechanisms such as point mutation, genome segment reassortment, genome segment recombination and interspecies transmission. Mutations occurring within the primer binding regions targeted by the current commonly employed sequence-dependent genotyping techniques lead to difficulties in genotyping novel mutant rotavirus strains. Therefore, use of sequence-independent techniques coupled with online rotavirus genotyping tools will help to understand the complete epidemiology of the circulating strains which, in turn, is vital for developing intervention measures such as vaccine and anti-viral therapies. In this study, sequence-independent cDNA synthesis that uses a single set of oligonucleotides that do not require prior sequence knowledge of the rotavirus strains, 454® pyrosequencing, and an online rotavirus genotyping tool, RotaC, were used to swiftly characterise the whole genome of rotaviruses. The robustness of this approach was demonstrated in characterising the complete genetic constellations and evolutionary origin of selected human rotavirus strains that emerged in the past two decades worldwide, human rotavirus strains frequently detected in Africa, and the whole genomes of some common strains frequently detected in bovine species. Most of the characterised strains emerged either through intra- or interspecies genome segment reassortment processes. The methods used in this study also allowed determination of the whole consensus genome sequence of multiple rotavirus variants present in a single stool sample and the elucidation of the evolutionary mechanisms that explained their origin. The 454® pyrosequence-generated data revealed evidence of intergenotype rotavirus genome segment recombination between the genome segments 6 (VP6), 8 (NSP2) and 10 (NSP4) of Wa-like and DS-1-like origin. The use of next generation sequencing technology combined with sequence-independent amplification of the rotavirus genomes allowed the determination of the consensus nucleotide sequence for each of the genome segments of the selected study strains directly from stool sample. The consensus nucleotide sequences of the genome segments encoding VP2, VP4, VP6 and VP7 of some of the study strains were codon optimised for insect cell expression and used to generate recombinant baculoviruses. The Bac-to-Bac baculovirus expression system was used to generate chimaeric rotavirus virus-like particles (RV-VLPs). These chimaeric RV-VLPs contained inner capsids (VP2 and VP6) derived from a South African RVA/Humanwt/ ZAF/GR10924/1999/G9P[6] strain, on to which outer capsid layer proteins composed of various combinations of VP4 and VP7 were assembled. The outer capsid proteins were derived from the dsRNA of G2, G8, G9 or G12 strains associated with either P[4], P[6] or P[8] genotypes that were directly extracted from human stool faecal specimens. The structures of these chimaeric RV-VLPs were morphologically evaluated using transmission electron microscopy (TEM). Based on the size and morphology of the particles, doublelayered (dRV-VLPs) and triple-layered RV-VLPs (tRV-VLPs) were produced. Recombinant rotavirus proteins readily assembled into dRV-VLPs, whereas approximately 10 – 30% of the assembled RV-VLPs from insect expressed recombinant VP2/6/7/4 were chimaeric tRVVLPs. These RV-VLPs will be evaluated in future animal studies as potential non-live rotavirus vaccine candidates. The novel approach of producing RV-VLPs introduced in this study, namely by using the consensus nucleotide sequence derived from dsRNA extracted directly from clinical specimens, should speed up vaccine research and development by bypassing the need to adapt the viruses to tissue culture and circumventing some other problems associated with cell culture adaptation as well. Thus, it is now possible to generate RV-VLPs for evaluation as non-live vaccine candidates for any human or animal field rotavirus strain. / Thesis (PhD (Biochemistry))--North-West University, Potchefstroom Campus, 2012

Despite the global licensure of two live-attenuated rotavirus vaccines, Rotarix® and RotaTeq®, rotavirus remains the major cause of severe dehydrating diarrhoea in young mammals and the need for further development of additional rotavirus vaccines, especially vaccines effective against regional strains in developing country settings, is increasing. The design and formulation of new effective multivalent rotavirus vaccines is complicated by the wide rotavirus strain diversity. Novel rotavirus strains emerge periodically due to the propensity of rotaviruses to evolve using mechanisms such as point mutation, genome segment reassortment, genome segment recombination and interspecies transmission. Mutations occurring within the primer binding regions targeted by the current commonly employed sequence-dependent genotyping techniques lead to difficulties in genotyping novel mutant rotavirus strains. Therefore, use of sequence-independent techniques coupled with online rotavirus genotyping tools will help to understand the complete epidemiology of the circulating strains which, in turn, is vital for developing intervention measures such as vaccine and anti-viral therapies. In this study, sequence-independent cDNA synthesis that uses a single set of oligonucleotides that do not require prior sequence knowledge of the rotavirus strains, 454® pyrosequencing, and an online rotavirus genotyping tool, RotaC, were used to swiftly characterise the whole genome of rotaviruses. The robustness of this approach was demonstrated in characterising the complete genetic constellations and evolutionary origin of selected human rotavirus strains that emerged in the past two decades worldwide, human rotavirus strains frequently detected in Africa, and the whole genomes of some common strains frequently detected in bovine species. Most of the characterised strains emerged either through intra- or interspecies genome segment reassortment processes. The methods used in this study also allowed determination of the whole consensus genome sequence of multiple rotavirus variants present in a single stool sample and the elucidation of the evolutionary mechanisms that explained their origin. The 454® pyrosequence-generated data revealed evidence of intergenotype rotavirus genome segment recombination between the genome segments 6 (VP6), 8 (NSP2) and 10 (NSP4) of Wa-like and DS-1-like origin. The use of next generation sequencing technology combined with sequence-independent amplification of the rotavirus genomes allowed the determination of the consensus nucleotide sequence for each of the genome segments of the selected study strains directly from stool sample. The consensus nucleotide sequences of the genome segments encoding VP2, VP4, VP6 and VP7 of some of the study strains were codon optimised for insect cell expression and used to generate recombinant baculoviruses. The Bac-to-Bac baculovirus expression system was used to generate chimaeric rotavirus virus-like particles (RV-VLPs). These chimaeric RV-VLPs contained inner capsids (VP2 and VP6) derived from a South African RVA/Humanwt/ ZAF/GR10924/1999/G9P[6] strain, on to which outer capsid layer proteins composed of various combinations of VP4 and VP7 were assembled. The outer capsid proteins were derived from the dsRNA of G2, G8, G9 or G12 strains associated with either P[4], P[6] or P[8] genotypes that were directly extracted from human stool faecal specimens. The structures of these chimaeric RV-VLPs were morphologically evaluated using transmission electron microscopy (TEM). Based on the size and morphology of the particles, doublelayered (dRV-VLPs) and triple-layered RV-VLPs (tRV-VLPs) were produced. Recombinant rotavirus proteins readily assembled into dRV-VLPs, whereas approximately 10 – 30% of the assembled RV-VLPs from insect expressed recombinant VP2/6/7/4 were chimaeric tRVVLPs. These RV-VLPs will be evaluated in future animal studies as potential non-live rotavirus vaccine candidates. The novel approach of producing RV-VLPs introduced in this study, namely by using the consensus nucleotide sequence derived from dsRNA extracted directly from clinical specimens, should speed up vaccine research and development by bypassing the need to adapt the viruses to tissue culture and circumventing some other problems associated with cell culture adaptation as well. Thus, it is now possible to generate RV-VLPs for evaluation as non-live vaccine candidates for any human or animal field rotavirus strain. / Thesis (PhD (Biochemistry))--North-West University, Potchefstroom Campus, 2012