Investigation of the genetic basis for virus tropism and virulence of classical swine fever virus

Classical swine fever virus (CSFV) is the causative agent of classical swine fever, an economically important and highly contagious disease of pigs. CSFV is a positive‐stranded RNA virus and CSFV strains vary considerably in their virulence, with high, moderate and low virulence strains. The E2 glycoprotein is the major surface component of the virion and modifications introduced into this glycoprotein appear to have an important effect on CSFV virulence. E2 shows a high variability among different CSFV strains, and a specific amino acid motif within the E2 varies between strains of different virulence. In the highly virulent CSFV strains, Koslov and ALD, this motif comprises residues E761/S763/L764 in the polyprotein, and has been termed the ‘SL’‐motif. However, CSFV strains encoding R761/S763/L764 represent the predominant alleles across all published CSFV genomes. Positive‐strand RNA viruses evolve rapidly, due to error‐prone RNA replication and the lack of proof‐reading activity of the RNA‐dependent RNA polymerase. With error rates ranging from 10‐6 to 10‐4 mutations per nucleotide incorporated, this results in a virus population that exists as a quasispecies of different, but closely related variants. Certain variants, or haplotypes, emerge or change continuously and the variation within the virus population enables the virus to change and adapt during infection. The diversity and quasispecies composition of CSFV and their role as determinants of virulence have not been studied in depth. In this thesis, the focus has been on elucidating which haplotypes constitute the virus population through establishing a high‐throughput method for cloning of full‐length CSFV cDNA. Deep sequencing cannot easily resolve the different haplotypes that constitute the virus populations, therefore obtaining full‐length cDNA clones represents an alternative approach to identify the individual haplotypes present. Numerous clones were obtained, sequenced by NGS, and phylogenetic analysis revealed that they were all unique. A majority of the unique mutations were also observed in the serum samples of CSFV infected pigs, or in published sequences. Animal experiments were performed with infectious cDNA clones, containing substitutions in the ‘SL’‐motif. One pig infected with a specific substitution variant, exhibited earlier signs of disease compared to pigs infected with other variants. The virus populations in different tissues and serum were investigated by deep sequencing to identify adaptations and diversity of the populations. Several nonsynonymous mutations were identified in some variants. This laid the basis for further in vitro study of the role of the ‘SL’‐motif in the highly and moderately virulent strains Koslov and ALD/A76, respectively. The results of this analysis, based on growth kinetics and serial passaging in several different cell types, revealed no discernible difference in growth of these variants in these cells in vitro. This thesis is comprised of three parts: Part I, is an introduction to CSFV, virus genome and protein functions, virus lifecycle, virus quasispecies, role of the E2 glycoprotein in virulence with the focus on the ‘SL’‐motif, and molecular virology, with the application of cloning of full‐length viral cDNA and NGS. Part II, comprises the manuscripts published or in preparation. Manuscript I describes the establishment of a highthroughput cloning method for generation of full‐length cDNA clones from a virus population, to gain insight into which haplotypes constitute the virus population. Several clones were obtained, and NGS employed for whole genome sequencing and phylogenetic analysis. Manuscript II describes the role of residues 763 and 764 in the virulence of CSFV strain Koslov through two infection experiments of pigs. Unexpectedly, one variant, vKos_LP exhibited earlier signs of disease in infected pigs, compared to the other variants. The evolution of the viral population during in vivo infection was investigated by NGS and several interesting adaptations were observed for some of the variants, such as vKos_SP, which contained several linked non‐synonymous mutations at the consensus level of the sequenced CSFV genome. Manuscript III builds upon the results observed in manuscript II, with an in vitro study of residues 761, 763 and 764 in CSFV strains Koslov and ALD/A76 to elucidate their potential role in cell specificity. This was investigated by growth kinetics and serial passaging in different cell types, which revealed no discernible difference in the variants. One adaptation in the ‘SL’‐motif was observed: a K761R substitution in the 9th passage of vKos_KSL in primary plexus choroideus cells. NGS revealed that a selective sweep had taken place in this passage, and that the R761 variant contained several non‐synonymous mutations in the NS5A/B. Part III sums up the findings of this thesis and provides a general discussion and overall conclusions, as well as future perspectives.