Insights into the molecular, evolutionary and structural biology of circoviruses

There have been few attempts to unravel the long‐term evolutionary history of vertebratevirus‐host relationships and determinants of viral speciation. Circoviruses, such as Beak andfeather disease virus (BFDV) are highly mutable ssDNA viruses that occupy entangledmultispecies ecological niches and provide an excellent model for studying paleobiologicalsignals that can unlock longer term phylogenies. In this study a phylogenetic analysis of newfull length BFDV genome sequences (n=39) from Australian budgerigars and lorikeets revealeda host‐based, cryptic and sympatric differentiation of BFDV in the Psittaciformes tribe Loriini.Furthermore, the BFDV lineage in lorikeets was similar to endogenous circoviral motifs presentin the genome of kea (Nestor notabilis), an insertion event that likely occurred many millionsof years ago. Furthermore, recombination analysis of BFDV from lorikeets uncovered a largerecombinant section of the replicase associated protein gene (Rep) derived from an ancestralcircovirus lineage in finches, a sister Order to the parrots, alluding to a psittacopasserine andcretaceous origin of BFDV. Gondwanan vicariance estimates were used to calibrate the genuswide macroevolutionary timescale of circoviruses. Divergence of BFDV in the tribe Loriini datedback to more than 10 million years which coincided with the Papuan central range orogenythat triggered the radiation of the lorikeets and lories and sympatric segregation of anantecedent circoviral clade in Australian lorikeets. The time calibrated circovirus phylogenyunveiled the chronology of circovirus speciation coinciding with the divergence estimates oftheir respective hosts to highlight deep forces of host codivergence in the evolution anddistribution of contemporary circovirus lineages. As the ancestry of BFDV lineage in lorikeetspredates the host divergence this study shows that a virus even with a high mutation rate canretain phenotypic stability and wait millennia to change hosts or become a recognisable newvirus species, a process that can lag behind the evolution of potential hosts.Conservation efforts for the orange‐bellied parrot (Neophema chrysogaster), one of theworld’s most critically endangered bird species, have been hampered by BFDV spilloverinfection. To better understand the vulnerability of the orange‐bellied parrot to potential BFDVgenotypes from captive and wild reservoirs hosts new full length BFDV genomes and partialxxxRep sequences were analysed alongside all available BFDV sequence data in the Australianlandscape. By assessing reservoir‐based and pathotype stratification, three recent incursions ofBFDV into orange‐bellied parrots were characterised, highlighting multiple threats of pathogenre‐emergence. Insignificant geographic but strong host‐based BFDV stratification was detectedacross the Australian landscape which suggested forces of cryptic host adaptation as the maindrivers for establishing new infection. However, analysis for host‐virus cophylogenyhighlighted competing forces of co‐divergence and cross‐species transmissions in the currentdistribution of BFDV genetic population. Homology models of capsid protein (Cap) structuresin different BFDV subpopulations demonstrated surface plasticity to accommodate mutationsexerted by adaptive drivers during host jumps or spill over infections. In addition, severaladaptive markers on BFDV Cap were proposed as they were identified evolving directionallyand became fixed into host adapted lineages.A novel diagnostic and genotyping method was developed for BFDV using PCR amplification ofthe capisd gene (Cap) and subsequent high‐resolution melt (HRM‐Cap) curve analysis. Limits ofdetection, sensitivity, specificity and genotype discriminatory power of the new method wasanalysed and compared with some contemporary molecular techniques. The limit of detectionof the new HRM‐Cap method was lower (2×10−5 ng/reaction or 48 viral copies) than that for apreviously published HRM‐Rep technique and conventional BFDV PCR (2×10−6 ng/reaction or13 viral copies). However, when used in a diagnostic setting with 348 clinical samples therewas strong agreement between HRM‐Cap and conventional PCR (kappa = 0.87, P < 0.01, 98%specificity) while HRM‐Cap had a higher specificity (99.9%) than HRM‐Rep (80.3%). HRM‐Capdemonstrated higher positive predictive value and discriminatory power for detection and genotyping of BFDV isolates compared to the existing rapid genotyping method based on Rep.Since no structural information is currently available for BFDV replication associated protein(Rep), an E. coli based recombinant expression and purification method for different Reptargets was developed and optimised for downstream structural biology applications. Thereplicase domain and full length native Rep was synthesised using a two‐step purificationprocess resulting in high concentration and high purity. However, crystallisation trials werexxxiunsuccessful limiting the potential of X‐ray crystallography to resolve the atomic structure ofRep. However, results from negative staining electron microscopy suggested that the fulllength native Rep oligomerises to a higher order macromolecule with an accumulatedmolecular weight ~140 KDa forming 10‐15 nm homogenous particles highly suitable for cryo‐EM reconstruction.With several new circovirus species identified in recent years from a diverse range of hosts,there is a pressing need for better understanding of the structural biology of circovirus Capsand how they form virions. Cap proteins from six novel circoviruses were synthesised usingrecombinant expression system and the first atomic resolution structure was resolved frombat associated circovirus by X‐ray crystallography and cryo‐EM reconstruction at 2.7 Å and 3.1Å resolution, respectively, and for canine circovirus at 3.2 Å by cryo‐EM. Mutationalexperiments highlighted the critical role of the conserved arginine rich motifs (ARM) in selfassemblyof Cap into virus like particles across circoviruses. The minimum length of ARMdomain required for initiating macromolecular assembly was also elucidated. Structuralalignment highlighted the underlying physical constraints imposed by both icosahedralsymmetry and electrostatic potential while also locating plausible regions for surfaceplasticity. This information has potential applications in vaccine development and forbioengineering protein nanocages for drug or gene delivery.