Viral RNA polymerases: the error paradox at the single-molecule level

RNA viruses are characterized by a high mutation frequency, which ? paradoxically ? is key to their success as infectious agents of all living organisms. RNA virus populations are ?quasi-species?, clouds of mutants ready to adapt to changing circumstances like the host?s immune response, antiviral therapy, or a new host species. This potential for rapid evolution is attributed to efficient replication in combination with the lack of a proofreading mechanism for removal of nucleotides that are misincorporated by the viral RNA polymerase (RdRp). Conceptually, RNA viruses are faced with a choice between replication speed (to outrun host responses) and reliable genome replication, reflected in a trade-off between processivity and fidelity of their RdRp. The high mutation frequency has presumably restricted genome size expansion, to prevent the accumulation of too many deleterious mutations (?error catastrophe?). In this light, the extraordinary 27- to 32-kilobase (kb) genome size of coronaviruses is remarkable. The group occupies an isolated position at the upper end of the RNA virus genome size scale and speculations about proofreading during coronavirus replication have been numerous. Recent reports implicate a newly discovered viral exoribonuclease (ExoN) in such a function, although the molecular mechanism remains entirely obscure. Interestingly, a group of distant coronavirus relatives, the arteriviruses, which also belong to the order of nidoviruses, have genomes of only 12-16 kb and lack a homolog of the ExoN domain, supporting the hypothesis that ExoN acquisition was a key event in coronavirus genome expansion. We have recently developed the first in vitro activity assays using purified recombinant RdRps of both SARS-coronavirus and the arterivirus EAV. We will use these biochemical assays, in combination with real-time single-molecule techniques (magnetic tweezers) and deep sequencing analysis of in vivo and in vitro produced viral RNA, to characterize the basic RdRp properties (including processivity and fidelity) of the distantly related SARS-coronavirus and EAV RdRps. The well-characterized poliovirus RdRp will be included for comparison and as positive control. The existence of proofreading during coronavirus RNA synthesis will be evaluated, as well as the putative role of the ExoN exoribonuclease in preventing error catastrophe. The recent description of a unique second RdRp in coronaviruses, presumably synthesizing primers for the viral ?main? RdRp, will be the starting point for an in-depth analysis of coronavirus RdRp initiation and elongation, studies in which also the viral helicase will be taken into account. Finally, we will study another trademark of the nidovirus RdRp: the mechanism of discontinuous minus-strand RNA synthesis that operates to produce the corona- and arterivirus subgenomic mRNAs from which the viral structural proteins are expressed. Our multidisciplinary approach (biochemistry, biophysics, and virology) will explore novel experimental avenues for viral RdRps. Our results will be relevant to the (viral) RdRp field at large, will shed more light on the unusual evolutionary position of coronaviruses, and will directly address key aspects of nidovirus molecular biology. The characterization of viral RdRps is highly relevant for the design of live virus vaccines and antiviral drugs, and for assessing the evolutionary potential of RNA viruses as dynamic infectious agents.