reovirus capsid, where the enzymes responsible for this process are located. The viral mRNAs are exported to the cytoplasm for translation. Retroviral RNAs are synthesized in the cell nucleus, where translation does not take place. The architecture of membranous replication complexes of (+) strand RNA viruses may favor RNA synthesis and exclude translation.
Even if mechanisms exist for controlling whether the genomes of RNA viruses are translated or replicated, some ribosome-RNA polymerase collisions are likely to occur. The isolation of a polioviral mutant with a genome that contains an insertion of a 15-nucleotide sequence from 28S ribosomal RNA (rRNA) is consistent with this hypothesis. The RNA polymerase apparently copied 15 nucleotides of rRNA after colliding with a ribosome.
Origins of Diversity in RNA Virus Genomes
Misincorporation of Nucleotides
All nucleic acid polymerases insert incorrect nucleotides during chain elongation. DNA-directed DNA polymerases have proofreading capabilities in the form of exonuclease active sites that can correct such mistakes. Most RdRPs do not possess this capability. The result is that error frequencies in RNA replication can be as high as one misincorporation per 103 to 105 nucleotides polymerized, whereas the frequency of errors in DNA replication is about 10,000-fold lower. Many of the polymerization errors cause lethal amino acid changes, but other mutations that are not lethal are retained in the genomes of infectious virus particles. This phenomenon has led to the realization that RNA virus populations are quasispecies, or mixtures of many different genome sequences. The errors introduced during RNA replication have important consequences for viral pathogenesis and evolution (Volume II, Chapter 10). Because RNA viruses exist as mixtures of genotypically different viruses, viral mutants may be isolated readily. For example, live attenuated poliovirus vaccine strains are viral mutants that were isolated from an unmutagenized stock of wild-type virus.
Fidelity of copying by RdRPs is determined by how the template, primer, and NTP interact at the active site. Nucleotide binding occurs in two steps: first, the NTP is bound in such a way that the ribose cannot interact properly with the Asp of motif A and the Asn of motif B (Fig. 6.6). If the NTP is correctly base paired with the template, then there is a conformational change in the enzyme, which reorients the triphosphate and allows phosphoryl transfer to occur. The closed active-site polymerase structures reveal a network of hydrogen bonds from the 2′ hydroxyl of the base-paired NTP to motif B in the fingers domain and the top of motif A. This network links a base-paired NTP with structural interactions that stabilize the closed active site and promote catalysis. The conformational change after NTP binding is a major fidelity checkpoint for the picornaviral RdRP.
Further insight into fidelity control in RdRPs comes from the analysis of an altered poliovirus 3Dpol with higher fidelity than the wild-type enzyme. This variant was selected for resistance to ribavirin, an antiviral nucleoside analog that causes transition mutations. The single amino acid change, G64S, slows the conformational change that occurs on NTP base pairing, thereby reducing the elongation rate. Although this amino acid is remote from the active site, it participates in hydrogen bonding to motif A, which is important in holding the NTP in an appropriate conformation for catalysis. Subtle changes in the enzyme caused by this substitution make it more dependent on correct NTP base pairing in the active site, thereby increasing replication fidelity. Of great interest is the observation that a similar interaction between fingers and motif A can be observed in RdRPs from a wide variety of viruses. This mechanism of enhancing fidelity may therefore be conserved in all RdRPs.
The RdRP of members of the Nidovirales (Fig. 6.16) allows faithful replication of the large (up to 41-kb) RNA genomes. The RNA synthesis machinery includes proteins not found in other RNA viruses, such as ExoN, a 3′-5′ exonuclease. Inactivation of this enzyme does not impair viral replication but leads to 15- to 20-fold increases in mutation rates. This observation suggests that ExoN confers a proofreading function upon the viral RNA polymerase, similar to the activity associated with DNA synthesis (Chapter 9). Viruses lacking the ExoN gene display attenuated virulence in mice, and are being considered as vaccine candidates.
Segment Reassortment and RNA Recombination
Reassortment is the exchange of entire RNA molecules between genetically related viruses with segmented genomes. In cells coinfected with two different influenza viruses, the eight genome segments of each virus replicate. When new progeny virus particles are assembled, they can package RNA segments from either parental virus. Because reassortment is the simple exchange of RNA segments, it can occur at high frequencies.
In contrast to reassortment, recombination is the exchange of nucleotide sequences among different genomic RNA molecules (Fig. 6.27). Recombination, a feature of many RNA viruses, is an important mechanism for producing new genomes with selective growth advantages. This process has shaped the RNA virus world by rearranging genomes and creating new ones. RNA recombination was first discovered in cells infected with poliovirus and was subsequently observed for other (+) strand RNA viruses. The frequency of recombination can be relatively high: it has been estimated that 10 to 20% of polioviral genomic RNA molecules recombine in a single growth cycle. Recombinant polioviruses are readily isolated from the feces of individuals immunized with the three serotypes of Sabin vaccine. The genome of such viruses, which are recombinants of the vaccine strains with other enteroviruses found in the human intestine, may possess an improved ability to reproduce in the human alimentary tract and have a selective advantage over the parental viruses.
Recombination can occur by two different mechanisms: nonreplicative, the nonhomologous end joining of two different RNA molecules; or replicative, the switching of templates. Nonreplicative recombination is highly inefficient and thought to influence virus evolution minimally. Replicative recombination mainly occurs between nucleotide sequences of two parental genome RNA strands that have a high percentage of nucleotide identity. This mechanism of RNA recombination is coupled with the process of genome RNA replication: it occurs by template switching during (−) strand synthesis. The RNA polymerase first copies the 3′ end of one parental (+) strand, then switches templates and continues synthesis at the corresponding position on a second parental (+) strand. The exact mechanism of template exchange is not known, but it might be triggered by pausing of the polymerase during chain elongation or damage to the template. Template switching in poliovirus-infected cells occurs predominantly during (−) strand synthesis because the concentration of (+) strand acceptors for template switching is 30 to 70 times higher than that of (−) strand acceptors. A prediction of the replicative mechanism, which has been verified experimentally, is that recombination frequencies should be lower between the genomes of different poliovirus serotypes.
Figure 6.27 RNA recombination. Schematic representation of RNA recombination occurring during template switching by RdRP. Two parental genomes are shown as acceptor and donor. The RNA polymerase (purple oval) has copied the 3′ end of the donor genome and is switching to the acceptor genome. The resulting recombinant molecule is shown.
Alteration of amino acids within the poliovirus 3Dpol thumb domain that directly interact with the RNA duplex led to the identification of Leu420 as critical for replicative recombination. This amino acid is located within an α-helix of the thumb domain in the exit channel for product RNA. It interacts with the ribose group of the third nucleotide of the product RNA strand, away from the active site.
