For DNA bacteriophages, the ratio of infectious particles to total particles usually ap-proaches a value of one. Such is not the case for animal viruses. Typically fewer than 1% of the particles derived from a cell infected with an animal virus are infectious in other cells as determined by a plaque assay. Although some of this discrepancy may be attribut-able to inefficiencies in the assay procedures, it is clear that many defective particles are being produced. In part, this production of defective particles arises because the mutation rates for animal viruses are unusually high and because many infections occur at high multiplicities, where defective genomes are complemented by nondefective viruses and therefore propagated.
Many DNA viruses use the host DNA synthesis machinery for replicating their genomes. Therefore, they benefit from the built-in proofreading and other error-correcting mecha- nisms used by the cell. However, the largest animal viruses (adenoviruses, herpesviruses, and poxviruses) code for their own DNA polymerases, and these enzymes are not as ef- fective at proofreading as the cellular polymerases. The resulting higher error rates in DNA replication endow the viruses with the potential for a high rate of evolution, but they are also partially responsible for the high frequency of defective viral particles.
The replication of RNA viruses is characterized by even higher error rates because viral RNA polymerases do not possess any proofreading capabilities. The result is that error rates for RNA viruses commonly approach one mistake for every 2500 to 10,000 nucleotides polymerized. Such a high misincorporation rate means that even for the smallest RNA viruses, virtually every round of replication introduces one or more nu- cleotide changes somewhere in the genome. If it is assumed that errors are introduced at random, most of the members of a clone (eg, in a plaque) are genetically different from all other members of the clone. The resulting mixture of different genome sequences for a particular RNA virus has been referred to as a quasispecies to emphasize that the level of genetic variation is much greater than what normally exists in a species.
Because of the redundancy in the genetic code, some mutations are silent and are not reflected in changes at the protein level, but many occur in essential genes and contribute to the large number of defective particles found for RNA animal viruses. The concept of genetic stability takes on a new meaning in view of these considerations, and the RNA virus population as a whole maintains some degree of homogeneity only because of the high degree of fitness exhibited by a subset of the possible genome sequences. Thus, strong selective forces continually operate on a population to eliminate most mutants that fail to compete with the few very successful members of the population. However, any time the environment changes (eg, appearance of neutralizing antibodies), a new subset of the population is selected and maintained as long as the selective forces remain constant.
The high mutation rates found for RNA viruses endow them with a genetic plasticity that leads readily to the occurrence of genetic variants and permits rapid adaptation to new environmental conditions. The large number of serotypes of the rhinoviruses causing the common cold, for instance, likely reflect the potential to vary by mutation. Although rapid genetic change occurs for most if not all viruses, no medically important RNA virus has exhibited this phenomenon as conspicuously as influenza virus. Point mutations accu-mulate in the influenza genes coding for the two envelope proteins (hemagglutinin and neuraminidase), resulting in changes in the antigenic structure of the virions. These changes lead to new variants not recognized by the immune system of previously infected individuals. This phenomenon is called antigenic drift . Apparently, those domains of the two envelope proteins that are most important for immune recogni-tion are not essential for virus reproduction and, as a result, can tolerate amino acid changes leading to antigenic variation. This feature may distinguish influenza from other human RNA viruses that possess the same high mutation rates, but do not exhibit such high rates of antigenic drift. Antigenic drift in epidemic influenza viruses from year to year requires continual updating of the strains used to produce immunizing vaccines.
The retroviruses likewise show high rates of variation because they depend for their replication on two different polymerases, both of which are error prone. In the first step of the replication cycle, the reverse transcriptase that copies the RNA genome into double-stranded DNA lacks a proofreading capability. Once the viral DNA has integrated into the chromosome of the host cell, the DNA is transcribed by the host RNA polymerase II, which similarly is incapable of proofreading. Accordingly, the retroviruses, including hu-man immunodeficiency virus type 1 (HIV-1), the causative agent of acquired immunode-ficiency syndrome (AIDS), exhibit a high rate of mutation. This property gives them the ability to evolve rapidly in response to changing conditions in the infected host.
Retroviruses that exhibit high rates of antigenic variation such as HIV-1 pose particu-larly difficult problems for the development of effective vaccines. Attempts are being made to identify conserved, and therefore presumably essential, domains of the envelope proteins for these viruses, which might be useful in developing a genetically engineered vaccine.
In early studies with influenza virus, it was noted that serial passage of virus stocks at high multiplicities of infection led to a steady decline of infectious titer with each pas-sage. At the same time, the titer of noninfectious particles increased. As is discussed be-low, the noninfectious genomes interfere with the replication of the infectious virus and so are called defective interfering (DI) particles. Later, these observations were ex-tended to include virtually all DNA as well as RNA animal viruses. The phenomenon is now named after von Magnus, who described the initial observations with influenza virus.
A combination of two separate events lead to this phenomenon. First, deletion muta-tions occur at a significant frequency for all viruses. For DNA viruses, the mechanisms are not well understood, but deletions presumably occur as a result of mistakes in replication or by nonhomologous recombination. The basis for the occurrence of deletions in RNA viruses is better understood. All RNA replicases have a tendency to dissociate from the template RNA but remain bound to the end of the growing RNA chain. By reassociating with the same or a different template at a different location, the replicase “finishes” repli-cation, but in the process creates a shorter or longer RNA molecule. A subset of these vari-ants possess the proper signals for initiating RNA synthesis and continue replicating. Because the deletion variants in the population require less time to complete a replication cycle, they eventually predominate and constitute the DI particles.
Second, as their name implies, the DI particles interfere with the replication of nonde-fective particles. Interference occurs because the DI particles successfully compete with the nondefective genomes for a limited supply of replication enzymes. The virions re-leased at the end of the infection are therefore enriched for the DI particles. With each successive infection, the DI particles can predominate over the normal particles as long as the multiplicity of infection is high enough so that every cell is infected with at least one normal infectious particle. If this condition is satisfied, then the normal particle can com-plement any defects in the DI particles and provide all of the viral proteins required for the infection. Eventually, however, as serial passage is continued, the multiplicity of infectious particles drops below one, and the majority of the cells are infected only with DI particles. When this happens the proportion of DI particles in the progeny virus decreases.
In good laboratory practice, virus stocks are passaged at high dilutions to avoid the prob-lem of the emergence of high titers of DI particles. Nevertheless, the presence of DI particles is a major contributor to the low fraction of infectious virions found in all virus stocks.
Besides mutation, genetic recombination between related viruses is a major source of genomic variation. Bacterial cells as well as the nuclei of animal cells contain the enzymes necessary for homologous recombination of DNA. Thus, it is not surprising that recombi-nants arise from mixed infections involving two different strains of the same type of DNA virus. The larger bacteriophages such as and T4 code for their own recombination en-zymes, a fact that attests to the importance of recombination in the life cycles and possibly the evolution of these viruses. The fact that recombination has also been observed for cyto-plasmic poxviruses suggests that they too code for their own recombination enzymes.
As far as is known, cells do not possess the machinery to recombine RNA molecules. However, recombination among at least some RNA viruses has been observed by two different mechanisms. The first, which is unique to the viruses with segmented genomes (orthomyxoviruses and reoviruses), involves reassortment of segments during a mixed in-fection involving two different viral strains. Recombinant progeny viruses that differ from either parent can be accounted for by the formation of new combinations of the genomic segments that are free to mix with each other at some time during the infection. Reassort-ment of this type occurring during infections of the same cell by human and certain animal influenza viruses is believed to account for the occasional drastic change in the antigenicity of the human influenza A virus. These dramatic changes, called antigenic shifts, produce strains to which much of the human population lacks immunity and, thus, can have enor-mous epidemiologic and clinical consequences .
The second mechanism of RNA virus recombination is exemplified by the genetic re-combination between different forms of poliovirus. Because the poliovirus RNA genome is not segmented, reassortment cannot be invoked as the basis for the observed recombi-nants. In this case, it appears that recombination occurs during replication by a “copy choice” type of mechanism. During RNA synthesis, the replicase dissociates from one template and resumes copying a second template at the exact place where it left off on the first. The end result is a progeny RNA genome containing information from two different input RNA molecules. Strand switching during replication, therefore, generates a recom-binant virus. Although this is not frequently observed, it is likely that most of the RNA animal viruses are capable of this type of recombination.
A “copy choice” mechanism has also been invoked to explain a high rate of recombi-nation observed with retroviruses. Early after infection, the reverse transcriptase within the virion synthesizes a DNA copy of the RNA genome by a process called reverse transcription. In the course of reverse transcription, the enzyme is required to “jump” between two sites on the RNA genome . This propensity to switch templates apparently explains how the enzyme generates recombinant viruses. Because reverse transcription takes place in subviral particles, free mixing of RNA templates brought into the cell in different virus particles is not permitted. However, retroviruses are diploid, because each particle carries two copies of the genome. This arrangement ap-pears to be a situation ready-made for template switching during DNA synthesis and most likely accounts for retroviral recombination.
Occasionally, retroviruses package a cellular mRNA into the virion instead of a sec-ond RNA genome. This arrangement can lead to copy choice recombination between the viral genome and a cellular mRNA. The end result is sometimes the incorporation of a cellular gene into the viral genome. This mechanism is believed to account for the pro-duction of highly oncogenic retroviruses containing modified cellular genes .