Another intriguing set of genes belongs to tetraselmis virus 1, which infects green algae. These hosts, found in nutrient-rich marine and fresh waters, are photosynthetic. The viral genome encodes pyruvate formate-lyase and pyruvate formate-lyase-activating enzyme, which are key members of cellular anaerobic respiration pathways and allow energy production when no oxygen is available. Green algae may use this system in waters depleted of oxygen by exuberant algal growth. If this process occurs in cells, why does the viral genome carry some of the genes involved? The answer is not known, but it is possible that the extra metabolic demands placed on cells during virus replication—especially at night— require additional fermentation enzymes for energy production. The presence of these genes suggests that tetraselmis virus 1 can change host metabolism, perhaps facilitating its reproduction.
These large viruses therefore have sufficient coding capacity to escape some restrictions imposed by host cell biochemistry. The smallest genome of a free-living cell is predicted to comprise <300 genes (based on bacterial genome sequences). Remarkably, this number is smaller than the genetic content of large viral DNA genomes. Nevertheless, the big viruses are not cells: their reproduction absolutely requires the cellular translation machinery, as well as host cell systems to make membranes and generate energy.
The parameters that limit the size of viral genomes are largely unknown. There are cellular DNA and RNA molecules that are much longer than those found in virus particles. Consequently, the rate of nucleic acid synthesis is not likely to be limiting. Nor does the capsid volume appear to limit genome size: the icosahedral shell of Mimivirus, which houses a 1.2 millionbase-pair DNA genome, is constructed mainly of a single major capsid protein. For larger genomes, the solution is helical symmetry, which can in principle accommodate very large genomes. The Pandoraviruses, with the largest known DNA viral genomes (2,500 kbp), are housed in decidedly nonisometric ovoid particles 1 μm in length and 0.5 μm in diameter.
There is no reason to believe that the upper limit in viral particle and genome size has been discovered. The core compartment of a mimivirus particle is larger than needed to accommodate the 1,200-kbp DNA genome. A particle of this size could, in principle, house a genome of 6 million bp if the DNA were packed at the same density as in polyomaviruses. Indeed, if the genome were packed into the particle at the density reached in some bacteriophages, it could be >12 million bp, the size of that of the smallest free-living unicellular eukaryote.
In cells, DNAs are much longer than RNA molecules. RNA is less stable than DNA, but in the cell, much of the RNA is used for the synthesis of proteins and therefore need not exceed the size needed to specify the largest polypeptide. However, this constraint does not apply to viral genomes. Yet the largest viral single-molecule RNA genomes, the 41-kb (+) strand RNAs of the nidoviruses (Box 3.4), are dwarfed by the largest (2,500kbp) DNA virus genomes. Susceptibility of RNA to chemical and nuclease attack might limit the size of viral RNA genomes. However, the most likely explanation is that there are few known enzymes that can correct errors introduced during RNA synthesis. An exonuclease encoded in the coronavirus genome is one exception: its presence could explain the large size of these RNAs. DNA polymerases can eliminate errors during polymerization, a process known as proofreading, and remaining errors can also be corrected after DNA synthesis is complete. The average error frequencies for RNA genomes are about 1 misincorporation in 104 or 105 nucleotides polymerized. In an RNA viral genome of 10 kb, a mutation frequency of 1 in 104 would produce about 1 mutation in every replicated genome. Hence, very long viral RNA genomes, perhaps longer than 40 kb, would sustain too many mutations that would be lethal. Even the 7.5-kb genome of poliovirus exists at the edge of infectivity: treatment of the virus with the RNA mutagen ribavirin causes a >99% loss in a single round of replication.
When new viral genomes are discovered, often many of the putative genes are previously unknown. For example, >93% of the >2,500 genes of Pandoravirus salinus resemble nothing known, and 453 of the 663 predicted open reading frames of tetraselmis virus 1 show no sequence similarity to known proteins. The implication of these findings is clear: our exploration of global genome sequences is far from complete, and viruses with larger genomes might yet be discovered.
The Origin of Viral Genomes
The absence of bona fide viral fossils, i.e., ancient material from which viral nucleic acids can be recovered, might appear to make the origin of viral genomes an impenetrable mystery. The oldest viruses recovered from environmental samples, the 30,000-year-old Pithovirus sibericum and Mollivirus sibericum, isolated from Late Pleistocene Siberian permafrost, are simply too rare and too young to provide much information on viral evolution. However, the discovery of fragments of viral nucleic acids integrated into host genomes, coupled with the advances in determining genome sequences of viruses and their hosts, has provided an improved understanding of the evolutionary history of viruses, a topic discussed in depth in Volume II, Chapter 10.
How viruses with DNA or RNA genomes arose is a compelling question. A predominant hypothesis is that RNA viruses are relics of the “RNA world,” a period populated only by RNA molecules that catalyzed their own replication in the absence of proteins. During this time, billions of years ago, cellular life could have evolved from RNA, and the earliest cellular organisms might have had RNA genomes. Viruses with RNA genomes might have evolved during this time. Later, DNA replaced RNA as cellular genomes, perhaps through the action of reverse transcriptases. With the emergence of DNA genomes probably came the evolution of DNA viruses. However, those with RNA genomes were and remain evolutionarily competitive, and hence they continue to survive to this day.
Analysis of sequences of more than 4,000 RNA-dependent RNA polymerases is consistent with the hypothesis that the first RNA viruses to emerge after the evolution of translation were those with (+) strand RNA genomes. The last common ancestor of these viruses encoded only an RNA-dependent RNA polymerase and a single capsid protein. Double-stranded RNA viruses evolved from (+) strand RNA viruses on at least two different occasions, and (–) strand RNA viruses evolved from dsRNA viruses. The emergence of viruses with the latter genome types was likely facilitated by the capture of genes such as those encoding RNA helicases, to allow for the production of larger genomes.
Single-stranded DNA viruses of eukaryotes appear to have evolved from genes contributed from both bacterial plasmids and (+) strand RNA viruses. Different dsDNA viruses originated from bacteriophages at least twice. The larger eukaryotic DNA viruses form a monophyletic group based on analysis of 40 genes that derive from a last common ancestor. These viruses appear to have emerged from smaller DNA viruses by the capture of multiple eukaryotic and bacterial genes, such as those encoding translation system components.
There is no evidence that viruses are monophyletic, i.e., descended from a common ancestor: there is no single gene shared by all viruses. Nevertheless, viruses with different genomes and replication strategies do share a small set of viral hallmark genes that encode icosahedral capsid proteins, nucleic acid polymerases, helicases, integrases, and other enzymes. For example, as discussed above, the RNA-dependent RNA polymerase is the only viral hallmark protein conserved in RNA viruses. Examination of the sequences of viral capsid proteins reveals at least 20 distinct varieties that were derived from unrelated genes in ancestral cells on multiple occasions. The emerging evidence therefore suggests that viral replication enzymes arose from precellular self-replicating genetic elements, while capsid protein genes were captured from unrelated genes in cellular hosts.
The compositions of the eukaryotic and bacterial viromes differ substantially (Chapter 1, Fig. 1.13). In bacteria, most known viruses possess dsDNA genomes; fewer viruses have ssDNA genomes, and there is a very limited number of viruses with RNA genomes. In eukaryotes, most of the virome diversity is accounted for by RNA viruses, but ssDNA and dsDNA viruses are common (Chapter