Viruses must mutate to survive in the face of attack by their host's immune system. A new model suggests that the viral mutation rate is optimized in an evolutionary trade-off between adaptability and genomic integrity.
Evolution by natural selection requires genetic variation, and the ultimate source of this variation is mutation — random errors in genomic replication. In fact, because the fidelity of genomic replication is influenced by genetic variation, the mutation rate is itself subject to natural selection. Broad taxonomic data are consistent with this idea. Although mutation rates per nucleotide vary by factors of up to a million among living things, the variation in genomic mutation rates is less than a factor of a thousand. In RNA viruses, roughly one nucleotide per genome is incorrectly reproduced in each replication; for retroviruses this genomic mutation rate is one per ten replications; and it is one per 300 replications in DNA-based microbes, including DNA viruses and microorganisms1. Indeed, the remarkable similarity of genomic mutation rates within each of these groups may reflect deep underlying selective constraints, but its explanation remains a challenge to evolutionary biologists.
Writing in the journal Complexity, Christel Kamp and colleagues2 have now tackled the evolution of genomic mutation rates from a fresh angle. By incorporating the immune response as an explicit selective force in standard 'quasispecies' models, they calculate a viral genomic mutation rate that optimally balances the costs of too much and too little genetic variation.
Quasispecies theory3 was developed over 30 years ago as a means of describing evolution in populations of self-replicating RNA molecules with high mutation rates (quasispecies is the term applied to closely related genetic sequences that are affected as a group by natural selection). But it was soon recognized that quasispecies theory made a useful tool for the study of viral evolution. Its most fundamental prediction is the existence of an error threshold. If the mutation rate exceeds this threshold, then all genomic information is irretrievably lost and the population becomes extinct in a kind of mutational meltdown. In standard quasispecies theory, the simplifying assumption is made that evolutionary fitness is entirely genetically determined and thus constant irrespective of the environment. Under this assumption, a zero mutation rate is optimal and selection should always favour greater fidelity of replication.
But viruses in their natural environments typically face rapidly changing selection pressures as, for example, exerted by the immune response of the body under viral attack. So Kamp et al.2 have extended quasispecies theory to incorporate an adaptive immune response. In such an environment, a quasispecies becomes subject to a second mutational threshold, this time a kind of mutational 'freeze': if the mutation rate is too low, then the quasispecies does not keep pace with environmental change and becomes extinct. An optimal genomic mutation rate must therefore lie somewhere between mutational freeze and mutational meltdown (Fig. 1).
Kamp et al. have calculated this optimal genomic mutation rate by finding the mutation rate that maximizes viral growth rate in the presence of an immune response. They find that the optimal mutation rate is given by the ratio of the timespan required for the virus to go through an entire replication cycle to the timespan for the immune system to mount a response to a new viral mutant. This result is reminiscent of population-genetic theories that have concluded that a rate of mutation that mirrors the rate of change of the selecting environment is optimal for adaptive evolution4,5,6.
But Kamp et al. go one better than these earlier models in that they incorporate the immune response as an explicit selective force. Comparing their quantitative predictions with data for viral genomic mutation rates, they suggest that many viruses, including HIV, replicate at the optimal genomic mutation rate. Interestingly, their result offers an explanation for the intriguing constancy of genomic mutation rates within viral classes, because the variation in the duration of viral life cycles and the time to mount an immune response is probably considerably smaller than the variation in the rates of mutation per nucleotide.
The idea that viral mutation rates are optimal for escaping host immune responses is appealing7, but some questions remain. First, as previously mentioned, genomic mutation rates in RNA viruses are ten times higher than in retroviruses and 300 times higher than in DNA viruses. This doesn't fit the hypothesis of Kamp et al., because there is no clear evidence for systematic differences in the duration of viral life cycles or the dynamics of the immune responses to these classes of virus. Second, escape from the immune system is not a universal feature of viruses: many viruses may survive by transmission to new hosts before the immune response takes effect.
In addition, the model of Kamp et al. forgoes some of the realism of population-genetic models for the evolution of mutation rates (reviewed in ref. 8). Such models explicitly consider the fate of genes that modify replication and repair — changes in the frequency of these modifier genes, due to the rise or fall of linked beneficial or deleterious mutations, affect the evolution of the mutation rate. This indirect selective force is considerably weakened by genetic recombination, a process that breaks apart linked genes — and a factor not included by Kamp et al. in their model. But recombination is substantial in many viruses, and its effect should probably be considered explicitly in modelling the evolution of viral mutation rates.
Despite these shortcomings, the paper of Kamp et al.2 is clearly an important conceptual development in the study of mutation-rate evolution in viruses. Moreover, developing a fuller understanding of the evolutionary causes and consequences of viral mutation rates is worthwhile from both basic and applied perspectives. Drugs that increase genomic mutation rates can kill off viral populations by causing them to exceed their error threshold9,10. A quantitative theory that can predict how close to the error threshold a given viral population is — without the need to estimate its mutation rate directly — might have real therapeutic value.
Drake, J. W., Charlesworth, B., Charlesworth, D. & Crow, J. F. Genetics 148, 1667–1686 (1998).
Kamp, C., Wilke, C. O., Adami, C. & Bornholdt, S. Complexity (in the press); Preprint cond-mat/0209613 (2002), http://arXiv.org
Eigen, M. Naturwissenschaften 58, 465–523 (1971).
Kimura, M. Genet. Res. 9, 23–34 (1967).
Leigh, E. G. Genetics 73, 1–18 (1973).
Ishii, K., Matsuda, H., Iwasa, Y. & Sasaki, A. Genetics 121, 163–174 (1989).
Nowak, M. A. Nature 347, 522 (1990).
Sniegowski, P. D., Gerrish, P. J., Johnson, T. & Shaver, A. BioEssays 22, 1057–1066 (2000).
Loeb, L. A. et al. Proc. Natl Acad. Sci. USA 96, 1492–1497 (1999).
Crotty, S., Cameron, C. E. & Andino, R. Proc. Natl Acad. Sci. USA 98, 6895–6900 (2001).
About this article
Virus Research (2019)
MicroPro: using metagenomic unmapped reads to provide insights into human microbiota and disease associations
Genome Biology (2019)
Effect of different catalyst preparation methods on the synthesis of carbon nanotubes with the flame pyrolysis method
AIP Advances (2018)
Hepatitis B virus core antigen mutations predict post-operative prognosis of patients with primary hepatocellular carcinoma
Journal of General Virology (2017)