The recent series of posts on polymerase error rates and viral evolution has elicited many excellent and thought provoking comments from readers of virology blog. Here is one that I had not thought of before, and which I’ll use on an exam in my virology course:
Here’s a tough question. In the follow up blog to this, you say that the high mutation rates of RNA viruses is beneficial to survival in a complex environment. If this is true, why don’t DNA viruses evolve high mutation rates also? It would be simple for them to delete their proofreading domain.
There is no answer to this question, so I’ll speculate. I believe that DNA viruses have error correction mechanisms so that they can have very long genomes. RNA viral genomes are no longer than 27-31 kb. This limit is probably imposed by their high error rate: if RNA genomes were longer, they would likely sustain too many lethal mutations to survive. Error correction mechanisms allow for DNA viral genomes up to 1.2 million bases in length. Smaller DNA viruses don’t have their own DNA polymerases – they use those of the cell. Cellular DNA polymerases have error repair to avoid mutations that lead to diseases such as cancer. Both forms of reproduction are evolutionarily sustainable: shorter RNAs with lots of errors; longer DNAs with fewer errors.
The reader who posed the original question then came back with this retort:
If reduced fidelity is beneficial to RNA viruses, because of the complex environment they are in, why don’t DNA viruses do the same thing?
I think the same endpoint, in terms of surviving in complex environments, is achieved by both strategies. RNA viruses have high diversity; DNA genomes have many more gene products which allow them to survive in diverse situations. Both strategies appear to be evolutionarily sustainable.
It’s important to keep in mind that the goal of viral evolution is survival. Evolution does not move a viral genome from “simple” to “complex”, or along a trajectory aimed at “perfection”. Change is effected by elimination of the ill adapted of the moment, not on the prospect of building something better for the future.
While researching this subject I came across a series of papers on DNA synthesis by African swine fever virus, a virus with a DNA genome of 168-189 kb. The viral genome encodes a complete DNA replication apparatus, including DNA polymerase and DNA repair enzymes. Incredibly, the DNA repair pathway itself is error-prone, which is believed to contribute to the genetic variability of the virus. There is some controversy concerning the error rate for this virus, so we don’t know the consequence of this observation for fidelity of DNA replication. Nevertheless, these observations suggest that evolution has seen fit to tinker with, and perhaps increase, the error rates of certain DNA based organisms.
Lamarche, B., Kumar, S., & Tsai, M. (2006). ASFV DNA Polymerase X Is Extremely Error-Prone under Diverse Assay Conditions and within Multiple DNA Sequence Contexts. Biochemistry, 45 (49), 14826-14833 DOI: 10.1021/bi0613325
“It’s important to keep in mind that the goal of viral evolution is survival.” So I'm curious, why do (all, most, some?) viruses make the host ill, or worst, kill it? If the virus needs the host replicating mechanisms, killing it is not a very good survival strategy, is it. why isn't it more like a symbiosis, if at all possible?
I'm not entirely certain that there is a broad difference between fidelity in DNA and RNA viruses.
Before you shriek at what seems like my ignorance, let me explain. What I mean to say is that RNA viruses–let's use Influenza as an example–have copying mechanisms that produce a certain number of errors per genome copied. It's possible that it is this number, not the number of errors per base pair, that is important for proper evasion of host defenses without compromising genomic integrity. While a virus could probably sustain a few mutations per genome per copy cycle, more than that are probably bad…
Therefore, by virtue of their larger genomes, DNA viruses should have a lower error rate to avoid increasing the number of errors per genome.
Another thing that no one really seems to think about is that these mutations are not really “random” per se; they are based on thermodynamics, and that science teaches us that when God plays dice, they are loaded.
Not boring at all. it makes sense, even as an speculation. I haven't seen it like that before. I'll be thinking about your thoughts for a while…
thanks for the answer
It’s interesting that you bring up Retroviridae and they have reverse transcriptase. Moloney murine leukemia virus is thought to make errors every 1 in 30000 which is a little better that Influenza's RNA dependent RNA polymerase. Could this be a kind of middle ground between the RNA error rate and the DNA error rate? Since Retroviridae store their genome as RNA but translate new viron particles using DNA so they have error rates from both camps including the error rates from reverse transcriptase. This may give HIV some advantages that may not be realized by RNA or DNA only viruses.
Having lower error rates in transcription also enables more consistent assembly of very large genomes.
Lynn Margulis provides some fascinating insight into this.
Many years ago this biologist, influenced by the perspectives of Gregory Bateson, was the first to postulate that mitochondrial DNA had a distinct evolutionary history than the rest of cellular DNA. This view was met with widespread snickering and derision at the time.
Now it is not only mainstream thinking, but most geneticists believe that all sorts of other intracellular organelles also evolved with distinct and separate evolutionary trajectories.
In her latest book “Acquiring Genomes: A Theory of the Origins of Species”, Ms. Margulis revolutionizes evolutionary theory in a way that will likely reverberate down future generations of evolutionary biologists for hundreds of years.
What Ms. Margulis meticulously describes in this book are evolutionary genomic acquisition strategies extant throughout the natural world, whereby organisms incorporate the entire genome of other organisms into new forms.
Discussing examples from diverse species such as wood-digesting bacteria incorporated into termites to the symbiotic coevolution of bacteria inside certain species of blue-green algae, to nitrogen fixing bacterial colonies in legumes, her pioneering work is a compelling read.
Controversial? Yes. But she is a microbiologist basically with very few peers and her pioneering work in mitochondrial DNA makes her views non-trivial.
Why this is relevant to the discussion at hand is as genomes become exponentially more complex through genomic acquisition, reduced transcription error rates of sigma 9 (i.e. an error is 9 standard deviations from the norm) become necessary.
Here is a link to Lynn's book:
http://tinyurl.com/qopp7d
So we find DNA viruses that exploit both error-prone and high-fidelity lifestyle niches.
Doesn't the real question the become why haven't we found a high-fidelity viral RNA polymerase? Perhaps it just comes down to nucleic acid chemistry? Or maybe the truth is still out there…
Does the ecologist's concept of r/K selection bring something to this discussion?
I.e. the high fidelity reproductive strategy is K selected and the high diversity reproductive strategy is r selected. The ecologist model suggests that these strategies tell us something about the nature of the niche. The r strategies fit a situation where things are unstable; in that case lots of resources are expended on searching for an available niche. The maple tree drops millions and millions of seeds in the hope one will find a suitable bit of real estate for the next generation tree. Similar the virus is looking for a bit of real estate on a landscape with a very aggressive owner.
Ah, it's too much fun playing with ideas outside one's area of expertise :).
The r/K idea seems like a good hypothesis. If RNA viruses are known to have high 'fecundity' compared to DNA viruses then this could be an answer…….can any of the virologists answer this one???
Anthony's comment about nucleic acid chemistry also seems like a valid hypothesis. The question would be in this case whether the presence of a 2'OH in RNA prevents proofreading by RNA polymerase or reverse transcriptase…..I think I need to do some reading!
The r/K idea seems like a good hypothesis. If RNA viruses are known to have high 'fecundity' compared to DNA viruses then this could be an answer…….can any of the virologists answer this one???
Anthony's comment about nucleic acid chemistry also seems like a valid hypothesis. The question would be in this case whether the presence of a 2'OH in RNA prevents proofreading by RNA polymerase or reverse transcriptase…..I think I need to do some reading!