GENETICS - "Gene machinery allows a little self-editing"

[Theresa42]

This is not directly related to H5N1, but I thought it was interesting on a theoretical level given that influenza is an RNA-virus, etc., etc.

Gene machinery allows a little self-editing
Posted 7/30/2006

(...)

An explanation for how RNA might take such a leading role inside cells, one that points to where today's life originated to boot, comes this week in the journal, Science. A team led by Nikolay Zenkin of the Waksman Institute in Piscataway, N.J., reports that when DNA is read by a messenger RNA molecule, botched transcriptions of the genes trigger a reaction that "proofreads most of the misincorporation events," and evicts the botched gene off the messenger RNA molecule. How the Rube Goldberg machinery of genes actually works, transferring their message from DNA to messenger RNA and finally to proteins, with the help of other RNA varieties, may be explained by this self-editing.

Zenkin and colleagues discovered the proofreading effect by dumping bacterial RNA into batches of specially prepared synthetic DNA and RNA dosed to contain botched bacterial genes. Experiments showed the botched genes would readily "participate in their own excision," once written to the bacterial RNA, according to the study.

"The discovery of self-correcting RNA transcripts suggests a previously missing link in molecular evolution," says biologist Patrick Cramer of Germany's Ludwig-Maximilians-Universität München, in a commentary accompanying the study. Scientists had supposed the earliest microbial forms of life found perhaps 3.8 billion years ago on Earth must have had genes carried on RNA, a so-called "RNA world," but how these RNA snippets would assemble into a full-fledged genome for each creature seemed unclear. However, self-editing RNA makes large RNA genomes suddenly seem more inevitable for creatures living in an RNA world, Cramer says.

(...)

http://www.usatoday.com/tech/science/columnist/vergano/2006-07-30-rna-editing_x.htm


Also...

Something for You RNA World Enthusiasts
August 8, 2006

There is a paper in last week's Science that describes a proofreading mechanism in prokaryotic (i.e. bacterial) RNA Polymerase, the enzyme responsible for transcribing DNA into RNA. When RNA Pol incorporates the wrong base into a growing RNA, the enzyme moves two steps back and cleaves the dinucleotide fragment (the mismatched DNA-RNA base pair plus the preceding pair DNA-RNA pair).

So far that's OK. RNA Pol can correct mistakes and thus can transcribe with increased fidelity etc. But the cool part is that the data from this paper supports a model where the mismatched nucleotide participates in the RNA cleavage reaction. So the active site of RNA polymerase's proofreading activity ... is RNA!

(...)

http://scienceblogs.com/transcript/2006/08/something_for_you_rna_world_en.php


And, the research article in Science...

Transcript-Assisted Transcriptional Proofreading

Nikolay Zenkin,1* Yulia Yuzenkova,1 Konstantin Severinov1,2,3*

Fidelity of template-dependent nucleic acid synthesis is the main determinant of stable heredity and error-free gene expression. The mechanism (or mechanisms) ensuring fidelity of transcription by DNA-dependent RNA polymerases (RNAPs) is not fully understood. Here, we show that the 3' end–proximal nucleotide of the nascent transcript stimulates hydrolysis of the penultimate phosphodiester bond by providing active groups and coordination bonds to the RNAP active center. This stimulation is much higher in the case of misincorporated nucleotide. We show that during transcription elongation, the hydrolytic reaction stimulated by misincorporated nucleotides proofreads most of the misincorporation events and thus serves as an intrinsic mechanism of transcription fidelity.

http://www.sciencemag.org/cgi/content/abstract/sci;313/5786/518


Self-Correcting Messages
Patrick Cramer
Mistakes can occur as RNA polymerase copies DNA into transcripts. A proofreading mechanism that removes the incorrect RNA is triggered by the erroneous RNA itself.

http://www.sciencemag.org/cgi/content/summary/sci;313/5786/447

 

[wetDirt]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

Yes, it has been apparent for some time that the 'high native mutation rate' for RNA viruses was due for a good housecleaning. It never made sense to me that you could have anything like a stable genome if it was constantly mutating. And 'purifying mutation' as a correction mechanism sounds like something made up on the spot: "and then all the bad copies magically disappear". The article also described quite a bit of the process of RNA repliccation. Influenza has the complication that the RNA is stuck to a ribonucleoprotien (to hide it from attack), which means it's less mobile during transcription than otherwise. In particular, cleaving of the reverse RNA into littler bits and cutting them loose makes it easy to exchange pieces during forwards-RNA re-assembly, in other words, recombination. A very nice article, I need to read it again.

 

[Racter]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

It never made sense to me that you could have anything like a stable genome if it was constantly mutating.

All genomes are constantly mutating. Influenza just does it a lot more. Considering the mind-boggling numbers of individuals involved, it can afford to -- and in return, it gets the ability to explore new host possibilities, as well as to keep coming back year after year to reinfect hosts who already have immunity. What doesn't make sense to me is a polymerase with the ability to provide that flexibility while at the same time capable of flawlessly accurate transcription when theory seems to require it.

And 'purifying mutation' as a correction mechanism sounds like something made up on the spot: "and then all the bad copies magically disappear".

I assume you meant "purifying selection". It might look like magic, but it's the simplest of logic: what works, replicates; what works a little better replicates a little more; what doesn't work, disappears. The odds of a given mutation being an improvement get longer and longer as a population approaches an adaptive optimum, so more and more mutations get selected out. Thinking of purifying selection as a "correction mechanism" for an organism is like thinking of the ground as a "braking mechanism" for a falling rock.

 

[wetDirt]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

All genomes are constantly mutating. Influenza just does it a lot more. Considering the mind-boggling numbers of individuals involved, it can afford to -- and in return, it gets the ability to explore new host possibilities, as well as to keep coming back year after year to reinfect hosts who already have immunity. What doesn't make sense to me is a polymerase with the ability to provide that flexibility while at the same time capable of flawlessly accurate transcription when theory seems to require it.

I've done a lot of thought experiments on this, and I can't make it work. Dunno if you saw, but elsewhere I went over one of the papers that purportedly showed a high mutation rate, and it had several flaws that made it impossible to tell. The worst is that it was using the phenotype rather than sequencing to confirm that the strain had gone from low to high path. Very unreliable. Better explained by your next bit.
A thought experiment: A clumsy virus is replicating away in the nuculeus, and being distractible, makes lots of errors. Some of these are hopelessly ruined, but who cares. The nucleus releases the new virions, and they go out into the cell, get packaged, and the cell lyses, releasing the little virons to their respective fates. Now a few were so tweaked they couldn't be packaged, but a lot weren't, and were released.

An intrepid researcher then takes a sample from the lysed cell, amplifies and sequences it, and says they have the answer.

Mingus, here, on another thread, describes what dual infections look like in the lab, and I bet not a lot of researchers do what he did to confirm them.
It would be interesting to do the experiment, and see how many different sequences you could get from one cell. I bet you don't see many in single infections.

I assume you meant "purifying selection". It might look like magic, but it's the simplest of logic: what works, replicates; what works a little better replicates a little more; what doesn't work, disappears. The odds of a given mutation being an improvement get longer and longer as a population approaches an adaptive optimum, so more and more mutations get selected out. Thinking of purifying selection as a "correction mechanism" for an organism is like thinking of the ground as a "braking mechanism" for a falling rock.

My mistake, I did mean purifying selection. My point is that replication doesn't discriminate, so we should be seeing more mutants. I don't think we do. And I find it hard to think of PS as correction when there are better ways. In particular, I can't shake the feeling that there are correction mechanisms such as this article describes. It makes a lot more sense that the flu cloud in circulation is heterogenous, multiple infections are common, and most of the mutation seen are actually polymorphs. You don't see what you don't look for. In particular, Mingus' story about dual infections rang a really loud bell. You have to split the sample multiple ways to detect multiple infections, and because of cost, I bet it just isn't done.

 

[Racter]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

I've done a lot of thought experiments on this, and I can't make it work.

Can you be more specific? Exactly what doesn't work for you? Suppose that out of 100,000 virions, 1,000 are viable. Granted, that sounds wasteful, but is there any particular reason to suppose that 1,000 virions would not be enough to sustain infection? What if it were 10,000 out of 1,000,000?

The worst is that it was using the phenotype rather than sequencing to confirm that the strain had gone from low to high path.

Pathogenicity isn't a property of a genotype. It isn't even a property of a phenotype. It's a property that emerges out of the complex interaction between virus and host. We've come a long way, but we still can't just look at the code and go: "blond, brunette, redhead". The Rube Goldberg metaphor above is apt; virus is as virus does.

It would be interesting to do the experiment, and see how many different sequences you could get from one cell. I bet you don't see many in single infections.

It's been done. Plenty. That's just the thing, really. The claims of very high mutation rates in influenza are supported by mountains of empirical evidence. Claims that something else is going on seem to require more than just the insistance that it is obvious from looking at the sequences. The above-linked stories address a fascinating area of study, but one which still involves considerable speculation, and (as the OP notes) one not particularly relevant to the H5N1 issue. I think a good place to start would be with some reasonable definition for "dual infection", but I haven't seen one yet. So far, it seems to mean whatever it needs to mean to support a particular interpretation of the evidence.

And on a pedantic note: it's hard to see the selective value of viral-induced host cell lysis as an explicit design feature of an enveloped virus, as any virions thus released would lack the protein coating aquired during budding from the host cell.

 

[wetDirt]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

Well, I'm sitting and reading the 28 July 2006 Science, and there's this article p514 about clonal adaptive radiation in a constant environment using e coli. Now to make a long story short, they do the clone thing and let it run for several dozen generations, and sequence the kids. They end up with 41 or so strains that they make into a tree. Quotes:"How do we explain the number of mutations in the 26-day population?"(It's 30x too high) Basically: stress and selection. "clonal competetion...may have slowed the emergence of multiple types in efirst weeks." "Purifying selection... are inherenently unlikely when weakly beneficial mutations arise in larte populations subject to persistent selection" "...The large pool of genetic variation available...may be the source material for rarer lateral transfer in times of stress. Much of the microevolutioin seen within bacterial species may be ultimately sourced to clonal diversification." "multidirectional divergence is relevant...in populations...crossing to a new host..." "mutational periodic selections are unlikely to ensure the purity of ...species...in the absence of lateral gene transfer" = recombination "The sharing of a niche by a large number of diversifying members of the same species is a feasible evolutionary strategy. A single fitness solution, or survival of the fittest, is not the only answer in a competitive environment."

And the NEXT article is the one on the self-correction mechanism.

What this says to me is that it's not as simple as playing 52-pick-up and then cleaning up the mess. What this article describes looks awfully familiar to me in looking at H5N1: a big, diverse population with lots of conserved genome variability, lateral gene transfer that both serves to narrow the tree width and conserve beneficial mutations, an error checking mechanism that explains how long, critical sequences can be copied with near absolute fidelity indefinitely, and how survival of the swarm, not the clone, is the trick. This has more the feel of a real system, not the simple 52-pick-up method.

 

[Racter]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

I'd be cautious in attempting to extrapolate the implications of lateral gene transfer in bacteria to its correlate in segmented RNA viruses; the mechanics are quite different, and the consequences in terms of selection surely are as well. But a couple of points:

Purifying selection [is] inherently unlikely when weakly beneficial mutations arise in large populations subject to persistent selection"

Which is another way of saying that as long as there is more to be gained from further change in the direction of some fitness peak, positive selection will tend to drive that -- and conversely, purifying selection is inherently likely to be acting on a population already well-adapted to a stable environment, in which further changes will tend to be away from the fitness peak.

"A single fitness solution, or survival of the fittest, is not the only answer in a competitive environment."

Well, environments aren't competitive, organisms are; environments simply present different organisms with different opportunities for successful competition. A single fitness solution is not the only answer where local design space offers multiple solutions. But in a virus well-adapted to avian hosts, it's hard to see mammalian polymorphisms as part of any solution while the virus occupies those hosts.

What this says to me is that it's not as simple as playing 52-pick-up and then cleaning up the mess.

I'd like to be able to say that it's even simpler than that: playing 52-pickup and letting the mess clean itself up -- except that the problem remains of resolving the known mutability of influenza (which I regard as well-established) with the observed conservation of some portions beyond what is easily explained by selection. That this represents a highly sophisticated system of near-perfect yet selectively flexible error-correction seems to require not only considerable "self-knowledge" on the part of the virus -- which must distinguish between those portions of its genome that must be copied perfectly and those which require variation -- but an "eye to the future" as well; it assumes that the virus displays a capacity for intent in manipulating its own genome in a manner conducive to the aquisition of new host species. That sounds like the sort of extraordinary claim which is usually expected to be supported by extraordinary evidence. If such evidence exists, I'd love to see it. Truly. I've been awestruck before by the wonders of nature, and I always look forward to repeating the experience. But sequences aren't theory any more than fossils are theory.

...survival of the swarm, not the clone, is the trick

Too bad this doesn't apply to humans, eh? And by the way, why is that?

 

[wetDirt]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

I'd be cautious in attempting to extrapolate the implications of lateral gene transfer in bacteria to its correlate in segmented RNA viruses; the mechanics are quite different, and the consequences in terms of selection surely are as well. But a couple of points:

It's not much different from convergent evolution in my mind, except the other way around: Probably, the DNA world first learned the trick from the RNA world. Lateral gene transfer is known in the RNA world. And I seriously think that it hasn't been found in negative-sense RNA viruses because nobody has been looking for it very hard. Just like an assumption of a high mutation rate swamps the urge to locate error checking mechanisms. It won't be found if nobody looks for it.

Which is another way of saying that as long as there is more to be gained from further change in the direction of some fitness peak, positive selection will tend to drive that -- and conversely, purifying selection is inherently likely to be acting on a population already well-adapted to a stable environment, in which further changes will tend to be away from the fitness peak.

I don't think the fitness peak model works very well, because the landscape is constantly shifting. The bug crosses species lines, has to evade both avian and mamallian immune systems, with differing biochemistry and temperature sensitivities, and doubtless many other fitness problems. It has to deal with issues from cellular immunity to bird migration pathways and seasonality. It need, within a very tight genome space, to be able to react over a shorter timespan than a couple of weeks, which is probably the minimum you need to start up new strains by sorting out random mutations. This virus can turn on a dime. It just isn't consitent with your little fitness peaks.

Well, environments aren't competitive, organisms are; environments simply present different organisms with different opportunities for successful competition. A single fitness solution is not the only answer where local design space offers multiple solutions. But in a virus well-adapted to avian hosts, it's hard to see mammalian polymorphisms as part of any solution while the virus occupies those hosts.

Then you aren't seeing a wonderful trick: There wouldn't be a selective advantage in keeping them around if they weren't useful, so they must be useful. How long have humans had flocks? Forty thousand years? Twenty? That's plenty of time to have set up a system that can jump to mammals. If the polymorphisms aren't detrimental, then they just represent 'stored memory': the ability to laterally swap in chunks of code that has been proved to be useful in infecting mammals. In my mind, this is the perfect way to expand the genome without changing its basic size or properties.

I'd like to be able to say that it's even simpler than that: playing 52-pickup and letting the mess clean itself up -- except that the problem remains of resolving the known mutability of influenza (which I regard as well-established) with the observed conservation of some portions beyond what is easily explained by selection. That this represents a highly sophisticated system of near-perfect yet selectively flexible error-correction seems to require not only considerable "self-knowledge" on the part of the virus -- which must distinguish between those portions of its genome that must be copied perfectly and those which require variation -- but an "eye to the future" as well; it assumes that the virus displays a capacity for intent in manipulating its own genome in a manner conducive to the aquisition of new host species. That sounds like the sort of extraordinary claim which is usually expected to be supported by extraordinary evidence. If such evidence exists, I'd love to see it. Truly. I've been awestruck before by the wonders of nature, and I always look forward to repeating the experience. But sequences aren't theory any more than fossils are theory.

It's not intent. It just a convenient way to keep large amounts of known good code in circulation in a tiny apartment without tripping over it all the time.

Elsewhere I have proposed a simple, non'intentional' way to have fixed and variable regions. It's based on an article in Nature, from a couple weeks ago, regarding a salivary RNA virus: The viral genome is twisted and folded up during transcription. Same as a tie-dye t-shirt. The white parts are the invariant parts, they don't get jostled during transcription, while the colored parts are banged into and have to compete for parts and are in general more distracted. It probably isn't an accident that the most invariant code is in the NP segment, where screwing up is the most dangerous, and it's not hard to imagine that the folding process protects the NP segment more than HA and *brain slips* that other one, the neuraminidase thingy.

Too bad this doesn't apply to humans, eh? And by the way, why is that?

But it does apply to humans. Programmers all the time save good bits of code for reuse later, they just don't clutter up the current program with it. Well, good ones don't, I won't admit to what's in some of mine. But where codesize is an issue, I sure don't leave odd bits in. And who's to say it doesn't, anyway? Skin and eye color and latitude/insolation? Lateral gene transfer during sex? Humans are one of the most omnivorous, weedy, generalist species there is. One chimpanzee is no chimpanzee. The culture, not the individual, is the unit of selection. The polymorphisms are viral culture and long-term memory. There is no intent involved, but the swarm encodes the variability that the individual can't. To me, elegant simplicity.

 

[Racter]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

Hey, mind fixing the botched quotes? One passage you attribute to me doesn't appear anywhere in the thread.

I don't think the fitness peak model works very well, because the landscape is constantly shifting.

No environment is truly static; all fitness landscapes are constantly shifting. Influenza isn't a special case in this regard. Each host species, with the unique design problems it poses for the virus, represents an individual fitness peak.

The bug crosses species lines, has to evade both avian and mamallian immune systems...

The virus has crossed species boundaries. That doesn't mean it has to. It's doing just fine with birds. If it occasionally stumbles into a non-avian host, that's due to dumb luck, not selective pressure. The virus cannot be under selective pressure to further fine-tune and maintain its adaptiveness to avian hosts and at the same time be under selective pressure to adapt to mammalian or other hosts. In doing one, it may get the other for free, but that's not the same thing at all.

This virus can turn on a dime. It just isn't consitent with your little fitness peaks.

Whether you accept or reject the fundamental principles of population genetics, you must apply your conclusion to all organisms. Whether Sewall Wright's "adaptive landscape" is a valid model for considering varying degrees of fitness of different combinations of genes has nothing whatsoever to do with this particular virus.

There wouldn't be a selective advantage in keeping them around if they weren't useful, so they must be useful.

Assumptive leaps like that have often been made in biology, but the results have sometimes been rather embarassing. Caution is advised.

Elsewhere I have proposed a simple, non'intentional' way to have fixed and variable regions. It's based on an article in Nature, from a couple weeks ago, regarding a salivary RNA virus...

Yes, I remember seeing that, and I appreciate that effort. That's just the sort of thing I'm talking about: actual theory. I also do find the sort of thing addressed in the articles linked in the OP quite interesting, just not very helpful toward explaining that which I'd currently most like to see explained.

But it does apply to humans.

Missed my point you did. Let me put it this way. The tragedy of the commons applies at the level of genes, individuals, and sub-populations. "For the greater good" just doesn't fly in the biological world. The most successful replicators are those best suited to replicating in the current host; not those best suited to making the jump to some yet-to-be-encountered host.

 

[wetDirt]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

Hey, mind fixing the botched quotes? One passage you attribute to me doesn't appear anywhere in the thread.

*mutters imprecations* accidentaly deleted nearly complete reply, *struggles with html formatting, loses*

... all fitness landscapes are constantly shifting. Influenza isn't a special case in this regard. ...Whether you accept or reject the fundamental principles of population genetics, you must apply your conclusion to all organisms. Whether Sewall Wright's "adaptive landscape" is a valid model for considering varying degrees of fitness of different combinations of genes has nothing whatsoever to do with this particular virus... (I gathered these two bits together)

The problem I have with this is a scale problem. Fitness crosses about 10 orders of magnitude of scale. I have no idea which pair of glasses to use when looking at fitness landscapes. It's so general as to be almost worthless. Not that I have a problem with the concept, just that it isn't very helpful just now in my understanding.

The virus has crossed species boundaries. That doesn't mean it has to. It's doing just fine with birds. If it occasionally stumbles into a non-avian host, that's due to dumb luck, not selective pressure. The virus cannot be under selective pressure to further fine-tune and maintain its adaptiveness to avian hosts and at the same time be under selective pressure to adapt to mammalian or other hosts. In doing one, it may get the other for free, but that's not the same thing at all.

Why not? Just because you say so? Isn't this a mighty leap of faith? I can think of several reasons why there would be tremendous advantages: hint: they don't have legs or cars.
I'm thinking of the lateral transfer of antibiotic resistance between different species of bacteria, here, as an example of a more general application.

Assumptive leaps like that have often been made in biology, but the results have sometimes been rather embarassing. Caution is advised.

And sometimes not so embarassing. I've gotten more mileage out of saying 'how could this observation be true at the same time as this one' than I have out of saying 'that's impossible / that's an outlier/ that's wrong'.
But it does tell me you appear not to believe the 'travel history' stuff. Interesting.

Yes, I remember seeing that, and I appreciate that effort. That's just the sort of thing I'm talking about: actual theory. I also do find the sort of thing addressed in the articles linked in the OP quite interesting, just not very helpful toward explaining that which I'd currently most like to see explained.

Missed my point you did. Let me put it this way. The tragedy of the commons applies at the level of genes, individuals, and sub-populations. "For the greater good" just doesn't fly in the biological world. The most successful replicators are those best suited to replicating in the current host; not those best suited to making the jump to some yet-to-be-encountered host.

The problem here is a sort of linear thinking, of not using all the scales at the same time. The most successful replicators are the ones that successfully reproduce in a cell, successfully get transmitted, by any expedient means, and complete the cycle. I guess we differ on whether or not there is an advantage to being able to cross species boundaries. From the fact that humans are afflicted with not only Influenza A, but also B and C, I think the jump is made pretty often. The wonderful thing about being an RNA virus is that you have a working knowledge of practically all eukaryotic cells. What virus could be better placed for jumping species? I think jumping species is a consequence of the inherent design of the virus, and not that much tinkering is required to jump. It's a matter of working in the current context and at the same time working in a larger context.

 

[Racter]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

The problem I have with this is a scale problem. Fitness crosses about 10 orders of magnitude of scale. I have no idea which pair of glasses to use when looking at fitness landscapes.

Yes, I agree. The metaphor of a "fitness landscape" is simple, but the phenomena it attempts to model are anything but -- not the least complication is that for any organism, selection is surely acting on multiple traits simultaneously, and some of these traits may be incompatible with each other.

Some of the bloodiest battles in the history of theoretical biology have been fought over the proper boundaries for concepts like the "unit of selection". There is a fundamental incompatibility between "generalist" and "specialist" traits. The generalist is more likely to be the first to exploit a new niche, but once a population is established, selective pressure immediately begins to favor specialization. You can't have both at the same time (at least, not the best of both). Any argument for group selection therefore immediately encounters the problem of reconciling selection at the level of the group with selection at the level of the individual. Evolution cannot see the future. Any trait which confers upon an individual the ability to out-compete its fellows will increase in the population, whether it's good for the group or not. If the niche happens to be a dead-end, that's just too bad. As we speak, organisms highly specialized to vanishing ecological niches are going extinct at a rate often estimated at one every few minutes.

The virus cannot be under selective pressure to further fine-tune and maintain its adaptiveness to avian hosts and at the same time be under selective pressure to adapt to mammalian or other hosts. In doing one, it may get the other for free, but that's not the same thing at all.

Why not? Just because you say so? Isn't this a mighty leap of faith?

In biology, leaps of faith must be made at every turn. Again, the trick is to make them as cautiously as possible. One thing that requires perhaps the most faith is that we can successfully suspend our natural tendency to anthropomorphize. We are often humbled by the beauty and elegance in solutions to biological design problems; just as often, however, we are baffled by the apparent clumsiness and waste. What is most difficult to question is our ability to correctly identify "problems" and "solutions", because these often appear so intuitively obvious to us. Some fundamental principles apply to every design process. One of these is that every change encounters consequences. As any human engineer can attest, not all of these consequences can easily be predicted in advance -- and that's working with a device which really shines at predicting things in advance (the human brain).

We can easily see how a trait such as a nice, long neck might be an advantage to a grazing animal in a sparsely vegetated environment, as it would provide unique access to tree leaves. Long legs, for some reason, seem less obvious as a solution. Even less obvious is that the long legs solution demands a longer neck, lest the beast be made vulnerable to predators by being forced to lie down to drink. We should never be too sure we understand what's going on.

But it does tell me you appear not to believe the 'travel history' stuff. Interesting.

I am agnostic with regard to certain interpretations of the 'travel history' stuff.

I think jumping species is a consequence of the inherent design of the virus, and not that much tinkering is required to jump.

I also agree with that. I just happen to feel that this is due to influenza being inherently sloppy as a replicator, rather than it being an insidiously clever executor of fiendishly complicated plans for the future.

 

[wetDirt]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

The virus cannot be under selective pressure to further fine-tune and maintain its adaptiveness to avian hosts and at the same time be under selective pressure to adapt to mammalian or other hosts. In doing one, it may get the other for free, but that's not the same thing at all.
and
...believe the virus is a sloppy replicator...

I have no idea how you did that double-quote-thing, that's quite a trick

I don't think there is much incentive to fine-tune its adaptation to avian hosts. This virus does not appear to be doing quite that. It appears that, overall, the virus is pretty well adapted already, and most of the time, doesn't kill birds. Something wierd happend at Qinghai to cause it to go hi-path. I don't think that is quite pinned down, yet. To me, what the virus looks like it's doing is just duck and weave and stall for time, waiting for either naive hosts, or staying low-path and just crusing around in more-or-less healthy waterfowl. There aren't strong selective pressures when the prey isn't inconvenienced much. Trouble occurs when it encounters mammals. Now the virus is stressed. Wrong temperature, wrong organs, wierd immune systems, all kinds of stressors. An avian virus that finds itself in a mammal may on occasion encounter one of its long-lost cousins and do some quick lateral transfer, and speed up the tedious process of adaptation by just cherry-picking good code, rather than playing 52-pickup till a good idea turns up. Hybrid vigor and 'all that' ensues. And the recent literature says that stress causes more mistakes in transcription than calm. This I can buy. But it's not the same thing as inherently sloppy replication. It's the best of both worlds. There are two different ways to do the same thing, and they are context sensitive. It's not surprising to me that this hasn't been teased out in the lab before. The lab is inherently stressy, at least more stressy than a nice warm duck.

 

[Racter]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

I don't think there is much incentive to fine-tune its adaptation to avian hosts. This virus does not appear to be doing quite that. It appears that, overall, the virus is pretty well adapted already, and most of the time, doesn't kill birds.

I agree. At this point, it probably has little to gain from further adaptation, at least in those avian species it has already aquired; it's about maintenance (i.e., "stabilizing" or, if you prefer, "purifying" selection). Again, the simplest of logic demands that as an organism approaches an adaptive optimum, the difficulty of aquiring further adaptive changes increases, and the payoff decreases.

Something weird happened at Qinghai to cause it to go hi-path. I don't think that is quite pinned down, yet.

Again I agree. I don't think we have the first clue, actually.

Now the virus is stressed. Wrong temperature, wrong organs, wierd immune systems, all kinds of stressors. An avian virus that finds itself in a mammal may on occasion encounter one of its long-lost cousins and do some quick lateral transfer, and speed up the tedious process of adaptation by just cherry-picking good code, rather than playing 52-pickup till a good idea turns up.

One problem I have with that is that in order for it to work, the long-lost cousins holding the good code must also be stranded inside the strange host; simpler just to postulate that the virus carries a full library wherever it goes -- except that this incurs a cost in overhead which would be justifiable only during the sort of rare and unforseeable event we're considering.

Another problem is that it assumes that the virus can somehow tell when it is stressed, and thereby determine when in needs to trade away some of its "bad" code for some "good" code. When a bacterial cell is stressed, it may (say) initiate conjugation on this basis, but a bacterial cell is just that: a cell, with an autonomous identity and an internal chemistry. These seem to be the minimum requirement for the ability to make such a determination. The virus comandeers host cell mechanisms for replication, and with a gun to my head, I might concede that its control over the host cell might extend to defining "stressed" as some chemical state of the host cell. I assume that the literature to which you refer, in fact, concerns selenium deficiency, which might be a nice definition of "stressed" as far as the cell is concerned, but it isn't clear why this should stimulate the reported increase in the mutation rate of the virus. What's even less clear is how the virus might use host cell chemistry as a clue to its own success in responding to the challenges of a new environment such as those you mention -- and even less clear is how such sophisticated functionality, once developed, could possibly hope to survive the crossing of a species boundary, particularly bird-to-mammal.

Hint: nesting quotes is no trickier than nesting for...next loops

 

[wetDirt]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

I agree. At this point, it probably has little to gain from further adaptation, at least in those avian species it has already aquired; it's about maintenance (i.e., "stabilizing" or, if you prefer, "purifying" selection). Again, the simplest of logic demands that as an organism approaches an adaptive optimum, the difficulty of aquiring further adaptive changes increases, and the payoff decreases.

But now I disagree again. There is always going to be some pressure on the virus from the ducks, where ducks are defined as a generic host immune system, where always is defined as 'for the forseeable future'. I see the avian polymorphs as stored code for keeping things lively and preventing complacency on the part of the ducks. Much as human seasonal flu, it prefers (gets better results from) naive immune systems, which means that viral life is tough in wised-up immune systems.

One problem I have with that is that in order for it to work, the long-lost cousins holding the good code must also be stranded inside the strange host; simpler just to postulate that the virus carries a full library wherever it goes -- except that this incurs a cost in overhead which would be justifiable only during the sort of rare and unforseeable event we're considering.

You need to read this story by Mingus:

http://www.flutrackers.com/forum/showthread.php?t=8987

He thinks perhaps there are more recombination events than people might think. I have a hard time thinking that recombination is rare, particularly because I have a vivid imaginiation and seeing, say, the Sacramento delta, and its huge variety of birds, all pooping into the same waterways, makes it easy for me to imagine how recombination events can occur. You'd think it would be rare in factory-produced swine, because they, umm, don't get out much. People probably are less exposed than ducks, but more than swine. I think if we count them up, we might find that the number and persistence of polymorphims goes with the soiciability of the species.

Another problem is that it assumes that the virus can somehow tell when it is stressed, and thereby determine when in needs to trade away some of its "bad" code for some "good" code. When a bacterial cell is stressed, it may (say) initiate conjugation on this basis, but a bacterial cell is just that: a cell, with an autonomous identity and an internal chemistry. These seem to be the minimum requirement for the ability to make such a determination.

The virus comandeers host cell mechanisms for replication, and with a gun to my head, I might concede that its control over the host cell might extend to defining "stressed" as some chemical state of the host cell. I assume that the literature to which you refer, in fact, concerns selenium deficiency, which might be a nice definition of "stressed" as far as the cell is concerned, but it isn't clear why this should stimulate the reported increase in the mutation rate of the virus. What's even less clear is how the virus might use host cell chemistry as a clue to its own success in responding to the challenges of a new environment such as those you mention -- and even less clear is how such sophisticated functionality, once developed, could possibly hope to survive the crossing of a species boundary, particularly bird-to-mammal.

Far as I'm concerned, stress can be any BadThing(tm) that can be detected by an interested party. There was an article the other week discussing how viruses 'know' when it's time to quit reproducing and start enveloping and getting ready to leave the cell. It's simply the amount of crowding in the nucleus, the virus being able to tell when it's bumping into too many of its own cloned bits. At a certain concentration (the stressor), the whistle blows and the virus switches gears. Likewise, a chemical stressor could be running out of a key ingredient like selenium or magnesium. Thermal stress is another example. The ratio of two stressors, like the ratio of crowding to magnesium, could be a simple way to measure how mad the host is getting.

I also think we need to distinguish between single and dual infections, cause I think the rules change in dual infections, and get a lot more interesting. Not that I have much of an idea what the rules are, anyway, though.

What's even less clear is how the virus might use host cell chemistry as a clue to its own success in responding to the challenges of a new environment such as those you mention -- and even less clear is how such sophisticated functionality, once developed, could possibly hope to survive the crossing of a species boundary, particularly bird-to-mammal.

Like I said above. I don't think we've even scratched the surface on how the virus operates. I still see the virus as I see microcontrollers, it is just a cute little machine, so what it does is tighly constrained, and likely to include a lot of very clever tricks to pack so much into a small space. Its lack of codespace argues even more strongly for offsite storage in the form of polymorphisms, and for different reproductive strategies depending on whether it's on its own or has cousins in town. I just don't buy the high random mutation thing, I *know* 52-pickup never works in writing code, and the idea of having to rely on it for matters of life and death is just painful for me. *cringes*

Hint: nesting quotes is no trickier than nesting for...next loops

thanks, I'll try it

 

[Racter]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

I see the avian polymorphs as stored code for keeping things lively and preventing complacency on the part of the ducks.

Well, I'd say you (among others) are actually taking a somewhat stronger position than that. You're arguing for a type of "group selection". As you noted above, deciding "which pair of glasses" to use here is not easy; recognizing which pair is already being used (as the default) is hard enough.

Much as human seasonal flu, it prefers (gets better results from) naive immune systems, which means that viral life is tough in wised-up immune systems.

Life is tough, period. Forever hoping to run off to someplace where it'll be easy doesn't seem like much of a solution, but I suppose that could just be my Protestant work ethic talking. It seems to me that the same qualities that would make a virus a strong contender in an ongoing arms race with immune systems in a population of hosts with partial immunity would also provide its best chance of being able to take advantage of the rare opportunity to gain a foothold in a population of hosts with little or no immunity; it's problematic to postulate that the virus has some special design features enabling it to increase the chances of such an opportunity presenting itself, and it's really not necessary anyway. I mean, there's naive, and then there's NAIVE. We're talking about a new subtype here. Though it is a subtype that may pose a serious threat to us humans, I will argue that it did not -- and cannot have -- evolved specifically in order to do that.

Naive immune systems alone won't do the trick; first and foremost, the hosts must have receptors the virus can bind to, cellular mechanisms it can exploit, etc. A strain of seasonal flu will have a lot better luck having another go at a population of humans it already visited in previous years than it will with a totally naive population of oak trees. An immunologically naive population of humans stretches before the H5N1 virus like a vast banquet table, a fact that we can see so easily that it's hard to appreciate that the virus can't.

You need to read this story by Mingus:

What really jumps out at me is the very last line.

I also think we need to distinguish between single and dual infections, cause I think the rules change in dual infections, and get a lot more interesting. Not that I have much of an idea what the rules are, anyway, though.

I hear you talking there.

I have a hard time thinking that recombination is rare, particularly because I have a vivid imaginiation and seeing, say, the Sacramento delta, and its huge variety of birds, all pooping into the same waterways, makes it easy for me to imagine how recombination events can occur.

It's easy to imagine how reassortment events can occur, but to consider the frequency of recombination involves getting down with the mechanical details of template-switching. The sequence data are compelling as circumstantial evidence.

There was an article the other week discussing how viruses 'know' when it's time to quit reproducing and start enveloping and getting ready to leave the cell.

That's interesting stuff (well, to some of us anyway). But the virions get "enveloped" as they leave the cell, so I'm guessing it was actually about assembly and formation of the capsid prior to budding from the host cell. Influenza doesn't involve a latent phase, so this is an ongoing process that can last as long as the cell holds up.

It's simply the amount of crowding in the nucleus, the virus being able to tell when it's bumping into too many of its own cloned bits.

One of the things that has been suggested is that a drop in the availibility of free nucleoproteins might provide a clue. But we weren't talking about the switch from replication to assembly and release of virions; we were talking about the ability of the virus to determine how it was doing in a broader sense -- a much broader sense, as it involves not merely a single cell, not even merely a single host -- but, in the very broadest sense, not even merely a single host species.

I *know* 52-pickup never works in writing code, and the idea of having to rely on it for matters of life and death is just painful for me.

I understand. And I think this is important, because it illustrates how the bias of a human perspective (and particularly that of a human engineer) can be as much of a liability as an asset as far as interpreting what's going on here. The objection you raise ultimately goes right to the heart of evolutionary theory itself. Many people find it so hard to accept "natural selection acting on random mutation" as the sole mechanism for design in biological organisms that they prefer the mythologies of various ancient peoples, and are willing to overlook the logical difficulties those encounter.

The thing to keep in mind is that the human designer begins with goals, and evaluates success within the context of broader goals. We may speak of biological "mechanisms" and evaluate their success as "strategies" within the context of reproductive competition, but these are *metaphors*. As metaphors, if they are so compelling as to disguise from us the fact that they are only metaphors, then perhaps they are a bit too compelling. The map is not the territory (speaking metaphorically).

A single oyster may produce more than 100 million eggs in a single breeding season, only a tiny fraction of which will survive to reproductive age themselves. At no point will a subjective judgement be required to evaluate their success or failure. What is perhaps most significant is that of those that fail, even the ones that fail the most spectacularly will not stop the earth from turning or the sun from shining, nor they will cause the whole ecosystem to freeze, or crash.

 

[wetDirt]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

Well, I'd say you (among others) are actually taking a somewhat stronger position than that. You're arguing for a type of "group selection". As you noted above, deciding "which pair of glasses" to use here is not easy; recognizing which pair is already being used (as the default) is hard enough.

Well, I guess then I am. It's called genetic diversity. I note that mostly traits that don't have selection pressure on them just drift around if they are cheap, and tend to go away if they are expensive. Polymorphisms that are mostly neutral but handy in certain cases would tend to be conserved. I think your 'purifying selection' thing only works against expensive traits, not cheap ones. I agree it's hard to figure out what they do; however, I suspect the popularity of some is not related to teenage viral fads.

Life is tough, period. Forever hoping to run off to someplace where it'll be easy doesn't seem like much of a solution, but I suppose that could just be my Protestant work ethic talking. It seems to me that the same qualities that would make a virus a strong contender in an ongoing arms race with immune systems in a population of hosts with partial immunity would also provide its best chance of being able to take advantage of the rare opportunity to gain a foothold in a population of hosts with little or no immunity; it's problematic to postulate that the virus has some special design features enabling it to increase the chances of such an opportunity presenting itself, and it's really not necessary anyway. I mean, there's naive, and then there's NAIVE. We're talking about a new subtype here. Though it is a subtype that may pose a serious threat to us humans, I will argue that it did not -- and cannot have -- evolved specifically in order to do that.

I've been mulling this all day now. In my mind, what happened with H5N1 created a combination that resulted in a simpler, less specialized strain. It found a least common denominator where the same code would run on birds, swine, cats, dogs, hyenas, mice, and ferrets. Note that now the human clade has drifted off from the avian one and is either randomly or not randomly acquiring human-style polymorphisms. The bird strain is drifting in directions that better adapt it to birds, now that it has been around awhile and there aren't that many ducks and geese that haven't met up with it in its main haunts. It's not clear that the Karo strain would infect chickens, because none of them have gotten it back from the humans yet. I guess I don't agree that the virus is a highly evolved precision machine; it sure acts like an opportunistic generalist, an omnivore, if you will.

Naive immune systems alone won't do the trick; first and foremost, the hosts must have receptors the virus can bind to, cellular mechanisms it can exploit, etc. A strain of seasonal flu will have a lot better luck having another go at a population of humans it already visited in previous years than it will with a totally naive population of oak trees. An immunologically naive population of humans stretches before the H5N1 virus like a vast banquet table, a fact that we can see so easily that it's hard to appreciate that the virus can't.

It's easy to imagine how reassortment events can occur, but to consider the frequency of recombination involves getting down with the mechanical details of template-switching.

Cool. Can we? I don't have the foggiest how template-hopping occurs. Niman says that it can make two complementary copies at onece, but I'm at a complete loss how that might look. I've been rummaging around for a bead-and-stick cartoon that shows it. That article on the RNA-RNP complex may have done some damage to my brain that still hasn't healed, though.

The sequence data are compelling as circumstantial evidence.

Oh. I forgot. Yeah, circumstantial, I guess.

That's interesting stuff (well, to some of us anyway). But the virions get "enveloped" as they leave the cell, so I'm guessing it was actually about assembly and formation of the capsid prior to budding from the host cell. Influenza doesn't involve a latent phase, so this is an ongoing process that can last as long as the cell holds up.

Might have been. The article's at work and I can't see it from here.

But... we were talking about the ability of the virus to determine how it was doing in a broader sense -- a much broader sense, as it involves not merely a single cell, not even merely a single host -- but, in the very broadest sense, not even merely a single host species.

Well, I would have to agree that it's unlikely there is a way to provide that kind of feedback, and I'm not sure that's what I meant. I meant more of a cell environment feedback, and it was purely mechanical. By the way, do you know whether they control the kind of lighting in labs that do virus work? Is it all fluorescent? Do they check for the sensitivity of the virus to shortwave light? Just wondering.

I understand. And I think this is important, because it illustrates how the bias of a human perspective (and particularly that of a human engineer)

While I am forced to work with engineers, I rejoice that I am not one of them. *hackles go back down*

can be as much of a liability as an asset as far as interpreting what's going on here. The objection you raise ultimately goes right to the heart of evolutionary theory itself. Many people find it so hard to accept "natural selection acting on random mutation" as the sole mechanism for design in biological organisms that they prefer the mythologies of various ancient peoples, and are willing to overlook the logical difficulties those encounter.

The thing to keep in mind is that the human designer begins with goals, and evaluates success within the context of broader goals. We may speak of biological "mechanisms" and evaluate their success as "strategies" within the context of reproductive competition, but these are *metaphors*. As metaphors, if they are so compelling as to disguise from us the fact that they are only metaphors, then perhaps they are a bit too compelling. The map is not the territory (speaking metaphorically).

A single oyster may produce more than 100 million eggs in a single breeding season, only a tiny fraction of which will survive to reproductive age themselves. At no point will a subjective judgement be required to evaluate their success or failure. What is perhaps most significant is that of those that fail, even the ones that fail the most spectacularly will not stop the earth from turning or the sun from shining, nor they will cause the whole ecosystem to freeze, or crash.

I can't say this is a good characterization of my understanding of natural selection in general. I have no problem with the idea that the virus has to produce a heap of progeny to have a chance of sucess, much like the busy oyster. For both Mr Oyster, and H5N1, the cost of churning out all those copies must be very low. In the case of the virus, it's not using its own machinery, but the poor cell's. The oyster is stuck in the sand, what else has it got to do all day anyway? But given the uncertainty of fate, it would be better to not churn out largely defective progeny; that seems like a poor system to me, but if it works, well, OK. But only if that is actually what is happening.

I'm not saying that I have a problem with the general idea of random mutations. What I am really having a problem with is using a high native random mutation rate as a hypothesis, not a conclusion. I'm not ready to buy it till I see more proof that the rate is as high in vivo as it is claimed to be in vitro. So far just in the last month I have learned that RNA viruses have an error-correcting mechanism that I was told they didn't have. I learned that at least one study claiming the mutation rate was high used contaminated sera, and used the phenotype, not the genotype, for evidence of mutation. And I learned that the claim that dual infections are rare isn't exactly true, what's more true is that they are seldom looked-for. And I've learned that there are long strings of seemingly invariant code found here and there, which forces the idea that the 'mutation rate' is not homogenous and isotropic. This makes me uncomfortable with endorsing a high mutation rate, it's still early, and who knows what I might learn this week?

Furthermore, I am expected to believe that polymorphisms are randomly generated over and over sequentially along migration pathways. Note that Niman's polymorphisms as used in the travel histories are not single points, they are embedded in chunks of code 25 or so bytes long. What I am being asked to believe is that a 25-byte string in the same position has randomly arisen independently in time and space to make the appearance of travel around the globe over several years. You want me to believe this happens not just in one of the 8 gene segments, but simultaneously in all 8 segments, because he states that there are marker strings in all 8 segments that tend to travel together.
You must be a teacher or something, you sure make me think. And at my age, too.

 

[Racter]

Re: "Gene machinery allows a little self-editing"

Re: "Gene machinery allows a little self-editing"

It's called genetic diversity. I note that mostly traits that don't have selection pressure on them just drift around if they are cheap, and tend to go away if they are expensive.

Well, selective pressure is obviously going to be stronger on some traits than others, but the only portions of a genome that are invisible to selection are those which don't produce "traits" (which are emergent properties of phenotypes, built out of "proteins", interacting with their environments). When we start talking about wobble bases, I still get lost in the recursion: because the third bases are invisible to selection at the level of proteins, they may be exploited by the virus for use in a fiendishly complicated scheme by which it directs its own evolution, a scheme which is itself visible to selection on a higher level (or is it lower?), though critical elements of it remain invisible (the fact that its elements are invisible are what makes it visible). A frustrated voice inside my head keeps demanding to know: are the third bases visible to selection, or not?

Genetic diversity may well favor a population of organisms, but selection -- acting on individuals -- is constantly working to reduce that diversity, as it favors only those traits that benefit the individual, right here, right now. Maybe the profound fecundity and mutability of the virus (along with horizontal gene transfer) change the rules a little, allowing it to use poorly adapted mutants as storage space for code. This might enable it to stay one step ahead of selection, but it seems like a mighty delicate proposition.

Note that now the human clade has drifted off from the avian one and is either randomly or not randomly acquiring human-style polymorphisms.

A continuous chain of human infections would support that. Human-to-bird transmission would also support it, but I still don't see how it could hold up very long under selection in avian hosts. The only other thing I can think of is a yet-unidentified mammalian reservoir.

I don't have the foggiest how template-hopping occurs.

There is a lot of work currently being done in this area, and you can spend long hours reviewing the literature and come away more confused than when you started out. Maybe there's a Template Switching During Viral RNA Replication For Dummies out there somewhere, but most of the stuff I've seen looks something like this:
http://jvi.asm.org/cgi/content/full/80/5/2337

Any theory relying on template switching as part of what we might call a "heuristic method for optimizing its search of design space" owes a large debt, that being the obligation to provide the details. An in-depth discussion is probably beyond the scope of what we can realistically hope to cover here, but one general observation I'd make is that the autonomy of viral polymerase (which seems a vital component of any such theory) is being brought into question by the findings of a number of researchers. Here are a couple of links relevant to that:
http://users.path.ox.ac.uk/~efodor/index.htm
http://www.nature.com/emboj/journal/v21/n1/full/7594236a.html

I'm assigning this as homework, partly because I expect time constraints to be limiting my participation here for at least the next few days.

That article on the RNA-RNP complex may have done some damage to my brain that still hasn't healed, though.

No pain, no gain, as they say.

I am expected to believe that polymorphisms are randomly generated over and over sequentially along migration pathways.

No, I can't accept that either. Genetic information is clearly being exchanged in chunks smaller than what reassortment plus random mutation can easily explain.