There are tantalizing hints of a deeper teleology in Wagner’s book – hints that put his arguments under the critical scrutiny of the paleontologist Douglas Erwin, when the latter spoke at the University of Chicago’s Evolutionary Morphology Seminar (I was there).
The final paragraph of Arrival of the Fittest (p. 221) is loaded with teleological implications (my emphasis):
"When we begin to study nature’s libraries we aren’t just investigating life’s innovatibility of that of technology. We are shedding new light on one of the most durable and fascinating subjects in all of philosophy. And we learn that life’s creativity draws from a source that is older than life, and perhaps older than time."
Sound Platonic? It is: the title of this chapter is “Plato’s Cave.”
In other publications, Wagner has been outspokenly skeptical of the “randomness” premise in conventional evolutionary theory. Erwin definitely did not like Wagner’s alternative, however; in no sense, Erwin urged, could the pathways through sequence and function space have been preloaded into the universe. I wish Wagner had been present that evening to defend his thesis.
Axe’s study was focused on stable protein folds. His assay used a function but the study was designed specifically to test the viability of folds. In a general context a stable fold and function are not necessarily related. In a biological, i.e. designed, context we have come to expect the correlation of stable folds with function, but in the sequence space we have no reason to expect that because of many factors already mentioned.
However, it seems obvious that most real biological functions require stable folds, with the exception of “proteins” that even ID theorists would accept evolved like the antifreeze proteins in the icefish. The reverse, that most folds require functions, is certainly not true. So if you want a better generalized estimate of the prevalence of functional folds, you would want to study the folds and assume that every fold is potentially functional. That assumption is of course false, but it’s a generous assumption, meaning that to the degree it is false functions will be more rare than folds. That is why Axe designed his study on folds rather than function, and it makes sense that his general estimate would be in the ballpark.
Seems a simple matter to knock out the editing apparatus and see how it works in a model system. Has this been done? Or are they saying that it’s unlikely to work today but may have worked in the LUCA’s early lineage since we just don’t know what that was and can’t test it?
I would agree there is probably some correlation between functions and folds, though there are caveats as you note. But I don’t see any reason to think his estimate is even in the ballpark. If his estimate was really in the ballpark, a host of laboratory experimental results should not have obtained, and the immune system should be basically powerless to find and tune functional antibodies towards novel substrates.
There are experimental results that put his estimate off by something on the order of 65 orders of magnitude.
Axe’s study started not with a normal protein, but with a temperature-sensitive mutant, literally selected to be on the razor’s edge of stability.
How on earth could you have missed that?
And it was only one fold. Folds are structural classifications.
That doesn’t seem obvious to me, but then again I study how real pathogenic mutations affect real protein stability and real function. Hint: plenty of real pathogenic mutants make proteins MORE stable.
How about the prion protein?
Obviousness is a really, really bad criterion to use in biology.
Some theorists do try to model cellular information transfer in the absence of editing. Carter and Wills (2018), for instance, argue that the RNA World could not have been progenitor of extant cellular information transfer systems, but that “principles of self-organisation that transcend Darwinian natural selection” are required. As they elaborate:
The direct evolution of inherited genetic information coupled to encoded functional proteins, as is observed in real-world molecular biology, is far more plausible than any scenario in which there was an initial RNA World of ribozymes sophisticated enough to operate a genetic code. The preservation of encoded information processing during the historically necessary transition from any ribozymally operated code to the ancestral aaRS enzymes of molecular biology appears to be impossible, rendering the notion of an RNA Coding World scientifically superfluous.
From here:
I have much more experience with a parallel evolutionary problem: suspending the normal rules of metazoan development to allow for the diversification of body plans at events such as the Cambrian Explosion. No one ever describes how these proto-animals with labile ontogenies would develop from a fertilized egg, as every model system we actually know is rule-governed.
We were discussing function in sequence space. “Prevalence of functional folds” is a red herring, because your assumptions about stability vs. function are incorrect.
“On the razor’s edge of stability” counts as stable. If you want to study the limits of fold stability you would have to study the limits of fold stability.
I have a hard time understanding what you mean by “objectively false” when you admit one sentence later that what I said was objectively true.
That doesn’t follow from what he said. There’s a difference between the set of proteins currently used by extant life, and all possible functional proteins. It is thought that many proteins begin, when they first evolve from non-coding DNA, as intrinsically disordered proteins, at most having a short functional motif, but no stable defined higher order 3-dimensional structure and some examples are known.
This suggests there is a substantial portion of functional peptide sequences that don’t fold into stable structures, though they could potentially be evolved incrementally towards acquiring a stable fold.
In particular see the introduction and discussion in this paper:
1: Bungard D, Copple JS, Yan J, Chhun JJ, Kumirov VK, Foy SG, Masel J, Wysocki
VH, Cordes MHJ. Foldability of a Natural De Novo Evolved Protein. Structure. 2017
Nov 7;25(11):1687-1696.e4. DOI: 10.1016/j.str.2017.09.006
No, it doesn’t. It was selected to be far less stable than wild-type, so that it doesn’t work at a couple of degrees above normal temperature.
So, why did Axe start with a ts mutant instead of the far more stable wild-type enzyme? And why didn’t you know that he did?
Now another neologism. There’s no reason to use the word “fold” as an adjective for “stability.” Again, Axe’s and Gauger’s use of “functional fold” is just a red herring. Let’s talk about protein function and protein stability. Folds are structural classifications.
Because it is objectively false. TS mutants are unstable. Can you try to hold back from asserting things you clearly don’t know? Then when your false assertions are corrected, you become hostile. Try pretending to know less than you know instead of pretending to know more than you know.
That’s a straw man based on naive, dichotomous thinking. The functions of many, probably most, proteins involve transitions between different conformations. At best, one can call them metastable.
Proteins that are really, really stable, like PrP in the prion conformation, are toxic because they are insoluble.
Suggestion: read a primer about muscle contraction, particularly the central role of calcium concentration making a radical change in the structure of the troponin complex. It’s the most complex system that we fundamentally understand, and it’s full of instability.
So you think that functional conformational changes which are triggered specifically as part of their functional parameters means the protein is unstable. Well there’s just no talking to you then.
