|
Author
|
Topic: In praise of garbage disposals
|
Art
Member
Member # 179
|
posted 01. January 2005 01:13
In his recent paper (Proceedings of the Biological Society of Washington 117, no. 2, pp. 213-239), Stephen Meyer intimated, among other things, that functionality when it comes to proteins is a rare proposition, that functional sequences are exceedingly rare in sequence space. In his paper, he cites a number of studies, among them one by Axe (J Mol Biol. 341, 1295-315, 2004) and one by Taylor et al. (Taylor SV,Walter KU, Kast P, Hilvert D. 2001. Proceedings of the National Academy of Sciences, U.S.A. 98:10596-10601). However, as has been discussed on this board and on ARN, these two studies (as well as others – these frame a rough range of estimated frequencies) provide a range of “improbability” that, without considering other studies that give quite different answers, extend from “rather probable” to “fairly improbable”. Axe tries, in his 2004 paper, to rationalize the apparent discrepancy by claiming that some experimental subjects are more complicated than others. This explanation is unsatisfying (IMO), and other possibilities are both more likely and much more interesting. One of these is the focus of this essay.
The experiments that Meyer embraces are simple in concept but fraught with possible complications. Briefly, researchers systematically introduce patches of random mutations into test “subjects” (enzymes for which there is good structural information and a relatively easy growth assay for presumed function). Mutations (single or combinations) in the subjects that reduce or eliminate bacterial growth are taken to reflect alterations that have eliminated the ability of the mutant protein to acquire a functional fold. However, this experimental approach is not so simple. Bacterial growth in these experiments reflects, not just the ability to attain a functional fold, but also (among other things – the factors that come into play are more numerous than I list here) the ability to interact with the cell’s complement of chaperonins and proteases such that a functional fold is attained before the protein is broken down. This calls into question the extrapolation from growth to fold, and thus to “frequency in sequence space”. This consideration, though, affords a reasonable way to reconcile the very disparate conclusions that different studies (including studies of the same protein!) reach. Indeed, it is possible, probable even, that once matters other than a potential to attain an inherent fold are factored into things, the results of the disparate studies alluded to above will converge.
But that is but an aside, an extended introduction to another, more interesting topic. Michael Denton proposed, a few years ago, that the relatively small number of protein folds that are apparent in protein structure databases are “Platonic forms” – fundamental structures whose shapes (and perhaps functions) reflect underlying rules of chemistry, physics, and design (Denton MJ, Marshall CJ, Legge M. J Theor Biol. 219, 325-42, 2003). The preceding paragraph reveals an alternative perspective. It is not unreasonable to visualize the protein content of a cell as being sculpted by the chaperonin/protease systems of the cell – those proteins (or sequences) that fold rapidly and into structures that are not efficient substrates for the proteases in a cell will “survive”, while others will not. Extant protein folds, then, would not be “Platonic forms”, but rather the results of a microscopic Darwinian ballet. The number of possible protein folds would be much, much more numerous than the 10,000 or so that we can see in extant structures. This small number would represent those structures that survive the Darwinian gauntlet of folding and turnover. (Recent studies cited in other threads on ARN support this suggestion, insofar as they indicate that non-canonical folds are probably fairly easy to attain. Also, estimates of the number of extant folds are all over the board, and 10,000 is but a compromise. But specifics are not that important for this essay.)
This POV suggests a very important, even central, role that folding and turnover plays in biological systems. We might (and perhaps will in the future) see how the very early days of the RNA+Protein World were shaped by proteases. As interestingly, we can start to appreciate some differences between eukaryotes and prokaryotes, and how the perceived complexity of eukaryotes may reflect some interesting differences between prokaryotes and eukaryotes when it comes to protein turnover.
In general, a simple underlying principle applies to general protein turnover in living things – substrates for breakdown are recognized and delivered to a tightly-regulated protease. These processes are invariably ATP-dependent (and thus energy-consuming). In bacteria, several ATP-dependent protease systems have been studied (see Gottesman, Ann, Rev. Cell. Dev. Bio. 19, 565-587, 2003 for a recent review). As a rule (there are interesting exceptions that preclude a clean generalization), substrates for these systems are recognized by chaperonin-like proteins (that are generally considered to be subunits of the complexes) and delivered for breakdown; specificity resides in the substrate itself, and its ability to interact with (or be recognized by) the chaperonin subunit. Occasionally, though, substrates are modified by the addition of adaptor proteins that facilitate the delivery and breakdown. Adaptors are infrequent in bacteria, and it is currently thought that specificity in proteolysis in bacteria resides largely in the interaction between substrate and protease system. In the context of the above suggestion, the range of protein folds that can exist in bacteria would be largely determined by the ability of folds to “escape” from the protease systems in the cell.
The eukaryotic proteolytic system (which will be called here the proteasome) can, occasionally, recognize a substrate “directly”. However, the bulk of the proteins that are degraded by the proteasome are first tagged for breakdown by the covalent addition of a small protein called ubiquitin. Ubiquitin is added to proteins via a cascade of activities – Ub is activated by so-called activating enzymes (E1’s), transferred to ligases by transferring enzymes (E2’s), and then attached to substrates by ligases (E3’s). (The recently-awarded Nobel prize in Chemistry is pertinent to this discussion. Readers can visit this site for more detailed information on the subject.) Of particular interest are the determinants of specificity for this system – the so-called E3 ligases. Eukaryotic cells possess hundreds or thousands of different E3s, and each presumably has a different substrate (= target protein or peptide motif) specificity. These enzymes, along with the components of the proteasome core, afford a great range of combinatorial possibilities, and might be enlikened to antibodies in terms of the range of possible substrates.
In bacteria, the survival of a protein is determined by the inherent abilities of a protein to interact with, and be “shaped” by, the chaperonin and protease systems in the cell. If a protein cannot be efficiently folded (on its own accord or in conjunction with chaperonins) it is broken down. (Overexpression of poorly-“foldable” proteins can also lead to aggregation and removal by a sort of precipitation – but this is not really relevant to this essay.) In eukaryotes, protein breakdown is determined by recognition by enzymes that attach distinctive tags (ubiquitin) that initiate the proteolytic breakdown. Thus, protein turnover in eukaryotes (and thus protein stability) is not so much a matter of interactions with the folding and proteolytic systems per se, but rather one of recognition that is more akin to antibody-antigen recognition. This added layer in the turnover process affords a greater degree of permissible protein complexity in a eukaryotic cell, since a protein’s “life” is no longer a matter of ability to avoid breakdown. Briefly, in prokaryotes, breakdown may be seen as a default pathway (not the only pathway that a “non-functional” protein might take, but one that is pertinent to this essay), while in eukaryotes, breakdown is the pathway that requires active recruitment. This opens up new vistas for protein evolution in eukaryotes. The universe of protein folds in eukaryotes would still be shaped by turnover – the variety of modifying enzymes would still limit the ability of newly-evolved sequences to “survive”, according to the occurrence of recognizable motifs in accessible positions of the new protein.
In closing, I would note that this contrast between bacteria and eukaryotes is not absolute – tags for breakdown exist in bacteria, and the proteasome can occasionally bypass the Ub requirement for recognition. So it is not wise to think of bacteria and eukaryotes as being mutually exclusive in these respects. But the general dynamics of the systems are such that some important general differences can be discerned, differences that impact both the “universe” of protein folds in prokaryotes and eukaryotes, and the possibilities for protein evolution in these two kingdoms of life.
And finally, a postscript – keep in mind when reading this that it is a brainstorm, an attempt to reconcile seemingly disparate studies and follow the lead to new and provocative insight.
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 06. January 2005 20:19
As an intro to this reply, I just wanted to get some things just for the record from this ID proponent's point of view.
Art:
quote:
However, as has been discussed on this board and on ARN, these two studies (as well as others – these frame a rough range of estimated frequencies) provide a range of “improbability” that, without considering other studies that give quite different answers, extend from “rather probable” to “fairly improbable”.
Just for the record, I don't think that the links that Art posts give "quite different answers" so as to render it in the range he suggests. But I'll leave that for others to decide. For more information on that, I recommend seeing this and this just for starters.
Art writes:
quote:
Axe tries, in his 2004 paper, to rationalize the apparent discrepancy by claiming that some experimental subjects are more complicated than others.
Admittedly I only have acess to the abstract of this paper. But for the life of me I cannot see where Axe is rationalizing any discrepancy such as that described in the OP. Unless I'm looking at the wrong paper? However, this is also a side show.
Getting to the meat of the OP, the comments below are quick and dirty, free time is still a little scant. Nonetheless, the problem I see with the OP's suggestion is that even without very specific recognition regions, unfolded proteins are generally sensitive to cellular proteases. At some level it must be true that proteins must fold into forms that are reasonably resistant to proteolysis and peptidases. However, we know that unfolding proteins (mutations, or heat, for instance) allows many proteins that are not normally substrates to be degraded.
Moreover, we know that putting a "tag" on a protein that is recognized by the proteases allows processive proteolysis through the whole protein, which would otherwise be stable. Therefore, I'm not sure we can say that folds within the protein are necessarily limited by the necessity to be resistant to proteolysis.
Finally, the inability of certain forms of proteins (the aggregates involved in some prion-like diseases) to be recognized and degraded is clearly rather important in the diseases these proteins may play. Being too resistant, even when misfolded, is a dangerous thing.
[ 06. January 2005, 20:24: Message edited by: Nelson-Alonso ]
IP: Logged
|
|
Art
Member
Member # 179
|
posted 06. January 2005 23:10
Hi Nelson,
I think you are leaping beyond Meyer's paper to other aspects of the "CSI in proteins" issue. For this thread, I wish to consider just those studies mentioned by Meyer. These studies (typified by Axe [2004] and Taylor et al. [2001]) use essentially identical experimental approaches but yield estimates for the frequency of functional sequence between 1 in 10^24 and 1 in 10^77 - 50 orders of magnitude! It is this discrepancy that I think the contributions of things like folding rates and protease sensitivity can help explain.
(For those who are interested, some other studies that address the larger issue Nelson refers to are mentioned in this discussion.)
Beyond that, I think you touch on one observation that supports the distinction I made between prokaryotes and eukaryotes - that is the existence of things like prions. As far as I know, there is nothing like a prion in prokaryotes. I would explain this with my suggestion that, in bacteria, degradation is the default. Thus, misfolded or unfolded proteins simply will not accumulate. This would not be true in eukaryotes, because degradation is an active process. If prion-like entities are not tagged for breakdown, then they can accumulate, etc.
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 07. January 2005 02:41
Art write:
quote:
I think you are leaping beyond Meyer's paper to other aspects of the "CSI in proteins" issue. For this thread, I wish to consider just those studies mentioned by Meyer. These studies (typified by Axe [2004] and Taylor et al. [2001]) use essentially identical experimental approaches but yield estimates for the frequency of functional sequence between 1 in 10^24 and 1 in 10^77 - 50 orders of magnitude! It is this discrepancy that I think the contributions of things like folding rates and protease sensitivity can help explain.
Perhaps, like I said, I havn't read all of Axe's paper. But I did read all of Meyer's paper. I don't see anything in Axe's paper that supports your contention of a "rationalization" (your word)something that Axe had to try to reconcile. That of course was just a side a show.
But I'm still interested in where you got that from.
Art writes:
quote:
Beyond that, I think you touch on one observation that supports the distinction I made between prokaryotes and eukaryotes - that is the existence of things like prions. As far as I know, there is nothing like a prion in prokaryotes. I would explain this with my suggestion that, in bacteria, degradation is the default. Thus, misfolded or unfolded proteins simply will not accumulate. This would not be true in eukaryotes, because degradation is an active process. If prion-like entities are not tagged for breakdown, then they can accumulate,
This has absolutely nothing to do with what I was responding to unless I'm mistaken. My response was geared toward your "brainstorm" about evolution of protein folds starting from the ool, however, if you would like to start a new discussion based on prions I would be more than game. But if I don't see any relevance to this thread, I'm going to have to ask you to start a new one.
[ 07. January 2005, 02:45: Message edited by: Nelson-Alonso ]
IP: Logged
|
|
Art
Member
Member # 179
|
posted 07. January 2005 22:21
quote:
I don't see anything in Axe's paper that supports your contention of a "rationalization" (your word)something that Axe had to try to reconcile. That of course was just a side a show.
But I'm still interested in where you got that from.
I've got the preprint on my computer, so the page numbers aren't there. But from the 14th page of the paper (sorry for not trying to fix the formatting):
quote:
"As discussed in Introduction, the method applied in the study of chorismate mutase by Taylor and coworkers (9) should provide a more accurate estimate than the earlier (lambda)-repressor study. Their search for functional chorismate mutases was restricted to sequences matching the hydropathic pattern of a natural version of the enzyme. So, bearing in mind the difference between a single-sequence pattern and a multi-sequence signature, their estimated functional prevalence should be compared to the estimated prevalence among signature-compliant sequences in the present study. Scaling their figure gives 10K40 for a 153 residue sequence (10–24(153/93)Z10K40). This is significantly larger in logarithmic terms than the above estimate for the large domain (10–64). However, in view of the difference in fold complexity (Figure 1) and the fact that pattern-based randomization is more restrictive than a signature-based randomization, there is no reason to think the two estimates are inconsistent.
Here's the place to continue discussing the relationship of the "garbage disposal" to Axe's work.
Regarding my comment on prions:
quote:
This has absolutely nothing to do with what I was responding to unless I'm mistaken. My response was geared toward your "brainstorm" about evolution of protein folds starting from the ool, however, if you would like to start a new discussion based on prions I would be more than game. But if I don't see any relevance to this thread, I'm going to have to ask you to start a new one.
Nelson, I think you've misunderstood the gist of the OP. It's not much about the ool (although there are ramifications in this regard), but rather about how breakdown - the "garbage disposal" - can define the universe of protein folds in general.
To outline things for future reference:
1. The reference to Meyer's paper is a pretext. Use the above link to discuss matters to this subject.
2. One of the two main points of the OP is the alternative I lay out to Denton's "Platonic Forms" idea.
3. The second is the general (but not universal, absolute) difference between prokaryotes and eukaryotes in terms of targeting proteins for breakdown, and the possible implications of this difference.
[ 07. January 2005, 22:23: Message edited by: Art ]
IP: Logged
|
|
Nel
Member
Member # 614
|
posted 08. January 2005 02:38
Art writes:
quote:
I've got the preprint on my computer, so the page numbers aren't there. But from the 14th page of the paper (sorry for not trying to fix the formatting):
I was simply referring to the way you worded the range of probabilities and how Axe dealt with them from rather probable to fairly improbable, I think that description is inaccurate , even taking the Taylor paper into account. I'll leave this issue here, and perhaps continue in that thread.
Art writes:
quote:
Nelson, I think you've misunderstood the gist of the OP. It's not much about the ool (although there are ramifications in this regard), but rather about how breakdown - the "garbage disposal" - can define the universe of protein folds in general.
Yes it was the latter that I mostly addressed (I used the phrase "starting" from the ool). The issue here I think is, has the ability of the protease to recognize a fold evolved to fit what is available and should be degraded, or have the proteins evolved to generally be resistant (perhaps from a non-ID point of view). I realize that had this been purely about the ool, we wouldn't be focusing on the energy-dependant proteases,but perhaps more promiscuous ones with the ability to cut many exposed peptide bonds. My comments then, didn't have much to do with point 3, which is what you were discussing in your last reply to me, and I'll try to comment on that part in further replies (thanks for the clarification).
[ 09. January 2005, 04:55: Message edited by: Nelson-Alonso ]
IP: Logged
|
|
Art
Member
Member # 179
|
posted 15. January 2005 23:01
quote:
The issue here I think is, has the ability of the protease to recognize a fold evolved to fit what is available and should be degraded, or have the proteins evolved to generally be resistant (perhaps from a non-ID point of view).
An interesting thought. Merging this with prior themes, this would suggest that the scope of protein folds that exist in extant cells would reflect the range of folds that existed when the first proteases arose. (This follows from the presumption that, once proteases arose, new folds would not occur because they would be degraded by the proteases.)
The question arises, of course, as to what, in the prebiotic world would determine that a protein "should be degraded"?
IP: Logged
|
|
Art
Member
Member # 179
|
posted 15. July 2005 23:49
Light, the lid, and other things
Back in the late 80’s and early 90’s, scientists began in earnest to explore genetic approaches to understand the role(s) that light plays in affecting the development of plants. The “approach du jour” (still in wide use today – scientists know a good thing when they see one) was to mutagenize seed (Arabidopsis was used for this purpose), collect a few generations of the mutagenized population in bulk, and then screen individuals for phenotypes that are diagnostic for alterations in the responses of plants to light. Among the phenotypes and mutants found were the usual sort, ones that are rather intuitive. (For example, recessive mutations affected in one or another of the several photoreceptors plants have were found.) However, there was a class that proved to be most interesting.
Before describing these mutations, it helps to review, very briefly, the early stages of growth of plants such as Arabidopsis. Suitably-sown seed will germinate, unfurling into an embryonic plant. In the dark, the aerial part of the plant grows long and relatively straight, as well as largely uncolored. When light impinges on the seedling, a series of events is triggered that leads to greening, chloroplast development, and the other steps that lead to the formation of green, photosynthetic tissues (cotyledons). In the absence of light, the short, green cotyledon does not develop; instead, the so-called hypoctyl grows long and unpigmented.
Briefly, it proved possible to identify and characterize a number of recessive mutations that conferred upon plants (when homozygous) the “ability” to undergo hypocotyl shortening and cotyledon greening in the absence of light. The recessive nature of the mutations was provocative – this suggested that there existed a central repressor of global light responses in plants, one that was affected by a class of constitutive photomorphogenesis (COP) mutations. Extensive genetic and molecular analysis showed this to be the case – COP mutations relieved a global repression of photosynthesis-related gene expression, being affected in genes whose products normally “relieve” this repression in response to light.
So why mention these mutations on an ID board? First, to point out that there are aspects of living things that, IMO, an ID-based approach would miss. The ID paradigm, the metaphors that abound in ID literature, are ones of building up, creation, and the like. In this view, photosynthesis-related development should be a matter of activation of hundreds, even thousands of specific genes dedicated to this developmental program. It seems unreasonable, fantastic, even, to suppose that a single recessive mutation could turn on, or recruit, or create, all of these thousands of genes in one fell swoop, and in ways that recapitulate many or most of the processes that a normal plant undergoes when germinated in the light. (One might invoke mutational “activation” of global activators to explain the light-independent development of these mutants, but such mutations would be dominant; COP and DET mutations are recessive, and thus not consistent with the “building” or “creating” metaphor.)
OK, maybe some here would buy this (but not likely, at least on its own), but why stuff this essay in a thread entitled “In Praise of Garbage Disposals”? Aye, there’s the rub ![[Smile]](smile.gif) . For ‘tis the nature of the COP genes that reveals a really interesting story. Once again abridging a long, detailed, and most interesting series of studies, it turns out that the collection of COP genes define the components of a complex that was first termed the COP9 complex (after one of the genes identified by the mutational analysis). As it became possible to track down, clone, and sequence the various COP genes, it became apparent that their protein products were related to the subunits of another subcellular complex. This complex was known for years before as ………….
(drum roll please)
………… the so-called “lid” of the proteasome, the large protein-degrading machine that was mentioned in the OP of this thread. The lid of the proteasome is the complex that serves as the gate, through which ubiquitinated proteins pass on their way to ultimate demise. The lid is largely (but not entirely – we never say always or never in biology) what regulates or controls protein turnover, by distinguishing ubiquitinated (or, more accurately, polyubiquitinated) from unmodified proteins. This similarity lent itself nicely to a model that holds that the COP complex controls gene expression by controlling the degradation of factors that promote the photosynthetic genetic program. This model has been largely confirmed by numerous experiments. In short, the COP complex is (at least a specialized lid for) a garbage disposal for activators of gene expression. In dark-grown plants, the COP complex breaks down (or perhaps sequesters or in other ways inactivates) activators; in the light, the COP complex is switched off, thus activating the photosynthetic program. (Hopefully, one can see why COP mutations are recessive.)
The COP connection (the COP9 complex is now known as the COP9 signalsome, or CSN, and it exists not just in plants, but in all eukaryotic cells) turns out to be one of many variations of mechanisms that cells use to control gene expression by protein turnover. It seems as if every day brings a new report of the involvement of a ubiquitin ligase (the enzyme that attaches ubiquitin to proteins, thereby initiating the degradation of the modified polypeptides) in the control of gene expression. Protein turnover, degradation, the garbage disposal, is increasingly being recognized as a central, fundamental, essential mechanism for regulating the expression of genes.
So, what does all of this mean? Basically, cells are constantly making and throwing away perfectly good transcriptional activators. This is, IMO, a strategy that is contrary to sound engineering principles, “intelligent design”, or any other machine-based metaphor one might choose to apply to living things.
I’ll close this essay with two evolution-inspired thoughts. One relates to the evolutionary origins of genetic programs such as those that exist in different plant tissues. The reflexive, intuitive model one would probably adopt for, say, the evolutionary differentiation of plants from simpler ancestors would be to invoke the origination of either a distinct photosynthetic program in a non-photosynthetic ancestor, or a program that was responsible for what came to be roots in the ancient photosynthetic ancestors of plants. The realization of the central role that protein degradation plays in development lends itself to a different model. Rather than the creation or building up of a completely new genetic program, the evolutionary differentiation of green and non-green tissues more likely involved the turning off, in a subset of cells and tissues, of the photosynthetic program that existed in the unicellular ancestor of plants. This same theme, turning off genes, likely played (and plays) an important role in the evolution of animals as well.
Finally, a last word about the CSN and its relative, the lid of the proteasome. It turns out that both of these complexes have another relative inside the cell – the so-called eukaryotic translation initiation factor 3 (or eIF3). All three of these complexes possess a subset of very similar subunits. This leaves the close of this essay with a curious, even paradoxical, thought – protein synthesis and breakdown require intriguingly similar complexes. (We’ll leave the evolutionary implications of the eIF3-CSN-lid similarities for another essay.)
A few random links for your further reading pleasure:
This one has some interesting summary about the CSN in different eukaryotes
Dr. Deng’s lab at Yale, and an interesting table showing the players in Arabidopsis and humans.
(I may add more links in random edits in the future, so keep checking back.)
[made a few minor grammatical corrections - 8-24]
[ 24. August 2005, 23:15: Message edited by: Art ]
IP: Logged
|
|
|
Printer-friendly view of this topic