|
Author
|
Topic: The Utility of IC
|
Mike Gene
Member
Member # 149
|
posted 07. March 2002 22:52
We have two scenarios to choose from:
Scenario A. Originally, FliM was a full length protein about 300 amino acids in length and in the lineage that led to modern day Aquifex, a deletion occurred cutting out the N-terminal 130 amino acids. However, because of the redundancy, FliN was able to compensate for this loss.
Scenario B. Originally, FliM was about the size as it is in modern day Aquifex, about 180 amino acids. Then, after Aquifex split off the eubacterial tree, and prior to the last common ancestor of the rest of the bacterial lineages, FliM was expanded to include the N-terminal 130 amino acids.
Alex notes the two scenarios are equally parsimonius and there is thus no reason to prefer one over the other. This is true from a purely scientific perspective, but remember this is a forum where we can brainstorm with teleological thinking; where one can employ novel intuitions, speculations, hypotheses, conjectures as part of putting some preliminary teleological thoughts on the table. Thus, while we may not be able to go much further in science, we might benefit from tapping into ID thinking.
As such, if I am to propose that the original flagella were designed, this does provide some illumination. We can reasonably expect the flagellum's designers to better designers than we are. Thus, if our best designers (engineers) arrive at insights about the design process, we might expect the flagellum's designers to have likewise been aware of such insights. In this case, the concept of robustness is increasingly recognized as a feature of good design. The ability to design something such that key features are preserved despite the uncertainties entailed in deploying the design would seem rather important. And front-loading evolution depends on this ability. Now, it would also seem that the assembly process would be a key feature of the flagellum. In this light, a flagellum built with FliN and full-length FliM would be more robust than a flagellum with truncated FliM. One could lose FliN, or the N-terminus of FliM, and still have a functioning flagellum. From an engineering perspective, one might prefer Scenario A.
Furthermore, I toy around with the idea of front-loading evolution. This view proposes that the original life forms were designed to channel evolution. It may mean, although, not necessarily, that some original states were actually more complex than what we see today. There are reasons for this that I can explain another time. But this provides an impetus to prefer a loss explanation as the default explanation. Added to this is the intuition (exploited by many teleologists) that it is easier to lose information than acquire information. That is, if it is more likely that FliM would lose unselected sequence than acquire new selectable sequence, the balance should be tipped towards the former.
But might there be more?
Scenario A would work simply like this. The ancestors of Aquifex entered an environment where chemotaxis was superfluous (that they use inorganic carbon sources for biosynthesis and an inorganic chemical energy source may be relevant). If the Che-chemotaxis system is IC, loss of one gene would cause the whole system to disappear. B. subtilis, Thermotoga, spirochetes, and E. coli all have CheA,W,B and Z (the latter protein interacts with FliM). Aquifex has none. Once CheY is gone, there is no longer any selective pressure to maintain the N-terminus of FliM. A deletion occurs by chance, but because FliN can compensate, it is a neutral mutation. However, there is now strong selective pressure to maintain FliN.
In scenario B, we begin with the intact chemotaxis system. In this scenario, a 390-bp sequence encoding 130 amino acids is fused to the N-terminus of the truncated FliM. In a rather fortuitous manner, it not only does not interfere with the already-existing functions of FliM, but it also able to interact with CheY such that a novel function emerges - switching.
It would seem to me that scenario A is rather straightforward, while scenario B entails ad hoc assumptions. First, we should note that FliM possesses three functions - it participates as part of the rotor, it participates in assembly, and it participates as the switch. If we begin with a truncated version that carries out the first two functions, but not the third, there doesn't seem to be any reason for thinking this protein would be necessarily preadapted to switch. Admittedly, our understanding of the switching mechanism is very vague. But it would seem that glomming something on to the N-terminus of this protein had a significantly likelihood of gunking up the motor or gunking up the assembly process, unless FliM was front-loaded to accept such a modification.
Also, assembly is a core function and a chunk of this N-terminal sequence is involved in these early and crucial flagellar events. It doesn't make much sense to mess with the core of an assembly process that already works. Natural selection tends to add-on, not rewire.
Of course, this is all very fuzzy, so perhaps a different angle might help.
Since scenario B entails bringing new information to the system, we must account for it. Where did it come from? I BLASTed with the N-terminal 130 amino acid sequence from E coli, Bacillus, Trepenoma, and Thermatoga and turned up only FliM homologs. Thus, this N-terminal sequence is FliM-specific and flagellum-dependent. This renders any cooption scenario ad hoc, forcing us to propose that 1) The sequence was originally found somewhere else doing something else; 2) was incorporated into FliM; and 3) subsequently lost a) prior to the last common ancestor of the rest of the bacteria tree or b) independently in all of these very different lineages. One cannot rule cooption out as a possibility, but it is less parsimonious than the straightforward loss hypothesis.
However, perhaps the N-terminal sequence comes from someplace inside the already existing truncated FliM sequence. Here I used ClustalX again. I aligned the N-terminal 130 amino acids from E coli, Bacillus, Trepenoma, and Thermatoga. ClustalX scores invariant positions, positions that are highly conserved with amino acids of very similar chemical properties ("strong") and positions that are highly conserved with amino acids of similar chemical properties. The results were 23 invariant positions, 32 strong positions, and 17 weak positions (I'll score this as 23-32-17). Next, I attempted to align the N-terminal 130 amino acids of FliM from these four species with the C-terminal 200 amino acids from the same gene/species. The alignment was rather poor, with a score of 1-6-7. I then decided to flip the Thermotoga N130 sequence (reflecting an inversion) and align to the C200 sequences to see if the score was better. It was not: 1-6-6. Thus, there doesn't seem to be good support for the expansion of this gene.
Finally, what about an appeal to random sequences being incorporated into FliM? Let me borrow an approach from Art (on ARN), who in turn was borrowing from Yockey:
quote:
If we make the assumption (that is BAD, but necessary since we can only work with what nature gives us) that only those sequences that we find in nature can actually satisfy these functions, we can try and estimate the "information" needed to accomplish these functions using the formula that Yockey gave us thirty years or so ago.....Briefly, I aligned a bunch of sequences [and] then estimated two extremes for the possible information content of these proteins. One extreme was calculated by counting only the number of different amino acids seen at conserved positions (as determined by ClustalW; conserved means similar residues in a majority of the sequences aligned) and using this information to derive a multiplier for the position which is just the number divided by 20 (the total number of amino acids). Unconserved positions and positions at which one or more sequence had a deletion were not counted here (or, more accurately, these positions were assigned a multiplier of 1). The end result is an estimate of the fraction of sequences that would resemble the particular protein......The other extreme was to calculate only the information inherent at absolutely invariant positions. This gives us an absolute lower bound on the information of the protein (or, more accurately, the function), again assuming that one and only one sequence family can satisfy the function. [1]
If we apply this approach to the N-terminal 130 amino acids, we get a lower bound (considering only invariant positions) of 1.2 x 10^-30 and an upper bound of 2.1 X 10^-75. These numbers are small enough to tentatively exclude random sequences as the source of this FliM sequence.
The bottom here is these considerations indicate an attempt to explain the origin of this sequence is faced with many difficulties not faced by the loss hypothesis.
Finally, there is one other consideration. Thermotoga is also a very deep branching bacterium. In fact, one could argue that it branched off the tree shortly after Aquifex, but before the rest of the lineages. And what is most significant is that Thermotoga lacks FliN. Now, if we stick to scenario A, we simply see independent examples of loss as a consequence of the original robustness of the system. But if we revert back to scenario B, now we have to argue that the N-terminal addition to FliM occurred prior to Thermotoga splitting off the main bacterial tree, as IC indicates that FliN could not be lost without the complete FliM gene being present.
In the end, I think the original ID considerations I raised work well enough for a teleologist to prefer scenario A as a tentative hypothesis. The above offers some (albeit weak) support. However, I would note an asymmetry in the evidential status. There is not much room for scenario A to be better supported. The only further solid thing we could have is the FliM N-terminal sequence already present in Aquifex and then there would be no dispute. On the other hand, there are many more ways scenario B could have been supported, yet don't exist. For example, the N-terminal addition to FliM could have been found exclusively in some bacterial crown group (i.e., one of the divisions of proteobacteria). Or, the FliM N-terminal sequence could have been poorly conserved, rendering independent acquisitions or acquisition from random sources more likely. Or the FliM N-terminal sequence could have been specifically involved in switching and nothing else (unlike the observable situation where it is involved in assembly), making it easier to envision it as an add-on.
1. This quote from Art comes from an exchange we had almost two years ago on the ARN forum. Art used this approach to argue the information content of two chaperone proteins was not high.
[ 07 March 2002, 23:33: Message edited by: Mike Gene ]
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 07. March 2002 23:23
As a brief aside.....
Janitor: Mike Gene may recognize my “voice,” since I’ve repeatedly begged the IDers to stop thinking like theologians and start thinking like engineers.
That's what I am trying to do, although I am hindered by the simple fact that I am not an engineer.
I wonder if he shares my impression that the "default," analytical position of biologists is not “Darwinism” or “naturalism,” whatever those are, but a sort of “design-teleological” instrumentalism?
Yes. Biology is permeated by the form/function relationship. If you want to teach kids about homeostasis, you don't point to anything they learned in a geology, physics, or chemistry class. You point to the thermostat/furnace system in their house. If a physicist claimed machines fueled the sun, or a geologist claimed machines form mountains, or a meteorologist claimed machines formed rain, or a chemist claimed machines formed crystals, their colleagues would think them wacky. But leading biologists claim machines form cells.
In fact, you might want to have some fun with the recent article in Science (3-1-02) entitled Reverse Engineering of Biological Complexity by Csete and Doyle. Here's just one of the very interesting insights they provide:
quote:
Experiments, modeling and simulation, and theory all have fragilities, but they are complementary, and through the right protocols they have the potential to create a robust “closedloop” systems biology (59). Biologists’ frustrating experience with theory has been primarily in an open-loop mode, where simple and attractive ideas can be wrong but receive enormous attention. Biology is the only science where feedback control and protocols play a dominant role, so it should not be surprising that there would be popular theories, coming from within science, that did not emphasize these issues. Biologists and engineers now have enough examples of complex systems that they can close the loop and eliminate specious theories (60). We should compare notes.
IP: Logged
|
|
nobody
Member
Member # 145
|
posted 09. March 2002 05:31
Mike,
Did you notice this over at ARN a couple months ago? It seems like it might relate, somewhat, to this discussion.
http://www.arn.org/ubb/Forum1/HTML/001662.html
--------------------------------------------------------------------------------
A paper published today on the web (web link to Proc. Natl. Acad. Sci. USA below) says that data suggest that a new type of cell, the Chronocyte, existed before the Eukarya, but was not in the Archaeal or Bacterial domains. Yet the Chronocyte clearly was, as the authors state, more complex than the other organisms (the word “complex” is used in the abstract, below, twice to describe this organism).
This raises the question, where did the Chronocyte come from, since it is not a member of either of the Archaea or the Bacteria and came before the Eucarya. This seems inconsistent with Darwinian evolution, but is consistent with design.
quote:
--------------------------------------------------------------------------------
The origin of the eukaryotic cell: A genomic investigation
Hyman Hartman, and Alexei Fedorov§
Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139; and § Department of Molecular and Cellular Biology, Harvard University, Cambridge, MA 02138
Communicated by Carl R. Woese, University of Illinois at Urbana-Champaign, Urbana, IL, December 10, 2001 (received for review October 25, 2001)
We have collected a set of 347 proteins that are found in eukaryotic cells but have no significant homology to proteins in Archaea and Bacteria. We call these proteins eukaryotic signature proteins (ESPs). The dominant hypothesis for the formation of the eukaryotic cell is that it is a fusion of an archaeon with a bacterium. If this hypothesis is accepted then the three cellular domains, Eukarya, Archaea, and Bacteria, would collapse into two cellular domains. We have used the existence of this set of ESPs to test this hypothesis. The evidence of the ESPs implicates a third cell (chronocyte) in the formation of the eukaryotic cell. The chronocyte had a cytoskeleton that enabled it to engulf prokaryotic cells and a complex internal membrane system where lipids and proteins were synthesized. It also had a complex internal signaling system involving calcium ions, calmodulin, inositol phosphates, ubiquitin, cyclin, and GTP-binding proteins. The nucleus was formed when a number of archaea and bacteria were engulfed by a chronocyte. This formation of the nucleus would restore the three cellular domains as the Chronocyte was not a cell that belonged to the Archaea or to the Bacteria.
To whom reprint requests should be addressed. E-mail: hhartman@mit.edu.
--------------------------------------------------------------------------------
www.pnas.org/cgi/doi/pnas.032658599 or www.pnas.org/cgi/content/abstract/032658599v1
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 10. March 2002 23:31
Hi Nobody,
Yes, I did see this mentioned before. However, I have not read the paper so it's best not to comment until then. The original suggestion was that this might fit into my thesis that the original cells may actually have been more complex than what we see. I don't think the paper will support that hypothesis (but neither would it contradict it). Does it fit in this thread? Actually, in an oblique way, yes. Perhaps I make the connection tomorrow night.
IP: Logged
|
|
rafe gutman
Member
Member # 134
|
posted 12. March 2002 22:54
hi mike, i still haven't bothered to look up any info on the flagella yet, so my knowledge of it is based entirely on this thread, but why couldn't the N-terminal portion of FliM come from FliN? if FliN is present in aquifex, but not in thermatoga, but thermatoga has the longer FliM, then could a possible scenario be this?
aquifex has ancestral flagella with shortened FliM, but possessing FliN. a genetic event places a portion of FliN onto FliM, extending it by ~180 aa. because of this redundancy, thermatoga is able to sustain a deletion of FliN.
if thermatoga branched later than aquifex, but before any of the other groups, then a possibility exists that what was first just a redundancy, later became essential.
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 14. March 2002 17:46
Good catch, Rafe. Yes, it is possible that the Aquifex state reflects the ancestral state. After Aquifex branched off the tree, FliN was duplicated and fused to the N-terminus of FliM. Then, Thermotoga lost FliN. Thus, according to my hypothesis, we have two independent examples of loss. With this hypothesis, we have an example of duplication and fusion, followed by loss.
But let's not lose sight of the big picture I am painting. It is often complained that ID doesn't come up with anything different than traditional science (apart from nitpicking traditional science). Yet as I explain above, where traditional science leaves us in this ambiguous state, a teleological approach allows me to take a step further and actually find reason to prefer one over the other.
My focus in this thread was on the utility of IC. I hope its clear to some that a research program, with IC-in-hand, can lead to interesting hypotheses. Perhaps a similar problem, with a different system, using a similar approach, may in fact open itself up to a better resolution.
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 20. March 2002 01:01
Another Example Where IC is Useful.
Let's move away from flagella and turn our attention to Ubiquitin (Ub), a small protein composed of 76 amino acids. Thus far, it has been universally found in eukaryotic organisms. It serves as a type of "tag" that gets attached to other proteins to label them for degradation (in an ATP-dependent fashion). Ub is also a very important protein, as it is employed in a wide range of basic biological processes, including cell division, the regulation of transcription, DNA repair, and embryogenesis. It's importance is reflected in its extreme sequence conservation.
When we compare the amino acid sequence of Ub from humans with yeast, we find only three differences, at positions 19, 24, and 28. They are as follows. Human19 - Proline, Yeast19 - Serine; Human24 - Aspartate, Yeast24 - Glutamate; Human28 - Alanine; Yeast28 - Serine.
I also checked out some other organisms. Tomatoes differ from humans in 3 positions also: positions 19, 24, and 56. In positions 19 and 24, the amino acids are as in yeast. At position 56, humans have serine
and tomatoes have alanine.
Fruit flies have an identical sequence to humans.
I also checked out some single-celled parastitic protists. Leishmania differ from humans at two positions: 14 and 51. At 14, humans have threonine and leishmania has alanine. At 51, humans have aspartate while leishmania has glutamate. In Eimeria (coccidia), there is only one difference between humans at position 17. Humans have glutamate and these critters have aspartate.
Thus, positions 14, 17, 19, 24, 28, 51, and 56 can vary. But all these amino acid substitutions are quite conservative (ala/thr/ser; glu/asp).
These data strongly indicate that Ub today is essentially the same as it was in the last common ancestor of all eukaryotes (which usually dates 1.5-2 billion years ago) and natural selection has maintained the sequence since this time. Anyone who would claim the Ub sequence was different in the past is making a unsupported claim.
So where does IC fit into the picture. As I mentioned, Ub functions mainly as a tag. What actually degrades the tagged-protein is a protein complex known as the proteasome. The business-end of the proteasome is the 20S chamber - a hollow chamber composed of four rings of proteins. Proteins-to-be-degraded are unfolded and threaded into this chamber where proteolysis occurs.
Yet there are two sets of intermediaries between a protein-to-be-degraded and a protein fed into the proteasome. First, the protein that is to be degraded is first labeled with Ub by a set of proteins that fall into three classes (E1, E2, and E3). Secondly, proteasomes have what is called an 19S cap attached to both ends of the degradative chamber. The 19S cap itself is composed to several different proteins and functions to capture, unfold, and thread the ubiquitinated protein.
Now, I have not look at the ubiquitin pathway in detail, but from a distant perspective, it has the echoes of IC. That is, it involves four themes - (1) Ub; (2) the E1/E2/E3 enzymes that attach Ub to proteins; (3) the 19S cap that captures, unfolds, and threads the protein; and (4) the 20S chamber that actually degrades the protein.
So where does IC utility come into play?
It starts when you BLAST with Ub sequence. If you look only among eukarya, you of course get all kinds of hits (as expected). If you search among eubacterial sequence, you get no significant hits. That is, bacterial have no Ub homologs. If you search among archaebacteria, you get no significant hits. But something interesting does come up. It appears as if someone has sequenced a Ub-fragment (18 amino acids) from Thermoplasma acidophilum:
AUTHORS Wolf,S., Lottspeich,F. and Baumeister,W.
TITLE Ubiquitin found in the archaebacterium Thermoplasma acidophilum
JOURNAL FEBS Lett. 326 (1-3), 42-44 (1993)
From a purely evolutionary perspective, this would make sense. Archaea and eukarya are often said to be more closely related (depending on which system is considered). And archaea have a very similar proteasome.
Yet from an IC perspective, this does not make sense. The archaeal proteasome does not have a 19S cap. And no one has reported the existence of E1/E2/E3 enzymes from archaea. Thus, there is nothing to link the Ub to the proteasome.
Is the concept of IC strong enough to call the evolution view into question? Or does the evolutionary view swamp out IC considerations and provides another clue to the evolution of IC?
Luckily, the genome of Thermoplasma has been sequenced. So its time to find the whole Ub gene by BLASTing the Thermoplasma genome with Ub sequence. And nothing significant is there. When you BLAST with the fragment sequence, the closest thing is a "conserved hypothetical protein" where 5/18 positions are identical. Bottom line - no Ub.
If you sniff around some more, you'll find confirmation on talk.origins (of all places), where someone who originally cited the Wolf study posted:
quote:
Well, I did, and now I must retract what I wrote above. By chance, one of the world's authorities on ubiquitin and protein degradation, Alexander Varshavsky, works at Caltech, and was very helpful to me. So it turns out that the above paper was wrong, and their sequence data was incorrect. Subsequent sequencing of archaebacterial genomes has *not* turned up any ubiquitin genes.
Thus, it turns out the IC perspective was powerful enough to turn back the evolution-friendly finding. Where traditional evolutionary thinking would cause us to embrace the originally sequence "unbiquitin" fragment, the IC perspective would provide cause for skepticism that would have won out in the end.
IP: Logged
|
|
Drosera
Member
Member # 139
|
posted 20. March 2002 03:01
Mike Gene writes,
quote:
Thus, it turns out the IC perspective was powerful enough to turn back the evolution-friendly finding. Where traditional evolutionary thinking would cause us to embrace the originally sequence "unbiquitin" fragment, the IC perspective would provide cause for skepticism that would have won out in the end.
Hmm, seems like the evidence was never very strong to begin with. 18 amino acids is exactly the kind of scale where chance similarity can play a plausible role.
Regardless, I'm very confused about what you think 'irreducible complexity' (IC) and the "IC perspective" is 'predicting' or suggesting or whatever. Isn't IC just a definition, like basically:
- an IC system contains multiple interacting parts that are all required for a given function
...and this, not surprisingly, allows one to predict that, in fact, all of the parts are indeed required. Of course, that's all by definition in the first place, so I don't see what the big deal is when it turns out that the parts are in fact required. Rigorously speaking you can't really conclude IC until you know this already anyway, I would think. And if it turns out that all of the parts aren't required, then it's not a failed prediction, it's just a misapplied definition.
An additional difficulty is that there appears to be no real standards for what constitutes a 'function' in the IC sense (how specific can one be?...if one really splits hairs then probably any protein can be considered "essential" for some very-specifically defined function, e.g. "increasing reaction X 5%" or something like that).
The reason I bring this up is that typing 'Archaebacteria ubiquitin' into Pubmed brought up two articles. The first is the Wolf et al 1993 article Mike referenced. But the second is this, from 1999 in the Journal of Biological Chemistry:
An Archaebacterial ATPase, Homologous to ATPases in the Eukaryotic 26 S Proteasome, Activates Protein Breakdown by 20 S Proteasomes.
The abstract is here.
Here is Mike Gene's (admittedly preliminary) discussion of the degradation pathway:
quote:
So where does IC fit into the picture. As I mentioned, Ub functions mainly as a tag. What actually degrades the tagged-protein is a protein complex known as the proteasome. The business-end of the proteasome is the 20S chamber - a hollow chamber composed of four rings of proteins. Proteins-to-be-degraded are unfolded and threaded into this chamber where proteolysis occurs.
Yet there are two sets of intermediaries between a protein-to-be-degraded and a protein fed into the proteasome. First, the protein that is to be degraded is first labeled with Ub by a set of proteins that fall into three classes (E1, E2, and E3). Secondly, proteasomes have what is called an 19S cap attached to both ends of the degradative chamber. The 19S cap itself is composed to several different proteins and functions to capture, unfold, and thread the ubiquitinated protein.
Now, I have not look at the ubiquitin pathway in detail, but from a distant perspective, it has the echoes of IC. That is, it involves four themes - (1) Ub; (2) the E1/E2/E3 enzymes that attach Ub to proteins; (3) the 19S cap that captures, unfolds, and threads the protein; and (4) the 20S chamber that actually degrades the protein.
......
From a purely evolutionary perspective, this [the proposed ubiquitin homology] would make sense. Archaea and eukarya are often said to be more closely related (depending on which system is considered). And archaea have a very similar proteasome.
Yet from an IC perspective, this does not make sense. The archaeal proteasome does not have a 19S cap. And no one has reported the existence of E1/E2/E3 enzymes from archaea. Thus, there is nothing to link the Ub to the proteasome.
Is the concept of IC strong enough to call the evolution view into question? Or does the evolutionary view swamp out IC considerations and provides another clue to the evolution of IC?
Here are some very interesting quotes from the JBC article:
(...sorry about the longish quotes but I think they're directly relevant...)
(bolds added)
quote:
Abstract: In eukaryotes, the 20 S proteasome is the proteolytic core of the 26 S proteasome, which degrades ubiquitinated proteins in an ATP-dependent process. Archaebacteria lack ubiquitin and 26 S proteasomes but do contain 20 S proteasomes. Many archaebacteria, such as Methanococcus jannaschii, also contain a gene (S4) that is highly homologous to the six ATPases in the 19 S (PA700) component of the eukaryotic 26 S proteasome. To test if this putative ATPase may regulate proteasome function, we expressed it in Escherichia coli and purified the 50-kDa product as a 650-kDa complex with ATPase activity. When mixed with the well characterized 20 S proteasomes from Thermoplasma acidophilum and ATP, this complex stimulated degradation of several unfolded proteins 8-25-fold. It also stimulated proteolysis by 20 S proteasomes from another archaebacterium and mammals. This effect required ATP hydrolysis since ADP and the nonhydrolyzable analog, 5'-adenylyl ,-imidophosphate, were ineffective. CTP and to a lesser extent GTP and UTP were also hydrolyzed and also stimulated proteolysis. We therefore named this complex PAN for proteasome-activating nucleotidase. However, PAN did not promote the degradation of small peptides, which, unlike proteins, should readily diffuse into the proteasome. This ATPase complex appears to have been the evolutionary precursor of the eukaryotic 19 S complex, before the coupling of proteasome function to ubiquitination.
Introduction: The 26 S proteasome, which is the major site of protein breakdown in mammalian cells, is composed of the 20 S proteasome (molecular mass of 700 kDa) and two 19 S regulatory complexes (700 kDa) (1, 2). The 20 S proteasome is a cylindrical proteolytic complex composed of four stacked, seven-membered rings (3). In the presence of ATP, the 19 S complex (also called PA700) becomes associated with each end of the 20 S cylinder (4, 5). The resulting 26 S particle degrades ubiquitinated and certain non-ubiquitinated proteins in an ATP-dependent process (1). Six of the approximately 18 subunits of the 19 S complex are ATPases whose precise functions in protein degradation and in 26 S assembly remain unclear (4, 6-9).
Within the 19 S regulatory complex, these ATPases form a subcomplex ("base"), which binds directly to the 20 S core particle (10). These ATPases are all members of the large AAA family, which contains more than 100 ATPases that are involved in diverse cellular processes, including protein degradation, cell division, peroxisome biogenesis, vesicle transport, and meiosis (11, 12). Only eukaryotic cells contain ubiquitin or 26 S proteasome complexes. The initial report of the finding of ubiquitin in the archaebacterium Thermoplasma acidophilum (13) and the cyanobacterium Anabaena variabilis (14) has not been confirmed, and the sequencing of several prokaryotic genomes has not revealed genes for ubiquitin or homologs of ubiquitin-conjugating enzymes (15). However, 20 S proteasomes are present in archaebacteria (16-19) and in actinomycetes (20-22). The 20 S proteasome from archaebacteria, although containing only one type of -subunit in the outer rings and one type of -subunit in the central two rings, is quite similar in architecture to the eukaryotic proteasome and is clearly the evolutionary ancestor of the eukaryotic particle (2), which contains seven distinct but homologous -subunits and seven distinct but homologous -subunits.
The present studies were undertaken to investigate whether the archaebacterial 20 S proteasome, like the eukaryotic particle, might also function in an ATP-dependent manner in association with a regulatory ATPase complex. In bacteria, such as Escherichia coli, which also lack ubiquitin, most intracellular protein degradation requires ATP and is catalyzed by large ATP-hydrolyzing proteolytic complexes (23, 24). Several of these enzymes (ClpAP, ClpXP, and HslVU) are composed of central proteolytic particles (ClpP and HslV) whose function in protein degradation requires ATP hydrolysis by an associated ring-shaped ATPase complex (ClpA, ClpX, and HslU) (24). Prior attempts to find a larger form of the archaebacterial proteasome that functions in an ATP-dependent fashion have not been successful (25). However, the sequencing of the genome of Methanococcus jannaschii and other archaebacteria have revealed a gene (S4) (26), whose predicted protein sequence is similar to that of the eukaryotic 26 S ATPases.
[From the end of the conclusion] PAN shows extensive homologies to the 26 S proteasome and is therefore the most likely evolutionary precursor to the 19 S complex and, in particular, to its base, the portion which contains its six ATPases that associates with the 20 S proteasome (10). The one notable potential difference between these ATPase complexes is that PAN does not enhance peptide hydrolysis by its 20 S particle whereas the 19 S (PA700) complex does so, perhaps because this effect may involve a non-ATPase subunit. Recent findings have suggested that the other proteins of the 19 S complex that comprise its "lid" are homologous to the COP9-signalosome-like complex, which functions in signal transduction and translation (10, 63, 64). The combination of these components with PAN must have been the critical step during the evolution of eukaryotic cells that allowed the coupling of proteasome function to ubiquitin conjugation and the establishment of the ubiquitin-proteasome pathway.
I'm no expert on this topic, but I gathered from the above-quoted material that:
1) At least some proteins in fact can be degraded by the eukaryotic proteasome despite being non-ubiquitinated. So is ubiquitin really required? What is really the IC system here?
2) Archaebacteria, and eubacteria I think, still have regulated, ATP-dependent proteasome function despite lacking ubiquitin. They appear to have some but not all of the components of the eukaryote system. How is this possible from the IC point of view? In fact, in at least Methanococcus jannaschii the system appears to be rather intermediate between eukaryotes and eubacteria, which is in accord with a large amount of other close similarities between eukaryotes and archaebacteria & an indication of their close relationship.
3) The statement "The archaeal proteasome does not have a 19S cap" appears to be contradicted by the article, and this finding appears to be very unexpected using the "IC perspective", in which the potentially IC system was described thusly:
quote:
Yet there are two sets of intermediaries between a protein-to-be-degraded and a protein fed into the proteasome. First, the protein that is to be degraded is first labeled with Ub by a set of proteins that fall into three classes (E1, E2, and E3). Secondly, proteasomes have what is called an 19S cap attached to both ends of the degradative chamber. The 19S cap itself is composed to several different proteins and functions to capture, unfold, and thread the ubiquitinated protein.
...but we apparently have the situation in this article where the ubiquitin is missing but the 19S cap is present and fully functional! The "IC perspective", insofar as the somewhat fuzzy idea predicts/suggests what to expect at all, appears to have mislead us rather badly here.
In fact, it seems to me that in the ubiquitin pathway case, scientists are rapidly approaching having a reasonably thorough explanation for how the evolution of this complex, interacting, lotsa-parts-required for at least most function in eukaryotes, actually occurred. Is this not a rather strong indication that irreducible complexity ain't all it was advertised to be?
Drosera
IP: Logged
|
|
James A. Barham
Member
Member # 50
|
posted 20. March 2002 08:04
To Mike and Janitor:
We all three seem to agree that goal-directed functionality is objectively real (not just a "projection" of the human mind or other form of illusion), and that it is absolutely central to biology (which scarcely anyone denies).
Given this agreement, I have a question about terminology. Isn't the word "design" problematic, in that it seems to assume in advance that organisms are like manmade machines, and that the functional organization has been imposed on them from the outside. Shouldn't we leave open the possibility that the functional organization of living things arises somehow sponatneously from within? If you agree that this is a theoretical possibility, then shouldn't we employ a more neutral term?
My preferred term is "immanent teleology," but I know that word is a red flag for most scientists. The trouble is that the other most natural term, "function," does not really place enough stress on the goal-directed aspect of biological organization. The Darwinians are all comfortable talking about "functions," since they imagine (foolishly, in my opinion) that they have a way of "reducing" functions to mechanisms.
At any rate, I think that anyone who wants to leave open the possibility of explaining biological functionality naturalistically ought to eschew the term "design."
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 20. March 2002 22:29
Drosera says he is confused about the point of my last posting. The whole point revolved around how we interpret the claim that ubiquitin was isolated from archaea. Since the function of Ub is apparently tied up with irreducible complexity as described above (I qualify it as apparent for lack of looking deeply into this system), the report in question is suspect. That is, to couple Ub to the proteasome, two intermediaries are required, yet they are lacking in archaea.
Drosera claims the evidence was never very strong to begin with. Indeed. But let us not overlook that the evidence was considered strong enough to find its way into the peer-reviewed literature (a very respectable journal, no less), complete with the title, "Ubiquitin found in the archaebacterium Thermoplasma acidophilum." This speaks to the issue of the peer reviewed literature and the things the scientific community is looking for. As I mentioned above, since the finding of Ub in archaea fit so nicely within the traditional evolutionary paradigm, apparently there was no cause for skepticism among the reviewers to reject this paper on the basis that "18 amino acids is exactly the kind of scale where chance similarity can play a plausible role."
Drosera then draws attention to another paper and asks three questions:
At least some proteins in fact can be degraded by the eukaryotic proteasome despite being non-ubiquitinated. So is ubiquitin really required? What is really the IC system here?
Keep in mind that I originally qualified my claims by noting that I have not look at the ubiquitin pathway in detail, but from a distant perspective, it has the echoes of IC. With this in mind, the ability to degrade proteins without ubiquitin is not relevant. The IC system of interest would come into play when Ub is added to the mix. Why? As mentioned above, Ub is essentially a "tag." A tag by itself has no function. Instead, it must be both applied and detected. Thus, the IC system in question is Ub-dependent protein degradation. To carry out the function, we need four ingredients: the tag (Ub); the ability to add the tag (E1/E2/E3); the ability to detect and employ the tag (19S cap); protein degradation (the proteasome).
Here, we can separate generic protein degradation from Ub-dependent protein degradation. The former does not require Ub. The latter does. Yet, an IC interaction (a subsystem, if you will), appears to be required to convert the former into the latter, namely, the first three ingredients listed above (which could very well expand into a rather large IC system if we begin scoring gene products/proteins).
Archaebacteria, and eubacteria I think, still have regulated, ATP-dependent proteasome function despite lacking ubiquitin. They appear to have some but not all of the components of the eukaryote system. How is this possible from the IC point of view? In fact, in at least Methanococcus jannaschii the system appears to be rather intermediate between eukaryotes and eubacteria, which is in accord with a large amount of other close similarities between eukaryotes and archaebacteria & an indication of their close relationship.
The ATP-dependent proteasome does not require ubiquitin. The difference is that in tying Ub to the ATPase proteasome, we require Ub, a mechanism to conjugate it to other proteins, and a mechanism is interpret and exploit this conjugation event. This is where the utility of IC came into play in my above posting.
The statement "The archaeal proteasome does not have a 19S cap" appears to be contradicted by the article, and this finding appears to be very unexpected using the "IC perspective", ...but we apparently have the situation in this article where the ubiquitin is missing but the 19S cap is present and fully functional! The "IC perspective", insofar as the somewhat fuzzy idea predicts/suggests what to expect at all, appears to have mislead us rather badly here.
Not at all. The 19S cap is more than PAN. Recall that I originally defined the cap as something that captures, unfolds, and threads the protein into the degradative chamber. The cap is made up of about 18 proteins. PAN constitutes 6 of the 18, leaving 12 proteins to account for. These 12 proteins probably play the "capture" role that is Ub-dependent. PAN, on the other hand, probably functions as a generic unfoldase and translocase. Thus, contrary to Drosera's claim, the IC picture is becoming clearer, enabling us to better score the IC components involved in the Ub-tagging and recognition.
Drosera claims that scientists are rapidly coming up with a reasonably thorough explanation for the evolution of this complex. This point is off topic in this thread (a discussion of the utility of IC), but I thought I'd comment (further discussion can be carried out in a seperate, new thread).
I'm not quite sure what "complex" he has in mind. Arguing for a similarity between the archaeal and eukaryal proteasome is not an explanation for its origin. And everything about the Ub branch of the eukaryal system is elusive. In fact, it would seem to me the problem for non-teleological gradualists is very, very thorny here. The problem goes beyond accounting for the highly conserved Ub with no homologs in bacteria. And it goes beyond the lack of homologs for all the machinery used to tag, detect, and capture. It gets really thorny because Ub has its fingers in vital, core cellular processes among eukaryotes. As just one example, if you scroll up to the study highlighted by nobody, you'll notice that not only is Ub a eukaryotic specific protein, but so too are the cyclins. Cyclins play essential roles in maintaining the very cell cyle of eukaryotic cells (a system chock full of quality control and proofreading steps). And their function is tied up in their degradation in a Ub-dependent fashion. Integrating Ub into this system, in a meaningful way, would seem to be no small task. In other words, to explain the origin of Ub and all its associated machinery may also entail explaining the origin of other eukaryal-specific features, such as their cyclin-dependent cell cycle, vesicle transport, and perhaps even various classes of transcription events.
Finally, let me add one more comment in reply to the assertion that the "IC perspective", insofar as the somewhat fuzzy idea predicts/suggests what to expect at all, appears to have mislead us rather badly here.
Let me say up front that the IC perspective, and inferences built around IC, are not going to be infallible. But this is not relevant as a useful concept is not required to be infallible. We need only look to evolutionary theory itself where, for example, the notion that complex structures evolve from simpler structures, has misled us on several occasions.
IP: Logged
|
|
nobody
Member
Member # 145
|
posted 21. March 2002 00:01
quote:
We need only look to evolutionary theory itself where, for example, the notion that complex structures evolve from simpler structures, has misled us on several occasions.
Hi Mike,
Please pardon a stupid question, but would you mind posting a few examples of what you had in mind here?
Thank you.
P.S. I really like your newly posted "Rules".
![[Smile]](smile.gif)
IP: Logged
|
|
kyle7
Member
Member # 191
|
posted 21. March 2002 02:53
Having read through the posts, there is a good article in Technology Review showing how engineering analysis is being applied to cells.
the virtual cell
I think that biology and engineering will converge together forming a whole new field. This will further the ID debate as we begin to see the actual complexity of life.
Note: there are three pages to the link!
[ 21 March 2002, 02:59: Message edited by: kyle7 ]
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 21. March 2002 21:02
James,
I understand your concern about terminology. I often use the term "teleological" (as anyone from ARN knows), as I am quite open to the possibility that what I score as something designed may arise as you suggest. Yet my working hypothesis is that something like the flagellum was indeed bioengineered. It is this hypothesis which then drives my terminology here. However, I should also point out that I recognize that manmade machines and living organisms differ in some rather fundamental ways. But I also think these differences become less relevant when comparing manmade machines with the molecular machines. In other words, I think of life not as something that emerges from chemistry, but as something that emerges from nanotechnology.
Of course, this is off topic for this thread. Perhaps we could continue this in another thread (although my replies are likely to be delayed due to other obligations).
IP: Logged
|
|
Mike Gene
Member
Member # 149
|
posted 21. March 2002 21:18
Nobody,
Let me give you a great example from the 80s and early 90s. Microsporidia are unicellular eukaryotes. They are also the simplest and smallest eukaryote, existing as
intracellular parasites. After the endosymbiont hypothesis really took root (this hypothesis proposes that mitochondria were originally bacteria), scientists began to look for the descendents of those eukaryotes that existed prior to acquiring the mitochondria (a reasonable expectation). Microsporidia and several others seemed to fit perfectly. For example, microsporidia not only lack mitochondria (suggesting they branched off prior to other eukarya acquiring mitochondria), but they also lack golgi bodies, peroxisomes, and cilia. Their genomes are smaller than some bacteria. And it got even better. They have 70S ribosomes, the same size as bacteria. Then it got better yet. Sequence analysis of rRNA showed microsporidia to be the deepest branching of all eukarya, practically confirming it split off very early on. So impressive was this evidence that Cavalier-Smith invented a whole new kingdom for microsporidia and other unicellular organisms also lacking mitochondria. The kingdom was called Archezoa.
Yet, the kingdom was short-lived. The first cracks in the wall developed when mitochondria genes were found in the genome of microsporidia, suggesting that this organism once had mitochondria and lost them. Then,further sequence analysis of several protein-coding genes (tubulin, TATA-binding protein, and others) indicate that microsporidia are close relatives to fungi. Today, they are considered degenerate fungi and even Cavalier-Smith has quietly expanded the kingdom fungi to include microsporidia.
There are several lessons in all of this. It is entirely correct when someone notes there is
absolutely nothing about evolutionary theory that says that organisms will, or must, become complex. But it is also a fact (to the best that I can tell) that no Darwinist ever raised any skepticism about microspordia as a primitive Achezoan. Yet there was always a good reason to do so. As is also well known, many parasites exemplify a loss of complexity and microsporidia are intracellular parasites. So why did no one ever raise any skepticism? Because microsporidia seemed to fit so well into the “simple-to-complex” continuum thought to be required for the prokaryotic-to-eukaryotic transition.
How does this fit into this thread? It speaks to the issue of utility. This scientific blunder does not mean the "simple-to-complex" thesis that permeates evolutionary thinking is useless. As I mentioned above, usefulness does not mean infallibility. Thus, it does not concern me that the IC perspective may sometimes get it wrong.
BTW, I'm glad you liked my rules. As such, I should wrap up this thread in a few more postings.
IP: Logged
|
|
James A. Barham
Member
Member # 50
|
posted 22. March 2002 08:42
Mike:
I guess I would have to agree that evolution is a matter of "nanotechnology" in some sense, since I am certainly arguing that the teleological organization of the cell transcends biochemistry. That is to say, I view the living state of matter as an emergent level in which the cell as a whole is controlling the biochemical reactions occurring within it to keep itself in existence (which is what I mean by "resisting" the second law without "violating" it, as birds resist the law of gravity without violating it). But the question is, How is this cellular, nanotechnology possible? (Evolutionary change I regard as a secondary problem, which we cannot hope to understand until we understand better how the "nanotechnology" itself works; that is to say, our knowledge of diachronic change in cell functionality must eventually flow from an improved knowledge of synchronic cellular functioning.)
Can this nanotechnology arise spontaneously within the cell, with the functional "causality" (for lack of a better word) being exerted from the inside, or must we posit a supernatural force being applied from the outside to constrain the reactions, in the same way that human beings constrain intrinsically inert matter in our machines by establishing a set of boundary conditions (and intervening periodically to fix them, as they inevitably degrade)? That is, is the teleology inherent in the idea of "nanotechnology" intrinsic (meaning organisms are sui generis) or is it extrinsic (meaning organisms are machines)?
If the former, then what physical principles can we possibly appeal to to understand such an inside-out form of nanoengineering? The answer is, of course, that no one knows, but I think there are hints in the condensed-matter physics literature. So, in short, "nanotechnology," yes, but not in the manner of human technology. Vocabulary is very tricky out here on the frontier of our understanding.
IP: Logged
|
|
|
Topic Closed
Printer-friendly view of this topic
