theyeti

I'm not playing games, I'm trying (and it's not always easy) to have a mature discussion, in which hopefully both of us will learn from each other. Maybe we're just talking past each other, but I don't understand your objections. If you have no beef with the occurance of partial or whole chromosomal duplications, then why berate me for bringing them up?

theyeti

DNAunion

DNAunion: No, I did not ask for examples of how chromosomes - or genes (which you also keep focussing on) - could be replicated.

I quoted material that said that nondisjunction would have led to genomic expansion in the first eukaryotic cell divisions. I then asked about the effects of having AN EXTRA, WHOLE, COMPLETE, SINGLE chromsome, or of having AN EXTRA, COMPLETE, HAPLOID SET of chromsomes in lower eukaryotes would be. The topic was nondisjunction and its effects; it was not the various ways that chromosomes/genes can get replicated.

In case you are still not following the topic... If a single chromosome was involved in nondisjuction, then one cell would end up with an ADDITIONAL, SINGLE, COMPLETE, WHOLE chromosome. If nondisjunction was severe - because the process had not been tuned by evolution yet, for example - then ALL chromosomes might be involved in nondisjunction: if they were, then one cell would end with A WHOLE, COMPLETE, EXTRA HAPLOID SET of chromsomes.

DNAunion: The topic did not deal with gene duplications; it dealt with nondisjunction. The topic did not deal with IC, as I pointed out.

theyeti

Okay, maybe it's my fault for the misunderstanding. I thought the issue was simply one of genomic expansion, and I didn't know that it was nondisjunction in particular that you were interested in. However, I belive that severe nondisjunction will result in an extra haploid set of chromosomes only in sexually reproducing species. In an asexual diploid organism, which the eukaryotic ancestor probably was, and which yeast is most of the time, nondisjunction would result in a tetraploid. Is this right? (I ask that question sincerely, since I'm not up on my cell cycle lit.)

theyeti

theyeti

Just pubbing around PubMed, I found some interesting tidbits that I didn't know about before. The common house mouse has a great deal of chromosomal variation within its species. There are 24 different known karyotypes that live in the central Alps alone. There are also 40 different known fusion variants. Here is a <a href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?SUBMIT=y" target="_blank">page</a> with articles about it; see especially the first two.

Another example is with the harvestman (some sort of spider). There are 10 different known karyotypes living in a hybrid zone between 2n = 16 and 2n = 22 populations. (<a href="http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&list_uids=109371 94&dopt=Abstract" target="_blank">article</a>). So again we have an example of chromosomal differences within an interbreeding species. So while most chromosomal changes are harmful, at least in humans, they obviously do happen and result in viable organisms.

theyeti

DNAunion

DNAunion: Rereading some material, I found something that may help answer one of the questions I raised earlier. If anyone would like to discuss it a bit more, then let's.


DNAunion: Here's the new material.

DNAunion: They have no observable examples to base this on. But we can still ask, does the scenario make sense?

theyeti

Here's an article that answers some questions about where tubulin came from among others. <a href="http://www.the-scientist.com/yr2001/nov/palevitz_p18_011126.html" target="_blank">Deciphering Protein Evolution</a>.

And then with actin...

So the components of the cytoskeleton appear to have bacterial homolouges.

theyeti

edited to add: Sorry, I now notice that DNAUnion already mentioned those proteins.

[ March 17, 2002: Message edited by: theyeti ]</p>

Nic Tamzek

Hey DNA,

Sorry I haven't had time for a long discussion, the best I'm going to be able to do is to point you in various directions for places to look, & give rough descriptions of what I understand authors to be saying, so that you don't get immeadiately swamped by the smaller-scale details in various articles (e.g. the euk. fossil debate). I've been looking at this question in my spare time for a year or so, in connection with the 'how the cilium evolved' question, and so I think I've got some grasp of "if I were to take several months and study this question and write a review article, I would focus on X, Y, and Z, even though at the moment I only partially understand X, Y, and Z." The literature on the evolution of euks/mitosis is fairly large, and the question has been greatly complexified by all of the sequence data that is coming in, and interpretation thereof, plus the fact that replication is only partially understood anyhow, and then mostly in model organisms, etc.

What I'm trying to say is that there's not going to be any short answers to your questions, but that to me, an interested amateur looking in from the outside, it appears that scientists are working on the question and progress is being made.


Ah, yes. As far as I know this idea can be traced back to Cavalier-Smith also. Here's the basic idea:

- in prokaryotes, the DNA & the origin of DNA replication is attached to the cell wall (this origin must replicate & then one origin must be moved down to the other end of the cell before replication. How this physically occurs is I think still poorly understood, although perhaps recent work has cleared things up).

- Origin of phagocytosis (phagocytosis = food vesicles being formed from the infolding and pinching off of the out membrane)

Brief discussion:

To have phagocytosis you have to have to first have a cell that doesn't have a rigid wall but can hold together with a flexible membrane...Cavalier-Smith argues that this occured in the common ancestor of archaea and euks ("the neomuran"), as an adaptation to hyperthermophily). Once you have the strong flexible membrane, you are preadapted for phagocytosis evolving in a gradual manner, e.g.:

- bacteria stick to outside
- bacteria stuck to outside partially enveloped in a pocket
- pocket pinches off & forms a food vesicle.

...all of these stages are functional & can gradually convert from one to the next. TCS argues that the ancestors of archaebacteria stayed in their hyperthermophilic environments, whereas the ancestors of eukaryotes took their radical new adaptive niche (eating prokaryotes) & ran with it.

- Anyhow, once you get full phagocytosis, you have the problem that sooner or later, those replication origins (which were attached to the cell membrane) will eventually get internalized as vesicles are pinched off from the membrane. Ergo, now there is selection pressure for internal division, aka mitosis, via pleuromitosis and several other stages discussed by Cavalier-Smith in various places (like the 1987 articles), and somewhat by another references you should look up DNAunion:

The Major Transitions in Evolution
Maynard-Smith and Szathmary, 1995,

...which is the more technical (although still accessible) version that preceded the "Origins of Life" book that you referenced. Things are discussed in somewhat more detail there, although they are reviewing lit & so reference some of the TCS articles & other work that I've been mentioning.


Anyhow, one upshot of the Cavalier-Smith scenario is indeed that prokaryote replication origins and mitosis centrosomes should be in some sense homologous. This is a very queer prediction that would be totally unexpected except that it is suggested by this particular theory, so if it is confirmed it would be a big success. However at the moment it's my understanding that neither prok. replication origins nor euk centrosomes are particularly well understood, so we'll have to wait awhile to test the prediction. The homology discovered in the late 90's between eukaryote and prokaryote cytoskeleton components (well, whether or not proks really have 'cytoskeletons' is a matter of definition, but you know what I mean) is encouraging however.

My above account also hopefully explains better my assertion that "mitosis was a by-product of phagocytosis", so hopefully that makes sense at least in outline now. The evolution of mitosis also had all kinds of side effects, e.g. linear chromosomes & allowing of much bigger genomes in eukaryotes than prokaryotes, which allows complex multicellularity, etc.

Nic

DNAunion

DNAunion: Thanks. That was good "dense" article on exactly what I was interested in. And it did contain some info that I did not find in my quick scan of abstracts.

theyeti

Glad I could be of help. <a href="http://iccbweb.med.harvard.edu/mitchisonlab/Pages/People/Laura.htm" target="_blank">Here's</a> a page of a person that studies FtsZ specifically in regards to its tubulin similarities.

At the bottom is a list of publications that might be worth checking out.

theyeti

Nic Tamzek

DNAunion cites several of the many articles published supporting the homology of eukaryote and prokaryote cytoskeletal components, but then asks the reasonable question:


OK. The following is a brief, unreferenced discussion of what I think the answer is. I have tried to explain this several times to evolution skeptics without much success, and I have yet to find one single handy easily-citable reference that puts all this together in authoritative fashion, although I do think it would be supportable if I had a month to write up a literature review, and I think that most of the experts subscribe to something close to this.

The question: tubulins and similar components are highly conserved and slowly diverging in eukaryotes, yet the prokaryotic homologs are much more divergent than would be expected based on a protein-clocklike-divergence assumption (e.g., if you assume a clock then for some protein pairs the time of divergence between proks and euks is something like *20 billion years*, which is clearly wrong as life, let alone the universe, ain't that old).

The best references on this is the special issue of Philos Trans R Soc Lond B Biol Sci., (1995) -- DNAunion has already referenced one article by Nasmyth from that issue, the articles by Doolittle (can't find the title), and Mitchison:

...are good, although somewhat dated as the FtsZ-tubulin homology was just a strong suspicion in 1995, and the actin homology was similarly weak at the time.

Anyhow, the above refs give a place to start.

OK, so, "why the dichotomy" as DNAunion asks.


The short answer:

Proteins do not always diverge in a clocklike manner. However, steady vs. rapid spurts of divergence do not occur completely at random.

Under the following conditions, proteins are mostly likely to evolve in a gradual, steady 'clocklike' manner:

- Function remains constant. In this situation, most of the substitutions will be selectively neutral, and hence the fixation of substitutions is a stochastic process, which (like radioactive decay) actually leads to an average, somewhat clock-like behavior.

- Protein divergence will be most clock-like in the *early stages* of divergence under stabilizing selection (described in previous paragraph). Doolittle (1995) has a useful graph, where, under constant mutation, sequence similarity decays at a fairly constant rate from 100% similarity to ~50% similarity, but then the curve begins to flatten out, reaching flatness at very low levels of similarity (whatever the similarity of randomly generated sequences is, ~10% or something depending on similarity criteria).

The reason for this, of course, is that when sequences are 100% similar, any change will result in divergence, but when sequence similarity is only 50%, any change has a 1/2 chance of substituting an already-changed position in the sequence. This is the curve flattening out. And of course, once the two sequences are random with respect to each other, further changes will on average result in no further sequence similarity decay -- you can't get any more divergent than random similarity. At this point there is no clocklike behavior at all (although structural similarity may remain as it is widely observed that sequence similarity can decay to randomness but structures will remain quite similar)

- OK, so that describes the situation under clock-like divergence, e.g. if function is constant, then stabilizing selection simply maintains that function, which results in only neutral mutations getting through, which is stochastic, which results in clock-like behavior on average, especially for the early stages of divergence (the later stages are more and more sensitive to rapid changes -- e.g. 5% divergence means a lot more error in 'clock-time' if you're going from 45%-40% than if you're going from 100% to 95%).


BUT, what if:

(a) the function *changes*, or

(b) the organism goes through some radical change in habitat that effects most or all of the proteins in the organism.

In the case of (a), selection will be *directional* as the protein is being optimized for a new function, and this can change large chunks of the protein sequence in a geological instant. Selected substitutions will swamp neutral substitutions, clock-like behavior is lost, and sequence similarity will suddenly decrease rapidly (and recall that the similarity signal weakens the closer one gets to randomness anyhow).

In the case of (b), the proteins may retain the same basic function, but they may have to perform it under radically changed conditions. Two very likely cases wherein this might occur are:

- movement from mesophilic (regular temperature) to thermophilic (high temperature) or hyperthermophilic (really high temperature, e.g. 90-100 degress C or more) environments.

- movement from a free-living environment to living inside another critter as a disease or parasite.

In the case of temperature change (which of course can occur gradually as there is always a gradient between hot water and cold water), every single protein will have to adapt to avoid being denatured by temperature. Similarly, the DNA itself might adapt its G+C/A+T ratio to be more resistant to 'melting' at high temperatures. Both of these things would increase the rate of sequence divergence dramatically.

In the case of evolution into parasitism/disease, again the environment has changed suddenly in terms of temperature and biochemical environment, and innumerable examples show that genomes, biochemistry and morphology may radically change in these circumstances. This is one reason why studying diseases and parasites may not give the most unbiased information about early microbial evolution, which of course occurred before there were any large multicellular critters, let alone humans, to live in.

There may be other cases where rapid sequence divergence would be expected, e.g. adaptation to acidic or salty conditions perhaps.


Alright, so, to get back to the large sequence divergence between prokaryotes and eukaryotes. Eukaryotes are widely thought to be most closely related to archaebacteria (based on numerous detailed similarities of biochemistry, Cavalier-Smith has a long list in his Jan. 2002 article). Plus, adaptation to thermophily appears to be a likely way to develop a cell without a cell wall but with a strong flexible membrane (which is pre-adapted to phagocytosis as described before). Therefore, it stands to reason that their common ancestor went through a stage that had to adapt to high temperatures, probably even hyperthermophilic conditions (which archaebacteria can reach but which eubacteria basically have not). So there is reason for divergence #1.

Reason #2 is that, on the scenario for the evolution of euks that I described in my previous post, many of the cytoskeletal and replication proteins went through fairly radical changes of function as an adaptation to phagocytosis in the ancestor of eukaryotes. This too would contribute to rapid sequence divergence.

The nice thing about the above explanation for the large sequence divergence is that it falls quite nicely out of Cavalier-Smith's general scenario for the evolution of archaebacterial and eukaryotes, which was delineated way back in the 1980's. It also explains the more clock-like behavior within euks and archaeabacteria, as once the adaptations had been completed, selection would return to stabilizing for the new niches.

So, that's probably the explanation. I didn't have the time to collect a zillion citations to armor this explanation against the slings and arrows of IDists, but I do think that this is a likely solution & well-supported by the preponderance of the literature if not all of it.

Thanks, nic

PS: Here is one highly relevant article:

...unfortunately, my #*&@ university no longer appears to have online access to BioMedNet reviews, so I don't have text access. I quoted half the article on ARN once while debating Mike Gene though, so you might be able to dig it up there when ARN is back.

Nic