|
Author
|
Topic: Nelson and Ross: Problems with Characterizing the Protostome-Deuterostome Ancestor
|
Moderator
Administrator
Member # 1
|
posted 03. August 2004 16:59
Problems with Characterizing the Protostome-Deuterostome Ancestor
by Paul Nelson and Marcus Ross
Abstract: Since Darwin’s time, the origins and relationships of the bilaterian animals have remained unsolved problems in historical biology (Conway Morris 2000). One of the central difficulties is characterizing the common ancestor of the protostomes and deuterostomes. We argue that an unresolved conceptual puzzle has plagued the many attempts to describe this Urbilaterian, or, in Erwin and Davidson’s (2002) terminology, the protostome-deuterostome ancestor (PDA). Any organism sophisticated enough to be a realistic candidate for the PDA, with such characters as an anterior-posterior axis, gut, and sensory organs, must itself have been constructed by a developmental process, or by what we term an ontogenetic network (Ross and Nelson 2002). But the more biologically plausible the PDA becomes, as a functioning organism within a population of other such organisms, the more it will tend to “pull” (in its characters) towards one or another of the known bilaterian groups. As this happens, and the organism loses its descriptive generality, it will cease to be a good candidate Urbilaterian.
To read the entire poster, click here.
[ 03. August 2004, 22:55: Message edited by: Moderator ]
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 04. August 2004 11:13
Professor Paul Myers commented on this poster at his blog, http://www.pharyngula.org. His comments are in bold; my replies follow.
The first gigantic problem is that there was absolutely no data on the poster.
The relevant data are well-known (see below). Myers and I talked about much of it during our conversation (nearly two hours, by my estimate --when we finally checked the time, it was 11:30 pm, 30 minutes past the scheduled end of the poster session). Rather, the poster focused on an unsolved conceptual problem arising from data familiar to any developmental biologist or evo-devo theorist. Only a handful of biologists (e.g., Eric Davidson; see below) currently are trying to solve the problem, which deserves to be much better known.
Nelson's poster was all words and speculation. There was no substance there, no details to grapple with. It was awkward to discuss, because there really was no handle to grasp.
Myers grasped the problem, because he suggested a solution, which he sketched on a piece of paper. His proposed solution fails, for reasons given in the poster (see below) –- but that there is a real problem cannot be denied.
What was it about? Well, the gist of it, as near as I can tell (and I would hope Nelson will chime in with comments to correct anything) was the usual creationist argument from personal incredulity. The center of the poster was a drawing, a simple oval. We were supposed to imagine that this was the egg of the hypothetical pre-Cambrian common ancestor of both protostomes and deuterostomes. Next, we were supposed to imagine what the adult of this organism would look like. Then Nelson had a few photos of an adult fly, nematode, and sea squirt. Finally, we were supposed to imagine that modification of the development of modern organisms was impossible, that the modification of the development of the protostome/deuterostome ancestor was similarly impossible, and therefore, evolution was impossible.
Let’s consider some of the data -– evidence –- that Myers and I discussed, and readers can judge for themselves if this is an argument from incredulity or a genuine evolutionary puzzle.
In the Drosophila system, the anterior-posterior (A-P, or head-to-tail) and dorsal-ventral (D-V, or back-to-front) axes of the egg (body plan) are specified in the mother’s ovary. For instance, Mom deposits bicoid mRNAs at the anterior pole of the egg. After fertilization, these mRNAs will be translated into DNA-binding proteins, which will then diffuse along a gradient towards the posterior pole of the egg. Myers has described this system in some detail at his blog, in this and subsequent posts:
http://pharyngula.org/index/weblog/comments/transcription_factors_and_morphogens/
The diffusion of bicoid begins the process (or is an important aspect, anyway) of specifying the body plan of the fly. The events occur in a single large cell, the syncitium. Cellularization –- the formation of cells within the syncitium –- is yet to come.
But not all insects establish their body plans this way. In some wasps, for instance, studied by Miodrag Grbic and Michael Strand, complete celluarization is the first noteworthy event. That is, the embryo divides into two (and then many more) distinct cells immediately, in what is known as holoblastic cleavage. "Initial cleavage events in these tiny eggs differ from the canonical type of insect syncytial cleavage. [The] wasps undergo total (holoblastic) cleavage in which nuclear division is immediately followed by cytoplasmic division, forming individual cells (blastomeres)" (Grbic 2003, p. 635).
Here's the problem. If we suppose that the syncitial developmental architecture is primitive for long germ-band insects, as is widely accepted, then how did complete cellularization evolve in some wasps? In particular, if the segmentation cascade requires the diffusion of maternal transcription factors throughout the syncitium, what would happen if a cell membrane suddenly arose in the path of those diffusing morphogens?
quote:
The celluarized environment of the C. floridanum [a wasp] embryo violates the most important paradigm of Drosophila patterning: Initiation of the segmentation cascade begins with the diffusion of transcription factors in a syncytial environment. Even though C. floridanum lacks a syncytial stage, it still undergoes long germ-band development. ... Early cellularization of the C. floridanum egg, however, makes it unlikely that protein gradients could form by diffusion as occurs during syncytial cleavage in other long germ-band species....embryonic cells become dye uncoupled at the very beginning of embryogenesis and no syncytial stage ever exists during germ-band formation. (Strand and Grbic 1997, p. 140)
"Dye uncoupled" refers to experiments where Grbic and Strand injected early blastomeres with a dye smaller in molecular dimensions than the bicoid protein. (Maybe Micah can help me figure out how to post an image from one of these experiments.) The dye stayed in the injected blastomere and not did not diffuse. If the dye couldn't get past the cell membrane, then it is likely that the bicoid protein couldn't either.
Reviewing these fundamental differences, Grbic muses, "It is hard to conceptualize the evolution of a novel stage that disrupts one of the crucial paradigms of Drosophila development, maternal specification of the embryonic axis" (Grbic 2003, p. 640).
These are differences of developmental architecture just within Insecta. As one moves to large groups -- e.g., the phyla, or the superphylum Ecdysozoa -- the differences becomes much more dramatic.
[more in a bit]
[here's more, 8-5-04]
I won’t run through those differences here, but will cite Davidson’s (1990, pp. 365-66) summary:
quote:
Classical authors…and those of their successors who have attempted to deal with more than one embryonic form, have been struck by the amazing variety in the modes of embryonic development that exist in the various phylogenetic reaches of the Animal Kingdom…All embryos do indeed achieve the imposition of spatial patterns of differential gene expression, and yet some begin this process by intercellular interaction, and others even before there are any cells that could carry out such interactions; some rely on lineages that are autonomously committed to given functions from the moment they appear, others deal wholly in plastic, malleable cell fate assignments; some utilize eggs that before fertilization are cytoskeletally organized in both axes, some in one axis only, some apparently in neither; for some kinds of embryos every individual has a different cell lineage, while for others development depends on a set of rigidly reproducible canonical cell lineages; and some embryos display truly amazing regulative capacities….the differences among taxa in their modes of embryonic development are anything but trivial and superficial (certain hopeful reductionist delusions of recent years to the contrary).
Davidson concludes his review paper with this:
quote:
It is clear that a casualty of these arguments is the 19th century concept that early development must be an evolutionarily conserved process. We see that during evolution the regulatory genetic elements controlling embryogenesis have been reassembled in many different combinations. (p. 384)
A 19th century view, maybe –- but Davidson’s own view earlier in his career. He and I talked about this in China at the notorious 1999 Chengjiang meeting. In 1971, in a classic paper with Roy Britten, Davidson argued that the functional logic of embryogenesis requires conservation of early stages. To build, or more importantly to modify, a complex structure –- e.g., a novel body plan –- “must require a succession of events that amount to the modification of a whole pattern of regulatory relationships” (1971, p. 131). They go on:
quote:
The processes of development appear to be sequential, not only in the obvious time course of the events, but in the basic molecular and cellular mechanisms. In other words, the later stages are built on a foundation consisting of the events occurring earlier in development. As a result, changes in the parts of the developmental program operative at a given stage might result in drastic alterations of later developmental events. We expect, therefore, that the regulatory programs active earlier in development would also have been elaborated at early stages of evolution. Clearly, there would be greater freedom for modification and improvement by natural selection of what are now early developmental stages before the more complex and dependent later stages of development were superimposed on them. As a corollary, we would expect that, once the later stages evolved, the earlier stages of the developmental regulatory program would be more or less fixed. One can imagine modest alterations or additions to the early parts of the developmental program, but it would be very unlikely that such programs could be supplanted. Therefore the basic developmental patterns would be expected to have been elaborated earlier in evolution and be more widespread, phylogenetically. (p. 131)
Davidson no longer holds this view –- as Myers said to me in Calgary, “Evidence matters” –- because of the comparative data, showing fundamental differences in ontogenetic architectures, described in Davidson’s more recent work.
Note the adjective, however: comparative data. The functional logic of Davidson's 1971 argument still holds, and indeed is supported by the experimental evidence from model systems.
But...monophyly of the animals. If one assumes that two (or more) taxa showing divergent early development share a common ancestor, then one has prima facie evidence that early development can vary dramatically. In his plenary lecture at Calgary, for instance, Brian Hall said (I’ll paraphrase), ‘Once upon a time the neo-Darwinians thought that early development was functionally constrained, and thus unlikely to vary much. Now of course we know better: We’ve got hundreds of examples of divergent early development.’ What Hall didn’t show (because, to my knowledge, it doesn’t exist) was experimental evidence showing heritable changes in such characters as cleavage patterns or modes of gastrulation.
Indeed, in conversations at meetings over the past few years, I’ve had some evo-devo biologists tell me that they would never expect to observe such variations (in part because of temporal asymmetry arguments, as discussed in part 3 of my SDB poster: evolution worked differently in the past). Myers suggested that, in this respect, neither Drosophila nor C. elegans was an adequate model system for trying to understand early metazoan evolution.
OK –- so what would be a good current model? Or even a model in principle? That’s the point of the thought experiment in the poster: Describe a simple metazoan, with a few hundred cells falling into a handful of specialized cell types, which undergoes a process of cell division and differentiation from an egg. How would this hypothetical organism escape the functional constraints that we know exist for all (real) developing systems?
Here’s where Myers drew what he calls a “crude triploblastic worm.” It would be relatively easy, he argued, for this organism to vary so that it would qualify as a reasonable PDA (protostome-deuterostome ancestor).
Maybe. But problems arise as soon as one tries to put flesh on the bones of the hypothesis. Suppose the crude worm were something like known flatworms. A flatworm egg wants to turn into a flatworm, with its gut, muscles, germ line, sensory organs, and epidermis in more or less the right place –- and not something else. (I would say the constraints we see in developing organisms are a generic feature of causally asymmetric systems, where entrenched upstream features govern downstream consequences, but that’s another post.)
In fact, this is the way we discover how development works in model systems. Perturb this cell lineage: watch what happens to the organism. Delete that gene: watch what happens. And so on. Developing systems have normal targets, with a range of viable variants in the immediate neighborhood of the target. We find the target and possible variants mainly by causing embryos to tell us, “All right –- I needed that gene to form my gut [or whatever]. Alas now I’m dead.”
Focusing on the PDA makes this problem especially interesting. If one jettisons monophyly for the metazoans, the problem goes away (in large measure); there’s no longer any theoretical demand to derive disparate developmental architectures and body plans from a common ancestor.
But monophyly is not up for grabs, so the problem persists.
Back to Myers:
No data. No experiments. No predictions. Just a request to build up a model of evolution and development in our imaginations. I'm sure it works for creationists, and being liberated from demands for evidence makes it easy to compound one's biases and come up with the answer Nelson wants, but in this mob of good practical naturalists who expect at least a nod towards some data, it fell flat. I'm afraid that when I was free to just imagine the adult protostome/deuterostome ancestor that would arise from his evocative, all-powerful Oval of Infinite Potential, I had no problem scribbling up a cartoon of a crude triploblastic worm, and saw no obstacle to incremental specialization of it's component parts in development and evolution.
If the problem were that easy, it would have been solved by now. But metazoan phylogeny is highly unstable [anyone want to place bets on how long Ecdysozoa survives as a clade?] and candidates for the PDA have come and gone. I’d argue the problem won’t yield because it presupposes an organism that we can’t even imagine (honestly). Put it this way: Suppose Myers and I and anyone else who cared to join us had a whole week, with chalk, blackboard, and all the data of comparative embryology, anatomy, and paleontology, to flesh out the PDA. Where would we find ourselves at the end of the week?
At best, with a list of non-specific characters: A-P axis, gut, sense organs, etc. Right back where we started.
Now you could argue that that's just compounding my atheistical, materialistic biases, and is as meaningless as Nelson's assumption of a conclusion. I would make two arguments against that, though. One of the purposes of a scientific presentation is to share the evidence and logic that leads to a particular conclusion in such a way that there isn't much room for argument, and that argument is at least directed towards constructive, alternative hypotheses. An open thought experiment that encourages unfettered guesswork is not science.
I wasn’t encouraging guesswork. I wanted a specific answer. The fact that none is forthcoming –- that the thought experiment leads only to guesswork, 145 years after the Origin –- is the point of the poster.
I do have to commend Nelson for having the guts to expose the hollow vacuum at the heart of anti-evolutionary thinking to the critical eyes of a swarm of practicing scientists, but there is another troubling problem here. This presentation is going to go on a list at the Discovery Institute of Intelligent Design forays into mainstream scientific venues.
The poster said nothing about intelligent design; for that matter, the poster said nothing about evolution “being impossible.” The evolutionary puzzle the poster describes would exist whether I showed up at SDB or not. If the bilaterians are monophyletic, then we must describe their common ancestor. That’s the problem. It existed for Darwin and it exists today. The problem will either be solved or it will persist. I predict it will persist, and will worsen, because of the conceptual and evidential difficulties outlined in the poster.
I was glad to meet Paul Myers, to hear his criticisms of the poster, and look forward to seeing him at SDB 2005 (if not sooner –- hey, Paul, invite me up to UM-Morris; I promise your students will have the liveliest discussion of their semester).
References
Britten, Roy J. and Davidson, Eric. 1971. Repetitive and non-repetitive DNA sequences and a speculation on the origins of evolutionary novelty. Quarterly Review of Biology 46:111-131.
Davidson, Eric. 1990. How embryos work: a comparative view of diverse modes of cell fate specification. Development 108:365-389.
Grbic, Miodrag. 2003. Polyembryony in parasitic wasps: evolution of a novel mode of development. Int. J. Dev. Biol. 47:633-642.
Strand, Michael and Miodrag Grbic. 1997. The Development and Evolution of Polyembryonic Insects. Current Topics in Developmental Biology 35:121-159.
[ 05. August 2004, 13:11: Message edited by: Paul A. Nelson ]
IP: Logged
|
|
andyg
Member
Member # 415
|
posted 04. August 2004 20:37
Paul,
Would it be possible to post your responses to The Panda's Thumb as well?
Thanks,
Andy
IP: Logged
|
|
Paul A. Nelson
Member
Member # 26
|
posted 06. August 2004 15:40
Andy,
I posted (actually, accidentally double-posted) the reply at The Panda's Thumb.
PN
IP: Logged
|
|
|
Printer-friendly view of this topic