Front-loading and choanoflagellates
by Krauze
Last year, I wrote: "I'll make a bold prediction: Once we start sequencing the oldest branches of our family tree, we will see that we have inherited more of our complexity from the microcopes [sic] than previously thought." This prediction flows naturally from my thoughts about front-loading - the conjecture that the first organisms were designed with a future state in mind. So, if the first eukaryotes (a large group of organisms, which includes plants, animals, fungi, and several one-celled organisms) were front-loaded for multicellularity, we would expect them to contain genes required for multicellular life. (For a longer introduction to front-loading, including answers to the most common questions, see my three part series.)
So, how has my prediction fared? To see this, we first need to look at some small critters with a big name: Choanoflagellates. They are small, single-celled eukaryotes that use their whip-like flagella to move through the water (these flagella are constructed in a way entirely different from the bacterial flagellum, and the two structures are not believed to be related). Furthermore, they are thought to be closely related to the organisms that gave rise to animals and sponges (these two comprise a group called Metazoa). So, by looking at these critters, we can get a good picture of how our ancestors looked.
Evolution of multicellularity. A. Proposed stages in the evolution of choanoflagellates into simple, multicellular organisms. B. Phylogeny of eukaryotes, showing how choanoflagellates are related to metazoans. Most animals belong to the Bilateria. The numbers represent (1) a unicellular flagellate, (2) a colonial group of cells, and (3) a simple, multicellular organism. From King, 2004.
In the article, "The Unicellular Ancestry of Animal Development", Berkely biologist Nicole King reviews some evidence that is relevant to front-loading. When the first eukaryote genomes were sequenced, many genes appeared to be unique to animals, including genes for cell-cell adhesion, receptors, and signal molecules. This wasn't that surprising, as the ability of cells to stick together and communicate with each other is exactly what's required if you want to be multicellular. But by studying choanoflagellates, King and her colleagues discovered that some of these genes were older than expected:
Surveys of expressed sequence tags (ESTs) and full-length cDNA sequences reveal choanoflagellates to express multiple members of gene families previously thought to be unique to animals. Despite the apparent simplicity of their lifestyle, choanoflagellates express a surprising diversity of animal signaling and adhesion gene homologs including TKs, G protein-coupled receptors, cadherins, and C-type lectins (King et al., 2003). In addition, several predicted polypeptides from choanoflagellates contain multiple protein-protein interaction domains (e.g., EGF, SH2, TNFR, and CCP) that typically function in animal signaling and adhesion proteins.
That this was indeed a surprising finding is supported by this article in ScienceMatters, quoting King: "I was surprised to learn that so much of animal biology was in place before the origin of animals," King says. "And I think that's what motivates most scientists - not learning that you were right, but learning that you were wrong." But from the perspective of front-loaded evolution, this isn't unsuspected at all. If the first eukaryotes were designed with multicellularity in mind, it only makes sense that they would contain genes required for this way of life.
In closing, let me point out some caveats. First, the obvious one that since this research was published in 2003, it strictly speaking doesn't count as a confirmed prediction on my part. Perhaps a better term would be retrodiction. Paleontologists often make retrodictions about which fossils they will find, even though the fossils already exist. Similarly, I only heard about these findings a couple of days ago, when Daniel Morgan linked to the ScienceMatters article, highlighting King's research.
The second caveat is more technical. The phylogeny of the lower branches of the eukaryote tree is still not completely ironed out, and the view that choanoflagellates are the ancestors of animals may change in the future. Maldonado (2004) suggests that choanoflagellates are descended from sponges that have been simplified. If this is the case, those genes mentioned before aren't evidence of front-loading, but simply the vestiges of a multicellular past. More research is required to rule out this possibility.
Thirdly, choanoflagellates may be old, but they aren't the oldest eukaryotes. From the first eukaryotes to choanoflagellates are billions of years, and it's possible that the genes weren't present in the original population of designed organisms, but are the result of subsequent evolution. To really start testing my prediction, we'll need to take a closer look at some of the deepest branching eukaryotes, such as slime molds or amoeba.
References
King N., 2004, "The Unicellular Ancestry of Animal Development", Developmental Cell 7(3):313-325
King N., Hittinger C.T., & Carroll S.B., 2003, "Evolution of Key Cell Signaling and Adhesion Protein Families Predates Animal Origins", Science 301(5631):361-363
Maldonado M., 2004, "Choanoflagellates, choanocytes, and animal multicellularity", Invertebrate Biology 123(1):1"“22
January 16th, 2006 at 4:06 pm
Hey Krauze, you stated before that, front-loading has deep implications for ID. What are they? They would you describe it, especially from Caroll's book, sounds telic to the core.
Comment by Benji — January 16, 2006 @ 4:06 pm
January 16th, 2006 at 4:39 pm
They would you describe it, especially from Caroll's book, sounds telic to the core.
The way you describe it sounds telic to the core! The way I asked it sound drunk to the core! (Smile…) ***
Comment by Benji — January 16, 2006 @ 4:39 pm
January 16th, 2006 at 4:40 pm
I.E, Krauss, your description of Caroll's book sounds telic to the core.
Comment by Benji — January 16, 2006 @ 4:40 pm
January 16th, 2006 at 6:07 pm
Krauze, interesting stuff.
What if the 'genes for multicellularity' are in fact reused 'genes for unicellular polarity'. With an effective toolkit for gene expression regulation, you may only need a few flexible components. Maybe the first eukaryotes were already so complex that they consisted of all the components needed for multicellular life. Compare it to the Java core classes. You can do an afwul lot of complex programming with them. It would be technically front-loading, no?
This question was inspired by PZ Myers' page on segmentation, where he assumes that syncytial development is convergent evolution from a short-germ ancestor. I can quite easily imagine that syncytial development can reuse an existing anterior-posterior gradient in a single cell. A single cell with a mutation that causes a temporal block in cytokenesis becomes syncytial. After the formation of the cellular blastoderm, the intially locally acting proteins (like bicoid) would then affect on entire cells at either end of the blastoderm. It seems quite difficult to go from a cellular stage to a syncytial stage, though.
I would also not directly assume that a colonial form as suggested by the picture was an intermediate to multicellularity. The two are fundamentally different things and a 2-cell intermediate to multicellularity seems to me easier.
Comment by AdR — January 16, 2006 @ 6:07 pm
January 16th, 2006 at 7:21 pm
Hi Benji,
"Hey Krauze, you stated before that, front-loading has deep implications for ID. What are they?"
Well, front-loading is a particular expression of the telic approach. You might say that with front-loading, we are taking the baby steps towards that elusive "theory of design" .
"[The way] you describe it, especially from Caroll's book, sounds telic to the core."
Of course, Carroll neither advocates front-loading nor intelligent design. But yes, the terms he uses to describe the gene regulation of the developing embryo are quite telic. We could say that biology has been well served by a "methodological designism".
Comment by Krauze — January 16, 2006 @ 7:21 pm
January 16th, 2006 at 7:28 pm
I thought, Caroll, was an avowed atheist who distances himself from teleological thinking? Does he know of the implcations that his work has?
Comment by Benji — January 16, 2006 @ 7:28 pm
January 16th, 2006 at 7:30 pm
Hi AdR,
Thanks for your comments. Now all we need is Mesk and KC, and we'll have the perfect thread.
"What if the "˜genes for multicellularity' are in fact reused "˜genes for unicellular polarity'."
Interesting possibility. Considering that multicellular polarity is constructed from the polarity of the fertilized cell, you might be on to something here.
"Compare it to the Java core classes. You can do an afwul lot of complex programming with them. It would be technically front-loading, no?"
Indeed. This also raises the possibility that rather than being front-loaded for a (more or less) specific outcome (multicellularity), the first eukaryotes were given some versatile, Lego-like building blocks, of which multicellularity is just one of many possible uses. I suspect that many of these questions will be easier to answer in the future, as the prize of whole-genome sequencing becomes lower, giving us more examples of the uses to which various genes have been put.
"I would also not directly assume that a colonial form as suggested by the picture was an intermediate to multicellularity. The two are fundamentally different things and a 2-cell intermediate to multicellularity seems to me easier."
Could you elaborate on this?
Comment by Krauze — January 16, 2006 @ 7:30 pm
January 16th, 2006 at 7:39 pm
Krauze,
Do you believe-and I asked this before-that specified complexity can be detected in the fossil record, thus rendering ID?
What are your thoughts on this"¦
Comment by Benji — January 16, 2006 @ 7:39 pm
January 16th, 2006 at 7:40 pm
I'm not sure, but I think Steve Meyer tried to do that. I don't know if he used 'SC' as a marker?
Comment by Benji — January 16, 2006 @ 7:40 pm
January 16th, 2006 at 8:10 pm
Hi Benji,
"Do you believe-and I asked this before-that specified complexity can be detected in the fossil record, thus rendering ID?"
The way I see it, specified complexity is characteristic of life. And since fossils are the result of life, they also signify specified complexity. However, I don't think specified complexity by itself proves intelligent design, merely that it is one of many pieces of evidence suggesting it.
Comment by Krauze — January 16, 2006 @ 8:10 pm
January 16th, 2006 at 10:11 pm
Hmmm…I don't think I understand? Do you advocate ID or not? Well, I think you consider yourself an ID evolutionist, right? What does that mean? Do you believe in the designer? To most christians, like Behe and Dembski, the designer, namely, is God.
Comment by Benji — January 16, 2006 @ 10:11 pm
January 16th, 2006 at 10:13 pm
Oh…you believe ID needs more than CSI. At least that's the way it looks.
Comment by Benji — January 16, 2006 @ 10:13 pm
January 16th, 2006 at 10:29 pm
Krauze,
In a previous post you acknowledged the problem of protecting unexpressed front-loaded information against mutational degradation. Your argument in that post was that the genes for multicellularity were already being expressed in the single-celled ancestor of animals, and were therefore protected against corruption.
The problem I see is that ALL of the genetic information ultimately expressed in animals, up to and including modern forms, would have to be similarly expressed in all ancestors going back to the earliest common one in order to protect it from degradation. Do you really believe that the earliest ancestor of choanoflagellates (assuming they are confirmed as the ancestor of animals) expressed all of this information?
Also, you commented on King's quote:
"'I was surprised to learn that so much of animal biology was in place before the origin of animals,' King says. 'And I think that's what motivates most scientists - not learning that you were right, but learning that you were wrong.' But from the perspective of front-loaded evolution, this isn't unsuspected at all. If the first eukaryotes were designed with multicellularity in mind, it only makes sense that they would contain genes required for this way of life."
I would point out that we would expect this from the Darwinian perspective as well as from the front-loaded perspective. The Darwinian perspective expects genomic change to be gradual, and does not expect a large number of mutations to happen at the same time by chance. I would argue that contra King, the result is actually expected and not a surprise from the Darwinian perspective.
I look forward to hearing your comments.
Comment by keiths — January 16, 2006 @ 10:29 pm
January 17th, 2006 at 4:03 pm
AdR: "I would also not directly assume that a colonial form as suggested by the picture was an intermediate to multicellularity. The two are fundamentally different things and a 2-cell intermediate to multicellularity seems to me easier."
krauze: Although it should be quite easy for single cells to adhere to each other and thus form a colony, they are not genetically one organism. In the end, multicellular organisms arise from a first single cell, and are therefore genetically and developmentally coupled to each other.
I don't really see how the genetic switch can be made from a colony to a multicellular organism. The mutations that caused the cooperation in the colony would be unrelated to the mutations necessary for multicellularity arising from a single cell/zygote. Mechanistically, it just doesn't seem to make sense.
It is interesting however, how the evolution of Dictyostelium went. They do form aggregates, but they are amoebas and it seems like a high degree of cooperation, but fundamentally different from "˜real' multicellularity.
He mentions both possibilities, but I find the survival "˜drive' for evolution dubious and speculative. To me, this is a clear example for a tendency to increasing complexity. There is no need for an adaptation to an external factor, just multicellularity per se is enough.
He uses terms like "˜Selection for fixation', 'The benefits of staying together', "˜Escape from predation', which seem to me neo-Lamarckian.
Comment by AdR — January 17, 2006 @ 4:03 pm
January 17th, 2006 at 4:57 pm
Hi Keiths,
"The problem I see is that ALL of the genetic information ultimately expressed in animals, up to and including modern forms, would have to be similarly expressed in all ancestors going back to the earliest common one in order to protect it from degradation."
Front-loading doesn't require that the first organisms had a copy of every gene. New genes can have arisen since, for example by duplication and convergence of existing genes. Indeed, as this article concludes, "the principal process by which new gene functions arise has been by making use of preexisting genes."
"The Darwinian perspective expects genomic change to be gradual, and does not expect a large number of mutations to happen at the same time by chance. I would argue that contra King, the result is actually expected and not a surprise from the Darwinian perspective."
Gradualism alone wouldn't lead one to expect this. Those genes could also have arisen gradually after the choanoflagellate lineage diverged from the lineage leading to animals (i.e. between point 1 and 3 on the diagram).
Comment by Krauze — January 17, 2006 @ 4:57 pm
January 17th, 2006 at 6:53 pm
Hi AdR,
"Although it should be quite easy for single cells to adhere to each other and thus form a colony, they are not genetically one organism. In the end, multicellular organisms arise from a first single cell, and are therefore genetically and developmentally coupled to each other."
As you yourself quote Nicole King ("she", not "he") as saying, such a colony could be formed by cells failing to detach after mitosis. If so, a colony of cells with nearly-identical genomes could indeed arise from a single cell.
"The mutations that caused the cooperation in the colony would be unrelated to the mutations necessary for multicellularity arising from a single cell/zygote. Mechanistically, it just doesn't seem to make sense."
Can't you elaborate on this as well?
As for your comments on selection pressures, I think it may be immature to do anything more than speculate about the causes behind the origin of multicellularity.
Comment by Krauze — January 17, 2006 @ 6:53 pm
January 18th, 2006 at 5:35 am
Hi Krauze,
Thanks for the reply.
I'm new to this blog so I don't know if you agree with Dembski's ideas regarding specified complexity. But if you do, then you believe that gene duplication and convergence, being undirected natural processes, cannot create additional specified complexity. That would mean that all of the specified complexity involved in wiring a primate brain, for example, was already contained and expressed in the genome of the earliest eukaryotic ancestor of animals. That sounds like a stretch to me. Is that what you're saying?
That's true. The point I was trying to make is that under the Darwinian interpretation, almost all of the mutations necessary for colonialism must have already existed in the final non-colonial ancestor, whether or not it was similar to the choanoflagellates. If they weren't already present, then multiple coordinated near-simultaneous mutations would have been needed to achieve colonialism, which is improbable. For that reason, it didn't surprise me to learn that many of the mutations were already in place in the choanaflagellates.
On the other hand, if the mutations necessary for multicellularity are quite distinct from the ones required for colonialism, then I can better understand King's surprise.
Comment by keiths — January 18, 2006 @ 5:35 am
January 18th, 2006 at 3:56 pm
Krauze,
I define a colony as an aggregation of individual single cells. Multicellularity is for me the coordinated development starting with a single cell. The genes responsible for multicellularity would be involved in timing of cell divisions, sequential gene expression, and of course the conservation of the germ line. The (only?) gradual way for such a complex scheme seems to start with 2 cells as an intermediate and build from there Genes for colony-formation would be more in the line of adhesion factors, gap junctions etc. I am aware though that there are scenario's where these factors would facilitate the evolution of multicellularity.
Anyway, I do think you touched upon one of the fundamental aspects of evolution. Multicellularity is one of the milestones in evolution. From an engineering approach because it increases the potential for further complexity and adds potential new interfaces. I have the feeling that there are only a few fundamental design patterns necessary. For instance that there is initially only the syncytial development/belated cytokinesis possible. All other mechanisms would lead to organisms that are too loosely-coupled to evolve into multicellularity. But I am open to all possibilities, as this is a complex subject.
I wonder if the different developmental routes (meroblastic, syncytial, holoblastic etc) that exist for multicellularity will reflect their evolutionary origin.
Comment by AdR — January 18, 2006 @ 3:56 pm
January 18th, 2006 at 4:18 pm
Hi Keiths (or is it "Keith S"),
"That would mean that all of the specified complexity involved in wiring a primate brain, for example, was already contained and expressed in the genome of the earliest eukaryotic ancestor of animals."
I believe that life is capable of producing specified complexity, just like the chess-playing computer Deep Blue is. I don't spend much time thinking about whether my opinions agree with those of Dembski.
"The point I was trying to make is that under the Darwinian interpretation, almost all of the mutations necessary for colonialism must have already existed in the final non-colonial ancestor, whether or not it was similar to the choanoflagellates."
True, but choanoflagellates aren't the final non-colonial ancestors, they're just the closest surviving ancestor. As the figure in my post also illustrate, there was a considerate period of time after the lineages leading to choanoflagellates and metazoans split, but before Metazoa arose, in which those genes could have evolved.
Comment by Krauze — January 18, 2006 @ 4:18 pm
January 18th, 2006 at 5:07 pm
Hi Krauze,
My name is Legion: for we are many.
Just kidding. It's Keith S.
Sorry to press you on this, but I just want to understand your position. I had assumed that as an intelligent design proponent, you found random mutation plus natural selection insufficient to account for all of life's complexity. I thought you saw front-loading as a way to get specified complexity into the organism without relying on RM + NS.
If life is capable of producing specified complexity, as you say, why exclude RM + NS as the source of that complexity? Is there a different, specific mechanism you have in mind?
I apologize if you've already covered this in earlier posts. Like I said, I'm a newcomer to the blog.
Regards,
Keith S.
Comment by keiths — January 18, 2006 @ 5:07 pm
January 18th, 2006 at 5:34 pm
Hi Keith,
"My name is Legion: for we are many."
Ah. Your mother's name wouldn't be "Rosemary" by any chance?
"Sorry to press you on this, but I just want to understand your position. I had assumed that as an intelligent design proponent, you found random mutation plus natural selection insufficient to account for all of life's complexity."
It's quite okay. To clarify, I don't think in terms of sufficiency. That is, rather than concerning myself with what could have happened, I'm more interested in finding out what actually did happen. Was random mutation plus natural selection the major cause of finch beak variations? Indeed, it seems so. Was random mutation plus natural selection the major cause of the bacterial flagellum? There isn't much evidence to suggest so. Etc. But I don't think there are any in principle reasons why natural selection shouldn't lead to biological complexity, if that's what you're asking.
"I thought you saw front-loading as a way to get specified complexity into the organism without relying on RM + NS."
Then, I am sad to say, you were wrong. I don't view front-loading as a means to some desired end ("not relying on RM+NS"), any more than the average evolutionary biologist sees evolution as "a way to explain the wonders of life without having to stand accountable to a deity" (to cite a common creationist stereotype). Rather, I'm interested in front-loading because it's an interesting idea, and because it seems to have some potential overlooked by most (as the post that started this thread illustrates).
Comment by Krauze — January 18, 2006 @ 5:34 pm
January 19th, 2006 at 5:48 am
Hi Krauze,
Thanks for the clarifications.
If I understand your positions, you acknowledge the need for the continuous expression of front-loaded information so that natural selection can protect it from degradation over time by weeding out any deleterious mutations. You acknowledge that all of the genes for complicated animal development might not be there in the front-loaded ancestor, but that gene duplication and convergence can account for the extra genes needed. You also see no reason why RM + NS cannot, in principle, lead to increased biological complexity.
A few questions:
1) A front-loaded mechanism would seemingly need some way of deciding when, and in what order, to "turn on" individual chunks of front-loaded information. Environmental information would also need to factor into this determination in many cases. Do you have a proposal for what the coordinating mechanism might be?
2) Since you argue that some of the front-loaded information is implicit and must be made explicit via gene duplication, that would seem to mean that the duplication events themselves would need to be front-loaded somehow so that they would occur at the proper times. How would this work?
3) Random mutations are sometimes beneficial. Beneficial random mutations can alter the course of an evolutionary lineage. How would a predetermined front-loaded scheme accommodate these deviations?
4) A common ancestor must have the necessary front-loaded information for all possible descendants. How would the coordinating mechanism decide which branch of the descendant tree it was on, and therefore which information to express?
Regards,
Keith
Comment by keiths — January 19, 2006 @ 5:48 am
January 19th, 2006 at 9:31 pm
Hi Keith,
It seems that your questions are based on the misunderstanding that front-loading involves programming the evolution of every detail of every species. So instead of answering them, I'll adress the misunderstanding. Or rather, I'll let Mike adress it, as he's already written about this at length.
Go here (post number 3).
Comment by Krauze — January 19, 2006 @ 9:31 pm
January 20th, 2006 at 3:57 pm
Very interesting discussion, as always. I don't want to distract from the topic, but somewhat tangential to the topic:
"In closing, let me point out some caveats. First, the obvious one that since this research was published in 2003, it strictly speaking doesn't count as a confirmed prediction on my part. Perhaps a better term would be retrodiction.""”Krauze
I've seen other IDers do this, particularly Mike Gene. Prediction and "retrodiction" (a truly awful term) are formally identical. In conventional language we make a distinction, but they are just the same thing, as on the basis of what we do know, we make a statement of fact (a proposition) about what we don't (currently) know"”but have a reasonable prospect of discovering or confirming. It doesn't matter whether we "predict" the past or predict the future. What we are predicting is a change in the ever-present, current, state of knowledge"”whether that knowledge is about the past or the future.
(I was just having an interesting discussion about the laws of mechanics. In mechanics there is no distinction between "past" and "future," so how in the world could physicists ever make a "prediction"? But they do! They can do that because what they are really predicting is a change in their current knowledge, irrespective of whether that knowledge is about the past or the future.)
So I don't think you or anyone else should so casually discount "retrodictions." The problem with "retrodictions" is that I don't know and cannot in any conceivable way make any determinations about the current state of your knowledge. I kinda gotta trust what you report about the state of your knowledge. My impression, Krauze, you're a pretty trustworthy guy.
I will accept your "retrodiction" as effectively equivalent to a "prediction." So you, Krauze, might say that a design-theoretic heuristic does lead to "predictions." Or is that too much to assert?
(Nevermind that last question. Carry on. LOL)
Comment by Rock — January 20, 2006 @ 3:57 pm
May 2nd, 2006 at 2:04 pm
Krauze,
The link to Daniel Morgan's site is broken, but his post is still there, just with a different URL.
http://danielmorgan.blogspot.c...
Comment by tomfool — May 2, 2006 @ 2:04 pm