Molecular Similarities Echo Design Themes?
By Mike Gene
Examples of apparent molecular convergence are increasingly easy to find. Sometimes, they have long been right under our noses, yet no one (ASAIK) has pointed some of them out. Let me offer yet another example of apparent molecular convergence and then comment on its possible relevance to ID.
One of the defining characteristics of the eukaryotic cell is its cytoskeleton. The cytoskeleton is not a static cellular skeleton, but is instead a dynamic structural substratum upon which almost every cellular process depends (to one extent or another). Yet what is perhaps most interesting to me is that the cytoskeleton appears to be composed of two independent, yet very similar subsystems: the actin/myosin system and the tubulin/kinesin/dynein system.
Both systems can be defined as having three basic components:
Let's consider the first two. In the case of #1, both actin and tubulin are monomeric proteins that can form long polymers that are inherently unstable. This inherent instability is actually a good design, as it allows cells to constantly remodel their internal state such that the remodeling can respond very quickly to changing conditions. [For example, during mitosis, the tubulin fibers (known as microtubules) of the cytoskeleton dissociate into their monomeric state and quickly reassemble in a different arrangement to serve as the spindle fibers during metaphase and anaphase.] If, however, you want stable fibers, simply "weld" them together with a rich toolkit of linker proteins.
The mechanism behind this dynamic instability is strikingly similar for both actin and tubulin, where both monomers hydrolyze nucleoside triphosphates after being incorporated into a growing filament (ATP in the case of actin; GTP in the case of tubulin). Both filaments demonstrate a structural polarity at their ends (called "+" and "-"), where assembly occurs more often at the positive end. In both cases, monomers bound to a nuceloside triphosphate bind to the ends, and upon incorporation, eventually carry out hydrolysis, which in turn, decreases the stability of the polymer where the affinity between monomers is decreased. Finally, the monomers in both fibers contain binding sites for their respective motor proteins. And that takes us to #2.
Myosin, kinesin, and dynein have roughly the same molecular morphology. All possess globular "heads" that bind to their respective fibers and also hydrolyze ATP. In each case, the hydrolysis of ATP is coupled to conformation changes that allows these heads to "walk" along their respective fibers. The three dimensional structures of the kinesin and myosin heads are also quite similar. The globular heads for all of these proteins are connected to helical "stalks" or "necks" which in turn are connected to adaptor regions that can bind cellular cargo. Motor proteins attach to all sorts of material through their adapter ends and then carry this cargo to various destinations along the fibrous tracks (it is becoming more clear everyday that simple diffusion plays a less important role in cell biology than anyone ever suspected).
Keep in mind that the systems are independent. Myosin motor proteins use actin filaments as their tracks and kinesins and dyneins use microtubules. Yet so similar are these two systems that biologists have proposed essentially the identical mechanism of action for the sliding that takes place among dyneins and microtubules in flagella and actin and myosin in muscle cells.
But in spite of all these similarities, there is really no good evidence that the various members of the two classes ever shared a common ancestral form. The amino acid sequences are very different between system components. Yet within a system, the amino acid sequences are highly conserved. For example, actin is among the most highly conserved of eukaryotic proteins. Typical actins are about 350 amino acids in length, and when the sequences from amoeba and vertebrates are compared, there is an 80% identity.
When all is said, we are apparently faced with two independent systems that show striking similarities at various levels: the activities of individual components, the behavior of multimeric complexes, and the overall mechanism of action. Yet such similarities are not obviously explained by common descent. Thus, what explains this remarkable convergence?
Standard explanations would invoke natural selection and similar selection pressures. But this is really not that convincing. First of all, we essentially have to assume both systems originated at about the same time. Otherwise, if one system gets a significant head start, it is hard to see how another could get a selective foothold in its original, primitive state simply because both systems are so similar. Secondly, and perhaps more importantly, is that natural selection does not entail the level of specificity that is entailed in the similarities. Natural selection merely retains that which "works." If there are lots of ways to make a cytoskeleton work (as the contingency-views of Gould would imply), the specific similarities are hard to explain. What we should have are two very different systems that happen to spit out the same general function.
Clearly, this convergence is telling us there are not many ways to make a cytoskeleton with motor proteins. On the contrary, there seems to be only one basic way of making a eukaryotic cytoskeletal system. And if this is true, the design inference begins to surface.
Even if we were to grant a huge role to natural selection in the design of these two independent systems, it would become clear that natural selection is simply reflecting the constraints of physical laws. The basic fabric of Nature herself appears to entail features that force natural selection into choosing this specific cytoskeletal design. Thus, Nature may be designed so that natural selection finds intended outcomes.
On the other hand, there is no reason to grant natural selection a creative role in the development of these systems since there is no good evidence that they evolved. Instead, we can note that in many ways, the similarities in these systems are conceptual. That is, a core design concept may have been implemented twice such that the similarities map to Mind's intervention.
ID allows us the freedom to explore the world with this perspective. And once we do so, we immediately confront a seemingly difficult problem. Why would any intelligent designer use two different systems to carry out the same ends? Why the redundancy? Would not an intelligent designer simply pick the best one and implement it alone?
The first thing to note is that mere redundancy is not a problem for ID, as redundancy confers a degree of robustness. That is, if one system goes down, another "back-up" system can be employed to restore function. What is problematic for ID is not redundancy, but needless redundancy. Thus, are the myosin/actin and kinesin/tubulin systems needlessly redundant? If we are dealing with intelligent design, we can predict that the answer is no. And as it turns out, there is plenty of evidence to support this prediction. In fact, the evidence strongly indicates the two systems are not even redundant.
Our first clues come from some important differences. Although I have previously highlighted the uncanny similarities in the two subsystems, distinct differences exist. For starters, recall that tubulin hydrolyzes GTP while actin hydrolyzes ATP. This could possibly allow the cell to fine tune and regulate the two systems independently. ID research could thus attempt to determine if such regulation exists.
Secondly, actin and tubulin polymerize in a different way. Actin filaments polymerize simply by adding individual monomers. In the case of tubulin, two slightly different forms of the protein form a dimer and the tubulin-dimer is the basic unit that polymerizes. This dimer is also asymmetric. One form (alpha) binds to GTP but does not hydrolyze it. The other form (beta) is the one which both binds and hydrolyzes ATP. To form a microtubule (MT) filament, you need both beta and alpha in a dimer form. From the perspective of ID, there must be a good reason for this. One possible reason is that not only do alpha-beta dimers link up to form a filament, but they also interact laterally to form a sheet that then folds into a tube. In other words, microtubules are actually much larger than actin filaments, as MTs form a hollow cylinder from 13 distinct tubulin filaments.
These differences might then reflect themselves in the fact that actin and MTs seem to be specialized for different cellular roles. Actin is important in the amoeboid type movement than many cells employ. It typically forms a network just beneath the plasma membrane. MTs, on the other hand, radiate outward from near the center of the cell (from something called a microtubule organizing center). They seem to be intimately associated with the Golgi bodies and endoplasmic reticulum. All flagella and cilia employ MTs and not actin. Perhaps the clearest example of a complementary, rather than redundant role, for the two systems can be seen during cell division. MTs play the crucial role in separating newly formed chromosomes while actin plays the crucial role in splitting one cell into two. In fact, when considered from this angle, the process of cell division depends on both actin/myosin and MTs/kinesin in an irreducibly complex manner.
This is quite interesting because if we are dealing with an irreducibly complex system (described by biochemist Mike Behe), then we have independent reasons to add to our observations of convergence that led to the initial design inference. What's more, we now have reason for thinking the two systems did not evolve independently since both systems are needed for a most basic cellular function. But it gets better than this.
Cell biologists have increasingly come to accept the tensegrity model of cellular architecture. This model suggests that the cytoskeleton is composed of both compression bearing elements that are intimately linked with tension bearing elements to provide the basic structure. A good analogy of such a structure is a tent. To put up a tent, you need both the poles which serves as compression bearing elements and the ropes, which serve as tension bearing elements. In the case of the eukaryotic cytoskeleton, the MTs serve as the compression-bearing elements (which explains why they are much thicker). The actin filaments serve as the contractile elements that generate the tension. Thus, it is not just cell division, but the structure itself which reflects an irreducibly complex state. Although the tensegrity model was not proposed by an ID proponent, it is clearly something that could have been as it would help clear up the problems I mentioned at the beginning. Also, the model originally stems from a scientist (Donald Ingber) whose impetus was to think of the cell in terms of architecture rather than chemical composition.
The eukaryotic cytoskeleton is a true marvel in more ways than one and I have merely scratched the surface. But there are two final things worth mentioning. First, it just underscores just how radically different eukaryotes are from bacteria. The notion that these cell forms shared a common ancestor is quite weak and lacks appreciation for so many of the ways these cells differ. Secondly, this sophisticated state is not something that can be attributed to "billions of years of evolution." On the contrary, if we assume the standard naturalistic views, it arose very early in the history of life and has remained essentially unchanged since. Whatever spawned these cytoskeletal systems got it right from the start.
I first floated the above argument in 2000. At this time, someone responded with the following abstract:
Kull FJ; Vale RD; Fletterick RJ
J Muscle Res Cell Motil, 1998 Nov, 19:8, 877-86
Recent studies have shown surprising structural and functional similarities between the motor domains of kinesin and myosin. Common features have also been described for motor proteins and G proteins. ...Using secondary structure topology, comparison of functional domains and active site chemistry as criteria for relatedness, we propose a set of rules for determining potential evolutionary relationships between proteins showing little or no sequence identity. These rules were used to explore the evolutionary relationship between kinesin and myosin, as well as between motor proteins and other phosphate-loop (P-loop) containing nucleotide-binding proteins. We demonstrate that kinesin and myosin show significant chemical conservations within and outside of the active site, and present an evolutionary development that produce their respective topologies from an ancestral protein. We also show that, when compared with various other P-loop-containing proteins, the cytoskeletal motors are most similar to G proteins with respect to topology and active site chemistry. We conclude that kinesin and myosin and G proteins are directly related via divergent evolution from a common core nucleotide-binding motif, and describe the likely topology of this ancestor. These proteins use similar chemical and physical mechanisms to both sense the state of the nucleotide bound in the active site, and then transmit these changes to protein partners. The different topologies can be accounted for by genetic insertions that add to the edge of a progenitor protein structure and do not disrupt the hydrophobic core.
I think the above abstract makes my point. Here we can see that we have two different proteins whose motor domains show "surprising structural and functional similarities" and "show significant chemical conservation within and outside of the active site," yet also show "little or no sequence identity." In fact, according to one Nature paper, there is "virtually no amino acid sequence identity" between kinesin and myosin.
Yet, the authors conclude that these proteins "are directly related via divergent evolution." Why? If you approach the world with non-teleological filters, similarities are best explained by common descent. Thus, even though the sequence similarities don't exist, structural/functional similarities are sufficient.
But if you approach this topic without making an assumption about whether the data must be explained in accord with non-teleology or teleology, mere similarity (especially at this level) does not lead us to infer a common ancestor. For example, when the authors "describe the likely topology of this ancestor," they have no evidence that such an ancestral protein really existed. They simply assume its existence as the cause for the observed similarity. In fact, the only evidence of its existence is the similarity itself (like arguing in a circle). Yet the very same data could be interpreted as the likely topology that reflects the overall conceptual design schematic tying form to function.
The bottom line here is that we have no compelling reason to attribute these similarities to common descent. That is, without sequence similarities, there is no reason to think divergent evolution is a more persuasive explanation than design. In fact, you'll have to consider me pretty hard core on this point - if there is little or no sequence identity between two proteins, you'll need strong independent evidence for attributing structural/functional similarity to common descent. I accept common ancestry because of sequence analysis, not in spite of it.
In the end, however, we have two alternative explanations. The structural/functional similarities derive either from common descent or design. I tentatively favor the last for two reasons.