The hypothesis of Common Design is actually a creationist concept, where similar structures or features are explained in terms of a designer creating similar functions. Biologist Edward Max explains it as follows:

Thus the similarities between species in anatomy and protein structure can be interpreted in two entirely different ways. The evolutionists say that the similarity between features of, for example, humans and apes reflects the fact that these features were inherited from a common ancestor; that is, the similar features of humans and apes are determined by modern copies of genes that once existed in species that was ancestral to both apes and humans. The creationists say that apes and humans were created independently but were designed with similar features so that they would function similarly. Both the gene copying and the independent creation views seem consistent with the similarity data, but which view is correct? [1]

Max successfully argues that we can tease apart the explanations of common design from common descent by looking for similarities that are not attributed to function. Since designers design to elicit a function, similarities not tied to function are unlikely to stem from design. Max then uses this logic to make a powerful case for evolution. But what if we fail to find similarities that are not tied to function?

Consider two impressive molecular machines. The first machine is the bacterial flagellum. The flagellum is composed of about 20 different protein parts and functions to convert a current of ions into rotary motion in order to propel the cell through a liquid medium. The other machine is known as the F-ATP synthase. It is composed of eight different protein parts and it uses the energy supplied by a proton current to synthesize ATP molecules. Remarkably, the machine can run in reverse, where ATP is broken down and the subsequent energy that is released can be used to create a gradient of protons. For our purposes, what is relevant is that the two machines appear to share at least one part. In the F-ATP synthase, it is called the beta subunit. In the flagellum, it is called FliI. When you compare these two proteins they share the same amino acids at approximately 30% of the positions. What is more, both beta and FliI can specifically bind ATP and break it down. Further analysis shows the two proteins to share a common ATP-binding domain, which accounts for a large percent of the similar amino acids. The two proteins are too similar to be explained by chance and we don’t know of any law that would account for their independent origin. Thus, scientists interpret the similarities as evidence of homology, where both are derived from a common ancestral state. This is the consensus view. But does the similarity between the two components really reflect common descent or might it reflect common design?

I am reluctant to advance the old thesis of common design because of its nasty ad hoc flavor. Nevertheless, we are exploring the world through the Design Matrix and I simply cannot subscribe to the notion that a designer would always reinvent the wheel every time a machine is invented. The reuse of components modified for different contexts would not be surprising. It fact, we might expect it. So perhaps the similarities between the beta subunit of the F-ATP synthase and the flagellum’s FliI are indeed examples of common design. After all, while I mentioned the flagellum was a rotary motor, I neglected to mention the same was true of the F-ATP synthase. The beta subunit is actually part of a protein ring that caps a stem-like structure making the whole thing look like a mushroom. To visualize it working, simply imagine a flow of current traveling through the stem of the mushroom such that the stem rotates within the cap, allowing the cap to synthesize ATP. In addition, both the flagellum and the ATP synthase are embedded in the membrane, where current passes through the channels created by the machine components to do the work. It is not difficult to believe that common design explains two membrane-embedded rotary motors using a similar component to extract energy. But with a closer look, things begin to get shaky.

The problem lies in the functional comparison of FliI and the beta subunit. The beta subunit’s ATP-related functions are intimately coupled to the rotary motion of other machine components. But FliI plays a role that is not directly associated with rotary motion in the flagellum. Instead, it breaks down ATP to provide the energy to shoot the flagellar parts across the membrane as part of a construction scheme. In other words, FliI uses ATP to build the flagellum, while the beta subunit can use ATP to create a current of ions across a membrane. Furthermore, it was long thought that FliI functioned as a monomer, where one copy of FliI functions with other proteins to secrete flagellar proteins. In contrast, the beta subunit functions as part of a six-member ring composed of three copies of beta and three copies of a very similar protein known as the alpha subunit. Given the functional differences between the two, why do they still share so many amino acids? Furthermore, given that there is more than one way to bind ATP, why do they use the same binding domain? Given the similarities not explained by function, the explanatory balance is tipped towards common descent.

Science is like perpetual childhood, where there is always something new to discover and wonder about. While it is true for many years that scientists thought FliI functioned as a single copy, more sophisticated experimental work showed that it too forms a six-member ring, using six copies of FliI [2]. Suddenly, the theme of common design resurfaces. Might the sequence and domain similarity be tied to the share structure of a ring? The FliI ring sits at the very base of the flagellum, where it acts as the intermediary between the cytoplasm and the embedded protein secretion apparatus, recruiting flagellar proteins. Since many flagellar proteins are probably unfolded as they enter the secretion apparatus, one can envision FliI snagging and threading the amino acids chains of flagellar proteins into the flagellum while it is under construction [3]. The explanatory balances now levels off, where it appears that both common descent and common design are plausible explanations.

Adding to the ambiguity when it comes to choosing between common descent or common design at the molecular level are the rather limited ways in which to build a functional protein. The theme of modularity repeats itself like a fractal image throughout biology and proteins are no exception. Proteins are typically constructed of smaller, distinct folded regions called domains or modules. Consider plasminogen activator, a protein that activates another protein in order to break down blood clots. Four different domains are used to build the plasminogen activator: one fibronectin type III domain, one epidermal growth factor domain, one kringle domain and one serine protease domain [4]. Each of these domains plays a functional role by itself. For example, the fibronectin type III domain is a generic binding domain very useful for scaffolding purposes [5]. As a result, you often find the fibronectin domain in a variety of different proteins in all different branches of life (front-loaded?) mediating interactions between the protein and other molecules [6]. The serine protease domain is the business end of the protein, as it is the enzymatic part that activates another protein. Since the vast majority of proteins are built from multiple domains, just how many domains are there?

Michael Denton, Senior Research Fellow in the Department of Biochemistry at the University of Otago in New Zealand, surveyed the literature and proposed that protein domains (or folds) are a function of natural law and not contingency. Denton and colleagues discuss how scientists have determined the forms of various protein and note:

It also became apparent that the 3D structures of individual folds were essentially invariant, some such as the Globin fold and the Rossman fold for example, having remained essentially unchanged for thousands of millions of years. Both their invariance and the typological classification schemes into which they could be grouped argued for their being a finite set of ‘‘real timeless structures’’ determined by physics rather than being mutable ‘‘Lego-like’’ aggregates of amino acids determined by selection. [7]

And also add:

Consideration of the various physical constraints which restrict the folded spatial arrangements of linear polymers of amino acids, the laws of fold form, suggests that the total number of permissible folds is bound to be restricted to a very small number. One recent estimate based on possible arrangements of typical structural elements gave a maximum of 4000 folds. Based on similar considerations, the authors of another recent paper suggested that the maximum is likely to be no more than a few thousand. A different type of estimate based on the rate of discovery of new folds, rather than permissible spatial arrangements, suggests that the total number of folds utilized by organisms on earth might not be more than 1000. In many recent reports the total number of different folds is often cited to be somewhat less than 1000. [7]

Two important themes emerge from Denton’s argument. If functional protein folds are determined by natural law, this is yet another tool that could be exploited by a front-loading designer. Having a catalog of the folds and knowledge of what type of mutations could crystallize the de novo appearance of various folds, a front-loading designer could structure the original components of the designed cells such that they again served as bait for the emergence of certain folds. A wonderful illustration of such a possibility concerns a rather simple protein known as the arc repressor, where experiments showed that that “new protein folds can evolve from existing folds without drastic or large-scale mutagenesis” [8]. On the other hand, a front-loading designer may simply have chosen folds that are most adaptable. Knowing that evolution would find it easier to borrow and reuse the folds it was given, rather than stumble upon novel and/or rare folds, the designer again has yet another line with which to reach into the future.

The second theme concerns the blurring between common design and common descent. If there are only a few thousand protein folds, then our degree of confidence about homology is greatly weakened if the main pillar of this inference is based on structural similarity. That is, if we were dealing with a nearly-infinite number of potential protein folds, then the fact that two proteins share folds would be strongly suggestive of common descent. But if the number of structures is quite limited, then an origin through convergence, or common design, is equally plausible. This is quite significant in our post-genomic age, given that biologists seem to be relying more and more on structural similarity to infer ‘homology.’

For example, let’s assume there are only about 1000 different protein folds. Let’s now imagine that an intelligent designer sought to seed this planet with microbial life forms containing proteins whose average number of domains was three (typical of modern proteins). This would mean that our designer could only design about 300-350 proteins before having to reuse a fold. Yet E. coli alone has over 5000 genes that code for proteins! Since we would not be able to design a heterogeneous pool of bacteria, let alone a single type of bacteria, from such a limited number of proteins, design had to reuse protein folds. Different proteins sharing the same fold may look evolved, as if the shared folds reflected common descent at some point in the past, yet the sharing is to be expected from the design of life.

1. Max, EE. 1986-2003. Plagiarized Errors and Molecular Genetics: Another argument in the evolution-creation controversy

2. Claret L, Calder SR, Higgins M, Hughes C. 2003. Oligomerization and activation of the FliI ATPase central to bacterial flagellum assembly. Mol Microbiol. 48:1349-55.

3. I explore this is a little more depth here.

4. Ny T, Elgh F, Lund B. 1984. The structure of the human tissue-type plasminogen activator gene: correlation of intron and exon structures to functional and structural domains. Proc Natl Acad Sci U S A. 81(17):5355-9.

5. Koide A, Bailey CW, Huang X, Koide S. 1998. The fibronectin type III domain as a scaffold for novel binding proteins. J Mol Biol. 284(4):1141-51.

6. Here .

7. Michael J. Denton, Craig J. Marshall and Michael Legge. 2002. The Protein Folds as Platonic Forms: New Support for the Pre-Darwinian Conception of Evolution by Natural Law. J. Theor. Biol. 219, 325–342.

8. Cordes MH, Walsh NP, McKnight CJ, Sauer RT. 1999. Evolution of a protein fold in vitro. Science. 284(5412):325-8.