t f p g+ YouTube icon

A Student’s Review of Behe’s “Two Binding Site Rule”

Bookmark and Share

November 11, 2010 Tags: Design

Today's entry was written by Kelsey Luoma. Please note the views expressed here are those of the author, not necessarily of The BioLogos Foundation. You can read more about what we believe here.

A Student’s Review of Behe’s “Two Binding Site Rule”

As part of their two and a half month long internship with BioLogos this summer, our three student interns were each asked to read Michael Behe's The Edge of Evolution and reflect on one of the arguments presented by Behe in the book. Today, we present the first of these essays, by Kelsey. Our goal in posting these papers is to show that it is not just experienced and current scientists who take issue with the arguments made against evolution; it is our future scientists as well. We ask that you keep any discussion focused on the book and the arguments presented, rather than directing them towards the author.

In his critique of Darwinian theory, The Edge of Evolution, Michael Behe claims that while random mutation and natural selection do result in genetic change, they cannot account for all of life’s diversity. The book begins with an affirmation of Darwin’s accomplishments and an assertion that all of life does in fact share a common ancestor. After these claims are made, however, it goes on to state that evolutionary theory cannot explain very much of the detail we see in ourselves and in the life forms around us. The author states that “the great process of science has shown that life is enormously elegant and intricate, especially at its molecular foundation” (p. 6). He then compares Darwinian evolution to a “long, blind-folded stroll outdoors” trying to find “the peak of a distant evolutionary mountain” (p. 6-7). Using specific examples and complex statistical calculations, Behe presents a case for a particular strain of thought called Intelligent Design. He concludes that, based on the inability of evolutionary theory to fully explain life, “the elegant, coherent, functional systems upon which life depends are the result of deliberate intelligent design” (p. 166). He asserts that at certain points in evolutionary history some sort of intelligent force stepped in and caused change to occur in a nonrandom way.

While Behe presents his ideas in an articulate and convincing manner, he relies on only a few weakly supported arguments. In fact, many of the arguments he uses are misleading and illogical. In this post, I will isolate one such misleading argument- that “complexes of just three or more different proteins are beyond the edge of evolution”- and present evidence to show that Behe may have been wrong (p. 135).

To develop the argument that protein interactions involving two or more binding sites are beyond the scope of Darwinian Evolution, Behe relies on several shaky pillars of thought. First, he rests on the claim that any significant protein interaction must involve at least five or six amino acid mutations. Next, he assumes that each of these mutations must occur at the same time rather than in a step-by-step manner. Based on these two “facts”, the author concludes that evolution of two protein-protein binding sites involves a miniscule 1 in 1040 chance. Since fewer than 1040 individuals have existed on earth since life began, he claims, such an interaction never could have evolved via random mutation and natural selection alone.

When I began to research the evidence beyond Behe’s text, I found that the assumptions he makes did not match up with the results of other current studies. The statistic upon which Behe relies -- that one protein interaction requires 1020 generations to evolve -- may not be as solid as he would have us believe. Behe sounds convincing within the pages of his book; as the reader becomes absorbed in the text, the argument seems to make sense. When, however, this data is stacked up against the work of other scientists, the truth of the matter becomes quite apparent. A cornucopia of studies and experiments show that the evolution of protein-protein interactions through random events is, in actuality, completely plausible.

Let’s begin by taking a look at the number of amino acids involved in protein-protein interactions. In The Edge of Evolution, Behe claims that in order to get a new binding site, there must be a change in “five or six amino acids in a coherent patch in the right way” (p. 134). This means, the author continues, that “reaching the goal requires five or six coherent mutational steps” (p. 134). A cursory glance at the database of other recent protein-protein interaction studies reveals that this is not always the case. Many other research projects and experiments have shown that oftentimes two proteins may be bound together via only two or three amino acids.

One research article by Neduva and Russel discusses the topic. The authors explain that very short “linear motifs” can form significant protein-protein interactions (Neduva & Russel, 2006). These small “motifs” are “short (3-8 residue)” stretches of “multiple peptides that share a common sequence pattern” (Neduva et al, 2005). Despite their small size, they are known to participate in protein interactions. In this way, they are similar to much larger chunks of proteins, called domains, which also carry out specific functions and bind to other proteins. They are different, however, because they are much smaller and they also arise as a result of single point mutations (Neduva et al, 2005).

Because they are so small, linear motifs have very few amino acids that are actually necessary for them to function. In fact, of the 3-8 amino acids that compose a given motif, generally only 2-3 are necessary for the motif’s role within the protein. Hold on a minute. This point is significant. Behe tells us that, at the very least, 5-6 amino acids are required for protein-protein binding to occur. This is the basis for his establishment of evolution’s “edge”. As we can see, however, studies of linear motifs are telling a different story. Linear motifs need only 2-3 amino acids to function. Many motifs form protein-protein binding interactions. Thus, a protein-protein binding interaction can evolve through mutations in only two or three amino acid molecules (Neduva and Russel, 2006).

Many other studies confirm this data. Linear motifs are not an obscure discovery studied by only a few mad scientists sequestered in a basement. Behe should have known about these results. Another study headed by Neduva, for example, also examined the role of short linear motifs in protein-protein interactions. This study reinforces the conclusion that only 2 or 3 amino acids are necessary to form binding interactions. This time, researchers tested direct protein binding with labeled peptides. Their intent was to find new motifs using genome-scale interaction studies. They did not conduct this study to prove that Behe was wrong, or even to show that proteins can interact via only 2 or 3 amino acids. This is, however, one of the conclusions that can be drawn. The study uncovered linear motifs in flies, yeast cells, humans and other organisms which were involved in protein-protein interactions but only consisted of a few amino acid residues. It was found that five or six amino acids were not necessary for two proteins to interact. Motifs consisting of as few as three amino acid residues could participate in protein interaction networks (Neduva et al, 2005).

Somehow, Behe failed to take note of these studies and the others like it when he wrote his book. While the mere fact that proteins need only evolve two or three amino acid associations for a new interaction to form does not in itself prove or disprove anything, it does begin to chip away at the validity of Behe’s argument.

Finally, in a study by Gruenger, protein sequences were altered and engineered in order to determine the necessary alterations required to form higher order protein complexes. In many cases, changing just one side chain of a protein resulted in a new protein-protein interaction and the development of a new multi-protein complex. Although these mutations were designed by humans, the fact that we can develop a new high-order protein complex simply by changing one side chain shows that Behe’s assumption is egregiously erroneous. One or two, rather than four or five, amino acid changes can produce a new significant protein-protein interaction (Grueninger et al, 2008).

Another of Behe’s central claims is that mutations cannot occur one by one over a long period of time, but rather must all take place at once. This idea is founded on an incomplete understanding of how random mutation and natural selection actually function. As long as each genetic mutation confers some type of survival or reproductive advantage, or at least causes no harm to the cell, changes can occur one at a time and gradually produce a significant alteration. The mutations do not have to occur simultaneously. In fact, isolated mutations occur on a relatively regular basis. According to one estimate, the rate of new mutations per generation in the human genome is about 175 (Nachman and Crowell, 2000). These individual mutations can accumulate in a stepwise, gradual process and result in a new protein-protein interaction.

Let’s take a look at one of many studies which represent a step by step mechanism of mutations. In this case, researchers were interested in how the duplication of gene sequences allows step by step mutations in proteins to occur more easily. In their paper, these researchers remind us that evolution tends to modify already existing genes rather than invent completely new ones. This shortens the process of evolution so that only a few small changes may result in new genes (gene duplication) or in new interactions. Also, when a gene duplicates, one of the copies can freely mutate without inhibiting the cell’s ability to function. It often undergoes “rewiring events” in addition to changes in gene regulation and expression. These phenomena can result in modified protein-protein interactions.

Behe did not take events such as gene duplication into consideration when he made the radical claims in his book. He states, for example, that “three or four” of the “five or six” amino acid changes necessary for a new protein interaction “might cause trouble if they occur singly” (p. 134). This statement fails to account for mutations which occur in gene duplicates rather than in solitary genes. When two copies of a given gene exist, “The genes involved have different probabilities of being retained related to how they were generated” (Robertson, David and Lovell, 2009, paragraph 1). As Behe noted in his book, when a normal gene undergoes a mutation, there is a chance that its function will be destroyed or it will cause harm to the cell. Duplicated genes are different. Mutations in duplicates are not as dangerous because there is an extra gene that codes for the same thing. If one gene copy accrues a mutation that destroys its intended function, the cell will continue to survive because it has another gene that codes for the exact same thing. In this way, the duplicated gene can accumulate mutations more easily than one would expect. (Robertson, David & Lovell, 2009). Behe’s “three or four” amino acids that “might cause trouble if they occur singly”, then, do not apply in the many cases in which duplicated genes mutate and change over time. If Behe had taken special cases like this into consideration, he might have been more careful in his radical claim that a single amino acid change only has a one in 1020 chance of occurring.

Let’s look at another one final example in which Behe’s ideas about protein mutation are contradicted. This time, researchers demonstrated that mutational pathways followed by proteins are not always simple and easy to track. Their study does not agree with Behe’s claims. It shows that many protein mutation pathways aren’t linear. They are so complex, in fact, that the genetic changes could not have all occurred at the same time. The results show that genetic changes take place one step at a time. In some cases, the researchers explain, evolution backtracks itself or follows circular pathways. A mutation may, for example, occur, disappear due to a reversion mutation, and then reappear later on. In fact, as the study shows, reversion trajectories can increase the number of possible evolutionary pathways in a cell by 50%.

The important point to get from this is not that reversion mutations are weird or that they somehow prove evolutionary theory. The fact that protein mutation pathways can go backwards or take other unexpected turns before arriving at an adaptation is the significant part of this story. It shows that mutations occur one at a time and often take place in unexpected order. In the case of reversion trajectories, there is absolutely no way that all of the point mutations involved could occur simultaneously as Behe tells us they must. Rather, one mutation occurs, is selected for by natural selection, and eventually becomes part of a longer story which results in significant cellular change (DePristo, Hartl and Weinrich, 2007).

I’m sure you have heard the Latin phrase “caveat emptor”- “let the buyer beware”. Latin may be a dead language, but these words continue to rattle sporadically in and out of our conversations. Coincidentally, The Edge of Evolution presents us with an ideal opportunity to utter those ageless words. Behe’s claims are issued as truth, and presented with rhetoric and skill, but they may be subject to defect. As I have shown in this paper, two of his central claims directly defy a great deal of scientific data published by other intelligent scientists. Behe claims that mutations must occur simultaneously in order for an interaction to result. Represented in this paper, however, are several reliable studies which show that mutations actually occur one at a time. Behe also states that five or six amino acid mutations must occur to form a protein-protein complex. But, as you remember, other studies that we examined show how short linear motifs can form interactions with as few as three amino acids. “Let the reader beware”, then, of buying into the ideas that Behe has to sell. They may not be as accurate as they appear.

References

Behe, M.J. (2007). The Edge of Evolution. New York: Free Press.

DePristo, M.A., Hartl, D.L., & Weinreich, D.M. (2007). Mutational reversions during adaptive protein evolution . Molecular Biology and Evolution, 24(8).

Grueninger, D., Treiber, N., Ziegler, M.O.P., Koetter, J.W.A., Schulze, M.S., & Schulz, G.E. (2008). Designed protein-protein association. Science, 319(5860)

Nachman, M.W., & Crowell, S.L. (2000). Estimation of the mutation rate per nucleotide in humans. Genetics, 156.

Neduva V, Linding R, Su-Angrand I, Stark A, Masi Fd, et al. (2005) Systematic Discovery of New Recognition Peptides Mediating Protein Interaction Networks. PLoS Biol 3(12): e405

Neduva, V., & Russel, R.B. (2006). Dilimot: discovery of linear motifs in proteins. Nucleic Acids Research, (34).

Neduva , V., & Russel, R.B. (2005). Linear motifs: evolutionary interaction switches. Science Direct, 579(15).

Robertson, David L., & Lovell, Simon C. (2009). Evolution in protein interaction networks: coevolution, rewiring and the role of duplication . Biochemical Society Interactions, 37.


Kelsey Luoma is a graduate of Point Loma Nazarene University in San Diego, California, where she received a bachelor's degree in biology. She plans to continue her education in medical school. As an evangelical Christian and student of biology, Luoma is very interested in resolving the conflict between faith and science. She has spent two summers working as a student intern for BioLogos. In the future, she hopes to serve internationally as a physician.


View the archived discussion of this post

This article is now closed for new comments. The archived comments are shown below.

Loading...
Page 2 of 3   « 1 2 3 »
Jason Hine - #39561

November 12th 2010

You make good points clearly, Kelsey.  I would encourage you to continue to refine your techniques for excellent peer review by keeping the focus of your review on the peer’s work.  This, in my estimation, is a key measure of maturity for any scholar, whether scientist or theologian: To respect the individual, even while questioning that individual’s work.  By doing so, you help preserve the open and constructive dialogs that can help move the enterprises of good science (and good theology) forward. 

Kudos and best wishes for your future endeavors!


hmm - #39573

November 12th 2010

John:

“So, Bilbo and hmm and pds, are you going to state that the thousands of papers and thousands of scientists who did the work identifying these motifs are wrong, and Behe is right in claiming that 5-6 residues are required?”

I’ve not stated my own thoughts. The only thing I’ve said about Behe in this discussion has been the question, has Behe really argued, that mutations “must” occur simultaneously. I was not so sure. Nobody replied.

My personal belief is that nature is not blind. For me the questions like, what can happen without guidance, are not personally very relevant - because I think that everything is somehow guided in our world by God and maybe orher forces. How do you see God at work in our world? Do you see Him effecting mutations? Do you see him guiding evolution?


John - #39578

November 12th 2010

hmm:
“I’ve not stated my own thoughts.”

That’s why I was asking, of course. Do binding sites require 5 residues or not? Kelsey has provided quite a bit of documentation showing that Behe is wrong.

“The only thing I’ve said about Behe in this discussion has been the question, has Behe really argued, that mutations “must” occur simultaneously. I was not so sure. Nobody replied.”

Behe wants his readers to think that without having to explicitly state a blatant falsehood. In the case of chloroquine, does the figure of 10^-20 have anything to do with the mutations actually happening?

Should I take your failure to reply to Kelsey’s point about the number of residues in binding sites as agreement with Kelsey?

“My personal belief is that nature is not blind.”

So why defend Behe’s misrepresentations of the facts of nature?

For me the questions like, what can happen without guidance, are not personally very relevant - because I think that everything is somehow guided in our world by God and maybe orher forces. How do you see God at work in our world?

It’s a mystery to me. I sure don’t see God working through Behe, nor does Kelsey!


hmm - #39580

November 12th 2010

John wrote:

“I sure don’t see God working through Behe, nor does Kelsey!”

I’ve not ever said so about anyone. How can you claim, that God is not working through someone?

IS THIS KIND OF ARGUMENTATION PERMITTED IN BIOLOGOS???


Bilbo - #39581

November 12th 2010

Behe’s endnote [13], explaining why he thought 5-6 amino acid residues would be needed:

Even though protein-binding sites often involve a score of amino acids on each of the partners, experiments have shown that only a fraction of those are important for having the two proteins stick to each other. (For example, see Braden, B.C., and Poljak, R.J. 1995. Structural features of the reactions between antibodies and protein antigens. FASEB J.  9:9-16: Lo Conte, L., Chothia, C., and Janin, J. 1999. The atomic structure of protein-protein recognition sites. J. Mol. Biol.  285:2177-98; Ma, B., Elkayam, T., Wolfson, H., and Nussinov, R. 2003. Protein-protein interactions: Structurally conserved residues distinguish between binding sites and exposed protein surfaces. Proc. Natl. Acad. Sci. USA 100:5772-77.) In terms of the swimming pool analogy, the five or six residues represent bumps and magnets that are aligned very nicely; if enough are aligned, then it doesn’t matter so much if other features aren’t aligned, as long as they don’t block the surfaces from coming together.


John - #39681

November 13th 2010

Bilbo, do any of the papers Behe cited address the minimum number of specific residues, as the ones cited by Kelsey did?


John - #39683

November 13th 2010

…that should be, as the DATA from the ones cited by Kelsey did?

Because, as Kelsey also points out, Behe’s assumptions have been falsified by scientists in the process of doing other things, not addressing Behe.


Bilbo - #39693

November 13th 2010

Hi John, 

I may be misunderstanding this chart from this paper: Ma, B., Elkayam, T., Wolfson, H., and Nussinov, R. 2003. Protein-protein interactions: Structurally conserved residues distinguish between binding sites and exposed protein surfaces. Proc. Natl. Acad. Sci., but it seems to show a ratio between the number of contact residues between proteins and the number of conserved residues, which I take to be the number required in order for there to be a binding site.  It seems to work out to about 20%, which would mean about 5-6 residues per 20.


http://www.ncbi.nlm.nih.gov/core/lw/2.0/html/tileshop_pmc/tileshop_pmc_inline.html?title=An external file that holds a picture, illustration, etc.
Object name is pq1030237004.jpg [Object name is pq1030237004.jpg]&p=PMC3&id=156276_pq1030237004.jpg


Bilbo - #39701

November 13th 2010

This comment was deleted again, without any explanation as to what was wrong with it:

“Bilbo - #39686

November 13th 2010
I’ll try again, John.  This time I’ll copy my comment and save it, so that I can prove that I was being polite.
You wrote:  “You did no evaluation. You assumed that Behe’s false claims were correct.”
The point of my evaluation is that even if Behe’s claims are false, I don’t think there are many, if any, reasonable biologists who think that there would be two functionless proteins waiting around for a third protein to evolve that gives a function to all three.  I may be mistaken, but I think this would apply to linear motifs as well.
Let’s see how long this comment lasts.”

Could the moderator please explain why it was deleted.


John - #39705

November 13th 2010

Hi Bilbo,

Again, there’s absolutely zero relevance to Behe’s claim regarding the minimum number of residues. In fact, if you read the first sentence of the Methods, you’ll see that they explicitly chose much, much larger interfaces:

“The initial dataset of interfaces with 1,629 two-chain interface entries in the Protein Data Base…”

Now, what does “two-chain interface” mean? How can Behe possibly cite this as evidence for *minimum* residues specifying a binding site when it is explicitly limited to two-chain interfaces? How can Behe be a competent biochemist if he is unaware of the hundreds (more likely thousands) of papers describing short linear motifs that are both necessary (shown by changing one of the conserved residues) and sufficient (synthetic peptides) for protein-protein binding?

And how can you, Bilbo, make any credible claim about what Behe was thinking when he wrote this?

Why not look for the truth first, as Kelsey did, before tying yourself in knots to cling to the idea that Behe is looking for the truth?


John - #39706

November 13th 2010

Bilbo:
“The point of my evaluation is that even if Behe’s claims are false, I don’t think there are many, if any, reasonable biologists who think that there would be two functionless proteins waiting around for a third protein to evolve that gives a function to all three.”

Behe is using a straw man fallacy. Co-option is the norm. Nice try at moving the goalposts, though. Behe refuses to test his hypothesis that IC structures cannot evolve, so there’s no reason for any of us to treat it as valid.

If a biochemist’s claims about basic biochemistry are false, why would you trust his claims about genetics and evolution, a subject in which he has no training or record of empirical scientific contribution?


Bilbo - #39708

November 13th 2010

Hi John,

If I understood the chart, the ratio of conserved residues to total contact residues pretty much confirms Behe’s point.  But perhaps I don’t understand it.  However, your comments haven’t enlightened me as to why I don’t (which may not be your fault.  Just my ignorance.).

Yes, short linear motifs would require fewer residues.  However, from the information I have been able to glean, it’s not clear how many slms there are, though the number found is in the low hundreds, nor whether they could have accounted for the many thousands of larger binding-sites found in proteins. 

You wrote:  Nice try at moving the goalposts,  

I wasn’t moving the goalposts.  Behe’s argument concludes that the edge of evolution is at three proteins, by which he meant that if three proteins are needed before there is function, it wouldn’t happen via random mutation.  That seems like a reasonable conclusion to me, regardless of the truth of the premises of his argument.


John - #39717

November 13th 2010

Bilbo,

Forget the chart. They started by only considering two-chain interfaces.

THEY EXCLUDED SHORT LINEAR MOTIFS. Short linear motifs are not components of two-chain interfaces BY DEFINITION.

“However, from the information I have been able to glean, it’s not clear how many slms there are…”

You’re moving the goalposts again.

It’s not relevant to Behe’s claim, which is how many residues are need to get the association started. Once you have binding, it’s easy to expand the binding site to two whole chains if there’s selection for it over time. The hurdle Behe claims simply doesn’t exist.

“I wasn’t moving the goalposts.”

That’s exactly what you continue to do!


John - #39718

November 13th 2010

Bilbo:
“Behe’s argument concludes that the edge of evolution is at three proteins,…”

But as Kelsey has conclusively shown, at least one of the major premises of Behe’s argument is false, rendering Behe’s conclusion false by many orders of magnitude.

“... by which he meant that if three proteins are needed before there is function, it wouldn’t happen via random mutation.”

It happens by mutation (that is ONLY random wrt fitness) AND NATURAL SELECTION.

“That seems like a reasonable conclusion to me, regardless of the truth of the premises of his argument.”

What an incredible admission! So that part of Behe’s book is therefore meaningless filler because its veracity is completely irrelevant?

What other premises don’t have to be true, Bilbo?


Steve Matheson - #39752

November 14th 2010

@ hmm - #39573
You write: “I’ve not stated my own thoughts. The only thing I’ve said about Behe in this discussion has been the question, has Behe really argued, that mutations “must” occur simultaneously. I was not so sure. Nobody replied.”

I have explained why Behe does in fact argue for simultaneous occurrence of the mutations. It is revealed in the way he did his ludicrous calculations.

His work in EoE is folk science, and cannot be taken seriously by anyone knowledgeable about the basics of evolutionary/population genetics.


John - #39770

November 14th 2010

Steve wrote:
“I have explained why Behe does in fact argue for simultaneous occurrence of the mutations. It is revealed in the way he did his ludicrous calculations.”

Yes, but for Behe’s defenders, it’s only about words. Data and math don’t matter, unless they can be used to advantage. Look at how Bilbo abandoned Behe’s need for the premise he thought was so important just yesterday, when he thought he could defend it by attacking Kelsey!


Bilbo - #39848

November 14th 2010

John - #39770:  “Look at how Bilbo abandoned Behe’s need for the premise he thought was so important just yesterday, when he thought he could defend it by attacking Kelsey!

Actually I “abandoned Behe’s need for the premise” a while ago:

http://telicthoughts.com/the-real-issue-of-behes-edge/

It’s not relevant to Behe’s claim, which is how many residues are need to get the association started. Once you have binding, it’s easy to expand the binding site to two whole chains if there’s selection for it over time.

If that is in fact how most protein binding sites evolved, then we might have a good argument against Behe.  Though, as I stated before, I think we might still have a problem.  Suppose, for example, that we need three short linear motifs before there is function.  What is the probability that such a thing would happen?  I suspect there is still a huge hurdle to cross.

It happens by mutation (that is ONLY random wrt fitness) AND NATURAL SELECTION.

Yes, but if three proteins must evolve before there is selective pressure, then it seems that we can only rely upon random mutation.


Cist - #39859

November 14th 2010

Thats whats called a double FAIL for Biologos.


John - #39862

November 14th 2010

Bilbo wrote:
“Actually I “abandoned Behe’s need for the premise” a while ago:”

So why did you need to attack Kelsey about it?

“Suppose, for example, that we need three short linear motifs before there is function.”

That’s a ridiculous supposition, Bilbo. We already know that single motifs are sufficient in many, many cases, all of which Behe ignored because they would have altered his edge by many orders of magnitude.

“What is the probability that such a thing would happen?  I suspect there is still a huge hurdle to cross.”

You really ought to read the papers cited by Kelsey and the papers cited in those papers. Your position, and Behe’s position, is preposterous.


John - #39864

November 14th 2010

Bilbo:
“Yes, but if three proteins must evolve before there is selective pressure, then it seems that we can only rely upon random mutation.”

Why would you think that would be the case, based on what real scientists have actually measured in real systems instead of the musings of former scientist Behe, who doesn’t bother to read the primary scientific literature on the very subjects he writes about? How do you explain Behe’s false assumption on HIV Vpu, when he even included Vpu in a figure in his book? Did that function involve the evolution of three proteins?

And why do you use the silly term “random mutation,” when mutations are only random wrt fitness, and not direction, location, etc.?

Why wouldn’t selection just act on existing polymorphisms in a population? You aren’t denying the existence of a vast amount of heritable variation that exists right now (regardless of how you think it came to be), are you? Because it’s clear that Behe is doing precisely that to fool his audience.


Page 2 of 3   « 1 2 3 »