Immune System Evolution #1: Comments on Annotated Bibliography

This post contains my commentary on the Annotated Bibliography on the Evolution of the Immune System, now online in the NCSE Kitzmiller archive. The Annotated Bibliography describes the significance of each publication listed in the Supplementary Material for the recent Nature Immunology article on the “immune system cross” during Behe’s testimony in Kitzmiller v. Dover. For article, click here. For the full Annotated Bibliography, click here. The Annotated Bibliography has reached an approximately final state, but I would still be interested in any additional comments people might have. My overall point with all this, of course, is that unless and until ID proponents (1) acknowledge the existence of this scientific literature, (2) admit that their previous statements about the nonexistence of this literature were wrong, and (3) substantively rebut this literature, providing a better and more detailed explanation for the immune system, then they aren’t even beginning to be scientifically serious.

As you read through this, keep in mind the Discovery Institute’s hiliarious commentary in their recent book attempting to rebut to the Kitzmiller decision, Traipsing Into Evolution:

Consider [Judge Jones’s] skewed summary of the evidence relating to the irreducible complexity of the immune system. He cited Kenneth Miller’s speculative assertions as if they were facts, while refusing even to mention biochemist Michael Behe’s detailed rebuttal during the trial. (Traipsing, p. 45, italics added)

Speculative? I guess in ID-Land, dozens of publications in top journals confirming key expectations is “speculation”, whereas the vague statement that divine intervention occurred sometime, somewhere, for unspecified reasons is considered rock-solid. Even better, Traipsing then quotes from Behe’s “detailed response” to the immune system section of Jones’s opinion. However, the book neglects to point out that Behe tried exactly the same silly arguments in his direct testimony at trial, and they were specifically debunked on cross-examination. For more on Behe’s “detailed rebuttal”, see here.


The purpose of this annotated bibliography is to give the nonspecialist reader some sense of the scholarly weight of the evolutionary immunology literature (here is another method). In particular, this bibliography focuses attention on the scientific work that developed, tested, and established the “transposon hypothesis” for the origin of receptor rearrangement in the adaptive immune system. Contrary to the impression one might get from the ID literature, the transposon hypothesis was not idle speculation, it was not thought up yesterday over breakfast, it is not particularly vague, it is not untestable, and it is not scientifically useless. In contrast, by reading through the bibliography in chronological order, we can see that the transposon hypothesis was explicitly proposed and published in top journals, it was carefully and seriously discussed for decades in the professional literature, it has inspired a very productive research program (both experimental and comparative), it has been tested with diverse evidence by researchers working in many different labs, and it has been dramatically confirmed.

In short, it is a classic case of serious, advancing evolutionary science. Without ever getting mentioned in a newspaper article or cable news show, hundreds of PhD scientists have devoted their careers to working out how and why the immune system evolved. Any one of these researchers has quietly produced more research results than the entirety of the intelligent design movement with their collection of op-eds, webpages, in-house publications, books published by InterVarsity Press, one or two books by slightly more rigorous publishers but with dubious peer-review, and one or two review articles slipped into obscure journals. Unlike “intelligent design” proponents, evolutionary immunologists repeatedly published their key work in top journals. For some reason, they have not felt the need to write law review articles, lobby school boards and legislatures to get their view taught, or muck around with state science standards, because they know that scientific hypotheses succeed by convincing other scientists through research.

For more along these lines, see Bottaro et al. (2006) and the Kitzmiller decision.

Why the Immune System?

The decision to make the evolution of the immune system a key example in both Kenneth Miller’s direct testimony and Michael Behe’s cross-examination was based on several factors. First, among the major systems that Behe discusses in Darwin’s Black Box, such as the eukaryotic cilium, the bacterial flagellum, the blood-clotting cascade, and the immune system, Behe makes particularly florid claims about the absence of literature on the origin of the immune system – for example:

“As scientists we yearn to understand how this magnificent mechanism came to be, but the complexity of the system dooms all Darwinian explanations to frustruation.” (Darwin’s Black Box, p. 139)

“We can look high or we can look low, in books or in journals, but the result is the same. The scientific literature has no answers to the question of the origin of the immune system.” (Darwin’s Black Box, p. 138)

These are claims crying out for refutation. As it happens, there is probably more scientific literature directly on the evolution of the immune system than on the other three systems put together (we can chalk this up to (1) the long tradition of evolutionary and comparative studies in immunology; (2) the age of the discipline (going back to the 1800’s); and (3) the massive amount of medical research money available for studies of the immune system, for obvious reasons). So, while Behe makes various mistakes on the other systems, many also discussed at trial, he is most dramatically wrong about the immune system literature.

Second, although the immune system is very complex, everyone is familiar with (a) its importance, (b) vaccinations and immunity to disease, and (c) many people have heard of antibodies. So it is actually not so difficult for a nonspecialist, such as a lawyer or judge, to realize the importance of immune system research.

Third, those in the community of “creationism watchers” knew that the ID advocates were particularly vulnerable on the immune system, given their past flailings in response to internet challenges. Some of this is mentioned in Bottaro et al. (referencing these Panda’s Thumb posts: [1][2]) and Michael Behe’s weak response. For an earlier episode often recalled by creationism watchers see this celebrated discussion on an ID bulletin board, where several pro-evolution posters challenged several leaders of the ID movement to admit that Behe’s 1996 claim was wrong.

Fourth, Behe’s dismissal of several famous articles on evolutionary immunology during his deposition also increased confidence that he was not at all familiar with the literature, and instead would brush off any challenges as not meeting his requirement for infinite detail.

Fifth, plaintiffs’ attorney Eric Rothschild had the strength of will to continue with the immune system gambit, despite Nick Matzke’s early attempt at explaining the evolution of V(D)J recombination to him on a whiteboard:

The beginnings of the immune system “battle plan” for the Behe cross-examination. Whiteboard drawings from May 2005, NCSE office. Click to see larger image.

Scientific Background

The annotated bibliography will not make much sense without a little background understanding of immunology. If you can understand this basic terminology, you should be well on your way to understanding the transposon hypothesis:

  • Adaptive immune system (AIS): The portion of the immune system that generates receptors that specifically recognize and bind to a particular pathogen. This system has “memory”, so that if a pathogen returns, it will be recognized much more quickly the second time around. Vaccines expose a person to a killed or weakened form of a pathogen, or key pieces of a pathogen. This “trains” the immune system to recognize common pathogens such as smallpox or polio, so that immunity can be developed without the person having to catch and fight off the actual disease. The adaptive immune response is restricted to jawed vertebrates (i.e., all vertebrates except lampreys and hagfish), and therefore evolved 450-500 million years ago in the common ancestor of cartilagenous fish (sharks and rays) and other “fish” (bony fish, including tetrapods). (Note: there is increasing evidence of different forms of adaptive responses in other organisms.)
  • Innate immune system: The portion of the immune system that uses general, nonspecific responses to fight pathogens. For example, a receptor molecule called TLR5 recognizes flagellin, the major component of bacteria flagella and a common conserved feature. The innate immune system is the first line of defense, and many of its features are shared between vertebrates and invertebrates
  • Immunoglobulins (Igs), also called antibodies: These are the receptor proteins of the immune system. They have a “Y” shape, where the top of the “Y” recognizes foreign molecules called antigens. The genes that encode immunoglobulins are composed of four basic kinds of segments: V (variable), D (diversity), J (joining), and C (constant). Vertebrates have hundreds of different copies of V, D, and J gene segments. Through a process called V(D)J recombination, different copies of these segments are spliced together, producing billions of different unique immunoglobulin receptors. Most of these receptors will not recognize a particular pathogen, but a few of them will, even if it is a completely novel pathogen.
  • Recombination Activating Gene (RAG): A RAG is a gene that codes for a RAG protein (the names of genes are italicized, and the corresponding proteins are not; sometimes PDF and HTML documents will be missing this formatting however). RAG proteins have the ability to recognize specific DNA sequences at two locations called Recombination Signal Sequences (RSSs), bring the pieces together, and cut the DNA at the RSS sites. DNA repair enzymes then repair the DNA and join the two segments, originally distant from each other, together. In the vertebrate immune system, two RAGs, RAG-1 and RAG-2, cooperate in this process.
  • Transposon: A transposon, or transposable genetic element, is a “jumping gene.” The transposon is usually a gene that codes for a protein. This protein, in turn, recognizes certain signal sequences (sound familiar?) in the transposon DNA allowing it to snip out the transposon DNA and transplant it someplace else. There are many different types of transposons with various degrees of relationship to each other, and they are found in all groups of organisms. Mutations can “break” transposons, resulting in “fossil” transposons that are thought to usually be “junk DNA” (although it may play some non-coding, structural role in cells). About 3% of the human genome is composed of the remains of DNA transposons, and about half of the genome is made of various other kinds of copied repetitive DNA.

With that terminology in hand, we can discuss the transposon hypothesis. The transposon hypothesis is based on similarities between transposons and RAGs, and suggests that RAG-1 and RAG-2, and the RSSs, are actually descended from “free-living” transposons. The model states that in an early vertebrate, already equipped with an innate immune system, a transposon inserted into a non-rearranging receptor gene. When this receptor gene was expressed, the transposon was activated, and snipped itself out of the receptor. However, because this excision process was inexact, the resulting receptor would be variable, and therefore some copies of the receptor protein would be better able to recognize new or mutated pathogens. Natural selection would therefore spread this variant. An extended process of gene duplication and diversification would elaborate this basic system into the modern V(D)J recombination system.

The transposon hypothesis suggests a number of specific observations that would corroborate the model if found:

  • sequence similarities between V(D)J RSSs and transposon recognition sequences
  • RAGs should operate by mechanisms similar to transposase mechanisms
  • RAGs might still be able to perform the same functions that transposons perform, such as DNA excision and insertion
  • immunoglobulins should have relatives that are non-rearranging receptors
  • if RAGs are descended from transposons, then “free-living” transposon relatives of RAG might still exist “in the wild”

All of these observations have been published in the last 10 years. The crowning achievements, and the easiest successes to understand, were the identification of transposon relatives of RAG-1 in sea urchins, lancelets, and cnidarians (Kapitonov and Jurka, 2005) and a RAG1-RAG2 homolog serving a non-immune function in sea urchins (Fugmann et al. 2006).

If the above, very short, summary did not completely sink in, other resources are available. For an excellent moderate-length introduction to evolutionary immunology, please see Matt Inlay’s online article, “Evolving Immunity.” Pay particular attention to the graphics. Next, read our Nature Immunology essay for a brief update on the scientific progress of the transposon hypothesis. See also these Panda’s Thumb blogposts by Matt Inlay or Andrea Bottaro: “New discovery of missing link between adaptive immune system and transposons,” “The Revenge of Calvin and Hobbes,” and “Behe’s meaningless complexity.” Once you have given those sources a try, take several minutes to work through Figure 1 and Figure 2 of Lewis and Wu’s (2000) commentary article, “The Old and the Restless”, published in the Journal of Experimental Medicine. Figure 1 is a model of how receptor rearrangement works in modern organisms. Figure 2 is a model of how it evolved.

Hopefully, after you have worked through this material you can re-read the above summary and make some headway.

How the Bibliography was Assembled

This bibliography was originally developed for the Behe cross-examination in the Kitzmiller case, discussed in Bottaro et al. (2006). The annotations are based on notes that were taken while the bibliography was being assembled (in an all-night session the weekend before Behe testified, it must be said). The annotations have subsequently been revised and updated.

While assembling the exhibit, it was very important to make sure it was not vulnerable to some of the critiques that can be leveled against lists of articles that are uncomprehendingly culled from automated literature searches. When one searches online databases for publications that involve the keywords “immune system” and “evolution” – two immense research topics – one will get results on a wide variety of topics. These include:

  1. the evolution of disease organisms
  2. the evolution of modern populations in response to diseases
  3. the evolution of immune system cells when an organism’s adaptive immune response is triggered (such as when someone is vaccinated)
  4. the use of customized antibodies for innumerable biomedical purposes
  5. the “evolution” of chemical reactions or biochemical reactions (this is a nonbiological definition of evolution, for example, the “evolution” of hydrogen gas during hydrolysis)
  6. clinical research on AIDS, allergies, etc.
  7. the evolution of the immune system within closely-related organisms – e.g., mammals, birds, or tetrapods
  8. “deep” comparative immunology between vertebrates and invertebrates
  9. the selective forces driving immune system evolution and receptor diversity
  10. the origin of the innate immune system
  11. the origin of the adaptive immune system

Only the last four topics directly address the evolutionary origin of the immune system, although many of the other topics can have some relevance. Among these four topics, this bibliography focused mostly on topic #11, and included a smattering of work on topics #8, 9, and 10. Within topic #11, the origin of the V(D)J recombination system was emphasized, as it is widely seen as the “key” system, the most remarkable feature of adaptive immunity, and the biggest evolutionary “puzzle” according to ID advocates.

This bibliography also focused on the review literature rather than the research literature. First, the research literature is mostly opaque to nonspecialists, and the implications of findings are made much clearer in the review literature. Second, at his deposition, Michael Behe expressed the opinion that if scientists had made any progress worth talking about in evolutionary immunology, he would have expected to see it in the review literature. A few of the most famous research articles were included, but most of the articles are articles reviewing the research literature – many of the review articles cite several hundred other articles.

This literature collection should not be considered comprehensive or complete – it is merely a sample of the literature on the evolution of the adaptive immune system, focusing on V(D)J recombination, with a small sampling of literature on the evolution of innate immunity, MHC, large-scale comparative immunology, etc. For some idea of how much more scientific literature there is on this topic, see the longer unannotated bibliography.

The Larger Context

Matt Inlay (author of “Evolving Immunity” and coauthor of Bottaro et al. 2006) points out that Michael Behe’s book Darwin’s Black Box was ironically timed. In 1996, the year Darwin’s Black Box was was published, research on V(D)J recombination was in the midst of a dramatic upswing:

Research progress on V(D)J recombination over the last 25 years. Results of a PubMed search on the term “V(D)J recombination”, plotted as a function of time (blue dots). Numbers of articles per year are shown on the y-axis, and the years of publication on the x-axis. Publications for 2006 were projected based on Jan-Apr numbers. Major breakthroughs in the transposon model are listed in text boxes. The vertical line denotes 1996, the year Darwin’s Black Box was published. The trendline was calculated via moving average with a period of 3. Graph and caption by Matt Inlay.

Inlay has added to the graph boxes showing the chronology of the major research findings supporting the transposon model. The viewer can see that they came fast and furious after 1996, almost as if on cue.

Major Themes

The story told by this bibliography, and by Ken Miller’s testimony in Kitzmiller, and by Bottaro et al. (2006), is basically of the proposal, development, testing, and confirmation of Sakano et al.’s 1979 suggestion in Nature that V(D)J recombination might have emerged when a transposon – a piece of DNA that codes for a protein that can chop out that piece of DNA and insert it somewhere else – inserted into the middle of a receptor gene. This hypothesis has been gathering steam for awhile, and really took off in the last ten years.

A subplot in the story is how evolution has played a key role in immunology throughout its development. Sakano et al.’s hypothesis was not born in a vacuum; it occurred in a field where comparative immunology, informed by evolutionary modeling, had been a core part of the discipline for a hundred years, going back to the work of Metchnikoff in the late 1800s.

If one reads chronologically through the bibliography, one will see the researchers themselves telling the same story, and citing the same papers that have been repeatedly cited in the various Behe immune system rebuttals. But you will also see some rather remarkable comments about Sakano et al.’s hypothesis. Some of the best ones will be highlighted.

For example, as the transposon hypothesis began to pick up steam in the mid-1990’s, it did attract some critisms. E.g., Lewis and Wu (1997) wrote,

“It is nonetheless worth noting that in spite of various similarities, there are fundamental differences between transposition and V(D)J recombination. For one, there are no reports of a transposase (mutant or otherwise) that is able to mediate site-specific inversion. Conversely, it has never been demonstrated that V(D)J recombination can cause the integration of one piece of DNA into another.” (Lewis and Wu 1997, p. 162)

However, the very next year, two labs discovered that the RAG proteins actually could cause the integration reaction. In a research article in Nature entitled “Transposition mediated by RAG1 and RAG2 and its implications for the evolution of the immune system”, Agrawal et al. write in their abstract,

“Here we show that RAG1 and RAG2 together form a transposase capable of excising a piece of DNA containing recombination signals from a donor site and inserting it into a target DNA molecule. […] The results support the theory that RAG1 and RAG2 were once components of a transposable element, and that the split nature of immunoglobulin and T-cell-receptor genes derives from germline insertion of this element into an ancestral receptor gene soon after the evolutionary divergence of jawed and jawless vertebrates.”

In a commentary article in the same issue of Nature, Plasterk (1998) said, “the authors show that the RAG1 and RAG2 proteins (which mediate V(D)J joining) can still catalyse a full transposition reaction. A similar result has been independently obtained by Martin Gellert and co-workers, and is reported in tomorrow’s issue of Cell.” This second article was Hiom et al. (1998).

In 1999, Schatz reviewed the progress of Sakano et al.’s transposon hypothesis and pronounced it a “prophetic hypothesis”:

“When examined in their germline configuration, the RSSs flanking V and J constitute an inverted repeat, much like that found at the end of transposons, and this realization led to the prophetic hypothesis, in 1979, that insertion of a transposable element into an ancient receptor gene exon was responsible for the generation of split antigen receptor genes during evolution ([Sakano et al., 1979]).” (Schatz 1999, p. 170, bold added)

Also in 1999, Susanna Lewis – the same Lewis as Lewis and Wu (1997) – engaged in a little prophesy of her own. If the RAGs evolved from transposons, she said, it would be really handy to find one:

“It would be extremely useful if a contemporary version of the original RAG transposon could be identified. A distant cousin, with credentials, would greatly facilitate any attempts to reconstruct the lost history between the time the first RAG element took up residence in the vertebrate genome and the emergence of a developmental recombination system.” (Lewis 1999, p. 65)

Six years later, exactly this was reported by Kapitonov and Jurka (2005). It is worth pointing out that, apart from the transposon hypothesis, there was no reason whatsoever to suspect that transposons similar to RAG should exist.

Full Annotated Bibliography

For the actual Annotated Bibliography, please click here.


Thanks to Andrea Bottaro, Matt Inlay, and the Panda’s Thumb crew for suggestions and corrections. Any mistakes that remain are my own.