Evolution Icon Evolution
Life Sciences Icon Life Sciences

A Fishy Story About AntiFreeze Gene Evolution

Casey Luskin

A new paper in Proceedings of the National Academy of Sciences claims to explain the origin of an antifreeze protein in an Antarctic fish. Though touted by Darwinian activists (note: I’m talking about Richard B. Hoppe, NOT the authors of the paper), this paper presents little more than a just-so story that shows no interest in testing the plausibility of the complex mutational account it gives.

The Darwinian blogosphere is excited about a new study which purports to explain the evolutionary origins of an antifreeze gene in an Antarctic fish, the Antarctic eelpout. Actually, this isn’t an entirely new story — claims of explaining the genetic evolution of antifreeze proteins have been made for years. Last year I reported another similar study:

These papers base their claims of evolution purely upon circumstantial evidence–comparisons of sequence similarity–and then tell a tale of deletion, reshuffling, and amplification. Explanation of these genes by “cobbling” via “[d]uplication, divergence, and exon shuffling” and “de novo” recruitment of non-coding sequences does not account for how such a complex gene could actually originate. This story does not address … the odds of spontaneously producing this functional gene.

The new story is little different. Published in the Proceedings of the National Academy of Sciences (PNAS), the new study looks at the origin of an antifreeze protein, AFPIII, a relatively short protein (66-amino-acids in length) which helps prevent freezing of water inside the fish by binding to incipient ice crystals. The authors claim AFPIII evolved from a duplicate copy of another gene in the fish, SAS-B. They summarize their evolutionary story: “AFPIII 5’FR, intron1 (I1), exon2 (E2; icebinding mature AFPIII), and 3’FR were derived from the 5’FR, I5, E6 (SAS C-terminal domain), and 3’FR, respectively, of the ancestral LdSAS-B,” and then “[a]ccelerated adaptive changes subsequently occurred in the nascent AFPIII gene.”

The Gene Evolution Game
Sounds simple and compelling, right? Don’t be too impressed. If you go back and read my article, “How to Play the Gene Evolution Game,” you’ll find that by using a combination of three magic wands — Gene Duplication, Natural Selection, and Rearrangement — it’s a simple matter to concoct a just-so story to “explain” the origin just about any gene sequence — no details required:

This summary of these 3 simple rules of the Gene Evolution Game will help you explain anything:

Gene Evolution Game Rule 1: Whenever you find sequence homology between two genes, just invoke a duplication event of some hypothetical, ancient ancestral gene, and you can explain how two different genes came to share their similarities.

Gene Evolution Game Rule 2: When you need to explain how a gene acquired some new function, or evolved differences from another gene, just invoke the magic wand of natural selection. No need to demonstrate that there is any benefit to the new gene, or that a step-wise path to adaptation exists. Finally, natural selection is especially useful when part of your gene appears unique–since natural selection can change anything, just conclude that natural selection changed your gene so much that it no longer resembles its ancestor.

Gene Evolution Game Rule 3: When a gene seems to be composed of the parts of several genes, just invoke duplications and rearrangements of all the DNA sequences you need, so you can get them all together in the right place. If you need to delete parts of a gene, or invert them, or transpose to a new location, just invoke different types of rearrangements as often and as liberally as you wish, and ba-da-bing, you’ve got your new gene!

This new study in PNAS about AFPIII uses Gene Evolution Game Rules 1, 2 and 3 to concoct a story that goes something like this:

Step 1: The purported ancestral gene, SAS-B, has a couple exons with high sequence homology to AFPIII. Using Gene Duplication Game Rule 1, Darwinian storytellers propose SAS-B underwent a duplication event producing SAS-B’.

Step 2: SAS-B has 6 exons, whereas AFPIII only has 2 exons. To get something that looks kind of like AFPIII out of a hypothetical copy of SAS-B, you have to get rid of a few exons (as well as the intervening introns) in SAS-B’. AFPIII’s exons 1 and 2 show sequence similarity to exons 1 and 6 in SAS-B, respectively. Using Gene Evolution Game Rule 3, the PNAS study’s authors propose that SAS-B lost exons 2, 3, 4, and 5, and introns 1, 2, 3, and 4. It also lost a portion of exon 1, which is missing in AFPIII.

Step 3: The exons in AFPIII aren’t identical to the exons in SAS-B, so you also have to explain how the differences evolved. Here’s where Gene Evolution Game Rule 2 comes in–natural selection, the magic wand which can change a DNA sequence into essentially anything you want. Thus, the PNAS article proposes “positive Darwinian selection” did all the work:

Accelerated adaptive changes subsequently occurred in the nascent AFPIII gene, indicated by nearly two-thirds of the residues in AFPIII experiencing positive Darwinian selection (Table 1); this resulted in rapid optimization to a full-activity AFP capable of preventing freezing of the fish body fluids.

Step 4: The AFPIII antifreeze gene is obviously important to Antarctic eelpouts, as they have “30 AFPIII genes.” The multitude of copies of this gene was thus said to be achieved through many rounds of duplicating the original AFPIII gene after it first evolved.

PNAS Authors Ask the Wrong Questions…
This story is slightly better than some, because here the investigators knew the function of the gene they were studying, and, based upon previous studies, they also knew of a few amino acids in the protein which were important for that function. Many other studies are much worse, for they invoke natural selection when they don’t even know the function being selected for. At least this study knew the function of the gene being studied.

In that regard, the authors felt it significant that some of the amino acids which were important for the antifreeze function had experienced “positive selection.” How do they know that they experienced positive selection? A high ratio of non-synonymous (i.e. amino acid changing) to synonymous (i.e. silent) nucleotide differences is taken as evidence of the force of “positive selection.”

What does this tell us specifically? It’s hard to say. Although other studies had determined that a few of the 13 amino acids in AFPIII under positive selection were chemically important for its ice-binding function (e.g. residues at positions 14, 16, and 44), some of the amino acids supposedly undergoing selection didn’t appear important for the protein’s antifreeze function. Moreover, some of the amino acids important for the ice binding function were not thought to be undergoing positive selection (e.g. residues at positions 9, 15, and 18).

If these results seem a little ambiguous or difficult to interpret, you’re not alone. Indeed it’s no coincidence that such methods of inferring positive selection have been criticized by other evolutionary biologists. For example, a 2007 article by evolutionary biologist Michael Lynch in Proceedings of the National Academy of Sciences USA goes to the heart of some of the assumptions inherent in many claims of genetic evolution. Lynch provides a list of “myths” promoted by biologists, one of which is the belief that “[c]haracterization of interspecific differences at the molecular and/or cellular levels is tantamount to identifying the mechanisms of evolution.” Likewise, biologist Austin Hughes warns that many inferences of positive selection are based upon questionable statistical analyses of genes:

A major hindrance to progress has been confusion regarding the role of positive (Darwinian) selection, i.e., natural selection favoring adaptive mutations. In particular, problems have arisen from the widespread use of certain poorly conceived statistical methods to test for positive selection. Thousands of papers are published every year claiming evidence of adaptive evolution on the basis of computational analyses alone, with no evidence whatsoever regarding the phenotypic effects of allegedly adaptive mutations. … Contrary to a widespread impression, natural selection does not leave any unambiguous ”signature” on the genome, certainly not one that is still detectable after tens or hundreds of millions of years. To biologists schooled in Neo-Darwinian thought processes, it is virtually axiomatic that any adaptive change must have been fixed as a result of natural selection. But it is important to remember that reality can be more complicated than simplistic textbook scenarios. … In recent years the literature of evolutionary biology has been glutted with extravagant claims of positive selection on the basis of computational analyses alone … This vast outpouring of pseudo-Darwinian hype has been genuinely harmful to the credibility of evolutionary biology as a science.

(Austin L. Hughes, “The origin of adaptive phenotypes,” Proceedings of the National Academy of Sciences USA, Vol. 105(36):13193-13194 (Sept. 9, 2008) (internal citations removed).)

Such criticisms seem applicable here. While previous studies had identified some of the supposedly “positively selected” amino acids as important to the antifreeze function, the authors make no attempt to provide a step-by-step explanation of how SAS-B’ changed into AFPIII. Were there selectively neutral — but necessary — steps encountered along the evolutionary pathway? Are there intermediate stages that are in fact maladaptive or non-functional? As I noted in “The Gene Evolution Game,” there are a number of important questions which should be addressed before a just-so story of genetic evolution can be made plausible:

  • Did the rearranged gene product … start out functional? If not, how quickly could it gain function? How was it preserved from loss until it became functional?
  • Are proteins really as malleable as this story would suppose or would the new combined gene encounter folding or other contextual problems?
  • What mutational pathway was taken to evolve Gene A … into a new gene with function B?
  • What selective advantages were gained at each small step of this evolutionary pathway?
  • Were any “large steps” (i.e., multiple specific mutations) ever required to gain a selective advantage along the evolutionary pathway? Would such “large steps” be likely to occur?
  • Could all of this happen on a reasonable timescale?

Alas, the paper addresses zero of these questions. Given their complicated story of duplication, deletion, and natural selection (and then lots more duplications), the least they could have done is performed some calculations to address the plausibility of their complex just-so story. For example, they could have performed calculations showing that the claimed mutations could occur on a reasonable evolutionary timescale. But the PNAS authors don’t seem interested in actually testing the plausibility of their evolutionary story. Instead, they just assert:

Thus, the evolution of Antarctic eelpout AFPIII has entailed tapping into two inconspicuous functionalities in the same cytoplasmic ancestor and remarkably, transforming them into a quintessential lifesaving antifreeze function. The partly nonprotein coding origin of AFPIII signal peptide represents an example of recruiting a hidden function by incorporating a translation start site, and sheds light on how products of duplicated genes could be targeted to different cellular localizations.

So, starting with the sheer dumb luck that the initial SAS-B gene had a “hidden function” that prevented ice crystal growth and “a sequence in the SAS-B gene that, when translated into a new protein, could — with a few modifications — direct the cell to secrete the protein” into the blood, obviously multiple mutations ranging from gene duplications to large deletions to numerous individual base substitutions would be necessary to evolve AFPIII.

…So We’ll Ask a Right Question

And there’s one last radical series of mutations necessary to explain AFPIII in Antarctic eelpouts, for eelpouts have over 30 copies of the AFPIII gene. Perhaps all this is plausible. But perhaps it isn’t. The paper doesn’t even attempt to tell us. Let’s try one simple analysis of the last stage.

In his 2005 textbook Evolution, Douglas Futuyma states that a high estimate of the gene duplication rate is “about 0.01 duplication per gene per million years.” (p. 470) A given gene will thus be duplicated about once every 100 million years. The present paper speculates that the antifreeze gene evolved in response to cooling temperatures in the Antarctic deep ocean water over the past 50 million years. What are we to make, then, of the fact that Antarctic eelpouts have over 30 AFPIII genes, all of which are said to have resulted from a duplication of a single AFPIII gene which evolved at some point in the past 50 million years in response to changing ocean temperatures?

Even if one invokes the magic wand of “positive selection,” this gene was apparently duplicating at a rate far higher than the average gene duplication rate. It should have taken some 3 billion years just to accomplish the last step of this little story, which took place in far less than 50 million years, as these repeated duplications are the very last step of the story. In other words, in the last stage alone it seems that this paper requires far too much genetic evolutionary change too quickly.

The paper doesn’t address any questions relating to the plausibility of its story and instead just asserts that story as true because, well, they played the Gene Evolution Game and this is what they got.

The Darwinian Choir Cheers Uncritically
Over at Pandas Thumb Richard Hoppe, in true PandasThumb fashion, taunts:

this is precisely the kind of evidence that Disco ‘Tute attack mouse Casey Luskin asked for a year ago:

Many scientific papers purporting to show the evolution of “new genetic information” do little more than identify molecular similarities and differences between existing genes and then tell evolutionary just-so stories of duplication, rearrangement, and subsequent divergence based upon vague appeals to “positive selection” that purport to explain how the gene arose. But exactly how the gene arose is never explained. In particular, whether chance mutations and unguided natural selection are sufficient to produce the relevant genetic changes is almost never assessed.

There it is, Casey.

I don’t usually feel the need to dignify this kind of behavior with a response, but for whatever it’s worth, Dr. Hoppe’s post uncritically capitulates to all the hype about this paper in its self-scripted ScienceDaily press release. Apparently the good Dr. Hoppe wasn’t interested in thinking critically about what they really found and had no idea what I was talking about. This PNAS paper does exactly what I said the other papers do (quite inadequately), and it doesn’t do exactly what I said the other papers fail to do but should be doing: It does not assess whether chance mutations and unguided selection are sufficient to produce the gene in question under the time allowed.


Casey Luskin

Associate Director, Center for Science and Culture
Casey Luskin is a geologist and an attorney with graduate degrees in science and law, giving him expertise in both the scientific and legal dimensions of the debate over evolution. He earned his PhD in Geology from the University of Johannesburg, and BS and MS degrees in Earth Sciences from the University of California, San Diego, where he studied evolution extensively at both the graduate and undergraduate levels. His law degree is from the University of San Diego, where he focused his studies on First Amendment law, education law, and environmental law.



Richard B. Hoppescience