The evolution of “irreducibly complex” antifreeze proteins in a polar fish (and a fish-slap at Behe)

A new paper in Proceedings of the National Academy of Sciences shows how a functional protein (an antifreeze protein in the blood of an Arctic fish) can be assembled out of scraps of genome that have no function at all. Moreover, the protein doesn’t become functional—e.g., being secreted into the fish blood to keep it from freezing—until the very last step of gene assembly, so the sequence looks “irreducibly complex”. But, contra the IDers, we can construct a perfectly naturalistic evolutionary sequence by looking at DNA in relatives and putative ancestors. This shows that the appearance of “irreducible complexity”—the existence of an adaptation that doesn’t seem to function until all its parts are in place—does not require a Behe-ian creationist Designer, but can arise from natural processes. But we already knew that.

Click on the segment below to read the paper; the reference is at bottom and the pdf is here.

The molecular mechanism and reconstruction of the evolutionary path is complicated, but I’ll try to present it in a stepwise fashion. (Remember that this is complex and I may get some stuff wrong; but I’ll do my best). Zhuang et al. used known phylogenies of fish related to the two “antifreeze” fish in the family Gadidae, a group of codfish. These codfish have functional antifreeze glycoproteins (AFGPs) that act to keep their blood from freezing when the cod swim in super-cold polar waters. The proteins do this by keeping ice crystals from forming and acting as sites of nucleation that could turn the fish into popsicles.

The fish’s AFGPs consist of three bits: the antifreeze protein itself, which consists of repeats of the amino acid sequence threonine-alanine-alanine (Thr-Ala-Ala), a second secretory protein that gives a signal to the genome to enable the antifreeze protein to be secreted into the blood, and a promoter region that is necessary to allow the DNA sequence to be transcribed into RNA (which then makes the antifreeze protein).

What’s remarkable about this configuration is that every bit of it, including the two proteins themselves and the promoter sequence, was cobbled together via translocations and duplications of DNA (this happens passively in the genome) until all the elements were in place. And the “functional” gene couldn’t function right up to the very end, when the promoter sequence moved to the right place to allow the protein-coding region to produce an RNA transcript. This entire series of steps was reconstructed by sequencing the DNA of relatives that don’t have functional AFGPs, so we could see the evolutionary order in which things were assembled, and where the functional bits originally came from.

Here’s how it occurred; this is Figure 4 from the paper, and I give its caption:

Evolutionary mechanism of the gadid AFGP gene from noncoding DNA. The color codes of the sequence components follow Fig. 1. (A) The ancestral noncoding DNA contained latent signal peptide-coding exons with a 5′ Kozak motif, adjacent to a duplication-prone 27-nt GCA-rich sequence. (B) The 27-nt GCA(Ala)-rich sequence duplicated forming four tandem copies. (C) A 9-nt in the midst of the four 27-nt duplicates became the three codons for one AFGP Thr-Ala-Ala unit and underwent microsatellitelike duplication forming a proto-ORF. (D) A proximal upstream regulatory region acquired through a putative translocation event. (E) A 1-nt frameshift led to a contiguous SP, a propeptide, and a Thr-Ala-Ala-like cds in a read-through ORF. (F) Intragenic (Thr-Ala-Ala)n cds amplification, fulfilling the antifreeze function under natural selection.

The evolutionary stages posited (and supported by sequence and phylogenetic analysis) are A-F. First, there is a sequence of GCAGCAGCA in an ancestor, a sequence that would normally code for repeated alanines, but wasn’t functional (this is found in a relative). It expanded through duplication: A —>B.

Then, a mutation from a guanine to a cytosine base in another ancestor converted one Ala-Ala-Ala sequence to a Threonine-Alanine-Alanine amino acid triplet, which itself expanded through successive duplications (B —>C). The gene now had four Thr-Ala-Ala units, but was still nonfunctional. But this was to be the core of the functional protein in the future; it’s the dark blue bit seen in C through F above.

Another part of the genome had a nonfunctional sequence that could serve as the secretory protein to get the dark blue protein secreted into the blood. A deletion of a single nucleotide (C—>E) rendered it capable of producing a signal protein (the purple bit in D-F). But the entire system was still nonfunctional because it lacked a promoter region.

Finally, the system became functional when the protogene moved to a location near a nonfunctional DNA region that could serve as a promoter for the nascent gene. Now a gene producing a repeated Thr-Ala-Ala protein could function and secrete it into the blood.

Further, natural selection could now act on the functioning gene to make it more effective, simply by selecting for those genes that had even more duplications of the Thr-Ala-Ala segment, so we had a big protein of repeated units that could act as an antifreeze in fish blood (E—>F). (More repeats = better antifreeze protection.)

It’s a bit more complicated than this, but this is the essence of how the final protein came to be. And it’s not speculation, because all the bits can be found in other species or posited in ancestors, and so this reconstruction is fairly sound. Moreover, it involves processes known to operate in the DNA: the moving of bits around by translocation, duplication of sequences, etc. No divine intervention is required to do this, even though the protein isn’t functional until it’s put together with the secretory protein and the promoter.

One might ask this reasonable question: “Well, if the nascent antifreeze protein is just sitting there and not doing anything before it becomes active, why isn’t it inactivated by mutations?” That’s a good question, and one answer is that the process took place reasonably quickly so that mutations (which are, after all, rare) didn’t have time to turn the dark blue protein core into gibberish. And once that core formed, duplication of the Thr-Ala-Ala would be rapid, promoted by natural selection because more repeats confer greater antifreeze activity.

So here we have an “irreducibly” complex system, functional as an antifreeze system only at the very end, but one that formed purely through natural and well-known genomic processes. No God or alien designer required. It’s a good example of how hard work (sequencing and phylogenetic reconstruction) can dispel the objection “we don’t understand how this irreducibly complex system formed, so God must have done it.”

The whole paper, besides being a really lovely piece of work, is a slap in the face of IDers like Michael Behe—a fish slap like the one below:


Zhuang, X., C. Yang, K. R. Murphy, and C.-H. C. Cheng. 2019. Molecular mechanism and history of non-sense to sense evolution of antifreeze glycoprotein gene in northern gadids. Proceedings of the National Academy of Sciences 116:4400-4405.


  1. Daniel Sharp
    Posted March 14, 2019 at 11:10 am | Permalink

    One would think that after all these years and all these humiliations, Behe would give up. I wonder if he’s ever actually asked himself, given how many experts have refuted him time and time again, ‘What if I’m wrong?’. Shall we hold out for that day? No, I don’t think so. Faith is a great shield for delusion. Alas. Luckily there are more intelligent people doing fascinating work like this, and they will be the ones better remembered than the fatuous ID brigade.

    • Mark R.
      Posted March 14, 2019 at 1:38 pm | Permalink

      Michael J. Behe (/ˈbiːhiː/ BEE-hee; born January 18, 1952) is an American biochemist, author, and advocate of the pseudoscientific[2] principle of intelligent design (ID).

      From Wiki…I like that in the first sentence of his bio, he’s called out as an advocate of pseudoscience.

    • gravelinspector-Aidan
      Posted March 14, 2019 at 10:01 pm | Permalink

      One would think that after all these years and all these humiliations, Behe would give up.

      One might think that until you realise that they’ve got a persecution complex. They like to be spurned, condemned and attacked in one unimportant-to-them part of their life because it makes them appear to be “Michael, Saviour of Our Parish Church” (sung got the tune of Flash, with as many camp posturings and pelvic thrusts as you care to gratuitously insert into the presentation) to the part of their life that they do care about.

      I’m sorry – I can’t keep a straight face. I have a mental image of seeing a Creationist starting to go off at the deep end onto another of their time-warping manglings of physics/ geology/ whatever, and the anti-creationists in the audience bursting into a chorus of “The Time Warp”. Most definitely with pelvic thrusts.
      It sounds more fun than pulling the wings off flies, and much more ethically defensible.

  2. Posted March 14, 2019 at 12:03 pm | Permalink

    Very well explained, Jerry. There were plenty of examples of evolutionary novelty way back in 2005 when I testified against Behe in the Kitzmiller trial, and every year since then more have appeared in the literature. If only because of the detailed step-by-step narrative, this is one of the best ever.

    • Posted March 14, 2019 at 2:45 pm | Permalink

      Good tweet. I’m gonna retweet it.

      Yeah, this isn’t really completely new, but it’s a good example of how you can reconstruct the evolution of “irreducibly complex” stuff.

    • Roger
      Posted March 14, 2019 at 3:01 pm | Permalink

      Then what the heck is he on about. Somebody let him know!

  3. Gregory Kusnick
    Posted March 14, 2019 at 12:05 pm | Permalink

    one answer is that the process took place reasonably quickly

    Another answer might be that if mutation had disrupted this gene complex before it became functional, there’d be nothing to talk about. That probably happens all the time, with duplications, translocations, and frameshifts randomly throwing together patterns that almost make some sort of functional sense, but that then dissolve and disperse again before natural selection can get a grip and conserve them. It’s only the winners of this lottery who make the news.

    • rickflick
      Posted March 14, 2019 at 2:44 pm | Permalink

      Good point. There is a level of pure happenstance in the process. Many brilliant genes were actually almost-genes before being destroyed. Often the critical, make-or-break event is when the configuration becomes just useful enough to warrant natural selection getting a grip on it, before subsequent refinements occur to produce what we see now.

    • Posted March 14, 2019 at 2:47 pm | Permalink

      Yes, certainly possible, but Behe would argue that this could NEVER take place.

    • infiniteimprobabilit
      Posted March 14, 2019 at 5:32 pm | Permalink

      I like Gregory’s answer.

      Because, unlike e.g. the eye, where each stage is a little bit better than the one before, this particular mechanism seems to confer no advantage at all before it’s fully complete.

      In the fullness of time, some of those steps probably occurred on many occasions, and were lost again because they gave no advantage (or lingered on as junk DNA?). Eventually, they all happened together and bingo! – it worked and gave a big advantage for Natural Selection to work on.

      We only get to see the ones that were successful. Just like we only hear about the Lotto winners.

      (Note I am NOT a biologist)


    • darrelle
      Posted March 15, 2019 at 6:50 am | Permalink

      Good to see you Gregory.

  4. Jon Gallant
    Posted March 14, 2019 at 12:06 pm | Permalink

    Perhaps Dr. Behe might respond that the Intelligent Designer is really an Intelligent Duplicator and Translocator (or ID&T), who oversees these processes so as to generate seemingly irreducibly complex outcomes.

    I gather that Dr. Behe’s latest book is based on the notion that nucleotide substitution mutations are the only (or principal) DNA change that occurs in nature. If so, he has not kept up with much recent and not-so-recent Genetics.

    • W.T. Effingham
      Posted March 14, 2019 at 12:55 pm | Permalink

      +1. Good Points. It’s as if one of the “best perks” of spreading misinformation is keeping the Target Audience® in mind knowing that any cutting-edge science is way beyond their comprehension.

  5. ThyroidPlanet
    Posted March 14, 2019 at 12:39 pm | Permalink


  6. Ken Kukec
    Posted March 14, 2019 at 12:44 pm | Permalink

    “… Behe might respond that the Intelligent Designer is really an Intelligent Duplicator and Translocator …”

    Don’t make William of Occam cut a bitch. 🙂

    • W.T. Effingham
      Posted March 14, 2019 at 1:00 pm | Permalink

      In this case, Occam would be (understandably) tempted to use a crusty, rusty razor.

  7. Mike Anderson
    Posted March 14, 2019 at 12:53 pm | Permalink

    a functional protein (an antifreeze protein in the blood of an Arctic fish) can be assembled out of scraps of genome that have no function at all. Moreover, the protein doesn’t become functional—e.g., being secreted into the fish blood to keep it from freezing—until the very last step of gene assembly

    The wonder of God’s genius never ceases to amaze. That Guy is so clever!

  8. Posted March 14, 2019 at 12:59 pm | Permalink

    The other problem area for good molecular evolution of course are the various scientists that claim that junk DNA actually has function.
    Cue the attempt to miss-use this discovery as a case in point for their side.

    • Steve Gerrard
      Posted March 14, 2019 at 1:58 pm | Permalink

      I’d be willing to accept “playground for messing around with genetic stuff without any penalties” as a “function” of junk DNA. It’s not what anyone usually means by “functional”, but I could live with it.

  9. Mark R.
    Posted March 14, 2019 at 1:35 pm | Permalink

    Thanks for this concise explanation. Will Behe read this? He sure needs to.

    • Posted March 14, 2019 at 2:48 pm | Permalink

      This assumes that Behe is open minded about this stuff. I am not sure that’s a good assumption. In fact, given his history (he’s over at Evolution News squealing about the critics of his book), he’s surely not open minded.

      • Adam L
        Posted March 14, 2019 at 3:20 pm | Permalink

        Did you notice you have your very own tag over at Evolution News?

        Seems you really get under their skin! Bravo!

      • Mark R.
        Posted March 14, 2019 at 3:42 pm | Permalink

        Yeah, I see your point. He’s both close-minded and thin-skinned…a bad combination.

  10. Curt Nelson
    Posted March 14, 2019 at 2:31 pm | Permalink

    The fish slapping dance is hilarious. I remember hearing once about the creative differences among the Monty Python troupe, that Michael Palin especially liked silliness and the fish dance was his idea. (So glad I could finally contribute to the evolution discussions.)

    • Torbjörn Larsson
      Posted March 14, 2019 at 5:32 pm | Permalink

      I remembered that. “Palin has stated that the sketch summarizes concisely what Python is all about.”

      [ ]

      As Jerry points out, it also summarizes concisely what creationism is all about. Strutting and fretting until the Trouth comes out.

  11. James Heard
    Posted March 14, 2019 at 2:38 pm | Permalink

    I’m still waiting for Behe to come up with a reasonable “god did it” explanation instead of attacking Natural phenomena in the name of religion.

    • Torbjörn Larsson
      Posted March 14, 2019 at 5:45 pm | Permalink

      I’m still waiting for visiting *Ken Miller* to stop attacking natural phenomena and find his parochial magic in quantum fluctuations now that we know religion is swindle analogous to astrology*.

      *) I.e. we now know intercessory prayer (horoscopes) doesn’t work thanks to Pew Institute meta-analysis, and we know for sure ‘god agencies’ (‘star signs’) are man made thanks to Planck showing us the universe was, is, and always will be 100 % natural.

      • Posted March 15, 2019 at 6:36 am | Permalink

        I give up, Torbjörn. When did I ever “attack natural phenomena?” On the contrary, I’ve been a public defender of science my whole career, and I’ve made it clear that evolution is a wholly natural phenomenon.

  12. norm walsh
    Posted March 14, 2019 at 3:00 pm | Permalink

    Happy pi day everyone!

  13. Posted March 14, 2019 at 4:12 pm | Permalink

    Fascinating. If stuff works, no matter how it comes together, it gets reproduced.

  14. Posted March 14, 2019 at 4:38 pm | Permalink

    What a fantastic paper! Reading it is like lifting the hood and watching the natural selection engine in action. Consider the statement,

    “With this paper, we fully resolved the lasting question of how two unrelated groups of fish at opposite poles: the Antartic notothenoid fishes and the northern codfishes, invented a near-identical AFGP. The notothenoid AFGP evolved within the structural framework of a preexisting gene ancestor but constructed a new AFGP cds from de novo expansion of a rudimentary partly non-sense tripeptide-coding element. Northern gadid displayed even greater evolutionary ingenuity, constructing all parts of a functional AFGP gene entirely from noncoding DNA.”

    It’s amazing that we’ve now developed such a level of insight into the workings of DNA that it’s possible to make such a statement. Prof. CC’s summary made it much easier for me to understand and realize the significance of the very jargony paper. I hope he doesn’t lose patience with us and stop writing about science.

    • Posted March 15, 2019 at 3:22 am | Permalink

      “Prof. CC’s summary made it much easier for me to understand and realize the significance of the very jargony paper.”

      I thoroughly agree with this. I’ve tried to read most of the papers referenced by our gracious host and without his zeroing in on key salient points it would require a lot more investment of time on my part to understand them. The diagram really did it’s job for me. I wonder how many examples of this type of diagram there are?

  15. Posted March 14, 2019 at 5:41 pm | Permalink

    This is an exciting discovery which, to tell you the truth, surprised me. However, even if all new genes after the initial abiogenesis originated from preexisting ones, evolution theory would still be quite OK.

  16. Hempenstein
    Posted March 22, 2019 at 8:17 am | Permalink

    In addition to brilliantly and painstakingly showing how the Arctic Northern gadids diddit, seems to me that the thing that really drives an oak stake into the ID gang is the independence of this route from that of the Antarctic notothenioid fishes, to arrive at nearly the same solution.

    If there was a Creator, why would it use the same route on both hemispheres? Oh, right, I know the answer! There were really two Creators, and one cheated, looking over the other’s shoulder.

    • Hempenstein
      Posted March 22, 2019 at 8:59 pm | Permalink

      …why wouldn’t it use….

  17. Posted March 23, 2019 at 11:10 am | Permalink

    Readers might like to contrast Jerry’s article with this misrepresentation by Dr Ross Anderson on the creationist website CreationEvolution ( as an interesting example of creationists disinformation.

%d bloggers like this: