After a cute introduction with outboard motors , Wells moves into Michael Behe’s use of irreducible complexity. Irreducibly complex systems as defined by Michael Behe are:
a single system composed of several interacting parts that contribute to the basic function, and where the removal of any one of the parts causes the system effectively to cease functioning.
(Behe 1996, p. 39)
This has been redefined a couple of times, but in the end the original definition is the one “intelligent design” activists continue to return to and the one used by Wells. To illustrate the “IC” concept, Wells uses Behe’s example of a mousetrap. Remove one component, and the mousetrap cannot function. Unfortunately for Wells and Behe, a cottage industry has sprung up making reducibly complex mousetraps, so this illustration has lost its force. Even more so since we have experimental and “in the wild” systems that have evolved “IC”.
The fact that there are complex systems that don’t work when you remove bits is in itself unremarkable. Behe’s contention is that systems that are “irreducibly complex” are unevolvable, and Wells presents this contention as fact. However, IC systems are far from unevolvable. Firstly, way back in the 1930’s Muller (1936) showed that evolution would produce such “IC” systems by incrementally adding parts that were dispensable at first but later became indispensable. In the 1970’s Cairns-Smith showed that systems could be assembled around scaffolding, which when removed left an “IC” system. Indeed there are several mechanisms whereby “IC” systems can come into being. The clear flaw in Behe’s contention is the fact that just because a modern system is not decomposable into smaller parts does not mean that it was not initially constructed from smaller parts.
There are other practical issues. Behe’s definitions of “part”, “system”, and “function” are arbitrarily flexible, making it easy for Behe to elide between different definitions and evade counter examples, as it suits him. For Behe, in the clotting system “part” means single enzymes, whereas for the flagella it is entire assemblies of proteins. We’ll see the implications of this later. However, Wells does not address any of these highly pertinent issues.
Then there is Behe’s admission that irreducibly complex systems may evolve indirectly, indeed, indirect evolution of systems by co-option is extensive in evolution. A beautiful example is this is the pentacholrophenyl (PCP) degradation pathway in bacteria. PCP is a highly toxic chemical which is not present in nature; it is only produced by humans and has only been present in any concentration in the environment within the last 60 years as an industrial waste product. Recently some bacteria have evolved the ability to metabolise PCP. The PCP pathway is irreducibly complex, in that removing any one of the three enzymes needed to break down PCP stops degradation, with subsequent cell death (Copley, 2000). Yet the PCP pathway was cobbled together out of two enzymes that broke down dicholrophenol (which is produced by fungi), and a mutation of maleylacetoacetate isomerase, that normally metabolizes amino acids. The mutation of maleylacetoacetate isomerase spontaneously formed a irreducibly complex degradation pathway from pre-existing intermediates (Copely 2000). An irreducibly complex system developed literally under our feet.
As well as having real, biological systems with irreducibly complexity evolving in front of us, we also have evidence from studies of computer models that IC systems can evolve. Yet you won’t find any of this in Wells’s book. You will only find a brief, irrelevant reference to the Kitzmiller trial, where Behe’s claims that the irreducibly complex immune system was unevolvable were systematically shredded, but no reference to the substantial biological evidence tendered in the trial. Some old favorites do get trotted out: the visual system, the clotting system and the bacterial flagellum. The real biological evidence is ignored, and tall tales are spun, so lets look at what Wells says in the light of evidence. Speaking of light, lets start with Wells misrepresentation of the visual system.
A little light on the subject
When light strikes the human retina it is absorbed by a molecule [retinal] which alters an attached protein [opsin], which then initiates what biochemists call a “cascade”—a precisely integrated series of molecular reactions that in this case causes a nerve impulse to be transmitted to the brain. If any molecule in this cascade is missing or defective, no nerve impulse is transmitted, the person is blind. Since the light sensing mechanism doesn’t function unless every part is present, it is irreducibly complex. The fossil record cannot tell us, and no evolutionary biologist has explained, how all these molecules assembled themselves to produce the light sensitive spot that was the starting point for Darwin’s speculation.
To start off with, Wells’s statement “If any molecule in this cascade is missing or defective, no nerve impulse is transmitted, the person is blind” is flat out wrong, arrestin and retinoid binding protein can be completely absent in vertebrates with no effect or mild night blindness (Gonzalez-Ferdanez, 2000). Other proteins can be damaged with only mild effects on vision. But nonetheless Wells is still using the flawed concept that if a modern system breaks if a component is damaged, it couldn’t have evolved. Lets see about that.
A light spot does not need to be as complicated or as tightly regulated as a vertebrate (or cephalopod) eye. One of the iconic light spots is that of the eukaryotic single-celled organism Euglena. The Euglena signal transduction cascade consists of a single protein, the light harvesting protein and the protein that generates the single signaling molecule are one and the same (Iseki 2002). Incidentally, this gives an insight into how signaling cascades develop. If gene duplication of the light sensing enzyme were to occur, and with subsequent mutation so that one copy remains light sensitive, and the other just signals, then you have a mini-cascade happening. Duplication and divergence underlies a large amount of evolutionary novelty.
But to return to light spots. As we saw, Euglena has a single-protein system, which is not irreducibly complex at all, and we can see how systems could be tacked on to it to form a cascade. Vertebrates use the molecule opsin, rather than the Euglena protein. Opsin is a very old molecule; bacteria have their own version, bacteriorhodopsin, which in its simple form is a single molecule that pumps ions across a membrane (Spudich 1998). Furthermore, changing ion concentrations in cells is a classic way to modify cells excitability and function. Once again we have a one step “cascade”, and a light spot based on this bacteriorhodopsin is something eminently evolvable.
In more advanced systems, bacteriorhodoposin is linked to the chemosenory signaling pathway. No new system was involved; an old one was co-opted. This theme of co-option is present in the visual system of multi-cellular organisms. Both invertebrate and vertebrate eyes, from simple to complex, use opsin to capture light. But in non-chordate invertebrates, opsin is linked to an enzyme called phospolipase C, and in chordates and vertebrates, it is linked to a phosphodiesterase enzyme (Nilsson 2004). In both cases, the light sensing molecule has co-opted existing signaling systems to work through (Nilsson 2004, Hisatomi 2002). You may ask how the phospholipase C and phosphodiesterase cascades were put together. These are relatively simple and highly flexible systems cobbled together from proteins with other functions. (See Nilsson 2004, Hisatomi 2002; we have already seen that you can mess around with these cascades quite a bit.)
Of course, you will learn none of this from Wells. Rather than go to the effort of finding out what researchers do know about visual system evolution, he spends his time criticizing Jerry Coyne because Coyne had the temerity to use an anatomical analogy to demonstrate how irreducibly complex systems could evolve. Wells blandly reproduces Behe’s fulmination that anatomy is quite irrelevant to molecular evolution. This is ironic, since Behe’s canonical example of irreducible complexity is a mouse trap. Coynes point is valid because it is the logic of part assembly, rather than the nature of the parts themselves, that is important. A modern human eye acts as an irreducibly complex system, remove a lens, for example, and you are effectively blind. But in simpler organisms, lensless eyes work adequately, and the modern human eye can be reached by minor modification of these systems. Having one photosensitive enzyme is inadequate for modern mammalian vision, but it is quite adequate for a simple photospot such as in Euglena, and subsequent enzymes can be tacked on.
This misunderstanding that modern dependence means systems are unevolvable extends to the clotting system.
The reducibly complex clotting system
Now before I go on, I’ll remind you of Behe’s definition of “irreducible complexity”: “a single system composed of several interacting parts that contribute to the basic function, and where the removal of any one of the parts causes the system effectively to cease functioning”.
I’d like to emphasize that the blood clotting system did not evolve in modern terrestrial vertebrates, as it sometimes seems to be implied in Behe’s writings. Instead it evolved in far more primitive organisms with low pressure blood systems, where a cut is less lethal than in vertebrates with high pressure circulatory systems. Clotting in the chordate lineage (the group of animals that includes vertebrates) arose somewhere between chordates and the earliest jawless fish. Primitive chordates such as the sea squirt have no clotting system; they plug up any wounds with mobile blood cells called haematocytes. Their equivalent in humans, platelets, are quite important in clotting but get left out of Behe’s descriptions (Jiang & Doolittle, 2003, Davidson et al., 2003, Yong & Doolittle 1987). This turns out to be important later on.
Have a look at the diagram of the clotting systems for a variety of chordates.
The reducibly complex clotting system. Jawed fish lack the extrinsic clotting system, hagfish lack factor X and other components of the vertebrate system, and the primitive chordate Amphioxus has no fibrinogen, but a thrombin like enzyme that produces a clot. Sea Squirts have no coagulation system, but plug holes with haematocytes. Click on the image for a full scale version.
The first thing you see is that several chordates have very much reduced clotting systems. Jawed fish lack the entire intrinsic pathway and have much less than the dozen proteins that Wells claims are essential for clotting to occur. Jawless fish, such as the hagfish, lack Factor X and some of the accessory enzymes and probably have only half of Wells’s “essential” proteins (Davidson et al. 2003). Without a complete genome of Amphioxus, we don’t know how many of the vertebrate clotting components it has but certainly less than jawless fish. We do know that it has an enzyme that forms clots and acts like thrombin in vertebrate coagulation assays, and we do know it lacks fibrinogen. However, it still forms clots. It is thus very clear that the clotting system is reducible. Why would Wells think otherwise? He is merely parroting Behe (as he does all the way through this chapter), without realizing Behe’s deep misunderstanding of biology. Here’s an example of how badly Behe misunderstands things from his chapter in Of Pandas and People:
Why is the blood clotting system incompatible with a nonintelligent evolutionary view of nature? Macroevolution means a change from a simpler to a more complex state. Let us try to envision such a change for blood clotting. Assume that we initially start with an organism that contains just a primitive version of thrombin and fibrinogen. The thrombin would immediately cut all the fibrin, causing a massive clot and the speedy death of the organism.
(Behe 1993, p. 145)
Several issues with Behe’s account immediately arise. A “primitive version” of thrombin would hardly chop up fibrinogen with the spectacular efficiency of our modern ones. Furthermore, it is the amplification of thrombin activity by the cascade that results in spectacular fast clotting in advanced vertebrates. Without the cascade, there would be no massive, system wide clot. Clotting would proceed in a more leisurely fashion, but this is no problem for organisms with low-pressure blood systems like Amphioxus. Behe is either being disingenuous or lacks even basic understanding of clotting physiology. If Wells was really investigating the clotting system, rather than parroting Behe, he would have noticed these points.
Such a simple “one step” system is not just theoretical speculation. Shrimp and some crustaceans have a one-step clotting enzyme, and they singularly fail to have lethal, system wide coagulation. Many other crustaceans operate with a simple two-component system (Theopold et al., 2004).
We know that the Amphioxus thrombin-like system is much reduced compared to jawless fish and jawed fish, let alone vertebrates, although the full details of the system are not clear. We do know about the trypsin-stimulated polyphenoloxidase system though (Pang et al., 2004), and this can serve as a model for primitive clotting systems in chordates.
When Amphioxus is wounded, tissue damage causes local calcium levels to be high. This activates trypsin released from the tissue (as well as circulating trypsin). The activated trypsin in turn activates polypheoloxidase in the circulation of Amphioxus, which the produces cross-linked melanin. While everyone knows that melanin is the pigment that makes our skins brown (and go browner after exposure to sunlight), fewer will be aware that it is an important component of the innate defense system in invertebrates. Furthermore, sticky masses of melanin are used to entangle and immobilize bacteria and can also help plug the wound (Pang et al., 2004).
Hmmm, a trypsin-like molecule, activated by high calcium on tissue injury, that makes gluggy stuff. That sounds just like thrombin! (Thrombin appears to be derived from tryspin-like molecules.) Fibrinogen is also used to limit bacterial invasion. Thrombin is not just used in the clotting system (despite Behe’s claim that it is—see Wang et al., 2005); it plays an important role in tissue remodeling and activating immune cells. It is highly plausible that a proto-thrombin was released during tissue damage and initially initiated wound repair, then started making gluggy stuff that limited bacterial invasion, and then was modified to make gluggy stuff that sealed blood vessels. (There are parallel examples seen in modern invertebrates). Given there is a calcium-activated trypsin that activates polyphenol oxidase locally, without glugging up the entire circulation with melanin, this suggests that locally released calcium could control a proto-thrombin easily (and calcium is critical to modern thrombin) conta Behe and in agreement with invertebrate “one step” systems.
From such a simple calcium-activated system, it is not too far to the simple hagfish system, and thence to the more complicated jawed fish system, and thence to us.
Behe (1996) wrote:
In fact, having a primitive, poorly controlled clotting system would probably be more dangerous to an animal, and therefore less advantageous, than having no such system at all!
Tell that to sea-squirts or Amphioxus (or any number of crustaceans), who handle clotting with a small component clotting system. (As noted above, sea squirts just use haemocytes, a very, very primitive precursor of platelets.)
Of course, you won’t find a discussion of this reducible clotting system from Wells. He chooses instead to misquote Russell Doolittle.
But (Doolittle claimed) “when these two lines of mice were crossed … [then] for all practical purposes, the mice lacking both genes were normal! He concluded: “Contrary to claims about irreducible complexity, the entire ensemble of proteins is not needed.”
But this is not what Doolittle said; Wells leaves out Doolittle’s very important qualifier.
And what do you think happened when these two lines of mice were crossed? For all practical purposes, the mice lacking both genes were normal! 6 Contrary to claims about irreducible complexity, the entire ensemble of proteins is not needed. Music and harmony can arise from a smaller orchestra. No one doubts that mice deprived of these two genes would be compromised in the wild, but the mere fact that they appear normal in the laboratory setting is a striking example of the point and counterpoint, step-by-step scenario in reverse!”
I have written a long article about Behe’s misquotation of Doolittle, and as Wells blindly follows Behe I direct you to this article (try and guess which one is the fibrinogen knockout mouse). But both Behe and Wells ignore a key point. Mice that have no fibrinogen, the key final step in the clotting cascade, still form clots. These aren’t good clots by any means, but the mice live as long as normal mice, and heal wounds as fast as normal mice, so they are good enough for the purpose in a laboratory setting. If clotting was really as irreducibly complex as Wells claims, clotting should utterly fail, and the mice should die at or shortly after birth, as we need the clotting system to keep our blood vessels intact. So if a mouse, with a modern, high pressure circulation system, can form clots without fibrinogen, then a primitive, low pressure clotting system would work quite well, just as it does in the fibrinogen free Amphioxus.
But of course Wells ignores this crucial aspect and proceeds to “the” bacterial flagellum.
In a spin with Flagella
The flagellum is beloved of ID promoters, as, at least in diagrams, actually looks like a human-built machine. But looks are deceptive, and there is a wealth of information to show that they evolved.
The relation ship of Type II secretory systems to type IV secretory/motility systems and the archebacterial flagellum. Homologous proteins are indicated by colour, the GspM/FlagG homolg Y1 has been omitted due to uncertainly as to its location in the membrane (modified from Musgrave 2004). Click to enlarge.
Again, Wells simply re-iterates Behes’ arguments. Behe’s contention is that the flagellum consists of a “motor”, a “shaft”, and a “propeller”. (Note that these “parts” are complexes of molecules, so when you demonstrate a flagellum with fewer proteins, Behe just says, “it’s still got a shaft”, whereas he insists that the “parts” of the clotting system are individual proteins and must all be present.) Behe asserts, and Wells concurs, that having a motor evolve without out a shaft or propeller would be impossible. On the contrary, it is easy to show how one can build up a flagellum from systems evolved for other purposes.
I have previously presented a continuously functional evolutionary pathway from a simple ancestor to a functional flagellum (Musgrave 2004). It is based on elaboration of a secretory system. The flagellar filament must be secreted to project outside the bacterial cell, so it makes sense that secretory systems form the heart of the flagellum. The type II secretory system features a small “piston” made up of helically arranged proteins. Up and down movement of the piston (powered by a “motor”) pushes materials outside of the cell. The type IV secretory system is an elaboration of the type II, except now the piston is a long filament, and that filament can stick to surfaces. The back and forth movement of the filament pulls the bacterium along, resulting in gliding motility. The flagellum is an elaboration of the Type IV secretory system, but now the filament freely rotates, rather than being stuck to a surface, and drives the bacteria along. Note that this fully functioning flagellum has only two of the three “parts” Behe insists are necessary for the irreducibly complex flagellum (the shaft and propeller are one and the same). We have all these real, functional intermediates leading to a functional flagellum, but you won’t find this out from Wells.
Oh sorry, that’s the archebacterial flagellum. Wells won’t tell you that there is more than one sort of flagellum or that flagellar motility is a minority amongst motility systems. Why ever would he ignore things like that?
Now the eubacterial flagellum is similar to the archebacterial flagellum in the sense that it built around a secretory system, but it’s a bit more complicated. Nick Matzke has a marvelously detailed article (Matzke 2003, Pallen & Matzke 2006) about the evolution of the eubacterial flagellum. The basic story is similar to that of the archebacterial flagellum. The core of the eubacterial flagellum is a type III secretory system. Virtually all the proteins in the flagellum can be accounted for as parts of existing systems or internal duplications (as predicted by evolutionary biology). Importantly, several gliding motility systems use similar motors and guidance systems to eubacterial flagellum, so a sequence of secretory system to gliding motility to swimming motility similar to archebacterial flagella is plausible. (Eubacterial flagella are also used in gliding motility, but there are other plausible pathways to swimming.)
Furthermore, the eubacterial flagellum is still a secretory system (Musgrave 2004, Matzke 2003, Pallen & Matzke 2006) and is even used by some bacteria to attach to cells and inject them with toxins (just like type III secretion systems) (Musgrave 2004, Matzke 2003). You can remove the “motor” or the “propeller” from the eubacterial flagellum and it still functions as a secretion system (Musgrave 2004, Matzke 2003). Indeed, some bacteria with paralyzed flagella use them as anchors to attach to cells and inject toxins into them. So you can see how you could build a eubacterial flagellum piecemeal around a core of a simple secretory system by direct Darwinian processes, then a small functional shift adds motility to this system (Musgrave 2004, Matzke 2003).
While Behe doesn’t explicitly say his subsystems have no function and concedes that some subsystems might have independent functions, his entire argument collapses if they do. His argument is that you have to have all “three” parts of the flagella (motor, shaft, propeller) in place at once for there to be any selectable activity. A motor by itself, he says, cannot be selected for on its own. But the motors of the flagella are variants of motors that are happily just pumping hydrogen ions in another part of the cell, so they are selectable by natural selection.
Similarly, to Behe a motor plus shaft is useless, as there is no selectable function without the propeller. But in real bacteria motors plus shafts are pumping proteins, attaching to surfaces, and so on, until that time a mutation in the shaft protein makes it curly, and gives selectable motion.
Wells follows Behe in critiquing Miller when he made a similar point but ignores the fact that we have a whole series of functional intermediates that exist in nature, from pure motors, to motor driven secretory systems, to rotating motor driven secretory systems used for gliding motility, to swimming flagella. The E. Coli EPC type III secretory system is visually identical to flagella.
Wells tries to object to Miller’s arguments about co-option by quoting Scott Minnich and Stephen Meyer.
… the flagellar motor consists of several dozen proteins that are not present in the [type III secretory system] but are “unique to the motor and are not found in any other living system”. They asked: “From whence, then, were these proteins co-opted”
It cannot be emphasized enough how wrong this statement is. The vast majority of the motor proteins are found in, or related to, other systems. There are 42 canonical proteins that make up a flagellum. Of these only 20 are required for motility (not several dozen). In turn, only one two of these, a rod connector protein and a cap protein, have not yet been found in other bacteria. Importantly, the actual motor proteins, and those that connect the motor to the body of the flagellum, all have relatives doing work in other systems (e.g. the MotAB motor proteins have relatives that drive secretion and gliding motility, Pallen & Matzke 2006).
“Intelligent design” activists often promote irreducible complexity as a “show stopper” for Darwinian evolution, and Wells follows this well worn path. Yet again and again irreducible complexity has been shown to be no barrier at all. We have computer models and real biological systems where irreducibly complex systems have evolved. Wells engages with none of this data, using misdirection and selective quotation to ignore substantive criticisms. In the ten years since Behe’s book was published (Behe 1996), “intelligent design” activists have produced no positive evidence for “intelligent design”. In contrast, back in 1987, Doolittle predicted based on evolutionary principles that fish would lack the extrinsic clotting system. Doolittle was right, once again evolutionary biology delivers testable predictions, while all Wells can deliver on Behe’s behalf is bluff and misquotations.
- Behe, MJ. Darwin’s Black Box: The Biochemical Challenge to Evolution (New York: The Free Press, 1996)
- Copley SD. Evolution of a metabolic pathway for degradation of a toxic xenobiotic: the patchwork approach. Trends Biochem Sci. 2000 Jun;25(6):261-5.
- Davidson CJ, Tuddenham EG, McVey JH. 450 million years of hemostasis. J Thromb Haemost. 2003 Jul;1(7):1487-94
- Doolittle RF & Feng DF (1987) Cold Spring Harbor Symposium on Quantitative Biology, 52, 869-874.
- Yong J & Doolittle RF (2003) Proceedings of the National Academy of Sciences, USA, 100, 7527-7532.
- Ganfornina MD, Sanchez D. Generation of evolutionary novelty by functional shift. Bioessays. 1999 May;21(5):432-9.
- Gonzalez-Fernandez F, Evolution of the visual cycle: the role of retinoid-binding proteins Journal of Endocrinology 2002 175 (1): 75.
- Hisatomi O, Tokunaga F. Molecular evolution of proteins involved in vertebrate phototransduction. Comp Biochem Physiol B Biochem Mol Biol. 2002 Dec;133(4):509-22.
- Iseki M et al., A blue-light-activated adenylyl cyclase mediates photoavoidance in Euglena gracilis Nature, 2002, 415, 1047-1051
- Jiang Y, Doolittle RF. The evolution of vertebrate blood coagulation as viewed from a comparison of puffer fish and sea squirt genomes. Proc Natl Acad Sci U S A. 2003 Jun 24;100(13):7527-32.
- Evolution of the Bacterial Flagellum (2004) IF Musgrave, pp 72-84, In “Why Intelligent Design Fails”, ed. M Young and T Edis
- Evolution in (Brownian) space: (2003) NJ Matzke Last accessed 22/08/05. warning, big file
- Muller. 1939. “Reversibility in evolution considered from the standpoint of genetics.” Biological Reviews of the Cambridge Philosophical Society. 14: 261-280.
- Nilsson DE Eye evolution: a question of genetic promiscuity, Current Opinion in Neurobiology 2004, 14:407–414
- Pang Q, Zhang S, Wang C, Shi X, Sun Y. Presence of prophenoloxidase in the humoral fluid of amphioxus Branchiostoma belcheri tsingtauense. Fish Shellfish Immunol. 2004 Nov;17(5):477-87.
- Pallen MJ and Matzke NJ From The Origin of Species to the origin of bacterial flagella Nature Reviews Microbiology 2006, advance online publication 2-7
- Spudich JL, Variations on a molecular switch: transport and sensory signalling by archaeal rhodopsins. Molecular Microbiology (1998) 28(6), 1051–1058
- Thornhill RH, Ussery DW. A classification of possible routes of Darwinian evolution. J Theor Biol. 2000 Mar 21;203(2):111-6.
- Theopold U et al., Coagulation in arthropods: defence, wound closure and healing, TRENDS in Immunology 2004 25 (6), 289-294
- True JR, Carroll SB. Gene co-option in physiological and morphological evolution. Annu Rev Cell Dev Biol. 2002;18:53-80.
- Wang X et al., Ovarian cancer, the coagulation pathway, and inflammation, Journal of Translational Medicine 2005, 3:25
Darwin’s Black Box: Irreducible Complexity or Irreproducible Irreducibility? by Keith Robinson
Irreducible complexity demystified by Pete Dunkelberg.
A Darwinian explanation of the blood clotting cascade by Kenneth Miller
Generalized review of Behe and irreducible complexity