The Design Matrix

I am going to start an on-going series of essays that correct common misunderstandings about the hypothesis of front-loaded evolution as laid out in The Design Matrix. Look for updates to this series as the months go by.

Today, we’ll start with the following erroneous belief:

Claim #1: The original front-loaded state had all the information needed to build complex, multi-cellular organisms (including humans).

This view of front-loading is often favored by those who think that evolution is incapable of generating new “information.” As such, the idea is that all the information for life is invested in a rather remarkable original state and that the history of evolution has been merely a history of differential information loss and maintenance.

Yet this view is flawed precisely because it denies that evolution can create new information. Front-loading is simply about biasing and channeling the type of new information that the blind watchmaker would stumble upon and use. For a simple example, consider the Wiki description as linked to above:

For instance, say a certain gene codes to produce the protein mentioned above (or any protein, for matter). A duplication mutation causes duplication of this gene, so now there are two copies of this gene in the genome. This mutation is clearly not a harmful mutation, since it supplies a protein that is already there, and as such has little or no effect on the organism. Then a second mutation causes the copy to make a different version of the protein, which adds some new ability or function. That mutation would clearly be selected for, because it is beneficial. Now the “information” in the second gene has been added, without any loss in genome space. A similar mechanism can work by using some of the unused DNA that makes up the majority of all genomes.

As I explain in The Design Matrix, gene duplication is precisely a mechanism we would expect from front-loading. And in this case, the new information in the second gene was not a string of purely random nucleotides, but a slightly altered version of a pre-existing gene, such that the past constrained what was found in the future. The original sequence, working in conjunction with the genetic code and the rest of the cellular architecture, biased what the second functional gene could look like.

Instead of Claim 1, the proper formulation is this:

The original front-loaded state had sufficient information that would bias evolutionary trajectories needed to evolve complex, multicellular organisms.

We have seen the single-celled organism, Tetrahymena, possesses an insulin receptor. Yet since insulin receptors belong to a class known receptor tyrosine kinases (RTKs), let’s talk about this class.

We can think of an RTK as a communication device, since these membrane proteins transmit signals from the cell’s environment into the cell. Event X outside the cell is translated into Event Y inside the cell. Specifically, the signaling molecules (such as hormones) bind the extracellular portion of the receptor protein. This binding event is then somehow communicated to the contents inside the cell. But how?

The RTKs span the membrane only through a single alpha helix, which means that you probably can’t transmit a conformational change from the external part of the protein to the internal part. Instead, the transmission strategy employs dimerization, where two RTKs bind to a signaling molecule which in turn leads the two RTKs to stick to each other.
Once they are stuck to each other, we can shift our focus to the part of the RTK that is under the membrane and exposed to the cell’s cytoplasm. The cytoplasmic components of each RTK now attach phosphate groups to each other (phosphates are added to the amino acid tyrosine). Once this happens, they become docking and activation sites for a variety of intracellular signaling proteins. These activated intracellular proteins can then kick off a cascade of events that can spread and/or amplify the signal, resulting in dramatic changes in the cell’s metabolism or gene expression. Of course, if you can turn on a switch, you better have a way turn off the switch, so your cells also possess protein tyrosine phosphatases that can strip the phosphates off the receptor’s tyrosines when needed.

Thus, RTKs can couple two seemly unrelated events - the binding of some molecule to the outside portion of the receptor and phosphorylation of the inside part of the protein. The latter event is a common way for cells to turn things ON and OFF, meaning that the binding of some molecule on the surface of the cell can radically alter what happens inside the cell. The modularity of this strategy is essentially conventional, where the relationship between the signaling molecule and intracellular events is determined solely by the identity of the binding domain of the RTK and the identity of the intracellular molecules that become activated by the RTK. If you swap extracellular binding domains, for example, it would simply mean that a different signaling molecule could elicit the same response. Such a process would clearly facilitate multicellular life, as signaling molecules released from one type of cell in your body would control the activity of a cell in another part of your body. And the potential for permutations needed to control a myriad of cells and processes is built into the basic design of this process.

Yet, as is usually the case, RTKs don’t function in isolation, but instead function as part of a system. We’ll look more closely at this in the next entry.

Over at Telic Thoughts, David E Levin writes:

It seems to me that front-loading of genetic information makes the very strong prediction that we should find in the genomes of simple species remnants of genes whose functions are specific to complex species. If all of the genetic information to make vertebrates (for example) was front-loaded into the earliest bacterial species, followed by functional loss of information from the genomes of species that did not need particular genes, we should see remnants of at least some of those lost genes. Are there, for example, remnants of metazoan-specific genes found in the genomes of protozoa or bacteria? As far as I am aware, there are not. For instance, a search of genomes for a large class of metazoan-specific genes that encode tyrosine kinase receptors, a distinctly metazoan innovation (from the evolutionary perspective), reveals nothing in the way of related pseudogenes or gene remnants in any bacterial or protozoan genome. This is the sort of evidence that one would have to produce for the idea of front-loading to be taken seriously. In the absence of evidential support for the idea, the rest of the discussion is meaningless. By contrast, there is plenty of evidence for the generation of new information through established evolutionary mechanisms, such as gene duplication (followed by mutational specialization), exon shuffling, and exonization of SINES, just to name a few. Where is the evidence for front-loading? Somebody help me to understand why this idea is worthy of serious consideration. (emphasis added)

Yes, mainstream evolutionary theory did predict that tyrosine kinases were indeed “a distinctly metazoan innovation.” And yes, if metazoans were front-loaded to appear, we would expect to find remnants of such genes in various unicellular organisms. But Levin is simply wrong in asserting “a search of genomes for a large class of metazoan-specific genes that encode tyrosine kinase receptors, a distinctly metazoan innovation (from the evolutionary perspective), reveals nothing in the way of related pseudogenes or gene remnants in any bacterial or protozoan genome.” On the contrary, scientists were surprised to find that single-celled choanoflagellates are loaded with tyrosine kinases and Tetrahymena possesses a receptor tyrosine kinase that probably responds to human insulin.

I will be adding some more neat stuff to the story about tyrosine kinases and front-loading shortly, but this example alone should help people to see why some of us do indeed think front-loading evolution, a hypothesis that has become steadily more plausible, is worthy of serious consideration. The sort of evidence that one would have to produce for the idea of front-loading to be taken seriously exists and continues to grow.

Relevant Reading:
Front-loading epithelial tissue

Old vs. New Ways of Viewing Evolution

A recent editorial from NewScientist is entitled Intelligence isn’t all it’s cracked up to be.
The author makes a couple of points that carry implications that he/she might be uncomfortable with.

First, we have this:

One reason might be that by fixating on intellect, we further the idea that we stand apart from nature as the only creatures that orient themselves by deliberate reasoning. Yet each day brings more evidence to undermine this view. For example, modern psychological studies show that when people make good decisions it is often by “gut feeling” rather than conscious calculation. It seems our vaunted intelligence is not all it’s cracked up to be…… As Socrates knew, the really intelligent know the limits of their own ability, an idea we seem to be relearning.

The author is entirely correct in cautioning us not to overestimate our intelligence. But let’s not forget a crucial point – our intelligence lays on the foundation of science, for it is our intelligence that expresses itself via the scientific method. Our intelligence allows us to mentally envision and form a hypothesis. Our intelligence allows us to formulate the hypothesis in a testable form and then carefully design an experiment. Our intelligence then comes into play when analyzing and interpreting the results of an experiment. It is our intelligence that then converts objective data, obtained by design, into the subjective concept of evidence.

Thus, we can legitimately rephrase the last sentence to read, “As Socrates knew, the really intelligent know the limits of science, an idea we seem to be relearning.”

Secondly, the author notes:

Such thinking appears to be moving towards the mainstream, as societies increasingly face complex problems that overwhelm the human mind. Engineers are finding that their task is not so much to find solutions as to design systems that can discover their own.

This is further evidence of the convergence of engineering and biology. If the task of the engineer is to design systems that can discover their own solutions, that, in a nutshell, goes a long way in describing the design and front-loading of evolution. The perspective of evolution as something that was designed is slowly coming into better focus as our own technology advances and converges.

In The Design Matrix, I explore the manner in which cytosine deamination, one of the most common DNA lesions, might facilitate evolution by substituting a pool of random amino acids for a pool of hydrophobic amino acids (I first described this back in 2002 ). But the story doesn’t stop there.

In 2005, David Orren published a paper (The irresistible resistance of nonsense: Evolutionary adaptation of termination codons to minimize the effects of common DNA damage. DNA Repair 4: 1208-1212) showing that all three termination codons are completely resistant to cytosine deamination (in the excerpt below, TM = transcriptional mutagenesis):

However, the sequences of termination codons also protect them from the effects of cytosine deamination of DNA ( Fig. 2). The UAA ochre codon derived from template ATT is obviously not susceptible to changes caused by cytosine deamination. The UAG (amber) and UGA (opal) termination codons are generated from template ATC and ACT, respectively. However, cytosine deamination to uracil in template ATC and ACT yields ATU and AUT, respectively. Thus, termination codon sequences are resistant or well-adapted to TM resulting from not only guanine lesions but also cytosine deamination. Intriguingly, the predominance of the ochre codon as the termination signal for prokaryotic genes [9] might be explained by cytosine deamination events that ultimately result in convergence of ATC and ACT sequences to ATT in the template DNA.

So the same code that appears to funnel cytosine deamination events toward a particular type of amino acid substitution also completely prevents cytosine deamination from eliminating the stop codons.

Orren notes:

Thus, in the face of common DNA damage, the design of the genetic code promotes mutations that favor substitutions over C-terminal extensions in even greater proportions than would be predicted by a completely random induction of codon changes. In evolutionary terms, this tendency towards less disruptive changes in protein structure and function may have more often allowed survival while promoting adaptation to environmental conditions.

And

Thus, the genetic code may be tailored to facilitate evolution via TM and retromutagenesis in prokaryotes and possibly in unicellular eukaryotes.

The puzzle pieces are slowing fitting together.

The blog Evolution Engineered looks at “the evolution-ID debate from an engineer’s perspective.” Recently, he provided some interesting reviews of different parts of The Design Matrix. In this posting, he surveys the ways in which the design perspective can guide research, after noting that I do not consider ID to be science. Since many people have difficulty with that distinction, let me address it.

As I have just noted, the DM approach is much more like a police investigation than science (keeping in mind that science is incapable of determining whether or not life was designed). Thus, the four criteria of the DM (Analogy, Discontinuity, Rationality, and Foresight) can all be used, as part of an investigation, to inspire and guide testable predictions about biotic reality. But here is the catch. Science can only address the questions raised by the criteria themselves (i.e., Is the analogy with an artifact strengthened? Does the molecular machine have system-dependent parts? Is the system a frozen accident? Etc.) Science is not determining whether or not a system is designed. It’s like a police investigation, where science can address whether or not the blood sample came from a suspect, or whether or not the victim died from a blow to the head (etc.), but it cannot determine whether or not the suspect is guilty.

Yet the scientific data can be fitted into the DM. That is, any data from testable hypotheses about the four criteria can be used when scoring with those criteria. The scoring process itself is not science , but the information can be used to assess the strength of a design inference as part of an investigation. As I noted in the book:

While the Design Matrix score might not be perfect, it is a significant step to scale the strength of the teleological signal. Furthermore, the Design Matrix score helps us move beyond the realm of suspicion, as the score itself can serve as both an impetus and a focal point for new research ideas that can in turn feed back into the score, either strengthening it or weakening it over time…… Thus, the Design Matrix can not only bring focus for research, but is also receptive to the findings of research.

With this in mind, let us consider some criticisms raised by someone on JJS P.Eng’s blog.

If these are indeed “working design hypotheses”, they should lead to predictions about the mechanism. For example, can ID “theory” postulate HOW transcriptional proofreading (one of the examples you cite) come about?

This criticism doesn’t work. With regard to the transcriptional proofreading, predictions about mechanisms are not needed from the investigative perspective. What is relevant is that the Rationality criterion, working with background information about other forms of proofreading, predicted that transcription itself would be proofread. Whether or not such proofreading occurred could then be tested. This information could then be fed back into The Matrix.

This criticism about mechanism, while common, is rooted in a faulty understanding of detecting a teleological cause (mind) as if it were just another non-teleological cause. This is explained on pp. 242-243. I should also mention that JJS P.Eng provides an insightful observation:

You are treating design as a mechanism, and any engineer can tell you that is patently false. Engineers make use of mechanisms in their designs.

It would be interesting to see an engineer expand on this theme.

JJS P.Eng also adds:

So the question really is “Can a myopic tinkerer/Blind Watchmaker create natural objects that appear to be designed?” Depending on the object under investigation, the answer can be possible, plausible, or probable. If the answer is only possible, this could raise suspicions of a rational designer. Thus more experiments would be required.

Here I would make a small, but significant change. Instead of the question being “Can a myopic tinkerer/Blind Watchmaker create natural objects that appear to be designed?”, it should be “Did a myopic tinkerer/Blind Watchmaker create natural objects that appear to be designed?”

Let’s get back to the skeptic and his second criticism:

Secondly, it is abundantly clear that all of the observations and bits of evidence used by Mr(s) Gene were provided by scientists who were not laboring under the design paradigm. In fact, it is abundantly clear that no scientist who advocates design has ever contributed one bit of evidence for it. Does this parasitism bother you?

First, it is not “abundantly clear” the observations stem from a non-teleological perspective, as chapter 3 outlines the many ways in which non-teleologists have heavily borrowed from the design perspective. Second, this criticism would have teeth if I was advocating that the DM is science. But since this is not the argument of the DM, the criticism fails. Data, whether generated by simple observation or obtained from any experiment, are always open to reinterpretation and there is nothing wrong with this at all. This is explained in chapter 6.

As for the third criticism:

When ID scientists come up with an observation that clearly supports teleology and also cannot be accommodated by evolutionary thinking, you will have a real reason to blog about this stuff. Until then, it’s just blowing smoke.

Yet the need to come up with something that “also cannot be accommodated by evolutionary thinking” is simply god-of-the-gaps thinking that stems from the traditional template that portrays design and evolution as mutually exclusive concepts. The DM does not rely on god-of-the-gaps thinking (chapter 10) nor the traditional template (chapter 2), does not view evolution and design as mutually exclusive (chaper 2) and outlines the logic of designing through evolution (chapter 7).

The skeptic’s criticisms may apply to the “ID is a growing field of science that shows evolution did not happen” that is common in the ID movement, but all of the criticisms fail against the argument(s) of The Design Matrix.

In his book, Chance and Necessity (1971) Nobel Laureate Jacques Monod puts his finger on the subjective element that is necessary to detect design (design being an expression of another mind):

Hence it is through reference to our own activity, conscious and projective, intentional and purposive-it is as makers of artifacts-that we judge of a given object’s “naturalness” or “artificialness.”

Yet Monod likewise explains why teleology can never be science without changing science itself:

The cornerstone of the scientific method is the postulate that nature is objective. In other words, the systematic denial that “true” knowledge can be got at by interpreting phenomena in terms of final causes – that is to say, of “purpose.” An exact date may be given for the discovery of this canon. The formulation by Galileo and Descartes of the principle of inertia laid the groundwork not only for mechanics, but for the epistemology of modern science, by abolishing Aristotelian physics and cosmology. To be sure, neither reason, nor logic, nor observation, nor even the idea of the systematic confrontation had been ignored by Descartes’ predecessors. But science as we understand it today could not have been developed upon those foundations alone. It required the unbending stricture implicit in the postulate of objectivity – ironclad, pure, forever undemonstrable. For it is obviously impossible to imagine an experiment which could prove the nonexistence anywhere in nature of a purpose, of a pursued end.

But the postulate of objectivity is consubstantial with science; it has guided the whole of its prodigious development for three centuries. There is no way to be rid of it, even tentatively or in a limited area, without departing from the domain of science.

Of course, as the guy in the middle, there is the other side of this observation – if life was indeed designed by an intelligent agent, science cannot incorporate this and must come up with another explanation that fits the canon, even if it means a reliance on promissory notes without an expiration date.

To overturn the canon and redefine science, one would need very powerful evidence. But what if the evidence was merely suggestive and weak? Then you should “depart from the domain of science” and carry on.

While not perfect, an analogy between The Design Matrix and a police investigation is useful in many ways. Here are three:

1. The analogy helps us appreciate what the DM is. A police investigation is an attempt to use empirical facts and our knowledge of other intelligent agents (humans) to determine what happened in the past. It is not purely subjective, but neither is it purely objective or science. It exists somewhere on the continuum between the subjective and objective and makes use of both. This is all similar to The Design Matrix.

2. The analogy helps us understand how the DM relates to science. A police investigation can make use of scientific findings without itself becoming science. Just because a detective may use DNA fingerprinting as part of his investigation does not mean the detective is trying be a scientist or imply his investigation is science. Likewise, The Design Matrix makes use of scientific findings without trying to imply it itself is science.

3. Finally, the analogy helps us understand how the DM relates to critics/skeptics. Most critics/skeptics seek to pass judgment, wanting precise definitions, evidence, and even a complete theory. But this is akin to playing the role of judge or jury, something that does not come into play until after the investigation is completed and deemed successful. The judge and jury do not pass judgment on the investigation while it is in progress. Only the investigators themselves are making judgment calls. The Design Matrix outlines the characteristics of a good investigator on page 33.

The Design Matrix will not help you win the Culture War or Convince the Skeptics. But if you have an open mind and share in the suspicion of design, and would like to join the investigative team, The Design Matrix is your training manual.

.

If you wanted to build a muticellular organism, it stands to reason that you would need to include an array of sensors that allowed this complex organism to detect its environment so it could respond. In our case, we have many senses. Consider the sense of smell and taste.

Smell and taste are very similar, in that both senses detect chemicals in the same way. With the sense of smell, molecules (which we call odorants) bind to receptors on the cilia of receptor cells in the roof of the nasal cavity. With the sense of taste, molecules (which we call tastants) bind receptors on the cilia of receptor cells in taste buds in your tongue. If enough tastants or odorants bind to the receptor, this triggers an electrical signal, known as an action potential, that is sent to the brain. The brain is thus notified about what is in your mouth or nose.

With taste, there are four primary taste sensations – sweet, sour, bitter, and salty – where each sensation represents to class of molecules (for example, sour represents acidity). Of course, the food we taste is not simply a combination of these four taste sensations (which sensation does steak taste like?). The food we taste is a combination of signals that some from both the taste buds and the more diverse receptors in the nasal cavity. This is why food tastes like it smells and why we can’t taste most foods when we have a head cold (the receptors in the nasal cavity are buried with inflammed tissue and secretions).

The two senses are also similar in that they both undergo rapid adaptation. For example, if you walk into a room and it stinks, you will stop noticing the stink after remaining in the room for some time. The stinky molecules remain in the room, but your olfactory receptors in your nose stop sending signals to the brain. Your brain, which must determine whether you will react to changes in the environment, is most concerned with new smells and tastes.

Both taste and smell receptors behave as sentries. Since the nose and mouth act at the interface of your inner body and the environment, we need some sensors that not only inform us about the environment, but whether the contents of that environment are helpful or threatening to our body. Things that smell or taste bad will trigger an avoidance response.

So how would a designer front-load such a sentry system into a single cell, something without a nose, mouth, or brain?

Once again, Tetrahymena serves as a nice model for thinking about front-loading, as a recent study measured whether or not Tetrahymena would detect and react to odorants and tastants commonly detected by vertebrates. [1]

Among the molecules analyzed were as follows:

menthol - cool taste
capsaicin – hot; active component of chili peppers
carvacrol - oregano
eugenol - cloves
piperine - black pepper
chloroquine – bitter
denatonium benzoate - very bitter
allyl isothiocyanate – horseradish

When these molecules are added to Tetrahymena’s media, they all trigger changes in the swimming behavior of the cells. In other words, the single-celled Tetrahymena also “smells” or “tastes” these molecules.

What’s more, Tetrahymena show adaptation to carvacrol, eugenol, quinacrine, and capsaicin. And previous work has shown that the altered swimming behavior is associated with generating actions potentials. So the basic foundation for the sense of smell and taste are represented in this single-celled organism: the ability to couple the binding of specific odorants and tastants in the environment with the generation of action pontentials, altered behavior, and sensory adaptation.

Remember the basic problem that a front-loading designer has – how does one design nonexistent multicellular organism for the future through a unicellular lifestyle? The logic of the design plan is laid out in my book, The Design Matrix. Yet by considering Tetrahymena alone, we can see how the foundations for both the endocrine system and two of the special senses could be implanted in a single-celled organism, such that this information could then serve as a seeing-eye dog for the blind watchmaker during subsequent evolution.

1. Rodgers LF, Markle KL, Hennessey TM. 2008. Responses of the ciliates tetrahymena and paramecium to vertebrate odorants and tastants. J Eukaryot Microbiol. 55:27-33.

Human insulin has the ability to mimic a growth factor for the single celled organism, Tetrahymena. But did you know that Tetrahymena’s ability to respond to mammalian insulin has been extensively studied?

For example, Tetrahymena produce an insulin-like molecule that is detectable with an antibody that specifically binds human insulin. It turns out that when Tetrahymena cells are starved, they increase the expression of this insulin. [1] In fact, you can stimulate Tetrahymena to uptake glucose by adding mammalian insulin to the culture, [2] the very type of response we see in mammalian cells. Mammalian insulin has also been shown to stimulate an adenyl cyclase signaling system in Tetrahymena, [3] a circuit that has also been studied mammalian cells [4]. The discovery of an insulin-activated adenyl cyclase in Tetrahymena led the researchers to conclude that “at the earliest stages of evolutionary development of unicellular organisms the molecular mechanisms providing hormonal regulation of various cellular processes have already been formed in these organisms.” Echoes of front-loading?

But it gets better. Not only does this protozoan possess insulin-receptors, insulin-like molecules, and respond to mammalian insulin, but it turns out that Tetrahymena respond to mammalian-hormones and possess many mammalian-like hormones:

Phagocytosis in Tetrahymena is enhanced by histamine and serotonin but no reaction has been observed to the chemically related indoleacetic acid, insulin influences its sugar uptake and triiodothyronine enhances its cell division. Tetrahymena also contains (produces, stores and secretes) vertebrate-hormone-like molecules, such as insulin, relaxin, adrenocorticotrophic hormone, endothelin, serotonin and histamine. The first encounter that Tetrahymena makes with a hormone provokes hormonal imprinting, following which the capacity of its binding sites and its hormone content change for hundreds of generations. A signal transduction system, similar to the mammalian system, is also present in Tetrahymena. Thus, the unicellular Tetrahymena has each component of an endocrine system. (emphasis added). [5]

Given that the endocrine system (our system of glands and hormones) plays an essential role in animal life, that a single-celled organism can possess “each component of an endocrine system” clearly adds to the growing plausibility of front-loading, as the core functional elements of this system can be packaged in a unicellular lifetyle. This point is further emphasized by the fact that human or bovine insulin can influence cell behavior in Hydra, plants, and even protozoa.

1. Csaba G, Kovács P, Pállinger E. 2007. Effect of starvation on insulin production and insulin binding in Tetrahymena. Cell Biochem Funct. 25:473-7.

2. Csaba G, Lantos T. 1975. Effect of insulin on the glucose uptake of protozoa. Experientia. 31:1097-8.

3. Shpakov AO, Derkach KV, Uspenskaia ZI, Shpakova EA, Kuznetsova LA, Plesneva SA, Pertseva MN. 2004. Regulation of Adenylyl Cyclase Signaling System in Cell Cultures of Infusoria Dileptus anser and Tetrahymena pyriformis by Peptides of Insulin Superfamily Journal of Evolutionary Biochemistry and Physiology 40: 364-373.

4. Pertseva MN, Plesneva SA, Shpakov AO, Rusakov YuI , Kuznetsova LA. 1995. Involvement of the adenylyl cyclase signaling system in the action of insulin and mollusk insulin-like peptide. Comp Biochem Physiol B Biochem 112:689-95.

5. G. Csaba . P. Kovács . Éva Pállinger. 2005. How does the unicellular Tetrahymena utilise the hormones that it produces? Paying a visit to the realm of atto-and zeptomolar concentrations. Cell Tissue Res 327: 199–203.

We’ve seen that Tetrahymena cells cannot exist in a sea of food unless the start-up concentration exceeds 100 ml/cells. While these data suggest Tetrahymena produces some type of growth factor that is endogenously synthesized and secreted, might something produced by a multicellular organism be able to substitute for this protozoan growth factor?

Below are the results from an experiment conducted by Søren Christensen’s lab (1):

The data shown in (A) are cells that are added to media. Note that a starting concentration of 400 and 40 cells/ml leads to cell death, but at 4000 cells/ml, you see growth of the population. The data in (B) show growth at all starting concentrations. So what is different? The researchers simply added human insulin in (B). We’ve seen that mammalian insulin enhances cell division in Hydra and plants, thus, as we would expect from a front-loading perspective, human insulin also can substitute for a Tetrahymena growth factor and stimulate cell survival and division. In fact, insulin appears to function in two different realms of concentration. Here is an excerpt from the study:

Insulin stimulates cell survival and activates proliferation in a biphasic manner at low initial cell densities of T. thermophila in conical culture flasks: at 400 cells/ml it activates proliferation in two separate intervals; down to nanomolar concentrations and again in the low pico- and femtomolar range. The reasons for this pattern are unknown but it is possible that the biphasic response may be due to receptors with different affinity states for insulin as described for insulin receptors in mammalian cell systems (for ref. see Gammeltoft, 1984). If this is true, then the effects of insulin observed at about 10-14 M may be brought about via receptors with a very high affinity to insulin or insulin-related material. One might also construe that the biphasic response may be due to events of an inhibitory/ desensitization action between separate receptor systems. Thus, like e.g. glucagon in hepatocytes (Housley et al., 1987), insulin or insulin-related material in T. thermophila may act through two functionally and presumably structurally distinct receptor populations, which we here call ILR1 and ILR2. ILR1 may be activated in order to stimulate proliferation down to about 10-7 M and inhibit the activity of ILR2 down to about 10-11 M. ILR1 may not be activated at lower concentrations and insulin induces cell survival and proliferation through the activation of ILR2 at about 10-11–10-14 M.

Insulin has some other interesting effects on Tetrahymena cells and we’ll look at these next time.

1. Figure from Søren T. Christensen, Helene Quie, Kåre Kemp And Leif Rasmussen. 1996. Insulin Produces A Biphasic Response In Tetrahymena Thermophila By Stimulating Cell Survival And Activating Proliferation In Two Separate Concentration Intervals. Cell Biology International 20: 437–444.

In the last entry, I raised a puzzle. Why is it that Tetrahymena cells survive and reproduce when a fresh culture contains at least 1000 cells/ml, but die when the culture only contains 100 cells/ml?

The clue was found in the following sentence:

All you need is some media, which would be a solution that contains all the ingredients needed for cell growth, and some cells, obtained from another previous culture.

Note the highlighted portion. When you start up a new culture of cells, you not only introduce the cells into the new media, but also some of the old surrounding media.

What would thus explain the puzzle is this: Tetrahymena secrete some “growth factor” into their surrounding environment, where cells stimulate each other (or themselves) to survive and/or divide. When you transfer one cell into 1 microliter, enough of the growth factor is transferred with it so the cell survives and divides. But if the same cell is put into 10 microliters, the greater volume dilutes the growth factor enough such that it is now insufficient to stimulate survival. The cell dies in a sea of food.

This, by itself, is interesting because it shows that single-celled organisms display a level of inter-dependence that would foreshadow the evolution of multicellularity. In fact, there are many examples of this. Among bacteria, there is a phenomenon known as quorum sensing , where bacteria secrete hormone-like molecules that allow them to coordinate and function as a population.

But as you might guess, this story gets more interesting than this.

Setting up a culture of cells is a relatively simple task. All you need is some media, which would be a solution that contains all the ingredients needed for cell growth, and some cells, obtained from another previous culture. Put simply, you fill a container with media and add a small amount of cells. These cells then do what cells do – they divide and form a large population of cells over time. In other words, the machinery within the cells converts the simple biomolecules in the media into new cells.

But now I have a puzzle for you.

Lets begin by making a media with the following ingredients: amino acids, glucose, vitamins, nucleosides, salts and citrate. Next, let’s transfer a single Tetrahymena cell to 1 microliter of the media. That corresponds to a density of 1000 cells per ml. What happens? The cell does what cells do – it divides and forms a population of cells.

But what happens if you transfer a single Tetrahymena cell to 10 microliters of media (which corresponds of 100 cells per mi.)? Answer – it dies.

So why does this single-celled organism die when it is surrounded by an abundance of food and there are no predators or toxins around?

I’ll give ya the answer.

We have seen that Hydra possesses an insulin-receptor and that insulin obtained from cows has the ability to induce a cellular response in this simple animal. We have also seen that bovine insulin likewise can influence the development of plants.

This all suggests that the last common ancestor of both plants and animals could have responded to mammalian insulin. Yet front-loading would lead us to predict that the same theme would hold true among single-celled protozoa.

About two years ago, I suggested that Tetrahymena might represent a useful model for exploring front-loading. And sure enough, an insulin-like receptor was indeed identified in this organism in 2003 (Søren T. Christensen, Charles F. Guerra, Aashir Awan, Denys N. Wheatley and Peter Satir 2003. Insulin receptorlike proteins in Tetrahymena thermophila ciliary membranes Curr Biol. 13:R50-2).

The researchers used antibodies raised against the phosphotyrosine kinase domain of a human insulin receptor and identified a protein that reacted with the antibody. It is most interesting to me that the receptor localizes to the cilia of Tetrahymena, as this is an angle I will be exploring in a month or so.

They were then able to use PCR to identify the gene, which they named TtPTK1. It was determined that “the TtPTK1 kinase domain has ~30% identity and 65% homology to human, mouse and various invertebrate insulin receptor beta–subunit kinase domains.” In conclusion, the authors note:

Our observations suggest that cilia in Tetrahymena have signal transduction components resembling those of insulin receptor-like systems based on phosphotyrosine signaling. Our results agree with the notion that signaling through phosphorylation via PTKs is not a defining character of metazoan cells, as was previously proposed [13], but is present in unicellular eukaryotes.

This finding clearly shows that the RTKs are not restricted to animals and choanoflagellates and are likely to be found in many other protozoa. But an insulin receptor in a single-celled organism? Like Hydra and plants, might mammalian insulin influence the behavior of this humble protozoan?

We’ll consider this in the next installment.

As readers know, The Design Matrix is neither anti-science nor anti-evolution. What is does represent is a challenge to the non-teleological perspective of biology and evolution. And it is not a challenge in the sense that I try to refute the non-teleological perspective, but that it lays out a positive, potentially fruitful teleological perspective of biology and evolution that embraces random variation and natural selection. What this portends is that a non-teleological perspective is not needed.

Chapter 6 (Ducks and Rabbits) plays a pivotal role in this story and recent research continues to emphasize that theme.

A new paper was published that shows that the act of envisioning reality with our minds shapes what our eyes see.

To test how imagery affects perception, Pearson, Tong and co-author Colin Clifford of the University of Sydney had subjects imagine simple patterns of vertical or horizontal stripes, which are strongly represented in the primary visual areas of the brain. They then presented a green horizontal grated pattern to one eye and a red vertical grated pattern to the other to induce what is called binocular rivalry. During binocular rivalry, an individual will often alternately perceive each stimulus, with the images appearing to switch back and forth before their eyes. The subjects generally reported they had seen the image they had been imagining, proving the researcher’s hypothesis that imagery would influence the binocular rivalry battle.

Additional experiments found that the effect of imagery on perception was approximately the same as showing the research subject a faint representation of one of the patterns between trials. Stronger shifts in perception were found if subjects either viewed or imagined a particular pattern for longer periods of time. They found that both imagery and perception can lead to a build-up of a “perceptual trace” that influences subsequent perception.

This too echoes The Design Matrix:

The findings may also help settle a longstanding debate in the research community over whether mental imagery is visual—that one imagines something just as one sees it—or more abstract.

“More recently, with advances in human brain imaging, we now know that when you imagine something parts of the visual brain do light up and you see activity there,” Pearson said. “So there’s more and more evidence suggesting that there is a huge overlap between mental imagery and seeing the same thing. Our work shows that not only are imagery and vision related, but imagery directly influences what we see.”

We tend to see what we expect to see because this is how our brains work. And remember that both the teleological and non-teleological perspective are just that – perspectives. If one’s brain has been trained to view data from a non-teleological perspective and expectation, this mindset itself becomes a powerful “perceptual trace” and would allow one to assign non-teleological significance to data that may in reality stem from a teleological cause.

Recent research concerning tyrosine kinases continues to strengthen the case for front-loading evolution:

When it comes to cellular communication networks, a primitive single-celled microbe that answers to the name of Monosiga brevicollis has a leg up on animals composed of billions of cells. It commands a signaling network more elaborate and diverse than found in any multicellular organism higher up on the evolutionary tree, researchers at the Salk Institute for Biological Studies have discovered.
[…]
This treasure trove of diverse and novel tyrosine kinases took the study’s lead author Gerard Manning, who heads the Razavi-Newman Center for Bioinformatics, by surprise since it was long thought that tyrosine kinases are restricted to multicellular animals where they handle communication between cells.
“We were absolutely stunned,” says Manning. “Based on past work, we had expected maybe a handful of these kinases but instead discovered that this primitive organism has a record number of them. Two other essential parts of the tyrosine kinase network - PTP and SH2 genes - are also more numerous than in any other genome, showing that it is the whole network that is elaborated here.”
[…]
The Monosiga kinases are more divergent than anything previously seen in animals, which may help scientists understand the fundamentals of how all tyrosine kinase signaling works. Despite their extreme diversity, Monosiga kinases time and again arrive at the same solution to a problem, as do animal kinases, but using a distinct method for instance to create a sensor structure that emerges from the cell, or to target a kinase to a specific part of the cell. “This convergent evolution suggests that there are only a limited number of ways build a functional network from these components,” says Manning.

With all this new information, one obvious question remains unanswered: what is a single-celled organism doing with all this communications gear? “We don’t have a clue!” says Manning, “but this discovery is the first step in finding out.”

That a single-celled organism contains a signaling network more elaborate and diverse than found in any multicellular organism clearly indicates the plausibility of such an ancestral, front-loaded state. What’s more, note that the system has been set up such that similar outputs are reached through convergent evolution. I’ll be commenting on this in more detail a little later.

We have previously seen that insulin obtained from cows has the ability to induce cellular changes in Hydra. This led someone to ask me whether Hydra itself produces insulin. While it is assumed that Hydra must produce some insulin-like molecule to react with its insulin-like receptor, no evidence for such an intrinsic ligand has yet to be discovered.

That bovine insulin can activate a cnidarian RTK is intriguing enough, as it opens up some doors from the perspective of front-loading evolution, but what if we traveled further back in time? Could mammalian insulin have an analogous effect on something that is not an animal?

Here is a link to a review article from several Brazilian labs that argues for the presence in insulin in plants. Let me pick out some interesting excerpts.

First, there is evidence that extrinsically added bovine insulin does indeed influence plant development. The authors first cite some earlier work:

After a long period in which no report is found in the literature of any plant physiological work related to insulin, Goodman and Davis (1993) reported that added insulin, insulin like growth factors I and II (IGF-I and IGF-II) accelerate the post-germinative development of fat-storing seeds (sunflower, watermelon and cucumber). They also measured increased activities of enzymes necessary for the conversion of fat to carbohydrate like fatty acyl CoA dehydrogenase, citrate synthase, malate dehydrogenase, isocitrate lyase, and malate synthase. No mechanism is suggested by the authors to explain this increase in enzyme activities although they hint at the possible increase in protein synthesis. The authors call attention to “the possibility that there are hormones and/or growth factors that have a regulatory role in both plants and animals” and some of these could be insulin-, and IGF-like proteins (Goodman and Davis, 1993).

They then cite some of their own work:

We found by immunofluorescence microscopy analysis that insulin, insulin receptor and phosphoserine proteins are localized to an internal tissue layer of the seed coat but not in cotyledon tissues of C. ensiformis. This region is assumed to be important in sugar transport to the embryo. We then employed bovine insulin to test if it has any effect on germination of C. ensiformis seeds. The results showed that insulin, vanadyl sulfate (an insulin mimetic compound), pinitol (a chiro inositol analogue) and glucose were able to accelerate C. ensiformis seed radicle and epicotyl development and on the contrary, tyrphostin (an inhibitor of insulin receptor kinase activity) inhibited these processes (Oliveira AEA, Ribeiro ES, da Cunha M, Gomes VM, Fernandes KVS, Xavier-Filho J - Insulin accelerates germination and development of Canavalia ensiformis (Jack bean) seeds. Submitted for publication).

C. ensiformis is an annual, semi-domesticated legume with long germination and developmental times making it a less than ideal model for germination and developmental studies. Therefore we utilized common bean (Phaseolus vulgaris) as a more convenient model plant. We showed that increasing concentrations of added bovine insulin (for 72 h) promote an increase in the mass and size of radicles and epicotyls of P. vulgaris and also in the number of lateral roots. Additionally we extracted and purified a protein from embryonic axes (48 h), which cross-reacted with an anti-human insulin antibody (Santos, 2003).

Thus, not only does cow insulin influence development of Hydra, but it also influences the development of plants. And it gets even more interesting.

Our laboratory has mostly been directed to the elucidation of the biochemical basis of bruchid (insect) resistance shown by some legume seeds (Macedo et al., 1993; Fernandes et al., 1993; Xavier-Filho et al., 1996; Sales et al., 2000). As such, investigation of potentially toxic proteins from the seed coat of the legume Canavalia ensiformis to Callosobruchus maculatus (cowpea weevil) led Elenir Oliveira to isolate and purify a number of proteins from this material. One of these purified proteins was submitted to sequencing as an assignment for a training course. The resulting analysis showed unambiguously that the protein had the same amino acid sequence as bovine insulin (Table 1). To control for potential contamination, the analysis was repeated with different samples of the protein obtained from different batches of seeds and the amino acid sequencing analysis was also performed by two independent laboratories. After obtaining a total of seven analyses for the sequence we were convinced that the seed coat of C. ensiformis indeed contained a protein with a sequence equal to that of bovine insulin (Oliveira et al., 1999a). In this manuscript we suggested that molecules of insulin in seed coat tissues survive desiccation after maturation of the seed and, together with other proteins can be easily extracted. The high solubility of these proteins is certainly due to the lack of tannins and pigments in this tissue (Oliveira et al., 1999a; Oliveira et al., 1999b). These insulin molecules (and also a peptide fragment of a receptor-like-kinase accompanying the protein, see Table 2) in the seed coat seemed to be remains of constituents of signaling pathways probably involved in the transport of carbohydrate (Oliveira et al., 1999a). Contrary to our expectancies our results were received with disbelief.

and

We also choose a second fast growing plant, cowpea (Vigna unguiculata) to test for the presence of insulin during development. The protein was detected (by Western blotting) both in empty pods and seed coats but not in the embryo. Insulin was measured by an ELISA assay using an anti-human insulin antibody. The highest concentrations (about 0.5 ng.mg-1 of protein) of this protein were found in seed coats of 16 and 18 DAP (days after pollination) in which case the values were 1.6 to 4.0 times higher than the values found for isolated pods of any day. Insulin was isolated from 10 DAP empty pods by the method of Khanna et al. (Khanna et al., 1976), purified by C4-HPLC and submitted to N-terminal amino acid sequencing. The amino acid sequence was found to be equal to the sequence of bovine insulin and to the sequence of the insulin isolated from C. ensiformis seed coat (see above and Table 1) (Venâncio, 2001; Venâncio et al., 2003).

Here is Table 1:

Yet despite all these data, the authors acknowledge:

We know that up to now no gene sequence was found for insulin in the genome of Arabidopsis (Anon, 2000) or in any other plant genome already published. We do not have any explanation for the conflicting results and the others already referred to above.

This raises the interesting question of conflicting evidence. On one hand, we have a good bit of biochemical and cytological evidence that indicates insulin plays a role in plant development and that some plants not only possess insulin, but their insulin is the same as bovine insulin. On the other hand, sequence data to support this activity does not exist. When it comes to the sequence dilemma, one must ask whether this is this one widespread, yet subtle, contamination problem. Or is this an annotation problem? Or, if you really want to bake your noodle, might some plant genomes code for pieces of insulin sequence, such that insulin is created through elaborate RNA-processing events?

Whatever the answer, what does appear to be on solid ground is the theme whereby a mammalian hormone has the ability to influence not only the development of a simple animal like Hydria, but also of some plants. Yet does the story stop here?

Hydra vulgaris is a member of the phylum Cnidaria. It appears to be a relatively simple animal and has a small number of cell types (you can read more about its basic biology here). Yet, as we have seen, it turns out that cnidarians actually possess a rather complex genetic tool kit.

We have also seen that receptor tyrosine kinases (RTK) would play important roles in facilitating the evolution of multicellular life. Added to this is the recent discovery that one example of an RTK, the insulin receptor, plays an important role, along with its ligand insulin, in the development of the nervous system.

So let’s begin the process of tying this together.

Could it be possible that the protein hormone insulin, that is spread throughout the body of mammals via the circulatory system, would actually play a role in the development or life of Hydra? In 1996, Steele et al. (1) identified a gene for a receptor tyrosine kinase that was very similar to the insulin receptor in mammals, called HTK7. They found that is was expressed in ectodermal cells (the cell type that can generate nervous tissue) at both ends of Hydra’s tube.

But what is most striking of all is that they found insulin, obtained from cows, had the ability to induce both DNA replication and cell division in Hydra’s ectodermal cells.

Okay, from the perspective, there is nothing all that surprising about finding insulin receptors, and responsiveness to mammalian insulin, in cnidarians. This is just another example of deep homology that is consistent with such a system being in place with the last common ancestor of all animals.

What’s more interesting this time around is that we are talking about a hormone and its receptor. Here, the function is simple - BIND. What makes this interesting is that cows and Hydra last shared a common ancestor at least 600 million years ago. This in turn means there are 1.2 billion years of evolution that separate the Hydra insulin receptor and the bovine insulin.

Each lineage would possess an independent history of mutations in the receptor followed by secondary, suppressor mutations in the ligand. Each lineage would possess an independent history of mutations in the ligand followed by secondary, suppressor mutations in the receptor. Yet despite two separate spans of co-evolution between receptor and ligand, the ligand from cows retains the ability to function with the receptor from Hydra.

All that this indicates a fairly strong selective constraint on a seemingly simply biochemical function (BIND). So where do we go from here?

1. R. E. Steele, Pauline Lieu, Ninh H. Mai, M. Andrew Shenk and Michael P. Sarras Jr. 1996. Response to insulin and the expression pattern of a gene encoding an insulin receptor homologue suggest a role for an insulin-like molecule in regulating growth and patterning in Hydra Development Genes and Evolution 206:247-259