EEB 122. Principles of Evolution, Ecology and Behavior?

Lecture 08.?The Expression of Variation: Reaction Norms

https://oyc.yale.edu/ecology-and-evolutionary-biology/eeb-122/lecture-8

Last time we were discussing developmental control genes and the way they lay down basic patterns in body plans. They provide insight into the deep history of developmental constraint and phylogenetic constraint, and they also set up patterns that then interact, during the course of development of individual organisms, interact with the environment to determine what the phenotype actually looks like.?

Today we're going to?talk about developmental plasticity and reaction norms, and in the process we are going to complete our?assemblage?of all of the tools we need to understand?microevolution.?I want you to think about organisms or?genomes at least, as having the potential to produce many different things. The actual thing that is realized depends upon the particular environments encountered, the particular history of that individual organism, and this can have profound effects on the way it looks, the way it behaves, and how long it lives.?

I'll tell you where a reaction norm fits in the evolutionary process; where it came from; how it interacts with genetics; how you can actually visualize the simultaneous effects of genes and environment by making reaction norm plots. That's an important thing. There's been a long controversy in our general culture about Nature versus Nurture. Today I'm giving you the tool to take that issue apart and understand it rigorously. You will end up seeing that all aspects of all organisms are determined both by genes and by environment.?Then I'll show you how this kind of immediate, short-term phenotypic plasticity interacts with developmental control genes and phylogenetic constraints, and I'll do that with the butterfly wing.

A reaction norm is a property of a genotype. What it does is to?describe the set of phenotypes into which one genotype can be mapped, as the environment varies.?

• In the simplest case you have one trait and you have one environmental variable, and this is the way that that genotype, one genotype, would react to this one environmental variable.?Organisms have lots of traits, and there are lots of environmental variables, and so you can immediately see that this simple picture can be generalized into an N-dimensional reaction surface.

• It can get very?complex if we're not just dealing about temperature, but food, population density, presence or absence of members of the other sex,?happening over the whole course of the organism's life.?So each genotype has the potential to end up anywhere along this reaction surface, depending upon the environmental history.

• We can think of the evolutionary process basically as being a cycle that moves between genotype space and phenotype space.?And reaction norms develop at the stage where the genotypes that are present in the just fertilized zygote are being translated into the phenotypes of the adults.

When the gametes are then mapped into genotypes and?when the zygotes are formed--this is the Hardy-Weinberg Law; this is basically population genetics up here. Down here, when the phenotypes that are being produced by the reaction norms then undergo behavior and ecology to determine a surviving set of organisms that can mate and reproduce and have babies, that's natural selection?down here.?

In every generation there's genetics, in every generation there's development, and in every generation there's ecology and behavior. So they're all necessary components of understanding the microevolutionary process.?

These are water fleas,?Daphnia,?that are reproducing asexually. There is?a series of different phenotypes that are all produced by the same genotype, the genotype is being copied exactly, and in the middle of the summer they are producing these helmets and spines.?

• There are spines, helmets and neck teeth in these Daphnia, and they are induced by dissolved molecules that are associated with predators.

• And the predator's efficiency in eating those Daphnia is affected by the production of those spines and helmets?on the Daphnia.

• Making a spine or a helmet has a reproductive cost. So if the predator's not around, you don't want to make the spine, because it's costing you babies. So it is a contingent plastic reaction. You modify the development of your offspring so that they're safer, but your offspring won't be able to have as many babies because they're better at not being eaten. There are bent shells in barnacles that do the same thing; they make them resistant to snail predators, but they reduce the barnacle's fecundity.

• If the cost were not there, then the organisms would make the defensive structure all the time. It was?not?cost-free?and they're forced to compromise, and they try to minimize the cost of the defensive structures by not producing them unless they get a signal that there's danger.

• When Daphnia smell midge larvae?in the water--a midge larva is a little invertebrate predator that swims around and it catches Daphnia with its fore legs,?and if Daphnia makes a little neck tooth, it makes it harder for the midge larvae to handle it. Evidently this neck tooth actually, although it looks very small, cost Daphnia something, because they only make it when they smell midge larvae in the water.?

Barnacles?are essentially shrimp that swim around as larvae and glue themselves to the substrate and spend their lives stuck to the basement, kicking food into their mouth with their feet. So the feet of the barnacle would be sticking out here.?Snails?parasitized?by?castrating?digenetic?trematodes?reproduce earlier. By the way, this digenetic trematode is also called?schistosomiasis. So this snail is an intermediate host for a serious human disease.?

• If the barnacle smells the snails, when it's growing up, it grows up in a clumped-over form. It bends as it grows, and instead of feeding freely out of the top of its body, it feeds very inefficiently, and it pays a price in not being able to make more babies.?

• This is the data on schistosomiasis, and the neat thing about this experiment is that the reaction in the snail is induced just by water in which there have been parasites, not by the parasites themselves. In other words, you just give the snail a whiff, a little bit of scent that a parasite is likely to get into its body, and its reaction is that "I am going to die from a parasite, so I better start reproducing." So it shifts its reproduction earlier in life; and you can see the extent of that shift right here.?

So these things that I'm describing are all induced responses, they are all plastic reactions to signals in the environment, and they all shape the reaction norms of these organisms.?

Here is one reaction norm. I have sketched a common one for many poikilothermic or cold-blooded?organisms; they don't regulate their body temperature. The higher the temperature, the smaller they are at maturity.?This general relationship describes how tadpoles grow, how many fish grow.?

If you look at a population, it can be conceived of as a bundle of reaction norms. So there are many genotypes out there. So if we have a very small population with just about five or six genotypes in it--the green dotted lines are the individual reaction norms for the different genotypes, and there could be a population mean reaction norm. You just calculate the mean value across all the environments and all the genotypes, and that describes how that population responds.?

This is important when you're trying to summarize this kind of complexity in ecology. You want to know how one population might react as the environment changes, so that you can analyze its impact on another one: a predator acting on a prey; a parasite acting on a host; a grazer acting on a plant.?

Traits can have very different expression patterns; it's not as though all traits have very dramatic reaction norms. And I just chose the example of five digits in many tetrapods, including ourselves, to indicate that you could have three different genotypes, and you could change population density a lot, and the number of digits on the hand wouldn't change. Everybody would have five fingers. There are some things that are just not sensitive to the environment. So think of the individual organism as a mosaic of sensitivities. Some of it is not sensitive at all to changes in the environment, or almost insensitive, and other parts of it are quite sensitive.?

For example, fecundity. If you increase population density, fecundity will go down in individual organisms, because they're having to compete harder to get food. So if you restrict food, by any mechanism, fecundity will drop--and an increase in population density is one way to do it--and the genotypes in the population can react differently to that increase. In all three cases fecundity decreased, but genotype 1 was quite sensitive, and genotype 3 was much less sensitive to the shift in population density.

As a matter of fact,?if you have a fluctuating population, and this population is going between low density and high density, you have a method of maintaining genetic variation right there, because the reaction norms cross, and the guys that were good at one density are lousy at the other. So if the population cycles back and forth between them, one time G1 is favored, the next time G3 is favored, and so forth.?

For example, if you have this sort of a reaction norm pattern for four genotypes, and you select upward here, you're going to lead to no response over here at all, because they all happen to converge at this point. So selection here doesn't make any difference to what you observe in this part of the environment. But in this case, the crossing reaction norm case that we had in the last picture with fecundity, if you selectupward in this environment, you're going to have a downward response.?

If we select at low population density, and population density is low for a long time, it's going to produce a shift in the population over here, because G1 will be favored, and it has low fecundity at high density.?

We can just look at a sketch of a reaction norm and we get a sense for how sensitive that trait is to changes in the environment. This is not a very plastic trait, it's pretty insensitive, and we can see that because it has a shallow slope. This trait's very sensitive. You change the environment a little, it changes a lot.?

Now it's not just spines and helmets that have reaction norms. This is a picture of an Affymetrix GeneChip for Drosophila melanogaster--it's got 13,500 genes--and what the chip is doing is it's picking up the messenger RNA, which is being expressed in the organism; and the intensity of light that you see at a given spot is a measure of the concentration of messenger RNA for that particular gene.?These things have reaction norms and?these concepts are general. They're not limited to morphology. They apply to any aspect of the phenotype, and this is now a very popular way to measure phenotypes.?

So to sum up on reaction norms.

• A reaction norm is a description of how genes are mapped into the phenotype as a function of the environment. They are properties of genotypes.

• If you really want a proper, rigorous way of measuring a reaction norm, you have to be able to clone the organism, so you can get the same genotype replicated and then test it in different environments.?Identical twins are probably?as far as you can go?in humans.?But in Daphnia, or in plants, it's possible to get genotypes replicated, up to a hundred individuals sometimes, and then you can make a very accurate measure of a reaction norm.

• You can think of a population as a bundle of individual reaction norms; and that's an important concept because when we come to ecology we're going to be thinking about how predators interact with prey, and about how competitors interact with each other. And when we do that, normally the way that biologists have done it in the past is they've chunked those things as species, where they have a species typical property.?So all the species 1 are supposed to behave one way, and all of species 2 are supposed to behave another way.

• But the differences between the individuals in those species are really important, and when the two species are interacting, it's not like they're all identical individuals interacting. They are different, and when the species interact it's bundles of reaction norms interacting with bundles of reaction norms. And this produces important effects. For example, it tends to stabilize ecological interactions.?

There's a real easy way to talk about the sensitivity of phenotypes to the environment. You just make a reaction norm plot and look at the slope. If the slope is steep, those organisms are very sensitive to changes in environment; if it's flat, they are not. ?

I've been talking a lot about phenotypic plasticity, and I've shown you these wonderful examples of Daphnia reacting sensitively to predators and so forth. Does that mean that organisms are really plastic? Can I just pick up a bunch of clay and mold it into anything that I want, depending on the environment that I expose it to? No I can't.?

And that's because?the large-scale structure is determined by things that are hard to change, and those are developmental patterns that have a deep evolutionary history, and they set up a rigid framework within which the plasticity is expressed. So the things that change slowly--those are the developmental control genes--are constraining the things that change rapidly.

So let's do this with the example of Distal-less. Distal-less is a developmental control gene.?

• The pictures here basically are showing you how the Drosophila larva gets set up very early in development. The first thing that happens is that an anterior/posterior axis gets laid down. That's done by the Hox genes. Then the dorsoventral axis is determined by Sog and Chordin and Decapentaplegic and things like that.

• Then, after the basic axes of the organisms are laid down and segments are formed, other things turn on that determine whether you'll be dealing with a head, a gut or a tail. Interestingly, the name for the gene that induces heart formation is Tinman, from the Wizard of Oz, who didn't have a heart.

• And what we're worried about today is this gene here, Distal-less, which determines body wall outgrowth. And if you look at the body of a fly, this is where the action of certain mutations takes place. If you get mutations in Distal-less, these are the parts of the body which are going to be affected. They are all extremities, all out-pocketings of the body wall, which are then being developed into antennae or mouth parts or legs.?

Now in order to tell you about this deep developmental constraint in butterfly wings, I first want you to notice that there's something that's called a Nymphalid groundplan.?

• The Nymphalidae are a large family of butterflies,?in the middle of the wing you could have stripes; in the outer part of the wing you could have what they called border eyespots, or border ocelli; right on the edge of the wing you could have bands, and so forth.?And we're going to focus on the eyespots.

• And you can see that simply by varying the location where colors are expressed, and by varying the size of the different elements, you generate a huge number of patterns. You can even use them to write numbers on wings. Evolution has written numbers on the back wing of this particular butterfly; this is an '89 butterfly.

• The model system in which this is best studied is in a butterfly called Bicyclus.?Bicyclus has a number of neat features. One of them is that it is developmentally plastic.?In the wet season it looks like this, and in the dry season it looks like this. And, in fact, these are two brothers who have been produced in the laboratory, with this one being raised under wet season conditions and this one being raised under dry season conditions. So one genotype can elicit a range of phenotypes, and you can see that in the process the eyespots change considerably in their size and intensity.?

Now it turns out that you can fish the Distal-less gene out of Drosophila, and you can use that segment of DNA to recognize the homolog gene in the butterfly, and you can then put a reporter onto the homolog, and you can ask that gene to express its reporter when it's being expressed, so that you can see visually where the gene's being expressed. When that's done, you can see that every place that an eyespot is going to form in the adult wing, you can see the gene being expressed in the wing disc, in the developing pupa.?

• The way that butterflies and flies and other holometabolous insects develop is that after the caterpillar or the larva has fed for awhile, and it's starting to form its pupa, the cells reorganize in the pupa, into structures that are going to be parts of the adult, and the wing disc, that's going to be the wing in the adult, looks like this in the pupa, and it's sitting right on the surface of the pupa.

• So this is actually an exceedingly neat system to work in because you can actually do cell manipulations, as well as genetic manipulations. You can manipulate both the developmental biology and the underlying genetic structure, in butterfly wings.?

• So Distal-less is actually telling the wing disc where to make eyespots, and the Nymphalid groundplan says you can only make those eyespots in certain places. And the Nymphalid groundplan, the butterfly wing groundplan, is arguably about 100,000,000 years old; it's ancient. So does that mean that you can't change the eyespots? No it doesn't.

• Almost everything about the eyespots has a reaction norm, except their location and number. Within a given species you're always going to get the same number, and they're always going to be in the same place, but whether they're big or so small that you can't even see them depends on the environment in which they're expressed.?

Can we think of macroevolution as having constructed a vase, within which the reaction norms sit? And the answer is no.?

• And the answer is no because some of the genes that are controlling the shape and the position of the eyespots--so things like Distal-less--are also involved in determining the slopes and the shapes of the reaction norms. These two things are genetically entangled, and their entanglement is a case of the same gene having two different functions at different times in development, and natural selection will operate on it throughout the lifecycle.

• So it's not as though there are some things that are constraints, that are not being changed, and there are other things that are genes that are sort of tweaking the constraints a little bit. In fact, the same genes are involved in producing both things. So if we want to shift the slope of the reaction norm by selecting on phenotypic plasticity in Bicyclus, we are going to be selecting on genes that are also determining the location and number of eyespots.?

To summarize my overview of it, what I want to emphasize is that the phenotype, the whole organism that you see, and the whole lifecycle of that organism that you see, is a mosaic of parts, and their pattern of determination varies tremendously in evolutionary age.?

• So if you just look at my own body, the parts of me that are extremely old are the fact that I have four limbs and five fingers.

• And the parts of me that are evolutionarily relatively young are the size of my cerebral cortex and some other aspects of me. If you were to look into the plasticity of my cerebral cortex, you would discover that it is incredibly plastic, and that when I am a little baby and I'm just born, I have billions more connections in my nerve cells than I do when I'm seven-years-old, and that a great deal of my mental development, between birth and the age of seven, has essentially been the remodeling of my cortex by plastic interactions with the environment. And in fact that's what a lot of learning is about; it's about plastic response to environment. So I am myself, as are you, a mosaic of things, of very different evolutionary ages.

• The basic developmental patterns that we see in animals are mostly about 500,000,000 years old. In plants they're a bit younger. The HOX control of body symmetry and body pattern in animals is arguably about 600,000,000 years old; maybe a little less, maybe 550,000,000. The ABC pattern of flower development in flowering plants is probably somewhere between about 95 and 135,000,000 years old; that's something that happened in the Cretaceous.?

Now let's shift timescale and go down to one generation, one organism, encountering a specific environment. Its plastic reaction to the environment has evolved relatively recently, and it implements specific contingency plans. Daphnia that come from lakes that do not have fish in them and haven't had fish in them for a long time, don't react when you put the smell of a fish into the water. The Daphnia that come from lakes that have had fish in them for along time react, and react strongly and quickly. So the plastic reaction is something that can evolve.?

I want to caution you though, it is not as though all the fine details of the plastic response are adaptive; they are not necessarily all adaptive. For example, think about temperature. If we are studying the plastic reactions of organisms to temperature, it may very well be that things that live in the Arctic have a different reaction norm than things that live in the tropics, because they've encountered a different temperature regime, and that that's an evolved reaction. But it's also quite possible that it's just biophysically impossible to do something when it gets colder; that doesn't have to evolve.?

So I want you also to be able to think of the necessity of taking something like a plastic reaction norm and dissecting it analytically so that you can figure out what part of it's adaptive and what part of it is just there because that's what?organisms are built out of. They are biochemical systems, and biochemistry?has reaction rates that change with temperature and with a lot of other things.?So it's?not all adaptive. Everything that you see in organisms has an evolutionary history. It doesn't have to be an adaptive history. It might be drift. Things might happen in phenotypes that are byproducts of stuff that's going on somewhere else in the organism.

The thing you actually see, the organism you analyze, is just one point on a multidimensional reaction surface. It could have been a lot of other things, and all those other things that it could have been are important when we think about evolutionary ecology, when we think about population dynamics, when we think about interactions between hosts and parasites, because they represent all those other potential interactions that could be going on in other circumstances.?

So by thinking about reaction norms, we can express the genetic variation in the population,?the developmental reaction to the environment and?the way all of those different genetic combinations will react to the environment; and we have the potential to visualize the dynamic over generations, as both the gene frequencies and the environmental circumstances change.?

Every phenotype is the product of both genetic and environmental influences, and the way they interact to produce the phenotype is extremely important. So it is almost never the case that you can claim that only Nature, or only Nurture, accounts for what you see in organisms.?