Session 3: Evidence of Evolution

Transcript of Part 3: Genetics of Morphology

00:00:00.04	Hi, my name is Hopi Hoekstra and I'm a professor at Harvard
00:00:03.10	University. And in this second segment, what I would like to do is
00:00:06.15	tell you a story about the genetic basis of evolutionary
00:00:09.22	change. And in particular, we're going to focus on the story
00:00:12.06	of a morphological trait. I'm going to tell you this story in the context
00:00:17.15	of making links between environment, phenotype, and
00:00:20.08	genotype. In particular, I'm going to tell you today a three-part story. The first
00:00:24.18	part of the story, I'm going to tell you about a role for natural
00:00:28.09	selection in driving differences and traits in a natural population.
00:00:31.14	And then I'm going to take you from the field into the lab,
00:00:34.11	where we've been doing work to understand the genetic basis
00:00:37.06	or the genes that control this phenotypic difference. And it's
00:00:40.12	not just that we want to know the genes, but I'll tell you
00:00:42.07	about how changes in these genes through development
00:00:44.25	actually produce variations in the phenotype. And then once we
00:00:48.02	made all the links, then we'll have a more complete understanding
00:00:50.14	of the process of adaptation. And here's where I think things can get
00:00:53.24	really fun. Because we can go back out into the wild
00:00:56.22	and ask how these traits evolved in natural populations. So let's
00:01:00.10	start today at the level of phenotype. So the phenotype I'm going to
00:01:03.23	tell you about is one that we've been studying in my group for almost
00:01:06.03	a decade. And that is color variation. So why do we study color?
00:01:10.25	Well, color is one of the primary ways in which organisms
00:01:14.11	interact with their environment. And it varies tremendously
00:01:18.04	between species, and it can also vary between species. And in
00:01:22.04	particular, we know at least in some cases, even small changes
00:01:25.03	in color difference can have a huge impact on fitness.
00:01:27.22	The ability of organisms to reproduce and survive in the wild.
00:01:31.03	So here I've just given you a number of examples of how
00:01:34.21	color is used. It can be used, for example, in terms of reproduction
00:01:38.21	in that it can be used for mate choice. So we have the canonical example of
00:01:41.29	the peacock's tail. But flowers also use color to attract pollinators
00:01:46.26	which is the way that they reproduce. Color can also be
00:01:49.20	used not to attract other species or other individuals,
00:01:52.18	but it can be a way to warn them of let's say a distasteful
00:01:56.29	poison. As is the case in these poison frogs.
00:02:00.29	Now, no good biological story comes without having
00:02:04.16	cheaters. So we also have mimics, those species that
00:02:08.14	are themselves not toxic, but mimic those that are toxic.
00:02:13.09	And thereby are also avoided by predators. But by
00:02:17.00	far, color is used most commonly in terms of camouflage or
00:02:21.23	in terms of crypsis. So that's what I'm going to focus on today.
00:02:25.00	In addition to being ecologically relevant, the other reason we
00:02:29.23	study color is because we can rely on nearly a decade's worth of work
00:02:34.22	by geneticists who have been tracking down genes involved
00:02:37.16	in spontaneous mutations that occur in laboratory populations
00:02:40.28	of mice. So here what I've done is I've made a list, which again
00:02:44.04	I don't expect you to read, but just to appreciate that we know a lot
00:02:47.25	of genes. And if you mutate that gene, it's going to have an effect on
00:02:51.08	color. And here are some of the mutant color differences
00:02:55.00	that have spontaneously arose in laboratory colonies of
00:02:57.26	mice. So in some sense, we have a little bit of a headstart because
00:03:00.26	we know a number of what I'll refer to as candidate genes,
00:03:03.29	genes that contribute to color differences in mammals, but may
00:03:08.05	also contribute to natural variation in color. Now unfortunately,
00:03:12.06	laboratory mice or their wild equivalents, while very common
00:03:15.14	in nature, for example, in our houses, vary a lot in laboratory
00:03:19.23	populations. They don't actually vary that much in the wild.
00:03:22.15	So, instead of studying these mice, we've decided to do
00:03:25.17	a study on a closely related group of mice, mice in the genus
00:03:29.20	peromyscus called deer mice, shown here in their slightly
00:03:32.21	flattened form. So this is a common way that mice are kept
00:03:36.25	in museum collections. And it nicely illustrates the variation in
00:03:40.24	color that I'm going to talk about today. So in this first image,
00:03:43.28	you can see that these deer mice vary tremendously in their
00:03:47.00	dorsal coat. They vary from almost completely white to
00:03:50.23	almost completely black. But they also vary not just in the
00:03:54.28	pigments and individual hairs on their dorsal coat, but they
00:03:57.19	also differ in the distribution of pigments across the body.
00:04:00.10	Or variation in color pattern, as shown here. So these candidate
00:04:05.04	genes that I showed you in the previous slide, many of those
00:04:07.23	are implicated in variation in dorsal color. But in fact,
00:04:11.04	we know very little about how patterns are made. So by studying
00:04:14.10	these mice, not only can we take advantage of this vast
00:04:17.00	range of variation in the color, but we also may learn something
00:04:20.02	new about color patterning. So you may have already noticed
00:04:24.09	that these slides were both taken by someone called Sumner back
00:04:28.07	in the 1920's. Well, this is Francis Sumner, shown here.
00:04:32.09	In his field regalia. This is actually the outfit that he would wear
00:04:36.16	when he would go out into the wild and trap mice. He's a
00:04:39.10	classic natural historian, he was associated with initially the
00:04:42.13	University of Michigan natural history museum. And he spent
00:04:46.29	a bulk of his career trying to answer the question of why
00:04:50.10	populations vary so much in nature. And to do this,
00:04:53.25	he would drive around the U.S. and catch mice, and document
00:04:57.20	variation in a number of traits, including color variation.
00:05:02.17	So Francis Sumner's favorite species of deer mouse, and
00:05:05.28	mine as well, is a particular species call peromyscus polionotus.
00:05:09.17	Peromyscus polionotus is often referred to as the old field
00:05:13.19	mouse. And that's because much of its range in the southeastern
00:05:16.27	U.S., as seen here, so Alabama, Georgia, South Carolina,
00:05:20.08	and Northern Florida. It occurs in these old fields, which are
00:05:24.13	really overgrown agricultural fields. Now throughout
00:05:28.13	most of its range, it may look to you like a typical mouse.
00:05:31.13	It's got a dark brown coat, a gray scruffy belly, and a striped
00:05:35.01	tail. But what's particularly interesting to us about these mice
00:05:38.19	is that they've recently invaded these sandbar islands and
00:05:44.15	sandy dune habitats on the Gulf coast of Florida, as well as
00:05:48.03	the Atlantic coast of Florida. So each one of these
00:05:51.16	numbers on this map refers to a different subspecies
00:05:54.13	of what I'll refer to as "beach mice," because the mice actually live
00:05:57.20	on the beach. Now, the first part of the story, I'm going to
00:06:00.29	focus on one of these subspecies. Here number 3.
00:06:04.04	And that is the Santa Rosa Island beach mouse. So let me show you
00:06:07.06	a picture of their habitat. So unlike their mainland counterparts
00:06:12.02	that live in dark loamy soils, these mice live on these beautiful
00:06:15.14	sandy islands off the Gulf coast of Florida. And here's a picture
00:06:20.03	of one of our field sites. So you'll notice that there are two
00:06:24.05	dramatic differences in habitat that these mice occupy.
00:06:26.25	So first, you can tell that the soil, in this case, this granulated
00:06:33.02	sand that's almost like walking on hills of granulated sugar.
00:06:36.07	It's much lighter in color than the dark loamy soils of the mainland
00:06:40.00	subspecies. But in addition, these beach habitats also have
00:06:45.17	much less vegetative cover, so these mice are exposed to
00:06:48.11	really high levels of predation. And I'll tell you more about
00:06:51.05	their predators in a minute. So it may not be surprising then,
00:06:55.01	when we go out and catch mice in these beautiful beaches,
00:06:58.08	the mice look different. So here's a picture of one of those beach
00:07:01.29	mice. And I should mention that this is not to scale.
00:07:04.02	So both of these mice are about the same size and they're about
00:07:06.23	the size of a ping pong ball. But what this slide does serve to illustrate
00:07:10.17	is the dramatic differences in pigmentation. So this particular
00:07:13.25	mouse, you can see is lacking pigment on its nose,on its sides, and if you could see the tail, it's also missing that
00:07:19.17	strong tail stripe. The other thing I want to mention to you
00:07:22.23	about this system is that we know something about the geological
00:07:25.14	age of these islands. They're about 6000 years old, which suggests
00:07:30.12	that the difference in coat color that you see here may have
00:07:33.17	evolved in just a few thousand years. So you may all be
00:07:37.24	thinking that this makes perfect sense, right? That these mice are
00:07:41.00	running around in these beautiful white sand beaches and having
00:07:43.04	a light color coat would make them more camouflaged.
00:07:46.15	And that's a great idea, but we wanted to actually prove that.
00:07:49.23	So as scientists, what we wanted to do is empirically
00:07:53.00	demonstrate that color matters for survival. So actually
00:07:56.07	do an experiment. We wanted to know how much it matters,
00:07:59.03	in other words, we want to estimate the strength of selection.
00:08:01.27	How favorable is it to actually match the background?
00:08:05.17	And then finally, we wanted to know the agent of selection.
00:08:07.25	Who's doing the selection? And in this case, who are the
00:08:10.17	predators? So first I want to tell you about the experiments we did
00:08:15.10	to try to make this link between color variation and the differences
00:08:18.25	in environment that these mice live in. And in particular, implicate a role
00:08:22.09	for natural selection. Now if I could do any experiment
00:08:25.13	in the world, here's what I'd love to do. I'd love to catch
00:08:28.25	let's say 100 light mice and 100 dark mice, maybe give them all
00:08:32.16	a tag. And let's say release half of them, equal numbers of
00:08:36.11	light and dark, in dark habitat. The other half, equal numbers
00:08:40.06	of light and dark, in light habitat. And then come back, let's say a few
00:08:43.21	months later, and see who survived. And I'd have the expectation
00:08:46.22	that mice that are lighter would survive better in light habitat,
00:08:49.11	and those that are darker in dark habitat. Now for a number of reasons,
00:08:53.03	that experiment is quite hard to do. So instead, what we
00:08:56.16	did is what I'd argue is the next best experiment. And in some ways,
00:08:59.17	it may be even better. So here's the experiment that we actually
00:09:02.22	did. Instead of using live mice, we made mice. So here's
00:09:07.05	a picture of my postdoc, Sacha Vignieri, who along with
00:09:10.05	an undergraduate from Harvard, Joanna Larson, made
00:09:13.09	hundreds of plasticine mice. Half of them were painted
00:09:17.13	dark to mimic the mainland mice, and half of them were painted light
00:09:20.18	to mimic the beach mice. Now this experiment in some
00:09:24.28	ways, as I mentioned may even be better than using live
00:09:27.20	mice, because here the only difference in these mice is
00:09:31.04	their color. So they're made from the same mold, they all look the same,
00:09:34.09	they all smell the same. Whereas with live mice, the beach mice
00:09:37.09	and mainland mice may differ in let's say smell, in escape
00:09:39.28	behavior, in activity patterns. So here, the experiment completely
00:09:44.08	focuses on the difference in color and not correlated traits.
00:09:47.05	But the downside of this experiment is would it actually
00:09:50.18	work? Could we actually fool predators into attacking these
00:09:54.00	plasticine models of mice and not live mice?
00:09:56.29	Well, I wouldn't be telling you about this experiment if it didn't actually
00:09:59.26	work. So what Sacha did was she released equal numbers
00:10:03.22	of these light and dark mice in both light and dark habitats, where
00:10:06.20	live mice of the species actually occurred. And then counted
00:10:10.29	predation events. So here's a predation event. So what you're
00:10:13.29	looking at here is a dark mouse that was put out on light
00:10:16.16	soil. And you may notice that it's missing part of its
00:10:20.03	left ear, and it's got a big chunk taken out of its back.
00:10:23.05	And this is a classic predation event, and in fact, not only
00:10:26.26	can we tell it's been attacked, but we can actually say
00:10:30.08	something about who's doing the attacking. Because these marks
00:10:33.02	are consistent with an avian predator. So here's the results of
00:10:38.10	the larger experiment that Sacha did, where she was counting
00:10:41.14	the number of predation events in different habitat types.
00:10:44.23	So, what you're looking at here are the results of this experiment.
00:10:50.08	So let's first focus on this far panel at the light habitat.
00:10:54.17	What you can see is that there's cryptic and what we'll refer
00:10:57.11	to as non-cryptic mice. And these bars indicate the relative level of
00:11:03.01	predation in both of these two types of mice in this particular habitat.
00:11:06.28	And what you can immediately see is the level of predation here in
00:11:10.19	these cryptic mice is much lower, in fact, it's about half
00:11:14.00	of non-cryptic mice. So what this means is that both mice
00:11:17.22	were attacked by predators, but the mismatched mice were attacked
00:11:21.15	about twice as often. Now when we look at the dark habitat
00:11:24.28	we see a very similar pattern, but in reverse. Here we can
00:11:28.17	see the dark mice survived much better, and while still attacked,
00:11:33.22	they were attacked about half as often as the mismatched mice.
00:11:36.24	So what this first thing tells us is that color seems to matter.
00:11:40.10	And in fact it matters a lot. We can take these numbers and sort
00:11:44.16	of translate that into a selection intensity. I'm not going to go
00:11:48.12	into details about this, but let me just say that color matters
00:11:51.17	a great deal for these mice. And in fact, mice that match their
00:11:55.21	habitat have about a 50% increased chance of survival
00:11:59.21	compared to those that are mismatched. And the final thing
00:12:02.27	I want to say, as I mentioned, we can tell in some cases
00:12:05.08	who's doing the predating. And about half the attacks
00:12:09.16	were due to avian predators, and about half of the attacks
00:12:12.14	were due to mammalian predators like coyotes and foxes.
00:12:16.03	So together, what this experiment is telling us is that
00:12:20.06	the differences between color variation are tightly linked
00:12:22.17	to the environment and that it's natural selection that
00:12:26.01	is playing a role in driving these color differences. So now that
00:12:30.18	we've implicated a role for natural selection, the next thing I
00:12:34.00	want to do is to take you from the field into the lab,
00:12:36.15	where I'll tell you about how we're going about identifying
00:12:39.15	the genes that are responsible for these differences in adaptive color
00:12:42.12	variation. So the first thing I want to do is give you a general sense
00:12:47.23	or an overview of the approach that we're taking. And I don't
00:12:50.22	want you to get caught up in the details, but more appreciate
00:12:53.01	this general approach. So what we're able to do is
00:12:57.05	take these mice from the field and bring them into the lab, so
00:13:00.13	we have both dark mice and light mice. And they have
00:13:03.27	differences in their genomes, in their chromosomes, so
00:13:06.24	I'm going to illustrate this by the dark mice having dark chromosomes
00:13:09.27	and the light mouse having light chromosomes. And what we can
00:13:12.29	do is take a dark mouse and a light mouse, one male and one
00:13:15.05	female. Put them in a cage together and they'll actually
00:13:17.11	reproduce. And they'll produce what we refer to as hybrids.
00:13:21.04	Then we take those hybrid individuals, and we breed them
00:13:24.20	together. And then what happens is in this next generation
00:13:27.16	is their genomes get shuffled. So some individuals are going to have
00:13:31.16	different parts of their chromosomes that come from the light parent,
00:13:34.08	and some from the dark parent. So we're effectively shuffling
00:13:38.06	up their genomes. Now in this population, this second generation of hybrids,
00:13:42.13	what we do in all those individuals is we measure their coat
00:13:45.16	color pattern, and then using genetics or molecular biology,
00:13:49.16	we're able to sort of characterize their chromosomes and determine what regions
00:13:53.08	come from each of the parents. And then what we do is
00:13:56.16	-- I'll just simplify and say we do a nice statistical analysis
00:13:59.20	and ask, are there regions of the genome that seem to be
00:14:03.05	correlated with different aspects of different color variation?
00:14:06.20	So for example, in these chromosomes here, do all these mice
00:14:10.08	that have the light allele from this parent in this region
00:14:15.04	of the chromosome, if all those mice have, let's say, light
00:14:18.13	tails. That suggests that a gene controlling tail color
00:14:21.20	may reside in this part of the chromosome. And this is what we
00:14:25.11	refer to as a QTL analysis, or quantitative trait locus analysis.
00:14:29.03	But it's really just this simple statistical association.
00:14:32.10	And then what we do with this statistical association is that
00:14:36.06	we take that region of the chromosome, and we lookfor its homologous region, the same region in the mouse
00:14:43.01	genome, which we have the complete genome sequence.
00:14:45.12	We know where all the genes are. And we look for
00:14:48.02	candidate genes, remember that list of candidate genes
00:14:51.07	that I showed you earlier in the talk, and ask do any of those
00:14:54.04	candidate genes fall within this particular chromosome region.
00:14:57.16	With the ultimate goal of finding a mutation in that gene
00:15:01.04	whether it's in the protein structure of that gene or
00:15:04.10	maybe a mutation that controls the regulation of this gene,
00:15:07.21	that we can then link to the color differences between the parents.
00:15:10.24	So now that I've given you this overview, next what I want to do is walk
00:15:14.02	you through each of the pieces. So as I mentioned the first
00:15:17.02	thing we do is this cross. So we brought mice in from the wild,
00:15:21.15	we brought in a mainland species and one of these beach
00:15:24.09	mouse subspecies. And here they are, shown again in their
00:15:26.28	flattened form. You can see a dark parent, and over here
00:15:29.13	the light parent. And here is that F1 hybrid shown here. And you can
00:15:34.16	already see this F1 hybrid has traits from both of the parents.
00:15:38.26	So for example, it doesn't have a tail stripe like the light parent
00:15:41.29	but it has a fully pigmented face like the dark parent.
00:15:45.01	That suggests both dominant and recessive alleles
00:15:49.01	contribute to this light color adaptive beach mouse
00:15:52.23	phenotype. Then as I mentioned, we take these F1 hybrids and we breed
00:15:56.06	them together, and that's when we get these F2s.
00:15:58.23	And this is where we've now with recombination shuffled
00:16:02.00	up their genomes. So, for example, presumably this mouse
00:16:05.22	has more pigment alleles from the light parent, and the mouse over here
00:16:10.08	has more pigment alleles from the dark parent. But what you can
00:16:13.05	see is that there's a continuum of variation. And what this
00:16:16.11	immediately tells us that this color difference we observe
00:16:19.11	between beach mice and mainland mice is not controlled by
00:16:22.06	a single gene, but in fact is controlled by a handful of genes.
00:16:26.10	And the reason I say a handful and not hundreds is because
00:16:29.04	you can see in this population we get mice that look like
00:16:31.15	the dark parent, and we get mice that look like the light parent.
00:16:34.16	And this suggests there are not hundreds of genes, because it would
00:16:37.26	be a very small chance that we would get all the light
00:16:40.11	-- all hundred of those light alleles in one individual.
00:16:43.06	But instead, probably just a handful. And from this variation,
00:16:46.27	we estimate there are about 3-5 genes. But I'll tell you more about
00:16:50.05	those genes in just a minute. So as I mentioned, we then take
00:16:53.19	the color variation in these mice and we measure them in
00:16:57.01	all these individuals. And then we also genotype them to
00:17:00.08	figure out what regions of the genome come from the light and dark parent.
00:17:03.06	We did this using molecular techniques. And here's the results
00:17:08.10	of that. So what this is is each one of those lines shown
00:17:11.29	here represents a chromosome. And each one of these markers is
00:17:15.14	a difference between the light and the dark parent.
00:17:17.12	So in each one of these markers we can tell whether a particular
00:17:20.08	individual has that region of the genome comes from the dark
00:17:23.05	parent or the light parent. Then what we do is the statistical
00:17:27.05	analysis. And what we found was, there are three regions
00:17:30.22	of the genome, which I've highlighted here, that seem to
00:17:33.27	control color. That is there's genes in these regions of
00:17:36.15	the genome that control different aspects of color patterning.
00:17:39.25	And lucky for us, in each one of these regions, there's one
00:17:44.17	of these candidate genes that I mentioned earlier.
00:17:47.00	Oh, I should also mention that the differences in the size
00:17:50.12	of the arrows reflects the amount of variation that a particular
00:17:53.14	locus explains. So the Agouti locus way over here, explains
00:17:58.12	a larger proportion of variation compared to Corin, which explains
00:18:02.05	the smallest amount. So what this is telling us is that there's
00:18:06.06	3 regions of the genome, and each one of those contains
00:18:08.29	what I'll refer to as a candidate gene. So next what I'd like
00:18:12.13	to do is just tell you about what of these genes. And then I'll
00:18:15.22	summarize and tell you about all three of them at the end.
00:18:17.28	So the gene I'm going to focus on is this gene up here,
00:18:21.07	Mc1r, or the melanocortin 1 receptor. And one of the
00:18:24.26	reasons we focused on this gene is because we actually
00:18:27.13	know a lot about the structure and the function of this gene.
00:18:30.03	So, the melanocortin receptor is a classic g-protein coupled
00:18:35.24	receptor. That is, it's found in the membrane, and it's got
00:18:39.18	extracellular and intracellular regions. Each one of these
00:18:43.22	little circles represents a different amino acid, and what
00:18:46.06	I've done is I've color coded those amino acids that we know
00:18:49.23	when you change that amino acid, it has an effect on
00:18:52.29	color, and on particular species. And those that make
00:18:56.20	-- when you make that change, an individual darker, it's shown
00:18:59.24	in black and those lighter is shown in gray. So you can see
00:19:03.12	summarized over a number of different studies, that
00:19:07.13	there are multiple mutations in this receptor that can affect
00:19:11.06	color. And it can either make an individual lighter or darker.
00:19:14.20	So the first thing to note is that there are many different
00:19:17.17	mutations. The second thing to note is that they're found throughout
00:19:20.28	the receptor. And the third thing to note is that even a single
00:19:23.27	amino acid mutation can have an effect on color. So one change
00:19:27.26	can have a big effect on phenotype. Now, the first thing we did is
00:19:32.29	we sequenced this gene in both the mainland and the beach
00:19:37.24	form, and asked are there mutational differences between
00:19:40.15	the two? And in fact, we found one and it's highlighted
00:19:44.03	here in red. And the mutation is a single nucleotide change that
00:19:48.03	caused the change in amino acid, whereas the mainland species had
00:19:51.20	an arginine at position 65, beach mice had a cysteine.
00:19:56.02	And this is a charged changing mutation, so it actually
00:19:58.21	changed the charge of that amino acid so that it's more likely
00:20:01.06	to have an effect on the structure and function of melanocortin
00:20:04.14	1 receptor. Now the unfortunate thing, though, was that it didn't
00:20:09.03	overlap with any of the other mutations that had been previously
00:20:11.07	characterized in other species. But that's okay because we can do an experiment
00:20:14.18	to test whether this mutational change had an effect on the way
00:20:18.03	that this receptor functions. So first thing I want to do is tell you a little bit
00:20:21.25	about what this receptor does. And then I'll tell you about our
00:20:24.25	experiment. So what I'm showing you here is a melanocyte.
00:20:28.19	Now a melanocyte is a pigmentation producing cell.
00:20:31.07	And in mammals, we produce two types of pigment,
00:20:34.03	a dark brown to black eumelanin, and a yellow to blonde
00:20:37.19	pheomelanin. So you can look at your own hair and
00:20:39.12	determine whether you have eumelanin or you have pheomelanin.
00:20:42.25	Now, this melanocyte has the ability to produce both types
00:20:47.21	of pigments. But which pigment it produces at any one time
00:20:50.14	is largely controlled by the melanocortin-1 receptor.
00:20:54.05	Which essentially acts like a switch. When it's turned on,
00:20:57.11	that is when let's say alpha-MSH, which activates Mc1r,
00:21:02.03	is around, Mc1r turns on and it signals by increasing intracellular
00:21:06.11	cyclic AMP levels, and you get the production of dark pigment.
00:21:10.00	By contrast, when Mc1r is turned off, then you get the production
00:21:15.15	of less cyclic AMP intracellularly, and the result is light
00:21:19.21	pigment. So Mc1r is very much a switch that determines which pigment is
00:21:24.27	produced. Now this leads to a nice prediction. So in beach mice we have
00:21:29.03	a mutation in Mc1r that we think leads to light coloration.
00:21:32.14	So our prediction is this, is that mutation of arginine to cysteine
00:21:36.18	change at position 65 will reduce receptor activity, which will result
00:21:41.13	in lighter pigmentation. So we can test this doing an experiment.
00:21:45.02	And so what we did is we took both the light allele and the dark
00:21:49.03	allele, that remember, differs by one nucleotide change,
00:21:51.26	we cloned it into an expression vector, we put it into
00:21:55.03	cells. And then we added alpha-MSH in increasing amounts.
00:22:01.14	To activate Mc1r, and then as a proxy for activity, we measured
00:22:05.13	cyclic AMP. So here are the results of this experiment.
00:22:09.17	So here what we did is you can see we added increasing
00:22:14.16	amounts of alpha-MSH, which turned on Mc1r, and you can
00:22:18.21	see that its signaling higher and higher cyclic AMP
00:22:21.22	until it sort of plateaus out. And this is a normal sigmoidal response curve
00:22:26.09	that we see in the mainland mouse. Now when we did the same
00:22:30.02	experiment with the beach mouse allele that differs by again
00:22:32.22	that one nucleotide change, what you can see is there's a dramatically different
00:22:36.14	pattern of activity. And that is no matter how much alpha-MSH
00:22:40.28	we added, there's still a relatively low level of receptor activity.
00:22:44.21	So what this suggests is that one nucleotide change changes the
00:22:48.26	function of the receptor, and it's in the direction we expect.
00:22:53.04	That it has lower activity, which is consistent with producing
00:22:55.21	lighter pigmentation. So at this point, it's worthwhile stepping
00:22:59.07	back and sort of thinking about what I've just showed you. So here I've
00:23:02.22	showed you a single nucleotide change affects the activity of the receptor,
00:23:07.04	we know this receptor affects color variation, and we know
00:23:11.08	that color variation affects survival in the wild. So what we
00:23:14.13	have done here is made a link between a single nucleotide
00:23:17.14	change and fitness or survival in the wild. But I don't want
00:23:22.24	to leave you with the impression that this is the whole story, because
00:23:26.28	as I mentioned, color difference is controlled not by a single
00:23:29.03	gene, but by multiple genes. And so the two other genes,
00:23:33.23	if you are paying attention you might have noticed, actually
00:23:37.02	interact with Mc1r. So Agouti, for example, represses Mc1r
00:23:42.02	activity. And what we see in beach mice, is that higher
00:23:45.26	levels of Agouti expression are associated with lighter pigmentation.
00:23:49.07	And so the allele in beach mice has increased expression
00:23:52.13	of Agouti, which leads to lower activity of Mc1r and light pigment.
00:23:57.02	Corin is the third gene, and while this gene is just newly
00:24:02.04	discovered in the last few years. We don't know its exact function
00:24:05.14	but we know it interacts with Agouti, and again, increased
00:24:09.05	expression of Corin is associated with this light pigmentation
00:24:11.26	in beach mice. So these genes are interacting together
00:24:15.11	to produce light pigmentation. So what we've done in this second part
00:24:20.28	is to tell you about the genes that affect coloration and a little
00:24:25.08	bit about how changes in those genes actually cause
00:24:28.09	these differences in color patterning. And now that we have a much
00:24:32.20	more complete picture of adaptation, here's where I think things can get
00:24:36.08	really fun. So I'm going to take you now back from the lab into the
00:24:39.16	wild. And we're going to talk about how these traits may have
00:24:42.05	evolved in natural populations of beach mice. So to do this, this
00:24:46.14	involves us actually going back to the field. So here's a picture
00:24:49.15	of me and my postdoc, Vera Domingues, when we're out
00:24:53.04	on the Atlantic coast catching mice. Here's a mouse
00:24:57.07	currently being weighed. So we hang it by its tail on a little
00:25:00.09	pesola. We take measurements including their weight,
00:25:03.15	the size of their ears and feet, et cetera. And what Vera's doing here
00:25:07.20	is measuring their coat color. And what's nice about this
00:25:11.23	work in the wild is we take these measurements, we measure their
00:25:14.21	coat color, we also give them a little ear tag, and we take a little
00:25:18.05	snippet of DNA from their tail, and we release them back
00:25:21.19	in the wild. So here we have a DNA sample from each of
00:25:24.12	these mice and we have a record of their coloration.
00:25:26.24	So next what I'd like to do is tell you about what we've learned
00:25:30.02	about natural populations of these mice and how these color
00:25:33.15	differences may have evolved. So just by way of reminder,
00:25:36.29	so far what I've done is I've focused on only one of these
00:25:40.22	populations, that is population number 3, the Santa Rosa
00:25:42.29	Island beach mouse. But next what I'd like to do is tell you about
00:25:46.02	variation among these subspecies. So the five subspecies
00:25:49.26	on the Gulf coast and the three subspecies on the Atlantic coast.
00:25:53.13	So, when we go out and catch these mice and record their
00:25:58.12	color differences. We find some very striking patterns, which I'm going to
00:26:02.19	show you in the next slide. So here what I'm showing you
00:26:05.18	are cartoons that represent the different subspecies of
00:26:10.12	beach mice. So each one of these cartoons shows you the typical
00:26:14.09	color of a beach mouse from each of these populations, compared
00:26:17.07	here to a mainland mouse. So the first thing you may notice
00:26:21.19	is that all the beach mice are much lighter in color
00:26:26.00	compared to the mainland mouse. But that each of these
00:26:28.07	subspecies differs in their color pattern, and in fact,
00:26:31.23	they're so distinct that if you went for let's say spring break
00:26:35.06	down to the Gulf coast of Florida and you brought me back a
00:26:38.13	beach mouse, I would say with about 95% certainty
00:26:42.03	just by looking at the color of the beach mouse, I could tell you
00:26:44.16	what subspecies it is. But, instead if you went to Florida and you
00:26:49.09	didn't tell me if you went to the Atlantic coast or the Gulf
00:26:51.28	coast, and you just brought me back a beach mouse, I'd probably
00:26:55.08	have a 50/50 chance of knowing what subspecies it was
00:26:58.17	and that's because the mice on the Atlantic coast are very similar
00:27:02.09	to mice on the Gulf coast. So let me highlight that here.
00:27:05.26	So for example, these two subspecies, even though they're
00:27:08.29	separated by over 300 kilometers, are very similar
00:27:12.18	in their overall color pattern. And in fact, I can't tell them
00:27:16.10	apart. Likewise, these mice are very similar. And these mice
00:27:20.26	are very similar. So what this suggests is that on the Atlantic
00:27:26.02	coast and Gulf coast, the mice have convergent color patterns.
00:27:31.08	And so what we wanted to do first was ask the question,
00:27:34.16	did these mice evolve these similar color differences independently?
00:27:39.14	And if so, did they use the same genes? So the first
00:27:44.02	thing I want to show you is a tree, or a topology that shows
00:27:48.24	you the relationships among these different subspecies.
00:27:52.00	So this is a simplified version of a tree that we generated
00:27:54.28	using molecular data, but it highlights the relationships
00:27:58.12	among these subspecies within peromyscus polionotus.
00:28:02.07	And what you can see is the Gulf coast beach mice, shown here,
00:28:06.08	all five of those subspecies cluster together. They're very
00:28:09.09	closely related. But they're actually not that closely related
00:28:12.16	to Atlantic coast beach mice, shown here. In other words,
00:28:17.11	it looks like light coloration has evolved independently
00:28:20.02	on two coasts. So the Gulf coast beach mice probably
00:28:23.07	arose from a dark colored ancestor sometime in the past,
00:28:26.22	that was probably from the Panhandle of Florida. Whereas,
00:28:29.23	the Atlantic coast beach mice independently evolved light
00:28:32.20	coloration, probably from an ancestor in Central Florida
00:28:35.26	that was dark in color. Okay, but here's the cool part,
00:28:40.04	now that we now at least in one of these Gulf coast subspecies
00:28:43.07	that the melanocortin-1 receptor is involved, we can
00:28:45.22	ask in these independently evolved light colored beach mice
00:28:48.22	on the Atlantic coast, is that same gene and same mutation
00:28:51.15	involved? So to do this, we returned back to this melanocortin-1 gene.
00:28:56.21	And we sequenced the DNA in the mice that we collected
00:28:59.19	on the Atlantic coast and asked, do we find that same
00:29:02.24	arginine to cysteine change at position 65? So we simply
00:29:07.28	genotyped that one particular site and asked, is it present in the
00:29:12.00	Atlantic coast? Now, despite the fact that mice from the Gulf
00:29:15.09	coast and the Atlantic coast are so similar in coat color,
00:29:18.06	let me just say that we never found that cysteine change in any of the
00:29:22.09	mice from the Atlantic coast. So what that tells us is that
00:29:26.01	the same mutation isn't involved in that convergently evolved
00:29:29.20	light coloration on the Atlantic coast. So you may be thinking,
00:29:32.13	well, it's not the same mutation, but maybe it's a new mutation.
00:29:36.09	Remember there's lots of mutations in Mc1r that can cause
00:29:39.08	color differences, I told you this earlier. So maybe a new mutation
00:29:42.12	in Mc1r is causing light coloration on the Atlantic coast. So
00:29:45.23	it may not be the same mutation, but it could be the same gene.
00:29:48.18	So we went back to the Atlantic coast mice and sequenced
00:29:50.20	the entire melanocortin-1 receptor, and asked are there any
00:29:54.08	new mutations that are correlated with color. Well, in fact,
00:29:57.08	we found four new mutations in the melanocortin-1 receptor.
00:30:01.02	But none of them were perfectly correlated with color, and
00:30:04.11	when we did those pharmacological assays like the ones
00:30:07.04	I showed you earlier, none of them had an effect on the
00:30:11.17	activity of Mc1r. So what this tells us is that it's not just
00:30:16.12	the same mutation, it's also not the same gene that's
00:30:19.11	responsible for the convergent evolution of light coloration in the
00:30:23.10	Atlantic coast mice. So this is the case for melanocortin-1 receptor,
00:30:26.25	but we're still checking these populations for changes in Agouti and Corin.
00:30:31.27	So what I've done today is I've told you a story about
00:30:37.00	how identifying not only the ultimate causes of phenotypic
00:30:41.12	variation, but also the genetic causes can tell us something about how
00:30:44.13	traits evolved in the wild. And this story revolved around a single
00:30:47.24	species, in which different mutations, at least in regards to
00:30:52.04	the melanocortin-1 receptor, are involved in generating
00:30:55.04	similar coat color patterns on two coasts of Florida.
00:30:57.25	But I want to end by telling you a story about convergent
00:31:01.14	evolution. Because sometimes the same genes and same
00:31:04.13	mutations are involved in similar phenotypes in very different
00:31:08.01	organisms. So, if you think about mammals, think about what
00:31:13.02	is the most different mammal from a mouse. Often the answer
00:31:17.03	I get is an elephant, and it's close. I'm going to tell you a story about
00:31:21.11	mammoths. So, mammoths were the subject of intense
00:31:26.17	study, especially about five years ago when there was a big
00:31:30.06	interest in sequencing ancient DNA. And mammoths were
00:31:33.28	a great target because they occurred in Siberia, and
00:31:38.05	their DNA was essentially frozen in permafrost about 14,000
00:31:43.03	years ago. So this is almost like keeping DNA in a giant freezer,
00:31:46.15	which is the best conditions possible. So about five years ago,
00:31:50.07	the goal was to sequence an entire gene from an extinct
00:31:53.25	organism. Now today of course, we're sequencing entire
00:31:57.03	genomes of extinct organisms, but just five years ago, we
00:32:01.00	wanted to sequence an entire gene from the nuclear genome.
00:32:04.29	So this is what a group in Germany, in Leipzig, headed by
00:32:10.12	Michael Hofreiter did. And the gene they chose to study
00:32:14.00	was the melanocortin-1 receptor, because it's a very
00:32:17.13	simple gene. Remember I told you, we know a lot about its
00:32:20.27	structure function, it's only about 1,000 base pairs in length,
00:32:23.05	and there's no introns. So it's very simple and it's a great starting
00:32:26.05	point. So they sequenced the melanocortin-1 receptor,
00:32:29.10	in DNA extracted from mammoths. And here's what they found.
00:32:35.14	They found a mutation, and what mutation was it?
00:32:39.01	Well, it was an arginine to cysteine change at the exact
00:32:41.19	same position that we found in beach mice. Now of course
00:32:45.15	the DNA it was extracted from was bone, so they didn't know
00:32:48.18	the phenotype or the coat color of mammoths, but based on our
00:32:52.06	work in beach mice, what this suggests is that mammoths
00:32:55.01	like beach mice, may have been polymorphic in color. Now the
00:32:59.07	question of why they were polymorphic in color, we don't know the
00:33:01.26	answer to. This could be due to survival differences, other
00:33:06.25	suggested that it's due to sexual selection. So when this
00:33:10.07	work was published, the press line was that blonde mammoths have
00:33:14.15	more fun. But I'm not going to give you any explanation, I'll leave that
00:33:19.08	up to your imagination. So in this case, we have very
00:33:22.21	divergent organisms, mice and mammoths that use
00:33:26.03	not only the same gene, but the same mutation.
00:33:29.12	And in fact, as I alluded to earlier, melanocortin-1 receptor
00:33:33.06	is involved in a number of different color variants.
00:33:36.05	And a number of organisms that you may be familiar
00:33:38.24	with. So in this case, it was the same mutation, but we know
00:33:42.25	that other changes in the melanocortin receptor can also
00:33:46.21	cause differences in color through different mutations. So I just wanted to
00:33:51.02	give you some examples. So some work that I've been doing with a colleague
00:33:54.12	at UC Berkeley, Erica Rosenblum, shows that changes in the
00:33:57.26	melanocortin-1 receptor is responsible for the production of these
00:34:00.22	very adorable lizards that differ in color, when they occur
00:34:05.14	in White Sands, New Mexico. So you can guess which
00:34:09.02	lizard occurs on white sands and which one occurs off white sands.
00:34:13.06	Color differences are also involved in animals you may see
00:34:17.09	every day, including cows. Also, many of you may know
00:34:22.12	of or even have a labrador dog. Well, they come in
00:34:26.21	generally two colors, there's the black labs and blonde labs,
00:34:30.12	again this is caused by a change in the melanocortin-1 receptor. And much more recently, again work out of Germany,
00:34:36.23	this time Svante Paabo's group, has shown that color
00:34:41.06	changes in the melanocortin receptor, in addition to being
00:34:45.20	responsible for human color differences, may also be
00:34:48.29	responsible for color differences or hair differences in neanderthals.
00:34:53.03	So they showed that changes in the melanocortin-1 receptor
00:34:56.02	or variants in the melanocortin-1 receptor were found in
00:34:59.24	neanderthals, suggesting even back in those days,
00:35:05.03	there could have been redheads as well. So what I hope to have done
00:35:09.07	today is to suggest to you that by making the link between
00:35:13.23	environment, phenotype -- in this case a morphological trait,
00:35:17.22	and genotype, we're able to say something about
00:35:21.05	how organisms evolve in the wild. The work that I presented
00:35:25.12	today was done by a large number of people. These are some
00:35:30.22	of the people in my lab group that contributed to the work,
00:35:34.05	shown here looking their mousiest. And in particular,
00:35:38.00	the work I talked about today was done by folks like
00:35:41.21	Cynthia Steiner, Marie Manceau, Vera Domingues, Sacha
00:35:44.23	Vignieri, Holger Rompler, and Lynne Mullen, as well as
00:35:47.29	an undergraduate, Joanna Larson. And we were funded by
00:35:52.03	a number of sources shown here, as well. And with that,
00:35:55.10	thank you for your attention. I hope you enjoyed this segment,
00:35:58.03	and you'll stick around for segment three. Thank you very much.