The Vertebrate Retina, Photoreceptors, and Color Vision
Transcript of Part 3: Human Color Vision and its Variations
00:00:02.02 I'm Jeremy Nathans. I'm a professor at the Johns Hopkins Medical School, 00:00:05.15 and an investigator of the Howard Hughes Medical Institute. 00:00:08.05 This is the second of three lectures on the vertebrate retina, 00:00:11.23 its structure, function, and evolution. 00:00:14.05 In this lecture, we're going to look at human color vision and its variations. 00:00:18.19 Now, why should we study color vision? 00:00:22.21 Well, one reason - the reason I like most - is because it's beautiful. 00:00:26.09 We see a world of color out there, as seen in this upper photograph, 00:00:31.17 and if we were deprived of that color, the world would be a far duller place. 00:00:36.16 And the story I want to tell you is really a centuries-long story, 00:00:42.05 It begins in 1665, in the dormitory room of Isaac Newton, 00:00:47.10 who was an undergraduate at Cambridge University at that time. 00:00:50.05 And Newton did a very simple, yet profound experiment. 00:00:54.02 He drilled a hole in his wooden shutter to admit a beam of light, 00:01:00.12 and he passed that beam of light through a prism and asked what happened to it. 00:01:06.03 And the answer was that the beam was split up into its component colors... 00:01:10.19 the colors of the rainbow, from red at the least refrangible to violet at the most refrangible, 00:01:19.11 and in between, orange, yellow, green, and blue. 00:01:22.09 Now that had been known for some time, but Newton appreciated 00:01:27.04 the physical significance of that phenomenon. 00:01:29.16 That is, that white light is a summation of different lights, 00:01:33.21 that those lights could be separated physically in an objective fashion, 00:01:37.17 and that their properties could be objectified, in the sense that each kind of ray 00:01:43.05 had a particular angle of refraction when passed from air to glass and then glass back to air again. 00:01:49.24 And furthermore, Newton realized that this method could be used 00:01:53.17 to purify the individual rays of light to obtain spectrally pure samples with which to work. 00:02:01.01 So, for example, by blocking off all but the orange ray of light as shown here, 00:02:05.26 Newton was able to obtain a beam of spectrally-pure orange. 00:02:10.03 Nowadays, perhaps, we would do this with a tuneable laser to obtain a spectrally-pure light. 00:02:15.23 But, Newton appreciated the power of this purification method, and he used it to good effect, 00:02:21.20 and in particular, he used it to recombine lights of different colors. 00:02:27.28 Now, I can't resist showing you Newton's original sketch. 00:02:32.06 This is his quick pen sketch of the experiment. 00:02:36.10 It's reversed left to right from the image that I showed you earlier, 00:02:39.05 but here's the beam of light coming in the window, 00:02:42.11 it's refracted by the prism, and then various beams are obtained from it 00:02:48.00 by using their specific angles of refraction. 00:02:50.26 And in particular, Newton performed the following very simple but profound experiment 00:02:57.19 diagrammed here, essentially with projectors, as we might do it today. 00:03:01.27 But Newton did it with mirrors and those beams of spectrally purified light. 00:03:07.19 Newton asks, what if you combine together a spectrally pure red and a spectrally pure green light? 00:03:17.27 What would it look like? 00:03:19.22 The answer is it looks like neither. It looks yellow. 00:03:22.14 That is, the superposition of red and green looks just like a spectrally pure yellow. 00:03:28.26 But, it is not physically identical to a spectrally pure yellow, and Newton could prove this 00:03:36.14 by passing this yellow, such as we see here, through another prism 00:03:41.28 and a spectrally pure yellow, identical to it in all appearances, 00:03:45.28 also through a prism. And what he saw was that they behaved quite differently. 00:03:49.20 A spectrally pure yellow retains its same initial angle of refraction, 00:03:54.15 and remains a single homogeneous beam when passed through a prism, 00:03:58.07 but a synthetic yellow, such as we see here, 00:04:00.21 one obtained by superposition of red and green lights, 00:04:04.05 in fact splits up back into its component red and green lights, 00:04:07.10 each with their own original angles of refraction when passed through a prism. 00:04:12.11 And Newton realized that this meant that human color vision was imperfect 00:04:17.19 in the sense that it could not distinguish certainly physically distinct combinations of lights. 00:04:23.06 Now, as far as we know, Newton never examined the physiologic basis for that phenomenon, 00:04:32.25 although he worked on the physical rules that govern 00:04:37.10 the recombining of lights of different colors. 00:04:40.18 Now we would refer to them as different wavelengths 00:04:43.13 instead of the angles of refraction that he used. 00:04:47.11 But he had remarkable physical insights into the color vision system, 00:04:53.20 and let me just give you one of these. 00:04:55.27 This appeared in his Opticks, his wonderful book describing his early experiments. 00:05:02.23 And here, Newton asks, "May not the harmony and discord of Colours" 00:05:08.11 (that is, the similarities and differences of the colors) 00:05:11.04 "arise from the proportions of the Vibrations propogated through the" 00:05:15.28 "fibres of the optick Nerves into the Brain as the harmony and discord of Sounds" 00:05:20.29 "arise from the proportions of the Vibrations of the Air?" 00:05:24.24 Now this is a remarkable insight at a time when it was 00:05:27.28 not at all clear what the physical nature of light was. 00:05:31.06 Newton had a sense, really by analogy, that light might consist of some sort of wave, 00:05:36.29 but I should say that Newton also in the Opticks discussed 00:05:41.22 the evidence that light has a particle-like characteristic. 00:05:46.24 He comes down somewhat in the middle and says that it is both particle-like and wave-like 00:05:52.20 a remarkable judgment in light of the fact that this is 2+ centuries before the quantum theory 00:06:00.07 emerged to tell us that in fact that light does have both particle and wave properties. 00:06:05.07 The physiologic underpinnings of the color mixing of the kind that Newton saw 00:06:11.08 was not really laid out until the 19th century. 00:06:14.25 Its first clear description we owe to Thomas Young, 00:06:19.24 the universal genius who worked on many things, including the Rosetta stone and vision. 00:06:25.28 And Young, in the Bakerian Lecture of 1802, before the Royal Society of London, 00:06:32.23 suggested the following as a model for how color vision might work. 00:06:38.27 And he says, as follows: "As it is almost impossible to conceive" 00:06:44.26 "each sensitive point of the retina to contain an infinite number of particles" 00:06:49.28 (that is, an infinite variety of receptors) 00:06:52.18 "each capable of vibrating in perfect unison with every possible undulation," 00:06:58.08 (that is, every possible wavelength of light, because at this point, 00:07:01.25 it was clear that light had wave properties) 00:07:04.10 "it becomes necessary to suppose the number limited," 00:07:07.20 "for instance, to the three principle colours, red, yellow, and blue," 00:07:11.24 "and that each particle is capable of being put in motion more or less forcibly" 00:07:17.10 "by undulations differing less or more from perfect unison." 00:07:22.04 Now this very prescient statement in fact hit the nail exactly on the head. 00:07:27.11 Because what Young is saying here is that each of the three receptors 00:07:30.24 would have a broad, bell-shaped curve. 00:07:34.03 There would be a place where it is maximally sensitive, 00:07:37.03 here it is matching the so-called undulation of the stimulus, 00:07:42.29 and then there are other parts of the spectrum 00:07:45.15 where it would be less sensitive because the match was less perfect. 00:07:48.00 And so, by having that sort of broad sensitivity, the limited number of receptors, 00:07:55.19 the three receptors that he's postulating, would be able to cover the spectrum, 00:07:58.28 and between them, and by comparing their degrees of excitation, 00:08:03.03 one could determine the color appearance of any given light. 00:08:07.13 That was exactly right. 00:08:09.17 Now, the physiologic evidence - the behavioral evidence, really, from testing humans - 00:08:16.02 that would support this model was not forthcoming until another 50 years or so. 00:08:21.14 And it came from two equally remarkable scientists, 00:08:25.12 Hermann von Helmholtz, shown at the left, 00:08:27.21 and James Clerk Maxwell, shown at the right. 00:08:29.23 They're both shown here in their 20s. Helmholtz, at the time of this photograph, 00:08:34.06 had just invented the ophthaloscope and revolutionized ophthalmology. 00:08:38.16 Maxwell, shown to the right, is holding his famous color top. 00:08:42.11 He's an undergraduate in this photograph. He had been interested in color for many years, 00:08:48.16 starting as a young man, and had devised a series of tops with different colored patterns 00:08:54.08 that he could spin at high speed to effect a uniform mixing of the colors. 00:08:59.05 We now know, if we can fast forward another 150 years or so, 00:09:07.23 that the Young model is exactly right, and there are, in fact, three light sensors, 00:09:15.29 and they have broad, bell-shaped curves, and these curves 00:09:19.02 have now been identified by a number of methods, 00:09:21.27 and here's just one of them. 00:09:23.25 In this case, the individual proteins that constitute these receptors and their chromophores 00:09:31.20 have been joined together and the absorbance spectra determined in a spectrophotometer, 00:09:35.22 and we see that Rho, which stands for rhodopsin, 00:09:40.10 the one in rods, is this curve peaking at about 500nm. 00:09:46.10 This one is not involved in color vision. I've put it here just for completeness. 00:09:49.20 It's involved in dim light vision only, but it has the same overall shape 00:09:54.07 because these different pigments work, in terms of their photochemistry, 00:09:57.23 really in the same way. 00:09:59.14 But the other pigment curves - the longer wave, the so-called L curve, 00:10:03.28 the medium wave (the M curve), and the short wave (the S curve) 00:10:08.09 are offset by substantial numbers of nanometers 00:10:12.24 from one another and from the rhodopsin curves. 00:10:15.07 And together, they span, essentially the full range of the visible spectrum. 00:10:20.03 Now what does this get us, in terms of color vision ability? 00:10:26.00 We can ask that in a number of ways. 00:10:28.00 One way we could ask that question is by simply asking, at any wavelength 00:10:34.25 along the entire visible spectrum, how different do two lights have to be to tell them apart? 00:10:41.19 So that's shown here, for example. 00:10:45.00 Here, on the horizontal axis is wavelength, and on the vertical axis 00:10:48.13 is the differential - the delta lambda - the differential between two lights 00:10:52.21 that one needs to discriminate them. 00:10:55.24 I think you can appreciate that the curve is not a simple flat curve. 00:10:59.11 It goes up and down - there are hills and valleys. 00:11:01.11 There are some places, for example right here, where a difference of perhaps half a nanometer 00:11:06.00 is sufficient to distinguish two lights, and other places 00:11:09.12 where a difference of say 6 nanometers or so is required. 00:11:13.01 This curve of hills and valleys is completely understood 00:11:19.06 by reference to the original absorbance curves of the pigments. 00:11:23.26 So, for example, if we ask how we can distinguish, say, 00:11:29.17 a light that is at say, 600 or so nanometers, 00:11:32.22 we can distinguish it because it differentially excites the long and medium wave pigments, 00:11:39.15 and a light of say 595 nm - a little bit blue-shifted relative to it 00:11:45.26 has a different ratio of excitation of those two pigments. 00:11:49.15 And that's a sufficiently great difference that we can tell them apart. 00:11:52.26 I should add that this set of curves also explains Newton's color mixing data 00:11:57.29 in a very simple way. For example, if we have a sum of red and green lights 00:12:03.17 those give a certain degree of excitation of those three different sensors 00:12:10.15 that is the same, as it turns out, as a spectrally pure yellow light. 00:12:14.09 So, of course, the system cannot tell them apart. 00:12:16.19 Because, in the end, a detector is really just a photon counter. 00:12:20.19 Once a given number of photons is captured and that number is then sent to the brain, 00:12:26.29 there's no way to tell what the source of those photons were. 00:12:29.26 Was it a long wave light that was captured inefficiently, but was very bright, for example, 00:12:34.28 or was it a shorter wave light which was captured more efficiently, but was dimmer. 00:12:38.10 That's completely lost within the system, and any pair of lights 00:12:42.19 which give the same total number of photons captured per unit time 00:12:47.27 for each of the three receptors, will, of course, then be indistinguishable. 00:12:51.29 As part of Maxwell's experimental approach to analyzing human color vision, 00:12:57.18 he brought to us a conceptualization of color perception, 00:13:02.14 which is extremely useful and is shown here. 00:13:04.29 Maxwell thought that we could conceptualize color space as a 3D space, 00:13:12.10 in which each of the three axes, as shown here (the L, the M, and the S axes), 00:13:19.24 correspond to the degrees of excitation of each of the three classes of cones. 00:13:24.19 So, for example, a white light, which equally excites all three cone classes, 00:13:30.25 as shown by this corner of this square here, 00:13:34.14 and this vector coming from the origin, 00:13:37.00 would be represented by a vector in this direction, and lights of other wavelengths 00:13:44.12 would be represented by distinctive vectors in their own unique directions. 00:13:48.16 So, in Maxwell's conceptualization, our color vision is a 3D color vision. 00:13:54.16 We have three degrees of freedom. If an individual were to have only two receptors, 00:13:59.23 that individual would have a 2D color space. 00:14:01.20 So, for example, if one is missing the L receptor, 00:14:05.17 the color space would consist of the M and S axes, and the plane defined by them, 00:14:11.12 but not the 3D S, M, and L space. 00:14:16.02 Now, let's talk a little bit about the chemistry of color reception, 00:14:24.10 and in particular, the ways in which different wavelengths of light might be selectively absorbed 00:14:31.16 by the different visual pigments. 00:14:33.08 We saw, in the first lecture, the overall structure of visual pigment. 00:14:37.10 It's a G-protein-coupled receptor. It has 7 transmembrane segments 00:14:40.24 and sits in the membrane of the outer segments. 00:14:44.06 And, as you recall from that first lecture, the actual light absorbing part 00:14:48.27 is a small molecular weight vitamin A derivative, called 11-cis-retinal. 00:14:54.27 That's true for all of the human pigments. 00:14:57.11 They're all built on the same basic architecture 00:15:00.01 and they all use exactly the same 11-cis retinal chromophore. 00:15:04.06 So, the difference between the pigments is a difference in the proteins, 00:15:09.15 and what the proteins do is they impose a somewhat different electronic environment 00:15:14.12 around that retinal chromophore which has the effect of tuning its absorbance spectrum 00:15:19.27 to either longer or shorter wavelengths. 00:15:21.26 So, one can think about this spectral tuning challenge as really 00:15:26.14 a question of how to modify the pi-electron system of the retinal chromophore 00:15:34.16 by virtue of the surrounding amino acid side chains. 00:15:38.13 And we can learn something that at least constrains the possible mechanisms 00:15:44.23 by examining the basic photochemistry of retinal itself. 00:15:47.19 The critical point, with respect to spectral tuning, is that the degree 00:15:55.00 of pi electron delocalization within retinal determines its absorbance spectrum. 00:16:00.20 Now, retinal being a conjugated polyene, as shown here, 00:16:03.18 with 6 double bonds, has a substantial amount of resonance and therefore 00:16:09.12 delocalization of electrons. 00:16:10.23 I've shown below just one resonance structure that one could draw. 00:16:14.22 Not a particularly favorable one because it places a positive charge on this carbon here, 00:16:19.06 as a result of moving the pi electrons over towards the nitrogen, 00:16:23.03 which now has a free electron pair, 00:16:25.06 but this, of course, sort of resonance structure, even though it is relatively unfavorable, 00:16:30.15 will contribute to some extent in that equilibrium ensemble of all resonance structures. 00:16:35.13 Just to get a sense of how substantial this resonance effect is, 00:16:41.24 in the absorbance spectra of these kinds of compounds, 00:16:44.27 let's look at two model compounds which are shown here. 00:16:48.18 A polyene at the top, and a cyanin at the bottom. 00:16:51.17 Now, these are drawn with the same number of double bonds as retinal has, 00:16:55.16 that is 6, and I think that you can see, if we look at this polyene at the top, 00:17:00.11 where the chain ends in carbons, here, that a resonance structure of 00:17:06.29 the sort that I've drawn just below it, 00:17:10.07 in which one carbon now carries a positive charge and the other carries a negative charge, 00:17:15.22 would be quite unfavorable. 00:17:18.04 That involves, of course, a charge separation, 00:17:21.18 in addition to the general unfavorability of placing charges on carbons, 00:17:25.26 and so, in general, polyenes have very little pi electron delocalization. 00:17:31.26 We know this for a fact because when one looks at the crystal structure of a polyene, 00:17:36.27 one sees that the lengths of the bonds are highly alternating. 00:17:39.28 There are lengths that are very long - the canonical single bond lengths, 00:17:43.28 and there are others in between that are very short - the double bond lengths, 00:17:47.07 and so this back and forth - long, short, long, short, long, short - 00:17:50.27 tells us that, really, the single bonds have virtually exclusive single bond character, 00:17:56.20 the double bonds have exclusive double bond character, 00:17:58.14 and to go along with that, the absorbance spectra of the polyenes 00:18:03.17 is at quite short wavelengths. For example, here, 00:18:06.26 we're down in the UV at about 360 nm for the absorbance maximum. 00:18:14.27 At the other end of the spectrum, so to speak, for pi electron delocalization, 00:18:23.25 would be a model compound of the sort shown down here. 00:18:26.15 This is a cyanin. Again, six double bonds, but now we have an NH2 at both ends, 00:18:33.29 in one case, there's a positive charge, a free electron pair, 00:18:38.09 but I think you can appreciate that these sort of resonance structures that are drawn below 00:18:42.26 in which each single bond has now been switched to a double bond, 00:18:45.28 each double bond now switched to a single bond, 00:18:48.04 simply is a mirror image of the one up above. 00:18:51.21 Now the free electron pair is at the right, the positive charge is at the left, 00:18:55.27 but everything else is the same in chemical terms. 00:19:00.18 And therefore, these two must be equal in their energies, and therefore equally likely to occur. 00:19:05.27 And therefore, the pi electrons are maximally delocalized in a cyanin-type compound. 00:19:13.02 And again, we know this is true because, from crystal structures of cyanins, 00:19:17.08 we see that all of the bond lengths are the same, 00:19:19.28 and they're mid-way between the lengths of single and double bonds. 00:19:23.16 Going along with that, we see that the absorbance maximum of a cyanin is quite long. 00:19:30.02 In this case, it's 750nm - far to the red. This would be a very deep red compound. 00:19:35.25 So, with these two model compounds, we see a difference 00:19:44.07 in absorbance that spans essentially the entire visible range. 00:19:47.19 All visual pigments that are known in nature are somewhere 00:19:52.04 within the extremes defined by these two model compounds 00:19:56.00 and, in essence, what visual pigments are doing is they're creating with the retinal chromophore 00:20:02.00 some sort of hybrid between a polyene and a cyanin. 00:20:06.00 A hybrid in the sense that the electrons are delocalized to a degree that is intermediate 00:20:10.22 between the delocalization seen with a polyene and the delocalization seen with a cyanin. 00:20:16.17 And, correspondingly, the absorbance spectra is intermediate 00:20:22.08 between the extremes shown here - 360nm and 750nm. 00:20:28.00 So, now that focuses our thinking a little bit, because really what we can imagine 00:20:32.14 the protein must do is make the chromophore look more polyene-like or more cyanin-like. 00:20:39.05 How it does that we don't know in complete detail at this point, 00:20:46.09 but we have a few clues. 00:20:47.11 First, we have the amino acid sequences of the proteins. 00:20:50.00 And when we look at those sequences, and here they're shown with the proteins splayed out in 2D, 00:20:55.02 with each dot in each of these schematics indicating an amino acid difference. 00:20:59.21 So, here, for example, we're comparing the M cone pigment 00:21:02.26 (the middle wave sensitive pigment) with rhodopsin. 00:21:05.18 Here we're comparing the S or short-wave pigment with rhodopsin. 00:21:11.06 Here we're comparing the medium pigment (the M pigment) with the short-wave pigment 00:21:15.08 and here, in the lower-right, we're comparing the medium pigment 00:21:19.03 with the longer-wave pigment (M vs. L). 00:21:22.17 So, I think you can appreciate that, although they have the same overall structure, 00:21:28.00 they are 7-transmembrane-pass G-protein coupled receptors, 00:21:30.29 there are obviously many amino acids that are identical - the open circles indicate identities 00:21:36.27 there are also, for at least three of these pairwise comparisons, 00:21:41.12 many amino acid differences. 00:21:44.14 I think you can also appreciate that there's an asymmetry here. 00:21:46.25 The M vs. L comparison shows that these two pigments are extremely similar - one to another. 00:21:53.14 There are only about 15 amino acid differences, depending which variants 00:21:58.01 (there are a few allelic variations that can make that number a little bigger or smaller) 00:22:02.17 but roughly 15 amino acid substitutions that distinguish the M and L pigments. 00:22:08.27 out of a total of 364 amino acids. 00:22:13.02 That asymmetry is going to be a focus of much of the remainder of this lecture 00:22:18.24 and the next lecture. 00:22:20.14 Now let's just look at what that sequence difference looks like in a dendrogram form. 00:22:28.19 Here we're looking at sequence divergence among human and mouse rod and cone pigments. 00:22:33.26 We've picked the mouse as a representative non-primate mammal. 00:22:37.26 And here, let me just say that the non-primate mammals have only two types of cones 00:22:42.20 and rods - very much like our rods - whereas a subset of primates, like ourselves, 00:22:49.09 have three cone pigments. 00:22:51.03 Now, if we look at these amino acid sequences, and here we're plotting on the horizontal axis, 00:22:56.14 the degree of sequence divergence - the typical dendrogram-type plot - 00:23:02.02 and on the vertical axis, we're plotting the 00:23:04.20 wavelength of maximal absorbance of the corresponding pigment. 00:23:07.26 So, let's just walk through this, starting with the mouse example. 00:23:11.01 The mouse rhodopsin, like typical mammalian rhodopsins, 00:23:14.17 absorbs at about 500 nm. And its sequence is rather different. 00:23:19.06 That is, this dendrogram line is rather long, 00:23:21.07 (this horizontal line) compared to the mouse medium wave pigment, 00:23:26.18 the only example that the mouse has of a longer-wave sensitive pigment. 00:23:30.21 And it's also quite different from the shorter wave pigment that the mouse has, 00:23:34.05 which actually absorbs in the near UV - it's absorbing at wavelengths 00:23:39.05 somewhat shorter than the one that we have. 00:23:40.28 In the case of the human pigment set, we see that again, 00:23:46.29 we have a rhodopsin typical of mammalian rhodopsins, 00:23:49.08 and we have a shorter-wave pigment, and these differ substantially in sequence. 00:23:54.09 And then in what must have been a very recent evolutionary event, 00:23:58.09 there has been a split on the arm of the tree to give a long and a medium wave pigment, 00:24:06.09 whereas our immediate mammalian ancestors did not enjoy that pair of pigments. 00:24:14.21 When the genes for the long and medium pigments were first identified, 00:24:19.12 not only were their sequences found to be very similar, 00:24:23.16 (and that, of course, is the source of the sequence comparison that we saw 00:24:26.16 just a few slides ago.) 00:24:28.13 But also the arrangement of genes was such that they were found 00:24:33.14 to be adjacent to one another on the DNA. 00:24:36.11 So here is the human arrangement of a long-wave pigment gene, shown with a red arrow, 00:24:41.08 and a medium-wave (an M pigment gene) shown with a green arrow, 00:24:43.23 and that's not a surprising arrangement, given their recent evolutionary history. 00:24:50.06 Genes that duplicate are not infrequently found to be adjacent to one another. 00:24:54.08 And we imagine that, at some point in evolution, there was an erroneous 00:24:58.22 recombination event which paired a segment downstream of one gene 00:25:03.27 and upstream of another, and as a result, we got a chromosome... 00:25:08.03 or our ancestors got a chromosome with both genes adjacent on the DNA. 00:25:13.29 Now, I've drawn here one aspect of that ancient arrangement 00:25:20.09 which I'll justify in the third lecture, 00:25:22.04 and that is that the original pair of pigments which underwent this recombination 00:25:27.09 were actually not identical - one was already an L pigment gene, 00:25:32.18 and the other was already an M pigment gene. 00:25:35.05 That is, they were already different from each other at the time of the duplication. 00:25:38.19 And, let me reserve that, as I said, that story for the third lecture. 00:25:43.08 Now, if we look in modern-day humans, although we see 00:25:48.13 this arrangement of a single L and a single M pigment gene, 00:25:51.09 on some chromosomes... in fact, only about 20% of human X-chromosomes 00:25:57.19 (it turns out, these are on the X chromosome.) 00:25:59.18 Only about 20% of human X-chromosomes actually have that arrangement. 00:26:03.17 When DNA from a large number of different humans was examined 00:26:09.03 to look at the arrangement of L and M genes, 00:26:12.14 it became apparent that the basic arrangement of one L gene and one M gene, 00:26:17.29 as shown at the top here, is actually not the most common of arrangements in the population. 00:26:24.22 In fact, most people have one L and two M genes, 00:26:29.14 some have one L and three M genes, 00:26:31.07 and there are a few even who have four or five M pigment genes. 00:26:35.01 All of these people, as far as we can tell, have perfectly normal color vision. 00:26:39.11 Now this arrangement almost certainly arises because 00:26:44.24 the genes are homologous to one another, 00:26:47.05 and are capable of recombination, and we'll discuss that in just a minute. 00:26:52.03 This recombination, in fact, gives rise to a number of variations in gene structure, 00:26:58.12 and the sort of variation which appeared first in terms of scientific attention 00:27:04.17 was the kind that affected color vision and has been called, perhaps incorrectly, 00:27:09.06 color blindness. Now, color blindness doesn't mean a complete inability to see colors. 00:27:14.11 As generally used, it means simply a reduction of color vision from trichromacy, 00:27:20.09 (from having three receptors) down to having only two. 00:27:23.21 Perhaps the most famous dichromat is shown here, John Dalton, 00:27:27.04 the great British chemist who, in his first scientific paper, in 1794, 00:27:33.05 described his own color vision in meticulous detail. 00:27:37.02 Dalton was a dichromat and he reported that what other people 00:27:42.29 called red or orange or yellow or green 00:27:46.04 appeared to him to be just a single kind of color, differing only in intensity. 00:27:50.25 Let's see what the world looks like to people who are missing one of the three receptors. 00:27:57.22 And, I'm showing here, photographs of a fruit stand, taken and processed so that we see 00:28:07.22 what it would look like to a normal trichromat, here on the upper left, 00:28:10.13 or to any of the three types of dichromats. 00:28:14.17 On the upper right, is a so-called protanope's view - someone missing the long-wave sensor. 00:28:21.29 On the lower right is the deuteranope's view - that's someone missing the medium-wave sensor. 00:28:27.00 And over here on the lower left is a tritanope's view - 00:28:29.28 that's someone missing the short-wave sensor. 00:28:31.23 Now, I think you can appreciate that, as Dalton reported, 00:28:35.23 without the long or medium wave sensor, the fruit looks very much washed out in its colors. 00:28:42.26 That is, the longer wavelength colors - the ones that 00:28:45.11 particularly characterize the ripeness of fruits, 00:28:49.08 the green, yellow, orange, and red colors, 00:28:52.15 really are not distinguishable readily by an individual 00:28:56.25 who is missing either the long or medium wave pigment. 00:29:01.12 What's the origin of these not-uncommon color vision anomalies? 00:29:06.25 It turns out that they arise from unequal recombination between the members of the 00:29:14.09 L and M tandem array, and that this recombination occurs quite readily 00:29:20.10 because the genes not only are adjacent to one another, 00:29:24.00 not only are they in a head-to-tail tandem array, 00:29:26.08 but they also have retained a very high degree of sequence identity 00:29:30.14 (approximately 98% identity at the DNA level). 00:29:33.21 And so, as a result, mispairing can occur at meiosis, of the kind shown here 00:29:38.19 For example, a region distal to this M pigment gene, 00:29:42.22 is nearly identical in sequence just distal to this L pigment gene. 00:29:48.02 And if those two regions were to pair at meiosis 00:29:51.12 and a recombination event were to occur there, 00:29:53.21 we would get the two meiotic products shown at the bottom. 00:29:57.04 One, in which only a single L pigment gene remains, 00:30:01.03 and the other in which an additional M pigment gene has been acquired. 00:30:05.00 The one at the top, if inherited by a man, since this is on the X chromosome, 00:30:12.15 (for whom this is, of course, the only X chromosome) 00:30:15.22 would cause dichromacy - this gentleman would be missing 00:30:20.07 M pigment and would have only the S and L pigments. 00:30:24.08 As far as we know, as I mentioned earlier, people with extra 00:30:28.01 pigment genes have perfectly normal trichromatic color vision. 00:30:32.21 In fact, current evidence would suggest that only the first gene in the array, 00:30:36.25 when there are multiple M pigment genes, only the first M pigment gene 00:30:40.19 is expressed, and more distal M pigment genes are not expressed. 00:30:44.23 Now, what if recombination occurs at an intragenic location, rather than between the genes? 00:30:54.06 That is, within the genes. 00:30:56.07 Well, if this region of the M pigment gene pairs with 00:31:00.07 the corresponding region of the L pigment gene, 00:31:01.29 and again these regions are almost identical in sequence, 00:31:04.19 it would create meiotic products in which there would be hybrid genes. 00:31:10.13 So, for example, one product in which the M pigment gene was lost, 00:31:15.09 would retain, not a normal L pigment gene, 00:31:17.25 but an L/M hybrid. And, the reciprocal product 00:31:22.08 would be present on the other chromosomal product of this recombination. 00:31:26.24 with an M pigment gene picked up as a more distal gene of that recombination event. 00:31:32.28 Now again, the color vision of these individuals would be affected. 00:31:38.10 Of course, the person who inherits this chromosome at the top, 00:31:42.05 (the male who inherits that chromosome) would be missing one of the longer wave genes, 00:31:47.22 and in this case, would have the L/M hybrid instead of either the normal L or normal M gene. 00:31:53.11 In the case of the individual who gets the chromosome shown at the bottom, 00:31:57.13 recall that I just mentioned that only the first two genes in the array are expressed, 00:32:05.01 so this person would express the L pigment gene (that would be a normal gene), 00:32:08.23 but the M/L hybrid would now be expressed instead of the normal M pigment gene. 00:32:13.27 And that hybrid, we've come to realize, would in general have an absorbance spectrum 00:32:20.14 different from either of the two parental pigments. 00:32:23.19 So, this hybrid (and this is also true of the hybrid diagrammed here) 00:32:27.10 would be shifted in its absorbance spectrum, and the individual down here 00:32:31.01 would be an anomolous trichromat - one who had three pigments, but not the normal set of 3. 00:32:37.16 And this person would make different judgments 00:32:41.05 when asked to quantitatively mix different colored lights. 00:32:44.09 Now, more complex events are also possible 00:32:51.19 because many of the starting arrangements involve arrays that have multiple M pigment genes. 00:32:57.27 So, for example, if we start with an array that has two M pigment genes, 00:33:02.22 and a second array that has only one and envision the kind of recombination events 00:33:09.04 - in this case intragenic recombination events - that one might encounter, 00:33:13.17 we can, for example, see a meiotic product that has a hybrid gene, 00:33:18.18 an L/M hybrid instead of the L pigment gene, paired with an M pigment gene. 00:33:23.02 Now this person would also be an anomalous trichromat, 00:33:25.29 in this case, having replaced the normal L pigment gene with a spectrally-shifted hybrid. 00:33:32.01 And the person down below would be the reciprocal kind of anomalous trichromat, 00:33:36.12 now having a hybrid M/L and an unexpressed M paired with a normal L pigment. 00:33:43.07 These are quite common in the population - roughly 8 or 10% of human X chromosomes 00:33:50.19 have some sort of rearrangement that affects trichromatic color vision. 00:33:54.25 And, therefore, roughly 8-10% of males who carry those chromosomes 00:33:59.12 show that color vision anomaly. 00:34:02.18 There are also more subtle variations in color vision, 00:34:07.29 and let me show you the one that is the most common 00:34:12.02 and perhaps the most fascinating. 00:34:13.28 And that is a single nucleotide variation which divides the apparently color normal population 00:34:22.13 into two distinct perceptual groups. 00:34:25.25 This is in males, and among females, into 3 perceptual groups. 00:34:28.26 This is a single nucleotide difference in the L pigment gene, 00:34:35.20 which results in a single amino acid substitution, 00:34:38.13 it's a serine vs. alanine difference at position 180. 00:34:42.25 And, if you look at, in this psychophysical test, a mixing of red and green lights... 00:34:50.11 and here the result is plotted on the horizontal axis 00:34:52.25 as a ratio of red to red+green lights, and on the vertical axis 00:34:57.13 the number of individuals who had a given test result. 00:35:02.14 And here the mixture of red and green lights is being compared 00:35:06.04 to a standard, spectrally pure yellow light. 00:35:08.12 in just the way that Newton did this color mixing experiment. 00:35:11.27 What we see is that some people - let's look at the male curve at the top. 00:35:15.06 (males are at the top, females are at the bottom) 00:35:17.13 Some males require more red in the mixture to match the standard yellow, 00:35:23.29 others require less, and the distribution is bimodal. 00:35:27.18 And in the female set, there are some that test like the males who require more red, 00:35:33.24 some who test like the males who require less red, 00:35:36.22 and others who are in the middle - these are the heterozygotes in the middle. 00:35:40.01 And what we're seeing is the effect of this polymorphic variation 00:35:47.13 and its effect on the absorbance spectrum of the L pigment 00:35:53.04 which is now shown here, determined analytically by producing 00:35:57.09 the recombinant proteins and reconstituting them with the retinal chromophore. 00:36:00.24 So, the serine version - the one that has serine as position 180 - 00:36:04.08 is a little bit red-shifted (that's this upper curve) compared to the absorbance spectrum 00:36:08.28 of the alanine at position 180. 00:36:10.26 This is a 4 or 5 nm shift, and it precisely accounts for this variation in spectral sensitivity 00:36:18.19 between individuals who have either the alanine version or the serine version 00:36:26.11 of the L pigment. 00:36:28.00 So, even though we think, perhaps, simplistically as the world of trichromats 00:36:35.18 as a single unitary group of individuals with respect to their color vision, 00:36:41.07 that's not so - individuals are seeing the world somewhat differently 00:36:45.20 based on this genotypic different, and although each of them is internally consistent, 00:36:50.17 if you gather them together in a laboratory and ask them to quantify color mixing, 00:36:57.23 you'll see that they divide up into these distinct perceptual groups. 00:37:01.17 We saw at the beginning of this lecture, and also in the last slide, 00:37:05.11 how human visual pigments can be produced in the laboratory, 00:37:09.19 and their absorbance spectra measured directly after they are joined to 11-cis-retinal. 00:37:13.25 That's a wonderfully direct technique, and it gives us a very precise measure 00:37:18.28 of the absorbance spectra of those pigments. 00:37:21.03 Now, we'd like to ask, if we go back to the living human eye, 00:37:24.21 can we analyze the pigments' spectral sensitivities in the same precise way? 00:37:28.27 And, in this example, I will show how males who have only a single gene on the X chromosome 00:37:36.10 provide a perfect group for this sort of study. 00:37:39.04 So, these gentleman have lost, by homologous recombination, 00:37:43.26 all but one of the genes on the X chromosome. 00:37:46.14 What remains behind is either a normal L pigment gene 00:37:50.14 because they've lost the M pigment genes by homologous intergenic recombination, 00:37:55.04 or what remains behind is a single, hybrid gene, 00:37:58.21 containing either more or less of the parental contributions, 00:38:02.07 depending on the exact point of recombination. 00:38:04.16 But again, the other genes in the array have been lost by this recombination mechanism, 00:38:10.07 and these subjects are, in a sense, the perfect ones to study, because 00:38:14.03 they have only two cone pigments that one needs to examine, 00:38:19.25 the S pigment, and the remaining longer-wave pigment. 00:38:24.20 And, for simplification, one can design the study such that the S pigment 00:38:29.10 is really not in any way interfering with the sensitivity 00:38:33.00 measurements for the longer wave pigments. 00:38:35.08 So, one can use longer wave lights and selectively desensitize the shorter pigments. 00:38:39.09 So, in a sense, we have, in at least the longer wave end of the spectrum, 00:38:44.17 we have the very simplest arrangement one could possibly have. 00:38:47.06 We have someone who has just a single X chromosome, and on that X chromosome 00:38:50.29 is the single visual pigment gene, and we're going to ask, what is its spectral sensitivity? 00:38:56.07 And that is shown on this next slide. 00:38:59.08 This is determined by the method called flicker photometry. 00:39:03.22 It's a completely safe and non-invasive method. 00:39:06.27 The individual views a screen, on which is alternating about 20 times per second, 00:39:12.15 a standard light - a light of a standard wavelength (say 600 nm), 00:39:16.29 and with it, in alternation, is a test light whose wavelength is variable. 00:39:22.27 It could be 610 nm, 620, 630, 640, 650, and so on, 00:39:29.15 in 10 nm increments up and down the wavelength axis. 00:39:33.01 And what the subject has to do is to adjust a dial which controls the intensities 00:39:40.09 (the relative intensities) of these two lights 00:39:42.19 so that the sense of flicker is completely eliminated. 00:39:45.22 Now, let's ask how that could possibly happen. 00:39:48.21 After all, these are two differently colored lights - lights of different wavelengths 00:39:51.21 that are being presented alternating one with another. 00:39:55.02 How could the sense that they're flickering back and forth not be seen? 00:39:59.11 Well, remember that these subjects have only a single visual pigment 00:40:03.16 in the long wavelength end of the spectrum. 00:40:05.12 And, we've suppressed the S-cone pigment, so that it's really irrelevant to this measure. 00:40:10.20 And I should say, the rods have also been suppressed 00:40:14.07 by the high intensity of the lights being used here, 00:40:17.03 so for that one remaining pigment, if the number of photons captured per 00:40:22.01 second by these two different stimuli is the same, 00:40:26.03 then at that point, for the individual, there's no sense of flicker. 00:40:30.28 just like a continuous illuminant. 00:40:32.24 And that's the sense that is obtained when dial is set to just the right location. 00:40:38.15 I think you can appreciate that that gives us the relative efficiency of photon capture 00:40:44.03 for those two different stimuli. 00:40:46.12 And that, of course, is done for each of the different test wavelengths, 00:40:50.11 along the whole wavelength axis, and then we can plot out the full sensitivity curve. 00:40:55.20 And here they are. In green, we see the curve of sensitivity for the M pigment. 00:41:01.16 That is, these are individuals who have an L/M hybrid, but the L/M hybrid 00:41:07.05 consists almost completely of M sequences. 00:41:10.13 And, in particular, it consists of M sequences in those regions 00:41:13.14 which are critical for the spectral difference 00:41:16.09 between L and M pigments - so, de facto, these are individuals 00:41:19.12 who have only a M pigment and no L pigment. 00:41:21.21 And here is their sensitivity spectrum in green. 00:41:24.14 And then, in yellow and red are the two different variants of the L pigment: 00:41:30.22 one that has alanine at position 180, or serine at position 180. 00:41:34.20 I think you can appreciate that the serine at position 180, 00:41:37.08 just as it was in the in vitro spectrum, 00:41:40.12 is offset to the right, compared to the alanine at 180 version. 00:41:45.05 So, here, this is a very satisfying result. In the living human eye, we can see 00:41:49.07 the full spectral sensitivity of the pigments that underlie normal color vision 00:41:55.00 in the human population. 00:42:00.08 And, my only regret in describing this experiment is that most interesting of dichromats, 00:42:05.25 John Dalton, didn't live long enough to participate. 00:42:09.06 Dalton was born in 1766. I think if he had been born two centuries later, 00:42:15.11 he would have been an eager participant in this experiment. 00:42:18.12 Dalton maintained a lifelong interest in color vision and color blindness 00:42:23.18 and, even though he didn't participate in this experiment, 00:42:26.06 you may be interested to know that he made substantial contributions to 00:42:29.29 the science of color vision, and in particular, to understanding color blindness 00:42:34.02 even after his death, which seems surprising. 00:42:37.28 But, Dalton suggested, during his lifetime that perhaps the source of his color anomaly 00:42:44.24 was a selective filtering out of longer wavelengths of light 00:42:49.29 by some color medium, even in his vitreous - the fluid within the eyeball - 00:42:54.29 or perhaps in the lens, and so he stipulated in his will, 00:42:58.12 that when he died, his eyes should be cut out, 00:43:02.04 and a little hole cut in the back of the eyeball, and 00:43:06.26 by viewing the world through that hole, it would possible to tell whether the world was 00:43:11.05 being filtered by some sort of color filter within his eyes. 00:43:14.10 And so that was duly undertaken, upon his death, 00:43:18.09 and the result was that, except for a bit of yellowing of the lens, 00:43:21.17 which was to be expected in someone who is somewhat older, 00:43:24.29 the path was completely clear. 00:43:27.01 There was no filtering at all by the hypothesized color filter. 00:43:32.23 And, therefore, Dalton's color blindness must have been due 00:43:37.21 to some difference within the retina, 00:43:39.17 or possibly the brain. And, this was a major conceptual leap 00:43:46.01 in understanding the origins of color blindness, 00:43:47.29 and it would have been a substantial contribution if nothing more had been done, 00:43:54.22 but in fact, Dalton continued to make contributions to the world of color blindness research, 00:44:00.18 and 150 years later, his eyes, which had been preserved 00:44:04.25 by the Manchester Literary and Philosophical Society, 00:44:08.15 were examined by DNA analysis, and a small tissue sample was taken for that analysis, 00:44:14.25 and it was shown that Dalton was, in fact, missing the M pigment gene. 00:44:20.27 And, he was a dichromat of just the sort we've seen before 00:44:24.17 who has only an L pigment gene and an S pigment gene. 00:44:28.05 I think Dalton would have been very pleased with that result.
