 Good morning, TAs. Any announcements? No news is good news. Today I'm going to talk with you about olfaction, and so Ralph Adolf's raised the question in every lecture. He asked what is vision, he's going to ask what is touch, and so the question is what is olfaction? Any comments? Well, I presume that one could say that olfaction is knowing what is where by smelling. So here is a remarkably gifted olfactory animal with a nose large enough to house all of those olfactory receptors. We're going to talk about olfaction among vertebrates, worms, and insects today. Proust has this wonderful passage at the beginning of Remembrance of Things Past about the Madeline that he tasted when he was a grown man, but that reminded her of the lime blossom that her aunt used to give him, that his aunt used to give him, reminded him of the grey house on the street where her room was, where the pavilion stood, the entire set of very pleasant emotional memories associated with his childhood just evoked by the little piece of Madeline, which is a, well you can get Madeline at Starbucks these days, they're the little scallop shaped cookies. So the nose can in principle classify thousands of different compounds, but what is it that's special about the memories associated with olfaction? We're going to discuss that mapping of the compounds to the brain, you can classify thousands of different compounds, possibly matching to memory templates, categorized on one's previous experience is it, and the other sensory stimuli, but what is special about the olfactory system that is able to bring emotional memories? We're not sure, certainly odorants or volatile compounds can be detected by the olfactory sensory neurons in the nose. Part of the answer may be that outputs from the olfactory both go directly to the cortex, olfactory cortex and as we'll see other regions of cortex without passing through the thalamus. So the generalization that we stated in the last lecture that nearly all sensory information goes through the thalamus does not apply to the olfactory system. And perhaps the olfactory stimulus goes to areas of the limbic system, which is the built around the ventricles, Ralph will talk more about the limbic system, that it mediate emotional and motivational responses. Certainly the amygdala and the anterior hypothalamus are involved in these responses. So we will go through the olfactory system, in this case using my conventional method from the receptors to the brain, and from the molecules to larger processing. Here is the beginning of the first figure in the chapter from Kandel on sensory processing. We're going to talk about the ends of the olfactory sensory neurons, the sensory cilia. We'll talk about the olfactory neurons themselves, which are in the olfactory epithelium. We will talk about the olfactory bulb, which is the first way station, and then projections directly to olfactory cortex. There is the outside world is right here, and the outside world does need to project into the brain. It does so through these holes in a plate called cripper form. We will talk about this in detail later on. First of all, in terms of what the olfactory system can distinguish, it's really quite amazing. Carvone, for instance, has a stereo center. It is a rather simple molecule, three double bonds, but it exists as two isomers. So although a single odorant receptor can respond to many different compounds, we see a lot of selectivity in odorant binding, or we can study a lot of selectivity. So we can distinguish the stereo isomers of carvone, and the R isomer smells like spearmint, and the S isomer like caraway. So this tells us that we need molecules capable of distinguishing stereochemistry. This is very likely on first principles, and we'll see this has been wonderfully corroborated by experiments. This is very likely then to involve proteins. As you know, proteins are made of stereo specific amino acids, and so it's possible for proteins to distinguish many, many examples of stereo specific compounds. In comparison, the two major branches, the two major aspects of the mammalian olfactory system, the main olfactory system, and the accessory olfactory system. So the main olfactory system starts with the main olfactory epithelium, and epithelium, for those of you who need a little bit of refreshment, is essentially a single layer of cells. The skin has epithelia, and many organs have epithelia, or endothelia, again, single layers of cells. The vomeronasal organ is also in the nose, also an epithelium, but in a different place in the nose, more anterior. Main olfactory epithelium projects to the main olfactory bulb, as I showed you on the previous slide, which then projects to various places in cortex. The accessory olfactory nucleus, pyramidal, sorry, piriform cortex, otherwise called visual cortex. The olfactory tubercle, the lateral amygdala, and entorhinal cortex, so these last two are the amygdala and the cortex, are again in the limbic system. The VNO vomeronasal organ neurons project to another olfactory bulb, the accessory olfactory bulb, and these go directly to the ventral amygdala and to the hypothalamus, which are heavily involved in emotion and motivated behavior. If we look in more detail at the main olfactory epithelium, we highlight, this is like one of the figures in Kandel, we highlight the cells of the main olfactory epithelium among mammals. Certainly the olfactory neurons are the key to this story. I should point out that there are lots of other cells in the olfactory epithelium. There are these so-called basal cells, supporting cells, and although we won't discuss it in great detail, the olfactory epithelium is capable of regenerating if it's damaged. So there are many ways of damaging the olfactory epithelium, and these basal cells then become like stem cells and they regenerate and find their ways to the proper regions in the central nervous system. This is one of the few places where you will remember that when you cut the optic nerve, the ganglion cells can find their way back, but here, if you destroy the entire neuron, the neuron that takes over can still find its way back. So the olfactory neuron has a dendrite. The dendrite has extensions themselves called cilia, or ciliary extension. In the transduction, the sensory component, the molecules involved are right here in the cilia. The cilia actually are embedded in a mucus layer, and the olfactory neurons not only can be destroyed, but they turn over naturally and are replaced every 60 days. Then the new cells find their way to the correct glomeruli. So if we look at olfactory receptors among vertebrates and in most other phyla, they are, with an exception that I will get to later in the talk, seven helix G-protein coupled receptors. In mice, more than a thousand genes, and perhaps in my dog Kobe also, more than a thousand genes, two to three percent of the genes, encode sensory receptors. We make less use of our olfactory system. We have about 350 functional odorant receptors. Many of the other odorant receptors are pseudo genes. They have acquired mutations and do not get expressed. Receptor sequences are quite variable, especially in the regions that are the putative odorant binding helices, so the binding site for the odorant. This makes the diverse repertoire, and in mammals, as I'm going to show you, each neuron probably expresses only a single receptor. Now how that gets done, what are the mechanisms that allow a neuron to express a single receptor is a topic in allelic exclusion that those of you who've studied immunology have probably studied. The mechanisms may be different in the olfactory system than in the immune system. However, as in the immune system, the olfactory receptors are clustered together. There are a couple of clusters of olfactory receptors, and they presumably are able to interact with each other's expression. Moving on in candel, we have the cilia in the olfactory epithelium, and we have the G-protein pathway. The start of the G-protein pathway in olfaction among vertebrates is rather similar to the start of other G-protein pathways. We have an odorant binding to a receptor, in this case in the transmembrane domain. We have it activating G-proteins. We have the G-protein dissociating and activating an effector. The entire process is fairly localized. It's rare for it to be this fast. Usually, olfaction is a bit more leisurely than vision. If we look at our usual map of the G-protein coupled receptor pathway, leading from receptors to G-proteins to effectors to intracellular messengers, there are a couple of differences. You may remember that we have been talking about the intracellular messengers. When we first brought up the G-protein pathway, we pointed out that the intracellular messenger typically activates kinase, may eventually get into the nucleus, but that a few ion channels in the olfactory system and in the retina are activated directly by the intracellular messengers. We talked about the fact that this occurs in the retina, although in the retina, as you will remember, activating rhodopsin decreases the concentration of the intracellular messenger, which is a cyclic nucleotide cyclic GNP. Here in the olfactory system, what are the differences? Well, we have a receptor I've showed it to you. We have a G-protein. We have a new G-protein that I have not put on the usual diagram. It's called G-olf, but I put it in the same square where I usually put Gs, the stimulatory G-protein, because in fact G-olf is very similar to Gs and does activate an enzyme. The enzyme that it activates is the denital cyclase, just like the enzyme that activates Gs, and the resulting cyclic AMP goes on to activate a channel. Why don't we stop here and have a bit of a quiz? Show me that's what I was doing at the bookstore, getting index cards. So this one is on the visual system, olfactory transduction. We've talked about the receptor, the G-protein, the effector, which is an enzyme, cyclic AMP, and as usual, you can look at this pathway using the patch clamp, in this case using the excised inside-out patch. When one has a patch from an olfactory receptor cell, or since it's tough to patch clamp acilium, if you take the olfactory receptor transduction channels and put them in a heterologous system, no channel openings with no cyclic AMP, and then add cyclic AMP to the bath, bathing the inside of the patch, in this excised inside-out, we get openings due to cyclic AMP. So this tells us that we have ion channels activated by an intracellular messenger, in this case, cyclic AMP. Any questions? So more about olfactory channels. They are permeable both to sodium and to calcium, so odor and binding causes depolarization of the olfactory neuron via sodium entry exactly, or as though this channel were a glutamate receptor, which it's not. In some cases, calcium also enters and activates a chloride channel, and in some cells, as you remember, the chloride equilibrium potential is near zero, rather than near the potassium equilibrium potential. That's true in this case as well, and so this process can increase the depolarization and stimulates the olfactory neuron to fire action potentials. So here again we see a difference between olfactory neurons and your response to the question in the quiz that you just took. In the retina, you need to go through several sets of neurons, several stages in the transduction pathway, until you find a cell that fires spikes, the retinal ganglion cells, but here are the primary neurons due fire action potentials. They have voltage-gated sodium channels, and so they can get through to the olfactory bulb. Oh, I think I made a mistake here, but the olfactory epithelium itself is this highly involved epithelium, so it has a vastly increased area, so that lots of olfactory neurons can be exposed to the olfactant. Interestingly enough, there are subzones of the olfactory epithelium. So this is a coronal section through the rat nose, which has been in-situ hybridized for the olfactory receptors. You see these little white spots here. There are distinct expression zones for among the receptors, that is, they are more or less expressed at random, but only in one of these four expression zones. On the other hand, another gene class important for olfactory transduction, which is expressed in all olfactory neurons, indeed does hybridize for all neurons. So we have subsets, but within those subsets, expression seems to be random, but each cell expresses only one olfactory receptor type. Something like one 250th of the neurons within a zone express each receptor. There are about a thousand of them, at least in the rat. And neurons within each expression zone send axons to a different quadrant of the olfactory bulb. And so the olfactory bulb is developing in parallel with the four regions of the olfactory epithelium. So in a little more detail, here we have another figure from Kandel. Here is the olfactory sensory neuron. It is sending its axon through this cribiform plate, and glomerulus at the other side of the cribiform plate. I think glomerulus is like a ball of yarn, and so the glomerulus is a tight, knit bunch of synapses and processes. Looks like a ball of yarn. The mitral cells, on the other hand, in the olfactory bulb, are the major components, the major projection neurons of the olfactory bulb. So here's a mitral cell, and they get it their names because they look triangular. They have a peaked top, like a bishop's mitre, a bishop's hat, that's why they're called olfactory cells, glomeruli, olfactory sensory neurons. And the cribiform plate gets its name apparently. This is not a top, this is not a word I use often, but it is perforated, and apparently in Latin the word for perforation is crib. So, well, I guess like a child's crib, which is perforated. Okay, any questions about the details here? Now, remarkably enough, although the, so this is a stain for a cell expressing a single type of olfactory receptor, and so there are roughly a thousand of these. This is done by making a knock-in mouse that has beta-galactosidase stuck into the coding region for one of the receptors. You can also do this with GFP, of course. But in this older diagram, only a few of the receptors in the animal are expressing beta-galactosidase, like Z. And remarkably enough, although they come from different regions of the turbinate, they all converge on a single glomerulus. They could send to other glomeruli as well, but many converge on a single glomerulus. Now, how does that happen? What is the chemo affinity? This is not a solved problem. There's a little bit of a hint, well, it is clear that the olfactory, the only thing that differs from one olfactory receptor neuron to the other is the olfactory receptor itself. Otherwise, they express the same genes. And indeed, it has been shown that the RNA for the olfactory receptor is expressed in the synaptic terminal as well as down here in the cell body. And so the thinking is that perhaps the olfactory receptor is also expressed in the synaptic terminal and helps to find the post-synaptic cell. But in fact, really, there are no good ideas about how this happens. Lots of experiments, but lots of ideas, but no proof. So this is a topic that one of you will have to solve as a graduate student or as a neuroscientist. So here is an example which really, olfactory system has been revolutionized more than other systems by the ability to image calcium and to see the activity in the projections of the odorant receptor. So this is an image of the olfactory bulb in a fish. It's slightly more straightforward to study olfaction in a fish because the stimuli are soluble amino acids typically. And so you can wash them through. So here are amino acids. Here are the glomeruli being studied. These are the delta F, the fluorescence associated with a stimulus, meaning that the glomerulus is activated. And so clearly specific glomeruli, or reasonably specific glomeruli, are also activated in response to odorants. In mice, the studies like this have moved also. And for instance, there is the nucleon-expressing I7, which is a receptor for octenol. And sure enough, if one now looks at the glomeruli to which I7 express, one can see calcium transients in those glomeruli. And interestingly enough, if you move the I7 glomerulus to the wrong place in the bulb by transferring the I7 gene into the genomic locus for another receptor, the axons find the moved glomerulus as well. So there are some remarkable aspects of olfactory perception and of olfactory maps, which we do not understand. An individual the piriform cortex, which is part of the olfactory cortex, receives projections from mitral cells. And each piriform cortex neuron receives projections from many glomeruli. The mitral cells also project to the olfactory tubercle and other areas. And so although integration of olfactory responses and identification of odorants may probably take place mostly in the cortex, there's some integration that occurs in the bulb, as you could see from the tightly wound glomerulus. Now let's stop, let's move from discussing the olfactory main olfactory system, main olfactory bulb, and olfactory cortex to this additional fascinating part of the olfactory system, the vomeronasal organ, or the accessory olfactory system. It is thought to respond to pheromones. We'll define pheromones a little bit later, mostly in rodents and among vertebrates that are not primates. It is very well developed. It's cup shaped near the front of the nose. And as we'll see in a moment, the neurons have basal and apical regions. So the apical region is here where the odorants come in. At the apex of the cells and the basal region is out here. The lumen right here, lumen meaning hole, is where the pheromones enter. The transduction channel and the receptors are on the microvilli at the edge of the lumen, very reasonable. So here is an antibody to a channel called a trip channel, which is thought to help depolarize the cilia in response to the first calcium input. So trip channel meaning, well, I won't tell you where the word comes from because that'll confuse you even more. But it's another of these channels that are activated by intracellular messengers, in this case by calcium. So the VNO receptor molecules, and this comes from the book too. The chapter in the book was written by or in association with Richard Axel, who won the Nobel Prize for discovering aspects of this system. So there are two distinct families of VNO, vomeronasal organ, G-protein coupled receptors. They are quite distinct families. That is, they are not related to main olfactory epithelium receptors. They arose independently. The two families are appropriately enough, expressed, again, each VNO neuron expresses only one receptor, as in the main olfactory organ. The two families are called appropriately enough vomeronasal receptor number one, and vomeronasal receptor number two. The protein sequence looks entirely different. V1R receptors have seven transmembrane regions, so do the V2Rs. The V1 receptors have much shorter transmembrane domain, n-terminal domains, than the V2 receptors, and that typically means that the ligand, in this case the odorant, in the case with the short extracellular domain, is binding in the membrane, whereas with a long extracellular domain, the hypothesis is usually that the exciting molecule, in this case the pheromone, is binding out here in the long extracellular region. So these two layers of vomeronasal receptors occur in different regions of the VMO. Remember it's cup shaped. Here's the lumen. Here's where the molecules, the pheromones, come in. The two regions and the two receptor subtypes also are coupled to different G-proteins. There is a GI expressed in the V1 region. GI is one of the transduction molecules we've learned about. I stands for inhibitory, but in this case it's probably not inhibiting cyclates. And there is a GO expressed in the other set of vomeronasal cells. So this is quite a complex system. And in the vomeronasal cell, we have a different intracellular messenger expressed. Now, instead of cyclic AMP and cyclic GMP, we have diacylglycerol expressed as well as IP3. And you may remember that this is appropriate to the GQ family of G-proteins. The GQ activates a phospholipase, which splits up phosphatidyl inocital 4-5 bisphosphate into diacylglycerol and inocital trisphosphate. So this is the transduction pathway that is followed by the D2 vomeronasal signal transduction. And this activates the TRIP-C2 channel directly with diacylglycerol. So yet another type of ion channel being activated by an intracellular in this case, a membrane-bound ligand, it's called the TRIP-C2. All right, I'll tell you what TRIP-C TRIP stands for. It stands for transient receptor potential, because they were first discovered in the retina of insects rather than in the olfactory system. But here, as you can see, they're in the accessory olfactory system in the vomeronasal organ. So what do vomeronasal organ receptors respond to? Well, in many cases, they respond to urine. Fascinating. And if you've taken your dog for a walk, you understand this, right? Dog stops at every bush. My wife calls it the nose book. Some neurons selectively respond to urine from mice of the same sex, others to urine of the opposite sex. Responses are rather narrowly tuned among when people do physiology on the VNO organ. Each neuron seems to respond to only a single compound. Now, maybe the experimenter did not visit enough trees or enough fire hydrants, can't tell, but there was only one compound that could be extracted from urine that activated the VNO. Interestingly enough, mice can produce ultrasonic calls, whistling in response to contact with urine from the opposite sex. And so there actually has been a renaissance in looking at mouse calls over the last few years here at Caltech. This was started by Paul Patterson and Elaine Chao, now being carried out by Wei Li Wu. But that is usually in response to stress. Here, mice produce ultrasonic calls, whistling in response to urine from the opposite sex. And apparently, if you destroy either the VNO, the vomeronasal organ, or the main olfactory epithelium, these calls stop. So in trip C2 knockout mice, which may be the transduction channel for part of the VNO, the neurons don't respond to urine, and mice don't vocalize in response to urine. So the presumption is that the diacylglycerol, which is released by the GQ pathway, is activating the trip C channel. And so now let's look at the accessory olfactory bulb projections to the brain. They have the mitral cells, have dendrites that arborize in multiple glomeruli, not just one. The ALB projects to the amygdala directly, and the hypothalamus via the amygdala. And so the projections from the rostral and cartil ALB halves superimposed in the amygdala. So the assumption then is that a lot of the integration of pheromone signals takes place primarily in the accessory olfactory bulb, which is probably different from the main olfactory system. And the main olfactory system as we responded, the integration of signals may take place in the cortex. Here it seems to take place in the bulb instead. So what are the generalities that we can make about the Omenol factory system and about the vomeronasal system? Well first, the main olfactory system among land-based animals mediates cortical responses to volatile odorants, and these cortical responses drive conscious behavior, food seeking, predator avoidance, et cetera. Pretty much unconsciously is the way the vomeronasal system works. So unconscious responses, water-soluble pheromone compounds in urine and in secretions of other individuals. So remember we said in previous lecture vision is knowing what is where by seeing, and in this lecture olfaction is knowing what is where by smelling. Well clearly if a person's perceptions are unconscious that twists the definition of knowing what is where, and it's not possible really to know what is where these are unconscious responses. Confusingly enough, even though we don't have a vomeronasal organs, we can detect and respond to some pheromones, including ones that control the menstrual cycle. So there are lots of mysteries to be understood yet about the roles of odorants and pheromones in human olfaction. Let's look at some knockout mouse data, which are fascinating. So trip two is being abbreviated here as trip C2. Here on the right, here on the left, we have a trip C2 wild type mouse that it has two copies of the trip C2 gene, or just one plus minus, so it's the heterozygote. There doesn't seem to be much difference in the behavior of the homozygote and heterozygote. That is, it seems as the one copy of the gene for the trip C2 channel is enough. And so these mice attack intruders introduced into their territory, especially they get a special kick out of attacking intruders which have been swabbed with urine from males. All right, so these dots on the ethogram are attacks by the subject mouse. And the control intruder evokes some attacks. The intruder swabbed with male pheromone gets a lot of attacks. However, the trip C2 knockout mouse does not sense the intruder or does not respond with attacks to the intruder, even when the intruder is swabbed with the urine. So we say then that at least in this case, also trip C2 knockout males mate normally with females, but they also mount males which control mice never do. And so the trip C2 knockout phenotype suggests that at least for this behavioral assay, the default pathway in the absence of VNO input is to mate with anything. And furthermore, apparently VNO input causes male mice to fight rather than to attempt to mate. So what are the chemical nature of pheromones? Well, pheromones include prostaglandins in fish, asteroid and rostinone in pigs, and proteins in hamsters such as a protein called aphrodysin. But usually individual compounds don't elicit strong responses. Natural pheromones are mixtures of many substances. Maybe they are all carried by proteins plus the proteins carry the pheromones plus small organic compounds. So it is a mystery how an individual vomeronasal receptor can respond to this mixture of pheromone molecules, but it does. Now let's go to a couple of model systems. We have a little bit of time. I think we can do this. First of all, the worm nematode, C. elegans, entirely different olfactory system. The behavioral assay here is that C. elegans can find and go toward volatile attractants and repellents. Here is the nose of the worm C. elegans. There are several well-described neurons and remarkably the entire set of olfactory processing in C. elegans uses only two pairs of neurons called AWA and AWC. So there are lots of olfactory receptors, so therefore each chemo-sensory neuron has to get inputs from many of these. So here is a AWC neuron expressing GFP. You can see its cell body here, then goes out to the periphery, and it has the cilia, which do the sensing. The dendrite is quite long. So the odor 10 receptor is expressed in AWA, including its dendrite, and AWA ODR10, among other ligands, responds to diacetyl, which is 2, 3-butane diome, and worms that are knockouts for odor 10 are not attracted to diacetyl. So the phenotype of the odor 10 knockout doesn't respond to acetyl. It also does not respond to 2, 3-pentane diome, interestingly enough differs by only one group. ODR10 mutants still chemotax to 2, 3-pentane diome, and so this is remarkable specificity. And AWC cells, the other sensory cell, has receptors for 2, 3-pentane diome. So we have a duality, one receptor for one type, one receptor for the other, and if you destroy AWC, the other olfactory sensory cell, the animal can't chemotax to pentane diome. More than responding to one cell though, odor 10 actually diacetyl, we have to understand where diacetyl comes from, and it comes from bacteria. These bacteria typically live on citrate as their carbon source, and sure enough, odor 10 also recognizes the metabolic intermediates, citrate and pyruvate. So diacetyl has become a volatile signature for some bacterial species. So other bacteria don't make diacetyl, but they make acetyl in or lactate, and diacetyl attraction therefore allows the worm to recognize some food sources at some distance in the soil, sort of like a dog figuring out where you are from afar, and so it is possible that the non-volatile interactions of citrate and pyruvate, which can also be sensed by odor 10, can provide, first of all, a long-distance attraction, and second, a more local taste-like attraction after the worm arrives at a bacterial colony. So remarkable aspects. Moving on to insect odorant receptors. We've had a surprise recently, although the drosophila olfactory receptor seems to have seven transmembrane domains and was thought for many years to be a G-protein-coupled receptor. It may instead be part of an ion channel. So here we have a gene family of 60 genes encoding insect olfactory receptors. For insect olfactory receptors, we also have an auxiliary subunit, which also looks like, has seven transmembrane domains and looks a bit like a G-protein-coupled receptor. However, it seems to occur in most olfactory receptors. So we have a variable subunit, presumably dimerizing with an invariable subunit. The names are various because this fixed receptor has been discussed, has been discovered by several groups, got several names, but the terminology is converging on orco, which means olfactory receptor perhaps common or something like that. So it may very well be that instead of the second messenger operation of vertebrate olfactory receptors, vomeronasal organ, even worm olfactory receptors, the insect olfactory receptor may be a direct ligand-gated ion channel activated by the odorant. And you can look in the book for the structure of the drosophila olfactory system, which is quite different from the one I've described here. And of course, Biddy Hong here at Caltech, one of our wonderful assistant professors, is working on this system. So here's a nice summary of the similarities and differences in olfactory receptors among vertebrates and insects. C. elegans is not here. As I've just told you, vertebrates have GPCRs, insects are not. Insects have a smaller repertoire. We can go down the list here about topology. One of the striking hints that the insect seven helix receptors may not be GPCRs is that the sense of their seven helices is backwards. Start out with the n-terminal domain inside, go backwards. Among vertebrates, there are lots of pseudogenes, especially among non-olfactory animals like us, very few in insects. Vertebrate olfactory receptors are monomers, each of them a G-protein-coupled receptor. Insects are at least a heterodimer. One receptor, one neuron, generally true. Gene selection among vertebrates is this highly interesting stochastic selection, very analogous to, but probably mechanistically different from the immune system. Among insects, it's deterministic. Zonal expression in both cases. Instructional role, I wouldn't take that too seriously. Nobody knows how the vertebrate receptors find their topics. A topic expression of the receptor, meaning you can also see it in the axon terminal, not known among insects. Some odorants can be inhibitory. That's easy to understand if the receptor is an ion channel. Glamaryli in both cases, and the number of glomeruli per receptor type is variable among vertebrates. Probably there is just one glomerulus per receptor type among insects. So complex, interesting parallel evolutions of the way animals respond to cells, and I think that's it for today, and Ralph will be talking next week. I'll be at office hours this afternoon.