 Thank you, Anna, very much. Thank you to all the organizers. I've had the pleasure of being here in Trieste before. As Anna mentioned, we've worked together for many years, so I visited this place a number of times. It's just a wonderful, fabulous institution. And I have to say the mission of ICTP and CISA has, in my opinion, never been more important since the Cold War in the world we are currently building. So we really do need to support this organization importantly. So and I particularly want to welcome the students and the postdocs to this. We ran a school here several years ago and it was, I think, quite successful and it was a brilliant time having interaction with the students, postdocs from all over the world. All right. Let's see how much I can tell you in an hour. I'm going to talk about how biology perceives chemistry. There are two screens here, which is always a mess for a speaker. So I'm going to sort of concentrate on this one because I think it's a little easier, maybe. I don't know if it is or not. I'll flip back and forth, probably. What? This is, oh, this is in the way. Let's see how long that lasts, all right? Well, it's on a short, see it's on a short wire. It can't fall very far. Oh, there you go. Okay. There goes everything. What happened? Oh, I see we lost the connection, though. Yeah. Let's see if we can try this again. All right. So I'm going to sort of with some introductory material since I'm also the first speaker and I'm sure not everybody in the room is so familiar with the olfactory system. Those of you who are, please, your patience is requested. I think I'll be able to make a point out of some of this introductory material as well, however. So let me start at the very beginning, of course, as sensory scientists, we all accept the idea that basically what the brain knows about the world comes to us through these little holes in our head. Virtually all of the information the brain knows about what's going on out there comes into us through these little holes and I guess the skin that covers your body. And of course, behind each of these little holes in the head are some specialized tissue that's sensitive to mechanical stimuli, to electromagnetic stimuli, or our interest here is to chemical stimuli in particular. So of course, we know something about, you know, many things about, for example, color vision and the use of three receptors that are maximally sensitive at different wavelengths and because they overlap, we're able to see millions of different shades or hues of colors and the brain interprets it in various ways. Similarly for the auditory system, very complex sounds are nonetheless broken down by the basilar membrane in the ear into their constituent sinusoidal waves and then the brain builds them back up again in a kind of a Fourier transform sort of mechanism we imagine into the complex sounds that they are. So how does this work for olfaction is of course our big question, how do we take something like a rose which has a very unified kind of smell, something that we would all identify with no trouble immediately and yet it's composed of a disparate group of chemicals. These are just five of the major constituents of the smell of a rose and you can quickly see that unlike, say, vision or sound, these do not vary along a continuous physical dimension like wavelength or frequency but rather there are multiple chemical dimensions here. So there are different functional groups, there are aromatics and aliphatic molecules, different molecular weights, different functional groups, all sorts of differences, volatility and so forth and yet somehow or another in your brain they all add up to the notion of a rose and so that's sort of the overriding question of course in any sensory system but in particular in olfaction and it remains somewhat mysterious although I think we have a lot of interesting ideas about it. Quick introduction to the olfactory system, you'll hear more about this I think from Charles Greer in the later talk. This is a sort of a cartoon of a rodent head in a sagittal section so you see the nostril, the opening, the nares, the nasal cavity which can be quite large but it's only at the back of the nasal cavity that you have this epithelium that covers some of these cartilaginous outcroppings called turbinates and so this increases the surface area available to things you snip up your nose. We call this the main olfactory epithelium, the sensory epithelium. There's also this other system, the vomeronasal organ which Tim are you going to talk about that? Nope, okay. Is anybody going to talk about the VNO? Sure, I don't know so we may not talk about this but it's a sort of an independent but integrated second olfactory system used by many animals but I won't say any more about it now because I'll just confuse everything. The cells in the main olfactory epithelium which is what I'll be talking about primarily today are primary sensory neurons, they're a true neuron I'll show you that in a moment and they send their axons through this very thin bone back to the structure here called the main olfactory bolt. So their axons just go back to the main olfactory bolt where they synapse with second-order neurons of various sorts. There'll be much discussion about all of that later. And then from the main olfactory bolt, these secondary neurons, these, yes, this next level of neuron, send their axons to an area, primarily to an area of cortex known as piriform cortex or olfactory cortex. There are several other places where they go as well but the piriform cortex is their main target. The thing I'd like to point out about this is that this is a very shallow circuit. Basically, you have two synapses from the outside world to cortex. Just two straight through synapses. Now, there's a lot of processing that goes on horizontally, if you will, and a lot of interneurons and things like that. But the straight through pathway is only two synapses. That's quite remarkable if you think about it. In the visual system, you would still be in the outer retina, let alone anything close to cortex. Here's a rather dramatic picture because everybody has to show a picture by Ramoni Kahal or should, I think. This is from Ramoni Kahal's work in the olfactory system, which he did quite a bit, and I imagine you'll see some more pictures of this from Charlie. So here are the receptor neurons, the olfactory sensory neurons out in the nose, if you will, in the epithelium. You can see they're a bipolar neuron. Talk about this more in a moment. Again, they send this single unbranched axon back into these structures called glomeruli. This is now the olfactory bulb area. And those glomeruli are the spherical structures. You'll see great Gilmore this later too. And these second order neurons, the mitral cells, send a single dendrite into one of these glomeruli where they synapse extensively with axons of the olfactory sensory neurons. They then send their axon down something called the lateral olfactory tract. They bundle together and go back to pure form cortex where they synapse when the dendrites primarily have pyramidal cells. So once again, two synapses to the cortex, one here in the bulb and one back here in the cortex itself. It gets you from the outside world to a fairly high level region of the brain. I'm gonna talk to you primarily about these sensory neurons out in the periphery because of course, they somehow or another limit what the brain is going to find out. So whatever processing may go on with their information, this is the raw information about the chemical world, is embodied in the activity of these sensory neurons. So they're a very simple, it's a very simple tissue which makes it experimentally very available if you will and convenient. There are only three types of cells in this tissue, these supporting cells, the basal cells and what I'm gonna talk about today, these olfactory sensory neurons. Again, a simple bipolar neuron, a long axon that goes back to the olfactory bulb unbranched. There are no connections between these cells that we know of in the epithelium. So there's no gap junctions or anything of that sort. They also send a single dendrite or rather thick dendrite up to the surface of the tissue where they extend from their, this swelling called the olfactory dendritic knob. And then from that are these up to a dozen or so very fine hairlike structures known as cilia and again increasing the surface area in contact with odors that come in here. And of course, this is all in a layer of, a thin layer of mucus. Again, what's most interesting to us is what goes on out here because this is where the odor molecules, the chemicals of the environment come into contact with the brain or the nervous system in its first instantiation. So what happens out there? I'll tell you very quickly because I don't wanna get into this in a long way but there is a group of molecules, a group of enzymes, mostly in the lipid membrane of the cilia. The most important perhaps is the receptor and you'll hear plenty about that today from me and others. So there's a receptor, a G-protein couple type receptor. So that's hooked up to a G-protein which has three pieces and when it finds an odor the alpha subunit comes off, activates yet another molecule called an adenyl cyclase producing the second messenger cyclic AMP which opens a special channel. When that channel opens, positive ions flow into the cell, sodium, some potassium and in particular calcium and that calcium actually serves to open a chloride channel which in these cells is somewhat unusual as chloride goes out of the cell because of a high concentration inside the cell and so that further depolarizes or activates the cell. The important thing for our purposes here today is that every olfactory sensory neuron has identical, this machinery as far as we know, identically. They're all precisely the same in possessing all of these enzymes. What makes one different from another is the particular receptor they express from a large family, of course first discovered and cloned by Linda who's trying to sit quietly in the corner but we'll talk much more about that later. And so these receptors essentially are what makes any olfactory sensory neuron different from any other one. Otherwise this machinery here, the only important thing to know about it for today at least is that the result of the activation of all this machinery by the binding of an odor to the receptor is an increase in the intracellular calcium. And increasing intracellular calcium is something that we can now, by a variety of technical methods, recognize quite easily and quite high resolution. So whether it's with a calcium sensitive dye like FURA or for most of the data I'll show you today, genetic modification called GCAMP6, which we can induce olfactory neurons to express and when they bind calcium at a certain concentration they fluoresce. And so now we can visualize optically whether or not a cell is active by just looking at the calcium level in the cell. So an increase in calcium essentially is a proxy for the binding of an odor to a receptor and the activation of this whole pathway and presumably the spiking of the cell and sending its message to the bulb. Okay, that's pretty much the introduction for the most part. Now a favorite trope in the olfactory field for many years has been this idea of getting from molecules to perception that we can go all the way from chemistry out there all the way up to perception. So in the case for example of a rose, how do you get from these molecules to the perception of a rose? I would say for the most part for many years now this field that's generally meant that we're going from chemistry, these molecules out here to psychology or what's sometimes called psychophysics. I suppose I should explain that here at the ICTP. This does not mean crazy physicists like you would find in the ICTP perhaps, but rather the combination of psychology and physics to give us the word psychophysics. And this comes from early work in the late 1800s I suppose even going back to then when a number of physicists believe that they can understand sensory systems and the psychology of perception by thinking about the physical stimuli and how a physiological entity would interact with that. It's been a very, very useful area of study. It's been very important. And we use it in olfaction as well. So olfactory psychophysics, I guess you might call it psychochemics or something like that, but we still use the term psychophysics. But typically to get from molecules to perception, what we mean is we're getting from chemistry to psychology. What I'd like to suggest is that what we often forget because it's been difficult but is now possible is that to get from chemistry to psychology, you need a little biology. You need a little wet stuff in between. And today I'd like to talk about the role of biology and all that. Well, I'd like to, but I don't think I'm going to. What happened? Oh, I see it wants to update. Ah, let's see. I don't really know exactly how to, let me see what I can do here. It wants to update my God damn computer. Well, I don't have the, I have to turn, I'll be right back. Don't go away. Don't go away. Everything's going to be fine, honestly. Hope I don't get any other messages I'm not aware of coming here today. Who knows what's there. This should come on in a second. There it is. Okay, chemistry to find out. Okay, so in order to talk about this, I think we have to go back and ask a very fundamental question, which is simply, it's going to sound almost silly, what is an odor? What actually is, do we mean by an odor or an odorant perhaps would be a better word? I'll use, I'll try and be regular about this and use odorant to refer to molecules and odor to refer to a perception. But I'm not so good at that, so I'll do my best. So what is an odorant actually is what this slide should say. So these are typical chemical characteristics of odors. They're almost all organic compounds. They're fairly low molecular weights. They tend to be fairly volatile. They're both aromatic and aliphatic in structure. That is they may have an aromatic or benzene ring or they may be a straight molecule. They may be saturated or unsaturated. And they use a variety of so-called functional groups, aldehydes, alcohols, et cetera, et cetera. The difficulty with this is that while this is very good at explaining the characteristics of virtually every odorant we know, it also explains the characteristics of many chemicals which are odorless as far as we know. They're odorless to us and odorless to many other animals as well. So the large number of odorless chemicals that nonetheless fulfill all of these chemical characteristics. And so I'd like to say, to start sort of back at the beginning in a way, that the best way to define an odorant is actually, or an odor, is actually in a kind of operational way. And that is they bind to odor receptors. A chemical that binds to an odor receptor is by definition an odor. That may sound trivial. It may continue to sound trivial. I hope not through the rest of the talk. I hope I'm gonna be able to show you that we can take this someplace. All right, so one of the things that's come out of the psychophysics and molecules to perception thing, the chemistry and the psychophysics, are a variety of sort of paradoxical, I don't know what you call them, olfactory, I can't think of the word for this, but phenomenon, thank you. Okay, paradoxical olfactory phenomenon. I'm gonna go through a couple of them to suggest what some of the difficulties are. So here you have a molecule, this is well known, this molecule, a carbon, which exists as a chiral center and exists as an enantiomeric pair. That means the only difference between these two molecules, from a chemical perspective, is that they rotate polarized light in a different direction. In every other way, in every other physiochemical characteristic, they're precisely the same. And yet one of them smells like caraway or rye bread, rye seeds, and the other like spear lint. So it's hard to imagine what the logical basis of that distinction is, or precisely how it gets done. But it can't be any of the physical chemical properties of these molecules that I showed you before. Odor discrimination, we're quite good at making discriminations down to the atomic level. So here are two molecules, hexal and heptalacetate, that differ really only by a single carbon atom and the associated hydrogens that go with it. So a single atom difference between these two molecules, and yet one smells like banana and the other pair, a distinction that we would never have any trouble making whatsoever. So this extremely small difference between these two very simple molecules, nonetheless leads to a clear discrimination ability. On the other hand, you have these two molecules which don't look chemically at all like each other. Yes, they're found in both of those vegetables. So they can be extracted from those vegetables. And if I gave you a vial, there are other chemicals of course in both bananas and pears. But if I gave you a vial of either one of those, you would immediately recognize it as a banana or a pear odor. But they also do exist in those fruits. Yes, here are two molecules that are wildly different from a chemical perspective. And yet they give a very virtually identical smell of sandalwood, very difficult to discriminate, and they even cross adapt each other. So here's just the opposite situation where wildly different chemicals nonetheless give you virtually the same smell. Here's another case of that. This is Moscow well-known ingredient in the fragrance industry. Now the original must comes from the anal gland of the civet. So you can imagine there was a lot of work being done on trying to make a synthetic musk because you can't get that much musk out of the anal gland of the civet and it's not much fun to do, I imagine. So as a result of that, there's been a lot of chemical work based on trying to make synthetic versions of musk and they've been very successful. So all of these are synthetic versions of musk and they're indistinguishable to most people and I believe most animals from naturally occurring musk and yet you can see they're quite different structures. They're even, so many of them have cyclic qualities but this is, here's an acyclic musk. This nitro musk is actually a very close relative of TNT. Yes, yes, I don't know the answer to that, I'm sorry to say. I know the perfumers are very happy to use these synthetic versions. They don't care, the ones I know. Absolutely, in fact, I was going to put them together and ask you. Yes, please ask. I think there are many questions that come up and especially the students here are not going to be able to. Yeah, well they're asking, they're asking. I would disagree that these are readily discriminable, readily discriminable, but they would all be categorized as musk. Right, okay. So they'll all fall into that grouping. I think this is something that pervasive and thinking well-faction is that confusion between categorization and musk. Sorry, can you just open the box here? Yes. All the musk in there? They're all categorized as musk, I don't know if I'll go into that. Yeah, I agree. But I mean, that will make that a really good point that if you tell someone this is categorized, you can change it and put that thing in that box because they are readily discriminable. Okay. There are cases where a chemical, this is relatively rare. Yes. Do I know the chemical pathways by which they were generated? I, oh sorry, yeah. So the question was what's the chemical logic behind creating that? I can't tell you in this case, that's sort of what I'm going to get to towards the end of the talk. So I don't know whether those, whether these various musks were found by accident or in some cases I think there were simply attempts to recreate the original. In some cases it was entirely an accident. That's typically the way it goes in any case in this, yeah. Also, can you show that? Yes, that's right. That's right, so the definition of an odor as binding to a receptor is not animal independent because different animals have different sets of receptors. Although there's a fair amount of overlap, it's also true that that's right. Yes, so I'm saying that both of those things happen that there are molecules that look very similar or have similar structures but smell different and molecules that have very different structures but smell very similar. I mean it could just be a category thing, that's okay with me. The point is simply that it's often difficult to predict. It would be difficult to predict from any of these that these would be acceptable as a sort of a musk category. So I think it only has to go that far. Another case of this sort of stuff is well known although it doesn't occur as often as people tend to think is that the concentration will alter the quality of an odor. So actually most of the time, and I'm sure Dima will speak about this, one of the remarkable things about the system is that you can change the concentration of an odor over six or even eight orders of magnitude and not change its quality. Its intensity, of course, will change. Maybe it's pleasantness or unpleasantness but not its quality. But there are a few cases where changing the concentration will also make a significant change in the quality. Indole and scatol, two well known chemicals that have very low concentrations have a sort of a fruity grapefruit sweet smell but at higher concentrations smell like crap, so to speak, literally, in fact, smell like crap. And so there you have a change in quality with concentration which you don't have for every odor either. I said that mostly doesn't happen. No, most odors, if you give most odors over seven or eight orders of magnitude they don't change their quality. I mean there are a few rare cases that do it. It's often cited about that but in point of fact what the system does much like the visual system does will still see blue light as blue in over a wide range of intensities. I don't think it's due to learning. And you would at a high concentration as well. Well, you may but I think Dima will talk about, I don't wanna get into concentration if that's all right. I'm just using these as examples of why it's difficult to take a purely chemical approach to understanding odors. That there are too many things that don't simply work out in that scheme. No. Still speak. Yeah. So there are a number of things that interfere it seems to me or give a lie to this simple idea that if we could totally just do this biochemical analysis that we need somehow or another to take account of the receptors in order to get to a perceptual area. Okay, so once again what is an odor? I'm gonna suggest they bind to receptors. One of the things that we spent a lot of time doing over the years is classifying odors. I think we'll have a talk about this as well. This goes all the way back at least to Linnaeus, believe it or not, the great classifier and organizer who had a system for classifying odors. He believed there were seven primary odors. So the search for primary odors has gone on for many years beginning at least with Linnaeus may be going back to Lucretius and something like 50 BC or something like that. But I won't go that far with this. There have been numerous other schemes. I'm gonna show you a couple of them here. I have a feeling I have a double of this slide but all the way down to a piano scale, one of my favorite. So, yeah, sorry, this is the right version of that slide. Sorry. So there are all these different schemes that have come up. They go from six or seven primary odors all the way up to a couple of hundred primary odors. This is a continued search, although with the discovery of the receptors and this large family of odor receptors, there was a lot less interest for a long time in trying to understand whether there would be primary odors. The thought being, well, with this gigantic receptor family, we're not like vision where there are just three or four light receptors, but we have thousands. So why worry about primaries? I think we're now coming back around to the idea that primaries are likely to be important. So what does it mean that they bind odor receptors? Let's talk for a moment about the receptor. Do you have a question? No, you're just scratching your armpit, that's all right. Well, he was doing this. Sorry. How can you tell the difference, you know? So all right, so let's talk about the receptors for a moment. Let me give you a very, very quick background on the receptors for those of you, again, not in the field. They were first discovered, really, by Linda Buck, then working with Richard Axel and published this amazing paper, Seminole Paper and Cell in 1991, called the novel multi-gene family man-code odorant receptors, a molecular basis for odor recognition. The receptor that we're talking about is a G-protein coupled receptor. This is not a real picture of it, but rather a cartoon based on the typical structure of these receptors that have seven transmembrane alpha helices that snake back and forth in and out of the membrane and external and an internal region as well. I think most important to recognize is that these are class A GPCRs like modopsin, dopamine receptors, sartone receptors, and so forth. And it's important to recognize that the binding region for the ligand for these receptors, be it a drug or an odorant, is not out here in the extracellular region, but rather about a third to a half of the way into the membrane. So in the somewhat empathic pocket of both hydrophobic and charged areas. And a given odor molecule has to find its way down into that pocket. If you're familiar with rodoxin, you know that retinal sits in exactly that same place, for example. This is sort of the quick list of what you need to know about odor receptors. They're proteins of about 300 plus amino acids. These are all class A GPCRs, as I mentioned. As is typical of class A GPCRs, one gene encodes one protein. There's a single coding exon. I think you'll hear more about this later. So in other words, there's no alternative splicing like you find in ion channels and many other sorts of things. They contain conserved and variable regions. The variable regions tend to be in these transmembrane areas, which is interesting because that's what may give them the diversity. It enables them to detect many different kinds of molecules. And of course, they're the largest family of G-protein coupled receptors by an order or two of magnitude. They range anywhere from being about 480-some odd receptors in humans. I think it is. Is that the right number now? I never know what the number in humans is anymore. 380, it's gone down. It was 4-some things. 425. Anyway, it's, let's say, 400 in humans, up to nearly 4,000 in elephants. So mice, rodents, typically have about 1,000 or 1,200 functional receptors. Yes. Why do we need so many odor receptor types? Yep. Yeah. I'm actually the wrong person to ask that. You've got to ask. That's a higher level question to be asked. I don't know why. Well, it's hard to know that. And I think efficiency is a dangerous word in biology personally. So I know there's always this idea that evolution is cut fruit and ruthless and efficient, does everything in the most optimal way. But that's just not true. And so it may simply be that we can have this many receptors. There's very little penalty to it. I mean, so the next largest family of receptors are serotonin receptors. There are 15 of them. So as I say, it's an order of magnitude or more difference. The thing is, there are only so many ways you can make a serotonin receptor, and then it doesn't bind serotonin. So it's no good. And so that would be selected against one, would assume. But with odor receptors, if you make duplications and you make more of them, and they bind something out there, why not keep it? And so you can imagine this pure hand waving here. But you can imagine there'd be very little negative selection on expanding the family beyond the wiring that has to come in to play and all that, right? Right, so it may be a reflection. So again, it may be a reflection of the fact that with the wavelengths are on a continuum, whereas there's no continuum precisely here. It's a more discrete sort of stimulus. So it may require a different strategy. I'm sure many people will talk about that idea today. Importantly, they don't easily express heterologously. I should update this slide. So for a long time, we couldn't get them to express in other kinds of cells at all where you could screen them with various odorants and see what they responded to. But they're very difficult to express in other cell types besides olfactory neurons. Heromazinami has managed to get some level of expression. But still, the large majority of odor receptors are orphan receptors. That is, we do not know what their ligand is because we can't get them expressed easily. Finally, there's this notion, and it's an important one. There's some controversy about it a little bit. And that is that each olfactory sensory neuron picks one of these olfactory genes, these odor receptor genes, and expresses only that one. Not only actually it's a monogene, but it's mono-allelic. Now, there's some evidence that early on in development, they may express multiple receptors. But it appears that a mature olfactory neuron has chosen one allele of one receptor and all of the protein it makes comes off of that. What would you think? It's not controversial in a sense that they, oh, in the mature state, yes. But I meant just that early on, there's now some evidence that there's multiple receptors expressed. So this is important for what I'm going to tell you now because what it means is that even though we may not know which receptor is being expressed by a particular neuron, if that neuron is activated by an odor that we put on it, then we know that it's activated some particular receptor. And we can, to some extent, identify receptors then by just their receptive field, if you will, the particular odors they bind or do not find or are activated by or not activated by. So we rely a great deal in the work I'm going to show you now on this fact that they express only one receptor gene. So this is really the strategy, right? They express one gene. We dissociate these olfactory sensory neurons from the epithelium, typically of a mouse. We either use some of the early work. I'll show you a sephuro. Now we use GCaMP6, which is a gene that can be expressed only in olfactory, mature olfactory sensory neurons. And we'll fluoresce green when activated by increased calcium. You use some panel of odors to challenge them. We analyze it with now a variety of ways, but in particular cluster analysis. And we're able to identify some receptors just by their pharmacological profile. Again, we do not know what particular gene this receptor is, but for some things that doesn't matter. So a couple of quick ideas about what you can do. So you can look at the cells in a dish, put all these different odors on them, and then just look at the percentage of cells that respond. Now to some extent, this is a rough estimate of how many receptor types are able to bind that particular odor. So for example, you can go from hexosilicylate, in which we see only 0.2% of the cells in a dish that light up all the way down to gamma undecolactone where nearly 6% do. And yes. These are mice. These are mice. Everything I'm going to show, almost everything, except for one slide, is going to be mice from here on in. So you could make a rough estimate that 0.2% of the receptors, the 0.2% of the cells, responding to this reflects something like between two and five receptors that are capable of binding hexosilicylate and maybe other things as well. All the way down to here, where you can imagine that as many as 50 or 60 receptors could be sensitive to gamma undecolactone, perhaps at different concentrations. Yeah. That's right. So it makes an underlying assumption that each of the receptor types is more or less equivalently expressed, which we know is, by the way, not true. There are a few receptors that seem to be way overexpressed. But for many that have been looked at, it is approximately true. So yes, these are strictly approximate numbers. And I just mean to give you a quick idea of, yeah. That's right. That's right. So I think I just mentioned that, but that's exactly right. That of these 50 or 60 receptors that may recognize gamma undecolactone, there may be some that only recognize it at a fairly high concentration. This is a long time ago, but this experiment was done using mostly high concentrations, because we were interested in specificity. So if cells don't respond at a high concentration of odor, then they really don't respond to it. This is a long time ago, so I can't remember. No, I think probably I'm guessing 300 micromolar. That's generally what we consider a high concentration. All right, so one way to classify odors is indeed by using their chemical constituents, right? And that's kind of the standard method. We've used that. Many other people have used it. It's been used most recently. I think Leslie will talk a fair amount about this. Using things like functional group, chain length, saturation, steerer groups, et cetera, all the common sort of chemical things. Here's one example of an experiment. Using this idea and that somewhat to our surprise, actually, suggested a possible kind of rule about receptor binding. This experiment was done some years ago by Ricardo Araneda in the lab. So using three simple aliphatic eight carbon molecules, but with different functional groups, an aldehyde and alcohol and an acid, we tried this on a large number of cells. Again, just looking at the calcium imaging, this is a record of about 150 out of, I seem to remember it was nearly 2,000 cells overall that were looked at, but we selected this group. I'll show you why in a moment. So what this means is that each row here, which is hard to see, of course, means as a cell, and then each column, there are three columns really, is one of the different functional groups, molecules of different functional groups. So either 30 or 300 micromolar responsiveness. And so that means all of these cells responded to all three of the odors. So we saw a variety of patterns, of course. So we saw, for example, that some cells out of the group respond to all three of these odors. They do not discriminate between those functional groups. Some cells responded only to the alcohol and the aldehyde, but not at all to the acid. A somewhat smaller group in this sample at least responded to the aldehyde and the acid, but not the alcohol. The interesting thing was that looking at over nearly 2,000 cells, we never ever saw a cell that responded to the alcohol and the acid, but not the aldehyde. So you can respond to all three, these two, or these two, but not this one and that one. Now I haven't put them up there in any old order. They're up there in order of increasing electronegativity of this functional group. So this suggests that for whatever reason, we don't know the reason yet, but whatever reason, there's a sort of a receptor rule here that it's very difficult somehow to build a receptor that combined a group of a lower electronegativity and a higher one, but not the middle one, but not being able to skip the middle one. We see this in other areas too and I'll try and point them out later. All right, so that was sort of earlier work and now I'd like to talk about the remaining 10 minutes that I'll have to rush through all of this so we have some more time for questions at the end. How we can consider classifying orders by thinking more biologically, which is I think the not so standard method. And we use to do this technique called medicinal chemistry. This is borrowed from the pharmaceutical industry and medicinal chemistry is a kind of, it combines both chemistry and biology. I'm not suggesting that we throw chemistry out the window, just that we think about it in biological terms as well. So note, for example, that chemicals are all named by organic chemists and classified by organic chemists or physical chemists. Rightfully so, they've extracted them, discovered them, synthesized them, whatever it is. But their naming scheme, their nomenclature is of course directed towards chemistry, towards reaction sites, towards places where you can or cannot saturate or put double bonds in a molecule and so forth and so on. These may or may not be, of course, biologically relevant. Medicinal chemistry tries to consider the biological side of this as well. So it combines organic chemistry and biology, making minor alterations in chemical constituents that are based on biological function, not based necessarily just on the chemistry. And this leads to a term called bioisosterism. So isosterism being the chemical ID of molecules that are somehow or another sterically similar, but we're interested in a bioisosterism here. So this, in our case, defines an odor by its ability to activate a receptor or a set of receptors rather than its chemical characteristics. And then we can look back at the chemical characteristics and see which, if any, were important and in what way. Let me give you a very quick example of how this could work. So here's a very simple little experiment. This only involves data from 30 cells. There were more in the original population, but this is sort of the key of it. So we looked at three chemicals, acetophenone, sorry, acetophenone, cyclohexanine, and this is a ketone, but chemists call it a tiglet. It's a tiglet-like group, so it's easier to call it tig for tiglet. We're looking at 30 cells here, and this is a Venn diagram. The cells have responded to one or more of these three odors. So you can see that about 18 of them responded to only one of the three odors and would be considered, I suppose, narrowly tuned. But then if you look at this diagram here, you'll see that 35% shared a responsivity of those two, 23% there and as many as 50% responded to both of these. Four out of them, in fact, responded to all three. So the question here might be the way we would typically organize this is we'd say 18 of the 30 receptors, those on the outside here, that responded only to a single odor, we would consider as being narrowly tuned, at least as regards these three odors. And 12 of the 30 would somehow or another be less specific or broadly tuned, and those are the ones in there. Let me see if I can change your opinion about this. Here are the three odors. These are the structures of the three odors, and typically the thing I think your eye is drawn to immediately is the big steric ring. In this case, it's aromatic ring. This is not aromatic, but nonetheless, it's a large steric sort of structure. But let's look at this tiglet group here. Now we take that tiglet group as a template, we see that it fits quite well on these other two molecules as well. And so perhaps what's actually being recognized here is this so-called tiglet motif, because I don't have another name for it. And that this big aromatic ring or this ring here are there as much to maintain the steric position, the conformer of this backbone. It's this backbone that's really the critical thing and is most critical in that position. Because of course, it could rotate around many of those bonds and find itself in different conformers. But this forces the molecule to be presented in this particular conformer. And so we could say instead, thinking about this tiglet group, that actually all 30 of these 30 receptors are specific, they're specific to this tiglet group. And so you see whether the naming or the classifying of these things becomes difficult because it depends kind of to some extent on how you look at it or what you're looking at or what you think represents the diverse group of odors. So we originally chose those three odors because we thought they represented a chemically diverse group of odors. But what's chemically diverse may not always be biologically diverse. I think that's the point here. Yeah, okay. All right, so let me just point out that one interesting thing here is that looking at medicinal chemistry with the olfactory system is kind of more interesting in some ways than what the pharmaceutical industry does. Well, the pharmaceutical industry, of course, they use a single receptor and a bunch of ligands and they're looking around for some new drug or some slightly better drug or something like that. But we're using, in fact, these same principles, but we're using as many as 1,000 different receptors simultaneously to look at various ligands. And so we get a different, I think, an increased view and a larger view of how receptors can work in biology. So even beyond what they do in olfaction. I'll give you an example of two sets of experiments. And I'm gonna go through this sort of quickly. Maybe I'll skip through this very quickly and we'll do the second one more. There's one last thing I wanna show you. So here's a pseudo-phenone, that molecule that I talked about before. And what we did was we made a series of hetero-atomic, what are called hetero-atom substitutions, where we simply took a carbon atom out of here, a methyl group of the carbon atom out of here and changed it for either a nitrogen, a sulfur, or an oxygen, in some cases, two of them. So these are medicinal chemistry type changes and then we look to see who responds to what. So this is here, the five, the six chemicals up top. Here are cells down here. There were a total of 276 cells. These numbers here, the number of cells that responded in this pattern. The 28 of the cells only recognized the pseudo-phenone, didn't care about any of the other molecules. A few of them recognized all the molecules and then there are all these different patterns. I think we had 36 different response patterns. Okay. Yeah. I know it's getting late, but just to make it clear, this is the binary recipe, the binary way of presenting. This is one high concentration. Yes, relatively. This is not the highest concentration, but yes. These were done, I believe, at 100 micromolar. I'm sorry, I should know that off the top of my head, but I don't remember. I would say it's responding, you could have responded kind of weekly. Yes, we set a threshold, yes. So this is strictly a binary classification. Okay. So we can then build a tree out of this and you can build a tree using chemistry. You can use this program called E-Dragon, which is a program that chemists use. I think Leslie, I assume you're gonna talk about this later, right? So we use a slightly earlier version of it than what Leslie has used most recently. ours has about 1,666 chemical descriptors and you can feed these molecules into that program and it will shuffle them around and do whatever it does and come up with a tree of relatedness based on these 1,666 chemical features. And this is the way those six molecules come up on this tree. Now, all right, so that's doing that. Okay, now, here's the biological tree based on that set of responses that I just showed you and I think what you can see is that it's somewhat different. They don't necessarily overlap entirely. I mean, there's similarities. So for example, one and two, these two chemicals are by biology very closely related. Cells are responded to chemical one, also responded to chemical two. That's not true here. Chemical one and chemical two, although they happen to be close here, are actually totally separate branches of the E-Dragon derived tree. The same thing is true of five and six, which are very close in biology. Cells are responded to five, tended to respond to six, but sorry, it was not true in chemistry. So we wanted to extend this and we used another class of molecules called esters. Esters are an important molecule in the fragrance and flavor industry. They're used extensively. There's a large chemistry around esters. I don't want to confuse you with this slide. I just want to show you that the nice thing about esters is you can do a lot of things to them. You can, there's this thing called, this is the ether oxygen. You can move that up and down the train. You can move the carbonyl group back and forth. You can change them around in a whole lot of ways, very simply, and they're all available off the shelf for the most part. And so you can look at these minor changes in the molecule and see what the biological changes. So here we used, in fact, six of these molecules. We actually used five esters and a ketone as a sort of an outside chemical because it's related but doesn't have the ether oxygen. And these are, this is again the same kind of response thing. So 872 cells responding, 58 different response patterns using these slightly different ester molecules. So again, we come up with a tree and we have the classification by cells responses and the chemistry classification. And you can see once again, there are differences. Once again, one and two are very close together but here they're on completely separate branches. Most surprising to us was four. So as we expected, the chemistry classification classifies four, that's the ketone, which we included as an outside odorant, a non-esteroidorant. And so of course, not unexpected, the chemistry classifies it there. But unexpectedly to us, four was actually recognized by a large number of odor receptors and some of them were the same as recognized other odors as well. So esters and ketones were in this case recognized together, yeah. How do what? I mean, I don't know the inside to be dragon personally but that's, we used the dragon to do the chemistry classification. Yeah, so these are all Euclidean distances. They come out as Euclidean distances, yes. I think that's what Leslie's gonna talk about. Yes. Well, but that's, well, I'll leave it up to talk about that because they just published a big paper on this using E-Dragon and machine learning together to come up with a classification scheme. Oh, that's right. Okay, so which I would say does not fit precisely with what I'm showing. There's no question that we don't agree on this. Pardon me? Right, they don't use receptors, they do this with humans. But so I'll show you something about that too. So we have a mini controversy here of some sort. Well, you'll see in a minute. Just give me another couple minutes, okay? All right, all right. So you can see that there are several differences then between the way you get a classification scheme out of biological activity versus the chemical program, the way chemists presumably would classify these owners. So for example, we find that the relative position of the carbonyl group is what's important and not curiously the functional or the ester group and a surprising tolerance of what are called reverse esters where you've moved the ether, oxygen and the carbonyl group to opposite sides. And that's not commonly true in much of the rest of biology is a famous example of that in Parkinson's disease where drug users took MPTP, which is the reverse ester of maripidine, I think it's called, which is what gets you high and instead they got Parkinson's disease and that's the result of an enzyme not recognizing MPTP and all it is is the reverse ester of maripidine. So, but the olfactory system seems to be quite good at recognizing reverse esters. The chemistry classification seems to be primarily based on functional group and the size of the arms, the difference in the arm length depending on where the ether oxygen is, which did not seem to matter very much to the neurons. So, we thought, well, we should look at behavior as well because if this is what's going on in the neurons maybe it's completely different by the time it gets to quote perception, okay? So we used a very simple test with mice called the habituation, dishabituation test. You have dip a Q-tip in one of these odors, you put it in the cage of the mouse, the mouse will rear up and snip it, they're interested in it, you do it several times and eventually they become habituated to it and no longer care about it. They spend much less time, they sniff it once and they're gone. Then if you put a second odor in and they rear up and sniff it and spend some time at it, you can say, well, that second odor differs from the first. If they don't show any additional interest in it, they quote, believe it to be the same odor, they're habituated to it. So here are two examples here, so you can see this is odor one and two and here they initially spent a lot of time with odor one, habituated to it and they seem to think that odor two is pretty similar to odor one because there's no increased activity. Same thing over here for five and six and if you remember that was the way the classification scheme worked out in the biology. On the other hand, here for odor one and five, they seem to see them as different. So you have odor one, eventually they habituate to it, but if you put odor five in, they're not habituated to that at all. They see that as a new odor, same thing here. Yeah, when there is a, right, so of course I can't talk to the mice as you know, so I don't know the answer to that. I just know that, I mean, the test is commonly used in that way. No, no, no. No, so it's true. I mean, of course, if there's more at stake then discrimination may be possible. So I'm not saying that, and I don't think our tree suggests that odor one and odor two are identical. That's certainly not the case. There are neurons that see two that don't see one that see one and don't see two. Just that on average, you would classify them as closer than say five and six or so you would classify one and two as closer than one and five. And that's the way the tree classifies them by simple OSN activation, neuronal activation, and that it works out the same way in the behavior. That's all I can really say about that. I'd also say, so we also try this in humans. The only time we've messed around with humans and I'll never do this again. But it kind of worked out for us with humans. So in this case, we just simply gave people, there were two tests, we gave people seven vials that included five of the odors. The ketone is not in here because ketones are sort of dangerous. And so you can't have people sniffing ketones. So they included the five odors and at least one of the odors. In fact, it was odor one twice and a blank. And then people were asked to, about 15 or 20 subjects in a year. I think they're 20 subjects. People were asked to simply group them the way they felt they were most similar. Put the most similar odors together in a group. In a second, somewhat easier test, we just gave them three vials. One vial had one odor and the other vial had the same two odors in it. And so all we said was, can you tell us which of these vials is the same? Which two are the same and which are different? So importantly here, because Leslie brought this up earlier and she'll bring it up again, I'm sure, we are not asking people to make a discrimination in the sense of what do you think this smells like? We're not asking them to make any kind of a perceptual judgment about it. Does it smell the same or does it smell different? That's the only thing we're asking. We're not suggesting what these odors are to people. And the way humans classify them is so that sometimes they miss the identical one and think they're slightly different, but they typically put one and three together and five and six together. So we go back to this again, here's mouse and here's human. And you'll see there are differences and similarities. So the mouse puts one and two together, but humans do not. One and two are easily discriminable, it seems, by humans. On the other hand, mice put five and six together and so do humans. So we actually would expect this to tell you the truth because mice and humans have different receptors. So we still think this is based on receptor because the dragon sees it differently as it were. The dragon always never puts one and two quite together and six and five always have three in the middle which is not the case for either the mouse or the human. Four doesn't count for human. So we see differences between mice and humans which we would expect. We also see some similarities, but both of them are different from the pure chemical classification scheme. All right, I only have a couple of seconds left here. So I'm gonna, I told you all this, I don't have to go over the conclusions. I just wanna show you a little bit of very recent work is unpublished and I think it's extremely important since we're talking about coding here. This is an old paper, a relatively old paper by David Lang and Lang and Francis, 1989 or 83 or something like that, talking about odors that are presented together. Sorry, you probably can't read this, but the important thing here is that some odors can reduce the intensities of others to a level where the suppressed component cannot be perceived. Does it always happen or does happen sometimes? The point here is that they had no idea what the mechanism would be and so we wanted to explore this because after all, everything I've told you about so far has been done of what we call monomolecular stimuli. A single odor is presented and then another odor is presented and another one and another one, but one at a time. But that's not the way we smell the world nor does any other animal smell the world. They generally smell it like the way you smell the rose with a group of odors. So we wanted to look at blends and we adopted a technique that comes out of the laboratory of Elizabeth Hillman, a biomedical engineer, well-versed in microscopy. Her graduate student, Wednesday Lee, who's also fortunately the husband of a graduate student in my lab, Lu Shu. So, and they did a series of experiments using this new kind of microscopy called SCAPE. You have to have an acronym, you know. So, swept confocally aligned planar excitation. This is a kind of a variation on light sheet microscopy but it has two important differences or several important differences but the main important difference is that it used a rapidly moving mirror to scan a piece of living tissue. And so you don't, nothing else moves. The tissue doesn't have to be moved around as you do with light sheet often or cameras and objectives. It's a single objective. You get three-dimensional volume. You can see right through the volume of a living piece of tissue. It has very high speed. In our case, we're looking at about five volumes per second. It can go up over 20 volumes per second. And importantly, you can get single cell resolution as well as whole tissue reservoir, a significant chunk of tissue resolution. This is our typical, we have a chamber. We crack open a mouse head basically, put it in there. The turbinates are exposed and we're looking at an area of the mostly turbinate two prime and three for most of the information we see. So this is what it looks like basically. Here's the dimensions, the three dimensions of it. I'm gonna show you, I hope, a quick movie of this. Let's see if it works. You'll see in a second we put the stimulus on. You can see cells light up. Of course, it looks much better on the computer than there. But you can see that you can actually get, so first of all, we're looking at a relatively wide swath of epithelium, but at the same time we have single, so if you look at each of the optical slices, we have single cell resolution as well. So we have both the possibility of doing what we've been doing for years on dissociated cells or with patch clamping cells, but we can also do what people did for years with the EOG or field recording, which is look at a large response of many cells, even though you don't know which cells it was. So here we have both of those things. So we did two simple blends of three odors. I'll tell you essentially about this one, including citriolacetophenone and benzylacetate. There was a second odor set. These odors, I'm not gonna have time to talk about that. Here's the tissue. I'm just gonna show you here in three slices. So these are the dimensions of the tissue, but we're gonna show you, I'm gonna show you now what it looks like sort of in three slices. Oops, I hope I didn't mess this up. Okay, right. So here, I'm sorry. So here are three slices at three different Z levels that do not overlap. And each one of these lights is of course, a cell that reacted to either citriolacetophenone or benzyldehyde. And one can then just simply go ahead and count those cells. Or you can look at the cells that were activated by the entire blend. Here's the way a typical experiment would go, all right? So the MSO, which give a response at all. Here's the blend. Here's all three odors together. There's the blend. And then we give the odors individually. Now there's some rundown, so, but curiously it's virtually always a linear rundown. So we can pretty easily take that into account. Then, so here's the blend. Here's all three together. And interestingly in this case, citriol, sorry, citriol gives virtually no response alone, but acetophenone gives quite a much larger response than you get to the blend. And benzyldehyde gives an intermediate response. Then we try them in pairwise fashions and so forth. What I'd like to suggest to you as to what's going on is that the response to the blend is often, not always, but is often smaller than the response to one or more of the individual components. Suggesting that one or more of the individual components is actually acting as an antagonist. That it's suppressing the response to one of the other odors. All right, here's sort of the bigger picture of the diagram. There's a whole lot more data than this, but I'm just gonna show you because for simplification's sake, cells that responded most specifically, most dominantly to either citriol, or most dominantly to acetophenone, or most dominantly to benzyldehyde. So the interesting cells are these. These cells responded very well to citriol and much less to the blend. These cells similarly responded very well to acetophenone, but much less to the blend. No, there are plenty of cells that responded to both or the other way, the blend was higher than the individual component. We're interested in these cells here. Same thing for benzyldehyde where there's a group of cells responded very well to benzyldehyde, but not at all or very poorly to the blend. So in each of these cases, this suggests some sort of suppression or inhibition. And indeed these are sort of the cell counts. So what you can see is that for the most part, there's no effect. This one would kind of hope to maybe see that the blend is a reciprocal combinatorial code of the odor of cells that were activated. But in fact, we see a fair amount of suppression as much as 15 or 20%, 15 to 20% suppression in many cases. So, I mean, this is a surprise. We thought we'd see some suppression, but 15 to 20% is a rather significant amount of suppression, yeah. Of the cells that we see respond with less of a response to the blend than to any one or more of the individual odors. Suggesting that one or more of the individual odors is actually acting as an antagonist, yeah. Yes. The same concentration of each of the odor. The same concentration of each of the odor. Yes, yes, yes, yes, so you mean who cares? So can I, yes, so I was about to say that we and others have showed antagonism, including Tuhara's lab and several other people have showed. And it's not surprising. These are G protein coupled receptors. Half the drugs on the market are antagonists of GPCRs, whether they're beta blockers or things like that. So it's not surprising that GPCRs, which odor receptors are class, should have antagonists. What surprises us is that in a simple blend of only three odors, we see as high a level of suppression as 10 or 20%. Imagine a much larger blend of odors. Imagine a perfume that has 80, 90 or 100 components in it and how many of them are acting both as agonists and antagonists. There's also some enhancement effects that we really cannot explain at all. And so I'm just gonna sweep those under the rug for the moment. The same thing was true of the other odor blend. I'm not gonna go through the data on this, but we saw, again, virtually the same thing. So of course, one question, sorry, that's the wrong side, this is the slide. One question is, how's this occurring? And we wanted to know for sure that it really is antagonism, some sort of competitive antagonism at the receptor level. So these are a whole slew of cells. I didn't have time to put all the dose response curve together, data together, but they're essentially dose response curves. So we're using citrall cells that do not respond to citrall and then do respond to acetylphenone at 10, 30, 100 or 300 microliters, micro molar, and then with 100 micro molar citrall in the mixture. So initially we see a whole variety of things. So here's an example of something that looks very much like competitive antagonism. If you did the dose response curve, that's what it would show would be a rightward shift. So you have no response to citrall, you have a good response to acetylphenone, it's at a low concentration, it's blocked by citrall, but as you increase the concentration of acetylphenone, the citrall block is overcome. And so that's a kind of a classic case of competitive antagonism. I just want to show one kind of interesting and unexpected bit of data, which is this cell here. So this is a curious cell because it shows a block by citrall of acetylphenone, but no matter how high a concentration of acetylphenone you put in, the citrall block remains. So it can't be overcome. Now I have to say this is what you would expect if you saw an allosteric effect, that is the citrall is actually binding, not in the binding pocket, but somewhere else, and it's not competitive with acetylphenone. That would be, that's a big claim I have to say because there's no case that we know of of a Class A GPCR that has an allosteric binding site. And this has been a holy grail of the pharmaceutical industry for years to try and find one because it would give you some specificity. People have been searching for years. Now let me be clear. Let me be clear. Class A GPCRs have no identified allosteric molecules, small molecules. There are larger molecules, so if you talk to some people they believe the cholesterol can act allosterically on GPCRs, but that's in the lipid. But small molecule allosteric interactors, if you will, have not been discovered for Class A GPCRs. Class B and Class C unquestionably there are. Many of them are drugs. Tim, I'm sorry. I agree, this is just something I'm, as I say, this is an extreme claim when we want to look into further. But one could imagine that with such a large number of receptors, you do finally have some sort of an allosteric interaction that one could look at. Well, it's not activated, yes. Understand it's not activated by citral at all. If you present these cells all, if you give them citral, they don't respond to it at all. Okay, so citral is only being a blocker here, and then normally as you increase the aceto-phenone, you would expect it to be able to bump out the citral. So this cell does not respond to citral at all. That's a blank response there. I have to stop soon. And then I express it really well. All right, because yeah, sorry. All right, so this is the end actually, summary. So I have two quick points to make here really. One is that odor character I believe is receptor dependent and that it's the combination of chemistry and biology that will give us a much clearer idea of how we should classify odors and how we should think about them and how they become an odor perceptually. In the same sense that I would say a molecule, an odor molecule doesn't contain a molecule as an odor property anymore than a photon or a particular wavelength contains blueness or redness or greenness, right? It's the way, it's the receptor that we have that makes them blue, red or green and then how we process that. And then this idea of agonist and antagonist. So I think the crucial thing here is that we have for many years believed, I think, implicitly that if we could just express odor receptors or look at all of them, we could put a big matrix up here on the wall. With the odor receptors here and all the different chemicals there, we could just check the boxes off and say, okay, this receptor responds to these five things, this receptor to some overlapping group of them but also different ones. And then eventually this would be, quote, the odor code. This would be the olfactory code. I think that's no longer a reasonable sort of idea because we have to consider at an important level the notion of antagonism as well. I think it's a great puzzle and I put this puzzle out for the rest of the coding of people and the higher brain center people here. How does the brain know that an odor is both an antagonist and an agonist? I know of no other case in a sensory system where at the stimulus level, you can find an interaction in the stimulus that does that. So blue light does not inhibit green cones. Just leaves them alone. But here you have a situation where receptors are actually being altered, their activity is being altered by the particular set of other molecules in there. And so how does a blend become a perception? Okay, I think that's it. I have a fancy little graphic from Magritte that I've altered in some way with the help of, sorry. This is it, altered with the help of Bolak Zepicek where we can take a molecule and turn, yeah, you get it. All right, and of course, the people that helped me do this, all right. Thank you.