2003 LECTURE SERIES

How Time Flies: The Molecular Architecture of Memory

Dr. Thomas J. Carew
Bren Professor and Chair, Department of Neurobiology and Behavior
University of California, Irvine

First of all, Jim, thank you for that lovely introduction, it was beautiful. I appreciate it. And let me also tell you at the outset how delighted I am to have the chance to speak to you tonight. I have come to this series now for many years, and it is a privilege to have a chance now to contribute to it.

So the series has a theme. The deep theme of the series has to do with learning memory and the brain, and I guess the first thing to sort of think about is why is that so interesting. What is it about the brain, what is it about memory that’s so fascinating? Perhaps put into more sort of a local and focused way, why are you here tonight? And I can tell why the seasoned veterans are here tonight, they’re here for the cookies. But I can also tell you that they’re here and have been loyal to this series for many years because they’re interested in learning and memory as well.

And I think one of the reasons we’re fascinated by memory is that we all use it every day, every moment of our day, to get to work, to get home, to do what we do professionally, to do what we do with our families and our friends. It’s such an integral part of our life that we probably take it for granted, but we’d be paralyzed without our memories at any moment in our day. So, at a personal level, we can appreciate it, but I think there’s more to memory than just the fact that we use a lot of it. I think in a real way, memories define us. We are, in a real way, the sum total of our memories. It’s as if an experience is a thread that weaves one more thread into a tapestry and that memory and memory and memory addition to that tapestry over time gives it texture and color and depth, and no one thread is visible anymore, but yet that tapestry captures the experiences of our lives.

And so in a real way, we are our memories, and I think this can be best appreciated by considering what happens if we don’t have memory. Now, there’s the normal foibles we all have of forgetting where we put our car in the parking lot, or forgetting the keys, that kind of stuff, forgetting to pay a bill, but--that was an aside--but those are the normal stuff of memory. What I’m talking about when I talk about the deep theme that captures ourselves is more the tragedy that accompanies severe loss of memory, where you don’t recognize your wife of 30 years anymore, you don’t recognize your children, you don’t know what you do for a living, you don’t know where you live anymore.

Now we’re talking not about foibles of memory, now we’re talking about deep pathology, where you no longer are the person who you used to be. And we’re talking about Alzheimer’s disease, we’re talking about stroke, we’re talking about accidents that can happen to compromise memory. And in these cases, it’s not that a few threads have come out of the tapestry, the whole thing’s coming unraveled. And so one can actually lose who they are by virtue of losing their memory, so it’s no small matter, this stuff we talk about in learning and memory. From a clinical point of view, it’s profound.

From a scientific point of view, what Dr. McGaugh and many folks in this room, Dr. Weinberger, and many of us, as well as scientists all over the world do, is study memory for another reason, and that’s because it’s just an amazingly interesting challenge. It’s one of the great frontiers in modern neuroscience, to understand the brain and brain mechanisms of learning.

To try to tell you what we’re talking about, let me just ask you to do something with me for a minute. Let me ask you all to close your eyes for a second. You all have your eyes closed? Look around, is everybody’s eyes closed? At least you’re paying attention. Now, I want you to imagine you’ve gotten home from the talk tonight, you’ve gotten up to your front door, you put your key in the front door’s lock, you open the door, you flip on the lights, and you walk to the kitchen. And as you walk to the kitchen, you probably have to make a right or a left or go through a hall, depending, and as you go into your kitchen, I want you to look on the walls. Look at the art, look at pictures of your family, your loved ones, your pets, whatever you have on your walls. And now you get to the kitchen, you flip the lights on. Okay, you can open your eyes.

What you’ve just done is extraordinary, because what you’ve just done is to walk through your house by virtue of only one thing, it’s your memory. The only thing that enables you to do that is your memory, and that was effortless, it was easy to do, and yet, if I asked you how you did it, you wouldn’t know. If you asked me how you did it, I wouldn’t know. Nobody knows. And it’s that challenge in neuroscience that we’re confronted with daily and are fascinated by, is just how is it that we can store information in our heads and get out that information at some later time.

And when we talked about memory, just as another brief piece of introduction, it’s not just one thing. There are lots of different things that go into learning and memory. There’s different content in memory, there’s different time courses in memory, so I’ll give you an example. I’m gonna give you a fact, all right? And the fact is that my sister-in-law’s name is Bernadette, all right. Now, I’m gonna ask you, what is my sister-in-law’s name?

[audience]: Bernadette.

Half these Barclay folks, you know, you just can’t slip anything by them. All right. So that memory, that memory was about 10 seconds old. Now, if I happen to cross you a year from now at another one of this series, and you happen to remember, I say, do you remember my sister-in-law? They say yeah, Bernadette, it’ll feel exactly the same to you, it’s the same fact, and yet it’s profoundly different. Because the mechanisms in your brain that allowed you to have that 15-second memory, that involves things like modification and phosphorylation. Now they’re starting to bolt for the cookies already. Don’t worry, it’s gonna be okay. I’ll get you there. But if a year from now I ask you that and you have that memory, there is simply no question that your genes have been activated to synthesize proteins and change neurons in your brain to endow you with that memory in a rather permanent store.

So that duration of memory is terribly interesting to me. And it’s not just in terms of facts. Let’s talk about a motor habit. Do any of you drive a stick shift? How many people drive a stick shift? Lots of people. You ever gone to an airport and rented a car? Of course, they’re automatic. So you get in the car, you turn on the car, you start pawing in the air with that left foot? So what’s up with that, all right? That’s a motor memory, that’s a memory that’s deeply ingrained. And it doesn’t just happen any old time, it happens when you sit in a car and turn on the ignition, because it’s surrounded by a constellation of stimuli that evoke that memory. You don’t do that at home with your pawing, or you might, but that’s where your dinner guests leave early, let me tell you, if you can shake the habit it’s probably better for you. But when you do it in the car, it’s because you’re looking for that phantom clutch. That’s a motor memory that’s deeply ingrained over years of practice and the like.

So memories can differ in content, and they can differ in their time course, and what I wanna talk to you about tonight is something that my laboratory’s been involved in for some time, and that’s the temporal domain of memory. We’re real interested in why it is that our brains sometimes allocate memories to something that’s gonna last a few seconds, and others in the other continuum last a lifetime. Why is it when you call information and get a phone number and you start to dial that number, God forbid it’s busy, ‘cause you’re already starting to throw that number away. You use it and you lose it.

But on the other end, there are memories that are virtually permanent, the memory of the birth of a son, or the memory of a wonderful thing in your family, or the memory of a tragedy such as 9/11. These are memories that are gonna endure, and we’re real interested in how the brain makes that choice, what kinds of mechanisms are recruited in the service of having a memory last a few minutes, a few days, a few weeks, a few years, or permanently.

So, to set the stage, this is sort of what we’re interested in talking about tonight, and that is looking at, asking the question, in the temporal domain of memory, why an experience gives rise to these different kinds of temporal domains, and I will tell you that this is absolutely arbitrary. There’s lots and lots of other domains we might talk about, but this is what I’m gonna focus on tonight, and in fact, I’m gonna tell you two stories. I’m gonna tell you about this intermediate term memory and long-term memory, and I’m not just gonna tell you a story, I’m gonna ask you to join me in experiments investigating them.

Rather than just give you charts or graphs and pictures and say okay, that’s how it works, I’d rather bring you along and do the experiments with you. So I’d like to introduce you to the kinds of approaches we use in the laboratory, the kinds of hypotheses we test, how we go about the questions we raise, the answers we get, and we’re gonna do that together in a minute. But in order to do that, in order to do those experiments together, the first thing we have to do is have a crash course in neurobiology.

So for the first thing, we need a little background. Let me get you to know your brain. So let me show you what your brain looks like. This is what your brain looks like. It’s thrown into all these folds called the cortex. The wonderful thing about the brain you probably didn’t know is it’s color-coded. That’s for neurosurgeons, just to make sure that they don’t screw up and take out the wrong part. So Lynn Stanton does not need this color-coded, he knows the occipital lobe is back here, but a neurosurgeon, it’s good to just not be too careful. All right, so this brain contains something like 10 to the 13th neurons, that’s about 10 trillion neurons. A boatload of neurons, lots of neurons, and those neurons conspire to make that brain an organized machine 24/7. It runs the railroad all the time. We’re interested in how this brain of yours and mine is involved in memory.

So let me just tell you a little bit about the brain. To use a metaphor of a computer, the brain has hardware. And the hardware of the brain can be seen as the elementary computational unit, which is a neuron, and I’ll show you what that is. The neurons connect to one another with something called a synapse, that’s where one neuron talks to another. Those synapses can form little circuits, and they, in turn, can form large systems. That’s the hardware, that’s your computer’s hardware.
So let me just first introduce you to the neuron, the elementary building block of the brain. So this is a nerve cell. The blue part in the middle is the nucleus, that’s where the genes are, your DNA. This is called the soma, or cell body. And the two parts you need to know about are the input parts called dendrites, these little spiny things coming off the top, and an axon that courses away, and that’s the information-carrying unit that’s talking at long-distance to other cells. So that’s the computational unit of the brain.

Now, I mentioned that these units can be put together into circuits, so here’s a few neurons, and there’s one in the foreground. I call your attention to this guy in the upper right-hand corner, and down here. This fellow in the upper right-hand corner is talking to the one in the foreground by means of this thing called a synapse. It’s got a long axon that has a myelin sheath on it, it’s sort of insulation, Saran Wrap around it. An electrical signal leaves this neuron, going at about 120 meters per second, so it’s really coursing quickly, gets to the end, and causes the release of neurotransmitters, these chemicals that are stored in little baggies at the end of the neuron. By that means, this neuron has communicated with that neuron, either to turn it on, to excite it, or to inhibit it, either way. They do both.

There’s another input to the same neuron here on this little spiny part of the dendrite, so there’s at least two synapses that are illustrated in this little circuit. But now, I mentioned that these circuits can be thrown into systems. A system is indicated here just diagrammatically. This is a diagrammatic representation of the visual system of the cortex in a primate. And each one of these boxes represents not 10,000, not 100,000, but millions of neurons. So there’s tens of millions of neurons illustrated in this little diagram, and it’s simply to illustrate how systems can interconnect and be richly interwoven.

So that’s the hardware of your brain. But in addition to the hardware, of interest to us tonight is the software of the brain. The software of the brain is inside each neuron, and that has to do with their genes, their proteins, and so-called signaling cascades, and we’re gonna talk a lot about those tonight. Things called kinases, that are the signaling devices in these cascades, and I’ll introduce you to that in a moment. And this software in the brain is inside every cell in your body, but in the neurons of your brain, they’re the ones that are gonna moment-to-moment adjust that neuron in the face of incoming information.

So let’s take a look at the software of the brain. First I wanna introduce you to genes and proteins. The genes reside in this neuron right here in the center called the nucleus. It sorta looks like a T-ball because there are pores in the nucleus that allow traffic in and out of it. In addition, I wanna call your attention, you probably can’t see it too well, but there’s little beaded structures called ribosomes out here, outside of the nucleus in what’s called the cytosol. And it’s that location that’s gonna be critical for the software of the neuron, inside this neuron, and we’re gonna talk about how genes make proteins, and there’s two steps. A gene synthesizes something called messenger RNA. Messenger RNA, in turn, is translated into protein. Those two steps are called transcription and translation. Let me just show that to you and come back to it.

So this is a picture inside the nucleus of a cell, and this is DNA. Your DNA is very much like a ladder that’s gotten twisted. Imagine a ladder that’s sort of made out of rubber, so you take a ladder at home, but when you get up on your roof, you twist it. On the outside of the ladder are sugars and phosphates, that’s the backbone. The inside, the rungs of the ladder, are actually called bases. There’s two bases that join by a hydrogen bond, so the rungs of the ladder are base pairs, and they’re designed to match up with one another. There’s a very strict code, certain base pairs like other base pairs. Now, when the gene’s gonna be transcribed, when you wanna turn it on, you unwind the gene a little bit, this DNA, and so now, imagine you’ve taken, essentially, an enzyme that acts like a little buzz saw and cuts right down the middle of that ladder, such that the base pairs are now unrequited, they don’t have their match.

What happens is, with these bases that now don’t have their partners, along comes another enzyme that synthesizes something called messenger RNA, and that aligns right along those same ladder rungs, such that they recapitulate the code perfectly. That’s called a transcript. So the transcript of the gene now leaves the nucleus, this red ribbon, messenger RNA--other things happen I’m leaving out--and it comes to this little machine out in the cell, the cytosol, this is called a ribosome. And at the ribosome, this ribbon gets run through the ribosome a little bit.

I’ve used an analogy with students that wasn’t worth a darn. Some of my colleagues, my older colleagues in this room will appreciate. Remember the old dryers, when you used to wring out something, you’d just put it through the wringer and you turn it and it goes through, like you’d wring out a T-shirt? Young students don’t know about those anymore, but we do, the old folks do. In any event, that’s sort of how this works, this ribbon goes through the equivalent of a molecular wringer, and what happens is that every three base pairs form something called a codon, and that’s a docking site for a small ferry boat called transfer RNA, that’s gonna come and dock right on that site, and it brings with it an amino acid. Amino acids are the backbones of proteins.

You know some amino acids, you know glutamate. You probably know tryptophan. You know glycine. So there’s lots of amino acids. And what happens is they get stacked on top of each other, a buddy, each codon gets one little amino acid, and it gets connected, and you spin off a long chain of amino acids, and that’s a protein. And the best way to think of a protein is to think of a strand of pearls. If you hold a strand of pearls in the air, that’s not what a protein looks like. We would call that a denatured protein, that’s not a good protein. If you drop that strand of pearls into your hand such that the amino acids interact, that’s what a protein’s like. It folds on itself, there’s lots of three-dimensional structure, terribly important for the operation.

So this is called transcription, this is called translation, so we’ve just turned the gene, it synthesized message, we transcribed it, and now that message is turned into protein by something called translation. Why am I telling you this, you’re asking yourself. It’s because this software is what we’re gonna change when we teach you something. This software is what is gonna allow you to remember seeing and hearing me tonight, hearing Dr. McGaugh tonight, and so forth.
Now, we’ve got a protein, but there’s one other thing that happens. That’s that long word called phosphorylation. All that means is you take a phosphate group and stick it on the protein. And that dramatically changes the way that protein works.

Well, I knew I was giving a talk at the Barclay, I knew that it would be an important group of folks I wanted to talk to. When Dr. McGaugh or Dr. Weinberger or I go out and talk to people at Stanford or Yale or Harvard, they’re just a bunch of science geeks, that’s not a big deal. But this mattered to me tonight, I wanted to actually make a hit with you, so I asked my lab six months ago to try to sequence a protein in toto, and not only sequence the protein, but to also capture that in the state of phosphorylation. So they used reverse high voltage x-ray crystallography, and the protein they sequenced is shown here.

This is the protein. It is a complex 97-amino acid protein, and it’s a globular protein, and as you notice, there are these green parts to the protein. That’s where the phosphorylation event’s gonna occur. This protein we call choclacin, we’ve cloned it, the protein melts in your mouth, not in your hand. And now, this is the protein and now what we found, which was to our delight, is those proteins that stick the phosphates on, their names are kinases, it turns out they’re red. Thank heavens. So there’s a red protein called the kinase, and it sticks these phosphate molecules onto the sites of phosphorylation of the protein.

So this is called a kinase, that’s called a kinase, those are the proteins involved in sticking the phosphate molecule on this protein, and when that happens, now it’s phosphorylated, and what happens when a protein’s phosphorylated is it changes its charge, and it can change its shape. So that a protein that used to look like this, now is a channel, to allow ions to flow through it, for example. So phosphorylation can dramatically alter the state of a protein.

Now, we don’t leave it like that forever. There’s another series of proteins around called phosphatase, it turns out they’re blue in [inaudible], and those phosphatases take off the phosphate, and the protein goes back to its normal status. And this is the state of events that a protein can go into the phosphorylated state, back to the dephosphorylated state again, and that happens all the time, all the time in your brain. That’s what happened when we got that 15 seconds of Bernadette out of your heads, something like that.

Okay. So now we can go and talk about the three parts of your software. We have transcription, translation, and phosphorylation. Why are we interested in that? It’s because that software is gonna be the core of your memory, and the core hypothesis for the evening is as follows. Learning reprograms the software of your neurons, and memory is maintained by this reprogrammed software.

So what we’re gonna do tonight, in a couple of experiments, it won’t last too long, is we’re gonna ask how this reprogramming occurs. Now, that brings us to the question of software. We’re gonna look at genes, we’re gonna look at proteins, we’re gonna look at these so-called signaling cascades. How do we study them? How do we study genes, proteins, and signaling cascades? I’ll show you. You take them, and now you have very specific drugs that can block transcription, or block translation, or block phosphorylation, so that we can interrogate the nervous system any time we want. Is that particular part of your software active in the instantiation of a memory?

So what do we wanna know? We wanna know about these different learning experiences, and we wanna know what makes them last so long, and I’m gonna focus on intermediate and long-term. Now, when you think about studying memory, immediately, the model organism that pops into your mind is gonna be this, the marine mollusk Aplysia California. This little critter lives about a mile from here, off Crystal Cove. You can go see him at mid to low tide. They’re a marine mollusk. This is the front, this is the back, top to bottom, this is what it tastes and smells the seawater with, and these little flaps on its back are called peripodia. Let me show you a little better view of the animal in its natural environment. This is the critter, again, this is the front of the animal, these little rabbit ears are what it tastes the water with. It’s got a dinky little eye there, it’s not good for much, but it’s very good at smelling and tasting, and it’s very good a feeling the water, water currents. And this little funnel on the back, this exhalant funnel, is attached to a breathing apparatus called a gill, that’s hidden by these flaps, and I’ll draw that for you in a minute, but this is what Aplysia looks like when it’s out in the wild. There’s another guy hiding from view.

Now, when we were at Yale, this is what they looked like. Once we got here, about three years ago, and brought them into the laboratory, occasionally we got one that we simply didn’t understand. They’re a little more relaxed, they seem more phlegmatic, their reflexes were a little more mellow, but the question that I have to explain to you is if we even consider this some sort of low and slow isoform or this animal, how can I look you in the eye and tell you we’re gonna study learning and memory from such a humble animal? I’ll tell you why I can do that. I can do that because of something called phenotypes.

You know that animals are shaped differently. So a goat doesn’t look like an elephant, doesn’t look like a frog, doesn’t look like your daughter, right? Better not. So there’s no question that animals are shaped differently. Birds are different than Aplysia. But when you get in their brains and, more importantly, when you get inside the brain to the software, the molecules that are involved in this kind of animal, they’re the same. Now, you may be actually offended to know that you share a large part of your genome with Drosophila, that’s that critter you swat on your peaches. Well, you’re swatting a distant cousin when you do that. Next time, you think about that before you do that, because they have a lot of the same genes you have, and they have a lot of the same molecular machinery you have to make memory happen. So that’s why we can study Aplysia, because they have this molecular machinery, and they have a real big advantage. We don’t just love the animal for its body, we love it for its mind.

This is the mind of Aplysia. It’s a few neurons, unlike us that have 10 to the 13th, they have, like, I don’t know, 20,000, right? If we’re counting? And the other thing about them is that many of their neurons are extraordinarily large. This guy over here has the largest cell body in the animal kingdom. It’s about a millimeter across, you can see it with your naked eye. Many of these others are very large and identifiable as unique individuals. That’s R2, R15, that guy’s name is L7, left upper quadrant cells, shows you what I do with my spare time, since I know the names of all these neurons, right? But nonetheless, this is the brain of Aplysia, it’s extremely tractable for the kinds of molecular studies that we wanna do.

So here’s what we study, we study a very simple reflex. This is what the animal looks like, and that gill I told you about is now diagrammed inside the animal, and that’s the animal in the relaxed state. Now, if we just touch the animal on the tail, or anywhere, touch its siphon with a little water jet, it shows a withdrawal reflex. And that simple little reflex actually shows a variety of different forms of memory, and the kind of memory I’m gonna tell you about tonight is called memory for sensitization. What that means is that normally, this reflex might last, I don’t know, 10 seconds, 15 seconds, it’s a very mild stimulus. But if we shock the animal’s tail once, that’ll increase the animal’s awareness. It’s like bringing up the gain in that reflex, it’s like arousing the animal. If we do that one time, that reflex, instead of lasting, you know, 10 seconds, it might last 50, 30, or to 60 seconds, and that memory might last about, I don’t know, 15, 20 minutes. But if we do that over and over, we can produce a memory for that event that doesn’t last minutes or hours, it can last days and weeks.

And it’s a little like you remembering getting aroused by, let’s say you’re walking through a parking lot at night and a car backfires and goes off and it sort of arouses you, and now a door opens, you might startle to that when you normally wouldn’t because, and the memory for that can endure, and especially if it’s a very startling event, so it’s called sensitization.

And what we can do, in studying this reflex and memory for sensitization is actually take out--Aplysia are like Legos, you sorta pop off the parts that you want, so we take off the tail, we take off the mantle, so this is the input, we’ve touched the tail to trigger the reflex, and this is what we measure, that’s the output, but we can take the nervous system and put it in a dish and study the neurons while the animal’s actually learning, either in the intact animal, or while this so-called reduced preparation allows us to still produce memory and study what’s happening in the neurons. And when you do that, you can record from neurons, like this sensory neuron that brings information in, or the motor neuron that brings information back out again, and study the way the sensory neuron talks to the motor neuron.

And here’s an example of that where we’ve injected current in the sensory neuron, and it talks to the motor neuron, releases a chemical, and that little bump right there is called a synaptic potential. Remember when I told you that the cells connected by means of synapses? That is the functional reflection where we’ve turned on this, that’s that action potential, that electrical signal gives rise to the release of transmitter.

Now we shock the tail to produce sensitization. And what happens is there’s more action potentials, and the volume goes way up. This cell’s talking a lot louder than that, so look at how much bigger that is than that, so that’s at the cellular, synaptic level, or a reflection that the animal’s aroused. And what happens during tail shock is the release of a hormone in the animal, a biogenic amine called serotonin. You’ve heard of serotonin, you’ve heard of Prozac, right? So serotonin is in your brains and Aplysia’s brains as well. Serotonin surrounds the sensory neurons, surrounds the motor neurons, they’re in red and green ‘cause we’ve injected dye in them, and the blue is an antibody against serotonin, and it richly invests the cell bodies, richly invests the synapse, so serotonin’s all over the place, and it’s released, the shading underneath this figure indicates that it’s released while the animal’s learning.

So we can measure the release of serotonin while the animal’s being taught something. And here’s what happens. If we put one pulse of serotonin on or shock the tail once, either one, we get a big increase, here’s the normal way the sensory neurons talk to the motor neuron is dramatically enhanced, and that stays up, and goes down in about 15 minutes. But if we do that five times in a right-of-way, spaced, then the synaptic potential stays enhanced a long time, and, interestingly, the next day, it’s still enhanced, so that lasts 24 hours or longer.

And when we plot that, we find that there’s these two phases of facilitation. And what I mean by facilitation, I mean that the synapse gets stronger for that length of time. One phase lasts about an hour or so after five of them, the other phase comes on later the next day. And we call this phase intermediate-term facilitation, we call that phase long-term facilitation. And what I’ve told you now is that we’ve identified three different kinds of phases to facilitation; the short one that lasts just a little while, if you just give one pulse of serotonin, but we’ve got this intermediate one that lasts about an hour or two, and this long-term one with this peculiar time course in between where nothing’s there.

And now we’re gonna do a couple experiments together. We’re gonna ask whether or not these observations at a synaptic level can predict the existence of memory in the animal. Can it give us some clues about real memory that exists in the animal? We’re gonna do that in two kinds of memory. The first one is intermediate-term. So let me tell you what we know about intermediate-term facilitation. That’s this kind of facilitation, the kind that’s produced out here about an hour after we give serotonin. What we know about it is a lot.

There’s a lot on this diagram that I’m not gonna talk about, but on the bottom, let me just pay attention to this guy. On the bottom, we know that when we produce this kind of facilitation in this time domain, first of all, it has this peculiar time course. It goes up, it comes down, and it comes back the next day, that’s number one. Number two, to induce it, we have to have protein synthesis. Remember that translation I talked to you about? Well, we have a drug that can block that, and we know that this synaptic event, this facilitation absolutely needs the synthesis of new proteins. And once we get that instantiated, once we’ve got that facilitation going, in order for it to be reflected, for the nervous system to tell us it knows it, it has to have the action of a particular kind of kinase, one of those signaling molecules (called PKA, it’s not important you know the name).

So here are the properties of the synapse that we know. This is the work of Michael Sutton in my lab. The question I wanna ask with you tonight is, does that correspond to memory? I’ve only told you about synapses so far, but this lecture’s about memory, this series is about memory. So what would be the prediction, if that synaptic event, ‘cause the intermediate-term facilitation in fact underlies memory, what would we require of that memory, what should it look like? Well, it should have these predictions. One is the memory should have a bi-phasic retention profile. What does that mean? It means it should, let me just go back, it means it should go up and should come down, and then it should come back the next day. A very funky prediction, a very strange prediction for memory, but that’s what the synapses told us. Secondly, it should require translation, protein synthesis to induce it, and third, it should require this signaling cascade, this kinase, for the memory to be expressed.

So first let’s test this hypothesis. Does memory have a bi-phasic retention profile? What we did was take intact animals, shock their tail, and we plotted the memory for that. If you just do it once, you get a little short-term memory. Just look at the red guys. The red guys are the memory for that intermediate-term memory, so five tail shocks gives you a very nice memory that lasts about an hour and a half. But the key prediction was, does that come back the next day as long-term. The answer is yeah. If you track that memory over time, it comes up, it goes down, by about three hours it’s gone, but with nothing but time intervening, lo and behold, the animal has a memory the next day. Strange, huh?

We may wanna talk about this in discussion, it’s a very interesting feature. It’s going on in your heads too, it’s just not as extreme. So this memory course we predicted exists. So we take this prediction, say yep, got it. Not only did the synapse do that, the animal’s memory does that for us. So let’s look at this prediction. The memory should require translation to induce it. We should need protein synthesis to induce this memory. Well how do we test that? You know how to do that. We’re gonna block translation. We’re gonna block translation with a drug, and see if we can block the memory. And what we do is we take this preparation I told you about, where we can now look at the nervous system and change the molecular environment of the nervous system while we induce the memory, and we’re gonna ask whether if we change that molecular environment and block translation, can we block the memory induction.

Here’s what that really looks like. This is the tail of the animal, this is the siphon, and this is the brain in a single little well, in a chamber that we can change all the environment, the neighborhood of these neurons, we can change this environment at will, reversibly and rapidly. So what we did was first ask, we’re gonna use a drug, it’s called emetine, that blocks translation, and we’re gonna say does it affect the baseline responding of the animal. No, the yellow guys are the drug group, the red guys are normal animals, and the baseline reflex is not changed at all.

What about short-term memory? If we give a single shock, we see perfectly fine short-term memory in the yellow guys there, the emetine guys, and perfectly normal in the red, so this translation inhibitor doesn’t touch short-term memory. Well, what about our intermediate-term memory that folks were interested in? Turns out here’s the normal memory that we produce, and it’s completely blocked if we block translation during induction. So now we can go back to our prediction and say yep, got that one too. That is, that memory, to produce that memory requires protein synthesis.

We’ll turn to the last prediction. What about this business that in order to express the memory, we need this signaling cascade. Its name is protein kinase A, PKA, doesn’t matter if you know its name. How are we gonna study that? Well, you know how to do that. We’re gonna block phosphorylation. We’re gonna block that signaling cascade when we produce the memory, and see if expression of the memory, asking you, do you remember? After we produced it, do you remember? And if the animal says yeah, I remember, and now we block this kinase, then the question is, do you still remember? And the answer is like this. Here’s the experiment. We train the animals. Once the memory’s going, now we put on this blocker of protein kinase A, and here’s the results.

This red line is the normal guys, that’s our normal memory we’ve come to know and love. The blue guys are the guys we trained just the same, and right here where this green bar starts, we’ve started to block that kinase with the drug, and look what happens to the memory. Now, this is the real deal, this is not a kinase, this is not a synapse, this is memory. The memory goes into the tank, the memory now is completely back down to normal. And interestingly, as soon as we remove the drug, the memory comes back to the state that the control animals have. So we’ve blocked the expression of the memory. What I just told you is you need this protein kinase signaling cascade in order to say yeah, I remember. You don’t need it to build in the memory, you need it to say yeah, I remember.

So what we’ve done now, we can return to this question mark we had, can we attribute memory to the synaptic changes, and we say yeah, we can say that we’ve identified a kind of memory that has these features that were completely predicted not by other behavior, but by the events in the software of those neurons.

Okay, halfway through. We’ve talked about intermediate, now let’s visit long-term memory. Long-term memory is typically defined by not just how long it lasts, but the fact that it needs to turn on genes and proteins, and we’ll talk about that in a second. So what we’re gonna do is now look at these different parts of the software again, and we’re especially gonna focus on phosphorylation, on kinases, and I’m gonna introduce you to two of them, two of our favorite. One is called tyrosine kinase, and one is called MAP kinase. Absolutely doesn’t matter that you remember their names, but it is important to know that they’re different, so there’s lots and lots of different kinases. They all have the same job, they stick phosphates on molecules, but they do it in different places.

And I’m gonna show you a diagram, hold onto your seat belts, I’m gonna show you a diagram of how these kinases interact. What they’re called doesn’t matter, I’ll just show you that there are these tyrosine kinases up here at the beginning, where it’s had input to the cell, where things that make the cell grow or make the cell learn, will turn them on, and then very often they funnel through this cascade, this cassette of three kinases, and come up to this guy called MAP kinase. If I gave you the real names of all of these, you’d really bolt for the cookies, so I’m not gonna tell you. So I’m gonna tell you about tyrosine kinases and ask if they do anything for long-term memory, and then I’m gonna say, well, if they do, do they act through this guy called MAP kinase.

So let’s first look at this part of the synaptic profile, out here 24 hours after memory’s induced. This is the work of Angela Pissou in the lab. First, Angela was interested in tyrosine kinases. So what questions can we ask? You probably already have some in your mind. One is, let’s block ‘em, we know how to block ‘em. Does that block the induction? Now we’re not talking about memory, we’re talking about just making that synapse stronger 24 hours later. What happens if we block tyrosine kinases? I’ll show you. We’re gonna block ‘em with a drug that only blocks tyrosine kinases, and here’s what happens. Up top is what normally happens, we get five pulses of serotonin, here’s the baseline. See how much bigger that synapse is the next day? That’s this white bar. So that’s our normal long-term facilitation, 24 hours later. However, if we do exactly that same thing in the presence of a drug that blocks tyrosine kinases, we completely block the induction of long-term. So what I’ve told you is yeah, blocking tyrosine kinases blocks the induction of long-term. Keep in the back of your mind, we’re gonna revisit this question when we look at long-term memory in just a moment.

Well, what about another kind of experiment? And this is a really cool experiment, I love this experiment. Can we enhance long-term facilitation by enhancing tyrosine kinases? So, in science, blocking something is convincing. But if you can also turn it on its head and say okay, if I think I got it right, if I now tune up that cascade, can I actually improve matters? That’s another level of proof that’s satisfying. So we’re gonna ask if we enhance, now, we’re not gonna block it, what if we make ‘em better, can we actually enhance long-term facilitation? And the answer is yeah.

See, here’s how you do it. What we did was first use one pulse of serotonin. One never gives you long-term, trust me. That’s what it looks like the next day, all right, one pulse of serotonin, look, nothing happens 24 hours later. But if we put, if we do exactly that in the presence of a drug that tunes up tyrosine kinase, there’s a neat trick, if you wanna talk about it in discussion, I’ll tell you, but you can tune up the tyrosine kinases that already live in that nerve cell. Look what happens. That single pulse of serotonin never gives you long-term normally, now gives you beautiful long-term. Look at that long-term. Isn’t that nice long-term? Don’t you like that long-term? So what I’ve just told you is we’ve improved long-term facilitation by improving tyrosine kinases, so now you’ve got two facts. I’ve told you you inhibit it and you make the synapse not strong the next day; if you activate it, you facilitate the process.

One last question, and that is I’ve told you about tyrosine kinases. What about these guys, MAP kinases? Remember them? Tyrosine kinases are up here. The question is, do they funnel through this MAP kinase? MAP kinase is sort of like the neck of an hourglass, lots of roads converge on MAP kinase and then again diverge. Very important place in the software of your nervous system. So the question is, how are we gonna ask about MAP kinase? Remember this improvement I just showed you that you liked so much? Does that require--you did like it, come on, in your soul. All right, if you get this improvement, does that require MAP kinase? Let’s ask the question. We’re gonna block phosphorylation not of tyrosine kinases, we’re gonna block MAP kinase. How are we gonna ask it?

Well, here’s our normal gain of function, here’s our normal improvement, so here’s our pre-, one pulse of serotonin, never gives you long-term, but we made it all better because we gave that single pulse of serotonin in conjunction with this activator of tyrosine kinase, so that’s the normal case. And now we do exactly that same thing, but now we block MAP kinase. Here’s our normal, new and improved facilitation completely blocked if we block MAP kinase. So what does that mean? It means that yeah, the enhancement that tyrosine kinase gives us acts through MAP kinase.

So now I’ve told you a bunch of molecular stories inside the neurons. Let’s go to the last part of this experiment. The heavy lifting is can we take these molecular events and predict the existence of a memory we didn’t know existed before? Or at least can we predict how that memory should be built, what it should be made out of. So how do we do that? Well, there’s a couple predictions for memory. If this is true, up here in yellow, here’s the prediction. What is the memory--we’re not talking about synapses anymore--the memory should be enhanced by increasing tyrosine kinase activation, we can make the memory better. And, if we can make the memory better, the next prediction is that improved memory should require MAP kinase activation to be put into place. Those are directly predicted from the molecular stuff we just did together.

So let’s first ask the question, does the memory require, or can we improve memory by activating tyrosine kinases, how are we gonna do that? Gonna go back to the books, we’re gonna go back to the preparation, where we use the ability to block or activate drugs in the nervous system while we study memory. And what we’re gonna do is give a couple tail shocks, never gives you long-term, only gives you short-term. In fact, on this diagram you can see, this is all plotted what happens the next day. Zero means there’s no change from the day before. So a couple tail shocks never give you long-term memory, trust me, all right? However, if you do exactly that, give a couple tail shocks, never give you long-term memory, in the presence of that drug that tunes up tyrosine kinases, boom! You get improved memory. You essentially have changed the rules such that now that sub-threshold training induces long-term memory in the animal.

So we’ve just improved memory by improving the activity of tyrosine kinases in the central nervous system. All right? Got it? We can check that one off. That prediction was fulfilled. That improved memory, if we’re right, should act through MAP kinase. Well, let’s check that out. How are we gonna do that? You know how to do that, we’re gonna block MAP kinase when we improve this memory. So we do exactly the same thing, we take this improved memory and we train the animal now with the improved memory technique, but in one group we’re gonna block MAP kinase, and here’s what happens. Instead of getting--this is the normal improved memory, in the control group, but if we do that when we block MAP kinase, that improved memory’s completely abolished.

So not only did we improve memory with tyrosine kinases, that improved memory absolutely relies on MAP kinase for its induction, so we can come back here and say yeah, we’ve nailed both of these hypotheses, we’ve predicted, from our molecular software, we predicted what the memory should look like in the intact animal, and we nailed those predictions. So now we can return to this core hypothesis. What I’ve done tonight is give you a few examples of how we’ve literally reprogrammed the software of the neurons by turning up protein kinase A, turning up tyrosine kinase, turning up MAP kinase. The names of them don’t matter at all, but understand that all those signaling cascades are sloshing around in your neurons as we speak. They’re involved in a lot of brain function, including the induction of memories.

So I’ve talked to you a lot about this software. I wanna just end by talking about these signaling cascades a little more. Not in terms of experiments anymore, you’re probably exhausted doing all these experiments, but what we will do is just consider it a little bit theoretically. What’s up with all these kinases? And just to really boggle your mind, this is just a drawing, I’m not gonna tell you what it is, of some of the kinases that we’ve identified in our nervous system, and Aplysia’s important memory, and this is the tip of the molecular iceberg, there’s literally dozens more. So what do we need ‘em all for, what are all these kinases doing?

I’m gonna suggest to you that these kinases, these signaling cascades are very well suited to broker the temporal domain of memory. So for immediate and short-term memory, simple phosphorylation itself can carry the game. Just how long the substrate’s got the phosphate stuck on, that could be the duration of the memory. If it lasts ten minutes, it lasts ten minutes. However, when we get into longer memories, the kinase itself can be turned on persistently, and it can also broker the protein synthesis step that can give you intermediate-term memory. And not only can you get intermediate, if you get to long-term memory, you can start turning on transcription with kinases. They’re terribly important in turning on genes and ultimately giving you a permanent structural change in the brain. So these kinases are really, really well suited to handle the temporal domain, the temporal demands of memory.

Well, so far I’ve told you how long you’re remembering something. I haven’t told you what you’re remembering. There’s another whole domain, and that’s the content, and some of the things you can remember are reflex memories, I’ve talked about that today in the animal we’ve been studying. Motor memories, that’s the pawing at the clutch, right, with your left foot, that’s a motor memory. Remembering how to swing a golf club, that’s a motor memory. Spatial memory, finding your car in the parking lot, big deal, that’s another kind of memory. Emotional memory, something wonderful, the birth of a child, something terrible, the loss of a loved one, there are profound memories in our brains. Another kind of memory’s called executive memory, that has to do with planning, figuring things out, carrying out planning in memory. And this is simply a tasting menu, there’s lots of different contents to memory.

So what’s up with these? How do these work? Well, I’m gonna suggest to you, this is hardly novel, that these work by virtue of the circuits that are activated by the experience that gives rise to them, so reflex memories are mediated by spinal cord circuits, motor memories by circuits in the cerebellum, spatial memories by circuits in what’s called the hippocampus, known to be important for spatial memory. Emotional memories go straight through the amygdala, very important structure that this university has a deep affection for. And finally, executive memory has to do with the frontal cortex of the brain, and I’m gonna say to you that these different kinds of contents are subserved in the brain by virtue of the experiences that activate particular circuits, and it’s the circuits that hold the content memory, okay?

Now we’ve gotta put two things together. We’ve got the content here, but we’ve got these cassettes, these temporal cassettes that can be used for the time-dependent virtue of memory. How might that work? Well, I’m gonna tell you, one-way to think about it is you take these cassettes and you import them into neurons in the spinal cord, put ‘em in the cerebellum, put ‘em in the hippocampus, so you can have an intermediate-term memory in the hippocampus, an intermediate-term memory in the cerebellum, an intermediate-term memory in the amygdala by virtue of the kinases that live in those neurons, and you can take this sort of temporal scheme and import it to all these different structures.

So I’m gonna suggest to you that the kinases themselves are absolutely agnostic to where they live. They don’t know they’re in the cerebellum or the amygdala, they just know they’re in the neurons, and their job is to code the temporal domain of memory. Now, it happens that if they live in the cerebellum, they’re gonna code the temporal domain for a motor memory. If the same kinases live in cells in the amygdala, they’re gonna code the temporal domain of an emotional memory, so that’s the idea.

This kind of model is helpful in our thinking, it’s certainly wrong. It’s certainly wrong in its details, it’s probably wrong conceptually, but what I’m showing it to you for is to tell you that we use this in science as a way to sort of formulate our thinking, get our head on straight about what prediction should be, and then test the hypothesis, and sometimes they’ll be borne out and you’ll feel good about it, sometimes you’re wrong, and that’s fine too, ‘cause the game is not always being right, the game is to figure out the truth of the matter, so sometimes when your hypothesis isn’t confirmed, it tells you a lot more. ‘Cause it tells you whoops, got that one wrong, have to go back to the drawing board.

So this is certainly gonna be wrong in its details, at least, but what it does do is illustrate for you the way my labmates and I sort of sit around thinking about both the content and the temporal domains of memory. It helps us at least figure out experiments that we might do next to try to push that field forward.

So with your permission, I’m gonna make a couple real closing comments, very quickly. First, I have to thank the folks that are actually in the trenches in my laboratory: Nigel Emtage, Yelena Mollushagen, Stefan Marnesco, Ernie Muller, they provided a lot of the background work; Mike Sutton, Angela Purcell, Shiv Sharma, Carolyn Sherf, Martha Bagnall, Adam Bristol, Kate Reisner, Joanna Schofhausen, Justin Show, Christine Kulkman, that wonderful undergraduate that won that award, Nicole Betard, Paul Hofstetter. These are all colleagues of mine that I’m absolutely blessed to work with. This research team makes it fun to come to work in the morning. They’re equal parts brilliant and irreverent, so they keep us honest, and they’re fabulous colleagues to work with, so I wanna thank them.

The final comment I wanna make is a postscript, and it has to do with progress. I’ve talked to you tonight about the fact that the field is moving extremely rapidly, the progress has been extraordinary. I sometimes kid with my friends that if we could take a ten-year time travel back, so go back ten years and have a crystal ball to see what we know now, we wouldn’t even have the vocabulary as scientists to describe what we know now. We wouldn’t have the things I just told you about, transcription, translation, some of the kinds of approaches, we had hints of it but we simply didn’t understand it, so there’s been tremendously rapid advances in the field on two levels, technically and conceptually. We think about the problem better, and the techniques are extraordinarily powerful for studying memory, and again, we can talk about that in discussion if you like.

What I wanna tell you is that some of that progress, a good bunch of it happens right here in your own backyard, at the University of California Irvine, at the Center for Learning and Memory, led by Jim McGaugh, at the Center for Brain Aging and Dementia, led by Karl Kopmann and Frank LaFerla, at the Irvine Center for Studying Spinal Cord Injury, led by Ozzie Stewart and his colleagues. These are hotbeds of science going on 24/7 here, all in one-way or another asking how the brain changes in response to the normal things that inform it, and in response to the abnormal things that we’d love to fix. So I’m proud to be a part of this community, it’s one of the places where this progress is being made.

But finally, just a post postscript about progress, I absolutely stand by this statement, it’s been extraordinary, it’s been rapid, it continues to be exciting, but I can also assure you that despite this progress, as we go on this journey trying to understand brain mechanisms in behavior and understand the brain mechanisms in memory, at least a component of that journey will always be a magical mystery tour. Thank you.