 Okay, we're going to get started. So thank you all for braving the cold weather to be here. We're looking forward to Dr. Bassler's second lecture as part of the Slembach Lectures. And so to honor Professor Slembach, who had many important contributions to both biochemistry and nutrition, the Slembach Endowment sponsors a two-part seminar series, and so this is a second lecture in the seminar series. And so yesterday, Dr. Bassler told us two very fascinating stories. One about developing novel approaches to image and visualize the spatial temporal changes in biofilm growth. And the second fascinating story was with inter-kingdom signaling between the host and also the microbiome. So today, Dr. Bassler will tell us a new set of stories and her talk is entitled How Bacteria That Slewish Self from Other. And please join me in welcoming Dr. Bassler. Thanks, Ophelia, for round two. Actually, it's a real luxury and a privilege to... Well, it's always a luxury and a privilege to talk here, but it's really fun to get to tell two stories, right? Because usually you just pick basically the one thing to talk about. And so this was fun for me because I got to put both these seminars together. And so I just want to say this is sort of the end of these two days. It's just been absolutely brain candy for me. I learned, there was like maximal learning of science for these two days. And so I just really appreciate the generosity and inviting me and that everybody's spending so much time with me and taking such care to explain all your projects to me in these ways that I could understand and learn a lot and have a lot of new ideas. Anyway, I just have always admired this place and I admired even more after these two days. So many, many thanks, especially to Ophelia, but to everybody for these two fun days. Okay, so let's see. So theoretically, these slides should be different from yesterday. So just for those of you, the first two slides are repeats and then I promise they're all different after that. I still work on quorum sensing. So they won't be that different, but the first two slides will be different. So let me just, if you guys can all text an email if you were here yesterday while I do the first, like give the context part. So what my gang works on is we try to think about how bacteria get any bang for their buck. So there are these itty bitty critters and yet they can kill us, right? So they're notorious pathogens. And so if they get inner on us, they can either make us ill or they can kill us. And then increasingly, and of course we talked about this a lot yesterday, we now understand that all higher organisms live in consortia with these microbiomes, which give all of these traits to the higher organisms that are not encoded in the genome. And a few of those like digesting plant food, making micronutrients, educating an immune system, those are some of the traits that we're only now learning about that are actually beneficial, right? So what's very clear in my field is that some bacteria on this earth are awful, some are magical, right? But the question is these are all big tasks. So bringing down a higher organism, you know, killing it, that's hard, keeping it alive, that's hard too. And so the question from my gang's perspective has always been how can they manage these amazing feats? And of course, at least part of that, what we now understand is by communicating with these chemical quorum sensing languages, synchronizing gene expression, which is really behavior, and acting in groups to carry out tasks that would be ineffective if a single bacterium acted alone because the bacteria are too small individually to have an effect, right? So that is quorum sensing, and here's the second and final repeat slide. I made it in different colors, so that maybe you wouldn't notice that it's a repeat. It's probably not working that well. Actually, someone told me that no one remembers anything that you say in a seminar 15 minutes after, so I could just give the same seminar every single day, and everybody would be like, oh, you've made so much progress. But anyway, I'm not gonna actually try that today. So here's the second repeat. So are you actually taping this? Anyway, so here's the, I thought about that last night too, okay, so here's the second, right? So here's quorum sensing at the simple version of it. So bacteria are either alone or they're part of communities, and they have to be able to distinguish between those two situations and appropriately behave based on which of those places they find themselves. So the way they do it is with these chemical communication systems, right? So if this is the bacterial cell, and it's in dilute suspension, the world is big, these bacteria are tiny, right? And so what, so they wanna be acting, going solo. And so what they do is they have some genes expressed that are at their disposal, but among the things they do is they make and release these small molecules, these hormone-like molecules that we call auto-inducers, again, depicted by these triangles. And at low cell density, the auto-inducers diffuse away so the bacteria can't detect them, they don't build up in concentration, and that says act as an individual. But as the bacteria grow and divide, since everybody is making some auto-inducer, the auto-inducer increases in step with cell density. And when the extracellular concentration of the auto-inducer reaches this critical threshold level, the bacteria detect it, and then in unison, all of the bacteria change their behaviors and they begin to carry out collective tasks, right? And so again, and this is gonna become important during today's talk, typically somewhere between 200 and 600 genes in each species of bacteria will be regulated, so it's a big deal when the bacterium decides to change from going in alone to acting as part of a community, a quarter of its genome can be involved in these, under these regulatory circuits, right? And then very often, this is also gonna be important today, the kinds of products that are, or the tasks that are controlled by quorum sensing are the release of exoproducts. So these public goods, so that they put stuff like exoenzymes outside of them or virulence factors, they put them outside of themselves and you never get your own product back. So this only works when the collective does it in synchrony and I get your protein and you get mine, right? And so they share in the collective benefit of making these public goods. And they have to be careful somehow not to be given those away willy-nilly. They're expensive and you don't wanna get cheated, right? So that's kinda gonna be what today is about, right? So anyway, they carry out these collective behaviors, hundreds of genes will be controlled by these communication circuits. And then one last time, and this is actually gonna be for the whole talk, this all was discovered initially by Woody Hastings in a couple of bioluminescent bacteria. And so the first gene, even though hundreds of genes are turned on, one of the sets of genes that turns on is the bioluminescence genes, which is encoded by luciferase, right? And so that is shown on this slide. So the favorite bacterium that I work on from these very first two famous quorum-sensing bacterium is this one named Vibrio Harvii. So this is a free-living bacterium that has quorum-sensing and it turns on hundreds and hundreds of genes, but among them are these luciferase enzymes, right? And so what you're looking at is what we see in the lab. So this is a flask, this is a pH plate, this is like the ridiculous stuff that the graduate students in my lab spend their time doing, right, this is why there will not be that much progress in today's talk. Anyway, and then, but being serious, right, Bonnie? I think we blew that yesterday. Anyway, being serious, right, the way we took this picture, we just turned the lights off in the room and this is what we see, right? So we're not doing anything to the cells, they make this beautiful bioluminescence at high cell density. And what was amazing about this early on in the field, right, is that you can just see it with your eyes and what else happens is that it turns on 10,000-fold. So this is a built-in reporter of the bacteria communicating and it's an exquisitely sensitive reporter, right? And so I will be used, so hundreds of genes turn on in Harvey eye, but what you can see is bioluminescence and so this is the only reporter that I will be using for all of today's talk. So in everything that I'll be telling you as the readout, we will be measuring quantitatively light production in this bacterium as the reporter for quorum sensing, right? And it's built in. Okay, so that's what we're gonna measure as one of the outputs, right? And then just one more time, there are multiple molecules involved in quorum sensing. So there is a molecule, a homocerein lactone that we call auto-inducer one because it's the first one that was discovered which has this ring, this nitrogen group and a four carbon chain and you're gonna have to notice this chemistry today. There's a hydroxyl group on the third carbon. So this molecule, as far as we can tell, Vibrio-Harvii and only Vibrio-Harvii makes this molecule and so it allows it to count its siblings, right? It's this a private secret conversation that Vibrio-Harvii has with its clone mates to understand how many of me are present. There's a second molecule which is for the genera Vibrio. So this molecule, which we call CAI-1, right? Is this fatty acid molecule and as far as we can tell, all Vibrios make this molecule but nobody but Vibrios has the synthase nor as from sampling studies, nobody else makes this molecule. So this molecule says I'm a Vibrio, right? So this molecule says you're my twin. This molecule says you're related to me. And then the third molecule that we discovered is this molecule that we call auto-inducer 2. It's a five carbon molecule decorated by a bunch of oxygens. This seems to be generic. We think it's sort of the bacterial trade language that is made broadly in the bacterial world by both gram negatives and gram positives. They all make this identical molecule. So we think there's no species information in it. So this is for inter-species communication. It just says I'm a bacterium, right? So what we think Vibrio-Harvii and probably other bacteria are doing is they're measuring the buildup of these molecules to say how many cells are around me and then encoded in those molecules. It's not just cell number. It's how closely or how distantly am I related to you? And then the bacteria tailor their behaviors based on both the total amount of molecule and then the combination, the blend and the ratios of the different molecules. And so then they behave appropriately both based on cell number and who the neighbors are. That's from yesterday. All right. And so now, in much more simple form that I told you yesterday, everything now is about Vibrio-Harvii for the rest of today. Okay, so the way that it works is that at low cell density, so when the bacteria are alone and the auto-inducers are not there, these three receptors that respond specifically to those three molecules, they are kinases. And through a series of sort of gruesome steps, they send phosphate to this protein called LuxO. So low cell density, they're kinases. They send phosphate to LuxO. And LuxO, again, indirectly represses quorum sensing. So that means no light and none of these public goods, which I'll call these goodies, okay? So low cell density means kinase, no light, cells are dark, no public goods. At high cell densities, when the auto-inducers accumulate and they get bound by their detectors, the receptors switch from kinases to phosphatases. Phosphate gets pulled out of the circuit. LuxO is dephosphorylated, so it can't repress. So that means make bioluminescence, like I showed you on that slide, and make all of these invisible to us, but these public goods. Okay, so for the whole talk, no auto-inducer means kinase, no light, plus auto-inducer means phosphatase and bioluminescence. Okay? All right, so that's how it works. And I told you a lot about that yesterday. And so then the problem that we started to think about when we were trying to understand this is how do bacteria really interpret this blend of molecules? Like, we really wanna understand that. And then, so for today's talk, I'm only gonna talk about this receptor and how it does its job. And so here's the problem, and we knew this problem even when we were figuring that stuff out. Okay, so I already showed you that the molecule that Vibrio-Harvii detects is this homocerein lactone with four carbons and there's a hydroxyl group on the third carbon. Okay? So that orange receptor detects that molecule. So what's the problem? So the problem is these are other Vibrios that live in the ocean, that live with Vibrio-Harvii, and each of them has their own intra-species communication molecule, right? So Dishari, Mark Mendel loves this guy. This is its auto-inducer, right? And Anguilarem does this one. And so what you can see is that the right-hand parts of the molecules are identical, but these chain links can change and then these decorations can change. And so the first thing we wondered, right, that orange receptor, which is called luxn, right, it's supposed to detect this molecule. So what we wondered is, can it detect these other molecules and or not, you know, doesn't get confused, right? So we just wanted to understand how good is this, right? Because it's swimming around in all of these molecules. So how can quorum sensing happen if this is about intra-species communication? So Jaboki, who was a graduate student in the lab, wanted to take this on as his project. And so what he decided to do was to build a Vibrio-Harvii reporter strain that has exactly one receptor. So it has this luxn receptor and the other two receptors are gone, right? So there's not gonna be any input from those other two systems, they're deleted, right? And what you know is that if there's no, and oh, he also knocked out all the auto-inducer synthesis, right? So we're gonna add auto-inducers with pipettes. We can synthesize all of these molecules. Okay, so what you know, in this strain, there's one receptor, luxn, and there's no endogenously produced auto-inducer. So that means kinase, no light, right? That's what I've told you. So then, if we add auto-inducer one with a pipette, right? The receptor will detect the auto-inducer, it converts to phosphatase, that means make light, right? So everybody has, was with me, right? So then what Jabo was gonna ask with this reporter strain, what happens if I add one of these non-cognate auto-inducers? And just like yesterday, everything will be color-coded, endogenous will be orange, blue will be non-cognate, okay? So can the receptor detect them? So he did that experiment. And so then again, just to reiterate, what are the things we're gonna change, right? So all homocerein lactones are different in only two ways. They can either have a hydroxyl, a ketone or hydrogen on this third carbon, and then the tails can be different lengths, right? So we can change those two features. And so we will be changing those two features. And so Laura Miller, who was a fabulous chemist and a postdoc in the lab, she made us a whole set of different homocerein lactones that are different in those features. And then Jabo tested them, okay? So RLU, this is relative light units. So this is light production, and then we're gonna add different molecules. And you have to notice in all of my graphs, these are log scales, right? So if you add just solvent, right? No auto-inducer, the receptor is a kinase, no light. If we add auto-inducer one, right? C4, three hydroxyl group, light turns on 5,000 fold. This is why we like bioluminescence, okay? Okay, so now what about all these other auto-inducers? And so Laura made us a bunch of them. And so here what we're doing is C4, C6, C8, C10, C12, right? Here it's ketone, there it's hydrogen, right? So those are the, change the C3, change the length of the tail. What you are supposed to see is that nothing happens, right? This receptor is absolutely perfect, right? It detects this auto-inducer and none of these others. Okay, so the bacteria don't get confused. So they're swimming around in all these homocerein lactones that other bacteria are producing. And in fact, this receptor is not detecting them. So that has to be, first of all, how quorum sensing works, right? Is that the receptor can't get confused or this couldn't work? Okay, so that's part one, right? That Jabo answered that question. Okay, so then the question, I'm not gonna know these slides by heart because I will be start telling you yesterday's slides in any moment, okay? So how's it do it? Okay, so here's what we understand about Lux N. This is our high resolution structure of Lux N, right? It has nine transmembrane domains, right? So it's sitting in the cytoplasmic membrane, right? And then on the inside of the cell, the enzymes, this kinase and the phosphatase, right? It's a two component protein, right? So somehow, so we don't have a clue how it detects its own auto-inducer, nor do we understand how it discriminates against all those other auto-inducers, right? Because, you know, it's a protein with all these transmembrane domains and we've never gotten a structure of it. And so Jabo wondered if he could start to figure that out in the absence of a crystal structure or now maybe a cryoem structure, you know, could he figure that out, right? And think about how auto-inducer detection and discrimination works. And so what he decided to do was to make a library of mutants, right? So he had cloned Lux N onto a plasmid and then he just took the DNA that encoded the nine transmembrane regions. We know that's where the auto-inducer gets detected and he did a random mutagenesis of that, right? So he got thousands of Lux N and then he reassembled the genes, of Lux N genes, each with a different mutation somewhere in those transmembrane domains, right? And then he could put them into his Vibrio Harvii strain, right? And then what he has now, remember, there's no other receptor there and it's not making auto-inducer. So now he just plates them out, right? And he's got theoretically a Vibrio Harvii that has a Lux N protein all with a mutation, you know, we don't know where, but somewhere in this transmembrane region, right? And then he could add auto-inducer with a pipette, right, and most of them will be perfectly fine. But occasionally, he suspected that he would find a Vibrio Harvii that had a point mutation in the transmembrane domain that could no longer respond to the endogenous auto-inducer. And if he found that, that would mean that perhaps that amino acid is part of detecting the homocerein lactone auto-inducer, right, so he did that experiment and he measured light production of all of his mutants and this is just, you know, a bunch of different mutants and what you can see, so here's wild type, it turns on light, not every mutant was dark, we got plenty of mutants, they're not on here, which made wild type light, but he got mutants that were dark, okay? So those presumably are defective in auto-inducer detection and so he could map them sort of onto the transmembrane domain and what was really cool about that experiment is that these are the mutants that don't, if you have a mutation anywhere in these amino acids, they cannot turn on light in response to this auto-inducer and what was neat was that they all clustered to the outside of the protein, which is where the auto-inducer should be and to just these few transmembrane regions and so in our imagination, we think, you know, it's not a flat structure, it's 3D, we assume that somehow those couple of transmembrane helices fold up to make a binding site that allows the auto-inducer to bind, okay? So that actually worked and so then what Jaubot wondered is, okay, so maybe these orange residues are defining the auto-inducer binding site, right? But how do you get discrimination against all those other auto-inducers that I showed you, right? So remember, you know, these auto-inducers can change both at the carbon, the third carbon, what moiety is there and then how long or how short they are and so he wondered, could he take that kind of a strategy and push it further to actually figure out how the receptor not only binds a homoceryl lactone but only responds to the right homoceryl lactone, C4 with that hydroxyl group, right? So he built a different kind of reporter strain. So now what he knows, right, is he has this strain that has wild-type luxen, there's no auto-inducer, so that means dark, right? And so what he wondered and what I've also shown you, sorry, is that if we add any of those non-cognate auto-inducers, light doesn't turn on, that's that first bar graph that I showed you, right? So this receptor does not turn on light in response, does not signal in response to those non-cognate auto-inducers. So then what he wondered is could he find in that library of point mutants that he made, could he find now mutant luxens that you give a non-cognate auto-inducer and they turn on light? So he's just gonna screen through for that and he can say yes or no, he can find such a beast and then afterwards he can ask, do they or do they not still respond to the cognate auto-inducer, right? So have you altered the specificity or do you have a mutant that can now respond to multiple auto-inducers, right? So same library, right? So he plates it out and this time he adds a non-cognate auto-inducer and looks for a guy that turns bright, right? Everybody gets that, right? So this is the same library and it's just the opposite, just look for a bright guy when we add the wrong auto-inducer. Okay, so now he's gonna actually do this experiment, he's gonna do it twice. So the first thing he's gonna do is he's gonna ask, can I get a mutant that now responds to an auto-inducer with a C4 tail but where we've removed that hydroxyl group? So remember, I told you, wild type luxen responds to this and not to that, right? And so he's screened through his, actually I'm just gonna show you that again, right? Yeah, so anyway, so here's light production. So these are the data, a dose response curve of the data that I showed you in that bar graph, right? So light production turns on in response to the cognate auto-inducer but if you add DMSO, that's black or if you add the non-cognate auto-inducer only when you get to ridiculous concentrations does light start to turn on, right? So that's kind of the fleshed out version of that very first graph I showed you, okay? So the receptor doesn't like the green auto-inducer. So he fishes through his mutants and sure enough, it turns out if you have, if you change histidine at 210 to asparagine, now what you can see is that in fact the receptor prefers the non-cognate auto-inducer to the cognate auto-inducer, right? So it actually responds to both but it likes the green auto-inducer better. So indeed, you can make a one amino acid change in the receptor and you can change the specificity of the receptor. So that shows from here to here but you remember there's three choices for what can be on the third carbon, right? So it turns out, here's the wild type again, it doesn't respond to any of the non-cognate auto-inducers. Jabbo's mutant actually doesn't care what's on the third carbon, right? So it prefers a keto group or a hydrogen over the wild type auto-inducer, right? So it's lost specificity for the third carbon moiety and it turns out that it doesn't matter what you change that amino acid to. See now what you can see is that whether we change it to, what is that, threonine, you know, it doesn't matter what we change the amino acid to 10 at, it now has no specificity for what's on that third carbon, right? So what those results together mean is that this histidine discriminates against anything but a hydroxyl group. It says you must have that hydroxyl group, okay? So now he's got that mutant so where is that, right? So we can go back to our topology map and histidine too, it had better be right in the middle of the binding site or he was fired, right? So of course it's in there, right? It had to be, we think the orange thing is the binding site so that hydroxyl group better be right there, right? Okay, so sure enough, so what we would think now is that somehow a homeocerein lactone slots into those orange amino acids and then this histidine, the green one says, is there a hydroxyl group on C3 or not? And if there is, I might think about signaling, if there's not, I will not signal, okay? So that's that and now we can do the same trick with the tail, right? So remember the tail is four carbons long and the endogenous auto-inducer, so we just doubled it. Laura made a molecule with, this has eight carbons. So it's just easy to double it. Same trick, I've already shown you this in the Bargrak, here's light production, the wild type luxe end responds to the cognate auto-inducer, absolutely no response if the tail is longer. Jabo fishes through the library and sure enough he finds one amino acid, leucine 166 that if you change it to alanine now you swap the preference, right? It loves the long homocerein lactone and it tolerates the cognate one, okay? This works like, this is about steric hindrance. So then Jabo did the same thing, he changed the amino acid to a few different things. So here, this is wild type, right? It likes C4 and hates all these longer molecules. When he changes that leucine, right? Leucine's a big amino acid. When he changes it to alanine, a little amino acid. Now C4, C6, C8 all work but not C10, right? So it reduces the discrimination, right? If we changed it to valine, so valine's smaller than leucine but bigger than alanine, right? Now the receptor will accept C4, C6 but not C8, not C10. So this is really like puzzle pieces. You have a long auto-inducer, you need a tiny amino acid, you have a short auto-inducer, you need a big amino acid so that the space gets filled up, right? That's how we imagine this has to work, okay? Where is it? Okay, yeah, right, it is, right? So here's leucine 166, right? So now what we really think happens, right? Is the auto-inducer slots in there and then there's a double test. One is like what's on the third carbon and then the blue amino acid says how long or how short that tail is. So there's two tests that the auto-inducer has to go through and then signaling happens, okay? So then the last one of these gruesome things that I will show you is that he can do a trick and make a double change, right? So we can change the hydroxyl group to a hydrogen and we can change the length of the tail from C4 to C8, right? So that will be the green and the blue molecule is a doubly modified molecule, right? And so then what you'll remember, so this is the hardest graph but it's actually really easy to remember. So what you'll remember when we change histidine at 210 to asparagine, now you change the preference from orange to green, right? These are the data you already saw, right? So now this receptor likes that molecule but it won't touch this long one, right? So this is about C3 and then the converse is true here. When we change the leucine at 166, now it prefers the long molecule but there still has to be a hydroxyl group on it. It won't touch this molecule, right? But if you make a doubly changed receptor now the favorite molecule is the doubly changed molecule. So this receptor which has two changes likes a molecule that has a long tail and a hydroxyl group, right? And it won't touch the others. So in fact the two gatekeeper amino acids they act independently. Each of them tests their own moiety, tail or decoration and you have to pass the two tests, right? For signaling to happen. Okay, so that in the absence of crystal structure is how we think the receptor thinks about these molecules. The orange amino acids say I'm a homocerin lactone then the blue one says my tail is four and the green one says I have a hydroxyl there. Okay, so now let's go back to all these vibrios, right? So this is Vibrio Harvii and I've told you how the receptor accepts this molecule and discriminates against all others, right, so now you understand at least as well as we understand how it does that. Okay, so what about this guy? So this guy, Vibrio Fischeri, its endogenous auto-inducer is C8 with a hydrogen. That's the molecule we have been using in our experiment and that's what this one loves. So how does that work, right? It actually uses the molecule that we had to change the receptor to get it to accept, right? So now that we knew all of this, Jaubot could actually look at the luxe N from this organism called Vibrio Fischeri and ask, what does it look like, right? Okay, so here's the auto-inducer Vibrio Harvii uses. Here's the auto-inducer Vibrio Fischeri uses. And so you'll remember in Vibrio Harvii, in the wild type at 166, which is about the tail, right? There's a leucine there and at 210, there's a histidine. And then when we changed it to accept this molecule, we put an alanine and asparagine. And so in the Vibrio Fischeri receptor, there's an alanine and asparagine. So Vibrio Fischeri is the doubly mutated receptor, right? And so it actually has the two amino acids in those places that accept the orthogonal molecule. And now you can say, why did you force us to actually learn all of this when it's like that? Okay, so in Jaubot's defense, right? You'll remember that when he started, we just had this topology map. We didn't know where the binds that we... There's tons of polymorphisms in these luxn-n proteins. So we didn't know where to look and which amino acid to look. And his experiment told us to look at those two. And again, in his defense, it took him a year to do it. It took nature four billion years to figure this out, right? But he only took a year, right? But anyway, but now by doing this, because we don't have crystal structures, right? By doing this, we can actually now look at these luxn-n's and we can actually kind of predict what the auto-inducer could be and what it won't be, right? Because now at least Jaubot has pinpointed where these auto-inducers bind and at least some of these gatekeepers that are restricting or enabling different auto-inducers to be the signals, right? Okay, so that's how we think that works. Okay, so now you get it, right? That the auto-inducer presumably touches maybe these amino acids, the two checks get done. But then after that happens, right? There's another event, right? So the auto-inducer slots in there and luxn-n says, okay, you're the right one. But then somehow that chemical information that's sitting in the receptor has to get turned into a signaling event, which is about changing a kinase or a phosphatase activity on the inside of the protein, right? So Jaubot wondered, could he actually figure out what the switch is, right? So these two component proteins have been studied forever and we don't understand how they take sensory information and convert that into enzyme activity, right? So he wondered if this strategy of his might actually work. And so what he wanted to do was to get mutants that signaled independently of auto-inducers. So here's the idea. He's got his reporter and you know that luxn-n without any exogenously added auto-inducer, it's a kinase, means no light. If we add the cognate auto-inducer, light turns on. And so he wondered, could he just look through his library? He's really made a lot of use of this one library. If anybody wants to be a graduate student or postdoc in my lab, you need to make one library, which took about a week and then mine it for three years, okay? So he took the same library and he just asked, are there any guys in there that if I don't add any auto-inducer, they make light? And so those guys would be guys that have switched, right? And so he did that experiment and sure enough, he could find many mutants, point mutants in his library that were constitutively bright. So he maps those exactly the way I'm telling you and they all also cluster. So these are the silver ones. And so all of those that are constitutively bright are right here. And so the receptor's modular, I mean, right? So these few amino acids on the outside. So remember, this is the outside and this is the inside. These guys, you know, allow that auto-inducer to cuddle in there and they do the double checks. And when it passes the two tests, well, there's three tests. One, your homocerein lactone, that's orange. Two, you have a hydroxyl group that's green. Three, you have a C4 tail. When you pass all those tests, then that information presumably gets communicated to these amino acids that are sitting right under it but they're on the inside of the cell and then that drives the kinase and phosphatase, right? So that we believe is the switch. And so we think that a crystal structure would certainly help us write our cryoEM but we think that those are gonna be the key amino acids that are actually allowing auto-inducer detection, discrimination, and then signal transduction, okay? So how we think this works now then is that at low cell density, when the bacteria are alone, obviously you don't wanna be making all these goodies, all these public goods. So the receptor is a kinase in the absence of the auto-inducer. The bacteria start growing. The cognate auto-inducer increases in proportion to cell number. It binds the receptors. I told you how that happens. The receptor converts to a phosphatase. You make light and you make all these hundreds of public goods. However, if the wrong bacterium, a non-sibling, is wandering around in your vicinity and it's donating homocerein lactone auto-inducers to the environment but it's not the right one, there can be tons of homocerein lactone there and then this receptor will not flip the switch so that it presumably, the bacteria, don't lose all their public goods to these competitors. Okay, so that's how we think it works. And so then that would mean, if that's all there was to it, that this receptor is still kinda dumb. Like it's really good at detecting its own auto-inducer but it's impervious to everybody else. It just doesn't detect them. And that's actually not true. So there's one, this is Xiao Bo's work. The most fantastic twist of this is that this receptor indeed does not signal in response to a non-cognate auto-inducer. I've shown you that like 50 times already today but it is absolutely keeping track of those auto-inducers and I'll show you that right now. Okay, so in this experiment, what Xiao Bo has done, this is your, it's the same reporter strain, light production, right? And if you add DMSO, nothing happens. If he adds auto-inducer one, that's the cognate auto-inducer, he's added it at the EC50. So this is half maximal, right? Light turns on. Okay, so he's gonna keep that auto-inducer there, the AI-1, the cognate auto-inducer and now he's gonna add a non-cognate auto-inducer. So if he adds C6, light goes down a little bit. C8, it goes down more, C10, C12. So what, so AI-1 is here at this level but if you add in a longer auto-inducer, they are competitive inhibitors. So the receptor is not inert. Every one of those fits into the binding site but they're antagonists. And the longer they are, the more potent they are. So sort of the farther away they are from the real signal which might have to do with how far you are from me as a species, the more potent they are. So every one of these auto-inducers is bound by that receptor in the binding site because we can show this is competitive. Okay, so the receptor is not inert to these other auto-inducers. So the way we really think this happens is as follows. When you're alone, you don't wanna make public goods. When you're at high cell density and it's your kin, your siblings, of course, you wanna do quorum sensing, everybody gets to share in these public goods. If Vibrio-Harvii is at high cell density but it's in a fierce competition with some other bacterium, you actually wanna keep track of that. And so what happens is that the non-cognate auto-inducers compete away the cognate auto-inducers and quorum sensing doesn't happen. So in fact, you are unwilling, Vibrio-Harvii is unwilling, even if that's at the right cell density to actually engage in quorum sensing if there's someone else in the environment that's making homocerein lactones. So we think then that this is about public goods and it's about not allowing itself to be cheated by competitors, even though they're gonna be Vibrios in the environment. So the receptor detects every one of these auto-inducers but they're antagonists. So that's how it works in a test tube, not a test tube, in a bacterium that has one receptor and we're adding auto-inducers, cognate or non-cognate, but we're adding them all with pipettes. And so what you'll remember from the very beginning of my talk and also from yesterday's talk is that in fact there's three systems that are in Vibrio-Harvii and they make endogenous auto-inducers at whatever level they make. And so what Jaubot wanted to do to sort of finish this up was to ask the question right now that we know that the non-cognate auto-inducers are antagonists for lux N. What we wanted to know is does that actually matter in a wild type cell that has all of these systems running? So he wanted to do an experiment to ask whether antagonizing one system made a dent in the context of a wild type bacteria. So what he's gonna do now is have wild type Vibrio-Harvii so it's gonna make all three auto-inducers and it has all three auto-inducer synthesis, all three receptors. And then he's gonna exagerously add a non-cognate auto-inducer. So here's the experiment. What he does is he grows Vibrio-Harvii to high cell density. So it's very bright, quorum sensing has happened. And then he dilutes it a thousand fold. So what happens on that dilution event is all the auto-inducers disappear and then he's gonna let Vibrio-Harvii grow and let auto-inducers accumulate again. And he's either gonna add solvents or nothing or he's gonna add our most potent inhibitor which was the C12, right? That was the longest molecule. And he's gonna measure bioluminescence as a readout of quorum sensing. Okay, so here's the control. Again, this is light production per cell and again this is on a log scale over time, right? So he dilutes the cell. So the way this works, this classic U-shaped curve is the hallmark of a quorum sensing behavior. So the way it works is he takes this overnight and he dilutes some. So right after dilution the cells are very bright because quorum sensing has been happening. The auto-inducers disappear upon dilution and so light production per cell declines dramatically, right? Because the auto-inducers have disappeared so the bacteria has stopped making luciferase, right? But on the x-axis, right, they're growing. So they're growing and so they're making their own auto-inducers which are building back up. Eventually at this time the auto-inducers hit and the cell density hits the critical amount that the bacteria turn on light again, right? So you get it, the auto-inducers disappear, light comes off, they make their own auto-inducers and it comes on again. So that is the classic wild type bioluminescence quorum sensing behavior. Okay, so that was this flask. So now what happens if this is going on but we add an inhibitor of lux N? So that's the red bar. So what you can see is that even though two systems are working perfectly fine, in fact, light production keeps falling 100 fold more, right? And then it comes on later. So what happens is that indeed if you simply inhibit that one system with a non-cognate auto-inducer, right? In fact, it does matter, right? And you can inhibit quorum sensing and then what's happening, right? Is that there's a fixed amount of this C12 added, right? But they're still making their own auto-inducer. So eventually they have grown to a higher cell density. They've made more auto-inducers and they must say, okay, there's actually a competitor here but now we've won, right? We have grown even more, right? And even though we're gonna lose some of these public goods because there's a competitor present, it's okay because now we have actually out-competed the competitor and so now we can turn on quorum sensing. So in fact, it's delayed but it does come on presumably when the bacteria detect that there's much more of the C4-homocerin lactone meaning self than there is of their kin, right? And so it does matter to simply knock down one system. It affects the entire cascade. Okay, so that is where, and then Jabo escaped and graduated, right? And so that is where we are and I'll tell you a new graduate student, Makayla Aikoff came to the lab and so we're getting closer, like the first part was all done with pipettes and fake molecules. This experiment was done with mostly endogenously made molecules and one fake one but what we'd really like to do is to ask, how does this play out, you know, in a minimal competition, right? With bacteria that really see each other and that are making their own auto-inducers in real time. And so what you'll remember, right, is that we have this guy, Debrille Fischeri and it's auto-inducer that turns on its quorum-sensing cascade is this C8 auto-inducer which is a potent inhibitor of Vibrio-Harvii, right? So now we have two bacteria that really live together in the ocean, right? That we could actually put them together and ask if we just let them grow and make their own auto-inducers, does this auto-inducer inhibit Vibrio-Harvii? And so here's the experiment that Makayla is doing right now. So what we would, and a prediction, which is a prediction, right, of what's gonna happen, right? So if we mix them together, right, and Vibrio-Fischeri is supposed to be purple, right? We would predict that because it makes that long auto-inducer, it will inhibit Vibrio-Harvii from making light. So these gray cells are dark. So first we can measure does or does not Vibrio-Harvii make light, we just measure that. And then what we can do is to plate the cells out, we can tell the difference between them and just count them and ask, how well does Vibrio-Harvii survive in this mixed culture? And then the trick that Makayla is doing is by making those two mutations at L166 and at H210, we can convert the Vibrio-Harvii auto-inducer receptor essentially into the Vibrio-Fischeri auto-inducer receptor. And so we can ask what happens to you if instead of being inhibited by your competitor, right, you actually use that auto-inducer to do quorum sensing. So what we would predict then is that Vibrio-Harvii will be bright because both auto-inducers are now auto-inducer, but it could pay a big penalty for responding to a non-cognate auto-inducer, right? And so we're gonna count whether or not Vibrio-Harvii gets out-competed here when its receptor is sloppy versus here when its receptor works perfectly fine. Okay, so the jury is out on that experiment. We haven't done it yet. She's building all those strains. And these are the kinds of things that we would like to do going forward to try to ask about whether or not, you know, in slightly more realistic environments, whether or not having a very perfect or a sloppy receptor actually matters. And remember, it's not just gonna be making light. It's gonna be making all these public goods and Vibrio-Fischeri would probably enjoy those, okay? So that is where we are on this second story. And there are only two heroes in today's story. So this is Zhaobo Ki. He was a graduate student in the lab. He's now a postdoc at Harvard. He's working on the microbiome. And this is Laura Miller. She is a fantastic chemist and she was a postdoc and she is a professor in San Francisco. And so one more time, thank you for listening to both those stories and I'm happy to take questions. Yeah, maybe I'll do this. I can hear you. Yeah, well it takes one to know one, so don't worry about that. There we go, that works. Really cool talk. So you showed basically that for Vibrio-Harvii, which is a small tail, that the bigger the tail gets, the more it is inhibited. Do you have any ideas for the reverse? So for a bacterium such as Fischeri, who usually does use a big greasy tail, are they inhibited by the small tails? And also how is that gonna play out in your potential competition experiments? Yeah, that's a great question. So Zhaobo did that experiment. So it turns out that, right? So how could, that's a great question. So how is Vibrio-Fischeri discriminating against, it has the L and the A, which accepts all the tails. So it turns out that Vibrio-Fischeri has an amino acid that Harvii doesn't have that discriminates against short. So there's a third, right? So they have a different problem. So long guys have a different, guys that use long auto induced, I actually cannot give two seminars today as in a row, that have long auto inducers have the opposite problem, which is how do you keep the short ones out? And so there's a third amino acid, right? That Fischeri has that actually makes it so the C4 doesn't elicit signaling, right? Yeah. This might have been considered already, but I wonder since these auto inducers and the receptors are so specific, did some people engineer some bacteria so that they could detect those auto inducers and being used as probes for the presence of these bacteria and also maybe as probiotics to produce a molecule specifically against these bacteria? Right. So the answer is sort of. And so the synthetic biologists have mind not this, there's other circuits that the receptors are more tractable, that there are cytoplasmic receptors that bind auto inducers or these. And so the synthetic biologists have been playing with these for a long time. And you can actually engineer in one receptor and one synthase and then get some bacterium that never does quorum sensing to actually in front of whatever promoter you want and you add an auto inducer and it will do a trick, right? And so I think that's a really profitable way and for sure people are using these as sensors, right? Where you engineer in either you use the one that comes with it or you can engineer in one of these systems with an output. And so those are being used a little bit and they work pretty, pretty well. The EC50s for all of these are sort of nanomolar. So you have to have a real concentration of if you want to use it to probe like some environment is that molecule there. If it's nanomolar, it works, right? And then this idea, of course, I think there's a great idea like could you later engineer microbiome or probiotic bacteria for plants or agriculture, right? To actually respond like reliably respond to these and then do beneficial jobs. I think that is what synthetic biologists would love to do with these circuits, right? And you see that you can mix and match them, right? You can have agonism, you can have antagonism, right? Like I think we know enough about these now that you can actually do some pretty interesting combos, right? Getting them to stay around in genomes if you really like put these out in the dirt or put them in, you know, that's harder. But I think that is the ideal and ideal, yeah. And they're really nice circuits, you know, because the molecules are cheap, you know, and there's this exquisite specificity at least in some of these receptors, yeah. How often do you think bacteria cheat and make receptors that make their neighbors make these public goods and just benefit from that? All the time. I assume they do it all the time. You know, I think like this eavesdropping and freeriding, it's gotta be big business, right, in bacteria. And so my view is the best work on that was done by Richard Novick in Staph aureus, right? And you can see, like the idea is that you're specific for one molecule or you're specific for another. And in order to go between those, there's a sloppy phase, you know, where you, like what we made, where you kind of respond to both. And you can, if you look through these genomes, you can see that happening, right? And so maybe they're cheaters for some billion years and then actually there's a benefit to becoming specific and so you do the second switch, right? But people think that might be how speciation happened, you know, where if I can talk to you, you're my family and if I can't, you know, and so, but you can see this running through these genomes, right, that some of these receptors are sloppy and some are specific. And then your real question about how that's really playing out in nature, like are the sloppy receptors actually just cheaters? You know, people need to really do those experiments, not in a test tube like this, but actually really ask how is that playing out? You know, ours is a fake experiment where we put these two bacteria, you know, but it would be fun to actually understand those things. But I assume they are trying their darndest and there's other mechanisms, you know, some bacteria eat the other guys auto-inducer, so one guy is trying to go one, two, three, four and then they like miscount. There's guys that make lactanases, right? So there's guys that talk with peptides, that's their auto-inducer, and they make lactanases that cut the rings and so their molecule is preserved, but the other guy is mute, right? And they all live together in the dirt and so people are starting to look, you know, in the natural world and think about like how these guys have been duking it out forever, right? And there's all kinds of shenanigans, you know, out there, right? And like I think that like that there's actually, it's actually, there's a lot of people that are trying to bring those into the lab now and actually use those as the basis of applications because they work out in the real world, right? And so I think that there's all kinds of, those kinds of fantastic, you know, manipulation strategies out there, right? And yeah, and this is one, this would be one of them where these receptors are a little bit sloppy, yeah. Thank you, thank you for a fabulous talk. So I have three quick small questions. So first is- That is probably an untruth. Go ahead, three quick small questions, what does that mean? Okay, go ahead. Yeah, so what is the half-life of these auto-inducers? And can we use this non-cognate auto-inducer to inhibit the quorum sensing dependent processes like biofilm formation? Can we use non-cognate auto-inducers to inhibit biofilm formation, say? And the third, how do they correlate with other second messengers? Say cyclic di-GMP or say PPGPP? Okay, well, I'm not answering PPG. In this audience, the PPGPP question, I'd be thrown out. Okay, let me see. So the half-life, I'm not gonna remember this, but, and of course, this is back to your question. I don't know what the half-life is in nature, right? They are very stable if you just let them sit around in a bottle of LB, right? So, and some are like the longer ones tend to be less stable than the shorter ones, but they're reasonable to work with. These are not delicate molecules, right, that you have to worry about, right? So they are reasonable to work with. Then your question, the answer is you should be able, if a non-cognate auto-inducer is an antagonist, you should be able to inhibit biofilm formation, right? As one of the readouts. Now, like in Pseudomonas, which is activated by that, I don't think, so they're not gonna be biofilm dispersal agents, right? It's not gonna be like the biofilm formed and I add this auto-inducer and then, but if you did it early enough, it might work. And one thing that we did was to dangle them off of plastics, right, you know, like could you make catheters that are resistant to, you know, infection, right? And that works, they were like this big, but in fact, people are starting to make products where you put non-cognate or inhibitors, anti-chorms as they molecules onto surfaces, and in fact, biofilms, at least in an academic lab, they don't form. So I think that idea of making these kinds of products or disinfectants or things, that's pretty good with these anti-chorms sensing molecules. And then the third question was, what? Oh, cyclic dyne DNP, yeah, but right, I'm trying to avoid. Right, so, quorum sensing controls, at least in Vibrios, quorum sensing controls, so oh, sorry, so cyclic di guanylate is this fantastic second messenger molecule that's a cytoplasmic molecule that actually controls biofilm formation and controls motility in lots of different bacteria. And at least in Vibrios, quorum sensing, so the outside auto-inducers, they go through the pathway I showed you and that controls the enzymes that make and break down cyclic di guanylate. So that is how biofilms are controlled, is that the enzymes that are required to make this, so this extracellular chemical information gets converted to intracellular chemical information in the form of cyclic di guanylate and that is what is impinging on biofilm formation. I don't know about PPGPP, right? Yeah, so, what? No, but if you know, if you added auto-inducer, does that change PPGPP levels? Inside, I don't know if anybody's ever actually looked at that, right? So I don't know, honestly, believe me or not, don't know the answer to that question, but this is about cyclic di guanylate, yeah. So you made the point yesterday and this morning or this afternoon at the beginning? No, it seems like 24, I know. Trust me, yeah. So that the cells want to have an idea about how many cousins are there, how many siblings and how many others are in there, but they all focus into the same kinase, phosphatase pathway. So are there other targets or do they hit other different sites on the same protein? Yeah, so they all, so they, there's one phosphorylation site, right? So I get your question, once the information is inside, how do you know which auto-inducer it came from? And so the way that you get some specificity, they're just signaling through one pathway, is that there's feedback loops. So when an auto-inducer engages with a particular receptor, the bacteria puts more of that receptor on the surface and so then it doesn't respond to the other two molecules, so what you do is you increase the dynamic range of responding to it, like they pay attention to one auto-inducer and they ignore the others, right? So part of it is they do it with feedback so that they can actually, it's not exactly linear the way the information comes in and then that new molecule that I showed you yesterday actually comes in and doesn't use the same pathway. And so you're right, there has to be, so the three I showed you, they all get integrated at that protein lux O and there's some tricks that allow them to pay attention and to ignore, right? But it's not perfect, it's sloppy because it all funnels to one protein but that new molecule does not use that circuit, right? And it actually controls targets, some that are the same as that and some that are separate, right? And so the question is whether or not those first three auto-inducers actually have other targets that we've just never found. We don't have any evidence for that but I'm wondering if we have to do that like in the microbiome or somewhere like that, right? Yeah, because otherwise you're right. It's like you ante in and how do you know whose chip it was, right? Yeah. All right, one more question. Hi, at the back. So I think the most exciting thing about the competition experiment is that you're using two vibrios. So you're gonna have two positive signals that say, yes, share goods in one signal that's gonna say, no. So what do you think's actually gonna happen? I don't know, well, okay, so prediction, right? You're right, so you get a question and they'll both be making auto-inducer two and they'll both be making the vibrio signal, they do, right? But what Xiaobo showed, right, is at least in the Jindup experiment if you simply inhibit luxe N, the whole system shuts down. So I think vibrio fissure I will be doing fine and I think vibrio harvii is gonna get inhibited, right? Based on the experiment that I showed you in the test tube, right? It doesn't matter that the other two systems are running. If you inhibit luxe N, right? All of quorum sensing is inhibited, at least delayed, right? And then I'm gonna bet when we make the sloppy receptor and it's given away, so we have three positive signals, then vibrio harvii is gonna lose in that one, right? Because it's just given it away, right? I mean, I imagine it probably will not work like that but that's my prediction, right? Yeah, we'll see, right? Let's thank Dr. Basel for an excellent seminar. Thank you for having me. Thank you.