 Can you get the volume control on this? Better now? No. Hi. Is this better? A little? And now? I can't hear, I can't hear any difference on my end. Yes, okay. Okay, well, I'll start by saying, well, extending a really enthusiastic thank you to the organizers for putting this meeting together. This has been a really wonderful experience, both to see colleagues that I hadn't seen for sometimes years and to meet new colleagues that I hope to see often in the future, so this has been fantastic. And also thank you for extending the invitation for me to prevent, present some of my work. Just to get us started, I'll note, I have a background in zoology and evolution, and I was brought into this enthusiasm by collective behavior, the evolution of cooperation and conflict, I think like many of us here. But really early on in my trajectory, I was generously introduced to microbial systems. And I remember clearly I saw something like this under the microscope and fell in love with these structures immediately. This is Vibrio cholera growing on a piece of glass. This is a maximum intensity production of a three-dimensional structure growing toward you. This is an example of what is often called biopharm formation, and it's very common among microbes. But up until about 10 years ago, we hadn't really, we hadn't started to experimentally assess what kinds of evolutionary ecological population dynamic processes occur in here from a social evolution perspective. So what I tried to do is combine that background with molecular genetic tools to see what's going on in here. And before I tell you about some of these experiments, I'll just give you a short background on what biopharm's are. And I should also say at the beginning, the term biopharm is a bit of a mess. It refers, it's an umbrella term. It's used a little bit differently by different people. So I like to explicitly say what I consider to be a biopharm, and that's a group of cells that's often on the surface, but not necessarily, and embedded in a secreted matrix of extracellular polymers. So these polymers are made by the bacteria themselves, and they have an enormous impact on the structure and ecology of these systems, as I'm going to tell you about. This is a typical canonical life cycle of biopharm that you'll often see in the literature. Cells first encounter a surface, and they can explore the surface with remarkable sophistication as we're learning now from lots of groups. And at some point, if the conditions are correct, if they deem them correct, for example, if nutrient concentration is sufficient, then they'll commit to more permanent attachments, they'll begin secreting extracellular material, and that binds cells together and allows them to gain protection from various threats and increased ability to exploit local resources. If these resources run out, then we will often see dispersal events, but this step is actually, we don't understand very well, but I'll talk about that later. I think many would argue that biopharm formation is a common, if not ubiquitous part of microbial life in the wild. Biopharm biologists often say that all bacteria live most of their lives this way. I'm not sure if that's really true, but there's certainly a lot of evidence that wild isolates of many species do this at least some of the time. So this is like an alternative or other feature of bacterial life that we need to consider in addition to their planktonic world, which we're more used to thinking about. From an applied perspective, biopharm's often cause infections that are very difficult to remove, and in fact, if they become mature enough, then they just have to be cut out surgically, which can be quite devastating for patients. They also grow in the interior of industrial pipes and probably pipes in your house, and they cause considerable damage economically. And so understanding how these work has practical implications for mitigating these harmful effects. But mostly, what I'm gonna talk about today is just the basic biology of these systems. And the perspective that I'm coming from, as I said, is a social evolution perspective. So we have cells immobilized on a surface like this, or immobilized in a flock that's floating around. And as I mentioned, the cells are bound together by a secreted matrix of various materials, which I'll talk about in a moment. But what's salient here is that the cells are immobilized in close contact with each other, and the spatial structure has an enormous impact on who can affect who, as Kareel noted at the beginning today. Couldn't agree more, and so I'm gonna talk today about how this structure arises. I'm not gonna go through this in detail because I think the audience is familiar with, from many wonderful talks here, about the different kinds of interactive behavior from consumption of material to secretion of material to direct contact. All of these can be either helpful or harmful to neighbors, depending on the details of what we're thinking about. But I want to pay attention in particular to the matrix material itself. This is its own class of interactive behavior. And I have a special affection for it because it's unique in defining for biofum growth. And it's also a little bit nebulous in its categorization with respect to these types of features. So it is a secretive product, but it's often a complex of many different things that some easily diffuse and some don't. Some of these things are taken back up by cells or by other species, and of course it mediates the proximity that's required for direct contact. So I think understanding the matrix was my first step to getting into biofum work and I'm still doing it. I think I probably always will be. It's very complicated, it turns out. Okay, so what does the matrix do? Very broadly it attaches cells to each other into surfaces that confer structural strength and resilience against exogenous stress, including antibiotic treatments, famously. It has other functions that we are going to discover. We don't know a lot about how the matrix is really organized at a very detailed level and it's a very active area of research in the biofilm biochemistry community and it's increasingly active in the ecology community too. But as I said, 10 years ago we didn't really. That is, that's, in principle, yeah. So some things that bacteria secrete, of course, are responded to by other bacteria. We get into, it's not clear if any of those molecules also serve as a structural element in matrices like this. So I think it's unclear. But they certainly might retain other signaling molecules which is quite important. The matrix concentrates solutes on the basis of their diffusivity, hydroviscity, electrostatic properties. Okay, but I'm interested here and I'm gonna go back, this is a little bit old, but I'm gonna go back to the earliest product that I did on this just to give you a sense of the trajectory of the thinking. And my questions are how does the matrix influence population dynamics, ecology, and evolution in biophones? How can we join these perspectives? And the first question I asked here was the really basic one. Does this material serve as a public good? Is it what happens when you mix cells that can produce it in cells that can't? And to study this question, I'm going to use the model system, Vibrio cholera, which is famous for causing pandemic cholera in humans. Biophone formation is really important for this pathogen both inside the host, where it helps them survive passage through the gut and is also presumably putatively involved in pathogenesis directly. It's also involved in escape from the host. This material is, these biophones chunks are basically better able to survive passage into the environment. And then once cholera is in a marine environment, it uses biophone formation to attach to shed chitin, which it eats. So the matrix is involved in all elements of its life. So the way I'm going to approach this question is I've made mutants that either produce matrix constitutively or not at all. I'm not gonna really go into the details of these deletions, but suffice it to say that I've basically made slightly cartoonish versions of the real bacteria in order to isolate this behavior and assess its competitive properties. I've also made them emotile, that's the flaw A mutation. And you'll see that the only difference between them is this VPSL deletion, which is what causes this strain to not make matrix and also gains a big growth rate advantage as a result. It's very expensive to make this material. So I'm gonna grow these two strains together in a microfluidic device, which is a little tunnel where we can flow liquid through. And for comparison, I'm also gonna grow them in liquid environments and we'll see what happens in these two types of setups. Okay, so if we look at bio-information of these two strains on their own, we see that the EPS plus, so EPS stands for extracellular polymeric substances. It's kind of a moniker, it's synonymous with matrix production, so I'll use them together. This strain makes more buy from than the other, which is not shocking, this is well known. So my thoughts, by default, was that maybe this strain that does not produce the material would get pulled up by the other when they're in co-culture, but it was actually the other way around. And we had an indication of a competitive or suppressive interaction from one strain to another. And to get a full, to see whether or not this is frequency dependent, we do this experiment for many different starting conditions of the population, so we have an initial abundance of the EPS producing strain and then a final abundance that this is after, I think, 48 hours. But the time of the experiment affects the magnitude of the results, not the sign. So we see that the secreting strain always wins no matter where it starts. So there's no frequency dependence here, it simply wins. If we do the same experiment in liquid, then we get the opposite result. The other strain wins, and this is consistent with the fact that it has a much higher growth rate. And it also says that the competitive advantage of the secreting strain has something to do with space in particular, of the biofilm environment. So to get a feeling for this, and a lot of the data I'm gonna show you are image heavy, because this is how we, this is how we get a feeling for what's happening in these systems, there's a lot of microscopy. All of that, basically all of the data except for liquid controls are microscopy. Raghu gave a more thorough explanation of this yesterday in his wonderful talk, but I'm going to just tell you really quickly again for a refresher how we image these things. So we use something called a confocal microscope that's able to isolate optical sections of this sample. So it takes light only from a chosen focal plane, and by taking many of these samples successively, we can build a three-dimensional picture of the system. And so this is what it looks like. This is the progress of competition between the two, where the EPS secreting strain starts at a very low abundance. So clearly there's a spatial competition between them along the substratum. There's no motility here, remember? So there's nothing happening in the planktonic phase except for removal of cells that fall off. And I'm not showing you here, but the clusters of red cells are displacing the blue ones from the glass. They're not just growing over them. Okay, so this was cool. It's surprising on intuition, but also consistent with a really great and interesting prediction made by Kevin Foster and Joel Xavier in a 2007 PNAS paper, which has become something of a landmark in the field. So I'd encourage you to look at that if you're interested. I want to show you this just to show you how far we've come and our ability to, yes. We're not, it's difficult to say. I think like someone with the right tools could see, could tell to what extent they're really pushing. But I think the best like null expectation is that the blue cells fall off more easily and whenever they fall off, on average they're just replaced by neighboring red cells if they're there. And so that's my guess. But in my like heart of hearts I think they're also pushing a little bit. Yeah. Okay, so this is to show you how far we've come and our ability to resolve what is happening in these structures so the microscopes improved, the fluorescent reporter constructs have improved, the image processing has improved, everything has improved. So this is only over about seven years and this is work that was innovated by Canute Dresher while he was at Princeton and has now developed further in his group at MPI Marburg with an extraordinarily talented image processing postdoc named Raimo Hartman. So I just wanted to show you this, but I also, there's something in here that I think is intriguing, it leads to the next question I have. So if we look, I've been really interesting, I've been stressing here competition along the surface, but in fact, we have to also think about the interface with the planktonic phase. There's an entire volume of these red clusters and we're not really, at this point we're not inspecting what's happening on the outer surface. And I want to add at this point that our interpretation is that the red cells are basically, this trade is cooperative, but it's limited in the extent to where the benefit goes. So the cells that are producing the material seem to retain its benefits more than they distribute it elsewhere. So the scales of competition and cooperation are a common and important theme. Okay, so I've shown you the matrix producing cells that have a strong competitive advantage just by a growth disadvantage or growth rate disadvantage. But what about incoming cells that are introduced after a biofilm or has already grown, right? Presumably cells that encounter in new environments don't always just have a clean slate to occupy, there's something there already and that's the question I'm gonna ask here. Okay, so this is a repetition. I'm asking what determines group entry and group departure? I'm gonna focus on entry though. Okay, so a quick rundown of what this experiment looks like. I'm going, this is the vocabulary I'm going to use. So I have a resident strain. This is the biofilm that was grown in advance and an invading strain which is introduced after the fact. I grow these resident biofilms and then I introduce the invading strain later and then image and then we see how the system evolves, or excuse me, how it develops what the population dynamics are after this. There's no exploration of evolutionary change at all to this point. We study this process for different types of planktonic strains that vary in motility and matrix production just for completeness but I'm only gonna show you the ones that are both motile and able to produce matrix, able to produce biofilm. That's the most realistic case and it's the most interesting. So I'm gonna focus on that one but the other cases are the same. So this is a controlled experiments. The invading strain is introduced by itself and then allowed to grow over the course of one day. It grows fine. It's a biofilm producing strain of cholera. This is what happens if there's a resident strain in the present already. So it's quite different. The ability to colonize the surface, the ability to colonize overall is reduced. Colonization is limited to where the residence is not located and the subsequent growth of this strain is actually reduced, there are fewer cells here than there were at the beginning. So there's something about the presence of an existing biofilm that discourages attachment of new ones and this is true even if the strain that's introduced is the same as the one that has grown there in the first place. So this is an important control. If you inoculate these two together, then they coexist. This is a time dependent thing. It's who's there first and who's there second that determines the outcome of this experiment. So this is in a molecular biology context. We need to, we want to see if we can break this phenotype, the invasion resistance and we naturally look to the biofilm matrix itself because this controls the structure of the biofilm residence population. And the image on the left, which I showed earlier but didn't explain, is some really wonderful work from the labs of Fenad Yildiz and Stephen Chu in California that was incredibly innovative and beautiful localization of different components of the matrix of biofilm. So the cells here are in blue. There are three different components shown. Green, gray, red. They're all interesting but I'm going to show you just one because it's the most interesting particularly with respect to this invasion resistance phenotype. That protein's called RBMA and it causes cells to stick to each other and also a bit to the surface. The gene name's not terrifically important but we need to use something just to refer to it. So this experiment is essentially the same as what I showed already but instead of just imaging the day of colonization and the day after we do a long time series and see what happens for many days after we try and invade a biofilm. This is a recapitulation of what I showed earlier. This is the invading strain does well if there's no one present. It colonizes well and it grows well. If the parental strain, this is a technical description of the strain's phenotype but it's the same as what we were looking at in the first project I showed you. If this strain is present already then the invading strain is incapacitated and it's ability to grow, excuse me, to colonize and to grow. Although I should say after talking with people here my interpretation of like elimination versus low abundance is different. So they're not gone. They're just not able to compete nearly as well as if they're introduced at the same time as the resident. If we use this matrix mutant, the result is dramatically different. So first of all the invading strain colonizes somewhat more and it's able to grow in order of magnitude more than with the wild type presence. Some may be interested that in this case the RBMA resident is displaced itself. It's outcompeted. I'm not showing that here but it is. And also varied I think most interestingly that the invading strain is able to percolate through the entire system this now which is dramatically different from what we saw before. This is dependent on motility. So they're not just well, I'll leave it at that. It sells without a flagellum, can't get in. Okay, so as a social evolutionist I'm now interested in whether or not this ability to resist invasion can be given from one strain to another. This also motivates different interpretations of how the matrix is structured. So what I've done here is mixed the two, the mutants and the parental resident together and thrown the invading strain on. And what we can see is that the invading strain can only enter where the mutant is. And we can repeat this just to get a fuller picture of what's happening. We can make the resident biophome with different combinations, different frequencies of the mutants and the parental strain and the ability of the invading one to grow to accumulate biomass is roughly a linear function of how much of the mutant was present in the first place. So this tells us that the invading strain simply occupies what volume was available by the mutant strain. So this was exciting to me because it's an unusual case I think of an ecologically or motivated experiment giving us an idea of how the matrix is structured. So this predicts that the RBMA protein which is important for invasion resistance is not shared from cells that are making it to cells that are not. And we can test that idea directly using some molecular genetic tricks. So we go on to the chromosome of cholera. We introduce an epitope at the end of this gene and this way when it's expressed it has a little tag on it and we can introduce an antibody that's specific to that tag with a fluorescent molecule conjugated to it and then localize where that protein component is in the biophomes. So we do that and sure enough here's our producing strain mutant and here's where the RBMA protein is. It only localizes to the producing cells as we predict. So just for, this was just for kicks basically but I was also curious whether or not this protein could be shared from cells that make it to other cells that make it and we can test that idea as well by mixing this strain which produces the tagged version of the protein with a wild type strain that makes an untagged version. And actually we see the same thing which I found to be very surprising and this indicates that this RBMA protein really localizes to the cell lineages that are producing it and it defines cell lineages as they expand from a surface during biophome growth which I think has important implications for understanding how spatial structure arises in color in particular but I think it's also it's important to explore for this effect in other species too. And I'm gonna show this, I have the same picture that Carill did but we both love this I'm sure. This is the fish experiments by the Borese group and if we look close up at two strains here we can see well-defined boundaries between different clusters of two different strains and this repeats throughout their samples so this is kind of consistent with the fact that with our prediction that the clonal bunches of biophomes tend to, there's indentation on this scale but not on the scale of 5, 10, 15 cellings and if we look at a larger scale however we do see on this scale extensive mixing between different species and this is a biophome grown on a person's tooth, this is a real in situ biophome. So there's many cases right here is a good one where a cluster of this blue strain is clearly not permitting entry of the orange cells around it but there are other areas where there is more mixture and so all I mean to say by this is that the results I'm conveying here are not meant to convince you that I think that all biophomes are monospecies that's empirically false so that's not defensible but rather that in order to understand how this structure arises we need to have a bit of mechanistic understanding about how biophomes are assembled and how they work, how the matrix works in particular. Okay, so what I've talked about so far is how the cholera matrix orient cells so as to discourage entry of new cells into the community. So I've just described now how it's two different ways that the matrix can give it a competitive advantage to the cells that are making it and there are many more so we've done other studies indicating that this is with my colleague Knutresher which indicate that producing a lot of matrix and forming thick biophomes allow you to also stabilize other cooperative secretion phenotypes like chitinase secretion which digests polymers outside of cholera in marine environments. So there are a lot of reasons to make matrix but among wild isolates there are many strains that don't so why is that? So I'm gonna talk a little bit about why this might be the case and there's some intuitive reasons why we can think about trade-offs for example and something I didn't tell you about my first project is that cells that produce matrix like this are great at competing for space on surfaces but they're terrible at leaving. They're unable to disperse very well whereas the blue strain, the non-matrix secreting strain whenever a cell division event occurs here more likely than not that cell is released into the outflow and so they're excellent at dispersal. So this trade-off between local competition and dispersal can obviously the balance of this trade-off or the balance of patch residence time and meta-population properties can select for strains that either don't make matrix or make it very rarely and they will be just fine and in fact while type cholera isn't actually either of these things. It switches between matrix production and non-production and it does that in a way so as to get as much of the benefits of production for local competition as possible while maintaining the ability to leave. So this is not like the entire picture but it does help us to understand the core ecological factors and the core selective forces that govern why bacteria do this, why they make matrix and why they don't and when they will and when they won't. Good on time. So I think I'm gonna show you one more project that has to do with the same question but the purpose here is to show you that the complex interactions between environmental conditions and biofemme architecture are both difficult to anticipate and they can have a very strong influence on population dynamics. So for this bit of the talk I'm gonna tell you about a different model organism. This is Pseudomonas originosa which we've heard about from several great talks already. On the left is Pseudomonas growing in a microfluidic device like the one I've already talked to you about and its biofilms are different to be sure but have a related structure to what we see in cholera with these mushroom tower type of things. If you put Pseudomonas into a more complicated flow environment, this serpentine type of chamber that we have on the right, then we see something quite different. So first of all, cells occupy the side surfaces of this environment, so the walls of this serpentine chamber but not only do they grow biofilms there but they kind of exude matrix material into the liquid phase above them and this has to do with the way flow around these corners generates forces on these corners as it does so. But what's really interesting, what's extra interesting about this is that those streamers as Knud calls them in this really great paper that he described these systems in, these streamers will catch debris that's flowing through the system and that includes other cells. So the red cells here are introduced after the green one has already been allowed to produce biofilm inside this chamber. And Knud and Howard Stone and Bunny were really interested in how this streamer phenomenon can clog chambers that has importance, for example, industrial implications. But I was interested in whether or not this would change our competition results because if there's a structure in the liquid, in the liquid catching cells that fall off easily then maybe our competition results with respect to matrix accretion is just not, it's not gonna hold here. So this is how I'm going to ask, this is how I'm gonna study that question. So we have simple flow environments with just a plain piece of glass with flow wafting over it and we have more complex flow environments. Instead of using a chamber like this, I'm going to use one that has intermittent column obstacles and this is meant to represent a cross-section of a packed soil, for example, an idealized packed soil. And the key difference here is that flow is relatively simple here but it's complicated here. It has to wind in between these columns and it creates the forces that cause these streamers to depart from biophones as they grow. So a lot of this work was done by a really enterprising undergrad that I had advised that when I was still a Princeton, Deirdre, she did fantastic work. And what we did here is a similar approach to the beginning, we have a wild type strain that naturally makes matrix and also devotes energy to growth. And we have a matrix mutants. This is a PEL mutant in a PA-14 background, just for those interested. PEL is the primary polysaccharide of its biophil matrix, they don't make PSL. So we have a PEL mutant that's otherwise the same as wild type, but it can't make biophones very well. Okay, and what I'm going to do here is show you how the system changes. So with the change in wild type frequency as a function of its initial frequency, it's the same kind of plot that I showed at the beginning, slightly different format for clarity here. And I'm going to do this experiment in two different chambers, the simple one, and what we see here is the same results as we got in the Vibrio cholera case, the matrix secreting wild type strain outcompetes the other. But in a more complicated flow environment, we see a beautiful shift to negative frequency dependence selection. So now we have coexistence of the two, which was consistent with what I was thinking might happen. So I was very proud. I thought, okay, we guessed this correctly, that's great. But if we look in detail at these chambers, we can find those streamer structures and there are no other, there's no red cells in them. So that's not actually the explanation for what's happening. And this was a little depressing because we really thought we got it, but it took a while to kind of tease apart another explanation for this. But what we noticed was that if we looked closely here and we could see that many regions between these columns are actually completely blocked by biofilm growth. And this was related to and consistent with what others had found about the consequences of streamers in biofilms. And so we guess that maybe because flow is blocked, there might be regions of these chambers where a matrix deficient strain which is susceptible to shear, removal by shear, could be happily growing because they no longer have to worry about being removed. And we can test that idea by performing this experiment for the same thing where we inoculate the two strains together. And then after imaging them, we can flow fluorescent beads to the system. And this is a trick that fluid dynamics researchers use to understand where flow is going in a system and where it's not. So when we do this, we see good evidence for this interpretation that places where the Pell A, the matrix mutant strain is accumulating are the places where flow has been blocked. And this we can assess it's messy, like these are noisy experiments, but over the entire chamber, this has a strong statistical signature. So this is a great example, I think, of how relatively unintuitive feedbacks between environmental flow conditions and matrix and biofilm architecture can have important consequences for the population dynamics of matrix accretion, and I'm guessing many other traits of interest. But this is, bacteria often grow in environments like this, so we have to think about it. This is the message here. Okay, so just to recap, I've discussed how matrix accretion exerts a major influence on population dynamics in biofilms. I would speculate that it's important for ecology and evolution in biofilms very broadly. And the interaction between and the environment and biofilm structure can be very, they're very difficult to anticipate, so we just have to do experiments and see what happens. I have the right amount of time, so I wanna tell you about one more project. I'm interested in general in how the biofilm growth process alters fundamental ecological properties of bacterial systems. I've done this with respect to the invasion of colonizing cells into an existing biofilm population, but as a segue into Bruce's talk later, I want to also think about other important features of bacterial life, and one of the most important is exposure to bacteriophages. This is a ubiquitous threat to bacteria in the natural environment. We know there's decades and decades of very beautiful work on the ecology and evolution of these systems, but we're just starting to learn now about how biofilm production influences this interaction. And I would argue that in order to get a full picture, we have to look really closely at infection dynamics on a single cell level. Okay, so I'm sure this is probably familiar, but just as a refresher, this is a typical litic phage life cycle. You have a cell on the left, everything's fine. A bacteriophage attaches to it, it injects its genome. It quickly hijacks the hosts into a cytoplasmic machinery to make more copies of itself. Eventually it degrades the host cell wall, the host lices and these new phages pop out and they're ready to infect new hosts. So this can be devastating, this is often devastating for natural bacterial populations. Okay, so what I want to do is track this process under the microscope. And there are different ways to achieve this, but what I thought we could do is to manipulate the phage genome such that when it's introduced to the host, not only is the phage proceeding with its infection, but we'll get a signature of the infection. The host changes color, so now we can see which hosts are infected and which aren't and we get a little window basically before they die. And importantly, the reason why I wanted to do this instead of just tagging the phages directly is because this method allows the next generation of phages to also make their host fluorescent. So we get this tool that's maintained throughout the duration of an infection. So these are just some pictures to show you that the system works. From left to right we have a small group of cells growing happily, there's a division event, there's actually another one here and we get a strong signal of phage infection. These dotted outlines are added after the fact to show you where a cell used to be before it laced. And after cells laced, of course, all the fluorescent protein just quickly diffuses away and we get no more signal. Okay, so now we wanna use this tool to get an idea for how phage infection can proceed in biofilm populations. And the way that I do this experiment is pretty simple. So we grow a biofilm in a microfluidic device. This is E. coli, I forgot to say the last slide. This is E. coli biofilms and phage T7. The reason we're able to do this is because T7 is so easily genetically manipulable. E. coli is nice too. Okay, so then we introduce phages. We give them a pulse. We also do experiments where we throw them in continuously. We look under the microscope and we just see what happens. Okay, so here we have a biofilm. Phages are introduced and sure enough, the population is obliterated basically by phage infection. So at this point, I was really, I was quite thrilled that this system was working. This took a lot of time to optimize the reporter system and the microscopy and every little feature of this that takes, it can be quite picky. So that was amazing, but this is actually, I was kind of frightened at this point because this is what happens in a liquid culture. Most of the cells die, some are left over, some might be resistant. There's probably, you can see cells rounded up here. These have had their cell walls digested but they haven't laced yet. There may be a resistant cell up here. On a large enough spatial scale, we would certainly see resistant cells the same way that you get in a planktonic population when you do these experiments. But I was a little worried we wouldn't get anything, like anything else. And so we thought, okay, well let's try and push this a little bit further. And of course we did in the end, but we had to grow biophones for a long time. So what I'm showing you here are plots of biofilm population size normalized to the initial point at which phages were introduced. And we grow biofilms for different lengths of time before we introduce the phages. So in this case we've grown for 12 hours, phages introduced, biofilm dies. Phages introduced after 24 hours, biofilm dies. And so on and so on. Okay, so we were concerned, but if we let these biofilms grow for long enough, then we see something else. So the biofilms no longer die when we add phages. And this stays the same as long as if you keep going. So there's an abrupt transition here before which biofilms are killed by phage infection the same way we would expect in a liquid culture of susceptible hosts. And after which there is no killing. And it's not because the cells are not susceptible anymore. They haven't been exposed to phages before this. So there was no selection for resistance. Also if you take these biofilms and disrupt them back into a liquid culture, you can infect them again. So there's something specific to the biofilm architecture that causes this. Here's just some time series for you. So this is exactly what I've just described. You can notice a big difference between this purple and green case is the size of the biofilm at the beginning. So there's a transition in the size and we're also interested in other kinds of architectural transitions. And naturally we look again to the biofilm matrix which is different in this case than it was for cholera. Why is this one growing? So this one has a short period of growth before the fate is to kill it, right? So why is this slope steeper? Yeah, well I mean, so that's a good question. I think that our interpretation is that when biofilms get this big and you're growing them in nutrient poor media, the nutrients diffusing into the population are depleted before they reach the rest of the entire population. So it's actually not the entire population growing. And so if we normalize to the initial population size, then I think that's responsible for this effect. Exactly, yeah. We introduced them in the same medium that the bacteria are growing in. Yeah, these are subject to flow. Well, these are growing. These are growing and so are these. Like these populations are expanding. So these cells on the outer edges are certainly growing without question, which you can see here. Like these populations are continuing to grow. Right, yes. So this is a cartoon picture of the E. coli biofilm matrix that has a bunch of different pieces. In line with the experiments that I showed you before, we're interested in how this matrix architecture controls this result that we found. And so we systematically deleted these things and saw what happens. That's the kind of bread and butter approach. So now I'm going to show you plots stacked up on top of each other because all of these biofilms were grown for 72 hours. Then phages were added. And it's the same type of plot format that I showed you in the slide before. So wild type survives phage exposure. So do biofilms lacking flagella. So do biofilms lacking cellulose. But not biofilms lacking these things called curly fibers, which are these amyloid structures that poke out of the cells and entangle them, and especially in the top most layers of the biofilm. This is great work by Regina Henges group in Berlin to show how curly localizes in biofilms. And we can use other molecular genetic tricks to plots or to track the production of this material as a function of time in biofilms. So this is a time series in which the entire biofilm at each time point has been collapsed into one vertical line in which we plot the intensity of curly production as a function of height. And looking at this figure in comparison with the one from the slide before, we can see that this matrix protein is produced right in the correct time for this transition from susceptibility to tolerance to occur. Yeah, that's a great question. Yes, they do. Yeah, so there's sporadic infection along the outer periphery, but it just, it never progresses into the middle. And eventually, you just don't detect it very much anymore. It's like the, I'm gonna, I think you'll get in, I'll touch this in the next slide, but yes, that's a great question, thank you. Okay, so this is the last experimental slide I have. We speculate on the basis of this result that curly fibers would change the permeability of biofilms to phages. The phages can't get in and they can't cause the infection. And we can test this directly by this time staining phages so we can localize them by microscopy. And then looking to see where they diffuse to in biofilms. So on the left, we have a wild type. And this is a slightly confusing picture, but this is an intensity projection in Z and X. So we're just looking sideways at a biofilm that grows up. You're looking straight into the side and you're looking kind of through the whole thing. And we do that so that you can see just how many phages there are along the whole surface. And there's a lot. So these biofilms are growing, but they're covered in phages that are not infecting the cells inside. In biofilm lacking curly, sure enough the phages are able to diffuse in. And this was a really ambitious effort from Lucia, who is a fantastic and really hardworking student. She wanted to see if we could recapitulate this result in vitro. So there's a lot of stuff going on in these biofilms. There's lots of cells and the cells are heterogeneous in their surface structure. There's other things that we can't completely account for. So her thought was that maybe we can purify this matrix components and add it to beads, a proxy for cells, and then see if this thing is recapitulated. You can do this. Okay, so luckily if you purify CSGA, which is the monomer components of curly, it will spontaneously polymerize in media. So that's what this orange stuff is, that those are curly fibers in the absence of cells and the blue things are phages. So clearly curly by itself doesn't prevent phages from entering like a cluster. And if we take those beads that I mentioned, the beads don't prevent phages from entering either. The phages are able to diffuse through the pores between packed beads. I think this is really awesome. If you put these two together, then the curly fibers spontaneously form clumps with the beads and these curly bead clumps are not permissive to phage diffusion. So we think that there's more to interpret this fully and this paper is in revision now. There are many, many more experiments to support this interpretation, but I just want to show you this to give you a sense of how much we can do by thinking about matrix material and how it can combine with other objects and even in a simple way to have significant impacts on biofilm ecology. So this is gonna be, the ecological implications of this are going to be an important part of my lab moving forward. I've also worked with Vani Bucci at UMass Dartmouth to develop a simulation framework that's specific to the system. So we'll be developing an interaction between our experimental and our computational frameworks for this problem, but I'll leave it at that. And I want to say thank you to all of my very cherished colleagues at Princeton who have had the great pleasure to work with for many years. And my colleagues in Germany where I've had a fantastic time working for the past few years, Canoe Treasurer is a close friend and the group leader at the Max Planck. Lucia, Praveen and Raimo were all instrumental for the phage projects. Lucia for all of the experiments I've executed pretty much, Praveen is a genius geneticist and helped translate the phage visualization idea into a real, into reality. And Raimo is an image processing specialist, as I said. So thank you for your time. Thanks again to the organizers. Thank you.