 I tried to keep the time, and all right, that's Sean. Great, hey everybody, my name's Sean, and I'm coming from Tokyo, and it's really, been a really fantastic week for me to be here. I wasn't able to join the first week, but I'm excited to watch the videos, and I'm primarily an experimentalist, and so I have a few goals in mind for today's talk, and one goal is to introduce an experimental technique that my group and other groups use called nanosims, nano-secondary ion mass spectrometry, introduce that as a tool to probe a number of different questions, but in this talk, we're gonna be probing phenotypic heterogeneity within cultures. I wanna introduce this as a tool, and I want especially the introduction of the tool part of my time today to be really discussion based, because this is a tool that's not necessarily that hard to use, but riding a bike is also not that hard to do once you know how to do it, and so if you don't know how to use the nanosims, at first I think it's kind of like, how do you do this? So I wanna have lots of questions about that, especially if people wanna use this tool and develop this as a kind of a community resource that we can apply. Then I wanna introduce some data that we've acquired using this tool, and show you what these spreads of data look like when we look at isotope incorporation. This is an anabolic activity of the cell. We're looking at the accumulation of material into cells, tracing that by using stable isotopes. What does that distribution look like and what types of parameters change the shape of that distribution? And it's during this part of the talk where I want to start asking you questions, and so part of my goal to come to this meeting is to present some things that I don't know how to explain, and as an experimental scientist, this is the, when I can obtain data that I don't know how to explain, this is the really fun time to interact with modelers and theoreticians, so that's part of my goal as well. Those are the main goals I have in mind. Great, and I can also just briefly introduce my institute. I'm from the Earth Life Science Institute at Tokyo Tech. We're an institute that has planetary scientists and geologists and a little bit of chemistry, also a little bit of microbiology in there. So we're trying to find out how life started and how it might start on other planets and how many might find it if it started on other planets as well. Okay, great. So this is a really, really basic introduction slide, and I don't need it for this room. We know that if we have an isogenic culture, we know that the cells can behave very, very differently within the flask or whatever medium that is. And so how can we study this variability? There's been a few different talks this week where people have looked at this variability, and I'm going to introduce and focus on this nanosynth technique, but I think one of the fun things for us to think about doing and to learn how to do is to work across our kind of instrumental toolboxes. So I don't know how to run a mother machine, but I know how to run the nanosynths. And so this is maybe another goal that I have in mind about being here is to try to teach each other different tools so that we can really acquire multidimensional data sets that can allow us to understand cells and cultures in higher levels of detail. So nanosynths experiment looks something like this. We take some stable isotopes. So first of all, let's take a really big step back. So take carbon. 99% of the carbon on planet Earth about is an atomic mass of 12. There's 1% of the carbon that has an atomic mass of 13. It's stable, never changes. There's a little tiny, tiny bit of deuterium in the water that I have waiting for me on my desk chair over there. Little tiny bit of deuterium, much less natural abundance than 1%. Look at nitrogen, natural abundance of nitrogen on planet Earth, take 15 over 14. You get a ratio of 0.0036 about. This has to do with how the planet has formed. It's different if you go to Mars, different if you go to Venus. But on Earth we have these stable isotope values. And through the tools of chemistry, we can purify isotopes so we can now go on Sigma and buy 100% labeled 13C glucose, or 100% labeled 13C acetate, or... You can't get it for everything, but you can get it for a lot of compounds that you might be interested in, like N2 or ammonium. So nanosynths experiment is the type of experiment where you take cells that are natural abundance and there are going to be small variations in the cells and between cells that are occurring in this kind of natural abundance range. This is going to be due to kinetic isotope fractionation that happens in metabolism and also it's gonna be due to equilibrium isotope effects. But for nanosynths experiment, we're just gonna ignore these really small variations and we're gonna focus on really large variations. The precision of the nanosynth instrument, it's about 10 arcs per thousand. It's not, we can't see the kinetic isotope effects with the nanosynths and we can't really see equilibrium isotope effects. But if I take a 10% labeled strength of ammonium and I put that into a culture of cells where the natural abundance 15 over 14 is around 0.0036, we can certainly see that with the nanosynth. So we can see like atom percent enrichments with the nanosynths. We can't see these really subtle variations that geochemists study. It'd be great to do that with the nanosynths, but we just don't get the secondary ion counts that would be sufficient. We don't have the counting statistics to have a really, really high level of precision. It's a bit of an aside. Type of experiment we do, we dump in a lot of stable isotope. This is the type of amount of isotope that for us as biologists, we're fine. I've talked to people who have told me they would feel comfortable eating a 13C sandwich. The kinetic isotope effect for 13C is not so big. Deuterium is, right? We saw this interesting paper that Eric showed yesterday. Deuterium will change you. 13C, I don't know anybody who has actually eaten that sandwich, but it's a lot more modest of an effect. If you walk down the hall in my institute, there's a stable isotope geochemistry lab, and if I tell them I'm gonna use a 1% label, they tell me don't come into my laboratory for a week because we're worried that it's the winter time, I'm starting to grow out a beard, and they're worried that a little bit of acetate is stuck on me, and it's gonna get in their lab, and it's gonna get into the instrument, and their instrument is gonna record the enriched levels of 13C for history. So that's the level that they operate, and the 9SMS were really operating at this really, really wild, enriched area of isotopes. Dump in a lot of isotopes. It's typical for us to use 10% label strength, 20% label strength, and then the other thing we have to think about is how long we're gonna incubate to be able to see a signal. Typically when we design experiments, we incubate for a third or a half of a doubling time. So we want to try to get a picture, kind of an instantaneous picture, if you will, about how fast cells are eating and accumulating these isotopes. And so depending on the questions that you have, you might incubate some deuterium and some ammonium and maybe some CO2 of some autotroph, et cetera. And then the nanosims, we'll have, I'll show a series of pictures to try to help us understand how the instrument actually works physically. But let's see this, this works, I heard it had a leg. It does. So the way that the instrument works is there's a small cylinder and it's filled with cesium carbonate. And around the cylinder of cesium carbonate is a tungsten filament and that tungsten filament heats up, puts a lot of energy into this little cylinder of cesium carbonate. And the cesium carbonate cylinder has a little pinhole on it. And so cesium ions just kind of come spewing out of this little tiny cylinder. Those are positively charged. You've got a cesium carbonate salt, you're heating it up, you're supplying all this kinetic energy. Now ions are just starting to spew out of it. Just do that by heat. And as these ions are spewing out, you have, you put a voltage right near the source. So you've got positive ions coming out, you put positive potential there and you shoot these cesium ions away from the ion source. Then the nano sims has a whole bunch of ion optics, all of these plates and little tiny quadruples that are focusing these cesium ions to smaller and smaller and smaller and smaller and smaller beam size up until the point where you get a beam size of about 70 to 80 nanometers, 50 nanometers is pretty achievable. You can go smaller, you can go to 10 nanometers but you pay a price for that and the price that you pay is the amount of secondary ion that you get out. So you can see more, you can see it higher resolution laterally but the higher and higher you see in your lateral resolution, the less and less secondary ion signal you get out. So there's a trade off here. In practice, yeah, 80, 70 nanometers spatial resolution, you get good secondary ion counts out. What do I mean by good? I mean at the label strengths that we're operating with, the precision of the instrument, it amply covers the ion count. So we don't have to worry so much about the amount of air, the amount of signal that we can add with the stable isotope is so big, it's not like we're measuring the light basically. Yeah. Yeah. Yeah. Right. Yeah. So then so you take this yellow orange arrow and you raster that over your sample and you dwell for a period of time at every point and you count ions. You can count, it depends on the Sims instrument but on the newer ones you can count seven different ions at a time. It's a, what's it called? There's a little cup. It's a, got it in a slide over here. What kind of detectors are these? They're ferritic cups, if I remember right. This is blurry, but if we go, so yeah, this is blur, what's going on here? We've got a, here's where our little, up on the top, there's a primary ion beam source. That's this cesium carbonate. It comes shooting out, it takes the corner to the left and you've got a sample. Here's this little orange noodle on the left, which is the cell and then this, that starts fragmenting the cell or anything that's down there. Yeah, yeah, secondary ions come out and then in the bottom right hand corner of this picture, this kind of blue zone where there's, these air lines taking a corner, there's a big magnet and it's like a meter long and so remember your Lorentz force, you've got ions, you're passing ions across a magnet and they take a radius, they take a corner and because the ions are different masses, they take a different radius. So the small ions take the corner tight and the heavy ions take it wide and then you have these little verity cups and you're counting current. So yeah, you measure current, it's a mass spec, yeah. So you go, you've got this primary ion beam and it's just rastering over the sample and then you're recording seven different masses as you raster around and that allows you to generate a picture and it's a picture made out of ions. Yeah, thanks for asking that. It's, let me go, so I'm gonna show this picture in a few different ways to try to help us, everybody build a picture in their mind. Primary ions come in, the place that the primary ions come from, it's got positive charge and it's forcing these positive ions away and the sample itself, here's a little patch of cells that might be a sample that's negatively charged and so the sample itself is sucking in these, sorry, it's sucking in the positive ions and they're stuck there and they're stuck there up until the point that you put so much positive charge in there that it can't take it anymore and then the positive charge will start to leave but since the sample is negatively charged, anything that's negatively charged wants to leave. So it's rejected and pushed away from the sample and so you can actually run the instrument in different polarities. You can shoot negative ions at the sample and then collect positive ions that are forced away from the sample but usually for ions like carbon and what we collect for nitrogen, we collect the cyanide ion which happens to be a really stable molecular fragment. Those are negative ions so those are ejected from the sample and those go through a whole bunch of ion optics and into this thing that's called a mass spec which is really just a giant magnet and then has these Faraday cups positioned at really, really specific locations to catch these ions at different radii. Any other question about the instrument? Yeah. At the end of the day, do you get the concentration of those negative ions in the region where the region that was bombarded? Sorry, I just realized I left a picture of water over here. This is like my PhD advisor. He would come into the lab. I'll get your questions in a second. He'd come into the lab and I would have these flasks on my bench like this and I just never change. He'd come in and be like, you're gonna pay for that. Can you say your question one more time? Yeah, so when you have finally all those ions which are coming out at different locations, what are you probing about the cell there, the concentration of those ions in the cell where this bombardment took place? Is that what? Yeah, so you raster around and I just said you raster around and you stay in one place for some amount of time. So you can raster really quick or you can raster really slow and you can raster again and again and again. And so either way you do it, you end up, you can burn through, you can collect all of the material that a cell is made out of, you can collect all of it and you can count all of it. Turning that into a concentration is dubious, just like doing quantitative mass spec is hard. I want to take ion counts and convert it into femtomoles of carbon. Maybe possible, but it's gonna take some dedication to do that. But essentially what we do is exactly what you said. We don't know for sure about the ionization response rate of every different cell. So if you do this with a cell of one biomass and a cell of a different biomass, or the same types of cells on different types of substrates, you might get a different secondary ion response, you get a different current and you might be misled about how much material you've actually scooped up by shooting this primary ion. But we do exactly what you say, it's just hard to quantify that. So what we do, instead of saying, I've measured this much carbon, as we say, I've got an amount of carbon 12 and amount of carbon 13. Now I can take the ratio of those and if I've done this type of stable isotope probing experiment, I can monitor the ratio of change going from natural abundance to enrichment. And that tells me something about how much the cell has been needing. Thanks for the question. Yeah, go ahead. So the samples have to be fixed, right? That's right. The cells are all dead, which is maybe the biggest drawback of this technique. Cells are dead. Do you have to fix them? No. If you fix them, it's been shown and it's really, really important what's been shown. You fix them with aldehydes. So your aldehydes are, they're natural abundance. And so you've done the stable isotope experiment and if you fix with aldehydes, you introduce, you dilute the carbon. You add some non-labeled carbon and so that gives you an offset. So you can look at kind of raw cells, but they're dead, they're dry, where it minus 10 to the minus 15th tour. It's really dry. So when you rasterize several times at the same spot, you don't expect stuff to have gone away in the meantime. Because I mean like if they aren't fixed, but like I guess if they are dry then there is no. And the second question is like, what limits the number of secondary ion you get? Because you said you have seven, like you can detect up to seven, but like. Great, yeah. That's purely logistical. And I'm sorry, this diagram's really kind of junky, but it's the one that Kamika makes. But it's actually good because it's got all of the details in it. So you've got your secondary ion beam coming out here on the bottom, going to the right. And that's getting focused. And here's your big magnet down here. It's a big black magnet. And the only thing that limits you is, well, so you've, how many trolleys? How many ferrity cups can you put in there? How, and how much of an arc is available to you? So you could imagine, so it's actually just, it's completely instrumental to design. So, they started out making these with five. Now they make them with seven. You could go to 10 or 15. You do run into problems with, like, with, yeah, what's going on with the current and how much space you have available here. But for example, the instrument at Caltech, the data that I'm gonna show in a slide or two, the instrument at Caltech, the operator of the Sims, he's wonderful. And he's the type of person that we could say, we wanna measure deuterium and protons and 13C at the same time. You can't do that with the commercially purchased instrument. It doesn't, you can't, you don't have a big enough radius. You don't have a big enough width to catch the HD and way over here, 12, 13, and then the 15N species and 14N species. So, Youngbin, he's awesome. He said, oh, okay, well next time I'm doing maintenance on the instruments. There's this little metal piece and I'm gonna file it down. And we'll be able to push that over just a little bit and it'll be enough to capture the whole width. And so that's the only instrument in the world that I know that can go, that can capture this really big range. So I'm still using that instrument because of that. And I send samples from Tokyo. Any other, yeah, go. So can you remind me what the resolution is that you can probe the cells? Is it like, can you distinguish what goes into the membrane from the cytosol or is it the comparison between cells? Let me show a picture. So it's something like, yeah, 70, 80 nanometers or so. So this, I don't know from where you're sitting, how much you can see or how good your eyes are. So this is a pseudomonas here. And I don't know if you look really closely, you can see this is a 14N carbon map. We catch the cyanide ion. That has to do with like molecular orbital theory and one ion being more stable than another. But you can treat this as a nitrogen ion image. And you can see it's brighter here and a little bit dimmer there. Here's some, maybe a couple cells that are clumped up. And you can see it's a little bit less bright in the center. I don't know if you can see that from where you are. So you can see a little bit of internal structure inside of the cell. And it's definitely good enough to see between cells. Yes, if I have time, I'll go into a little bit of looking at cellular aggregates that are multi-species and using this technique as a way to try to understand what cells might be sharing with one another when they grow in consortia. So you can see the cell boundary really clearly when they're growing in consortia. That's exactly what they are. Yep, exactly. Exactly. Yeah, just like you were talking about yesterday. Yeah, it's all poisons. Great. Okay, this is great. Thanks for asking questions. Yeah, this is good. So I have a technical question because I don't really know this form of ionization. Do you see any effects of which species ionize first and for which ones you kind of need to kind of bombard for a longer time with CS plus? Yeah, that's a really, really fun question and it's something I'd like to know more about. And we have done a little bit of work but I don't know systematically about that. I said on this image, the positive ions, they're kind of getting sucked into this negative plate and then these negative ions are getting rejected but the kind of ion extraction efficiency of those negative ions, it's not equal for different species. And so we did a test and I don't have the data in these slides but yeah, it looks like for example, the 13 ions, the 13 C ions start to get rejected a little bit sooner than the 14 in carbon, the cyanide ion. I don't know why that is but it's really, really interesting and it's super important for these types of studies and I will be the person to admit, I think this is being filmed but I will admit that most microbiologists, we shoot primary ions into the sample until we get what we think is an acceptable secondary on count rate and then we capture that and we use those data but there is something underlying that and it's exactly our question and I don't think it's systematically understood. It could even be that in the first rejected ions are from different types of biomolecules and I don't have any information about that. It's really interesting to think about and I think it's a really important question. So as a follow up, this would also mean that maybe you could kind of suppress the signal of certain ions depending on the presence of other ions. Yeah, yeah, and vice versa. Yeah, so I've thought about what types of compounds could we add to get greater ionization out because the cyanide ion, for example, it's great, you get these really high intensities out, your counting statistics are good, everything's fine. Deuterium's kind of like you're scraping by, carbon's okay, but it'd be really fun to think about adding molecules that will increase the secondary ion. Yeah, great question. So I'm not sure how dumb this question can be but this thing is about mass spectroscopy always got my curiosity in the sense that you're always measuring the ratio, mass to charge ratio, that's the thing that you're measuring, but in this case that you have isotopes and ions, so you're always changing mass, you're always changing, actually your things have different masses and different charges, can you get wrong reads just because of that, I mean, or is this thing always well defined? I have that mass and that charge and I exactly know what it is. Okay, so in this case, we've got what we call a magnetic sector instrument and so we've got this stream of ions and it's got everything in there and we take that stream of ions and accelerate them, we've got voltage, we're accelerating ions and we've got a magnet and then the ions are taking the corner at a different radius according to their mass. So what we do with this instrument is we position these little catchers in this, these Faraday cups, at what we think is the right radius. So you can put it in the wrong place and catch some ions, so the question is how do you know what we're seeing, right? So when we start off, what we do is we put in some standard things. We know this material has tons of carbon and lots of nitrogen and so basically there has to be some initial kind of field studies where we're like, okay, this is really where protons are and you move, so first what you do is you physically move that Faraday cup and you get it close, but then after that you can't move things mechanically good enough and so then you start to tune the ion beam that's going in there to get rid of, to skim off ions that might be really close in mass and there are like contaminating ions that are really, really close, like there's this boron species that always messes up our nitrogen. There's a lot of boron and a lot of glass that we use so we're now moving to stainless steel surfaces to get rid of that, but you can see that you can't position it physically, you can't position that Faraday cup properly, but you can use the electronics of the instrument to skim off those ions just a little tiny bit heavier. It's a great question and we've collected junk data sets where it's impossible that there's this much 15N and it's like, well, and then you look and it's boron so we're going to stainless steel. This is really great, maybe I should say, I think we're all very eager to learn about how this machine is really working so it's great to learn about that but if we're taking away your ability to talk about what you actually wanted to talk about, you should feel free to completely ignore us and move on. But I'm gonna continue and I'm gonna ask you, so okay, so you're shooting these ions at the cells, I presume they just go through the membranes, maybe come out the other end even, some of them get stuck inside, I'm just, you're saying, okay, so now the cells ejects ions so what is really happening? Are they opening pores and then these things are accelerated in the electric field? Are the cells already bothered by the electric field? You put them in, right, because they also have this membrane potential that they're trying to, I mean it's trying to understand what's happening to these poor cells as you're shooting the ion beam at them and why are they letting the ions out? Such a fun question, nobody really knows but I'll just talk about how I think about it for a moment. So, and I'll talk a little bit about observations. So at first when you start putting these cesium ions, you're shooting cesium ions onto the surface and at first you don't get any secondary ions out. Where are they, what's going on? At first the positive ions just stick. So there's kind of a reservoir capacity of the material to just take up cesium ions and they're just sticking in there. You're not getting any secondary ions out. What's going on? Eventually you put so many cesium ions in that it's like the way I think about it and this is speculation. There's just not room to have anything else in there. Something's gotta leave and then you start to get your secondary ions out. What do you get out first? As you raster and you count, so if you, it's possible that you just raster, you stay here until there's no more ions and then you move along and you stay here until there's no more ions and then you move along. But you can also raster for very, very, very quickly and then take many, many frames of data. And when you do that, if you raster quickly and you take many, many frames of data, you can really see that you're burning through the cell. And so that tells me that it's really a surface analytical technique. So at first, maybe this is going to this former question a little bit. At first you are seeing the outside of the cell and then you're getting into the cell and then finally you're back on the membrane. And so it is interesting to think about kind of doing a depth profile as you burn through the cell and try to recover some of the three-dimensional. Okay, maybe one or two more questions and then I'll get some data. But this is, one or two more because this is part of why I'm here is to like, I wanna learn about how, so when I'm gonna kind of blow my cover here and say I'll introduce some data, things that I'm confused about and then I wanna try to introduce a few ways that I think we can go, how can we, for example, do lineage resolved isotope and corporation measurements? And to do that, I need to learn how like mother machines work and things like that. So I'm really, and vice versa, we can share this stuff together. So is there one more question? I was just wondering like, regarding this profile and structures inside the cell and so on, I realized the treatment that you need to apply in the cells before you get them through the machine might completely affect, like by dry them. For example, you get rid of the cytoplasm, how the things are then sort of gonna be on the membrane or like on the side and so on. The inner structure, I'm not sure how meaningful it is then because you transform it a lot by sort of removing all the water in there. That's right. Yeah, so we're not looking at the metabolome. Yeah, we're really looking at the large biomolecules. Yeah, and you can show that actually really nicely with, if you do this deuterium label, you can show that you wash out all of the exchangeable protons. So you get rid of the metabolome and all exchangeable, do some water rinsing or something. So what are you left with? You're left with the big molecules of biochemistry. Okay, cool. So let's look a little bit at some data together. And then I'm gonna, yeah, so I wanna introduce the type of data that we can observe with this instrument, what it looks like. Introduce a few brief findings and then get into some questions that I have and maybe have more of a discussion about how we could use these data together. And then if there is time, I'll go a little bit into space, but I think this first part might actually be enough. Here's what it looks like. This is for, these are images of staff. Here on the left is the 12C ion image. Middle is the cyanide image and on the right is the proton image. And I put some little notes on the bottom of the slide. So if we add a 13C compound, we'll have a corresponding map of 13C. And we can put those images right on top of each other and then now we've got for every cell, we've got an isotope ratio of that cell. Do the same thing with 15N and we can do the same thing with deuterium. And conceptually, we might think about adding 13C and using that to estimate both the biosynthetic rate and perhaps compound preference if we've got a complex media. If you've got a really defined media, it's gonna be tightly related to biosynthetic rate, but if you've got a complex media, you might be able to distinguish some nutrient preference. And the same is true with a 15N label that you might add if it's a complex media. You might see that sometimes the cells are eating, some of the cells are eating ammonium and some of them are eating amino acids and you can find that. Deuterium, especially from water, I think is rather interesting. When cells synthesize their fatty acids, for example, they require protons and if you add protons in the form of deuterium, you can use this as a sort of a general tracer for biosynthetic rate and that's how we've used that in the past. And I like that because it seems like if you go out in the environment, you get some complex community, you don't know anything about nutrient preferences, you don't know much about even the media or if you're in ocean water or something, but if you add deuterium, you kind of have this general marker of biosynthetic activity. How much material is being incorporated into this system that's growing? Question? Sorry. You were commenting on that, but I think I missed it. But you said that you need to use a lot of isotopes to start with, right? And we've seen yesterday and you were also mentioning that the deuterium is changing sort of all the rates in your cells. So how is this sort of not affecting then, like how is this a good metric for like gross rates and rates of biosynthetic rates when we know that it's changing the rates? Yeah, so with deuterium, the natural abundance of deuterium is really, really low. And so even if we use a one or 5% label strength of water, it's like unbelievably huge compared to natural abundance. I forget 10,000 times higher or something. It's really, really high. But there will be a phenotype with 1% D2O. Like it's, deuterium is really different from protons. So yeah, but it's a good question. It's a really great question with deuterium and then with 13C there are those who will say they'll have a 13C sandwich for lunch. So I don't know. Okay, so I said you can start to use these tracers to estimate growth rates and you can do that like this. Conceptually the way I think about it is you start out with a cell and it's made of natural abundance material and then the cell's growing and it divides. And it's taking up, so we kind of never use 100% label strength because we don't need to. So it's taking up natural abundance label and it's also taking up the stable isotope that you've got in there. And so if the cell divides, if the mother and the daughter cell are equal mass, you've doubled your mass and you can think about the, you can relate the accumulation of isotope to a doubling rate. And one of the questions that I have that I wanna discuss with people in the room is whether or not the mass of mother and daughter cells is equal to one. And what is the asymmetry of material between mother and daughter cells when there is division? But if we just think about it from a really naive perspective that there is no asymmetry in division and the mother cell is exactly the same mass as the two cells that are created after that, then we can use some really simple math to say there's been this much isotope incorporation and that is gonna be related to this doubling time just by knowing the time, the labeling strength and the amount of label that was incorporated into the sample, the cell. Okay, there's some questions about spatial resolution. How good is it? What is it? You know, we've got these planktonic cells on the left. We can see them. Here's a consortia on the right from a methane cold seed. So these, this is a consortia made out of amyarchaea and sulfate reducing delta protea bacteria. And in the right, I'm just playing an isotope ratio image. And in this case, it's showing the atom percent incorporation of 15 N. So we can see that some cells, in this case the experiment was using ammonium, we can see that some cells within this consortia are taking up a lot more ammonium than the other cells. So presumably they're growing faster. Okay, let's go back to planktonic cells and grow cells in a chemistat. So here is a chemistat experiment, same 15 N images as before. But if we look at the distribution, if we look at the accumulation of 15 N and 14 N, or deuterium protons, we can make a histogram plot of what that accumulation is on the right. What is shown is a plot that we have used deuterium incorporation per cell to estimate the doubling time. And here in the middle is the average of the distribution and the x-axis is a log transformed, is log transformed, I can get the pointer out, maybe I can't. It's log transformed mu, specific growth rate. So it looks like for these cells, this is staff in chemistat in a complex media. The deuterium incorporation and the estimated doubling time of these cells, every little kind of hatch mark here is one individual cell. It looks like there's this log normal distribution of growth rates at the single cell level. Yeah, great, yeah, thanks. How's the experiment done? Yeah, we've got a chemistat. You've got a chemistat, they're growing in there and they're eating non-isotope labeled initially. Then you give some media where now a certain percentage is isotope labeled and you know this percentage and then you just give a certain amount of stuff or you give it for a certain amount of time or it's the same thing maybe. And then you wait some time and then you take them out. Yeah, chemistat, steady state, switch over to isotope, wait, in this case it was a half of a doubling time and then take out a little bit and then go. And then these are the, this is the distribution of the estimate of growth rate assuming everybody is taking up this rate at the same time, at the same rate and using this formula that you have. And so you look per cell and you estimate per cell what fraction have the isotope and then you sort of log transform that and divide it by mu, by T to get the mu. Yeah. But do you know what the error bar is on this number that you measure, how accurate can you say? Per cell, yeah, I don't, I'd have to go, oh, we do know, it's much smaller than this. What is, yeah, I have to dig into the supplement. So, but are we really seeing like from one tail of the distribution to the other tail, like eight times 32, is it 256? Yeah. 256 fold rate, a change in growth rate? Apparently. So, this I think is really different from what you see in a mother machine, isn't it? What is the, so I'm confused. 30%. This is right, so this is an area that I'm here to talk, yeah. Many orders of magnitude. Yeah, so this seems weird. So, it's good to talk about what this might mean, yeah. So, I had another question, but now that I heard this description. So, you changed the feed, but then it will take some time before the medium in the chemostat is completely replenished, right? So, at the first, I don't know what the volume is in the chemostat, but the first hour or something, only a small part will be labeled, right? So, most of the cells will still be eating the other, the original medium or not. Yeah. So, I'm rewinding back to when we did this experiment. I try to remember exactly how we did it. And I may have to come back to you, but yeah, it, yeah. Perhaps we grew, yeah, I'm gonna have to come back to and exactly how we designed the experiment, because the point you raised is totally valid. Because the dilution rate is typically how long it takes, not only before the cells grow, but also how quickly you. Yeah. That's true. That's true, yeah. But not the actual growth rate. Yeah. Right, but we did use this really simple formula where we say like, okay, we've got, we start with natural abundance, then we measured these cells and we know the labeling strength and we know the time. And so then we can go to divisions. So, it may have been that we actually just dropped in water and closed the chemostat and grew them for half of the doubling time. Yeah, good question. And this mu bar that you plot here, is it somewhat comparable to the dilution rate of the chemostat? That's a really interesting question. It is, but let me go forward a little bit. So you've run the chemostat with different dilution rates. And so top, we're growing fast and bottom we're growing slower. And yeah, it's comparable, but there's an offset and the offset changes as you change your chemostat and forest growth rate. And so there's something going on here with the isotopes, right? We're using the isotopes as a marker of how many divisions, how much material has been incorporated and how many divisions that might be associated with. But there's also turnover. So there's isotope uptake in excess of division that's going on as well. And so that turnover changes with the dilution rate here, which is related to your question. So I'm missing something here. As I understand, you have a single snapshot, isn't it? Yeah, we do. Now that single snapshot has cells and their daughters as well, right? So, and you can identify who's the mother cell, who's the daughter cell of a particular pair that you might be looking at, right? We can't identify that. You can't. Which is a big problem that we have, this planktonic culture. So I don't understand how by a single snapshot in which you have a variation of your isotope across the cells, you are able to deduce the growth rate. What's the calculation that you're doing? What are you estimating and what are you? How are you calculating the growth rate? Yeah, so we start with some basal level of, in this case, deuterium. And then we measure this. And so basically we're saying if we measure this, how much division would that be? So if it was 100% label strength and we found two cells, or just one cell, we found a cell and it was 50% labeled. If we started out at zero and we don't, because natural abundance is not zero, but let's pretend that we start out with zero D. And then we find a cell that has 50, and we do 100% label experiment and we find a cell that's 50% labeled. And we say, okay, it started out zero or 100% protons. And now it's 50% protons and 50% deuterium. So let me say it, divided one time. And so we get kind of shades in between that. We find 5% deuterium. It's like, well, it hasn't divided, but it's accumulating up the substrate at this rate, because we know time in experiments. Does that help? No. I'm not completely sure about it, but I'm concerned, I guess, there are different cells, right? And in the time period that you have exposed them to your radio isotope. Stable isotope. Right, a stable, okay, the different isotope. You know, different cells would be in different phases of their growth trajectory. Yeah. And they might start encountering it at different times of their, at different phases of their trajectory. Yes. Plus, so, I mean, the amount that they would have, I mean, the variability of the amount of the stuff that you're measuring that they would have could depend on a lot of other things. Yeah. And how would you just zero in from that amount in that cell on the growth rate of that cell? Yeah, that's an awesome question. And I don't think we can get to that from a planktonic study that we're doing. We can pretend like it doesn't really exist and show the data like this, but that's really sweeping us under the rug, that as a cell's dividing, you know, the mother machine, we need to watch some videos after lunch or something and really watch this, right? But here we're sweeping that totally under the rug and it's wrong. It should. Yes, I agree. Yep, yep, yep, yep. Yes. Exactly. Yeah. Yeah, I agree, so I'm feeling okay. But then every now and again I've seen a paper or I've talked to somebody who says that the actual growth of the cell is not exponential. So is that, okay, okay, so I'm happy, you're not worried about this. I'm also worried about the x-axis. Sorry, just a brief presentation, I just arrived very, very late as a microbiologist. I'm not a mathematician or informatician or whatever, but actually that is a chemo state. So actually your growth curve will be always in the log phase. So usually all the bacteria inside should be always dividing because in the normal microbiology usual, if you use a chemo state, the bacteria inside should always grow in a log phase. So they are continuously dividing. So it will be great. Second question, do you have some insights also in the phosphorus incorporation and not only on the hydrogen incorporation? For example, for the DNA, you know, sometimes. I don't know if you have some insights or you plan to do it. Right, right, yeah, so we're in a chemo state. We think all the cells are dividing. We think that the ratio, yeah, we're just looking at ratio, so it does seem like we can kind of sweep the sound of the rug, but I share Sanjoy's feeling of looking down and feeling. So where I'm trying to go now is to do this in a lineage resolved fashion where we can have the complete lineage history of the cells. We have all of the cells individual doubling times microscopically and then we have the isotopes labeled on top of that. That's gonna take a year to finish up, but we've got the methods sorted out now. Great question about phosphorus. Big bummer, State of the Universe says that there's no stable isotope of peak. Great question, yeah, big bummer. Great, so cells grow slower and the variability increases. Why? Let's rewind to Eric's talk yesterday. Really cool thinking about that and we see that here as well. Why do we see this gigantic distribution? Like, yeah, some orders of magnitude higher, 100-fold difference, slow to fast, I don't know. And I don't know if it has something to do with Sanjoy's question or not. Yeah, it's something I'm confused about. We've got multiple isotopes. So we can use deuterium as a proxy for biosynthetic activity. We've got a complex media, multiple nitrogen sources. We've added 15 anemonium. Here's a plot where we think we know the cell-specific growth rate for every individual cell based off of deuterium incorporation. And so we know how much nitrogen should be incorporated into the cells to achieve this division. But on the left side of this plot, we see a bunch of cells that haven't incorporated any label from ammonium. So we think that these cells on the left that are growing really slow are taking nitrogen from amino acids in this complex media. As the cells grow faster and faster, it looks like they start to derive more of their biosynthetically-required nitrogen from the label from ammonium. So it seems like there's some nutrient preference and it's growth rate dependent. But we're doing this chemostat experiment where we've got cells growing along, doubling times, going slow. We've got this kind of medium rate in the middle. We've got this fast rate shown on the right. And apparently, this nutrient preference not only growth rate dependent, but as you go from slow on the left to kind of this medium growth rate in the middle and fast on the right, it inverts. So why is that? Why do the cells like ammonium or amino acids and they like it in different ways depending on the bulk dilution rates? I don't know. This is a question that I'm confused about. But we can start to resolve nutrient preferences. I don't know why that is in this case. So the basic technique here allows us to get a picture of what we think is the growth rate diversity within the culture. Why is it so different from what we see from mother machines? I don't know. It's worrying. It's interesting. Maybe we can do experiments. I've had some conversations with people in the room about making really long mother machines where we can fix cell lineages in place and then look at the isotope incorporation along the lineage. I'm doing that with agar pads right now, agar pad experiments. There's a lot of logistics to solve to get the material into the nanosomes. But at least with agar pads, I think we're just there to start doing experiments. But I'm really happy to brainstorm with others about how we might be able to actually have microscopy on top of this and get away from planktonic growth, which I think confounds us. Question? So I think I've missed one point for your self-specific growth rate. Do you also use the light image or whatever for two kind of segment cells? And this is how you then determine which one is a cell to derive the growth rate from? Yeah, thanks for asking that. Yeah, Sue, we have these ion images. Yes. And we actually just circle all of the cells. And then as a follow-up question, do you see any consequences of cell size, orientation, et cetera, on the growth rate that you're calculating? Super great question. In this case, none. Okay. We see this apparently huge variability in growth rate and we've got pixels, so we know cell size and there's no relationship between growth rate and cell size in this experiment. Okay. I don't know why. Yeah, great question. Sorry, I just thought of one thing. Is stuff known like, I'm thinking there are some organisms like cell 11 that grow really slow, but if you look at them a single cell level, it's known that like 80% of them are dormant and only 20% of them are actually growing. Do you know something like that also happens is the staphylococcus and that's why you see this huge variation in growth rate? I don't know. Is there a microbiologist in the room? There is. Yeah, I don't know. It's a great question. Yeah. Okay. Great. So I've already been up here for about an hour. You have 10 minutes. Oh, I have 10 minutes. Okay. It's almost lunchtime. We've already gone through some of these things. So I've tried to give you an overview of the instrument, its capabilities, some data that we've acquired and we've already talked about things that I don't understand. And so what do we want to incorporate next to these types of experiments? I'm trying to incorporate cell age and lineage history. Like I mentioned with egg or pads, maybe it's fun to think about with a mother machine. I'm really curious about the distribution of resources and biomolecules between mother and daughter cells. Is it 50-50 and if it is, what is the distribution of new material compared to old material when cells divide? This is a question I think that we could answer with this technique. One thing that is completely ignored so far is variability in the actual cell composition. So I know when I've done transmission electron microscopy studies of cells, I can find some cells that have a ton of compartments and storage granules inside of them and other cells that don't. And this is gonna change all of our ratios and our whole notion of how ratios of isotopes can be correlated to growth rate. And so the variability in cell biomass is something that I think needs to be defined and investigated. And I think that's harder to do. How do we get a good picture of the C to N ratio in a cell? The nanosomes kind of gives us that, but does anybody have an idea about how much variability in biomass composition there is between cells and the culture and how to access that information? So we looked into this a little bit using modeling, but because all these proteins that you're making are pretty large and they all take 20 amino acids, turns out that if you just look at the proteome, you take proteomic studies and you just calculate which amino acids are in there and the frequencies, they're really constant across conditions. Of course, it could change that you use more lipids or less, but proteomic, I'm pretty sure it remains the same. Yeah, I feel good about the proteins too, but I worry about glycogen or fatty acid storage, yeah, polyphosphate, blah, blah, blah. Okay, what do people wanna do? You wanna ask them questions and talk or do you wanna see some more? There's at least one question. I do have a few slides to show how we can apply this to these, to multicellular consortia and those might be interesting to talk about in a show. Yeah, but a question. Yeah, I was also getting there. So my understanding about the system is that you need to fix these cells in order to image them and then understand the ratios, right? So you were talking a bit about mother machines, so I'm wondering how you can do that there because there you have cells in these microfluidic devices and they're extremely tiny. How can you retrieve cells and then look at them? Have you thought about that? Well, luckily I was brainstorming the other day with, you know, about this exact question and luckily we came into this room and I started doodling on this white board out and it's still here. And it doesn't look very fancy, right? Just looks like some drainage system or something. So my idea that I had with Fijo when we were talking was he's using PDMS. And so I said, okay, what if you make your kind of exit channel really, really long and you've got glass on top and glass on bottom. You've got PDMS in between and the daughter cells are all coming out but you make that quite long so you can kind of save those cells. And here on the top are these wells where you're putting in nutrients. You've got glass on top and glass on bottom. And so the idea that we had the other night was fix the cells in place chemically, embed them in plastic. So you've got the spatial structure there and then peel the glass off and then see if you can put that in this instrument. That's just one idea, but I don't know how to do it. Oh, cool. Yeah, that'd be great. Yeah, cool, thanks for that. Yeah, well I guess it all runs in the family because I did my postdoc in Victoria's lab and that's where I started noodling around with the nanosomes and yeah, so yeah, it's a good idea to chat with her and brainstorm. Okay, do people want to see a couple like spatial imaging? Okay, so this is, so we're in the planktonic world, right? And we're pretty confused because we don't know where the cells have come from, right? And this leads to this question, okay, can we do it lineage in a lineage-resolved way? Let's rewind to my postdoc and you mentioned Victoria. So Victoria's lab looks at anaerobic methane oxidation and that can happen in a few different ways, but one way that it can happen is that you can have an archaeocell that can do C1 metabolism, take methane and go to CO2. Where do the electrons go? Good question. They go somehow, probably, over to these sulfate-reducing bacteria. So here's an image where we're using fluorescence in situ hybridization, fish. We can paint cells based on their ribosomal sequences, their ribosomal RNA using a probe that hybridizes to a specific sequence and the ribosome has a little fluorophore on it. So in this case, the archaea, their painted red and these sulfate-reducing bacteria are in green. These are natural samples, really complex community. Here's a picture on the right. It's really just seafloor sediment, complicated mud, but the fish probes really allow us to go in and find cells of specific phylogenetic affiliation. Bad thing about fish probes is that the phylogenetic specificity is pretty broad. So I'm calling these sulfate-reducing bacteria and amyarchea, but the fish probe resolution is, I don't know, genus or larger or something like that. It's not very good. It's kind of embarrassing. And you might even notice funny things like the cells are a lot of different shapes and sizes and even the fish probe seems to be interacting with the cell in different ways and what's going on with that, I don't know. Is it different types of cells or is it the same type of cell in different states? Not sure. But when I got into Victoria's lab, I'm not trained as a microbiologist at all, so I was just playing around with the microscope and having fun. I noticed these archaea, bacteria, consortia come in all sorts of shapes and sizes. And this got me really confused because if you think about the need to donate electrons to a centrifugal partner, remember, one cell is taking electrons from methane and putatively trafficking those electrons to a partner. If that happens in any way in a process that's diffusion dependent, it seems like space should matter and size should matter. But you've got so, you've got clumps of cells like this kind of big tanker up top where the archaea and the bacteria kind of hemispherically separated from one another. And then you've got things to the bottom left where they're all really, really intermixed. And so people have had a blast doing all sorts of reaction diffusion modeling and shown that small, well-mixed consortia that are centrifugal should really out-compete these big hemispherically separated types up top. But you go into nature and you find those. And so I started wondering like, okay, is there any relationship with what's going on with the size and spatial dependency here? There's an anisomes in the basement of Caltech. I could go and use it. So I decided to do a nice probe experiment. In this case, Young Bin had not yet filed the instrument down. Probably highly illegal compared to if you ask the Kameka company to do this. He had not filed it down. And so deuterium was not yet available on the instrument. So I just used ammonium to monitor growth rate. What do we have here in this picture? We've got three-dimensional consortia. The anisomes is awful with depth. It goes out of focus, I don't know, like 20 nanometers or something. It has no ability to track focus. You can do that manually. But if you've got anything that is not flat, it's awful to try to get spatial resolution on. And so what I did is I put these consortia into plastic and I cut them. And I've got one sheet, and that's what we're looking at here. It's a sheet, and I've concentrated these consortia. And then all of these little kind of stars in the sky here are these different consortia. And I've painted them using this fish approach. Then we can put that in the anisomes. And we can get these phylogeny activity pairs. From the instrument and from the microscope. And then we can start with all of these shapes and sizes, start asking really basic questions. Like, what's going on between archaea and bacteria that are right next to each other in comparison to archaea and bacteria that are very far away from each other? What's going on with the size of the aggregate in terms of its biosynthetic activity? What's going on with the cell distance to the surface of the aggregate? Start asking all of these types of questions. And we did. It was really interesting. But it's time for lunch, so I don't know what to do. I mean, should I keep going for five or 10 minutes? I don't know what to do. And we have to walk up the hill today, right? We're gonna burn calories. We need to, yeah. Do I have some questions? Yeah. The data, yeah, what a cliffhanger, right? It's really, really interesting. Let's do it the other way around. There are no more questions, and you're gonna show the data. If somebody has a question, they can ask him after the... All right, it's my fault. I've taken responsibility. I can go fast, and when people are like, we need to eat. Okay. So the first plot I wanna show is a plot that shows the kind of biosynthetic activity the uptake of the 15N for sulfate-reducing bacteria on the y-axis and for anemia archaea on the x-axis. And in this plot, every data point is one consortium. And so what this data plot tells us is that the archaea biosynthetic activity is correlated to the bacteria activity. Doesn't have to be that case. It could be that these things are just glommed on to one another and happen to be spatially next to each other, but they're growing at different rates and they break off and they do all sorts of stuff. But I thought this was kind of a really kind of a test of this centrophic hypothesis. These cells are together, they're clumped on to one another, and they're growing in a proportional rate. Blue and red, great, thanks for asking. Blue and red are two different phylogenetic affiliations. So there's these different fish probes that we can use and there's a big mixed community. There's a few different types of sulfur-reducing bacteria in here paired with the same anise. We think it's the same anise and we don't have the phylogenetic resolution to prove that. But different phylogenetic groups. Okay, so we think these things are, we really think they're kind of co-breathing together. They're centrophic. But what about our predictions from modeling? Picture on the left here by Bernard Schenk. We're predicting that well-mixed communities that are sharing electrons, something or molecule that's subject to diffusion is gonna be much better off than a situation like the one on the right. But in the natural world, we find all these shapes and sizes, which is confusing. So we wanted to look a little bit at spatial positioning and we derived kind of a size-dependent way of capturing mixiness. If you will. Here on the top left, there's a checkered board and on the top right, there's all the squares that have been separated. And so we did this in a way that's kind of normalized to size and I can't get into the equation right now. It's been a few years since time for lunch. But this is a size normalized way of capturing the mixiness between two types of colors. And we tested that against the natural consortia. Here's on the left, there's this thing that's all mixed up. And then on the right, there's this one that we call Mickey Mouse. It's like these, you can see these pink, peaky ears up on the top. It's a little bit dim on the screen. But that one's really, really separated. So we've got this metric that captures the degree of spatial heterogeneity in here. So we asked, okay, if you look at the activity of these cells in the consortia, does it have anything to do with how segregated or how well-mixed they are in the consortia? It doesn't have anything to do with that. Got these two phylogenetic groups. It doesn't matter if the cells are like this kind of Mickey Mouse type. The archaea are totally, you know, segregated from the bacteria, compared to very, very well-mixed. A lot of growth rate, heterogeneity, and diversity in here. But it doesn't have anything to do with the kind of bulk mixiness that we find in the consortia. We started trying to look inside of the consortia, specifically, and start to think about distance activity relationships. And that's kind of shown, illustrated nicely in this nice picture from Beth Orkut. Christophe Millay. Think about these cells packed in like this. You find consortia that look like this. You might think that the cells in the interior are at some sort of a disadvantage compared to the cells that are right at the interface of their centrophic partner. Here's an example in nature that looks kind of similar to that previous picture. We started trying to analyze these distance relationships. And so we've got every single cell, we've got the phylogeny, we've got every single cell's ammonium uptake rate. We've got every single cell's position, and so we can put grids on all of these consortia and start to ask questions like, what's the activity, you know, correlation of these two cells that are really close to one another compared to very far away from each other? We can do that across all of these consortia by using a z-score to normalize. And we see that the distance to the centrophic partner doesn't really have anything to do with the cell activity, either for archaea or for bacteria. It doesn't look like space is mattering here. Another thing we might ask is, is activity related to your proximity to the surface? You might think that access to nutrients is better on the surface of these consortia compared to buried in the middle. Look at that. You can be buried in the middle, nine or 10 microns inside, and have the same activity as being on the surface. It doesn't look like there's anything to do with spatial arrangement, neither at the whole consortia relationship, how well you are mixed or how kind of clumpy you are, or whether or not you're nestled up to your centrophic partner, or you're very far away from your centrophic partner. So we've got these data, and we didn't know what to do for like a year, and we're just talking about it. What does this mean? What does it mean? We start thinking, I don't know, there's this kind of new fill. This was a while back, kind of still it was newer then, of an electron microbiology. We started thinking maybe the cells are not sharing anything like a molecule, and maybe they're sharing electrons directly. And that might look something like this. Whoopsies. A cell might incorporate something like this beta, this barrel protein into a membrane, and then fill it up with heme groups, and then stack a whole bunch of heme binding proteins on top, and make a little nanowire, or make a blanket, a conductive blanket on top of the cell surface, and that might allow electrons to be trafficked around these consortia. Remember percolation theory. Can I get material from here to here, or is there a path to go from here to here, depending on the density of different types in the conduit you have? This is the kind of problem that we have here. Look in the genomes of these anemiarchia, you find these proteins that are some of the largest multi-heme cytochrome-containing proteins that have been found in archaea, and they've got a surface layer domain. So it looks like they're embedded in the outside of the cell. These archaea have this liquid crystalline cover on top of them. Maybe it helps hold the cell together. That's called an S-layer. These proteins, S-layer domain, and then it's studded, it's full of all of these heme groups. This is all genomic prediction, by the way. Wanted to test that genomic prediction, looked into the really old literature where people were trying to measure the intracellular compartment pH of different compartments in eukaryotic cells, and you could use peroxidases. Little, you take a protein with a different pKa, and it has a heme group on it. Depending on the pKa, that protein will be localized in different compartments, according to the pH of that compartment. And then the transmission electron microscopy, folks in the 70s, they used this peroxidase assay to paint, in this case, with osmium, something that's a high Z element, something that has a lot of contrast with transmission electron microscopy. So we brought out this stain that hasn't been used for a long time, kind of brought it out of history and used it, I think, for the first time in archaea, and we could paint in between the cells. See images, if you look in these archaea, between them, without this stain, you can see there's kind of something that looks like some sheath. It's just some higher Z number containing elements, a little bit less electrons going through, but we could paint the outside of these archaea and see this really black area. The blackness comes from a peroxidase activity, where you add hydrogen peroxidase as an oxidant, and this molecule called diaminobenzidine that's the reductant, makes this big, not defined molecule that sticks to osmium, it's osmophilic and you can see it with the electron microscopy. Okay, and that allowed us all to, it allowed us to generate this model where we think that the archaea are centrifugally related to the bacteria based on interspecies electron transfer, and that's why we think there's no distance dependencies. It's time for lunch. This is the whole thing that I've mentioned. I went really fast in the last part of the talk, so if people wanna talk more, let's have lunch and talk more, and I'm around for the rest of the week. And just wanna show up another picture of the Institute. We have beautiful cherry trees at the end of March, getting earlier and earlier every year. This summer was the hottest in history in Japan. It was really hot. So please come over. We're studying the origin evolution of life. It's a space that reminds me a lot of this space. I've had a really, really good time being here. We've got labs in the basement, so if people are interested in doing lab experiments or theory, I think it's a really fun place to be. And we'd really welcome people to come over, and so feel free to keep in touch. Thanks for bearing with me during the long talk, and thanks for all the questions. So one more question, only one. You have one more question. So what do you mean later? Oh, good question. Great, great. Potentially the most important question. That's the most important question, right? One more question. No, scientifically, no, so you can have lunch, I think, and you have some comments? No. On lunch. Let's thank Sean again. And so there are...