Plant Physiology Webinar: Highlighting Focus Issue on Plant Cell Atlas, Feb. 2022
Hello, and welcome to our next webinar in our series. Today we have three speakers speaking on the topic of the plant cell Atlas. And the first one is is Hilda. The second one is Daniel and the third one is Jen and we are hosted and moderated today by Kenneth. This webinar is brought to you to celebrate the February 2022 Plant Physiology Focus Issue on the topic of "The Plant Cell Atlas". And this Focus Issue was edited by Kenneth Birnbaum, Julia Bailey-Seres, Marissa Oteguie, and Sue Rhee and it's currently online, you can find it at the Plant Physiology website. You can find this and and many other recordings of webinars available on YouTube shortly on the Plantae YouTube channel. We'd like to make a special thank you to ASPB
members. If you're not a member of ASPB, we encourage you to join. The membership supports this series, we have a code Presents10 that will give you 10% off your membership dues. Please put questions for the speakers into the questions box. And they will be read out by Kenneth after each speaker speaks. If you have technical questions, feel free to email me
at mwilliams@aspb.org. And now I'm going to turn the microphone and the camera over to Kenneth. Thank you. Okay, great. Thank you, Mary. And thanks. Thank you all for coming. I first just wanted to explain what is the Plant Cell
Atlas, I think that would be a good thing. And technically we define this as the quest of the mission to identify all plant cell types and characterise their function. So this is a big task. And clearly sort of implied in this there's there's a technical side that we are looking for innovations or adaptations to techniques that are going to address specific biological questions in the plant kingdom. So very much
geared toward the plant kingdom. And you know, you will recognise by the name, that it's sort of fashioned after the Human Cell Atlas. But in some senses, we want to go beyond that, as you'll see today. And of course, these are very much geared
toward toward plant questions. Okay, so the Focus Issue that that Mary was talking about had 17 articles and reviews, and we, we encourage you to go look at that as Mary did. But today, we can't give you all 17. So we're giving you, let's call it a tasting of some of the things that we think could and will inspire you. And hopefully interest you and sort of our little treats out of that. And so let me stop talking and turn it over to the people are going to talk about the science. And
so the first speaker up today is Hilde Nelissen, she is a group leader at the VIB in Ghent in the Plant Science biology department. She's interested in growth. And she uses the maize leaf as a model, as you'll see. And she's going to tell us about spatial transcriptomics and PLASTOCHRON1, and I will turn it over now to Hilde. Go ahead. Okay, thank you, Mary and Ken. Good morning, everybody. And
good afternoon, if you're in Europe or in these parts of the world. I'm very grateful with this opportunity to highlight the paper that we had in this special issue in breakthrough technologies, which was mainly focused on the in situ sequencing approach, which is only one of the ways to do spatial transcriptomics. I've altered the title of this presentation a little bit because I want to go even beyond what we published in this one paper to show you what the possibilities are with spatial transcriptomics. So as Ken already indicated, we're very much interested how a plant grows and use maize as a model system. So typically, a lot of the traits that are measured are endpoint traits. And so for
instance, final plant height. But if you measure this, it actually has a lot of growth processes that were occurring in order to achieve this final plant height. And this growth occurs in the individual organs, but also we've shown and others also that there is a high coordination between the growth processes in the different organs, for instance, between leaf, stem, and inflorescences. And so we've mainly chosen the maize leaf to study these growth processes. Because it is quite easily accessible because of rapid turnover time. And also because it has a nice spatial organisation, because the dividing cells are at the base of the leaf followed by the elongating cells and then the mature cells. And so this allows
us because of the size of the maize leaf to really sample for the cells that are enriched for these dividing cells or elongating cells. And as such, we can do a lot of molecular downstream analysis. And so we combine this macroscopic phenotypes with the cellular analysis and molecular analysis to further understand how these growth processes are coordinated. And one of our pet genes that we have identified as such as called PLASTOCHRON1 or PLA1 in short, and it's part of a family of cytochrome P450s, ehich are known also in other plant species to stimulate organ growth. And in maize, we identified this, when we trrf to overexpress PLASTOCHRON1 we found that the plants were producing leaves which were much broader, and they kept on growing, these plants have difficulties to make inflorescences. So we had to go
back and make another design. And then we chose to express PLASTOCHRON1 ectopically within this division zone of the maize leaf, using a specific promoter that we identified through our profiling studies. And when we did this targeted overexpression, we actually saw that we could indeed stimulate organ growth in a way which was beneficial, like it's shown here for seedling vigour. But also later on, when we tested these plants in field trials, we could show that it had a positive effect on other organs. And so it affected final biomass and
seed yield. So this really showed to us that PLASTOCHRON1 had a lot of potential. So we further wanted to understand how its mode of action is. And so when we looked at the cellular
analysis, we saw that PLASTOCHRON1 had modes of action through making cells dividing for a longer period of time. And so we were thinking maybe it keeps the cells in undifferentiated state for a longer period of time, but we couldn't really understand how it was doing so. On top of that, there were some other clues that we got that also pointed for us towards a role in the shoot apical meristem, besides the possibility that it regulates undifferentiated cells. PLASTOCHRON means that it regulates actually the time between leaves to be formed. And as you can see this in this case, this is a knockout mutant of PLASTOCHRON1, you see that shorter leaves are being formed, but that more leaves are being formed at the same amount of time as the segregating wild type. And this becomes progressively worse as plants
develop. And as leaf initiation is also a process that takes place in the shoot apical meristem, we had another hint that PLA might function in the shoot apical meristem. And finally, there was a third hint that led us to the shoot aplical meristem. And that was when we profiled the expression of
PLASTOCHRON1 over this leaf developmental gradient that I explained. And then we found that PLASTOCHRON1 was very highly expressed in the dividing cells, but that it dropped almost zero even in cells that we could still identify as dividing. So this means that PLASTOCHRON1 discriminates between dividing cells close to the shoot apical meristem, and more distant. And so for this reason, we really thought that we should check out the shoot apical meristem. And we did so
by using in situ hybridization with PLASTOCHRON1 as a probe. And that, in this transverse section, we found that PLASTOCHRON1 was not expressed in the emerging primordia in the P0, which you can see from this horseshoe shaped expression profile. And we could confirm this in transverse section. So the absence of PLA1 the emerging primordia. However, we see that PLA is specifically expressed in the cells that subtend the emerging primordia. And when we zoom out a bit, we can see that this expression goes down as the primordia become older. And so to us it looked more like PLASTOCHRON1 was a spiral around the shoot apex. And so this was very intriguing to us. And we
wanted to know how the expression of PLASTOCHRON1 would correlate to known marker genes. Very much at that time, we got the opportunity to actually test in situ sequencing technology which is a spatial transcriptomics technology that allows to visualise the expression of multiple genes simultaneously on the same tissue context. And based on for instance, the laser capture microdissection date of the Scanlon lab, we picked a panel of 90 genes with high and low expression levels. Based upon literature, we took genes with very specific or broad expression levels. And we looked
whether we could apply this technology also in plants because until then it was mainly used in animal tissue. And as you can see from the image, this really worked well. For the majority of the genes, we could not find any correlation between the genes that did not function and the level of expression or the specificity. And so when we did this, and we brought this
technology to plants, we implemented another improvement. And that was in order to map the expression level to the individual cells. So what you see here are general expression profiles, but when you can segment different cells in the background, you can actually map all these count to the individual cells. And in animal tissue, this segmentation is
typically done by DAPI staining to visualise the nuclei, and then it's thought that the cell is a certain radius around the nucleus. However, in plants, we know that the nucleus is not always nicely in the centre of the cell. And so we complemented the DAPI staining with the cell wall stain. So once we had evaluated the technology in plants, we could use it to do some biology and then circle back to our favourite gene PLASTOCHRON1 . And so here I only showed three of the 19 genes because it becomes a little bit crowded sometimes. And what I want to show you here is that the
expression of PLASTOCHRON1 , which is in yellow, which is again, this spiral around the shoot apical meristem or down the shoot apex, it actually coincides with a boundary that has been formed by the expression of the red gene and the blue gene. And the red gene is actually Rough Sheath 1 which is an ortholog or a homologue of KNOTTED1, which is a marker for indeterminant cells. And the blue marker is actually a market for the epidermis, which means that really a cell fate has been acquired, and these are determinants cells. And so you can see that there is a sharp demarcation between the red and the blue genes. And that actually PLASTOCHRON1 is expressed at the boundary between these indeterminant and determinant cells. And so this fit actually with the phenotype that we saw in our targeted overexpression line where we could show that indeed, we saw a longer and higher expression in all the leaf primordia of PLASTOCHRON1. It also ties in
with what is known in literature, because in the single cell data sets that were published over the past couple of years, PLASTOCHRON1 always turn out to be in the same cell clusters as far as instance KNOTTED1. And so we are now further investigated the link investigating the link between KNOTTED1 and PLASTOCHRON1 because also about 10 years ago, PLA1 was identified as one of the putative KNOTTED1 targets. And so, this has led for us to a working hypothesis that we are further following up on.
Sorry, three minutes. Yes, once we had now the system optimised for in situ sequencing, we also saw whether we could apply this beyond the paper that we published and for other technologies of spatial transcriptomics. And here I showed that we could also make it work for molecular cartography, which typically has a smaller region of interest that can be imaged with the higher resolution. When we complemented this with the
segmentation, we were able to really map the expression counts to the single cells. And as such, we can do a clustering and a co-expression analysis. And this analysis showed us that plus two from one PLASTOCHRON1 was coexpressed with Angustifolia3 or GRF-Interacting Factor1. This has been shown in the past in Arabidopsis to be a growth regulator, but also by the team of Sarah Hake. It was shown that GIF1 plays an important role in meristem determinacy, and therefore, we actually decided to make a double mutant between PLASTOCHRON1 and AN3 or GIF1. And as you can see, in the
double mutant, we actually have a more severe phenotype than in the individual mutants, and especially when you pay attention to the amount of leaves that are being formed. Then again, we went back to the spatial transcriptomics to further investigate this genetic interaction. And so we use spatial transcriptomics with the same gene panel that I showed earlier in wild type and compare that to the mutant background, this is one flavour I can give you, where you can see that the expression of PLASTOCHRON1 is altered and is more expanded into the dome of the shoot apical meristem in this gif1 or angustifolia mutants. And we could actually confirm this
again with classical or single gene in situ hybridizations. So with this, I hope I have convinced you that spatial transcriptomics is a very nice tool that we now can, a targeted approach that we can use in developmental biology, either to compare expression profiles between different genes, we can do a gene targeted approach, we can look in which cell types is our gene expressed. But you can also do a cell targeted approach where you click on certain genes, and you know which genes are defining the this type of cell. However, when you map it
to a single cell level, you can even go to higher or the co-expression levels. And I think it goes without saying that you can use this technology to for instance, annotate single cell clusters that you obtained. And so with this, I would like to thank the team that did this, and especially also all the collaborators, especially the people of the platforms that are mentioned, and the funding agencies. And then the people in
the team that did the work was pioneered by Reinout who graduated two months ago, and is now taken over by Jessica and Denia with the assistance of Julia and Kirin. And with this, I would like to take questions, and hope I stayed in time. That was great Hilde. And you're right on time. So thank you.
Perfect. Okay, so we have a few questions piling in the q&a if you put them in the chat, put them back in the q&a, and I will try to get to them. And so quickly, we have a couple of questions on resolution. And I guess we'll compare the two techniques, in situ sequencing and spatial, other spatial transcriptomics. And I guess we're probably talking about sub cellular level. Yeah, these providers don't don't the sub cellular level, and how this we didn't really look at it, I think we found it already a challenge to map it to individual cells. But I think it
is possible to go to sub cellular level. Yeah. And Hilde you mention, sorry. Yeah, we didn't do it yet. Okay. But you mentioned that the second technique had higher resolution than the in situ sequencing. Yeah, that is correct. In our hands, it had a higher resolution, which means that we had per cell more counts on at the gene level, which then allows to do much more. This
gives you much more statistical power to do this coexpression analysis. However, the in situ sequencing has a big, has the advantage that the region that you can image is larger. So for some applications, you don't need always need to go to a single cell level. So it depends, I think, which provider to use, which question you're trying to answer. Okay. Okay, good. We have another question about PLA1 . And whether it's doing the same thing in roots.
Well to our knowledge PLASTOCHRON1 is not expressed very highly in roots, also, and this I definitely know that the counterpart in arabidopsis is not expressed in the roots. However, we do see root phenotypes in our overexpression lines, especially also in arabidopsis, when you use an over, a 35S overexpression line, you definitely see phenotypes in the roots. But that's of course, because you ectopically express. Normally, it's not expressed in the roots.
Okay, and maybe one last question here on PLA1, which is do you have a hypothesis on how the localised expression of PLA1 can affect leaf growth so much further from the shoot apical meristem? So the mobile signal? Yes, so so that is indeed, one of the big hypotheses that we have, because PLASTOCHRON1 is a Cytochrome P450, which means it's an enzyme it hydroxylizes something, unfortunately, or maybe fortunately, because that gives us something to work on. We don't know exactly what the molecule is that is being catalysed by PLASTOCHRON1 and orthologues. And so, definitely it has been shown in other species that it is mobile signal. It's thought to be, yeah, maybe a type of hormone. Yeah, so definitely, that's one of the ideas. On the other hand, as you see that this is on the boundary between cells, you could also see it like a shuttling so that it starts to initiate another signal cascade downstream into into the leaves.
Right. Interesting. Okay, cool. All right. Good. I think I think we should move on and stay on time. Thank you very much. Hilde, that was great, interesting. And while we're
getting the slides up, let me introduce our next speaker. It's Daniel Kierzkowski. He's an assistant Professor up in Montreal. And his work in general focuses on Molecular Cellular and tissue level mechanisms, really looking at plant shape, plant organ shape, and biomechanics. And he's going to talk to us about the sorry, he's going to talk to us about more about now we're going on the imaging side and tracking imaging of the growing stamen and the dynamics and I don't see the slides up are we having an issue? It's coming. Okay. So thank you very much, Ken, for the nice
introduction. Thanks for for inviting me here and thanks to everybody for being here with us. So today, I would like to show you how live imaging can help us to understand growth and development of organs. Specifically, flora organs that
are hidden inside the floral bud. One of the most fascinating question developmental biology is to understand how a group of undifferentiated cells can give rise to this tremendous variability of organ shapes that we see in nature. Organ forms emerge from intertwined interactions between different developmental processes. Cells have to grow in order to make the bigger organs. Organs have to be patterned, so different parts of the organ has to get some specific identity, so the cells have to differentiate as well. Those processes are controlled at cellular level via molecules like hormones and genes, but also by physical forces, especially in plant tissues, where cells are interconnected by stiff cell walls. If you look at the organ, for example, this arabidopsis
leaf, it's very difficult to infer how this organ has developed, what are the growth patterns leading to the emergence of this specific shape. One of the reason for this is for example, is that different growth patterns can actually generate the same forms as exemplified in this in this simulation. So, we really need to go and observe cellular behaviour over time to understand development of organs in plants. So, how do we do that actually in in our lab, so, we
use confocal time lapse imaging combined with MorphGraphX software which is the image analysis software developed by Richard Smith group. And we acquire sequential images of the living tissue in this example, it is shoot apical meristem of tomato, with some some time intervals. Using cell outlines, for example, membrane marker or cell wall stain, we can now segment those images on the surface into cells. And this helps us to generate heat maps ofdifferent growth parameters. So, what you can see here for example, is a heat map of growth, other way area extension, where you can see that cells that are growing very quickly are marked in red and cells that are growing slowly are marked in blue. So, for
example, emerging, future emerging primordium here, where I show you is growing very quickly. So we have all quantitative information for all cell behaviours of entire organ during its development. And such an approach has been extensively used before to monitor leaf growth and recently has been also used to look at the growth of floral organs, more specifically sepals. So sepals were very easy to access and people have been looking at the growth of the of those organs in the past. But we are interested in more challenging, more difficult organs to access. So we want to look at the development of stamens so male reproductive organs in flowers.
So in arabidopsis, we have six stamens, four longer stamens and two shorter stamens, and they are basically composed of two distinct regions, at the bottom you have a stalk that is the filament, this filament will be elongating a lot to deliver pollen to the stigma. And at the top of this filament you have anthers, which where actually the the pollen grains are developing. So if you do a cross section for this anther, we have four lobes, in each lobe will contain the locule where the germ lines will develop. Until now, stamen development has been basically inferred from fixed tissues. So basically, people were using histological
sections to understand how those cells are working together in order to create the forms that we see. And the reason for this is basically, that entire development of the stamen is hidden inside the floral bud so we don't have easy access for confocal imaging. So how can we look at the stamen developments with confocal imaging, if this is hidden. So my postdoc Sylvia Silveira in my lab has developed a new approach to to do so. So she is dissecting inflorescence of Arabidopsis, she's getting into the flower when were we the floral bud when we know that the stamens are initiating, she's very gently removing abaxial sepal uncovering the initiating stamen primordia and now we are able to follow this the growth of those organs with confocal microscope. So here what you can see is that we can now follow the development of this organ from its very early initiation, basically a very tiny primordium until the final shape of the stamen is fully developed. So the same what you have seen in
the movie is depicted on this image. So we can see we are actually observing the same stamen from one day of initiation, a tiny primordium 30 microns, until 10 days after initiation when you have a fully developed organ with different parts and it's nearly ready to deliver pollen. So it's one week and a half that of imaging that we can now perform. So what can you learn from that? Quite a lot of things of course, I don't have time to go into details in this talk, but I will give you some examples. So for example, we can try to understand how many cells are giving rise to the primordium to the stamen in the epidermis. So what you can see here is this floral meristem
where before before any signs of stamen primordium initiation and one day later, we have a clear bulge indicated by the positive Gaussian curvature. So, if you now mark all the cells of this very early primordium and trace them back to the floral meristem, we can say for example, that okay, around 20 cells in the epidermis will be contributing to this to this organ, which is the number very similar to leaves or sepals. We can also now precisely determine the timing when different parts of the organ are specified. So, filaments when actually the filament and anther are getting distinct fates. To do so, we can
use the cell clones basically, we mark the cells with specific colour at the very early stages, so at four days after initiation in this case, and we look how they develop over time. And we look at their sector, so their progeny. And so if we look at this progeny at the light last stage, we can see for example, at this specific time, that's that all the sectors that developed from single source at the very early stages will be either located at the filament or out on the anther, indicating that this is exactly the stage when the two fates are specified, thee filament and anther. So we can see it here coloured in in blue for anther and for here in yellow for filaments. And this is very clear now for example, that at this very early stage, when the framework is still very tiny and looking like a finger, like a finger, we already have those fates specified and the filament that will be very, very long structure up to sort of few millimetres is actually developing only from two to three rows of cells. So how this
filament elongation occurs, we can now look in detail into that. And so for example, what we decided to do is to see which part of the very early filament is contributing the most to the final length of the structure. So we marked the early filament, which we divided into five equal parts 20%. And we have looked at it what happens over time with them. And strikingly, the two top parts of green and yellow sector that represented about 40% of the initial length, they don't grow, this part is not growing. So it just this
part is not contributing to the filament. In contrast, the bottom parts like the red one, this is the part where the growth actually occurs. So we can look at this in even more detail. So what you can see here is the heatmap of growth for every single cell during the entire elongation of the filament. So again, just to remind you, red cells are growing fast, yellow are growing slowly. And we can see that at
early stages with very fast growth, which is slowing down progressively to increase again, just a bit, just before anthesis. If you look at the cell proliferation, so cell divisions, we can see that they occur at early stages, and they progressively are eliminated. So the last stage of growth, there's no divisions at all anymore. So we can actually
looking by that sort of data, we can distinguish three phases of growth of the filament. In the first phase, cells are growing very quickly, cells are dividing a lot and all the structures elongating. Later on in the middle phase, the cell divisions are actually restricted to the bottom, the growth is also restricted the proximal part of the filament, and the top part is not growing. So we have sort of a restriction of growth
towards the proximal part of the of the of the structure. Later, we have acceleration of growth. So this last stage is there is no divisions, we have very, very fast growth. But this growth is not any more located at the bottom, it's more displaced over the tip of the filament. So it's like it's this, it's very
similar behaviour that has been observed in the hypocotyl of arabidopsis for example. So here are the three different stages of the elongation of the filament. So what I have shown you here very quickly is that our approach enables really precise tracking of cells to understand the origin of different parts of an organ or an organ itself. And this enables us to also very precisely in characterise the growth and development of the organ at cellular resolution in the quantitative way to basically put numbers to every single cell and understand how they interact with each other to generate a specific form. So this research actually provided
the first quantitative atlas of cellular growth dynamics of any kind of the internal reproductive organs in plants from the initiation until the full development. So this is the first quantitative atlas of internal growth for the first time in Arabidopsis. So this work was done mainly by Sylvia Silveira, a postdoc in my group who has designed the approach and had all the imaging and analysis, and Sylvia was helped by Constance and Andrea in this work. This work was done in collaboration with Anne-Lise Routier lab from University of Montreal. And of course, it was possible thanks to the funding from different Canadian funding agencies and from from the from Quebec. So I would like to thank you very much for your attention, and they will be very happy to get any questions you have. Thank you very much again.
Okay, great. Thank you, Dan, I didn't even give you your three minute warning, because you were so tidy with your talk. You were on conclusions. Anyway. All right. Let me let me start with some questions here. And let's start with this one, which is very intriguing. Can we also examine fertilisation and
reproductive processes using live imaging like this? So I guess that, you know, fertilisation, yes, you probably can do it. But this is not the same approach. Quantitative analysis of pollen tube imaging is done already. Yeah. But what I think we could, what we will be able to do in the future is to for example, to look at the different development of the internal tissues. So for example, the germ lines, you
know, like, we hope that one day we'll be able to go inside the developing anther and to understand how those different layers and the cells etc. are forming, to generate to generate the germlines. Okay, and another question here. Thank you. Does the microdissection that I guess you have to do on the outer organs, change the physics component to which you referred earlier in the talk, I guess.
So, you you, of course, this is you know, you cannot call this technique non invasive, yeah, we cut the tip off the plant, we dissect it and so on. So, for sure you, you are expected to have some alterations, and we can in some time lapse imaging that we do, we can see, for example, that locules that are very elongated will be less elongated in the in vitro cultures than on the plant. But in general the growth the older part on components of the organs, especially if such a floral organ like like stamens are in place. So we believe that it is a good extrapolation, and we are not actually affecting so much the growth or the shape of the structure in in vitro cultures by removing other floral organs. Okay, you have a lot of questions here. Let me just ask
one of the first thing we have some time because you ended early. How do you maybe want to go into a little more detail of how you measure the growth speed in in your metrics that you show with the heat maps. Okay, so you have consecutive images, because our series are taken every 24 hours. We segment cells, specific cell that say at the time zero, and then we are looking at the progeny of those cells at time one. We have our additional MorphoGraphx after
can give you the basically the surface area of the cell at the beginning and their progeny at the end. And then in this way, you can calculate the area extension, in percent, you can calculate the growth anisotropy you can calculate how many cells originate from this one single cell etc. And you can do it also for very prolonged time, then we calculate how many cells do you get from one single cell at t zero? And then how many cells you get from it five days later. Okay. And then here's a question, and I was also wondering, you do you do this tracing back to the origin of when a cell gets its identity? But of course, it would be nice to see the the factors, the induction, the hormones or whatever, that are perhaps responsible. So the question
here is, have you tried looking at the proteins, I guess, fluorescent or otherwise, that are also, you know, in play during this growth process, Of course, for this specific research that is already published, we haven't. We have concentrated our efforts fully on quantifying the growth and proliferation. But yes, we are looking at this as well, you know, we are interested in especially on auxin signalling concerned how the how the auxin would be correlating with growth in organs not only stamens, but also in other organs? Because especially auxin is, on one side is a hormone that is involved in patterning of an organ, but at the same time, it's also involved in the differentiation of different tissues, so accelerating or deaccelerating differentiation. So yes, it is on our table, and we will be doing this in the future. And we are doing it already. Okay, good. And you just answered a couple of other questions in the answer to that. So you were very, you're very
efficient with the questions. Let let me move on to our last well, let me first thank you very much, Dan. That was a great talk really interesting. And a lot of the comments also reflect
that. And now I'll get to our last speaker. Jen Brophy. She's an assistant professor at Stanford, and she's also a Chan Zuckerberg Biohub investigator. She's been interested in synthetic biology for a while developing tools for engineering domesticated bacteria and soil microbes. In plants, she did spatial pattern engineering with synthetic circuits. And now in her own lab, she's interested in applying synthetic biology to both plants and bacteria, and in particular, interested in issues of climate change. And so Jen seems like she's ready to go. So
I'll turn it over to her. Hey, thank you, Ken, and hello everybody. I'm going to tell you about what I wrote in a review for this Plant Cell Atlas issue.
It is a perspective piece, presenting an idea of how we may be able to engineer some synthetic plant development. I just want to make it really clear at the beginning here that this is an idea of how we may go about engineering plant form, and that it might not work, but I thought it would be useful to write about so that we have the opportunity to discuss this potential approach. Okay, so, I want to start by talking about why we might want to engineer plant form. I think a lot of us are familiar with this idea that a lot of the foods that we eat today are derivatives of wild varieties in which we have changed the structure of the plant to increase both the size and some of the features of the edible portions of plants. So if
we can continue to engineer plant form, we may be able to do things like continue to boost yield. We think it's fun to consider other methods or other applications of engineering plant form, for example, waste reduction, can we take some of the forms that plants grow into, and reduce them down into shapes that are all suited for market so that we don't end up wasting any of the produce that we grow? Or can we change the growth programmes so that the plants that we produce are, are ideal for mechanised harvest. So I'm showing here saffron flowers, which are one of the most expensive food crops we have today, in part because it requires hand harvest of all of the portions of the plant that we like to eat. And so if we can change the structure of this flower, so that it presents the portions that we want to collect, in a way that make them amenable to mechanised harvest, we may be able to reduce the cost and labour that go into producing some of the food crops that we rely on. And then finally, maybe most dear to my heart is stress tolerance, we can change the structure of the plant, we may be able to produce varieties that are robust to environmental stress. So this is an example in which a root system may be a little bit too short to reach water, during droughts where the upper layers of soil have worn out, or dried out. And if we can engineer the root structure to reach water
that's deeper in the soil, we may have plants that are more robust to something like drought stress. So this is the why of engineer plant form. The question is how. So I'm going to present an idea of how to engineer plant form, but I just want to highlight some really recent and interesting methods for engineering plant form that have been really impactful. So first, there's been a number of papers both out of Cold Spring Harbour, as in Zach Lippman's lab and Dave Jackson's lab to engineer the promoter regions of important genes to change the structure of tomato and corn. There's also been efforts to express developmental genes and really specific spatial patterns. So that you can alter the structure of the resulting plants, and methods for kind of dynamically changing gene expression by delivering viruses that encode Cas9 activators, and repressors. And
those Cas9 activators and repressors, up or downregulate specific genes as the plant is growing and modify its structure. Now, this is all great, but we know that structure is encoded by a really complex network of genes. And so it becomes difficult to imagine how you might go about changing the structure especially because as you modify expression of one gene, you may affect several different aspects of development. So I want to introduce a concept of refactoring to tell you how I think we can do this. So
refactoring is a term borrowed from computer science, where you change the underlying code for a specific process, such that the process remains the same, but the code is different. And I think that this is interesting, because in biology and synthetic biology, specifically, we've applied this concept of refactoring to get control over complex biological systems. So in biology, refactoring means taking all of the genes that encode for a specific function, and recoding the DNA sequence for each of the proteins. So you can use the degeneracy in the
genetic code. So there are multiple codons that encode for each amino acid or most of the amino acids, which means that the very different DNA sequences can encode for the same protein. So you use that kind of flexibility in the genetic code to recode each of the proteins and then express them from promoters and other regulatory elements that are synthetic that you have designed and know the exact expression levels of and by doing this, you can then more easily modify expression of each of these proteins. And you don't have to worry about native regulation of the promoters that usually drive expression of these genes, or things like small RNAs or methylation patterns that may modify the DNA itself or the sequences themselves. So this process of taking a bunch of native genes and recoding them synthetically has been used to do a bunch of kind of cool things in bacterial systems. So we've been able to
engineer M13 phage, a specific type of bacteriaphage, to function as an anti tumour agent, to recode the entire gene cluster that makes type three secretion systems, which are what salmonella bacteria used to inject effector proteins into mammalian cells. And by recoding the whole cluster under these synthetic controllable elements, these authors were able to move the ability to make this type three secretion system from salmonella into other types of bacteria, and to secrete non pathogenic but other types of proteins through the needle. There are several other examples in which refactoring has been really useful for allowing us to modify complex biological systems. So I think that we can use a refactoring type approach
to try to engineer development and this refactoring type of approach is going to be based on auxin, or at least I think it would be cool if it was based on auxin. So auuxinis a key plant hormone in development, it is synthesised and then moved around the plant in ways that change or dictate the structure of the resulting plant. So this is just examples of, of different portions of the plant. And the yellow arrows are showing sort of where auxin is flowing within those tissues. And then some of the proteins that are responding to that auxin concentration change. So in tissues where there are cells where there's low auxin, I'm going to present just a couple of the molecular tools that respond to auxin. Cells that are
low auxin, genes that drive development, and some of them are kept off by the binding of auxin response factors, these ARF proteins to DNA, and the binding of Aux/IAA proteins to those ARF transcription factors, the Aux/IAA recruit repressors that keep the expression of the downstream genes off, and I should say like this is the canonical auxin response pathway. There are several variations on this pathway, but I'm just presenting to you the canonical one. So in tissues where there's high auxin, and this Aux/IAA protein that becomes bound to this TIR1 E3 ligase ubiquitin ligase, which then leads to its degradation. The degradation of the Aux/IAA frees up this ARF to recruit an RNA polymerase and initiate the transcription of downstream genes. So what is exciting is that we might be able to use mutant versions of these Auz/IAAs and ARFs, in order to control development. So if you mutate specific regions of these Aux/IAAs, you can make them immune to TIR1, and the presence of auxin, so that even when there's a high amount of auxin in the plant, these proteins will continue to repress the ARFs that they usually bind to and prevent the transcription of downstream genes. And sort of
similarly, but in an opposite direction, if you truncate off the region of the ARF that is bound by the Aux/IAA, you would have a version of this protein that cannot be repressed and therefore wherever it is expressed, it should initiate the transcription of its downstream genes. We are fortunate in that these mutations that would do each of these things have been well characterised for a large number of Aux/IAAs and ARF transcription factors. So it should be relatively straightforward to figure out how to modify these proteins in order to make them insensitive to auxin, and therefore useful for engineering development. Now, you have to be careful because messing with auxin response regulators can have a lot of pleiotropic effects. These are just two examples where you've got that kind of gain of function or auxin insensitive mutant version of an Aux/IAA, and that causes a bunch of developmental defects in this plant. So this is the the plant containing the mutant version.
And it is expressed from its native promoter, which is active in many cell types. And you can see that the plant doesn't produce any lateral roots, it has a hard time sensing gravity, doesn't have any root hairs, and has just a bunch of different pleiotropic effects. And this is an example of an ARF knockout where the root system fails to initiate which is obviously a very large phenotype and not super useful for being able to modify specific aspects of development. So what I think you
need to do is to carefully control the expression of mutant ARFs and Aux/IAAs in order to make specific changes to the development of the plant. This is just a little cartoon suggesting that you would have this synthetic upstream gene expression programme that eventually expresses an ARF or Aux/IAA that is insensitive to auxin. These are just two examples of how this may work. So in this example, you would be expressing the Aux/IAA mutant version in the xylem pole pericycle cells that would usually initiate lateral roots. But if you repress auxin signalling in those cell types, you should be able to make a plant that doesn't have any lateral roots. And restricting expression of the gene to just
those cell types might enable you to have a very specific effect on development. So you can modify lateral root branching without modifying other aspects of the plant. Here's just a different example, where you may affect gravitropism by expressing this same protein in a different tissue, in this case, in the epidermis, and elongation zone. So what's exciting I think about this strategy is that you wouldn't need to do a knockout. So in the refactoring pipeline
that I showed you earlier, when people take a bunch of genes that encode for a complex biological process, and then recode, and reintroduce them, they do it into a knockout background, you have to remove all of the native genes that typically encode for this process. But here, because each of the Aux/IAA and ARF mutant proteins are dominant, you should be able to introduce this into a wild type background, and modify development, which is nice because that means that the kind of plant material that you start with to do your engineering can be unmodified. Okay, so here's just an example of maybe a slightly more complex synthetic programme. This is taking a background, or the background plant that I showed you before, where there's no lateral roots being produced and showing how you might be able to reintroduce root branching under synthetic control, using the kind of timed expression of the ARF proteins that are known to be important for lateral root development. I should point out again, here, we really don't know if expressing these mutant ARF proteins, even in the right cell types in the right order would produce lateral roots, or if there are more aspects of the development that we would need to be able to control. But it's just an example to show you how
we might start to do these things. And my hope is that if you started to do this type of engineering that it would elucidate potentially ports of development that we we don't currently understand and don't know that we don't know. Okay, so why would we engineer auxin response to change plant development, there are a couple of reasons that I want to highlight that I have been sort of driving my interest in this space. The first is that auxin is sort of the key driver of
development, many of the environmental signals that alter a plant's development funnel down into auxin response regulators, whether that's modifying the expression of these regulators or their post transcriptional modifications, all kinds of different things. And if we can engineer plant development to proceed independently of these of auxin, then we can engineer it to proceed independently of these other environmental cues. And so you get something maybe that looks like this, where the amount of water and nutrients, the presence of microbes, the growth of the shoot system may usually impact the growth of the root system. But if we genetically determine the root system, so it always grows into the same shape, then we can start to probe structure-function relationships, and say, how does the shape of the root system affect the plant's ability to survive in a specific environment, for example.
And then you get situations like this where you in a wild type plant may have a root structure that grows differently in a bunch of different environments, but an engineered plant that always grows exactly the same root system can be used to see what the effect is on the shoot. And obviously, most of these plants will be very ill-adapted or unable to survive in these environments. But I think it would be really interesting to use this type of engineering in order to test some of our hypotheses about what ideal root structures look like in different environments. Okay, the other reason to try to engineer plant development with auxin response regulators is that it may enable the engineering of diverse plant species. So we can come up with a scheme for arabidopsis or for some of our model plants, where we show that the expression of a mutant Aux/IAA and ARFs can be used to modify development in sort of a predictive manner. Then we can look for Aux/IAAs and ARFs in other plant species, make the same type of modifications and then drive modifications to their development as well. And this is useful because these genes are
really highly conserved. And so unlike some of the developmental regulators that we have been able to modify in our model systems, which don't have homologues in these other species, Aux/IAAs and ARFs definitely do. In fact, genomic sequencing has revealed families of ARFs in all plant species, or all of the flowering plant species that we've looked at to date. And this is where this sort of ties back into the Plant Cell Atlas. So you'll notice that there are large families of ARFs and Aux/IAAs in each of these plant species, and it's really difficult to know which member of the family is important for the development of which structure in a plant. To a
certain extent, they're interchangeable, but they're not completely identical. And so you can use some single cell omics to determine which Aux/IAAs and ARFs may be expressed in the tissue that you're interested in modifying, and its expression or modification in response to different conditions or over developmental time, which give you some idea of how it's acting, and how to modify its expression in order to change development in a hopefully predictable manner. Okay, so that's my crazy scheme for how to modify plant development. Just in summary, the goal would be to engineer plant form by carefully controlling the expression of mutant auxin response regulators. And that you might want to use this to
test structure-function relationships, or produce some new plant varieties. I'm just showing here a cartoon of one of the kind of original goals, original, but very ambitious goals of synthetic biology, which is to engineer trees so that they grow into homes by modifying their their growth programme. So with that, I just want to say thank you, my lab just opened. So my thank you slide, it has just one person on
it. And but I also wanted to highlight Jose Dinneny and his lab, where I recently finished a postdoc, and got into plant biology. So that's it. Thank you. Okay, great, Jen. That's some really interesting food for thought. I'll start here with a question. So you're really you're focused on auxin, and clearly through that hormone, you can control a lot. But we know that it's got such
intricate feedback mechanisms. And so that works on both ends on the refactoring side, figuring out the entire pathway in any given process is going to be ... we don't have that information yet. Right. And then when you're, later on, when you're just manipulating a few tweaking parts of the auxin pathway, knowing the consequences is going to be hard to predict. So how do you how do you deal with it's kind of an open ended question, but how do we deal with that in this approach? Yeah, so one of the things that I like about these auxin response regulators is that they are so powerful and able to modify so many different aspects of development, that they should have an impact at, you know, kind of wherever you express them. And the trick is to refine that impact down into the change
in development that you want. And so we have some ideas based on like decades of auxin biology research, like which tissues are expressing which kind of auxin responsive transcription factors and things like that. So the idea is to start by testing some of the models of development that have been developed by expressing these things in a subset of tissues, and then looking at what the impact is, by doing that on the whole plant. And if you see that there are, you know, more pleiotropic effects than you were expecting, then you can maybe try to modify it the expression level even further by tuning not just, you know, the tissue in which it's expressed, but the point in development at which it's expressed. And, you know, quite honestly, I don't know that this would work. But I think it would be really fun to kind of put some of that knowledge to work and see, okay, you know, this is our model of how the development proceeds. If we try to introduce an auxin response regulator to
modify that development, and we we don't see what we were expecting, you know, can we do some transcriptomics to see okay, well, you know, in normal progression of development, you may know that auxin signalling activates this like set of genes. But when we introduce our truncated ARF for example, that is supposed to kind of mimic that auxin bolus that leads to gene expression, we see only a subset of those genes expressed or we see, you know, all the genes expressed that we were expecting plus a whole bunch more, it gives us an opportunity to start to try to maybe troubleshoot and refine and and broaden maybe our understanding of how these processes are working. Cool. So it's kind of also a experimental platform to learn more about the plants. Okay. Let me move on to some other questions. Here's an interesting one. Um, basically, I'll boil this down, is you sort of propose making either more simplified are replicating circuits, do you ever want to make them more complex? Yeah, I think if you, if you could get the simple things to work ... one of the reasons I focused on the simple stuff is
because even the simple things are very hard. But it would be great to see what would happen if you could modify sort of a simple circuit to kind of include more environmental inputs, or more decision making points that then give you you know, not just a modified version of the original plant, but a completely new form. I don't have the best idea at the moment of what that looks like. I think the challenge of trying to just, you know, recapitulate normal development using synthetic regulation is going to be challenging enough. And that we would learn a lot by modifying our expression programmes to change dynamics of these pathways or tissues in which they're expressed, which I think still constitute relatively simple circuits based by this kind of metric, but would be pretty informative. So maybe, if we get there, we'd love to see more complex stuff happen. Be awesome.
Okay, cool. All right, here's kind of a related idea, or thought, and that has to do with how do you, how do you deal or factor in or break apart the cross talk with other hormones and auxin? Yeah, so the refactoring process of recoding the genes is one way in which I was hoping to try to break the interaction between auxin and other hormones. So you know, the other hormones can lead to the expression of small RNAs that regulate auxin response factor and Aux/IAA levels. And, and also that they can lead to post translational modifications. So if you change, you know, residues within these mutant Aux/IAAs and ARFs that are known to be phosphorylated, simulated all these other things, and then you should start to wall off these regulators from most of the other inputs that would change their activity. And there's going to be a limit to this, but
it's a place to start, and may also help us figure out, you know, other mechanisms by which these other hormones are modifying development through ARFs and Aux/IAAs. Cool. Again, like another experimental, using it as an experimental system, the probing system. Cool. Okay. Ah, we have more questions, but ah, we're over time now. And I guess we promised to let everybody go. And I'm not sure if the speakers
can answer keep answering the questions in the q&a, but maybe they can. And so maybe they will. And I think Mary wants to close off with a couple of calls. Yes, thank you. And I can send the questions and chat to the speakers if they would like to continue answering them. And first of all, I'd like to say
thank you to Hilde, Daniel and Jenn. Great talks. And thank you, Kenneth for fielding the questions, expertly. Thank you, attendees. It was a lot of chat, a lot of questions. It was very, very stimulating. The recording will be posted in the next few
days. And this is really one that I think we want to all encourage our students to watch because, you know, there's no "plants arer boring" going on here. Right? This is exciting stuff. This is the future. I'm really excited. And so that's
it. Thank you, editors. Thank you Plant Phys. Thank you, ASPB. Thank you, everyone, and I'll see you next time. Bye. Bye. Thank you. Thank you. Thank you. Bye
2022-02-17 01:41
