 It's a modified leaf and it's been modified over evolutionary time to form this cup and the pitcher produces nectar around the edge and on the lid here that nectar attracts mainly insects which fall in and drown in the pool inside and inside there's this inside there's this whole little aquatic ecosystem so there's this pool that's formed partially from secretions from the plant but mostly from rainwater and in that pool there's a whole community of organisms that thrive so there are insects that are specialized on pitcher plant systems these aquatic insects there's mites there's rotifers there's protozoa there's bacteria there's fungal yeasts and there's in this system there's predators and prey there's decomposers there's primary producers so you have this whole dynamic ecosystem going on but it's on a smaller scale and there are many different reasons why I think these make great model ecosystems one is that you have this small easily defined area and that's one thing in community ecology that can be hard to find is that you have this this small sort of isolated community it's like a little aquatic island um there they have this tube-like shape so bringing them into the lab is not such a stretch you know it's not quite the same as when you have the ocean and you bring a little bit of the ocean and it's very different this is already sort of that contained tube-like shape it's naturally replicated so um in the habitat you can have many different pitchers that are growing on the same plant in the same environment and it has this host microbiome association so we think that in a lot of ways the picture community is sort of like a gut because it's it's helping its host to break down and access nutrients from its prey from its food so it's this nice kind of plant analog of an animal's gut and because it has this host microbiome association it also has some pretty clear ecosystem functions which can be difficult to find as well right what what functions really matter in a lot of ecosystems it's hard sometimes for us to determine that but in this system this breakdown and decomposition of insect prey is something that we know is really um allowing nutrients to enter into this aquatic pool and also for the host plant to be able to absorb so these plants grow in really low nutrient ecosystems and they get a lot of their nitrogen and phosphorus from the insects that they eat this is showing an evolutionary tree of plants of flowering plants um and mapped onto it are the separate evolutions of carnivory and plants so you can see that carnivory has evolved multiple times independently in plants and i want to highlight the pitcher plants so here are the napenthecie these are found in southeast asia there's many different species um these are the cephalotaceae this is a family that's found just in western australia and there's only a single species in this family and then there's the cerisinaceae and these are found in the americas and there are many different species as well um and these different groups are are in fundamentally different groups of plants and they've evolved completely independently but they have the same cup-shaped leaves um they have a lot of the same sort of colors the red and green coloration um and they have the same function of trapping and digesting insects i started this experiment by just trying to understand what organisms live in the pitchers who's there um this was relatively well known for insects but not so much on the microbial level and so uh we collected over 700 different samples in north america southeast asia and australia from pitcher plants and also from their surroundings so so from the bog water from the soil that surrounds them and we did meta barcoding of the bacteria and the eukaryotes and then for from a subset of our samples we did shotgun metagenomes to look at overall functional potential and what we found in terms of who's there is that for eukaryotes it's mostly these specialized mosquitoes, midges, mites and then fungi, protozoan algae and in terms of bacteria it's mostly protobacteria, bacteria dedes, actinobacteria and acetylbacteria so one of the questions i was interested in is if succession is really leading to change in these pitcher communities over time and this was a an experiment that i did with an undergrad um elizabeth van sen she was looking specifically in one species napenthes mirabilis and she started out by looking at pitchers that had not yet opened and before they opened their sterile or very close to sterile and then the community colonizes from the environment and so what she found was that consistent with this um in terms of both bacteria and eukaryotes there was sort of this increase that happened across weeks so she investigated the same pitchers over time from when they opened and then the following four weeks so in terms of just the richness we see this this increase happening over time and then in terms of the composition we see that there is this trajectory happening so here this is an nmds plot and the most important thing to know about these plots is that each point is representing the entire bacterial community from a single pitcher and then if these two points are close together those bacterial communities are more similar to each other if they're further apart they're more dissimilar so what you can see is that there's this sort of if there's a lot of variability within a week right but you also see that there's this progression happening over time so we do see this successional change in the community composition across weeks we also wanted to know to what extent environmental filtering may be altering these pitcher communities and we wanted to look at it on two different scales the first was at the genus level so looking among the different species within a single genus and then the other was looking at pitcher plants plants in general so sort of this pitcher plant habit and for this first part for looking across the different the two genera we focused on the napenthes and serocinia and i was comparing these eight different species of napenthes and these six different species of serocinia and so these are you can see they have different shapes and structures and coloration and i targeted locations where there were multiple species that were all growing together in the same habitat and you do find this with these plants that there's sites where you can have you know three or more different species growing together so what we did find was that within each genus the host species identity influences the pitcher communities and i wanted to know more about what might be driving this what we found was that certain characteristics of the pitcher's drive composition differences for the napenthes the the strongest driver was the ph or the acidity of the pitcher fluid and then apenthes can actually do this amazing thing where they excrete hydrogen ions into the fluid and really acidify it so you get these super acidic pitchers around 1.5 so here's an nmds plot and these are now colored by the acidity of the pitcher fluid and you can see that there's this progression from acidic to more neutral pitchers and a strong correlation between the composition of the community and the acidity of that pitcher fluid which might be expected in the serocinia they don't have these extremes in acidity and instead we see that shape and volume has a stronger effect so these some pitchers have these tall sort of narrow fluted shape and others are more squat and round and that had a stronger effect on the communities and i want to highlight that these both of these effects were much stronger than the effects of dispersal so for example we would find two pitchers that were right next to each other but if they had very different ph levels they would have very different communities if they had very different shapes they would have different communities whereas like a pitcher in massachusetts and in florida that has similar shape they tended to have more similar community so we do see that on this genus level environmental filtering alters these pitcher communities when you look on a kind of narrow scale we definitely find that these characteristics of particular pitchers have a really strong effect on the composition so i also wanted to look at the pitcher plant habit in general and i was comparing the napenthes and the serocinia again um and here i just want to highlight how distantly related these are to each other so they're as distantly related as cacti are to blueberries when people first found these plants they thought that maybe they came from the same family but then now we really know that this is a very clear example of convergent evolution and i was curious to what extent this convergent evolution might also drive convergence in terms of the communities that associate with them so what we found was that the napenthes and serocinia even though they're on opposite sides of the world um generally are colonized by organisms that come from the same phylogenic clades so here what i'm showing you is a evolutionary tree of bacteria and these are the bacteria that are found in across the study um on the outside here in brown are all the different bacterial species that are found in bog water or soil that was surrounding the pitcher plant habitat and then here in red and in blue are all the different bacteria that are found in at least 10 percent of either the napenthes or the serocinia so i wanted to focus on on organisms that were found repeatedly and commonly within these pictures that seem to be um associates instead of just sort of randomly showing up there from the environment and what you see is that they tend to come from these same groups when you look really broadly um you don't necessarily have the same asv's or species um but you do see that they generally come from these similar groups and you see the same thing for eukaryotes although it's a little more restricted here um you see it for the histiomated mites the dipterin insects and the rotifers so you see different organisms but they come from these same groups and this suggests that potentially they fulfill similar functions within these little aquatic pools and to look a little bit more at potential function we hypothesized that these pitcher plants would be enriched in genes for prey degradation so we looked at chitonase genes since chiton is the main biopolymer in these insect exoskeletons and we did find that in the metagenomes of the napenthes and the serocinia they were enriched in chitonases as compared to other published metagenomes from phylospheres lake and soil habitats and then conversely we hypothesized that they would be depleted in enzymes for degrading cellulose because cellulose is plant material and we thought that these these pitcher plant communities are not degrading plant material as much as they are degrading insect material and we did find that in general they were depleted so it does suggest that we may be seeing sort of a convergence in the types of interactions that appear on this very broad scale looking at the pitcher plant habit and the organisms that associate with it so we see it on both of these scales and if we come back to this slide what we've seen so far is that succession does have an impact on changing the composition of these communities over time and environmental filtering plays a really large role in what organisms are surviving and thriving within pitchers and then dispersal was not as important so we did see some effects of dispersal and of distance but it was not it did not have as large of an impact as as this environmental filtering did I can pause here for a minute does anyone have questions I don't see any question in the chat or in the participant list so I think we can move on and so I really wanted to also think about historical contingency and it's a little bit more difficult to examine because you can imagine that if I was trying to just look for it out in the wild um you don't really know the history of these different pitchers you don't know sort of what slightly slight differences in environmental characteristics all right I was thinking I was interested in asking if historical contingency is also affecting the composition and function in pitcher communities the common perception for bacterial communities is that environmental filtering dominates and I think part of the reason that this is the common perception is that it is much easier to measure and also composition and function have been found to converge under the same environmental conditions but a lot of times this is because the experiments that are really controlled so you need a really controlled experiment to be able to understand if if contingencies are playing a role but in a lot of these controlled experiments they've been done in artificial media with single or few carbon sources and I think that this also restricts your ability to see these these effects so I worked on this experiment with Gabriel Leventhal and Manny Grolka we were all postdocs in Otto Cordero's lab at the time and we started out with 10 unique pitchers these were all coming from the same species now and from the same bog and fluid filtered the fluid through three micron filters to be able to focus on the bacterial component of the community and then use this sterilized acidified ground cricket and water so I wanted to have a really complex medium that was also really similar to what these communities would be consuming in their natural habitat then I would grow them for three days and every three days I would take half the volume and move it to a new tube and did this 21 times over 63 days during the experiment we measured the relative abundances with metabarcoating the carbon dioxide production with a micro rasp the functional fingerprint with these eco plates which looks at about 30 different substrates to see how the communities can metabolize them and then we also looked at chitinase activity because we already knew how to identify chitinases as an important ecosystem function in this in these types of communities what we found was that these communities equilibrate over time so this is showing that the dissimilarity of the communities between adjacent days was pretty high at the beginning and then decreased over time so the communities became more similar to each other over adjacent days in the experiment but even though they equilibrated around here after about three weeks in the lab we found that there were still differences in terms of the richness of these communities so they were different in the effective number of species that were present and there were pretty large differences and it wasn't just the richness it was also the composition so these communities were still composed of very different species and I'm not showing the family level here but we also saw differences of the family level not as extreme as the differences at the species level but but there were there was not complete convergence at the family level so what we found was that the historical contingencies do have a lasting effect on the community composition and here I'm showing an NMDS plot where the the sort of beginning point of the experiment is here next to the microcosm name and then you can see over time there's this initial large shift and this initial large shift is due to environmental selection because basically we're taking these out of pictures and we're moving them into a lab environment so very quickly there's this fast initial change and that's because these communities are adapting now to a pretty different environment we're trying to keep it as similar as as it was to its picture but obviously it's going to be different when you move into a lab but now all these communities are you know getting exactly the same food source they have exactly the same temperature they're in the same volume they have the same light and so despite that we see that there are still these persistent differences in the communities and these are due to historical contingency because all the conditions are being held exactly the same now so what's left is sort of the history of these different communities so what about community function um we found when we looked at carbon dioxide that the carbon dioxide did converge over time so this is the variance and the percent CO2 produced across different microcosms at the beginning it was higher and then it very quickly decreased and became similar over time so sort of overall they had similar levels of metabolic activity but historical contingency affects the more specific functions so when you look at this functional fingerprint from the eco plates we see that these distinct differences remain at the end of the experiment here we're seeing composition in the circles and function are the triangles this is a percrustes plot they're both break herds to similarities and then we see you know to what extent do they correlate with each other and you actually see that function strongly correlates with the composition in these communities so these the compositional differences remain and these functional differences remain and in terms of chitonese says this is also true so here i'm showing you five out of the 10 different communities and both mo3 and mo9 had really high chitonese production um compared with the other communities and then we we had cultured um some of these strains so we cultured a bunch of the strains from these five different communities measured their chitonese activity alone as a just um as an individual strain and then mapped that back to the corresponding asvs and what you can see is that for mo3 it was really one strain that contributed to this high chitonese activity whereas in mo4 there were three different strains that each had really high chitonese activity and together they contributed to this high activity of the community so i thought this was cool because it means that you can take the the activity from the individual strains and that this can be directly related to the community function we think that species interactions are contributing to the historical contingency and this is because we found that asvs that are shared or species so asvs or species that are shared among different microcosms have different dynamics depending on which microcosm they're in and one thing here i'm going to show you is that the the trajectories the sort of uh trajectories of their dynamics over time are more similar when the communities that they are in are more similar so we saw this interesting correlation that if they have a more similar community they also have more similar dynamics within that community and um and this seems to relate to species interactions that are likely happening within these complex communities i wanted to highlight a couple other interesting findings from this experiment because i think they're related to some of the things we're talking about in in this workshop one thing we found was that the early richness of the community predicted the final richness so not necessarily on day zero because on day zero i think we had a lot of things that were not metabolically active or that were not adapted to living in this lab environment but on day three so after the communities had been in the lab just for one transfer we found that they were strongly correlated with the richness at the end of the experiment and if we normalized the richness of the communities by the richness on day three and then looked at how that changed over time you can see that they equilibrate at this universal rate so we do see these strong differences in the um the richness of these different communities and we see that species interactions likely matter but they do also seem to equilibrate at this they lose as these over time at this stable rate we also found that the community assembly trajectory was really reproducible so this was using filtered and unfiltered samples that were in the same condition so i told you that i'd filtered the you know the primary samples for this experiment were filtered but i also had a set that were unfiltered because i was curious about protozoa and what would happen to them in this lab context and what we found was that if you compare the filtered and the unfiltered communities they actually follow really similar trajectories when you look at a break artist similarity so this is exciting to me because i think it suggests that if you had enough information about the the history of communities then you could really predict how they change over time when they come into the same environment so what we see here was that um historical contingencies can also play a role in determining the composition of these communities and can affect ecosystem function so one thing that maybe you've noticed i left out was evolution and we continued this experiment to look at community dynamics over evolutionary time scales um this was led by Akshit Goyal at MIT and um we've just posted the pre-print um this month earlier this month so we continued this experiment for about a year looking at more than 300 generations and did additional deep metagenomic sequencing and mapped back to genomes from the strains that i had cultured from these different communities we found that most of the variation and most of the evolution arose from pre-existing genetic variants belonging to the same species so there were these strains that you could not tell apart by their 16 s um but they actually played large roles in in the change in these communities over time so i'm going to just briefly talk about one aspect of this paper and then i'll end but we found that even highly related strains that were about 100 snips apart can decouple in their dynamics so this is an example of what it looks like when you have coupled major and minor strains of the same species and this is what it might look like if you have decoupled strains and so the majority were coupled but there was this subset that were decoupled and these decoupled ones this is was the was the where basically it split so at about 100 snips is where we started to see differences in the trajectories of these these strains and this is an example of a consequence of the strain decoupling so this is one ASV and this major strain stays pretty similar over time but you see here that after you know around 200 generations this minor strain started growing we found that most of the community interactions happening were at the strain level so this is looking at the interaction strength or the species species correlation and this is the strain strain correlation or interaction strength so the majority were happening at the strain level and this is an example of these hidden interactions between minor strains so when you look at just the species or the ASVs here you don't see any correlation but if you look at these minor strains they had a pretty strong negative correlation so they were highly negatively correlated if you want to know more about this study there's a lot more to it also looking at different the different gene families and gene groups that were that were behind these dynamics you can look at this pre-print so i want to end just by coming back to this and thinking about these really complex and interesting microbial communities that can drive a lot of important dynamics happening in our ecosystems and in our own bodies and i hope i've convinced you that picture plants can be used as these cool model ecosystems to be able to better understand some of these processes that may be behind these these interesting dynamics i want to thank the labs where this work was done many places gave me permissions to do field work and funding sources and with that thank all of you for listening and i can answer some questions great thanks a lot for the very nice talk there are in fact a few questions in the chat but yes i can start from those and then if someone wants to ask a question can raise the end and talk so martina is asking do you see any evidence of gene loss in species evolving in the same picture yeah so there was some evidence of gene loss and part of it was or you mean in the same picture or in this in this in the microcosms in the lab environment in the microcosm in the evolution experiment yeah so there was some um there was some loss of phage um resistance and potentially that might be because phages we also you know didn't directly measure phages but um but in them in the metagenomes we didn't see a lot of activity of phages there so um so there seemed to be some loss in sort of defensive um genes and let's see there was another group but i'm forgetting it off the top of my head right now but yeah we did see some particular gene loss happening uh great there is a question from Simon maybe please yeah thank you very much i have two questions one of them is um all the experiments you did at least in the lab would seem to suggest that this is an ideal system for thinking about a metapopulation approach with some aspects of island biogeography within the package and maybe taking into account uh how the picture plants grow over time and whether the community changes just like a patch dynamic approach so i wondered whether you thought about taking a modeling approach incorporating that and and the second question is to what extent can you look at um phylogenies of these species that are occupying the plants and plants itself yeah yeah those are really interesting questions um in terms of thinking about the metapopulation dynamics and the patch dynamics yeah i think that would be really interesting and i think one way to do that would be to have these sorts of separate communities in the lab but have some level some sort of low level dispersal happening among them and one of the first experiments we've done that kind of addresses that is we did um co-lesscence experiments where we combined all of these different communities in in all the possible combinations and then looked at how that changed over time and i'm still analyzing that data so i haven't dug into it that much yet but that was one way of thinking about but that's like a very large scale dispersal and i think it would be really cool to be able to think more about um yeah in the field yeah yeah i think you could also do it in the field yeah it's it would definitely be harder to control but you can yeah i think that would be a great way as well i think about co-evolution of different species so i think it's really hard to measure right because you want to see if there's reciprocal change in different species happening over time and that reciprocal change can be can be difficult to find um but i i would love to look more into that and i think that the picture so the picture plants each picture is an individual leaf and in some they only last a couple of months in some they last up to two years depending on the species so i think that there is time potentially in a picture to have co-evolution of the organisms inside happening and then there may be some potential to have co-evolution happening between the organisms that are associated with the picture and the plant obviously the timescale of the lifespan is really different um and so i meant over a longer period of time by looking at the phylogeny yeah so i think that you could probably my guess is that you would probably see something like that happening with the insect associates because you have these insects that seem really specialized on the picture habitat their particular characteristics of like the mosquitoes no longer are blood feeding that that tend to be picture inhabitants um so they have sort of different characteristics from from what you might expect um yeah i would love to be able to look at that more it's something that i've talked about a lot with nomi pierce particularly with the insects but i don't know so much with the microbes because you don't have this persistent vertical transmission happening it seems like a lot of the microbes are coming from the surrounding environment so may also have other habitats that they're thriving in they're not so specialized as the insect inhabitants for example thank you very much thanks for your question yes i think we have time for a very quick question from Mercedes yes thank you very fascinating to see in the evolutionary experiments the different levels of organization the importance of within species variation to the community which i think is something we don't think enough in community ecology my question was have you found any relationship between strain diversity um the the size the the sort of diversity of the assemble community and also how these change over succession or whatever what how they're changing time so you're thinking that maybe in communities that have fewer species they have more strains no the other way around that somehow there is a synergy between the diversification within the species and and the richness of the assembly so and perhaps a feedback in both directions which would kind of lead us back to the question on the on the species pool at the at the larger level right yeah yeah that's a good question i i haven't dug into that i should look into that more yeah actually it was leading the um the analysis for the strain dynamics so i'll ask him for for his tables because i don't have all the details of exactly how many we're in it would be fascinating because it's a very unexplored area which i think is super important this kind of you know and potentially interactions at these level different levels of organization yeah thank you that's a great idea