 because we are already a little bit late. So next speaker is Mike Brockhurst and the title is exactly the same as before. Plasmid mediated horizontal gene transfer in microbiomes. Great, thank you. Thanks for the invitation. Just to pick up Connie's question, yes we have measured fitness costs on the rhizosphere. Yeah the plasmids are costly, the bacteria do grow and the compensatory mutations do compensate for costs on plant roots. So yeah we can make that up later if you like. So yeah I'm going to talk about our work looking at more complex communities. So as you'll pick up throughout the talk I collaborate a lot with Jamie and some of this this all this work is on the same system as Jamie's and actually I had to throw this talk together quite quickly realizing what Jamie was talking about and that we were in the same session. So excuse me if this is a bit rough and ready but it allows me to talk about some new data that we haven't published yet. Great so this is just an opening slide just to thank everyone involved in the work. I'm recording, do you want to be recorded since you are saying it's new data or do you want to stop? It's fine it's on bioarchive so it's fine. Ellie and Jamie really developed this system when they were in my lab and then a few other PhD students and postdocs have picked up the system and here are some other collaborators particularly Cat Coyt who's done some modeling which I'll talk about today. Okay so I don't need to explain plasmids as being important but you know they really are important in terms of bacterial evolution in terms of carrying genes around and evolutionary innovation as we all know and we all know what plasmids are so they sit separate from the chromosome and they're semi-autonomous and they replicate themselves. So this autonomous nature is really interesting as an evolutionary biologist because it means you can have different fitness interests between the chromosome and the plasmid. We can have these context dependent fitness benefits as we've heard about so in a drug-containing environment an antibiotic resistance gene is beneficial whereas we can think possibly about some constitutive fitness costs of plasmid carriage that need to be outweighed for the plasmid to be beneficial and part of the reason why the conflicts of fitness interests between chromosomes and plasmids can come about is because plasmids are capable both of vertical and horizontal transmission at least when they're mobilizable or conjugative and I'll talk only really about conjugative plasmids today. So that means we can get different levels of selection operating in populations on the plasmid versus the chromosome giving them different evolutionary properties. Also as Jamie touched upon plasmids act as vehicles for other mobile elements and so often how transposed ones nested within them which themselves can have their own fitness interests can hit rides on plasmids and so we can think you have this kind of Russian doll model that people have proposed in terms of understanding the interactions between mobile elements that become nested within plasmids and I'll sort of think about those ideas as we go through. As Jamie mentioned the system we work on at least were isolated from this sugar beet field they isolated loads of plasmids actually and the way they captured them was by them encoding mercury resistance and so just from the original capture of these plasmids we see something really interesting about their evolution so these divergent plasmids all encode basically the same transposon providing mercury resistance and so even in the field we know that that transposon is probably jumping around hitching rides on different plasmids and that was revealed by this exogenous capture experiment back in the 90s by Andrew Lee. So one thing that's interesting about the collection is that most of the plasmids are pretty big and they encode a bunch of other stuff that we don't really know anything about and we focus pretty much exclusively on the mercury resistance trait. So as Jamie explained we can measure the fitness costs of these plasmids and effectively what happens is in the absence of mercury we see these quite large fitness costs but variable among the plasmids and then as you increase the amount of mercury ions in the medium then you can recover a fitness cost for plasmid carriage. So in this region of parameter space we see a benefit of plasmid carriage and we expect them to be on the positive selection at least at least initially whereas here we have a cost and we expect the plasmids to be on the negative selection. So the first thing I'm going to sort of take you through is how the mode of transmission of plasmids varies between those two conditions where you have positive selection for the plasmid encoded trait versus negative selection for the plasmid encoded trait and so we can do that as a very simple experiment with a competition essentially a mixture of strains half of which have the plasmid half of which don't but they have different markers and then we can track transfer of the plasmid over multiple rounds of growth. So the first thing we can see with that experiment which we do across this gradient of mercury selection is that plasmids are able to spread under a broader range of selective conditions than the same genes encoded on the chromosome. So you've shown me on the bottom the data for the chromosome encoded genes and on the top the data for the plasmid encoded genes across three different levels of mercury so yellow is where there is none and then the two blues are where we've got some mercury present and so the chromosomal copy of mercury resistance genes spreads only in the presence of mercury and in the absence of mercury it just stays at 50% whereas the plasmid encoded version can spread where mercury is around and also spread to fixation where mercury isn't around and that's due to the this ability of plasmids to cause infectious transmission and if we look in a bit more detail in those plasmid containing populations so this is just showing you the six replicates here and the three different levels of mercury and showing you in the lines the original donors shaded for plasmid carriage with the solid line and the pink and the original recipients in the dotted line and shaded when they become transconjugates in blue then we can see that in the absence of mercury the plasmid essentially spreads completely through horizontal transmission so the plasmid barriers that we see at the end are basically all transconjugates that gained the plasmid during the experiment whereas when we select with mercury we see global expansion of the original plasmid donor and so you move as you move from this negative to positive selection regime you move from a high horizontal gene transfer regime where you've got lots of infectious transmission and conjugation maintaining the plasmid into a low horizontal gene transfer regime where the plasmid essentially spreads via vertical transmission and so that can be important under more complex selection regimes so this is quite simple we've had just constant selection at different levels but we can look also at pulse selection at different frequencies of imposing mercury selection for those mercury resistance genes and so if you do that in populations where we've introduced a different plasmid than I've just the data I've just shown you so in this case PQBR103 that Jamie mentioned which is a very large plasmid with quite a low conjugation rate and then just showing you some transfer experiments here where we transferred populations for 30 odd transfers the plasmid started at 50 percent in the population and when we imposed continual selection for the plasmid it rose high frequency and then it was sustained when we withdrew selection so the gray is showing you when we added mercury and that's a productive compensatory evolution here when we had slightly less frequent selection for the plasmid we still saw this robust maintenance of the plasmid but when we didn't select for the plasmid to begin with then it felt a very low and sometimes undetectable frequencies but then when we imposed a pulse mercury selection later in the experiment then those very rare individuals that had the plasmid swept high frequency under positive selection but then did subsequently and slowly decline but what's interesting is that the only regime where we really saw appreciable levels of transconduct against in the environment was this regime where we had a very rare selection for the plasmid so showing you again that when you're in a rare positive selection regime then or then essentially conjugation becomes much more important and you enter a more of a high horizontal gene transfer regime um okay so that's just sort of thinking about populations how selection varies the mode of inheritance of a plasmid and also how the frequency of selection for the plasmid alters um relative importance of those modes of inheritance so showing that when you have less positive selection infectious transmission becomes more important but also when you have even rare selection infectious transmission is really important as a mode of maintaining the plasmid in the absence of of any benefit that that plasmid encodes okay so Jamie did some really nice experiments following up that two species system that he talked about just now where we looked at much more diverse kind of natural soil communities these were actually washes that we took from the potting soil that we use in our experiments and so I'm not going to go into much detail here simply to say that we introduced two modes of fluorescence with different plasmids and then we tried to detect those plasmids in the background community after seven days so to do that we used this technique called um epic PCR where essentially you make concatenated PCR products in an in an emulsion so the emulsion ensures that you're only doing that on single cells and you're essentially able to link a taxonomic marker to um the 16s gene sorry you lick the 16s gene to a gene of interest so 16s is your taxonomic marker and you're trying to track the plasmid so by doing that we were able to um detect transmission of two of our plasmids um into members of this community these experiments were really hard to do um and this technique is quite challenging but so this top plot here is showing you the frequency of the proportion of reeds that were due to spw 25 in yellow and then in white showing you the proportion of reeds in the epic PCR that were essentially transconjugate so other species that had gained the plasmid and then showing you some some kind of feel for the diversity of those other species that gained the plasmid and say they're mostly close relatives of the pseudomonas but we do see some surprisingly distantly related species so we're a little bit wary about this data um just because of the difficulties of doing epic PCR but this is really to prove that this these plasmids do disseminate broadly um even in this relatively simple soil community although there's lots of species there but it's a lab a lab system that we've devised um rather than a completely natural system okay so I moved to Manchester recently and at Manchester is this person here he's called um Catcoyt and she has done over the last few years some amazing modeling so she's developed a whole body of ecological theory for understanding the microbiomes complex microbiomes and so most of her work is focused on the human microbiome but I persuaded her to think about our kind of simpler soil microbiome and so what her models typically do um is model lots of different species and then embed a kind of network of interactions among those species and so those interactions might be positive interactions shown by blue arrow over here or negative interactions shown by the the red the red bar there so she can model complex microbiomes that have different levels of negative versus positive interactions in them and then she overlaid onto that model uh essentially a resistance gene and that resistance gene could have different levels of mobility um so it could spread within a species or it could spread between species and so that's effectively modeling a plasmid um in this complex microbiome model so what we've used this model to look at um is how plasmids affect the stability of complex microbiome communities and so we measure stability um essentially is the change in abundance so um it's just a log of this this value here so abundance after exposure to an antimicrobial divided by the abundance before exposure um so we can measure stability in our model um we do that under a bunch of different conditions so we do it in the absence of the resistance gene and in the presence of the resistance gene and we have a we either allow some mild exposure to this antimicrobial stress prior to the perturbation so you could you either have communities that have some mild exposure to the stress or not and then we expose them to this very high level of the stress and we look at their stability we then calculated two metrics so delta r is essentially the change in stability due to a resistance gene being present in the community whereas delta e is the change in stability due to that prior selection week selection with the with the stress um and then we measured three different compartments of the community so we measured stability of the entire community we measured stability just of the background species so these are the species that acts as recipients for the plasmid and don't have the plasmid to begin with and don't have any resistance genes um and then we also measured stability of the focal species so that's just to explain how the model works so what we then do is model the susceptible community and then we model a community where we different communities where we add resistance genes um and then we have a another parameter that controls the mobility of those resistance genes so how likely are they to transmit to other cells and other species in our community um and so i'm just going to show you first off very simple model where we don't have any species interactions going on so this is kind of an ecology free model that's just looking at the properties of the system um when you when you have different levels of mobility okay so if we look first at delta r so this is the the increase in stability due to the resistance gene um then as you increase mobility you see an increase in total community stability and you see an increase in background community stability and basically no effects on the focal species which already has the resistance genes so remember this is this is stability here so we're measuring the ability of them to survive this very strong persuasion so this makes sense as you increase mobility the genes that provide resistance to the persuasion are more able to invade the community and so they provide a broader stability they allow more of the species to survive the persuasion and we can see that effect by looking at the the percentage of resistant cells in the the community um and so yeah as you increase mobility you see more and more resistance in the community um and prior exposure allows the resistance genes on the plasmid to invade lower levels of mobility so they're they're effectively driving driving those those plasmid borne genes into the community at slightly lower levels of mobility um and then finally thinking about this delta e property so this is the benefit this is the the change in stability due to prior exposure and so we see that prior exposure is most important that these intermediate mobility levels so that's effectively capturing this difference here so where prior exposure selects for um genes that become beneficial upon the perturbation later in in the experiment so they're driving these genes through lower levels of mobility driving them into the community at lower levels of mobility so next cap took that model and she put in all of the network of species interactions and so we can test how horizontal gene transfer by plasmids affects um stability of different kinds of microbial communities so she does that by changing the level the percentage of positive interactions in the community so over here we've got a just a competitive community and over here we've got a cooperative community and you can have everything in between um okay so firstly just looking at um the total community so this is the entire community and what we're plotting here are the percentage of positive interactions in that community and then again um ability of the resistance gene um and so this is showing you delta r so the the change in stability due to the resistance gene and you can see that in the total community adding resistance genes is usually good so it increases the stability of the total community and it does so most strongly in the presence of highly mobile resistance genes in a cooperative community um and in terms of delta e we see again that the the effect of prior selection is basically felt most strongly at these intermediate mobility levels so it's more interesting when you look at the background community so here we see um some interesting effects of the network of interactions in the community so as the community becomes more competitive um then we see that introducing resistance genes with low mobility actually destabilizes the background community and so this is because these cells have a really the cells that we've introduced with resistance genes have a really strong advantage when they're surviving because they can survive this perturbation um they don't tend to give those resistance genes away and so they they experience really great competitive release when there's this perturbation and they're able to out compete everything um after the so upon the perturbation but where you have um highly mobile resistance genes and where you have a more cooperative community then you still see these strong benefits in the background community of introducing resistance genes particularly when they're mobile um and then we look when we look at the focal species so this is the where we've the species that originally carries the resistance genes we see this quite interesting effect of prior selection which is where you can see a strong cost in terms of stability for the focal species so this is where in competitive communities there's the cost of giving your resistance genes away to other species so essentially you're no longer benefiting from the competitive release the benefit of being resistance in the absence of non-resistant competitors so finally just to wrap up um I know I've run out of time but we tested this idea in our experimental system so we took a community in soil we introduced our resistance genes either on the chromosome or on two different plasmids and then we transferred them and exposed them to a moat preprospation with or without this weak prior selection so what we see is that essentially adding resistance genes does improve the stability of the total community um but much more interestingly if we look at the background species we can see that introducing chromosomal resistance genes strongly decreases the stability of the background community as predicted by the model um whereas introducing mobile resistance genes actually increases the stability of the background community and so as our model predicted that's because these genes these resistance genes when they're mobile on plasmids move into the background community um and so if we look before and after the pulse then we can see strong selection and increase in those resistance genes in the background and then finally this neat little detail in the focal species that if they give away their resistance genes because they're mobile then we see this weakening of um focal species stability uh when you actually select for those genes before the perspiration okay so just to sum up infectious transmission of plasmids is really important it enables resistance genes to survive and spread without positive selection it increases the stability of microbiomes um against perturbation by antimicrobials but when they encode resistance genes but the precise effects of that depend on the interplay between mobility and the network of interactions in the community thanks for listening thank you very much Michael a okay see if we have questions I don't know if Olivia wants to speak up herself or should I read the question is I find the pattern of rare selection being associated with higher horizontal gene transfer very interesting do you have an explanation or hypothesis for this observation so yes our explanation for that is um in that experiment the the plasmid has a low conjugation rate and in when you don't select for it it declines the very low frequency but it doesn't go extinct and the reason it doesn't go extinct is because that low level of conjugation allows it to persist in the population even though it's costly um and so it's essentially that very low level conjugation and persistence that then allows those very rare cells to reinvade um the community when you select for the genes on the plasmid um and so what's interesting is that that declined to very very low numbers so we couldn't even detect them when we were placing out neat cultures they were so rare um that that decline essentially prevents compensatory evolution happening because the the cells are too rare for them to acquire a compensatory mutation and so at that point conjugation and transmission is really the only game in town for plasmids to survive um in those situations where they're costly and they're not being selected and so um yeah the the answer there is is really that they persist at extremely low levels through conjugation um and just I supposed to follow up Dan Rankin did a model many years ago that sort of predicted that that the very rare pulses of um selection for plasmids could or selection for genes encoded on plasmids could select for their mobility so um that was a nice vindication to his model okay we have one more question by Marco Consentino uh hi I very nice talk um I have a question so does your experimental system would your experimental system allow to address the question of the reduction in diversity in a community when you introduce a beneficial trait uh through a plasmid because this will be the subject of my talk tomorrow but it's very theoretical but okay uh yes it does allow that so um so to explain the the the papers on bioarchive but to explain why it can test that the um when we have prior selection in that system then um some of the slightly unexpected results to do with stability are essentially to do with um the the community being all of the very very susceptible species being selected out of that community and diversity declining um in the presence of resistance genes so yeah that it's a slightly convoluted explanation but yeah it would we could design an experiment to explicitly look at that in a slightly more sophisticated way than than that experiment but um but yeah it would allow that definitely very interesting so you could measure the the amount of taxa as a function of time for example after you introduce a beneficial yeah yeah we didn't track it in enough um at high enough frequency to look at the rate of loss of species and things like that but um but it wouldn't be checked that challenging to do that okay very interesting thanks okay