 because it's working? I think so, yeah. So I'm going to be talking about a model of an RNA that we've owned, that is implicated in antibiotic tolerance, which is a precursor to the development of full antibiotic resistance. I'm sure everyone's heard of AMR, or antimicrobial resistance, in which 2019 was just under 5 million deaths globally. Byddwch y bydd y 50 ychydig yn cymryd a chyfnodd ar 10 miliwn. Yn ymgyrch yn ymgyrch oherwydd unrhyw ychydig. Felly, ymgyrch o'r ddweud i gyd, rydym ni'n wneud yn ymgyrch o'r ddweud. Felly, bydd Covid, ydy'r 2019, mae'n 3.7 miliwn. Felly, mae'n gwybod bod ysgrifennu gwirioneddol yn ymgyrch. Felly, rydyn ni'n gwirioneddol ar y cyfnodd angharthu. Felly, mae'n gwirioneddol ar y cyfnodd... Ymgyrch ar y cyfnodd... Ows, achos though we have a group of bacterial cells and some of them can be susceptible and some of them can just randomly be tolerant, so as tolerance can arise randomly. When you add an antibiotic obviously the susceptible cells will die and the tolerant cells will remain alive. So one thing about tolerance is that it's transient meaning that the cells can flip between a tolerant state and a susceptible state. Obviously if they all flip to susceptible then the antibiotic treatment will be successful, however if we still remain with tolerant cells the treatment won't be successful and the longer these tolerant cells stay alive then the more likely they are to acquire an actual genetic mutation which makes them fully resistant. So tolerance is, yeah, it's reversible, it's a transient process but resistance is an irreversible process so once we've reached this stage it's too late. So we want to learn more about this tolerance and why it occurs to hopefully stop the development of full blown antibiotic resistance. So we're studying this system called the RTC system which causes this kind of tolerance upon exposure to ribosome targeting antibiotics. So on the left here you can see a plot of fluorescence which is representing growth and time down here and we have the black line represents a non-inducing antibiotic so when we introduce the antibiotic at time zero the cells start to die and they don't really rescue their growth because the RTC system hasn't been activated. However in these three activating antibiotics which are all targets of ribosomal RNA we see that especially in the gentamicin and chloramphenyl you see that the on antibiotic exposure again just like the non-inducing antibiotic the cells die but they're actually after about four to six hours they're able to rescue their growth. So that's like a key feature of this RTC induced tolerance is that cells can rescue their growth after a period of antibiotic exposure. So what actually is the RTC system and why do we see this happen? So firstly it's really highly conserved across all domains of life so it's essential and it's an RNA repair system. So firstly why repair RNAs like they're quite transient molecules especially like mRNAs however other RNAs are not and they don't have such a high turnover rate such as ribosomal RNAs and transfer RNAs. So it does actually make sense for the cell to repair an RNA rather than just make new ones. So the RTC system has two RNA repair proteins there's an RTC A which is an RNA cyclase and RTC B which is an RNA ligase. So I'm going to go through the mechanisms of actions of these two proteins and then introduce that to the bigger picture of what we see with the system. So this is a bit of RNA with it's a healthy bit of RNA with a phosphodiester bond making up the backbone. This RNA could be damaged and this bond could break. So now we're left with a fraction of RNA with a 3' phosphate end. So that's the substrate for the RTC A reaction which as it's a cyclase it converts a 3' phosphate to a 2' 3' cyclic phosphate. Now this cyclic phosphate is one of the substrates for the RTC B reaction. So the RTC B protein is a ligase and it basically ligates or joins together a cyclic phosphate end with a 5' hydroxy end of RNA. And that repair is always perfect and just reforms this phosphodiester bond perfectly and we then have another healthy bit of RNA. So that's the mechanisms of both RTC A and RTC B. If we think back to what I was saying before about how this system is activated in exposure to ribosome targeting antibiotics. We come up with a hypothesis that the damage to get from this perfect bit of RNA to the bond breaking is caused by ribosome targeting antibiotics. And secondly that each bit each one of these RNAs here represents a type of ribosome that we're. So the one with the phosphodiester bond that is perfect and healthy is labelled as a healthy ribosome and these are going to make up the model which you'll see in a second. So the one with the 3' phosphate is what we're calling a damaged ribosome and then the one with a cyclic phosphate is what we've labelled as a tagged ribosome. And that will be clear as why we've called it that in a little bit. So back to like RNA repair. If we think about this system like this and the fact that we're now thinking of these RNAs as part of the bigger ribosome. It makes way more sense for a cell to repair a tiny portion of part of the ribosome. Then it's way more energy for the cell to make a whole new ribosome. So that's why we've hypothesised this in this way. So now talking about the expression of the RTC proteins. So RTC and B are co-expressed from a really tightly controlled promoter. And that because of this really tight regulation it requires the regulator protein which is the third protein in the RTC system which is RTCR. So this protein is constitutively expressed from just like a housekeeping op-on. And so RTCR first requires its own activation to then initiate the transcription of the RTCA and B genes. So if we combine this op-on here with the system of ribosomes that I just spoke about on the last slide. We come up with the RTC model which combines the RNA repair functions of RTCA and RTCB with the whole expression of these proteins as well. So what we end up with is, as I said before, this tagged ribosome is called a tagged ribosome because it's recognised by RTCR and it's required for the activation of RTCR. Once RTCR is active it can then activate the expression of the RTCB and A genes to the proteins. And then RTCA goes round and converts the damaged ribosome to a tagged ribosome and RTCB converts the tagged ribosome back to a healthy ribosome which are used in translation of the proteins. So firstly this model we've constrained most of the parameters from the literature. There's three that are unknown which are the damage rate here through ribosome targeting antibiotics and then the maximal rates of transcription of both RTCR and RTCA and B. So by looking at this model there's two that are worth noting at the stage. Firstly there's a positive feedback loop through the fact that RTCA is actually required for its own translation essentially because it produces the signal that activates the transcription of its gene. We have included some, there is some baseline expression so we do have a very small amount of these proteins at the beginning so it works. And secondly there's also ultra sensitivity and that's included in the fact that RTCR when it becomes active is a hexamer. So we assume that it requires six of these RT, the tagged ribosomes, to bind to it for its activation and therefore you get this shape curve that the more, the bone binding you get the more, the quicker the activation essentially. So both the presence of positive feedback and the presence of ultra sensitivity both indicate or both kind of like a perfect storm for, or could indicate the presence of bi-stability in the system. So the next thing we did was a stability analysis. Very interesting, it's the first time I heard about the system. How does it get started? Do you need, because I mean if something you have 100% healthy you get one broken ribosome, how do you get the first tagged ribosome? We have included like a baseline expression of the RTCA and B proteins so you can get the first tagged ribosome. Well a tagged ribosome is the kickstarter of the activation of the whole system. Sorry I'm not quite following. Okay so yeah as I said positive feedback and ultra sensitivity indicate that there could be bi-stability. So we did a stability analysis. Firstly this isn't bi-stability, this is a mono-stable. This is what it would look like if it was mono-stable. So this is, these are steady state values and RTCB on the left axis and healthy ribosomes on the right side. And as damage rate increases the healthy ribosomes obviously decrease, they're being damaged and there's kind of a peak in the expression of the RTCB protein and then once the healthy ribosomes and there's no longer enough of them to translate then we also get a decrease in RTCB. However it doesn't actually look like this because there is bi-stability present. So this is what it actually looks like. So as I said before these are steady state values and the grey area represents a region where there's the coexistence of two stable cell states. So the dotted dash line is an unstable steady state and then we have the stable steady state that's on the upper region and the one that's on the lower region. So that means that we can have the presence of two sub-populations sort of at the same time. So if we look at the damage rate and if we go from an area of low damage rate and we go up then we'll be in the on state of the RTC system. So the system has been activated and that's where we see the tolerance cells arising if you think back to that first slide where the blue cells were tolerant. And then if we go from an area of high damage and we decrease the damage then we will end up in the off state and then those cells will be susceptible. So it's depending upon the initial conditions of the system that determine whether we will be in the on state or the off state. And obviously to reduce the amount of tolerance and therefore resistance we want to be in the off state. So the next thing we did was sensitivity analysis by perturbing the initial conditions in the model. So we perturbed the one at a time for each species initial condition. And what we found was that in general it's really hard to switch the system from on to off. But it was RTCB that was the protein that was most likely to require the least amount of perturbation to switch the system off. So that's why here we've got RTCB steady state on the y axis and then the damage rate again here similar to the other plot. The blue or green region represents when the system will be on and then the red region here represents when it will be off. So if we add a higher damage rate we require much less of a perturbation to get the system to switch off. As we decrease that damage rate the perturbation that's required to switch the system off is much larger. And actually within this range of parameters and all the parameters that we've used here below 1.5 we can't switch the system off by perturbing one single species. However what this does tell us is that RTCB by decreasing the amount that we have that we are able to switch off. So we introduced to the model RTCB inhibition. Just by assuming an inhibitor comes in, binds to the RTCB protein, when they're bound together RTCB is now inactive and it cannot convert the tag ribosomes back to healthy ribosomes. So when we introduced that and compared to this bigger curve here, everything else apart from the introduction of inhibition is exactly the same. So what we see is firstly the region of bi-stability becomes far less and at the same time as you increase RTC inhibition, obviously the amount of active protein decreases but the region of bi-stability and the region where RTCB is on is much much smaller. So we're way more likely to be in the off state now we've inhibited RTCB. If you think back to the model, the schematic of the model, when this was not my initial thought that would, I didn't think this would be how it worked. So here I thought initially that you'd have to inhibit RTCA because that produces the ligand that binds to cause the activation of these proteins. And so if you knocked out RTCA then you wouldn't have any of the proteins and then the system wouldn't be activated. So as a comparison I also did this exactly the same inhibition but with the RTCA protein and what we see here is that yes the protein is decreased so it probably is the system is maybe less strongly on but it doesn't really change the amount, the region of bi-stability and we're just as likely to be in the on state as we were before in comparison. So takeaway from here is that quite counter-intuitively it's the RTCB protein that you have to perturb and inhibit to cause the system to be off. So next we did another sensitivity analysis but instead of looking at the initial conditions we're now looking at the parameters in the model. So here we have again the RTCB steady state here and the damage rate on the x-axis and in this plot here I've changed the growth rate which in the model is dilution at the moment. So there's two values of dilution, one where in the purple region we do see bi-stability and then in the grey area the grey line is not bi-stable within this range of damage. If we then introduce another parameter so now we have the dilution on the y-axis and ATP on the x-axis. This purple region here represents so for every combination of parameters that lies within this region there is bi-stability present. We've perturbed both the dilution rate and ATP because although at the moment they're fixed in the model that's not necessarily very realistic they're not always going to be a fixed number in reality. So if we then also change one of the parameters that controls RTC inducibility we can see that this range of where bi-stability is present varies across a big range. So as I just said it's not necessarily realistic that the dilution rate and ATP will are constant values. They likely do vary over time obviously as bacteria grow the growth rate changes. So we took the growth rate from experimental data and we wanted to see what it would look like on this plot. Basically plotting growth rate against ATP both varied over time. So from experimental data of growth rate this model here of bacterial growth which is coarse grained model that takes into consideration key processes and looks at how they affect growth. So we inferred the ATP values from the experimental growth rate which gives us this black curve here. What this shows us at the moment is that the black curve intersects a number of the different purple regions which confirms that our assumptions of having a fixed ATP and a fixed growth rate in the RTC model aren't necessarily correct and that we probably benefit from varying these parameters over time to have a look at the system in a more dynamic way. So that is what we plan to do next. We're going to combine the RTC model with this model of bacterial growth. That will allow us to analyse the system dynamically and also observe how RTC affects growth because it's very linked to the ribosomes. We can connect them by the ribosomes and see how all the important cellular reactions affect the RTC system and vice versa. So to conclude I think RNA repair is a really interesting area that is maybe not being studied so much. The RTCB is the protein that needs to be inhibited to block the RTC system and prevent the onset of RTC induced tolerance. Finally, it's important to consider this system within the dynamic physiology of the cell. So that's what we plan to do next. Thanks for listening and thank you to everyone at Edinburgh and especially Andrea for helping all of us. Questions? We are not considering... Tolerance is like a precursor to full resistance. We've not modelled that transition. We're just modelling the fact, the system and that we know that that system induces antibiotic tolerance. Tolerance just arises... Resistance just arises by the fact that tolerance cells are around for longer so they're more likely to acquire a mutation. I don't think so. We haven't looked into it like that. Any more questions? I promise to sit more in the front next time. Can you speculate on the intuition why inhibiting RTCB works so much better than RTCA? You said it was counterintuitive but it must also mean that you have some intuition now. Originally I said because I thought that if you inhibit RTCA then you don't get any expression of the proteins. But I think the reason why inhibiting RTCB works so much better is because that actually produces the healthy ribosomes. So with no RTCB you can't go from a tagged ribosome to a healthy ribosome and therefore you don't get translation of the proteins again. Yes, but then what I would see from the diagram is that now the tag ribosomes accumulate which would even more activate this feedback loop. So you would say that these things accumulate and this would not shut it off while if you would inhibit RTCA you would not accumulate these tagged ribosomes and then you would stop the feedback loop. So it's still very unintuitive. But you still can't get expression of the proteins if you don't have any healthy ribosomes. Ah, yeah, that is true. This might be very wrong since I don't quite understand biology just yet. This ultrasensitivity, is there a way to tune that and wouldn't that also just break down your feedback loop? Yeah, we're not like super sure so we set the ultrasensitivity with some parameters and they were based on some assumptions so they could be changed. But I think you'd still get activation of it, it still wouldn't change the fact. So if there are not more questions we can have a break now and we can have side-think coffee.