 Okay, so first I would like to thank the organizers for this kind of invitation and I will try to move, to pick your attention to the engineering we are doing, to try to evolve or to generate microbes, bacteria that could be used for some therapeutic applications. So our approach is, let's say, a modest one, trying to generate modules that can be combined in, let's say, a chassis of your choice, and so it's not kind of a radical issue like complete engineering of the genome or changing the basis, etc. So we more focus on developing these particular modules. So I will concentrate today in our work that we have been doing to provide targeting addition properties to E. coli and also to provide the expression of injection to non-pathogenic E. coli strains. So as you probably know, the use of bacteria in medicine has been quite limited and certainly most application that you still see in the market comes from the use of probiotics, natural strains that have certain properties. So they are considered natural, so they are from joggers or from, let's say, certain isolates from a patient, etc., they are certain strains that are considered beneficial and are given for certain pathologies or to protect your immune system. And there are other major application which is basically the use of attenuated pathogens to induce immune responses and then produce vaccines or to induce as vaccines or induce a vaccination. Since recombinant technology appears, there was clearly an interest in the modifying bacteria to produce molecules, not only in vitro but also in vivo, maybe to enhance these properties that are sometimes found in beneficial bacteria. And in fact, if you review the literature, there's quite a few examples in which the people have been using traditional recombinant engineering to modify probiotic strains or to modify non-pathogens or even attenuated strains of pathogens to deliver certain molecules of interest like, for instance, interleukins or antibodies against inflammatory diseases, some cytotoxic proteins against cancer because bacteria are able to grow in solitumor preferentially when they are administered systemically, as you will see later. And also other molecules that could intervene against infections, viral or bacterial infections so you could have a probiotic that could protect your causal surfaces against this type of infection or even deliver certain molecules to cells in your gut so you can pre-program the metabolites or some molecules produced by the epithelial cells that are in contact with these microbes. And this is our area in which there's been work, there's many groups that are working on developing this type of applications. So well our consensus, let's say for us, it would be to have, so there's, with this synthetic biology you can generate modules, modules of different types and so ideally synthetic biology can provide a way to combine all these modules in a perfect chassis so you will not need to use a natural strain which is partially beneficial but will have also other properties that are unknown. So we could ideally think on introducing in a chassis that is maybe indicated for a certain application all number of modules to sense the environment and response to that or for instance by delivering a protein we can also introduce modules that would allow the bacteria to move towards certain signals so targeting the bacteria to certain areas or introduce reporters that could indicate us that the presence of a certain molecule or the presence of a certain disease and all could be contained by certain mechanisms of containment so you can introduce in your bacteria all these type of things. I would talk as I mentioned previously on the module that we have developed for targeting the specific addition of E. coli and two seals and surfaces and also later about a delivery module based on injectisomes. So why we focus first on this type of modules to target the additional bacteria to certain cells because we consider that this is an essential property if you want to use bacteria for therapeutic applications is to have the possibility to deliver your bacteria to the type of cells that you are interested in or maybe in some cases you will have a pathological disease and cell that is infected by a virus so you have an antigen of a virus there or it can be a tumor cell and you will have certain protein express but also in addition to this type of application you think that these modules could have other applications also in vitro let's say on abiotic surfaces so you can maybe develop this type of addition against surfaces so that you can help your bacteria to attach to a certain surface for instance in a biosensor sorry I think it moves in the wrong direction yeah okay so if you see what's in nature and so what happened how the bacteria adhere to different surfaces is they have developed a number usually of proteins or polymers of proteins that are assembled on the surface of the bacteria and that allows the bacteria to recognize certain structures on the surface usually as any strain any natural strain carries multiple additions against different with different specificities and so it's difficult to you to reprogram a natural strain is you do not delete partially the major additions that it contains another problem with natural additions also is that the specificity it is kind of broad and low and this is because it frequently recognizes carbohydrate and as you have heard from the previous talk carbohydrate are frequently found in many proteins so bacteria will have the tendency to bind to more than one tissue in my oil so this are more than one cell types using natural addisons so you have to really engineer something new if you want to be more specific and so we thought of just based on what is known about addisons as I told you that can be anchor on the surface we play with different possibilities but we wanted to reduce this to let's say two modular parts one is a domain that will carry the addition property and the second would be a domain that anchors this to the surface of the bacteria so what will be the ideal properties of this type of domains will be high affinity and specificity of course and additionally that it would be something that you can select from a library okay so you can later just simply by changing this addition model you will be able to target two different surfaces so you will have a method of selection and a large library of this type of domains so we thought of using why not immunoglobulin domain because natural addisons in bacteria frequently have IG like domains like infimbria this is a female domain of a pilae type one pilae in icola and it carries an IG like 4 okay so it's possible to use a human or other type of antibody like domains to replace for natural addisons and we decided to use after testing different types this type of antibody molecules that are derived from camel antibodies which carry all this specificity in a single domain of the variable VH domain these antibodies are found in in camelids they were discovered in in Belgium and by the group of search Wilderman years ago and this has very interesting properties in binding and solubility etc and so there are one thing that attract our attention also was that they are simple and you can express them very well in in bacteria okay and in terms of selection this system also allows to easily select the specificity the specificity you want I just simply summarize the standard methodology that has been that is developed but there are other alternatives just after the immunization of an animal of this type you generate a library of genes from the beer cells you cloned it in icola you can infect with bacteriophages and make a pool of bacteriophage and select the specificity of the clone that is recognized in the antigen of your interest as is a simple scheme for basic technology that it was developed time years ago so we decided to use this type of antibody like molecules as addition domains and we later of course co-developed a way to anchor these domains on the surface of E. coli and we tested different anchor domains I'm going to make the story very short we the domain of the the proteins that end up to be the the best ones would were derived from entero pathogenic E. coli entero hemorrhagic E. coli strains that are called intimins intimins are proteins that are expressed on the surface of E. coli of these strains and allow the bacteria to attach to the surface of enterocytes it carries a barrel a domain that anchors in the outer membrane and it displays Ig like domains toward the surface and this is shown here this is this the outer membrane of the bacteria here is the hot cell and this bacteria attached to the enterocytes using intimins which display this Ig like domains for contacting a specific receptor that the bacteria inject into the enterocytes so it's a protein that has a certain specificity a clear specificity for a specific protein that is not found normally in in the mammalian cell so we end up with this type of constructions with an outer membrane anchor domain and this antibody domain and we just started to to express these modules with plasmids and standard plasmids that you can induce in a strain that lacks the major fimbria type in E. coli so they do not attach to mannose residues and we found that they were correctly expressed display on the surface we introduced tax for immune detection and that they bind correctly the epitope on solution and we also tested if they could somehow derives the addition of the bacteria to to a surface coated by the antigen recognized by the antibody in this case so we coat elisa plates with different antigens toward which we have a nano bodies against them and as you can see if you incubate these surfaces with back with E. coli bacteria expressing these additions on on the surface you have bacteria attached to the surface of the plastic coated with a fibrinogen or gfp if you have addition these additions recognizing gfp or human fibrinogen on the surface so we decide to move from the plasmid to the chromosome to avoid antibiotics and to avoid the complexity of making induction etc so what we did essentially was to re combine our plasmids into the chromosome and substitute the p-tac promoters by constitutive promoters and delete at simultaneously that we introduce other reporters like bioluminescence in our strain we delete other additions that could be interfered with the addition process and one thing that we tested and for us was very important was to see at the same time that we introduced the this this construct we remove all antibiotics okay but one thing that we tested is that not only the expression of our constant it was okay in the chromosome but also the growth and the viability of the bacteria was perfect so it was essentially growing at the same rate and it has the same viability than the strain that doesn't carry these constructs in the chromosome and as you see the expression of the construct is maintained throughout days of continuous culture without any requirement of any selection pressure so it's just growing the strains in lb okay so we wanted to test this with cells and so since we had this type of and certain antibodies that nanobodies against gfp and also against the translocated intimine receptor is an antigen of this effect which is not found as i told you before in in hila cells or in human cells in any part of the of your body so we thought this could be nice examples to test the the addition of these strains to our cells so we make stable transfectance of hila cells that express on the surface the antigens recognized by the adhesions in this case was gfp displayed on the surface anchored by an transmembrine domain and and this is the region of of this antigen that is here recognized by by this adhesion and it was anchored to another fluorescent protein that is not cross doesn't cross recognized by cross we are with the nanobody against gfp and we tested initially growing cells in vitro on plates the hila and transfected cells our cell that expresses gfp on or here on the surface and we incubate these wells with our strains the one that is a control or the one that expresses the adhesions against gfp or tear and initially as i told you before we introduced this bioluminescence so you can see all wells that are producing light so and once you was to just see what are the bacteria that remain on the wells after washing you just see that only the bacteria that expresses the adhesion against gfp binds to hila cells expressing gfp on the surface and only bacteria expressing this antibody against tear binds to the well that has carries hila cells with tear on the surface as you take a look at what is inside you would be able to see it better so you will have cells this is a field in which you can see some untransfected cells and other cells that are expressing the gfp and this is a stain for ecoli and as you see ecoli binds to cells that expresses the gfp on the surface and the control strain that recognize a different antigen will not bind to these cells and a nice control is this non-untransfected cells that as you see are clean of bacteria this is more or less the same but with confocal microscopy selecting a feeling where you have a transfected and non-transfected cells and you can clearly see the number of bacteria surrounding your target cell and the same happened in the case of the other adhesion against tear you have a cell that is expressing tear it's fully covered with bacteria so another important thing is the speed at how this happens and we make a video to see how in in time in a time frame of less than three minutes of infection you have here a cell that is non-transfected and the bacteria go scan a little bit the surface and then escapes and whereas in cells in which they are expressing of the target bacteria go and attaches to to the surface and stop moving in and this is something that occurs as you see in less than three minutes so it's very fast we wanted to test this a little bit in vivo and we as you probably know bacteria especially anaerobic bacteria has been used since more than a century ago to try to eliminate tumors because certain bacteria specifically anaerobic can grow in certain anoxic areas in solid tumors and this is a world that was initiated more than a hundred years ago but as I said there's been a group like Neil Forrest from Michigan who are engineering salmonella strains and others to be able to deliver specifically molecules of interest in the tumor. One issue with this type of let's say experimental approach against cancer is the dose that is safe to be administered of course you need a bacteria that is not pathogenic but also you need a low dose to avoid a systemic infection but at the same time the dose has to be high enough so that you warranty the colonization of a tumor by the bacteria. So we decided to test whether the expression of these glues this adhesion on the surface of the bacteria helps the colonization of the tumor and so we established with ourselves a model experimental model based on Hila cell expressing GFP so you can do the bisocutaneous injection you a newt mice you can generate a solid tumor in the flank of the mice and later you can introduce your bacteria by systemic infection in the tailbone of the animal and it is reported that in this type of approaches with bacteria and colonizing tumors three to five days after injection there's a peak of the proliferation of bacteria in the tumors whereas other organ usually have less less number of bacteria as you will see and so we decided to test whether our bacteria were able to colonize these tumors with expressing the GFP and it would be a better targeting of our strains and initially we just simply use the dose that is reported by the people that is using Salmonella or even E. coli k12 to access a tumor a solid tumor that is in the range of 10 to the 7 and as you see both are strained that is targeting the tumor that expresses GFP and the strain that is a control colonize it equally well six out of six animals where the tumors of these animals were colonized efficiently and with having a bacterial proliferation toward 10 to the power of eight per gram of tissue or tumor whereas other organs like liver and spleen were almost clean of bacteria okay so that means that at a high dose there's no difference so we wanted to test what happened if we reduce the dose okay if we really force a condition that is not optimal for the bacteria to colonize those tumor and so we reduce the dose to one percent of standard dose using 10 to the 5 and as you see from this slide now then there's a significant difference between the strain that carries an adhesion that targets the tumor in which in these animals eight out of nine animals were colonized the tumors of this animal were colonized by the bacteria efficiently whereas in the controls only two out of nine or three out of nine we use in the wild type strain that carries non-specific adhesions natural strain were colonized so this engineered strain significantly enhance the colonization at lower doses so you will you are able to target more efficiently bacteria using a lower dose to a tumor and this is a control of a specificity we introduce a group in which the tumor were generated by Hila cells that were not expressed in GFP and in this case our strain was not able to colonize this tumor more efficiently was like in the control identical to the control so to conclude a little bit this part of the talk I show you these constructs this adhesion that can be expressed constitutively from the chromosome of the bacteria and that can allow you to target specific cells or surfaces or tumors in vivo and we believe this we are developing adhesions that will target real tumors with specific markers and that in areas in which bacteria may access and may deliver specific molecules toward the tumor cell or in the tumor environment so I then will focus on one of these delivery machines machineries and as you probably know bacteria had evolved a number of protein secretion systems especially grand negative bacteria to translocate proteins across the inner and outer membrane and one of these some of them are simple but many are some others are extraordinary complex and allow direct translocation from the bacteria to another cell then this another cell in and for this particular type of protein delivery systems usually is a eukaryotic cell or a mammalian cell plant cell and so many pathogens from especially grand negative strains from pseudomonas or from E. coli salmonella carries complex machineries that are called injectisomes for the type 3 secretion system that allows the translocation of a specific proteins toward the cytosol of the infected cell and these effector proteins that are translocated by this machinery are can make multiple functions in the whole cell but helps the infection to proceed they can target cell cycle they can target inflammation they can target the cytoskeleton and mitochondria so they have multiple possibilities to target and usually these machineries are coupled to dedicated ATPase that energizes the system for for translocation they carry also defectors that are translocated carries a short and terminal sequence that is not cleave but is recognized by the machinery to be translocated toward the cytosol of a mammalian cell so in a way this is like a syringe a molecular syringe for translocation of proteins and there's clearly an interest in using this type of devices to toward a specific delivery of molecules in the cytosol of mammalian cells so we decided to try to assemble this type of injection and based on the injection that are expressed by enteropathogenic E. coli strains which as I mentioned previously helped the bacteria these pathogenic strains to attach to the enterocytes so they are syringe like complexes that as I showed you before but the significant difference of this type of injection is that it extends a filament a very long filament that can be up to almost a micron long to access the enterocyte even from a distance so he can inject proteins not even with a close contact he can inject protein initially from from a distance and then it introduces for instance this translocated intimine receptor and then it binds intimately to the enterocyte later okay so these filaments have a channel in which the proteins are supposed to be translocated through this pore of the of this molecule so this is a scheme of how EPEC attaches and injects the translocated intimine receptor and form this attaching and effacing this particular interaction which in this particular case the tear is also able to signal to signal and to reorganize acting on the surface of the enterocyte and so the bacteria will be let's say sit on a pedestal of acting so it's something that is very easily seen is that the bacteria remains extracellular but the acting polymerizes just underneath the bacteria having a like a pedestal like structure bacteria not only introduced here introduces as I mentioned previously many other effectors that has other functions this type of bacteria but as so initially our interest was with antibodies but because as you know antibody molecules are usually targeting a stacellular receptor or soluble antigens they're simply because they are accessible but there are many other targets inside the cell which are important and which are difficult to access using an antibody strategy so we decided to so we thought that since E. coli can be used for selection of antibody molecules why not use the same strain also for delivery but as I said the this type of injectison could deliver many other proteins of interest not only the effectors so we did some some work with the pathogenic strain just to test that the syringes were capable of translocating small antibody fragments like the nanobodies that I mentioned previously and as you see we tag the nanobodies with a sequence that is recognized by the but from a natural effector and we could see the secretion in this case toward the medium specific dependent on on the type 3 system of the antibodies fragments and we made different experiments to show that they were active they were binding the antigen with similar affinity and these are just simply a mutant in which you knock out the ATP ATPase and then without this ATPase activity you don't see any secretion of the molecule or other proteins that are forming this filament and the pore and we also test that these antibody molecules could be with this system could be injected into mammalian cells not only secreted with this with the system with its different experiment maybe the more visual is this one in which you use a reporter that is an enzyme beta lactamase that can degrade a substrate a fluorescent substrate in infected cells so this what you see here are infected hella cells that are you don't see the bacteria you see just simply see the hella cells and if the beta lactamase is inside the cytosol of the cell then the this substrate of that is green of beta lactamase turns blue because of the enzymatic activity of the beta lactamase and so we use this assay to test whether we could translocate also the nanobodies and we test several of them and we also see a translocation of fusion between the nanobodies and the enzyme inside the the hella cells so we it was clear it was clear that there was a potential to translocate the terrologous proteins using the injectes of epec but the problem was that epec is a pathogen so you are not going to use a pathogen at all for introducing anything especially because as I told you epec not only introduces the protein that you are interested in it introduces over 26 effectors other proteins that you are not interested in so we decided that the best way to go would be to move the injectesomes to a non-pathogenic strain and see if we could assemble these complexes the complex delivery into a k12 strain generating something that we call as synthetic injector e coli strain just to continue with this epec ehec cf strain the nomenclature that they saw so how we introduce this into k12 was so how we envision the introduction was not easy it was previously reported that you can clone the whole island into a cosmic and then you could see some assembly of of the tie three but it was very weak and partially because the regulation is not maintained and so the the loci that encodes all the structural proteins of this injectesome is located in a region of about 35 kilobases that is called the locus of enterocyte effacement and it is it carries the structural genes required for the assembly of this machinery but it also carries other genes that might not be they are not interested for for synthetic approach like regulators or the effectors that are being injected and that you don't want to have there so we decided to pick from this loci only the genes that were required for translocation and since we wanted to have a good expression and a controllable expression we decided also to remove the promoters from the our constructs and so we what we decided was to remove the regulators and remove the the effectors and promoters regions and so we just decided to amplify the genes that were coding essential genes to assemble the injectesome in a functional way so we amplify this engineer operons without any as you said only the open reading frames with no promoters no regulators and no effectors and we introduce these frames or fragments into vectors that carry a PTAG promoter to induce the system by addition of IPTG initially and as we are not we don't want to to work with plasmid or antibiotic resistance as I mentioned previously so we decided to integrate all our constructs in the chromosome at different position and in fact what we did was to integrate them in replacing natural adhesions found in bacteria so our strain was further deleted of natural adhesion factor and during the process of integration so we integrate these five operons in the chromosome replacing this cryptic adhesions found in k12 and we've indeed generate an additional strain that lacks the promoter in the first operon of the system so that we could have a control of the expression of a control of the specificity of the translocation of any protein that we wanted so this is an experiment that in which you can see growing the strain and comparing with the production of these injectesomes in epec to compare it with the production in the engineer strain via with addition of IPTG and these are proteins these ESPD, BE and A are proteins of the filament that are in fact translocated through the injectesome and form this filament to help to recognize the mammalian cell and then form a pore for protein translocation so once you see this protein is that the Type III system is active and as you see the three proteins BE, BE and A are also found specifically when we induce in the supernatants of the our engineer strain and you can recognize them by specific antibodies so we detect by Westerblot different components but we decided to go as the farther and try to purify the injectesomes of our engineer strain and compare with what we could purify from enteropathogenic strain and as you see we could detect filamentous structures of different lengths that are compatible with the antibody and that are cross-recognized by antibodies against this injectesome from epec and from our engineer strain which are very similar in they look very similar and so our strain is probably assembling bona fide injectesome of epec so we tested by different methods how can we if our strain is able to translocate proteins into mammalian cells acting really as a syringe and since I previously mentioned the the one that is more easy to test probably is the tear effector why because one is the first protein that is translocated naturally by the pathogen it inserts in the membrane and then it polymerizes acting so there's a simple readout to detect if there's been translocation of this protein by the type 3 is that the acting polymerizes underneath the bacteria so we decided to to test whether our strain was able to translocate tear and for that we introduce tear which is a specific chaperone that is required for its secretion and also intimate now on the surface of the bacteria using amplifying this fragment of the leaf so we will have a strain that in addition to of expressing all the components of the injectesome will carries this from the chromosome carries the the effector protein and the intimate for specific recognition of of tear and this is a confocal image in which we compare the pathogen the epec strain in which I don't know if you are able to see something but the bacteria are stained in both cases with this cyan color so it's what well what you see here are cells infected that almost it's difficult to see the acting because it's a stain in red and it's hard to see but then you have to believe me that the bacteria this in this case is epec there's a small accumulation of acting underneath the bacteria okay this is a strain that lacks the the tpa's so in this particular strain you will see bacteria attached but there's no acting accumulation there's no red spots underneath and these are the engineering strain in which you see the bacteria in cyan and then all this red is the polymerization of acting underneath the the bacteria so meaning that our work our engineering strain was capable of translocating this protein and into the mammalian cell in a functional form so it was able to to signal acting polymerization and this is the specific control in which without the promoter you don't see any even bacteria attached to the cells in this particular case so this summarizes I don't know what happened with this like a not so intense so uh yeah no no you can see better no yeah I don't know what it was like a tired of projecting things so you well then you can see the image so this is a summary of what we we have now but we are working on systems that may allow us to induce the system by not IPTG but other inducer that can be used in vivo but essentially the the the strain it is in a way in which we you can assemble this specific uh injectisomes on the surface and then you can produce the protein of interest in a different position and then this protein will be injected into the mammalian cell and so what could be the application of this especially we combine it with the targeting that I mentioned previously so we could deliver proteins specifically to specific cells and not only antibody fragments we can also think on natural effectors found in in these pathogens this pathogen has like a catalog of proteins that trigger different pathways so you can choose something that acts on for cell killing or for uh in in inflammation for instance and and so you can target a particular pathways using this delivery of protein delivery in the cytosol of bacteria and of course you can introduce other type of things like enzyme peptides for immunizations or why not transcription factor has been shown that certain transcription factor can be translated through this system and can modify the genome also of the of the cell so I think there's uh I say an opportunity to to do a real application with this type of of bacterial engineering so I would like to although you have been seeing the people that has been involved in my presentation I would like to thank uh especially Carlos for his work in synthetic adhesions valence valentio has developed this system also for selection so we can select against different targets and the adhesions and the nanobodies that are of our interest using directly the bacterial display system and David was mainly responsible of the engineering in that of the type 3 in in k12 and I would like also to to thank a friend professor got frankle from imperial college who is an expert in epec etc and who introduced me to all this world initially and to victor of course who has been my lesson mentor in the past and who is still inspiring many ideas in in our group and so thank you for your attention and