 such a great meeting. So we will change a little bit the environment that we will talk about. And first of all, maybe just before I start, few nice pictures of Slovenia, the neighboring country of Italy where I come from. But what I want to talk about today is soil bacteria. And we work with soil bacterium bacillus satellites for quite a while. And the approach we use in our work is actually that we isolate bacteria from a small space. Like from soil grains. And from each grain we isolate different strains of bacillus satellites. And the aim was to really understand social interactions between these strains. At that time, when we started to do this work, we were studying quorum sensing and diversity of quorum sensing specificity types that we find in bacillus, namely ferotypes. And from this picture, you can see from the color codes that even within the species you have in a small area, many so-called ferotypes. And the same is true if you look at the rhizosphere of one plant. It is the diversity of these social types is present. Now, in addition, when we have now these populations of bacteria, we try to understand how they behave in multicellular groups. Namely, we study swarming or also biofilms, being colonies or biofilms on plant roots or actually as pellicles. And the two social behaviors we addressed most often is one which I already mentioned to you. This is so-called quorum sensing. I don't need to repeat what it is, but it's a cell density-dependent behavior, which depends on accumulation of signals. And in-gram positives, the signaling molecules are peptides. And I just want to mention my student, Miha, Shpatsa Panku is in the audience. And he's here. He was presenting a poster on the role of these peptides in inducing proteases and bioform regulation. Actually, protease is mainly on this poster. And how these proteases regulate potentially the stability of these peptides. But today, I will talk about the other topic we do in our lab and try to explore. And this is so-called kin discrimination. And the work was presented, part of it, by Maya Bolyashic, who is also in the lab. And if you have questions about it, you can also ask her. But a lot of work on this topic that I will talk about today has been done by my more senior, two senior researchers in the lab, Apalonsa Stevonich and Barbara Kreiger. So kin discrimination is a mechanism where you have recognition between cells and then differential behavior upon this recognition between the interacting species or strains, actually. But so it is all based on cooperative behaviors which have been discussed during this conference as such that are susceptible to exploitation by cheaters. And a long time ago, Hamilton proposed that kin discrimination might be a mechanism, a sortment mechanism which can protect these populations from invasion by cheaters which are not kin. And so the process then results in the fact that two populations can, if they are non-kin, differentially behave towards the other. Or if they are kin, they can assist each other and thus better coordinate the group behavior. Now, what I would like to talk to you about today are three topics. First, I will explain how we discovered this phenomenon in the population of Bacellus satellites, actually community of strains that we isolated. And I will talk about a mechanism that we find occurs during strain interactions that are non-kin. And at the end, I would like to touch upon on the topic what are the social consequences when these strains interact in multicellular settings. Now, if we have a population of strains which are coming from the same environment, we ask ourselves, could we detect that between these strains a differential treatment which would correlate also with their phylogenetic relatedness? And so the first thing, of course, that was necessary to do was the determination of phylogenetic relatedness between these strains. And as I told you before, they are all Bacellus satellites. And therefore, they are highly similar at the level of phylogenetic genes. But still, the 39 strains that we use to address this question have been previously determined by us and in collaboration with Fred Cohen from United States that even at this small area of one cubic centimeter, you can find three ecological types of Bacellus satellites. So we took these strains. And then we asked ourselves whether we can, during their cooperative swarming, evaluate interactions. And at that time, it was already known that, for example, during protus mirabilis swarming, you can detect these boundaries that come between strains when they touch upon each other, as well as in a mix of bacterial compatibility tests. So we used the strains. We inoculated them at different spots or areas of the agar plate that induces, an agar medium that induces swarming. And so we indeed found this phenotype that we predicted. It will be there. So there is this boundary, dramatic boundary between strains that are different. And we don't see such a boundary. We see merging between the swarms of the same strain. So Barbara went on and tested in all possible combinations the 39 strains that we had. And of course, she did it many times. And at the end, she came out with the following metrics. Before that, I would also like to say that when we did these interactions, actually, Barbara, I did them. There were three different phenotypes, very strong boundary, a little less pronounced one, and no boundary, which we call merging. And the relatedness, actually I will say this first. So when we compared these strains, and now the color codes are labeled according to the ecotype. And between ecotypes, strains are more phylogenetically related. What we saw is that these groups that you see here, so in here, the strains always merged. But when you have two circles, strains formed a boundary. And you can see that boundaries would always occur between ecotypes. We don't see any merging between ecotypes ever. But within one ecotype, for example, the yellow one, you can see that there are more kin groups, as we call them. So the diversity of social interactions at the level of recognition or this differential treatment is much higher. It's 12 kin groups among 39 strains. And this is very similar, which was found already by Voss and Belliser in 2009, where they looked at a mix of bacterial isolates from a soil patch of a little bit bigger patch of 16 centimeters by 16 centimeters. Now if we looked and compared the phylogenetic genes of these strains very carefully in pairwise combinations and ask ourselves what kind of identity will correlate with the merging. So this is here indicated yellow merging with the intermediate phenotype or the boundary, which is indicated blue. And from this picture, you can already see that majority of interactions were resulting in the boundary formation. And merging would only occur in higher frequencies. And here you can see the frequencies. When there was 100% identity between phylogenetic genes, and then it dropped significantly. And below 99.4% identity between two genes in the pairwise combinations, you would never, ever see anymore the merging. So phylogenetic genes are not involved in these interactions. They are just reflection of the evolution of this population, suggesting that as the population evolves, there is a higher and higher frequency possibility that they will be antagonistic. But even if they are 100% identical at the phylogenetic genes, you can still see, for example, here one boundary or several intermediate types. So the next question, when we saw this nice distribution of boundaries and merging, and which suggested to us that actually the mechanism behind correlates with the evolution of these strains, we asked ourselves, could we detect also mechanisms that are responsible for these phenotypes for so-called kin discrimination? And the approach we used here in collaboration with Roberto Coulter's lab in the United States, I would especially like to acknowledge Nikol Lyons. He's a very talented postdoc. And here is Barbara from my lab who has also done the same kind of experiments. And we both did transposon metagenesis in different strains and as well as reverse genetics and then compared our results and talked about what we see. And the approach we used with transposon metagenesis was such that a strain, a parental strain would be used. It would be mutated. And then we would look for boundary formation between ancestor and the mutant. And doing these kind of experiments, it was found that the law side that were responsible for this boundary formation could be grouped into three groups functionally. And this is toxin immunity genes, which is not that surprising. For example, in Bacillus Vap A is the most interesting, maybe for this in this respect, which is a contact dependent toxin immunity function and is also polymorphic between strains. And then these are like killing also toxin immunity genes. And then there is a group of surface components, genes that are involved, for example, in synthesis of tychoic acid, or genes involved in synthesis of extracellular polymers, the polysaccharides that are present on the surface. And finally, here are regulatory genes, many of which actually regulate these functions. So at the end, we could conclude that there are three groups of genes, antimicrobial cell surface, and the response regulation genes that are controlling this function. The question was then, OK, we see these antimicrobial genes that might be important. Are the antagonistic interactions the main drivers of kin discrimination between Bacillus strains? Is there any cell damage in the cell boundary that we see? And this is another picture of the cell boundary. So here, Nick has done an experiment where he looked at cells in the boundary or between two kin strains. And he found that between kin strains, there is less, for example, propidium iodide stain cells, indicating less dead cells than if you look at the number of propidium iodide stain cells in the boundary. And we have done an experiment where we looked with electron microscopy what's happening actually at the boundary. And you can see these cells here are much different than the cells in the swarm. So in the boundary, the cells look like kind of empty bags, suggesting that there is some antagonism going on. This correlates also with the transcriptional activities that were tested within the boundary. And I will not go into the details. I just want to show you one example, which is, for example, expression of sigma w gene between two strains when they interact and they are non-kin. Sigma w is a sigma factor responsible for the patient to the stress which occurs at the envelope level. And you can see that very high induction of the sigma w in one of the strains. And that's what happens usually. When two strains meet, one responds widely, the other actually doesn't change its genes that much. So it seems like one is attacking. And the other is less, is then responding to this attack. Now, if strains are compared, which are phylogenetically, they are still very closely related within the species or subspecies. But if we compare strains and say, OK, let's compare in such a way that they would decrease in phylogenetic relatedness, we see that the boundaries here, they become more prominent with the distance. So the very closely related strains, they still merge. And then you can see some boundary and the increase in boundary prominence. And Nick did one very interesting comparison using bioinformatics, taking the genome of a lab strain that has been used to study biofilms a lot in the bacillus community of scientists. And here are listed genes that are actually known genes in coding antibiotic or antagonistic functions. Doesn't really matter their names. What's important here is that here are the strains listed also in the order. So that phylogeny relatedness drops in this direction. And the gray areas represent the similarity, and white areas represent like missing, that the strain is missing, the gene. And you can see that as the phylogenetic relatedness drops, so does the common number of antagonistic genes in these genomes. Compared, of course, here towards to the reference strain. And so with the boundary prominence, as I told you before, the boundary prominence increases. Now, we wanted to know if we, for example, have two non-kin and then inactivate genes that we predict are involved in one, could we actually merge them back? Could we make non-kin, kin again? And that has been a very, very difficult task. And the only time we were successful to do this was when we used very closely related strain, PC216, which is only different between this lab strain in this area of sp-beta-prophage, and mainly the sublensin locus, which encodes also in antibiotic. So here is the interaction between wild type strain and PC26 strain. You see some boundary not very strong. If we now inactivate the sublensin gene, which is present in this focal strain, then we found that, indeed, you can reverse the phenotype and the strains become like mergers or kin, as we call them also. So if we found now these three types of groups of genes, how does this knowledge correlates with the knowledge that is already available in the literature on kin discrimination in other organisms? And there are two mechanisms which have been studied so far. Mostly, for example, polymorphic receptors have been indicated as means that strains, which are kin, can find each other and therefore increase their probability of interaction. Indictiostelium, in myxocococcus santus, for example, the trareceptor responsible for the outer membrane exchange or flow 1 gene in saccharomyces cerevisiae. Or there are also indications of immunity to the effectors strategies in proteos mirabilis. For example, this group which has done the contact-dependent killing as proposed, the contact-dependent killing as the mechanism to line formation, but also from the laboratory of Karina Gibbs, who proposed that actually these IDS genes that she discovered in collaboration with Pete Greenberg not the killing machine, but rather proteins that are able to help two strains to identify each other as kin if they are the same. But in bacillus, we have a little bit different system as at least at this point when we look at it. And we call it combinatorial system. Many loci are involved in kin discrimination. That's what we find. So far, we have not found any such strong receptors that would bring the cells together. But they might exist, that it's very possible. And so what we believe is that strains forming boundary encode a different set of kin discrimination loci. And this is not one or two or three. We think it's many more loci that are involved. And that these loci are highly dynamic in their existence. It's not that one strain will have them three loci all the time, they will change due to horizontal gene transfer. That's our hypothesis. And why we think there are many loci? Because it's very difficult to switch from non-kin to kin because mutations in kin discrimination loci, and these are the same mutations that we introduce into different strains of bacillus satellites, can give very different phenotypes. So that suggests that they have other partners with them. And that there is also a symmetry in the expression of KD loci when transcriptomic is performed. But I have not talked about this much. So to conclude with this, I would say that number of shared antibicrobials and immunities is the key determinant of the bacillus satellites kin discrimination system. OK. So why? What would that do? Maybe this kind of system would prevent mixing, random mixing, which would be detrimental to the community of different strain and would stabilize the cooperation within the group of kin cells. So if we propose now that when kin meet, they actually are happy they are doing well. And when non-kin meet, then the outcome is a bit different. They don't want that non-kin get close to them. What would actually happen in more real settings if we now really mix them together and not try to compete them one against each other on the agar surface as we did here? So we did experiments with plants which are more realistic systems. Like we used Arabidopsis thaliana. And we inoculated these plants. We actually put the plant into the medium that was then inoculated with strains. And after a certain period of time, then the spatial distribution of these strains has been looked at. And you can see that strains were labeled with fluorescent markers constitutively expressed. And here the names of the strains indicate if they are the same. So here are kin or self. So you can see that in this kind of combination, you see always both types of, I mean, both cells present. They are kin. Or if you have even different strains, but they are kin defined by the swarming assay, they also coexist and form this cluster on the root. This is not the case when non-kin are mixed in this setting. Because then we see only one strain prevailing on the root. And here again, another one has one in this competition. If we do these assays on the swarming plate, this time mixing strains and inoculating them in the middle of the plate, again, the strains are labeled with different markers. So the prediction would be that, again, there would be competition between non-kin and maybe coexistence of kin. And after 24 hours, if we look under the dissecting microscope at these mixtures, yes, indeed, the kin always stay together. They swarm together. And that's why the color on the plate of the colony is yellow. But we don't see this phenotype if we mix non-kin. In non-kin combinations, always one or the other wins. Sometimes we see sectoring. But this sectoring is always present only when we mix non-kin, not when we mix kin. So these strains, these cells are moving. They are not fixed. They are mixing. But not very well if they are non-kin. They then move also sometimes in these separate sectors over the plate. The territoriality is achieved, obviously, through competition. Because who wins in the assay between non-kin is very much dependent on the frequency of one of the strains. So if one would be here, it's only 4 to 1 ratio, and this one would always win. You see, if you mix kin, this does not happen. You just see the ratio preserved between kin. And finally, maybe I will say a few words also about mixing of kin and non-kin in a static culture. Here, a static culture, we grow cells statically, and they form pellicles on the surface of the liquid medium. Again, they are labeled with different fluorescent markers. And Maya, Barbara, and Istok have done this work. What we see here is a little bit different than in swarming assays. We see coexistence of kin, of course, that's predicted. They mix well. But also, we see that non-kin coexist in this pellicle. Only the size of the patches is very different. It's larger when non-kin are mixed. So here, mixing is very, very close. And if we go at 12 micrometers distance, we would, in kin combinations, always see two colors. But this would not be in non-kin combinations, where it would be one or the other. And such mixing also has consequences for fitness. Not as dramatic as in swarming assay, where one is just wiped out almost most often. But here, what we see is that if we mix kin, so the fitness does not change much. Both strains have the coexist well. And in non-kin combinations that form like this very strong boundary, one would be a loser. The other wouldn't gain so much, but one would be the loser. In this kind of boundary pairs, there is still good coexistence. So with this, I can slowly come to my conclusions. And this is that kin discrimination that we discovered shapes multicellular behaviors. It shapes how they coexist, how they swarm. They swarm separately. So it also affects territoriality. And finally, we believe at the moment that antimicrobials and immunity pairs, which have to be the same in kin and are different in non-kin, are the main molecular determinants of kin discrimination in bacillus. So with this in mind, I would leave you with a take home message. And that's gains and losses of social life depend on who you are interacting with. So pick wisely.