Emergent Nonequilibrium Statistical Mechanics from Death and Birth in Biofilms
Kalziqi, Arben L.
MetadataShow full item record
This thesis experimentally explores the statistical mechanics which emerge in the study of bacterial biofilms, highly nonequilibrium communities in which a vast number of bacteria make their homes, and which are of tremendous importance in ecology, medicine, and the economy. In the first set of experiments, we found that local, contact-based killing between cells results. We inoculated multiple, well-mixed strains of V. cholerae on agar pads, then incubated them for 24 hours. When we chose a strain pairing where cells could not kill each other, we found that the strains remained well-mixed regardless of temperature. However, when we mixed together two strains which could kill each other on contact via the Type VI Secretion System (“T6SS”), we found that they underwent an order-disorder transition reminiscent to that seen in the Ising model of an electron spin lattice, with higher temperatures corresponding to later timepoints in this transition. Because spatial assortment is a common means by which bacteria solve public goods dilemmas, we hypothesized that bacteria which could kill non-kin might be more cooperate with their kin. Though a phylogenetic analysis, we found that the number of different T6SS toxins strongly correlated with the number of genes dedicated to the production of external goods, a proxy for cooperativity. Thus, intercellular killing leads to Model A coarsening and (possibly) to the evolution of cooperation. In the next set of experiments, we used genetically modified strains of V. cholerae which secreted no exopolysaccharides (“EPSes”), and thus formed tissue-like (“Matrix-”) biofilms resembling simple stacks of cells. We inoculated biofilms with “nonkiller” or “mutual killer” pairings, and used a white-light interferometer to measure their surface topographies with ~nanometer precision. Surprisingly, we found that surface of biofilms with killing were significantly rougher than those without. A 2015 paper by Risler, Peilloux, and Prost suggested that in the homeostatic limit, the surface fluctuation spectra of a tissue surface may resemble those of a thermal permeable membrane, with an activity- mediated effective temperature. Our biofilm measurements served as experimental support for this theory, and provide further evidence of an effective fluctuation-response relationship driven by birth and death which may exist in cellular solids. Further, we performed minimal simulations which both recapitulated the aforementioned topographical difference and suggested the killing serves to fluidize biofilms. The final set of experiments served as a theoretical and experimental expansion of the previous set. We grew biofilms that could produce EPSes (“Matrix+”), and were thus less tissue-like and more similar to the typical biofilms which are found in nature. We tested the mechanical properties of Matrix- and Matrix+ biofilms, and found that the latter had a higher viscosity by a factor of roughly three. Next, we measured surface topographies and found that while the topographies of Matrix- and Matrix+ biofilms looked similar, Matrix+ biofilms had an effective temperature that was roughly three times higher. To probe whether the effective temperature derived in the second set of experiments had a kinetic interpretation, we used the generalized Stokes-Einstein relation to convert extracted effective temperatures into effective diffusivities, and found that cellular diffusion inside Matrix- and Matrix+ biofilms occurred at the same (viscosity-independent) rate. Simulations, analytical results, and experimental PIV all agree with this result, lending yet more credence to the effective fluctuation-response relationship suggested by Risler et al.