The increase of bacterial resistance against various disinfection processes is a worrying phenomenon. As a consequence, the search for alternative techniques for fighting micro-organisms has become one of the major issues of these last few years. In this domain, water treatment is an important preoccupation for many states, either to guarantee the quality of drinking water supply or to detoxify industrial effluents, charged with chemically or biologically pollutants. The financing of this thesis project was provided by the European project AQUACAT. This project aims at evaluating an emerging technology, the titanium dioxide (TiO2) photocatalysis, as an alternative to the traditional water treatments. This technique is based on the illumination of the catalyst, which thus generates reactive oxygen species (ROS), particularly hydroxyls radicals. TiO2 photocatalysis can be used as a water treatment by detoxifying organic compounds and inactivate micro-organisms. Moreover, its bactericidal activity have a promising biomedical future, by killing cancerous cells for example. Although the bactericidal action of the TiO2 photocatalysis is well documented, its mode of action on bacteria was not yet well defined. The objective of this thesis is to contribute to a better comprehension of the mode of action of the TiO2 photocatalysis on the bacterium Escherichia coli. Few microbiologists ever worked on this subject. A review of published articles on the mode of action of TiO2 photocatalysis showed some experimental failures. Moreover, dogmas related to TiO2 photocatalysis showed contradictions, which will be discussed in this thesis. This thesis is divided in four chapters, including two major sections: i) interaction between TiO2 particles, bacteria, biomolecules and salts and ii) the defence mechanisms of the bacterial cells against TiO2 photocatalysis. The first chapter shows that bacterial cells are adsorbed onto TiO2 particles. So as to demonstrate this feature, bacterial cells and TiO2 particles were mix in the presence of two salts affecting at different degrees the effectiveness of TiO2 photocatalysis rate. In the presence of NaCl-KCl, photocatalysis was very effective and the cultivability of the cells decreased at the beginning of the illumination phase. On the contrary, in a phosphate solution, a latency phase of 20 minutes was observed. Using flow cytometry and microscopic observations, we observed very distinctly that the bactericidal effect begun when the cells started to be adsorbed on TiO2 particles. Bacterial adsorption on TiO2 particles could be correlated to the loss of cell cultivability and also to the loss of membrane integrity, measured by flow cytometry. We then studied the interaction between biomolecules (proteins and DNA), bacterial cells and TiO2 particles. Biomolecules strongly adsorbed onto TiO2 particles in a NaCl-KCl solution (up to 60 µg protein per ml of catalyst). A desorption buffer composed of phosphate (50mM, pH 7.0) and SDS (0.1%) is able to desorb 12% only of the proteins. When bacterial cells were in presence of a cell crude extract corresponding to the quantity of proteins belonging to their population (equal to 12µg protein/ml for 2×107 bacteria /ml), it was observed that the effectiveness of the treatment was decreased, suggesting that cell crude extract protected bacterial cells from the photocatalytical treatment. So as to improve our understanding of the mode of action of the TiO2 on the bacterial cells, we compared the survival of E. coli mutants with the wild type one. Knowing that TiO2 generated ROS, we assayed mutant affected in their defence mechanisms against these molecules. Thus, the sensitivity of mutants defected in the induction of DNA repair and of DNA protection systems suggested the induction of cytoplasmic damage by the photocatalytical process. We could also highlight the importance of the Fenton reaction and the formation of hydroxyls radicals in this inactivation process. The hypothesis is that these hydroxyl radicals were also produced in the cells, following an increase in reactive iron and hydrogen peroxide. Interestingly, the loss of E. coli cultivability after TiO2 illumination appeared not only during the illumination, but also when cells were incubate on Petri plates, whereas TiO2 was not illuminated. Thus bacteria underwent a secondary stress, leading also to their loss of cultivability. Finally, we identified genes expressed in the response to the treatment (primary stress) and after a short incubation in a rich culture medium to 37 °C (secondary stress). Thus after an illumination time (100% of bacteria were cultivable) and an other one which show first signs of letality (80% of the cultivable bacteria), no difference in expression of the mRNA could be observed with the transcriptome analysis (DNA chips) of the E. coli genome. On the other hand, as soon as the cell were resuspended in a rich medium after the treatment during only 5 minutes, we could observe differences, by a factor higher than two, in the induction of certain genes. The totality of genes induced during the TiO2 photocatalysis were repressed by Fur and were implied in the iron transport.