G-quadruplexes are topologically diverse four-stranded structures formed by certain G-rich nucleic acids in physiological conditions in vitro. During my thesis, I studied their in vivo formation, and the pathological consequences of their persistency. The experimental system used during my thesis rely on the measurement of the stability of human tandemly-repeated sequences called "minisatellite" in mitotically growing S. cerevisiae. The unitary motif of CEB1 form a G-quadruplex in vitro, efficiently unwound by the purified Pif1 helicase. In a Pif1-deficient strain, CEB1 is frequently rearranged (expansion and contraction). At first, we showed that this CEB1 instability was dependent on its G-quadruplex-forming motifs, by directed mutagenesis. I confirmed this result by treating cells with a specific G-quadruplex ligand: it specifically induces the instability of CEB1, but not its G-mutated counterpart. In collaboration with J. Lopes, I showed that the CEB1 instability occurred during replication, upon formation of G-quadruplexes in the leading strand template. The ability to form G-quadruplexes, among other GC-rich minisatellite features, induces gross chromosomal rearrangements. The last aspect of my thesis consisted in the study of the structure-function relationship of G-quadruplexes and genomic instability
