Heparan Sulfate (HS) are complex polysaccharides involved in many biological processes. The structure of HS is regulated at the cell surface by unique extracellular endosulfatases, the Sulfs. Sulfs dramatically change HS functional properties, thereby being implicated in many physiopathological processes including cancer. Sulfs features two domains: a catalytic domain (CAT) that comprises the active site, and an hydrophilic basic domain (HD) responsible for HS binding. The aim of my PhD project is to characterize the structural and the functional properties of the human for HSulf-2, which remains poorly understood. In this context, we have first studied the enzyme/substrate recognition mechanisms. We identified two novel HS binding motifs on these enzymes implicated in their activity. In addition, using natural and synthetic oligosaccharides, we demonstrated that the HD is not essential for HS recognition, but is directs the processive and orientated desulfation of the polysaccharide. Moreover, we showed that a tetrasaccharide is the minimal oligosaccharide size required for HSulf-2 activity. Our results enabled us to propose a new model depicting the desulfation process of HS by the Sulfs. Second, we have shown that HSulf-2 is a proteoglycan, given that it harbors a unique PTM (Chondroitin Sulfate, CS chain) on its HD domain. This chain decreases enzyme activity and HS binding in vitro. In the tumoral microenvironment, using a murine orthotropic mammary tumor model, we showed that the CS chain is lost by proteolytic processing, leading to the activation of HSulf-2, and the promotion of tumor growth, vascularization and metastasis. Finally, we have undertaken the structural characterization of the Sulfs. For this, we decided to study separately the two domains found in these enzymes (CAT and HD). Crystallogenesis assays were undertaken for the CAT domain to solve its structure by X-ray crystallography, but were unsuccessful. Regarding the HD, we set up a protocol of production and purification of recombinant HD and we initiated NMR studies and other biophysics analyses in order to structurally characterize the domain and to identify the HS binding sites. Our preliminary results suggest that the HD is an unstructured domain, except for its N- and C-terminal parts. Overall, our data provide significant insights into this critical regulatory step of HS function.