These enzymes are divided into two classes depending on the metal content at the active site. They have catalytic dinuclear metal centers (Figure 1) in which either one nickel and one iron or two iron atoms are connected through two thiolate bridges. The bridging thiolates have different origins in the two enzymes. In NiFe-hydrogenases, they are provided by two cysteine residues, so that Ni is coordinated by four cysteinates in a non-regular configuration. In FeFe-hydrogenases the bridge is made from a small organic dithiolate ligand. X-ray crystallography and FTIR spectroscopy have clearly shown that the two kinds of hydrogenase are characterized by active sites that contain cyanide (CN-) and carbon monoxide (CO) molecules as ligands of the iron centers. Besides the curiosity for the presence of such toxic molecules as constituents of a native enzyme, these units make hydrogenases some of the extremely rare examples of organometallic centers in biology. In the case of the FeFe-hydrogenase active site the bridging dithiomethylamine ligand plays a very critical function as it participates in proton exchange, facilitating proton reduction to H2 or proton uptake from H2. The proton-exchange site in NiFe-hydrogenases is less clear, but is likely to reside on the terminal Ni-bound cysteinate residues. Finally, what the three-dimensional structures have also revealed is the presence of an array of iron-sulfur clusters (Figure 1), distant from each other by less than 15 Å, allowing an electrical communication between the active site and the protein surface where redox partners are expected to bind for accepting (H2 oxidation) or providing (H2 formation) electrons. At this surface, a so-called distal cluster is exposed to solvent and plays a crucial role in connecting the hydrogenase to its redox partners.