The cytoplasmic, NAD-linked hydrogenase of Alcaligenes eutrophus H16 consists of four non-identical subunits. From the mutant strain HF14, defective in this enzyme, a protein was isolated that reacted with specific antibodies raised against the wild-type hydrogenase; the reaction type was of partial identity. The same protein was also tested with specific antibodies raised against each of the four denatured subunits of the wild-type hydrogenase and was found to be completely identical with the second largest subunit; it reacted weakly with the antibody against the largest subunit and not at all with the antibody against the small subunits. In SDS-polyacrylamide gel electrophoresis the protein of the mutant migrated as a single polypeptide and corresponded to the second largest subunit of soluble hydrogenase with Mr = 56,000. The mutant enzyme strongly differed from the wild-type hydrogenase in its binding behaviour to chromatographic gels. It had pronounced hydrophobic properties and bound strongly to phenyl-Sepharose; it had high affinity to triazin dye gels. Enzymatically the polypeptide was totally inactive with NAD as electron acceptor, but showed weak residual activities with methylene blue, ferricyanide and cytochrome c. The protein also contained nickel; however, because of the instability of this polypeptide the amount varied between 0.2-1.4 nickel per enzyme molecule. As shown by ESR studies, the iron is organized in a [4Fe-4S] cluster but is partially present also in the 3Fe-form. No nickel signal could be detected. The role of the polypeptide subunit for hydrogen activation in the intact hydrogenase is discussed.