Summary Biosynthesis of the molybdenum cofactor (MoCo) can be divided into (1) the formation of a precursor and (2) the latter's subsequent conversion, by molybdopterin synthase, into the organic moiety of MoCo. These two steps are reflected by the complementation groups A and B and the two formally distinguished types of MoCo deficiency that have an identical phenotype. Both types of MoCo deficiency result in a pleiotropic loss of all molybdoenzyme activities and cause severe neurological damage. MOCS1 is defective in patients with group A deficiency and has been shown to encode two enzymes for early synthesis via a bicistronic transcript with two consecutive open reading frames (ORFs). MOCS2 encodes the small and large subunits of molybdopterin synthase via a single transcript with two overlapping reading frames. This gene was mapped to 5q and comprises seven exons. The coding sequence and all splice site–junction sequences were screened for mutations, in MoCo-deficient patients in whom a previous search for MOCS1 mutations had been negative. In seven of the eight patients whom we investigated, we identified MOCS2 mutations that, by their nature, are most likely responsible for the deficiency. Three different frameshift mutations were observed, with one of them found on 7 of 14 identified alleles. Furthermore, a start-codon mutation and a missense mutation of a highly conserved amino acid residue were found. The locations of the mutations confirm the functional role of both ORFs. One of the patients with identified MOCS2 mutations had been classified as type B, in complementation studies. These findings support the hypothetical mechanism, for both forms of MoCo deficiency, that formerly had been established by cell-culture experiments.