The complete amino acid sequence of glucose oxidase from Penicillium amagasakiense was determined by Edman degradation and mass spectrometry of peptide fragments derived from three different specific proteolytic digests and a cyanogen bromide cleavage. The complete sequence of each monomer comprises 587 amino acid residues, contains three cysteine residues, and seven potential N-glycosylation sites, of which at least five were confirmed to be glycosylated. Glucose oxidase from P. amagasakiense shows a high degree of identity (66%) and 79% similarity to glucose oxidase from Aspergillus niger, and is a member of the glucose-methanol-choline (GMC) oxidoreductase family. The tertiary structures of glucose oxidase from A. niger and cholesterol oxidase from Brevibacterium sterolicum were superimposed to provide a template for the sequence comparison of members of the GMC family. The general topology of the GMC oxidoreductases is conserved, with the exception of the presence of an active site lid in cholesterol oxidase and the insertion of additional structural elements in the substrate-binding domain of alcohol oxidase. The overall structure can be divided into five distinct sequence regions: FAD-binding domain, extended FAD-binding domain, flavin attachment loop and intermediate region, FAD covering lid, and substrate-binding domain. The FAD-binding and the extended FAD-binding domains are composed of several separate sequence regions. The other three regions each comprise a single contiguous sequence. Four major consensus patterns have been identified, including the nucleotide-binding consensus sequence close to their N-termini. The functions of the two motifs recently selected by the Genetics Computer Group, Madison, Wisconsin, as additional signature patterns of the GMC oxidoreductases are discussed. The other consensus patterns belong to either the FAD-binding or the extended FAD-binding domain. In addition, the roles of conserved residues are discussed wherever possible.