The plasmid-borne raf operon encodes functions required for inducible uptake and utilization of raffinose by Escherichia coli. Raf functions include active transport (Raf permease), alpha-galactosidase, and sucrose hydrolase, which are negatively controlled by the Raf repressor. We have defined the order and extent of the three structural genes, rafA, rafB, and rafD; these are contained in a 5,284-base-pair nucleotide sequence. By comparisons of derived primary structures with known subunit molecular weights and an N-terminal peptide sequence, rafA was assigned to alpha-galactosidase (708 amino acids), rafB was assigned to Raf permease (425 amino acids), and rafD was assigned to sucrose hydrolase (476 amino acids). Transcription was shown to initiate 13 nucleotides upstream of rafA; a putative promoter, a ribosome-binding site, and a transcription termination signal were identified. Striking similarities between Raf permease and lacY-encoded lactose permease, revealed by high sequence conservation (76%), overlapping substrate specificities, and similar transport kinetics, suggest a common origin of these transport systems. alpha-Galactosidase and sucrose hydrolase are not related to host enzymes but have their counterparts in other species. We propose a modular origin of the raf operon and discuss selective forces that favored the given gene organization also found in the E. coli lac operon.