Anthracyclines are aromatic polyketide antibiotics, and several of these compounds are widely used as anti-tumor drugs in chemotherapy. Aclacinomycin-10-hydroxylase (RdmB) is one of the tailoring enzymes that modify the polyketide backbone in the biosynthesis of these metabolites. RdmB, a S-adenosyl-L-methionine-dependent methyltransferase homolog, catalyses the hydroxylation of 15-demethoxy-epsilon-rhodomycin to beta-rhodomycin, one step in rhodomycin biosynthesis in Streptomyces purpurascens. The crystal structure of RdmB, determined by multiwavelength anomalous diffraction to 2.1A resolution, reveals that the enzyme subunit has a fold similar to methyltransferases and binds S-adenosyl-L-methionine. The N-terminal domain, which consists almost exclusively of alpha-helices, is involved in dimerization. The C-terminal domain contains a typical alpha/beta nucleotide-binding fold, which binds S-adenosyl-L-methionine, and several of the residues interacting with the cofactor are conserved in O-methyltransferases. Adjacent to the S-adenosyl-L-methionine molecule there is a large cleft extending to the enzyme surface of sufficient size to bind the substrate. Analysis of the putative substrate-binding pocket suggests that there is no enzymatic group in proximity of the substrate 15-demethoxy-epsilon-rhodomycin, which could assist in proton abstraction and thus facilitate methyl transfer. The lack of a suitably positioned catalytic base might thus be one of the features responsible for the inability of the enzyme to act as a methyltransferase.