A recent article in TiBS, by Macheroux and co-workers[1xStehr, M. et al. Trends Biochem. Sci. 1998; 23: 56–57Abstract | Full Text PDF | PubMed | Scopus (31)See all References[1], pertains to the cofactor- and substrate-binding domains in flavin-dependent N-hydroxylating enzymes. We would like to restrict our comments primarily to lysine-N6-hydroxylase, IucD, an enzyme that we have been investigating for some years.Studies of IucD–PhoA and IucD–LacZ fusion proteins[2xHerrero, M., de Lorenzo, V., and Neilands, J.B. J. Bacteriol. 1988; 170: 56–64PubMedSee all References[2]support the observation that IucD is normally membrane bound[3xGoh, C.J., Szczepan, E.W., Menhart, N., and Viswantha, T. Biochim. Biophys. Acta. 1989; 990: 240–245Crossref | PubMedSee all References[3]. These studies suggest that at least two domains in IucD attach to the cytoplasmic side of the plasma membrane. The first membrane-attachment domain appears to reside in the first 25 residues of the protein, and its sequence resembles that of the signal peptide[2xHerrero, M., de Lorenzo, V., and Neilands, J.B. J. Bacteriol. 1988; 170: 56–64PubMedSee all References[2]. Although Diekmann and co-workers[4xPlattner, H.J. et al. Biol. Metals. 1989; 2: 1–5Crossref | PubMed | Scopus (13)See all References, 5xMacheroux, P., Plattner, H.J., Romaguera, A., and Diekmann, H. Eur. J. Biochem. 1993; 213: 995–1002Crossref | PubMed | Scopus (15)See all References]have described a soluble form of IucD[4xPlattner, H.J. et al. Biol. Metals. 1989; 2: 1–5Crossref | PubMed | Scopus (13)See all References, 5xMacheroux, P., Plattner, H.J., Romaguera, A., and Diekmann, H. Eur. J. Biochem. 1993; 213: 995–1002Crossref | PubMed | Scopus (15)See all References], an inability to isolate membrane-free preparations of the enzyme prompted us to develop recombinant, cytoplasmic forms of the protein by using strategies based on the gene-fusion approach[6xThariath, A.M., Socha, D., Valvano, M.A., and Viswanatha, T. J. Bacteriol. 1993; 175: 589–596PubMedSee all References[6].We contructed three different plasmids, pAT2, pAT3 and pAT5, which encode modified IucD forms (rIucDs) that possess altered N-termini[6xThariath, A.M., Socha, D., Valvano, M.A., and Viswanatha, T. J. Bacteriol. 1993; 175: 589–596PubMedSee all References[6]. Escherichia coli cells transformed with pAT5 and pAT3 produced cytoplasmic forms of lysine-N6-hydroxylase[6xThariath, A.M., Socha, D., Valvano, M.A., and Viswanatha, T. J. Bacteriol. 1993; 175: 589–596PubMedSee all References, 7xThariath, A.M., Fatum, K.L., Valvano, M.A., and Viswanatha, T. Biochim. Biophys. Acta. 1993; 1203: 27–35Crossref | PubMed | Scopus (17)See all References, 8xSee all References]. The rIucD encoded by pAT2 can be produced either as a membrane-associated form (rIucD 426) or as a cytoplasmic form (rIucD 456)—production of the latter form being dependent on expression in a supK strain of E. coli[6xThariath, A.M., Socha, D., Valvano, M.A., and Viswanatha, T. J. Bacteriol. 1993; 175: 589–596PubMedSee all References[6]. The incorporation of the desired fusion and the absence of any other mutations in the IucD component were confirmed by determination of the nucleotide sequence of each construct. There were a few discrepancies relative to the previously reported nucleotide sequence of IucD[2xHerrero, M., de Lorenzo, V., and Neilands, J.B. J. Bacteriol. 1988; 170: 56–64PubMedSee all References[2]; however, these are consistent with the revised sequence that has recently been published[9xMarrone, L. and Viswanatha, T. Biochim. Biophys. Acta. 1997; 1343: 263–277Crossref | PubMed | Scopus (8)See all References[9].Our rIucD 398 construct contains 19 amino acid residues derived from β-galactosidase; the rest of the fusion protein comprises residues Val48–Thr426 of wild-type IucD. Thus, rIucD 398—which is catalytically active—lacks the N-terminal segment that Macheroux and co-workers[1xStehr, M. et al. Trends Biochem. Sci. 1998; 23: 56–57Abstract | Full Text PDF | PubMed | Scopus (31)See all References[1]identified as the FAD-binding domain. We therefore think that it would be prudent to defer assignment of specific cofactor- and substrate-binding domains until a thorough characterization based on the elucidation of three-dimensional structure of the protein has been achieved. The N5-ornithine hydroxylases PvdA and Sid1 remain to be characterized. Such a paucity of information calls for caution to be exercised in the assignment of specific functional domains in this protein.