This paper explores the interaction between 4-amino-antifolates and aldehyde oxidase (aldehyde: O2 oxidoreductase, EC 1.2.3.1) that was purified 60- to 120-fold from rabbit liver with yields of 5-15%. The purification procedure consisted of one heat and two ammonium sulfate precipitations followed by chromatography on hydroxylapatite and then Sephacryl S-200. Analysis of initial rates of hydroxylation of methotrexate, aminopterin and dichloromethotrexate indicated an order of affinities of dichloromethotrexate (10 microM) greater than methotrexate (35 microM) greater than aminopterin (272 microM). There was no difference in the Vmax of methotrexate and dichloromethotrexate (248 and 231 nmoles/min/mg protein respectively); aminopterin (130 nmoles/min/mg protein) was less than that of the other two. The Vmax/Km ratios were 24.1, 7.20 and 0.48 for dichloromethotrexate, methotrexate and aminopterin respectively. This enzyme preparation also mediated the hydroxylation of methotrexate polyglutamyl derivatives with a decrease in the rates of hydroxylation, as the total number of glutamyl residues was increased to four, a consequence of a marked increase in Km values and/or decrease in Vmax; the ratios of the Vmax/Km for the di-, tri-, and tetraglutamates were 0.94, 0.31 and 0.21 respectively. This low activity of the polyglutamyl derivatives of methotrexate for aldehyde oxidase is consistent with the observations that the predominant forms of 4-amino-antifolate polyglutamates found in human liver after administration of methotrexate are the polyglutamyl derivatives of the parent compound. Finally, substrate inhibition for methotrexate and dichloromethotrexate was observed at concentrations in excess of 150 and 30 microM, respectively, about 5- and 3-fold higher than their respective Km values. Hence, while dichloromethotrexate had the lowest Km for aldehyde oxidase amongst the 4-amino-antifolates studied, the actual rates of hydroxylation depended upon the concentration employed because of substrate inhibition. Aminopterin was a very poor substrate for this enzyme at low and saturating concentrations. These properties of the hydroxylation of 4-amino-antifolates may be of importance in the design of clinical regimens with these agents--in particular, regimens that employ infusion of these drugs into the hepatic artery. However, the relevance of these observations to the hydroxylation of 4-amino-antifolates by human liver remains to be established.