The selenium analog of methimazole. Measurement of its inhibitory effect on type I 5'-deiodinase and of its antithyroid activity.

Methimazole (MMI), unlike propylthiouracil (PTU) is a poor inhibitor of type I iodothyronine deiodinase (ID-1). Inhibition of the enzyme by PTU was attributed initially to formation of a mixed disulfide between PTU and a cysteine residue at the active site. Presumably, MMI was unable to form a stable mixed disulfide and thus did not inhibit the enzyme. However, it has been demonstrated recently that ID-1 is a selenium-containing enzyme, with selenocysteine, rather than cysteine, at the active site. This observation raised the possibility that the selenium analog of MMI, methyl selenoimidazole (MSeI), might be a better inhibitor of ID-1 than MMI itself, as formation of the Se-Se bond with the enzyme would be expected to occur more readily than formation of the S-SE bond. To test this possibility, we developed a procedure for the synthesis of MSeI and compared MSeI with MMI and PTU for inhibition of ID-1 and for antithyroid activity. For inhibition of ID-1, MMI and MSeI were tested at concentrations of 10-300 microM. No significant inhibition was observed with MMI. MSeI showed slight but significant inhibition only in the 100-300 microM range. PTU, on the other hand, showed marked inhibition at 1 microM. Thus, replacement of the sulfur in MMI with selenium only marginally increases its inhibitory effect on ID-1. As an inhibitor of ID-1, MSeI is much less than 1% as potent as PTU. MMI and MSeI were also compared for antithyroid activity, both in vivo and in vitro. As an inhibitor of the catalytic activity of thyroid peroxidase, MMI was 4-5 times more potent than MSeI in a guaiacol assay, but only twice as potent in an iodination assay. In in vivo experiments with rats, MMI was at least 50 times more potent than MSeI in inhibiting thyroidal organic iodine formation. The relatively low potency of MSeI in vivo suggests that it is much less well concentrated by the thyroid than in MMI.

Synthesis and DNA-sequence selectivity of a series of mono- and difunctional 9-aminoacridine nitrogen mustards.

The aim of this work was to identify nitrogen mustards that would react selectively with DNA, particularly in G-rich regions. A series of mono- and difunctional nitrogen mustards was synthesized in which the (2-chloroethyl)amino functions were connected to the N9 of 9-aminoacridine by way of a spacer chain consisting of two to six methylene units. The length of the spacer chain connecting the alkylating and putative DNA-intercalating groups was found to affect the preference for the alkylation of different guanine-N7 positions in a DNA sequence. All of the compounds reacted preferentially at G's that are followed by G as do most other types of nitrogen mustards, but the degree of selectivity was greater. The compounds reacted at much lower concentrations than were required for comparable reaction by mechlorethamine (HN2), consistent with initial noncovalent binding to DNA prior to guanine-N7 alkylation. The degree of DNA-sequence selectivity increased as the spacer-chain length decreased below four methylene units. Most strikingly, long spacer compounds reacted strongly at 5'-GT-3' sequences, whereas this reaction was almost completely suppressed when the spacer length was reduced to two or three methylenes. Mono- and difunctional compounds of a given spacer length showed no consistent difference in DNA-sequence preference.