An investigation of the mechanisms of inactivation of the pyridoxal 5‘-phosphate (PLP)-dependent pig brain γ-aminobutyric acid (GABA) aminotransferase by 4-amino-5-fluoro-5-hexenoic acid (2), a monofluorinated analogue of the anticonvulsant drug vigabatrin, is described. Inactivation of [3H]PLP-reconstituted GABA aminotransferase with 2 followed by denaturation released the coenzyme in two forms, one as PLP and the other in a modified form in the ratio 7:3. All enzyme activity was lost upon inactivation by 2, but about 30% of the activity returned upon incubation with PLP, consistent with the formation and release of 30% of the coenzyme in a modified form, as noted above. Inactivation of GABA aminotransferase with [2-3H]-2 followed by gel filtration resulted in the attachment of 0.7 equiv of tritium to the enzyme, even though complete inactivation occurred. This also is consistent with the above results that about 30% of inactivation is the result of release of a modified coenzyme, leaving 30% of the enzyme as its apoenzyme form. Isolation and mass spectral analysis of the modified coenzyme gave peaks consistent with a modified coenzyme formed from a reaction with the inactivator (27). Denaturation of the enzyme containing 0.7 equiv of radioactivity from the above experiment led to release of 0.2−0.3 equiv of the radioactivity as γ-acetyl-GABA (20). Treatment of the denatured enzyme with sodium periodate generated 0.2−0.25 equiv of succinic acid, leaving 0.15 equiv of radioactivity still covalently bound to the enzyme. Analysis of amine metabolites shows the formation of 0.5 equiv of 20. Analysis of the nonamine metabolites resulted in the identification of 1 equiv of 4-oxo-5-hexenoic acid (24). After inactivation, 2.6 ± 0.1 equiv of fluoride ions was detected, consistent with the loss of 1 fluoride ion to produce inactivation, 1 fluoride ion to generate the 4-oxo-5-hexenoic acid, and 0.5 fluoride ion released in the production of γ-acetyl-GABA. Normal transamination also occurs; 6.3 ± 0.6 transamination events occurred during inactivation, as measured by the conversion of [14C]-α-ketoglutarate to [14C]glutamate. These results indicate that there are, at least, three different inactivation mechanisms in effect (Schemes -). All of these mechanisms begin with Schiff base formation between 2 and the active site PLP followed by removal of the γ-proton and elimination of the fluoride ion. It is from this conjugated allene intermediate (17) that all of the inactivation pathways and metabolites result, except for the normal transamination product. The partition ratio, the amount of inactivator converted to a product per inactivation event, is about 8; 6.5 transaminations, 0.5 conversion to 20, and 1.0 conversion to 24 per 1.0 inactivation event.

The mechanisms of inactivation of γ-aminobutyric acid (GABA) aminotransferase by (Z)- (2)- and (E)-4-amino-6-fluoro-5-hexenoic acid (3) were studied. The kinetic constants of inactivation for 2 and 3 were approximately the same. Inactivation of [7-3H]PLP-reconstituted GABA aminotransferase by 2 and 3 also gave similar results for the two isomers:  63% (2) and 66% (3) of the radioactivity remained covalently attached to the enzyme; 31% (2) and 29% (3) were released as PLP; 5% (2) and 4% (3) of the radioactivity emerged as PMP. Treatment of GABA aminotransferase with either [3H]-2 or [3H]-3 led to the incorporation of 1.0 equiv of tritium into the enzyme after gel filtration. Urea denaturation at pH 7.0, however, released about 0.3 equiv of the tritium from the enzyme, and urea denaturation at pH 2.4 released 0.35 equiv of the tritium. About 85% of the released radioactivity was identified as 4-amino-6-oxohexanoic acid (31) and the remainder as a mixture of 4-oxo-5-hexenoic acid (10) and the product of Michael addition of β-mercaptoethanol to 10. Neither inactivator produced any amine metabolites during inactivation. The first divergence from similarity between the two isomers was in the isolation of nonamine metabolites. Inactivator 2 generated two nonamine metabolites, whereas 3 produced only one. The additional metabolite with 2 was identified as 10. The metabolite in common may be the normal transamination product or something derived from it. To confirm this possibility, it was shown that both isomers undergo transamination; 2 is transaminated 1.4 ± 0.3 times, and 3 is transaminated 0.7 ± 0.3 time. Fluoride ion release also was monitored, and it was found that 2 released 1.4 ± 0.2 F- and 3 released 0.9 ± 0.05 F-. The additional F- release for 2 is expected, given that it produces an additional metabolite that requires release of F- for its formation. Absorption spectra of GABA aminotransferase inactivated with 2, 3, and 4-amino-5-fluoropentanoic acid (35) showed an absorbance at 430 nm that was missing in the spectrum of native enzyme. Taken together, these results indicate that 2 and 3 inactivate GABA aminotransferase by multiple mechanisms, but at least the major inactivation mechanism is different for the two isomers. The results for 2 can be rationalized by three different mechanisms. All are initiated by Schiff base formation of the inactivator with the active site PLP followed by γ-proton removal. The major pathway (about 65%) proceeds by isomerization of the fluorovinyl double bond (Scheme ) and produces a ternary complex between the enzyme, the coenzyme, and the inactivator (13); a small amount of inactivation (about 3%) may result from formation of a weakly stable covalent adduct with an active site nucleophile (14). The other two inactivation pathways proceed by isomerization of the aldimine double bond of the Schiff base with PLP. Scheme (about 30%) results in the formation of a weakly stable adduct that decomposes upon denaturation at neutral pH to 4-amino-6-oxohexanoic acid (31). Scheme (about 5...

No abstract

No abstract