Effect of homocysteine on brain glutamate decarboxylase and interaction with pyridoxal phosphate

There is ample experimental evidence which indicates that 4-aminobutyric acid is an inhibitory neurotransmitter in mammalian brain. Changes in 4-aminobutyric acid concentrations in the central nervous system have therefore prompted speculation regarding its involvement in neurological aberrations. Individuals homozygous for cystathionine /3-synthase deficiency (an autosomal recessive inherited disorder of methionine metabolism) are affected by a wide variety of clinical abnormalities, which include mental dysfunction of varying type and severity. A considerable proportion of affected patients experience intermittent convulsive episodes. As a consequence of the genetically determined enzyme deficiency, the normal condensation of serine and homocysteine to produce cystathionine is markedly affected, resulting in an accumulation of methionine, homocysteine and the disulphide homocystine in tissues and body fluids. The biochemical aetiology of the convulsive episodes is not well defined, although it has been shown that administration of homocysteine to experimental animals results in the production of severe convulsive seizures. It has also been reported that homocysteine inhibits several pyridoxal phosphate-dependent bacterial enzymes. In view of these facts, together with the suggestion that the convulsant properties of many compounds may act through a disturbance of brain 4-aminobutyric acid metabolism, we decided to ascertain whether the convulsant action of homocysteine might function in this manner, possibly by interference with the 4-aminobutyric acid-synthesizing enzyme glutamate decarboxylase (EC 4.1.1.15) which has an essential requirement for pyridoxal phosphate. Glutamate decarboxylase was partially purified from mouse brain, after homogenization and ultrasonication, by (NH,),SO, precipitation (60% satd.) and gel-exclusion chromatography. The active fractions eluted from a Sephadex G-200 column were concentrated by ultrafiltration, by using a macrosolute concentrator (Amicon, type B 15), and used for enzyme assays in which glutamate decarboxylase activity was quantified by measuring the evolution of I4CO2 from L-[ l-14Clglutamic acid. Assay mixtures, in a total volume of 0.3ml, contained (final concns.) 0.2 M-potassium phosphate, pH 6.8, 0.025 pCi of 1 ''Clglutamic acid and partially pure glutamate decarboxylase. Glutamic acid and pyridoxal phosphate concentrations were varied according to the experiment, and homocysteine, when added, was present at final concentrations of 0.1-10mM. Incubations were allowed to proceed for 90min at 37OC in reaction vessels fitted with air-tight stoppers from which were suspended small glass wells. Evolution of I4CO2 was determined essentially by the method of Wu et al. (1973), except that ethanolamine was used as a C0,-trapping agent. The rate of I4CO, evolution was markedly decreased in the presence of homocysteine when the concentration of either glutamate or pyridoxal phosphate was varied in the assay, the rate decreasing with increasing concentrations of homocysteine. The observed decrease in I4CO2 evolution can be interpreted as follows. Either, there may be direct inhibition of glutamate decarboxylase by homocysteine binding at the glutamatebinding site, or homocysteine may interact with free pyridoxal phosphate to form a complex which could still bind to the coenzyme-binding site but block the binding of glutamate. Brain glutamate decarboxylase is particularly sensitive to depletion of pyridoxal phosphate both in uitro (Roberts et al., 1964) and in uiuo (Minard, 1967); moreover, the activity of some decarboxylases may be regulated by a decarboxylation-dependent transamination reaction (O'Leary & Baughn, 1977) whereby the resulting inactive apo-decarboxylase is normally re-activated by endogenous pyridoxal phosphate. The formation of a homocysteine-pyridoxal phosphate complex could interfere with the regulatory control of the physiological action of 4-aminobutyrate, by preventing re-formation of the holoenzyme. In this study, we have shown that there is a direct interaction between homocysteine and pyridoxal phosphate. The absorption spectrum of pyridoxal phosphate has a characteristic peak at 388 nm, representative of the free aldehyde. Addition of homocysteine (0.1-0.5 KIM) to a solution of 0.1 12 M-pyridoxal phosphate (determined spectrophotometrically ; cjes = 4.9 x lo3 M ~ . cm-') produces a spectral change resulting in the appearance and concentration-dependent increase of an absorbance peak at 328nm and decrease in absorbance at 388nm. The presence of this absorption peak at 328nm suggests that a new product has been formed, since homocysteine does not absorb in this region of the U.V. spectrum. The spectral changes observed have a single isosbestic point at 345nm, indicating an equilibrium between homocysteine, pyridoxal phosphate and the homocysteine-pyridoxal phosphate complex. The formation of this complex was a time-dependent process, attaining an apparent equilibrium after 40min. It was thus possible to calculate a dissociation constant of 0.48 mM for this reaction, by using the A388 values obtained with various concentrations of homocy steine. It is proposed that the interaction of homocysteine and pyridoxal phosphate produces a thiazine derivative ofthe complex, characterized by formation of a cyclic structure in which the nitrogen and sulphur of homocysteine enter into a ring formation with the carbonyl group of pyridoxal phosphate. This hypothesis was confirmed by U.V. and mass-spectroscopic examination of a crystalline product synthesized by nonenzymic condensation of homocysteine and pyridoxal phosphate. Our results demonstrate that mouse brain glutamate decarboxylase activity is decreased in the presence of homocysteine. The observed interaction, and formation of a complex, between homocysteine and pyridoxal phosphate suggests one mechanism which could result in lower glutamate decarboxylase activity. Further experiments are necessary ( a ) to establish that glutamate decarboxylase activity is inhibited in uiuo by administration of homocysteine, (b) to show unequivocally that homocysteineinduced convulsions act through a disturbance of brain 4-aminobutyrate metabolism (this could be achieved by the use of specific 4-aminobutyrate transaminase inhibitors) and (c) to elucidate the exact mechanism(s) of glutamate decarboxylase inhibition.