A Mechanism of Sulfite Neurotoxicity


  1. Xin Zhang,
  2. Annette Shoba Vincent,
  3. Barry Halliwell and
  4. Kim Ping Wong

+Author Affiliations

  1. Department of Biochemistry, Faculty of Medicine, National University of Singapore, 8 Medical Drive, Singapore 117597, Singapore
  1.  To whom correspondence should be addressed. Tel.: 65-6874-3244; Fax: 65-6779-1453; E-mail: bchsitkp@nus.edu.sg.


Exposure of Neuro-2a and PC12 cells to micromolar concentrations of sulfite caused an increase in reactive oxygen species and a decrease in ATP. Likewise, the biosynthesis of ATP in intact rat brain mitochondria from the oxidation of glutamate was inhibited by micromolar sulfite. Glutamate-driven respiration increased the mitochondrial membrane potential (MMP), and this was abolished by sulfite but the MMP generated by oxidation of malate and succinate was not affected. The increased rate of production of NADH from exogenous NAD+ and glutamate added to rat brain mitochondrial extracts was inhibited by sulfite, and mitochondria preincubated with sulfite failed to reduce NAD+. Glutamate dehydrogenase (GDH) in rat brain mitochondrial extract was inhibited dose-dependently by sulfite as was the activity of a purified enzyme. An increase in the Km (glutamate) and a decrease in Vmax resulting in an attenuation in Vmax/Km (glutamate) at 100 μM sulfite suggest a mixed type of inhibition. However, uncompetitive inhibition was noted with decreases in both Km(NAD+) and Vmax, whereas Vmax/Km (NAD+) remained relatively constant. We propose that GDH is one target of action of sulfite, leading to a decrease in α-ketoglutarate and a diminished flux through the tricarboxylic acid cycle accompanied by a decrease in NADH through the mitochondrial electron transport chain, a decreased MMP, and a decrease in ATP synthesis. Because glutamate is a major metabolite in the brain, inhibition of GDH by sulfite could contribute to the severe phenotype of sulfite oxidase deficiency in human infants.

Sulfite is formed from sulfur dioxide, an environmental pollutant. It is also generated endogenously by the metabolism of sulfur-containing amino acids such as methionine and cysteine and from sulfate in response to bacterial lipopolysaccharide (1). The most common exogenous source is sulfiting agents used as preservatives in dried fruits and vegetables (2) and in wine where millimolar concentrations have been reported (3). Sulfite is also used as a stabilizer in many drugs administered to patients (4). Interestingly, its presence in a dexamethasone preparation increased the neurotoxicity of excitotoxic agents (5) suggesting that it could have an effect on neuronal cells. Humans can oxidize sulfite (Formula) to sulfate (Formula) by sulfite oxidase (SO),1 a mitochondrial enzyme. However, the SO activity in human liver was reported to be only 10% that of rat liver (6). Although SO is expressed in human lung (7), its activity was reported to be low (8), which may be why some asthmatic subjects react adversely to sulfite in food or atmospheric sulfur dioxide (9). The comprehensive literature over a 33-year period also shows a relationship between sulfite as food additives and asthma (10).

Sulfite oxidase is a dimeric metallohemoprotein with molybdenum and protoheme as prosthetic groups (1112). The catalytic molybdenum centers of SO from various species of animals appear to be identical as examined by electron paramagnetic resonance (1314). However, SO activity from human liver was almost 10 to 20 times lower than levels found in rat and chicken liver, and it was suggested that the decreased reactivity of the human enzyme could be due to nonfunctional molybdenum centers (6). A deficiency of SO in humans could be due to a mutation in the SO gene or in any of the several genes encoding the synthesis of molybdopterins (1518). The associated severe neurological dysfunction characterized by dislocation of ocular lenses, mental retardation, and attenuated growth of the brain suggests that neuronal cells are highly susceptible to sulfite toxicity. Indeed, SO activity measured in whole brain of some laboratory animals was consistently low compared with other tissues (1920). Measurement of the expression of SO in human tissues in our laboratory concurred with this observation (21). Four missense mutations were characterized in cell lines from patients with isolated sulfite oxidase deficiency (22). A substitution of G to A of the cDNA of liver SO resulted in an Arg-to-Gln substitution at amino acid residue 190 (23). Another arginine residue (Arg-160) was recently reported to be essential for the binding of sulfite near the molybdenum cofactors in human SO (24), whereas the residue Tyr-343 was proposed to mediate the substrate specificity and catalytic activity of the molybdoprotein (25). Despite the advances made in the molecular biology of SO, there is little information on the mechanism by which accumulation of sulfite affects neuronal function. Cell death was observed in CSM 14.1.4 (a rat neuronal cell line) following exposure to 5 mM sulfite (26). However, the mechanism of toxicity was not elucidated, although free radicals were implicated as increased toxicity of sulfite was observed when intracellular reduced glutathione was compromised (27). One electron oxidation of sulfite would produce a sulfite radical (Formula), capable of damaging DNA, lipids, and proteins (2829). In this study, we examined the effects of Formula on rat brain mitochondria and Neuro-2a and PC12 cells and attempted to elucidate its mechanism of action following our earlier observation that micromolar concentrations of Formula produced an increase in reactive oxygen species (ROS) in Madin-Darby canine kidney (MDCK) and opossum kidney (OK) cells. The sulfite-mediated oxidative stress was accompanied by a depletion of intracellular ATP, and this was thought to be due to its inhibitory action on mitochondrial glutamate dehydrogenase (30).