Document a1ML12rjy3VXw9p38e3X2d5oN

la typhimur- xiunds con3 and lheir unpson Inst. 6*1 A com ind propvl 378. ome P-450 Edited by Esiabrook. 75). Ligand Trans. J, and iigand sane cyto.'40. of hepatic cal piperorpl. Phar ma of the d by feme am. Acad. i Bot. 19, 'Her. E. C. arcinogen iectrophiii. Cancer y. E. c. & nuta genic ie known 01. Gener al metabSiochem. a rfi g & jw. 6*' * f f \ i * i * I I i fj Toluol Vol 19 pp *67 lo 64L 1tl Printed in Crcn Bntam All rifhu reserved 1111`im Copyright O UBI *--ymem Ptm Lad SULPHITE TOXICITY: A CRITICAL REVIEW OF IN VITRO AND IN VIVO DATA PLAINTIFF'S EXHIBIT A. F. Gunnison Institute of Environmental Medicine. New York University Medical Center. 550 First Avenue. New York, NY 10016, USA (Received 12 February 1981) Summary--Data on sulphite chemistry and toxicity in in vitro systems are reviewed from the perspective of potential mammalian toxicity. The observed toxicity of ingested sulphite in mammals is also summar ized and the conclusions reached are compared with the results of the in vitro experiments. Information on sulphite metabolism is included to reconcile the different conclusions that may be drawn from these 'two sets of data. Consideration of data from all sources facilitates the selection of the specific reactions of sulphite most likely lo be of toxicological significance in mammals. Introduction of sulphite among adults in Belgium was in the range Sulphur dioxide (SOj), sulphurous acid (H2SOj) and salts of sulphite (SOJ ~) and bisulphite (HSOJ) exhibit antioxidant and antimicrobial properties which make them logical choices as preservatives for certain foods and beverages. These quadrivalent-sul phur (S") substances exist in a pH-sensitive equilib rium as described by the following series of equations: ofOOOl 1-0016 mmol/kg. Because of the high concen tration of sulphite in many wines and the wide per sonal variation in wine consumption, it follows that individual consumption of sulphite also varies con siderably. Therefore, there is probably a small per centage of the population that takes in amounts of sulphite greatly in excess of these estimated averages. Sulphite is also taken into the body during the inhalation of air polluted with' S02. Although the so2 + HjO *--- h2so3==== hso3- so; - focus of this article is on ingested sulphite, inhaled SO] cannot be totally dissociated from this source since it also adds to the body burden of 'exogenous' Which of the above chemical species predominates sulphite, albeit usually to a minor degree. One can depends upon the pH and the acid dissociation con calculate, for example, that the daily intake of sulphite stants. the latter in turn being somewhat dependent due to inhalation of atmospheric SO] at the maxi on temperature and ionic strength. For example, at mum 24-hr average permitted by the EPA (0T4 ppm) 25:C the pK, of HSO; is 6-25 at high ionic strength would be approximately 25 times less than the sul and 7 2 at low ionic strength (Shapiro, 1977). Under phite taken in by drinking 250 ml of wine containing physiological conditions (considered to be pH 7 4 and sulphite at a level of 5 mu. However, although the 37`C) an essentially exclusive mixture of SOj * and subject is outside the scope of this article, it is impor HSOJ will result, with the former species predomin tant to emphasize here that the localized effects of ating, no matter which of the S" species is initially SO] on the pulmonary system may be of much introduced. For convenience and to avoid confusion, greater significance than the absolute amount of SO] 'sulphite' will be used throughout this paper for refer absorbed. ring to any of these readily interconvertible S* spe The subject material of this paper is organized into cies, However, when specific reference is made to one several areas. The first of these consists of in vitro species only, the chemical formula of that compound reactions of sulphite with biological compounds or ion <e.g. HSOj") will be used. under approximately physiological conditions. These Sulphite is used extensively in wine making as a data are considered from the perspective of their im selective inhibitor of yeasts and bacteria and is plications for mammalian toxicity. Next, the in vitro present in finished wines in concentrations up to ap modification of enzyme activity by sulphite and the proximately 6 mil. although part of this sulphite is in toxicity of sulphite in biologically active in vitro assay combined form. Foods and other beverages to which systems are considered, followed by a review of toxi sulphite is often added include dehydrated fruits, veg cological data gathered from in vivo experiments in etables and soups, as welt as fruit juices and beer mammals. The metabolism of sulphite in mammals is (Institute of Food Technologists and Committee on summarized and finally some general observations Public Information. 1976). The mean daily intake of and conclusions are presented. The literature refer sulphite from the diet has been estimated by various ences, although intended to be sufficiently extensive to sources to be between 00014 and OI4 mmol/kg body be representative, are not necessarily comprehensive. weight tin the United States), although the latter Greater detail in some subject areas can be obtained figure was reported as a probable over-estimation. from two excellent reviews, one by Shapiro (1977) and Bigwood 11973) estimated that the mean daily intake the other by Hayatsu (1976k 921001 RowVerK 098,'i 668 A. F. Gunnison Reactions of sulphite in vitro Addition to aldehydes and ketones Sulphite reacts reversibly with open-chain alde hydes and ketones to form hydroxysulphonate com pounds (Petering & Shih. 1975: Schroeter, 1966). The stability of these adducts varies considerably depend ing upon the pH and especially upon the reactive species. For example, at pH 7 the apparent dissoci ation constant (KJ of glucose hydroxysulphonate is approximately 2 2 m (Vas. 1949), while that of the formaldehyde adduct is 8 x 10"5 m (Dasgupta, DeCesare & Ullrey. 1980). Vas (1949) determined that the velocity constant for the decomposition of glucose hydroxysulphonate at pH 61 and 20C is approxi mately l/min (the value at pH 7 was not determined). Extrapolation of these constants to the in vivo situ ation, indicates that sulphite would be required to be present continuously if even a small amount of glu cose were to be maintained as the hydroxysulphonate adduct. Other physiological aldehydes and ketones, such as pyruvate, x-ketoglutarate and acetaldehyde, form hydroxysulphonate adducts that are considerably more stable than the glucose adduct. Burroughs & Sparks (1973) give the dissociation constants for pyru vate and acetaldehyde hydroxysulphonates at pH 7 and 20 C as 4-6 x 10"* m and 2-8 x IO"kM. respect ively. Ionic addition to C-C double bonds The sulphite ion adds to some C-C double bonds forming sulphonic acid compounds. Under suitable conditions this reaction has been shown to occur with several molecules of extreme biological importance. Addition to pyridine and flavin nucleotides. Sulphite adds reversibly to the 3-4 double bond of the pyridine ring of nicotinamide adenine dinucleotide (NAD), forming a suiphonate group at the active site for reduction by H". The stability of (his adduct alone at pH 7 5 (25 C) is only moderate, its dissociation con stant being approximately 3 x I0"J m (Shih & Peter ing. 1973). However, when NAD* is associated with certain enzymes, its sulphite adduct is much more stable. An analogous sulphite adduct is formed with flavin adenine dmudeotide (FAD) and flavin mononucleo tide (FMN) at the N, atom of (he isoalloxazine ring, the usual site of reduction of the ring by H'. These adducts are even less stable than the NAD-sulphite adduct (dissociation constants approximately 2 m). but as with the latter, their stability can be extensively enhanced when they are bound to protein (Muller St Massay. 1969k Since the sulphite adducts of NAD' and flavin coenzymes cannot accept H ' from the sub strate. they cannot function in their usual capacity. Examples of this inhibition of enzyme function will be discussed later. Addition to menadione. Vitamin Kj (menadione) is a water-soluble synthetic form of vitamin 1C. Sulphite, at pH 7-4. adds to the 2-3 double bond of the naphthoquinone ring of menadione to form a suiphon ate adduct Since the natural forms of vitamin K (K, and K]| are fat soluble, the reaction rates of mena dione are not necessarily applicable to them. As with other reactions of this type, the addition of sulphite to menadione is reversible. The dissociation constant for the adduct is 10"` m (Shih St Petering. 1973), indicating considerable stability of the adduct in the pres ence of free sulphite. However, the sulphite adduct is a source of vitamin K when fed to animals (Nir. Kafri St Cohen. 1978). suggesting that the adduct readily dissociates in the body. Addition to uracil and cytosine. Sulphite adds rever sibly to the 5-6 double bonds of uracil, uridine or uridine 5'-phosphate forming the 5.6-dihydro-6-sulphonate adduct (Hayatsu. 1976). The rate of the for ward reaction (formation of the adduct) is most rapid at a pH of approximately 7 and the equilibrium shifts in the direction of the reverse reaction above and below that pH. The dependency of the reaction rate on sulphite concentration at pH 7 has been demon strated in several studies (Hayatsu. Wataya. Kai & Iida, 1970; Pitman & Jain. 1979; Shapiro. Welcher. Nelson & Di Fate. 1976). According to Hayatsu et al. (1970k 86% of the uridine reactant at 22;C in I M-sulphite (excess) was in the form of the adduct after 0-5 hr. while in Ol M-sulphite only 12% of the undine had reacted in the same period of time. At this pH. the equilibrium of the reaction was decisively in the direction of the adduct and in I M-sulphite essentially all of the uridine had been convened to the suipho nate after 1 hr. However, uridine can be regenerated from its suiphonate adduct at pH 7 by removal of free sulphite. The regeneration process (i.e. reversal of adduct formation) approximates first-order kinetics. At physiological pH and in the absence of sulphite, the half-life of 5.6~dihydrouracil-6-sulphonate is several hours. Regeneration occurs more rapidly as the pH rises and at or above pH 11 it is extremely rapid, the half-life of the suiphonate compound being a few seconds (Rork St Pitman. 1974). Suiphonate adducts of cytosine and its derivatives, analogous to those of uracil are formed during incu bation with sulphite under the appropriate con ditions. The extent of adduct formation is determined primarily by pH and the concentration of sulphite. The cytidine adduct is sable at_acid pH. and in high concentrations of sulphite (about 0-5m) approxi mately 80% of cytidine is in this form at equilibrium (Shapiro. Di Fate & Welcher, 1974k Compared to uridine adducts, cytidine adducts are relatively un sable at physiological pH. only 10% of the cytidine existing in the combined form at equilibrium under conditions of excess sulphite. Therefore, a high con centration of sulphite is required to maintain the pres ence of dihydrocytosine 6-sulphonate at physiological pH. Recent dau suggest however, that this adduct may be considerably more sable in situ in bacterio phage DNA (Sklyadneva. Chekanovskaya. Nikolaeva Sl Tikchonenko. 1979ak These dau will be discussed later. Deamination of 5,Mihydrocytosine-6-sulphonate Under the appropriate conditions, 5.6-dihydrocytosine-6-sulphonate can be deaminated to the corre sponding uracil adduct This deamination is catalysed by basic substances (including sulphite) and occurs at an optimal rate at pH 5 (Shapiro et al. 1974). Thus. via the deamination reaction and the addition reac tions described above, cytosine can be converted to uracil with obvious toxicological implications. The 921001 Row Vet.. f*' <; i n .tc it for ir :catie presidjct is r. Kafri readily s res er mine or o-6-sulthe fort rapid n shifts ve and on rate '.emonKai Sc elcher, >u et at. 1 M-Suli after iridine his pH. in the ntially ilphonerated of free sal of netics. .Iphue. ate is lly as emely being natives, incucon- mtned ifphite. high roxilOrium red to un dine under i con>restical -duct terio''aeva ssed A * k i it i I I 1 1 i * I f ! i * ii. i i torrelysed rs at tus. acrd to The 1 Critical review of sulphite tonicity 669 optimal pH for the overall conversion is 5. and at physiological pH the rate of conversion declines to ap proximately 1% of that observed at optimal pH. Further, at low sulphite concentrations compatible with the in vivo situation, the rate of conversion can be expected t<T be extremely low. Slae Sc Shapiro (1978) have estimated this rate to be 4-0 x 10" "/sec for a sulphite concentration of 10"1 him. Transamination of cytosine Sulphite is capable of catalysing the transamination of cytosine and its derivatives with primary and secondary amines to produce N4-substituted cyto sines The reactive intermediate for transamination is the 5.6-dihydrocytosme-6-sulphonate adduct dis cussed previously. Shapiro Sc Gazit (1977) have car ried out transamination reactions at physiological temperature and pH between polylysine and cytidine, polycytidyiic acid and lysine, and polylysine and polycytidyiic acid. High concentrations of sulphite were used (approximately IM) and the reactions were allowed to proceed for hours or days. Crosslinking between monomers and polymers of lysine and cyto sine was observed. The significance of these reactions lies in their logical extension to the crosslinking of nucleic acids and proteins. Although attempts to crosslink double-stranded native DNA with poly lysine were not successful, evidence for crosslinking was found in a similar experiment involving ONA and histones (Shapiro Sc Gazit, 1977). Sklyadneva, Shie & Tikchonenko (1979b) were successful in crosslinking lysine with DNA isolated from S,, bacterio phage; 40?- of the cytosine residues were transaminated during a 24-hr reaction in the dark at pH 6 25. 025 M-sulphite and 25'C. Sulphitolysis reactions Sulphitolysis of thiamine. The cleavage of thiamine by sulphite was described by Williams, Waterman. Keresztesy Sc Buchman (1935). This reaction is irre versible and involves a nucleophilic attack by sulphite on the quaternary nitrogen of the thiazole ring to yield pyrimidine sulphonic acid and 4-methyl-5-0hydroxyethylthiazole. The kinetics of this reaction have been studied in detail by Leichter (1969) who showed that thiamine sulphitolysis is first order with respect to each reactant. At pH 5 and 25`C, the halflife of 10 fM-thiamine in the presence of 1 OmM-sulphite is approximately 13 hr. Petering Sc Shih (1975) have calculated from Leichters data a second-order rate constant of M x 10" j/m. sec for the sulphitolysis of thiamine at pH 7 and 25'C. Sulphitolysis ofdisulphide bonds. Sulphite can rever sibly lyse disulphide bonds by a nucleophilic displace ment mechanism resulting in the formation of thiol and S-sulphonate compounds (Cedi. 1963). Under non-denaturing conditions and at physiological pH. the reaction between sulphite and S-S bonds in free cystine goes essentially to completion (cysteine 5-sulphonate is stable at this pH) while the disulphides of most proteins are unreactive. The unreactive protein disulphide bonds are apparently protected from nu cleophilic attack either stericaily or by the unfavour able electronic environment of neighbouring amino acids. Quantitative reaction of all disulphide bonds in a specific protein can be accomplished by denaturing the protein and using high concentrations of sulphite reactant. McArdle (1967) demonstrated the reaction of sul phite with non-mercaptalbumm at physiological pH. produdng an S-sulphonate compound at the albumin-cysteine (or albumin-glutathione) site. Greg ory (1981) has recently confirmed this observation using both purified rabbit-plasma albumin and fresh whole rabbit plasma. Gregory further demonstrated, at physiological pH, the partial sulphitolysis of disul phides present in rabbit-plasma fibronectin protein. Gunnison Sc Palmes (1978) and Gunnison Sc Benton (1971) have shown that chemically stable plasmaprotein S-sulphonate compounds are formed in a matter of minutes or hours during the incubation of sulphite (approximately 0-5 mM) with the plasmas of several species of mammals at physiological pH. Free-radical reactions Sulphite can be oxidized to sulphate by free oxygen (autoxidation) via a free-radical chain mechanism which is initiated by superoxide-anion (OD or HSO}' radicals (McCord Sc Fridovich. 1969; Yang. 1970). Once initiated, chain-propagating reactions generate highly reactive free-radical intermediates such as sulphur-oxygen species (Hayon, Treinin Sc Wilt 1972). Of -, and the peroxide (HO]-) and hy droxyl (OH-) radicals. The chain-initiating radical Of - can be generated by certain enzymatic reactions and both Of - and HSOy can be produced by the reaction of transition-metal ions with oxygen (Yang, 1970). Fridovich Sc Handler (1961) have shown that the catalytic action of several oxidative enzymes (i.e. xanthine oxidase, liver aldehyde oxidase, cytochrome oxidase, lipoxidase and peroxidase) can initiate the aerobic oxidation of sulphite, while diverse other oxi dase enzymes cannot It appears that enzymes capable of initiation are those that effect the univalent reduc tion of oxygen to produce the superoxide anion. Free radicals generated by the aerobic oxidation of sulphite initiate several reactions of potential biologi cal significance (Hayatsu. 1976). The rate of these reactions is, in general, enhanced by conditions that favour autoxidation of sulphite, such as the presence of transition-metal ions (e.g Mr* ", Fe* * *) and oxy gen, and is inhibited by free-radical scavengers (e.g hydroquinone). Sulphite autoxidation occurs readily at pH 7 and reaction of the free radicals generated with various substrates proceeds more rapidly at sul phite concentrations of approximately 1-20 mM than at higher concentrations. At 1 m. for example, sulphite ions can compete effectively with potential substrates for the free radicals generated by sulphite autoxida tion. while at 20 mM this competition by sulphite is relatively ineffective due to low concentration. Con ditions favourable for the aerobic oxidation of sul phite (stated above) have been used in investigations of the reactivity of the free-radical intermediates of sulphite autoxidation. and will be referred to in the following paragraphs as a sulphite/free-radical en vironment Significant cleavage of the glycosidic linkages of uridine and cytidine. but not of purine nucleosides or pyrimidine deoxyribonudeosides. occurred in a sulphite/free-radical environment (Kitamura Sc Hayatsu. 1974). In this same system there was also extensive S21001 RowVerK 09-4 1 670 A. F Gunnison fission of the chains of polyuridylic acid [polyfUl] and polycytidylic acid [polyfC)], but not of polyadenylic acid [polyfA)] or poly(U):poly(A). Breaking of DNA phosphodiester linkages was also observed when double-stranded DNA of phage T7 was incu bated in a sulphite/free-radical environment and sub sequently treated with alkali (Hayatsu & Miller. 1972). 4-Thiouracil and 6-isopentenyladenosine (ipa) are among the minor base constituents of yeast transfer RNAs. Both of these bases are modified as a result of autoxidation of sulphite, probably directly by the sul phite-ion radical (SOf - l forming uracil 4-sulphonate and a sulphonated product of ipa. respectively. The site of sulphonale formation in the latter molecule is the unsaturated bond of the isopentenyl side chain and is an example of the addition of the sulphite-ion radical to an olefinic double bond. This class of reac tion is described in some detail by Stacey & Hams (1963). The optimum pH for the reaction is in the range 5-7. and the speed of the reaction is highly dependent upon the solubility of the olefin in aqueous medium. Ally) alcohol, for example, readily undergoes addition of sulphite via a free-radical mechanism. Methionine and certain other dialkyl sulphides are oxidized to their respective sulphoxides (R--S--R) in i a sulphite/free-radical environment (Yang. 1970). A reaction scheme has been proposed in which superox ide anion and hydroxyl radicals are responsible for sulphide oxidation. Yang (1973) has also shown that tryptophan is destroyed by free radicals generated during the aerobic oxidation of sulphite. Kaplan. McJilton Sc Luchtel (1975) have demon strated that 0-5 to 10 mM-sulphite can induce the oxi dation of a heterogeneous mixture of unsaturated fatty acids contained in com oil in a dose-dependent fashion, presumably by a free-radical mechanism. Although the oxidation products probably consisted mainly of peroxides, the possible presence of other oxidation products could not be excluded by the ana lytical method used (i.e. reactivity with thiobarbituric actd. TBA). The precise chemistry of the reactions taking place was not investigated. The mechanism of the oxidative reactions) was not dear since the for mation of TBA-reactive materials was inhibited by the presence of an antioxidant, indicating a free-radi cal mechanism, and yet Mn * * also inhibited rather than enhanced the reaction, suggesting that sulphite autoxidation was unimportant. Similar results were obtained by Inouye. Ikeda. Ishida. Ogata. Akiyama A Utsumi (1978). who measured lipid peroxidation (TBA-reactive material) in rat-liver homogenate. Addition of 2 mM-sulphite to the homogenate greatly increased the TBA-reactive material while addition of 2mM-Mn* *. even in the presence of sulphite, suppressed it. In rim modification of enzyme activity Sulphite inhibits the in trim activity of several enzymes with either NAD or flavin nucleotide cofac tors by adding to the active site of the co(actor, as previously discussed. G. Pfleiderer. D. Jeckel A T. Widand. in 1956. first postulated an enzymatically inactive complex between lactate dehydrogenase (LDH). NAD* and sulphite (from Ciaccio. 1966). Parker. Lodola A Holbrook (1978) estimated the dissociation constant for this complex <LDH-NAD*-SO,' ' ciSO,' ' + LDH-NAD*) to be approximately 10'7 m (as compared to 1-5 x I0'2 m for the non-enzymatic adduct) and con cluded that the binding of NAD* to LDH activated the nicotinamide ring for attack by sulphite by a factor of approximately 10s. The effect on the activi ties of LDH and of several other NAD*-dependent dehydrogenases (malate. alcohol glutamate and x-glycerophosphate) is apparent from the in rim experi ments of Ciaccio (1966) in which 50% inhibition was observed in the presence of 003-0-5 mM-sulphite. The sulphite adducts of flavins bound to enzymes also inactivate the enzymes and exhibit greatly enhanced stabilities relative to adducts of free FAD. FMN and model isoalloxazines. A prime example is the binding of sulphite to FAD-giucose oxidase. The dissociation constant of this complex is approxi mately 7 x 10'* m at pH 7 (Swoboda A Massey. 1966) compared to dissociation constants of approxi mately 2 m for sulphite adducts of free flavin. Although the FAD-sulphite adduct of glucose oxi dase exhibits considerable stability in the presence of sulphite, it is. nevertheless, unstable in the absence of sulphite, as was demonstrated by the regeneration of active enzyme by dialysis. This reversibility is an im portant property with regard to the toxicological sig nificance of the sulphite adducts of flavins (and NAD). Massey. Muller. Feldberg. Schuman, Sullivan. HowelL Mayhew. Matthews A Foust (1969) have cata logued a number of flavoproteins in terms of their reac tivity with sulphite. i.e. the formation of flavin nucleo tide-sulphite adducts. A series of oxidase enzymes (dand L-amino acid oxidase, oxynitrilase. lactate oxi dase and glycollate oxidase) reacted readily with sul phite. forming complexes with dissociation constants ranging from 10*2 to 10*7 m. Presumably the complexed enzymes were inactive although this was not measured. It should be noted that the flavoprotein dehydrogenases tested did not form adducts even when incubated in 20 mM-sulphite for several hours. Incubation of cytochrome oxidase with 05 or 5 mMsulphite at pH 7 for 4 hr inhibited its activity by 37 and 80% respectively (Cooperstein. 1963). The mech anism of inhibition was believed to involve one or more disulphide bonds, since other disulphide bondreducing agents, such as cysteine and reduced gluta thione. were also inhibitory. The inhibition could be reversed by incubation with oxidized glutathione. x-Glucan phosphoryiase (a-l.4-glucan:orthophosphate glucosyttransfcrase: EC 14.1.1) from rabbit muscle, which catalyses the reversible formation of glucose l-phosphate from glycogen, is substantially inhibited at pH 6 by sulphite concentrations in the 10-30 nut range (Kamogawa A FukuL 1973). This in hibitory effect is highly specific and completely revers ible by dialysis. Sulphite acts as a competitive inhibi tor (K, m 7 mst) with respect to glucose l-phosphate (K, - 11 mu) in glycogen synthesis and with respect to inorganic phosphate (PJ in glycogen degradation The authors speculate that this competition for the phosphate-binding site of the enzyme may be due to the structural similarity of HSOJ and phosphate. Harkness A Roth (1969) have reported a striking 921001 RowVarR O ' . o. 1966). imatcd ihe complex D *) to i red to ti and con-I activated I ite by a ie activi- -ependent : and i-glyi--t experii tion was i hue. io enzymes bit greatly re FAD. ample is xidase. The Is approxii Massey, i approxiirt flavin, lucose oxtsence of >sence of. ... cation of y is an im''"gical sig. i NAD). n. Howhave cataf their reacnudeormes (octate oxii with suli anstants > he comis was not avoprotein . ns even hours. . or 5 mwvity by 37 TVe mechone or ' : bondced glutai could be >ne. > lophos>m rabbit tnation of l antially in the i. This in dy revers- inhibi- i osphate ... respect gradation. ' for the due to ..ate. a striking 2- ix 1 * 4 4 % t 44 * i i 3 :1 Critical review of sulphite tonicity 471 sulphite-induced enhancement of activity of Tsuruo & Hayatsu. 1974). Treatment with alkali, 2.3- diphosphoglyccnc acid (2.3-DPG) phosphatase. which removed sulphite from the dihydro-6-sulpho- This enzyme is believed by these authors to be the nate moiety, restored the template activity of DNA. physiological phosphatase catalysing the conversion The modified DNA was prepared by heat denaturof 2.3-DPG to 3-phosphoglyceric add and P,. Since ation and reaction with approximately I M-sulphite at 2.3- DPG plays a major role in regulation of oxygen pH 6. In experiments in which calf-thymus DNA was affinity for haemoglobin, any factor that affects its con not heat-denatured (Shapiro, Braverman. Louis & centration in the red cell could be of physiological Servis, 1973), sulphite did not convert residues of importance. In the experiments of Harkness & Roth cytosine to uracil, implying that single-stranded DNA (1969) incubation of the purified phosphatase enzyme is a requisite for this reaction. with 20 mM sulphite for 1 hr at pH 7-8 and 37C pro Sulphite has been shown to inhibit the transform duced a 37-fold increase in enzyme activity; at ing activity of DNA isolated from a bacterium (strains 2- 5 mM-sulphite, activity was increased by about of Bacillus subtilis) under conditions that favour the 15-fold. Incubations with lower concentrations were aerobic oxidation of sulphite, suggesting a free radi not performed. cal-mediated mechanism of DNA alteration (Inoue. Sulphite is a potent inhibitor of most sulphatase Hayatsu &Tanooka. 1972). Inhibition was greatest at enzymes (Roy. I960). For example, the K, values for approximately 20 mM-sulphite and decreased pro sulphite inhibition of aryl sulphatases A, B and C gressively as the sulphite concentration increased to from ox.liver are 2#iM. 05 mM and 01 mM. respect I m. Other factors that decreased the rate of sulphite ively. If precautions are taken to minimize the aut- auloxidation. such as free-radical scavengers and oxidation of sulphite, then K, for inhibition of sulpha elimination of oxygen, also decreased the sulphite- tase A decreases to approximately 02 fM (Roy, 1976). mediated inhibition of transforming activity. Inactiva The early work on the kinetics of aryl sulphatases was tion of the DNA transforming activity was strongly performed with nitrocatechol sulphate and other inhibited by 4-thiouridine which is known to react unphysiological substrates without certain knowledge with the sulphite-ion radical. Thus, this radical was of the true physiological substrates. More recently, assumed to be the species primarily responsible for the physiological substrates for the sulphatases have the inactivation. been thought to include lipids containing galactosyl A free-radical mechanism was also implicated in 3- sulphate residues (such as cerebroside sulphate), the sulphite-mediated inactivation of bacteriophage mucopolysaccharides containing N-acetylgalactos- lambda (Kudo. Miura & Hayatsu. 1978). Inactivation amine 4-sulphate residues (such as dermatan sul of phage infcctivity of indicator bacteria was ob phate). heparin sulphate, chondroitin sulphates and served in sulphite concentrations of 01-10 mM in steroid sulphates (Roy, 1976). incubations at pH 7 and 37'C. After a 4-hr incubation in 10 mM-sulphite, the infectivity of the phage had de creased to 10'* of its initial value. Phage inactivation Sulphite toxicity in biologically active, in vitro test sys was attributed not to DNA damage but to alteration tem of coat proteins; this affected their adhesion of bac Modification ofactivities of RNA, DNA and associated proteins teria and their ability to inject DNA. Several lines of evidence suggested that tryptophan in the coat pro teins was modified by reaction with the sulphite-ion Shapiro & Braverman (1972) have demonstrated radical. that conversion of uracil to the 6-sulphonate adduct Turchinsky, Kusova & Budowski (1974) have interferes with hydrogen binding to adenine and demonstrated the sulphite-catalysed-crosslinking of reduces the ability of poly(U) to form a helical the maturation and coat proteins with the nucleic complex with polyfA). The uracil-sulphonate adducts acids of the RNA bacteriophage, MS2. The MS2 were formed in poiy(U) by reaction with I M-sulphite phages were treated with 1 M-sulphite at pH 7 for at pH 7, followed by stabilization at pH 4 and dialysis 0-5-4 hr, with consequent covalent association of ap to remove excess sulphite. The modification of uracil proximately 1% of the protein with RNA. The mech residues of polyfU) also inhibited its ability to code for anism of crossiinking was presumably by transamina phenylalanine incorporation into protein in an tion of S.6-dihydrocytostne-6-sulphonate as discussed Escherichia coti cell-free protein-synthesizing system in a previous section. operating at approximately physiological pH. A rela Sklyadneva el al. (1979a) have also presented evi tively small percentage of adduct formation (26%) dence for the sulphite-catalysed transamination of caused a much larger decrease in incorporation (54%). cytosine bases with protein in bacteriophage DNA. leading the authors to suggest that a single uracil These authors propose that the intermediate. saturation might be sufficient to block translation at that point. This same research group later showed 5,6-dihydrocytosine-6-sulphonate. is stabilized m situ by polar groups of protein, especially the amino that sulphite could similarly modify natural mes groups of basic amino adds. This stabilization is senger RNA (from coliphage MS2) and ribosomal effective at refrigerator temperatures for up to 4 RNA from . coli, leading to decreases in the incor months. When phage panicles are disintegrated, how poration of amino adds into protein (Braverman. Shapiro & Szer. 1975). in related experiments, uraal-sulphonate adducts formed in calf thymus DNA by deamination of cyto sine residues, interfered with the DNA polymerase reaction, thus inactivating DNA as a template (Kai. ever, the sulphonate adduct in the DNA becomes un stable and, depending upon the speafic conditions, either reverts to a cytosine residue or undergoes trans amination. This situ stability of cytosine-sulphonate adducts may have important implications for the in vino rate of deamination of cytosine to uracil. 921001 RowVerK ojum 672 A F. Gunnison Chromosome damage and mutagenesis of C:G to A:T transitions (i.e. hot spots). Further As anticipated from chemical data, sulphite was more, in a strain of . coli that lacks uracil-DNA shown to cause mutations, presumably by deamina glycosidase. the rate of spontaneous transitions at tion of cytosine to uracil. Cultures of several mutant cytosine residues is elevated to the rate observed at strains of E. coli were treated with 1 M-sulphite at S-methylcytosine residues (Duncan St Miller, 1980). pH 5 2 for 30 min and the frequency of back mutation On the other hand. Wang. Gehrke St Ehrlich (1980) was determined (Mukai. Hawryiuk St Shapiro, 1970). have recently demonstrated that, at pH 5-5 and 3 m- Only those mutants that were cytosine-guanidine sulphite. under conditions where more than 96?; of (C.G) at the mutant site showed an increase in rever cytosine residues in single-stranded DNA were con sion frequency. When incubations were performed at verted to uracil only 2-3% conversion of 5-methyl- pH 7 or 8, sulphite had no measurable effect on rever cytosine residues to thymine occurred. Wang St sion frequency, a result consistent with chemical data Ehrlich (1980) concluded that it is much more likely on the pH profile for deamination. that cytosine rather than 5-methylcytosine residues A similar specificity for C:G to A:T (adenine- are involved in sulphite-induced mutagenesis. This thymine) transitions was reported by Summers & point is the subject of continuing research and is still Drake (1971) using bacteriophage T4rll as a test sys open to question. Preliminary experimentation tem (although T4 contains the cytosine analogue necessary for calculating the rate of 5-methylcytosine 5-hydroxymethylcytosine). At pH 5, inactivation and deamination under physiological conditions is now mutation frequency of the phage showed excellent underway in the laboratory of Dr R. Shapiro (per dose-response relationships with both sulphite con sonal communication, 19801 centration (02-0-9 m) and treatment time. The rever Recently, Mallon & Rossman (1981) have demon sion frequency resulting from a 4-hr treatment with strated an enhancement of UV mutagenicity resulting 09 M-sulphite was approximately 110/107. Recent from exposure at physiological pH to much lower duplication of these experiments, however, revealed sulphite concentrations than are required for measur the initial findings to be in error. Sulphite was not able conversion of cytosine to uracil. Cells from a capable of causing a measurable rate of reversion of Chinese hamster line, V79. exposed to IOmM-sulphite T4 phage, although a 10-20-fold lo*er mutation rate at pH 7-4, either during or immediately following UV than that initially reported could not be excluded irradiation, showed an approximately twofold in (J. W. Drake, personal communication 1981). crease in mutation frequency over that caused by UV An increase in mutation frequency of phage lambda treatment alone. In similar experiments with . coli, incubated in 3 M-sulphite at pH 5-6 was observed by 100 mM-sulphite caused an eight-fold increase. In both Hayatsu St Miura (1970). Maximum mutation fre cases, exposure to sulphite alone had no effect on quency was produced after 1-S hr. while inactivation mutation frequency. Since the co-mutagenic effect of of the phage continued for the 3 hr duration of the sulphite was essentially of equal potency whether sul- experiment. ; phite exposure occurred during or immediately fol In his review of the genetic effects of sulphite. lowing UV irradiation, Mallon St Rossman (1981) Shapiro (1977) cites two reports in which sulphite ap speculated that sulphite might have been affecting a parently caused mutations in Saccharomyces cerevisiae DNA repair process. Subsequent experiments utiliz and Micrococcus aureus at much lower concentrations ing two . coli strains deficient in the excision repair than are usually required (i.e. 5 and 10 dim, respect process demonstrated that the presence of sulphite ively). In' the former case, however, a very low pH (3-6) did not enhance UV mutagenesis, implying that the was required. function of sulphite in the initial experiments was to In the above experiments, sulphite-induced muta- ~ inhibit excision repair. genicity was observed in cells containing double- Perry Sc Evans (1975) have demonstrated a positive stranded DNA. This appears to be inconsistent with correlation between mutagenesis and sister chromatid other experiments in which the cytosine bases in iso exchange and suggest the latter as a highly sensitive lated double-stranded DNA were inert to sulphite technique for assaying the chromosome mutagenicity (Shapiro. 1977; Shapiro et al. 1973). In reality, how of environmental agents. MacRae St Stich (1979) ever. DNA. especially in certain growth phases, showed that sulphite induces dose-related sister chro always exists partially in its reactive single-stranded matid exchange in Chinese hamster ovary cells at font) (BjurseU. Gussander St Lindahl, 1979). concentrations between approximately 003 and 7 mM. The proposed mechanism of sulphite-induced The potency of this induction, however, was much mutagenesis, that is conversion of cytosine to uracil, is lower than that exhibited by the strong mutagenic not consistent with the existence of uracil-DNA gly- agents shown to possess this ability. MacRae St Stich cosidasc. This enzyme, discovered initially in . coli (1979) suggested that cleavage of the DNA chain by (Lindahl 1974), catalyses the excision from DNA of free radicals, probably hydroxyl radicals, generated by uracil bases produced by deamination of cytosine, autoxidation of sulphite (Hayatsu St Miller. 1972) thus apparently preventing C:G to A:T transition. might be responsible for this action of sulphite. Recent data suggest that sulphite-induced mutations In a different type of assay, damage to the chromo at C:G sites actually involve deamination of 5- somes of mammalian oocytes was observed following methylcytosme to thymine. Since the newly-formed in cairo exposure to sulphite (Jagiello. Lin St Ducayen. thymine in DNA is not recognized by a glycosidase. it 1975). In the same experiments there was also an inhi cannot be excised and repaired. Coulondre. Miller, bition of entry of oocytes into meiosis when they were Farabaugh St Gilbert (1978) have demonstrated that cultured in the presence of sulphite. Mouse oocytes, in S-methylcytosine residues in . coli are associated general were more sensitive than those of the cow or with a high rate of spontaneous mutation consisting ewe and when exposed to sulphite concentrations of 921001 RowVerK Furthericil-DNA iitions at served at 1980). ch (1980) and 3 mi 96% of vere con5-methylWang Sc ore likely residues :sis. This nd s still neruation Icytosine s is now >iro (per- : demonresulting ch lower - measurs from a -sulphite ing UV .oid ind by UV h . co/i. :. In both effect on effect of :ther suljteiy fol io (1981) Tecting a its utilizon repair sulphite that the > was to positive tromatid sensitive agenicity h (1979) ter chro- cells at d 7 mM. s much jtagenic Sc Stich .ham by rated by r. 1972) te. chromoollowing "Hicayen. an inhiey were cytes. in . cow or ..ions of * T 4Sr f f *X -# tSt 1 5 : 4t * I* 1V -MJb f i 1! -I 't : .. i i Criuc.il review of sulphite toxicity 673 approximately 02mM or above showed a slight inhi bition of entry into meiosis which became complete at a sulphite concentration of 6 urn. Chromosome 'fuzzi ness' also occurred at 0-2 rim and higher concen trations. but the authors ascribed no genetic signifi cance to this, type of aberration. More significant genetically was the fragmentation of chromosomes and the anaphase lagging that occurred sporadically in cow and ewe oocytes at concentrations of sulphite of approximately 3 rim and above. Effects on mammalian cells in culture The adhesion of cultured Chinese hamster cells (cell line Don) to the substratum was slightly inhibited by incubation with 10mM-sulphite and severely inhibited by 30 or SO mM-sulphite (Kudo. Hayatsu, Yokoiyama Sc Kuroda. 1980). A possible explanation for this ob servation has been offered by Gregory (1981) who found that sulphite concentrations at or above 30 mM progressively cleaved disulphide-linked rabbit-plasma fibronectin dimers into monomers. A plasma fibronectin-type protein is also present in disulphidebonded aggregates in relatively large quantities on the surface of cells (cell surface protein: CSP) where it functions in cell-substratum adhesion and cell-cel! in teractions. Disruption of CSP by strong disulphide reducing agents has been shown to destroy its func tions (Ali Sc Hynes. 1978). An observation by Kikugawa Sc fizuka (1972) that 7 5 mM-sulphite inhibits ADP- and collagen-induced aggregation of rabbit platelets may well be linked to the above finding of Kudo et al. (1980). by similar underlying mechanisms. Thompson Sc Pace (1962) measured cell prolifer ation in mouse fibroblasts, mouse-liver cells and HeLa cells exposed in culture to initial sulphite concen trations ranging from approximately 1 to 20 mM for periods up to 9 days. There was complete inhibition of growth in all cell lines cultured in the high sulphite concentration, while HeLa cells, which were the most sensitive to sulphite, showed marked growth inhi bition even at 1 mM-sulphite. Results consistent with these findings were obtained by Das Sc Runeckles (1974), although synchronous cultures of Chlorella pyrenoidosa were used. In these studies, exposure of cells to initial sulphite concentrations of 0-5-2 m.M for a period of 48 hr caused a progressive decrease in cell number and in DNA content, expressed as a percent age of dry weight but not in RNA or protein content. During the 48-hr exposure period, there was a decline in pH from 6-6 to approximately 4 (measured in a culture containing about 2 mM-sulphite) which may have affected entry of sulphite into the cell. The authors concluded that sulphite affected active growth of the cells by impairing DNA synthesis. Inhibition of DNA synthesis (measured by [JH]thymidine incorporation) was also observed by Chin. BisseU Sc Bassham (1977) in chick-embryo fibroblasts cultured in the presence of sulphite for 18 hr. While 005 mM-sulphite caused no measurable inhibition. Ol and 1-0 mM-sulphite caused a 14 and 52?; reduction in thymidine incorporation, respect ively. After incubation for 48 hr. decreased cell vi ability was observed in 05 and 1-0 mM-sulphite. Sul phite did not affect cell-membrane permeability to mannitol (passive diffusion) or to 2-deoxyglucosc (earner-mediated transport). In addition, no discern ible effect of sulphite on glucose metabolism was revealed by monitoring selected intermediates of the glycolysis pathway, glycogen synthesis, the pentose shunt and the tricarboxylic acid cycle. Timson (1973) cultured human lymphocytes in 01-10 mM-sulphite and concluded that exposure to 10 mM-sulphite for 72 hr was cytotoxic while lower concentrations had an antimitotic effect which was possibly due to inhibition of DNA synthesis during the early stages of mitosis. In this system, exposure to 01 mM-sulphite over a period of 72 hr caused 43% inhibition of mitosis, and a similar exposure to 10 mM-sulphite produced a 63% inhibition. In a perplexing study which is somewhat difficult to relate to others of its kind. Schneider Sc Calkins (1970) exposed human lymphocytes in culture to sul phite by bubbling SO; through the medium. Although the sulphite concentration was not measured either at the onset of exposure or during the incubation period, which lasted for up to 3 days, it is possible to calculate from the information given that the initial concentration was approximately 0005 rim. assuming complete absorption of S02 by the medium. The pH during the incubation period varied between approximately 6-6 and 7-2. There was a significant decrease both in DNA synthesis, as measured by the percentage of cells that incorporated [JH]thymidine, and in the mitotic index of cells exposed to sulphite in comparison with appropriate control cells. These effects were most apparent in cells exposed before DNA synthesis was initiated. In addition, chromoso mal abnormalities, consisting mainly of a reduction in number and in clumping and fuzziness, occurred at a higher incidence in the sulphite-exposed cultures. These experiments demonstrated detrimental effects at an estimated concentration of sulphite approxi mately two orders of magnitude below those reported by other investigators as causing similar changes. It is felt, however, that the results of Schneider Sc Calkins (1970) are open to question because the authors neg lected to test other concentrations of sulphite to dem onstrate conclusively that the effects observed were truly a function of sulphite exposure. The delivery of the sulphite dose by bubbling S02 through the medium introduced the possibility of side effects due to physical damage of ceils and/or oxygenation of the culture medium. Although these side effects were sup posedly controlled for in cultures that received air only, the possibility of a synergistic effect between sul phite and the bubbling of air through the medium was not considered. A 50% reduction of the intracellular concentration of 2,3-DPG in human erythrocytes resulted from incubation of the cells for 4 hr with 4 mM-sulphite at pH 7-5 and 37*C (Parker, 1969). Essentially no intra cellular 2J-DPG remained when the concentration of sulphite was increased to 20 nut These data can be explained by the sulphite-induced activation of 2,3-DPG phosphatase of human erythrocytes dis cussed previously (Harkness Sc Roth, 1969). The de pletion of 2J-DPG from erythrocytes was ac companied by a reversible increase in their passive permeability to Na and K ions, a change that was thought to be a direct effect of sulphite and not to be caused indirectly by 2J-DPG levels. 921001 RowVerl 0 H H.' > 674 A. F Gunmson Mammalian toxicity Chrome sulphite feeding studies In studies in which sulphite was administered to animals (usually rats) in the diet or dnnking-water, the concentration of sulphite was often expressed in differenr units, making direct comparisons difficult. Therefore, in this section, an attempt is made to express (he sulphite exposure in all experiments in terms of mmol consumed/kg body weight/day. Since some investigators have not given the data on food or water consumption required for these calculations, it has been assumed where necessary that rats ingest 100 ml water and 67 g solid diet/kg body weight/day. Further, the instability of sulphite added to the diet or drinking-water has been noted and taken into account by some investigators in their calculations of intake, and ignored by others. No attempt is made here to adjust intake figures for this variable unless adequate data on stability have been given. It is well documented that sulphite, when pre-mixed with the diet, can cleave the thiamine molecules con tained therein and destroy their activity, thereby caus ing a deficiency of thiamine in the organism (Bhagat St Lockett, 1964; Fitzhugh. Knudsen St Nelson. 1946). Thiamine deficiency has not resulted, however, when sulphite has been administered in fluids, although there is evidence that thiamine can be destroyed in the stomach when ingested simultaneously with sul phite (Lhuissier, 1966k Furthermore, the possibility that sulphite may inhibit the synthesis of thiamine by the bacterial flora of the intestine cannot be excluded (Cremer St Hotzel 1966). In spite of this destruction by sulphite, recent data show that thiamine is not destroyed systemically by sulphite (Gunnison, Dulak. Chiang. Zaccardi St Farruggella, 1981a). One of the earliest attempts to investigate the chronic toxicity of sulphite comprehensively was pub lished in 1946 by Fitzhugh et al. These investigators administered sulphite to rats for approximately 1 yr by incorporating it into their diet at several concen trations. resulting in nominal intakes in the treatment groups ranging from 008 to 13 mmol/kg/day. Unfor tunately, these estimates of sulphite intake are mean ingful only as ceiling values, since sulphited food was sometimes left in feeder cups for up to a week between changes, during which lime as much as 75% of the sulphite was lost due to chemical reaction. In these experiments the investigators attempted, not en tirely successfully, to separate the effects of sulphiteinduced thiamine deficiency from those due to the ageing of the sulphited diet and to the direct toxicity of sulphite. They concluded that, in addition to the toxic signs attributable to thiamine deficiency, such as polyneuritis, the stunting of growth and atrophy of organs (the tatter two due to inanition), other toxic changes were produced by ingestion of the sulphited diets. At a nominal sulphite intake of 1-6 mmol/kg/ day or more, the growth rate and average survival time of rats was decreased and pathological changes including gastric squamous epithelial hyperplasia, bleached incisor teeth, brown uteri, calcified renal tubular casts and atrophy of bone and bone marrow were observed. It was not dear, however, which of these signs could be prevented or mitigated by thiamine therapy, or whether residual toxicity was caused by the interaction of sulphite with constituents of the diet and/or the direct action of sulphite on the organism. In I960. Lockett St Natoff. administering sulphite chronically to rats in their drinking-water at an intake rate of approximately 1-2 mmol/kg/day, ob served none of the toxic signs listed by Fitzhugh et al. (1946). Their obvious conclusion, therefore, was that sulphite was not directly responsible for this toxidty. Bhagat St Lockett (1964) later attempted to clarify the roles of thiamine destruction and storage of diet in the development of toxidty resulting from feeding sulphited diets to rats. In these experiments, the growth rate of young rats over a 5-7-wk period was used as an assay for toxidty. Diets contained approxi mately 130 /ig thiamine/100 g and 06% sodium meta bisulphite at the time of mixing (equivalent to an intake of approximately 5-9 mmol sulphite/kg/day. assuming no loss of sulphite prior to use) However, subsequent storage at room temperature resulted in the rapid and parallel destruction of both sulphite and thiamine content, and after 8 days only approxi mately 20% of their initial concentrations remained. Young rats fed this diet within 2 months of its prep aration showed a decreased growth rate, which could be corrected by thiamine supplementation in spite of weak antithiamine activity resulting from the ex posure of dietary yeast to sulphite. Diets that were stored at room temperature for 75 days or longer caused toxidty in the form of a reduced growth rate and diarrhoea, which could not be completely cor rected by thiamine supplementation. Although resi dual sulphite concentrations in the diets at the time of consumption were not measured, it seems almost cer tain that the sulphite would have been too low after 75 days of storage to be a factor in the development of the observed toxidty. The most thorough investigation of chronic sul phite toxidty to date is the work of TiL Feron St de Groot (1972a) and Til, Feron, de Groot A van der Wal (1972b) using rats and pigs. Sulphite was admin istered in the diet, and losses due to chemical reaction prior to feeding were kept to a minimum by frequent diet preparation and storage at low temperature. The loss of sulphite that did occur was measured and the dietary concentrations were corrected accordingly. In addition, thiamine was added to the diet to compen sate for its sulphite-mediated destruction, thus ensur ing that any toxicity observed during the experiment would not be due to thiamine deficiency. Sulphite was administered to three generations of rats for periods up to 2 yr at intake rates of approxi mately 0-7, 1-5. 3, 6, and 13 mmol/kg/day Tor the five treatment groups, and the health of these animals was compared with that of a matched control group receiving the same diet with no added sulphite (Til et al. 1972a). Slight growth retardation was observed in the F,- and Fj-generation rats of the high-dose group and marginally reduced haemaloent. haemoglobin and erythrocyte counts also occurred in F0 rats at this treatment level Occult blood was present in the faeces of approximately 20-50% of rats ingesting 6 mmol/kg/day. Abnormal morphology of the forestomach was observed in some Fi-generation rats ingesting 3 mmol/kg/day. and in rats of the two high est treatment groups more severe hyperplasia and inflammation of both the fore- and glandular stomach 92V101 RowV* onstnuenu hue on the nimstermg ig-water at -g 'day. obhugh er al. was that is toxicity. 1 to clarify age of diet >m feeding nents. the icriod was d approximm metaent to an ite/kg/day. However, esulted in h sulphite / approxiremained. f us prepiich could n spite of i the exthat were t longer >wth rate etely cor'ugh resi de time of Imost cerlow after elopment ic sul& de an der s adminI reaction frequent ure. The . and the lingly. In compenis ensurjeriment ations of lpproxithe five mals was >1 group e (Til er trved in se group noglobin s at this in the iigesting he foreon rats o high*ia and tomach 's* :si i VV* , . Critical review of sulphite toxicity *75 was observed. There were no dose-related trends in tumour frequencies among F0- and F,-generation rats that died during the experiment or were killed at 2 yr of age. Likewise, sulphite had no effect on the indices of reproduction and early development investigated, i.e. on fertility (% of females with litters), mean number of pups per litter, birth weight and mortality prior to weaning. In F2 generation females, kidney weight relative to body weight was increased in the high dosage group, although kidney function, as measured by phenol red excretion, urine specific grav ity and glutamic-oxalacetic transaminase activity in the unne, was not adversely affected by sulphite feed ing. Increased relative weights of kidneys have also been observed in other chronic and subchronic sul phite toxicity studies. The no-effect level determined from the experiments of Til er al. (1972a) was equival ent to an intake of I $ mmol/kg/day. This level of consumption, adjusted by a 100-fold safety factor, has been adopted by WHO as its maximum acceptable daily intake (ADI) level, i.e. 0-70 mg (as S02)/kg body weight (Joint FAO/WHO Expert Committee on Food Additives, 1974). In a companion experiment to the one described above, Dutch Landrace pigs were fed sulphite from weaning for periods of up to 48 continuous weeks (Til er al. 1972b). The same nominal concentrations of dieiary sulphite were used for these animals as for the rats, but actual intakes were considerably less, ap proximately 0-1, 0-3. 06, Hi and 3-6 mmol/kg/day in the five treatment groups. The thiamine added to the diet was sufficient to prevent deficiency in ail groups except that on the highest dose, in which a slight reduction in hepatic thiamine level occurred. Growth and food consumption were significantly decreased only in the highest dose group (3-* mmol/kg/day). Paired-feeding studies showed that this decreased rate of growth was due solely to decreased food consump tion and was not a direct cause of ingested sulphite. Increases in the relative weights of liver and kidney in this same treatment group, however, were attributable to sulphite consumption, but were not accompanied by histological changes. As in the experiment with rats, the most damaging effects of sulphite were the inflammatory and hyperplastic changes in the stomach mucosa of animals in the two highest dosage groups, although no occult blood was found in the faeces of these animals. Several other chronic feeding studies of merit conducted in rats will not be reviewed here in any depth (Oman, Causeret St Hugot. 196$; Lanteaume, RameL Girard. Jaulmes, Gasq St Ranau, 196$: Lock ett St Naioft, I960), in all of these studies sulphite was either added to the drinking-water or given in solution by gastric intubation, resulting in daily intakes of between 00$ and 2 mmol/kg. Usually several generations of rats were treated and all ex periments lasted for at least 1 yr. in these studies, investigators measured the effect of sulphite on such parameters as fertility, the general health and growth rate of offspring, organ weights, haematology, food intake and growth rate of adults, histological appearance of major organs and tumour incidence. In general, no consistent trends attributable to sul phite exposure were apparent in any of these par ameters. > Short-term studies 1 - Investigations by Til et al. (I972a| of the toxicity of high doses of sulphite ingested with the diet over rela tively short periods, showed that the food intake, food efficiency and growth rate were drastically reduced in young male rats ingesting approximately SOmmol sulphite/kg/day for 8 wk. In addition, severe anaemia, increased spleen weight and a slightly increased leuco cyte count were observed after only 3 wk. Gunnison er al. (1981a) have confirmed these observations and have shown that the anaemia results not from the systemic activity of sulphite following ingestion but from the interaction of sulphite with a constituents) of the diet, possibly cyanocobalamine. In other short-term studies of up to 4 months dur ation, daily sulphite intakes ranging from approxi mately 0-$ to 6 mmol/kg caused a slight increase in the excretion of calcium and had no effect on the hepatic stores of vitamin A (Joint FAO/WHO Expert Committee on Food Additives, 1974; Lanteaume, Morin, Palluel & Pallaget, 1978). The effect of sul phite on calcium excretion is probably of little or no physiological significance. In thiamine-deficient rats, small amounts of sul phite administered separately from the diet depressed weight gain and survival, while in rats not deficient in thiamine, sulphite intake up to 6-2 mmol/kg/day had no effect on weight gain over a 3-month period (Cremer St HotzeL 1966). This observation may help to explain the toxicity of sulphite reported by Fitzhugh er al. (1946) and discussed earlier. In contrast to the rat data, approximately 01 mmol sulphite/kg/day admin istered in imbibed fluids for 25 consecutive days to human volunteers with thiamine deficiency (deter mined by biochemical signs) caused no clinical, neuro physiological or biochemical alterations compared with controls (Cremer St HotzeL 1970). The bulk of evidence accumulated from both long- and short-term studies supports the view that sulphite administered in fluids separately from the solid diet does not measurably reduce the thiamine status of the animal In spite of evidence of sulphite-induced chromo some aberrations resulting flora in vitro exposure (see previous section), sulphite did not induce a detectable increase in dominant-lethal mutations in either male or female germ ceils of mice receiving repeated daily intraperitonea! injections of 2-9-4-8 mmol/kg (Gener- oso. Huff Sl Cain, 1978). Nor were Jagiello er al. (197$) successful in inducing chromosome aberrations in mouse oocytes cultured in vitro following a single intravenous injection of up to approximately 2 mmol sulphite/kg. Reviewing the data from experiments designed to evaluate the mammalian toxicity of ingested sulphites leads to the conclusion that apart from the indirect toxicity resulting from destruction of dietary thiamine or other changes in the diet, and the direct irritant effect on the gastro-intcstinal tract at relatively high intake levels, no serious adverse effects were observed as a result of chronically administered sulphite. This conclusion is surprising in the light of the reactivity of sulphite with many biologically important molecules and the toxicity of sulphite observed in biologically active in vitro test systems. The difference between the toxic potential of sulphite perceived Grom in vitro data 921001 RowVerK UU&-, 676 A. F Gunnison and the toxicity observed in in vivo experiments has been noted previously (Shapiro. 1977); the probable explanation becomes apparent with an understanding of sulphite metabolism. Mammalian sulphite metabolism The primary route of sulphite metabolism in mam mals is its enzymatically mediated oxidation to sul phate. The enzyme involved. sulphite:cytochrome c oxidoreductase (EC 1.8.3. U, termed sulphite oxidase, is apparently ubiquitous among mammalian species and is present at high levels in the liver and in lower concentrations m most of the other tissues of the body. Sulphite oxidase, located in the mitochondrial intermembranous space, exists as a dimer of identical subunits, each consisting of a molybdenum ion (Mo"*) and a haem molecule in addition to the apoenzyme. The in vivo oxidation of sulphite involves the transfer of i pair of electrons from sulphite to the Mo6* Ions and then to the haems associated with the enzyme molecule. The electron pair is then passed to cytochrome c of the respiratory chain, eventually reducing 05 02 to H20 and producing I molecule of ATP in the process. Although sulphite can be autoxidizcd by a free-radical chain mechanism under appro priate conditions, it is thought that this reaction sequence does not proceed readily in mammalian tissues for a variety of reasons (Cohen it Fridovich, 1971). In addition to being the major metabolic pathway of exogenous sulphite (ingested sulphite and inhaled S02). enzymatically mediated oxidation of sulphite to sulphate is the terminal step in the catabolism of sul phur-containing amino acids. Thus, an enzyme that apparently evolved to protect the tissues of the body from the insult of endogenously-produced sulphite also functions in the metabolism (and presumably the detoxification) of this same substance originating from exogenous sources. Assuming that the sulphate excreted by animals in 'sulphur balance' originates primarily from the catabolism of sulphur-containing amino acids, it can be estimated from data on sul phate excretion that the daily quantity of endogenous sulphite generated by humans is approximately 03-04 mmol/kg (Institute of Food Technologists and Committee on Public Information, 1976). Under nor mal circumstances this is considerably greater than the estimated intake of exogenous sulphite (see Intro duction). The capacity of mammalian sulphite oxidase for sulphite oxidation is extremely high compared with the normal sulphite load from endogenous and ex ogenous sources. For example, Cohen. Drew, Johnson Sl Rajagopalan (1973) have estimated by in vitro assay that the sulphite oxidase contained in the tissues of the rat is theoretically capable of oxidizing sulphite at the rate of approximately 750 mmol/kg/day. Also, Oshino it Chance (1975) and Wilkins, Greene St Weller (1968) demonstrated that the perfused livers of rats and dogs can oxidize sulphite for short periods of time at rates of at least 58 and 18 mmol/kg/day. re spectively. Using established pharmacokinetic tech niques. Gunnison, Bresnahan St Palmes (1977) inves tigated the rate of sulphite oxidation in intact rats following rapid intravenous delivery. Elimination of sulphite which occurred predominantly by metab olism to sulphate, was characterized by first-order rate constants of the order of 07-l/min which is equivalent to a half-life for sulphite of approximately 1 min. Further, Gibson St Strong (1973) were unable to detea sulphite in the urine of rats following admin istration of approximately 6 mmol/kg by gastric intu bation. Since sulphite is readily absorbed from the gastro-iniestinal tract (Bhagat St Lockett. I960), this observation attests to the capacity of rats to metab olize systemic sulphite rapidly. Gunnison er al. (1977) compared the activity of sul phite oxidase in rats with that in rabbits and rhesus monkeys using in vivo kinetic methods, and demon strated that rats exhibit approximately three and five times greater aaivity, respeaively, than the latter two species. In addition, Johnson St Rajagopalan (1976a,b) have shown by in vitro assay that rat liver possesses approximately 10-20 times more sulphite oxidase aaivity than does human liver. These com parisons suggest that the rat may be a poor species for the evaluation of sulphite toxicity in humans since the opportunity for potentially damaging reaaions of sulphite is comparatively less in the rat due to more rapid metabolism of sulphite to sulphate. Although the rapid rate of oxidation of sulphite by sulphite oxidase certainly minimizes the quantitative importance of other metabolic pathways, rats injected intraperitoneally with approximately 3 mmol/kg/day and rabbits and rhesus monkeys ingesting approxi mately 2 mmol/kg/day nevertheless metabolized a portion of this exogenous sulphite to 5-sulphonate compounds in the plasma (Gunnison St Palmes, 1978). Indeed, the rats and rabbits possessed detect able concentrations of plasma 5-sulphonate com pounds prior to exposure to exogenous sulphite, indi cating that the sulphite generated endogenously from sulphur-containing amino add catabolism was also partially metabolized via this route. Because of their relative biological stability, these 5-sulphonate metab olites were evident in plasma, which does not usually contain detectable free sulphite. When sulphite is present Jn the tissues in suffi ciently high concentration, it "reacts with 0-mercaptopyruvate (a normal intermediate in sulphur-amino add catabolism) forming inorganic thiosulphate (S2Oj ~). This metabolite is, at most, marginally de tectable in the urine of normal humans and rats, while in both these species large quantities are exacted into the urine of individuals that are deficient in sulphite oxidase and have, therefore, relatively high systemic levels of sulphite. Throughout the literature on the investigation of mammalian sulphite toxidty there is an almost total absence of any attempt to determine tissue sulphite concentrations. This seems inexplicable, particularly since concentration is the most logical and appro priate basis for the correlation of findings from in m<o experiments with data gathered in vitro The limited data existing on in duo sulphite concentration comes from studies of sulphite metabolism in normal and sulphite oxidate-deficient mammals. Gunnison & Palmes (1973 St 1978) determined free sulphite in the plasma of several spedcs, both prior to and during administration of sulphite. Sulphite ongtnating from endogenous sources was not detectable (i.e. was less than 3 jot) and when exogenous sulphite was adminis- 921001 RowVeirl. first-order which is oximately ere unable ing adminstr c intuJ fr. m the I960), this to metab- v tty of sul* tnd rhesus id demonce and five c latter two jagopalan i rat liver sulphite <se comor species rtans since acr ons of ue tc more Jlphite by antttative -its injected nol/kg/day anproxiX3 ed a sulp nonate & Palmes, id detect*te com pile. indiously from was also of their e rietabi usually in suffitercaptohur-amtno iiosulphate tally de ls, while creted into n sulphite systemic gaiion of nost total sulphite ncularly u approom in tiro limited i comes --ul and imson St ''7 in the during ig from was less - idmtrus- -*-T *; iff i ^\ i 'If i <r` f t Critical review of sulphite tonicity *77 tered in the drinking-water of rats, rabbits and rhesus monkeys at a rate of approximately 2 mmol/kg/day and of mice at approximately 6 mmol/kg/day, plasma sulphite was detected only in the rhesus monkeys. Of the plasma samples collected from six rhesus monkeys at various times during the light cycle (i.e. primary drinking time), sulphite ranging in concentration from 8 to 67 jim was detected in 10 of 23. Gunnison, Farruggella, Chiang, Dulak, Zaccardi St Birkner (1981b) administered sulphite to rats (2-9 mmol/kg) by gastric intubation and measured plasma-sulphite concen tration with respea to time following intubation. Peak concentrations ranged from 70 to 800 jim and detectable concentrations persisted for i-3 hr depend ing on the dose. Free sulphite has been reported in the plasma of a child diagnosed as deficient in sulphite oxidase (Shih, Abroms, Johnson. Carney. MandelL Robb, Cloherty St Rajagopalan. 1977). The plasma concentration in creased from 14 jiM to about 130 jim and then de creased to 2 jim when the child's diet was first enriched with cysteine and then restricted in sulphuramino acid content. In sulphite oxidase-deficient rats possessing approximately 1% of the enzyme aaivity of normal adults, plasma-sulphite concentrations ranged from undetectable to 60 jim with a mean of 18 jiM (Gunnison, et at. 1981b). Gunnison & Farruggella (1979) maintained ap proximately steady-state plasma-sulphite concen trations in the range of 400-650 jim for up to 6 hr by zero order intravenous infusion of rabbits at a rate of approximately 09 mmol/kg/hr. The purpose of these infusions was to investigate the kinetics of 5-sulphonate formation in the aorta and lung. However, dur ing the course of performing the experiments, it was also learned that rabbits could not survive for longer than 2 or 3 hr when plasma-sulphite levels were main tained in the range of 700-1000 jim (unpublished data). It is worthy of note that in the majority of the in vitro experiments discussed earlier, sulphite concen trations were either dose to or exceeded the concen trations shown to be lethal to rabbits, and were, in addition, almost always at least one order of magni tude greater than those observed in animals and humans severely deficient in sulphite oxidase. This fact does not diminish the value of the in vitro data since the primary purpose of such experiments is usually to identify, or define more accurately under optimal conditions for their development, potentially toxic changes which may also appear m ore under more prolonged but less severe exposure conditions. Nevertheless, the relationship of sulphite concen trations used in vitro to those attainable in vivo should be considered in prediaions of the likelihood of the occurrence of a particular toxic effect in vivo. Alternative mammalian models for evaluation of sul phite toxicity A case has been made against the use of the rat for the evaluation of sulphite toxicity in humans because of the lower activity of sulphite oxidase in the latter species. In addition, it is known that a genetic defi ciency of this enzyme in humans can lower its activity still further. These cases of severe deficiency result in grave health effects which can lead to death (Irrevcrre. Mudd. Heizer St Laster. 1967). Although occurrences of extreme deficiency are apparently rare, they do illustrate the crucial role of this enzyme in human health and raise questions concerning the pattern of normal variation in sulphite oxidase aaivity within the human population, as well as the significance of possible minor (i.e. subclinical) deficiencies of sulphite oxidase on the metabolism of sulphite and ultimately on the development of chronic toxic effects. It is clear that the sulphite oxidase-competent rat cannot be used to investigate these questions. In the course of their thorough investigations into the nature and functioning of sulphite oxidase. John son, Rajagopalan St Cohen (1974) found that rats deficient in this enzyme could be produced by mani pulation of the tungsten (W) and Mo content of the diet. In an internal environment of relatively high W and low Mo, W either replaces Mo or prevents its incorporation into newly synthesized apoenzyme mol ecules, causing these molecules to be inaaive (John son, Cohen St Rajagopalan. 1974). This results in the progressive loss of sulphite-oxidase aaivity to a lower steady-state level, the magnitude of which is depen dent upon the W:Mo intake ratio. By manipulation of this ratio, steady-state sulphite-oxidase aaivities can be attained ova a wide range (Gunnison et at. 1981b). This phenomenon has been exploited in our laboratory where groups of sulphite oxidase-defident female rats, possessing approximately 1-2% of the ac tivity exhibited by normal female adults, have been charaaerized metabolically and used to investigate sulphite toxicity (Gunnison er at. 1981a.b). These defi cient rats are considered to be models for humans having approximately 10% of the sulphite-oxidase ac tivity ascribed to normal individuals. The model has hcipal to answer some questions regarding the primary toxicity of sulphite, specifically concerning the systemic dcstruaion of thiamine and the develop ment of anaemia, as mentioned previously. More im portantly, using this model a serious concern has been raised regarding the possible involvement of sulphite in the aetiology of early breast cancer. A low inci dence of mammary adenocarcinoma (4/149) was ob served in sulphite oxidase-defident rats of less than 5 months of age which had been treated with a high W/low Mo regime for only 40-65 days. The tissues of these rats were, of course, exposed to elevated concencentrations of endogenously-generated sulphite. No tumours were found in age-matched controls. Although the difference between the sulphite oxidasedeficient and control groups was not statistically sig nificant, the authors believed that because of the early age at which the tumours developed, they were very likely to be treatment related. If this is true, then the role of excess W and a deficiency of Mo (apart from its effea on sulphite-oxidase activity) in the produc tion of early mammary tumours must be considered in addition to that of systemic sulphite. These faaors are germane to the evaluation of any toxic effea ob served using this model. The direa toxicity of excess W and of Mo deficiency can be at least partially de termined by the use of controls, but the possibility of synagism with sulphite must also be considered and is more difficult to evaluate, in spite of these drawbacks, the sulphite oxidase-deficient rat shows 321001 RowVetK 678 A F Gunnison promise as a model for the evaluation of human sul phite toxicity. It is believed by this author that sulphite oxidase- competent mammals are adequate models for evalu ation of the localized effects of inhaled sulphur di oxide on the upper respiratory tract of humans, in terms of accurately reflecting tissue sulphite concen tration. There is recent evidence that the upper air ways of the respiratory tract directly exposed to inhaled S02 (including the nasal passages, trachea and major bronchi) can build up considerable local ized concentrations of sulphite without measurably affecting the overall systemic concentration of sul phite (Gunnison. Zaccardi. Dulak & Chiang. 1981). It is probable that the capacity of the animal to oxidize sulphite enzymatically has little bearing on the con centrations of sulphite that develop in the upper res piratory tract of animals inhaling S02. Although the voluminous literature on the potential toxicology of inhaled 3>Oj is. in general not within the scope of this review, an important chronic study involving S02 will be discussed here because it suggests an active role for sulphite in the origin of bronchogenic tumours. With out citing particular references, it is accurate to state that numerous investigations in a variety of mammals exposed chronically to concentrations of S02 several times greater than those ordinarily observed in urban environments have revealed no irreversible toxic re sponse in the respiratory tract. In contrast, however, lifetime intermittent exposure of rats to benzo[o]pyrene (BP), a known carcinogen, in conjunction with S02 suggests that S02 may be acting as a cocarcino gen (Laskin. Kuschner. Sellakumar & Katz. 1976). The data from this study are summarized in Table 1. Although the significance of the temporal aspect of SOz exposure relative to that of BP is difficult to assess, consideration of the data in toto strongly sug gests a role for S02 in the aetiology of these broncho genic squamous-cell carcinomas. Discussion Although in citro research has been extremely use ful in increasing our knowledge of sulphite chemistry and of the potentially toxic reactions of sulphite, in ntro systems have, in general been poor models for predicting mammalian toxicity, largely because in nro mammalian sulphite concentrations have been greatly overestimated. Extrapolations of in ritro data to the m tiro situation have often been erroneous (Inouye et at. 1978; Kaplan et al. 1975: Schneider & Calkins. 19701 primarily because of ignorance of sulphite metabolism. Now. however, quantitative data relating exogenous sulphite exposure to in vivo sulphite con centrations in mammals are available, and provide a basis for evaluation of the in viva implications of in vitro data. Thus, whereas plasma-sulphite concen trations of up to approximately 100 were observed in a rhesus monkey ingesting 2 mmol sulphite/kg/day. most of the sulphite toxicity demonstrated in vitro resulted from concentrations that were considerably higher. Further, because of its rapid metabolic clearance by sulphite oxidase, chronically ingested sulphite does not accumulate in the tissues and reach an elevated steady-state concentration but is rapidly eliminated after absorption, giving sporadic brief episodes of ele vated tissue-sulphite levels; these will not sustain sul phite adducts which are unstable in the absence of free sulphite. Therefore, the toxicological significance of the reversible reactions of sulphite described in the earlier sections is probably slight or nil under con ditions of intermittent exposure of body tissues to sul phite--the expected pattern of human exposure Compared to reversible reactions, the irreversible reactions of sulphite are of potentially greater toxico logical significance. The sulphitolysis of thiamine, for example, although a relatively slow reaction, results in significant destruction of the vitamin in stored diets as well as in the gut. The sulphite-catalysed deamination of cytosine and S-methylcytosine residues in DNA is an irreversible reaction of possible toxicological significance. Shapiro (1977) has calculated, on the basis of the in vitro rate of conversion of cytosine to uracil under physiological conditions, that a sulphite concentration of only 03 im is required to double the spontaneous muta tion rate in humans. The toxicological implications of this calculation are now unclear, however, in the light of the apparent lack of involvement of cytosine resi dues in C:G to A:T transitions. As pointed out earlier, this is an area of current investigation. Another group of reactions with toxic potential are the reactions of free radicals, produced by the autoxidation of sulphite, with tryptophan, methionine (and certain other sulphides!. DNA and olefins. Again, reactions with the first three substrates are apparently Table I. Squamous^ell carcinomas in ran chronically exposed to henzo{i]piTene and SO;. jio>\ anti in Exposure conditions (on 5 days wk) No. of rats Per group With carcinomas Incidence IM Filtered air 15 0 0 10 ppm SOj (6 hr) 15 0 0 BPt(lhr) 50 1 3 10 ppm S02 (6 hr) followed by BPt (l hr) 30 2 7 BPt with 4 ppm S02 (1 hr) 45 4 9 10 ppm SOj (6 hri followed by BP* with 4 ppm SO. 11 hr) 46 9 20 *D*ii from Laskia ft al 1976. 921001 RowVerK lomr > (Inouye et Sl Calkins, f sulphite a relating u.phite con 'd provide a ''tons of in concenobserved 'hite/kg/day. ted in vitro isiderably clearance by ilphite does elevated liminated iooes of elesustam sul'tsence of nificance -t.-ed in the under con s' "ts to sulx ire. eversible rater toxicohiamme. for )' results in j I diets as cytosine and eversible n Shapiro in vitro rate ihysiological o of only ie is mutap..nations of . in the light yfine resi st ted out ti. .. xnential are th autoxik ne (and hi Again. apparently 3 ' .,, ^ --c~. " .? ' Critical review of sulphite toxicity irreversible, while information on the reversibility of the latter is not available. The implications of these reactions include damage of membranes due to freeradical attack on their unsaturated lipids, and chromosome aberrations resulting from sulphiteinduced DNA chain breaks. Although free-radical reactions proceed at lower concentrations than many of the ionic reactions of sulphite, it is difficult to speculate on their occurrence in vivo because of the presence of numerous free-radical scavengers which prevent significant sulphite autoxidation, even at favourable sulphite concentrations, and because superoxide dismutase may inhibit the initiation of autoxidation by Of. Most of our knowledge of mammalian sulphite tox icity originates from experiments in which sulphite oxidase-competent rats were fed large quantities of sulphite. From these data most toxicologists would legitimately conclude that the hazard to humans of sulphite consumption at present levels is very low. However, the role of sulphite-oxidase activity in sul phite toxicity has not as yet been adequately investi gated. We have pointed out previously that, with re spect to sulphite-oxidase capacity, the rat is a poor model for humans. There are, in addition, essentially no data on the variability of sulphite-oxidase activity in the human population. Further, little is known of the quantitative relationships between intake of sul phur-containing amino adds, sulphite-oxidase activity and endogenous sulphite concentration. Given the relatively low mean sulphite-oxidase ca pacity of humans, the possibility of significant genetic variation in sulphite-oxidase activity among individ uals, the uncharacterized variable of dietary sulphuramino add content and the generally low human con sumption of sulphites, it appears that the importance of endogenous sulphite generation relative to exogen ous sulphite intake may have been underestimated in most previous investigations of sulphite toxidty. Further, endogenous sulphite is produced intracellularly in relatively dose proximity to DNA and other potential target molecules, while ingested sulphite must traverse several barriers before reaching these sites. In human sulphite oxidase-deficiency disease as well as in sulphite oxidase-deficient rats, steady-state sulphite concentrations in the micromolar range have resulted from purely endogenous sources. Under these conditions, reversible reactions of sulphite can be of physiological significance provided the sulphite adduct is reasonably stable. As an example, it was previously sated that sulphite reacts readily, although reversibly, with NAD when associated with LDH. forming an enzymatically inactive adduct with a pHindependent dissociation constant of approximately 10" * vt (Parker er al. 19781. Using this K.. one can calculate that in the presence of 50 pM SOj ". more than 99% of the NAD bound to LDH would be in the form of the sulphite adduct and, therefore, presum ably inactive. There is some discrepancy, however, between this calculated value and the experimenally determined inhibition of LDH in vitro. Ciaccio (1966) and Oshino St Chance (1975) demonstrated 50% inhi bition of the enzyme by 350 and 200 /a* sulphite, re spectively. Similar data for other enzyme-cofactorsulphite adducts and other types of sulphite addition products suggest that metabolic disturbances may result from steady-state sulphite concentrations in the micromolar range, as can be expected to occur in animals sufficiently deficient in sulphite oxidase. A series of enzymes that are inhibited by micro molar concentrations of sulphite are the sulphatases, most notably sulphatase A. Cerebroside sulphatase, which is thought to be identical to sulphatase A. is deficient in cases of metachromatic leucodystrophy (MLD), a genetically determined disease in which myelin degeneration is associated with accumulation of cerebroside sulphate (Moser, 1972). There is also a series of genetically determined diseases (mucopoly saccharidoses) many of which are caused by. or as sociated with, functional deficiencies of one or more other sulphatase enzymes (Roy, 1976). Nearly all of these diseases are characterized by mental retardation as well as by physical defects, especially those of bone. The possibility of inhibition of sulphatase enzymes in vivo by sulphite was recognized by the investigators who first identified sulphite oxidase-deficiency disease. Analysis of the urine and tissues from a patient who died of this disease revealed no significant increase in tissue or urinary sulphate esters as would be expected if sulphatases had been inhibited (Percy, Mudd, Irreverre & Laster, 1968). However, these results are diffi cult to interpret since the severely decreased avail ability of sulphate in this patient would have resulted in a reduced rate of sulphate ester synthesis. Several experiments reviewed in this paper have suggested, or are consistent with, a role for sulphite as a cocarcinogen. Certainly the development of bron chogenic squamous-cell carcinomas following inhala tion exposures to BP and SOj (Laskin et al. 1976) and the in vitro demonstration of sulphite enhance ment of UV mutagenicity (Malkm St Rossman. 1981) support this hypothesis. In addition, the appearance of mammary adenocarcinomas in young sulphite oxi dase-deficient rats might result from an interaction of sulphite and tungsten. Although this is admittedly speculative, the cocarcinogenesis hypothesis should continue to be investigated. The in vitro investigations of sulphite reactivity and toxicity reviewed in this paper have suggested an array of mechanisms that could produce toxicity in mammals. Consideration of this information and the data from in vivo evaluations of sulphite toxicity in the light of mammalian sulphite metabolism, makes possible the elimination of some of these mechanisms as unlikely and the selection of others as being of greater potential and as deserving further investiga tion. AcknowMgtmenti--This investigation was supported by Grant No. ES 00613 of the National Institute of Environ mental Health Sciences and is part of a centre programme supported by the National Institute of Environmental Health Sciences. Grant No. ES 00260. I thank Dr Robert Shapiro for his advice in the preparation of this manuscript and Jane T. 'Gunnison and Dr E. D. Palmes for their editorial suggestions. tCFUCNCES Ali. 1. U. A Hynes. R. O. (1978k Role of disulfide bonds in the attachment and function oflarge, external, transforma tion-sensitive glycoprotein at the cell surface. Biot him. biophyx. Acta 510. 140. i ... . - r <\ iUUl 680 A. F. Gunnison Bhagai. B 4 Lockett. M. F. (I960) The absorption and elimination of metabisulphite and thiosulphate by rats. J. Pharm. Pharmac 12. 690. Bhagat. B. 4 Lockett. M. F. (1964). The effect of sulphite in solid diets on the growth of rats. Fd Cosmet. Toxicol. 2. I. Bigwood. E.J. (1973). The acceptable daily intake of food additives. CRC Crii. Rer. Toxicol 2. 41. Bjursell G_ Gussander. E. Sl Lindahl T. (I979L Long regions of single-stranded DNA m human cells. Nature. Land. 280. 420. Bravcrman. B.. Shapiro. R. 8c Szer. W. (1975). Modification of E. coli ribosomes and coliphage MS2 RNA by bisul fite: effects on rtbosomal binding and protein synthesis. Nucleic Acids Rev. 2. 501. Burroughs. L F. Sc Sparks. A. H. (1973). Sulphite-binding power of wines and ciders. I. Equilibrium constants for the dissociation of carbonyl bisulphite compounds. J. Su. FJ Auric. 24, 187. Cecil. R. (1963). Intramolecular bonds in proteins. I. The role of sulfur in proteins. In The Proteins Vol. I. Edited by H. Neurath. 2nd Ed. p. 379. Academic Press. New York. Chin. S- BisseiL M. 3. & Bassham. J. A. (1977). The conse quences of bisulfite exposure in primary chick embryo fibroblast in culture. Bull. mrir. Contain. Toxicol. IS. 749. Ciaccto. E. I. (1966). The inhibition of lactate dehydro genase by 3-acetylpyridine adenine dinucleotide and bisulfite. J. hint. Chem. 241. 1581. Cluzan. R_ Causcret. J. & Hugot. O. (19651. Le metabisul fite de potassium. Etude de toxicite a long terme sur le rat. Aunts Biol. Anim. Biochim. Biopins. 5. 267. Cohen. H. J.. Drew. R. T. Johnson. J. L. Sc Rajagopalan. K. V. (I973L Molecular basis of the biological function of molybdenum. The relationship between sulfite oxidase and the acute toxicity of bisulfite and SO). Proc. natn. Acad. Sci. U S A. 70. 3655. Cohen. H. J. 8c Fridovich. 1.(19711 Hepatic sulfite oxidase. Purification and properties. J. hitd. Chem. 246. 359. Coopersiein. S. J. (1963). Reversible inactivation of cyto chrome oxidase by disulfide bond reagents. J. hiol. Chem. 238. 3606. Coulondre. C- Miller. J. H.. Farabaugh. P J. 4 Gilbert. W. (1978). Molecular basis of base substitution hotspots in Escherichia coli. Nature. Land. 274. 775. Cremcr. H. D. 4 Hotzel D. (19663 Thiamin-Sulfit-Antagonismus in vivo--cine Anlithiaminwtrkung. Bihlthta Nutr. Dteia' 8. 126. cited Biol. Ahstr. 1968. 79. 591. Cremer. H.-D. 4 Hotzel D. (19701 Thiaminmangel und Unbedenklichkeil von Sulfit fur den Menschcn. 4. Mitteilung. Schlussfolgerungen und Zusammenfassung der Ergebnisse eines Stoffwechselversuchs am Menschen. hit. Z. VitamForseh. 40. 52: cited from Chem. Ahstr. 1970. 73. 107. Das. G. 4 Runccklcs. V C. (1974). Effects of bisulfite on metabolic devetopmeni in synchronous Chlorella p\renotdosa. Ennr. Res. 7, 353. Dasgupta. P. K_ DeCesare. K. Sc Ullrey. J. C. (1980). De termination of atmospheric sulfur dioxide without tetrachloromcrcurate (II) and the mechanism of the Schiff reaction Amtlyl Chem. 52. 1912. Duncan. B K. 4 Miller. J. H. (19801. Mutagenic deamina tion of evtosme residues in DNA. Nature. Land. 287. 560 Fitxhueh. O G. Knudsen. L F 4 Nelson. A. A. (1946). The chronic toxtcilv of sulfites. J. Pharmac. exp. Ther. 86. 37. Fridovich. I. Sl Handler. P. (19611. Detection of free radi cals generated during enzymic oxidations by the in itiation of sulfite oxidation. J. hiol. Chem. 236. 1836. Generoso. W. M_ Huff. S. W 4 Cain. K T. (1978). Tests on induction of chromosome aberrations in mouse germ cells with sodium bisulfite. Mutation Res. 56. 363. Gibson. W B. 4 Strong. F. M. (1973). Metabolism and elimination of sulphite by rats, mice and monkeys. Fd Cosmet. Toxicol II. 185. Gregory. R. E. (1981). Isolation and Identification of Plasma Protein S-Sulfonates Following in citro and m nro Exposure to Sulfite. Ph.D. Thesis. New York Uni versity. Gunnison. A. F. 4 Benton. A. W. (1971). Sulfur dioxide: sulfite. Interaction with mammalian serum and plasma. Archs emir. Hlth 22. 381. Gunnison. A. F,, Bresnahan. C. A. 4 Palmes. E. D. (1977). Comparative sulfite metabolism in the rat. rabbit and rhesus monkey. Toxic. ap,rl. Pharmac. 42. 99. Gunnison. A. F_ Dulak. L, Chiang. G_ Zaccardi. J. 4 Farruggella. T. J. (1981a). A sulphite oxidase-deficient rat model: subchronic toxicology. Fd Cosmet. Toxicol 19, 221. Gunnison. A. F. 4 Farruggella. T. J. (1979). Preferential S-sulfonate formation in lung and aorta. Chemtco-Biol Interactions 2S. 271. Gunnison. A. F. Farruggella. T. J_ Chiang. G.. Dulak. L_ Zaccardi. J. 4 Birkner. J. (1981b). A sulphite oxidasedeficient rat model: metabolic characterization. Fd Cos met. Toxicol. 19, 209. Gunnison. A. F. 4 Palmes. E. D. (1973k. Persistence of plasma 5-sulfonates following exposure of rabbits to sul fite and sulfur dioxide. Toxic, appl. Pharmac. 24, 266. Gunnison. A. F. 4 Palmes. E. D. (19781 Species variability in plasma S-sulfonate levels during and following sulfite administration. Chemico-Biol. Interactions 21. 315. Gunnison. A. F_ Zaccardi J. Dulak. L 4 Chiang. G. (19811. Tissue distribution of S-sulfonate metabolites fol lowing exposure to sulfur dioxide. Encir. Res. 24. 432. Harkness. D. R. 4 Roth. S. (19691 Purification and proper ties of 2.3-diphosphoglyceric acid phosphatase from human erythrocytes. Biochem. hiophvs. Res. Commun. 34. 849. Hayatsu. H. (1976L Bisulfite modification of nucleic acids and their constituents. Prog. Noel. Acid. Res. mol. Bml. 16. 75. Hayatsu. H. 4 Miller. R. C. Jr (1972). The cleavage of DNA by the oxygen-dependent reaction of bisulfite. Bio chem. hiophys. Res. Commun. 46. 120: Hayatsu. H. 4 Miura. A. (1970). The mutagenic action of sodium bisulfite. Biochem. htophrs. Res. Commun. 39. 156. Hayatsu. H. Wataya. Y. Kal K. 4 lida. S. (I970L Reac tion of sodium bisulfite with uracil Sytosine. and their derivatives. Biochemistry. N. >'. 9. 2858. Hayon. E. Treinin. A. 4 Will J. (19723 Electronic spectra, photochemistry, and autoxidation mechanism of the sulfite-btsulfite-pyrosulfite systems. The SO;. SO;. SO; and SO; radicals. J. Am. chem. Soc. 94. 47. Inoue. M . Hayatsu. H. 4 Tanooka. H. (1972). Concen tration effect of bisulfite on the inactivation of trans forming activity of DNA. Chemtco-Biol. Interactions 5. 85. Inouye. B. Ikcda. M,, Ishida. T. Ogata. M_ Akiyama. J. 4 Utsumi. K. 11978). Participation of superoxide free radi cal and MnJ * in sulfite oxidation. Toxic, appl. Pharmac 46. 29 Institute of Food Technologists and Committee on Public Information (19761 Sulfites as food additives. .Vutr Ret 34. 58. Irreverrc. F. Mudd. S. H. Hetzer. W. D. 4 Lastcr. L. 119673 Sulfite oxidase deficiency: studies of a patient with mental retardation, dislocated ocular lenses, and abnormal urinary excretion of S-sulfo-u-cysteine, sulfite and thiosulfate. Biochem. Med. I. 187. Jagteilo. G. M. Lin. J. S. 4 Ducayen. M. B. (1975). SO; and its metabolites: effects on mammalian egg chromo somes. Encir. Res. 9. 84. Johnson. J. L_ Cohen. H. J. 4 Rajagopalan. K. V. (19741 Molecular basts of the biological function of molybde- 921001 RowVerK 100 ' 'onkeys. Fd '\ * ' uy, 'fication of " "'O and in York Uni- __ `ur dioxide: "4 plasma. IT T * ---------- D (1977). rabbit and ;ardi. J. Sl deficient rat Toxicol. 19, T .. f 1 * 8 I Preferential 'emico-Bid. Dulak. L. >e oxidase'n. Fd Cos- sistence of 'bits to sul24, 266. variability *mg sulfite JU.hiang. G. bolites fol24.431 nd properitase from immun. 34, ' -7* _ -4 fx-! -'<S i -A cleic acids nil. Biol. leavage of ulfite. Bio- action of ilium. 39. 701 Reacand their spectra. ' the sul- T so; Concenof irans'urrious 5. jmj, J. free radiPharmac. >n Public >uir Rrr .aster. L. 1 patient ises. ami r. sulfite '51 SO, chromo- f 119741 molybde Critical review of sulphite toxicity 681 num. Molybdenum-free sulfite osidase from livers of tungsten-treated rats. J. biol. Chem. 249. 5046. Johnson. J. Lt Rajagopalan. K. V. (1976a). Purification and properties of sulfite osidase from human liver. J. din. /meu. 58, $43. Johnson. J. L. dt Rajagopalan. K. V. (1976b). Human sul fite oxidase deficiency. Characterization of the molecular defect in a multicomponent system. J. din. Incest. 51, 551. Johnson. J. L, Rajagopalan. K. V. it Cohen. H. J. (1974). Molecular basis of the biological function of molybde num. Effect of tungsten on xanthine oxidase and sulfite oxidase in the rat. J. biol. Chem. 249. 859. Joint FAO/WHO Expert Committee on Food Additives (1974). Toxicological Evaluation of Some Food Addi tives Including Anti-caking Agents. Antimicrobials. Antioxidants, Emulsifiers and Thickening Agents. WHO Fd Add. Ser. No. 5, p. 130. Kai. K. Tsuruo. T. & Hayatsu, H. (1974). The effect of bisulfite modification on .the template activity of ONA for'DNA polymerase I. iVucleic Acids Res. I. 889. Kamogawa. A. & Fukui. T. (1973k Inhibition of a-gtucan phosphorylase by bisulfite competition at the phosphate binding site. Biochrn. biophys. Acta 302, 158. Kaplan. D, McJilton. C. A LuchteL D. (1975k Bisulfite induced lipid oxidation. Archs encir. Hlth 30, 507. Kikugawa. K. A lizuka. K. (1972k Inhibition of platelet aggregation by bisulfite-sulfite. J. pharm Sd. 61. 1904. Kitamura. N. A Hayatsu. H. (1974k Cleavage of the glycosidic linkage of pyrimidine nbonucleosides by the bisul fite-oxygen system. Nucleic Adds Res. I. 75. Kudo. 1, Hayatsu. H. Yokoiyama, A. A Kuroda. Y. (1980k Effect of bisulfite on cell-to-substratum adhesion of Chinese hamster cells in culture. J. Pharmacobto-dyn. 3, 74. Kudo. I, Miura. A. A Hayatsu. H. (1978). Inactivation of bacteriophage /. during the aerobic oxidation of bisulfite. Encir. Res. 16, 205. Lanteaume. M.-TK. Morin. P,, Palluel. G. <t Pallaget. C. (1978k Recherchcs de certains effets metaboliques ou toxiques pouvant resuller a moyen lerme de I'ingestion d'actde sulfureux par le rat. Annls Nutr. Aliment. 32. 129. Lanteaume. M.-T_ Ramel. P_ Girard. P,, Jaulmes, P. Gasq. M. A Ranau. J. (1965). Effets physiologiques a long lerme de ('anhydride sulfureux ou des sulfites uti lises pour le traitemenl des vins rouges. (Observations portant sur quatre-generations de rat). Annls Falsif. Expert, chm. 58, 16. Laskin. S_ Kuschner. M_ Seltakumar, A. A Katz. G. V. (1976k Combined carcinogen-irritant animal inhalation studies. In Air Pollution and the Lung. Edited by E. F Aharooson. A. Bcn-David A M. A. Klmgbcrg. p. 190. John Wiley A Sons. New York. Lcicfclcr, J. (1969k The Cleavage of Thiamine by Residual Sulfite in Casein. Ph.O. Thesis. University of California. Berkeley. Lhutsster. M. (1966). Influence des sulfites sur I'efficacite biologique de la vitamtne B, Bull. Acad. nain. Med. 150. 242. LindahL T. (1974k An N-glycosidase from Escherichia coh that releases free uracil from ONA containing deaminated evtostne residues. Proc. natn. Acad. Sci. US A. 71.3649. Lockett. M. F. A Natoff. I. L (1960k A study of the tox icity of sulphite. I. J. pharm. Pharmac. 12. 488 McArdle. C. (1967k The reaction of sulphite ions with an mtcrmolecular disulphide bond in bovine serum nonmercaptalbumin. Bioehem. J. 102. 211. McCord. J. M. A Fridovich. I. (1969k The utility of superoxide dismutase in studying free radical reactions. I. Radicals generated by the interaction of sulfite, dimethyl sulfoxide, and oxygen. J. hiol. Chem. 244, 6056. MacRac. W. O. A Such. H. F. (1979k Induction of sister chromatid exchanges in Chinese hamster cells by the reducing agents bisulfite and ascorbic acid. Toxicology 13. 167. Mallon. R. G. A Rossman. T. B. (19811. Bisulfite (sulfur dioxide) is a comutagen in . cofi and in Chinese ham ster cells. Mutation Res. IS, 125. Massey. V. Muller. F_ Feldberg. R. Schuman. M. Sulli van. P,, Howell L. G. Mayhew. S. G. Matthews. R. G. A Foust. G. P. (1969k The reactivity of (Uvoprotems with sulfite. Possible relevance to the problem of oxygen reactivity. J. biol. Chem. 244, 3999. Moser. H. W. (1972). Sulfattde lipidosis: metachromatic leukodystrophy. In The Metabolic Basis of Inherited Dis ease. Edited by J. B. Stanbury. J. B. Wyngaarden Sc O. S Fredrickson. 3rd Ed. p. 688. McGraw-Hill Book Co.. New York. Mukai, F,, Hawryluk. I. A Shapiro. R. (1970). The muta genic specificity of sodium bisulfite. Bioehem. hiophys. Res. Conunun. 39. 983. Miilier. F. A Massey, V. (1969k Flavin-sulfite complexes and their structures. J. bid. Chem. 244, 4007. Nir. I. Kafti I. A Cohen. R. (1978k Effect of pelleting on the vitamin K activity of various vitamin K, prep arations in chicks. Poultry Sd. 57. 206. Oshino. N. A Chance. B. (1975k The properties of sulfite oxidation in perfused rat liver: interaction of sulfite oxi dase with the mitochondrial respiratory chain. Archs Bioehem. Biophys. 170, 514. Parker. D. M. Lodoia. A. A Holbrook. J. J. (1978). Use of the sulphite adduct of nicotinamide-adenine dinudcotide to study ionizations and the kinetics of lactate dehydro genase and malate dehydrogenase. Bioehem. J. 173. 959. Parker. J. C. (1969k Influence of Z3-diphosphoglycerate metabolism on sodium-potassium permeability in human red blood ceils: studies with bisulfite and other redox agents. J. din. Incest. 41. 117. Percy, A. K, Mudd, S. H,, Irreverre. F. A Laster. L. (1968). Sulfite oxidase deficiency: sulfate esters in tissues and urine. Bioehem. Med. 2, 198. Perry. P. A Evans. H. J. (1975k Cytological detection of mutagen-carcinogen exposure by sister chromatid exchange. Nature. Land. 258. 121. Petering. D. H. A Shih. N. T. (1975k Biochemistry of bisulfite-sulphur dioxide. Encir. Res. 9. 55. Pitman. (. H. A Jain. N. B. (1979). Covalent addition of bisulfite ion to N-alkylated uracils and thiouracils. Aust. J. Chem. 32, 545. Rork. G. S. A Pitman, 1. H. (1974). Elimination of bisulfite ion from a series of uracil-bisulfite adducts. Evidence for a two-step mechanism. J. Am. chem Sac. 96. 5464 Roy. A. B. (1960k The synthesis and hydrolysis of sulfate esters. In Adcances in Enzymohgy. VoL 22. Edited by F. F. Nord. p. 205. Interscience Publishers. Inc. New York. Roy. A. B. 11976k Sulphatases, lysosomes and disease. -tux. J. exp. Bid. Med. Set 54, 111. Schneider. L. K. 4c Calkins, C. A. (1970k Sulfur dioxide- mduced lymphocyte delects in human peripheral blood cultures Encir. Res. 3, 473. Schrocter, L. C. (1966k Sulfur Dioxide. Applicants in Foods. Becerages. and Pharmaceuticals. 1st Ed. Perga- mon Press Oxford. Shapiro. R. (1977k Genetic effects of bisulfite (sulfur di- oxidek Mutation Res. 39,149. Shapiro. R. A Bravcrman. B. (1972k Modification of poly- uridybc add by bisulfite: effect on double helix forma tion and coding properties. Bioehem. htophw. Res Com mon. 47, 544. Shapiro. R_ Bravcrman. B. Louis J. B. A Sems. R E. (1973k Nucleic aod reactivity and conformation II Reaction of cytosine and uracil with sodium bisulfite. J bud. Chem 24*. 4060. Shapiro. R, Di Fate. V. A Wekher. M. 11974k Deamma- - 921001 RowVarK 100) 682 A. F. Gunnison lion of cytosine derivatives by bisulfite. Mechanism of the reaction. J. Am. chem. Soc 96, 906. Shapiro. R. A Gaxit, A. (1977k. Crosslinking of nucleic acids and proteins by bisulfite. In Advances in Experi mental Medicine and Biology. Vol. HAA. Protein Crosslinking. Biochemical and Molecular Aspecti. Edited by M. Friedman, p. 633. Plenum Press, New York. Shapiro. R_ Wefcher. M,, Nelson. V. A Di Fate. V. (19761 Reaction of uracil and ihymine derivatives with sodium bisulfite. Studies on the mechanism and reduction of the adduct. Bioclum. btophys. Acta 425, 113. Shih. N. T. A Petering. D. H. (1973). Model reactions for the study of the interaction of sulfur dioxide with mam malian organisms. Biochem. htophvs. Res. Common. 55, 1319. Shih. V. E_ Abroms. 1. F,, Johnson. J. L, Carney. M_ MandelL R. Robb. R. Cloherty. J. P. A Rajagopalan. K. V. (1977). Sulfite oxidase deficiency. Biochemical and clinical investigations of hereditary metabolic disorder in sulfur metabolism. New-Engl. i. Med. 297, 1022. Sklyadneva. V. B,, Chekanovskaya. L A. Nikolaeva. I. A. Sc Tikchonenko. T. I. (1979a). The secondary structure of bacteriophage DNA in situ. VIII. The reaction of sodium bisulphite with intraphage cytosine as a prebe for studying the DNA-protein interaction. Biochem. bio- phys. Acta 565. 31. Sklyadneva. V. B_ Shie. M. A Tikchonenko. T. I. (1979b). Alteration of the DNA secondary structure in. the DNApolylysine complex evidenced by sodium bisulfite modi fication. FEBS Lett. 107, 129. SUe. S. Si Shapiro. R. (1978). Deamination of cytidine by bisulphite: mechanism at neutral pH. J. org. Chem. 43, 4197. Stacey. F. W. A Harris. J. F. Jr (1963). Formation of carbon-heiero atom bonds by free radical chain additions to carbon-carbon multiple bonds. In Organic Reactions. Vol. 13. Edited by A. C. Cope. R. Adams. A. H. Blau. V. Boekelheide. T. L Cairns. D. Y. Curtin A C. Niemann, p. 196. John Wiley A Sons. Inc. New York. Summers. G. A. A Drake. J. W. (1971). Bisulfite mutagen esis in bacteriophage T4. Genetics. N. Y. 68. 603. Swoboda. B. E. P. A Massey. V. (1966). On the reaction of the glucose oxidase from Asperigillus niger with bisulfite. J. biol. Chem. 241. 3409. Thompson. J. R. A Pace. D. M. (1962). The effects of sul phur dioxide upon established cell lines cultivated in rifro. Can. J. Biochem. Physiol. 40, 207. TiL H. P. Feron. V. J. A de Groot. A. P. (1972a) The toxicity of sulphite. L Long-term feeding and multigeneration studies in rats. Fd Cosmet. Toxicol. !, 291. TiL H. P. Feron. V. J. de Groot. A. P. A van der WaL P. (1972b) The toxicity of sulphite. II. Short- and long term feeding studies in pigs. Fd Cosmet. Toxicol. 10, 463. Timson. J. (1973) Action of sodium sulphite on the mitosis of human lymphocytes. Chromosomes Today 4, 211. Turchinsky. M. F_ Kusova. K. S. A Budowsky. E. I. (1974) Conversion of non-covalent interactions in nudeopro- teins into covalent bonds: bisulfide-induced formation of polynucleotide-protein crosslinks in MS2 bacteriophage virtons. FEBS Lett. 38, 304. .Vas. K. (1949) The equilibrium between glucose and sul phurous add. J. Soc. chem. Ind. 68 340. Wang. R. Y. H. A Ehrlich. M. (1980) Comparison of bisul fite modification of 3-methyideoxycytidine and deoxycytidine residues. (Abstr.) Fedn Proc. Fedn Am. Socs exp. Biol. 39. 1882. Wang. R. Y. H, Gehrke. C. W. A Ehrlich. M. (1980) Com parison of bisulfite modification of 5-methyldeoxycytidine and deoxycytidinc residues. Nucleic Acids Res. 8, 4777. Wilkins. J. W. Jr. Greene. J. A- Jr A Weller. J. M. (1968) Toxidty of intraperitoneal bisulfite. Clin. Pharmoc. Ther. 9.328. Williams. R. R. Waterman. R. E_ Kcresztesy. J. C. A Buchman. E R. (1935) Studies on crystalline vitamin B,. 111. Cleavage of vitamin with sulfite. J. Am. chem. Soc. SI, 536. Yang. S. F. (1970) Sulfoxide formation from methionine or its sulfide analogs during aerobic oxidation of sulfite. Biochemistry. N.Y. 9. 5008. Yang. S. F. (1973) Destruction of tryptophan during the aerobic oxidation of sulfite ions. Enrir. Res. 6. 395. 921001 RowVerK 1(U)7