Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
THE EFFECT OF DRUGS ON PHYSIOLOGICALLY ACTIVE THIOL SYSTEMS Maxwell Schubert The Study Group For Rheumatic Diseases The Department of Chemistry New York University - Bellevue Medical Center New York City
270 There are at the present time many compounds of widely different chemical structures that are being introduced into living animals to produce rather specific biological effects, thera- peutic or otherwise, and that on more or less solid grounds are believed to produce their effects by inhibiting or in some other way regulating the activity of thiol systems. It is the purpose of this survey to bring together in one discussion work that has been going on in several fields as widely separated as mercurial diuretics and fungal antibiotics. A presentation will first be made of the kinds of chemical reaction known to occur between the type compounds used and simple thiols. This will be followed by an examination of the basis in each field for ascribing the drug action to thiols and, as far as possible, an assessment of the present validity of this basis. It seems particularly appropriate to attempt this in a symposium on Chemical-Biological Correla- tion since there are present so many with detailed experience in particular special fields that we need to touch and so there is every opportunity for a lively discussion on this subject which has many controversial aspects. Since it is easier to classify the drugs used than the effects aimed at, or we might say the biological targets, we will arrange the subsequent discussion along the lines of the chemical types of the compounds. In Table 1 are listed the type compounds to which discussion will be limited together with the specific targets or the general effects they are aimed at. There is little at the start that can be said about any common features of the targets, except that they are all living, metabolizing cells or systems of cells. But chemically there is a common feature to all of the drugs. They all could react with simple thiol compounds. This fact, that a large number of drugs of varied chemical structure could react with thiol compounds has led to a tendency to group these compounds together on the basis of this common property. They are sometimes called thiol reagents. Bacq5 has gone so far as to call them thioloprives. While there may be something to say for the introduction of such a new term, the word will not be used in this review because it is felt that it could have a tendency to narrow the possibilities of the mechanisms of drug action considered. To cite a single specific instance, in his analysis of the action of vesicants, Bacq has not left room for the important possibilities suggested by the subsequent work of PirielOl in which is presented evidence for the action of mustard gas on collagen, a protein containing no thiols at all. TABLE 1 Drug Target of Effect Hg++ ; S406r As +++.; As++ +++ Arscnyl halides as lewisite Alkyl halides as; iodoacetate, sulfur mustards, nitrogen mustards Unsaturated carbonyls Disulfide; sulfoxide-sulfide Qui nones Nap hthoquinone s Alloxan Proximal tubule of kidney Trypanosomes, spirochetes Schistosomes, Leishmanias, Filarias Pyruvate oxidase, vesication Vesication, lacrimation, Lundsgaard effect Antibiotic Antibiotic Antibiotic, spermicide Antiplasmodic, antibiotics, clot-mechan ism p-Cells of the pancreas
271 Before beginning the chemical-biological correlation it will be best to survey carefully the chemical base. This seems particularly desirable since there are several kinds of reaction known to occur easily with simple thiol-disulfide systems, such as cysteine-cystine, and which could conceivably also occur in more complex biological systems but which have never been described or considered in such material. The compounds listed in Table 1 can be separated into two groups depending on their action with thiols ; some act as oxidants, some act to form a covalent bond with an electron pair of the sulfur atom of the thiol. The oxidant action is most simply represented as a complete removal of an electron pair from a thiol molecule or ion as: Â© 0 Â© R-S-H^Â±H + R-S-i=Â±R-S +2â¬ (I) This is immediately followed by condensation of the positive ion formed with a remaining negative mercaptide ion because of the tendency of the latter with its electron-rich sulfur nucleus to share an electron pair with an electron-poor atomic nucleus: o Â© R-S- + R-Sâ- R-S-S-R (H) In this way a disulfide is formed as a secondary product of the primary oxidation reaction. A mechanism such as this might explain the great difficulty which has always been encountered in trying to measure oxidation-reduction potentials of thiol-disulfide systems. It has long been known that the disulfide exerts no effect on these potentials. If the reaction of equation (II) is only very slightly reversible then the disulfide might exert little effect on reaction (I). It will also be shown later that when disulfides do dissociate they seem to do so by a mechanism different from the reverse of (it) and representable by the first of equations (XII). The oxidants of Table 1 that can bring about the reaction (I) are pentavalent arsenic and pentavalent antimony. Chloropicrin has also been reported sometimes to oxidize thiols smoothly and completely to disulfides though the reduction product simultaneously formed from the chloropicrin is not known. 33,67 Hellstrom^3 has made some interesting observations on oxidations brought about by certain halogenated acids according to one or the other of the following: 2 HOOC-CH2-SH + X-CH2-COOHâÂ»(HOOC-CH2-S)2 + CH3-COOH + HX (III) 2 HOOC-CH2-SH + 2 X-CH2-COOHâÂ»(HOOC-CH2S)2 + 2 CH3-COOH + X2 These reactions were studied mainly in acid solutions and there is no data on how far into the neutral range they may persist. This particular type of oxidation has not been reported for any biochemical system, nor has it even been found to occur in the physiological pH range. Among the oxidants that can convert thiols to disulfides there is one group of special interest and that is compounds that themselves have the disulfide structure. These include not only the organic disulfides but also the inorganic salt, sodium tetrathionate, which has a disulfide structure. The special interest attaching to organic disulfides as oxidants is that they are not certainly known to oxidize any organic compounds other than thiols and therefore their reduction in an organic system can be considered a test for the presence of thiols. This oxida- tion presumably occurs in two separate and successive steps as follows: R-S-S-R + R'-S-H^R-S-H + R-S-S-R1 + R'-S-H?Â±R-S-H + R'-S-S-R1 (IV) Mixed disulfides, the product of the intermediate step, are more commonly known among aromatic disulfides and even these are made by reactions other than reaction (IV). There is another aspect of this disulfide-thiol interaction of interest and that is the apparently parallel reaction between thiols and the sulfoxide-sulfides, or as they are sometimes called, alkyl thio- sulfinates. 128
272 R-S-SO-R + R'-S-HâÂ»R-S-S-R' + R-SOH (V) Finally there is an interesting analog of reaction (IV). This is the reduction of disulfides by bisulfite according to the reaction investigated by Clarke;" R-S-S-R + Na-SO2-ONaâ>R-S-S O2-ONa + R-S-Na (VI) But this reaction differs from that of equation (IV) in that even with excess sulfite the analog of the second half of reaction (IV) does not occur. All the other compounds listed in Table 1 undergo the second type of reaction with thiols, that is the formation of a covalent link between the electron-rich sulfur nucleus of the thiol and an electron-poor nucleus of the compound. This is another aspect of the same tendency already illustrated in equation (II), of a mercaptan to share one of its free electron pairs with an electro- philic atomic nucleus. Among such atomic nuclei are the heavy metal ions which form very slightly dissociated mercaptides according to the type equation; 2 R-S-H + Hg+^^rtR-S-Hg-S-R + 2 H+ (VII) Trivalent arsenic and antimony, bivalent cadmium and lead, jmonovalent gold and silver readily form analogous mercaptides. Though all reactions of the type illustrated for mercury in equation (VII) are reversible, the equilibria are usually displaced far to the right. Such a simple mercaptide is not the only type of compound formed between thiol and metal. Some metals can, under favorable conditions, form chelate complex compounds when there is near to the thiol group another group also capable of complex formation. The simplest example of this is illustrated by compounds of the type formed from BAL and cadmium chloride. These were studies by Gilman and coworkersSO and are of several forms. Mixing cadmium chloride and BAL in neutral or acid solution precipitates the white compound (VIII). At alkaline reaction and with excess thiol a water soluble cadmium complex is formed to which should probably be assigned the structure (IX). Conversion of (VIII) to (IX) occurs at pH 7 to 8. Analogous compounds are Cd 1 Na2 I Cd' S-CH-CH2OH HOCH2CH-SX XS-CH-CH2OH j (VIII) (IX) formed with mercury. Such metal complex compounds are quite stable and are similar to the simple mercaptides of mercury, arsenic and antimony in being only very slightly dissociated. Complexes of the type (IX) have the coordinating metal in the tetracovalent condition, they are stable anionic complexes. That all four sulfur atoms are coordinated with the metal nucleus is shown by the acid strength of the complex which is such that at pH 8 the anion exists as a disodium salt whereas mercaptans are not completely ionized until a pH of about 11 is reached. The low dissociation of such complex ions into metal ion and BAL in neutral to alkaline solution is shown by the fact that on addition of sulfide ion no heavy metal sulfides precipitate. The conversion of the water insoluble complex (VIII) to the water soluble complex (IX) is completely analogous to the changes occurring among complexes formed between the divalent metal ions, iron or cobalt, and the thiol acids, cysteine or thioglycollic acid. 114,115 These complexes must be made in the complete absence of air. At pH 4 to 6 they precipitate as crystalline complexes of the form (X) while in the presence of excess thiol acid and at pH 8 to 10, water soluble complexes of the form (XI) are produced which can be separated in crystalline condition by addition of alcohol. Compounds of the type (XI) are the primary stage involved in the iron
273 S-CH2 Fex Na2 r COâ0 O-CO (X1 (X\1 catalyzed oxidation of thiol acids by oxygen. Such complexes may also be involved in the oxida- tion of protein thiols even when oxidation is carried out by oxidanls other than air. This is indicated by the work of AnsonZ who showed that oxidation of denatured egg albumin thiols by ferricyanide is inhibited by cyanide. Both the complex types represented by (VIII) or (X) and (IX) or (XI) are of interest in considerations of the action of drugs in tissues since they could be formed in the physiological pH range. Besides these reactions between metal ions and thiols there is another kind that can occur between some metal ions and disulfides'. This has been described in a series of studies by several workers for the metals silver and mercury. 105,127 It is an oxidation-reduction reaction in which part of the disulfide molecule is reduced to thiol and part is oxidized to a higher level. At first this higher level was thought to be a sulfonic acid but the most recent work of Lavine'4 has supplied reasonable evidence that the higher level is a sulfinic acid, at least for the case of the disulfide cystine. The reaction written in two ways to account for these two opinions would be; 2 R-S-S-R + 2 H2OâÂ»3 R-S-H + R-S-O2H (XII) 3 R-S-S-R + 3 H2OâÂ»5 R-S-H + R-S-O3H Though this reaction has been studied mainly in acid solution this is primarily for reasons of experimental convenience. There is no reason to think it can not also occur at neutral pH range. The function of the metal ion appears to be to upset equilibria and to drive the reaction toward the right by removal of the thiol as insoluble mercaptide. Whether the formation of a soluble complex with low dissociation would be equally effective in driving the reaction forward has never been studied. There is still uncertainty regarding many phases of the reaction; it is conceivable that with different disulfides or in different pH ranges or with different metals, one or the other of reactions (XII) would predominate. Though the main driving mechanism of the reaction seems to be the formation of the. slightly dissociated mercaptide, insolubility of the metal sulfinate or sulfonate would also help to determine its direction. There has been no study of possible application of this reaction to such disulfide containing proteins as keratin or insulin. Further- more, if the driving mechanism of the reaction is the formation of slightly dissociated metal mercaptides then it could be expected that trivalent arsenic and antimony compounds might bring about similar changes. Such possibilities have never been described but would be rather easy to test. One of the commonest types of nuclear condensation of a thiol compound is with alkyl halides. This reaction is probably preceded or at least assisted by ionization of the halogen; R-CH2-C1ââ¢C1" + R-CH2+ + R-S~âÂ» R-CH2-S-R (XIII) This is really similar to the nuclear condensations already described as for example in (II) where condensation of thiol occurs with an ozidized, electron-poor sulfur nucleus or in (VII) where it occurs with a metal ion. In (XIII) the condensation occurs with an electron-poor carbon atom and yields a thio ether. This is the type of condensation that takes place when alkyl halides such as chloracetate, mustard gas or bromacetophenone react with thiols. The speed of such reactions varies enormously with both the alkyl halide and the thiol undergoing reaction. The possibility of alternative oxidative reactions that can occur under certain conditions with particular halogen acids has already been mentioned in equation (III). The extent and limitations of such oxidative reactions has not yet been explored. The reactions of mustards with thiols was first investigated by Helfrich and Reid and appeared to be simply of the type (XIII). More recently, as a result of intensive study of both sulfur and nitrogen mustards, the reaction has been found to be far
274 more complex. Here it can be only briefly noted that while both sulfur and nitrogen mustards appear to react in aqueous media by intermediate formation of sulfonium and ammonium compounds respectively there is a fundamental difference. The sulfur mustards' ^ appear to react by way of intermediate forms such as (XIV) while the nitrogen mustards" are believed to Â©XCH2-CH2âCl CH2-^CH2 CH2-CH2-S ^NQ S NCH2-CH2-C1 CH3 XH2-CH2-C1 XCH2-CH2âCl (XIV) (XV) react by way of intermediates of the type (XV). In each case the conclusions are based on studies of the kinetics of the reactions and on the isolation of intermediate compounds or their derivatives. The unexpectedly high reactivity of both sulfur and nitrogen mustards is presumably related to the formation of the positive onium ions which would be expected to have a greater tendency to act as alkylating agents because of a greater tendency to yield positively charged alkyl ions than the simple alkyl halides of reaction (XIII). The alkylating action of a sulfonium compound can easily be shown to occur by warming thetine and thiourea in neutral aqueous solu- tion. The following reaction occurs ;1Z4 OdOC-CH2-S (CH2-COCn dOC-CH2-S (CH2-COCn2 + S=C(NH2)2â Â»S(CH2COe2 +xlC-CH2-S-C(NH2 (XVI) These considerations suggest another kind of reaction than can take place between sulfur compounds and active alkyl halides and that has not often been considered in biochemical studies. This is the interaction of the alkyl halides with thioethers, either such as are formed as a result of reaction (XIII) or such naturally occurring ones as methionine. This reaction, which leads to the production of sulfonium salts, has been studied for the particular case of the action of mustard gas on methionine, yielding the disulfonium ionj'31 CH3 S(CH2-CH2-S-CH2-CH2-CHNH2-COOH)2 (XVII) Such sulfonium salts could themselves again act as alkylating agents, transferring one of the combined alkyl groups to some other nucleus capable of attaching an alkyl group as in (XVI). As with the action of mercury salts, it has, in the past, been generally assumed that alkyl halides react only with thiols. More recently the possibility of reaction with thioethers has been considered. That alkyl halides could also react with disulfides has scarcely ever been considered. Yet it was long ago shown'.33 that disulfides can also react with alkyl halides according to the equation; CH3-S-S-CH3 + CH31â>(CH3)3SI + (CH3)3SI3 (XVIII) Though this reaction was carried out at higher temperatures it is altogether likely that with more reactive alkyl halides, such as the mustards, a similar reaction could take place under physio- logical conditions. The formation of sulfides according to equation (XIII) has usually been considered to be practically irreversible but recently Peters and Wakelin99 have described a procedure in which a thio ether in the presence of a mercury or silver salt at acid reaction undergoes a scission such that on subsequently making the solution alkaline a positive nitroprusside test can be shown. The reaction was studied in particular for the case represented in (XIX);
275 SO2(CH2-CH2-S-CH2CHNH2COOH)2âÂ»HS-CH2-CHNH2COOH (XIX) + HO-CH2-CH2-SO2-CH2CH2-S-CH2-CHNH2-COOH To what extent this reaction is related to the particular structure shown is not at present appar- ent. It seems likely that the presence of the thioether in a position p with respect to a sulfone might make it especially likely to undergo a hydrolytic reaction. Of the two thioether groups present only one is split. The reaction may be related to a hydrolytic process 1ong ago studied by Otto in which disulfones are split in alkali; R-SO2-CH2-CH2-SO2-RâfcR-SOOH + R-SO2-CH2-CH2-OH (XX) This splitting of a thioether with a silver or mercury ion according to (XIX) is also reminiscent of the similar splitting of disulfides by the same metal ions. It must also always be kept in mind that these alkyl halides can alkylate atomic nuclei other than sulfur nuclei. 87, 118 Frequently such alkylations at nitrogen or oxygen nuclei are slower, but just as the speed of sulfur alkylation can vary greatly as a result of nearby structural features, so it may be expected that alkylation of other nuclei will also be greatly influenced by neighboring structural elements. We know that protein structures and protein denaturation have great effects on alkylation speeds though we have as yet only vaguely formulated theories as to why this is so. RosnerllO showed that iodoacetate continues to react with denatured egg albumin long after the reaction with thiol groups has been completed. Anson and Stanley3 showed that tobacco mosaic virus could be almost completely inactivated by iodoacetamide under conditions such that few, if any, thiols react. Two am mo acid residues that would appear most likely to be involved in such a reaction through their side chains are lysine and arginine. The next type of reaction we will consider is that in which the thiol attaches itself to an ethylenic structure. Such a structure has, among its resonance forms, one in which there is an electron-poor carbon nucleus and this will be the one to which the thiol sulfur will attach itself; (XXI) This reaction, as most of the preceding, involves attachment of the electron-rich sulfur nucleus to an electron-poor nucleus. The process has often been studied from the point of view of Markownikoff's rule. ^ These reactions are generally not readily reversible. The reaction of some mustards with thiols has been thought to go through a stage in which the alkylhalide of the mustard is first converted to an alkene. Â° Di vinyl sulfone in some respects reacts to give the same products as mustard sulfone. For instance with glycine both sulfones yield the same sulfonazane ;1 8 ^CH^i' CHj S02 ^,N-CH2-COOH (XXII) XCH2-CH2 The reaction of both sulfones with thiosulfate, an inorganic thiol, led to the conclusion that mustard sulfone is first converted to divinyl sulfone and so both sulfones finally yield the same thiosulfonate;130
Z76 S02(CH2-CH2C1)2âÂ»S02(CH=CH2)2âÂ»S02(CH2CH2-S-S02-ONa)2 (XXIII) Somewhat similar to the addition of thiols to alkenes is the addition of thiols to carbonyl groups ;116, 117 v . * f^-r"* ucc * ^V" twiv\ '*--. j.* â O- H '>**Â»Â». V"*X IAJUVJ R' R R SR Though reaction (XXIV) is quite analogous to reaction (XXI) yet there is an important difference in degree of stability of the products. Reaction (XXIV) is far more readily reversible, so much so that the products often give positive nitroprusside reactions and always react with iodine to regenerate the carbonyl compound and produce the disulfide of the thiol. The reaction with iodine is so rapid that titration to a sharp end-point can be performed just as with free thiol. Generally reaction (XXIV) occurs only with more reactive carbonyl compounds such as occur in aldehydes but reactive ketonic carbonyl groups such as that of pyruvic acid also form such addition com- pounds which can in most cases be called semi-mercaptals. Acetylation of the hydroxyl group generally stabilizes the products. Such acetylated semi-mercaptals no longer give a nitro- prusside test, even at pH 10. They are only very slowly oxidized by iodine and cannot be titrated to a sharp end-point. An interesting property of the semi-mercaptals formed in reaction (XXIV) is the ease with which they react with amino compounds ;1 * ' R yOH R /NHR1 ^C + R'NH2â* C (XXV) R SR R SR In the particular case of cysteine the thiol and amino groups are in the same molecule and appropriately spaced so thiazolidine compounds result;lO8, 117 HS â CH2 S CH2 R-CHO + I , â¢ R -f.H l " (XXVI) H2N-CHâCOOH NH-CH-COOH Such thiazolidines also dissociate sufficiently so that on treatment with iodine they yield cystine and the original aldehyde. On treatment in neutral aqueous solution with iodoacetate they yield S-carboxymethyl cysteine. Acetylation of these thiazolidines, which occurs at the nitrogen atom, stabilizes them so that the products are only very slowly oxidized by iodine to yield cystine. Such thiazolidines can also be formed from the half hidden aldehydes of reducing sugars, 120 but curiously pyruvic acid seems to react with cysteine to produce only a semi-mercaptal and not a thiazolidine derivative. LiebermanSO has found that 3-ketosteroids also yield such thiazolidines with cysteine. Application of these reactions to protein thiols have been made in the study of keratine. 88 Related to the formation of semi-mercaptals and thiazolidines is the reaction in which aldehydes give cyclic mercaptals directly with HS-CH2 S-CH2 R-CHO + j ââ¢R-CrT^ | ' (XXVII) -CH2OH itr"' I " H-CH2OH NS-CH2- Though there are no drugs in use which are thought to act as a result of reactions of the form of either (XXI) or (XXIV), these reactions have been discussed at length because of the fact that some antibiotics have structures combining both the ethylene linkage and the carbonyl group, thus uniting the possibilities of both reactions (XXI) and (XXIV) in one molecule. Addition of thiol to such a system of double bonds in conjugate position has been studied for several cases. 17,90 The condensation of thioglycollate and benzalacetophenone gives a p-keto-sulfide ;
277 C6H5-CH=CH-CO-C6H5âÂ» S-CH2-COOH (XXV1H) Such thioethers seem to dissociate reversibly far more readily than those of equation (XXI) but do so only at rather alkaline reaction. The thioether of (XXVIII) has a carbonyl group in the same position relative to the sulfur that in (XIX) is occupied by a sulfone group. This condition may be the cause of the more ready dissociation of these thioethers. The quinone molecule is one that in the smallest possible compass combines just about all the possibilities for reaction with thiols so far discussed. Quinones are strong oxidants ; they could be imagined to react with thiols by oxidizing them to disulfide; they have reactive ethylenic links and also reactive carbonyl groups and so might possibly react in accordance with equations (XXI) or (XXIV); they furthermore contain two separate and complete a,p unsaturated carbonyl groupings and so might be considered along with the antibiotics of this general structure. It is best, however, to consider quinone as a special system, among whose resonance isomers is one, with an electron-poor nucleus, which can attach to itself the electron-rich mercaptide ion; , o O O 9=9, + HSR - n | (XXIX) The resulting substituted hydroquinone has been isolated in several cases. .' In the presence of remainmg quinone this monosubstituted hydroquinone can itself become oxidized to a quinone and if there is more thiol present the whole cycle of reactions can recur leading to a disubstituted hydroquinone. In fact continued repitition of this whole sequence will yield, if the supply of thiol holds out, a tetra-alkylthio-substituted hydroquinone and such a compound with the structure (XXX) has been shown to be formed in good yield in the reaction between quinone and thioglycoll- ate. 121 The reaction of quinone with cysteine is far more complex but by working in quite dilute solution Kuhn and Beinert'' have isolated a crystalline product to which they ascribe the structure (XXXI). No proof is offered for this structure and another possibility that would need to be H-COOH OH H (XXX) (XXXI) (XXXII) considered is the structure (XXXII). In view of the high tendency of quinones to undergo nuclear condensation not only with thiols but also with amines this structure might seem even more probable. However formula (XXXII) contains two hydrogen and one oxygen atom more than formula (XXXI) and the analytical data fits (XXXI) better. Substitution reactions similar to (XXIX) have also been found to occur between thiols and naphthoquinones. However if the quinone used has all available positions adjacent to the quinone carbonyl groups substituted then these conden- sation reactions cannot occur. Such completely substituted quinones or naphthoquinones could however atill react as oxidants of thiols. Equation (XXIX) shows the condensation of an electron-rich thiol with a quinone. This type of condensation can also occur with great ease with other electron-rich nuclei, such as the nitrogen of amines. Amino-substituted hydroquinones have long been known to be formed by the action of many kinds of amines, even tertiary amines such as pyridine, with unsubstituted quinones. ' " It is probably such condensations that account for the tanning properties of quinone
278 which could condense with the free basic side chains of the lysine or the arginine residues of collagen. There remains to be discussed the complex chemistry of alloxan, but here we are on far less certain ground, for the chemistry of alloxan is even less well understood than that of quinones. Alloxan, like quinone, could be imagined to react with thiols either because of its properties as an oxidant or because it has an active carbonyl group. On reduction alloxan goes through two separate steps, absorbing a single electron at each step. The first step, which can be brought about by hydrogen sulfide at room temperature, yields the bimolecular form alloxantin and the second step, which can be brought about by the same reductant at elevated temperature, yields dialuric acid: ^.NU-CO NH-CO OH H CO-NH NH-CO co Nrojrsco' /câoâvc ^cojÂ±co^ ^CHOH (xxxm) NH-CO T^H-CO XCO-NH NH-CO On the other hand alloxan has one carbonyl group located between two adjacent carbonyl groups, a structure which also occurs in mesoxalic acid and ninhydrin. This structure is known to have a particularly marked tendency to add molecules such as water and alcohols. In conformity with this, alloxan is known to form stable addition compounds with water, alcohol and phenols. Addition of thiols has not been studied. In the cold alloxan reacts with sodium bisulfite to produce a normal bisulfite-carbonyl addition compound. At higher temperatures this breaks up to give the reduction product alloxantin. It is not possible at present to make a decision between the two simple and obvious reactions that could occur between alloxan and thiols; reduction of alloxan to dialuric acid and oxidation of thiol to disulfide or condensation of the two compounds to a semi- mercaptol. Other reactions, in addition to these two, could only be imagined. This finishes the survey of the known chemical reactions that could occur between the drugs of Table 1 and thiols, or their close relatives, the disulfides. It might be well to summar- ize this briefly as regards specificity. The oxidants, as oxidants, appear to be fairly specific as a group; disulfide, sulfide-sulfoxide, tetrathionate, pentavalent arsenic and antimony do not oxidize any other known tissue components nearly as rapidly as they oxidize thiols. The work of Purr'07 suggests an exception; the possibility that a mixture of ferrous iron and ascorbate may be oxidized by disulfide. Also rather specific for thiols are the trivalent arsenic and antimony compounds, including lewisite. These react very easily with thiols and with little else except hydroxy acids and phenolic compounds. Mercury raises the interesting possibility of being able to react with either thiols or disulfides. It is also known to form complexes with amino com- pounds. With the alkyl haltdes the variety of possible reactions suddenly becomes quite broad and includes any electrophilic nucleus. The unsaturated carbonyl compounds and the quinones also can react with a great variety of different molecular types besides thiols and do not even show any great preference for thiols. Of alloxan we can only say at present that it reacts with many kinds of molecules. With this background of the possible ways thiols or disulfides could react with the com- pounds listed in Table 1 we are in a better position to examine some of the drugs used and to judge the value of explanations offered for their mode of action. There remains however one important feature to keep in mind. All the preceding disucussion has concerned the interaction of simple thiol compounds with the reagents listed in Table 1. In the biological problems we must face however, the thiols are for the most part protein thiols and the reactivity or accessibility to reagents of protein thiols is known to be greatly dependent on the state of the protein. For those proteins which can exist in either a native or a denatured state, the thiols may be partly or entirely unreactive to many thiol reagents when the protein is in the native state. It seems appropriate to begin the examination of drugs that could conceivably have some action on a physiological thiol system with those that contain mercury. The older name of the thiol group is mercaptan, a word apparently synthesized from a Latin phrase to record the general property of the thiol compounds to capture or combine with mercury. But though this property of mercury to combine tightly with thiols has been known since the dawn of chemistry, it is only within the last few years that thought and experiment have been directed to an examina- tion of the possibility of explaining the action of the extensively used mercurial diuretics in terms of the combination of mercury and physiologically active thiols. Handley and LaForgeS' first
279 showed that the diurectic action of both organic mercurials, mercuhydrin and mersalyl, and of inorganic mercury compounds is prevented or abolished by BAL, thioglycollate and glutathione. Cysteine in contrast had little effect in preventing mercury diuresis. In dogs the administration of BAL caused an abrupt drop in urine output to p re-mercurial level. Sussman and Schackl3? came to substantially the same results using mercuhydrin and further pointed out that BAL inhibition of diuresis might be due to formation of a stable, mercury-thiol complex and not to an antidiuretic action of BAL, inself. Farah and Maresh.^S thought these effects of reversing mercur ial diuresis with thiols suggested the possibility that the mechanism of action of mercurial diuretics is an inactivation of thiol enzymes responsible directly or indirectly for the resorption of salt and water from the lumen of the kidney tubules. In addition to their diuretic effect, mercurials were known to have another important though far less marked physiological eff'ct. The use of mercurials as diuretics was known to be associated with occasional toxic effects on the heart and in animals such toxic effects had been measured. DeGraff and Lehman.^ had shown a protective effect against this cardiotoxicity of mercurials following administration of sodium thiosulfate. Lehman^ related the cardiotoxicity of mercurials to the degree of instability of the complexes, that is the ease with which such complexes dissociated, presumably to yield mercuric ion. Long and Farah81.82 had shown that the monothiols, cysteine, glutathione and thioglycollate, markedly reduce the cardiac toxicity of mersalyl. They further found that BAL could suppress completely both the cardiotoxic and the diuretic actions of mersalyl. In the reversal of either the cardiotoxic action of this mercurial or its acute toxicity by thiols, BAL was, in each case, effective in about a fifth of the equivalent amounts of cysteine or glutathione. Farah and Maresh3' found it possible to produce an important differential effect. By the use of the monothiols, cysteine and glutathione, they could abolish the cardiac toxicity of mersalyl without reducing its diuretic action. All of this work vjas consistent with the idea that both the diuretic effect and the cardiotoxic action of mercurials was due to inactivation or disturbance, by the drug, of physio- logically essential thiol systems. But the finding of Farah and Maresh, besides its obvious clinical importance, raised new ideas concerning the mechanism of mercurial action. If mono- thiols could compete successfully with the cardiac system that was poisoned by the mercury but could not compete with the renal system, it would appear as if these two physiological systems exerted quite different degrees of mercury binding. The renal system seemed to bind the mercury in a manner far less readily reversible. This could be interpreted as an application, at a bio- logical level, of phenomena already described at a biochemical level; that is that inhibition of some enzymes by arsenic or mercury could be reversed by monothiols or dithiols while inhibition of other enzymes by mercury could not be reversed by monothiols but only by dithiols. This reasoning had been the starting point for the extensive research programs centering about BAL. Stocken and Thompson1 35 have discussed many of the phenomena leading to the conclusions that the arsenic acceptor of many living cells can form more stable compounds with arsenic than any simple thiols so far investigated. Barron and his co-workersl2 in a study of a great variety of enzymes, found that inhibition by lewisite or by other arsenicals could be more effectively reversed by BAL than by glutathione. If inhibition by mercury is due to combination with thiol enzymes it would seem quite natural to expect for mercury a similar more complete reversal of effect with BAL than with a monothiol, provided ring compounds can be formed. Yet the work of Thompson and Whittakerl4O does not bear this out for the case of pyruvate-oxidase inhibited by mercury or by antimony. In both cases inhibition is reversed by monothiols and by BAL, yet there is not the same strikingly greater effect by dithiols. Barron and Kalnitskyll studied the reversal of mercury inhibition of succinoxidase brought about by a variety of thiols and also found the contrast in efficiency of dithiol and monothiol less pronounced than in the case of reversal of lewisite inhibition. In fact for mercury inhibition they found 1, 3 dimercaptopropanol far more effective than BAL in reactivating succinoxidase. The effect of thiols in reversing inhibition of enzyme action by heavy metals is thought to be a reflection of the competition of thiol and enzyme for the metal. The existence of such a competitive effect by no means proves that the enzyme combines with the metal through thiols of its own. If the enzyme-metal compound dissociates rather readily, then monothiols as glutathione and cysteine suffice to take the metal away from the enzyme. In the case of more tightly bound enzyme-metal compounds monothiols are ineffective metal removers and only BAL type dithiols are effective. But both among dithiols and among monothiols there is a great variation in ability to reverse mercury or arsenic inhibit- ion of enzyme activity. Thiols cannot be put in a linear sequence with regard to effectiveness in reversing metal inhibition, nor can it even be said that all dithiols are more effective inhibition reversers than all monothiols. With organic mercurials the possibility of formation of five membered rings, such as lewisite can form with BAL, may not even exist.
280 The work so far reported strongly suggests the possibility that mercurial diuretics bring about their action by inhibiting physiologically essential thiols of the kidney tubule but can not be regarded as proof. There is nowhere yet any direct proof that mercury reacts with any physio- logically functional kidney thiols. All that is proved by reactivation experiments with monothiols and dithiols is that whatever it is that mercury does can be reversed more or less effectively by these agents. The possibility of mercury inhibition by complex formation with amide, amino or heterocyclic nitrogen compounds must still be recognized and would be just as consistent with all inhibition reversal experiments. The use of theophylline with mercurials is probably itself an expression of the combination of mercury with an amide. Yet this thiol theory of the action of mercurial diuretics is a good working theory because it will suggest the further experiments needed to establish or to disqualify itself. It can also be of value as a guide to identify the enzyme systems of the kidney tubule responsible for salt and water resorption and to lead to an under- standing of this process. This thiol theory of mercurial diuresis has gotten valuable support from another line of work, that on the nephrotoxic action of tetrathionate. The tetrathionate ion has already been described as a member of a class of oxidants, the disulfides, which are rather specific for thiols. That tetrathionate can also react with protein thiols has been shown by the work of Anson2 on denatured egg albumin. In this protein all the thiols are exposed. On the other hand unrease contains both exposed and hidden thiols and only the latter condition its activity. 38,62 Since tetrathionate does not inactivate this enzyme Fischer and Goffart39 conclude that tetrathionate reacts only with exposed thiols. Oilman and his co-workers showed that tetrathionate produces discrete lesions of cells of the proximal tubules of the kidney with nuclear degeneration within a half hour after injection. They also showed that in rabbit and dog, tetrathionate is rapidly reduced to thiosulfate. After confirming the ability of this inorganic disulfide to oxidize organic thiols such as I.J/-.1 . cysteine and glutathione, they showed that cysteine or thiosorbitol protect rabbits against the nephrotoxic action of tetrathionate. In searching for a mechanism, the conclusion was reached that the effect of tetrathionate is extracellular. This was based on the fact that tetrathionate appears to be distributed in extracellular fluid. Goffart and Fischer56 carried this idea further and concluded that the attack of tetrathionate is not on protein thiols of the kidney at all but on glutathione. This conclusion was based on an estimate of the thiols of kidney tissue. The amount of Prussian blue deposited in kidney tissue after treatment with ferricyanide followed by ferric salt was estimated and it was confirmed that kidney tissue is rich in thiols. After intravenous lethal doses of tetrathionate the kidney slices still show as strong a Prussian blue deposit as normal kidney. On the other hand estimation of the glutathione content of rabbit kidney showed a disappearance of 80 percent of the reduced glutathione after injection of a lethal dose of tetrathionate. They conclude, as did Oilman and his co-workers, that the tetra- thionate acts on kidney cells by removing diffusible glutathione rather than by penetrating these cells itself. Philips et al. 100 studied the effect of tetrathionate on succinoxidase, a typical thiol enzyme, taking material from several sources. Tetrathionate is a powerful inhibitor of the succinic dehydrogenase component of succinic oxidase. On other grounds succinic dehydrogenase is known to be strongly inhibited by thiol reagents. ' * An assay was then made of the activity of the succinoxidase system of rabbit kidney cortex 60 minutes after an intravenous nephrotoxic dose of tetrathionate. This assay showed that no specific inactivation of succinic dehydrogenase in vivo had occurred as a result of tetrathionate injection. So it was concluded that inactivation of some other thiol enzymes of kidney cortex must be more directly concerned with the nephrotoxic action of tetrathionate. Clinical evidence'5 shows some diminution of tubular resorptive capacity after tetrathionate administration indicating a similarity of the action of this drug in man and laboratory animals. There are now known several types of compounds that are nephrotoxic and that have in common a high degree of reactivity with respect to thiols. Of these, mercurials could act by mercaptide or complex formation, tetrathionate by oxidation. The convergence of the results of work with mercury compounds and tetrathionate gives weight to the theory that a thiol system is involved in their physiological action. The nephrotoxic effects of cadmium and alloxan point in the same direction. Oilman" finds that the water soluble cadmium-BAL complex exerts a marked nephrotoxicity. It is for this reason that prophylactic administration of BAL, to rabbits with intravenous lethal doses of cadmium chloride did not save the animals from fatal renal insufficiency. The higher renal toxicity of cadmium-BAL complex than of simple cadmium ion is ascribed to its ability to penetrate cells of the renal tubules where liberation of cadmium by metabolic transformation of the complex follows. Nephrotoxic action of antimonials will be mentioned later.
281 Closely associated with the effect on the kidney of the drugs described is their effect on the heart. This has been observed in the clinical use of mercurials in cases of cardiac insuf- ficiency. The effect of mercurials on the heart can occur by various mechanisms. Mercuric chloride removes the inhibitory effect on the rhythmical contractions of heart muscle due to the action of the vagus nerve; this effect is restored after washing the heart with cysteine. ?0 At another level of action, Bailey and Perry? found actin-myosin interaction to depend on the thiols of myosin. These thiols are also concerned in ATPase activity of myosin. A variety of thiol reagents, including iodosobenzoate, hydrogen peroxide or iodine, iodoacetamide or chlormercuri- benzoate, reduce both enzyme activity and ability to form actomyosin at about the same rate. The groups concerned are.mainly, but not entirely, those giving the nitroprusside test and are accessible to oxidants. Godeaux52 found that the supercontraction of myosin fibers in a solution of ATP containing potassium chloride and magnesium chloride was inhibited completely and instantly by mercuric chloride; the supercontraction reaction was also inhibited more or less rapidly by hydrogen peroxide, potassium iodate, chloropicrin, mustard gas, chloracetone and iodoacetamide. He furthermore believed the order in which these compounds arranged themselves with regard to this action on myosin fibers to be identical with the order of their action on protein thiols. While chloropicrin blocks thiols completely and inhibits myosin supercontraction completely, monochloracetone can block only 60 percent of the protein thiols and only at high concentration (. 1 M) does it inhibit superprecipitation completely. 53,54 This slower acting vesicant, he also found, could block 40 percent of the myosin thiols without having any effect on superprecipitation, but when it blocked the next 20 percent of the myosin thiols the superprecip- itation was completely inhibited Under these conditions the hidden thiols have not yet reacted. Polis and Meyerhof' point out that iodoacetate attacks only superficially located thiols while mercury compounds also attack less accessible thiols. They show that low concentrations of iodoacetamide, iodoacetate, iodosobenzoate and phenylmercuric hydroxide activate ATPase by 30-80 percent. Glutathione antagonized both activating and inhibiting influences. Fredericq^ finds that chloropicrin, while reacting to destroy all the thiols of native myosin, does so in two distinct stages, the first is an almost instantaneous reaction, the second is a gradual one. Another aspect of the action of mercurial compounds supposed to be related to thiols is their use as antiseptics and parasiticidal agents. Though this bactericidal effect of mercuric chloride is antagonized by thiols36 it cannot be reversed by thiols. 139 That mercury compounds can also inhibit enzymes supposed not to require thiols has raised suspicions as to whether in its antibacterial action mercury really attacks essential thiols. The next drugs of Table 1 include the trivalent and pentavalent arsenicals. This is the group of drugs most thoroughly explored with regard to action on thiol systems. The arsenicals are the drugs which, together with the disulfides, are probably most nearly specific in reacting readily with thiols and not with any other known tissue components. Since the aspects of arsen - icals in their role as therapeutic agents for the treatment of infections with spirochaetes or tryp- anosomes has already been analyzed by Eagle and Doak it will not be discussed here. The aspects of arsenicals related to the vesicant action of lewisite will be discussed later in connection with the halogenated vesicants. It is curious that though the arsenicals constitute the group of drugs whose mode of action has longest been suspected to be related to thiols, and have been most thoroughly studied in this connection, the antimonials have scarcely received any attention at all from this point of view and generally have been far less studied in the laboratory. Even the preparation of new compounds has lagged far behind the more impressive array of available arsenicals. Yet clinically anti- monials have been used almost as long as arsenicals. Disregarding its more ancient uses, antimony in the form of tartar emetic came into clinical use in 1908 in the treatment of trypanos- omiasis, in 1913 in leishmaniasis and in 1915 in schistosomiases. Certainly it was natural, because of the early promise of arsenicals and because of the close chemical similarity of arsenic and antimony, to test antimonials for anti-parasitic properties. Though antimonials have been used in combatting a number of disease conditions, it is principally in schistosomiasis, filariasis and some forms of leishmania infections, particularly kala-azar, that they have come to be most widely used. Because so much less work has been done with the mechanism of action of antimonials in biological systems and even on isolated enzyme systems we can only guess at its mode of action by analogy, based on the similarity of chemical reactions of antimony compounds to those of arsenic compounds. Certainly in simple chemical systems we know that pentavalent antimonials are as readily reduced.by thiols to the trivalent conditions as are the pentavalent arsenicals. We also know that trivalent antimony forms mercaptides at least as readily as the trivalent arsenic compounds. That the formation of such mercaptides also extends to protein
282 thiols appears probable from the claims of several patents in which antimony is reported to form combinations with the products resulting from partial hydrolysis, reduction and dialysis of keratins. 136, 144 The function of the partial hydrolysis is merely to render the kerateine water soluble. Even before the advent of BAL the protective effect of monothiols on antimonial toxicity to mice had been noted by Launoy. ?3 More extended studies by Chen and Geiling26 showed that cysteine not only reduced the toxicity to mice of trivalent antimonials but also reduced their toxicity to trypanosomes. Studies of the effect of BAL on antimony toxicity has led to contra- dictory results. Braun et al. 19 found BAL to decrease the toxicity of tartar emetic, fuadin and neostam, but Gammill et al. *4 while finding that BAL decreased the toxicity of tartar emetic, reported that it increased the toxicity of fuadin and neostam. Sandground'' ' has reported findings of considerable interest on the detoxication to rats of the pentavalent antimonial stibosan with Â£-aminobenzoic acid. At the same time Â£-aminobenzoate did not interfere with the trypanocidal potency of stibosan. This was an extension of previous work on similar detoxication of penta- valent arsenicals. Subsequently Sandgroundl 1 2 showed that a variety of substituted benzoates also provided great protection to rats against carbarsone and arsanilic acid. Williamson and Lourie'43 found that p-aminobenzoate antagonizes the trypanocidal activity of tryparsamide. McChesney et al. 86 broadened the field of protective agents still further when they reported that ascorbate and some of its analogs protected rats against toxic effects of neoarsphenamine. All these detoxication studies have been made in vivo. Whether they have any relation to thiols is hot at present known. They can serve as leads for future work and in the meantime may give rise to some caution against too ready acceptance of the thiol theory of antimonial and arsenical drug action. A question that must constantly recur with reference to the action of drugs presumed to operate by affecting a thiol system is that if one such drug produces a certain physiological effect, why do not all the others produce the same effect, at least in some measure? If mercurials damage kidney tissue and more specifically the tubules, why do not antimonials also do so? Franz41 found that one of the earliest changes, observable within a few hours after administration of tartar emetic intravenously to rabbits, was a swelling and degeneration of kidney epithelium. Harris60 found the most conspicious injury produced in rats given intravenous lethal doses of tryparsamide or stibosan was a necrosis of the convoluted tubules. p-Aminobenzoate given intravenously or orally to such poisoned rats resulted in a marked decrease in such lesions. As regards the direct action of antimonials on the parasites against which they are used, we have such data as that of Gellhorn and Van Dyke*' on the accumulation of antimony in those organs as the spleen and liver in which the leishmania parasites congregate in hamsters. Some interesting findings were made in a screening study of drug effects on experimental infections of Schistosoma mansoni in mice. 122 It was shown that whereas almost any antimony compound had some effect on the parasitic worms, arsenic and bismuth compounds were completely without effect. A single mercurial had a weak effect in ridding the mice of these parasites. The specifi- city of antimony in this case, though arsenic, bismuth and mercury as well as antimony all react similarly with thiols, is not explained. Furthermore though all antimonials showed some effect in disturbing the worms or actually killing them their efficacy varied widely. It seemed that fat soluble antimonials were more effective and those tried were mostly of the type Sb(SR)3 in which antimony is already combined with thiol. Finally, one of the most efficacious drugs foundl23 was the antimony complex of BAL, a complex in which antimony might be expected to be tightly enough bound so that it could not be toxic to either host or parasite. This drug is given by mouth and actually is scarcely toxic to the mouse but appears highly effective in ridding the mouse of the schistosome worms. The schistosome worm is a more highly organized animal than the bacteria or protozoa. It has a cuticle and a digestive tract and it is likely that the drug reaches the worm through the digestive tract. The worm lives in the bloodstream of its host; in the adult stage it lives in the veins of the mesenteries. It would seem that the most plausible path of the drug fed to the mouse would be to penetrate the intestine, enter the bloodstream, be ingested by the worm and penetrate its intestine. The visible effect of the drug is a degeneration of the sexual organs of the parasite. If antimonials are toxic to animals because of a thiol interaction it might be expected that they would be toxic to all animals. Yet when given to a host carrying parasites it is sometimes possible to kill the parasite without killing the host, that is, the therapeutic index of the drug is greater than one. If BAL forms a very firm complex with antimony and detoxifies the antimony with respect to the host, it might be thought that BAL would also detoxify the antimony with respect to the parasite. This is not always the case. We already know that though BAL can detoxify arsenic, antimony and mercury compounds in man, it intensifies the toxicity of lead and seems to cause no great change in the toxicity of cadmium. * ' BAL seems incapable of releasing cadmium from lung tissue. '25 The mapharside- BAL .compound is more toxic to rats than
283 mapharside itself, 95 yet the presence of excess BAL reverses these relative toxicities. Phenyl- thioarsenites were found to have about the same toxicity as the parent phenylarsenoxides to amebic parasites, but to have a considerably reduced toxicity to their hosts. ' With respect to schistosome worms Bueding21 has found that BAL increases the effect of mercury in inhibiting glycolysis, that is BAL intensifies the toxicity of mercury to the worm. It seems clear that we are in no position to make useful predictions concerning the effect of heavy metals combined with BAL in animals. Each species and each metal must be tested empirically and there is reasonable hope that useful combinations can be found to detroy parasites, whether bacterial, protozoal or helminthic, and at the same time to be relatively harmless to the host, whether man or beast. The next group of drugs to consider is that of the alkyl halides, including the war gases. The target of these drugs is perhaps best defined by their descriptive classification as lacri- mators, sternutators, lung irritants and vesicants but though these are the immediate and obvious effects most of the compounds have potentialities for slower, far more generalized systemic toxicity, both cytotoxic and nucleotoxic. This group also, includes arsenicals of the lewisite type and the muscle poison, lodoacetate, which Lundsgaard brought into prominence. All the members of this group are either alkyl halides or arsenyl halides. They are all capable of reaction with thiols in such a way as to attach directly to the sulfur the alkyl group as in (XIII) or the arsenyl group in a manner analogous to equation (VII). The products are called in the first case thioethers, in the second arsenic mercaptides or thioarsenites. It has already been pointed out that different reaction mechanisms are known to occur, as for the sulfur and nitrogen mustards, (XIV, XV, XXIII); and that even totally different end products may result, as shown in equations (III). There is no question that all these compounds could react with thiols, there is even not much doubt that in biological systems they do react with thiols. However, whether their main action in biological systems, that is whether the action that produces the observed biologi- cal effect is directly due to interference with physiologically functioning thiols is not at all clear. Chemically the group is characterized as a highly electrophilic group. Many of the compounds are capable of reacting with ele'ctron rich atomic nuclei other than thiol sulfur. Of particular importance would be nitrogen in amino, imino or heterocyclic form. For instance the tertiary amino nitrogen of pyridine reacts easily with iodoacetate and so do other tertiary amines. In the reduction of the pyridine nucleotides during their natural functioning cycle the quaternary pyridin- ;um salt is reduced to a tertiary amino compound and in this condition could possibly react with the active alkylating agents of this group. This has never been shown to occur. The arsenic compounds of this group, compounds of the lewisite type, are on the other hand probably far more specific for either thiol or hydroxy compounds since arsenic seems to have little affinity for nitrogen and few compounds are known with an arsenic-nitrogen link. The action of lewisite on kerateine investigated by Stocken and Thompsonl 34 led to the conclusion that only the thiols of the protein react with lewisite. On the other hand it has already been pointed out that another vesicant, mustard gas itself, readily combines with collagen, a protein containing no thiols, yielding a sulfur containing product. Mustard in aqueous solution has been shown to react with carboxylate ions to yield esters. 89 Binet and Wellers^ found the action of mustard gas on yeast and dog skin did not affect the thiol content and conclude that the toxicity of mustard is not due to reaction with tissue thiols. The amount of mustard sulfur combining with proteins may be ten times the equivalent of the thiols that react. ' Goffart^ studied the effect of a variety of vesi- cants on skin and concluded that while chloropicrin, chloracetophenone and iodoacetamide react rapidly with skin thiols, mustard does not act in the same way. The effect of mustard on yeast growth can not be interpreted simply as a binding of glutathione. 69 [n a study of the effect of lacrimators. Mack-worth" found them to react far more rapidly with protein thiols than with simple thiols and made the interesting observation that denatured enzymes containing no demonstrable thiol could protect active enzymes against inhibition by lacrimators. Yet under the same conditions, denatured egg albumin, known to contain active thiol, afforded no protection. Peters and Wakelin98 showed that mustard gas does not abolish fixed thiols of muscle though mustard aulfone and phenyldichloroarsine readily do so. Examining the reaction of mustard with kerateine they find only a quarter of the combined mustard to be bound to thiol. Thus mustard differs from lewisite which seems to attack only thiols. These same authors'?? point out that the toxic dose of mustard gas in vivo is of a far lower order of magnitude than the concentration required to inhibit enzyme systems. They found N, N-diethyldithiocarbamate, which might have been expected to decrease the inhibition of the pyruvate oxidase system by mustard, actually increased the inhibition. That the potentiating effect of the dithiocarbamate might be due to formation of the compound (C2H5)2N-CS-S(CH2)2S(CH2)2C1 seemed possible from the high toxicity of this compound to pyruvate oxidase. A similar report of increase in toxicity of mustard is that of Kucharik and Telbisz' 1 , that a mixture of mustard and glutathione is more toxic to paramecia than mustard alone. It is findings such as these that emphasize the objections to the
284 simple concept that all potential thiol reagents really produce their biological effects by reaction with physiologically essential thiols. Certainly it cannot be said that the drugs of this group produce their effects by similar mechanisms or even that they all react with the same targets. Yet there is a curious similarity in their immediate effects which permits grouping them as vesicants. Perhaps even more complicated than the vesicants and their relatives are the antibiotics. Among these compounds are found a wide variety of structures. Of these, two types have been listed in Table 1 as potentially capable of reacting with thiols and in fact known to react with the simpler ones. These are the a, p or p, \ unsaturated carbonyl compounds on the one hand and the disulfides or alkylthiosulfinates on the other hand. The curious property that many antibiotics have of becoming inactivated on incubation with simple thiols, even when there is no obvious possibility of interaction known, was first pointed out by Cavillito and Bailey. Z3 Geiger and Conn first pointed out a structural feature of some antibiotics, the unsaturated ketones, that could account not only for the inactivation of these antibiotics by cysteine, but for the antibacterial action itself. This idea was worked out in particular for clavacin and penicillic acid and served as a guide in a search for other antibiotics among relevant organic compounds and led to the find- ing of some acrylophenones and dibenzoylethylenes with marked antibacterial activity. 46 Such antibacterial unsaturated ketones when mixed in excess with thiols abolish the nitroprusside reaction of the thiols and when mixed with excess thiol lose their antibacterial properties. These unsaturated ketones were further found to have an inactivating effect on enzymes known to have essential thiols. Cavallito and Haskell^ extended this structural idea to include a. p and ^.-, unsaturated lactones and for the ^ P*Y -angelica lactones isolated products of reaction with cysteine and related aminothiols. Broderson and KjaerZO made an extensive study of unsaturated lactones and found only a few to have antibacterial properties. They concluded that this structure by itself was not sufficient to give a compound antibacterial properties. Rinderknecht et al. 109 in general supported the results of Geiger and presented additional data along the same line. CavallitoZ2 made a study of thiol structure as related to inactivation of a number of different antibiotics and found two main groups of antibiotics; one composed of such substances as gliotoxin or diallyl sulfoxidesulfide, readily inactivated by any of a half dozen thiols tried; and a second group, including penicillin and streptomycin which was rapidly inactivated only by cysteine or a derivative of cysteine having both thiol and amino group free. On the basis of this difference in specificity to inactivation by thiols he suggested the interesting idea that these latter antibiotics might block the growth of a polypeptide chain by blocking the free cysteinyl end when this was formed at the end of such a chain. Bailey and Cavallito" made an important distinction among those antibiotics inactivatable by incubation with cysteine, dividing them into two distinct groups on the basis of whether the bacteriostasis they produce is, or is not, reversed by addition of cysteine. Those in which this bacteriostasis is readily reversed by cysteine are characterized as showing little specificity to inactivation by thiols; those in which bacteriostasis is not reversed by cysteine are more specific and inactivated only by cysteine or close analogs. This classifica- tion is identical with that just described based on the specificity of the thiol inhibitor. Thus a reasonable case has been made that the antibiotics with unsaturated ketone or unsaturated lactone structure could exert their action through non-specific binding of essential thiols, reversible by addition of cysteine. For the antibiotics gliotoxin and allylsulfoxidesulfide the picture is some- what similar though in this case the reaction with thiols, while also non-specific, is of the thiol- disulfide oxidation-reduction type. Z4 There remain the curious antibiotics penicillin and streptomycin. Both are inactivated by incubation with cysteine and both are rather specific in not being readily inactivated by other thiols. In the case of streptomycin the inactivation by cysteine might be due to thiazolidine forma- tion with the free carbonyl group of the biose component since streptomycin is also inactivated by carbonyl reagents such as hydroxylamine and semicarbazide. Furthermore dihydrostreptomycin is not inactivated by cysteine. Several workersl6,48, 141 have shown that streptomycin is inactiv- ated by a large number of reducing agents and even loses activity in the presence of bacteria under reducing conditions. From existing data no clear case can be made that streptomycin exerts its antibacterial effect by reaction with essential bacterial thiols though the possibility is still open. With the penicillins the relation to thiols is quite different. Penicillin can undergo no obvious reaction with thiols and it is in fact not rapidly inactivated by thiols generally but only by cysteine and cysteinyl glycine. The structure of penicillin has been shown to be sufficiently similar to that of glutathione37 to suggest the possibility of an anti-essential-metabolite function.
285 Pratt and Dufrenoy'03, 104 have also suggested a thiol involvement of penicillin action but along totally different lines. The argument rests on observations of patterns produced by redox potential indicators flooded over penicillin assay plates. The pattern with 2, 6-dichlorphenol- indophenol is as follows: the zone of inhibited growth is stained blue; the uninhibited areas are faintly blue and between these two areas is a narrow colorless ring that indicates a zone of enhanced growth. The colors of these zones is interpreted as being related to their glutathione contents in the sense that the less the color the higher the glutathione content and the higher the glutathione content the more active the growth. Then in the central zone of dark blue color there is no glutathione and no growth, in the outer areas of faint blue color there is a certain amount of glutathione associated with normal growth and in the narrow zone between these there is no blue color at all showing a high glutathione concentration and greatly enhanced growth. These authors then show that flooding plates with formaldehyde, which is assumed to block the glutathione thiols, results in the formerly narrow colorless ring of high glutathione content now staining deep blue showing the glutathione to have been blocked. These indications that penicillin inhibition is correlated with disappearance of bacterial glutathione is then linked with the observation of Gale and Taylor4.* that penicillin inhibits the uptake of glutamic acid by Staph. aureus. The possibility is suggested that penicillin prevents synthesis of the essential glutathione in the cell by preventing the entrance into the cell of a necessary component of the glutathione. Another curious feature of penicillin is that its antibacterial action is enhanced both in vitro and in vivo by the presence of low concentrations of cobalt salts. In line with the suggestion that the action of penicillin may lie in blocking glutathione formation by keeping glutamic acid out of the cell, it would seem reasonable to interpret the enhancing effect of cobalt by a related mechanism as inhibiting glutathione formation by removal of cysteine in the form of the highly stable complexes it is known to form with cobalt ions. Another suggestion for the mechanism of the inhibition of penicillin action by cysteine is that of a chain reaction of penicillin destruction initiated by cysteine. '26 This is based on the decreasing ratio of cysteine required to antagonize increasing concentrations of penicillin. Quinor.es, both benzoquinones and naphthoquinones, are known to have a strong anti- bacterial effect. 4,91 In addition to this they have been recognized as the most effective com- pounds for killing spermatozoa" and the cercarial stage of schistosome parasites. 113 Some naphthoquinones have highly fungicidal properties. 138 Finally some naphthoquinones have a chemotherapeutic as well as a prophylactic effect in Plasmodium infections in ducks and chickens. ^ ' The mechanism of the antibacterial action of quinones has been considered from the point of view of their high oxidation potential, 92 as well as from the point of view of their general reactivity. 58 The ease with which unsubstituted or partially substituted quinones react with simple thiols has already been discussed. In addition they condense rapidly with a variety of amines, including even cyclic tertiary amines such as pyridine. Colwell and McCall29 have considered it probable that the antibacterial and antifungal action of napthoquinones is due to reaction with thiols. Foote and co-workers4O have tried to relate the inhibition of spore germina- tion ai some fungi with the inhibition of carboxylase activity by a number of naphthoquinones. Carboxylase is a thiol enzyme. They went even further and, from the failure of prolonged incu- bation of naphthoquinones with carboxylase to cause any more inhibition than when these compon- ents were freshly mixed, concluded that the naphthoquinones are able to react with and to inhibit only the functioning enzyme and not the resting enzyme. They tried to give this idea some plausibility by pointing out that sluggishly reacting thiols of the enzyme might become available or reactive only during the functioning of the enzyme. The explanation of the antibacterial action of quinones as due to a condensation with essential thiols has been criticized by pointing out that many completely substituted quinones, such as spinulosin, are quite potent antibiotics. It is probable that such completely substituted quinones cannot condense with thiols by reactions of the form (XXIX). Halogen substituted quinones, however, seem to have a reactivity toward thiols similar to alkyl halides and far greater than aryl halides. For instance, chloranil reacts as readily with thioglycollate as quinone itself and yields identically the same product. 121 The mechanism of the reaction of thiols with halogenated quinones is altogether different from the reaction with unsubstituted quinones. In the former case no hydroquinone is produced, in the latter hydroquinone may be produced. It is conceivable that other sufficiently negative groups might be displaced just as are halogens from substituted quinones. Certainly the fungicidal properties of the 2,3-dichlor 1 ,4-naphthoquinone' 38 could be due to thiol interaction, by replace- ment of halogen by thiol. Hoffman-Ostenhof" has studied the inhibiting effect of a variety of quinones on known thiol enzymes such as urease and papain. He found a close parallelism in the effect of many quinones on these enzymes and thought that consequently the mechanism of quinone inhibition for these enzymes is probably similar. He further concluded that there was no such parallel between the inhibition of these enzymes by quinones and the antibacterial potency of the
286 same quinones, since many of the most potent antibiotics had little or no inhibiting effect on the enzymes. His conclusion is that the antibiotic action of quinones is probably complex. Geiger^S has made some observations that may help to unravel some of this complexity. In the first place he finds a difference in the requirements of gram-negative and gram-positive bacteria as regards the structure of quinones that will inhibit their growth. Whereas gram-positive organisms were stopped by all quinones tried, the gram-negative organisms were inhibited only by quinones with at least one unsubstituted ortho position. In the second place the antibacterial activity of quinones can be blocked by thiols only in the case of the gram-negative bacteria and not with gram-positive bacteria. These findings become of even greater interest when considered together with the work of Henry, Stacey and Teece on the isolation of a thiol containing nucleoprotein from gram- positive bacteria with specific gram staining properties. Such a nucleoprotein could not be isolated from gram-negative bacteria. Of great interest is the observation that under the influ- ence of bacteriostatic concentrations of penicillin, cells of Staph. aureus gradually lose their positive reaction to the Gram stain. ^ Another interesting physiological mechanism relating quinones and thiols was suggested in detail by Lyons. 84 He formulated a theory for the mode of action of Vitamin K as a functional part of prothrombin. Fibrinogen, he thought, had hidden thiols which were unmasked by one component of prothrombin only to be converted immediately to the disulfide of fibrin by an oxida- tion by the naphthoquinone component of prothrombin. One piece of evidence for this scheme was the claim to have found naphthoquinone in prothrombin by a colorimetric test. Damel was unable to confirm this. Yet some sort of thiol mechanism is implied in reports^" of the inhibition of fibrinogen coagulation by iodoacetamide. The last of the drugs of Table 1 to discuss is alloxan. This compound long known and studied has sprung into prominence during the last few years because with it a diabetic condition can be produced in experimental animals. Injection intravenously of proper doses of alloxan into several species of laboratory animals leads to rapid and selective necrosis of the p-cells of the islets of Langerhans. 93 Previous administration of glutathione or cysteine completely blocks this necrotizing effect of alloxan on the p-cells and also prevents development of diabetes. ?5 That thiols protect animals against development of diabetes only if administered immediately before or at the same time as the alloxan led to the opinion that the effect produced by alloxan is due to combination with thiols rather than to oxidation of them to disulfides. ?? The opinion that glutathione combines with alloxan was further supported by the appearance of an additional maximum in the absorption spectrum at 305,* p on mixing glutathione and alloxan. Yet it was also shown94 that glutathione can reduce alloxan to dialuric acid. The fact is we do not have clear data at present as to which reaction takes place or, if both take place, which predominates. Alloxan, like ninhydrin which it resembles structurally, also reacts with amino acids having a free amino group. 79 PurrlOk claims alloxan converts protein thiol to disulfide without affecting other protein groups. Hopkins and co-workers66 found alloxan at low concentration could inhibit succinic dehydrogenase. Hexosediphosphatase inhibition by alloxan.could not be reversed by cysteine, suggesting that the reaction is not an oxidative one. 142 As with many other thiol reagents already discussed there is no question that alloxan can and does react with tissue thiols. The critical question that is not yet sufficiently answered is this; does the diabetes produced in animals by alloxan result from reaction with tissue thiols? Lazarow76 gives a degree of plausibility to an affirmative answer with the followmg observations. Since (3-cells synthesize insulin, a disulfi'de protein, they may be expected to have a higher redo* potential than other tissues. This appears to be so from results of staining with Janus green. Since the p"-cells have a weaker reducing potential they probably contain less thiol, either as cell protein or as reduced glutathione. This would make them less capable of destroying alloxan than other cells with a greater thiol reserve. Or looking at the same matter in another way, the synthesis of insulin, containing 12 percent cystine, by the p-cells is likely to compete with glutathione synthesis, so these cells are apt to be left with only a small glutathione reserve to protect them from alloxan. Summarizing the broad survey attempted, we find few sharp conclusions. Arsenic compounds, whether trivalent or pentavalent, are known to react with no tissue components other than thiols and the evidence that they do in fact react with tissue thiols is quite good. Of anti- monials we can say almost the same although some antimony compounds can also react with hydroxy acids. If there is less evidence that they do react with tissue thiols this is largely because much less work has been done with antimonials. Mercury compounds can equally readily
287 react with thiols and in many cases are known to react with tissue thiols. But since mercurials could also react with disulfides and a great variety of amino compounds we are here on less certain ground. Tetrathionate and the disulfide antibiotics are not certainly known to react with anything but thiols. With the alkyl halides, unsaturated carbonyl antibiotics, quinones and alloxan we are on altogether uncertain ground and no generalization is at present likely to be good except that they are all capable of reacting with thiols. In some individual cases there is a fair amount of evidence that they do produce their effects by reaction with tissue thiols. In a few cases, notably the sulfur and nitrogen mustards, there is good evidence that at least some of their effects are unrelated to thiols. For most of these compounds the great need is for more evidence.
288 BIBLIOGRAPHY 1. Anderson, H. H. , Hansen, E. L. , Sah, P. P. T. and Cafiso, J. R. , J. Pharm. Exptl. Ther., 91,112 (1947). 2. Anson. M. L. , J. Gen. Physiol. , 24, 399 (1941). 3. Anson, M. L. and Stanley, W. M. , J. Gen. Physiol., 24, 679(1941). 4. Armstrong, W. D. , Spink, W. W. and Kahnke. J. , Proc. Soc. Exptl. Biol. Med. , 53, 230 (1942). 5. Bacq, Z.M., Experientia, 2, 349, 385(1946). 6. Bailey, J. H. and Cavallito, C. J. , J. Bact. , 55, 175(1948). 1 Bailey, K. and Perry. S. V. , Biochim. Biophys. Acta, 1, 506(1947). 8. Banks, T. E. , Boursnell, J. C. , Francis, G. E. , Hopwood, F. L. and Wormall, A., Biochem. J. , 40, 734 (1946). 9. Banks, T. E. , Boursnell. J. C. , Francis, G. E. , Hopwood, F. L. and WormaU, A., Biochem. J. , 40, 745 (1946). 10. Barnett, E. B. , Cook, J. W. and Peck. W. C. , J. Chem. Soc., 1035(1924). 11. Barron, E. S. G. and Kalnitsky, G. , Biochem. J. , 41, 346(1947). 12. Barron, E. S. G. , Miller, Z. B. , Bartlett, G. R. , Meyer, J. and Singer, T. P. , Biochem. J. , 4J, 69 (1947). 13. Barron, E. S. G. and Singer, T. P. . J. Biol. Chem., 157. 221 (1945). 14. Binet, L. and Wellers, G. , Bull. Soc. Chim. Biol. , 28, 751 (1946). 15. Bloom, N. , Forbes, G. and Policoff, L. , Proc. Soc. Exptl. Biol. Med., 72, 207(1949). 16. Bondi, A., Dietz, C. C. and Spaulding, E. H. , Science, 103. 399(1946). 17. Bougault, J. and Chabrier, P. , Compt. Rend. 224. 395 (1947). 18. Boursnell, J. C. , Francis, G. E. and Wormall, A. , Biochem. J. . 40. 737 (1946). 19. Braun, H. A. , Lusky, L. M. and Calvery, H. O. , J. Pharm. Exptl. Ther. . Supt. 87, 119 (1946). 20. Broderson, R. and Kjaer, A. , Acta Pharmacol. Toxicol. , 2, 109 (1946). 21. Bueding, E. , Physiol. Rev. 29., 195(1949). 22. Cavallito, C. J. , J. Biol. Chem. 164. 29(1946). 23. Cavallito, C. J. and Bailey, J. H. , Science, 100, 390(1944). 24. Cavallito, C.J., Bailey, J. H. and Warner, W. F. , J. Am. Chem. Soc., 68, 715(1946). 25. Cavallito. C. J. andHaskell, T. H. , J. Am. Chem. Soc., 67, 1991 (1945).
289 Ik. Chen. G. and Geiling, E. M. K. . J. Infect. Dis., 2Â£, 152(1945). 27. Clarke. D. H. and Theiler, M. . J. Infect. Dis., 82, 138(1948). 28. Clarke. H T , J. Biol. Chem. , 97. 235 (1936). 29. Colwell, C.A. and McCall, M. , J. Bact. , 51, 659 (1946). 30. Cunneen, J. I. , J. Chem. Soc. , 134(1947). 31. Dam, H. , Vitamins and Hormones, 6, 27 (1948). 32. De Graff, AC. and Lehman, R. A. . J. Am. Med. Assoc. , 119, 998 (1942). 33. Desreux, V., Fredericq, E. and Fischer, P., Bull. Soc. Chim. Biol., 28, 493(1946). 34. Dufrenoy, J. and Pratt. R. , J. Bact. , 54, 283 (1947). 35. Farah, A. and Maresh, G. , J. Pharm. Exptl. Ther. , 9J. 73(1948). 36. Fildes, P. , Brit. J. Exptl. Path. , 21, 67 (1940). 37. Fischer, E. , Science, 106. 146 (1947). 38. Fischer, P., Bull. Soc. Chim. Biol.. 2JJ. 240(1946). 39. Fischer. P. and Gof fart, M. . Compt. Rend. Soc. Biol., 141. 527(1947). 40. Foote, M. W. , Little, J. E. and Sproston, T. J. . J. Biol. Chem , 181 . 481 (4949). 41. Franz, G. , Arch. Exptl. Path. Pharm., 186, 661 (1937). 42. Fredericq, E. , Bull. Soc. Chim. Belg. . 54, 265 (1945). 43. Gale. E. F. and Taylor, E.S. , J. Gen. Microbiol. . 1, 314(1947). 44. Gammill, J.F., Southam, C. M. .and Van Dyke, H. B. . Proc. Soc. Exptl. Biol. Med., 64, 13 (1947). 45. Geiger, W. B. , Arch. Biochem. , \Â±. 23(1946). 46. Geiger, W. B. , Arch. Biochem., 16, 423(1948). 47. Geiger. W. B. and Conn, J. E. . J. Am. Chem. Soc., 67, 112(1945). 48. Geiger. W. B. , Green. S. R. and Waksman, S. A. , Proc. Soc. Exptl. Biol. Med. , 6_1 , 187 (1946). 49. Gellhorn, A. and Van Dyke, H. B. , J. Pharm. Exptl. Ther. , 88. 162 (1946). 50. Gilman, A. . Philips, F. S. , Allen, R. P. and Koelle, E.S. , J. Pharm. Exptl. Ther. , Supt. 87. 85 (1946). 51. Gilman, A. , Philips, F. S. , Koelle, E.S. . Allen, R. P. and St. John, E. , Am. J. Physiol. 147, 115 (1946). 52. Godeaux, J. , Bull. Soc. Roy. Sci. Liege, 1_3, 219 (1944). 53. Godeaux, J. , Compt. Rend. Soc. Biol.. 140, 675(1946). 54 Godeaux, J. , Compt. Rend. Soc. Biol. , 140, 678 (1946). 55. Goffard, M. . Arch. Intern. Pharmacodyn. , 71. 9 (1947).
290 56. Goffart, M. and Fischer. P., Compt. Rend. Soc. Biol. , 141, 528 (1947). 57. Golumbic, C. , Fruton, J. S. and Bergmann, M. , J. Org. Chem. . j_l , 518 (1946). 58. Gulland, J. M. , Biochem. J. , 26, 32 (1932). 59 Handley, C. A. and LaForge, M. , Proc. Soc. Exptl. Biol. Med. , 65, 74 (1947). 60. Harris, P. N. , J. Pharm. Exptl. Ther. , 82, 254 (1944). 61. Helfrich, O B. and Reid, E. E. . J. Am. Chem. Soc., 42. 1208 (1920). 62. Hellerman, L. . Chinard. F. P. and Deitz, V. R. , J. Biol. Chem., 147, 443(1943). 63. Hellstrom, N. . Ark. Kemi Mineral. Geol. , 13A. Â»6 (1939). 64. Henry. H. , Stacey, M. andTeece, E.G., Nature. 156. 720(1945). 65. Hoffman-Ostenhof. O. , Science, 105. 549 (1947). 66. Hopkins, F. G. , Morgan, E. J. and Lutwak-Mann, C. . Biochem. J. , 3_2. 1829(1938). 67. Jackson, K. E. , Chem. Rev., 14, 251 (1934). 68. Jeener, R. , Experientia. 3.. 243 (1947). 69. Kinsey. V. E. and Grant, W. M. , J. Cell. Comp. Physiol. , 29, 289(1947). 70. Koschotojanz, C. S. and Turpajew, T. M. , Nature. 158. 837(1946). 71. Kucharik. J. and Telbisz, A., Arch. Intern. Pharmacodyn. , 75. 157(1947). 72. Kuhn, R. and Beinert, H. . Ber. , 77, 606 (1944). 73. Launoy, L. , Bull. Soc. Chim. Biol. , 1_7. 1022 (1935). 74. Lavine. T. F. , J. Biol. Chem., 117. 309(1937). 75. Lazarow, A., Proc. Soc. Exptl. Biol. Med.. 61, 441 (1946). 76. Lazarow, A. , Physiol. Rev. , 2%, 48 (1949). 77. Lazarow, A., Patterson, J. W. and Levey, S. , Science, 108. 308(1948). 78. Lehman, R.A. , Proc. Soc. Exptl. Biol. Med.. 64, 428(1947). 79. Lieben. F. and Edel, E. , Biochem. Zeit. . 244. 403 (1932). 80. Lieberman, S. , Experientia, 2. 411 (1946). 81. Long, W.K. and Farah. A.. Science, 104, 220(1946). 82. Long, W.K. and Farah, A., J. Pharm. Exptl. Ther.. 88, 388(1946). 83. Longcope, W. T. and Luetscher, J. A. , Ann. Int. Med., 31, 545(1949). 84. Lyons, R.N. , Austral. J. Exptl. Biol. Med. Sci. , 23, 131 (1945). 85. Mackworth, J. F. , Biochem. J. . 42, 82 (1948). 86. McChesney, E. W. , Barlow. O. W. and Klinck, G. H. , J. Pharm. Exptl. Ther., 80. 81 (1944).
291 87. Michaelis, L. and Schubert, M. , J. Biol. Chem. , 106, 331 (1934). 88. Middlebrook, W. R. and Phillips, H. , Biochem. J. , 41, 218(1947). 89. Moore, S. , Stein, W. H. and Fruton, J. S. , J. Org. Chem., U 675 (1946). 90. Nicolet, B. H. , J. Am. Chem. Soc. , 53, 3066 (1931). 91. Oxford, A. E. , Chem. Ind. . 6_1, 189(1942). 92. Page, J.E. and Robinson, F. A. , Brit. J. Exptl. Path., 24, 89 (1943). 93. Palay. S. L. and Lazarow, A., Anat. Rec. , 96, 55(1946). 94. Patterson, J. W. , Lazarow, A. and Levey, S. , J. Biol. Chem. , 177, 197 (1949). 95. Peters, R. A. and Stocken, L.A. , Biochem. J. , 41, 53(1947). 96. Peters, R.A. , Stocken, L. A. , Thompson, R. H. S. . Woodward, F. N. , Millidge, A. F. and Gasson, E. J. , U.S. Pat. 2,432, 797 (1947). 97. Peters, R.A. and Wakelin, R. W. , Biochem. J. , 4J, 545 (1947). 98. Peters, R.A. and Wakelin, R. W. , Biochem. J. , 4J, 550 (1947). 99. Peters, R.A. and Wakelin, R. W. , Biochem. J. , 4JL, 555 (1947). 100. Philips, F.S. , Gilman, A., Koelle, E. S. and Allen, R. , J. Biol. Chem., 167. 209(1947). 101. Pirie, A., Biochem. J..4J, 185(1947). 102. Polis, B. D. and Meyerhof, O. , J. Biol. Chem., 169, 389(1947). 103. Pratt, R. andDufrenoy, J. , Bact. Rev., 12, 79(1948). 104. Pratt, R. andDufrenoy, J. , J. Am. Chem. Soc., 70, 1671 (1948). 105. Preisler, P. W. and Preisler, D. B. , J. Biol. Chem., 89, 631 (1930). 106. Purr, A.. Biochem. J. , 2J, 5 (1935). 107. Purr, A. , Biochem. J. , 29., 13 (1935). 108. Ratner, S. and Clarke, H. T. , J. Am. Chem. Soc., 52, 200 (1937). 109. Rinderknecht, H. , Ward, J. L. , Bergel, F. and Morrison, A. L. , Biochem. J. , 4J, 463 (1947). 110. Rosner, L. , J. Biol. Chem. , 132, 657 (1940). 111. Sandground, J. H. , Proc. Soc. Exptl. Biol. Med. , 52, 188(1943). 112. Sandground, J. H. , J. Pharm. Exptl. Ther. , 80, 393(1944). 113. Schreiber, F. G. and Schubert, M. , J. Parasitol. , IS, 364(1949). 114. Schubert, M. , J. Am. Chem. Soc. , 54, 4077 (1932). 115. Schubert. M. , J. Am. Chem. Soc. , 15, 4563 (1933). 116. Schubert, M. . J. Biol. Chem., 11I. 671 (1935).
292 117. Schubert, M. . J. Biol. Chem., 114, 341 (1936). 118. Schubert, M. , J. Biol. Chem. , 116. 437 (1936). 119. Schubert, M. . J. Biol. Chem.. 121, 539(1937). 120. Schubert. M. , J. Biol. Chem., 130. 601 (1939). 121. Schubert, M. . J. Am. Chem. Soc. , 69, 712 (1947). 122. Schubert, M. , Am. J. Trop. Med. , 28. 137(1948). 123. Schubert, M. , Am. J. Trop. Med., In press (1950). 1.'.4, Schubert, M. , Unpublished data. 125. Simon, F. P. . Potts, A.M. and Gerard, R. W. , Arch. Biochem. , 12, 283(1947). 126. Simon, R. D. , Brit. J. Exptl. Path., 2_2, 202 (1948). 127. Simonsen, D. G. , J. Biol. Chem., 94, 323(1934). 128. Small. L. D. , Bailey, J. H. and Cavallito, C. J. . J. Am. Chem. Soc., 63, 1710(1947). 129. Snell, J. M. and Weissberger, A., J. Am. Chem. Soc., 6.1. 450(1939). 130. Stahmann, M. A. , Golumbic, C. , Stein, W. H. and Fruton, J. S. , J. Org. Chem., 1_1, 719 (1S46). 131. Stein, W. H. and Moore, S. , J. Org. Chem. , 1_1 , 681 (1946). 132. Stein, W. H. , Moore, S. and Bergmann, M. . J. Org. Chem.. JJ, 664(1946). 133. Steinkopf, W. and Muller, S. , Ber. , Â£6, 1926 (1923). 134. Stocken, L. A. and Thompson. R. H. S. . Biochem. J. , 40, 529(1946). 135. Stocken, L. A. and Thompson, R.H.S. , Biochem. J. , 40, 535(1946). 136. Sturm. E. , U.S. Pat. 2, 1 72, 71 7 and 2, 180, 262 (1940). 137. Sussman, R. M. and Schack, J. A. , Proc. Soc. Exptl. Biol. Med.. 6^, 247(1947). 138. Ter Horst, VI. P. and Felix, E. L. , Ind. Eng. Chem., 35., 1255(1943). 139. Thomas, G. W. and Cook, E. S. , J. Bact. . 54, 527 (1947). 140. Thompson, R.H.S. and Whittaker, V. P. , Biochem. J. , 4_1, 342(1947). 141. Van Dolah, R. W. and Christenson, G. L. , Arch. Biochem., 1.2, 7(1947). 142. Walsh, E.G. and Walsh, G. , Nature, 161. 976 (1948). 143. Williamson, J. andLourie, E. M. , Ann. Trop. Med. Parasitol. , 40, 255(1946). 144. Wulfing, J.A. . French Pat. 823,824(1938).
293 DISCUSSION DR. HARRIS L. FRIEDMAN (Lakeside Laboratories, Milwaukee. Wisconsin): I would like to make a few comments, very briefly since the time is short, as to the role of thiol groups and the effects of mercurials. We have been very interested in this, as have many others. There are many conflicting results which need to be explained if the mechanism by inhibition of thiol group is correct. Handleyi/of Baylor has recently shown that mercurial diuretics do inhibit the succinic dehydrogenase enzyme which occurs in the kidney, and there is enough mercury ion in the kidney to account for the inhibition; but still, as Dr. Schubert has pointed out, it cannot be deduced that this is what actually happens. Recently, also, Ruskin2/and co-workers have shown that vitamin C can detoxify the action of a mercurial on the heart just as well as BAL or thioglycolic acid. I would like to ask Dr. Schubert, in this case, how that can be tied in with the sulfhydryl groups. i/C.A. Handley. Fed. Proc. 8, 299 (1949) & A. Ruskin, J. E. Johnson, Proc. Soc. Exp. Biol. Med. , 72, 577(1949) DR. SCHUBERT: 1 am glad to hear about that ascorbic acid effect. The only answer I can give is another case, in which antimony compounds have been detoxified also with ascorbic acid and, in fact, with a large number of substitute benzolc acids. Ascorbic acid, incidentally, could react with thiols. 1 once studied that a little bit, and the only thing I could find in that particular case was that oxidized ascorbic acid very readily oxidized thiols. Reduced, on the other hand, it is not very likely to interact with thiols, but it could interact with disulfides, and there is some work by Purr in which he tries to show that disulfides are reduced by a mixture of ferrous iron and ascorbic acid. That is the only inter- action I know of between those two. . Of course, the other thing which is known is that thiols generally protect ascorbic acid against oxidation. That is a common effect of many of these so-called sluggishly acting reduct- ants. You get the same sort of thing with hydroquinone. It can also inhibit oxidation of sulfur compounds. DR. F. S. PHILIPS (Sloan-Kettering Institute, New York, New York): I was pleased to hear Dr. Schubert's discussion, because I have felt for some time that there are many questions to be answered before one can attribute the mecrianism of action of a compound to sulfhydryl inactivation. Everything Dr. Schubert has said has been explicit, but I should like to emphasize some other points of view than those developed in his presentation. The case of the nitrogen and sulfur mustards should be emphasized since there is an interest in determining the common factors in radiation poisoning, and the action of the p-chlor- alkylamines and sulfides. Biologically, the physical agents, i.e. penetrating radiations, and the above chemical agents have many properties in common. There is a growing belief among some workers that radiant energy may cause cellular damage through reaction, which involve sulfhydryl groups. It would, therefore, be of importance if it could be established that the physiological mechanism of action of mustards also involves reaction with sulfhydryl groups. However, it should be stressed that mustards probably do not act biologically through inactivation of sulfhydryl groups. I take as authority for this statement. Dr. Schubert himself, and among others, R. A. Peters, who has stated quite definitely that, though mustards may react with simple sulfhydryl derivatives their reaction with sulfhydryl in complex compounds, such as proteins, is not very striking, and much less than that which occurs with carboxyl, amino, or phosphoryl groups. There is an aspect to the understanding of the properties of agents which react with sulfhydryl groups, which is often not considered by investigators who hypothecate sulfhydryl inactivation as a mechanism of biological action. I believe Dr. Schubert raised the same point.
E94 One may well ask why, for example, alloxan should be the only sulfhydryl reactive reagent which damages the p-cells of the pancreas, whereas other sulfhydryl oxidizing agents do not damage this particular cell; indeed, very few other agents have been found to affect the (J-cell. Moreover, alloxan has as another major site of action the proximal tubule of the kidney. It is likely that both the proximal tubule of the kidney and the p-cell of the pancreas are of about equal sensitivity to the agent. Nevertheless, tetrathionate, a powerful sulfhydryl oxidant, does not touch the p-cell of the pancreas while it causes extensive, alloxan-like damage in proximal tubules. Other examples of the role of cellular distribution in the mechanism of action of sulfhydryl reactive agents may be stated. When ionic cadmium is injected into an animal, its actions on the proximal tubule of the kidney are not striking; its toxicity would suggest other sites of action which, at present, are not well understood. Nevertheless, if cadmium is introduced in the form of a cadmiumdibal complex, it exerts a selective toxic action in this region of the kidney. One might also ask why iodo acetate and a number of other "sulfhydryl inhibitors" do not attack this particular region of the kidney to the same extent. It appears, then, that an important feature in understanding the mechanism of action of sulfhydryl reactive agents concerns not merely their affect on enzymes in cells after having entered certain cells, but also the factors which determine specificity of cellular distribution and cellular sites of action. DR. SCHUBERT: I will just take a second; I think there are some other cases. I think the discussion Dr. Philips gave is very much to the point, and I am quite in agree- ment with everything he said. I think there are two things which need to be considered. One, of course, is the thing which was discussed yesterday, which is the matter of cell permeability, and this is the thing which Oilman discusses in connection with the high toxicity of cadmium-BAL complexes, and the fact that BA1. does not protect against cadmium poisoning. I would just like to point out one additional example of that character, and that is that, among antimony compounds, we found one of the most effective for getting rid of the Schistosoma worm was the antimony-BAL, compound itself. If BAL. is given with antimony, and can take the antimony away from enzyme systems, you are going to have a little difficulty explaining why that should be better for killing schistosome worms than a compound in which you did not have the BAL attached to the antimony at all. DR. CHALMERS L. GEMMILL (University of Virginia. Charlottesville, Virginia): I would just like to ask Dr. Schubert if he could give us some help with the recent observation!/ we have made. The inhibition of invertase by mercuric chloride, can be reversed with BAL or increased with cysteine; in other words, you can work in both ways depending on whether a di- thiol or a monothiol is used. I would also like to state that ascorbic acid has been described as a reactivator of the mercuric chloride inhibited invertase system. -L/Gemmill, C. L. and E. Bowman, J. Pharm. Exp. Therap. In press, (1950). DR. SCHUBERT: I do not believe I can say anything very intelligent about that case; I did not know about that one, either. I am glad to know about it, but that is about all I can say. I think, of course, that invertase is one of the enzymes which, presumably, contain no essential thiols. Just by way of emphasizing this point that these metals can combine with groups other than thiols, antimony, as far as we now know, combines only with sulfur compounds; yet, recently, we had occasion, when working with collagen, to find that collagen is able to attach to itself at neutral reaction quite large amounts of antimony. So, possibly, invertase can combine, through groups other than thiol groups, with the mercury. Certainly it is likely that occurs, but why you should get the activation with the glutathione and not with BAL that is not clear.