Food Additives & Contaminants
ISSN: 0265-203X (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/tfac19
Nucleophilic reactions of sorbic acid G. D. Khandelwal & B. L. Wedzicha To cite this article: G. D. Khandelwal & B. L. Wedzicha (1990) Nucleophilic reactions of sorbic acid, Food Additives & Contaminants, 7:5, 685-694, DOI: 10.1080/02652039009373934 To link to this article: http://dx.doi.org/10.1080/02652039009373934
Published online: 10 Jan 2009.
Submit your article to this journal
Article views: 22
View related articles
Citing articles: 6 View citing articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tfac20 Download by: [University of Cambridge]
Date: 05 November 2015, At: 19:30
FOOD ADDITIVES AND CONTAMINANTS, 1990, VOL. 7, NO. 5, 6 8 5 - 6 9 4
Symposium Paper Nucleophilic reactions of sorbic acid G. D. KHANDELWAL and B. L. WEDZICHA Procter Department of Food Science, University of Leeds, Leeds, LS2 9JT, UK
Downloaded by [University of Cambridge] at 19:30 05 November 2015
(Received 9 February 1990; accepted 1 March 1990) The conjugated dienoic acid structure of sorbic acid renders it susceptible to nucleophilic attack. Nucleophiles known to react with sorbic acid include sulphite ion and amines. These attack the molecule in position 5 and, in the cse of amines, cyclization to form substituted dihydropyridones may follow. Recent investigations show that thiols in general can also add to sorbic acid. Cysteine, for example, reacts slowly with sorbic acid at 80°C and pH 5.5, leading to 5-substituted 3-hexenoic acid. In general, reaction products are difficult to isolate from aqueous reaction mixtures as they are susceptible to acid- and basecatalysed hydrolysis. A synthesis of model compounds may be carried out by reaction of sorbate esters with the appropriate thiol (or its ester if it is an acid) in the presence of the corresponding sodium alkoxide. It is interesting that alkyl thiols give di-adducts with sorbate ester whilst low molecular weight thiols containing an oxygen atom give a monoadduct. The mechanism of this reaction and its implications to the preparation of samples for toxicological evaluation are discussed. The reaction of sorbic acid with nitrite ion is unusual and its mechanism is considered.
Introduction Sorbic acid, the trans-trans form of 2,4-hexadienoic acid, has the following structure: CH3
H C=C
H
H
C=C / H
\ COOH
Proton and 13C NMR data give chemical shifts at the carbon atoms of the conjugated diene, as follows (Leraux and Vauthier 1970, Reger and Habib 1980): Carbon atom H chemical shift/ppm 13 C chemical shift/ppm l
2 5- 81 118- 1
3 7- 41 147- 2
4 6- 25 129- 6
5 6- 25 140- 6
indicating the lowest electron density to be associated with position 3. Nucleophilic attack is therefore most likely in this position. The simplest example of nucleophilic addition to a diene is the reaction of butadiene with H + Nu" where Nu" represents a nucleophile. This involves the initial protonation of the diene at Ci and subsequent attack by the nucleophile to 0265-203X/90 $3.00 © 1990 Taylor & Francis Ltd.
686
G. D. Khandelwal and B. L. Wedzicha
give the 1,2- or 1,4-addition product as follows: Nu"
CH3-CH-CH=CH2
CH2-CH-CH=CH2 »
Nu (1,2-addition)
I
Downloaded by [University of Cambridge] at 19:30 05 November 2015
Nu"
CH2-CH=CH-CH2 I H II
CH3-CH=CH-CH2 Nu
(1,4-addition)
Protonation at Ci is preferred because, unlike protonation at C2, it yields a secondary carbonium ion stabilized by delocalization of the positive charge (Sykes 1965). The final product depends on the relative rates of reaction of the nucleophile with the possible carbonium ions I and II. With sorbic acid we have a polarized double-bond system. A simpler example is the a,/3-unsaturated acid, crotonic acid, which may add H+Nu~ by a 1,4mechanism if the nucleophile is weak, i.e.: CH3CH—CH C
-OH
/OH CH 3 CH-CH=C X I OH Nu
CH 3 CH=CH-C,
Nu"
CH3CH-CH=C /
OH
OH
CH 3 CH-CH 2 -C I Nu On the other hand, strong nucleophiles such as amines and thiols will give 1,4addition products without the need for the initial protonation, the addition of H + taking place as the final stage in the mechanism. Extending the argument to sorbic acid, it is anticipated that a strong nucleophile will attack the molecule in positions 3 and 5 as follows: OH
The much greater extent to which the charge on the intermediate arising from attack at position 5 is delocalized, suggests that this should be preferred despite
Nucleophilic reactions of sorbic acid
687
lower electron density at position 3 of sorbic acid. The reaction is completed by addition of H + to positions 2 or 4:
OH
OH
Downloaded by [University of Cambridge] at 19:30 05 November 2015
III
IV
It is likely that which product is formed will depend on the relative stabilities of resonance forms III and IV and their relative rates of reaction with H + . It is seen that the two products are, respectively, the result of 1,2- and 1,4-addition across the diene part of the molecule. As is known for butadiene, the tendency for nucleophiles to add 1,2- and 1,4- across the diene may depend also on the polarity of the solvent; for example 1,2-addition to butadiene tends to occur preferentially at lower temperatures in non-polar solvents (Sykes 1965). A consequence of 1,2addition to the diene of sorbic acid is that the molecule is then capable of adding a second molecule of nucleophile to form a diadduct, as follows:
OH
OH
Nu
OH Nu
Nu
O
Foods contain numerous nucleophiles (amines, thiols), and sulphite ion, which is present in foods preserved with 'sulphur dioxide' is known for its nucleophilic reactivity (Wedzicha 1984). It is therefore of interest, regarding food uses of sorbic acid, to consider the nature of reaction products when sorbic acid undergoes reaction with nucleophiles. An additional and unusual reaction is that of sorbic acid with nitrite ion; this will also be reviewed.
Reaction with sulphite ion
The reaction between sorbic acid and sulphur(iv) oxospecies is well documented. Hagglund and Ringbom (1926) suggest that it proceeds as a result of
688
G. D. Khandelwal and B. L. Wedzicha
4,5-addition of HSO3~ to give the following product V: CH3-CH-CH2-CH=CH-COOH SO3~
Downloaded by [University of Cambridge] at 19:30 05 November 2015
V Heintze (1976) reports that when solutions containing l-8mM sorbic acid and l-6mMHSO3~ at pH3-5 are left at room temperature for 65-100 days, a 1:1 addition product is formed in approximately 60% yield. However, when such reaction mixtures are analysed for sulphur(iv) species, by a distillation method, all the added sulphur(iv) is recovered. The adduct appears to be labile. Hagglund and Ringbom (1926) also report that crotonic acid reacts with HSO3~, and it is reasonable to expect the product shown above to add a second sulphite ion. The fact that sorbic acid forms only a 1:1 adduct is possibly indicative of the reaction with HSO3- proceeding by 1,4-addition across the diene, rather than the 1,2-addition required for compound V. Reaction with amines Sorbic acid reacts with ammonia, alkylamines, aromatic amines and benzylamine at high temperature (e.g. 150-200°C) to form allegedly the amino acid VI (Fischer and Schlotterbeck 1904), and dihydropyridones VII and VIII (Mosher 1950, Shamma and Rosenstock 1961, Verbiscar and Campbell 1964, Kheddis etal. 1981). Intermediate IX has also been identified (Verbiscar and Campbell 1964, Kheddis etal. 1981). OH
NHR (or Ar) VI
NHR (or Ar)
O
R (or Ar) VIII
I
where R = H or alkyl Compound VIII isomerises to VII at high temperature. The dihydropyridones are seen to be lactams (internal amides) which could be easily formed by cyclization of compounds such as the diadduct X, formed between sorbic acid and the amine, as follows (Verbiscar and Campbell 1964, Kheddis etal. 1981):
Nucleophilic reactions of sorbic acid
689
Me
Me
NHR V-NHR
-> +NHR
NHR
-H* -OH-
-• XI
Downloaded by [University of Cambridge] at 19:30 05 November 2015
© O OH The product then eliminates a molecule of amine to form the dihydropyridones, with structure VII being more thermodynamically stable. It is conceivable that the dihydropyridones could also be formed by cyclization of monoadducts with the double-bond in positions 2 and 3 to give VII and VIII, respectively, but this is considered unlikely (Kheddis et al. 1981). The double-bond would need to be cis. It is important to give consideration to sorbic arid-amino compound adducts in relation to food uses of sorbic acid because in some situations (e.g. grilling of cheese) the preservative may be exposed to severe heat treatment. A large number of 3,4-unsaturated lactones have also been found to be physiologically active (Haynes 1948). There is, however, no evidence to suggest that sorbic acid reacts with simple amines such as butylamine or glycine even after prolonged heating at 100°C in water at pH 5-7 (Wedzicha and Brook 1989). Reaction with nitrite ion
In a series of papers, Namiki and co-workers (Kito and Namiki 1978, Osawa et al. 1979, Osawa and Namiki 1982) reported that the reaction of sorbic acid with nitrite ion produced ethyl nitrolic acid XI and 1,4-dinitro-2-methylpyrrole XII as the main reaction products. Another product XIII containing a furoxan ring has been isolated. The overall scheme of the reaction is illustrated in figure 1.
OH NO 2
XII
XI
XIII
Figure 1. Products formed from the reaction of nitrite ion with sorbic acid (Osawa and Namiki 1982).
690
G. D. Khandelwal and B. L. Wedzicha
Downloaded by [University of Cambridge] at 19:30 05 November 2015
The identification of the reaction products was based on the usual spectroscopic data and X-ray crystallography, but no detailed mechanism for the formation of ethyl nitrolic acid and 1,4-dinitro-2-methylpyrrole has been proposed. It is regarded that the principal nitrite-derived reactant is N2O3. It is well established (Kooyman et al. 1951) that N2O3 may be polarized as NO2+NO~ for electrophilic nitration and as NO+NC>2~ for nucleophilic nitration. Such polarized species are known to add across polarized double bonds (Park and Williams 1976). Evidence of the initial product A of addition of N2O3 to sorbic acid and its conversion to the oxime B is the isolation of XIII (Osawa et al. 1979). The only way we can conceive the formation of structures which might eventually degrade to a nitrolic acid-type structure is through a Beckmann rearrangement of B to give the following type of intermediate:
where X could be NO2.
Reaction with thiols Experimental approach Sorbic acid reacts slowly with thiols in aqueous solution at 80° C and pH3*7-5'7 (Wedzicha and Brook 1989). Kinetic measurements have shown (Wedzicha and Zeb 1990) the reaction of sorbic acid with mercaptoethanol, mercaptoacetic acid, cysteine and glutathione to be of second order. The reaction product was believed (Wedzicha and Brook 1989) to be the result of 1,4-addition across the diene, with the double-bond in position 3 of the sorbic acid chain, but no systematic study of the structure of sorbic acid-thiol reaction products had been carried out. The main reason for the lack of structural evidence is that it is very difficult to isolate the adducts, and there is a tendency for decomposition to take place yielding sorbic acid in high yield, even under mildly acid or alkaline conditions. We (Khandelwal and Wedzicha 1990) approached the problem by synthesizing methyl and ethyl esters of the adducts in question by nucleophilic addition of thiolate anions to the ester of sorbic acid. Such addition was carried out in the corresponding alcohol in the presence of an appropriate sodium alkoxide or triethylamine, as catalyst. The products were easily purified by distillation. The ultimate purpose was to prepare sufficiently pure samples for structural analysis using a reaction that gave the derivative of the same product as when the reactions of sorbic acid and the corresponding thiol were carried out in water. Ethyl sorbate and cysteine ethyl ester hydrochloride were suspended in dry ethanol, and sodium ethoxide was added. After 48 h stirring at room temperature the reaction mixture was worked up and product XIV isolated in good yield. Simi-
Nucleophilic reactions of sorbic acid
691
larly, methyl sorbate reacted with cysteine ethyl ester hydrochloride in dichloromethane in the presence of triethylamine. If, on the other hand, sorbic acid is allowed to react with cysteine in aqueous media and the product is esterified, it is identical to that formed from esters of sorbic acid and cysteine.
OEt
Downloaded by [University of Cambridge] at 19:30 05 November 2015
SR'
SR'
O
OEt
XV
XIV
The work was further extended to include ethyl-2-mercaptoacetate, mercaptoethanol and methyl-2-mercaptoacetate under similar conditions, and reactions always gave the monoadducts with the same structure as XIV, with the corresponding thiol attached to position 5. The work also included the reaction of alkyl thiols (ethanethiol, 2-methyl-2-propanethiol, butanethiol) with sorbate esters, in which case only di-adducts XV were isolated. No mono-adducts could be found even when the molar ratio of thiol to sorbate was 1:1 or less. Thus, it is seen that nucleophilic attack by alkyl thiols leads to the formation of di-adducts whilst reaction with mercaptoethanol, esters of 2-mercaptoacetic acid and cysteine ethyl ester hydrochloride give mono-adducts. Mechanism A mechanism for the attack on sorbate ester by the thiolate anion is given in figure 2. The anionic intermediate combines with H + to form either B or C.
SR'
OR
TS
SR'
H
H (B)
O
SR'
O
OR
OR
B (C)
Figure 2. Mechanism for the addition of thiolate anion to sorbate esters and possible rearrangement of the 3-ene product to a 2-ene product. The transition state is labelled TS and the two sets of arrows show electron shifts for the addition of H + to positions 2 and 4.
692
G. D. Khandelwal and B. L. Wedzicha
Downloaded by [University of Cambridge] at 19:30 05 November 2015
Product B is required if a di-adduct is to be formed. One question is whether, in the presence of a strong base such as sodium alkoxide, C may be converted to B; an acid/base-catalysed mechanism for isomerization is suggested in figure 2 and the driving force for the conversion could be the conjugated nature of the a,f3unsaturated ester so formed. Our results show that methyl-3-hexenoate is completely unreactive towards thiols under the same conditions as used for synthesis and, unless there is some participation from the substituent in position 5, the isomerization of the 3-ene to the 2-ene seems an unlikely possibility. When methyl 2-hexenoate is treated with sodium methoxide in the presence of thiol, product XVI is obtained and, in the absence of any nucleophile, product XVII is obtained. SR
O
XVI
OMe
O
XVII
An explanation for the mechanism is to say that the 3-hexenoate derivative is the thermodynamically favoured reaction product but the 2-hexenoate derivative may be formed as an initial kinetically favoured product. Thus, in the case of very reactive nucleophiles, e.g. alkylthiol, the formation of di-adduct is possible, whilst in the case of less reactive nucleophiles an irreversible isomerization to the more stable product predominates. The main difference between nucleophiles which cause mono-addition and those which cause di-adducts to be formed is the presence of oxygen in the former. It is hard to envisage that the electronegativity of the oxygen or the presence of lone pairs of electrons on this atom might be responsible for the creation of a field which causes repulsion of negative charge within the anionic transition state (TS in figure 2) away from position 4 to facilitate pick-up of a hydrogen ion at position 2. An intriguing question is, if given a range of oxygen-containing thiols of different hydrocarbon chain lengths, what is the length of chain before the thiol begins to behave as an alkyl thiol? Unfortunately, the longest chain mercaptocarboxylic acid available is 3-mercaptopropionic, because longer-chain acids tend to cyclize. The experiments have yet to be carried out with mercaptoalcohols of varying chain lengths, but it is clear that the size of the ester group (up to butyl) of the sorbate does not affect the stereochemistry of the addition reaction. The mono-adducts prepared do not add a second oxygen-containing thiol molecule, but a further question is whether they are able to add an alkyl thiol. The reaction of methyl sorbate-methyl thioglycollate reaction product with ethane thiol gave rise to product XVIII, whose 'H NMR spectrum was identical to that of an authentic sample reported by Blenderman and Joullie (1983). This indicates that displacement of the original nucleophile had taken place, and that ethanethiol had not added to the double-bond. To add even greater interest to the debate regarding the factors which control mono- and di-adduct formation, we have recently found that if sorbaldehyde is used in place of sorbates, the stereochemistry of addition is reversed, i.e. alkyl thiols react to form mono-adducts and the other thiols form diadducts. An intermediate situation exists with methyl hexadienyl ketone, where both types of adduct are obtained with both types of thiol.
Nucleophilic reactions of sorbic acid
693
SEt OMe XVIII
Downloaded by [University of Cambridge] at 19:30 05 November 2015
The detailed analysis of spectroscopic evidence and a full discussion of possible structural interpretations, which allowed the assignment of all structures presented here, is published elsewhere (Khandelwal and Wedzicha 1990). Hydrolysis of reaction products An important application of the synthetic work reported here is the preparation of samples for toxicological evaluation. For this to be achieved it is necessary to hydrolyse the esters to form the reaction products obtainable had sorbic acid itself been involved in the reactions. Attempts using acid and base hydrolysis of mono-adducts were unsuccessful because sorbic acid was obtained in high yield as one of the products. Currently we are examining the possibility of using esterase enzymes under neutral conditions to avoid breakdown of the 1:1 adducts. On the other hand, the desired alkyl thiol-sorbic acid reaction products were obtained when di-adducts were hydrolysed under normal conditions of ester hydrolysis. The instability of mono-adducts in acid or base solution has implications to their formation in foods. First, analysis of sorbic acid often involves steam distillation from acidified solution. The mono-adducts are decomposed under such conditions. Thus, the mono-addition reaction with thiols may not lead to a measurable loss of thiol; indeed, there may be no evidence that it may have occurred in a particular model or food system unless the investigator adapts the analytical procedure to allow for potential reversibility. However, the concentration of the available preservative may decrease as a result of such 'reversible' interactions. There may be a need for specifying free and reversibly bound sorbic acid content. It is interesting to note that the reaction of sorbic acid with sulphite ion is also reversible when mixtures are analysed for sulphur(iv) species (Heintze 1976). This may suggest that the structure of the reaction product should be shown with the double-bond in position 3, unlike structure V suggested by Hagglund and Ringbom (1926). In foods, possible reactions are between sorbic acid and cysteine residues of peptides and proteins, or thiols which are present as flavour components. In the case of reactions with proteins, subsequent proteolysis would release the mono-adduct of cysteine with sorbic acid, but it seems possible that the acid environment of the stomach could be conducive to the breakdown of this substance to sorbic acid and cysteine. The same could be true of the sorbic acid-sulphite ion adduct and, in both cases, the binding of sorbic acid might be of no toxicological significance. Acknowledgement
This work is being supported by the Ministry of Agriculture, Fisheries and Food. We are grateful to Dr D. J. McWeeny for his continuing interest. References BLENDERMAN, W. G., and JOULLIÉ, M. M., 1983, Lithium/ammonia reductions of 2-thiophenecarboxylic acids. Journal of Organic Chemistry, 42, 3206-3213.
Downloaded by [University of Cambridge] at 19:30 05 November 2015
694
G. D. Khandelwal and B. L. Wedzicha
FISCHER, E., and SCHLOTTERBECK, F., 1904, Verwandlung der Sorbinsäure in Aminosäuren. Berichte der Deutschen Chemische Gesellschaft,37,2357-2361. HÄGGLUND, E., and RINGBOM, A., 1926, ber die sulfitaddition an ungesättigte verbindungen. Zeitschrift für Anorganische und Allgemeine Chemie, 150, 231-253. HAYNES, L. J., 1948, Physiologically active unsaturated lactones. Quarterly Reviews (London), 2, 46-72. HEINTZE, K., 1976, Über die gegenseitige Beeinflussung von Sorbinsäure und schwefliger Säure. Die industrielle Obst- und Gem Üseverwet, 61, 555-556. KHANDELWAL, G. D., and WEDZICHA, B. L., 1990, Derivatives of sorbic acid-thiol adducts. Food Chemistry, 37, 159-169. KHEDDIS, B., BAHIBAH, D., HAMDI, M., and PÉRIÉ, J. J., 1981, Dihydropyridones-2 dérivées de l'acide sorbique: Synthèse et analyse conformationnelle. Bulletin de la Societé Chimique de France, 3-4(II), 135-140. KITO, Y., and NAMIKI, M., 1978, A new N-nitropyrrole. 1,4-dinitro-2-methylpyrrole, formed by the reaction of sorbic acid with sodium nitrite. Tetrahedron, 343, 505-508. KOOYMAN, E. C., FARENHORST, E., and WERNER, E. G. C., 1951, The nitrosation of methallylchloride. Recueil des Travaux Chimiques des Pays-Bas, 70, 689-695. LERAUX, Y., and VAUTHIER, E., 1970, Ètude par spectroscopie et moments dipolaires de diènes conjugués du type CH3CH=CH—CH=CH—Y. Comptes Rendus des Seances de l'Academie des Sciences, Sŕie C, 271, 1333-1336. MOSHER, H. S., 1950, Heterocyclic Compounds, Vol. 1, edited by R. C. Elderfield (New York: John Wiley & Sons), p. 651. OSAWA, T., and NAMIKI, M., 1982, Mutagen formation in the reaction of nitrite with food components analogous to sorbic acid. Agricultural and Biological Chemistry, 46, 2299-2304. OSAWA, T., KITO, Y., and NAMIKI, M., 1979, A new furoxan derivative and its precursors formed by
the reaction of sorbic acid with sodium nitrite. Tetrahedron Letters, 45, 4399-4402. PARK, J. R., and WILLIAMS, D. L. H., 1976, Reaction of dinitrogen trioxide with 2-methylpropene and other alkenes. Journal of the Chemical Society, Perkin Transactions, II, 828-832. REGER, D. L., and HABIB, M. M., 1980, Carbon-13 NMR investigation into the interaction of sodium sorbate and sorbic acid with various micelle-forming surfactants. Journal of Physical Chemistry, 84, 177-192. SHAMMA, M., and ROSENSTOCK, P. D., 1961, The synthesis and properties of some α,β-unsaturated valerolactams. Journal of Organic Chemistry, 26, 718-725. SYKES, P., 1965, A Guidebook to Mechanism in Organic Chemistry, 2nd edn (London: Longmans), pp. 149-155. VERBISCAR, A. J., and CAMPBELL, K. N., 1964, Unsaturated six-membered ring lactams. Journal of Organic Chemistry, 64, 2472-2474. WEDZICHA, B. L., 1984, Chemistry of Sulphur Dioxide in Food (London: Elsevier Applied Science). WEDZICHA, B. L., and BROOK, M. A., 1989, Reaction of sorbic acid with nucleophiles: preliminary studies. Food Chemistry, 31, 29-40. WEDZICHA, B. L., and ZEB, A., 1990, Kinetics of the reaction between sorbic acid and thiols. International Journal of Food Science and Technology, 25, 230-232.
ISSN: 0265-203X (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/tfac19
Nucleophilic reactions of sorbic acid G. D. Khandelwal & B. L. Wedzicha To cite this article: G. D. Khandelwal & B. L. Wedzicha (1990) Nucleophilic reactions of sorbic acid, Food Additives & Contaminants, 7:5, 685-694, DOI: 10.1080/02652039009373934 To link to this article: http://dx.doi.org/10.1080/02652039009373934
Published online: 10 Jan 2009.
Submit your article to this journal
Article views: 22
View related articles
Citing articles: 6 View citing articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tfac20 Download by: [University of Cambridge]
Date: 05 November 2015, At: 19:30
FOOD ADDITIVES AND CONTAMINANTS, 1990, VOL. 7, NO. 5, 6 8 5 - 6 9 4
Symposium Paper Nucleophilic reactions of sorbic acid G. D. KHANDELWAL and B. L. WEDZICHA Procter Department of Food Science, University of Leeds, Leeds, LS2 9JT, UK
Downloaded by [University of Cambridge] at 19:30 05 November 2015
(Received 9 February 1990; accepted 1 March 1990) The conjugated dienoic acid structure of sorbic acid renders it susceptible to nucleophilic attack. Nucleophiles known to react with sorbic acid include sulphite ion and amines. These attack the molecule in position 5 and, in the cse of amines, cyclization to form substituted dihydropyridones may follow. Recent investigations show that thiols in general can also add to sorbic acid. Cysteine, for example, reacts slowly with sorbic acid at 80°C and pH 5.5, leading to 5-substituted 3-hexenoic acid. In general, reaction products are difficult to isolate from aqueous reaction mixtures as they are susceptible to acid- and basecatalysed hydrolysis. A synthesis of model compounds may be carried out by reaction of sorbate esters with the appropriate thiol (or its ester if it is an acid) in the presence of the corresponding sodium alkoxide. It is interesting that alkyl thiols give di-adducts with sorbate ester whilst low molecular weight thiols containing an oxygen atom give a monoadduct. The mechanism of this reaction and its implications to the preparation of samples for toxicological evaluation are discussed. The reaction of sorbic acid with nitrite ion is unusual and its mechanism is considered.
Introduction Sorbic acid, the trans-trans form of 2,4-hexadienoic acid, has the following structure: CH3
H C=C
H
H
C=C / H
\ COOH
Proton and 13C NMR data give chemical shifts at the carbon atoms of the conjugated diene, as follows (Leraux and Vauthier 1970, Reger and Habib 1980): Carbon atom H chemical shift/ppm 13 C chemical shift/ppm l
2 5- 81 118- 1
3 7- 41 147- 2
4 6- 25 129- 6
5 6- 25 140- 6
indicating the lowest electron density to be associated with position 3. Nucleophilic attack is therefore most likely in this position. The simplest example of nucleophilic addition to a diene is the reaction of butadiene with H + Nu" where Nu" represents a nucleophile. This involves the initial protonation of the diene at Ci and subsequent attack by the nucleophile to 0265-203X/90 $3.00 © 1990 Taylor & Francis Ltd.
686
G. D. Khandelwal and B. L. Wedzicha
give the 1,2- or 1,4-addition product as follows: Nu"
CH3-CH-CH=CH2
CH2-CH-CH=CH2 »
Nu (1,2-addition)
I
Downloaded by [University of Cambridge] at 19:30 05 November 2015
Nu"
CH2-CH=CH-CH2 I H II
CH3-CH=CH-CH2 Nu
(1,4-addition)
Protonation at Ci is preferred because, unlike protonation at C2, it yields a secondary carbonium ion stabilized by delocalization of the positive charge (Sykes 1965). The final product depends on the relative rates of reaction of the nucleophile with the possible carbonium ions I and II. With sorbic acid we have a polarized double-bond system. A simpler example is the a,/3-unsaturated acid, crotonic acid, which may add H+Nu~ by a 1,4mechanism if the nucleophile is weak, i.e.: CH3CH—CH C
-OH
/OH CH 3 CH-CH=C X I OH Nu
CH 3 CH=CH-C,
Nu"
CH3CH-CH=C /
OH
OH
CH 3 CH-CH 2 -C I Nu On the other hand, strong nucleophiles such as amines and thiols will give 1,4addition products without the need for the initial protonation, the addition of H + taking place as the final stage in the mechanism. Extending the argument to sorbic acid, it is anticipated that a strong nucleophile will attack the molecule in positions 3 and 5 as follows: OH
The much greater extent to which the charge on the intermediate arising from attack at position 5 is delocalized, suggests that this should be preferred despite
Nucleophilic reactions of sorbic acid
687
lower electron density at position 3 of sorbic acid. The reaction is completed by addition of H + to positions 2 or 4:
OH
OH
Downloaded by [University of Cambridge] at 19:30 05 November 2015
III
IV
It is likely that which product is formed will depend on the relative stabilities of resonance forms III and IV and their relative rates of reaction with H + . It is seen that the two products are, respectively, the result of 1,2- and 1,4-addition across the diene part of the molecule. As is known for butadiene, the tendency for nucleophiles to add 1,2- and 1,4- across the diene may depend also on the polarity of the solvent; for example 1,2-addition to butadiene tends to occur preferentially at lower temperatures in non-polar solvents (Sykes 1965). A consequence of 1,2addition to the diene of sorbic acid is that the molecule is then capable of adding a second molecule of nucleophile to form a diadduct, as follows:
OH
OH
Nu
OH Nu
Nu
O
Foods contain numerous nucleophiles (amines, thiols), and sulphite ion, which is present in foods preserved with 'sulphur dioxide' is known for its nucleophilic reactivity (Wedzicha 1984). It is therefore of interest, regarding food uses of sorbic acid, to consider the nature of reaction products when sorbic acid undergoes reaction with nucleophiles. An additional and unusual reaction is that of sorbic acid with nitrite ion; this will also be reviewed.
Reaction with sulphite ion
The reaction between sorbic acid and sulphur(iv) oxospecies is well documented. Hagglund and Ringbom (1926) suggest that it proceeds as a result of
688
G. D. Khandelwal and B. L. Wedzicha
4,5-addition of HSO3~ to give the following product V: CH3-CH-CH2-CH=CH-COOH SO3~
Downloaded by [University of Cambridge] at 19:30 05 November 2015
V Heintze (1976) reports that when solutions containing l-8mM sorbic acid and l-6mMHSO3~ at pH3-5 are left at room temperature for 65-100 days, a 1:1 addition product is formed in approximately 60% yield. However, when such reaction mixtures are analysed for sulphur(iv) species, by a distillation method, all the added sulphur(iv) is recovered. The adduct appears to be labile. Hagglund and Ringbom (1926) also report that crotonic acid reacts with HSO3~, and it is reasonable to expect the product shown above to add a second sulphite ion. The fact that sorbic acid forms only a 1:1 adduct is possibly indicative of the reaction with HSO3- proceeding by 1,4-addition across the diene, rather than the 1,2-addition required for compound V. Reaction with amines Sorbic acid reacts with ammonia, alkylamines, aromatic amines and benzylamine at high temperature (e.g. 150-200°C) to form allegedly the amino acid VI (Fischer and Schlotterbeck 1904), and dihydropyridones VII and VIII (Mosher 1950, Shamma and Rosenstock 1961, Verbiscar and Campbell 1964, Kheddis etal. 1981). Intermediate IX has also been identified (Verbiscar and Campbell 1964, Kheddis etal. 1981). OH
NHR (or Ar) VI
NHR (or Ar)
O
R (or Ar) VIII
I
where R = H or alkyl Compound VIII isomerises to VII at high temperature. The dihydropyridones are seen to be lactams (internal amides) which could be easily formed by cyclization of compounds such as the diadduct X, formed between sorbic acid and the amine, as follows (Verbiscar and Campbell 1964, Kheddis etal. 1981):
Nucleophilic reactions of sorbic acid
689
Me
Me
NHR V-NHR
-> +NHR
NHR
-H* -OH-
-• XI
Downloaded by [University of Cambridge] at 19:30 05 November 2015
© O OH The product then eliminates a molecule of amine to form the dihydropyridones, with structure VII being more thermodynamically stable. It is conceivable that the dihydropyridones could also be formed by cyclization of monoadducts with the double-bond in positions 2 and 3 to give VII and VIII, respectively, but this is considered unlikely (Kheddis et al. 1981). The double-bond would need to be cis. It is important to give consideration to sorbic arid-amino compound adducts in relation to food uses of sorbic acid because in some situations (e.g. grilling of cheese) the preservative may be exposed to severe heat treatment. A large number of 3,4-unsaturated lactones have also been found to be physiologically active (Haynes 1948). There is, however, no evidence to suggest that sorbic acid reacts with simple amines such as butylamine or glycine even after prolonged heating at 100°C in water at pH 5-7 (Wedzicha and Brook 1989). Reaction with nitrite ion
In a series of papers, Namiki and co-workers (Kito and Namiki 1978, Osawa et al. 1979, Osawa and Namiki 1982) reported that the reaction of sorbic acid with nitrite ion produced ethyl nitrolic acid XI and 1,4-dinitro-2-methylpyrrole XII as the main reaction products. Another product XIII containing a furoxan ring has been isolated. The overall scheme of the reaction is illustrated in figure 1.
OH NO 2
XII
XI
XIII
Figure 1. Products formed from the reaction of nitrite ion with sorbic acid (Osawa and Namiki 1982).
690
G. D. Khandelwal and B. L. Wedzicha
Downloaded by [University of Cambridge] at 19:30 05 November 2015
The identification of the reaction products was based on the usual spectroscopic data and X-ray crystallography, but no detailed mechanism for the formation of ethyl nitrolic acid and 1,4-dinitro-2-methylpyrrole has been proposed. It is regarded that the principal nitrite-derived reactant is N2O3. It is well established (Kooyman et al. 1951) that N2O3 may be polarized as NO2+NO~ for electrophilic nitration and as NO+NC>2~ for nucleophilic nitration. Such polarized species are known to add across polarized double bonds (Park and Williams 1976). Evidence of the initial product A of addition of N2O3 to sorbic acid and its conversion to the oxime B is the isolation of XIII (Osawa et al. 1979). The only way we can conceive the formation of structures which might eventually degrade to a nitrolic acid-type structure is through a Beckmann rearrangement of B to give the following type of intermediate:
where X could be NO2.
Reaction with thiols Experimental approach Sorbic acid reacts slowly with thiols in aqueous solution at 80° C and pH3*7-5'7 (Wedzicha and Brook 1989). Kinetic measurements have shown (Wedzicha and Zeb 1990) the reaction of sorbic acid with mercaptoethanol, mercaptoacetic acid, cysteine and glutathione to be of second order. The reaction product was believed (Wedzicha and Brook 1989) to be the result of 1,4-addition across the diene, with the double-bond in position 3 of the sorbic acid chain, but no systematic study of the structure of sorbic acid-thiol reaction products had been carried out. The main reason for the lack of structural evidence is that it is very difficult to isolate the adducts, and there is a tendency for decomposition to take place yielding sorbic acid in high yield, even under mildly acid or alkaline conditions. We (Khandelwal and Wedzicha 1990) approached the problem by synthesizing methyl and ethyl esters of the adducts in question by nucleophilic addition of thiolate anions to the ester of sorbic acid. Such addition was carried out in the corresponding alcohol in the presence of an appropriate sodium alkoxide or triethylamine, as catalyst. The products were easily purified by distillation. The ultimate purpose was to prepare sufficiently pure samples for structural analysis using a reaction that gave the derivative of the same product as when the reactions of sorbic acid and the corresponding thiol were carried out in water. Ethyl sorbate and cysteine ethyl ester hydrochloride were suspended in dry ethanol, and sodium ethoxide was added. After 48 h stirring at room temperature the reaction mixture was worked up and product XIV isolated in good yield. Simi-
Nucleophilic reactions of sorbic acid
691
larly, methyl sorbate reacted with cysteine ethyl ester hydrochloride in dichloromethane in the presence of triethylamine. If, on the other hand, sorbic acid is allowed to react with cysteine in aqueous media and the product is esterified, it is identical to that formed from esters of sorbic acid and cysteine.
OEt
Downloaded by [University of Cambridge] at 19:30 05 November 2015
SR'
SR'
O
OEt
XV
XIV
The work was further extended to include ethyl-2-mercaptoacetate, mercaptoethanol and methyl-2-mercaptoacetate under similar conditions, and reactions always gave the monoadducts with the same structure as XIV, with the corresponding thiol attached to position 5. The work also included the reaction of alkyl thiols (ethanethiol, 2-methyl-2-propanethiol, butanethiol) with sorbate esters, in which case only di-adducts XV were isolated. No mono-adducts could be found even when the molar ratio of thiol to sorbate was 1:1 or less. Thus, it is seen that nucleophilic attack by alkyl thiols leads to the formation of di-adducts whilst reaction with mercaptoethanol, esters of 2-mercaptoacetic acid and cysteine ethyl ester hydrochloride give mono-adducts. Mechanism A mechanism for the attack on sorbate ester by the thiolate anion is given in figure 2. The anionic intermediate combines with H + to form either B or C.
SR'
OR
TS
SR'
H
H (B)
O
SR'
O
OR
OR
B (C)
Figure 2. Mechanism for the addition of thiolate anion to sorbate esters and possible rearrangement of the 3-ene product to a 2-ene product. The transition state is labelled TS and the two sets of arrows show electron shifts for the addition of H + to positions 2 and 4.
692
G. D. Khandelwal and B. L. Wedzicha
Downloaded by [University of Cambridge] at 19:30 05 November 2015
Product B is required if a di-adduct is to be formed. One question is whether, in the presence of a strong base such as sodium alkoxide, C may be converted to B; an acid/base-catalysed mechanism for isomerization is suggested in figure 2 and the driving force for the conversion could be the conjugated nature of the a,f3unsaturated ester so formed. Our results show that methyl-3-hexenoate is completely unreactive towards thiols under the same conditions as used for synthesis and, unless there is some participation from the substituent in position 5, the isomerization of the 3-ene to the 2-ene seems an unlikely possibility. When methyl 2-hexenoate is treated with sodium methoxide in the presence of thiol, product XVI is obtained and, in the absence of any nucleophile, product XVII is obtained. SR
O
XVI
OMe
O
XVII
An explanation for the mechanism is to say that the 3-hexenoate derivative is the thermodynamically favoured reaction product but the 2-hexenoate derivative may be formed as an initial kinetically favoured product. Thus, in the case of very reactive nucleophiles, e.g. alkylthiol, the formation of di-adduct is possible, whilst in the case of less reactive nucleophiles an irreversible isomerization to the more stable product predominates. The main difference between nucleophiles which cause mono-addition and those which cause di-adducts to be formed is the presence of oxygen in the former. It is hard to envisage that the electronegativity of the oxygen or the presence of lone pairs of electrons on this atom might be responsible for the creation of a field which causes repulsion of negative charge within the anionic transition state (TS in figure 2) away from position 4 to facilitate pick-up of a hydrogen ion at position 2. An intriguing question is, if given a range of oxygen-containing thiols of different hydrocarbon chain lengths, what is the length of chain before the thiol begins to behave as an alkyl thiol? Unfortunately, the longest chain mercaptocarboxylic acid available is 3-mercaptopropionic, because longer-chain acids tend to cyclize. The experiments have yet to be carried out with mercaptoalcohols of varying chain lengths, but it is clear that the size of the ester group (up to butyl) of the sorbate does not affect the stereochemistry of the addition reaction. The mono-adducts prepared do not add a second oxygen-containing thiol molecule, but a further question is whether they are able to add an alkyl thiol. The reaction of methyl sorbate-methyl thioglycollate reaction product with ethane thiol gave rise to product XVIII, whose 'H NMR spectrum was identical to that of an authentic sample reported by Blenderman and Joullie (1983). This indicates that displacement of the original nucleophile had taken place, and that ethanethiol had not added to the double-bond. To add even greater interest to the debate regarding the factors which control mono- and di-adduct formation, we have recently found that if sorbaldehyde is used in place of sorbates, the stereochemistry of addition is reversed, i.e. alkyl thiols react to form mono-adducts and the other thiols form diadducts. An intermediate situation exists with methyl hexadienyl ketone, where both types of adduct are obtained with both types of thiol.
Nucleophilic reactions of sorbic acid
693
SEt OMe XVIII
Downloaded by [University of Cambridge] at 19:30 05 November 2015
The detailed analysis of spectroscopic evidence and a full discussion of possible structural interpretations, which allowed the assignment of all structures presented here, is published elsewhere (Khandelwal and Wedzicha 1990). Hydrolysis of reaction products An important application of the synthetic work reported here is the preparation of samples for toxicological evaluation. For this to be achieved it is necessary to hydrolyse the esters to form the reaction products obtainable had sorbic acid itself been involved in the reactions. Attempts using acid and base hydrolysis of mono-adducts were unsuccessful because sorbic acid was obtained in high yield as one of the products. Currently we are examining the possibility of using esterase enzymes under neutral conditions to avoid breakdown of the 1:1 adducts. On the other hand, the desired alkyl thiol-sorbic acid reaction products were obtained when di-adducts were hydrolysed under normal conditions of ester hydrolysis. The instability of mono-adducts in acid or base solution has implications to their formation in foods. First, analysis of sorbic acid often involves steam distillation from acidified solution. The mono-adducts are decomposed under such conditions. Thus, the mono-addition reaction with thiols may not lead to a measurable loss of thiol; indeed, there may be no evidence that it may have occurred in a particular model or food system unless the investigator adapts the analytical procedure to allow for potential reversibility. However, the concentration of the available preservative may decrease as a result of such 'reversible' interactions. There may be a need for specifying free and reversibly bound sorbic acid content. It is interesting to note that the reaction of sorbic acid with sulphite ion is also reversible when mixtures are analysed for sulphur(iv) species (Heintze 1976). This may suggest that the structure of the reaction product should be shown with the double-bond in position 3, unlike structure V suggested by Hagglund and Ringbom (1926). In foods, possible reactions are between sorbic acid and cysteine residues of peptides and proteins, or thiols which are present as flavour components. In the case of reactions with proteins, subsequent proteolysis would release the mono-adduct of cysteine with sorbic acid, but it seems possible that the acid environment of the stomach could be conducive to the breakdown of this substance to sorbic acid and cysteine. The same could be true of the sorbic acid-sulphite ion adduct and, in both cases, the binding of sorbic acid might be of no toxicological significance. Acknowledgement
This work is being supported by the Ministry of Agriculture, Fisheries and Food. We are grateful to Dr D. J. McWeeny for his continuing interest. References BLENDERMAN, W. G., and JOULLIÉ, M. M., 1983, Lithium/ammonia reductions of 2-thiophenecarboxylic acids. Journal of Organic Chemistry, 42, 3206-3213.
Downloaded by [University of Cambridge] at 19:30 05 November 2015
694
G. D. Khandelwal and B. L. Wedzicha
FISCHER, E., and SCHLOTTERBECK, F., 1904, Verwandlung der Sorbinsäure in Aminosäuren. Berichte der Deutschen Chemische Gesellschaft,37,2357-2361. HÄGGLUND, E., and RINGBOM, A., 1926, ber die sulfitaddition an ungesättigte verbindungen. Zeitschrift für Anorganische und Allgemeine Chemie, 150, 231-253. HAYNES, L. J., 1948, Physiologically active unsaturated lactones. Quarterly Reviews (London), 2, 46-72. HEINTZE, K., 1976, Über die gegenseitige Beeinflussung von Sorbinsäure und schwefliger Säure. Die industrielle Obst- und Gem Üseverwet, 61, 555-556. KHANDELWAL, G. D., and WEDZICHA, B. L., 1990, Derivatives of sorbic acid-thiol adducts. Food Chemistry, 37, 159-169. KHEDDIS, B., BAHIBAH, D., HAMDI, M., and PÉRIÉ, J. J., 1981, Dihydropyridones-2 dérivées de l'acide sorbique: Synthèse et analyse conformationnelle. Bulletin de la Societé Chimique de France, 3-4(II), 135-140. KITO, Y., and NAMIKI, M., 1978, A new N-nitropyrrole. 1,4-dinitro-2-methylpyrrole, formed by the reaction of sorbic acid with sodium nitrite. Tetrahedron, 343, 505-508. KOOYMAN, E. C., FARENHORST, E., and WERNER, E. G. C., 1951, The nitrosation of methallylchloride. Recueil des Travaux Chimiques des Pays-Bas, 70, 689-695. LERAUX, Y., and VAUTHIER, E., 1970, Ètude par spectroscopie et moments dipolaires de diènes conjugués du type CH3CH=CH—CH=CH—Y. Comptes Rendus des Seances de l'Academie des Sciences, Sŕie C, 271, 1333-1336. MOSHER, H. S., 1950, Heterocyclic Compounds, Vol. 1, edited by R. C. Elderfield (New York: John Wiley & Sons), p. 651. OSAWA, T., and NAMIKI, M., 1982, Mutagen formation in the reaction of nitrite with food components analogous to sorbic acid. Agricultural and Biological Chemistry, 46, 2299-2304. OSAWA, T., KITO, Y., and NAMIKI, M., 1979, A new furoxan derivative and its precursors formed by
the reaction of sorbic acid with sodium nitrite. Tetrahedron Letters, 45, 4399-4402. PARK, J. R., and WILLIAMS, D. L. H., 1976, Reaction of dinitrogen trioxide with 2-methylpropene and other alkenes. Journal of the Chemical Society, Perkin Transactions, II, 828-832. REGER, D. L., and HABIB, M. M., 1980, Carbon-13 NMR investigation into the interaction of sodium sorbate and sorbic acid with various micelle-forming surfactants. Journal of Physical Chemistry, 84, 177-192. SHAMMA, M., and ROSENSTOCK, P. D., 1961, The synthesis and properties of some α,β-unsaturated valerolactams. Journal of Organic Chemistry, 26, 718-725. SYKES, P., 1965, A Guidebook to Mechanism in Organic Chemistry, 2nd edn (London: Longmans), pp. 149-155. VERBISCAR, A. J., and CAMPBELL, K. N., 1964, Unsaturated six-membered ring lactams. Journal of Organic Chemistry, 64, 2472-2474. WEDZICHA, B. L., 1984, Chemistry of Sulphur Dioxide in Food (London: Elsevier Applied Science). WEDZICHA, B. L., and BROOK, M. A., 1989, Reaction of sorbic acid with nucleophiles: preliminary studies. Food Chemistry, 31, 29-40. WEDZICHA, B. L., and ZEB, A., 1990, Kinetics of the reaction between sorbic acid and thiols. International Journal of Food Science and Technology, 25, 230-232.
