1,2,3-Dithiazole chemistry in heterocyclic synthesis

The chemistry of various 5 H -1,2,3-dithiazoles is investigated with emphasis on assisted ring opening and ring closure reactions leading to new heterocycles. Thus on treatment with catalytic tetraalkylammonium iodide N -(2-chloropyrid-3-yl)- and N -(4-chloropyrid-3-yl)-4-chloro-1,2,3-dithiazol-5 H -imines 19 and 20 give thiazolo[5,4-b ]pyridine-2-carbonitrile 16 and thiazolo-[4,5-c ]pyridine-2-carbonitrile 17 respectively. Similar treatment of bisdithiazoles 29 and 30 afford high yielding routes to 1,3,4-thiadiazole-2,5-dicarbonitrile 31 and thiazole-2,4,5-tricarbonitrile 32 respectively. N -(Pyrid-3-yl)-4-chloro-1,2,3-dithiazol-5 H -imine 36 reacts with secondary alkylamines to give as main product pyrido[2,3-d ]pyrimidines 37 and several minor byproducts including a deep green quinoidal 2,2’-bithiazole 40 . Dithiazolylidenacetonitriles 43 react with either anhydrous HBr or tetraalkylammonium chloride to afford a series of 3-halo-4-substituted-isothiazole-5-carbonitriles 45 and 52 respectively. The reactions of dithiazoles 43 with tetraalkylammonium chloride are complicated owing to the formation of isothiazolo-pentathiepin-8-carbonitrile 53 , isothiazolodithiin-4,5,7-tricarbonitrile 54 , tetracyanothiophene 56 and an unidentified compound 55 whose possible structures are proposed. The mechanistic rationales for the formation of the identified products are proposed.


Introduction
1,2,3-Dithiazole is one of the four possible dithiazole systems all of which are reported in the literature as cations.The salts are planar, 6π and therefore aromatic and can be adequately represented by three canonical forms where the charge (identifying the three most electrophilic sites) is distributed at C-5 or on either of the ring sulfur atoms (Scheme 1).1,2,3-Dithiazolium salts are commonly prepared from substituted acetonitrile derivatives to afford the 5-substituted-4-chloro-1,2,3-dithiazolium chlorides 1. 1,2 One example has appeared in the literature where phenylacetoxime was treated with disulfur dichloride to afford the 5-chloro-4-phenyl-1,2,3-dithiazolium chloride 2 and development of this route could provide access to 5chloro-4-substituted-1,2,3-dithiazolium salts. 3 Our interest focuses on the chemistry of the readily available 4,5-dichloro-1,2,3dithiazolium chloride 3 prepared from chloroacetonitrile and disulfur dichloride. 5Dithiazolium chloride 3 on treatment with nucleophilic species affords neutral 5H-dithiazoles 4 (eg.treatment of dithiazolium 3 with H 2 O, aniline, diethyl malonate or H 2 S provides 1,2,3-dithiazolone 5, ARKAT dithiazolimine 6, dithiazolylidene 7 and dithiazolethione 8 respectively in good yields (Scheme 2). 5  1,2,3-Dithiazoles have uses in both biological and material sciences: N-aryldithiazolimines show interesting antifungal, 6 antibacterial, 7 and herbicidal 8 activities.A search for organic conductors based on neutral radicals has led to the preparation of two 1,2,3dithiazolyl radicals 9 1 and 10 9 and also a tetrathiadiazafulvalene analogue 11 10 has been prepared and studied.Our interest in 1,2,3-dithiazole chemistry revolves around the construction of dithiazole systems that can be converted into new heterocyclic systems via ring transformation.The majority of these ring transformations involve the initial preparation of a neutral dithiazole which supports a potentially nucleophilic side chain or substituent capable of attacking the electrophilic dithiazole at either S-1 (Path A) or at C-5 (Path B) with subsequent ring opening.Dithiazoles, however, can also be ring opened with the use of soft nucleophiles to afford the disulfide intermediate 12 (Path C) (Scheme 3).This disulfide can be a source of both electrophilic and nucleophilic sulfur.Various dithiazolylidenes have been prepared specifically with these mechanistic possibilities in mind in order to broaden the capability of dithiazolium chloride 3 as a useful synthetic tool.Emphasis will be made on the in situ generation of disulfides 12 endowed with electrophilic traps for nucleophilic sulfur (Path C).

Scheme 4
Where the aryl group is electron deficient the major product is the imidoyl chloride carbonitrile 14. 11 The mechanism that has been proposed by Rees is shown in Scheme 4. To our knowledge no examples of the thermolysis of N-heteroaryl-1,2,3-dithiazol-5H-imines have appeared in the academic or patent literature.
Thermolysis of N-(pyrid-3-yl)-4-chloro-1,2,3-dithiazol-5H-imine 15 gave thiazolo- [5,4-b]pyridine-2-carbonitrile 16 and thiazolo [4,5-c]pyridine-2-carbonitrile 17 in low yields as might be expected based on the mechanism to give benzothiazole 13.Repeating the reaction in the presence of a soft nucleophile benzyltriethylammonium iodide at 132 o C in chlorobenzene gave improved but still moderate yields of both isomers.Under these conditions the dithiazole ring is anticipated to suffer an assisted nucleophilic ring opening-ring closure (ANRORC) 12 like mechanism involving the intermediate disulfide 18.The nucleophilic sulfur that is generated (S-1) is trapped at the electrophilic pyridyl C-2 and C-4 positions and a subsequent oxidation restores the aromaticity of the thiazolopyridine systems.Introducing a chlorine substituent at either C-2 or at C-4 on the pyridyl ring was expected to assist in directing the ring closure and furthermore adjusts the oxidation level of the starting systems thus avoiding the need for oxidative rearomatisation.The 2-chloro and 4-chloroderivatives 19 and 20 were readily prepared starting from the corresponding aminochloropyridines and dithiazolium chloride 3. Gratifyingly treatment with catalytic iodide (5 mol %) gave the expected single isomers in near quantitative yields.Thermolysis of either the 2-chloro and 4-chloro-derivatives 19 and 20 did not yield regiospecific ring closures but mixtures of products were isolated adding further support to the ANRORC type mechanism proposed above.We attempted to extrapolate the above success to the preparation of 3,1-benzothiazin-4ones starting from the readily available methyl ester 21 where the electrophilic trap was now the carboxylate group.Initial thermolysis of the ester 21 gave some of the desired benzothiazinone 22 together with the benzothiazolecarbonitrile 23.On treatment with various soft nucleophilies the major product was however the Narylcyanothioformamide 24.In our case the methyl ester 21 on treatment with triphenylphosphine gave only a moderate yield of the cyanothioformamide 24.

Synthesis of percyano thiazole and 1,3,4-thiadiazole
The 1,2,3-dithiazole ring can act as both a source of nucleophilic sulfur at S-1 and as electrophilic trap at the ring carbon C-5.
On treatment with soft nucleophiles bisdithiazole systems such as 27 can therefore be considered as possible precursors to heterocyclic systems of type 28 (Scheme 5).Two bisdithiazoles of this type compound 29 14 and compound 30 15 have been reported in the literature and each could provide rapid routes to the corresponding percyano-1,3,4-thiadiazole 31 and percyanothiazole 32 respectively.Treatment of either bisdithiazole 29 or 30 with soft nucleophiles such as chloride, bromide or iodide gave the expected percyano-1,3,4-thiadiazole 31 and thiazole 32 systems respectively.These percyano heterocycles suffered hydrolysis during chromatographic isolation to afford a moderate quantity of the corresponding carboxamides 33 and 34 respectively.Chromatography could be avoided by using polymer bound triphenylphosphine as the soft nucleophile since the polymer resin could be separated by filtration to afford a clean solution of the desired percyano heterocycle.Recrystallisation of this gave pure product free of any impurities or carboxamide.Surprisingly free triphenylphosphine gave only traces of product.

Synthesis of fully substituted pyrido[2,3-d]pyrimidines 37: unexpected byproducts
The dithiazolimine 36 prepared from the available fully substituted 2-aminopyridine 35 gave on treatment with cyclic secondary amines the expected fully substituted pyrido [2,3-d]pyrimidines 37 in moderate to good yields together with minor quantities of the guanidine 38 and the 4-aminosubstituted dithiazolimine 39 (Scheme 6).This is the first time such a product type has been observed in 1,2,3-dithiazole chemistry, which raises new questions about the possibilities of the use of 1,2,3-dithiazoles in synthesis.A possible precursor to this compound is the 4-aminosubstituted dithiazolimine 39 which could have been cleaved to afford the ring opened intermediate 41.These intermediates have been prepared starting from the 4-aminosubstituted dithiazolimines by Kim using alkali bases in alcohol. 16The two central carbons forming C-2 of the thiazole rings are presumed to be derived from the solvent CH 2 Cl 2 .However, attempts to improve the yield of the 2,2'-bithiazole 40 by replacing the solvent with different sources of carbon such at CHCl 3 , CCl 4 , CH 2 ClBr, CH 2 ClI, CH 2 Br 2 , and CHBr 3 gave very similar reaction mixtures but surprisingly no trace of the green product, though formation of the green product 40 in the presence of CH 2 Cl 2 was reproducible.The formation of the pyrido[2,3-d]pyrimidine 37 from dithiazolimine 36 is in itself mechanistically interesting as two possible pathways can be envisaged.In the first (Path A) the nitrile group neighbouring the dithiazolimine is initially attacked by the amine and cyclises onto the dithiazole C-5 carbon which instigates opening of the dithiazole ring.A second possibility (Path B) involves attack by the amine at the dithiazole ring sulfur S-2 to afford the disulfide intermediate 42 which then suffers ring closure to afford either the 4-aminodithiazolimine 39 or the pyrido [2,3-d]pyrimidine 37 depending on which nitrile is attacked preferably by the incoming amine (Scheme 7).Further studies are needed to determine which is the predominant mechanism.

Chemistry of dithiazolylidenacetonitriles 43: Formation of isothiazoles
8][19] Treatment of dithiazoles 43 with anhydrous HBr afforded the 3-bromoisothiazoles 45 in moderate to good yields.However, with the dithiazolylidenemalononitrile 43 (X = CN) the analogous treatment with anhydrous HCl gives only a trace of the expected isothiazole 44. 19Here we argue that the reaction mechanism proceeds by formation of the imidoyl bromide 46 which then cyclises onto the dithiazole ring sulfur S-1.Anhydrous HCl being a weaker acid than HBr is unable to drive this transformation.During the investigation of this reaction mechanism two unexpected 3H-pyrroles were isolated. 18The first, a deep blue colored 3H-pyrrole 49 whose structure was determined by single crystal X-ray crystallography, was from the reaction of triphenylphosphine with dithiazolylidenemalononitrile 43 (X = CN).The second, on treatment of the dithiazole 43 (X = CN) with excess morpholine, was the orange colored 3H-pyrrole 50.
In light of difficulties in displacing the 4-chloro substituent of neutral 5H-1,2,3-dithiazoles a rational mechanism for the formation of these 3H-pyrroles 49 and 50 required the involvement of the ring opened disulfide [cf. the proposed mechanism of the bismorpholino-3H-pyrrole 50 (Scheme 9)].As such the dithiazolylidenemalononitrile 43 (X = CN) can ring open to afford an intermediate disulfide 51 that could afford isothiazoles, 3H-pyrroles and even (although this has not been observed yet) 1,2,3-dithiazole ring systems depending on the cyclisation modes of the tricyanovinyl group (Scheme 10).In order to investigate the reaction further, a series of substituted dithiazolylidenacetonitriles was prepared and treated with benzyltriethylammonium chloride.Unfortunately for the halo and non-substituted acetonitrile derivatives 43 (X = H, Cl, or Br) the geometry of the attached nitrile group has not been determined (ie.cis or trans with respect to the dithiazole ring sulfur) but in each case only one isomer was observed by NMR and by chromatography.Owing to a strong non-bonding interaction between the ester carbonyl group with the dithiazole ring sulfur S-1 the ethyl carboxylate 43 (X = CO 2 Et) is known to have a trans geometry for the nitrile group and the ring sulfur, 20 however, this system was unreactive to halide.

8
Scheme 8 Scheme 9 Scheme 10 Rees et al. reported a quantitative conversion of the iminocarboxylic acid 25 into benzothiazinone 22 on treatment with triphenylphosphine and proposed that the phosphonium salt byproducts help activate the carboxylic acid towards attack (cf.intermediate 26).