N -Benzoyl- N -methylsulfonylanthranilates: unexpected cyclization reaction to 4-alkoxy-2,1-benzothiazines

The synthesis of 4-alkoxy-1 H -2,1-benzothiazine 2,2-dioxide derivatives was achieved through an unexpected base-mediated cyclization of N -benzoyl- N -methylsulfonylanthranilates. A reaction mechanism involving the sulfene species has been postulated. The obtained 4-alkoxy derivatives can be further functionalized at the N-1 position allowing unsymmetrically N , O - disubstituted 1 H -2,1-benzothiazines 2,2-dioxide to be prepared

The synthetic route to obtain the 1H-2,1-benzothiazine-4(3H)-one 2,2-dioxide skeleton was first reported simultaneously by two different groups, Loev and co-workers 3 and Rossi and coworkers. 4Both groups described a virtually identical five-step procedure which entails: (1) the conversion of sulfoacetic acid into its carboxylic methyl half-ester and then (2) into the sulfonylchloride which was then (3) reacted with aniline to give methyl N-phenylsulfamoylacetate; followed (4) by hydrolysis of this ester-amide and (5) cyclization with PPA.An improved 3-step synthesis was then developed by Lombardino which entails the base-mediated cyclization of a N-alkyl-N-(methylsulfonyl)anthranilate, prepared from the coupling of an anthranilate and methanesulfonyl chloride (MsCl), followed by N-alkylation of the resulting sulfonamide. 5This general method results in a relatively high overall yield and offers an easier entry into the 2,1benzothiazine 2,2-dioxide heterocyclic system.It must be pointed out that the cyclization only occurs for the N-alkylated intermediates since the presence of a strongly acidic sulfonamide hydrogen prevents further formation of the methyl carbanion that is essential for the cyclization reaction. 6ore recently, a solid-phase method for the synthesis of 1H-2,1-benzothiazine-4(3H)-one 2,2-dioxide derivatives utilizing polymer-bound anthranilic acid derivatives was reported by Jeon and co-workers. 7This procedure is suitable for the combinatorial generation of drug-like heterocyclic compound libraries even though it is not associated with high yields of the target compounds.

Results and Discussion
The more convenient Lombardino procedure was chosen to obtain compound 1.Actually, through a serendipitous discovery we found an interesting way to synthesize 4-alkoxy-1H-2,1benzothiazine 2,2-dioxide derivatives.
Thus, as depicted in Scheme 1, the N-sulfonylanthranilate 4a was prepared by the reaction of methyl 5-chloro-2-nitrobenzoate 8 with 2-bromophenol, in DMSO and in the presence of K2CO3, followed by the catalytic reduction to amino derivative 3a 9 and treatment with MsCl in pyridine and CH2Cl2.The reaction of N-(methylsulfonyl)anthranilate 4a with benzoyl chloride and NaH in dry THF gave the N-benzoyl-N-(methylsulfonyl)anthranilate 5a.When this key intermediate 5a was submitted to the cyclization step by treatment with NaH in dry DMF, the desired cyclization product 1 was not obtained while 6-(2-bromophenoxy)-4-methoxy-1H-2,1benzothiazine 2,2-dioxide 7a was formed as the major reaction product (40% yield) together with N-benzoylanthranilate 9a (20% yield).2b The structure of 4-methoxy derivative 7a was assigned only by NMR experiments since it was not possible to obtain this compound in a suitable crystal form for X-ray analysis.Its 1 H NMR spectrum shows the MeO and H-3 signals as singlets at  3.79 and  6.67, respectively together with a broad singlet at  11.44 due to a NH exchangeable proton.Moreover, 13 C NMR spectrum indicates two diagnostic signals: a singlet of MeO at 57.25 ppm and a singlet of C-3 at 100.57ppm.To definitively confirm the exact structure of compound 7a, a bidimensional NOESY spectrum was performed.The results indicate two main interactions: the first between the 4-OMe and the H-3 proton and the second between the same OMe group with the H-5 proton (Figure 1).Besides the 1 H NMR and 13 C NMR experiments, the structure of the N-benzoylanthranilate 9a was unambiguously assigned through X -ray crystallography (Figure 2).The same behaviour was observed when the ethyl ester derivative 5b was used in the cyclization step; the corresponding 4-ethoxy derivative 7b was obtained (35% yield) together with N-benzoylanthranilate 9b (25% yield) (Scheme 1).In order to evaluate the reaction reproducibility and to exclude any influence of the 2-bromophenoxy substituent on the reaction trend, the cyclization was repeated using both unsubstituted methyl and ethyl N-benzoyl-N-(methylsulfonyl)anthranilate, 6a and 6b respectively; again 4-alkoxy 2,1-benzothiazines 8a (36%) and 8b (33%) were obtained together with N-benzoylanthranilate 10a 10 (20%) and 10b 11 (25%).
However, to obtain the target compound 1, 4-alkoxy derivative 7a was elaborated further by N-benzoylation, under basic conditions in THF, to derivative 11 which was subsequently O-demethylated in 48% HBr (Scheme 1).The latter reaction was carefully monitored to minimize the N-debenzoylation reaction.
Turning back our attention to the cyclization step, the unexpected formation of both 4-alkoxy derivatives 7 and 8 together with N-benzoylanthranilates 9 and 10 led us to speculate on the reaction mechanism.In a first hypothesis, it was supposed that desulfonylation occurs together with cyclization on the ester group followed by dehydration.Thus, N-benzoylanthranilate 9a was obtained by desulfonylation.On the other hand, 4-alkoxy-N-benzoyl intermediate (i.e., 11) had to be obtained as a result of cyclization/dehydration. Since 4-methoxy 2,1-benzothiazine 7a was obtained, only an additional N-debenzoylation of 4-alkoxy-N-benzoyl intermediate 11 could explain the formation of derivative 7a.Nevertheless, it must be noted that it has never been reported that cyclization on the ester group is followed by dehydration when the N-alkyl-Nmethylsulfonylanthranilate substrate is used.This observation led us to further speculate that the following mechanism, depicted in Scheme 2 for methyl ester 5a (Ar = 2-BrPh), might be operative: under basic cyclization conditions the methylsulfonyl group is removed which generates the sulfene species (II) 12 together with 9a in its anion form (I); the highly reactive sulfene (II) can react with I giving a simultaneous nucleophilic/electrophilic reaction with the methyl ester and with the nitrogen anion, respectively; this makes it possible for the benzoyl to migrate from the nitrogen to the oxygen atom (III).The benzoyl migration allows the C-4 position to be occupied by a benzoyloxy group, which is a better leaving group than the methoxy one.The benzoate leaves, via cB1 elimination, permitting the unexpected 6-(2-bromophenoxy)-4-methoxy-1H-2,1-benzothiazine 2,2-dioxide 7a to be obtained.The low yield observed for compound 7a, and in general for the other 4-alkoxy 2,1-benzothiazines, could be due to the instability of the sulfene species which hampers a quantitative reaction with intermediate I; unreacted I furnishes benzamide 9a after acidic workup.Scheme 2. Postulated reaction mechanism.

Conclusions
In conclusion, a reliable procedure for synthesizing 4-alkoxy-1H-2,1-benzothiazines 2,2-dioxide was serendipitously discovered through an unexpected cyclization reaction carried out on N-benzoyl-N-(methylsulfonyl)anthranilates.Nevertheless, this cyclization step results in a low yield which needs to be improved to have an efficient synthetic method.Despite this limitation, the procedure herein reported is, to the best of our knowledge, the only way to achieve 4-alkoxy-1H-2,1-benzothiazines 2,2-dioxide.