Synthesis of heterocycles from anthranilic acid and its derivatives

Anthranilic acid (2-aminobenzoic acid, Aa) is the biochemical precursor to the amino acid tryptophan, as well as a catabolic product of tryptophan in animals. It is also integrated into many alkaloids isolated from plants. Aa is produced industrially for production of dyestuffs and pharmaceuticals. The dissertation gives a historical background and a short review on the reactivity of Aa. The synthesis of several types of nitrogen heterocycles from Aa is discussed.

Treatment of anthranilonitrile (2-aminobenzonitrile, a derivative of Aa) with organomagnesium compounds gave deprotonation and addition to the nitrile triple bond to form amine-imine complexed dianions. Capture of these intermediate with acyl halides normally gave aromatic quinazolines, a type of heterocyclic compounds that is considered to be highly interesting as scaffolds for development of new drugs. When the acyl halide was a tertiary 2-haloacyl halide, the reaction instead gave 1,4-benzodiazepine-3-ones via rearrangement. These compounds are isomeric to the common benzodiazepine drugs (such as diazepam, Valium®) which are 1,4-benzodiazepine-2-ones.

Capture of the dianions with aldehydes or ketones, led to 1,2-dihydroquinazolines. Unsubstituted imine anions could be formed by treatment of anthranilonitrile with diisobutylaluminium hydride. Also in this case capture with aldehydes gave 1,2-dihydroquinazolines.

Several different dicarboxylic acid derivatives of Aa were treated with dehydrating reagents, and the resulting products were more or less complex 1,3-benzoxazinones, one of which required X-ray crystallography confirm its structure. During the work on preparation of N-substituted derivatives of Aa, necessary for synthesis of 1,4-benzodiazepine-3,5-diones, it was noted that many of the obtained products were in fact not N-substituted, but O-substituted. This challenged the established notion that Aa reacts nucleophilically at the N-terminal under most conditions. Several grave errors in the recent literature were revealed.

In 1976 researchers from the group that originally developed the common benzodiazepine drugs published a retraction of a claim of synthesis of a benzodiazepine by Gärtner in 1904. They found that the method actually gave a 6-membered ring system, not a 7-membered 1,4-benzodiazepine-3,5-dione as originally claimed. Because the 1,4-benzodiazepine skeleton is highly interesting as a scaffold for development of new drugs, a few publications on synthesis of this target has appeared. However, repetition of several of the described syntheses failed to yield the poorly described products.

Studies on how to ring close N-carbamoyl derivatives of Aa were undertaken. It became clear that Umpolung of the substrates by N-derivatisation was a necessary prerequisite for ring closure. The introduction of the N-nitroso group was developed to this end, leading to N1-nitroso substituted 1,4-benzodiazepine-3,5-diones. The nitroso group could be removed after ring closure. Heating of one of these compounds induced a ring contraction rearrangement. A proposed mechanism involves elimination of HNO (nitrosyl) and proton mediated loss of CO.

List of scientific papers

I. Wiklund P, Bergman J (2003). Ring forming reactions of imines of 2-aminobenzaldehyde and related compounds. Org Biomol Chem. 1(2): 367-72.
https://pubmed.ncbi.nlm.nih.gov/12929432

II. Wiklund P, Romero I, Bergman J (2003). Products from dehydration of dicarboxylic acids derived from anthranilic acid. Org Biomol Chem. 1(19): 3396-403.
https://pubmed.ncbi.nlm.nih.gov/14584803

III. Wiklund P, Bergman J (2004). Alkylation and acylation of basic salts of anthranilic acid. Tetrahedron Letters. 45: 969-72.

IV. Wiklund P, Rogers-Evans M, Bergman J (2004). "Synthesis of 1,4-benzodiazepine-3,5-diones." (Manuscript)