Concise Synthesis of 6-Cyanobenzo[b]furan, a Useful Building Block

Abstract A new three-step synthesis of 6-cyanobenzo[b]furan (6) was developed, starting from commercially available 6-hydroxybenzo[b]furan-3-one (18). Key steps in this process were the first step, which was the reductive dehydration of 18 to produce 6-hydroxybenzo[b]furan (19), and the last step, which converted the aryl triflate 20 to the aryl cyanide 6 in a palladium-catalyzed cross-coupling protocol. Overall yield for this new synthesis was 49%. GRAPHICAL ABSTRACT


INTRODUCTION
Benzo [b]furans are important components of drugs and biologically active compounds. Examples of benzo[b]furan-containing drugs include (−)-1-(benzofuran-2-yl)-2-propylaminopentane ((−)-BPAP), [1] dronedarone, [2] and amiodarone. [3] Consequently, there has been significant effort directed toward the design of new syntheses for benzo [b]furans. [4,5] We have described previously the syntheses of specific benzo[b]furan building blocks. 5-Formylbenzo[b]furan-2-carbonitrile was a previously unknown compound that we required for the preparation of a compound library. We recently reported [6] a seven-step, practical synthesis for this new compound in 17% overall yield. More recently, we developed a new synthesis for 6-cyanobenzo[b]furan (6), which is shown in Scheme 1. [7] Starting from 3-hydroxybenzonitrile (1), we prepared triiodo compound 2 in 88% yield by treatment with iodine and then selectively removing two iodo groups from 2 with N-methylmorpholine to give 3-hydroxy-4-iodobenzonitrile (3) in 28% yield. Compound 3 underwent a Sonagashira reaction, [8] using the procedure of Wishka et al. [9] to produce the intermediate phenylacetylene 4 in 79% yield. Subsequent cyclization of 4 with cuprous iodide catalysis gave a 93% yield of 6-cyanobenzo[b]furan (6), after treatment of the initial mixture with sodium hydroxide to convert the 2-trimethylsilylbenzofuran intermediate 5 to the desired product 6. Although this sequence was relatively short, and four of the five steps proceeded in good yield, the deiodination of compound 2 to produce 3-hydroxy-4-iodobenzonitrile (3) was a poor-yielding reaction that we were unable to improve, and this prompted us to evaluate new methods for the preparation of 6. This report describes an improved, concise synthesis of 6-cyanobenzo[b]furan (6).

RESULTS AND DISCUSSION
In Scheme 2 is shown another known route to 6-cyanobenzo[b]furan (6), starting from 4-nitrosalicylic acid (7). Reduction of 7 with borane-dimethylsulfide complex gave benzyl alcohol 8 in 95% yield. [10] Oxidation of 7 with manganese dioxide furnished 4-nitrosalicylaldehyde (9) in 75% yield. [10] Aldehyde 9 was then used to install the fused furan ring, by treatment with 2-bromomalonic acid diethyl ester under basic conditions. [11] This reaction was reported to produce ester 10 in 59% yield, which actually is a reasonably good yield in view of the complexity of this conversion that involves alkylation, condensation, hydrolysis and decarboxylation processes. Hydrolysis of 10 with potassium hydroxide afforded the corresponding acid 11 in 92% yield, [12] and subsequent decarboxylation with copper and quinoline at high temperature gave 6-nitrobenzo[b]furan (12) in 63% yield. [12] Chemical reduction of the nitro group of 12 with ferric chloride in the presence of N,N-dimethylhydrazine gave 6-aminobenzo[b] furan (13) in 79% yield. [12] Conversion of the amino group in 13 to a cyano group to produce the desired compound 6 was accomplished in a two-step Sandmeyer reaction, in the presence of sodium cyanide. [13] The yield for this reaction was not reported.

SYNTHESIS OF 6-CYANOBENZO[b]FURAN 1309
The overall process for the preparation of benzofuran 6 shown in Scheme 2 is unattractive because of its length, poor yields with two of the conversions, and an unknown yield for the last two-step Sandmeyer reaction. Also, the Sandmeyer conversion involved the use of sodium cyanide under acidic conditions, which we wished to avoid. Although this chemistry is basically a compilation of known reactions that would undoubtedly work to produce product, it underscored the need for a better synthesis of benzofuran 6.
We looked for cost-effective, commercially available starting materials that potentially could be converted to 6-cyanobenzo[b]furan (6). In Fig. 1 are shown four such materials that we identified and attempted to convert to compound 6. All of these potential starting materials had structural elements common to 6, and we made several attempts to use these materials to devise a new route to benzofuran 6.
We first attempted to alkylate 3-hydroxybenzoic acid (14) with a two-carbon synthon, such as bromoacetaldehyde diethyl acetal, reasoning that a subsequent annulation of this unit to the 4-position of the benzene ring could provide a precursor to compound 6. However, the alkylation conditions that we explored all gave mixtures of alkylated products, where alkylation had occurred to produce the desired ether but had also partially produced a dialkylated ether-ester product. Because this complication would lengthen an already long planned sequence, we moved next to 3-bromophenol (15), where only monoalkylation of the phenol could occur.  Alkylation of 15 with bromoacetaldehyde diethyl acetal, using sodium hydride in dimethylformamide, gave us the desired ether in 98% yield. However, when we treated this ether with acidic reagents (e.g., p-toluenesulfonic acid in toluene or xylene; or polyphosphoric acid) in an attempt to produce the desired benzofuran, we observed mixtures of 4-bromobenzo[b]furan and 6-bromobenzo[b]furan. These mixtures were very difficult to separate and yields of either isomer were low (less than 2%). 5-Bromo-2-chlorophenol (16) was a potentially attractive starting material, because problems of regiochemistry would no longer be an issue. We were able to replace the bromo group with cyano in compound 16, using zinc cyanide and Pd(Ph 3 P) 4 in dimethylformamide, in 94% yield. We planned to attempt a Sonagashira reaction with this compound (2-chloro-5-cyanophenol), but became discouraged by literature reports that described the lower reactivity of chlorobenzenes. [14] However, it was very useful to develop the cyanation conditions with compound 16, because this replacement reaction was a critical feature of our ultimate, successful route to compound 6.
3-Cyanophenol (17) was also used as a starting point for alkylations that could ultimately produce the desired benzofuran. Treatment of 17 with bromoacetaldehyde diethyl acetal gave a 72% yield of the expected O-alkylated product. Treatment of this ether with polyphosphoric acid, however, gave a mixture from which only 4-cyanobenzo[b]furan could be isolated. We were also able to alkylate 17 with allyl bromide, using potassium carbonate in acetone at 0 °C, to provide a 94% yield of the expected propargyl ether. We had planned to do a Claisen rearrangement with this ether using diethylaluminum chloride, [15] evaluate the regiochemistry of the product or products, and subsequently convert the desired o-allylphenol, if formed, by ozonolysis to 2-hydroxy-4-cyanophenylacetaldehyde, which could then be cyclized to benzofuran 6. However, we did not proceed with this plan because of the result we obtained with the diethyl acetal.
In Scheme 3 is shown the new procedure for the synthesis of 6-cyanobenzo[b] furan (6), starting from commercially available 6-hydroxybenzo[b]furan-3-one (18). We believe that this concise method is now the best procedure that has been reported for the synthesis of 6. The first step in this process involved conversion of 6-hydroxybenzo[b] furan (18) to 6-hydroxybenzo[b]furan (19) using a reductive dehydration protocol. Thus, treatment of 18 (which can be prepared from resorcinol in two steps [16] ) with lithium borohydride in tetrahydrofuran, followed by treatment with acid, produced 6-hydroxybenzo[b]furan (19) in 68% yield. Compound 19 has previously been prepared from 18 in a three-step process from 18, by protection of the hydroxyl group using tert-butylsilyl chloride, reduction of the ketone with sodium borohydride, and dehydration of the resulting alcohol by treatment with acid. [17] Conversion of 19 to the triflate 20 proceeded in 88% yield; compound 20 then underwent a cross-coupling

1311
reaction with zinc cyanide, in the presence of palladium-tetrakis(triphenylphosphine) palladium(0), to give 6 in 82% yield. Palladium-catalyzed cyanation of aryl triflates is useful methodology for the preparation of aryl cyanides. [18] The conditions shown in Scheme 3 for the reductive dehydration of compound 18 are the optimized conditions that were found after several attempts. In Scheme 4 is shown our first attempt, with lithium aluminum hydride in tetrahydrofuran, [19] which made us realize that overreduction to the dihydrobenzofuran 21 was a problematic side reaction that needed to be minimized. To improve the yield of the reductive dehydration process, we screened various other conditions, using different reducing reagents, as shown in Table 1. All reactions in Table 1 were performed at room temperature, which helped to minimize the production of 21. The best conditions found were LiBH 4 in tetrahydrofuran (THF) followed by a HCl/H 2 O quench, which provided 19 in 68% yield (first entry). Interestingly, it was observed that if the reaction mixture was quenched with water and kept neutral, we could isolate the intermediate 2,3-dihydrobenzo[b]furan-3,6-diol (22). Thus, addition of HCl was necessary for the elimination of the secondary alcohol to give 19.

FUNDING
Research reported in this publication was supported by the National Institutes of Health under Award No. U01AI082052. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

SUPPLEMENTAL MATERIAL
Supplemental data for this article can be accessed on the publisher's website.