The furan ring is a pervasive structural motif found in a broad range of natural products and pharmacologically active compounds.^[1] Fused furans are a class of polycyclic furan derivatives with multifaceted significance and can be classified into either b-fused furans (2,3-fused furans) or c-fused furans (3,4-fused furans), which exist in a vast number of biologically active molecules.^[2] For example, the antiproliferative activity of natural product asperfuranone was found to be active against human non-small A549 lung cancer cells.^[3] Axially chiral furan-based scaffolds have been recognized as a class of important heteroaryl atropisomers and constitutes the core units of many chiral phosphine ligands and bioactive molecules.^[4] Over the past decades, considerable efforts has been devoted to the synthesis of fused furans from non-cyclic precursors, especially for benzo[b]furans due to their benzenoid structure^[5] and potential in the construction of axially chiral scaffolds.^[6] In contrast, methods to achieve the c-fused furans are quite limited.^[7] Tandem reaction is an efficient strategy for the one-pot synthesis of c-fused furans.^[8] In this context, Wang and coworkers^[8a] have reported the construction of oxa-/aza-[n.3.1]skeletons through an Au(I)- catalyzed intramolecular cross-cycloaddition of alkynylcyclopropane ketones (Scheme 1A). Similarly, Nishiyama et al.^[8b] also have reported a Ru-catalyzed transfer oxygenative cyclization of α,ω-diynes for the synthesis of c-fused furans (Scheme 1B), where the O atom in bicyclic furan products was transferred from dimethyl sulfoxide (DMSO).

In recent years, we have been committed to developing novel tandem reactions for the synthesis of key structural motifs, especially tandem rearrangements via oxonium ion^[9] or sulfonium ion^[10] intermediates. Inspired by the stereodivergent synthesis of 1,4-dicarbonyl compounds via [3,3]-sulfonium rearrangement of sulfoxides reported by Maulide and coworkers,^[11] we^[10] have reported a novel tandem [3,3]-sulfonium rearrangement strategy for the construction of tetrasubstituted furans (Scheme 2A). This transition-metal-free tandem reaction was initiated by the cross-coupling of heterosubstituted alkynes and alkynyl sulfoxides in the presence of boron trifluoride diethyl etherate (BF₃•Et₂O) followed by a [3,3]-sigmatropic sulfonium rearrangement and heterocyclization to furnish tetrasubstituted furans as the final product with 100% atom economy. Encouraged by the success of this intermolecular sulfonium rearrangement, we naturally envisioned that an intramolecular [3,3]-sulfonium rearrangement could proceed via a similar reaction pathway, providing cycloalkano[c]furans as products (Scheme 2B). The proposed mechanism of the tandem intramolecular rearrangement was illustrated in Scheme 3. Electrophilic activation of alkylthio/alkylsulfinyl substituted diynes 1 by Lewis acid initiated the intramolecular cross-coupling of alkynyl sulfides and alkynyl sulfoxides, thereby generating the O-al- kenylsulfoxonium ion intermediate 3. The consecutive [3,3]-sigmatropic sulfonium rearrangement gives thioketene intermediate 4, followed by a 5-exo-dig heterocyc- lization and aromatization process to deliver cycloalkano-[c]furans 6. The positively charged sulfur could be an extra driving force for the rearrangement.^[12] However, this intramolecular tandem reaction faces a significant challenge from the potential competing intermolecular rearrangement which leads to undesired intermolecular cross- coupling products or oligomers. Accordingly, minimizing the interference from intermolecular reactions and ensuring the designed intramolecular reaction pathway are crucial for achieving selective access to cycloalkano[c]furans 6.

Our research began with the preparation of diyne sulfide sulfones 1 which serve as starting materials. Compounds 1 can be synthesized through the nucleophilic substitution of terminal diynes to disulfides^[13] and the successive mono- oxidation with m-CPBA or Oxone.^[14]

To begin with, a pilot experiment was conducted to evaluate the validity of the strategy, using methyl(8-(meth- ylsulfinyl)octa-1,7-diyn-1-yl)sulfane (1a) as the substrate in 1,2-dichloroethane (DCE). BF₃•Et₂O was chosen as the catalyst based on the previous performances.^[10] Unexpectedly, product 6a' was isolated instead of the predicted product 6a with a yield of 45% (Scheme 4). High-resolu- tion mass spectrometry (HRMS) revealed that the exact mass of 6a' was twice as many as the bicyclic 6a’s, suggesting that this unlooked-for product might result from the double reaction of 1a. To verify the hypothesis, the ethyl analogue 6b' was synthesized. Its structure was unambiguously identified by single-crystal X-ray crystallography, revealing a 12-membered ring scaffold. Obviously, 6a' and 6b' were formed via intramolecular rearrangement of the cross-dimeric intermediate, which was originated from sulfoxide 1 through an intermolecular sulfonium rearrangement similar to our previously described sequence. The occurrence of this undesired reaction led us to seek conditions that suppress the competitive dimerization of sulfoxide 1 (Table 1). To our delight, decreasing the substrate concentration to 0.01 mol/L delivered the desired product 6a with a 50% isolated yield without detection of dimeric byproduct 6a' (Entry 3).

After examining the substrate concentration, the influence of other reaction parameters was further investigated. Selected conditions for optimization are summarized in Table 2. Elevating reaction temperature was proved to enhance the yield (Entries 4 and 5). In addition, extending reaction time led to the complete conversion of 1a with slightly improved yields (Entries 2 and 3). The application of anhydrous sodium sulfate as a drying agent led to a slightly improved isolated yield (62%) of 6a (Entry 6). Increasing the catalyst dosage from 0.1 equiv. to 1.0 equiv. did not further improve the yields (Entries 7~9), prompting us to explore several other catalysts. Reactions with Brønsted acid TfOH or Tf₂NH gave slightly inferior yields with complex product mixtures (Entries 10 and 11). Other catalysts including several transition-metal catalysts commonly employed for alkyne activation^[15] were also attempted, which worked even worse than BF₃•Et₂O (Entries 12~19). As for solvent selection, 1,2-dichloroethane (DCE) turned out to be the optimal choice for its high boiling point. Other common nucleophilic solvents were not involved here to prevent possible interference with heterocyclization process. Based on these examinations above, the optimized reaction conditions were determined to be: 0.5 mmol of alkynyl sulfoxide 1, 0.1 equiv. of BF₃•Et₂O and 2.0 g of anhydrous Na₂SO₄ in 50 mL of DCE at 70 ℃ for 12 h.

Employing the optimal conditions shown above, the substrate scope of this intramolecular rearrangement was examined (Table 3). The impact of steric hindrance on the RS substituents was first investigated (6a~6d). Introduction of a bulky group such as phenyl group (6d) led to a significantly decreased yield of 33%. Subsequently, the influence of the carbon tether length was examined (6e~6j). The result showed that sulfoxides were tolerated regardless of extending or shortening the carbon tethers, suggesting the flexibility on the fused-ring size. To our surprise, 1g and 1h could also undergo the reaction, affording furan products containing medium-sized rings (6g and 6h), although with relatively lower separation yields. It should be noted that medium-sized rings are typically challenging to synthesize due to the unfavorable transannular steric interactions and entropic effects.^[16] For macrocycle products, 12- and 14-membered ring products (6i and 6j) were also furnished smoothly with slightly higher isolated yields than their medium-ring counterparts (42% and 46%). To further increase the product diversity, several substrates containing benzo ring (6k) or heteroatoms (6l~6n) in carbon tethers were employed and no significant drops in isolated yields were found in corresponding c-fused furan products. Additionally, the structures of cycloalkano[c]furans 6f and 6j have been unambiguously confirmed by single-crystal X-ray crystallography.

Finally, subsequent post modification of the cycloalkano[c]furan products was investigated. The Ni-NHC catalyzed cross-coupling of aryl sulfides with Grignard reagents, as reported by Yorimitsu and coworkers,^[17] was utilized for the conversion of alkylthio groups in cycloalkano[c]furan products (Table 4). Both alkyl and aryl Grignard reagents were adaptable, providing the correspond- ing cycloalkano[c]furan products 7a~7d. Additionally, by decreasing the amount of Grignard reagent, partial modification of the alkylthio groups has also been achieved, furnishing product 7e with one alkylthio substituent untouched.

Next, the c‑fused furan ring was employed as a diene in Diels-Alder reactions (Table 5).^[18] To start with, the methylthio-substituted cycloalkano[c]furan product 3b was selected as a representative substrate. To our delight, 3b underwent Diels-Alder cycloaddition with N-benzylmalei- mide accompanied by an acid-promoted aromatization^[19] successfully, affording arylated product 8a containing a benzene core. cis-Disubstituted alkenes, monosubstituted alkenes and alkynes participated well in this tandem reaction, affording the corresponding benzoannulated products (8b~8d). A point worth mentioning is that the Diels-Alder reactions with dimethyl maleate or dimethyl acetylenedicarboxylate gave the same product 8d, while the isolated yield with the alkyne was relatively lower. However, reaction with methyl ethenyl sulphone bearing an electron- withdrawing sulfonyl group led to the formation of 8e, which lacks a hydroxyl substituent on the phenyl substituent, indicating an embedded oxidation process by the oxygen in air during aromatization.

In summary, a direct and convenient protocol has been described for the synthesis of cycloalkano[c]furans, which are of high biological interests. The sequence extends intermolecular [3,3]-rearrangements of sulfonium ions to the intramolecular mode, embedded within a bicycle-generat- ing furan transformation. Mechanistically, this reaction was initiated by a Lewis acid-catalyzed intramolecular cross-coupling, followed by a [3,3]-rearrangement of the resulting O-alkenylsulfoxonium ion intermediates and subsequent heterocyclization to furnish cycloalkano[c]- furans. The advantages include the easy access of starting materials, mild conditions with the flexibility on the annulated ring size. Further functionalization of the furan products was achieved by Ni-NHC-catalyzed cross-coupling of the alkylthio groups and Diels-Alder cyclization of furan rings, which enriched the structural diversity of the resulting furan products.