FPDs are highly reactive compounds [25,26] containing an highly electrophilic enone fragment which facilitates the course of cycloaddition and nucleophilic addition reactions. In recent years, some types of cycloaddition reactions were investigated for FPDs: the [4 + 2] cycloaddition with alkenes resulting in pyran-annulated products [27-34] and the [3 + 2] cycloaddition with nitrones resulting in isoxazole-annulated products [35-37] (Scheme 2). However, formal [4 + 1] cycloaddition reactions for FPDs remain to be unknown.

Scheme 2: Cycloaddition reactions of [e]-fused 1H-pyrrole-2,3-diones.

To evaluate the possibility of synthesizing the target spirooxindole compounds, we initially investigated a reaction of benzoxazine-containing FPD 1a with diazooxindole 2a in anhydrous acetonitrile at room temperature (Scheme 3). The reaction came to an end in 24 hours, with the color of the solution being turned from purple to red. The starting FPD 1a is bright violet; thus, the disappearance of the violet color was used as an indicator of the reaction’s completion. Product 3aa was isolated as yellow crystals in 73% yield and characterized by NMR, IR, and mass spectra, and single crystal X-ray analysis (CCDC 2201614). As evinced by the NMR data, only one diastereomer of product 3aa was obtained. Contrary to the isoxazole-annulated products of a [3 + 2] cycloaddition of nitrones to FPDs [35], product 3aa appears to be stable on storage in solution, which was confirmed by the fact that the NMR data remained unvaried after keeping the product in solution for one day.

Scheme 3: The model reaction of FPD 1a and diazooxindole 2a.

Next, the conditions (Table 1) of the model reaction of FPD 1a and diazooxindole 2a were optimized.

Table 1: Reaction of FPD 1a and diazooxindole 2a in different solvents.a

Entry Solvent Temperature, °C Yieldb, % 1 toluene 25 3 2 chloroform 25 48 3 ethyl acetate 25 29 4 1,4-dioxane 25 9 5 DMSO 25 32 6 acetonitrile 25 82 7 acetonitrile 25 73c 8 acetonitrile 83 50c

aA suspension of FPD 1a (100 µmol, 32.0 mg) and diazooxindole 2a (100 µmol, 16.0 mg) in the corresponding solvent (1 mL) was stirred in an oven-dried closed microreaction vial for 24 hours; bUPLC yield (the chromatograms were recorded immediately after sample preparation); cisolated yield.

The best yield of product 3aа (Table 1, entries 6 and 7) was obtained by the reaction performed in acetonitrile at room temperature, therefore, these conditions were taken as a standard for further reactions.

Next, the reagent scope of the reaction was explored by involving diazooxindoles 2a–d into the reaction with FPD 1a (Table 2).

Table 2: Reaction of FPD 1a and diazooxindoles 2a–d.a

Entry Diazooxindole Product R1 R2 Time, hb Yieldc, % drd 1 2a 3aa H H 24 73 >99:1 2 2b 3ab H Br 168e –e – 3 2b 3ab H Br 6f 60 >99:1 4 2c 3ac H OMe 24 61 >99:1 5 2d 3ad Bn H 24 58 50:1

aA suspension of FPD 1a (500 µmol, 160.0 mg) and diazooxindole 2a–d (500 µmol) in acetonitrile (3 mL) was stirred in an oven-dried closed microreaction vial for the given time; breaction time was monitored by the disappearance of the dark violet color of FPD 1a; cisolated yield; dratio was determined by 1H NMR in isolated product; eno disappearance of dark violet color of FPD 1a, the reaction was monitored by UPLC-MS (the reaction not completed within a week); fthe reaction was carried out in refluxing solvent.

Compared to substrate 2a, the presence of substituents in diazo compounds 2b–d led to decreased reaction yields (Table 2, entries 2–5). The reaction with diazooxindole 2b having an electron-withdrawing group (-Br) in the C(5) position (Table 2, entries 2 and 3) required additional heating to obtain the product 3ab. On the other hand, the reaction of diazooxindoles 2c and 2d (Table 2, entries 4 and 5) bearing an EDG (electron-donating group) in positions N(1) (Bn-) or C(5) (MeO-) did not require heating and proceeded under conditions similar to the ones with unsubstituted diazooxindole 2a.

Next, we investigated the substrate scope using different FPDs 2 (Table 3).

Table 3: Reaction of FPDs 1a–j and diazooxindole 2a.a

Entry Product R1 R2 X Time, hb Yieldc, % drd 1 3aa H Ph O 24 73 >99:1 2 3ba H C6H4Cl-4 O 24 71 >99:1 3 3ca H C6H4Br-4 O 24 62 >99:1 4 3da H C6H4Me-4 O 24 73 >99:1 5 3ea H C6H4OMe-4 O 24 78 >99:1 6 3fa H C6H4NO2-4 O 24 48 >99:1 7 3ga Cl Ph O 24 86 >99:1 8 3ha H C6H4Cl-4 NH 6e 70 50:1 9 3ia H Ph N–Ph 6e 85 >99:1 10 3ja H C6H4Cl-4 N–Bn 6e 63 >99:1

aA suspension of FPD 1a–j (500 µmol) and diazooxindole 2a (500 µmol, 80.0 mg) in acetonitrile (3 mL) was stirred in an oven-dried closed microreaction vial; breaction time was monitored by the disappearance of the dark violet color of FPD 1; cisolated yield; dratio was determined by 1H NMR in isolated product; ethe reaction was carried out in refluxing solvent.

FPDs 1а–f successfully reacted with diazooxindole 2а under the previously developed conditions and gave good product yields (Table 3, entries 1–6). Neither the yield, nor the reaction rate were observed as being markedly affected by electron-donating or weak electron-withdrawing groups present in the aroyl substituent of FPDs 1. However, the presence of a strong electron-withdrawing group (–NO2) in the aroyl fragment of the FPD 1f, significantly decreased the yield of the target product 3fa (Table 3, entry 6). The introduction of an electron-withdrawing group (–Cl) into the benzoxazine fragment of FPD 1g increased the yield of the target reaction product, without affecting the reaction rate (Table 3, entry 7). Quinoxaline-annulated FPDs 1h–j required heating, as these compounds reacted too slowly at room temperature (Table 3, entries 8–10). It should be noted that FPDs 1h–j gave yields of the target products close to that of the products obtained from FPDs annulated with a benzoxazine fragment. The structures of products 3aa, 3ab, and 3ha were approved by single crystal X-ray analysis (CCDC 2201614, CCDC 2201616, CCDC 2201615).

We also decided to study the effect the benzo-annulated moiety in FPDs has on inducing the reaction. Under the same conditions, FPD 1k containing a morpholine fragment (Scheme 4) was involved in the reaction with 2a which gave the expected product 3ka in a fairly good yield of 56% and dr 99:1. The characteristic signals in the NMR spectra of the products 3aa and 3ka appeared to be the same; thus, the structure of product 3ka was ascertained to be similar to that of product 3aa.

Scheme 4: The reaction of FPD 1k with diazooxindole 2a.

With the above observations and the reported literature [24] as a basis, the formation of spirofuranoxindoles 3aa–ka was assumed as proceeding via two stages: (a) the nucleophilic Michael attack of the negatively charged [38] C(3) atom of diazooxindoles 2 at the C(3a) atom of FPDs 1 (Scheme 5), and (b) further intramolecular SN2 attack by the oxygen of the aroyl group with ensuing elimination of a nitrogen molecule. To verify our assumption, 3-bromooxindole (4) was involved in the reaction with FPD 1i in the presence of 1.1 equiv of TEA. In this case, the base-promoted deprotonation of 3-bromooxindole (4) affords a highly nucleophilic intermediate, which undergoes Michael addition to FPD 1i, followed by intramolecular SN2 attack [39-43] by the oxygen of the aroyl group (Scheme 5) to give the same diastereomer 3ia with a good 54% yield.

Scheme 5: A) Plausible mechanism of formal [4 + 1] cycloaddition of FPDs 1 with diazooxindoles 2 (negative charge delocalization is colored in blue); B) plausible base-promoted reaction mechanism of FPD 1i and 3-bromooxindole (4, negative charge delocalization is colored blue).