Multicomponent Reaction Discovery: Three-Component Synthesis of Spirooxindoles

Abstract

The Lewis acid catalyzed, three-component reaction of isatin and two 1,3-dicarbonyl compounds is reported. Reactions proceed with high efficiency under mild reaction conditions and with good functional group tolerance to afford spirooxindole pyranochromenedione derivatives.

High-throughput screening is an efficient method for discovery of new reactions and synthesis of complex chemotypes. ¹ We have recently reported several novel Rh(I)-catalyzed transformations identified using reaction screening in which 1,3-dicarbonyl substrates exhibited diversified reactivities as nucleophiles with electrophiles derived from o-alkynylbenzaldehydes.² To further explore modes of reactivity of 1,3-dicarbonyl compounds under Rh(I) catalysis, we have screened their reactions with a diverse set of electrophiles. Our analysis led to the discovery of the selective, catalyzed condensation of N-methylisatin 1 with two molecules of 1,3-cyclohexanedione 2 to give spirooxindole ^{3 , 4 , 5} pyranochromenedione 3 in 46% yield (Scheme 1).

Scheme 1.

Reactions of N-methylisatin and Various 1,3-Dicarbonyls

Similarly, reaction of 4-hydroxy-6-methyl-2-pyrone 4 with N-methylisatin 1 resulted in the production of condensation product 5 in 36% yield (dr = 5:1). The structure of 5 (major diastereoisomer shown) was confirmed by X-ray crystal structure analysis. ⁶ Unexpectedly, when 1, 2, and 4 were added in a ~1:1:1 ratio in the presence of 10 mol % of Rh(cod)₂BF₄ in Cl(CH₂)₂Cl at 60 °C, the three-component coupling product 6 was obtained in 32% after 24 h.⁷ Only trace amounts of 3 and 5 were detected along with recovered starting materials. While we were delighted by the selectivity of this three-component reaction, the reaction proceeded very slowly. The isatin was only fully consumed after prolonged reaction time (3 days) affording spirooxindole 6 in 78% yield.

In order to accelerate the three-component reaction, we next examined additives including Ag(I) salts and observed a significant rate acceleration. Upon further examination, we discovered that the Rh(I) catalyst was not required for this reaction; when AgBF₄ was employed as the catalyst, the yield of 6 was improved to 62% (Table 1, entry 2). Consequently, we suspected that the reaction was operating under Lewis acid catalysis. After screening numerous Lewis acids, we found that SnCl₄⁸ provided the three-component product in good yield after 12 h (Table 1, entry 6), while other catalysts including CuOTf, BF₃•Et₂O and TiCl₄ gave low to moderate yields (Table 1, entries 3-5). ⁹ Other Sn(IV) catalyst systems and desiccants (4Å molecular sieves and MgSO₄) did not improve the overall efficiency of the transformation.

Table 1.

Optimization of Reaction Conditions for the Formation of 6 from 1, 2 and 4

entrya	catalystb	time (h)	yield (%)c
1	Rh(cod)2BF4 (10 mol %)	24	32
2	AgBF4 (10 mol %)	24	62
3	(CuOTf)•C6H6 (10 mo l%)	24	38
4	BF3•OEt2 (10 mol %)	24	65
5	TiCl4 (10 mol %)	24	61
6	SnCl4 (10 mol %)	12	76

^aConditions: Cl(CH₂)₂Cl, 60 °C; 1 (1.0 equiv), 2 (1.1 equiv), 4 (1.1 equiv)

^bThe three-component product is not observed in the absence of catalyst, only 27 (vide infra) is observed

^cYield (UPLC analysis) using biphenyl as an internal standard.

Subsequent experiments led to the finding that we could further accelerate the three-component reaction to produce spirooxindoles using microwave irradiation. In the event, irradiation of 1, 2, and 4 in DCE at 80 °C with 10 mol % of SnCl₄•5H₂O as catalyst afforded an 80% yield of 6 in only 80 min (Table 2, entry 2). As a result of these studies, we examined the reaction with a broad range of substrates to determine the reaction specificity and scope. Various substituted isatins were reacted with 1,3-cyclohexanedione 2 and 4-hydroxy-6-methyl-2-pyrone 4 under both thermal and microwave conditions (Table 2). From the results, it is evident that most of the reactions provided the desired spirooxindole products in good to excellent yields employing both electron-deficient (Table 2, entry 6) and electron-rich (Table 2, entry 4) isatins as substrates. Bromo-substitution on the isatin was tolerated including the sterically demanding 4-bromo derivative. In addition, unprotected isatins also participated in the reaction affording the desired coupling product 7 in moderate yield (Table 2, entry 1).

Table 2.

Three-Component Reaction with Different Isatins

entry	R	R′	protocol	time	product	yield (%) a
1	H	H	A B	16 h 80 min	7	63 78
2	H	Me	A B	12 h 80 min	6	75 80
3	H	Ph	A B	20 h 80 min	8	81 56
4	5-OMe	Me	A B	18 h 80 min	9	81 65
5	5-Me	Me	A B	18 h 80 min	10	89 77
6	5-NO2	Me	A B	18 h 80 min	11	62 85
7	4-Br	Me	A B	24 h 80 min	12	75b 52c
8	5-Br	Me	A B	18 h 80 min	13	90 80
9	6-Br	Me	A B	18 h 80 min	14	93 84
10	7-Br	Me	A B	18 h 80 min	15	65 67

Protocol A: SnCl₄ (10 mol %), Cl(CH₂)₂Cl, 60 °C. Protocol B: SnCl₄•5H₂O (10 mol %), Cl(CH₂)₂Cl, μW, 80 °C.

^aIsolated yields of compounds purified by chromatography on SiO₂

^b20 mol % SnCl₄ at 80 °C

^c20 mol % SnCl₄•5H₂O.

We next examined a number of 1,3-dicarbonyl compounds in the three-component reaction with N-methylisatin 1 and 4-hydroxy-6-methyl-2-pyrone 4. The results listed in Table 3 show that good to excellent yields of coupling products were obtained under the optimized conditions using dimedone, 1,3-cyclopentanedione, 1,3-indanedione, and 1-oxaspiro[5.5]undecane-2,4-dione as coupling partners (Table 3, entries 1-5). In addition, different 1,3-dicarbonyl reagents, substituted pyrone and 4-hydroxycoumarin were reacted with N-methylisatin 1 and 1,3-cyclohexanedione 2. Cyclic substrates afforded the desired spirooxindole products in moderate to good yields (Table 4, entries 1-5). However, reaction of 2,4-pentanedione catalyzed by SnCl₄ (10 mol %) did not result in formation of the desired product. When Cu(OTf)₂ (10 mol %) was used as catalyst, product 26 was obtained in 50% yield (Table 4, entry 6). The less reactive 4-hydroxycoumarin also gave a good yield of the three component coupling product (Table 4, entry 5). However, when 5,5-dimethyl-1,3-cyclohexanedione was reacted with N-methylisatin and 1,3-cyclohexanedione, the yield of the three-component product was low (Table 4, entry 3), due to increased production of the corresponding two-component reaction product.⁶ The yield of the three-component product was diminished when the 1,3-dicarbonyl compounds are similar in pKa.¹¹

Table 3.

Reaction of 4 with Different 1,3-Dicarbonyl Reagents

entry	1,3-dicarbonyl	product	protocol	time	yield (%)a
1			A B	12 h 80 min	93 93
2			A B	16 h 80 min	71 70
3			A B	12 h 80 min	60b,c 39b,c
4			A B	16 h 80 min	68 65
5			A B	16 h 80 min	70 41

Protocol A: SnCl₄ (10 mol %), Cl(CH₂)₂Cl, 60 °C. Protocol B: SnCl₄•5H₂O (10 mol %), Cl(CH₂)₂Cl, μW, 80 °C.

^aIsolated yields of compounds purified by chromatography on SiO₂

^b2.0 equiv 1,3-indanedione

^cSee reference 10.

Table 4.

Reaction of 2 with Different 1, 3-Dicarbonyl Reagents

entry	1 3-dicarbonyl	product	protocol	time	yield (%)a
1			A B	16 h 80 min	75 70
2			A B	16 h 80 min	60e 41e
3			A B	16 h 80 min	51 47
4			A B	16 h 80 min	63b,f 47b,f
5			A B	48 h 4h	84c 65c
6			A B	24 h 2.5h	50b,d 32b,d

Protocol A: SnCl₄ (10 mol %), Cl(CH₂)₂Cl, 60 °C. Protocol B: SnCl₄•5H₂O (10 mol %), Cl(CH₂)₂Cl, μW, 80 °C.

^aIsolated yields of compounds purified by chromatography on SiO₂

^b2.0 equiv 1,3-dicarbonyl reagent was used

^cSolvent: Cl(CH₂)₂Cl/1,4-dioxane, 3:1; 2.5 equiv 4-hydroxycoumarin

^dCu(OTf)₂ (10 mol %) was used as catalyst

^e1.5 equiv of each reagent was used.

^fSee reference 10.

On the basis of these results, we propose two plausible pathways for the reaction (Scheme 2). N-Methylisatin 1 may coordinate with SnCl₄ and react with 2 to afford the aldol adduct 27.¹² Dehydration of 27 by SnCl₄ affords indolenium intermediate 29¹³ which may be subsequently attacked by 4 to furnish 31 (path A).¹⁴ Alternatively, the reaction may be initiated by aldol reaction of 4 with 1 to afford 28, followed by dehydration and nucleophilic attack of 2, to afford intermediate 31 (path B). Prior to addition of SnCl₄, mixing of 1, 2 and 4 provided an equilibrium mixture of 1:27:28 (5:3:2). In addition, in the absence of SnCl₄, mixing 1 and 2 affords a 1:1 equilibrium mixture (1:27) while mixing 1 and 4 affords only an 4:1 mixture (1:28).⁶ The position of the complex equilibrium during the three-component reaction appears to be related to the pKa of the corresponding dicarbonyl compounds such that the aldol adduct of the weaker carbon acid is favored, thereby biasing the reaction towards path A.¹⁵ In combination with the enhanced nucleophilicity of the more acidic 1,3-dicarbonyl compound (vide infra), the three-component reaction pathway thus predominates. Nucleophilic addition of the hydroxyl group of 31 to the adjacent carbonyl group, followed by dehydration catalyzed by SnCl₄, would provide the desired coupling product 6.

Scheme 2.

Proposed Mechanism

To gain insight into the reaction pathways, we independently prepared two of the proposed intermediates along the mechanistic pathway. Intermediates 27 and 32 were independently prepared and converted to 6 under the three-component coupling conditions.⁶ In addition, compound 33, used as a model of 27, was treated with 4 and 2 respectively under the three-component reaction conditions (Scheme 3). The products of this reaction support the formation of indolenium intermediate 36 (similar to the proposed intermediate 29 in Scheme 2). The reaction of pyrone afforded the desired product 34 in good yield, while cyclohexandione afforded only trace amounts of 35. These results suggest that hydroxypyrone 4 is a much better nucleophile for indolenium electrophiles such as 36 in comparison to 2. By analogy, indolenium ions formed in our three-component reaction should react more quickly with 4 than 2.

Scheme 3.

Reactions via an Indolenium Intermediate

In addition, time course experiments indicate that intermediate 27 is formed as the primary product upon mixing 1, 2, and 4.⁶ Upon addition of SnCl₄, disappearance of 27 corresponds to the appearance of the three-component product 6. On the basis of these results, in combination with the indolenium reactivity profile (Scheme 3), we propose that Path A (Scheme 2) is the major path for the formation of the three-component reaction products such as 6. By monitoring the reaction using UPLC, kinetic profiles of the model reaction were obtained and clearly demonstrate the predominance of three-component product 6 compared to two-component products 3 and 5 throughout the reaction (Figure 1).

Figure 1.

Kinetic Plot of the Three-Component Reaction

^a UPLC yields with biphenyl as internal standard (λ= 254 nm).

In summary, using a reaction screening approach we have identified the Lewis acid-catalyzed, three-component coupling of substituted isatins and two pKa differentiated 1,3-dicarbonyls to synthesize various spirooxindoles bearing a pyranochromenedione ring system. Mechanistic studies indicate the likely involvement of an indolenium ion derived from the isatin and one 1,3 dicarbonyl component. Further studies to intercept this intermediate are currently underway and will be reported in due course.