Stereoselective Synthesis of Spirocyclic Oxindoles via Prins Cyclizations

Abstract

The synthesis of spirocyclic oxindole pyran and oxepene frameworks using highly stereoselective Prins cyclizations of homoallylic and bis-homoallylic alcohols and isatin ketals is described.

Exo-methylene pyrans are present in a variety of biologically active natural products.¹ We recently employed the intramolecular silyl-modified Sakurai (ISMS) reaction² to construct the exo-methylene pyran subunit of the macrocyclic core of (-)-zampanolide (1, Figure 1a).³ We envisioned use of this powerful methodology to access a variety of exo-methylene tetrahydropyrans (3 and 4, Figure 1b) using diverse ketals and acetals (5 and 6, Figure 1b) for diversity-oriented synthesis⁴ and chemical library development (Figure 1b). As part of our studies, we also considered preparation of pyran spirooxindole hybrid molecules⁵ (7, Figure 1b) from allylsilanes 8 and isatin ketals 9 in order to merge fragments of two biologically interesting motifs. The spirooxindole core structure is represented in numerous pharmacological agents and alkaloids⁶ including the anticancer agent MI-63 (2, Figure 1a).⁷ In this communication, we report how our initial synthesis plan evolved to identify stereoselective, Lewis acid-mediated Prins cyclizations to access both enantioenriched spirocyclic oxindole pyrans and oxepenes.

Figure 1.

a) Representative exo-methylene pyran spirooxindole molecules. b) Initial synthesis plan.

In initial studies, allyl silane 12 and derived silyl ether 13 (Scheme 1) were prepared employing Cu(I)-catalyzed⁸ ring-opening of chiral, non-racemic epoxide 11 with vinyl Grignard 10.⁹ After optimization of ISMS reaction conditions, ketals 14 and acetal 15¹⁰ were successfully employed in the synthesis of spirocyclic exo-methylene pyran 16 and 2,6-syn-disubstituted pyran 17¹¹ (Table 1, entries 1 and 2). Subjection of N-methyl isatin dimethylketal 18¹² (Table 1) and allyl silane 13 to optimized ISMS reaction conditions (Table 1, entry 3) did not afford the desired spiroannulated product 19 and led only to recovery of silyl ether¹³ 20 (Scheme 2) and N-methyl isatin. Warming of the reaction to room temperature afforded product 21, presumably from direct allylation of the derived isatin oxonium ion with allylsilane 13 (Table 1, entry 4). Attempts to convert compound 21 into the desired spiroannulated product under acidic conditions afforded a complex mixture of products. When CH₂Cl₂ was used as solvent, spirooxindole 22 bearing an endocyclic alkene¹⁴ was isolated in moderate yield. A control experiment employing ketal 14 in the ISMS reaction with CH₂Cl₂ as a solvent afforded product 16 and its endocyclic olefin isomer (1:2 ratio).

Scheme 1.

Preparation of allyl silane 13

Table 1.

Use of allylsilane 13 in ISMS reactions

entry	acetal/ketal	T[°C]	solvent	product	yield (%)a
1	14	−78	Et2O	16	80
2	15	−78	Et2O	17	84 dr>20:1
3	18	−78	Et2O	19	<2b
4	18	−78 to rt	Et2O	21	63 dr= 1:1
5	18	−78 to rt	CH2Cl2	22	36

Reaction conditions: allylsilane 13 (1.1 equiv), TMSOTf (0.3 equiv), 2,6-tBu-4-Me-pyridine (DBMP) (0.05 equiv), 4 Å MS, 14 h.

^aIsolated yields.

^bNot observed, N-methyl isatin and desilylated allyl silane 20 recovered

Scheme 2.

Spirooxindole pyrans via Prins cyclizations

Based on these observations and given that spiroannulation did not occur at low temperature (−78 °C), we performed the reaction at higher temperatures which afforded spirooxindoles 22 and 23 in good overall yield (Scheme 2, entry 1). The apparent isomerization of the double bond (endocyclic vs exocyclic olefin) suggested the possibility of a mechanistic pathway which was different than the expected ISMS reaction. Intramolecular Prins-type cyclization¹⁵ of homoallylic silyl ether 20 derived from desilylation of 13 could account for such an outcome. To support this hypothesis, homoallylic silyl ether 20 was synthesized and employed in the reaction to afford products 22 and 23 in good overall yield (Scheme 2, entries 2 and 3). An additional spirooxindole pyran synthesis sequence is shown in Scheme 3. Preparation of homoallylic alcohol 24¹⁶ by epoxide ring-opening, followed by treatment with isatin ketal 18 and TMSOTf, afforded spirooxindoles 27 and 28 (Scheme 3). The relative stereochemistry and alkene position of major stereoisomer 27 were confirmed by x-ray crystallographic analysis.¹¹

Scheme 3.

Diastereoselective synthesis of spirooxindoles

In order to explain the stereochemical outcome of the Prins cyclizations, we propose a chair transition state (Figure 2) in which the larger aryl substituent of the oxindole moiety adopts a pseudo-equatorial orientation¹⁷ (TS-1) leading to the observed diastereoisomer (cf. 27 and 28, Scheme 3). An alternative chair (TS-2) leading to the disfavored diastereoisomer has significant steric interactions between the isatin carbonyl oxygen and the R substituent on the chiral center. Examination of molecular models of the proposed intermediate tertiary carbocations 29 and 30 obtained using Spartan conformational searches (AM1) followed by DFT minimization (performed using a 6-31G* basis set; E = 8.25 Kcal/mol)¹¹ shows destabilizing 1,3-diaxial interactions in carbocation 30 which is derived from Prins cyclization through TS2.

Figure 2.

Proposed transition states.

In order to confirm the relative stereochemistry at the spiro center of minor regioisomer 23 generated during the Prins cyclization, we subjected both regioisomers 22 and 23 to metal-catalyzed hydrogenation. Interestingly, using catalytic amounts of Pd/C, a mixture of chromatographically separable diastereoisomers 31 and 32 were observed by ¹H NMR analysis of crude samples, indicating that regioisomers 22 and 23 had the same relative stereochemistry at the spiro center (Table 2).

Table 2.

Amide-directed hydrogenation

spirooxindole	catalyst	H2	solvent	conversiona(%)	ratioa
22	Pd/C	50 psi	MeOH	96	2:1
23	Pd/C	50 psi	MeOH	99	7:1
22	Wilkinson’s catalyst RhCl(PPh3)3	50 psi	EtOH/Benzene	-	-
22 + 23	Crabtree’s catalyst [Ir(cod)py(PCy3]PF6	1 atm	CH2Cl2	99	only 32

^aConversion and ratios of 31:32 determined by ¹H NMR analysis of crude samples.

In light of the poor diastereoselectivity observed using standard hydrogenation conditions, we next evaluated the possibility of amide-directed hydrogenation.¹⁸ While use of Wilkinson’s catalyst did not generate the desired hydrogenated product, use of Crabtree’s catalyst¹⁹ led to the production of 32 in excellent diastereoselectivity (dr > 30:1) indicating complete substrate control in the amide-directed hydrogenation (Table 2).

In order to broaden the scope of the methodology to access spirocyclic oxindoles, we prepared a series of homoallylic alcohols (24, 33–36) and isatin ketals (18, 37,²⁰ 38) for examination in the Prins cyclization (Table 3). Cyclizations were found to be successful with isatin ketals bearing NH functionality to afford spirooxindole products 39–43. Introduction of a bulky bromine substituent on the 4-position of the isatin ketal (Table 3, entries 2, 4 and 5) resulted in improved diastereoselectivity and noticeably influenced the product olefin regiochemistry (cf. entries 3 and 4), which may be explained by highly regioselective elimination of a carbocation intermediate distal from the bromo-oxindole moiety (cf. 29, Figure 2).

Table 3.

Prins-type spiro-annulation

entry	alcohol	isatin	product	yields (%)a, f
1	33	37	39	78b dr=13:1 rr = 2.6:1 ee > 99%
2	34	38	40	70c dr>20:1 rr=13:1 ee > 99%
3	35	37	41	85b, d dr> 10:1 rr = 4.3:1 ee > 99%
4	36	38	42	88c dr>20:1 rr=10:1 ee > 99%
5	24	38	43	78b dr = 20:1 rr = 10:1 ee > 99%

^aIsolated yields.

^bReaction conditions: TMSOTf (1.0 equiv), −40 °C to rt, 14 h, CH₂Cl₂.

^cTMSOTf (1.0 equiv), −40 °C to 0 °C, 3 h, CH₂Cl₂.

^dIsolated as an inseparable mixture of regioisomers. Hydrogenation was necessary to facilitate products separation.

^frr = regioisomeric ratio; major isomer shown.

Considering the well-documented racemization observed during Prins cyclization due to competitive oxonia-Cope rearrangement (Figure 3),²¹ we also measured the enantiomeric excess of the spirocyclic products. In all cases, we did not observed erosion in enantiopurity (Tables 3 and 4). These findings are consistent with the observations that stabilization of the intermediate tetrahydropyranyl cation raises the transition states energy for ring-opening and effectively eliminates the oxonia-Cope rearrangement.21a, b

Figure 3.

Possible racemization of Prins cyclization products via 2-oxonia-Cope rearrangement.

Table 4.

Diastereoselective synthesis of spirooxindole oxepenes

entry	R	R1	R2	product	yields (%)a, b
1	OPh	H	Me	46	51 dr>20:1 rr = 10:1 ee > 99%
2	OPh	H	H	47	41 dr>20:1 rr = 10:1 ee > 99%
3	OPh	Br	H	48	43 dr>20:1 rr = 13:1 ee > 99%
4	H	H	H	49	56 dr>20:1 rr = 16:1 ee > 99%
5	H	Br	H	50	57 dr>20:1 rr = 20:1 ee > 99%

^aIsolated yields

^brr = regioisomeric ratio; major isomer shown. Reaction conditions: TMSOTf (1.0 equiv), −40 °C to rt, 3 h, CH₂Cl_2.

Finally, we extended the methodology to intramolecular Prins cyclization of bis-homoallylic alcohols (Table 4, 44, 45²²) and isatin ketals (Table 4, 18, 37, 38) to generate spirocyclic oxindole oxepenes^{14b, 23} 46–50 in high diastereo- and regioselectivity (Table 4). The relative stereochemistry of spirooxindole 46 was confirmed via X-ray crystallographic analysis.¹¹

In conclusion, enantiopure spirocyclic oxindole pyransand oxepenes have been efficiently synthesized by highly stereoselective Prins-type cyclizations of both homoallylic and bis-homoallylic alcohols and isatin ketals. The protocol is highly complementary to related annulations involving chiral organosilanes.²⁴ Further studies involving related annulations and library synthesis applications are in progress and will be reported in due course.