Double Diastereoselective, Nucleophile-Catalyzed Aldol Lactonizations (NCAL) Leading to β-Lactone Fused Carbocycles and Extensions to β-Lactone Fused Tetrahydrofurans

Abstract

A double diastereoselective variant of the nucleophile catalyzed aldol-lactonization (NCAL) process is described. This strategy delivers β-lactone-fused carbocycles with good to excellent diastereoselectivities using cinchona alkaloid catalysts with enantioenriched aldehyde acids, which gave low diastereoselectivity based on substrate control alone. β-Lactone fused tetrahydrofurans are also prepared for the first time via the NCAL process, however diastereoselectivity was only modestly improved when applying double diastereodifferentiation to these systems.

The significant utility of double diastereodifferentiation¹ has been demonstrated on numerous occasions in the context of natural product total synthesis. Therefore, the importance of this strategy cannot be overstated.² Furthermore, the degree to which a chiral reagent can influence stereochemical outcomes by overcoming inherent substrate bias is revealed through the study of double diastereodifferentiation.

The development of concise enantioselective routes to β-lactones continues to be an active area of research due to their varied reactivity and applications.³ Our recent contributions to this area have focused on intramolecular, nucleophile catalyzed aldol-lactonization (NCAL) processes of aldehyde acids that provide convenient access to bicyclic β-lactones bearing two or more stereogenic centers (Scheme 1a)⁴ building on elegant work by Wynberg and coworkers.⁵ This process was previously rendered enantioselective with the use of O-acetyl quinidine (O-AcQD) or β-isocupreidine (β-ICPD) as chiral nucleophilic promoters (chiral Lewis bases) to obtain enantioenriched β-lactone fused cyclopentanes.^4,6 More recently, the NCAL process was extended to ketoacid substrates⁷ (Scheme 1b), and a single example of a stoichiometric, enantioselective reaction made use of tetramisole as a chiral promoter.⁸ In addition, the utility of the NCAL process was demonstrated in a concise, bioinspired, racemic⁹ and subsequently enantioselective¹⁰ total synthesis of salinosporamide A. Herein, we demonstrate the utility and powerful stereochemical influence of cinchona alkaloid catalysts in double diastereoselective NCAL reactions for the preparation of variously substituted carbocycle-fused β-lactones. We also report the first examples of tetrahydrofuran-fused β-lactones obtained through the NCAL process.

Scheme 1.

Formation of Bicyclic and Tricyclic β-Lactones via the NCAL Process From (a) Aldehyde Acids (1) and (b) Ketoacids (4)

We previously found that, with respect to the carboxylic acid, β-substituted aldehyde acid and ketoacid substrates provided bicyclic β-lactones (i.e. 6) with high diastereoselectivity (Figure 1). This observation provided evidence for a NCAL process proceeding via ammonium enolate intermediates based on A^1,3-strain arguments, since a [2+2] pathway proceeding through a ketene intermediate would be expected to afford low diastereoselectivity. However, substrates bearing substituents at other positions (i.e. γ, δ) gave low diastereoselectivities as would be predicted based on the absence of A^1,3 strain.⁷ This led us to consider double diastereodifferentiation with chiral nucleophiles, O-TMS quinidine (O-TMSQD) and O-TMS quinine (O-TMSQN),¹¹ with enantioenriched substrates to determine if catalyst control could override the low diastereoselectivities obtained with substrate control alone (Figure 2).

Figure 1.

Observed diastereoselectivites for NCAL reactions leading to β-lactone fused carbocycles 6-8 and proposed selectivity models.

Figure 2.

Optically active nucleophiles (Lewis Bases) employed in the NCAL process.

We initiated double diastereodifferentiation studies with previously studied carbocyclic substrates. Subjecting enantiomerically enriched aldehyde acid (+)- 9a¹² (87% ee, chiral HPLC) to standard NCAL conditions with Et₃N as the nucleophile resulted in a 2:1 mixture of anti/syn β-lactones 10a:10a’, respectively (Table 1, entry 1). Use of 10 mol % O-TMSQD led to an increased level of diastereoselection and complete reversal in diastereoselection to 1:7 anti/syn β-lactones 10, and both relative and absolute stereochemistry of the major β-lactone 10a’ was confirmed by X-ray analysis (Table 1, entry 2).¹³ Alternatively, diastereomeric β-lactone 10a’ could be obtained with high diastereoselectivity (dr >19:1) employing O-TMSQN indicative of the matched case (Table 1, entry 3). Commercially available dimeric catalysts, hydroquinidine 1,4-phthalazinediyl diether ((DHQD)₂PHAL) and hydroquinine 1,4-phthalazinediyl diether ((DHQ)₂PHAL), were also studied and provided similar results (Table 1, entries 4 and 5).

Table 1.

Catalyst Screening for Double Diastereodifferentiation with Aldehyde Acid (+)-9a

entry	nucleophile (mol %)	base	% yielda	dr (10a:10a’)b
1	Et3N (100)c	Et3N	58	2:1
2	OTMSQD (10)d	i-Pr2NEt	73	1:7
3	OTMSQN (10)d	i-Pr2NEt	60	>19:1
4	(DHQD)2PHAL (10)d	i-Pr2NEt	59	1:4
5	(DHQ)2PHAL (10)d	i-Pr2NEt	53	>19:1

^aIsolated yield.

^bDiastereomeric ratio determined by ¹H NMR (500 MHz) of the crude reaction mixture.

^cReaction was run for 48 h at 0.05 M.

^dReaction was run for 72 h at 0.2 M.

Since double diastereoselectivity was possible with γ-substituted aldehyde acid substrates, we next studied other enantiomerically enriched substrates with alternate substitution patterns. As previously reported and to serve as a reference point, both enantiomers of the unsubstituted bicyclic β-lactone 10b:10b’ were obtained with good enantioselectivity employing the pseudoenantiomeric catalysts, O-AcQD and β-ICPD, in 92% and 90% ee, respectively.^4a This example demonstrates the potential for significant catalyst control in the NCAL process of aldehyde acids. As expected, anti-silyloxy-β-lactone 10c’ derived from aldehyde acid 9c was obtained with high diastereoselectivity (dr >19:1) with Et₃N as the nucleophilic promoter due to the β-substituent (Table 2, entry 2). This substrate also provided excellent efficiency. However, not surprisingly, diastereoselectivity could not be altered with either O-TMSQD or O-TMSQN due to the strong conformational bias exerted by allylic 1,3-strain (see Figure 1) but rather led only to greatly reduced conversion. Use of Et₃N with the γ,δ-substituted aldehyde acid 9d gave low diastereoselectivity leading to a 2:1 mixture of β-lactones 10d:10d’ (Table 2, entry 3). Reversed diastereoselectivity (dr 1:3) was obtained with O-TMSQD providing β-lactone 10d’ as the major diastereomer but gave low yield (32%). However, use of O-TMSQN gave both improved yields and diastereoselectivity (dr 10:1) with β-lactone 10d as the major diastereomer suggestive of a matched case. Double diastereodifferentiation was also possible with cyclohexyl-fused β-lactones 10e:10e’, which improved a 2:1 diastereomeric ratio obtained with Et₃N to complete catalyst control with O-TMSQN leading to high diastereoselectivity (>19:1); however, conversions were low (Table 2, entry 4).

Table 2.

Double Diastereoselective NCAL Reactions with Enantioenriched Aldehyde Acids 9b-e

entry	carbocycle fused bicyclic β-lactones	NEt3 % yield (dr)a,b	O- TMSQD % yield (dr)b,c	O- TMSQN % yield (dr)b,c
1d		55	82, 92% eee	42, 90% eef
2		84 (1:>19)	33 (1:>19)	23 (1:>19)
3		45 (2:1)	32 (1:3)	55 (10:1)
4		38 (2:1)	31 (1:>19)	10 (>19:1)

^aReaction run for 12-24 h at 0.05 M.

^bYields refer to isolated, purified products and diastereomeric ratios were determined by ¹H NMR (500 MHz) analysis (integration) of the crude reaction mixtures.

^cReaction run for 48-72 h at 0.20 M.

^dPreviously described (see refs. ^4b,c) enantioselective NCAL process with an achiral substrate shown for comparison.

^eReaction run with O-AcQD.

^fReaction run with β-ICPD.

In the context of a total synthesis effort, we explored the NCAL process for the synthesis of tetrahydrofuran-fused β-lactones¹⁴ (e.g. 12). Initial studies toward these bicycles provided access to racemic tetrahydrofuran- and tetrahydropyran-fused β-lactones 12a-d:12a’-d’ (Table 3). However, after extensive optimization studies, only low yields were obtained using Et₃N leading to γ- or δ-substituted tetrahydrofurans along with attendant low diastereoselectivies as expected in analogy to cyclopentyl systems (Table 3, entries 1, 2). More tractable ketoacid substrates (e.g. 11, R = Me) delivered the highest yields with the use of 4-pyrrolidinopyridine (PPY) leading to β-lactones 12c:12c’ as a 1:1 mixture of diastereomers (Table 3, entry 3). The relative stereochemistry of β-lactone 12c’ was confirmed by single crystal X-ray analysis (Figure 3). In the case of tetrahydropyran-fused β-lactone 12d, high diastereoselectivity was observed, however, in low yield (Table 3, entry 4). The presence of an oxygen atom in the tether likely alters reactive rotamer populations and a change in mechanism is possible, especially with α-oxygenated acid substrates leading to ketene intermediates e.g. [2+2] cycloadditions, which may account for lower yields in these reactions.

Table 3.

Tetrahydrofuran- and Tetrahydropyran-fused β-Lactones 12 via the NCAL Process

entry	tetrahydrofuran bicyclic β-lactones	% yield (dr)a
1		25 (1:1)
2		31b (2:1)
3		54c,d (1:1)
4b		26 (>19:1)

^aYields refer to isolated, purified products and diastereomeric ratios were determined by ¹H NMR analysis (integration) of the crude reaction mixtures.

^bReaction was run at −30 °C.

^cPPY (1.5 equiv) was used as nucleophile and i-Pr₂NEt (2.0 equiv) as base.

^dReaction run at 0 °C.

Figure 3.

Single crystal X-ray structure (ORTEP representation) of tetrahydrofuran-fused β-lactone 12c’.

While cinchona alkaloid catalysts proved useful for double diastereodifferentation leading to carbocyclic systems, they were less successful when applied to aldehyde acids bearing oxo-linkages (Scheme 2). For example, racemic aldehyde acid 11e provided 47% yield of a 1:1 mixture of diastereomeric tetrahydrofuran-fused β-lactones 12e:12e’ employing Et₃N.¹⁵ Attempted double diastereodifferentation with O-TMSQD and racemic gave low enantioselectivity for both diastereomers 2:1) indicating that effective catalyst control did not occur with this substrate and thus diastereoselectivity would not be expected to improve when employing enantioenriched substrate. Similar results were obtained with O-TMSQN (not shown). These initial results obtained with cinchona alkaloid catalysts along with the consistently low yields of tetrahydrofuran-fused β-lactones obtained via the NCAL process, despite extensive optimization studies, precluded further investigations of double diastereodifferentiation with this class of substrates.

Scheme 2.

Attempted Double Diastereoselective Synthesis of Tetrahydrofuran-Fused β-Lactones 12e/12e’

In summary, double diastereodifferentiation via the NCAL process is possible with cinchona alkaloid catalysts and enantioenriched aldehyde acids. In particular, carbocycle-fused β-lactones were highly amenable to double diastereodifferentiation leading to improvements in diastereoselectivities from 1:1-2 to >19:1 in several cases. This process thus enables access to highly functionalized carbocycles with existing stereocenters with high diastereoselectivity. Tetrahydrofuran-fused β-lactones were accessed for the first time via the NCAL process; however, prospects for double diastereodifferentiation were precluded due to low reaction yields. These studies reveal the exquisite stereochemical control exerted by the cinchona alkaloids in the NCAL process given the ability of these catalysts to override inherent substrate bias and in some cases reverse diastereoselectivity obtained from substrate control alone.