Enantioselective Catalysis of an Anionic Oxy-Cope Rearrangement Enabled by Synergistic Ion Binding

Abstract

Charge-accelerated rearrangements present interesting challenges to enantioselective catalysis, due in large part to the competing requirements for maximizing reactivity (ion-pair separation) and stereochemical communication. Herein, we describe application of a synergistic ion-binding strategy to catalyze the anionic oxy-Cope rearrangement of a symmetric bis-styrenyl allyl alcohol in up to 75:25 e.r. Structure–reactivity–selectivity relationship studies, including linear free-energy-relationship analyses, with bifunctional urea catalysts indicate that H-bonding and cation-binding interactions act cooperatively to promote the chemo- and enantioselective [3,3]-rearrangement. Implications for catalyst designs applicable to other transformations involving oxyanionic intermediates are discussed.

1. Introduction

Since Cope and Hardy’s discovery of the thermal isomerization of 1,5-dienes,^[1] the pericyclic transformation that came to bear Cope’s name has come to epitomize the mechanistic intrigue and synthetic potential of sigmatropic rearrangements more generally (Scheme 1A).^[2,3] The [3,3]-sigmatropic rearrangement is among the principal pericyclic reactions taught in introductory organic chemistry courses as an illustration of orbital symmetry arguments^[4] and conformational analysis^[5] for predicting reaction outcomes and product stereochemistry. Although extensive experimental and theoretical analyses of the all-hydrocarbon Cope rearrangement have inspired profound mechanistic insights^[6] and vigorous debates,^[7] the reaction is typically of only limited synthetic utility in unbiased systems due to its reversibility.

Scheme 1.

Precedent and Design Strategy

The utility of the Cope rearrangement is greatly expanded through the introduction of a hydroxyl group at the C3 position of the 1,5-diene, which provides significant thermodynamic driving force for the reaction along with an avenue for the formation of a versatile δ,ε-unsaturated carbonyl product (the oxy-Cope rearrangement).^[8] Evans’s discovery that extraordinary rate accelerations in the range of 10¹⁰ to 10¹⁷ -fold could be achieved upon deprotonation of 1,5-hexadiene-3-ols, thereby enabling efficient reactions at even cryogenic temperatures, had a transformative impact on the synthetic utility of the transformation.^[9-11] Subsequently, the “anionic oxy-Cope” rearrangement has been applied broadly and with great creativity as a reliably diastereoselective method for C–C bond-formation in target-oriented synthesis (Figure 1B).^[12-15]

Figure 1.

Differential Linear Free-Energy Relationships. All reactions conducted under the standard conditions shown in Scheme 3 with catalysts 12a–m varied at the aryl pyrrolidine moiety. Dashed lines represent the linear regressions to the blue data points with the parameters shown. ^a X = H (in red) is excluded due to steric differences. The effective σ_p value for 3,4,5-trifluorphenylbenzene was approximated as: σ_p,effective = σ_p(F) + 2 σ_m(F) = 0.736. ^b The angle, θ, was calculated from the optimized structure of the corresponding 2-arylpyrrolidine acetamide optimized using density functional theory: B3LYP/6-31G(d).

The basis for the remarkable efficiency of the anionic oxy-Cope rearrangement has been the subject of significant analysis. The rate acceleration can be attributed, in part, to a HOMO-raising effect, wherein the anion is stabilized through extensive delocalization across the rearrangement transition structure relative to the charge-localized ground state.^[16] Additional experimental and computational studies provide evidence that overlap between the high-lying oxyanion nonbonding (n_O) orbital and the adjacent σ*_CC (or the lowest-lying antibonding molecular orbital, ψ⁴_π) in the substrate weakens the C–C bond and increases the dissociative character of the rearrangement (Figure 1B).^[16d,17,18] In this extreme, the transition structure is best described as a weakly interacting complex between an α,β-unsaturated carbonyl compound and an allyl anion. This model accounts for the rate-enhancement achieved with crown complexes or other Lewis basic additives that increase the ionic character on O. However, hyper-basic or polar conditions also increase the propensity for substrate decomposition through dissociative pathways.^[9b,19] While the diastereochemical preferences of the rearranging anion have been the focus of significant study,^{[9b,12,13,20]} little attention has been granted to the role of the counterion since Evans’s seminal studies.^[21,22] The extent and geometrical features of cation participation in the rearrangement transition structure, and their influence on the rate of rearrangement relative to decomposition, remain unclear.

The importance of electrostatic interactions in modulating reactivity and selectivity in both enzymes^[23] and abiological catalyst systems is well appreciated.^[24-28] Recently, we reported a synergistic ion-binding strategy for enantioselective catalysis of the [2,3]-Wittig rearrangement.^[28] In this system, a polyfunctional chiral thiourea catalyst simultaneously engages both the reactive anion and its cesium countercation to promote stereoselectivity-determining rearrangement. On the basis of experimental and computational results, enantioinduction was attributed to selective transition-state attraction between Cs⁺ and the migrating allyl anion. Given the mechanistic similarities between the two charge-accelerated sigmatropic rearrangements,^[11,16,18b] we considered whether a comparable strategy might be applied to the catalysis of a desymmetrizing anionic oxy-Cope rearrangement (Figure 1C). To our knowledge, no catalytic, enantioselective variants of the oxy-Cope rearrangement have been reported.^[29-31]

Synergistic binding of catalyst to both the alkoxide and the metal counterion was anticipated to provide a basis for enhanced stereochemical communication in the rearrangement process, but the implications on reactivity were less certain.^[32,33] Whereas the rate of the [2,3]-Wittig rearrangement was enhanced by H-bonding interactions with the incipient alkoxide, the anionic oxy-Cope rearrangement would be expected to experience deceleration as a result of H-bonding interactions. Consequently, we anticipated that rate-acceleration due to the catalyst would necessarily rely on the stabilizing effects of simultaneous cation-binding to compensate for H-bonding interactions necessary for substrate recognition.^[34] Product/decomposition and enantiomer ratios should be dictated by the weak interactions governing competing reaction pathways.

Herein we describe the development of a catalytic, enantioselective oxy-Cope rearrangement of a symmetric tertiary alcohol. The mechanistic implications of relationships between enantioselectivity and chemoselectivity are highlighted in a discussion of the competing noncovalent interactions responsible for enantioinduction. Ultimately, we provide evidence that both cation-binding and anion-binding interactions are responsible for rate-acceleration over the uncatalyzed pathway. This finding is unexpected given long-held assumptions about the factors governing reactivity in the anionic oxy-Cope rearrangement and suggests that a synergistic ion-binding strategy may be effective for asymmetric catalysis of transformations involving oxyanionic intermediates more generally.

2. Results and Discussion

2.1. Urea-Catalyzed, Enantioselective oxy-Cope Rearrangement

The anionic oxy-Cope rearrangement typically is effected with stoichiometric quantities of a strong Brønsted base (such as KH or NaH) in polar ethereal solvents such as tetrahydrofuran or 1,4-dioxane. Often 18-crown-6 or another Lewis base is added to promote charge separation of the intermediate metal alkoxide ion pair. We anticipated that a catalytic system would require balancing the acidities of substrate, product, and H-bond donor to achieve turnover under mildly basic conditions that avoid irreversible catalyst deprotonation while maximizing the strength of noncovalent catalyst–substrate interactions.^[35,36] With these considerations in mind, we elected to examine the enantioselective rearrangement of symmetrically substituted, tertiary 1,5-dien-3-ols in nonpolar solvents that should disfavor charge separation in the absence of a catalyst.

Bis-styrenyl alcohol 1a was selected as a model substrate due to its ease of preparation and purification on multi-gram scale, as well as the common occurrence of chalcone derivatives as leads for medicinal chemistry applications.^[37,38] An initial survey was conducted in dichloromethane at ambient temperature (~23 °C) by varying the Brønsted base promoter employed in conjunction with thiourea catalyst 3b (Scheme 2; Table 1); the latter had been identified as optimal in the enantioselective [2,3]-Wittig rearrangement reported previously.^[28] Very low levels of substrate consumption occurred in the absence of base (Table 1, entry 1). Strong bases, such as sodium hydride and potassium tert-butoxide, promoted substrate decomposition to numerous, unidentified products (entries 2–4). Weaker bases such as cesium carbonate (entry 5), were ineffective, as was saturated aqueous cesium hydroxide (entry 6). By contrast, cesium hydroxide monohydrate (CsOH•H₂O) afforded ketone 2a as the major product with measurable, albeit low, levels of enantioinduction (entries 7, 8). Notably, the thiourea catalyst was found to be necessary for both reactivity and enantioselectivity; its exclusion lead to only low conversion to intractable decomposition products (entry 9). With this catalyst/base combination, evaluation of standard solvents revealed that little to no substrate consumption occurs in alkane solvents (entry 10), likely due to the poor solubility of both the substrate and base. Reactions conducted in dichloromethane or toluene were more productive, with modestly improved enantioselectivity observed in the latter (entry 11).

Scheme 2.

Model system for the development of an enantioselective anionic oxy-Cope rearrangement.

Table 1.

Optimization of reaction conditions.^a

Entry	Cat (mol %)	Base (equiv)	Solvent (M)	Time (h)	Conv b (%)	Yield 2a b (%)	2a:decomp b	er c
A. Evaluation of Brønsted base promoters.d
1	3b (10)	none	CH2Cl2 (0.05 M)	24	3	0	--	--
2	3b (10)	NaH (0.5)	“	8	12	0	0:1	--
3	3b (10)	KOtBu (0.5)	“	8	100	<2	0:1	--
4	3b (10)	NaOtBu (0.5)	“	8	10	10	1:0	49:51
5	3b (10)	Cs2CO3 (0.5)	“	8	0	0	--	--
6	3b (10)	sat. aq.CsOH (1:20 v/v)	“	24	0	0	--	--
7	3b (10)	CsOH•H2O (0.5)	“	8	39	25	1.8:1	56:44
8	3b (10)	CsOH•H2O (1.0)	“	24	32	23	2.5:1	56:44
9	none	“	“	24	4	0	0:1	--
10e	3b (10)	“	C6H12 (0.05 M)	24	0	0	--	--
11	3b (10)	“	PhMe (0.05 M)	24	26	16	1.6:1	59:41
B. Evaluation of reaction conditions for improved product yield. f
12	3a (10)	CsOH•H2O (1.0)	PhMe (0.05 M)	48	89	14	1:5.4	55:45
13g	3b (10)	“		48	42	18	1:1.3	61:39
14	4a (10)	“	“	48	94	18	1:4.2	66:34
15	4a (10)	CsOH•H2O (0.25)	ClPh (0.05 M)	24	87	40	1:1.2	65:35
16	4a (10)	“	ClPh (0.025 M)	24	84	49	1.4:1	65:35
17	4a (5)	“		16	61	53	7.2:1	65:35
18	none	“	“	16	5	0	0:1	--

^aSee Scheme 1 for conditions and catalysts.

^bPercent remaining starting material and product yield were determined from the ¹H NMR spectrum of the crude reaction mixture, integrated relative to a mesitylene internal standard added at the end of the reaction.

^cEnantiomeric excess was determined by CSP-HPLC (ODH, 15% IPA/hexanes, 1 mL/min) monitoring at 230 nm.

^dReactions were conducted on 0.05 mmol scale in 1-dram oven-dried vials under N₂ atmosphere using reagents stored on the bench-top and were quenched with the addition of saturated aqueous ammonium chloride solution.

^e35 °C.

^fReactions were conducted on 0.05 mmol scale in 10-mL round-bottom flasks under N₂ atmosphere using reagents stored in an N₂ atmosphere glovebox and were quenched with the addition of 0.1 M hydrochloric acid.

^gCatalyst 3b was not dried and stored in the glovebox; the low conversion relative to catalyst 3a may be ascribed simply to residual moisture.

While these results proved informative, some inconsistency in the reaction outcome was observed using CsOH•H₂O (95% purity with up to 5% Cs₂CO₃) stored on the benchtop. This variability was attributed to the inhibitory effect of water, presumably due to its propensity to associate with the alkali metal cation. Indeed, reactions conducted using dry CsOH•H₂O (>99% purity) and under anhydrous conditions afforded reproducible results. With these modifications, the urea analogue of thiourea 3a was found to afford improved enantioselectivity (entries 12, 14), but decomposition remained competitive with the productive rearrangement, particularly under rigorously dry conditions. Accordingly, the reaction conditions were modified further to improve upon the poor product/decomposition ratios. By employing chlorobenzene as the reaction solvent, reducing the absolute concentration, decreasing the catalyst loading, and shortening the reaction time, clean conversion to product 2b could be achieved while maintaining near-complete suppression of background decomposition. We hypothesize that these changes slowed intermolecular decomposition pathways involving reaction of the enolate formed during or after the rearrangement with other molecules of substrate or product. Accordingly, these conditions were selected as an appropriate platform for further optimization.

With reaction conditions affording clean product formation in hand, we performed a series of catalyst structure–chemoselectivity–enantioselectivity relationship studies (Scheme 3). Simplified catalysts 5–8 were prepared and assessed for their efficacy in promoting the enantioselective oxy-Cope rearrangement. Catalyst 5, which bears an indoline amide instead of a 2-arylpyrrolidine amide, exhibited decreased activity, chemoselectivity, and enantioselectivity relative to 4a. In contrast, catalyst 6, which bears an unsubstituted pyrroldine amide, showed a more modest decrease in activity and enantioselectivity, while preserving the good chemoselectivity observed with 4a. The indoline amide motif in 5 exists predominantly as a single amide rotamer and is expected to display a steric profile similar to that of the constrained arylpyrrolidine amide in 4a;^[39] however, N-aryl amides are substantially less Lewis basic than analogous N-alkyl derivatives.^[40] Taken together, the results with catalysts 5 and 6 thus implicate a direct role for the amide carbonyl in the product-determining transition state.

Scheme 3.

Catalyst structure–chemoselectivity–enantioselectivity relationship studies ^a

^a Conversion (yield) determined from the ¹H NMR spectrum of the crude reaction mixture, integrated relative to a mesitylene internal standard added at the end of the reaction.

The reaction outcome proved to be even more responsive to perturbation of the catalyst 2-phenylpyrrole moiety. Replacing the phenyl group with a methyl group, as in catalyst 7, led to improved reactivity at the expense of chemo- and enantioselectivity. In contrast, ablation of the methyl substituent, as in catalyst 8, afforded substantially improved enantioselectivity, albeit with a reduction in chemoselectivity. Ablation of the entire pyrrole moiety, as in catalyst 9, led to an inversion in the sense of enantioinduction, with further erosion of chemoselectivity. Taken together, these observations reveal that the pyrrole and its aryl substituent play a critical role in defining the catalyst features necessary for enantioinduction. This observation parallels those made in the development of enantioselective Claisen and [2,3]-Wittig rearrangements,^[28,41] for which cation-π interactions with the catalyst arylpyrrole moiety were invoked as defining conformational control elements. However, despite the apparent crucial role of the phenyl substituent in enantiocontrol, the use of an expanded aromatic group in its stead, as in catalyst 10 or 11, afforded no improvement, possibly due to competing steric effects.

Given the information gleaned from catalyst modification, two additional changes to the catalyst structure were investigated in order to decrease the steric profile proximal to the active site while increasing the electron density of the pyrrole moiety.^[42] Both of these modifications, manifested in catalysts 12 and 13, afforded improvements in reactivity and enantiocontrol without influencing product/decomposition ratios substantially.^[43] While product enantioselectivity remained modest, even with these improvements, catalyst 13 marks the most effective catalyst for the enantioselective oxy-Cope rearrangement identified to date.

2.2. Noncovalent Interactions Implicated in Stereoinduction

The catalysts described herein are similar to those optimized for the enantioselective [2,3]-Wittig rearrangement,^[28] for which the arylpyrrolidine, pyrrole, and amide carbonyl were proposed to define an “aromatic box” enabling recognition and organization of Cs⁺ (Scheme 1C).^[44] Given the anticipated importance of cation-binding for promoting reactivity in the anionic oxy-Cope rearrangement, we sought to determine if catalyst 13 exhibited similar cation specificity (Scheme 4; Table 2). Potassium hydroxide (KOH) and sodium hydroxide (NaOH) were thus evaluated as Brønsted base co-catalysts, but negligible product formation was observed with those modifications (Table 2, entries 2 & 3). While anhydrous CsOH was not available, precluding a direct comparison, lyophilization of benzene from CsOH•H₂O prior to its use and rigorous exclusion of water from the reaction medium had little impact on the reaction outcome (entry 4). We thus conclude that the catalyst does not associate with NaX or KX complexes (X⁻ = ⁻OH, [1a – H]⁻, [2a – H]⁻, etc.) in a manner analogous to its interactions with the reactive CsX ion pair. We hypothesize that the disparate reactivity of different alkali metal alkoxides arises from differences in the van der Waals radii of the metal ions. Like Cs⁺, K⁺ and Na⁺ can form strong cation-π and Lewis acid–base interactions with each of the motifs comprising the catalyst’s “aromatic box”;^[45-47] however, these cooperative interactions apparently cannot be sustained simultaneously with the smaller cations.

Scheme 4.

The effect of countercation on reaction outcome.

Table 2.

Cation identity is crucial for reactivity, chemoselectivity, and enantioselectivity.^a

Entry	Base (equiv)	Conv (%) b	Yield 2a (%) b	2a :decomp b	er c
1	CsOH • H2O (0.25)	76	60	3.75:1	75:25
2	KOH (0.25)	4	4	1:0	62:38
3	NaOH (0.25)	0	0	--	--
4 d	CsOH • H2O (0.25)	80	60	3.0:1	75:25

^aReactions were conducted on 0.05 mmol scale in 10-mL round-bottom flasks under N₂ atmosphere using reagents stored in an N₂ atmosphere glovebox and were quenched with the addition of deionized water. See Scheme 4 for conditions.

^bPercent remaining starting material and product yield were determined from the ¹H NMR spectrum of the crude reaction mixture, integrated relative to a mesitylene internal standard added at the end of the reaction.

^cEnantiomeric ratio was determined by CSP-HPLC (ODH, 15% isopropanol/hexanes, 1 mL/min) monitoring at 230 nm.

^dIn an N₂ atmosphere glovebox CsOH • H₂O (0.25) and 13 were premixed in anhydrous benzene for 15 minutes before the mixture was lyophilized and subjected to standard reaction conditions.

While the catalyst optimization described above provided strong indications that the pyrrole and amide moieties are crucial for catalyst function, the role of the 2-arylpyrrolidine remained to be defined. We recognized that if the aryl group interacts with Cs⁺ or the reacting ion pair in the product-determining transition state(s), chemoselectivity and/or enantioselectivity should be sensitive to systematic variation of the steric and electronic properties of that arene. Accordingly, a series of substituted 2-phenylpyrrolidine amido-urea catalysts (12b–g) was prepared and evaluated in the model system (Figure 1A). While catalysts 12a–g all afforded product 2a in comparable yield, a strong correlation was observed between the electronic properties of the substituted arene (as described by the Hammett σ_p value)^[48] and enantioselectivity, wherein catalysts bearing the most electron-donating substituents achieved the highest enantioselectivity. Combined with the importance of the cesium cation, this correlation suggests that a cation-π or similar electrostatic interaction is stronger in the transition state leading to the major enantiomer of product than in the transition state leading to the minor enantiomer of product, thereby contributing to enantiodifferentiation.^[49,50]

Given this insight, we hypothesized that expanded aromatic substituents at pyrrolidine might confer a similar advantage, given that they are more polarizable and possess larger quadrupole moments than phenyl derivatives alone. Both physical parameters are known to correlate directly with cation-π interactions.^[51] Moreover, the enantioselectivity obtained with other reactions catalyzed by arylpyrrolidine amido-thioureas have, in several cases, been shown to improve with increased aromatic expanse.^[52] However, these modifications did not afford any improvement in the anionic oxy-Cope rearrangement. Instead enantioselectivity observed with catalysts 12a and 12h–j was found to correlate inversely with the angle θ defining the span of each arene across its widest point (Figure 1B), such that catalysts bearing the largest (or widest) arylpyrrolidines afforded product 2a with the lowest e.r. Consistent with this detrimental steric effect, electron-rich 3,5-disubstituted phenylpyrroldine catalysts 12l and 12m exhibited lower enantioselectivity than would be expected based on electronic parameters alone (Figure 1C). While the linear free-energy relationships between catalyst structure and enantioselectivity provide compelling evidence for a specific role of the catalyst arylpyrrolidine moiety in enantioinduction, the competing steric and electronic substituent effects limit the scope of further variations at this site.

2.3. Mechanistic Requirement for Synergistic Ion-Binding

With the development of polyfunctional urea catalysts exhibiting pronounced influence over chemo- and enantioselectivity in the anionic oxy-Cope rearrangement, we sought to evaluate the extent to which anion-binding, cation-binding, and the potential synergistic interplay between the two accounted for catalyst control over the reaction outcome. Toward that aim, an achiral N,N’-dialkylurea and an achiral crown ether were evaluated as catalysts, both in isolation and together (Scheme 5; Table 3). The simple urea alone promoted the reaction to a negligible extent under standard reaction conditions (Table 3, entry 1). The use of 18-crown-6 alone lead to rapid conversion, primarily to decomposition products (entry 2). Remarkably, the combination of the two catalysts resulted in significantly improved chemoselectivity for product formation over decomposition with only a modest reduction in reactivity relative to use of 18-crown-6 alone (entry 3). While these results corroborated our initial hypothesis that cation-binding would be essential for substrate activation, the cooperative impact of the urea on chemoselectivity suggested that H-bonding interactions served a more productive role than anticipated previously.^[53]

Scheme 5.

H-Bonding attenuates Lewis base-promoted decomposition.

Table 3.

H-Bonding attenuates decomposition pathways.^a

Entry	Cat (mol %)	Time (h)	Conv (%) b	Yield 2a (%) b	2a:decomp b
1	N,N’-bis(octyl)urea (10)	20	8	<2%	<1:4
2	18-crown-6 (25)	3	100	12%	1:7.3
3	N,N’-bis(octyl)urea (10) + 18-crown-6 (25)	3	>90	36%	1:1.5

^aReactions were conducted on 0.05 mmol scale in 10-mL round-bottom flasks under N₂ atmosphere using reagents stored in an N₂ atmosphere glovebox and were quenched with the addition of 0.1 M hydrochloric acid.

^bPercent remaining starting material and product yield were determined from the ¹H NMR spectrum of the crude reaction mixture, integrated relative to a mesitylene internal standard added at the end of the reaction.

Given the unexpected role of hydrogen-bonding in modulating the chemoselectivity in the achiral model system, we sought to examine the function of the H-bond donor within the full chiral catalyst. To probe each of the two N–H units independently, a series of amide and N,N,N’-trialkylurea analogues of catalyst 4a were developed (Scheme 6). On the basis of approximate pK_a values determined in organic media, potential H-bonding interactions with each of these motifs should be comparable in strength (slightly stronger with 15–17; slightly weaker with 18 and 19) to those with dual H-bond donor 4a.^[35,36,54] However, amides 15–17 were not competent catalysts for the anionic oxy-Cope rearrangement, affording low conversion of substrate 1a to decomposition products under standard conditions. While N,N,N’-trialkylureas 18 and 19 were modestly more active than the amide analogues, decomposition still predominated over the formation of 2a (with negligible enantioselectivity). These results thus highlight the critical, albeit counterintuitive, role of dual H-bonding interactions in dictating productive reaction outcomes.

Scheme 6.

Single-H-bond donors are ineffective catalysts.

^a Conversion (yield) determined from the ¹H NMR spectrum of the crude reaction mixture, integrated relative to a mesitylene internal standard added at the end of the reaction.

Given that the apparent benefit of H-bonding interactions in anionic oxy-Cope rearrangements is unexpected in the context of literature precedent, we sought to ascertain whether the chiral urea was instead functioning as a pre-catalyst to a Brønsted-basic cesium ureate.^[55,56] In order to evaluate this hypothesis, the conjugate base of urea 8 was prepared independently and subjected to the standard reaction conditions, both in the presence and absence of additional CsOH•H₂O (Scheme 7; Table 4). While conversion to product 2a and enantioselectivity were found to be comparable using 8 and its conjugate base (20) in the presence of CsOH•H₂O (Table 4, entries 1 & 2), no reaction was observed upon treating the substrate 2a with 20 in the absence of additional CsOH•H₂O (entry 3). These results thus suggest that, while reversible catalyst deprotonation is possible under the reaction conditions (i.e. 20 can be protonated to reenter the productive catalytic cycle), its cesium ureate conjugate base is not the active catalytic species.

Scheme 7.

The urea is not a Brønsted base precatalyst.

Table 4.

Urea deprotonation is reversible but not requisite. ^a

Entry	Cat	Base (mol%)	Conv (%) b	Yield 2a (%) b	2a:decomp b	er c
1	8	CsOH•H2O (25)	64	48	2.9:1	70:30
2 d	20	CsOH•H2O (25)	82	63	3.3:1	70:30
3 d	20	none	0	0	--	--

^aReactions were conducted on 0.05 mmol scale in 10-mL round-bottom flasks under N₂ atmosphere using reagents stored in an N₂ atmosphere glovebox and were quenched with the addition of 0.1 M hydrochloric acid.

^bPercent remaining starting material and product yield were determined from the ¹H NMR spectrum of the crude reaction mixture, integrated relative to a mesitylene internal standard added at the end of the reaction.

^cEnantiomeric ratio was determined by CSP-HPLC (ODH, 15% isopropanol/hexanes, 1 mL/min) monitoring at 230 nm

^d20 h.

Conclusion

The development of the enantioselective anionic oxy-Cope rearrangement described here serves as a functional probe of the nature of the transition state for the emblematic transformation. Competing dissociative pathways are manifest in product/decomposition and enantiomer ratios wherein the latter is particularly sensitive to interactions between the rearranging anion and a polyfunctional urea catalyst. It has long been assumed that maximum charge separation between the reacting oxyanion and its counteraction is necessary for productive rearrangement. However, the results described herein paint a more nuanced picture. Finely balanced cation-binding and anion-binding interactions are shown to activate the substrate while attenuating the “naked” anionic character that leads to decomposition pathways. While only modest enantioselectivities were ultimately achieved with this catalyst system, the fundamental insights gleaned through this study suggest that a revision to long-standing notions about charge-accelerated sigmatropic rearrangements may be in order. Furthermore, these results highlight opportunities for the application of a synergistic ion-binding strategy for the enantioselective catalysis of transformations involving oxyanionic intermediates more generally.

3. Experimental section

3.1. Procedures, Materials, Instrumentation, and Software

Unless otherwise noted, all reactions were performed under a nitrogen atmosphere in round- bottom flasks, which were flame-dried under vacuum and cooled under a nitrogen atmosphere prior to use. The 10-mL round-bottom flasks employed for catalytic experiments were washed in a base bath, rinsed with copious deionized water, and dried >12 hours in a 250 °C oven prior to cooling under vacuum in a glovebox antechamber. Air- and/or moisture- sensitive liquids were transferred with stainless steel cannulae or with plastic Norm-Ject or glass Hamilton gas-tight syringes fitted with stainless-steel needles. Cooling baths were prepared from acetone/dry ice (−78 °C), acetonitrile/dry ice (−40 °C), or ice/water (0 °C) mixtures; a submersion cooling apparatus (Neslab CC 100) in an acetone bath was used to maintain other cooling bath temperatures. Reactions were monitored by thin-layer chromatography (TLC) on Silica Gel 60 F254 plates (EMD), visualized under UV light (254 nm) or with potassium permanganate or ceric ammonium molybdate (CAM) stain, which developed upon heating. Flash chromatography was performed using SiliaFlash P60 (230–400 mesh, SiliCycle), in a glass column or 3-mL pipette. Organic solutions were concentrated under reduced pressure on a Büchi rotary evaporator.

Reagent-grade cesium hydroxide monohydrate (>99% purity) and cesium carbonate were purchased from Strem and stored in a glovebox under N₂ atmosphere. Other commercially available reagents were purchased from Sigma–Aldrich, Alfa Aesar, Strem, Oakwood, Matrix Scientific, Ark Pharm Inc., or TCI and used without purification, unless otherwise indicated. Extraction and chromatography solvents (EMD or BDH) were reagent grade and used without purification. When employed in a reaction, acetonitrile, dichloromethane, toluene, tetrahydrofuran, tert-butyl methyl ether, ethyl ether, and 1,4-dioxane (PURE SOLV) first were dried by passage through columns of activated alumina.^[57] Anhydrous chlorobenzene was purchased in a SureSeal bottle from Sigma–Aldrich and was used as received. Cyclohexane was dried by distillation from calcium hydride, degassed by three freeze-pump-thaw cycles under vacuum, and stored > 48 hours over activated 4 Å molecular sieves (30 g MS per 100 mL cyclohexane) in a glovebox under N₂ atmosphere prior to use. Pyridine, triethylamine, N,N-diisopropyl ethylamine, and N,N-diisopropylamine were distilled from calcium hydride under nitrogen atmosphere prior to use. Deuterated solvents—namely CDCl₃, CD₃OD, and C₆D₆ (Cambridge Isotope Laboratories)—and HPLC solvents (EMD) were used without purification. Molecular sieves (4Å, 4–12 mesh beads, Sigma-Aldrich) were activated by heating between 250 °C and 300 °C in vacuo for 24 hours and were stored in a glovebox under nitrogen atmosphere prior to use.

Proton nuclear magnetic resonance (¹H NMR) spectra and proton-decoupled carbon nuclear magnetic resonance (¹³C {¹H} NMR) spectra were recorded at 25 °C (unless stated otherwise) on Varian-Mercury-400 (400 MHz), Varian Unity/Inova 500 (500 MHz), or Varian Unity/Inova 600 (600 MHz) spectrometers at the Harvard University nuclear magnetic resonance facility. Chemical shifts for proton are reported in parts per million downfield from tetramethylsilane and are referenced to residual protium in the NMR solvent according to values reported in the literature: δ(CDCl₃) = 7.26, δ(CD₃OD) = 3.30, δ(C₆D₆) = 7.15, δ(C₆D₅CD₃). Chemical shifts for carbon are reported in parts per million downfield from tetramethylsilane and are referenced to the carbon resonances of the solvent according to values reported in the literature: δ(CDCl₃) = 77.00, δ(C₆D₆) = 128.06 Data are represented as follows: chemical shift, integration, multiplicity (br = broad, s = singlet, d = doublet, t = triplet, q = quartet, qn = quintet, sp = septet, m = multiplet), coupling constants in Hertz (Hz). Note that ¹³C NMR spectra were obtained without ¹⁹F decoupling. Standard optical rotation measurements were performed on a Jasco P-2000 digital polarimeter using a 1 mL cell with a 5 cm path length. Infrared spectra were recorded using a Bruker Optics Tensor 27 FT-IR spectrometer. Data are represented as follows: frequency of absorption (cm^–1), intensity of absorption (s = strong, m = medium, w = weak, br = broad). High-resolution mass spectrometry was measured using a Bruker micrOTOF-QII™ ESI-Qq-TOF mass spectrometer calibrated using an aqueous sodium formate solution (prepared via adding 1 mL of 1 M aq. NaOH in 100 mL of 1% aq. formic acid). Chiral stationary phase high performance liquid chromatography (CSP-HPLC) was performed using an Agilent 1200 quaternary HPLC system with commercially available ChiralPak OD-H or AD-H columns. Integration was typically performed at 230 nm.

Curve fitting and data processing was performed using the MatPlotLib^[58] package in iPython Notebooks.^[59] NMR spectra were processed with ACD/NMR Processor^[60] or MNova.^[61] Computations were performed with Gaussian 09 on the Odyssey cluster supported by Harvard University’s Faculty of Arts and Sciences Research Computing group.

3.2. Catalyst Synthesis and Characterization Data

Catalysts 3a, 3b, and 4a were prepared as reported previously.^[28] Ureas 7, 8, 10, and 11 were synthesized, through the general procedure described below. Fragments 21 and 22 were prepared as reported previously,^[62] and spectral data were in agreement with the published values. The syntheses of catalysts 5, 6, 9, and 12–20 are detailed in the supplementary materials.

An undried 25-mL conical flask was charged with a PTFE-coated magnetic stir bar and 21 (0.20 mmol, 1.0 equiv), then flushed with N₂. Anhydrous dichloromethane (1.4 mL) was added, and the resulting solution was cooled to 0 °C. Trifluoroacetic acid (0.7 mL) was added slowly at 0 °C under N₂. The flask was maintained for 30 minutes at 0 °C, then removed from the ice bath and maintained for 1.5 hours at ambient temperature. After 2 hours in total, the reaction solution was again cooled to 0 °C, diluted with dichloromethane (6 mL), and basicified (to pH ≈ 10 by pH paper) with the gradual addition of saturated aqueous sodium carbonate. The layers were mixed and partitioned, and the aqueous layer was extracted with additional dichloromethane (4 × 5 mL). The combined organic extracts were dried over sodium sulfate, filtered, and concentrated to afford the free amine as a colorless oil, which was used immediately in the following steps.

The 50-mL undried conical flask containing the amine isolated above was charged with a PTFE-coated magnetic stir bar, evacuated, and back-filled with N₂. Anhydrous dichloromethane (2.0 mL) was added, and the resulting solution was cooled to 0 °C. Saturated aqueous sodium bicarbonate (2.0 mL) was added at 0 °C under N₂, and the mixture was stirred vigorously. Stirring was halted while phosgene solution (15 wt% in toluene, 0.145 mL, 0.22 mmol, 1.1 equiv) was added directly to the organic layer; then vigorous stirring was resumed. The flask was maintained for 10–15 minutes at 0 °C; then, the reaction mixture was diluted with dichloromethane (4 mL) and transferred to a separatory funnel. The layers were partitioned, and the aqueous layer was extracted with additional dichloromethane (2 × 8 mL). The combined organic extracts were dried over sodium sulfate, filtered, and concentrated to afford the isothiocyanate as a white to pale yellow solid, which was used immediately in the following step.

The 50-mL undried conical flask containing the isocyanate prepared above was charged with a PTFE-coated magnetic stir bar and flushed with N₂. Anhydrous dichloromethane (1.0 mL) was added under N₂, followed by 22 (0.24 mmol, 1.2 equiv) as a solution in anhydrous dichloromethane (3 × 0.3 mL to effect a quantitative transfer). Following the addition, the septum was sealed with plastic paraffin film, and the reaction was maintained for 20 hours. After 20 hours, the reaction was concentrated, then loaded onto a silica gel column with toluene and purified by flash chromatography, eluting with ~15% ethyl acetate in hexanes. The fractions containing the desired product by TLC were combined and concentrated. The resulting residue was redissolved in ethyl ether, filtered through a plug of a lint-free laboratory wipe to remove silica and sand, and concentrated to afford a pale yellow foam. The foam was dried under high vacuum (< 1 torr), then redissolved in benzene. The solution was frozen in liquid nitrogen, and lyophilized to afford urea 7, 8, 10, or 11 as a white or taupe powder. The powder was crushed and dried under high vacuum overnight, then in a vacuum desiccator over phosphorous pentoxide (~ 50 torr) for several days prior to use.

3.2.1. 1-((S)-3,3-dimethyl-1-((R)-2-methyl-2-phenylpyrrolidin-1-yl)-1-oxobutan-2-yl)-3-((1R,2R)-2-(2,5-dimethyl-1H-pyrrol-1-yl)cyclohexyl)urea (7)

Urea 7 was prepared from 21 and 22b according to the general procedure described above (133 mg, 0.27 mmol, 55% yield) ¹H NMR (600 MHz, CDCl₃): δ 7.24 (t, J = 7.6 Hz, 2 H), 7.15 (t, J = 7.6 Hz, 1 H), 7.07 (d, J = 7.6 Hz, 2 H), 5.74 (s, 2 H), 4.75 (d, J = 9.4 Hz, 1 H), 4.51 (d, J = 9.4 Hz, 1 H), 4.13–4.01 (m, 3 H), 3.83 (dt, J = 5.9, 10.9 Hz, 1 H), 3.76 (td, J = 8.2, 9.7 Hz, 1 H), 2.41–2.32 (m, 1 H), 2.01 (dd, J = 6.2, 7.6 Hz, 2 H), 1.95–1.89 (m, 2 H), 1.85 (s, 3 H), 1.85–1.75 (m, 3 H), 1.42–1.31 (m, 3 H), 0.96 (s, 9 H) ppm. ¹³C NMR (126 MHz, C₆D₆): δ = 171.6, 158.4, 146.9, 128.7, 126.4, 125.8, 67.6, 60.2, 58.7, 53.7, 50.5, 45.0, 35.9, 34.7, 32.9, 27.3, 26.6, 26.4, 26.0, 22.3, 15.4 (br) ppm. FT-IR (thin-film, CDCl₃-cast): 3355, 2951, 2933, 2868, 1635, 1563, 1548, 1520, 1505, 1495, 1480, 1446, 1416, 1398, 1368 cm^–1. HRMS (ESI): for C₃₀H₄₄N₄O₂, [M + H]⁺ calculated m/z = 493.3523, found m/z = 493.3534; [M + Na]⁺ calculated m/z = 515.3362, found m/z = 515.3348. [α]²²_D = +65.6 ° (c. 1.0, CHCl₃)

3.2.2. 1-((S)-3,3-dimethyl-1-((R)-2-methyl-2-phenylpyrrolidin-1-yl)-1-oxobutan-2-yl)-3-((1R,2R)-2-(2-phenyl-1H-pyrrol-1-yl)cyclohexyl)urea (8)

Urea 8 was prepared from 21 and 22c according to the general procedure described above (220 mg, 0.41 mmol, 83% yield) ¹H NMR (600 MHz, CDCl₃): δ 7.40 (t, J = 7.6 Hz, 2 H), 7.35–7.30 (m, 3 H), 7.22 (t, J = 7.6 Hz, 2 H), 7.13 (t, J = 7.6 Hz, 1 H), 7.08 (d, J = 7.6 Hz, 2 H), 6.92 (dd, J = 1.8, 2.9 Hz, 1 H), 6.23 (t, J = 2.9 Hz, 1 H), 6.15 (dd, J = 1.8, 3.5 Hz, 1 H), 4.63 (d, J = 9.4 Hz, 1 H), 4.50 (d, J = 9.4 Hz, 1 H), 4.07 (ddd, J = 5.3, 7.0, 10.6 Hz, 1 H), 3.86 (ddd, J = 3.5, 10.8, 11.2 Hz, 2 H), 3.83–3.77 (m, 1 H), 3.75 (dt, J = 1.8, 7.6 Hz, 2 H), 2.17 (d, J = 12.3 Hz, 2 H), 2.01 (dd, J = 6.2, 7.3 Hz, 2 H), 1.86 (s, 3 H), 1.85–1.77 (m, 4 H), 1.72 (d, J = 12.3 Hz, 1 H), 1.40 (d, J = 11.7 Hz, 1 H), 1.33–1.24 (m, 1 H), 1.12 (ddt, J = 4.1, 10.0, 13.5 Hz, 1 H), 0.95 (s, 9 H) ppm. ¹³C NMR (126MHz, C₆D₆): δ 171.1, 158.0, 146.9, 135.3, 135.3, 130.2, 129.2, 128.8, 127.4, 126.5, 125.8, 120.0, 109.9, 109.4, 67.7, 59.7, 58.6, 54.9, 50.4, 44.9, 36.1, 35.3, 34.6, 27.3, 26.3, 25.9, 25.4, 22.4 ppm. FT-IR (thin-film, CDCl₃-cast): 3357, 2933, 2864, 1630, 1603, 1550, 1494, 1478, 1466, 1446, 1415, 1367, 1308, 1286, 1275 cm^–1. HRMS (ESI): for C₃₄H₄₄N₄O₂, [M + H]⁺ calculated m/z = 541.3543, found m/z = 541.3542; [M + Na]⁺ calculated m/z = 563.3362, found m/z = 563.3360. [α]²²_D = +60.9 ° (c. 1.0, CHCl₃)

3.2.3. 1-((S)-3,3-dimethyl-1-((R)-2-methyl-2-phenylpyrrolidin-1-yl)-1-oxobutan-2-yl)-3-((1R,2R)-2-(2-methyl-5-(naphthalen-2-yl)-1H-pyrrol-1-yl)cyclohexyl)urea (10)

Urea 10 was prepared from 21 and 22d according to the general procedure described above. (51 mg, 0.084 mmol, 84% yield). ¹H NMR (600 MHz, CDCl₃): δ 7.90–7.83 (m, 2 H), 7.78 (d, J = 6.5 Hz, 2 H), 7.48 (d, J = 6.3 Hz, 1 H), 7.48 (d, J = 6.3 Hz, 1 H), 7.41 (d, J = 5.9 Hz, 1 H), 7.24 (dd, J = 7.3, 7.6 Hz, 2 H), 7.15 (d, J = 7.6 Hz, 2 H), 7.10 (t, J = 7.3 Hz, 1 H), 6.15 (br s, 1 H), 5.97 (d, J = 2.9 Hz, 1 H), 4.70 (d, J = 7.6 Hz, 1 H), 4.53 (d, J = 9.4 Hz, 1 H), 4.12 (br s, 1 H), 4.06 (t, J = 5.9 Hz, 1 H), 3.98 (dt, J = 2.9, 11.2 Hz, 1 H), 3.78 (td, J = 7.6, 10.0 Hz, 1 H), 3.70 (br s, 1 H), 2.53 (br s, 3 H), 2.31–2.09 (m, 2 H), 2.07–2.00 (m, 2 H), 1.88 (s, 3 H), 1.86–1.79 (m, 2 H), 1.73–1.52 (m, 3 H), 1.41–1.14 (m, 2 H), 0.95 (s, 9 H), 0.91–0.85 (m, 1 H) ppm. ¹³C NMR (126 MHz, C₆D₆) 171.2, 157.1, 146.9, 136.6, 134.6, 133.1, 131.2, 129.0, 128.8, 127.0, 126.6, 126.5, 125.9, 110.5, 109.7, 67.8, 61.6, 58.6, 52.5, 50.4, 44.9, 36.5, 34.9, 33.1, 27.3, 26.4, 26.3, 25.6, 22.4, 16.0 ppm. FT-IR (thin-film, CDCl₃-cast): 3371, 2953, 2929, 2868, 1637, 1601, 1539, 1495, 1479, 1461, 1446, 1414, 1392, 1367, 1305 cm^–1. HRMS (ESI): for C₃₉H₄₈N₄O₂, [M + H]⁺ calculated m/z = 605.3856, found m/z = 605.3869; [M + Na]⁺ calculated m/z = 627.3675, found m/z = 627.3695. [α]²⁴_D = +68.7 ° (c. 1.0, CH₂Cl₂)

3.2.4. 1-((S)-3,3-dimethyl-1-((R)-2-methyl-2-phenylpyrrolidin-1-yl)-1-oxobutan-2-yl)-3-((1R,2R)-2-(2-methyl-5-(naphthalen-1-yl)-1H-pyrrol-1-yl)cyclohexyl)urea (11)

Urea 11 was prepared from 21 and 22e according to the general procedure described above (55 mg, 0.091 mmol, 91% yield). ¹H NMR (600 MHz, CDCl₃): δ 7.87 (t, J = 9.4 Hz, 2 H), 7.81–7.32 (m, 5 H), 7.22 (dd, J = 7.1, 7.6 Hz, 2 H), 7.13 (d, J = 7.6 Hz, 2 H), 7.15 (t, J = 7.6 Hz, 1 H), 7.06 (t, J = 7.1 Hz, 1 H), 6.16* (s, 1 H), 6.09 (s, 1 H), 6.01** (s, 1 H), 4.87–4.41 (m, 2 H), 4.21* (br s, 1 H), 4.17–4.08 (m, 1 H), 4.09–3.91** (m, 1 H), 3.87–3.67 (m, 1 H), 3.47–3.16 (m, 1 H), 2.59* (br s, 3 H), 2.56** (br s, 3 H), 2.26–1.97 (m, 7 H), 1.90 (s, 3 H), 1.95–1.80 (m, 3 H), 1.95–1.80 (m, 2 H), 0.98** (s, 9 H), 0.93* (s, 9 H) ppm where * and ** denote clearly distinguishable major and minor rotamers. ¹³C NMR (126 MHz, C₆D₆) δ 170.9, 156.9, 146.7, 135.3, 134.7, 134.4, 133.7, 132.5, 130.8, 129.4, 129.0, 128.9, 127.6, 127.4, 127.2, 126.6, 125.7, 125.5, 110.2, 110.1, 67.7, 60.9, 58.2, 51.4, 50.4, 44.8, 36.4, 34.8, 33.0, 27.1, 26.2, 26.1, 25.2, 22.5, 15.8 ppm, where only the resonances for the major rotamer are listed. FT-IR (thin-film, CDCl₃-cast): 3411, 2952, 2932, 2866, 1642, 1525, 1415, 1393, 1299, 760, 699 cm^–1. HRMS (ESI): for C₃₉H₄₈ N₄O₂, [M + H]⁺ calculated m/z = 605.3856, found m/z = 605.3863; [M + Na]⁺ calculated m/z = 627.3675, found m/z = 627.3666. [α]²⁴_D = +29.3 ° (c. 1.0, CH₂Cl₂)

3.3. Substrate Synthesis and Characterization Data

Symmetric, substituted 1,5-diene-2-ol substrates (1) were prepared via addition of allylmagnesium chloride to the corresponding dibenzylidineacetones (23), which were prepared as reported previously.^[63] A representative procedure is described for 1a (where Ar = 4-methoxyphenyl).

3.3.1. (E)-1-(4-methoxyphenyl)-3-((E)-4-methoxystyryl)hexa-1,5-dien-3-ol (1a):

A 200-mL flame-dried oblong flask was charged with 23a (Ar = 4-methoxyphenyl, 5.95 g, 20 mmol, 1.0 equiv) and a PTFE-coated magnetic stir bar. The flask was evacuated and backfilled with N₂ before anhydrous tetrahydrofuran (60 mL) was added. The mixture was stirred to homogeneity, then cooled to 0 °C in an ice/water bath. Allyl magnesium chloride (2.0 M solution in tetrahydrofuran, 14 mL, 28 mmol, 1.4 equiv.) was then added slowly while stirring at 0 °C. The reaction was maintained as the ice bath was allowed to warm to ambient temperature (~23 °C). After 5 hours, no remaining starting material could be observed by TLC. The reaction mixture was quenched with the gradual addition of saturated aqueous ammonium chloride (80 mL). The mixture was extracted with ethyl ether (3 × 100 mL); then, the combined organic extracts were washed brine, dried over sodium sulfate, filtered, and concentrated to afford a viscous pale yellow oil. The crude material was recrystallized at –10 °C from a saturated dichloromethane solution layered with hexanes. The white flakey crystals were collected by filtration, rinsing with hexanes to afford 1a (5.75 g, 17.1 mmol, 85% yield) as a white crystalline solid. ¹H NMR (600 MHz, CDCl₃): δ 7.35 (d, J = 8.8 Hz, 4 H), 6.87 (d, J = 8.8 Hz, 4 H), 6.63 (d, J = 15.8 Hz, 2 H), 6.24 (d, J = 16.4 Hz, 2 H), 5.88 (tdd, J = 7.6, 10.6, 16.4 Hz, 1 H), 5.21 (d, J = 10.6 Hz, 1 H), 5.22 (d, J = 16.4 Hz, 1 H), 3.82 (s, 6 H), 2.58 (d, J = 7.6 Hz, 2 H), 2.01 (s, 1 H) ppm. ¹³C NMR (126 MHz, CDCl₃): δ 159.1, 133.1, 131.7, 129.5, 128.0, 127.6, 119.6, 113.9, 74.6, 55.2, 46.5 ppm. FT-IR (thin-film, CDCl₃-cast): 3448, 3031, 3003, 2932, 2904, 2834, 1606, 1510, 1463, 1440, 1419, 1244, 1174, 1106, 1032, 971,919, 819 cm^–1. HRMS (ESI): for C₂₂H₂₄O₃, [M – OH]⁺ calculated m/z = 319.16982, found m/z = 318.1702; [M + Na]⁺ calculated m/z = 359.1623, found m/z = 359.1616.

3.4. Standard Catalytic Procedure and Product Characterization

3.4.1. Standard Catalytic Procedure

In an N₂ atmosphere glovebox, an oven-dried, 10-mL round-bottom flask was charged with cesium hydroxide monohydrate (2.1 mg, 0.0125 mmol, 25 mol%, Strem), catalyst (0.0025 mmol, 5 mol%), and a PTFE-coated magnetic stir bar. A solution of 1a (2.0 mL of 0.025 M solution, 0.05 mmol, 1.0 equiv) in anhydrous chlorobenzene was added. The flask was sealed with a rubber septum, removed from the glovebox, and stirred (750 rpm) at ambient temperature (~ 23 °C) for 16–20 hours over which time it developed a pale yellow hue. The reaction was quenched with the addition of deionized water (2.0 mL). The layers were mixed and partitioned; the organic layer was filtered through a plug of sodium sulfate. The aqueous layer was extracted with additional ethyl ether (3 × 2 mL); each time, the extracted organic layer was filtered through the same plug of sodium sulfate. The filtrate was concentrated to afford a yellow oil. Mesitylene (5.0 μL, 0.036 mmol, 0.36 equiv) was added as an internal standard, and the mixture was diluted with CDCl₃ (0.75 mL). The conversion and product yield were determined from relative integrations of the allylic resonances and the mesitylene benzylic resonance in the ¹H NMR spectrum of the crude reaction mixture. The crude material was purified by silica gel chromatography (on a pipette column), eluting with 10% ethyl acetate in hexanes to afford 2a as a pale yellow oil, and the enantiomeric excess was determined by CSP-HPLC analysis.

Note: Substrate 1a is mildly acid-sensitive, and will undergo elimination upon extended exposure to chloroform. Accordingly, CDCl₃ was stored over anhydrous potassium carbonate, and NMR samples were kept only for the time required for analysis.

3.4.2. (E)-1,5-bis(4-methoxyphenyl)octa-1,7-dien-3-one (2a)

¹H NMR (600 MHz, CDCl₃): δ 7.45 (d, J = 8.2 Hz, 2 H), 7.43 (d, J = 15.8 Hz, 1 H), 7.15 (d, J = 8.2 Hz, 2 H), 6.90 (d, J = 8.2 Hz, 2 H), 6.83 (d, J = 8.2 Hz, 2 H), 6.54 (d, J = 15.8 Hz, 1 H), 5.69 (tdd, J = 7.0, 10.1, 17.0 Hz, 1 H), 5.01 (d, J = 17.0 Hz, 1 H), 4.97 (d, J = 10.0 Hz, 1 H), 3.85 (s, 3 H), 3.77 (s, 3 H), 3.33 (quin, J = 7.2 Hz, 1 H), 2.93 (dd, J = 7.2, 13.0 Hz, 2 H), 2.41 (td, J = 7.0, 7.2 Hz, 2 H) ppm. ¹³C NMR (126 MHz, CDCl₃): δ 199.1, 161.6, 158.0, 142.3, 136.4, 136.3, 130.0, 128.5, 127.2, 124.3, 116.6, 114.4, 113.8, 55.4, 55.2, 47.2, 40.8, 40.4 ppm. FT-IR (thin film, CDCl₃-cast): 3742, 3668, 3033, 3000, 2910, 2835, 2161, 2060, 1885, 1681, 1650, 1596, 1572, 1510, 1462 cm^–1. HRMS (ESI): for C₂₂H₂₄O₃, [M + H]⁺ calculated m/z = 337.1804, found m/z = 337.1802; [M + Na]⁺ calculated m/z = 359.1623, found m/z = 359.1629; [2M + Na]⁺ calculated m/z = 695.3349, found m/z = 695.3351. CSP-HPLC: ODH, 1 mL/min, 15 % isopropanol in hexanes, 230 nm, R_t(major) = 12.9 min; R_t(minor) = 18.8 min

Scheme 8.

General procedure for the synthesis of conformationally constrained urea catalysts.