A Case Study in Catalyst Generality: Simultaneous, Highly-Enantioselective Brønsted- and Lewis-Acid Mechanisms in Hydrogen-Bond-Donor Catalyzed Oxetane Openings

Abstract

Generality in asymmetric catalysis can be manifested in dramatic and valuable ways, such as high enantioselectivity across a wide assortment of substrates in a given reaction (broad substrate scope) or as applicability of a given chiral framework across a variety of mechanistically distinct reactions (privileged catalysts). Reactions and catalysts that display such generality hold special utility, because they can be applied broadly and sometimes even predictably in new applications. Despite the great value of such systems, the factors that underlie generality are not well understood. Here, we report a detailed investigation of an asymmetric hydrogen-bond-donor catalyzed oxetane opening with TMSBr that is shown to possess unexpected mechanistic generality. Careful analysis of the role of adventitious protic impurities revealed the participation of competing pathways involving addition of either TMSBr or HBr in the enantiodetermining, ring-opening event. The optimal catalyst induces high enantioselectivity in both pathways, thereby achieving precise stereocontrol in fundamentally different mechanisms under the same conditions and with the same chiral framework. The basis for that generality is analyzed using a combination of experimental and computational methods, which indicate that proximally localized catalyst components cooperatively stabilize and precisely orient dipolar enantiodetermining transition states in both pathways. Generality across different mechanisms is rarely considered in catalyst discovery efforts, but we suggest that it may play a role in the identification of so-called privileged catalysts.

Figure 1.

A) The Sharpless epoxidation represents a prototypical example of substrate generality in an asymmetric catalytic reaction. Even though the reaction scope is limited to allylic alcohols, it tolerates extensive variation in the alkenyl substituents, thereby enabling its broad application in synthesis. B) The cinchona alkaloids are prototypical privileged chiral frameworks for asymmetric catalysis. Such systems are capable of inducing high enantioselectivity across a range of mechanistically unrelated reactions (refs. 4a, 10). C) The addition of TMSBr to oxetanes catalyzed by chiral squaramide 1a can proceed simultaneously by fundamentally different, yet highly enantioselective Lewis- and Brønsted-acid mechanisms.

Figure 2.

A) Chiral squaramide-catalyzed addition of TMSBr to oxetanes (ref 9l). B) Original mechanistic hypothesis for the oxetane ring-opening reaction. C) Catalyst 1a reproducibly affords highly enantioenriched products, but reactions catalyzed by other, structurally-similar H-bond donors fail to provide reproducible levels of enantioselectivity.

Figure 3.

A) Effect of HBr (generated by the photochemical reaction of Br₂ in toluene) on reaction rate (black dots) and enantioselectivity (red dots) ([1a] = 0.002 M, [2]₀ = 0.1 M, [TMSBr]₀ = 0.2 M). The reaction with [HBr]=0 was conducted in the presence of TMSCHN₂ as a base (see text). Reaction rates were determined via in situ FTIR monitoring using a ReactIR (see General Procedure for ReactIR experiments in the Supporting Information). B) Enantioselective addition of HBr to 2 promoted by 1a.

Figure 4.

A) Reaction time-course measured using in situ FTIR monitoring. Upon the addition of TMSCHN₂ (green - 4000 s) the rate of consumption of 2 (black) decreases but continues at a lowered rate. Upon the addition of HCl (8000 s), the TMSCHN₂ is rapidly consumed. Following complete consumption of TMSCHN₂ the rate of oxetane consumption increases. B) The reaction run in the presence of TMSCHN₂ proceeds with high enantioselectivity, and at a rate that is slower but still competitive with the HBr co-catalyzed pathway.

Figure 5.

The addition of trace water allows for high reactivity and enantioselectivity with reduced loadings of squaramide 1a in a gram-scale reaction of oxetane 2 with TMSBr.

Figure 6.

A) Proposed reaction mechanisms consisting of competing Brønsted and Lewis acid pathways. B) Computed lowest energy major (R) and minor (S) transition states for the Brønsted acid pathway (ΔΔE^‡ = 2.6 kcal/mol). C) Computed lowest energy major (R) and minor (S) transition states for the Lewis acid pathway (ΔΔE^‡ = 2.5 kcal/mol). Transition states were optimized at SMD (Et₂O) – B97D/def2-SVP. The electronic energies were corrected by single-point refinement at SMD (Et₂O) – B97D3/def2-TZVP

Figure 7.

Predicted ΔΔE^‡ vs experimental ΔΔG^‡ for the Lewis acid pathways for catalysts 1a-1g and the Brønsted acid pathway for catalyst 1a. See Figure 8 for catalyst structures, Table S11 for tabulated data, and Figures S7 and S34 for experimental details.

Figure 8.

A) Enantioselectivity for the reaction of 2 catalyzed by 1a-1g determined both for the HBr-promoted reaction and with TMSBr in the presence of TMSCHN₂ (average of two runs, see Figures S7 and S33 for details). B) Graphical representation of relative enantioselectivities of the Brønsted and Lewis acid mechanisms for reactions of 2 catalyzed by 1a-1g.

Figure 9.

A) Electrostatic potential map of catalyst 1a using the geometry from TS_silyl-R (scale from −0.03 to +0.03). B) Electrostatic potential map of the Lewis acid transition state using the geometry from TS_silyl-R (scale from −0.06 to +0.06). C) Electrostatic potential map of the Brønsted acid transition state using the geometry from TS_protic-R (scale from −0.06 to +0.08). All ESPs were computed at SMD (Et2O) – B97-D3/Def2-TZVP and plotted with a density isovalue of 0.0004.