Zimmermann-Traxler model

The asymmetric synthesis of (—)-denticulatin A (30) shows an interesting application of the boron aldol chemistry (Scheme 6) [23]. In a group-selective aldol reaction between the weso-aldehyde 27 and (5)-28, the hydroxyalde-hyde 29 was formed with > 90 % de, which spontaneously cyclized to the lactol 31. The configuration at the stereocenters of C-2 and C-3 in 29 is in accordance with the induction through the sultam auxiliary as well as with preference of an a-chiral aldehyde to react to the ant/-Felkin diastereomer in an aldol reaction which is controlled by the Zimmermann-Traxler model [24, 25]. 

The two syn aldols (57) and (58) are enantiomers (provided there is no additional asymmetry in A or B), as are the anti aldols (59) and (60). The syn anti outcome depends fundamentally on the geometry of the enolate and can be predicted on the basis of the six-membered cyclic transition state known as the Zimmermann-Traxler model. 

These examples have been discussed in the framework of the Zimmermann-Traxler chair models, without considering the distortions of the chairs that can noticeably modify the relative free-energy differences. Moreover, in some aldol reactions, boat-shaped transition models better account for the observed stereoselectivities [122, 123], Ab initio calculations have supported such boat transition models however, the calculated energy differences are rather small [41, 124], Molecular mechanics calculations have also been applied to these problems, by, among others, Bemardi, Gennari and Paterson [124, 125, 126], and some general trends were highlighted. 

Under kinetic control, the reactions of prochiral aldehydes with Z-enolates generally lead to syn aldols, while E-enolates lead to anti aldols. The presence of bulky R groups on the enolates, however, may alter these selectivities. The highest diastereoselectivities are observed with boron or titanium enolates. These selectivity trends are interpreted by a concerted cyclic mechanism. The favored transition state resembles a distorted chair, in line with the Zimmermann-Traxler proposals [57, 160, 253] (Figure 6.70). This model has been supported by theoretical studies [9, 40, 41, 125, 1249], Transition states analogous to C2 and C4 (Figure 6.70) are destabilized by 1,3-ecIipsing interactions between the C-R, M-L and C-R bonds, so that models Ci and C3 are more favorable. For the sake of simplification, only the reaction on one face of the enolates is shown in these models, but enolate face selectivity will be discussed later. In some cases, boatlike transition-state models are invoked to interpret selectivity inversions [401, 402, 666], Moreover, Heathcock and coworkers [105] obtained evidence for the influence of an excess of n-B BOTf on the stereoselectivity of the aldol reactions of Z-enolates. In such reactions, anti aldols can be formed preferentially (see bdow). 

Agami s model was subsequently challenged by List, Lerner, and Barbas III in 2000 [8a], when they proposed a one-proline enamine mechanism for the proline-catalyzed intermolecular aldol reaction between ketones and aldehydes. Shortly afterwards, on the basis of DFT calculations, Houk and co-workers proposed a very similar mechanism for the Hajos-Parrish intramolecular aldol [19]. Using the B3LYP/6-31H-G(2df,p) level of DFT theory, Houk and co-workers [20] have seen that the energy difference between the two possible chair Zimmermann-Traxler-like transition states, which differ in the orientation of the enamine with... 

---