Cram chelate model, additions

In his original paper,2 Cram disclosed an alternative model that rationalizes the preferred stereochemical course of nucleophilic additions to chiral carbonyl compounds containing an a heteroatom that is capable of forming a complex with the organometallic reagent. This model, known as the Cram cyclic or Cram chelate model, has been extensively studied by Cram9 and by others,410... 

As a Stereochemical Prohe in Nucleophilic Additions. Historically, the more synthetically available enantiomer, (4R)-2,2-dimethyl-l,3-dioxolane-4-carhoxaldehyde, has been the compound of choice to probe stereochemistry in nucleophilic additions. Nevertheless, several studies have employed the (45)-aldeh-yde as a substrate. In analogy to its enantiomer, the reagent exhibits a moderate si enantiofacial preference for the addition of nucleophiles at the carbonyl, affording anti products. This preference for addition is predicted by Felkin-Ahn transition-state analysis, and stands in contrast to that predicted by the Cram chelate model. Thus addition of the lithium (Z)-enolate shown (eq 1) to the reagent affords an 81 19 ratio of products with the 3,4-anti relationship predominating as a result of preferential si-face addition, while the 2,3-syn relationship in each of the diastere-omers is ascribed to a Zimmerman-Traxler-type chair transition state in the aldol reaction. ... 

When the boron ligands and the aldehyde are both chiral, the inherent stereoselectivities of each partner may be either matched or mismatched (Chapter 1). In principle, a chiral aldehyde could derive facial selectivity from either the Felkin-Anh-Heathcock model (Figures 4.8 and 4.10) or the Cram-chelate model (Figure 4.11). However, because the boron of these reagents can accept only one additional ligand, chelation is not possible. Therefore only the Felkin-Anh-Heathcock effects... 

Scheme 2.11 Cram chelate model for the Grignard addition.

In additions of hydride donors to a-chiral carbonyl compounds, whether Cram or anti-Cram selectivity, or Felkin-Anh or Cram chelate selectivity occurs is the result of kinetic control. The rate-determining step in either of these additions is the formation of a tetrahedral intermediate. It takes place irreversibly. The tetrahedral intermediate that is accessible via the most stable transition state is produced most rapidly. However, in contrast to what is found in many other considerations in this book, this intermediate does not represent a good transition state model for its formation reaction. The reason for this deviation is that it is produced in an... 

After Cram had discovered the selectivities now named after him, he proposed the transition state model for the formation of Cram chelate products that is still valid today. However, his explanation for the preferred formation of Cram products was different from current views. Cram assumed that the transition state for the addition of nucleophiles to a-alkylated carbonyl compounds was so early that he could model it with the carbonyl compound alone. His reasoning was that the preferred conformation of the free a-chiral carbonyl compound defines two sterically differently encumbered half-spaces on both sides of the plane of the C=0 double bond. The nucleophile was believed to approach from the less hindered half-space. 

Less acidic than Ti and Zi chloroderivatives, MeTi(OPr )3 perfoims chelation-controlled addition to chiral alkoxy ketones as well as or better than organomagnesium compounds, but fails to chelate to aldehydes or hindered ketones. Should the formation of a cyclic chelation intermediate be forbidden, the reaction is subject to nonchelation control, according to Ae Felkin-Anh (or Comforth) model. Under these circumstances, the ratio of the diastereomeric products is inverted in favor of the anti-Cram product(s). In the case of benzil (83 Scheme 7) this can be accounted for by the unlikely formation of a cyclic intermediate such as (85), and thus the preferential intermediacy of the open chain intermediate (86) that leads to the threo compound (88). This view is substantiated by the fact that replacement of titanium with zirconium, which is characterized by longer M—O bonds, restores the possibility of having a cyclic intermediate and, as a consequence, leads to the erythro meso) compound (87) thus paralleling the action of Mg and Li complexes. 

Despite the great deal of attention devoted to nucleophilic additions to a-chiral carbonyls, the source of stereoselectivity in these reactions (predicted by Cram s rules of asymmetric induction ) remains largely unresolved. Neither direct structural studies nor correlation of reactant and product stereochemistries have yielded any conclusive support for a single comprehensive model. Similarly, the effect of Lewis acids on these systems is only understood at the level of chelation-controlled additions (vide infra). 

Threo diastereoselectivity is consistent with a chelation-controlled (Cram cyclic model) organolithium addition (Figure 8a). Since five-membered chelation of lithium is tenuous, an alternative six-membered chelate involving the dimethylamino nitrogen atom of the thermodynamically less stable (Z)-hydrazone (in equilibrium with the ( )-isomer) cannot be discounted. The trityl ether (entry 4, Table 9) eliminates the chelation effect of the oxygen atom such that the erythro diastereomer predominates (via normal Felkin-Ahn addition) (Figure 8b). 

Scheme 4.4. Eliel s asymmetric addition to carbonyls using Cram s chelate model. Table 4.4. Asymmetric addition of nucleophiles to oxathianes and oxazines.

The addition of benzylmagnesium chloride to 130 at —78 °C is strongly influenced by chelation of the a-hydroxy center with the magnesium cation diastereofacial selectivity consistent with the Cram cyclic model therefore results in a 66% yield of the syn isomer 131 only. Interestingly, and for reasons not quite clear, the addition of allylmagnesium chloride proceeds with high diastereoselectivity to provide the anti isomer 133 as the major diaster-eomer (Scheme 31) [37]. 

The stereoselectivity of the addition of pinacolone enolsilane 1 to P-alkoxy aldehydes bearing two stereocenters depends on the ability of the metal to form intermediate chelates. Those metals that monocoordinate the carbonyl group form Fel-kin products and the stereochemistry of these aldols is predicted by the Felkin-Anh s model. For metals chelating both the carbonyl and alkoxy groups, anti-Felkin products are obtained. In these cases the cyclic-Cram s model has to be used to predict the stereochemical outcome of the reaction. Therefore, non-chelated (Felkin-Ahn) and chelated models (cyclic-Cram) have been successively applied to understand the stereochemistry of the final reaction products. 

With a-alkyl-substituted chiral carbonyl compounds bearing an alkoxy group in the -position, the diastereoselectivity of nucleophilic addition reactions is influenced not only by steric factors, which can be described by the models of Cram and Felkin (see Section 1.3.1.1.), but also by a possible coordination of the nucleophile counterion with the /J-oxygen atom. Thus, coordination of the metal cation with the carbonyl oxygen and the /J-alkoxy substituent leads to a chelated transition state 1 which implies attack of the nucleophile from the least hindered side, opposite to the pseudoequatorial substituent R1. Therefore, the anb-diastereomer 2 should be formed in excess. With respect to the stereogenic center in the a-position, the predominant formation of the anft-diastereomer means that anti-Cram selectivity has occurred. 

For carbonyl addition, three transition state models have been proposed the Felkin-Anh model36, the chelate Cram model37 and the dipolar Cornforth model37 . 

Very few examples of asymmetric 1,4-induction are reported in connection with the addition of acidic Ti complexes to chiral y-alkoxycarbonyl compounds. According to the Cram model, the chelation is expected to afford a flexible seven-membered ring intermediate, resulting in less efficient induction (equation 34). An early example of asymmetric 1,4-induction is provided by the reaction of o-phthalal-dehyde (96 equation 35) with 2 equiv. of MeTi(OPr )3 which affords an 83 17 molar mixture of racemic- 97) and meso-(98), whereas Ae analogous reaction with MeMgl leads to a 1 1 mixture of the above diastereomers. 

---