Cyclic-Cram’s model

As outlined in Section D.2.3.5., the stereochemical outcome of the addition of nucleophilic reagents to chiral aldehydes or ketones is rationalized most plausibly by the Cram-Felkin-Anh model. On the other hand, the corresponding reactions of oxygen- or nitrogen-heterosub-stituted aldehydes or ketones may be interpreted either by the same transition state hypothesis or, alternatively, by Cram s cyclic model. 

If a chiral aldehyde, e.g., methyl (27 ,4S)-4-formyl-2-methylpentanoate (syn-1) is attacked by an achiral enolate (see Section 1.3.4.3.1.), the induced stereoselectivity is directed by the aldehyde ( inherent aldehyde selectivity ). Predictions of the stereochemical outcome are possible (at least for 1,2- and 1,3-induction) based on the Cram—Felkin Anh model or Cram s cyclic model (see Sections 1.3.4.3.1. and 1.3.4.3.2.). If, however, the enantiomerically pure aldehyde 1 is allowed to react with both enantiomers of the boron enolate l-rerr-butyldimethylsilyloxy-2-dibutylboranyloxy-1-cyclohexyl-2-butene (2), it must be expected that the diastereofacial selec-tivitics of the aldehyde and enolate will be consonant in one of the combinations ( matched pair 29), but will be dissonant in the other combination ( mismatched pair 29). This would lead to different ratios of the adducts 3a/3b and 4a/4b. 

The stereochemical course of this reaction can be rationalized by Cram s cyclic model of asymmetric induction in which lithium is coordinated between the imine nitrogen and the 2-alkoxy group. 

These results may be explained either by Cram s cyclic model in the case of lithium alkyls or by Cornforth s dipolar model if copper-boron trifluoride reagents are used. Boron trifluoride causes double complexation of both nitrogen and oxygen atoms which results in the formation of an adduct with rigid antiperiplanar conformation due to electrostatic repulsion (see 4 and 5)9. 

The computational support for Felkin s torsional strain model and its success in interpretation of experimental diastereoselectivities has led to its widespread adoption. It appears to be the preeminent open transition state involved in reductions when chelation is not important. Complementary selectivity observed in reductions that do involve chelation may be understood in terms of Cram s cyclic model. 

Novel nonchelation phenomena are observed with a steroidal a-hydroxy aldehyde. The reaction of a lithium or magnesium alkynide with the aldehyde gives the (20/, 22/ )-diastereomer piedominantiy, the formation of which was explained by Cram s cyclic model. When BF3-OEt2 is added to the lidiium alkynide prior to the addition of the aldehyde, the stereoselectivity is inverted, and the (20) ,225)-isomer is obtained as the principal product. Transformation of a-alkoxy aldehyde to the boron ate complex is suggested. Other l wis acids, such as B(OMe)3, AlCh, etc., are less effective (equation 29). °... 

A high degree of stereoselectivity can be realized under chelation control, where an oxygen atom of an ether function (or more generally a Lewis base) in the a-, P- or possibly y-position of carbonyl compounds can serve as an anchor for the metal center of a Lewis acid. Since Cram s pioneering work on chelation control in Grignard-type addition to chiral alkoxy carbonyl substrates [30], a number of studies on related subjects have appeared [31], and related transition state structures have been calculated [32], Chelation control involves Cram s cyclic model and requires a Lewis acid bearing two coordination sites (usually transition metal-centered Lewis acids). 

Scheme 1.2. Intraligand vs. interligand asymmetric induction (a) Diastereoselective addition via Cram s cyclic model ([49], cf., Section 4.2). (b) Asymmetric synthesis of a pure enantiomer via diastereoselective addition to a carbonyl with a chiral auxiliary [50].

The selectivity of the reactions illustrated in Scheme 1.4 are rationalized by Cram s cyclic model, discussed in Section 4.2. 

Figure 4,11. Cram s cyclic model for asymmetric induction. L and S are large and small substituents, respectively [2,6]).

If the organometallic reagent is capable of chelation, the second model becomes operative. This model, sometimes called Cram s cyclic model [147] involves the assistance of a che-... 

A short synthesis of L-( — )-rhodinose (635), the trideoxyhexose subunit of the antibiotic streptolydigin, takes advantage of the propensity of Grignard reagents to add to lactaldehydes under chelation control (Cram s cyclic model) to produce 5y -diols. 

A study of the stereochemical outcome of the addition of lithium enolates to a-alkoxyaldehydes has shown that the predominant product is not that predicted by application of Cram s cyclic model for asymmetric induction. Assuming the alkoxy-group to be the largest group a- to the aldehyde, the major product is that predicted by Felkin s model (Scheme 59). ... 

The additions of nucleophiles to aldehydes and ketones are promoted by coordination of a Lewis acid to the oxygen atom of the carbonyl group. The coordination with the metal enhances the electrophilicity of the C=0 group facilitating the attack of the nucleophile. From a stereochemical point of view, the presence of a Lewis acid is particularly important when a substituent with a heteroatom able to coordinate with the metal is placed next to the carbonyl group. In such cases, the prediction of the stereoselectivity of the reaction requires a chelated reactive conformation as that represented in Figure 4.2. This model is known as Cram s cyclic model and again the attack of flie nucleophile takes place preferentially from the less-hindered side. 

The stereochemical outcome of these reactions can be rationalized by means of a chelated Cram s cyclic model M ( /( ) (Scheme 11), where the N-Cbz group is the chelating ligand and the ju-tolylthio residue acts as the stereocontrolling large group. 

---