External stereoselection

To date, direct asymmetric synthesis of optically active chiral-at-metal complexes, which by definition leads to a mixture of enantiomers in unequal amounts thanks to an external chiral auxiUary, has never been achieved. The most studied strategy is currently indirect asymmetric synthesis, which involves (i) the stereoselective formation of the chiral-at-metal complex thanks to a chiral inductor located either on the ligand or on the counterion and then (ii) removal of this internal chiral auxiliary (Fig. 4). Indeed, when the isomerization of the stereogenic metal center is possible in solution, in-... 

A theoretical analysis is presented for the binding of the four dia-stereoisomers of benzo[a]pyrene diol epoxides (BPDEs) to N2(g), N6(a), 06(G) and NU(c). Molecular models for binding and stereoselectivity involving intercalation, intercalative covalently and externally bound forms are presented. Molecular mechanics calculations provide the energetics which suggest possible structures for the formation of each of the principal DNA-BPDE complexes. Stereographic projections are used to illustrate the molecular structures and steric fits. The results of previous calculations on intercalation and adduct formation of BPDE l(+) in kinked DNA (37) are summarized and extended to include the four diastereoisomers l( ) and II( ). The theoretical model is consistent with the observed experimental data. 

NU(C) base atoms (5) The stereoselectivity of the BPDEs during intercalative covalent binding in kinked DNA and (6) The possible reorientation of the complex to yield an externally bound adduct. The energetics for each of these processes will be presented to identify the important steps that influence the binding of specific isomers. It will be shown that the orientation of each diastereoisomer of BPDE about specific base atoms in kinked receptor sites in the duplex DNA during covalent bond formation is the determining factor in stereoselectivity. 

Although we postulate that this receptor site results in stereoselectivity, it may not be the final state. The orientation of the long axis of the pyrene moiety is approximately 80°-90° and this implies quasi-intercalation of site IQ (56-58). The kinked site proposed by Hogan et al. (50) and studied by Taylor and Miller (MO for l(+)-N2(G) binding in retrospect appears to represent different binding sites. The orientation of the pyrene moiety of a(BPDE) = 1+3° and the local DNA axis in the kink of y(DNA) = 29° (50) will be explained by external binding in the next section. The intercalative covalent binding in a kinked site is an intermediate step between intercalation and the final structure for the externally bound BPDE-N2(G) adduct, but it becomes the final structure for the quasi intercalated BPDE-N6(g), 06(g) and N +(c) adducts. 

Externally and internally bound adducts. The internally bound BPDEs have been presented in a step in which the favored proton assisted Sjj2 reaction results in the trans addition and the stereoselectivity of isomers toward base atoms. The DNA is dynamic. We propose a relaxation which yields both internally and externally bound adducts the former favoring binding to n6(a), 06(g) and N t(C), and the latter favoring N2(G). 

The stereoselected Cda conformation of the BPDE i(-) and Il(-) adducts to N6(a) were chosen for study in a reoriented complex with an externally bound pyrene moiety. In Figure 13, the adduct is shown in its optimum orientation in B-DNA with adenine after an anti - syn transformation for which the non-bonded contacts are poor, and with the normal anti base orientation with favorable contacts. The fit improves for the anti base as ax 30°. The orientation of the pyrene moiety is a(BPDE) =31° and the local helical axis of the DNA is oriented at y(DNA) = 15° Calculations were not performed with externally bound BPDE-DNA adducts to 06(G) and NU(C). Calculations of externally bound BPDE I(-)-N6(a) adducts with kinked DNA with ax + 30° yields an orientation a(BPDE) = 31° in good agreement with experimental results for the externally bound component (51). 

The main features of this proposed mechanism are (l) the stereoselectivity of the BPDEs by the DNA during intercalative covalent binding and (2) the final orientation of the bound pyrene which may be oriented internally (intercalative covalent) or externally (outside the helix). The stereoselectivity occurs during covalent bond formation and after intercalation. Relaxation of the DNA allows the adduct to adjust to its final orientation. If the experimental measurements are assumed to be made on the DNA-adducts after the final orientation has been achieved, then the following interpretations can be made. 

The l(+) and Il(+) isomers are stereoselected by N2(G), whereas the i(-) and II(—) isomers are stereoselected by the n6(a) and 06(G) during intercalative covalent steps with trans addition. The l( + )-and Il(+)-N2(G) adducts are rearranged to an externally bound form with the pyrene in the minor groove, but the I(-)-N6(a) and II(-)-06(g) adducts remain quasi intercalated. This is determined by the relative energy change between the two forms as we see from Table XIII. However, there is a superposition of the two types of sites, IQ and IIX (51 57,58), and BPDE i(-) DNA adducts exhibit both types of binding. By symmetry, the cis BPDE l(-)-N2(G) adduct is predicted to behave similarily to the trans l(+)-N2(G) adduct. It should be externally bound. 

Catalytic properties of external chiral additives such as (2S,3/ )-4-dimethyl-amino-l,2-diphenyl-3- methyl-2-butoxide (A 16) (574, 575) and 2-magnesium-3-zinc salts of dialkyl (f ,f )-tartrate (A17) were employed in the highly stereoselective addition of organozinc reagents to derivatives of 3,4-dihydro-isoquinoline-A-oxide (Scheme 2.147) (576). 

In the enantioselective synthesis, the asymmetry (i.e., the stereoselectivity) is induced by the external chiral catalyst, while the diastereoselective synthesis does not require a chiral catalyst. The stereogenic center already present in the molecule is able to induce stereoselectivity, assuming that the synthesis starts with a single enantiomer. For instance, imagine that an a,/ -substituted product is formed, and that the reactant already contains a stereogenic carbon at a. If the reaction of (aS) leads, e.g., largely to (aS, / R) and hardly to the (aS, /IS) diastereomer (i.e., stereoisomers that are not mirror-images of each other), the reaction is diastereoselective (Scheme 14.2). 

The asymmetric synthesis of allenes by stereoselective manipulations of enantio-merically pure or enriched substrates relies on the availability of such optically active substrates. In contrast, a direct synthesis of allenes by the reaction of prochiral substrates in the presence of an external asymmetric catalyst is an almost ideal process [102]. Most of the catalytic asymmetric syntheses in organic chemistry involve the creation of chiral tetrahedral carbon centers [103], whereas the asymmetric synthesis of allenes requires the construction of an axis of chirality. 

If stoichiometric quantities of the chiral auxiliary are used (i.e., if the chiral auxiliary is covalently bonded to the molecule bearing the prochiral centres) there are in principle three possible ways of achieving stereoselection in an aldol adduct i) condensation of a chiral aldehyde with an achiral enolate ii) condensation of an achiral aldehyde with a chiral enolate, and iii) condensation of two chiral components. Whereas Evans [14] adopted the second solution, Masamune studied the "double asymmetric induction" approach [22aj. In this context, the relevant work of Heathcock on "relative stereoselective induction" and the "Cram s rule problem" must be also considered [23]. The use of catalytic amounts of an external chiral auxiliary in order to create a local chiral environment, will not be considered here. 

The lactone concept is not restricted to the simple model biaryl synthesis presented here. It has been successfully expanded to a broad series of structurally diverse biaryl substrates (e.g., lactones with additional stereocenters and functional groups, configurationally stable lactones, seven-membered lactones, and again configurationally unstable biaryl hydroxy aldehydes ), to different activation modes in the ring-opening step (e.g., use of metallated nucleophiles, carbonyl activation by Lewis acids, (Ti -complexation, etc.), and for various strategies of stereoselection (e.g., external vs. internal asymmetric induction). ... 

The observed stereoselectivity of this ring cleavage reaction [60] favors a mechanism, Eq. (65), in which formation of a Iead(IV) homoenolate by cleavage of the external bond with inversion of configuration is followed by rapid cleavage of the second bond. 

Sung et al. [71] attempted to explain the reasons for the observed stereoselectivities on the basis of the mechanism outlined in Scheme 1.25 it is assumed that the cyclic intermediate 68 is formed under thermodynamic control and can equilibrate to the more stable isomer before being attacked by the external nucleophile. When the aminoacid has a bulky side chain (for example Y = i-Pr), the cyclohexyl intermediate 68 will preferentially dispose this chain in the equatorial position. Between the two possible diastereoisomers, the most favored one will be that with R1 in the equatorial position 68a and therefore this will be formed preferen-... 

Reduction using a large excess of Raney nickel in methanol is extremely regioselective only the internal double bond is affected. This is possibly because of the highly hindered nature of the external disubstituted double bond, which is flanked by the tricyclic ring system and the, V-BOC substituent. The transformation occurred stereoselectively from the less hindered or-face of the molecule producing the required R configuration. 

---