using allyl iodide and benzyl iodide and obtained the desired
products in 99.5 and 93% ee (Table 2, entries 3 and 5).
Application of two substituted benzyl iodides in this protocol
also gave excellent % ee and satisfactory yields (Table 2,
entries 7 and 9).
intermediate that racemizes slowly on the reaction time
7c
scale. To account for formation of an enantiomerically pure
enolate from (S)-6b, we have computationally modeled its
reaction with dimeric lithium dimethylamide. Dimer-based
mechanisms for deprotonation of ketones and esters by LDA
and LiHMDS have been identified experimentally,13 and
open dimer mechanisms for deprotonation of ketones have
We reasoned that the moderate alkylation yields attained
in the 1 min deprotonation protocol might be due to
incomplete enolate formation. Since lengthening the depro-
tonation time would increase the extent of enolate racem-
ization, we sought conditions that would allow enolate to
form in the presence of electrophile. Because this in situ
protocol would also allow instantaneous trapping of the
enolate, we anticipated an improvement in enantioselectivity.
This in situ strategy has been successfully applied to the
synthesis of quaternary Ala-derived N-PMB 1,4-benzodiaz-
14
been studied computationally. Transition structures leading
to the (M)- and (P)-enolates were located at B3LYP/6-31G*;
single-point energies were calculated at B3LYP/6-31+G*
(Figure 2). Transition state (M)-(S)-12b leading to the (M)-
8
b
epin-2-ones. An obvious key requirement for this protocol
is that the base (present in excess) not react with the
electrophile at -100 °C. This requirement was met through
the use of KHMDS8b and allylic and benzylic bromides
(rather than iodides). As can be seen in Table 2, application
of the in situ protocol for reaction of allyl bromide gave
significantly improved enantioselectivity (Table 2, entry 2,
cf. “sequential” and “in situ” columns).
Similarly, the in situ protocol for benzylic bromides
generally gave higher yields and enantioselectivities than the
sequential protocol with the corresponding iodides (Table
2, cf. entries 4 and 5, 6 and 7, 8 and 9). Finally, application
of the in situ protocol to methylation gave very poor yields,
suggesting competitive reaction of KHMDS with MeOTf
(Table 2, entry 1, cf. “sequential” and “in situ” columns). In
summary, for methylation and allylation, the sequential 1
min LDA protocol is recommended. In all other cases, the
in situ KHMDS protocol gives superior yields (75-92%)
and enantioselectivity (95->99.5% ee).
As was the case for Ala- and Phe-derived 1,4-benzodiaz-
8
epin-2-ones, the deprotonation/alkylation of 6b appears
uniformly retentive. The absolute configurations of (+)-7b
and 9b were determined by hydrolysis to the known
quaternary amino acids, (S)-(-)-R-Me-Pro-OH and (R)-(-
)
-R-Bn-Pro-OH. Retentive (R)-stereochemistry is assigned
Figure 2. B3LYP/6-31G* transition structures for deprotonation
of (S)-6b by (LiNMe ) . Selected bond lengths in Å; relative free
energies at B3LYP/6-31+G*.
to (+)-10b and (+)-11b based on the positive rotation of
2
2
1
13
(R)-9b. H and C NMR spectroscopy of 7b-11b indicate
a single conformation of the BZD ring, in sharp contrast to
quaternary 1,4-benzodiazepin-2-ones, which exist as mixtures
8,10,11
enolate is favored by 11.9 kcal/mol (173 K). Examination
of the corresponding explicit bis(Me O) solvates (M)-(S)-
2
of the (M)- and (P)-conformers.
B3LYP/6-31G* cal-
culations12 of the possible equatorial proline (M)-(S)- and
and (P)-(S)-13b similarly indicated a 9.7 kcal/mol (173 K)
preference for formation of the (M)-enolate. Thus, the
formation of an enantiopure (M)-enolate by deprotonation
of (S)-6b appears feasible.
Because the solution structure of the enolate intermediate
derived from (S)-6b is not yet known, we considered two
limiting structures: free anion (M)-14b and the corresponding
axial proline (P)-(S)-conformations of 7b indicate a 25.8 kcal/
mol preference for the former at 25 °C. This strong
preference is due in large part to amide resonance, which is
retained in the equatorial-proline (M)-(S)-conformation, but
compromised in the axial-proline (P)-(S)-conformation (O-
C5-N4-C3 dihedrals of -174.6 and -73.7°, respectively).
Successful memory of chirality transformation requires the
formation of an enantiopure conformationally chiral reactive
2 3
Li(OMe ) salt (M)-15b. Both the equilibrium geometries
of these species and their corresponding ring-inversion
transition structures 14b* and 15b* were located at B3LYP/
6-31G* (Figure 3). B3LYP/6-31+G*//B3LYP/6-31G* ac-
tivation free energies for racemization of 14b and 15b are
12.2 and 16.0 kcal/mol, corresponding to racemization half-
(
10) Lam, P. C.-H.; Carlier, P. R. J. Org. Chem. 2005, 70, 1530-
538.
11) We use the helical descriptors (M)- and (P)- to assign the chirality
of the ring, according to the sign of the C2-N1-C7-C6 dihedral angle
M ) minus, P ) positive).
1
(
(
Org. Lett., Vol. 7, No. 23, 2005
5307