Molecular recognition of carbonyl compounds using aluminum tris(2,6-diphenylphenoxide) (ATPH): New regio- and stereoselective alkylation of α,β-unsaturated carbonyl compounds [19]

Full text of DOI:10.1021/ja000789d

J. Am. Chem. Soc. 2000, 122, 7847-7848
Molecular Recognition of Carbonyl Compounds
7847
Using Aluminum Tris(2,6-diphenylphenoxide)
(ATPH): New Regio- and Stereoselective Alkylation
of r,â-Unsaturated Carbonyl Compounds
Susumu Saito, Masahito Shiozawa, Takashi Nagahara,
Masakazu Nakadai, and Hisashi Yamamoto*
Graduate School of Engineering, Nagoya UniVersity, CREST
Japan Science and Technology Corporation (JST)
Furo-cho, Chikusa, Nagoya 464-8603, Japan
ReceiVed March 6, 2000
The regio- and stereoselective synthesis of R,â-unsaturated
carbonyl compounds is an important goal in organic synthesis.1
Carbonyl alkylation by dienolates of R,â-unsaturated carbonyl
compounds is a widely used general method,2 but problems are
frequently encountered controlling the regioselectivity (of both
the deprotonation and alkylation steps) and stereoselectivity of
the olefin geometry of the new double bond. We report here an
entirely new paradigm to solve this problem. The new method
depends strongly on aluminum tris(2,6-diphenylphenoxide) (AT-
PH)3,4 as the key reagent (Figure 1).
Figure 2. X-ray crystal structures [cylinder (upper) and CPK (lower)
models] of (a) ATPH-1 and (b) ATPH-3 complexes.
(E)-γ-products exclusively.4a,b None of the R-alkylated product
was obtained in either case. Varying the lithium amide or aldehyde
does not affect the E:Z selectivities (Scheme 1).
To explain the above striking results, the X-ray crystal
structures of the ATPH-1 and ATPH-3 complexes were
obtained (Figure 2). These two complexes adopt an s-trans
conformation.5 Whereas aldehyde 3 (Al-OdC angle (θ) of 193.9
(4)°) favors the anti-complexation (θ > 180°), ester 1 (θ ) 136.2
(3)°) shows syn-complexation (θ < 180°) (Table 1).6 The (Z)-
γ-methyl of 1 and the (E)-γ-methyl of 3 occupy sterically less
hindered space, i.e., rather outside of the cavity of ATPH (Figures
2 and 3).7
Figure 1. Molecular structure of ATPH.
If interconversion of the extended dienolate conformers (s-trans
a s-cis)5 is possible,8 and if the extended dienolates resemble
the corresponding ATPH-carbonyl complexes,8 the approach of
Sequential treatment of a toluene solution of ATPH (3.3 equiv)
with methyl 3-methyl-2-butenoate (1) (2.0 equiv) and benzalde-
hyde (1.0 equiv) at -78 °C was followed by deprotonation with
a THF solution of LTMP (2.3 equiv). The reaction mixture was
stirred for 30 min and quenched with aqueous NH4Cl to give
homoallyl alcohol 2a in 91% isolated yield (Scheme 1). The
(5) For the general aspects of the conformation (s-cis vs s-trans) of
unsaturated carbonyl compounds, see: Stereochemistry of Organic Com-
pounds; Eliel, E. L.; Wilen, S. H. Eds.; John Wiley & Sons: New York;
1994; Chapter 10.2., p 615. Also see the depiction below for two major
conformations of an R,â-unsaturated carbonyl compound and diene, respec-
tively. The definition for the conformations (s-trans and s-cis, where “s”
denotes a “single bond”) of an extended dienolate is based on that for a diene.
Scheme 1
(6) For a discussion of the coordination aptitude of a variety of Lewis acids
toward carbonyl compounds (e.g., syn- vs anti-coordination), see: (a)
Schreiber, S. L. in ComprehensiVe Organic Synthesis; Trost, B. M., Ed.;
Pergamon Press: Oxford; 1991; Vol. 1, Chapter 1.10 and references therein.
(b) Shambayati, S.; Crowe, W. E.; Schreiber, S. L. Angew. Chem., Int. Ed.
Engl. 1990, 29, 256. See below for the general definition of syn and anti
complexations of R,â-unsaturated carbonyl compounds, which denotes the
complexation mode of the Lewis acid (MLn) with respect to the orientation
of the R,â-double bond.
predominant alkylation site was at the (Z)-γ position of 1 ((Z)-
γ:(E)-γ ) 13:1). In sharp contrast, senecialdehyde (3) gave the
(1) See ref 3 of Supporting Information.
(7) To confirm this further, we measured 1H NMR shift changes (∆δ) from
free (1 and 3) to bound (ATPH-1 and ATPH-3) substrates. See Supporting
Information.
(2) See ref 4 of Supporting Information.
(3) See ref 5 of Supporting Information.
(4) The present γ-aldolization is not substrate-specific, but rather showed
substrate generality using â-substituted-R,â-unsaturated carbonyl compounds,
see: (a) Saito, S.; Shiozawa, M.; Ito, M.; Yamamoto, H. J. Am. Chem. Soc.
1998, 120, 813. (b) Saito, S.; Shiozawa, M.; Yamamoto, H. Angew. Chem.,
Int. Ed. 1999, 38, 1769.
(8) Full experimental details (NOE and NMR studies) corroborating the
mechanistic models in Figure 3 will be addressed later in a full paper. In fact,
a rapid equilibrium (s-trans a s-cis) faster than the rate of aldolization is
envisioned by complete reversal of the olefin configuration of esters (the E-γ-
methyl of (E)-6 was delivered to the Z-γ-methylene of 9).
10.1021/ja000789d CCC: $19.00 © 2000 American Chemical Society
Published on Web 07/26/2000

7848 J. Am. Chem. Soc., Vol. 122, No. 32, 2000
Table 1. Selected Metrical Data for ATPH Complexes
Communications to the Editor
emphasize that the (Z)-γ positions are under the significant
influence of the congested environment extended by the cavity
of ATPH, since the approach of not only electrophiles but also
of LDA is circumvented. Thus, both (E)- and (Z)-aldehydes 5
must be encapsulated effectively in an anti, s-trans coordination
similar to ATPH-3, and s-trans dienolate conformers I1 and I2
are more reactive species.10
In contrast, (E)- and (Z)-acid esters 6 showed γ-methyl-selective
deprotonation-aldolization regardless of the original olefin con-
figurations of the esters to give identical (Z)-product 9 (Scheme
3). Despite fast equilibrium between the corresponding dienolate
property
ATPH-1 (R ) OMe)
ATPH-3 (R ) H)
CdO, Å
1.249(5)
1.833(3)
136.2(3)
4.2
1.128(6)
1.810(3)
193.9(4)
16.9
Al-O (sp2), Å
θ, (deg)
φ, (deg)
Scheme 3
conformers,8 s-cis conformer I3 seems more reactive.11 These
results are in agreement with the coordination bias of ATPH-1.
Furthermore, the observed kinetic attack of LDA or LTMP, which
prefers deprotonation at a methyl over methylene, indicates that
ATPH has a negligible steric influence on the ester deprotonation
steps. Indeed, the cavity can no longer maintain its original
propeller-like structure (Figure 1) due to the coordination bias of
esters with relatively small θ (Figure 2 (a) and Table 1).
In summary, we have demonstrated that several â,â-disubsti-
tuted-R,â-unsaturated carbonyl substrates can be differentiated
by complexation with ATPH. We achieved diverse selectivity and
reactivity during the reactions, which involve the control of: (1)
syn- vs anti-complexation; (2) s-cis vs s-trans conformation of
carbonyl compounds; (3) the deprotonation site; (4) s-cis vs s-trans
conformation of extended dienolates; and (5) the alkylation site,
depending on the R,â-unsaturated carbonyl compounds.
Figure 3. Illustrative depictions of ATPH-3 and ATPH-1 complexes
and their enolate intermedates as more reactive conformers. E+ denotes
ATPH‚OdCHR1 or R1CHO, or other reactive species.
electrophiles (E+)9 should be influenced by significant steric
interactions with ATPH (Figure 3).
By taking advantage of the distinct recognition of ATPH with
carbonyl compounds, we next examined the regio- and stereo-
selectivity of substrates 5 and 6, which have different substituents
at the â-positions. In cases where (E)- and (Z)-aldehydes 5 were
subjected to general conditions using ATPH and LDA, anomalous
regio- and chemoselective outcomes were accommodated: (E)-
γ-methylene-selective aldolization predominated with (E)-5 to give
(E)-product 7, whereas the (E)-γ-methyl of (Z)-5 was alkylated
exclusively to give (E)-product 8 (Scheme 2). These experiments
Supporting Information Available: Experimental procedures, spec-
tral and analytical data for all new compounds and metrical data for ATPH
complexes (PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
Scheme 2
JA000789D
(10) To confirm the influence of ATPH on the dienolate equilibria,
deprotonation of ATPH-(E)-5 or ATPH-(Z)-5 with LDA was followed by
immediate workup (after ∼15 s) with excess MeOH at -78 °C. The original
E/Z ratios of (E)-5 (E:Z ) >99:1) and (Z)-5 (E:Z ) <3:97) were partially
preserved to give E/Z ratios of 80:20 and 19:81, respectively, suggesting that
the equilibrium shifted to I1 and I2. Note that the rate of aldolization between
two carbonyl species at -78 °C is rapid enough that these quenching
experiments do not directly reflect the actual product distributions.
(11) A similar workup experiment10 with the extended dienolates derived
from ATPH-(E)-6 and ATPH-(Z)-6 both resulted in the regeneration of (E)-6
and (Z)-6 in a ratio of ∼80:20.
(9) At present, we have no clear evidence to identify the actual electrophilic
species. Although ATPH‚R1CHO is formally possible, decomplexed R1CHO
or other possibilities are also nonnegligible.