An amino alcohol ligand for highly enantioselective addition of organozinc reagents to aldehydes: Serendipity rules

Article abstract of DOI:10.1021/ol0259488

(matrix presented) Amino alcohol 4 (or its enantiomer) is prepared in two simple steps. Commercial (1R,2S)-2-amino-1,2-diphenylethanol is dialkylated with bis(2-bromoethyl) ether. Subsequent hydrogenation over 5% Rh on alumina in the presence of morpholine unexpectedly stops at the hexahydro derivative 4. Amino alcohol 4 promotes the enantioselective addition of diethylzinc to aldehydes at room temperature in up to 99% enantiomeric excess.

Full text of DOI:10.1021/ol0259488

ORGANIC
LETTERS
2002
Vol. 4, No. 13
2133-2136
An Amino Alcohol Ligand for Highly
Enantioselective Addition of Organozinc
Reagents to Aldehydes: Serendipity
Rules
William A. Nugent
Bristol-Myers Squibb Company, Process Research and DeVelopment Department,
Chambers Works, P.O. Box 269, Deepwater, New Jersey 08023-0269
william.nugent@bms.com
Received March 29, 2002
ABSTRACT
Amino alcohol 4 (or its enantiomer) is prepared in two simple steps. Commercial (1R,2S)-2-amino-1,2-diphenylethanol is dialkylated with
bis(2-bromoethyl) ether. Subsequent hydrogenation over 5% Rh on alumina in the presence of morpholine unexpectedly stops at the hexahydro
derivative 4. Amino alcohol 4 promotes the enantioselective addition of diethylzinc to aldehydes at room temperature in up to 99% enantiomeric
excess.
The enantioselective addition of organozinc reagents to
aldehydes1 catalyzed by enantiopure â-amino alcohols
remains the focus of much research activity as evidenced
by hundreds of research papers and several comprehensive
review articles.2 Perhaps because of its archetypical natures
this transformation represents the enantioselective counterpart
of the Grignard addition reaction3sit remains a favored
testing ground for novel amino alcohol ligands.
Several years ago we reported4 the use of parallel synthesis
techniques to explore the structure-activity relationship of
â-amino alcohol ligands in asymmetric catalysis. The reaction
selected for study was the addition of diethylzinc to benz-
aldehyde (eq 1). Some relevant observations from that study
our study, we found that a six-membered ring on nitrogen
(piperidinyl or morpholinyl) in the tertiary amine substructure
was optimal in all cases. As might be expected, both the
N-terminal and O-terminal substituents play a significant role
in controlling the enantioselectivity of eq 1. The prototypical
“stripped down” ligand 1 provided virtually no enantio-
selectivity. Introduction of a phenyl group at the N-terminal
position (ligand 2) sharply increases enantioselectivity to 83%
while sterically “inflating” the methyl group of 1 to the level
of a cyclohexyl substituent (ligand 3) increases the enantio-
selectivity to 67%. This led to the proposal that ligand 4,
which incorporates both of these structural features, might
provide even higher ee values for eq 1.
(1) Oguni, N.; Omi, T. Tetrahedron Lett. 1984, 25, 2823.
(2) Pu, L.; Yu, H.-B. Chem. ReV. 2001, 101, 757. Soai, K. Niwa, S.
Chem ReV. 1992, 92, 833. Noyori, R. Asymmetric Catalysis in Organic
Synthesis; Wiley: New York, 1994; Chapter 5.
(3) Grignard, V. C. R. Hebd. Sceances Acad. Sci. 1900, 130, 1322.
(4) Nugent, W. A.; Licini, G.; Bonchio, M.; Bortolini, O.; Finn, M. G.;
McClelland, B. W. Pure Appl. Chem. 1998, 70, 1071.
are shown in Figure 1. For the series of amino alcohols in
10.1021/ol0259488 CCC: $22.00 © 2002 American Chemical Society
Published on Web 05/25/2002

Scheme 1. Preparation of Amino Alcohol 4 and Its Fully
Hydrogenated Analogue 8
Figure 1. Observed enantioselectivity for addition of diethylzinc
to benzaldehyde (eq 1) using a series of â-amino alcohols prepared
in parallel synthesis studies.
This would be the case if we assume that the contributions
from the O-terminal and N-terminal substituents are inde-
pendent and additive. This concept was first applied to
modular ligand design by Snapper and Hoveyda5 who noted
“The described strategy assumes that the influence of each
ligand subunit is independent and additive: it is impossible
to rule out cooperative effects without individually testing
each combination”.
In fact, the literature contains examples of ligands which
share the flat/bulky structural motif with 4 and which provide
high enantiomeric excesses in eq 1. A notable example is
ligand 5 which was reported6 by Pericas and co-workers to
presence of triethylamine (DMSO, 25 °C) resulted in
dialkylation of the amine nitrogen to afford (1R,2S)-7 in 74%
yield after crystallization from hot toluene. Attempted
hydrogenation of both phenyl rings of 7 using 5% Rh/
alumina catalyst (4% HOAc/MeOH, 25 °C, 60 psi) in the
presence of morpholine (1 equiv) instead reproducibly
afforded 4 (71% yield after crystallization from heptane).
For details, see the Supporting Information.
In the absence of the morpholine additive, hydrogenation
of 7 was sluggish. Complete conversion was achieved by
raising the reaction temperature to 50 °C but the product
was that derived from benzylic C-N cleavage, 1,2-dicyclo-
hexylethanol. The fully saturated dicyclohexyl amino alcohol
8 could be prepared by reversing the order of the steps, i.e.,
hydrogenation of 67 followed by dialkylation with 2-bro-
moethyl ether.
A sample of (1S,2R)-4 was prepared in analogous fashion
from (1S,2R)-6. An X-ray crystal structure (Figure 2)
confirmed that hydrogenation had occurred at the O-terminal
end, leaving the N-terminal phenyl ring intact.
Ligand 4 promotes the enantioselective addition of dieth-
ylzinc to benzaldehyde (eq 1). Using a 2-fold excess of
diethylzinc in 2:1 hexane/toluene and 5% of 4 as ligand, the
addition proceeded in 98% yield and 99% enantiomeric
excess (ee) Under identical conditions, ligands 7 and 8
promoted eq 1 in 89% and 92% ee, respectively.
promote eq 1 at room temperature in 91% enantiomeric
excess. The synthesis of 4 appeared complicated compared
with known ligands such as 5 and was not pursued in our
initial report. However, we have subsequently prepared 4 as
a result of a “happy accident”.
We previously described4 the synthesis of amino alcohol
7 by reaction of morpholine with enantiopure stilbene oxide
(90 °C, 7 d). More recently, we developed a faster route as
shown in Scheme 1. Treatment of commercial (1R,2S)-2-
amino-1,2-diphenylethanol with 2-bromoethyl ether in the
The addition of diethylzinc to a series of aldehydes using
(1R,2S)-4 as catalyst is summarized in Table 1. The reactions
(5) Cole, B. M.; Shimizu, K. D.; Krueger, C. A.; Harrity, J. P. A. Snapper,
M. L.; Hoveyda, A. H. Angew. Chem., Int. Ed. Engl. 1996, 35, 1668.
(6) Vidal-Ferran, A.; Moyano, A.; Pericas, M. A.; Riera, A. J. Org. Chem.
1997, 62, 4970.
(7) Complete hydrogenation of 2-amino-1,2-diphenylethanol to its di-
cyclohexyl analogue has been reported: Pankova, M.; Sicher, J. Collect.
Czech. Chem. Commun. 1965, 30, 388.
2134
Org. Lett., Vol. 4, No. 13, 2002

addition of diethylzinc to aromatic aldehydes proceeds in
high yield and with enantioselectivities in the range 98% to
99% ee. Several R-branched aliphatic aldehydes (entries
7-11) also afforded the corresponding acetates in 98% to
99% ee.
The diminished yield in the case of the tertiary alkyl
derivatives (entries 10 and 11) is the result of the ubiquitous
competing reductive pathway shown in eq 3. This side
reaction was intitially noted by Noyori and co-workers9 for
the case of diethylzinc addition to benzaldehyde. In our
experience, detectable amounts of primary alcohols are
always formed during the addition of organozinc reagents
to aldehydes in the presence of â-amino alcohols. This
pathway represents a significant side reaction in the case of
sterically bulky aldehydes bearing tertiary alkyl substituents.
In an effort to identify the limitations of ligand 4, we
extended our studies to include a heterocyclic aldehyde (entry
12) and an R,â-unsaturated aldehyde10 (entry 13); somewhat
surprisingly, these additions still proceeded with synthetically
useful enantiomeric excesses. Only in the case of a straight-
chain aliphatic aldehyde (entry 14) did the ee drop below
90%.
Figure 2. ORTEP diagram of amino alcohol (1S,2R)-4 showing
erythro structure and location of cyclohexyl substituent.
Table 1. Asymmetric Addition of Diethylzinc to Aldehydesa
entry
aldehyde in eq 2
yield (%)b
ee (%)
To determine the absolute stereochemistry of eq 2, the
product acetate esters were isolated in three cases where the
sign of optical rotation for the product has been assigned.
The products from benzaldehyde, hexanal, and cyclohexane-
carboxaldehyde all exhibited (+)-rotation, indicating that the
(R)-enantiomer had been formed in each case.11 On the basis
of the consistent order of elution observed for the product
enantiomers in Table 1, these are also tentatively assigned
as having (R)-stereochemistry.
The success of the flat/bulky structural motif for orga-
nozinc additions raised the question of whether further
increasing the steric bulk of the O-terminal substituent might
further enhance enantioselectivity. To this end, amino alcohol
9 was prepared from trans-â-bromostyrene by a sequence
of Ni-catalyzed cross-coupling with tert-BuMgCl, MCPBA
epoxidation, ammonia addition, and resolution (mandelic
acid) followed by alkylation with 2-bromoethyl ether. (The
complexity of this synthesis further underscores the efficiency
of our serendipitous route to 4.) Using representative
aldehydes and the standard conditions of Table 1, enantio-
1
2
3
4
5
6
7
8
9
10
11
12
13
14
benzaldehyde
m-tolualdehyde
p-tolualdehyde
m-anisaldehyde
p-fluorobenzaldehyde
p-chlorobenzaldehyde
isobutyraldehyde
98
97
96
97
98
97
93
91
93
73
78
94
95
96
99
99
98
98
99
98
98
99
98
99
99
96
94
87
2-ethylbutyraldehyde
cyclohexanecarboxaldehyde
trimethylacetaldehydec
2,2-dimethyl-4-pentenalc
3-thiophenecarboxaldehyde
methacrolein
hexanald
a All reactions contained aldehyde (3.0 mmol), Et2Zn (6.0 mmol), amino
alcohol (1R,2S)-4 (0.15 mmol), and tert-butylbenzene as internal standard
(300 µL) in 2:1 hexane-toluene (9 mL) at room temperature for 3 h except
as indicated; for details, see the Supporting Information. b Yield and ee
determined by gas chromatography on Cyclodex B stationary phase (J&W
Scientific). c 24 h run at room temperature. d 3 h run at 0 °C.
in Table 1 were quenched by addition of acetic anhydride
(eq 2), which quantitatively converts the intermediate zinc
(8) Nugent, W. A. Chem. Commun. 1999, 1369.
(9) Kitamura, M.; Okada, S.; Suga, S.; Noyori, R. J. Am. Chem. Soc.
1989, 111, 4028.
(10) Oguni and co-workers have previously reported enantioselective
addition of organozinc reagents to methacrolein using amino alcohol
catalysts: Hayashi, M.; Kaneko, T.; Oguni, N. J. Chem. Soc., Perkin Trans.
1 1991, 25.
(11) For assignments see, the following. (R)-(+)-1-Phenyl-1-propyl
acetate: Faraldos, J.; Arroyo, E.; Herradon, B. Synlett 1997, 367. (R)-(+)-
3-Octyl acetate: Mihailovic, M. Lj.; Mamuzic, R. I.; Zigic-Mamuzic, Lj.;
Bosnjak, J.; Cekovic, Z. Tetrahedron 1967, 23, 215, (S)-(-)-1-Cyclohexyl-
1-propyl acetate: Levene, P. A.; Marker, R. E. J. Biol. Chem. 1932, 97,
379.
alkoxides to the corresponding acetate esters and allows
direct analysis of the product mixtures by chiral capillary
column gas chromatography.8 As exemplified by entries 1-6,
Org. Lett., Vol. 4, No. 13, 2002
2135

lent methods14 are now available to prepare this epoxide. In
addition to organozinc additions, â-amino alcohols are
broadly useful chiral auxiliaries in asymmetric synthesis15
and we anticipate additional uses for 4, reflecting its ready
synthesis. In truthsat least in this casesserendipity rules.
Acknowledgment. This work was carried out in DuPont
Central Research under the sponsorship of the DuPont
Pharmaceuticals Company. The author thanks Mr. David M.
Lattomus for skilled technical assistance. Thanks are also
due to Dr. William J. Marshall for determining the X-ray
crystal structure of (1S,2R)-4.
selectivities for addition of diethylzinc were generally lower
for ligand 9 than those observed for ligand 4. For details,
see the Supporting Information.
The reason hydrogenation of 7 stops at the stage of the
mono-cyclohexyl derivative 4 is unclear at this point. The
formation of 1,2-dicyclohexylethanol as the overwhelming
product in the absence of morpholine is especially striking
when one considers that morpholine is the expected coprod-
uct of this hydrogenolysis. (The importance of morpholine
as an additive came to light when we changed the synthesis
of 7 as noted above. Evidently, excess morpholine in the
crude 7 used in initial studies was sufficient to promote the
selective hydrogenation.) Conceivably, the directing effect12
of the hydroxyl group provides a significant acceleration for
hydrogenation of the O-terminal phenyl ring in the presence
of the morpholine catalyst poison.
Supporting Information Available: Detailed descrip-
tions of experimental procedures, characterization of new
compounds, and X-ray data for compound 4. This material
is available free of charge via the Internet at http://pubs.acs.org.
OL0259488
(13) Although addition of morpholine to trans-stilbene oxide was rather
slow at 90 °C in our initial studies, a subsequent report by Pericas suggests
that this addition may be greatly accelerated in the presence of lithium
perchlorate. We have not yet tested this possibility. Sola, L.; Reddy, K. S.;
Vidal-Ferran, A.; Moyano, A.; Pericas, M. A.; Rieras, A.; Alvarez-Larena,
A.; Piniella, J.-F. J. Org. Chem. 1998, 63, 7078.
(14) See, for example: Chang, H.-T.; Sharpless, K. B. J. Org. Chem.
1996, 61, 6456. Tu, Y.; Wang, Z.-X.; Shi, Y. J. Am. Chem. Soc. 1996,
118, 9806. Chang, S.; Galvin, J. M.; Jacobsen, E. N. J. Am. Chem. Soc.
1994, 116, 6937.
Both enantiomers of 2-amino-1,2-diphenylethanol are
commercially available. As noted above, 4 can also be
prepared13 from enantiopure trans-stilbene oxide and excel-
(12) Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. ReV. 1993, 93,
1307.
(15) For a review, see: Ager, D. J.; Prakash, I.; Schaad, D. R. Chem.
ReV. 1996, 96, 835.
2136
Org. Lett., Vol. 4, No. 13, 2002