Origins of the double asymmetric induction on proline-catalyzed aldol reactions

Article abstract of DOI:10.1021/jo800934b

(Chemical Equation Presented) Computational studies to elucidate the origin of the double asymmetric induction on proline-catalyzed aldol reaction have been performed using HF/6-31G(d) calculations. The computed transition structures explain the experimen

Full text of DOI:10.1021/jo800934b

Origins of the Double Asymmetric Induction on Proline-Catalyzed
Aldol Reactions
Fe´lix Caldero´n,† Elisa G. Doyagu¨ez,† Paul Ha-Yeon Cheong,‡
Alfonso Ferna´ndez-Mayoralas,*,† and K. N. Houk*,‡
Instituto de Qu´ımica Orga´nica General, CSIC, Juan de la CierVa 3, 28006 Madrid, Spain, and
Department of Chemistry and Biochemistry, UniVersity of California, Los Angeles, California 90095-1569
mayoralas@iqog.csic.es; houk@chem.ucla.edu
ReceiVed May 14, 2008
Computational studies to elucidate the origin of the double asymmetric induction on proline-catalyzed
aldol reaction have been performed using HF/6-31G(d) calculations. The computed transition structures
explain the experimental data obtained.
Introduction
induction. When (R)-aldehyde 2 and ketone 3 reacted in the
presence of (S)-proline, anti-(S, S) aldol 4 was the only product
obtained. However, (R)-proline catalysis afforded the anti(R,
R) aldol as major product with lower stereoselectivity (d.r. 5:1).
This topicity was confirmed when the aldehyde enantiomer ent-2
was used, affording the enantiomeric aldol products. These
results indicate that the inherent selectivities of the two chiral
species are mutually reinforcing with the (R)-aldehyde/
(S)-proline or with (S)-aldehyde/(R)-proline matched pairs. The
mismatched case results upon reversing the absolute configu-
ration of either of the two chiral components.
We observed similar effects in the synthesis of 5-membered
ring azasugars 15, a potent R-galactosidase inhibitor (Scheme
2).7,8 The reaction of 8 and 3 was first tried following the
described conditions,6 however very low yields of desired
products were obtained. It is known that enamine-based direct
cross-aldol reactions involving linear aldehydes as acceptor and
ketones as donor is difficult.9 The problem lies in the difficulties
for the catalyst to differentiate between the aldehyde and the
ketone to form the nucleophilic enamine, resulting in undesirable
aldehyde-aldehyde dimers. To overcome this problem, the al-
dehyde was slowly added to a preformed solution of proline
From the discoveries of Hajos, Parrish, Eder, Sauer, and
Wiechert1,2 in the early 1970s, to the recent discoveries of List,
Barbas, and Lerner,3,4 the proline-catalyzed aldol reaction has
been shown to be a versatile method for C-C bond formation.
Interesting asymmetric syntheses have been accomplished using
this methodology, such as the synthesis of carbohydrates,
steroids, and useful building blocks.5 We have recently con-
tributed to this field by reporting a novel synthesis of six-
membered azasugars from diethyl tartrate (1), where the key
step is a proline-catalyzed aldol reaction between the chiral
aldehyde 2 with dioxanone 3 (Scheme 1).6 During the course
of this study we observed an interesting double asymmetric
† CSIC.
‡ University of California.
(1) (a) Hajos, Z. G.; Parrish, D. R. J. Org. Chem. 1973, 38, 3239–3243. (b)
Hajos, Z. G.; Parrish, D. R. J. Org. Chem. 1974, 39, 1615–1621.
(2) (a) Eder, U.; Sauer, G.; Wiechert, R. Angew. Chem., Int. Ed. Engl. 1971,
10, 496–497.
(3) (a) List, B.; Lerner, R. A.; Barbas, C. F., III. J. Am. Chem. Soc. 2000,
122, 2395–2396. (b) Notz, W.; List, B. J. Am. Chem. Soc. 2000, 122, 7386–
7387. (c) Bui, T.; Carbas, C. F., III. Tetrahedron Lett. 2000, 41, 6951–6954. (d)
Sakthivel, K.; Notz, W.; Bui, T.; Barbas, C. F., III. J. Am. Chem. Soc. 2001,
123, 5260–5267.
(4) For a complete revision see: (a) Alleman, C.; Gordillo, R.; Clemente,
F. R.; Ha-Yeon Cheong, P.; Houk, K. N. Acc. Chem. Res. 2004, 37, 558–569.
(5) (a) Berkessel, A.; Gro¨ger, H. In Asymmetric Organocatalysis; VCH:
Weinheim, 2004. (b) Dalko, P. I.; Moisan, L. Angew. Chem., Int. Ed. 2004, 43,
5138–5175. (c) Caldero´n, F.; Ferna´ndez, R.; Sa´nchez, F.; Ferna´ndez-Mayoralas,
A. AdV. Synth. Cat. 2005, 347, 1395–1403.
(6) Caldero´n, F.; Doyagu¨ez, E. G.; Ferna´ndez-Mayoralas, A. J. Org. Chem.
2006, 71, 6258–6261.
(7) Stu¨tz, A. E. In Iminosugars as Glycosidase Inhibitors: Nojirimycin and
Beyond; Wiley-VCH: Weinheim, Germany, 1999.
(8) Mengel, A.; Reiser, O. Chem. ReV. 1999, 99, 1191–1223.
(9) List, B.; Pojarliev, P.; Castello, C. Org. Lett. 2001, 3, 573–575.
7916 J. Org. Chem. 2008, 73, 7916–7920
10.1021/jo800934b CCC: $40.75 2008 American Chemical Society
Published on Web 09/24/2008

Proline-Catalyzed Aldol Reactions
SCHEME 1. Synthesis of Azasugars from Diethyltartrate
SCHEME 2a
SCHEME 3. Model System for Theoretical Study
R-azide aldehyde 8 was modeled with (R)-2-azidopropanal 18
(Scheme 3b). The three rotamers of 17 and 18, involving the
rotation around C1-C2 bond, were studied. The reaction of these
model electrophiles and the (S)- and (R)-proline-enamine of
dioxanone 19 and 20 were computed using HF/6-31G* calcula-
tions.10 The influence of the enamine ring conformation in the
course of the nucleophilc attack was studied for all cases using
semiempirical PM3.
Previous quantum mechanical studies have shown that the
proline aldol mechanism involves a nucleophilic attack of
a Reaction conditions: (a) p-TsOH, 2.2-dimethoxypropane, 24 h, r.t., 50%;
(b) PPTS, DMF, 2 h, r.t., 100%; (c) BnBr, NaH, 3 h, r.t., 85%. (d) HCl,
MeOH, 2 h, r.t., 82%; (e) NaIO4, MeOH/water, 75 min, r.t., 83%.
and ketone. Using this protocol, the reaction of 8 and 3 in the
presence of (S)-proline gave anti-(3S, 4S)-aldol 14 with good
yield (60%, Scheme 2) and excellent stereoselectivity (d.r. >
20:1). On the other hand, the reaction of 8 and 3 in the presence
of (R)-proline gave anti-(3R, 4R)-aldol 16 as major product but
with low yield (10%) and moderate stereoselectivity (d.r. 5:1).
Results and Discussions
To explain the origins of the double asymmetric induction
of these proline-catalyzed aldol additions, we explored the
stereoselectivity of this reaction using quantum mechanical
calculations on a model system. The R-hydroxy aldehyde ent-2
was modeled with (S)-2-methoxypropanal 17 (Scheme 3a), and
FIGURE 1. Transition state models for the proline-catalyzed aldol
reaction.
J. Org. Chem. Vol. 73, No. 20, 2008 7917

Caldero´n et al.
FIGURE 2. Most stable transition structures for aldol reaction of 17 with 19 (∆H298 ) kcal/mol) from (S)-proline. First row structures correspond
to re-facial attack and the second the si-attack. An optimized structure for TS S-si-2 was not found.
FIGURE 3. Most stable transition states for aldol reaction of 17 with 20 (∆H298 ) kcal/mol) from (R)-proline. First row structures correspond to
re-facial attack and the second the si-attack.
the neutral enamine and simultaneous proton transfer from the
carboxylic acid to the carbonyl acceptor.4,11 Transition states
with an approximate dihedral angle +60° ((S)-Pro) and -60°
((R)-Pro) relative to the carbonyl group and the enamine double
bond are favored by almost 10 kcal/mol. Transition states with
the carboxylic acid group and the enamine double bond syn
are 2-10 kcal/mol higher in energy than the transition states
involving the anti-enamine (Figure 1a). Only the anti-enamine
was considered in this study.
It is well-known that the nucleophilic attack follows a partial-
Zimmerman-Traxler-like12 transition state but the metal that
forms the sixth atom in the chair is missing. The re-face of the
acceptor aldehyde, where the aldehyde substituent is located in
a pseudoequatorial arrangement, is preferred for the nucleophilic
7918 J. Org. Chem. Vol. 73, No. 20, 2008

Proline-Catalyzed Aldol Reactions
repulsion between the methyl of 17 and the enamine of 19 in
[TS S-re-2] that is not present in [TS R-si-3].
This hypothesis was confirmed from a similar theoretical
study with model R-azide-aldehyde, 18. The most stable
transition structure in this case was also one in which there was
minimal steric interaction between the proline enamine and the
aldehyde substrate (Figure 4).
In Table 1, the computational results involving the model
systems shown in Scheme 3 are compared with experimental
results. There is an excellent agreement between the computed
and the experimental stereoselectivity.
In conclusion, as previously reported, the stereoselectivity
of the aldol reaction catalyzed by proline is governed by the
preference for a transition state where the aldehyde substituent
is located in a pseudoequatorial position (partial-Zimmerman-
Traxler transition state). However, when an aldehyde bearing a
R-stereogenic center is used as acceptor, steric factors have been
shown to be determining from the computational evidence
obtained. The reaction catalyzed by (R)-proline follows the
normal Felkin-Anh course, with nucleophilic attack anti to the
largest group, the medium group inside, and the smallest outside,
as in TS R-si-3; however, with (S)-proline, this TS is sterically
hindered, and a lower (mismatched) stereoselectivity is observed.
HF/6-31G(d) has shown to be an excellent method/basis set
for these calculations.
FIGURE 4. Most stable transition structures for aldol reaction of 18
with 19 and 20. First row: (S)-proline: re (a) and si (b) facial attack.
Second row: (R)-proline: re (c) and si (d) facial attack.
TABLE 1. Predicted and Experimental ∆H Values (kcal/mol) for
the Transition States of Proline-Catalyzed Reactions (17 versus 2,
and 18 versus 8)
Experimental section
∆H298
anti/syn
predicted
for 17
∆Hexp
anti/syna
experimental
for 2
∆H298
anti/syn
predicted
for 18
∆Hexp
anti/syna
experimental
for 8
General Procedure for the Catalytic Asymmetric Aldol
Reaction of Ketone 3 and Aldehyde 8 Using (R)- or (S)-
Proline. A solution of (R)- or (S)-proline (5.5 mg, 0.048 mmol) in
DMF (0.2 mL) was stirred for 24 h. Afterward, ketone 3 (0.13 g,
1 mmol) was added, and the mixture was allowed to react for 3 h.
Then, aldehyde 8 was added portionwise (0.10 g, 0.48 mmol) in
three additions of 0.16 mmol solved in 33 µL of DMF during 48 h.
After the last addition, the mixture was stirred at room temperature
for 96 h. Then, NH4Cl (sat) (1 mL) was added and the mixture
was extracted with AcOEt (3 × 2 mL). The organic layer was dried
(Na2SO4) and the solvent was evaporated. The residue was purified
by column chromatography (hexane/AcOEt 4:1). Yields and anti/
sin diastereoselectivities are shown in scheme 3.
(3S,4S,5R)-5-Azido-6-O-benzyl-1,3,4-trihydroxy-1,3-O-isopro-
pylidene-hexan-2-one (14). Column chromatography (hexane/
AcOEt 4:1). [R]D25 -22.1° (c 0.14, CH2Cl2); 1H NMR (300 MHz,
CDCl3, 298 K): δ 7.4-7.3 (m, 5H), 4.7-4.5 (m, 2H), 4.38 (dd, J
) 8.7, 1.5 Hz, 1H), 4.28 (d, J ) 17.7 Hz, 1H), 4.07 (d, J ) 17.7
Hz, 1H), 4.0-3.9 (m, 1H), 3.9-3.8 (m, 2H), 3.76 (d, 1H, J ) 3.9
Hz), 1.51 (s, 3H), 1.41 (s, 3H). 13C NMR (75 MHz, CDCl3, 298
K): δ 212.5 (C), 138.1 (C), 128.9 (CH), 128.3 (CH), 127.4 (CH),
101.9 (C), 74.1 (CH2), 72.5 (CH), 70.3-70.2 (CH, CH2), 66.9
(CH2), 60.8 (CH), 23.9 (CH3), 23.8 (CH3). EM (IES-EM): m/z 358.0
[M+ + 23]; Anal. Calcd. for C16H21N3O5: C, 57.30; H, 6.31; N
12.53. Found: C 57.28; H 6.21; N 12.86.
(S)-Proline
(R)-Proline
1.42
>3
0.95
>3
>3
1.15
>3
0.95
a Calculated from the ratio of aldol products as determined by 1H
NMR.
attack in the (S)-proline transition state. The same reasoning
can be applied to the transition states of (R)-proline but, in this
case, the attack on the si-face of the aldehyde places the
substituent in equatorial orientation.
First, we examined the TS corresponding to the nucleophilic
attack of 19 and 20 on (S)-aldehyde 17.
As shown in Figure 2, the most stable transition state [TS
S-re-2] involves the addition of the anti-enamine to the re-face
of 17. The si-face attacks are disfavored due to the steric
interactions between the substituents of 17 and the enamine of
19. Other re-face attack TS [TS S-re-1] and [TS S-re-3] are
higher in energy than [TS S-re-2] because the former involves
repulsion between the oxygens of the methoxy group of the
aldehyde and the enamine, and the latter involves the repulsion
between the largest group (Me) and the proline-ring.
The transition states with (R)-proline are shown in Figure 3.
In the transition structures involving the (R)-proline enamine,
the facial selectivity of 17 switches as compared to the (S)-
proline case, and now the si-face attacks are favored. A reface
attack involves substantial steric interactions between the
enamine and the substituents of 17. The most stable transition
structure is [TS R-si-3]. This transition structure is more stable
than analogous [TS R-si-1] and [TS R-si-2] due to the steric
interactions between the oxygen of 20 and either the oxygen or
methyl of 17.
(3R,4R,5R)-5-Azido-6-O-benzyl-1,3,4-trihydroxy-1,3-O-isopro-
pylidene-hexan-2-one (16). Column chromatography (hexane/
25
1
AcOEt 4:1). [R]D +25.1° (c 0.7, MeOH); H NMR (300 MHz,
CDCl3, 298 K): δ (major isomer) 7.4-7.3 (m, 5H), 4.7-4.5 (m,
2H), 4.39 (dt, J ) 8.7, 1.5 Hz, 1H), 4.28 (d, J ) 17.7 Hz, 1H),
4.09 (d, J ) 17.7, 1H), 4.0-3.9 (m, 1H), 3.9-3.8 (m, 2H), 3.77
(dd, J ) 3.9, 1.5 Hz, 1H), 1.51 (s, 3H), 1.41 (s, 3H). 13C NMR (75
MHz, CDCl3, 298 K): δ (major isomer) 212.5 (C), 137.8 (C), 128.8
(CH), 128.1 (CH), 127.2 (CH), 101.8 (C), 73.8 (CH2), 72.3 (CH),
70.1-70.0 (CH, CH2), 66.7 (CH2), 60.7 (CH), 23.8 (CH3), 23.7
(CH3). EM (IES-EM): m/z 358.0 [M+ + 23]; Anal. Calcd. For
C16H21N3O5: C, 57.30; H, 6.31; N 12.53. Found: C 57.25; H 6.27;
N 12.72.
The stereoselectivity is lower in the case of (S)-proline (Figure
2), as compared to (R)-proline (Figure 3), because there is steric
J. Org. Chem. Vol. 73, No. 20, 2008 7919

Caldero´n et al.
(2R,3S,4R,5S)-3,4-dihydroxy-2,5-bis(hydroxymethyl)pyrroli-
dine hydrochloride [2,5-dideoxy-2,5-imino-D-galactitol (15) •HCl].
To a solution of 14 (0.055 mmol) in MeOH (2.5 mL), Pd/C (10%)
was added (20 mg). The suspension was stirred under H2 (45 psi)
during 18 h at room temperature. Afterward, conc. HCl (0.3 mL)
was added, and the mixture was stirred under H2 (45 psi) for 6 h
at room temperature. After this time, the reaction mixture was
filtered through Celite, and the solvent was evaporated to give 15.
Spectroscopic data for 15 are consistent with those described in
the literature. 13 EM (IES-EM): m/z 164.1 [M+-Cl-], 217.3 [M+ +
23]; Anal. Calcd. for C6H14ClNO4: C, 36.10; H, 7.07; N, 7.02.
Found: C, 36.38; H, 7.02; N, 7.35.
Acknowledgment. We thank Dr. F. R. Clemente, Dr. C. D.
Anderson, and C. Y. Legault for their helpful discussions and
suggestions and the UCLA Academic Technology Service and
CESGA for computer resources. F.C. and E.G.D. thank Min-
isterio de Educacio´n y Ciencia and CSIC for predoctoral
fellowships (BQU2001-1503 and I3P-BPD2005, respectively).
Financial support by the Spanish Government is gratefully
acknowledged (grant CTQ2004-03523/BQU).
(10) All the transition states were fully optimized and characterized by
frequency analysis at the HF level of theory with the 6-31G(d) basis set as
implemented in Gaussian 98 (see Supporting Information for a complete author
list). Previous work in related reactions has shown the efficiency of this method
and this basis set: Ha-Yeon Cheong, P.; Ahang, H.; Thayumanavan, R.; Tanaka,
Supporting Information Available: Experimental details for
the synthesis of compounds 8. 1H and 13C NMR copies of new
compounds. Cartesian coordinates of all reported structures. This
material is available free of charge via the Internet at
http://pubs.acs.org.
F.; Houk, K. N.; Barbas, C. F. Org. Lett. 2006, 8, 811–814. Enthalpies, ∆H298
,
were computed for the gas phase.
(11) (a) Clemente, F. R.; Houk, K. N. Angew. Chem., Int. Ed. 2004, 43,
5766–5768. (b) Bahmanyar, S.; Houk, K. N.; Martin, H. J.; List, B. J. Am. Chem.
Soc. 2003, 125, 2475–279. (c) Bahmanyar, S.; Houk, K. N. J. Am. Chem. Soc.
2001, 123, 11273–11283. (d) Bahmanyar, S.; Houk, K. N. J. Am. Chem. Soc.
2001, 123, 12911–12912.
(12) Zimmerman, H. E.; Traxler, M. D. J. Am. Chem. Soc. 1957, 79, 1920–
1923.
JO800934B
(13) (a) Singh, S.; Han, H. Tetrahedron Lett. 2005, 45, 6349. (b) Doyagu¨ez,
E. G.; Caldero´n, F.; Sa´nchez, F.; Ferna´ndez-Mayoralas, A. J. Org. Chem. 2007,
72, 9353–9356.
7920 J. Org. Chem. Vol. 73, No. 20, 2008