with esters of phenylglyoxylic acid. Using several aldehydes,
the expected 1,2-diols were isolated with a high degree of
syn- diastereoselectivity.9 Later on, the same authors pub-
lished an aldol reaction of mandelic acid esters with
aldehydes in the presence of TiCl4 and amines.10 This method
is limited to the use of only p-bromobenzaldehyde. We have
been able to optimize and extend this procedure by using
catalytic amounts of amines in the presence of titanium(IV)
alkoxides. An optimized protocol is given in Table 1. Aldol
When used with catalytic amounts of cinchona alkaloids,
aldol adducts were isolated in high yields and with a high
degree of syn diastereoselectivity. The syn-configured aldol
products were found in their racemic form. The use of larger
amounts of 1,2-aminoalcohols was connected with a change
in diastereoselectivity. By using equimolar amounts, we
obtained a preference for the anti-configured aldol adducts.
The best results so far were obtained by using N-methyl-
ephedrine. By reacting racemic methyl mandelate 4 with
aldehydes 1a-k in the presence of titanium(IV) tert-butoxide
and optically active methylephedrine, we isolated the anti-
configured aldol adducts 5a-k in high yields as well as
enantioselectivities (Table 2). By using (+)-N-methylephed-
Table 1. Catalytic and Diastereoselective Aldol Additiona
Table 2. Enantioselective Aldol Additiona
yieldb
(%)
ratioc
(syn/anti)
entry
R
compd
1
2
3
4
5
6
7
8
Ph
3a
3b
3c
3d
3e
3f
88
86
78
75
80
70
82
60
88:12 (95:5)d
95:05
94:06
79:21
80:20
55:45
57:43
50:50
t-Bu
iso-Pr
Et
c-Hex
Me
yieldb
compd (%)
ratioc
eed (anti)
entry
R
(syn/anti) (configuration)e
1
2
3
4
5
6
Ph
5a
5c
5d
5e
5i
98
88
89
85
92
82
4:96
42:58
51:49
61:39
40:60
10:90
93 (S,S)
72 (S,S)
61 (S,S)
58 (S,S)
67 (S,S)
92 (S,S)
iso-Pr
Et
c-Hex
n-Pr
Ph-CH2-CH2-
Ph-(CHdCH)-
3g
3h
a Reaction conditions: 1 equiv of aldehyde, 1 equiv of isopropyl
mandelate, 1 equiv of Ti(OiPr)4, rt, 5 mol % of Et3N. b Isolated yields.c The
diastereoselectivity was determined by 1H NMR analysis and by X-ray
structure analysis of 3c. d Diisopropyl ethylamine.
p-BrC6H4-
5k
a Reaction conditions: 1 equiv of aldehyde, 1 equiv of methyl mandelate,
1 equiv of Ti(OtBu)4, rt, 2 equiv of (-)-N-methylephedrine. b Isolated yields.
c The diastereoselectivity was determined by H NMR analysis. d Enanti-
1
oselectivities were determined by HPLC on Chiralpak AS and 1H NMR
analysis using the Mosher ester technique. d The absolute configuration was
established by X-ray structure analysis of 5k.
adducts 3a-h of even enolizable aldehydes 1c-g and
mandelic acid esters were isolated with high yields and with
a high degree of syn diastereoselectivity after 1-2 h at room
temperature. The yields of this reaction did not depend on
the influence of mandelic esters or titanium(IV) alkoxides
used. A slightly increased diastereoselectivity was observed
when using diisopropyl ethylamine (entry 1, Table 1).
Meerwein-Ponndorf reductions11 or Tishchenko products12
were not observed under these conditions. When using chiral
mandelic acid esters, no enantioselectivities in the aldol
adducts could be detected. Complete racemization was
observed in each of our reactions.13
Next, we focused our attention on the enantioselective
execution of this procedure. Several chiral amines and
diamines were used in these reactions without any success
in regards to enantioselectivities. In further experiments, we
tested several chiral 1,2-aminoalcohols in these reactions.
rine or (-)-N-methylephedrine, we were able to obtain both
anti-configured enantiomeres, (R,R)- and (S,S)-5a-k. No
transesterification14 was detectable under these reaction
conditions.
As pointed out above, the enantio- and diastereoselection
that were observed did not depend on the configuration of
the mandelic acid esters used. The same enantio- as well as
(13) For a study of controlled racemization of mandelic acid derivatives,
see: (a) Ebbers, E. J.; Ariaans, G. J. A.; Bruggink, A.; Zwanenburg, B.
Tetrahedron: Asymmetry 1999, 10, 3701-3718. (b) The assumption of a
chiral induction is based on the following observations. The aldol reaction
is completed to the full within 1 h. After the same period, only 20% of the
starting chiral isopropyl mandelate has been racemized in comparative
experiments of controlled racemization. See Supporting Information. The
reason for the total racemization we observed after 1 h is still unclear.
Controlled experiments of chiral aldol adduct 3a in the presence of Ti(Oi-
Pr)4 also indicate a racemization. An exhaustive racemization of 3a at room
temperature was detected after 15 min. Currently, we are not able to
discriminate between a racemization of chiral aldol products and a racemic
aldol addition.
(14) (a) Seebach, D.; Hungerbuehler, E.; Naef, R.; Schnurrenberger, P.;
Weidmann, B.; Zueger, M. Synthesis 1982, 138-140. (b) Seebach, D.;
Weidmann, B.; Widler, L. In Modern Synthetic Methods; Scheffold, R.,
Ed.; Verlag Sauerlaender: Aarau, 1983; pp 217-353. (c) Shapiro, G.; Marzi,
M. J. Org. Chem. 1997, 62, 7096-7097. (d) dos Santos, A. R.; Ferreira,
M. de L. G.; Kaiser, C. R.; Ferezou, J.-P. Eur. J. Org. Chem. 2005, 15,
3348-3359.
(9) (a) Clerici, A.; Porta, O. J. Org. Chem. 1982, 47, 2852-2856. (b)
Clerici, A.; Porta, O. J. Org. Chem. 1983, 48, 1690-1694. (c) Clerici, A.;
Clerici, L.; Malpezzi, L.; Porta, O. Tetrahedron 1995, 51, 13385-13400.
(10) (a) Clerici, A.; Clerici, L.; Porta, O. Tetrahedron Lett. 1995, 36,
5955-5958. (b) Clerici, A.; Pastori, N.; Porta, O. J. Org. Chem. 2005, 70,
4174-4176.
(11) For reviews, see: (a) de Graauw, C. F.; Peters, J. A.; van Bekkum,
H.; Huskens, J. Synthesis 1994, 1007-1017. (b) Graves, D. R.; Campbell,
E. J.; Nguyen, S. T. Tetrahedron: Asymmetry 2005, 16, 3460-3468.
(12) For an overview, see: Mahrwald, R. Curr. Org. Chem. 2003, 7,
1713-1723.
5354
Org. Lett., Vol. 8, No. 23, 2006