High Catalytic Activity of Chiral Amino Alcohol Ligands Anchored to Polystyrene Resins

Article abstract of DOI:10.1021/jo980708k

Enantiomerically pure (2S,3S)-2,3-epoxy-3-phenylpropanol (3) has been anchored to Merrifield resins with different degrees of cross-linking and functionalization. The resulting epoxy-functionalized resins 4 have been submitted to completely regioselective

Full text of DOI:10.1021/jo980708k

J . Org. Chem. 1998, 63, 6309-6318
6309
High Ca ta lytic Activity of Ch ir a l Am in o Alcoh ol Liga n d s An ch or ed
to P olystyr en e Resin s
†
‡
†
,†
†
Anton Vidal-Ferran, Nick Bampos, Albert Moyano, Miquel A. Peric a` s,* Antoni Riera, and
‡
J eremy K. M. Sanders
Unitat de Recerca en S ´ı ntesi Asim e` trica (URSA), Departament de Qu ı´ mica Org a` nica, Universitat de
Barcelona, Mart ı´ i Franqu e` s 1-11, 08028 Barcelona, Spain and University Chemical Laboratory,
Lensfield Road, Cambridge CB2 1EW, UK
Received April 15, 1998
Enantiomerically pure (2S,3S)-2,3-epoxy-3-phenylpropanol (3) has been anchored to Merrifield resins
with different degrees of cross-linking and functionalization. The resulting epoxy-functionalized
resins 4 have been submitted to completely regioselective (C-3 attack) and stereospecific ring-opening
with secondary amines [piperidine (a ), N-methylpiperazine (b), and cis-2,6-dimethylpiperidine (c)]
in the presence of lithium perchlorate to afford (2R,3R)-3-(dialkylamino)-2-hydroxy-3-phenylpropoxy
resin ethers 5a -c. The progress of these two processes has been monitored by 13C gel-phase NMR
spectroscopy. Polymer-supported amino alcohols 5a -c have been evaluated as catalytic ligands
in the enantioselective addition of diethylzinc to benzaldehyde, best results being obtained with
the cis-2,6-dimethylpiperidine containing ligand 5c. Analogously, (2R,3R)-3-(cis-2,6-dimethyl-
piperidino)-3-phenyl-1,2-propanediol (11) has been anchored to a 2-chlorotrityl chloride resin (Barlos
resin) in dichloromethane in the presence of diisopropylethylamine, and the anchoring process has
been also monitored by 13C gel-phase NMR spectroscopy. The resulting resin 12 has subsequently
been used as a chiral ligand in the catalytic addition at 0 °C of diethylzinc to a family of fourteen
representative aromatic and aliphatic aldehydes 8a -n , to afford the corresponding (S)-1-substituted
1
-propanols 10a -n with a mean enantiomeric excess of 92%.
In tr od u ction
to narrow the gap between the homogeneous and het-
erogeneous approaches, is fully warranted.
The strategy of attaching a chiral ligand onto a polymer
support offers several advantages in catalytic asymmetric
synthesis over the use of the same ligand in solution.
These advantages include the easy recovery and potential
recycling of the chiral catalyst, a highly simplified
product purification, and the possibility of carrying out
the desired transformation in continuous mode in a flow
reactor. According to this, many examples of asymmetric
catalysts that employ polymer bound chiral ligands have
been reported.1
We have recently described the synthesis of a family
of enantiomerically pure (1R,2R)-1-(dialkylamino)-1-
phenyl-3-(R-oxy)-2-propanols 1 from a readily available
21
enantiopure building-block. An iterative process, aimed
at the fine-tuning of their catalytic properties in the
enantioselective addition of diethylzinc to benzaldehyde,
allowed the identification of structural features key to
high catalytic activity and enantioselectivity. These
turned out to be the presence of a bulky protecting group
and choice of the dialkylamino substituent as a nitrogen-
-5
Chiral 1,2-amino alcohols have been attached to poly-
mers and subsequently used to catalyze the enantiose-
lective addition of organozinc derivatives to aldehydes
(6) Itsuno, S.; Fr e´ chet, J . M. J . J . Org. Chem. 1987, 52, 4140-4142.
(
7) Soai, K.; Niwa, S.; Watanabe, M. J . Org. Chem. 1988, 53, 927-
9
28.
8) Soai, K.; Niwa, S.; Watanabe, M. J . Chem. Soc., Perkin Trans. 1
1989, 109-113.
9) Watanabe, M.; Araki, S.; Butsugan, Y.; Uemura, M. Chem.
6
-19
in a heterogeneous system
or even in a continuous
(
reactor.20 Although some of these heterogeneous cata-
lysts induce the addition of diethylzinc to different
aldehydes with notable results, their catalytic activity
(
Express 1990, 5, 761-4.
(10) Itsuno, S.; Sakurai, Y.; Ito, K.; Maruyama, T.; Nakahama, S.;
Fr e´ chet, J . M. J . J . Org. Chem. 1990, 55, 304-310.
(turnover number) and enantioselectivity (enantiomeric
(
11) Soai, K.; Watanabe, M.; Yamamoto, A. J . Org. Chem. 1990, 55,
excess of the resulting alcohols) are consistently lower
than those recorded with their homogeneous counter-
parts. Thus, continuation of efforts in this area, aimed
4
1
8
832-4835.
(12) Soai, K.; Watanabe, M. Tetrahedron: Asymmetry 1991, 2, 97-
00.
(13) Watanabe, M.; Soai, K. J . Chem. Soc., Perkin Trans. 1 1994,
37-842.
*
To whom correspondence should be addressed. E-mail mpb@
(14) Liu, G.; Ellman, J . A. J . Org. Chem. 1995, 60, 7712-7713.
ursa.qo.ub.es.
(15) Dreisbach, C.; Wischnewski, G.; Kragl, U.; Wandrey, C. J .
†
University of Barcelona
University of Cambridge
Chem. Soc., Perkin Trans. 1 1995, 875-878.
(16) El moualij, N.; Caze, C. Eur. Polym. J . 1995, 31, 193-198.
(17) Seebach, D.; Marti, R. E.; Hintermann, T. Helv. Chim. Acta
1996, 79, 1710-1740.
‡
(1) Noyori, R. Asymmetric Catalysis in Organic Synthesis; J ohn
Wiley & Sons: New York, 1994, Ch. 8, pp 347-364.
2) Hermkens, P. H. H.; Ottenheijm, H. C. J .; Rees, D. Tetrahedron
996, 52, 4527-4554.
3) Blackburn, C.; Albericio, F.; Kates, S. A. Drugs Future 1997, 22,
007-1025.
4) Fr u¨ chtel, J . S.; J ung, G. Angew. Chem., Int. Ed. Engl. 1997, 35,
(
(18) Huang, W. S.; Hu, Q. S.; Zheng, X. F.; Anderson, J .; Pu, L. J .
Am. Chem. Soc. 1997, 119, 4313-4314.
1
1
1
1
(
(19) Halm, C.; Kurth, M. J . Angew. Chem., Int. Ed. Engl. 1998, 37,
510-512.
(
(20) Kragl, U.; Dreisbach, C. Angew. Chem., Int. Ed. Engl. 1996,
35, 642-644.
17-42.
(5) Shuttleworth, S. J .; Allin, S. M.; Sharma, P. K. Synthesis 1997,
(21) Vidal-Ferran, A.; Moyano, A.; Peric a` s, M. A.; Riera, A. J . Org.
Chem. 1997, 62, 4970-4982.
217-1239.
S0022-3263(98)00708-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 08/07/1998

6
310 J . Org. Chem., Vol. 63, No. 18, 1998
Vidal-Ferran et al.
Sch em e 1
containing six-membered ring.21 Compounds 2a ,b arose
from this process as those possessing the optimal struc-
tural features. A subsequent molecular modeling study
led to the design and synthesis of ligand 2c containing
the cis-2,6-dimethylpiperidino residue,22 which plays the
role of increasing the energy gap (with respect to 2a )
among the diastereomeric transition states leading to the
S and R enantiomers of 1-phenyl-1-propanol. In full
agreement with theoretical predictions, ligand 2c induced
high enantioselectivity in the addition of diethylzinc to
a broad range of aromatic and aliphatic aldehydes.
With the aim of incorporating the advantages of
heterogeneous catalysis onto ligands 2, we planned
anchoring them to polymers possessing a benzylic func-
tionalization through the primary hydroxyl group. As a
leading principle, we considered that the polymer body
would behave as a very bulky substituent, thus fulfilling
the structural requirement for high enantioselectivity we
had previously optimized. We report in the present paper
the anchoring of ligands 2 to chloromethyl- (Merrifield)
and to chloro-(2-chloro)trityl-substituted (Barlos) polymer
backbones, and the catalytic properties of the resulting
functionalized resins in the enantioselective addition of
diethylzinc to aldehydes.
romethane, dimethylformamide, and toluene. The last
one is a suitable solvent to use when testing the catalytic
properties of chiral amino alcohols in the enantioselective
addition of diethylzinc to aldehydes.
The linkage of the (2R,3R)-3-(dialkylamino)-2-hydroxy-
3
-phenylpropoxy moiety to the Merrifield resins was
26
achieved in two steps from enantiopure epoxy alcohol 3,
using as a relay the epoxy-functionalized resin 4, as
indicated in Scheme 1. It is worth noting that this
scheme is characterized by a high versatility, since many
different ligands 5 are available from a single resin 4 by
simple variation of the secondary amine used in the ring-
opening process. A range of Merrifield resins with
different initial levels of functionalization (1.2-3.5 mmol
of active chloride per gram of resin) and cross linkage
(1-2% divinylbenzene) was used in order to test the effect
of these parameters on the catalytic properties of the final
resins.
Epoxy alcohol 3 was first deprotonated with an excess
of sodium hydride at 0 °C under nitrogen in DMF, and
the resulting alkoxide was allowed to react with the
27
2
Merrifield resins in DMF at 0 °C under N for 48 h.
The progress of the coupling reactions could be efficiently
13
28
monitored by C gel-phase NMR.
This simple tech-
nique, which can be run on a regular NMR spectrometer
2
9-31
and has been widely used in solid-phase synthesis,
affords a direct measure of the incorporation of ligand
into the polymer, thus allowing a fast optimization of
reaction conditions. We have represented in Figure 1 the
1
3
C gel-phase NMR spectrum of one of the starting
Merrifield resins (2% DVB, f ) 3.5) (a), along with the
spectrum recorded at room temperature (b) and at 50 °C
c) of the same resin at the end of the anchoring process.
o
(
1
3
For comparison purposes, the solution C NMR spectrum
of the closely related monomeric ether 6 has also been
included in this figure (d). The chloromethyl carbon
Resu lts a n d Discu ssion
Atta ch m en t of th e (2R,3R)-3-(Dia lk yla m in o)-2-
h ydr oxy-3-ph en ylpr opoxy Moiety to Mer r ifield Res-
in s. Our first choice for a polymer backbone was the
widely used Merrifield resin (chloromethylated styrene-
divinylbenzene copolymer), since the chloromethyl groups
contained in its structure provide a convenient means
by which to tether the ligand to the solid support. In
fact, Leznoff has reported the O-linkage of related
molecules to such polymers.23 Moreover, as far as the
choice of support is concerned, the Merrifield resin
seemed to be appropriate for our purposes, as it forms
(
24) Sarin, V. K.; Kent, S. B. H.; Merrifield, R. B. J . Am. Chem. Soc.
1
980, 102, 5463-5470.
(25) Merrifield, R. B. Peptide Synthesis, Structures and Applications;
Gutte, B., Ed.; Academic Press Inc.: San Diego, 1995; pp 93-169.
(26) Gao, Y.; Hanson, R. M.; Klunder, J . M.; Ko, S. Y.; Masamune,
H.; Sharpless, K. B. J . Am. Chem. Soc. 1987, 109, 5765-5780.
(27) These reaction conditions (HNa/DMF/0 °C) were analogous to
those employed for the model transformation in solution (the reaction
of 3 with benzyl chloride), which furnished the corresponding benzyl-
protected epoxy alcohol 6 in very high yield.
(28) The final resins gave poor quality gel phase 1H NMR spectra,
rendering this technique unsuccessful in monitoring reaction progress.
In addition, the final resins were also studied by magic angle spinning
(
MAS) NMR; in the resulting spectra, the chain signals were quite
gels which swell and become highly solvated in organic
broad and overlapped significantly with the resin backbone signals.
The results of this study will be reported in due course.
solvents of intermediate polarity2
4,25
such as dichlo-
(
29) Giralt, E.; Rizo, J .; Pedroso, E. Tetrahedron 1984, 40, 4141-
152.
(30) Look, G. C.; Holmes, C. P.; Chinn, J . P.; Gallop, M. A. J . Org.
4
(
22) Vidal-Ferran, A.; Moyano, A.; Peric a` s, M. A.; Riera, A. Tetra-
hedron Lett. 1997, 38, 8773-8776.
23) McArthur, C. R.; Worster, P. M.; J iang, J .-L.; Leznoff, C. C.
Can. J . Chem. 1982, 60, 1836-1841.
Chem. 1994, 59, 7588-7590.
(31) Murphy, M. M.; Schullek, J . R.; Gordon, E. M.; Gallop, M. A.
J . Am. Chem. Soc. 1995, 117, 7029-7030.
(

Chiral Amino Alcohol Ligands Anchored to Polystyrene Resins
J . Org. Chem., Vol. 63, No. 18, 1998 6311
with a family of six-membered ring cyclic amines: pi-
peridine, N-methylpiperazine, and cis-2,6-dimethylpi-
peridine, to afford the functionalized resins 5a -c. The
amines employed are those which provided the highest
2
1,22
catalytic activity in the original monomeric ligands.
Crotti’s method, which involves the use of a large excess
of nucleophile in a 5-10 M solution of lithium perchlorate
in acetonitrile (55 °C) and greatly favors attack of
nucleophile at the benzylic position, was used for the
reactions.3
3-38
Under these conditions, the ring-opening
of the monomeric benzylated epoxy alcohol 6 proceeds
with complete regio- and stereoespecificity, the only
isolated product being the one arising from attack at C-3
(7).
The ring-opening reaction of the epoxide-functionalized
resins with secondary amines was again monitored by
1
3
C NMR gel-phase spectroscopy. In this way, it could
be easily established that the reaction takes place by
regioselective attack at the benzylic position. Also in this
case, the 13C chemical shifts of the (2R,3R)-3-(dialkyl-
amino)-3-phenyl-2-hydroxypropoxy skeleton in the func-
tionalized polymers were closely similar to those observed
in solution for the analogous ligand 7 (Figure 2). In the
case of ligand 5a , for instance, line widths ranged from
1
4.5 to 206 Hz at room temperature. As line widths are
sensitive to the nucleus mobility, it is worth mentioning
that the closer the nucleus is to the anchoring point, the
broader the corresponding signal. Peaks around 73 ppm
F igu r e 1. 13C gel-phase NMR spectra of modified polystyrene-
2
%-co-divinylbenzene in CDCl
3
at 75.4 MHz (solvent peaks are
2
(corresponding to CH according to DEPT assignment in
marked with an asterisk): (a) Merrifield resin: 3.5 mmol of
chlorine per gram of resin, 1.5 Hz line broadening, room
temperature. (b) Epoxy alcohol bound polymer, 1.5 Hz line
broadening, room temperature. (c) Epoxy alcohol bound poly-
mer, 1.5 Hz line broadening, 50 °C. (d) Model compound (6) in
solution.
the model compound in solution) are the broadest (line
width of 206 Hz for two overlapped broad signals),
suggesting that they are due to carbons in the neighbor-
hood of the polymer backbone. As expected, peaks at 71.6
ppm (line width 34.1 Hz) and 68.2 ppm (line width 32.3
Hz), which correspond to methine carbons further away
from the support, are not as broad. Next, the signals at
51.7 ppm (line width 22 Hz), 26.1 ppm (line width 13.4
Hz) and 24.4 ppm (line width 14.5 Hz) corresponding to
the remote piperidine ring, are sharper than those arising
from the propoxymethyl chain, thus reflecting that these
carbons are further away from the polymer skeleton.
Finally, the sharpest signals are those due to the phenyl
groups, appearing together with the aromatic hump
which can be attributed to the polymer. As can be seen
from the examples described above, the relative distance
of a carbon from the support (an indication of its mobility)
signal in the starting resin appears in the 13C gel-phase
NMR spectrum at 46.3 ppm.29 Once the reaction is
complete (48 h), this signal has completely disappeared,
and four new broad resonances attributed to the 2,3-
epoxy-3-phenylpropoxymethyl aliphatic skeleton have
appeared at 73.1, 69.7, 60.9, and 55.8 ppm. It is worth
pointing out that these chemical shifts are fully coinci-
dent with the ones observed in solution for the model
compound 6. The important line broadening (line widths
at half-height ranged from 39 to 117 Hz) observed in the
gel-phase experiment at room temperature, which can
be attributed to chemical shift anisotropy and restricted
mobility of the chains,32 could be partially mitigated by
(33) Chini, M.; Crotti, P.; Macchia, F. J . Org. Chem. 1991, 56, 5939-
5
942.
recording the spectrum at 50 °C. Under these conditions,
slightly sharper signals (half-height line widths ranged
between 35 and 69 Hz) were observed.
The 2,3-epoxy-3-phenylpropoxymethyl functionalized
resins 4 were subsequently submitted to ring-opening
(
34) Chini, M.; Crotti, P.; Flippin, L. A.; Macchia, F. J . Org. Chem.
1991, 56, 7043-7048.
35) Chini, M.; Crotti, P.; Flippin, L. A.; Gardelli, C.; Giovani, E.;
Macchia, F.; Pineschi, M. J . Org. Chem. 1993, 58, 1221-1227.
36) Calvani, F.; Crotti, P.; Gardelli, C.; Pineschi, M. Tetrahedron
1994, 50, 12999-13022.
37) Colombini, M.; Crotti, P.; Di Bussolo, V.; Favero, L.; Gardelli,
(
(
(
C.; Macchia, F.; Pineschi, M. Tetrahedron 1995, 51, 8089-8112.
(38) Azzena, F.; Calvani, F.; Crotti, P.; Gardelli, C.; Macchia, F.;
Pineschi, M. Tetrahedron 1995, 51, 10601-10626.
(
32) Albericio, F.; Pons, M.; Pedroso, E.; Giralt, E. J . Org. Chem.
1
989, 54, 360-366.

6
312 J . Org. Chem., Vol. 63, No. 18, 1998
Vidal-Ferran et al.
Ta ble 1. Liga n d An ch or in d to Mer r ifield Resin s
functionalized resins
starting
Merrifield resin
5a
5b
5c
DVB (%)
foa
fb
fmaxc
fd
fmaxc
fb
fmaxc
1
2
2
1.2
2.3
3.5
0.8
1.6
1.8
1.0
1.6
2.1
0.9
1.6
1.7
1.0
1.5
2.0
1.0
1.6
1.7
1.0
1.5
2.0
a
fo ) mmol Cl/g resin (initial substitution level). b f ) mmol
ligand/g resin (calculated by elemental analysis of nitrogen with
the following formula: f ) 0.714% N). c fmax ) maximum ligand
substitution level (mmol ligand/g resin), calculated with the
following formula:
fo
fmax
)
MW - 35.45
1
+ fo
1
000
Where fo is the initial substitution level and MW is the molecular
weight of the anchored group. It is implicit in this formula that
both the anchoring and ring opening steps are taking place
quantitatively. d f ) mmol ligand/g resin (calculated by elemental
analysis of nitrogen with the following formula: f ) 0.357% N).
Ta ble 2. Ca ta lytic Beh a vior of th e Resin -An ch or ed
Liga n d s 5a -c in th e Ad d ition of Et2Zn to Ben za ld eh yd e
a
ligand
%DVB; f)
amount
(%)
ee of resulting
alcohol (%)
b
(
5
5
5
5
5
5
5
5
5
a (1; 0.8)
a (2; 1.6)
a (2; 1.8)
b (1; 0.9)
b (2; 1.6)
b (2; 1.7)
c (1; 1.0)
c (2; 1.6)
c (2; 1.7)
5
4
3
5
4
3
3
4
3
36
22
20
39
20
19
69
57
57
c
a
All reactions were performed in toluene at room temperature
for 24 h using a Et2Zn/PhCHO molar ratio of 2/1. Conversions were
b
higher than 95% in all cases (determined by GC). Amount of
catalyst (% molar with respect to benzaldehyde). c Mean of two
experiments.
F igu r e 2. 13C gel-phase NMR spectra of modified polystyrene-
%-co-divinylbenzene in CDCl at 75.4 MHz (solvent peaks are
2
3
Sch em e 2
marked with an asterisk): (a) Starting material in the regio-
selective ring-opening step, 1.5 Hz line broadening, room
temperature. (b) Amino alcohol bound polymer, 1.5 Hz line
broadening, room temperature. (c) Amino alcohol bound
polymer, 1.5 Hz line broadening, 50 °C. (d) Model compound
(7) in solution.
is a valuable parameter for peak assignment in 13C gel-
phase NMR spectra.
Determination of the level of functionalization of the
final resins 5a -c was relatively simple. As neither the
starting Merrifield resin nor the intermediate 2,3-epoxy-
Ca ta lytic P r op er ties of th e Am in o Alcoh ol Su b-
stitu ted Mer r ifield Resin s in th e En a n tioselective
Ad d ition of Dieth ylzin c to Ben za ld eh yd e. All the
modified resins 5a -c which, for a given amino alcohol
residue, include different levels of cross-linking and
ligand functionalization, were tested in the addition
reaction of diethylzinc to benzaldehyde (Scheme 2) to
determine how these factors affect the catalytic properties
of the resin. The results are summarized in Table 2.
3
-phenylpropoxymethyl-functionalized polymer contain
nitrogen atoms, an analytical determination for this
element should provide the level of functionalization in
the final resins. According to elemental analysis data,
the content of amino alcohol units in the final polymers
are very close (within experimental error) to the maxi-
mum value which can be calculated from the initial
ligand substitution level (cf. Table 1). This result indi-
cates that the ring-opening process of epoxides grafted
onto the polymer backbone is taking place with high
efficiency under Crotti’s conditions. This interesting
observation, that could be of great value for the prepara-
tion of other amino alcohol functionalized resins, is
2
The addition reaction of Et Zn to benzaldehyde was
studied at room temperature in toluene and using a
3-5% molar amount of ligand as indicated in Table 2.
All the substituted resins promoted the enantioselective
addition of diethylzinc to benzaldehyde, but with enan-
tioselectivities significantly lower than those observed
with the related homogeneous ligands.2
1,22
The predomi-
1
3
confirmed by C gel-phase NMR, since no residual
signals corresponding to 4 can be observed in the final
products 5.
nant enantiomer of 1-phenylpropanol (10a ) formed in the
reactions possessed the S configuration, in full agreement
3
9
with the empirical model established by Noyori. Con-

Chiral Amino Alcohol Ligands Anchored to Polystyrene Resins
J . Org. Chem., Vol. 63, No. 18, 1998 6313
Sch em e 3
with the optimal amino alcohol bound Merrifield resin
(69% ee with benzaldehyde using 5c, compared with 95%
ee for the same reaction in solution with ligand 2c),
suggested that the polystyrene skeleton was perhaps not
behaving as a sufficiently bulky substituent for the
primary hydroxy group in our amino diol ligands. This
suggestion was based on the consistent observation that
in solution, for ligands containing the same amino
residue, the induced enantioselectivity increases with the
bulk of the R-oxy group. In an attempt to circumvent
this difficulty, we considered the possibility of anchoring
our amino alcohol optimized ligand to commercial 2-chlo-
rotrityl chloride resin (Barlos resin).42 This polymer,
which is increasingly considered as a convenient support
F igu r e 3.
sistently with the results obtained in solution, for a given
level of cross-linking and similar functionalization, the
most efficient resin was the one bearing a cis-2,6-
dimethylpiperidino residue 5c. In the present case,
however, the influence of the dialkylamino moiety on the
catalytic activity is much higher than in solution, and
the resins containing other amino residues, 5a and 5b,
provided only mediocre results. An important observa-
tion arises when resins containing the same 3-(dialkyl-
amino)-2-hydroxy-3-phenylpropoxy residue, but variable
degrees of cross-linking and functionalization, are com-
pared. In all cases, the best result was obtained with the
polymeric ligand arising from chloromethylated polysty-
rene-1%-co-divinylbenzene with an initial substitution
level of 1.2 mmol of chlorine per gram of resin. For 5c
with these specifications, the addition of diethylzinc to
benzaldehyde took place with 69% enantioselectivity. An
explanation to this behavior could be as follows: As the
amino alcohol residues are increasingly diluted over the
polymer backbone, the formation of dimeric zinc alkox-
ides becomes less favorable (Figure 3). According to the
commonly accepted mechanism for the amino alcohol-
43
in solid-phase synthesis, possesses essentially the same
backbone as the Merrifield resin (polystyrene-1%-co-
divinylbenzene) except for a 2-chlorotrityl chloride anchor
instead of a chloromethyl one. If the 3-amino 1,2-diol
1
1 could be attached by the primary hydroxyl group to
the 2-chlorotrityl anchor, the resulting polymer supported
ligand would have a very similar steric and electronic
environment in the resin as it has in the structure 2c.
The anchoring reaction of 11 should proceed in a chemose-
lective way according to literature precedents, as second-
ary hydroxyl groups react very slowly with trityl ha-
44
lides.
Amino diol 11 was prepared by regio- and stereoselec-
tive ring-opening of enantiopure epoxy alcohol 3 by cis-
2
,6-dimethylpiperidine in the presence of titanium tet-
21,45,46
raisopropoxide (Scheme 3).
The reaction proceeded
smoothly at room temperature, amino diol 11 being
formed in 73% yield.
Two different reaction conditions were used to attach
1
1 to the 2-chlorotrityl chloride anchor in the Barlos
43
resin: pyridine in a 1:1 mixture of DMF/CH
2
Cl
In both cases,
an excess of nucleophile (the primary hydroxyl group in
1) and base are required; Table 3 shows the various
reaction conditions tested. The degree of anchoring of
1 in the resulting resins 12 was best established by
2
,
and
promoted catalytic addition of Et
2
Zn to aldehydes,40 these
42
2 2
diisopropylethylamine (DIEA) in CH Cl .
dimeric species are catalytically inactive, requiring dis-
sociation as a preliminary step to catalysis.41 As a
consequence, if a ligand for the enantioselective addition
of diethylzinc to aldehydes is to be anchored to a solid
support, it is advisable to use a low functionalized resin
in order to favor a maximal concentration of monomeric
zinc alkoxide species and, hence, a maximal catalytic
activity.
1
1
nitrogen elemental analysis. The highest degree of
anchoring (1.2 mmol of ligand per gram of resin, entry
III) was achieved with the reaction conditions which
involve the use of a large excess of nucleophile and a long
reaction time. However, the much milder conditions used
in entry II yielded a resin with almost the same degree
P r ep a r a tion a n d Ca ta lytic P r op er ties of Mod ified
Ba r los Resin s in th e En a n tioselective Ad d ition of
Dieth ylzin c to Ben za ld eh yd e. The results obtained
(
42) Barlos, K.; Gatos, D.; Kallitsis, J .; Papaphotiu, G.; Sotiriu, P.;
(
39) Noyori, R.; Kitamura, M. Angew. Chem., Int. Ed. Engl. 1991,
0, 49-69.
40) Noyori, R. Asymmetric Catalysis in Organic Synthesis; J ohn
Wiley & Sons: New York, 1994; Ch. 5, pp 255-297.
41) In situations where the dimeric zinc alkoxide predominates,
the background, noncatalyzed addition becomes important. In fact, we
have seen that the noncatalyzed addition of Et Zn (0.56 M) to
Wenging, Y.; Sch a¨ fer, W. Tetrahedron Lett. 1989, 30, 3943-3946.
(43) Wenschuh, H.; Beyermann, M.; Haber, H.; Seydel, J . K.; Krause,
E.; Bienert, M.; Carpino, L. A.; El-Faham, A.; Albericio, F. J . Org.
Chem. 1995, 60, 405-410.
(44) Kocienski, P. J . Protecting Groups; Georg Thieme Verlag:
Stuttgart, 1994; p 58.
(45) Caron, M.; Sharpless, K. B. J . Org. Chem. 1985, 50, 1557-1560.
(46) Canas, M.; Poch, M.; Verdaguer, X.; Moyano, A.; Peric a` s, M.
A.; Riera, A. Tetrahedron Lett. 1991, 32, 6931-6934.
3
(
(
2
benzaldehyde (0.22 M) in toluene/hexane (2: 2.5) takes place to a 23%
extent at room temperature after 24 h.

6
314 J . Org. Chem., Vol. 63, No. 18, 1998
Vidal-Ferran et al.
Ta ble 3. P r ep a r a tion of F u n ction a lized Tr ityla ted
Ta ble 4. Ca ta lytic P r op er ties of F u n ction a lized
Tr ityla ted Resin s in th e Am in o Alcoh ol P r om oted
a
Resin s Sta r tin g fr om Ba r los r esin
a
Ad d ition of Dieth ylzin c to Ben za ld eh yd e
molar ratio
-
active Cl
f (mmol)
f (mmol ligand/
g of resin)
% molar
ligand
temp,
°C
ee
(%)
b
configc
positions:amino reactn ligand/g
entry
entry
base/solvent
diol:base
time (h) resin)b
I
II
0.9
1.1
1.1
1.2
4
5
5
5
rt
rt
0
79
92
93
94
S
S
S
S
I
II
py/CH2Cl2:DMF (1:1)
DIEA/CH2Cl2
III DIEA/CH2Cl2
1:2.2:4.8
1:1.4:2.1
1:2:3
96
24
96
0.9
1.1
1.2
III
IV
0
a
a
Polystyrene-1%-co-divinylbenzene. Specification of the resin:
Resins prepared as described in Table 3. The polymer needs
to be stirred smoothly in the appropriate solvent to swell properly
prior to use. The resin was left to stir for 24 h in toluene before
adding the aldehyde and Et2Zn in molar ratio of 1:2. Stirring was
further continued for 24 h in order to achieve conversions higher
than 90% (determined by GC). b Enantiomeric excess in the resin
promoted addition of diethylzinc to benzaldehyde (determined by
GC: â-DEX 120 column). c Configuration of the addition product
b
1
.34 mmol labile Cl/g of resin. Determined by elemental analysis
of nitrogen.
(1-phenylpropanol).
Sch em e 4
Sch em e 5
coincident with those for the model compound in solution,
3 (Figure 4). Signals in the Barlos modified resins were,
1
however, much broader than those in the modified
Merrifield. To increase chain mobility and consequently
reduce line widths, the spectrum of this resin was
recorded in toluene-d at 80 °C.
8
Gratifyingly enough, when an amount of functionalized
Barlos resin equivalent to 5% mol of amino alcohol units
(
f ) 1.1 mmol ligand per gram of resin, entry II, Table 3)
was used to catalytically induce the addition of Et Zn to
benzaldehyde at room temperature, (S)-1-phenylpropanol
0a of 92% ee was rapidly formed (Scheme 5, R)Ph). The
2
1
recorded improvement in stereoselectivity was in full
agreement with our expectations, as the steric and
electronic environment of the anchored amino alcohol in
the Barlos resin is very close to that in 2c. A slight
improvement in enantioselectivity (up to 93% ee) could
be achieved by simply performing the reaction at 0 °C,
and a further improvement (up to 94% ee) was recorded
by using the more functionalized resin at 0 °C (entry IV,
Table 4). According to these results, ligand 12 is of
particular interest as it is placed at the top among the
polymer-supported ligands suitable for the enantioselec-
F igu r e 4. 13C gel-phase NMR spectra of modified Barlos resin
(
o-chlorotritylated polystyrene-1%-co-divinylbenzene) at 75.4
MHz (solvent peaks are marked with an asterisk): (a) Modified
Barlos resin 12: CDCl , 3 Hz line broadening, room temper-
ature. (b) Modified Barlos resin 12: toluene-d , 3 Hz line
broadening, room temperature. (c) Modified Barlos resin 12:
toluene-d , 3 Hz line broadening, 80 °C. (d) Model compound
13) in solution.
3
8
8
(
of anchoring. On the other hand, the use of pyridine as
the base was not as effective in catalyzing the coupling,
since even after 4 days the degree of functionalization
was lower than that obtained with H u¨ nig’s base under
the milder conditions in entry II.
tive addition of Et
2
Zn to benzaldehyde. In effect, polymer-
supported ligands have been reported to induce in this
6
-20
reaction enantioselectivities between 10 and 98%;
the
The final resins were characterized by 13C gel-phase
NMR spectroscopy. With this technique, the correct
number of signals for the aliphatic backbone of the amino
alcohol substituent were observed at chemical shifts
selectivity displayed by 12 is therefore one of the highest
reported in the literature. It is worth noting the small
difference in the enantioselectivities recorded with the
ligand in solution (97%) and with the immobilized one

Chiral Amino Alcohol Ligands Anchored to Polystyrene Resins
J . Org. Chem., Vol. 63, No. 18, 1998 6315
Ta ble 5. Ca ta lytic En a n tioselective Ad d ition of Et2Zn to Ald eh yd esa 8a -n Lea d in g to Alcoh ols 10a -n b Med ia ted by
P olym er -Su p p or ted Ca ta lyst 12
enant excess (%)e
resulting
alcohol
conversionc
(%)
selectivityd
(%)
starting aldehyde
Benzaldehyde (8a )
o-Fluorobenzaldehyde (8b)
m-Fluorobenzaldehyde (8c)
p-Fluorobenzaldehyde (8d )
o-Tolualdehyde (8e)
m-Tolualdehyde (8f)
p-Tolualdehyde (8g)
o-Methoxybenzaldehyde (8h )
m-Methoxybenzaldehyde (8i)
p-Methoxybenzaldehyde (8j)
with 12
with 2c
(S)-10a
(S)-10b
(S)-10c
(S)-10d
(S)-10e
(S)-10f
(S)-10g
(S)-10h
(S)-10i
(S)-10j
(S)-10k
(S)-10l
(S)-10m
(S)-10n
99
96
98
>99
91
93
93
88
>99
86
86
82
86
99
98
96
96
99
84
94
94
92
98
95
84
75
90
94
94
88
94
95
91
94
94
86
94
94
86
90
98
90
97
97
97
98
90
1
2
3
-Naphthaldehyde (8k )
-Naphthaldehyde (8l)
-Cyclohexenecarbaldehyde (8m )
Isovaleraldehyde (8n )
90
a
The resin (1.2 mmol ligand/g resin) was left to stir 24 h in toluene to swell properly. A molar ratio of aldehyde, Et2Zn and anchored
ligand of 1:1.5:0.08 was used for 8b-n . See text for 8a . The mixture was stirred for 24 h at 0 °C after which it was quenched and analyzed.
b
The S enantiomer is preferentially obtained in all the cases. The absolute configuration has been established by comparing the sign of
c
the optical rotation except for 10c, 10i, and 10m in which the S configuration has been assumed. Determined by integration of residual
d
starting material in front of all new products in the gas chromatogram of the reaction crude. Determined by integration of the addition
products (both enantiomers) in front of other reaction products. e By GC using either a â-DEX 120 column (for 10a -m ) or a R-DEX 120
(for 10n ).
(
94%), which suggests that although the ligand is an-
ee. Even more importantly, the mean difference in
enantiomeric excess of the alcohols obtained with 12 and
with its homogeneous counterpart 2c for a common set
of aldehydes is only of 2.6% ee. With respect to individual
compounds, very high enantioselectivities were obtained
with para-substituted benzaldehydes (94-95% ee), ir-
respective of the electronic nature of the substituent. In
any case, and not unexpectedly, the most electron-
deficient substrate is also the fastest-reacting one. A
completely parallel behavior in reactivity can be found
in the ortho-substituted benzaldehydes, although the
enantioselectivity is slightly lower in this case (86-91%
ee). For the less studied meta-substituted analogues, in
turn, a uniformly high (94% ee) enantioselectivity was
observed. For the bulky 1- and 2-naphthaldehydes, a
slightly lower enantioselectivity (86-90%) was recorded,
as well as for isovaleraldehyde (8n ). The other aliphatic
aldehyde tested (8m ), however, provided the highest
enantioselectivity in the series investigated.
chored on an insoluble support, the chiral ligand chains
possess the required mobility to provide the catalytic sites
where the enantioselective reaction takes place. This is
remarkable considering that most ligands suffer a sig-
nificant drop in enantioselectivity when attached to a
polymer support. Although the differences in function-
alization between the prepared resins are small, the
slightly better result (Table 4, entry IV) recorded with
the most functionalized resin tends to indicate a different
scenario with respect to the situation in Merrifield type
resins. Most probably, the simple presence of the bulky
trityl substituent is a sufficient condition to prevent the
formation of dimeric species with ligands 2c and 12.47
From a practical point of view, it is worth mentioning
that the resin could be simply recovered by quenching
4
the reaction with satd NH Cl followed by filtration and
reconditioning by washing with 5% aq HCl (to eliminate
inorganic residues) and, immediately after the acidic
treatment, washing with pH 9 phosphate buffer, water,
methanol, and lastly toluene. After careful drying, it was
reused, yielding the final alcohol 10a with no decrease
in enantioselectivity.
Con clu sion s
In summary, we have succeeded in attaching (2S,3S)-
This highly active new functionalized resin was also
2
,3-epoxy-3-phenylpropanol to Merrifield resins. This
2
studied in the addition of Et Zn to a set of 13 representa-
chiral vector has allowed the anchoring of enantiomeri-
cally pure amino alcohols through the regio- and ste-
reospecific oxirane ring-opening with secondary amines.
tive aldehydes 8b-n including aliphatic ones (cyclic and
acyclic), three complete sets of ortho-, meta-, and para-
substituted aromatic aldehydes (including both electron-
donating and electron-withdrawing substituents), and the
naphthaldehydes (1- and 2-substituted). The results of
these reactions have been summarized in Table 5, where
the results obtained with the closely related homogeneous
ligand 2c have also been included for comparison pur-
poses.
A first point that deserves a comment is the uniformly
good performance of the anchored ligand 12 over the
range of studied aldehydes. Thus, for the 14 studied
aldehydes, the mean enantiomeric excess of the resulting
alcohols is 92%, with a standard deviation of only 3.6%
13
From the analytical point of view, C gel-phase NMR
spectroscopy has proved to be very useful in the charac-
terization of intermediates and final products in the solid
phase synthesis. Given the ready availability of this
analytical tool and the wide range of nucleophiles capable
of reacting with oxiranes, the methodology reported in
this paper could be a convenient vehicle for the controlled
integration on such resins of many different nucleophile
types through an enantiomerically pure handle.
The amino alcohol functionalized Merrifield resins
have shown to possess catalytic activity in the enantio-
selective addition of diethylzinc to benzaldehyde. As
observed for related ligands in solution, the cis-2,6-
dimethylpiperidino fragment imparts to the functional-
ized resins the optimal catalytic properties and, for all
(47) NMR studies on the relative stability of dimeric zinc alkoxides
provide support to this assumption. Vidal-Ferran, A.; Moyano, A.;
Peric a` s, M. A.; Riera, A. Unpublished results.

6
316 J . Org. Chem., Vol. 63, No. 18, 1998
Vidal-Ferran et al.
the prepared ligands, the lower the functionalization/
reticulation of the ligands, the higher the induced enan-
tioselectivity. This represents a nice confirmation on the
monomeric nature of the catalytically active species in
this reaction.
In an attempt to further improve the catalytic perfor-
mance of the resin-immobilized ligands, we have an-
chored (2R,3R)-3-(cis-2,6-dimethylpiperidino)-3-phenyl-
MHz, CDCl
3
, 1.5 Hz line broadening) δ 73.1 (70), 69.6 (54),
4
9
6
1.0 (37), 55.7 (45).
Resin 4 (2% DVB, fmax ) 2.5)48 fr om Ch lor om eth yla ted
P olystyr en e (2% DVB, f ) 3.5). Compound 3 (1.45 g, 9.65
o
mmol) in DMF (15 mL), sodium hydride (385 mg, ca. 12.83
mmol) in DMF (8 mL), and polymer support (2.5 g, 8.75 mmol)
in DMF (15 mL) were treated as described in the general
procedure and stirred 48 h at 0 °C. The functionalized resin
4 (2% DVB, fmax ) 2.5) (3.21 g; 92% yield) was used in the
1
3
next step without further purification. C gel-phase NMR (75
MHz, CDCl , 1.5 Hz line broadening) δ 73.0 (117), 69.5 (66),
1.0 (39), 55.6 (51).
Lith iu m P er ch lor a te In d u ced Regioselective Rin g-
1
,2-propanediol to a 2-chlorotrityl chloride (Barlos) resin.
The catalytic behavior exhibited by the resulting ligand
2 in the enantioselective addition of diethylzinc to a
3
4
9
6
1
broad and representative family of aldehydes converts
it in the most generally useful heterogeneous ligand for
this chemistry ever reported. In light of the exceptional
catalytic profile of 12, we are currently applying this
polymeric catalyst to other enantioselective processes
catalyzed by amino alcohols.
Op en in g of P olym er Bou n d Ep oxy Eth er s by Secon d a r y
Am in es. Resin 5a (1% DVB; f ) 0.8) fr om P ip er id in e a n d
4 (1% DVB, fm a x ) 1.1). Gen er a l P r oced u r e. Piperidine
(1.83 mL, 18.6 mmol) was added via syringe into a mixture of
polymer bound epoxy ether 4 (620 mg, 0.67 mmol assuming
1
00% anchoring) and LiClO
3 mL) at 55 °C under N . After smooth magnetic stirring for
8 h at 55 °C, the functionalized resin was filtered, washed
4
(1.98 g, 18.6 mmol) in acetonitrile
(
2
4
Exp er im en ta l Section
with dimethylformamide (2 × 10 mL), dimethylformamide:
water 1:1 (4 × 10 mL), water (4 × 10 mL), methanol (4 × 10
mL), and toluene (4 × 10 mL), and dried under vacuum to
constant weight to afford 630 mg (93% yield) of modified resin
5a (1% DVB; f ) 0.8): 13C gel-phase NMR (75 MHz, CDCl3,
1.5 Hz line broadening) δ 73.2 (72), 72.7 (86), 71.7 (14), 68.4
(16), 51.8 (11), 26.3 (7), 24.5 (7 Hz).49 Anal. Calcd for fmax: N,
1.40. Found: N, 1.10.
Gen er a l. 1H NMR and 13C NMR spectra in solution were
recorded in CDCl at 300 and 75.4 MHz, respectively. The
3
NMR gel samples were prepared as follows: the appropriate
mass of resin was placed in a 5 mm NMR tube, and the same
volume of solvent was added. When the solvent had been
absorbed, small additional fractions of solvent were added to
obtain a homogeneous gel. The so-prepared samples were
allowed to stand for 8-12 h before recording the spectra.
gel phase NMR spectra were recorded at 75.4 MHz in CDCl
or toluene-d . Standard conditions were used: 8.7 µs pulse
Resin 5b (1% DVB; f ) 0.9) fr om N-m eth ylp ip er a zin e
1
3
C
3
a n d 4 (1% DVB, f
24.8 mmol), polymer bound epoxy ether 4 (650 mg, 0.70 mmol
) 1.1): N-Methylpiperazine (2.75 mL,
m a x
8
assuming 100% anchoring), LiClO (2.64 g, 24.8 mmol), and
4
width (the 90° pulse width was 10 µs), acquisition time 1.815
s, 0.2 s delay between pulses, and 1.5 or 3 Hz line broadening).
Chemical shifts are quoted relative to solvent signals. Line
widths at half-height have been determined using the standard
spectrometer software and are given in parentheses. Elemen-
tal analyses were carried out by the “Servei d’An a` lisis El-
ementals del C.S.I.C. de Barcelona”. Tungsten(VI) oxide was
used in the resin analyses to ensure total combustion of the
samples. (2S,3S)-2,3-Epoxy-3-phenylpropanol, 3, was pre-
pared according to the procedure described by Sharpless et
acetonitrile (4 mL) were treated as described in the general
procedure to afford 705 mg (98% yield) of modified resin 5b
(1% DVB; f ) 0.9): 13C gel-phase NMR (75 MHz, CDCl , 1.5
3
Hz line broadening) δ 73.2 (87), 72.5 (42), 71.1 (13), 68.2 (11),
55.3 (8), 50.6 (16), 45.8 (13 Hz).49 Anal. Calcd for fmax: N,
2.80. Found: N, 2.42.
Resin 5c (1% DVB; f ) 1.0) fr om cis-2,6-d im eth ylp ip -
er id in e a n d 4 (1% DVB, f
eridine (2.49 mL, 18.6 mmol), polymer bound epoxy ether 4
m a x
) 1.1): cis-2,6-Dimethylpip-
(620 mg, 0.67 mmol assuming 100% anchoring), LiClO (1.98
4
g, 18.6 mmol), and acetonitrile (3 mL) were treated as
described in the general procedure to afford 660 mg (95% yield)
2
6
al. DMF and CH
2 2 2
Cl were distilled from CaH and stored
under N . Merrifield resins were commercially available.
2
1
3
An ch or in g (2S,3S)-2,3-Epoxy-3-ph en ylpr opan ol to Mer -
r ifield R esin s. R esin 4 (1% DVB, fm a x ) 1.1)48 fr om
of modified resin 5c (1% DVB; f ) 1.0): C gel-phase NMR
(75 MHz, CDCl
64.7 (11), 48.7 (27), 47.7 (28), 31.8 (12), 17.3 (14), 15.0 (10 Hz).
3
, 1.5 Hz line broadening) δ 73.1 (55), 68.1 (14),
49
Ch lor om eth yla ted P olystyr en e (1% DVB, f
o
) 1.23).
Gen er a l P r oced u r e. A solution of 3 (660 mg, 4.39 mmol) in
DMF (15 mL) was added via cannula to a suspension of sodium
hydride (175 mg, ca. 5.83 mmol) in DMF (20 mL) at 0 °C under
Anal. Calcd for fmax: N, 1.40. Found: N, 1.42.
Resin 5a (2% DVB; f ) 1.6) fr om p ip er id in e a n d 4 (2%
DVB, fm a x ) 1.8): Piperidine (2.44 mL, 24.8 mmol), polymer
bound epoxy ether 4 (650 mg, 1.18 mmol assuming 100%
anchoring), LiClO (2.64 g, 24.8 mmol), and acetonitrile (4 mL)
4
were treated as described in the general procedure to afford
N
2
. The mixture was stirred for 20 min, quickly poured onto
a suspension of the Merrifield resin (4.0 g, 4.9 mmol) in DMF
20 mL) at 0 °C, flushed with N , and smoothly stirred for 48
(
2
h at 0 °C. The functionalized resin was filtered, washed with
738 mg (98% yield) of modified resin 5a (2% DVB; f ) 1.6):
1
3
dimethylformamide (4 × 10 mL) and dichloromethane (4 ×
3
C gel-phase NMR (75 MHz, CDCl , 1.5 Hz line broadening)
1
4
0 mL), and dried under vacuum to constant weight to afford
.05 g (89% yield) of modified resin 4 (1% DVB, fmax ) 1.1),
δ 72.7 (186), 71.6 (26), 68.2 (26), 51.6 (18), 26.1 (11), 24.4 (11
Hz).49 Anal. Calcd for fmax: N, 2.24. Found: N, 2.31.
Resin 5b (2% DVB; f ) 1.6) fr om N-m eth ylp ip er a zin e
a n d 4 (2% DVB, fm a x ) 1.8): N-Methylpiperazine (2.75 mL,
24.8 mmol), polymer bound epoxy ether 4 (650 mg, 1.18 mmol
which was used in the next step without further purification.
1
3
C gel-phase NMR (75 MHz, CDCl
3
, 1.5 Hz line broadening)
4
9
δ 73.2 (22), 69.6 (28), 61.2 (12) and 55.8 (16).
Resin 4 (2% DVB, fm a x ) 1.8) fr om Ch lor om eth yla ted
P olystyr en e (2% DVB, f ) 2.3). Compound 3 (1.52 g, 10.12
4
8
assuming 100% anchoring), LiClO (2.64 g, 24.8 mmol), and
4
acetonitrile (4 mL) were treated as described in the general
procedure to afford 760 mg (99% yield) of modified resin 5b
o
mmol) in DMF (15 mL), sodium hydride (405 mg, ca. 13.5
mmol) in DMF (10 mL), and polymer support (4.0 g, 9.2 mmol)
in DMF (20 mL) were treated as described in the general
procedure and stirred 24 h at 0 °C. The functionalized resin
1
3
(
2% DVB; f ) 1.6): C gel-phase NMR (75 MHz, CDCl
Hz line broadening) δ 72.8 (107), 72.5 (101), 71.0 (22), 68.0
24), 55.0 (15), 50.3 (27), 45.6 (29 Hz).49 Anal. Anal. Calcd
3
, 1.5
(
for fmax: N, 4.20. Found: N, 4.50.
4
(2% DVB, fmax ) 1.8) (4.40 g; 87% yield) was used in the
13
Resin 5c (2% DVB; f ) 1.6) fr om cis-2,6-d im eth ylp ip -
er id in e a n d 4 (2% DVB, fm a x ) 1.8): cis-2,6-Dimethylpip-
eridine (3.3 mL, 24.8 mmol), polymer bound epoxy ether 4 (650
next step without further purification. C gel-phase NMR (75
(48) Maximum functionalization of the resin (mmol 2,3-epoxy-3-
4
mg, 1.18 mmol assuming 100% anchoring), LiClO (2.64 g, 24.8
phenylpropoxy groups/g resin) calculated with the formula shown in
Table 1.
mmol), and acetonitrile (4 mL) were treated as described in
the general procedure to afford 750 mg (96% yield) of modified
(49) Phenyl group signals are not reported as they significantly
1
3
overlap with the polymer peaks.
resin 5c (2% DVB; f ) 1.6): C gel-phase NMR (75 MHz,

Chiral Amino Alcohol Ligands Anchored to Polystyrene Resins
J . Org. Chem., Vol. 63, No. 18, 1998 6317
CDCl
(
3
, 1.5 Hz line broadening) δ 72.8 (114), 68.0 (29), 64.6
This anchoring step was also studied under more severe
conditions: DIEA (775 µL, 4.44 mmol), 11 (780 mg, 3.0 mmol),
2 2
and resin (1.1 g, 1.47 mmol of active Cl) in CH Cl (10 mL)
21), 48.4 (135), 47.7 (115), 31.7 (18), 17.2 (21), 14.9 (20 Hz).49
Anal. Calcd for fmax: N, 2.10. Found: N, 2.18.
Resin 5a (2% DVB; f ) 1.8) fr om p ip er id in e a n d 4 (2%
DVB, fm a x ) 2.5): Piperidine (3.1 mL, 31.0 mmol), polymer
bound epoxy ether 4 (650 mg, 1.62 mmol assuming 100%
were treated as described above during 48 h. A second fraction
of DIEA was added (775 µL, 4.44 mmol) and smooth magnetic
stirring continued for 48 h. The resin was purified as
described above to afford 1.43 g (99% yield) of 12 (1%, f ) 1.2).
anchoring), LiClO
were treated as described in the general procedure to afford
14 mg (91% yield) of modified resin 5a (2% DVB; f ) 1.8).
4
(3.30 g, 31.0 mmol), and acetonitrile (5 mL)
1
3
C gel-phase NMR set of data was fully coincident with the
7
one described above. Anal. Calcd for fmax: N, 1.40. Found:
N, 1.64.
1
3
C gel-phase NMR (75 MHz, CDCl
3
, 1.5 Hz line broadening)
δ 72.7 (206), 71.6 (34), 68.2 (32), 51.7 (22), 26.1 (13), 24.4 (14
With p yr id in e a s th e ba se: Pyridine (585 µL, 7.23 mmol)
was added via syringe into a mixture of 11 (865 mg, 3.28 mmol)
and the resin (913 mg, 1.22 mmol of active Cl) in CH Cl (5.5
4
9
Hz). Anal. Calcd for fmax: N, 2.94. Found: N, 2.53.
Resin 5b (2% DVB; f ) 1.7) fr om N-m eth ylp ip er a zin e
a n d 4 (2% DVB, fm a x ) 2.5): N-Methylpiperazine (3.4 mL,
2
2
mL) and DMF (5.5 mL) under N at room temperature. After
2
2
4.8 mmol), polymer bound epoxy ether 4 (650 mg, 1.62 mmol
smooth magnetic stirring for 72 h at room temperature, the
functionalized resin was filtered, washed with dimethylform-
amide (2 × 10 mL), dimethylformamide:water 1:1 (4 × 10 mL),
water (4 × 10 mL), pH 9 phosphate buffer (4 × 10 mL), water
4
assuming 100% anchoring), LiClO (3.30 g, 31.0 mmol), and
acetonitrile (5 mL) were treated as described in the general
procedure to afford 741 mg (91% yield) of modified resin 5b
1
3
(
3
2% DVB; f ) 1.7). C gel-phase NMR (75 MHz, CDCl , 1.5
(
8 × 10 mL), methanol (4 × 10 mL), and toluene (4 × 10 mL),
4
9
Hz line broadening) δ 72.6, 71.3, 68.1, 55.3, 50.6, 45.9, 21.4.
Anal. Calcd for fmax: N, 5.60. Found: N, 4.74.
and dried under vacuum to constant weight to afford 977 mg
(
was fully coincident with the one described above. Anal.
13
87% yield) of 12 (1%, f ) 0.9). C gel-phase NMR set of data
Resin 5c (2% DVB; f ) 1.7) fr om cis-2,6-d im eth ylp ip -
er id in e a n d 4 (2% DVB, fm a x ) 2.5): cis-2,6-Dimethylpip-
eridine (4.1 mL, 18.6 mmol), polymer bound epoxy ether 4 (650
mg, 1.62 mmol assuming 100% anchoring), LiClO (3.30 g, 31.0
4
mmol), and acetonitrile (5 mL) were treated as described in
the general procedure to afford 738 mg (88% yield) of modified
Calcd for fmax: N, 1.40. Found: N, 1.30.
Gen er a l P r oced u r e for th e En a n tioselective Am in o
Alcoh ol-Ca ta lyzed Ad d ition of Dieth ylzin c to Ald eh yd es.
A suspension of the appropriate polymer bound catalyst (3-
8
% molar as indicated in each case in Tables 2, 4, or 5) in
1
3
resin 5c (2% DVB; f ) 1.7).
CDCl , 1.5 Hz line broadening) δ 73.2, 68.3, 64.9, 48.8, 48.0,
2.0, 21.5, 17.6, 15.3.49 Anal. Calcd for fmax: N, 2.66. Found:
N, 2.36.
P r ep a r a tion of (2R,3R)-3-(cis-2,6-Dim eth ylp ip er id in o)-
-p h en yl-1,2-p r op a n ed iol. cis-2,6-Dimethylpiperidine (2.7
C gel-phase NMR (75 MHz,
toluene (2 mL) was allowed to stir smoothly under N at room
2
3
temperature in order to swell the polymer, after which (30 min
for a modified Merrifield resin and 24 h for a Barlos one) the
aldehyde 8 (1 mmol) was added at room temperature. The
mixture was stirred for 20 min and then cooled to the desired
temperature if necessary. Diethylzinc (amount indicated in
each case in Tables 2, 4, or 5) was added dropwise. The
3
3
i
mL, 20.0 mmol) and Ti(O Pr)
to a solution of 3 (2.0 g, 13.3 mmol) in CH
at room temperature. After 14 h of stirring at room
4
(5.9 mL, 20.0 mmol) were added
2
2
Cl (50 mL) under
mixture was stirred 24 h under N
was quenched by the addition of a saturated NH
10 mL). The resin was removed by filtration, the aqueous
solution was then extracted with CH Cl
2
, after which the reaction
N
2
4
Cl solution
temperature, a 10% solution of NaOH in brine (25 mL) was
added and vigorous stirring continued for a further 24 h. The
mixture was filtered through Celite and the residue washed
(
2
2
(3 × 10 mL), and the
combined organic extracts were dried and concentrated in
vacuo. The enantiomeric excesses were determined from the
crude mixture by GC analyses. Conditions of GC analyses:
â-DEX or R-DEX 120 column, 30 m length, 0.25 mm inner
diameter, isotherm temperature program, He as carrier gas
with CH
with CH
2
Cl
Cl
2
(3 × 10 mL). The aqueous solution was extracted
2
2
(3 × 25 mL), and the combined organic extracts
dried and concentrated in vacuo. The residual oil was recrys-
2
tallized from hexane:Et O 2:1 to give 2.45 g (73% yield) of 11:
2
3
1
mp 88 °C: [R]
D
) -16.6 (c ) 0.9 in CHCl
3
); H NMR (300
(
2.4 mL/min). For 1-phenylpropanol: â-DEX 120 column, 112
C, t R isomer 49.3 min, t S isomer 52.0 min. For 1-(o-tolyl)-
propanol: â-DEX 120 column, 120 °C, t R isomer 59.1 min,
S isomer 63.8 min. For 1-(m-tolyl)propanol: â-DEX 120
column, 120 °C, t R isomer 52.4 min, t S isomer 53.8 min.
For 1-(p-tolyl)propanol: â-DEX 120 column, 120 °C, t
isomer 50.4 min, t S isomer 53.3 min. For 1-(2-methoxyphen-
yl)propanol: â-DEX 120 column, 135 °C, t S isomer 48.1 min,
R isomer 54.1 min. For 1-(3-methoxyphenyl)propanol:
R isomer 66.2 min, t S isomer
MHz, CDCl ) δ 7.30-7.36 (m, 5H), 4.36 (ddd, 1H, J ) 7.5, 6.0,
3
°
R
R
and 6.0 Hz), 4.17 (br s, 1H, OH), 4.11 (d, 1H, J ) 7.5 Hz),
R
3
1
.46-3.58 (m, 2H), 2.78-2.98 (m, 2H), 2.61 (br s, 1H, OH),
.23-1.69 (m, 6H), 1.16 (d, 3H, J ) 6.6 Hz) and 1.15 (d, 3H,
t
R
1
3
R
R
J ) 6.6 Hz); C NMR (75 MHz, CDCl
3
) δ 136.8 (C), 129.4 (CH),
), 66.78 (CH),
), 20.2 (CH ), 19.8
); IR (KBr) 3401, 3083, 3064, 3021, 2985,
964, 2933, 2867, 2848, 2821, 2796, 1148, 1115, 1079, 1028,
R
R
1
5
28.1 (CH), 127.5 (CH), 68.6 (CH), 66.85 (CH
1.7 (CH), 49.8 (CH), 33.1 (CH ), 32.9 (CH
) and 18.3 (CH
2
R
2
2
3
R
(CH
3
2
t
R
2
7
1
-
1
+
+
2 5 2
H O ,
â-DEX 120 column, 135 °C, t
R
R
49, 706 cm ; MS (EI) m/z 263 (M , 0%), 202 (M - C
00%). Anal. Calcd for C16 : C, 72.97; H, 9.57; N, 5.32.
6
8.0 min. For 1-(4-methoxyphenyl)propanol: â-DEX 120
H
25NO
2
R R
column, 135 °C, t R isomer 68.2 min, t S isomer 70.4 min.
Found: C, 73.06; H, 9.53; N, 5.21.
For 1-(2-fluorophenyl)propanol: â-DEX 120 column, 112 °C,
R isomer 51.0 min, t S isomer 54.5 min. For 1-(3-
fluorophenyl)propanol: â-DEX 120 column, 112 °C, t R isomer
0.0 min, t S isomer 63.2 min. For 1-(4-fluorophenyl)-
propanol: â-DEX 120 column, 112 °C, t R isomer 57.3 min,
S isomer 61.2 min. For 1-(1-naphthyl)propanol: â-DEX 120
column, 160 °C, t S isomer 98.9 min, t R isomer 103.2 min.
For 1-(2-naphthyl)propanol: â-DEX 120 column, 160 °C, t
isomer 100.2 min, t S isomer 102.4 min. For 5-methyl-3-
hexanol: R-DEX 120 column, 65 °C, t R isomer 15.1 min, t
S isomer 15.5 min. For 1-(3-cyclohexenyl)propanol: â-DEX
An ch or in g of (2R,3R)-3-(cis-2,6-d im eth ylp ip er id in o)-
-p h en yl-1,2-p r op a n ed iol to a Ba r los Resin (in itia l su b-
stitu tion level: 1.34 m m ol Cl/g).
With d iisop r op yleth yla m in e (DIEA) a s th e ba se: DIEA
t
R
R
3
R
6
R
R
(
1
345 µL, 1.98 mmol) was added via syringe into a mixture of
1 (350 mg, 1.33 mmol) and the resin (705 mg, 0.94 mmol of
active Cl) in CH Cl (7 mL) under N at room temperature.
t
R
R
R
2
2
2
R
R
After smooth magnetic stirring for 24 h at room temperature,
the functionalized resin was filtered, washed with dimethyl-
formamide (2 × 10 mL), dimethylformamide:water 1:1 (4 ×
R
R
R
1
0 mL), water (4 × 10 mL), pH 9 phosphate buffer (4 × 10
1
7
20 column, 100 °C, t
4.2-74.4 min.
R
R
S isomer 69.8-70.6 min, t R isomer
mL), water (8 × 10 mL), methanol (4 × 10 mL), and toluene
(
4 × 10 mL) and dried under vacuum to constant weight to
1
3
afford 905 mg (99% yield) of 12 (1%, f ) 1.1). C gel-phase
NMR (75 MHz, toluene-d , 3 Hz line broadening, 80 °C) δ 87.4
38), 70.1 (50), 68.3 (66), 66.8 (47), 50.4 (43), 49.3 (42), 32.5
To establish the absolute configuration of the final com-
pounds, the alcohols were purified by bulb-to-bulb distillation
of the crude mixtures. The optical rotation was measured in
8
(
(
4
9
52), 19.2 (200), 16.5 (31).
Anal. Calcd for fmax: N, 1.40.
each case, and its sign was compared with the reported value
51
Found: N, 1.54.
[(S)-1-phenylpropanol,50 (S)-1-(o-tolyl)propanol,
(S)-1-(m-

6
318 J . Org. Chem., Vol. 63, No. 18, 1998
Vidal-Ferran et al.
tolyl)propanol,52 (S)-1-(p-tolyl)propanol, (R)-1-(2-methoxy-
53
(GRQ93-1083), MEC and the British Council in Madrid
5
4
53
phenyl)propanol, (S)-1-(4-methoxyphenyl)propanol, (R)-1-
2-fluorophenyl)propanol, (R)-1-(4-fluorophenyl)propanol,
R)-1-(1-naphthyl)propanol, (R)-1-(2-naphthyl)propanol, and
(HB1995-30) is gratefully acknowledged. Anton Vidal-
5
5
56
(
(
(
Ferran thanks MEC for a contract under the program:
Acciones para la Incorporaci o´ n a Espa n˜ a de Doctores y
Tecn o´ logos. We further thank F. Albericio, D. Andreu,
M. J . Duer, P. Lloyd-Williams and B. Ponsati for helpful
advice and Lipotec S.A. for a generous gift of the Barlos
Resin.
5
7
58
5
9
S)-5-methyl-3-hexanol ].
Ack n ow led gm en t. Financial support from CIRIT-
CICYT (QFN93-4407), DGICYT (PB95-0265), CIRIT
(
50) Kitamura, M.; Suga, S.; Kawai, K.; Noyori, R. J . Am. Chem.
Soc. 1986, 108, 6071-6072.
51) Chelucci, G.; Conti, S.; Falorni, M.; Giacomelli, G. Tetrahedron
991, 47, 8251-8257.
52) Chaloner, P. A.; Perera, S. A. R. Tetrahedron Lett. 1987, 28,
J O980708K
(
1
(56) Kang, J .; Lee, J . W.; Kim, J . I. J . Chem. Soc., Chem. Commun.
1994, 2009-2010.
(
3
013-3014.
(57) Niwa, S.; Soai, K. J . Chem. Soc., Perkin Trans. 1 1991, 2717-
(
(
53) Capillon, J .; Gu e´ tt e´ , J . P. Tetrahedron 1979, 35, 1817-1820.
54) Smaardijk, A. A.; Wynberg, H. J . Org. Chem. 1987, 52, 135-
2720.
(58) Oguni, N.; Omi, T.; Yamamoto, Y.; Nakamura, A. Chem. Lett.
1983, 841-842.
(59) Soai, K.; Ookawa, A.; Kaba, T.; Ogawa, K. J . Am. Chem. Soc.
1987, 109, 7111-7115.
1
37.
(55) Seebach, D.; Beck, A. K.; Schmidt, B.; Wang, Y. M. Tetrahedron
1
994, 50, 4363-4384.