N/N/O versus N/O/O and N/O amino isoborneols in the enantioselective ethylation of benzaldehyde

Article abstract of DOI:10.1016/j.tetasy.2008.01.012

Enantiopure 10-(4-methylpiperazin-1-yl)isoborneol, a N/N/O-tridentate chiral diamino isoborneol, has been obtained from enantiopure ketopinic acid and its catalytic activity (as chiral ligand in the enantioselective addition of diethylzinc to benzaldehyde) measured and compared with the previously reported data for related N/N/O, N/O/O, and N/O isoborneols. It is demonstrated that the pentacoordinated-zinc Oppolzer's model for transition states is useful for explaining the catalytic behavior of N/N/O-tridentate ligands. In such cases, stable syn- or anti-type transition states can be proposed depending on the catalyst's conformational flexibility.

Full text of DOI:10.1016/j.tetasy.2008.01.012

Available online at www.sciencedirect.com
Tetrahedron: Asymmetry 19 (2008) 269–272
N/N/O versus N/O/O and N/O amino isoborneols
in the enantioselective ethylation of benzaldehyde
´
´
´
Tomas de las Casas Engel, Beatriz Lora Maroto, Antonio Garcıa Martınez and
Santiago de la Moya Cerero*
Departamento de Quı´mica Orga´nica I de la Universidad Complutense de Madrid, Facultad de Ciencias Quı´micas,
Ciudad Universitaria s/n, 28040 Madrid, Spain
Received 20 December 2007; accepted 8 January 2008
Abstract—Enantiopure 10-(4-methylpiperazin-1-yl)isoborneol, a N/N/O-tridentate chiral diamino isoborneol, has been obtained from
enantiopure ketopinic acid and its catalytic activity (as chiral ligand in the enantioselective addition of diethylzinc to benzaldehyde) mea-
sured and compared with the previously reported data for related N/N/O, N/O/O, and N/O isoborneols. It is demonstrated that the
pentacoordinated-zinc Oppolzer’s model for transition states is useful for explaining the catalytic behavior of N/N/O-tridentate ligands.
In such cases, stable syn- or anti-type transition states can be proposed depending on the catalyst’s conformational flexibility.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
other chiral N/O-bidentate ligands, specially b-amino alco-
hols,5 have been prepared and tested in this reaction. On
the other hand, chiral ligands with three coordination sites
have been far less studied.6,7
The enantioselective addition of organozinc reagents to
aldehydes catalyzed by chiral ligands1 is one of the most
important asymmetric processes of synthetic utility, since
the optically active secondary alcohols obtained are the
key intermediates for the preparation of important biologi-
cally active compounds (including several synthetic drugs)
and new materials with interesting optical properties.2
It is expected that a tridentate chiral ligand should confer
more rigidity to the reactive zinc-chelate catalyst than a
bidentate one, and thus may enhance the enantioselectivity
of the addition reaction.7d Nevertheless, the role played
by the third coordination site is more complicated and
unknown. Thus, the very few reported comparison studies
between bidentate and tridentate ligands7 show different
tendencies, from unchanged enantioselectivities7b,d to
enantioselectivity reversal.7a
Since Noyori et al. demonstrated the high efficiency of
3-exo-(dimethylamino)isoborneol (DAIB) 1 (Fig. 1), an
N/O-bidentate b-amino alcohol, as a chiral ligand for the
enantioselective addition of diethylzinc to benzaldehyde,3
explaining such activity in terms of activated bimetallic
species formed from a reactive zinc-chelate catalyst,4 many
Herein we report the factors controlling the catalytic activ-
ity of tridentate amino alcohols as chiral ligands for the
enantioselective addition of diethylzinc to benzaldehyde,
by a comparison of the catalytic behavior of a series of
N/O, N/N/O, and N/O/O 10-aminoisoborneols described
previously by Oppolzer and Aoyama, with a new related
N/N/O ligand obtained and tested by us.
N
OH
1
Figure 1. DAIB, a N/O-bidentate chiral amino alcohol.
2. Results and discussion
*
In 1988, Oppolzer described that both N/O amino isobor-
Corresponding author. Tel.: +34 91 394 5090; fax: +34 91 394 4103;
e-mail: santmoya@quim.ucm.es
neol 2 and N/N/O diamino isoborneol 3 (Fig. 2) are able to
0957-4166/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2008.01.012

270
T. de las Casas Engel et al. / Tetrahedron: Asymmetry 19 (2008) 269–272
(Fig. 2) whereas application of the Oppolzer’s model to
tridentate diamino isoborneol 3 predicts the (pro-S) endo-
syn-6-Si (Fig. 4),7a,11 also in agreement with the experimen-
tal data shown in Figure 2.
OH
OH
N
N
N
2
3
Et
O
R
N/O
82% e.e. (pro-
N/N/O
87% e.e. (pro-S)
O
Et
Et
N
N
Zn
O
Zn
N
R
)
Zn
Et
Zn
R
O
Et
Et
H
H
R
Ph
S
Ph
endo-anti-
7-Re
endo-syn-
6
-Si
OH
OH
Figure 4. Proposed most stable TS for bidentate 2, 4 or 5 (endo-anti-7-Re,
by the application of Noyori’s transition-state model) and for tridentate 3
(endo-syn-6-Si, by the application of Oppolzer’s transition-state model).
Zinc-chelate catalyst in red.
N
N
O
4
5
N/O
N/O/O
On the other hand, N/O/O Aoyama’s ligand 5 (pro-R)
must act as a simple N/O-bidentate ligand, similar to anal-
ogous pro-R ligands 2 and 4 (see Fig. 2). This fact can be
easily explained due to the low coordinative character of
the second oxygen of 5 (ether oxygen), when compared
with the second nitrogen of 3 (amine nitrogen).
77% e.e. (pro-
R)
74% e.e. (pro-R)
Figure 2. Previously described catalytic activities for related isoborneol-
based N/O, N/N/O, and N/O/O ligands.
catalyze the enantioselective addition of diethylzinc to
benzaldehyde with high and similar enantioselectivities
(82% and 87% ee), but with a reversal of the sense of enan-
tioselection.7a Fifteen years later, Aoyama described the
same enantioselection grade (near 75% ee) and sense of
enantioselection (pro-R) for both N/O amino isoborneol
4 and N/O/O oxaamino isoborneol 5 (Fig. 2).8
For testing such hypotheses, we were interested in studying
the catalytic behavior of a new N/N/O ligand of the type of
3, that is, a 10-[(2-aminoethyl)amino]isoborneol. This new
analogue should act as a tridentate ligand and its catalytic
activity should be explained on the basis of a stable Oppol-
zer’s pro-S TS of the type of endo-syn-6-Si (see Fig. 4).
In the seminal work of Oppolzer, the enantioselection
reversal exhibited by 3 was explained on the basis of its
tridentate character, which makes possible the formation
of bimetallic transition states with a pentacoordinated
catalytic zinc atom.7a This pentacoordinated-zinc transi-
tion-state model of Oppolzer 6 is different to the tetracoor-
dinated-zinc model 7, initially proposed by Noyori for
bidentate DAIB4 (Fig. 3).
Following this line, we obtained (1R)-10-(4-methylpipera-
zin-1-yl)isoborneol 812 from commercial (1S)-ketopinic
acid in two steps (amidation with piperazine13 followed
by reduction with LAH,14 overall yield: 82%) and studied
its catalytic activity in the enantioselective addition of
diethylzinc to benzaldehyde15 (Fig. 5).
Surprisingly, instead of the expected (pro-S) catalytic activ-
ity for 8, a moderate (pro-R) enantioselection was observed
(Fig. 5, 50% ee). In principle, this experimental fact seemed
to indicate that 8 is acting in the reaction in a similar man-
ner as (pro-R) amino isoborneols 2, 4, or 5; that is, as a
simple N/O bidentate ligand. On the other hand, a compar-
ison between the use of bidentate ligands 2, 4, or 5, and
new ligand 8 shows a considerable enantioselection-degree
O
Zn
O
O
Zn
O
Zn
Zn
N
Ph
N
Ph
N
6
7
OH
Figure 3. Pentacoordinated-zinc Oppolzer’s transition-state model 6 and
tetracoordinated-zinc Noyori’s one 7. Zinc-chelate catalyst in red.
N
N
8
Application of Noyori’s model to bidentate amino iso-
borneols 2 or 4 predicts the (pro-R) TS (transition state)
endo-anti-7-Re9 (Fig. 4) to be the most stable,10 which is
in agreement with the sense of enantioselection found
N/N/O
50% e.e. (pro-
R)
Figure 5. Catalytic activity of N/N/O ligand 8.

T. de las Casas Engel et al. / Tetrahedron: Asymmetry 19 (2008) 269–272
271
diminution (Fig. 2). However, we believe that the extraor-
dinary similarity between the structures of 3 and 8 (see
Figs. 2 and 5) makes it not possible to discard a tridentate
behavior for 8. In this sense, ligand 8 must be able to
generate Oppolzer-type pentacoordinated-zinc transition
states (type 6 in Fig. 6), as tridentate ligand 3 does. Never-
theless, the mentioned similarity is not enough to maintain
the structure (absolute configuration) for the most stable
TS (controller of the enantioselection sense). Thus, while
endo-syn-6-Si was proposed by Oppolzer as the most stable
TS for N/N/O-tridentate ligand 3 (see Fig. 4),7a we now
propose endo-anti-60-Re as the most stable one for N/N/
O-tridentate ligand 8 (Fig. 6).16
tional flexibility for the catalyst energetically favors the for-
mation of syn-type transition states, whereas low catalyst
flexibility favors the formation of anti ones, and (3) the cat-
alyst’s flexibility can be modulated by the structure of the
nitrogen–nitrogen link.
Acknowledgments
This work was financially supported by the Ministry
of Education and Science, Spain (Research Project
CTQ2004-07244-C02) and UCM-CM (Research Project
910107).
Et
H
References
O
Et
N
Zn
N
Zn
1. (a) Soai, K.; Niwa, S. Chem. Rev. 1992, 92, 833; (b) Pu, L.;
Yu, H. H. L. Chem. Rev. 2001, 101, 767; (c) Catalytic
Asymmetric Synthesis; Ojima, I., Ed.; Wiley-VCH: New
York, 2000; (d) New Frontiers in Asymmetric Catalysis;
Mikami, K., Lautens, M., Eds.; Wiley-Interscience: Hobo-
ken, 2007.
R
O
Et
Ph
endo-anti-6´-Re
2. For example: (a) In drug synthesis: Kanai, M.; Ishii, A.;
Kuriyama, M.; Yasumoto, M.; Inomiya, N.; Otsuka, T.;
Ueda, H. Patent written in Japanese. Patent No. JP
2003226659; In liquid-crystal preparation: (b) Armstrong, J.
D.; McWilliams, J. C. Patent No. US 5977371.
3. Kitamura, M.; Suga, S.; Kawai, K.; Noyori, R. J. Am. Chem.
Soc. 1986, 108, 6071.
4. (a) Kitamura, M.; Okada, S.; Soga, S.; Noyori, R. J. Am.
Chem. Soc. 1989, 111, 4028; (b) Yamakawa, M.; Noyori, R.
J. Am. Chem. Soc. 1995, 117, 6327; (c) Yamakawa, M.;
Noyori, R. Organometallics 1999, 18, 128.
Figure 6. Proposed most stable TS for ligand 8 (endo-anti-60-Re, by
application of Oppolzer’s transition-state model). Zinc-chelate catalyst in
red.
This difference between 3 and 8 can be explained on the
basis of the catalyst’s conformational flexibility (with a
flexible six-membered zinc-chelate ring), and its modula-
tion by the type of geometry imposed by the second-nitro-
gen coordination. Thus, the simple ethylene bridge joining
the two nitrogens in ligand 3 gives place to a set of rela-
tively flexible pentacoordinated-zinc transition states of
type 6. This TS flexibility, joined to the fact that the cata-
lyst’s zinc-ethyl group (red ethyl) is folded over the catalyst
in such pentacoordinated-zinc transition states (note the
different disposition of such ethyl group in both tetracoor-
dinated-zinc endo-anti-7-Re and pentacoordinated-zinc
endo-syn-6-Si, Fig. 4), makes favorable a syn assembly of
the reacting species to the catalyst, which agrees with
Oppolzer’s suggestion of endo-syn-6-Si as the most stable
TS for 3. Nevertheless, the introduction of a second ethyl-
ene bridge between the nitrogens in ligand 8 makes the
corresponding set of pentacoordinated-zinc transition
states more rigid. This increase in rigidity would disfavor
the syn assemblies for the transition states, to the advan-
tage of the anti ones. This return to favored anti-type
transition states, initially proposed by Noyori for N/O-
bidentante ligands4 (i.e., endo-anti-7-Re), agrees with the
suggestion of endo-syn-60-Si as the most stable TS for 8
(note the minimal steric interaction between the close benz-
aldehyde’s hydrogen and catalyst’s zinc ethyl).
5. Some recent examples are: (a) Fontes, M.; Verdaguer, X.;
`
`
Sola, L.; Pericas, M. A.; Riera, A. J. Org. Chem. 2004, 69,
2532; (b) Anaya de Parrodi, C.; Juaristi, E. Synlett 2006,
2699; (c) Sugiyama, S.; Aoki, Y.; Ishii, K. Tetrahedron:
´
´
Asymmetry 2006, 17, 2847; (d) Garcıa Martınez, A.; Teso
´
´
Vilar, E.; Garcıa Fraile, A.; de la Moya Cerero, S.; Martınez
´
Ruiz, P.; Dıaz Morillo, C. Tetrahedron: Asymmetry 2007, 18,
742.
6. Some recent examples are: (a) Stanley, L. M.; Sibi, M. P.
Tetrahedron: Asymmetry 2004, 15, 3353; (b) Murtagh, K.;
Sweetman, B. A.; Guiry, P. J. Pure Appl. Chem. 2006, 78, 311;
(c) Tanaka, T.; Yasuda, Y.; Hayashi, M. J. Org. Chem. 2006,
71, 7091; (d) Kelsen, V.; Pierrat, P.; Gros, P. C. Tetrahedron
2007, 63, 10693; (e) Jiang, F.-Y.; Liu, B.; Dong, Z.-B.; Li,
J.-S. J. Organomet. Chem. 2007, 692, 4377.
7. (a) Oppolzer, W.; Radinov, R. N. Tetrahedron Lett. 1988, 29,
5645; (b) Tanaka, K.; Ushio, H.; Suzuki, H. J. Chem. Soc.,
Chem. Commun. 1989, 1700; (c) Collomb, P.; von Zellesky, A.
Tetrahedron: Asymmetry 1998, 9, 3911; (d) Le Goanvic, D.;
Holler, M.; Pale, P. Tetrahedron: Asymmetry 2002, 13, 119;
(e) Bisai, A.; Singh, P. K.; Singh, V. K. Tetrahedron 2007, 63,
598.
8. Hari, Y.; Aoyama, T. Synthesis 2005, 4, 583.
9. Endo versus exo indicates the coordination face of the
reacting species to the catalyst’s Zn–O bond. syn versus anti
indicates the relative disposition of both zinc-ethyl groups. Re
versus Si indicates the carbonyl face on which the ethyl
transference occurs.
10. Note: (1) the most favored coordination of the reacting
species by the less hindered catalyst’s endo face, (2) the most
favored less-hindered anti disposition for both zinc-ethyl
3. Conclusion
In conclusion, it has been demonstrated that: (1) Oppol-
zer’s transition-state model (pentacoordinated zinc) is a
useful model for explaining the catalytic behavior of N/
N/O-tridentate 10-amino isoborneols, (2) a high conforma-

272
T. de las Casas Engel et al. / Tetrahedron: Asymmetry 19 (2008) 269–272
groups, and (3) the minimal steric interaction for benzalde-
hyde’s phenyl group.
14. Standard LAH reduction in refluxing THF (for example, see
Ref. 8). Yield: 84%.
11. In this case, the geometry imposed by the pentacoordinated
zinc allows the syn disposition of both zinc-ethyl groups (both
syn groups are now more separated).
15. Under argon, diethylzinc (1.0 M in hexane, 2 mL, 2.0 mmol)
was added to ligand 8 (0.050 mmol) in anhydrous hexane
(1 mL) and the mixture stirred for 1 h at rt. After that, freshly
distilled benzaldehyde (1.0 mmol) was slowly added and the
resulting mixture stirred for 5 h at rt. Final treatment with
1 N HCl at 0 °C and standard work up (see Ref. 8) yielded the
resulting scalemic mixture of 1-phenylpropan-1-ol. Yield:
99%. The ee (50%) was determined GC (cyclodex-B). The
dominant configuration was determined by both the sign of
the mixture’s specific rotation and the elution time in chiral
GC.
16. Piperazine ring shows a preference for the boat conformation
in the coordination with different metals including Zn(II). For
example, see: (a) Bharati, K. S.; Sreedaran, S.; Rahiman, A.
K.; Rajesh, K.; Narayanan, V. Polyhedron 2007, 26, 3993; (b)
Valks, G. C.; McRobbie, G.; Lewis, E. A.; Hubin, T. J.;
Hunter, T. M.; Sadler, P. J.; Pannecouque, C.; De Clercq, E.;
Archibald, S. J. J. Med. Chem. 2006, 49, 6162; (c) Nuutinen,
J. M. J.; Ratilainen, J.; Rissanen, K.; Vaniotalo, P. J. Mass
Spetrom. 2001, 36, 902.
20
12. White solid. Mp: 65–66 °C. ½aꢀD ¼ ꢁ66:2 (c 0.38, CH2Cl2).
1H NMR (CDCl3, 200 MHz): 3.93 (dd, J = 7.87, 3.86, 1H),
2.77 (d, J = 13.22, 1H), 2.44 (br s, 8H), 2.27 (d, J = 13.22,
1H), 2.26 (s, 3H), 1.89–1.19 (several m, 7H), 1.13 (s, 3H),
1.08–0.84 (m, 1H), 0.79 (s, 3H). 13C NMR (CDCl3, 50 MHz):
78.2, 58.7, 55.3, 54.7, 50.8, 45.9, 44.7, 39.0, 34.0, 27.6, 20.4,
20.4. FTIR: 3207.5 (str.). MS (%): 252 (7), 234 (6), 219 (11),
113 (26), 100 (61), 71 (100), 57 (99). HRMS: 252.2203 (calcd
for C15H28N2O: 252.2202).
13. (1S)-Ketopinic acid (5.0 mmol) and SOCl2 (5.2 mmol) were
refluxed in anhydrous THF for 15 min. The reaction mixture
was then cooled down to 0 °C and 4-methylpiperazine
(10.5 mmol) slowly added. After that, the reaction mixture
was stirred for 1 h at rt. Standard work-up (extraction
and acid- and base-washing) yield pure amide in 98% yield.
Excess of the amine can be recovered from the aqueous acid
extract.