Enantioselective intramolecular aldol addition/dehydration reaction of a macrocyclic diketone: Synthesis of the musk odorants (R)-muscone and (R,Z)-5-muscenone

Article abstract of DOI:10.1002/anie.200604518

(Chemical Equation Presented) One hundred years after the first isolation of natural (R)-muscone [(R)-1], a short and efficient access to the outstanding odorants (R)-1 and (R,Z)-5-muscenone [(R)-2] has become possible thanks to an unprecedented, reversible intramolecular aldol addition/enantioselective dehydration (up to 76% ee; see scheme).

Full text of DOI:10.1002/anie.200604518

Angewandte
Chemie
DOI: 10.1002/anie.200604518
Musk Odorants
Enantioselective Intramolecular Aldol Addition/Dehydration
Reaction of a Macrocyclic Diketone: Synthesis of the Musk Odorants
(R)-Muscone and (R,Z)-5-Muscenone**
Oliver Knopff,* JØrôme Kuhne, and Charles Fehr
One hundred years after the isolation of natural (R)-muscone
[(R)-1] by Walbaum,[1] we describe a short and efficient
synthesis of (R)-1 and (R,Z)-5-muscenone [(R)-2] (up to
Scheme 1. Synthetic strategy for the synthesis of (R)-1 and (R)-2 from
3: a) EF=Eschenmoser fragmentation: H2NNHTs, cat. AcOH, toluene,
reflux; then AcOOH; b) H2, cat. Lindlar, EtOH; c) H2, cat. Pd/C,
EtOH.[2,6b–c]
76% ee) by using an unprecedented, enantioselective intra-
molecular aldol addition/dehydration reaction as the key step.
Compounds (R)-1 and (R)-2 are macrocyclic musks in
which the configuration at C3 influences very strongly the
odor character and the human olfactory threshold.[2] Musk
(R)-2 has an extremely low threshold (0.027 ngLÀ1) and
unsuccessful (low conversion of 3). In view of the low
reactivity of diketones, it is not surprising that there are only
two known direct asymmetric intramolecular aldol reactions,
from Agami et al.,[9] Hajos and Parrish,[10a] and Wiechert and
co-workers.[10b]
possesses
a
highly desired nitromusk character [(S)-2:
As 3 was inert to the reported reaction conditions (l-
proline, N,N-dimethylformamide), we decided to study sys-
tematically the reactivity of 3 towards metal isopropoxides,
MOiPr.[11] Surprisingly, quantitative formation of 4 at room
temperature was observed in the presence of two equivalents
of NaOiPr after one day.
3 ngLÀ1]. Compound (R)-1 is appreciated for its strong
animal musk character [(S)-1: weakly musky].
To our knowledge, neither of the two musk odorants has
ever been found in commercial fragrances, which indicates
that the published syntheses do not completely fulfill the
requirements for a large-scale preparation.[3] In view of the
growing interest from the fragrance industry[4] and academia[5]
in (R)-1, and the exceptional olfactory characteristics of (R)-
2, practical syntheses of both of these compounds are highly
desirable.
After screening a selection of Na alkoxides 8–11 (Table 1)
derived from chiral b-amino alcohols, we were delighted to
Table 1: The effect of alkoxides 8–11 on the yield and enantioselectivity of
the formation of 4 from 3 in THF.[12]
Our synthetic strategy was to develop an enantioselective
aldol condensation of the readily available macrocyclic
diketone 3[6a] for the formation of product (S)-4, which then
could be easily transformed into both (R)-1 and (R)-2
following the previously published synthesis (Scheme 1).[2,6b–c]
Over the last decade impressive achievements in direct
asymmetric aldol methodology have been made for the
reaction between ketones and aldehydes.[7] However, our
attempts with the use of alkoxides ([Zn]OR*,[8a] [Ba]OR*,[8b]
[Ca]OR*,[8c] [La]OR*,[8d] [Ti]OR*,[8e]) or l-proline[8f] were
Entry Alkoxide
t [days] c [molLÀ1 [a]
]
Conv. [%][b] ee [%][e]
1
2
3
4
5
6
7
4 equiv 8
4 equiv 9
4 equiv 10
4 equiv 11
4 equiv 8
2 equiv 8
8 equiv 8
3
3
3
3
1
1
2
0.1
0.1
0.1
0.1
0.8
0.8
1
88
53 (S)
85
36 (S)
49
25 (S)
[*] Dr. O. Knopff, J. Kuhne, Dr. C. Fehr
Firmenich SA
56
50 (S)
99 (95)[c]
64 (96)[d] (S)
56 (S)
Corporate R&D Division
B.P. 239, 1211, Gen›ve 8 (Switzerland)
Fax: (+41)22-780-3334
91
90
76 (S)
E-mail: oliver.knopff@firmenich.com
[a] Concentration of 3 at the beginning of the reaction. [b] Conversion
was determined by GC. [c] Yield of isolated products. [d] ee value after 2
recrystallizations. [e] Determined by chiral GC analysis (CHIRASIL DEX
CB) after reduction to the alcohol.
[**] We thank Dr. J.-Y. de Saint Laumer (Firmenich SA) for the energy
calculations, and B. Egger and E. Foures (Firmenich SA) for their
work in the laboratory.
Angew. Chem. Int. Ed. 2007, 46, 1307 –1310
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1307

Communications
obtain (S)-4 in an enantiomeric excess of 53% in the presence
of four equivalents of the Na alkoxide of (+)-N-methylephe-
drine (8) (Table 1, entry 1, 88% yield, 3 days in THF).[12]
Notably, transient aldol 7 was not observed during the
course of this reaction.
Interestingly, the enantiomeric excess was strongly de-
pendent on the configuration at C2 of the b-amino alcohol
(Table 1, entry 2, 9, 36 % ee) and the relative size of the amino
moiety (Table 1, entry 3, 10, 25% ee), whereas a more bulky
phenyl group at C2 gave nearly the same enantiomeric excess
(Table 1, entry 4, 11, 50% ee).
An even higher ee value and a higher reaction rate could
be obtained with four equivalents of 8 by performing the
Scheme 3. Simplified illustration of the dynamic kinetic resolution of
aldol intermediate (Æ)-7 mediated by Na alkoxide 8 [NaOR]n·[THF]m.[18]
reaction at
a higher concentration (Table 1, entry 5,
0.8 molLÀ1). After allowing the reaction to run for one day
at room temperature, we obtained (S)-4 in quantitative yield
and with an enantiomeric excess of 64%. Importantly, lower
quantities of 8 (2 equiv) gave a lower ee value (Table 1,
entry 6, 56% ee), and higher quantities of 8 (8 equiv) gave a
higher ee value (Table 1, entry 7, 76% ee). (+)-N-Methyl-
ephedrine could be easily recycled (up to 98%) and reused
for the formation of 8.
Subsequently, (S)-4 was transformed into the musk odor-
ants (R)-muscone (1) and (R,Z)-5-muscenone (2) in high
yield (two steps, 70%)[2] without any loss of enantiomeric
excess.
Owing to the fact that Na alkoxide 8 has already been
used by Plaquevent and co-workers[19] for enantioselective
dehydrohalogenations, it is reasonable to assume that (S)-4 is
formed by the enantiomer-differentiating dehydration of 7
(pathway A is faster than pathway B).[20,21] Interconversion
between the enantiomers of 7 through a retro-aldol/aldol
addition sequence enables the transformation of undesired
aldol (R)-7 to desired aldol (S)-7 (dynamic kinetic resolution).
Conformational analysis of 7 (Figure 1)[17] shows that its
sterically very demanding 11-membered ring hinders attack of
8 on one side of the bridgehead proton. Chelation of the
sodium cation of 8 to the oxygen and nitrogen atoms should
give a rigid conformer with a sterically demanding face
(methyl and phenyl groups).[22] We speculate that in the faster
deprotonation of 7 with 8, the methyl and phenyl groups of 8
point away from the bulky 11-membered ring (Figure 1 and
Scheme 3, pathway A).[19,23] This mechanism is consistent with
the observation that the Na alkoxide of (+)-N-methylpseu-
doephedrine (9) gives a lower enantiomeric excess[24] than 8
and that 11 (2 phenyl groups on the same side) affords high
enantioselectivities. The observation that sterically demand-
ing alkyl substituents on the nitrogen atom (10) lower the
enantiomeric excess and slow down the reaction rate could be
a result of a stronger steric interaction of the alkyl groups with
the 11-membered ring during the deprotonation step.
One way to answer the question of whether the aldol
reaction or the dehydration step (or both) were responsible
for the high enantioselectivity would be the determination of
the enantiomeric excess of aldol intermediate 7. As the
quantity of transient 7 was too low (< 1%),[13] racemic (Æ )-
7
[14] was prepared from 3 and treated with two equivalents of
Na alkoxide 8 to determine the enantioselectivity in the
dehydration step (Scheme 2). In less than one minute, 3 was
We have thus developed a short and efficient synthesis of
(R)-muscone (1) and (R,Z)-5-muscenone (2) (up to 76% ee)
from a macrocyclic diketone by using an unprecedented,
reversible intramolecular aldol addition/enantioselective
dehydration reaction.[25] This sodium ephedrate mediated
Scheme 2. Preparation of (Æ)-7: a) 1.5 equiv ZrCl3OPr, 1.7 equiv NBu3,
CH2Cl2, À108C. Reactivity of aldol (Æ)-7 towards Na alkoxide 8.
formed quantitatively,[15] which indicates that the retro-aldol
reaction is much faster than the aldol dehydration reaction.
Not surprisingly, relative energy calculations showed that
retro-aldol product 3 (À12.1 kcalmolÀ1) is much lower in
energy than aldol 7 (0 kcalmolÀ1) and dehydration product
(S)-4 (À2.7 kcalmolÀ1).[16,17]
On the basis of the experimental results, we suggest the
following mechanism for the formation of (S)-4: In the
presence of a large excess of 8, both enantiomers of transient
7 are formed from 3 in small amounts by reversible steps (Na
enolate formation, aldol addition, and protonation;
Scheme 3).[18]
Figure 1. Preferred conformation of aldol (1S,11S,14S)-7.[17] Simplified
illustration of the enantiomer-differentiating aldol dehydration of 7.[18]
1308
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 1307 –1310

Angewandte
Chemie
Chim. Acta 1967, 50, 2101; d) for a review of the synthesis of
reaction represents an efficient and hitherto unknown type of
dynamic kinetic resolution, which involves a new class of aldol
intermediates.[26] Currently, we are expanding this novel
reaction to other ring sizes and acyclic diketones.
macrocyclic musks, see: A. S. Williams, Synthesis 1999, 1707.
[7] a) B. Alcaide, P. Almendros, Angew. Chem. 2003, 115, 884;
Angew. Chem. Int. Ed. 2003, 42, 858; b) C. Palomo, M. Oiarbide,
J. M. Garcia, Chem. Eur. J. 2002, 8, 36; c) R. Mahrwald (Ed.)
Modern Aldol Reactions, Vols. 1 & 2, Wiley-VCH, Weinheim,
2004; for a review, see: T. D. Machajewski, C.-H. Wong, Angew.
Chem. 2000, 112, 1406; Angew. Chem. Int. Ed. 2000, 39, 1352.
[8] a) B. M. Trost, H. Ito, E. R. Silcoff, J. Am. Chem. Soc. 2001, 123,
3367; b) Y. M. A. Yamada, M. Shibasaki, Tetrahedron Lett. 1998,
39, 5561; c) T. Suzuki, N. Yamagiwa, Y. Matsuo, S. Sakamoto, K.
Yamaguchi, M. Shibasaki, R. Noyori, Tetrahedron Lett. 2001, 42,
4669; d) Y. M. A. Yamada, N. Yoshikawa, H. Sasai, M. Shiba-
saki, Angew. Chem. 1997, 109, 1942; Angew. Chem. Int. Ed. Engl.
1997, 36, 1871; e) R. Mahrwald, Org. Lett. 2000, 2, 4011; f) B.
List, R. A. Lerner, C. F. Barbas III, J. Am. Chem. Soc. 2000, 122,
2395.
[9] C. Agami, N. Platzer, H. Sevestre, Bull. Soc. Chim. Fr. 1987, 2,
358. There is also an example of an aldolase antibody catalyzed
aldol condensation: B. List, R. A. Lerner, C. F. Barbas III, Org.
Lett. 1999, 1, 59.
[10] a) Z. G. Hajos, D. R. Parrish, J. Org. Chem. 1974, 39, 16 15; b) U.
Eder, G. Sauer, R. Wiechert, Angew. Chem. 1971, 83, 492;
Angew. Chem. Int. Ed. Engl. 1971, 10, 496; c) H. Sasai, T. Susuki,
S. Arai, T. Arai, M. Shibasaki, J. Am. Chem. Soc. 1992, 114, 4418;
for a review, see: B. List, Tetrahedron 2002, 58, 5573.
[11] When Mg, Ca, and Zn isopropoxides were employed, no trace of
4 was detected, and with Li, Ba, La, Ti, and Zr isopropoxides
only 1–3% of 4 was obtained. The study was performed in N-
methylpyrrolidone (NMP).
Experimental Section
Sodium alkoxides 8–11 can be prepared from the corresponding b-
amino alcohols in several ways (for example by the addition of 1 equiv
of NaH in THF, followed by heating at reflux and stirring for
30 min).[19]
(S)-4 (64% ee): A mixture of (+)-N-methylephedrine (2.9 g,
16 mmol), NaH (60 wt% dispersion in mineral oil, 0.64 g, 16 mmol),
and 4–Š molecular sieves (0.8 g) in dry THF (5 mL) was heated at
reflux and stirred for 30 min. The mixture was cooled to room
temperature, 3 (4 mmol, 1.0 g) was added, and the mixture was
stirred. The reaction was followed by GC. To stop the reaction, the
mixture was hydrolyzed with an aqueous HCl solution (2n, 15 mL).
After extraction of the aqueous layer with diethyl ether, the organic
layer was washed with water, dried over MgSO4, and filtered. The
solvent was removed under vacuum, and the residue was purified by
flash chromatography. The ee value was determined by reduction of
(S)-4 to the corresponding alcohol (LiAlH4 in dry THF) and injection
onto a chiral GC column (CHIRASIL DEX CB).[2]
Received: November 3, 2006
Published online: January 9, 2007
Keywords: aldol reaction · amino alcohols ·
.
asymmetric synthesis · dynamic kinetic resolution · elimination
[12] THF was superior to tBuOMe, 1,4-dioxane, toluene, 1,2-
dimethoxyethane, NMP, CH2Cl2, and EtOAc.
[13] Reaction intermediates (aldol 7, Na enolate of 3) were not
detected during the slow formation of 4 from 3 with the use of
four equivalents of 8 (13C NMR in [D8]THF). It should be noted
that 8 acts in the presence of 4 like a chiral shift reagent
[different shift of the 13C NMR signals of the carbonyl function-
ality and the double bond of (S)-4 and (R)-4]. This suggests that
there is coordination of 8 to 4. No shift in the 13C NMR signal of
the keto group of diketone 3 was observed.
[14] Aldol product 7 with the indicated configuration (1R,11R,14R)/
(1S,11S,14S) is the major product. Only small amounts (1%) of 7
with the (1R,11S,14R)/(1S,11R,14S) configuration were isolated.
Calculation of the relative energy showed that aldol product
(1R,11S,14R)/(1S,11R,14S)-7 is higher in energy (0.9 kcal
molÀ1).[17]
[15] Upon continued stirring at room temperature, aldol condensa-
tion product (S)-4 was slowly formed from 3 in 53% ee (3 days,
88%).
[16] Aldol product 7 is stable to chromatography but it forms retro-
aldol product 3 during GC injection.
[17] The structures of the molecules have been optimized by using
classical multiple conformational analysis at the molecular
mechanic level (MMFF) with the use of MacroModel 8.1, Ed.
2003, Schrödinger Inc., Portland, OR, USA, 2000. The energy
minimal conformations were then calculated at higher level DFT
(B3LYP/6.31G**) by using the program Jaguar 5.5, Ed. 2003,
Schrödinger Inc., Portland, OR, USA, 2003.
[18] For clarity, the illustration is simplified. We propose that
aggregated intermediates are involved in the aldol and dehy-
dration reaction: P. G. Williard, Q.-Y. Liu, J. Am. Chem. Soc.
1993, 115, 3380, and references therein.
[1] H. J. Walbaum, J. Prakt. Chem. 1906, 73, 488; L. Ruzicka, Helv.
Chim. Acta 1926, 9, 715.
[2] C. Fehr, J. Galindo, O. Etter, Eur. J. Org. Chem. 2004, 1953, and
references therein.
[3] A recent literature search revealed 44 published and 19 patented
syntheses for (R)-muscone (1). Only one commercial source of
(R)-1 (86% ee) could be found, which offers the latter for
around 4800 E/kg (Toyotama). For (R,Z)-5-muscenone (2), only
one synthesis has been published.[2]
[4] a) A. Boix Camps, Tagasako Times 2006, 157, 26: “laevo-
muscone is the best musk chemical in the world”; b) Y.
Matsumura, H. Fukawa, A. Endo, JP 2002_335991, 2002;
[Chem. Abstr. 2002, 137, 383885]; c) G. Kim, M. S. Park, J. H.
Yoo, S. Y. Lee, KR 2000049980, 2000 [Chem. Abstr. 2002, 137,
124944].
[5] a) B. Bulic, U. Lücking, A. Pfaltz, Synlett 2006, 1031; b) M. Ito, S.
Kitahara, T. Ikariya, J. Am. Chem. Soc. 2005, 127, 6 172; c) C.
Fehr, J. Galindo, I. Farris, A. Cuenca, Helv. Chim. Acta 2004, 87,
1737; d) P. Scafato, S. Labano, G. Cunsolo, C. Rosini, Tetrahe-
dron: Asymmetry 2003, 14, 3873; e) P. K. Fraser, S. Woodward,
Chem. Eur. J. 2003, 9, 776; f) S. Fujimoto, K. Yoshikawa, M. Itoh,
T. Kitahara, Biosci. Biotechnol. Biochem. 2002, 66, 1389; g) T.
Yamamoto, M. Ogura, T. Kanisawa, Tetrahedron 2002, 58, 9209;
h) Y. H. Choi, J. Y. Choi, H. Y. Yang, Y. H. Kim, Tetrahedron:
Asymmetry 2002, 13, 801; i) V. P. Kamat, H. Hagiwara, T.
Katsumi, T. Hoshi, T. Suzuki, M. Ando, Tetrahedron 2000, 56,
4397; j) A. Alexakis, C. Benhaïm, X. Fournioux, A. van den
Heuvel, J.-M. LevÞque, S. March, S. Rosset, Synlett 1999, 1811,
and references therein.
[6] a) G. Ohloff, J. Becker, K. H. Schulte-Elte, Helv. Chim. Acta
1967, 50, 705; b) A. Eschenmoser, D. Felix, G. Ohloff, Helv.
Chim. Acta 1967, 50, 708; c) J. Schreiber, D. Felix, A. Eschen-
moser, M. Winter, F. Gautschi, K. H. Schulte-Elte, E. Sundt, G.
Ohloff, J. Kalvoda, H. Kaufmann, P. Wieland, G. Anner, Helv.
[19] M. Amadji, J. Vadecard, D. Cahard, L. Duhamel, P. Duhamel, J.-
C. Plaquevent, J. Org. Chem. 1998, 63, 5541, and references
therein.
Angew. Chem. Int. Ed. 2007, 46, 1307 –1310
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1309

Communications
[20] For an example of a kinetic resolution of a racemic aldol by l-
[23] The observation of a positive nonlinear effect (4 equiv of 8 in
THF) suggests the existence of dimers or higher aggregates of 8.
For other examples of nonlinear effects in the presence of b-
amino alcohols, see: C. Girard, H. B. Kagan, Angew. Chem.
1998, 110, 3088; Angew. Chem. Int. Ed. 1998, 37, 2922, and
references therein. The precoordination of alkoxide 8 to the keto
group of aldol product 7 is likely, but there is no experimental
proof.
proline-catalyzed dehydration, see: C. Agami, F. Meynier, C.
Puchot, J. Guilhem, C. Pascard, Tetrahedron 1984, 40, 1031.
[21] We have no evidence for the formation of reaction intermedi-
ates, which could act as a base and might compete with 8 in the
enantiomer-differentiating dehydration reaction.[13]
[22] For X-ray structures of Li ephedrates (tetramer) and a Na
aminoalkoxide (tetramer), see: a) M. A. Nichols, A. T. McPhail,
E. M. Arnett, J. Am. Chem. Soc. 1991, 113, 6 222; b) E. M. [24] D. Cahard, L. Ferron, J.-C. Plaquevent, Synlett 1999, 960.
Arnett, M. A. Nichols, A. T. McPhails, J. Am. Chem. Soc. 1990,
112, 7059; c) G. Müller, T. Schꢀtzle, Z. Naturforsch. B 2004, 59,
1400.
[25] O. Knopff, patent submitted (Firmenich SA, prior. 16.02.2004).
[26] For an example involving a dynamic kinetic resolution of an
aldol, see: F. F. Huerta, J.-E. Bꢀckvall, Org. Lett. 2001, 3, 1209.
1310
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 1307 –1310