(+)-(R,Z)-5-Muscenone and (-)-(R)-Muscone by enantioselective aldol reaction and grob fragmentation

Article abstract of DOI:10.1002/chem.200902774

(+)-(R,Z)-5-Muscenone ((R)-1) was synthesized by an enantioselective aldol reaction, catalyzed by new ephedrine-type Ti reagents (up to 70% enantiomeric excess). Substrate-directed diastereoselective reduction of the aldol product and Grob fragmentation of the tosylate of the resultant 1,3-diol afforded (+)-1. This approach also gave access to (-)-(R,E)-5-muscenone and (-)-(R)-muscone.

Full text of DOI:10.1002/chem.200902774

FULL PAPER
DOI: 10.1002/chem.200902774
(+)-(R,Z)-5-Muscenone and (À)-(R)-Muscone by Enantioselective Aldol
Reaction and Grob Fragmentation
Charles Fehr,*[a] Andrea K. Buzas,[a, b] Oliver Knopff,[a] and Jean-Yves de Saint Laumer[a]
Dedicated to Professor Albert Eschenmoser on the occasion of his 85th birthday
Abstract: (+)-(R,Z)-5-Muscenone ((R)-1) was synthesized by an enantioselective
aldol reaction, catalyzed by new ephedrine-type Ti reagents (up to 70% enantio-
meric excess). Substrate-directed diastereoselective reduction of the aldol product
and Grob fragmentation of the tosylate of the resultant 1,3-diol afforded (+)-1.
This approach also gave access to (À)-(R,E)-5-muscenone and (À)-(R)-muscone.
Keywords: aldol reaction · amino
alcohols · asymmetric synthesis ·
fragmentation · fragrances
Introduction
Our synthesis was based on a catalytic kinetic resolution
by enantioselective Corey–Bakshi–Shibata (CBS) reduction
of racemic bicyclic enone (Æ)-5, readily accessible by intra-
molecular aldolization of muscodione (4) under basic condi-
tions (Scheme 1). An Eschenmoser fragmentation
(H2NNHTs, cat. AcOH, toluene, then AcOOH), followed
by Lindlar hydrogenation then afforded (R)-1.
Muscenone, a highly appreciated Firmenich musk odorant,
is a racemic mixture of (Z)-5-muscenone (1), (E)-5-musce-
none (2) (ca. 80%), and (E)- and (Z)-4-muscenone (ca.
20%).[1] The 5-muscenones are more strongly scented than
the 4-muscenones, and 1 (ca. 45%) is particularly appreciat-
ed for its better top note and nitro-musk character.
Having identified (R)-1 as the ultimate target, we then
had the challenge to envisage an enantioselective intramo-
As the natural (R)-muscone 3 exhibits a much more pro-
nounced musk character than its enantiomer, some years
ago we prepared the then unknown enantiomers of 1 and
2.[2] (R,Z)-5-Muscenone ((R)-1) turned out to be much more
strongly scented than its enantiomer, and its threshold value
is more than 100 times lower than that of (S)-1.[2]
[a] Dr. C. Fehr, Dr. A. K. Buzas, Dr. O. Knopff,
Dr. J.-Y. de Saint Laumer
Firmenich SA, Corporate R&D Division, P.O. Box 239
1211 Geneva 8 (Switzerland)
Fax : (+41)22-780-3334
E-mail: charles.fehr@firmenich.com
[b] Dr. A. K. Buzas
Scheme 1. (R)-1 by kinetic resolution (CBS reduction) according to refer-
Postdoctoral fellow at Firmenich SA (November 2008–October 2009)
ence [2].
Chem. Eur. J. 2010, 16, 2487 – 2495
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2487

lecular aldol reaction of 4. It should be noted that, in princi-
ple, deprotonation already leads to a chiral product, since it
is the result of an enantiotopic group differentiation; if the
deprotonation is reversible, it would then, in principle, be
possible to effect the intramolecular aldol reaction in a ki-
netically controlled, enantioselective manner. If all preced-
ing steps are reversible, the irreversible dehydration can
take place enantioselectively. This pathway, a dynamic kinet-
ic resolution (DKR), has been successfully realized at Fir-
menich by using four (or eight) equivalents of the sodium
alkoxide of (+)-N-methylephedrine (9). Compound (S)-5
was isolated in almost quantitative yield with 64% enantio-
meric excess (ee) (Scheme 2).[3]
Scheme 3. New strategy: (R)-1 by enantioselective aldol reaction, diaste-
reoselective reduction, and Grob fragmentation.
diketones 11,[9b] and the transannular aldolizations of 16[7]
and 18[9c] (Scheme 4).
It is important to note that ketone–ketone aldolizations
are difficult to achieve because of the inherent propensity of
the aldol product to undergo a retro-aldol reaction. The lan-
thanide-catalyzed aldol addition of triketone 10 (n=1) to 12
is a typical example of this problem of reversibility.[6]
Another possible issue in the aldol reaction of 4 is the dia-
stereocontrol (7 versus 8). In contrast to the literature exam-
ples (17[7] and 19[9c]), which involve the formation of stable
cis-fused five-membered rings, we wanted to favor the less
stable cis-aldol product 7.
Herein, we describe the enantioselective aldolization of
muscodione 4, promoted by new chiral Ti reagents (70% ee)
and the stereocontrolled reduction/fragmentation of crystal-
lized aldol (À)-7 (ꢀ98% ee) to afford (R)-1 with an unal-
tered ee value.
Scheme 2. Reversible aldolization/enantioselective dehydration according
to reference [3].
In parallel, we explored the possibility of an enantioselec-
tive aldolization through the formation of the cis-fused aldol
product 7 (less stable than 8), out of four possible diastereo-
mers. Subsequent diastereoselective reduction and Grob
fragmentation of the corresponding tosylate should guaran-
tee the direct formation of the desired Z olefin (Scheme 3),
as an alternative to Eschenmos-
Results and Discussion
The enantioselective aldol reaction: The preliminary aldol
reactions of muscodione (4) with (S)-proline gave no reac-
tion; consequently we tested chiral Li and Mg amides in
er fragmentation of (S)-5 fol-
lowed by Lindlar hydrogena-
tion.
In recent years, considerable
progress has been achieved in
the area of enantioselective
aldol reactions,[4] based on
metal enolate chemistry[5–7] or
organocatalysis.[8,9]
However,
these reactions mostly involve
aldehydes, or sometimes a-ke-
toesters, as the electrophilic
partners. The rare examples of
ketone–ketone aldolizations are
restricted to intramolecular
cases, such as the desymmetri-
zations of triketones 10[6,9a] or Scheme 4. Published enantioselective ketone–ketone aldolizations.
2488
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2487 – 2495

Enantioselective Aldol and Grob Fragmentation
FULL PAPER
place of (S)-proline. Li amide 20 deprotonated 4, but no
aldol reaction ensued, and the Mg amides 21 and 22[10] also
showed a low reactivity (ca. 10% conversion). The 1,1’-bi(2-
naphthol) (BINOL)-derived mixed La–Li–BINOL alkoxide,
LLB-II,[6a] was also ineffective, even when doped with potas-
sium hexamethyl disilazide (KHMDS) or NaOtBu.
A significant improvement in both diastereoselectivity (al-
lowing the formation of the desired diastereomer 7) and
enantioselectivity (up to 70% ee) was attained with the Ti
reagents formed from TiCl4 and ephedrine-type amino alco-
hols (+)-26 to (À)-32. We found that the Ti reagents ob-
tained from TiCl4 and an amino alcohol possessing a tertiary
amino group (26 to 32) in N-methylpyrrolidone (NMP) and
TMEDA effectively promoted the aldol reaction of 4. In
most cases variable amounts of the less stable cis-diastereo-
mer 7 were formed together with 8. After having identified
a better selectivity with (+)-N-methylephedrine ((+)-26) (7/
8=45:55; 7: 36% ee, 8: 20% ee) than with (À)-N-methyl-
pseudoephedrine ((À)-33) (7/8=10:90; 7: <7% ee, 8:
7% ee) (Table 1; entries 1 and 2), we focused on the ephe-
drine series. The more highly substituted (+)-N-isopropyle-
phedrine ((+)-27)[14] gave 7 with 44–46% ee (7/8 ꢁ1:1) and
was used for several optimization experiments (Table 1; en-
tries 3 and 4).[15] Lastly, we varied the steric bulk of the N-
substituents to find the most efficient amino alcohol in
terms of enantioselectivity, by analogy with a study of Alex-
akis and co-workers on diamines.[16]
Finally, it was found that Ti or Zr enolates underwent the
desired aldol addition.[11] TiCl4 (1.2 equiv)/NBu3 (1.4 equiv)
in CH2Cl2 at À58C smoothly afforded aldol products 7 and
8 in a ratio of approximately 1:5. Recrystallization from
heptane afforded pure 8 (46%) and the minor isomer 7
(10%) was isolated after chromatography of the mother liq-
uors (Scheme 5). Treatment of 4 with ZrCl3OPr/NBu3 in
CH2Cl2 at À108C (or ZrCl4/N,N,N’,N’-tetramethyl-1,2-
ethane (TMEDA) in CH2Cl2 at 258C) also afforded 8 in
88% yield. In both 7 and 8 the OH group is axial and the
Me group is equatorial (trans with respect to the OH
group), as shown by NMR spectroscopy. New calculations
indicate that cis-aldol 7 is less stable than trans-aldol 8 by
approximately 1.8 kcalmolÀ1. The two other possible diaste-
reomers with a cis relationship between the OH and Me
groups were not observed.
Table 1. Enantioselective aldol reaction of 4.
Entry Amino
T
Time Conversion Yield of 7
Yield of 8
alcohol [8C] [h]
[%]
[%] (ee [%]) [%] (ee [%])
1[a]
(+)-26
(À)-33
(+)-27
(+)-27
(+)-28
(+)-28
(+)-29
(+)-30
(+)-31
(À)-32
(À)-32
(À)-32
(À)-32
8
2
ꢁ20
ꢁ9 (36)
ꢁ5 (<7)
29 (44)
33 (46)
8 (67)
ꢁ11 (20)
ꢁ45 (7)
32 (18)
34 (19)
23 (18)
36 (24)
24 (18)
30 (0)
39 (25)
27 (17)
52 (17)
12 (20)
12 (38)
2[a]
25 1.5 ꢁ50
3[b]
0
3
ꢁ65
ꢁ75
ꢁ35
ꢁ50
ꢁ50
ꢁ40
ꢁ70
ꢁ45
ꢁ85
4[b]
À20 72
5[b]
0
0
0
0
0
0
3
9
5
9
5
4
6[b]
7 (68)
7[b]
21 (40)
ꢁ1 (–)
23 (45)
10 (64)
21 (70)
9 (62)
8[b]
We next tested the aldol reaction with TiCl4/(À)-sparteine
((À)-23) in CH2Cl2 at À708C. After 90 min, the reaction was
quenched (ca. 65% conversion) and 8 was isolated in 42%
yield (7% ee). The combination Cl3TiOiPr/(À)-23/CH2Cl2/
À608C also afforded 8 with a low ee value (10%). Variation
of the reaction conditions (temperature, solvent (CH3CN))
did not change the outcome of the reaction. Replacement of
(À)-23 by diamine 24[12] or aminoether 25,[13] or the use of
9[b]
10[b]
11[b]
12[c]
13[c]
À20 96
0
7.5 ꢁ40
À20 72 ꢁ30
11 (68)
Reaction conditions: [a] Amino alcohol/TiCl4 (2.40 equiv), TMEDA
(2.80 equiv), NMP. [b] Amino alcohol/TiCl4 (0.80 equiv), TMEDA
(0.93 equiv), NMP, H2O (0.1 equiv). [c] [(À)-32·TiCl4 (0.40 equiv),
TMEDA (0.46 equiv), NMP, H2O (0.05 equiv).
[5c]
methyl mandelate or mandelic acid and TiCl4 gave 8 with
a maximum ee value of 12%.
We found that use of (+)-
28·TiCl4, possessing an N-isobu-
tyl substituent, gave 7 with a
much higher ee value (68% ee;
Table 1; entry 6) compared with
(+)-27·TiCl4 (bearing an N-iso-
propyl
group)
(44% ee;
Table 1; entry 3). Going from
N-isobutyl to N-isopentyl ((+)-
29·TiCl4), the ee value dropped
again
(40% ee;
Table 1;
entry 7). The analogous N-tert-
butyl-substituted reagents (+)-
30·TiCl4 and (+)-31·TiCl4 were
no more selective (Table 1; en-
Scheme 5. Synthesis of (Æ)-7 and (Æ)-8 by Ti- and Zr-mediated aldol reactions.
Chem. Eur. J. 2010, 16, 2487 – 2495
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2489

C. Fehr et al.
À
ment of one chlorine atom at Ti by an oxygen atom (Ti O
bond). Therefore, we assume that the active complex is an
adduct between the amino alcohol and TiCl4, in which the
OH bond (as part of a chelate) is still intact (e.g., (À)-
32·TiCl4) and can be considered as a combined Lewis acid–
Brønsted acid catalyst.[18] A broad signal between 3500 and
2400 cmÀ1 and a signal at 1602 cmÀ1 in the IR spectrum sup-
ports this assertion.
When the experiment of entry 10 (Table 1) was repeated
with (À)-32·TiCl4 of 50% ee, the aldol product 7 had an ee
value of 32%, thus indicating a linear relationship between
reagent ee and product ee values. Probably 7 and 8 are
formed by DKR from the (E)- and (Z)-Ti enolates (of low
ee value), which interconvert readily (Scheme 6).
Comparison of entries 5 and 6 in Table 1 (35 and 50%
conversion, respectively) shows that the formation of cis-
fused aldol 7 is more pronounced at low conversion. The
proportional increase of aldol product 8 in the course of the
reaction is certainly due to an equilibration through enoliza-
tion[19] and not (or only in part) to a retro-aldol pathway, as
the enantiomeric excess of 8 increases from 18 to 24%.
tries 8 and 9). However, the
commercial
(À)-N,N-dibutyl-
norephedrine ((À)-32) showed
good reactivity, even at À208C,
and
a high enantioselectivity
(21%;
70% ee;
Table 1;
entry 11); one recrystallization
from acetone/heptanes afforded
enantiopure
(À)-7
(10%;
Scheme 6. Postulated reaction course for the formation of enantio-enriched 7 and 8: reversible Ti enolate for-
mation followed by DKR.
ꢀ98% ee). In spite of the un-
favorable diastereoselectivity,
this result is outstanding and
bears comparison with pub-
In another series of experiments we replaced TMEDA by
lished transannular reactions that, probably due to the ring
sizes, only afforded the cis-diastereomers (see above). More-
over, the readily separated, undesired diastereomer 8 under-
goes rapid and quantitative retro-aldol reaction in the pres-
ence of NaOMe, and 4 can be recycled without any loss.
This new enantioselective aldol reaction is probably stoi-
chiometric, but using 0.8 molar equivalents of Ti reagent
(Table 1; entries 3–11) proved to be much more effective
than 2.4 equiv (Table 1; entries 1 and 2), because unreacted
4 is readily recovered. The amount of Ti complex can be re-
duced to 0.4 equiv, but the conversion is somewhat lower
(Table 1; entries 12 and 13).[17] Traces of water (added or
present in aged TMEDA) increased the rate of the reaction.
The complex formed from (À)-ephedrine ((À)-34) and
TiCl4 does not catalyze the aldol reaction: presumably the
resulting complex lacks electrophilic character due to the
(À)-23 (Table 2). Several differences were noticed, most re-
markably, the diastereoselectivity in favor of the cis-fused
aldol 7 was improved.[20] Additionally, the ee value of 7 was
always lower and the ee value of 8 was equal or higher. The
two experiments performed with (À)-27 and (+)-27
(Table 2; entries 1 and 2) show that there is no marked
matched or mismatched combination. When using (À)-23 as
the base, traces of water inhibited the aldol reaction.
Table 2. Enantioselective aldol reaction of 4 in the presence of (À)-23.
Entry Amino al-
Conversion Yield of 7 [%] (ee Yield of 8 [%]
cohol
[%]
[%])
(ee [%])
1[a]
2[a]
3[b]
4[c]
5[d]
(À)-27
(+)-27
(+)-28
(+)-28
(À)-32
ꢁ40
ꢁ40
ꢁ25
ꢁ40
ꢁ30
ꢁ25 (34)
ꢁ26 (36)
10 (56)
11 (52)
14 (45)
ꢁ9 (n.d.)[e]
ꢁ10 (23)
11 (23)
23 (31)
8 (32)
À
Ti N bond. On the other hand, the complex formed from
(À)-34·HCl and TiCl4 allowed a rapid aldol addition, afford-
ing 8 and minor amounts of 7, both with low ee values.
When the Li alkoxide derived from (+)-27 ((+)-27+
BuLi) was treated with TiCl4, complex 35 was formed, which
did not promote the desired aldol addition. Here the dimin-
ished electrophilicity of the complex is due to the replace-
Reaction conditions: [a] Amino alcohol/TiCl4 (0.80 equiv), (À)-23
(1.20 equiv), NMP, 08C, 3 h. [b] Amino alcohol/TiCl4 (0.80 equiv), (À)-23
(0.93 equiv), NMP,
0 to 208C over 4.5 h. [c] Amino alcohol/TiCl4
(0.80 equiv), (À)-23 (1.20 equiv), NMP, 08C (4 h), then 208C (3 h).
[d] Amino alcohol/TiCl4 (0.80 equiv), (À)-23 (0.93 equiv), NMP, 08C, 4 h.
[e] n.d.=none detected.
2490
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2487 – 2495

Enantioselective Aldol and Grob Fragmentation
FULL PAPER
Stereocontrolled access to (+)-(R,Z)-5-muscenone ((R)-1)
and (À)-(R,E)-5-muscenone((R)-2) by Grob fragmentation:
We presumed that hydroxy ketone (À)-7 (ꢀ98% ee) would
be ideally suited for the synthesis of enantiopure (R)-1 by
OH-directed reduction to diol 36, followed by a Grob frag-
mentation[21] of tosylate 37 (Scheme 7). Indeed, compound
37 possesses the required configuration for the formation of
(R)-1, since the bonds to be broken are in an antiperiplanar
sired trans-diol 36 in high yield (99% crude) and with excel-
lent diastereoselectivity (ꢀ98:2). Tosylate 37 (pyridine, p-
toluenesulfonyl chloride (TsCl)) was then submitted to frag-
mentation (KOtBu, tBuOH, 30 min), affording (+)-1 (ꢀ
98% Z) (67% from (À)-7, ꢀ98% ee).
Alternatively, (À)-7 (ꢀ98% ee) was reduced with lithium
tri-sec-butylborohydride (l-Selectride) in THF at À658C.[24]
Exclusive equatorial hydride addition afforded cis-diol 38
(100% crude). Tosylation (BuLi, TsCl; 100% crude) and
fragmentation of tosylate 39 (KOtBu, tBuOH, 15 h) afford-
ed (R)-2 (ꢀ98% E) (76% from (À)-7; ꢀ98% ee).
À
arrangement and the vicinal C C bonds of the future olefin
are synclinal to each other. Likewise, equatorial reduction
of (À)-7, followed by tosylation and Grob fragmentation
would give access to enantiopure (R)-2.[22] Here, the corre-
sponding tosylate 39 has to undergo a conformational
change to adopt the antiperiplanar arrangement needed for
the fragmentation to (R)-2.
We next applied the same reduction/fragmentation reac-
tions on (Æ)-8. The reductions of (Æ)-8 using either
NMe4BHACTHNUGTRNE(UNG OAc)3 or l-Selectride afforded diols 40 and 42, re-
spectively, with excellent yields and selectivities. Whereas
tosylate 41 underwent smooth fragmentation to afford 2 (ꢀ
98% E) as expected (76% from 40), tosylate 43 underwent
an unprecedented OH-assisted 1,2-hydride migration/elimi-
nation to afford epoxide 44 (34%; tentative assignment of
configuration)[25] and through a minor pathway, a syn-frag-
In principle, hydroxy ketone 8 could also be transformed
into two diasteromeric 1,3-diols, 40 and 42, respectively, and
hence into the target fragmentation products 1 and 2
(Scheme 7). Here, trans-diol 40 should lead to 2, and cis-diol
42 to 1. Whereas trans-hydroxy tosylate 41 is ideally suited
for fragmentation, cis-hydroxy tosylate 43 has to undergo a
conformational flip, which would force the trans-11-ring to
adopt a 1,2-diaxial orientation. Models and calculations
seem to indicate that such a conformational change is unfav-
orable, though not impossible.
mentation, produced (E)-macrocyclic ketone
2 (15%).
Indeed, calculations demonstrate that the trans-diaxial con-
former is highly disfavored with respect to the trans-diequa-
À
torial conformer due to the three axial C C bonds (versus
À
three equatorial C C-bonds). However, the ring strain in
In the event, hydroxy ketone (À)-7 (ꢀ98% ee) was re-
the 11-membered ring is approximately the same in both
conformations. On the other hand, the analogous conforma-
tional change of 39 (cis-fused
duced with NMe4BHACTHNUTRGNEUNG
(OAc)3 in AcOH[23] to afford the de-
bicycle) is energetically less dis-
À
favored (1 versus 2 axial C C
bonds).[26]
Conclusion
We have explored the intramo-
lecular, enantioselective aldol
reaction of muscodione 4, in
which both carbonyl functional-
ities are keto groups and have
found that the highly electro-
philic Ti catalysts, formed from
TiCl4 and an amino alcohol,
were able to promote these re-
actions with appreciable enan-
tioselectivity (up to 70% ee).
Enantiopure hydroxy ketone
(À)-7 (ꢀ98% ee) was obtained
by one recrystallization. Com-
pound (À)-7 was efficiently
transformed into (R)-1 (ꢀ98%
ee) by an internally directed
diastereoselective reduction, to-
Scheme 7. Synthesis of (R)-1 and (R)-2 from (À)-7 and synthesis of (Æ)-2 from (Æ)-8. Formation of epoxide
(OAc)3 (1.3 equiv), AcOH, 24 h; b) l-Selec-
tride (2.0 equiv), À608C, 20–30 min, then H2O2, 5% aqueous NaOH; c) TsCl (2.0–2.1 equiv), pyridine, 58C,
15 h; d) BuLi (1.0–1.5 equiv), 408C, 15 min; then TsCl (1.17–1.50 equiv), À208C, 15 min; e) KOtBu (3.0 equiv),
tBuOH, 25–358C, 30 min.; f) KOtBu (9.0–10.5 equiv), tBuOH, 25–358C, 2–15 h.
sylation and base treatment
(Grob fragmentation). More-
over, we have compared the re-
activity of the four diastereo-
(Æ)-44 by 1,2-hydride shift. Reagents and conditions: a) NMe4BH
ACHTUNGTRENNUNG
Chem. Eur. J. 2010, 16, 2487 – 2495
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2491

C. Fehr et al.
dried over Na2SO4, filtered, concentrated, and purified by bulb-to-bulb
distillation at 1208C and 3 mbar to afford (+)-29 as a white solid (7.78 g,
66%). M.p. 35–378C; [a]2D0 =À5.2 (c=0.96 in CHCl3); 1H NMR
(400 MHz, CDCl3): d=0.83–0.90 (m, 9H), 1.31 (m, 2H), 1.48–1.58 (hept,
J=6.8 Hz, 1H), 2.23 (s, 3H), 2.40–2.55 (m, 4H), 3.90 (brs, 1H), 4.80 (d,
J=4.2 Hz, 1H), 7.20–7.25 (m, 1H), 7.29–7.33 ppm (m, 4H); 13C NMR
(100 MHz, CDCl3): d=142.5 (s), 127.9 (d), 126.8 (d), 126.1 (d), 72.9 (d),
63.5 (d), 53.2 (t), 38.9 (q), 36.4 (t), 26.3 (d), 22.9 (q), 22.7 (q), 10.1 ppm
(q); MS: m/z (%): 235 (1) [M]+, 128 (100), 105 (10), 77 (10), 70 (10), 58
(22), 56 (8), 43 (11).
meric monotosylates 37, 39, 41, and 43. Interestingly, com-
pound 43 cannot adopt the required conformation for anti
fragmentation (trans-antiperiplanar arrangement of the
bonds to be broken) and preferentially undergoes an unpre-
cedented epoxidation/1,2-hydride migration, with concomi-
tant departure of the tosylate to afford epoxide 44. In addi-
tion, unexpectedly, syn fragmentation of 43 is also observed,
leading to 2.
Compound (+)-30: This compound was prepared by following the proce-
dure reported by Saavedra[27] (first part): KHCO3 (5.00 g, 50.0 mmol) was
added to a solution of (+)-ephedrine hydrochloride (10.08 g, 50.0 mmol)
in absolute ethanol (50 mL) at RT, and the resulting mixture was stirred
for 60 min at RT. Pivalaldehyde (6.45 g, 8.23 mL, 75 mmol) was added
dropwise under nitrogen and the resulting mixture was stirred at RT for
2 d. After consumption of (+)-ephedrine, an aqueous solution of 5%
HCl was added dropwise to adjust the solution to pH 1 and most of the
ethanol was evaporated under vacuum. The residue was made basic with
a 5% aqueous solution of NaOH and extracted with Et2O (3ꢃ50 mL).
The combined organic phases were washed with water, saturated aqueous
NaCl, dried over Na2SO4, filtered, and concentrated.
Reduction of the intermediate oxazolidine:[28] The crude oxazolidine
(11.05 g, 47.40 mmol) in dry CH3CN (235 mL) was treated at 08C with
NaBH3CN (5.97 g, 94.80 mmol) in one portion, followed by the addition
of trimethylsilylchloride (TMSCl; 25.7 g, 30.3 mL, 237 mmol) over
15 min. The resulting mixture was allowed to reach RT, then stirred at
RT for 4 h. NaBH3CN (5.97 g, 94.80 mmol) and TMSCl (12.87 g, 15 mL,
118.5 mmol) were added at RT and the reaction was stirred at RT over-
Experimental Section
General: Bulb-to-bulb distillation was performed with a Bꢁchi GKR-51
glass-oven; b.p. corresponds to the oven temperature. TLC was per-
formed on silica gel F-254 plates (Merck); detection with EtOH/anisalde-
hyde/H2SO4 18:1:1. Column chromatography was performed on silica 32–
63, 60 ꢂ (Brunschwig). GC spectra were recorded on a Varian 3500 in-
strument fitted with one of the following capillary columns: DB1 30 W
(15 mꢃ0.319 mm), DB-WAX 15W (15 mꢃ0.32 mm) or the chiral capilla-
ry column: CP-Chirasil-DEX CB (25 mꢃ0.25 mm) (Chrompack); carrier
gas He at 0.63 bar. Optical rotations were determined by using a 1 mL
cell and a Perkin–Elmer 241 polarimeter; c in g/100 mL solution. 1H and
13C NMR spectra were recorded on a Bruker DPX 400 instrument (1H
decoupling frequency at 2.5 ppm). Mass spectra were recorded on a Hew-
lett Packard MSD 5972 automated GC–MS instrument, electron energy
70 eV.
Preparation of the amino alcohols
night, before being quenched with
a saturated aqueous solution of
Compounds (+)-27 and (À)-27:[14] These compounds were prepared by
following the procedure reported by Saavedra[27] or in an autoclave as
follows: (+)-34 (120 g, 0.727 mol), 5% Pd/C (9.0 g), acetone (379 mL),
H2, 70 bar, 508C, 24 h. Yield of (+)-27 after filtration, concentration and
distillation: 135–143 g (90–95%).
K2CO3 and stirred at RT for 1 h. CH3CN was evaporated under vacuum,
and the residue was partitioned between water and Et2O. The organic
layer was separated, and the aqueous phase was extracted with Et2O (3ꢃ
40 mL). The combined organic phases were washed with water, a saturat-
ed aqueous solution of NaCl, dried over Na2SO4, filtered, and concentrat-
ed to give an orange colored oil. Flash column chromatography on silica
gel (cyclohexane/AcOEt, 8:2 to 1:1) and bulb-to-bulb distillation (1308C/
3 mbar) afforded (+)-30 as a pale yellow oil (1.5 g, 13%). [a]2D0 =+5.1
(c=0.98 in CHCl3); 1H NMR (400 MHz, CDCl3): d=0.83 (s, 9H), 0.94
(d, J=6.70 Hz, 3H), 2.19–2.30 (m, 2H), 2.27 (s, 3H), 2.76 (m, 1H), 3.20
(brs, 1H), 4.78 (d, J=4.8 Hz, 1H), 7.20–7.35 ppm (m, 5H); 13C NMR
(100 MHz, CDCl3): d=143.2 (s), 127.9 (d), 126.8 (d), 126.2 (d), 74.5 (d),
69.3 (t), 66.5 (d), 40.9 (q), 33.1 (s), 28.2 (q), 10.1 ppm (q); MS (GC–MS):
m/z (%): 235 (1) [M]+, 178 (10), 128 (100), 105 (5), 77 (10), 71 (10), 58
(40), 56 (8), 44 (11).
Accordingly, (À)-34 afforded (À)-27 with the same yield.
Compound (+)-28: This compound was prepared by following the proce-
dure reported by Saavedra:[27] Isobutyraldehyde (5.190 g; 6.57 mL,
72.09 mmol) was added dropwise to a solution of freshly distilled (+)-34
(7.93 g; 48.06 mmol) in absolute ethanol (50 mL) at RT, under nitrogen,
and the resulting mixture was stirred at RT for 3 h. After consumption of
(+)-34, NaBH4 (3.65 g, 96.1 mmol) was added and the reaction mixture
was stirred at RT over the weekend. A solution of 5% HCl was added
dropwise to adjust the pH to approximately 1, then most of the ethanol
was evaporated under vacuum. The residue was made basic with a 5%
aqueous solution of NaOH and extracted with Et2O (3ꢃ50 mL). The
combined organic layers were washed with saturated aqueous NaCl,
dried over Na2SO4, filtered, concentrated, and distilled by bulb-to-bulb
distillation at 1408C and 3 mbar to afford (+)-28 as a colorless oil (7.40 g,
Compound (+)-31: This compound was prepared by following the proce-
dure reported by Saavedra:[27] KHCO3 (2.60 g, 26 mmol) was added to a
solution of (+)-ephedrine hydrochloride (5.22 g, 26.0 mmol) in absolute
ethanol (30 mL) at RT, and the mixture was stirred for 30 min. 3,3-Dime-
thylbutanal (3.90 g, 4.90 mL, 39.0 mmol) was added dropwise under nitro-
gen and the resulting mixture was stirred at RT for 3 h. After consump-
tion of (+)-ephedrine, NaBH4 (1.976 g, 52 mmol) was added and the re-
action mixture was stirred at RT overnight then for 8 h at 558C. A solu-
tion of 5% HCl was added dropwise to adjust the solution to pH 1 and
most of the ethanol was evaporated under vacuum. The residue was
made basic with a 5% aqueous solution of NaOH and extracted with
CH2Cl2 (3ꢃ50 mL). The combined organic layers were washed with a sa-
turated aqueous solutions of NaCl, dried over Na2SO4, filtered, concen-
trated, and purified by bulb-to-bulb distillation at 1108C and 3 mbar to
afford (+)-31 as a white solid (6.16 g, 95%). M.p. 66–698C; [a]2D0 =+14.9
(c=1.00 in CHCl3); 1H NMR (400 MHz, CDCl3): d=0.88 (s, 9H), 1.35–
1.42 (m, 1H), 2.26 (s, 3H), 2.41–2.55 (m, 2H), 2.79–2.86 (m, 1H), 3.90
(brs, 1H), 4.80 (d, J=4.4 Hz, 1H), 7.20–7.24 (m, 1H), 7.30–7.32 ppm (m,
4H); 13C NMR (100 MHz, CDCl3): d=142.4 (s), 127.9 (d), 126.8 (d),
126.1 (d), 72.7 (t), 63.2 (d), 50.7 (t), 40.6 (t), 39.3 (q), 29.8 (s), 29.5 (q),
10.1 ppm (q); MS: m/z (%): 178 (10), 142 (100), 77 (8), 70 (10), 58 (15),
55 (10), 43 (10).
1
70%). [a]2D0 =+12.9 (c=1.07 in CHCl3); H NMR (400 MHz, CDCl3): d=
0.85–0.89 (m, 9H), 1.75 (hept, J=6.8 Hz, 1H), 2.17 (s, 3H), 2.18–2.27 (m,
4H), 3.73 (brs, 1H), 4.78 (d, J=4.5 Hz, 1H), 7.20–7.25 (m, 1H), 7.29–
7.33 ppm (m, 4H); 13C NMR (100 MHz, CDCl3): d=142.4 (s), 127.9 (d),
126.8 (d), 126.2 (d), 73.2 (d), 64.4 (d), 63.8 (t), 38.6 (q), 26.5 (d), 20.7 (q),
20.5 (q), 10.1 ppm (q); MS: m/z (%): 221 (1) [M]+, 114 (100), 105 (10),
77 (10), 70 (10), 58(30), 42 (8).
Compound (+)-29: This compound was prepared by following the proce-
dure reported by Saavedra:[27] KHCO3 (5.00 g, 50.0 mmol) was added to
a solution of (+)-ephedrine hydrochloride (10.08 g, 50.0 mmol) in abso-
lute ethanol (50 mL) at RT and the mixture was stirred for 30 min. 3-
Methyl butanal (6.45 g, 8.00 mL, 75.0 mmol) was added dropwise under
nitrogen and the resulting mixture was stirred at RT for 1 h. After con-
sumption of ephedrine, NaBH4 (3.80 g, 100 mmol) was added and the re-
action mixture was stirred at RT over the weekend. A solution of 5%
HCl was added dropwise to adjust the pH to 1 and most of the ethanol
was evaporated under vacuum. The residue was made basic with a 5%
aqueous solution of NaOH and extracted with Et2O (3ꢃ50 mL). The
combined organic layers were washed with saturated aqueous NaCl,
2492
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2487 – 2495

Enantioselective Aldol and Grob Fragmentation
FULL PAPER
Compound (À)-32:[29] This compound was prepared by following the pro-
cedure previously reported[29] or purchased from Aldrich. [a]2D0 =À24.7
(c=1.21 in CHCl3); (lit.:[29] [a]2D2 =À24.4 (c=2.00 in hexane)).
NMR spectra of 8: 1H NMR (400 MHz, CDCl3): d=1.02 (d, J=6.5 Hz,
3H), 1.08 (m, 1H), 1.20–1.73 (m, 18H), 1.85 (m, 1H), 1.96 (t, J=13 Hz,
1H), 2.14 (m, 1H), 2.36 (m, 1H), 2.41 ppm (m, 2H); 13C NMR
(100 MHz, CDCl3): d=211.0 (s), 79.7 (s), 55.7 (d), 49.9 (t), 45.0 (t), 39.7
(t), 28.7 (d), 26.9 (t), 26.4 (t), 26.2 (t), 26.0 (t), 25.2 (t), 25.0 (t), 22.2 (t),
22.1 (q), 19.3 ppm (t).
Compound (+)-32:[29] This compound was prepared by following the pro-
cedure previously reported[29] or purchased from Aldrich. [a]2D0 =+26.7
(c=1.82 in CHCl3); (lit.:[29] [a]2D5 =+24.4 (c=2.05 in hexane)).
Compound (+)-36: A solution of (À)-7 (1.54 g, 6.11 mmol; ꢀ98% ee) in
Aldol reactions
AcOH (35 mL) was treated with NMe4BHACTHNUTRGNEUNG(OAc)3 (1.61 g, 6.12 mmol)
Compound (Æ)-8: A solution of muscodione (4) (15.12 g, 60.0 mmol) in
CH2Cl2 (210 mL) was treated at 22–248C with a solution of ZrCl3OPr
(36% in AcOEt) (65.25 g, 91.6 mmol). After 5 min, the yellowish solu-
tion was treated at À10–08C with NBu3 (19.43 g, 25.0 mL, 105 mmol).
After 15 min, the reaction mixture was poured into water and extracted
with Et2O. The organic phases were washed with H2O, saturated aqueous
NaHCO3, and saturated aqueous NaCl, dried over Na2SO4, and evaporat-
ed. Recrystallization from heptane (345 mL) afforded 8 (11.93 g, 79%)
and 2.80 g of mother liquors, from which 8 (1.39 g, 9%) was recovered.
under cold water cooling. After stirring the resulting solution at RT for
90 min (90% conversion), additional NMe4BH(OAc)3 (561 mg,
AHCTUNGTRENNUNG
2.13 mmol) was added and stirring was continued for 17 h. The solution
was poured into a 30% aqueous solution of NaOH (130 mL) and extract-
ed with EtOAc. The organic phases were washed with H2O and a saturat-
ed aqueous solution of NaCl, dried over Na2SO4, and evaporated to
afford (+)-36 as a white solid (1.58 g, 100%). M.p. 101–1048C; [a]D20
=
+24.6 (c=0.79 in CHCl3); 1H NMR (400 MHz, CDCl3): d=0.89 (d, J=
6.5 Hz, 3H), 0.92 (m, 1H), 1.03 (“t”, J=13 Hz, 1H), 1.08 (“q”, J=12 Hz,
1H), 1.18–1.82 (m, 21H), 1.82–1.93 (m, 2H), 4.24 ppm (ddd, J=12, 4.5,
4.5 Hz, 1H); 13C NMR (100 MHz, CDCl3): d=77.4 (s), 70.0 (d), 48.7 (d),
43.5 (t), 39.1 (t), 38.2 (t), 28.0 (t), 27.0 (t), 26.4 (t), 25.8 (d), 25.3 (t), 23.9
(t), 23.5 (t), 22.7 (t), 21.9 (q), 18.7 ppm (t); MS: m/z (%): 236 [M+À18]
(24), 218 (29), 175 (13), 161 (21), 147 (30), 137 (32), 133 (33), 119 (37),
111 (83), 91 (83), 81 (83), 55 (94), 41 (100).
Compounds (À)-7 and (À)-8
Preparation of the (À)-32·TiCl4 (Table 1, entry 13 (large-scale experiment):
A solution of TiCl4 (1m in CH2Cl2; 54.8 mL, 54.8 mmol) was added to a
solution of (À)-32 (14.45 g, 54.8 mmol) in CH2Cl2 (55 mL) over 5 min,
under nitrogen. The temperature rose to 408C. The resulting brown mix-
ture was stirred for 15 min, concentrated at 408C under N2, and dried
under vacuum (6 mbar) for 2 h. All other Ti complexes were prepared ac-
cordingly.
Tosylate 37: A solution of crude 36 (474 mg, max. 1.87 mmol) in pyridine
(1.9 mL) was treated with TsCl (355 mg, 1.84 mmol) under ice cooling.
The reaction mixture was stirred at 08C for 2 h. Further TsCl was added
(355 mg, 1.84 mmol) and the mixture was kept in the refrigerator (58C)
overnight. The suspension was treated with AcOEt and water and the
phases were separated. The organic phases were washed with 5% HCl,
H2O, and a saturated aqueous solution of NaCl, dried over Na2SO4, and
evaporated. Tosylate 37 (923 mg, 100%) was used without purification.
1H NMR (400 MHz, [D8]THF): characteristic signals: d=0.86 (d, J=
6.5 Hz, 3H), 2.44 (s, 3H), 4.66 ppm (ddd, J=12, 5, 5 Hz, 1H); 13C NMR
(100 MHz, [D8]THF): characteristic signals: d=144.8 (s), 136.7 (s), 130.4
(2d), 128.5 (2d), 83.5 (d), 76.4 (s), 47.1 (d), 44.0 (t), 27.0 (2t), 26.5 (d),
23.5 (t), 22.0 (q), 21.5 (q), 19.6 ppm (t).
Aldol reaction (Table 1, entry 13 (large-scale experiment using 0.4 equiv of
(À)-32·TiCl4): The above brown complex (À)-32·TiCl4 (0.4equiv) was dis-
solved in NMP (137 mL) under N2 at RT. The temperature rose to 318C.
After 30 min 4 (34.6 g, 137.1 mmol) was added under stirring. After
30 min the dark solution was treated at 0–58C with
a solution of
TMEDA (7.39 g, 9.60 mL, 63.8 mmol) and H2O (123 mL, 6.85 mmol). The
turbid reaction mixture was kept in the freezer (À20–À188C) for 3 d,
before being poured onto a 5% HCl/ice mixture and extracted with Et2O
(3ꢃ100 mL). The combined organic layers were washed with water, satu-
rated aqueous NaCl, dried over Na2SO4, filtered, and concentrated
(34.6 g). The acidic aqueous phase was basified with a 5% aqueous solu-
tion of NaOH (300 mL) and extracted with Et2O (3ꢃ100 mL). The com-
bined organic layers were washed with water, saturated aqueous NaCl,
dried over Na2SO4, concentrated, and bulb-to-bulb distilled to afford
(À)-32 (12.86 g, 89%). The remaining heavy oil (34.2 g) was purified by
flash chromatography on silica gel (800 g) and cyclohexane/Et2O 1:1 to
Compound (+)-(R)-1:
A suspension of crude 37 (878 mg, max.
1.78 mmol) in tert-butanol (18 mL) was treated at 258C with KOtBu
(598 mg, 5.34 mmol). The reaction mixture was stirred at 25–358C for
30 min, poured into a 5% aqueous solution of HCl and extracted with
EtOAc. The organic phases were washed with H2O, a saturated aqueous
solution of NaHCO3, saturated aqueous NaCl, dried over Na2SO4, and
evaporated. Bulb-to-bulb distillation (100–1308C/0.02 mbar) afforded
(+)-(R)-1 (314 mg; 90% pure; >98% isomeric purity; 67% from (À)-7,
ꢀ98% ee), identical to an authentic sample.[1] [a]D20 =+11.6 (c=1.12 in
MeOH); (lit.:[2] [a]D20 =+11.7 (c=2.45 in MeOH). The ee value was deter-
mined by chiral GC of the reduced LiAlH4 product.
afford, successively, unreacted
4
(22.9 g, 66%), (À)-8 (4.20 g (12%,
38% ee), a mixture of 7 and 8 (537 mg, 1.5%) and (À)-7 (3.81 g, 11%;
68% ee). Recrystallization of (À)-7 from heptane/2% acetone (99 mL)
afforded enantiopure (À)-7 (1.94 g, 6%; ꢀ98% ee). M.p. 162–1648C;
[a]2D0 =À39 (c=0.5 in CHCl3).
(À)-8 (38% ee). [a]2D0 =À14.9 (c=1.46 in CHCl3) (extrapolation for enan-
tiomerically pure (À)-8: [a]2D0 =À39).
Compound (+)-38:
A
cooled (À708C) solution of (À)-7 (500 mg,
For the determination of the ee values of 7 and 8, the crude mixture con-
taining 7 and 8 (250 mg) in THF (5 mL) was treated with TMSOTf
(290 mL) and NEt3 (210 mL) (30 min, 08C) and the corresponding silyl
ethers were extracted with NH4Cl/Et2O and injected on the chiral GC
column. Order of elution: (+)-8, (À)-8, (+)-7, and (À)-7. The same ee
values were found after chromatography.
1.98 mmol; ꢀ98% ee) in THF (15 mL) was treated with l-Selectride (1m
in THF; 3.95 mL, 3.95 mmol). After stirring the resulting solution at
À608C for 20 min, the mixture was poured into a 5% aqueous solution
of NaOH (50 mL). The reaction flask was rinsed with AcOEt. The two-
phase system was treated with 35% H2O2 (2.8 mL), stirred for 90 min,
and extracted with EtOAc. The organic phases were washed with a satu-
rated aqueous solution of NaCl, a 10% aqueous solution of Na2SO3, and
a saturated aqueous solution of NaCl, dried over Na2SO4, and evaporated
to afford a pale oil (547 mg, 100%). [a]2D0 =+23.7 (c=0.75 in CHCl3);
1H NMR (400 MHz, CDCl3): d=0.91 (d, J=6.5 Hz, 3H), 1.00–1.65 (m,
23H), 1.73 (brd, J=14 Hz, 1H), 1.92 (brd, J=8 Hz, 1H), 2.15 (m, 1H),
3.93 ppm (d, J=2.5 Hz, 1H); 13C NMR (100 MHz, CDCl3): d=77.5 (s),
72.1 (d), 46.1 (d), 44.6 (t), 39.4 (t), 37.0 (t), 29.7 (t), 26.9 (t), 26.2 (t), 25.7
(t), 25.3 (t), 24.9 (t), 24.0 (t), 22.1 (q), 21.6 (d), 18.6 ppm (t); MS: m/z
(%): 236 (33) [MÀ18]+, 221 (13), 218 (20), 178 (14), 161 (15), 147 (22),
137 (28), 111 (100), 95 (61), 81 (75), 67 (61), 55 (75), 41 (66).
Aldol reaction (Table 1, entry 11 (using 0.8 equiv of (À)-32·TiCl4): The
complex (À)-32·TiCl4, prepared from (À)-32 (421 mg, 1.60 mmol) and
TiCl4 (1m in CH2Cl2; 1.60 mL, 1.60 mmol) was dissolved in NMP (2 mL)
and treated as above with 4 (504 mg, 2.00 mmol), TMEDA (216 mg,
0.28 mL, 1.86 mmol), and H2O (4 mL, 0.2 mmol) and kept in the freezer
for 96 h. Chromatographic purification as above afforded
4 (70 mg,
14%), (À)-8 (261 mg (52%; 17% ee), and (À)-7 (107 mg, 21%; 70% ee).
Recrystallization of (À)-7 as above afforded enantiopure (À)-7 (51 mg,
10%; ꢀ98% ee).
NMR spectra of 7: 1H NMR (400 MHz, CDCl3): d=1.01 (d, J=6.5 Hz,
3H), 1.20–1.68 (m, 20H), 1.74 (m, 1H), 2.18–2.45 ppm (m, 4H);
13C NMR (100 MHz, CDCl3): d=214.2 (s), 78.3 (s), 58.67 (d), 44.8 (t),
42.5 (t), 37.8 (t), 27.7 (d), 26.8 (t), 26.4 (t), 25.7 (t), 25.5(t), 25.1 (t), 25.0
(t), 24.9 (t), 22.0 (q), 18.6 ppm (t).
Tosylate 39: A solution of crude 38 (540 mg, max. 1.96 mmol) in THF
(20 mL) was treated at 25–418C (exothermic reaction) with BuLi (1.48m
in hexane; 2.0 mL, 2.96 mmol). The solution was stirred at 408C for
15 min, cooled at À208C, and treated at once with TsCl (438 mg,
Chem. Eur. J. 2010, 16, 2487 – 2495
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2493

C. Fehr et al.
2.30 mmol). The reaction mixture (À158C) was stirred for 15 min and
poured into a 5% aqueous solution of HCl. The product was extracted
with AcOEt and the organic phases were washed with a 5% aqueous so-
lution of HCl, H2O, and a saturated aqueous solution of NaCl, dried over
Na2SO4, and evaporated (893 mg, 100%). 1H NMR characteristic signals
(400 MHz, CDCl3): 0.86 (d, J=6.5 Hz, 3H), 4.69 ppm (d, J=2 Hz, 1H);
13C NMR (100 MHz, CDCl3): d=144.9 (s), 134.1 (s), 129.9 (2d), 128.1
(2d), 83.9 (d), 75.5 (s), 44.8 (d), 43.8 (t), 38.6 (t), 34.3 (t), 26.6 (t), 25.7–
24.8 (5t), 23.6 (t), 21.7 (2q), 21.6 (d), 18.3 ppm (t).
with 35% H2O2 (5.6 mL), stirred for 90 min, and extracted with EtOAc.
The organic phases were washed with
a 10% aqueous solution of
Na2SO3, H2O, and a saturated aqueous solution of NaCl, dried over
Na2SO4, and evaporated (986 mg, 98%). Bulb-to-bulb distillation (180–
2108C/0.05 mbar) over CaCO3 (45 mg) afforded 42 (986 mg; 98%).
1H NMR (400 MHz, CDCl3): d=0.93 (d, J=6.5 Hz, 3H), 1.07 (td, J=13,
2.5 Hz, 1H), 1.12–1.68 (m, 19H), 1.75 (m, 1H), 1.86 (m, 1H), 1.95 (m,
1H), 2.17 (m, 1H), 3.07 (br, 1H), 3.43 (br, 1H), 3.98 ppm (brs, 1H);
13C NMR (100 MHz, CDCl3): d=77.4 (s), 70.4 (d), 46.1 (t), 43.5 (d), 42.5
(t), 39.0 (t), 27.1 (t), 26.2 (2t), 25.5 (t), 25.2 (t), 25.0 (t), 22.5 (t), 22.0 (q),
21.6 (d), 21.5 ppm (t); MS: m/z (%): 236 (40) [MÀ18]+, 221 (12), 193
(10), 178 (13), 137 (19), 127 (41), 109 (48), 98 (61), 81 (67), 55 (100), 41
(99).
Compound (À)-(R)-2 from (+)-38: A suspension of crude 39 (438 mg,
max. 0.86 mmol) in tert-butanol (10 mL) was treated with KOtBu
(336 mg, 3.00 mmol) at 258C. The reaction mixture was stirred at 25–
308C for 30 min. As the reaction was very slow, a second and (after an-
other 30 min) a third portion of KOtBu (336 mg, 3.00 mmol) was added.
After 15 h, the mixture was poured into a 5% aqueous solution of
NaOH and extracted with diethyl ether. The organic phases were washed
with H2O and saturated aqueous NaCl, dried over Na2SO4, and evaporat-
ed (307 mg). Bulb-to-bulb distillation (100–1208C/0.02 mbar) afforded
209 mg of (R)-2 (74% pure; >98% isomeric purity; 76%, ꢀ98% ee).
Flash column chromatography on SiO2 (6 g, cyclohexane/AcOEt=98:2)
afforded (R)-2 (152 mg, 75%), identical to an authentic sample.[1] The ee
value was determined by chiral GC of the reduced LiAlH4 product.
[a]D20 =À3.6 (c=0.10 in MeOH); (lit.:[2] [a]2D0 =À3.3 (c=0.06 in MeOH).
Tosylate 43: A solution of 42 (986 mg, 3.88 mmol) in THF (40 mL) was
treated with BuLi (1.48m in hexane; 3.90.0 mL, 3.90 mmol) at 25–418C
(exothermic reaction). The solution was stirred at 408C for 15 min, dilut-
ed with THF (10 mL) (milky aspect unchanged), cooled at À208C, and
treated at once with TsCl (1.11 g, 5.83 mmol). The reaction mixture
(À158C) was stirred for 15 min (clear solution) and poured into a 5%
aqueous solution of HCl. The product was extracted with diethyl ether
and the organic phases were washed with H2O, a saturated aqueous solu-
tion of NaHCO3, and a saturated aqueous solution of NaCl, dried over
Na2SO4, and evaporated. Excess TsCl was removed at 1008C/0.01 mbar.
The residual viscous oil (1.57 g, 99%) was used without purification.
1H NMR characteristic signals (400 MHz, CDCl3): d=0.85 (d, J=6.5 Hz,
3H), 4.93 (brd, J=2 Hz, 1H); 13C NMR (100MHz, CDCl3): d=144.8 (s),
134.1 (s), 129.8 (2d), 127.7 (2d), 83.5 (d), 75.5 (s), 45.5 (t), 43.9 (d), 39.3
(t), 37.6 (t), 27.0 (t), 26.2 (t), 26.0 (t), 25.1 (t), 25.0 (t), 24.7 (t), 22.1 (t),
21.6 (2q), 21.6 (d), 21.4 ppm (t).
Compound (À)-(R)-3 from (À)-(R)-2: Hydrogenation of (À)-(R)-2
(150 mg, 0.636 mmol) over washed Raney Ni in EtOH according to refer-
ence [2] afforded (À)-(R)-3 (146 mg (ꢀ98% ee). [a]2D0 =À12.6 (c=0.48 in
MeOH); (lit.:[2] [a]2D0 =À12.7 (c=0.09 in MeOH).
Compound 40: A solution of (Æ)-8 (2.00 g, 7.94 mmol) in AcOH (45 mL)
was treated with NMe4BHACTHNUTRGNEUNG(OAc)3 under cold water cooling (95% pure;
Compound 44: A suspension of crude 43 (398 mg, 0.975 mmol) in tert-bu-
tanol (10 mL) was treated with KOtBu (336 mg, 3.00 mmol) at 258C. The
reaction mixture was stirred at 25–308C for 30 min. As the reaction was
very slow, a second and (after another hour) a third portion of KOtBu
(328 mg, 2.93 mmol) was added. After 2 h, the mixture was poured into a
5% aqueous solution of HCl and extracted with diethyl ether. The organ-
2.39 g, 8.64 mmol). After stirring the resulting solution at RT for 15 h
(90% conversion), additional NMe4BH(OAc)3 (1.20 g, 4.32 mmol) was
AHCTUNGTRENNUNG
added and stirring continued for 72 h. The solution was poured into 30%
NaOH (150 mL) and extracted with EtOAc. The organic phases were
washed with H2O and a saturated aqueous solution of NaCl, dried over
Na2SO4, and evaporated (1.95 g, 95% pure; 92%). 1H NMR (400 MHz,
CDCl3): d=0.92 (d, J=6.5 Hz, 3H), 0.93 (q, J=11.5 Hz, 1H), 1.16 (dd,
J=14, 13 Hz, 1H), 1.20–1.76 (m, 21H), 1.82–1.98 (m, 3H), 3.68 ppm (td,
J=11, 4 Hz, 1H); 13C NMR (100 MHz, CDCl3): d=76.5 (s), 74.7 (d), 49.2
(d), 45.6 (t), 44.5 (t), 40.0 (t), 27.9 (t), 27.2 (t), 26.5 (t), 26.1 (t), 26.0 (t),
25.7 (d), 25.1 (t), 24.5 (t), 22.0 (q), 21.8 ppm (t); MS: m/z (%): 236 (11)
[M]+-18, 221 (7), 166 (8), 137 (12), 128 (100), 109 (37), 98 (35), 69 (45),
55 (57), 41 (58).
ic phases were washed with H2O,
a saturated aqueous solution of
NaHCO3, and a saturated aqueous solution of NaCl, dried over Na2SO4,
and evaporated (226 mg). The crude material was purified by flash
column chromatography on SiO2 (5 g, cyclohexane/AcOEt=70:30) to
afford a mixture of 44 and 2 (112 mg; 70:30; yield 44/2=34:15%). Treat-
ment of this mixture with excess NaBH4 in MeOH/H2O (5:1) converted 2
into the corresponding alcohol, thus allowing ready purification of 44 by
flash column chromatography on SiO2 (cyclohexane/AcOEt=9:1).
1H NMR (400 MHz, CDCl3): d=0.85 (d, J=6.5 Hz, 3H), 0.90 (m, 1H),
1.25 (s, 1H), 1.22–1.88 (m, 22H), 1.94 ppm (ddd, J=15, 7, 3 Hz, 1H);
13C NMR (100 MHz, CDCl3): d=67.3 (s), 66.5 (s), 38.1 (t), 33.3 (t), 32.3
(t), 30.5 (t), 27.7 (t), 26.4 (2t), 25.4 (d), 23.5 (t), 22.8 (t), 21.9 (t), 21.6 (t),
21.6 (q), 21.4 ppm (t). MS: m/z (%): 236 (17) [M]+, 221 (15), 193 (7), 181
(8), 165 (9), 151 (14), 135 (15), 109 (36), 95 (63), 81 (88), 55 (100), 41
(85).
Tosylate 41: A solution of crude 40 (95% pure; 533 mg, max. 2.00 mmol)
in pyridine (2.1 mL) was treated with TsCl (800 mg, 4.20 mmol) under ice
cooling. The reaction mixture was stirred at 08C for 90 min and kept in
the refrigerator (58C) overnight. The suspension was treated with AcOEt
and water and the phases were separated. The organic phases were
washed with a 5% aqueous solution of HCl, H2O, and a saturated aque-
ous solution of NaCl, dried over Na2SO4, and evaporated (1.02 g).
1H NMR characteristic signals (400 MHz, CDCl3): d=0.86 (d, J=6.5 Hz,
3H), 2.44 (s, 3H), 4.66 ppm (ddd, J=11, 11, 4 Hz, 1H); 13C NMR
(100 MHz, CDCl3): d=144.4 (s), 134.7 (s), 129.6 (2d), 127.7 (2d), 87.7
(d), 76.6 (s), 46.7 (d), 45.0 (t), 41.6 (t), 40.1 (t), 27.0 (2t), 26.1 (t), 25.7 (t),
25.6 (d), 25.3 (2t), 24.7 (t), 22.0 (t), 21.6 (2q), 21.7 (q), 21.6 ppm (q).
Compound
2 from 41: A suspension of crude 41 (967 mg, max.
[1] E. Demole, C. Mahaim, P.-A. Blanc, EP 584477 to Firmenich (prior.
30. July 1992); Chem. Abstr. 1994, 120, 253101.
[2] C. Fehr, J. Galindo, O. Etter, Eur. J. Org. Chem. 2004, 1953.
[3] O. Knopff, J. Kuhne, C. Fehr, Angew. Chem. 2007, 119, 1329;
Angew. Chem. Int. Ed. 2007, 46, 1307.
1.89 mmol) in tert-butanol (20 mL) was treated with KOtBu (668 mg,
5.97 mmol) at 258C. The reaction mixture was stirred at 25–358C for
20 min, poured into a 5% aqueous solution of HCl and extracted with
EtOAc. The organic phases were washed with H2O, a saturated aqueous
solution of NaHCO3, and a saturated aqueous solution of NaCl, dried
over Na2SO4, and evaporated. Bulb-to-bulb distillation (100–1308C/
0.02 mbar) afforded 2 (351 mg; 96% pure; >98% isomeric purity; 76%
from (Æ)-8).
Compound 42: A cooled (À708C) solution of 8 (1.00 g, 3.96 mmol) in
THF (30 mL) was treated with l-Selectride (1m in THF; 7.90 mL,
7.90 mmol). After stirring the resulting solution at À658C for 30 min, the
mixture was poured into a 5% aqueous solution of NaOH (50 mL). The
reaction flask was rinsed with AcOEt. The two-phase system was treated
[4] Review: C. Palomo, M. Oiarbide, J. M. Garcia, Chem. Soc. Rev.
2004, 33, 65.
[5] a) For examples with lanthanide/alkali metal complexes, see: M. Shi-
basaki, N. Yoshikawa, Chem. Rev. 2002, 102, 2187; b) With Zn-eno-
lates: N. Kumagai, S. Matsunaga, T. Kinoshita, S. Harada, S. Okada,
S. Sakamoto, K. Yamaguchi, M. Shibasaki, J. Am. Chem. Soc. 2003,
125, 2169; B. M. Trost, H. Ito, E. R. Shilcoff, J. Am. Chem. Soc.
2001, 123, 3367; c) With Ti enolates: R. Mahrwald, B. Ziemer, Tetra-
hedron Lett. 2002, 43, 4459.
2494
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 2487 – 2495

Enantioselective Aldol and Grob Fragmentation
FULL PAPER
[6] a) M. Shibasaki, H. Sasai, T. Arai, Angew. Chem. Int. Ed. 1997, 36,
1237; b) H. Sasai, T. Suzuki, S. Arai, T. Arai, M. Shibasaki, J. Am.
Chem. Soc. 1992, 114, 4418 (footnote 11).
[17] Likewise, 0.2 molar equivalents of (À)-32·TiCl4 (08C, 9 h) gave a
conversion of approximately 20%.
[18] H. Yamamoto, Tetrahedron 2007, 63, 8377.
[7] M. Inoue, N. Lee, S. Kasuya, T. Sato, M. Hirama, M. Moriyama, Y.
Fukuyama, J. Org. Chem. 2007, 72, 3065.
[19] D. E. Ward, M. Sales, P. K. Sasmal, J. Org. Chem. 2004, 69, 4808.
[20] Subtle changes in the Ti reagent can greatly influence the diastereo-
selectivity, see: M. T. Crimmins, P. J. McDougall, Org. Lett. 2003, 5,
591; V. Rodriguez-Cisterna, C. Villar, P. Romea, F. Urpi, J. Org.
Chem. 2007, 72, 6631; M. B. Hodge, H. F. Olivo, Tetrahedron 2004,
60, 9397.
[21] P. S. Wharton, G. A. Hiegel, J. Org. Chem. 1965, 30, 3254. Review:
D. Caine, Org. Prep. Proced. Int. 1988, 20, 1. Early example: A. Es-
chenmoser, Helv. Chim. Acta 1952, 35, 1660.
[22] For another stereodivergent reduction/fragmentation, in which a 3-
hydroxy ketone is converted into either a Z- or an E-unsaturated
macrocyclic ketone, see: C. Fehr, J. Galindo, O. Etter, W. Thommen,
Angew. Chem. 2002, 114, 4705; Angew. Chem. Int. Ed. 2002, 41,
4523.
[23] D. A. Evans, A. H. Hoveyda, J. Org. Chem. 1990, 55, 5190; D. A.
Evans, D. A. Chapman, E. M. Carreira, J. Am. Chem. Soc. 1988,
110, 3560.
[8] P. I. Dalko, L. Moisan, Angew. Chem. 2004, 116, 5248; Angew.
Chem. Int. Ed. 2004, 43, 5138; B. List, Tetrahedron 2002, 58, 5573;
H. Torii, M. Nakadai, K. Ishihara, S. Saito, H. Yamamoto, Angew.
Chem. 2004, 116, 2017; Angew. Chem. Int. Ed. 2004, 43, 1983 and
references therein.
[9] a) L. Hoang, S. Bahmanyar, K. N. Houk, B. List, J. Am. Chem. Soc.
2002, 124, 10656; Z. G. Hayos, D. R. Parrish, J. Org. Chem. 1974, 39,
1615; U. Eder, G. Sauer, R. Wiechert, Angew. Chem. 1971, 83, 492;
Angew. Chem. Int. Ed. 1971, 10, 496; see also: T. Nagamine, K. Ino-
mata, Y. Endo, L. A. Paquette, J. Org. Chem. 2007, 72, 123; b) J.
Zhou, V. Wakchaure, P. Kraft, B. List, Angew. Chem. 2008, 120,
7768; Angew. Chem. Int. Ed. 2008, 47, 7656; see also: C. Agami, N.
Platzer, H. Sevestre, Bull. Soc. Chim. Fr. 1987, 2, 358; c) C. L. Chan-
dler, B. List, J. Am. Chem. Soc. 2008, 130, 6737.
[10] K. W. Henderson, W. J. Kerr, J. H. Moir, Synlett 2001, 1253.
[11] For aldol reactions of Ti enolates, see: D. A. Evans, D. L. Rieger,
M. T. Bilodeau, F. Urpi, J. Am. Chem. Soc. 1991, 113, 1047; Y.
Tanabe, N. Matsumoto, T. Higashi, T. Misaki, T. Itoh, M. Yamamoto,
K. Mitarai, Y. Nishii, Tetrahedron 2002, 58, 8269 and references
therein.
[24] H. C. Brown, S. Krishnamurthy, J. Am. Chem. Soc. 1972, 94, 7159.
[25] For 1,2-hydride shifts with concomitant departure of bromide, see:
E. L. Eliel, S. H. Wilen, Stereochemistry of Organic Compounds,
Wiley, New York, 1994, p.725.
[26] The energy calculations were performed on 39 (R=Me) and 43
(R=Me) by using the OPLS 2005 force field and applying a stan-
dard conformational search such as that implemented in Macromo-
del (Version 9.7, Schrçdinger, LLC, New York, NY, 2009). Further
DFT calculations were performed on the lower energy conformers
at the B3LYP/6-31G** level of theory by using the Jaguar program
package (Version 7.6, Schrçdinger, LLC, New York, NY, 2009). The
energy differences between the two conformers of 39 and of 43
[12] N. Cabello, J.-C. Kizirian, S. Gille, A. Alexakis, G. Bernardinelli, L.
Pinchard, J.-C. Caille, Eur. J. Org. Chem. 2005, 4835.
[13] Preparation: (À)-27, NaH (1.2 equiv), THF, reflux, 45 min, then MeI
(1.1 equiv), 458C, 90 min (94% yield).
[14] For most recent work, see: C. Fehr, H. Randall, Chem. Biodiversity
2008, 5, 942; C. Fehr, J. Galindo, I. Farris, A. Cuenca, Helv. Chim.
Acta 2004, 87, 1737; Angew. Chem. 1996, 108, 2726; Angew. Chem.
Int. Ed. 1996, 35, 2566.
were shown to be 10.0 and 12.6 kcalmolÀ1
, respectively. These
[15] Some other tested solvents, bases, ligands, catalysts: DMF proved in-
ferior to NMP; THF and acetonitrile gave rise to rapid and com-
plete aldolization, but the only product formed was almost racemic
8. No reaction took place with DMSO, N,N’dimethylpropyleneurea
(DMPU), or CH2Cl2 (+ added NMP). Replacement of TMEDA by
values should be only considered qualitatively, as the reactive spe-
cies are different (OK instead of OH, OTs instead of OMe).
[27] J. Saavedra, J. Org. Chem. 1985, 50, 2271.
[28] P. C. B. Page, H. Heaney, S. Reignier, G. A. Rassias, Synlett 2003, 22.
[29] K. Soai, S. Yokoyama, T. Hayasaka, J. Org. Chem. 1991, 56, 4264.
NBu3 was less efficient and 1,4-diazabicycloACTHNUTRGNEUGN[2.2.2]octane (DABCO)
led to a much slower reaction. The complexes formed from (À)-27
and LaCl3 or ZrCl4 and from (À)-32 and SnCl4 were all ineffective.
[16] J.-C. Kizirian, J.-C. Caille, A. Alexakis, Tetrahedron Lett. 2003, 44,
8893; Q. Perron, A. Alexakis, Tetrahedron: Asymmetry 2007, 18,
2503. We thank Prof. Alexakis for this valuable suggestion.
Received: October 7, 2009
Published online: January 14, 2010
Chem. Eur. J. 2010, 16, 2487 – 2495
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2495