Polyketide building blocks via diastereoselective nitrile oxide cycloadditions with homoallylic alcohols and monoprotected homoallylic diols

Article abstract of DOI:10.1002/chem.200900529

A modular approach to δ²-isoxazolines, latent aldol adducts and polyketide building blocks, is reported. The magnesium-mediated, hydroxyl-directed method allows for the diastereoselective access to a wide variety of masked β-hydroxy ketones, starting from readily available aliphatic and aromatic oximes, homoallylic alcohols and monoprotected homoallylic diols. The utility of the prepared δ²-isoxazolines as polyketide building blocks is demonstrated by their ready conversion into the corresponding β-hydroxy ketones. The anti-diastereoselectivity of the reaction was established by derivatization, NOE studies and comparison of known compounds. A rationale for the observed diastereoselectivity is proposed.

Full text of DOI:10.1002/chem.200900529

FULL PAPER
DOI: 10.1002/chem.200900529
Polyketide Building Blocks via Diastereoselective Nitrile
Oxide Cycloadditions with Homoallylic Alcohols and
Monoprotected Homoallylic Diols
Nina Lohse-Fraefel and Erick M. Carreira*[a]
Abstract: A modular approach to D2-
and monoprotected homoallylic diols.
The utility of the prepared D2-isoxazo-
lines as polyketide building blocks is
demonstrated by their ready conver-
sion into the corresponding b-hydroxy
isoxazolines, latent aldol adducts and
polyketide building blocks, is reported.
The magnesium-mediated, hydroxyl-di-
rected method allows for the diastereo-
selective access to a wide variety of
masked b-hydroxy ketones, starting
from readily available aliphatic and ar-
omatic oximes, homoallylic alcohols
ketones. The anti-diastereoselectivity of
the reaction was established by deriva-
tization, NOE studies and comparison
of known compounds. A rationale for
the observed diastereoselectivity is pro-
posed.
Keywords: allylic alcohols · cyclo-
addition
· nitrile oxide · poly-
ketides · synthetic methods
Introduction
Results and Discussion
Polyketide natural products are attractive synthetic targets
due to their potent biological activity and their stereochemi-
cal complexity. The modular synthesis of relevant building
blocks is a major challenge in organic chemistry, as it would
emulate the biosynthetic strategies.[1] Although the most
widely used strategies for addressing this objective are aldol
additions, allylation, crotylation and allenylmetal reactions,[2]
other methodologies have been reported.[3] We have recent-
ly disclosed an alternative approach to polyketide building
blocks via Kanemasa nitrile oxide cycloadditions[4] with al-
lylic alcohols.[5] Following a single reaction protocol and
using easily accessible optically active oximes and allylic al-
cohols, this method allows for the direct synthesis of an
entire palette of stereochemical permutations found in di-
propionate subunits and represents a novel approach to
Within this article we document in full detail[9] the successful
extension of the magnesium-mediated, hydroxyl-directed ni-
trile oxide cycloaddition methodology from allylic alcohols
to homoallylic alcohols and monoprotected homoallylic
diols. A wide variety of highly functionalized polyketide
building blocks can be synthesized from a readily available
set of starting materials due to the modular nature of this
protocol. The utility of the prepared D2-isoxazolines as
masked aldol adducts is demonstrated by their facile trans-
formation into the corresponding b-hydroxy ketones. The
anti-relationship found in these cycloadducts was confirmed
by conversion of a cycloadduct into a known isoxazoline for
which the relative configuration had been previously estab-
lished. Additionally, derivatization of two additional cyclo-
adducts and subsequent NOE experiments provided further
strong evidence for the proposed anti-relationship. The
requisite starting materials—oximes, homoallylic alcohol
(S)-2-methyl-3-butenol and monoprotected homoallylic al-
cohols—were readily prepared, as described below and in
the Supporting Information.
ACHTUNGTRENNUNGpolyketide building blocks. This magnesium-mediated nitrile
oxide cycloaddition was successfully applied in total synthe-
ses of complex polyketide natural products[6,7] and for the
preparation of stereochemically rich pentaketides.[8]
[a] N. Lohse-Fraefel, Prof. Dr. E. M. Carreira
Laboratorium fꢀr Organische Chemie, ETH Zꢀrich
HCI H335, 8093 Zꢀrich (Switzerland)
Fax : (+41)44-632-1328
Magnesium-mediated nitrile oxide cycloadditions with (S)-2-
methyl-3-butenol: The preliminary prospecting cycloaddi-
tion reaction was performed with 1 equiv of oxime 1 and
2 equiv of homoallylic alcohol (S)-2-methyl-3-butenol[10] in
the presence of 3.3 equiv of iPrOH and 3.0 equiv of EtMgBr
in CH2Cl2, in parallel to the conditions initially developed
E-mail: carreira@org.chem.ethz.ch
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200900529.
Chem. Eur. J. 2009, 15, 12065 – 12081
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
12065

for allylic alcohols.[5a] Chlorination of the oxime to the cor-
responding hydroximinoyl chloride was achieved using
1.0 equiv of tBuOCl. The reaction proceeded in a complete-
ly regioselective manner, in 83% yield and with dr of 7:1 in
favor of the anti-diastereomer.[11] Slow addition of the hy-
droximinoyl chloride to the magnesium alkoxides solution
proved to be advantageous with respect to the observed dia-
stereoselectivity. The optimal time was found to be 1 to
1.5 h; longer addition times did not result in further im-
provement of the diastereoselectivity. Lowering the reaction
temperature, varying the solvent (CH2Cl2, toluene and
hexane)[12] or the addition order of the reactants did not in-
crease the diastereomeric ratio but in some cases resulted in
a decrease in the isolated yield. While the initial reactions
were performed with 2.0 equiv of the homoallylic alcohol, it
was observed that identical results could be obtained with
only 1.3 equiv of the dipolarophile. With these optimized
conditions in hand, different aliphatic as well as aromatic
oximes were examined in the
consider this result as a matched/mismatched case, more de-
tailed studies and experiments would be necessary. Oxime 5,
quaternary at the a-position, underwent cycloaddition in tol-
uene in 89% yield and the diastereoselectivity of this reac-
tion was found to be 4:1 (entry 5). During the course of re-
lated studies in the area of nitrile oxide cycloadditions with
allylic alcohols it was found that in some cases a large
excess of tBuOH instead of 3.3 equiv of iPrOH as additive
in the reaction can dramatically improve the diastereoselec-
tivity.[13] In the case of homoallylic alcohols these conditions
proved unsuccessful; when 25 equiv of tBuOH were used in-
stead of 3.3 equiv of iPrOH both the diastereoselectivity
(2:1 vs. 4:1) and the yield (54% vs. 89%) decreased
(entry 6). Oximes 7[14] and 9, derived from (+)- and (ꢀ)-lac-
tates, respectively, were submitted to standard reaction con-
ditions in toluene (entries 7 and 8). The resulting yields
were similar to those of oximes 1 and 3 (82 and 83% vs. 78
and 83%), while the diastereoselectivities were slightly
magnesium-mediated cycload-
dition with (S)-2-methyl-3-bu-
tenol in the presence of iPrOH
and EtMgBr in either CH2Cl2
or toluene. As shown herein,
the protocol proved successful
for a broad range of aliphatic
and aromatic nitrile oxides.
Unless otherwise stated the ni-
trile oxides were generated in
situ according to standard pro-
cedures using tBuOCl. The hy-
droximinoyl chloride solutions
were generally added over 1 to
1.5 h to the solution of magne-
sium alkoxides at 08C. After
6 h at 08C the reaction mixture
was gradually warmed to RT
over 12 h.
Table 1. Nitrile oxide cycloadditions with (S)-2-methyl-3-butenol.
Entry
Oxime
Solvent
Cycloadduct
dr[a]
Yield [%][b]
1
2
toluene
CH2Cl2
7:1
7:1
78
83
3
4
toluene
CH2Cl2
9:1
9:1
83
85
5
6
4:1[c]
2:1[d]
89
54[d]
toluene
toluene
toluene
The 1,3-dipolar cycloaddi-
tion of the nitrile oxide derived
from oxime
1
and (S)-2-
methyl-3-butenol in toluene af-
forded cycloadduct 2 in a dr of
7:1 and in 78% yield (Table 1,
entry 1). When the reaction
was performed in CH2Cl2, iso-
xazoline 2 was formed with the
same diastereoselectivity and
in slightly better yield (83%,
entry 2). With the enantiomer-
ic oxime 3 under identical re-
action conditions, the cycload-
dition took place in a similar
yield (83% in toluene, 85% in
CH2Cl2) but with slightly
better diastereoselectivity (9:1)
(entries 3 and 4). However, to
7
8
6:1[e]
82
83
5:1[e]
9
toluene
toluene
10:1[e]
87
10
5:1[c]
66[f,g]
[a] Determined by 1H NMR analysis of the crude material. [b] Isolated yield. [c] Determined after filtration
over silica gel. [d] 25 equiv of tBuOH instead of 3.3 equiv of iPrOH. [e] Determined by 13C NMR analysis of
the crude material. [f] NCS used instead of tBuOCl. [g] Isolation of the hydroximinoyl chloride.
12066
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12065 – 12081

Polyketide Building Blocks
FULL PAPER
Table 3. Reaction with oxime 18.
lower (6:1 and 5:1 vs. 7:1 and 9:1). In the reaction with the
nitrile oxide derived from oxime 11,[8] unbranched at the a-
position, cycloadduct 12 was formed in 87% yield and with
a diastereomeric ratio of 10:1 (entry 9). The C-alkynyl ni-
trile oxide 13[15] underwent cycloaddition in toluene in 66%
yield and in 5:1 diastereoselectivity. As previously reported,
chlorination of this substrate was achieved using 1 equiv of
NCS instead of the generally used tBuOCl.[5]
Entry
Solvent
dr[a]
Yield [%][b]
Under the same conditions in toluene, oxime 15, un-
branched at the a-position, exhibited a drastically lower
yield (45%) than the substituted oximes 1 and 3 (78% and
83%), albeit in the same diastereoselectivity range (8:1)
(Table 2, entry 1).[16] In order to improve the yield of the re-
1
2
toluene
CH2Cl2
12:1
11.1
38
36
[a] Determined after derivatization with benzaldehyde, see Equation (6).
[b] Isolated yield.
Table 2. Nitrile oxide cycloadditions with (S)-2-methyl-3-butenol (16).
Table 4. Reaction with hydroximinoyl chloride 20.
Entry
equiv
oxime
equiv
tBuOCl
equiv
16
Solvent
dr[a]
Yield
[%][b]
1
2
3
1.0
1.3
1.0
1.0
1.3
1.3
1.3
1.0
1.3
toluene
toluene
CH2Cl2
8:1
45
52
67
7:1
8:1[c]
Entry
Solvent
dr[a]
Yield [%][b]
[a] Determined by 1H NMR analysis of the crude material. [b] Isolated
yield. [c] Determined after filtration over silica gel.
1
2
toluene
CH2Cl2
8:1
13.1
68
71
[a] Determined after derivatization with benzaldehyde, see Equation (6).
[b] Isolated yield.
action with oxime 15, the reaction in toluene was performed
under slightly different conditions: the oxime was used in
excess and the homoallylic alcohol as the limiting reagent
(entry 2). Under these conditions the yield could be slightly
improved (52% vs. 45%), nonetheless, it remained below a
satisfactory level. However, when the cycloaddition was con-
ducted in CH2Cl2 with the oxime as the limiting reagent and
excess homoallylic alcohol, the yield obtained was 67%, and
the diastereomeric ratio of 8:1 remained unchanged
(entry 3).
1
As a consequence of overlapping peaks in the H NMR as
well as in the 13C NMR spectra of cycloadduct 19, the diaste-
reoselectivity could not be determined from these data.
Condensation of the unpurified material with benzaldehyde,
LiCl and DBU in MeCN afforded alkene 21 as a single
olefin diastereomer [determined by 1H NMR, Eq. (6)], as
expected from the use of
AHCTUNGTREGrNNNU eACHTUGNTRENNNUG
a stabilized phosphonate
The phosphonate in cycloadduct 19 provides an excellent
starting point for further synthetic transformations.[6] When
oxime 18[6] was chlorinated in either toluene or CH2Cl2, and
the in situ prepared hydroximinoyl solution, as generally
done, was added slowly to the magnesium alkoxides solu-
tion, the isolated yields of the corresponding cycloadduct
were less then desirable—38 and 36%, respectively—while
the diastereoselectivity was excellent (12:1 and 11:1)
(Table 3, entries 1 and 2). The hydroximinoyl chloride 20,
derived from oxime 18, can be isolated and purified before
being used in nitrile oxide cycloadditions [Eq. (4)].[6] With
1 equiv of the purified hydroximinoyl chloride and 1.3 equiv
of (S)-2-methyl-3-butenol in either toluene or CH2Cl2, the
corresponding cycloadduct 19 could be isolated in 68 and
71% yield, respectively (Table 4), a significant improvement
compared to the conditions with the in situ generated hy-
droximinoyl chloride, and diastereoselectivity of 8:1 and
12:1, respectively.
agent.[5a]
When standard conditions were applied to the reaction
with benzaldehyde oxime[17] in toluene, the isolated yield of
38% was poor, however, the observed diastereoselectivity
was excellent (14:1) (Table 5, entry 1). We suspected that
part of the problem could be found in the chlorination step.
In a next experiment, the chlorination was carried out in
CH2Cl2 instead of toluene, with 1.1 equiv of tBuOCl instead
of 1 equiv and at 08C instead of ꢀ788C. After 30 min the
hydroximinoyl chloride could be isolated as an off-white
solid. Excess tBuOCl was removed under high vacuum and
the crude product was directly used without further purifica-
Chem. Eur. J. 2009, 15, 12065 – 12081
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12067

E. M. Carreira and N. Lohse-Fraefel
tion in the next step. In the reaction with the isolated hy-
droximinoyl chloride in toluene, the yield could be im-
proved from 38 to 56%, whereas the diastereoselectivity
was found to be slightly lower (10:1 vs. 14:1) (entry 2).
When the cycloaddition reaction with the previously isolated
hydroximinoyl chloride was carried out under otherwise
standard conditions in CH2Cl2, the resulting isoxazoline
could be isolated in 87% yield, while the diastereoselectivity
was found to be 7:1 (entry 3).
afford bis(isoxazoline) 25 in a dr of 5:1. The diastereomers
could easily be separated by flash chromatography. Com-
pound 25 is a masked pentaketide adduct comprised of al-
ternating propionate and acetate subunits. It includes five
stereocenters, is densely functionalized and might serve as a
useful building block for polyketide synthesis. The primary
alcohol, as well as the primary silyl ether, provide the neces-
sary functionality for further chemical transformations. The
two isoxazolines in 25 can potentially be cleaved to the cor-
responding b-hydroxy ketones and subsequently further ela-
borated to the corresponding syn- or anti-diols.[20]
Table 5. Reaction with benzonitrile oxide.
During the course of our studies, magnesium-mediated ni-
trile oxide cycloadditions with homoallylic alcohols have
been communicated.[21] Although the influence of different
substituents on the homoallylic alcohol was examined, these
studies were limited to the use of benzonitrile oxide. Ali-
phatic nitrile oxides, known to be more susceptible towards
dimerization, were not investigated.
Entry
equiv tBuOCl
Solvent
dr[a]
Yield [%][b]
The limitations of the Kanemasa nitrile oxide cycloaddi-
tion with homoallylic alcohols were found to preclude the
use of 1,2-disubstituted and secondary homoallylic alcohols.
In both cases the isolated yields of the desired cycloadducts
were poor due to the diminished reactivity of the dipolaro-
phile. Moreover, in reactions with secondary homoallylic al-
cohols, both syn- and anti-diastereomers were recovered
from the reaction mixture in nearly equal amounts.
1
2
3
1.0
toluene
toluene[d]
CH2Cl2
14:1
10:1
7:1
38
56
87
1.1[c]
1.1[c]
[a] Determined by 1H NMR analysis of the crude material. [b] Isolated
yield. [c] Isolation of the hydroximinoyl chloride. [d] Chlorination in
CH2Cl2.
The primary alcohol in cycloadducts arising from 1,3-dipo-
lar reactions between nitrile oxides and homoallylic alcohols
provides a very useful starting point for further functionali-
zation. The 1,3-relationship generally found in polyketides
is—unlike in products arising from nitrile oxide cycloaddi-
tions with allylic alcohols—already installed after the cyclo-
addition reaction. Accordingly, the primary alcohol in dia-
stereomerically pure cycloadduct 2[18] was conveniently oxi-
dized with the biphasic NaOCl/TEMPO/KBr system[19] to
the corresponding aldehyde 23. Further reaction of the un-
purified aldehyde with hydroxylamine hydrochloride in a
pyridine/EtOH mixture allowed access to oxime 24 in 95%
yield over two steps (Scheme 1). When the 1,3-dipolar cyclo-
addition was performed with diastereomerically pure oxime
24 and 1.3 equiv of (S)-2-methyl-3-butenol under the opti-
mized conditions, the reaction proceeded in 61% yield to
Magnesium-mediated nitrile oxide cycloadditions with
ACHUTNGERNmNUG onoAHCTUNGTRNENpUGN rotected homoallylic diols: Like other isoxazolines
derived from diastereoselective nitrile oxide cycloadditions
with allylic and homoallylic alcohols, cycloadducts 27, aris-
ing from monoprotected homoallylic diols 26, serve as
useful, densely functionalized building blocks for polyketide
synthesis [Eq. (8)].
The monoprotected homoallylic diols were readily pre-
pared as illustrated in Table 6 via a Wittig rearrangement.[22]
For the synthesis of iodomethyltributyltin 29, tributyltinhy-
dride was in a first step alkylated with paraformaldehyde to
afford hydroxymethyltributyltin 28 in 83% yield. The pri-
mary alcohol was readily converted via Appel reaction[23] to
the corresponding iodide in 88% yield (Scheme 2).[24]
The precursors for the Wittig rearrangement were pre-
pared in excellent yields by O-alkylation of 2 equiv of the
monoprotected (Z)-butenediol[25] with iodomethyltributyltin
in a THF/HMPA mixture at RT (Table 6). The presence of
HMPA was crucial to achieve high yields in this reaction.
Wittig rearrangement using 3 equiv of nBuLi at ꢀ788C in
THF smoothly afforded the desired monoprotected homoal-
lylic alcohols. Although Still and co-workers reported that
Scheme 1. Synthesis of a small polyketide fragment.
12068
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12065 – 12081

Polyketide Building Blocks
FULL PAPER
Table 6. Synthesis of monoprotected homoallylic diols.
peaks—by comparison of the differently shifted diastereo-
topic signals of C5 of the isoxazoline in the crude 13C NMR
spectrum after prolonged measurement times.
In a similar fashion the cycloaddition with optically active
oxime 1 as nitrile oxide precursor and different monopro-
tected homoallylic diols was examined. Again, as the results
summarized in Table 8 indicate, the same trends as outlined
above could be observed. Cycloadducts arising from mono-
Entry PG
Product Isolated Product Yield Product Isolated
yield
[%]
[%][a]
yield
[%]
AHCTUNGTREGpNNUN rotected homoallylic diols with bulkier protective groups
1
2
3
4
5
Tr
30a
96
90
87
89
58
31a
31b
31c
31d
31e
86
88
93
89
53
32a
32b
32c
32d
32e
83
63
73
78
43
showed better anti/syn diastereoselectivity than their coun-
terparts derived from dipolarophiles with less sterically de-
manding substituents. As in the above described series of
experiments, the yields of the 1,3-dipolar reactions were all
good (79 to 86%). With trityl ether 32a the cycloaddition
was performed both in CH2Cl2 and toluene under otherwise
identical conditions. By employing the chlorinated solvent
both yield (86 vs. 79%) and diastereoselectivity (25:1 vs.
12:1) were superior compared to the reaction in toluene (en-
tries 1 and 2). The diastereoselectivity of cycloadduct 40 was
TBDPS 30b
TIPS
TBS
TES
30c
30d
30e
Scheme 2. Preparation of 29.
1
determined by H NMR analysis of the mixture. As in simi-
these two reactions can be carried out in a one-pot proce-
dure,[26] we found it to be advantageous to isolate the stan-
nane intermediates.
lar cases with silicon protective groups described above, this
method of analysis could not be applied to the other cyclo-
addition products in Table 8 due to overlapping signals. The
diastereomeric ratios of the cycloadducts given in entries 3
and 4 were determined after protection of the crude primary
alcohols as the corresponding trityl ethers. In these deriva-
tives, the differently shifted diastereotopic signals of the hy-
In an initial series of experiments the cycloaddition be-
tween the nitrile oxide derived from oxime 11 and different-
ly monoprotected homoallylic diols 26 was examined. The
results in Table 7 clearly reveal the dependence of the dia-
stereomeric ratio in the corresponding cycloadduct on the
chosen protective group in the dipolarophile. With sterically
demanding protective groups, such as Tr or TBDPS, excel-
lent diastereoselectivities up to 21:1 could be obtained.
However, with less bulky sub-
1
drogen atom at C5 of the isoxazoline in the H NMR spec-
trum allowed for the determination of the diastereoselectiv-
ity of the previous cycloaddition.[27] For the determination of
both yield and diastereoselectivity of these cycloaddition re-
stituents, such as TBS or TES
ethers, the observed diastereo-
selectivity decreased (10:1,
entry 4, and 6:1, entry 5). The
yields of the reaction were
generally high (73% to 85%)
except for the reaction with
mono-TES-protected diol 32e
(53%) (entry 5). The lower
yield in the reaction with TES
derivative 32e might be due to
the relative instability of this
silyl ether under the reaction
conditions. The diastereoselec-
tivities of nitrile oxide cycload-
ditions with monoprotected
homoallylic diols were general-
ly determined by integration of
the differently shifted diaste-
reotopic signals for the hydro-
gen atom at C5 of the isoxazo-
line in the crude 1H NMR
spectrum or—in cases when
this method could not be ap-
plied due to overlapping
Table 7. Nitrile oxide cycloadditions with monoprotected homoallylic diols.
Entry
1
Homoallylic alcohol
Cycloadduct
dr anti/syn[a]
Yield [%][b]
82
19:1[c]
2
3
4
5
21:1
14:1
10:1
6:1
85
73
78
53
[a] Determined by 13C NMR analysis of the crude material. [b] Isolated yield. [c] Determined by 1H NMR
analysis of the crude material.
Chem. Eur. J. 2009, 15, 12065 – 12081
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12069

E. M. Carreira and N. Lohse-Fraefel
actions, equal amounts of the crude material were used for
derivatization and purification. Cycloadducts arising from
reactions with monoprotected homoallylic alcohols inherit
an anti-relationship (see below). A simple orthogonal pro-
tection–deprotection of the primary alcohols would allow
ready access to the corresponding syn-diastereomers.
ketones 44 and 46 in 94 and 96% yield (Scheme 3).
Confirmation of the relative configuration: The primary al-
cohols in isoxazolines 6 and 4 were protected as p-methoxy-
benzyl ethers under mild conditions (Schemes 3 and 4). Re-
ꢀ
ductive N O bond cleavage and subsequent imine hydroly-
sis, followed by DDQ oxida-
tion, afforded p-methoxyben-
zylidene acetals 48 and 49.
Isoxazolines 6 and 4 were both
submitted to the reaction con-
ditions as a mixture of diaste-
reomers. The same diastereo-
meric ratios were observed
again in the corresponding ace-
tals. NOE analysis with selec-
tive irradiation on the major
Table 8. Nitrile oxide cycloadditions with monoprotected homoallylic diols.
Entry
Homoallylic alcohol
Solvent
Cycloadduct
dr anti/syn[a]
Yield [%][b]
diastereomer revealed an anti-
relationship. Strong NOEs
were observed between all
three hydrogen atoms at C2, C4
and C6, providing strong evi-
dence for this relative configu-
ration as illustrated in Figure 1.
In a second approach to con-
firm the anti-diastereoselectiv-
ity of magnesium-mediated ni-
trile oxide cycloadditions with
homoallylic alcohols, cycload-
duct 12 was converted into
39a+39c/39b+39d
1
2
CH2Cl2
toluene
25:1
12:1
86
79
3
4
CH2Cl2
CH2Cl2
20:1[c]
6:1[c]
83
84
1
[a] Determined by H NMR analysis of the crude material. [b] Isolated yield of combined diastereomeric prod-
ucts. [c] Determined by H NMR analysis of the material after protection of the primary alcohol as trityl ether.
1
known
isoxazoline
52
Isoxazoline reduction to b-hy-
droxyketones: The utility of
cycloadducts arising from reac-
tions with homoallylic alcohols
as latent aldol adducts and po-
lyketide building blocks was
demonstrated by their straight-
forward transformation into
the corresponding b-hydroxy
ꢀ
ketones. The N O bond cleav-
age of the isoxazoline can be
effected by reduction with
Raney nickel and subsequent
hydrolysis of the resulting
imine to afford the corre-
sponding b-hydroxy ketone.[28]
The primary alcohols in cyclo-
adducts 4 and 8 were protected
Scheme 3. Reductive opening of cycloadducts.
as
their
p-methoxybenzyl
ethers under mild conditions.
Reduction with Raney nickel,
accompanied by boric acid, in
a
methanol/water
mixture
under hydrogen atmosphere
lead to the desired b-hydroxy
Scheme 4. Structural determination of relative configuration.
12070
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12065 – 12081

Polyketide Building Blocks
FULL PAPER
hydrogen atom of the syn-diastereomer of the examined b-
methyl ester isoxazolines occurs at slightly higher field than
the corresponding signal of the anti-diastereomer. In
1H NMR spectra of b-hydroxy isoxazolines we found the
major diastereotopic signal for the C5 hydrogen atom shifted
slightly upfield compared to the analogous minor diastereo-
topic signal. As is illustrated in Scheme 5, b-hydroxy isoxa-
zoline 12 was converted to the corresponding methyl ester.
After this transformation, the chemical shifts of the C5 hy-
drogen atoms of the heterocycle had inverted; the signal of
the major diastereomer occurred now more downfield than
the corresponding signal of the minor diastereomer. We
therefore propose an anti-diastereoselectivity of magnesium-
mediated nitrile oxide cycloadditions with homoallylic alco-
hols.
Figure 1. Observation of strong NOEs in p-methoxybenzylidene acetals
48 and 49.
(Scheme 5). Panek et al. had previously established the rela-
tive stereochemistry for this class of compounds by derivati-
zation and subsequent coupling constant analysis.[29] The pri-
mary alcohol in isoxazoline 12 was conveniently oxidized to
the corresponding aldehyde with the biphasic TEMPO/
NaOCl/KBr system.[19] Lindgren oxidation[30] to the carbox-
ylic acid, followed by methylation with TMSCH2N2 afforded
methyl ester 50 in 70% yield over three steps. Deprotection
of the silyl ether with TBAF in THF lead to alcohol 51 in
87% yield. At this point the diastereomers were separated
and the synthesis was continued with exclusively the minor
syn-isomer 51b. Protection of the primary alcohol as benzyl
ether provided isoxazoline 52. The spectral characteristics
(1H and 13C NMR, HRMS) of isoxazoline 52 were in all re-
spects identical to those previously reported, providing fur-
ther strong evidence for the syn-relationship in this minor
diastereomer and therefore for an anti-relationship in the
major diastereomer.
Model for the anti-diastereoselectivity: A rationale for the
observed anti-diastereoselectivity arises from the compari-
son of the possible transition states depicted in Figure 2. In
Figure 2. Proposed transition states leading to anti- or syn-cycloadducts.
analogy to the generally accepted transition state model for
Kanemasa nitrile oxide cycloadditions with allylic alco-
hols,[12] the comparison of the two chelated transitions states
provides a possible explanation for the observed diastereo-
selectivity in 1,3-dipolar reactions between nitrile oxides and
homoallylic alcohols. Transition state B (TS B) suffers from
a somewhat higher energy of formation as a consequence of
1,3-allylic interactions (H $ Me) which are absent in transi-
tion state A (TS A). The reaction is thus suggested to pro-
ceed through a transition-state similar to TS A, resulting in
the observed preference for the formation of the anti-diaste-
reomer. However, the energetic difference between the two
reaction transition states leading to anti- and syn-products
cannot solely arise from the energetic differential arising
from simple analysis of allylic interactions, which can be es-
timated to be at best as 0.5 kcalmolꢀ1. We suggest that the
eclipsing interaction between C-2 Me and C-4 is magnified
as C-4 undergoes rehybridization from Csp2 in the starting
material to Csp3 in the product, wherein the costly interac-
tion can be estimated by comparison of eclipsed gauche
butane (E=4.5–6 kcalmolꢀ1).
Scheme 5. Structural confirmation of relative configuration.
Kociolek and Hongfa assigned the relative configuration
of the products of 1,3-dipolar cycloadditions between nitrile
oxides and homoallylic alcohols as syn.[21] Their assignment
of the relative stereochemistry of b-hydroxy isoxazolines is
based on the comparison of the chemical shifts of the C5 hy-
drogen atom in the 1H NMR spectrum with analogous b-
methyl ester isoxazolines, compounds described by Panek
et al.[29] who demonstrated that the chemical shift of the C5
Chem. Eur. J. 2009, 15, 12065 – 12081
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12071

E. M. Carreira and N. Lohse-Fraefel
solution of (S)-2-methyl-3-butenol (56.0 mg, 0.650 mmol, 1.30 equiv) and
iPrOH (126 mL, 1.65 mmol, 3.30 equiv) in CH2Cl2 (10 mL) at 08C. The re-
action mixture turned momentarily cloudy and became clear and color-
less upon stirring at 08C for 30 min. At this time, the hydroximinoyl chlo-
ride in CH2Cl2 was added to the reaction dropwise via cannula over
60 min, followed by two rinses (1 mL each). The slightly yellow solution
was stirred at 08C for 6 h and was slowly allowed to warm to room tem-
perature overnight (large Dewar with ice, covered with aluminum foil).
After 18 h the reaction was quenched by addition of saturated, aqueous
NH4Cl (15 mL). The mixture was vigorously stirred for 20 min, then the
layers were separated and the aqueous phase was extracted with EtOAc
(3ꢂ15 mL). The combined organic solutions were washed with brine,
dried over anhydrous Na2SO4, filtered, and concentrated under reduced
pressure. Purification by flash chromatography (hexane/EtOAc 4:1) pro-
vided isoxazoline 2 (125 mg, 83%, dr 7:1 by integration of the signals at
4.40 ppm (major) and 4.65 ppm (minor) in the crude 1H NMR spectrum)
as a clear, colorless oil. [a]3D4 (c = 0.50, CHCl3) = +37.38; 1H NMR
Conclusion
We have described diastereoselective magnesium-mediated,
hydroxyl-directed nitrile oxide cycloadditions with homoal-
lylic alcohols and monoprotected homoallylic diols. These
reactions represent an alternative approach to masked poly-
ketide building blocks. While the reaction is anti-diastereo-
selective, the corresponding syn-diastereomers of reactions
with monoprotected homoallylic diols are likewise accessible
by a simple orthogonal protection-deprotection sequence.
The value of the presented nitrile oxide cycloaddition ap-
proach to polyketide building blocks was demonstrated by
the ready transformation of different isoxazolines to the cor-
responding b-hydroxy ketones. The observed anti-diastereo-
selectivity was confirmed by NOE studies as well as by the
conversion of a cycloadduct into a previously described iso-
xazoline, for which the relative configuration had been es-
tablished.
(300 MHz, CDCl3, * denotes minor diastereomeric peak): d
= 4.65*
(ddd, 1H, J=12.5, 8.1, 4.4 Hz), 4.40 (ddd, 1H, J=10.3, 8.4, 8.4 Hz), 3.71–
3.58 (m, 2H), 3.68 (d, 2H, J=6.2 Hz), 3.02 (dd, 1H, J=17.8, 10.9 Hz),
3.00* (dd, 1H, J=17.1, 10.6 Hz), 2.85* (dd, 1H, J=16.8, 8.1 Hz), 2.82–
2.71 (m, 1H), 2.79 (dd, 1H, J=17.1, 8.7 Hz), 2.33 (brs, 1H), 1.95–1.82
(m, 1H), 1.17 (d, 3H, J=7.2 Hz), 1.14* (d, 3H, J=7.1 Hz), 0.95* (d, 3H,
J=7.2 Hz), 0.90 (d, 3H, J=6.8 Hz), 0.90* (s, 9H), 0.89 (s, 9H), 0.06* (s,
6H), 0.05 ppm (s, 6H); 13C NMR (75 MHz, CDCl3, * denotes minor dia-
stereomeric peak): d = 161.8, 84.0, 81.5*, 66.4, 66.2, 65.0*, 40.4, 40.3,
39.7*, 38.2*, 36.0, 26.1, 18.4, 14.8, 13.5, 11.4*, ꢀ5.2 ppm; IR (thin film): n˜
= 3428, 2957, 2930, 2884, 2858, 2740, 2710, 2243, 1621, 1472, 1434, 1390,
1362, 1298, 1257, 1213, 1182, 1097, 1046, 1007, 974, 939, 909, 838, 815,
777, 734, 678, 646 cmꢀ1; elemental analysis calcd (%) for C15H31NO3Si: C
Experimental Section
General methods: All non-aqueous reactions were carried out using
oven-dried (1358C) or flame-dried glassware under a positive pressure of
dry argon unless otherwise stated. Tetrahydrofuran, diethyl ether, tolu-
ene, acetonitrile, and methylene chloride were purified by distillation and
dried by passage over activated alumina under an argon atmosphere
(H2O content <30 ppm, Karl-Fischer titration).[31] Pyridine was distilled
from KOH under an atmosphere of dry nitrogen. n-Butyl lithium was ti-
trated with sBuOH/phenanthroline.[32] All other commercially available
reagents were used without further purification. The reactions were all
magnetically stirred and monitored by thin-layer chromatography using
Merck silica gel 60 F254 plates and visualized using ceric ammonium mo-
lybdate or potassium permanganate stain. Chromatographic purification
of products (flash chromatography) was performed on E. Merck silica gel
60 (230–400 mesh) using a forced flow of eluant at 0.3–0.5 bar.[33] Concen-
tration under reduced pressure was performed by rotary evaporation at
308C at the appropriate pressure, unless otherwise stated. Yields refer to
chromatographically purified and spectroscopically pure compounds,
unless otherwise stated. Melting points were measured on a Bꢀchi 510
apparatus. All melting points were measured in open capillaries and are
uncorrected. Optical rotations were measured on a Jasco DIP-1000 polar-
imeter operating at the sodium D line with a 100 mm path length cell,
and are reported as follows: [a]TD, concentration (g per 100 mL), and sol-
vent. NMR spectra were recorded either on a Varian Mercury 300 spec-
trometer operating at 300 and 75 MHz for 1H and 13C acquisitions, re-
spectively, or on a Bruker DRX500 spectrometer operating at 500 MHz
59.76,
H 10.36, N 4.65; found: C 59.67, H 10.41, N 4.82; HRMS
(MALDI): m/z: calcd for C15H32NO3Si [M+H]+: 302.2146; found:
302.2142.
(R)-2-{(S)-3-[(S)-2-(tert-Butyldimethylsilanyloxy)-1-methylethyl]-4,5-di-
hydroisoxazol-5-yl}-propan-1-ol (4): To a solution of oxime 3 (109 mg,
0.500 mmol, 1.00 equiv) in CH2Cl2 (5 mL) at ꢀ788C was added tBuOCl
(59.0 mL, 0.500 mmol, 1.00 equiv) dropwise over 10 min. The resulting
deep blue solution was stirred 1 h at ꢀ788C. In a separate flask EtMgBr
(3.0m in Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added dropwise to a
solution of (S)-2-methyl-3-butenol (56.0 mg, 0.650 mmol, 1.30 equiv) and
iPrOH (126 mL, 1.65 mmol, 3.30 equiv) in CH2Cl2 (10 mL) at 08C. The re-
action mixture turned momentarily cloudy and became clear and color-
less upon stirring at 08C for 30 min. At this time, the hydroximinoyl chlo-
ride in CH2Cl2 was added to the reaction dropwise via cannula over
60 min, followed by two rinses (1 mL each). The slightly yellow solution
was stirred at 08C for 6 h and was slowly allowed to warm to room tem-
perature overnight (large Dewar with ice, covered with aluminum foil).
After 20 h the reaction was quenched by addition of saturated, aqueous
NH4Cl (15 mL). The mixture was vigorously stirred for 20 min, then the
layers were separated and the aqueous phase was extracted with EtOAc
(3ꢂ15 mL). The combined organic solutions were washed with brine,
dried over anhydrous Na2SO4, filtered, and concentrated under reduced
pressure. Purification by flash chromatography (hexane/EtOAc 4:1) pro-
vided isoxazoline 4 (128 mg, 85%, dr 9:1 by integration of the signals at
4.41 ppm (major) and 4.64 ppm (minor) in the crude 1H NMR spectrum)
as a clear, colorless oil. [a]2D5 (c = 0.98, CHCl3) = +36.08; 1H NMR
(300 MHz, CDCl3): d = 4.41 (ddd, 1H, J=10.3, 8.1, 8.1 Hz), 3.69–3.56
(m, 4H), 3.03 (dd, 1H, J=16.8, 10.3 Hz), 2.84–2.72 (m, 1H), 2.78 (dd,
1H, J=16.8, 7.8 Hz), 2.41 (brs, 1H), 1.94–1.81 (m, 1H), 1.14 (d, 3H, J=
7.2 Hz), 0.88 (d, 3H, J=6.9 Hz), 0.87 (s, 9H), 0.05 ppm (s, 6H);
13C NMR (75 MHz, CDCl3): d = 161.6, 83.7, 66.1, 66.0, 40.2, 39.6, 35.7,
25.9, 18.3, 14.6, 13.4, 5.3, ꢀ5.4 ppm; IR (thin film): n˜ = 3421, 2957, 2930,
2884, 2858, 2739, 2360, 2341, 1622, 1472, 1463, 1389, 1362, 1297, 1257,
1098, 1045, 1007, 972, 939, 897, 838, 815, 782, 754, 678, 668 cmꢀ1; elemen-
tal analysis calcd (%) for C15H31NO3Si: C 59.76, H 10.36, N 4.65; found:
1
for H acquisitions. Chemical shifts (d) are reported in ppm with the sol-
vent resonance as the internal standard relative to chloroform (d 7.26)
for 1H, and (d 77.0) 13C. All 13C spectra were measured with complete
proton decoupling. Data are reported as follows: s=singlet, d=doublet,
t=triplet, q=quartet, m=multiplet; coupling constants in Hz. IR spectra
were recorded on a PerkinElmer Spectrum RXI FT-IR spectrophotome-
ter. Absorptions are given in wavenumbers (cmꢀ1). Mass spectra were re-
corded by the MS service at ETH Zꢀrich. EI-MS (m/z): VG-TRIBRID
spectrometer. MALDI-MS and ESI-MS (m/z): IonSpec Ultima Fourier
Transform Mass Spectrometer. Elemental analyses were performed at
the Mikrolabor der ETH Zꢀrich.
(R)-2-{(S)-3-[(R)-2-(tert-Butyldimethylsilanyloxy)-1-methylethyl]-4,5-di-
hydroisoxazol-5-yl}-propan-1-ol (2): To a solution of oxime 1 (109 mg,
0.500 mmol, 1.00 equiv) in CH2Cl2 (5 mL) at ꢀ788C was added tBuOCl
(59.0 mL, 0.500 mmol, 1.00 equiv) dropwise over 10 min. The resulting
deep blue solution was stirred 1 h at ꢀ788C. In a separate flask EtMgBr
(3.0m in Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added dropwise to a
C
59.58, H 10.16, N 4.85; HRMS (MALDI): m/z: calcd for
C15H31NO3SiNa [M+Na]+: 324.1965; found: 324.1962.
(R)-2-{(S)-3-[2-(tert-Butyldimethylsilanyloxy)-1,1-dimethylethyl]-4,5-di-
hydroisoxazol-5-yl}-propan-1-ol (6): To a solution of oxime 5 (118 mg,
12072
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12065 – 12081

Polyketide Building Blocks
FULL PAPER
0.500 mmol, 1.00 equiv) in toluene (5 mL) at ꢀ788C was added tBuOCl
(59.0 mL, 0.500 mmol, 1.00 equiv) dropwise over 15 min. The solution was
stirred at ꢀ788C for 4 h. After 1 h the solution was still colorless and
turned light blue and cloudy over the course of 4 h. In a separate flask
EtMgBr (3.0m in Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added drop-
3H), 0.07* ppm (s, 3H); 13C NMR (75 MHz, CDCl3, * denotes minor dia-
stereomeric peak): d = 162.1, 83.9, 81.7*, 65.8, 65.1*, 64.7*, 64.5, 40.3,
40.0*, 36.6, 36.0*, 25.9, 22.4, 18.2*, 13.0, 11.3*, ꢀ4.6, ꢀ4.7 ppm; IR (thin
film): n˜ = 3435, 2957, 2931, 2887, 2858, 1627, 1472, 1464, 1408, 1389,
1372, 1362, 1340, 1312, 1294, 1258, 1216, 1116, 1086, 1026, 954, 881, 834,
778, 664 cmꢀ1; elemental analysis calcd (%) for C14H29NO3Si: C 58.49, H
10.17, N 4.87; found: C 58.56, H 10.41, N 5.05; HRMS (MALDI): m/z:
calcd for C14H29NO3SiNa [M+Na]+: 310.1809; found: 310.1813.
wise to
a solution of (S)-2-methyl-3-butenol (56.0 mg, 0.650 mmol,
1.30 equiv) and iPrOH (126 mL, 1.65 mmol, 3.30 equiv) in toluene
(10 mL) at 08C. The reaction mixture was stirred at 08C for 30 min. At
this time, about 2/3 of the cloudy hydroximinoyl chloride mixture was
added to the reaction dropwise via cannula over 60 min. Then the light
blue, turbid hydroximinoyl chloride mixture was allowed to warm to RT
for 1 min, whereupon it turned deep blue and clear. The solution was
cooled again to ꢀ788C and the addition via cannula was continued over
30 min, followed by two rinses (1 mL each). The slightly yellow solution
was stirred at 08C for 6 h and was slowly allowed to warm to room tem-
perature overnight (large Dewar with ice, covered with aluminum foil).
After 14 h the reaction was quenched by addition of saturated, aqueous
NH4Cl (15 mL). The mixture was vigorously stirred for 20 min, then the
layers were separated and the aqueous phase was extracted with EtOAc
(3ꢂ15 mL). The combined organic solutions were washed with brine,
dried over anhydrous Na2SO4, filtered, and concentrated under reduced
pressure. After filtration over silica gel (hexane/EtOAc 3:1) isoxazoline 6
(141 mg, 89%, dr 4:1 by integration of the signals at 4.39 ppm (major)
and 4.64 ppm (minor) in the 1H NMR spectrum) was isolated as a pale,
colorless oil. [a]3D0 (c = 1.08, CHCl3) = +17.18; 1H NMR (300 MHz,
CDCl3, * denotes minor diastereomeric peak): d = 4.64* (ddd, 1H, J=
12.8, 8.4, 4.4 Hz), 4.39 (ddd, 1H, J=10.0, 8.1, 8.1 Hz), 3.74–3.3.56 (m,
2H), 3.51* (s, 2H), 3.50 (s, 2H), 3.05 (dd, 1H, J=16.8, 10.3 Hz), 3.02*
(dd, 1H, J=16.8, 10.6 Hz), 2.85* (dd, 1H, J=16.8, 8.4 Hz), 2.80 (dd, 1H,
J=16.8, 8.4 Hz), 2.37 (dd, 1H, J=7.5, 4.1 Hz), 1.95–1.81 (m, 1H), 1.18 (s,
3H), 1.16 (s, 3H), 0.93* (d, 3H, J=7.2 Hz), 0.89 (s, 9H), 0.88 (d, 3H, J=
7.2 Hz), 0.05 ppm (s, 6H); 13C NMR (75 MHz, CDCl3, * denotes minor
diastereomeric peak): d = 164.6, 164.3*, 83.9, 81.5*, 70.9, 66, 0, 65.1*,
40.2, 39.9*, 39.2, 38.6, 38.2*, 26.0, 23.0, 18.4*, 13.3, 11.3*, ꢀ5.4 ppm; IR
(thin film): n˜ = 3414, 2957, 2931, 2886, 2858, 2246, 1611, 1472, 1392,
1363, 1282, 1254, 1217, 1186, 1099, 1046, 1006, 939, 838, 815, 757,
668 cmꢀ1; elemental analysis calcd (%) for C16H33NO3Si: C 60.91, H
10.54, N 4.44; found: C 61.16, H 10.71, N 4.36; HRMS (MALDI): m/z:
calcd for C16H33NO3SiNa [M+Na]+: 338.2122; found: 338.2121.
(R)-2-{(S)-3-[(R)-1-(tert-Butyldimethylsilanyloxy)-ethyl]-4,5-dihydroisox-
azol-5-yl}-propan-1-ol (10): To
a solution of oxime 9 (102 mg,
0.500 mmol, 1.00 equiv) in toluene (5 mL) at ꢀ788C was added tBuOCl
(59.0 mL, 0.500 mmol, 1.00 equiv) dropwise over 10 min. The mixture
turned yellow, then green and finally blue over the course of 15 min and
was stirred total 105 min ꢀ788C. In a separate flask EtMgBr (3.0m in
Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added dropwise to a solution of
(S)-2-methyl-3-butenol (56.0 mg, 0.650 mmol, 1.30 equiv) and iPrOH
(126 mL, 1.65 mmol, 3.30 equiv) in toluene (10 mL) at 08C. The mixture
was stirred at 08C for 1 h. At this time, the hydroximinoyl chloride in tol-
uene was added to the reaction dropwise via cannula over 2 h, followed
by two rinses (1 mL each). The slightly yellow solution was stirred at 08C
for 6 h and was slowly allowed to warm to room temperature overnight
(large Dewar with ice, covered with aluminum foil). After 18 h the reac-
tion was quenched by addition of saturated, aqueous NH4Cl (15 mL).
The mixture was vigorously stirred for 20 min, then the layers were sepa-
rated and the aqueous phase was extracted with EtOAc (3ꢂ15 mL). The
combined organic solutions were washed with brine, dried over anhy-
drous Na2SO4, filtered, and concentrated under reduced pressure. Purifi-
cation by flash chromatography (hexane/EtOAc 4:1) provided isoxazo-
line 10 (119 mg, 83%, dr 5:1 by integration of the signals at 84.0 ppm
(major) and 81.6 ppm (minor) in the crude 13C NMR spectrum (12 h
measurement time)) as a clear, colorless oil. [a]2D6 (c = 1.27, CHCl3)=+
1
59.58; H NMR (500 MHz, CDCl3, * denotes minor diastereomeric peak):
d = 4.73 (q, 1H, J=6.4 Hz), 4.70* (ddd, 1H, J=10.9, 8.5, 4.4 Hz), 4.45
(ddd, 1H, J=10.3, 8.6, 8.6 Hz), 3.71–3.59 (m, 2H), 3.07 (dd, 1H, J=17.5,
10.3 Hz), 3.06* (dd, 1H, J=17.8, 10.5 Hz), 2.86* (dd, 1H, J=17.3,
8.4 Hz), 2.79 (dd, 1H, J=17.1, 8.8 Hz), 2.14 (brs, 1H), 1.95–1.86 (m,
1H), 1.36* (d, 3H, J=6.4), 1.35 (d, 3H, J=6.5 Hz), 0.95* (d, 3H, J=
6.9 Hz), 0.92 (d, 3H, J=6.4 Hz), 0.90* (s, 9H), 0.89 (s, 9H), 0.10 (s, 3H),
0.07 ppm (s, 3H); 13C NMR (75 MHz, CDCl3, * denotes minor diastereo-
meric peak): d = 162.3, 161.8*, 84.0, 81.6*, 65.8, 65.1*, 64.6, 64.5*, 40.3,
40.0*, 36.9, 35.8*, 25.9, 22.4, 18.2, 13.2, 11.1*, ꢀ4.6, ꢀ4.7 ppm; IR (thin
film): n˜ = 3414, 2957, 2931, 2887, 2859, 1680, 1627, 1472, 1464, 1409,
1389, 1372, 1362, 1339, 1313, 1294, 1259, 1217, 116, 1094, 1042, 1025, 953,
880, 835, 813, 778, 757, 667 cmꢀ1; elemental analysis calcd (%) for
C14H29NO3Si: C 58.49, H 10.17, N 4.87; found: C 58.47, H 10.24, N 5.15;
HRMS (MALDI): m/z: calcd for C14H29NO3SiNa [M+Na]+: 310.1809;
found: 310.1811.
(R)-2-{(S)-3-[(S)-1-(tert-Butyldimethylsilanyloxy)-ethyl]-4,5-dihydroisox-
azol-5-yl}-propan-1-ol (8): To a solution of oxime 7 (102 mg, 0.500 mmol,
1.00 equiv) in toluene (5 mL) at ꢀ788C was added tBuOCl (59.0 mL,
0.500 mmol, 1.00 equiv) dropwise over 10 min. The mixture turned
yellow, then green and finally blue over the course of 15 min and was
stirred total 105 min at ꢀ788C. In a separate flask EtMgBr (3.0m in
Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added dropwise to a solution of
(S)-2-methyl-3-butenol (56.0 mg, 0.650 mmol, 1.30 equiv) and iPrOH
(126 mL, 1.65 mmol, 3.30 equiv) in toluene (10 mL) at 08C. The mixture
was stirred at 08C for 1 h. At this time, the hydroximinoyl chloride in tol-
uene was added to the reaction dropwise via cannula over 2 h, followed
by two rinses (1 mL each). The slightly yellow solution was stirred at 08C
for 6 h and was slowly allowed to warm to room temperature overnight
(large Dewar with ice, covered with aluminum foil). After 18 h the reac-
tion was quenched by addition of saturated, aqueous NH4Cl (15 mL).
The mixture was vigorously stirred for 20 min, then the layers were sepa-
rated and the aqueous phase was extracted with EtOAc (3ꢂ15 mL). The
combined organic solutions were washed with brine, dried over anhy-
drous Na2SO4, filtered, and concentrated under reduced pressure. Purifi-
cation by flash chromatography (hexane/EtOAc 4:1) provided isoxazo-
line 8 (114 mg, 82%, dr 6:1 by integration of the signals at 83.9 ppm
(major) and 81.7 ppm (minor) in the crude 13C NMR spectrum (12 h
measurement time)) as a clear, colorless oil. [a]2D8 (c = 0.99, CHCl3) =
ꢀ10.98; 1H NMR (500 MHz, CDCl3, * denotes minor diastereomeric
peak): d = 4.73 (q, 1H, J=6.4 Hz), 4.67* (ddd, 1H, J=13.2, 8.6, 4.5 Hz),
4.46 (ddd, 1H, J=10.4, 8.3, 8.3 Hz), 3.70–3.59 (m, 2H), 3.07 (dd, 1H, J=
17.1, 10.4 Hz), 3.05* (dd, 1H, J=17.4, 10.9 Hz), 2.84* (dd, 1H, J=17.3,
8.6 Hz), 2.81 (dd, 1H, J=17.2, 8.3 Hz), 2.15 (brs, 1H), 1.94–1.86 (m,
1H), 1.35 (d, 3H, J=6.451 Hz), 0.95* (d, 3H, J=6.9 Hz), 0.92 (d, 3H,
J=7.0 Hz), 0.89 (s, 9H), 0.88* (s, 9H), 0.10 (s, 3H), 0.10* (s, 3H), 0.07 (s,
(R)-2-[(S)-3-(tert-Butyldiphenylsilanyloxymethyl)-4,5-dihydroisoxazol-5-
yl]-propan-1-ol (12): To a solution of oxime 11 (314 mg, 1.00 mmol,
1.00 equiv) in toluene (10 mL) at ꢀ788C was added tBuOCl (118 mL,
1.00 mmol, 1.00 equiv) dropwise over 15 min. The mixture turned yellow,
then green and finally blue over the course of 2.5 h and was stirred total
2.5 h at ꢀ788C. In a separate flask EtMgBr (3.0m in Et2O, 1.00 mL,
3.00 mmol, 3.00 equiv) was added dropwise to a solution of (S)-2-methyl-
3-butenol (112.0 mg, 1.30 mmol, 1.30 equiv) and iPrOH (252 mL,
3.30 mmol, 3.30 equiv) in toluene (20 mL) at 08C. The mixture was
stirred at 08C for 1 h. At this time, the hydroximinoyl chloride in toluene
was added to the reaction dropwise via cannula over 2 h 15 min, followed
by two rinses (1 mL each). The slightly yellow solution was stirred at 08C
for 6 h and was slowly allowed to warm to room temperature overnight
(large Dewar with ice, covered with aluminum foil). After 16 h the reac-
tion was quenched by addition of saturated, aqueous NH4Cl (30 mL).
The mixture was vigorously stirred for 20 min, then the layers were sepa-
rated and the aqueous phase was extracted with EtOAc (3ꢂ30 mL). The
combined organic solutions were washed with brine, dried over anhy-
drous Na2SO4, filtered, and concentrated under reduced pressure. Purifi-
cation by flash chromatography (hexane/EtOAc 4:1 to 2:1) provided is-
AHCTUNGTRENNUNG
Chem. Eur. J. 2009, 15, 12065 – 12081
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12073

E. M. Carreira and N. Lohse-Fraefel
(12 h measurement time)) as a clear, slightly yellow oil. [a]3D0 (c = 0.46,
CHCl3)=+22.88; H NMR (500 MHz, CDCl3, * denotes minor diastereo-
the reaction was quenched by addition of saturated, aqueous NH4Cl
(15 mL). The mixture was vigorously stirred for 20 min, then the layers
were separated and the aqueous phase was extracted with EtOAc (3ꢂ
15 mL). The combined organic solutions were washed with brine, dried
over anhydrous Na2SO4, filtered, and concentrated under reduced pres-
sure. After filtration over silica gel (hexane/EtOAc 2:1) isoxazoline 17
(96 mg, 67%, dr 8:1 by integration of the signals at 4.37 ppm (major) and
1
meric peak): d = 7.67–7.64 (m, 4H), 7.49–7.38 (m, 6H), 4.71* (ddd, 1H,
J=10.9, 8.5, 4.5 Hz), 4.48 (ddd, 1H, J=10.4, 8.4, 8.4 Hz), 4.44 (s, 2H),
3.70–3.59 (m, 2H), 3.08 (dd, 1H, J=17.1, 10.4 Hz), 3.07* (dd, 1H, J=
17.2, 10.9 Hz), 2.87* (dd, 1H, J=17.3, 8.5 Hz), 2.80 (dd, 1H, J=17.2,
8.6 Hz), 2.17 (brs, 1H), 1.94–1.86 (m, 1H), 1.11* (s, 9H), 1.07 (s, 9H),
0.95* (d, 3H, J=6.9 Hz), 0.91 ppm (d, 3H, J=7.0 Hz); 13C NMR
1
4.60 ppm (minor) in the crude H NMR spectrum) was isolated as a pale,
(75 MHz, CDCl3, * denotes minor diastereomeric peak): d
=
158.4,
colorless oil. [a]2D8 (c 0.75, CHCl3)=+27.78; 1H NMR (500 MHz,
=
158.2*, 135.5, 132.6, 130.0, 127.8, 84.1, 81.9*, 65.7, 65.0*, 59.3, 40.2, 40.0*,
38.8, 38.1*, 26.9, 19.4, 13.1, 11.2* ppm; IR (thin film): n˜ =3436, 3071,
3041, 2947, 2932, 2885, 2858, 2363, 1628, 1589, 1472, 1428, 1390, 1363„
1334, 1301, 1259, 1211, 1113, 1083, 1000, 878, 824, 800„ 741, 702 cmꢀ1; ele-
mental analysis calcd (%) for C23H31NO3Si: C 69.48, H 7.86, N 3.52;
CDCl3, * denotes minor diastereomeric peak): d = 4.60* (ddd, 1H, J=
12.9, 8.4, 4.5), 4.37 (ddd, 1H, J=10.3, 8.3, 8.3 Hz), 3.78 (t, 2H, J=
6.4 Hz), 3.64–3.54 (m, 2H), 2.98 (dd, 1H, J=17.1, 10.3 Hz), 2.94* (dd,
1H, J=17.1, 10.8 Hz), 2.80* (dd, 1H, J=17.2, 8.4 Hz), 2.75 (dd, 1H, J=
17.1, 8.4 Hz), 2.55–2.45 (m, 2H), 2.22 (brdd, 1H, J=4.0 Hz), 1.87–1.78
(m, 1H), 0.89* (d, 3H, J=7.0 Hz), 0.85* (s, 9H), 0.84 (d, 3H, J=6.9 Hz),
0.83 (s, 9H), 0.01 ppm (s, 6H); 13C NMR (75 MHz, CDCl3, * denotes
found: C 69.38,
H 7.90, N 3.43; HRMS (MALDI): m/z: calcd for
C23H31NO3SiNa [M+Na]+: 420.1965; found: 420.1960.
minor diastereomeric peak): d
= 157.9, 83.7, 81.5*, 65.9, 64.9*, 60.6,
(R)-2-((S)-3-Trimethylsilanylethynyl-4,5-dihydroisoxazol-5-yl)-propan-1-
ol (14): To a solution of trimethylsilanylpropynal oxime[34] (70.6 mg,
0.500 mmol, 1.00 equiv) in CHCl3 (0.75 mL) at RT was added pyridine
(one drop), followed by NCS (66.8 mg, 0.500 mmol, 1.00 equiv). The solu-
tion was stirred 1.5 h at RT before Et2O (10 mL) was added. After filtra-
tion over Celite the mixture was concentrated, diluted with Et2O
(10 mL), filtered again over Celite and concentrated. In a separate flask
EtMgBr (3.0m in Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added drop-
59.3*, 41.6, 40.2, 39.7*, 31.2, 25.9, 25.7*, 18.2, 13.2, 11.2, ꢀ5.3, ꢀ5.4 ppm;
IR (thin film): n˜ =3413, 2956, 2930, 2884, 2858, 2362, 1626, 1472, 1434,
1388, 1361, 1313, 1257, 1223, 1103, 1049, 1007, 962, 939, 838, 811, 778,
719, 662 cmꢀ1; elemental analysis calcd (%) for C14H29NO3Si: C 58.49, H
10.17, N 4.87; found: C 58.56, H 10.27, N 5.05; HRMS (MALDI): m/z:
calcd for C14H29NO3SiNa [M+Na]+: 310.1809; found: 310.1812.
{1-[(S)-5-((R)-2-Hydroxy-1-methylethyl)-4,5-dihydroisoxazol-3-yl]ethyl}-
phosphonic acid diethyl ester (19): EtMgBr (3.0m in Et2O, 500 mL,
1.50 mmol, 3.00 equiv) was added dropwise to a solution of (S)-2-methyl-
3-butenol (56.0 mg, 0.650 mmol, 1.30 equiv) and iPrOH (126 mL,
1.65 mmol, 3.30 equiv) in CH2Cl2 (10 mL) at 08C. The mixture was
stirred at 08C for 30 min. At this time, previously prepared hydroximino-
yl chloride 20 (122 mg, 0.500 mmol, 1.00 equiv) in CH2Cl2 (5 mL) was
added to the reaction dropwise via cannula over 2 h, followed by two
rinses (1 mL each). The slightly yellow solution was stirred at 08C for 6 h
and was slowly allowed to warm to room temperature overnight (large
Dewar with ice, covered with aluminum foil). After 24 h the reaction was
quenched by addition of saturated, aqueous NH4Cl (15 mL). The mixture
was vigorously stirred for 20 min, then the layers were separated and the
aqueous phase was extracted with EtOAc (3ꢂ15 mL). The combined or-
ganic solutions were washed with brine, dried over anhydrous Na2SO4,
filtered, and concentrated under reduced pressure. Half of the crude ma-
terial was used for the determination of the dr by derivatization with
benzaldehyde to alkene 21 (see below). Analysis of the corresponding
crude 1H NMR spectrum (integration of the signals at 4.60 ppm (major)
and 4.81 ppm (minor)) revealed a diastereomeric ratio of 13:1. Purifica-
tion of half of the crude material by flash chromatography (15% acetone
in EtOAc) provided isoxazoline 19 (52 mg, 71%) as a clear, colorless oil.
1H NMR (300 MHz, CDCl3): d = 4.56–4.45 (m, 1H), 4.20–4.09 (m, 4H),
3.72–3.57 (m, 2H), 3.25–2.87 (m, 3H), 2.41–2.33 (m, 1H), 1.96–1.87 (m,
1H), 1.50–1.40 (m, 3H), 1.36–1.31 (m, 6H), 0.93–0.90 ppm (m, 3H, J=
7.2 Hz) (includes both diastereomers); 13C NMR (75 MHz, CDCl3): d =
156.9, 156.8, 85.1, 84.6, 65.7, 62.8, 62.7, 65.7, 62.8, 62.7, 40.2, 33.4, 33.2,
31.5, 31.3, 29.8, 21.6, 21.5, 16.6, 16.5, 13.3, 13.3, 12.4, 12.3 ppm (includes
both diastereomers); IR (thin film): n˜ =3412, 2982, 2927, 1646, 1617,
1458, 1392, 1333, 1250, 1218, 1164, 1093, 1047, 1021, 968, 900, 860, 799,
725 cmꢀ1; HRMS (MALDI): m/z: calcd for C12H24NO5PNa [M+Na]+:
316.1284; found: 316.1282.
wise to
a solution of (S)-2-methyl-3-butenol (56.0 mg, 0.650 mmol,
1.30 equiv) and iPrOH (126 mL, 1.65 mmol, 3.30 equiv) in toluene
(10 mL) at 08C. The mixture was stirred at 08C for 1 h. At this time, the
previously prepared hydroximinoyl chloride in toluene (5 mL) was added
to the reaction dropwise via cannula over 3.5 h, followed by two rinses
(1 mL each). The slightly yellow solution was stirred at 08C for 6 h and
was slowly allowed to warm to room temperature overnight (large
Dewar with ice, covered with aluminum foil). After 2.5 d the reaction
was quenched by addition of saturated, aqueous NH4Cl (15 mL). The
mixture was vigorously stirred for 20 min, then the layers were separated
and the aqueous phase was extracted with EtOAc (3ꢂ15 mL). The com-
bined organic solutions were washed with brine, dried over anhydrous
Na2SO4, filtered, and concentrated under reduced pressure. The mixture
was filtered over a shot pad of silica gel (hexane/EtOAc 1:2). The analy-
sis of the 1H NMR spectrum (integration of the signals at 4.60 ppm
(major) and 4.79 ppm (minor)) revealed a diastereomeric ratio of 5:1. Pu-
rification by flash chromatography (hexane/EtOAc 3:1) provided isoxa-
zoline 14 (63 mg, 66%) as a clear, colorless oil. [a]2D6 (c = 0.37, CHCl3)=
+53.48; 1H NMR (300 MHz, CDCl3, * denotes minor diastereomeric
peak): d = 4.79* (ddd, 1H, J=13.8, 9.0, 4.8 Hz), 4.60 (ddd, 1H, J=17.2,
9.1, 8.1 Hz), 3.68 (brs, 2H), 3.11* (dd, 1H, J=17.1, 11.1 Hz), 3.10 (dd,
1H, J=17.0, 10.7 Hz), 2.89* (dd, 1H, J=17.0, 9.0 Hz), 2.86 (dd, 1H, J=
17.0, 9.1 Hz), 2.01–1.89 (m, 1H), 1.85 (brs, 1H), 0.97* (d, 3H, J=6.9 Hz),
0.93 (d, 3H, J=7.0 Hz), 0.24 (s, 9H), 0.23* ppm (s, 9H); 13C NMR
(75 MHz, CDCl3, * denotes minor diastereomeric peak): d
= 143.3,
143.2*, 104.6, 104.3*, 93.2*, 93.1, 85.3, 83.5*, 65.3, 64.9*, 41.3, 40.9*, 40.0,
39.8*, 12.8, 11.1*, ꢀ0.4 ppm; IR (thin film): n˜ =3401, 2963, 2895, 2361,
2160, 1594, 1555, 1462, 1437, 1321, 1252, 1085, 1045, 995, 921, 846,
761 cmꢀ1; elemental analysis calcd (%) for C11H19NO2Si: C 58.63, H 8.50,
N 6.22; found: C 58.44, H 8.47, N 6.49; HRMS (MALDI): m/z: calcd for
C11H19NO2SiNa [M+Na]+: 248.1077; found: 248.1073.
(R)-2-{(S)-3-[2-(tert-Butyldimethylsilanyloxy)-ethyl]-4,5-dihydroisoxazol-
5-yl}-propan-1-ol (17): To a solution of oxime 15 (102 mg, 0.500 mmol,
1.00 equiv) in CH2Cl2 (5 mL) at ꢀ788C was added tBuOCl (59.0 mL,
0.500 mmol, 1.00 equiv) dropwise over 15 min. The resulting deep blue
solution was stirred 50 min at ꢀ788C. In a separate flask EtMgBr (3.0m
in Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added dropwise to a solution
of (S)-2-methyl-3-butenol (56.0 mg, 0.650 mmol, 1.30 equiv) and iPrOH
(126 mL, 1.65 mmol, 3.30 equiv) in CH2Cl2 (10 mL) at 08C. The reaction
mixture turned momentarily cloudy and became clear and colorless upon
stirring at 08C for 30 min. At this time, the hydroximinoyl chloride in
CH2Cl2 was added to the reaction dropwise via cannula over 75 min, fol-
lowed by two rinses (1 mL each). The slightly yellow solution was stirred
at 08C for 6 h and was slowly allowed to warm to room temperature
overnight (large Dewar with ice, covered with aluminum foil). After 48 h
(R)-2-[(S)-3-((E)-1-Methyl-2-phenylvinyl)-4,5-dihydroisoxazol-5-yl]pro-
pan-1-ol (21): To flame dried LiCl (12.7 mg, 0.300 mmol, 1.20 equiv) was
added azeotropically dried unpurified cycloadduct 19 (0.250 mmol,
1.00 equiv) in MeCN (3 mL) at RT, followed by dropwise addition of
DBU (41.1 mL, 0.275 mmol, 1.10 equiv). The yellow, slightly cloudy mix-
ture was stirred 15 min at RT, before freshly distilled benzaldehyde
(37.2 mL, 0.375 mmol, 1.50 equiv) was added dropwise. After the mixture
was stirred at RT for 15 h, H2O (3 mL) and brine (3 mL) were added and
the resulting solution extracted with EtOAc (3ꢂ10 mL). The combined
organic solutions were washed with brine, dried over anhydrous Na2SO4,
filtered, and concentrated under reduced pressure. Purification by flash
chromatography (hexane/EtOAc 1:1) provided isoxazoline 21 (29 mg,
47% yield over two steps, dr 13:1 by integration of the signals at
4.60 ppm (major) and 4.81 ppm (minor) in the crude 1H NMR spectrum)
12074
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12065 – 12081

Polyketide Building Blocks
FULL PAPER
as a white solid. M.p. 868C; [a]2D8 (c = 0.28, CHCl3)=+55.48; 1H NMR
(300 MHz, CDCl3, * denotes minor diastereomeric peak): d = 7.42–7.27
(m, 5H), 6.70 (s, 1H), 4.81* (ddd, 1H, J=13.7, 9.0, 5.0 Hz), 4.60 (ddd,
1H, J=10.3, 8.7, 8.7 Hz), 3.78–3.63 (m, 2H), 3.31 (dd, 1H, J=15.9,
10.3 Hz), 3.29* (dd, 1H, J=16.2, 10.6 Hz), 3.06* (dd, 1H, J=16.5,
9.0 Hz), 3.02 (dd, 1H, J=16.2, 8.7 Hz), 2.22 (d, 3H, J=1.3 Hz), 2.17* (d,
3H, J=1.6 Hz), 2.07–1.94 (m, 1H), 1.00* (d, 3H, J=6.9 Hz), 0.98 ppm
(d, 3H, J=6.9 Hz); 13C NMR (75 MHz, CDCl3): d = 160.4, 136.3, 133.4,
129.4, 129.1, 128.4, 127.7, 85.6, 66.3, 40.6, 38.6, 14.9, 13.5 ppm; IR (thin
film): n˜ =3402, 2962, 2924, 1560, 1490, 1442, 1372, 1239, 1086, 1028, 918,
864, 814, 758, 698, 618, 581 cmꢀ1; elemental analysis calcd (%) for
C15H19NO2: C 73.44, H 7.81, N 5.71; found: C 72.93, H 7.32, N 5.21;
HRMS (MALDI): m/z: calcd for C14H19NO2 [MꢀH]+: 244.1332; found:
244.1335.
for 20 min, then the layers were separated and the aqueous phase was ex-
tracted with EtOAc (3ꢂ15 mL). The combined organic solutions were
washed with brine, dried over anhydrous Na2SO4, filtered, and concen-
trated under reduced pressure. Purification by flash chromatography
(hexane/EtOAc 5:1 to 2:1) provided isoxazoline 34 (269 mg, 82%, dr
19:1 by integration of the signals at 4.70 ppm (major) and 4.79 ppm
(minor) in the crude 1H NMR spectrum) as a pale, thick oil. 1H NMR
(300 MHz, CDCl3): d = 7.67–7.63 (m, 4H), 7.48–7.22 (m, 21H), 4.70
(ddd, 1H, J=10.2, 8.8, 8.8 Hz), 4.41 (dd, 2H, J=15.7, 12.9 Hz), 3.96–3.83
(m, 2H), 3.33 (dd, 1H, J=9.6, 4.9 Hz), 3.19 (dd, 1H, J=9.6, 5.5 Hz), 3.00
(dd, 1H, J=17.3, 10.4), 2.71 (dd, 1H, J=17.3, 8.8 Hz), 2.29 (brs, 1H),
2.00–1.91 (m, 1H), 1.07 ppm (s, 9H); 13C NMR (75 MHz, CDCl3): d =
158.8, 143.7, 135.6, 132.7, 130.1, 128.7, 128.1, 128.0, 127.3, 87.2, 81.0, 62.8,
62.5, 59.3, 46.4, 39.5, 26.9, 19.3 ppm; IR (thin film): n˜ =3436, 3067, 3020,
2947, 2931, 2858, 2361, 1960, 1898, 1825, 1590, 1490, 1472, 1449, 1428,
1391, 1363, 1333, 1254, 1217, 1151, 1113, 1088, 1026, 933, 899, 876, 824,
(R)-2-[(S)-3-((E)-1-Methyl-2-phenylvinyl)-4,5-dihydroisoxazol-5-yl]pro-
pan-1-ol (22): To
a solution of syn-benzaldehyde oxime (60.6 mg,
798, 761, 745, 702, 668 cmꢀ1
; HRMS (MALDI): m/z: calcd for
0.500 mmol, 1.00 equiv) in CH2Cl2 (5 mL) at 08C was added tBuOCl
(62.2 mL, 0.550 mmol, 1.10 equiv) dropwise. After 1.5 h the solution was
concentrated under reduced pressure. The resulting off-white solid was
dried azeotropically with toluene (3ꢂ5 mL). Last traces of tBuOCl were
removed under high vacuum for 10 min. In a separate flask EtMgBr
(3.0m in Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added dropwise to a
solution of (S)-2-methyl-3-butenol (56.0 mg, 0.650 mmol, 1.30 equiv) and
iPrOH (126 mL, 1.65 mmol, 3.30 equiv) in CH2Cl2 (10 mL) at 08C. The re-
action mixture turned momentarily cloudy and became clear and color-
less upon stirring at 08C for 30 min. At this time, the hydroximinoyl chlo-
ride in CH2Cl2 (5 mL) at RT was added to the reaction dropwise via can-
nula over 1 h 45 min, followed by two rinses (1 mL each). The slightly
yellow solution was stirred at 08C for 6 h and was slowly allowed to
warm to room temperature overnight (large Dewar with ice, covered
with aluminum foil). After 14 h the reaction was quenched by addition of
saturated, aqueous NH4Cl (15 mL). The mixture was vigorously stirred
for 20 min, then the layers were separated and the aqueous phase was ex-
tracted with EtOAc (3ꢂ15 mL). The combined organic solutions were
washed with brine, dried over anhydrous Na2SO4, filtered, and concen-
trated under reduced pressure. Purification by flash chromatography
(hexane/EtOAc 3:1 to 1:1) provided isoxazoline 22 (89 mg, 87%, dr 7:1
by integration of the signals at 4.66 ppm (major) and 4.85 ppm (minor) in
the crude 1H NMR spectrum) as a white solid. M.p. 618C; [a]2D2 (c =
0.89, CHCl3)=+63.58; 1H NMR (300 MHz, CDCl3, * denotes minor dia-
C42H45NO4SiNa [M+Na]+: 678.3010; found: 678.3001.
(S)-3-(tert-Butyldiphenylsilanyloxy)-2-[(S)-3-(tert-butyldiphenylsilanyl-
ACHUTNGRENUoNG xyACHUTTGNRENmNUGN ethyl)-4,5-dihydroisoxazol-5-yl]-propan-1-ol (35): To a solution of
oxime 11 (157 mg, 0.500 mmol, 1.00 equiv) in CH2Cl2 (5 mL) at ꢀ788C
was added tBuOCl (59 mL, 0.500 mmol, 1.00 equiv) dropwise over 10 min.
The resulting green solution was stirred 2.5 h at ꢀ788C. In a separate
flask EtMgBr (3.0m in Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added
dropwise to a solution of monoprotected homoallylic diol 32b (221 mg,
0.650 mmol, 1.30 equiv) and iPrOH (126 mL, 1.65 mmol, 3.30 equiv) in
CH2Cl2 (10 mL) at 08C. The reaction mixture turned momentarily cloudy
and became clear and colorless upon stirring at 08C for 30 min. At this
time, the hydroximinoyl chloride in CH2Cl2 was added to the reaction
dropwise via cannula over 60 min, followed by two rinses (1 mL each).
The slightly yellow solution was stirred at 08C for 6 h and was slowly al-
lowed to warm to room temperature overnight (large Dewar with ice,
covered with aluminum foil). After 18 h the reaction was quenched by
addition of saturated, aqueous NH4Cl (15 mL). The mixture was vigo-
rously stirred for 20 min, then the layers were separated and the aqueous
phase was extracted with EtOAc (3ꢂ15 mL). The combined organic solu-
tions were washed with brine, dried over anhydrous Na2SO4, filtered, and
concentrated under reduced pressure. Purification by flash chromatogra-
phy (hexane/EtOAc 10:1) provided isoxazoline 35 (277 mg, 85%, dr 21:1
by integration of the signals at 80.6 ppm (major) and 79.4 ppm (minor) in
the crude 13C NMR spectrum (12 h measurement time)) as a pale, thick
oil. 1H NMR (300 MHz, CDCl3): d = 7.67–7.61 (m, 8H), 7.47–7.35 (m,
12H), 4.67 (ddd, 1H, J=10.3, 8,7, 8.7 Hz), 4.40 (dd, 2H, J=15.3,
12.8 Hz), 3.95–3.81 (m, 2H), 3.81–3.69 (m, 2H), 3.04 (dd, 1H, J=17.1,
10.6 Hz), 2.77 (dd, 1H, J=17.1, 9.0 Hz), 2.30 (dd, 1H, J=5.6, 6.5 Hz),
1.97–1.87 (m, 1H), 1.07 (s, 9H), 1.06 ppm (s, 9H); 13C NMR (75 MHz,
CDCl3): d = 158.7, 135.5, 135.5, 132.9, 132.9, 132.6, 130.0, 129.9, 127.9,
80.6, 63.3, 62.3, 59.4, 47.8, 39.6, 27.1, 27.0, 19.4, 19.4 ppm; IR (thin film):
n˜ =34.36, 3071, 3050, 3014, 2958, 2931, 2891, 2858, 2361, 1962, 1891,
1825, 1773, 1654, 1628, 1590, 1472, 1428, 1391, 1362, 1334, 1306, 1260,
1216, 1189, 1113, 1000, 998, 938, 880„ 824, 803, 741, 702, 668, 613 cmꢀ1; el-
emental analysis calcd (%) for C39H49NO4Si2: C,71.85, H 7.57, N 2.15;
stereomeric peak): d
= 7.70–7.63 (m, 2H), 7.43–7.37 (m, 3H), 4.85*
(ddd, 1H, J=13.4, 8.6, 4.7 Hz), 4.66 (ddd, 1H, J=10.3, 8.4, 8.4 Hz), 3.77–
3.66 (m, 2H), 3.40 (dd, 1H, J=16.5, 10.3 Hz), 3.39* (dd, 1H, J=16.8,
10.9 Hz), 3.17* (dd, 1H, J=16.8, 8.7 Hz), 3.13 (dd, 1H, J=16.5, 8.7 Hz),
2.26 (brs, 1H), 2.09–1.85 (m, 1H), 0.99* (d, 3H, J=6.8 Hz), 0.98 ppm (d,
3H, J=6.8 Hz); 13C NMR (75 MHz, CDCl3, * denotes minor diastereo-
meric peak): d = 156.8, 130.2, 129.5, 128.7, 126.7, 84.9, 82.8*, 65.9, 65.3*,
40.4, 40.0*, 38.9, 38.0*, 13.3, 11.5* ppm; IR (thin film): n˜ =3401, 3062,
2964, 2931, 2881, 1597, 1570, 1498, 1447, 1357, 1235, 1181, 1077, 1041,
996, 908, 855, 761, 692, 674 cmꢀ1; elemental analysis calcd (%) for
C12H15NO2: C 70.22, H 7.37, N 6.82; found: C 70.09, H 7.45, N 6.65;
HRMS (MALDI): m/z: calcd for C12H16NO2 [M+H]+: 206.1176; found:
206.1179.
found: C 71.69,
H 7.71, N 2.29; HRMS (MALDI): m/z: calcd for
C39H49NO4Si2Na [M+Na]+: 674.3092; found: 674.3100.
(S)-2-[(S)-3-(tert-Butyldiphenylsilanyloxymethyl)-4,5-dihydroisoxazol-5-
yl]-3-trityloxy-propan-1-ol (34): To
a
solution of oxime 11 (157 mg,
(S)-2-[(S)-3-(tert-Butyldiphenylsilanyloxymethyl)-4,5-dihydroisoxazol-5-
yl]-3-triisopropylsilanyloxypropan-1-ol (36): To a solution of oxime 11
(157 mg, 0.500 mmol, 1.00 equiv) in CH2Cl2 (5 mL) at ꢀ788C was added
tBuOCl (59 mL, 0.500 mmol, 1.00 equiv) dropwise over 10 min. The re-
sulting green solution was stirred 2.5 h at ꢀ788C. In a separate flask
EtMgBr (3.0m in Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added drop-
0.500 mmol, 1.00 equiv) in CH2Cl2 (5 mL) at ꢀ788C was added tBuOCl
(59 mL, 0.500 mmol, 1.00 equiv) dropwise over 10 min. The resulting deep
blue solution was stirred 2.5 h at ꢀ788C. In a separate flask EtMgBr
(3.0m in Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added dropwise to a
solution of monoprotected homoallylic diol 32a (224 mg, 0.650 mmol,
1.30 equiv) and iPrOH (126 mL, 1.65 mmol, 3.30 equiv) in CH2Cl2
(10 mL) at 08C. The reaction mixture turned momentarily cloudy and
became clear and colorless upon stirring at 08C for 30 min. At this time,
the hydroximinoyl chloride in CH2Cl2 was added to the reaction dropwise
via cannula over 90 min, followed by two rinses (1 mL each). The slightly
yellow solution was stirred at 08C for 6 h and was slowly allowed to
warm to room temperature overnight (large Dewar with ice, covered
with aluminum foil). After 18 h the reaction was quenched by addition of
saturated, aqueous NH4Cl (15 mL). The mixture was vigorously stirred
wise to
a solution of monoprotected homoallylic diol 32c (168 mg,
0.650 mmol, 1.30 equiv) and iPrOH (126 mL, 1.65 mmol, 3.30 equiv) in
CH2Cl2 (10 mL) at 08C. The reaction mixture turned momentarily cloudy
and became clear and colorless upon stirring at 08C for 30 min. At this
time, the hydroximinoyl chloride in CH2Cl2 was added to the reaction
dropwise via cannula over 90 min, followed by two rinses (1 mL each).
The slightly yellow solution was stirred at 08C for 6 h and was slowly al-
lowed to warm to room temperature overnight (large Dewar with ice,
covered with aluminum foil). After 18 h the reaction was quenched by
Chem. Eur. J. 2009, 15, 12065 – 12081
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12075

E. M. Carreira and N. Lohse-Fraefel
addition of saturated, aqueous NH4Cl (15 mL). The mixture was vigo-
rously stirred for 20 min, then the layers were separated and the aqueous
phase was extracted with EtOAc (3ꢂ15 mL). The combined organic solu-
tions were washed with brine, dried over anhydrous Na2SO4, filtered, and
concentrated under reduced pressure. Purification by flash chromatogra-
phy (hexane/EtOAc 10:1) provided isoxazoline 36 (209 mg, 73%, dr 14:1
by integration of the signals at 80.6 ppm (major) and 79.0 ppm (minor) in
the crude 13C NMR spectrum (12 h measurement time)) as a clear, color-
less oil. 1H NMR (300 MHz, CDCl3): d = 7.67–7.64 (m, 4H), 7.48–7.37
(m, 6H), 4.70 (ddd, 1H; J=16.5, 9.0, 7.5 Hz), 4.44 (s, 2H), 3.93–3.80 (m,
4H), 3.16 (dd, 1H, J=17.4, 10.6), 2.95 (dd, 1H, J=17.1, 9.0 Hz), 2.52 (t,
1H, J=5.9 Hz), 1.99–1.90 (m, 1H), 1.11–1.04 ppm (m, 30H); 13C NMR
(75 MHz, CDCl3): d = 158.8, 135.6, 132.7, 130.1, 127.9, 80.6, 63.7, 62.9,
59.4, 47.6, 39.5, 27.0, 19.5, 18.2, 12.1 ppm; IR (thin film): n˜ =3436, 3072,
3051, 2942, 2892, 2865, 2717, 2247, 1961, 1891, 1822, 1773, 1656, 1628,
1590, 1464, 1428, 1383, 1363, 1334, 1307, 1257, 1208, 1156, 1113, 1013,
998, 910, 883, 824, 795, 739, 702, 690, 660, 612 cmꢀ1; HRMS (MALDI):
m/z: calcd for C32H51NO4Si2Na [M+Na]+: 592.3249; found: 592.3241.
addition of saturated, aqueous NH4Cl (15 mL). The mixture was vigo-
rously stirred for 20 min, then the layers were separated and the aqueous
phase was extracted with EtOAc (3ꢂ15 mL). The combined organic solu-
tions were washed with brine, dried over anhydrous Na2SO4, filtered, and
concentrated under reduced pressure. Purification by flash chromatogra-
phy (hexane/EtOAc 10:1) provided isoxazoline 38 (139 mg, 53%, dr 6:1
by integration of the signals at 79.2 ppm (major) and 78.8 ppm (minor) in
the crude 13C NMR spectrum (12 h measurement time)) as a clear, color-
less oil. 1H NMR (300 MHz, CDCl3): d = 7.70–7.64 (m, 4H), 7.48–7.35
(m, 6H), 4.70 (ddd, 1H, J=16.6, 9.2, 7.3 Hz), 4.55 (d, 2H, J=1.2 Hz),
4.14–4.04 (m, 2H), 3.74 (dd, 2H, J=5.4, 0.9 Hz), 3.15–2.97 (m, 2H),
2.18–2.04 (m, 1H), 1.07 (s, 9H), 0.94 (t, 9H, J=8.0 Hz), 0.58 ppm (q, 6H,
J=8.0 Hz); 13C NMR (75 MHz, CDCl3): d = 158.6, 135.6, 133.0, 130.0,
127.9, 79.2, 60.5, 59.5, 57.4, 45.2, 38.7, 27.0, 19.5, 7.1, 4.5 ppm; IR (thin
film): n˜ =3338, 3072, 3050, 2956, 2932, 2876, 2859, 1960, 1889, 1822, 1768,
1657, 1590, 1472, 1428, 1390, 1362, 1308, 1258, 1240, 1209, 1186, 1114,
1006, 939, 909, 881, 823, 802, 740, 702, 612 ppm; HRMS (MALDI): m/z:
calcd for C29H45NO4Si2Na [M+Na]+: 550.2779; found: 550.2772.
(S)-3-(tert-Butyldimethylsilanyloxy)-2-[(S)-3-(tert-butyldiphenylsilanyl-
ACHTUNGTRENNUNGoxyACHTUNGTRENNUNGmethyl)-4,5-dihydroisoxazol-5-yl]-propan-1-ol (37): To a solution of
(S)-2-{(S)-3-[(R)-2-(tert-Butyldimethylsilanyloxy)-1-methylethyl]-4,5-di-
hydroisoxazol-5-yl}-3-trityloxypropan-1-ol (40): To a solution of oxime 1
(326 mg, 1.50 mmol, 1.00 equiv) in CH2Cl2 (15 mL) at ꢀ788C was added
tBuOCl (177 mL, 1.50 mmol, 1.00 equiv) dropwise over 10 min. The re-
sulting deep blue solution was stirred 1 h at ꢀ788C. In a separate flask
EtMgBr (3.0m in Et2O, 1.50 mL, 4.50 mmol, 3.00 equiv) was added drop-
wise to a solution of monoprotected homoallylic diol 32a (672 mg,
1.95 mmol, 1.30 equiv) and iPrOH (378 mL, 4.95 mmol, 3.30 equiv) in
CH2Cl2 (30 mL) at 08C. The reaction mixture turned momentarily cloudy
and became clear and colorless upon stirring at 08C for 30 min. At this
time, the hydroximinoyl chloride in CH2Cl2 was added to the reaction
dropwise via cannula over 60 min, followed by two rinses (1 mL each).
The slightly yellow solution was stirred at 08C for 6 h and was slowly al-
lowed to warm to room temperature overnight (large Dewar with ice,
covered with aluminum foil). After 18 h the reaction was quenched by
addition of saturated, aqueous NH4Cl (15 mL). The mixture was vigo-
rously stirred for 20 min, then the layers were separated and the aqueous
phase was extracted with EtOAc (3ꢂ15 mL). The combined organic solu-
tions were washed with brine, dried over anhydrous Na2SO4, filtered, and
concentrated under reduced pressure. Purification by flash chromatogra-
phy (hexane/EtOAc 8:1 to 6:1) provided isoxazoline 40 (721 mg, 86%, dr
25:1 by integration of the signals at 4.64 ppm (major) and 4.79 ppm
(minor) in the crude 1H NMR spectrum) as a pale, thick oil. 1H NMR
(300 MHz, CDCl3): d = 7.45–7.40 (m, 6H), 7.34–7.21 (m, 9H), 4.70–4.58
(m, 1H), 3.2–3.77 (m, 2H), 3.64–3.62 (m, 2H), 3.31–3.26 (m, 1H), 3.20–
3.13 (m, 1H), 3.01–2.82 (m, 1H), 2.80–2.63 (m, 2H), 2.48 (brt, 0.5H, J=
6.1 Hz), 2.38 (brt, 0.5H, J=6.0 Hz), 2.00–1.90 (m, 1H), 1.12 (d, 1.5H, J=
7.0 Hz), 1.11 (d, 1.5H, J=7.0 Hz), 0.87 (s, 9H), 0.04 ppm (s, 6H) (in-
cludes both anti-diastereomers); 13C NMR (75 MHz, CDCl3): d = 161.7,
161.6, 143.5, 128.5, 127.8, 127.0, 87.0, 86.9, 80.2, 80.0, 66.0, 62.7, 62.5, 46.4,
46.3, 40.2, 39.9, 35.8, 25.9, 18.3, 14.7, 14.6, ꢀ5.2, ꢀ5.3, ꢀ5.4 ppm (includes
both anti-diastereomers); IR (thin film): n˜ =3468, 3061, 3030, 2958, 2930,
2886, 2857, 2361, 2342, 1955, 1815, 1721, 1654, 1623, 1598, 1490, 1472,
1446, 1390, 1358, 1325, 1256, 1210, 1176, 1158, 1088, 1032, 1012, 969, 939,
912, 896, 838, 777, 760, 700, 668, 638 cmꢀ1; HRMS (MALDI): m/z: calcd
for C34H45NO4SiNa [M+Na]+: 583.3010; found: 578.3004.
the oxime 11 (107 mg, 0.340 mmol, 1.00 equiv) in CH2Cl2 (3.5 mL) at
ꢀ788C was added tBuOCl (40 mL, 0.340 mmol, 1.00 equiv) dropwise over
10 min. The resulting green solution was stirred 2.5 h at ꢀ788C. In a sep-
arate flask EtMgBr (3.0m in Et2O, 340 mL, 1.02 mmol, 3.00 equiv) was
added dropwise to a solution of monoprotected homoallylic diol 32d
(98.0 mg, 0.450 mmol, 1.30 equiv) and iPrOH (86.0 mL, 1.12 mmol,
3.30 equiv) in CH2Cl2 (7 mL) at 08C. The reaction mixture turned mo-
mentarily cloudy and became clear and colorless upon stirring at 08C for
30 min. At this time, the hydroximinoyl chloride in CH2Cl2 was added to
the reaction dropwise via cannula over 90 min, followed by two rinses
(1 mL each). The slightly yellow solution was stirred at 08C for 6 h and
was slowly allowed to warm to room temperature overnight (large
Dewar with ice, covered with aluminum foil). After 18 h the reaction was
quenched by addition of saturated, aqueous NH4Cl (15 mL). The mixture
was vigorously stirred for 20 min, then the layers were separated and the
aqueous phase was extracted with EtOAc (3ꢂ15 mL). The combined or-
ganic solutions were washed with brine, dried over anhydrous Na2SO4,
filtered, and concentrated under reduced pressure. Purification by flash
chromatography (hexane/EtOAc 10:1 to 4:1) provided isoxazoline 37
(141 mg, 78%, dr 10:1 by integration of the signals at 80.6 ppm (major)
and 79.0 ppm (minor) in the crude 13C NMR spectrum (12 h measure-
ment time)) as a clear, colorless oil. 1H NMR (300 MHz, CDCl3): d =
7.74–7.65 (4H), 7.49–7.38 (m, 6H), 4.66 (ddd, 1H, J=16.5, 9.1, 7.4 Hz),
4.45 (s, 2H), 3.83 (brs, 2H), 3.76 (dd, 2H, J=5.5, 1.9 Hz), 3.14 (dd, 1H,
J=17.3, 10.7 Hz), 2.93 (dd, 1H, J=17.3, 9.1 Hz), 2.47 (brs, 1H), 1.97–
1.88 (m, 1H), 1.07 (s, 9H), 0.90 (s, 9H), 0.08 (s, 3H), 0.07 ppm (s, 3H);
13C NMR (75 MHz, CDCl3): d = 158.8, 135.5, 132.6, 130.0, 127.9, 80.6,
63.1, 62.6, 59.4, 47.4, 39.4, 27.0, 26.1, 19.5, 18.4, ꢀ5.3, ꢀ5.3 ppm; IR (thin
film): n˜ =3430, 3072, 3050, 2855, 2931, 2889, 2858, 2247, 1958, 1893, 1827,
1712, 1662, 1624, 1585, 1470, 1426, 1390, 1362, 1334, 1255, 1209, 1113,
1007, 938, 910, 838, 779, 739, 702, 611 cmꢀ1; elemental analysis calcd (%)
for C29H45NO4Si2: C 65.99, H 8.59, N 2.65; found: C 66.13, H 8.50, N
2.67; HRMS (MALDI): m/z: calcd for C29H45NO4Si2Na [M+Na]+:
550.2779; found: 550.2772.
(S)-2-[(S)-3-(tert-Butyldiphenylsilanyloxymethyl)-4,5-dihydro-isoxazol-5-
(S)-2-{(S)-3-[(R)-2-(tert-Butyldimethylsilanyloxy)-1-methylethyl]-4,5-di-
hydroisoxazol-5-yl}-3-(tert-butyldiphenylsilanyloxy)-propan-1-ol (41): To
a solution of oxime 1 (109 mg, 0.500 mmol, 1.00 equiv) in CH2Cl2 (5 mL)
at ꢀ788C was added tBuOCl (59 mL, 0.500 mmol, 1.00 equiv) dropwise
over 10 min. The resulting deep blue solution was stirred 1 h at ꢀ788C.
In a separate flask EtMgBr (3.0m in Et2O, 500 mL, 1.50 mmol, 3.00 equiv)
was added dropwise to a solution of monoprotected homoallylic diol 32b
(221 mg, 0.65 mmol, 1.30 equiv) and iPrOH (126 mL, 1.65 mmol,
3.30 equiv) in CH2Cl2 (10 mL) at 08C. The reaction mixture turned mo-
mentarily cloudy and became clear and colorless upon stirring at 08C for
30 min. At this time, the hydroximinoyl chloride in CH2Cl2 was added to
the reaction dropwise via cannula over 60 min, followed by two rinses
(1 mL each). The slightly yellow solution was stirred at 08C for 6 h and
was slowly allowed to warm to room temperature overnight (large
Dewar with ice, covered with aluminum foil). After 18 h the reaction was
yl]-3-triethylsilanyloxy-propan-1-ol (38): To
a solution of oxime 11
(157 mg, 0.500 mmol, 1.00 equiv) in CH2Cl2 (5 mL) at ꢀ788C was added
tBuOCl (59 mL, 0.500 mmol, 1.00 equiv) dropwise over 10 min. The re-
sulting green solution was stirred 2.5 h at ꢀ788C. In a separate flask
EtMgBr (3.0m in Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was added drop-
wise to a solution of monoprotected homoallylic diol 32e (141 mg,
0.650 mmol, 1.30 equiv) and iPrOH (126 mL, 1.65 mmol, 3.30 equiv) in
CH2Cl2 (10 mL) at 08C. The reaction mixture turned momentarily cloudy
and became clear and colorless upon stirring at 08C for 30 min. At this
time, the hydroximinoyl chloride in CH2Cl2 was added to the reaction
dropwise via cannula over 90 min, followed by two rinses (1 mL each).
The slightly yellow solution was stirred at 08C for 6 h and was slowly al-
lowed to warm to room temperature overnight (large Dewar with ice,
covered with aluminum foil). After 18 h the reaction was quenched by
12076
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12065 – 12081

Polyketide Building Blocks
FULL PAPER
quenched by addition of saturated, aqueous NH4Cl (15 mL). The mixture
was vigorously stirred for 20 min, then the layers were separated and the
aqueous phase was extracted with EtOAc (3ꢂ15 mL). The combined or-
ganic solutions were washed with brine, dried over anhydrous Na2SO4,
filtered, and concentrated under reduced pressure. Half of the crude ma-
terial was used for the determination of the dr by derivatization to the
corresponding trityl ether (protection of the primary alcohol with TrCl,
corresponding trityl ether (protection of the primary alcohol with TrCl,
see below). Analysis of the corresponding crude H NMR spectrum (inte-
1
gration of the signals at 4.62 ppm (major) and 4.55 ppm (minor)) re-
vealed a diastereomeric ratio of 6:1. Purification of half of the crude ma-
terial by flash chromatography (hexane/EtOAc 10:1) provided isoxazo-
line 42 (91 mg, 84%) as a clear, colorless oil. 1H NMR (300 MHz,
CDCl3): d = 4.66–4.55 (m, 1H), 3.82–3.66 (m, 6H), 3.12–2.98 (m, 1H),
2.93–2.72 (m, 2H), 2.66 (brs, 0.5H), 2.57 (brs, 0.5H), 1.93–1.85 (m, 1H),
1.16 (d, 1.5H, J=7.2 Hz), 1.14 (d, 1.5H, J=7.2 Hz), 0.89 (s, 18H), 0.06 (s,
3H), 0.06 (s, 6H), 0.05 ppm (s, 3H) (includes both anti-diastereomers);
13C NMR (75 MHz, CDCl3): d = 162.1, 162.0, 79.8, 79.6, 66.2, 63.1, 63.0,
62.7, 62.5, 47.5, 47.4, 40.2, 39.8, 35.9, 26.0, 18.4, 18.3, 14.8, 14.7, ꢀ5.4,
ꢀ5.5 ppm (includes both anti-diastereomers); IR (thin film): n˜ =3436,
2957, 2930, 2886, 2858, 2246, 1620, 1579, 1472, 1434, 1390, 1362, 1310,
1257, 1216, 1095, 1006, 939, 911, 838, 814, 777, 734, 668 cmꢀ1; HRMS
(ESI): m/z: calcd for C21H45NO4Si2Na [M+Na]+: 454.2779; found:
454.2773.
1
see below). Analysis of the corresponding crude H NMR spectrum (inte-
gration of the signals at 4.63 ppm (major) and 4.40 ppm (minor)) re-
vealed a diastereomeric ratio of 20:1. Purification of half of the crude
material by flash chromatography (hexane/EtOAc 6:1) provided isoxazo-
line 41 (115 mg, 83%) as a clear, colorless oil. 1H NMR (300 MHz,
CDCl3): d = 7.69–7.61 (m, 4H), 7.48–7.36 (m, 6H), 4.68–4.56 (m, 1H),
3.95–3.58 (m, 6H), 2.98–2.83 (m, 1H), 2.81–2.68 (m, 2H), 2.53 (brs,
0.5H), 2.43 (brs, 0.5H), 1.95–1.86 (m, 1H), 1.12 (d, 1.5H, J=7.2 Hz),
1.11 (d, 1.5H, J=6.9 Hz), 1.06 (s, 9H), 0.87 (s, 9H), 0.04 ppm (s, 6H) (in-
cludes both anti-diastereomers); 13C NMR (75 MHz, CDCl3): d = 161.7,
161.6, 135.4, 132.8, 129.8, 127.7, 79.7, 79.6, 66.0, 65.9, 63.3, 63.2, 62.4, 62.2,
47.8, 47.7, 40.2, 39.9, 35.8, 26.9, 25.9, 19.3, 18.3, 14.7, 14.6, ꢀ5.3, ꢀ5.4 ppm
(includes both anti-diastereomers); IR (thin film): n˜ =3436, 2956, 2930,
2858, 2361, 1958, 1895, 1818, 1660, 1590, 1472, 1428, 1390, 1362, 1307,
(S)-3-(tert-Butyldimethylsilanyloxy)-2-{(R)-3-[(R)-2-(tert-butyldimethylsi-
lanyloxy)-1-methyl-ethyl]-4,5-dihydroisoxazol-5-yl}-propan-1-ol
(56):
1H NMR (300 MHz, CDCl3): d = 4.73–4.62 (m, 1H), 3.99–3.84 (m, 2H),
3.76–3.62 (m, 4H), 3.10–2.99 (m, 1H), 2.94–2.71 (m, 3H), 1.90–1.80 (m,
1H), 1.16 (d, 1.5H, J=6.9 Hz), 1.15 (d, 1.5H, J=7.2 Hz), 0.90 (s, 9H),
0.89 (s, 9H), 0.10 (s, 3H), 0.09 (s, 3H), 0.05 ppm (s, 6H) (includes both
syn-diastereomers); 13C NMR (75 MHz, CDCl3): d = 161.9, 161.8, 78.1,
77.4, 66.2, 64.0, 63.8, 46.3, 46.2, 39.5, 39.0, 36.2, 36.1, 26.1, 18.5, 18.4, 14.9,
ꢀ5.2, ꢀ5.3 ppm (includes both syn-diastereomers); IR (thin film): n˜
=3436, 2955, 2930, 2892, 2858, 1586, 1472, 1408, 1390, 1361, 1256, 1211,
1093, 1045, 1003, 972, 936, 837, 776 cmꢀ1; HRMS (ESI): m/z: calcd for
C21H45NO4Si2Na [M+Na]+: 454.2779; found: 454.2770.
1257, 1212, 1190, 1113, 1004, 936, 837, 778, 740, 702 cmꢀ1
; HRMS
(MALDI): m/z: calcd for C31H49NO4Si2Na [M+Na]+: 578.3092; found:
578.3096.
(S)-3-[(R)-2-(tert-Butyldimethylsilanyloxy)-1-methylethyl]-5-[(S)-1-(tert-
butyldiphenylsilanyloxymethyl)-2-trityloxyethyl]-4,5-dihydroisoxazole
(55): To unpurified primary alcohol 41 (0.250 mmol, 1.00 equivalent) in
dry pyridine (2 mL) at RT was added TrCl (209 mg, 0.750 mmol,
3.00 equiv). The mixture was stirred at RT for 20 h before all volatiles
were removed under reduced pressure. Last traces of pyridine were re-
moved by azeotropic coevaporation with cyclohexane (3ꢂ4 mL). Purifi-
cation by flash chromatography (hexane/EtOAc 10:1) provided 55
(134 mg, 67% over two steps, dr 20:1 by integration of the signals at
4.63 ppm (major) and 4.40 ppm (minor) in the crude 1H NMR spectrum)
as a pale, thick oil. 1H NMR (300 MHz, CDCl3): d = 7.61–7.58 (m 4H),
7.43–7.17 (m„ 21H), 4.68–4.58 (m, 1H), 3.92–3.83 (m, 1H), 3.78–3.72 (m,
1H), 3.64–3.49 (m, 2H), 3.33 (d, 1H, J=5.9 Hz), 3.32 (d, 1H, J=6.2 Hz),
2.82–2.56 (m, 3H), 2.10–2.00 (m, 1H), 1.07 (d, 1.5H, J=6.9 Hz), 1.02 (d,
1.5H, J=6.9 Hz), 0.96 (s, 9H), 0.86 (s, 9H), 0.02 (s, 3H), 0.01 ppm (s,
3H) (includes both anti-diastereomers); 13C NMR (75 MHz, CDCl3): d =
161.3, 144.2, 135.7, 135.6, 133.5, 133.3, 129.7, 128.7, 127.8, 127.3, 126.9,
86.8, 79.0, 66.0, 65.9, 61.9, 61.8, 61.0, 46.4, 39.5, 39.0, 36.1, 36.0, 27.0, 26.1,
19.4, 18.5, 14.9, 14.8, ꢀ5.1 ppm (includes both anti-diastereomers); IR
(thin film): n˜ =3088, 3058, 3015, 2953, 2929, 2880, 2856, 2361, 2341, 2320,
1958, 1818, 1776, 1656, 1597, 1568, 1542, 1490, 1472, 1447, 1428, 1390,
1361, 1325, 1256, 1216, 1185, 1157, 1113, 1081, 1032, 1010, 935, 898, 837,
(S)-3-[(R)-2-(tert-Butyldimethylsilanyloxy)-1-methylethyl]-5-[(S)-1-(tert-
butyldimethylsilanyloxymethyl)-2-trityloxyethyl]-4,5-dihydroisoxazole
(57): To unpurified primary alcohol 42 (0.250 mmol, 1.00 equivalent) in
dry pyridine (2 mL) at RT was added TrCl (209 mg, 0.750 mmol,
3.00 equiv). The mixture was stirred at RT for 20 h before all volatiles
were removed under reduced pressure. Last traces of pyridine were re-
moved by azeotropic coevaporation with cyclohexane (3ꢂ4 mL). Purifi-
cation by flash chromatography (hexane/EtOAc 10:1) provided 57
(104 mg, 62% over two steps, dr 6:1 by integration of the signals at
4.62 ppm (major) and 4.55 ppm (minor) in the crude 1H NMR spectrum)
1
as a pale, thick oil. H NMR (300 MHz, CDCl3): d = 7.43–7.39 (m, 6H),
7.31–7.18 (m, 9H), 4.62 (ddd, 1H, J=9.3, 9.3, 7.2 Hz), 3.77–3.55 (m, 4H),
3.25 (d, 2H, J=5.6 Hz), 2.95–2.77 (m, 2H), 2.76–2.59 (m, 1H), 2.09–1.99
(m, 1H), 1.11 (d, 1.5H, J=6.9 Hz), 1.08 (d, 1.5H, 7.2 Hz), 0.88 (s, 9H),
0.82 (s, 9H), 0.04 (s, 6H), ꢀ0.02 ppm (s, 6H) (includes both anti-diaste-
reomers); 13C NMR (75 MHz, CDCl3): d = 161.5, 144.2, 128.8, 127.8,
126.9, 86.7, 79.2, 66.2, 66.1, 61.2, 46.2, 39.4, 38.9, 36.2, 36.0, 26.1, 18.4,
14.9, 14.4, ꢀ5.1, ꢀ5.2 ppm (includes both anti-diastereomers); IR (thin
film): n˜ =3089, 3068, 3034, 2955, 2929, 2884, 2857, 2735, 2247, 1958, 1812,
1598, 1491, 1472, 1462, 1449, 1389, 1362, 1316, 1255, 1218, 1184, 1088,
1006, 977, 910, 837, 814, 776, 734, 706, 668, 648, 632 cmꢀ1; HRMS
(MALDI): m/z: calcd for C40H59NO4Si2Na [M+Na]+: 696.3875; found:
696.3868.
775, 762, 744, 700, 633 cmꢀ1
;
HRMS (ESI): m/z: calcd for
C50H63NO4Si2Na [M+Na]+: 820.4188; found: 820.4176.
(S)-3-(tert-Butyldimethylsilanyloxy)-2-{(S)-3-[(R)-2-(tert-butyldimethylsi-
lanyloxy)-1-methyl-ethyl]-4,5-dihydroisoxazol-5-yl}-propan-1-ol (42): To a
solution of oxime 1 (109 mg, 0.500 mmol, 1.00 equiv) in CH2Cl2 (5 mL) at
ꢀ788C was added tBuOCl (59 mL, 0.500 mmol, 1.00 equiv) dropwise over
10 min. The resulting deep blue solution was stirred 1 h at ꢀ788C. In a
separate flask EtMgBr (3.0m in Et2O, 500 mL, 1.50 mmol, 3.00 equiv) was
added dropwise to a solution of monoprotected homoallylic diol 32d
(141 mg, 0.65 mmol, 1.30 equiv) and iPrOH (126 mL, 1.65 mmol,
3.30 equiv) in CH2Cl2 (10 mL) at 08C. The reaction mixture turned mo-
mentarily cloudy and became clear and colorless upon stirring at 08C for
30 min. At this time, the hydroximinoyl chloride in CH2Cl2 was added to
the reaction dropwise via cannula over 60 min, followed by two rinses
(1 mL each). The slightly yellow solution was stirred at 08C for 6 h and
was slowly allowed to warm to room temperature overnight (large
Dewar with ice, covered with aluminum foil). After 18 h the reaction was
quenched by addition of saturated, aqueous NH4Cl (15 mL). The mixture
was vigorously stirred for 20 min, then the layers were separated and the
aqueous phase was extracted with EtOAc (3ꢂ15 mL). The combined or-
ganic solutions were washed with brine, dried over anhydrous Na2SO4,
filtered, and concentrated under reduced pressure. Half of the crude ma-
terial was used for the determination of the dr by derivatization to the
(R)-2-{(S)-3-[(R)-2-(tert-Butyldimethylsilanyloxy)-1-methyl-ethyl]-4,5-di-
hydroisoxazol-5-yl}-propionaldehyde oxime (24): To the diastereomerical-
ly pure primary alcohol 2 (85.1 mg, 0.282 mmol, 1.00 equiv) in CH2Cl2
(2.5 mL) at 08C was added TEMPO (0.900 mg, 5.60 mmol, 2.00 mol%)
and KBr (3.40 mg, 28.2 mmol, 10.0 mol%). The mixture was vigorously
stirred and NaOCl (0.61m in H2O, 509 mL, 0.310 mmol, 1.10 equiv) in
pH 8.6 buffer (1.8 mL) was added in portions. After 15 min TLC analysis
indicated complete consumption of the starting material. The reaction
was quenched by addition of MeOH (100 mL). CH2Cl2 (10 mL) and brine
(10 mL) were added. The layers were separated and the aqueous phase
extracted with CH2Cl2 (3ꢂ10 mL). The combined organic solutions were
washed with brine, dried over anhydrous Na2SO4, filtered, and concen-
trated under reduced pressure to give the intermediate aldehyde. The re-
sulting colorless oil was used immediately without further purification.
To a solution of the unpurified aldehyde in EtOH (2.2 mL) at RT was
added NH2OH·HCl (29.4 g, 0.423 mmol, 1.50 equiv) in pyridine (0.3 mL).
Chem. Eur. J. 2009, 15, 12065 – 12081
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12077

E. M. Carreira and N. Lohse-Fraefel
The mixture was stirred at RT for 12 h and subsequently concentrated
under reduced pressure. To the resulting residue was added EtOAc
(20 mL) and H2O (10 mL). The layers were separated and the organic
layer was washed with H2O (10 mL), brine (10 mL), dried over anhy-
drous Na2SO4, filtered, and concentrated. The pyridine was removed by
azetropical coevaporation with cyclohexane (3ꢂ10 mL). Purification by
flash chromatography (hexane/EtOAc 4:1) gave oxime 24 (84 mg, 95%
over two steps) as a clear, colorless oil. [a]2D5 (c = 0.64, CHCl3)=+77.48;
1H NMR (300 MHz, CDCl3, * denotes signal corresponding to the minor
oxime diastereomer): d = 9.04* (brs, 1H), 8.54 (brs, 1H), 7.29 (d, 1H,
J=7.2 Hz), 6.62 (d, 1H, J=7.8 Hz), 4.56 (ddd, 1H, J=10.6, 7.2, 4.7 Hz),
3.68–3.58 (m, 2H), 3.41–3.30* (m, 1H), 3.00 (dd, 1H, J=17.4, 10.6 Hz),
2.82–2.70 (m, 1H), 2.73 (dd, 1H, J=16.8, 6.9 Hz), 2.57–2.46 (m, 1H),
1.13 (d, 3H, J=7.2 Hz), 1.11 (d, 3H, J=7.2 Hz), 0.87 (s, 9H), 0.03 ppm
(s, 6H); 13C NMR (75 MHz, CDCl3, * denotes signal corresponding to
the minor oxime diastereomer): d = 161.4*, 161.2, 152.2*. 152.0, 81.5,
81.3*, 66.0, 39.4, 39.3*, 38.9, 35.8, 34.4*, 26.0, 18.4, 14.8, 14.7*, 14.3*, 14.1,
ꢀ5.2, ꢀ5.3 ppm; IR (thin film): n˜ =3324, 2955, 2930, 2886, 2857, 2368,
2360, 2341, 1648, 1624, 1587, 1462, 1436, 1408, 1389, 1356, 1300, 1256,
1220, 1182, 1098, 1046, 1007, 980, 938, 837, 815, 777, 668 cmꢀ1; HRMS
(ESI): m/z: calcd for C15H31N2O3Si [M+H]+: 315.2099; found: 315.2094.
(22 mg, 51%) as a = 1.26, CHCl3)=+17.48;
colorless oil. [a]3D0 (c
1H NMR (300 MHz, CDCl3, * denotes minor diastereomeric peak): d =
7.32–7.18 (m, 2H), 6.93–6.82 (m, 2H), 4.59* (ddd, 1H, J=13.1, 8.1,
5.0 Hz), 4.49 (ddd, 1H, J=15.9, 8.4, 6.9 Hz), 4.42 (s, 2H), 3.81* (s, 3H),
3.80 (s, 3H), 3.70–3.56 (m, 2H), 3.48* (dd, 1H, J=9.3, 5.3 Hz), 3.47 (dd,
1H, J=9.4, 5.6 Hz), 3.40 (dd, 1H, J=9.0, 5.9 Hz), 3.39* (dd, 1H, J=9.3,
6.2 Hz), 2.91 (dd, 1H, J=17.1, 10.6 Hz), 2.73 (dd, 1H, J=17.4, 8.7 Hz),
2.82–2.68 (m, 1H), 2.12–1.98 (m, 1H), 1.14 (d, 3H, J=6.9 Hz), 0.93 (d,
3H, J=6.8 Hz), 0.88 (s, 9H), 0.04 ppm (s, 6H); 13C NMR (75 MHz,
CDCl3, * denotes minor diastereomeric peak): d = 161.3, 159.0, 130.6,
129.1, 113.7, 81.5, 80.7*, 72.9, 72.5*, 71.9, 66.2, 55.4, 38.7*, 38.4*, 38.2,
37.9, 36.0*, 35.9, 26.0, 18.4, 14.9, 13.1*, 13.0, ꢀ5.2, ꢀ5.3 ppm; IR (thin
film): n˜ = 2956, 2931, 2901, 2857, 2740, 2547, 2362, 2062, 1882, 1715,
1613, 1586, 1514, 1464, 1442, 1389, 1361, 1302, 1249, 1174, 1096, 1037,
1008, 981, 939, 900, 874, 837, 778, 705, 679, 637 cmꢀ1; elemental analysis
calcd (%) for C23H39NO4Si: C 65.52, H 9.32, N 3.32; found: C 65.36, H
9.48,
N
3.60; HRMS (ESI): m/z: calcd for C23H40NO4Si [M+H]+:
422.2721; found: 422.2715.
(2R,5S,6S)-1-(tert-Butyldimethylsilanyloxy)-5-hydroxy-7-(4-methoxyben-
zyloxy)-2,6-dimethylheptan-3-one (44): To a solution of isoxazoline 43
(200 mg, 0.474 mmol) in 5:1 MeOH/H2O (18 mL) was added boric acid
(445 mg, 7.20 mmol) and W-2 Raney Nickel (20 mg). The reaction was
purged with H2 and vigorously stirred for 45 min. The mixture was fil-
tered through Celite, concentrated under reduced pressure and immedi-
ately purified by flash chromatography (hexane/EtOAc 15:1 to 10:1) to
afford b-hydroxy ketone 44 (189 mg, 94%) as a colorless oil. [a]3D0 (c =
0.97, CHCl3)=ꢀ32.78; 1H NMR (300 MHz, CDCl3, * denotes minor dia-
stereomeric peak): d = 7.26–7.21 (m, 2H), 6.88–6.84 (m, 2H), 4.42 (s,
2H), 4.07–3.98 (brs, 1H), 3.79 (s, 3H), 3.73* (dd, 1H, J=9.7, 7.8 Hz),
3.72 (dd, 1H, J=9.7, 7.5 Hz), 3.62 (dd, 1H, J=10.0, 5.3 Hz), 3.61* (dd,
1H, J=9.7, 5.3 Hz), 3.53–3.39 (m, 2H), 2.84–2.71 (m, 1H), 2.68–2.53 (m,
2H), 1.94–1.81 (m, 1H), 1.02* (d, 3H, J=7.2 Hz), 1.01 (d, 3H, J=
7.2 Hz), 0.92 (d, 3H, J=7.2 Hz), 0.91* (d, 3H, J=7.5 Hz), 0.86 (s, 9H),
0.03 (s, 3H), 0.02 ppm (s, 3H); 13C NMR (75 MHz, CDCl3, * denotes
minor diastereomeric peak): d = 214.6*, 214.4, 159.1, 130.3, 129.2, 113.8,
73.1, 73.0, 70.8*, 70.7, 69.0*, 65.7, 55.4, 49.3, 47.5, 47.0*, 38.6, 38.3*, 26.0,
18.4, 14.0, 13.9*, 13.0, 11.5*, ꢀ5.3 ppm; IR (thin film): n˜ =3500, 2955,
2929, 2904, 2884, 2856, 1706, 1612, 1586, 1513, 1462, 1407, 1386, 1361,
1301, 1246, 1088, 1035, 1006, 939, 834, 775, 667 cmꢀ1; elemental analysis
calcd (%) for C23H40O5Si: C 65.05, H 9.49; found: C 64.94, H 9.48;
HRMS (ESI): m/z: calcd for C23H40O5SiNa [M+Na]+: 447.2537; found:
447.2531.
(R)-2-[(S)-3-((R)-1-{3-[(R)-2-(tert-Butyldimethylsilanyloxy)-1-methyleth-
yl]-4,5-dihydroisoxazol-5-yl}-ethyl)-4,5-dihydroisoxazol-5-yl]-propan-1-ol
(25): To a solution of oxime 24 (118 mg, 0.374 mmol, 1.00 equiv) in
CH2Cl2 (3.7 mL) at ꢀ788C was added tBuOCl (47.0 mL, 0.412 mmol,
1.10 equiv) dropwise over 10 min. The resulting green solution was
stirred 3 h at ꢀ788C. In a separate flask EtMgBr (3.0m in Et2O, 374 mL,
1.12 mmol, 3.00 equiv) was added dropwise to a solution of (S)-2-methyl-
3-butenol (41.9 mg, 0.487 mmol, 1.30 equiv) and iPrOH (95.0 mL,
1.23 mmol, 3.30 equiv) in CH2Cl2 (7.4 mL) at 08C. The reaction mixture
turned momentarily cloudy and became clear and colorless upon stirring
at 08C for 30 min. At this time, the hydroximinoyl chloride in CH2Cl2
was added to the reaction dropwise via cannula over 90 min, followed by
two rinses (1 mL each). The slightly yellow solution was stirred at 08C
for 6 h and was slowly allowed to warm to room temperature overnight
(large Dewar with ice, covered with aluminum foil). After 15 h the reac-
tion was quenched by addition of saturated, aqueous NH4Cl (15 mL).
The mixture was vigorously stirred for 20 min, then the layers were sepa-
rated and the aqueous phase was extracted with EtOAc (3ꢂ15 mL). The
combined organic solutions were washed with brine, dried over anhy-
drous Na2SO4, filtered, and concentrated under reduced pressure. Purifi-
cation by flash chromatography (hexane/EtOAc 2:1 to 1:1) provided is-
ACHTUNGTRENNUNG
(R)-4-(tert-Butyldimethylsilanyloxy)-1-[(2R,4S,5R)-2-(4-methoxyphenyl)-
5-methyl-[1,3]dioxan-4-yl]-3-methylbutan-2-one (49): To b-hydroxy
ketone 44 (123 mg, 0.289 mmol, 1.00 equiv) and 4 ꢃ molecular sieves
(300 mg) in CH2Cl2 (3 mL) at ꢀ158C was added DDQ (78.8 mg,
0.347 mmol, 1.20 equiv). The mixture was stirred at ꢀ158C for 20 h
before the suspension was filtered over silica gel, and concentrated under
reduced pressure. Purification by chromatotron (hexane/EtOAc 30:1) af-
forded 49 (92 mg, 75%) as a colorless oil. [a]2D6 (c = 0.40, CHCl3)=
ꢀ52.28; 1H NMR (300 MHz, CDCl3, * denotes minor diastereomeric
peak): d = 7.40–7.33 (m, 2H), 6.88–6.83 (m, 2H), 5.49* (s, 1H), 5.46 (s,
1H), 4.15–3.97 (m, 2H), 3.79* (s, 3H), 3.79 (s, 3H), 3.78–3.49 (m, 3H),
2.93–2.76 (m, 1H), 2.88 (dd, 1H, J=16.2, 8.1 Hz), 2.71* (dd, 1H, J=16.1,
4.1 Hz), 2.70 (dd, 1H, J=16.3, 3.4 Hz), 1.94–1.78 (m, 1H), 1.18* (d, 3H,
J=6.9 Hz), 1.05* (d, 3H, J=7.2 Hz), 1.01 (d, 3H, J=6.8 Hz), 0.87 (s,
9H), 0.86* (s, 9H), 0.77 (d, 3H, J=6.9 Hz), 0.04 (s, 3H), 0.02 (s, 3H),
0.02* (s, 3H), 0.01* ppm (s, 3H); 13C NMR (75 MHz, CDCl3, * denotes
=
(300 MHz, CDCl3): d = 4.64 (ddd, 1H, J=12.5, 7.5, 5.0 Hz), 4.37 (ddd,
1H, J=10.0, 9.0, 9.0 Hz), 3.70–3.58 (m, 4H), 3.08 (dd, 1H, J=17.4,
10.0 Hz), 3.05 (dd, 1H, J=17.1, 10.6 Hz), 2.87 (dd, 1H, J=17.4, 7.5 Hz),
2.77 (dd, 1H, J=17.1, 9.3 Hz), 2.84–2.71 (m, 2H), 2.27 (brs, 1H), 1.97–
1.83 (m, 1H), 1.23 (d, 3H, J=7.2 Hz), 1.14 (d, 3H, J=6.9 Hz), 0.90 (d,
3H, J=7.4 Hz), 0.88 (s, 9H), 0.05 ppm (s, 6H); 13C NMR (75 MHz,
CDCl3): d = 162.0, 160.3, 84.4, 82.0, 66.3, 66.0, 40.4, 40.1, 39.1, 37.7, 35.9,
26.0, 18.3, 14.8, 14.4, 13.4, ꢀ5.3, ꢀ5.4 ppm; IR (thin film): n˜ =3430, 2956,
2929, 2872, 2857, 1720, 1621, 1582, 1462, 1435, 1388, 1357, 1322, 1293,
1257, 1219, 1093, 1046, 1013, 974, 940, 838, 812, 777, 670 cmꢀ1; HRMS
(ESI): m/z: calcd for C20H38N2O4SiNa [M+Na]+: 421.2493; found:
421.2486.
(S)-3-[(S)-2-(tert-Butyldimethylsilanyloxy)-1-methylethyl]-5-[(S)-2-(4-me-
thoxybenzyloxy)-1-methylethyl]-4,5-dihydroisoxazole (43): To isoxazoline
(310 mg, 1.03 mmol, 1.00 equiv) and CSA (23.8 mg, 0.103 mmol,
minor diastereomeric peak): d
= 211.5, 211.1*, 159.8*, 159.7, 131.3*,
4
131.0, 127.4*, 127.3, 113.6*, 113.5, 101.8*, 100.9, 79.2, 75.5*, 73.6*, 73.0,
65.6, 65.3*, 55.4, 49.7, 49.5*, 46.8, 45.9*, 34.4, 31.4*, 26.1, 18.5, 13.2*,
13.1*, 12.8, 12.6, ꢀ5.3 ppm; IR (thin film): n˜ = 3400, 2956, 2930, 2901,
2856, 2360, 2342, 2028, 1891, 1793, 1715, 1615, 1590, 1518, 1463, 1391,
1366, 1303, 1250, 1214, 1172, 1141, 1114, 1077, 1035, 1008, 979, 949, 938,
910, 891, 836, 777, 668 cmꢀ1; HRMS (ESI): m/z: calcd for C23H38O5SiNa
[M+Na]+: 423.2561; found: 423.2555. NOE-experiments: Irradiation at
5.46 ppm: enhanced signals at 4.05 ppm (strong), 3.79 (weak), 3.54
(strong). Irradiation at 0.77 ppm: enhanced signals at 4.10 ppm
0.100 equiv) in CH2Cl2 (9 mL) at RT was added p-methoxybenzyl 2,2,2-
trichloroacetimidate (466 mg, 1.65 mmol, 1.60 equiv). The mixture was
stirred for 18 h at RT before additional p-methoxybenzyl 2,2,2-trichloro-
ACHTUNGTRENNUNGacetimidate (291 mg, 1.03 mmol, 1.00 equiv) was added. After a total of
64 h H2O (10 mL) was added. The layers were separated and the aqueous
phase was extracted with CH2Cl2 (3ꢂ15 mL). The combined organic solu-
tions were washed with brine, dried over anhydrous Na2SO4, filtered, and
concentrated under reduced pressure. Purification by flash chromatogra-
phy (hexane/EtOAc 15:1 to 10:1) afforded p-methoxybenzyl ether 43
12078
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12065 – 12081

Polyketide Building Blocks
FULL PAPER
(medium), 4.04 ppm (medium), 3.53 ppm (medium), 2.70 ppm (medium),
1.86 ppm (strong).
(hexane/EtOAc 15:1) afforded p-methoxybenzyl ether 47 (97 mg, 52%)
as a colorless oil. [a]2D8 (c = 0.92, CHCl3)=+16.08; 1H NMR (300 MHz,
CDCl3, * denotes minor diastereomeric peak): d = 7.29–7.23 (m, 2H),
6.88–6.85 (m, 2H), 4.58* (ddd, 1H, J=12.8, 8.1, 4.7 Hz), 4.47 (ddd, 1H,
J=15.9, 8.7, 7.2 Hz), 4.42 (s, 2H), 3.80 (s, 3H), 3.49 (s, 2H), 3.48 (dd, 1H,
J=9.3, 5.3 Hz), 3.34 (dd, 1H, J=9.0, 6.2 Hz), 3.33* (dd, 1H, J=9.3,
5.9 Hz), 3.00* (dd, 1H, J=17.1, 10.9 Hz), 2.92 (dd, 1H, J=16.8, 10.3 Hz),
2.78* (dd, 1H, J=16.8, 7.8 Hz), 2.76 (dd, 1H, J=16.8, 8.7 Hz), 2.10–1.98
(m, 1H), 1.95–1.82* (m, 1H), 1.15 (s, 6H), 0.94 (d, 3H, J=6.9 Hz), 0.88
(s, 9H), 0.03 ppm (s, 6H); 13C NMR (75 MHz, CDCl3, * denotes minor
diastereomeric peak): d = 164.2, 164.1*, 159.0, 130.7, 129.1, 113.7, 81.7,
80.9*, 72.9, 72.6*, 71.9, 71.0, 55.4, 38.6, 38.0, 37.8, 26.0, 23.2, 23.1, 18.4,
13.1, 11.8*, ꢀ5.3 ppm; IR (thin film): n˜ =2956, 2991, 2857, 1613, 1586,
1514, 1464, 1391, 1362, 1302, 1249, 1173, 1097, 1037, 1007, 887, 837, 777,
670 cmꢀ1; elemental analysis calcd (%) for C24H41NO4Si: C 66.16, H 9.48,
N 3.21; found: C 66.32, H 9.50, N 3.34; HRMS (MALDI): m/z: calcd for
C24H41NO4SiNa [M+Na]+: 458.2697; found: 458.2694.
(S)-3-[(S)-1-(tert-Butyldimethylsilanyloxy)-ethyl]-5-[(R)-2-(4-methoxy-
benzyloxy)-1-methylethyl]-4,5-dihydroisoxazole (45): To isoxazoline
8
(87 mg, 0.303 mmol, 1.00 equiv) and CSA (7.00 mg, 0.0303 mmol,
0.100 equiv) in CH2Cl2 (2.7 mL) at RT was added p-methoxybenzyl 2,2,2-
trichloroacetimidate (137 mg, 0.484 mmol, 1.60 equiv). The mixture was
stirred for 20 h at RT before additional p-methoxybenzyl 2,2,2-trichloro-
ACHTUNGTRENNUNGacetimidate (137 mg, 0.484 mmol, 1.60 equiv) was added. After a total of
3 d, additional p-methoxybenzyl 2,2,2-trichloroacetimidate (137 mg,
0.484 mmol, 1.60 equiv) was added. After a total of seven days H2O
(10 mL) and CH2Cl2 (10 mL) was added. The layers were separated and
the aqueous phase was extracted with CH2Cl2 (3ꢂ15 mL). The combined
organic solutions were washed with brine, dried over anhydrous Na2SO4,
filtered, and concentrated under reduced pressure. Filtration over a short
pad of silica gel (hexane/EtOAc 2:1), followed by purification by flash
chromatography (hexane/EtOAc 12:1) afforded p-methoxybenzyl ether
45 (56 mg, 46%) as a colorless oil. [a]3D0 (c = 0.33, CHCl3)=ꢀ11.28;
4-(tert-Butyldimethylsilanyloxy)-1-[(2R,4S,5R)-2-(4-methoxyphenyl)-5-
methyl-[1,3]dioxan-4-yl]-3,3-dimethylbutan-2-one (48): To a solution of
isoxazoline 47 (90.0 mg, 0.207 mmol) in 5:1 MeOH/H2O (8.2 mL) was
added boric acid (203 mg, 3.28 mmol) and W-2 Raney Nickel (10 mg).
The reaction was purged with H2 and vigorously stirred for 50 min. The
mixture was filtered through Celite into a separating funnel containing
H2O (5 mL) and CH2Cl2 (10 mL). After separation, the aqueous layer
was extracted with CH2Cl2 (3ꢂ10 mL) and the combined organics were
washed with brine, dried over anhydrous Na2SO4, filtered, and concen-
trated under reduced pressure to give the intermediate b-hydroxy ketone
which was used without further purification. To the unpurified b-hydroxy
ketone (43.2 mg, 98.5 mmol, 1.00 equiv) and 4 ꢃ molecular sieves
(100 mg) in CH2Cl2 (1 mL) at ꢀ158C was added DDQ (26.8 mg,
0.118 mmol, 1.20 equiv). The mixture was stirred at ꢀ158C for 5 h and
3 h at 08C before the suspension was filtered over silica gel and concen-
trated under reduced pressure. Purification by flash chromatography
(hexane/EtOAc 5:1) afforded 48 (30 mg, 34% over two steps) as a color-
less oil. [a]3D1 (c = 0.35, CHCl3)=ꢀ32.98; 1H NMR (300 MHz, CDCl3, *
denotes minor diastereomeric peak): d = 7.41–7.32 (m, 2H), 6.89–6.82
(m, 2H), 5.49* (s, 1H), 5.46 (s, 1H), 4.48* (ddd, 1H, J=6.8, 6.8, 2.5 Hz),
4.16–3.97 (m, 2H), 3.79* (s, 3H), 3.78 (s, 3H), 3.66–3.50 (m, 3H), 2.98
(dd, 1H, J=17.1, 7.8 Hz), 2.92* (dd, 1H, J=11.8, 5.9 Hz), 2.64* (dd, 1H,
J=17.7, 6.9 Hz), 2.61 (dd, 1H, J=17.1, 3.4 Hz), 1.93–1.78 (m, 1H), 1.79–
1.70* (m, 1H), 1.17* (d, 3H, J=7.2 Hz), 1.10 (s, 3H), 1.09 (s, 3H),0.87 (s,
9H), 0.76 (d, 3H, J=6.5 Hz), 0.03 (s, 3H), 0.02 ppm (s, 3H); 13C NMR
1H NMR (300 MHz, CDCl3,
* denotes minor diastereomeric peak)
d7.32–7.22 (m, 2H), 6.90–6.82 (m, 2H), 4.69 (q, 1H, J=5.9 Hz), 4.65–4.51
(m, 1H), 4.43 (s, 2H), 4.39* (s, 2H), 3.81* (s, 3H), 3.80 (s, 3H), 3.48 (dd,
1H, J=9.0, 5.3 Hz), 3.42* (dd, 1H, J=9.0, 5.3 Hz), 3.41 (dd, 1H, J=9.3,
5.9 Hz), 3.35* (dd, 1H, J=9.0, 5.6 Hz), 3.02* (dd, 1H, J=17.4, 10.9 Hz),
2.95 (dd, 1H, J=17.1, 10.6 Hz), 2.80* (dd, 1H, J=17.1, 8.1 Hz), 2.79 (dd,
1H, J=17.4, 8.7 Hz), 2.13–2.00 (m, 1H), 1.99–1.86* (m, 1H), 1.34 (d, 3H,
J=6.5 Hz), 1.32* (d, 3H, J=6.2 Hz), 0.94 (d, 3H, J=6.9 Hz), 0.88 (s,
9H), 0.08 (s, 3H), 0.05 ppm (s, 3H); 13C NMR (75 MHz, CDCl3, * de-
notes minor diastereomeric peak) d161.8, 159.1, 130.6, 129.2, 113.8, 82.0,
81.4*, 73.1*, 73.0, 72.5*, 71.9, 64.8*, 64.7, 55.4, 38.8*, 38.0, 36.4*, 35.2,
25.9, 22.5, 18.3, 12.8, 11.9*, ꢀ4.5, ꢀ4.6 ppm; IR (thin film): n˜ =2954,
2930, 2892, 2856, 1739, 1612, 1586, 1512, 1463, 1441, 1370, 1362, 1301,
1246, 1217, 1172, 1111, 1085, 1034, 1006, 980, 952, 881, 865, 826, 812, 776,
753, 707, 666, 636 cmꢀ1; elemental analysis calcd (%) for C22H37NO4Si: C
64.82, H 9.15, N 3.44; found: C 64.85, H 9.25, N 3.64; HRMS (ESI): m/z:
calcd for C22H38NO4Si [M+H]+: 408.2565; found: 408.2559.
(2S,5S,6R)-2-(tert-Butyldimethylsilanyloxy)-5-hydroxy-7-(4-methoxyben-
zyloxy)-6-methyl-heptan-3-one (46): To
a solution of isoxazoline 45
(50.0 mg, 0.123 mmol) in 5:1 MeOH/H2O (5 mL) was added boric acid
(124 mg, 2.00 mmol) and W-2 Raney Nickel (5 mg). The reaction was
purged with H2 and vigorously stirred for 60 min. The mixture was fil-
tered through Celite, concentrated under reduced pressure and immedi-
ately purified by flash chromatography (hexane/EtOAc 15:1 to 10:1) to
afford b-hydroxy ketone 46 (49 mg, 96%) as a colorless oil. [a]2D6 (c =
0.46, CHCl3)=ꢀ23.38; 1H NMR (300 MHz, CDCl3, * denotes minor dia-
stereomeric peak): d = 7.26–7.21 (m, 2H), 6.89–6.84 (m, 2H), 4.43 (s,
2H), 4.20* (ddd, 1H, J=12.8, 6.9, 3.4 Hz), 4.14 (q, 1H, J=6.8 Hz), 4.03
(ddd, 1H, J=12.5, 6.2, 3.1 Hz), 3.80 (s, 3H), 3.49–3.45 (m, 2H), 2.84 (dd,
1H, J=17.7, 3.1 Hz), 2.83* (dd, 1H, J=17.7, 9.3 Hz), 2.68 (dd, 1H, J=
17.7, 9.3 Hz), 2.64* (dd, 1H, J=17.7, 2.8 Hz), 1.97–1.82 (m, 1H), 1.28 (d,
3H, J=6.5 Hz), 0.95* (d, 3H, J=6.8 Hz), 0.93 (d, 3H, J=6.9 Hz), 0.91 (s,
9H), 0.08 (s, 3H), 0.07 ppm (s, 3H); 13C NMR (75 MHz, CDCl3, * de-
notes minor diastereomeric peak): d = 214.6, 159.2, 130.3, 129.3, 113.8,
75.1, 73.5*, 73.2, 73.1, 70.7, 68.9*, 55.4, 42.0, 41.6*, 38.6, 38.4*, 26.0, 20.9,
18.3, 14.0, 11.6*, ꢀ4.4, ꢀ4.7 ppm; IR (thin film): n˜ =3500, 2954, 2930,
2888, 2856, 1713, 1612, 1586, 1513, 1463, 1443, 1407, 1389, 1363, 1336,
1301, 1247, 1204, 1172, 1114, 1087, 1036, 1006, 937, 831, 777, 705, 668,
(75 MHz, CDCl3, * denotes minor diastereomeric peak): d
= 212.4,
159.9, 131.6*, 131.3, 127.5*, 127.4, 113.8*, 113.6, 101.9*, 101.0, 79.0, 75.9*,
73.7*, 73.0, 70.1*, 69.9, 55.4, 49.9, 42.1, 41.2*, 34.4, 31.3*, 25.9, 21.5*, 21.4,
21.3, 18.3, 12.6, 11.5*, ꢀ5.5 ppm; IR (thin film): n˜ = 2957, 2930, 2895,
2857, 2719, 2361, 2034, 1711, 1616, 1590, 1519, 1464, 1393, 1365, 1343,
1303, 1251, 1171, 1143, 1114, 1036, 1010, 984, 938, 912, 837, 777, 668 cmꢀ1
;
HRMS (ESI): m/z: calcd for C24H40O5SiNa [M+Na]+: 459.2537; found:
459.2530. NOE-experiments: Irradiation at 5.49 ppm: enhanced signals at
4.48 ppm (strong), 3.78 ppm (medium) and 3.53 ppm (strong). Irradiation
at 5.46 ppm: enhanced signals at 4.10 ppm (strong), 3.78 ppm (weak) and
3.53 ppm (strong). Irradiation at 4.48 ppm: enhanced signals at 5.49 ppm
(strong), 3.78 ppm (weak) and 3.53 ppm (weak). Irradiation at 2.98 ppm:
enhanced signals at 3.62 ppm (medium), 2.61 ppm (strong) and 1.10
(weak). Irradiation at 1.75 ppm: enhanced signals at 4.48 ppm (medium),
4.10 (medium), 3.97 (medium), 3.79 (weak), 1.17 (medium). Irradiation
at 1.86 ppm: 4.10 (weak), 2.98 (weak), 2.61 (weak), 1.10 (weak), 0.76
(medium). Irradiation at 0.76 ppm: enhanced signals at 4.10 ppm
(medium), 3.53 ppm (medium), 2.61 ppm (medium), 1.93 ppm (strong).
Irradiation at 1.17 ppm: 3.97 (weak), 3.65 ppm (weak), 3.53 ppm (weak),
2.92 ppm (weak), 2.64 ppm (weak).
636 cmꢀ1
;
HRMS (ESI): m/z: calcd for C22H38O5SiNa [M+Na]+:
433.2381; found: 433.2374.
(S)-3-[2-(tert-Butyldimethylsilanyloxy)-1,1-dimethylethyl]-5-[(R)-2-(4-
methoxybenzyloxy)-1-methylethyl]-4,5-dihydroisoxazole (47): To primary
alcohol
6
(135 mg, 0.428 mmol, 1.00 equiv) and CSA (9.90 mg,
0.0428 mmol, 0.100 equiv) in CH2Cl2 (2.7 mL) at RT was added p-me-
thoxybenzyl 2,2,2-trichloroacetimidate (242 mg, 0.856 mmol, 2.00 equiv).
The mixture was stirred for 63 h at RT before H2O (10 mL) and CH2Cl2
(10 mL) was added. The layers were separated and the aqueous phase
was extracted with CH2Cl2 (3ꢂ10 mL). The combined organic solutions
were washed with brine, dried over anhydrous Na2SO4, filtered, and con-
centrated under reduced pressure. Purification by flash chromatography
(S)-2-[(S)-3-(tert-Butyldiphenylsilanyloxymethyl)-4,5-dihydroisoxazol-5-
yl]-propionic acid methyl ester (50): To primary alcohol 12 (152 mg,
0.382 mmol, 1.00 equiv) in CH2Cl2 (4 mL) at 08C was added TEMPO
(1.20 mg, 7.60 mmol, 2.00 mol%) and KBr (4.50 mg, 38.2 mmol,
10.0 mol%). The mixture was vigorously stirred and NaOCl (approx
1.5m in H2O, 382 mL, 0.573 mmol, 1.50 equiv) in pH 8.6 buffer (3 mL)
was added in portions. Additional aq. NaOCl solution (400 mL,
Chem. Eur. J. 2009, 15, 12065 – 12081
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12079

E. M. Carreira and N. Lohse-Fraefel
0.600 mmol, 1.57 equiv) was added after 10 min. After a total of 1.5 h
TLC analysis indicated complete consumption of the starting material.
The reaction was quenched by addition of MeOH (200 mL). H2O (10 mL)
and CH2Cl2 (10 mL) were added, the layers separated, and the aqueous
phase extracted with CH2Cl2 (3ꢂ10 mL). The combined organic solutions
were washed with brine, dried over anhydrous Na2SO4, filtered, and con-
centrated under reduced pressure to give the intermediate aldehyde
which was used without further purification. A solution of NaClO2
(207 mg, 2.29 mmol, 6.00 equiv) and 2-methyl-2-butene (404 mL,
3.82 mmol, 10.0 equiv) in pH 3.8 buffer (2.3 mL) and tBuOH (11.1 mL)
was added at 08C to the unpurified aldehyde. The mixture was stirred at
08C for 50 min, before pH 3.8 buffer (6 mL) and EtOAc (10 mL) were
added. The layers were separated and the aqueous phase was extracted
with EtOAc (3ꢂ10 mL). The combined organic solutions were dried over
anhydrous Na2SO4, filtered, and concentrated under reduced pressure to
give the intermediate carboxylic acid which was used without further pu-
rification. To the unpurified carboxylic acid in benzene/MeOH 7:1
(16 mL) at 08C was added dropwise TMSCH2N2 (2.0m in hexane,
1.53 mL, 3.06 mmol, 8.00 equiv). The yellow solution was stirred for
15 min at 08C before being quenched with acetic acid (dropwise addition
until solution stayed colorless). The volatiles were removed under re-
duced pressure. Purification by flash chromatography (hexane/EtOAc
8:1) provided methyl ester 50 (113 mg, 70% yield over three steps) as a
1.00 equiv) and benzyl 2,2,2-trichloroacetimidate (44.0 mL, 0.235 mmol,
2.00 equiv) in cyclohexane (0.5 mL) and CH2Cl2 (0.25 mL) at RT was
added trifluoromethanesulfonic acid (5.00 mL, 0.0572 mmol, 0.487 equiv).
The mixture was stirred for 15 h at RT. The crystalline trichloroacet-
AHCTUNGTREGaNNNU mide was removed by filtration (washings with pentane (10 mL), Et2O
dissolves the amide). The filtrate was washed with saturated, aqueous
NaHCO3 solution (5 mL) and brine, dried over anhydrous Na2SO4, fil-
tered, and concentrated under reduced pressure. Purification by flash
chromatography (hexane/EtOAc 4:1) provided isoxazoline 52 (20 mg,
62%) as a colorless oil. 1H NMR (300 MHz, CDCl3): d = 7.40–7.28 (m,
5H), 4.74 (ddd, 1H, J=10.6, 7.5, 7.5 Hz), 4.53 (s, 2H), 4.30 (s, 2H), 3.70
(s, 3H), 3.17 (dd, 1H, J=17.4, 10.6 Hz), 2.89 (dd, 1H, J=17.4, 7.5 Hz),
2.73–2.63 (m, 1H), 1.29 ppm (d, 3H, J=7.1 Hz); 13C NMR (75 MHz,
CDCl3): d = 173.9, 156.7, 137.3, 128.6, 128.1, 128.0, 81.5, 72.9, 64.7, 52.2,
44.4, 39.4, 13.8 ppm; HRMS (MALDI): m/z: calcd for C15H19NO4Na
[M+Na]+: 278.1387; found: 278.1393. These spectral characteristics are
identical in all respects to those previously reported.[35]
Acknowledgements
We thank ETH and the Swiss National Science Foundation for support
of the research described.
colorless oil. [a]3D4 (c
=
0.58, CHCl3)=+29.88; 1H NMR (300 MHz,
CDCl3, * denotes minor diastereomeric peak): d = 7.67–7.64 (m, 4H),
7.45–7.37 (m, 6H), 4.85 (ddd, 1H, J=10.9, 7.5, 7.5 Hz), 4.72* (ddd, 1H,
J=10.6, 7.5, 7.5 Hz), 4.43 (s, 2H), 3.72 (s, 3H), 3.71* (s, 3H), 3.17* (dd,
1H, J=17.4, 10.6), 3.10 (dd, 1H, J=17.7, 10.9), 2.86* (dd, 1H, J=17.4,
7.5 Hz), 2.85 (dd, 1H, J=17.4, 7.8 Hz), 2.77 (dq, 1H, J=7.0, 7.0 Hz),
2.64* (dq, 1H, J=7.0, 7.0 Hz), 1.28* (d, 3H, J=7.2 Hz), 1.15 (d, 3H, J=
7.2 Hz), 1.07 ppm (s, 9H); 13C NMR (75 MHz, CDCl3, * denotes minor
[1] a) D. E. Cane, C. T. Walsh, C. Khosla, Science 1998, 282, 63–68;
b) J.-Y. Ortholand, A. Ganesan, Curr. Opin. Chem. Biol. 2004, 8,
271–280.
[2] For aldol addition approaches to polyketide building blocks, see:
a) D. A. Evans, J. V. Nelson, T. R. Taber, Top. Stereochem. 1982, 13,
1; b) I. Paterson, J. A. Cannon, Tetrahedron Lett. 1992, 33, 797; c) I.
Paterson, in Modern Carbonyl Chemistry (Ed.: J. Otera), Wiley-
VCH, Weinheim, 2000; pp. 249–297; d) E. M. Carreira, in Compre-
hensive Asymmetric Catalysis, Vol. III (Eds.: E. N. Jacobsen, A.
Pfaltz, H. Yamamoto), Springer, Heidelberg, 1999, pp. 997–1065;
e) C. H. Heathcock, in Comprehensive Organic Synthesis, Vol. 2
(Eds.: B. Trost, I. Fleming), Pergamon, Oxford, 1991, pp. 133–238;
f) C. J. Cowden, I. Paterson, in Organic Reactions, Vol. 51 (Ed.:
L. A. Paquette), Wiley, New York, 1997, pp. 1–200. For the use of
allylation, crotylation and allenylmetal reagents in polyketide syn-
thesis, see: g) S. E. Denmark, N. G. Almstead, in Modern Carbonyl
Chemistry (Ed.: J. Otera), Wiley-VCH, Weinheim, 2000, pp. 299–
401; h) C. S. Chemler, W. R. Roush, in Modern Carbonyl Chemistry
(Ed.: J. Otera), Wiley-VCH, Weinheim, 2000, pp. 403–490; i) J. A.
Marshall, Z.-H. Lu, B. A. Johns, J. Org. Chem. 1998, 63, 817.
[3] For examples, see: a) A. B. Smith III; C. M. Adams, Acc. Chem. Res.
2004, 37, 365; C. M. Adams, Acc. Chem. Res. 2004, 37, 365; b) C.
Meyer, N. Blanchard, M. Defosseux, J. Cossy, J. Acc. Chem. Res.
2003, 36, 766; c) M. Lautens, J.-F. Paquin, Org. Lett. 2003, 5, 3391;
d) A. M. Misske, H. M. R. Hoffmann, Chem. Eur. J. 2000, 6, 3313;
e) B. Breit, S. K. Zahn, J. Org. Chem. 2001, 66, 4870; f) D. Myles,
S. J. Danishefsky, Pure Appl. Chem. 1989, 61, 1235.
[4] a) S. Kanemasa, S. Kobayashi, M. Nishiuchi, H. Yamamoto, E.
Wada, Tetrahedron Lett. 1991, 32, 6367; b) S. Kanemasa, M. Nishiu-
chi, E. Wada, Tetrahedron Lett. 1992, 33, 1357; c) S. Kanemasa, M.
Nishiuchi, A. Kamimura, K. Hori, J. Am. Chem. Soc. 1994, 116,
2324.
[5] a) J. W. Bode, N. Fraefel, D. Muri, E. M. Carreira, Angew. Chem.
2001, 113, 2128; Angew. Chem. Int. Ed. 2001, 40, 2082; b) P. Asch-
wanden, L. Kvaerno, R. W. Geisser, F. Kleinbeck, E. M. Carreira,
Org. Lett. 2005, 7, 5741.
[6] a) J. W. Bode, E. M. Carreira, J. Am. Chem. Soc. 2001, 123, 3611;
b) J. W. Bode, E. M. Carreira, J. Org. Chem. 2001, 66, 6410.
[7] a) D. Muri, N. Lohse-Fraefel, E. M. Carreira, Angew. Chem. 2005,
117, 4104; Angew. Chem. Int. Ed. 2005, 44, 4036; b) A. M. Szpilman,
D. M. Cereghetti, N. Wurtz, J. M. Manthorpe, E. M. Carreira,
Angew. Chem. 2008, 120, 4407; Angew. Chem. Int. Ed. 2008, 47,
4335; c) A. M. Szpilman, J. M. Manthorpe, E. M. Carreira, Angew.
diastereomeric peak): d
= 173.9*, 173.8, 158.2*, 157.9, 135.5, 132.6,
130.0, 127.9, 81.1*, 81.0, 59.2, 52.1, 44.3*, 43.6, 39.1*, 37.5, 29.9, 19.4,
13.6*, 12.1 ppm; IR (thin film): n˜ =3437, 3072, 3049, 2954, 2933, 2892,
2859, 1963, 1893, 1831, 1740, 1629, 1606, 1594, 1502, 1472, 1462, 1429.
1379, 1362, 1338, 1263, 1203, 1166, 1113, 998, 939, 873, 824, 800, 742, 703,
612 cmꢀ1; elemental analysis calcd (%) for C24H31NO4Si: C 67.73, H 7.34,
N 3.29; found: C; 67.58H, 7.57, N 3.54; HRMS (MALDI): m/z: calcd for
C24H31NO4SiNa [M+Na]+: 448.1915; found: 448.1914.
(S)-2-((S)-3-Hydroxymethyl-4,5-dihydroisoxazol-5-yl)-propionic
acid
methyl ester (51): To TBDPS ether 50 (111 mg, 0.260 mmol, 1.00 equiv)
in THF (2.5 mL) was added at 08C TBAF (1.0m in THF, 0.780 mL,
0.780 mmol, 3.00 equiv). The mixture was stirred at 08C for 4 h before
additional TBAF (1.0m in THF, 0.780 mL, 0.780 mmol, 3.00 equiv) was
added. The mixture was stirred for 12 h at RT. The reaction was
quenched by the addition of saturated, aqueous NaHCO3 solution
(10 mL) and Et2O (10 mL) was added. The layers were separated and the
aqueous phase was extracted with Et2O (3ꢂ10 mL). The combined or-
ganic solutions were dried over anhydrous Na2SO4, filtered, and concen-
trated under reduced pressure. Purification by flash chromatography
(hexane/EtOAc 1:1) provided 51 (128 mg, 85%) as a clear, colorless oil.
A second purification by flash chromatography (hexane/EtOAc 5:1) re-
sulted in separation of the diastereomers. [a]3D4 (c = 0.61, CHCl3)=+
1
13.58; H NMR (300 MHz, CDCl3, * denotes minor diastereomeric peak):
d = 4.88 (ddd, 1H, J=10.6, 7.5, 7.5 Hz), 4.76* (ddd, 1H, J=10.6, 7.5,
7.5 Hz), 4.40 (s, 2H), 3.71 (s, 3H), 3.70* (s, 3H), 3.19* (dd, 1H, J=17.4,
10.6 Hz), 3.12 (dd, 1H, J=17.4, 10.9 Hz), 2.93* (dd, 1H, J=17.4, 7.5 Hz),
2.88 (dd, 1H, J=17.4, 7.8 Hz), 2.79 (dq, 1H, J=7.2, 7.2 Hz), 2.71* (dq,
1H, J=7.2, 7.2 Hz), 2.18 (brs, 1H), 1.28* (d, 3H, J=7.2 Hz), 1.17 ppm
(d, 3H, J=7.2 Hz); 13C NMR (75 MHz, CDCl3, * denotes minor diaste-
reomeric peak): d = 174.0*, 173.8, 158.6*, 158.2, 81.7*, 81.5, 58.3, 52.2,
44.2*, 43.7, 38.8*, 37.6, 13.8*, 12.4 ppm; IR (thin film): n˜ =3419, 2978,
2954, 2885, 2361, 2342, 1734, 1629, 1460, 1437, 1380, 1337, 1266, 1206,
1168, 1051, 991, 922, 867, 668 cmꢀ1; elemental analysis calcd (%) for
C8H13NO4: C 51.33, H 7.00, N 7.48; found: C 51.40, H 7.10, N 7.62;
HRMS (MALDI): m/z: calcd for C8H13NO4Na [M+Na]+: 210.0737;
found: 210.0733.
(S)-2-((R)-3-Benzyloxymethyl-4,5-dihydroisoxazol-5-yl)-propionic
acid
methyl ester (52): To a solution of alcohol 51b (22.0 mg, 0.118 mmol,
12080
www.chemeurj.org
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12065 – 12081

Polyketide Building Blocks
FULL PAPER
Chem. 2008, 120, 4411; Angew. Chem. Int. Ed. 2008, 47, 4339;
Angew. Chem. 2008, 120, 4411; d) F. Kleinbeck, E. M. Carreira,
Angew. Chem. 2009, 121, 586; Angew. Chem. Int. Ed. 2009, 48, 578;
Angew. Chem. 2009, 121, 586.
[20] For transformations of similar bis(isoxazolines)to the corresponding
syn- and anti-diols, see ref. [8].
[21] M. G. Kociolek, C. Honga, Tetrahedron Lett. 2003, 44, 1811.
[22] For the synthesis of the corresponding THP derivative, see: A. P.
Kozikowski, J. Scripko, J. Am. Chem. Soc. 1984, 106, 353.
[23] R. Appel, Angew. Chem. 1975, 87, 863; Angew. Chem. Int. Ed. Engl.
1975, 14, 801.
[8] L. D. Fader, E. M. Carreira, Org. Lett. 2004, 6, 2485.
[9] This work has been communicated in preliminary form: N. Lohse-
Fraefel, E. M. Carreira, Org. Lett. 2005, 7, 2011.
[10] For the enantioselective synthesis of (S)-2-methyl-3-butenol, see:
M. J. Batchelor, R. J. Gillespie, J. M. C. Golec, C. J. R. Hedgecock,
S. D. Jones, R. Murdoch, Tetrahedron 1994, 50, 809.
[24] a) This synthesis has been previously described, see: J. ꢃhman, P.
Somfai, Synth. Commun. 1994, 24, 1117; b) although the iodide is
unstable towards prolonged storage at room temperature, it can be
stored for several weeks at ꢀ208C.
[11] The diastereoselectivities of nitrile oxide cycloadditions with (S)-2-
methyl-3-butenol were generally determined by integration of the
differently shifted diastereotopic signals for the hydrogen atom at
[25] The monoprotected (Z)-butenediols were synthesized according to
standard procedures with excess (Z)-butenediol.
1
C5 of the isoxazoline in the crude H NMR spectrum. In exceptional
[26] W. C. Still, A. Mitra, J. Am. Chem. Soc. 1978, 100, 1927.
[27] In all isoxazolines prepared the two possible anti-diastereomers
showed identical chemical shifts in the 1H NMR spectrum for the
hydrogen atom at C5 of the heterocycle. The same observation could
be made for the two possible syn-diastereomers.
cases this method could not be applied due to overlapping peaks
and the diastereoselectivity was determined by comparison of the
differently shifted diastereotopic signals of C5 of the isoxazoline in
the crude 13C NMR spectrum after prolonged measurement times.
[12] Kanemasa and co-workers have previously shown that coordinating
solvents such as Et2O and THF are disadvantageous for magnesium-
mediated nitrile oxide cycloadditions, see ref. [4].
[13] N. Becker, E. M. Carreira, Org. Lett. 2007, 9, 3857.
[14] M. Frederickson, R. Grigg, Z. Rankovic, M. Thornton-Pett, J. Red-
path, R. Crossley, Tetrahedron 1995, 51, 6835.
[28] D. P. Curran, J. Am. Chem. Soc. 1982, 104, 4024.
[29] J. S. Panek, R. T. Beresis, J. Am. Chem. Soc. 1993, 115, 7898.
[30] B. O. Lindgren, T. Nilsson, Acta Chem. Scand. 1973, 27, 888.
[31] A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K. Rosen, F. J.
Timmers, Organometallics 1996, 15, 1518.
[32] S. C. Watson, J. F. Eastham, J. Organomet. Chem. 1967, 9, 165.
[33] W. C. Still, M. Kahn, A. Mitra, J. Org. Chem. 1978, 43, 2923.
[34] This oxime has been previously reported, see: L. P. Safronova, A. S.
Medvedeva, N. S. Vyazankin, J. Gen. Chem. USSR 1983, 53, 1177.
[35] J. S. Panek, R. T. Beresis, J. Am. Chem. Soc. 1993, 115, 7898.
[15] L. P. Safronova, A. S. Medvedeva, N. S. Vyazankin, J. Gen. Chem.
USSR 1983, 53, 1177.
[16] The reason for the unsatisfactory yield might be found in enhanced
dimer formation. Aliphatic nitrile oxides are prone for dimerization.
Thereby higher substituted nitrile oxides exhibit less dimer forma-
tion than their less substituted or even unsubstituted counterparts.
[17] This oxime is commercially available.
[18] The diastereomers were separated by chromatography.
[19] P. L. Anelli, C. Biffi, F. Mantanari, S. Quici, J. Org. Chem. 1987, 52,
2559.
Received: February 26, 2009
Revised: July 3, 2009
Published online: September 24, 2009
Chem. Eur. J. 2009, 15, 12065 – 12081
ꢁ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12081