Assembly of a key dienic intermediate for tetrodotoxin via a Machetti-DeSarlo reaction

Article abstract of DOI:10.1021/jo401960p

A route to a racemic diene intermediate for the synthesis of tetrodotoxin is described. Key steps of the sequence leading to such a compound include the oxidative amidation of a phenol, a Cu(II)-catalyzed cyclocondensation of a nitroketone with an olefin (Machetti-DeSarlo reaction), and a nucleophilic fragmentation of the resulting isoxazoline. Several unusual reactions encountered in the course of this study are thoroughly discussed.

Full text of DOI:10.1021/jo401960p

Article
pubs.acs.org/joc
Assembly of a Key Dienic Intermediate for Tetrodotoxin via a
Machetti−DeSarlo Reaction
Jaclyn Chau, Sanjia Xu, and Marco A. Ciufolini*
Department of Chemistry, The University of British Columbia, 2036 Main Mall, Vancouver, British Columbia V6T 1Z1, Canada
S
* Supporting Information
ABSTRACT: A route to a racemic diene intermediate for the synthesis of
tetrodotoxin is described. Key steps of the sequence leading to such a
compound include the oxidative amidation of a phenol, a Cu(II)-catalyzed
cyclocondensation of a nitroketone with an olefin (Machetti−DeSarlo
reaction), and a nucleophilic fragmentation of the resulting isoxazoline.
Several unusual reactions encountered in the course of this study are thoroughly discussed.
stereoselective installation of OH and latent CHO groups on
dienone 5, recognized as the product of oxidative amidation of
phenolic ester 6.11
INTRODUCTION
■
Tetrodotoxin (TTX, 1; Figure 1),1 a neurotoxic agent2 initially
isolated3 from puffer fish,4 remains a privileged target for
Past efforts12 demonstrated the elaboration of 5 to nitrile 11,
as shown in Scheme 1. Key steps in this sequence included a
Scheme 1. Previously Described Route12 to Compound 11
Figure 1. Structure and retrosynthetic analysis of tetrodotoxin.
Torssell13 cyclization of nitroketone β-9 and a nucleophilic
fragmentation of the resultant isoxazoline β-10.14 Compounds
β-7−β-10 in Scheme 1 are so described to underscore the β
configuration of the tert-butyldiphenylsilyloxy (TBDPSO)
group. It subsequently transpired that it is advantageous to
carry out this sequence with the α-epimers of 7−10.15
Moreover, the Torssell step was found to scale up poorly,
complicating material throughput. Herein, we describe
remedies to these problems, as well as the elaboration of the
α-TBDPSO epimer of 11 to a diene of the type 3. We stress
that this work constituted a feasibility-level study, for which
issues of enantioselectivity were of secondary importance.
Accordingly, all experiments were carried out in the racemic
series.16
chemical synthesis on account of complex architecture, high
density of reactive functionalities packed into a small
framework, and sheer difficulty of a synthetic attack.5 The
first synthesis of ( )-1 was achieved in 1972;6 however,
enantioselective avenues to the natural product7 and its
congeners8 began to appear only in the early 2000s.
Interestingly, the best contemporary route7b to 1 relies on
new synthetic technology: the insertion of a nitrenoid into a
C−H bond.9 This suggests that key to an efficient approach is
the inclusion of new reactions as elements of the overall
strategy.
In that respect, we perceive opportunity in an oxidative
amidation of a phenol10 as an early step toward 1. To illustrate
(Figure 1), the TTX forerunner 2, where [CHO] indicates an
expressed or latent formyl group and P a protecting group,
could be made by bis-dihydroxylation of diene 3, which results
upon Wittig reaction of ketone 4. The latter arises via the
Received: September 4, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jo401960p | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
RESULTS AND DISCUSSION
Scheme 3. Preparation of Nitroketones 9
■
The elaboration of 5 to a nitroketone requires prior reduction
of the keto group.17 Our previous route relied on a DIBAL
reduction of 5 that preferentially yielded the alcohol β-13
(Scheme 2).18 Interestingly, a Corey−Bakshi−Shibata (CBS)19
Scheme 2. Reduction of Dienone 5 and Protection of the
a
Resultant Alcohols
had to be carried out with a 2.4:1 mixture of α- and β-9. This
step was challenging.32 The Torssell cyclization13 had
performed adequately on a small scale,12 but it proved
unsuitable for processing grams of material. In a search for
alternatives, we evaluated a Cu(II)-catalyzed process described
by Machetti and DeSarlo,33 a variant of which (5 mol % of
Cu(OAc)2,34 10 mol % of N-ethylpiperidine,35 CHCl3,36 30
°C,37 168 h) advanced the nitroketones to isoxazolines α-10
and β-10 in 40−50% yield (Scheme 4). A substrate
a
Reagents: (a) 1.0 equiv, 1 mol % of 12, THF, 0 °C, 1 h, >95% yield
of crude α- + β-13; (b) 1.0 equiv each, CH2Cl2, overnight, 85−90%
yield.
Scheme 4. Cu(II)-Catalyzed Formation of Isoxazolines 10
(Machetti−DeSarlo Reaction)
reduction20 preferentially afforded alcohol α-13,21 albeit with
weak selectivity.22 While the reason for this stereochemical
reversal is unclear, a hypothesis may be advanced as follows.
Nucleophiles tend to undergo 1,2- or 1,4-addition to enones of
the type 5 from the face of the π system anti to the
heteroatomic group,23 perhaps due to a Felkin−Anh-type
effect24 created by the electronegative heteroatom.25 Thus, the
stereochemical outcome of the CBS reduction is consonant
with the anticipated facial preference of 5, while the DIBAL
reduction occurs with opposite selectivity. Perhaps, the latter
reaction proceeds through a complex of the acetamide (or an
anionic form thereof) with DIBAL, leading to β-13 via
intramolecular hydride delivery to the keto carbonyl (Figure
2).26
concentration of 0.2 M minimized byproduct formation, but
attempts to accelerate the reaction by using more catalyst,
especially more base, promoted formation of side products (cf.
Scheme 5 for the nature of these). Isoxazoline α-10 was
purified to homogeneity, and no further work was done with β-
10.
The Machetti−DeSarlo reaction performed consistently on a
scale of up to 6 g of nitroketone; however, it was not free from
difficulties. First, the yield of isoxazoline was moderate. Second,
the reaction required an induction period of 15−20 h before
the isoxazoline would begin to form. Third, the reaction failed
to reach completion, unreacted nitroketone remaining even
after 7 days. Fourth, it tended to stall, and addition of more
catalyst to stalled reactions failed to revive them. Fifth,
significant quantities of acids 8 accompanied the desired
isoxazolines. The acids could only form by hydrolysis of the
nitroketone through a process that would release nitromethane,
but because reagents and solvents had been carefully dried, it
seemed likely that the water required to cleave 9 was that
liberated by the reaction itself.38 Numerous small-scale
reactions were thus carried out in CDCl3 solutions (NMR
Figure 2. Hypothesis for the stereochemical course of the reduction of
5.
The mixture of α- and β-13 was advanced to the more stable
tert-butyldiphenylsilyl (TBDPS) ethers27 α- and β-7 (Scheme
2), which were converted into nitroketones28 α-9 and β-9 as
described earlier.12 It proved expedient to eschew purification
of the various intermediates en route to the nitroketones. In
this way, dienone 5 produced a 2.4:1 mixture of α-9 (major)
and β-9, in 56% overall yield after chromatography over four
steps: CBS reduction, TBDPS protection, ester saponifica-
tion,29 and carbonyldiimidazole30 (CDI)-mediated condensa-
40
tube),39 with monitoring by H NMR, in order to garner a
better understanding of the process. These experiments
revealed that release of MeNO2 (singlet at 4.33 ppm) began
to occur after about 15−20 h, at about the same time that the
1
31
tion of the resultant acids with MeNO2 (Scheme 3).
1
Unfortunately, the separation of α and β diastereomers at any
stage of this sequence was difficult. Therefore, the crucial
conversion of the nitroketones into isoxazolines α-10 and β-10
isoxazoline was becoming apparent in the H NMR spectrum.
In an effort to contain/suppress MeNO2 release, i.e., the
formation of acids 8, the effect of adding drying agents to the
B
dx.doi.org/10.1021/jo401960p | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Scheme 5. Side Reactions Observed During the Cyclization
of Nitroketones 9
Scheme 6. Preparation of N-BOC Substrate 23
(Figure 3). All subsequent work was therefore carried out with
stereochemically homogeneous materials.
reaction mixture was examined. Molecular sieves inhibited
isoxazoline formation. It seems likely that this was due to
protonation of N-ethylpiperidine by the acidic molecular sieves,
an event that would deny the reaction an essential basic agent.
Powdered, anhydrous Na2SO4 had no effect, but CaSO4
(powdered white Drierite activated by heating under vacuum)
significantly diminished MeNO2 formation, without fully
suppressing it.41 However, reactions run in the presence of
CaSO4 still tended to stall and could not be resurrected by the
addition of more catalyst.
Figure 3. X-ray crystal structure of compound 21.
It should also be noted that the step leading to 20 afforded
byproducts 24−26 in 16%, 10%, and 4% yields, respectively
(Scheme 7). The isolation of 24 underscores the notion
A byproduct detected during NMR monitoring of the
reaction, compound 1742 (Scheme 5), and the observation that
running the reaction at higher temperatures in the presence of
CaSO4 induced only a greater extent of formation of 17
without accelerating isoxazoline formation, provided a clue as
to the source of the above problems. A sensible mechanism for
the formation of 17 starts with reversible deprotonation of the
AcNH group, either in the free nitroketone (cf. 14) or in a
Cu(II) chelate thereof (cf. 15). Such events lead to the release
of nitroacetone, a good ligand for Cu(II) that can sequester the
metal and bring the catalytic cycle leading to the isoxazolines to
a halt. Moreover, the O terminus of the anion of the acetamide
could add to the keto carbonyl (cf. 14 → 18), triggering release
of MeNO2 and formation of acids 8 upon hydrolysis of
azalactones 19. On such a basis, it seemed desirable to replace
the N-acetyl with a less N−H acidifying and less O-nucleophilic
BOC group.43
N-BOC-protected nitroketone 23 was thus prepared
according to Scheme 6, several aspects of which deserve
comment. First, the CBS reduction of 20 proceeded with
improved selectivity relative to 5 (4.9:1 in favor of the α epimer
vs 2.4:1). Second, the less polar N-BOC compounds were
easier to handle and purify than their N-acetyl congeners.
Finally, and in contrast to the N-acetyl series, the epimeric
alcohols obtained upon CBS reduction of 20 could be easily
separated at the stage of 21: the latter crystallized from the
mixture, enabling also a structural proof by X-ray diffractometry
Scheme 7. Byproducts Formed in the Step Leading to 20
advanced in Scheme 5 that deprotonation of the acetamide is
relatively facile and that such an event may be at the root of
many undesirable side reactions. In fact, 24 arguably arose
through Michael cyclization of the N anion of 5 at the O
terminus and interception of the emerging enolate by BOC2O.
The formation of the noteworthy cycloheptatriene 25 indicates
that deprotonation of the ester segment in 5 and/or 20 is also
facile. Indeed, the work of Carreno and collaborators44 supports
the sequence of events depicted in Scheme 8 for the formation
of 25. Finally, compound 26 reflects the tendency of enolate 27
to eliminate the anion of N-BOC-acetamide, leading to quinone
methide 29. The genesis of 26 may be accounted for by
C
dx.doi.org/10.1021/jo401960p | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
series, compound 21 was reproducibly advanced to isoxazoline
36 in 34% overall yield (after chromatography) over a four-step
sequence encompassing N-deacetylation, ester saponification,
nitroketone synthesis, and Cu(II)-catalyzed cyclization (76%
average yield per step), without purification of intermediate
products (Scheme 10). Chromatography of the final 36 also
Scheme 8. Presumed Sequence of Events Leading to the
Formation of Byproducts 25 and 26
Scheme 10. Preparation of Compound 36 in Four Steps
from 21 without Intermediate Purification
invoking a 1,6-addition of 27 to 29 and evolution of the
resultant 30 as per Scheme 8.
Elimination of N-BOC-acetamide seems responsible also for
the formation of significant quantities of 34 (Scheme 9) during
returned quantities of acid 22 (typically 25−30% yield based on
21), which was conveniently recycled into the sequence.
However, we emphasize that the stated yield of 36 refers to the
actual yield of the process outlined in Scheme 10, and it is not
based on recovered or recycled 22.
Scheme 9. Proposed Mechanism of Formation of 34 during
Hydrolysis of 21
The previously described12 nucleophilic isoxazoline fragmen-
tation (Li2CO3/MeOH)45 advanced α-10 to 37 (the structure
of which had been confirmed earlier by X-ray diffractometry;
Figure 4)46 in 98% yield after chromatography (Scheme 11).
attempts to reach 22 by simultaneous ester saponification/N-
deacetylation of 21 with aqueous LiOH. Compound 34
probably arose upon tautomerization of elimination product
35, followed by release of the silyl group under basic
conditions. It is this problem that mandated the N-
deacetylation of 21 with N2H4·H2O prior to basic hydrolysis
(Scheme 6).
The behavior of N-BOC nitroketone 23 under the
Machetti−DeSarlo conditions of Scheme 4 paralleled that of
the N-acetyl congener. However, and in sharp contrast to the
case of 9, an increase in the amount of base (30 mol % of N-
ethylpiperidine vs the original 10 mol %) accelerated the
conversion of 23 into the desired isoxazoline with no significant
increase in formation of the byproducts shown in Scheme 5.
This it is consistent with the diminished N−H acidity and O
nucleophilicity of the N-BOC compounds relative to N-Ac
materials. As before, an increase in the amount of Cu(OAc)2
(10 mol % vs 5 mol %) resulted only in the formation of more
of the N-BOC analogue of compound 17, while the addition of
CaSO4 diminished the extent of MeNO2 release without
affecting rates and yields. More beneficial was the use of MeCN
as a solvent in lieu of CHCl3, a modification that induced the
reaction to proceed to completion. Finally, the addition of silica
gel to the reaction medium andmore importantlythe
conduct of the reaction in more dilute solutions (0.04 M)
produced the best results. Just as seen earlier in the N-acetyl
Figure 4. X-ray crystal structure of compound 37.46
The fragmentation of 36 occurred as efficiently to afford 38,
which was free from contaminants (NMR) and therefore
required no further purification.47
The quantities of acetamido intermediates accumulated in
the course of these investigations served to explore one more
transformation: the elaboration of 37 into a diene of the type
3.48 This effort started with the protection of the free OH
group as a BOM derivative49 in preparation for release of the
TBDPS group, oxidation to a ketone, and a Wittig reaction
(Scheme 12). Desilylation of 32 and Parikh−Doering
oxidation50 of the resultant alcohol afforded ketone 40,51
D
dx.doi.org/10.1021/jo401960p | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Scheme 11. Nucleophilic Fragmentation of Isoxazolines α-10
and 36
EXPERIMENTAL SECTION
■
Experimental Protocols. Unless otherwise indicated, 1H (300
MHz) and 13C (75 MHz) NMR spectra were obtained from CDCl3
solutions at room temperature. Chemical shifts are reported in parts
per million (ppm) on the δ scale, coupling constants, J, in hertz (Hz),
and multiplicities as s (singlet), d (doublet), t (triplet), q (quartet), dd
(doublet of doublets), m (multiplet), c (complex), br (broad), ABq
(AB quartet), and app (apparent). Infrared (IR) spectra (cm−1) were
recorded from thin films deposited on NaCl plates. Mass spectra (m/
z) were obtained in the electrospray (ESI) or atmospheric pressure
chemical ionization (APCI) modes on a time-of-flight mass
spectrometer equipped with an electrospray ion source. Melting
points (uncorrected) were measured on a Mel-Temp apparatus.
Commercial reagents and solvents were used without further
purification except for THF (freshly distilled from Na/benzophenone
under N2), CH2Cl2 (freshly distilled from CaH2 under N2), CHCl3
(washed with H2O, treated with solid K2CO3, freshly distilled from
Na2SO4 under Ar), MeOH (freshly distilled from Mg/I2 under Ar),
1,2-dichloroethane and acetone (freshly distilled from CaSO4 under
Ar), and i-Pr2NEt and MeCN (distilled from CaH2 under Ar).
Commercial n-BuLi was titrated against Ph2CHCOOH.55 Flash
chromatography was performed on Silicycle 230−400 mesh silica
gel. All reactions were performed under an Ar atmosphere in oven-
dried flasks fitted with rubber septa for the introduction of substrates/
reagents/solvents via syringe and equipped with Teflon stirring bars.
Preparation of α-9 without Purification of the Various
Intermediates. Commercial BH3·SMe2 (∼10 M solution, 2.7 mL, 27
mmol, 1.0 equiv) was carefully added dropwise, via syringe, to a rapidly
stirred solution of 5 (6.0 g, 27 mmol, 1.0 equiv) and (S)-CBS catalyst
(0.075 g, 0.27 mmol, 0.01 equiv) in dry THF (135 mL) maintained
under Ar in a 500 mL round-bottom flask immersed in an ice bath.
The mixture was stirred for 45 min, during which time it was warmed
to room temperature; then it was cooled back to 0 °C, the flask was
opened, and MeOH (3.3 mL, 81 mmol, 3.0 equiv) was carefully added
dropwise via syringe at 0 °C (Caution! H2 evolution). The solution was
warmed to room temperature, and the solvent was removed under
reduced pressure. The residue of crude α-13 (major component of a
2.4/1 mixture with β-13) was dried under high vacuum, and then it
was taken up in dry CH2Cl2 (100 mL); to this solution were added
imidazole (1.83 g, 27 mmol, 1.0 equiv) and TBDPSCl (7.0 mL, 27
mmol, 1.0 equiv). The mixture was stirred for 12 h (overnight) at
room temperature under Ar, and then 0.05 N HCl (100 mL) was
added. The layers were separated, and the aqueous layer was extracted
with CH2Cl2 (2 × 100 mL). The extracts were combined, dried
(Na2SO4), filtered, and concentrated. The residue of crude α-7 (major
component of a 2.4:1 mixture with β-7) was dissolved in THF (100
mL) and treated with H2O (100 mL) and LiOH·H2O (3.4 g, 81 mmol,
3.0 equiv). The solution was stirred for 1.5 h at room temperature, and
then most of the THF was removed by rotary evaporation. Diethyl
ether (100 mL) was added to the aqueous suspension, and the layers
were separated to remove organic byproducts from previous steps (the
desired carboxylic acids were dissolved in the aqueous layer as the Li
carboxylates). The aqueous layer was acidified with 50% aqueous
AcOH (30 mL) and extracted with EtOAc (3 × 100 mL, check pH of
aqueous layer). The combined extracts were dried (Na2SO4), filtered,
and concentrated. The residue was resuspended in toluene, and the
resulting mixture was again concentrated (rotary evaporation) to
remove AcOH. Coevaporation with toluene was repeated three to four
times until all AcOH had been removed. The residue of crude α-8
(major component of a 2.4:1 mixture with β-8) was taken up in dry
THF (70 mL) and treated with CDI (4.36 g, 27 mmol, 1.0 equiv). The
solution was stirred at room temperature under Ar for 1 h, and then
MeNO2 (8.5 mL, 0.16 mol, 6.0 equiv) and t-BuOK (12.0 g, 0.11 mol,
4.0 equiv) were added at room temperature and the resulting mixture
was stirred at room temperature for 2 h. Aqueous 50% AcOH (30 mL)
was added, and the volatiles were removed under reduced pressure.
The residue was partitioned between H2O (70 mL) and CH2Cl2 (100
mL), the layers were separated, and the aqueous phase was extracted
with more CH2Cl2 (2 × 100 mL). The combined extracts were dried
Scheme 12. Aromatization of Ketone 63 under Wittig
Conditions
contact of which with Ph3PCH2 at −78 °C triggered rapid
aromatization to 42.52 Previous observations (Schemes 8 and
9) suggested that suppression of the ester functionality would
resolve the problem. Therefore, the ester was selectively
reduced (LiBH4)53 and the resultant primary alcohol was
blocked with a second BOM group (Scheme 13). Release of the
Scheme 13. Preparation of Diene 47
TBDPS group in 44 was best achieved with TBAF buffered
with acetic acid,54 while Dess−Martin periodinane (DMP) was
the reagent of choice for the oxidation of 45 to ketone 46,
which, to our relief, underwent smooth Wittig olefination to
furnish diene 47 in 87% yield.
CONCLUSION
■
The elaboration of 5 to isoxazoline 36 under the modified
Machetti−DeSarlo conditions detailed above alleviated material
throughput issues that affected the original route to 11. An
avenue to diene 47 is also described. Observations recorded in
the course of this research are key to the success of ongoing
synthetic efforts toward the natural product.
E
dx.doi.org/10.1021/jo401960p | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
(Na2SO4), filtered, and concentrated, and the residue was purified by
gradient column chromatography (1/1−7:/3 EtOAc/hexanes) to
furnish nitroketone α-9 (7.9 g, 16 mmol, 56% over four steps), light
yellow oil, as the major component of a 2.4/1 mixture with
Data for compound 24 are as follows. 1H NMR: 5.85 (d, J = 9.8 Hz,
1H); 5.77 (dd, J = 10.2 Hz, 1.9 Hz, 1H); 5.67 (app d, J = 5.1 Hz, 1H);
5.77 (d, J = 5.1 Hz, 1H); 3.64 (s, 3H); 2.79, 2.64 (ABq, JAB = 14.9 Hz,
2H); 1.96(s, 3H); 1.49 (s, 9H). 13C NMR: 170.0, 166.1, 150.8, 148.1,
132.4, 120.8, 106.3, 84.0, 80.4, 69.0, 52.0, 45.5, 27.8, 14.1. IR: 1754,
1739, 1662. MS: 346 [M + Na]+. HRMS: calcd for C16H22NO6 [M +
H]+ 324.1447; found 324.1453.
1
diastereomer β-9. Data for α-9 are as follows H NMR (acetone-d6):
7.75−7.72 (m, 4H); 7.53−7.37 (m, 6H); 6.20−6.14 (m, 2H); 5.88−
5.82 (m, 2H); 5.57 (s, 2H); 4.61−4.59 (m, 1H); 3.22 (s, 2H); 1.88 (s,
3H); 1.07 (s, 9H). Signals arising from diastereomer 912 were apparent
at 5.67 (s, 2H), 3.30 (s, 2H), 1.78 (s, 3H), and shoulder at 1.07 (s,
9H). 13C NMR (acetone-d6): 195.3, 170.8, 136.6, 134.5, 130.9, 130.2,
129.7, 128.8, 85.1, 64.4, 52.6, 48.4, 27.4, 23.7, 19.8. Signals arising from
diastereomer 912 were apparent at 195.5, 170.4, 130.8, 128.7, 52.2,
48.7, 27.3, 23.5, and 19.7. IR: 1736, 1700, 1656, 1558. MS: 515 [M +
Na]+; negative mode: 491 [M − H]−. HRMS: calcd for
C27H32N2O5SiNa [M + Na]+ 515.1978; found 515.1962.
1
Data for compound 25 are as follows. H NMR: 6.36 (dd, J = 7.0
Hz, 1.2 Hz, 1H); 6.21 (app d, J = 9.7 Hz, 1H); 6.05 (d, J = 7.0 Hz,
1H); 5.63 (dd, J = 9.7 Hz, 6.7 Hz, 1H); 3.69 (s, 3H); 3.28 (d, J = 6.7
Hz, 1H); 2.47 (s, 3H); 1.51(s, 9H); 1.43 (s, 9H). 13C NMR: 173.3,
170.9, 152.3, 151.5, 151.3, 124.2, 123.9, 122.8, 122.3, 117.2, 83.8, 83.7,
52.3, 47.8, 27.9, 27.7, 26.1. IR: 1739, 1708 (shoulder). MS: 446 [M +
Na]+. HRMS: calcd for C21H29NO8Na [M + Na]+ 446.1791; found
446.1793.
Data for compound 26 are as follows. 1H NMR: 7.10, 7.03 (app AB
q, JAB = 8.8 Hz, 8H); 3.82 (s, 6H); 1.54 (s, 18H). 13C NMR: 168.2,
151.5, 151.2, 138.2, 131.5, 131.0, 121.2, 83.9, 52.9, 27.8. IR: 1752,
1718. MS: 551 [M + Na]+. HRMS: calcd for C28H32O10Na [M + Na]+
551.1893; found 551.1888.
Preparation of α-10. A solution of α-9 (major component of a
2.4/1 mixture with diastereomer β-9; 226 mg, 459 μmol, 1 equiv),
Cu(OAc)2·H2O (5 mg, 25 μmol, 5.4 mol %), and N-ethylpiperidine (8
μL, 6.7 mg, 59 μmol, 13 mol %) in dry CHCl3 (3 mL) was stirred at
30 °C (oil bath temperature) under Ar for 7 days, and then it was
concentrated. The residue was dissolved in EtOAc (10 mL) and
washed three times with 0.01 M aqueous HCl solution (5 mL). The
organic phase was rinsed with brine, dried (Na2SO4), and evaporated,
and the residue was purified by flash column chromatography on silica
gel (50% EtOAc in hexanes) to give pure α-10 (60 mg, 27%, white
powder, mp 74−75 °C), plus a fraction containing α-10 and β-10 (12
mg, 6%), and pure isoxazoline β-1012 (30 mg, 14%), for an overall 47%
Preparation of 21. Commercial BH3·SMe2 complex (1.0 mL, 10.0
mmol, 1.0 equiv) was carefully syringed over 5 min into a cold (0 °C)
solution of 20 (3.20 g, 9.9 mmol, 1.0 equiv) and (S)-CBS catalyst (28
mg, 0.1 mmol, 0.01 equiv) in THF (74 mL), with good stirring under
Ar. The ice bath cooling the mixture was removed, and stirring was
continued for an additional 50 min. The reaction was quenched by
careful dropwise addition of MeOH (1.5 mL, 37 mmol, 3.7 equiv)
(Caution! H2 evolution). When gas evolution stopped, more MeOH
(10 mL) was added and stirring was continued for another 15 min.
The solution was concentrated, the residue was redissolved in MeOH
(15 mL), and the solution was again concentrated to dryness. The
latter operation was repeated once to ensure complete decomposition
of organoboron species. The residue was then filtered through a short
silica gel plug with 50% EtOAc/hexanes until no product (Rf = 0.13 in
30% EtOAc/hexanes) was observed in the eluate by TLC. The filtrate
was concentrated to dryness under vacuum, and the residue was
azeotroped with toluene (15 mL) and then dried under high vacuum
to constant mass (3.0 g). The crude reduction product was dissolved
in dry DMF (20 mL) and treated with imidazole (1.30 g, 19.1 mmol,
1.9 equiv) and TBDPSCl (2.9 mL, 10.0 mmol, 1.0 equiv). The
reaction flask was then immersed into an oil bath maintained at 70 °C
and the solution stirred for 24 h. The mixture was then cooled to room
temperature, and most of the DMF was removed under vacuum. The
residue was taken up in EtOAc (60 mL), and the solution was
successively washed with 0.02 M HCl solution (3 × 30 mL), DI water
(15 mL), and brine (15 mL), dried (Na2SO4), and concentrated. The
residue was redissolved in MeOH (15 mL), and the solution was again
concentrated to dryness. The latter operation was repeated once to
ensure complete removal of volatiles. The residue thus obtained was
dissolved in refluxing MeOH (15 mL) and the hot solution was
allowed to stand overnight, whereupon 21 crystallized as transparent
prisms. This solid was filtered and rinsed with cold MeOH (2 × 5 mL)
and then recrystallized again from MeOH (10 mL) to afford pure 21
(2.82 g, 5.0 mmol, 51%) as white prisms, mp 94−95 °C. The
combined mother liquors from such recrystallizations were concen-
trated and the residue was subjected to gradient chromatography (1/
9−1/4 EtOAc/hexanes) to obtain more 21 (0.45 g, 0.8 mmol, 8%; Rf
= 0.38 in 1/4 EtOAc/hexanes), its β epimer (0.44 g, 0.8 mmol, 8%; Rf
= 0.31 in 1:4 EtOAc/hexanes) as a pale yellow oil, and a mixture of the
two (0.56 g, 1.0 mmol, 10%; containing 48% of 21). In summary, a
76% yield of both diastereomers was obtained with a 4.9/1
diastereomeric ratio.
1
yield of isoxazolines. H NMR: 7.71−7.60 (m, 4H); 7.53−7.37 (m,
6H); 6.18 (dd, J = 9.96, 5.94, 1H); 5.6 (d, J = 9.96, 1H); 5.75 (br,
1H); 5.05 (app s, 2H); 4.24 (d, J = 5.94, 1H); 4.06 (d, J = 18.4, 1H);
2.97 (d, J = 18.4, 1H); 1.95 (s, 3H); 1.10 (s, 9H). 13C NMR: 191.4,
170.8, 161.7, 135.8, 135.6, 135.5, 134.9, 133.4, 132.6, 132.4, 130.5,
130.4, 128.1, 128.0, 85.1, 63.2, 54.2, 52.3, 27.0, 24.0, 19.1. IR: 3289,
1747, 1651, 1596, 1530. MS: 475 [M + H]+. HRMS: calcd for
C27H31N2O4Si [M + H]+ 475.2053; found 475.2059.
Larger Scale Preparation of α-10 Contaminated with Acids
α- and β-8. A solution of α-9 (major component of a 2.4/1 mixture
with diastereomer β-9; 3.5 g, 7.1 mmol, 1.0 equiv), Cu(OAc)2·H2O
(0.072 g, 0.36 mmol, 0.05 equiv), and N-ethylpiperidine (0.10 mL,
0.71 mmol, 0.10 equiv) in dry CHCl3 (25 mL) was stirred at 30 °C
(oil bath temperature) under Ar for 168 h, and then it was
concentrated. The residue was dissolved in a minimal volume of
EtOAc and filtered through a short plug of silica gel, rinsing with more
EtOAc. The solvent was evaporated, and the residue was purified by
flash column chromatography on silica gel (EtOAc/Et2O 1/4) to give
α-10 as the major component of a 5/1 mixture with acids α- and β-8.
The mass of this material amounted to 2.0 g of a pale yellow foam,
which therefore consisted of 84% (mass-wise) of α-10 and 16% of
acids, making the yield of oxazoline equal to 1.68 g, 3.5 mmol, 49%.
Preparation of 20. Solid DMAP (82 mg, 0.67 mmol, 0.04 equiv)
and solid Boc2O (6.6 g, 30.3 mmol, 1.8 equiv) were added to a
solution of 5 (3.75 g, 16.8 mmol, 1.0 equiv) in THF (100 mL). The
mixture was stirred under Ar at room temperature for 48 h, and then it
was concentrated under vacuum. The residue was dissolved in EtOAc
(50 mL), and the resulting solution was washed with 0.02 M HCl
solution (2 × 30 mL), DI water (10 mL), and brine (10 mL), and then
dried (Na2SO4) and concentrated. The residue was purified with
gradient chromatography (3/7−1/1−7/3 EtOAc/hexanes) to afford,
in order of elution: 25 (0.72 g, 1.7 mmol, 10%; Rf = 0.50 in 3/7
EtOAc/hexanes) as a pale yellow oil; 26 (0.17 g, 0.3 mmol, 4%; Rf =
0.34 in 3:7 EtOAc/hexanes) as white plates, mp 171 °C (dec); the
desired 20 (3.20 g, 9.9 mmol, 61%; Rf = 0.24 in 3/7 EtOAc/hexanes)
as a pale yellow oil; 24 (0.88 g, 2.7 mmol, 16%; Rf = 0.52 in EtOAc) as
a reddish oil.
1
Data for compound 21 are as follows. H NMR: 7.71−7.67 (m,
4H); 7.46−7.36 (m, 6H); 6.16 (dd, J = 10.3 Hz, 1.8 Hz, 2H); 5.81
(dd, J = 10.3 Hz, 2.9 Hz, 2H); 4.46−4.42 (m, 1H); 3.53 (s, 3H); 3.08
(s, 2H); 2.16 (s, 3H); 1.53 (s, 9H); 1.07 (s, 9H). 13C NMR: 170.5,
170.1, 153.9, 134.0, 133.7, 130.0, 129.5, 128.0, 127.9, 84.5, 63.1, 57.3,
51.6, 44.0, 27.6, 27.0, 25.2, 19.3. IR: 1741, 1682. MS: 586 [M + Na]+.
HRMS: calcd for C32H41NO6NaSi [M + Na]+ 586.2601; found
586.2600.
1
Data for compound 20 are as follows. H NMR: 7.26 (d, J = 10.2
Hz, 2H); 6.24 (d, J = 10.2 Hz, 2H); 3.66 (s, 3H); 3.22 (s, 2H); 2.28 (s,
3H); 1.46 (s, 9H). 13C NMR: 184.6, 172.2, 169.0, 152.7, 148.4, 128.0,
85.2, 58.7, 52.0, 42.5, 27.5, 26.3. IR: 1740, 1690, 1670, 1631. MS: 346
[M + Na]+. HRMS: calcd for C16H21NO6Na [M + Na]+ 346.1267;
found 346.1265.
F
dx.doi.org/10.1021/jo401960p | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Data for the β-epimer of 21 are as follows: mp 82−84 °C (MeOH).
1H NMR: 7.71−7.66 (m, 4H); 7.44−7.35 (m, 6H); 6.20 (dd, J = 10.3
Hz, 2.1 Hz, 2H); 5.89 (dd, J = 10.3 Hz, 2.7 Hz, 2H), 4.58−4.53 (m,
1H); 3.67 (s, 3H); 3.19 (s, 2H); 2.04 (s, 3H); 1.29 (s, 9H); 1.07 (s,
9H). 13C NMR: 170.4, 169.8, 153.6, 136.0, 133.9, 132.3, 129.9, 128.4,
127.9, 84.1, 63.8, 57.1, 51.8, 43.1, 27.4, 27.1, 24.9, 19.3. IR: 1741, 1682.
MS: 586 [M + Na]+. HRMS: calcd for C32H41NO6SiNa [M + Na]+
586.2601; found 586.2598.
chromatography (EtOAc/hexanes 1/1) to give 37 (1.32 g, 2.60
mmol, 98%) as an off-white solid, mp 166−168 °C.
Data for 37 are as follows. 1H NMR: 7.72−7.68 (m, 4H); 7.50−7.39
(m, 6H); 5.84 (br d, J = 10.29, 1H); 5.76 (br s, 1H); 5.72 (dd, J =
10.2, 2.6, 1H); 4.45−4.41 (m, 1H); 4.35 (d, J = 3.12, 1H); 4.07−4.03
(m, 1H); 3.67 (s, 3H); 3.45 (d, J = 16.08, 1 H); 2.87 (d, J = 16.08, 1
H); 1.98 (s, 3H); 1.10 (s, 9H). 13C NMR: 170.6, 170.3, 135.9, 135.7,
133.1, 132.0, 130.2, 130.2, 128.0, 127.9, 127.8, 117.4, 71.6, 70.2, 55.1,
51.9, 39.17, 38.7, 26.9, 24.1, 19.2. IR: 3370, 3304, 2252, 1720, 1641.
MS: 529 [M + Na]+. HRMS: calcd for C28H34N2O5SiNa [M + Na]+
529.2135; found 529.2122.
Preparation of 36 without Purification of Intermediates.
Hydrazine hydrate (225 μL, 4.4 mmol, 1.7 equiv) was added (syringe)
to a suspension of 21 (1.47 g, 2.61 mmol, 1.0 equiv) in THF (26 mL)
maintained under Ar in a heavy-walled glass tube fitted with a screw-
cap container. The mixture was heated to 60 °C and stirred for 24 h,
and then it was cooled to room temperature and concentrated in
vacuo and the residue was filtered through a short silica gel plug (10
mL) with 1/1 EtOAc/hexanes until no deacetylated product (Rf = 0.72
in 3/7 EtOAc/hexanes) eluted. The filtrate was concentrated to
dryness, and the residual light yellow oil was dissolved in THF (14
mL). Deionized water (14 mL) was added with good stirring, followed
by solid LiOH·H2O (332 mg, 7.90 mmol, 3.0 equiv, added in one
portion). The resulting suspension became a clear monophasic
solution within 4 h. Stirring was continued for another 1 h until all
starting ester had been consumed, and then the mixture was cooled in
an ice bath and acidified with 0.4 M HCl solution (20 mL, 8.0 mmol,
3.1 equiv), added slowly with vigorous stirring. The acidic solution was
extracted with EtOAc (2 × 35 mL), and the combined extracts were
washed with DI water (20 mL) and brine (20 mL) and then dried
(Na2SO4) and concentrated. The residue of crude 22 was dissolved in
THF (17 mL) and treated with carbonyldiimidazole (467 mg, 2.88
mmol, 1.1 equiv). After 3 h, to the mixture were added MeNO2 (840
μL, 15.6 mmol, 6.0 equiv) and t-BuOK (1.17 g, 10.4 mmol, 4.0 equiv),
and then the reaction flask was immersed in an oil bath maintained at
40 °C for 30 min. The mixture was cooled to room temperature and
quenched by adding 1/9 HOAc/H2O (20 mL) solution, and then it
was extracted with CH2Cl2 (30 mL, 2 × 10 mL). The combined
extracts were dried (Na2SO4) and concentrated under reduced
pressure. Complete removal of AcOH was achieved by azeotroping
with toluene (3 × 10 mL). The residue was filtered through a short
silica gel plug (10 mL) with 1/1 EtOAc/hexanes (removal of
imidazonium salts) until no more 23 (Rf = 0.46 in 3/7 EtOAc/
hexanes, streak) eluted. Concentration of the filtrate afforded crude
23,56 which was dissolved in freshly distilled MeCN (64 mL)
containing Cu(OAc)2·H2O (25 mg, 0.13 mmol, 0.05 equiv), silica gel
(14 mg), and N-ethylpiperidine (105 μL, 0.76 mmol, 0.3 equiv). The
resulting suspension was stirred (Ar) at 35 °C (oil bath) for 168 h,
whereupon TLC indicated complete consumption of 23. The mixture
was concentrated under vacuum, the residue was dissolved in EtOAc
(70 mL), and the solution was washed with 0.02 M HCl (3 × 20 mL)
and brine (20 mL), dried (Na2SO4), filtered, and concentrated.
Purification of the residue by flash chromatography (1/9−1/4 EtOAc/
hexanes) furnished 36 as an off-white foam (470 mg, 0.88 mmol, 34%;
Rf = 0.56 in 3/7 EtOAc/hexanes). Further elution with with 1/1
EtOAc/hexanes returned acid 22 (340 mg, 0.67 mmol, 26%; Rf = 0.37
in 7/3 EtOAc/hexanes, streak), mp 157−158 °C (CH2Cl2),57 which
was conveniently recycled.
Preparation of 38. Solid Li2CO3 (5 mg, 0.07 mmol, 0.5 equiv)
was added to a solution of 36 (72 mg, 0.14 mmol, 1.0 equiv, dried to
constant weight under high vacuum and then stored in a desiccator
overnight) in dry MeOH (2.5 mL). The suspension was stirred under
Ar for 1 h, and then it was diluted with CH2Cl2 (2.5 mL), filtered
through Celite with more CH2Cl2 (10 mL), and concentrated in
vacuo. Compound 38 (76 mg, 0.13 mmol, 99%; Rf = 0.46 in 3/7
EtOAc/hexanes) was obtained as an off-white foam. The NMR spectra
of this material reveal the presence of no impurities; consequently, no
further purification was carried out.
Data for 38 are as follows. 1H NMR: 7.73−7.69 (m, 4H); 7.48−7.38
(m, 6H); 5.85 (d, J = 10.3 Hz, 1H); 5.68 (dd, J = 10.2 Hz, 2.4 Hz,
1H); 4.77 (br, 1H); 4.46 (dd, J = 4.8, 2.0, 1H); 4.13−4.09 (m, 2H);
3.67 (s, 3H); 3.39, 2.82 (ABq, JAB = 15.4 Hz, 2H); 2.14 (d, J = 3.5,
1H) ; 1.45 (s, 9H); 1.10 (s, 9H). 13C NMR: 170.4, 153.9, 136.4, 135.8,
133.3, 133.3, 132.1, 130.4, 130.3, 128.2, 128.1, 127.9, 117.53, 80.8,
72.2, 70.2, 54.8, 51.9, 40.00, 39.8, 28.4, 27.1, 19.4. IR: 3418, 2246,
1721 (broad). MS: 571 [M + Li]+. HRMS: calcd for C31H40N2O6NaSi
[M + Na]+ 587.2553; found 587.2551.
Preparation of 39. A solution of 37 (0.303 g, 0.60 mmol, 1.0
equiv), BOMCl (0.12 mL, 0.90 mmol, 1.5 equiv), and DIPEA (0.19
mL, 1.08 mmol, 1.8 equiv) in 1,2-dichloroethane (3.0 mL) was stirred
at 75 °C under Ar for 48 h, and then it was cooled to room
temperature, diluted with CH2Cl2 (40 mL), and partitioned between
saturated aqueous NH4Cl solution (50 mL). The organic layer was
separated, and the aqueous phase was extracted with more CH2Cl2 (2
× 30 mL). The combined extracts were washed with H2O (150 mL),
dried (Na2SO4), and concentrated under vacuum, and the residue was
purified by column chromatography on silica gel (Et2O/hexanes 7/3)
to give 39 (0.341 g, 0.55 mmol, 91%) as a colorless oil.
Data for 39 are as follows. 1H NMR: 7.73−7.69 (m, 4H); 7.48−7.30
(m, 11H); 6.27 (br s, 1H); 5.77 (d, J = 10.32, 1H); 5.56 (dd, J = 10.32,
3.42, 1H); 4.74−4.67 (m, 4H); 4.53−4.49 (m, 1H); 4.46−4.43 (m,
1H); 4.16−4.13 (m, 1H); 3.68 (s, 3H); 3.28 (d, J = 15.63, 1H); 2.94
(d, J = 15.63, 1H); 1.99 (s, 3H); 1.10 (s, 9H). 13C NMR: 170.8, 170.2,
137.3, 136.0, 135.88, 133.6, 133.0, 130.4, 130.1, 130.0, 129.0, 128.5,
128.0, 127.9, 127.8, 127.8, 118.1, 94.5, 77.5, 76.4, 69.9, 68.2, 54.7, 54.5,
51.9, 40.0, 36.7, 30.4, 27.0, 24.0, 19.3. IR: 3291, 2247, 1737, 1665. MS:
649 [M + Na]+. HRMS: calcd for C36H42N2O6SiNa [M + Na]+
649.2710; found 649.2710.
Preparation of 43. Solid LiBH4 (0.079 g, 3.6 mmol, 10.0 equiv)
was added to a dry THF (3.0 mL) solution of 39 (0.227, 0.36 mmol,
1.0 equiv). The mixture was stirred at room temperature under Ar for
24 h, and then it was cooled to 0 °C and saturated aqueous NH4Cl
solution (2.0 mL) was carefully added (Caution! H2 evolution). The
solution was stirred at room temperature for a further 30 min, and
then it was diluted with EtOAc (20 mL). The layers were separated,
and the aqueous phase was extracted with more EtOAc (2 × 20 mL).
The combined extracts were dried (Na2SO4) and concentrated under
vacuum, and the residue was purified by column chromatography on
silica gel (100% EtOAc) to give 43 (0.152 g, 0.25 mmol, 70%) as a
colorless oil.
Data for 36 are as follows. 1H NMR: 7.70−7.61 (m, 4H); 7.52−7.38
(m, 6H); 6.15 (dd, J = 9.9 Hz, 6.0 Hz, 1H); 5.81 (dd, J = 9.9 Hz, 1.0
Hz, 1H); 5.09−5.02 (m, 2H); 4.83 (d, J = 11.1 Hz, 1H); 4.17 (dd, J =
6.0 Hz, 2.0 Hz, 1H); 3.91, 3.40 (ABq, JAB = 18.5 Hz, 2H); 1.45 (s,
9H); 1.09 (s, 9H). 13C NMR: 191.5, 161.7, 155.01, 135.8, 135.7, 134.7,
134.5, 132.6, 132.5, 130.6, 130.4, 128.2, 128.1, 85.1, 80.7, 62.9, 55.4,
54.7, 52.0, 28.4, 27.1, 19.1. IR: 1747, 1711. MS: 555 [M+Na]+, 587 [M
+ MeOH + Na]+. HRMS: calcd for C30H36N2O5NaSi [M + Na]+
555.2291; found 555.2300.
Preparation of 37. A solution of α-10 (major component of a 5/1
mixture with acids 8, 1.50 g, corresponding to 1.26 g of α-10, 2.65
mmol, 1.0 equiv) and Li2CO3 (117 mg, 1.58 mmol, 0.6 equiv) in dry
MeOH (63.0 mL; freshly distilled from Mg turnings) was stirred under
argon at room temperature for 1.5 h, and then the solvent was
removed under vacuum. The residue was purified by flash
Data for 43 are as follows. 1H NMR: 7.73−7.68 (m, 5H); 7.44−7.24
(m, 10H); 6.73 (br s, 1H); 5.82 (d, J = 10.4, 1H); 5.52 (dd, J = 10.4,
3.54, 1H); 4.74−4.45 (m, 5H); 4.35−4.33 (br m, 1H); 4.15−4.12 (br
m, 1H); 3.90−3.68 (br m, 2H); 2.42−2.33 (m, 1H); 2.20−2.12 (m,
1H); 1.97 (s, 3H); 1.09 (s, 9H). 13C NMR: 170.6, 137.3, 136.0, 135.9,
133.6, 133.1, 131.2, 130.1, 130.0, 128.5, 128.2, 127.9, 127.9, 127.8,
127.8, 119.1, 94.3, 77.4, 77.1, 69.9, 67.3, 58.4, 55.9, 39.2, 36.3, 27.0,
G
dx.doi.org/10.1021/jo401960p | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
24.2, 19.3. IR: 3332 (broad), 2245, 1662. MS: 621 [M + Na]+. HRMS:
calcd for C35H42N2O5SiNa [M + Na]+ 621.2761; found 621.2759.
Preparation of 45. A 1,2-dichloroethane (7.0 mL) solution of 43
(0.473 g, 0.79 mmol, 1.0 equiv), DIPEA (0.34 mL, 1.98 mmol, 2.5
equiv), and BOMCl (0.22 mL, 1.58 mmol, 2.0 equiv) was stirred
under Ar at 50 °C for 24 h, and then it was cooled to room
temperature, diluted with CH2Cl2 (20 mL), and partitioned with
saturated aqueous NH4Cl solution (15 mL). The organic layer was
separated, and the aqueous phase was extracted with more CH2Cl2 (2
× 20 mL). The combined extracts were washed with H2O (50 mL),
dried (Na2SO4), and concentrated under vacuum to afford crude 44 as
a pale yellow oil. This material was taken up in THF (5.0 mL) at room
temperature under Ar and treated with a 1 M solution of AcOH in
THF (2.0 mL, 1.98 mmol, 2.5 equiv) followed by a 1 M solution of
TBAF in THF (2.0 mL, 1.98 mmol, 2.5 equiv). The mixture was
stirred at room temperature for 15 h, and then it was diluted with
saturated aqueous NaHCO3 solution (10 mL) (Caution! CO2
evolution) and diluted with EtOAc (20 mL). The layers were
separated, and the aqueous phase was extracted with more EtOAc (2 ×
20 mL). The combined extracts were dried (Na2SO4) and
concentrated under vacuum, and the residual pale yellow oil was
purified by column chromatography on silica gel (EtOAc) to give 45
(0.329 g, 0.66 mmol, 84% over two steps) as a colorless oil.
5.27 (br s, 1H); 4.89−4.78 (m, 5H); 4.71−4.59 (m, 5H); 3.86−3.82
(m, 2H); 2.50−2.47 (m, 2H); 1.95 (s, 3H). 13C NMR: 170.1, 137.4,
137.3, 136.4, 131.3, 128.6, 128.5, 128.2, 128.0, 127.9, 127.8, 127.8,
119.4, 118.6, 94.9, 91.1, 71.6, 69.9, 69.9, 64.3, 55.6, 38.6, 37.2, 24.2. IR:
3300, 2244, 1659. MS: 499 [M + Na]+. HRMS: calcd for
C28H32N2O5Na [M + Na]+499.2209; found 499.2189.
ASSOCIATED CONTENT
* Supporting Information
■
S
Text giving additional characterization data, figures giving
proton and 13C NMR spectra, and CIF files giving X-ray
crystallographic data for 21 and 37. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail for M.A.C.: ciufi@chem.ubc.ca.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
Data for 45 are as follows. 1H NMR: 7.40−7.28 (m, 10H); 6.64 (br
s, 1H); 6.00 (dd, J = 10.23, 4.44, 1H); 5.81 (d, J = 10.23, 1H); 4.87 (br
s, 2H); 4.78 (br s, 2H); 4.67 (br s, 2H); 4.62 (br s, 2H); 4.53 (d, J =
3.0, 1H); 4.28−4.18 (m, 1H); 4.14−4.11 (m, 1H); 3.89−3.63 (m,
3H); 2.58−2.49 (m, 1H); 2.28−2.20 (m, 1H); 1.92 (s, 3H). 13C
NMR: 170.4, 137.4, 137.0, 130.2, 128.6, 128.5, 128.3, 128.0 (2
signals), 127.7, 118.5, 95.0, 94.8, 70.3, 70.0, 65.4, 63.9, 54.8, 36.8, 35.8,
24.1. IR: 3335, 2244, 1658. MS: 503 [M + Na]+. HRMS: calcd for
C27H32 N2O6Na [M + Na]+ 503.2158; found 503.2155.
Preparation of 46. Solid Dess−Martin periodinane (0.874 g, 2.06
mmol, 1.5 equiv) was added at room temperature to a CH2Cl2 (10
mL) solution of 45 (0.660 g, 1.37 mmol, 1.0 equiv), and the mixture
was stirred for 30 min. An aqueous solution of Na2S2O3/NaHCO3 (7/
1, 20 mL) was added, and the mixture was diluted with Et2O (20 mL)
and stirred until the layers turned clear (ca. 15 min). The layers were
separated, and the aqueous phase was extracted with more Et2O (2 ×
20 mL). The combined extracts were dried (Na2SO4) and
concentrated under vacuum, and the residue was purified by column
chromatography on silica gel (EtOAc/hexanes 7/3) to give 46 (0.618
g, 1.29 mmol, 94%) as a colorless oil.
■
We thank the University of British Columbia, the Canada
Research Chair program, NSERC, CIHR, and BCKDF for
support of our program.
REFERENCES
■
(1) Reviews: (a) Koert, U. Angew. Chem., Int. Ed. 2004, 43, 5572.
(b) Narahashi, T. Proc. Jpn. Acad., Ser. B 2008, 84. (c) Hucho, F.
Angew. Chem., Int. Ed. 1995, 34, 39.
(2) Reviews: (a) Hucho, F. Angew. Chem., Int. Ed. 1995, 34, 39.
(b) Kim, Y. H.; Kim, Y. B.; Yotsu-Yamashita, M. J. Toxicol 2003, 22,
521.
(3) Isolation and structural elucidation: (a) Woodward, R. B. Pure
Appl. Chem. 1964, 9, 49. (b) Tsuda, K.; Ikuma, S.; Kawamura, M.;
Tachikawa, R.; Sakai, K.; Tamura, C.; Amakasu, O. Chem. Pharm. Bull.
1964, 12, 1357. (c) Goto, T.; Kishi, Y.; Takahashi, S.; Hirata, Y.
Tetrahedron 1965, 21, 2059.
(4) The toxin has since been found in various other animals,
including certain blue-ringed octopi, sea stars, newts, and toads, and
worms. For reviews of TTX chemical and evolutionary ecology see:
(a) Williams, B. L. Mar. Drugs 2010, 8, 381. (b) Hanifin, C. T. Mar.
Drugs 2010, 8, 577.
(5) Recent reviews: (a) Chau, J.; Ciufolini, M. A. Mar. Drugs 2011, 9,
2046. (b) Nishikawa, T.; Isobe, M. Chem. Rec. 2013, 13, 286.
(6) (a) Kishi, Y.; Aratani, M.; Fukuyama, T.; Nakatsubo, F.; Goto, T.;
Inoue, S.; Tanino, H.; Sugiura, S.; Kakoi, H. J. Am. Chem. Soc. 1972,
94, 9217. (b) Kishi, Y.; Fukuyama, T.; Aratani, M.; Nakatsubo, F.;
Goto, T.; Inoue, S.; Tanino, H.; Sugiura, S.; Kakoi, H. J. Am. Chem.
Soc. 1972, 94, 9219−9221.
(7) (a) Ohyabu, N.; Nishikawa, T.; Isobe, M. J. Am. Chem. Soc. 2003,
125, 8798. (b) Hinman, A.; Du Bois, J. J. Am. Chem. Soc. 2003, 125,
11510. (c) Nishikawa, T.; Urabe, D.; Isobe, M. Angew. Chem., Int. Ed.
2004, 43, 4782. (d) Sato, K.; Akai, S.; Sugita, N.; Ohsawa, T.; Kogure,
T.; Shoji, H.; Yoshimura, J. J. Org. Chem. 2005, 70, 7496. (e) Sato, K.
I.; Akai, S.; Shoji, H.; Sugita, N.; Yoshida, S.; Nagai, Y.; Suzuki, K.;
Nakamura, Y.; Kajihara, Y.; Funabashi, M.; Yoshimura, J. J. Org. Chem.
2008, 73, 1234. (f) Akai, S.; Seki, H.; Sugita, N.; Kogure, T.;
Nishizawa, N.; Suzuki, K.; Nakamura, Y.; Kajihara, Y.; Yoshimura, J.;
Sato, K. Bull. Chem. Soc. Jpn. 2010, 83, 279.
(8) (a) Nishikawa, T.; Urabe, D.; Yoshida, K.; Iwabuchi, T.; Asai, M.;
Isobe, M. Chem. Eur. J. 2004, 10, 452. (b) Nishikawa, T.; Urabe, D.;
Yoshida, K.; Iwabuchi, T.; Asai, M.; Isobe, M. Pure Appl. Chem. 2003,
75, 251. (c) Nishikawa, T.; Asai, M.; Ohyabu, N.; Yamamoto, N.;
Isobe, M. Angew. Chem., Int. Ed. 1999, 38, 3081. (d) Asai, M.;
Nishikawa, T.; Ohyabu, N.; Yamamoto, N.; Isobe, M. Tetrahedron
2001, 57, 4543−4558. (e) Nishikawa, T.; Urabe, D.; Yoshida, K.;
Data for 46 are as follows. 1H NMR: 7.42−7.30 (m, 10H); 7.16 (br
s, 1H); 6.92 (d, J = 10.59, 1H); 6.07 (d, J = 10.59, 1H); 5.03 (d, J =
6.99, 1H); 4.88−4.83 (m, 4H); 4.73 (d, J = 4.29, 1H); 4.67−4.63 (m,
3H); 4.45 (d, J = 4.29, 1H), 3.93−3.78 (m, 2H); 2.53−2.48 (m, 2H);
1.95 (s, 3H). 13C NMR: 191.0, 170.7, 154.3, 137.5, 137.3, 128.8, 128.6,
128.4, 128.3, 128.1, 127.9, 126.8, 117.1, 95.3, 93.9, 71.5, 70.7, 70.4,
64.1, 55.5, 39.8, 38.3, 23.8. IR: 3347, 2246, 1678, 1660 (shoulder).
MS: 501 [M + Na]+. HRMS: calcd for C27H30N2O6Na [M + Na]+
501.2002; found 501.1989.
Preparation of 47. Commercial n-BuLi solution (2.5 M in
hexanes, 0.50 mL, 1.25 mmol, 3.0 equiv) was carefully added (syringe)
to a THF (3.0 mL) suspension of Ph3PCH3Br (0.747 g, 2.09 mmol,
5.0 equiv) at room temperature, under Ar, with good stirring. After 1
h, the mixture was cooled to −78 °C, and a THF (3.0 mL) solution of
46 (0.200 g, 0.42 mmol, 1.0 equiv) was added dropwise (syringe).
Upon completion of the addition, the reaction was stirred at −78 °C
for 5 min, and then it was warmed to 0 °C and stirred for 2.5 h. The
mixture was carefully treated with saturated aqueous NH4Cl solution
(20 mL), and then it was warmed to room temperature and diluted
with Et2O (20 mL). The layers were separated, and the aqueous phase
was extracted with more Et2O (2 × 20 mL). The combined extracts
were dried (Na2SO4) and concentrated under vacuum, and the
residual pale yellow oil was purified by column chromatography on
silica gel (Et2O) to give 47 (0.173 g, 0.36 mmol, 87%) as a colorless
oil.
Data for 47 are as follows. 1H NMR: 7.41−7.28 (m, 10H); 6.91 (br
s, 1H); 6.23 (d, J = 10.3, 1H); 5.80 (d, J = 10.3, 1H); 5.32 (br s, 1H);
H
dx.doi.org/10.1021/jo401960p | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Iwabuchi, T.; Asai, M.; Isobe, M. Org. Lett. 2002, 4, 2679.
(f) Nishikawa, T.; Asai, M.; Isobe, M. J. Am. Chem. Soc. 2002, 124,
7847.
(9) Review of C−H insertion reactions of carbenoids and nitrenoids:
Gutekunst, W. R.; Baran, P. S. Chem. Soc. Rev. 2011, 40, 1976.
(10) Review: Ciufolini, M. A.; Braun, N. A.; Canesi, S.; Ousmer, M.;
Chang, J.; Chai, D. Synthesis 2007, 3759.
(11) Chau, J.; Ciufolini, M. A. Org. Synth. 2013, 90, 190.
(12) Mendelsohn, B. A.; Ciufolini, M. A. Org. Lett. 2009, 11, 4736.
(13) (a) Andersen, S. H.; Das, N. B.; Jørgensen, R. D.; Kjeldsen, G.;
Knudsen, J. S.; Sharma, S. C.; Torssell, K. B. G. Acta. Chem. Scand. B
1982, 36, 1. (b) Torssell, K. B. G. Nitrile Oxides, Nitrones and
Nitronates in Organic Synthesis; VCH: New York, 1988.
(14) For a related reaction see: Itoh, T.; Watanabe, M.; Fukuyama, T.
Synlett 2002, 1323.
(31) Baker, D. C.; Putt, S. R. Synthesis 1978, 478. Compound 9 was
first made by Dr. Cyril Benhaim during his tenure in this group. The
procedure described herein (see the Experimental Section) incorpo-
rates minor modifications relative to the one detailed in ref 12.
(32) As described in ref 12, attempted dehydration of the
nitroketones to the corresponding α-oxonitrile oxide with, e.g., 4-
chlorophenyl isocyanate32a or BOC2O32b resulted in formation of the
desired isoxazolines as minor components of a complex mixture. We
thank our former coworkers, Dr. Cyril Benhaim and Mr. (now Dr.)
Brian Mendelsohn, for carrying out exploratory experiments in this
area. (a) Mukaiyama, T.; Hoshino, T. J. Am. Chem. Soc. 1960, 82,
5339. (b) Basel, Y.; Hassner, A. Synthesis 1997, 309 .
(33) (a) Cecchi, L.; De Sarlo, F.; Machetti, F. Chem. Eur. J. 2008, 14,
7903. (b) Trogu, E.; De Sarlo, F.; Machetti, F. Chem. Eur. J. 2009, 15,
7940.
(34) Control experiments confirmed that Cu(OAc)2 was required.
Although simpler nitroketones related to 9 can be induced to cyclize in
the presence of amine bases only, we had previously determined that
treatment of 9 with, e.g., plain Et3N produced no 10.12 Cecchi, L.; De
Sarlo, F.; Machetti, F. Tetrahedron Lett. 2005, 46, 7877.
(15) For full details see: Chau, J. M.S. Thesis, University of British
Columbia, 2012.
(16) Techniques that should enable the translation of the chemistry
described herein to an enantioselective regime have been described:
Mendelsohn, B.; Lee, S.; Kim, S.; Teyssier, F.; Aulakh, V. S.; Ciufolini,
M. A. Org. Lett. 2009, 11, 1539.
(17) This is to prevent an undesirable Michael cyclizations of
intermediates such as 2-(1-acetamido-4-oxocyclohexa-2,5-dien-1-yl)
acetic acid and N-(1-(3-nitro-2-oxopropyl)-4-oxocyclohexa-2,5-dien-1-
yl)acetamide.
(18) Confirmed by the X-ray crystal structure of a derivative:
Mendelsohn, B. Dissertation, University of British Columbia, 2010.
(19) Reviews: (a) Itsuno, S. Org. React. 1998, 52, 395. (b) Corey, E.
J.; Helal, C. J. Angew. Chem., Int. Ed. 1998, 37, 1986.
(20) The CBS reduction also alleviated technical difficulties arising
from the polar nature and aqueous solubility of 5 and of alcohols 13, as
well as the sensitivity of the latter to acidic and basic conditions and
their ability to coordinate Al(III) species.
(35) Inferior results were obtained with N-methylmorpholine (20%
yield of isoxazoline), imidazole (11% yield), or DABCO (no product
detected).
(36) The use of freshly distilled CHCl3 (removal of stabilizing
EtOH) was key to maximum efficiency.
(37) The published procedure33 calls for operation at 60 °C.
However, α-9 yielded numerous byproducts at such a temperature. A
cleaner reaction was observed at 30 °C, but at the detriment of rate.
Interestingly, epimer β-9 tolerated higher reaction temperatures,
advancing to β-10 in 40% yield after only 2.5 days at 60 °C, with
marginal formation of byproducts.
(38) The formation of the isoxazoline from the nitroketone must
necessarily generate 1 equiv of H2O.
(21) Indirect structural proof for α-13 was subsequently obtained by
single-crystal X-ray diffractometry of compound 37 (vide infra).
(22) The observation that DIBAL and CBS reductions produce
opposite stereochemical outcomes was initially made by our former
co-worker: Liang, H. Dissertation, University of British Columbia,
2009. Early reduction experiments using the CBS method were
carried out by our former coworker, Mr. (now Dr.) Brian Mendelsohn:
Mendelsohn, B. Ph.D. Dissertation; University of British Columbia,
Vancouver, 2010.
(23) Examples of such 1,2-additions are as follows. (a) Compound
10 in: Nagaoka, H.; Miyakoshi, T.; Yamada, Y. Tetrahedron Lett. 1984,
25, 3621. (b) Compound 3 in: Swiss, K. A.; Liotta, D. C.; Maryanoff,
C. A. J. Am. Chem. Soc. 1990, 112, 9393. Examples of such 1,4-
additions are as follows. (c) Compound 9 in: Canesi, S.; Bouchu, D.;
Ciufolini, M. A. Angew. Chem., Int. Ed. 2004, 43, 4336.
(24) (a) Eliel, E. L.; Wilen, S. H. Stereochemistry of Organic
Compounds; Wiley: New York, 1994; p 875.
(39) These reactions were carried out with ca. 30 mg of nitroketone
and 0.6 mg of Cu(OAc)2 (weighed as a solid) dissolved in 0.6 mL of a
stock solution of N-ethylpiperidine (3.0 μL) in CDCl3 (2 mL). This
solution thus contained 0.9 μL of N-ethylpiperidine.
(40) The amount of (paramagnetic) Cu(II) present in the mixture
induced an insignificant extent of line broadening, enabling close
1
monitoring of the progress of the reaction by 300 MHz H NMR.
(41) Nitromethane can be adsorbed/chemisorbed onto strongly
basic earth oxides such as CaO or MgO: Kheir, A. A.; Haw, J. F. J. Am.
Chem. Soc. 1994, 116, 817. It seemed unlikely that nonbasic CaSO4
might also adsorb/chemisorb MeNO2. However, if that were the case,
the NMR analysis described herein would produce meaningless results.
The following experiment unequivocally ruled out such a possibility.
Two 0.6 mL aliquots of a stock solution of MeNO2 (3 μL) in CDCl3
(1.5 mL) were separately syringed into two NMR tubes, one of which
contained powdered anhydrous CaSO4 (ca. 20 mg). Both solutions
were sonicated for 5 min, and a 1H NMR spectrum of each was
recorded. The ratio of the integrated areas under the signals of
CH3NO2 and residual CHCl3 were identical in both solutions,
signifying that no sequestration of MeNO2 had occurred. In contrast,
the signal of residual H2O had been largely suppressed in the sample
containing CaSO4.
(25) This preference may be reversed by the use of appropriate
techniques. See ref 23b as well as, e.g.: (a) Imbos, R.; Brilman, M. H.
G.; Pineschi, M.; Feringa, B. L. Org. Lett. 1999, 1, 623. (b) Merino, E.;
Melo, R. P. A.; Ortega-Guerra; Ribagorda, M.; Carreno, M. C. J. Org.
Chem. 2009, 74, 2824 and literature cited therein.
(26) For an analogous proposal, see Figure 1 in: Carreno, M. C.;
Perez Gonzalez, M.; Ribagorda, M.; Houk, K. N. J. Org. Chem. 1998,
63, 3687.
(27) The hydrophobic TBDPS group diminished the aqueous
solubility of all subsequent intermediates.
(42) This material was not thoroughly purified, and it was
characterized by 1H NMR and low- and high-resolution mass
spectrometry. The identity was confirmed by comparison with a
sample prepared from commercial 4-acetamidophenol (see the
Supporting Information).
(28) Review on the synthesis and chemistry of α-nitroketones:
Ballini, R.; Bosica, G.; Fiorini, D.; Palmieri, A. Tetrahedron 2005, 61,
8971.
(43) For relevant discussion see: Bodanszky, M. Principles of Peptide
Synthesis; Springer-Verlag: Berlin, Germany, 1993; see especially p
173ff.
(29) As detailed in ref 12, careful control of the pH during
acidification of the saponification mixture was necessary to prevent
cyclization of the acids to an undesired lactonic byproduct (cf.
compound 19 in ref 12).
(44) Carreno, M. C.; Ortega-Guerra, M.; Ribagorda, M.; Sanz-
Cuesta, M. J. Chem. Eur. J. 2008, 14, 621.
(45) This step occurred much more efficiently when MeOH freshly
distilled from Mg turnings was employed as the solvent, instead of
commercial dry MeOH.
(30) Staab, H. A. Angew. Chem., Int. Ed. Engl. 1962, 1, 351.
I
dx.doi.org/10.1021/jo401960p | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
(46) The X-ray crystal structure of 37 is described in the Ph.D.
dissertation of our former co-worker, Mr. (now Dr.) B. Mendelsohn,
who prepared it by the Torsell method.13,22 The compound
cocrystallized with one molecule of CH2Cl2, but only the structure
of 37 is shown in Figure 3. For full details, see the Supporting
Information.
No attempt was made to obtain a better spectrum by altering the
relaxation delay between pulses. IR: 3325, 1709, 1656. MS: 530 [M +
Na]+; negative ion mode 506 [M − H]−. HRMS: calcd for
C29H37NO5NaSi [M + Na]+ 530.2339; found 530.2346.
(47) Chromatographic purification of batches of isoxazolines
obtained from multigram-scale runs of the above sequences would
occasionally return product contaminated with 15−20% of acids 8/22.
This was somewhat surprising, since acids and isoxazolines exhibit
widely differing chromatographic mobilities (e.g., in 70% EtOAc/
hexanes: 22, Rf = 0.00; 36, Rf = 0.55). Rather than further purification
of the isoxazolines, it was expedient to remove the contaminants at the
stage of 37/38 to minimize losses. Indeed, the basic treatment and
subsequent chromatography of 37/38 eliminated all traces of residual
acids. See the Experimental Section for details.
(48) Recent work has revealed that the N-BOC compound 38 can be
elaborated to very advanced TTX intermediates without passing
through an analogous diene. Therefore, the sequence leading to the
latter was studied only in the N-acetyl series.
(49) It is worthy of note that while the BOM protection of 37
proceeded efficiently to afford 39 in 91% yield after 48 h, the same
reaction of its epimer 11 was slow and could not be forced to
completion without incurring unacceptable losses. Furthermore, the
chromatographic mobility of the BOM derivative of 11 was similar to
that of the starting alcohol, complicating purification. This is one of
the reasons it was advantageous to use α-diastereomers 7−10/21−23
in sequences leading to advanced TTX intermediates.
(50) (a) Parikh, J. R.; Doering, W. V. E. J. Am. Chem. Soc. 1967, 89,
5505. (b) Review: Tidwell, T. T. Org. React. 1990, 39, 297.
(51) Because this ketone proved to be a dead-end compound, the
sequence leading to it was not optimized, nor was 40 fully
characterized.
(52) This substance was not thoroughly purified, and it was
1
characterized only by 300 MHz H NMR (CDCl3: δ 7.45−7.32 (m,
5H); 7.15 (d, J = 8.46, 1H); 7.04 (d, J = 8.46, 1H); 5.32 (s, 2H); 4.93
(s, 2H); 3.76 (s, 2H); 3.74 (s, 3H)) and ESI mass spectrometry (m/z
350 (M + Na)+).
(53) This reduction was slow and required occasional addition of
more LiBH4 over 24 h.
(54) The action of unbuffered TBAF on 44 returned 45 in only
about 20% yield, as did treatment with HF·(pyridine).
(55) Kofron, W. G.; Baclawski, L. M. J. Org. Chem. 1976, 41, 1879.
(56) A small amount of nitroketone 23 was purified by flash
chromatography (1/9 EtOAc/hexanes). In CDCl3 solution, about 15%
of 23 appears to exist as the enol tautomer of the ketone (presumably,
1
internally H-bonded to the NO2 group) on the basis of H and 13C
1
NMR. Data for 23 are as follows. H NMR: 7.71−7.67 (m, 4H);
7.46−7.37 (m, 6H); 6.73 (s, 0.15 H, presumed enol tautomer); 5.99
(app dd, J = 10.3 Hz, 1.8 Hz, 2H); 5.81 (app dd, J = 10.3 Hz, 2.6 Hz,
2H); 5.26 (s, 1.7 H); 4.61 (br, 1H); 4.56 (m, 1H); 3.02 (s, 1.7 H);
2.77 (s, 0.3 H, presumed enol tautomer); 1.45 (s, 9H); 1.08 (s, 9H).
13C NMR: 193.3, 171.3 (presumed enol tautomer), 155.1, 154.6
(presumed enol tautomer), 136.0, 133.8, 130.3, 130.0, 128.3, 127.8,
118.1 (presumed enol tautomer), 83.9, 80.5, 63.5 (presumed enol
tautomer), 63.4, 52.2 (presumed enol tautomer), 51.1, 48.1, 42.4
(presumed enol tautomer), 28.4, 27.0, 19.3. IR: 3409, 1703, 1560,
1493. MS: 573 [M + Na]+. HRMS: calcd for C30H36N2O5NaSi [M +
Na]+ 573.2397; found 573.2390.
(57) Chromatographic purification of 22 is quite difficult. A pure
sample was obtained by cooling a hot saturated CH2Cl2 solution of the
acid, whereupon the compound crystallized as a powdery solid, which
1
was filtered and washed twice with CH2Cl2, mp 157−158 °C. H
NMR: 7.71−7.66 (m, 4H); 7.48−7.36 (m, 6H); 5.93 (dd, J = 10.3 Hz,
1.6 Hz, 2H); 5.81 (dd, J = 10.3 Hz, 2.5 Hz, 2H); 4.52 (m, 1H); 2.64
(br, 2H); 1.45 (s, 9H); 1.07 (s, 9H). 13C NMR: 173.1, 136.0, 133.9,
130.0, 129.5 (br, 2 signals), 127.8, 81.2 (br), 63.5, 51.1, 44.2 (br), 28.5,
27.0, 19.32. The carbonyl carbon of the BOC group at ca. 155 ppm
was barely visible, presumably due to spin saturation/slow relaxation.
J
dx.doi.org/10.1021/jo401960p | J. Org. Chem. XXXX, XXX, XXX−XXX