Total synthesis of fluoxetine and duloxetine through an in situ imine formation/borylation/transimination and reduction approach

Article abstract of DOI:10.1039/c4ob01142b

We report efficient, catalytic, asymmetric total syntheses of both (R)-fluoxetine and (S)-duloxetine from α,β-unsaturated aldehydes conducting five sequential one-pot steps (imine formation/copper mediated β-borylation/transimination/reduction/oxidation) followed by the specific ether group formation which deliver the desired products (R)-fluoxetine in 45% yield (96% ee) and (S)-duloxetine in 47% yield (94% ee). This journal is the Partner Organisations 2014.

Full text of DOI:10.1039/c4ob01142b

Organic &
Biomolecular Chemistry
View Article Online
View Journal | View Issue
PAPER
Total synthesis of fluoxetine and duloxetine
through an in situ imine formation/borylation/
transimination and reduction approach†
Cite this: Org. Biomol. Chem., 2014,
12, 6121
Adam D. J. Calow,a Elena Fernández*b and Andrew Whiting*a
We report efficient, catalytic, asymmetric total syntheses of both (R)-fluoxetine and (S)-duloxetine from
α,β-unsaturated aldehydes conducting five sequential one-pot steps (imine formation/copper mediated
β-borylation/transimination/reduction/oxidation) followed by the specific ether group formation
which deliver the desired products (R)-fluoxetine in 45% yield (96% ee) and (S)-duloxetine in 47% yield
(94% ee).
Received 3rd June 2014,
Accepted 30th June 2014
DOI: 10.1039/c4ob01142b
www.rsc.org/obc
Introduction
Fluoxetine 1 and duloxetine 2, developed by Eli Lilly,1 are top-
selling pharmaceuticals used for the treatment of major
depressive disorder (MDD)2 and other conditions.3 Fluoxetine
1 belongs to the selective serotonin reuptake inhibitor (SSRI)
class of anti-depressants4 and duloxetine 2 to the serotonin-
norepinephrine reuptake inhibitor (SNRI) class.5
Fig. 1 Molecular structure of fluoxetine 1 and duloxetine 2.
Due to the success and importance of these drugs, several
groups have been interested in their preparation. An original
asymmetric approach to fluoxetine 1 was developed by Brown
et al., using the chiral auxiliary diisopinocamphenylchloro-
borane, for the asymmetric reduction of ketone precursors.6
Sharpless et al. also developed a route to fluoxetine using an
asymmetric epoxidation of an allylic alcohol, followed by
ring-opening strategy.7 Corey et al. achieved an asymmetric
reduction using the chiral oxazaborolidine (CBS reduction) in
combination with borane to reduce a prochiral ketone in this
approach.8 In recent years, the advancement of asymmetric
catalytic hydrogenation has also proven highly effective for the
asymmetric reduction of ketones (e.g. Noyori et al.)9 and,
indeed, other groups have employed this methodology to the
synthesis of both fluoxetine 1 and duloxetine 2.10
pharmacological and pharmacokinetic properties depending
on the enantiomer of fluoxetine 1.12 This evidence suggests
that the (S)-enantiomer of fluoxetine 1 is more active in the
inhibition of serotonin than the (R)-enantiomer.12 Addition-
ally, one of the major metabolites of fluoxetine 1, norfluoxe-
tine (demethylated fluoxetine), is significantly more active as
an inhibitor. In contrast, duloxetine 2 is marketed as a single
(S)-enantiomer.13
Herein, we report an efficient, catalytic, asymmetric syn-
thesis of fluoxetine and duloxetine with key steps that involve:
(1) an in situ imine formation, (2) a copper-catalysed asym-
metric β-borylation protocol that requires a specific bulky
amine to block the imine functionality and prevent 1,2
addition versus 1,4 addition of the Cu-Bpin system, (3) a
sequential transimination reaction, (4) a reduction of CvN
bond and (5) a C–B oxidation protocol. Interestingly, since the
asymmetry is induced in the second step by using a cheap
chiral ligand (R/S)-dimethyl-BINAP [(R/S)-DM-BINAP)], another
key point is the prevalence of the asymmetric induction along
the following synthetic steps towards the target product.
It is interesting to note that fluoxetine 1, despite being
a chiral compound, is marketed as the racemic HCl-salt
(Fig. 1).11 However, studies have revealed evidence of differing
aCentre for Sustainable Chemical Processes, Department of Chemistry,
Durham University, South Road, Durham, DH1 3LE, UK.
E-mail: andy.whiting@durham.ac.uk; Fax: +44 (0)191 384 4737;
Tel: +44 (0)191 334 2081
bDepartament Química Física I Inorgànica, University Rovira I Virgili, C/Marcel.lí
Results and discussion
Domingo s/n 43007, Tarragona, Spain
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4ob01142b
In recent years we have been interested in the preparation of
γ-amino alcohols (e.g. 7-)14 via the β-borylation15 of α,β-unsatu-
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 6121–6127 | 6121

View Article Online
Paper
Organic & Biomolecular Chemistry
[Pd–C (10%)], because this has been employed for standard
debenzhydrylation in the literature [see eqn (1)].19 However,
using this methodology we encountered significant C–O bond
hydrogenolysis,20 i.e. cleavage of the benzylic hydroxyl-group,
which led to the formation of 10a as a significant product, in
addition to the formation of the desired 8a. We therefore con-
sidered transfer hydrogenation as a suitable method, due to
the practical ease of delivering stoichiometric amounts of
hydrogen in situ from the decomposition of ammonium
formate [see eqn (1)]. However, this resulted in the formation
of a mixture of 7a, 8a, and 10a. Increased loadings of
ammonium formate resulted in 10a being the primary
product, with complete N-benzhydryl group cleavage. Other
milder methods, such as hydrogenation via Wilkinson’s cata-
lysis and, indeed, refluxing TFA, resulted in no debenzhydryla-
tion. To our disappointment, conventional debenzhydrylation
methodologies appeared to be too harsh for substrate 7a due
to the presence of the benzylic hydroxyl-group, which appears
to undergo facile hydrogenolysis under palladium-catalysed
hydrogenation conditions. In addition to debenzhydrylation,
hydrolysis of compound 5a to the analogous aldehyde, with
subsequent reductive amination using methylamine–NaBH4
(to yield 12a) was attempted and indeed did work, but due to
the instability of the analogous β-boryl aldehyde the overall
conversion in this case was low (<20%) and, hence, we needed
to avoid the utilisation of such β-boryl aldehydes as intermedi-
ates in subsequent synthesis.
Scheme 1 Asymmetric one-pot methodology towards chiral γ-amino
alcohols.
ð1Þ
Scheme 2 Retrosynthetic analysis of fluoxetine 1.
rated imines (e.g. 4-) because such γ-amino alcohols have
applications as auxiliaries in synthetic and biochemical
systems.16 In this context, we demonstrated a novel protocol
for the asymmetric β-borylation of enal-derived α,β-unsaturated
aldimines.17 This methodology owes its success to the steri-
cally bulky N-benzhydryl substituent, which favours exclusive
1,4-boron addition (enals are prone to 1,2-boron addition to
the carbonyl). The resulting β-boryl imines 5- can be reduced
and oxidised in one-pot to yield N-benzhydryl γ-amino
alcohols 7- with ee values up to 97% (Scheme 1).
We therefore became interested in applying our one-pot
methodology (Scheme 1) to the total synthesis of some pharma-
ceuticals, such as fluoxetine 118 and duloxetine 2. By apply-
ing our retrosynthetic analysis to fluoxetine 1, one can clearly
see that γ-amino alcohol 7a is an appropriate precursor to
fluoxetine 1. Indeed, we considered that a debenzhydrylation,
N-methylation and, finally, a nucleophilic aromatic substi-
tution would result in the target compound 1 (Scheme 2).
Initially, we prepared compound 7a from cinnamaldehyde
3a using CuCl/L0 or CuC/L1 (L0 = PPh3 and L1 = (R)-
DM-BINAP) as the catalytic system.17 Our initial hypothesis to
transform 7a into 1 required a debenzhydrylation step using
hydrogen over a palladium-on-carbon heterogeneous catalyst
Intrigued by recent reports of transimination21 (also known
as imine-metathesis22), we wondered whether treating the
β-boryl imine 5a (Scheme 1) with an excess of methylamine,
would result in the formation of N-methyl imine 11a. More
specifically, could the equilibrium between N-benzhydryl
imine 5a and N-methyl imine 11a, on addition of methyl-
amine, be directed towards the formation of 11a as a result of
the difference in amine nucleophilicity of methylamine and
benzhydrylamine (Scheme 3)? If successful, this would bypass
the need for forming the parent β-boryl aldehyde23 simply by
the addition of cheap and readily available methylamine.
Continuing with the established one-pot methodology
(Scheme 4), we therefore treated the intermediate β-boryl
imine 5a with excess methylamine (4 equiv.), followed by
in situ reduction using NaBH4–MeOH. Subsequently, solvent
Scheme 3 Proposed transimination through amine exchange.
6122 | Org. Biomol. Chem., 2014, 12, 6121–6127
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 5 Yun et al.’s formal synthesis of fluoxetine.18
Scheme 4 Asymmetric synthesis of (R)-fluoxetine 1. Conditions: (a)
Ph2CHNH2 (1 equiv.), 3 Å-MS, THF, r.t., 6 h; (b) CuCl (3%), PPh3 L0 (6%),
NaOtBu (9%), B2pin2 (1.1 equiv.), MeOH (2.5 equiv.), r.t., 15 h or CuCl
(3%), (R)-DM-BINAP L1 (3%), NaOtBu (9%), B2pin2 (1.1 equiv.), MeOH (2.5
equiv.), r.t., 15 h; (c) MeNH2 (4 equiv.), r.t., 2 h; (d) NaBH4, MeOH, r.t., 3 h.
Remove solvent in vacuo; (e) H2O2, NaOH, THF, reflux, 1 h; (f) NaH (1.1
equiv.), DMA, 70 °C, 30 min. Ar–Cl (1.2 equiv.), 100 °C, 3 h.
removal (to prevent MeOH oxidation to formaldehyde, which
in the presence of γ-amino alcohols leads to the formation of
1,3-oxazines, as previously reported24) and replacement with
THF, followed by B–C oxidation with H2O2–NaOH of boronate
12a, gave the known precursor to fluoxetine, γ-amino alcohol
9a [54% yield when using PPh3 L0 and 61% when using
(R)-DM-BINAP L1, see Scheme 4]. This was achieved in five-steps,
all of which were conducted in one-pot, without intermediate
purification. Next, the addition of NaH to 9a resulted in the
in situ generation of the analogous Na-alkoxide of 9a which, on
addition of 4-chlorobenzotrifluoride at elevated temperature
(100 °C, 3 h), gave fluoxetine (rac)-1 in 74% yield [(R)-1 in 96%
ee when using L1] (Scheme 4). Determination of the enantio-
meric excess was carried out by chiral HPLC on the fluoxetine
N-acyl compound 13a (see ESI†), which is consistent with pre-
viously reported values of asymmetric induction (previously
found to be 97% ee).17 It is important to note that recent work
described by Yun et al. on the asymmetric β-borylation of
α,β-unsaturated amides 14, conducted to the formal synthesis
of (S)-fluoxetine with excellent enantioselectivity (99% ee),18
using in this case copper salts modified with a type of chiral
Josiphos ligand (Scheme 5). The intermediate compound 15
Scheme 6 Asymmetric synthesis of (S)-duloxetine 2. Conditions: (a)
Ph2CHNH2 (1 equiv.), 3 Å-MS, THF, r.t., 8 h; (b) CuCl (3%), PPh3 L0 (6%),
NaOtBu (9%), B2pin2 (1.1 equiv.), MeOH (2.5 equiv.), r.t., 15 h or CuCl
(3%), (S)-DM-BINAP L2 (3%), NaOtBu (9%), B2pin2 (1.1 equiv.), MeOH
(2.5 equiv.), r.t., 15 h; (c) MeNH2 (4 equiv.), r.t., 2 h; (d) NaBH4, MeOH, r.t.,
3 h. Remove solvent in vacuo; (e) H2O2, NaOH, THF, reflux, 1 h; (f) NaH
(1.1 equiv.), DMSO, 50 °C, 1.5 h. Ar–Cl (1.2 equiv.), 70 °C, 1.5 h.
dation to the aldehyde (without purification of the intermedi-
ate allylic alcohol) using Swern conditions.26
Hence enal 3b (Scheme 6) was transformed in situ to the
corresponding N-benzhydryl aldimine 4b in the presence of
3 Å-molecular sieves and THF. After 9 hours, the imine was
transferred directly to the pre-catalyst (copper salt, base,
ligand and B2pin2) mixture, followed by the addition of MeOH,
to give the intermediate β-boryl aldimine 5b. Subsequent tran-
simination was achieved through the addition of methylamine
(in THF) which, after in situ borohydride reduction gave 12b.
Again, to prevent the unwanted formation of oxazines
(through in situ formaldehyde formation24), the solvent was
removed in vacuo prior to C–B oxidation and, hence, oxidation
resulted in the formation of the known precursor γ-amino
alcohol 9b in good yield [47% yield when using PPh3 L0 and
57% when using (S)-DM-BINAP L2, see Scheme 6]. Finally,
addition of NaH to 9b resulted in the in situ generation of the
analogous alkoxide of 9b which, on addition of 1-fluoro-
naphthalene at elevated temperature (70 °C, 1.5 h), gave dulox-
etine (rac)-2 in 83% yield [(S)-2 in 94% ee when using L2]
(Scheme 6). The enantiomeric excess was again determined by
chiral HPLC on the N-acetamide 13b of 2 (see ESI†).
25
could be reduced using LiAlH4 to give 9a in quantitative
yields, which could be transformed to fluoxetine using known
procedures (e.g. Scheme 4).7
With these results in hand, we turned our attention to the
total synthesis of duloxetine, which is marketed as the (S)-2
enantiomer. Enal 3b is not commercially available and there-
fore had to be prepared via reduction of the parent acid
(DIBAL-H) to the analogous allylic alcohol, followed by oxi-
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 6121–6127 | 6123

View Article Online
Paper
Organic & Biomolecular Chemistry
α,β-unsaturated imine 4a in situ. After 6 h, an aliquot of the
solution containing the in situ-formed imine 4a (16.0 mL,
Conclusions
In conclusion, we have developed an efficient, catalytic, asym- 4.00 mmol) was transferred to a Schlenk-tube (under argon)
metric route to both (R)-fluoxetine and (S)-duloxetine (45 and containing CuCl (12.0 mg, 0.12 mmol), PPh3 (62.9 mg,
47% overall yield, 96 and 94% ee, respectively) through the 0.24 mmol) or (R)-DM-BINAP (88.2 mg, 0.12 mmol), NaOt-Bu
asymmetric copper-mediated β-borylation of α,β-unsaturated (34.6 mg, 0.36 mmol) and B2pin2 (1.12 g, 4.4 mmol). After
imines. Although this strategy involves six steps, the first five- 5 min, MeOH (400 μL, 10.0 mmol) was added to the solution
steps are conducted following a one-pot strategy. Importantly, and the reaction was stirred overnight. Methylamine (8 mL,
the asymmetric induction provided by CuCl, modified with a 16.0 mmol, 2 M THF solution) was added under argon and the
cheap chiral ligand (R/S)-DM-BINAP L1/L2, is high and is con- resulting solution was stirred for 1.5 h. NaBH4 (0.46 g,
stant along the following transformation towards the targeted 12.0 mmol) was added, followed by the drop-wise addition of
pharmaceuticals. Having demonstrated this approach, further MeOH (8.0 mL). The mixture was stirred for 3 h, followed by
applications are underway and will be communicated in due the removal of solvent under reduced pressure. THF (20 mL)
course.
was added to the resulting residue, followed by NaOH (2.4 mL,
w/v 20%) and H2O2 (1.1 mL, w/v 35%), and the solution was
heated to reflux for 1 h. After cooling, the resulting solution
was partitioned between EtOAc and brine. The aqueous layer
was extracted further with EtOAc (3×). The organic phase was
separated and dried over anhydrous MgSO4. After filtration the
organic phase was removed under reduced pressure to yield a
crude product. Purification by silica gel chromatography
(DCM → DCM–MeOH–NEt3, 5 : 1 : 1%) gave the pure product
as an off colourless oil, which formed an off colourless solid 2
on standing [356 mg, 54% when using PPh3 and 402 mg, 61%
when using (R)-DM-BINAP]: 1H-NMR (400 MHz, CDCl3):
δ 7.40–7.24 (m, 5H), 4.95 (dd, J = 8.7, 3.1 Hz, 1H), 3.65–3.4
(bs, 1H), 2.97–2.83 (m, 2H), 2.46, (s, 3H), 1.93–1.72 (m, 2H);
13C-NMR (101 MHz, CDCl3): δ 145.0, 128.2, 127.0, 125.6, 75.4,
50.3, 36.7, 35.9; LR-MS (ESI+) 166.5 [M + H]+; HR-MS (ESI+)
Calculated [C10H15NO + H]+ 166.1232, found 166.1228. All
spectroscopic values are consistent with those obtained in the
literature.27
Synthesis of 3-(methylamino)-1-(thiophen-2-yl)propan-1-ol
(9b). Benzhydrylamine (0.86 mL, 5.00 mmol) and (2E)-3-(thio-
phen-2-yl)prop-2-enal 3b (0.63 mL, 5.00 mmol) was added to a
stirring solution of THF (20 mL) and oven-dried 3 Å-MS (5.0 g)
for 6 h, to form the α,β-unsaturated imine 4b in situ. After 6 h,
an aliquot of the solution containing the in situ-formed imine
4b (12.0 mL, 3.0 mmol) was transferred to a Schlenk-tube
(under argon) containing CuCl (9.0 mg, 0.09 mmol), PPh3
(48.0 mg, 0.18 mmol) or (S)-DM-BINAP (66.1 mg, 0.09 mmol),
NaOt-Bu (27.0 mg, 0.27 mmol) and B2pin2 (0.84 g, 3.3 mmol).
After 5 min, MeOH (300 μL, 7.5 mmol) was added to the solu-
tion and the reaction was stirred overnight. Methylamine
(6 mL, 12.0 mmol, 2 M THF solution) was added under argon
and the resulting solution was stirred for 1.5 h. NaBH4 (0.34 g,
9.0 mmol) was added, followed by the drop-wise addition of
MeOH (6.0 mL). The mixture was stirred for 3 h, followed by
the removal of solvent under reduced pressure. THF (15 mL)
was added to the resulting residue, followed by NaOH (1.8 mL,
w/v 20%) and H2O2 (0.84 mL, w/v 35%), and the solution was
heated to reflux for 1 h. After cooling, the resulting solution
Experimental
General experimental
All reagents were used as received from the supplier without
further purification, unless stated. All solvents were used as
received from the supplier, except THF (freshly distilled) and
methanol (stored over molecular sieves). Molecular sieves, 3 Å
1–2 mm beads, were supplied from Alfa Aesar, and stored at
220 °C. Reactions were monitored by TLC analysis using
POLTFRAM® SIL G/UV254 (40 × 80 mm) TLC plates. Flash
column chromatography was carried out using Silica gel as
supplied from Sigma-Aldrich (230–400 mesh, 40–63 μm, 60 Å)
and monitored using TLC analysis. 1H NMR spectra were
recorded on a Varian-Mercury 500 MHz spectrometer, operat-
ing at ambient probe temperature unless specified elsewhere.
13C NMR spectra were recorded on a Varian Mercury 500 MHz
instrument, operating at 101 MHz, unless specified elsewhere.
Deuterated chloroform CDCl3 was used as solvent for all NMR
spectra, unless specified elsewhere. NMR peaks are reported as
singlet (s), doublet (d), triplet (t), quartet (q), broad (br), com-
binations thereof, or as a multiplet (m). Mass spectra for
liquid chromatography mass spectrometry (LCMS) were
obtained using a Waters (UK) TQD mass spectrometer (low
resolution ESI+, electrospray in positive ion mode, ESI+),
Waters (UK) Xevo QTOF mass spectrometer (low and high
resolution ASAP+) and a Waters (UK) LCT premier XE (high
resolution ESI+, electrospray in positive ion mode, ESI+) unless
stated elsewhere. HPLC analysis was carried out on an Agilent
1100 series instrument, fitted with a Perkin Elmer series 200
degasser. AS-H-CHIRALCEL column (250 × 4.6 mm) fitted with
guard cartridge (50 × 4.6 mm) was used to achieve chiral
resolution, unless stated elsewhere. Optical rotations were
measured using a JASCO P-1020 polarimeter with [α]D values
given in deg cm2 g−1
.
Experimental procedure
Synthesis of 3-(methylamino)-1-phenylpropan-1-ol (9a). was partitioned between EtOAc and brine. The aqueous layer
Benzhydrylamine (0.86 mL, 5.00 mmol) and cinnamaldehyde was extracted further with EtOAc (3×). The organic phase was
3a (0.63 mL, 5.00 mmol) was added to a stirring solution of separated and dried over anhydrous MgSO4. After filtration the
THF (20 mL) and oven-dried 3 Å-MS (5.0 g) for 6 h, to form the organic phase was removed under reduced pressure to yield a
6124 | Org. Biomol. Chem., 2014, 12, 6121–6127
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
crude product. Purification by silica gel chromatography (d, J = 8.3 Hz, 1H), 7.29 (d, J = 7.9 Hz, 1H), 7.21 (dd, J = 5.0,
(DCM → DCM–MeOH–NEt3, 5 : 1 : 1%) gave the pure product 1.2 Hz, 1H), 7.06 (d, J = 3.5 Hz, 1H), 6.94 (dd, J = 5.0, 3.5 Hz,
as an off colourless oil, which formed a pale yellow oil 9b on 1H), 6.86 (d, J = 7.2 Hz, 1H), 5.79 (dd, J = 7.7, 5.3 Hz, 1H),
standing [241 mg, 47% when using PPh3 and 292 mg, 57% 2.88–2.79 (m, 2H), 2.51–2.40 (m, 2H), 2.44 (s, 3H) ppm.
1
when using (S)-DM-BINAP]: H NMR (400 MHz, CDCl3): δ 7.20 13C NMR (101 MHz, CDCl3): δ 153.4, 145.3, 134.6, 127.5, 126.6,
(dd, J = 5.0, 1.2 Hz, 1H), 7.06 (dd, J = 5.0, 3.4, 1H), 6.93–6.91 126.3, 126.2, 125.7, 125.2, 124.7, 124.5, 122.1, 122.1, 120.6,
(m, 1H), 5.19 (dd, J = 8.4, 3.2 Hz, 1H), 4.68–4.32 (bs, 1H), 107.0, 74.8, 48.4, 39.1, 36.6 ppm. LRMS (ESI+) [M + H]+, 298.0.
3.02–2.83 (m, 2H), 2.45 (s, 3H), 2.05–1.86 (m, 2H) ppm. HRMS (ESI+) calculated [C18H19NOS + H]+ 298.1266, found
13C NMR (101 MHz, CDCl3): δ 149.7, 126.6, 123.7, 122.3, 71.9, 298.1263. [α]2D2 = +105.4 (1.0, MeOH). Enantiomeric excess was
50.1, 36.8, 35.9 ppm. LRMS (ESI+) [M + H]+, 171.9. HRMS determined by derivatisation to 13b. All spectroscopic values
(ESI+) calculated [C8H13NOS + H]+ 172.0796, found 172.0829. are consistent with those obtained in the literature.28
All spectroscopic values are consistent with those obtained
in the literature.13
Synthesis of (2E)-3-(thiophen-2-yl)prop-2-enal (3b). (2E)-3-
(Thiophen-2-yl)prop-2-enoic acid (3.0 g, 19.5 mmol) was dis-
Synthesis of fluoxetine, N-methyl-3-phenyl-3-[4-(trifluoro- solved in THF (80 mL) and cooled to −78 °C under argon.
methyl)phenoxy]propan-1-amine (1). 3-(Methylamino)-1-phenyl- DIBAL-H (58.5 mL, 1 M THF) was added slowly over 1 hour,
propan-1-ol 9a (330 mg, 2.00 mmol) was dissolved in dry and the resulting solution was allowed to react overnight,
dimethylacetamide (2.8 mL) and transferred to an oven-dried warming to room temperature. The resulting solution was
Schlenk-tube and purged with argon. NaH (100 mg, 2.2 mmol, quenched with a saturated aqueous potassium sodium tartrate
60% in mineral oil) was transferred directly to the solution solution and allowed to stir for 1 h. After, the resulting solu-
and heated (70 °C) under argon for 30–40 min, or until hydro- tion was partitioned between EtOAc and the aqueous layer was
gen evolution had ceased. 4-Chlorobenzotrifluoride (354 µL, extracted with EtOAc (3×). The organic phase was separated
2.4 mmol) was added under argon, and the resulting solution and dried over anhydrous MgSO4. After filtration the organic
was heated (100 °C) for 3 h. On cooling, the solution was parti- phase was removed under reduced pressure to yield a crude
tioned between toluene and H2O and washed (3× H2O). The allylic product [(2E)-3(thiophen-2-yl)prop-2-en-1-ol)]. In a sep-
organic phase was separated and dried over anhydrous MgSO4. arate vessel, DMSO (42.9 mmol, 3.0 mL) and DCM (40 mL)
After filtration the organic phase was removed under reduced were combined under argon and cooled to −78 °C). Oxalyl
pressure to yield a crude product. Purification by silica gel chloride (21.5 mmol, 1.8 mL) was added and the reaction
chromatography (DCM → DCM–MeOH–NEt3, 5 : 1 : 1%) gave mixture was stirred for 10 min. The crude allylic alcohol [(2E)-3-
the pure product as a yellow oil 1, (458 mg, 74%): 1H-NMR (thiophen-2-yl)prop-2-en-1-ol)] was added (in DCM, 12 mL) to
(400 MHz, CDCl3): δ 7.43 (d, J = 8.6 Hz, 2H), 7.39–7.24 (m, 5H), the solution at −78 °C, and allowed to stir for 10 min. Triethyl-
6.90 (d, J = 8.6 Hz, 2H), 5.31 (dd, J = 8.2, 4.7 Hz, 1H), 2.79–2.69 amine (97.5 mmol, 13.6 mL) was subsequently added, and the
(m, 2H), 2.43, (s, 3H), 2.26–1.95 (m, 2H) ppm. 13C-NMR solution allowed to warm to room temperature over 1.5 h.
(101 MHz, CDCl3): δ 160.5, 141.0, 128.8, 127.9, 126.8, 126.7, After, the resulting solution was partitioned quenched with
125.8, 115.8, 78.6, 48.2, 38.6, 29.7 ppm. LR-MS (ESI+) 309.3 water and partitioned between EtOAc and the aqueous layer
(57%) [M]+; HR-MS (ESI+) Calculated [C17H18NOF3 + H]+ was extracted with EtOAc (3×). The organic phase was separ-
310.1419, found 310.1411. [α]2D2 = +3.5 (1.0, HCCl3). Enantio- ated and dried over anhydrous MgSO4. After filtration the
meric excess was determined by derivatisation to 13a. All organic phase was removed under reduced pressure to yield a
spectroscopic values are consistent with those obtained in the crude brown oil. Purification by silica gel chromatography
1
literature.27
(hexane–EtOAc, 9 : 1) gave 3b as a yellow oil (996 mg, 37%). H
Synthesis of duloxetine, methyl[3-(naphthalene-1-yloxy)-3- NMR (400 MHz, CDCl3): δ 9.63 (d, J = 7.7 Hz, 1H), 7.58 (d, J =
(thiophen-2-yl)propyl]amine (2). 3-(Methylamino)-1-(thiophen- 15.6 Hz, 1H), 7.51 (d, J = 5.0 Hz, 1H), 7.37 (d, J = 3.7 Hz, 1H),
2-yl)propan-1-ol 9b (150 mg, 0.87 mmol) was dissolved in dry 7.11 (dd, J = 5.1, 3.6 Hz, 1H), 6.52 (dd, J = 15.6, 7.7 Hz, 1H)
DMSO (3.0 mL) and transferred to an oven-dried Schlenk-tube ppm. 13C NMR (101 MHz, CDCl3): δ 192.9, 144.4, 139.3, 132.0,
and purged with argon. NaH (43.5 mg, 0.96 mmol, 60% in 130.4, 128.5, 127.4 ppm. LRMS (ESI+) [M + H]+, 138.8. HRMS
mineral oil) was transferred directly to the solution and heated (ESI+) calculated [C7H6OS + H]+ 139.0218, found 139.0246. All
(60 °C) under argon for 1.5 h, or until hydrogen evolution had spectroscopic values are consistent with those obtained in the
ceased. 1-Fluoronaphthalene (154 µL, 1.2 mmol) was added literature.29
under argon, and the resulting solution was heated (70 °C) for
1.5 h. On cooling, the solution was partitioned between phenoxyl}propyl}acetamide (13a). Fluoxetine
Synthesis of N-methyl-N-{3-phenyl-3-[4-trifluoromethyl]
(200 mg,
1
toluene and H2O and washed (3× H2O). The organic phase was 0.65 mmol), DCM (4 mL), acetic anhydride (1 mL) and pyri-
separated and dried over anhydrous MgSO4. After filtration the dine (1 mL) were combined and allowed to stir over night. The
organic phase was removed under reduced pressure to yield a resulting solution was diluted in DCM (30 mL) and washed
crude product. Purification by silica gel chromatography (DCM with HCl (3 × 10 mL, w/v 20%) and water (3×). The organic
→ DCM–MeOH–NEt3, 5 : 1 : 1%) gave the pure product as a layer was separated and dried over anhydrous MgSO4. Fil-
yellow oil 2, (214 mg, 83%): 1H NMR (400 MHz, CDCl3): tration followed by the removal of solvent under vacuum
δ 8.38–8.33 (m, 1H), 7.80–7.76 (m, 1H), 7.51–7.46 (m, 2H), 7.39 yielded a crude yellow oil. Purification by silica gel chromato-
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 6121–6127 | 6125

View Article Online
Paper
Organic & Biomolecular Chemistry
graphy (hexane–DCM, 1 : 1 → DCM–MeOH, 9 : 1) gave 13a as a
yellow oil (220 mg, 96%). IR (neat): ν 3052, 2928, 1636, 1578,
1396, 1093, 771 cm−1. 1H NMR (400 MHz, CDCl3): δ 7.42 (d, J =
8.5 Hz, 2H), 7.38–7.27(m, 5H), 6.89 (d, J = 8.4 Hz, 2H), 5.21
(dd, J = 8.6, 4.3 Hz, 1H) 3.63–3.51 (m, 2H), 2.97 (s, 3H),
2.25–2.09 (m, 2H), 2.04 (s, 3H) ppm. 13C NMR (101 MHz,
CDCl3): δ 170.6, 160.3, 140.7, 129.1, 128.3, 126.9, 126.8, 125.7,
125.5, 115.6, 78.4, 47.1, 37.4, 36.6, 21.1 ppm. LRMS (ESI+)
[M + H]+, 351.9. HRMS (ESI+) calculated [C19H2ONO2F3 + H]+
352.1524 found 352.1515. Enantiomeric excess was deter-
mined by HPLC using an AS-H CHIRALCEL column (250 ×
4.6 mm) fitted with guard cartridge (50 × 4.6 mm), 25 °C,
1.0 mL min−1, 210 nm, hexane–IPA (9 : 1). tR (R) = 23.6 min;
tR (S) = 31.9 min.
3 F. S. Messiha, Neurosci. Biobehav. Rev., 1993, 17, 385–
396.
4 C. Hiemke and S. Härtter, Pharmacol. Ther., 2000, 85,
11–28.
5 C. J. Harmer, G. M. Goodwin and P. J. Cowen,
Am. J. Psychiatry, 2004, 161, 1256–1263.
6 M. Srebnik, P. V. Ramachandran and H. C. Brown, J. Org.
Chem., 1988, 53, 2916–2920.
7 Y. Gao and K. B. Sharpless, J. Org. Chem., 1988, 53, 4081–
4084.
8 E. J. Corey and G. A. Reichard, Tetrahedron Lett., 1989, 30,
5207–5210.
9 (a) D. Liu, W. Gao, C. Wang and X. Zhang, Angew. Chem.,
Int. Ed., 2005, 44, 1687–1689; (b) T. Ohkuma, D. Ishii,
H. Takeno and R. Noyori, J. Am. Chem. Soc., 2000, 122,
6519–6511; (c) R. Noyori, T. Ohkuma, M. Kitamura,
H. Takaya, N. Sayo, H. Kumobayashi and S. Akutagawa,
J. Am. Chem. Soc., 1987, 109, 5856–5858; (d) R. Noyori,
Angew. Chem., Int. Ed., 2002, 41, 2008–2022.
Synthesis of N-methyl-N-[3-(naphthalene-1-yloxy)-3-(thio-
phen-2-yl)propyl]acetamide (13b). Duloxetine
2 (166 mg,
0.56 mmol), DCM (3 mL), acetic anhydride (1 mL) and pyri-
dine (1 ml) were combined and allowed to stir over night. The
resulting solution was diluted in DCM (30 mL) and washed
with HCl (3 × 10 mL, w/v 20%) and water (3×). The organic 10 (a) P. Motloch, I. Valterová and M. Kotora, Adv. Synth.
layer was separated and dried over anhydrous MgSO4. Fil-
tration followed by the removal of solvent under vacuum
yielded a crude yellow oil. Purification by silica gel chromato-
graphy (hexane–DCM, 1 : 1 → DCM–MeOH, 9 : 1) gave 13b as a
Catal., 2014, 356, 199–204; (b) Â. de Fátima, A. A. M. Lapis
and R. A. Pilli, J. Braz. Chem. Soc., 2005, 16, 495–499;
(c) M. Iwata, R. Yazaki, N. Kumagai and M. Shibasaki,
Tetrahedron: Asymmetry, 2010, 21, 1688–1694.
yellow oil (150 mg, 79%). IR (neat): ν 2931, 1636, 1516, 1323, 11 D. T. Wong, K. W. Perry and F. P. Bymaster, Nat. Rev. Drug
1245, 1108, 835 cm−1. Observed as a mixture of rotamers,
Discovery, 2005, 4, 764–774.
major rotamer: 1H NMR (400 MHz, CDCl3): δ 8.40–8.30 (m, 12 C. J. Wenthur, M. R. Bennett and C. W. Lindsley, ACS
1H), 7.84–7.81 (m, 1H), 7.56–7.51 (m, 2H), 7.44 (d, J = 8.6 Hz, Chem. Neurosci., 2014, 5, 14–23.
1H), 7.30 (d, J = 8.0 Hz, 1H), 7.21 (dd, J = 5.0, 1.2 Hz, 1H), 7.11 13 Y. Suzuki, M. Iwata, R. Yazaki, N. Kumagai and
(d, J = 3.8 Hz, 1H), 6.97 (dd, J = 5.0, 3.5 Hz, 1H), 6.87 (d, J = 8.5 M. Shibasaki, J. Org. Chem., 2012, 77, 4496–4500.
Hz, 1H), 5.74 (dd, J = 8.0, 4.9 Hz, 1H), 3.82–3.61 (m, 2H), 3.00 14 (a) C. Solé, A. Whiting, H. Gulyás and E. Fernández, Adv.
(s, 3H) 2.57–2.45 (m, 2H), 2.06 (s, 3H) ppm. 13C NMR
(101 MHz, CDCl3): δ 170.7, 153.1, 144.8, 134.6, 127.7, 126.8,
126.5, 126.1, 125.7, 127.5, 124.9, 124.8, 122.0, 121.1, 106.9,
Synth. Catal., 2011, 353, 376–384; (b) C. Solé, A. Tatla,
J. A. Mata, A. Whiting, H. Gulyás and E. Fernández, Chem.
– Eur. J., 2011, 17, 14248–14257.
74.5, 45.1, 36.7, 33.3, 21.9 ppm. LRMS (ESI+) [M + Na]+, 361.3. 15 (a) J. Cid, H. Gulyás, J. J. Carbó and E. Fernández, Chem.
HRMS (ESI+) calculated [C20H21NO2S + H]+ 340.1371, found
340.1377. Enantiomeric excess determined by HPLC using an
Soc. Rev., 2012, 41, 3558–3570; (b) A. D. J. Calow and
A. Whiting, Org. Biomol. Chem., 2012, 10, 5485–5497.
AS-H CHIRALCEL column (250 × 4.6 mm) fitted with guard 16 H.-H. Blaser, Chem. Rev., 1992, 92, 935–952.
cartridge (50 × 4.6 mm), 25 °C, 1.0 mL min−1, 210 nm, 17 A. D. J. Calow, A. Batsanov, A. Pujol, C. Solé, E. Fernández
hexane–IPA (85 : 15). tR (S) = 29.2 min; tR (R) = 38.2 min.
and A. Whiting, Org. Lett., 2013, 15, 4810–4813.
18 H. Chea, H.-S. Sim and J. Yun, Adv. Synth. Catal., 2009, 351,
855–858.
19 (a) S. Ram and L. D. Spicer, Synth. Commun., 1987, 17,
415–418; (b) P. Bucks, Russ. Chem. Rev., 1983, 52, 2072–
2090.
20 (a) S. Rajagopal and A. F. Spatola, Appl. Catal., A, 1997, 152,
69–81; (b) R. J. Rahaim and R. E. Maleczka, Org. Lett., 2011,
13, 584–587.
Acknowledgements
A.D.J.C. would like to thank the EPSRC for a doctoral training
account.
21 M. Ciaccia, R. Cacciapaglia, P. Mencarelli, L. Mandolini
and S. Di Stefano, Chem. Sci., 2013, 4, 2253–2261.
Notes and references
1 (a) D. T. Wong, J. S. Horng, F. P. Bymaster, K. L. Hauser 22 G. K. Cantrell and T. Y. Meyer, Organometallics, 1997, 16,
and B. B. Molloy, Life Sci., 1974, 15, 471–479; 5381–5383.
(b) D. T. Wong, F. P. Bymaster and E. A. Engleman, Life Sci., 23 This β-boryl aldehyde eliminates to give cinnamaldehyde
1995, 57, 411–441.
2 L. Sghendo and J. Mifsud, J. Pharm. Pharmacol., 2012, 64,
317–325.
during column chromatography, as reported in the litera-
ture: I. Ibrahem, P. Breistein and A. Córdova, Angew. Chem.,
Int. Ed., 2011, 50, 12036–12041.
6126 | Org. Biomol. Chem., 2014, 12, 6121–6127
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
24 A. D. J. Calow, A. S. Batsanov, E. Fernández, C. Solé and 27 G. Wang, X. Liu and G. Zhao, Tetrahedron: Asymmetry, 2005,
A. Whiting, Chem. Commun., 2012, 48, 11401–11403. 16, 1873–1879.
25 H. Kakei, T. Nemoto, T. Ohshima and M. Shibasaki, Angew. 28 J. Deeter, J. Frazier, G. Staten, M. Staszak and L. Weigel,
Chem., Int. Ed., 2004, 116, 321–324. Tetrahedron Lett., 1990, 31, 7101–7104.
26 K. Omura and D. Swern, Tetrahedron, 1978, 34, 1651– 29 E. Kim, M. Koh, J. Ryu and S. B. Park, J. Am. Chem. Soc.,
1660.
2008, 130, 12206–12207.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem., 2014, 12, 6121–6127 | 6127