Continuous-Flow Synthesis of (-)-Oseltamivir Phosphate (Tamiflu)

Article abstract of DOI:10.1055/s-0039-1690878

Herein the anti-influenza drug (-)-oseltamivir phosphate is prepared in continuous flow from ethyl shikimate with 54% overall yield over nine steps and total residence time of 3.5 min from the individual steps. Although the procedure involved intermediate

Full text of DOI:10.1055/s-0039-1690878

SYNLETT
0
9
3
6
-
5
2
1
4
1
4
3
7
-
2
0
9
6
© Georg Thieme Verlag Stuttgart · New York
2020, 31, A–E
cluster
A
en
C. R. Sagandira, P. Watts
Cluster
Synlett
Continuous-Flow Synthesis of (–)-Oseltamivir Phosphate (Tamiflu)
Cloudius R. Sagandira
Paul Watts*
0
0
0
0-0
0
0
2-4
1
5
6-
5
6
7
9
HO
CO2Et
O
CO2Et
HO
Department of Chemistry, Nelson Mandela University,
University Way, Port Elizabeth, 6031, South Africa
Paul.Watts@mandela.ac.za
AcHN
OH
ethyl shikimate
NH2⋅H3PO4
9-step semitelescoped
continuous-flow synthesis
via azide chemistry
tr = 3.5 min
(–)-oseltamivir phosphate
54% overall yield
Published as part of the Cluster Integrated Synthesis Using
Continuous-Flow Technologies
Received: 27.02.2020
Accepted after revision: 13.03.2020
Published online: 24.04.2020
opment of azide-chemistry-free routes. Unfortunately,
these alternative azide-free routes are generally inefficient
compared to the azide-chemistry approaches. More than 60
synthetic routes have been developed towards Tamiflu to
date.1,7–10 However, most of these synthetic approaches suf-
fer from the use of potentially hazardous azide chemistry,
thus raising safety concerns and eventually ruled out for
large-scale synthesis in batch systems.1,2 Therefore, the de-
velopment of alternative practical and safe processes for
(–)-oseltamivir phosphate (1) synthesis, which can be
adapted at large scale, is imperative.
DOI: 10.1055/s-0039-1690878; Art ID: st-2020-u0112-c
Abstract Herein the anti-influenza drug (–)-oseltamivir phosphate is
prepared in continuous flow from ethyl shikimate with 54% overall yield
over nine steps and total residence time of 3.5 min from the individual
steps. Although the procedure involved intermediate isolation, the dan-
gerous azide chemistry and intermediates involved were elegantly han-
dled in situ. It is the first continuous-flow process for (–)-oseltamivir
phosphate involving azide chemistry and (–)-shikimic acid as precursor.
Key words azide chemistry, continuous flow, influenza, (–)-oseltami-
Continuous-flow synthesis has emerged as a useful
technique in synthetic chemistry, largely motivated by its
numerous advantages relative to batch; these include im-
proved synthetic efficiency, safety, and selectivity.11–19 The
only literature on continuous-flow total synthesis of (–)-os-
eltamivir phosphate (1) is a five-step procedure by Hayashi
and Ogasawara,10 which started from Michael addition and
avoided azide chemistry. Although their approach showed
ingenuity in synthesizing a compound with three chiral
centers through a multistep continuous-flow procedure
starting from the Michael reaction in a single passage, the
throughput of 58 mg per 15 h (total yield of 13%) was insuf-
ficient to meet demand.
Herein we report safe and efficient continuous-flow
conversion of (–)-shikimic acid (2) into (–)-oseltamivir
phosphate (1) via the hazardous azide chemistry. Our flow
synthetic route (Scheme 1) was inspired by the batch syn-
thesis by Shi,9 Kalashnikov et al.20 and Karpf et al.21 It ele-
gantly takes advantage of the chirality present in (–)-shi-
kimic acid (2) to introduce the groups at C3, C4, and C5
with the desired stereochemistry.2
vir phosphate, total synthesis
Influenza is a severe viral infection of the respiratory
system, which is responsible for a significant morbidity and
mortality due to both annual epidemics and predictable
pandemics.1 (–)-Oseltamivir phosphate (Tamiflu) is used for
treatment and prophylaxis of influenza.1–5 According to the
World Health Organization (WHO), this drug is one of the
most important anti-influenza drugs to guard against a
pandemic.6 This has prompted the research community to
focus on the development of new, better, and practical ap-
proaches to this drug. In the early years of discovery, (–)-
shikimic acid (2) was used as starting material and further-
more, the current industrial procedure uses (–)-shikimic
acid (2). There have been several new approaches to (–)-os-
eltamivir phosphate (1) from sources other than (–)-shi-
kimic acid (2) based on arguments regarding the availabili-
ty of (–)-shikimic acid (2) as well-potential risks of azide
chemistry on an industrial scale.1,7–9 The legitimate (–)-shi-
kimic acid (2) availability concerns in the early years of the
development of this drug have long been solved.9–11 Howev-
er, the concerns associated with the use of potentially haz-
ardous azide chemistry are yet to be addressed. There have
been enormous efforts in this regard mainly through devel-
We started from ethyl shikimate (3) already synthesized
in flow from shikimic acid (2).22 Trimesylation of ethyl shi-
kimate (3) using MsCl in the presence of triethylamine
© 2020. Thieme. All rights reserved. Synlett 2020, 31, A–E
Georg Thieme Verlag KG, Rüdigerstraße 14, 70469 Stuttgart, Germany

B
C. R. Sagandira, P. Watts
Cluster
Synlett
N3
CO2Et
HO
HO
CO2H
HO
HO
CO2Et
MsO
MsO
CO2Et
esterification
Ref 22
mesylation
azidation
MsO
OMs
OH
OH
ethyl shikimate (3)
OMs
5
(–)-shikimic acid (2)
4
2
3
2
5
2
5
O
CO2Et
aziridine
ring-opening
O
CO2Et
O
CO2Et
1
6
1
1
6
O
P
aziridination
N–P bond cleavage
3
3
EtO
P
N
EtO
4
7
5
OEt7
6
4
4
H2N
N
EtO
H
OMs
OMs
OMs
7a
7
6
O
AcHN
CO2Et
2
5
O
CO2Et
O
CO2Et
azide reduction
acetylation
azidation
3
1
6
4
AcHN
AcHN
NH2
10
N3
OMs
9
8
salt formation
O
CO2Et
AcHN
NH2⋅H3PO4
(–)-oseltamivir phosphate (1)
Scheme 1 Reaction sequence for the synthesis of (–)-oseltamivir phosphate (1) from (–)-shikimic acid (2)
(TEA) to afford O-trimesylate 4 was initially threatened by
reactor blockage due to the insoluble byproduct triethylam-
monium chloride in EtOAc.
Sonication was introduced to resolve this, and O-
mesylate 4 was rapidly afforded in excellent yield. Sonica-
tion was later avoided by using tributylamine (TBA) whose
ammonium salt is soluble in EtOAc instead of TEA and opti-
mization was performed (Table 1).
A slight excess of MsCl (1.5 equiv.) was necessary and
conversion improved with increase in the amount of TBA
(Table 1, entries 1, 2, and 5). Increase in residence time im-
proved conversion (Table 1, entries 2 and 4) whilst the ef-
fect of temperature increase was insignificant (Table 1, en-
tries 2 and 3); 92% isolated yield of O-mesylate was afforded
at 12 s residence time and room temperature using MsCl
(1.5 equiv.) and TBA (4 equiv.). DBU can be used to achieve
comparable results.
O-Trimesylate 4 in the presence of an appropriate azi-
dating agent undergoes highly regio- and stereoselective
nucleophilic substitution of allylic O-mesylate at the C-3
position affording azide 5. NaN3 (aq) is the commonly used
azidating agent, and azide 5a is a known minor side product
(Figure 1).9,20,21 Comprehensive optimization studies on this
reaction have been reported in a preliminary publication.23
Mesyl shikimate (4) azidation performed at 0 °C and 30
s residence time using NaN3 (aq) afforded 86% conversion
and 100% azide 5 selectivity (Table 2, entry 1). An increase
in temperature and residence time improved conversion
but promoted side product 5a formation, thus lowering
azide 5 selectivity (Table 2, entries 1–4). Using NaN3 (aq),
the best results were found at 25 °C and 30 s residence time
affording azide 5 in 91% isolated yield. Other azidating
agents such as Diphenylphosphoryl azide (DPPA), Trimeth-
ylsilyl azide (TMSA), and Tetrabutylammonium azide
Table 1 Ethyl Shikimate Mesylation Optimization
Entry Residence MsCl
TBA
Temp
(°C)
Conv. (%)c
time (s)
(equiv.)a
(equiv.)b
1
2
3
4
5
6
7
1
1
1
2
2
2
2
4
3
1
r.t
r.t
40
r.t
r.t
r.t
r.t
54
1.5
1.5
1.5
1.5
1.5
1.5
68
1
69
12
12
12
12
71
97 (89)d
84
52
a MsCl equiv. per OH on ethyl shikimate (3).
b TBA equiv. per OH on ethyl shikimate (3).
c Conversion determined by HPLC.
d Number in parentheses corresponds to isolated yield. All reactions were
quenched by 1 M HCl.
© 2020. Thieme. All rights reserved. Synlett 2020, 31, A–E

C
C. R. Sagandira, P. Watts
Cluster
Synlett
Table 3 Aziridine 6 Formation Optimizationa
O
N3
MsO
OEt
Entry
Residence time (s)
Temp (°C)
Conv. (%)b
NaN3 (aq)
OMs
1
2
3
4
12
12
12
60
100
150
190
150
7
O
10 Bar
5
MsO
MsO
58
OEt
+
O
100 (94)c
92
Thermally Controlled Reactor
N3
OEt
OMs
in acetonitrile
a Standard conditions: azide 5 (0.1 M, 1 equiv.), (EtO)3P (1.1 equiv.), aceto-
nitrile as solvent.
4
5a
b Conversion determined by HPLC.
Figure 1 Optimization of O-trimesylate 4 azidation in flow
c Number in parentheses corresponds to isolated yield.
(TBAA) can be used. Azide 5 isolated yield of 78% was
achieved using DPPA in acetonitrile at 25 °C and 30 s resi-
dence time.
O
N3
OEt
MsO
OMs
5
Table 2 Optimization of O-Trimesylate 4 Azidationa
25 °C and 30 s
DPPA (1.1 equiv.)
(EtO)3P
(0.055 M, 1.1 equiv)
in acetonitrile
Entry Azidating
agent
Temp (°C) Residence Conv. (%)c Azide 5
time (s)
selectivity (%)d
O
1
2
3
4
5
6
7
NaN3 (aq)
NaN3 (aq)
NaN3 (aq)
NaN3 (aq)
DPPAb
0
25
30
30
30
60
30
30
30
86
100
100
100
93
100
MsO
OEt
100 (91)
69
4
MsO
O
100
25
OMs
O
P
OEt
73
(0.1 M, 1 equiv.)
(TEA, 1.1 equiv.)
in acetonitrile
RO
N
10 bar
OR
25
89 (78)
90
OMs
TMSAb
25
100
100
190 °C and 10 s
6
TBAA
25
66
(74% isolated yield)
a Standard conditions: O-trimesylate 4 (0.1 M, 1 equiv.), azidating agent
(1.1 equiv.), acetonitrile as solvent.
Figure 2 Two-step reaction sequence to aziridine 6
b TBA ( 1.1 equiv.), quenched by 1 M HCl.
c Conversion determined by HPLC.
d Number in parentheses corresponds to isolated yield.
Pure aziridine 6 (1 equiv.) subsequently undergoes regio-
and stereoselective ring opening with 3-pentanol in the
presence of BF3·OEt2 (1.5 equiv.) at the allylic position to in-
troduce the N-based substituent on C-4 of the cyclic carbon
backbone. High temperatures and longer residence times
favored 3-pentyl ether 7 synthesis (Figure 3). Full conver-
sion (95% isolated yield) was achieved at 100 °C and 12 s
(Figure 3).
Azide 5 undergoes aziridination when treated with
(EtO)3P (1.1 equiv.) under water-free conditions to afford
aziridine 6. When the reactor was heated at 100 °C and 12 s
residence time, only a 7% conversion was obtained (Table 3,
entry 1). Temperature and residence time increase im-
proved the results, at 190 °C affording the highest conver-
sion and yield (Table 3, entry 3). The use of (MeO)3P also
produced comparable results.
With the azidation and aziridination reactions separate-
ly optimized, multistep synthesis of aziridine 6 from syn-
thetic mesyl shikimate (4) via azide 5 formation and con-
sumption in situ was performed to enhance process safety
(Figure 2). Although NaN3 (aq) was the most efficient azi-
dating agent, more affordable and its use is accompanied by
better atom efficiency compared to DPPA; we used DPPA for
telescoping because water (NaN3 (aq)) was detrimental in
the succeeding aziridination step. The total residence time
for this multistep synthesis was 25 s affording aziridine 6 in
84% yield.
Due to the safety concerns of handling aziridines,24–27
we further demonstrated in situ generation and consump-
tion of aziridine 6 to afford 3-pentyl ether 7 in 87% yield in
20 s residence time (Figure 4).
Acetylation of 3-pentyl ether 7 was achieved via a tan-
dem of reactions; N–P bond cleavage of 3-pentyl ether 7 by
H2SO4 and subsequently treated with Ac2O to afford acet-
amide 8 in flow. We began with N–P bond-cleavage optimi-
zation first (Table 4). Higher temperatures gave better re-
sults, the best being 170 °C and 3 s residence time using
H2SO4 (5 equiv., Table 4, entry 3). In the two-step process,
the reaction was neutralized in-line with NaOH (aq), and
the crude was treated with Ac2O (1.6 equiv.) in acetonitrile
at room temperature and 30 s residence time affording
acetamide 8 in 93% isolated yield.
© 2020. Thieme. All rights reserved. Synlett 2020, 31, A–E

D
C. R. Sagandira, P. Watts
Cluster
Synlett
Table 4 Optimisation of N–P Bond Cleavage of 3-pentyl ether 7a
With the pure azide 9 in hand, the last intermediate os-
eltamivir (10) was obtained via azide group reduction. Vari-
ous azide 9 reduction procedures such as P(n-Bu)3, Ph3P,
and H2 in the presence of Pd/C, Lindler catalyst, and Raney-
Ni to obtain oseltamivir (10) have been reported.8,20,21 In-
spired by the work of Fringuelli et al.28, we employed the
CoCl2-catalyzed chemoselective reduction of azide 9 using
NaBH4 in water (pH = 8) at ambient temperature to afford a
novel procedure towards oseltamivir (10). The formation of
the black cobalt boride precipitate which threatened our
flow procedure was resolved by introduction of sonication
to avoid blockages. Azide 9 reduction was performed and
optimized at room temperature using NaBH4 (2 equiv.) and
CoCl2 (0.1 equiv.) affording 96% conversion (93% yield) at 5 s
residence time (Figure 5)
Entry
Residence time (s)
Temp (°C)
Conv. (%)b
1
2
3
3
3
3
80
120
170
53
71
100
a Standard conditions: 3-pentyl ether 7 (0.1 M, 1 equiv.), H2SO4 (5 equiv.),
(1.1 equiv.), acetonitrile as solvent.
b Conversion determined by HPLC.
Acetamide 8 azidation was accomplished using various
azidating agents such as NaN3, DPPA, TMSA, and TBAA to af-
ford azide 9. NaN3 (aq, 3 equiv.) at 190 °C and 45 s residence
time gave the best yield (89%) as reported in a preliminary
publication.23
Figure 5 Optimization of oseltamivir (10) synthesis
Use of either ethanol or methanol as NaBH4 solvent gave
comparable results (see Supporting Information, Table 1).
In the final step, oseltamivir (10) was treated with H3PO4
(1.2 equiv.) to immediately afford (–)-oseltamivir phos-
phate (1) in 97% yield at 50 °C and 60 s residence time.
In summary, we successfully developed an efficient
nine-step total synthesis procedure in continuous flow for
(–)-oseltamivir phosphate (1) in remarkable 3.5 min total
residence time, 54% untelescoped and 48% semitelescoped
overall yield. This process is significantly shorter and high-
er-yielding than most of the reported procedures.1,2 Al-
though our procedure is not multistep like the only report-
ed (–)-oseltamivir phosphate (1) continuous-flow process
by Hayashi and Ogasawara (310 min residence time and 13
% total yield),10 we had superior total yield and residence
time. To the best our knowledge, this is the first (–)-shikim-
ic acid based synthetic procedure for (–)-oseltamivir phos-
phate (1) which takes advantage of continuous-flow tech-
nology to elegantly handle the hazardous azide chemistry.
Consequently, we believe that application of flow is the last
piece of the puzzle in solving the long-standing azide chem-
istry concerns in the large-scale synthesis of (–)-oseltamivir
Figure 3 Optimization of 3-pentyl ether 7 synthesis
O
O
OEt
EtO
P N
OEt
OMs
6
190 °C, 12 s
BF3⋅OEt2
(EtO)3P
in
in
3-pentanol/acetonitrile
(80:20)
acetonitrile
O
N3
OEt
MsO
OMs
O
5
10 bar
O
in acetonitrile
OEt
O
P
EtO
N
EtO
100 °C, 8 s
H
OMs
7
Figure 4 Two-step reaction sequence to 3-pentyl ether 7
© 2020. Thieme. All rights reserved. Synlett 2020, 31, A–E

E
C. R. Sagandira, P. Watts
Cluster
Synlett
phosphate in industry and academia. The route is scalable
and has great potential to become an alternative to the cur-
rent commercial route due to its operational simplicity, in-
expensive reagents, and lack of protecting group chemistry.
(8) Nie, L.; Ding, W.; Shi, X.; Quan, N.; Lu, X. Tetrahedron: Asymme-
try 2012, 23, 742.
(9) Nie, L.; Shi, X.; Ko, K. H.; Lu, W. J. Org. Chem. 2009, 74, 3970.
(10) Ogasawara, S.; Hayashi, Y. Synthesis 2017, 49, 424.
(11) Sagandira, C. R.; Watts, P. Eur. J. Org. Chem. 2017, 44, 6554.
(12) Brandt, J. C.; Wirth, T. Beilstein J. Org. Chem. 2009, 5, 1.
(13) Delville, M. E.; Nieuwland, P. J.; Janssen, P.; Koch, K.; van Hest, J.
C. M.; Rutjes, F. P. J. T. Chem. Eng. J. 2011, 167, 556.
Funding Information
(14) Baumann, M.; Baxendale, I. R.; Ley, S. V.; Nikbin, N.; Smith, C. D.
Org. Biomol. Chem. 2008, 6, 1587.
(15) Stevens, C. V. Chem. Soc. Rev. 2016, 45, 4892.
(16) Gutmann, B.; Roduit, J. P.; Kappe, C. O.; Roberge, D. Angew.
Chem. Int. Ed. 2010, 49, 7101.
We thank the National Research Foundation (NRF SARChI Grant),
Council for Scientific and Industrial Research (CSIR-DST Grant) and
Nelson Mandela University for financial support.
N
ati
o
n
a
lResearc
h
F
o
u
n
d
ati
o
n
()N
e
l
so
n
M
a
n
d
e
l
a
U
n
i
v
ersity()
(17) Ötvös, S. B.; Hatoss, G.; Georgiádes, A.; Kovács, S.; Mándity, I.;
Nováket, Z. RSC Adv. 2014, 4, 46666.
Supporting Information
(18) Smith, C. J.; Smith, C. D.; Nikbin, N.; Ley, S. V.; Baxendale, I. R.
Org. Biomol. Chem. 2011, 9, 1927.
Supporting information for this article is available online at
https://doi.org/10.1055/s-0039-1690878.
S
u
p
p
orit
n
gInformati
o
n
S
u
p
p
orti
n
gInformati
o
n
(19) Sprecher, H.; Payán, M. N. P.; Weber, M.; Yilmaz, G.; Wille, G.
J. Flow Chem. 2011, 2, 20.
(20) Kalashnikov, I. A.; Sysolyatin, S. V.; Sakovich, G. V.; Sonina, E. G.
Russ. Chem. Bull. Int. Ed. 2013, 62, 163.
(21) Karpf, M.; Trussardi, R. Angew. Chem. Int. Ed. 2009, 48, 5760.
(22) Sagandira, C. R.; Watts, P. J. Flow Chem. 2019, 9, 79.
(23) Sagandira, C. R.; Watts, P. Beilstein J. Org. Chem. 2019, 15, 2577.
(24) Hsueh, N.; Clarkson, G. J.; Shipman, M. Org. Lett. 2015, 17, 3632.
(25) Hsueh, N.; Clarkson, G. J.; Shipman, M. Org. Lett. 2016, 18, 4908.
(26) Penkova, M.; Nikolova, S. Bulg. Chem. Commun. 2017, 44, 105.
(27) Sweeney, J. B. Chem. Soc. Rev. 2002, 31, 247.
References
(1) Magano, J. Chem. Rev. 2009, 109, 4398.
(2) Magano, J. Tetrahedron 2011, 67, 7875.
(3) Abrecht, S.; Harrington, P.; Iding, H.; Karpf, M.; Wirz, B.; Zutter,
U. Chim. Int. J. Chem. 2004, 58, 621.
(4) Moscona, A. N. Engl. J. Med. 2005, 353, 1363.
(5) Ishikawa, H.; Suzuki, T.; Orita, H.; Uchimaru, T. Chem. Eur. J.
2010, 16, 12616.
(6) WHO Report Standard Guidelines for the Clinical Management
of Severe Influenza Virus Infections, 2017.
(28) Fringuelli, F.; Pizzo, F.; Vaccaro, L. Synthesis 2000, 646.
(7) Nie, L.; Shi, X. Tetrahedron: Asymmetry 2009, 20, 124.
© 2020. Thieme. All rights reserved. Synlett 2020, 31, A–E