Sugar-based synthesis of tamiflu and its inhibitory effects on cell secretion

Article abstract of DOI:10.1002/chem.200902048

Tamiflu is currently the most effective drug for the treatment of influenza, but the insufficient supply and side-effects of this drug demand urgent solutions. We present a practical synthesis of Tamiflu by using novel synthetic routes, cheap reagents, and the abundantly available starting material D-glucal. The strategy features a Claisen rearrangement of hexose to obtain the cyclohexene backbone and introduction of diamino groups through tandem intramolecular aziridination and ring opening. In addition, this synthetic protocol allows late-stage functionalization for the flexible synthesis of Tamiflu analogues. By using the synthesized Tamiflu and its active metabolite (oseltamivir carboxylate), we inves-tigated their influences on neuroendocrine PC12 cells in various aspects. It was discovered that oseltamivir carboxylate significantly inhibits the vesicular exocytosis (regulated secretion) of PC 12 cells, and suggests a mechanism underlying the Tamiflu side-effects, in particular its possible adverse influences on neurotransmitter release in the central nervous system.

Full text of DOI:10.1002/chem.200902048

FULL PAPER
DOI: 10.1002/chem.200902048
Sugar-Based Synthesis of Tamiflu and Its Inhibitory Effects on Cell Secretion
Jimei Ma,[a] Yanying Zhao,[b] Simon Ng,[a] Jing Zhang,[b] Jing Zeng,[a] Aung Than,[b]
Peng Chen,*[b] and Xue-Wei Liu*[a]
Abstract: Tamiflu is currently the most
effective drug for the treatment of in-
fluenza, but the insufficient supply and
side-effects of this drug demand urgent
solutions. We present a practical syn-
thesis of Tamiflu by using novel syn-
thetic routes, cheap reagents, and the
abundantly available starting material
d-glucal. The strategy features a Clai-
tandem intramolecular aziridination
and ring opening. In addition, this syn-
thetic protocol allows late-stage func-
tionalization for the flexible synthesis
of Tamiflu analogues. By using the syn-
thesized Tamiflu and its active metabo-
lite (oseltamivir carboxylate), we inves-
tigated their influences on neuroendo-
crine PC12 cells in various aspects. It
was discovered that oseltamivir carbox-
ACHUTNGRENyNUG lACTHNGUTRENNUaG te significantly inhibits the vesicular
exocytosis (regulated secretion) of
PC12 cells, and suggests a mechanism
underlying the Tamiflu side-effects, in
particular its possible adverse influen-
ces on neurotransmitter release in the
central nervous system.
Keywords: carbohydrates · cell se-
cretion · inhibitors · tamiflu · total
synthesis
ACHTUNGTRENNUNGsen rearrangement of hexose to obtain
the cyclohexene backbone and intro-
duction of diamino groups through
Introduction
effective weapons to combat various strains of influenza by
preventing virus replication. Both of these drugs mimic
sialic acid and competitively inhibit viral neuraminidase, the
enzyme that removes terminal sialic acid from glycoconju-
gates on the host cell surface to allow release and spread of
the influenza virus (Figure 1).[1] Compared with Relenza,
which relies on an inconvenient inhalation administration,
Tamiflu (prodrug) has high oral bioavailability and can be
readily hydrolyzed by hepatic esterase to give the free car-
boxylate of oseltamivir (GS4071, OC, 2) as the active com-
petitive inhibitor to neuraminidase. Tamiflu is currently the
most effective drug for the treatment of influenza, including
H1N1.
Influenza, a common and highly infectious disease, kills hun-
dreds of thousands of people annually and millions during
its outbreaks. At the time of writing, the recent outbreak of
influenza A-H1N1 (swine flu) is rapidly reaching global pan-
demic stage (over 100 countries) and creating great panic
worldwide, similar to that induced by SARS and bird flu a
few years ago.
In the limited drug arsenal against influenza, neuramini-
dase (or sialidase) inhibitors, specifically Tamiflu (oseltami-
vir phosphate, 1·H3PO4) and Relenza (Zanamivir), excel as
The current commercial synthesis involves a 10-step pro-
[a] J. Ma, S. Ng, J. Zeng, Prof. Dr. X.-W. Liu
Division of Chemistry and Biological Chemistry
School of Physical and Mathematical Sciences
Nanyang Technological University
21 Nanyang Link, Singapore 637371
Fax : (+65)67911961
cess of complex chemical reactions and relies on (À)-shi
ACHTUNGTRENNUNGkim-
ic acid,[2] and was thought to be the most proficient till now.
However, Tamiflu is undoubtedly a popular molecule of
high importance, and intensive efforts have been made to
develop alternative routes to Tamiflu with various strategies
for cyclohexenyl ring formation and to start from readily
available and less-expensive starting materials.[3] For exam-
ple, asymmetric Diels–Alder chemistry[4] and the Michael
reaction[5] have been exploited to generate the cyclohexene
moiety. Commercially available starting materials, such as
(À)-quinic acid,[6] l-serine,[7] xylose,[8] mesoaziridine,[9] sub-
stituted cyclohexanediene,[10] lactone,[11] pyridine,[12] 2,6-di-
methoxyphenol,[13] d-mannitol,[14] l-methionine,[15] and ethyl
E-mail: xuewei@ntu.edu.sg
[b] Y. Zhao, J. Zhang, A. Than, Prof. Dr. P. Chen
Division of Bioengineering
School of Chemical and Biomedical Engineering
Nanyang Technological University
70 Nanyang Avenue, Singapore 637457
Fax : (+65)67911761
E-mail: chenpeng@ntu.edu.sg
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200902048.
Chem. Eur. J. 2010, 16, 4533 – 4540
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4533

Herein we report an economical synthetic route to Tami-
flu that uses cheap and abundant d-glucal as the starting
material. Furthermore, we investigated the effects of Tami-
flu and its active form (OC) on the morphology, differentia-
tion, cytoskeleton organization, and vesicular exocytosis
(regulated secretion) of neuroendocrine PC12 cells. It was
discovered that OC significantly inhibits the vesicular exocy-
tosis (regulated secretion) of PC12 cells. Given that exocyto-
sis is a fundamental and ubiquitous function found in many
cell types and is responsible for various critical biological
processes, such as hormone secretion from endocrine cells
and neurotransmitter release from neurons, our results pos-
tulate a mechanism for the side-effects of Tamiflu, in partic-
ular its possible adverse influences on neurotransmitter re-
lease in the central nervous system.
Results and Discussion
Sugar-based synthesis of Tamiflu
Retrosynthetic analysis: Carbohydrates are structurally di-
verse and contain a wealth of stereochemical properties.
They have been serving as chiral pools for the total synthesis
of numerous bioactive natural products for decades. Herein
d-glucal, which is commercially available, was chosen as the
starting material for the Tamiflu synthesis to take advantage
of its naturally occurring stereocenters. As shown in
Scheme 1, compound 4 in our retrosynthetic analysis was
Figure 1. Top: Chemical structures of sialic acid and the two most popular
neuraminidase inhibitors (Zanamivir and oseltamivir). Bottom: The
working mechanism of neuraminidase inhibitors. The neuraminidase of
the influenza virus cleaves the binding of hemagglutinin with the cellular
receptor at the sialic acid residue to release the virus for further replica-
tion.
Scheme 1. Retrosynthetic analysis of (À)-oseltamivir (1).
benzoate,[16] have been used as alternative starting materials
for robust syntheses of Tamiflu.
identified as the pivotal intermediate that could be used to
synthesize 3, the prototype of the target molecule, through
tandem intramolecular aziridination, ring-opening by N-nu-
cleophiles, and olefin formation (Scheme 1). We envisioned
that the configurations at C4 and C5 of compound 3 could
be set by taking advantage of the regioselectivity and stereo-
selectivity in the formation and ring-opening of the aziri-
dine. Carbamate 4 could be derived from aldehyde 5
through known transformations. Aldehyde 5 could be gener-
ated from functionalized glucal 6 by employing a 3,3-sigma-
tropic rearrangement as a critical step to form the carbocy-
cle with the desired chiral configurations at C2 and C3.
Glucal 7 is commercially available or can be readily synthe-
sized from d-glucose.
The efficacy of oseltamivir as a powerful anti-influenza
drug notwithstanding, adverse side-effects have been report-
ed, particularly the mysterious psychiatric consequences that
have dismayingly led to 54 adolescent suicides in Japan as of
2007.[17] The molecular mechanisms underlying the Tamiflu
side-effects are yet to be unveiled, but the side-effects of
TamACHTUNGTRENNUNGiflu are not totally unexpected given that the Tamiflu in-
hibition of neuraminidase inherently lacks isoform or spe-
cies selectivity and that neuraminidases are implicated in
many essential biological processes, such as signal transduc-
tion,[18] cell proliferation and differentiation,[19] cell-to-cell
interaction,[20] immune responses,[21] and neuronal func-
tions.[22]
4534
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 4533 – 4540

Tamiflu Synthesis and Biology
FULL PAPER
Construction of cyclohexene core 10 (a precursor to 4):
Scheme 2 shows the construction of a six-membered carbon
backbone starting from commercially available d-glucal. In-
out by using synthesized carbocycle 10. Our strategy was to
take advantage of the hydroxyl group on C4 position, the
configuration of which is originally from the C4 of the sugar
ring. We first followed the strategy of intramolecular nitro-
gen delivery for the unnatural aminosugar synthesis, which
was successfully developed in our laboratory.[24] Specifically,
a sulfonamide ester was tethered onto the 4-hydroxy group
and generated rhodium-stabilized nitrene by using Rh2-
AHCTUNGRTEG(NNUN OAc)4/PhI=O. The active nitrene species was intramolecu-
larly delivered into a carbon–carbon double bond, followed
by an aziridine ring-opening reaction by using nitrogenous
nucleophiles. Unfortunately, this endeavor did not lead to
formation of the desired product, probably due to insur-
mountable hindrance. Inspired by the work of Padwa
et al.,[25] we attempted to form carbamate 4a by treating al-
cohol 10 with Cl3CCONCO/K2CO3 (yield 87%). Disap-
pointingly, intramolecular aziridination could not take place
despite the extensive screening of catalysts, such as [Cu-
Scheme 2. Synthesis of compound 10. Reagents and conditions: a) p-anis-
ACHTUNGTRENNUNGaldehyde diethyl acetal (1.5 equiv), PPTS (0.1 equiv), DMF, 258C, 2 h;
b) TBSCl (1.2 equiv), imidazole (2.4 equiv), DMAP (0.1 equiv), DMF,
258C, 3 h; c) DIBAL-H (1.2 equiv), dichloromethane, À15–08C, 2 h,
65% (3 steps); d) DMP (1.2 equiv), dichloromethane, 258C, 2 h followed
by methyltriphenylphosphonium bromide (1.8 equiv), nBuLi (1.5 equiv),
THF, À78–258C, 1 h, 67% (2 steps); e) diphenyl ether, 2108C, 2 h, 88%;
f) NaClO2 (3 equiv), NaH2PO4 (3 equiv), 2-methyl-2-butene (5 equiv),
tBuOH/H2O, 25 8C, 2 h followed by EtI (2 equiv), K2CO3 (1.5 equiv),
DMF, 258C, 3 h, 87% (2 steps); g) DDQ (1.5 equiv), dichloromethane/
H2O, 92%. Abbreviations: PPTS=pyridinium p-toluenesulfonate;
DMF=N,N-dimethylformamide; TBSCl=tert-butyldimethylsilyl chlo-
ride; DMAP=4-dimethylaminopyridine; DIBAL-H=diisobutylalumini-
um hydride; DMP=Dess–Martin periodinate; DDQ=2,3-dichloro-5,6-
dicyano-p-benzoquinone.
A
2ACHTUNGTREN(UNNG OAc)4, and [Rh2-
AHCTUNGTRENNUNG
stituted carbamate 4 was prepared in 77% yield by treating
alcohol 10 with CDI/NH2OH, followed by TsCl/Et3N
(Scheme 3).[26] The Rh
2ACTHNUTRGNE(NUG OAc)4-catalyzed aziridination in di-
stallation of 4,6-benzylidene acetal and silylation of 3-hy-
droxyl gave fully protected d-glucal, which underwent selec-
tive opening of the benzylidene acetal with DIBAL-H in di-
chloromethane at À158C to give free primary alcohol 8
(65% yield starting from d-glucal). The primary hydroxyl
group in 8 was oxidized to the aldehyde by using Dess–
Martin periodinate or Swern reagents and then subjected to
Wittig methylenation to give terminal olefin 6 in 67% yield.
The next step was the critical Claisen rearrangement reac-
tion, which allowed ready access to a carbocycle from the
sugar ring. This reaction was conducted in a sealed reaction
vessel at 2108C in diphenyl ether, and aldehyde 5 was dia-
stereoselectively achieved in an excellent yield of 88%. As
discussed by Bꢁchi and Powell,[23] this rearrangement is con-
trolled by a facial preference via a boat-like transition state
that results in the formation of single isomer 5. The oxida-
tion of 5 to ethyl ester 9 was found to be best carried out by
using NaClO2/NaH2PO4 in the presence of 2-methyl-2-
butene, followed by esterification (87% yield). Use of
Oxone or I2/KOH in ethanol gave poor yields of approxi-
mately 30%. The p-methoxybenzyl (PMB) group was suc-
cessively removed with 2,3-dichloro-5,6-dicyanobenzoqui-
none (DDQ) to give alcohol 10 in 92% yield.
Scheme 3. Synthesis of compound 11. Reagents and conditions for 4a:
a) Cl3CCONCO (2 equiv), K2CO3 (5 equiv), dichloromethane/MeOH,
87%; for 4: b) CDI (1.5 equiv), dichloromethane, 258C, 2 h; hydroxyl-
AHCTUNGERTGaNNUN mine hydrochloride (2 equiv), pyridine, 258C, 3 h followed by TsCl
(1.1 equiv), Et3N (1.05 equiv), Et2O, 25 8C, 12 h, 77% (2 steps);
c) (CuOTf)2·toluene (0.05 equiv), K2CO3 (3 equiv), MeCN, 258C, 12 h;
TMSN3 (2 equiv), TBAF (1 equiv), THF, 0–258C, 3 h, 82%. CDI=1,1’-
carbonyldiimidazole; TBAF=tetra-n-butylammonium fluoride.
chloromethane at room temperature proceeded in moderate
yield
(63%). Further optimization
showed that
(CuOTf)2·toluene was an ideal catalyst for this transforma-
tion, giving the highest yield (94%) as determined by crude
1H NMR spectroscopy (Table 1). With these optimized aziri-
dination conditions, ring opening with N-containing nucleo-
philes was sequentially conducted in a one-pot manner. Un-
satisfactorily, p-methoxybenzylamine and allyl amine were
introduced to compound 4 in a one-pot process that gave
two diastereomers (approximate ratio: anti/syn 3:1) after
ring-opening. To our surprise, when TMSN3 was used as the
nucleophile, compound 11 was formed stereoselectively and
regioselectively in 82% yield. Notably, this aziridine inter-
mediate has the provision for synthesis of Tamiflu analogues
by using different N-, O-, or S-containing nucleophiles. Opti-
mization of the reaction conditions is under investigation to
Copper-catalyzed intramolecular aziridination: The installa-
tion of the three contiguous chiral centers in the molecule
of Tamiflu, which is probably the most difficult part in all
Tamiflu synthesis methods proposed thus far, was carried
Chem. Eur. J. 2010, 16, 4533 – 4540
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4535

P. Chen, X. W. Liu et al.
Table 1. Optimization of aziridination of tosyl-carbamate 4.[a]
ed with MsCl/Et3N to produce an aziridine intermediate,
which was subjected to regioselective ring-opening with 3-
pentanol/BF3·Et2O to generate ether 14 in 52% yield. The
stereochemistry of 14 was confirmed by using X-ray crystal-
lography (see the Supporting Information). The azido group
in 14 was reduced to an amine with PPh3 in THF/H2O (90%
yield). Finally, the oseltamivir phosphate salt (Tamiflu) was
obtained by treating 1 with H3PO4 in EtOH (85% yield).
Hydrolysis of 1 with LiOH in water afforded oseltamivir
acid 2.
Catalyst
Rh2A(OAc)4
Cu(OTf)2
IprCuCl[c]
Cu(MeCN)4PF6
(CuOTf)2toluene
Solvent
Yield [%][b]
G
CH2Cl2
CH3CN
CH3CN
CH3CN
CH3CN
63
78
83
86
94
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
[a] All reactions were conducted at RT. [b] Yields, judged from crude
1H NMR data. [c] 1,3-bis(2,6-diisopropylphenyl)imidazolium copper chlo-
ride.[26]
Our synthetic route to Tamiflu is based on the cheap and
abundantly available starting material d-glucal. By taking
advantage of the naturally occurring chiral centers and
innate stereochemistry of the d-glucal scaffold, most of the
reactions in our synthesis occur in a regio- and stereoselec-
tive manner to readily produce the desired products in high
yields. Other highlights of this work include the 3,3-sigma-
tropic rearrangement reaction, which allows the construction
of a cyclohexene core with a conjugated carboxylate moiety.
Similar strategies can be employed to form other functional-
ized six-membered carbocycles from abundantly available
O-containing sugar scaffolds. In our method, vicinal diamino
groups on C4 and C5 were introduced by stereoselective in-
tramolecular nitrogen delivery with a tethered carbamate at
the C3 position, followed by regio- and stereoselective ring
opening of aziridine. Elegantly, this strategy fixes the stereo-
chemistry at both C4 and C5. The third key step is the in-
stallation of a pentyloxy group at C3, which was achieved
through regio- and stereoselective ring opening of the aziri-
dine intermediate. Furthermore, the functionalization on the
three contiguous chiral centers, C3, C4, and C5, which occur
at the late stage of our synthesis, allow prompt and efficient
synthesis of Tamiflu analogues by varying the nucleophiles
and alkyloxy groups.
improve the stereoselectivity when other nucleophiles are
applied.
Completion of synthesis of Tamiflu: Treatment of 11 with
1,8-diazabicycloACHTUNGTRENNUNG[5.4.0]undec-7-ene (DBU) in acetonitrile at
room temperature for 24 h failed to generate the desired
a,b-unsaturated carboxylate. Chirality at the C1 position
was partially inverted to give a pair of diastereoisomers
through a deprotonation/protonation process. Interestingly,
aromatization occurred when 11 was treated with DBU at
higher temperature. However, the elimination was successful
after acetylating the nitrogen in carbamate 11 with AcCl/
NaH, with an overall yield of 67% for the two steps
(Scheme 4). Substituted carbamate 12 was then hydrolyzed
by using Cs2CO3 in EtOH to provide alcohol 13 in 72%
yield.
Influences of Tamiflu and oseltamivir carboxylate on neuro-
endocrine PC12 cells: We examined the effects of Tamiflu
and OC on neuroendocrine PC12 cells, which are a popular
cell model to study vesicular exocytosis (regulated secretion)
of neurotransmitters/hormones and neural differentiation. It
was found that both Tamiflu and OC had no obvious influ-
ences on cell morphology, proliferation, differentiation in-
duced by nerve growth factor, or organization of cytoskele-
ton networks (see the Supporting Information). These obser-
vations support the view that Tamiflu is generally well toler-
ated. On the other hand, we discovered that OC exerts sig-
nificant inhibitory effects on vesicular exocytosis of PC12
cell. This suggests that Tamiflu treatment may adversely
affect neurotransmitter release in the central nervous system
and may thus explain the neuropsychiatric consequences
and other side-effects observed in some patients.
Scheme 4. Complete synthesis of Tamiflu. Reagents and conditions:
a) AcCl (2 equiv), NaH (1.5 equiv), THF, 0–258C, 2 h followed by DBU
(2 equiv), MeCN, 258C, 24 h, 67% (2 steps); b) Cs2CO3 (0.1 equiv),
EtOH, 258C, 3 h, 72%; c) DMP (1.5 equiv), dichloromethane, 258C, 2 h
followed by LiAlH
(tBuO)3 (3 equiv), THF, À20–258C, 3 h, 70%
(2 steps); d) MsCl (1.5 equiv), Et3N (3 equiv), CH2Cl2, 0–258C, 3 h;
.
e) BF3 Et2O (1 equiv), 4 ꢂ MS, 3-pentanol, À208C, 2 h, 52% (2 steps);
f) Ph3P (1.5 equiv), THF/H2O (1:1), reflux, 3 h, 90%; g) 85% H3PO4
(1 equiv), ethanol, 558C, 30 min, 85%; h) LiOH (1.5 equiv), THF/H2O
(10:1), 258C, 3 h, 80%. MS=molecular sieve.
OC impairs vesicle trafficking: The large dense core secreto-
ry vesicles (LDCVs) in PC12 cells were selectively labeled
by overexpression of enhanced green fluorescence protein
(EGFP)-conjugated neuropeptide-Y (NPY-EGFP).[27] Fluo-
rescently labeled vesicles in the subplasmalemmal region
The chirality of C3 was reversed by Dess–Martin oxida-
tion followed by reduction with LiAlH
ACHTUNGTERN(NNUG OtBu)3 to stereospe-
cifically give intermediate 3 in 70% yield. It was then treat-
4536
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 4533 – 4540

Tamiflu Synthesis and Biology
FULL PAPER
were resolved by using total internal reflection fluorescence
microscopy (TIRFM), which is instrumental in observing
phenomena that occur near the cell membrane by selective-
ly illuminating the thin section (<200 nm) just above the in-
terface between the glass coverslip and the adhered cell
membrane. Figure 2a presents typical TIRFM images of an
OC-treated (left) and an untreated PC12 cell.[28] Each bright
fluorescent dot indicates the footprint of a subplasmalemmal
vesicle. There was no appreciable difference in the total
number of visible vesicles between the treated and untreat-
ed cells.
All the secretory vesicles in neuroendocrine or neuron
cells move constantly. They move laterally (parallel to the
cell membrane) when in the subplasmalemmal region, in
which they travel around and interact with various secretory
factors and eventually fuse with the plasma membrane to
discharge their vesicular cargoes into the extracellular fluid
upon being triggered. In addition, they transit vertically be-
tween the subplasmalemmal region and the inner cytosol, in
which a large reserve pool of vesicles resides. The dynamics
of vesicle trafficking directly correlates with the competence
and kinetics of vesicular exocytosis.[27a,29] The motion of indi-
vidual LDCVs was tracked at 0.5 s time intervals over 2 min
to obtain information about the velocity and area coverage
of vesicle lateral movement, the time that vesicles dwell in
the subplasmalemmal region, and the rate of vesicle arrival
from the inner cytosol. Figure 2b depicts the typical motion
trajectories of a vesicle in the OC-treated cell (top)[30] and
the untreated cell (bottom). It is evident that vesicles under-
take confined Brownian motion.[27a]
Interestingly, the velocity of vesicle movement in OC-
treated cells is dramatically reduced (587 vesicles in 11 treat-
ed cells: (55.1Æ1.7) nmsÀ1 vs. 657 vesicles from 14 untreated
cells: (183.1Æ5.5) nmsÀ1; p<0.001; Figure 2c). The dashed
rectangles (Figure 2b) that just
encase the total vesicle foot-
print give a first-order estima-
tion of the coverage of vesicle
random motion and reflect how
vesicles are confined by physi-
cal barriers or molecular tether-
ing. In OC-treated cells the
vesicle movements were more
severely confined than in the
untreated
control
((0.27Æ
0.02) mm2 vs. (0.58Æ0.07) mm2;
p<0.001; Figure 2d). The re-
duction in motion area cannot
be attributed to a decrease in
the vesicle dwell time. On the
contrary, vesicles stay in the
subplasmalemmal region for
much longer in the treated cells
compared with the control
((63.8Æ1.8) s vs. (41.2Æ1.5) s;
p<0.001; Figure 2e). In addi-
tion to its inhibitory influences
on vesicle lateral trafficking,
OC also causes a large decrease
in the rate of vesicle arrival
from the inner cytosol ((23.3Æ
3.1) vesiclescellÀ1 arrived in
2 min
vs.
(52.1Æ3.6) vesi-
clescellÀ1; p<0.001; Figure 2f).
The rate of vesicle retrieval
back into the inner cytosol,
which is in balance with vesicle
arrival to keep the total
number of subplasmalemmal
vesicles constant, was similarly
affected. Taken together, it can
be seen that OC severely im-
pairs both lateral and vertical
trafficking of LDCV vesicles in
Figure 2. Oseltamivir carboxylate (OC) and Tamiflu reduce vesicle trafficking (both lateral and vertical move-
ment) in PC12 cells. a) Typical TIRFM images of a PC12 cell with (left) or without (right) OC treatment
(0.5 mm for 20 h). Individual footprints of the subplasmalemmal vesicles can be resolved as bright green dots
(highlighted by circles). The cell contours are outlined by the dashed lines. b) Motion trajectories of a vesicle
in an OC-treated cell (top) and a vesicle in an untreated cell (bottom). The dashed rectangle that exactly en-
cases the total vesicle footprints gives a first-order estimation of the area coverage of vesicle lateral motion
(scale bars=200 nm). c), d), e), and f) The statistics of the vesicle average velocity, motion area, dwell time,
and total number of vesicles arrived from the inner cytosol during 2 min imaging, respectively. The statistics,
shown as mean ÆSEM (standard error of the mean), are from 587 vesicles in 11 OC-treated cells, 564 vesicles
in 10 Tamiflu-treated cells, or 657 vesicles from 14 untreated control cells. Student t test: * p<0.05; ** p<
0.01; *** p<0.001 vs. control.
Chem. Eur. J. 2010, 16, 4533 – 4540
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4537

P. Chen, X. W. Liu et al.
PC12 cells. In comparison with OC, Tamiflu also similarly
suppressed vesicle trafficking, but to lesser extent
(Figure 2). Sufficient mobility is believed to be important to
ensure vesicle fusion competence and fast replenishment of
readily releasable vesicles during a continuous stimula-
cell, in response to a local perfusion of high-K+ solution
that invokes a voltage-activated Ca2+ current and in turn
triggers Ca2+-dependent exocytosis. Each current spike cor-
responds to a single vesicular release of catecholamine mol-
ecules (including dopamine, epinephrine, and norepineph-
a
A
ACHTUNGTRENrNUGN ine). As compared with the control, the average number of
ate may inhibit vesicular exocytosis.
amperometric spikes in response to high-K+ stimulation for
2 min was significantly inhibited by OC treatment (19 OC-
treated cells: (22.8Æ3.8) spikescellÀ1 vs. 40 control cells:
OC inhibits vesicular exocytosis: Vesicular exocytosis was as-
sayed by carbon fiber microelectrode (CFM)-based amper-
ometry, which is able to detect exocytosis with single vesicle
sensitivity and millisecond resolution.[31] Figure 3a[30] shows a
typical amperometric trace, recorded by using a voltage-
biased (700 mV) CFM positioned on the surface of a PC12
(45.4Æ4.1) spikescellÀ1
;
p<0.001; Figure 3a, bottom),
whereas Tamiflu did not appreciably affect the extent of
exocytosis (21 Tamiflu-treated cells: (41.9Æ4.7) spikescellÀ1
;
p>0.05).
The characteristic waveform of the amperometric spikes
reveals the kinetics of quantal vesicle fusion catalyzed by se-
cretory proteins, such as SNAREs.[31] Individual amperomet-
ric spikes were extracted and analyzed to investigate the ef-
fects of Tamiflu and OC on vesicle fusion kinetics. The
mean amperometric spikes averaged from all the recorded
signals from the control cells (1814 spikes from 40 cells),
TamACTHNUGRTENUNGiflu-treated cells (880 spikes from 21 cells), and OC-
treated cells (434 spikes from 19 cells) are displayed in Fig-
ure 3b. The mean amperometric amplitude from the OC-
treated cells was much smaller than that from the control
cells ((8.83Æ0.49) pA vs. (15.91Æ0.47) pA; p<0.001). In
comparison, the amplitude reduction caused by Tamiflu was
less ((10.53Æ0.33) pA; p<0.001).
The rise slope, which is defined by the slope of the am-
perometric current as it rises from 35 to 90% of its peak, re-
flects how fast the initial fusion pore formed between the
vesicular and the plasma membrane expands to quickly dis-
charge the molecules inside. It was observed that the rise
slope in the OC-treated cells was much less steep than that
in the control cells ((5.11Æ0.46) pAmsÀ1 vs. (10.58Æ
0.46) pAmsÀ1; p<0.001; Figure 3c). The half-width time
AHCTNUGTER(GNNUN t1/2) of the amperometric signal indicates the time of the
fusion process. It was found that the t1/2 of the OC-treated
cells was significantly longer than that of the control
((6.15Æ0.19) ms vs. (5.13Æ0.09) ms; p<0.001; Figure 3c).
From these observations, it can clearly be seen that the
fusion kinetics are greatly slowed by treatment with OC.
Furthermore, the size of the quantal vesicular release (Q),
which is the charge integration of the amperometric spike
and the indicator of the total number of released molecules,
was reduced by OC treatment ((69.69Æ3.19) fC vs. (87.87Æ
2.03) fC; p<0.001; Figure 3c). Similarly, Tamiflu treatment
also led to reduction in the rise slope and quantal size (Fig-
ure 3c). Taken together, OC significantly impaired not only
the vesicular fusion competence (Figure 3a) but also the ef-
ficiency of quantal vesicular release (Figure 3b and c).
Figure 3. Oseltamivir carboxylate inhibits vesicular exocytosis in PC12
cells. a) A representative amperometric recording from a untreated PC12
cell in response to high-K+ stimulation for 2 min (top)[30] and the statis-
tics of the average total spike number (bottom) from the control cells
(40 cells: (45.4Æ4.1) spikescellÀ1), Tamiflu-treated cells (21 cells: (41.9Æ
4.7) spikescellÀ1), and OC-treated cells (19 cells: (22.8Æ3.8) spikescellÀ1).
b) The average amperometric spikes averaged from 1814 fusion events
from the control cells (c), 880 fusion events recorded from Tamiflu-
treated cells (g), and 434 fusion events recorded from OC-treated
cells (a). c) Statistics of the rise slope, half-width time (t1/2), and quan-
tal size (Q) of the amperometric spike. The data is normalized to the
It should not be surprising that Tamiflu, particularly after
being hydrolyzed to its active form (OC), could have pro-
found influences in cell functions because its target, neur-
control, for which rise slope=(10.58Æ0.46) pAmsÀ1
;
t
1/2 =(5.13Æ
AHCTUNGTREGaNNNU minidases, are implicated in many essential cellular pro-
0.09) ms; Q=(87.87Æ2.03) fC. In a) and c), the statistics are represented
as mean ÆSEM; white column: control cells, dotted column: Tamiflu-
treated cells, slashed column: OC-treated cells. Student t test: *** p<
0.001 vs. control.
cesses in which they remove the terminal sialic acid from
glycoproteins and glycolipids.[18–19] The influences of Tamiflu
and OC on the central nervous system have been investigat-
4538
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 4533 – 4540

Tamiflu Synthesis and Biology
FULL PAPER
[2] a) C. U. Kim, W. Lew, M. A. Williams, H. Liu, L. Zhang, S. Swam-
ed by using rat brain synaptosomes, and it was found that
neither Tamiflu nor OC affected the release of monoamine
neurotransmitters.[32] In another study, it was reported that
Tamiflu increased dopamine levels in the rat medial prefron-
tal cortex.[33] For the reasons that require further investiga-
tion, these observations obtained from different experimen-
tal preparations by using biochemical assays are seemingly
discrepant with our findings on PC12 cells obtained by
single-cell recording or imaging with single vesicle sensitivity
and millisecond resolution. Our experiments demonstrate
that OC significantly inhibits exocytosis, which is a funda-
mental process in neurons and many other secretory cells.
Therefore, it is possible that, once the hydrolyzed Tamiflu
(OC) passes the blood–brain barrier,[34] it might inhibit the
release of neurotransmitters from presynaptic neurons and
consequently modulate neurotransmission in the central
nervous system. Excitation or depression of the central
nervous system may result depending on the types of neu-
rons and neurotransmitters being affected. The ability of
AHCTUNGTREGiNNUN nathan, N. Bischofberger, M. S. Chen, D. B. Mendel, C. Y. Tai,
W. G. Laver, R. C. Stevens, J. Am. Chem. Soc. 1997, 119, 681–690;
b) C. U. Kim, W. Lew, M. A. Williams, H. Wu, L. Zhang, X. Chen,
P. A. Escarpe, D. B. Mendel, W. G. Laver, R. C. Stevens, J. Med.
Chem. 1998, 41, 2451–2460; c) M. Federspiel, R. Fischer, M.
Hennig, H. J. Mair, T. Oberhauser, G. Rimmler, T. Albiez, J. Bruhin,
H. Estermann, C. Gandert, V. Gockel, S. Gotzo, U. Hoffmann, G.
Huber, G. Janatsch, S. Lauper, O. Rockel-Stabler, R. Trussardi,
A. G. Zwahlen, Org. Process Res. Dev. 1999, 3, 266–274; d) P. J.
Harrington, J. D. Brown, T. Foderaro, R. C. Hughes, Org. Process
Res. Dev. 2004, 8, 86–91; e) L.-D. Nie, X.-X. Shi, K. H. Ko, W.-D.
Lu, J. Org. Chem. 2009, 74, 3970–3973; f) M. Karpf, R. Trussardi,
Angew. Chem. 2009, 121, 5871–5873; Angew. Chem. Int. Ed. 2009,
48, 5760–5762.
[3] a) V. Farina, J. D. Brown, Angew. Chem. 2006, 118, 7488–7492;
Angew. Chem. Int. Ed. 2006, 45, 7330–7334; b) M. Shibasaki, M.
Kanai, Eur. J. Org. Chem. 2008, 1839–1850; c) J. Gong, W. Xu,
Curr. Med. Chem. 2008, 15, 3145–3159; d) M. Karpf, R. Trussardi, J.
Org. Chem. 2001, 66, 2044–2051; e) M. Javier, Chem. Rev. 2009, 109,
4398–4438; f) J. Andraos, Org. Process Res. Dev. 2009, 13, 161–185.
[4] a) Y. Y. Yeung, S. Hong, E. J. Corey, J. Am. Chem. Soc. 2006, 128,
6310–6311; b) K. Yamatsugu, S. Kamijo, Y. Suto, M. Kanai, M. Shi-
basaki, Tetrahedron Lett. 2007, 48, 1403–1406; c) K. Yamatsugu, L.
Yin, S. Kamijo, Y. Kimura, M. Kanai, M. Shibasaki, Angew. Chem.
2009, 121, 1090–1096; Angew. Chem. Int. Ed. 2009, 48, 1070–1076;
d) N. T. Kipassa, H. Okamura, K. Kina, T. Hamada, T. Iwagawa,
Org. Lett. 2008, 10, 815–816.
TamACHTUNGTRENNUNGiflu and OC to affect synaptic function and neuronal ex-
citability has been reported in two recent studies.[35]
Conclusion
[5] H. Ishikawa, T. Suzuki, Y. Hayashi, Angew. Chem. 2009, 121, 1330–
1333; Angew. Chem. Int. Ed. 2009, 48, 1304–1307.
We have developed an economic synthetic route to Tamiflu
by using cheap and abundantly available starting materials
and reagents. In addition, our method allows late-stage func-
tionalization for ready and flexible synthesis of Tamiflu ana-
logues that may be able to overcome virus resistance to
[6] J. C. Rohloff, K. M. Kent, M. J. Postich, M. W. Becker, H. H. Chap-
man, D. E. Kelly, W. Lew, M. S. Louie, L. R. McGee, E. J. Prisbe,
L. M. Schultze, R. H. Yu, L. Zhang, J. Org. Chem. 1998, 63, 4545–
4550.
[7] X. Cong, Z. J. Yao, J. Org. Chem. 2006, 71, 5365–5368.
[8] J. J. Shie, J. M. Fang, S. Y. Wang, K. C. Tsai, Y. S. E. Cheng, A. S.
Yang, S. C. Hsiao, C. Y. Su, C. H. Wong, J. Am. Chem. Soc. 2007,
129, 11892–11893.
TamACHTUNGTRENNUNGiflu and to avoid the side-effect problems of Tamiflu.
The practical synthetic strategy presented here may provide
not only an alternative to the synthesis of Tamiflu but also
the opportunity of designing Tamiflu-like molecular tools to
study the cellular functions mediated by neuraminidases.
Herein we show that the active Tamiflu metabolite (OC)
significantly inhibits the vesicular exocytosis in neuroendo-
crine PC12 cells. This inhibition on the ubiquitous exocytotic
process may explain why Tamiflu causes a variety of prob-
lems, including nausea, vomiting, diarrhea, headaches, verti-
go, insomnia, somnolence, and behavioral excitement. Fur-
ther investigations will be carried out to reveal the detailed
molecular mechanisms of how Tamiflu interferes with the
exocytotic pathways and possibly other cell functions that
involve neuraminidases.
[9] a) Y. Fukuta, T. Mita, N. Fukuda, M. Kanai, M. Shibasaki, J. Am.
Chem. Soc. 2006, 128, 6312–6313; b) T. Mita, N. Fukuda, F. X.
Roca, M. Kanai, M. Shibasaki, Org. Lett. 2007, 9, 259–262.
[10] a) K. M. Bromfield, H. Graden, D. P. Hagberg, T. Olsson, N. Kann,
Chem. Commun. 2007, 3183–3185; b) J.-J. Shie, J.-M. Fang, C.-H.
Wong, Angew. Chem. 2008, 120, 5872–5875; Angew. Chem. Int. Ed.
2008, 47, 5788–5791; c) M. Matveenko, A. C. Willis, M. G. Banwell,
Tetrahedron Lett. 2008, 49, 7018–7020.
[11] B. M. Trost, T. Zhang, Angew. Chem. 2008, 120, 3819–3821; Angew.
Chem. Int. Ed. 2008, 47, 3759–3761.
[12] a) N. Satoh, T. Akiba, S. Yokoshima, T. Fukuyama, Tetrahedron
2009, 65, 3239–3245; b) N. Satoh, T. Akiba, S. Yokoshima, T. Fu-
kuyama, Angew. Chem. 2007, 119, 5836–5838; Angew. Chem. Int.
Ed. 2007, 46, 5734–5736.
[13] U. Zutter, H. Iding, P. Spurr, B. Wirz, J. Org. Chem. 2008, 73, 4895–
4902.
[14] T. Mandai, T. Oshitari, Synlett 2009, 783–786.
[15] T. Oshitari, T. Mandai, Synlett 2009, 787–789.
[16] B. Sullivan, I. Carrera, M. Drouin, T. Hudlicky, Angew. Chem. 2009,
121, 4293–4295; Angew. Chem. Int. Ed. 2009, 48, 4229–4231.
[17] a) I. Fuyuno, Nature 2007, 446, 358–359; b) M. Long, Cell Res. 2007,
17, 309–310; c) R. Hama, Int. J. Risk Safety Med. 2008, 20, 5–36;
d) http://www.rocheusa.com/products/Tamiflu/pi.pdf; e) T. Jefferson,
M. Jones, P. Doshi, C. Del Mar, Lancet 2009, 374, 1312–1313.
[18] T. Miyagi, T. Wada, K. Yamaguchi, K. Hata, K. Shiozaki, J. Bio-
chem. 2008, 144, 279–285.
Acknowledgements
We gratefully acknowledge support from Nanyang Technological Univer-
sity (RG 50/08) and the support by an AcRF tier 2 grant (T206B3220)
from the Ministry of Education, Singapore. We also thank Dr. Yong-Xin
Li for X-ray analyses.
[19] E. Monti, A. Preti, B. Venerando, G. Borsani, Neurochem. Res.
2002, 27, 649–663.
[20] N. Papini, L. Anastasia, C. Tringali, G. Croci, R. Bresciani, K. Yama-
guchi, T. Miyagi, A. Preti, A. Prinetti, S. Prioni, S. Sonnino, G. Tetta-
[1] a) N. V. Joseph, Drug Dev. Res. 1999, 46, 176–196; b) A. Moscona,
N. Engl. J. Med. 2005, 353, 1363–1373; c) I. M. Lagoja, E. D. Clercq,
Med. Res. Rev. 2008, 28, 1–38.
Chem. Eur. J. 2010, 16, 4533 – 4540
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4539

P. Chen, X. W. Liu et al.
manti, B. Venerando, E. Monti, J. Biol. Chem. 2004, 279, 16989–
16995.
[21] X. L. Nan, I. Carubelli, N. M. Stamatos, J. Leukocyte Biol. 2007, 81,
284–296.
[22] a) J. A. Rodriguez, E. Piddini, T. Hasegawa, T. Miyagi, C. G. Dotti,
J. Neurosci. 2001, 21, 8387–8395; b) M. Saito, Y. Tanaka, C. P. Tang,
R. K. Yu, S. Ando, J. Neurosci. Res. 1995, 40, 401–406.
[23] G. Bꢁchi, J. E. Powell, J. Am. Chem. Soc. 1967, 89, 4559–4560.
[24] a) R. Lorpitthaya, Z.-Z. Xie, J.-L. Kuo, X.-W. Liu, Chem. Eur. J.
2008, 14, 1561–1570; b) R. Lorpitthaya, Z.-Z. Xie, K. B. Sophy, J.-L.
Kuo, X.-W. Liu, Chem. Eur. J. 2010, 16, 588–594.
[25] A. Padwa, A. C. Flick, C. A. Leverett, T. Stengel, J. Org. Chem.
2004, 69, 6377–6386.
[26] R. Liu, S. R. Herron, S. A. Fleming, J. Org. Chem. 2007, 72, 5587–
5591.
[27] a) E. M. Zhang, R. H. Xue, J. Soo, P. Chen, Pfluegers Arch. 2008,
457, 211–222; b) J. Zhang, D. L. Fu, M. B. Chan-Park, L. J. Li, P.
Chen, Adv. Mater. 2009, 21, 790–793.
[29] a) V. E. Degtyar, M. W. Allersma, D. AxelACTHNGUTERNrUNG od, R. W. Holz, Proc.
Natl. Acad. Sci. USA 2007, 104, 15929–15934; b) R. W. Holz, D. Ax-
elrod, Acta Physiol. 2008, 192, 303–307.
[30] C. Dubost, I. E. Marko, T. Ryckmans, Org. Lett. 2006, 8, 5137–5140.
[31] R. H. Xue, Y. Y. Zhao, P. Chen, Biochem. Biophys. Res. Commun.
2009, 380, 371–376.
[32] K. Satoh, R. Nonaka, A. Gata, D. Nakae, S. I. Uehara, Biol. Pharm.
Bull. 2007, 30, 1816–1818.
[33] T. Yoshino, K. Nisijima, K. Shioda, K. Yui, S. Kato, Neurosci. Lett.
2008, 438, 67–69.
[34] K. Morimoto, M. Nakakariya, Y. Shirasaka, C. Kakinuma, T. Fujita,
I. Tamai, T. Ogihara, Drug Metab. Dispos. 2008, 36, 6–9.
[35] a) Y. Izumi, K. Tokuda, K. A. OꢃDell, C. F. Zorumski, T. Narahashi,
Neurosci. Lett. 2007, 426, 54–58; b) A. Usami, T. Sasaki, N. Satoh,
T. Akiba, S. Yokoshima, T. Fukuyama, K. Yamatsugu, M. Kanai, M.
Shibasaki, N. Matsuki, Y. Ikegaya, J. Pharmacol. Sci. 2008, 106, 659–
662.
Received: July 23, 2009
[28] A. Burgess, S. R. Wainwright, L. S. Shihabuddin, U. Rutishauser, T.
Seki, I. Aubert, Dev. Neurobiol. 2008, 68, 1580–1590.
Revised: November 2, 2009
Published online: March 9, 2010
4540
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 4533 – 4540