Influenza neuraminidase inhibitors possessing a novel hydrophobic interaction in the enzyme active site: Design, synthesis, and structural analysis of carbocyclic sialic acid analogues with potent anti-influenza activity

Article abstract of DOI:10.1021/ja963036t

The design, synthesis, and in vitro evaluation of the novel carbocycles as transition-state-based inhibitors of influenza neuraminidase (NA) are described. The double bond position in the carbocyclic analogues plays an important role in NA inhibition as d

Full text of DOI:10.1021/ja963036t

J. Am. Chem. Soc. 1997, 119, 681-690
681
Influenza Neuraminidase Inhibitors Possessing a Novel
Hydrophobic Interaction in the Enzyme Active Site: Design,
Synthesis, and Structural Analysis of Carbocyclic Sialic Acid
Analogues with Potent Anti-Influenza Activity
Choung U. Kim,*,† Willard Lew,† Matthew A. Williams,† Hongtao Liu,†
Lijun Zhang,† S. Swaminathan,† Norbert Bischofberger,† Ming S. Chen,†
Dirk B. Mendel,† Chun Y. Tai,† W. Graeme Laver,‡ and Raymond C. Stevens§
Contribution from Gilead Sciences Inc., 353 Lakeside DriVe, Foster City, California 94404, John
Curtin School of Medical Research, The Australian National UniVersity, Canberra 260, Australia,
and Department of Chemistry, UniVersity of California, Berkeley, California 94720
ReceiVed August 28, 1996X
Abstract: The design, synthesis, and in vitro evaluation of the novel carbocycles as transition-state-based inhibitors
of influenza neuraminidase (NA) are described. The double bond position in the carbocyclic analogues plays an
important role in NA inhibition as demonstrated by the antiviral activity of 8 (IC50 ) 6.3 µM) vs 9 (IC50 > 200 µM).
Structure-activity studies of a series of carbocyclic analogues 6a-i identified the 3-pentyloxy moiety as an apparent
optimal group at the C3 position with an IC50 value of 1 nM for NA inhibition. The X-ray crystallographic structure
of 6h bound to NA revealed the presence of a large hydrophobic pocket in the region corresponding to the glycerol
subsite of sialic acid. The high antiviral potency observed for 6h appears to be attributed to a highly favorable
hydrophobic interaction in this pocket. The practical synthesis of 6 starting from (-)-quinic acid is also described.
Introduction
glycoproteins expressed by both influenza A and B viruses. HA
is known to mediate binding of viruses to target cells via
terminal sialic acid residue in glycoconjugates. This binding
is the first step of viral infection. In contrast to HA activity,
NA catalyzes removal of terminal sialic acids linked to
glycoproteins and glycolipids. Although biological conse-
quences of this activity are not completely understood, it has
been postulated that NA activity is necessary in the elution of
newly formed viruses from infected cells by digesting sialic
acids in the HA receptor.8,9 NA may also promote viral
movement through respiratory tract mucus, thus enhancing viral
infectivity.9,10 Therefore, NA has been considered to be a
suitable target for designing agents against influenza viruses.
In earlier studies, 2,3-didehydro-2-deoxy-N-acetylneuraminic
acid (Neu5Ac2en, 1) was found to be an influenza NA inhibitor
with a Ki of 4 µM.11 Biochemical studies12 indicate 1 is
considered a transition state-like analogue binding to the active
site of NA.13,14 Recently, on the basis of structural information
generated from the X-ray crystallographic study of 1 complexed
with NA, the following rationally designed NA inhibitors were
prepared: 2,3-dihydro-2,4-dideoxy-4-amino-N-acetylneuraminic
acid (4-amino-Neu5Ac2en, 2) and its guanidino analogue (4-
Despite considerable progress in elucidating the molecular
mechanism and cellular biology of influenza virus, influenza
infection continues to be the most serious respiratory disease
both in terms of morbidity and mortality.1 Current available
treatments have severe limitations. Amantidine and its analogue
rimantidine are the only compounds licensed for the treatment
and prophylaxis of influenza A infection. However, these
compounds are not effective against influenza B viruses, and
their clinical use has been limited by side effects and the rapid
emergence of resistant viral strains.2 Vaccine development has
been partially successful in the control of influenza epidemics
due to the highly variable mutation of influenza virus.3
The unique replication mechanism of influenza virus has
allowed investigators to identify a number of potential molecular
targets for drug design. Those targets include haemagglutinin,4
neuraminidase,5 M2 protein,6 and endonuclease.7 Haemagglu-
tinin (HA) and neuraminidase (NA) are two major surface
† Gilead Sciences Inc.
‡ The Australian National University.
§ University of California.
X Abstract published in AdVance ACS Abstracts, January 1, 1997.
(1) Lui, K.-J.; Kendal, A. P. Am. J. Public Health 1987, 77 (6), 712.
(2) Saul, H. New Scientist 1995, 26.
(8) Palese, P.; Jobita, K.; Ueda, M.; Compans, R. W. Virology 1974,
61, 397.
(3) (a) Couch, R. B.; Six, H. R. In AntiViral Chemotherapy: New
Direction for Clinical Application and Research; Mills J., Corey L., Eds.;
Elsevier Science Publishing: Oxford, 1986; pp 50-56. (b) Hayden, F. G.;
Belshe, R. B.; Clover, R. D.; Hay, A. J.; Oakes, M. G.; Soo, W. N. Engl.
J. Med. 1989, 321, 1696.
(4) (a) Sauter, N. K.; Hanson, J. E.; Glick, G. D.; Brown, J. H.; Crowther,
R. L.; Park, S. J.; Skehel, J. J.; Wiley, D. C. Biochemistry 1992, 31, 9609.
(b) Mammen, M.; Dahmann, G.; Whitesides, G. J. Med. Chem. 1995, 38,
4179.
(5) (a) Colman, P. M. In The influenza Viruses: Influenza Virus
neuraminidase, Enzyme and Antigen; Krug, R. M., Ed.; Plenum Press: New
York, 1989; pp 175-218. (b) Colman, P. M. Protein Sci. 1994, 3, 1687.
(6) Hay, A. J.; Wolstenholme A. J.; Skehel, J. J.; Smith, M. H. EMBO
J. 1985, 4, 3021.
(7) Hastings, J. C.; Selnick, H.; Wolanski, B.; Tomassini, J. E. Antimi-
crob. Agents Chemother. 1996, 40, 1304.
(9) Liu, C.; Eichelberger, M. C.; Compans, R. W.; Air G. M. J. Virol.
1995, 69, 1099.
(10) Klenk, H. D.; Rott, R. AdV. Virus Res. 1988, 34, 247.
(11) (a) Meinal, P.; Bodo, G.; Palese, P.; Schulman, J.; Tuppy, H.
Virology 1975, 58, 457. (b) Holzer, C. T.; von Itzstein, M.; Jin, B.; Pegg,
M. S.; Stewart, W. P.; Wu, W.-Y. Glycoconjugate J. 1993, 10, 40.
(12) Chong, A. K. J.; Pegg, M. S.; von Itzstein, M. Biochem. Int. 1991,
24, 165.
(13) Taylor, N. R.; von Itzstein, M. J. Med. Chem. 1994, 37, 616.
(14) Janakiraman, M. N.; White C. L.; Laver, W. G.; Air, G. M.; Luo,
M. Biochemistry 1994, 33, 8172.
(15) Palese, P.; Schulman, J. In Chemoprophylaxis and Virus Infections
of the Upper Respiratory Tract; Oxford, J. S., Ed.; CRC: Cleveland, 1977;
Vol. 1, pp 189-205.
(16) Nohle, U.; Beau, J. M.; Schauer, R. Eur. J. Biochem. 1982, 126,
543.
S0002-7863(96)03036-3 CCC: $14.00 © 1997 American Chemical Society

682 J. Am. Chem. Soc., Vol. 119, No. 4, 1997
Kim et al.
guanidino-Neu5Ac2en, 3).13,17,18 In comparison to 1, both 2
there are no subtypes in the type B virus. Although amino acid
sequence homology among NA from both type A and type B
virus strains has been found to be only 30%,23 enzyme activity
of NA among the different strains is the same, indicating the
highly conserved nature of the active site of the enzyme.24 The
X-ray crystallographic structures of NA have been determined
from three influenza subtypes: A/Tokyo,25 A/Tern,26 and
B/Beijing.27 The structures displayed a symmetrical folding
pattern of six four-stranded antiparallel â-sheets arranged like
blades of a propeller. NA exists as a mushroom-shaped spike
with a boxlike head on top of a long stalk containing a
hydrophobic region by which it is embedded in the viral
membrane.
and 3 are more potent NA inhibitors with Ki values of 10-8
M
and 10-10 M, respectively. Both the amino group in 2 and the
guanidino group in 3 are suggested to form salt bridges with
Glu119 in the NA active site, while the latter adds a strong
charge-charge interaction with Glu227.18 In addition, 3 also
exhibited potent antiviral activity against a variety of influenza
A and B strains in the cell culture assay.19 Compound 3 is
currently being evaluated in human clinical trials and has shown
efficacy in phase II challenge studies in both prophylaxis and
treatment of influenza virus infections.21,22 However, poor oral
bioavailability and rapid excretion precluded 3 as a potential
oral agent against influenza infection and 3 has to be admin-
istered by either intranasal or inhaled routes in clinical trials.20
Crystallographic studies of NA reveal that, in the active site,
the amino acids which line and surround the walls of the binding
pocket are highly conserved among all influenza strains
examined so far. The high-resolution crystallographic structure
of sialic acid (4a) complexed with NA revealed that sialic acid
binds the enzyme in a considerably deformed conformation due
to the strong ionic interactions between the carboxylate of the
substrate and Arg118, -292, and -371 in the active site of the
enzyme.28 In solution the carboxylate of sialic acid is axial,
but the deformation of the ring on binding put the carboxylate
into a pseudoequatorial position. This binding mode is very
similar to that found in the X-ray crystal structure of Neu5Ac2en
(1) complexed with NA.26 In this case, the double bond of
Neu5Ac2en constrains the pyranose ring of the sugar into a
planar structure around the ring oxygen. On the basis of this
structural information, it has been proposed that the catalytic
mechanism for the cleavage of sialic acid from glycoconjugates
(4b) implicates the formation of the C2 carbonium cation 5
which is stabilized by the neighboring oxygen atom as shown
in Scheme 1.13,14 Furthermore, kinetic isotope studies have
provided convincing evidence for the C2 carbonium cation
formation as an intermediate structurally similar to the transition
state in the sialic acid cleavage by NA.12
Transition-state mimics frequently are potent inhibitors for
the catalyzing enzyme. The concept of structural similarity to
the transition state has found wide application in drug design
over the years. The multitude of enzyme-inhibitor interactions
are governed by steric as well as electronic factors. In theory,
compounds that closely resemble the transition-state structure
should give high binding affinity toward the target enzyme.29
Using intermediate 5 as a key transition-state mimic is a
reasonable approximation in view of the X-ray crystallographic
studies described above. Considering the flat oxonium cation
in 5 as an isostere of the double bond, the cyclohexene scaffold
was selected as a replacement for the oxonium ring of 5 which
would keep the conformational changes to a minimum (Scheme
1). In addition, the carbocyclic system was expected to be
chemically versatile for the manipulation of side chains attached
to the ring. Earlier attempts by us and others to mimic the
intermediate 5 with the completely flat benzene ring did not
lead to potent NA inhibitors,30 suggesting the importance of
In the case of an influenza epidemic, oral administration may
be a more convenient and economical method for treatment and
prophylaxis. Therefore, it would be desirable to have a new
class of orally active NA inhibitors as potential agents against
influenza infection. Our first move toward this objective
involved the design of a new class of compounds by using a
carbocyclic template in place of the dihydropyran ring of the
Neu5Ac2en system. It is expected that the carbocyclic ring
would be chemically more stable than the dihydropyran ring
and easier to modify for optimization of antiviral and pharma-
cological properties.
The validity of this approach was verified by the discovery
of very potent NA inhibitors in this new carbocyclic series. New
lipophilic side chains at the C3 position of the carbocyclic system
imparted potent NA inhibitory activity. X-ray crystallographic
analysis of the carbocyclic analogue bound to NA confirmed
that there was in fact hydrophobic space in the glycerol-binding
subsite to accommodate bulky lipophilic groups. The discovery
of this hydrophobic pocket in the active site of NA was exploited
to increase the lipophilicity of inhibitors to optimize pharma-
cologic properties for potential oral bioavailability while
maintaining potent antiviral activity.
This work constitutes an example of rational drug design
based on information available from the crystal structure of
inhibitors complexed with the enzyme in the active site and
highlights the great importance of the hydrophobic interaction
for the high-affinity binding of inhibitors toward the enzyme.
Design
NA has been classified into nine subtypes for type A influenza
virus strains according to their serological properties. However,
(17) von Itzstein, M.; Wu, W.-Y.; Jin, B. Carbohydr. Res. 1994, 259,
301.
(18) von Itzstein, M.; Wu, W.-Y.; Kok, G. B.; Pegg, M. S.; Dyason, J.
C.; Jin, B.; Phan, T. V.; Smythe, M. L.; White, H. F.; Oliver, S. W.; Colman,
P. M.; Varghese, J. N.; Ryan, D. M.; Wood, J. M.; Bethell, R. C.; Hotham,
V. J.; Cameron, J. M.; Penn, C. R. Nature (London) 1993, 363, 418.
(19) Wood, J. M.; Bethell, R. C., Bethell, R. C.; Coates, J. A.; Healy,
N.; Hiscox, S. A.; Pearson, B. A.; Ryan, D. M.; Ticehurst, J.; Tilling, J.;
Walcott, S. M.; Penn, C. R. Antimicrob. Agents Chemother. 1993, 37, 1473.
(20) (a) Ryan, D. M.; Ticehurst, J.; Demsey, M. H.; Penn, C. R.
Antimicrob. Agents Chemother. 1994, 38, 2270. (b) Hayden, F. G.; Treanor,
J. J.; Betts, R. F.; Lobo, M.; Esinhart, J.; Hussey J. Am. Med. Assoc. 1996,
275, 295.
(23) (a) Air, G. M.; Laver, W. G; Luo, M.; Stray, S. J.; Legrone, G.;
Webster, R. G. Virology 1990, 177, 578. (b) Blok, J.; Air, G. M.; Laver,
W. G.; Ward, C. W.; Lilley, G. G.; Wood, E. F.; Roxburgh, C. M.; Inglis,
A. S. Virology 1982, 119, 109.
(24) Colman, P. M. Pept. Protein ReV. 1984, 4, 215.
(25) Varghese, J. H.; Colman, P. M. J. Mol. Biol. 1991, 221, 473.
(26) Bossart-Whitaker, P.; Carson, M.; Babu, Y. S.; Smith, C. D.; Laver,
W. G.; Air, G. M. J. Mol. Biol. 1993, 232, 1069.
(27) Burmeister, W. P.; Ruigrok, R. W.; Cusak, S. EMBO J. 1992, 11,
49.
(28) Varghese, J. N.; McKimm-Bresckhin, L.; Caldwell, J. B.; Kortt,
A.; Colman, P. M. Proteins: Struct. Funct. Genet. 1992, 14, 327.
(29) (a) Pauling, L. Chem. Eng. News 1946, 24, 1375. (b) Bartlett, P. A.
Stud. Org. Chem. 1985, 20, 439.
(21) Hayden, F. G.; Treanor, J. J.; Esinhart, J.; Eason C. U.; Hussey, E.
K. AntiViral Res. 1995, 26(3), A 300 (Abst. 140).
(22) Fromtling, R. A.; Castaner, J. Drugs Future 1996, 21 (4), 375.

Influenza Neuraminidase Inhibitors
J. Am. Chem. Soc., Vol. 119, No. 4, 1997 683
Scheme 1. Rational Design of Carbocyclic Transition-State
Scheme 2a
Analogues
a Reagents: (a) MeOCH2Cl, DIPEA, CH2Cl2; (b) NaN3, NH4Cl,
MeOH/H2O; (c) MeSO2Cl, Et3N, CH2Cl2; (d) Ph3P, THF then Et3N/
H2O; (e) NaN3, NH4Cl, DMF; (f) AcCl, pyridine; (g) H2, Lindlar
catalyst, EtOH; (h) KOH, THF/H2O; (i) CF3CO2H, CH2Cl2.
hydroxyls of 11 to the trans C4 and C5 amino groups of 8. The
approach selected for this transformation, as outlined in Scheme
2, relied on the conversion of 11 to aziridine 16 followed by
azide ion attack in a regio- and stereospecific manner as a key
reaction. The synthesis began with the preparation of epoxide
12 from 11 as described in the literature.32 Nucleophilic ring
opening of MOM-protected epoxide 13 with sodium azide in
the presence of ammonium chloride generated azido alcohol
14 in 86% yield. The ring opening of the epoxide was both
regio- and stereospecific as depicted in 14, and this could be
attributed to the steric and electronegative inductive influence
of the MOM group in 13. Conversion of azide 14 to aziridine
16 was efficiently accomplished in 78% yield via a two-step
sequence: (1) mesylation of the hydroxyl group in 14 and (2)
reduction of the azide functionality in 15 with triphenylphos-
phine in the presence of triethylamine and water. The aziridine
ring opening of 16 with sodium azide gave 17 exclusively. This
selective ring opening was again a consequence of the favored
azide ion attack at the C5 position due to the steric and
electronegative inductive effects of the MOM group. Finally,
acetylation of 17 followed by reduction of the azide group in
18 and saponification of the methyl ester and deprotection of
the MOM group in 19 with trifluoroacetic acid provided 8.
Our initial attempts to utilize shikimic acid (11) for the
synthesis of 9 proved to be rather difficult. Alternatively, (-)-
quinic acid (21) was chosen as a starting material (Scheme 3).
The cyclohexene intermediate 22, readily available from 21 by
literature methods,33 possesses considerable structural similarity
to 9. The C5 hydroxyl of 22 was protected as the pivaloyl ester,
and mild acid hydrolysis of 23 furnished diol 24. Exposure of
24 to thionyl chloride in the presence of triethylamine generated
the stereochemistry of substituents around the ring in the design
of inhibitors with high affinity for NA.
Importance of the Double Bond Position
In the initial stages of designing carbocyclic NA inhibitors,
the olefinic isomers as shown in structures 6 and 7 were
considered as two possible transition-state analogues. The
isomer 6 is structurally closer to transition-state 5 than isomer
7. However, it was difficult to assess a priori which isomer
would be a better NA inhibitor, especially in light of the potent
NA inhibitory activity displayed by Neu5Ac2en (1) and its basic
analogues 2 and 3, in which the double bond was located in
the position corresponding to isomer 7. Recently, carbocyclic
analogues of the Neu5Ac2en system have been reported to have
some interesting NA inhibitory activity.31 The molecular
modeling of isomers 6 and 7 indicated that these two molecules
overlay well. Therefore, in order for us to assess the choice
between isomers 6 and 7, we felt that the synthesis of both
isomers was necessary in order to compare NA activity before
embarking on extensive structure-activity relationship studies.
For this comparative study, simple isomers 8 and 9 were chosen.
(30) (a) Tedrzejas, M. J.; Singh, S.; Brouillette, W. J.; Laver, G. W.;
Air, G. M.; Luo, M. Biochemistry 1995, 34, 3144. (b) Williams, M.;
Bischofberger N.; Swaminathan, S.; Kim, C. U. Bioorg. Med. Chem. Lett.
1995, 5(9), 2251.
(31) Chandler, M.; Conroy, R.; Cooper, A. W. J.; Lamont, R. B.;
Scicinski, J. J.; Smart, J. E.; Storer, R.; Weir, N. G.; Wilson, R. D.; Wyatt,
P. G. J. Chem. Soc., Perkin Trans. 1 1995, 1189.
(32) McGowan, D. A.; Berchtold, G. A. J. Org. Chem. 1981, 46, 2381.
(33) Ulibarri, G.; Nadler, W.; Skrydtrup, T.; Audrian, H.; Chiaroni, A.;
Riche, C.; Grierson, D. S. J. Org. Chem. 1995, 60, 2753.
Synthesis and NA Activity of 8 and 9
A comparison of structural similarities revealed that (-)-
shikimic acid (11) and 8 share common structural features.
However, the conversion of 11 to 8 requires the effective
stereochemical control for transforming the trans C4 and C5

684 J. Am. Chem. Soc., Vol. 119, No. 4, 1997
Kim et al.
Scheme 3a
a Reagents: (a) (CH3)3COCl, pyridine; (b) CH3CO2H, H2O; (c) SOCl2, pyridine, CH2Cl2; (d) NaN3, DMF; (e) CH3SO2Cl, Et3N, CH2Cl2; (f)
Ph3P, THF then Et3N/H2O; (g) CH3COCl, pyridine, CH2Cl2; (h) NaN3, NH4Cl, DMF; (i) KOH, CH3OH, H2O; (j) H2, Lindlar catalyst.
cyclic sulfite 25 as a stable intermediate in 90% yield. The
sulfite ring opening of 25 with sodium azide gave azide 26 as
a single product. This complete regiospecific ring opening is
a consequence of favored azide ion attack at the allylic C3
position of 25. The â-hydroxy azide moiety in 26 was then
converted in a three-step sequence to aziridine 28, analogous
to the conversion of 14 to 16. As expected, ring opening of
acetylated aziridine 28 with sodium azide was completely
regiospecific, giving azide 29 as the only product. Saponifica-
tion of 29 followed by reduction of the azide functionality
provided 9.
The inhibitory activity of 8 and 9 was directly determined in
a NA enzymatic assay. While 8 proved to be a potent NA
inhibitor with an inhibitory concentration (IC50) of 6.3 µM, 9
did not exhibit inhibitory activity at concentrations up to 200
µM. This result demonstrated that the double bond position in
the design of carbocyclic NA inhibitors plays an important role
in NA activity. However, further structural investigation is
required in order to illustrate the binding difference(s) of 8 and
9 in the NA active site.
hydrophobic interaction would lead to new NA inhibitors with
increased lipophilicity while maintaining potent NA inhibitory
activity. This consideration is especially important for designing
orally bioavailable drugs since balancing lipophilicity and water
solubility could be as critical as the size of the molecule for its
absorption from the intestinal tract. For optimization of
hydrophobic interactions, the dimensions of the spacer would
be expected to play an important role in binding affinity to the
enzyme, as factors such as length, geometry, and conformational
mobility. On the basis of this premise, we undertook a
systematic modification of the R′ portion in 6 with various
aliphatic side chains. Previous studies5b suggested that the
carboxylate of Neu5Ac2en (1) and its analogues form strong
ionic interactions with three guanidino groups of Arg118, -292,
and -371. The acetamido moiety at the C5 of Neu5Ac2en and
its analogues interacts with Arg152 and Glu227, and the methyl
group fits nicely into a hydrophobic pocket formed by Trp180,
Ile224, and Arg226. The acetamido group was also demon-
strated to be optimal for antiviral activity.34 The basic amino
and guanidino groups at the C4 position of the Neu5Ac2en
system were optimal for NA activity as described above.11b,18
Therefore, in the carbocyclic structure 6, the C1 carboxylate,
C4 acetamido, and C5 amino groups were kept constant, while
the C3 aliphatic group was optimized for antiviral activity.
Replacement of the Glycerol Moiety with Lipophilic Side
Chains
Crystallographic studies of Neu5Ac and its analogues bound
to NA appear to indicate that the C7 hydroxyl of the glycerol
side chain does not interact with any amino acids of the NA
active site.5b This suggested that the C7 hydroxyl could be
eliminated from the carbocyclic analogues without losing
binding affinity to NA. Furthermore, it was also realized that,
in the transition-state intermediate 5, the oxonium double bond
is highly polarized and electron deficient. Taking into account
these features, in the carbocyclic structure, the CHOH group at
the C7 position of the glycerol side chain in the Neu5Ac system
was replaced with the oxygen atom as shown in structure 6.
This C3 oxygen atom would reduce the electron density of the
double bond via the σ bond electronegative inductive effect. In
addition to these rationales, our decision to have the C3 oxygen
atom was based on the synthetic practicality of modifying R′
groups to optimize NA inhibitory activity and pharmacological
properties.
Synthesis of Lipophilic Analogues 6
For structure-activity relationship studies of carbocyclic
analogues 6, a general and efficient route to introduce various
alkyl ethers at the C3 position was required. The approach we
selected, as outlined in Scheme 4, relied on the aziridine opening
of 31 with alcohols, which should in principle be highly
regioselective due to the preferred nucleophilic attack at the C3
allylic position. The requisite aziridine 31 was derived from
the trans amino alcohol 30 by the two-step, one-pot process:
(1) selective protection of the amino functionality with trityl
chloride and (2) mesylation of the hydroxyl in the presence of
triethylamine. Under these conditions, the mesylate intermediate
was converted to aziridine 31 in 86% overall yield. Of a number
of protecting groups tried on amine 30, best results were
obtained by using the trityl group. Treatment of 31 with various
alcohols in the presence of 1.5 equiv of BF3‚Et2O followed by
acetylation of the crude product provided the ethers 32 in 55-
80% yield. As expected, no other regio- and stereoisomers were
X-ray crystallographic structures of Neu5Ac and its analogues
complexed with NA show that the two terminal hydroxyls of
the glycerol side chain form a bidentate interaction with
Glu276.25,27 However, it is also noted that the C8 of the glycerol
chain makes hydrophobic contacts with the hydrocarbon chain
of Arg224.5b Therefore, we hoped that the optimization of this
(34) Smith, P. W.; Starkey, I. D.; Howes, P. D.; Sollis S. L.; Keeling, S.
P.; Cherry, P. C.; von Itzstein, M.; Wu, W. Y.; Jin, B. Eur. J. Med. Chem.
1996, 31, 143.

Influenza Neuraminidase Inhibitors
J. Am. Chem. Soc., Vol. 119, No. 4, 1997 685
Scheme 4a
a Reagents: (a) CH3OH, HCl; (b) TrCl, Et3N, CH2Cl2; (c) CH3SO2Cl, Et3N, CH2Cl2; (d) BF3‚Et2O; ROH; (e) Ac2O, DMAP, pyridine; (f) Ph3P,
THF then Et3N/H2O; (g) KOH, THF, H2O.
Scheme 5a
Table 1. Influenza Neuraminidase Inhibition and Plaque
Reduction by Carbocylic Analogues
enzymea
IC50 (nM)
plaqueb
EC50 (nM)
R
compd
H
CH3
CH3CH2
CH3CH2CH2
CH3CH2CH2CH2
(CH3)2CHCH2
CH3CH2(CH3)CH*
8
6a
6b
6c
6d
6e
6f
6300
3700
2000
180
300
200
10
NDc
ND
ND
ND
ND
80
a Reagents: (a) TsCl, DMAP, pyridine; (b) SO2Cl2, pyridine; (c)
p-TsOH, CH3OH; (d) DBU, THF.
(R)-isomer
6g
9
135
(S)-isomer
(CH3CH2)2CH
(CH3CH2CH2)2CH
6h
6i
2
1
16
150
1
16
observed in this aziridine ring-opening reaction. Finally,
reduction of the azide functionality and saponification of the
ester group in 32 gave 6. Thus the convergent and efficient
synthesis of various ether analogues 6 has been achieved from
intermediate 17 which is readily available from shikimic acid
as shown in Scheme 2.
ND
2500
15
3
a NA. b H1N1, A/ws. c ND ) not determined.
in quantitative yield by treatment of 35 with 1,8-diazabicyclo-
[5.4.0]undec-7-ene in tetrahydrofuran. This completed the
synthesis of epoxide 12 from quinic acid in good overall yield
without column chromatography.
Although shikimic acid as a chiral starting material is
convenient because of easy conversion to the key epoxy
intermediate 12, its high cost and low availability in large
quantities made it impractical for scale-up of 12. This became
a formidable challenge when multi-kilograms of material were
needed for evaluation of clinical candidates for further drug
development. (-)-Quinic acid (21), which was used as a
starting material for the synthesis of 9 (Scheme 3), appeared to
be an ideal starting material because of its low cost and
commercial abundance. Conversion of quinic acid to the
acetonide intermediate 33 (Scheme 5) proceeded in high yield
as reported.35 At this point we focused on the selective
dehydration of the C2 hydroxyl in 34. Although a similar
dehydration method has been reported in the literature,33 it was
difficult to reproduce on a large scale in our hands.
Thus, tosylate 34 was reacted with sulfuryl chloride in
pyridine, followed by acetonide cleavage in refluxing methanol
in the presence of p-toluenesulfonic acid. The desired diol 35
was then directly crystallized out of the reaction mixture in 54%
overall yield. This procedure was amenable to a several hundred
gram scale. In the dehydration and acetonide removal sequence
for the conversion of 34 to 35, the other olefinic regioisomer
of 35 was aromatized under the reaction conditions and easily
separated by crystallization. Finally, epoxide 12 was produced
Structure-Activity Relationships of Carbocyclic
Analogues 6
Two methods were used to evaluate the carbocyclic analogues
as inhibitors of NA (Table 1). The intrinsic activity of each
compound was assessed by measuring the inhibition of enzy-
matic activity. Compounds that exhibited potent NA inhibitory
activity were further evaluated in cell culture by a plaque
reduction assay using an influenza A (H1N1) strain. As shown
in Table 1, the length, size of branching, and geometry of the
alkyl groups in 6 profoundly influence the NA inhibitory
activity. In a series of linear alkyl analogues (6a-d), steady
increases in the enzyme inhibitory activity were observed up
to the n-propyl analogue 6c. The over 20-fold increase in the
NA inhibitory activity for 6c compared to the methyl counterpart
6a implicated a significant hydrophobic interaction of the
n-propyl group with amino acids in the active site. Branching
at the â-carbon of the n-propyl group (compound 6e) resulted
in no enhancement of NA inhibitory activity compared to that
of 6c. In contrast, when the methyl group was added at the
R-position of the n-propyl group (compounds 6f and 6g), NA
inhibitory activity increased almost 20-fold. This improved
(35) Shing, T. K.; Tang, Y. Tetrahedron 1990, 46, 6575.

686 J. Am. Chem. Soc., Vol. 119, No. 4, 1997
Kim et al.
Figure 1. X-ray structure of 6h bound to influenza neuraminidase.
potency undoubtedly arose from additional hydrophobic interac-
tions in the NA active site generated by the R-methyl group.
Remarkably, the absolute stereochemistry of the R-methyl group
(6f and 6g) did not influence the NA inhibitory activity as they
both exhibited almost equal potency. This result led us to
speculate that the methyl and ethyl groups in the sec-butyl
analogues 6f and 6g contribute almost the same binding energy
by an equal degree of interaction with amino acids in the active
site. This hypothesis was further supported by the NA activity
of 3-pentyl analogue 6h, which possesses two identical branched
ethyl groups and shows the NA inhibitory activity increase of
almost 10-fold that of either isomers 6f and 6g. Further
extension of the two ethyl groups to the 4-heptyl analogue 6i
resulted in a decrease of NA inhibitory activity, suggesting that
the 3-pentyl group in 6h might provide the optimum hydro-
phobic contact with the NA active site. Interestingly, the
enzymatic and plaque reduction activities of 6h were comparable
to those of 3, which exhibited over 100-fold increase in activity
compared to that of amino analogue 2. Replacement of the
amino group in 6h with the guanidino moiety resulted in a
significant increase in its enzymatic and cell culture activity
compared to 6h.36 Thus, carbocyclic NA inhibitors represented
by 6h and its guanidino analogue are more potent NA inhibitors
than any inhibitors reported previously, including compounds
2 and 3.
by Trp178 and Ile222 with the oxygen atom of the amide
interacting with Arg152. These binding features were not
significantly different from that found in the inhibitor-NA
complex of Neu5Ac2en and other sialic acid based inhibitors.25-27
The 3-pentyloxy side chain at the C4 position was situated
against a large hydrophobic surface created by the hydrocarbon
chains of Glu276, Ala246, Arg224, and Ile222. Contrary to
this, the glycerol side chain of sialic acid based inhibitors interact
with the carboxylate of Glu276 through two terminal hydroxyls
in a bidentate hydrogen bond donor-acceptor mode. In order
to accommodate the large 3-pentyl group of 6h, the carboxylate
of Glu276 is forced to orientate outward from the hydrophobic
pocket. The hydrophobic interaction observed for the 3-pen-
tyloxy group in the NA active site illustrates well the other
structure-activity relationships observed in the different alkyl
chains in Table 1. As discussed above, the remarkable degree
of correlation between the aliphatic side chains and NA
inhibitory activity result from the different degree of hydro-
phobic interactions due to different length, geometry, and
conformational rigidity of the alkyl chains. The incremental
entropy gain resulting from increased hydrophobic interaction
is believed to reflect the NA inhibitory activity of these
compounds. Further detailed structural studies in this series
are being investigated.
Conclusion
Carbocyclic analogues of sialic acid described in this study
represent a rational drug design of enzyme inhibitors based on
a transition-state mechanism. Similarity replacement of the
oxonium transition state arising from the sialic acid cleavage
by NA with a cyclohexene scaffold led to a new series of NA
inhibitors. It was clearly demonstrated that, in the design of a
transition-state analogue, the closer the structure is to the
transitition state, the more potent the enzyme inhibitor, as
indicated by the striking difference in NA inhibition of 8 vs 9.
The X-ray crystallographic study of 6h bound to NA revealed
the existence of a large hydrophobic pocket in the region
corresponding to the glycerol subsite of sialic acid. The finding
of this new hydrophobic pocket led us to a new series of potent
carbocyclic NA inhibitors 6, in which the clear structure-
X-ray Structure Analysis of 6h
Information on protein-ligand interactions generated by
X-ray crystallography plays a very important role in structure-
based enzyme inhibitor design. In the course of studying
structure-activity relationships of carbocyclic NA inhibitors,
the crystal structure of 6h bound to NA was investigated. As
shown in Figure 1, the carboxylate of inhibitor 6h was held
strongly by three arginine residues (Arg292, Arg 371, Arg 118).
The C3 amino group forms strong charge-charge type hydrogen
bond interactions with Glu119 and Asp151. The methyl group
of the C4 acetamide occupies a hydrophobic pocket generated
(36) Data not shown. A full account of structure-activity relationships
of carbocyclic analogues related to 6 will be published separately.

Influenza Neuraminidase Inhibitors
J. Am. Chem. Soc., Vol. 119, No. 4, 1997 687
ethyl acetate. The combined organic extracts were dried, filtered, and
evaporated to afford crude mesylate 15 (8.56 g, 99%), which was of
suitable purity to use directly in the next step. An analytical sample
was prepared by chromatography eluting with ethyl acetate/hexane (1:
1) and isolated as a white oil: 1H NMR (CDCl3) δ 6.85 (m, 1H), 4.82
(d, 1H, J ) 6.9), 4.73 (d, 1H, J ) 6.9), 4.67 (dd, 1H, J ) 3.9, 9.0),
4.53 (br t, 1H, J ) 4.2), 3.78 (s, 3H), 3.41 (s, 3H), 3.15 (s, 3H), 2.98
(dd, 1H, J ) 6.0, 18.6), 2.37 (m, 1H); 13C NMR (CDCl3) δ 165.6,
134.3, 129.6, 96.5, 78.4, 69.6, 55.8, 55.7, 52.1, 38.2, 29.1. Anal. Calcd
for C11H17N3O7S: C, 39.40; H, 5.11; N, 12.53. Found: C, 39.55; H,
5.17; N, 12.75.
Aziridine (16). To a solution of 15 (8.56 g, 25 mmol) in THF (150
mL) at 0 °C was added triphenylphosphine (8.2 g, 31 mmol), initially
adding a third of the amount while cooling and then after removing
the ice bath adding the remaining amount over a period of 15 min.
The reaction mixture was stirred at room temperature for 3 h, during
which time a white precipitate formed. To this suspension was added
triethylamine (5.2 mL, 37.5 mmol) and water (10 mL), and the mixture
was stirred at room temperature for 12 h. The reaction mixture was
concentrated to remove THF, and the residue was partitioned between
CH2Cl2 (200 mL) and brine (200 mL). The aqueous phase was
extracted with CH2Cl2, and the combined organic extracts were dried
(Na2SO4), filtered, and evaporated to give a crude product which was
purified by chromatography, eluting with ethyl acetate/methanol (9:1)
to give 16 (4.18 g, 78%) as an oil containing 5% triphenylphosphine
oxide: 1H NMR (CDCl3) δ 6.81 (m, 1H), 4.78 (s, 2H), 4.54 (m, 1H),
3.73 (s, 3H), 3.41 (s, 3H), 2.87 (dd, 1H), 2.64 (br s, 1H), 2.56-2.47
(m, 2H); 13C NMR (CDCl3) δ 166.9, 132.5, 128.0, 95.9, 69.5, 55.2,
51.6, 31.1, 27.7, 24.1.
Methyl (3R,4R,5S)-4-Amino-5-azido-3-(methoxymethoxy)-1-cy-
clohexene-1-carboxylate (17). To a solution of 16 (3.2 g, 15 mmol)
in DMF (30 mL) was added sodium azide (4.9 g, 75 mmol) and
ammonium chloride (1.6 g, 30 mmol), and the mixture was heated at
65-70 °C for 21 h. The reaction mixture was cooled to room
temperature, diluted with ethyl acetate (100 mL), and filtered. The
filtrate was evaporated, and the residue was partitioned between diethyl
ether (100 mL) and brine (100 mL). The organic phase was washed
again with brine, dried, filtered, and evaporated. The aqueous washings
were extracted with ethyl acetate and treated in the same manner as
described above. The crude product was chromatographed with 5%
methanol in CH2Cl2 to give 17 (2.95 g) as an oil which contained a
small amount of triphenylphosphine oxide from the previous step: 1H
NMR (CDCl3) δ 6.82 (t, 1H, J ) 2.3), 4.81 (d, 1H, J ) 7.2), 4.77 (d,
1H, J ) 6.9), 4.09-4.04 (m, 1H), 3.76 (s, 3H), 3.47 and 3.44 (m
overlapping s, 4H), 2.94-2.86 (m, 2H), 2.36-2.24 (m, 1H); 13C NMR
(CDCl3) δ 165.9, 137.3, 128.2, 96.5, 79.3, 61.5, 55.7, 55.6, 51.9, 29.5.
Amine 17 was characterized as the N-tert-butylcarbamate derivative.
To a solution of 17 in ethyl acetate was added an equivalent volume
of saturated sodium bicarbonate, followed by an excess of di-tert-butyl
dicarbonate. The mixture was vigorously stirred at room temperature
for 18 h. The organic phase was separated, washed with brine, dried,
filtered, and evaporated. The residue was chromatographed with
hexane/ethyl acetate (2:1) to afford the N-tert-butylcarbamate derivative
as a white solid: mp 78-80 °C; 1H NMR (CDCl3) δ 6.79 (t, 1H, J )
2.4), 4.90 (br s, 1H), 4.70 (d, 1H, J ) 7.0), 4.55 (br s, 1H), 4.01 (br s,
1H), 3.77 (s, 3H), 3.42 (m, 4H), 2.87 (dd, 1H, J ) 5.2, 17.8), 2.31-
2.22 (m, 1H). Anal. Calcd for C15H22N4O6: C, 50.84; H, 6.26; N,
15.81. Found: C, 50.73; H, 6.60; N, 15.58.
activity relationship was observed among analogues with various
alkyl chains (R′ in 6). This result demonstrated that, in the
design of enzyme inhibitors, consideration of hydrophobic
interactions in the enzyme active site might lead to a new
generation of structurally unique molecules. Finally, the ethyl
ester of 6h (designated as GS4104) exhibited good oral
bioavailability in several animals (mice, rats, and dogs) and
demonstrated oral efficacy in the mouse and ferret influenza
model. On the basis of potent in vitro/in vivo activity and very
favorable pharmacological properties, GS4104 has been selected
as a clinical candidate for the oral treatment and prophylaxis of
influenza infection.
Experimental Section
General. All reactions were conducted under a dry atmosphere of
argon unless otherwise noted. All reaction solvents were anhydrous
grade obtained from Aldrich Chemical Co. Melting points were
obtained on a Thomas-Hoover melting point apparatus and are
uncorrected. Elemental analyses were performed by Robertson Microlit
Laboratories, Inc., Madison, NJ. High-resolution mass spectra were
performed by the Mass Spectrometry Lab, University of California,
Berkeley. Column chromatography was carried out using 230-400
mesh silica gel. NMR spectra were recorded at 300 MHz unless
otherwise indicated; J values are reported in hertz. Reaction solutions
were dried over MgSO4 unless otherwise indicated.
Methyl (3R,4R,5S)-4,5-Epoxy-3-(methoxymethoxy)-1-cyclohex-
ene-1-carboxylate (13). To a solution of 12 (4.0 g, 23.5 mmol) in
CH2Cl2 (100 mL) was added N,N′-diisopropylethylamine (12.3 mL,
70.5 mmol) followed by chloromethyl methyl ether (3.6 mL, 47 mmol).
The solution was refluxed for 3.5 h, and the solvent was evaporated.
The residue was partitioned between ethyl acetate (200 mL) and water
(200 mL). The aqueous phase was extracted with ethyl acetate (100
mL), and the combined organic extracts were washed with brine (100
mL), dried, filtered, and evaporated to give 4.9 g (97%) of a solid
residue which was of suitable purity to use directly in the next step.
An analytical sample was recrystallized (ether/hexane) to give 12 as
1
colorless plates: mp 64-66 °C; H NMR (CDCl3) δ 6.73 (m, 1H),
4.87 (s, 2H), 4.59 (m, 1H), 3.75 (s, 3H), 3.57 (m, 1H), 3.48 (m, 4H),
3.07 (dd, 1H, J ) 1.2, 19.8), 2.47 (dq, 1H, J ) 2.7, 19.5). Anal. Calcd
for C10H14O5: C, 56.07; H, 6.59. Found: C, 56.15; H, 6.61.
Methyl (3R,4S,5R)-5-Azido-4-hydroxy-3-(methoxymethoxy)-1-
cyclohexene-1-carboxylate (14). To a solution of 13 (4.9 g, 22.9
mmol) in methanol/water (8:1, 175 mL) were added sodium azide (7.44
g, 114.5 mmol) and ammonium chloride (2.69 g, 50.4 mmol), and the
mixture was refluxed for 15 h. The reaction mixture was diluted with
water (75 mL) to dissolve precipitated salts, and the solution was
concentrated to remove methanol. The resulting aqueous phase
containing an oily residue was diluted to a volume of 200 mL with
water and was extracted with ethyl acetate (3 × 100 mL). The
combined organic extracts were washed with brine (100 mL), dried,
filtered, and evaporated. The crude was chromatographed (hexane/
ethyl acetate, 1:1) to give 14 (5.09 g, 86%) as a pale yellow oil which
was of sufficient purity to use in the next step without further
purification. An analytical sample was prepared by chromatography
(2.5% methanol in CH2Cl2): 1H NMR (CDCl3) δ 6.86 (m, 1H), 4.79
(s, 2H), 4.31 (br t, 1H, J ) 4.2), 3.90-3.75 (m, 5H), 3.43 (s, 3H), 2.92
(d, 1H, J ) 6.6), 2.87 (dd, 1H, J ) 5.4, 18.6), 2.21-2.30 (m, 1H); 13
C
Methyl (3R,4R,5S)-4-Acetamido-5-azido-3-(methoxymethoxy)-1-
cyclohexene-1-carboxylate (18). A solution of 17 (in CH2Cl2 (15 mL)
was treated with pyridine (4 mL) and acetyl chloride (150 µL).
Aqueous workup followed by chromatography of the residue gave 18
(350 mg, 76%) as a pale oil: 1H NMR (CDCl3) δ 6.78 (s, 1H), 6.39
(br d, 1H, J ) 7.8 Hz), 4.72 (d, 1H, J ) 6.9 Hz), 4.66 (d, 1H, J ) 6.9
Hz), 4.53 (br d, 1H, J ) 8.4 Hz), 4.00-3.90 (m, 1H), 3.80-3.65 (m,
1H), 3.75 (s, 3H), 3.37 (s, 3H), 2.85 (dd, 1H, J ) 5.4, 17.7 Hz), 2.35-
2.20 (m, 1H), 2.04 (s, 3H).
(3R,4R,5S)-4-Acetamido-5-amino-3-(methoxymethoxy)-1-cyclo-
hexene-1-carboxylic Acid (19). A solution of 18 (39 mg, 0.13 mmol)
in ethanol was stirred with Lindlar catalyst (39 mg) under 1 atm of
hydrogen. The reaction mixture was filtered through a Celite pad,
which was washed with hot ethanol. The filtrate was evaporated to
NMR (CDCl3) δ 166.2, 134.7, 130.3, 96.8, 72.5, 70.5, 58.6, 55.8, 52.1,
28.7. Anal. Calcd for C10H15N3O5: C, 46.69; H, 5.88; N, 16.33.
Found: C, 46.68; H, 5.89; N, 15.89.
Methyl (3R,4S,5R)-5-Azido-4-[(methylsulfonyl)oxy]-3-(meth-
oxymethoxy)-1-cyclohexene-1-carboxylate (15). To a solution of 14
(6.47 g, 25.2 mmol) in CH2Cl2 (100 mL) at 0 °C was added
triethylamine (4.4 mL, 31.5 mmol) followed by the addition of
methanesulfonyl chloride (2.14 mL, 27.7 mmol). The reaction was
stirred at 0 °C for 45 min and warmed to room temperature with stirring
for 15 min. Solvent was evaporated, and the residue was partitioned
between ethyl acetate (200 mL) and water (100 mL). The organic phase
was washed with water (100 mL), saturated sodium bicarbonate (100
mL), and brine (100 mL). The aqueous washes were extracted with

688 J. Am. Chem. Soc., Vol. 119, No. 4, 1997
Kim et al.
give a pale foam (33 mg), which was dissolved in THF (1 mL) and
was treated with aqueous 0.48 M potassium hydroxide (380 µL) for 1
h. The reaction was acidified to pH ) 4.0 with Amberlite IR-120.
The resin was filtered and washed with water, and the filtrate was
evaporated to give a pale solid which was purified by C18 column
chromatography eluting with water. The fractions containing the
product were pooled and lyophilzed to give 19 (20 mg, 60%) as a white
powder: 1H NMR (D2O) δ 6.65 (s, 1H), 4.87 (d, 1H, J ) 7.5 Hz),
4.76 (d, 1H, J ) 7.5 Hz), 4.47 (br d, 1H, J ) 8.7 Hz), 4.16 (dd, 1H,
J ) 11.4, 11.4 Hz), 3.70-3.55 (m, 1H), 3.43 (s, 3H), 2.95 (dd, 1H, J
) 5.7, 17.4 Hz), 2.60-2.45 (m, 1H), 2.11 (s, 3H).
(3R,4R,5S)-4-Acetamido-5-amino-3-hydroxy-1-cyclohexene-1-car-
boxylic acid (8). To 19 (4 mg, 0.015 mmol) was added 40%
trifluoroacetic acid in CH2Cl2 (1 mL, cooled to 0 °C prior to addition).
After being stirred at room temperature for 1.5 h, the reaction mixture
was concentrated to give a white foam. Coevaporation from water
several times followed by lyophilization gave trifluoroacetate salt 8
(4.9 mg, 100%) as a white solid: 1H NMR (D2O) δ 6.85 (m, 1H), 4.45
(m, 1H), 4.05 (dd, 1H, J ) 11.4, 11.4 Hz), 3.65-3.55 (m, 1H), 3.00-
2.90 (m, 1H), 2.60-2.45 (m, 1H), 2.09 (s, 3H); HRMS (FAB) calcd
for C9H15N2O4 (MH+) 215.1032, found 215.1027.
N-Acetylaziridine (28). To a solution of 27 (0.380 g, 0.94 mmol)
in THF (19 mL) was added triphenylphosphine (0.271 g, 1.04 mmol),
and the solution was stirred at room temperature for 2 h. To this
solution were added water (1.9 mL) and triethylamine (0.39 mL, 2.82
mmol), and the mixture was stirred for 14 h. The solvents were
evaporated, the crude residue was dissolved in CH2Cl2 (20 mL) and
cooled to 0 °C, and pyridine (0.68 mL, 8.4 mmol) was added followed
by the dropwise addition of acetyl chloride (0.30 mL, 4.2 mmol). The
mixture was stirred at 0 °C for 5 min and diluted with ethyl acetate.
The organic phase was washed with water and brine and dried. The
solvent was evaporated, and the residue was purified by chromatography
eluting with hexane/ethyl acetate (3:1) to give 28 (0.205 g, 83%) as an
oil: 1H NMR (CDCl3) δ 7.19 (m, 1H), 5.58 (m, 1H), 3.77 (s, 3H),
3.14 (m, 2H), 2.85 (dd, J ) 7.0, 1.6 Hz, 1H), 2.34 (m, 1H), 2.16 (s,
3H), 1.14 (s, 9H); HRMS (EI) calcd for C15H21NO5 (M+) 295.1420,
found 295.1420.
Methyl (3S,4S,5R)-4-Acetamido-3-azido-5-(pivaloyloxy)-1-cyclo-
hexene-1-carboxylate (29). A mixture of 28 (0.200 g, 0.68 mmol),
sodium azide (0.221 g, 3.4 mmol), and ammonium chloride (0.146 g,
2.7 mmol) in DMF (10 mL) was stirred at room temperature for 14 h.
The reaction was diluted with ethyl acetate, washed with water and
brine, and dried. The solvent was evaporated, and the residue was
purified by chromatography eluting with hexane/ethyl acetate (2:1) to
give a mixture of products which was dissolved in acetic anhydride (2
mL) and stirred for 2 h. Excess acetic anhydride was evaporated to
afford 29 (149 mg, 65%) as an oil: 1H NMR (CDCl3) δ 6.76 (m, 1H),
5.53 (d, 1H, J ) 8.5), 5.05 (m, 1H), 4.31 (m, 1H), 4.08 (m, 1H), 3.79
(s, 3H), 2.91 (m, 1H), 2.51 (m, 1H), 1.99 (s, 1H), 1.20 (s, 9H).
(3S,4S,5R)-4-Acetamido-3-amino-5-hydroxy-1-cyclohexene-1-car-
boxylic Acid (9). A solution of potassium hydroxide (123 mg, 2.2
mmol) in methanol/water (1:1, 4.4 mL) was added to 29 (149 mg, 0.44
mmol), and the mixture was stirred at room temperature for 3 h. The
mixture was cooled to 0 °C and acidified with Amberlite IR-120 to
pH ) 3.5. The mixture was filtered, washed with methanol, and
evaporated to give a solid residue (73 mg). The residue (8 mg) was
dissolved in ethanol (2 mL), treated with Lindlar catalyst (15 mg), and
stirred under 1 atm of hydrogen for 16 h. The reaction was filtered
through Celite, and the catalyst was washed with hot methanol/water
(1:1). The filtrate was evaporated, and the reside was purified by C8
column chromatography eluting with water to give 9 (6 mg) as a white
solid: 1H NMR (D2O) δ 6.28 (m, 1H), 4.06-3.85 (m, 3H), 2.83 (dd,
1H, J )17.7, 5.4), 2.35 (m, 1H), 2.06 (s, 3H); HRMS (FAB) calcd for
C9H15N2O4 (MH+) 215.1032, found 215.1025.
Methyl (3R,4R,5S)-4-Amino-5-azido-3-hydroxy-1-cyclohexene-1-
carboxylate Hydrochloride (30). A solution of 17 (2.59 g, 10.2 mmol)
in 5% hydrogen chloride in methanol (30 mL) was stirred for 3 h at
room temperature, and additional 5% hydrogen chloride in methanol
(10 mL) was added. After the mixture was stirred for 1 h, the solvent
was evaporated to give 30 (2.52 g, 99%) as a tan solid after drying
under high vacuum. An analytical sample was recrystallized from
methanol/ether: 1H NMR (CD3OD) δ 6.73 (t, 1H, J ) 2.1), 4.35 (m,
1H), 3.93 (s, 3H), 3.77 (s, 3H), 3.12-2.98 (m, 2H), 2.50-2.39 (m,
1H). Anal. Calcd for C8H13ClN4O3: C, 38.64; H, 5.27; N, 22.53.
Found: C, 38.72; H, 5.21; N, 22.45.
N-Tritylaziridine (31). To a suspension of 30 (2.52 g, 10.1 mmol)
in CH2Cl2 (50 mL) at 0 °C was added triethylamine (3.55 mL, 25.5
mmol) followed by the addition of solid trityl chloride (5.55 g, 12.8
mmol) in one portion. The mixture was stirred at 0 °C for 1 h. After
being stirred for 2 h at room temperature, the reaction mixture was
cooled to 0 °C, and triethylamine (3.6 mL, 25.5 mmol) and methane-
sulfonyl chloride (0.97 mL, 12.5 mmol) were added. The resulting
mixture was stirred for 1 h at 0 °C and for 22 h at room temperature.
The solvent was evaporated, and the residue was partitioned between
ether (200 mL) and water (200 mL). The organic phase was washed
with water (200 mL), and the combined aqueous phases were extracted
with ether (200 mL). The combined organic extracts were washed with
water and brine and dried. The crude product was chromatographed
eluting with hexane/CH2Cl2 (1:1) to give 31 (3.84 g, 86%) as a white
foam: 1H NMR (CDCl3) δ 7.4-7.23 (m, 16H), 4.32 (m, 1H), 3.81 (s,
3H), 3.06 (dt, 1H, J ) 1.8, 17.1), 2.94-2.86 (m, 1H), 2.12 (m, 1H),
1.85 (t, 1H, J ) 5.0); HRMS (FAB) calcd for C27H25N4O2 (MH+)
437.1977, found 437.1976.
Methyl (3S,4R,5R)-3,4-O-Isopropylidine-3,4-dihydroxy-5-(piv-
aloyloxy)-1-cyclohexene-1-carboxylate (23). To a solution of 22 (10.9
mmol) in CH2Cl2 (60 mL) at 0 °C was added pyridine (4.4 mL, 54.5
mmol) followed by addition of trimethylacetyl chloride (2.7 mL, 21.8
mmol). The mixture was warmed to room temperature and stirred for
14 h. The mixture was diluted with CH2Cl2, and the organic phase
was washed with water and brine and dried. The solvent was
evaporated, and the residue was purified by chromatography eluting
with hexane/ethyl acetate (9:1) to give 23 (2.32 g, 68%) as an oil: 1H
NMR (CDCl3) δ 6.72 (m, 1H), 5.04 (m, 1H), 4.76 (m, 1H), 4.40 (m,
1H), 3.77 (s, 3H), 2.72-2.49 (m, 2H), 1.37 (s, 3H), 1.35 (s, 3H), 1.23
(s, 9H).
Methyl (3S,4R,5R)-3,4-Dihydroxy-5-(pivaloyloxy)-1-cyclohexene-
1-carboxylate (24). A solution of 23 (2.32 g, 2.3 mmol) in acetone/
water (1:1, 100 mL) was heated at 55 °C for 16 h in the presence of
p-toluenesulfonic acid (10 mg). The solvents were evaporated, and
the residue was coevaporated sequentially from water and toluene to
give 24 as an oil which was used without further purification: 1H NMR
(CDCl3) δ 6.83 (m, 1H), 5.06 (m, 1H), 4.42 (m, 1H), 4.09 (m, 1H),
3.77 (s, 3H), 2.68-2.41 (m, 2H), 1.22 (s, 9H).
Cyclic Sulfite 25. To a solution of 24 (0.41 g, 1.5 mmol) in THF
(8 mL) at 0 °C was added triethylamine (0.83 mL, 6.0 mmol), followed
by dropwise addition of thionyl chloride (0.33 mL, 4.5 mmol). The
mixture was warmed to room temperature and stirred for 3 h. The
mixture was diluted with chloroform and was washed with water and
brine and dried. The solvents were evaporated, and the residue was
purified by chromatography eluting with hexane/ethyl acetate (5:1) to
give 25 (0.430 g, 90%) as an oil: 1H NMR (CDCl3) δ 6.89-6.85 (m,
1 H), 5.48-4.84 (m, 3 H), 3.80, 3.78 (s, 3 H), 2.90-2.60 (m, 2 H),
1.25, 1.19 (s, 9 H).
Methyl (3R,4S,5R)-3-Azido-4-hydroxy-5-(pivaloyloxy)-1-cyclo-
hexene-1-carboxylate (26). The mixture of 25 (0.40 g, 1.3 mmol)
and sodium azide (0.41 g, 6.29 mmol) in DMF (10 mL) was stirred at
room temperature for 20 h. The reaction mixture was then diluted
with ethyl acetate, washed with saturated ammonium chloride, water,
and brine, and dried. The solvents were evaporated to give 26 (0.338
g, 90%) as an oil: 1H NMR (CDCl3) δ 6.78 (m, 1H), 5.32 (m, 1H),
4.20 (m, 1H), 3.89 (m, 1H), 3.78 (s, 3H), 3.00-2.60 (m, 2H), 1.21 (s,
9H).
Methyl (3R,4S,5R)-3-Azido-4-[(methylsulfonyl)oxy]-5-(pivaloy-
loxy)-1-cyclohexene-1-carboxylate (27). To a solution of 26 (0.338
g, 1.1 mmol) in CH2Cl2 (11 mL) at 0 °C was added triethylamine (0.4
mL, 2.9 mmol) followed by dropwise addition of methanesulfonyl
chloride (0.18 mL, 2.3 mmol). The mixture was stirred at 0 °C for 30
min and was diluted with CH2Cl2. The organic phase was washed with
water and brine, dried, and evaporated. The residue was purified by
chromatography eluting with hexane/ethyl acetate (3:1) to give 27 (0.38
g, 82%) as an oil: 1H NMR (CDCl3) δ 6.82 (m, 1H), 5.44 (m, 1H),
4.76 (dd, J ) 7.3, 1.4 Hz, 1H), 4.48 (m, 1H), 3.80 (s, 3H), 3.11 (s,
3H), 2.82-2.61 (m, 2H), 1.21 (s, 9H).

Influenza Neuraminidase Inhibitors
J. Am. Chem. Soc., Vol. 119, No. 4, 1997 689
Representative Procedure: Alcoholysis of 31. Synthesis of
Methyl (3R,4R,5S)-4-Acetamido-5-azido-3-(1-ethylpropoxy)-1-cy-
clohexene-1-carboxylate (32h). To a solution of 31 (15 g, 34 mmol)
in 3-pentanol (230 mL) was added BF3‚Et2O (6.27 mL, 51 mmol). The
solution was heated at 70-75 °C for 2 h and then evaporated to give
a residue which was dissolved in dry pyridine (2.0 mL) and treated
with acetic anhydride (16 mL, 170 mmol) and (dimethylamino)pyridine
(200 mg, 1.6 mmol). The reaction mixture was stirred at room
temperature for 18 h and evaporated, and the residue was partitioned
between ethyl acetate and 1 M HCl. The organic phase was washed
with saturated sodium bicarbonate and brine and dried. The solvent
was evaporated, and the residue was chromatographed eluting with
hexane/ethyl acetate (1:1) to give 32h (7.66 g, 69%) as a white solid.
An analytical sample was recrystallized from hexane/ethyl acetate to
(s, 3H), 3.56-3.48 (m, 1H), 3.31-3.25 (m, 1H), 2.86 (dd, 1H, J )
5.7, 17.5 Hz), 2.28-2.17 (m, 1H), 2.04 (s, 3H), 1.61-1.36 (m, 2H),
1.17 (d, 3H, J ) 6.2 Hz), 0.91 (t, 3H, J ) 7.4 Hz). Anal. Calcd for
C14H22N4O4: C, 54.18; H, 7.15; N, 18.05. Found: C, 53.94; H, 7.10;
N, 18.32.
Methyl (3R,4R,5S)-4-Acetamido-5-azido-3-(1-propylbutoxy)-1-
cyclohexene-1-carboxylate (32i): isolated as a solid; 1H NMR (CDCl3,
500 MHz) δ 6.78 (t, 1H, J ) 2.0), 6.05 (d, 1H, J )7.0), 4.55 (m, 1H),
4.29-4.23 (m, 1H), 3.76 (s, 3H), 3.46-3.42 (m, 1H), 3.36-3.30 (m
1H), 2.85 (dd, 1H, J ) 6.0, 17.5), 2.26-2.16 (m, 1H), 2.03 (s, 3H),
1.47-1.24 (m, 8H), 0.91 (m, 6H).
Representative Procedure: Synthesis of 6 from 32. (3R,4R,5S)-
4-Acetamido-5-amino-3-(1-ethylpropoxy)-1-cyclohexene-1-carboxy-
lic Acid (6h). To a solution of azide 32h (1 g, 3.1 mmol) in THF (30
mL) were added triphenylphosphine (1.21 g, 4.6 mmol) and water (5.6
mL). The solution was heated at 50 °C for 10 h, THF was evaporated,
and the aqueous oily residue was partitioned between ethyl acetate and
brine. The organic phase was dried, filtered, and evaporated. Purifica-
tion of the residue by chromatography eluting with methanol/ethyl
acetate (1:1) gave 830 mg of an oil which was dissolved in THF (15
mL) and treated with 1 N potassium hydroxide (4 mL, 4.16 mmol).
The reaction mixture was stirred at room temperature for 40 min and
acidified to pH 6 with Dowex 50WX8. The resin was filtered and
washed with water and methanol. Solvents were evaporated, and the
residue was purified by C18 chromatography eluting with water and
then with 5% acetonitrile in water. Fractions containing the desired
product were pooled and lyophilized to afford 6h (600 mg, 75%) as a
white solid: 1H NMR (D2O) δ 6.50 (t, 1H, J ) 2.1 Hz), 4.30-4.26
(m, 1H), 4.03 (dd, 1H, J ) 9.0, 11.7 Hz), 3.58-3.48 (m, 2H), 2.88
(dd, 1H, J ) 5.4, 16.8 Hz), 2.53-2.41 (m, 1H), 1.62-1.40 (m, 4H),
0.90 (t, 3H, J ) 7.5 Hz), 0.85 (t, 3H, J ) 7.5 Hz). Anal. Calcd for
C14H24N2O4‚1.5H2O: C, 54.01; H, 8.73; N, 9.00. Found: C, 53.69;
H, 8.70; N, 8.71.
1
afford 32h as needles: mp 121-123 °C; H NMR (CDCl3) δ 6.79 (t,
1H, J ) 2.1 Hz), 5.92 (d, 1H, J ) 7.5), 4.58 (bd, 1H, J ) 8.7), 4.35-
4.25 (m, 1H), 3.77 (s, 3H), 3.36-3.25 (m, 2H), 2.85 (dd, 1H, J ) 5.7,
17.4), 2.29-2.18 (m, 1H), 2.04 (s, 3H), 1.60-1.45 (m, 4H), 0.91 (t,
3H, J ) 3.7), 0.90 (t, 3H, J ) 7.3); 13C NMR (CDCl3) δ 171.1, 166.2,
138.4, 127.7, 82.1, 73.8, 57.5, 57.2, 52.0, 30.4, 26.1, 25.5, 23.4, 9.5,
9.1. Anal. Calcd for C15H24N4O4: C, 55.54; H, 7.46; N, 17.27.
Found: C, 55.57; H, 7.31; N, 17.30.
Methyl (3R,4R,5S)-4-acetamido-5-azido-3-methoxy-1-cyclohex-
ene-1-carboxylate (32a): mp 145-146 °C (needles from hexane/ethyl
1
acetate); H NMR (CDCl3) δ 6.89 (t, 1H, J ) 2.4), 5.73 (d, 1H, J )
7.6), 4.40 (m, 1H), 4.12 (m, 1H), 3.78 (s, 3H), 3.58 (m, 1H), 3.42 (s,
3H), 2.89 (ddd, 1H, J ) 0.7, 5.7, 17.6), 2.34-2.22 (m, 1H), 2.07 (s,
3H). Anal. Calcd for C11H16N4O4: C, 49.25; H, 6.01; N, 20.88.
Found: C, 49.24; H, 5.98; N, 20.90.
Methyl (3R,4R,5S)-4-acetamido-5-azido-3-ethoxy-1-cyclohexene-
1-carboxylate (32b): mp 141-143 °C (needles from hexane/ethyl
1
acetate); H NMR (CDCl3) δ 6.86 (t, 1H, J ) 2.3), 6.08 (d, 1H, J )
7.9), 4.45 (m, 1H), 4.12 (m, 1H), 3.77 (s, 3H), 3.72-3.51 (m, 2H),
2.87 (dd, 1H, J ) 5.5, 17.1), 2.32-2.21 (m, 1H), 2.06 (s, 3H), 1.20 (t,
3H, J ) 7.0). Anal. Calcd for C12H18N4O4: C, 51.06; H, 6.43; N,
19.85. Found: C, 50.79; H, 6.21; N, 19.63.
(3R,4R,5S)-4-Acetamido-5-amino-3-methoxy-1-cyclohexene-1-
carboxylic Acid Hydrochloride (6a). Amino acid 6a was isolated as
the hydrochloride salt after being treated with 1 N HCl and lyophi-
lized: 1H NMR (D2O) δ 6.87 (s, 1H), 4.26-4.14 (m, 2H), 3.66-3.57
(m, 1H), 3.45 (s, 3H), 2.94 (dd, 1H, J ) 5.7, 17.4), 2.59-2.47 (m,
1H), 2.09 (s, 3H).
Methyl (3R,4R,5S)-4-acetamido-5-azido-3-(1-propoxy)-1-cyclo-
hexene-1-carboxylate (32c): mp 147-148 °C (needles from hexane/
ethyl acetate); 1H NMR (CDCl3) δ 6.85 (t, 1H, J ) 2.1), 5.81 (d, 1H,
J ) 6.3), 4.74 (m, 1H), 4.10 (m, 1H), 3.77 (s, 3H), 3.69-3.37 (m,
3H), 2.87 (dd, 1H, J ) 5.7, 17.1), 2.32-2.21 (m, 1H), 2.05 (s, 3H),
1.57 (m, 2H), 0.92 (t, 3H, J ) 7.2). Anal. Calcd for C13H20N4O4: C,
52.69; H, 6.80; N, 18.91. Found: C, 52.67; H, 6.56; N, 18.89.
Methyl (3R,4R,5S)-4-acetamido-5-azido-3-(1-butoxy)-1-cyclohex-
ene-1-carboxylate (32d): mp 136-137 °C (needles from hexane/ethyl
(3R,4R,5S)-4-Acetamido-5-amino-3-ethoxy-1-cyclohexene-1-car-
boxylic acid (6b): isolated as an amorphous white powder; 1H NMR
(D2O) δ 6.55 (s, 1H), 4.29 (m, 1H), 4.12 (dd, 1H), 3.82- 3.51 (m,
3H), 2.90 (dd, 1H), 2.58-2.43 (m, 1H), 2.07 (s, 3H), 0.90 (t, 3H).
(3R,4R,5S)-4-Acetamido-5-amino-3-propoxy-1-cyclohexene-1-car-
1
boxylic acid (6c): isolated as a solid; H NMR (D2O) δ 6.53 (t, 1H,
1
acetate); H NMR (CDCl3) δ 6.86 (t, 1H, J ) 2.3), 6.03 (d, 1H, J )
J ) 2.1), 4.29 (m, 1H), 4.16 (dd, 1H, J ) 8.7, 11.1), 3.78-3.72 (m,
2H), 3.62 (ddd, 1H), 2.95 (dd, 1H), 2.58-2.52 (m, 1H), 2.11 (s, 3H),
1.58 (q, 2H, J ) 7.3), 0.91 (t, 3H, J ) 7.3).
7.9), 4.43 (m, 1H), 4.09 (m, 1H), 3.77 (s, 3H), 3.68-3.41 (m, 4H),
2.87 (dd, 1H, J ) 5.2, 17.6), 2.32-2.21 (m, 1H), 2.06 (s, 3H), 1.54
(m, 2H), 1.36 (m, 2H), 0.91 (t, 3H, J ) 7.3). Anal. Calcd for
C14H22N4O4: C, 54.18; H, 7.15; N, 18.05. Found: C, 54.21; H, 6.98;
N, 17.94.
(3R,4R,5S)-4-Acetamido-5-amino-3-butoxy-1-cyclohexene-1-car-
boxylic acid (6d). Amino acid 6d was isolated as a hydrochloride
1
salt after being treated with 1N HCl and lyophilized: H NMR (D2O)
δ 6.81 (s, 1H), 4.28 (m, 1H), 4.15 (dd, 1H), 3.84-3.75 (m, 1H), 3.66-
3.53 (m, 2H), 2.95 (dd, 1H), 2.59-2.45 (m, 1H), 2.10 (s, 3H), 1.61-
1.50 (m, 2H), 1.42-1.23 (m, 2H), 0.90 (t, 3H).
(3R,4R,5S)-4-Acetamido-5-amino-3-(2-methylpropoxy)-1-cyclo-
hexene-1-carboxylic acid (6e): isolated as an amorphous white
powder; 1H NMR (D2O) δ 6.61 (br s, 1H), 4.25 (m, 1H), 4.13 (dd, 1H,
J ) 9.0, 11.3), 3.62-3.54 (m, 2H), 3.30 (dd, 1H, J ) 7.2, 9.3), 2.91
(dd, 1H, J ) 5.5, 17.2), 2.56-2.45 (m, 1H), 2.03 (s, 3H), 1.85-1.77
(m, 1H), 0.88 (2d, 6H).
Methyl (3R,4R,5S)-4-acetamido-5-amino-3-(2-methylpropoxy)-1-
cyclohexene-1-carboxylate (32e): isolated as a solid; 1H NMR (CDCl3)
δ 6.86 (t, 1H, J ) 2.2), 6.01 (d, 1H, J ) 7.9), 4.40 (m, 1H), 4.07 (m,
1H), 3.77 (s, 3H), 3.63 (m, 1H), 3.43 (dd, 1H, J ) 6.4, 8.8), 3.19 (dd,
1H, J ) 6.7, 8.8), 2.87 (dd, 1H, J ) 5.8, 17.3), 2.33-2.22 (m, 1H),
2.05 (s, 3H), 1.82 (m, 1H), 0.90 (2d, 6H).
Methyl (3R,4R,5S)-4-Acetamido-5-amino-3-(1(R)-methylpropoxy)-
1-cyclohexene-1-carboxylate (32f). Isolated as a crystalline solid
which was a 9:1 (32f:32g) diastereomeric mixture. Data for the major
diastereomer are reported: 1H NMR (CDCl3) δ 6.79 (t, 1H, J ) 2.2
Hz), 6.14 (d, 1H, J ) 7.3 Hz), 4.55 (m, 1H), 4.33-4.23 (m, 1H), 3.77
(s, 3H), 3.56-3.45 (m, 1H), 3.40-3.27 (m, 1H), 2.85 (dd, 1H, J )
5.5, 17.5 Hz), 2.30-2.15 (m, 1H), 2.04 (s, 3H), 1.59-1.40 (m, 2H),
1.10 (d, 3H, J ) 6.0 Hz), 0.91 (t, 3H, J ) 7.4 Hz). Anal. Calcd for
C14H22N4O4: C, 54.18; H, 7.15; N, 18.05. Found: C, 54.06; H, 7.11;
N, 17.96.
(3R,4R,5S)-4-Acetamido-5-amino-3-(1(R)-methylpropoxy)-1-cy-
clohexene-1-carboxylic acid (6f): isolated as an amorphous white
1
powder; H NMR (D2O) δ 6.52 (s, 1H), 4.28 (d, 1H, J ) 8.7), 4.04
(dd, 1H, J ) 8.8, 11.5), 3.74-3.65 (m, 1H), 3.50-3.60 (m, 1H), 2.90
(dd, 1H, J ) 5.5, 17.2), 2.50-2.40 (m, 1H) 2.10 (s, 3H), 1.60-1.45
(m, 2H), 1.14 (d, 3H, J ) 6.2), 0.91 (t, 3H, J ) 7.4).
(3R,4R,5S)-4-Acetamido-5-amino-3-(1(S)-methylpropoxy)-1-cy-
clohexene-1-carboxylic acid (6g): isolated as a white solid; 1H NMR
(D2O) δ 6.48 (t, 1H), 4.27 (m, 1H), 4.04 (dd, 1H, J ) 8.9, 11.6), 3.71-
3.65 (m, 1H), 3.59-3.50 (m, 1H), 2.90 (dd, 1H, J ) 5.5, 17.2), 2.50-
2.40 (m, 1H), 2.10 (s, 3H), 1.60-1.45 (m, 2H), 1.14 (d, 3H, J ) 6.2
Hz), 0.91 (t, 3H, J ) 7.4 Hz).
Methyl (3R,4R,5S)-4-Acetamido-5-amino-3-(1S-methylpropoxy)-
1-cyclohexene-1-carboxylate (32g). Isolated as a crystalline solid
which was a 9:1 (32g:32f) diastereomeric mixture. Data for the major
diastereomer are reported: 1H NMR (CDCl3) δ 6.78 (t, 1H, J ) 2.2
Hz), 5.82 (d, 1H, J ) 7.4 Hz), 4.61 (m, 1H), 4.35-4.26 (m, 1H), 3.76

690 J. Am. Chem. Soc., Vol. 119, No. 4, 1997
Kim et al.
(3R,4R,5S)-4-Acetamido-5-amino-3-(1-propylbutoxy)-1-cyclohex-
ene-1-carboxylic acid (6i): isolated as a white solid; 1H NMR (D2O,
500 MHz) δ 6.62 (t, 1H, J ) 2.5), 4.32 (m, 1H), 4.06 (dd, 1H, J )
8.5, 11.5), 3.69-3.65 (m, 1H), 3.61-3.56 (m, 1H), 2.93 (dd, 1H, J )
5.5, 17.0), 2.54-2.48 (m, 1H), 1.57-1.22 (m, 8H), 0.95-0.90 (m, 6H).
Methyl (1R,3R,4R,5R)-1-Hydroxy-3,4-O-isopropylidine-3,4-dihy-
droxy-5-[(p-tolylsulfonyl)oxy]cyclohexane-1-carboxylate (34). To a
solution of 33 (29.8 g, 121 mmol) and (dimethylamino)pyridine (500
mg) in pyridine (230 mL) was added p-toluenesulfonyl chloride (27.7
g, 145 mmol). The mixture was stirred at room temperature for 72 h.
The solvent was evaporated and the residue was diluted with water,
and extracted with ethyl acetate. The combined organic extracts were
washed with water, brine, and dried. Evaporation and purification by
chromatography eluting with hexane/ethyl acetate gave tosylate 34 (44.6
Plaque Reduction Assay. Plaque reduction assays were performed
as previously described.38 Confluent monolayers of MDCK cells in
six-well tissue culture plates were inoculated with 30 pfu of influenza
A/WS/33 virus. After 1 h, the inoculum was removed, and the cells
were overlaid with DMEM medium containing Earle’s salts, 15 mM
Hepes, pH 7.4, 0.2% BSA, 2 µg of TPCK-treated trypsin/mL, and 1%
agarose. After 2-3 days the overlay was removed, and the plaques
were visualized by staining the cell monolayer with 0.1% crystal violet
in 20% methanol.
Crystallization and Data Collection. Isolation, purification, and
crystallization of N9 neuraminidase has previously been reported.39
Crystals of the neuraminidase-inhibitor 6h complex were obtained by
soaking the neuramindase crystals overnight in 5 mM inhibitor 6h
solution (5 mM inhibitor, 2 volumes 1.4 M KH2PO4, 1 volume of 3 M
K2HPO4). Cubic-shaped crystals with approximate dimensions 0.2 ×
0.2 × 0.2 mm were used to collect room temperature data at the
Stanford Synchrotron Radiation Laboratory, beamline 7-1 (λ ) 1.08
Å) using a MAR30 Image Plate System. The crystal belong to the
cubic space group I432, a ) 182.9 Å. A complete data set to 2.7 Å
resolution was collected using 1° oscillations and a total of 30°. The
reflections were indexed using DENZO 1.5.11 and merged/scaled using
the program SCALEPACK.40 A total of 51 334 reflections were
collected; 12 383 reflections were unique. The reduced data set was
98% complete to 2.7 Å resolution with an Rmerge of 7.1%.
1
g, 92%) as a viscous oil; H NMR δ 7.84 (d, 2H, J ) 8.4), 7.33 (d,
2H, J ) 8.1), 4.76 (m, 1H), 4.42 (m, 1H), 4.05 (dd, 1H, J ) 5.5, 7.5),
3.80 (s, 3H), 2.44 (s, 3H), 2.35 (m, 1H), 2.24 (m, 2H), 1.96 (m, 1H),
1.26 (s, 3H), 1.13 (s, 3H). Anal. Calcd for C15H24O8S: C, 53.90; H,
5.91; S, 8.01. Found: C, 53.86; H, 5.91; S, 7.88.
Methyl (3R,4R,5R)-3,4-Dihydroxy-5-[(p-tolylsulfonyl)oxy]-1-cy-
clohexene-1-carboxylate (35). To a solution of 34 (44.6 g, 111.5
mmol) in CH2Cl2 (450 mL) at -78 °C was added pyridine (89 mL)
followed by slow addition of sulfuryl chloride (26.7 mL, 335 mmol).
The mixture was stirred at -78 °C for 5 h, and methanol (45 mL) was
added dropwise. The mixture was warmed to room temperature and
stirred for 12 h. Ethyl ether was added, and the organic phase was
washed with water and brine and dried. The crude residue (44.8 g)
was dissolved in methanol (500 mL), p-toluenesulfonic acid (1.06 g,
5.6 mmol) was added, and the solution was refluxed for 4 h. The
reaction mixture was cooled to room temperature and evaporated, and
the residue was dissolved in methanol (500 mL) and refluxed for 4 h.
Solvent was evaporated, and the residue was chromatographed eluting
with hexane/ethyl acetate to give 26.8 g of a solid which was
recrystallized from hexane/ethyl acetate to give 35 (20.5 g, 54%) as a
white solid; 1H NMR (CDCl3) δ 7.82 (d, 2H, J ) 8.3), 7.37 (d, 2H, J
) 8.3), 6.84 (m, 1H), 4.82 (dd, 1H, J ) 5.8, 7.4), 4.50 (m, 1H), 3.90
(dd, 1H, J ) 4.4, 8.2), 3.74 (s, 3H), 2.79 (dd, 1H, J ) 5.5, 18.2), 2.42
(dd, 1H, J ) 6.6, 18.2). Anal. Calcd for C15H18O7S: C, 52.60; H,
5.30; S, 9.37. Found: C, 52.75; H, 5.15; S, 9.21.
Structure Solution and Refinement. The neuraminidase com-
plexed with inhibitor 6h structure was solved by molecular replacement
using the influenza A subtype N9 neuraminidase crystal structure
containing residues 82-469 (PDB ID 1INY).41 Alternate rounds of
model building with the molecular graphics program O version 6.0.342
and refinement of the atomic coordinates using X-PLOR 3.143 were
performed until convergence of the free-R and R factor was achieved.
Examination of Fo - Fc maps showed unambiguous density for the
orientation of inhibitor 6h, sugar molecules, and calcium ion. Addition
of four N-acetyl-D-glucosamine sugars, five R-D-mannose sugars, one
calcium ion, and one inhibitor 6h molecule reduced the free-R and R
factor further. The model was refined through positional and overall
B factor refinement in X-PLOR until the minimization was complete.
The final model has an R factor of 16.3% (Rfree ) 22.8) in the 6-2.7
Å resolution range, with rms deviations of 0.012 Å and 1.83° for the
bond lengths and angles, respectively. The coordinates will be
deposited in the Protein Data Bank, Brookhaven National Laboratory.
Methyl (3R,4R,5S)-4,5-Epoxy-3-hydroxy-1-cyclohexene-1-car-
boxylate (12). To a solution 35 (20.0 g, 58.5 mmol) in THF (300
mL) at 0 °C was added 1,8-diazabicyclo[5.4.0]undec-7-ene (8.75 mL,
58.5 mmol). The reaction mixture was warmed to room temperature
and stirred for 12 h. Solvent was evaporated, and the residue was
purified by chromatography eluting with hexane/ethyl acetate (1:3) to
give epoxide 12 (9.72 g, 100%) as a white solid which was identical
to an authentic sample by NMR.32
Structure Analysis. Visual inspection of the structures was aided
with the programs O42 and INSIGHTII44 on an INDIGO Silicon
Graphics workstation. The stereochemistry of the models were
examined with PROCHECK 2.1.4.45
Neuraminidase Enzyme Assay. Neuraminidase enzyme activity
was determined using minor modifications to the literature method.37
Influenza A/PR/8/34 (H1N1), purified on sucrose density gradients,
was used as the source of enzyme and 4-(methylumbelliferyl)-R-D-N-
acetylneuraminic acid was used as substrate in a reaction buffer
containing 33 mM MES, pH 6.5, and 4 mM calcium chloride. Virus
was mixed with various inhibitor concentrations and incubated at room
temperature for 30 min before substrate was added to a final
concentration of 10 µM. Reactions were stopped after 8 min at 37 °C
with the addition of 1.5 volumes of 0.014 mM sodium hydroxide in
83% ethanol. Fluorescence was quantitated in a Perkin-Elmer fluo-
rimeter (Model LS50B) with an excitation wavelength of 360 nm,
emission wavelength of 448 nm, and slit width of 2.5.
JA963036T
(39) Laver, W. G.; Colman, P. M.; Webster, R. G.; Hinshaw, V. S.; Air,
G. M. Virology 1984, 137, 314.
(40) Otwinowski, Z.; Minor, W. The HKL Program Suite, in preparation,
1996.
(41) White, C. L.; Janakiraman, M. N.; Laver, W. G.; Philippon, C.;
Vasella, A.; Air, G. M.; Luo, M. J. Mol. Biol. 1995, 245, 623.
(42) Jones, T. A.; Zou, J.-Y.; Cowan, S. W.; Kjeldgaad, M. Acta
Crystallogr. 1991, A47, 110.
(43) Brunger, A. T. The X-PLOR 3.1 Software; Yale University: New
Haven, CT, 1993.
(44) Biosym Technologies, 1992.
(45) Laskowski, R. A.; MacArthur, M. W.; Moss, D. S.; Thornton, J.
M. J. Appl. Crystallogr. 1993, 26, 283.
(37) Potier, M.; Mameli, L.; Belisle, M.; Dallaire, L.; Melancon, S. B.
Anal. Biochem. 1979, 94, 287.
(38) Hayden, F. G.; Cote, K. M.; Douglas, R. G., Jr. Antimicrob. Agents
Chemother. 1980, 17, 865.
(46) During preparation of the manuscript, sialic acid based neuramini-
dase inhibitors with alkyl amide side chains at the C6 position were reported.
Sollis, S. L.; Smith, P. W.; Howes, P. D.; Cherry, P. C.; Bethell, R. C.
Bioorg. Med. Chem. Lett. 1996, 6, 1805.