Design, synthesis, and structural analysis of influenza neuraminidase inhibitors containing pyrrolidine cores

Article abstract of DOI:10.1021/jm000468c

The discovery of (±)-(2S,3R,4R)-2-(trifluoroacetamido)methyl-3-amino-1- (N′-ethyl-N′-isopropylcarbamyl)pyrrolidine-4-carboxylic acid (A-192558, 20e) as a potent inhibitor of influenza neuraminidase (NA) is described. Efficient syntheses of two core structures, cis-3-(allyloxycarbonyl)amino-1-(9′-fluorenylmethoxycarbonyl) pyrrolidine-4-carboxylic acid (7) and tert-butyl (±)-(2S,3R,4R)-2-aminomethyl-3-bis(tert-butyloxycarbonyl)amino-1- (N′-ethyl-N′-isopropylcarbamyl)pyrrolidine-4-carboxylate (18b), were developed. Starting with these core structures and using available structural information of the NA active site as the guide, analogues were synthesized in both the tri- and tetrasubstituted pyrrolidine series by means of high-throughput parallel synthesis in solid or solution phase for expeditious SAR. These studies accelerated the identification of (±)-(2S,3R,4R)-2-(trifluoroacetamido)methyl-3-amino-1- (N-ethyl-N-isopropylcarbamyl)pyrrolidine-4-carboxylate (20e, A-192558) as the most potent NA inhibitor in this series (IC₅₀ = 0.2 μM against NA A and 8 μM against NA B). The X-ray crystallographic structure of A-192558 bound to NA revealed the predicted interaction of the carboxylic group with the positively charged pocket (Arg118, Arg292, Arg371) and interaction of the trifluoro-acetamino residue with the hydrophobic pocket (Ile222, Trp178) of the enzyme active site. Surprisingly, the ethyl and isopropyl groups of the urea functionality induced a conformational change of Glu276, turning the Glu276/Glu277 hydrophilic pocket, which normally accommodates the triglycerol side chain of substrate sialic acid, into an induced hydrophobic pocket.

Full text of DOI:10.1021/jm000468c

1192
J . Med. Chem. 2001, 44, 1192-1201
Design , Syn th esis, a n d Str u ctu r a l An a lysis of In flu en za Neu r a m in id a se
In h ibitor s Con ta in in g P yr r olid in e Cor es
Gary T. Wang,*,† Yuanwei Chen,† Sheldon Wang,† Robert Gentles,† Thomas Sowin,† Warren Kati,‡
Steve Muchmore,§ Vincent Giranda,§ Kent Stewart,§ Hing Sham,‡ Dale Kempf,‡ and W. Graeme Laver#
Pharmaceutical Product Division, Abbott Laboratories, Abbott Park, Illinois 60064, and J ohn Curtin School of Medical
Research, The Australian National University, Canbera 260, Australia
Received November 2, 2000
The discovery of (()-(2S,3R,4R)-2-(trifluoroacetamido)methyl-3-amino-1-(N′-ethyl-N′-isopro-
pylcarbamyl)pyrrolidine-4-carboxylic acid (A-192558, 20e) as a potent inhibitor of influenza
neuraminidase (NA) is described. Efficient syntheses of two core structures, cis-3-(allyloxy-
carbonyl)amino-1-(9′-fluorenylmethoxycarbonyl)pyrrolidine-4-carboxylic acid (7) and tert-butyl
(()-(2S,3R,4R)-2-aminomethyl-3-bis(tert-butyloxycarbonyl)amino-1-(N′-ethyl-N′-isopropylcar-
bamyl)pyrrolidine-4-carboxylate (18b), were developed. Starting with these core structures and
using available structural information of the NA active site as the guide, analogues were
synthesized in both the tri- and tetrasubstituted pyrrolidine series by means of high-throughput
parallel synthesis in solid or solution phase for expeditious SAR. These studies accelerated
the identification of (()-(2S,3R,4R)-2-(trifluoroacetamido)methyl-3-amino-1-(N-ethyl-N-isopro-
pylcarbamyl)pyrrolidine-4-carboxylate (20e, A-192558) as the most potent NA inhibitor in this
series (IC50 ) 0.2 µM against NA A and 8 µM against NA B). The X-ray crystallographic
structure of A-192558 bound to NA revealed the predicted interaction of the carboxylic group
with the positively charged pocket (Arg118, Arg292, Arg371) and interaction of the trifluoro-
acetamino residue with the hydrophobic pocket (Ile222, Trp178) of the enzyme active site.
Surprisingly, the ethyl and isopropyl groups of the urea functionality induced a conformational
change of Glu276, turning the Glu276/Glu277 hydrophilic pocket, which normally accommodates
the triglycerol side chain of substrate sialic acid, into an induced hydrophobic pocket.
In tr od u ction
Influenza infection has been the cause of some of the
worst epidemics in human history and continues to be
a major health concern. In the United States alone,
20 000-40 000 people die from the flu, and the illness
costs as much as $12 billion a year in health care and
lost productivity.1 Despite these facts, effective and safe
antiinfluenza therapeutics are lacking, making anti-
influenza a high-priority and attractive area of drug
discovery. In recent years, virology studies of influenza
virus have resulted in an improved understanding of
the replication mechanism of the virus. Several molec-
ular targets have been identified for drug intervention
including haemaagglutinin, neuraminidase (sialidase),
M2 protein, and endonuclease.2 Among those potential
targets, neuraminidase (NA) appears to be particularly
attractive.3 NA is one of two glycoproteins expressed on
the virus surface and catalyzes the cleavage of sialic acid
residues from glycoproteins, glycolipids, and oligosac-
charides. The catalytic activity of NA is essential for
influenza virus replication and infectivity. It has been
shown that NA is required for elution of newly synthe-
sized virons from infected cells and that it facilitates
F igu r e 1. Structure of A-87380 and GG167.
the movement of the virus through the mucus of the
respiratory tract.3 The active site of NA is highly
conserved across all influenza A and B virus strains,
rendering broad-spectrum antiinfluenza agents possible.
Indeed, the effectiveness of NA inhibitors as antiinflu-
enza agents has been demonstrated both in animal
models and in human clinical trails by several research
groups and highlighted by recent FDA approval of
zanamivir and Tamiflu.4-7
Our effort in this area started with the identification
of cis-N-Boc-3-aminopyrrolidine 4-carboxylic acid (1,
Figure 1) as a modestly active NA inhibitor (IC50 ) 50
µM against NA A/Tokyo). Compound 1 was evaluated
structurally as a NA inhibitor lead by comparing 1 with
NA inhibitors known in the literature at the time,
primarily GG167 (2, Figure 1) and the related com-
pounds4,5 and by computer modeling. On the basis of
the literature data,8,9 we have derived an “airplane”
model of the NA active site as illustrated in Figure 2 to
summarize the basic structural requirements of a potent
* Address correspondence to: Dr. Gary T. Wang, D-4CP, AP10,
Abbott Laboratories, Abbott Park, IL 60064. Phone: 847-937-2489.
Fax: 847-935-0310. E-mail: gary.t.wang@abbott.com.
† Medicinal Chemistry Technologies, Abbott Laboratories.
‡ Anti-viral Research, Abbott Laboratories.
§ Structural Biology, Abbott Laboratories.
#
The Australian National University.
10.1021/jm000468c CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/21/2001

Influenza NA Inhibitors with Pyrrolidine Cores
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 8 1193
Sch em e 1a
a
Reagents and conditions: (i) t-BuOK, THF, rt overnight, 85-
95%; (ii) BnNH3OAc, HOAc, MeOH, reflux 1 h, 70-80%; (iii)
NaBH4, HOAc, 10 °C, 92%; (iv) H2, Pd/C, 90%; (v) allyl chlorofor-
mate, DIEA, DCM, overnight; (vi) TFA, rt 3 h; (vii) Fmoc-Cl,
NaHCO3, dioxane-H2O. Overall yield for steps v-vii was 93%.
F igu r e 2. Working model of the NA active site shown in two-
dimensional illustration. The center of site 2 is about 6 Å from
site 1 and about 4 Å from site 3, while site 4 is about 6 Å from
site 1 and 5 Å from site 3. Sites 1 and 3 are separated by 9-10
Å or about 7 single bond lengths. With GG167, the carboxylate
and acetyl groups bind to sites 1 and 3, respectively, while
the guanidino group interacts with site 2 and the triglycerol
side chain interacts with site 4.
combinatorial chemistry, particularly high-throughput
parallel synthesis, was used for rapid and comprehen-
sive optimization of a given scaffold.11 In this paper, we
report our effort on the tri- and tetrasubstituted pyr-
rolidine series using this approach, which led to the
discovery of A-192558.
NA inhibitor. The active site of NA has four main well-
conserved binding sites. The positively charged site 1
consists of Arg118, Arg292, and Arg371 and interacts
with the carboxylate of sialic acid or GG167 via charge-
charge interaction and hydrogen bonding. The nega-
tively charged site 2 consists of Glu119, Glu227, and
Asp151 and interacts with the guanidino group of
GG167. The small hydrophobic pocket consisting of
Ile222 and Trp178 (site 3) accommodates the acetyl
group of sialic acid or GG167, and site 4, consisting of
Glu276 and Glu277, binds to the triglycerol side chain
of sialic acid or GG167. In the two-dimensional sense,
sites 3 and 1 are situated at the head and tail of the
“airplane” respectively, separated by 9-10 Å or 6-7
single bond lengths, while sites 2 and 4 are situated at
two wings of the “airplane”. In the three-dimensional
sense, these binding pockets are clearly off-set from the
plane defined by the ring of the cyclic nucleus (as
illustrated by GG167) since aromatic (planar) mimics
of GG167 are poor inhibitors of NA.10 Literature data
indicated that occupation of all four pockets is necessary
for potent inhibition.8
Analysis of compound 1 using this model clearly
indicated some structural deficiencies. Aside from the
fact that a five-membered ring is used to present ligands
for an enzyme active site that normally accommodates
a six-membered ring in both the substrate and the most
potent known inhibitor (GG167),11 compound 1 has only
three substituents and therefore can interact with three
out of the four binding pockets. Nevertheless, compound
1 has a rigid (cyclic) scaffold, a positively charged group,
and a carboxylate, all in common with GG167. Thus,
we concluded that compound 1 can serve as a valid lead
and set out to optimize the structure, with the ultimate
goal of discovering clinically effective oral NA inhibitors.
Our strategy was to combine the strength of traditional
medicinal chemistry approaches with combinatorial
chemistry. Specifically, structural (X-ray crystallogra-
phy) analysis, computer modeling, and tradition syn-
thesis was used to explore a variety of scaffolds, while
Ch em istr y
Syn th esis of th e cis-3-Am in op yr r olid in e-4-ca r -
boxylic Acid Cor e. Our synthesis of the required
trisubstituted pyrrolidine core 7 is shown in Scheme 1.
Tandem Michael addition and Dieckmann condensation
of Boc-glycine methyl ester and tert-butyl acrylate
following a modification of the procedure of Kuhn and
Oswald gave the desired pyrrolidine â-keto ester 3 in
good yield.12 Compound 3 was converted to N-benzyl
enamine 4 by treating with benzylamine acetate in
methanol. Reduction of enamine 4 was accomplished
with NaBH4 in acetic acid in excellent yield, leading to
a separable mixture of cis- and trans-N-benzylamino
esters in 35:1 ratio, from which the cis-isomer 5 was
isolated. Debenzylation of 5 by hydrogenolysis then gave
compound 6. The cis-configuration of compound 6 was
readily established from the large NOE observed be-
tween protons on C3 and C4. Further manipulation of
protecting groups gave the orthogonally protected di-
amino acid 7.
High -Th r ou gh p u t P a r a llel Syn th esis of Tr isu b-
stitu ted P yr r olid in e An a logu es. Since a free car-
boxylic acid residue is required for NA inhibitory
activity, our SAR effort for the trisubstituted pyrrolidine
core was focused on substitution on two amino groups.
With the orthogonal protected core 7 in hand, solid-
phase synthesis appeared perfectly suited for this
purpose.
Compound 7 was coupled to standard p-hydroxy-
methylphenoxymethyl resin (Wang resin) following the
standard procedure, giving resin-bound core 8. Concep-
tually, a combinatorial library using mix-and-split
synthesis method can be prepared to co-optimize sub-
stitution on both the ring nitrogen and the exocyclic
amino group. However, we chose to take the parallel
synthesis approach based on a number of consider-
ations. Structurally, substitution on the exocyclic amino
group may not be tolerated, and therefore, large num-

1194 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 8
Wang et al.
Sch em e 2a
Sch em e 4a
a
Reagents and conditions: (i) 20% piperidine, DMF, rt 1 h; (ii)
a
Reagents and conditions: (i) 20% piperidine, DMF, rt 1 h; (ii)
h; (iii)
Pd(PPh3)4, Bu3SnH, DCM, H2O, 15 min; (iv) N,N′-bis(Boc)-1-
guanylpyrazole, DIEA, DMF, 50 °C (13a ); RCHO, DMA, NaBH-
(OAc)3, HOAc (13b-d ); Me2NCOCl, DIEA, DCM (13e); (v) 95%
TFA, 2.5 h, rt.
RCO2H, DIC, DCM; or ROCOCl, DIEA, DCM; or RSO2Cl, DIEA,
DCM; or RR′′NCOCl, DIEA, DCM; 2-24 h; (iii) Pd(PPh3)4,
Bu3SnH, DCM, H2O, 15 min; (iv) 95% TFA, 2.5 h, rt.
N,N-diisopropylcarbamyl chloride, pyridine, DCM,
4
Sch em e 3a
structure 18 also utilized the tandem Michael addition
and Dieckmann condensation between amino acid and
acrylate as the key step (Scheme 5).15 Initially, N-Cbz-
serine methyl ester with or without side chain hydroxyl
protection was tried as the condensation partner, but
the reaction was complicated by â-elimination of the
hydroxyl group. This problem was circumvented by
employing benzyl N-Cbz-R-glutamate, which reacted
with tert-butyl acrylate in the presence of t-BuOK to give
tert-butyl N-Cbz-2-(carboxymethyl)-3-oxopyrrolidine-4-
carboxylate in acceptable yield. This very polar material
was treated directly with ammonium acetate in reflux-
ing methanol, leading to enamino ester 14a which was
subsequently converted to the corresponding benzyl
ester 14b with benzyl bromide. Convertion of acid 14a
to ester 14b seemed an unnecessary step, but it facili-
tated purification of the material at this stage.17 The
enamino group of 14b was then protected as the bis-
(Boc) derivative to give 15, and hydrogenation of 15 with
Pd/C in acetic acid gave two diastereomers 16a ,b in good
yield. The diastereomeric amino acids 16a ,b were
inseparable chromatographically, and the ratio was
determined by NMR studies to be 3:1. This observed
selectivity in favor of the desired (()-(2S,3R,4R) dia-
stereomer contradicted what would be predicted based
on steric factors. We postulated that the debenzylation
of the side chain benzyl ester precedes the reduction of
the double bond and the free carboxylate group partially
controls the relative sterochemistry by chelating with
the catalyst, giving the observed stereoselectivity. In
contrast, hydrogenation of the analogous enamine 21
(derived from N-Cbz-L-leucine methyl ester and tert-
butyl acrylate following similar procedures) under iden-
tical conditions gave two separatable diastereomers
22a ,b in a ratio of 1:1.2, as predicted based on steric
consideration (Scheme 6).
a
Reagents and conditions: (i) 20% piperidine, DMF, rt 1 h; (ii)
20% phosgene, DCM, DIEA, 0 °C, 1 h; (iii) RNHR′′, pyridine, DCM,
3-8 h; (iv) Pd(PPh3)4, Bu3SnH, DCM, H2O, 15 min; (v) 95% TFA,
2.5 h, rt.
bers of analogues with subsitution at this position may
not be warranted. Technically, parallel synthesis of
discrete compounds allows routine characterization of
the products as well as purification if necessary and,
therefore, is well-suited for SAR studies.
Thus, resin-bound core 8, after deprotection of the
ring nitrogen, was functionalized with commercially
available carboxylic acids, chloroformates, sulfonyl chlo-
rides, or carbamyl chlorides to give resin-bound amides,
carbamates, sulfonamides, or ureas (9), respectively
(Scheme 2). Additionally, the resin-bound core 8, after
removal of the Fmoc protection with piperidine, was
converted to the corresponding carbamyl chloride on
resin using a method which we have described13 and
then reacted with primary or secondary amines, allow-
ing preparation of a broader range of ureas (Scheme 3).
The Alloc group of the resin-bound compounds 9 was
then removed with Pd(PPh3)4 and n-Bu3SnH in DCM
with a catalytic amount of water. The desired products
10 were obtained by cleavage with TFA. All the opera-
tions of solid-phase synthesis, including the cleavage
step, were carried out fully automated in batches of 32
on a commercial instrument (Advanced ChemTech 396),
allowing preparation of about 550 analogues over the
course of 2 months. All the products were analyzed by
HPLC with both UV and an evaporative light scattering
detectors and mass spectrometry before accepted for
enzyme inhibition assay.
To examine the effect of substitution on the exocyclic
amino group, the substitution on the ring nitrogen was
fixed as N,N-diisopropylurea (vide infra) and the exo-
cyclic amino group was converted into a guanidino
moiety using N,N′-bis(Boc)-1-guanylpyrazole (13a ) or
N,N-dimethylurea using N,N-dimethylcarbamyl chlo-
ride (13e), or alkylated via reductive amination (13b-
d ) (Scheme 4). Guanidination of primary amines on a
solid support with N,N′-bis(Boc)-1-guanylpyrazole has
been described recently.14
We again chose to fix the ring nitrogen substitution
based on the results of optimizing the trisubstituted
pyrrolidines 10 (Scheme 3 and Table 1). Accordingly, a
mixture of 16a ,b was converted to the corresponding
N-ethyl-N-isopropylureas, from which the desired (()-
(2S,3R,4R) diastereomer 17 (higher Rf, major isomer)
was isolated. Curtius rearrangement using diphenylphos-
phoryl azide in the presence of benzyl alcohol gave the
Cbz-protected amine 18a which upon hydrogenolysis
gave the free amine 18b.
With compound 18b on hand, a library of analogues
then became readily accessible. Since we intended to
project a ligand into the hydrophobic pocket (site 3,
Figure 1) via the C2 aminomethyl group and the small
Syn th esis of Tetr a su bstitu ted P yr r olid in es. The
synthesis of the desired tetrasubstituted pyrrolidine core

Influenza NA Inhibitors with Pyrrolidine Cores
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 8 1195
Sch em e 5a
a
Reagents and conditions: (i) t-BuOK, THF, rt 24 h; (ii) H4NOAc, MeOH, reflux 3 h; (iii) BnBr, DIEA, DCM, reflux 2 days (30-50%
yield for steps i-iii); (iv) (Boc)2O, DMAP, DCM, 24 h (40-50% yield); (v) H2, Pd, HOAc (80% yield); (vi) N-ethyl-N-isopropylcarbamyl
chloride, pyridine, DCM, 4 h (82% yield); (vii) DPPA, Et3N, BnOH, toluene, 100 °C, 16 h; (viii) H2, Pd (70% yield for steps vii and viii);
(ix) RCO2H, EDC, DCM; or ROCOCl, DIEA, DCM; or RSO2Cl, DIEA, DCM; or RR′′NCOCl, DIEA, DCM; 2-24 h; (x) 95% TFA, 2.5 h, rt.
Ta ble 1. NA Inhibitory Activity of Selected Trisubstituted
Pyrrolidines (()-10
Sch em e 6
volume of this pocket, a relatively small set of analogues
was required to fully optimize the structure. Conse-
quently, solution-phase parallel synthesis starting with
18b was apparently the method of choice for library
synthesis. Thus, 18b was treated with a set of amine-
reactive reagents (carboxylic acids, sulfonyl chlorides,
chloroformates, and carbamyl chlorides) to give 19 (a
library of 70 compounds). Compounds 19 were purified
by extraction and chromatography using silica gel
cartridges. TFA deprotection then gave the designed
inhibitors 20. All samples of 20 were analyzed with
HPLC and mass spectrometry. Samples with HPLC
purity better than 75% were accepted for biological
testing (70 compounds out of 76 attempted syntheses
met this criteria).
Resu lts a n d Discu ssion
Tr isu bstitu ted P yr r olid in e Ser ies. On the basis
of computer modeling, we anticipated that the -CO2H
group of our lead structure 1 interacts with the posi-
tively charged site 1 of the NA active site (Figure 2),
the exocyclic NH2 group binds to the negatively charged
site 2, and the Boc group interacts with the hydrophobic
site 3 via one of the methyl groups. That leaves the polar
site 4 unoccupied. Consequently, we focused our initial
effort primarily on optimizing the ring nitrogen substi-
tution of 1, with the goal of identifying substituents that
can interact with both sites 3 and 4 of the NA active
site. Accordingly, larger numbers (∼550) of analogues
of 1 were synthesized, including amides (∼250), car-
bamates (∼25), sulfonamides (∼25), and ureas (∼250).
Automated solid-phase parallel synthesis (Schemes 2
and 3) turned out to be an excellent approach for this
purpose due to its efficiency and simplicity.
In selecting monomers for the functionalization of the
ring nitrogen of 8, considerations were given to achiev-
ing both diversity and the structural requirements of
NA active site. This was particularly important for
carboxylic acids (amide formation) and amines (urea

1196 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 8
Wang et al.
Ta ble 2. NA Inhibitory Activity of Selected Trisubstituted
Pyrrolidines (()-13
formation), for which large pools were commercially
available. Thus, large numbers of aliphatic carboxylic
acids and amines with subtle structural variations were
employed to fine-tune the interaction with the hydro-
phobic pocket (Figure 2). At the same time, a variety of
monomers with aromatic or polar functional groups
were chosen in order to discover unexpected active
structures.
F igu r e 3. X-ray crystallographic structure (active site only)
of the complex of trisubstituted pyrrolidine 10i with NA A.
Table 2). For the GG167 (Figure 1) series of compounds,
it has been reported that conversion of 4-amino-
Neu5Ac2en to 4-guanidino-Neu5Ac2en (GG167) re-
sulted in a 40-fold increase in NA inhibitory activity.8b
In contrast, when 10f (IC50 ) 4 µM) was converted to
the corresponding guanidine (13a ), a small loss of
activity was observed. Alkylation or acylation of the
exocyclic amino group led to inactive compounds.
To gain insight of the interaction of this series of
compounds with the NA active site, the structure of
compound 10i bound to N9 A/Tern NA was solved using
X-ray crystallography to a resolution of 2.8 Å (Figure
3). As expected, the carboxylate of 10i interacts strongly
with the three arginines of site 1 (Arg118, Arg292,
Arg371) through charge-charge interaction. The exo-
cyclic 3-amino group forms hydrogen bonds with Glu119
of site 2 and interacts electrostatically with Asp151. The
most surprising discovery from this structure is, how-
ever, the disposition of the urea functionality of 10i.
Contrary to our initial expectation, the isopropyl group
does not occupy the hydrophobic pocket formed by the
side chains of Ile222 and Trp178 (site 3, Figure 2).
Instead, this group is projected into site 4 and forces a
conformational change of the side chain of Glu276,
creating a new (induced) hydrophobic pocket. Since our
observation, this same phenomenon has also been
reported with several compounds containing six-mem-
bered rings with hydrophobic residues replacing the
triglycerol side chain of GG167 (Figure 1).4,6 The 3′-
hydroxypropyl group, the second N-substituent of the
urea functionality of 10i, indeed points toward the
solvent.
The vast majority of compounds 10 synthesized were
inactive as NA inhibitors (IC50 > 100 µM). Compounds
10 with IC50 lower than 50 µM against NA A/Tokyo are
summarized in Table 1, along with several closely
related inactive compounds. Of all the amides synthe-
sized, the only active compound was the 2-ethylbutyric
acid derivative 10a (IC50 ) 22 µM). The ureas turned
out to be more productive, with nearly a dozen com-
pounds showing IC50 below 50 µM. The most active
compounds (10e,l) showed IC50 of about 1.5 µM, repre-
senting about 30-fold improvement in potency over our
lead 1. Further examination of the data in Table 1
indicated a clear preference for R-branched amides or
ureas derived from aliphatic secondary amines. In the
urea series, isopropyl was clearly optimal as one of the
N-substituents (10b-f), while the other N-substituent
can be a variety of groups larger than methyl. It is
noteworthy that ureas of primary amines, including
propylamine, are inevitably inactive. Similar observa-
tions have been reported in pyrancarboxamides related
to GG167.6b No active compounds were identified from
the carbamate or sulfonamide series, most likely a
reflection of the fact that all the chloroformates and
sulfonyl chlorides employed are either aromatic or
aliphatic with straight chains.
Compounds 10h -n were specially designed for con-
comitant interaction with both the hydrophobic site 2
(via the isopropyl group) and the polar site 4 of the
active site. Evidently, this strategy failed to yield the
desired outcome. Despite the variation in both charge
characteristics and the chain length of the second
N-substituent, the activity of these compounds fluctu-
ates within a range of 1.3-20 µM. This result seemingly
suggests that the polar residues have no interaction
with the enzyme and point toward the solvent. This was
proven by X-ray crystal structure (vide infra).
Tetr a su bstitu ted P yr r olid in e Ser ies. Even though
our effort with the trisubstituted pyrrolidine series led
to only 30-fold enhancement in NA inhibitory activity
(Table 1), we were encouraged by the findings of the
X-ray structure of the 10i and NA A complex (Figure
3). It has been reported that GG167 is not orally
bioavailable.5 The extreme hydrophilicity of the com-
pound (calculated log P ) -7.1) is undoubtedly a
contributing factor. Thus, placing a hydrophobic group
A limited number of compounds with substitution on
the exocyclic amino group were prepared (Scheme 4 and

Influenza NA Inhibitors with Pyrrolidine Cores
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 8 1197
Ta ble 3. NA Inhibitory Activity of Selected Tetrasubstituted
Pyrrolidines (()-20
F igu r e 4. X-ray crystallographic structure (active site only)
of the complex of tetrasubstituted pyrrolidine 20e with NA A.
Compound 10i was also shown for comparison.
(e.g. isopropyl) into site 4 to replace polar residues such
as the triglycerol side chain of GG167 represented an
important step toward the goal of orally bioavailable NA
inhibitors. Furthermore, the X-ray structure (Figure 3)
revealed that compound 10i did not occupy the hydro-
phobic site 3. This hydrophobic interaction was known
to be very important, contributing about 3 orders of
magnitude to the binding affinity of GG167.8b Thus, it
was conceivable that modified pyrrolidines permitting
this interaction would bring the desired enhancement
in potency. From the structure shown in Figure 3, it
appeared that a lipophilic residue placed at the C2
position of the pyrrolidine ring trans to the 3-amino and
4-carboxylate would be capable of interacting with site
3. The tetrasubstituted pyrrolidine core 18b was de-
signed on the basis of this premise.
This rotation turned the exocyclic amino group away
from a position for optimal interaction with all three
acidic amino acid residues of site 2 (Asp151, Glu119,
Glu227), diminishing binding contribution from this
site.
The data in Table 3 also indicate that the tetra-
substituted pyrrolidines are substantially more potent
inhibitors of NA A than NA B, an undesirable feature.
This phenomenon has been reported in other series of
NA inhibitors where the induced hydrophobic pocket is
seen.6,7 Therefore, one of the challenges for future work
is to improve the potency against both NA A and NA B
simultaneously.
Con clu sion
With the synthesis of tetrasubstituted pyrrolidine core
18b accomplished (Scheme 5), solution-phase parallel
synthesis was utilized to prepare a library of 70 deriva-
tives 20, including amides, sulfonamides, and some
ureas. The SARs of these compounds are summarized
in Table 3. Not surprisingly, compounds 20 with an
acetamido (20a ) or trifluoroacetamido (20e) group were
found to be the only actives from the series. The activity
falls off rapidly with the increasing size of the C2
substituents, reflecting the small size as well as the
rigidity of site 3 of the NA active site (Figure 2). The
trifluoroacetamino-functionalized 20e is 26- and 40-fold
more active than the acetamino-functionalized 20a
against NA A and NA B, respectively.
The most active tetrasubstituted pyrrolidine, 20e (A-
192558), was soaked into a crystal of N9 A/Tern NA,
and the structure of the complex was determined, as
shown in Figure 4. The expected interaction of the
carboxylate with the positively charged site 1 was
readily observed. The urea functionality again interacts
with site 4 hydrophobically by inducing a conforma-
tional change of Glu276, as with compound 10i (Figure
3). The C2 trifluoroacetamido group occupies the hy-
drophobic pocket (site 3), meeting our design goal.
However, the pyrrolidine ring of 20e unexpectedly
rotated by about 90° relative to the trisubstituted 10i.
Starting from the initial lead 1, a sub-micromolar NA
inhibitor 20e was discovered employing a combination
of structure-based drug design and combinatorial chem-
istry (parallel synthesis). Compound 20e represents a
250-fold improvement in terms of activity against NA
A over the initial lead 1.
Several reasons can be cited to explain why the
tetrasubstituted pyrrolidines 20 failed to show further
improvement in NA inhibitory activity. Fixing the
substitution on the ring nitrogen of 20 as the N-ethyl-
N-isopropyl might be one of the factors. This residue
was optimized for the trisubstituted series 10 but may
not be optimal for the tetrasubstituted series 20. We
addressed this question by synthesizing a combinatorial
library which varied substitutions on both the ring
nitrogen and the C2 aminomethyl group simulta-
neously, as described elsewhere.18 Another important
factor is the questionable contribution to binding from
the exocyclic amino group. As mentioned previously, the
X-ray crystallographic structures (Figures 3 and 4)
indicated that this amino group was not in the optimal
position for maximum interaction with site 2 in either
the tri- or tetrasubstituted series. Indeed, control com-
pounds in which the exocyclic amino group of 20e was
removed had essentially identical NA inhibitory activity

1198 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 8
Wang et al.
at 10-15 °C for 6 h, when TLC (30% EtOAc-hexane) indicated
complete conversion of the starting material. Acetic acid was
evaporated. The residue was cooled with an ice bath and 200
mL of saturated aqueous Na2CO3 was carefully added. The
mixture was then extracted with DCM (4 × 200 mL). The
combined organic solution was dried (MgSO4), filtered and
concentrated. The crude product was purified by chromatog-
raphy (silica gel, 10% EtOAc-hexane, then 30% EtOAc) to give
5 (33.0 g, 92%, white solid) and 1.3 g (3.7%) of the trans-isomer.
MS (DCI-NH3): m/z 377 (M + H), base peak. 1H NMR
(CDCl3): 1.48 (s, 9H), 1.49 (s, 9H), 2.17 (bs, 1H), 3.08-3.12
(m, 1H), 3.45-3.55 (m, 4H), 3.60-3.78 (m, 1H), 3.85 (s, 2H),
7.32 (m, 5H). In DMSO-d6, strong NOE was observed between
C3 H (δ 3.38) and C4 H (δ 3.09). Anal. (C21H32N2O4) C, H, N.
as the parent 20e, indicating that the exocyclic amino
group makes negligible contribution to binding.19 This
indicates that even the tetrasubstituted pyrrolidine 20e
still acts as a three-site binder, consistent with its high-
nanomolar activity. These observations clearly illustrate
the challenge of designing potent NA inhibitors based
on a five-membered ring scaffold: how to occupy all four
binding pockets of the NA active site simultaneously in
an optimal way. Indeed, our group came to the realiza-
tion that designing compounds using GG167 as a
template and focusing on solving the oral bioavailability
problem would be a more productive avenue,20 as
reported subsequently by the Gilead group.6,7 Within
the perimeter of the five-membered ring scaffold, the
only way to address this challenge is to synthesize a
series of core structures varying the substitution pattern
in order to identify a structure that can project all four
required substituents into the corresponding binding
pockets. The results of these efforts will be described
in future reports.
ter t-Bu tyl cis-N-Boc-3-a m in op yr r olid in e-4-ca r boxyla te
(6). Compound 5 (11.5 g, 30.6 mmol) was hydrogenated in
MeOH (250 mL) with 10% Pd/C overnight. After filtration of
the catalyst, evaporation of the solvent gave the crude product
which was purified by column chromatography (5% MeOH in
EtOAc) to give a white solid (7.8 g, 90.5%). MS (DCI-NH3):
1
m/z 287 (M + H), base peak. H NMR (CDCl3): 1.48 (s, 9H),
1.49 (s, 9H), 3.00 (m, 1H), 3.30 (m, 1H), 3.50-3.70 (m, 3H),
3.90 (m, 1H). Anal. (C14H26N2O4) C, H, N.
cis-3-(Allyloxyca r bon yl)a m in o-1-(9′-flu or en ylm eth oxy-
ca r bon yl)p yr r olid in e-4-ca r boxylic Acid (7). To a solution
of compound 6 (8.0 g, 27.9 mmol) in DCM (140 mL) was added
diallyl dicarbonate (11.5 mL, 30 mmol) and the solution was
stirred for 4 h. The solvent was evaporated and the residue
was directly purified by column chromatography (5-30%
EtOAc in hexane) to give 11.1 g of a white solid (∼100%). MS
(DCI-NH3): m/z 371 (M + H), m/z 388 (M + NH4). 1H NMR
(CDCl3): 1.47 (s, 9H), 1.48 (s, 9H), 3.12 (m, 1H), 3.30-3.48
(m, 1H), 3.50-3.70 (m, 3H), 4.4-4.5 (m, 1H), 4.58 (d, 2H),
5.20-5.38 (d. d, 2H, 5.80-6.00 (m, 1H).
The material from above was treated with 100 mL of 95%
TFA-water at ambient temperature for 2.5 h. The solvent was
evaporated under vacuum. The oily material was dissolved in
250 mL of 10% Na2CO3 and 150 mL of dioxane. N-(9-
Fluorenylmethoxycarbonyloxy)succinimide (10.4 g, 30.7 mmol)
was added. The solution was stirred at room temperature
overnight, washed with diethyl ether (2 × 300 mL), and
adjusted to pH 1-2 with concentrated HCl. The mixture was
then extracted with EtOAc (3 × 150 mL). The combined
organic solution was dried (MgSO4), filtered and concentrated.
Column chromatography gave a white solid, 11.4 g (93% from
6). MS (FAB+ K): m/z 475 (M + K), m/z 513 (M + 2K - H).
1H NMR (CDCl3): 3.25-3.35 (m, 1H), 3.40-3.50 (m, 1H),
3.65-3.80 (m, 3H), 4.15-4.25 (m, 1H), 4.35-4.45 (d, 2H),
4.50-4.70 (m, 2H), 5.15-5.40 (m, 2H), 5.55-5.65 (m, 1H),
5.80-5.98 (m, 1H), 6.85 (bs, 1H), 7.28-7.34 (t, 2H), 7.35-7.42
(t, 2H), 7.54, 7.56 (d, 2H), 7.74, 7.76 (d, 2H).
Loadin g of 7 to p-Hydr oxym eth ylph en oxym eth yl Resin
(8). A mixture of p-hydroxymethylphenoxymethyl resin (Wang
resin, 0.73 mmol/g, 15.0 g, 10.95 mmol; NovaBiochem, San
Diego, CA) and 100 mL of DCM was shaken for 0.5 h.
Compound 7 (7.17 g, 16.43 mmol) was then added, followed
by diisopropylcarbodiimide (2.57 mL, 16.43 mmol) and DMAP
(134 mg, 1.1 mmol). The mixture was shaken overnight. The
liquid was drained and the resin was washed 3 times with
DCM. The resin was then treated with 10% Ac2O in DCM (100
mL) for 0.5 h, drained, and washed again with DCM 3 times,
and dried in vacuo. The dry resin weighed 19.2 g. Fmoc test
gave a loading of 0.52 mmol/g.21
Exp er im en ta l Section
Gen er a l Meth od s. Commercially available solvents were
used as received and reagents were purchased from Aldrich
Chem. Co. (Milwaukee, WI), unless noted otherwise. Column
chromatography was performed with the solvent systems
indicated using E. Merck silica gel 60 (70-230 mesh). Unless
noted otherwise, all the reactions were conducted under an
1
atmosphere of nitrogen. H NMR spectra were recorded on a
GE AE-300 (300 MHz) using tetramethylsilane as an internal
standard. Elementary analyses were performed by Robertson
Micolit Laboratories, Madison, NJ. Abbreviations: Cbz, (benzyl-
oxy)carbonyl; Fmoc, fluorenylmethyloxycarbonyl; DMF, N,N-
dimethylformamide; NMP, N-methylpyrrolidone; DMAP, 4-(di-
methylamino)pyridine; EDCI, 1-ethyl-3-(3′-dimethylamino-
propyl)carbodiimide hydrochloride; HOBt, 1-hydroxybenzo-
triazole; TFA, trifluoroacetic acid.
ter t-Bu tyl N-Boc-3-oxop yr r olid in e-4-ca r boxyla te P o-
ta ssiu m Sa lt (3). A solution of Boc-glycine methyl ester (37
g, 0.199 mol) and tert-butyl acrylate (29.75 mL, 0.199 mol) in
400 mL of anhydrous THF (Aldrich SureSeal Grade) was
cooled to 0 °C. Potassium t-butoxide (25.86 g, 0.219 mol) was
added portionwise. The mixture was allowed to warm to
ambient temperature and stirred with a mechanical stirrer
for 24 h. Hexane (300 mL) was added to the thick mixture
and the solid was collected by filtration and washed twice with
diethyl ether. The dry solid weighed 56 g and was used directly
for the next step. MS (DCI-NH3): m/z 303 (M + NH4), base
peak. This material is a mixture of two regioisomers.
ter t-Bu tyl N-Boc-3-ben zyla m in o-3-p yr r olin e-4-ca r box-
yla te (4). A solution of compound 3 (41 g, 0.127 mol) and
benzylamine acetate (106 g, 0.63 mol) in 400 mL of MeOH and
7.3 mL of HOAc was refluxed for 1.5 h. The mixture was
concentrated and the residue was partitioned between 300 mL
of water and 300 mL of EtOAc. Two layers were separated
and the aqueous layer was extracted with EtOAc (3 × 135 mL).
The combined organic solution was dried (MgSO4), filtered and
concentrated to give a pale yellow solid. Column chromatog-
raphy on silica gel eluting with 10% EtOAc in hexane gave 4
as a white solid (37 g, 78.7%). Further elution with 30% EtOAc
in hexane give the minor regioisomer (6.0 g). MS (DCI-NH3):
m/z 375 (M + H), m/z 319 (M - C4H9). 1H NMR (CDCl3): 1.47
(s, 9H), 1.48 (s, 9H), 4.13 (m, 2H), 4.20 (m, 1H), 4.25 (m, 1H),
4.30 (m, 3H), 7.30 (m, 5H). Anal. (C21H30N2O4) C, H, N.
ter t-Bu tyl cis-N-Boc-3-ben zyla m in op yr r olid in e-4-ca r -
boxyla te (5). A three-necked flask, fitted with a thermometer
and rubber septum, was charged with HOAc and cooled to 10-
15 °C. NaBH4 (10.8 g, 0.284 mol) was added portionwise slowly
so that the internal temperature was kept below 20 °C. The
solution was stirred for 30 min. Compound 4 (35.5 g, 94.8
mmol) was then added in one portion. The solution was stirred
N-Acyla ted cis-3-(Allyloxyca r bon yl)a m in op yr r olid in e-
4-ca r boxyla te Wa n g Resin (9). For removal of the Fmoc,
resin 8 (19.0 g, 0.52 mmol/g) was treated with 190 mL of 25%
piperidine in NMP for 1 h. After draining, the resin was
washed with DMF (3 × 100 mL), MeOH (3 × 100 mL) and
DCM (3 × 100 mL) and dried.
On an Advanced ChemTech 396 multiple peptide synthe-
sizer (Advanced ChemTech, Louisville, KY), 80-100 mg of the
above deprotected resin was placed in each reaction vessel (RV)
(typically in batches of 32 reactions). The instrument was
programmed to add 1.0 mL of 80% DCM-NMP. For coupling

Influenza NA Inhibitors with Pyrrolidine Cores
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 8 1199
of carboxylic acids, the instrument was programmed to add 3
equiv of selected carboxylic acids in 0.25 mL of NMP and 3
equiv of diisopropylcarbodiimide to the corresponding RV. For
coupling of chloroformates, sulfonyl chlorides or carbamyl
chlorides, the instrument was programmed to add 3 equiv of
each monomer in 0.25 mL NMP and 3 equiv of diisopropyl-
ethylamine to the corresponding RV. The RV block was then
programmed to shake intermittently for 2-24 h, followed by
programmed washing with NMP (5 × 1 mL), MeOH (5 × 1
mL) and DCM (5 × 1 mL). Randomly chosen resins were
subjected to the Chloranil Test to confirm the completion of
acylation.22
N-Acyla ted cis-3-Am in op yr r olid in e-4-ca r boxylic Acid
(10). At the end of synthesis of resins 9 described above, 1.5
mL of a solution of (PPh3)4Pd in DCM (10 mg/mL) was
manually added to each RV via a syringe, followed by manual
addition of 0.1 mL of water and 0.1 mL of n-Bu3SnH. The
instrument was programmed to shake the RV block for 0.5 h,
wash the RVs with NMP (5 × 1.0 mL), MeOH (5 × 1.0 mL)
and DCM (5 × 1.0 mL). The instrument was then fitted with
a cleavage block and programmed to deliver 1.5 mL of TFA to
each RV, shake the RV block for 2.5 h, and drain the TFA
solutions into receiving vessels. The TFA solutions were then
concentrated on a speedvac to dryness. Each resulting products
were split into three portions, one for HPLC analysis, one for
MS analysis and the major portion for registration and
biological testing.
temperature, the solvent was evaporated. The residue was
taken up in 300 mL of water and the solution was adjusted to
pH 2-3, followed by extration with EtOAc (4 × 200 mL). The
EtOAc solution was then dried, filtered and concentrated to
give a brown oil (∼130 g). MS (DCI-NH3): m/z 395 (M + H) of
the keto ester as base peak.
The crude keto ester from above was mixed with 500 mL of
MeOH and 107 g of NH4OAc. The solution was refluxed for 5
h. After evaporation of the MeOH, the residue was mixed with
water (100 mL) and extracted with EtOAc (4 × 150 mL). The
combined EtOAc solution was dried, filtered and concentrated
to give crude 14a as a brown oil, which was directly used for
the next step.
To a solution of crude 14a from above (∼89 g, ∼0.237 mol)
in 500 mL of DCM were added benzyl bromide (49 g, 0.28 mol)
and DIEA (40.9 mL, 0.47 mol). The solution was refluxed for
2 days. DCM was evaporated and the residue was dissolved
with water (200 mL) and EtOAc (300 mL). Two layers were
separated and the aqueous layer was further extracted with
EtOAc (3 × 100 mL). The combined EtOAc solution was
worked up by drying, filtration and solvent removal. Flash
chromatography of the crude product on a silica gel column
using 20% EtOAc in hexane gave 14b as a clear oil (36 g, 27.8%
for three steps). Another product with lower Rf was also
isolated (15 g) which was identified as the benzyl N-Cbz-2-
(benzyloxycarbonyl)methyl-3-amino-3-pyrroline-4-carbox-
ylate. 14b MS (DCI-NH3): m/z 467 (M + H), m/z 484 (M +
NH4). 1H NMR (CDCl3): 1.43 (s, 9H), 2.62 (d, 1H), 2.66 (d,
1H), 3.38 (d, 1H), 4.10 (d, 1H), 4.24 (d, 1H), 5.16 (bs, 4H), 6.0
(bs, 2H), 7.20-7.42 (m, 10H).
ter t-Bu tyl N-Cbz-2-(ben zyloxyca r bon yl)m eth yl-3-bis-
(bu tyloxyca r bon yl)a m in o-3-p yr r olin e-4-ca r boxyla te (15)
To a solution of 14b (14 g, 30 mmol) in 100 mL of DCM at 0
°C was added di-tert-butyl dicarbonate (20 g, 90 mmol),
followed by DMAP (0.65 g). The solution was stirred at
ambient temperature overnight and concentrated. Flash chro-
matography on silica using 15% EtOAc as the eluent gave 15
as a thick oil which solidified upon standing (9.8 g, 49.0%).
MS (DCI-NH3): m/z 667 (M + H), m/z 684 (M + NH4), m/z
567 (M - Boc + H), base peak. 1H NMR (CDCl3): 1.43 (s, 27H),
2.70-2.82, 3.00-3.08 (m, 2H), 4.22-4.40 (m, 1H), 4.40-4.50
(m, 1H), 5.00 (bs.1H), 5.10-5.30 (m, 4H), 7.20-7.40 (m, 10H).
(()-(2S,3R,4R)- a n d (()-(2R,3R,4R)-ter t-Bu tyl 2-Ca r -
boxym eth yl-3-bis(bu tyloxycar bon yl)am in o-3-pyr r olidin e-
4-ca r boxyla te (m ixtu r e of 16a ,b). A solution of compound
15 (19 g, 28.5 mmol) in HOAc (500 mL) was mixed with 40 g
of 10% Pd/C and the mixture was hydrogenated at 4 atm of
H2 overnight. After removal of the catalyst, the solution was
concentrated and dried in vacuo to give a solid (14 g). This
crude mixture of 16a ,b was used directly for the next step.
ter t-Bu tyl (()-(2S,3R,4R)-2-Ca r boxym eth yl-3-bis(ter t-
bu tyloxycar bon yl)am in o-1-(N′-eth yl-N′-isopr opylcar bam -
yl)p yr r olid in e-4-ca r boxyla te (17). To a solution of 16 (16.0
g, 31.7 mmol) in DCM (150 mL) was added DIEA (11 mL, 79
mmol), followed by N,N-isopropylethylcarbamyl chloride (5.2
g, 38 mmol). The solution was stirred at ambient temperature
for 24 h when complete reaction was observed. The solution
was then washed with cold 0.01 N HCl (3 × 80 mL) and brine
(3 × 80 mL). After drying (MgSO4), the solution was concen-
trated and the residue purified by flash chromatography on a
silica gel column using 60% EtOAc in hexane with 2% HOAc
as the solvent, giving the desired (()-(2S,3R,4R) isomer 16
(higher Rf, 9.0 g, 50.9%, ∼90% purity) and the (()-(2R,3R,4R)
isomer (lower Rf, 3.5 g). 16 MS (DCI-NH3): m/z 558 (M + H),
base peak. 1H NMR (CDCl3): 1.04-1.18 (m, 5H), 1.22 (d, 3H),
1.45 (s, 9H), 1.50 (s, 18H), 2.60-2.80 (d.q, 2H), 2.90-3.00 (m,
1H), 3.30-3.40 (m, 2H), 3.50-3.60 (t, 1H), 3.70-3.80 (d. d,
1H), 3.90-4.00 (m, 1H), 4.60-4.70 (m, 1H), 4.70-4.80 (m, 1H).
cis-3-Am in o-1-(N′-eth yl-N′-isop r op ylca r ba m yl)p yr r oli-
d in e-4-ca r boxylic a cid (10e): 1.06 (t, J ) 7.1 Hz, 3H), 1.16
(2 d, J ) 6.4 Hz, 6H), 3.09-3.23 (m, 2H), 3.43-3.50 (m, 1H),
3.53-3.58 (d.d, J ) 4.1 and 11.9 Hz, 1H), 3.74-3.83 (m, 2H),
3.86-3.99 (hep., J ) 6.4 Hz, 1H), 4.10-4.15 (m, 1H). HRMS:
C
11H22N3O3, calcd 244.1661, found 244.1664.
P r ep a r a tion of Ca r ba m yl Ch lor id e on Resin (11). This
was carried out according to previously described procedure.13
cis-3-Am in o-1-(N′,N′-d iisop r op ylca r ba m yl)p yr r olid in e-
4-ca r boxylic Acid on Wa n g Resin (12). Resin 8 was
deprotected with piperidine as described above. The depro-
tected resin (2.0 g) was treated with 5 equiv each of diisopro-
pylcarbamyl chloride and diisopropylethylamine in DCM (8
mL) in a peptide flask overnight. The mixture was then
drained and the resin was washed with NMP, MeOH and DCM
(3 times each) and vacuum-dried. Chloranil test was negative.
This resin was then subjected to the (PPh3)4Pd-H2O-Bu3SnH
procedure described above to remove the Alloc group, and then
washed and dried.
P r ep a r a tion of 13. 13a : To a mixture of resin 12 (100 mg,
0.05 mmol) and NMP (1.5 mL) were added N,N′-bis(Boc)-1-
guanylpyrazole (77.5 mg, 0.25 mmol) and DIEA (44 µL, 0.25
mmol). The mixture was shaken at 50 °C overnight, when the
ninhydrin test of the resin turned negative. The mixture was
then filtered and the resin was washed with NMP, MeOH and
DCM (3 times each) and vacuum-dried. This resin was then
treated with 95% TFA-water to effect deprotection and
cleavage. The TFA solution was concentrated to give an oil
from which 9.0 mg of a solid was obtained upon trituration
with Et2O-hexane. MS (DCI-NH3): m/z 300 (M + H) and m/z
317 (M + NH4). The parent amine was not observed.
13b-d : To a mixture of resin 12 (100 mg, 0.05 mmol) and
5% HOAc in NMP (1.5 mL) was added 20 equiv of the required
aldehyde. The mixture was shaken for 2 h and drained, and
this treatment was repeated one more time. The resin was
then suspended in 2.0 mL of NMP and NaBH3CN (16 mg, 0.25
mmol) was added. The mixture was then shaken overnight
and drained. The resin was washed with NMP, MeOH and
DCM (3 times each) and vacuum-dried. TFA cleavage (rt, 2.5
h) gave the desired products which were characterized by
HPLC and MS.
ter t-Bu tyl N-Cbz-2-(ben zyloxycar bon yl)m eth yl-3-am in o-
3-p yr r olin e-4-ca r boxyla te (14b). A solution of N-Cbz-L-Asp-
R-OBn (NovaBiochem, San Diego, CA; 100 g, 0.278 mol) and
tert-butyl acrylate (90 mL, 0.56 mol) in 500 mL of THF was
cooled to 0 °C and t-BuOK (72 g, 0.56 mol) was added
portionwise over 0.5 h. After stirring overnight at ambient
ter t-Bu tyl (()-(2S,3R,4R)-2-(Ben zyloxyca r bon yl)a m i-
n om eth yl-3-bis(ter t-bu tyloxyca r bon yl)a m in o-1-(N′-eth yl-
N′-isop r op ylca r b a m yl)p yr r olid in e-4-ca r boxyla t e (18a ).
Compound 17 (4.0 g, 7.2 mmol), triethylamine (5.0 mL, 35.9
mmol), benzyl alcohol (6.5 mL, 63 mmol) and diphenylphos-
phoryl azide (1.5 mL, 11.6 mmol) were mixed with 50 mL of

1200 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 8
Wang et al.
toluene. The solution was stirred at 100 °C for 48 h, when TLC
indicated complete reaction. The solution was diluted with 100
mL of EtOAc and washed with cold 0.01 N HCl (2 × 150 mL)
and brine (2 × 150 mL). After drying (MgSO4), the solvent
was evaporated and the residue was purified by flash chro-
matography (20% EtOAc-hexane, then 30% EtOAc-hexane),
giving 18b as a white solid (2.9 g, 61.0%). MS (DCI-NH3): m/z
663 (M + H), base peak. 1H NMR (CDCl3): 1.04-1.18 (m, 5H),
1.20 (d, 3H), 1.45 (s, 9H), 1.50 (s, 18H), 2.80-2.96 (m, 1H),
3.20-3.42 (m, 4H), 3.43-3.56 (bt, 1H), 3.72-3.82 (bt, 1H),
4.50-4.60 (m, 1H), 4.62-4.72 (m, 1H), 5.08 (s, 2H), 5.50 (bs,
1H), 7.35 (bs, 5H).
ter t-Bu tyl (()-(2S,3R,4R)-2-Am in om eth yl-3-bis(ter t-bu -
tyloxycar bon yl)am in o-1-(N′-eth yl-N′-isopr opylcar bam yl)-
p yr r olid in e-4-ca r boxyla te (18b). A solution of 18a (2.5 g,
3.8 mmol) in 100 mL of 4.4 wt % of formic acid in MeOH was
purged with nitrogen. Palladium black (0.2 g) was added. After
15 min of stirring at ambient temperature, TLC indicated
complete reaction. The catalyst was removed by filtration and
the filtrate was concentrated. The residue was further dried
under vacuum to give a thick oil (2.1 g, ∼100%). MS (DCI-
NH3): m/z 529 (M + H), base peak. This material was used
directly for solution-phase synthesis of library of compounds
20.
Syn th esis of Libr a r y of Com p ou n d s 20. (a ) Rea ction
of 18b w ith Acid Ch lor id es, Su lfon yl Ch lor id es, a n d
Ca r ba m yl Ch lor id es. Reactions were carried out in batches
of 48 or smaller numbers when appropriate. A stock solution
of 18b in DCM (25 mg/mL) was aliquotted into the reaction
vessels (screw-cap tubes) at 1 mL/tube. Pyridine (5 equiv) was
added, followed by the desired acid chlorides, sulfonyl chlo-
rides, and carbamyl chlorides (1 equiv). The solutions were
agitated by shaking on a shaker. After 6 h at ambient
temperature, the reaction mixtures were directly purified by
chromatography using 2.0-g silica gel cartridges (Alltech,
Deerfield, IL) in batches of 24 with a 24-port vacuum manifold,
eluting with 50% EtOAc in hexane. The fractions containing
compound 19 were concentrated in a speedvac. The dried
compounds 19 were then treated with 2.0 mL of 80% TFA in
DCM for 4 h. Evaporation of TFA/DCM gave the TFA salt of
compounds 20 which were analyzed by reverse-phase HPLC
and MS (typical yield: 10 mg). The vast majority of compounds
20 prepared were over 90% pure by evaporative light scatter-
ing detection (ELS) on HPLC, even though samples with purity
over 75% purity were all accepted for NA inhibition testing.
NA In h ibition Assa y. NA from A/Tokyo/3/67 virus was
prepared by digestion of the virus with Pronase and purified
as described previously.23 NA inhibition assays were conducted
in 50 mM sodium citrate, 10 mM calcium choride, 0.1 mg/mL
BSA (pH 6.0) buffer with 5% DMSO. Reaction mixtures
included NA, inhibitor, and 30 µM 4-methylumbelliferylsialic
acid substrate in a total volume of 200 µL and were contained
in white 96-well U-shaped plates. The reactions were initiated
by the addition of enzyme and allowed to proceed for 30 min
at room temperature. The fluorescence for each well of the
plate was measured once each minute during the reaction
period by a Fluoroskan II plate reader (ICN Biomedical)
equipped with excitation and emission filters of 355 ( 35 and
460 ( 25 nm, respectively. The plate reader was under the
control of DeltaSoft II software (Biometallics) and a Macintosh
computer. Reaction velocities were linear over the 30-min
reaction period. Compounds were initially tested for influenza
NA inhibition at 0.5 mM. Those compounds which exhibited
more than 50% inhibition were evaluated further using eight
concentrations of compound. The reaction velocities for the
dose-response study were fit to the following equation using
a nonlinear regression program (Kaleidagraph) to determine
the overall Ki value:24 (1 - Vi/Vo) ) [I]/{[I] + Ki(1 + [S]/Km)},
where Vi and Vo represent inhibited and uninhibited reaction
velocities, respectively, and Km ) 40 µM which was established
from previous experiments.
X-r a y Cr ysta llogr a p h y. Isolation, purification, and crys-
tallization of N9 Tern NA has been described previously.25 The
crystals were exposed to the compounds of interest by soaking
in solutions that contained 2.3 M NaPO4, pH 7.5, and ligand
concentration approximately 10 mM. The soakings were
carried out at 4 °C and for 12-h duration. Once soaked, the
crystals were then transferred serially to a cryoprotectant, 2.3
M NaPO4, which contained 13% (v/v) and then 26% (v/v)
glycerol. The crystals were then mounted on an R-Axis II-C
area detector, which used X-rays generated by a Rigaku RU-
200 rotating anode X-ray generator fitted with Yale focusing
mirrors. Crystals were typically exposed for 20 min/0.5°
oscillation, and a total of 25° of data were collected. Data were
integrated and scaled using the HKL suite of programs.26 Maps
were calculated using the XPLOR suite of programs27 and then
displayed in the QUANTA molecular modeling program.28
Ack n ow led gm en t. We thank Dr. Daniel Norbeck,
Dr. J ames Summers, and Dr. William Kohlbrenner for
their support. We are grateful to Ayda Saldivar and H.
J an Huffaker for testing the NA inhibitory activity of
some of the compounds reported in this paper.
(b) Rea ction of 18b w ith Ca r boxylic Acid s. These
reactions were carried out following the same procedure as
described above with 25 mg of 18b per reaction, except that 1
equiv of EDCI and catalytic amount (3 mg) of HOBt were used
as the coupling reagents.
Su p p or tin g In for m a tion Ava ila ble: Additional analyti-
cal data for compounds in Tables 1-3. This material is
available free of charge via the Internet at http://pubs.acs.org.
(()-(2S ,3R ,4R )-2-(T r i flu o r o a c e t a m i d o )m e t h y l-3-
a m in o-1-(N′-et h yl-N′-isop r op ylca r b a m yl)p yr r olid in e-4-
ca r boxylic Acid (20e). To a solution of 18b (0.47 g, 0.89
mmol) in 10 mL of DCM were added HOBt (24 mg, 0.1 mmol),
EDCI (0.2 g, 1.1 mmol) and TFA (82 µL, 1.1 mmol). The
solution was stirred at room temperature overnight and then
concentrated. The residue was directly purified by flash
chromatography on a silica gel column eluting with 10-25%
EtOAc in hexane, giving 19 (X ) CF3CO-) as a white solid
(0.51 g, 91%). MS (DCI-NH3): m/z 625 (M + H), base peak.
1H NMR (CDCl3): 1.04-1.20 (m, 5H), 1.25 (d, 3H), 1.45 (s,
9H), 1.55 (s, 18H), 2.90-3.00 (m, 1H), 3.28-3.40 (m, 3H),
3.50-3.62 (m, 2H), 3.75-3.85 (dd, 1H), 3.85-3.96 (m, 1H),
4.50-4.60 (dd, 1H), 4.70-4.80 (m, 1H), 8.28 (bs, 1H). This
material (19, X ) CF3CO-, 0.28 g, 0.45 mmol) was treated
with 10 mL of 80% TFA in DCM for 6 h. After evaporation of
the solvents, the residue was treated with 10 mL of 1 N HCl
in diethyl ether, resulting in a white solid (20e, HCl salt) (175
mg, 97%). HPLC (ELS) indicated 98% purity. MS (DCI-NH3):
m/z 369 (M + H), base peak. 1H NMR (DMSO-d6): 0.9-1.1
(m, 5 H), 1.10-1.18 (d, 3H), 2.80-3.00 (m, 1H), 3.10-3.30 (m,
3H), 3.50-3.60 (m, 2H), 3.70-3.80 (m, 1H), 3.80-3.90 (m, 1H),
4.35-4.45 (d.d, 1H), 8.30-8.50 (bs, 2H), 9.32-9.42 (bs, 1H).
HRMS: C14H24N4O4F3, calcd 369.1750, found 369.1760.
Refer en ces
(1) (a) Service, R. F. Researchers Seek New Weapon Against the
Flu. Science 1997, 275, 756-757. (b) Glezen, W. P. Serious
Morbidity and Motality Associated with Influenza Epidemics.
Epidemiol. Rev. 1982, 4, 25-44. (c) Nicholson, K. G. Impact of
Influenza and Respiratory Syncytial Virus on Mortality in
England and Wales from J anuary 1975 to December 1990.
Epidemiol. Infect. 1996, 116, 51-63.
(2) Reviews: (a) Meanwell, N. A.; Krystal, M. Taking Aim at A
Moving Target-Inhibitors of Influenza Virus Part 1: Virus
Adsorption, Entry and Uncoating. Drug Discovery Today 1996,
1, 316-324. (b) Meanwell, N. A.; Krystal, M. Taking Aim at A
Moving Target-Inhibitors of Influenza Virus Part 2: Viral
Replication, Packaging and Release. Drug Discovery Today 1996,
1, 388-397. (c) Itzstein, M. v.; Barry, J . G.; Chong, A. K. The
Development of Potential Anti-influenza Drugs. Curr. Opin.
Ther. Patents 1993, 3, 1755. (d) Colema, P. M. In The Influenza
Viruses: Neuraminidase: Enzyme and Antigen; Krug, R. M., Ed.;
Plenum Press: New York, 1989; pp 175-218.
(3) (a) Bramford, M. Neuraminidase Inhibitors as Potential Anti-
Influenza Drugs. J . Enzyme Inhib. 1995, 10, 1-16. (b) Air, G.
M.; Laver, W. G. The Neuraminidase of Influenza Virus.
Proteins: Strut. Funct. Genet. 1989, 6, 341-356.

Influenza NA Inhibitors with Pyrrolidine Cores
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 8 1201
(4) (a) Smith, P. W.; Sollis, S. L.; Howes, P. D.; Cherry, P. C.;
Starkey, I. D.; Cobley, K. N.; Weston, H.; Scicinski, J .; Merritt,
A. Whittington, A.; Wyatt, P.; Taylor, N.; Green, D.; Bethell, R.;
Madar, S.; Fenton, R. J .; Morley, P. J .; Pateman, T.; Beresford,
A. Dihydropyrancarboxamides Related to Zanamivir: A New
Series of Inhibitors of Influenza Virus Sialidases 1. Discovery,
Synthesis, Biological Activity, and Structure-Activity Relation-
ship of 4-Guanidino- and 4-Amio-4H-pyran-6-carboxamides. J .
Med. Chem. 1998, 41, 787-797. (b) Taylor, N. R.; Cleasby, A.;
Singh, O.; Skarzynski, T.; Wonacott, A. J .; Smith, P. W.; Sollis,
S. L.; Howes, P. D.; Cherry, P. C.; Bethell, R.; Colman, P.;
Varghese, J . Dihydropyrancarboxamides Related to Zanamivir:
A New Series of Inhibitors of Influenza Virus Sialidases 2.
Crystallographic and Molecular Modeling Study of Complexes
of 4-Amio-4H-pyran-6-carboxamides and Sialidase from Influ-
enza Virus Types A and B. J . Med. Chem. 1998, 41, 798-807.
(5) (a) Hayden, F. G.; Rollins, B. R.; Madren, L. K. Anti-influenza
Virus Activity of the Neuraminidase Inhibitor 4-Guanidino-
Neu5Ac2en in Cell Culture and in Human Respiratory Epi-
thelium. Antiviral Res. 1994, 25, 123-131. (b) Woods, J . M.;
Bethell, R. C.; Coates, J . A. V.; Healy, N.; Hiscox, S. A.; Pearson,
B. A.; Ryan, D. M.; Ticehurst, J .; Tilling, J .; Walcott, S. M.; Penn,
C. R. 4-Guanidino-2,4-Dideoxy-2,3-Dehydro-N-Acetylneuraminic
Acid Is a Highly Effective Inhibitor of Both Sialidase (Neuramini-
dase) and Growth of a Wide Range of Influenza A and B Viruese
in vitro. Antimicrob. Agents Chemother. 1993, 37, 1473-1479.
(c) Ryan, D. M.; Ticehurst, J .; Dempsey, M. H.; Penn, C. R.
Inhibition of Influenza Virus Replication in Mice by GG167 (4-
Guanidino-2,4-Dideoxy-2,3-Dehydro-N-Acetylneuraminic Acid)
is Consistent with Extracellular Activity of Viral Neuraminidase
(Sialidase). Antimicrob. Agents Chemother. 1994, 38, 2270-2275.
(6) (a) Babu, Y. S.; Chand, P.; Bantia, S.; Kotian, P.; Dehghani, A.;
El-Kattan, Y.; Lin, T.-S.; Hutchinson, T. L.; Elliot, A. J .; Parker,
C. D.; Ananth, S. L.; Horn, L. L., Laver, G. W.; Montgomery, J .
A. BCX-1812 (RWJ -270201): Discovery of a Novel Highly Potent,
Orally Active, and Selective Influenza Neuraminidase Inhibitor
through Structure-Based Drug Design. J . Med. Chem. 2000, 43,
3482-3486. (b) Lew, W.; Wu, H.; Chen, X. Graves, B. J .; Escarpe,
P. E.; MacArthur, H. L.; Mendel, D. B.; Kim, K. U. Carbocyclic
Influenza Neuraminidase Inhibitors Possessing a C3-Cyclic
Amine Side Chain: Synthesis and Inhibitory Activity. Bioorg.
Med. Chem. Lett. 2000, 10, 1257-1260. (c) Kim, C. U.; Lew, W.;
Williams, M.; Liu, H.; Zhang, L.; Swaminathan, S.; Bischofberg,
N.; Chen, M. S.; Mendel, D. B.; Tai, C. Y.; Laver, W. G.; Stevens,
R. C. 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. J . Am. Chem.
Soc. 1997, 119, 681-690.
(7) (a) Mendel, D. B.; Tai, C. Y.; Escarpe, P. A.; Li, W.; Sidwell, R.
W.; Huffman, J . H.; Sweet, C.; Lakeman, K. J .; Merson, J .; Lacy,
S. A.; Lew, W.; Williams, M. A.; Zhang, L.; Chen, M. S.;
Bischofberger, N.; Kim, C. U. Oral Administration of a Prodrug
of the Influenza Virus Neuraminidase Inhibitor GS4071 Protects
Mice and Ferrets against Influenza Infection. Antimicrob. Agents
Chemother. 1998, 42, 640-646. (b) Li, W.; Escarpe, P. A.;
Eisenberg, E. J .; Cundy, K. C.; Sweet, C.; J akeman, K. J .;
Merson, J .; Lew, W.; Milliams, M.; Zhang, L.; Kim. C. U.;
Bischofberger, N.; Chen, M. S.; Mendel, D. B. Identification of
GS4104 as an Orally Bioavailable Prodrug of the Influenza Virus
Neuraminidase Inhibitor GS4071. Antimicrob. Agents Chemo-
ther. 1998, 42, 647-653.
(8) (a) Itzstein, M. v.; Dyason, J . C.; Olive, S. W.; White, H. F.; Wu,
W.-Y.; Kok, G. B.; Pegg, M. S. A Study of the Active Site of
Influenza Virus Sialidase: An Approach to the Rational Design
of Novel Anti-influenza Drugs. J . Med. Chem. 1996, 39, 388-
391. (b) Taylor, N. R.; Itzstein, M. v. Molecular Modeling Studies
on Ligand Binding to Sialidase From Influenza Virus and the
Mechanism of Catalysis. J . Med. Chem. 1994, 37, 616-624.
(9) (a) Varghese, J . N.; Epa, G. V.; Colma, P. M. Three-Dimensional
Structure of the Complex of 4-Guanidino-Neu5Ac2en and In-
fluenza Virus Neuraminidase. Protein Sci. 1995, 4, 1081-1087.
(b) Itzstein, M. v.; Wu, W.-Y.; Kok, G. B.; Pegg, M. S.; Dyason,
J . C.; J in, B. Phan, T. V.; Smythe, M. L.; White, H. F.; Oliver, S.
W.; Colman, P. M.; Varghese, J . N.; Ryan, D. M.; Woods, J . M.;
Bethell, R. C.; Hotham, V. J .; Cameron, J . M.; Penn, C. R.
Rational Design of Potent Sialidase-Based Inhibitor of Influenza
Virus Replication. Nature 1993, 363, 418-402. (c) Crennel, S.
J .; Garman, E. F.; Phiilppon, C.; Vasella, A.; Laver, W. G.; Vimr,
E. R.; Taylor, G. L. The Structure of Salmonella typhimurium
LT2 Neuraminidase and Its Complex with Three Inhibitors at
High Resolution. J . Mol. Biol. 1996, 259, 262-280.
(10) (a) Chand, P.; Babu, Y.; Bantia, Chu, N.; Cole, B.; Kotian, P. L.;
Laver, W. G.; Montgomery, J . A.; Pathak, V. P.; Petty, S. L.;
Shrout, D. P.; Walsh, D. A.; Walsh, G. M. Design and Synthesis
of Benzoic Acid Derivatives as Influenza Neuraminidase Inhibi-
tors Using Structure-Based Drug Design. J . Med. Chem. 1997,
40, 4030-4052. (b) Wiilliam, M.; Bischofberger, N.; Swami-
nathan, S.; Kim, C. U. Synthesis and Influenza Inhibitory
Activity of Aromatic Analogues of Sialic Acid. Bioorg. Med.
Chem. Lett. 1995, 5, 2251-254. (c) Singh, S.; J edrzejas, M. J .;
Air, G. M.; Luo, M.; Laver, W. G.; Brouillette, W. J . Structure-
Based Inhibitors of Influenza Virus Sialidase. A Benzoic Acid
Lead with Novel Interaction. J . Med. Chem. 1995, 38, 3217-
3225.
(11) When this work was in progress, whether potent NA inhibitors
based on a five-membered ring scaffold could be identified was
an open question. However, highly potent (sub-nanomolar) NA
inhibitors containing a cyclopentane scaffold have been reported
very recently.6a
(12) (a) Rapoport, H.; Willson, C. D. Synthesis of Pyrrole-3-carboxylic
Acid. J . Org. Chem. 1960, 26, 1102-1104. (b) Kuhn, R.; Osswald,
G. Neue Synthese von â-Pyrrolidonen: Darstellung von DL-γ-
Oxo-prolin, DL-allo-Hydroxy-prolin und 4-Athoxy-pyrrol-carbon-
sure-(2). Chem. Ber. 1956, 89, 1423-1442.
(13) Wang, G. T.; Wang, S.; Chen, Y.-c.; Sciotti, R.; Sowin, T. Solid
Phase Synthesis of Ureas of Secondary Amines via Carbamyl
Chloride. Tetrahedron Lett. 1997, 38, 1895-1898.
(14) Robinson, S.; Roskamp, E. J. Solid-Phase Synthesis of Guanidines.
Tetrahedron 1997, 53, 6697-6705.
(15) We also briefly investigated the direct regioselective C2 func-
tionalization of the dianion derived from the â-keto ester 3 as
described by Giles et al.16 The results were unsatisfactory.
(16) Giles, M.; Hadley, M. S.; Callagher, T. â-Ketoester Dianions as
Regiospecific Enolate Equivalents for N-Substituted Pyrrolidin-
3-ones. J . Chem. Soc., Chem. Commun. 1990, 1047-1048.
(17) Even though enantiomerically pure benzyl N-Cbz-L-R-glutamate
was employed in the synthesis, we have found that compounds
14 and 15 as well as all the downstream derivatives were
racemic. The epimerization evidently occurred during the t-
BuOK catalyzed condensation step.
(18) Gentles, R.; Chen, Y.; Woodall, L.; Wang, G., Wang, S.; Hochlow-
ski, J .; Sowin, T.; Voorbach, M.; Cheng, X.; Gao, L.; Beutel, B.;
Pavlik, P.; Martin, Y.; Stewart, K.; Hutchins, C. Discovery of a
Novel Hydrophobic Binding Pocket in Neuraminidase Active Site
via A Combinatorial Library of Pyrrolidines. J . Med. Chem., in
preparation.
(19) Chen, Y.; Betebenner, D.; Wang, G.; Gentles, R.; Sowin, T.;
Sham, H.; Kempf, D. Bioorg. Med. Chem. Lett., in preparation.
(20) Wang, G.; Wang, S.; Progress Toward The Synthesis of Func-
tionalized (()-(3R,4R,5S)-3,4,5-Triamino-Cyclohexene-1-Carbox-
ylic Acid. J . Org. Chem., in preparation.
(21) Meienhofer, J .; Waki, M.; Edgar, P. H.; Lambros, T. J .; Makofske,
R. C.; Chang, C. D. Solid-Phase Synthesis Without Repetitive
Acidolysis. Int. J . Pept. Protein Res. 1979, 13, 35-42.
(22) Vojkovcky, T. Detection of Secondary Amines on Solid Phase.
Pept. Res. 1995, 8, 236-237.
(23) Kati, W. M.; Saldivar, A. S.; Mohamadi, F.; Sham, H. L.; Laver,
W. G.; Kohlbrenner, W. E. Biochem. Biophys. Res. Commun.
1998, 244, 408-413.
(24) Segel, I. H. Enzyme Kinetics; Wiley-Interscience: New York,
1975; pp 105-106.
(25) Laver, W. G.; Coleman, P. M.; Webster, R. G.; Hinshaw, V. S.;
Air, G. M.; Influenza Virus Neuraminidase with Hemaggutinin
Activity. Virology 1984, 137, 314-323.
(26) Gewirth, D. The HKL Manual, 5th ed.; Yale University Press:
New Haven, CT, 1996.
(27) Brunger, A. T. Nat. Struct. Biol. 1997, 4, 862-865.
(28) The Quanta Manual; Biosym/MSI Inc., 1994.
J M000468C