Synthesis and evaluation of a new series of substituted acyl(thio)urea and thiadiazolo [2,3-a] pyrimidine derivatives as potent inhibitors of influenza virus neuraminidase

Article abstract of DOI:10.1016/j.bmc.2006.08.034

A series of substituted acyl(thio)urea and 2H-1,2,4-thiadiazolo [2,3-a] pyrimidine derivatives were prepared and both of their cell culture and enzymatic activity toward influenza virus were tested. Their in vitro neuraminidase inhibitory activities were in good agreement with the corresponding activities in cultured cells and they were evaluated as potent neuraminidase inhibitors. Of the analogues that demonstrated IC₅₀s < 0.1 μM, 16 and 60 were further investigated as candidates with the most potential for future development. The molecular docking work of the representative compound was described to provide more insight into their mechanism of action and further rationalize the observations of this new series herein, which represents a novel class of highly potent and selective inhibitors of influenza virus.

Full text of DOI:10.1016/j.bmc.2006.08.034

Bioorganic & Medicinal Chemistry 14 (2006) 8574–8581
Synthesis and evaluation of a new series of substituted
acyl(thio)urea and thiadiazolo [2,3-a] pyrimidine derivatives
as potent inhibitors of influenza virus neuraminidase
Chuanwen Sun,a,b Xiaodong Zhang,b Hai Huanga and Pei Zhoua,*
aDepartment of Biosynthetic Drugs, School of Pharmacy, Fudan University, Shanghai 200032, China
bBio-X Life Science Research Center, Department of Life Science and technology,
Shanghai Jiaotong University, Shanghai 200030, China
Received 15 June 2006; revised 21 August 2006; accepted 22 August 2006
Available online 18 September 2006
Abstract—A series of substituted acyl(thio)urea and 2H-1,2,4-thiadiazolo [2,3-a] pyrimidine derivatives were prepared and both of
their cell culture and enzymatic activity toward influenza virus were tested. Their in vitro neuraminidase inhibitory activities were in
good agreement with the corresponding activities in cultured cells and they were evaluated as potent neuraminidase inhibitors. Of
the analogues that demonstrated IC50s < 0.1 lM, 16 and 60 were further investigated as candidates with the most potential for
future development. The molecular docking work of the representative compound was described to provide more insight into their
mechanism of action and further rationalize the observations of this new series herein, which represents a novel class of highly
potent and selective inhibitors of influenza virus.
Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction
viral replicative cycles has revealed several potential
molecular targets (M2 proteins,3 endonuclease,4
hemagglutin,5 and neuraminidase6) that can be used
for anti-flu drugs design, there are currently only a few
licensed drugs available for influenza treatment. M2
inhibitors such as amantadine and rimantadine, which
act specifically against influenza A virus by blocking
the ion channel of the M2 protein, provide only limited
protection due to a narrow spectrum of activity.7 The
only neuraminidase (NA) inhibitors which received
FDA approval are zanamivir and oseltamivir. Though
zanamivir displays excellent antiviral activity when
administered intranasally, it is less effective when deliv-
ered systemically. It has very low oral bioavailability
and is rapidly eliminated by renal excretion.8 Oseltami-
vir is orally active, but it has been reported to cause
vomiting and nausea. Thus, there is still a great need
to design, screen, and identify new agents for the chemo-
therapy of influenza virus infection, and research for
more effective drugs following systemic administration
is a high priority.
Influenza virus commonly known as flu is the conta-
gious etiologic agent that causes an acute respiratory
infection, hence it has always been a major threat1 to
human health worldwide and cause for economic costs.
Each year about 120 million people in North America,
Europe, and Asia are infected, while in the USA alone
more than 200,000 patients are admitted to hospitals
because of influenza and there are approximately
30,000–40,000 influenza-related deaths.
Currently, the developments of vaccines for influenza
virus are of restricted usefulness as they are not suscep-
tible to the high mutability of the virus. Effective chemo-
therapy for influenza virus is also limited due to newly
discovered drug resistance2 in mutant strains. The obvi-
ous choice remains the development of effective drugs,
which are not susceptible to mutation. Although the
Keywords: Influenza virus; Neuraminidase inhibitor; Acyl(thio)ureas
derivatives; Pyrimidine derivatives; Antiviral activity.
During the past decade, thiourea derivatives have been
reported effective against HIV9,10 and to have bacterici-
dal action.11 Few studies have been done concerning
evaluations of substituted acyl(thio)urea and 2H-1,2,4-
*
Corresponding author. Tel.: +86 021 62932620; fax: +86 021
64321926; e-mail addresses: sunwillin@sjtu.edu.cn;
edu.cn
pzhou@shmu.
0968-0896/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.08.034

C. Sun et al. / Bioorg. Med. Chem. 14 (2006) 8574–8581
8575
F
O
F
S
O
describe further details of our new discovery: the in vitro
antiviral properties and neuraminidase inhibitory activ-
ities of this new compound series. The docking work
was also reported to rationalize the observations herein.
Me
C
N
N
O
Figure 1. Initial influenza inhibitor, 10.
2. Chemistry
thiadiazolo [2,3-a] pyrimidine for their antiviral activity.
In our early attempts to screen for activity against differ-
ent viruses, acyl(thio)urea 10 (Fig. 1) was revealed as a
weak inhibitor of influenza virus. As part of the pro-
gram to maximize this activity and identify more effec-
tive influenza virus inhibitors as well as searching for
the specific influenza target(s),12 a new class of substitut-
ed acyl(thio)urea and 2H-1,2,4-thiadiazolo [2,3-a]
pyrimidine derivatives (designated SATPs, series I –
XII) were prepared and a highly specific anti-influenza
virus activity in cell culture was discovered. Their in vitro
inhibitory activities of influenza neuraminidase were
also investigated and found to correlate well with their
antiviral efficacy in cell culture, thus they were evaluated
as effective neuraminidase inhibitors. In this study, we
Compounds were synthesized using methods presented
in Schemes 1–3. Scheme 1 presents the method used to
prepare the reported compound series III, IV, VI, and
VII. The intermediate N0-(4,6-disubstituted pyrimidine-
2-yl)-thiourea II was obtained by the reaction of a hot
aqueous solution of 2-amino-4,6-disubstitued pyrimi-
dine with potassium thiocyanate (KSCN). And 5-aryl-
2-furoic acid was allowed to react with SOCl2 to provide
the intermediate 5-aryl-2-furoyl chloride I. N0-(4,6-Di-
substituted pyrimidine-2-yl)-N-(5-aryl-2-furoyl) thio-
urea III was prepared by the reaction of the
intermediate I with II, followed by oxidizing cyclization
via dropwise treatment with bromine as oxidant to
produce the target thiadiazolo [2,3-a] pyrimidines IV.
Scheme 1. Reagents: (a) SOCl2; (b) HCl; (c) KSCN/CH3CN; (d) Et3N; (e) Br2.
Scheme 2. Reagents: (a) SOCl2; (b) HCl; (c) KSCN/CH3CN; (d) Et3N.

8576
C. Sun et al. / Bioorg. Med. Chem. 14 (2006) 8574–8581
Scheme 3. Reagents: (a) NH3(H2O); (b) (COCl)2; (c) 2-amino-4,6-disubtituted pyrimine; (d) H2NCONHC(CH3)3.
Table 1. SAR of polysubstituted pyrimidinyl acyl(thio)urea analogues
The substituted nicotinic acid was acylated with thionyl
dichlorideto give substituted nicotinyl chloride V.13,14
The preparation of N0-(4, 6-disubstituted pyrimidine-2-
yl)-N-substituted b-pyridinecarbonyl thiourea VI and
thiadiazolo [2,3-a] pyrimidines VII was quite similar to
the procedure for the acyl(thio)urea derivation III
described above.
X
N
O
S
R
O
N
Y
N
N
H
H
15-25
a,b
IC50
Compound
R
X
Y
Scheme 2 presents a synthesis of the reported compound
series X. The intermediate VIII (substituted formyl chlo-
ride) was prepared by acylation with substituted formic
acid using thionyl dichloride. And the intermediate IX
N0-(tert-butylaminocarbonyl)-thiourea was synthesized
by the reaction of a hot aqueous solution of tert-butylu-
rea containing hydrochloric acid with potassium thiocy-
anate. The target product X was synthesized by the
reaction of intermediate VIII with equal molar of IX
(Scheme 2) in (CH3CH2O)3N and acetonitrile, and the
mixture was refluxed for 3 h. After removal of the vola-
tiles in vacuo, the resulting precipitate was collected by
filtration and desiccated. The final compound X was
recrystallized from acetonitrile.
15
16
17
18
19
20
21
22
23
24
25
2-Cl
2-Cl
OEt
Me
1.65
0.08
0.32
1.77
OEt
OH
OMe
Cl
OEt
Me
OMe
Cl
2-Cl
2-Cl
2-Cl
2-Cl
14.5
>20
>20
1.66
>20
2.3
0.36
Cl
OMe
Cl
OEt
Me
Cl
2-Cl
4-NO2
4-NO2
4-NO2
4-NO2
OEt
OEt
OH
OEt
Me
Me
a IC50 of influenza (H1N1) virus, lM; MTS IC50 > 300 lM for all
compounds. Data are means of three assays.
b cf., the cell culture inhibitory activity of zanamivir; IC50, 0.05 lM.
analogues, except for 22 (22 vs 19). Other corresponding
4-nitrophenyl derivatives (23, 24) had no appreciable
activity toward influenza virus when compared with
their 2-chlorophenyl counterparts (15, 16). Interestingly,
the 4-nitrophenyl derivative 25 was an exception, which
was found to inhibit the influenza virus in vitro at
0.36 lM and showed activity comparable to that of 17.
Nonetheless, substitutions on the pyrimidine ring
profoundly affect the ability to inhibiting influenza virus
in vitro.
The method for the synthesis of compounds series XI
and XII is shown in Scheme 3. They are prepared start-
ing from the substituted formic acid, followed by treat-
ment with ammonia liquor and subsequent reaction with
(COCl)2. Then the substituted acylisocyanates were
converted to the target acylurea derivatives XI, XII by
reacting with 2-amino-4,6-substituted pyrimidines and
tert-butylurea, respectively.
3. Results and discussion
Introduction of the tert-butylaminocarbonyl group
afforded improved activity. Replacement of polysubsti-
tuted pyrimidine by tert-butylaminocarbonyl (26, 27)
further increased the potency 8- to 10-fold over ana-
logue 19, 24 (Table 2). With regard to the substituent
Q, replacement of 5-aryl-2-furyl by groups such as phen-
yl or methoxyl (28, 29) resulted in a slight decrease in the
potency, while other substituents (30 ꢀ 32) showed
activity that was equipotent to that of analogue 26.
Introduction of a specific functional cycloalkyl such as
chrysanthemoyl (33, 34) resulted in a significant increase
in potency, and analogue 34 was the most potent com-
pound in this series and inhibited influenza virus in vitro
at a concentration of 0.26 lM.
The SAR of different substituents in the pyrimidine ring
is shown in Table 1. Different substituents at the 4- or 6-
position in the pyrimidine ring had greatest influence on
activity. In this series of 5-aryl-2-furoyl thioureas bear-
ing a substituted pyrimidine ring, there was a large pref-
erence for electron-donating groups such as Me, OMe,
OEt, and OH in this portion of the molecule (Table 1;
15–18). Substitution by electron-withdrawing groups
such as Cl (19–20) led to poor activity (IC50 > 20 lM).
For instance, the acyl(thio)urea 16 was the most active
compound in the series with IC50 of 0.08 lM. However,
the 4-nitrophenyl substitution did not offer much
improvement
over
2-chlorophenyl
substituted

C. Sun et al. / Bioorg. Med. Chem. 14 (2006) 8574–8581
Table 2. tert-Butylaminocarbonyl acyl(thio)urea analogue SAR
8577
activity in this series was the test compound 44,
although it was 2-fold less potent than the correspond-
ing furyl substituted analogue 36. This result suggests
that a furan ring at this position provides improved
activity and should be included in the rational design
of new influenza virus inhibitors. Variation of the X, Y
substituents in the pyrimidine ring (39–43) and other
analogues with different aromatic substituents as Q
(45, 46, and 52–54) was also examined. They did not
achieve good efficiency against influenza virus except
for 53. Additionally, substitution of chloro- at the 2-
or 6-position in the pyridine ring was not a major deter-
minant in activity (39 vs 42) in this series. Analogues
sharing the same functional cycloalkyl substituents (50,
51) as 33 show similar inhibitory activity and replace-
ment of methyl by methoxy in the pyrimidine ring did
increase the potency slightly (50 vs 51) but was not sig-
nificant. Investigation of ethoxy substituents at the C-4,
6 positions of pyrimidine ring produced analogue 49
that has significant improvement in the potency, while
analogue with substituents as chlorine (48) was again
not encouraging.
O
S
O
Me
Me
Me
Q
N
H
N
N
H
H
26-34
d
Compound
Q
IC50
26
27
5-(2-Cl–Ph)–2-Furyl
5-(4-NO2–Ph)–2-Furyl
Ph
1.42
1.30
1.79
1.83
1.67
1.43
1.35
0.51
0.26
28
29
OMe
(2,4-Cl2–Ph)–OCH2
2,6-F2–Ph
30
31
32a S-(+)
33a cis-(ꢁ)
34a trans-(ꢁ)
2-Me-1-(4-Cl–Ph)–Pr
*CFPCb
*DCPCc
a Compounds 32–34 are pure enantiomers as indicated.
*CFPC, 3-(2-chloro-3,3,3-trifluoropropenyl)-2,2-dimethyl cyclo-
propyl.
*DCPC, 3-(2,2-dichloro ethenyl)-2,2-dimethyl cyclopropyl.
d IC50 of influenza (H1N1) virus, lM.
b
c
Other acyl(thio)ureas bearing a substituted pyrimidine
ring without 5-aryl-2-furyl were synthesized and resulted
in lowered potency (Table 3). This demonstrated a large
advantage in activity of the existence of a five-membered
ring (furan) over corresponding analogues without the
furan ring in most instances, and this trend continued
for other series of compounds. The compounds contain-
ing furan ring (35, 36) were 15-fold more potent than the
corresponding 6-chloro-3-pyridyl analogues (37, 38),
which showed no activity up to 20 lM. However, the
only analogue with no furan ring that retained good
The formation of 2H-1, 2, 4-thiadiazolo [2,3-a] pyrimi-
dine ring had relatively small effect on potency against
influenza virus (Table 4, 56, 57 vs 25, 23). However,
compounds 58 and 60 in this series inhibited the virus
with an IC50 of 0.35 and 0.09 lM, respectively.
Compound 58 showed improved potency over the corre-
sponding un-ringed analogues such as 15. The high
potency of 60 bearing the same ethoxy substituents as
the active analogues 16, 49 is further support of a pref-
erence for electron-donating groups such as ethoxy in
their activity.
Table 3. SAR of aryl and chrysanthemoyl Q groups
The necessity for a thiourea functionality (Table 5) was
also investigated. The corresponding ureas (63–69) were
found to be less effective than their thiourea counter-
parts were. Replacement of the thiourea bridge with
urea diminished or eliminated activity altogether in most
cases. For instance, 64, 65 had a 10-fold reduction in
activity compared to 31, 32. In this respect, it follows
that the thiourea bridge is an important contribution
to inhibitory activity against influenza virus.
X
O
S
N
Y
N
Q
N
N
H
H
35-54
a
Compound
Q
X
Y
IC50
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
5-(4-NO2–Ph)–2-Furyl
5-(2-Cl–Ph)–2-Furyl
6-Cl-3-py
OMe
OMe
OMe
OMe
OEt
Me
Me
Cl
1.22
1.29
Me
Cl
>20
Table 4. Effect of 2H-1,2,4-thiadiazolo [2,3-a] pyrimidine ring on
inhibition of influenza virus
6-Cl-3-py
6-Cl-3-py
>20
>20
OEt
OH
Me
OEt
Cl
6-Cl-3-py
2-Cl-3-py
8.58
7.19
>20
>20
2.59
>20
18.5
2.1
>20
X
O
Me
Q
S
N
2-Cl-3-py
2-Cl-3-py
OEt
OEt
OMe
OMe
OMe
Me
N
N
Y
N
2-Cl-3-py
3-py
Cl
55-62
OMe
OMe
Me
Cl
5,6-Cl2-3-Py
Ph
a
Compound
Q
X
Y
IC50
55
56
57
58
59
60
61
62
(2,4-Cl2–Ph)–OCH2
OMe
Me
OMe
OH
>20
2-Me-1-(4-Cl–Ph)–Pr
2-Me-1-(4-Cl–Ph)–Pr
*CFPCb
*CFPCb
Cl
5-(4-NO2–Ph)–2-Furyl
5-(4-NO2–Ph)–2-Furyl
5-(2-Cl–Ph)–2-Furyl
6-Cl-3-py
0.67
5.25
0.35
9.32
0.09
2.41
OEt
OMe
Me
OEt
OMe
Me
Me
Cl
0.31
0.97
0.58
1.36
5.1
OEt
OEt
OMe
OEt
OMe
OMe
OEt
Me
Cl
2-F-4-Cl–Ph
2-F-4-Cl-Ph
(2,4-Cl2–Ph)–OCH2
Me
OMe
OMe
6-Cl-3-py
2-Cl-3-py
3-py
OEt
OMe
OMe
OMe
1.89
>20
a IC50 of influenza (H1N1) virus, lM.
a IC50 of influenza (H1N1) virus, lM.
b
*CFPC, 3-(2-chloro-3,3,3-trifluropropenyl)-2,2-dimethyl cyclopropyl.

8578
C. Sun et al. / Bioorg. Med. Chem. 14 (2006) 8574–8581
Table 6. In vitro influenza neuraminidase inhibitory activity of the
representative compounds of PAFHs
Table 5. Thiourea versus urea comparison
O
O
a
a,b
IC50 (lM)
Compound
IC50 (lM)
Compound
W
Q
N
N
15
16
17
18
19
22
25
26
27
28
29
30
31
32
33
34
35
36
1.3
0.06
0.21
1.3
40
41
44
46
47
49
50
51
52
53
54
56
57
58
59
60
61
65
11.0
9.5
H
H
63-69
2.8
n.ac
1.6
a
Compound
Q
W
IC50
15.9
1.2
63
64
65
66
67
68
69
Ph
CONH(t-Bu)
CONH(t-Bu)
2-Me-1-(4-Cl–Ph)–Pr CONH(t-Bu)
>20
0.27
0.85
0.43
4.2
2,6-F2–Ph
14.1
15.5
0.25
1.2
2,6-F2–Ph
2,6-F2–Ph
2,6-F2–Ph
2,6-F2–Ph
4,6-(Me)2–2-Pymb
>20
>20
15.9
>20
1.1
1.5
4,6-(OMe)2–2-Pym
4-Cl-6-(OMe)–2-Pym
4-Cl-6-Me–2-Pym
1.1
1.6
1.5
1.4
0.69
4.2
a IC50 of influenza (H1N1) virus, lM.
1.2
1.3
b Pym, pyrimidinyl.
0.26
12.2
0.05
2.3
0.46
0.19
1.3
It was noted that the tert-butylacyl(thio)urea analogues
have improved water solubility while retaining good
lipophilicity, which may contribute partially to the in-
crease in its anti-viral potency. Although some of the
other SATPs have a relative low solubility (<10 mg/
mL at room temperature), their high lipophilicity is
expected to lead to an efficient penetration of the SATPs
through cellular membranes and biological barriers.
This property may be advantageous for this novel class
of influenza inhibitors, as it may facilitate the uptake of
the lipophilic SATPs into cells or biological compart-
ments where the virus accumulates.
1.3
18.7
a The concentration of inhibitor that requires to reduce the NA activity
by 50%.
b cf., neuraminidase inhibitory activity of zanamivir; IC50, 0.02 lM.
c n.a., not active in the concentration range tested (0.01–20 lM), other
compounds which have IC50 value > 20 lM are not shown.
spread of the virus from infected cells, and effectively re-
tard its propagation. They have a broader antiviral spec-
trum, better tolerance, proven efficacy in reducing
respiratory events, and less potential for rapid emer-
gence of clinically important resistance than is seen with
the M2 inhibitors. Therefore, the NAIs represent an
important advance over the M2 inhibitors in influenza
therapy. Although not currently approved for preven-
tion in some countries, they do seem to be effective for
chemoprophylaxis and could be used for long-term pro-
tection of those not responding to vaccine, or ineffective
due to antigenically novel viruses. In this regard, this
new compound series evaluated herein as neuraminidase
inhibitors has the potential to become the candidates for
future development of new effective drugs.
Mechanistic investigations into the mode of inhibition
led to enzymatic studies for the activity of this set of
compounds toward influenza neuraminidase, which is
an essential surface glycoprotein that is required for
the release of newly formed virus. It was found that
(Table 6) analogues which were highly active inhibitors
in cell culture also showed good inhibitory activities in
this enzymatic studies, and the concentrations at which
the 50% inhibition of neuraminidase activity was
achieved were in the same range as those observed in their
antiviral evaluations in cultured cells. Besides, Com-
pound 60 (IC50 = 0.05 lM) was again encouraging and
it shows the highest inhibitory activity toward the influen-
za neuraminidase. The fact that the neuraminidase inhib-
itory activities were in good agreement with the
corresponding antiviral efficacy in cultured cells suggests
that the SATPs belongs to a novel class of NA inhibitors.
3.1. Molecular docking
As an efficient approach for investigating protein–ligand
interactions, molecular docking plays a key role in ra-
tional drug design,15 especially when the crystal struc-
ture of the protein has been published. Since the
crystal structure of influenza virus neuraminidase com-
plexed with zanamivir is available, the interactions in
this crystal complex with the corresponding entry code
1a4g in PDB were first explored. For further SAR anal-
ysis, molecular docking has been performed to investi-
gate the key interactions responsible for the high
potency of this class of inhibitors as well as the binding
mode. FlexX, which is a reliable method in complex val-
idation tests among various docking algorithms,16 was
chosen as the docking protocol. FlexX scores were em-
ployed for evaluation, since they consist of a variety of
score terms as well as the relative affinity between the en-
zyme and inhibitors. Quite a few analogues reported
herein, including the most promising compound 60,
were docked to neuraminidase. Since most of the active
For comparison, both cell culture and enzymatic activity
of zanamivir, which is used as a positive control, were
also evaluated in the same system with all reported com-
pounds. It was noted that zanamivir exerted higher
potency than most of SATPs in inhibiting the enzymatic
activity. However, the influenza strain tested in these as-
says seemed to be insensitive to these well-known NA
inhibitors, as neither of zanamivir’s cell culture (Table
1) nor enzymatic activity (Table 6) was satisfying when
compared with the data that had been previously report-
ed in 1998,17 which also implies an increasing need for
developing novel flu inhibitors with new structures.
In general, the NA inhibitors known as NAI do not
inhibit virus replication but do prevent the release and

C. Sun et al. / Bioorg. Med. Chem. 14 (2006) 8574–8581
8579
compounds shared quite similar binding modes, the
docking results of the representative analogues are dis-
cussed and shown in Figure 3. Besides, their total FlexX
scores are also provided in Table 7 to quantify and
investigate the docking results.
binding mode (Fig. 3a). The acyl backbone of the li-
gand also makes charge-charge-based hydrogen bonds
to the guanidino groups of arginine residues 115, 291,
and 373 which are colored in green in the binding cav-
ity. The pyridine ring of 60 does not appear to make
strong interactions to any active-site residues, although
there may be a weak electrostatic interaction between
the ring nitrogen and Asp 148. In addition, the acyl
nitrogen on the backbone of 60 forms hydrogen bonds
to the hydroxyl of Tyr 408 colored in yellow, while the
oxethyl fragment improves the activity by making a
targeted hydrogen bond with Asn 293 colored in pur-
ple. These hydrogen bonds may further strengthen
the binding of the ligand, resulting in a strong network
of hydrogen-bonding interactions.
The crystal structure of neuraminidase-zanamivir com-
plex (Fig. 2) demonstrated the pattern of protein–ligand
interactions, which consist of strong charge-charge- and
charge-partial charge-based hydrogen bonds. The car-
boxylic acid of zanamivir makes charge-charge-based
hydrogen bonds to a cluster of three arginine residues,
115, 291, and 373, including a planar salt bridge to
373. In addition to the hydrogen bonds from the termi-
nal hydroxyl groups of the glycerol moiety to the car-
boxyl of Tyr 408, there is an electrostatic interaction
between the guanidino substituent and Asp 148. These
interactions were suggested to account for its inhibitory
potency and selectivity on neuraminidase17 from influ-
enza virus.
The docking results of the other three representative
analogues are shown in Figures 3b–d. Most of the ac-
tive compounds shared quite similar binding mode with
60. Furthermore, the total FlexX score of compounds
60, 28, 22, and 19 (Table 7) are in good agreement with
previous enzymatic inhibitory activity, since a lower
score value may indicate a more energy favorable state
and more strong affinity for the binding complex. It
was noted that compound 19 (Fig. 3d) did not bind
to the pocket at the entrance of the binding cavity sur-
rounded by the three arginine cluster, which is in
accordance with its relatively low neuraminidase inhib-
itory activity.
With respect to compound 60, the top 10 molecular
docking results of FlexX were analyzed, and they
bound in a closely similar manner to influenza neur-
aminidase with their scaffold well suited to interact
with the active site of the neuraminidase. Based on
the comparison with the crystal complex mentioned
above, though 60 shared different structural properties
with those of zanamivir, they seemed to bear a similar
The oxyethyl groups of 60 occupy a small pocket at the
entrance of the binding cavity, which may prevent the
substrate of NA from entering into the active site. This
may explain the high inhibitory activity of compound 60
toward the influenza neuraminidase. Moreover, this
docking result further demonstrates the importance of
these approaches to understanding the interactions be-
tween the enzyme and SATPs, which may provide more
insight into the mechanism of action of these newly dis-
covered inhibitors as well as some guidelines for struc-
tural optimization in drug design.
Table 7. The total FlexX scores of the representative analogues
Compound
Total FlexX score
19
22
28
60
ꢁ22.56
ꢁ33.62
ꢁ27.60
ꢁ36.23
4. Conclusion
In this paper, we have described the synthesis of a series
of substituted acyl(thio)urea and 2H-1,2,4-thiadiazolo
[2,3-a] pyrimidine. Evaluation of their in vitro antiviral
activity revealed compounds with high potent cellular
activity against wild-type influenza virus in cultured
MDCK cells. Meanwhile, the neuraminidase inhibitory
activities of these compounds correlate well with their
in vitro antiviral efficacy in cell culture, which reveals
that their specific influenza target was the neuraminidase
from influenza virus. Analogues 16 and 60 (Fig. 4) inhib-
ited the influenza virus with an IC50 of 0.08 and
0.09 lM, respectively, and these novel inhibitors were
investigated as candidate compounds with the most po-
tential for future development. This research leads to a
better understanding of SAR of influenza virus inhibi-
tors and thereby provides some insight into the rational
design of anti-flu agents and the discovery of new effec-
tive drugs.
Figure 2. Crystal structure of zanamivir bound in the active site of
influenza neuraminidase. Key amino acid residues forming the subsite
are highlighted and the major protein–ligand salt bridges, hydrogen-
bonding interactions are shown using Sybyl 6.9.

8580
C. Sun et al. / Bioorg. Med. Chem. 14 (2006) 8574–8581
Figure 3. Stereo view showing the binding of representative compounds to the active site of influenza neuraminidase. (a) The hydrogen-bonding
interaction between compound 60 and Arg 115, 291, 373 (atoms colored green) is illustrated by dashed lines, as are other associated hydrogen bonds
(see text); other amino acid residues near the active site are shown in line ribbons. (b–d) The docking poses and interactions of compounds 28, 22, and
19, respectively.
units (PFU)/well. Each compound was serially diluted
and tested against the virus to calculate its IC50 (Tables
1–5) and all values are means of at least three indepen-
dent experiments on different days. Additionally, gener-
al cytotoxicity was measured using the MTS cellular
toxicity assay18,19 to distinguish between specific anti-vi-
ral activity and non-specific host cell toxicity. All the
compounds reported herein have MTS IC50 values of
>300 lM and are considered to be noncytotoxic (these
data are not included in the tables).
5.2. Enzyme assay
Figure 4. Advanced potential development candidates 16, 60.
The in vitro enzyme assay is based on the method
reported by Vonitzstein et al.20 with some improvence.
The neuraminidase from the H1N1 strain of influenza
5. Experimental
was obtained by the method described by Tai et al.21
Values for the IC50 were measured via a spectrofluoro-
metric technique that uses 20-(4-methylumbelliferyl)-a-
D-acetylneuraminic acid as substrate. This substrate
was cleaved by neuraminidase to yield a fluorescent
product, which can be quantified. The assay mixture
contained inhibitors at various concentrations and en-
zyme in 33 mM MES (2-(N-morpholino) ethanesulfonic
5.1. In vitro assay in MDCK cells
The MDCK cells were grown in 96-well microtiter plates
in minimal essential medium (MEM) containing 10%
FBS; influenza Beijing/30/95 virus (H1N1) were ampli-
fied and titrated in MDCK cells at 50 plaque-forming

C. Sun et al. / Bioorg. Med. Chem. 14 (2006) 8574–8581
8581
acid) buffer, 4 mM CaCl2 at pH 6.5 (total vol-
ume = 100 lL). There action was started by the addition
of 10 lL of the substrate to a final concentration of
20 lM. After 15 min at 37 °C, 150 lL of 14 mM
NaOH/83% ethanol was added to 0.1 mL of the reaction
mixture to terminate the reaction. A blank was run with
the same substrate solution with no enzyme. Fluores-
cence was read using an Aminco–Bowman fluorescence
spectrophotometer (excitation, 355 nm, emission,
460 nm) and substrate blanks were subtracted from the
sample readings. The IC50 was calculated by plotting
percent inhibition versus the inhibitor concentration,
and determination of each point was performed in
triplicate.
Calcd (%) for C20H19ClN4O4S: C, 53.75; H, 4.26; N,
12.54. Found: C, 53.64; H, 4.35; N, 12.41.
5.6. 5,7-Diethoxyl-2-(6-chloro-3-pyridinecarbonylimino)-
2H-1,2,4-thiadiazolo[2, 3-a]pyrimidine (compound 50)
Yield: 72%, mp: 176–178 °C; IR (KBr plate, cmꢁ1): 1690
(C@O), 1630 (C@N); 1H NMR (DMSO-d6): dHppm
7.60–8.95 (m, 3H, pyridine-H), 5.65 (s, 1H, pyrimi-
dine-5-H), 1.16 (t, 3H, CH3), 4.18 (q, 2H, OCH2); Anal.
Calcd (%) for C20H17ClN4O4S: C, 47.43; H, 3.69; N,
18.45. Found: C, 47.51; H, 3.60; N, 18.36.
Acknowledgments
5.3. General chemistry methods
We thank Dr. Wei Li for revising the manuscript. We
also gratefully acknowledge research specialist Hai
Huang for valuable advice.
All solvents and reagents were commercially available
and used as received. Melting points were obtained in
open capillary tubes in a 8100 digital melting point
apparatus and are uncorrected. IR spectra were run as
KBr pellets on a Bruker Eqinox 55 FTIR spectrometer.
1 H NMR spectra were recorded on a Varian-XL-400
nuclear magnetic resonance apparatus, Elemental analy-
sis was performed on the Germany Elementar Vario EL
elemental analyzer.
References and notes
1. Monto, A. S.; Iacuzio, I. A.; LaMontague, J. R. Threat. J.
Infect. Dis. 1997, 176, 51–58.
2. Hayden, F. G.; Belshe, R. B.; Glover, R. D. N. Engl.
J. Med. 1989, 321, 1696–1702.
3. Hay, A. J.; Wolstenholme, A. J.; Skehel, J. J. EMBO
J. 1985, 4, 3021.
5.4. Molecular docking procedure
4. Hastings, J. C.; Selnick, H.; Wolanski, B.; Tomassini, J. E.
Antimicrob. Agents Chemother. 1996, 40, 1304.
5. Mammen, M.; Dahmann, G.; Whitesides, G. M. J. Med.
Chem. 1995, 38, 4179.
6. Colman, P. M. In The Influenza Viruses: Influenza Virus
Neuraminidase, Enzyme and Antigen; Krug, R. M., Ed.;
Plenum: New York, 1989; pp 175–218.
7. Erik De Clercq J. Clin. Virol. 2001, 22, 73–89.
8. Ryan, D. M.; Ticehurst, J.; Dempsey, M. H. Antimicrob.
Agents Chemother. 1995, 39, 2583–2584.
9. Venkatachalam, T. K.; Mao, C. Bioorg. Med. Chem. 2004,
12, 4275–4284.
10. Zhang, H.; Vrang, L.; Backbro, K. Antivir. Res. 1995, 28,
331–342.
11. Sarkis, G. Y.; Faisal, E. D. J. Heterocycl. Chem. 1985, 22,
137.
12. Sun, C. W.; Huang, H.; Feng, M. Q. Bioorg. Med. Chem.
Lett. 2006, 1, 162–166.
13. Krutosikova, A.; Kovac, J.; Sykora, V. Coll. Czech. Chem.
Commun. 1994, 39, 1892.
All calculations were carried out on a R14000 SGI Fuel
workstation using the software package SYBYL version
6.9 (Tripos, St. Louis, MO, USA). FlexX 1.11.1 within
SYBYL package was employed to explore the interac-
tion between the ligand and enzyme. The crystal struc-
ture of influenza virus neuraminidase complexed with
zanamivir was retrieved from PDB with corresponding
entry code 1a4g. Then the protein was prepared by
removing heteroatoms and water molecules and adding
all hydrogen atoms. The active site of 1a4 g was defined
˚
as residues with at least one atom within a radius of 9 A
from any atom of zanamivir. Then all compounds were
sketched using sybyl with all hydrogen atoms added and
Gasteiger-Huckel22 charges were used to parameterize
the compounds. Furthermore, their conformers with
low energy were ensured by RANDOM searches avail-
able in SYBYL. Then the compounds were docked to
neuraminidase from influenza virus by FlexX facilities.
FlexX scoring function was employed to evaluate the
docking pose of the compounds. The top 10 docking re-
sults of each compound were used for further discussion.
14. Oliveto, E. P. Chem. Heteerocycl. Compd. 1962, 14, 178–
179.
15. Drews, J. Science 2000, 287, 1960–1964.
16. Kontoyianni, M.; McClellan, L. M.; Sokol, G. S. J. Med.
Chem. 2004, 47, 558.
5.5. Physical data for two key compounds: N0-(4, 6-
diethoxylpyrimidin-2-yl)-N-[5-(2-chlorophenyl)-2-
furoyl]thiourea (compound 16)
17. Neil, R. T.; Cleasby, A.; Singh, O. J. Med. Chem. 1998, 41,
798–807.
18. Barltrop, J. A.; Owen, T. C. Bioorg. Med. Chem. Lett.
1991, 1, 611–614.
19. Mosmann, T. J. Immunol. Methods 1983, 65, 55–63.
20. Vonitzstein, L.M.; Wu, W.Y.; Phen, T.V.; Betty, J.E.P.
Appl. 92309684.6, 1993.
21. Tai, C. Y.; Escarpe, P. A.; Sidwell, R. W., et al.
Antimicrob. Agents Chemother. 1998, 42, 3234–3241.
22. Gasteiger, J.; Marsili, M. Tetrahedron 1980, 36, 3219.
Yield: 86%, mp: 175–177 °C; IR(KBr plate, cmꢁ1):
1632(C@O), 3195 (N@H), 1235(C@S) 1H NMR
(DMSO-d6): dHppm 7.24–7.86 (m, 6H, Ar–H), 6.15 (s,
1H, p0yrimidine-5-H), 12.08 (s, 1H, N0–H), 12.46 (s,
1H, N –H), 1.22 (t, 6H, CH3), 4.08 (q, 4H, OCH2) Anal.