Antimicrobial and cytotoxic arylazoenamines. Part III: Antiviral activity of selected classes of arylazoenamines

Article abstract of DOI:10.1016/j.bmc.2008.08.028

Eighty-five arylazoenamines, characterized by different types of aryl and basic moieties, have been synthesized and evaluated in cell-based assays for cytotoxicity and antiviral activity against a panel of ten RNA and DNA viruses. The most commonly affected viruses were, in decreasing order, CVB-2, RSV, BVDV, YFV, and Sb-1; the remaining viruses were either not affected (HIV-1, VSV, and VV) or susceptible only to a very few compounds (Reo-1 and HSV-1). Thirty-five compounds exhibited high activity, with EC₅₀ in the range 0.8-10 μM, and other 28 compounds had EC₅₀ between 11 and 30 μM, thus indicating that the arylazoenamine molecular pattern is an interesting novel pharmacophore for antiviral agents against ssRNA viruses. Moreover, some compounds (as 28, 32, 42, and 53) appear of high interest, being devoid of toxicity on the human MT-4 cells (CC₅₀ > 100 μM). A ligand-based computational approach was employed to identify highly predictive pharmacophore models for the most frequently affected viruses CVB-2, RSV, and BVDV. These models should allow the design of second generation of more potent inhibitors of these human and veterinary pathogens.

Full text of DOI:10.1016/j.bmc.2008.08.028

Bioorganic & Medicinal Chemistry 16 (2008) 8447–8465
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Antimicrobial and cytotoxic arylazoenamines. Part III: Antiviral activity
of selected classes of arylazoenamines q
c
Michele Tonelli a, Vito Boido a, Caterina Canu a, Anna Sparatore b, Fabio Sparatore a, , Maria Silvia Paneni ,
Maurizio Fermeglia c, Sabrina Pricl c, Paolo La Colla d, Laura Casula d, Cristina Ibba d, David Collu d,
Roberta Loddo d
*
a Dipartimento di Scienze Farmaceutiche, Università di Genova, Viale Benedetto XV 3, 16132 Genova, Italy
b Istituto di Chimica Farmaceutica, Università di Milano, Viale Abruzzi 42, 20131 Milano, Italy
c Dipartimento di Ingegneria Chimica, dell’Ambiente e delle Materie prime, Università di Trieste, Via Valerio 10, 34127 Trieste, Italy
d Dipartimento di Scienze e Tecnologie Biomediche, Università di Cagliari, Cittadella Universitaria, S.S. 554, Km 4.500, 09042 Monserrato, Cagliari, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Eighty-five arylazoenamines, characterized by different types of aryl and basic moieties, have been syn-
thesized and evaluated in cell-based assays for cytotoxicity and antiviral activity against a panel of ten
RNA and DNA viruses.
The most commonly affected viruses were, in decreasing order, CVB-2, RSV, BVDV, YFV, and Sb-1; the
remaining viruses were either not affected (HIV-1, VSV, and VV) or susceptible only to a very few com-
pounds (Reo-1 and HSV-1).
Received 19 May 2008
Accepted 13 August 2008
Available online 15 August 2008
Keywords:
Arylazoenamino compounds
Antiviral activity
RNA and DNA viruses
Pharmacophore models
Thirty-five compounds exhibited high activity, with EC50 in the range 0.8–10
lM, and other 28 com-
pounds had EC50 between 11 and 30 M, thus indicating that the arylazoenamine molecular pattern is
l
an interesting novel pharmacophore for antiviral agents against ssRNA viruses. Moreover, some com-
pounds (as 28, 32, 42, and 53) appear of high interest, being devoid of toxicity on the human MT-4 cells
(CC50 > 100 lM).
A ligand-based computational approach was employed to identify highly predictive pharmacophore
models for the most frequently affected viruses CVB-2, RSV, and BVDV. These models should allow the
design of second generation of more potent inhibitors of these human and veterinary pathogens.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
trimethylindoline,4 as well as by reacting glyoxal monophenylhyd-
razone with secondary amine (piperidine5 and pyrrolidine6). More-
Since 20 years, we have shown that by reacting
a
-(tert-amino)-
over, unstable arylazoenamines were supposed to be formed, at
first, in the reaction of diazonium salts with unsaturated amines,
that, finally, gives rise to b-phenylhydrazones of substituted
glyoxals.7
ketones with excess of arylhydrazines, in the presence of acids, sta-
ble, red-orange arylazoenamines are formed, instead of (or in
addition to) the expected indoles.1,2
It was assumed that the essential step for the formation of
this kind of compounds is the tautomerization of the initially
formed arylhydrazone into the corresponding arylhydrazoen-
amine, which is then oxidized by the excess of arylhydrazine,
with simultaneous formation of an arylamine and ammonia
(Scheme 1).
To the best of our knowledge, biological activities of arylazoen-
amines resulted unexplored until 1990,8,9 though they can be con-
sidered either as sub-structures of ortho (and para) aminoazo
compounds, or as vinylogues of aryltriazenes (either linear or cyc-
lic), many of which are endowed with either antimicrobial, anti-
parasitic or cytotoxic activities (Fig. 1).
Arylazoenamino compounds have been already obtained in the
past through different pathways, as by reacting arylhydrazines
with 2-(bromomethyl)-1-methylquinolinium bromide, or with
1,2,3,3-tetramethylindoleninium iodide,3 by coupling diazonium
salts with the so-called methylene bases as 2-methylene-1,3,3-
Representative compounds of the first type are the following:
2,6-diamino-3-phenylazo pyridine (pyridium), once used as uri-
nary antiseptic and analgesic; 5-(4-amino-1-naphthylazo)uracil,
highly effective against Schistosoma mansonii10,11; 1-[3-(4-pheny-
lazo-5,6,7,8-tetrahydronaphth-1-yl)amino]-propylpiperidines12
and 1-[3-(4-phenylazo-5,6,7,8-tetrahydronaphth-1-yl)aminolup-
inane13 very active against Mycobacterium tuberculosis; and Congo
Red, recently shown to be able to inhibit the deposition of prion
protein in amyloid plaques.14,15
q
For notes I and II, see Refs. 8 and 9.
* Corresponding author. Tel.: +39 010 512721; fax: +39 010 3538358.
E-mail address: sparator@unige.it (F. Sparatore).
0968-0896/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.08.028

8448
M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
NHNH2
R'
NH N=C CH
R"
R'
C=
R
R'"
R'"
N
NH NH
C
N
R""
R""
R
R
R"
R
R'
_
N
R'
N
C
R'"
R'"
+
N=N
N
+NH3
NH2
C
N
C
C
R"
+
R""
R""
R
R"
R
Scheme 1.
H2N
N=N
N
N H
NH2 R-NH
N=N
O
R = H; CH2CH2N(C2H5)2
NH
O
;
N=N
CH2
NH
R
R = CH2 CH2 CH2
N
H
N
NH2
H2N
N=N
N=N
SO2ONa
R
SO2ONa
Cl
N=N
N
R
CH3
CH3
H2N
N
N=N
N
Et
N
H2N
O
C
O
NH2
N
N
R
C
N
N
NH2
N
N
CH3
N
N
CH3
N
N
H
N
N=N
N
O
C
O
CH3
R = H; CONHCH 2CH2
CH3
CH3
N
O
N
O
F
F
O
N=N
C
NH
F
N=N C
NH
F
F
N
N
N
C2H5
CH2
N
N
N
H
N
H
N
CH2)n
Figure 1. Some biologically active azocompounds reminding arylazoenamines.
Triazenes as a class of compounds16 hold an important position
as carcinogens and/or anticancer agents: 3,3-dimethyl-1-phenyl-
triazene and, even better, dacarbazine17 are able to methylate
DNA, thus the former resulted in a powerful carcinogen, while
the latter is in clinical use for the treatment of malignant mela-
noma and Hodgkin’s disease. A structurally more complex triazene

M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
8449
combines antitumor potential with inhibition of Pneumocystis cari-
2. Chemistry
nii.18 Temozolamide,19 closely related to dacarbazine, is currently
used to treat malignant glyomas. A recent development of this cyc-
lic acyltriazene is represented by the synthesis of benzotetrazepi-
nones,20–24 that are very promising agents against breast cancer.
It is worth noting that these compounds are weak alkylating agents
and may damage DNA by a novel mechanism.
The syntheses of arylazoenamines of general structures A–D
were carried out as indicated in Scheme 1 by heating the ethanolic
solution of relevant carbonyl derivative with the suitable arylhydr-
azine (or its hydrochloride), concd hydrochloric acid, and 85%
phosphoric acid in a molar ratio of 1:2:3:5. Yields were fairly good
except when the arylhydrazine is unsubstituted or is substituted
with electron-releasing groups. In the last cases a small amount
of the corresponding indole was formed and separated by column
chromatography.
Starting from 1-methyl-3-piperidone two isomeric compounds,
which can be separated by column chromatography, are simulta-
neously formed (Scheme 2). In all cases, the higher-melting isomer
shows the lower Rf in the T.L.C. As previously demonstrated,9 the
low-melting isomer corresponds to the expected 3-arylazo-1-
methyl-1,4,5,6-tetrahydropyridine (C), while the high-melting
one corresponds to 2-arylazo-methylene-1-methylpyrrolidine
(D). The mechanism of this ring contraction has already been
discussed.9
The two isomeric compounds could also be obtained starting
from the corresponding 2-formyl-1-methylpyrrolidine arylhydraz-
one, but this procedure is more expensive and does not offer any
advantage.
The required octahydro-2H-quinolizin-1-one, octahydro-indoli-
zin-8-one, and 1-methyl-3-piperidone were prepared as described
in the literature.34–36 1-Benzyl-3-piperidone is commercially
available.
Arylazoenamines could be also seen as simple azocompounds,
whose reactivity is somewhat enhanced by the conjugation with
the ene double bond, and as such they could be related to the even
more activated acyldiazenes. Simple aromatic azocompounds and
acyldiazenes are endowed with a variety of biological activities25,26
(mainlyantifungaland cytotoxic), generallyrelatedtotheirabilityto
oxidize biological thiols, though other mechanisms should be sup-
posed in some cases. Antifungal27 and antiproliferative28 activities
have been found in a peculiar class of cyclic azocompounds (3,3-
disubstituted-3,4-dihydro-1,2,4-benzotriazines29), which more-
over can display several other pharmacological activities depending
on the substituents that are present on the bicyclic system.30,31
On this base, some years ago, thirty arylazoenamines were pre-
pared and tested8,9 for antimicrobial activity against eight Gram+
and Gramꢀ bacterial strains and against eight yeast like fungi.
While the antibacterial activity was very poor [MIC in the range
from 1 to 2 mg/ml (norfloxacine: MIC = 4–8
occasional values of 512 g/ml against Pseudomonas aeruginosa],
a moderate antifungal activity was displayed by most compounds,
with MIC from 32 to 512 g/ml against the whole set of tested fun-
gi (miconazole nitrate: MIC = 0.25–2 g/ml).
lg/ml) and only two
l
l
l
Moreover, a few of these compounds, which were additionally
tested against bovine viral diarrhoea virus (BVDV) and coxackie
virus type B2 (CVB-2), exhibited an interesting antiviral activity.
Therefore many new arylazoenamines have now been synthe-
sized, and together with part of the previously described com-
pounds were evaluated in cell-based assays for cytotoxicity and
antiviral activity against a large panel of RNA and DNA viruses.
On the whole, 85 arylazoenamines have been considered, which
may be allotted in seven groups (A–G) in relation to the nature of
the basic moiety (Fig. 2).
Compounds of structure E were obtained by coupling aryldiazo-
nium salts with 2-methylene-1,3,3-trimethylindoline as indicated
by König4 (Scheme 3).
Finally, compounds of structures F and G were prepared by con-
densing glyoxal monoarylhydrazone with the suitable secondary
cyclic amine or piperazine (all commercially available) as indicated
by Severin et al.5,6 (Scheme 3).
3. Results and discussion
3.1. Biological activity—general considerations
Twenty-nine out of these compounds have been already de-
scribed: nine of structure A,2,8 ten of structure C,9 six of structure
D,9 two of structure E3,32 and two of structure F5,6; one additional
compound of structure F (phenylazoethenemorpholine) is indi-
cated by Sommer33 as already known, but we failed to find out
any data concerning its preparation and characterization.
Title compounds were evaluated for antiviral activity against
viruses representative of two of the three genera of Flaviviridae
family, that is, Flaviviruses (yellow fever virus, YFV) and Pestivirus-
es (bovine viral diarrhoea virus, BVDV), as Hepaciviruses can
N=N Ar
N=N Ar
N
N
A: 1-17
B: 18-23
N=N Ar
CH3
CH3
N
N
R
CH N=N Ar
N
CH
=N Ar
N
R
CH3
D: 44-63
C: 24-43
E: 64, 65
X
CH=CH N=N Ar
N
Heteroaryl/Aryl
N
CH=CH N=N Ar
N
F: 66-73
G: 74-85
Figure 2. Structures of the investigated compounds.

8450
M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
R
R
NH
N
N
N
N
CH3
N
C
CH3
R
or
R
N
N
CH
N
NH
N
CH
N
CH3
D
CH3
Scheme 2.
CH3
CH3
CH3
CH N=N
+
-
CH3
CH2
R
N2
+ Cl
N
N
R
CH3
CH3
R
+
NH N=CH
CH=O
N=N CH=CH
N
HN
X
X
Scheme 3.
hardly be used in routine cell-based assays. Compounds were also
tested against representatives of other virus families. Among
ssRNA+ were a retrovirus (human immunodeficiency virus type 1,
HIV-1) and two Picornaviruses (coxsackie virus type B2, CVB-2
and poliovirus type-1, Sabin strain, Sb-1); among ssRNAꢀ were a
Paramixoviridae (respiratory syncytial virus, RSV) and a Rhabdo-
viridae (vesicular stomatitis virus, VSV) representative. Among
double-stranded RNA (dsRNA) viruses was a Reoviridae represen-
tative (Reo-1). Two representatives of DNA virus families were also
included: herpes symplex type-1, HSV-1 (Herpesviridae), and vac-
cinia virus, VV (Poxviridae).
On the whole, 35 compounds have shown an EC50 6 10 lM
against at least one virus, 28 had EC50 between 11 and 30 lM,
and only 17 had a EC50 in the range 31–100 lM; thus, the aryl-
azoenamine molecular pattern is, indeed, an interesting novel
pharmacophore for antiviral agents against ssRNA viruses.
Cytotoxicity and antiviral activities of test and reference com-
pounds are reported in Tables 2–5.
3.2. Cytotoxicity on host cells
Test compounds showed different degrees of cytotoxicity
against the confluent cell monolayers (in stationary growth) used
to support the multiplication of the different viruses.
The most susceptible to toxicity were the exponentially grow-
ing lymphoblastoid human cells (MT-4) used to grow HIV-1,
while the non-human host cell lines exhibited a progressively re-
duced sensitivity in the order MDBK > BHK > Vero-76. Indeed,
NM-108 (20-b-methylguanosine), 6-azauridine, AZT (30-azido-
thymidine), and ACG (acyclo-guanosine) were used as reference
inhibitors of ssRNA+, ssRNAꢀ, HIV-1, and DNA viruses, respectively.
Interestingly, 80 over 85 tested compounds exhibited antiviral
activity against one or more viruses: in particular, 27 compounds
exhibited a selective activity against a single virus, while 33, 15,
4, and 1 were, respectively, active against two, three, four, and five
viruses. None of the test compounds inhibited the multiplication of
HIV-1, VSV, and VV, and only three compounds exhibited a modest
activity against Reo-1. Similarly, only four compounds displayed
moderate activity against HSV-1, but it must be noted that, be-
cause of shortage of products, only 50 over 85 compounds were
tested against it. In both cases, the observed activity was secondary
to the activity on other viruses.
only 25.9% of compounds showed no toxicity (CC50 P 100
against MT-4 versus 47.1% on Vero-76 cells; on the contrary
45.9% of compounds exhibited a CC50630 M for MT-4 compared
with 10.6–11.8% in the case of BHK and Vero-76 cells, respec-
tively. CC50 so low as 2 M were found for MT-4 cells, while
the lowest CC50 values for BHK and Vero-76 were, respectively,
10 and 11 M.
lM)
l
l
l
Cytotoxicity is clearly influenced by the nature of both the ami-
no heads and the arylazo moieties; however, the former seem to be
more important than the latter. Considering together the four cell
lines, it is observed that the azoethene derivatives of structure F
On the other hand, an increasing number of compounds exhib-
ited antiviral activity, sometimes of very high level, against, in the
order, Sb-1 (12), YFV (16), BVDV (28), RSV (34), and CVB-2 (62) (Ta-
ble 1). Twenty-eight of the 62 active compounds inhibited the mul-
tiplication of CVB-2 more strongly than the reference compound
and G are the most toxic, exhibiting for about 54% a CC50 6 30
while the tetrahydropyridines of structure C are the least toxic,
with only 2.5% of compounds with CC50 6 30 M. Viceversa,
ꢁ61% of compounds of the last type showed no toxicity
(CC50 P 100 M), while only 15% of compounds F and G have a
CC50 in this range. The remaining types of compounds (A, B, D,
lM,
NM-108 (EC50 = 20
compounds exhibited a EC50 value comparable with that of the
respective reference drug, NM-108 (EC50 = 1.7 M) and 6-azauri-
dine (EC50 = 1.2 M).
lM). Concerning BVDV and RSV, only a few
l
l
l
l

M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
8451
Table 1
Number of active compounds on susceptible viruses and range of their EC50
Virus
No. of active over
No. of active compounds (range of EC50 lM)
85 tested compoundsa
Sb-1b
YFVb
12
16
28
34
62
2(8)
6(12–27)
4(12–23)
11(11–21)
15(11–30)
1(45)
3(53–60)
4(7–10)
4(6–10)
6(6–9)
3(36–48)
2(40, 46)
1(40)
5(65–100)
6(60–100)
6(60–100)
11(51–100)
BVDVb
RSVc
5(0.8–5.5)
6(0.9–5)
10(0.9–5)
CVB–2b
10(6–10)
18(12–30)
13(32–50)
a
Compounds with EC50 > 100 lM or higher than CC50 are considered inactive.
Single-stranded, positive RNA virus.
Single-stranded, negative RNA virus.
b
c
and E) stay in the middle between the above extremes, with only
small differences among them.
The high cytotoxicity observed for a few compounds, particu-
larly against MT-4 cells (i.e., 7, 12, 55, 69, with CC50 = 2–4 lM),
while lowering their interest as antivirals, may warrant deeper
investigations to obtain novel antiproliferative agents.
Compounds endowed with the highest activity against a given
virus may also be selective for it (26: EC50 = 8 M for Sb-1; 32:
EC50 = 0.8 M for BVDV; 84: EC50 = 0.9 M for RSV), or may exhibit
simultaneously good activity on one or two other viruses, as com-
pound 75 (EC50 = 0.9 M for CVB-2) and 47 (EC50 = 7 M for YFV)
that inhibit also, respectively, RSV and BVDV and CVB-2 with
EC50 in the range 5–16 M.
l
l
l
l
l
l
3.3. Structure–activity relationships
Generally, compounds acting on three or more viruses display
only a moderate or even a modest degree of activity. Interesting
exceptions are represented by compounds 28, 47, 76, 83, and, par-
Also the antiviral activity, as already seen for cytotoxicity, is
influenced by the nature of the amino head and the arylazo moi-
ety; while the former seems to mainly address the activity versus
the different types of viruses, the latter principally regulate the
intensity of the antiviral action. However, in some cases the defini-
tion of the real role of each structural feature may result very dif-
ficult and even puzzling. Indeed, the shift of a substituent from a
position to another of the aromatic moiety may produce a qualita-
tive more than a quantitative variation of activity. Moreover, not
all the possible combinations of the basic heads and aromatic moi-
eties could be explored.
Among the five compounds completely devoid of antiviral
activity, the lack of activity can be related either to the nature of
the basic head (indoline: 64, 65), or to the aromatic moiety (3,5-
bis-trifluoromethylphenyl: 36, 55; pentafluorophenyl: 41), since
the corresponding aromatic moieties of the former and the basic
heads of the latter are commonly present in very active
compounds.
Despite these shortcomings, some general rules can be drawn.
The arylazohexahydroquinolizines A and the arylazoethenam-
ines F definitely appear as the most suitable structures for activity
against CVB-2; indeed, only one compound (16) of the first type
and no one of the second failed to inhibit its multiplication.
Also most of the structurally related arylazohexahydroindoli-
zines B and arylazoethenepiperazines G are active against CVB-2,
though acting also against Sb-1 and RSV, respectively. About a half
of compounds active against RSV is found in groups F and G; the
remaining being evenly distributed among the others.
ticularly, 81, that exhibited a EC50 = 7 lM against CVB-2, YFV and
RSV. Most of the nineteen large spectrum compounds belong
(13/19) to groups C and D; the remaining are found in groups G
(3/19), A (2/19), and B (1/19).
3-(4-Chlorophenylazo)-1-methyltetrahydropyridine (28) is ac-
tive on Sb-1, CVB-2, RSV and Reo virus, with EC50 = 8, 5, 20 and
32 lM, respectively. The shift of chlorine from para to ortho posi-
tion (26) produces the loss of the activity on CVB-2, RSV and Reo
viruses, leaving unchanged that on Sb-1; however when the chlo-
rine is placed on meta position (27) the activity on CVB-2 and RSV
is maintained (though reduced), but that on Sb-1 and Reo is lost,
while appears a good activity on BVDV. The activity on Sb-1 is lost
also when the 4-chlorophenylazo substituent of compound 28 is
attached on the methylenepyrrolidine moiety (46) that is isomeric
to the tetrahydropyridine ring of 28. Curiously, compound 46
exhibits the same spectrum of activity (CVB-2, RSV, and BVDV) of
the meta chloro derivative 27. Differently from what is observed
for the isomeric 4-chlorophenylazo derivatives 28 and 46, the iso-
meric 3-nitrophenylazo compounds 32 and 50 maintain a high and
selective activity against BVDV.
The varied activities shown by the four isomeric compounds 26,
27, 28, and 46 demonstrate the intriguing complexity of the struc-
ture–activity relationships for this kind of compounds. Therefore,
in an attempt to reach a better understanding of the structural
requirements for antiviral activity on the 85 arylazoenamines A–
G, a ligand-based computational approach was employed in order
to derive a highly predictive pharmacophore model for the most
frequently affected viruses CVB-2, RSV, and BVDV. Results of this
study are discussed in the next section.
Interestingly, no one of compounds of F and G groups was active
against BVDV.
The arylazotetrahydropyridines C and the arylazomethylene-
pyrrolidines D share in almost equal measure the activity against
CVB-2 and BVDV or on both types of viruses. It is worth noting that
the exchange of the methyl group with a benzyl abolishes the
activity against BVDV, leaving or enhancing that against CVB-2.
Sole activity on YFV is never found in any tested compound, but
in association with activity on CVB-2 and/or BVDV is most com-
monly found among arylazoenamines F and arylazoethenepipera-
Despite the high antiviral activity of many compounds, the cor-
responding selectivity indexes (SI) are most commonly below 10,
due to the high toxicity on the host cells often observed. Neverthe-
less, 17 compounds exhibited a selectivity index in the range from
>10 to >125, for at least one of the viruses CVB-2, BVDV, RSV, and
Sb-1. The reference compound NM-108 exhibits a SI >5 and >58.8
for CVB-2 and BVDV viruses, respectively.
Thus, taking into account the EC50 and SI values, the best com-
pounds can be arranged in the following order of decreasing po-
tency and selectivity; for CVB-2: 75 > 66 P 53 P 28; for BVDV:
32 > 22 > 42 = 50 > 16; for RSV: 84 > 23, and for Sb-1: 26 P 28. In
these sequences even the least potent compounds compare well
with the reference compounds. It must be observed, however, that
in view of a therapeutical application the cytotoxicity on human
zines G. Activity against YFV is shown by only
compounds of group A, but by none of group B.
1 over 17
Selective activity on Sb-1 is shown by a single compound of
group C, while activity on Sb-1 associated with activity on CVB-2
and BVDV is exhibited by several compounds of the same and
other groups, especially by hexahydroindolizines B.

Table 2
Cytotoxicity against MT-4, MDBK, BHK and Vero-76 cell lines and BVDV, YFV, CVB-2, RSV, and Sb-1 inhibitory activity of arylazoenamines of structures A and B
N=N Ar
N=N Ar
N
N
A: 1-17
B: 18-23
M) YFV EC50
a
b
c
d
e
f
g
h
i
Compound
Aryl=
MT-4 CC50
(
lM)
MDBK CC50
(lM)
BVDV EC50
(lM)
BHK CC50
(
l
(l
M)
Vero-76 CC50
(
lM)
CVB-2 EC50
(l
M)
RSV EC50
(
lM)
Sb-1 EC50 (lM)
1
2
3
4
5
6
7
8
4-F-Phl
2-Cl-Ph
3-Cl-Ph
4-Cl-Ph
3-Br-Ph
4-Br-Ph
3-CF3-Ph
3-NO2-Ph
4-NO2-Ph
3,4-DiCl-Ph
3,5-DiCF3-Ph
3-CF3-4-F-Ph
3-CF3-4-Cl-Ph
3-CF3-4-Br-Phm
3-NO2-4-Cl-Ph
1-Naphthyl
7-Chloro-1-
quinolyl
94
78
16
42
9
56
>100
20
60
10
59
8
>100
>100
30
31
>100
35
>100
18
>100
40
>56
>100
>20
>60
>10
>59
>8
>100
93
46
69
49
62
40
87
>100
54
19
60
41
>100
10
100
>93
>46
>69
>49
P62
>40
>87
P100
>54
>19
>60
>41
80
>100
>100
60
80
>100
92
27
32
22
19
18
10
21
71
26
38
41
45
64
42
51
>90
51
>100
60
18
>100
>100
>60
>80
>100
>92
25
>100
>92
>100
>90
>100
>100
>100
>100
>100
>100
75
27
4
>100
90
>100
20
37
91
35
34
3
19
34
33
12
15
>100
86
9
100
>100
100
>100
>100
100
100
90
>100
>100
>100
>100
>100
25
>100
>90
>85
10
11
12
13
14
15
16
17
>30
>31
>100
>35
61
>18
6
5
>10
>20
>48
20
48
>90
>85
85
18
19
20
21
22
23
4-Cl-Ph
>100
>100
26
30
18
52
>100
2
30
68
67
21
>30
>68
>67
>21
3.5
>100
1.7
>100
34
95
61
51
51
>100
90
>34
>95
>61
>51
>51
>100
1
75
10
54
30
9
44
>100
20
>75
>100
>100
>100
>90
7
20
60
45
12
>90
>100
>100
>20
3-NO2-Ph
3,4-DiCl-Ph
3-CF3-4-Cl-Phn
1-Naphthyl
1-Phthalazyl
>100
>100
100
90
>100
>100
20
91
>100
>100
>100
NM-108 (20-b-methylguanosine)
>100
1.2
6-Azauridine
>100
26
>20
a
Compound concentration (
Compound concentration (
Compound concentration (
Compound concentration (
Compound concentration (
Compound concentration (
Compound concentration (
Compound concentration (
Compound concentration (
l
l
l
l
l
l
l
l
l
M) required to reduce the viability of mock-infected MT-4 (CD4+ human T cells containing an integrated HTLV-1 genome) cells by 50%, as determined by the MTT method.
M) required to reduce the viability of mock-infected MDBK (bovine normal kidney) cells by 50%, as determined by the MTT method.
M) required to achieve 50% protection of MDBK cells from BVDV (bovine viral diarrhoea virus) induced cytopathogenicity, as determined by the MTT method.
M) required to reduce the viability of mock-infected BHK (hamster normal kidney fibroblast) monolayers by 50%, as determined by the MTT method.
M) required to achieve 50% protection of BHK cells (kidney fibroblast) from YFV (yellow fever virus) induced cytopathogenicity, as determined by the MTT method.
M) required to reduce the viability of mock-infected VERO-76 (monkey normal kidney) monolayers by 50%.
M) required to reduce the plaque number of CVB 2 (coxsackie virus B2) by 50% in VERO-76 monolayers.
M) required to reduce the plaque number of RSV (respiratory syncitial virus) by 50% in VERO-76 monolayers.
M) required to reduce the plaque number of Sb-1 (poliovirus type-1, Sabin strain) by 50% in VERO-76 monolayers.
b
c
d
e
f
g
h
i
l
Active also on Reo-1 virus (EC50 = 34
Active also on HSV-1 virus (EC50 = 25
Active also on HSV-1 virus (EC50 = 33
lM).
m
n
l
M).
l
M).

Table 3
Cytotoxicity against MT-4, MDBK, BHK, and Vero-76 cell lines and BVDV, YFV, CVB-2, RSV, and Sb-1 inhibitory activity of arylazoenamines of structure C
N=N Ar
N
R
C: 24-43
a
b
c
d
e
f
g
h
i
Compound
Aryl=
MT-4 CC50
(l
M)
MDBK CC50
(l
M)
BVDV EC50
(lM)
BHK CC50
(l
M)
YFV EC50
(lM)
Vero-76 CC50
(l
M)
CVB-2 EC50
(lM)
RSV EC50
(
lM)
Sb-1 EC50 (lM)
R = CH3
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Phenyl
4-F-Ph
2-Cl-Ph
>100
P100
>100
>100
>100
>100
83
>100
>100
>100
>100
43
>100
>100
62
50
71
>100
48
>100
87
>100
>48
100
12
>58
19
60
90
0.8
11
>100
7
>64
>100
>28
40
>61
>100
5
>100
90
>100
85
>100
>100
>100
>100
95
>100
>100
38
82
80
44
>100
>100
>100
>100
>100
>90
>100
>85
>100
71
100
>100
>95
>100
>100
>38
>82
>80
>44
>100
>100
>100
>100
P100
77
>100
75
>100
53
>100
34
5
27
>100
>100
>100
>100
>71
14
>100
35
34
75
30
>100
40
20
>100
>77
8
>75
8
>100
>100
>100
>100
>100
25
>100
>100
>70
3-Cl-Ph
4-Cl-Phl
58
90
3-Br-Ph
>100
>100
>100
>100
62
>100
73
64
>100
28
>100
61
>100
>100
>100
100
>100
>100
>100
71
>100
>100
70
90
80
>100
>100
P100
18
18
4-Br-Phm
3-CF3-Ph
>100
>100
>100
>71
>100
>100
>70
>90
>80
>100
100
>100
3-NO2-Ph
4-NO2-Ph
4-CH3-Ph
3,4-DiCl-Ph
3,5-DiCF3-Ph
3-CF3-4-F-Ph
3-CF3-4-Cl-Ph
3-CF3-4-Br-Ph
3-NO2-4-Cl-Ph
PentaF-Ph
1-Naphthyl
>90
>80
>100
>100
>100
40
>100
>100
76
>100
>100
R = CH2-C6H5
43
NM-108 (20-b-
3-CF3-4-Cl-Ph
43
>100
28
>100
P28
1.7
32
90
22
1
48
>100
4
20
>48
>100
>48
>100
methylguanosine)
6-azauridine
2
>100
>100
>100
26
20
>20
1.2
>20
For the meaning of a–i: see Table 2.
lActive also on Reo-1 virus (EC50 = 32
l
M). mActive also on Reo-1 virus (EC50 = 76
lM).

Table 4
Cytotoxicity against MT-4, MDBK, BHK, and Vero-76 cell lines and BVDV, YFV, CVB-2, RSV, and Sb-1 inhibitory activity of arylazoenamines of structures D and E
CH3
CH3
H
N
R
CH N=N Ar
N
CH3
N=N Ar
D: 44-63
E: 64, 65
a
b
c
d
e
f
g
h
i
Compound
Aryl=
MT-4 CC50
(l
M)
MDBK CC50
(l
M)
BVDV EC50
(lM)
BHK CC50
(l
M)
YFV EC50
(
lM)
Vero-76 CC50
(lM)
CVB-2 EC50
(lM)
RSV EC50
(lM)
Sb-1 EC50 (lM)
R=CH3
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
Phenyl
4-F-Ph
4-Cl-Ph
3-Br-Ph
4-Br-Ph
3-CF3-Ph
3-NO2-Ph
4-NO2-Phl
4-CH3-Ph
4-CH3O-Ph
3,4-diCl-Phm
3,5-diCF3-Ph
3-CF3-4-F-Phn
3-CF3-4-Br-Ph
PentaF-Ph
>100
100
>100
55
23
49
>100
59
85
52
>100
90
>100
>100
>100
67
90
36
10
18
21
9
11
21
5.5
19
20
18
16
>36
100
46
>100
80
55
41
>100
50
>100
>100
>100
77
23
48
>100
36
P55
7
82
75
92
80
100
>100
>100
>100
>100
100
100
93
19
>75
24
16
36
>100
>100
79
60
5
>100
>93
71
100
>100
20
12
20
>80
20
>100
>100
75
>100
>100
>100
>93
>100
>100
75
>82
>75
>92
>80
>100
>50
>100
65
>100
>77
>23
>48
>41
>100
>100
>100
>100
>100
>100
>100
>100
53
>93
>100
60
25
>100
>100
>100
39
2
11
>100
>100
>100
41
>100
>100
>100
P100
>100
24
8
>100
>100
R=CH2-C6H5
59
60
61
62
63
64
65
Phenyl
4-Cl-Ph
4-CH3-Ph
3,4-diCl-Pho
3-CF3-4-Cl-Ph
Phenylp
22
17
10
19
20
30
55
>100
74
18
22
79
14
20
72
>100
>74
>18
>22
>79
>14
>20
>72
1.7
31
20
25
50
22
47
21
90
>31
>20
>25
>50
>22
>47
>21
1
87
70
50
55
80
73
>100
>100
16
23
8
21
30
>87
>70
>50
>55
>80
>73
>100
>100
>70
>50
>55
>80
>73
>100
>100
30
>73
>100
20
4-Cl-Ph
NM-108 (20-b-
methylguanosine)
6-Azauridine
2
>100
>100
>100
26
20
>20
1.2
>20
For the meaning ofa–i: see Table 2; lmixture (2:1) with C isomer as seen from NMR spectrum; mactive also on HSV-1 virus (EC50 = 40
l
M); nactive also on HSV-1 virus (EC50 = 60
l
M); omixture (3:1) with C isomer as seen from NMR
spectrum; pas perchlorate.

Table 5
Cytotoxicity against MT-4, MDBK, BHK, and Vero-76 cell lines and BVDV, YFV, CVB-2, RSV, and Sb-1 inhibitory activity of arylazoenamines of structures F and G
X
CH=CH N=N Ar
N
Heteroaryl/Aryl
N
CH=CH N=N
N
F: 66-73
G: 74-85
a
b
c
d
e
f
g
h
i
Compound
X=
Ar=
MT-4 CC50
(lM)
MDBK CC50
(lM)
BVDV EC50
(l
M)
BHK CC50
(l
M)
YFV EC50
(lM)
Vero-76 CC50
M)
CVB-2 EC50
(l
M)
RSV EC50
(
lM)
Sb-1 EC50 (lM)
(
l
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
/
/
/
Phenyl
4-Cl-Ph
4-EtO-CO-Ph
Ph
4-EtO-CO-Ph
Ph
8
63
30
100
7
>100
28
14
13
P100
10
5
10
13
13
7
7.5
9
20
8
19
>63
>30
>100
>7
>100
>28
>14
>13
>100
>10
>5
>10
>13
>13
>7
>7.5
>9
82
>82
46
>100
23
>100
>100
>100
>100
>75
>100
42
>62
20
>69
10
7.5
9
70
22
>100
15
>100
30
30
30
22
55
50
34
35
14
11
45
11
45
4
4
50
7
35
7
6
12
6
0.9
5
6
30
>22
>100
>15
>100
15
>30
4
7
5
11
6
9
5
5
7
4
13
0.9
6
>100
1.2
>70
>22
>100
>15
27
61
26
2
33
7
>100
>100
67
>100
>100
>100
>100
75
>100
60
62
50
69
47
29
43
48
49
46
90
>100
CH2
CH2
O
>30
>30
>30
>22
>55
>50
>34
>35
>14
>11
>45
>11
>45
>11
>38
>100
>20
S
Ph
19
17
19
17
15
32
17
16
19
21
30
41
21
29
>100
2
EtO-CON Ph
2-Pyridinyl
2-Pyrimidinyl
Phenyl
2-F-Ph
4-F-Ph
2-Cl-Ph
3-Cl-Ph
4-Cl-Ph
5
5
>11
7
>11
14
>11
5
3-CF3-Ph
4-NO2-Ph
4-CH3CO-Ph
4-CH3O-Ph
>20
>8
>19
1.7
12
>49
>46
1
11
38
>100
20
NM-108 (20-b-methylguanosine)
>100
>100
20
>20
6-Azauridine
>100
26
For the meaning ofa–i: see Table 2.

8456
M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
cells is more significant than that on host cells, and therefore,
among the foregoing sequences, compounds 53, 28, 32, and 42 ap-
pear of high interest being devoid of toxicity on the human MT-4
of 84 disposes over the RAr feature of Hypo1_RSV, whereas one
nitrogen atom of the aza group again maps the remaining HBD
feature.
cells (CC50 > 100
l
M).
As expected, for the majority of the less active compounds we
found that they do not map adequately or eventually miss some
of the relevant hypothesis features; for some of these molecules,
their lack of affinity could be primarily ascribed to their inability
to achieve an energetically favorable conformation required by
the pharmacophore and shared by the active compounds.
Taken together, all these evidences concur to indicate that each
Hypo1 model provides reasonable pharmacophoric characteristics
of the three virus inhibitors for the components of their corre-
sponding activities.
3.4. Molecular modeling
3.4.1. Pharmacophore models generation
It is a widely accepted concept that three-dimensional (3D)
pharmacophore modeling is a well-behaved approach to quantita-
tively explore the common chemical characteristics among a con-
siderable number of different structures. Indeed, a computed
pharmacophore model can, in principle, be as good as the informa-
tion data input. Nevertheless, to achieve such a quality, at least
three must obey rules should be respected in a three-dimensional
quantitative structure–activity relationship (3D-QSAR) generation
using CATALYST: (i) the training set must be constituted by a wide
population (at least 16 items) of structurally diverse representative
covering at least four orders of magnitude of activity, (ii) the most
active compound should inevitably be included in the training set,
and (iii) all biological data must be obtained by homogeneous
procedures.
In our case, three training sets (TrS)—consisting of 18, 43, and
21 compounds for BVDV, CVB-2, and RSV, respectively—were pre-
pared by considering structural diversities and by the widest cov-
erage of in vitro activity range possible (see Table 6). The molecules
in each TrS were selected according to the following criteria: (i)
each TrS for each virus should contain structures from each class
of active compounds and (ii) each TrS should cover the molecular
bioactivities (EC50) as widely as possible. Should there be only
one compound with maximum or minimum order of bioactivity
in a class, this compound had to be assigned to the training set.
Compounds 51 and 62, although active against BVDV and/or
CVB-2, were not included in the corresponding training sets as they
were experimentally tested as a mixture of isomers.
A set of 10 hypotheses was automatically generated for each of
the three BVDV_TrS, CVB-2_TrS, and RSV_TrS training sets using
the HypoGen module in CATALYST. Hydrophobic aromatic (HAr)
sites, hydrogen bond acceptor (HBA), hydrogen bond donor
(HBD), positive ionizable (PI), and ring aromatic (RAr) moieties
were selected as the functional features to describe the antiviral
activity of the compounds. The best predictive hypothesis (Hypo1)
for each training set obtained from HypoGen has three features for
BVDV, and two features for CVB-2 and RSV. In detail, Hypo1_BVDV
is characterized by two hydrogen bond donors and one hydropho-
bic aromatic, while both Hypo1_CVB-2 and Hypo1_RSV have one
hydrogen bond donor and one aromatic ring (see Fig. 3). These
three hypotheses were also all characterized by the highest cost
differences, the lowest root-mean-square (RMSD) deviations, and
the highest correlation coefficients (see below).
3.4.2. Cost analysis
When generating an hypothesis, CATALYST performs a fixed cost
calculation, which represents the simplest model that fits all data
perfectly, and a null cost calculation, which presumes that there
is no relationship in the data set, and that the experimental activ-
ities are normally distributed about their mean. Accordingly, a gen-
erated pharmacophore hypothesis with a score that is substantially
lower than that of the corresponding null hypothesis is likely to be
statistically significant. Two further criteria should be satisfied in
assessing the validity of a given hypothesis: (a) the cost of the
hypothesis should be close to the fixed cost and (b) the configura-
tional cost, that is the magnitude of the hypothesis space for a given
TrS of compounds, should be less than 17. If this last cost exceeds
17, there are more degrees of freedom in the TrS that the CATALYST
algorithm can properly handle and, consequently, the correspond-
ing pharmacophore is likely to be poorly meaningful.
The quality of the three statistically most significant hypotheses
out of the ten generated by HypoGen for each TrSs (Hypo1_BVDV,
Hypo1_CVB-2, and Hypo1_RSV, respectively) was evaluated by
considering the relevant cost functions (see Table 7).
Detailedly, for BVDV the cost range between Hypo1_BVDV and
the fixed cost is 8.59, while that between the null hypothesis and
Hypo1_BVDV is 77.69 (see Table 7, first row). According to these
scores, criterion (a) is satisfied as the total cost of Hypo1_BVDV
is much closer to the fixed cost than to the null cost. Furthermore,
a high correlation coefficient (q) of 0.98, a RMSD value of 0.50, and
a configurational cost of 12.78 (criterion (b)) allow us to conclude
that Hypo1_BVDV is a reliable pharmacophore model with high
predictivity for our compounds. Treating the 43 test compounds
active against CVB-2 with HypoGen yielded the corresponding best
hypothesis Hypo1_CVB-2 (see Table 7, second row), characterized
by a total cost of 107.92. The total fixed cost for this run was
89.62 and the cost of the null hypothesis was 173.57. Thus, the
large difference between the null and total hypothesis costs
(65.65) and the closeness of the hypothesis cost to the fixed total
cost (18.30), coupled with a low configurational cost (13.56), a
In harmony with the predicted activities, the three most active
compounds in the three TrSs (i.e., compounds 32 in BVDV_TrS, 75
in CVB-2_TrS, and 84 in RSV_TrS in Table 6) can be nicely mapped
onto the corresponding Hypo1 models by the best fit values, as can
be appreciated from Figure 4. As we can see from this figure, com-
pound 32 fits very well all features of the pharmacophore model
Hypo1_BVDV. In fact, the first HBA is located on the nitro group,
the phenyl moiety overlaps with the HAr feature, and the second
HBA is mapped by the aza group. Noticeably, although the other
two hypotheses possess the same chemical features, the different
relative orientations and distances of these groups in the two cases
impose different mapping modes. Thus, in the case of compound
75, it is the aromatic ring bound to the aza group that overlaps
neatly over the RAr site, while one of the two nitrogen of the
pyrimidine substituent serves as an HBA in Hypo1_CVB-2. On the
other hand, the aromatic moiety of the acetylphenyl substituent
good RMSD (0.94), and a high q value (0.97), all speak in favor of
the statistical significance of the top-ranked hypothesis. Finally,
also for Hypo1_RSV the ample difference between the null and to-
tal hypothesis cost (79.85), the high correlation coefficient (0.95),
and the excellent RMSD value (0.38) (see Table 7, last row) indi-
cates a reliable predictivity of the corresponding 3D-QSAR pharma-
cophore and confirm that it is not a chance correlation.
3.4.3. Validation of the pharmacophore models
To cross-validate the observed correlations, we prepared three
test sets (TsS) of further 9, 17, and 12 compounds for BVDV,
CVB-2, and RSV, respectively. Interestingly, good correlation coeffi-
cients (0.97, 0.95, and 0.96) were observed when a regression anal-
ysis was performed by mapping each test set onto the features of
the corresponding best pharmacophore hypothesis. The predicted
and the experimental EC50 values for the three TsSs along with

M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
8457
Table 6
Estimated and experimentally determined activity values of the training set compounds for BVDV, CVB-2, and RSV
BVDV
EC50,exp.
CVB-2
EC50,exp.
M)
RSV
EC50,exp.
Training set
compounds
EC50,est.
M)
Errora
Training set
compounds
EC50,est.
Errora
Training set
compounds
EC50,est.
M)
Errora
(lM)
(
l
(l
(lM)
(l
M)
(l
32
22
42
17
50
47
33
27
45
53
52
49
39
30
14
9
0.8
3.5
5
5
5.5
9
0.9
7.9
3
7.8
4.6
1.1
2.3
75
43
66
28
53
76
79
85
72
77
69
71
61
21
18
6
73
35
83
59
47
4
60
46
29
1
0.9
4
4
5
5
5
5
5
6
6
7
7
8
1.6
5.8
3.1
7
2.2
5.8
6
5.3
5.9
6
1.8
1.5
84
73
82
75
79
77
23
76
45
83
71
30
44
48
4
0.9
4
4
5
5
6
7
11
12
13
15
18
20
20
25
30
30
40
75
75
1.1
1.2
1.8
1.0
7.2
4.1
4.5
5
5
10
9.6
15
10
18.9
19
34
32
34
29
34
33
68
65
93
ꢀ1.7
ꢀ1.3
1.6
1.4
ꢀ1.1
ꢀ0.8
1.7
1.2
ꢀ2.3
1.2
1.2
1.0
15
13
19
12
19
26
16
41
55
58
93
94
96
ꢀ1.2
1.4
11
12
18
18
20
21
40
60
61
86
90
1.6
1.1
ꢀ1.1
ꢀ1.5
ꢀ1.0
1.3
1.1
1.0
ꢀ1.3
1.3
1.1
1.7
1.6
1.3
5
10
ꢀ1.4
1.4
1.5
ꢀ1.3
1.0
12
15
9
16
10
13
ꢀ1.1
ꢀ1.1
1.1
1.0
ꢀ1.0
9
1.7
10
10
12
14
14
16
16
19
23
24
27
27
32
34
35
36
42
44
45
50
51
51
53
54
60
64
71
76
100
ꢀ1.1
1.4
1.6
59
66
27
15
58
41
ꢀ1.0
31
56
ꢀ1.2
ꢀ1.1
ꢀ2.0
ꢀ1.5
1.3
1.1
100
ꢀ1.2
ꢀ1.1
ꢀ1.1
ꢀ1.1
6.9
11
20
23
13
28
26
28
35
31
31
30
38
42
44
58
51
56
48
55
67
70
81
66
100
1.2
ꢀ1.8
1.2
ꢀ1.0
1.0
1.1
ꢀ1.1
ꢀ0.9
ꢀ1.2
ꢀ1.1
ꢀ1.0
ꢀ1.0
1.2
2
38
37
48
14
22
12
68
17
15
25
19
52
13
8
1.0
1.1
ꢀ1.1
1.0
1.1
1.1
1.1
42
57
ꢀ1.2
108
1.1
a
Value in the error column represent the ratio of the estimated activity to measured activity, or its negative inverse if the ratio is less than one.
the respective errors are shown in Table 8. The actual activity val-
ues are within the limits of uncertainty (i.e., 2.0), thus confirming
the validity of the original 3D pharmacophores.
4. Conclusions
Several sets of arylazoenamines (A–G, for a total of 85 com-
pounds), characterized by different types of aryl and basic moie-
ties, have been synthesized and assayed for antiviral activity
against a panel of ten RNA and DNA viruses. Eighty over 85 com-
pounds exhibited selective or large spectrum antiviral activity; in
particular, 12 were active on Sb-1, 16 on YFV, 28 on BVDV, 34 on
RSV, and 62 on CVB-2. Two viruses (Reo-1 and HSV-1) were sus-
ceptible only to very few compounds; the remaining (HIV-1, VSV,
and VV) were not affected by any compound.
Thirty-five compounds have shown high activity, with EC50 in
the range 0.8–10
pounds had EC50 between 11 and 30
arylazoenamine molecular pattern is an interesting novel pharma-
A final evaluation of the statistical relevance of the 3D pharma-
cophore models, the Fisher method was applied using the Cat-
Scramble program available in the CATALYST suite of programs.
According to the validation procedure, the experimental activity
of the compounds in each training test was scrambled randomly,
and the resulting new training set was used for a new HypoGen
run. The parameters adopted in running these calculations were
the same employed in the initial HypoGen calculations and, since
a 98% confidence level was selected, 49 random hypotheses were
generated. The resulting data clearly indicate that all values gener-
ated after randomization produced hypotheses with no predictive
values similar or close to the corresponding Hypo1s. Indeed, none
of the outcome hypotheses has lower cost score, better correlation
or smaller root-mean-square deviation than the initial one. Table 9
lists the first 10 lowest total score values of the resulting 49
hypotheses for BVDV, CVB-2, and RSV, respectively. In conclusion,
there is a 98% chance for the best hypothesis to represent a true
correlation in the training set activity data for each antiviral class
of compounds.
l
M, against at least one virus and other 28 com-
lM, thus indicating that the
cophore for antiviral agents against ssRNA viruses.
Some interesting correlations between the structure and the
type of affected virus were shown. The arylazohexahydroquinoli-
zines A and the aryazoethenamines F appear as the most suitable
structures for activity on CVB-2. The arylazotetrahydropyridines
C and the arylazomethylenepyrrolidines D share in almost equal
measure the activity on CVB-2 and BVDV, while the RSV was

8458
M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
Figure 3. Top scoring HypoGen pharmacophore for BVDV (top, left), CVB-2 (top, right), and RSV (bottom, center). Hypothesis features are color-coded as follows: cyan,
hydrophobic aromatic (HAr); green, hydrogen bond acceptor (HBA), dark gold, aromatic ring (RAr). The distances (Å) between all features in the corresponding hypotheses are
also shown.
Figure 4. Mapping of the three most active compounds of the training sets onto the corresponding top-ranked 3D pharmacophore models (top, left), compound 32 onto
Hypo1_BVDV; (top, right), compound 75 onto Hypo1_CVB-2; (bottom, center) compound 84 onto Hypo1_RSV. Hypothesis features are color-coded as follows: cyan,
hydrophobic aromatic (HAr); green, hydrogen bond acceptor (HBA), light blue/gray, aromatic ring (RAr).
Table 7
Results for the best pharmacophore hypotheses (Hypo1) generated by CATALYST HypoGen using the training sets for BVDV, CVB-2, and RSV
Hypotheses
Total fixed cost
Total Hypo1 cost
Null cost
Configurational cost
RMSDa
q
Features
Hypo1_BVDV
Hypo1_CVB-2
Hypo1_RSV
93.89
89.62
103.01
102.48
107.92
121.15
180.17
173.57
201.00
12.78
13.56
14.23
0.50
0.42
0.38
0.98
0.97
0.95
HBA, HBA, HAr
HBA, RA
HBA, RA
All cost units are in bits.
a
Defined in terms of the deviation of the log(estimated activity), from the log(experimental activity), normalized by the log(uncertainty).

M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
8459
Table 8
Estimated and experimentally determined activity values of the test set compounds for BVDV, CVB-2, and RSV
BVDV
EC50,exp.
M)
CVB-2
EC50,exp.
M)
RSV
EC50,exp.
M)
Test set
compounds
EC50,est
M)
Errora
Test set
compounds
EC50,est
M)
Errora
Test set
compounds
EC50,est.
(lM)
Errora
(
l
(
l
(
l
(
l
(
l
16
35
44
48
54
29
46
57
26
6
7
8
9
1.3
1.3
1.9
67
78
74
81
5
44
7
3
4
5
6
7
3.5
ꢀ1.1
1.3
ꢀ1.1
ꢀ1.1
1.0
80
85
81
74
78
3
29
28
46
25
2
5
6
7
7
9
18
18
20
20
30
60
75
4.9
7
8.5
6
12
16
20
19
24
24
51
64
ꢀ1.0
1.2
1.2
6.4
5.5
6.6
10
11
16
19
21
46
100
19
13
14
13
25
40
94
1.2
ꢀ1.2
ꢀ1.1
ꢀ1.5
1.2
18
19
21
22
26
30
30
34
35
38
40
41
71
18
21
19
28
25
28
29
38
28
33
44
38
62
1.3
1.1
ꢀ1.1
1.1
ꢀ1.1
ꢀ1.2
1.3
ꢀ1.1
ꢀ1.1
9
ꢀ1.0
ꢀ1.1
ꢀ1.0
1.1
ꢀ1.3
ꢀ1.2
1.1
1.2
20
63
27
70
10
39
11
56
ꢀ1.3
ꢀ1.2
ꢀ1.2
24
ꢀ1.1
ꢀ1.1
a
Values in the error column represent the ratio of the estimated activity to measured activity, or its negative inverse if the ratio is less than one.
Table 9
Output parameters of the 10 lowest cost hypotheses resulting from the statistical evaluation according to the Cat-Scramble validation procedure for BVDV, CVB-2, and RSV
BVDV CVB-2 RSV
Hypothesis
q
RMSD
Total cost
Hypothesis
1
q
RMSD
Total cost
Hypothesis
1
q
RMSD
Total cost
1
2
3
4
5
6
7
8
0.77
0.76
0.76
0.74
0.73
0.71
0.71
0.69
0.66
0.64
0.98
0.75
0.78
0.79
0.80
0.81
0.84
0.88
0.92
1.01
1.11
0.50
131.66
133.14
133.75
135.22
137.29
140.01
141.56
142.81
144.35
145.22
102.48
0.84
0.84
0.81
0.79
0.78
0.75
0.75
0.73
0.73
0.72
0.97
0.91
0.93
1.02
1.01
1.05
1.13
1.21
1.24
1.25
1.26
0.42
158.04
159.51
181.09
183.63
184.64
190.07
192.55
193.78
194.16
198.23
107.92
0.74
0.72
0.68
0.66
0.63
0.61
0.57
0.55
0.52
0.47
0.95
0.68
0.75
0.84
0.90
0.96
0.99
1.05
1.08
1.14
1.21
0.38
155.35
157.87
161.69
168.21
169.97
170.30
174.08
174.53
178.04
181.69
121.15
2
3
4
5
6
7
8
10
2
3
4
5
6
7
8
10
9
10
Hypo1
Hypo1
Hypo1
mainly affected by arylazoethenepiperazines G. Many compounds
exhibited a selectivity index (CC50 for host cells/EC50 for a given
virus) in the range from >10 to >125.
5. Experimental
5.1. General
Taking into account the EC50 and SI values, the best compounds
can be arranged in the following order of decreasing potency and
Chemicals, solvents, and reagents used for syntheses were pur-
chased from Sigma–Aldrich, Fluka, or Lancaster, and were used
without any further purification, unless otherwise specified. Not
commercially available arylhydrazines, as 4-F-3-CF3,37 4-Cl-3-
CF3,38 4-Br-3-CF3,39 4-ethoxycarbonyl-phenylhydrazine,40 and 1-
naphthylhydrazine,40 were prepared according to Enders.40
Column chromatography (CC): neutral alumina (Al2O3), activity
1 (Merck). Mps: Büchi apparatus, uncorrected. 1H NMR spectra:
Varian Gemini-200 spectrometer; CDCl3; d in ppm rel to Me4Si as
internal standard. J in Hz; Ind = hexahydroindolizine, Pip = piperi-
dine, Py = tetrahydropyridine, Pyrr = pyrrolidine, Pz = piperazine,
Q = hexahydroquinolizine. Elemental analyses were performed on
a Carlo Erba EA-1110 CHNS-O instrument in the Microanalysis Lab-
oratory of the Department of Pharmaceutical Sciences of Genoa
University.
selectivity;
for
CVB-2:
75 > 66 P 53 P 28;
for
BVDV:
32 > 22 > 42 = 50 > 16; for RSV: 84 > 23, and for Sb-1: 26 P 28. In
these sequences, even the least potent compounds still compare
well with the reference drugs NM-108 and 6-azauridine.
Compounds 53, 28, 32, and 42 appear of particular interest
being also devoid of toxicity on the human MT-4 cells
(CC50 > 100 lM).
A ligand-based computational approach was employed to iden-
tify molecular structure requirements in the available series of
compounds to act as effective viral inhibitors. Three highly predic-
tive pharmacophore models were generated based on 18, 43, and
21 training set compounds for BVDV, CVB-2, and RSV, respectively,
which consist of two hydrogen bond donors and one hydrophobic
aromatic in the case of BVDV (Hypo1_BVDV), and one hydrogen
bond donor and one ring aromatic for the two other viral species
(Hypo1_CVB-2 and Hypo1_RSV, respectively). The application of
our 3D pharmacophore models to three test sets of different com-
pounds showed that each model is able to accurately predict the
activity of the compounds toward the respective virus. Since the
target proteins for the antiviral activities of our series of com-
pounds is as yet unknown, our three-dimensional QSAR pharmaco-
phore models should constitute useful tools in the design of second
generations of different, more potent inhibitors of these human
and veterinary pathogens.
5.2. Aryl/heteroarylazoenamines of structures A, B, C, and D (1–
63): general method
To a solution of a-aminoketone (3 mmol) {hexahydro-2H-quin-
olizin-1-one,34 hexahydro-8(5H)-indolizinone,35 1-Methyl-3-pip-
eridone36} in 15 mL of abs. ethanol were added, in the following
order, aryl/heteroarylhydrazine hydrochloride (6 mmol) or the cor-
responding free base, 0.75 mL of conc. hydrochloric acid, and
0.9 mL of 85% phosphoric acid. The mixture was refluxed with stir-

8460
M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
ring for 3 h. After cooling, ethanol was evaporated at reduced pres-
sure, water was added, and the acid solution was basified with 2 M
NaOH and extracted with Et2O. After drying, the solvent was re-
moved obtaining a red-orange oil from which the volatile sub-
stances were removed by heating until 80–100 °C under vacuum
(0.2 torr). The residue was purified by CC (CH2Cl2).
The low (C) and high (D) melting point isomers, obtained in the
reaction with 1-methyl-3-piperidone, were separated by eluting
firstly with CH2Cl2 and then CH2Cl2 + 2%Et2NH.
3.15–3.30 (m, 6H, C(2, 4, 6) of Q); 7.32–7.78 (m, 5 arom. H);
8.74–8.84 (m, 1 arom. H). Anal. calcd for C19H21N3: C 78.32, H
7.26, N, 14.42; found: C 78.67, H 7.03, N 14.02.
5.2.8. 1-[4-(7-Chloro)quinolylazo]-3,4,6,7,8,9-hexahydro-2H-
quinolizine (17)
Yield: 58%. CC (CH2Cl2). Mp 203–206 °C (Et2O). 1H NMR (CDCl3):
1.74–2.14 (m, 6H, C(3, 7, 8) of Q); 2.89 (t, J = 6.4, 2H, C(9) of Q);
3.29–3.50 (m, 6H, C(2, 4, 6) of Q); 7.35–8.84 (m, 5 arom. H). Anal.
calcd for C18H19ClN4: C 66.15, H 5.86, N 17.14; found: C 66.04, H
6.04, N 17.09.
In the case of phthalazylazoenamine 23, the reaction mixture
was neutralized and then shaken with CH2Cl2 which extracted
the compound as hydrochloride.
Compounds 1–4, 6, 9, 10, 13, 15, 24, 26–28, 32-35, 38, 40, 44,
5.2.9. 8-(4-Chlorophenylazo)-1,2,3,5,6,7-hexahydroindolizine
(18)
46, and 51–54 have been already described.2,8,9
Yield: 25%. Mp 165–167 °C (Et2O). 1H NMR (CDCl3): 1.95–2.20
(m, 4H, C(2, 6) of Ind); 2.65 (t, J = 6.1, 2H, C(7) of Ind); 3.22 (t,
J = 7.8, 2H, C(1) of Ind); 3.33 (t, J = 6.0, 2H, C(5) of Ind); 3.55 (t,
J = 6.6, 2H, C(3) of Ind); 7.25 (d, J = 2.2, 2 arom. H); 7.50 (d,
J = 2.2, 2 arom. H). Anal. calcd for C14H16ClN3: C 64.24, H 6.16, N
16,05: found: C 64.45, H 6.20, N 16.19.
5.2.1. 1-(3-Bromophenylazo)-3,4,6,7,8,9-hexahydro-2H-
quinolizine (5)
Yield:42%. Mp 115–117 °C (Et2O). 1H NMR(CDCl3): 1.61–1.88(m,
6H, C(3, 7, 8) of Q); 2.59 (t, J = 6.5, 2H, C(9) of Q); 3.08–3.36 (m, 6H,
C(2, 4, 6) of Q); 7.03–7.63 (m, 4 arom. H). Anal. calcd for C15H18BrN3:
C 56.26, H 5.67, N 13.12; found: C 56.45, H 5.69, N 13.09.
5.2.10. 8-(3-Nitrophenylazo)-1,2,3,5,6,7-hexahydroindolizine
(19)
5.2.2. 1-(3-Trifluoromethylphenylazo)-3,4,6,7,8,9-hexahydro-
2H-quinolizine (7)
Yield: 43%. Mp 157–159 °C (Et2O). 1H NMR (CDCl3): 1.98–2.22
(m, 4H, C(2, 6) of Ind); 2.68 (t, J = 6.2, 2H, C(7) of Ind); 3.29 (t,
J = 7.9, 2H, C(1) of Ind); 3.38 (t, J = 6.0, 2H, C(5) of Ind); 3.64 (t,
J = 6.6, 2H, C(3) of Ind); 7.37–8.40 (m, 4 arom. H). Anal. calcd for
Yield: 25%. Mp 102–103 °C (Et2O). 1H NMR (CDCl3): 1.60–1.97
(m, 6H, C(3, 7, 8) of Q); 2.63 (t, J = 6.4, 2H, C(9) of Q); 3.07–3.33
(m, 6H, C(2, 4, 6) of Q); 7.22–7.77 (m, 4 arom. H). Anal. calcd for
C16H18F3N3: C 62.13, H 5.87, N 13.58; found: C 62.00, H 6.04, N
C14H16N4O2: C 61.75, H 5.92. N 20.58; found: C 61.78, H 6.04, N
13.23.
20.45.
5.2.3. 1-(3-Nitrophenylazo)-3,4,6,7,8,9-hexahydro-2H-
quinolizine (8)
5.2.11. 8-(3,4-Dichlorophenylazo)-1,2,3,5,6,7-
hexahydroindolizine (20)
Yield: 46%. CC (Et2O). Mp 132–135°C (Et2O). 1H NMR (CDCl3):
1.63–2.00 (m, 6H, C(3, 7, 8) of Q); 2.63 (t, J = 6.4, 2H, C(9) of Q);
3.12–3.37 (m, 6H, C(2, 4, 6) of Q); 7.30–8.32 (m, 4 arom. H). Anal.
calcd for C15H18N4O2: C 62.92, H 6.34; N 19.57; found: C 62.61, H
6.39, N 19.58.
Yield: 30%. CC (Et2O). Mp 176–177 °C (Et2O). 1H NMR (CDCl3):
1.87–2.11 (m, 4H, C(2, 6) of Ind); 2.59 (t, J = 6.4, 2H, C(7) of Ind);
3.17 (t, J = 7.8, 2H, C(1) of Ind); 3.29 (t, J = 5.9, 2H, C(5) of Ind);
3.54 (t, J = 6.6, 2H, C(3) of Ind); 7.18–7.34 (m, 2 arom. H); 7.58 (d,
J = 2.0, 1 arom. H). Anal. calcd for C14H15Cl2N3: C 56.77, H 5.10, N
14.19; found: C 56.43, H 5.08, N 14.00.
5.2.4. 1-(3,5-Bis-trifluoromethylphenylazo)-3,4,6,7,8,9-
hexahydro-2H-quinolizine (11)
Yield: 72%. Mp 147–149 °C (Et2O). 1H NMR (CDCl3): 1.64–1.99
(m, 6H, C(3, 7, 8) of Q); 2.63 (t, J = 6.5, 2H, C(9) of Q); 3.12–3.36
(m, 6H, C(2, 4, 6) of Q); 7.40–7.90 (m, 3 arom. H). Anal. calcd for
5.2.12. 8-(4-Chloro-3-trifluoromethylphenylazo)-1,2,3,5,6,7-
hexahydroindolizine (21)
Yield: 81%. Mp 161–162.5 °C (Et2O). 1H NMR (CDCl3): 1.88–2.12
(m, 4H, C(2, 6) of Ind); 2.59 (t, J = 6.4, 2H, C(7) of Ind); 3.17 (t,
J = 7.9, 2H, C(1) of Ind); 3.29 (t, J = 5.8, 2H, C(5) of Ind); 3.54 (t,
J = 7.1, 2H, C(3) of Ind); 7.32 (d, J = 8.6, 1 arom. H); 7.56 (dd,
J = 8.6, 2.0, 1 arom. H); 7.82 (d, J = 2.4, 1 arom. H). Anal. calcd for
C17H17F6N3: C 54,11, H 4.54, N 11.11; found: C 54.00, H 4.72, N
11.02.
5.2.5. 1-(4-Fluoro-3-trifluoromethylphenylazo)-3,4,6,7,8,9-
hexahydro-2H-quinolizine (12)
C15H15ClF3N3: C 54.64, H 4.59, N 12.74; found: C 54.47, H 4.57, N
Yield: 88%. Mp 133–134 °C (Et2O). 1H NMR (CDCl3): 1.60–1.98
(m, 6H, C(3, 7, 8) of Q); 2.58 (t, J = 6.5, 2H, C(9) of Q); 3.07–3.34
(m, 6H, C(2, 4, 6) of Q); 6.98–7.80 (m, 3 arom. H). Anal. calcd for
12.69.
5.2.13. 8-(1-Naphthylazo)-1,2,3,5,6,7-hexahydroindolizine (22)
Yield: 32%. CC (CH2Cl2). Mp 148–151 °C (CH2Cl2). 1H NMR
(CDCl3): 1.92–2.10 (m, 4H, C(2, 6) of Ind); 2.77 (t, J = 6.4, 2H, C(7)
of Ind); 3.17–3.30 (m, 4H, C(1, 5) of Ind); 3.46 (t, J = 6.9, 2H, C(3)
of Ind); 7.33–7.78 (m, 6 arom. H); 8.78–8.85 (m, 1 arom. H). Anal.
calcd for C18H19N3: C 77.95, H 6.90, N 15.15; found: C 78.01, H 6.94,
15.13.
C
16H17F4N3: C 58.71, H 5.23, N 12.84; found: C 58.51, H 5.43, N
12.59.
5.2.6. 1-(4-Bromo-3-trifluoromethylphenylazo)-3,4,6,7,8,9-
hexahydro-2H-quinolizine (14)
Yield: 37%. CC (CH2Cl2). Mp 155–157 °C (Et2O). 1H NMR (CDCl3):
1.62–1.98 (m, 6H, C(3, 7, 8) of Q); 2.60 (t, J = 6.5, 2H, C(9) of Q);
3.07–3.35 (m, 6H, C(2, 4, 6) of Q); 7.40–7.84 (m, 3 arom. H). Anal.
calcd for C16H17BrF3N3: C 49.50, H 4.41, N 10.82; found: C 49.53, H
4.32, N 10.74.
5.2.14. 8-(1-Phthalazinylazo)-1,2,3,5,6,7-hexahydroindolizine
hydrochloride (23)
Yield: 49%. Mp 174–176 °C (Et2O). 1H NMR (CDCl3): 1.90–2.10
(m, 4H, C(2, 6) of Ind); 2.72 (t, J = 6.2, 1H, C(7) of Ind); 2.93 (t,
J = 6.2, 1H, C(7) of Ind); 3.08 (t, J = 7.0, 2H, C(1) of Ind); 3.73–3.86
(m, 2H, C(5) of Ind); 4.10–4.28 (m, 2H, C(3) of Ind); 7.88–8.72
(m, 4 arom. H); 9.10 (d, J = 8.0, 1 arom. H). Anal. calcd for
5.2.7. 1-(1-Naphthylazo)-3,4,6,7,8,9-hexahydro-2H-quinolizine
(16)
Yield: 58%. CC (CH2Cl2). Mp 153–155 °C (Et2O). 1H NMR (CDCl3):
1.63–2.00 (m, 6H, C(3, 7, 8) of Q); 2.79 (t, J = 6.4, 2H, C(9) of Q);

M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
8461
C16H17N5 + HCl+H2O: C 57.57, H 6.04, N 20.97; found: C 57.39, H
6.25, 20.25.
(t, J = 6.0, 2H, C(6) of Py); 7.27 (s, 1H, C(2) of Py). Anal. calcd for
12H10F5N3: C 49.49, H 3.46, N 14.43; found: C 49.51, H 3.32, N 14.00.
C
5.2.23. 1-Methyl-3-(1-naphthylazo)-1,4,5,6-tetrahydropyridine
(42)
5.2.15. 3-(4-Fluorophenylazo)-1-methyl-1,4,5,6-
tetrahydropyridine (25)
Yield: 42%. Mp 116–119 °C (Et2O). 1H NMR (CDCl3): 2.08 (q, 2H,
C(5) of Py); 2.82 (t, J = 6.7, 2H, C(4) of Py); 3.14 (s, CH3N); 3.29 (t,
J = 5.7, 2H, C(6) of Py); 7.35 (s, 1H, C(2) of Py); 7.44–7.90 (m, 6
arom. H); 8.79–8.87 (m, 1 arom. H). Anal. calcd for C16H17N3: C
76.46, H 6.82, N 16.72; found: C 76.46, H 6.74, N 16.39.
Yield: 25%. Mp 82–84 °C (pentane). 1H NMR (CDCl3): 1.98 (q,
2H, C(5) of Py); 2.60 (t, J = 6.4, 2H, C(4) of Py); 3.06 (s, CH3N);
3.22 (t, J = 5.8, 2H, C(6) of Py); 6.97–7.08 (m, 2 arom. H); 7.20 (s,
1H, C(2) of Py); 7.49–7.59 (m, 2 arom. H). Anal. calcd for
C12H14FN3: C 65.73, H 6.44, N 19.16; found. C 65.52, H 6.15, N
19.28.
5.2.24. 1-Benzyl-3-(4-chloro-3-trifluoromethylphenylazo)-
1,4,5,6-tetrahydropyridine (43)
5.2.16. 3-(3-Bromophenylazo)-1-methyl-1,4,5,6-
tetrahydropyridine (29)
Yield: 27%. Mp 118–119 °C (Et2O). 1H NMR (CDCl3): 1.96 (q, 2H,
C(5) of Py); 2.64 (t, J = 5.8, 2H, C(4) of Py); 3.23 (t, J = 5.7, 2H, C(6) of
Py); 4.48 (s, 2H, CH2C6H5); 7.22–8.02 (m, 8 arom. H and 1H, C(2) of
Py). Anal. calcd for C19H17ClF3N3: C, 60.08, H 4.51, N 11.06; found:
C 59.74, H 4.47, N 10.95.
Yield: 20%. Mp 96–99 °C (Et2O). 1H NMR (CDCl3): 1.99 (q, 2H,
C(5) of Py); 2.61 (t, J = 6.4, 2H, C(4) of Py); 3.12 (s, CH3N); 3.26 (t,
J = 5.8, 2H, C(6) of Py); 7.20 (s, 1H, C(2) of Py); 7.24–7.77 (m, 4
arom. H). Anal. calcd for C12H14BrN3: C 51.44, H 5.04, N 15.00;
found: C 51.59, H 5.07, N 14.82.
5.2.25. 2-(4-Fluorophenylazo)methylene-1-methylpyrrolidine
(45)
Yield: 23%. Mp 107–108 °C (pentane). 1H NMR (CDCl3): 2.07 (q,
2H, C(4) of Pyrr); 2.98 (s, CH3N); 3.27 (t, J = 7.7, 2H, C(3) of Pyrr);
3.57 (t, J = 7.0, 2H, C(5) of Pyrr); 7.98–8.12 (m, 1H, @CH– and 2
arom. H); 7.50–7.61 (m, 2 arom. H). Anal. calcd for C12H14FN3: C
65.73, H 6.44, N 19.16; found: C 65.60, H 6.12, N 19.31.
5.2.17. 3-(4-Bromophenylazo)-1-methyl-1,4,5,6-
tetrahydropyridine (30)
Yield: 22%. Mp 112–114 °C (Et2O). 1H NMR (CDCl3): 1.92 (q, 2H,
C(5) of Py); 2.54 (t, J = 6.4, 2H, C(4) of Py); 3.04 (s, CH3N); 3.18 (t,
J = 5.8, 2H, C(6) of Py); 7.17 (s, 1H, C(2) of Py); 7.34–7.42 (m, 4
arom. H). Anal. calcd for C12H14BrN3: C 51.44, H 5.04, N 15.00;
found: C 51. 40, H 4.52, N 14.78.
5.2.26. 2-(3-Bromophenylazo)methylene-1-methylpyrrolidine
(47)
Yield: 22%. Mp 103–107 °C (Et2O). 1H NMR (CDCl3): 2.08 (q, 2H,
C(4) of Pyrr); 3.01 (s, CH3N); 3.30 (t, J = 7.8, 2H, C(3) of Pyrr); 3.60
(t, J = 7.0, 2H, C(5) of Pyrr); 7.14 (s, 1H, @CH–); 7.20–7.76 (m, 4
arom. H). Anal. calcd for C12H14BrN3: C 51.44, H 5.04, N 15.00;
found: C 51.41, H 4.98, N 14.87.
5.2.18. 1-Methyl-3-(3-trifluoromethylphenylazo)-1,4,5,6-
tetrahydropyridine (31)
Yield: 20%. Mp 76–77 °C (pentane). 1H NMR (CDCl3): 1.93 (q,
2H, C(5) of Py); 2.57 (t, J = 6.5, 2H, C(4) of Py); 3.06 (s, CH3N);
3.20 (t, J = 5.8, 2H, C(6) of Py); 7.23 (s, 1H, C(2) of Py); 7.27–7.81
(m, 4 arom. H). Anal. calcd for C13H14F3N3: C 57.99, H 5.24, N
15.61; found: 57.65, H 5.38, N 15.75.
5.2.27. 2-(4-Bromophenylazo)methylene-1-methylpyrrolidine
(48)
Yield: 33%. Mp 150–152.5 °C (Et2O). 1H NMR (CDCl3): 2.02 (q,
2H, C(4) of Pyrr); 2.95 (s, CH3N); 3.20 (t, J = 7.8, 2H, C(3) of Pyrr);
3.53 (t, J = 7.0, 2H, C(5) of Pyrr); 7.09 (s, 1H, @CH–); 7.34–7.42
(m, 4 arom. H). Anal. calcd for C12H14BrN3: C 51.44, H 5.04, N
15.00; found: C 51.07, H 5.17, N 14.80.
5.2.19. 3-(3,5-Bis-trifluoromethylphenylazo)-1-methyl-1,4,5,6-
tetrahydropyridine (36)
Yield: 20%. Mp 86–87 °C (pentane). 1H NMR (CDCl3): 2.02 (q,
2H, C(5) of Py); 2.64 (t, J = 6.7, 2H, C(4) of Py); 3.18 (s, CH3N);
3.31 (t, J = 5.8, 2H, C(6) of Py); 7.38 (s, 1H, C(2) of Py); 7.55–8.03
(m, 3 arom. H). Anal. calcd for C14H13F6N3: C 49.86, H 3.89, N
12.46; found: C 49.91, H 3.60, N 12.30.
5.2.28. 1-Methyl-2-(3-trifluoromethylphenylazo)methylenepy-
rrolidine (49)
Yield: 15%. Mp 107–108 °C (Et2O). 1H NMR (CDCl3): 2.10 (q, 2H,
C(4) of Pyrr); 3.03 (s, CH3N); 3.32 (t, J = 7.9, 2H, C(3) of Pyrr); 3.62
(t, J = 7.0, 2H, C(5) of Pyrr); 7.18 (s, 1H, @CH–); 7.33–7.86 (m, 4
arom. H). Anal. calcd for C13H14F3N3: C 57.99, H 5.24, N 15.61;
found: C 58.11, H 5.55, N 15.33.
5.2.20. 3-(4-Fluoro-3-trifluoromethylphenylazo)-1-methyl-
1,4,5,6-tetrahydropyridine (37)
Yield: 30%. Mp 81–84 °C (pentane). 1H NMR (CDCl3): 2.00 (q,
2H, C(5) of Py); 2.61 (t, J = 6.7, 2H, C(4) of Py); 3.13 (s, CH3N);
3.26 (t, J = 5.9, 2H, C(6) of Py); 7.10–7.87 (m, 1H, C(2) of Py and 3
arom. H). Anal. calcd for C13H13F4N3: C 54.36, H 4.56, N 14.63;
found: C 54.03, H 4.22, N 13.67.
5.2.29. 1-Methyl-2-(3-nitrophenylazo)methylenepyrrolidine
(50)
Yield: 30%. Mp 117–119 °C (Et2O). 1H NMR (CDCl3): 2.06 (q, 2H,
C(4) of Pyrr); 3.00 (s, CH3N); 3.27 (t, J = 7.8, 2H, C(3) of Pyrr); 3.60
(t, J = 7.0, 2H, C(5) of Pyrr); 7.18–8.32 (m, 1H, @CH– and 4 arom. H).
Anal. calcd For C12H14N4O2: C 58.53, H 5.73, N 22.75; found: C
58.44, H 5.89, N 22.73.
5.2.21. 3-(4-Bromo-3-trifluoromethylphenylazo)-1-methyl-
1,4,5,6-tetrahydropyridine (39)
Yield: 34%. Mp 94–95 °C (Et2O). 1H NMR (CDCl3): 2.01 (q, 2H,
C(5) of Py); 2.63 (t, J = 6.4, 2H, C(4) of Py); 3.16 (s, CH3N); 3.30 (t,
J = 5.7, 2H, C(6) of Py); 7.33 (s, 1H, C(2) of Py); 7.59–7.93 (m, 3
arom. H). Anal. calcd for C13H13BrF3N3: C 44.85, H 3.76, N, 12.07,
found: C 44.95, H 4.06 N 11.76.
5.2.30. 1-Methyl-2-(4-nitrophenylazo)methylenepyrrolidine
and 1-Methyl-3-(4-nitrophenylazo)-1,4,5,6-tetrahydropyridine
(mixture 2:1) (51)
Yield: 60%. Mp 180 °C (EtOH). 1H NMR (CDCl3): 1.98 (q,
0.33 ꢂ 2H, C(5) of Py); 2.14 (q, 0.67 ꢂ 2H, C(4) of Pyrr); 2.63 (t,
J = 6.4, 0.33 ꢂ 2H, C(4) of Py); 3.10 (s, 0.67 ꢂ 3H, NCH3 of Pyrr);
3.18 (s, 0.33 ꢂ 3H, NCH3 of Py); 3.22–3.37 (m, 0.67 ꢂ 2H, C(3) of
5.2.22. 1-Methyl-3-pentafluorophenylazo-1,4,5,6-
tetrahydropyridine (41)
Yield: 24%. Mp 100–101 °C (pentane). 1H NMR (CDCl3): 1.95 (q,
2H, C(5) of Py); 2.55 (t, J = 6.5, 2H, C(4) of Py); 3.11 (s, CH3N); 3.25

8462
M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
Pyrr) and 0.33 ꢂ 2H, C(6) of Py); 3.70 (t, J = 7.0, 0.67 ꢂ 2H, C(5) of
Pyrr); 7.36–7.40 (m, 1H, @CH– of Pyrr and Py); 7.58 (d, J = 9.0, 4
arom. H); 8.16 (d, J = 9.2, 4 arom. H). Anal. calcd for C12H14N4O2:
C 58.52, H 5.73, N 22.75; found: C 58.75, H 5.72, N 22.87.
(t, J = 7.7, 0.75 ꢂ 2H, C(3) of Pyrr); 3.59 (t, J = 7.1, 0.75 ꢂ 2H, C(5)
of Pyrr); 4.45 (s, 0.25 ꢂ 2H, CH2C6H5 of Py); 4.54 (s, 0.75 ꢂ 2H,
CH2C6H5 of Pyrr); 7.20–7.70 (m, 1H, @CH– of Pyrr and Py and 8
arom. H). Anal. calcd for C18H17Cl2N3: C 62.43, H 4.95, N 12.13;
found: C 62.68, H 4.96, N 12.16.
5.2.31. 2-(3,5-Bis-trifluoromethylphenylazo)methylene-1-
methylpyrrolidine (55)
5.2.39. 1-Benzyl-2-(4-chloro-3-trifluoromethylphenylazo)-
methylenepyrrolidine (63)
Yield: 15%. Mp 128–130 °C (Et2O). 1H NMR (CDCl3): 2.07(q, 2H,
C(4) of Pyrr); 3.01 (s, CH3N); 3.28 (t, J = 7.8, 2H, C(3) of Pyrr); 3.61
(t, J = 7.0, 2H, C(5) of Pyrr); 7.19 (s, 1H, @CH–); 7.45–7.70 (m, 3
arom. H). Anal. calcd for C14H13F6N3: C 49.86, H 3.89, N 12.46;
found: C 50.42, H 3.56, N 12.35.
Yield: 37%. Mp 114–116 °C (Et2O). 1H NMR (CDCl3): 2.12 (q, 2H,
C(4) of Pyrr); 3.38 (t, J = 7.8, 2H, C(3) of Pyrr); 3.62 (t, J = 7.1, 2H,
C(5) of Pyrr); 4.47 (s, 2H, CH2C6H5); 7.15–7.98 (m, 1H, @CH– and
8 arom. H). Anal. calcd for C19H17ClF3N3: C 60.08, H 4.51, N
11.06; found: C 59.95, H 4.49, N 10.95.
5.2.32. 2-(4-Fluoro-3-trifluoromethylphenylazo)methylene-1-
methylpyrrolidine (56)
5.3. 2-(Arylazomethylene)-1,3,3-trimethylindoline (64–65)
Yield: 18%. Mp 120–123 °C (Et2O). 1H NMR (CDCl3): 2.03 (q, 2H,
C(4) of Pyrr); 2.96 (s, CH3N); 3.23 (t, J = 7.8, 2H, C(3) of Pyrr); 3.56
(t, J = 7.0, 2H, C(5) of Pyrr); 7.00–7.75 (m, 1H, @CH– and 3 arom. H).
Anal. calcd for C13H13F4N3: C 54.36, H 4.56, N 14.63; found: C
54.34, H 4.56, N 14.60.
The title compounds were prepared through the method used by
König and Muller4 with minor modification. A solution of aniline or
4-chloroaniline (15 mmol) in 6 mL of conc. hydrochloridric acid and
15 mL of water was cooled in an ice bath and a solution of sodium ni-
trite (15 mmol) in 3.5 mL of cold water was added dropwise. The
resulting solution of diazonium salt was then poured slowly into a
stirred emulsion of 1,3,3-trimethyl-2-methyleneindoline41 in
15 mL of water and subsequently added of sodium acetate (10 g).
After 1 h at 0 °C with stirring, the aqueous phase was extracted with
CH2Cl2, theorganicphasewasdriedwithsodiumsulfateandconcen-
trated to dryness, obtaining, in the case of 64, an oil which crystal-
lized when treated with the equivalent amount of 1N perchloric
acid in abs. ethanol. Compound 65 was a crystalline solid.32
5.2.33. 2-(4-Bromo-3-trifluoromethylphenylazo)methylene-1-
methylpyrrolidine (57)
Yield: 43%. Mp 126–128 °C (Et2O). 1H NMR (CDCl3): 2.12 (q, 2H,
C(4) of Pyrr); 3.05 (s, CH3N); 3.31 (t, J = 7.9, 2H, C(3) of Pyrr); 3.65
(t, J = 7.1, 2H, C(5) of Pyrr); 7.19 (s, 1H, @CH–); 7.50–7.90 (m, 3
arom. H). Anal. calcd for C13H13BrF3N3: C 44.85, H 3.76, N 12.07;
found: C 44.76, H 3.89, N 11.89.
5.2.34. 1-Methyl-2-(pentafluorophenylazo)methylenepyrroli-
dine (58)
5.3.1. 2-(Phenylazomethylene)-1,3,3-trimethylindoline
perchlorate (64)
Yield: 64%. Mp 239–240 °C. Anal. calcd for C18H19N3+HClO4: C
57.22, H 5.34, N 11.12; found: C 57.30, H 5.29, N 11.12. Already de-
scribed as hydroiodide,3 hydrochloride,4 and free base,3,42 which,
however, in our hands did not crystallize well.
Yield: 15%. Mp 154–155 °C (pentane). 1H NMR (CDCl3): 2.10 (q,
2H, C(4) of Pyrr); 3.07 (s, CH3N); 3.24 (t, J = 7.8, 2H, C(3) of Pyrr);
3.67 (t, J = 7.1, 2H, C(5) of Pyrr); 7.23 (s, 1H, @CH–). Anal. calcd
for C12H10F5N3: C 49.49, H 3.46, N 14.43; found: C 49.58, H 3.22,
N 14.18.
5.2.35. 1-Benzyl-2-phenylazomethylenepyrrolidine (59)
Yield: 43%. Mp 121–123 °C (Et2O). 1H NMR (CDCl3): 2.08 (q, 2H,
C(4) of Pyrr); 3.38 (t, J = 7.7, 2H, C(3) of Pyrr); 3.56 (t, J = 7.0, 2H,
C(5) of Pyrr); 4.53 (s, 2H, CH2C6H5); 7.03–7.73 (m, 1H, @CH– and
10 arom. H). Anal. calcd for C18H19N3: C 77.95, H 6.90, N 15.15;
found: C 77,74, H 6.89, N 15.15.
5.3.2. 2-(4-Chlorophenylazomethylene)-1,3,3-trimethyl-
indoline (65)
Yield: 50%. Mp 132–134 °C (MeOH) lit.32 141.5 °C. 1H NMR
(CDCl3): 1.76 (s, 6H, 2 CH3 of Indoline); 3.35 (s, CH3N); 6.84–7.33
(m, 5 arom. H); 7.38 (d, J = 9.2, 2 arom. H); 7.61 (d, J = 9.2, 2 arom.
H). Anal. calcd for C18H18ClN3: C 69.33, H 5.82, N 13.48; found: C
69.13, H 5.85, N 13.47.
5.2.36. 1-Benzyl-2-(4-chlorophenylazo)methylenepyrrolidine
(60)
Yield: 32%. Mp 127–129 °C (EtOH). 1H NMR (CDCl3): 2.09 (q, 2H,
C(4) of Pyrr); 3.37 (t, J = 7.8, 2H, C(3) of Pyrr); 3.57 (t, J = 7.1, 2H,
C(5) of Pyrr); 4.53 (s, 2H, CH2C6H5); 6.98–7.76 (m, 1H, @CH– and
9 arom. H). Anal. calcd for C18H18ClN3: C 69.34, H 5.82, N 13.48;
found: C 69.02, H 5.78, N 13.40.
5.4. 2-Amino-1-arylazoethenes of structure F and G (66–85):
general method
a) A solution of the relevant glyoxal monoarylhydrazone
(6.75 mmol) and secondary amine (6.5 mmol) in 60 mL of
anhydrous toluene was stirred at room temperature for
3 h. After removing the solvent under vacuum, the solid
was crystallized with the suitable solvent.
5.2.37. 1-Benzyl-2-(4-methylphenylazo)methylenepyrrolidine
(61)
Yield: 50%. Mp 122–124 °C (Et2O). 1H NMR (CDCl3): 2.07 (q, 2H,
C(4) of Pyrr); 2.34 (s, CH3); 3.37 (t, J = 7.6, 2H, C(3) of Pyrr); 3.53 (t,
J = 7.0, 2H, C(5) of Pyrr); 4.51 (s, 2H, CH2C6H5); 7.07–7.63 (m, 1H,
@CH– and 9 arom. H). Anal. calcd for C19H21N3: C 78.32, H 7.26,
N 14.42; found: C 78.02, H 7.24, N 14.38.
b) The required glyoxal monophenylhydrazone was prepared
as described by Severin et al.5; in the same way, glyoxal 4-
chlorophenylhydrazone43
and
glyoxal
4-ethoxycar-
bonylphenylhydrazone were prepared. The latter hydrazone
is new. Yield: 84%. Mp 163–165 °C (Et2O). Anal. calcd for
C
11H12N2O3: C 59.99, H 5.49, N 12.72; found: C 60.19, H
5.55, N 12.71. 1H NMR (CDCl3): 1.26 (t, J = 7.0, CH3 of COOEt);
4.25 (q, J = 7.0, CH2 of COOEt); 7.24 (d, J = 8.7, 2 arom. H);
7.36 (d, J = 9.3, 1H, @CH–); 7.88 (d, J = 8.8, 2 arom. H); 9.48
(d, J = 10.0, 1H, CHO); 10.97 (s, NH, collapses with D2O).
5.2.38. 1-Benzyl-2-(3,4-dichlorophenylazo)methylenepyrroli-
dine and 1-benzyl-3-(3,4-dichlorophenylazo)-1,4,5,6-tetrahy-
dropyridine (mixture 3:1) (62)
Yield: 57%. Mp 113–115 °C. 1H NMR (CDCl3): 1.94 (q, 0.25 ꢂ 2H,
C(5) of Py); 2.10 (q, 0.75 ꢂ 2H, C(4) of Pyrr); 2.61 (t, J = 6.5,
0.25 ꢂ 2H, C(4) of Py); 3.21 (t, J = 6.0, 0.25 ꢂ 2H, C(6) of Py); 3.38
Compounds 66 and 69 were already described.5,6

M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
8463
5.4.1. 1-(4-Chlorophenylazo)-2-(1-pyrrolidino)ethene (67)
5.4.10. 2-[4-(2-Fluorophenyl)piperazin-1-yl]-1-phenylazoe-
Yield: 41%. Mp 138–140 °C (Et2O/pentane). 1H NMR (CDCl3):
1.91–2.08 (m, 4H, C(3, 4) of Pyrr); 3.25–3.57 (m, 4H, C(2, 5) of Pyrr);
7.08 (d, J = 10.8, 1 vinylic H); 7.30 (d, J = 8.7, 2 arom. H); 7.50 (d,
J = 8.6, 2 arom. H); 7.66 (d, J = 10.8, 1 vinylic H). Anal. calcd for
thene (77)
Yield: 68%. Mp 129–131 °C (CH2Cl2/Et2O). 1H NMR (CDCl3): 3.16
(t, J = 5.0, 4H, C(3, 5) of Pz); 3.54 (t, J = 5.0, 4H, C(2, 6) of Pz); 6.85–
7.66 (m, 9 arom. H and 2 vinylic H). Anal. calcd for C18H19FN4: C
69.66, H 6.17, N 18.05; found: C 69.24, H 5.98, N 17.85.
C12H14ClN3: C 61.15, H 5.99, N 17.83; found: C 60.81, H 5.98, N 17.32.
5.4.2. 1-(4-Carbethoxyphenylazo)-2-(1-pyrrolidino)ethene (68)
Yield:80%. Mp 163–164 °C (Et2O). 1H NMR(CDCl3): 1.37(t, J = 7.1,
CH3 of COOEt); 1.84–2,10 (m, 4H, C(3, 4) of Pyrr); 3.35–3.54 (m, 4H,
C(2, 5) of Pyrr); 4.34 (q, J = 7.0, CH2 of COOEt); 7.17 (d, J = 10.8, 1
vinylic H); 7.58 (d, J = 8.6, 2 arom. H); 7.71 (d, J = 10.6, 1 vinylic H);
8.02 (d, J = 8.6, 2 arom. H). Anal. calcd for C15H19N3O2: C 65.91, H
7.01, N 15.37; found: C 66.08, H 7.06, N 15.44.
5.4.11. 2-[4-(4-Fluorophenyl)piperazin-1-yl]-1-phenylazoeth-
ene (78)
Yield: 68%. Mp 168–170 °C (toluene). 1H NMR (CDCl3): 3.17 (t,
J = 5.0, 4H, C(3, 5) of Pz); 3.51 (t, J = 5.2, 4H, C(2, 6) of Pz); 6.85–
7.66 (m,
18H19FN4 + 0.5 H2O: C 67.69, H 6.31, N 17.54; found: C 67.82, H
6.14, N 17.82.
9 arom. H and 2 vinylic H). Anal. calcd for
C
5.4.3. 1-(4-Carbethoxyphenylazo)-2-(1-piperidino)ethene (70)
Yield: 63%. Mp 122–124 °C (Et2O). 1H NMR (CDCl3): 1.37 (t,
J = 7.1, CH3 of COOEt); 1.60–1.74 (m, 6H, C(3, 4, 5) of Pip); 3.30–
3.44 (m, 4H, C(2, 6) of Pip); 4.34 (q, J = 7.0, CH2 of COOEt); 7.30–
7.50 (m, 2 vinylic H); 7.58 (d, J = 8.8, 2 arom. H); 8.03 (d, J = 8.8,
2 arom. H). Anal. calcd for C16H21N3O2: C 66.88, H 7.37, N 14.62;
found: C 66.60, H 7.27, N 14.59.
5.4.12. 2-[4-(2-Chlorophenyl)piperazin-1-yl]-1-phenylazoeth-
ene (79)
Yield: 49%. Mp 83–85 °C (CH2Cl2/Et2O). 1H NMR (CDCl3): 3.13
(t, J = 4.8, 4H, C(3, 5) of Pz); 3.55 (t, J = 4.9, 4H, C(2, 6) of Pz);
6.97–7.65 (m, 9 arom. H and 2 vinylic H). Anal. calcd for
C18H19ClN4: C 66.15, H 5.86, N 17.14; found: 66.00, H 5.96, N 17.30.
5.4.13. 2-[4-(3-Chlorophenyl)piperazin-1-yl]-1-phenylazoe-
thene (80)
5.4.4. 2-(4-Morpholino)-1-phenylazoethene (71)
Yield: 54%. Mp 106 °C (Et2O). 1H NMR (CDCl3): 3.37 (t, J = 4.9,
4H, morpholine); 3.78 (t, J = 4.9, 4H, morpholine); 7.15 (d,
J = 11.0, 1 vinylic H); 7.35–7.70 (m, 5 arom. H and 1 vinylic H).
Anal. calcd for C12H15N3O: C 66.34, H 6.96, N 19.34; found: C
66.03, H 6.85, N 19.25.
Yield: 75%. Mp 134–135 °C (CH2Cl2/Et2O). 1H NMR (CDCl3): 3.27
(t, J = 5.1, 4H, C(3, 5) of Pz); 3.50 (t, J = 5.2, 4H, C(2, 6) of Pz); 6.76–
7.63 (m, 9 arom. H and 2 vinylic H). Anal. calcd for C18H19ClN4: C
66.15, H 5.86, N 17.14; found: 66.19, H 5.92, N 17.28.
5.4.14. 2-[4-(4-Chlorophenyl)piperazin-1-yl]-1-phenylazoe-
thene (81)
5.4.5. 1-Phenylazo-2-(4-thiomorpholino)ethene (72)
Yield: 48%. Mp 88–89 °C (Et2O). 1H NMR (CDCl3): 2.76 (t, J = 5.0,
4H, thiomorpholine); 3.70 (t, J = 5.1, 4H, thiomorpholine); 7.10–
7.70 (m, 5 arom. H and 2 vinylic H). Anal. calcd for C12H15N3S: C
61.77, H 6.48, N 18.01, S 13.74; found: C 61.47, H 6.46, N 18.35,
S 13.43.
Yield: 87%. Mp 197–198 °C (toluene). 1H NMR (CDCl3): 3.22 (t,
J = 5.1, 4H, C(3, 5) of Pz); 3.51 (t, J = 5.2, 4H, C(2, 6) of Pz); 6.80–
7.65 (m, 9 arom. H and 2 vinylic H). Anal. calcd for C18H19ClN4: C
66.15, H 5.86, N 17.14; found: 65.84, H 6.00, N 17.39.
5.4.15. 1-Phenylazo-2-[4-(3-trifluoromethylphenyl)piperazin-
1-yl]ethene (82)
5.4.6. 2-[(4-Ethoxycarbonyl)piperazin-1-yl]-1-phenylazoethene
(73)
Yield: 69%. Mp 146–148 °C (CH2Cl2/Et2O). 1H NMR (CDCl3): 3.32
(t, J = 5.1, 4H, C(3, 5) of Pz); 3.53 (t, J = 5.2, 4H, C(2, 6) of Pz); 7.04–
7.64 (m, 9 arom. H and 2 vinylic H). Anal. calcd for C19H19F3N4: C
63.32, H 5.31, N 15.55; found: C 63.06, H 5.39, N 15.91.
Yield: 53%. Mp 106–107 °C (CH2Cl2/Et2O). 1H NMR (CDCl3): 1.29
(t, J = 7.1, CH3 of COOEt); 3.35 (t, J = 5.2, 4H, C(3, 5) of Pz); 3.60 (t,
J = 5.1, 4H, C(2, 6) of Pz); 4.18 (q, J = 7.1, CH2 of COOEt); 7.11–7.66
(m, 5 arom. H and 2 vinylic H). Anal. calcd for C15H20N4O2: C 62.48,
H 6.99, N 19.43; found: C 62.47, H 7.00, N 19.53.
5.4.16. 2-[4-(4-Nitrophenyl)piperazin-1-yl]-1-phenylazoethene
(83)
5.4.7. 1-Phenylazo-2-[4-(pyridin-2-yl)piperazin-1-yl]ethene
(74)
Yield: 90%. Mp 215–217 °C (CH2Cl2/Et2O). 1H NMR (CDCl3):
3.50–3.61 (m, 8H, C(2, 3, 5, 6) of Pz); 6.86 (d, J = 9.4, 2 arom. H,
p-NO2-phenyl); 7.18–7.65 (m, 5 arom. H and 2 vinylic H); 8.17
(d, J = 9.4, 2 arom. H, p-NO2-phenyl). Anal. calcd for C18H19N5O2:
C 64.08, H 5.68, N 20.76; found: C 64.01, H 5.71, N 20.55.
Yield: 75%. Mp 114–116 °C (CH2Cl2/Et2O). 1H NMR (CDCl3): 3.50
(t, J = 5.1, 4H, C(3, 5) of Pz); 3.71 (t, J = 5.2, 4H, C(2, 6) of Pz); 6.65–
7.65 (m, 5 arom. H, 2 vinylic H and 3H, C(3, 4, 5) of Pyridine); 8.20–
8.26 (m, 1H, C(6) of Pyridine). Anal. calcd for C17H19N5: C 69.60, H
6.53, N 23.87; found: 69.62, H 6.57, N 24.01.
5.4.17. 2-[4-(4-Acetylphenyl)piperazin-1-yl]-1-
phenylazoethene (84)
5.4.8. 1-Phenylazo-2-[4-(pyrimidin-2-yl)piperazin-1-yl]ethene
(75)
Yield: 79%. Mp 188–190 °C (toluene). 1H NMR (CDCl3): 2.55 (s,
CH3CO); 3.45–3.60 (m, 8H, C(2, 3, 5, 6) of Pz); 6.86 (d, J = 9.4, 2
arom. H, p-NO2-phenyl); 7.18–7.65 (m, 5 arom. H and 2 vinylic
H); 8.17 (d, J = 9.4, 2 arom. H, p-NO2-phenyl). Anal. calcd for
Yield: 74%. Mp 140–142 °C (CH2Cl2/Et2O). 1H NMR (CDCl3): 3.46
(t, J = 5.2, 4H, C(3, 5) of Pz); 3.98 (t, J = 5.3, 4H, C(2, 6) of Pz); 6.57 (t,
J = 5.0, 1H, C(5) of Pyrimidine); 7.20–7.65 (m, 5 arom. H and 2 vinylic
H); 8.35 (d, J = 4.9, 2H, C(4, 6) of Pyrimidine). Anal. calcd for
C20H22N4O + 0.5H2O: C 69.64, H 6.75, N 16.31; found: C 70.34, H
6.58, N 16.28.
C16H18N6: C 65.29, H 6.16, N 28.55; found: C 65.01, H 6.14, N 28.79.
5.4.18. 2-[4-(4-Methoxyphenyl)piperazin-1-yl]-1-phenylazoe-
thene (85)
5.4.9. 1-Phenylazo-2-(4-phenylpiperazin-1-yl)ethene (76)
Yield: 77%. Mp 159–161 °C (CH2Cl2/Et2O). 1H NMR (CDCl3): 3.29
(t, J = 5.4, 4H, C(3, 5) of Pz); 3.54 (t, J = 5.5, 4H, C(2, 6) of Pz); 6.90–
7.67 (m, 10 arom. H and 2 vinylic H). Anal. calcd for C18H20N4: C
73.94, H 6.89, N 19.16; found: C 73.64, H 6.97, N 19.40.
Yield: 60%. Mp 173–175 °C (toluene). 1H NMR (CDCl3): 3.16 (t,
J = 5.0, 4H, C(3, 5) of Pz); 3.53 (t, J = 5.1, 4H, C(2, 6) of Pz); 3.79 (s,
CH3O); 6.97–7.65 (m, 9 arom. H and 2 vinylic H). Anal. calcd for
C19H22N4O: C 70.78, H 6.88, N 17.38; found: C 70.51, H 6.81, N 17.39.

8464
M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
5.5. Cell-based assays
To this end, Vero-76 cells were seeded in 24-well plates at a den-
sity of 2 ꢂ 105 cells/well and were allowed to form confluent mon-
olayers by incubating overnight in growth medium at 37 °C in a
humidified CO2 (5%) atmosphere. Then, monolayers were infected
5.5.1. Compounds
Compounds were dissolved in DMSO at 100 mM and then di-
luted in culture medium.
with 250
lL of proper virus dilutions to give 50–100 PFU/well. Fol-
lowing removal of unadsorbed virus, 500
lL of Dulbecco’s modified
5.5.2. Cells and viruses
Eagle’s medium supplemented with 1% inactivated FCS and 0.75%
methyl-cellulose, without or with serial dilutions of test com-
pounds, were added. Cultures were incubated at 37 °C for 2 (Sb-1
and VSV), 3 (CVB-2, VV and HSV-1) or 5 days (RSV) and then fixed
with PBS containing 50% ethanol and 0.8% crystal violet, washed
and air-dried. Plaques were then counted. EC50 (50% effective con-
centration) was calculated by linear regression technique.
Cell lines were purchased from American Type Culture Collec-
tion (ATCC). The absence of mycoplasma contamination was
checked periodically by the Hoechst staining method. Cell lines
supporting the multiplication of RNA and DNA viruses were the
following: CD4+ human T cells containing an integrated HTLV-1
genome (MT-4); Madin Darby bovine kidney (MDBK); baby ham-
ster kidney (BHK-21) and monkey kidney (Vero-76) cells.
5.6. Molecular modeling
5.5.3. Cytotoxicity assays
For cytotoxicity tests, run in parallel with antiviral assays,
MDBK and BHK cells were resuspended in 96 multiwell plates at
an initial density of 6 ꢂ 105 and 1 ꢂ 106 cells/mL, respectively, in
maintenance medium, without or with serial dilutions of test com-
pounds. Cell viability was determined after 48–96 h at 37 °C in a
humidified CO2 (5%) atmosphere by the 3-(4,5-dimethylthiazol-
2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) method.44 Vero-76
cells were resuspended in 24 multiwell plates at an initial density
of 4 ꢂ 105 cells/mL. The cell number of Vero-76 monolayers was
determined by staining with the crystal violet dye.
For cytotoxicity evaluations, exponentially growing cells de-
rived from human hematological tumors [CD4+ human T-cells con-
taining an integrated HTLV-1 genome (MT-4)] were seeded at an
initial density of 1 ꢂ 105 cells/mL in 96-well plates in RPMI-1640
medium supplemented with 10% fetal calf serum (FCS), 100 U/mL
5.6.1. Pharmacophore modeling
The model structures of all compounds were built using the
CATALYST 2D–3D sketcher,45 and the conformational search of each
compound was carried out using the Poling algorithm46–48 and
generalized CHARMM force field49 as implemented in the program.
Molecular flexibility was taking into account by considering each
compound as a collection of conformers representing a different
area of conformational space accessible to the molecule within a
given energy range. The ‘best quality’ generation option was
adopted to select representative conformers over a 0–20 kcal/mol
interval above the computed global energy minimum in the con-
formational space. Since the number of conformers generated for
each compound has been limited to a maximum of 250, this search
procedure should identify the best three-dimensional arrangement
of chemical functions explaining the activity variations among the
compound training set.
penicillin G and 100 lg/ml streptomycin. Cell cultures were then
incubated at 37 °C in a humidified, 5% CO2 atmosphere in the ab-
sence or presence of serial dilutions of test compounds. Cell viabil-
ity was determined after 96 h at 37 °C by the MTT method.
Based on the conformations for each compound, CATALYST 4.9
software package was used to generate plausible three-dimen-
sional pharmacophore models. During hypotheses generation, the
software attempts to minimize a cost function of two main terms:
the first penalizes the deviation between the estimated activities of
the training set molecules and their experimental values, while the
second penalizes the complexity of the hypothesis. The uncertain
factor for each compound represents the ratio range of uncertainty
in the activity value based on the expected statistical straggling of
biological data collection. Uncertainty influences the first step—
also called the constructive phase—of the hypothesis generating
process.50 In this work, an uncertainty of 2.0 was preferred over
the default factor of 3.0, as the experimental activities of our com-
pounds barely span the required four orders of magnitude.
An analysis of the functional groups characterizing our com-
pounds suggested that hydrophobic aromatic (HAr) sites, hydrogen
bond acceptors (HBA), hydrogen bond donors (HBD), positive ion-
izable points (PI), and ring aromatic (RAr) sites could effectively
map their critical chemical features and, hence, describe the anti-
viral activity of our compounds. Accordingly, these five features
were selected to constitute the essential information in the auto-
mated hypothesis generation process.
Two validation procedures were followed to determine the sta-
tistical relevance and the validity of the proposed 3D pharmaco-
phore models: the test set prediction model, and the Cat-
Scramble method. In this work, the former procedure consisted in
the collection of further, different compounds into three sets
(TsS), and in performing a regression analysis by mapping each
TsS onto the best pharmacophore hypothesis for each virus. The
obtainment of high correlation coefficients generated using the
test set compounds revealed the good correlation between the ac-
tual and estimated activities and, hence, the predictive validity of
the corresponding 3D hypothesis. The Cat-Scramble validation pro-
cedure is based on Fisher’s randomization test.51 The goal of this
5.5.4. Antiviral assay
Activity of compounds against human immunodeficiency virus
type-1 (HIV-1) was based on inhibition of virus-induced cytopath-
ogenicity in MT-4 cells acutely infected with a multiplicity of infec-
tion (m.o.i.) of 0.01. Briefly, 50
were added to each well of flat-bottom microtitre trays containing
l
L of RPMI containing 1 ꢂ 104 MT-4
50
lL of RPMI, without or with serial dilutions of test compounds.
Then, 20
lL of a HIV-1 suspension containing 100 CCID50 were
added. After a 4-day incubation, cell viability was determined by
the MTT method.
Activity of compounds against yellow fever virus (YFV) and Reo
virus type-1 (Reo-1) was based on inhibition of virus-induced cyto-
pathogenicity in acutely infected BHK-21 cells. Activities against
bovine viral diarrhoea virus (BVDV), in infected MDBK cells, were
also based on inhibition of virus-induced cytopathogenicity.
BHK and MDBK cells were seeded in 96-well plates at a density
of 5 ꢂ 104 and 3 ꢂ 104 cells/well, respectively, and were allowed to
form confluent monolayers by incubating overnight in growth
medium at 37 °C in a humidified CO2 (5%) atmosphere. Cell mono-
layers were then infected with 50
lL of a proper virus dilution (in
serum-free medium) to give a m.o.i = 0.01. One hour later, 50
lL of
MEM Earle’s medium, supplemented with inactivated fetal calf ser-
um (FCS), 1% final concentration, without or with serial dilutions of
test compounds, were added. After 3–4 days of incubation at 37 °C,
cell viability was determined by the MTT method.
Activity of compounds against coxsackie virus type B2 (CVB-2),
Polio Virus type-1 Sabin strain (Sb-1), vesicular stomatitis virus
(VSV), vaccinia virus (VV), herpes virus 1 (HSV-1) and respiratory
syncytial virus (RSV), A-2 strain, in infected Vero-76 cells, was
determined by plaque reduction assays in Vero-76 cell monolayers.

M. Tonelli et al. / Bioorg. Med. Chem. 16 (2008) 8447–8465
8465
Florio, T. Int. J. Immunopathol. Pharmacol. 2006, 19, 339. and references
reported therein.
16. Kimball, D. B.; Haley, M. M. Angew. Chem. Int. Ed. 2002, 41, 3338.
17. Gesher, J. A.; Hickman, J. A.; Simmonds, R. J.; Stevens, M. F. G.; Vaughan, K.
Biochem. Pharmacol. 1981, 30, 89.
18. Stevens, M. F. G.; Phillip, K. S.; Rathbone, D. L.; O’Shea, D. M.; Queener, S. F.;
Schwalbe, C. H.; Lambert, P. A. J. Med. Chem. 1997, 40, 1886.
19. Prous, J.; Graul, A.; Castañer, J. Drugs Future 1994, 19, 746.
20. Jean-Claude, B. J.; Just, G. J. Chem. Soc. Perkin Trans. I 1991, 2525.
21. Jean-Claude, B. J.; Just, G. Heterocycles 1998, 48, 1347.
22. Jean-Claude, B. J.; Mustafa, A.; Damian, Z.; De Marte, J.; Vasilescu, D. E.; Yen, R.;
Chan, T. H.; Leyland-Jones, B. Biochem. Pharmacol. 1999, 288, 753.
23. Jean-Claude, B. J.; Mustafa, A.; Watson, A. J.; Damian, Z.; Vasilescu, D.; Chan, T.
H.; Leyland-Jones, B. J. Pharmacol. Exp. Ther. 1999, 288, 484.
24. Dumont-Hornebeck, B.; Strube, Y. N.; Vasilescu, D.; Jean-Claude, B. J. Bioorg.
Med. Chem. Lett. 2000, 10, 2325.
25. Miyadera, T. In The Chemistry of Hydrazo, Azo and Azoxy Groups; Patai, S., Ed.; J.
Wiley: London, New York, 2002; pp 495–539.
26. Polanc, S. J. Heterocyclic Chem. 2005, 42, 401.
27. Reich, M. F.; Fabio, P. F.; Lee, V. J.; Kuck, N. A.; Testa, R. T. J. Med. Chem. 1989, 32,
2474.
28. Barbieri, F.; Sparatore, A.; Alama, A.; Novelli, F.; Bruzzo, C.; Sparatore, F. Oncol.
Res. 2003, 13, 455.
type of validation is to check whether there is a strong correlation
between the chemical structures and the biological activity. This is
done by randomizing the activity data associated with the training
set compounds, generating pharmacophore hypothesis using the
same features and parameters employed to develop the original
pharmacophore model. The statistical significance is calculated
according to the following formula:
significance ¼ 100 ꢂ ½1 ꢀ ð1 þ x=yÞꢃ
where x is the total number of hypotheses having a total cost lower
than the original (best) hypothesis, and y is the total number of
HypoGen runs (initial + random runs). Thus, for instance, 49 random
spreadsheets (i.e., 49 HypoGen runs) have to be generated to obtain
a 98% confidence level. Should any randomized data set result in the
generation of a 3D pharmacophore with similar or even better cost
values, root-mean-square deviations, and correlation coefficients,
then it is likely that the original hypothesis does reflect a chance
correlation.
29. Sparatore, F.; Cerri, R. J. Heterocyclic Chem. 1979, 16, 1001.
30. Sparatore, F.; Boido, A. Farmaco 1991, 46(Suppl.), 233.
31. Novelli, F.; Sparatore, F. Farmaco 2002, 57, 871.
Acknowledgments
32. Gault, F.; Wahl, H. Compt. Rend. 1955, 240, 983.
33. Sommer, S. Angew. Chem. Int. Ed. 1977, 16, 58.
Financial support from Italian MIUR (FIRB RBNE01J3SK01) is
gratefully acknowledged. The authors thank O. Gagliardo for per-
forming elemental analyses.
34. Reckhow, W. A.; Tarbell, D. S. J. Am. Chem. Soc. 1952, 74, 4960.
35. Kunieda, T.; Koga, K.; Yamada, S.-I. Chem. Pharm. Bull. 1967, 15, 337.
36. Hirsch, J. A.; Jarmas, A. A. J. Org. Chem. 1978, 43, 4106.
37. Haylor, B. D. G.; Smith, P. H. G. Eur. Pat. Appl. EP 634,413, 1995; Chem. Abstr.
1995, 122, 214088.
References and notes
38. Nagarajan, K.; Talwalker, P. K.; Kulkarni, C. L.; Venkateswarlu, A.; Prabhu, S. S.;
Nayak, G. V. Ind. J. Chem. Sect. B 1984, 23B, 1243.
39. Challand, S. R.; Copp, F. C.; Denyer, C. V.; Eakins, K. E.; Walker, J. M. G.;
Whittaker, N.; Caldwell, A. G. Eur. Pat. Appl. EP 22,578, 1981; Chem. Abstr. 1981,
95, 62193.
40. Enders, E. In Houben-Weyl; Müller, E., Ed.; Methoden der organischen Chemie;
G. Thieme Verlag: Stuttgart, 1967; Vol. 10/II, pp 201–214.
41. Plancher, G. Gazz. Chim. It. 1898, 28 II, 418.
1. Sparatore, F.; Boido, V. Farmaco 1988, 43, 1099.
2. Boido Canu, C.; Boido, V.; Sparatore, F.; Sparatore, A.; Susanna, V.; Russo, S.;
Cenicola, M. L.; Marmo, E. Farmaco, Ed. Sci. 1988, 43, 819.
3. Rosenhauer, E. Chem. Ber. 1924, 57, 1192.
4. König, W.; Müller, J. Chem. Ber. 1924, 57, 144.
5. Severin, T.; Adam, R.; Lerche, H. Chem. Ber. 1975, 108, 1756.
6. Lerche, H.; Severin, T. Chem. Ber. 1978, 111, 1195.
7. Crary, J. W.; Quayle, O. R.; Lester, C. T. J. Am. Chem. Soc. 1956, 78, 5584.
8. Sparatore, A.; Canu Boido, C.; Boido, V.; Sparatore, F.; Debbia, E.; Pesce Schito, A.
Farmaco 1990, 45, 867.
42. Colonna, M.; Bruni, P. Gazz. Chim. It. 1965, 95, 857.
43. Clark, J. D.; Heise, J. D.; Shah, A. S.; Peterson, J. C.; Chou, S. K.; Levine, J.; Karakas,
A. M.; Ma, Y.; Ng, K.-Y.; Patelis, L.; Springer, J. R.; Stano, D. R.; Wettach, R. H.;
Dutra, G. A. Org. Proc. Res. Dev. 2004, 8, 176.
44. Pauwels, R.; Balzarini, J.; Baba, M.; Snoeck, R.; Schols, D.; Herdewijn, P.;
Desmyster, J.; De Clercq, E. J. Virol. Methods 1988, 20, 309.
45. CATALYST 4.9; Accelrys Inc., San Diego, CA, USA, 2004.
46. Smellie, A.; Teig, S. L.; Towbin, P. J. Comput. Chem. 1994, 16, 171.
47. Smelie, A.; Kahn, S. D.; Teig, S. L. J. Chem. Inf. Comput. Sci. 1995, 35, 285.
48. Smelie, A.; Kahn, S. D.; Teig, S. L. J. Chem. Inf. Comput. Sci. 1995, 35, 295.
49. Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J.; Swaminathan, S.;
Karplus, M. J. J. Comput. Chem. 1983, 4, 187.
9. Canu Boido, C.; Boido, V.; Sparatore, F.; Sparatore, A.; Bombieri, G.; Benetollo,
F.; Debbia, E.; Pesce Schito, A. Farmaco 1993, 48, 749.
10. Elslager, E. F.; Worth, D. F. J. Med. Chem. 1963, 6, 444.
11. Elslager, E. F.; Capps, D. B.; Kurz, D. H.; Short, F. W.; Werbel, L. M.; Worth, D. F. J.
Med. Chem. 1966, 9, 378.
12. Werbel, L. M.; Elslager, E. F.; Fisher, M. W.; Gavrilis, Z. B.; Phillips, A. A. J. Med.
Chem. 1968, 11, 411.
13. Vazzana, I.; Sparatore, F.; Fadda, G.; Manca, C. Farmaco 1993, 48, 737.
14. Milhavet, O.; Mange, A.; Casanova, D.; Lehmann, S. J. Neurochem. 2000, 74, 222.
15. Corsaro, A.; Paludi, D.; Villa, V.; D’Arrigo, C.; Chiovitti, K.; Thellung, S.;
Russo, C.; Di Cola, D.; Ballerini, P.; Patrone, E.; Schettini, G.; Aceto, A.;
50. Güner, O. F. In Pharmacophore Perception Development and Use in Drug Design;
International University Line: La Jolla, CA, USA, 2000; p 173.
51. Fisher, R. The Design of Experiments; Hafner Publishing: New York, NY, USA, 1966.