Geometrically and conformationally restrained cinnamoyl compounds as inhibitors of HIV-1 integrase: Synthesis, biological evaluation, and molecular modeling

Article abstract of DOI:10.1021/jm9707232

Various cinnammoyl-based structures were synthesized and tested in enzyme assays as inhibitors of the HIV-1 integrase (IN). The majority of compounds were designed as geometrically or conformationally constrained analogues of caffeic acid phenethyl ester (CAPE) and were characterized by a syn disposition of the carbonyl group with respect to the vinylic double bond. Since the cinnamoyl moiety present in flavones such as quercetin (inactive on HIV-1-infected cells) is frozen in an anti arrangement, it was hoped that fixing our compounds in a syn disposition could favor anti-HIV-1 activity in cell-based assays. Geometrical and conformational properties of the designed compounds were taken into account through analysis of X-ray structures available from the Cambridge Structural Database. The polyhydroxylated analogues were prepared by reacting 3,4-bis(tetrahydropyran- 2-yloxy)benzaldehyde with various compounds having active methylene groups such as 2-propanone, cyclopentanone, cyclohexanone, 1,3-diacetylbenzene, 2,4- dihydroxyacetophenone, 2,3-dihydro-1-indanone, 2,3-dihydro-1,3-indandione, and others. While active against both 3'-processing and strand-transfer reactions, the new compounds, curcumin included, failed to inhibit the HIV-1 multiplication in acutely infected MT-4 cells. Nevertheless, they specifically inhibited the enzymatic reactions associated with IN, being totally inactive against other viral (HIV-1 reverse transcriptase) and cellular (RNA polymerase II) nucleic acid-processing enzymes. On the other hand, title compounds were endowed with remarkable antiproliferative activity, whose potency correlated neither with the presence of catechols (possible source of reactive quinones) nor with inhibition of topoisomerases. The SARs developed for our compounds led to novel findings concerning the molecular determinants of IN inhibitory activity within the class of cinnamoyl-based structures. We hypothesize that these compounds bind to IN featuring the cinnamoyl residue C=C-C=O in a syn disposition, differently from flavone derivatives characterized by an anti arrangement about the same fragment. Certain inhibitors, lacking one of the two pharmacophoric catechol hydroxyls, retain moderate potency thanks to nonpharmacophoric fragments (i.e., a m-methoxy group in curcumin) which favorably interact with an 'accessory' region of IN. This region is supposed to be located adjacent to the binding site accommodating the pharmacophoric dihydroxycinnamoyl moiety. Disruption of coplanarity in the inhibitor structure abolishes activity owing to poor shape complementarity with the target or an exceedingly high strain energy of the coplanar conformation.

Full text of DOI:10.1021/jm9707232

3948
J . Med. Chem. 1998, 41, 3948-3960
Geom etr ica lly a n d Con for m a tion a lly Restr a in ed Cin n a m oyl Com p ou n d s a s
In h ibitor s of HIV-1 In tegr a se: Syn th esis, Biologica l Eva lu a tion , a n d Molecu la r
Mod elin g
Marino Artico,*,† Roberto Di Santo,† Roberta Costi,† Ettore Novellino,‡ Giovanni Greco,§ Silvio Massa,|
Enzo Tramontano, Maria E. Marongiu, Antonella De Montis, and Paolo La Colla*,
Istituto Pasteur - Fondazione Cenci Bolognetti, Dipartimento di Studi Farmaceutici, Universita` di Roma “La Sapienza”,
Piazzale A. Moro 5, I-00185 Roma, Italy, Dipartimento di Scienze Farmaceutiche, Universita` di Salerno, Piazza Vittorio
Emanuele 9, I-84084 Penta (Salerno), Italy, Dipartimento di Chimica Farmaceutica e Tossicologica, Universita` di Napoli
“Federico II”, Via D. Montesano 49, I-80131 Napoli, Italy, Dipartimento Farmaco Chimico Tecnologico, Universita` di Siena,
Banchi di Sotto 55, I-53100 Siena, Italy, and Dipartimento di Biologia Sperimentale, Universita` di Cagliari, Viale Regina
Margherita 45, I-09124 Cagliari, Italy
Received October 23, 1997
Various cinnammoyl-based structures were synthesized and tested in enzyme assays as
inhibitors of the HIV-1 integrase (IN). The majority of compounds were designed as
geometrically or conformationally constrained analogues of caffeic acid phenethyl ester (CAPE)
and were characterized by a syn disposition of the carbonyl group with respect to the vinylic
double bond. Since the cinnamoyl moiety present in flavones such as quercetin (inactive on
HIV-1-infected cells) is frozen in an anti arrangement, it was hoped that fixing our compounds
in a syn disposition could favor anti-HIV-1 activity in cell-based assays. Geometrical and
conformational properties of the designed compounds were taken into account through analysis
of X-ray structures available from the Cambridge Structural Database. The polyhydroxylated
analogues were prepared by reacting 3,4-bis(tetrahydropyran-2-yloxy)benzaldehyde with various
compounds having active methylene groups such as 2-propanone, cyclopentanone, cyclohex-
anone, 1,3-diacetylbenzene, 2,4-dihydroxyacetophenone, 2,3-dihydro-1-indanone, 2,3-dihydro-
1,3-indandione, and others. While active against both 3′-processing and strand-transfer
reactions, the new compounds, curcumin included, failed to inhibit the HIV-1 multiplication
in acutely infected MT-4 cells. Nevertheless, they specifically inhibited the enzymatic reactions
associated with IN, being totally inactive against other viral (HIV-1 reverse transcriptase) and
cellular (RNA polymerase II) nucleic acid-processing enzymes. On the other hand, title
compounds were endowed with remarkable antiproliferative activity, whose potency correlated
neither with the presence of catechols (possible source of reactive quinones) nor with inhibition
of topoisomerases. The SARs developed for our compounds led to novel findings concerning
the molecular determinants of IN inhibitory activity within the class of cinnamoyl-based
structures. We hypothesize that these compounds bind to IN featuring the cinnamoyl residue
CdC-CdO in a syn disposition, differently from flavone derivatives characterized by an anti
arrangement about the same fragment. Certain inhibitors, lacking one of the two pharma-
cophoric catechol hydroxyls, retain moderate potency thanks to nonpharmacophoric fragments
(i.e., a m-methoxy group in curcumin) which favorably interact with an “accessory” region of
IN. This region is supposed to be located adjacent to the binding site accommodating the
pharmacophoric dihydroxycinnamoyl moiety. Disruption of coplanarity in the inhibitor
structure abolishes activity owing to poor shape complementarity with the target or an
exceedingly high strain energy of the coplanar conformation.
In tr od u ction
In addition to RT and PR, another target potentially
amenable to a selective chemotherapeutic intervention
is the HIV-1 integrase (IN), an enzyme which has no
counterparts in the host cell and catalyzes an essential
step in the retroviral life cycle.4,5 In fact, following
reverse transcription of the viral DNA genome into a
double-stranded DNA (dsDNA), IN catalyzes integration
of the latter into the host chromosome through coordi-
nated reactions of processing and joining.6 Initially the
enzyme recognizes the LTR termini of the viral dsDNA,
of which it removes the last two nucleotides (GT) leaving
two recessed 3′-OH ends. Then, IN catalyzes joining of
the processed viral 3′-ends to the 5′-ends of strand
breaks in the host DNA. Removal of mispaired nucle-
Monotherapy with available drugs, which target the
retroviral enzymes reverse transcriptase (RT)1 and
protease (PR),2 has failed to provide long-term suppres-
sion of HIV-1 replication in infected individuals. To
obtain a more effective treatment of AIDS, efforts are
actually directed toward the design of combination
therapies and the exploitation of new targets.3
* Corresponding authors.
† Universita` di Roma “La Sapienza”.
‡ Universita` di Salerno.
§ Universita` di Napoli “Federico II”.
| Universita` di Siena.
Universita` di Cagliari.
S0022-2623(97)00723-1 CCC: $15.00 © 1998 American Chemical Society
Published on Web 09/17/1998

Cinnamoyl Compounds as Inhibitors of HIV-1 IN
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21 3949
Ch a r t 1. HIV-1 Integrase Inhibitors
Ch a r t 2. Cinnamoyl-Based Derivatives Investigated as
IN Inhibitors
F igu r e 1. Torsional angles τ1 and τ2 for 3,4-dihydroxycin-
namoyl-like structures are defined by thick lines. Syn and anti
mutual dispositions of CdC and CdO groups correspond to τ2
values of about 0° and 180°, respectively.
otides and gap repair, which are carried out by cellular
enzymes, leads to provirus formation.7
ascribed to the presence of an o-methoxyhydroxy, rather
than the o-dihydroxy system which is known to correlate
with potent anti-IN activity.
Many different classes of compounds have been
reported to inhibit the HIV-1 IN in enzyme assays, and
among them, natural and synthetic polyhydroxylated
molecules have emerged as potent IN inhibitors (Chart
1). They include flavones such as quercetin (1),8 caffeic
acid phenethyl ester (CAPE) (2) and related amides,8,9
bis-catechols,10 curcumin (3),11 a tyrphostin (4) and its
potent 6,7-dihydroxynaphtho-2-yl derivative (5),12 (-)-
arctigenin,13 and L-chicoric (6) and dicaffeoylquinic
acids.14 However, very few of the above compounds
have shown specific anti-IN activity, and still fewer
proved active in cell-based assays. Notable exceptions
are L-chicoric acid and some dicaffeoylquinic acids
isolated from Bolivian medicinal plants which, in ad-
dition to being very potent and specific IN inhibitors in
enzyme assays, inhibit the HIV-1 multiplication in
acutely infected cells at noncytotoxic concentrations.14
Interestingly, the latter compounds are characterized
by a 3,4-dihydroxycinnamoyl moiety which is also
present in most of the above IN inhibitors, sometimes
incorporated in a ring structure as in quercetin and 6,7-
dihydroxynaphtho-2-yltyrphostin. The only other poly-
hydroxylated IN inhibitor which has been reported
Although the 3D structure of HIV-1 IN has been
partly solved by X-ray crystallography,16 target structure-
based design of novel inhibitors is currently hampered
by the fact that one of the most important domains of
this enzyme, namely, the one containing the catalytic
residue Glu-152, has been found disordered to such a
great extent that its structure could not be determined.
Therefore, design of novel IN inhibitors and interpreta-
tion of structure-activity relationships (SARs) for this
class of compounds are still ligand structure-based.
With the aim of obtaining more potent anti-IN inhibi-
tors, hopefully active also in HIV-1-infected cells, we
synthesized various polyhydroxylated cinnamoyl-con-
taining derivatives (7-13) (Chart 2) and tested them
for activity against HIV-1 IN, RT, and cellular RNA
polymerase in enzyme assays as well as for anti-HIV-1
antiproliferative and antitopoisomerase activity in cell-
based assays.
Most of our compounds were designed as geometri-
cally or conformationally constrained cinnamoyl struc-
tures characterized by a syn disposition of the carbonyl
group with respect to the vinylic double bond (Figure
1). The rationale of “freezing” the CdC-CdO moiety
in a syn arrangement relied on the observation that
flavones such as quercetin (1) are inactive on HIV-1-
infected cells (in all of them the CdC-CdO fragment
moderately active in cell-based assays is curcumin,15
a
pigment isolated from the rhizomes of Curcuma longa,
which exhibits a variety of pharmacological properties,
antiinflammatory and antitumor activities included.
However, the low potency of curcumin can be likely

3950 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21
Artico et al.
Sch em e 1
Sch em e 2
reacting 3,4-dihydroxybenzaldehyde with piperidin-4-
one hydrochloride in glacial acetic acid under a stream
of gaseous hydrochloric acid (Scheme 2).
Resu lts a n d Discu ssion
Molecu la r Design . As stated, analysis of crystal
structures retrieved from the Cambridge Structural
Database (CSD)16 allowed us to take into account
geometrical and conformational features of the designed
compounds. These were synthesized and tested in order
to address the following SAR issues.
The effect of constraining in a syn disposition the Cd
C and CdO fragments of the cinnamoyl chain (Figure
1) was evaluated by synthesis of the geometrically
constrained cyclovalone analogues 7a -f and indan
derivatives 8a ,b. Compounds 9a -d , forced into a syn
arrangement by conformational factors, were synthe-
sized as open-chain counterparts of 7a -f. Through
7a -f and 9a -d we also investigated symmetrical cin-
namoyl moieties sharing the same carbonyl group.
Chalcone derivatives 10a -d , stable as syn conformers,
represented a further step toward molecular simplifica-
tion as they formally result from removal of one of the
two vinylic moieties in compounds 9a -d .
Effects on anti-IN activity related to an increased
distance between the two terminal aromatic moieties
were evaluated by synthesis of curcumin analogues
11a -d and especially of 1,3-bis(cinnamoyl)benzene
derivatives 12a ,b. In these molecules, both syn and anti
conformations are energetically feasible.
Preparation and testing of 13, which cannot attain a
fully coplanar conformation, provided information con-
cerning shape-dependent effects relevant to IN inhibi-
tory activity.
is geometrically fixed in anti) whereas some flexible
CAPE derivatives, like L-chicoric acid (6), display anti-
HIV-1 activity in cultured cells. It was hoped that
stabilization of the syn arrangement in CAPE ana-
logues, through geometrical (bond connectivity) or con-
formational constraints, could yield novel IN inhibitors
effective in cell-based assays. The design of the new
inhibitors was assisted by analysis of X-ray structures
retrieved from the Cambridge Structural Database
(CSD)17 of compounds coincident or congeneric with
those synthesized and tested.
Ch em istr y
Curcumin and related compounds 11a -d were pre-
pared according to Babu and Rajasekharan by conden-
sation of the proper arylaldehyde with acetylacetone and
boric acid in the presence of 1,2,3,4-tetrahydroquino-
line.18 For the synthesis of the other arylidene deriva-
tives, the use of 3,4-dihydroxybenzaldehyde required
preventive protection of the hydroxyl groups of the
catechol system, preferably with easily removable groups
such as tetrahydropyran-2-yl.19 We therefore reacted
3,4-dihydroxybenzaldehyde with 3,4-dihydro-R-pyran in
the presence of pyridinium p-toluenesulfonate to achieve
3,4-bis(tetrahydropyran-2-yloxy)benzaldehyde (Scheme
1).
A correlation has been repeatedly found between the
anti-IN potency of polyhydroxylated aryl compounds
and the number and/or mutual positioning of hydroxyl
groups, the most active compounds being those contain-
ing two pairs of o-bis-hydroxy substituents.21 To verify
whether such a pattern held likewise in our series, so
as to support a mechanism of action similar to that of
other CAPE analogues, the designed compounds differed
also in the number of free hydroxyl groups.
An ti-IN Assa ys. All the synthesized compounds
were initially tested in a dual 3′-processing and strand-
transfer assay together with aurintricarboxylic acid
(ATA), used as reference compound.22 In this study, a
purified HIV-1 IN expressed in Escherichia coli was
incubated with a 21-mer double-stranded oligonucle-
otide corresponding to the U5 end of the HIV-1 LTR,
labeled at the 3′-end of the plus strand. This allowed a
Polyhydroxylated derivatives were then prepared by
reacting 3,4-bis(tetrahydropyran-2-yloxy)benzaldehyde
with various compounds having active methylene groups
such as 2-propanone, cyclopentanone, cyclohexanone
(reported in Scheme 1 as a representative example of
the general procedure), 1,3-diacetylbenzene, 2,4-dihy-
droxyacetophenone, 2,3-dihydro-1-indanone, 2,3-dihy-
dro-1,3-indandione, and others. Condensation was car-
ried out in alkaline medium according to the published
procedure.20
The crude chalcone-like derivatives obtained from
condensation were subjected to deprotection by hydro-
lytic cleavage in the presence of p-toluenesulfonic acid
as a catalyst. This procedure was used for achieving
the mono- and bis-arylidene derivatives 7-13 described
in Tables 1 and 2, whereas 3,5-bis(3,4-dihydroxyben-
zylidene)piperidin-4-one (7f) was easily obtained by

Cinnamoyl Compounds as Inhibitors of HIV-1 IN
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21 3951
F igu r e 2. Inhibition of HIV-1 IN-catalyzed 3′-processing and strand-transfer reactions by curcumin and related derivatives.
The DNA strand-transfer products (upper pannel) migrate more slowly than 21-mer substrate and 3′-processing products G, C,
and L (lower panel). The upper panel is a darker exposure of the gel in order to show the integration products. Lane 1, without
DNA; lane 2, +IN without drugs; lanes 3-6, with 12b (100, 50, 25, 12.5 µM); lanes 7-10, with 9c (1.2, 0.6, 0.3, 0.15 µM); lanes
11-14, with 7d (0.6, 0.3, 0.15, 0.07 µM); lanes 15-18, with 10d (10, 5, 2.5, 1.2 µM); lanes 19-22, with 8a (10, 5, 2.5, 1.2 µM);
lanes 23-25, with 3 (50, 25, 12.5 µM); lanes 26-28, with ATA (1.2, 0.6, 0.3 µM).
Ta ble 1. Chemical and Physical Data of Derivatives 7-13
rapid and quantitative recovery, by gel filtration, of the
acid-soluble dinucleotides released by IN in the 3′-
recrystna yield reaction chromatb
compd
mp (°C)
solvent
(%)
time (h)
system
ref
processing reaction and, thus, a very accurate compari-
son of the anti-IN potencies of test compounds. On the
other hand, separation of reaction products by denatur-
ing PAGE and densitometric quantification allowed us
to measure the extent of inhibition of both 3′-processing
and strand-transfer reactions. A typical experiment is
shown in Figure 2.
7a
7b
7c
7d
7e
7f
8a
8b
9a
9b
9c
9d
10a
10b
10c
10d
11a
11b
11c
11d
12a
12b
13
179-181
244-246
>300
a
b
b
c
d
e
c
c
e
e
f
c
c
c
a
d
e
g
h
i
22
99
42
46
98
92
53
56
64
18
24
74
66
39
93
100
55
10
26
19
48
26
24
0.5
7
1
3.5
1
48
4
3
1
72
2
2
15
19c
12
15
4
4
4
4
4
33
40
41
A
B
B
B
>300
223-225
>300
262-264
264-266
68-70
42
The IC50 values presented in the SAR study of Table
2 refer to 3′-processing reactions analyzed by gel filtra-
tion. However, it should be noted that essentially
similar IC50 values were obtained when: (a) 3′-process-
ing and strand-transfer reactions were monitored by
PAGE and densitometric quantification; (b) the strand-
transfer reaction was carried out with a 5′-labeled
substrate (see Experimental Section) corresponding to
the product of the 3′-processing reaction (data not
shown).
C
C
D
43
44
114-116
221-223
248-250
212-214
232-234
209-210
199-201
128-130
193-195
216-218
136-137
>300
45
46
47
19
48
18
20
49
50
E
E
F
G
H
H
C
A
I
c
c
f
Sequence-specific removal of the labeled terminal
dinucleotide by intervention of nucleophiles, such as
glycerol, 3′-OH groups of the substrate DNA, and water,
generates 5′-glycerol-phosphate-GTOH (G), cyclic-GT (C),
and 5′-phosphate-GTOH (L), respectively, as reaction
products.23 When formation of the latter was analyzed
by denaturing PAGE gels, like those illustrated in
Figure 2, all the active compounds were found to inhibit
the 3′-processing reaction in a dose-dependent manner
(lower panel). Moreover, the fact that our compounds
inhibited the formation of G, C, and L to the same extent
suggested that all of them blocked nucleophilic attack,
likely binding at the IN active site.
210-211
160-162
4
6
A
a
Recrystallization solvents: a, acetic acid; b, methanol/water;
c, methanol; d, tetrahydrofuran/toluene; e, ethanol; f, ethanol/
b
toluene; g, toluene; h, toluene/benzene; i, cyclohexane. Chro-
matographic system silica gel: A, chloroform/methanol, 20:1, B,
chloroform/methanol; C, chloroform; D, chloroform/ethyl acetate,
1:1; E, diethyl ether/n-hexane, 1:1; F, diethyl ether/n-hexane, 3:2;
G, diethyl ether/n-hexane, 2:1; H, chloroform/ethyl acetate, 9:1; I,
chloroform/methanol, 20:1. c Reflux time.
A longer exposure of the gel (Figure 2, upper panel)
allowed us to see that active compounds also inhibited
the strand-transfer reaction in a dose-dependent man-

3952 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21
Artico et al.
Ta ble 2. Structures and Biological Properties of Cinnamoyl-Based Derivatives

Cinnamoyl Compounds as Inhibitors of HIV-1 IN
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21 3953
Ta ble 2 (Continued)
a
Compound concentration (µM) required to reduce HIV-1 IN 3′-processing activity by 50%. Data represent mean values ( SD for three
independent determinations. Compound concentration (µM) required to reduce the exponential growth of MT-4/KB cells by 50%. c Fold
b
increase in protein-linked DNA breaks (PLDB) with respect to untreated controls. Compound concentrations were 10 × CC50 for KB cells.
d
CC50 and IC50 are expressed as µg/mL.
ner, as evidenced by a decrease in the amount of larger
reaction products.
IN In h ibitor y Activities. The IC50 values listed in
Table 2 range from 0.2 to above 100 µM (threshold value
for compounds considered inactive). The highest po-
tency was obtained with cyclovalone analogues 7b-f in

3954 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21
Artico et al.
which the symmetric dihydroxycinnamoyl chains are
fixed in a syn arrangement (Figure 1).
of protein-linked DNA breaks (PLDB) in KB cells, but
again, no correlation could be found with either anti-
proliferative effect or activity against IN.
Methylation of the two m-hydroxyls in 7b yielded 7a
which is devoid of activity. Compounds 9a -d , open-
chain analogues of the cyclovalone compounds, dis-
played submicromolar activity when the bis-catechol
moieties were preserved (9c); otherwise activity de-
creased (9b) or was lost (9a ,d ). Indan derivatives 8a ,b,
both geometrically constrained in a syn disposition,
turned out active and moderately active, respectively.
Chalcones 10a ,d , featuring a single o-hydroxycinnamoyl
residue, exhibited good activity. Hydroxylation of 10a
at the position ortho to the carbonyl group, so as to give
10b, caused a significant drop of activity. Lack of the
catechol system in the chalcone series, as in 10c,
destroyed activity. Examination of anti-IN activity
values within the series of curcumin (11a -d ) and 1,3-
bis(cinnamoyl)benzene (12a ,b) derivatives revealed that
activity required the presence of o-hydroxyls (11b and
12a ), whereas activity was relatively independent of the
distance between the catechol units and the nature of
the central linker. Compound 13, the only one in our
set predicted noncoplanar by molecular modeling analy-
sis, resulted totally inactive.
Selectivity of IN In h ibition in En zym e Assa ys.
The fact that our compounds active against IN were
cytotoxic for cultured cell lines at concentrations sig-
nificantly higher than those required for enzyme inhibi-
tion raised the possibility that they acted nonspecifi-
cally. In fact, a number of IN inhibitors have been
reported to affect other viral enzymes in cell-free assays.
Among them are ATA and flavones, which show a fairly
similar structure-activity dependence for inhibition of
both RT and IN.8,24,25
When evaluated for their capability to inhibit the
recombinant HIV-1 RT and the cellular RNA poly-
merase II, none of the compounds were found active at
30 µM, the highest testable concentration (results not
shown), thus suggesting that those compounds which
turned out as potent IN inhibitors in enzyme assays
exhibit a fairly high degree of selectivity.
Str u ctu r e-Activity Rela tion sh ip s. The compu-
tational work was largely based on crystal structures
retrieved from the CSD17 of compounds strictly conge-
neric or coincident with those listed in Table 2. The
crystal structures analyzed are reported in Table 3
together with their CSD identification code and values
of relevant torsional angles. Model building and ma-
nipulation were realized using the SYBYL software
package.26 Computational protocols are detailed in the
Experimental Section.
Conformations of structures containing a cinnamoyl
moiety can be described through the torsional angles τ1
and τ2 defined in Figure 1. Owing to π-conjugation, τ1
and τ2 values close to 0° or 180° are associated with
energetically stable planar arrangements. The impor-
tance of an extended planar conformation was first
recognized by Burke et al.9,27 who proposed 5,6- and 6,7-
dihydroxynaphth-2-oyl nuclei (14a ,b) as conformation-
ally constrained mimetics of CAPE (2) and tyrphostin
(4) analogues differing in the value of τ1 (fixed to 0° in
14a and to 180° in 14b) (Chart 3). The 5,6-dihydroxy
substitution pattern is associated with highest inhibi-
tory activity when, in 14a ,b, R is different from OH
(esters and amides). These data suggest that a confor-
mation with a τ1 value of about 0° is favored for
interaction with IN.
The 3,4-dihydroxycinnamoyl moiety in Figure 1 be-
comes superimposable with that incorporated in the
structures of quercetin (1) only if τ2 takes a value of
about 180° (X-ray coordinates of quercetin yield τ1 and
τ2 values of 7.0° and -178.5°, respectively28). Contrary
to quercetin, cyclovalone analogues 7a -f display a
relative orientation of the carbonyl defined by τ2 values
falling around 0°. This was inferred from crystal
structures of congeners of 7a -c,f exhibiting a total or
nearly total coplanar arrangement about sp2-hybridized
atoms (kadmux, bzcptk20, and vabfin in Table 3). The
remarkable inhibitory activities of the cyclovalone ana-
logues bearing 3,4-dihydroxyls (7b-f) suggest that
flexible cinnamoyl analogues likewise bind to IN with
the carbonyl function oriented in syn with respect to
the vinylic double bond (τ2 value close to 0°). These data
justify speculation that the binding modes of quercetin
and cyclovalone analogues (7b -f) are different. Our
An tir etr ovir a l Effect. Since curcumin has been
reported15 active against the HIV-1 multiplication in
cell-based assays, we tested our inhibitors for their
capability to selectively inhibit the HIV-1-induced cy-
topathic effect (CPE) in de novo infected MT-4 cells.
However, under conditions where nevirapine and AZT,
which were used as reference compounds, showed EC50’s
of 0.01 and 0.25 µM, respectively, curcumin was unable
to prevent the virus-induced CPE. Similarly, none of
the compounds under investigation inhibited the HIV-1
multiplication at concentrations below those cytotoxic
for MT-4 cells (Table 2).
An tip r olifer a tive Activity. Due to the fact that
some curcumin analogues showed antiproliferative ac-
tivity against suspension cultures of lymphoid MT-4
cells at concentrations as low as 1.4 µM, we evaluated
whether they were also capable of inhibiting the pro-
liferation of cells (KB) derived from a solid human
tumor, growing in monolayer (Table 2). Title com-
pounds proved inhibitory against KB cells at concentra-
tions comparable to those active against MT-4 cells,
which ranged between 1.4 and 50 µM. In no case,
however, did the potency of antiproliferative activity
correlate with the presence of catechol groups, thus
suggesting that test compounds inhibited cell prolifera-
tion by mechanisms different from the emergence of
reactive quinone species.
Since some polyhydroxylated compounds active against
IN, namely, flavones, have been reported to be topo-
isomerase inhibitors,8 we investigated whether the
antiproliferative activity of our IN inhibitors correlated
with inhibition of cellular DNA binding proteins such
as topo I and topo II. To this end we used an intact cell
assay (see Experimental Section), rather than an in
vitro enzyme assay, to take into account the capability
of test compounds to penetrate the plasma membrane
and to accumulate into the nuclear compartment. Eto-
poside (VP-16) and camptothecin (CPT) were used as
reference drugs. As shown in Table 2, only compounds
7c and 8b were able to significantly increase the amount

Cinnamoyl Compounds as Inhibitors of HIV-1 IN
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21 3955
Ta ble 3. Geometries Retrieved from the CSD of Compound
Ch a r t 3. Naphthoic Acids (RdOH), Flavones, and
AG593
Coincident or Congeneric with Those Listed in Table 2a
pharmacophoric heteroatoms (three) and is the smallest
in size among the IN inhibitors structurally related to
CAPE. A comparison of the IC50 values of 10a ,b,d
suggests that the presence of the 2′-hydroxyl ortho to
the carbonyl (10b) is responsible for a 20-fold drop of
activity. Inspection of the crystal structures of 10b
(butein) reveals that the carbonyl oxygen is engaged in
a hydrogen bond with the 2′-hydroxy group. Intramo-
lecular hydrogen bond in 10b, making the carbonyl
oxygen less available to accept a hydrogen bond from
IN, would account for the above results.
1
According to H NMR measurements, curcumin (3)
and 11a -d exist exclusively as enolic tautomers (hep-
tatrienolone structure). Our findings are in agreement
with those reported by Sui et al. for dibenzoylcur-
cumin.29 In the solid state curcumin (3) and its bis-
desmethoxy derivative 11d are enols displaying a planar
conformation about at least one of the two cinnamoyl
residues with a τ2 value close to 0° (binmeq and ganjag,
respectively, in Table 3). Provided that curcumin and
11b inhibit IN in an extended coplanar arrangement
mimicking that of cyclovalone analogues 7b-f, there
should be available a complementary cleft in the enzyme
which is roughly flat in shape and of a size of ap-
proximately 20 × 7 Å (corresponding to the maximum
length and width of 11b). Our model contrasts with
that proposed by Mazumder et al.,11 who hypothesized
that curcumin binds as a diketo tautomer in a folded
conformation presenting the two phenyls stacked above
each other, thus surrogating the catechol pattern com-
monly believed to be vital for activity. Moreover,
Mazumder’s model does not apply to 9b which, like
curcumin, is moderately active and bears the same
substituents on the phenyl rings. In fact, conforma-
tional analyses, performed on 9b using the SYBYL/
SEARCH module, did not identify any folded confor-
mation for this compound (results not shown). Taken
together, these data suggest that intramolecular stack-
ing of phenyl rings is not obligatory for bis-monohy-
droxylated inhibitors such as curcumin. Lack of free
catechol hydroxyls, in curcumin as well as in compounds
a
The CSD reference code is given for each structure. τ1 and τ2
values are defined in Figure 1 (values in parentheses refer to the
equivalent moiety oriented rightward). Cocrystallized species are
not displayed for sake of clarity. Some structures lack one or more
hydrogens.
hypothesis is supported by SARs reported for flavone
derivatives.8 Specifically, 3,4-dihydroxyflavone (15a )
was found devoid of activity although it matches the
general formula of Figure 1 with τ2 fixed to about 180°.
Additionally, baicalein (15b) proved to be active al-
though it lacks the pharmacophoric 3,4-dihydroxycin-
namoyl moiety.
Indan derivatives 8a ,b, inhibiting IN with different
potency, are constrained in geometries very similar to
those of cyclovalone analogues 7a -f (see mbyino and
bzyind in Table 3). It is not clear why the presence in
8b of an endocyclic methylene in place of a carbonyl as
in 8a causes a 30-fold drop of activity. In fact, a strongly
detrimental effect of the methylene is not entirely
consistent with 7c and 9c being equipotent. Evidently,
subtle steric and electrostatic effects come into play
whose nature requires additional investigation to be
elucidated.
Compounds 9a -d and 10a -d attain coplanar con-
formations judging from the X-ray structures of pevkik,
legbuu, and butein in Table 3. It is worth noting the
fairly good activity exhibited by 10d , which features only
a carbonyl and two catechol hydroxyls. To our knowl-
edge this compound possesses the lowest number of

3956 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21
Artico et al.
9b and 12b, can be nevertheless associated with moder-
ate inhibitory activity (IC50 values ranging from 30 to
37 µM) thanks to other chemical functions favorably
interacting with a flat “accessory” area of IN adjacent
to the “pharmacophoric" binding site. Comparisons
between curcumin and 11c as well as between 9b and
9d suggest that one of the two m-methoxy groups in
curcumin and 9b might be involved in the above-
hypothesized attractive interactions. The inactivity of
7a indicates that a m-methoxy substituent must have
a proper 3D orientation in order to exert a positive effect
on potency (7a and 9b share in fact the same bis-
cinnamoyl substructure). Probably, the cyclohexanone
skeleton in 7a is not ideal for ensuring a full coplanarity
of the two cinnamoyl moieties, whereas the less steri-
cally encumbered backbone of 9b can more easily allow
both arylidene moieties to lie on the same plane. This
view is suggested by inspection of the crystal structures
of kadmux and pevkik as models of 7a and 9b, respec-
tively.
F igu r e 3. Molecular models of the inactive compound 13
(black) and the active CAPE (gray) overlayed about the
common cinnamoyl residue. Conformations of 13 and CAPE
are 3.2 and 0.0 kcal/mol higher in energy, respectively, than
their corresponding global minimum conformers (Tripos force
field). The terminal phenyl rings of the two compounds are
not coincident in space.
strain energy (difference with respect to the global
minimum conformer). Attainment of a fully staggered
conformation in CAPE (2), which also contains a phen-
ethyl group, was penalized by only 1.7 kcal/mol of strain
energy under the same computational settings. Twist-
ing the terminal phenyl ring of about 90° in both 13
and CAPE lowered the internal energies to 3.2 and 0.0
kcal/mol, respectively. An overlay of these two “semi-
planar” conformations about the common cinnamoyl
frame shows two distinct locations for the terminal
phenyl rings (Figure 3). Inactivity of 13 may in
principle depend on an exceedingly high energetic cost
to attain coplanarity. Alternatively, the phenyl ring in
the semiplanar conformation of this compound may be
sterically repelled upon binding.
Full coplanarity of the inhibitor structure is not an
absolute requirement for high activity. In fact, compu-
tational studies performed in our laboratories estab-
lished that certain IN inhibitors, such as AG593 (16),12
cannot assume an entirely coplanar conformation (re-
sults not shown). Potency of tyrphostins has been
reported to be poorly sensitive to the nature and the
length of the linker between amide nitrogens,12 sug-
gesting that the linker is not part of the basic pharma-
cophore. An intramolecular loop may, in these cases,
orient the terminal moieties into distinct binding sites,
one of the two being complementary to the basic
pharmacophore. The latter hypothesis was formulated
earlier by Zhao et al.27 who proposed the existence of
two putative binding sites within IN, one being of “high
affinity”.
Con clu sion s
Among the numerous polyhydroxylated compounds
that have been shown to inhibit the HIV-1 IN in enzyme
assays, very few of them are effective anti-HIV-1 agents
when tested in cell-based assays. We therefore devel-
oped new IN inhibitors containing a cinnamoyl moiety
with the CdC and CdO double bonds attaining a syn
disposition (τ2 about 0° in Figure 1). It was hoped that
a backbone “frozen” in such a 3D arrangement, by
geometrical or conformational constraints, could reduce
cytotoxicity while retaining activity on infected cells.
In cell-free assays, our compounds inhibited neither
the HIV-1 RT nor the cellular RNA polymerase II, thus
suggesting that they specifically target the HIV-1 IN.
Nevertheless, none of them were effective in preventing
the HIV-1 multiplication in acutely infected cells. While
this manuscript was in preparation, new data on
curcumin analogues have been published31 which are
consistent with the above observations. Moreover, we
have demonstrated that the antiproliferative potency of
cinnamoyl-based structures, which does not correlate
with the capability of inducing single-strand DNA
breaks, occurs at concentrations up to 100 times higher
than those effective against the viral enzyme. This
suggests that in cell-based assays, cytotoxicity may not
mask antiviral activity. Therefore, unless inhibition of
cell proliferation is carried by title compounds through
interaction with cell receptors on the plasma membrane,
the possibility cannot be ruled out that the in vitro IN
reaction fails to mimic the integration process occurring
in the infected cells.
The 1,3-bis(cinnamoyl)benzene derivatives 12a ,b can
access low-energy conformations entirely coplanar in
which the C1-C(dO) and C3-C(dO) bonds of the 1,3-
dicarbonylbenzene moiety are eclipsed with the benzene
C2 atom. A substructure search in the CSD, using the
central linker CO-C6H4-CO as a query, identified 13
out of 53 hits characterized by values of the torsional
angles τ3 (C2-C1-CdO) and τ4 (C2-C3-CdO) within
(10°. The remaining 40 conformations were loosely
clustered around the following pairs of τ3 and τ4
values: 0°/180°, 180°/0°, and 180°/180°. Using the
SYBYL/SEARCH routine, we detected a low energy
conformation for 12b very similar to that proposed by
Mazumder et al.11 for curcumin (3). To sum up, flex-
ibility of compounds 12a ,b does not allow a straight-
forward prediction of their bioactive conformations. We
do not know whether these compounds, the largest in
size of our series, fit partially or entirely into the flat
“accessory” binding site.
The poor activity of the partially saturated ketone 13
(IC50 value can be estimated slightly above 100 µM) is
likely related to disruption of coplanarity. Molecular
mechanics calculations, performed on 13, based on the
Tripos force field,30 revealed that the most stable
conformations of this compound are folded so as to
present the whole phenethyl chain out of the plane of
the cinnamoyl unity. A transoid fully coplanar confor-
mation of 13 was achieved at the cost of 4.9 kcal/mol of
Examination of SARs developed from our compounds,
interpreted in the light of crystallographic data and
molecular modeling analyses, led to novel findings

Cinnamoyl Compounds as Inhibitors of HIV-1 IN
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21 3957
mL), and dried. Evaporation of solvent furnished crude
protected product which was suspended in methanol (40 mL)
and treated with a catalytic amount of p-toluenesulfonic acid.
After stirring at room temperature for 15 h the solvent was
removed; the residue was treated with water and neutralized
with a saturated solution of sodium hydrogen carbonate. The
suspension was extracted with ethyl acetate (3 × 50 mL), and
the organic extracts were collected, washed with brine (3 ×
100 mL), and dried. Evaporation of the solvent gave crude
product which was chromatographed on silica gel column
(chloroform/methanol, 10:1, as eluent) to obtain pure 7b (1.7
g, 99% yield).
concerning the molecular determinants of IN inhibitory
activity within the class of cinnamoyl-containing struc-
tures. The importance of at least one catechol unit is
suggestive of a binding mode similar to that of CAPE
analogues. All these compounds are hypothesized to act
in a conformation with the carbonyl group oriented in
syn with respect to the vinylic double bond (τ1 and τ2
values of about 0° in Figure 1). This implies that the
binding mode of the above inhibitors is different from
that of flavone derivatives, such as quercetin, character-
ized by an anti disposition of the CdC-CdO moiety
incorporated in the pyranone nucleus (τ2 value of about
180°). Contrary to that hypothesized by Mazumder et
al.,11 we propose that bis-monohydroxylated inhibitors
such as curcumin do not attain a folded conformation
with the two phenyl rings stacking above each other.
Lack of one of the two catechol hydroxyls, both crucial
for high potency, may be compensated by the presence
of other substructures which favorably interact with a
flat “accessory” area of IN adjacent to the “pharmaco-
phoric site”. Lack of coplanarity in the inhibitor struc-
ture may abolish activity only if the out-of-plane moiety
is oriented toward the steric boundaries of the binding
site or a coplanar arrangement is associated with an
exceedingly high strain energy.
This procedure was used for the synthesis of compounds 7c-
e, 8a ,b, 9c,d , 10a -d , 12a ,b, and 13 starting from 3,4-bis-
(tetrahydropyran-2-yloxy)benzaldehyde (for reaction times at
50 °C and chromatographic systems, see Table 1) and for
compounds 9a ,b starting from veratraldehyde and vanillic
aldehyde, respectively; compound 7a was obtained as reported
previously.33 Spectroscopic data for derivatives 7d ,e, 8b, 9c,
12a ,b, and 13 are reported below.
1
7d : IR cm-1 3200 (OH) and 1580 (CO); H NMR (DMSO-
d6) δ 4.86 (s, 4H, CH2), 6.79-6.86 (m, 6H, benzene H), 7.49 (s,
2H, dCH-), 9.10 (br s, 4H, OH).
7e: IR cm-1 3440, 3200 (OH), and 1570 (CO); 1H NMR
(DMSO-d6) δ 3.95 (s, 4H, CH2), 6.79-6.95 (m, 6H, benzene H),
7.45 (s, 2H, dCH-), 9.24 and 9.57 (2br s, 4H, OH).
8a : IR cm-1 3440, 3200 (OH), and 1660 (CO); 1H NMR
(DMSO-d6) δ 6.93 (d, J o ) 8.2 Hz, 1H, benzene C5-H), 7.67-
7.93 (m, 6H, benzene C2-H, C6-H, and indan H), 8.38 (s,
1H, dCH-), 9.64 and 10.56 (2br s, 2H, OH).
Exp er im en ta l Section
9c: IR cm-1 3460, 3250 (OH), and 1580 (CO); 1H NMR
(DMSO-d6) δ 6.78 (d, J o ) 8.1 Hz, 2H, benzene C5-H), 6.99
(d, J t ) 15.9 Hz, 2H, C2-H and C4-H), 7.07 (d, J o ) 8.1 Hz,
2H, benzene C6-H), 7.14 (s, 2H, benzene C2-H), 7.56 (d, J t
) 15.9 Hz, 2H, C1-H and C5-H), 9.15 (br s, 4H, OH).
Ch em istr y. Melting points were determined on a Bu¨chi
530 melting point apparatus and are uncorrected. Infrared
(IR) spectra (Nujol mull) were recorded on a Perkin-Elmer 297
instrument. 1H NMR spectra were recorded at 200 MHz on a
Bruker AC 200 spectrometer using tetramethylsilane (Me4Si)
as internal reference. All compounds were routinely checked
1
12a : IR cm-1 3300 (OH) and 1640 (CO); H NMR (DMSO-
d6) δ 6.83 (d, J o1 ) 8.0 Hz, 2H, benzene C5-H), 7.23 (d, J o1
)
1
8.0 Hz, 2H, benzene C6-H), 7.31 (s, 2H, benzene C2-H), 7.68
(m, 5H, C2-H, C3-H, and benzoyl C5-H), 8.35 (d, J o2 ) 7.7
Hz, 2H, benzoyl C4-H and C6-H), 8.65 (s, 1H, benzoyl C2-
H), 9.19 and 9.78 (2br s, 4H, OH).
by TLC and H NMR. TLC was performed by using aluminum-
baked silica gel plates (Merck DC-Alufolien Kieselgel 60 F254).
Developed plates were visualized by UV light. Solvents were
reagent grade and, when necessary, were purified and dried
by standard methods. Concentration of solutions after reac-
tions and extractions involved the use of a rotary evaporator
operating at a reduced pressure of approximately 20 Torr.
Organic solutions were dried over anhydrous sodium sulfate.
Analytical results agreed to within (0.40% of the theoretical
values.
1
12b: IR cm-1 3300 (OH) and 1635 (CO); H NMR (DMSO-
d6) δ 6.86 (d, J o1 ) 8.5 Hz, 4H, benzene C3-H and C5-H),
7.71-7.82 (m, 9H, C2-H, C3-H, benzene C2-H, C6-H, and
benzoyl C5-H), 8.38 (d, J o2 ) 7.7 Hz, 2H, benzoyl C4-H and
C6-H), 8.68 (s, 1H, benzoyl C2-H), 10.20 (br s, 2H, OH).
13: IR cm-1 3480, 3160 (OH), and 1630 (CO); 1H NMR
(DMSO-d6) δ 2.86-2.98 (m, 4H, CH2), 6.58 (d, J t ) 16.1 Hz,
1H, C2-H), 6.77 (d, J o ) 8.0 Hz, 1H, hydroxybenzene C5-H),
7.01 (d, J o ) 8.0 Hz, 1H, hydroxybenzene C6-H), 7.07 (s, 1H,
hydroxybenzene C2-H), 7.15-7.30 (m, 5H, benzene H), 7.47
(d, J t ) 16.1 Hz, 1H, C3-H), 9.15 and 9.60 (2br s, 2H, OH).
3,5-Bis(3,4-d ih yd r oxyb en zylid en e)-4-p ip er id on e Hy-
d r och lor id e (7f). A solution of 4-piperidone hydrate hydro-
chloride (1.0 g, 6.5 mmol) in glacial acetic acid (40 mL) was
saturated with anhydrous hydrogen chloride and then treated
with a suspension of 3,4-dihydroxybenzaldehyde (2.7 g, 19.5
mmol) in the same solvent (50 mL). After stirring at room
temperature for 48 h the precipitate which formed was filtered
and washed with water, ethanol, and light petroleum ether,
in turn, to obtain pure 7f (2.2 g, 92% yield).
Syn th eses. Specific examples presented below illustrate
general synthetic procedures.
3,4-Bis(tetr a h yd r op yr a n -2-yloxy)ben za ld eh yd e. A so-
lution of 3,4-dihydro-R-pyran (2.02 mL, 1.87 g, 22 mmol) in
dichloromethane (12 mL) was added dropwise onto a well-
stirred suspension of 3,4-dihydroxybenzaldehyde (1.0 g, 7.2
mmol) and pyridinium p-toluenesulfonate (0.04 g, 0.16 mmol)
in the same solvent (34 mL). The mixture was stirred at room
temperature for 1.5 h, then was washed with brine (3 × 50
mL), and dried. Evaporation of the solvent gave crude product
which was purified by column chromatography on silica gel
(cyclohexane/ethyl acetate, 9:2, as eluent) to obtain 3,4-bis-
(tetrahydropyran-2-yloxy)benzaldehyde as a yellowish oil (1.1
g, 50% yield).
By this procedure were prepared the following compounds:
4-(tetrahydropyran-2-yloxy)benzaldehyde (100% yield, oil);19
2-hydroxy-4-(tetrahydropyran-2-yloxy)acetophenone (100% yield,
mp 67-69 °C);19 4-(tetrahydropyran-2-yloxy)acetophenone
(100% yield, mp 85-87 °C).32
2,6-Bis(3,4-d ih yd r oxyben zylid en e)cycloh exa n on e (7b).
A mixture of 3,4-bis(tetrahydropyran-2-yloxy)benzaldehyde
(3.1 g, 10.2 mmol), cyclohexanone (0.5 g, 5.1 mmol), and barium
hydroxide octahydrate (4.8 g, 15.2 mmol) in methanol (100 mL)
was stirred at 50 °C for 7 h. After cooling the solvent was
evaporated, and the residue was treated with 1 M hydrochloric
acid (100 mL) and extracted with ethyl acetate (3 × 50 mL).
The organic layer was separated, washed with brine (3 × 100
1,7-Bis(4-h yd r oxyp h e n yl)-1,6-h e p t a d ie n e -3,5-d ion e
(11c). A solution of 4-hydroxybenzaldehyde (2.4 g, 19.8 mmol)
and acetylacetone (1.0 g, 9.9 mmol) in N,N′-dimethylforma-
mide (2 mL) was treated with boric acid (2.0 g, 32.0 mmol),
and the mixture was heated at 100 °C for 5 min. After this
time a solution of 1,2,3,4-tetrahydroquinoline (2.0 mL, 2.1 g,
15.8 mmol) and acetic acid (0.6 mL) in N,N′-dimethylforma-
mide (2 mL) was added. The resulting mixture was heated
at 100 °C for 4 h, then cooled, diluted with 20% acetic acid
(100 mL), and stirred at room temperature for 1 h. The
precipitate which formed was filtered, washed with ethyl
acetate and then with light petroleum ether, and finally
chromatographed to afford pure 11c (790 mg, 26% yield).

3958 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21
Artico et al.
By this procedure were prepared compounds 11a ,b,d (for
chromatographic systems, see Table 1); oily derivatives 11a ,d
were extracted with ethyl acetate before chromatographic
purification. Spectroscopic data of these derivatives together
with IR and 1H NMR spectra of curcumin (3) are reported
below. The presence of a singlet (1H) for C4-H in the 1H NMR
spectra of curcumin (3), 11b,c, and 11a ,d (δ ) 5.94, 6.08, 6.04,
5.82, and 5.86 ppm, respectively), accounts for the hep-
tatrienolone structure.
obtained by IPTG induction of the E. coli strain BL21(DE3)
containing the pINSD‚His vector. Protein purification was
carried out following the Novagen procedure, except for the
presence of 5 mM CHAPS in binding buffer (5 mM imidazole,
0.5 M NaCl, 20 mM Tris-HCl, pH 7.9), washing buffer (60 mM
imidazole, 0.5 M NaCl, 20 mM Tris-HCl, pH 7.9), and elute
buffer (1 M imidazole, 0.5 M NaCl, 50 mM Tris-HCl, pH 7.9,
5 mM â-mercaptoethanol).
The following oligos representing the terminal 21 nucle-
otides of the HIV-1 U5 LTR were used in this study: A, 5′-
GTG TGG AAA ATC TCT AGC AGT-3′ (plus strand); B, 5′-
ACT GTC AGA GAT TTT CCA CAC-3′ (minus strand); C, 5′-
GTG TGG AAA ATC TCT AGC A-3′ (plus strand two
nucleotides shorter than A). For standard 3′-processing as-
says, C was annealed with B in 0.1 M NaCl by heating at 80
°C and slowly cooling to room temperature overnight. This
double-stranded substrate was labeled by introducing at the
3′-end of C the two missing nucleotides using [R-32P]dGTP,
cold dTTP, and Klenow polymerase. Unincorporated [R-32P]-
dGTP was separated from the duplex substrate by two
consecutive runs through G-25 Sephadex quick spin columns.
Alternatively, C was labeled at the 5′-end using [γ-32P]ATP
and T4 polynucleotide kinase, purified through G-25 Sephadex
quick spin column, and then annealed with unlabeled B in
0.1 M NaCl as stated above. ATA, used as a standard in our
IN assays, was purchased from Aldrich (12,326.9; batch 68853-
104).
1
3: IR cm-1 3300 (OH) and 1620 (CO); H NMR (CD3OD) δ
3.90 (s, 6H, OCH3), 5.94 (s, 1H, C4-H), 6.62 (d, J t ) 15.8 Hz,
2H, C2-H and C6-H), 6.81 (d, J o ) 8.0 Hz, 2H, benzene C5-
H), 7.09 (d, J o ) 8.0 Hz, 2H, benzene C6-H), 7.19 (s, 2H,
benzene C2-H), 7.56 (d, J t ) 15.8 Hz, 2H, C1-H and C7-H).
1
11a : IR cm-1 1610 (CO); H NMR (CDCl3) δ 3.92 and 3.94
(2s, 12H, OCH3), 5.82 (s, 1H, C4-H), 6.49 (d, J t ) 15.6 Hz,
2H, C2-H and C6-H), 6.87 (d, J o ) 8.0 Hz, 2H, benzene C5-
H), 7.09 (d, J m ) 1.9 Hz, 2H, benzene C2-H), 7.16 (d, J o ) 8.0
Hz, J m ) 1.9 Hz, 2H, benzene C6-H), 7.60 (d, J t ) 15.6 Hz,
2H, C1-H and C7-H), 15.80 (br s, 1H, OH).
1
11b: IR cm-1 3300 (OH) and 1590 (CO); H NMR (DMSO-
d6) δ 6.08 (s, 1H, C4-H), 6.58 (d, J t ) 15.8 Hz, 2H, C2-H and
C6-H), 6.78 (d, J o ) 8.1 Hz, 2H, benzene C5-H), 7.02 (d, J o
) 8.1 Hz, 2H, benzene C6-H), 7.08 (s, 2H, benzene C2-H),
7.46 (d, J t ) 15.8 Hz, 2H, C1-H and C7-H), 8.71, 9.22, and
9.69 (3br s, 5H, OH).
1
11c: IR cm-1 3300 (OH) and 1620 (CO); H NMR (DMSO-
3′-P r ocessin g Assa y. Standard reaction conditions were
40 mM NaCl, 10 mM MnCl2, 25 mM Tris-HCl, pH 7.5, 1 mM
DTT, 2% glycerol, 1 nM duplex B:C labeled at the 3′-end, 5
nM IN (considered as monomer). Incubation was carried out
at 37 °C for 30 min in a volume of 50 µL. The reaction was
stopped by adding an equal volume of 0.5 M Na2HPO4. Each
sample (90 µL) was added to a well of a Multiscreen 96-well
filtration plate preloaded with 100 µL of DEAE Sephacel slurry
(equilibrated in 0.5 M Na2HPO4) which had been vacuum-
dried. Vacuum was applied after incubation at room temper-
ature for 10 min. The uncleaved, labeled substrate was
retained in the resin, whereas labeled dinucleotide (GT)
products were collected at the bottom of 96-well filtration
plates and counted for radioactivity. When reactions were
analyzed by denaturing PAGE, 6 µL of each sample was taken
before stopping the reaction and added to 3 µL of sample buffer
(96% formamide, 20 mM EDTA, 0.08% bromophenol blue,
0.25% xylene cyanol). Samples were heated at 100 °C for 3
min, layered onto a denaturing 15% polyacrylamide gel (7 M
urea, 0.09 M Tris borate, pH 8.3, EDTA 2 mM, 15% acryla-
mide), and run for 1 h at 65 W. Reaction products were
visualized by autoradiography and quantified by densitometric
analysis.
Biologica l Assa ys. Compounds were solubilized in DMSO
at 200 mM and then further diluted. MT-4 cells were grown
at 37 °C in a 5% CO2 atmosphere in RPMI 1640 medium
supplemented with 10% fetal calf serum (FCS), 100 IU/mL
penicillin G, and 100 µg/mL streptomycin. Cell cultures were
checked periodically for the absence of mycoplasma contami-
nation with a MycoTect Kit (Gibco). HIV-1 stock solutions had
titers of 4.5 × 106 50% cell culture infectious dose (CCID50)/
mL. Cytotoxicity evaluation was based on the viability of
mock-infected cells, as monitored by the MTT method. Activity
of the compounds against the HIV-1 multiplication in acutely
infected cells was based on inhibition of virus-induced cyto-
pathic effect in MT-4 cells and was determined by the 3-(4,5-
dimethylthiazol-1-yl)-2,5-diphenyltetrazolium bromide (MTT)
method.37 Anti-RNA polymerase II assays were performed as
previously described.38 Purified rRT was assayed for its RNA-
dependent polymerase-associated activity in a 50-µL volume
containing 50 mM Tris-HCl (pH 7.8), 80 mM KCl, 6 mM MgCl2,
1 mM DTT, 0.1 mg mL-1 BSA, 0.5 OD260 unit mL-1 template:
primer [poly(rC)‚oligo(dG)12-18], and 10 µM [3H]dGTP (1 Ci
mmol-1). After incubation for 30 min at 37 °C, the samples
were spotted on glass fiber filters (Whatman GF/A), and the
acid-insoluble radioactivity was determined.
d6) δ 6.04 (s, 1H, C4-H), 6.71 (d, J t ) 15.8 Hz, 2H, C2-H and
C6-H), 6.82 (d, J o ) 8.5 Hz, 4H, benzene C3-H and C5-H),
7.51-7.60 (m, 6H, C1-H, C7-H, benzene C2-H, and C6-
H), 10.10 (br s, 3H, OH).
11d : IR cm-1 1610 (CO); 1H NMR (CDCl3) δ 5.85 (s, 1H,
C4-H), 6.63 (d, J t ) 15.6 Hz, 2H, C2-H and C6-H), 7.38 (m,
6H, benzene C3-H, C4-H, and C5-H), 7.57 (m, 4H, benzene
C2-H and C6-H), 7.67 (d, J t ) 15.6 Hz, 2H, C1-H and C7-
H), 15.90 (br s, 1H, OH).
Molecu la r Mod elin g. The October 1996 and April 1997
releases (3D graphics 5.12 and 5.13 versions for UNIX
platforms) of the CSD17 were searched through substructure
queries using the CSD/QUEST graphical routine. Molecular
modeling was performed with use of the software package
SYBYL26 running on a Silicon Graphics Indigo XS24 worksta-
tion. Molecular models of compounds 13 and CAPE (3 in
Chart 1), whose coordinates were not available from the CSD,
were assembled using standard bond lengths and bond angles
of the SYBYL fragment library. Conformational energies were
calculated according to the molecular mechanics Tripos force
field30 including electrostatics. Partial atomic charges were
assigned using the Gasteiger-Hu¨ckel method.34,35 Geometry
optimizations were performed with the SYBYL/MAXIMIN2
minimizer by applying the Broyden-Fletcher-Goldfarb-
Shanno algorithm36 and setting an energy difference of 0.001
kcal/mol as convergence criterion. Structures of compound 13
and CAPE were superimposed by minimizing the root-mean-
square distance between the non-hydrogen atoms of the
cinnamoyl moiety using the SYBYL/FIT command. Confor-
mational analyses on 9b, 12b, 13, 16, and CAPE were carried
out using the SYBYL/SEARCH module. Properly selected
torsional angles were scanned by 20° increments throughout
0-340°, or 0-160° for symmetrically substituted phenyl rings.
A van der Waals scaling factor of 0.75 was applied to “soften”
steric contacts. Such scaling was aimed at minimizing the
typical limitation of systematic rigid search consisting of lack
of relaxation for each conformation. The number of output
conformations to be examined was reduced by setting the
“energy window” (energy difference between the generated
conformation and the current minimum) to 10.0 kcal/mol.
Results from conformational analyses were evaluated by
means of the SYBYL/MOLECULAR SPREADSHEET routine.
Differences in energies between conformations, as in the
modeling of 13 and CAPE, were computed on geometry-
optimized structures.
In tegr a se a n d DNA Su bstr a tes. Expression of the HIV-1
IN protein with an amino-terminal polyhistidine tag was
Qu a n tita tion of P r otein -Lin k ed DNA Br ea k s. A modi-

Cinnamoyl Compounds as Inhibitors of HIV-1 IN
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21 3959
fied in vivo K-SDS coprecipitation assay39 was used for
quantification of protein-linked DNA breaks (PLDB) levels.
Briefly, KB cells were seeded at a density of 2.5 × 105/mL and
labeled with [14C]thymidine (0.75 µCi/mL) for 48 h. Monolay-
ers were washed twice, and after trypsinization, cells were
resuspended in aliquots of 3 × 105/mL and were incubated for
1 h at 37 °C. Then, duplicate samples were treated with test
drugs at concentrations 10 times higher than respective CC50
values and further incubated for 1 h. Cells were collected by
centrifugation at 2000g for 10 min and resuspended in 1 mL
of warm lysis buffer (1.5% SDS, 5 mM EDTA, 0.4 mg/mL
salmon sperm DNA), and the viscous cell lysates were sheared
by passage through a 22-gauge needle five times. After 10-
min incubation at 65 °C, KCl was added to a final concentra-
tion of 100 mM; samples were chilled in an ice bath for 10
min and centrifuged at 3500g for 10 min at 4 °C. Pellets were
resuspended in 1 mL of warm washing buffer (10 mM Tris-
HCl, pH 8.0, 100 mM KCl, 1 mM EDTA, 0.1 mg/mL salmon
sperm DNA) and again incubated for 10 min at 65 °C, chilled
in an ice bath, and centrifuged as above. After a second
washing step, pellets were solubilized at 65 °C in 400 µL of
water and counted for radioactivity.
(13) Eich, E.; Pertz, H.; Kaloga, M.; Schultz, J .; Fesen, M. R.;
Mazumder, A.; Pommier, Y. (-)-Arctigenin as a Lead Structure
for Inhibitors of Human Immunodeficiency Virus Type-1 Inte-
grase. J . Med. Chem. 1996, 39, 86-95.
(14) Robinson, W. E., J r.; Reinecke, M. G.; Abdel-Malek, S.; J ia, Q.;
Chow, S. A. Inhibitors of HIV-1 Replication that Inhibit HIV
Integrase. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 6326-6331.
(15) Li, C. J .; Zhang, L. J .; Dezube, B. J .; Crumpacker, C. S.; Pardee,
A. B. Three Inhibitors of Type 1 Human Immunodeficiency Virus
Long Terminal Repeat-Directed Gene Expression and Virus
Replication. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 1839-1842.
(16) Dyda, F.; Hickman, A. B.; J enkins, T. M.; Engelman, A.; Craigie,
R.; Davies, D. R. Crystal Structure of the Catalytic Domain of
HIV-1 Integrase: Similarity to Other Polynucleotidyl Trans-
ferases. Science 1994, 266, 1981-1986.
(17) Bernstein, F. C.; Koetzle, T. F.; Williams, G. J . B.; Meyer, E. F.,
J r.; Brice, M. D.; Rodgers, J . R.; Kennard, O.; Shimanouchi, T.;
Tasumi, T. The Protein Data Bank: A Computer Based Archival
File for Macromolecular Structures. J . Mol. Biol. 1977, 112,
535-542.
(18) Babu, K. V. D.; Rajasekharan, K. N. Simplified Condition for
Synthesis of Curcumin I and Other Curcuminoids. OPPI Briefs
1994, 26, 674-677.
(19) Severi, F.; Costantino, L.; Benvenuti, S.; Vampa, G.; Mucci, A.
Synthesis and Description of Chalcone-like Compounds, Inhibi-
tors of Aldose Reductase. Med. Chem. Res. 1996, 6, 128-136.
(20) Oosawa, T.; Ochi, H. J pn. Kokai Tokkyo Koho J P 07,206,763
[95,206,753]; Chem. Abstr. 1996, 124, 8511s.
(21) Zhao, H.; Neamati, N.; Mazumder, A.; Sunder, S.; Pommier, Y.;
Burke, T. R., J r. Arylamide Inhibitors of HIV-1 Integrase. J .
Med. Chem. 1997, 40, 1186-1194.
(22) Cushman, M.; Sherman, P. Inhibition of HIV-1 Integration
Protein by Aurintricarboxylic Acid Monomers, Monomer Ana-
logues and Polymer Fractions. Biochem. Biophys. Res. Commun.
1992, 185, 85-90.
(23) Engelman, A.; Mizouchi, K.; Craigle, R. HIV-1 DNA Integration
Mechanism of Viral DNA Cleavage and DNA Strand Transfer.
Cell 1991, 67, 1211-1221.
Ack n ow led gm en t. The authors thank Italian Min-
istero della Sanita`-Istituto Superiore di Sanita`-IX Pro-
getto AIDS 1996 (Grants 9403-07 and 9403-59) for
financial aid. Thanks are also due to Foundation “Isti-
tuto Pasteur - Fondazione Cenci Bolognetti”, University
of Rome “La Sapienza” (Italy), Assessorato Igiene
Sanita`, Regione Sardegna, and MURST (Rome) for
partial support.
(24) Ono, K.; Nakane, H.; Fukushima, M.; Chermann, J .-C.; Barre´-
Sinoussi, F. Differential Inhibitory Effects of Various Flavonoids
on the Activities of Reverse Transcriptase and Cellular DNA and
RNA Polymerases. Eur. J . Biochem. 1990, 190, 469-476.
(25) Chu, S. C.; Hsiech, Y. S.; Lin, J . Y. Inhibitory Effects of
Flavonoids on Moloney Murine Leukemia Virus Reverse Tran-
scriptase Activity. J . Nat. Prod. 1992, 55, 179-183.
(26) SYBYL Molecular Modeling System (version 6.2), Tripos Assoc.,
St. Louis, MO.
(27) Zhao, H.; Neamati, N.; Hong, H.; Mazumder, A.; Wang, S.;
Sunder, S.; Milne, G. W. A.; Pommier, Y.; Burke, T. R., J r.
Coumarin-Based Inhibitors of HIV-1 Integrase. J . Med. Chem.
1997, 40, 242-249.
(28) Rossi, M.; Rickles, L. F.; Halpin, W. A. The Crystal and
Molecular Structure of Quercetin: a Biologically Active and
Naturally Occurring Flavonoid. Bioorg. Chem. 1986, 14, 55-
69.
(29) Sui, Z.; Salto, R.; Li, J .; Craik, C.; Ortiz de Montellano, P. R.
Inhibition of the HIV-1 and HIV-2 Proteases by Curcumin and
Curcumin Boron Complexes. Bioorg. Med. Chem. 1993, 1, 415-
422.
Refer en ces
(1) Artico, M. Nonnucleoside anti-HIV-1 Reverse Transcriptase
Inhibitors (NNRTIs): A Chemical Survey from Lead Compounds
to Selected Drugs for Clinical Trials. Farmaco 1996, 51, 305-
331.
(2) Appelt, K.; Clare, M.; Pettit, S. C.; Michael, S. F.; Swanstrom,
R.; Moore, M. C.; Dreyer, G. B.; Erickson, J . W. Therapeutic
Approaches to HIV. Perspect. Drug Discovery Des. 1993, 1, 23-
128.
(3) De Clercq, E. Toward Improved anti-HIV Chemotherapy: Thera-
peutic Strategies for Intervention with HIV Infections. J . Med.
Chem. 1995, 38, 2491-2517.
(4) La Femina, R. L.; Schneider, C. L.; Robbins, H. L.; Callahan, P.
L.; LeGrow, K.; Roth, E.; Schleif, W. A.; Emini, E. A. Require-
ment of Active Human Immunodeficiency Virus Type 1 Integrase
Enzyme for Productive Infection of Human T-Lymphoid Cells.
J . Virol. 1992, 66, 7414-7419.
(5) Sakai, H.; Kawamura, M.; Sakuragi, J .; Sakuragi, S.; Shibata,
R.; Ishimoto, A.; Ono, N.; Ueda, S.; Adachi, A. Integration is
Essential for Efficient Gene Expression of Human Immunode-
ficiency Virus type 1. J . Virol. 1993, 67, 1169-1174.
(6) Brown, P. O.; Bowerman, B.; Varmus, H. F.; Bishop, J . M.
Retroviral Integration: Structure of the Initial Covalent Product
and its Precursor, and a Role for the Viral Integrase Protein.
Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 2525-2529.
(7) Bushman, F. D.; Fujiwara, T.; Craigie, R. Retroviral DNA
Integration Directed by HIV Integration Protein in Vitro. Science
1990, 249, 1555-1558.
(8) Fesen, M. K.; Pommier, Y.; Leteurtre, F.; Hiroguchi, S.; Young,
J .; Kohn, K. W. Inhibition of HIV-1 Integrase by Flavones,
Caffeic Acid Phenethyl Ester (CAPE) and Related Compounds.
Biochem. Pharmacol. 1994, 48, 595-608.
(9) Burke, T. R., J r.; Fesen, M. R.; Mazumder, A.; Wang, Y.;
Carothers, A. M.; Grunberger, D.; Driscoll, J .; Khon, K.; Pom-
mier, Y. Hydroxylated Aromatic Inhibitors of HIV-1 Integrase.
J . Med. Chem. 1995, 38, 4171-4178.
(30) Clark, M.; Cramer, R. D., III; Van Opdenbosch, N. Validation
of the General Purpose Tripos 5.2 Force Field. J . Comput. Chem.
1989, 10, 982-1012.
(31) Mazumder, A.; Neamati, N.; Sunder, S.; Schulz, J .; Pertz, H.;
Eich, E.; Pommier, Y. Curcumin analogues with altered poten-
cies against HIV-1 integrase as probes for biochemical mecha-
nisms of drug action. J . Med. Chem. 1997, 40, 3057-3063.
(32) Burckhalter, J . H.; Dixon, W. D.; Black, M.; Westland, R. D.;
Werbel, L. M.; DeWald, H. A.; Dice, J . R.; Rodney, G.; Kaump,
D. H. 2-(2-Pyridyl)-1,2-diaryl Alkanols as Hypocholesteremic
Agents. J . Med. Chem. 1967, 10, 565-575.
(33) Samdahl, B. Diveratrylidene and Divanillydene Compounds of
Cyclohexanone, Comprising Several New Indicators. J . Pharm.
Chim. 1928, 7, 162-173.
(34) Gasteiger, J .; Marsili, M. Iterative Partial Equalization of Orbital
Electronegativity. Tetrahedron 1980, 36, 3219-3228.
(35) Purcel, V. P.; Singer, J . A. A Brief Review and Table of
Semiepirical Parameters Used in the Hu¨ckel Molecular Orbital
Method. J . Chem. Eng. Data 1967, 12, 235-246.
(10) LaFemina, R. L.; Graham, P. L.; LeGrow, K.; Hastings, J . C.;
Wolfe, A.; Young, S. D.; Emini, E. A.; Hazuda, D. J . Inhibition
of Human Immunodeficiency Virus Integrase by Bis-Catechols.
Antimicrob. Agents Chemother. 1995, 39, 320-324.
(11) Mazumder, A.; Raghavan, K.; Weistein, J .; Kohn, K. W.; Pom-
mier, Y. Inhibition of Human Immunodeficiency Virus Type-1
Integrase by Curcumin. Biochem. Pharmacol. 1995, 49, 1165-
1170.
(36) Head, J .; Zerner, M. C. A Broyden-Fletcher-Goldfarb-Shanno
Optimization Procedure for Molecular Geometries. Chem. Phys.
Lett. 1985, 122, 264-274.
(37) Pauwels, R.; Balzarini, J .; Baba, M.; Snoeck, R.; Schols, D.;
Herdewijn, P.; Desmyster, J .; De Clercq, E. Rapid and Auto-
mated Tetrazolium-Based Assay for the Detection of anti-HIV
Compounds. J . Virol. Methods 1988, 20, 309-321.
(12) Mazumder, A.; Gazit, A.; Levitzki, A.; Nicklaus, M.; Yung, J .;
Kohlhagen, G.; Pommier, Y. Effects of Tyrphostins, Protein
Kinase Inhibitors, on Human Immunodeficiency Virus Type 1
Integrase. Biochemistry 1995, 34, 15111-15122.

3960 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21
Artico et al.
(38) Tomassini, J .; Davies, M. E.; Amstrong, M. E.; Hastings, J .;
Hazuda, D.; Lewis, J .; McClemens, W.; Wolfe, A.; Selnick, H.;
Baldwin, J .; Bourgeois, M.; Ponticello, G.; Radilowski, E.; Smith,
G.; Tebben, A. Inhibition of the cap-dependent endonuclease of
influenza virus by 4-substituted-2,4-dioxobutanoic acid com-
pounds. Antimicrob. Agents Chemother. 1994, 38, 2827-2837.
(39) Lee, K. H.; Imakura, Y.; Haurna, M.; Beers, S. A.; Thurston, L.
S.; Dai, H. J .; Chen, C. H. Antitumor Agent 107. New Cytotoxic
4-Alkylamino Analogues of 4′-Demethyl-epipodophyllotoxin as
Inhibitors of Human DNA Topoisomerase II. J . Nat. Prod. 1989,
52, 606-613.
(44) Glaser, E.; Tramer, E. Condensation Product of Vanillin and
Salicylaldehyde with Acetone and the Mechanism of this Con-
densation. J . Prakt. Chem. 1927, 116, 331-346.
(45) McGookin, A.; Sinclair, D. J . Isomerism of the Styryl Alkyl
Ketones. IV. 4-Hydroxy- and some Derivatives of 2-Hydroxy-
styryl Alkyl Ketones. J . Chem. Soc. 1928, 1170-1177.
(46) Klinke, P.; Gibian, H. Chalcones. Chem. Ber. 1961, 94, 26-38.
(47) Dziedzic, S. Z.; Hudson, B. J . F. Polyhydroxy Chalcones and
Flavanones as Antioxidants for Edible Oils. Food Chem. 1983,
12, 205-212.
(48) Geissman, T. A.; Clinton, R. O. Flavanones and Related Com-
pounds. I. The Preparation of Polyhydroxychalcones and Fla-
vanones. J . Am. Chem. Soc. 1946, 68, 697-700.
(49) J entzsch, K.; Spiegel, P.; Kamitz, R. Qualitative and Quantita-
tive Studies of Curcuma Dyes in Different Zingiberaceae Drugs.
2. Quantitative studies. Sci. Pharm. 1970, 38, 50-58.
(50) Pabon, H. J . J . A Synthesis of Curcumin and Related Com-
pounds. Recl. Trav. Chim. Pays Bas 1964, 83, 379-386.
(40) Samdahl, B. Some New Indicators, Derivatives of Cyclohex-
anone. J . Pharm. Chim. 1930, 11, 8-19.
(41) Samdhal, B.; Hansen, B. Condensation of Cyclopentanone with
Phenol Aldehydes. J . Pharm. Chim. 1928, 7, 162-173.
(42) Ionescu, M. V. New Reaction for Aromatic Aldehydes. Bull. Soc.
Chim. 1930, 47, 210-214.
(43) Haworth, R. D.; Lamberton, A. H. Some Derivatives of Catechol
and Pyrogallol. II. J . Chem. Soc. 1946, 1003-1005.
J M9707232