Rational Design and Synthesis of Novel Dimeric Diketoacid-Containing Inhibitors of HIV-1 Integrase: Implication for Binding to Two Metal Ions on the Active Site of Integrase

Article abstract of DOI:10.1021/jm030559k

Discovery of diketoacid-containing compounds as HIV-1 integrase (IN) inhibitors played a major role in validating this enzyme as an important target for the development of therapeutics against HIV infection. In fact, S-1360, the first clinically used IN inhibitor containing a triazole ring as a bioisostere of a carboxylic acid moiety belongs to this class of compounds. To understand the role of divalent metal-chelating in the inhibition of IN (J. Med. Chem. 2002, 45, 5661-5670), we designed and synthesized a series of novel dimeric diketo-containing compounds with the notion that such dimeric compounds may simultaneously bind to two divalent metal ions on the active site of IN. We rationalized that the two diketo subunits separated by uniquely designed linkers can potentially chelate two metal ions that are either provided from one IN active site or two active sites juxtaposed together in a higher order tetramer. Herein, we show that all the new compounds are highly potent against purified IN with varied selectivity for strand transfer, and that some of the analogues exert potent inhibition of the cytopathic effect of HIV-1 in infected CEM cells. This study represents the first attempt to rationally target two divalent metal ions on the active site of IN and may have potential implications for the design of second generation diketoacid-containing class of inhibitors.

Full text of DOI:10.1021/jm030559k

J . Med. Chem. 2004, 47, 2561-2573
2561
Ra tion a l Design a n d Syn th esis of Novel Dim er ic Dik etoa cid -Con ta in in g
In h ibitor s of HIV-1 In tegr a se: Im p lica tion for Bin d in g to Tw o Meta l Ion s on th e
Active Site of In tegr a se
Ya-Qiu Long,*,† Xiao-Hua J iang,† Raveendra Dayam,‡ Tino Sanchez,‡ Robert Shoemaker,| Shizuko Sei,§ and
Nouri Neamati*,‡
State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Shanghai Institutes for Biological Sciences,
CAS, 555 Zuchongzhi Road, Shanghai 201203, China, Laboratory of Antiviral Drug Mechanisms, SAIC-Frederick,
Screening Technologies Branch, DTP, DCTD, National Cancer Institute at Frederick, Frederick, Maryland 21702, and
Department of Pharmaceutical Sciences, School of Pharmacy, University of Southern California, Los Angeles, California 90089
Received November 4, 2003
Discovery of diketoacid-containing compounds as HIV-1 integrase (IN) inhibitors played a major
role in validating this enzyme as an important target for the development of therapeutics against
HIV infection. In fact, S-1360, the first clinically used IN inhibitor containing a triazole ring
as a bioisostere of a carboxylic acid moiety belongs to this class of compounds. To understand
the role of divalent metal-chelating in the inhibition of IN (J . Med. Chem. 2002, 45, 5661-
5670), we designed and synthesized a series of novel dimeric diketo-containing compounds
with the notion that such dimeric compounds may simultaneously bind to two divalent metal
ions on the active site of IN. We rationalized that the two diketo subunits separated by uniquely
designed linkers can potentially chelate two metal ions that are either provided from one IN
active site or two active sites juxtaposed together in a higher order tetramer. Herein, we show
that all the new compounds are highly potent against purified IN with varied selectivity for
strand transfer, and that some of the analogues exert potent inhibition of the cytopathic effect
of HIV-1 in infected CEM cells. This study represents the first attempt to rationally target
two divalent metal ions on the active site of IN and may have potential implications for the
design of second generation diketoacid-containing class of inhibitors.
In tr od u ction
Integration of viral cDNA into the host genome is a
vital step in the HIV replication cycle. This process is
carried out in two distinct enzymatic reactions by HIV-1
integrase (IN).1,2 In the first step while in the cytoplasm
of an infected cell, IN selectively cleaves the terminal
dinucleotides (...CA8GT) from the 3′-ends of the viral
DNA in a reaction known as 3′-processing. In the second
step, IN transfers viral DNA into host genome wherein
terminal 3′-OH of the viral DNA attacks as a nucleo-
phile on the host DNA. This latter step, referred to as
strand transfer or integration occurs in the nucleus of
the infected cell. Currently, several FDA approved
HIV-1 reverse transcriptase and protease inhibitors are
available for the treatment of symptomatic HIV disease
and acquired immune deficiency syndrome (AIDS).3
Highly active combination antiretroviral therapy has
been very effective in bringing down the viral load;
however emergence of multidrug resistant viral strains
in infected patients can complicate the response to the
treatment.4 Because of its vital role in the viral replica-
tion cycle, the addition of an IN inhibitor to existing
components of combination regimen is expected to
improve the outcome of therapy.
F igu r e 1. Structures of selected HIV-1 IN inhibitors. Com-
pound 1 (S-1360) is in Phase II clinical trials.
In the past several years, numerous compounds with
diverse structural features have been reported as IN
inhibitors,5-7 of which the most developed are a class
of compounds with â-diketoacids.8 Indeed, S-1360 (1)
was the first drug to enter into clinical trials.9,10 Several
reported â-diketoacids selectively inhibit the strand
transfer reaction of IN and exhibit potent antiviral
effects against HIV-infected cells. Further studies have
indicated that, unlike numerous other IN inhibitors
whose antiviral effects can be attributable to non-IN-
dependent phenomena,11,12 members of the â-diketoacid
family disrupt viral infectivity in a manner consistent
with inhibition of integration.8 It is believed that the
â-diketoacid inhibitors function by competing with
substrate DNA in binding to the IN active site.13
Further structure-activity relationship (SAR) studies
were subsequently carried out on â-diketoacids and
several novel compounds were found to be active against
* To whom correspondence should be addressed. Y.-Q.L., Phone: 86-
21-50806876; fax: +86-21-50807088; e-mail: yqlong@mail.shcnc.ac.cn.
N.N., Phone: 323-442-2341; fax: 323-442-1390; e-mail: neamati@usc.
edu.
† Shanghai Institute of Materia Medica.
§ Laboratory of Antiviral Drug Mechanisms, SAIC-Frederick,
National Cancer Institute.
| Screening Technologies Branch, DTP, DCTD, National Cancer
Institute.
‡ University of Southern California.
10.1021/jm030559k CCC: $27.50 © 2004 American Chemical Society
Published on Web 04/10/2004

2562 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10
Long et al.
IN consists of three distinct structural domains: the
zinc binding N-terminal, the catalytic core and the DNA
binding C-terminal.20 Structures of each individual
domain of IN and the core plus N terminus and core
plus C-terminus were previously determined; neverthe-
less, the X-ray structure of the full-length enzyme has
proven to be difficult. The relative spatial arrangement
of these structural domains in an intact IN and their
interaction with DNA substrates has remained largely
unknown. Recently, an X-ray structure of a dimer of IN
core domain complexed with an inhibitor was reported,
wherein the inhibitor interacted with two crystal-
lograpically related active sites.21
Indeed, purified integrases of various retroviruses
have been reported to exist as dimers, tetramers, and
oligomers in solution.22-24 It was demonstrated that IN
exists in equilibrium between dimeric and tetrameric
species in solution.25 The biochemical and structural
studies suggest that a multimer, probably a tetramer
of IN, is required for the full integration reaction.26,27
Two of the four active sites of an active tetramer are
involved in the integration reaction, and the other two
active site bearing units structurally support the active
tetramer. The dimer of dimers observed in the X-ray
structure of the IN core and N-terminal domain is a
model for a tetrameric unit of IN (Figure 2).27 In a dimer
unit of IN, two active sites are located on the opposite
sides of the dimer (Figure 3). These sites never come
closer because the monomer-monomer interface is
highly hydrophobic in nature and thus energetically
F igu r e 2. The tetrameric arrangement of IN (core and
N-terminal domains; PDB 1K6Y). Each of the colored ribbons
represents a monomeric unit of IN. The two adjacent active
sites (one from each dimer) are believed to participate together
in the integration reactions of IN. The prominent amino acid
residues on the IN active site are rendered as red stick models.
CPK model shows the phosphate ion near the IN active site.
IN14-17 (for recent reviews see 5,18). Compounds 2 and
3 belong to such a new class of â-diketoacid containing
compounds collectively referred here as aryl â-diketoac-
ids (ADK).19
F igu r e 3. (A) Two dimeric units of IN (N-terminal and core domains; PDB 1K6Y) form a tetrameric complex (B) in such a way
that two of the four active sites of tetramer come close and the two adjacent active sites are believed to participate together in the
integration reactions of IN. (C) The newly designed benzyloxy-linked ADK dimer (5c) is believed to interact with two neighboring
active sites of the functional tetramer of IN. The red stick models represent prominent amino acid residues inside the active site
of IN while the ribbon models represent N-terminal and core domains of IN. CPK model shows the phosphate ion which is found
near the active site of IN (PDB 1K6Y).

Dimeric Diketoacid Containing Inhibitors of HIV-1 Integrase
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10 2563
F igu r e 4. (A) The bound conformation of the compound 3 inside the IN active site. The diketoacid functionality of 3 interacting
with Mg2+ ion while the aryl part of the compound occupied an area close to E152, N155, K156, and K159. The violet ball-and-
stick model represents 3 while the ribbon model shows active site region of the IN. (B) The carboxylate oxygen atoms of the
catalytically important amino acid residues E152, D116, and D64 (DD(35)E) are shown as balls on apexes of a triangle (PDB
1BIS). The Mg2+ is shown as magenta. (C) The distances between the nearest carboxylate oxygen of E152, D116, and D64 are
given in Å.
unfavorable. In the tetramer arrangement, two of the
four active sites (one from each dimer) are located close
to each other (Figure 2). These active sites are separated
by ∼40 Å, which is much longer than the five base pair
spacing of the sites of insertion of the two viral DNA
ends into host DNA. However, the C-terminal domain
is missing in this tetramer model, and the relative
arrangement of intact IN as a monomer within an active
tetramer is unknown. In the functionally relevant
tetrameric arrangement of intact IN, these two adjacent
active sites probably move close together with the help
of certain conformational changes within a dimeric unit
of intact IN in order to facilitate integration reactions.
Recently, models were proposed for the full-length IN
dimer and a tetramer complexed with substrate DNA.
The model of the tetrameric arrangement of full-length
IN supports the possible involvement of two neighboring
active sites from two dimers in integration reactions.28,29
Although, the requirement of a metal cofactor for
catalysis in IN and many other DNA processing en-
zymes is well accepted, the number of metal ions and
in some cases the type of metal have been the subject
of debate for more than three decades.30 It is generally
accepted that Mg2+ is a more reasonable cofactor for
integration because of its 1 000 000-fold abundance over
Mn2+ in cells.30,31 There are also growing and popular
notions in support of a two-metal ions theory (for
excellent reviews see32,33). Even though the X-ray
structures show one metal ion on the active site of IN,
it is believed that the second metal ion comes with DNA.
However, currently there are no IN-DNA complex
structures available. After a careful review of the
literature we concluded that the second metal ion most
likely binds to E152, whereas the first one binds to D64
and D116. The second metal ion can potentially coor-
dinate the carboxylate oxygen atoms of E152 and
possibly D64. In a recent molecular dynamics study
second Mg2+ ion was placed between E152 and D64, and
it was found that the second Mg2+ was more stable when
it was coordinated with an HPO4 ion which mimics a
-2
portion of the DNA backbone.34 We further hypothesized
that the two metal ions may not have to be present at
the active site of IN through its entire enzymatic
reaction. For example, the first metal ion, which binds
D64 and D116, could play a major role in the 3′-
processing reaction, and the second one, which binds
to E152, is maybe more relevant during the strand
transfer reaction. Sequential metal dependency can also
be explained by other situations. Because the smallest
subunit required for catalysis is a tetramer of IN,
individual monomers could share metal ions. In this
scenario, the number of metal ions for a tetrameric
complex must be greater than two and less than four.
With an aim to discover novel compounds that could
interact with two metal ions31 within an active site or
on adjacently located active sites of a functionally active
tetramer of intact IN, we designed two sets of aryl
diketoacid dimers (ADK dimers). These dimers dis-
played an interesting and enhanced inhibitory activity
against IN. The selectivity of our novel dimers for strand
transfer over 3′-processing reaction varied from 1- to 29-
fold. In this study, we describe design, synthesis, SAR,
and active site bound conformations of two different sets
of ADK dimers.
Design
The rationale behind the design of ADK dimers came
mainly from our observations on the bound conforma-
tion of compound 3 inside the active site of IN, the
relative positions of three catalytically important amino
acids residues E152, D116, and D64 at the IN active
site, and the tetrameric arrangement of active intact
IN. In its bound conformation, the diketoacid moiety of
3 occupied an area near amino acid residues D64, C65
and Mg2+. Subsequently, two of the four oxygen atoms
of the diketoacid moiety formed coordination bonds with

2564 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10
Long et al.
Sch em e 1a
Ta ble 1. Distances between the Nearest Carboxylate Oxygen
Atoms of the Amino Acid Residues E152, D116, and D64 from
the Known Crystal Structures of IN
amino acid residues (distance in Å)
PDB accession codes ref E152-D116 E152-D64 D116-D64
1B92 (A)a
1B9D (A)
1BI4 (A)
1BIS (B)
1BIU (A)
1BIZ (A)
1BL3 (C)
1EX4 (A)
1K6Y (A)
1QS4 (A)
2ITG (A)
1EXQ (A)
39
39
40
38
38
38
40
37
27
21
35
37
5.62b
5.56b
9.62
10.90
9.40
7.40
7.71
5.91
11.90
10.48
12.54
11.64
9.31
9.25
6.57
8.43
7.68
10.26
5.00
4.91
8.65
8.28
13.88
7.53
4.46
4.47
3.45
4.29
2.93
4.13
3.06
3.71
3.77
2.95
4.15
4.76
a
Reagents: (a) n-BuLi, THF, -78 °C, 95%; (b) Et3SiH, BF3Et2O,
0 °C, CH2,Cl2, 0 °C, 84%; (c) n-BuLi, DMAc, THF, -78 °C, 63%;
(d) CH3ONa, MeO2CCO2Me, toluene/DME ) 1:1, 0 °C -60 °C; (e)
THF/CH3OH ) 1:1, 1 N NaOH, rt; (d,e two steps overall yield
53.3%) (f) EDCl, HOBt, DMF, 0 °C, piperazine.
a
The letter in parentheses indicates the subunit of the crystal
b
structure. Please note that several amino acid residues are
missing in the active site region of these structures and D64
adopted very different orientation.
Sch em e 2a
Mg2+ (Figure 4a). The aryl part of 3 occupied a cavity
near residue E152, N155, K156, and K159. The varia-
tion in the distance between the nearest pair of car-
boxylate oxygen atoms of residues E152, D116, and D64
in the reported X-ray structures indicates the flexible
nature of a part of the active site which harbors residue
E152 (Table 1 and Figure 4b).21,27,35-40 The closest
distance between carboxylate oxygen atoms of D64 and
D116 is 2.95 Å in PDB 1QS4, in which Mg2+ is
coordinated with the carboxylate oxygen atoms of D64
and D116. The difference in the distances between E152
and D116 (5.6 to 12.5 Å) and E152 and D64 (4.9 to 13.8
Å) indicates a possibility that in the presence of a second
metal ion the E152 may move closer to D64, and these
two amino acid residues hold the second metal ion
wherein D64 acts as a bridge between the two metal
ions. We rationalized that the diketo functionality of the
â-diketoacid class of compounds is responsible for metal
binding as well as anti-IN activity in isolated enzyme
assays and infected cells. On the basis of the bound
conformation of 3 and the probable location of the second
metal atom between E152 and D64 inside the IN active
site, we considered that the introduction of a second
diketo moiety onto 3 would significantly affect its
activity. Thus, a compound with properly separated
diketo moieties could chelate both metal ions at the IN
active site. We also hypothesize that the second diketo
moiety could interact with the second metal ion on the
adjacent active site on the surface of the tetrameric unit.
To evaluate these concepts, we designed two sets of ADK
dimers, wherein two aryl diketo monomers are con-
nected by various linkers such as cyclic/acyclic diamines
and benzyloxy isomers.
a
Reagents: (a) Benzoxyl peroxide, NBS, CCl4; (b) 3-hydroxy-
acetophenone, K2CO3, DMF; (c) MeOOCCOOtBu, CH3ONa; (d) 1
N NaOH, THF/CH3OH.
crude adduct 6 to triethylsilane in the presence of boron
trifluoride etherate,41 provided the corresponding 2-fluo-
robenzylphenyl bromide 7. Bromide 7 was then lithi-
ated, and the resulting solution was treated with
N-methoxy-N-methylacetamide to provide ketone 8.42
Treatment of 8 with dimethyl oxalate and sodium
methoxide provided the intermediate ester adduct 9,43
which were hydrolyzed in situ to provide the diketo acid
3. Various diamines were coupled with 3 in the presence
of HOBt and EDCI to afford corresponding amide-linked
dimers 4a -e.
Ben zyloxy-Lin k ed ADK Dim er s. The synthesis of
benzyloxy-linked ADK dimers 5a -d was achieved in a
similar fashion by coupling bis(methyl ketones) 11a -c
and 16 with oxalylation agent in the presence of base,
with hydrolysis of the resulting bis(methyl esters)
12a -c and 17 being subsequently carried out in the
presence of 1 N NaOH in MeOH-THF at room tem-
perature (Scheme 2, 3). Among the methods investi-
gated for oxalylation, tert-butyl methyl oxalate in the
presence of sodium methoxide proved to be the most
effective.44
Syn th esis
Am id e-Lin k ed ADK Dim er s. The synthesis of the
amide-linked ADK dimers was accomplished by coupling
the monomer 3 with various diamines. The length and
orientation of the linker was designed as described
above. Scheme 1 depicts the chemistry employed in the
preparation of this series of ADK dimers (synthesis of
the piperazine-linked dimer 4a is shown here as an
example). For the preparation of compounds 4a -e,
treatment of the 2-fluorobenzaldehyde with 3-bromo-l-
lithiobenzene, followed by exposure of the resulting
For the synthesis of the benzyloxyl-linked ADK
dimers 5a -c, the required bis-acetophenones were
prepared by treatment of m (o, p)-xylene with NBS in
the presence of benzoxyl peroxide, followed by coupling

Dimeric Diketoacid Containing Inhibitors of HIV-1 Integrase
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10 2565
Ta ble 2. Inhibition of HIV-1 Integrase Catalytic Activities, Antiviral Activities and GOLD Scores of a Series of Amide-Linked ADK
Dimers (4a -e)
Sch em e 3a
linked dimers (4a -e) showed a much better selectivity
for the strand transfer versus 3′-processing reaction
(Figure 5 and Tables 2-3) and a stronger anti-viral
efficacy in HIV-infected cells than the benzyloxy-linked
ADK dimers (5a -d ).
In h ibitor y Activity of Am ide-Lin ked ADK Dim er s
a ga in st IN a n d HIV-1 Rep lica tion . 4-[3-(2-Fluo-
robenzyl)phenyl]-2-hydroxy-4-oxo-but-2-enoic acid (com-
pound 3) was selected as the monomer, which formed a
series of symmetric dimers with various diamines
having different length and orientation. The intrinsic
selectivity for strand transfer of these dimeric inhibitors
decreased as the length of the linear linker increased
from two carbon to six carbon chain (Table 2, compounds
4b-e). Although 4b-d showed high potency and selec-
tivity for strand transfer, they did not show significant
antiviral activity. This lack of antiviral activity was due
to cytotoxicity for reasons that are not obvious. Com-
pounds 4b-d with CC50 values e12 µM were among
the most toxic analogues. Interestingly, the highly rigid
analogue 4a and the highly flexible 4e showed some-
what decreased selectivity for strand transfer as com-
pared to 4b and 4c. Both compounds exhibit antiviral
activity. For example, up to 100% protection of CEM
cells from HIV-induced cytopathic effect was achieved
with compound 4a at 2 µM. Although, compound 4e
(EC50 ) 16 ( 5 µM) was 20-fold less potent than 4a
(EC50 ) 0.8 ( 0.3 µM) it was 11-fold less toxic than 4a .
In conclusion, the size and shape of the linker had
dramatic influence on the strand transfer selectivity,
the antiviral activity as well as the toxicity. Further
studies are underway in our laboratory to better un-
derstand the effect of the linker geometry on activity.
a
Reagents: (a) J ones reagent; (b) 1,3-dibromobenzene, n-BuLi,
-78 °C; (c) (CH3)2SiCl2, NaI, CH3CN; (d) n-BuLi, THF, -78 °C,
N-methoxy-N-methylacetamide; (e) CH3ONa, t-BuCO2CO2Me,
toluene/DME/THF; (f) 1 N NaOH, THF/CH3OH.
of the resulting R,R′-dibromo-m (o, p)-xylene 10 with
3-hydroxyacetophenone and potassium carbonate. For
the preparation of benzyl-linked dimer 5d , the required
bis-methyl ketone 16 was prepared in four steps (Scheme
3, steps a-d). Oxidation of (3-bromophenyl)(2-fluo-
rophenyl)methanol 6 with J ones reagent provided ke-
tone 13, which was subsequently coupled with 3-bromo-
l-lithiobenzene. The resulting tertiary alcohol 14 was
exposed to dimethylsilane chloride in the presence of
sodium iodide and then to a mixture of N-methoxy-N-
methylacetamide and n-BuLi in THF at -78 °C to
provide the desired bis-acetophenone 16.
In h ibitor y Activity of Ben zyloxy-Lin k ed ADK
Dim er s a ga in st IN a n d HIV-1 Rep lica tion in Cell
Cu ltu r e. Two units of 2,4-dioxo-4-phenylbutyric acid
were built into the phenyl ring in various substitution
patterns, which provided a quick survey of the optimum
relative orientation of bis-aryl diketoacid. Among the
Resu lts a n d Discu ssion
The data from two sets of ADK dimers tested against
purified IN and in HIV-1 infected CEM cells are
summarized in Tables 2 and 3. All ADK dimers exhib-
ited potent inhibitory activity against IN. The amide-

2566 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10
Long et al.
F igu r e 5. (A) A representative gel showing inhibition of purified IN by selected compounds. A 21-mer blunt-end oligonucleotide
corresponding to the U5 end of the HIV-1 LTR, 5′ end-labeled with 32P, is reacted with purified IN. The initial step involves
nucleolytic cleavage of two bases from the 3′-end, resulting in a 19-mer oligonucleotide. Subsequently, 3′ ends are covalently
joined at several sites to another identical oligonucleotide that serves as the target DNA. This reaction is referred to as strand
transfer and the products formed migrate slower than the original substrate (shown in the figure as STP for strand transfer
products). Drug concentrations in µM are indicated above each lane. (B) Quantitation of panel A.
Ta ble 3. Inhibition of HIV-1 Integrase Catalytic Activities, Antiviral Activities and GOLD Scores of a Series of Benzyl-Linked ADK
Dimers (5a -d )
benzyl-linked bis-phenyl diketoacids (Table 3, com-
phenyl diketoacid units is in ortho position. The ortho-
pounds 5a -c), the optimum relative orientation of the
substituted bis-phenyl diketoacid 5a exhibited the most

Dimeric Diketoacid Containing Inhibitors of HIV-1 Integrase
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10 2567
Ta ble 4. H-Bonding and Mg2+-Ligand Interactions Observed
potent antiviral activity. An interesting feature among
these three compounds is their lack of toxicity. Thus,
the lack of antiviral activity could be due to two possible
explanations. First, these compounds are not stable in
cell culture (6 days at 37 °C). Second, the compounds
are stable but are not cell permeable. We favor the
second explanation because under similar conditions
they retain their anti-IN activity. It is interesting to note
that compounds 5a and 5d showed moderate selectivity
against strand transfer but showed potent antiviral
activity in HIV-1 infected CEM cells (Figure 5). On the
other hand, compound 5b showed good selectivity for
strand transfer but was inactive as an antiviral agent.
Compound 5c with no selectivity for strand transfer was
also inactive in cell-based assays (Figure 5). In conclu-
sion, we did not observe a correlation between strand
transfer selectivity and antiviral activity. Our future
structure-activity relationship studies will shed more
light on the nature of strand transfer selectivity and
antiviral activity.
with ADK Dimers
compd H-bonding interactions (Å)* Mg2+-ligand interactions (Å)
4a
aCdO‚‚‚HO D116 (2.07)
bHO‚‚‚OH C65 (1.96)
aCdO‚‚‚Mg2+ (3.01)
bHO‚‚‚Mg2+ (1.93)
cCdO‚‚‚Mg2+ (2.12)
aCdO‚‚‚Mg2+ (1.91)
bHO‚‚‚Mg2+ (2.09)
aCdO‚‚‚Mg2+ (2.50)
bHO‚‚‚Mg2+ (1.95)
aCdO‚‚‚Mg2+ (2.06)
bHO‚‚‚Mg2+ (2.72)
cCdO‚‚‚Mg2+ (1.90)
a′CdO‚‚‚Mg2+ (2.04)
bHO‚‚‚Mg2+ (2.02)
bHO‚‚‚Mg2+ (2.37)
cCdO‚‚‚Mg2+ (1.77)
a′CdO‚‚‚HN K159 (2.27)
c′CdO‚‚‚HO D64 (2.44)
4b
4c
4d
c′CdO‚‚‚HN N62 (2.28)
b′HO‚‚‚HN K159 (2.08)
4e
5a
aCdO‚‚‚HO D64 (2.28)
b′HO‚‚‚HN K159 (1.71)
-Ph-O′‚‚‚HN Q62 (2.86)
b′OH‚‚‚OdC D64 (2.05)
b′HO‚‚‚HN N155 (2.42)
c′O‚‚‚HN N155 (2.19)
d′O‚‚‚HN K159 (2.04)
b′HO‚‚‚HN Q62 (1.95)
c′O‚‚‚HN Q62 (2.30)
5b
5c
bHO‚‚‚Mg2+ (1.65)
dO‚‚‚Mg2+ (1.91)
b′HO‚‚‚OdC D64 (2.04)
Ph-O‚‚‚HN K159 (1.67)
bHO‚‚‚Mg2+ (2.22)
dO‚‚‚Mg2+ (1.66)
d′O‚‚‚Mg2+ (2.41)
Dock in g Stu d ies. To predict bioactively bound con-
formation of both sets of ADK dimers, we docked all the
compounds into IN active site using GOLD. Currently,
there is no structure available for active tetrameric
arrangement of intact IN; therefore, we used an active
site from a monomeric subunit of IN for docking
purposes. The active molecules bound well into IN active
site, and formed favorable H-bonding and steric interac-
tions with various active site amino acid residues inside
the active site of IN (Tables 2 and 3). The bound
conformation of the monomer 3 (with one diketoacid
functionality) inside the active site of IN is shown in
Figure 4a. The diketoacid moiety of 3 interacts with
amino acid residues D64, C65 and with the Mg2+ ion
while the aryl part of the compound occupied an area
close to E152, N155, K156, and K159. An H-bonding
interaction is observed between fluoro group of 3 and
ꢀ-amine of K159. The binding location of 3 is very
similar with the bound location of 5CITEP in IN-
5CITEP complex crystal structure, but overall position-
ing of the various functional groups is very different.
The diketo moiety of 5CITEP is positioned between
E152 and D64, D116 while the diketoacid moiety of 3
interacts with D64, C65 and Mg2+ ion. Recent modeling
studies revealed that a part (residues 138-149) of IN
active site is highly flexible, and remarkable differences
are observed in the predicted binding conformation of
5CITEP when compared to bound conformation of
5CITEP in the IN-5CITEP complex crystal structure,
as the bound orientation of 5CITEP in crystal structure
is influenced by crystal packing effects.45,46
bOH‚‚‚OdC D64 (2.74)
5d
3b
c′O‚‚‚HN N155 (2.06)
d′O‚‚‚HN K159 (2.08)
bOH‚‚‚OdC C64 (1.41)
F‚‚‚HN K159 (2.80)
bHO‚‚‚Mg2+ (2.31)
aO‚‚‚Mg2+ (1.44)
bHO‚‚‚Mg2+ (1.76)
dO‚‚‚Mg2+ (2.31)
a
See Tables 2 and 3 for structures of 4a -5d . Superscripted
letters indicate the position of H-bond donor or acceptor atoms.
See Figure 1 for compound 3.
b
piperazine linker. The carbonyl oxygen atom from the
second diketo moiety (Table 2, superscripted letters a′,
b′, and c′) formed a strong H-bonding interaction with
ꢀ-amine of K159 (4a CdO‚‚‚HN K159 (2.27 Å)). This
interaction was not observed with the other dimers
which possessed a flexible acyclic linker of varying
length. Due to difference in the flexibility and size of
the linker, the second diketo moiety and its associated
aryl part of the dimers adopted quite different bound
conformation. The second diketo moiety and its associ-
ated aryl part of 4a occupied an area surrounded by
amino acid residues K159, K156, N155, E152, I151, Q62,
and D64. In compound 4d , the second diketo moiety and
its associated aryl part of the compound occupied the
same area as 4a and further extended into a space near
amino acid residues Q148, P142, I141, and H114.
Bou n d Con for m ation s of Ben zyloxy-Lin ked ADK
Dim er s. A part of the compound (the diketoacid moiety
whose oxygen atoms are denoted by a, b, c, and d (see
Table 3)) in all the dimers adopted a quite similar bound
conformation and occupied almost same location inside
the active site of IN. This part of the compound occupied
a space near amino acid residues D64, D116 and Mg2+
ion. The oxygen atoms of the diketoacid moiety coordi-
nate with Mg2+ ion. The bound conformations of 5a and
5c inside the active site of IN are shown in Figure 7.
Due to the difference in the linker (benzyloxy isomers),
the bound orientation of the second diketoacid moiety
(Table 3, a′, b′, c′, and d′) of these dimers is different.
The second diketoacid moiety and central benzyloxy
group of 5a occupied a wide cavity surrounded by amino
acid residues P142, I151, E152, N155, K159, D64, Q62
and formed several H-bonding interactions with N155,
K159, and D64 (Table 4). In case of compound 5b, the
Bou n d Con for m a t ion of Am id e-Lin k ed ADK
Dim er s. All the amide-linked ADK dimers bound to
almost similar locations inside the IN active site. One
of the two diketo moieties (Table 2, superscripted letters
a, b, and c) of all the dimers occupied a cavity close to
D64, H67, E92, D116, and Mg2+. The oxygen atoms of
this diketo group formed coordinate bonds with Mg2+
in all the dimers (Table 4). Due to strong metal ligand
interactions, this portion of the compound in all the
dimers adopted very similar bound conformation. The
bound conformations of compounds 4a and 4d are
shown in Figure 6. The compound 4a adopted somewhat
different bound conformation, as it has a constrained

2568 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10
Long et al.
F igu r e 6. The bound conformations of amide-linked ADK dimers (A) 4a and (B) 4d inside the active site of IN. The violet ball-
and-stick models represent 4a and 4d while the ribbon model represents active site region of the IN. The prominent amino acid
residues are rendered as stick models. The Mg2+ ion is shown as magenta.
F igu r e 7. The bound conformations of benzyloxy-linked ADK dimers (A) 5a and (B) 5c inside the active site of IN. The violet
ball-and-stick models represent 5a and 5c while the ribbon model represents the active site region of IN. The prominent amino
acid residues at IN active site are rendered as stick models. The Mg2+ ion is shown as magenta.
second diketoacid moiety and central benzyloxy group
occupied a cavity surrounded by N155, E152, I151, D64,
Q62 and formed H-bonding interactions with D64 and
Q62. The central benzyloxy group of 5c occupied an area
near N155, E152, I151, and one of the carboxylate
oxygen atoms of the second diketoacid moiety coordi-
nates with Mg2+ ion (Figure 7B). An H-bonding interac-
tion is observed between the enol hydroxyl of the second
diketoacid moiety and D64 (dimer OH‚‚‚OdC D64 (2.41
Å)). The second diketoacid moiety of 5d occupied an area
near N155, K159, and the metal ligand and H-bonding
interactions are given in Table 4.
In summary, the bound conformation of one of the two
diketo groups, which interacts with Mg2+ ion, is very
similar in both sets of ADK dimers. At least two of the
three (amide-linked dimers) or four (benzyloxyl-linked
dimers) oxygen atoms of this diketo moiety coordinate
with Mg2+ ion (Table 4). Except 4a , the second diketo
functionality of the amide-linked dimers occupied a
cavity close to E152. In the presence of the second metal

Dimeric Diketoacid Containing Inhibitors of HIV-1 Integrase
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10 2569
bovine serum albumin, 10 mM 2-mercaptoethanol, 10% di-
methyl sulfoxide, and 25 mM MOPS, pH 7.2) at 30 °C for 30
min. Then, 20 nM of the 5′-end 32P-labeled linear oligonucle-
otide substrate was added, and incubation was continued for
an additional 1 h. Reactions were quenched by the addition of
an equal volume (16 µL) of loading dye (98% deionized
formamide, 10 mM EDTA, 0.025% xylene cyanol and 0.025%
bromophenol blue). An aliquot (5 µl) was electrophoresed on a
denaturing 20% polyacrylamide gel (0.09 M Tris-borate pH 8.3,
2 mM EDTA, 20% acrylamide, 8 M urea).
Gels were dried, exposed in a PhosphorImager cassette, and
analyzed using a Typhoon 8610 Variable Mode Imager (Am-
ersham Biosciences) and quantitated using ImageQuant 5.2.
Percent inhibition (% I) was calculated using the following
equation:
ion, the second diketo moiety of these dimers could
interact with the second metal ion, which further
provides energetically favorable interactions between
this set of dimers and IN active site. We also speculate
that the second diketoacid moiety of the benzyloxy-
linked dimers could interact with an Mg2+ ion located
on the adjacent active site of a functionally relevant
tetrameric IN as the length (the distance between two
terminal carboxylate oxygen atoms) of this set of dimers
in their extended conformation spans from 18.25 Å (5d )
to 27.18 (5c), which is close to the length of five base
pairs on the target DNA.
Con clu sion s
We have successfully designed and synthesized a
series of novel diketo containing compounds to target
two metal ions on the active site of IN. Our initial
docking studies indicated that one of the diketo groups
of ADK dimers interacts with Mg2+ ion, while the second
diketo moiety occupies an area near E152. Our study
shows that the selectivity for strand transfer is impor-
tant for antiviral activity but not essential. Moreover,
the free carboxylate group is not essential for activity
either against purified IN or in cell-based assays.
Further studies are in progress to optimize the size and
the shape of the linker and will be reported in due
course.
% I ) 100 × [1 - (D - C)/(N - C)]
where C, N, and D are the fractions of 21-mer substrate
converted to 19-mer (3′-processing product) or strand transfer
products for DNA alone, DNA plus IN, and IN plus drug,
respectively. The IC50 values were determined by plotting the
logarithm of drug concentration versus percent inhibition to
obtain concentration that produced 50% inhibition.
An ti-HIV Assa ys in Cu ltu r ed Cells. The anti-HIV activity
was evaluated in human T cell line CEM-SS infected with
HIV-1 as described by Weislow et al.47 In brief, cells were
plated in 96-well plates at 5 × 103 cells/well and infected with
HIV-1RF (MOI ) 0.3). Serial dilutions of compounds were then
immediately added to the cells in a final volume of 200 µL. In
each experiment, AZT and dextran sulfate were included as
control compounds for anti-HIV activity. The cells were
maintained at 37 °C with 5% CO2-containing humidified air
for 6 days. Cell viability was quantified by absorbance at 450
nm after 4 h incubation with 2,3-bis[2-methoxy-4-nitro-5-
sulfophenyl]-5-[(phenylamino)carbonyl]-2H-tetrazolium hy-
droxide (XTT) at 0.2 mg/mL. Antiviral activity was graded
based on the degree of anti-HIV protection as active (80-100%
protection), moderate (50-79% protection), and inactive (0-
49% protection). Toxicity of the compounds was determined
simultaneously on the same plate in uninfected CEM-SS cells.
Molecu la r Mod elin g. The structures of all the compounds
(Tables 1 and 2) were built and minimized using Catalyst
(Accelrys, Inc). All compounds were modeled in their enol
tautomeric form as this form is considered biologically rel-
evant. Compounds 3, 5a -d were modeled as carboxylates
(dianionic form). The poling algorithm implemented within
Catalyst was used to generate conformations for all the
compounds.48 For each compound all feasible unique confor-
mations were generated over a 20 kcal/mol range of energies
using the best flexible conformation generation method in
Catalyst. The subunit B of the core domain X-ray structure of
IN (PDB 1BIS) in which all the active site amino acid residues
were resolved was chosen for docking purpose. A Mg2+ ion was
placed in the active site between carboxylate oxygen atoms of
amino acid residues D64 and D116 considering the geometry
of the Mg2+ ion that was present in the subunit A of the IN in
PDB 1BIS and subunit A in IN-5CITEP complex X-ray
structure (PDB 1SQ4). All the water molecules present in
protein were removed, and hydrogen atoms were added to the
protein considering appropriate ionization states for both the
acidic and basic amino acid residues. Docking was preformed
using version 1.2 of the GOLD: Genetic Optimization for
Ligand Docking (Cambridge Crystallographic Data Centre)
software package.49 A 12 Å radius active site was defined
considering the carboxylate oxygen atom (OD1) of amino acid
residue D64 as the center of the active site. All conformers of
the compounds were docked into the active site of the IN. On
the basis of the GOLD fitness score, for each molecule, a bound
conformation with high fitness score was considered as the
best bound conformation. All docking runs were carried out
using standard default settings with a population size of 100,
a maximum number of 100 000 operations, and a mutation
and crossover rate of 95. The fitness function that was
Exp er im en ta l Section
Syn th esis. Gen er a l Syn th etic Meth od s. Unless other-
wise stated, the1H NMR spectra were recorded on a Varian
400-MHz spectrometer. The data are reported in parts per
million relative to TMS and referenced to the solvent in which
they were run. Elemental analyses were obtained using a
Vario EL spectrometer. Melting points (uncorrected) were
determined on a Buchi-510 capillary apparatus. IR spectra
were recorded on Bio-Rad FTS-185 spectrometers. HRMS (EI)
spectra were obtained on a Finnigan MAT 95 mass spectrom-
eter. The MS and HRMS (ESI) spectra were obtained on an
APEXIII 7.0 T FTMS mass spectrometer. The solvent was
removed by rotary evaporation under reduced pressure, and
flash column chromatography was performed on silica gel H
(10-40 µm). Anhydrous solvents were obtained by redistilla-
tion over sodium wire.
Ma ter ia ls, Ch em ica ls, a n d En zym es. All compounds
were dissolved in DMSO, and the stock solutions were stored
at -20 °C. The γ [32P]-ATP was purchased from either
Amersham Biosciences or ICN. The expression systems for the
wild-type IN and soluble mutant INF185KC280S were generous
gifts of Dr. Robert Craigie, Laboratory of Molecular Biology,
NIDDK, NIH, Bethesda, MD.
P r ep a r a tion of Oligon u cleotid e Su bstr a tes. The oligo-
nucleotides 21top, 5′-GTGTGGAAAATCTCTAGCAGT-3′ and
21bot, 5′-ACTGCTAGAGATTTTCCACAC-3′ were purchased
from Norris Cancer Center Core Facility (University of South-
ern California) and purified by UV shadowing on polyacryla-
mide gel. To analyze the extent of 3′-processing and strand
transfer using 5′-end-labeled substrates, 21top was 5′-end-
labeled using T4 polynucleotide kinase (Epicenter, Madison,
WI) and γ [32P]-ATP (Amersham Biosciences or ICN). The
kinase was heat-inactivated, and 21bot was added in 1.5-molar
excess. The mixture was heated at 95 °C, allowed to cool slowly
to room temperature, and run through a spin 25 mini-column
(USA Scientific) to separate annealed double-stranded oligo-
nucleotide from unincorporated material.
In tegr a se Assa ys. To determine the extent of 3′-processing
and strand transfer, wild-type IN was preincubated at a final
concentration of 200 nM with the inhibitor in reaction buffer
(50 mM NaCl, 1 mM HEPES, pH 7.5, 50 µM EDTA, 50 µM
dithiothreitol, 10% glycerol (w/v), 7.5 mM MnCl2, 0.1 mg/mL

2570 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10
Long et al.
implemented in GOLD consisted basically of H-bonding,
complex energy, and ligand internal energy terms.
1H NMR (CDCl3): δ 7.88-7.84 (m, 2H); 7.49-7.42 (m, 2H);
7.25-7.17 (m, 2H); 7.15 (s, 1H); 7.10-7.04 (m, 2H); 4.08 (s,
2H). EI-MS m/z: 300 (M+).
(3-Br om op h en yl)(2-flu or op h en yl)m eth a n ol (6). To a
cold solution (-78 °C) of 1.6 M n-BuLi in hexanes (6.5 mL,
10.4 mmol) and 20 mL of anhydrous THF was added 1,3-
dibromobenzene (1.2 mL, 10.0 mmol) under N2. The stirred
solution was kept below -70 °C. After the reaction was stirred
for 2.5 h at this temperature, 2-fluorobenzaldehyde (1.06 mL,
10.4 mmol) was added slowly to keep the temperature below
-50 °C. After 0.5 h being stirred at -78 ∼ -50 °C, the reaction
mixture was allowed to room temperature and stirred for 3 h.
The reaction mixture was poured into aqueous ammonium
chloride under stirring and extracted with ether. The organic
layer was washed with saturated aqueous NaHCO3 and brine
and dried over Na2SO4. After removal of the solvent, the
residue was chromatographed (15:1 ratio of petroleum ether:
4-[3-(2-Flu or oben zyl)ph en yl]-1-(4-{4-[3-(2-flu or oben zyl)-
p h en yl]-2-h yd r oxy-4-oxob u t -2-en oyl}p ip er a zin -1-yl)-2-
h yd r oxy-bu t-2-en e-1,4-d ion e (4a ). A solution of compound
3 (0.15 g, 0.5 mmol), EDCI (0.101 g, 0.525 mmol), and HOBt
(0.071 g, 0.525 mmol) in dry DMF (1.5 mL) was stirred at 0
°C for 15 min. To this solution was added piperazine (0.02 g,
0.23 mmol) in dry DMF. The reaction mixture was stirred for
2.5 h at room temperature and then poured into ice-water
and extracted with dichloromethane. The combined organic
layers were washed with water and brine and dried over Na2-
SO4. The concentration provided the residue which was
purified by chromatography using CH2Cl2/CH3OH (60:1) as
1
eluent to give compound 4a (0.056 g, 40%) as brown solid. H
1
NMR (CDCl3): δ 7.79-7.76 (m, 4H); 7.44-7.38 (m, 4H); 7.25-
7.19 (m, 2H); 7.17-7.13 (m, 2H), 7.09-7.03 (m, 4H); 4.05 (s,
4H), 3.73-3.78 (m, 8H). ESI-MS m/z: 651.1 (M+ + H, 100);
674.2 (M+ + Na, 12.0). HR-ESIMS calcd. for C38H32O6N2-
F2+Na: 673.2122, found: 673.2123. Anal. (C38H32O6N2F2) C,
H, N.
4-[3-(2-F lu or ob en zyl)p h en yl]-2-h yd r oxy-4-oxob u t -2-
en oic Acid (2-{4-[3-(2-flu or oben zyl)p h en yl]-2-h yd r oxy-4-
oxobu t-2-en oyla m in o}eth yl)a m id e (4b). The treatment of
aryl diketoacid 3 with ethane-1,2-diamine as described above
for the preparation of 4a provided the desired product 4b as
a pale yellow solid (1.83 g, 65% yield) after chromatography
(CH2Cl2:CH3OH ) 60:1). 1H NMR (CDCl3): δ 7.87-7.83 (m,
6H); 7.75-7.42 (m, 4H); 7.19 (s, 2H); 7.26-7.02 (m, 6H), 4.05
(s, 4H); 3.64 (m, 4H). ESI-MS m/z: 625.2 (M+ + H), 647.2 (M+
+ Na). HR-ESI-MS calcd for C36H30O6N2F2 (M+Na)+: 647.1970,
found: 647.1975.
ether) to provide 2.81 g of 6 as yellow oil in yield of 95%. H
NMR (CDCl3): δ 7.59 (s, 1H); 7.39-7.53 (m, 2H); 7.20-7.27
(m, 2H); 7.10-7.19 (m, 2H); 6.95-7.03 (m, 1H); 6.02 (d, 1H);
2.83 (d, 1H).
(3-Br om op h en yl)(2-flu or op h en yl)m et h a n e (7). To a
solution of compound 6 (1.183 g, 4.21 mmol) and triethylsilane
(0.81 mL, 5.05 mmol) in dichloromethane (15 mL) was added
neat boron trifluoride etherate (0.64 mL, 5.05 mmol) under
N2 at 0 °C. The reaction mixture was stirred at room-
temperature overnight and then poured into saturated aque-
ous NaHCO3 and extracted with ether. The combined organic
phases were washed with brine and dried over Na2SO4. After
evaporation, the residue was purified by column chromatog-
raphy using petroleum ether to afford compound 7 (0.971 g,
yield 84%) as yellow oil. 1H NMR (CDCl3): δ 7.35-7.30 (m,
2H); 7.20-7.00 (m, 6H); 3.95 (s, 2H).
1-[3-(2-F lu or oben zyl)p h en yl]eth a n on e (8). To a solution
of compound 7 (0.67 g, 2.53 mmol) in THF (10 mL) was added
dropwise 1.6 M n-butyllithium in hexanes (1.9 mL, 3.03 mmol)
under N2 at -78 °C. The reaction mixture was stirred for 2 h
at this temperature, and then N,N-dimethylacetamide (0.35
mL, 3.80 mmol) was added slowly so that the temperature
remained below -50 °C. After 0.5 h being stirred at this
temperature, the reaction mixture was poured into aqueous
HCl (1N)/ice under vigorous stirring and extracted with ether.
The combined organic phases were washed with brine and
dried over Na2SO4, and the solvent was removed under
vacuum. The residue was purified by chromatography using
petroleum ether/ethyl acetate (10:1) as eluent to give com-
4-[3-(2-F lu or ob en zyl)p h en yl]-2-h yd r oxy-4-oxob u t -2-
en oic Acid (3-{4-[3-(2-flu or oben zyl)p h en yl]-2-h yd r oxy-4-
oxob u t -2-en oyla m in o}p r op yl)a m id e (4c). Compound 4c
was obtained as pale yellow solid in 68% yield according to
1
the same procedure as 4a . H NMR (CDCl3): δ 7.88-7.84 (m,
6H); 7.44-7.38 (m, 4H); 7.21 (s, 1H); 7.24-7.13 (m, 3H), 7.09-
7.02 (m, 3H); 4.06 (s, 4H); 3.47 (q, 4H, J ) 6.42 Hz); 1.84 (m,
2H). ESI-MS m/z: 661.4 (M+ + Na, 100). HR-ESI-MS calcd
for C37H32O6N2F2 (M + Na)+: 661.2126, found: 661.2105. Anal.
(C37H32O6N2F2) C, H, N.
4-[3-(2-F lu or ob en zyl)p h en yl]-2-h yd r oxy-4-oxob u t -2-
en oic Acid (5-{4-[3-(2-flu or oben zyl)p h en yl]-2-h yd r oxy-4-
oxobu t-2-en oyla m in o}p en tyl)a m id e (4d ). Treatment of
aryl diketoacid 3 with pentane-1,5-diamine as described above
for the preparation of 4a provided the desired product 4d as
a pale yellow solid (1.2 g, 70% yield) after chromatography
(CH2Cl2:CH3OH ) 60:1). 1H NMR (CDCl3): δ 7.87-7.82 (m,
4H); 7.44-7.38 (m, 4H); 7.28-7.25 (m, 4H); 7.19 (s, 1H); 7.24-
7.13 (m, 3H), 7.09-7.02 (m, 3H); 4.06 (s, 4H); 3.41 (q, 4H, J )
6.86 Hz); 1.69-1.62 (m, 4H); 1.48-1.40 (m, 2H). ESI-MS m/z:
667.2 (M + H) +, 689.3 (M + Na) +. HR-ESI-MS calcd for
1
pound 8 (0.288 g, 63% yield) as yellow oil. H NMR (CDCl3):
δ 7.83-7.77 (m, 2H); 7.37-7.19 (m, 2H); 7.19-7.15 (m, 2H);
7.14-7.03 (m, 2H); 4.04 (s, 2H); 2.57 (s, 3H). EI-MS m/z: 228
(M+).
4-[3-(2-F lu or ob en zyl)p h en yl]-2,4-d ioxob u t yr ic Acid
Meth yl Ester (9). To a cold solution (0 °C) of CH3ONa (2.91
g, 53.95 mmol) in dry toluene were added dimethyl oxalate
(3.18 g, 26.97 mmol) and compound 8 in dry DME (45 mL)
under N2. The solution was stirred for 0.5 h at 0 °C and then
heated to 60 °C overnight. The reaction mixture was quenched
with 1.0 N HCl and extracted with EtOAc. The combined
organic layers were washed with saturated aqueous NaHCO3
and brine and dried over Na2SO4. The solvent was removed
under vacuum. The residue was chromatographed using
petroleum ether/ethyl acetate (4:1) to give compound 9 as a
C
39H36O6N2F2 (M + Na)+: 689.2439, found: 689.2445. Anal.
(C39H36O6N2F2) C, H, N.
4-[3-(2-F lu or ob en zyl)p h en yl]-2-h yd r oxy-4-oxob u t -2-
en oic Acid (6-{4-[3-(2-flu or oben zyl)p h en yl]-2-h yd r oxy-4-
oxobu t-2-en oyla m in o}h exyl)a m id e (4e). Compound 4e was
prepared by coupling aryl diketoacid 3 with hexane-1,6-
diamine in a manner similar to 4a . Purification from chroma-
tography (CH2Cl2/CH3OH ) 60:1) afforded a pale yellow solid
in yield of 70%. 1H NMR (CDCl3): δ 7.87-7.82 (m, 4H); 7.44-
7.38 (m, 4H); 7.19(s, 2H); 7.27-7.13 (m, 5H), 7.09-7.02 (m,
3H); 4.06 (s, 4H); 3.39 (q, 4H, J ) 6.73); 1.63-1.60 (m, 4H);
1.43-1.40 (m, 4H). ESI-MS m/z: 681.3 (M + H)+, 703.4 (M +
Na)+. HRMS (ESI) calcd for C40H38O6N2F2 (M + Na)+: 703.2596,
found: 703.2577. Anal. (C40H38O6N2F2) C, H, N.
r,r′-Dibr om o-O-xylen e (10a ). Benzoxyl peroxide (18 mg)
was added to a solution of NBS (6.675 g, 37.5 mmol) and
O-xylene (1.85 mL, 15.0 mmol) in CCl4 (30 mL). The reaction
was refluxed for 36 h with stirring. The precipitate was filtered
off at room temperature, and the solvent was removed under
reduced pressure. The residue was chromatographed with
1
pale yellow solid. Mp 58-59 °C. H NMR (CDCl3): δ 3.94 (s,
3H), 4.07 (s, 3H), 7.05 (s, 1H), 7.03-7.14 (m, 2H), 7.21-7.23
(m, 1H), 7.40-7.47 (m, 2H), 7.82-7.87 (m, 2H). EI-MS m/z:
314 (M+). IR (film): 3121, 1726, 1625, 1585, 1266, 1235 cm-1
.
4-[3-(2-Flu or oben zyl)ph en yl]-2,4-dioxobu tyr ic Acid (3).19
A solution of compound 9 in THF/CH3OH (1:1) (40.0 mL) was
stirred with 1 N NaOH (40.0 mL, 40.0 mmol) for 1 h at room
temperature and then extracted with ether. The water phase
was acidified with 2 N HCl to pH 1-2 and extracted with ethyl
acetate. The combined organic layers were washed with
NaHCO3 and brine and dried over Na2SO4, and the solvent
was removed under vacuum. The residue was recrystallized
from petroleum ether and dichloromethane to give compound
3 as a light yellow solid (1.79 g, two steps overall yield 53.35%).

Dimeric Diketoacid Containing Inhibitors of HIV-1 Integrase
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10 2571
petroleum ether to give R,R′-dibromo-o-xylene as a white solid
(1.527 g, yield 33.5%). 1H NMR (CDCl3): δ 7.39-7.36 (m, 2H);
7.33-7.30 (m, 2H); 4.67 (s, 4H).
δ 3.95 (s, 6H); 5.16 (s, 4H); 7.06 (s, 2H); 7.24-7.21 (m, 2H);
7.45-7.39 (m, 5H); 7.56 (s, 1H); 7.61-7.58 (m, 4H) (an enol
proton is not shown in the spectrum). 13C NMR (100 MHz,
CDCl3): 190.32 (×2), 168.55 (×2), 162.32 (×2), 158.75 (×2),
136.69 (×2), 136.10 (×2), 129.79 (×2), 128.87 (×2), 127.11 (×2),
126.33 (×2), 120.70, 120.62, 113.26 (×2), 98.23 (×2), 69.97
(×2), 53.26 (×2). Anal. (C30H26O10) C, H.
1-{3-[2-(3-Acet ylp h en oxym et h yl)b en zyloxy]p h en yl}-
eth a n on e (11a ). To the solution of 3-hydroxyacetophenone
(0.34 g, 2.5 mmol) in dry DMF (4 mL) were added potassium
carbonate (0.71 g, 5.14 mmol) and 10a (0.264 g, 1.0 mmol)
under N2. The mixture was stirred at 60 °C for 4 h. It was
diluted with water and the aqueous mixture extracted with
ethyl acetate. The combined organic layers were washed with
2N NaOH and brine and dried over Na2SO4. The solvent was
removed under reduced pressure. The residue was chromato-
graphed by PE/EtOAC (3:1) to give compound 11a as a white
4-(3-{3-[3-(3-Ca r boxy-3-oxop r op ion yl)p h en oxym eth yl]-
ben zyloxy}p h en yl)-2,4-d ioxobu tyr ic Acid (5b). Compound
5b was obtained from the hydrolysis of compound 12b with 1
N aqueous NaOH solution (3.2 mL, 3.2 mmol) in THF/CH3-
OH (1:1) (4.0 mL) analogously to compound 5a . Recrystalli-
zation from petroleum ether/dichloromethane afforded com-
pound 5b as a pale yellow solid (0.122 g, yield 73.5%). 1H NMR
(DMSO-d6): δ 7.68-7.63 (m, 4H); 7.51-7.48 (m, 6H); 7.36-
7.33 (dd, 2H); 7.09 (s, 2H); 5.23 (s, 4H). HR-ESI-MS calcd for
1
solid (0.34 g, yield 90.9%). H NMR (CDCl3): δ 7.56-7.53 (m,
6H); 7.42-7.34 (m, 4H); 7.17-7.15 (m, 2H), 5.23 (s, 4H); 2.56
(s, 6H). EI-MS (m/z, %): 374 (M+).
C
28H22O10 (M + Na)+: 541.111; found: 541.1108. Anal.
4-(3-{2-[3-(3-M e t h o x y c a r b o n y l-3-o x o p r o p i o n y l)-
ph en oxym eth yl]ben zyloxy}ph en yl)-2,4-dioxobu tyr ic Acid
Meth yl Ester (12a ). To a solution of CH3ONa (167 mg, 3.1
mmol) in 1.5 mL of dry toluene were added dropwise compound
11a (116 mg, 0.31 mmol) and tert-butyl methyl oxalate (198
mg, 1.24 mmol) in THF/DME (1.5 mL/1.5 mL) at 0 °C. After
being stirred at 0 °C for 0.5 h, the reaction mixture was poured
into HCl (1 N)/ice under vigorous stirring and extracted with
CH2Cl2. The combined organic phases were washed with 1 N
HCl brine and dried over Na2SO4, and the solvent was removed
under vacuum. The residue was purified by chromatography
using CHCl3/CH3OH (10:1) as eluent to give compound 12a
(0.158 g, yield 93.5%) as yellow solid. Mp 131-134 °C. 1H NMR
(CDCl3): δ 3.94 (s, 6H), 5.26 (s, 4H), 7.03 (s, 2H), 7.19 (dt, 2H,
J ) 8.4, 1.9 Hz), 7.36-7.43 (m, 4H), 7.52-7.56 (m, 4H), 7.58
(C28H22O10‚1.6H2O) C, H.
r,r′-Dibr om o-p-xylen e (10c). Bromination of p-xylene
(5.56 mL, 112.4 mmol) with NBS (20.1 g, 45.0 mmol) and
benzoxyl peroxide (18 mg) provided the desired R,R′-dibromo-
p-xylene 10c as a white solid (6.12 g, yield 51.0%) after
1
chromatography (petroleum ether). H NMR (CDCl3): δ 7.37
(m, 4H); 4.48 (s, 4H).
1-{3-[4-(3-Acet ylp h en oxym et h yl)b en zyloxy]p h en yl}-
eth a n on e (11c). Compound 11c was prepared as a white solid
in 36% yield according to the same procedure as 11a . 1H NMR
(CDCl3): δ 7.57-7.54 (m, 4H); 7.47 (s, 4H); 7.39-7.35 (t, 2H);
7.19-7.16 (m, 2H); 5.12 (s, 4H); 2.58 (s, 6H). EI-MS m/z: 374
(M+).
4-(3-{4-[3-(3-M e t h o x y c a r b o n y l-3-o x o p r o p i o n y l)-
ph en oxym eth yl]ben zyloxy}ph en yl)-2,4-dioxobu tyr ic Acid
Meth yl Ester (12c). Compound 12c was obtained from the
oxalylation of compound 11c in a fashion similar to the
preparation of 12a . Purification from chromatography afforded
pure 11c (0.414 g, yield 100%) as yellow oil. 1H NMR (CDCl3):
δ 7.67-7.62 (m, 4H); 7.51-7.47 (m, 6H); 7.36-7.34 (m, 2H);
7.08 (s, 2H); 5.23(s, 4H), 3.85 (s, 6H).
4-(3-{4-[3-(3-Ca r boxy-3-oxop r op ion yl)p h en oxym eth yl]-
ben zyloxy}p h en yl)-2,4-d ioxobu tyr ic Acid (5c). Treatment
of compound 12c (0.163 g, 0.298 mmol) with 1 N NaOH as
described above for the preparation of 5a provided the desired
product 5c as a pale yellow solid (0.107 g, yield 69.2%) after
recrystallization from petroleum ether and dichloromethane.
1H NMR (DMSO-d6): δ 7.66-7.61(m, 4H); 7.50-7.46 (m, 5H);
7.34-7.31(m, 3H); 7.08 (s, 2H); 5.21(s, 4H). HR-ESI-MS calcd
for C28H22O10(M + Na)+: 541.1111; found: 541.1116. Anal.
(C28H22O10‚1.5H2O) C, H.
(s, 2H), 15.19 (br, 1H). 13C NMR (100 MHz, DMSO-d6):
δ
190.09 (×2), 168.47 (×2), 162.02 (×2), 158.68 (×2), 135.78 (×2),
135.02 (×2), 130.31 (×2), 128.98 (×2), 128.18 (×2), 120.82 (×2),
113.45 (×2), 98.36 (×2), 67.67 (×2), 53.07 (×2). IR (film): 3111,
2952, 1734, 1594, 1436, 1283 cm-1
.
4-(3-{2-[3-(3-Ca r boxy-3-oxop r op ion yl)p h en oxym eth yl]-
ben zyloxy}p h en yl)-2,4-d ioxobu tyr ic Acid (5a ). A solution
of compound 12a (0.164 g, 0.30 mmol) in THF/CH3OH (1:1)
(4.0 mL) was treated with 1 N NaOH (3.0 mL, 3.0 mmol). The
reaction mixture was stirred for 1 h at room temperature and
then extracted with ether. The water phase was acidified with
2 N HCl to pH 1-2 and extracted with ethyl acetate. The
combined organic layers were washed with brine and dried
over Na2SO4, and the solvent was removed under vacuum. The
residue was recrystallized from petroleum ether/dichlo-
romethane to give compound 5a as a pale yellow solid (0.109
g, yield 70.0%). 1H NMR (CDCl3): δ 7.64-7.61 (m, 3H); 7.59-
7.55 (m, 3H); 7.49-7.45 (m, 2H); 7.41-7.39 (q, 2H, J ) 13.40);
7.35-7.32 (m, 2H); 7.06 (s, 2H), 5.36 (s, 4H). HR-ESI-MS calcd
for C28H22O10 (M + Na)+: 541.1111; found 541.1113. Anal.
(C28H22O10‚1.7H2O) C, H.
r,r′-Dibr om o-m -xylen e (10b). The treatment of m-xylene
(1.85 mL, 15.0 mmol) with benzoxyl peroxide (23 mg) and NBS
(6.675 g, 37.5 mmol) as described above for the preparation of
10a provided the desired product 10b as a white solid (1.878
g, yield 47.4%) after chromatography with petroleum ether.
1H NMR (CDCl3): δ 7.40 (s, 1H); 7.31 (m, 3H); 4.46 (s, 4H).
1-{3-[3-(3-Acet yl-p h en oxym et h yl)b en zyloxy]p h en yl}-
eth a n on e (11b). Compound 11b was prepared from the
coupling of 10b (0.396 g, 1.5 mmol) and 3-hydroxyacetophen-
one (0.51 g, 3.75 mmol) in the presence of potassium carbonate
(1.1 g, 7.6 mmol) analogously to compound 11a . Purification
by chromatography (PE/EtOAC ) 4:1) produced compound 11b
as a white solid (0.414 g, yield 74.0%). 1H NMR (CDCl3): δ
7.58-7.54 (m, 5H); 7.43-7.36 (m, 5H); 7.20-7.17 (m, 2H); 5.14
(s, 4H); 2.60 (s, 6H). EI-MS (m/z, %): 374 (M+).
(3-Br om op h en yl)(2-flu or op h en yl)m eth a n on e (13). To
a solution of compound 6 (3.838 g, 13.66 mmol) in acetone 40
mL at 0 °C was added J ones reagent (10.21 mL) dropwise with
stirring. The resulting orange solution was stirred under these
conditions for 25 min. The cooling bath was then removed, and
2-propanol 10 mL was added dropwise, whereupon a green
precipitate formed immediately. The mixture was stirred at
room temperature for 10 min and filtered through a short plug
of Celite. The flask and the Celite pad were washed with ether.
The organic layer was washed with saturated aqueous NaH-
CO3 and brine and dried over Na2SO4. The solvent was
removed under vacuum. The residue was purified by chroma-
tography using petroleum ether/ethyl acetate (8:1) as elutent
to afford compound 13 (3.757 g, yield 98.6%) as a white solid.
1H NMR (300 MHz, CDCl3): δ 7.98 (s, 1H); 7.97-7.71 (m, 2H);
7.60-7.53 (m, 2H); 7.38-7.29 (m, 2H); 7.21-7.15 (m, 1H). EI-
MS m/z: 278 (M+). IR (film): 1658, 1608, 1563, 1480, 1453,
1297 cm-1
.
Bis(3-br om op h en yl)(2-flu or op h en yl)m eth a n ol (14). To
an oven-dried round-bottom flask were added a solution of 1.6
M n-BuLi in hexanes (6.5 mL, 10.4 mmol) and 35 mL of dried
THF under N2. The stirred solution was cooled to -78 °C, and
to this solution was added 1,3-dibromobenzene (2.28 mL, 18.85
mmol) at such a rate as not allows the temperature to exceed
-70 °C. After the reaction was stirred 3.5 h at this tempera-
ture, compound 13 in 5 mL of THF was added slowly so that
the temperature remained bellowed -50 °C. After being stirred
4-(3-{3-[3-(3-M e t h o x y c a r b o n y l-3-o x o p r o p i o n y l)-
ph en oxym eth yl]ben zyloxy}ph en yl)-2,4-dioxobu tyr ic Acid
Meth yl Ester (12b). Oxalylation of compound 11b (238 mg,
0.636 mmol) with MeOOCCOOtBu (407 mg, 2.54 mmol) in the
presence of CH3ONa (343 mg, 6.36 mmol) were performed in
a fashion similar to the preparation of 12a to give compound
12b (0.343 g, yield 98.7%) as a yellow solid. 1H NMR (CDCl3):

2572 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10
Long et al.
at -50 °C for 0.5 h, the reaction mixture was allowed to warm
slowly to room temperature. Then the reaction mixture was
poured into saturated aqueous NH4Cl under stirring and
extracted with ether. The organic layer was washed with
saturated aqueous NaHCO3 and brine and dried over Na2SO4,
and the solvent was removed under vacuum. The residue was
purified by chromatography using petroleum ether/ethyl ac-
etate (40:1) as eluent to afford compound 14 (5.58 g, yield 95%)
as white glue. 1H NMR (300 MHz, CDCl3): δ 7.67-7.61 (m,
3H); 7.53-6.96 (m, 8H); 6.68-6.67 (m, 1H); 4.41 (s, 1H). EI-
MS m/z: 434 (M+). IR (film): 1658, 1608, 1563, 1480, 1453,
ether:dichloromethane to give compound 5d as a pale yellow
solid (0.086 g, 76% yield). H NMR (CDCl3): δ 7.92-7.80 (m,
4H); 7.49-7.45 (m, 3H); 7.35-7.35 (m, 5H); 7.12-6.91 (m, 3H);
7.00 (s, 1H); 6.09 (s, 1H). ESI-MS m/z: 489.1 (M+ - H). Anal.
(C27H19FO8‚1.6H2O) C, H.
1
Ack n ow led gm en t. We thank Paula Roberts, Quan-
En Yang, Omar Ragab, and Carney Chen for their
excellent technical assistance. The work in Y.-Q. L.’s
laboratory was supported by grants from Shanghai
Municipal Committee of Science and Technology, China
(02QB14056 and 03DZ19219) and the Chinese Academy
of Sciences (KSCX1-SW-11). The work in N.N.’s labora-
tory was supported by funds from the GlaxoSmithKline
Drug Discovery Award. The work in Screening Tech-
nologies Blanch was supported by funds from the NCl,
under contract number NO1-CO-12400.
1297 cm-1
.
Br om o-3-((3-b r om op h en yl)(2-flu or op h en yl)m et h yl)-
ben zen e (15). The compound 14 (5.58 g, 12.79 mmol) and
sodium idodide (7.67 g, 51.16 mmol) were stirred in CH3CN
(35 mL) and placed under N2. To the stirred mixture at room
temperature was added dichlorodimethylsilane (3.08 mL, 25.59
mmol) dropwise. The reaction mixture turns dark brown
almost immediately, and TLC confirms complete reduction.
The mixture was diluted with ethyl acetate and washed with
water, saturated aqueous NaHCO3, 10% aqueous sodium
thiosulfate, and brine. The solvent was removed under vacuum.
The residue was purified by chromatography using petroleum
ether as elutent to afford compound 15 (5.37 g, yield 100%) as
a white solid. 1H NMR (CDCl3): δ 7.39-7.36 (m, 2H); 7.35-
7.13 (m, 5H); 7.12-7.01 (m, 4H); 6.91-6.86 (m, 1H); 5.74 (s,
1H). EI-MS (m/z, %): 418 (M+).
Note Ad d ed a fter ASAP P ostin g. The version of
this paper posted April 10, 2004, was missing a double
bond in the general structure of Table 3. The keto enol
double bond was added to the structure in the corrected
version posted April 13, 2004.
1-{3-[(3-Acetylph en yl)(2-flu or oph en yl)m eth yl]ph en yl}-
eth a n on e (16). To a solution of compound 15 (1.049 g, 2.497
mmol) in THF (10 mL) was added dropwise of 1.6 M n-
butyllithium in hexanes (3.9 mL, 6.244 mmol) under N2 at -78
°C. The reaction mixture was stirred for 45 min at this
temperature, and then N-methoxy-N-methylacetamide (0.70
mL, 6.244 mmol) was added slowly so that the temperature
remained below -50 °C. After being stirred for 2 h at this
temperature, the reaction mixture was poured into aqueous
HCl (1 N)/ice under vigorous stirring and extracted with
ether.The combined organic phases were washed with brine
and dried over Na2SO4, and the solvent was removed under
vacuum. The residue was purified by chromatography using
petroleum ether/ethyl acetate (10:1) as eluent to give com-
pound 16 (0.522 g, yield 40.2%) as a white solid. 1H NMR (400
MHz, CDCl3): δ 7.86-7.83 (m, 2H); 7.74-7.73 (m, 2H); 7.43-
7.40 (t, 2H); 7.30-7.24 (m, 3H); 7.10-7.04 (m, 2H); 6.91-6.87
(td, 1H); 5.94(s, 1H); 2.55 (s, 6H). EI-MS m/z: 346 (M+).
Refer en ces
(1) Brown, P. O. Integration; Cold Spring Harbor Press: Cold Spring
Harbor, 1999.
(2) Asante-Appiah, E.; Skalka, A. M. HIV-1 integrase: structural
organization, conformational changes, and catalysis. Adv. Virus
Res. 1999, 52, 351-369.
(3) De Clercq, E. Strategies in the design of antiviral drugs. Nat.
Rev. Drug Discovery 2002, 1, 13-25.
(4) Richman, D. D. HIV chemotherapy. Nature 2001, 410, 995-
1001.
(5) Neamati, N. Patented small molecule inhibitors of HIV-1 inte-
grase: a ten-year saga. Expert Opin. Ther. Pat. 2002, 12, 709-
724.
(6) Neamati, N. Structure-based HIV-1 integrase inhibitor design:
a future perspective. Expert Opin. Invest. Drugs 2001, 10, 281-
296.
(7) Dayam, R.; Neamati, N. Small-Molecule HIV-1 Integrase Inhibi-
tors: the 2001-2002 Update. Curr. Pharm. Des. 2003, 9, 1789-
1802.
(8) Hazuda, D. J .; Felock, P.; Witmer, M.; Wolfe, A.; Stillmock, K.;
Grobler, J . A.; Espeseth, A.; Gabryelski, L.; Schleif, W.; Blau,
C.; Miller, M. D. Inhibitors of strand transfer that prevent
integration and inhibit HIV- 1 replication in cells. Science 2000,
287, 646-650.
(9) Yoshinaga, T.; Sato, A.; Fujishita, T.; Fujiwara, T. in Vitro
Activity of a New HIV-1 Integrase Inhibitor in Clinical Develop-
ment. 9th Conference on retroviruses and opportunistic infec-
tions: Seattle, 2002.
(10) Billich, A. S-1360 Shionogi-GlaxoSmithKline. Curr. Opin. In-
vestig. Drugs 2003, 4, 206-209.
(11) Este, J . A.; Cabrera, C.; Schols, D.; Cherepanov, P.; Gutierrez,
A.; Witvrouw, M.; Pannecouque, C.; Debyser, Z.; Rando, R. F.;
Clotet, B.; Desmyter, J .; De Clercq, E. Human immunodeficiency
virus glycoprotein gp120 as the primary target for the antiviral
action of AR177 (Zintevir). Mol. Pharmacol. 1998, 53, 340-345.
(12) Pluymers, W.; Neamati, N.; Pannecouque, C.; Fikkert, V.;
Marchand, C.; Burke, T. R., J r.; Pommier, Y.; Schols, D.; De
Clercq, E.; Debyser, Z.; Witvrouw, M. Viral entry as the primary
target for the anti-HIV activity of chicoric acid and its tetraacetyl
esters. Mol. Pharmacol. 2000, 58, 641-648.
(13) Espeseth, A. S.; Felock, P.; Wolfe, A.; Witmer, M.; Grobler, J .;
Anthony, N.; Egbertson, M.; Melamed, J . Y.; Young, S.; Hamill,
T.; Cole, J . L.; Hazuda, D. J . HIV-1 integrase inhibitors that
compete with the target DNA substrate define a unique strand
transfer conformation for integrase. Proc. Natl. Acad. Sci. U S
A 2000, 97, 11244-11249.
(14) Zhuang, L.; Wai, J . S.; Embrey, M. W.; Fisher, T. E.; Egbertson,
M. S.; Payne, L. S.; Guare, J . P., J r.; Vacca, J . P.; Hazuda, D. J .;
Felock, P. J .; Wolfe, A. L.; Stillmock, K. A.; Witmer, M. V.; Moyer,
G.; Schleif, W. A.; Gabryelski, L. J .; Leonard, Y. M.; Lynch, J .
J ., J r.; Michelson, S. R.; Young, S. D. Design and synthesis of
8-hydroxy-[1,6]naphthyridines as novel inhibitors of HIV-1
integrase in vitro and in infected cells. J . Med. Chem. 2003, 46,
453-456.
4-(3-{(2-F lu o r o p h e n y l)[3-(3-m e t h o x y c a r b o n y l-3-
oxop r op ion yl)p h en yl]m et h yl}p h en yl)-2,4-d ioxobu t yr ic
Acid Meth yl Ester (17). To an oven-dried round-bottom flask
was added CH3ONa (146 mg, 2.7 mmol) in toluene under N2.
The solution was cooled to 0 °C. To this solution were added
MeOOCCOOtBu (173 mg, 4.0 mmol) and compound 16 in dried
DME/THF (1:1) (2 mL). The solution was stirred at room
temperature for 2 h. The reaction mixture was quenched with
0.1 N HCl and extracted with EtOAc. The combined organic
layers were washed with saturated aqueous NaHCO3 and
brine and dried over Na2SO4, and the solvent was removed
under vacuum. The crude product was purified from column
chromatography (CHCl3:CH3OH ) 35:1) to give 119 mg of 17
1
in yield of 85%. Mp 56-58 °C. H NMR (CDCl3): δ 7.90-7.88
(d, 2H); 7.78 (s, 2H); 7.49-7.45 (m, 3H); 7.35-7.33 (m, 2H);
7.12-7.06 (m, 2H); 7.01 (s, 1H); 6.91-6.87 (td, 1H); 5.97 (s,
1H); 3.89 (s, 6H). IR (film): 2954, 1736, 1602, 1488, 1436, 1273
cm-1. EI-MS m/z: 518 (M+). HRMS calcd for C29H23FO8:
518.1337, found 518.1385.
4-{3-[[3-(3-Ca r b oxy-3-oxop r op ion yl)p h en yl](2-flu or o-
p h en yl)m eth yl]p h en yl}-2,4-d ioxobu tyr ic Acid (5d ). A
solution of compound 17 (119 mg) in THF/CH3OH (1:1) (4.0
mL) was treated with 1 N NaOH (5.0 mL, 5.0 mmol). The
reaction mixture was stirred for 1 h at room temperature then
extracted with ether. The water phase was acidified with 2 N
HCl to pH 1-2 and extracted with ethyl acetate. The combined
organic layers were washed with saturated aqueous NaHCO3
and brine and dried over Na2SO4, and the solvent was removed
under vacuum. The residue was recrystallized from petroleum

Dimeric Diketoacid Containing Inhibitors of HIV-1 Integrase
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 10 2573
(15) Pais, G. C.; Zhang, X.; Marchand, C.; Neamati, N.; Cowansage,
K.; Svarovskaia, E. S.; Pathak, V. K.; Tang, Y.; Nicklaus, M.;
Pommier, Y.; Burke, T. R., J r. Structure Activity of 3-Aryl-1,3-
diketo-Containing Compounds as HIV-1 Integrase Inhibitors. J .
Med. Chem. 2002, 45, 3184-3194.
(16) Marchand, C.; Zhang, X.; Pais, G. C.; Cowansage, K.; Neamati,
N.; Burke, T. R., J r.; Pommier, Y. Structural determinants for
HIV-1 integrase inhibition by beta-diketo acids. J . Biol. Chem.
2002, 277, 12596-12603.
(17) Pluymers, W.; Pais, G.; Van Maele, B.; Pannecouque, C.; Fikkert,
V.; Burke, T. R., J r.; De Clercq, E.; Witvrouw, M.; Neamati, N.;
Debyser, Z. Inhibition of Human Immunodeficiency Virus Type
1 Integration by Diketo Derivatives. Antimicrob Agents Chemoth-
er. 2002, 46, 3292-3297.
(18) Pais, G. C. G.; Burke, T. R. Novel aryl diketo-containing
inhibitors of HIV-1 integrase. Drugs Future 2002, 27, 1101-
1111.
(32) Steitz, T. A. DNA polymerases: structural diversity and common
mechanisms. J . Biol. Chem. 1999, 274, 17395-17398.
(33) Steitz, T. A. A mechanism for all polymerases. Nature 1998, 391,
231-232.
(34) Lins, R. D.; Adesokan, A.; Soares, T. A.; Briggs, J . M. Investiga-
tions on human immunodeficiency virus type 1 integrase/DNA
binding interactions via molecular dynamics and electrostatics
calculations. Pharmacol. Ther. 2000, 85, 123-131.
(35) Bujacz, G.; Alexandratos, J .; Qing, Z. L.; Clement-Mella, C.;
Wlodawer, A. The catalytic domain of human immunodeficiency
virus integrase: ordered active site in the F185H mutant. FEBS
Lett. 1996, 398, 175-178.
(36) 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.
(19) Wai, J . S.; Egbertson, M. S.; Payne, L. S.; Fisher, T. E.; Embrey,
M. W.; Tran, L. O.; Melamed, J . Y.; Langford, H. M.; Guare, J .
P., J r.; Zhuang, L.; Grey, V. E.; Vacca, J . P.; Holloway, M. K.;
Naylor-Olsen, A. M.; Hazuda, D. J .; Felock, P. J .; Wolfe, A. L.;
Stillmock, K. A.; Schleif, W. A.; Gabryelski, L. J .; Young, S. D.
4-Aryl-2,4-dioxobutanoic acid inhibitors of HIV-1 integrase and
viral replication in cells. J . Med. Chem. 2000, 43, 4923-
4926.
(37) Chen, J . C.; Krucinski, J .; Miercke, L. J .; Finer-Moore, J . S.;
Tang, A. H.; Leavitt, A. D.; Stroud, R. M. Crystal structure of
the HIV-1 integrase catalytic core and C-terminal domains: a
model for viral DNA binding. Proc. Natl. Acad. Sci. U.S.A. 2000,
97, 8233-8238.
(38) Goldgur, Y.; Dyda, F.; Hickman, A. B.; J enkins, T. M.; Craigie,
R.; Davies, D. R. Three new structures of the core domain of
HIV-1 integrase: an active site that binds magnesium. Proc.
Natl. Acad. Sci. U.S.A. 1998, 95, 9150-9154.
(20) Esposito, D.; Craigie, R. HIV integrase structure and function.
Adv. Virus Res. 1999, 52, 319-333.
(39) Greenwald, J .; Le, V.; Butler, S. L.; Bushman, F. D.; Choe, S.
The mobility of an HIV-1 integrase active site loop is correlated
with catalytic activity. Biochemistry 1999, 38, 8892-8898.
(40) Maignan, S.; Guilloteau, J . P.; Zhou-Liu, Q.; Clement-Mella, C.;
Mikol, V. Crystal structures of the catalytic domain of HIV-1
integrase free and complexed with its metal cofactor: high level
of similarity of the active site with other viral integrases. J . Mol.
Biol. 1998, 282, 359-368.
(41) Smonou, I. One step reduction of diaryl ketones to hydrocarbons
by etherated boron trifluoride-triethylsilane system. Synth.
Commun. 1994, 24, 1999-2002.
(42) Nahm, S.; Weinreb, S. M. N-Methoxy-N-methylamides as effec-
tive acylating agents. Tetrahedron Lett. 1981, 22, 3815-3818.
(43) Bachmann, W. E.; Cole, W.; Wilds, A. L. The total synthesis of
the sex hormone Equilenin and its stereoisomers. J . Am. Chem.
Soc. 1940, 62, 824-839.
(44) J iang, X. H.; Song, L. D.; Long, Y. Q. Highly Efficient Preparation
of Aryl beta-Diketo Acids with tert-Butyl Methyl Oxalate. J . Org.
Chem. 2003, 68, 7555-7558.
(45) Sotriffer, C. A.; Ni, H.; McCammon, A. J . HIV-1 Integrase
inhibitor interactions at the active site: prediction of binding
modes unaffected by crystal packing. J . Am. Chem. Soc 2000,
122, 6136-6137.
(46) Barreca, M. L.; Lee, K. W.; Chimirri, A.; Briggs, J . M. Molecular
Dynamics Studies of the Wild-Type and Double Mutant HIV-1
Integrase Complexed with the 5CITEP Inhibitor: Mechanism
for Inhibition and Drug Resistance. Biophys. J . 2003, 84, 1450-
1463.
(47) Weislow, O. W.; Kiser, R.; Fine, D.; Bader, J .; Shoemaker, R.
H.; Boyd, M. R. New soluble-formazan assay for HIV-1 cytopathic
effects: application to high-flux screening of synthetic and
natural products for AIDS antiviral activity. J . Natl. Cancer Inst.
1989, 81, 577-586.
(21) Goldgur, Y.; Craigie, R.; Cohen, G. H.; Fujiwara, T.; Yoshinaga,
T.; Fujishita, T.; Sugimoto, H.; Endo, T.; Murai, H.; Davies, D.
R. Structure of the HIV-1 integrase catalytic domain complexed
with an inhibitor: a platform for antiviral drug design. Proc.
Natl. Acad Sci. U S A 1999, 96, 13040-13043.
(22) Sherman, P. A.; Fyfe, J . A. Human immunodeficiency virus
integration protein expressed in Escherichia coli possesses
selective DNA cleaving activity. Proc. Natl. Acad. Sci. U S A
1990, 87, 5119-5123.
(23) J ones, K. S.; Coleman, J .; Merkel, G. W.; Laue, T. M.; Skalka,
A. M. Retroviral integrase functions as a multimer and can turn
over catalytically. J . Biol. Chem. 1992, 267, 16037-16040.
(24) Vincent, K. A.; Ellison, V.; Chow, S. A.; Brown, P. O. Charac-
terization of human immunodeficiency virus type 1 integrase
expressed in Escherichia coli and analysis of variants with
amino- terminal mutations. J . Virol. 1993, 67, 425-437.
(25) J enkins, T. M.; Engelman, A.; Ghirlando, R.; Craigie, R. A soluble
active mutant of HIV-1 integrase: involvement of both the core
and carboxyl-terminal domains in multimerization. J . Biol.
Chem. 1996, 271, 7712-7718.
(26) Heuer, T. S.; Brown, P. O. Photocross-linking studies suggest a
model for the architecture of an active human immunodeficiency
virus type 1 integrase-DNA complex. Biochemistry 1998, 37,
6667-6678.
(27) Wang, J . Y.; Ling, H.; Yang, W.; Craigie, R. Structure of a two-
domain fragment of HIV-1 integrase: implications for domain
organization in the intact protein. EMBO J . 2001, 20, 7333-
7343.
(28) De Luca, L.; Pedretti, A.; Vistoli, G.; Barreca, M. L.; Villa, L.;
Monforte, P.; Chimirri, A. Analysis of the full-length integrase-
DNA complex by a modified approach for DNA docking. Biochem.
Biophys. Res. Commun. 2003, 310, 1083-1088.
(29) Podtelezhnikov, A. A.; Gao, K.; Bushman, F. D.; McCammon, J .
A. Modeling HIV-1 integrase complexes based on their hydro-
dynamic properties. Biopolymers 2003, 68, 110-120.
(30) Cowan, J . A. The Biological Chemistry of Magnesium; VCH:
1995; pp 1-23.
(48) Smellie, A.; Teig, S. L.; Towbin, P. Poling - Promoting Confor-
mational Variation. J . Comput. Chem. 1995, 16, 171-187.
(49) CCDC, C. C. E., UK GOLD; 1.2 ed.
(31) Neamati, N.; Lin, Z.; Karki, R. G.; Orr, A.; Cowansage, K.;
Strumberg, D.; Pais, G. C.; Voigt, J . H.; Nicklaus, M. C.;
Winslow, H. E.; Zhao, H.; Turpin, J . A.; Yi, J .; Skalka, A. M.;
Burke, T. R., J r.; Pommier, Y. Metal-dependent inhibition of
HIV-1 integrase. J . Med. Chem. 2002, 45, 5661-5670.
J M030559K