Computational and synthetic approaches for developing Lavendustin B derivatives as allosteric inhibitors of HIV-1 integrase

Article abstract of DOI:10.1016/j.ejmech.2016.07.077

Through structure-based virtual screening and subsequent activity assays of selected natural products, Lavendustin B was previously identified as an inhibitor of HIV-1 integrase (IN) interaction with its cognate cellular cofactor, lens epithelium-derived growth factor (LEDGF/p75). In order to improve the inhibitory potency we have employed in silico-based approaches. Particularly, a series of new analogues was designed and docked into the LEDGF/p75 binding pocket of HIV-1 IN. To identify promising leads we used the Molecular Mechanics energies combined with the Generalized Born and Surface Area continuum solvation (MM-GBSA) method, molecular dynamics simulations and analysis of hydrogen bond occupancies. On the basis of these studies, six analogues of Lavendustine B, containing the benzylamino-hydroxybenzoic scaffold, were selected for synthesis and structure activity-relationship (SAR) studies. Our results demonstrated a good correlation between computational and experimental data, and all six analogues displayed an improved potency for inhibiting IN binding to LEDGF/p75 in?vitro to respect to the parent compound Lavendustin B. Additionally, these analogs show to inhibit weakly LEDGF/p75-independent IN catalytic activity suggesting a multimodal allosteric mechanism of action. Nevertheless, for the synthesized compounds similar profiles for HIV-1 inhibition and cytoxicity were highlighted. Taken together, our studies elucidated the mode of action of Lavendustin B analogs and provided a path for their further development as a new promising class of HIV-1 integrase inhibitors.

Full text of DOI:10.1016/j.ejmech.2016.07.077

Accepted Manuscript
Computational and synthetic approaches for developing Lavendustin B derivatives as
allosteric inhibitors of HIV-1 integrase
Fatima E. Agharbaoui, Ashley C. Hoyte, Stefania Ferro, Rosaria Gitto, Maria Rosa
Buemi, James R. Fuchs, Mamuka Kvaratskhelia, Laura De Luca
PII:
S0223-5234(16)30642-0
10.1016/j.ejmech.2016.07.077
EJMECH 8795
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 26 May 2016
Revised Date: 25 July 2016
Accepted Date: 31 July 2016
Please cite this article as: F.E. Agharbaoui, A.C. Hoyte, S. Ferro, R. Gitto, M.R. Buemi, J.R. Fuchs,
M. Kvaratskhelia, L. De Luca, Computational and synthetic approaches for developing Lavendustin B
derivatives as allosteric inhibitors of HIV-1 integrase, European Journal of Medicinal Chemistry (2016),
doi: 10.1016/j.ejmech.2016.07.077.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
GRAPHICAL ABSTRACT
Computational and synthetic approaches for developing Lavendustin B
derivatives as allosteric inhibitors of HIV-1 integrase
Fatima E. Agharbaouia,b,c, Ashley C. Hoyteb, Stefania Ferroa, Rosaria Gittoa, Maria Rosa Buemia,
James R. Fuchsc, Mamuka Kvaratskheliab, Laura De Lucaa,*
A computational approach applying docking, rescoring, ultra short MD and hydrogen bonds
analysis has been applied for the design and the optimization of Lavendustin B derivatives as IN-
LEDGF/p75 inhibitors. The selected derivatives were then synthetized and evaluated using HTRF
assays.

ACCEPTED MANUSCRIPT
Computational and synthetic approaches for developing Lavendustin B derivatives as
allosteric inhibitors of HIV-1 integrase
Fatima E. Agharbaouia,b,c, Ashley C. Hoyteb, Stefania Ferroa, Rosaria Gittoa, Maria Rosa Buemia,
James R. Fuchsc, Mamuka Kvaratskheliab, Laura De Lucaa,*
aDipartimento di Scienze Chimiche, Biologiche, Farmaceutiche e Ambientali (CHIBIOFARAM)
Polo Universitario SS. Annunziata, Università di Messina, Viale Annunziata I-98168 Messina,
b
Italy; Center for Retrovirus Research and College of Pharmacy, The Ohio State University,
C
Columbus, OH 43210, USA; Division of Medicinal Chemistry and Pharmacognosy, College of
Pharmacy, The Ohio State University, Columbus, OH, 43210, USA.
Abstract
Through structure-based virtual screening and subsequent activity assays of selected natural
products, Lavendustin B was previously identified as an inhibitor of HIV-1 integrase (IN)
interaction with its cognate cellular cofactor, lens epithelium-derived growth factor (LEDGF/p75).
In order to improve the inhibitory potency we have employed in silico-based approaches.
Particularly, a series of new analogues was designed and docked into the LEDGF/p75 binding
pocket of HIV-1 IN. To identify promising leads we used the Molecular Mechanics energies
combined with the Generalized Born and Surface Area continuum solvation (MM-GBSA) method,
molecular dynamics simulations and analysis of hydrogen bond occupancies. On the basis of these
studies, six analogues of Lavendustine B, containing the benzylamino-hydroxybenzoic scaffold,
were selected for synthesis and structure activity-relationship (SAR) studies. Our results
demonstrated a good correlation between computational and experimental data, and all six
analogues displayed an improved potency for inhibiting IN binding to LEDGF/p75 in vitro to
respect to the parent compound Lavendustin B. Additionally, these analogs show to inhibit weakly
LEDGF/p75-independent IN catalytic activity suggesting a multimodal allosteric mechanism of
action. Nevertheless, for the synthesized compounds similar profiles for HIV-1 inhibition and
cytoxicity were highlighted. Taken together, our studies elucidated the mode of action of
Lavendustin B analogs and provided a path for their further development as a new promising class
of HIV-1 integrase inhibitors.
Keywords
HIV-1; Integrase; LEDGF/p75; ALLINI; computational studies; IN-LEDGF/p75 binding.
Abbreviations
1

ACCEPTED MANUSCRIPT
allosteric integrase inhibitor (ALLINI); catalytic core domain (CCD); C-terminal domain (CTD)
IN-binding domain (IBD); Integrase (IN); Integrase lens epithelium-derived growth factor
Allosteric Inhibitors (INLAIs); viral DNA (vDNA); Genetic Optimization for Ligand Docking
(GOLD); lens epithelium-derived growth factor (LEDGF/p75); lens epithelium-derived growth
factor Integrase Inhibitor (LEDGINs); molecular dynamic (MD); Molecular Mechanics energies
combined with the Generalized Born and Surface Area continuum solvation (MM-GBSA),
Multimerization Integrase Inhibitors (MINIs); Natural Products (NPs); N-terminal domain (NTD),
Non Catalytic Integrase Inhibitors (NCINIs); protein-protein interaction (PPI); protein-protein
interaction inhibitors (PPIIs); 3’processing (3’P); strand-transfer (ST); vesicular stomatitis virus g
(VSV-g).
*Corresponding author. E-mail address: ldeluca@unime.it; fagharbaoui@unime.it
1. Introduction
An essential step of the retroviral lifecycle is the insertion of the reverse-transcribed viral genome
into the host chromosome. This process is catalyzed by HIV-1 integrase (IN), that has gained
popularity as a promising target for the discovery of novel anti-HIV drugs.
IN is comprised of three domains: the N-terminal domain (NTD), the catalytic core domain (CCD)
that coordinates two catalytic Mg2+ ions and the C-terminal domain (CTD) [1, 2]. Initial drug
discovery efforts for IN inhibitors have focused on small molecules able to inhibit the catalytic
activity of IN and have resulted in three FDA approved IN inhibitors currently in clinical use,
raltegravir, elvitegravir and dolutegravir [3-8]. These compounds all share a similar mechanism of
action in that all bind at the IN active site in the presence of the viral DNA and inhibit the strand
transfer (ST) activity. While these inhibitors have been highly effective against HIV, resistance
mutations have emerged in patients [9-11]. Therefore, there is a continual need for the development
of new IN inhibitors with innovative scaffolds that target alternative sites of the enzyme.
The integration process comprises two catalytic steps: the first is a hydrolytic reaction termed
3’processing (3’ P), followed by a transesterification reaction (also reffered as to ST) [12-14]. The
cellular chromatin associated protein lens epithelium-derived growth factor (LEDGF/p75)
associates with IN and significantly enhances integration efficacy by tethering preintegration
complexes to active genes during integration [15-19]. LEDGF/p75 is a transcriptional co-activator
strongly associated with chromatin throughout the cell cycle [20-22]. Its C-terminal domain
contains the IN-binding domain (IBD), allowing it to not only interact with natural cellular binding
partners, but also HIV-1 IN [18, 19, 23, 24].
Recent efforts have led to the discovery of a new class of allosteric IN inhibitors (ALLINIs, also
known as LEDGINs, NCINIs, INLAIs, or MINIs) [25-30] targeting the IN dimer interface at the
2

ACCEPTED MANUSCRIPT
principal LEDGF/p75 binding pocket. Interestingly, two alternative approaches have identified a
similar class of quinoline-based ALLINIs: a high throughput screen was used to discover
compounds inhibiting 3’-processing activity of IN [26] and the rational drug design was exploited
to develop small molecules to block the IN-LEDGF/p75 interaction [25]. Of note, the rational
design approach was made possible by the crystal structure of a CCD-CCD dimer bound to the IBD
[31]. Furthermore, the ability to solve the structures of ALLINIs bound at the CCD-CCD dimer [25-
28] has facilitated the rapid expansion of this class of inhibitors.
ALLINIs exhibit a multimodal mechanism of action in that they not only inhibit IN-LEDGF/p75
interaction but they also promote higher order aberrant IN multimerization, resulting in inactive
protein [2, 32, 33]. Surprisingly, ALLINIs exhibit higher potency when present during virion
morphogenesis compared with the early stage of viral replication [29, 34-37]. In the virions, where
due to the lack of competing LEDGF/p75 binding to the IN dimer, ALLINIs potently induce
aberrant IN multimerization and result in eccentric, non-infectious virions; whereas during the early
stage of HIV-1 replication LEDGF/p75 effectively competes with ALLINI binding to IN and
reduces the inhibitor potency [38]. Selection of HIV-1 phenotypes under the genetic pressure of
various ALLINIs have identified substitutions at the IN dimer interface at the inhibitor binding sites
that confer resistance to these compounds [25, 28, 39, 40]. Collectively, the studies with ALLINIs
have shown that the potent inhibitors that target IN sites distinct from the active site can be
developed. At the same time, there is a need to further improve these compounds to overcome the
resistance seen in cell culture assays.
Natural Products (NPs) have historically been an extraordinary source for new medicines and are
continuing to be the origin of lead compounds for drug discovery [41, 42]. Previously, we have
reported the application of a structure-based virtual screening strategy for the identification of NPs
as potential protein-protein interaction inhibitors (PPIIs) targeting the IN-LEDGF/p75 protein
complex [43]. Among them, we focused our interest on the Lavendustin B (Figure 1A), which
inhibited IN binding to LEDGF/p75 in Alphascreen assay [43]. This novel scaffold is unique from
all reported ALLINIs and could represent an encouraging new hit compound warranting further
improvement and investigation. Therefore, to exploit this novel scaffold and improve its potency as
an IN inhibitor, we have employed in silico approaches to identify promising Lavendustin B
derivatives and examine inhibitory activities using in vitro and cell based assays.
2. Results and discussion
2.1 Rational design
3

ACCEPTED MANUSCRIPT
By using a combination of docking and ultrashort molecular dynamics (MD), we have generated a
weighted ensemble of protein-ligand configurations for IN-LEDGF/p75 protein-protein interaction
inhibitors. Therefore we estimated their binding affinities by averaging snapshots taken from the
MD trajectories, together with the presence of fundamental hydrogen bonds [44]. These in silico
studies, followed by experimental analysis of selected compounds, have led to the identification of
Lavendustin B with an IC50 of 94.07 µM for inhibiting IN binding to LEDGF/p75 in vitro [43]. In
silico docking studies (Figure 1 B) have highlighted the following interactions: the carboxylic group
of Lavendustin B establishes H-bond interactions with the backbone nitrogen atoms of Glu170 and
His171 residues, similar to the interactions seen with LEDGF/p75 hotspot residue of Asp366 [24].
Additionally, there is the formation of a potential hydrogen bond with the hydroxyl group side chain
of the Thr174 residue. The remaining portion of Lavendustin B is housed within the dimer interface
cleft comprised of IN subunit A residues of Thr174, Gln168, Ala169 and Met178 and IN subunit B
residues of Ala128, Ala129, Trp131 and Trp132 allowing the molecule to establish hydrophobic
contacts with both subunits. We have used these in silico predicated interactions as the basis for our
current study.
In order to design new analogs, we utilized a published X-ray crystal structure of the active
compound KF115 (PDB-400J) which shares a similar binding mode with Lavendustin B (Figure
1A and C) [28]. KF115 is a pyridine-based inhibitor in the class of ALLINIs that preferentially
promoted aberrant IN multimerization over inhibiting the IN-LEDGF/p75 interaction. The
superimposition of the crystal structure of KF115 bound to the CCD-CCD dimer with the docked
model of Lavendustin B (Figure 1B) reveals a high degree of similarity with the carboxylic groups
of the both compounds interacting with Glu170, His171 and Thr174. In addition, the 4-
chlorophenyl and the 3,4-dimethylphenyl groups of KF115 occupy the hydrophobic pockets in a
similar manner of the two 2-hydroxyphenyl portions of Lavendustin B. Considering these results,
structural modifications on Lavendustin B were introduced in silico: the 2-hydroxy group was
removed, and halogen atoms (chlorine and fluorine) and methyl substituent were added to explore
the hydrophobic areas. The planned modifications on Lavendustin B are depicted in Figure 1 D for
compounds (1-10).
Figure 1
2.2 Docking and molecular dynamics (MD)
4

ACCEPTED MANUSCRIPT
Before carrying out the synthesis of the designed compounds (1-10) we wanted to predict the
potential binding mode of the analogs by means of the reported computational procedure [43, 45].
First a docking simulation into the principal LEDGF/p75 binding pocket on IN [24] was performed
using GOLD (Genetic Optimization for Ligand Docking) [46]. In order to take into account the
flexible side chain of residue Gln95 two different conformations of IN CCD were used (PDB ID:
3LPU [25] and 2B4J [32]). More than two clusters were taken for additional analysis. To eliminate
potentially unfavorable contacts, the geometry of the system was minimized using the steepest
descent algorithm followed by a conjugate gradient. The solvent effects were considered through
the generalized Born implicit solvent model. The output complex was employed to estimate ligand
binding free energy using the MM-GBSA method. The obtained results for both CCD
conformations (complex 1 and 2) are shown in Table 1.
Table 1
Since the calculated binding energy of the complex 1 and 2 were similar we decided to consider
only the conformation of the protein retrieved by 2B4J (complex 1) for further or more complete
analysis. Figure 2 shows the binding orientations of the designed analogs. We observed that six
compounds, namely (1-3, 5, 6 and 8), share the binding mode with parent Lavendustin B. The other
derivatives (4, 7, 9 and 10) assume a binding pose for which they mimic the carboxylic
functionality of Lavendustin B but adopt a different orientation for the aromatic portion. Key
binding interactions for compound (2), which has been selected for subsequent studies, are
highlighted in Supplemental Figure 1.
Figure 2
The docked complexes were further analyzed using both Ultrashort Molecular Dynamics
simulations and sander module of AMBER 11 [47]. Additionally, the models were used to estimate
the binding affinities by averaging snapshots taken from the MD trajectories using the MM-GBSA
method (Table 1).
By comparing the binding energies compound (2) was predicted as the most potent derivative of the
series, followed by derivatives (1) and (5). By contrast, compounds (4, 7, 9 and 10), which were
predicted to bind differently than KF115 or Lavendustin B, displayed the weakest binding energies.
Therefore, these compounds were predicted to be less active. To explore this hypothesis, analysis of
5

ACCEPTED MANUSCRIPT
hydrogen bond occupancies and distance calculations from the hot spot amino acids in the
hydrophobic pocket were carried out using Ptraj module from AMBER11 [47]
Table 2
The obtained results (Table 2) show that (1) and (2) established an extra hydrogen bond interaction
with Gln95 residue. Moreover, the carboxylic acid is closer to the hot spot amino acids and has the
stronger hydrogen bond occupancies in comparison with Lavendustin B. In contrast, compounds (4,
7, 9 and 10) showed the weaker hydrogen bonding abilities, especially with residue His171.
Collectively our in silico approach, in combination with our previous studies [48], suggested the
following criteria for the synthesis. The analogs need to bind to the CCD-CCD dimer in a similar
mode to Lavendustin B and KF115. Furthermore, the derivatives are expected to exhibit improved
binding energy and higher hydrogen bond occupancies in the binding pocket than parent
Lavendustin B. This criterion allowed us select compounds (1-3, 5, 6 and 8) for chemical synthesis
and in vitro evaluation.
2.3 Chemistry
The picked 5-(dibenzylamino)-2-hydroxybenzoic acids (1-3, 5, 6 and 8) were synthesized following
the multistep procedure depicted in scheme 1.
Scheme 1
Commercially available 2-hydroxy-5-nitro-benzoic acid (11) was reduced, using zinc dust in acid
medium, to form the corresponding 5-amino-2-hydroxy-benzoic acid (12) with high yield and
successively converted into the 5-amino-2-hydroxybenzoate (13) by esterification reaction.
In the next step, methyl 5-(benzylamino)-2-hydroxybenzoates (14-19) were obtained by reaction of
intermediate 13 with the suitable benzaldehyde, by means of a reductive amination, in the presence
of sodium cyanoborohydride.
For the synthesis of 5-(dibenzylamino)-2-hydroxybenzoates (20-25) the obtained intermediates (14-
19) were treated in dimethylformamide with the appropriate benzyl bromide and sodium hydride,
under argon atmosphere. Finally, the target compounds (1–3, 5, 6, and 8) were prepared by
hydrolysis in basic medium.
6

ACCEPTED MANUSCRIPT
2.4 In vitro screening of synthesized compounds 1-3, 5, 6 and 8
All the synthetized derivatives were tested in homogeneous time-resolved fluorescence (HTRF)-
based assays to evaluate their inhibitory effects on the IN-LEDGF/p75 binding, LEDGF/p75
dependent integration, and LEDGF/p75 independent 3’-processing activity (Table 3).
Table 3
As result all the target compounds were able to inhibit LEDGF/p75 dependent IN activity
displaying IC50 values ranging from 3.78 µM to 18.50 µM (Table 3 and Figure 3B) with (2) being
the most potent in the series. Since this assay does not delineate whether the inhibitor affected IN
binding to LEDGF/p75 or impaired IN activity in a LEDGF/p75 independent manner we conducted
additional assays as follow. We firstly tested our analogs for their ability to inhibit IN-LEDGF/p75
binding (Table 3 and Fig 3C). All the tested compounds were able to disrupt IN binding to
LEDGF/p75 with IC50 values ranging from 3.28 µM to 27.59 µM and among them compound (2)
showed the best activity. In addition, we also analyzed the analogs for their ability to inhibit IN
3’processing activity in the absence of LEDGF/p75. The obtained date showed a weak inhibitions
for compounds 1,2,5 and 6. The results, summarized in Table 3, show that our compounds were
able to impair IN activity. Specifically, derivative (2) again displayed the best IC50 value (25.67
µM), Figure 3D. Collectively, these results lead us to think that Lavendustin B and its analog 5-
[bis(2-chlorobenzyl)amino]-2-hydroxybenzoic acid (2) can share a common mechanism of action
with pyridine- or quinoline-based ALLINIs. For example, similarly to reported ALLINIs [26-29,
32, 33] compound (2) inhibits both LEDGF/p75 dependent and independent activities of IN
indicating a multimodal mechanism of action. At the same time we note the following differences.
The quinoline-based ALLINIs inhibit IN-LEDGF/p75 binding and IN activity in LEDGF/p75
independent reactions with comparable IC50 values [2, 27, 32, 33]. Pyridine-based ALLINIs are
significantly more potent (up to 58-fold) for inducing abberent IN multimerization in the absence of
LEDGF/p75 compared with inhibition of IN-LEDGF/p75 binding [28]. In contrast, compound (2)
was ~8-fold more potent for inhibiting IN-LEDGF/p75 interactions compared with affecting IN
activity in the LEDGF/p75 independnet manner. Future studies are worrented to clarify the
observed differences between these structurally distinct classes of inhibitors.
7

ACCEPTED MANUSCRIPT
Figure 3
2.5 Antiviral and cytotoxicity assays
Antiviral activities and cellular toxicity of the most potent analog (2) were examinated founding
similar profiles for HIV-1 inhibition and cytoxicity (see Figure 4). Unfortunately, these findings do
not allow us to delineate any potential antiviral activity of 2 from its cellular toxicity.
Figure 4
3. Conclusions
By means of a in silico methodology a new series of Lavendustin B analogs were designed.
The selected 5-(dibenzylamino)-2-hydroxybenzoic acids (1-3, 5, 6 and 8) inhibited IN-LEDGF/p75
binding as well as impaired IN activity through allosteric mechanisms, and their improved
inhibitory potency has been confirmed using a combination of HTRF-based assays.Additionally, the
most active derivative (2) exhibited a multimodal mechanism of action, similar to the reported
ALLINIs. Unfortunately, compound (2) displayed significant cellular toxicity precluding our efforts
to delinate its potential antiviral activity.
4. Experimental section
4.1 Docking simulation
3D structure of each ligand was constructed using Discovery Studio 2.5.5 [49] and minimized using
CHARMm force field, followed by Smart Minimizer algorithm performing 1000 steps of Steepest
Descent with a root mean square (RMS) gradient tolerance of 3, followed by Conjugate Gradient
minimization, until the RMS gradient for potential energy was less than 0.05 kcal ꢀmolꢀÅ. For
docking simulations the crystal structure of the dimeric Catalytic Core Domain (CCD) of HIV-1 IN
complexed with the Integrase Binding Domain (IBD) of LEDGF/p75 was retrieved from RCSB
Protein Data Bank (PDB:2B4J) [31]. The LEDGF/p75 structure and the water molecules from the
X-ray crystallography were removed and the missing hydrogens were replaced. Validation of the
docking protocol was performed by docking the native co-crystallized ligands of the two crystal
structures with the PDB codes 3LPT and 3LPU, into LEDGF/p75 binding site. The comparison of
docking results with the co-crystallized form showed success rates with the docked ligand strictly
8

ACCEPTED MANUSCRIPT
superimposed with the crystallized conformation with RMSD = 1.01 Å indicating that the used
scoring function is successful. These values were small enough and supported the hypothesis that
experimental binding modes could be reproduced with accuracy using this protocol. The standard
default settings were used in all calculations. Docking studies were performed using the genetic
optimization for ligand docking (GOLD) software package version 4.1.1 from the Cambridge
Crystallographic Data Centre (CCDC) [46] and as described in our previous paper [48]. For the
prediction of ligand binding positions GoldScore fitness function was used. For each ligand 100
independent runs and a maximum of 15000 genetic operations were performed using the default
operator weights and a population size of 100 chromosomes. Default cutoff values of 2.5 Å for
hydrogen bonds and 4.0 Å for van der Waals interactions were employed. Automatic bond settings
were used, allowing the torsion angles of all acyclic, rotatable bonds in the ligand to vary except for
amide bonds. Results differing by less than 0.75 Å in ligand-all-atom RMSD were clustered
together. Results differing by less than 1.00 Å in ligand-all atom RMSD were clustered together. A
20.0 Å radius active site was drawn on the original position of the LEDGF/p75 IBD dipeptide
Ile365-Asp366 and automated cavity detection was used. Two hydrogen bond constraints were used
to specify that two protein atoms should be hydrogen-bonded to the ligand, namely NH backbone of
Glu170 and His171 with a constraint weight of 5. Binding energy of the minimized complex was
calculated using the MM-GBSA method [50] implemented in the AMBER program.
4.2 Molecular dynamic simulation
The starting model for simulations of IN-LEDGF/p75 was prepared as described in our previous
paper [48]. In brief, from the X-ray structure 2B4J of IN CCD (chains A and B) in complex with the
IBD of LEDGF (chains C and D) [31] was used. First, chain D and water molecules were removed
from the structure. Then, the missing residues of the CCD of IN were added by superimposing
chain C of the HIV-1 IN 1BL3 [51] structure and energy-minimized using Maestro [52] with a
RMSD of 0.30 Å. From the resulting complex, the chain C of IBD of LEDGF/p75 was castoff in
order to simulate IN-inhibitor complexes. MD simulations were carried out using the sander module
of AMBER 11 [53] and parm 99.dat and frcmod.ff03 parameter files [54, 55]. These parameters
were assigned to the designed ligands, while partial charges were calculated using the AM1-BCC
method as implemented in the Antechamber suite of AMBER 11. The geometry of the system was
minimized in order to remove any bad contacts using the steepest descent algorithm for the first 250
steps before switching to the conjugate gradient algorithm for the remaining 250 steps. Solvent
effects were taken into account by using the generalized Born implicit solvent model. The
9

ACCEPTED MANUSCRIPT
minimized structure was the input for MD runs using constant-temperature Langevin dynamics at
300 K for 100 ps with a time step of 1fs and a distance cutoff of 12.0 Å for the nonbonded
interactions. Snapshots of the complexes during the simulations and the average structures were
obtained with the Ptraj module of the AMBER 11 suite [53]. The hydrogen bonds were detected
when the acceptor-donor atom distance was lower than 3.5 Å and the acceptor-H-donor angle was
more than 120°. The MM-GBSA method [50] implemented in the AMBER program was used to
evaluate the ligand-protein interaction free energies of the minimized complex and the 100
snapshots extracted at 1 ps intervals. For MM-GBSA analysis, snapshots at 40 ps intervals were
extracted from the last 4 ns of the MD trajectory, and the binding free energies were averaged over
the ensemble of conformers produced (100 snapshots for each trajectory).
4.3 Chemistry
All starting materials and reagents commercially available (Sigma-Aldrich and Alfa Aesar) were
used without further purification. Anhydrous solvents CH3OH, DMSO, CH2Cl2, THF and DMF
were used directly from their Sure-Seal bottles and were purchased from Sigma Aldrich. 4Å
molecular sieves also purchased from Sigma Aldrich, were activated by heating to 300°C under
vacuum overnight. All reactions were performed in oven-dried glassware and were monitored for
completeness by thin layer chromatography (TLC) using silica gel then visualized by UV light, or
1
developed by treatment with KMnO4 stain. A 300MHz Bruker NMR was utilized to obtain H and
13C NMR spectra in CDCl3 or DMSO-d6. All NMR chemical shifts (δ) are reported in parts per
million after calibrations to residual isotopic solvent and coupling constants (J) are reported in Hz.
All exchangeable protons were confirmed by addition of D2O. Melting points were determined on a
BUCHI Melting Point B-545 apparatus and are uncorrected. Mass spectrometry analysis were
realized on Bruker MicrOTOF (ESI) equipped with and Agilent 1200 LC.
Procedure for the synthesis of 5-amino-2-hydroxy-benzoic acid (12)
A mixture of 2-hydroxy-5-nitrobenzoic acid (1 mmol, 183 mg) and conc. hydrochloric acid (18 ml)
was placed in an ice bath water. Zinc powder (3 mmol, 196 mg) was added dropwise through
condenser and the reaction was refluxed for 4 h. Then the mixture was cooled to room temperature,
diluited with water and washed with ethyl acetate (3 x 25ml). The obtained organic solution was
evaporated in vacuum to give the crude product. The residue was purified by crystallization with
1
diethyl ether. Yield: 88%, Mp: 280-282°C. HNMR, (DMSO-d6): δ = 6.70 (dd, J=8.8, J=1.8, 1H,
10

ACCEPTED MANUSCRIPT
ArH), 6.92 (dd, J=8.8, J=2.9, 1H, ArH), 7.24 (s, 1H, ArH).Anal. Calcd for C7H7NO3: C(54.90%)
H(4.61%) N(9.15%). Found: C: 54.80; H: 4.50, N: 9.10.
Procedure for the synthesis of methyl 5-amino-2-hydroxybenzoate (13)
To a solution of 5-amino-2-hydroxybenzoic acid (20 mmol, 3.062 g) in methanol (40 mL) sulfuric
acid (4.5 mL) was added at 0°C. The reaction mixture was refluxed at 80 °C for 24 h. The solvent
was then removed, the reaction mixture was neutralized with saturated sodium bicarbonate aqueous
solution until pH =7. The product was extracted with ethyl acetate (3 x 25ml), the organic layers
were combined, dried over anhydrous Na2SO4 and concentrated to provide the methyl 5-amino-2-
hydroxybenzoate (13) as a pale yellow solid. Yield: 90%, Mp: 97-99 °C. 1HNMR, (DMSO-d6): δ =
3.87 (s, 3H, CH3), 6.92 (d, J=8.2, 1H, ArH), 7.15 (m, 1H, ArH), 7.39 (m, 1H, ArH), 10.18 (bs, 1H,
OH). Anal. Calcd for C8H9NO3 : C(57.48%) H(5.43%) N(8.38%). Found: C: 57.67; H: 5.63, N:
8.34.
General procedure for the synthesis of methyl 5-(benzylamino)-2-hydroxybenzoates (14-19)
To a stirred solution of methyl 5-amino-2-hydroxybenzoate (13) (1.49 mmol, 250 mg) in MeOH
(7ml), over 4A° molecular sieves, benzaldehyde (1.49 mmol, 0.151 ml) and AcOH (0.159 ml) were
added. The mixture was heated at 40°C and agitated for 30 min. Then it was stirred for 1h at room
temperature. The solution was cooled to 5-10°C and NaCNBH3 (1.94 mmol, 121 mg), was slowly
added. The resulted mixture was stirred for 2h at room temperature before being quenched by the
addition of water. The solvent was evaporated under pressure and the crude mixture was taken up in
CH2Cl2, washed with water and brine, dried (Na2SO4) and evaporated. The residue was purified by
chromatography (eluent: Hexane/Ethyl acetate: 8/2) to give intermediates 14-19.
1
Methyl 5-(benzylamino)-2-hydroxybenzoate (14) Yield: 67%, Oil. HNMR, (CDCl3): δ = 3.87 (s,
3H, CH3), 4.25 (s, 2H, CH2), 6.79-6.87 (m, 2H, ArH), 7.09 (s, 1H, ArH), 7.28-7.36 (m, 5H, ArH),
10.24 (bs, 1H, OH). Anal. Calcd for C15H15NO3 : C(70.02%) H(5.88%) N(5.44%). Found: C: 70.12;
H: 5.80, N: 5.54.
1
Methyl 5-[(2-chlorobenzyl)amino]-2-hydroxybenzoate (15) Yield: 66%, Oil. HNMR, (CDCl3): δ =
4.07 (s, 3H, CH3), 4.57 (s, 2H, CH2), 7.03 (s, 2H, ArH), 7.21-7.46 (m, 4H, ArH), 7.55 (m, 1H, ArH),
10.38 (bs, 1H, OH). Anal. Calcd for C15H14ClNO3 : C(61.76%) H(4.84%) N(4.80%). Found: C:
61.76; H: 4.84, N: 4.80.
11

ACCEPTED MANUSCRIPT
1
Methyl 5-[(3-chlorobenzyl)amino]-2-hydroxybenzoate (16) Yield: 62%, Oil. HNMR, (CDCl3): δ =
3.93 (s, 3H, CH3), 4.28 (s, 2H, CH2), 6.81-6.89 (m, 2H, ArH), 7.07 (s, 1H, ArH), 7.25-7.29 (m, 3H,
ArH), 7.38 (s, 1H, ArH), 10.21 (bs, 1H, OH). Anal. Calcd for C15H14ClNO3 : C(61.76%) H(4.84%)
N(4.80%). Found: C: 61.59; H: 4.73, N: 4.68.
1
Methyl 5-[(2-fluorobenzyl)amino]-2-hydroxybenzoate (17) Yield: 65%, Oil. HNMR, (CDCl3): δ =
4.08 (s, 3H, CH3), 4.52 (s, 2H, CH2), 7.02 (s, 2H, ArH), 7.20-7.29 (m, 3H, ArH) 7.39-7.55 (m, 2H,
ArH), 10.36 (bs, 1H, OH). Anal. Calcd for C15H14FNO3: C(65.45%) H(5.13%) N(5.09%). Found:
C: 65.37; H: 5.01, N: 4.97.
1
Methyl 5-[(3-fluorobenzyl)amino]-2-hydroxybenzoate (18) Yield: 62%, Oil. HNMR, (CDCl3): δ =
3.91 (s, 3H, CH3), 4.27 (s, 2H, CH2), 6.80-6.87 (m, 2H, ArH), 7.06 (s, 1H, ArH), 7.22-7.41 (m, 4H,
ArH), 10.19 (s, 1H, OH). Anal. Calcd for C15H14FNO3: C(65.45%) H(5.13%) N(5.09%). Found: C:
65.37; H: 5.01, N: 4.97.
Methyl 2-hydroxy-5-[(2-methylbenzyl)amino]benzoate (19) Yield: 63%, Oil. 1HNMR, (CDCl3): δ =
2.30 (s, 3H, CH3), 3.84 (s, 3H, CH3), 4.50 (s, 2H, CH2), 6.80-6.86 (m, 2H, ArH), 7.05-7.14 (m, 2H,
ArH), 7.30-7.36 (m, 3H, ArH), 10.21 (bs, 1H, OH). Anal. Calcd for C16H17NO3 : C(70.83%)
H(6.32%) N(5.16%). Found: C: 70.68; H: 6.21, N: 5.29.
General procedure for the synthesis of methyl 5-(dibenzylamino)-2-hydroxybenzoates (20-25)
The appropriate methyl 5-(benzylamino)-2-hydroxybenzoate (14-19) (0.3 mmol) was dissolved in
anhydrous DMF and was added under argon to a suspension of NaH (0.3 mmol, 29 mg) in
anhydrous DMF. The reaction mixture was stirred at 0°C under for few minutes, followed by the
addition of a solution of suitable benzyl bromide (0.4 mmol) in anhydrous DMF. After 3 h, the
reaction mixture was extracted by ethyl acetate (3 x 30ml) and washed with water (2 x 20ml) then
with brine (2 x 20ml). The organic layer was dried (Na2SO4), filtered and the solvent removed
under reduced pressure. The resultant product was purified via flash chromatography (Hexane/Ethyl
acetate: 98/2) affording the title compounds 20-25.
Methyl 5-(dibenzylamino)-2-hydroxybenzoate (20) Yield: 65%, Mp: 159-161 °C. 1HNMR, (CDCl3):
δ=3.68 (s, 3H, CH3), 4.53 (s, 4H, CH2), 6.85 (d, J=9.0, 1H, ArH), 7.00 (dd, J=9.0, J=3.2, 1H, ArH),
7.24-7.36 (m, 11H, ArH), 10.18 (bs, 1H, OH). Anal. Calcd for C22H21NO3: C(76.06%) H(6.09%)
N(4.03%). Found: C: 75.09; H: 6.06, N: 4.13.
Methyl 5-[bis(2-chlorobenzyl)amino]-2-hydroxybenzoate (21) Yield: 62%, Mp: 175-177 °C.
1HNMR, (CDCl3): δ = 3.87 (s, 3H, CH3), 4.66 (s, 4H, CH2), 6.83 (m, 2H, ArH), 7.09 (m, 1H, ArH),
12

ACCEPTED MANUSCRIPT
7.20-7.28 (m, 6H, ArH), 7.38-7.43 (m, 2H, ArH), 10.17 (bs, 1H, OH). Anal. Calcd for
C22H19Cl2NO3: C(63.47%) H(4.60%) N(3.36%). Found: C: 63.28; H: 4.54, N: 3.52.
Methyl 5-[bis(3-chlorobenzyl)amino]-2-hydroxybenzoate (22) Yield: 60%, Mp: 179-181 °C.
1HNMR, (CDCl3): δ = 3.93 (s, 3H, CH3), 4.50 (s, 4H, CH2), 6.89 (d, J=9.1, 1H, ArH), 7.00 (dd,
J=9.1, J=3.0, 1H, ArH), 7.13-7.18 (m, 2H, ArH), 7.24-7.32 (m, 7H, ArH), 10.25 (bs, 1H, OH).
Anal. Calcd for C22H19Cl2NO3: C(63.47%) H(4.60%) N(3.36%). Found: C: 63. 57; H: 4.50, N:
3.12.
Methyl 5-[bis(2-fluorobenzyl)amino)-2-hydroxybenzoate (23) Yield: 61%, Mp: 177-179 °C.
1HNMR, (CDCl3): δ = 3.92 (s, 3H, CH3), 4.63 (s, 4H, CH2), 6.87 (d, J=9.2, 1H, ArH), 7.00 (dd,
J=9.2, J=3.2, 1H, ArH), 7.05-7.13 (m, 5H, ArH), 7.24-7.30 (m, 2H, ArH), 7.28-7.30 (m, 2H,
ArH),10.22 (bs, 1H, OH). Anal. Calcd for C22H19F2NO3: C(68.92%) H(5.00%) N(3.65%). Found:
C: 68.74; H: 4.93, N: 3.32.
Methyl 5-[bis(3-fluorobenzyl)amino]-2-hydroxybenzoate (24) Yield: 58%, Mp: 172-174 °C.
1HNMR, (CDCl3): δ = 3.92 (s, 3H, CH3), 4.51 (s, 4H, CH2), 6.89 (d, J=9.1, 1H, ArH), 7.00 (dd,
J=9.1, J=3.0, 1H, ArH), 7.14-7.17 (m, 2H, ArH), 7.25-7.32 (m, 7H, ArH), 10.27 (bs, 1H, OH).
Anal. Calcd for C22H19F2NO3: C(68.92%) H(5.00%) N(3.65%). Found: C: 68.79; H: 4.98, N: 3.45.
Methyl 5-[bis(2-methylbenzyl)amino]-2-hydroxybenzoate (25) Yield: 61%, Mp: 164-166 °C.
1HNMR, (CDCl3): δ = 2.37 (s, 6H, CH3), 3.95 (s, 3H, CH3), 4.60 (s, 4H, CH2), 6.91 (s, 1H, ArH),
6.92 (s, 1H, ArH), 7.19-7.37 (m, 9H, ArH), 10.27 (bs, 1H, OH). Anal. Calcd for C24H25NO3 :
C(76.77%) H(6.71%) N(3.73%). Found: C: 76.65; H: 6.68, N: 3.89.
General procedure for the synthesis of 5-(dibenzylamino)-2-hydroxybenzoic acids (1-3, 5-6 and 8)
The appropriate methyl-2-hydroxybenzoate (20-25) was dissolved in a mixture of THF (2 ml) and
MeOH (2 ml), and 2M NaOH (4 ml) was added. The reaction was refluxed for 24h. After this time
the solvent was removed under reduced pressure and the residue was acidified to pH 2 with 2N
HCl. The solution was extracted with ethyl acetate (3 x 10ml), dried over Na2SO4, filtered and the
solvent removed under reduced pressure. The final products (1-3, 5-6 and 8) were crystallized with
a mixture of Hexane/Ethyl acetate/Ethanol (1/1/1) and some drops of methanol.
1
5-(Dibenzylamino)-2-hydroxybenzoic acid (1) Yield: 59%, Mp: 140-142 °C. HNMR, (DMSO-d6):
δ = 4.60 (s, 4H, CH2), 6.76 (d, J = 9, 1H, ArH), 7.00 (dd, J = 9, J = 3.2, 1H, ArH), 7.05 (s, 1H,
ArH), 7.23-7.35 (m, 10H, ArH), 10.53 (bs, 1H, OH), 13.65 (bs, 1H, OH).Anal. Calcd for
C21H19NO3: C(75.66%) H(5.74%) N(4.20%). Found: C: 75.54; H: 5.60, N: 4.15. ESI(+), CH3OH,
HR-MS : ion [M+H]+ , m/z 333, C21H19NO3, m/z theory 333,1365, m/z found 334,14326.
13

ACCEPTED MANUSCRIPT
1
5-[Bis(2-chlorobenzyl)amino]-2-hydroxybenzoic acid (2) Yield: 53%, Mp: 148-150 °C. HNMR,
(DMSO-d6): δ = 4.68 (s, 4H, CH2), 6.75-6.84 (m, 2H, ArH), 6.94 (s, 1H, ArH),7.23-7.31 (m, 6H,
ArH), 7.47 (m, 2H, ArH). Anal. Calcd for C21H17Cl2NO3: C(62.70%) H(4.26%) N(3.48%). Found:
C: 62.58; H: 4.35, N: 3.35. ESI(+), CH3OH, HR-MS : ion [M+H]+ , m/z 401, C21H17Cl2NO3, m/z
theory 401,0585, m/z found 402,0651.
1
5-[Bis(3-chlorobenzyl)amino]-2-hydroxybenzoic acid (3) Yield: 50%, Mp: 147-149 °C. HNMR,
(DMSO-d6): δ = 4.64 (s, 4H, CH2), 6.78 (d, J = 8.9, 1H, ArH), 6.97- 7.05 (m, 2H, ArH), 7.21-7.39
(m, 8H, ArH), 10.55 (bs, 1H, OH). Anal. Calcd for C21H17Cl2NO3: C(62.70%) H(4.26%) N(3.48%).
Found: C: 62.86; H: 4.10, N: 3.25. ESI(+), CH3OH, HR-MS : ion [M+H]+ , m/z 401,
C21H17Cl2NO3, m/z theory 401,0585, m/z found 402,0651.
1
5-[Bis(2-fluorobenzyl)amino]-2-hydroxybenzoic acid (5) Yield: 52%, Mp: 137-139 °C, HNMR,
(DMSO-d6): δ = 4.65 (s, 4H, CH2), 6.79 (d, J = 9.1, 1H, ArH), 6.99 (dd, J = 9.1, J = 3.2, 1H, ArH),
7.06-7.34 (m, 9H, ArH), 10.59 (bs, 1H, OH), 13.69 (bs, 1H, OH). Anal. Calcd for C21H17F2NO3 :
C(68.29%) H(4.64%) N(3.79%). Found: C: 68.07; H: 4.46, N: 3.58. ESI(+), CH3OH, HR-MS : ion
[M+H]+ , m/z 369, C21H17F2NO3, m/z theory 369,3678, m/z found 370,1241.
1
5-[Bis(3-fluorobenzyl)amino]-2-hydroxybenzoic acid (6) Yield: 49%, Mp: 141-143 °C HNMR,
(DMSO-d6): δ = 4.63 (s, 4H, CH2), 6.78 (d, J = 8.8, 1H, ArH), 6.99 (dd, J = 8.8, J = 3.2, 1H, ArH),
7.04 (d, J = 3.2, 1H, ArH), 7.20-7.38 (m, 8H, ArH), 10.52 (bs, 1H, OH), 13.69 (bs, 1H, OH). Anal.
Calcd for C21H17F2NO3 : C(68.29%) H(4.64%) N(3.79%). Found: C: 68.14; H: 4.59, N: 3.61.
ESI(+), CH3OH, HR-MS : ion [M+H]+, m/z 369, C21H17F2NO3, m/z theory 369,3678, m/z found
370,1249.
1
5-[Bis(2-methylbenzyl)amino]-2-hydroxybenzoic acid (8) Yield: 48%, Mp: 136-138 °C. HNMR,
(CDCl3): δ= 2.28 (s, 6H, CH3), 4.53 (s, 4H, CH2), 6.83 (d, J = 9.2, 1H, ArH), 7.85 (m, 1H, ArH)
7.14-7.20 (m, 9H, ArH), 9.79 (bs, 1H, OH). Anal. Calcd for C23H23NO3: C(76.43%) H(6.41%)
N(3.88%). Found: C: 76.26; H: 6.30, N: 3.95. ESI(+), CH3OH, HR-MS : ion [M+H]+ , m/z 361,
C23H23NO3, m/z theory 361,1678, m/z found 362,1745.
4.4 Recombinant proteins and HTRF-based Assays
His-tagged LEDGF/p75, FLAG-tagged IN and His-tagged IN were constructed and purified as
described previously [32]. Homogeneous time resolved fluorescence energy transfer (HTRF) based
LEDGF/p75dependent IN activity, IN-LEDGF/p75 binding, and LEDGF/p75 independent
3’processing assays were carried out as previously described [32]. The HTRF signal was recorded
using a Perkin Elmer Multimode EnSpire plate reader. The fitted dose-response curves were
14

ACCEPTED MANUSCRIPT
generated to calculate IC50 using Origin software (OriginLab, Inc.). All fitted curves displayed a R2
of 0.99 or greater.
4.5 Antiviral Activity and Cytotoxicity Assays
The indicated concentrations of the test compounds or diluent control (DMSO) were added directly
to the target cells and the cells were infected with untreated virions. HeLa TZM-bl cells (2^105
cells/well of a 6-well plate in 2 ml of complete medium) were pre-incubated with the indicated
concentrations of the test inhibitor or diluent control (DMSO) for 2 h. The cells were then infected
with HIV-1 virions equivalent to 4 ng of HIV-1 p24 as determined by HIV-1 Gag p24 ELISA
(ZeptoMetrix) following manufacturer’s protocol. Two hours post-infection the culture supernatant
was removed, washed once with complete medium, and then fresh complete medium was added
with the inhibitor concentration maintained. The cells were cultured for 48 h and the cell extracts
were prepared using 16 reporter lysis buffer (Promega). Luciferase activity was determined using a
commercially available kit (Promega). The cytotoxicity assays were performed as described
previously [28]. The fitted dose-response curves were generated to calculate EC50 or CC50 using
Origin software (OriginLab, Inc.). All fitted curves displayed a R2 of 0.97 or greater.
Acknowledgements
We are grateful to Dr. Ross C. Larue for critical reading of the manuscript and helpful comments.
These studies were in part supported by University of Messina Research&Mobility 2015 Project
(project code RES_AND_MOB_2015_DE_LUCA) and by National Institutes of Health grant
R01AI110310 (to M.K. and J.R.F.)
References
[1] M. Jaskolski, J.N. Alexandratos, G. Bujacz, A. Wlodawer, Piecing together the structure of
retroviral integrase, an important target in AIDS therapy, FEBS J 276 (2009) 2926-2946.
[2] L. Feng, R.C. Larue, A. Slaughter, J.J. Kessl, M. Kvaratskhelia, HIV-1 integrase
multimerization as a therapeutic target, Curr Top Microbiol Immunol 389 (2015) 93-119.
[3] T. Wills, V. Vega, Elvitegravir: a once-daily inhibitor of HIV-1 integrase, Expert Opin Investig
Drugs 21 (2012) 395-401.
[4] FDA approves raltegravir tablets, AIDS Patient Care STDS 21 (2007) 889.
[5] S.L. Karmon, M. Markowitz, Next-generation integrase inhibitors : where to after raltegravir?,
Drugs 73 (2013) 213-228.
[6] V. Summa, A. Petrocchi, F. Bonelli, B. Crescenzi, M. Donghi, M. Ferrara, F. Fiore, C. Gardelli,
O. Gonzalez Paz, D.J. Hazuda, P. Jones, O. Kinzel, R. Laufer, E. Monteagudo, E. Muraglia, E. Nizi,
F. Orvieto, P. Pace, G. Pescatore, R. Scarpelli, K. Stillmock, M.V. Witmer, M. Rowley, Discovery
15

ACCEPTED MANUSCRIPT
of raltegravir, a potent, selective orally bioavailable HIV-integrase inhibitor for the treatment of
HIV-AIDS infection, J Med Chem 51 (2008) 5843-5855.
[7] J. van Lunzen, F. Maggiolo, J.R. Arribas, A. Rakhmanova, P. Yeni, B. Young, J.K. Rockstroh,
S. Almond, I. Song, C. Brothers, S. Min, Once daily dolutegravir (S/GSK1349572) in combination
therapy in antiretroviral-naive adults with HIV: planned interim 48 week results from SPRING-1, a
dose-ranging, randomised, phase 2b trial, Lancet Infect Dis 12 (2012) 111-118.
[8] K.K. Pandey, Critical appraisal of elvitegravir in the treatment of HIV-1/AIDS, HIV AIDS
(Auckl) 6 (2014) 81-90.
[9] N. Sichtig, S. Sierra, R. Kaiser, M. Daumer, S. Reuter, E. Schulter, A. Altmann, G. Fatkenheuer,
U. Dittmer, H. Pfister, S. Esser, Evolution of raltegravir resistance during therapy, J Antimicrob
Chemother 64 (2009) 25-32.
[10] R.T. Steigbigel, D.A. Cooper, P.N. Kumar, J.E. Eron, M. Schechter, M. Markowitz, M.R.
Loutfy, J.L. Lennox, J.M. Gatell, J.K. Rockstroh, C. Katlama, P. Yeni, A. Lazzarin, B. Clotet, J.
Zhao, J. Chen, D.M. Ryan, R.R. Rhodes, J.A. Killar, L.R. Gilde, K.M. Strohmaier, A.R. Meibohm,
M.D. Miller, D.J. Hazuda, M.L. Nessly, M.J. DiNubile, R.D. Isaacs, B.Y. Nguyen, H. Teppler, B.S.
Teams, Raltegravir with optimized background therapy for resistant HIV-1 infection, N Engl J Med
359 (2008) 339-354.
[11] M. Metifiot, N. Vandegraaff, K. Maddali, A. Naumova, X. Zhang, D. Rhodes, C. Marchand, Y.
Pommier, Elvitegravir overcomes resistance to raltegravir induced by integrase mutation Y143,
AIDS 25 (2011) 1175-1178.
[12] M. Li, M. Mizuuchi, T.R. Burke, Jr., R. Craigie, Retroviral DNA integration: reaction pathway
and critical intermediates, EMBO J 25 (2006) 1295-1304.
[13] L. Krishnan, A. Engelman, Retroviral integrase proteins and HIV-1 DNA integration, J Biol
Chem 287 (2012) 40858-40866.
[14] A. Engelman, P. Cherepanov, The structural biology of HIV-1: mechanistic and therapeutic
insights, Nat Rev Microbiol 10 (2012) 279-290.
[15] P. Cherepanov, G. Maertens, P. Proost, B. Devreese, J. Van Beeumen, Y. Engelborghs, E. De
Clercq, Z. Debyser, HIV-1 integrase forms stable tetramers and associates with LEDGF/p75 protein
in human cells, J Biol Chem 278 (2003) 372-381.
[16] A. Ciuffi, M. Llano, E. Poeschla, C. Hoffmann, J. Leipzig, P. Shinn, J.R. Ecker, F. Bushman,
A role for LEDGF/p75 in targeting HIV DNA integration, Nat Med 11 (2005) 1287-1289.
[17] M. Llano, D.T. Saenz, A. Meehan, P. Wongthida, M. Peretz, W.H. Walker, W. Teo, E.M.
Poeschla, An essential role for LEDGF/p75 in HIV integration, Science 314 (2006) 461-464.
[18] M.C. Shun, N.K. Raghavendra, N. Vandegraaff, J.E. Daigle, S. Hughes, P. Kellam, P.
Cherepanov, A. Engelman, LEDGF/p75 functions downstream from preintegration complex
formation to effect gene-specific HIV-1 integration, Genes Dev 21 (2007) 1767-1778.
[19] K. Busschots, J. Vercammen, S. Emiliani, R. Benarous, Y. Engelborghs, F. Christ, Z. Debyser,
The interaction of LEDGF/p75 with integrase is lentivirus-specific and promotes DNA binding, J
Biol Chem 280 (2005) 17841-17847.
[20] M. Llano, J. Morrison, E.M. Poeschla, Virological and cellular roles of the transcriptional
coactivator LEDGF/p75, Curr Top Microbiol Immunol 339 (2009) 125-146.
[21] L.Q. Al-Mawsawi, N. Neamati, Blocking interactions between HIV-1 integrase and cellular
cofactors: an emerging anti-retroviral strategy, Trends Pharmacol Sci 28 (2007) 526-535.
[22] A. Engelman, P. Cherepanov, The lentiviral integrase binding protein LEDGF/p75 and HIV-1
replication, PLoS Pathog 4 (2008) e1000046.
[23] P. Cherepanov, E. Devroe, P.A. Silver, A. Engelman, Identification of an evolutionarily
conserved domain in human lens epithelium-derived growth factor/transcriptional co-activator p75
(LEDGF/p75) that binds HIV-1 integrase, J Biol Chem 279 (2004) 48883-48892.
16

ACCEPTED MANUSCRIPT
[24] P. Cherepanov, A.L. Ambrosio, S. Rahman, T. Ellenberger, A. Engelman, Structural basis for
the recognition between HIV-1 integrase and transcriptional coactivator p75, Proc. Natl. Acad. Sci.
U S A 102 (2005) 17308-17313.
[25] F. Christ, A. Voet, A. Marchand, S. Nicolet, B.A. Desimmie, D. Marchand, D. Bardiot, N.J.
Van der Veken, B. Van Remoortel, S.V. Strelkov, M. De Maeyer, P. Chaltin, Z. Debyser, Rational
design of small-molecule inhibitors of the LEDGF/p75-integrase interaction and HIV replication,
Nat Chem Biol 6 (2010) 442-448.
[26] L.D. Fader, E. Malenfant, M. Parisien, R. Carson, F. Bilodeau, S. Landry, M. Pesant, C.
Brochu, S. Morin, C. Chabot, T. Halmos, Y. Bousquet, M.D. Bailey, S.H. Kawai, R. Coulombe, S.
LaPlante, A. Jakalian, P.K. Bhardwaj, D. Wernic, P. Schroeder, M. Amad, P. Edwards, M. Garneau,
J. Duan, M. Cordingley, R. Bethell, S.W. Mason, M. Bos, P. Bonneau, M.A. Poupart, A.M.
Faucher, B. Simoneau, C. Fenwick, C. Yoakim, Y. Tsantrizos, Discovery of BI 224436, a
Noncatalytic Site Integrase Inhibitor (NCINI) of HIV-1, ACS Med Chem Lett 5 (2014) 422-427.
[27] M. Tsiang, G.S. Jones, A. Niedziela-Majka, E. Kan, E.B. Lansdon, W. Huang, M. Hung, D.
Samuel, N. Novikov, Y. Xu, M. Mitchell, H. Guo, K. Babaoglu, X. Liu, R. Geleziunas, R.
Sakowicz, New class of HIV-1 integrase (IN) inhibitors with a dual mode of action, J Biol Chem
287 (2012) 21189-21203.
[28] A. Sharma, A. Slaughter, N. Jena, L. Feng, J.J. Kessl, H.J. Fadel, N. Malani, F. Male, L. Wu,
E. Poeschla, F.D. Bushman, J.R. Fuchs, M. Kvaratskhelia, A New Class of Multimerization
Selective Inhibitors of HIV-1 Integrase, PLoS Pathog 10 (2014) e1004171.
[29] E. Le Rouzic, D. Bonnard, S. Chasset, J.M. Bruneau, F. Chevreuil, F. Le Strat, J. Nguyen, R.
Beauvoir, C. Amadori, J. Brias, S. Vomscheid, S. Eiler, N. Levy, O. Delelis, E. Deprez, A. Saib, A.
Zamborlini, S. Emiliani, M. Ruff, B. Ledoussal, F. Moreau, R. Benarous, Dual inhibition of HIV-1
replication by integrase-LEDGF allosteric inhibitors is predominant at the post-integration stage,
Retrovirology 10 (2013) 144.
[30] N. van Bel, Y. van der Velden, D. Bonnard, E. Le Rouzic, A.T. Das, R. Benarous, B.
Berkhout, The allosteric HIV-1 integrase inhibitor BI-D affects virion maturation but does not
influence packaging of a functional RNA genome, PLoS One 9 (2014) e103552.
[31] P. Cherepanov, A.L. Ambrosio, S. Rahman, T. Ellenberger, A. Engelman, Structural basis for
the recognition between HIV-1 integrase and transcriptional coactivator p75, Proc Natl Acad Sci U
S A 102 (2005) 17308-17313.
[32] J.J. Kessl, N. Jena, Y. Koh, H. Taskent-Sezgin, A. Slaughter, L. Feng, S. de Silva, L. Wu, S.F.
Le Grice, A. Engelman, J.R. Fuchs, M. Kvaratskhelia, A multimode, cooperative mechanism of
action of allosteric HIV-1 integrase inhibitors, J Biol Chem 287 (2012) 16801-16811.
[33] K.A. Jurado, A. Engelman, Multimodal mechanism of action of allosteric HIV-1 integrase
inhibitors, Expert Rev Mol Med 15 (2013) e14.
[34] K.A. Jurado, H. Wang, A. Slaughter, L. Feng, J.J. Kessl, Y. Koh, W. Wang, A. Ballandras-
Colas, P.A. Patel, J.R. Fuchs, M. Kvaratskhelia, A. Engelman, Allosteric integrase inhibitor potency
is determined through the inhibition of HIV-1 particle maturation, Proc Natl Acad Sci U S A 110
(2013) 8690-8695.
[35] M. Balakrishnan, S.R. Yant, L. Tsai, C. O'Sullivan, R.A. Bam, A. Tsai, A. Niedziela-Majka,
K.M. Stray, R. Sakowicz, T. Cihlar, Non-Catalytic Site HIV-1 Integrase Inhibitors Disrupt Core
Maturation and Induce a Reverse Transcription Block in Target Cells, PLoS One 8 (2013) e74163.
[36] B.A. Desimmie, R. Schrijvers, J. Demeulemeester, D. Borrenberghs, C. Weydert, W. Thys, S.
Vets, B. Van Remoortel, J. Hofkens, J. De Rijck, J. Hendrix, N. Bannert, R. Gijsbers, F. Christ, Z.
Debyser, LEDGINs inhibit late stage HIV-1 replication by modulating integrase multimerization in
the virions, Retrovirology 10 (2013) 57.
17

ACCEPTED MANUSCRIPT
[37] J. Fontana, K.A. Jurado, N. Cheng, N.L. Ly, J.R. Fuchs, R.J. Gorelick, A.N. Engelman, A.C.
Steven, Distribution and Redistribution of HIV-1 Nucleocapsid Protein in Immature, Mature, and
Integrase-Inhibited Virions: a Role for Integrase in Maturation, J Virol 89 (2015) 9765-9780.
[38] L. Feng, V. Dharmarajan, E. Serrao, A. Hoyte, R.C. Larue, A. Slaughter, A. Sharma, M.R.
Plumb, J.J. Kessl, J.R. Fuchs, F.D. Bushman, A.N. Engelman, P.R. Griffin, M. Kvaratskhelia, The
Competitive Interplay between Allosteric HIV-1 Integrase Inhibitor BI/D and LEDGF/p75 during
the Early Stage of HIV-1 Replication Adversely Affects Inhibitor Potency, ACS Chem Biol (2016).
[39] L. Feng, A. Sharma, A. Slaughter, N. Jena, Y. Koh, N. Shkriabai, R.C. Larue, P.A. Patel, H.
Mitsuya, J.J. Kessl, A. Engelman, J.R. Fuchs, M. Kvaratskhelia, The A128T resistance mutation
reveals aberrant protein multimerization as the primary mechanism of action of allosteric HIV-1
integrase inhibitors, J Biol Chem 288 (2013) 15813-15820.
[40] A. Slaughter, K.A. Jurado, N. Deng, L. Feng, J.J. Kessl, N. Shkriabai, R.C. Larue, H.J. Fadel,
P.A. Patel, N. Jena, J.R. Fuchs, E. Poeschla, R.M. Levy, A. Engelman, M. Kvaratskhelia, The
mechanism of H171T resistance reveals the importance of N inverted question mark -protonated
His171 for the binding of allosteric inhibitor BI-D to HIV-1 integrase, Retrovirology 11 (2014) 100.
[41] D.J. Newman, G.M. Cragg, Natural products as sources of new drugs over the 30 years from
1981 to 2010, J Nat Prod 75 (2012) 311-335.
[42] Y.W. Chin, M.J. Balunas, H.B. Chai, A.D. Kinghorn, Drug discovery from natural sources,
AAPS J 8 (2006) E239-253.
[43] L. De Luca, F. Morreale, F. Christ, Z. Debyser, S. Ferro, R. Gitto, New scaffolds of natural
origin as Integrase-LEDGF/p75 interaction inhibitors: virtual screening and activity assays, Eur J
Med Chem 68 (2013) 405-411.
[44] L. De Luca, S. Ferro, F. Morreale, F. Christ, Z. Debyser, A. Chimirri, R. Gitto, Fragment
hopping approach directed at design of HIV IN-LEDGF/p75 interaction inhibitors, J Enzyme Inhib
Med Chem 28 (2013) 1002-1009.
[45] S. Ferro, L. De Luca, F.E. Agharbaoui, F. Christ, Z. Debyser, R. Gitto, Optimization of
rhodanine scaffold for the development of protein-protein interaction inhibitors, Bioorg Med Chem
23 (2015) 3208-3214.
[46] G. Jones, P. Willett, R.C. Glen, A.R. Leach, R. Taylor, Development and validation of a
genetic algorithm for flexible docking, J Mol Biol 267 (1997) 727-748.
[47] D.A. Case, T.A. Darden, T.E. Cheatham, III, C.L. Simmerling, J. Wang, R.E. Duke, R. Luo,
R.C. Walker, W. Zhang, K.M. Merz, B. Roberts, B. Wang, S. Hayik, A. Roitberg, G. Seabra, I.
Kolossváry, K.F. Wong, F. Paesani, J. Vanicek, J. Liu, X. Wu, S.R. Brozell, T. Steinbrecher, H.
Gohlke, Q. Cai, X. Ye, J. Wang, M.-J. Hsieh, G. Cui, D.R. Roe, D.H. Mathews, M.G. Seetin, C.
Sagui, V. Babin, T. Luchko, S. Gusarov, A. Kovalenko, P.A. Kollman, AMBER 11, in, University
of California, San Francisco, 2010.
[48] L. De Luca, F. Morreale, A. Chimirri, Insight into the fundamental interactions between
LEDGF binding site inhibitors and integrase combining docking and molecular dynamics
simulations, J Chem Inf Model 52 (2012) 3245-3254.
[49] D.S. Accelrys.
[50] P.A. Kollman, I. Massova, C. Reyes, B. Kuhn, S. Huo, L. Chong, M. Lee, T. Lee, Y. Duan, W.
Wang, O. Donini, P. Cieplak, J. Srinivasan, D.A. Case, T.E. Cheatham, 3rd, Calculating structures
and free energies of complex molecules: combining molecular mechanics and continuum models,
Acc Chem Res 33 (2000) 889-897.
[51] S. Maignan, J.P. Guilloteau, Q. Zhou-Liu, C. Clement-Mella, V. Mikol, 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 282 (1998) 359-368.
[52] v.S.L.N.Y. Maestro, NY, 2009.
18

ACCEPTED MANUSCRIPT
[53] D.A.D. Case, T. A.; Cheatham, T. E.; III; Simmerling, C. L.; Wang, J.; Duke, R. E.; Luo, R.;
Walker, R. C.; Zhang, W.; Merz, K. M.; Roberts, B.; Wang, B.; Hayik, S.; Roitberg, A.; Seabra, G.;
Kolossváry, I.; Wong, K. F.; Paesani, F.; Vanicek, J.; Liu, J.; Wu, X.; Brozell, S. R.; Steinbrecher,
T.; Gohlke, H.; Cai, Q.; Ye, X.; Wang, J.; Hsieh, M.-J.; Cui, G.; Roe, D. R.; Mathews, D. H.;
Seetin, M. G.; Sagui, C.; Babin, V.; Luchko, T.; Gusarov, S.; Kovalenko, A.; Kollman, P. A.;
University of California: San Francisco, 2010.
[54] Y. Duan, C. Wu, S. Chowdhury, M.C. Lee, G. Xiong, W. Zhang, R. Yang, P. Cieplak, R. Luo,
T. Lee, J. Caldwell, J. Wang, P. Kollman, A point-charge force field for molecular mechanics
simulations of proteins based on condensed-phase quantum mechanical calculations, J Comput
Chem 24 (2003) 1999-2012.
[55] J. Wang, R.M. Wolf, J.W. Caldwell, P.A. Kollman, D.A. Case, Development and testing of a
general amber force field, J Comput Chem 25 (2004) 1157-1174.
[56] DeLano, W.L., DeLano Scientific LLC, The PyMOL Molecular Graphics System, San Carlos,
CA, USA, 2008.
19

ACCEPTED MANUSCRIPT
Table 1. Binding free energy estimation
∆G (complex 1) ∆G (complex 2) ∆G (Snapshots)
bind
bind
bind
Compound
R
Kcal/mol
Kcal/mol
2B4J-Ligand
2B4J-Ligand
3LPU-Ligand
Kcal/mol
-18,09
-22,41
-24,99
-21,68
-17,75
-22,47
Lavendustin B
-15,39
-21,54
-25,73
-24,74
-21,33
-20,45
-14,77
-19,41
-25,73
-24,41
-19,02
-20,64
H
1
2
3
4
5
2-Cl
3-Cl
4-Cl
2-F
3-F
4-F
2-CH
-20,10
-15,87
-22,24
-20,51
-16,00
-23,35
-20,08
-15,54
-21,40
6
7
8
3
3
3
3-CH
4-CH
-16,29
-12,86
-15,09
-11,38
-14,00
-12,65
9
10
∆G
is the calculated binding free energy
bind
20

ACCEPTED MANUSCRIPT
Table 2. Hydrogen bonds analysis from the results of MD simulation for IN in complex with the
designed compounds
Compound
IN in complex with
Lavendustin B
Donor
Acceptor Occupancy (%)a Distance (Å)b
Thr174(A) OH
Glu170(A) NH
Glu170(A) NH
His171(A) NH
Thr174(A) OH
Thr174(A) OH
Glu170(A) NH
Glu170(A) NH
His171(A) NH
Gln95(A) NH
Thr174(A) OH
Thr174(A) OH
Glu170(A) NH
Glu170(A) NH
His171(A) NH
Gln95(A) NH
Thr174(A) OH
Glu170(A) NH
Glu170(A) NH
His171(A) NH
Thr174(A) OH
Glu170(A) NH
His171(A) NH
His171(A) NH
Thr174(A) OH
Glu170(A) NH
Glu170(A) NH
His171(A) NH
Thr174(A) OH
Thr174(A) OH
Glu170(A) NH
His171(A) NH
His171(A) NH
Thr174(A) OH
Glu170(A) NH
His171(A) NH
Thr174(A) OH
Thr174(A) OH
Glu170(A) NH
Glu170(A) NH
His171(A) NH
Glu170(A) NH
His171(A) NH
Thr174(A) OH
Thr174(A) OH
O3
O1
O2
O2
O
11.80
78.20
36.60
63.30
21.90
10.40
90.40
64.30
76.70
10.40
76.60
34.90
98.80
78.80
84.30
37.20
72.30
86.40
61.20
77.70
21.30
63.60
25.60
24.50
24.30
95.60
76.70
82.10
78.40
17.90
57.40
21.20
10.40
45.50
32.40
23.70
54.80
17.40
96.80
49.00
46.00
21.60
10.30
28.57
23.81
3.092 ( 0.21)
2.930 ( 0.18)
3.051 ( 0.22)
3.230 ( 0.18)
3.077 ( 0.21)
3.093 ( 0.26)
2.923 ( 0.17)
3.108 ( 0.22)
3.102 ( 0.19)
3.140 ( 0.18)
3. 029 ( 0.20)
2.896 ( 0.25)
2.966 ( 0.18)
3.070 ( 0.20)
3.086 ( 0.19)
3.102 ( 0.19)
3.024 ( 0.20)
2.907 ( 0.16)
3.083 ( 0.20)
3.073 ( 0.18)
2.958 ( 0.26)
2.944 ( 0.17)
3.206 ( 0.18)
3.168 ( 0.20)
3.043 ( 0.27)
2.960 ( 0.18)
3.060 ( 0.20)
3.093 ( 0.18)
2.999 ( 0.26)
3.093 ( 0.18)
2.933 ( 0.16)
3.173 ( 0.18)
3.267 ( 0.17)
2.978 ( 0.27)
2.970 ( 0.20)
3.076 ( 0.19)
2.950 ( 0.26)
3.164 ( 0.19)
2.866 ( 0.14)
3.173 ( 0.21)
3.210 ( 0.18)
2.970 ( 0.19)
3.243 ( 0.17)
3.206 ( 0.19)
3.316 ( 0.19)
Lavendustine B
O1
O2
O1
O1
O
1
2
H
O
O1
O2
O1
O1
O
O2
O1
O
2Cl
3
4
5
3Cl
4Cl
2F
O
O1
O1
O1
O
O2
O1
O2
O1
O1
O
O1
O1
O
O
O
O1
O2
O
O1
O2
O2
O
O1
O
O1
6
7
8
3F
4F
2CH
3
3CH
9
3
4CH
10
3
21

ACCEPTED MANUSCRIPT
Glu170(A) NH
His171(A) NH
His171(A) NH
O1
O
O1
50.00
28.57
26.19
2.884 ( 0.16)
3.065 ( 0.18)
3.273( 0.14)
a The listed donor and acceptor pairs correspond to the hydrogen bonds occupancies during the simulation.
b The average distance with standard error (SE =standard deviation/N1/2) in parentheses between hydrogen-
acceptor atom and hydrogen-donor atom in the investigated time period.
22

ACCEPTED MANUSCRIPT
Table 3: Inhibition of IN-LEDGF/p75 binding, LEDGF/p75 dependent integration, and LEDGF/p75
independent 3’-processing activity of the synthetized compounds
LEDGF/p75
IN-LEDGF/p75
Compound
R
Dependent IN
Activity (IC50µM)
7.93 ± 0.79
3’P (IC50 µM)
Binding (IC50 µM)
8.75 ± 2.36
3.28 ± 0.80
19.50 ± 2.55
17.84 ± 4.25
20.83 ± 2.46
27.59 ± 3.98
94.07
69.97 ±10.27
25.62±1.88
>70
1
H
3.78 ± 0.35
13.48 ± 4.45
9.00 ± 0.24
18.50 ± 2.19
14.66 ± 0.69
-
2
2-Cl
3-Cl
2-F
3
63.05 ±5.16
68.56
5
6
3-F
>70
8
2-CH3
a Lavendustin B
-
Data for IC50 are given as the mean ± SD from at least three independent experiments.
a The IC50 of Lavendustin B is reported elsewhere and was evaluated by alpha screen assays [43].
23

ACCEPTED MANUSCRIPT
Figure Captions
Figure 1: A) Chemical structures of 5-[bis(2-hydroxybenzy l)amino]-2-hydroxybenzoic acid (Lavendustin
B) B) Superimposition of Lavendustin B (green) and compound KF115 (cyan) in complex with IN CCD
(PDB code 2B4J). Key residues of the pocket are presented. The figure was created using PyMOL [56]. C)
Chemical structures of KF115. D) Substitutions on the parent molecule.
Figure 2: Superimposition of IN CCD with A) Lavendustin B (violet) and compound 1 (cyan); B)
compounds 2 (green), 3 (cyan) and 4 (magenta); C) compounds 5 (green), 6 (cyan) and 7 (magenta); D)
compounds 8 (green), 9 (cyan) and 10 (magenta). Key residues of the pocket are presented. The figure was
created using PyMOL [56].
Figure 3: A) chemical structure of compound 2; B) Curve fitting of inhibition of LEDGF/p75 dependent IN
activity by compound 2 (black squares); C) Curve fitting of dose-dependent inhibition of IN-LEDGF
binding by compound 2 (black squares); D) Curve fitting of dose-dependent inhibition of 3’processing by
compound 2 (black squares).The average values from three independent experiments are shown for each
assay.
Figure 4: A) Luciferase quantification to measure the inhibition by compound 2 over the indicated
concentrations. VSV-g pseudotyped virus was used to infect HEK293T cells under drug treatment. B)
Cytotoxicity effect of compound 2 over the indicated concentrations without viral infection.
24

ACCEPTED MANUSCRIPT
Scheme 1
Reagents and conditions: (i) Zinc dust, 0-5°C, HClconc., reflux, 4h; (ii) MeOH, H2SO4, 0°C, 80°C, 24h; (iii)
MeOH, AcOH, 30’, 40°C, rt 1h, 5-10 °C NaCNBH3, 2 h rt; (iv) NaH, DMF, Argon 3h rt; (v) NaOH, MeOH,
THF, 60°C, 24h, HCl 2N.
25

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT